In situ surface modification of forward osmosis membrane by polydopamine/polyethyleneimine-silver nanoparticle for anti-fouling improvement in municipal wastewater treatment

Increasing water scarcity and resource depletion with the growing world population become major challenges in the 21st century. To address these problems, membrane-based technology has unique advantages by recovering fresh water and nutrients from municipal wastewater (Yusuf et al. 2020). Forward osmosis (FO) is a new burgeoning osmotic pressure-driven membrane separation process, which has obtained wide attention and various applications in wastewater treatment/reclamation, seawater desalination and power generation (Lee & Hsieh 2019; Suwaileh et al. 2020). FO relies on the natural osmotic pressure gradient across the membrane to drive water molecules in feed solution (FS) transferring to draw solution (DS) (Goh et al. 2019). As a state-of-the-art membrane technology, it is ascribed to its no need for hydraulic pressure, lower membrane fouling propensity as well as high water flux recovery (Wang & Liu 2021).

In spite of its advantages and wide applications, membrane fouling, especially biofouling, remains a major issue for FO, which is ascribed that the membrane is in contact with complex raw wastewater directly during long-term operation (Sun et al. 2018; Bao et al. 2019; Jung et al. 2020). The foulants on the membrane surface give rise to a decrease in water permeability, an increase in the frequency of cleaning and a shortened membrane life (She et al. 2016; Ibrar et al. 2019). Therefore, various methods are proposed to alleviate membrane fouling, including the pretreatment of FS (Pramanik et al. 2019), optimization of operation parameters (Sun et al. 2019), cleaning strategies of membranes (Lotfi et al. 2018; Kim et al. 2020) and modification of membrane (Ibrar et al. 2019; Firouzjaei et al. 2020).

The membrane surface properties play a fundamental role in membrane fouling control, thus efficient modifications to enhance the fouling resistance of membranes are in high demand (Xu et al. 2017). In general, a hydrophilic membrane surface can improve water flux and reduce the adhesion of organic and biological substances so as to enhance the fouling resistance of the membrane surface (Goh et al. 2019). Inspired by mussels, the research of dopamine (DA) in forming multi-functional coating on the membrane surface has attracted much attention (Qi et al. 2018; Song et al. 2018). Under alkaline conditions, dopamine can form a material-independent polydopamine (PDA) coating through self-polymerizing cross-linking (Yang et al. 2015). The presence of functional groups in PDA, such as amine, catechol and imine, improves the hydrophilicity of the membrane surface and decreases the adsorption of organic foulants and microorganisms on the membrane surface (Karkhanechi et al. 2014). However, the aggregates are liable to form in the process of self-polymerization, which leads to an inhomogeneous coating and increases the roughness of the membrane surface. In addition, the stability of the PDA layer is insufficient (Yang et al. 2016a). Polyethyleneimine (PEI), with superior chemical and thermal stability, is a functional polymer containing abundant amino groups, which can react with PDA through Michael addition and Schiff base reaction, resulting in a smooth and stable PDA/PEI coating in an alkaline environment (Yang et al. 2014).

