Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Evaluation of the removal of disinfection by-products precursors in Seymour water with ozone, UV and… Chin, Adeline S. 2003

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2004-0100.pdf [ 4.96MB ]
Metadata
JSON: 831-1.0063516.json
JSON-LD: 831-1.0063516-ld.json
RDF/XML (Pretty): 831-1.0063516-rdf.xml
RDF/JSON: 831-1.0063516-rdf.json
Turtle: 831-1.0063516-turtle.txt
N-Triples: 831-1.0063516-rdf-ntriples.txt
Original Record: 831-1.0063516-source.json
Full Text
831-1.0063516-fulltext.txt
Citation
831-1.0063516.ris

Full Text

Evaluation of the Removal of Disinfection By-Products Precursors in Seymour Water with Ozone, UV and the Advanced Oxidation Process Ozone-UV B.Sc. Life Sciences, McMaster University, 1997 B.Sc. Honours Neural Computation, McMaster University, 1999 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF T H E REQUIREMENTS FOR T H E D E G R E E OF MASTER OF APPLIED SCIENCE by Adeline S. Chin in The Faculty of Graduate Studies (Department of Civil Engineering) We accept this thesis as conforming to the required standard Em T H E UNIVERSITY OF BRITISH COLUMBIA December 1, 2003 © Adeline S. Chin, 2003 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my depart-ment or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of Civ i l Engineering The University Of British Columbia Vancouver, Canada Date December 8 ) 6*005 Abstract ii Abstract This research evaluated the treatment potential of ozone (O3), ultraviolet irradia-tion (UV) and the combined O3-UV advanced oxidation process (AOP) in the re-moval of two classes of disinfection byproduct (DBP) precursors (quantified as the trihalomethane formation potential (THMFP) and haloacetic acid formation poten-tial(HAAFP)) from water samples obtained from the Seymour Reservoir near Vancou-ver, Canada. The water from this source, which provides 40% of the drinking water for the Greater Vancouver Regional District, is characterized by low pH (6.4) and low total organic carbon (1.3 — 3.3mg/L) [1]. Laboratory batch scale experiments were carried out at various UV and O3 dosages in order to monitor the reaction kinetics. The combined O3-UV treatment significantly reduced the concentration of both DBP precursors. The reduction is attributed to the mineralization of the total organic carbon (at a rate of 0.044±0.02 minute - 1) and changes in organic constituents (~ 78% decreased in U V 2 5 4 absorbance). After 30 minutes of O 3 - U V treatment, THMFP and H A A F P were reduced by approximately 80% and 70%, respectively. Ozone treatment alone provided roughly 50% reduction of both T H M F P and H A A F P after 30 minutes of ozonation. Therefore, the O3 treatment proved not to be as effective as the combined O3-UV AOP. Ultraviolet treatment alone proved to be relatively ineffective in reducing the concentrations of the DBP precursors studied. These results indicate that the combined O3-UV AOP is a useful treatment option that could be implemented to reduce the concentration of DBP precursors in Seymour's water. Contents iii Contents Abstract ii Contents iii List of Tables v List of Figures vi Acknowledgements viii 1 Introduction 1 2 Relevant Background Information 3 2.1 Natural Organic Matter (NOM) 3 2.2 Disinfection Byproducts (DBPs) 4 2.3 Treatment Options 7 2.3.1 Ultraviolet Irradiation (UV) 7 2.3.2 Ozone (0 3) 9 2.3.3 0 3 - U V Advanced Oxidation Process (AOP) 14 3 Objectives 18 4 Experiment Design 19 4.1 Seymour Source Water Characterization 19 4.2 Laboratory Scale 20 4.3 Pilot Scale 23 5 Methodology 24 5.1 Laboratory Preparation 24 5.1.1 Glassware 24 5.1.2 Reagents and Blanks 24 5.2 Analytical Methods 24 5.2.1 Total Organic Carbon (TOC) 25 5.2.2 pH and Temperature 25 5.2.3 Ultraviolet Absorption at 254 nm (UV 2 5 4 ) 25 5.2.4 Disinfection Byproduct Formation Potentials (DBPFPs) . . . 25 5.2.5 Trihalomethane (THM) 26 5.2.6 Haloacetic Acid (HAA) 27 Contents iv 5.2.7 Residual Aqueous Ozone 30 5.2.8 Residual Ozone in Gas Phase . 30 6 Results and Discussion 31 6.1 Ultraviolet Irradiation (UV) 31 6.2 Ozone (03) 36 6.3 Combined Ozone-Ultraviolet AOP 47 6.4 Comparison of the 0 3 and 0 3 -UV Treatments 59 6.5 Pilot Plant Results 61 7 Conclusions 65 8 Recommendations 67 8.1 Disinfection Potential 67 8.2 Ozone Disinfection Byproduct Measurement 67 . 8.3 Feasibility of AOP without Filtration 67 8.4 Reactor Design and Kinetic Modelling 68 8.5 Water Quality Factors 68 8.6 Turbidity 69 Bibliography 70 A Ozone Generator Characterization 77 B Hydraulic Characterization of the Systems 79 B.l Laboratory System 79 B.2 Pilot Plant System 79 C Mass Balance with Ozone and Combined Ozone-UV 81 List of Tables v List of Tables 2.1 A list of common halogenated disinfectant byproducts (DBPs) 4 2.2 Factors affecting T H M and HAA formation in raw water 6 2.3 Summary of literature review for the ozone treatment 13 2.4 Theoretical amount of oxidant required to yield a mole of -OH 15 2.5 Theoretical yield of -OH based on photolysis of O3 or H2C>2 16 2.6 Summary of literature review for O3-UV AOP 17 4.1 Seymour source water quality parameters 19 5.1 GC temperature program for CHC1 3 analysis 27 5.2 GC temperature program for HAA3 analysis '29 5.3 Quantification ions and retention times for HAA3 analysis 29 6.1 Pseudo first order reaction fit parameters for TOC (O3-UV). . . . . . 48 6.2 Pseudo first order reaction fit parameters for chloroform formation potential (0 3 -UV) 56 6.3 Pseudo first order reaction fit parameters for HAA3FP (0 3 -UV). . . . 56 List of Figures vi List of Figures 2.1 Nomenclature for natural organic matter (NOM) 3 2.2 Electronic configuration of molecular ozone 10 2.3 Mechanisms of ozone decomposition 12 2.4 O3-UV reaction pathways in water 15 4.1 A schematic of the laboratory scale batch apparatus 20 4.2 Picture of the laboratory scale batch apparatus 22 6.1 T O C and U V 2 5 4 versus time (UV) 33 6.2 DBPFP versus time (UV) 34 6.3 Speciation of HAA versus time (UV) 35 6.4 Ozone consumption versus time 37 6.5 T O C versus time and ozone consumption (O3) 39 6.6 UV2B4 versus time and ozone consumption (O3) 40 6.7 THMFP versus time and ozone consumption (O3) 43 6.8 HAAFP versus time and ozone consumption (O3) 44 6.9 pH versus time and ozone consumption (03) 45 6.10 Speciation of HAA versus time (03) 46 6.11 Ozone consumption and UV dosage versus time 47 6.12 T O C versus time and ozone consumption (O3-UV) 50 6.13 U V 2 5 4 absorbance versus time and ozone consumption (O3 -UV). . . . 51 6.14 Chloroform formation potential versus time and ozone dosage (O3-UV). 52 6.15 HAA3FP versus time and ozone dosage (O3-UV) 53 6.16 Speciation of HAA versus time (O3-UV) 54 6.17 pH versus time and ozone consumption (O3-UV) 55 6.18 Correlation between TOC and DBPFP ( O 3 -UV) 58 6.19 Comparison of O3 and O3 — UV results 60 6.20 T O C concentration versus time (pilot scale) 62 6.21 U V 2 5 4 absorbance versus time (pilot scale) 63 6.22 THMFP versus time (pilot scale) 63 6.23 HAAFP vs. time (pilot scale) 64 6.24 pH versus time (pilot scale) 64 A . l Schematic of the in-gas measurement system 77 A. 2 Ozone generator performance 78 B. l Hydraulic characteristics of lab batch system 80 B.2 Hydraulic characteristics of pilot plant batch system 80 List of Figures vii C l Mass balance for ozone in the laboratory batch system 82 Acknowledgements viii Acknowledgements This project would not have been possible without the assistance and support of the following people, whom I sincerely thank: • Dr. Don Mavinic for the inception of the project • Dr. Pierre Berube for his help and guidance throughout this project • Dr. Madjid Mohsenic for the use of his UV lamp and sharing his knowledge of AOP • Dr. Jim Atwater for his funding • Paula Parkinson for all her patience and help in interpreting and troubleshoot-ing the analytical methods and results • Susan Harper for ensuring that I had everything that I needed in the laboratory • Bill for building my reactor and miscellaneous parts • Harald, Scott and John for their technical help • Mary, Cythe, Ayesha, Christine, Sandi and Diane for their administrative as-sistance • GVRD for their use of the pilot plant • Peter Zadorozny and Ron Stracke from the GVRD Chemistry Lab for their help with the HAA method • Trojan Technologies for their donation of their UV Max C lamp • Richard for his love, encouragement and humour through this ordeal Chapter 1. Introduction 1 Chapter 1 Introduction Chlorine has been used extensively as a disinfectant against waterborne pathogens because it is an inexpensive process that produces long-lived disinfection residuals. However, the drawback with the use of chlorine as a disinfectant is that chlorine re-acts with naturally occurring organic matter (NOM) to form disinfection by-products (DBPs). These DBPs can be carcinogenic and may cause reproductive defects [2, 3]. Experiments on laboratory animals have demonstrated these effects [3]. Epidemiology studies have shown correlations between DPBs and reproductive defects in humans [3]. Elevated concentrations of trihalomethanes (THMs), a class of DBPs, are sus-pected to be the cause of a significant number of miscarriages by women living in Cheaksphere, Virginia [4]. Regulatory limits concerning concentrations of DPBs in water supplies have been imposed in some jurisdictions. The United States Environmental Protection Agency (USEPA) approved a Stage I Disinfection/DBP Rule in November 1998 which spec-ified the maximum contamination level for two particular classes of DBPs: a total T H M concentration of 80 fJ-g/L and a sum concentration of five key haloacetic acids (HAA5) of 60 pg/L, with compliance based on annual averages [2]. The Stage II Mi-crobial/DBP Rule, proposed in August 2003, did not alter the previously approved maximum contamination levels for DBPs, but required that the measured concentra-tions of DBPs must not be exceeded for all tests [3]. In Canada, a interim maximum acceptable concentration of 100 pg/Loi total THMs is suggested under the Guidelines for Canadian Drinking Water Quality [5]. At present there are no federal guidelines concerning haloacetic acids (HAA) concentrations. However, these are only guide-lines and the responsibility lies within the provincial and municipal governments to decide on the acceptable levels of DBPs in treated water. Water utilities are faced with the challenge of providing adequate disinfection of their raw water, but at the same time they must minimize the concentration of DPBs found in the treated water. This balance between disinfection and DBPs has forced many water utilities to re-evaluate their current treatment technology and has been the driving force behind recent research and development in the field of drinking water treatment. Chapter 1. Introduction 2 Advanced oxidation processes (AOPs) are a class of emerging technologies that shows promise for drinking water treatment [6, 7]. The principle underlying AOPs is the generation of hydroxyl radicals that are capable of oxidizing natural organic matter in the water matrix [8, 9]. Not only do the hydroxyl radicals oxidize a portion of organic compounds into carbon dioxide, but there is evidence that they also trans-forms DBP precursors into forms that do not readily react with chlorine. The most common AOPs are combined ozone-UV (O3-UV), hydrogen peroxide-UV (H2O2-UV) and hydrogen peroxide-ozone (H2O2-O3) [9]. This thesis summarizes research into the effectiveness of the O3-UV AOP in re-moving DBP precursors, specifically T H M formation potential (THMFP) and HAA formation potential (HAAFP), for raw water from the Greater Vancouver Regional District's Seymour Reservoir. The results are compared to those from treatment of the same source water using 0 3 and UV separately. Chapter 2. Relevant Background Information 3 Chapter 2 Relevant Background Information 2.1 Natural Organic Matter (NOM) The term natural organic matter (NOM) refers to the general group of substances derived from the decomposition of detrital materials in either terrestrial or aquatic environments [10]. These organic materials are generally composed of 50 — 60% carbon, 4 — 6% hydrogen, 23 — 40% oxygen and ~ 0.3% nitrogen by weight (with only trace quantities of phosphorus and sulfur). Any particular component of NOM is typically described as being either humic (nonpolar) or nonhumic (polar) [11, 12]. Humic substances (HS) are further subdivided into two categories, humic acids and fulvic acids, where the former describes the hydrophobic fraction that precipitates in a solution of pH 2.0 or less and the latter is the soluble fraction in the same solution [10]. In general, fulvic acids have lower molecular weight (less than 2000 Daltons) and are less aromatic compared to humic acids. They also have comparatively higher oxygen content, higher carboxylic acid and lower phenolic content [10, 12]. Humic acids are predominantly darker in color and have larger molecular weight (2000 to > 10000 Daltons). Figure 2.1 depicts the nomenclature hierarchy described above. Natural Organic Matter (NOM) Non-humic Substances Humic Substances Fulvic Acid Humic Acid Figure 2.1: Nomenclature for natural organic matter ( N O M ) [10]. The composition of NOM found within any given aquatic water body varies consid-erably, depending on the geography of the local watershed [13, 14]. Furthermore, the Chapter 2. Relevant Background Information 4 composition can vary with time due to factors such as seasonal fluctuations [12, 15]. 