This study was performed to synthesize membranes of polyethersulfone (PES) blended with graphene oxide (GO) and PES blended with GO functionalized with photoactive semiconductor catalyst (TiO2 and ZnO). The antifouling and self-cleaning properties of composite membranes were also investigated. The GO was prepared from natural graphite powder by oxidation method at low temperature. TiO2 and ZnO nanopowders were synthesized by anhydrous sol–gel method. The surface of TiO2 and ZnO nanopowders was modified by a surfactant (myristic acid) to obtain a homogeneously dispersed mixture in a solvent, and then GO was functionalized by loading with these metal oxide nanopowders. The PES membranes blended with GO and functionalized GO into the casting solution were prepared via phase inversion method and tested for their antifouling as well as self-cleaning properties. The composite membranes were synthesized as 14%wt. of PES polymer with three different concentrations (0.5, 1.0, and 2.0%wt.) of GO, GO-TiO2, and GO-ZnO. The functionalization of membranes improved hydrophilicity property of membranes as compared to neat PES membrane. However, the lowest flux was obtained by functionalized membranes with GO-TiO2. The results showed that functionalized membranes demonstrated better self-cleaning property than neat PES membrane. Moreover, the flux recovery rate of functionalized membranes over five cycles was higher than that of neat membrane.

INTRODUCTION

Textile processing industry is one of the common sectors in developing countries. The sector uses excessive amounts of water and produces wastewater resulting in pollution loading being too much. The wastewater contains high amount of suspended and dissolved solids, non-reacted dyestuffs (color) and other chemicals that are used in the different stages of dyeing and other processing. The presence of even small amounts of dye in water (e.g. 10–20 mg/L) is highly visible and affects the water transparency (Rajkumar & Kim 2006).

So far, many wastewater treatment technologies have been developed for color removal, including physical, chemical and biological processes. All of them have some advantages and drawbacks (Alver & Metin 2012; Kurt et al. 2012; De Jager et al. 2014; Hayat et al. 2015; Dehghani et al. 2016). The conventional chemical coagulation process generates a large volume of hazardous sludge and has a problem of sludge disposal. However, the biological treatment of textile wastewater shows low degradation efficiency due to presence of biologically inert high molecular weight dyestuffs (Rajkumar et al. 2007). Additionally, a combination of two or more treatment methods for the further treatment was investigated by different researchers in order to satisfy the legal requirements (Doumic et al. 2015; Jung et al. 2015; Punzi et al. 2015).

Recently, membrane technologies have gained significance because of important advantages such as easy operation, small footprint, high quality of effluent water, and cost-effectiveness (Lee et al. 2016a). Thus, membranes have been used intensively for color removal, wastewater treatment, and water recovery in the past decade (Kurt et al. 2012). However, wastewater has organic and inorganic pollutants. These pollutants can cause serious membrane fouling, which raises operating costs, decreases membrane life time, and increases maintenance (Dizge et al. 2011; Jegatheesan et al. 2016). Complicated interactions between membrane surface and dye molecules as well as microbes existing in textile wastewater can cause membrane fouling. This type of fouling can be considered reversible or irreversible and it is very difficult to remove from the membrane surface due to the interactions between membrane polymer chains and dye molecules or microbes (Drews 2010; Chidambaram et al. 2015; Chidambaram & Noel 2015). Therefore, many efforts have been made to prevent membrane fouling, such as modification of membrane materials, improvement of biomass characteristics, and optimization of operating conditions (Chang et al. 2002; Torretta et al. 2013; Zinadini et al. 2014; Zinadini et al. 2015). Among these methods, modification of membrane materials especially blending modification can significantly prevent the membrane fouling (Ozay et al. 2016). Moreover, self-cleaning membrane production with inorganic nanoparticles such as TiO2 and ZnO has been widely studied. Additionally, ZrO2, Al2O3, and SiO2 nanoparticles, among others, were used successfully and decreased the biofouling (Richards et al. 2012; Safarpour et al. 2014; Xu et al. 2014a, 2014b; Xu et al. 2016).

Application of graphene-based materials has increased considerably in the last decade because of their extraordinary features such as large specific surface area, thermal conductivity, and good optical transparency (Geim 2009; Dreyer et al. 2010; Zhu et al. 2010). Moreover, graphene oxide (GO) is an ideal nanomaterial for attaching a semiconductor catalyst such as TiO2 and ZnO. Its large surface area and high charge carrier mobility enhance further photocatalytic efficiency of semiconductor catalysts (Jiang et al. 2011; Wang et al. 2013). Indeed, the functionalized structures which are constructed by modifying GO with different materials have led to speeding up these trends in this field (Liu et al. 2010a, 2010b, 2010c; Guo & Dong 2011; Tang et al. 2013; Huang et al. 2014; Lee et al. 2016b). Among these materials, semiconductor catalysts which show photocatalytic properties attract much more attention. A photocatalyst can be defined in a simple manner as a semiconductor which generates free radicals after photon absorption of radiation. TiO2 and ZnO are the most commonly used semiconductors for photocatalytic studies due to their non-toxic behaviors, stability, inexpensiveness, high photosensitivity and photocatalytic degradation activity against organic pollutants (Sakthivel et al. 2003; Chakrabarti & Dutta 2004; Gaya & Abdullah 2007; Ocakoglu et al. 2012; Ocakoglu et al. 2013; Huang et al. 2014; Doruk et al. 2016). These remarkable properties are taken into account for the selection of these materials in this study. The fabrication of functionalized structures combining graphene-based materials and photocatalytic nanostructures could expand the application fields such as environmental areas.

