Preparation of a novel positively charged nanofiltration composite membrane incorporated with silver nanoparticles for pharmaceuticals and personal care product rejection and antibacterial properties

Pharmaceuticals and personal care products (PPCPs) are a diverse group of chemicals used extensively in medicines and personal hygiene products. To date, environmental pollution by PPCPs has been an increasing public concern due to their universal occurrence and potential adverse effects on non-targeted human beings and organisms (Kolpin et al. 2002; Dougherty et al. 2010). Especially in wastewater treatment plants, PPCPs cannot be removed completely during treatment so that considerable amounts of such chemicals remain in the treated wastewater and then enter the environment through disposal or reuse of treated wastewater and biosolids (Kinney et al. 2006). Therefore, how to improve the removal efficiency of PPCPs has attracted much attention.

Membrane technology has been widely used in many fields, including water treatment, protein separation and reverse osmosis (RO) pretreatment (Halpern et al. 2005). Nanofiltration (NF) is a liquid-phase pressure-driven separation process employing semipermeable membranes with the pore size in the nanometre range, between that of the ultrafiltration (UF) and RO membranes. The main separation and transport mechanism of the NF process involves size exclusion effect, convection and Donnan exclusion between the membrane and the treated solution. Due to its unique ability of separating and fractionating ionic and small molecule organic species, NF has a smaller occupation area and a better rejection of PPCPs with a molecular weight of 150–1,000 Da. In this study, we focused on the development of a positively charged membrane for the removal of PPCPs through easy modification of a UF membrane using a dip-coating and photoreduction technique. Compared with interfacial polymerization (Fang et al. 2013), the self-assembly process (Ouyang et al. 2008) and UV-assisted grafting polymerization (Deng et al. 2011; Zhong et al. 2012) that have been adopted for preparing positively charged NF membranes, the approach adopted in our study is simple and avoids using hazardous materials. In our early study, we reported a method to prepare a pot blending of biopolymer–TiO₂ composite membranes which had enhanced mechanical strength (Huang et al. 2015).

To date, NF technology has emerged as a valuable tool in purifying PPCPs. However, one of its disadvantages is the contamination on the membrane surface that decreases the fluid flux and reduces its efficacy. Membrane fouling leads to the rise in manufacturing and energy costs, damage or destruction of membranes, reduction in operating lifetimes, and secondary contamination of water by the metabolic products of microorganisms. Different strategies have been taken to address this issue, e.g., feed water pre-treatment, managing operational performance and development of modified antibacterial membranes (Gao et al. 2011). However, it still remains a challenge to realize the antibacterial property of the NF membrane surface by using a simple strategy.

Chitosan (CTS) is a kind of polysaccharide with several properties such as hemostatic activity (Bonferoni et al. 2009), non-toxicity (Bhattarai et al. 2005), biodegradability, and antibacterial activity (Deng et al. 2011). N-[(2-hydroxy-3-trimethylammonium)propyl] chloride chitosan (HTCC), a water-soluble derivative of CTS, was found with better bacteria inhibition properties than CTS. The introduction of positively charged quaternary ammonium groups significantly enhanced the interaction between the CTS molecule and the bacterial cell wall, inhibiting the bacterial growth and reproduction as well (Qin et al. 2004). To date, HTCC has been widely applied for the preparation of membranes. It is also well known that silver has an inherent antimicrobial property and is capable of causing a bacteriostatic or a bactericidal impact (Thiel et al. 2007). Moreover, silver is very effective in purification systems for disinfecting water or air (Sondi & Salopek-Sondi 2004; Zhang et al. 2004). Recently, the hybrids of silver nanoparticles with amphiphilic hyperbranched macromolecules have been reported as an effective antimicrobial surface coating. They may affect the cell integrity and metabolism through direct interaction with cell membranes, releasing dissolved silver species, and generating reactive oxygen species (Jin et al. 2010; Marambio-Jones & Hoek 2010). Nano-Ag particles have been incorporated into nanocomposite microfiltration (Gunawan et al. 2011; Diagne et al. 2012), UF (Zodrow et al. 2009), and RO (Huang et al. 2012) membranes with various methods including membrane surface modification (Diagne et al. 2012), phase-inversion (Zodrow et al. 2009; Huang et al. 2012) and interfacial polymerization.

In our study, we report a simple method to synthesize a kind of novel composite membrane with silver nanoparticles which act as an antibacterial agent. The HTCC-Ag/polyethersulfone (PES) membrane was prepared through photochemical reduction technique on the base of an HTCC/PES membrane. The obtained composite membranes were characterized by scanning electron microscopy (SEM), X-ray photoelectron spectroscopy (XPS), and atomic force microscopy (AFM). Meanwhile, we studied the rejection of PPCPs by HTCC/PES positively charged membranes and investigated the antibacterial activities of the membranes.

Chitosan (weight-averaged molecular weight ≈ 150 kDa, Sinopharm Chemical Reagent Co., Ltd) was dissolved in the solvent acetic acid (CH₃COOH ≥99.5%, Sinopharm Chemical Reagent Co., Ltd). Glycidyl trimethylammonium chloride (GTMAC, 96% purity) was from Dongying Guofeng Fine Chemical Co., Ltd. Epichlorohydrin (ECH) (Sinopharm Chemical Reagent Co., Ltd) was used as a cross-linking agent. Ethanol, sodium chloride, silver nitrate, nutrient agar, glucose, sucrose and polyethylene glycol (PEG) (analytical grade) were used as received without any further purification; the PES UF membrane with molecular weight cut-off (MWCO) of 100 kDa was kindly provided by Ande Membrane Separation Technology & Engineering (Beijing) Co., Ltd. All aqueous solutions were prepared using purified water with a resistance of 18.2 MΩ.cm.

