Engineering design of a biofilm formed on a pH-sensitive ZnO/PSf nanocomposite membrane with antibacterial properties

Abstract

In this research, zinc oxide (ZnO) nanoparticles were used on a matrix of a polysulfone ultrafiltration membrane to make a nanocomposite membrane with a positive surface charge for filtration of biological macromolecules. The engineering of the biofilm structure developed onto the membrane surface leads to increased membrane flux, elevated rejection of protein, and reinforced antifouling properties of the membrane. AFM results indicated that the formed biofilm has increased roughness at higher pH levels. Further, FTIR analysis proved that the extent of biological macromolecules deposited on the membrane surface is greater at higher pHs. Through engineering the conditions of nanocomposite fabrication and adjustment of pH of the protein solution, the best antifouling performance was related to the nanocomposite containing 0.5 wt% of nanoparticles at pH of 8.9. Antibacterial tests proved the antibacterial properties of the nanocomposites containing ZnO nanoparticles.

---

1. Introduction

Proteins are among the important natural macromolecules of great importance in nutritional, health, and pharmaceutical areas thanks to their special structure and properties developed by their amine groups.¹ BSA protein is the most abundant protein in blood plasma. It is responsible for carrying medications and nutrition in the body. BSA has several functional groups including binding, transport, delivery of fatty acids, bilirubin, etc.² Based on the mentioned properties, BSA is used as a natural biological macromolecule in research as the representative of proteins in experiments related to biological fouling in membrane bioreactor systems.³ Separation of proteins to use their concentrated product containing high protein percentage or to use the residual medium containing lower protein percentage, is of great interest today. Various methods have been proposed for separation of proteins including liquid chromatography, electrophoretic, and membrane separation methods.⁴ Similar to separation of other compounds, the problem of separating proteins by membrane filtration processes is the fouling developed in the pores and surface of the membrane during the filtration process, gradually decreasing the flux and yield and eventually the membrane durability.^3,5 One of the reasons of fouling development is the hydrophobicity property in polymer membranes resulting in attachment of hydrophobic compounds to the surface and pores, thereby development of fouling.^6,7 Therefore, in order to reduce membrane fouling, various methods are in use to enhance the membrane hydrophilicity.^8,9 One of the popular methods is using hydrophilic polymer additives such as polyethylene glycol (PEG) and poly vinyl pyrrolidone (PVP).^6,10 Another method is application of metal oxide nanoparticles which donate special and various properties to membranes and composites thanks to their special characteristics related to their nano size.^11–14 So far, different nanoparticles including titanium oxide, silver, and alumina have been used for this purpose,^5,15–20 but usage of ZnO nanoparticles has been understudied.^21–23 ZnO nanoparticles enjoy various properties including reasonable price, stability, antibacterial and antifungal properties, simpler production than TiO₂, lower toxicity than TiO₂, greater area than TiO₂ due to the crystal structure of ZnO nanoparticles, higher thermal resistance of the membrane, and easy attachment to hydroxyl groups (–OH), as well as development of hydrophilicity and thus are of great interest in this research.²⁴ The mentioned properties along with the photo oxidizing properties against biological and chemical genera, self-sterilization, as well as the antibacterial properties of this nanoparticle make it highly applicable in cosmetic and pharmaceutical areas as well as separation of natural macromolecules.²⁵ Considering the sensitive nature of protein to acidity, the effect of the acidity of filtration environment on the results of separation and extent of fouling is of great importance. This is because as the pH changes and approaches the isoelectric point of the protein, accumulation of protein molecules takes place.²⁶ On the other hand, fabrication of a membrane with a surface sensitive to acidity reinforces the extent of sensitivity of filtration results to acidity.²⁷

The aim of this research is to synthesize a membrane to separate BSA protein (as a model protein representative of the proteins available in activated sludge environment). For this purpose and in order to lower the membrane's biological fouling, PEG, a hydrophilic polymer additive, together with ZnO hydrophilic nanoparticles were used. The density, porosity, and surface charge of the layer deposited on the membrane surface influence the flux, separation yield, and extent of fouling. The BSA's natural sensitivity to acidity and the membrane surface allow for the possibility of engineering the biological film developed in the membrane surface. Therefore, the objectives specified in this regard are synthesizing a nanocomposite membrane containing different weight percentages of nanoparticles, investigation of the fouling and resistances present in the filtration of the BSA protein solution by these membranes, examination of the effect of surface charge of membranes and the acidity of the protein solution on fouling, and eventually determination of the optimal level of nanoparticles in the membrane and the best pH in the filtration of the desired protein, in order to reach conditions in which the protein filtration is done with a greater flux and separation yield. Next, based on the obtained results and observations, filtration was also conducted in a membrane bioreactor (MBR) system. Different resistances were obtained and the developed biofilm structure was examined at different pHs.

2. Materials and methods

2.1. Materials

BASF polysulfide and N-methylpyrrolidone (NMP) solution were purchased from Ultrason and Merck, respectively. ZnO nanoparticles were synthesized with a size of 20 nm, through wet chemical method.²⁸ Polyethylene glycol (PEG) was purchased from Merck as a pore maker with a molecular weight of 20 000 g mol⁻¹. BSA protein of V fraction type was supplied by Merck. KH₂PO₄, K₂HPO₄, glacial acetic acid, and C₂H₃NaO₂·3H₂O were purchased from Merck to prepare phosphate and acetate buffer. The activated sludge was provided from the urban treatment plant of Tehran's Ekbatan Town.

2.2. Methods

2.2.1. Preparation of the polymer membrane. To synthesize the nanocomposite membrane through blending method, the nanoparticles were dispersed by an ultrasonic bath in NMP solution at 70 °C. The solution was then stirred by a magnetic stirrer at constant and uniform rate, to which polysulfide grains were gradually added at the same temperature of 70 °C, while the solution was being continuously stirred. Once all polymer grains were dissolved, the pore maker (PEG) was added and finally after the complete dissolution of the pore maker and achieving a homogenous mixture, the polymeric solution was cast on a smooth glass plate by a casting knife with a thickness of 350 μm and then immersed in a non-solvent bath. The membrane was synthesized through phase inversion method (the details can be found in the research by Homayoonfal et al.^29–31). Table 1 indicates the composition of the synthesized membranes.

