Fouling reduction in nanofiltration membranes in the treatment of municipal sewage – effect of coagulant type used for prior chemically enhanced primary treatment

Water consumption is rapidly increasing because of a rapid rise in population, urbanization, and growing industrialization. The result is the production of large volumes of wastewater (Homer-Dixon 2010). Management, treatment and disposal of wastewater are critical because improper disposal leads to the contamination of ground and surface waters (Tabraiz et al. 2016). Treatment is, therefore, essential prior to disposal or reuse. Physico-chemical, biological and membrane-based treatment methods have been employed in the past (Donkadokula et al. 2020). However, membrane-based wastewater treatment systems are preferred due to the high quality of the effluent produced. A membrane bioreactor (MBR) is an attractive option for such systems (Zeeshan et al. 2017; Al-Asheh et al. 2021). The membranes investigated in MBR were microfiltration, ultrafiltration and nanofiltration (NF) (Choi et al. 2007; Tabraiz et al. 2017; Petropoulos et al. 2021). The major problem with a MBR is membrane fouling due to direct contact between biomass and the membrane. Multiple studies suggested mitigation measures to reduce it (Choi et al. 2007; Ang et al. 2016; Tabraiz et al. 2021; El Batouti et al. 2022).

Chemically enhanced primary treatment (CEPT) has been effectively used for wastewater treatment. In CEPT, metal salts such as ferric chloride, aluminium sulfate, or polymers are used. Various studies demonstrated that CEPT has many benefits as the initial first step for subsequent treatments. These include higher removal rates of total suspended solids (TSS) and biological oxygen demand (BOD) compared to conventional wastewater treatment. In addition, it reduces hydraulic retention time only used twice for the settling tanks, thus reducing footprint and costs (Maktabifard et al. 2018; Chua et al. 2020; Shewa & Dagnew 2020).

In Hong Kong, about 75% of wastewater is treated using CEPT, producing wastewater fit for direct discharge into the sea (Ju et al. 2016). CEPT can reduce chemical oxygen demand (COD) by 43.1–95.6%, TSS by 70.0–99.5% and phosphate by 40.0–99.3%, depending on the coagulants and wastewater matrix (Shewa & Dagnew 2020). CEPT also removes a significant amount of heavy metals (Johnson et al. 2008).

Previous studies have tested a combination of CEPT and membrane technology for wastewater treatment and produced good results (Haydar & Aziz 2009). However, the effect of CEPT on membrane fouling was not investigated. Therefore This study therefore focused on this aspect. Coagulants used were alum, ferric chloride and their 1:1 mixture on fouling of NF membrane. Furthermore, CEPT-NF treated effluent quality was also evaluated for non-potable reuse.

Composite samples of municipal sewage were collected every 24 hours for the wastewater characterization. The parameters tested and the testing procedure used are shown in Table 1. Ten wastewater samples were taken, and their mean value was reported.

The coagulants used are shown in Table 2. Their optimum dose was determined using a jar test apparatus. Triplicate trials for each coagulant were performed, and their averaged concentration was adopted as the optimum dose in CEPT. Turbidity was used to measure the coagulant's performance.

In stage-1, three cases were studied (Table 2).

Percentage reduction in permeate flux was used to gauge the membrane fouling. The NF membrane was considered fouled when the permeate flux was reduced by 15% from its design or initial permeate flux (Lim & Bai 2003; Chon et al. 2013; Zhang et al. 2010). Permeate volume was monitored after every 4 h in the pilot-scale plant operation. Averaged membrane fouling rates and the fouling rates at different stages (maturation, steady and jump) were also calculated as a ratio of percentage flux reduction to time.

Triplicate trials were performed to evaluate the performance of CEPT-NF using nine effluent parameters. These include turbidity, TSS, total dissolved solids (TDS), BOD, COD and total hardness of chlorides and heavy metals. The test procedures used were per the Standard Methods (Rice et al. 2012).