Two grams of CTS was dissolved in 100 mL of 2% (v/v) acetic acid. The solution was then adjusted to pH = 9.0 by 6 mg/mL NaOH, and stirred for 2 h to allow CTS alkalinization. The precipitate obtained was transferred into a flask, followed by the addition of isopropyl (30 mL). After the mixture was stirred at 60 °C for 4 h, 7.53 of GTMAC was added and stirred at 80 °C for 8 h. Finally, the viscous solution was cooled to ambient temperature and poured into acetone. The HTCC obtained was washed three times by 500 mL acetone and dried at 80 °C for 12 h. The degree of quaternization of HTCC was 88.7% and the solubility in water was above 250 mg/mL.

The surface morphologies of composite membranes were obtained by using SEM (JEOLJSM-6380LV; Japan). The existence of Ag nanoparticles was examined by X-ray diffraction (XRD, D8 Advance, Germany). AFM (Dimension310), operated in air atmosphere and in the tapping mode, was used to measure the surface images of the air-dried membranes. The root mean square roughness values in 4 μm × 4 μm area of the membrane surfaces were calculated at least three times for each independent membrane sample. The surface chemical compositions of the membranes were analyzed by XPS (PHI Quantera, Japan) using Mg Kα as radiation source.

The influences of ionic strength on rejection performance towards PPCPs was also investigated with a series of sodium chloride solutions mixed in 100 mL samples of pH 7.0 at 37 °C.

All results presented are average data with standard deviation from at least three samples of each membrane type.

We investigated the antibacterial property of HTCC-Ag/PES membrane using the Gram-negative Escherichia coli Rosetta and E. coli DH5α and Gram-positive Bacillus subtilis with the inhibition zone method. Nanofibrous mats were cut onto disks with a diameter of 1 cm and then sterilized with a UV lamp for 30 min. The diluted bacterial suspension (50 μl) with bacterial colonies of 5–10 × 10⁵ CFU/ml (CFU: colony forming units) was added into the agar culture medium uniformly (CLSI 2012a). The membranes mentioned above were sequentially placed on the plate. After the active layer was affixed to the surface of the medium, the plates were inverted and incubated at 37 °C for 24 h. CFU results were obtained from three measurements of three independent membranes based on triplicates (CLSI 2012b).

Figure 3(a) illustrates the XPS peak fitting curve of the samples. The two remarkable peaks around 372.5 eV and 366.7 eV correspond to Ag3d_3/2 and Ag 3d_5/2, respectively, according to Sawada et al. (2012), which confirmed that nano-Ag was indeed deposited on the surface of composite membranes. XRD analysis was employed to investigate the architecture of the HTCC-Ag/PES membrane so as to illustrate the existence of the nano-Ag particles deposited on the film surface. As can be seen in Figure 3(b), in the XRD pattern of the HTCC-Ag/PES composites exists a peak with a maximum intensity at 2θ = ∼38^◦, which corresponds to the zero valence silver (Nguyen et al. 2011). From the XRD and XPS analysis, it is clear that Ag nanoparticles exist in composite membranes.

Figure 4(b) shows the MWCO of the composite membrane, which was determined through permeation tests using 1 g/L glucose (198 Da), sucrose (342 Da) and three kinds of standard PEG with molecular weights of 600, 800, and 1,000 Da, respectively. The MWCO was the molecular weight of organic substance with a retention of 90% (Afonso et al. 2001). Figure 4(b) illustrates the observed rejections of the studied membrane towards five different fractions at ΔP = 4.0 bar. From the rejection behavior, the MWCO of the composite membrane was found to be around 941 Da.

As shown in Figure 7(b), the rejection of atenolol and carbamazepine decreases while the rejection of ibuprofen increases with an increasing concentration of Na⁺. The reason is that there exists an interaction between the dissolved salts such as sodium chloride and the membrane surfaces. Furthermore, the existence of Na⁺ would reduce the thickness of the electric double layer, thus weakening the electrostatic repulsion. When the ionic content increased, the rejection of the ibuprofen and atenolol by the composite membranes changed steeply while that of carbamazepine changed very little, which was not only characteristic of a charged NF membrane but also explained by electrostatic effect with the dissolved salts.

As mentioned above, the order in which the three strains were inhibited by the HTCC-Ag/PES membranes is as follows: E. coli DH5α > E. coli Rosetta> Bacillus subtilis. Overall, the inhibition effect of Gram-positive bacteria is weaker than that of Gram-negative bacteria.

In summary, the positively charged HTCC-Ag/PES membranes were successfully prepared via photoreduction technique. Our results showed that silver nanoparticles assembled on the surface of the membranes. The PPCP rejection by the membranes followed the sequence: atenolol > carbamazepine >ibuprofen, revealing the characteristic of positively charged NF membranes. We found that the PPCP rejection was mainly influenced by the electrostatic repulsion. It was also proved that the existence of nano-Ag particles helped to improve antibacterial activity against both Gram-negative and Gram-positive bacteria. Our method could provide new ideas and theoretical support for removal of emerging trace contaminants by CTS-based NF membranes.

This work was financially supported by National Natural Science Foundation of China (Grant No. 51208259) and the Jiangsu Key Laboratory of Chemical Pollution Control and Resources Reuse, Nanjing University of Science and Technology (Grant No. 30920140122008).