Table 1 The composition of the membranes synthesized in this research

No.	Membrane type	PSF%	NMP%	PEG%	(ZnO/PSF)%
1	PEG-NCM0	16	75	9	0
2	PEG-NCM0.15	16	75	9	0.15
3	PEG-NCM0.3	16	75	9	0.3
4	PEG-NCM0.4	16	75	9	0.4
5	PEG-NCM0.5	16	75	9	0.5
6	PEG-NCM0.6	16	75	9	0.6
7	PEG-NCM1	16	75	9	1
8	PEG-NCM1.5	16	75	9	1.5
9	PEG-NCM2	16	75	9	2

---

When naming the membranes, first the name of the pore former was written as PEG. Next, the nanocomposite membrane was written in short as NCM (Nano Composite Membrane). Finally, the percentage of nanoparticles was provided. Note that the PEG-NCM0 membrane equals a blank membrane devoid of nanoparticle.

2.2.2. The filtration tests of BSA protein and determination of the filtration resistances. To evaluate the performance of filtration and separation of BSA biomolecule and calculate the present resistances, a dead end system was employed. BSA protein was fabricated in acetate buffer at pH = 3.5 and in phosphate buffer at pHs = 7, 8.3, and 8.9. It then underwent filtration analysis in the dead end filtration system, according to Fig. 1. The effective area of the membrane in this system was 4.52 cm² and the constant pressure exerted by nitrogen capsule was 300 kPa. The fluid used during the experiment was uniformly stirred with a suitable rate by a magnetic stirrer to prevent development of concentration polarization around the membrane. Each tenth of a second, the weight of the treated fluid was recorded by a balance as well as computer and finally converted to the flux.

Fig. 1 A schema of the dead end system to test the membrane fouling.

Eventually, the filtration in the membrane bioreactor was conducted in a system shown in Fig. 2. It was performed at the constant pressure developed equal to −0.3 bar and membrane effective area of 23.75 cm². The bioreactor had an active volume of 7.5 L and the COD of the sludge in the initial state was 1000 mg L⁻¹. Feeding to the reactor was carried out based on ratios of 5, 1, and 100 for nitrogen, phosphorus, and carbon, respectively. The sources for supplying these three substances were ammonium nitrate, ammonium phosphate, and glucose, respectively. Furthermore, the MLSS of the utilized sludge was 7 and its pH varied between 6 and 7.

Fig. 2 A schema of the membrane bioreactor system to test the membrane fouling.

In a dead end system and MBR, one can determine the degree of membrane fouling through the serial resistances present across the fluid passage route. The model of serial resistances can be expressed as follows:¹⁵

In this relation, J is the output flux (m³ m⁻² s⁻¹), ΔP_T is the pressure of both sides of the membrane (Pa), μ is the viscosity of the output solution (Pa s), and R_t is the total resistance of the filtration (m⁻¹).

The total resistance is defined as the sum of the membrane's intrinsic resistance (R_m), the resistance of the biofilm developed on the surface of the membrane (R_C), and the resistance of fouling developed in response to pores clogging and irreversible absorption of contaminants on the walls of pores and the membrane surface (R_P).

R_t = R_m + R_C + R_P (2)

The mentioned resistances can be calculated by the following equations through experimental data:

In these relations, J_S is the flux of the desired solution under steady conditions, J_w1 is the initial flux of water, and J_w2 is the final flux of water after removal of the biofilm layer.

In order to express the extent of rejection of a compound in the filtration fluid by the membrane, membrane rejection measurement is used, which is calculable as follows:

where, C_p is the concentration of the desired compound in the treated fluid and C_f is the concentration of the compound of interest in the initial fluid or the feed.

During filtration tests, to minimize the experimental error, all flux behaviors (J_S, J_w1 and J_w2) were measured 3 times and averaged for each membrane sample.

2.3. The analyses for determination of membranes' properties

2.3.1. Scanning electron microscopy (SEM) analysis. In order to observe the size of nanoparticles, SEM images were taken by Hitachi, S-4160 apparatus, Japan. All of the samples were dried, then covered with a thin gold layer to make them conductive and prepared to be observed by SEM. The obtained SEM images were processed by ImageJ software for analyzing the particle size distribution.

2.3.2. Thermo gravimetric (TGA) analysis. Thermo gravimetric analysis (TGA) of the membrane samples was carried out using a Thermal Analysis (TA) TGA Q50 Thermo gravimetric Analyzer (made in USA). The temperature range of 25–800 °C was used under air atmosphere with a heating rate of 10 K min⁻¹.

2.3.3. Mechanical resistance analysis. For evaluating the membrane mechanical resistance, a Testing Machine (INSTRON 5566, USA) was employed. To this end, the samples with a thickness of 300 μm and dimension of 0.5 cm × 3 cm were put in the device. The stretching was kept at the speed of 1 mm min⁻¹ and then sample elongation was recorded against the exerted pressure.

2.3.4. Dynamic light scattering (DLS) analysis. DLS test was used to measure the surface charge and zeta potential of the particles present in the solution or the membrane surface as well as their electrostatic interaction with the protein molecules. For this purpose, Brookhavan Device (90Plus/BI-MAS) and Electro Kinetic Analyzer (EKA) (USA) were used to measure zeta potential in the solutions and on the surface of membranes, respectively.

2.3.5. FTIR analysis. The ATR-FTIR analysis was conducted to confirm the presence of ZnO nanoparticles on the membrane, and detect the functional groups present on the membrane surface. For this test, the membranes were first slightly thinned under a pressure of 10 bar so that light could pass through them. This test was performed by Perkin Elmer Spectrum2 Device (USA).