The aim of this study was to create an antifouling and self-cleaning membrane surface. For this purpose, the polyethersulfone (PES) membranes blended with GO and with GO functionalized by photoactive semiconductor catalyst (TiO2 and ZnO) were synthesized via the phase inversion method and successfully used in methylene blue dye filtration. GO and semiconductor catalysts were prepared according to Hummers' and sol–gel methods, respectively. The characterization of the prepared GO and functionalized GO with TiO2 and ZnO catalysts were characterized by Fourier transform infrared spectroscopy (FTIR), scanning electron microscopy (SEM), and X-ray powder diffraction (XRD). Also, the structures of membranes and properties were investigated using energy-dispersive X-ray spectroscopy (SEM-EDX), atomic force microscopy (AFM), and static contact angle measurements. Additionally, antifouling properties of the prepared composite membranes were tested by filtration of Escherichia coli as a model bacteria solution. Finally, the performance of self-cleaning properties of the composite membranes was examined by filtrated methylene blue dye as a model solution. After that, the dye-contaminated membranes were subjected to visible light photocatalytic irradiation and the membrane surfaces were cleaned without using any chemical cleaning agents.

EXPERIMENTAL

Chemical reagents

Graphite powder (325 mesh, 99%) was purchased from Alfa Aesar. Titanium(IV) chloride (TiCl4) (99.9%), benzyl alcohol (99.8%), zinc nitrate hexahydrate (Zn(NO3)2.6H2O), polyethylene glycol (PEG300), and N-methyl-2-pyrrolidione (NMP) were purchased from Sigma-Aldrich. Sulfuric acid (98%, H2SO4), hydrochloric acid (38%, HCl), hydrogen peroxide (30%, H2O2), potassium permanganate (KMnO4), ammonia (NH3) and tetrahydrofuran were purchased from Merck. Polyethersulfone (Ultrason E 6020 P) as a polymer was kindly provided by BASF Company from Turkey. Endo agar used in preparing plates for the antimicrobial membrane filtration test was purchased from Merck. Endo agar plates were prepared according to Merck Microbiology Manual (Merck 2005) E. coli (ATCC 25922) bacteria was obtained from our Biotechnology Research Organization for Science and Technology (Mersin, Turkey). All chemicals were analytical grade and a two-stage Millipore Direct-Q3UV purification system was used to obtain distilled water.

Characterization of GO, GO–TiO2 and GO–ZnO nanopowders

A schematic diagram of GO–TiO2 and GO–ZnO nanocomposite formation is shown in Figure S1 (Supporting Information, available with the online version of this paper). However, the synthesis of GO, GO–TiO2 and GO–ZnO nanopowders is described in detail in the Supporting Information. XRD analysis was undertaken using a Rigaku Smartlab model X-ray diffractometer at Cu-Kα radiation (λ = 1.54 Å). The XRD analysis of dried TiO2 nanoparticles and ZnO nanorods samples were carried out continuous scans from 10 to 90° at 2° scan rate at 2θ min−1 in ambient air. SEM images of GO–TiO2 and GO–ZnO nanopowders were taken using a Zeiss Supra 55. The presence of alkyl chains on the metal oxide surfaces was confirmed by FTIR-attenuated total reflectance spectroscopy. The characterization of synthesized nanopowders (GO, GO–TiO2 and GO–ZnO) is given in the Supporting Information (Figures S2–S6, available with the online version of this paper).

Preparation of neat PES membrane

Neat PES membranes were synthesized by phase inversion method. The PES beads were dried at 80 °C in an oven for 2 h before use. The casting solution was prepared by dissolving PES beads (14% w/w) in NMP solvent at room temperature. The composition of casting solution is given in Table 1. The prepared solution was stirred vigorously at 60 °C for 6 h and then agitated at room temperature overnight to obtain a clear homogeneous solution. Afterwards, the polymer solution was ultrasonicated for 20 min to remove air bubbles from the solutions. The bubble-free solution was cast onto a glass plate with a casting knife of about 200 µm gap at 100 mm/s. The glass plate was immediately immersed into a coagulation bath including distilled water. After that, the formed membranes were stored in distilled water for 1 day for complete removal of the residual solvent. Thickness of the prepared membranes was measured by a digital micrometer and it was calculated as 163 ± 25 µm.

