Remote community drinking water supply : mechanisms of uranium retention and adsorption by ultrafiltration, nanofiltration and reverse osmosis
Schulte-Herbruggen, Helfrid Maria Albertina
Worldwide, around 884 million people lack access to safe drinking water. To address this, groundwater sources such as boreholes and wells are often installed in remote locations especially in developing countries. However, the natural chemical composition of groundwater may be a source of toxicity to human health. Uranium is naturally present in the environment, and concentrations above the World Health Organisation (WHO) drinking water guideline (15 μg/L) are found in various parts of the world. Uranium has a complex aqueous chemistry and its speciation, which varies according to pH and available ligands, determines its behaviour (e.g. mobility, reactivity or sorption tendency). Nanofiltration and reverse osmosis have proved effective in removing uranium from water, although fundamental removal mechanisms are not well understood. Even the more porous ultrafiltration (UF) has been shown to remove uranium when used in combination with complexation/coagulation methods. To address the water purification needs of remotely located communities with no or unreliable access to energy, a renewable energy powered membrane system was designed using UF as pre-treatment to remove particles, bacteria and viruses and NF/RO to remove ions. The system was trialled in the Australian outback, using natural groundwater high in uranium (>300 μg/L). Results showed that pH had a large effect on the uranium behaviour in the system and, curiously, interaction by sorption or precipitation to the membranes was observed at certain pH values. However, due to the complexity of the water and the combination of UF and NF/RO membranes, the mechanisms of the uranium retention and interaction with the membrane were not clear. Further systematic study was needed to investigate the uranium behaviour with the membranes. Laboratory studies were carried out with one membrane type at a time: UF, NF and RO. It was postulated that pH, organic matter and inorganic ions such as calcium have an important influence on uranium retention and interaction with membranes. Results show that uranium behaviour in the membrane systems was highly pH dependent. During the UF experiments, increased adsorption of uranium occurred in uranium-only solutions at pH 5-7. From the UF experiments with organic matter it could be concluded that organic matter did not increase retention (size exclusion) of uranium, however it did increase the adsorption. Humic acid increased adsorption to 80-95% at pH 3-5, alginic acid at pH 3 while tannic acid caused a nearly 100% adsorption at pH 10-11. Further investigating uranium behaviour with NF and RO membranes, it was found that uranium showed the same increase in affinity to the membrane at pH 5-7, with about 50% being taken up by NF and 30% by RO membranes. The effect of pressure on uranium-membrane interaction was investigated for NF and RO at pH 6 and 8.5. Pressure and consequent concentration polarisation only increased uranium affinity to the NF membrane at pH 8.5 where the uranium species and MWCO of the membrane were similar. There was no or little effect of pressure on the affinity of uranium to the NF membrane at pH 6 or to the RO membrane. At pH 6, STEM-EDX results showed that uranium was distributed through-out the polyamide active layer of the NF membrane while FTIR results confirmed that uranium bound to carboxyl groups in the polyamide. At pH 8.5 however, FTIR results showed that uranium did not form chemical bonds with the membrane, but was rather attracted to the surface through hydrogen bonding and loosely forming a layer on top of the membrane visible in SEM. It was concluded that at least three different characteristics of the uranium species and membranes played a role for the interaction: 1) uranium species valency and membrane charge, 2) uranium species size relative to the membrane pore size, and 3) the reactivity of the uranium species towards the membrane functional groups. The effect of calcium on uranium retention and uranium-membrane interaction in NF and RO was also investigated. Calcium affects uranium speciation by forming a neutral complex with uranium at pH 8-9, causing a decrease in adsorption to the membrane. Calcium also precipitates at pH 10. SEM and TEM images showed that the precipitation of calcium carbonate (CaCO3) as calcite caused co-precipitation of uranium, trapping it on the surface of the membrane. About 48-55% of the calcium precipitated which caused a 26-35% co-precipitation of uranium, compared to <5% adsorption in the absence of calcium at pH 10. Finally the chemical drinking water quality of mainly boreholes and wells across a West African country, Ghana, was investigated (199 samples in total from “improved” sources). In addition, the user water costs were documented and the scope for advanced treatment explored. The WHO guidelines for chemical water quality were exceeded in 38% of the samples. The main contaminants were nitrate (21%), managanese (11%) and fluoride (7%), while heavy metals such as lead, arsenic and uranium were localised to mining areas. It was concluded that when taking the cost of unsuccessful borehole development into account, alternative treatment may be a suitable option where inorganic contamination is high. The findings from this study show the importance of the water quality conditions (pH, organic matter and calcium) on the behaviour of contaminants such as uranium in membrane systems and explain the mechanisms of adsorption and co-precipitation of uranium to the membranes at certain pH values. These are important considerations when selecting appropriate membranes for water treatment and also for the maintenance of membranes. The study also showed that there is need for advanced treatment of drinking water in e.g. Ghana, but highlights the importance of strategies on local and national level to ensure long-term sustainability and integration of any such treatment.