2.3.6. Atomic force microscopy (AFM) analysis. AFM test was employed to investigate the changes in the morphology of the surface of membranes synthesized under various conditions. The parameter of roughness is important in membrane fouling, which is among the properties of the membrane itself and measurable by AFM images.

In this research, AFM test was conducted by Veeco Device, Auto Probe-cp-research (USA). The test was done at the contact state with a scanning rate of 2 Hz and resolution of 256 × 256. In this research, AFM images were taken from the surface of membranes or the protein biofilm developed on them. All these were evaluated further.

2.3.7. Antibacterial assay analysis. When conducting this test, the membranes of interest were placed in plates containing nutrient agar culture medium, on which E. coli was cultured as an index as to whether water is hygienic and antibacterial or not by Kirby–Bauer antibiotic testing (KB testing or disc diffusion antibiotic sensitivity testing) method.^32,33 Following around 24 h of exposure at 30 °C in an incubator, a region of the membrane surroundings in which no bacteria was grown was introduced as the inhibition zone.³⁴

3. Results and discussion

The polysulfone nanocomposite membranes containing ZnO nanoparticles were synthesized through blending of ZnO nanoparticles with casting solution followed by phase inversion method. The characterization of ZnO nanoparticles was performed in order to detect size distribution (SEM images), functional groups (FTIR analysis), and surface charge (DLS analysis) of ZnO nanoparticles. Moreover, the structural analyses of the nanocomposite were examined in terms of TGA analysis, mechanical strength analysis and AFM analysis. Finally, the filtration tests of BSA protein and filtration in the MBR system were conducted to examine the properties of rejection, membrane flux, and filtration resistances. The effect of surface charge of the nanocomposite and the acidity of the protein solution were also investigated. All of the results are individually presented next.

3.1. SEM analysis

Fig. 3 shows the SEM image of ZnO nanoparticles and particle size distribution of them measured by the image processor software (ImageJ).

Fig. 3 SEM image of ZnO nanoparticle (a) and particle size distribution of nanoparticles (b) prepared by the image processor software (ImageJ).

As can be seen from SEM image presented in Fig. 3(a), the sample size of ZnO nanoparticles is approximately 76 nm. Moreover, the size distribution of nanoparticles (Fig. 3(b)) demonstrates that ZnO nanoparticles have a size distribution with a mean size of about 60 nm.

3.2. TGA analysis

In order to investigate thermal behavior of membrane samples and the quantity of ZnO entrapped in the PSf matrix, thermo gravimetric analysis was performed. The obtained results are presented in Fig. 4, demonstrating that the thermal degradation temperature increases with the growth in ZnO loading, confirming the positive influence of inorganic nanoparticles in thermal stability of blended nanocomposite membranes.³⁵

Fig. 4 TGA curve of blank membrane PEG-NCM0 (a) and different nanocomposite membrane: PEG-NCM0.3 (b), PEG-NCM0.5 (c) and PEG-NCM2 (d).

TGA curve clearly shows that all blank and nanocomposite membranes are stable in a wide range of operating temperatures from 25–90 °C, which is typical of common membrane operations. As exhibited in the TGA curve, residual masses reported after TGA analysis for PEG-NCM0, PEG-NCM0.3, PEG-NCM0.5 and PEG-NCM2 were 17.4%, 21.9%, 30%, 31.9%, respectively, which confirm the trend of nanoparticle content entrapped in the polysulfone matrix.

3.3. Mechanical resistance analysis

Fig. 5 illustrates the effect of nanoparticles loading on the membrane mechanical strength. As obviously seen in Fig. 5, the initial slope of tensile curves increases with more nanoparticle loading, confirming that the presence of nanoparticles with a high mechanical resistance enhances membrane Young's modulus. Moreover, ZnO nanoparticles diminish the resistance to break due to lowered uniformity of polymeric film.³⁶

Fig. 5 The mechanical strength behavior of blank membrane PEG-NCM0 (a) and different nanocomposite membranes PEG-NCM0.3 (b), PEG-NCM0.5 (c) and PEG-NCM2 (d).

3.4. Measurement of surface charge

The aim of this test is to measure the surface charge for different samples including the solution of nanoparticles and the synthesized membranes. In this test, the surface charge present in the ZnO nanoparticles solution dispersed in the NMP solvent was measured along with the polymer solution devoid of nanoparticles including the PSF polymer and NMP solvent, and the polymer solution containing dispersed nanoparticles. In addition to the three mentioned solutions, the surface charge of the synthesized membranes: PEG-NCM0.5 and PEG-NCM2 was also measured, with the results expressed in Table 2. As can be seen from the information in the table, presence of ZnO nanoparticles onto the synthesized membrane following phase inversion causes increased surface charge of the membrane. The greater the presence of nanoparticles within the membrane structure, the larger the positive number claimed by zeta potential. Based on what is discussed further, the surface charge of membrane influences the behavior filtration and fouling of nanocomposite membranes.

Table 2 The surface charge of the solutions of nanoparticles and the synthesized membranes

Membrane type or solution	Zeta potential (mV)
ZnO in NMP solution	−1.56
PSF in NMP solution	−2.1
ZnO & PSF in NMP solution	−5.52
PEG-blank membrane	−4
PEG-NCM0.5 membrane	+10.79
PEG-NCM2 membrane	+17.26
Activated sludge	−14

---

Therefore, with the surface charge results, it can be expected that with the elevation of pH and anionization of the BSA protein solution, absorption of protein grows on the surface.^37,38 The results obtained from FTIR test together with AFM images also confirmed this.

3.5. FTIR analysis

In order to prove the presence of nanoparticles in the synthesized nanocomposite membranes and identify the nature of the biofilm formed on the surface of membranes, FTIR analysis was done on the membrane surface. Fig. 6 demonstrates the FTIR analysis of the powder of ZnO nanoparticles, raw membrane, and composite membrane.

Fig. 6 FTIR spectra of ZnO nanoparticles, the blended nanocomposite membrane and the blank membrane without nanoparticles.