Nevertheless, biofouling cannot be mitigated effectively by the enhancement in membrane hydrophilicity due to the self-replicating ability of live bacteria and the formation of biofilm on the membrane surface (Qi et al. 2018; Firouzjaei et al. 2020). Modifying the membrane surface with effective biocides is therefore thought to be a good alternative to endow the membrane surface with excellent antibiofouling activity. Most biocides can be classified into three main categories based on different molecular structures: inorganic biocides, organic biocides and organic–inorganic compound biocides. Metal ions such as silver, copper and zinc ions as well as metal oxides like zinc oxide, titanium dioxide and copper oxide are the most common inorganic biocides (Zhu et al. 2018). The advantages of inorganic biocides include long-lasting antibacterial effects, a wide range of applications, high heat resistance and easy processing. Nevertheless, some inorganic biocides are usually expensive and do not achieve the desired antimicrobial activity immediately when initially applied. Organic biocides mainly include natural antimicrobial materials such as chitosan and lysozyme and synthetic antimicrobial materials like polyquaterniums and polyquaterniums (Gerba & Müller 2015). Organic biocides tend to have a better bactericidal effect compared to inorganic biocides, while at the same time, they also have obvious disadvantages, such as poor heat resistance, easy decomposition and poor antimicrobial durability. The composite biocides containing both organic and inorganic antimicrobial materials combine the advantages of both organic and inorganic biocides, but its preparation process is relatively complex (Seyedpour et al. 2019). Silver nanoparticles (Ag NPs), exhibiting a broad-spectrum antibacterial activity even at low concentrations, were widely employed to be embedded in the membrane surface, which renders the membrane direct, effective and long-lasting antibacterial effect (Zhang et al. 2016a). Qiu & He (2018) obtained a long-term antibiofouling membrane via second interfacial polymerization of zwitterion and in situ synthesis of Ag NPs by sodium borohydride (NaBH₄). However, the preparation of Ag NPs involves additional chemical agents and complicate synthesis procedures, which are time-consuming, cost-ineffective and environment-unfriendly (Zhang et al. 2016a). From this viewpoint, PDA/PEI coating layer has enormous potential to reduce the coated silver ions into silver nanoparticles in situ due to the strong metal adsorbing and chelating capacities of active groups on PDA/PEI and the great reducibility of dopamine monomer (Yang et al. 2016b).

In this study, cellulose triacetate (CTA) FO membrane was coated by PDA/PEI layer to improve the hydrophilicity of the membrane surface and then in situ reduction of Ag NPs onto the PDA/PEI coating was aimed to obtain high antibacterial properties. After the modification, the surface morphology, surface chemical composition, surface roughness, surface hydrophilicity and mass transfer performance of membranes were systematically characterized. Besides, the antiadhesion and antibacterial properties of modified membranes against Escherichia coli and Staphylococcus aureus were experimentally evaluated. Finally, the performance of the modified membrane in the treatment of municipal wastewater was studied and the antifouling mechanism of the modified membrane was revealed by analyzing the type and structure of the fouling layer on the membrane surface.

A CTA membrane was supplied by Hydration Technology Innovations (HTI) as the basis matrix for modification. Dopamine (DA, 98%), polyethyleneimine (PEI, Mw = 600) and tri (hydroxymethyl) amino methane hydrochloride (Tris–HCl, > 99%) were received from Aladdin Reagent Co., Ltd (Shanghai, China) to prepare PDA/PEI coating layer. Silver nitrate (AgNO₃) from Sigma-Aldrich (China, Shanghai) was used to form Ag NPs in situ. Gram-negative E. coli (ATCC25922) and Gram-positive Staphylococcus aureus (S. aureus, CMCC(B)26003) provided by Luwei Technology Co., Ltd (Shanghai, China) were used for antiadhesion and antibacterial experiments. All reagents in this study used were of analytical grade and deionized water was used in all experiments.

DA (0.2 g) and PEI (0.3 g) were dissolved in Tris–HCl solution (100 mL, pH = 8.5). The nascent CTA FO membrane was placed on a customized membrane holder with an active layer facing the clean glass plate. This PDA/PEI solution then was poured into the membrane holder with shaking at 50 rpm for 24 h at room temperature. After reaction, the excessive PDA/PEI solution was removed. The modified membrane was rinsed with DI water and stored in sterile DI water at 4 °C before use. Hereon, the PDA/PEI coated CTA membrane will be designed as PDA/PEI CTA membrane.