Table 1

The composition of the casting solution

Membrane sample PES (%wt.) NMP (%wt.) Sample (%wt.) 
Neat PES 14 86.0 0.0 
PES/GO-0.5 14 85.5 0.5 
PES/GO-1.0 14 85.0 1.0 
PES/GO-2.0 14 84.0 2.0 
PES/GO-ZnO-0.5 14 85.5 0.5 
PES/GO-ZnO-1.0 14 85.0 1.0 
PES/GO-ZnO-2.0 14 84.0 2.0 
PES/GO-TiO2-0.5 14 85.5 0.5 
PES/GO-TiO2-1.0 14 85.0 1.0 
PES/GO-TiO2-2.0 14 84.0 2.0 
Membrane sample PES (%wt.) NMP (%wt.) Sample (%wt.) 
Neat PES 14 86.0 0.0 
PES/GO-0.5 14 85.5 0.5 
PES/GO-1.0 14 85.0 1.0 
PES/GO-2.0 14 84.0 2.0 
PES/GO-ZnO-0.5 14 85.5 0.5 
PES/GO-ZnO-1.0 14 85.0 1.0 
PES/GO-ZnO-2.0 14 84.0 2.0 
PES/GO-TiO2-0.5 14 85.5 0.5 
PES/GO-TiO2-1.0 14 85.0 1.0 
PES/GO-TiO2-2.0 14 84.0 2.0 

Preparation of functionalized PES membrane

The neat and functionalized GO with photocatalytic semiconductor catalyst (TiO2 and ZnO) composite membranes were prepared according to the phase inversion method. In this method, the GO, GO–TiO2 and GO–ZnO nanopowders were added to the casting solution in order to prepare functionalized membranes. Different ratios of GO, GO–TiO2 and GO–ZnO nanopowders (0.5, 1.0, 2.0%wt.) were dispersed in NMP solvent using an ultrasonication bath for 15 min, and PES beads were then added to this solution. The PES membranes with different GO content were prepared, and the casting solution composition of various membranes is presented in Table 1. The casting solutions were ultrasonicated for 20 min to remove air bubbles from the solutions. Then, the casting solutions of PES/GO, PES/GO–TiO2, and PES/GO–ZnO were spread with a casting knife gap setting of 200 μm at 100 mm/s casting shear on a glass plate. The glass plates were immediately dipped in a deionized water bath to obtain polymer precipitation. After membrane formation, the membranes were taken from the water bath and stored in fresh deionized water for at least 1 day to guarantee the complete phase inversion.

Characterization methods for synthesized membranes

The surface and cross-section morphologies of the prepared membranes were characterized by SEM-EDX using a Zeiss/Supra 55 FE-SEM model operating at 15 kV. All samples were dried and coated under vacuum with a thin layer of platinum–palladium by a sputtering system.

AFM was used to characterize the surface roughness of prepared membranes. The average of the surface roughness (Ra) values was calculated as the standard deviation of all the height values within the given area. The AFM analyses were performed on a Park System XE-100 SPM AFM microscope. Small squares (0.5 × 0.5 cm) of dried samples were cut and 5 μm × 5 μm areas were scanned by contact mode in the air. Three different regions for each sample were scanned, and the average values of Ra are presented.

The contact angle of the neat and composite membranes was measured by a KSV CAM 200 goniometer (KSV Instruments). The analyses were performed for at least five different locations on the membrane surfaces.

The performance of the prepared membranes was measured by a dead-end flat sheet membrane module with a filtration area of 14.6 cm2 at a temperature of 25 ± 1 °C and an operating pressure of 5 bar.

The permeation flux was measured by collecting the filtered water at determined intervals and using the following equation: 
formula
1
where Jw,1 is permeate flux (L/(m2·h)); V the volume of permeate pure water (L), A the effective area of the membrane (m2), and Δt the filtration time (h).
The hydraulic permeability, Lp (L/(m2·h·bar)), through the membrane was determined from the slope of the plot of the water flux as a function of transmembrane pressure. 
formula
2

where ΔP refers to transmembrane pressure.

In order to calculate the membrane porosity, the dry membranes were cut to a definite size (14.6 cm2) and then immersed in distilled water for 1 day. Water on the surface of the membranes was removed carefully with a clean cloth and the membranes were weighed. Then, the membranes were dried for 24 h in a desiccator and weighed again to measure the membranes' weight in dry state. The membrane porosity (ε) was determined by gravimetric method, as defined in the following equation (Zhao et al. 2014):

 
formula
3
where Ww is the weight of the wet membrane (g); Wd the weight of the dry membrane (g); ρw the density of pure water at room temperature (0.998 g/cm3); A the effective area of the membrane (cm2), and the membrane thickness (cm).
Furthermore, the mean pore radius on the basis of the pure water flux and porosity data was determined using the filtration velocity method. According to the Guerout–Elford–Ferry equation, rp can be determined by (Xu et al. 2014a, 2014b; Zhao et al. 2014): 
formula
4
where rp is mean pore radius (nm); η water viscosity (8.9 × 10−4 Pa.s); the membrane thickness (m); Q the volume of the permeate water per unit time (m3/s); ΔP the operating pressure (0.1 MPa).
The flux recovery ratio (FRR) can be defined as follows (Vatanpour et al. 2012). 
formula
5
where Jw,2 is the water permeate flux of cleaned membranes (L/(m2·h)).