As can be observed in the graphs plotted in Fig. 6, at wave numbers of 3500 cm⁻¹ and 500 cm⁻¹ in the ZnO nanoparticles graph, absorption can be seen which is related to the hydroxyl groups (OH) and the Zn–O bonds.³⁹ Absence of these peaks in the spectrum related to the raw membrane and its presence in the graph of nanocomposite membranes confirm the presence of ZnO nanoparticles in the structure of the nanocomposite membrane.

To examine the presence of biofilm on the membrane surface at different acidities, FTIR was also conducted on the membrane surface after the filtration of BSA solution. As can be seen in Fig. 7, in the FTIR spectrum, following the filtration operation of BSA protein, extra peaks can be observed at various pHs. Comparison of these peaks with the spectrum available for BSA in references⁴⁰ confirms BSA deposits on the membrane surface.

Fig. 7 FTIR spectra of BSA, nanocomposite membrane after filtration of BSA solution at pH = 3.5 and pH = 8.5.

A characteristics peak at the wave number of 3000 cm⁻¹ in the spectrum related to BSA, also observed in the spectrum of the membrane tested at pH = 3.5, is associated with the NH group of BSA. In a similar vein, the peaks at the wave number of 1700 cm⁻¹ and 1550 cm⁻¹ are related to CO, and CC groups⁴¹ of BSA, respectively.

The notable point in the diagram of Fig. 7 is greater transmission (lower absorbance) for characteristic peaks at pH = 3.5 in comparison with filtration at pH = 8.5. It seems that at higher acidity levels, the extent of BSA deposit is greater on the membrane surface. This can be due to altered surface charge of BSA in response to the changes in acidity and the tendency to greater absorption on the membrane surface in response to electrostatic attraction. In fact, at different pHs, the thickness of biofilm deposited on the membrane changes, resulting in different spectra for the membrane after the filtration of BSA solution.

3.6. AFM analysis

To investigate the surface roughness of the synthesized nanocomposites and the roughness of the biofilm developed on the nanocomposites, AFM images were used. AFM images indicated that presence of nanoparticles causes increased surface roughness of membrane. Furthermore, increased concentration of ZnO nanoparticles within the membrane structure was followed by greater roughness for the membrane. This can be owing to presence of nanoparticles within the membrane surface during the phase inversion process, which is itself caused by the hydrophilicity of nanoparticles and their tendency to be present in the coagulation bath.^13,30,42 Therefore, according to Fig. 8, in response to increased presence of nanoparticles, a membrane with a rougher surface is achieved. It is expected that during the filtration process more biofilm be placed on such a rough membrane.

Fig. 8 3D images of AFM from the surface of (A) raw polysulfone membranes and nanocomposite membranes prepared through blending method containing (B) 0.5 and (C) 2 wt% ZnO nanoparticles.

Moreover, the morphology of the biofilm deposited on the membrane surface was studied by AFM analysis. Indeed, the effect of the extent of nanoparticles in the membrane structure (Fig. 9) as well as the effect of the acidity of filtration environment (Fig. 10) on the morphology of the biofilm formed on the membrane surface were examined by AFM analysis.

Fig. 9 3D images of AFM from the morphology of the biofilm of activated sludge formed on the surface of nanocomposite membranes prepared through blending method containing (A) 0.3 and (B) 0.5 wt% ZnO nanoparticles.

Fig. 10 3D images of AFM from the morphology of the biofilm of the BSA formed on the surface of nanocomposite membranes prepared through blending method containing 0.5 wt% ZnO nanoparticles at pH = 3 (A), and pH = 8.3 (B).

As can be observed in Fig. 9, increased presence of nanoparticles through the blending method in the membrane structure results in decreased roughness of the biofilm developed on the membrane surface. In other words, as the presence of nanoparticles intensifies, less biofilm is developed on the membrane surface. This can be due to the hydrophilicity of the membrane surface in response to the presence of nanoparticles, inhibiting deposition of biofilm on the membrane surface.

According to Fig. 10, with the increase in the acidity of the filtration environment, the roughness of the biofilm formed on the membrane surface grows. This can be owing to greater deposition of soluble particles on the membrane surface or greater porosity of the biofilm deposited on the membrane surface.⁴³ As was observed in FTIR results, the extent of particles deposited on the membrane surface is greater at pH = 8.3 than at pH = 3.

3.7. The antibacterial test

In order to evaluate the antibacterial properties of the synthesized nanocomposite films, antibacterial test was done based on the method described in Section 2.3.7. Fig. 11 demonstrates the images of plates containing E. coli-k12. The inhibition zone around different membrane samples has been specified by an arrow on which the diameter of inhibited region has been written (as shown in Fig. 11). Considering the diameter of samples as 1.5 cm, the size embedded in the figure minus 1.5 cm presents the inhibition zone dimension. It shows that as the extent of nanoparticles increases, so does the antibacterial property. It also suggests that the inhibitions zone becomes wider. This phenomenon can be clearly seen in the Fig. 11 for PEG-NCM0.5 and PEG-NCM2 membranes.

Fig. 11 The antibacterial test of the membranes synthesized by E. coli bacteria (1) the PEG-NCM0 blank membrane, (2) PEG-NCM0.3 membrane, (3) PEG-NCM0.5 membrane, (4) PEG-NCM1 membrane, (5) PEG-NCM2 membrane, and (6) polysulfone membrane containing alumina nanoparticles in the previous research.

