Peak fitting and modelling of NOM removal by coagulation from River Murray water in South Australia

Abstract

Removal of natural organic matter (NOM) remains a major challenge confronting the water industry in Australia, in the supply of safe drinking water. Its treatability can vary depending on its character and this may be enhanced by optimising the inorganic coagulant dose and depending on the coagulant type, the coagulation pH. Generally, in coagulation modelling input parameter like pH, UV abs 254nm, turbidity and colour have been used in modelling approaches. However UV abs 254nm and colour does not comprehensively describe the character of NOM and improved characterization techniques should lead to better assessment of treatability of water by conventional treatment process. Investigation was conducted on data obtained from high performance size exclusion chromatography (HPSEC) of NOM to assess the potential of using this for estimating the removal of dissolved organic carbon (DOC) under simulated conventional treatment process. Controlled jar testing using various doses from low to high (60 to 200 mg/L) and at different coagulation pH levels, from 5.5 to 8.5 was conducted under laboratory conditions. Using raw water, collected from the Morgan Water Treatment Plant (WTP), water quality parameters such as pH, UV abs 254nm, DOC, turbidity, colour and HPSEC before and after treatment were recorded. The basis of this study was to; • Determine the removal of DOC in relation to its character by application of various coagulant doses. • Determine the applicability of using a peak fitting technique to analyse HPSEC data for determination of treatability of NOM in water A peak fitting technique was used to evaluate data of HPSEC analysis to explore the character of organics peak fitting data was obtained for modelling after resolving a range of peaks, representing organics with various molecular weight profiles from the HPSEC chromatogram. The fitted peak areas of organics in raw water were calculated and compared with those of treated water to investigate the removable or non- removable components by using low to high alum doses. Much less removal of lower molecular weight (non- humic substances) compounds occurred compared with removals of higher molecular (humic substances) compounds showing the lower molecular weight compounds are generally recalcitrant to removal by coagulation. Models were developed that relate HPSEC peak area removal with alum dose, and molecular weight with their removals at controlled pH. The basis of this study was to establish suitable tools for evaluation and identification of coagulable NOM fractions and optimum treatment conditions. Here, we report the development of a model specially designed using peak fitting data for HPSEC molecular weight profile of treated water, to enable prediction of percentage area removal by coagulation. The key aim is to provide reliable information of the character of NOM in order to optimize treatment for its removal. The model was developed from data of waters from the Morgan Water Treatment Plant (South Australia).