Malodorous dimethylpolysulfides in Perth drinking water.

Abstract

The formation of an objectionable "swampy" odour in drinking water distribution systems in Perth, Western Australia, was first described by Wajon and co-authors in the mid-1980s (Wajon et al., 1985; Wajon et al., 1986; Wajon et al, 1988). These authors established that the odour, variously described as "swampy", "sewage" or "cooked vegetable" was caused by dimethyltrisulfide (DMTS) which has an odour threshold concentration of 10 nanograms per litre (ng/L). Investigations described in the present Thesis extend the work of Wajon and co-workers in attempting to establish the origin and cause of DMTS formation in Perth drinking water distribution systems.The DMTS problem appeared to be confined to water originating from a particular type of groundwater, specifically groundwater sourced from shallow, unconfined aquifers, which contain relatively high concentrations of sulfide, dissolved natural organic matter (NOM) and dissolved iron. DMTS was not present in the groundwater, but only formed in the distribution system, after treatment of groundwater via alum coagulation-filtration and oxidation processes. One objective of the present work was to determine the reasons for the observed association between DMTS formation and this specific groundwater type. A primary focus was to investigate the chemistry and biochemistry of sulfur species and NOM which might act as precursors to DMTS. The work was driven by the view that increased understanding of the problem might lead to more effective and acceptable treatment solutions than those presently in use.The observation that DMTS forms in distributed water that originates from groundwater, but not in water from surface sources has led to the hypothesis that groundwater NOM may contain precursor(s) to DMTS For example, it was proposed that methyl esters and ethers within humic substances might be a source of methyl groups that could participate in DMTS formation in distributed water (Wajon and Heitz, 1995; Wajon and Wilmot, 1992). Further, comparison of levels of reduced sulfur with levels of dissolved organic carbon (DOC) in groundwaters feeding Wanneroo GWTP revealed that a positive correlation between these two parameters existed. This observation provided further impetus to examine the nature of NOM in these groundwater systems. In the present study (discussed in Chapter 3), NOM from two Perth drinking water sources was isolated and characterised, with the aim of identifying major differences in structure and/or functional groups that might influence DMTS formation. NOM was isolated from water samples using ultrafiltration, and characterised using pyrolysis gas chromatography-mass spectrometry (Py-GC-MS) and offline- thermochemolysis/methylation (TCM). Pyrolysis of groundwater NOM yielded a high proportion of organosulfur compounds, primarily methyl thiophenes and sulfur gases, but did not yield detectable amounts of methoxy-aromatic compounds. Analysis by TCM yielded sulfur compounds tentatively identified as the methyl esters of methylthiopropanoate and methylthiobutanoate, compounds that may arise as degradation products of dimethylsulfoniopropionate (DMSP), an algal odmoregulator Compounds such as DMPS could potentially undergo reactions to form DMTS in distributed water.The task of investigating the formation of nanogram-per-litre concentrations of DMTS demanded the development of new analytical procedures that could be used to determine similarly low concentrations of DMTS precursors. Evidence existed to suggest that inorganic polysulfides could be plausible precursor compounds, and since no technique existed to analyse and quantify individual polysulfide homologues a new technique needed to be developed and verified. The technique, first used in a semiquantitative manner by Wajon and Heitz (1995), utilizes methyl iodide to derivatise polysulfides in-situ. The technique was developed further and shown to be quantitative and specific for inorganic polysulfides. Further, a new procedure for the determination of d i methyl polysulfides (DMPSs; CH3SnCH3, where n = 2-5), based on purge and trap was developed. In this new procedure analytes were trapped on a "Grob" activated charcoal tube, which was integrated into a commercially available, automated purge and trap instrument. Perdeuterated analogues of the DMPS analytes were synthesized and used as internal standards. These modifications resulted in a more rapid and robust procedure than the previously used procedures, vii which were based on closed loop stripping analysis (CLSA). Validation of the precision, accuracy, linearity and robustness of the new procedures for both inorganic polysulfides and dimethylpolysulfides is described in Chapter 4.Previous authors (Wajon and Heitz, 1995; Wajon and Wilmot, 1992; Wilmot and Wajon, 1997) hypothesized that DMTS could arise in the distribution system from residual polysulfides or other reduced sulfur compounds originating from groundwater. The latter authors showed that a