When explaining the antibacterial property of the synthesized nanocomposite, ROS molecules should be mentioned. Reactive Oxygen Species (ROS) are active molecules containing oxygen that have oxygen or peroxide ion. They are the byproduct of the production of oxygen through cellular metabolism. When microorganisms are exposed to environmental stressors such as UV ray and heat, the level of their produced ROS increases. As a result, the cells are no longer able to protect themselves against ROS by enzymes and antioxidants, thereby causing cellular death. Metal oxides such as zinc oxide and their nanoparticles liberate metal ions, thereby increasing the production of ROS by cells. This results in their antibacterial properties, but the precise mechanism of the effect of metal oxide nanoparticles in production of ROS is still unknown.¹⁶

On the other hand, metal nanoparticles especially ZnO lead to increased permeability of the membrane of bacterial cell. This causes defects in the membrane transfer system, in response to which nanoparticles enter the cell. It thus affects enzymes, proteins, nucleic acids especially DNA, causing its destruction. Furthermore, ZnO nanoparticles prevent attachment of bacteria to the surface of host cell and formation of biofilm.³⁴ Other researchers have reported the same behavior; Zodrow et al. observed more than 90% reduction in E. coli grown on the membrane surface during dead end filtration by nanocomposite membrane containing Ag nanoparticles.¹⁷

The results indicated that the membranes that have nanoparticles in a combinational form in their structure enjoy antibacterial properties. The comparison between the membranes synthesized in previous studies of this research group^42,44 and the current research revealed that the membranes containing ZnO nanoparticles had a greater antibacterial property in comparison with the membranes containing alumina nanoparticles (sample 6 in Fig. 11).

3.8. Measurement of the pure water flux of the synthesized nanocomposite membranes

The results of the pure water flux through the synthesized nanocomposite membranes are presented in Fig. 12. According to the diagram, the PEG-NCM0.5 has the best flux. Based on these results and another research done by this group, the ratio of 0.5% nanoparticles to polymer can be considered as the threshold level.⁴² With the increase in the amount of nanoparticles for the nanocomposites synthesized below the threshold level, membrane coagulation occurs faster in the coagulation bath due to the hydrophilicity of the nanoparticles. As a result, the size of pores grows and thus greater flux, as 503 L h⁻¹ m⁻², can be seen. However, for the membranes synthesized above the threshold level, owing to increased viscosity of the polymer solution, the coagulation rate declines and thus the size of pores together with flux also diminish.

Fig. 12 The effect of elevation of the nanoparticle concentration onto the structure on the flux of pure water.

Other researchers have, also seen high pure water flux for membrane in the presence of other nanoparticles. The pure water flux was reported as 750 L h⁻¹ m⁻² for PSf/Al₂O₃,⁴⁴ 60.6 L h⁻¹ m⁻² for PSf/rGO/TiO₂ (ref. 45) and 130 L h⁻¹ m⁻² for polysulfone/GO/TiO₂ (ref. 46) nanocomposite membrane. For PES/cobalt⁴⁷ and PES/rGO/TiO₂ (ref. 48) nanocomposite membrane, pure water flux was measured as 160.07 and 420 L h⁻¹ m⁻², respectively.

3.9. Investigation of the performance of nanocomposite membranes during filtration of BSA solution

Based on the results obtained from filtration of pure water, the nanocomposite membrane of PEG-NCM0.5 has had the greatest flux. The filtration of BSA solution also indicated that the PEG-NCM0.5 membrane has had the lowest membrane resistance, the lowest total resistance, and the lowest biofilm resistance (Table 3). The total, biofilm, pore, and membrane resistance for PEG-NCM0.5 and PEG-NCM2 were the lowest at pH = 8.9.

Table 3 The resistances present for synthesized membranes in the filtration of BSA protein

pH	Membrane type	Rt × 10^11 (m^−1)	Rm × 10^11 (m^−1)	RP × 10^11 (m^−1)	RC × 10^11 (m^−1)
3.5	PEG-blank	121.3	17.32	75.98	28
	PEG-NCM0.5	60.65	8.08	35.32	20.22
	PEG-NCM2	80.86	9.33	28.57	42.96
7	PEG-blank	105.47	50.54	30.32	24.61
	PEG-NCM0.5	48.52	7.58	26.11	14.83
	PEG-NCM2	71.35	8.08	24.7	38.57
8.3	PEG-blank	93.3	43.32	28.03	21.95
	PEG-NCM0.5	43.32	6.38	25.12	11.82
	PEG-NCM2	45.94	8.66	19.81	17.47
8.9	PEG-blank	121.3	48.52	27.29	45.49
	PEG-NCM0.5	37.9	5.77	22.43	9.7
	PEG-NCM2	40.43	7.58	19.07	13.78

---

Based on the information related to BSA protein in literature,^37,38 as was mentioned previously, with the increase of pH up to around 4.7, the surface charge of BSA protein declines, reaches zero, and then becomes negative. However, according to the results of zeta potential, in the synthesized membranes, as the amount of nanoparticles increased within the structure of polymer membrane, the membrane surface charge moved towards positive numbers (+10.79 for the membrane containing 0.5% nanoparticles and +17.26 for the membrane containing 2% nanoparticles). At higher acidity levels, the extent of negative charge of the protein grows, followed by intensification of the attraction between the protein and the positive surface of the nanocomposite membrane. This results in increased deposition of BSA on the membrane surface, previously confirmed by FTIR results in Fig. 7. On the other hand, as the acidity grows, the flux of the membrane increases, but the resistances of filtration decrease. Similar to these results have been expressed by Priyananda et al., where the porosity of BSA deposited on polysulfone membrane increases with the development of the absolute value of zeta potential of BSA protein.⁴⁹ Fig. 13 demonstrates the effect of pH on the variations of flux for the three tested membranes.

Fig. 13 The effect of acidity on the flux of BSA solution of the membranes.

It appears that at higher acidity levels, although the amount of protein deposited on the membrane surface grows, the structure of the biofilm deposited on the membrane surface, in terms of porosity, is such that the membrane flux drops less. AFM image of PEG-NCM0.5 membrane together with the protein biofilm on it at pHs = 3.5 and 8.3 in Fig. 10 indicated that the biofilm developed at pH = 8.3 is rougher than at pH = 3.5, with higher roughness representing greater porosity of the biofilm.⁴³ Therefore, due to the negative surface charge of BSA particles and the positive surface of membrane, the arrangement of the biofilm on the surface of nanocomposite membrane is such that it provides a path for the passage of feed flow through the membrane. However, at higher acidity levels, in spite of lower biofilm deposition on the membrane surface, fouling is higher, whereas for a raw membrane, elevation of acidity up to 8.3 leads to increased flux. Indeed, the filtration resistances reach their lowest points for the PEG-blank membrane at pH = 8.3. After this pH at pH = 8.9, however, the filtration resistance for this membrane grows. This shows that acidity in the raw membrane and the amount of deposited biofilm are directly related, but the arrangement of the biofilm on the surface of this membrane does not have a desirable structure and brings about increased fouling.

