Impact of Ozone on Disinfection By-Products: Comparison of
Research Report 9 Three Surface Waters with Differing Character
Research Report 9
Impact of Ozone on Disinfection By-Products: Comparison of three Surface Waters with Differing Character
Chris Kostakis and Brenton C. Nicholson
Cooperative Research Centre for Water Quality and Treatment
Research Report no. 9
November 2001
CRC for Water Quality and Treatment Research Report 9-2001
© CRC for Water Quality and Treatment, 2001
ISBN 1 876616 10 5
DISCLAIMER
• The Cooperative Research Centre for Water Quality and Treatment and individual contributors are not responsible for the outcomes of any actions taken on the basis of information in this research report, nor for any errors and omissions.
• The Cooperative Research Centre for Water Quality and Treatment and individual contributors disclaim all and any liability to any person in respect of anything, and the consequences of anything, done or omitted to be done by a person in reliance upon the whole or any part of this research report.
• The research report does not purport to be a comprehensive statement and analysis of its subject matter, and if further expert advice is required the services of a competent professional should be sought.
The Cooperative Research Centre for Water Quality and Treatment can be contacted at:
CRC for Water Quality and Treatment Private Mail Bag 3 Salisbury South Australia, 5108 AUSTRALIA
Phone: 61 8 8259 0337 Fax: 61 8 8259 0228
ii CRC for Water Quality and Treatment Research Report 9-2001
FOREWORD
This report is the result of work carried out in CRC Project 3.2.3 “Alternative Disinfection Regimes” and was also submitted by Chris Kostakis to fulfil the requirements for the degree of Master of Applied Science in Chemistry at the University of South Australia. The degree was conferred in November 2001.
Research Officer: Chris Kostakis
Project Leader: Dr Brenton C. Nicholson
Research Node: Australian Water Quality Centre
CRC for Water Quality and Treatment Project No. 3.2.3 – Alternative Disinfection Regimes
iii CRC for Water Quality and Treatment Research Report 9-2001
Executive Summary
The provision of microbiologically safe drinking water is still commonly achieved by chlorination in most water treatment plants. The discovery of chloroform and many other halogenated organic compounds in drinking water as a result of chlorination has sparked public concern and the introduction of stringent regulations on maximum contaminant levels of these disinfection by-products in treated water. Water utilities overseas, and possibly Australian utilities in the future, are being compelled to adopt alternative water treatment regimes in order to reduce the level of these chlorinated disinfection by-products. The use of ozone in water treatment for the management of various water quality issues including disinfection by-product control has increased in popularity and is becoming a widespread practice. Ozone is shown to react with natural organic matter present in water to form lower molecular weight oxygenated compounds such as aldehydes, ketoacids and carboxylic acids. Analytical procedures for the determination of these classes of ozonation by-products were optimised and validated in this work. Bench scale ozonation experiments were performed on waters collected from the Myponga, Hope Valley and Tod Reservoirs. Doses of 0.5, 1, 2, 3 and 5mg of ozone per mg of dissolved organic carbon content were applied and aldehydes, ketoacids and carboxylic acids were found to generally increase with increasing ozone dose. The formation of these compounds was shown to depend on the organic and inorganic character of the water. Bacterial regrowth potentials were determined and the biodegradability of the organic matter was shown to increase in ozonated waters. Raw waters and their ozonated samples were subsequently chlorinated to determine the effects of pre-ozonation on the formation potentials of trihalomethanes, chloroacetic acids and other chlorinated disinfection by-products. The effects of ozone dose on chlorinated by-product formation following chlorination differed in each water, illustrating the dependence on source water characteristics. The character of the organic matter and high bromide levels are shown to significantly contribute to the outcome of halogenated by-product formation. Finally, water samples were collected at each stage of the water treatment process from the Edenhope water treatment plant in Western Victoria, Australia. The plant utilises advanced treatment technologies such as ozonation and biologically activated carbon. Aldehydes, ketoacids, chlorinated disinfection by-products and other water quality parameters were monitored throughout the treatment process and the observed distribution of the various products is discussed.
iv CRC for Water Quality and Treatment Research Report 9-2001
Acknowledgments
This work was funded by the CRC for Water Quality and Treatment and supported by SA Water as part of CRC Project 3.2.3 “Alternative Disinfection Regimes” The contribution of the other supervisors in this project, namely Associate Professor Dennis Mulcahy and Dr Gunter Klass of the University of South Australia is gratefully acknowledged. The support and cooperation of co-workers and postgraduate students at the Australian Water Quality Centre is also acknowledged as without this, work of such quality would not have been possible. Chris Kostakis would also like to thank Mr Noel Sims, Mr Peter Felgate, Dr Hilton Kobus and Dr Paul Kirkbride of State Forensic Science for their encouragement, motivation and for some necessary time off given to complete this report.
v CRC for Water Quality and Treatment Research Report 9-2001
Table of Contents Page
FOREWORD…………………………………………………………………………………………..………..iii EXECUTIVE SUMMARY…………………………………………………………………………………..….iv ACKNOWLEDGMENTS……………………………………………………………………………………….v TABLE OF CONTENTS……………………………………………………………………………………….vi LIST OF TABLES…………………………………………………………………………………………...... viii LIST OF FIGURES…………………………………………………………………………………………….ix LIST OF ABBREVIATIONS………………………………………………………………………………...... xi
1 INTRODUCTION………………………………………………………………………………….…..1 1.1 Overview………………………………………………………………………………………….……1 1.2 Natural Organic Matter in Water………………………………………………………………….....1 1.3 NOM: Precursors to Disinfection By-products………………………………………………….….3 1.3.1 Chlorine and its Use in Water Treatment………………………………………………....3 1.3.2 Chlorinated Waters in South Australia……………………………………………….……6 1.4 Alternative Disinfectants…………………………………………………………………………..….7 1.4.1 Effectiveness of Disinfection……………………………………………………………..…7 1.4.2 Chloramines……………………………………………………………………………….…7 1.4.3 Chlorine Dioxide……………………………………………………………………….…….8 1.4.4 Ozone……………………………………………………………………………………..…..8 1.5 Reactions of Ozone…………………………………………………………………………….….….8 1.5.1 Reactions of Ozone With Organic Compounds…………………………………………..8 1.5.2 Reactions of Ozone With NOM……………………………………………………………..9 1.5.3 Aldehyde and Ketoacid Analyses…………………………………………………………11 1.5.4 Biodegradable Organic Matter…………………………………………………………….11 1.6 Research Objectives…………………………………………………………………………………12
2 THE WATER SOURCES AND THEIR CHEMICAL CHARACTERISTICS………………………………………………………………………………...13 2.1 Hope Valley Raw Water……………………………………………………………………………..13 2.2 Myponga Raw Water………………………………………………………………………………… 14 2.3 Tod Raw Water……………………………………………………………………………………….15 2.4 Reagent Water……………………………………………………...... 16
3 OZONATION: APPLICATION AND BY-PRODUCT ANALYSIS…………...... 17 3.1 Ozonation Experiments…………………………...... ……………...... ……17 3.2 Raw Waters and Their Ozone Demand…………………...... …………18 3.3 Ozonation By-Products and Evaluation of The Methodology……...... 21 3.3.1 Determination of Aldehydes in Water……………………...... …..21 3.3.1.1 Analytical Procedure………………………………...... 22 3.3.1.2 Sample Dechlorination…………………………...... …..22 3.3.1.3 Background Contamination………………………...... 23 3.3.1.4 Recoveries and Limits of Quantitation……………...... 24 3.3.2 Determination of Ketoacids in Water………………………...... 25 3.3.2.1 Buffering the pH of the Reaction…………………...... 26 3.3.2.2 Decarboxylation of Oxalacetic Acid……………...... ….26 3.3.2.3 Analytical Procedure……………………………...... …27 3.3.2.4 Recoveries and Limits of Quantitation…………...... …...27 3.3.3 Determination of Carboxylic Acids in Water………………...... 28 3.3.3.1 Identification of Carboxylic Acids………………...... 30 3.3.3.2 Developmental Work…………………………...…...... 30 3.3.3.3 Analytical Procedure……………………………...... …31 3.4 Determination of Bacterial Regrowth Potentials…………………...... 32
4 CHLORINATION: APPLICATION AND BY-PRODUCT ANALYSIS……………...... ………..34 4.1 Chlorination Experiments…………………………………...... ……….34 4.2 Chlorination By-Products and Methodology…………………...... ……34
vi CRC for Water Quality and Treatment Research Report 9-2001
4.2.1 Determination of Trihalomethanes in Water………………...... …34 4.2.2 Determination of Chloroacetic Acids in Water……………...... …35 4.2.3 Determination of Haloacetonitriles, Chloroketones, Chloropicrin and Chloral Hydrate………………………...... …..35 4.2.4 Determination of Adsorbable Organic Halogens in Water…...... 37
5 EFFECTS OF OZONATION ON NATURAL WATERS………...... …38 5.1 Effects of Ozonation on UV254 Absorbance and DOC Levels…...... ….38 5.2 Effects of Ozone on Aldehyde Formation………………………...... …39 5.3 Effects of Chlorine on Aldehyde Formation……………………...... …42 5.4 Effects of Ozone on Ketoacid Formation…………………………...... 44 5.5 Effects of Chlorine on Ketoacid Formation………………………...... 46 5.6 Effects of Ozone on Carboxylic Acid Formation…………………...... 47 5.7 Impact of Ozone on Biodegradable Organic Matter………………...... 48
6 EFFECTS OF OZONATION ON CHLORINATED DBPs……...... …55 6.1 Effects of Ozone on Trihalomethane Formation…………………...... 55 6.2 Effects of Ozone on Chloroacetic Acid Formation Potentials……...... 58 6.3 Effects of Ozone on Other Chlorinated By-products……………...... …61 6.3.1 Chloroketones……………………………………………...... …..61 6.3.2 Chloral Hydrate and Chloropicrin…………………………...... …62 6.3.3 Haloacetonitriles……………………………………………...... 63 6.4 Relationship Between Total Organic Halogens and Pre-ozonation...... 63 6.5 Chlorination of Individual Ozonation By-products………………...... 67
7 OZONE AND ITS FULL SCALE USE AT EDENHOPE……...... …..68 7.1 Treatment Train Process…………………………………………...... …68 7.2 UV254 Reduction and DOC Levels Through the Plant………….…...... 69 7.3 Aldehyde Levels Through the Plant…………………………….…...... 70 7.4 Ketoacid Levels Through the Plant………………………………...... 72 7.5 Biodegradable Organic Matter Through the Treatment Process………………………………...72 7.6 Chlorinated DBP Formation in Finished Water…………………...... …75
8 SUMMARY AND CONCLUSIONS………………………...... ………77 9 RECOMMENDATIONS…………………………………………...... 80 10 REFERENCES……………………………………………………………...... 81 APPENDICES I – Additional Methods Development – Aldehydes……...... …….87 II – Additional Methods Development – Ketoacids………...... ….89 III – Mass Spectra and Identification of Carboxylic Acids……………………….………...... …..91 IV – Raw Data – Ozonation By-product Formation…………...... 93 V – Raw Data – Chlorination By-product Formation………...... 95
vii CRC for Water Quality and Treatment Research Report 9-2001
List of Tables
Table Page
1.1 Stage 1 of the D-DBP Rule – Maximum Contaminant Levels and Maximum Residual Disinfection Levels…...………………………...... 6 1.2 C.t Values for 99% Inactivation of Microorganisms by Disinfectants at 5oC…………………………………………………...... …..7 1.3 Standard Oxidation Potentials of Common Oxidants Used in Water Treatment……….….....8 2.1 Raw Water Quality Data……………………………………………...... ….…16 3.1 Transferred Ozone Doses and Ozone Residuals………………………...... …19 3.2 Evaluation of Reagent Waters…………………………………………...... …24 3.3 Aldehyde Recoveries From Spiked Reagent and Raw Water…………...... …24 3.4 Limits of Quantitation for Each Aldehyde………………………………...... 25 3.5 Ketoacid Recoveries From Spiked Reagent and Raw Water…………...... ….28 3.6 Limits of Quantitation for Each Ketoacid………………………………...... 28 3.7 Major Mass Spectral Ions of Each Derivatised Acid……………………...... 30 3.8 Optimum Conditions for the Analysis of Carboxylic Acids…………...... …..31 4.1 Volumes of Quenching Agents Used for Specific Analyses…………...... ….34 4.2 Limits of Quantitation for Various Chlorinated Compounds…………...... …36 5.1 Effect of Ozone Dosage on SUVA Values for Each Surface Water…...... ….39 5.2 Bacterial Regrowth Potential Data for Myponga Water………………...... 49 5.3 Correlation Coefficients for Linear Regression of Ozonation By-products versus BRP…………………………………………...... …….49 6.1 Chlorination of Individual Ozonation By-products………………………...... 67 7.1 Raw water Quality Data – Lake Wallace Water………………………...... …68 7.2 Treatment Plant Design Criteria – Edenhope Treatment Plant…………...... 70 7.3 BRP Through Edenhope Treatment Plant………………………………...... 74 7.4 Chlorinated DBP in Treated Water……………………………..………...... 75
viii CRC for Water Quality and Treatment Research Report 9-2001
List of Figures Figure Page 1.1 Fuchs model for Humic Acid…………………………………………...... …2 1.2 Proposed Model for Fulvic Acid……………………………………...... ……2 1.3 Proposed Tetramer Structure for Humic Acid…………………………...... …3 1.4 Common Disinfection By-products in Drinking Water………………...... ….5 1.5 Reaction of Ozone With Olefins…………………………………………...... 9 1.6 Aqueous Decomposition of Ozone……..……………………………...... …..10 2.1 Area Served by the Hope Valley Reservoir……………………………...... …14 2.2 Area Served by the Myponga Reservoir…………………………………...... 15 2.3 Tod Reservoir Located on Eyre Peninsula………………………………...... 16 3.1 Experimental Set-up for Semi-batch Ozonation…………………………...... 18 3.2 Ozone Demand in Myponga Raw Water……………………………...... ……20 3.3 Ozone Demand in Hope Valley Raw Water……………………………...... 20 3.4 Ozone Demand in Tod Raw Water………………………………………...... 21 3.5 PFBHA Derivatisation of Carbonyl Compounds………………………...... 21 3.6 Potential Organic Ozonation By-products: The Aldehydes…………...... …22 3.7 Analysis of Aldehydes in Water…………………………………………...... 23 3.8 Potential Organic Ozonation By-products: The Ketoacids………………...... 25 3.9 Double Derivatisation of Ketoacids……………………………………...... …25 3.10 Reaction Mechanism Towards Oxime Formation………………………...... 26 3.11 Decarboxylation of Oxalacetic Acid Under Aqueous Conditions……...... …..27 3.12 Analysis of Ketoacids in Water…………………………………………...... 29 3.13 Derivatisation of Carboxylic Acids………………………………………...... 30 3.14 Analysis of Carboxylic Acids in Water…………………………………...... 32 3.15 A Typical Bacterial Growth Curve………………………………….…...... …33 4.1 Common Chlorinated Disinfection By-products………………………...... …36 5.1 UV254 and DOC Reductions With Increasing Ozone Dosage……………...... 39 5.2 Formaldehyde Levels With Increasing Ozone Dose…………………...... …..40 5.3 Acetaldehyde Levels With Increasing Ozone Dose……….……………...... 41 5.4 Glyoxal Levels With Increasing Ozone Dose……………………..…...... …...41 5.5 Methylglyoxal Levels With Increasing Ozone Dose……...……………...... 42 5.6 Effects of Chlorination Following Ozone Treatment on Aldehyde Formation……………..…43 5.7 Ketoacid Formation in Different Waters With Increasing Ozone Dose…...... 45 5.8 Comparison of Total Aldehyde and Ketoacid Formation………………...... 46 5.9 Effects of Chlorination and Its Destruction of Ketoacids………………...... 47 5.10 Influence of Ozone Dosage on Carboxylic Acid Formation……………...... 48 5.11 Relationship Between Total Aldehydes, Ketoacids and BRP in Myponga Water………..…50 5.12 Comparison of Formaldehyde Levels Before and After BRP Assay……...... 51 5.13 Comparison of Acetaldehyde Levels Before and After BRP Assay……...... 51 5.14 Comparison of Glyoxal Levels Before and After BRP Assay…………...... 52 5.15 Comparison of Methylglyoxal Levels Before and After BRP Assay…...... ….52 5.16 Comparison of Pyruvic Acid Levels Before and After BRP Assay……...... 53 5.17 Comparison of Ketomalonic Acid Levels Before and After BRP Assay...... 53 5.18 Comparison of Glyoxylic Acid Levels Before and After BRP Assay…...... 54 6.1 Influence of Pre-ozonation on Chloroform and Bromodichloromethane Formation Potentials……………………………...... 56 6.2 Influence of Pre-ozonation on Chlorodibromomethane and Bromoform Formation Potentials………………………………………...... 56 6.3 Distribution of THM Species In Each Raw Water Control……………...... …57 6.4 Influence of Pre-ozonation on Trichloroacetic Acid Formation Potentials……………………59 6.5 Influence of Pre-ozonation on Dichloroacetic Acid Formation Potentials……………………59 6.6 Influence of Pre-ozonation on Monochloroacetic Acid Formation Potentials………..………59 6.7 Total Chloroacetic Acid Levels With Increasing Ozone Dosage………...... 60 6.8 Effects of Pre-ozonation on Chloroketone Formation…………………...... ….61 6.9 Effects of Pre-ozonation on Chloral Hydrate and Chloropicrin Formation…………..……….62 6.10 Effects of Ozone on Haloacetonitrile Formation Potentials……………...... 64 6.11 Effect of Ozone on Total Organic Halide Formation in Myponga Water…………...... ….…64 6.12 Effect of Ozone on Total Organic Halide Formation in Hope Valley Water……………….…65
ix CRC for Water Quality and Treatment Research Report 9-2001
6.13 Relationship Between TOX and Total THMs – Myponga Water………...... 66 6.14 Relationship Between TOX and Total THMs – Hope Valley Water…...... …66 7.1 Schematic Representation of Edenhope Water Treatment Plant………...... 69 7.2 UV254 and DOC Levels Through Stages of the Treatment Process……...... …71 7.3 Monitoring the Aldehyde Levels Through the Treatment Process……...... …71 7.4 Monitoring the Ketoacid Levels Through the Treatment Process………...... 73 7.5 Bacterial Growth Curves Through Treatment Process…………………...... 74 7.6 Contribution of Individual Haloacetonitriles Between Tod Water and Lake Wallace Water………………………………...... ……76
x CRC for Water Quality and Treatment Research Report 9-2001
List of Abbreviations
AOC - assimilable organic carbon AOX - adsorbable organic halides AWQC - Australian Water Quality Centre BRP - bacterial regrowth potential CAA - chloroacetic acids CH - chloral hydrate CP - chloropicrin DBAN - dibromoacetonitrile DBP - disinfection by-product DCAA - dichloroacetic acid DCAN - dichloroacetonitrile 11DCA - 1,1-dichloroacetone 13DCA - 1,3-dichloroacetone DCC - dicyclohexylcarbodiimide DOC - dissolved organic carbon EPA - Environmental Protection Agency GC-ECD - gas chromatography with electron capture detection HAA - haloacetic acid KHP - potassium hydrogen phthalate LOQ - limit of quantitation MCAA - monochloroacetic acid MCL - maximum contaminant level MTBE - methyl tert-butyl ether NaOH - sodium hydroxide NOM - natural organic matter PFBHA - pentafluorobenzylhydroxylamine PFP - pentafluorophenol R - correlation coefficient RSD - relative standard deviation TBAHS -tetrabutylammonium hydrogen sulfate 111TCA -1,1,1-trichloroacetone TCAA - trichloroacetic acid TCAN - trichloroacetonitrile THM - trihalomethanes UV - ultraviolet
xi CRC for Water Quality and Treatment Research Report 9-2001
1 Introduction
1.1 Overview This review has been designed to give the reader a brief but broad background of some current issues facing water treatment utilities. It begins by discussing the origin, complexity and water quality problems of natural organic matter in water and then focuses on the important problem of chlorinated disinfection by-products. Alternative disinfection regimes are discussed paying attention to the use of pre-ozonation as a means of disinfection by-product control. The use of ozone as an oxidant and pre- disinfectant has its own potential concerns which are discussed. The research objectives of this project arising from these issues are then presented.
1.2 Natural Organic Matter in Water Natural organic matter (NOM) is ubiquitous and found in groundwater, river water, lakes, swamps and seawater. It is comprised of variable proportions of the soluble and insoluble remains of terrestrial and aquatic plant, animal and microbial material. Natural processes such as microbial degradation, chemical polymerisation and oxidation that take place essentially produce a complex heterogeneous mixture of substances varying in size and structure and containing a range of functional groups [1]. Humic substances, yellow to brown in colour, are the major components of NOM and can account for 50 to 80% of the total NOM present [2]. Humic substances are comprised of humic and fulvic acids that are operationally defined on the basis of their solubility; fulvic acid is completely soluble in aqueous solution whereas humic acid is soluble in alkali but precipitates in acidic conditions. This operational division between fulvic and humic acids is not precise and for each there is a solubility overlap arising from differences in molecular size and distribution of functional groups. Humic substances still remain unidentified. Over the decades many authors have proposed model structures for humic substances [3-5]. Some of these hypothetical models are presented in Figures 1.1-1.3. Figure 1.1 depicts an early model by Fuchs featuring a highly condensed aromatic system heavily substituted by carboxyl and hydroxyl groups [3]. In its time the model was quite influential due to its highly conjugated system which could account for the colour observed in humic materials. Schnitzer and Khan in 1972 proposed a structure composed of small aromatic compounds substituted with carboxyl and hydroxyl groups and held together by intermolecular hydrogen bonding [4]. Steelink applied elemental analysis to determine the atomic ratios of carbon, hydrogen, nitrogen and oxygen in humic substances. This was used as a key guide in proposing hypothetical structures for humic and fulvic acids [5]. Figure 1.3 displays a proposed structure for humic acid. The numerous hypothetical models that were proposed over the years were based on technology and information available at that time. These structures are now recognised as quite naive. As NOM is composed of gross mixtures that are innately diverse and heterogeneous, many scientists believe that structural models have little meaning. The extensive interest with NOM in water has been primarily due to the role NOM plays in a range of aesthetic, environmental and health issues. NOM affects the aesthetic quality of water by imparting colour and causing taste and odour problems. NOM also exhibits high complexation and solubilisation capabilities and so can bind trace metals and organic pollutants such as pesticides and polyaromatic hydrocarbons [6,7]. This ultimately effects the pollutant’s fate and transport in the environment. Hiraide and coworkers have reported that 50-70% of dissolved iron and copper in river water are present as complexes with humic substances [8]. Furthermore NOM are precursors for haloforms in waters treated with chlorine [9,10]. Because of the above-mentioned problems associated with NOM and the problems it poses towards the production of quality drinking water, investigations into determining the character and origin of NOM have developed into a major research field.
1 CRC for Water Quality and Treatment Research Report 9-2001
O
HO COOH COOH HO OH HOH HO
OH COOH
HO COOH COOH CH3O O
Figure 1.1: Fuchs Model for Humic Acid [3].
O OH OH O OOH OH C O C OH C C HO O OH C C O OH OH O HO C C OH O O C C OH O OH C OH OH OH O
HO O O C OH OH O OOHOHO C OH CC HO C C O OH HO OH C C O O C C O C O HO OH HO O OH C OH HO O OH OH
Figure 1.2: Proposed Model for Fulvic Acid (Adapted from Schnitzer and Khan, 1972)[4].
2 CRC for Water Quality and Treatment Research Report 9-2001
HOH2C COOH H CCH HOCH2 OH O H C OH H COH H
CH2OH O HO OC OH H C O
COOH COOH C36H36O18 n
Figure 1.3: Proposed Tetramer Structure for Humic Acid (Adapted from Steelink, 1985)[5].
Characterisation of and research on NOM has generally been conducted using samples that have been isolated and subsequently concentrated from the bulk water. Numerous isolation and concentration techniques have emerged over time including reverse osmosis [11], ultrafiltration [12] and adsorption onto macroporous resins [2,13-15], each having their own strengths and weaknesses. Problems arising from these extraction techniques include: (i) Recovery losses: What fractions are missed in the isolation procedure? (ii) Extreme pH alterations in some techniques: What changes occur to the NOM structure? (iii) The inorganic matrix may be removed or concentrated: What changes occur to the NOM structure? (iv) Significant contamination from isolating media affecting results of subsequent tests. At present there is no universally standardised method for isolating NOM and consequently a direct comparison of the results between samples is difficult when the isolation procedures are different. While isolation techniques may greatly facilitate attempts to identify and characterise NOM one must critically question how well these reconstituted samples compare to bulk water. In the natural environment, humic substances interact with both organic and inorganic species. In isolating these materials many interactions are disrupted and changes in the chemical structure and natural reactivity of the compounds may occur. Furthermore, fulvic acids used in many studies are not fully representative of the organic matrix of ‘real waters’. Aminosugars, proteic materials and polysaccharides are also predominant constituents but are lost in the isolation of fulvic acids [16].
