Removal of natural steroid hormones from municipal MBR effluent with UF/NF/RO membranes

Abstract

In the context of water scarcity, the increasing risks linked with the presence of natural steroid hormones and many emerging anthropogenic micropollutants (MPs) persevering through municipal wastewater treatment works (MWWTWs) are of concern for their endocrine-disrupting activities detected in receiving surface waters. For the last decades, the use of membrane technology has grown considerably in wastewater treatment and has proven to be an effective method for the removal of a wide variety of contaminants from wastewater. In this study, domestic wastewater treated by a full-scale membrane bioreactor (MBR) at MWWTWs in the City of Cape Town (CoCT), South Africa, was used directly as the influent to a reverse osmosis (RO) pilot-plant for the removal of selected natural steroid hormones; 17β-estradiol (E2) and Testosterone (T) as a potential indirect water recycling application. Three commercially available ultrafiltration/nanofiltration/reverse osmosis (UF/NF/RO) membranes, namely UA60, NF270 and XLE, were selected for this investigation. Membrane surface modification was investigated to minimize fouling during the MBR-UF/NF/RO treatment processes. To enhance the resistance to flux decline, a thin film composite reverse osmosis membrane was grafted with polyvinyl alcohol (PVA) through cross-link with glutaraldehyde (GA). To assess the effect of surface modification on the membrane surface, analyses using attenuated total reflectance Fourier transform infrared spectroscopy (ATR-FTIR) and scanning electron microscopy (SEM) were done. Escherichia coli (E. coli) bacterial solution was used as biofouling to evaluate the influence of the surface modification initiated on antifouling properties of the membrane. A model MBR secondary effluent with a bench-scale NF/RO system was used for these fouling tests. The pure water flux decreased somewhat as a result of the morphological structure and chemical property change due to membrane surface modification. The membrane resistance to fouling was better and the biofouling model used exposed the anti-biofouling capacity of the membrane, although a small decrease of salt rejection was detected. The flux decline and flux recovery ratios improved with an increase in PVA concentration. The sterilization ratio increased from 33.8 to 36.8% and the pure water flux decline, reduced from 46.04 to 25.94 % after modification. The effects of operating conditions on the removal of inorganics by MBR-UF/NF/RO were evaluated. Experimental runs were conducted on a pilot plant in a continuous system, varying the pH, as well as the permeate flux and the percentage recovery in the case of RO membrane. Chemical analysis of different inorganics was conducted to calculate the percentage removal. Results revealed considerable effects of pH control on the removal of the inorganics of interest as well as the carbon oxygen demand (COD). Adjustment of flux and recovery for the RO membrane was shown to be a factor of consideration for the improvement of inorganics removal in the advanced treatment of domestic secondary effluent. It was shown that the quality of water obtained with the RO membrane could meet quality requirements for reuse application in cooling systems and irrigation among others. The UF and NF reduction of inorganics was shown to be limited to meet the required water standard for some of the reuse applications. The NF membrane was found to produce water suitable for restricted irrigation and cooling systems, while the UF produced permeate that was found appropriate for cooling systems only. Estrogenicity and androgenicity were assessed using the enzyme-linked immunosorbent assays (ELISA) and the recombinant yeast estrogen receptor binding assays (YES). The influent pH and flux did not have an influence on the rejection of E2 and T, which was most likely ruled by adsorption, size exclusion and diffusion simultaneously. Size exclusion was seemingly dominant, especially with NF and RO membranes. T with a smaller partitioning coefficient (log Kow) value was most likely adsorbed on the membranes and then passed through it to give a low rejection with all three membranes. RO and NF membrane processes exhibited excellent removal rates (>95%) for E2 and T. It was revealed that RO showed higher removal percentages when compared with NF and UF. All the E2 effluent samples with the MBR/UF, MBR/NF and MBR/RO, were higher than the lowest trigger value of 0.4ng/L of the test, but less the than USEPA and WHO of 0.7 ng/L as well as less than the predicted no-effect concentration (PNEC) values for fish (1 ng E2/L). Consequently, domestic secondary wastewater treated by full-scale MBR followed with a pilot-plant, NF or RO is acceptable for the effective removal of natural hormones (E2 and T). Ultimately, a multi-barrier tactic using MBR followed by RO or NF could prove the most effective in pollutant removal followed with a disinfectant at the end.