2.2 Disinfection Byproducts (DBPs) Disinfection byproducts (DBPs) are formed when a disinfecting agent reacts with NOM. Given that many municipal and regional drinking water providers in developed nations use chlorine as a disinfectant, the most frequently encountered DBPs are chlorinated byproducts. However, if bromine is to be found in the source water, then brominated DBPs can often be present as well. Table 2.1 lists the major classes of halogenated organic DBPs of concern that have been identified thus far. Table 2.1: A list of common halogenated disinfectant byproducts (DBPs) [2]. DBP Class Individual DBPs Chemical Formula Trihalomethanes (THMs) Chloroform Bromodichloromethane Dibromochloromethane Bromoform CHCl3 CHCl2Br CHClBr2 CHBr3 Haloacetic Acids (HAAs) Monochloroacetic acid (MCAA) Dichloroacetic acid (DC A A) Trichloroacetic acid (TCAA) Bromochloroacetic acid Dibromochloroacetic acid Monobromoacetic acid(MBAA) Dibromoacetic acid (DBAA) Tribromoacetic acid (TBAA) CH2ClCOOH CHCl2COOH CCl3COOH CHBrClCOOH CBr2ClCOOH CH2BrCOOH CBr2COOH CBr3COOH Haloacetonitriles Trichloroacetonitrile Dichloroacetonitrile Bromochloroacetonitrile Dibromoacetonitrile CCl3C = N CHCl2C = N CHBrClC = N CHBr2C = N Haloketones 1,1-Dichloroacetone 1,1,1-Trichloroacetone CHC12C0CH3 CC13C0CH3 Cyanogen halides Cyanogen chloride Cyanogen bromide CIC = N BrC = N Halopicrins Chloropicrin Bromopicrin CC13N03 CBr3N03 If ozone is used as a disinfectant, then other classes of DBPs can be formed, such as aldehydes, aldoketoacids and carboxylic acids [2] . However, these classes constitute only a small subset of the total list of DBPs that can be encountered in chlorinated Chapter 2. Relevant Background Information 5 water. However, according to Richardson (2002) [16] the current list of DBPs is probably incomplete, with over 50% of total number of halogenated DBPs (TOX) and more than 60% of the total number of DPBs produced by ozonation currently unidentified. In addition, less is known about DBPs produced by treatment of water containing NOM with disinfectants other than chlorine and ozone. Two classes of halogenated DBPs that have been extensively studied are the trihalomethanes (THMs) and the haloacetic acids (HAAs), both of which have been identified as having potential health risks for humans. As a result, many jurisdictions regulate their presence in drinking water. The ultimate concentrations and mass ratios of each types of DBPs in treated water are dependent on a number of factors such as temperature, pH, chlorine dose, and the concentration and chemical makeup of the NOM in a particular water source [17, 18]. Furthermore, the total time that the chlorine has to react with NOM significantly affects the final concentrations of DBPs [18-20]. This latter point is particularly relevant given that biofilm growing inside distribution systems can contribute to the formation of DBPs over time [21, 22]. A summary of general trends in the formation of halogenated DBPs is shown in Table 2.2. More detailed information can be found in each of the references cited therein. As stated above, the reactions that produce DBPs are affected by many factors. However, the production of THMs and HAAs has been reported to have two distinct temporal phases: a rapid formation followed by a slower rate, that can be mod-elled as a second order process [17, 18, 24, 31, 32]. Reckhow et. al. (1990) [30] found that, although raw water having different concentrations of NOM precursor materials yielded different final concentrations of chlorinated DBPs after the same treatment procedure, the formation curves had the same general shape for T C A A , D C A A and CHClz. Therefore, the DBP formation potential (DBPFP) is measured after a sufficient period of time (usually a week) to ensure that the reaction has reached completion and the maximum concentration is represented. Chapter 2. Relevant Background Information 6 Table 2.2: Factors affecting T H M and HAA formation in raw water. Factor Impact on T H M Concentra-tion Impact on H A A Concentra-tion Temperature Linear increase with increas-ing temperature [17, 23]. Linear increase with increas-ing temperature [17]. pH Curvilinear increase with in-creasing pH. No effect for pH >9.5 [17, 23]. Decrease with increasing pH, particularly for T C A A [17, 23, 24]. Chlorine Dose Rapid and curvilinear in-crease with increasing chlo-rine dose [17, 25, 26]. Curvilinear increase with in-creasing chlorine dose, espe-cially for T C A A [17, 24, 27]. Higher chlorine dosage results in higher proportion of tri-halogenated HAAs [8]. Bromide Generally shifts to the for-mation of brominated species [17, 28] but large chlorine dose can out-compete the for-mation of brominated THMs [29]. Shift the reactions to produce more brominated species [17]. Quantity k. Qual-ity of NOM Increases with TOC concen-tration, provided that the NOM are suitable precursor materials. Humic acids are found to be more reactive than fulvic acids [17, 30]. Increases with TOC concen-tration, provided that the NOM are suitable precursor materials. Humic acids are found to be more reactive than fulvic acids [17, 30]. Pipe Material/ Biofilm Increases with exposure to biofilm material as a function of time [21, 22]. Increases with exposure to biofilm material as a function of time. There is a tendency to produced more DCAA and less T C A A compared to the results found from bottle in-cubation [22]. Chapter 2. Relevant Background Information 7 2.3 Treatment Options The two most commonly used approaches to reducing the formation of DBPs are to: 1. use a non-chlorine based disinfectant, and 2. physically remove NOM from the water prior to chlorination. The use of a non-chlorine based disinfectant is feasible for the 'primary disinfec-tion' step of drinking water treatment. However, non-chlorine based disinfectants do not provide a lasting disinfection residual. Therefore, chlorine-based disinfectants are commonly used as a 'secondary disinfectant' to provide long lasting protection against the growth of pathogens throughout the water distribution system. The physical removal of NOM is typically achieved by first chemically converting the NOM into particulate material (e.g. coagulation/flocculation) and then using a physical removal technology (e.g. gravity settling, sand filtration, granular activated carbon filtration and membrane filtration) to remove the particulate NOM [33-37]. However, this approach is not effective at removing all of the NOM, especially the smaller fulvic acids. In addition, this approach requires an extensive physical infras-tructure. A third approach, which is gaining popularity, is to prevent the formation of DBPs by chemically converting the NOM into forms that do not readily react with chlorine. Specifically, treatment of raw water with O3, UV and O3-UV combined have been suggested as methods for provoking these chemical changes. The mechanisms underlying each of these treatment processes are briefly discussed below. 2.3.1 Ultraviolet Irradiation (UV) Ultraviolet irradiation utilizes the principles of photochemistry, whereby photons of a specific energy (hi/) are used to raise a molecule (M) to an excited electronic state [38, 39]. The excited electronic state can result in ionization, where an electron is expelled from the molecule: M + hv->M+ + e- (2.1) or in deactivation: M + hu^M* (2.2) where the molecule relaxes into an alternate groundstate configuration (M*) via the rupture of atomic bonds. Alternatively, the photoexcited molecule can transfer energy Chapter 2. Relevant Background Information 8 to other molecules with lower activation energies and result in deactivation occurring to these secondary molecules [39]. Low pressure mercury (Hg) lamps that emit strongly at a wavelength of 253.7 nm are a common choice for disinfection. Radiation at a wavelength of 253.7 nm is effec-tive at breaking bonds in nucleic acids and proteins, thereby leading to inactivation of microorganisms. This wavelength is also capable of breaking a large variety of other organic molecular bonds [39]. Low pressure Hg lamps are the choice for many water utilities because they emit most efficiently at 253.7 nm. Medium and high-pressure Hg lamps have the advantage of covering a broader wavelength spectrum (thereby opening the possibility for a greater variety of photochemical reactions) but they con-sume more energy in order to produce the same intensity of 253.7 nm radiation as their low pressure counterparts. There have been many studies conducted on the use of UV lamps for disinfection, but there is a noticeable knowledge gap in terms of its potential use as an oxidizing agent for the removal of DBP precursors. UV irradiation studies conducted by Malley, Shaw <k Ropp (1995) [40] on ground and surface water did not find any changes in TOC, UV254 absorbance, THMFP or H A A F P with UV dosages of 0.06 - 0.2 W • s/cm2. However, they did report increased concentrations of aldehydes. In that report, they quoted a study performed by Mogometry Watson Inc. (1994) [40] in which reclaimed wastewater was irradiated with a UV dosage of 2 — 2.8 W • s/cm2 (10 times the dosage that is required for practical disinfection) and found minimal effects on DBP formation potential (DBPFP). Legrini et al. (1993) [41] concluded that UV irradiation at 253.7nm alone is not effective in reducing the organics component in water. However, other studies performed with low pressure mercury lamps at higher UV dosages did find changes in the organic constituents and THMFPs . Li et al. (1996) [42] irradiated humic acids with a UV dose of 0.01 W/cm2 for up to 6 hours. There was a 67% reduction in TOC after 6 hours which corresponds to a dosage of 216 W • s/cm2. They also found an enhanced yield of T H M F P with UV irradiation. Mole, Fielding & Lund (1999) [43], irradiated Thames River water at UV dosage range of 1.1 to 13 i y • s/cm2 and found a 9% decrease in T H M F P at the highest UV dosage of 13 W • s/cm2. They also reported a decline in UV254 absorbance with increasing UV dosage. Another study that proported a change in THMFP with UV irradiation was con-ducted by Kleiser & Frimmel (2000) [44]. River Ruhr (Germany) water was irradiated with a low pressure Hg lamp of 3.3 W over a period of 240 minutes (288W • s/cm2). Chapter 2. Relevant Background Information 9 The authors reported that there was a 7% and 13% reduction of T H M F P after 30 and 240 minutes irradiation. Thomson, Roddick & Drikas (2002) [45] used UV as a pretreatment option for highly coloured water in East Moorable System, Victoria, Australia. They found a 16% reduction in DOC at a UV dose of 26 W • s/cm2 and the UV254 absorbance was reduced by half, which implied significant changes to the NOM. Although, they did not actually measured the specific disinfectant by-products, they incubated the sample for 8 days to determine the chlorine demand with titration. The chlorine demand for the irradiated water was only slightly higher than that for the raw water. A study conducted by Gallard &c von Gunten (2002) [32], supported that view. Lake Zurich water was irradiated at a dosage of 2160W • s/cm2 and it was found that there was no change in THMFP but a increased in chlorine demand. In summary, it appears that given a sufficiently high UV dosage, UV irradiation is able to reduce the concentration of TOC and reduce the CV254 absorbance to some extent. But in terms of THMFP, the results are not consistent. Regardless, it is questionable that the high costs associated with providing the high dosages required to obtained these changes warrants the benefits obtained. 2.3.2 Ozone (O s) The use of O3 as an oxidant to remove colour, taste and odour due to micropollutants in drinking water has been extensively studied [8]. However, few researchers have examined the feasibility of using O3 specifically as an oxidant in the removal of DBP precursors from source water. The details concerning the mechanism by which DPB precursors become oxidized in the presence of O3 are a subject of speculation and are not easily measured, particularly for heterogeneous mixtures such as NOM in raw water. However, it has been determined that there are two general mechanisms through which 0 3 can react with NOM: processes that directly involve molecular 0 3 and indirect processes involving hydroxyl radicals (-OH) that are formed by the decomposition of 0 3 in water [9]. As shown in Fig. 2.2, molecular 0 3 can act as a dipole. This makes it an ex-tremely selective reagent that prefers to react with unsaturated aromatic and aliphatic compounds as well as some specific functional groups. In general, molecular 0 3 will interact with NOM in one of the following ways [8] Chapter 2. Relevant Background Information 10 5 + o < • 0 o: .o: p. . q : • '8" 5 + ' 8" Figure 2.2: Electronic configuration of molecular ozone. (Adapted from Langlais et. al. (1991) [8].) 1. Cyclo addition (Criegee mechanism) - O3 reacts with double bonds to form cyclic structures. Next, water reacts with the carbon-carbon bond, which then decomposes the structure into carbonyl compounds (aldehydes and ketones). 2. Electrophilic reactions - This type of reaction is restricted to those NOM molecules that have sites with high electronic density (particularly aromatic compounds). The result is typically hydroxylated byproducts, which are susceptible to further ozonation. 3. Nucleophilic reactions - This type of reaction occurs when a NOM molecule contains sites showing an electronic deficit, particularly carbon atoms carrying electron withdrawing groups. As stated above, hydroxyl radicals (-OH) are formed from the decomposition of O3 in water. The decomposition rate is affected by pH, O3 concentration and the chemical content of the water matrix. The ozone decomposition process can be represented through a series of reactions using either the Hoigne, Staehelin and Bader [HSB] mechanism or the Gordon, Tomiyasu and Fukutomi [GTF] mechanism. These mechanisms and their associated reaction rates are described in detail in Refs. [8, 46]. In both mechanisms, the hydroxide ion (OH - ) is fundamental in initiating ozone decomposition in water. However, many other compounds are capable of initiating the process, so long as they are capable of forming a superoxide ion (O2) from an ozone molecule. Examples include hydroperoxide ions (HO2), cations, some acids, humic substances and UV radiation at 253.7 nm. However, in order to sustain the breakdown of 0 3 in water, the superoxide ions must be regenerated. This is the role of promoters, which are capable of regenerating the superoxide anion from hydroxyl radicals. Common organic promoters include aryl groups, formic acid, glyoxylic acid, Chapter 2. Relevant Background Information 11 primary alcohol and humic acids. Finally, there are inhibitors which are capable of consuming -OH without regenerating the superoxide anions. Some of the more com-mon inhibitors include bicarbonate and carbonate ions, alkyl groups, tertiary alcohols and humic substances [8]. Figure 2.3 summarizes the direct and indirect mechanisms by which O3 is decomposed and the roles of initiators, promoters and inhibitors in the whole cycle. The difficulty in describing how O3 influences the production of DBP precursors from NOM lies in determining which pathways dominate the reactions cited above, particularly when O3 is in the presence of a heterogeneous mixture of NOM. The pathway is largely dependent on the nature and quality of the water ma-trix. As shown in Fig. 2.3, humic substances (HS) (which can be a major constituent of NOM) can act as initiator, promoter and inhibitor. However, knowing the general quality of the source water can give some insight into which processes dominate. It has been generally observed that under acidic conditions (pH< 4) the direct molecular O3-NOM processes dominate the reactions in raw water, but at pH > 9 the indirect -OH-NOM processes are more prominent [9]. This is because at high pH the excess OH~ will serve to initiate O3 decomposition into -OH [46]. Similarly, if there is a large concentration of bicarbonate ions in the water source, it will end up scavenging the hydroxyl radicals and therefore provide conditions more favourable for the direct O3-NOM pathway [8]. However, in most cases, the combined effect of both the direct and indirect pathways must be considered. Relevant