1.3 NOM: Precursors to Disinfection By-Products
1.3.1 Chlorine and its use in water treatment The practice of using chlorine as a water disinfectant has been successful since the first decade of the 20th century and still remains the major disinfectant in use around the world. Its use has been important in providing safe drinking water and since its inception has caused a large drop in mortality rates by preventing the spread of pathogenic water-borne diseases such as typhoid and cholera. Disinfection with chlorine is responsible for reducing the risk of death from several per hundred thousand (for typhoid and cholera in the late 1890s to early 1900s) to virtually no occurrences of these diseases today [17]. Chlorine has remained the disinfectant of choice as it fulfils many of the requirements of an ideal disinfectant. It has excellent bactericidal properties with short to moderate contact times [18] and a stable disinfectant residual can be maintained throughout the distribution system preventing potential bacterial regrowth. Other benefits include its strong oxidising power that enables effective colour removal. Chlorine is also relatively cheap and easy to apply. It can be applied
3 CRC for Water Quality and Treatment Research Report 9-2001 as either gaseous chlorine or as a hypochlorite salt. However, it fails the requirements of an ideal disinfectant as chlorine reacts with NOM in water to form undesirable chlorinated by-products. Rook [9] and Bellar & coworkers [19] independently reported that aqueous chlorine reacts with NOM in water to form chloroform and other trihalomethanes (THMs). These findings coincided in time with results linking carcinogenic properties with drinking water contaminants [10]. Toxicology and carcinogenesis studies have shown THMs induce cancers in animals [20-22]. In 1979, these concerns prompted the United States Environmental Protection Agency (USEPA) to issue regulations in the United States limiting the concentration of total THMs to 100 µg/L [23]. In monitoring, total THMs are comprised of chloroform (CHCl3), bromodichloromethane (CHBrCl2), dibromochloromethane (CHBr2Cl) and bromoform (CHBr3). The brominated THMs arise from the presence of bromide in water. Inorganic bromide is oxidised to hypobromous acid by hypochlorous acid (in water chlorine reacts to form hypochlorous acid) which reacts with NOM in competition with chlorine [32,109,110]. These revelations sparked increasing attention towards the investigation and understanding of the reactions between chlorine and NOM. The identification of a range of other chlorinated disinfection by- products (DBPs) exhibiting similar adverse health effects soon followed causing increasing concerns in the quality of drinking water. Non-volatile halogenated compounds such as di- and trichloroacetic acid were first reported in the 1980s and are another class of chlorinated DBPs that occur frequently in chlorinated waters [24]. A number of studies have shown haloacetic acids (HAAs) form at concentrations comparable to THMs [25,26]. The HAAs are of health concern as they have been found to cause liver tumours in mice [27]. Other chlorinated DBPs that have been found to form include the semi-volatile haloketones [10] and haloacetonitriles [28]. Recent bioassay tests reveal that some analogues of both classes demonstrate mutagenic properties [29]. A list of the abovementioned DBPs and their respective chemical structures along with others that will be monitored in this study are presented in Figure 1.4.
4 CRC for Water Quality and Treatment Research Report 9-2001
Figure 1.4: Common Disinfection By-products in Drinking Water.
In 1992, regulatory negotiations took place in the United States between water authorities and health organisations This was prompted by the formation and potential health effects of these chlorinated DBPs. The issue was to reduce human exposure to disinfectants and DBPs while simultaneously minimising exposure to pathogenic microorganisms. A two stage Disinfectant-Disinfection By-Product (D-DBP) Rule was proposed by the USEPA and took effect in December 1998. The Stage 1 D-DBP rule [30] has decreased the maximum allowable contaminant level (MCL) of total THMs by 20% to 80 micrograms/litre (µg/L). In addition to reducing total THM levels in finished drinking water, regulations have been set to monitor five haloacetic acids suspected of causing health problems. The five major HAAs include monochloroacetic acid (MCAA), dichloroacetic acid (DCAA), trichloroacetic acid (TCAA), monobromoacetic acid (MBAA) and dibromoacetic acid (DBAA). An MCL of 60µg/L for the total concentration of these five HAAs has been set. As some disinfectants are also suspected to be toxic at higher concentrations, the Stage 1 D-DBP Rule sets limits for disinfectant residual levels present in drinking water. The Stage 1 D-DBP Rule requirements are listed in Table 1.1.
5 CRC for Water Quality and Treatment Research Report 9-2001
Table 1.1: Stage 1 of the D-DBP Rule – Maximum Contaminant Levels and Maximum Residual Disinfectant Levels. Maximum contaminant levels for DBPs 80 µg/L THMsa 60 µg/L HAAsb 10 µg/L bromate 1.0 mg/L chlorite Maximum residual disinfectant levels 4 mg/L chlorine 4 mg/L chloramines 0.8 mg/L chlorine dioxide a. sum of the four THMs: chloroform, bromodichloromethane, dibromochloromethane, bromoform b. sum of five HAAs: monochloroacetic acid, dichloroacetic acid, trichloroacetic acid, monobromoacetic acid, dibromoacetic acid.
Stage 2 regulations were initially proposed to further reduce the MCLs for total THMs and total HAAs to 40µg/L and 30µg/L respectively, and MCLs for other chlorinated DBPs were to be set [30]. However, in recent negotiations a Stage 2 standard was developed that will change (if promulgated) the Stage 1 THMs and HAAs MCLs from system-wide averages to locational running averages. The MCLs would remain the same but this would ensure that no single distribution location would exceed the proposed levels [104]. In addition to the USEPA, other health organisations world-wide including the World Health Organisation (WHO), European Economic Community (EEC), Environment Canada, NZ Department of Health, and the National Health and Medical Research Council, Australia (NH&MRC) have given consideration to limits for DBPs. Limits vary considerably depending on the basis for their derivation and as to their purpose. Countries such as the United States have set standards and utilities must comply with these standards as they are legally binding. In Europe, member countries must meet MCLs set by the EEC but there are no penalties for non-compliance. Regulations in Canada are similar to Australia whereby non-enforceable guideline levels have been set.
1.3.2 Chlorinated Waters in South Australia Australian surface waters, especially South Australian waters, are generally high in NOM. Moreover, the levels of THMs in South Australian finished drinking waters are much higher in relation to the standards that apply overseas [31-33]. THMs have been monitored in South Australian water supplies since 1976 and levels up to 1122 µg/L have been detected [31]. The high THM levels in South Australian waters are due to a number of factors. Levels of NOM in many surface waters tend to be high due to the nature of vegetation, the soil character, rainfall patterns and high temperatures, particularly in summer. In addition, long detention times through distribution pipelines and the high chlorine doses required to maintain a disinfectant residual are all contributing factors favouring the formation of high levels of chlorinated DBPs [31]. Furthermore, during periods of increased salinity within the River Murray (primary water source for the city of Adelaide during dry conditions) elevated bromide levels lead to high levels of brominated DBPs [31]. The Australian Guideline Level for total THMs is set at 250 µg/L [34]. THM levels in Australian waters have been reported to frequently be greater than 100 µg/L with some South Australian finished waters approaching if not surpassing the guideline value [33,35]. With the advent of stringent regulations overseas to limit the public exposure to these chlorinated DBPs, utilities have made the move to, or given consideration to, more advanced alternative water treatment practices.
6 CRC for Water Quality and Treatment Research Report 9-2001
1.4 Alternative Disinfectants Disinfection is imperative in providing microbiologically safe drinking water. Water chlorination has been used in the United States since 1908 and was first employed in Adelaide during 1953. As indicated previously, it remains the most common disinfectant in water treatment as it is highly effective and relatively cheap. However the concerns that chlorine reacts with NOM to form a wide range of chlorinated DBPs that may potentially cause adverse health effects and the promulgation of regulations on the levels of these compounds in finished drinking water has prompted water utilities to re-evaluate their water treatment processes. Water treatment modifications for controlling DBP formation can involve the use of alternative disinfectants. This section briefly details the types of chemical agents that are used as disinfectants in water treatment.
1.4.1 Effectiveness of Disinfection The relative effectiveness of disinfectants against a range of microorganisms under specific conditions has been determined and a C.T value calculated for each. The C.T value is calculated by multiplying the concentration of disinfectant (in mg/L) by the contact time (in minutes) that achieves a 99% inactivation of the microorganism being tested. A low C.T value indicating a strong disinfectant. Table 1.2 summarises the C.T values for chlorine, chloramine, chlorine dioxide and ozone against microorganisms including E. coli, viruses and protozoan cysts. Although C.T values have yet to be established for Cryptosporidium, several researchers have demonstrated that ozone can inactivate this pathogenic microorganism, whereas chlorine cannot [105]. The values clearly illustrate the high effectiveness of ozone, chlorine dioxide and chlorine and the weaker susceptibility of microorganisms towards chloramine.
Table 1.2: C.t Values for 99% Inactivation of Microoganisms by Disinfectants at 5oC. (Adapted from Hoff [18].)
Microorganism Chlorine Chloramine Chlorine dioxide Ozone pH 6-7 pH 8-9 pH 6-7 pH 6-7 E. coli 0.034 – 0.05 95 – 180 0.4 –0.75 0.02 Polio I 1.1 – 2.5 768 – 3740 0.2 – 6.7 0.1 – 0.2 Rotavirus 0.01 – 0.05 3806 – 6470 0.2 – 2.2 0.006 – 0.6
Phage f2 0.08 – 0.18 - - - G. intestinalis cysts 47 - >150 - - 0.5 – 0.6 G. muris cysts 30 – 630 - 7.2 – 18.5 1.8 – 2.0
1.4.2 Chloramines
Chloramines (NHxCly) are formed when chlorine and ammonia are added to water. Dichloramine and trichloramine are stronger disinfectants than monochloramine yet are less stable and possess extremely offensive odours and consequently, monochloramine is the disinfectant of choice. A high excess of chlorine to ammonia favours the formation of di- and trichloramine and so controlling the ratio of chlorine to ammonia is important to ensure that monochloramine is the major species formed. The use of monochloramine as a disinfectant is effective in reducing the levels of chlorinated DBPs [36]. The extent of THM and DBP reduction is influenced by the order of chlorine and ammonia addition and whether pre-formed chloramine is used. By adding the ammonia first, chlorine is converted to chloramine before it has the ability to react with NOM and form appreciable levels of THMs and other chlorinated compounds. The disinfection efficiency of monochloramine however is weak and requires 25 to 100 times the contact time of free chlorine for equivalent disinfection. Hence the required contact time for inactivation of viruses and protozoan cysts with chloramine post- disinfection is rarely achievable. The oxidising power of monochloramine is also weaker and less likely to reduce taste and odour causing compounds that may be present in raw waters [106]. Disinfection with chloramine results in the formation of cyanogen chloride at low micrograms per litre
7 CRC for Water Quality and Treatment Research Report 9-2001 levels [37]. Cyanogen chloride is highly irritant and poisonous. It is rapidly metabolised in the body to form cyanide. The Australian guideline value for cyanogen chloride is set at 80 µg/L and is based on the cyanide value [34].
1.4.3 Chlorine dioxide Chlorine dioxide is a reactive gas and a highly effective disinfectant with a biocidal activity comparable to chlorine. It is also a strong oxidant and has the ability to oxidise iron and manganese complexes and reduce colour without the production of chlorinated DBPs that are formed by chlorine [107]. The reactivity of chlorine dioxide makes handling and transportation a problem and so chlorine dioxide is generally generated on site by the acidification of sodium chlorite. A main disadvantage of the use of chlorine dioxide as a disinfectant is that it dissociates in aqueous solution forming chlorite and chlorate. Chlorate, chlorite and chlorine dioxide are all potentially toxic compounds and consequently the maximum residual concentration has been set at 0.4 mg/L. At this low dosage adequate disinfection is generally not accomplished [34].
1.4.4 Ozone Ozone is a reactive gas generated by passing an electric discharge through dry air or oxygen and must be generated on site. Ozone is gaining popularity in the public water supply industry as an alternative primary disinfectant and oxidant. Because of its high reactivity and hence short lifetime, ozone cannot be used as the sole disinfectant and so a post disinfectant is required to maintain a disinfectant residual in the distribution system to prevent potential bacterial regrowth. Ozone is a powerful oxidant having a standard electrode potential in acidic solution of 2.07V. It is a stronger oxidising agent than other oxidants currently used in water treatment. Table 1.3 lists the oxidation potentials of these common oxidants.
Table 1.3: Standard Oxidation Potentials of Common Oxidants Used in Water Treatment.
Disinfectant Standard Oxidation Potential Species
O3 -2.07 HOCl -1.49
Cl2 -1.36
ClO2 -1.15
NH2Cl -0.74
Ozone’s oxidising ability makes it effective for the removal of pathogenic microorganisms and algal toxins [38,39,105], taste and odour causing compounds [40,41], colour [42] and resistant pesticides such as the triazine herbicides [43].
1.5 Reactions of Ozone In aqueous conditions, ozone is a very reactive molecule yet is selective in the chemical reactions it undergoes. It reacts rapidly with some types of organic and inorganic molecules and more slowly with others. Substances found in raw waters that are most reactive with ozone include metal ions existing in their lower oxidation states such as Mn2+, Cu+ and Fe2+ and substances containing sp2 hybridised carbon atoms such as unsaturated fatty acids, terpenes and aromatic compounds [44,45].
1.5.1 Reactions of ozone with organic compounds Ozone is well known to react with organic compounds via two mechanistic pathways. The first pathway, referred to as ozonolysis, involves molecular ozone adding to sites of unsaturation, that is,
8 CRC for Water Quality and Treatment Research Report 9-2001 carbon-carbon double bonds. The mechanistic scheme is outlined in Figure 1.5 and involves the addition of ozone to the double bond to form an ozonide (1). In an aqueous environment hydrolysis of the ozonide cleaves the original carbon-carbon double bond to yield the cleaved products (2). The reaction results in the formation of aldehydes and ketones.
O3 O CC C C C C O O OO O Ozonide 1
H2O
CO
Cleavage products 2 OC
Figure 1.5: Reaction of Ozone with Olefins.
Under acidic conditions ozone does not react appreciably with water and is present as the ozone molecule with molecular ozone reactions dominating. However, as the pH is elevated ozone spontaneously decomposes in water to form many reactive free radical species such as the hydroxyl radical [46]. Ozone decomposition is catalysed by the hydroxide ion and the currently accepted mechanism for the decomposition of ozone is presented in Figure 1.6. The reaction scheme clearly illustrates the complexity of the decomposition process and the many reactive intermediates that form.
1.5.2 Reactions of ozone with NOM Several researchers have investigated how ozonation alters NOM in raw waters. Ozonation is shown to break long chain high molecular weight NOM into smaller oxygenated fragments [47,48]. Gloor and coworkers using exclusion chromatography with dissolved organic carbon detection noticed that compounds with molecular weights greater than 1500 Daltons were reduced following ozonation and an increase in the lower molecular weight region had been observed [48]. The change in molecular weight distribution of NOM was similarly observed by Edwards and coworkers [47]. Accompanying the oxidative fragmentation of NOM, the organic acidity was found to increase as a result of ozonation with substantial quantities of oxalic acid formed [47,49].
9 CRC for Water Quality and Treatment Research Report 9-2001
initiation: - - O3 + OH HO2 + O2
propagation/termination: - - HO2 + O3 O3 + HO2 - - HO2 + OH O2 + H2O - - O2 + O3 O3 + O2 - - O3 + H2O OH + O2 + OH - - O3 + OH O2 + HO2 - - O3 + OH O3 + OH
OH + O3 HO2 + O2 2- - - OH + CO3 OH + CO3 - - CO3 + O3 products (CO2 + O2 + O2) - + HO2 = O2 + H - - HO2 + OH O2 + ( + H2O ) - + H2O2 = HO2 + H
Figure 1.6: Aqueous Decomposition of Ozone.
In recent years there has been interest in using ozone as a viable alternative to chlorination for oxidation and disinfection purposes. The main reason for this interest is the need to eliminate the formation of THMs and other organochlorine compounds formed by the reactions of chlorine with naturally occurring organic material in water sources. As NOM is the major source of precursor material in the formation of chlorination by-products, ozone is gaining favour in drinking water treatment to oxidise the NOM to a more hydrophilic form that is no longer attacked by chlorine on subsequent chlorination hence reducing the quantity of DBPs formed. However there are conflicting reports in the literature on the impact of preozonation on THM precursors. Several studies [50-54] have shown modest to large decreases in THM concentrations following chlorination of ozonated waters, whereas others [55,56,57] have reported actual increases in THM precursors or chlorinated organic material. The outcome probably depends on the degree of oxidation of the NOM achieved. Ozone’s strong oxidising power is highly effective in the management of various water quality problems, however because of its oxidising ability, it is inevitable that by-products are produced by the reactions of ozone with NOM. Studies have shown ozonation results in the formation of aldehydes and dialdehydes [58-63]. The formation of these compounds is significant as aldehydes can potentially cause adverse health effects [64]. Formaldehyde and acetaldehyde have produced respiratory tumours in animals after inhalation exposures. Glyoxal has been shown to promote stomach tumours. Aldehydes may also cause immediate water quality deterioration due to their objectionable organoleptic properties. Low molecular weight aliphatic aldehydes have low odour thresholds and their presence is responsible for the characteristic fruity odours that often appear following ozonation [65]. Another important concern is that aldehydes may react with secondary disinfectants to form secondary by-products. McKnight and Reckhow [102] found that acetaldehyde produced by ozonation can react with chlorine to form chloroacetaldehyde. This substitution reaction then proceeds rapidly to form the trichlorinated product chloral hydrate. Pedersen and co-workers performed a study that demonstrated that formaldehyde can react with chloramines to form cyanogen chloride [108]. Other studies has shown that pre-ozonation can increase the formation of chloropicrin upon post-chlorination [103].
10 CRC for Water Quality and Treatment Research Report 9-2001
Other classes of by-products formed by the ozonation of NOM are ketoacids [63,66,67,68] and carboxylic acids [85]. Ketoacids reported in the literature include glyoxylic acid, pyruvic acid and ketomalonic acid. Glyoxylic acid is a known mutagen [64]. With water utilities considering the use of ozone, a better understanding of by-product formation is required as the trends reported in the literature are far from consistent and generalisations are difficult to make. Xie and Reckhow [68] reported that ketoacids were present at substantially greater levels than aldehydes following ozonation. Garcia-Araya and coworkers [67] also reported similar findings. However, Andrews and Huck observed that ketoacid levels were at least ten times lower than aldehyde levels following ozonation of both river and lake waters [63].
1.5.3 Aldehyde and ketoacid analysis Ozone has been used in drinking water treatment in some European countries for over eight decades and has only recently gained increasing popularity in Northern America. In spite of its successful use throughout the decades, there is only limited information available in the literature on the formation and occurrence of ozonation DBPs during water treatment. By-products formed as a result of ozonation are usually oxygenated and polar in nature. Analysing for these compounds had proven to be difficult because of their chemical nature and extremely high solubility in water. Early work on the analysis of aldehydes from aqueous media was achieved by the use of either conventional solvent extractions or closed-loop stripping techniques [69]. These techniques however failed to detect the lower molecular weight, polar compounds such as formaldehyde, glyoxal etc. Over the last decade several groups of researchers have investigated the potential use of derivatising reagents for making highly polar and hydrophilic aldehydes amenable to solvent extraction which can then be followed by chromatographic separation and detection [62,70,71]. A successful method for the analysis of aldehydes and other carbonyl compounds uses pentafluorobenzyl hydroxylamine (PFBHA) as the derivatising agent. This was first described by Yamada and Somiya [70] and later improved by Glaze and coworkers [62] and Sclimenti and coworkers [71]. In the current work the procedure has been modified and method validation was also involved. A second major class of ozonation by-products, first reported by Xie and Reckhow, are the ketoacids [66]. Knowledge on the formation and occurrence of ketoacids in drinking water treatment is limited and only a few studies have been reported [63,66-68]. Analysis of ketoacids from water is achieved by a double derivatisation method with PFBHA and diazomethane. Once again, the procedure was modified in this work and relevant method validation was performed in our laboratory.
1.5.4 Biodegradable organic matter Numerous studies have shown that ozonation increases the biodegradability of dissolved natural organic matter [58,72-75]. Ozonation reacts with high molecular weight fulvic and humic acids forming lower molecular weight, more polar, oxygenated products as described in section 1.5.2. These ozonation by-products, namely aldehydes, ketoacids and carboxylic acids, are readily biodegradable and constitute an important fraction of the biodegradable organic carbon and as a consequence can lead to increased microbial activity in drinking water distribution networks [76,77]. Aesthetic concerns may be associated with microbial growth causing deterioration of taste, odour and colour of drinking water as well as deterioration in the microbiological quality of the water [78]. A variety of tests have been developed to estimate the amount of biodegradable organic matter (BOM) in water [79,80]. These techniques measure the BOM by means of the growth of an inoculum (assimilable organic carbon, AOC) or by means of the activity of indigenous bacteria (biodegradable organic carbon, BDOC). BDOC refers to the portion of the organic carbon in water that can be mineralised by heterotrophic organisms. BDOC is determined by the difference in the DOC of the water sample after and prior to inoculation. AOC refers to the portion of BDOC that can be converted to cell mass and expressed as a carbon concentration by means of a conversion factor or calibration. A method for measuring AOC has been reported by Drikas and coworkers [79] and involves a microbiological assay that is dependent on the rate of bacterial growth. Water samples to be tested are first sterilised by filtration and inoculated with indigenous microorganisms. Biomass and hence bacterial growth is measured by monitoring changes in the turbidity of the sample. Chapter 3.4 describes the procedure in further detail.
11 CRC for Water Quality and Treatment Research Report 9-2001
1.6 Research objectives In Australia DBP issues have received little attention compared with overseas, indicative of the Australian guideline value for trihalomethanes (THMs) which is well above overseas standards. However public concerns and the tightening of regulations overseas may influence Australian guidelines in the future. Although ozone is gaining in popularity overseas, research on the production of ozonation by-products is still in its early stages. Moreover, there is a lack of information about ozonation by-products and the impact of preozonation in Australia. Surveys and studies have been performed overseas in order to monitor local situations, yet the work described here is the first conducted in Australia and is quite timely as the interest in the use of ozone as a pre-disinfectant is dramatically increasing. The relative yields of several classes of ozonation by-products reported in the literature do not seem to follow any trends. Furthermore, there are conflicting reports on the impact of ozone on the formation of chlorinated DBPs following chlorination of ozonated water. In addition, the objective of this research is not only to have the practical benefit of understanding the effects of ozonation on the formation of chlorinated organic compounds but also to shed light on the discrepancies presented in the literature regarding ozonation followed by chlorination as a treatment regime. The primary experimental goals of this research were to: (i) evaluate analytical methods for DBPs by assessing analyte recoveries, method precision, accuracy, ruggedness and simplicity. (ii) determine the ozonation by-products formed from the ozonation of three South Australian surface waters of different character. (iii) determine the effects of ozone dosage on ozonation by-products. (iv) determine the impact of ozonation on chlorinated DBPs. (v) evaluate the effects of ozone on assimilable organic carbon formation. (vi) assess the DBPs produced through the treatment process of a small scale water treatment plant utilising ozone.
12 CRC for Water Quality and Treatment Research Report 9-2001
2 The Water Sources and Their Chemical Characteristics
Three South Australian reservoir waters differing in character were investigated in this study, namely Hope Valley, Myponga and Tod Reservoirs. Hope Valley Reservoir is one of six systems serving the metropolitan community of Adelaide. All metropolitan systems, except Myponga Reservoir, use River Murray water to supplement their natural catchment inflow supplies. Adelaide is heavily reliant on the River Murray as a water source. In a dry year over 80% of Adelaide’s water supply could be from the River Murray while in a wet year it may be as little as 20%. Myponga Reservoir is the sixth and last constructed system to serve Adelaide. The Myponga reservoir, unlike the other metropolitan reservoirs, cannot be augmented with water pumped from the River Murray. The natural catchment surrounding the area is the sole origin of its water. Tod Reservoir, together with water from a number of bores, serves a large part of the Eyre Peninsula. The Peninsula is a large area that is low in rainfall and restricted to only one surface water source. Water quality parameters for each water source are listed in Table 2.1.