Overall, the best antifouling performance has been related to the PEG-NCM0.5 membrane at pH = 8.9. The best reduction in the pore resistance occurs at pH = 3.5, showing 58% reduction. This phenomenon can be attributed to the repulsion between the positive charge of the protein and the membrane surface. The membrane resistance indicated the greatest reduction at pH = 8.9 by 88%. Further, the biofilm resistance at pH = 8.9 revealed 79% reduction, since at higher pHs, the protein becomes more anionic, and due to the structure of the obtained biofilm, lowered fouling is followed. In sum, the total resistance showed 69% reduction at pH = 8.9.

3.10. Investigation of the filtration performance of nanocomposite membranes in a membrane bioreactor

The performance of the filtration of the synthesized nanocomposite membranes in the membrane bioreactor was investigated. According to Table 4, obtained through the test of some samples of the nanocomposite membranes in the membrane bioreactor, it can be concluded that expectedly based on filtration of pure water, the PEG-NCM0.5 has the lowest membrane resistance. Furthermore, the total resistance and biofilm resistance are also lowest for this membrane. This has also been shown in AFM images in Fig. 9, in which the biofilm of the sludge formed on the PEG-NCM0.5 is less than on the PEG-NCM0.3.

Table 4 The resistances present in the filtration of activated sludge for the membranes

Membrane type	Rt × 10^11 (m^−1)	Rm × 10^11 (m^−1)	RP × 10^11 (m^−1)	RC × 10^11 (m^−1)
PEG-blank	242.6	202.1	22.5	18
PEG-NCM0.15	202.16	121.3	65.3	15.56
PEG-NCM0.3	121.3	93.3	16.9	8.06
PEG-NCM0.5	93.3	80.86	5.74	6.7
PEG-NCM1.5	173.28	93.3	7.7	77.02

---

Finally, for the membrane containing 0.5 wt% nanoparticles, the total resistance, biofilm resistance, membrane resistance, and pore resistance in the filtration of activated sludge showed 61.55, 62.76, 60, and 74.5% reduction respectively in comparison with the raw membrane.

In fact the presence of 0.5 wt% of ZnO in the polysulfone matrix results in total resistance of 93.3 × 10¹¹ m⁻¹. For Al₂O₃/PSf nanocomposite membrane in 0.03 wt% of alumina, total resistance was reported as 120 × 10¹¹ m⁻¹.⁴⁴ For Fe₃O₄/PSf nanocomposite membrane total filtration resistance in MBR system was reported about 50 × 10¹¹ m⁻¹.⁵⁰

According to the obtained data, it can be concluded that in the activated sludge filtration system, the biofilm and total resistance are influenced by the protein compounds present in the sludge, where the representative of these compounds is BSA protein, confirming this issue. The membrane resistance is lowest in both the sludge system and the BSA solution for the membrane containing 0.5% nanoparticles. The reason of the dramatic increase in the biofilm resistance for the PEG-NCM1.5 membrane can be attributed to the electrostatic interactions between the membrane and sludge. This is because with the increase in the amount of nanoparticles, the surface charge of the membrane becomes a positive number with a larger absolute value. Based on Table 2, this number should be within the range of +10 and +17 mV. Further, according to that very table, the surface charge of active sludge has been reported to be −14 mV. Hence, the sum of these charges approaches zero, causing absorption of active sludge on the membrane and formation of resistant biofilm.

Table 5 reports the level of COD removed by the membranes tested in the bioreactor. The extent of removal has grown up to an order of 1.69. As can be observed in the data reported in Table 5, the maximum COD removal is obtained for the PEG-NCM0.5 membrane (about 88%) which shows 36% growth in comparison with raw membrane.

Table 5 The percentage of COD removal

Membrane type	COD removal%
PEG-blank	52
PEG-NCM0.15	53
PEG-NCM0.3	62
PEG-NCM0.5	88
PEG-NCM1.5	74

---

3.11. Examination of the structure of the biofilm formed on the surface of membrane in a membrane bioreactor with the changes in pH

The structure of the biofilm developed on PEG-blank, PEG-NCM0.5, and PEG-NCM2 membranes in the membrane bioreactor was investigated at different pH levels of 3.7 (acidic), 8.3 (basic), and natural pH of activated sludge (pH = 6.5). The results are provided in Fig. 14.

Fig. 14 The effect of pH on the biofilm of the sludge formed on the membranes.

According to Fig. 14, the level of biofilm formed on the surface of PEG-NCM0.5 membrane is the lowest, where based on the data reported in Table 4, this variety of membrane had also the lowest biofilm resistance. In addition, the amount of the biofilm developed on the PEG-NCM2 membrane indicates that at alkaline pHs, more biofilm has deposited on the nanocomposite membrane containing 2 wt% nanoparticles, owing to the positive surface charge of this membrane. On the other hand, according to the images of Fig. 14, at the first glance, the amount of biofilm with the increase in pH is the same as the amount developed on the raw membrane. Nevertheless, the table of resistances (Table 4) shows that the resistance developed on the PEG-NCM2 membrane is lower, compared with the raw membrane. Based on AFM images (Fig. 9), this can be due to the porosity of the biofilm formed on the surface of PEG-NCM2 nanocomposite membrane, where it has greater flux during the filtration of activated sludge. This also can be a proof that activated sludge benefits from a greatest share of the soluble proteins in comparison with polysaccharide compounds and NOM compounds.¹¹ In the research by Priyananda et al., it has been stated that the deposited layer is porous when the concentration of BSA solution is low, i.e. lower than the gel forming concentration. This porous deposit is also labile.⁴⁹ This ability was observed during simple washing of the biofilm layer off the membrane surface.

