Mitigating biofouling of polyamide thin-film composite RO membrane modified with AgNPs and pSBMA: an evaluation close to the actual SWRO desalination application scenario

Polyamide thin-film composite (PATFC) membranes are widely used in mainstream membrane technologies such as reverse osmosis (RO) and nanofiltration (Pan et al. 2016), but they are susceptible to biological fouling during application (Suresh et al. 2022). The development of biofouling-resistant membranes not only inhibits the formation and growth of biofilms at their source but also alleviates the pressure on pretreatment and membrane cleaning processes, reduces the use of chemical reagents, and minimizes byproduct formation (Lin et al. 2010). Therefore, this approach is considered the most effective method for preventing and managing biofouling in membranes.

There are two commonly used approaches for preparing anti-biofouling membranes. One method involves incorporating hydrophilic materials during membrane fabrication via interfacial polymerization (Agnihotri et al. 2013; Mariën & Vankelecom 2018; Vatanpour et al. 2022), while the other entails grafting hydrophilic materials onto the membrane surface post-fabrication (Firouzjaei et al. 2020; Zhao et al. 2021). The latter approach is particularly advantageous as it allows for simple and facile surface modification while maintaining the structural properties of the substrate membranes. Moreover, post-fabrication modification enables large-scale modification of commercially available membranes that are of consistent quality and available at a low cost. Hydrophilic polymers and nanomaterials are two types of hydrophilic materials that have been widely studied for enhancing the fouling resistance of PATFC (pressure-assisted thin-film composite) membranes.

Among hydrophilic polymeric materials, polyethylene glycol (PEG)-based polymers have been the most extensively explored for antifouling membrane surface modification. The neutral charge of PEG reduces the interactions between foulants and the membrane, thereby effectively preventing foulant adsorption (Kang et al. 2007; Zhang et al. 2019). However, PEG tends to oxidize in complex media, compromising its antifouling property and suitability for long-term applications (Ray et al. 2017). Zwitterionic polymers (ZPs) have emerged as promising alternatives to PEG derivatives for low-fouling surfaces, as they form denser and tighter hydration layers than other hydrophilic materials (Lin et al. 2019). Consequently, PATFC membranes coated with ZPs exhibit significantly reduced adsorption of proteins and bacteria (He et al. 2016; Cui et al. 2020). While ZPs can delay or prevent microbial attachment, they cannot inactivate bacterial cells. Thus, to combat biofouling, several studies propose the fabrication of dual-function membranes by combining ‘offensive’ and ‘defensive’ strategies (Zou et al. 2021). These strategies rely on strong antimicrobial agents, such as organic antimicrobial compounds or biocidal nanoparticles, to inhibit bacterial proliferation, which is considered a promising solution to biofouling in membrane-based technologies.

Surface modification using hydrophilic nanomaterials has been actively investigated to impart ‘offensive’ biofouling resistance to TFC membranes. Various nanomaterials, including metal oxide nanoparticles and carbon-based nanomaterials, have been used to develop antifouling TFC membranes (Zhang et al. 2017; Pejman et al. 2020; Liu et al. 2021; Ngo et al. 2022; Feng et al. 2023). Silver nanoparticles (AgNPs) have received particular attention due to their broad-spectrum antimicrobial activity (Dakal et al. 2016; Salleh et al. 2020). Unlike organic antimicrobial agents that inactivate bacterial cells through contact-mediated mechanisms, AgNPs' toxicity is driven by the release of Ag⁺ ions (Durán et al. 2016). The physicochemical properties of AgNPs can be tailored to achieve improved reactivity, and their dissolution-dependent mechanism of toxicity makes them less susceptible to the presence of other chemical foulants in the feed stream (Ben-Sasson et al. 2016). Despite concerns about their long-term efficiency, the regenerative capability of AgNPs has proven to be a key element in the preferential design of silver-modified membranes for biofouling mitigation (Liu et al. 2017).

Studies on the co-modification of PATFC membranes with ZPs and AgNPs for anti-biofouling applications show promising results but reveal significant gaps (Liu et al. 2021; Zhang et al. 2018; Yi et al. 2019)). First, there is a lack of systematic optimization of modification conditions (e.g., reaction time of the grafting initiator, loading sequence, and the ratio of zwitterions to nanoparticles) specifically for seawater reverse osmosis (SWRO) desalination, the most common application of PATFC RO membranes. Second, current research typically evaluates the performance of modified membranes using short-term static contact experiments with Escherichia coli and NaCl filtration tests. However, these methods fail to account for variables such as shear forces from water flow, background water composition, and physiological differences between E. coli and other microorganisms that contribute to membrane fouling. Third, the long-term anti-biofouling efficacy of membranes co-modified with ZPs and AgNPs remains largely unexplored. Fourth, while membrane stability has been examined under mild conditions (e.g., NaCl or NaHCO₃), it has yet to be tested under more rigorous chemical cleaning processes. Finally, there is limited mechanistic insight into the biofouling resistance of AgNP- and ZPs-co-modified membranes, with the resistance often attributed to Ag⁺ disinfection and surface hydrophilicity and roughness, without deeper exploration of interactions with biofilm and ZPs/AgNPs.

