Disinfection may cause a dilemma for drinking water treatment plants using chlorine for the maintenance of a disinfectant residual in distribution systems. On one hand the chlorine residual should ensure microbially safe drinking water, but on the other hand harmful disinfection by-products (DBPs) can be formed from the reaction of natural organic matter (NOM) with chlorine. Many utilities are looking to the combined use of ozonation for primary disinfection, followed by chlorine or chloramines as a means of minimizing DBP formation while maintaining a stable disinfectant residual. However, these combinations may lead to a new spectrum of by-products that differs from that produced when a single disinfectant is used. The formation of mixed ozonation/chlorination and ozonation/chloramination by-products is the subject of this dissertation. A diverse collection of precursor compounds that produce a large amount of “unknown” total organic halides (TOX) was identified by performing bench scale tests to simulate chlorination of known ozonation by-products. Simple mono- and di-carboxylic acids were not found to react with chlorine. Di-aldehydes, α-keto-acids, and α-hydroxy-acids are oxidized by chlorine but do not show TOX formation. However, chlorine does become incorporated in β-diketones. Oxalacetic acid, 3-methyl-2,4-pentanedione, acetonedicarboxylic acid, and malic acid were found to form more “unknown TOX” than common chlorination by-products. Usually, the chlorine demand as well as the TOX increase with decreasing pH and increasing chlorination time. The identification of “unknown TOX” was the second major goal of this work. Quenched samples from the model compound studies were derivatized with pentafluorobenzylhydroxylamine (PFBHA), extracted with methyl-tert-butyl-ether (MtBE), and silylated with bis-(trimethylsilyl)-triflouroacetamine (BSTFA). Malic and acetonedicarboxylic acids were each found to produce a previously-unknown byproduct after reacting with chlorine. The identity of this and other new by-products was suggested based on the mass spectra. Surprisingly, mono-chlorinated species were found to be more abundant than di-chlorinated species in all cases. The third phase of this research showed that a substantial amount of “unknown TOX” is also formed in distribution systems where chlorine is used as final disinfectant. The “unknown TOX” ranged between 60% and 80% of the measured TOX.

Water utilities in the U.S. and Europe are investigating a wide range of treatment process improvements to meet more stringent water quality regulations (e.g., the U.S. D/DBP Rule). A relatively new process that has shown promise is catalytic ozonation, the use of a metal catalyst in conjunction with ozone. Laboratory-scale experiments were performed using a stirred batch and semi-continuous ozonation apparatus to study the catalytic ozonation process. Initial experiments were designed to evaluate dissolved metal ion assisted ozonation of oxalic acid at pH 6. The catalytic properties of dissolved cobalt (II) were subsequently studied in more detail. The results suggested that the first step in the catalytic ozonation reaction pathway is the formation of a cobalt (II) oxalate complex. Cobalt (II) oxalate is then oxidized by ozone. Further experiments showed that the rate of cobalt (II) assisted ozonation of oxalic acid increased with decreasing pH over the pH range of 5.3 to 6.7. The dissolved cobalt (II) studies provided the basis for understanding the catalytic properties of cobalt (II) oxide. It is anticipated that solid phase catalysts have more practical applications than dissolved catalysts in water treatment. The catalytic reactivity of two model di-carboxylic acids, oxalate and malonate, were compared in a series of laboratory experiments. In-situ attenuated total reflectance infrared spectroscopy indicated that oxalate formed an inner sphere complex and malonate formed an outer sphere complex with cobalt (II) oxide. Catalytic ozonation was evaluated for removal of pCBA, a non-adsorbing model micropollutant that does not react directly with molecular ozone. Cobalt (II) oxide and a mixed metal oxide (copper and zinc oxide with a calcium aluminate binder) either did not change the removal or they inhibited the removal of pCBA in deionized water compared to ozone alone. Alumina supported ruthenium also accelerated the removal of pCBA from a natural water, but it may follow a different catalytic reaction pathway. (Abstract shortened by UMI.)