Effects of chlorine disinfection on RO membrane biofouling at low feed water temperature for wastewater reclamation

In recent decades, to meet increasing demands for high-quality reclaimed water, reverse osmosis (RO) processes have become widely applied in water-scarce countries due to their features of small footprint, automatic operation and high-quality permeate (Radcliffe & Page 2020; Tseng et al. 2021; Zhao et al. 2021). Nevertheless, RO membrane fouling has always been a major obstacle for the large-scale application of RO systems. Membrane fouling deteriorates permeate and shortens membrane lifetime, which leads to membrane failure and higher operational costs (Shenvi et al. 2015; Jiang et al. 2017; Kehrein et al. 2021). Among the types of RO membrane fouling, biofouling, which represents the ‘Achilles’ heel’ of the RO process, is severe and more difficult to prevent due to bacterial regrowth (Flemming et al. 1997; Khedr 2000).

In order to alleviate biofouling of the RO membrane, multiple approaches have been investigated and proposed, including membrane surface modification, antiscalant and biocide application, feed water pretreatments and chemical cleaning (Nguyen et al. 2012; Yu et al. 2017; Yu et al. 2021). Among numerous approaches to alleviate biofouling, chlorination of feed water is a commonly used and economical technique to inhibit biofilm formation in RO systems (Zhang et al. 2015). However, chlorine disinfection could not inactivate all the bacteria completely. Studies have shown that chlorine disinfection could not always effectively control RO membrane fouling in some projects or lab-scale studies at room temperature (Winters 1997; Obaid & Ben Hamida 1998; Khan et al. 2015; Wang et al. 2019; Luo et al. 2021a). Furthermore, metagenomics analysis revealed that chlorine-resistant bacteria (CRB) could build stable biofilm to resist chlorine and cause RO membrane biofouling (Wang et al. 2021).

Feed water temperature is another important factor affecting biofilm formation on the RO membrane (Farhat et al. 2016). Temperature is well known as a key factor in governing microbial growth (Ratkowsky et al. 1982). Water temperature could modify hydrophobicity and microbial cell surface charge, which in turn affects bacterial cell attachment (Pompermayer & Gaylarde 2000). Water temperature could also affect the characteristics of extracellular polymeric substances (EPS) produced by bacteria (Lewis et al. 1989). Morimatsu et al. (2012) showed that low temperature could increase the viscosity of EPS matrix. Apparently, the water temperature could also affect the microbial community structure (Lindstrom et al. 2005). Wang et al. (2021) showed that chlorine disinfection affected RO membrane biofouling by changing microbial community and EPS characteristics on the membrane. The research findings from Farhat et al. (2016) showed that the membrane apparatus with low-temperature feed water was less susceptible to biofouling than that with high-temperature feed water. However, the biofilm at low temperature was thicker than that at high temperature, which meant that the biofouling formed at low temperature might be more difficult to clean (Farhat et al. 2016). It is worth noting that except for a few low-latitude countries, most countries must deal with low-temperature operation of RO systems. Therefore, it is necessary to investigate the alleviation of RO membrane biofouling at low temperature. Nevertheless, the effect of chlorine disinfection on membrane biofouling at low feed water temperature has not been reported in the literature, which still needs systematic investigation.

In this study, the main objective was to investigate the effects of chlorine disinfection on the biofouling of the RO membrane at low feed water temperature (10 °C) for wastewater reclamation. The permeate flux with and without chlorination during the operation process was recorded using a lab-scale cross-flow RO apparatus. After the operation, the foulants on the membrane were analyzed comprehensively using microscopic and DNA sequencing-based techniques to elucidate the effects of chlorination at low temperature.

The feed water of the lab-scale RO apparatus was collected from the effluent of a municipal wastewater treatment plant in Qingdao, China. The treatment plant adopted the moving-bed biofilm reactor as the secondary treatment, with coagulation and filtration as tertiary treatments. The water samples were filtered by filter paper and stored at 4 °C. The general water quality parameters are shown in Table 1.