The two-way analysis of variance (ANOVA) was applied to the results. The difference in results was considered significant if the significance level (P value) was below or equal to 0.05.

Pakistan has no standards to evaluate the reuse of treated wastewater for non-potable use, e.g. irrigation, fish and other aquatic life. Therefore, treated wastewater was evaluated on the US Environmental Protection Agency (USEPA) guidelines for non-potable use.

Results of the initial characterization of municipal sewage are given in Table 3. The wastewater used in this study was real municipal wastewater. Real wastewater influent exhibits variation with respect to time.

The wastewater had high BOD, COD, hardness, and TSS. ANOVA analysis demonstrated a stable wastewater composition (P < 0.05).

The wastewater samples drawn from the jar test apparatus were also tested for TSS (Table 4) and COD (Table 5). It was observed that removal in case-3 was 75% for TSS and COD, which was the best among the three cases.

A significant reduction in TSS was observed (Table 4) in stage-1 (CEPT); thus, the effluent of stage-1 was suitable for NF membrane. The high concentration of TSS in the wastewater clogs the membrane when directly applied, resulting in membrane fouling (Almoalimi & Liu 2022). An insignificant change in pH of wastewater was observed after stage-2, and the pH after stage-2 was between 7.01 and 7.42.

Reduction in the chloride was also significant, and better removal was noted in case-3. It was further noted that the removal of chloride increased with time. Typically, pore size reduces during operation due to pore plugging (Nataraj et al. 2008). Due to reduced pore size monovalent ions such as chloride, removal was observed after some period of membrane run. Significant removal of heavy metals was also observed. Typical retentates of NF membrane are divalent ions, and around 97% Pb and 95% Cu were removed in CEPT-NF (Figure 4(f) and 4(g)).

The effluent values of CEPT-NF for case-3 are compared with the (USEPA) guidelines (Table 6). The treated wastewater meets the guidelines’ values and can therefore be used for non-potable purposes (Table 6).

Rates for percentage flux reduction were also high for case-1 in the maturation stages (Table 7). In steady and jump phases, flux reduction rates were higher in case-2 compared to case-1 and case-3. Averaged flux reduction was low in case-3; therefore, fouling was low in case-3. The membrane fouled in 46 min in case-1, 55 min in case-2 and 70 min in case-3. Thus maximum operation time was achieved in case-3, showing a delay in fouling compared to case-1 and case-2.

The membrane fouling was higher in case-1 due to less removal of TSS and COD in CEPT, resulting in high averaged flux reduction rates. Case-3 showed better TSS and COD removal in CEPT, with low membrane fouling rates (Kim et al. 2005).

In case-2, the fouling rates were also high but lower than in case-1 (Table 7), as indicated by the maturation, steady and jump stage. Thus it was concluded that the fouling rates in NF were strongly correlated with the efficiency of CEPT.

The following conclusions can be drawn from the current study:

• (1)

  The type of coagulant plays a major role in the performance and fouling time of NF membrane. A 1:1 mixture of alum and ferric chloride gave the best results for NF membrane.

• (2)

  After CEPT-NF treatment, the percentage reduction in turbidity, TDS, COD, 1:1 alum-ferric chloride mixture was 97%, 86% and 88%, respectively (with an effluent concentration of 0.88 NTU, 41 mg/L and 101 mg/L, respectively).

• (3)

  The NF fouling for the 1:1 alum-ferric chloride mixture was observed after 70 mins of its operation, while it was 46 mins and 55 mins for alum and ferric chloride, respectively.

• (4)

  After CEPT-NF treatment with the 1:1 mixture, the effluent was fit for non-potable use, i.e. irrigation and propagation of fish and aquatic life as per USEPA guidelines.

• (5)

  The reduction of COD, TSS and chlorides were dependent on coagulants and more effective removal was achieved using alum and ferric chloride as a mixture. However, removal of heavy metals, TDS, total hardness and turbidity was independent of coagulants.

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

The authors declare there is no conflict.