literature concerning the effectiveness of O3 in removing DBP precursors from raw water is summarized in Table 2.3. Due to variations in sample raw water and in the ozone dosage applied in each study, only relative reductions observed in each study have been reported here. Results from the literature indicate that ozone treatment does not significantly reduce the concentration of total organic carbon, but it does change the constituents as shown by the reduction of U V absorbance at 254 nm and a removal of T H M and HAA precursors. Chapter 2. Relevant Background Information 12 Initiation 0 H - , H 0 2 - , Fe 2 + , HCOO-, UV, HS, ... t-BuOH, HS, ... Figure 2.3: Mechanisms of ozone decomposition. The initiation, promotion, and inhibition of radical-type chain reaction (Adapted from Langlais et al. (1991) [8]). Ozone breaks down into superoxide ion (O2 ) which then picks up another oxygen from another ozone molecule to form O3". From that it reacts with a H + and form the hydroxyl radical (-OH). If there are inhibitors present in the water matrix, it will scavenged the -OH and end the chain reactions. On the other hand, if promoters are present in the water, it will drive the reactions to form the superoxide ion to regenerate more -OH. Certain compounds, as shown in the initiation box, are also capable of starting this chain reaction to produce hydroxyl radicals. It is of interest to note that humic substance (HS) can act as initiator, promoter and inhibitor Chapter 2. Relevant Background Information 13 Table 2.3: Summary of literature review for the ozone treatment. Reference Source Water (Ozone Dosage) Observations Galapate et al. (2001) [47] Minaga Reservior, Japan (0.85 - 3.0 mg 03/mg DOC) • 5 - 16% reduction in DOC • 47 — 72% decrease in UV260 absorbance • 6 - 43% reduction of T H M precursors Ko et al. (2000) [48] River Ruhr, Germany (3.5 to 12.5 mg 0 3 ) • no noticeable decrease in DOC • decrease in <7V_54 with increased O3 dose • 60 - 70% reduction in THMFP, 65% in T C A A F P but no change in D C A A F P Chang et al. (2000) [49] Te-Chi Reservoir, Taiwan (5.2 mg/L) • SUVA (ratio of t / V W D O C ) decreased with increased ozone dosage • overall 35% reduction in THMFP Westerhoff et al. (1999) [50] hydrophobic isolates of 3 water sources (2.6 mg 03/mg DOC) • 2 - 10% reduction in DOC • 60 - 70% reduction in UV2M absorbance Hu et al. (1999) [51] groundwater in petro-chemical industry, Singapore (2.1mg/L) • 23.7% reduction in T H M F P • 34.5% reduction in HAA5FP Friedman et al. (1997) [52] Capilano, Seymour and Coquitlam Reservoirs, Vancouver, B.C., Canada (3 mg/L) • 51 - 71% reduction in THMFP • 37 - 48% reduction in HAAFP Singer & Chang (1989) [53] 7 water treatment plants (WTPs) with preozonation in U.S.A. (0.13 - 1.1 mg 03/mg DOC) • preozonation alone had almost negligible effects on the overall TOC • reduction in C/V254 absorbance • average of 10 — 15% reduction in T H M F P Chapter 2. Relevant Background Information 14 2.3.3 0 3 -UV Advanced Oxidation Process (AOP) Advanced oxidation process (AOP) is the term used to describe the oxidation process carried out by the highly reactive hydroxyl radicals [9, 46]. The most common pro-cesses used to generate -OH is through the use of combined catalytic oxidants such as O3-UV , H2O2-UV and H 2 0 2 -03 . The generation of the hydroxyl radicals from these three combined catalytic processes will be briefly described below. The use of the heterogeneous photocatalytic process T i 0 2 — U V , ultrasound irradiation and high electron beam irradiation have also been investigated [6, 54]. Ozone alone, when used under conditions that favours the formation of hydroxyl radicals, can also be considered as an AOP, provided that the production of -OH is self-sustained. These latter options will not be discussed in this thesis. Hydrogen Peroxide-Ozone (H2O2-O3) When H2O2 is added to water in the presence of 0 3 , it dissociates into the hy-droperoxide ion (HO2—) before it reacts with molecular ozone to form the superoxide ion (O2) and hydroxyl radical (• OH) [8, 46]. H 2 0 2 — » H 0 2 + H + (2.3) H 0 2 + 0 3 —> O H " + 0 2 + 0 2 (2.4) The 0 2 and -OH then participate in the indirect ozone pathway, described in the section above. (See Fig.2.3). Hydrogen Peroxide-UV Irradiation(H 20 2-UV) Direct photolysis of hydrogen peroxide leads to formation of 2 hydroxyl radicals. H 2 0 2 + hu —> 2 • OH (2.5) It should be noted that this reaction is relatively slow compared to the H2O2-O3 pathway [9]. Ozone-UV Irradiation (O3-UV) Ozone added to water in the presence of UV results in the formation of hydrogen peroxide. The hydrogen peroxide can then react with either ozone or UV in the same manner as the reactions described above [9, 33]. In essence, O3-UV is a combination Chapter 2. Relevant Background Information 15 of H 2 0 2 - U V and H 2 0 2 - 0 3 . This is summarized in Fig. 2.4. Although all of the above three described processes ultimately result in the gen-eration of hydroxyl radicals, the energy required per mass of -OH produced differs between them. Table 2.4 compares the theoretical amounts of oxidants and UV re-quired for the formation of a mole of -OH. 2 - O H {Hp2- UV pathway) O H + 0 2 + 0 2 (H 2 0 2 -0 3 pathway) •OH Figure 2.4: O3-UV reaction pathways in water. Table 2.4: Theoretical amount of oxidant required to yield one mole of -OH. (Adapted from Gottschalk et al. (2000) [9].) (a) moles of photons required for each mole of -OH formed (b) assumes that superoxide is formed which yields one -OH per 0 2 , which may not be the case in some raw water. (c) hydrogen peroxide formed in situ AOP Process 0 3 U V ^ H 2 0 2 O3-OH-W 1.5 — — O3-UV 1.5 0.5 (0.5)^ H 2 0 2 - 0 3 6 ) 1.0 — 0.5 H 2 0 2 - U V — 0.5 0.5 Chapter 2. Relevant Background Information 16 Based upon the above information, H2O2-UV appears to be the most efficient process for it requires the least amount of oxidant per mole of -OH. But in fact, the photolysis of O3 yields more -OH than that of H2O2 because O3 absorbs UV more efficiently than H 2 0 2 at a wavelength of 254nm [9]. Table 2.5 compares the theoretical formation of -OH from the photolysis of O 3 and H2C>2 based on its molar absorbance or extinction (6254) coefficient. Table 2.5: Theoretical yield of -OH based on photolysis of O 3 or H 2 0 2 . (Adapted from Gottschalk et al. (2000) [9].) (a) assume 10cm pathlength; cone. (O3) = cone.(H2O2) = 10~ 4 M Process e254nm(M _ 1C 'm- 1) Stoichiometry •OH /photon^ H 2 0 2 0 3 20 3300 H 2 0 2 2-OH 0 3 -> 2-OH 0.09 2.00 Given this information, it appears that the O3-UV process provides the maximum yield of -OH. However, it does come at an additional cost; compared to the photolysis of H2O2, the 0 3 - U V process has the higher initial capital and operation costs due to the energy and resources required to produce ozone and power the lamp, and so in some cases H2O2-UV or H 2 02-0 3 are preferred [6]. Furthermore, its was reported by Wallace et al. (1988) [55] that the THMFP results obtained from using 0 3 - H 2 0 2 were so similar to those from using O3-UV that they questioned if the actual stoichiometric yield of -OH for O3-UV was two times that of O3-H2O2. Hence, very little research has been performed on O3-UV as an option for the removal of DBP precursors in drinking water. Table 2.6 presents a summary of a literature review pertaining to the O3-UV process. Note that since the set-up and source water used are different for each experiment, only the relative performance of the 0 3 - U V process is reported for each case presented. Chapter 2. Relevant Background Information 17 Table 2.6: Summary of literature review for O3-UV AOP. Reference Source Water (Ozone and UV Dosage) Observations Amirsardari et al. (2001) [56] North Pine Reservoir, Brisbane, Australia (14.7 mg 03/L followed by a 20W, 1.3L UV lamp) • 90% reduction in THMFP • 75% decrease in UV254 • 55% removal in. TOC Backlund (1994) [57] Lake Savojarvi, Finland (1400 mg 0 3 with 2 low pressure UV lamp) • 68% reduction in THMFP Kusakabe et al. (1990) [58] Commercial humic acid (0.24-0.46, mol 0 3 / m 3 & UV dosage 3 - 13.7 W/m2) • showed greater reduction in TOC and CHCl3FP with combined O3-UV than O3 alone • greater reduction in TOC and CHCI3FP with increase treatment time Wallace et al. (1988) [55] Lake water supply to Shreveport, L.A, USA (20 mg/L 0 3 & UV dosage 0.67 W/L) • 10% reduction in THMFP Sierka &; Amy (1985) [59] Commercial humic acid (0 - 43.2 mg 0 3 & UV lamp) • decrease in UV254 with increase time • 59% removal of TOC • 86% reduction of THMFP Glaze et al. (1982) [60, 61] 2 lakes - Texas/Louisiana and L.A., USA (0.14- 6.4 mg 03/L-min k a medium-pressure UV lamp of 0.1 -0AW/L) • greater reduction in TOC and T H M F P with increased treatment time • example of Cross Lake, LA water showed an ~ 50% re-duction in TOC and ~ 86% reduction in THMFP for 0 3 dose of 126 mg/L and UV dose of 0.2W/L Chapter 3. Objectives 18 Chapter 3 Objectives The objective of this project was to evaluate the effectiveness of the three treatment options (ozone, UV and combined O 3 - U V AOP) in the removal of total organic carbon (TOC) and two specific disinfection by-product precursors (THMFP and HAAFP) from water samples taken from the Seymour reservoir. Each of the above three treatment processes were applied to samples of raw water in a laboratory scale batch system for varying total treatment times and dosages of O 3 and/or UV to quantify the reaction rates. Finally, the process was repeated in a pilot scale batch system to validate the results obtained at the laboratory scale. The specific tasks completed during the course of this project were as follows: • Set-up and calibrate the laboratory scale batch system. • Perform a mass balance to account for the flow/consumption of ozone in the system. • Run the system for varying total treatment times and dosages. • Test the raw and treated samples for TOC, UV absorbance at 254nm and pH. • Dose the raw and treated samples with sodium hypochlorite (bleach) and incu-bate for a week to quantify the long-term DBP formation potential. • Test the chlorinated sample for THMFP and HAAFP. • Determine the DBP removal rates for each treatment. • Retrofit a pilot plant batch system to emulate the laboratory system. • Validate the results obtained from the laboratory setup with those obtained from the pilot scale apparatus. Chapter 4. Experiment Design 19 Chapter 4 Experiment Design 4.1 Seymour Source Water Characterization The source water used for this project is from one of the three watersheds owned by the Greater Vancouver Regional District (GVRD). The Seymour watershed is situ-ated in the mountains of North Vancouver. It is roughly 18 000 hectares in size and provides 40% of the total water supply to the region [1]. Typical of coastal mountain areas, the reservoir is recharged from precipitation and snow melt. Because the water-shed is protected, it is very pristine and contain negligible amounts of anthropogenic pollutants. Thus, any analytes found in the water are likely derived from the natural surroundings. This natural water is characterized by low organic carbon content, low pH and alkalinity. Table 4.1 summarizes some of the key water quality parameters. Table 4.1: Seymour source water quality parameters. (Adapted from the GVRD Annual Report, 2002 [1]) Water Quality Parameter Average Measurement Total Organic Carbon (mg/L) 2.0 Dissolved Organic Carbon (mg/L) 1.9 Bromate (mg/L) < 0.01 Total Cyanide (mg/L) < 0.005 Total Calcium (mg/L) 1.8 Phosphorus Total (mg/L) < 0.005 Sulphate (mg/L) 1.5 UV254 (Abs/cm) 0.74 Temperature (°C) 3 - 1 7 Turbidity (NTU) 1.3 True Colour (TCU) 13 pH 6.4 Conductivity (p,Q,~1cm~1) 13 Alkalinity as CaCOz (mg/L) 3.7 Hardness as CaC03 (mg/L) 5.1 Chapter 4. Experiment Design 20 4.2 Laboratory Scale A batch system was assembled on a laboratory bench in the Environmental Engineer-ing Laboratory at the University of British Columbia. Figure 4.1 shows a schematic of this system. FLOWMETER \ RECYCLING PUMP Figure 4.1: A schematic of the laboratory scale batch apparatus. Ozone was generated by coronal discharge through compressed air. The air was filtered through molecular sieves and dried by flowing through a desiccator prior to being fed into the O3 generator (Azcozon VMUS-4, Surrey, British Columbia). The O 3 was bubbled through a 6.2L plexiglass contactor that was 0.115 m in diameter and 0.4 m tall, that was filled with water to be treated. The contactor was set up to introduce a counter-current flow of O 3 and water. Ozone was bubbled from the base, and the water was recycled from the bottom to the top of the reactor with the aid of a pump (Masterflex Inc.) at a rate of 1 L/min. An annular UV reactor 0.045 m in diameter and 0.28 m tall was added into the water circuit as needed. A low-pressure UV lamp (Trojan Inc.) that emitted pri-marily at 253.7nm was enclosed in a 0.02 m diameter x 0.28 m height x2mm thick plexiglass sleeve inserted inside the annular UV reactor, giving a net volume capacity of 0.3 L. The lamp's intensity was 1.45 x 10~2W/cm2 and, within the sleeve, had an average intensity of 9.69 x 10 - 3 W/cm 2 measured with a research radiometer sensor Chapter 4. Experiment Design 21 (International Light Model SED240). Off gas from the reactor was collected in two KI traps (glass bottles with a diffuser filled with potassium iodide) connected in series. Al l tubing from the compressed air to the reactor were 1/4" OD teflon tubing (Nalgene) with stainless steel compression fittings (Swagelok). Quarter-inch OD soft tubing and nylon barbed fittings were used for the water recycle and off-gas lines. See Figure 4.2 for a picture of the laboratory scale apparatus. Raw water from the Seymour reservoir was equilibrated to room temperature prior to use. Since the water in the Seymour reservoir ranges in temperature over 3 — 13°C [1] and the solubility of O3 varies with temperature, it was decided to let the water acclimatize to room temperature (20 to 23°C) prior to beginning any experiments. This eliminates one of the external factor that can affect the net efficiency of the treatment processes. Raw water were added to the plexiglass contactor such that it was filled to the top to reduce the amount of headspace available to trap O3. The ozone/air flow rate into the reactor was maintained at 0.29 L/min as monitored with a calibrated flowmeter (Cole Palmer) and verified with a bubble meter (Hewlett Packard). This particular flowrate was chosen because it was within the range that gave the maximum yield of 0 3 from the ozone generator, given the constraint that the pressure of the air supplied to the ozone generator must never exceed 5 psig. See Appendix A for characterization of the ozone generator. Dye tests of the laboratory scale system confirmed that under the experimental conditions used, the contactor behaved like a completely mixed continuous stirred tank (CSTR) in that its contents