In total, in this study, positive charge of the membrane surface was used for engineering of biofilm formed on the membrane surface. In fact attractive force at alkaline pH and repulsive force at acidic pH manage particle deposition on membrane surface. Eventually at alkaline pHs deposition of substance on the membrane surface increase and on the other side the porosity of formed layer enhances. In fact the denser biofilm forms at pH of iso electric point. Any increase in pHs results in the porous film and any decrease in pH deduces to the thin film formation. These changes in biofilm structure at different pHs is schematically shown in Fig. 15.

Fig. 15 The schematic illustration from engineering of biofilm formed on ZnO/PSf nanocomposite membrane.

As seen in the Fig. 15, the interaction between membrane surface and BSA molecule influence biofilm formed on the membrane surface. After biofilm formation on the membrane surface, other filtration steps is performed in the presence of an engineered and pH sensitive membrane surface.

Altogether, out of the obtained results in this study, it was observed that the presence of nanoparticles could modify the surface property (hydrophilicity and roughness) and structure of bio film formed on the membrane, thereby manipulating the membrane filtration performance. The obtained results in this study were compared with a sample of commercial membrane (UP020, MICRODYN-NADIR, Germany), with the results provided in Fig. 16.

Fig. 16 The comparison between surface properties and filtration performance of blank membrane, nanocomposite membrane (PEG-NCM0.5) and MICRODYN-NADIR commercial membrane.

On the whole, from the data reported in Fig. 16, the nanocomposite membrane has a higher surface hydrophilicity and lower surface roughness rather than both blank membrane and commercial membrane. The nanocomposite membrane with modified surface properties is expected to have better filtration performance, compared to blank and commercial counterparts.

PEG-NCM0.5 has 52% and 140% greater pure water flux rather than blank and Nadir membranes, respectively. The comparison of the performance of membranes in the MBR system shows that the membrane flux in the bioreactor medium (MLSS = 7) is about 130 L h⁻¹ m⁻² for nanocomposite membrane, which is 44% higher than Nadir membrane. Considering the COD removal, the PEG-NCM0.5 nanocomposite membrane has a removal efficiency of about 38% more than Nadir membrane.

4. Conclusion

In this research, polysulfone ultrafiltration nanocomposite membranes containing ZnO nanoparticles synthesized through phase inversion method were investigated in the filtration of BSA protein solution. In the conducted experiments, the effect of pH of the protein solution and the weight percentage of ZnO nanoparticles on the developed biofilm was investigated. Eventually, optimal conditions for formation of a porous biofilm were obtained, whereby at a certain pH and with a specific amount of nanoparticles, the flux of protein solution becomes greater and the resistances against the filtration are lowest.

Considering the positive surface charge of the synthesized membranes and the negative surface charge of the protein at pHs above the isoelectric point (pH = 4.7), the extent of electrostatic absorption between the membrane and protein grows. The higher the weight percentage of nanoparticles in the membrane, the greater the level of surface charge and zeta potential, thereby increasing the protein absorption. The FTIR and AFM results also proved this issue. However, the filtration results indicate greater flux for the membrane at higher pHs, where based on previous studies, this behavior can be attributed to greater porosity developed for the formed biofilm. Hence, in the nanocomposite membranes fabricated with positive surface charge, the acidity of the BSA protein solution and the resistance of the formed biofilm are directly related. Further, the thickness and roughness of the biofilm layer have also increased, suggesting development of further biofilm. We, however, experience increased flux, owing to the porosity of the biofilm developed at higher acidity levels. Therefore, in this research through engineering the conditions of nanocomposite fabrication and changing the acidity of the environment containing protein, we achieved conditions in which despite increased thickness of the biofilm, the flux passing through the nanocomposite membrane is larger. This behavior occurs at pH = 8.9 for the PEG-NCM0.5 membrane. Moreover, the filtration of activated sludge in the MBR system revealed that the behavior of the developed biofilm at different pHs is similar to that of the BSA protein biofilm, implying greater portion of protein in activated sludge in comparison with other compounds. The results indicated that with the increase in pH, although the extent of activated sludge biofilm developed on the surface of nanocomposite membrane grows, since it has a greater porosity, it brings about greater flux as well.