To address the above issues, taking SWRO desalination as the design application scenario, PATFC membranes modified with zwitterionic poly(sulfobetaine methacrylate) (pSBMA) and AgNPs were prepared and used to study the following aspects: (1) aiming for anti-biofouling while also considering salt rejection rate, water flux, and resistance to organic fouling, the key modification parameters (AgNO₃ concentration, contact time of membrane with initiator, and atom transfer radical polymerization (ATRP) reaction time) were optimized using response surface methodology (RSM) through an improved anti-biofouling testing method and (2) the optimized modified membranes were subjected to long-term continuous run test with RO unit feed water from actual SWRO desalination plant to evaluate the durability of the positive effects of the co-modification on the PATFC membrane. Our results will provide a valuable theoretical and technical reference for AgNPs or/and pSBMA modification technology for the preparation of anti-biofouling membranes applied in SWRO desalination.

All reagents (Text S1, Supporting Information) were used as received without further purification. All solutions were prepared with reagent-grade chemicals and purified water (18.2 MΩ cm) produced by a Smart-N system.

The PATFC RO membrane used in the experiment was obtained by disassembling the LC-LE 4040 roll-type membrane assembly. The membrane was then cut to an appropriate size and soaked in a 25% aqueous solution of isopropyl alcohol (IPA) for 30 min to remove the surface coating of the commercial film. The membrane was subsequently rinsed with ultra-pure water for more than 24 h, with the water being changed several times during the process. The pretreated membranes were stored in a refrigerator at 4 °C prior to use.

The ARGET ATRP method was used for the grafting of the amphoteric polymer pSBMA on PATFC membranes (Min & Matyjaszewski 2005). The AgNPs were loaded on PATFC membranes through AgNO₃ reduction with NaBH₄ (Ben-Sasson et al. 2014). Successful loading of AgNPs can be confirmed by SEM, EDS, and XPS analysis (Figure S1(a)–(b), Supporting Information), while successful grafting of pSBMA can be verified by XPS and FTIR analysis (Figure S1(b)–(c), Supporting Information). Modification parameters include AgNO₃ concentration ([AgNO₃]₀), contact time of membrane with α-bromoisobutyryl bromide (BiBB) and dopamine hydrochloride (DA) (T_BiBB-DA), and ATRP reaction time (T_ATRP). The adjustment range of [AgNO₃]₀ (2–5 mM), T_BiBB-DA (2–10 min), and T_ATRP (30–90 min) is initially determined by pre-experiments (Figure S2–S4, Supporting Information). Experiments have shown that the loading order of AgNPs and pSBMA has little effect on membrane performance (Figure S5, Supporting Information), which differs from the findings reported by Liu et al. (2017). To reduce the leaching of AgNPs, the sequence of loading AgNPs first and then grafting pSBMA for the preparation of AgNPs–pSBMA–PATFC membrane was adopted.

A high-pressure flat membrane equipment (TQFM-24, Shanghai Tongqin Environmental Protection Technology Co., Ltd) was employed for continuous flow testing experiments. The membrane cell area of this high-pressure flat testing apparatus is 24 cm² (4 cm), with an adjustable filtration pressure range of 1–60 bar and a cross-flow rate of 1 L/min with a recovery rate of 60%. Integrated with a thermostatic water bath (25 °C), electronic balance, and computer, this setup allows for filtration experiments at specified temperatures and pressures while measuring membrane flux. The water flux and salt rejection rate was determined using a 2,000 mg/L NaCl test solution (Text S2, Supporting Information).

In order to investigate the performance of AgNPs–pSBMA–PATFC membranes during long-term operation, the TQFM-24 high-pressure flat membrane equipment was used for continuous flow testing of AgNPs–pSBMA–PATFC membranes for a period of 6 months. RO feed water from SWRO desalination plant (Table S1, Supporting Information) was used for testing. The membrane flux and desalination rate were recorded during the operation. Since it was not possible to pause the operation to cut membrane samples, only the membrane samples at the end of the operation were characterized.