are completely mix almost instantaneously. Ap-pendix B l details the dye test. For each type of treatment applied (ozone, UV and combined O 3 - U V AOP), a series of batch runs lasting 5,10,30, 60 minutes were performed. Longer runs (up to 10 hours) were also performed to determine what the ultimate oxidation capacity of the treatment. Following the completion of each batch run, samples were taken from the reactor and analyzed for TOC, pH, UV254 and disinfection by-product forma-tion potentials for T H M and HAA. This experimental approach enabled the removal kinetics for the DBP precursors to be investigated for each type of treatment applied. Chapter 4. Experiment Design 22 Figure 4.2: Picture of the laboratory scale batch apparatus. Chapter 4. Experiment Design 23 4.3 Pilot Scale The pilot scale set-up was assembled from existing equipment at the GVRD water treatment test facility at the Seymour dam. A separate UV reactor was installed in the plumbing in series with one of the available ozone contactor columns. The contactor had a circumference and height of approximately 0.805 m and 3.3 m, respectively. From these measurements, the volume was determined to be approxi-mately 175L. A centrifugal pump was used to recycle water from the contactor to a Trojan UV Max C (Trojan, London, Ontario) UV reactor and then back to the ozone contactor. The recycle flow rate was set at 7.5 gpm. An air compressor (Thomas Industries Inc, L.A. , USA) was used to generate compressed air on site to supply the ozone generator (PCI-Wedeco, New Jersey, USA). Ozone was bubbled up from the bottom of the contactor in a counter current flow configuration. Ozone monitors (PCI-Wedeco, New Jersey, USA) were used to record the O3 concentration in the in-gas entering and off-gas leaving the ozone column. The system hydraulics were parameterized using a rhodamine dye test and the measured profile is shown in Appendix B2. Due to time limitations, only one set of experiments was performed at the pilot scale. Samples measuring 1.5L in volume were collected at preselected time intervals of 5, 10, 30 and 60 minutes following the start of the test. Purging of the ozone from samples collected for analysis was achieved using an aquarium diffuser, where atmospheric air was bubbled into the sample immediately after removal from the reactor. Samples were then stored in a refrigerator before being transported back to the laboratory for analysis. The samples were analyzed for pH, temperature, TOC, UV absorbance at 254nm and, after chlorination and incubation for a week, T H M and HAA formation potential. Chapter 5. Methodology 24 Chapter 5 Methodology 5.1 Laboratory Preparation 5.1.1 Glassware Due to the low concentrations of analytes that were being encountered, all glass-ware used were subjected to a rigorous cleaning regime to minimize contamination. Al l glassware were washed with detergent and rinsed several times with tap water and then further with distilled water. Al l sample collection vials (40 mL) bottles (125mL,500mL), autosampler vials (2mL) and non-volumetric glassware were then baked in a muffle oven (Thermolyne 30400 Furnace) at 400°C for at least 4 hours. After cooling, they were stored in a clean container with aluminum foil covering all openings. Volumetric glassware were baked at 105°C in a V W R Scientific 1350 F M Forced Air Oven for at least 1 hour. Septa and lids were also washed with detergent, rinsed with tap water and then rinsed with distilled water before being baked at 105°C for at least an hour. 5.1.2 Reagents and Blanks All reagents and chemicals used were of laboratory quality, with the exception of sodium hypochlorite (Javex bleach). Blanks and water used to make reagent stan-dards was distilled using a Millipore Aqua-Q Ultra-pure water system. 5 . 2 Analytical Methods All samples were subjected to the following procedures which were procured from either Standard Methods [62] or the US Environmental Protection Agency [63]. Changes or alterations to the methods cited therein are elaborated upon in this thesis. Chapter 5. Methodology 25 5.2.1 Total Organic Carbon (TOC) Samples for measuring TOC were collected in 40 mL amber vials with aluminum foil used as a lid. TOC was determined by the Persulfate-Ultraviolet Oxidation Method 5310C [62] with the aid of the Dohrman Phoenix 8000 UV-Persulfate analyzer (Dohrman). Various types of filters and pore sizes were used to measure the dissolved organic carbon (DOC) in the water samples. However, it was realized early in the course of these experiments that the measured DOC was always equal or greater than the measured TOC (unfiltered samples). Thus, it was concluded that the majority, if not all, of the carbon in the water taken from the Seymour reservoir is dissolved. Hence, DOC was not measured in all samples and will not be reported herein. 5.2.2 pH and Temperature The pH and the temperature of the samples were measured with the Beckman $44 pH metre. Due to the low conductivity of the water, pH pHIX adjustor solution (Orion Research) was added to the samples. The instrument was calibrated prior to use with 2 standard buffer solutions (pH 4.00 and pH 7.00). 5.2.3 Ultraviolet Absorption at 254 nm (UV254) The ultraviolet absorbance of samples at 254 nm was measured by a UV 300 UV-Visible spectrometer (Spectronic Unicam) with a 1cm pathlength quartz culvert, in accordance with Standard Method 5910B [62] . This is a surrogate parameter used to determine the types of carbon compounds in the water. Compounds with aromatic or double bonds absorb more strongly at this wavelength. However, ozone absorbs strongly at this wavelength as well. Hence, U V 2 5 4 was also used as a means to determine if ozone has been purged from water samples prior to being subjected to further analysis. 5.2.4 Disinfection Byproduct Formation Potentials (DBPFPs) To quantify the DBPFPs, chlorine in the form of sodium hypochlorite (bleach) was added to each sample and incubated in the Innova 4230 Refrigerated Incubator Shaker (New Brunswick Scientific, Edison, N.J., USA) at 25°C for 8 ± 2 days. Standard Chapter 5. Methodology 26 Method 5710B Trihalomethane Formation Potential [62] was employed to determine the formation potential for both T H M and HAA. This method outlines the amount of chlorine dosing solution needed to maintain a chlorine residual of at least 3--5 mg/L. However, to ensure that the maximum possible formation potential was achieved, two times the suggested amount was used to dose the samples. As a result, this approach tends to overestimate the concentration of DBPs that are likely to be formed in a water treatment plant. To ensure that there was a sufficient chlorine residual maintained, the free and combined chlorine residuals were measured by the DPD (N,N-diethyl-p-phenylenediamine) colourimetric method with a Hach CN-66 Field Kit, after incubation. The typical range for total chlorine is 10 — 15mg/L. Samples that met this requirement were then tested for the T H M and HAA concentrations. 5.2.5 Trihalomethane ( T H M ) Because of the undetectable bromine concentration in the raw water, only the chlo-rinated compounds were singled out for measurement. For THMs, only chloroform (CHCI3) was isolated from the water samples using the liquid-liquid extraction gas chromatography (GC) method 6232B [62]. Pentane (Fischer Chemical, HPLC grade) was cleaned with basic alumina (Brockman Activity 1, 60-325 mesh) stored at 105°C, prior to use to eliminate background interference. Three millilitres of pentane were used to extract chloroform from each 30 mL sample in a 40mX amber vial. A small quantity of 1,2-dibromopropane was added to the pentane to provide an internal standard. Samples were analyzed using a Hewlett Packard 6890 Series Gas Chromatographer (GC) equipped with a N i 6 3 electron capture detector (ECD) and a Hewlett Packard 7673 autosampler. Prepurified helium gas carried the volatilized sample through a 28 m (length) x 0.53 mm (ID) x 3 pm (film thickness) capillary column (DB624, J & W Scientific 125-1334) housed within the GC oven subjected to the temperature program shown in Table 5.1. One microlitre aliquots were injected into the GC column using a microsyringe (Hamilton Model 701N, 10pL). In order to quantify statistical errors, every 10 t / l sample was duplicated and typically there was one duplicate for every sample set run through the GC. The response peaks from the GC were calibrated by running samples of chloroform (Fisher Chemical) solutions of known concentrations through the apparatus. The retention time of chloroform in this particular GC was deter-Chapter 5. Methodology 27 Table 5.1: GC temperature program for CHCI3 analysis. Parameters Settings Injector Type Splitless Temperature 90°C Detector Type ECD Temperature 260°C Oven Initial temperature 30°C, holding for 3 minutes Ramp 5°C/min Final temperature 90°C mined to be 7.2 minutes and that of the internal standard (1,2-dibromopropane) was 16.4 minutes. A post-run data processing program was written (GCAREAS.exe) to accelerate data analysis. 5.2.6 Haloacetic Acid (HAA) The method employed for quantifying HAA was a modified version of the Environ-mental Protection Agency (EPA) Method 552.2 [63] and the method described by Xie (2001) [64]. Below is a summary of the chemicals and procedures used by this author to extract a measure of three key chlorinated HAAs: monochloroacetic acid (MCAA), dichloroacetic acid (DCAA) and trichloroacetic acid (TCAA), which will henceforth be referred to as HA A3. Monobromoacetic (MB A A) was used as a surrogate to en-sure that the extraction procedure worked and 1,2-dibromopropane was used as an internal standard. A stock HAA3 solution was made from pure chemicals (Aldrich) using methyl-tert-butyl ether (MTBE) as the solvent. Aqueous calibration standards were made from the stock HAA3 mix. The internal standard was incorporated in the M T B E extraction solvent at a concentration of 360 u.g/L. The HAA3 sample extraction procedure was implemented as follows: 1. Remove amber sample collection vial from cold storage and allow it to equili-brate to room temperature. 2. Remove 15 mL of sample from the 40 mL amber collection vial for residual chlorine analysis. Chapter 5. Methodology 28 3. Add 20 \x/L of 48 mg/L surrogate to the remaining sample to yield a concen-tration of about 32 [ig/L. (Note that any concentration can be used so long as it provides a measurable response using the GC). 4. Adjust the pH to less than 0.5 by the addition of approximately 2mL of con-centrated H2SO4. 5. Quickly add 12 g of muffled sodium sulfate and shake until almost dissolved (approximately 1 minute). It is advisable to tilt the amber vial when adding the sodium sulfate, so that all of it does not collect on the bottom of the vial. 6. Add 3mL of the extraction solvent (MTBE with internal standard) and place on a mechanical shaker for 30 minutes. 7. Allow sample to phase separate for about 5 minutes. 8. Using a pasteur pipette, transfer as much of the upper M T B E layer (no water phase) to a 10 mL COD tube with lid. (Note: any glassware can be used provided that it is long, narrow and has a lid to it - this makes the transfer after methylation easier). 9. Add 1 mL of 10% H 2 S 0 4 in methanol (Fisher) to the COD tube. 10. Shake and cap the COD tube and place in heating bath at 50°C for 2 hours. 11. Remove tube and allow to cool. 12. Slowly add AmL of saturated sodium bicarbonate (just the supernatant) to the COD tube. 13. Shake the tube for 2 minutes, frequently venting the CO2 off gas. 14. Transfer the upper M T B E layer into an autosampler vial and analyze immedi-ately. Samples were analyzed using a Hewlett Packard 6890 Series Gas Chromatography System equipped with a Hewlett Packard 5973 Mass Selective Detector and a Hewlett Packard 6890 Series autosampler. Although equipment and temperature programs similar to those described by Xie (2001) were employed initially, the retention times and results obtained from the method presented herein did not agree with the results quoted in Ref. [64]. Rather, the relevant retention times observed by this author were considerably shorter and the peaks of interest were partially obscured by background Chapter 5. Methodology 29 signals. Consequently, a modified temperature program for the GC was designed and single-ion monitoring for the mass spectroscropy (MS) was employed to enhanced the sensitivity of the signals. A 30 m x 0.25mm(ID) x 0.25 pm (film thickness) capillary column (DB5MS, 5% phenylmethyl silicon, J & W Scientific) was used. The modified temperature program is summarized in Table 5.2. Table 5.2: GC temperature program for HAA3 analysis. Parameters Settings Injector Type Injection Volume Temperature Oven Initial temperature Rampl Ramp 2 Final temperature Splitless 2pL 200°C 35°C, holding for 10 minutes 5°C/min and hold for 5 minutes 10°C/min 70°C A full scan, which measures the total ion current, was first employed to identify the associated retention times of the analytes of interest. Given that, the mass spectrum of each analyte could be located in the instrument internal library, which shows the relative abundance of the fragmented compound as a function of ion (mass/charge). To increase the sensitivity of the detector, single ion monitoring was implemented by choosing a large ion with a big abundance that unique to that analyte (or very different from the other analytes). Table 5.3 lists the quantification ion and retention time for each compound of interest. Table 5.3: Quantification ions and retention times for HA A3 analysis. Compound Monitored Ion (m/z) Retention Time (minutes) MCAA-methyl ester 108 6.2 MBAA-methyl ester 152 10.1 DCAA-methyl ester 83 10.9 1,2-dibromopropane-methyl ester 121 11.5 TCAA-methyl ester 141 16.2 Chapter 5. Methodology 30 A post-run data processing program was written (GCMSAREAS.exe) to expedite data analysis. 5.2.7 Residual Aqueous Ozone Ozone concentration in the aqueous phase was measured using Indigo Colourimetric Method 45OO-O3B [62]. Because ozone degrades fairly quickly, care was taken to make this measurement as soon as an experiment was completed. For the laboratory scale experiments, an Indigo II solution was made and analyzed with a spectrophotometer (Spectronic Unicam) at a wavelength of 600 nm with a 2 cm pathlength culvert. For the pilot plant experiments, a Hach Field Kit with medium concentration ampules was used for convenience. 