Microbial experimentation and antifouling property of the membranes

In this study, antibacterial activity of the prepared membranes containing GO, GO–TiO2 and GO–ZnO nanopowders against E. coli bacteria was investigated using the agar diffusion method. All types of membrane were filtered with 100 mL of E. coli suspension using a dead-end filtration system set-up. The bacterial load of the influent was about 1.2 × 106 colony-forming units (CFU) per millilitre. The CFU is a measure of viable bacterial cells. Bacteria were grown aerobically on agar plates at 37 °C overnight. Endo agar cup plate method was used to determine antibacterial activity of the GO, GO–TiO2 and GO–ZnO nanopowders against the test pathogen, E. coli. However, in order to test antibacterial activity of prepared membranes, they were cut into circular disks and put on the agar plate. The neat PES membrane was considered as the control sample (Ozay et al. 2016).

Self-cleaning activity of the membranes

In order to analyze the effect of semiconductor nanopowders for membrane cleaning, PES-GO composite membrane functionalized with TiO2 and ZnO nano powders was used to remove methylene blue dye from the membrane surface. Methylene blue dye (25 mg/L) was filtered for 2 h. After that, the membranes were exposed to UV light for 2 h. Then, the flux of membranes exposed to UV light was measured and compared with distilled water flux. The quartz photoreactor was surrounded by six UVA (Philips TL 8W Actinic BL) type lamps emitting waves at 365 nm predominantly. Lamps were placed in a hexagonal position inside an aluminum-foil-coated tube for equal reflection. The light intensity of the lamps was measured for UVA as 3.5 mW/cm2, respectively, using a UV-light meter (Lutron UVA-365 sensors) inside the quartz reactor (Doruk et al. 2016).

RESULTS AND DISCUSSION

Characterization of neat PES and PES/GO, PES/GO–TiO2, and PES/GO–ZnO blended membranes

Pure water permeation flux and hydrophilicity properties of membranes

The water hydraulic permeability (Lp) of water and contact angle of PES membranes blended with GO functionalized with TiO2 and ZnO nanopowder are illustrated in Figure 1. According to Figure 1, Lp of PES/GO-ZnO membranes was higher than that of PES/GO membranes. Pure water permeation flux of neat PES membrane was 43.8 ± 1.7 L/(m2·h·bar). Higher flux than that of neat membrane was obtained by PES/GO-0.5, PES/GO-1.0, and PES/GO-2.0: 55.9 ± 1.8, 60.1 ± 1.1, and 62.0 ± 1.5 L/(m2·h·bar), respectively. However, PES/GO-ZnO-0.5, PES/GO-ZnO-1.0, and PES/GO-ZnO-2.0 had 62.1 ± 0.7, 65.6 ± 1.1, and 66.5 ± 1.9 L/(m2·h·bar), respectively. Also, lowest flux was obtained as 47.3 ± 1.5, 21.6 ± 1.0, 19.5 ± 0.9 L/(m2·h·bar) for PES/GO-TiO2-0.5, PES/GO-TiO2-1.0, and PES/GO-TiO2-2.0 membranes, respectively. Through blending of GO and ZnO nanopowders into the polymer matrix, Lp of the membranes increased with increasing of nanopowder amount from 0.5 to 2.0%wt. The GO and GO-ZnO nanopowders might improve hydrophilicity property of membranes. However, lowest flux was obtained by GO-TiO2 membranes and Lp of the membranes decreased with increasing TiO2 amount. It was well described by Teow et al. (2012) that TiO2 has strong aggregation tendency in the polymer matrix. So, the TiO2 aggregation may have blocked the membrane's pores in our study.
Figure 1

Comparison of water hydraulic permeability and static contact angle of neat and GO blended PES membranes.

Figure 1

Comparison of water hydraulic permeability and static contact angle of neat and GO blended PES membranes.

It is well known that surface hydrophilicity is a significant factor in determining the flux and antifouling performance of membranes (Xu et al. 2016). The contact angle results of various membranes are compared in Figure 1. As shown in this figure, static contact angle decreased with addition of the GO and for functionalized GO-ZnO membranes. Neat PES membrane showed the highest water contact angle of 66.2 ± 1.3°. Blending of 0.5, 1.0, and 2.0%wt. GO nanopowders reduced the water contact angles to 56.4 ± 0.9°, 53.1 ± 1.1°, and 51.5 ± 0.8°, respectively. However, blending of 0.5, 1.0, and 2.0%wt. GO-ZnO nanopowders enhanced the water contact angles to 54.8 ± 1.2°, 50.6 ± 0.7°, and 49.2 ± 1.0°, respectively. Finally, blending of 0.5, 1.0, and 2.0%wt. GO-TiO2 nanopowders had 55.1 ± 1.2°, 51.7 ± 1.5°, and 50.5 ± 0.9°, respectively. The decreasing of contact angle can be due to the presence of carboxylic acid and hydroxyl groups on GO nanosheets, which increase hydrophilicity of GO. Composite membranes of GO functionalized with TiO2 and ZnO had an almost similar trend as the GO blended membranes. An increasing of nanopowder amount from 0.5 to 2.0%wt. caused a little decreasing of water contact angle of the composite membranes. Additionally, the ZnO nanorods can have a needle-spike structure and this structure decrease water contact angle.