References

1. P. M. Ajayan, L. S. Schadler and P. V. Braun, Nanocomposite science and technology, John Wiley & Sons, 2006.
2. A. Bhogale, N. Patel, P. Sarpotdar, J. Mariam, P. M. Dongre, A. Miotello and D. C. Kothari, Colloids Surf., B, 2013, 102, 257–264.
3. A. Asatekin, A. Menniti, S. Kang, M. Elimelech, E. Morgenroth and A. M. Mayes, J. Membr. Sci., 2006, 285, 81–89.
4. P. Kanagaraj, A. Nagendran, D. Rana, T. Matsuura and S. Neelakandan, Int. J. Biol. Macromol., 2015, 72, 223–229.
5. J. Garcia-Ivars, M.-I. Alcaina-Miranda, M.-I. Iborra-Clar, J.-A. Mendoza-Roca and L. Pastor-Alcañiz, Sep. Purif. Technol., 2014, 128, 45–57.
6. A. L. Ahmad, A. A. Abdulkarim, B. S. Ooi and S. Ismail, Chem. Eng. J., 2013, 223, 246–267.
7. S. Mafirad, M. Mehrnia and M. Sarrafzadeh, Water Sci. Technol., 2011, 64, 1154–1160.
8. J. H. Jhaveri and Z. V. P. Murthy, Desalination, 2016, 379, 137–154.
9. B. P. Tripathi, N. C. Dubey, R. Subair, S. Choudhury and M. Stamm, RSC Adv., 2016, 6, 4448–4457.
10. G. Arthanareeswaran, D. Mohan and M. Raajenthiren, J. Membr. Sci., 2010, 350, 130–138.
11. P. Le-Clech, V. Chen and T. A. G. Fane, J. Membr. Sci., 2006, 284, 17–53.
12. J. Mansouri, S. Harrisson and V. Chen, J. Mater. Chem., 2010, 20, 4567–4586.
13. M. Homayoonfal, M. R. Mehrnia, Y. M. Mojtahedi and A. F. Ismail, Desalin. Water Treat., 2013, 51, 3295–3316.
14. C. Carl, A. J. Poole, M. J. Vucko, M. R. Williams, S. Whalan and R. de Nys, Biofouling, 2012, 28, 1077–1091.
15. T.-H. Bae and T.-M. Tak, J. Membr. Sci., 2005, 249, 1–8.
16. A. M. C. Ng, C. M. N. Chan, M. Y. Guo, Y. H. Leung, A. B. Djurišić, X. Hu, W. K. Chan, F. C. C. Leung and S. Y. Tong, Appl. Microbiol. Biotechnol., 2013, 97, 5565–5573.
17. K. Zodrow, L. Brunet, S. Mahendra, D. Li, A. Zhang, Q. Li and P. J. Alvarez, Water Res., 2009, 43, 715–723.
18. A. Mollahosseini and A. Rahimpour, Biofouling, 2013, 29, 537–548.
19. R. Zhang, Y. Liu, M. He, Y. Su, X. Zhao, M. Elimelech and Z. Jiang, Chem. Soc. Rev., 2016, 45, 5888–5924.
20. A. Akbari and M. Homayoonfal, Korean J. Chem. Eng., 2016, 33, 2439–2452.
21. M. Al-Fori, S. Dobretsov, M. T. Z. Myint and J. Dutta, Biofouling, 2014, 30, 871–882.
22. A. Pal, T. K. Dey, A. Singhal, R. C. Bindal and P. K. Tewari, RSC Adv., 2015, 5, 34134–34151.
23. S. Javdaneh, M. R. Mehrnia and M. Homayoonfal, Korean J. Chem. Eng., 2016, 33, 3184–3193.
24. S. Balta, A. Sotto, P. Luis, L. Benea, B. Van der Bruggen and J. Kim, J. Membr. Sci., 2012, 389, 155–161.
25. M. Malini, M. Thirumavalavan, W.-Y. Yang, J.-F. Lee and G. Annadurai, Int. J. Biol. Macromol., 2015, 80, 121–129.
26. S. Salgın, U. Salgın and S. Bahadır, Int. J. Electrochem. Sci., 2012, 7, 12404–12414.
27. M. Homayoonfal and M. R. Mehrnia, Sep. Purif. Technol., 2014, 130, 74–83.
28. R. Habibi, J. T. Daryan and A. M. Rashidi, J. Exp. Nanosci., 2009, 4, 35–45.
29. M. Homayoonfal, A. Akbari and M. R. Mehrnia, Desalination, 2010, 263, 217–225.
30. M. Homayoonfal, M. R. Mehrnia, M. Shariaty-Niassar, A. Akbari, A. F. Ismail and T. Matsuura, Desalination, 2014, 354, 125–142.
31. Y. M. Mojtahedi, M. R. Mehrnia and M. Homayoonfal, Desalin. Water Treat., 2013, 51, 6736–6742.
32. S. T. Odonkor and J. K. Ampofo, Microbiol. Res., 2013, 4, e2.
33. S. Edberg, E. Rice, R. Karlin and M. Allen, J. Appl. Microbiol., 2000, 88, 106S–116S.
34. D. Shailaja, Int. J. Bioassays, 2013, 2, 914–917.
35. M. J. Martínez-Morlanes, A. M. Martos, A. Várez and B. Levenfeld, J. Membr. Sci., 2015, 492, 371–379.
36. Y. Ma, F. Shi, Z. Wang, M. Wu, J. Ma and C. Gao, Desalination, 2012, 286, 131–137.
37. T. Xu, R. Fu and L. Yan, J. Colloid Interface Sci., 2003, 262, 342–350.
38. V. L. Vilker, C. K. Colton and K. A. Smith, J. Colloid Interface Sci., 1981, 79, 548–566.
39. M. Andres-Verges and C. Serna, J. Mater. Sci. Lett., 1988, 7, 970–972.
40. B. Mukherjee, K. Santra, G. Pattnaik and S. Ghosh, Int. J. Nanomed., 2008, 3, 487.
41. N. Brandes, P. B. Welzel, C. Werner and L. W. Kroh, J. Colloid Interface Sci., 2006, 299, 56–69.
42. M. R. Mehrnia, Y. M. Mojtahedi and M. Homayoonfal, Desalination, 2015, 372, 75–88.
43. W. S. Ang and M. Elimelech, J. Membr. Sci., 2007, 296, 83–92.
44. M. Homayoonfal, M. R. Mehrnia, S. Rahmani and Y. M. Mojtahedi, J. Ind. Eng. Chem., 2014, 22, 357–367.
45. M. Safarpour, V. Vatanpour, A. Khataee and M. Esmaeili, Sep. Purif. Technol., 2015, 154, 96–107.
46. M. Kumar, Z. Gholamvand, A. Morrissey, K. Nolan, M. Ulbricht and J. Lawler, J. Membr. Sci., 2016, 506, 38–49.
47. L. Gzara, Z. Ahmad Rehan, S. B. Khan, K. A. Alamry, M. H. Albeirutty, M. S. El-Shahawi, M. I. Rashid, A. Figoli, E. Drioli and A. M. Asiri, J. Taiwan Inst. Chem. Eng., 2016, 65, 405–419.
48. V. Vatanpour, A. Shockravi, H. Zarrabi, Z. Nikjavan and A. Javadi, J. Ind. Eng. Chem., 2015, 30, 342–352.
49. P. Priyananda and V. Chen, J. Membr. Sci., 2006, 273, 58–67.
50. M. R. Mehrnia and M. Homayoonfal, J. Membr. Sci., 2016, 520, 881–894.

---