5.2.8 Residual Ozone in Gas Phase Standard Method 2350.E Ozone Demand/Requirement - Semibatch Method [62] was used to determine the amount of ozone collected in the gas phase. Sodium thiosulfate solution (0.1 M) was used as the titrant and was standardized using dichromate. Chapter 6. Results and Discussion 31 Chapter 6 Results and Discussion This chapter presents the experimental results obtained in the laboratory and the pilot scale. Each section presents the results from the work done for one treatment method at the laboratory scale. The final section presents the results from work done with the pilot scale set-up. The laboratory scale experiments were performed in 3 separate series of batches - series 1 (April 2003), series 2 (May 2003) and series 3 (June 2003). Note that no HAA3FP data were taken from series 1 due to problems with the HAA3 extraction procedures. Furthermore, the pH was not measured during all of the batch experiments in series 1. Al l data presented herein will be shown graphically. In cases where multiple measurements were taken, the value presented graphically corresponds to the average value of the measurements made, and the error bars represents one standard deviation of these measurements. Data analysis was based on the 70% confidence interval (i.e. one standard deviation) of the measurements made. 6.1 Ultraviolet Irradiation (UV) The UV dosage was calculated as follows: UV dosage (W • s/cm2) = Intensity (W/cm2) x Retention Time (s) (6.1) where, the retention time was calculated as: ^ . ™ . / \ T-, • m . / \ volume of UV reactor (L) ,„ „. Retention Time (s) = Experiment Time(s) x — — - (6.2) volume of the total system (L) The UV dosage used in this experiment ranged from 0 to 1.6 W • s/cm2. For comparison purposes, a UV dose of 4.0 x 10 - 2 W • • • /cm2 is typically used to achieved 2-log inactivation for most bacteria, viruses, protozoa and phages in drinking water treatment applications [39, 65]. The UV dosage is shown as the upper x-axis and experiment time as the lower x-axis in the following figures. Chapter 6. Results and Discussion 32 Figure 6.1 illustrates the TOC concentration and U V 2 5 4 absorbance of samples as a function of UV dosage and experiment time. As presented, the TOC concen-tration was not significantly affected by the UV dosage. The U V 2 5 4 absorbance, a surrogate parameter for the aromaticity of the compounds, decreased slightly with time. However, the overall reduction was limited (~ 11 ± 6%) even after 30 minutes of irradiation for all 3 batches measured. The disinfection by-product formation potentials as a function of UV dosage are shown in Fig. 6.2. There was no significant impact of treatment using UV irradiation on the concentration of chloroform precursors. On the other hand, the reduction of HAA3FP was significant. However, the reduction was relatively small (15.5 ± 3%), which suggests that UV treatment only has minor effects on the removal of HAA precursors. The speciation of HAA3 is shown in Fig. 6.3 for series 2 and 3. The ratio of the M C A A F P : D C A A F P : T C A A F P for the raw water was approximately 2 : 29 : 69. These ratios did not change substantially during UV treatment. The results indicate that over the range of the UV dosages investigated, UV treat-ment had negligible effects on the TOC concentration, UV254 absorbance, chloroform formation potential ( C H C I 3 F P ) and HAA3FP. These observations are consistent with those found in the literature [40, 41]. However, other studies that employed much higher UV dosages found significant reductions of TOC [42, 44, 45] and up to 50% reductions in UV254 [45]. However, there have been conflicting findings regarding THMFP, as stated in Section 2.3.1 of this thesis. The disagreement primarily ap-pears to be related to the wide range of UV dosages and variations in the quality of the raw source water used in these studies. More experiments at higher UV dosages could be conducted to definitively determine whether UV treatment can have an ef-fect on the DBPFP of the raw water from the Seymour reservoir. However, since the specific focus on this project was on the O 3 - U V AOP, high intensity UV studies were not pursued. Chapter 6. Results and Discussion 33 UV Dosage (W*s/cm2) 0.0 0.4 0.8 1.2 0.00 0.10 0 20 40 60 Experiment Time (minutes) (a) UV Dosage (W*s/cm2) 0.0 0.4 0.8 1.2 1.6 0.00 0 20 40 60 Experiment Time (minutes) (b) Figure 6.1: T O C and UV254 versus time (UV) for • series 1, V series 2 and series 3. (a) TOC. (b) UV254 absorbance. Chapter 6. Results and Discussion 34 UV Dosage (W*s/cm2) 0.4 0.8 1.2 1.6 20 40 Experiment Time (minutes) (a) 60 0.0 UV Dosage (W*s/cm2) 0.4 0.8 1.2 1.6 400 D) 3. •5 300 c o o a. c ~ 200 A E CO ^ 100 0 20 40 60 Experiment Time (minutes) (b) Figure 6.2: D B P F P versus time (UV) for • series 1, V series 2 and • series 3. (a) Chloroform formation potential, (b) HAA3 formation potential. The dotted line indicates the USEPA Stage II Microbial/DBP Rule(2003) reg-ulatory limit [3]. Chapter 6. Results and Discussion 35 Figure 6.3: Speciation of H A A versus time ( U V ) for (a) Series 2. (b) Series 3. Chapter 6. Results and Discussion 36 6.2 Ozone ( 0 3 ) Ozone consumption during O3 treatment was denned as follows: Mconsumed Mproduced MaqUeous Mgaseous ^sample V s a m p / e where (6.3) • the total mass of O 3 produced, Mpr.0duce<i, was determined by multiplying the pre-determined ozone production rate of the ozone generator (see Appendix A) by the experiment time; • Maqueous represents the mass of O 3 dissolved in the water sample, which was measured by Indigo Method [62] (see Chapter 5) immediately after the comple-tion of an experiment; • Mgaseous represents the mass of excess O 3 off-gas that was collected in the KI traps over the duration of an experiment. This quantity was determined by titration; • and the volume of the sample Vsampie was 6.2 L. A mass balance analysis based on the ozonation of distilled water (which should theoretically consume no O 3 ) indicated that 85 to 95% of the MprodUCed could be recov-ered from Maqueous and Mgaseous using the Indigo Method and KI traps, respectively. The concentration of O 3 in the distilled water reached saturation after approximately 10 minutes for ozone input of rate of approximately 12.44 mg/min. See Appendix C for the details of this mass balance recovery. Although a positive relationship was observed between ozone consumption and ozonation time, the relationship was not linear or constant, for the different series of batches as illustrated in Figure 6.4. Even though care was taken to perform all experiments as consistently as possible, there may have been variations in operating conditions and in the water matrix used in the different batches that could account for the differences in ozone consumption observed at long experiment times (> 30 minutes). Chapter 6. Results and Discussion 37 v • s _i i i i i i i 0 20 40 60 Experiment Time (minutes) Figure 6.4: Ozone consumpt ion versus t ime for • series 1, V series 2 and • series 3. Note that in this section, the experimental results for each of the parameters monitored have been plotted versus two different independent variables: time and dosage. Both types of plots are presented since they provide different information. The time plots provide information on reaction kinetics, while the dosage plots provide information on the extent of treatment, that is the dosage required. Presenting the information in terms of time and intensity on different graph was not necessary for the UV treatment results since the output of the UV lamp was constant with time. Figure 6.5(a) and (b) depict the TOC concentration as function of experiment time and ozone dosage, respectively. With the exception of the results from batch 2, there were no significant changes in TOC concentration with ozone treatment. The data from batch 2 show a 15% reduction in TOC within the first 5 minutes and a slower decrease with further ozonation time. Although the reduction is significant, it is relatively small (from 1.9mg/L to 1.6 mg/L). Note that the last point in batch 3 is an outlier as there is no rationale for a large increase in TOC concentration with treatment. This could have been due to an instrument error or perhaps due to leaching of carbon from another source. Ozonation had a significant and rapid effect on UV254 absorbance as illustrated in Fig. 6.6. Al l 3 series of batches showed that the UV254 absorbance decreased very a E ra co O E a E 3 (A C o o a> c o N O 0.03 0.02 0.01 0.00 Chapter 6. Results and Discussion 38 rapidly with ozone treatment (~ 73% reduction after 5 minutes) and a very slight additional decline with increased ozone dosage (~ 87% reduction after 60 minutes). These results suggests that although ozonation was not capable of removing (i.e. mineralizing) organic matter, it was capable of changing the chemical characteristics (i.e. aromatic nature) of this organic matter. Chapter 6. Results and Discussion 39 o.oo 1 — L 20 40 Experiment Time (minutes) (a) 60 2.00 1.50 ro E, O 1.00 0.50 h 0.00 0.00 0.01 0.02 0.03 Ozone Consumption (mg 03 /mL sample) (b) Figure 6.5: T O C versus time and ozone consumption (O3) for • series 1, V series 2 and • series 3. (a) experiment time and (b) ozone consumption Chapter 6. Results and Discussion 40 0.10 0.08 E « o c re . Q L _ o V) < in C M > 3 0.06 h 0.04 0.02 0.00 20 40 Experiment Time (minutes) (a) 0.10 0.00 0.01 0.02 0.03 Ozone Consumption (mg 03 /mL sample) (b) Figure 6.6: C/V254 versus time and ozone consumption (O3) for • series 1, V series 2 and • series 3. (a) experiment time and (b) ozone consumption Chapter 6. Results and Discussion 41 The likely change in the chemical characteristics of the organic material in the raw water that occurred during ozone treatment led to a significant reduction in the concentration of DBP precursors. As shown in Figs. 6.7 and 6.8, the chloroform and HAA3 formation potentials were significantly reduced as a function of ozone treat-ment. A 50% reduction was achieved for chloroform precursors, but the observed residual concentrations of approximately 150 fig/L is still above the USEPA regu-latory limit of 80 fig/L for THMs [3]. Similarly, there was a 50% reduction of the HA A3 formation potential for both series 2 and 3. However, the observed residual concentration of approximately 175 u,g/L is above the USEPA DBP Stage II regula-tory limit of 60(j,g/L [3]. Increasing the treatment time beyond 5 to 10 minutes did not yield significantly better results. This suggests that ozone rapidly oxidize T H M and HAA precursors into non-reactive forms that are resistant to further oxidation. The ratios T H M F P / T O C and HAA3FP/TOC were relatively constant after roughly 5 minutes at 8.4 ± 0.8 and 11.15 ± 2.6 for all batches. These observations imply that the majority of DBP precursors were oxidized into stable forms within the first few minutes of treatment. Consequently, prolonged ozone treatment did not yield further changes in the organic material. The residual products are suspected to be oxalic acids, which are long-lived intermediates that are relatively non-reactive with O3 [66]. This hypothesis is supported by the observed decrease in pH after ozone treatment, as shown in Fig. 6.9. It should be noted that these results are consistent with those from a similar study performed by the GVRD on the same water source as part of a program to upgrade its primary disinfection facilities to provide 3-log inactivation for Giardia cysts [52]. The concentrations of individual components of the HAA3 for each batch are shown in Fig. 6.10. The general trend is that M C A A and D C A A precursors were not affected by ozone treatment. Only the concentration of T C A A precursors seem to have been affected by ozonation. An attempt was made at parameterizing the kinetics of DBPFP reduction. The net reaction is presumed to be a cumulative reaction involving both molecular ozone (direct pathway) and hydroxyl radicals (indirect pathway). It is reasonable to as-sume that both of these pathways lead to second order reactions [8, 24], and their cumulative effect can be modelled as follows: = (fcO 3[0 3] + k.OH[-OE}) [^(t) - $residual] (6.4) where <&(£) represents the disinfectant byproduct formation potential (DBPFP) at Chapter 6. Results and Discussion 42 time t and $residual is the residual DBPFP as t —> oo. The reaction constants and concentrations for ozone and hydroxyl radicals are represented as (ko3,k.oii) and ([O3],[-OH]), respectively. Since the rate at which O 3 was being injected into the system was roughly constant and no attempts were made to distinguish the relative contributions from the direct and indirect pathways, Eq. 6.4 was simplified to include one effective reaction constant for both the O3 and -OH reactions = &o3[03] + fc.o/r[-OH]): - ^ = k*[$(t)-$reaidual] (6.5) Integrating Eq. 6.5 then yields $(t) = ^residual + = 0) - Residual) e-*** = ^residual + * J o _ _ e _ * * ' (6.6) where $ i o s s = $( i = 0) - $residuai for simplicity. The chloroform and HAA3 formation potential data from each batch were fit to Eq. 6.6 using the statistical software program Sigma-Plot (SPSS). However, it was found in all cases that the initial drop between datum at t = 0 and the next point at t — 5 minutes was so large that the software was not able to provide a satisfactory fit. There were simply not enough data taken within the first five minutes to capture meaningful information regarding k§. Therefore, it was not possible to accurately develop rate kinetic for treatment using O3. Since the present procedures used to analyze for T H M and H A A are labour inten-sive and expensive, many researchers have suggested the use of surrogate parameters (such as U V 2 5 4 , TOC and SUVA) to predict DBPFP [18, 26, 67]. However, cor-relation analysis between these surrogates ( U V 2 5 4 , TOC and SUVA) and DBPFP (THMFP and HAAFP) for the ozone treatment on Seymour's water did not result in any strong, positive correlations. In summary, ozone treatment did not significantly reduce TOC concentration but it did change its constituents into less reactive non-aromatic forms. In essence, it removed some of the DBP precursors. Both chloroform and HAA3 formation potentials were reduced by approximately 50% very quickly with ozone treatment, but there were no added benefits.from increasing ozonation time beyond approximately 5 minutes. It is suspected that ozone treatment oxidizes the NOM into oxalic acids, which are long-lived and ozone-resistant. This hypothesis is supported by a decline in pH during O3 treatment. However, the levels of T H M and HAA3 precursors reported herein exceeded USEPA regulations. Chapter 6. Results and Discussion 43 400 20 40 Experiment Time (minutes) (a) 400 0.00 0.01 0.02 0.03 Ozone Consumption (mg 03/ml sample) (b) Figure 6.7: T H M F P versus time and ozone consumption (O3) for • series 1, V series 2 and • series 3. (a) experiment time, (b) ozone consumption. • The dotted line indicates the USEPA Stage II Microbial/DBP Rule(2003) regulatory limit [3]. Chapter 6. Results and Discussion 44 400 20 40 Experiment Time (minutes) (a) 60 D) 1 300 c o Q. ~ 200 ro i V o co 2 100 0 I i i i i i i i i I 0.00 0.01 0.02 0.03 Ozone Consumption (mg 0 3 / m l sample) (b) Figure 6.8: H A A F P versus time and ozone consumption ( 0 3 ) for • series 1, V series 2 and • series 3. (a) experiment time, (b) ozone consumption. Dotted line denotes the USEPA Stage II Microbial/DBP Rule(2003) reg-ulatory limit [3]. Chapter 6. Results and Discussion 45 8.00 6.00 X Q. 4.00 2.00 8.00 6.00 4.00 0 20 40 60 Experiment Time (minutes) (a) 2.00 0.00 0.01 0.02 0.03 Ozone Consumption (mg 0 3 /ml_ sample) (b) Figure 6.9: p H versus time and ozone consumption (O3) for V series 2 and series 3. (a) experiment time, (b) ozone consumption. Chapter 6. Results and Discussion 46 400 ^ 300 3 . C o ro "H 200 o o c o u 2 X 100 M C A A D C A A T C A A 20 40 60 Experiment Time (minutes) (a) 20 40 Experiment Time (minutes) (b) Figure 6.10: Speciation of H A A versus time (O3) for (a) series 2 and (b) series 3. Chapter 6. Results and Discussion 47 6.3 Combined Ozone-Ultraviolet AOP For O3-UV AOP treatments, the ozone consumption and UV dosage were determined using the procedures described in the two previous sections. Figure 6.11 illustrates the relationship between experiment time, ozone consumption and U V dosage for the 3 series of batches. UV Dosage ( W * s / c m z ) 0.0 0.4 0.8 1.2 1.6 t V -0 20 40 60 Experiment Time (minutes) Figure 6.11: Ozone consumption and U V dosage versus time for • series 1, V series 2 and • series 3. 