Additionally, at the end of the operation, the membranes were subjected to a simulated cleaning process. The experiments followed the chemical cleaning agent recommendations provided in the technical documentation of FilmTec RO membranes (Dupont 2023). A cleaning method preferred by inorganic colloids, silica, biofilm, organic foulants was selected. Specifically, modified membrane samples (2 cm × 2 cm) were sequentially immersed in cleaning solutions 0.2% HCI (25 °C and pH 1) and 0.1% NaOH and 0.025% sodium salt of dodecylsulfate (Na-DSS) for 2 h each. The container for cleaning solution was placed on a shaker (35 °C and 150 rpm) to ensure adequate contact between the cleaning solution and the membrane. Residual AgNPs on the membrane surface were dissolved using 10 mL of 10% HNO₃, and the resulting Ag⁺ concentration was determined by ICP-MS. Changes in pSBMA loading were evaluated by comparing the intensity of its characteristic FTIR peaks.

The biofilm on the membrane surface was characterized by the confocal laser scanning microscopy (LSM780, ZEISS, Germany) after staining with SYTO9 and propidium iodide (PI) (LIVE/DEAD Biofilm Viability kit, FilmTracer). Live microorganisms are displayed with green fluorescence and dead with red fluorescence. The obtained images were analyzed using ZEN2010 (Carl Zeiss Microscopy GmbH, Jena, Germany). Extracellular polymeric substances (EPSs) were analyzed by fluorescence excitation and emission matrix (EEM) spectra (Liu et al. 2023). Parallel factor analysis (PARAFAC) was performed via MATLAB 9.11 using the drEEM toolbox.

Acyl-homoserine-lactones (AHLs) were concentrated using liquid–liquid extraction with ethyl acetate followed by rotary evaporation (Liu et al. 2022). The residue was then recovered with 0.5 mL of methanol for testing. Signal molecules were analyzed using HPLC-MS/MS (Agilent 6400, USA), with a linear gradient of solvent A (methanol) and solvent B (0.1% formic acid and 2 mM acetonitrile) as the mobile phase. The flow rate was maintained at 0.3 mL/min, and mass spectra were measured in ESI mode. To confirm the identity of the putative AHLs, a full scan (m/z 100–350) coupled with a SIM scan was performed. The obtained mass spectra were compared with reference standards.

In order to facilitate rapid optimization of the preparation conditions, the inhibition rate obtained from 3-h static contact experiments was used as an indicator to represent the anti-biofouling performance of the modified membranes. The water flux and salt rejection rate was used to characterize the separation ability of membrane.

In order to obtain the optimal modification parameters with the antibacterial rate and water flux as the response variable and [AgNO₃]₀ (A), T_BiBB-DA (B), and T_ATRP (C) as influencing factors, RSM was employed to analyze the optimal modification parameters (Text S1 and Tables S1–S3, Supporting Information). The optimal recommended values for [AgNO₃]₀, T_BiBB-DA, and T_ATRP are 5 mM, 10 min, and 44.4 min, respectively (Table S4, Supporting Information). Under these optimized modification conditions, the experimental measurements for antimicrobial activity and water flux were respectively 93.0% and 4.35 L m⁻² h⁻¹ bar⁻¹, which were consistent with the model predictions.

The study has demonstrated that the sequential modification of PATFC membranes with AgNPs and pSBMA is an effective approach for enhancing biofouling resistance from a short-term operational perspective (>90% improvement) but not significantly effective from a long-term operational perspective (<25%, based on biovolume). The modification optimization results indicate that the selection of test solutions (including microorganisms and background components) and test methods needs to be specifically designed to ensure that the performance parameters obtained from batch experiments are more accurately translated to actual operational performance. When both AgNPs and pSBMA are loaded onto the PATFC membrane, the AgNPs exhibit different functions compared to when AgNPs are loaded alone. In the former case, the improvement is due to the quorum sensing inhibition and the bactericidal action of silver ions released by AgNPs. In the latter case, the bactericidal action of silver ions is shielded by halide ions competition. Besides forming a hydration layer on the membrane surface, pSBMA also provides a negative charge barrier that weakens the impact of halide ions in seawater on AgNPs. Although this study improved the testing methods for anti-biofouling modifications, the significant differences between performance data from batch experiments and actual operational indicate that the existing testing methods require substantial changes. Moreover, the 180-day operation period in this study only corresponds to one cleaning cycle of an actual SWRO desalination plant. Subsequent research needs to evaluate the AgNPs–pSBMA–PATFC membrane from a full life-cycle perspective. Finally, based on the experimental results, it is recommended that AgNPs be loaded onto the membrane before pSBMA to ensure the effectiveness of AgNPs. However, this poses challenges for the replenishment of AgNPs loss over time.

This research is funded by the National Key R&D Program of China (No. 2023YFF0614501), Zhejiang Public welfare technology research program (No. LGG21E080011), and the Jiyang 533 Talent Program (No. JY21E0102).

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

The authors declare there is no conflict.