Table 1

Feed water characteristics

Characteristics .	Concentration .
pH	7.0–8.0
Conductivity (μS/cm)	3,560 ± 20
DOC (mg/L)	8.88 ± 0.18
-N (mg/L)	1.06 ± 0.12
UV254 (cm−1)	0.123 ± 0.002
Hardness (mg/L CaCO3)	443.50 ± 31.50
Cl− (mg/L)	900.17
F− (mg/L)	0.27
(mg/L)	188.45
(mg/L)	7.57

Characteristics .	Concentration .
pH	7.0–8.0
Conductivity (μS/cm)	3,560 ± 20
DOC (mg/L)	8.88 ± 0.18
-N (mg/L)	1.06 ± 0.12
UV254 (cm−1)	0.123 ± 0.002
Hardness (mg/L CaCO3)	443.50 ± 31.50
Cl− (mg/L)	900.17
F− (mg/L)	0.27
(mg/L)	188.45
(mg/L)	7.57

View Large

To investigate the effects of chlorination on the RO membrane fouling at low temperature, one set of experiments was run for 27 days at 10 °C of feed water temperature. The water samples disinfected by 0 (without chlorination) and 5 mg-Cl₂/L chlorine were set as the feed water. NaClO solution (2.30 g/L) was added to the sample to simulate chlorine dosage 5 mg-Cl₂/L. After 30 min of the lightproof reaction, the excess of Na₂S₂O₃ solution (2%) was added for dechlorination. For the water sample without chlorination, an equal volume of NaClO and Na₂S₂O₃ solutions was mixed and then added into the water sample to achieve the same ionic strength as compared with that in the water sample with 5 mg-Cl₂/L chlorine. After disinfection, the residual bacteria were counted by the heterotrophic plate count (HPC) method (Yu et al. 2018). Chlorine disinfection of the water samples was performed daily during the 27-day continuous operation of the experiment.

where J/J₀ is the normalized flux at time t. J_pss is the final normalized flux at the equilibrium phase where the flux would not decline. k is the membrane surface ‘blocked’ per unit of total volume permeated through the membrane, which could represent the fouling potential of the water sample (m⁻¹).

The water temperature was monitored in real time by a temperature recorder (Jucsan, China) and uploaded to a computer. Conductivity and pH were measured by a multi-parameter controller (LEICI, China). UV absorbance at 254 nm (UV₂₅₄) was measured by a Uvmini-1240 spectrophotometer (Shimadzu, Japan). Dissolved organic carbon (DOC) was measured by a Multi N/C 2100 analyzer (Analytik Jena, Germany). The total hardness was determined by EDTA titration. The anion concentration was determined by a Dionex ICS-5000 ion chromatograph (Thermo Fisher, USA).

At the end of the running of the RO apparatus, the membrane coupons were taken out from the RO cell and cut into small pieces of a certain area for further analysis. Photographs of the fouled membranes were taken with a digital camera. Scanning electron microscope (SEM, QUANTA FEG250) analysis was used to investigate the microscopic details of the membrane images. The surface roughness of the membranes was measured by AFM (Agilent 5400) with a scan area of 50 μm × 50 μm. The roughness parameters (Ra) were calculated using the Nano-Navi Station analysis platform.

The biofilm accumulated on the membrane was analyzed by a confocal laser scanning microscope (CLSM, OLYMPUS FV1000). Fouled membrane samples were stained by four types of fluorescent dyes. Total cells, dyed cells, proteins, and polysaccharides were labeled by SYTO9 (Molecular Probes, ex 480/em 500 nm, shown as green), propidium iodide (PI, Molecular Probes, ex 490/em 635 nm, shown as red), SYPRO^® Orange Protein Gel Stain (Thermo Fisher, ex/em 470/570, shown as blue), and tetramethylrhodamine conjugates of concanavalin A (TRITC-ConA, Thermo Fisher, ex/em 555/580, shown as yellow), respectively. The staining step was adopted from the method described by Yuan et al. (2015).

The total amount of foulant on the membrane was characterized by turbidity. The foulant on the membrane was scraped and suspended in ultrapure water, then measured by turbidity using an ultravioletvisible spectrophotometer (DR6000, HACH). The dissolved organic matter in the foulant were characterized by the DOC concentration using NaOH solution (pH 12) as a solvent, following the method as described by Wang et al. (2019).