2.1 Hope Valley Raw Water The construction of Hope Valley Reservoir began in 1869 and was completed in 1872. The reservoir is located about 12km northeast of the city of Adelaide and directly supplies reticulated water to over 200000 people. It is an off-stream storage reservoir receiving water from the River Torrens via an aqueduct. The area served by Hope Valley Reservoir is shown in Figure 2.1. The boundaries shown are not fixed and may vary from time to time depending on the demand and capacity of the reservoir. The reservoir has a capacity of 3470ML and a surface area of about 600000m2. Although the capacity of the reservoir is small it serves a large area. This is possible as the reservoir is also supplied with water from the River Murray via two larger storage reservoirs located upstream. The amount of River Murray water depends upon seasonal and demand factors. It can be close to 80% at times during summer periods. Without this additional backup the water in the Hope Valley Reservoir during peak summer demand periods would last only about two weeks. Hope Valley Reservoir in effect functions as a service reservoir collecting water from other supply sources and distributing the water to the community. With this is mind and the small capacity of the reservoir the water’s residence time is not long. Turbidity of the water is hence fairly high, as the short stay in the reservoir does not allow appreciable settling to take place. Hope Valley raw water can be characterised as slightly coloured with a moderate dissolved organic carbon (DOC) content.
13 CRC for Water Quality and Treatment Research Report 9-2001
Figure 2.1: Area Served by the Hope Valley Reservoir. (The lighter shaded area is supplied by the reservoir)
2.2 Myponga Raw Water Myponga Reservoir is situated on the Fleurieu Peninsula about 55km south of Adelaide. The large capacity reservoir was completed in 1962 and positioned onstream of the Myponga River. It has a capacity of 26800ML, a surface area of about 2.8km2 and utilises a catchment area spanning 124km2. The purpose of the reservoir was to supply the areas immediately north and south of the district and parts of the Encounter Bay region. The area served by the reservoir is shown in Figure 2.2. Unlike all other metropolitan reservoirs, Myponga Reservoir is not augmented with water pumped from the River Murray. Myponga raw water is highly coloured with a high DOC content due to the sandy catchment. The water’s salinity is comparable to Hope Valley water but its alkalinity is much lower.
14 CRC for Water Quality and Treatment Research Report 9-2001
Figure 2.2: Area Served by the Myponga Reservoir.
2.3 Tod Raw Water Eyre Peninsula in South Australia is a dry region with only 3% of the land area having an annual rainfall of 500mm or more. The low rainfall is the reason for the lack of perennial flowing streams on the peninsula. The Tod River is the only stream providing reliable flows in normal rainfall years in the region. Tod River Reservoir is situated 27km north of Port Lincoln and offstream to the river. Water is diverted into the reservoir by a weir constructed across the river. The reservoir has a capacity of 11300ML and a surface area of 1.34km2. An extensive pipeline system from the Tod River reservoir assures water supply to a large area on the peninsula. A number of regions on the Peninsula contain underground water basins and are used as supplementary sources of water. Low rainfall, low river flowrate and a high evaporation rate subsequently results in Tod raw water having a high salinity. As a result the water hardness is very high and alkalinity is also significant. The NOM content and colour of the raw water however is low. Figure 2.3 shows the state of South Australia and the location of Tod Reservoir.
15 CRC for Water Quality and Treatment Research Report 9-2001
Figure 2.3: Tod Reservoir Located on Eyre Peninsula.
2.4 Reagent Water Water used to prepare standard solutions, reagent solutions, sample blanks and the washing of glassware is prepared by passing consumer tap water through large-scale ion exchange resins followed by reverse osmosis (RO). The RO treated water is then passed through a Milli-Q PlusTM water purification system (Millipore Corporation Ltd.). The resultant water has a resistivity of greater than 18MΩ.cm and an organic content of less than 100µg/L as DOC.
Table 2.1: Raw Water Quality Data.
Myponga Hope Valley Tod Sample Date 12th Oct. 98 18th Nov. 98 20th Jan. 98 pH 7.83 7.95 8.3 Colour (HU) 70.2 26.9 9.8 Turbidity (NTU) 2.7 5.4 4.9
UV254 0.444 0.160 0.223 DOC (mg/L) 12.0 6.3 4.1 SUVA (L/mg.cm) (x10-2) 3.70 2.54 5.44
Alkalinity as Ca2CO3 (mg/L) 54 91 156 Total Hardness (mg/L) 102 141 741 Br - (mg/L) 0.56 0.37 4.5 Ca2+ (mg/L) 19.2 24.9 92 Mg2+ (mg/L) 13.1 19.1 124
16 CRC for Water Quality and Treatment Research Report 9-2001
- HCO3 (mg/L) 66 111 190 Total Dissolved Solids (mg/L) 350 310 3200
17 CRC for Water Quality and Treatment Research Report 9-2001
3 Ozonation: Application and By-Product Analysis
3.1 Ozonation Experiments Ozone was generated from ultra high purity oxygen using an Ozonia LN103 ozone generator. Oxygen passes through a high voltage that dissociates oxygen molecules (O2) into atomic oxygen. Collisions between atomic oxygen and molecular oxygen generates ozone. The ozone gas produced then passes through an Orec O3DM-110 ozone monitor that calculates and monitors the ozone concentration. Bench-scale ozonation experiments were performed at room temperature in semi-batch mode. That is, a continuous stream of ozone gas was bubbled through a static liquid volume. A glass cylindrical reactor was used with a capacity of 1L (620mm in height). All materials in contact with ozone and the water sample were made of glass, teflon or stainless steel. The ozonation set-up is illustrated in Figure 3.1. Ozone gas enters the glass reactor through ‘pin-size’ holes located at the end of teflon tubing that is positioned at the bottom of the reactor. The reactor is fitted with a magnetic stirrer and with vigorous stirring, bubbles ranging in size from approximately 0.5mm- 4mm in diameter are formed. Both the liquid depth within the reactor and the bubble path length through the liquid is 580mm. An efficient ozone transfer from the gas phase to solution was obtained by optimising various parameters. A high concentration of ozone in the gas phase, long bubble residence times and a high gas to liquid surface area were all optimised to maximise ozone transfer between phases. The time required for ozonation was determined by the raw water’s dissolved organic carbon (DOC) level, the ozone gas concentration, flow rate into the reactor, the reactor volume and the desired ozone dose to dissolved organic carbon (O3:DOC) mass ratio. Ozonation time began when bubbles first appeared in the reaction vessel and stopped by diverting the flow by means of a stainless steel 3- way valve. The contact times were calculated to achieve applied O3:DOC mass ratios of 0.5:1, 1:1, 2:1, 3:1, and 5:1 mgO3:mgDOC. Data in this report are presented in terms of applied ozone dose. The applied ozone doses of 1:1 and 2:1 were selected to reflect current drinking water treatment practices [58]. The high dose range of 3:1 and 5:1 were also investigated to provide a measure of the maximum yields of by-products that could be achieved. Gaseous ozone concentrations were determined by the use of a potassium iodide trap. The effluent gas from the reactor is bubbled through a potassium iodide (KI) solution. The concentration of ozone can be determined by titrating the KI solution against a standard thiosulfate solution and using starch as an indicator. The titration was performed in accordance with 2350-Ozone Demand – Semi-Batch Method [80]. This procedure was also used to assess the accuracy of the ozone monitor. By using the same ozonation scheme outlined in Figure 3.1 and replacing the raw water sample with a KI solution, the ozone concentration of the influent gas was determined and compared to the reading calculated by the monitor. The accuracy of the monitor was assessed prior to ozonation experiments for each water source and the readout was consistently greater than 90% of the concentration determined by titration.
18 CRC for Water Quality and Treatment Research Report 9-2001
Figure 3.1: Experimental Set-up for Semi-batch Ozonation.
Following ozonation, an aliquot of the ozonated sample was immediately taken for an aqueous ozone residual determination. Ozone residuals were measured using potassium indigo trisulfonate as per method 4500-O3 B Indigo Colorimetric Method [80]. The raw waters were ozonated at each ozone dose in duplicate as two litres was required to provide a sufficient volume to complete appropriate analyses intended for this study. Each of the ozonated waters were carefully transferred to a 2.5L teflon-lined screw capped amber coloured glass bottle. The ozonated waters were stored overnight at 4oC and residual dissolved ozone was not quenched but allowed to dissipate. When required, the ozonated waters were then chlorinated within 24 hours and analysed for all determinands within a further 48 hours.
3.2 Raw Waters and Their Ozone Demand Raw waters were ozonated at five applied ozone doses, at their ambient pH and at room temperature (nominally at 20oC). The ozone consumption and dissolved ozone residuals were determined for each dose and are displayed in Table 3.1.
19 CRC for Water Quality and Treatment Research Report 9-2001
Table 3.1: Transferred Ozone Doses and Ozone Residualsa.
Myponga Raw Water b O3 (In) O3 Residual O3 (Out) O3 Consumed 54.5 1.6 18.0 36.5 32.7 0.4 10.8 21.9 21.8 0.04 5.7 16.1 10.9 <0.01 2.2 8.7 5.4 <0.01 0.9 4.5 Hope Valley Raw Water b O3 (In) O3 Residual O3 (Out) O3 Consumed 21.0 1.29 9.8 11.2 12.6 0.28 5.3 7.3 8.4 0.10 2.1 6.3 4.2 0.01 0.6 3.6 2.1 <0.01 0.1 2.0 Tod Raw Water b O3 (In) O3 Residual O3 (Out) O3 Consumed 20.5 0.7 6.4 14.1 12.3 0.3 3.5 8.8 8.2 0.04 2.1 6.1 4.1 <0.01 1.1 3.0 2.0 <0.01 0.2 1.8 aAll ozone levels expressed in milligrams. b O3 residual is not subtracted to determine O3 consumed as the residual is allowed to dissipate.
The fate of ozone in water is complex. Ozone can oxidise reduced inorganic species such as Fe2+ and Mn2+ and metal complexes, as well as react with organic compounds via a molecular or a free radical pathway to form oxygenated organic compounds. Ozone demands are therefore significantly affected by the chemical and physical characteristics of the water. The reactivity of ozone is also influenced by temperature [81]. Temperature strongly affects the reaction kinetics and thus the demand applied in a given contact time. Ozonation was carried out at room temperature, typically 20oC. The determination of ozone demands is necessary for water utilities intending to implement ozone in their treatment process. Ozone demands play a role in determining the ozone dose required and hence the cost and feasibility in the use of ozone. When water is dosed with ozone, there is an immediate demand, which is referred to as the ozone demand. Aqueous ozone concentrations at doses below the O3 demand are not detectable. Applied ozone doses greater than the ozone demand generate an ozone residual and aqueous ozone may be determined. Ozone demands for particular waters must be met to obtain the benefits that come with ozone in its use in water treatment. For ozone to be fully effective as a disinfectant, the ozone dose applied should be greater than the demand [81]. Franklin McKnight and coworkers have reported that manganese oxidation is not complete unless the ozone dose exceeds the water’s demand [82]. Carlson and coworkers also showed that bromate begins to form, from the oxidation of bromide, in appreciable amounts only with doses greater than the immediate demand [81]. The ozone demands in Myponga, Hope Valley and Tod raw waters were found to be 16.2, 6.0 and 5.5mg/L respectively. The demands were determined by applying increasing ozone doses and calculating the aqueous ozone concentration immediately after the required contact time had elapsed. Figures 3.2 to 3.4 show the results of the ozone demand study.
20 CRC for Water Quality and Treatment Research Report 9-2001
Each water, once the demand was met, showed a strong linear relationship between the ozone dose and the ozone residual. It can be seen that ozone demand is highest in the water with the highest NOM content (Myponga water). The DOC concentrations of Hope Valley and Tod water are 6.3 and 4.1mg/L respectively. Interestingly, the ozone demands are quite similar. The ozone demand in Tod water would be expected to be lower due to the water’s lower DOC content. Tod water’s higher alkalinity and greater presence of inorganic metal ions requires a greater concentration of ozone to achieve the water’s immediate demand. This clearly indicates ozone’s ability to rapidly oxidise inorganic species. The nature of the DOC may also influence the demand. The slope of the curve once the demand is met also supports this view. Only approximately half the aqueous ozone residual is found in Tod water compared with Hope Valley water at equivalent ozone doses above the demand.
1.8
1.6
1.4
1.2
1.0
0.8
0.6 Ozone Demand
Aqueous (mg/L) Ozone 0.4
0.2
0.0 0 5 10 15 20 25 30 35 40 Transferred Dose (mg/L)
Figure 3.2: Ozone Demand in Myponga Raw Water.
1.4
1.2
1.0
0.8
0.6 Ozone Demand 0.4 Aqueous (mg/L) Ozone 0.2
0.0 024681012 Transferred Ozone (mg/L)
Figure 3.3: Ozone Demand in Hope Valley Raw Water.
21 CRC for Water Quality and Treatment Research Report 9-2001
0.8
0.7
0.6
0.5
0.4 Ozone Demand 0.3
0.2 Aqueous Ozone (mg/L) 0.1
0.0 0 2 4 6 8 10 12 14 16 Transferred Dose (mg/L) Figure 3.4: Ozone Demand in Tod Raw Water
3.3 Ozonation By-Products and Evaluation of the Methodology Classes of ozonation by-products investigated in this study include aldehydes, ketoacids and carboxylic acids. Because method development and validation of these analytical methods were an essential part of this project, the success or failure of each method is discussed. All extracts were analysed using a dual column gas chromatograph with electron capture detection (model 3400 Cx; Varian Instrument Group, Sunnyvale, California) equipped with a split/splitless injector (Varian 1080 injector), dual electron capture detectors and an autosampler (Varian 8200). The analytical columns used were a DB-1 and a DB-1701 (J&W Scientific). Appendix 1 lists the column dimensions and chromatographic conditions applied.
3.3.1 Determination of aldehydes in water Low molecular weight aldehydes are polar and difficult to extract from aqueous solutions. Without prior derivatisation, these analytes cannot be detected at the low parts per billion (ppb) range. Pentafluorobenzyl hydroxylamine (PFBHA) is found to be a useful derivatising reagent for carbonyl containing compounds. PFBHA (I) effectively derivatises highly polar and hydrophilic compounds making them amenable to solvent extraction followed by gas chromatographic separation. Aldehydes are derivatised, in aqueous solution, to form their corresponding pentafluorobenzyl oximes, (II) and (III). The reaction scheme is shown in Figure 3.5. Note that with the exception of symmetrical ketones and formaldehyde, two geometrical isomers of the derivatives are formed.
CH2 O R1 F5 NC CH2 ONH2 II R2 O F5 I C OXIMES 45oC, 1 hr. 45 min. R1 R2 R= H, alkyl CH2 O R2 F5 NC III R1 I : PFBHA ; II & III : E- and Z- isomer
Figure 3.5: PFBHA Derivatisation of Carbonyl Compounds.
22 CRC for Water Quality and Treatment Research Report 9-2001
The oximes are extracted with hexane, analysed by gas chromatography with electron capture detection (GC-ECD) and quantified using procedural standard calibration. Previous literature has reported the addition of the peak areas of both isomers prior to quantitation [62]. This has a tendency to lead to erroneous results if chromatographic interferences co-elute with one of the two isomers. Quantitation in this work involved averaging the results obtained from both isomers and both columns and reporting the value if the mean was within 10% of the four values. Other researchers have included an extraction clean up step to reduce the level of unreacted derivatising reagent, PFBHA, in the final extract [62,70,71]. This step was removed from the procedure as a significant amount of PFBHA still remained in the final extract and since complete resolution between PFBHA and the analytes of interest was obtained, this step was not deemed necessary. Eight aldehydes were monitored throughout the ozonation and chlorination experiments. These are shown in Figure 3.6.
O O formaldehyde heptaldehyde HC H3C(CH2)5C H H
O O acetaldehyde benzaldehyde H3CC C H H
O propionaldehyde glyoxal H O H3CCH2 C CC H O H
butyraldehyde O methyl glyoxal H3C O H3CCH2CH2 C CC H OH
Figure 3.6: Potential Organic Ozonation By-products: The Aldehydes.
3.3.1.1 Analytical procedure Water samples for aldehyde analysis were collected and stored in 100mL teflon sealed glass bottles free of any headspace. Aliquots (20mL) of standards (see Appendix I) and samples were transferred to 40mL teflon sealed glass reaction vials. To this, 1mL of a KHP/NaOH buffer system (pH6) followed by 1mL 10mg/mL PFBHA solution was added. The reaction vials were then placed in a water bath set at 45oC for 1hr 45min. The samples and standards were allowed to cool to room temperature and two drops of concentrated sulphuric acid were added to quench the derivatising reaction. The oxime derivatives were then extracted with hexane (4mL) containing 1,2-dibromopropane (500µg/L) and analysed by GC-ECD. 1,2-dibromopropane is used as an internal reference standard to determine relative retention times for each oxime derivative. Samples were stored headspace free at 4oC and analysed within 48 hours. Residual ozone in the water samples was not quenched but allowed to dissipate in solution prior to aldehyde analysis and sample preservation was not necessary as holding times were kept as short as possible. The procedure is schematically outlined in Figure 3.7.
3.3.1.2 Sample dechlorination Aldehydes were also monitored following chlorination to see whether chlorination had any effect on the aldehyde levels formed from ozonation. Residual chlorine, dissolved ozone and other oxidising substances interfere with the PFBHA derivatisation and so dechlorination is required prior to analysis. Munch and coworkers [83] showed that typical dechlorinating agents such as sodium sulfite, sodium
23 CRC for Water Quality and Treatment Research Report 9-2001 thiosulfate and ascorbic acid when added to chlorinated tap water cause increases in the levels of certain aldehydes and decreases in others. They found that ammonium chloride was the most suitable dechlorinating agent. Ammonium chloride removes free chlorine by forming the less reactive chloroamine that does not interfere with the derivatisation.
20mL water sample
1mL pH6 buffer 1mL PFBHA solution (10mg/mL)
Derivatisation: 45oC 1hr 45min. (allow to cool ~15min.)
H2SO4conc. (2 drops) hexane (4mL) containing 500µg/L 1,2- dibromopropane
Extraction (shake for 90sec)
Aqueous layer Organic layer
Waste GC-ECD analysis
Figure 3.7: Analysis of Aldehydes in Water.
3.3.1.3 Background contamination Formaldehyde is a ubiquitous chemical and present at trace levels in the environment. Reagent water can be a major source of formaldehyde contamination yet contamination from the atmosphere may also be significant. Munch and coworkers investigated the potential of formaldehyde contamination
24 CRC for Water Quality and Treatment Research Report 9-2001 after exposure to room air [83]. Uncapped bottles of reagent water open to air showed significant increased levels of formaldehyde over 24, 48 and 72 hour periods. Trace levels of formaldehyde were present in the reagent water used in this study. Levels were negligible compared to formaldehyde levels produced following ozonation nevertheless a considerable effort was made to remove the background formaldehyde from Milli-Q water. Several treatment processes were applied and are listed below: (i) Milli-Q water was refluxed with acidified potassium permanganate for one hour and distilled. (ii) Milli-Q water was irradiated with ultra-violet light for one hour. (iii) Milli-Q water was passed through an anion exchange resin in the bisulfite form. Aldehydes react and form a complex with the bisulfite and hence retained on the stationary phase. Formaldehyde levels for Milli-Q water and Milli-Q water subsequently treated are shown in Table 3.2. Formaldehyde levels in Milli-Q water were found to be low and at trace levels and further treatment to remove formaldehyde was not successful. The water quality produced by our laboratory’s water purification system was satisfactory and further time-consuming treatment was not investigated.
Table 3.2: Evaluation of Reagent Waters.
Treated water Formaldehyde (µg/L) Milli-Q (MQ) water 1.9 MQ distilled over permanganate 1.7 MQ irradiated with UV 1.9 MQ passed through ion exchange column 2.1
3.3.1.4 Recoveries and limits of quantitation Recoveries were determined by spiking both reagent water and raw water with known amounts of aldehydes at two concentrations. Two concentrations were chosen to assess recoveries at both ends of the calibration range. Raw water was spiked to determine whether there are any interferences caused by the presence of other components in the sample matrix. Ten replicates were used to calculate the standard deviation of the recoveries in reagent water and raw water at both spiking levels. The limits of quantitation (LOQ) were set at a level of half the concentration of the lowest calibration standard, which represented a signal to noise ratio of approximately 10 to 1. The recoveries and limits of quantitation are shown in Tables 3.3 and 3.4 respectively.
Table 3.3: Aldehyde Recoveries from Spiked Reagent and Raw Water.
Reagent water Raw water - Hope Valley reservoir Analyte conc. rec. SD conc. rec. SD Conc. rec. SD Conc. rec. SD (a) (a) (a) (a) (ppb) (%) (ppb) (%) (ppb) (%) (ppb) (%) Formaldehyde 2 108 7 50 83 4 2 133 14 50 87 5 Acetaldehyde 2 94 10 50 91 7 2 112 12 50 92 4 Propionaldehyde 2 94 6 50 87 3 2 108 11 50 90 4 Butyraldehyde 2 86 6 50 88 4 2 101 12 50 91 4 Heptaldehyde 2 93 15 50 87 4 2 96 22 50 98 3 Benzaldehyde 2 86 9 50 90 5 2 89 6 50 93 3 Glyoxal 2 92 6 50 94 3 2 96 9 50 97 2 Methyl glyoxal 2 107 4 50 96 5 2 89 7 50 91 6 (a):Standard deviation of the percent recoveries, rec.: recoveries, conc.: concentration, ppb=µg/L.
25 CRC for Water Quality and Treatment Research Report 9-2001
Table 3.4: Limits of Quantitation for Each Aldehyde.
ANALYTE LOQ ( g/L) ANALYTE LOQ ( g/L) Formaldehyde 1.0 Heptaldehyde 0.5 Acetaldehyde 0.5 Benzaldehyde 0.5 Propionaldehyde 0.5 Glyoxal 0.5 Butyraldehyde 0.5 Methyl glyoxal 0.5
3.3.2 Determination of ketoacids in water A second major class of ozonation by-products reported in overseas studies are the ketoacids [66,67,68]. The method involves a double derivatisation as first reported by Xie and Reckhow [66]. The ketoacids (1) are derivatised using PFBHA to form the corresponding oxime derivatives (2). The oxime derivatives, once extracted from the water using methyl tert-butyl ether (MTBE), are further derivatised with diazomethane (CH2N2). Methylation of the carboxyl groups (3) is required to enable chromatographic separation of these compounds. Figure 3.8 lists the ketoacids monitored in this research and Figure 3.9 outlines the reaction sequence.
H O H3CCH2 O Glyoxylic acid CC 2-Ketobutyric acid CC OOH O OH
H C O O Pyruvic acid 3 Oxalacetic acid CC CCH2 O OOH HO CC O OH O Ketomalonic acid HO C O CC OOH
Figure 3.8: Potential Organic Ozonation By-products: The Ketoacids.
CH O NH R 2 2 F5 CH2 O R O C F5 N C C O 45oC, 1hr. 45min. 2 CO HO 1 HO
CH2N2 CH2 O R F5 N C 3 CO H3CO
Figure 3.9: Double Derivatisation of Ketoacids.
26 CRC for Water Quality and Treatment Research Report 9-2001
3.3.2.1 Buffering the pH of the reaction The reaction between PFBHA and carbonyl compounds is pH dependent and consequently samples with different pH values containing equivalent ketoacid levels might be expected to form varying amounts of oxime derivatives. The reaction mechanism of the derivatisation, shown in Figure 3.10, underlines the importance of pH control. It can be seen that oxime formation involves nucleophilic attack by the basic nitrogen compound (PFBHA) on the carbonyl carbon. The solution must be acidic enough to allow an appreciable fraction of the carbonyl group to be protonated making it more susceptible to nucleophilic attack. However, high acidity will also protonate the amine and cause it to lose its nucleophilicity. Therefore, reaction pH should be controlled and the should be the same for both standards and samples. Ketoacid analyses reported in the literature [66,67,68] have not controlled the pH of the reaction mixture. The United States Environmental Protection Agency has reported markedly diverse results from interlaboratory proficiency tests on aldehyde analyses using PFBHA [83]. The large variation in results was due to laboratories not controlling the sample pH by means of a pH buffer. This work assessed the use of the buffer system currently utilised in the aldehyde procedure (section 3.3.1). The buffer combines potassium hydrogen phthalate (KHP) and sodium hydroxide (NaOH) to form a pH 6 buffer system. This buffer was found to be unsuccessful and highly incompatible due to a more polar extracting solvent being used for ketoacid extractions. MTBE readily extracted KHP from the aqueous phase causing severe adverse effects such as scavenging the methylating agent (CH2N2) and causing severe chromatographic interferences. An alternative pH 6 phosphate buffer - 2- system (H2PO4 /HPO4 ) was substituted and found to be successful. The procedure and recoveries are discussed in Sections 3.3.2.3 and 3.2.3.4 respectively.
R1 + H+ COH H R1 H+ R1 R1 R + CO 2 .. N C OH NCOH R2 NH H R 2 HR2 H
R _ 1 R2 H2O + NC NC OH2 R1 HR2 OXIME derivative Figure 3.10: Reaction Mechanism Towards Oxime Formation.