0 0. 0.08 E rs (A _ l E 0.06 O O) E ion 0.04 a E 3 (A C O 0.02 L> V c 0 N O 0.00 2 An ozone mass balance for the O3-UV system was performed using distilled water as the sample. The results, shown in Appendix C, indicate that the concentration of ozone in the aqueous phase was maintained at 2.5 mg/L, which is a factor of 2 lower than the concentration measured with ozone treatment alone. Furthermore, the percent recovery was relatively poor, having stabilized at approximately 60% after 30 minutes of treatment. The discrepancies between the experiments performed with ozone and those performed with O3-UV is likely due to the conversion of the dissolved ozone into hydroxyl radicals via photolysis. As illustrated in Fig. 6.12, treatment had a significant impact on the TOC concen-tration. Although the reduction was extensive (57 ± 18%), complete mineralization was not possible even after 10 hours of treatment. The residual after 10 hours of Chapter 6. Results and Discussion 48 treatment was 0.3 mg/L TOC. As the U V 2 5 4 absorbance data in Fig. 6.13 suggests, the chemical makeup of the compounds in the samples changed significantly with O3-UV treatment. There was a reduction of approximately 78% in U V 2 5 4 after treatment for 5 minutes, and the absorbance did not decrease with further treatment. The rapid change in U V 2 5 4 ab-sorbance and the relatively slower reduction in the TOC concentration suggests that the chemical characteristics (i.e. aromatic nature) of the organic material changed very rapidly, but that the mineralization of the organic material proceeded at a much lower rate. The impact of O3-UV treatment on the TOC, T H M F P and H A A F P is likely due to both the direct (molecular 0 3 ) and indirect ( O H radical) oxidation pathways, although the indirect pathway is suspected to dominate in the AOP. The net reaction involving O3-UV can be modelled in a manner similar to Eq. 6.4 (introduced in Section 6.2), where where $(t) represents the TOC, THMFP and H A A F P at time t and $residuai is the respective residual concentrations as t —> 00. Again, since there is no clear way of differentiating the effects of the direct and indirect reactions pathways, an effective rate constant (fc$ = fco3[03] + fc.o/f[-OH]) was used. It is assumed that the O3 entering the contactor is constant as a function of time and is in excess for the amount of NOM initially present in the sample. Therefore, the model can be simplified as shown in Eq. 6.5 which integrates to Eq. 6.6. The data of Fig. 6.12(a) have been fit to Eq. 6.6 using the statistical software program SigmaPlot (SPSS). The best fit parameters and their standard errors are listed in Table 6.1. Table 6.1: Pseudo first order reaction fit parameters for TOC ( O 3 - U V ) . Series # ^residual [P9/L) §ioss (pg/L) k$ (minute 1 ) R2 1 0.796 ±0 .4 0.989 ±0 .4 0.027 ±0.02 0.95 2 0.434 ±0 .3 1.378 ±0 .3 0.023 ±0 .01 0.91 3 0.324 ±0 .1 1.399 ± 0 . 1 0.037 ±0 .01 0.98 Although the rate constant (&$) for TOC measured for the 3 series differ slightly, they are not statistically different. Therefore, the rate constant from the 3 series were combined to produce a global k$ of 0.044 ± 0.02 minute - 1 for TOC decay during the O3-UV treatment of the raw water from the Seymour. The concentrations of disinfectant byproduct precursors were also reduced signifi-cantly with combined O3-UV treatment, as shown in Fig. 6.14 (chloroform formation potential) and Fig. 6.15 for (HAA3 formation potential). The effect upon each of Chapter 6. Results and Discussion 49 the individual HAA3 precursors is shown in Fig. 6.16. As with ozone treatment alone, the bulk of the reduction of HAA3FP can be attributed to the reduction of the concentration of T C A A precursors. However, unlike the ozone treatment alone, the combined O3-UV treatment does result in noticeable decreases in the concentrations of both M C A A and DCAA. The measured pH as a function of experiment time and ozone dosage is shown in Fig. 6.17. As with the ozone treatment alone, there is a noticeable decrease in pH upon treatment. Again, it is hypothesized that this is probably due to the generation of oxalic acids. This will be discussed in more detail later in this section. The concentrations of the above mentioned precursors show roughly exponential behaviour as a function of treatment time, with the most significant changes having occurred within the first 10 minutes for each batch. Longer run times of 800 minutes confirmed that there are residual DBPFPs. These formation potentials can be fit with the pseudo first order reaction model Eq. 6.6 with <f> representing either the chloroform formation potential or HAA3FP. The best fit parameters and their standard errors are shown in Tables 6.2 and 6.3 for chloroform formation potential and HAA3 formation potential, respectively. Chapter 6. Results and Discussion 50 UV Dosage (W*s/cmz) 0.0 0.4 0.8 1.2 2.00 r 1.50 o> E i.oo y o.5o y o.oo 20 40 Experiment Time (minutes) (a) 60 2.00 1.50 O) E, o 1.00 0.50 0.00 0.00 0.02 0.04 0.06 0.08 Ozone Consumption (mg 03/mL sample) (b) Figure 6.12: T O C versus time and ozone consumption ( O 3 - U V ) for • series 1, V series 2 and • series 3. (a) ozone consumption, (b) experiment time. Chapter 6. Results and Discussion 51 UV Dosage (W*s/cm2) 0.4 0.8 1.2 20 40 Experiment Time (minutes) (a) 1.6 0.10 „ 0.08 • E S 0.06 c re J3 L. o (A | 0.04 *t in CM > 3 0.02 0.00 0.00 0.02 0.04 0.06 0.08 Ozone Dosage (mg 03/mL sample) (b) Figure 6.13: UV254 absorbance versus time and ozone consumption ( O 3 - U V ) for • series 1, V series 2 and • series 3. (a) ozone consumption, (b) experiment time. Chapter 6. Results and Discussion 52 UV Dosage (W*s/cmz) 0.4 0.8 1.2 0 20 40 Experiment Time (minutes) (a) 400 O) « 300 c a> o CL C o 200 rs E k . o u_ E o o £ 100 h 0.00 0.02 0.04 0.06 0.08 Ozone Consumption (mg 03/ml sample) (b) Figure 6.14: Chloroform formation potential versus time and ozone dosage ( O 3 - U V ) for • series 1, V series 2 and • series 3. (a) ozone consumption, (b) experiment time. The dotted line indicates the USEPA Stage II Microbial/DBP Rule(2003) regulatory limit [3]. Chapter 6. Results and Discussion 53 UV Dosage (W*s/cm2) 0.0 0.4 0.8 1.2 20 40 Experiment Time (minutes) (a) 60 0.00 0.02 0.04 0.06 0.08 Ozone Consumption (mg O3/1T1I sample) (b) Figure 6.15: HAA3FP versus time and ozone dosage ( O 3 - U V ) for • series 1, V series 2 and • series 3. (a) ozone consumption, (b) experiment time. The dotted line indicates the USEPA Stage II Microbial/DBP Rule(2003) regulatory limit [3]. Chapter 6. Results and Discussion 54 Figure 6.16: Speciation of H A A versus time ( O 3 - U V ) for (a) series 2 and (b) series 3. Chapter 6. Results and Discussion 55 8.00 UV Dosage (W*s/cm2) 0.0 0.4 0.8 1.2 1.6 6.00 X a. 4.00 2.00 0 20 40 60 Experiment Time (minutes) (a) 0.00 0.02 0.04 0.06 0.08 Ozone Consumption (mg 0 3/ml_ sample) (b) Figure 6.17: p H versus time and ozone consumption ( O 3 - U V ) for • series 1, V series 2 and • series 3. (a) ozone consumption, (b) experiment time. Chapter 6. Results and Discussion 56 Table 6.2: Pseudo first order reaction fit parameters for chloroform formation poten-tial (0 3 -UV). Series # ^residual {h-9/'L) k<j> (minute 1 ) K1 1 37.53 ± 16 184.80 ± 25 0.128 ±0.04 0.96 2 51.53 ± 17 234.54 ± 30 0.111 ±0.05 0.95 3 42.94 ± 1 9 250.28 ± 38 0.101 ±0.03 0.96 Table 6.3: Pseudo first order reaction fit parameters for HAA3FP ( O 3 - U V ) . Series # ^residual {f-9j'L) $ioss (u-g/L) fc$ (minute 1 ) R* 2 3 60.79 ± 2 1 42.94 ± 19 313.67 ± 3 9 250.38 ± 38 0.125 ±0.04 0.107 ±0.04 0.96 0.92 The rate constants for the chloroform formation potential and HAA3FP obtained for the individual series were statistically similar. Therefore, global rate kinetics were developed from all the data obtained for chloroform and HA A3 formation potential. The global rate for chloroform and HAA3 formation potential are A;$ = 0.120 ± 0.028 minute - 1 and = 0.147 ± 0.041 minute - 1, respectively. The results from O3-UV treatment indicate that the USEPA regulatory levels are met after 30 minutes for chloroform and after 1 hour for HAA3. However, it should be recognized that these minimum dosage times are rather conservative since the conditions (incubation time, pH, chlorine concentration and temperature) used to prepare samples for DBPFP were far more favorable for reactions with chlorine than those found in a municipal water supply system. The actual conditions within a municipal water supply pipeline are expected to produce less DBPs, since the water in the distribution system is typically colder than the incubation temperature and the retention time within the distribution system is typically shorter than the incubation time of 7 days used during sample preparation. Correlation analyses were performed between the surrogate parameters (TOC, U V 2 5 4 and SUVA) and the DBPFP (THMFP and HAAFP) . Total organic carbon produced a positive correlation with THMFP and H A A F P (see Fig. 6.18) for the raw and treated Seymour's water. Both U V 2 5 4 and SUVA were not well correlated to either DBPFP. Therefore, TOC can potentially be used as a surrogate to predict THMFP and H A A F P in Seymour's water. Chapter 6. Results and Discussion 57 In summary, O3-UV treatment had a significant impact on the reduction of both T H M and HAA precursors. This reduction is attributed to the transformation of organic constituents into a least reactive form (as shown by a 78% decreased in UV254 absorbance after 5 minutes) and to the mineralization of TOC (at a rate of 0.044 ± 0.02 minute - 1). DBPFP decreased with increased treatment time. The USEPA regulatory limit for T H M and HAA can be met within 30 minutes and 60 minutes, respectively, of O3-UV treatment. Figure 6.18: Correlation between T O C and D B P F P ( O 3 - U V ) for • series 1, V series 2 and • series 3. (a) chloroform formation potential, (b) HAA3 formation potential. Chapter 6. Results and Discussion 59 6.4 Comparison of the O3 and O3 -UV Treatments By comparing the results for the ozone alone and the O3-UV AOP, it can be con-cluded that the latter treatment is more effective at mineralizing TOC water samples. (See Fig. 6.19 (a)). However, the treatments seem to be equally effective in reducing DBPFPs on short timescales, with approximately 50% reduction of chloroform for-mation potential and HAA3FP within the first 5 minutes. However, the additional benefits of the O 3 - U V AOP are seen with further treatment, with up to approxi-mately 83% reduction in chloroform formation potential and approximately 88% for HAA3FP after 60 minutes. Figures 6.19 (c) and (d) compare the reduction of these two DBPFPs for ozone and O3-UV treatment. The question then arises as to whether the reduction of the above mentioned for-mation potentials is, in fact, primarily due to the mineralization of TOC. It should be noted that the reaction time constant k$ for both the chloroform formation potential and HAA3FP are roughly a factor of 3 greater than that for kroc- Therefore, TOC is reduced at a slower rate than either of the measured DBPFPs. This suggests that the reduction of the DBPFPs in the presence of combined O 3 - U V is predominantly attributable to the conversion of organic compounds into forms that do not react with chlorine, as opposed to the removal of organic carbon from the water. The intermediates or other by-products formed during combined O3-UV treat-ment are of interest, although no attempt was made to investigate the formation of such substances. However, it should be noted that the UVz^i absorbance results are essentially identical to those obtained using ozone treatment alone, as shown in Fig. 6.19 (b). It is speculated that a number of acids, such as oxalic acids, are formed during both forms of treatment. This hypothesis is supported by a decline in pH from approximately 6.2 to approximately 4.2 after 0 3 - U V treatment, as shown in Fig. 6.17. However, this decline is not quite as pronounced as that found for the ozone treatment only, where a decline from approximately 6.2 to approximately 3.8 occurred . Further research is required, to confirm this hypothesis. In summary, it appears that O 3 - U V AOP treatment does provide some advantages over ozone treatment alone. The AOP is able to mineralize more organic carbon and reduce the chloroform formation potential and HAA3FP to a greater extent than treatment with ozone alone. This makes AOP a feasible option as a stand alone treatment to prevent the formation of DBP during the chlorination step of drinking water treatment for raw water from the Seymour reservoir. Chapter 6. Results and Discussion 60 Figure 6.19: Comparison of 0 3 and 0 3 - UV results for (a) TOC, (b) UV Ab-sorbance at 254nm, (c) Chloroform Formation Potential and (d) HAA3 Formation Potential. Data taken from all series of batches for o 0 3 and T 0 3 - U V treatments. The solid line represents the best fit estimate for the 0 3 - U V data and the dashed line for the ozone data. Chapter 6. Results and Discussion 61 6.5 Pilot Plant Results Due to time constraints, only one series of experiments was performed for each type of treatment at the pilot scale (August 2003). Samples were taken without disrupting the flow within the system through a valve connected near the recycling pump. The temperature of the raw water in the pilot system was 18 — 20°C. Unfortunately, the Hach Field Kit used to measure ozone concentration in the aqueous phase malfunctioned. Therefore the only indication that ozone was dissolved in the water was through the use of indigo powder. Hence, it was not possible to accurately estimate ozone consumption. For this reason, the results are presented as a function