The removal of anionic surfactant from commercial laundry wastewater with reverse osmosis membrane

Abstract

Fresh, clean water has always been critical for the world's social development. The current water scarcity will only worsen unless measures are put in place to either reduce water usage or clean and reuse greywater. In areas with limited water resources, affordable technologies can be used to treat greywater and increase the water supply. Greywater sources that can be reused include domestic, hospital and industrial laundry wastewater. These wastewaters contain different chemicals such as organic and inorganic constituents, which make it difficult to treat. Microfiltration and ultrafiltration are examples of physical filtration processes that can reduce turbidity and pathogens sufficiently, but struggle to remove organics. Therefore, implementing an additional step such as reverse osmosis (RO) could be the solution in the removal of harmful chemicals in greywater. Unfortunately, the salts that are removed from the water, precipitate on the membrane surface, thus, decreasing the overall process efficiency, due to fouling and scaling. Scaling causes decline in permeation flux, degeneration of membranes, production loss and higher operating costs. This occurrence of fouling cannot be completely isolated; however, it can be minimised. There are two approaches for dealing with the fouling effect, namely, minimization and remediation. Remediation focuses more on frequent chemical cleaning. By using suitable pre-treatment measurements upstream of RO, scale formation can be minimised. In this study, the use of a commercial antiscalant was examined in the treatment of laundry wastewater influent. The removal of anionic surfactants and COD’s from this effluent with a low-pressure, extra low energy, reverse osmosis membrane for reuse application was investigated. The effect of different laundry detergent feed concentrations on operational parameters such as the membrane salt rejection and permeate flow rate (flux) was also analysed. The effect of different antiscalant concentrations to minimise scaling was also evaluated. Membrane fouling and remediation was evaluated by selected membrane surface characteristics. Model laundry wastewater was treated using a bench-scale reverse osmosis unit. The effects of laundry detergent concentration and antiscalant dosage on the permeate flow rate (flux) and rejection characteristics of the membrane were examined. Removal efficiencies for surfactant and COD concentration were analysed as an indication of membrane performance. A detailed examination of membrane fouling was done by investigating membrane surface characteristics using Scanning Electron Microscopy (SEM); Attenuated Total Reflection-Fourier Transform Infrared Spectroscopy (ATR-FTIR) and Energy Dispersive X-Ray Spectroscopy (EDX), before and after antiscalant addition. Design Expert 11 was used to generate a predictive model to describe the behaviour of permeate flux decline over time. ATR-FTIR revealed all the characteristic peaks on a virgin extra low energy (XLE) polyamide thin film composite membrane, in its clean state. It was observed that more foulant is deposited onto the surface of membranes with lower or no antiscalant dosage compared to the higher antiscalant dosed membranes. A morphological change of the membranes was observed using SEM analysis. The hindered attachment of scalant on the surface of the membranes resulted in a much lower rate of flux decline when compared to membranes with no antiscalant addition. EDX revealed that the amount of carbon decreased with an increase in laundry detergent amount (concentration). This could be due to the carbonyl group present in the PA layer being masked by the foulant layer. The flux decline could be associated with the fouling phenomenon caused by the accumulation of anionic surfactant molecules on the membrane surface, where the build-up of a concentration polarisation layer and/ the or the entrapment in the polyamide layer. Surfactant rejection exceeded 99.8% in almost all the experimental runs over a range of varied feed concentrations. An average COD removal throughout was 91-96%. It must be noted that the COD removal during the Percentage removal (COD and average EC) of the membranes are all significantly high, between 96-98% removal for average EC and between 91-96% removal for COD, however it was observed that membranes with membranes with no anti-scalant addition performed slightly better than membranes with anti-scalant dosing. It was observed that the predictive model successfully described the permeate flux decline of laundry wastewater using an RO membrane within the design space of the model. It can be confirmed that the membrane performance investigated using model laundry wastewater could be improved when using commercial antiscalant.