3.3.2.2 Decarboxylation of Oxalacetic acid Oxalacetic acid is a β-ketoacid and as such it might be expected to decarboxylate [93]. Oxalacetic acid was indeed found to decarboxylate. This was observed in the calibration standards that were prepared in this research. The breakdown of oxalacetic acid yields the ketoacid pyruvic acid, a well- known ozonation by-product. Oxalacetic acid is reported to form in ozonated waters and has been quantified [84]. Consequently, if oxalacetic acid is incorporated into a calibration mixture containing pyruvic acid then levels of pyruvic acid in samples will be underestimated and oxalacetic acid levels overestimated. Current studies found significant decarboxylation to occur immediately and the amount of decarboxylation increased with increasing standard solution storage time. There was no mention by the authors concerning this observation and so there is likelihood that the levels of oxalacetic acid they have reported have been overestimated. It was decided that it was not practical to prepare individual oxalacetic acid standards immediately prior to each analysis and so this was not pursued any further. Figure 3.11 illustrates the mechanism of decarboxylation as it occurs under aqueous conditions.
27 CRC for Water Quality and Treatment Research Report 9-2001
O O C CH2 C OH HO α β C O OXALACETIC ACID
O O O C CH2 C - - OH CH C O C - CO 2 OH 2 C O O
O O- CH C 3 OH CH2 C OH C C O O PYRUVIC ACID Figure 3.11: Decarboxylation of Oxalacetic Acid Under Aqueous Conditions.
3.3.2.3 Analytical Procedure Ketoacid and aldehyde analyses were conducted concurrently and so samples for both analyses were collected and stored in the same containers. As mentioned in Section 3.3.1.1, samples were stored in 100mL teflon sealed glass bottles free of any headspace. The PFBHA derivatisation procedure was - 2- identical to that of the aldehyde method with the exception that a H2PO4 /HPO4 pH6 buffer (0.2M, 1mL) was added and MTBE (4mL) was used as the extracting solvent. MTBE was also spiked with 1,2-dibromopropane (500µg/L) used as an internal reference standard. An aliquot of the MTBE extract (1.0mL) was transferred to a 2.0mL GC vial and 500µL of cold diazomethane solution (see Appendix II) was added. The GC vial was sealed, agitated and allowed to stand at room temperature for 60 minutes. Silica (10-20mg) was then added to the vial to quench any unreacted diazomethane. The vial was agitated and allowed to stand for a further 15 minutes. The MTBE extract was transferred to a new GC vial and analysed by GC-ECD. A schematic representation of the procedure is outlined in Figure 3.12. Gas chromatographic conditions are as listed in Appendix II.
3.3.2.4 Recoveries and limits of quantitation Recoveries were determined by spiking both reagent water and raw water with known amounts of ketoacids at two concentrations. Two concentrations were chosen to assess recoveries at both ends of the calibration range. Raw water was spiked to determine whether there are any interferences caused by the presence of other components in the sample matrix. The recoveries and LOQ are shown in Tables 3.5 and 3.6 respectively. The number of replicates and estimation of the LOQ are the same as mentioned in Section 3.3.1.4.
28 CRC for Water Quality and Treatment Research Report 9-2001
Table 3.5: Ketoacid Recoveries from Spiked Reagent and Raw Water.
Reagent water Raw water - Hope Valley reservoir Analyte conc rec. SD conc rec. SD conc rec. SD Conc rec. SD . (a) . (a) . (a) . (a) (%) (%) (%) (%) (ppb) (ppb) (ppb) (ppb) Glyoxylic acid 2 100 7 50 98 3 2 94 8 50 110 7 Ketomalonic 2 89 11 50 94 3 2 99 7 50 108 14 acid 2-ketobutyric 2 89 10 50 99 5 2 90 14 50 94 9 acid Pyruvic acid 2 101 9 50 91 3 2 87 6 50 103 9
(a):Standard deviation of the percent recoveries, rec.: recoveries, conc.: concentration, ppb=µg/L.
Table 3.6: Limits of Quantitation for Each Ketoacid. ANALYTE LOQ ( g/L) Glyoxylic acid 0.5 Ketomalonic acid 0.5 2-ketobutyric acid 1.5 Pyruvic acid 0.5
3.3.3 Determination of carboxylic acids in water Andrews and Huck [85] have reported a method for the analysis of low molecular weight carboxylic acids in waters. The procedure had been adapted from that described by Wong and coworkers [86]. It involves the biphasic derivatisation of the carboxylic acid (aqueous phase) and pentafluorophenol (organic phase) with dicyclohexylcarbodiimide (DCC) as a coupling agent. Pentafluorophenyl esters (1) are formed and the organic phase is analysed by GC-ECD. The reaction scheme is illustrated in Figure 3.13.
29 CRC for Water Quality and Treatment Research Report 9-2001
20mL water sample
1mL pH6 phosphate buffer 1mL PFBHA solution (10mg/mL)
Derivatisation: 45oC 1hr 45min. (allow to cool ~15min.)
H2SO4conc. (2 drops) MTBE (4mL) containing 500µg/L 1,2-dibromopropane
Extraction (shake for 90sec)
Aqueous layer Organic layer
500µL CH2N2 in MTBE
Waste Methylation 1hr, room temp.
10-20mg silica 15min.
GC-ECD analysis
Figure 3.12: Analysis of Ketoacids in Water.
30 CRC for Water Quality and Treatment Research Report 9-2001
O OH O F5 RC RC DCC OH O o 70 C, 3 hrs. F5 Toluene (1) Figure 3.13: Derivatisation of Carboxylic Acids.
The formation of pentafluorophenyl esters is dependent on the presence of tetrabutylammonium hydrogensulfate (TBAHS) and the sample pH. TBAHS acts as an ion-pair reagent and carboxylic acids must be in their ionised form for derivatisation to occur. The tetrabutylammonium cation forms an ion-pair complex with carboxylate anions. Thermal energy and mechanical agitation assists the migration of the ion-pair complex to the organic layer where derivatisation occurs with pentafluorophenol in the presence of DCC.
3.3.3.1 Identification of carboxylic acids Authentic derivatives were prepared by derivatising pure commercially available carboxylic acids. Each acid (5 mmoles) was esterified using pentafluorophenol (6 mmoles) in diethyl ether (5mL) and in the presence of dicyclohexylcarbodiimide (6 mmoles). The reaction mixture was stirred for 1 hour at room temperature. The resultant solution was gravity filtered to remove the newly formed insoluble dicyclohexylurea, diluted and analysed by GC-MS and GC-ECD to confirm retention times and mass spectra. The major ions of each pentafluorophenyl ester are listed in Table 3.7 and their respective spectra can be found in Appendix III.
Table 3.7 Major Mass Spectral Ions of Each Derivatised Acid.
Analyte MWt. Major ions* Pentafluorophenylformate 212 184, 136, 117, 155, 93, 69, …., 212 Pentafluorophenylacetate 226 184, 155, 117, 136, 183, 69, 226 Pentafluorophenylpropionate 240 57, 184, 155, 136, 117, 167, …, 240 Pentafluorophenylbenzoate 288 105, 77, 51, 155, 106, 183, 117
* base peak is highlighted in bold.
3.3.3.2 Developmental work Increasing concentrations of carboxylic acids were spiked in reagent water and analysed using the procedure reported by Andrews and Huck [85]. Following analysis by GC-ECD, detection and good linearity was observed for all analytes. The method appeared to be sound and ideal for our research needs. To further assess the reliability of the method, acids were spiked in raw waters. This was to determine whether there are any matrix interferences that may potentially affect the analysis. Hope Valley and Myponga raw water were used to determine recoveries and matrix effects. Recoveries using Hope Valley water were very low, typically less than 20%, and a negligible detection for all four acid derivatives was observed in Myponga water. The current method was therefore unsatisfactory and developmental work was required. The first approach was to assess the buffer reagent and determine whether a 0.2M phosphate system has a sufficient buffering capacity. One millilitre of buffer was added to 20mL reagent water, Hope Valley water and Myponga water. Addition of buffer was found to modify the pH of all three water samples to a similar level. This indicated that the lack of derivative detection was not due to the buffer system. Further assessment of the buffer system was however conducted. The ionic strength of the buffer system was modified to see if an increase in strength would have any effect on recoveries.
31 CRC for Water Quality and Treatment Research Report 9-2001
Carboxylic acids were spiked in reagent water and the pH adjusted using a 0.2M, 1M and 2M pH6 - 2- H2PO4 /HPO4 system. Carboxylic acid detection was observed with the 0.2M system as previously observed. Surprisingly, there was no detection of any acid derivatives using either a 1M or 2M buffer system. It was now clear that the method was extremely sensitive to the ionic strength of the water sample. A series of reactions varying different parameters was conducted in an effort to increase the robustness of the method. The acids were spiked in reagent water at a concentration of 500 µg/L. Twelve reaction mixtures were required to assess pH and the concentrations of TBAHS, PFP and DCC. The results are summarised in Table 3.8. The procedure reported by Andrews and Huck [85] was used as the control. Reaction mixtures yielding higher responses than the control were assigned a greater number of stars. The ideal pH for derivatisation was found to be at pH6. Initial judgment was that higher pHs should increase yields of derivatised product as all organic acids would be present in their ionised form. On the contrary, biphasic derivatisation was shown to be unsuccessful at pHs greater than 7. This once again highlights the sensitivity of the method towards the presence of inorganic anions in solution such as the hydroxide ion. The high level of inorganic ions is believed to prevent ion-pairing between the tetrabutylammonium cation and the carboxylate anions. Recoveries were enhanced with higher concentrations of ion-pair reagent. The reaction mixture that yielded the greatest recoveries was reaction 4 and these conditions were used to once again spike natural waters and assess matrix interferences. Acids were spiked in reagent water, tap water and three natural waters. Hope Valley Reservoir, Happy Valley Reservoir and Myponga Reservoir water were used. All waters were spiked with acids at a concentration of 400 µg/L and analysed using the conditions and procedure outlined in reaction 4 (Table 3.8) and Section 3.3.3.3.
3.3.3.3 Analytical procedure Water samples for carboxylic acid analysis were collected and stored in 40mL teflon-sealed glass vials. Aliquots (20mL) of calibration standards and samples were transferred to 40mL teflon sealed glass reaction vials. To this, 1mL pH6 phosphate buffer (0.2M), 1mL TBAHS solution (20mg/L), 3mL toluene, 150µL DCC (5mg/mL) and 150µL PFP (5mg/mL) were added. The reaction vials were then placed in a water bath set at 70oC for 3 hours. The samples and standards were allowed to cool to room temperature and the organic layer was transferred to 2mL glass GC vials for GC-ECD analysis. The procedure is schematically presented in Figure 3.14.
Table 3.8: Optimum Conditions for the Analysis of Carboxylic Acids. Buffer1 TBAHS DCC PFP Result pH (mg/L) (µL) (µL) CONTROL 6 10 150 150 *** 2 6 10 300 300 **** 3 6 20 150 150 *** 4 6 20 300 300 ***** 5 7 10 150 150 ND 6 7 10 300 300 ND 7 7 20 150 150 * 8 7 20 300 300 ** 9 8 10 150 150 ND 10 8 10 300 300 ND 11 8 20 150 150 ND 12 8 20 300 300 ND ***: Relative response obtained for the control; ND: not detected; DCC and PFP were prepared as a 2.5mg/mL solution in toluene. 1. 0.2M buffer system.
32 CRC for Water Quality and Treatment Research Report 9-2001
3.4 Determination of Bacterial Regrowth Potentials An objective of this study was to determine the effect of ozone on bacterial regrowth. It is well documented that ozonation increases the biodegradability of natural organic matter (NOM) [58,72,74,75,87]. Thus ozonation without subsequent biological stabilisation may result in bacterial regrowth in drinking water that can lead to the deterioration of the water quality. Over the last few years there has been increasing interest in the measurement of biodegradable organic matter (BOM) in raw and treated waters. Current determinations use biological assays to measure BOM and consequently the potential for bacterial regrowth [73,79,80,87]. Bacterial regrowth potentials (BRP) were determined as per the method reported by Drikas and coworkers [79]. The bacterial regrowth is measured by adding a suspension of indigenous bacteria to a sterile water sample and measuring the change in the turbidity of the sample. The increase in turbidity is directly proportional to the bacterial regrowth. Samples were stored in glass bottles at 4oC and BRP analyses were performed within 48 hours of sampling. Figure 3.15 shows a typical bacterial growth profile. The bacterial regrowth curves in this work are measured by automated monitoring of the sample’s turbidity over time.
20mL water sample 1mL pH6 buffer (0.2M) 1mL TBAHS (20mg/mL) 3mL toluene 150 µL DCC (5mg/mL) 150 µL PFP (5mg/mL)
Derivatisation: 70oC/3hours agitating water bath
allow to cool
GC-ECD analysis
Figure 3.14: Analysis of Carboxylic Acids in Water.
33 CRC for Water Quality and Treatment Research Report 9-2001
Figure 3.15: A Typical Bacterial Growth Curve.
The growth curve can be divided into three phases. Phase A is termed the lag phase. The duration of the lag phase can be short or long depending on how quickly the bacteria acclimatise themselves to the medium. In phase B the inoculum has adapted itself to the substrate and growth then proceeds exponentially. The slope in this phase relates to the growth rate (µ). A steeper slope signifies a greater substrate quality and hence the ease with which the substrate can be biodegraded. The regrowth factor (f) is defined as the ratio of turbidity after and before the experiment and is proportional to the substrate quantity. Finally, phase C is known as the stationary phase and occurs due to the exhaustion of essential nutrients and/or the bacteriostatic concentrations of metabolic wastes. A change in turbidity is associated with bacterial biomass and is generally reported in terms of a carbon concentration as acetate equivalents in µg/L [73]. The term is referred to as acetate carbon equivalents (ACE) and is derived from the growth factor and a constant that has been determined by using known concentrations of acetate as the sole carbon source.
34 CRC for Water Quality and Treatment Research Report 9-2001
4 Chlorination: Application and By-Product Analysis
4.1 Chlorination Experiments Sample chlorinations were performed according to the procedure currently used at the Australian Water Quality Centre to determine chlorinated by-product formation potentials [35]. Concentrated chlorine water was prepared by gently sparging chlorine gas in chlorine demand-free reagent water for 1-2 minutes. The chlorine water strength was determined by titrating a known micro-aliquot of the chlorine water using the DPD-Ferrous Titrimetric method [80]. The chlorine water can be stored in a glass-stoppered flask for several weeks at 4oC but the chlorine water strength must always be measured immediately prior to use. Solution concentrations were in the range of 2000-2800 mg/L as free Cl2. Aliquots of water samples (1L) were transferred to 1L amber coloured glass-stoppered bottles and - 2- buffered to a pH of 7.2 ± 0.1 with a phosphate buffer (H2PO4 /HPO4 ) system. The samples were then chlorinated by the addition of a known volume of concentrated chlorine water to achieve a resultant dose of 20mg/L free chlorine. The chlorinated waters were then incubated at 35oC for 4 hours. Following the prescribed incubation period, residual chlorine was measured by the DPD-Ferrous Titrimetric method. Immediately following chlorine residual determinations, aliquots of each chlorinated sample were transferred to headspace-free glass bottles and/or vials pre-spiked with a suitable quenching agent. The samples were then stored at 4oC and analysed within 48 hours. The type and volume of quenching agent used for each particular analysis are listed in Table 4.1.
Table 4.1: Volumes of Quenching Agents Used for Specific Analyses. Analysis Quenching Storage volume agent1
THMs 100 µL Na2SO3 20mL glass vial HAAs 250 µL NH4Cl 100mL crimp capped glass bottle CKs, HANs, CH, CP
AOXs 200 µL Na2SO3 200mL glass bottle Aldehydes 250 µL NH4Cl 100mL crimp capped glass bottle Ketoacids
Bacterial Regrowth Potentials 250 µL Na2SO3 250mL glass stoppered bottle
DOC 150 µL NH4Cl 40mL glass vials 1 Na2SO3:10% w/v; NH4Cl: 20% w/v.
4.2 Chlorination By-Products and Methodology A range of chlorinated DBPs were monitored in this study and are displayed in Figure 4.1. Adsorbable organic halogens (AOX) were analysed at the University of South Australia, Mawson Lakes Campus. The remaining chlorinated compounds were analysed by the methods currently employed at the Australian Water Quality Centre. Detailed description of the methodologies and chromatographic conditions may be found in the relevant method manuals [35].
4.2.1 Determination of trihalomethanes in water An aliquot of water sample (5mL) and sodium sulfate solution, 0.24g/mL (5mL) are sealed in a vapour tight sample vial. A Perkin-Elmer HS40 headspace autosampler agitates and heats the vial to 35oC promoting the transfer of volatile components into the vapour phase. A defined volume of headspace is transferred onto the column (DB-624; 30m x 0.32mm x 1.8µm film thickness; J&W Scientific,
35 CRC for Water Quality and Treatment Research Report 9-2001
California, USA) and analysed by gas chromatography with electron capture detection. A Varian 3400 gas chromatograph was used with the following chromatographic conditions: an initial column temperature of 60oC held for 1min, then increased to a final temperature of 180oC at a rate of 30oC/min. The injector and detector temperatures were set at 100oC and 300oC respectively. Limits of quantitation for each trihalomethane were set at 1.0 µg/L. The method’s precision is better than 5% and is currently accredited by the National Association of Testing Authorities (NATA).
4.2.2 Determination of chloroacetic acids in waters Water samples and calibration standards (30mL) were adjusted to pH 0.5 with concentrated sulfuric acid (1.0mL). Sodium sulfate (6g) was added to facilitate extraction of the non-dissociated acids. The acids were extracted with MTBE (3mL) containing 500µg/L 1,2-dibromopropane as an internal reference standard. The acids were then methylated with diazomethane to form the methyl esters. Limits of quantitation for each chloroacetic acid are 2.5, 0.1 and 0.1 µg/L for MCAA, DCAA and TCAA respectively.
4.2.3 Determination of haloacetonitriles, haloketones, chloropicrin and chloral hydrate Sample aliquots and calibration standards (35mL) were transferred to 40mL teflon-faced screw cap glass vials. Sodium chloride (approximately 8g) was added to each vial and the analytes extracted with MTBE (2mL) containing 500µg/L 1,2-dibromopropane used as an internal reference standard. The MTBE extracts were analysed by gas chromatography with electron capture detection. Limits of quantitation (LOQ) for each analyte are listed in Table 4.2.
36 CRC for Water Quality and Treatment Research Report 9-2001
Figure 4.1: Common Chlorinated Disinfection By-products.
Table 4.2: Limits of Quantitation of Various Chlorinated By-products.
Analyte LOQ (µg/L) Analyte LOQ (µg/L) Trichloroacetonitrile 0.2 1,1-dichloropropanone 0.1 Dichloroacetonitrile 0.1 1,3-dichloropropanone 0.5 Bromochloroacetonitrile 1.0 1,1,1-trichloropropanone 0.1 Dibromoacetonitrile 1.0 Chloropicrin 1.0 Chloral hydrate 0.1
37 CRC for Water Quality and Treatment Research Report 9-2001
4.2.4 Determination of adsorbable organic halogens in water The analysis measures the total quantity of halogenated organic material in a water sample and is based on the method outlined in 5320B-Dissolved organic halogen [80]. The analyses were carried out in the laboratories at the University of South Australia, Mawson Lakes Campus. The instrument is a commercial Euroglas organohalogen analyser ECS 2000 manufactured in the Netherlands by Euroglas Analytical Instruments B.V. Dissolved organic material in the water is isolated and concentrated by adsorption onto activated carbon. The activated carbon is then rinsed with a potassium nitrate solution to remove any inorganic halides present. The activated carbon containing the adsorbed organic material is pyrolysed in a furnace (800oC) oxidising the organic material to carbon dioxide and any bound halogens to hydrogen halides. The halide content is quantified by coulometric titration using a silver electrode. The LOQ for this method has been set at 2 µg/L.
38 CRC for Water Quality and Treatment Research Report 9-2001
5 Effects of Ozonation on Natural Waters
This chapter discusses the results of the ozonation experiments performed on the natural surface waters used in this study. Each surface water was ozonated at five ozone doses. The ozone doses are expressed as a ratio of ozone dose per unit of dissolved organic carbon. The ratios of applied ozone dose to DOC were 0.5:1, 1:1, 2:1, 3:1 and 5:1 respectively. Following ozonation, parameters determined included UV254 absorbance, DOC concentration, ozonation by-products and bacterial regrowth potentials.
5.1 Effects of Ozonation on UV254 Absorbance and DOC Levels Absorption in the ultra-violet range is dependent upon the electronic structure of molecules and at a wavelength of 254nm, is largely limited to conjugated systems such as conjugated carbon-carbon double bonds and aromatic rings. A variety of spectroscopic techniques have shown humic substances to contain a vast array of conjugated sites [1]. The high level of conjugation gives rise to the yellow to brown coloration typical of humic and fulvic acids. Ozone is known to react directly with aromatics and unsaturated aliphatic substances. This is clearly illustrated by the rapid loss of colour in the water samples investigated and the reduction in UV absorbance at 254nm. Increasing the ozone dosage causes further reductions in UV254 absorbance. Figure 5.1 illustrates the large drop in UV absorbance for the three waters which is an indication of the significant role that ozonation plays in altering the chemical structure of NOM. For all waters the largest reduction in UV254 absorption was found to occur at low ozone doses of 1-2:1 (O3:DOC). This trend was consistent with the observation of others [58,88]. The UV254 absorption profile at increasing ozone doses typifies that although ozone is a very reactive oxidant, it still exhibits some selectivity attacking the most reactive sites first. A 2:1 applied ozone dose to DOC ratio, comparable to ozone doses used in drinking water treatment plants, caused UV254 reductions of 42, 51 and 56% for Tod, Hope Valley and Myponga raw water respectively. Specific UV absorbance (SUVA) is defined as the ratio of UV absorbance at 254nm to the DOC (in mg/L) and provides an insight into the relative number of unsaturated chemical bonds and the degree of aromaticity of NOM in specific waters. SUVA values for the raw waters at each applied ozone dose are listed in Table 5.1. Higher SUVA values indicate the presence of humic material. The declining SUVA values following ozonation confirms the alteration of the structure and functional group content of NOM. Overall loss of dissolved organic carbon (DOC) in waters following ozonation is difficult to predict. The loss of DOC can be attributed to organic carbon oxidation producing either volatile organic compounds or mineralisation to CO2. Miltner and coworkers reported a less than 10% DOC removal over the range of ozone doses examined yet a significant removal of UV254 absorbing compounds was observed [75]. Based on literature, it appears that DOC levels do not significantly change due to ozonation [47,75,89]. The DOC loss observed in this study is much greater than most results reported in the literature. It is possible that the greater loss of DOC is chiefly a function of the experimental ozonation conditions used and not the character of NOM. Figure 5.1 highlights the impact of ozonation on DOC levels for all three surface waters. The greatest effect of organic carbon reduction by ozonation was observed with Myponga water. A linear relationship was apparent with increasing ozone doses. At the maximum applied ozone dose of 5:1 (O3:DOC) approximately a 30% reduction in organic carbon content was observed. Adequate DOC removal was also observed for Hope Valley water with a greater than 20% reduction of organic carbon content at the maximum applied ozone dose. The DOC removal over the ozone dose range however differed between Myponga and Hope Valley water. A substantial portion of the DOC loss in Hope Valley water was observed to occur at the lowest applied ozone dose. This may be due, in part, to the low SUVA value of this water. In Myponga water, ozone is first consumed during the oxidation of humic substances. Because the Hope Valley water is lower in humic substances (as evidenced by its low SUVA value) ozonation is most likely attacking portions of the non-humic substances and may be resulting in mineralisation of some of those portions of the NOM. The dissimilar trend in DOC removal by ozonation is indicative of the differences in the NOM character of the waters. Negligible DOC reduction was observed for Tod water. A reason for this is that ozone may be preferentially reacting with inorganic species present in the water. Metal ions existing in their lower
39 CRC for Water Quality and Treatment Research Report 9-2001 oxidation states are readily oxidised by ozone. In addition, many divalent inorganic anions such as carbonates, sulphates and phosphates act as radical scavengers and subsequently terminate radical chain reactions brought about by the decomposition of ozone in water. Tod water hence highlights the impact of high levels of inorganic species on pre-ozonation.
Table 5.1: Effect of ozone dosage on SUVA values for each surface water. SUVA (L/mg.cm x10-2) Ozone Dose Myponga Hope Valley Tod (O3:DOC) Reservoir Reservoir Reservoir Raw 3.70 2.54 5.44 0.5:1 3.05 2.42 5.13 1:1 2.61 1.86 4.32 2:1 1.80 1.46 3.45 3:1 1.40 1.26 2.82 5:1 1.14 1.04 2.71
100
90 UV Reduction 80 Myponga Hope Valley 70 Tod 60
50
40 DOC Reduction 30
20
Absorbance and DOC Reduction (%) Reduction DOC and Absorbance 10 254
UV 0 0:1 1:1 2:1 3:1 4:1 5:1 Applied Ozone Dose to DOC Ratio
Figure 5.1: UV254 and DOC Reductions with Increasing Ozone Dosage.
