Study of Arsenic removal from aqueous solution using spiral wound nanofiltration membrane

Abstract

newlineArsenic contamination in drinking water is recognized as major health hazard worldwide. In natural water, arsenic is available in two dominant toxic inorganic forms, i.e. arsenic (III) and arsenic (V) and both are toxic in nature. However, arsenic (III) is more toxic than arsenic (V). In this work, arsenic (III) removal was studied using HFN300 Polyethersulfone Nanofiltration membrane in spiral wound configuration at Pilot scale. The influence of various operating parameters such as transmembrane pressure (2.98-8.46 bar), initial feed concentration (10-50 mg/L), feed flowrate (9-17 LPM) and initial solution pH (4-10) on membrane performance were examined. The permeate concentration was analysed using Atomic Absorption Spectrophotometer for arsenic (III) concentration. It was found that rejection was dependent on pressure, feed concentration, and solution pH. Results showed that more than 96.4 % arsenic (III) rejection was achieved for 50 mg L-1 of feed at optimized conditions. The rejection percentage was found to increase from 84 % to 93 %, while increasing initial solution pH from 4 to 10. The effect of co-ion concentration [monovalent (Cl-) and divalent (SO42-)] in feed also studied on membrane performance. The experimental results showed that increasing co-ion concentration (40 200 mg/L) in feed decreases the rejection efficiency of trivalent arsenic. The competitive nature between ions can be attributed to the decrease in rejection efficiency along with electro-migration repulsion of co-ions. Correspondingly, the permeate flux was also decreased with increasing co-ion concentration due to concentration polarization effect. newlineThe membrane performance was modelled over a wide range of solution pHs (4-10). Arsenic (III) is assumed to be in uncharged solute form at below pH 9.2. Therefore, film theory combined Spiegler-Kedem (FTCSK) model which covers the transport mechanism of diffusion and convection was suitably selected to predict the membrane performance at pH 8. Accordingly, membrane parameter