The porosity and mean pore radius of the composite membranes are presented in Figures 2 and 3, respectively. Based on the results of porosity and mean pore radius of membranes presented in the figures, GO and GO-ZnO blended composite membranes showed an increase in porosity (Figure 2) and mean pore radius (Figure 3) compared with that of neat PES membrane. However, GO-TiO2 blended composite membranes displayed a decrease in porosity and mean pore radius. As can be seen from Figure 3, the neat PES membrane had a pore radius of 22.7 ± 2.2 nm. It can be also seen that mean pore radius of composite membranes was in the range between 7.37 ± 0.9 and 33.2 ± 1.6 nm.
Figure 2

The porosity of membranes of neat PES, PES/GO, and PES/GO functionalized with semiconductor catalysts (TiO2 and ZnO).

Figure 2

The porosity of membranes of neat PES, PES/GO, and PES/GO functionalized with semiconductor catalysts (TiO2 and ZnO).

Figure 3

The mean pore radius of membranes of neat PES, PES/GO, and PRS/GO functionalized with semiconductor catalysts.

Figure 3

The mean pore radius of membranes of neat PES, PES/GO, and PRS/GO functionalized with semiconductor catalysts.

The increased hydrophilicity together with the bigger pore and higher porosity enabled the GO and GO-ZnO nanopowder blended membranes to have a good pure water permeation flux compared with the other GO-TiO2 nanopowders blended membranes.

Morphology of the GO and functionalized GO blended PES membranes

The AFM analyses were used to understand changes in the surface roughness of the nanocomposite PES membranes. Figure 4 indicates the three-dimensional AFM images of the neat PES membrane and the semiconductor catalyst (TiO2 and ZnO) blended composite membranes. The brightest area represents the highest point of the membrane surface and the dark regions indicate valleys or membrane pores. Values of the root mean square roughness (Rq), the average roughness (Ra), and 10-point average roughness (Rz) of various membranes are given in Table 2. The AFM images of the membranes synthesized by the solution blending method (Figure 4(b)4(j)) indicate that the surface roughness is higher than that of prepared PES membrane (Figure 4(a)). As expected, the concentration of GO, TiO2 and ZnO affect the surface roughness. AFM measurements performed in contact mode give more consistent roughness statistics. The addition of GO-TiO2 and GO-ZnO nanopowder structures to PES membrane leads to increase in the Ra value, depending upon the ratio, size and shape of metal oxide nano-structures, compared to neat PES membrane (Table 2). Scanning on micro-scale regions of each PES/GO, PES/GO–TiO2 and PES/GO–ZnO blended membrane reveals different roughness on the membrane surfaces due to the various formations such as aggregations of additive materials (GO, GO–TiO2, GO–ZnO).
Table 2

Rq, Ra, and Rz values of the composite membranes

Membrane sample Rq (nm) Ra (nm) Rz (nm) 
Neat PES 4.453 3.547 20.129 
PES/GO-0.5 5.523 4.321 18.572 
PES/GO-1.0 6.794 5.695 27.712 
PES/GO-2.0 5.046 3.902 22.916 
PES/GO-ZnO-0.5 6.728 5.231 24.342 
PES/GO-ZnO-1.0 6.320 5.409 19.918 
PES/GO-ZnO-2.0 5.560 4.546 24.444 
PES/GO-TiO2-0.5 5.050 4.104 21.824 
PES/GO-TiO2-1.0 7.236 5.525 27.701 
PES/GO-TiO2-2.0 5.516 4.316 21.977 
Membrane sample Rq (nm) Ra (nm) Rz (nm) 
Neat PES 4.453 3.547 20.129 
PES/GO-0.5 5.523 4.321 18.572 
PES/GO-1.0 6.794 5.695 27.712 
PES/GO-2.0 5.046 3.902 22.916 
PES/GO-ZnO-0.5 6.728 5.231 24.342 
PES/GO-ZnO-1.0 6.320 5.409 19.918 
PES/GO-ZnO-2.0 5.560 4.546 24.444 
PES/GO-TiO2-0.5 5.050 4.104 21.824 
PES/GO-TiO2-1.0 7.236 5.525 27.701 
PES/GO-TiO2-2.0 5.516 4.316 21.977 
Figure 4

Three-dimensional AFM images of membranes of neat PES, PES/GO, and PES/GO functionalized with semiconductor catalysts.

Figure 4

Three-dimensional AFM images of membranes of neat PES, PES/GO, and PES/GO functionalized with semiconductor catalysts.

Figure 5 shows the morphologies of top-layer and cross-section of the membranes. Neat membrane had a dense skin layer and a support layer with finger-like structure (Figure 5(a) and 5(b)). As shown in Figure 5(c)5(n), the surface of GO and functionalized GO-ZnO was smooth and the agglomerated GO and GO-ZnO were not detected on the membrane surface. However, GO-TiO2 agglomeration was obviously observed on the membrane surface (Figure 5(o)5(t)). In addition, cracks were distinguished on all the synthesized membranes and supplied ultrafiltration properties. The cross-sectional images showed that some narrow pore structures were obtained when the GO-TiO2 nanopowder content increased, especially in 2.0%wt. GO blended membranes (Figure 5(p)5(t)). In addition, wider pore structures were seen for functionalized GO-ZnO composite membrane. These wider pores improved the water fluxes of the membranes.
Figure 5

Top-surface and cross-section SEM images of the neat PES, PES/GO, and PES/GO functionalized with semiconductor catalysts.

