Advancing the analytical characterization of natural organic matter in water

Abstract

The right to clean water is a fundamental human right, the responsibility for which rests upon water service institutions to ensure that the water supplied to consumers is free of and safe from any harmful substances, which is also important for sustainable livelihoods and a healthy economy. As such, when treating water for drinking purposes, it is crucial to consider the composition of natural organic matter (NOM) in the source water and during the treatment process. NOM is a complex mixture of organic compounds, such as humic and non-humic substances with a range of chemical properties, which are present in various environments of the ecosystem such as soils, natural surface waters, ground waters and sediments. NOM may contribute to the yellow-brown colour of water; it can incur an unpleasant taste or odour to water; and it contributes to the bacterial regrowth and corrosion in water distribution networks. Additionally, NOM in water may react with disinfectants to produce potentially carcinogenic disinfection by-products (DBPs). Therefore, there is a growing need for NOM research into combining conventional and advanced analytical techniques to provide a robust tool for its efficient characterization, thereby guiding drinking water treatment. A comparative study was undertaken to gain a better understanding of the composition of NOM in water across different seasons by characterizing NOM present in source water and after rapid gravity sand filtration at two drinking water treatment plants of Africa’s largest bulk water provider, situated in Gauteng. This was achieved by means of conventional techniques: dissolved organic carbon (DOC), ultraviolet absorbance at 254 nm (UV254) and specific ultraviolet absorbance (SUVA), as well as advanced techniques such as solid phase extraction (SPE) and fluorescence excitation-emission matrices (FEEM). Further, a novel approach using ultra-performance liquid chromatography - high resolution mass spectrometry (UPLC-HRMS) was developed to enhance the understanding of the composition and concentration of NOM in water. Before the characterization of the sampled water, aspects of sample storage and preparation were explored to ensure sample integrity was maintained during transportation and storage, thereby contributing to accurate reporting of results. As such, a range of storage conditions and durations were assessed and it was found that storing water samples for up to 34 days in the dark, at room temperature or at 5 oC gave more similar results to the initial sample measurements than storage under direct sunlight or UV radiation, as determined by DOC and UV254 measurements. As a result, it was recommended that samples are stored at 5 oC in amber storage containers for a maximum of 34 days to maintain sample integrity and minimise degradation. Additionally, syringe and membrane filters from different brands were pre-washed with ultrapure water to investigate the presence of possible residual contaminants from the filter material that may leach into the water samples during sample preparation and thereby impact on NOM analysis. The results indicated that there were indeed retained UV-active contaminants in the filter materials of which 81 to 91% were removed after washing with 25 mL of ultrapure water. Furthermore, there were no significant differences in the UV254 measurements of the 25 mL pre-wash eluate fraction across the different brands. For this reason, the deciding factors for the chosen filter employed in this study were the amount of UV-active contaminants present in each filter in the first wash, the cost of the filters, their ease of use and the amount of plastic/consumable waste generated. Following the seasonal collection of the water samples, characterization using the conventional techniques indicated changes in NOM character during the sampling period with season which were also due to the location of the point of water abstraction for each treatment plant with differences in discharges into the source water. The Upper Vaal catchment area source water NOM consisted mainly of hydrophobic organic material of high molecular mass and aromaticity. Moreover, the obtained results indicated that both treatment plants were consistent with removing NOM content across different seasons, regardless of the differing raw water quality. DOC and UV254 results obtained from the cost-effective, portable, battery operated UV254 Go! analyzer were compared to results obtained from conventional laboratory based instruments, to assess its use for field-based measurements. There was a strong correlation between the UV254 measurements obtained with these instruments, which showed the potential of the UV254 Go! as a convenient and cost-effective means to directly determine NOM aromaticity in the field. However, the UV254 Go! reported higher DOC measurements than the conventional total organic carbon (TOC) analyzer, as a result more experiments are needed to determine an empirical correction factor to correct for the difference. Fractionation using the established modified polarity rapid assessment method (mPRAM) sorbents (namely C18, CN and NH2) and alternative sorbents (hydrophilic-lipophilic-balanced (HLB) and Carbon S (CS)), which had not been previously investigated for this application, showed promising results. The obtained fractions were measured for DOC and UV254 which showed that the water samples from both treatment plants were mainly composed of non-polar hydrophobic (HPO) components across all seasons, with UV254 measurements decreasing after treatment. Regardless of its water-wettable properties, the alternative HLB sorbent fractions showed a stronger retention of the somewhat less polar NOM components. The results indicated a stronger elution solvent is required. The alternative CS sorbent showed stronger retention than the HLB sorbent and holds potential in the fractionation of non-humic acid and non-fulvic NOM components in water. The observed variability in the respective fractions highlighted the importance of detailed NOM monitoring and characterization. FEEM analysis of the unfractionated seasonal water samples indicated that NOM components varied per treatment plant, season and between the raw and treated water, and further indicated that the source water from the Upper Vaal catchment area is primarily composed of aromatic and marine humic acid-like components. UPLC-HRMS analysis for NOM characterization revealed that NOM components are better ionized in the negative electrospray ionization (ESI) mode and thus better represent the composition of NOM than ionization in the positive ESI mode. Due to the extensive number of peaks in the ESI mass spectra, chemometrics was employed to deconvolute the data, revealing similarities and differences between seasonal water samples and SPE polarity fractions. Principal component analysis (PCA) score plots showed a clear difference in NOM composition between the seasons with winter and autumn indicating chemical similarity. Additionally, PCA highlighted that the raw water and water after sand filtration (treated) were similar to each other but dissimilar to the mPRAM fractions. With regards to the HLB sorbent, PCA showed poor separation of the respective fractions, suggesting either lack of strong retention of NOM components or very strong retention of some components which were not eluted with the solvents used. Unlike the mPRAM sorbents, the HLB and CS sorbents could successfully differentiate between the winter and autumn water fractions, showing potential in the use of these sorbents for NOM fractionation. The CS sorbent fractions were dissimilar to each of the mPRAM sorbent fractions, proving that the sorbent may add a new dimension to the PRAM technique, as it provides a novel separation of NOM components. Overall, this study contributes to informing drinking water utilities about seasonal NOM variability and fluctuating source water quality, and supports the implementation of adaptive treatment processes to maintain safe drinking water. It provides novel approaches to more comprehensive NOM monitoring and characterization, providing valuable insight into the efficiency of the treatment processes employed at leading drinking water treatment plants, thereby confirming the quality of water supplied to South African consumers.