Surface morphologies of the virgin membrane, modified membranes and fouled membranes were identified by scanning electron microscopes (SEM, Model XL 30, Philips, Netherlands) equipped with an energy-dispersive X-ray spectroscope (EDX). The Attenuated Total Reflection-Fourier transform infrared spectroscopy (ATR-FTIR, Spectrum One B, Perkin Elmer Inc, American) analysis was carried out to identify the major functional groups on the surface of virgin, modified and fouled membranes. The chemical composition changes between CTA nascent membrane and modified membranes were characterized by X-ray photoelectron spectroscopy (XPS, K-Alpha, Al Kα radiation). Atomic force microscopy (AFM, BioScope, Veeco Inc.) was employed to measure the membrane surface roughness of the CTA, PDA/PEI CTA and PDA/PEI-nAg CTA membranes and obtain 3D morphological images. The nascent and modified membranes’ surface hydrophilicity was evaluated through a water contact angle performed by a contact angle goniometer (SL200B, Solon Tech Co., Ltd, China). The structure and composition of the fouling layer were observed by confocal laser scanning microscopy (CLSM, Nikon C2). The fouled membrane pieces of 0.5 × 0.5 cm² were stained with SYTO 9, PI, fluorescein isothiocyanate (FITC), concanavalin A (Con A) and calcofluor white (CW) in series to label the living bacteria, dead bacteria, proteins, α-D-glucopyranose polysaccharides (α-polysaccharides) and β-D-glucopyranose polysaccharides (β-polysaccharides) in the fouling layer, respectively, and the three-dimensional images were obtained by NIS-Viewer software. The specific steps were provided in the previous literature (Sun et al. 2019).

Anti-adhesive tests were conducted to observe the attached cells on the fouled membranes by SEM. In brief, after sterilizing with an ultraviolet lamp for 15 min, small pieces (1 × 1 cm) of virgin membrane and modified membranes were kept in contact with 5 mL E. coli and S. aureus suspension (10⁸ ∼ 10⁹ CFU/mL), respectively, for 12 h at 37 °C, and then gently washed with phosphate-buffered solution (PBS, pH = 7.5) to remove the loosely attached bacteria. Chemical fixation pretreatment of the fouled membranes was then carried out by the method described by Xiong et al. (2016) to observe the adhesion of bacteria on the membrane surface.

The diffusion inhibition zone method was employed to evaluate the bacterial inhibition behaviors of the modified membranes. Briefly, 200 μL bacterial suspension (E. coli and S. aureus) was spread on an LB medium. The active layer of membrane coupons (the original membrane and the modified membranes) with a diameter of 10 mm was contacted with bacteria and incubated at 37 °C for 24 h. The size of the inhibition zone around the samples was determined to characterize the antibacterial activity.

The antibacterial property of the modified membrane was further evaluated by the colony-forming unit. A small piece (1 × 1 cm) of each membrane was kept in contact with 5 mL E. coli and S. aureus suspension, respectively, for 24 h at 37 °C while shaking at 100 rpm. The membranes were then taken out and washed with PBS to remove the free cells from the membrane surface. After rinsing, the membranes were immersed in 0.85% sterilized normal saline (5 mL) for 5 min with sonication and vortex to remove the attached bacteria from the membrane surface and dissolve in the solution. The suspension (1 mL) after serially diluted was inoculated in LB agar medium, and the colonies of viable bacteria on the membrane surface were counted after inoculating for 24 h at 37 °C.

The dynamic fouling experiment was performed by the laboratory-scale FO system with cross-flow side stream configuration run in parallel (Fig. S1). Co-current recirculation of the FS and DS was employed on each side of the membrane module (length × width × height = 10 cm × 4 cm × 0.3 cm) by two peristaltic pumps. The effective FO membrane area was 20 cm². In order to evaluate the antifouling performance of modified membranes in more realistic settings, raw municipal wastewater was adopted as a feed solution in the dynamic fouling experiment, which was obtained from the septic tank effluent in the adjacent residential area. Prior to use, municipal wastewater was filtered with a 200 mesh screen and the wastewater displayed the following main quality: 150 ∼ 280 mg/L chemical oxygen demand (COD), 35 ∼ 60 mg/L ammonia nitrogen (⁠-N), 38 ∼ 70 mg/L total nitrogen (TN) and 1 ∼ 4 mg/L phosphate (⁠-P) (more detailed information illustrated in Table S1). The standard synthetic seawater was prepared as a DS according to the detailed composition listed in Table S2.