Figure 5

Top-surface and cross-section SEM images of the neat PES, PES/GO, and PES/GO functionalized with semiconductor catalysts.

The SEM-EDX images of the neat and GO, GO–TiO2 and GO–ZnO nanopowder blended composite membranes are shown in Figure 6. The PES element structure can be clearly observed for neat membrane (Figure 6(a)) and GO blended composite membrane (Figure 6(b)). The SEM-EDX images prove for GO–TiO2 and GO–ZnO blended composite membranes that TiO2 and ZnO nanopowders successfully penetrated the polymer chain and the surface of PES membrane (Figure 6(c) and 6(d)).
Figure 6

The SEM-EDX images from the surface of the (a) neat PES membrane, (b) PES/GO-2.0, (c) PES/GO-TiO2-2.0, and (d) PES/GO-ZnO-2.0.

Figure 6

The SEM-EDX images from the surface of the (a) neat PES membrane, (b) PES/GO-2.0, (c) PES/GO-TiO2-2.0, and (d) PES/GO-ZnO-2.0.

Antimicrobial performance of neat PES and PES/GO, PES/GO–TiO2 and PES/GO–ZnO blended membranes

GO and ZnO nanomaterials are well-known antimicrobial agents (Bao et al. 2002; Liu et al. 2009; Liu et al. 2011; Vatanpour et al. 2012; Ocakoglu et al. 2014; Xu et al. 2014a, 2014b; Zhao et al. 2014). As shown in Figure 7, neat PES and PES/GO-0.5 membranes have little or no antimicrobial activity, demonstrated by metallic sheen formation, which is an indicator of E. coli growth, around the membranes after overnight incubation at 37 °C (Wehr & Frank 2004). It was easily recognized on the neat membrane that a great deal of CFUs appeared. An effective antimicrobial activity was readily seen for all of the PES membranes except for the two mentioned above. We assert that there is a correlation between antimicrobial activity and PES membrane nanopowder concentration for all of the membrane types. We also suggest that 0.5%wt. nanopowder concentrations for all membranes, except PES/GO-0.5, are capable of producing antimicrobial affectivity on E. coli bacteria. The other reason for similar performance in the antibacterial activity was the distribution of PES/GO, PES/GO–ZnO and PES/GO–TiO2 membranes. GO surface functionalization with semiconductor catalysts was effective in providing almost the same strong antimicrobial activity, comparable to that of other membranes (Tiraferri et al. 2011). In our study, PES/GO-2.0, PES/GO-TiO2-2.0 and PES/GO-ZnO-2.0 functionalization imparted strong antimicrobial activity to the membranes (Figure 7). Moreover, it can be clearly seen from Figure 8 that any living bacteria were not detected in the membrane permeate (PES/GO-ZnO-2.0) after contact with membrane and E. coli for 1.5 h. GO inactivates bacteria upon direct cell contact by inducing membrane damage, mediated by physical disruption, charge transfer and formation of reactive oxygen species (Wehr & Frank 2004; Akhavan & Ghaderi 2010; Liu et al. 2011; Tiraferri et al. 2011) and extraction of lipid from the cell membrane (Tu et al. 2013). We think the same mechanism may be working for the other PES/GO-TiO2-2.0 and PES/GO-ZnO-2.0 membranes. Our results are consistent with previous literature (Perreault et al. 2014; Perreault et al. 2015).
Figure 7

Results of E. coli growth in an inhibition test.

Figure 7

Results of E. coli growth in an inhibition test.

Figure 8

Results of E. coli growth in membrane permeate (left) and initial solution (right).

Figure 8

Results of E. coli growth in membrane permeate (left) and initial solution (right).

Self-cleaning properties of PES/GO–TiO2 and PES/GO–ZnO nanopowder blended membranes

The self-cleaning performance of the neat and functionalized with semiconductor catalysts GO composite membranes was tested by measuring the water flux recovery after the membrane was fouled by methylene blue dye solution (25 ppm). For cleaning of fouled membranes, they were removed from the dead-end module and rinsed with water for 2 min. Then, the membranes were immersed in the photocatalytic reactor filled with distilled water, and UV light was exposed on membrane surface. After that, water fluxes of the cleaned membranes with UV light were measured with distilled water. Figure 9 shows the FRR for the cycle of dye fouling–cleaning experiments in order to test the performance of the membrane self-cleaning property. It can be seen from this figure that the flux recovery values in those five cycles were 90.2%, 84.5%, 78.3%, 72.4%, and 66.1% for PES/GO-ZnO-2.0; 71.1%, 64.6%, 59.4%, 52.1%, and 46.5% for PES/GO-TiO2-2.0; and 35.7%, 30.3%, 24.1%, 18.7%, and 13.1% for neat PES membrane. It was concluded that reversible fouling was the dominant phenomenon and exposure to UV light degraded the dye molecules on the surface of GO-ZnO and GO-TiO2 nanopowder membranes. It can be also seen that GO-ZnO had better self-cleaning efficiency than GO-TiO2 in this study. It might be that TiO2 supplied low photocatalytic efficiency due to the rapid charge recombination rate within TiO2 particles (Anandan et al. 2012).
Figure 9

FRR of neat PES, PES/GO-TiO2-2.0, and PES/GO-ZnO-2.0.