5.2 Effects of Ozone on Aldehyde Formation Aldehyde formation was determined for each source water at each applied ozone dose. Levels of propionaldehyde, butyraldehyde, heptaldehyde and benzaldehyde on most occasions were not detected and if formed were present at insignificant concentrations close to their limit of detection. For this reason these four aldehydes are not discussed herein, as their negligible levels do not pose a threat to the quality of drinking water. Generally, increasing ozone dose increased aldehyde levels in each of the raw waters. The results are displayed in Figures 5.2 – 5.5 and expressed as micrograms per milligram of DOC content. The raw data can be found in Appendix IV.
40 CRC for Water Quality and Treatment Research Report 9-2001
Formaldehyde was found to be the predominant aldehyde at the higher ozone doses for the three water waters. At the higher ozone doses formaldehyde levels in Myponga water were approximately three times greater than glyoxal levels, the second most prevalent aldehyde formed. Interestingly, the trend in formation for each aldehyde species differs considerably suggesting that formation is dependent on a number of factors. Myponga water for instance does not form any appreciable formaldehyde until an ozone dose to DOC ratio of 2:1 is applied. At this dose the level of formaldehyde is increased by twenty fold and subsequent increases in ozone linearly increases formaldehyde levels. Acetaldehyde follows a similar trend in Myponga water. That is, following an ozone dose of 2:1 acetaldehyde suddenly begins to form at sizeable levels however, unlike formaldehyde further increases in ozone dose does not produce much higher acetaldehyde levels. Formation of glyoxal and methyl glyoxal in Myponga water increases linearly with ozone dose. By-product formation is different in Hope Valley water. Compared with Myponga water, formaldehyde levels increase linearly with ozone dose. Glyoxal and methyl glyoxal levels also increase linearly with ozone dose, which is the same as observed with Myponga water. Acetaldehyde formation in Hope Valley water is similar to formation in Myponga water. That is, low levels are formed up to a certain dose, then as the dose increases there is a significant increase but higher doses do not increase the concentration further. With Tod Reservoir water, all aldehydes form steadily as the ozone dose increases but at doses of 2:1 and higher, formation has levelled off. Although formaldehyde and acetaldehyde are not formed at appreciable levels at the low ozone doses, the rate of UV absorbance reduction at 254nm is substantial at these doses. Ozone is thus destroying the conjugated π-electron system of the NOM by cleaving bonds at sites of unsaturation. These newly formed lower molecular weight compounds could be possible precursors to formaldehyde and acetaldehyde which is why these low molecular weight aldehydes are not observed to form at the lower ozone doses. Glyoxal and methyl glyoxal do not follow the same trend as formaldehyde and acetaldehyde. As can be seen by Figures 5.4 and 5.5 these ozonation by-products appear to form immediately upon ozonation and increase linearly with increasing ozone dose. This suggests the formation of these dialdehydes originate from alternative precursors.
20 Formaldehyde 18
16 )
C
14 O
n
D
io
t 12 g a
m
m
r
r
10 e
o
f p
e 8 e
d d
y y
h 6 h
e e
d d
l 4 l a A
g
2 µ (
0 Tod 5:1 3:1 Hope Valley 2:1 Ap 1:1 plied Myponga Ozone 0.5:1 Dose 0:1 to DO C Ratio
Figure 5.2: Formaldehyde Levels with Increasing Ozone Dose.
41 CRC for Water Quality and Treatment Research Report 9-2001
8 Acetaldehyde 7
) 6
OC
n
D
o
i
t 5 g
a
m
m
r
r
e 4 o
p
F
e
e
d 3 d y
y
h
h
e
e
d
d
l
2 l
a
A
g
1 µ (
0 Tod Hope Valley 5:1 3:1 2:1 Appl 1:1 Myponga ied Ozo ne Dos 0.5:1 e to DO 0:1 C Ratio
Figure 5.3: Acetaldehyde Levels with Increasing Ozone Dose.
7
Glyoxal 6
)
C
O
5 n
D
o
i
t
g
a
4 m
m
r r
e o
p
F
3 e
e
d
d
y
y
h
h
2 e e
d
d
l
l
a
A
1 g
µ
( 0 Tod 5:1 3:1 Hope Valley 2:1 Appli ed Oz 1:1 one D 0.5:1 ose to Myponga DOC 0:1 Ratio
Figure 5.4: Glyoxal Levels with Increasing Ozone Dose.
42 CRC for Water Quality and Treatment Research Report 9-2001
6
)
C
Methyl glyoxal
5 O
n
D
o
i
t
g
a
4 m
m
r
r
e
o
p
F
e
3 e
d
d
y
y
h
h
e 2 e d
d
l
l
a
A
g
1 µ (
0 Tod Hope Valley
5:1 3:1 Myponga 2:1 1:1 A 0.5:1 pplied Ozo 0:1 ne dose to DOC Ratio
Figure 5.5: Methylglyoxal Levels with Increasing Ozone Dose.
Miltner and coworkers [75] in a pilot scale study on river water with a DOC level of 1.4 mg/L reported no significant increases in acetaldehyde yields beyond a transferred O3 to DOC ratio of 0.5mg/mg. Furthermore they observed formaldehyde, methyl glyoxal and glyoxal had not reached maximum yields at the highest transferred O3 to DOC ratio applied of 2.8mg/mg. Their results are consistent with the findings observed in this study using Myponga and Hope Valley water. Sclimenti and coworkers however observed that methyl glyoxal, and not formaldehyde, was the major ozonation by- product [71]. Their findings showed that the level of methyl glyoxal was twice that of formaldehyde and close to four times the amount of glyoxal formed. It is important to note that generalisations may be difficult to make as the pH of standards and samples were not controlled in the work reported by Sclimenti and coworkers. Controlling the pH for both standards and samples has been shown to be necessary if one is to obtain a level of accuracy in the quantitation of aldehydes [83]. Further work reported in the literature by Schmidt and coworkers showed a steady increase in glyoxal and methyl glyoxal with increasing O3 to DOC ratios of 0.5, 1, 2 and 3 mg/mg DOC [77]. Interestingly, formaldehyde levels reached a maximum at 1:1 and dramatically decreased at the higher ozone doses. This was not the case in this study. It is evident that the character of the water and nature of the DOC plays an enormous part in the levels and relative ratios of aldehydes produced following ozonation. This is illustrated by the results obtained by international researchers mentioned above. As this is the first study attempted in Australia, comparisons with other Australian waters is unfortunately not possible.
5.3 Effects of Chlorine On Aldehyde Formation Chlorination studies were conducted on each ozonated water and the levels of aldehydes were determined. This was to investigate the effect of chlorination on the levels of aldehyde species. Figure 5.6 displays the aldehyde levels following ozonation and chlorination of Myponga raw water. Similar trends were observed in all waters.
43 CRC for Water Quality and Treatment Research Report 9-2001
Formaldehyde levels (O3) 220 70 Glyoxal (O ) Formaldehyde levels (O3+Cl2) 3 Glyoxal (O +Cl ) Acetaldehyde levels (O3) 3 2 200 Methyl glyoxal (O ) Acetaldehyde levels (O3+Cl2) 3 60 Methyl glyoxal (O +Cl ) 180 3 2
160 50 140 40 120
100 30 80
60 20 Aldehyde levels in ug/L. Aldehyde
40 10 20
0 0 0:1 0.5:1 1:1 2:1 3:1 5:1 0:1 0.5:1 1:1 2:1 3:1 5:1 Applied ozone dose to DOC Ratio
Figure 5.6: Effects of Chlorination Following Ozone Treatment on Aldehyde Formation. (Myponga Reservoir)
Apart from the chlorinated DBPs described earlier in Section 1.3.1, aqueous chlorine reacts with NOM in waters to form formaldehyde and acetaldehyde. Levels of 33µg/L and 23µg/L of formaldehyde and acetaldehyde respectively were formed by chlorine in the unozonated Myponga raw water. Similar increases in formaldehyde and acetaldehyde levels were observed for each ozonated water following chlorination. This suggests that chlorine is oxidising certain organic species unaltered by ozone, to form formaldehyde and acetaldehyde. (Unaltered by ozone as ozone may have already reacted due to its high reactivity and fast reaction rates). Interestingly, the 1:1 applied O3 to DOC dosed sample did not follow the same trend with a substantial formation of formaldehyde observed following chlorination of this sample. Figure 5.6 shows a dramatic jump in formaldehyde formation from applied ozone doses of 1:1 to 2:1. A possible explanation for the significant increase in formaldehyde levels following chlorination of the 1:1 ozone dosed sample is that formaldehyde precursors were produced by ozone at an applied dose of 1:1. And any further increase in oxidant (chlorine) concentration appears to have the same effect as what is observed for the 2:1 O3 to DOC treated sample. The effects that are observed with methyl glyoxal for each ozonated water are of particular significance. Methyl glyoxal almost completely disappears following chlorination, a strong indication that methyl glyoxal is reacting with chlorine to form a chlorinated by-product. If it were being removed by oxidation then ozone should show a similar effect, ie. concentrations should be diminishing at the higher ozone doses. The results for glyoxal are more difficult to interpret. Glyoxal seems to form upon chlorination of raw water however, it too is shown to react with chlorine and possibly form a chlorinated product. Similar effects have been observed by other authors on waters, which have a different character [111,112].
44 CRC for Water Quality and Treatment Research Report 9-2001
5.4 Effects of Ozone on Ketoacid Formation Various ketoacids are shown to form from the ozonation of NOM in waters. Investigating the formation of ketoacids in water treatment is essential as their presence pose a threat for a number of reasons. Glyoxylic acid is a known mutagen [90], pyruvic acid is claimed to be an efficient THM precursor [57] and ketoacids are biodegradable consequently increasing the assimilable organic carbon present and contributing to bacterial regrowth [74,87]. Limited information exists in the literature on ketoacid formation. Work that has been published does not necessarily follow any obvious trend. Moreover, the methodology that has been used does not include a pH buffer to ensure the conditions of derivatisation remain the same for both standards and samples. The lack of a common method by all researchers leaves some doubt as to whether the ketoacid levels obtained are correct or have been influenced by the derivatisation conditions. Ketoacid levels were determined at each ozone dose for Myponga, Hope Valley and Tod raw waters and are graphically represented in figure 5.7. Similar ketoacid profiles are observed for both Myponga and Hope Valley water at all ozone doses apart from ketomalonic acid which is formed at approximately double the concentration. In addition, ketoacid levels in these waters increased linearly with increasing ozone dosage. Levels of ketoacids far exceed the levels of aldehydes that are formed by ozonation in these two waters. Figure 5.8 illustrates this by comparing the sum of ketoacids versus the sum of aldehydes at each ozone dose. Tod raw water was atypical in that ketoacid formation increased and then decreased with increasing ozone dose. Ketoacid formation was found to be 4-6 times greater than aldehyde formation at the lower applied ozone doses in Myponga water. Xie and Reckhow reported similar findings in that ketoacid levels far exceeded aldehyde levels [68]. They also observed that at lower ozone doses glyoxylic acid formation was greater than pyruvic acid and ketomalonic acid formation with pyruvic acid and ketomalonic acid levels being similar at these doses. In this study, glyoxylic and pyruvic acid were formed at similar levels in both Myponga and Hope Valley waters. Xie and Reckhow reported that at doses higher than 3:1, ketomalonic acid was the major by-product followed by glyoxylic acid and pyruvic acid being the minor by-product. Markedly opposite effects were observed with Myponga water at the higher ozone doses. At an applied O3 to DOC dose of 3:1, pyruvic acid had a 1.5 times greater yield than glyoxylic acid and ketomalonic acid was least formed. The yield of ketomalonic acid being only 20% of that of pyruvic acid.
45 CRC for Water Quality and Treatment Research Report 9-2001
50 Glyoxylic Acid 25 Pyruvic Acid 40
20
30 Myponga 15 Hope Valley 20 Tod 10
10 5
0 0 0:1 0.5:1 1:1 2:1 3:1 5:1 0:1 0.5:1 1:1 2:1 3:1 5:1
1.2 16 2-Ketobutyric Acid Ketomalonic Acid 1.0 14
12 0.8 10
0.6 8
Ketoacid Formation (ug ketoacid per mg DOC) 6 0.4 4 0.2 2
0.0 0 0:1 0.5:1 1:1 2:1 3:1 5:1 0:1 0.5:1 1:1 2:1 3:1 5:1 Applied Ozone Dose to DOC Ratio
Figure 5.7: Ketoacid Formation in Different Waters with Increasing Ozone Dose.
Krasner [59] and Griffini and coworkers [74] had also observed significantly greater levels of ketoacids than aldehydes. Andrews and Huck however found the levels of aldehydes to be much greater than ketoacids following ozonation of lake and river water [84], while this study found ketoacids to be the major product, except for Tod Reservoir water at the highest dose. Other authors have reported that glyoxylic and pyruvic acid reach a maximum and decrease following higher ozonation doses. Garcia-Araya and coworkers [67] report glyoxylic acid to be the major by- product at doses less than 2:1 but higher ozone doses causes a dramatic reduction in glyoxylic acid levels. The reduction in formation was put down to the acid’s high reactivity towards ozone. Schmidt and coworkers published similar results reporting that glyoxylic and pyruvic acid pass a maximum and at higher doses ketomalonic acid is the major ozonation by-product [77]. The ozonation of Tod water is observed to follow these trends. In Tod water, the ketoacids were observed to pass a maximum at an applied O3 to DOC ratio of 2:1 with levels significantly dropping at the higher ozone doses. However the other waters showed no evidence of decline with increasing ozone dose. The ketoacid levels from these water sources continue to rise with increasing ozone dose. Overall these results indicate that the water chemistry and/or nature of the organic material markedly influences by-product formation.
46 CRC for Water Quality and Treatment Research Report 9-2001
Myponga Raw Water Hope Valley Raw Water Tod Raw Water 800 400 200 Total aldehydes 700 Total ketoacids 180 350
g/L) 160 µ 600 300 140 500 250 120
400 200 100
80 300 150 60 200 100 40 100 50 20 Aldehyde & Ketoacid ( Levels Aldehyde
0 0 0 0.5:1 1:1 2:1 3:1 5:1 0.5:1 1:1 2:1 3:1 5:1 0.5:1 1:1 2:1 3:1 5:1 Applied Ozone Dose to DOC Ratio
Figure 5.8: Comparison of Total Aldehyde and Ketoacid Formation.
5.5 Effects of Chlorine on Ketoacid Formation It was shown in Section 5.3 that chlorination of raw and ozonated waters form significant levels of formaldehyde and acetaldehyde. Surprisingly, ozonation by-products such as ketoacids do not form following chlorination. Furthermore, all these ketoacids are completely destroyed by the action of chlorine. An applied O3 to DOC dose of 5:1 for Myponga water produced a sizeable pyruvic acid yield of 210 µg/L. Following chlorination, the level of pyruvic acid present in the water sample dropped to less than 5 µg/L. Similar effects were observed for all the ketoacids and in all three water sources. Figure 5.9 illustrates the destruction of ketoacids with chlorine in Myponga water at all applied ozone concentrations. As indicated previously, the loss is unlikely to be due to oxidation as levels should be reduced at higher ozone doses. The loss then appears to be due to the formation of chlorinated products.
47 CRC for Water Quality and Treatment Research Report 9-2001
60 220 Glyoxylic Acid (O3) Glyoxylic Acid (O +Cl ) 2-Ketobutyric Acid (O3) 200 3 2 2-Ketobutyric Acid (O3+Cl2) Pyruvic Acid (O3) 50 Ketomalonic Acid (O3) 180 Pyruvic Acid (O3+Cl2) Ketomalonic Acid (O3+Cl2) 160 40 140
120 30 100
80 20 60 Ketoacid Levels (ug/L) Levels Ketoacid
40 10 20
0 0
0:1 0.5:1 1:1 2:1 3:1 5:1 0:1 0.5:1 1:1 2:1 3:1 5:1 Applied Ozone Dose to DOC Ratio
Figure 5.9: Effects of Chlorination and its Destruction of Ketoacids.
5.6 Effects of Ozone on Carboxylic Acid Formation. Carboxylic acids are another class of ozonation by-products. Carboxylic acids are an important group of ozonation by-products as they have been shown to contribute to the biodegradable organic matter [91]. The determination of formic, acetic, propionic and benzoic acid were unfortunately limited to ozonated Myponga and Hope Valley waters. The biphasic derivatisation was extremely susceptible towards moderate to high levels of salinity and consequently re-analysis of the ozonated samples following either chlorination or bacterial regrowth assays proved unsuccessful. The additions of a buffer and a free chlorine quenching agent for chlorination studies and an inorganic nutrient solution for bacterial regrowth assays naturally increased the ionic strength of the water sample and consequently carboxylic acid derivatisation was unsuccessful. Furthermore, carboxylic acid analysis using this method was also unsuccessful for Tod water due to its high salinity. Carboxylic acid levels for Myponga and Hope Valley raw water at each ozone dose are displayed in Figure 5.10. It can be seen that formic acid formation in Myponga water follows a similar trend to formaldehyde formation indicating ozone is further oxidising the aldehyde. Formic acid levels however do not seem to be increasing at the higher ozone doses. It is likely that at these higher ozone doses formic acid is being oxidised by ozone to form carbon dioxide, which may account for the loss of DOC observed at the higher ozone doses. Similarly, a dramatic jump in acetic acid formation at the higher applied ozone doses was also observed. Again, this mimicked the trend seen with acetaldehyde formation. Acetic acid levels continued to significantly increase at the maximum applied ozone dose of 5:1 in Myponga water.
48 CRC for Water Quality and Treatment Research Report 9-2001
Myponga Raw Water 70 Formic Acid 60 Acetic Acid 50 Propionic Acid 40 Benzoic Acid
10
0
Hope Valley Raw Water 65
60 40
g AcidCarboxylic per mg DOC 30 µ 20 10 0 0:1 0.5:1 1:1 2:1 3:1 5:1 Applied Ozone Dose to DOC Ratio
Figure 5.10: Influence of Ozone Dosage on Carboxylic Acid Formation.
Levels and trends of formic acid formation in Hope Valley water were different to what was observed with ozonated Myponga water. These differences are further evidence that the NOM in both waters are quite distinct. Formic acid levels in Hope Valley water increased at low ozone doses analogous to the steady formaldehyde increases observed with this water. Higher ozone doses however had no effect on formic acid levels. The lack of formic acid formation at the higher ozone doses observed in both waters, may be attributable to carbon dioxide production. The lower level of formic acid formation in Hope Valley water typifies the weaker impact ozone has on the reduction of DOC in this water. Acetic acid however is observed to steadily increase with ozone dose. At the highest applied ozone dosage, acetic acid levels in Hope Valley water are comparable to levels detected in ozonated Myponga water. In both ozonated waters, benzoic acid was not produced and a trace level of propionic acid was detected in both waters at the higher ozone concentrations. The substantial quantities of formic and acetic acid produced with ozonation are consistent with the work conducted by others who have reported the acidity of water containing NOM to increase following ozonation [47,49]. Oxalic acid, a dicarboxylic acid, has been shown to also form in vast amounts following ozone treatment [47].
5.7 Impact of Ozone on Biodegradable Organic Matter Myponga raw water and selected ozonated aliquots were sterile filtered through a 0.2 micron polycarbonate membrane filter and inoculated with indigenous bacteria in order to determine their potential in supporting a bacterial regrowth. The methodology used was based on the current procedure routinely used at the Australian Water Quality Centre [35]. The results of the bioassay are listed in Table 5.2 and include the growth rate, the growth factor and the acetate carbon equivalent (ACE). All of the oxidised samples showed a higher growth factor than raw water and the BRP increased with higher applied ozone doses. The increase in BRP is due to the oxidative destruction of high molecular weight substances by ozone to for low molecular, biodegradable oxygenated compounds such as aldehydes, ketoacids and carboxylic acids.
49 CRC for Water Quality and Treatment Research Report 9-2001
Table 5.2: Bacterial Regrowth Potential Data for Myponga Water.
O3:DOC Growth ACE Growth rate factor (f) (µg/L) (µ, hr-1) 0:1 4.4 69 0.163 0.5:1 13.9 257 0.235 2:1 45.2 884 0.192 5:1 67.7 1334 0.236
Measurements to determine the ability of a water type to support bacterial growth are currently accomplished using biological assays. These assays are intrinsically slow and analysis time generally spans several days due to their dependence on the rate of bacterial growth. For this reason, the analysis of specific compounds to be used as alternative surrogates for estimating bacterial regrowth in water samples would be highly beneficial. Ozonation by-products such as aldehydes and ketoacids are shown to have a direct relationship with increases in biodegradable matter. Table 5.3 lists the correlation coefficients for linear regression of each aldehyde and ketoacid. These compounds were observed to significantly increase with increasing ozone doses. A strong correlation with individual aldehydes and ketoacids also means a strong correlation with total aldehydes and total ketoacids as well as total aldehydes plus total ketoacids as illustrated in Figure 5.11. A strong relationship exists between aldehydes, ketoacids and the bacterial regrowth potentials in ozonated Myponga water. This clearly suggests that aldehyde and ketoacid species are components of biodegradable organic matter (BOM). Therefore, estimating increases in BOM following ozonation can be achieved by determining the aldehyde and/or ketoacid levels relative to the unozonated sample. The results in this study agree well with the findings reported by Miltner and coworkers [75] and Van de Kooij and coworkers [73] who independently reported the similarity in behaviour of ozonation by-products and assimilable organic carbon (AOC). Both authors had used a culture of Pseudomonas fluorescens strain P17 and Spirillum strain NOX as their inoculum for AOC studies as opposed to indigenous microorganisms isolated from the water source as was the case in this work.
Table 5.3: Correlation Coefficients for Linear Regression of Ozonation By-products Versus BRP for Myponga Water. Ozonation By-product Correlation coefficient Formaldehyde 0.994 Acetaldehyde 0.973 Glyoxal 0.997 Methyl glyoxal 0.991 Pyruvic acid 0.960 Glyoxylic acid 0.987 Ketomalonic acid 0.996
50 CRC for Water Quality and Treatment Research Report 9-2001
300 600
250 500
200 g/L) g/L) µ µ 400
150 300
100 200 Total Ketoacids ( Total ( aldehydes 100 50
0 0 R=0.980 R=0.994
0 400 800 1200 0 400 800 1200 Acetate carbon equivalent (µg/L)
Figure 5.11: Relationship Between Total Aldehydes, Ketoacids and BRP in Myponga Water.
All inoculated samples were re-analysed for aldehydes and ketoacids following the regrowth process. This was to determine whether any changes in the levels of these compounds were evident. Substantial decreases in by-product concentrations were found for all compounds except glyoxal which showed an 8% reduction and methylglyoxal which showed a 54% reduction. This finding further confirms that these ozonation by-products apart from glyoxal and to some extent methylglyoxal are components of AOC and a rich nutrient source at that. Figures 5.12-5.18 compares the pre- and post- BRP levels for the common aldehydes and ketoacids monitored in this study. Overall, the ketoacids were found to be a far superior nutrient source than the aldehydes. Levels of glyoxylic, pyruvic and ketomalonic acids declined by 97%, 96% and 93% respectively during the BRP assay. Formaldehyde and acetaldehyde were also easily biodegradable with reductions of 86% and 74% respectively. Dialdehydes were observed to be less biodegradable with reductions of 8% for glyoxal and 54% for methylglyoxal.
51 CRC for Water Quality and Treatment Research Report 9-2001
160 )
L 0
/ 14
g
µ (
120 n
o
i t
a 100
m
r o
F 0
8
e d
y 60
h
e
d l
a 40
m r
o 20 F 0
O3 5 2 :1 O + BRP 0 :1 3 .5 0: : 1 1 io o DOC Rat one Dose t Applied Oz
Figure 5.12: Comparison of Formaldehyde Levels Before and After BRP Assay.
50
)
L
/ g
µ 40
(
n
o
i t
a 30
m
r
o
F
e
d 20
y
h
e
d
l a
t 0
e 1
c A
0
O3
5: O + BRP 2: 1 3 0. 1 0: 5: 1 1 io o DOC Rat one Dose t Applied Oz
Figure 5.13: Comparison of Acetaldehyde Levels Before and After BRP Assay.
52 CRC for Water Quality and Treatment Research Report 9-2001
50
)
L /
g 40
µ
(
n
o i
t 30
ma
r
o
F
l 20
a
x
o
y l
G 10
0
O3
5: O + BRP 2: 1 3 0. 1 0 5: :1 1 o o DOC Rati one Dose t Applied Oz
Figure 5.14: Comparison of Glyoxal Levels Before and After BRP Assay.