of time only (Figures 6.20- 6.24). In addition, the off-gas ozone monitor at the plant malfunctioned during the test. Therefore, there were no measurements for the ozone off-gas. Only the O 3 concentration in the in-gas (recorded by a second ozone monitor) could be measured. KI traps were used to calibrate the ozone monitors, but the results were too inconsistent to report. Figure 6.20 summarizes the TOC results from the 3 treatments at the pilot scale. As expected, UV did not affect the concentration of TOC and there was a slight de-crease in TOC with increased ozonation time (45% reduction after 1 hour). The great-est reduction occurred with O3-UV. Almost complete mineralization was achieved af-ter an hour of O3-UV treatment. The data were fitted using a pseudo first order rate of = 0.030 ± 0.008 minutes - 1 for TOC removal using O 3 - U V AOP, as determined from the least squares fit of Eq. 6.6 to the 0 3 - U V data. These results are similar to what was found in the lab scale. The composition of the NOM in the water was likely altered with ozone and O 3 - U V treatment, as suggested by the reduction in U V 2 5 4 absorbance in Fig. 6.21. There was no change in U V 2 5 4 absorbance with UV treatment only. These results agree qualitatively with the results obtained from laboratory scale experiments. Similarly, the chloroform formation potential results were compatible with the findings from the laboratory scale experiments. As shown in Fig. 6.22, UV treatment had no effect upon the chloroform precursors. There was approximately 40% reduc-tion in chloroform formation potential with ozone treatment, with slight improvement with increased ozonation time. However, the final concentration still exceeded the USEPA regulatory limit of 80 pg/L [3]. The most significant reductions occurred with the combined 0 3 - U V treatment, with the most significant changes occurring within the first 10 minutes, and 87% of the chloroform precursors removed after 30 minutes of treatment. The effective first order rate constant was k$ = 0.110 ± 0.011 minute - 1 Chapter 6. Results and Discussion 62 for chloroform removal using O3-UV AOP, as determined from a least squares fit of Eq. 6.6 to the O3-UV data. The USEPA regulatory limit was satisfied after roughly 10 minutes of treatment. 1.5 * • • • • 1.0 9 0 T O O 0.5 T 0.0 T 0 20 40 Experiment Time (minutes) 60 Figure 6.20: T O C concentration versus time (pilot scale) for • UV, o O3 and T O 3 - U V treatment. Some of the HAA3FP data had to be neglected because of the malfunction of the analytical equipment (i.e. GC-MS) during sample analysis. The results of HAA removal for O 3 - U V treatment are shown in Fig. 6.23. As expected, there is an initial rapid decline within the first 10 minutes of treatment. M C A A concentrations were too small to be detected, but both DCAA and T C A A reduced with increased O3-UV treatment time. The only analytical parameter measured during the pilot plant experiments that did not consistently follow the same trend as the results from the laboratory scale experiments was pH. Instead of decreasing with ozone treatment, the pH was observed to increase, as seen in Fig. 6.24. Furthermore, pH appeared to be increasing with O3-UV treatment time. Further studies would have to be performed to investigate if this phenomenon is reproducible and significant, and the mechanisms governing this phenomenon. Chapter 6. Results and Discussion 63 0.08 E 0) o c re ja k_ o w < IO > 3 0.06 0.04 0.02 20 40 Experiment Time (minutes) Figure 6.21: UV254 absorbance versus time (pilot scale) for • UV, o O3 and T O 3 - U V treatment. 150 250 ^ 2 0 0 re c Q) O CL C O E o LL E o 2 o .c o 100 20 40 Experiment Time (minutes) Figure 6.22: T H M F P versus time (pilot scale) for • UV, o 0 3 and T O 3 — UV treatment. Chapter 6. Results and Discussion 64 300 20 40 Experiment Time (minutes) Figure 6.23: H A A F P vs. time (pilot scale) for 0 3 - U V treatment 7.2 7.0 T T 6.8 O • 1 O 6.6 -6.4 • • • O -• 6.2 0 20 40 60 Experiment Time (minutes) Figure 6.24: p H versus time (pilot scale) for • UV, o 0 3 and T O 3 - U V treatment. Chapter 7. Conclusions 65 Chapter 7 Conclusions The following key conclusions can be drawn from the results presented herein: 1. UV treatment on its own had very little impact on the organic constituents in Seymour's water. The UV dosage used (less than 1.6 W • s/cm2) did not result in any significant reduction of TOC or in any changes to the UV254 absorbance. There was no significant reduction of chloroform formation potential and minor reduction on HAA3 formation potentials. 2. Ozone treatment on its own did not significantly reduce TOC in the water. However, ozone treatment likely changed the organic constituents into com-pounds with fewer electron-rich bonds, as suggested by a 75% reduction in UV254 absorbance. 3. Ozone treatment significantly reduced the chloroform and HAA3 formation po-tentials by approximately 50%. The reduction in HAA3FP is attributed pri-marily to the removal of T C A A precursors. The concentrations of M C A A and DCAA found in chlorinated samples did not change significantly with ozone treatment alone. The majority of the reductions in DBPFPs occurred within the first 5 minutes of treatment with very little reduction upon further ozone dosage. However, the levels remaining did not meet the EPA Stage II Micro-bial/DBP regulations for both THMs and HAAs. 4. The combined O 3 - U V AOP significantly mineralizes organic carbon, as demon-strated by a decrease in TOC. The kinetics underlying this reduction follow a first order decay with a residual TOC of approximately 0.3 mg/L remaining after 10 hours of treatment. This 78% reduction in TOC was mainly attributed to the oxidative capacity of the non-selective, fast-reacting OH radicals. 5. Ozone-UV treatment significantly reduced T H M and HAA precursors. The kinetics underlying both the THMFP and HAA3FP with 0 3 - U V AOP follow pseudo first order exponential decays, with a fast decline happening in the first 10 minutes. The residual chloroform and HAA3 precursors were measured to be Chapter 7. Conclusions 66 approximately 12pg/L and 30pg/L, respectively after 10 hours of treatment. Al l 3 components of the HAA3 were reduced with increasing ozone and UV dosages, with T C A A having the greatest reduction. USEPA regulation levels were met after 30 and 60 minutes of the combined O3-UV treatment for T H M and HAA, respectively. 6. The reduction in DBPFP with O 3 — UV process was not primarily attributed to the reduction of TOC, but rather to a change in organic constituents of the raw water into compounds that are less reactive with chlorine. 7. Pilot plant results validated the laboratory scale data for TOC, UV254 and T H M F P for all 3 treatment processes. Chapter 8. Recommendations 67 Chapter 8 Recommendations The following recommendations are suggested to expand on the scope of the present work. 8.1 Disinfection Potential Since both ozone and UV are traditionally used as disinfectants, it is important to determine if the combined AOP can provide disinfection as well as oxidation of disinfectant byproduct precursors. The monitoring of indicator organisms before and after treatment can offer insights to disinfection ability. 8.2 Ozone Disinfection Byproduct Measurement One of the concerns associated with using ozone is the creation of ozonation dis-infection byproducts such as aldehydes, ketoacids and carboxylic acids. Although, these classes of DBPs are not presently regulated by USEPA, the health risks asso-ciated with them may very well prove to be of concern. More research is required to understand the potential health risks posed by these classes of DBPs. Hence, mon-itoring the concentration of these ozonation DBPs would be of interest. However, as suggested by Richardson [16], only a small fraction of the possible DBPs have been identified thus far. Consequently, specific ozonation DBPs of concern must be identified and the analytical methods needed to detect them must be developed. 8.3 Feasibility of AOP without Filtration One of the benefits of using ozone is that it breaks large aromatic molecules into smaller aliphatic molecules of lower molecular weight. This is desirable if ozonation is followed by filtration using either granulated activated carbon (GAC) or a biofilter. However, if there is no filtration in the treatment train, then the smaller aliphatic Chapter 8. Recommendations 68 molecules could actually fuel microbial regrowth in a water distribution system. This is clearly not desirable. It was shown in these experiments that the combined O 3 - U V AOP, unlike ozone treatment alone, has the ability to oxidize TOC in addition to breaking down large compounds into smaller molecules. Given the low concentration of NOM found in the Seymour reservoir water, it may be possible that O3-UV AOP has the ability to oxidize enough TOC to avoid fuelling microbial regrowth in a water distribution system. Testing could be performed in a simulated distribution system (SDS) set-up to investigate this hypothesis. Furthermore, an SDS would also be useful in estimating chlorinated DBPFP in a real distribution system. 8.4 Reactor Design and Kinetic Modelling The design of a reactor plays a tremendous role in determining the ultimate efficiency of a treatment processes. A review of the literature reveals that there are many different experiment designs and a spread of 0 3 and UV dosages applied to source water of varying quality. There are no clear answers as to what system works best for a given source water, but there must be some general principles that can be used by engineers to optimize O3-UV AOP in any given apparatus. For example, there is the question as to whether a UV reactor submerged in the ozone bubble column is better than having separate O3 and UV reactors, as used in this study. In order to make more progress in modelling the reaction kinetics, more data is needed. The use of automated on-line data collection systems (ozone analyzers, pcba probes, etc . . .) would allow for continuous monitoring of key parameters. Modifying the reactor configuration and changing the 0 3 and UV dosages in a systematic way will yield further information. 8.5 Water Quality Factors As mentioned in Chapter 2, there are many factors that can affect the formation of DBPs. While there have been attempts to incorporate all of these factors into regression models [18, 33], such models typically have limited predictive power. More experimental data that looks at different factors such as pH, temperature, etc., needs to be generated in order to have a better understanding and to form a predictive model. It is recommended that experiments addressing each of these factors individ-ually and in combinations be performed. Chapter 8. Recommendations 69 8.6 Turbidity Seymour has encountered several turbidity events in the past during which the tur-bidity measures had exceeded Health Canada regulatory limits. Turbidity typically hinders disinfection and has a negative impact upon water aesthetics. It would be of interest to monitor the performance of O 3 - U V AOP under turbid conditions. It can be expected that with increasing turbidity less UV light will be available for photolysis of O 3 to produce -OH radicals. Thus it is expected that there will be no AOP at high turbidity. This was one of the original questions posed in the research proposal. Bibliography 70 Bibliography [1] GVRD. Water. The Greater Vancouver Regional District Quality Control Annual Report 2002. Volume 1. 2002. [2] United States Environmental Protection Agency. National primary drinking wa-ter regulations: Disinfectants and disinfection byproducts rule. Federal Register, 63(241):69389-69476, December 16 1998. [3] United States Environmental Protection Agency. National primary drinking water regulations: Stage 2 disinfectants and disinfection byproducts rule. Federal Register, 68(159):49548-49681, August 18 2003. [4] A. Huslin. In a glass of water, a mother's worst fear. Washington Post, page C01, Feburary, 10 2002. [5] Federal-Provincial-Territorial Committee of Drinking Water. Summary of guide-lines for Canadian drinking water quality. Technical report, Health Canda, April 2003. [6] H. Zhou and D.W. Smith. Advanced technologies in water and wastewater treat-ment. Canadian Journal of Civil Engineering, 28(Sl):49-66, 2001. [7] T. Oppenlander. Photochemical Purification of Water and Air. Wiley-Vch, 2003. [8] B. Langlais, D.A. Reckhow, and D.R. Brink, editors. Ozone in Water Treatment: Application and Engineering. AWWA Research Foundation, 1991. [9] C. Gottschalk, J.A. Libra, and A. Saupe. Ozonation of Water and Wastewater: A Practical Guide to Understanding Ozone and its Application. Wiley-Vch, 2000. [10] D.L. Macalady and J.F. Randville. Perspective in Environmental Chemistry. Oxford University Press, 1998. [11] S.E. Barrett, S.W. Krasner, and G.L. Amy. Natural Organic Matter and Dis-infection By-Products: Characterization and Control in Drinking Water, chap-ter Natural Organic Matter and Disinfection By-Products Characterization and Bibliography 71 Control in Drinking Water - An Overview, pages 1-12. 761. American Chemical Society, 2000. [12] D.M. Owen, G.L. Amy, Z.K. Chowdhury, R.J. Paode, G. McCoy, and K. Viscosil. NOM characterization and treatability. Journal AWWA, 1:48-63, 1995. [13] J-P. Croue, D. Violleau, and L. Labouyrie. Natural Organic Matter and Disin-fection By-Products: Characterization and Control in Drinking Water, chapter Disinfection By-Product Formation Potentials of Hydrophobic and Hydrophillic NOM Fractions: A Comparison Between a Low- and a High-Humic Water, pages 139-153. 