Figure 9

FRR of neat PES, PES/GO-TiO2-2.0, and PES/GO-ZnO-2.0.

The photos of neat membranes and blended membranes of GO functionalized with photoactive semiconductor (TiO2 and ZnO) exposed to UV light are shown in Figure 10. The surfaces of fouled neat PES, PES/GO-ZnO-2.0, and PES/GO-TiO2-2.0 membranes are shown in Figure 10(a)10(c). It can be clearly seen that PES/GO-ZnO-2.0 (Figure 10(e)) membrane had the cleanest membrane surface when compared with neat (Figure 10(d)) and PES/GO-TiO2-2.0 (Figure 10(f)) membranes.
Figure 10

Photographs of membranes fouled with methylene blue dye: (a) neat PES, (b) PES/GO-ZnO-2.0, (c) PES/GO-TiO2-2.0; and membranes cleaned with UV light: (d) neat PES, (e) PES/GO-ZnO-2.0, (f) PES/GO-TiO2-2.0.

Figure 10

Photographs of membranes fouled with methylene blue dye: (a) neat PES, (b) PES/GO-ZnO-2.0, (c) PES/GO-TiO2-2.0; and membranes cleaned with UV light: (d) neat PES, (e) PES/GO-ZnO-2.0, (f) PES/GO-TiO2-2.0.

CONCLUSIONS

The PES membranes blended with GO and with GO functionalized with photocatalytic semiconductor catalyst (TiO2 and ZnO) were prepared by direct addition of the nanopowders in the casting solution. The effects of blended GO and functionalized GO composite membranes on the morphology were examined by water hydraulic permeability, antibacterial resistance and self-cleaning properties of the functionalized membrane surface. The results showed that the carboxylic acid, hydroxyl, and other functional groups formed on the graphene supplied hydrophilicity and increased water flux when compared with neat membrane. Moreover, GO functionalized with ZnO composite membrane brought more hydrophilicity to the PES membrane. But the water flux of GO membrane functionalized with TiO2 decreased due to blocking of the membrane pores by round-shape TiO2 nanoparticles. A significant improvement was observed in biofouling prevention in the GO and functionalized GO composite membrane. The cross-section SEM images showed that all of the synthesized membranes had finger-like structure. The self-cleaning properties of the functionalized GO with TiO2 and ZnO semiconductor catalyst were also investigated. The results indicated that methylene blue dye was successfully degraded by TiO2 and ZnO when exposed to UV light on the membrane surface. Semiconductor catalysts protected the membrane surface from binding dye molecules into the polymer chain. It was concluded from this study that the GO is excellent antifouling material. Moreover, functionalized GO with ZnO nanopowders showed unique self-cleaning properties, which is promising for new applications for this type of membrane.

ACKNOWLEDGEMENT

This research was supported by the Department of Scientific Research Projects of Mersin University (project number: 2017-1-TP2-2010).