40
)
L
/
g
µ
(
n
o 30 i
t
ma
r
o
F
l 20
a
x
o
y
l
g l y 0
h 1
t Me
0
O3 5 2 :1 O + BRP 0 :1 3 0 .5 :1 :1 Ratio one to DOC Applied Oz
Figure 5.15: Comparison of Methylglyoxal Levels Before and After BRP Assay.
53 CRC for Water Quality and Treatment Research Report 9-2001
300
) L
/ 250
g
µ
(
n o
i 200
t
ma r
o 150
F
d
i
c a
100
c
i
v
u r
y 50 P
0
O3
5: O + BRP 2: 1 3 0. 1 0 5: :1 1 o o DOC Rati one Dose t Applied Oz
Figure 5.16: Comparison of Pyruvic acid Levels Before and After BRP Assay.
140
)
L /
g 120
µ
(
n
o i
t 100
ma r
o 80
F
d i
c 0
a 6
c
i
n o
l 40
ma o
t 20
e K 0
O3
5: O + BRP 2: 1 3 0. 1 0 5: :1 1 o o DOC Rati one Dose t Applied Oz
Figure 5.17: Comparison of Ketomalonic Acid Levels Before and After BRP Assay.
54 CRC for Water Quality and Treatment Research Report 9-2001
200
) 180
L /
g 0
µ 16
(
n
o 140
i t
120
ma
r o
F 100
d i
c 80
a
c i
l 60
y x
o 40
y l
G 20 0
O3
5: O + BRP 2: 1 3 0. 1 0 5: :1 1 o o DOC Rati one Dose t Applied Oz
Figure 5.18: Comparison of Glyoxylic Acid Levels Before and After BRP Assay.
55 CRC for Water Quality and Treatment Research Report 9-2001
6 Effects of Ozonation on Chlorinated Disinfection By-Products
All ozonated waters were subsequently chlorinated and the effects of ozone were evaluated by measuring the formation potentials of trihalomethanes (THMs), chloroacetic acids (CAAs), chloroketones, haloacetonitriles, chloral hydrate, chloropicrin and adsorbable organic halides (AOX). Chlorination was performed under extreme conditions in order to achieve a substantial production of disinfection by-products (DBPs). Raw water from each source was subjected to the same chlorination conditions without pre-ozonation in order to provide a control. The conditions were standardised by keeping variables such as temperature, pH and incubation time constant and the chlorine dose in excess. Maintaining such standard conditions enabled the direct comparison of the effects of ozone dosage on chlorinated DBP formation. The chlorination conditions used are as specified earlier in Section 4.1.
6.1 Effects Of Ozone On Trihalomethane Formation It has well been established that aqueous chlorine reacts with natural organic matter (NOM) in waters to form THMs [9,19]. The four major THMs formed in chlorinated waters are chloroform (CHCl3), bromodichloromethane (CHCl2Br), chlorodibromomethane (CHClBr2) and bromoform (CHBr3). Factors affecting THM formation include NOM concentration, pH, bromide ion levels, temperature and chlorine dose [31]. The chemical nature of NOM, and in particular the presence of inorganic species such as bromide play a significant role in THM formation and speciation. This contribution is complex as shown by THM formation in the controls. The levels of each THM at each applied ozone dose are presented in Figures 6.1 and 6.2. The levels are expressed in molar amounts and not by weight as bromoform is considerably heavier than chloroform. Chloroform levels differ markedly between all three controls. Chloroform is the predominant THM formed upon the chlorination of Myponga raw water and generally the major THM present in chlorinated waters containing low levels of bromide [31,32]. Chloroform is shown to contribute 74% towards the total THMs formed in Myponga water. The bromide level in Hope Valley water, listed in Table 2.1, is notably less, yet chloroform contributes only 54% towards total THMs. Furthermore total THM levels for Myponga and Hope Valley controls were 0.42 and 0.27 µmol/mg DOC respectively. These differences illustrate the dependence of THM formation on source water characteristics. The greater degree of unsaturation in the organics present in Myponga water as indicated by the higher SUVA value may play a vital role in the effects observed between these two surface waters. The DOC content of Tod water is approximately three times less than Myponga water. Total THM formation of Tod water, without prior ozonation, was found to be 1.5 times greater. In terms of total THM formation in micromoles per unit DOC, Tod water produced a sizeable 2.6 times the level of total THMs compared to Myponga water. The inorganic content in Tod water is seen to play a vital role in the levels and types of THMs that form. Salinity and bromide levels in Tod water are very high and approximately ten times the levels found in Myponga and Hope Valley water. Alkalinity is also notably higher. Figure 6.3 compares the distribution of individual THM formation for each control following chlorination.
56 CRC for Water Quality and Treatment Research Report 9-2001
0.12 0.32 Chloroform Bromodichloromethane 0.11 0.28 Myponga 0.10 Hope Valley 0.24 Tod 0.09
mol/mg DOC) mol/mg 0.20 µ 0.08 0.16 0.07 0.12 0.06 0.08 0.05 0.04
Trihalomethane Levels ( Trihalomethane 0.04 0.00 0:1 1:1 2:1 3:1 4:1 5:1 0:1 1:1 2:1 3:1 4:1 5:1 Applied Ozone Dose to DOC Ratio
Figure 6.1: Influence of Pre-ozonation on Chloroform and Bromodichloromethane Formation Potentials
0.50 0.55 Chlorodibromomethane Bromoform 0.45 0.50 0.45 0.40 0.40 Myponga 0.35 Hope Valley 0.35 mol/mg DOC) mol/mg Tod µ 0.30 0.30
0.25 0.25
0.20 0.015 0.15 0.010 0.10 0.005 Trihalomethane Levels ( Trihalomethane 0.05
0.00 0.000 0:1 1:1 2:1 3:1 4:1 5:1 0:1 1:1 2:1 3:1 4:1 5:1 Applied Ozone Dose to DOC Ratio
Figure 6.2: Influence of Pre-ozonation on Chlorodibromomethane and Bromoform Formation Potentials.
57 CRC for Water Quality and Treatment Research Report 9-2001
Myponga control
4% 0% 22%
74%
Tod control
1% 10% 47%
42%
Hope Valley control
13% 1%
54% 32%
CHCl3 CHCl2Br CHClBr2 CHBr3
Figure 6.3: Distribution of THM Species for Each Raw Water Control.
The very high contribution of bromide in Tod water influences the production of THMs by forming bromine-containing species. As has been shown by other authors, bromide ions are readily oxidised by aqueous chlorine to hypobromous acid, which is a more effective halogen-substituting agent than its chlorine counterpart [97,98]. The presence of high levels of bromide as found in Tod water not only influences THM speciation but also markedly enhances the total THMs formed.
58 CRC for Water Quality and Treatment Research Report 9-2001
Pre-ozonation of raw water is observed to decrease the total THM levels formed following chlorination. Ozone reacts and oxidises NOM forming oxygenated lower molecular weight products that are less reactive towards substitution reactions with chlorine. Surprisingly though, individual THMs behave differently with exposure of the NOM to ozone. At low ozone doses, consistent with operating conditions in treatment plants, chloroform levels steadily decrease as the ozone dose is increased. This is seen in both Myponga and Hope Valley as shown in Figure 6.1. Higher ozone doses however have a smaller influence on chloroform reduction. Chloroform formation per unit DOC appears to plateau at the higher ozone doses. As bromide levels are relatively low in both Myponga and Hope Valley waters, the formation of di- and tri-brominated THMs are small compared with chloroform and subsequently the THM formation is largely represented by chloroform. It is important to note the negligible formation of chloroform in Tod water. Chloroform levels in this water were found to be less than 1% of the total THMs produced for all ozone doses and including the control. Water aliquots for chlorination experiments are dosed with excess chlorine at levels of 20mg/L Cl2. This level far exceeds the bromide concentration in Tod water on a molar basis yet brominated THMs are the predominant by-products formed. This is due to the greater substituting power of bromine. Chlorine can act by oxidation and substitution, with the latter responsible for the formation of halogenated DBPs. Kinetic studies have shown that the rate of bromine substitution is faster than chlorine substitution [92].
Inspection of CHCl2Br, CHClBr2 and CHBr3 for any trends associated with increasing ozone dosage are far more complex and difficult to interpret. Appreciable levels of CHCl2Br were produced for all water controls yet markedly different effects were observed when each water was pre-ozonated. A decline in CHCl2Br was observed for all waters as ozone dose increased. The decline was more pronounced with Tod, and with Myponga the decline was quite small. Tod appeared to level out at applied ozone doses of greater than 2:1.
The profile for CHClBr2 formation across the full ozone dose range in Tod water was similar to CHCl2Br. Significant reductions were achieved at the lower ozone doses and once again levelling out at applied ozone doses of greater than 2:1. CHClBr2 levels in Tod water were approximately five times greater than CHCl2Br throughout the whole dose range.
Levels of CHBr3 and CHClBr2 in chlorinated Myponga and Hope Valley waters were low indicative of waters low in bromide. Both waters exhibited similar trends in formation towards these two THMs. At the low ozone doses an increase in formation was observed for CHBr3 and CHClBr2 changing to a decrease in formation at the higher doses. CHBr3 was the predominant THM in Tod water following chlorination. This was expected due to the water’s very high bromide content. CHBr3 formation was observed to decrease with increasing ozone doses of up to 2:1. For example, an applied ozone dose of 2:1 resulted in a 39% reduction of CHBr3 relative to the control. However further increases in ozone concentration resulted in a steady increase in CHBr3 formation.
6.2 Effects of Ozone on Chloroacetic Acid Formation Potentials Chloroacetic acids (CAAs) were first reported in chlorinated waters in the 1980s [24]. Trichloroacetic acid (TCAA) is shown to cause liver tumours in mice [27] and hence their presence in finished drinking water is of significance due to their potential health effects. TCAA and DCAA were formed at significant levels in the Myponga and Hope Valley controls. TCAA formation in the Myponga control was approximately half the concentration of chloroform whereas the Hope Valley control, levels of TCAA surpassed the chloroform levels indicating characteristic differences between the NOM of each water type. Figures 6.4-6.6 presents the chloroacetic acid levels expressed as micrograms per mg DOC for each ozone concentration and Figure 6.7 illustrates the decline in total CAA formation for both Hope Valley and Myponga water at increasing ozone doses. Actual chloroacetic acid levels for each water source can be found in Appendix V. The decline in total CAA formation is shown to be due to the destruction of TCAA precursors by preozonation. At the maximum applied ozone dose of 5:1, Myponga water showed an 82% reduction in TCAA and a 59% reduction in CHCl3 clearly highlighting that TCAA precursors are more susceptible to ozone attack than the corresponding THM precursors. The reduction of TCAA with ozone is very similar for both Hope Valley and Myponga water suggesting common precursors are present in both sources.
59 CRC for Water Quality and Treatment Research Report 9-2001
22
20 n
o Trichloroacetic acid i t
a 18 )
C
m
r
16 O o
D
F
g 14 d
i
c m
12 A r
e
c
i
p
t 10 e
g
c
µ
a
8 n
o i
r (
o 6 l h 4 C
Myponga
Hope Valley 0:1 1:1 2:1 3:1 4:1 Ap 5:1 plied Ozon e Dose to DOC Ratio
Figure 6.4: Influence of Pre-ozonation on Trichloroacetic Acid Formation Potentials.
22
n 20
o
i
t )
a 18
C
m
r
O o
D 16 F
g Dichloroacetic acid d
i m
c 14
r
A
e
c i
p 12
t
e
g
c µ
10
a
n
i
o
( r
o 8
l h
C 6
Myponga
Hope Valley 4:1 5:1 2:1 3:1 0:1 1:1 o DOC Ratio Applied Ozone Dose t
Figure 6.5: Influence of Pre-ozonation on Dichloroacetic Acid Formation Potentials.
60 CRC for Water Quality and Treatment Research Report 9-2001
6.0
n 5.5 o
i Monochloroacetic acid
t
)
a 5.0 C
rm 4.5
O
o
D F
4.0
g
d 3.5
m
ci
r
A
e 3.0
c c
i
p t
2.5
e
g
c
µ
a 2.0
n
o (i
r 1.5
o l
h 1.0 C 0.5 0.0 Myponga
5:1 Hope Valley 3:1 4:1 2:1 o 1:1 OC Rati 0:1 ose to D Ozone D Applied
Figure 6.6: Influence of Pre-ozonation on Monochloroacetic Acid Formation Potentials.
500
450 Myponga 400 Hope Valley
350
300
250
200
150
100 Total Chloroacetic Acid Formation (ug/L) Total ChloroaceticFormation Acid 0:1 0.5:1 1:1 2:1 3:1 5:1 Applied Ozone Dose to DOC Ratio
Figure 6.7: Total Chloroacetic Acid Levels with Increasing Ozone Dosage.
The effect of pre-ozonation on DCAA and MCAA formation however were negligible suggesting that TCAA is formed from distinct precursors. Similar effects of pre-ozonation on TCAA, DCAA and MCAA levels were reported by Westrick and coworkers [94]. The results in this study highlight the dependence of the character of NOM on CAA formation. This is further confirmed by results reported in the literature. Andrews and Huck for instance reported a decrease in CAAs with pre-ozonation of lake water, yet an increase in CAAs was observed following pre-ozonation of a river water and its isolated NOM fractions [95]. Miller and Uden reported comparable levels of TCAA and chloroform
61 CRC for Water Quality and Treatment Research Report 9-2001 were formed on chlorination of aquatic fulvic acids, yet TCAA levels were far greater than chloroform levels with the chlorination of fulvic acids extracted from soil [96]. Surprisingly, Simpson and Hayes reported that MCAA levels were generally much greater than DCAA levels and TCAA was present as the minor species in a range of drinking waters across Australia [33]. Levels of chloroacetic acids in Tod control and respective pre-ozonated sample were low (see Appendix V) indicative of the predominant formation of brominated acids and a mixture of bromo- and chloro- acids as was observed with trihalomethane levels for this water. Due to time constraints the introduction of these halogenated species into the method was not possible. It is well documented that waters high in bromide following chlorination produce predominantly bromine-containing haloacetic acids and the effects of pre-ozonation on haloacetic acid formation is dependent on bromide concentration [92,97-99]. Bromide levels in Tod water are exceptionally high and unfortunately the brominated acetic acids were not investigated.
6.3 Effects of Ozone on Other Chlorinated By-products
6.3.1 Chloroketones Chloroketones are another class of chlorinated by-products investigated in this study. Mutagenicity tests have shown both 1,1-Dichloropropanone (11DCP) and 1,3-Dichloropropanone (13DCP) to be mutagenic. 13DCP also induced skin tumours in mice [29]. Levels of chloroketones, if at all formed, are very low in Australian reticulated supplies. 1,1,1-Trichloropropanone (111TCP) has been detected at concentrations up to 20 µg/L and no data are available for other chloroketones [34]. There are no set Australian guideline values for chloroketones in drinking water due to the inadequate data. 111TCP and 11DCP were found to form at very low levels under the chlorination conditions employed. 13DCP was not detected above the limit of quantitation for all water sources. Levels of 111TCP were greater than 11DCP for all waters and at each ozone dose. The results are graphically illustrated in Figure 6.8.
1,1-dichloropropanone 1.0 Myponga Hope Valley 0.8 Tod 0.6
0.4
0.2
0.0 1,1,1-trichloropropanone 3.5 3.0 2.5 2.0 1.5 1.0
Haloketone levels in ug per mg DOC 0.5 0.0 0:1 0.5:1 1:1 2:1 3:1 5:1 Applied Ozone Dosage to DOC
Figure 6.8: Effects of Pre-ozonation on Chloroketone Formation.
The effects of ozone on chloroketones seems to have an opposite effect to what is observed with the more dominant chlorinated compounds such as CHCl3 and TCAA. Ozonation increases chloroketone precursors in both Myponga and Hope Valley water as 11DCP and 111TCP were found to increase at the higher ozone doses. Myponga and Hope Valley water both showed identical trends in the
62 CRC for Water Quality and Treatment Research Report 9-2001 formation of 111TCP. 11DCP formation was slightly greater for Myponga than Hope Valley water. As is illustrated in Figure 6.8, the increase in 111TCP is linear over the applied dose range. A 2.8 fold increase in 111TCP formation was observed at an applied ozone to DOC ratio of 5:1 relative to the control. Increases in 111TCP upon preozonation has also been reported by Reckhow and Singer [100]. The results for Tod water however are difficult to interpret as a common trend over the dose range is not obvious. Both 11DCP and 111TCP decrease, then increase, with increasing dose, however the changes are quite small. The salinity and high bromide levels in Tod water play a major part in the lack of chloroketone formation and the difficulty in interpreting the effects of ozone. Due to time constraints the brominated analogues of these ketone acetone derivatives were not investigated. Levels of the brominated analogues for Tod water would undoubtedly be high and possibly of health concern.
6.3.2 Chloral hydrate and Chloropicrin Chloral hydrate and chloropicrin are among other chlorinated compounds that may form as by- products during chlorination of raw waters containing NOM. Chloral hydrate formation, unlike 111TCP, differs markedly between Myponga and Hope Valley water. Figure 6.9 highlights this observation. Myponga water forms approximately double the concentration of chloral hydrate per mg of DOC than Hope Valley water consistent with the formation of other by-products. Chloral hydrate is observed to markedly increase upon ozonation up to applied doses of 2:1. Further increases in ozone dosage had no effect on chloral hydrate formation. Similar findings were observed by Smith and coworkers [89] and Andrews and Huck [95]. A trace level of chloral hydrate was found to form in chlorinated Tod water. Preozonation of Tod water did not seem to enhance the small amount of chloral hydrate present in the control.
Chloral Hydrate Chloropicrin 1.50 16
14 1.25
12 1.00 Myponga 10 Hope Valley Tod 0.75 8
6 0.50 (ug per mg DOC) 4 0.25
Chlorinated By-product Formation By-product Chlorinated 2
0 0.00 0:1 0.5:1 1:1 2:1 3:1 5:1 0:1 0.5:1 1:1 2:1 3:1 5:1 Applied Ozone Dose to DOC Ratio
Figure 6.9: Effects of Pre-ozonation on Chloral Hydrate and Chloropicrin Formation.
63 CRC for Water Quality and Treatment Research Report 9-2001
Chloropicrin levels on the other hand were observed to be unaffected by preozonation at the lower ozone doses in Myponga water. Application of an ozone dose to DOC ratio of 2:1 however increased chloropicrin levels in this water by over 100%. Further increases in ozone dose substantially increased chloropicrin levels. Chloropicrin formation increased by a factor of 7 between applied ozone doses of 1:1 and 5:1. An increase in chloropicrin formation upon preozonation was also observed by Miltner and coworkers [75] and Becke and coworkers [101. Surprisingly, the results obtained for Hope Valley water were quite unexpected. Chloropicrin levels were observed to decrease as the ozone dose increased to 1:1 but further increases in ozonation increased levels slightly. Chloropicrin was not detected in the chlorination of Tod raw water. The trend observed for Tod water with THM formation, that is, a high level of brominated analogues and a negligible amount of chloroform formation would more than likely be seen with respect to bromopicrin formation. Once again, due to time constraints bromopicrin was unable to be screened for and hence quantified.
6.3.3 Haloacetonitriles Haloacetonitriles pose concerns due to their potential mutagenic properties. Dichloroacetonitrile (DCAN) and bromochloroacetonitrile (BCAN) are direct-acting mutagens in bacteria and all four haloacetonitriles investigated in this study induced DNA damage in mammalian cells [29]. Trichloroacetonitrile (TCAN) was formed at trace levels in both Myponga and Hope Valley controls and in which, upon preozonation, the levels fell below the limit of detection at applied ozone doses of 1-2:1. TCAN was not detected in the Tod control. DCAN was found to form at higher levels and also formed upon chlorination of Tod water. Similar to TCAN, preozonation reduced the formation of DCAN in both Myponga and Hope Valley water. A greater reduction was seen at the lower ozone doses. Figure 6.10 illustrates the effects of ozonation on haloacetonitrile formation. Levels of BCAN and dibromoacetonitrile (DBAN) were greatest for Tod water. This was expected due to the high level of bromide in Tod water. The impact of ozone at all doses had a negligible effect on DBAN and BCAN formation in Tod water. Bromide levels for Myponga and Hope Valley water were measured to be 0.56 and 0.37 mg/L respectively and surprisingly the levels of BCAN and DBAN were greater in Hope Valley than Myponga water. These findings were consistent with the results obtained with CHClBr2 and CHBr3 in Section 6.1 with these two water sources. The findings suggest that sizeable differences of bromide concentration at the low levels do not necessarily cause changes in bromide substitution in the chlorinated by-products formed and that the character of NOM plays more of a vital role.
6.4 Relationship Between Adsorbable Organic Halogens and Preozonation Adsorbable organic halogen (AOX) analysis is a measurement of the halogenated organic material in a water sample. In this work the origin of halogenated compounds arises from the chlorination of NOM. AOX analysis is therefore a useful method for estimating the formation of chlorinated organic compounds in total following disinfection with chlorine. The method detects organohalogen compounds apart from fluorinated compounds. Chlorinated disinfection by-products (DBPs) include the classes mentioned earlier in this chapter and other chlorinated compounds not previously mentioned including partially chlorinated humic substances. The results are reported as chloride in µg per litre. AOX analysis was conducted on both Hope Valley and Myponga controls and on each ozonated sample. Figure 6.11 and 6.12 graphically illustrates the effects of ozone on TOX formation.
64 CRC for Water Quality and Treatment Research Report 9-2001
Trichloroacetonitrile 3.0 Dichloroacetonitrile 0.06
2.5 0.05 Myponga
Hope Valley 2.0 0.04 Tod 1.5 0.03
1.0 0.02
0.01 0.5
0.00 0.0 0:1 0.5:1 1:1 2:1 3:1 5:1 0:1 0.5:1 1:1 2:1 3:1 5:1 Dibromoacetonitrile 3.5 Bromochloroacetonitrile 16
14 3.0
12 2.5 10
2.0 8
Haloacetonitrile formation mg DOC per ug in 1.5 6
1.0 4
2 0.5 0 0:1 0.5:1 1:1 2:1 3:1 5:1 0:1 0.5:1 1:1 2:1 3:1 5:1
Applied Ozone Dose to DOC Ratio
Figure 6.10: Effects of Ozone on Haloacetonitrile Formation Potentials.
1500 Myponga water source 1400
1300
1200
1100
1000
900
800
Adsorbable Organic Halides (ug/L) Halides Organic Adsorbable 700
0:11:12:13:14:15:1 Applied Ozone (mg/L) to DOC (mg/L)
Figure 6.11: Effect of Ozone on Adsorbable Organic Halide Formation in Myponga Water.
65 CRC for Water Quality and Treatment Research Report 9-2001
700 Hope Valley water source 650
600
550
500
450
400
350
300 Adsorbable Organic Halides (ug/L)
250 0:1 1:1 2:1 3:1 4:1 5:1 Applied Ozone Dose to DOC Ratio
Figure 6.12: Effect of Ozone on Adsorbable Organic Halide Formation in Hope Valley Water.
Figure 6.11 highlights the effects of ozone on AOX formation in Myponga surface water. Ozonation is shown to destroy chlorinated DBP precursors in an almost linear relationship. An applied ozone dose of 5:1 reduced total chlorinated compounds by over 50%. Hope Valley water similarly exhibited substantial reductions in AOX with increasing ozone doses. A 5:1 ozone dose also decreased amounts of chlorinated organic compounds by over 50%. An interesting observation is the trends in AOX reduction between the two surface waters. The reduction in AOX by ozonation in Myponga water follows a more steady trend. Hope Valley on the other hand is seen to first follow a rapid and then more steady reduction in AOX formation as the ozone dose increased depicted in Figure 6.12. An applied ozone dose of 0.5:1 decreased AOX formation in Hope Valley water by 30% and a maximum applied dose of 5:1 caused a 56% decrease relative to the control. The different profiles in AOX removal are indicative of the different character of NOM between the waters.
Trihalomethanes are one of the major classes of chlorinated DBPs and their analysis has commonly been used as a surrogate for AOX. Greater levels of THMs indicate a greater level of total halogenated organic compounds. Figure 6.13 and 6.14 illustrates the relationship between AOX and total THMs for both Myponga and Hope Valley water. The linearity achieved between AOX and total THMs for Myponga water is exceptional having a line of best fit R2 value of 0.995. The relationship for Hope Valley water was not as strong further confirming the differences in NOM character between the water sources. The points on the curves in Figures 6.13 and 6.14 represent the THM levels for each of the ozonated samples.
66 CRC for Water Quality and Treatment Research Report 9-2001
1500 Myponga water source 1400 0:1
1300 1:1 0.5:1 1200
1100
1000
2:1 900
800 3:1 5:1 Adsorbable Organic Halides (ug/L) Halides Organic Adsorbable 700 R= 0.995
300 350 400 450 500 550 600 650 700 Total Trihalomethanes (ug/L)
Figure 6.13: Relationship Between AOX and Total THMs – Myponga Water.