The pristine and modified membranes were installed with the active layer directly facing the municipal wastewater (FS) and the support layer in contact with the synthetic seawater (DS). Firstly, a baseline experiment was carried out using DI as FS and 2 L synthetic seawater was used as DS for 24 h. Subsequently, the FS was replaced with raw municipal wastewater with avolume of 1 L. The system was operated for 7 days with across-flow velocity of 10 cm/s. A stirring device was adopted to provide mixing in the FS tank. And the DS tank was located on a digital balance (Sartorius, BSA3202S) in order to calculate the membrane flux based on the weight changes of the DS with the extension of the operation time. At the end of every day, the FS and DS were replaced with fresh ones to avoid significant changes in their concentrations. All experiments were carried out at room temperature. Water samples were collected from every day detected on the basis of the Chinese NEPB standard methods (CEPB 2002). The dynamic antifouling performance of the modified membrane in the treatment of raw municipal sewage was discussed by analyzing the flux change and the structure and composition of fouling layer on the membrane surface.

The elementary composition change of the CTA membrane surface after modification was further analyzed by XPS to elucidate the structure of the membrane surface precisely. As shown in Figure 3(b), after the PDA/PEI deposition, an absorption peak of N element (396.6 eV) emerged. It indicated that PDA/PEI with substantial amino groups was introduced to the membrane surface successfully. As for PDA/PEI-n Ag CTA membrane, not only the absorption peak of N element but also a new peak with the binding energy of 365.1 eV corresponding to Ag 3d appeared on the membrane surface (Yang et al. 2016b). This peak verified that the silver ions were fixed and then reduced to Ag NPs in situ successfully on the membrane surface by the catechol groups of the PDA/PEI coating layer.

In order to discuss the hydrophilicity of the virgin and modified CTA membrane, the measurement of water contact angles was conducted. As presented in Fig. S2, the water contact angle value dropped from 69.5° to 45.3° after modification by PDA/PEI coating, indicating an enhancement of surface hydrophilicity. It was attributed to the introduction of various hydrophilic functional groups by PDA/PEI coating, such as amino and hydroxyl functional groups. After the in situ reduction of AgNPs on the membrane surface which led to a decreased water contact angle, the hydrophilicity of the PDA/PEI-n Ag CTA membrane was further improved due to the hydrophilic properties of silver nanoparticles.

Bacteria tend to adhere to the membrane surface and use the surrounding nutrients to reproduce to form the biofilm on the membrane surface, thus affecting the operating performance of the membrane. Therefore, to improve the antifouling performance of the membrane, the rate of bacteria attached to the membrane surface should first be reduced. Figure S3 shows the adhesion of E. coli and S. aureus on the CTA membrane, PDA/PEI CTA membrane and PDA/PEI-n Ag CTA membrane determined by SEM.

As demonstrated in Fig. S3, the number of E. coli attached to modified CTA membranes was decreased in comparison with the pristine CTA membrane. Furthermore, the E. coli attached on the surface of the PDA/PEI-nAg CTA-modified membrane was significantly deformed. S. aureus almost covered evenly the surface of the pristine CTA membrane. In contrast, the number of S. aureus on the PDA/PEI CTA membrane surface was greatly reduced due to the improvement of hydrophilicity. However, by SEM observation, the number of S. aureus attached to the PDA/PEI CTA membrane surface was not remarkably lower due to the higher roughness. The results revealed that the better bacterial adhesion resistance of the PDA/PEI CTA membrane and PDA/PEI-n Ag CTA membrane was attributed to the improved hydrophilicity of membranes to some extent.