REFERENCES

REFERENCES
Anandan
S.
Rao
T. N.
Sathish
M.
Rangappa
D.
Honma
I.
Miyauchi
M.
2012
Superhydrophilic graphene-loaded TiO2 thin film for self-cleaning applications
.
ACS Appl. Mater. Interfaces
5
,
207
212
.
Chang
I. S.
Le Clech
P.
Jefferson
B.
Judd
S.
2002
Membrane fouling in membrane bioreactors for wastewater treatment
.
J. Environ. Eng.
128
(
11
),
1018
1029
.
Chidambaram
T.
Noel
M.
2015
Membrane processes for dye wastewater treatment: recent progress in fouling control
.
Crit. Rev. Environ. Sci. Technol.
45
,
1007
1040
.
Dizge
N.
Soydemir
G.
Karagunduz
A.
Keskinler
B.
2011
Influence of type and pore size of membranes on cross flow microfiltration of biological suspension
.
J. Membr. Sci.
366
(
1–2
),
278
285
.
Doumic
L. I.
Soares
P. A.
Ayude
M. A.
Cassanello
M.
Boaventura
R. A. R.
Vilar
V. J. P.
2015
Enhancement of a solar photo-Fenton reaction by using ferrioxalate complexes for the treatment of a synthetic cotton-textile dyeing wastewater
.
Chem. Eng. J.
277
,
86
96
.
Dreyer
D. R.
Park
S.
Bielawski
C. W.
Ruoff
R. S.
2010
The chemistry of graphene oxide
.
Chem. Soc. Rev.
39
,
228
240
.
Geim
A. K.
2009
Graphene: status and prospects
.
Science
324
,
1530
1534
.
Hayat
H.
Mahmood
Q.
Pervez
A.
Bhatti
Z. A.
Baig
S. A.
2015
Comparative decolorization of dyes in textile wastewater using biological and chemical treatment
.
Sep. Purif. Technol.
154
,
49
153
.
Jegatheesan
V.
Pramanik
B. K.
Chen
J.
Navaratna
D.
Chang
C. Y.
Shu
L.
2016
Treatment of textile wastewater with membrane bioreactor: a critical review
.
Bioresour. Technol.
204
,
202
212
.
Lee
A.
Elam
J. W.
Darling
S. B.
2016a
Membrane materials for water purification: design, development, and application
.
Environ. Sci.: Water Res. Technol.
2
,
17
42
.
Liu
Y.
He
L.
Mustapha
A.
Li
H.
Hu
Z. Q.
Lin
M.
2009
Antibacterial activities of zinc oxide nanoparticles against Escherichia coli O157:H
.
J. Appl. Microbiol.
107
,
1193
1201
.
Liu
Y.
Li
M.
Fang
Q.
Lv
Q.
Wu
M.
Cao
S.
2010a
Structural and photoluminescence properties of polyethylene glycol (PEG)-assisted growth Co-doped ZnO nanorod arrays compared with pure ZnO nanorod arrays
.
Chin. J. Phys.
48
(
4
),
523
531
.
Merck
2005
Merck Microbiology Manual
, 12th edn.
Merck KGaA, Darmstadt
,
Germany
.
Ocakoglu
K.
Harputlu
E.
Guloglu
P.
Erten-Ela
S.
2012
The photovoltaic performance of new ruthenium complexes in DSSCs based on nanorod ZnO electrode
.
Synth. Met.
162
,
2125
2133
.
Ocakoglu
K.
Zafer
C.
Varlikli
C.
Icli
S.
2013
Preparation of dye sensitized titanium oxide nanoparticles for solar cell applications
.
Mater. Sci. Semicond. Process.
16
,
1688
1694
.
Ozay
Y.
Dizge
N.
Gulsen
H. E.
Akarsu
C.
Harputlu
E.
Ozer
E.
Unyayar
A.
Ocakoglu
K.
2016
Investigation of electroactive and antibacterial properties of polyethersulfone membranes blended with copper nanoparticles
.
Clean – Soil, Air, Water
44
(
8
),
930
937
.
Perreault
F.
Tousley
M. E.
Elimelech
M.
2014
Thin-film composite polyamide membranes functionalized with biocidal graphene oxide nanosheets
.
Environ. Sci. Technol. Lett.
1
(
1
),
71
76
.
Perreault
F.
Fonseca de Faria
A.
Nejati
S.
Elimelech
M.
2015
Antimicrobial properties of graphene oxide nanosheets: why size matters
.
ACS Nano
9
(
7
),
7226
7236
.
Punzi
M.
Nilsson
F.
Anbalagan
A.
Svensson
B. M.
Jönsson
K.
Mattiasson
B.
Jonstrup
M.
2015
Combined anaerobic–ozonation process for treatment of textile wastewater: removal of acute toxicity and mutagenicity
.
J. Hazard. Mater.
292
,
52
60
.
Richards
H. L.
Baker
P. G. L.
Iwuoha
E.
2012
Metal nanoparticle modified polysulfone membranes for use in wastewater treatment: a critical review
.
J. Surf. Eng. Mater. Adv. Technol.
2
,
183
193
.
Safarpour
M.
Khataee
A.
Vatanpour
V.
2014
Preparation of a novel polyvinylidene fluoride (PVDF) ultrafiltration membrane modified with reduced graphene oxide/titanium dioxide (TiO2) nanocomposite with enhanced hydrophilicity and antifouling properties
.
Ind. Eng. Chem. Res.
53
,
13370
13382
.
Sakthivel
S.
Neppolian
B.
Shankar
M. V.
Arabindoo
B.
Palanichamy
M.
Murugesan
V.
2003
Solar photocatalytic degradation of azo dye: comparison of photocatalytic efficiency of ZnO and TiO2
.
Solar Energy Mater. Solar Cells
,
77
65
82
.
Torretta
V.
Urbini
G.
Raboni
M.
Copelli
S.
Viotti
P.
Luciano
A.
Mancini
G.
2013
Effect of powdered activated carbon to reduce fouling in membrane bioreactors: a sustainable solution
.
Case study. Sustainability
5
,
1501
1509
.
Tu
Y.
Lv
M.
Xiu
P.
Huynh
T.
Zhang
M.
Castelli
M.
Liu
Z.
Huang
Q.
Fan
C.
Fang
H.
Zhou
R.
2013
Destructive extraction of phospholipids from Escherichia coli membranes by graphene nanosheets
.
Nat. Nanotechnol.
8
,
594
601
.
Wehr
H. M.
Frank
J. H.
2004
Standard Methods for the Examination of Dairy Products
, 17th edn.
American Public Health Association
,
Washington, DC
.
Zhu
Y. W.
Murali
S.
Cai
W. W.
Li
X. S.
Suk
J. W.
Potts
J. R.
Ruoff
R. S.
2010
Graphene and graphene oxide: synthesis, properties, and applications
.
Adv. Mater.
22
,
3906
3924
.
Zinadini
S.
Zinatizadeh
A. A.
Rahimi
M.
Vatanpour
V.
Zangeneh
H.
2014
Preparation of a novel antifouling mixed matrix PES membrane by embedding graphene oxide nanoplates
.
J. Membr. Sci.
453
,
292
301
.

Supplementary data