700 Hope Valley water source
0:1 650
600
550
500
0.5:1 450 1:1 400 2:1
350 3:1 5:1 300 Adsorbable Organic HalideAdsorbable Organic (ug/L) R= 0.927 250 100 120 140 160 180 200 220 240 260 Total Trihalomethanes (ug/L)
Figure 6.14: Relationship Between AOX and Total THMs – Hope Valley Water.
67 CRC for Water Quality and Treatment Research Report 9-2001
6.5 Chlorination of Individual Ozonation By-products The detection of aldehyde and ketoacid losses following chlorination in this work, reported earlier in Sections 5.3 and 5.5, prompted the investigation to determine whether these compounds react with chlorine to form chloroform and/or mono-, di- and trichloroacetic acid. Each aldehyde, ketoacid and carboxylic acid monitored in this study was spiked into individual reaction vials containing reagent water and chlorinated following the conditions outlined in Section 4.1. Following chlorination and the removal of free chlorine by the addition of a suitable quenching agent (see Table 4.1), the solutions were analysed to determine whether chloroform or chloroacetic acids were produced. The results are summarised in Table 6.1.
Table 6.1: Chlorination of Individual Ozonation By-products.
Ozonation By- Conc. CHCl3 MCAA DCAA TCAA product1 µmol/L (µmol/L) (µmol/L) (µmol/L) (µmol/L) Formaldehyde 16.7 ND <0.03 0.02 0.02 Acetaldehyde 11.4 0.01 <0.03 0.02 0.02 Propionaldehyde 8.6 0.01 <0.03 0.01 0.01 Butyraldehyde 6.9 0.01 <0.03 0.02 0.03 Heptaldehyde 4.4 0.02 <0.03 0.06 0.04 Benzaldehyde 4.7 0.02 <0.03 0.05 0.05 Glyoxal 8.6 0.01 <0.03 0.03 0.03 Methylglyoxal 6.9 0.02 <0.03 0.04 0.04 Pyruvic acid 5.7 0.01 <0.03 0.02 0.05 Glyoxylic acid 6.8 0.01 <0.03 0.02 0.03 Ketomalonic acid 4.2 0.01 <0.03 0.03 0.03 2-ketobutyric acid 4.9 0.01 <0.03 0.02 0.02 Formic acid 10.9 0.02 <0.03 0.04 0.04 Acetic acid 8.3 ND <0.03 0.01 0.01 Propionic acid 6.8 0.02 <0.03 0.05 0.05 Benzoic acid 4.1 0.02 <0.03 0.06 0.06 1 All analytes prepared at a concentration of 500µg/L.
As previously mentioned in Sections 5.3 and 5.5, a range of aldehydes and ketoacids were shown to substantially decrease in concentration following chlorination. These ozonation products are believed to react with chlorine and subsequently form chlorinated by-products. The reaction between aqueous chlorine and the compounds listed in Table 6.1 are shown to form negligible levels of chloroform, dichloroacetic acid and trichloroacetic acid. The identities of the products are unknown but warrant further attention.
68 CRC for Water Quality and Treatment Research Report 9-2001
7 Ozone and its Full Scale Use at Edenhope
Edenhope is a small country town in Western Victoria with a population of around 1000. Edenhope draws its water supply from Lake Wallace, a natural shallow eutrophic lake. The water is low in colour yet high in DOC content, and alkalinity as well as salinity are also high. Lake Wallace has commonly suffered severe blue-green algal blooms that have necessitated shutdown of the water supply and the introduction of emergency groundwater supplies due to confirmed toxicity of these blooms. To alleviate problems associated with algal blooms the town constructed a new advanced water treatment plant that utilises an ozone/biological activated carbon (BAC) process. The plant was commissioned on 22nd January 1998 and is one of only a few plants in Australia utilising ozone. The water quality parameters for Lake Wallace are listed in table 7.1.
Table 7.1: Raw water quality data.
Lake Wallace Sample Date 5th Feb. 99 pH 8.45 Colour (HU) 19.7 Turbidity (NTU) 4.5
UV254 0.447 DOC (mg/L) 23.8 SUVA (L/mg.cm, x10-2) 1.87
Alkalinity as Ca2CO3 (mg/L) 189 Total Hardness (mg/L) 210 Br - (mg/L) 1.49 Ca2+ (mg/L) 30.5 Mg2+ (mg/L) 32.4 - HCO3 (mg/L) 212 Total Dissolved Solids (mg/L) 780
7.1 Treatment Train Process The maximum capacity of the Edenhope plant is 3ML/day with normal operation of around 2.5ML/day. In the first stage of treatment raw water is dosed with an optimum level of alum allowing flocculation to occur. After about 20 minutes of ‘conditioning’ in the flocculation tank air is injected from below. Air bubbles rise and carry the flocs to the surface to form a scum (float) on the surface of the flotation tank. The bubbles of air rising to the surface continuously buoy the scum and it is periodically skimmed off to a sludge tank and discharged to the sewer. This process is collectively termed dissolved air flotation (DAF). About 90% of the turbidity in the water is removed by flotation. The remaining 10% is trapped when the treated water passes through sand filter beds located at the bottom of the tank. The second stage of treatment is an advanced treatment stage utilising ozone and BAC. Ozone is bubbled through the water with a contact time of 5 minutes. The ozone residual after 5 minutes is monitored and maintained constantly at approximately 0.1mg/L by automated regulation. The ozonated water is then passed through biologically activated carbon filters with an empty bed contact time of 15 minutes. A dose of chlorine is added to the treated water, and electric pumps then transfer the water to an elevated tank providing water pressure to the town. The treatment train is schematically represented in figure 7.1 and a summary of treatment plant design criteria are listed in table 7.2.
69 CRC for Water Quality and Treatment Research Report 9-2001
Figure 7.1: Schematic Representation of Edenhope Water Treatment Plant LW1: Lake Wallace raw water; LW2: following DAF; LW3: following ozonation; LW4: following BAC; LW5: water from public tap.
7.2 UV254 Reduction and DOC Levels through the Plant Flocculation in conjunction with DAF is very effective for DOC reduction as illustrated below in figure 7.2. Removal of the flocculated material by DAF and filtration achieves a 30% and 46% reduction in DOC and UV254 absorbance respectively. Ozonation then follows the conventional treatment and further reduction in UV254 absorbance is observed. Ozone readily adds to sites of unsaturation consequently destroying conjugated π-electron systems. The destruction of the conjugated sites reduce absorption of electromagnetic radiation at the higher wavelengths. A further 30% reduction in UV absorption at 254nm is observed following ozonation. The reduction in UV254 absorbance and negligible decrease of DOC clearly displays ozone’s ability to alter the character of NOM, yet its inability to completely oxidise organic compounds to CO2. DOC pre- and post-ozonation levels were determined to be 16.8 and 16.7mg/L respectively. Further decreases in these two parameters are observed following water passage through a biologically active carbon bed. Organics are removed as a result of microbial processes and adsorption onto the activated carbon.
70 CRC for Water Quality and Treatment Research Report 9-2001
Table 7.2: Treatment Plant Design Criteria – Edenhope Water Treatment Plant
Nominal Plant Capacity 3 ML/day Coagulant Dose Range Alum 50-200 mg/L Flocculation Stages 2 Flocculation Time 20 minutes total Flotation Rate (inc. DAF recycle) 8 m/hr Filtration Rate (inc. DAF recycle) 9 m/hr Filter Media Mono medium sand Disinfection Sodium hypochlorite pH adjustment Soda ash Advanced Treatment Module Ozonation pH Not greater than 6.5 Ozonation Capacity 800 kg/hour Maximum Ozone Dose 9 mg/L @ 2 ML/day Ozone contact Time 5 minutes @ 2 ML/day BAC Vessels 2 Empty Bed Contact Time 15 minutes @ 2 ML/day Activated Carbon Coal-based
7.3 Aldehyde Levels Through the Plant Low levels of formaldehyde, acetaldehyde and propionaldehyde are present in the raw water indicating the ubiquitous nature of these simple aldehydes. Flocculation through the DAF process removed 30% of the DOC but did not remove any of the aldehydes present in the water. Ozonation produced appreciable levels of formaldehyde, acetaldehyde, glyoxal and methylglyoxal. Formaldehyde was the major aldehyde formed at a level almost twice that of acetaldehyde, the second most prominent aldehyde. The passage of water following ozonation is through a bed of biologically activated carbon. Aldehydes are shown to be readily biodegradable by their removal through these microbial containing beds. The levels of aldehydes through the treatment process are shown in figure 7.3.
71 CRC for Water Quality and Treatment Research Report 9-2001
25 0.50 24 0.45 23 UV254 22 DOC 0.40
21 UV 0.35 20 254 Absorbance 19 0.30 18 0.25 17 16 0.20
Dissolved Organic Carbon (mg/L)Dissolved Organic 15 0.15 14 13 0.10 LW1 LW2 LW3 LW4 Treatment Process LW1: raw water; LW2: following DAF; LW3: following ozonation; LW4: following BAC
Figure 7.2: UV254 and DOC levels through stages of the treatment process
70 60 50 Formaldehyde Acetaldehyde 40 Propionaldehyde 30 20 10 0 25 glyoxal 20 methyl glyoxal
Aldehyde Formation, ppb. 15
10
5
0 LW 1LW 2LW 3LW 4LW 5 Treatment Train Process LW1:Lake Wallace raw water; LW2:following DAF; LW3:after ozonation; LW4:through BAC filters; LW5:public tap (post chlorination)
Figure 7.3: Monitoring the aldehyde levels through the treatment process.
72 CRC for Water Quality and Treatment Research Report 9-2001
7.4 Ketoacid Levels Through the Plant Levels of ketoacids increased following ozonation of Lake Wallace water within the treatment process. Pyruvic acid was formed at the highest level on ozonation followed closely by glyoxylic acid. Ketomalonic acid was significantly less and a trace level of 2-ketobutyric acid was also formed. An appreciable level of pyruvic acid and a smaller amount of glyoxylic acid were present in the influent raw water. Interestingly, the DAF stage markedly reduced their presence to trace levels although this stage had no effect on aldehyde concentrations. The sum of ketoacids formed following ozonation was significantly less than the total aldehydes. In contrast the relationship between aldehydes and ketoacids with ozonation experiments performed on Hope Valley, Myponga and Tod raw water at all doses was different. Ketoacid levels were higher than aldehydes for these waters clearly demonstrating the behaviour of different waters is unique. Consequently bench scale and pilot plant tests on specific waters intending to use ozone in the treatment process is vital in determining levels and relative abundances of different classes of compounds. Ketoacids, like aldehydes, are readily biodegradable and this is demonstrated in figure 7.4. Each ketoacid is reduced following the passage through the biologically activated carbon bed. A greater aldehyde than ketoacid removal was evident following analysis of the water after the BAC stage. A 77% removal of the total aldehydes formed was observed following BAC treatment compared with 60% of the ketoacids indicating aldehyde biodegradation occurs with more ease and at a faster rate. Ketoacid levels were monitored in the final treated water following chlorination and surprisingly glyoxylic and pyruvic acid increased by 1.8 and 2.4 times respectively. The increase in ketoacid levels following chlorination was not expected as previous work in this research revealed total elimination of ketoacids following chlorination. These observations were detailed earlier in chapter 5.5. Aldehydes also increased as did DOC and UV254. This suggests that due to the detention time within the system, the water from the public tap in the town represented raw water of a slightly different character.
7.5 Biodegradable Organic Matter Through the Treatment Process Bacterial regrowth potentials (BRP) were monitored throughout the treatment process as described in section 3.4. BRP data expressed as acetate carbon equivalents (ACE) following each treatment stage are listed in table 7.2 and bacterial growth curves of each sample are presented in figure 7.5. The BRP for Lake Wallace raw water was higher than Myponga raw water (see section 5.7). The dissimilar character of NOM between both sources best explains the difference in the waters’ ability to support bacterial growth. Approximately 60% of the NOM present in Lake Wallace raw water appears to have apparent molecular weights of less than 1000 Da. The high proportions of lower molecular weight organics, low SUVA value and low colour present in Lake Wallace water makes it more easily biodegradable than Myponga raw water. Myponga water is a highly coloured water with a moderate SUVA value of 0.037 L/mg.cm.
73 CRC for Water Quality and Treatment Research Report 9-2001
50
40 glyoxylic acid pyruvic acid 30 2-ketobutyric acid ketomalonic acid 20
10
0 Total Ketoacids 100
80
Ketoacid Levels, ppb. 60
40
20
0 LW 1 LW 2 LW 3 LW 4 LW 5 Treatment Process LW1:Raw water; LW2:following DAF; LW3:after ozonation; LW4:through BAC filters; LW5: public tap (post chlorination).
Figure 7.4: Monitoring the ketoacid levels through the treatment process.
The potential for bacterial regrowth is substantially increased following ozonation. This is expected as ozone reacts with NOM to form lower molecular weight oxygenated compounds that are susceptible to microbial degradation. For example, sizeable increases in aldehyde and ketoacid levels were observed following ozonation and these compounds have been shown to be major contributors to assimilable organic carbon (AOC). This was clearly observed in the earlier work in section 5.7 with ozonated Myponga water. Following ozonation, the water passes through a bed of biologically activated carbon (BAC). The use of biological filtration is extremely useful and serves a variety of purposes. Foremost the use of BAC reduces the biodegradable organic matter (BOM) present at higher levels in the ozonated water compared to the raw water. The removal of BOM ensures a biologically stable water and coliform regrowth control. This is clearly illustrated by the ACE of the water following BAC treatment listed in table 7.2 and the removal of assimilable organic compounds such as aldehydes and ketoacids mentioned earlier in the chapter. Secondly, the activated carbon has a high efficiency for removing, by adsorption, unwanted organic compounds that may pose a threat to the treated water quality. These include pesticides and taste and odour-causing compounds. And finally, a reduction in chlorinated DBPs in the product water is achieved by the removal of DBP precursors by a reduction in the DOC of the water. A reduction in DOC lowers the required chlorine demand of the treated water and hence results in the reduction of DBP formation.
74 CRC for Water Quality and Treatment Research Report 9-2001
Table 7.3: Bacterial regrowth potentials through a treatment plant – Edenhope filtration plant.
Treatment process ACE (µg/L) Lake Wallace raw 198 After DAF 239 After ozonation 889 After BAC 346 Distribution system 324
Figure 7.5: Bacterial Growth Curves Through Treatment Process
75 CRC for Water Quality and Treatment Research Report 9-2001
7.6 Chlorinated DBP Formation in Finished Water Lake Wallace treated water supplied to the consumer’s tap contains a notable level of total trihalomethanes (THMs). The level just falls below the Australian guideline level currently set at 250 µg/L however one must keep in mind that the Australian guideline level is more than double current legally enforced overseas limits. Table 7.4 lists the resultant levels of chlorinated DBPs formed following chlorination of Lake Wallace treated water.
Table 7.4: Chlorinated disinfection by-products in the finished water
DBP Raw water Public tap Chloroform <1.0 6.1 Bromodichloromethane <1.0 19 Chlorodibromomethane <1.0 61 Bromoform <1.0 124 Monochloroacetic acid <2.5 <2.5 Dichloroacetic acid <0.1 6.5 Trichloroacetic acid <0.1 1.7 Trichloroacetonitrile <0.2 <0.2 Dichloroacetonitrile <0.1 2.2 Bromochloroacetonitrile <1.0 3.5 Dibromoacetonitrile <1.0 24 1,1-dichloroacetone <0.1 1.2 1,3-dichloroacetone <0.5 <0.5 1,1,1-trichloroacetone <0.1 1.2 Chloral hydrate <0.1 2.5 Chloropicrin <1.0 <1.0
The individual THM levels formed in the finished water is consistent with the fact that Lake Wallace raw water is relatively saline and contains a reasonably high level of bromide. Bromoform is the predominant THM and comprises 59% of the total THMs formed. Dibromochloromethane is the second most prominent THM to form and chloroform being the minor THM, present at only 3% relative to total THMs. The change in speciation from chlorine to bromine containing species in waters with moderate to high levels of bromide content was similarly observed for Tod water in chapter 6 and has been observed and well documented by other authors [92,97-99]. The THM levels formed following full-scale treatment of Lake Wallace water is a clear indication of the importance to study the effects of preozonation in local situations. Advanced water treatment technology, that is, the incorporation of ozone and biologically activated carbon (BAC) in the treatment process cannot be assumed to dramatically reduce total THMs to a level below current overseas standards. It has been well observed in this study that a substantial increase in THM levels with a high proportion of brominated species is seen in waters with high levels of bromide regardless if preozonation is applied. Utilities treating water containing moderate to high levels of bromide must carefully consider implementing expensive advanced treatment techniques such as ozone and BAC to control DBP levels in the distribution system.
76 CRC for Water Quality and Treatment Research Report 9-2001
Chloroacetic acid levels formed in Lake Wallace finished water are low. This compares well with the levels found in chlorinated Tod water, section 6.2. Unfortunately due to time constraints the current methodology did not screen for the brominated analogues which may form at sizeable levels and possibly pose health risk concerns. Similar trends between Lake Wallace water and Tod water were also seen for other halogenated DBPs. Figure 7.6 depicts the similar contributions of individual haloacetonitriles between both waters. Dibromoacetonitrile (DBAN) was the major haloacetonitrile formed in both waters and present in similar relative ratios of 81% and 74% respectively for Lake Wallace treated water and Tod water ozonated at an applied dose of 2:1 and subsequently chlorinated. Trichloroacetonitrile (TCAN) in both waters was below the limit of quantitation. These two surface waters differ considerably in their NOM makeup. Lake Wallace raw water has a very high dissolved organic carbon content of 23.8 mg/L compared with 4.1 mg/L for Tod water. Their SUVA values also differ and were determined to be 5.4 and 1.9 for Tod and Lake Wallace water respectively. Low SUVA values can indicate that the NOM may be of an algogenic origin instead of a terrestrial one. This is indeed the case, as Lake Wallace is well known for its severe algal blooms. The prospect of algal blooms is the reason for ozone implementation in the plant. Ozone is effective in the destruction of harmful algal toxins and taste and odour-causing algal metabolites. The highly different NOM character between both sources yet the similar relative contribution of brominated species following chlorination is illustrative of the significance of the inorganic parameters in determining the by-products formed.
DCAN 7% 6% Tod Water (2:1) Lake Wallace Water
BCAN 12% 20%
DBAN 81% 74%
0 20 40 60 80 100 120 140 160 % Contribution
Figure 7.6 Contribution of individual haloacetonitriles between Tod water and Lake Wallace water.
77 CRC for Water Quality and Treatment Research Report 9-2001
8 SUMMARY AND CONCLUSIONS
The ozone demands of three South Australian waters, namely Myponga Reservoir, Hope Valley Reservoir and Tod Reservoir water, were determined by applying increasing ozone doses and calculating the aqueous ozone concentration immediately after the required contact time had elapsed. Once the water’s demand was met, a linear relationship between the applied ozone dose and the aqueous ozone residual was observed in all waters. Ozone demand was highest in Myponga Reservoir water, the water with the highest NOM content. Based on NOM content, the ozone demand in Tod Reservoir water would be expected to be lower than Hope Valley Reservoir water. The ozone demands for these two waters were found to be similar. High levels of salinity and alkalinity were associated with higher ozone demands for Tod Reservoir water. The nature of the DOC may also influence the demand. Analytical method for the determination of aldehydes, ketoacids and carboxylic acids were developed and validated in this study. The analysis of both aldehydes and ketoacids involved aqueous phase derivatisation of their respective carbonyl functional groups using pentafluorobenzylhydroxylamine (PFBHA). Derivatisation with PFBHA increased the hydrophobicity of the polar analytes enabling these classes of compounds to be extracted from the aqueous phase and amenable to gas chromatographic separation. Trace levels of formaldehyde were present in the laboratory’s reagent water and considerable effort was made to eliminate the background formaldehyde. Irradiation with UV light, chemical oxidation to form CO2 and the use of anion exchange column chromatography in order to remove or reduce the background formaldehyde was not successful. Formaldehyde levels in the reagent water were negligible compared to the levels formed following ozonation and so further time-consuming treatment was not investigated. Improvement on the analysis of ketoacids was achieved by the incorporation of a buffer to control the pH of all standards and samples. As the derivatisation to form the corresponding oxime is pH dependent, pH control ensures uniformity in the rates of derivatisation for all standards and samples. The lack of pH control by means of a buffer has been shown to cause large variations in results through interlaboratory proficiency tests. Bench scale ozonation experiments were conducted on three South Australian surface waters at applied ozone to DOC mass ratios of 0.5:1, 1:1, 2:1, 3:1 and 5:1. Following ozonation of each surface water, parameters that were determined included UV254 absorbance, DOC concentrations, ozonation by-products and bacterial regrowth potentials. Ozonation was observed to significantly alter the chemical structure of the NOM in all waters indicative of the large reduction in UV254 absorbance. For all waters the largest reduction in UV254 absorbance was found to occur at low applied ozone doses of 1-2:1. At an applied ozone dose of 2:1 UV254 absorbance reductions of 42, 51 and 56% were observed for Tod, Hope Valley and Myponga raw water respectively. Increasing ozone dose caused further reductions in UV254 absorbance. DOC removal was seen in both Myponga and Hope Valley raw waters with greater than 20% reductions observed at the maximum applied ozone dose. Negligible DOC reduction was observed for Tod water. Formaldehyde, acetaldehyde, glyoxal and methyl glyoxal were formed upon the ozonation of Myponga, Hope Valley and Tod waters. Formaldehyde was found to be the predominant aldehyde at the higher ozone doses for all three waters. The trend in formation for each aldehyde species differed considerably suggesting that formation is dependent on a number of factors and each species, particularly dialdehydes, are believed to originate from alternative precursors. Comparing aldehyde levels, relative ratios of each aldehyde in each surface water and results obtained by international researchers, it is evident that the character of the water and the nature of the DOC plays a vital role in the formation of these compounds. Various ketoacids were also shown to form from the ozonation of NOM in waters. Similar ketoacid profiles were observed for both Myponga and Hope Valley water at all ozone doses except for ketomalonic acid which was formed at approximately double the concentration in Myponga water. Ketoacid levels increased with increasing ozone dosage in both waters. Ketoacid levels in Myponga and Hope Valley water exceeded the aldehyde levels at all ozone doses. Tod water was atypical in that ketoacid formation increased and then decreased with increasing ozone dose. Aldehyde and ketoacid levels were determined following chlorination of the raw and ozonated waters to investigate the effect of chlorine on these compounds. Formaldehyde and acetaldehyde levels increased in the raw and all ozonated waters following chlorination. Methyl glyoxal and all ketoacids
78 CRC for Water Quality and Treatment Research Report 9-2001 were found to substantially diminish following chlorination suggesting that these ozonation by- products react with chlorine to form unidentified chlorination by-products. Glyoxal levels were found to only slightly reduce after chlorination. The chlorination by-products were not identified but were not the more common DBPs such as chloroform or the chlorinated acids. The trend in formic acid and acetic acid formation following ozonation in both Myponga and Hope Valley waters were similar to formaldehyde and acetaldehyde. Levels of these carboxylic acids at the higher ozone doses far exceeded total aldehyde and total ketoacid levels. Higher ozone doses however had no effect on formic acid levels. The lack of formic acid formation at the higher ozone doses may be attributable to CO2 production. Benzoic acid was not produced in either Myponga or Hope Valley ozonated waters. Bacterial regrowth potentials were determined for Myponga raw water and selected ozonated samples. The bacterial regrowth potential was seen to increase with increasing ozone dose indicating ozone reacts with NOM present in the water to form compounds that are easily biodegradable and subsequently increasing the water’s biodegradable organic content. A strong relationship was observed between aldehydes, ketoacids and the bacterial regrowth potential highlighting that aldehyde and ketoacid species are significant components of biodegradable organic matter. This was confirmed by high reductions in levels of formaldehyde, acetaldehyde and the ketoacids. The dialdehydes, glyoxal in particular, were only partly degraded. The analysis of these compounds can therefore be used as surrogates for estimating bacterial regrowth in water samples. All ozonated waters were subsequently chlorinated and the effects of ozone were evaluated by measuring the formation potentials of trihalomethanes (THMs), chloroacetic acids, chloroketones, haloacetonitriles, chloral hydrate, chloropicrin and total organic halides. Chloroform formation potentials differ markedly between all three raw water controls indicative of the different source water characteristics of each water. Chloroform contributes 74%, 54% and less than 1% towards total THMs in Myponga, Hope Valley and Tod raw water controls respectively. Total THM formation in Tod raw water was 1.5 times greater than total THMs in the Myponga control yet the DOC concentration of Tod is approximately three times less than Myponga water. The high bromide level in Tod water is seen to play a vital role in the levels and types of THMs that form. Bromoform was found to be the predominant THM formed in Tod water. Pre-ozonation of Myponga and Hope Valley water decreased total THM formation following chlorination. The greater the applied ozone dose, the greater reduction in total THMs was observed for both waters. Formation of di- and tri-brominated THM species are small compared to chloroform levels in both Myponga and Hope Valley water and so THM reduction is largely represented by the reduction of chloroform levels in the ozonated samples. A negligible formation of chloroform was observed in Tod water. Bromoform, the predominant THM in Tod water, was observed to decrease with increasing ozone doses of up to 2:1. However further increases in ozone concentration resulted in a steady increase in concentrations. Trichloroacetic acid and dichloroacetic acid were formed at significant levels in the Myponga and Hope Valley controls. Trichloroacetic acid formation in the Myponga control was approximately half the concentration of chloroform whereas in the Hope Valley control, levels of trichloroacetic acid surpassed the chloroform levels indicating characteristic differences between the NOM of each water type. Total chloroacetic acid formation decreased with pre-ozonation in Myponga and Hope Valley waters. The decline in total chloroacetic acid formation was due to the destruction of trichloroacetic acid precursors by pre-ozonation. The effect of pre-ozonation on dichloroacetic acid and monochloroacetic acid formation was negligible suggesting that trichloroacetic acid is formed from distinct precursors. Levels of chloroacetic acids in Tod water were very low. The high bromide content in Tod water suggests brominated acetic acids and a mixture of brominated and chlorinated acids would form at significant levels. However these haloacetic acid analogues were not monitored. Levels of other disinfection by-products monitored in this study were formed at very low levels under the chlorination conditions employed. The effects of ozone on chloroketone formation seemed to have an opposite effect to what was observed with the more dominant chlorinated compounds such as chloroform and trichloroacetic acid. Ozonation increased chloroketone precursors in both Myponga and Hope Valley waters. This was similarly observed for chloral hydrate. Chloropicrin levels on the other hand were observed to be unaffected by pre-ozonation at the lower ozone doses. Application of higher ozone doses however significantly increased chloropicrin formation. AOX analysis is a useful method for estimating the formation of total chlorinated organic compounds following disinfection with chlorine. Pre-ozonation of Myponga and Hope Valley water is shown to
79 CRC for Water Quality and Treatment Research Report 9-2001 destroy chlorinated disinfection by-product precursors. Substantial reductions in AOX were observed with increasing ozone doses. AOX reductions of greater than 50% were seen for both waters at the maximum applied ozone dose of 5:1. Aldehydes, ketoacids, chlorination by-products and other water quality parameters were determined through each treatment stage at the Edenhope Water Filtration Plant. The plant uses advanced water treatment processes such as flocculation, dissolved air flotation, ozone contactors, biologically activated carbon (BAC) and chlorination to supply the community with safe drinking water. A 30% reduction in the DOC concentration and a 46% reduction in UV254 absorbance were achieved following passage of the water through the first stage of treatment, that is, flocculation and dissolved air flotation. Ozonation then follows and a further significant reduction in UV254 absorbance was observed. A negligible change in DOC concentration was seen following ozonation, again clearly indicating ozone’s ability to alter the character of NOM through oxidative reactions without complete mineralisation to CO2. Aldehydes and ketoacids were formed at sizeable levels following ozonation. Formaldehyde was the major aldehyde formed followed by acetaldehyde and glyoxal. Pyruvic acid was the predominant ketoacid detected followed by glyoxylic acid. As expected, the bacterial regrowth potential of the water following ozonation significantly increased. Following BAC treatment, levels of aldehydes, ketoacids and the bacterial regrowth potential of the water had significantly diminished. A small reduction in DOC concentration and UV254 was also observed. DBP levels in the treated water following chlorination were found to be significantly high with the brominated analogues predominating consistent with raw waters containing high levels of bromide. The high THM levels formed following full-scale treatment of Lake Wallace water is a clear indication of the importance to study the effects of pre-ozonation in local situations. Water utilities must carefully consider the implementation of expensive treatment techniques such as ozonation and BAC as a means to control DBP levels. Reductions of DBP levels by the incorporation of ozone cannot be assumed. It has been well observed in this study that substantial increases in disinfection by-products with a high proportion of brominated species is seen in waters with moderate to high levels of bromide regardless if pre-ozonation is applied.