small proportion of sulfide in the groundwater was not completely oxidised to sulfate during the water treatment process and proposed that this residual reduced sulfur fraction, which they referred to as non-sulfide reduced sulfur (NSRS) could contain precursors to DMTS. In a review of the chemistry of sulfide oxidation (Chapter 2) it was shown that the most likely forms of sulfur comprising the NSRS that enters the Wanneroo distribution system are organosulfur compounds and elemental sulfur, probably associated with organic matter in the form of a sulfur sol.Analysis of inorganic polysulfides in treated water, using the newly described method in Chapter 4, revealed that small amounts of these compounds (20-80 ng/L) were occasionally present in some samples. However, it was concluded that, since inorganic polysulfides could not survive water treatment processes, these compounds probably arose from traces of biofilm or pipe sediment that may have entered the water during sampling. It was proposed that the presence of biofilm particulates in water samples probably also accounted for observations that DMTS appeared to form in some water samples during storage of the sample. These studies are discussed in Chapter 5.The primary method of control of DMTS formation in the distribution system has been to maintain free chlorine residuals. However, the mechanisms by which this occurs have not been studied; the effectiveness of DMTS oxidation by chlorine, or how chlorine affects microbial processes that might form DMTS is not known. These issues are addressed in the final section of Chapter 5. Experiments to determine the effectiveness of oxidation of dimethyldisulfide (DMDS) and DIVITS (5 mu g/L) by free chlorine (0.2 to 0.6 mg/L) in distributed water showed that these substances are rapidly and completely oxidised in water containing a chlorine residual of more than 0.4 mg/L. However, slow regeneration of traces of DMDS and DIVITS after dissipation of free chlorine to non-detectable levels showed that these compounds were incompletely oxidised at the lower chlorine concentrations~ This provides some rationale for field observations that DIVITS occurs even where low, but measurable, chlorine residuals appear to exist (<0.2 mg/L).As was established in a review of the chemistry of reduced sulfur compounds Chapter 2), reducing conditions not present in the oxic bulk water are required for DMTS to form and to persist. It was therefore proposed that microbial reduction processes could generate anoxic microniches in the distribution system, within which DMTS production could occur. This hypothesis was investigated in Chapter 6; the new methods for analysis of organic and inorganic polysulfides were applied to the study of biofilms and deposits of colloidal material found in distribution pipes and storage reservoirs. The study demonstrated that these materials contained concentrations of methylated and inorganic polysulfides four to six orders of magnitude higher than those ever found in the bulk water phase. The results indicated that reducing conditions most probably exist within the biofilms and pipewall deposits, where these polysulfides were formed. The iron-rich pipe slimes appeared to protect the sulfur compounds against the oxidative effects of chlorine and dissolved oxygen. It was concluded that the organic and inorganic polysulfides most probably arise through microbial sulfate reduction processes that occur in anoxic microenvironments within the slimes and deposits.Microbial processes that lead to the formation of polysulfides and dimethylpolysuifides under conditions approximately representative of those in distribution systems were investigated in work described in Chapter 7. The aim of this work was to investigate the role of biofilms in the formation of DMTS and to determine the nature of chemical precursors which might stimulate these processes. Biofilms, artificially generated on synthetic supports within chambers filled with water from Wanneroo GWTP, were exposed to compounds thought to be potential DMTS precursors. The response of the systems in terms of production of methylated sulfur compounds was monitored. Conclusions of the study were that, under the test conditions, production of DMDS and DMTS could occur via several mechanisms and that these dimethyloligosulfides could be formed even without the addition of compounds containing sulfur or methyl moieties. DMTS did not form in the absence of biofilms and it was therefore concluded that minimisation of biofilm activity was a key in preventing DMTS formation. Outcomes of the work imply that environments within distribution systems are complex and dynamic, as perhaps manifested by the intermittent nature of the DMTS problem.Finally, in Chapter 8 the conclusions to the present studies are summarised. It is shown how they underpin the rationale for proposed new treatment solutions aimed at preventing DMTS problems in the Wanneroo zone, primarily by minimising microbial activity and biofilm formation within distribution systems.