761. American Chemical Society, 2000. [14] M.L. Pomes, C L . Larive, E.M. Thurman, W.R. Green, W.H. Orem, C.E. Rostad, T.B. Coplen, B.J. Cutak, and A . M . Dixon. Sources and haloacetic acid/THM formation potentials of aquatic humic substances in the wakarusa river, and clin-ton lake near lawrence, kansas. Environmental Science and Technology, 34:4278-4286, 2000. [15] E.H. Goslan, D.A. Fearing, J. Banks, D. Wilson, P. Hills, A.T. Campbell, and S.A. Parson. Seasonal variations in the disinfection by-product precursors profile of a reservior water. Journal of Water Supply: Research and Technology-AQUA, 51:475-482, 2002. [16] S. D. Richardson. The role of GC-MS and LC-MS in the discovery of drinking water disinfection by-products. Journal of Environmental Monitoring, 1:1-9, 2002. [17] International Programme on Chemical Safety. Disinfectants and Disinfectant By-products. Environment Health Criteria 216. World Health Organization, 2000. [18] G.L. Amy, P.A. Chadik, and Z.K. Chowdry. Developing models for predicting trihalomethane formation potentia and kinetics. Journal AWWA, 79:89-97, 1987. [19] J. Brereton. Impacts of Tuberculated Iron and Surface Biofilm on Trihalomethane Formation in Chlorinated Drinking Water. PhD thesis, University of British Columbia, 1998. [20] R .M. Clark. Chlorine demand and T T H M formation kinetics: A second-order model. Journal of Environmental Engineering, January: 16-24, 1998. Bibliography 72 [21] K.C.W. Chan, D.S. Mavinic, and J.A. Brereton. Trihalomethane formation in drinking water and production within a polyvinyl chloride pipe environment. Journal of Environmental Engineering and Science, 1:293-302, 2002. [22] L.A. Rossman, R.A. Brown, P.C. Singer, and J.R. Nuckols. DBP formation kinetics in a simulated distribution system. Water Research, 35:3483-3489, 2001. [23] A.A. Stevens, C.J. Slocum, D.R. Seeger, and G.G. Robeck. Chlorination of organics in drinking water. Journal AWWA, 68, 1976. [24] K.S. Kim, B.S. Oh, and J-W. Kang. Kinetic study of HAAs formation in chlo-rinated drinking water: Effect of pH, HOC1, and ozone dosage. Proceedings of IOA Conference, 2003. [25] R.R. Trussell and M.D. Umphres. The formation of trihalomethanes. Journal AWWA, 70:604-612, 1978. [26] M.C. Kavanaugh, A.R. Trussell, J. Cromer, and R.R. Trussell. An empirical kinetic model of T H M formation: Application to meet the proposed T H M stan-dards. Journal AWWA, 72:578-582, 1980. [27] D.A. Reckhow and P.C. Singer. Water Chlorination: Chemistry, Environmental Impace and Health Effects, volume 5, chapter Mechanisms of Organic Halide Formation During Fulvic Acid Chlorination and Implications with Respects to Preozonation. Lewis Publishers, 1985. [28] G.L Amy, L. Tan, and M.K Davis. The effects of ozonation and activated carbon adsorption on trihalomethane speciation. Water Research, 25:191, 1991. [29] J .M. Symons, S.W. Krasner, L.A. Simms, and M.J . Sclimenti. Measurements of T H M and precursors concentration revisited: The effect of bromide ion. Journal AWWA, 85:51, 1993. [30] D.A. Reckhow, P.C. Singer, and R.L. Malcolm. Chlorination of humic materials: Byproduct formation and chemical interpretations. Environmental Science and Technology, 24:1655-1664, 1990. [31] W.J. Miller and P.C. Uden. Characterization of nonvolatile aqueous chlorination products of humic substances. Environmental Science and Technology, 17:150-152, 1983. Bibliography 73 [32] H. Gallard and U. von Gunten. Chlorination of natural organic matter: Kinetics of chlorination and of T H M formation. Water Research, 36:65-74, 2002. [33] P.C. Singer. Formation and Control of Disinfection By-Products in Drinking Water. AWWA Research Foundation amd American Water Works Association, 1999. [34] M.S. Siddiqui, G.L. Amy, and B.D. Murphy. Ozone enhanced removal of natural organic matter from drinking water sources. Water Research, 12:3098-3106, 1997. [35] P.R. Berube, D.S. Mavinic, E.R. Hall, S.E. Kenway, and K. Roett. Evaluation of adsorption and coagulation as membrane pretreatment steps for the removal of organic material and disinfection-by-product precursors. Journal of Environ-mental Engineering and Science, 1:465-476, 2002. [36] Y . Amirsardari, Q. Yu, and P. Williams. Effect of ozonation and coagulation on turbidity and TOC removal by simulated direct filtration for potable water treatment. Environmental Technology, 18:1143-1150, 1997. [37] T. Chaiket, P.C. Singer, A. Miles, M . Moran, and C. Pallotta. Effectiveness of coagulation, ozonation, and biofiltration in controlling DBPs. Journal AWWA, 94:81-95, 2002. [38] J.P. Mally Jr. Formation and Control of Disinfection By-Products in Drinking Water, chapter Control of Disinfection By-Product Formation Using Ultraviolet Light, pages 223-235. American Water Works Association, 1999. [39] W.J. Masschelein. Ultraviolet Light in Water and Wasterwater Sanitation. Lewis Publishers, 2002. [40] J.P. Malley Jr., J.P. Shaw, and J.R. Ropp. Evaluation of By-Products Produced by Treatment of Groundwater With Ultraviolet Irradiation. AWWA Research Foundation and American Water Works Association, 1995. [41] O. Legrini, E. Oliveros, and A M Braun. Photochemical processes for water treatment. Chemical Reviews, 93:671-698, 1993. [42] J.W. Li , Z. Yu, M. Gao, L. Zhang, X . Cai, and F. L i . Effects of UV on charac-teristic and THMFP of humic acid. Water Research, 30:347-350, 1996. Bibliography 74 [43] N. Mole, M . Fielding, and D. Lunt. Disinfection By-Products in Drinking Water: Current Issues, chapter The Influence of UV Disinfection on the Formation of Disinfection By-Products, pages 54-61. Royal Society of Chemistry, 1999. [44] G. Kleiser and F.H. Frimmel. Removal of precursors for disinfection by-products (DBPs) - differences between ozone- and OH-radical-induced oxidation. The Science of the Total Environment, 256:1-9, 2000. [45] J. Thomson, F.A. Roddick, and M. Dirkas. UV photooxidation facilitating bio-logical treatment for the removal of NOM from drinking water. Journal of Water-Supply: Research and Technology - AQUA, 51:297-306, 2002. [46] J. Hoigne. Quality and Treatment of Drinking Water II, chapter Chemistry of Aqueous Ozone and the Transformation of Pollutants by Ozonation and Ad-vanced Oxidation Process, pages 83-142. Springer, 1998. [47] R.P. Galapate, A .U. Baes, and M . Okada. Transformation of dissolved organic matter during ozonation: Effects on trihalomethane formation potential. Water Research, 35:2201-2206, 2001. [48] Y-W. Ko, G. Abbt-Braun, and F.H. Frimmel. Effect of preozonation on the for-mation of chlorinated disinfection by-products for River Ruhr. Acta hydrochirn. hydrobio., 28:256-261, 2000. [49] C-N. Chang, Y-S. Ma, and F-F. Zing. Reducing the formation of DBPs by preozonation. Chemosphere, 46:21-30, 2002. [50] P. Westerhoff, J. Debroux, G. Aiken, and G. Amy. Ozone-induced changes in natural organic matter(NOM) structure. Ozone Science and Engineering, 21:551-570, 1999. [51] J.Y. Hu, Z.S. Wang, W.J. Ng, and S.L. Ong. Disinfection by-products in water produced by ozonation and chlorination. Environmental Monitoring and Assess-ment, 59:81-93, 1999. [52] M . Friedman, J. Smith, D. Neden, I. Singh, R.A. Hulsey, and P. Cheng. Sec-ondary water quality impact of high concentration of ozone addition to an unfil-tered water supply. Proceedings - AWWA Water Quality Technology Conference, Bellevue, WA, USA, P6P:l-4, 1997. Bibliography 75 [53] P.C.Singer and S. Chang. Impact of ozone on the removal of particles, TOC and THM Precursors. AWWA Research Foundation, 1989. [54] S.-K. Han, S.-N. Nam, and J.-W. Kang. OH radical monitoring technologies for AOP advanced oxidation process. Water Science & Technology, 46:7-12, 2002. [55] J.L. Wallace, B. Vahadi, J.B. Fenandes, and B .M. Boyden. The combination of O3/H2O2 and O3/UV radiation for reduction of T H M F P in surface water. Ozone Science and Engineering, 10:103-112, 1988. [56] Y . Amirsardari, Q. Yu, and P. Williams. Effect of ozonation and U V irradiation with direct filtration and disinfection and disinfection by-product precursors in drinking water treatment. Environmental Technology, 22:1015-1023, 2001. [57] P. Backlund. Destruction of natural mutagen and T H M precursors in water by ozonation, UV-irradiation and photolytic ozonation. Ozone Science and Engi-neering, 20:113-120, 1994. [58] K. Kusakabe, S. Aso, J. Hayashi, K. Isomura, and S. Morooka. Decomposition of humic acid and reduction of trihalomethane formation potential in water by ozone with UV irradiation. Water Research, 21:781-785, 1990. [59] R.A. Sierka and G.L. Amy. Catalytic effects of UV light and/or ultrasound on the ozone oxidation of humic acid and T H M precursors. Ozone Science and Engineering, 7:47-62, 1985. [60] W.H. Glaze, G.R. Peyton, S. Lin, R.Y. Huang, and J.L. Burieson. Destruc-tion of pollutants in water with ozone in combination with ultraviolet radiation. 2. natural trihalomethane precursors. Environmental Science and Technology, 16:454-458, 1982. [61] G.R. Peyton, F.Y. Huang, J.L. Burieson, and W.H. Glaze. Destruction of pol-lutants in water with ozone in combination with ultraviolet radiation. 1. gen-eral principles and oxidation of tetrachloroethylene. Environmental Science and Technology, 16:448-453, 1982. [62] APHA, AWWA, and WEF. Standard Methods for the Examination of Water and Wastewater, 18th ed. APHA, AWWA k WEF, 1992. [63] D.J. Munch, J.W. Munch, and A . M . Pawlecki. Method 552.2: Determination of haloacetic acid and dalapon in drinking water by liquid-liquid extraction, Bibliography 76 derivatization and gas chromatography with electron capture detection, revision 1.0. Technical report, Environmental Protection Agency, 1995. [64] Y . Xie. Analyzing haloacetic acids using gas chromatography/mass spectrometry. Water Research, 35:1599-1602, 2001. [65] United State Environmental Protection Agency. UV disinfection guidance man-ual proposal draft. Technical Report 815-D-03-007, USEPA, June 2003. [66] D.A. Reckhow. Formation and Control of Disinfection By-Products in Drink-ing Water, chapter Control of Disinfection By-Product Formation Using Ozone, pages 179-203. American Water Works Association, 1999. [67] G.L. Amy, P.A. Chadik, and R.A. Sierka. Ozonation of aquatic organic matter and humic substances: Analysis of surrogate paramters for predicting effects on THMFP. Environmental Technology Letters, 7:99-108, 1986. Appendix A. Ozone Generator Characterization 77 Appendix A Ozone Generator Characterization Ozone production by the Azcozon VMUS-4 ozone generator (Azco Industries, Surrey, British Columbia) as a function of air flow rates was measured in the laboratory. Each measurement consisted of setting a particular flow rate from the gas cannister and then flowing the output of the ozone generator through 2 potassium iodide (KI) traps for 5 minutes. A needle valve was used to control the air flow rates, a rotameter (Cole Palmer) was used as a bench mark scale and a soap film flowmeter (Hewlett Packard) was used to verify the flow rate with the aid of a stop watch. Figure A . l shows the setup for these measurements. eCPPRESSEP •'.mn. F L O W M E T E R i F L O W M E T E R 4> A A ,Kl.fRAP'S:r Figure A . l : Schematic of the in-gas measurement system The ozone production rate (mg/min) was determined from titration of the KI and was converted into a concentration by dividing the ozone production rate by the Appendix A. Ozone Generator Characterization 78 air flow rate. The results are presented in Fig. A.2. 50 0.2 0.4 0.6 0.8 Flow Rate (L/min) Figure A.2: Ozone generator performance for • ozone production rate (mg/L) and o ozone concentration as a function of air flow rate. The general trend is that with increasing flow rate there is an increase in ozone production, but a decrease in ozone concentration. However, there is a minimum air flow rate that must be met in order to generate any O3. An air flow rate of 0.29 L/min was regarded as being optimal for performing the laboratory scale batch experiments. The corresponding ozone production rate was 12.44±0.18 mg/min and the ozone concentration was 43.5 ± 0.64 mg/L. Further rationale for choosing such a low flow rate was to improve ozone transfer efficiency by slowing the rate at which air was forced through the bubble column. Appendix B. Hydraulic Characterization of the Systems 79 Appendix B Hydraulic Characterization of the Systems Rhodamine dye was used to determine the hydraulic characteristics of the systems. A slug of rhodamine was injected at time zero and samples were collected in COD vials at equal time intervals from collection ports. The absorbance of the sample was read at 520 nm as a means to determine how much dye has reached the collection port at that time. B.l Laboratory System The batch system was set-up with both the ozone and UV reactors in the plumbing with the air and recycle flow rates set as during all experiments. A slug of rhodamine was injected immediately before the pump head and samples were collected at the base of the ozone reactor. Samples were taken continuously after injection, with 18 samples taken in one minute. Figure B . l illustrates the mixing profile. The dye was essentially mixed in 40 seconds, which was much less than the retention time of the system of 6.2 minutes. Hence it can be concluded that the hydraulics of this system is that of a completely mixed reactor. B.2 Pilot Plant System Similarly, a slug of rhodamine was used to test the pilot plant setup. The dye was injected immediately before the pump head and samples were collected in 30 second intervals at the base of the ozone column. The hydraulic profile is shown in Figure B.2. Complete mixing occurred after 200 seconds. The retention time for the system was approximately 375 seconds. Appendix B. Hydraulic Characterization of the Systems 80 40 80 Time (seconds) 120 Figure B . l : Hydraulic characteristics of lab batch system using rhodamine dye testing. 100 200 300 Time (seconds) 400 Figure B.2: Hydraulic characteristics of pilot plant batch system using rho-damine dye testing. Appendix C. Mass Balance with Ozone and Combined Ozone-UV 81 Appendix C Mass Balance with Ozone and Combined Ozone-UV To account for all the ozone in the system, a mass balance using distilled water was performed for both the ozone alone and combined O3-UV treatments. Distilled water was used because it contains essentially no material which could consume O 3 . The reactor was filled to the top with distilled water and subjected to ozonation for pre-specified times of 2, 5, 10 and 30 minutes. The ozone concentration left in the aqueous phase was measured using Indigo Method [62] and the concentration reported for both treatments are shown graphically in Fig. C.l(a). It appears that for the ozone treatment, ozone saturation was achieved after 10 minutes at a concentration of 5.2 ± 0.3 mg/L. The combined O 3 - U V process achieved a lower constant ozone concentration of 2.53 ± 0.2mg/L for all ozonation times. Since distilled water contains essentially no analytes, all O 3 injected into the system is expected to be present in the aqeous phase or in the off-gas. The percent O 3 recovery was defined by the following equation: Figure C.l(b) shows the percent recovery for both ozone and combined O 3 - U V treatments at each prespecified time. The recovery for the ozone treatment was approximately 85% or greater after 5 minutes. However, the recovery for the combined O3-UV process actually decreased with treatment time, which was presumably due to the photolysis of O 3 to -OH. The resulting concentration of -OH was not measured. recovery (%) = O3 in aqueous(mg) + 0 3 in off-gas(m<7) ( C l ) O3 produced (mg) Appendix C. Mass Balance with Ozone and Combined Ozone-UV 82 UV Dosage (W*s/cm2) 0.0 0.2 0.4 0.6 0.8 6 10 20 Ozonation Time (minutes) (a) UV Dosage (W*s/cm2) 0.2 0.4 0.6 30 10 20 Ozonation Time (minutes) (b) Figure C l : Mass balance for ozone in the laboratory scale batch system using distilled distilled water for • ozone alone and o combined O3-UV treatment, (a) Ozone concentration in the aqueous phase and (b) percent ozone recovery as a function of ozonation time. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0063516/manifest

Comment

Related Items