The microbial density in the foulants was characterized by the HPC method. A piece of the fouled membrane was cut with the area measured. The bacteria on the coupon were suspended by 0.9% NaCl solution and then cultured in R2A agar plates at 25 °C for 7 days. For microbial community analysis, a piece of the fouled membrane was cut, and the total genomic DNA was extracted using an E.Z.N.A.^® Water DNA Kit (Omega Bio-tek, USA) following the manufacturer's protocol. The V4 hypervariable region of the bacterial 16S rRNA gene was amplified and then sequenced using an Illumina Miseq platform (PE300, Illumina, USA) (Caporaso et al. 2012). The microbial community analysis, including operational taxonomic unit (OTU) classification and alpha and beta diversity analyses, was performed as described in the previous study based on the SILVA Taxonomy database (GTDB 2022) (Yu et al. 2017; Parks et al. 2018). All analyzed sequences were submitted to the NCBI SRA database under an accession number PRJNA806491.

Different concentrations (0 and 5 mg-Cl₂/L) of chlorine were added to the water samples for disinfection. After chlorination, the residual bacterial concentrations in the water samples are shown in Table 2. The results showed that some bacteria still survived after disinfection with 5 mg-Cl₂/L chlorine in 30 min, which was consistent with the previous study (Wang et al. 2019).

The flux decline curves with and without chlorination were fitted using Equation (2) via the intermediate blocking model. The final normalized flux J_pss of water samples with and without chlorination was 0.0291 and 0.0295, respectively. The fouling potential parameter k of water samples with and without chlorination was 0.135 and 0.109 m⁻¹, respectively (Table 3). These results indicated that chlorine disinfection would increase the RO membrane fouling potential of the water sample at low feed water temperature (10 °C), which was consistent with the previous study at ambient temperature (25 °C) (Wang et al. 2019).

The densities of total foulants, dissolved organic matter and bacteria on RO membranes were characterized by turbidity, the DOC concentration and the HPC method, respectively. As shown in Table 5, the total foulant on the RO membrane with chlorination was 40.77 ± 0.82 FTU/cm², which was much higher than that without chlorination (29.87 ± 0.60 FTU/cm²). Similar to the density of total foulants, the density of dissolved organic matter and bacteria on the RO membrane with chlorination was much higher than that without chlorination (Table 5). These results indicated that chlorine disinfection of the feed water would accelerate the accumulation of foulants on the RO membrane surface at low temperature, which was consistent with the previous study at ambient temperature (Wang et al. 2019).

The alpha diversity of fouled membranes and water samples was analyzed and is shown in Table 6. According to the Shannon and Simpson indices, not surprisingly, microbial community diversity in RO feed water decreased significantly after chlorination. Furthermore, the results in Table 6 showed that, under the unchlorinated condition, compared with the feed water, the low temperature resulted in the decrease of microbial community diversity on the RO membrane surface. Meanwhile, under the chlorinated condition, with the operation of the apparatus, the microbial community diversity on the membrane surface was restored, but it could not reach the level on the RO membrane surface without chlorination.

In this work, the effects of chlorine disinfection on the RO membrane biofouling at low feed water temperature (10 °C) were investigated. At low temperature, chlorination could not alleviate the RO membrane fouling. Besides, it would increase the RO membrane fouling potential of the feed water. At low temperature, the biofilm on the membrane with chlorination was thinner and denser than that without chlorination. In addition, the membrane with chlorination contained more foulants and dissolved organic matter than that without chlorination. Chlorination failed to continuously prevent the bacteria accumulation on the RO membrane at low temperature but screened out bacteria that were potentially more suitable for the low-temperature membrane environment.

This study was supported by the Research Institute for Environmental Innovation (Suzhou), Tsinghua (No. 00010272), the Higher Educational Science and Technology Program of Shandong Province, China (2020KJD003) and the Taishan Scholars Program of Shandong Province, China (No. tsqn201812091).

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

The authors declare there is no conflict.