Furthermore, the number of viable bacterial colonies adhering to the membrane's surface is also shown in Figure 6. The percentages of bacterial content in Figure 6 were based on the number of viable bacterial colonies on the surface of the original CTA membrane. After the deposition of PDA/PEI, the viable bacterial colonies of E. coli and S. aureus on the membrane surface decreased by 6 and 10%, respectively, which was in agreement with the results obtained from the SEM images (Fig. S3). It was due to the increased hydrophilicity of the membrane surface leading to reduce the adhesion of bacteria. The relative content of viable bacteria colonies on the surface of the PDA/PEI-n Ag decorated membrane was greatly decreased, and the mortality rates of E. coli and S. aureus were reaching about 90 and 85%, respectively. It was ascribed that the introduction of Ag NPs on the modified membrane surface. Ag NPs can directly contact bacteria and permeate into cells, resulting in inactivating important cellular functions. Furthermore, Ag NPs could release silver ions in contact with the bacterial suspension. And then Ag⁺ could destroy the cellular membrane or even permeate inside the bacteria, interfering with the normal metabolism of the bacteria followed by the death of the bacteria (Zhang et al. 2016a). As mentioned in section 3.3.1, the overall structure of the cell had been destroyed (Fig. S3). Therefore, the immobilization of Ag NPs on the PDA/PEI-n Ag modified membrane surface endowed the membrane with superior antibacterial property toward E. coli and S. aureus.

The number of living bacteria in the biofilm formed on the fouled surface of virgin CTA, PDA/PEI CTA and PDA/PEI-n Ag CTA membrane was further determined by plate counting method, and the concentrations of viable bacteria are displayed in Fig. S5. It was found that the concentrations of viable bacteria on the surface of virgin CTA, PDA/PEI CTA and PDA/PEI-nAg CTA membrane were (11.7 ± 1.56) × 10⁵, (7.1 ± 0.14) × 10⁵ and (1.3 ± 0.5) × 10⁵ CFU/mL, respectively. The observations were in accordance with the results of CLSM results. These results demonstrated that the introduction of Ag NPs based on the secondary reaction platform of PDA/PEI coating could inhibit the formation and development of biofilm on the membrane surface.

Increase in the hydrophilicity of the membrane surface can form a hydrophilic hydration layer, which can generate a steric hindrance effect and energy barrier to prevent the adsorption and deposition of hydrophobic foulants. However, the function of hydrophilic materials cannot effectively kill the bacteria that have attached to the membrane surface, which can rapidly reproduce and form biofilm under appropriate conditions leading to severe biofouling. Therefore, to effectively control membrane fouling, it is essential to endow the membrane surface with antiadhesion and antibiofouling abilities simultaneously. Based on the analysis of the above experimental results, a variety of hydrophilic functional groups, such as catechol, and amino, introduced by the PDA/PEI could reduce the deposition of foulants and bacteria (Figures 9 and 10). The in situ formation of silver nanoparticles by the reducibility of PDA could further improve the hydrophilic and antimicrobial properties of the membrane surface to control the formation and development of biofilm and mitigate membrane fouling efficaciously.

In this work, PDA/PEI was coated on the surface of the CTA membrane, followed by the in situ reduction of AgNPs to obtain the high antifouling and antibacterial PDA/PEI-nAg CTA membrane. The hydrophilicity of the PDA/PEI-n Ag CTA membrane was significantly improved without affecting the mass transport and integrity. After exposure to E. coli and S. aureus for 24 h, the PDA/PEI-nAg CTA membrane displayed superior antibacterial properties with the bactericidal ratio reaching 90 and 85%, respectively. In dynamic fouling experiments for municipal wastewater treatment, the decline of water flux of PDA/PEI-n Ag CTA membrane was significantly reduced, and the number of viable bacteria and organic substances attached to the membrane surface was remarkably decreased in comparison with the virgin CTA membrane. This facile and eco-friendly membrane surface modification approach has great potential in mitigating the FO membrane fouling.

The authors gratefully acknowledge the financial support from Science and Technology Project of Shaanxi Province, China (Grant No. 2021JQ-222) and the Young Talent fund of University Association for Science and Technology in Shaanxi, China (Grant No. 20200413).

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.