80 CRC for Water Quality and Treatment Research Report 9-2001
9 Recommendations
• The discovery that formic acid and acetic acid formation was substantially greater than total aldehyde and ketoacid levels demonstrates the importance to identify and quantify this class of compounds. The method used in this study for the analysis of low molecular weight carboxylic acids was limited to various water types. A more reliable method capable of analysing a wider range of carboxylic acids and its application to waters with high salinities needs to be developed. • Future ozonation and chlorination experiments on fractionated natural organic matter (NOM) isolated from the waters used in this research may be beneficial. Certain NOM fractions may give rise to greater amounts of ozonation and/or chlorination by-products. This knowledge will provide an insight into which fractions need to be removed in order to minimise the formation of disinfection by-products. Effects of individual parameters such as bromide ion content, alkalinity, pH and salinity should also be investigated. Fractionation techniques include ultrafiltration (fractionates the NOM by apparent molecular size) or the use of adsorption and ion-exchange chromatography (fractionating the NOM on the basis of functionality). • This research showed that dialdehyde and ketoacid levels significantly diminished following chlorination suggesting that these carbonyl compounds take part in substitution reactions with chlorine. Chlorination of individual aldehydes, ketoacids and carboxylic acids were shown to form negligible levels of chloroform and/or chloroacetic acids, however other unknown significant sized peaks were present. Due to time constraints, mass spectral identification was not possible. Identification of these unknown chlorinated compounds should be investigated. • The addition of alum and the use of jar tests to simulate the flocculation stage in water treatment plants could be performed on these source waters. Comparison of ozonation by-products in water samples without flocculation, with ozonation prior to flocculation and with ozonation after flocculation would identify dominant by-product precursors.
81 CRC for Water Quality and Treatment Research Report 9-2001
10 References
1. Thurman, E.M., Malcolm, R.L. 1983. Structural Study of Humic Substances: New Approaches and Methods. In Aquatic and Terrestrial Humic Materials. Christman, R.F., Gjessing, E.T. Eds., Ann Arbor Sci., Michigan, Ch. 1, 1-23. 2. Malcolm, R.L. 1991. Factors to be Considered in the Isolation and Characterisation of Aquatic Humic Substances. In Humic Substances in the Aquatic and Terrestrial Environment, Boren, H., Allard, B., Eds., Wiley, London, 369-391. 3. Fuchs, W. 1931. Die Chemie Der Kohle. Springer Berlin. 4. Schnitzer, M., Khan, S.V. 1972. Characterisation of Humic Substances by Chemical Methods. In Humic Substances in the Environment, Marcel Dekker Inc., New York, NY. Ch.3., 29-54. 5. Steelink, C. 1985. Implications of Elemental Characteristics of Humic Substances. In Humic Substances in Soil, Sediment and Water: Geochemistry, Isolation and Characterisation, Aiken, G.R., McKnight, D.M., Wershaw, R.L., MacCarthy, P., Eds., Wiley-InterScience, New York, NY, 457-476. 6. McCarthy, J.F., Zachara, J.M. 1989. Subsurface Transport of Contaminants. Environ. Sci. Technol., 23, 496-502. 7. Khan, S.U. 1980. Determining the Role of Humic Substances in the Fate of Pesticides in the Environment. J. Environ. Sci. Health, B15, 1071-1090. 8. Hiraide, M., Ueda, T., Mizuike, A., 1989. Anal. Chim. Acta, 227, 421. 9. Rook, J.J. 1974. Formation of Haloforms during Chlorination of Natural Water. J. Water Treat. Exam., 23, 234-243. 10. Rook, J.J. 1977. Chlorination of Fulvic Acids in Natural Waters. Environ. Sci. Technol., 11, 478-482. 11. Serkiz, S.M., Perdue, E.M. 1990. Isolation of Dissolved Organic Matter from the Suwannee River using Reverse Osmosis. Water Res., 24, 911-916. 12. Koechling, M.T., Summers, R.S. 1995. Evaluation of Ultrafiltration and Hydrophobic Separation for Characterising Natural Organic Matter. Proc. AWWA Water Quality Technology Conference, New Orleans, LA, 1913-1924. 13. Thurman, E.M., Malcolm, R.L. 1981. Preparative Isolation of Aquatic Humic Substances. Environ. Sci. Technol., 15, 463-466. 14. Leenheer, J.A. 1981. Comprehensive Approach to Preparative Isolation and Fractionation of Dissolved Organic Carbon from Waters and Wastewaters. Environ. Sci. Technol., 15, 578- 587. 15. Morran, J., Bursill, D.B., Drikas, M., Nguyen, H. 1996. A New Process for Removing Natural Organic Matter. Proc. IWSA Workshop on Influence of Natural Organic Matter Characteristics in Drinking Water Treatment and Quality, Poitiers, France, September. 16. Mallevialle, J. 1996. Why is Natural Organic Matter Problematic? Proc. IWSA Workshop on Influence of Natural Organic Matter Characteristics in Drinking Water Treatment and Quality, Poitiers, France, September. 17. Aiken, E.W., Hoff, J.C., Lippy, E.C. 1982. Waterborne Outbreak Control: Which Disinfectant? Environ. Health Perspect., 46, 7-12. 18. Hoff, J.C. Inactivation of Microbial Agents by Chemical Disinfectants. Report EPA/600/2- 86/067, Water Engineering Research Laboratory, USEPA, Cincinatti, Ohio, USA. 19. Bellar, T.A., Lichtenberg. J.J., Kroner, R.C. 1974. The Occurrences of Organohalides in Chlorinated Drinking Waters. J. Am. Water Works Assoc., 66, 703-706. 20. NCI. 1976. Report on the Carcinogenesis Bioassay of Chloroform. NTISPB-264-018. Bethesda, MD: National Cancer Institute.
82 CRC for Water Quality and Treatment Research Report 9-2001
21. NTP. 1986. Toxicology and Carcinogenesis Studies of Bromodichloromethane in F344/N Rats and B6C3F1 Mice. Technical Report Series No. 321, NIH, Publication No. 88-2537, CAS No. 75-27-4. National Toxicology Program, US Department of Health and Human Services. 22. NTP. 1989. Toxicology and Carcinogenesis Studies of Tribromomethane (Bromoform) in F344/N Rats and B6C3F1 Mice. Technical Report Series No. 350, NIH, Publication No. 89- 2805, CAS No. 75-25-2. National Toxicology Program, US Department of Health and Human Services. 23. National Interim Primary Drinking Water Regulations: Control of Trihalomethanes in Drinking Water. 1979. Federal Register. 44(231), 68624-68707. 24. Uden, P.C., Miller, J.W. 1983. Chlorinated Acids and Chloral in Drinking Water. J. Am. Water Works Assoc., 75, 524-527. 25. Christman, R.F., Norwood, D.L., Millington, D.S., Johnson, J.D. 1983. Identity and Yields of Major Halogenated Products of Aquatic Fulvic Acid Chlorination. Environ. Sci. Technol. 17, 625-628. 26. Hargesheimer, E.E., Satchwill, T. 1989. Gas Chromatographic Analysis of Chlorinated Acids in Drinking Water. Aqua, 38, 345-351. 27. DeAngelo, A.B., McMillan, L.P. 1987. Carcinogenicity of Chlorinated Acetic Acids. In Water Chlorination: Chemistry, Environmental Impact and Health Effects. Jolley, R.L., et al., Eds., Vol. 6, Lewis Publ., Chelsea, Michigan, 193-199. 28. Trehy, M.L., Bieber, T.I. 1981. Detection, Identification and Quantitative Analysis of Dihaloacetonitriles in Chlorinated Natural Waters. In Advances in the Identification and Analysis of Organic Pollutants. Keith, L.H. Ed., Vol 2, Ann Arbor Science Publ., Ann Arbor, Michigan, 941-975. 29. Bull, R.J., Robinson, M. 1985. Carcinogenic Activity of Haloacetonitrile and Haloacetone Derivatives in the Mouse Skin and Lung. In Water Chlorination: Chemistry, Environmental Impact and Health Effects, Jolley, R.L., et al., Eds., Vol. 5, Lewis Publ., Chelsea, Michigan, 221-227. 30. Pontius, F.W. 1997. Future Directions in Water Quality and Regulations. J. Am. Water Works Assoc., 89(3), 40-54. 31. Bursill, D.B. 1983. Trihalomethanes in South Australian Water Supplies. Proc. AWWA 10th Federal Convention, Sydney, 4.1-4.13. 32. Nicholson, B.C., Hayes, K.P., Bursill, D.B. 1984. By-products of Chlorination, Water, 9(3), 11- 15. 33. Simpson, K.L., Hayes, K.P. 1998. Drinking Water Disinfection By-products: An Australian Perspective. Water Res., 32, 1522-1528. 34. NHMRC/ARMCANZ. 1996. Australian Drinking Water Guidelines, National Health and Medical Research Council/ Agricultural and Resource Management Council of Australia and New Zealand. 35. AWQC, Australian Water Quality Centre, SA Water Corporation. Method manuals. 36. Nicholson, B.C., Hayes, K.P., Thomas, P.M. 1990. Chlorination By-products of Drinking Water Disinfection. In Halogenated Organics and the Environment, Adelaide, Australia, 1-10. 37. Krasner, S.W., McGuire, M.J., Jacangelo, J.G., Patania, N.L., Reagan, K.M., Aieta, E.M. 1989. The Occurrence of Disinfection By-products in US Drinking Water. J. Am. Water Works Assoc., 81(8), 41-53. 38. Rice, R.G., Robson, C.M., Miller, G.W., Hill, A.G. 1981. Uses of Ozone in Drinking Water Treatment. J. Am. Water Works Assoc., 73, 44-57. 39. Rositano, J., Nicholson, B.C., Pieronne, P. 1998. Destruction of Cyanobacterial Toxins by Ozone, Ozone Sci. Eng., 20, 223-228. 40. McGuire, M.J., Gaston, J.M. 1998. Comparing Peroxone and Ozone for Controlling Taste and Odor Compounds, Ozone Sci. Eng., 20, 215.
83 CRC for Water Quality and Treatment Research Report 9-2001
41. Ferguson, D.W., McGuire, M.J., Koch, B., Wolfe, R.L., Aieta, E.M. 1990. Disinfection By- products and Microorganisms. J. Am. Water Works Assoc., 82(4), 181-191. 42. Dimitriou, M.A., Ivanco, E. 1989. Ozone in Water Treatment, Proc. 9th Ozone Congress, New York. 43. Husley, R.A., Long, B.W., Adams, C.D. 1990. Proc. IOA Spring Conf. on New Developments: Ozone in Water and Wastewater Treatment, Shreveport, La.. 44. Bailey, P.S. 1978. Ozonation in Organic Chemistry, Vol. 1. Olefinic Compounds. Academic Press, New York, NY. 45. Bailey, P.S. 1982. Ozonation in Organic Chemistry, Vol. 2. Non-olefinic Compounds. Academic Press, New York, NY. 46. Hoigne, J., Bader, H. 1976. The Role of Hydroxyl Radical Reactions in Ozonation Processes in Aqueous Solutions. Water Res., 10, 377-386. 47. Edwards, M., Benjamin, M.M. 1992. Transformation of NOM by Ozone and its Effect on Iron and Aluminium Solubility. J. Am. Water Works Assoc., 84(6), 56-66. 48. Gloor, R., Leidner, H., Wuhrmann, K., Fleischmann, Th. 1981. Exclusion Chromatography with Carbon Detection. A Tool for Further Characterisation of Dissolved Organic Carbon. Water Res., 15, 457-462. 49. Owen, D.M., Amy, G.L., Chowdhury, Z.K., Paode, R., McCoy, G., Viscosil, K. 1995. NOM Characterisation and Treatability. J. Am. Water Works Assoc., 87(1), 46-63. 50. Trussell, R.R., Umphres, M.D. 1978. The Formation of Trihalomethanes, J. Am. Water Works Assoc., 70, 604-612. 51. Glaze, W.H., et al., 1982. Destruction of Pollutants in Water with Ozone in Combination with Ultraviolet Radiation: 2: Natural Trihalomethane Precursors. Environ. Sci Technol., 16, 454- 458. 52. Robertson, J.L., Oda, A. 1983. Combined Application of Ozone and Chlorine or Chloramine to Reduce Production of Chlorinated Organics in Drinking Water Disinfection. Ozone Sci. Eng., 5, 79. 53. Reckhow, D.A., Singer, P.C. 1984. The Removal of Organic Halide Precursors by Preozonation and Alum Coagulation. J. Am. Water Works Assoc., 76(4), 151. 54. Singer, P.C. 1988. Formation and Control of Trihalomethanes. Proc. AWWA, Water Quality Technology Conference. St. Louis, Mo. 55. Umphres, M.D., Trussell, R.R., Trussell, A.R., Leong, L.Y.C., Tate, C.H. 1979. The Effects of Preozonation on the Formation of Trihalomethanes, OZONews, 6(3), Part 2, Mar. 56. Stieglitz, L., Roth, W., Kuhn, W., Leger, W, 1976. Behaviour of Organohalogen Compounds during Drinking Water Preparation, Vom Wasser, 47, 347-377. 57. Reckhow, D.A., Singer, P.C. 1985. Mechanisms of Organic Halide Formation During Fulvic Acid Chlorination and Implications with Respect to Preozonation. In Water Chlorination: Chemistry, Environmental Impact and Health Effects. Jolley, R.L., et al., Eds., Vol. 5. Lewis Publ., Chelsea, Mich., 1229-1257. 58. Paode, R.D., Amy, G.L., Krasner, S.W., Summers, R.S., Rice, E.W. 1997. Predicting the Formation of Aldehydes and BOM. J. Am. Water Works Assoc., 89(6), 79-93. 59. Najm, I.N., Krasner, S.W., 1995. Effects of Bromide and NOM on By-Product Formation, J. Am. Water Works Assoc., 87(1), 106-115. 60. Jacangelo, J.G., Patania, N.L., Reagan, K.M., Marco Aieta, E., Krasner, S.W., McGuire, M.J. 1989. Ozonation: Assessing Its Role in the Formation and Control of Disinfection By-products. J. Am. Water Works Assoc., 81(8), 74-84. 61. Gilli, G., Scursatone, E., Palin, L., Bono, R., Carraro, E., Meucci, L. 1990. Water Disinfection: A Relationship Between Ozone and Aldehyde Production. Ozone Sci. Eng., 12, 231-241.
84 CRC for Water Quality and Treatment Research Report 9-2001
62. Glaze, W.H., Koga, M., Cancilla, D. 1989. Ozonation Byproducts. 2. Improvement of an Aqueous Phase Derivatization Method for the Detection of Formaldehyde and Other Carbonyl Compounds Formed by the Ozonation of Drinking Water. Environ. Sci. Technol., 23, 838-847. 63. Andrews, S.A., Huck, P.M., Coutts, R.T. 1993. Quantitation of Ozonation By-products of Fractionated Aquatic Natural Organic Matter, Vom Wasser, 81, 151-165. 64. Bull, R.J., Kopfler, F.C. 1991. Health Effects of Disinfectants and Disinfection By-products. Am. Water Works Research Foundation and Am. Water Works Assoc., Denver. 65. Hrudey, S.E., Gac, A., Daignault, S.A. 1988. Potent Odour-causing Chemicals Arising From Drinking Water Disinfection. Water Sci. Technol., 20, 55-61. 66. Xie, Y., Reckhow, D.A. 1992. A New Class of Ozonation By-Products: The Ketoacids. Proc. AWWA Annual Conf., 251-265. 67. Garcia-Araya, J.F., Croue, J.P., Beltran, F.J., Legube, B. 1995. Origin and Conditions of Ketoacid Formation During Ozonation of Natural Organic Matter in Water. Ozone Sci. Eng., 17, 647-656. 68. Xie, Y., Reckhow, D.A. 1992. Identification and Quantification of Ozonation By-products: Ketoacids in Drinking Water, Proc. IOA Pan American Committee, Pasadena Conference. 69. Glaze, W.H., Koga, M., Cancilla, D., Wang, K., McGuire, M.J., Liang, S., Davis, M.K., Tate, C.H., Aieta, E.M.. 1989. Evaluation of Ozonation By-products from Two California Surface Waters. J. Am. Water Works Assoc., 81(8), 66-73. 70. Yamada, H., Somiya, I. 1989. The Determination of Carbonyl Products in Ozonated Water by the PFBOA Method. Ozone Sci. Eng., 11, 127. 71. Sclimenti, M.J., Krasner, S.W., Glaze, W.H. 1990. Ozone Disinfection By-Products: Optimisation of the PFBHA Derivatization Method for the Analysis of Aldehydes, Proc. AWWA Water Quality Technology Conf., 477-501. 72. Ribas, F., Frias, J., Huguet, J.M., Lucena, F. 1997. Efficiency of Various Water Treatment Processes in the Removal of Biodegradable and Refractory Organic Matter. Water Res., 31, 639-649. 73. Van der Kooij, D., Visser, A., Hijnen, W.A.M. 1982. Determining the Concentration of Easily Assimilable Organic Carbon in Drinking Water. J. Am. Water Works Assoc., 74, 540-545. 74. Griffini, O., Bao, M.L., Barbieri, K., Birruni, D., Santianni, D., Pantani, F. 1999. Formation and Removal of Biodegradable Ozonation By-products during Ozonation-Biofiltration Treatment: Pilot Scale Evaluation. Ozone Sci. Eng., 21, 79-98. 75. Miltner, R.J., Shukairy, H.M., Summers, R.S. 1992. Disinfection By-product Formation and Control by Ozonation and Biotreatment. J. Am. Water Works Assoc., 84(11), 53-62. 76. Glaze, W.H. 1986. Reaction Products of Ozone: A Review. Environ. Health Perspect., 69, 151-157. 77. Schmidt, W., Petzoldt, H., Hambsch, B., Werner, P. 1995. Influence of Pre-ozonation of Bank Filtrate on the Bacterial Regrowth Potential and the Formation of Disinfection By-products by Chlorination. Vom Wasser, 84, 301. 78. Lee, S.H., O’Connor, J.T., Banerli, S.K. 1980. Biologically Mediated Corrosion and Its Effect On Water Quality in Distribution Systems. J. Am. Water Works Assoc., 72, 636-645. 79. Drikas, M., Tapang, K.C., Bennet, L.E. 1993. Bacterial Regrowth Potential of Treated Water Supplies. Water, 20(4), 35-39. 80. APHA-AWWA-WEF: 1992. Standard Methods for the Examination of Water and Wastewater, 18th Ed. and Supplement, American Public Health Association, AWWA, Water Environment Federation, Washington D.C., 1992. 81. Carlson, M. 1993. The Effect of Organic Carbon on Ozone Demand and Decay. Proc. AWWA Water Quality Technology Conference, Miami, Fl., 321-342.
85 CRC for Water Quality and Treatment Research Report 9-2001
82. Franklin McKnight, K., Carlson, M., Fortin, P., Ziesemer, C. 1993. Comparison of Ozone Efficiency for Manganese Oxidation Between Raw and Settled Water, Ozone Sci. Eng., 15, 331-341. 83. Munch, J.W., Munch, D.J., Winslow, S.D. 1998. A User’s Guide to Aldehyde Analysis Using PFBHA Derivatization and GC/ECD Detection: Avoiding the Pitfalls. Proc. AWWA Water Quality Technology Conference, San Diego, California, 4D-3. 84. Andrews, S.A., Huck, P.M. 1996. Using Fractionated Natural Organic Matter to Study Ozonation By-product Formation. In Disinfection By-products in Water Treatment: The Chemistry of Their Formation and Control. Minear, R.A., Amy, G.L., Eds., Lewis Publishers, New York, 411-447. 85. Andrews, S.A. 1993. Organic By-product Formation From the Ozonation and Chlorination of Aquatic Natural Organic Matter, PhD Thesis, Department of Civil Engineering, University of Alberta. 86. Wong, J.T.F., Baker, G.B., Coutts, R.T. 1988. Rapid and Simple Procedure for the Determination of Urinary Phenylacetic Acid Using Derivatisation in Aqueous Medium Followed By Electron Capture Gas Chromatography. J. Chromatog., 428, 140-146. 87. Cipparone, L.A., Diehl, A.C., Speitel, J.E. 1997. Ozonation and BDOC Removal: Effect on Water Quality. J. Am. Water Works Assoc., 89(2), 84-97. 88. Anderson, L.J., Johnson, J.D., Christman, R.F. 1986. Extent of Ozone’s Reaction with Isolated Aquatic Fulvic Acid. Environ. Sci. Technol., 20, 739-742. 89. Smith, M.E., Cowman, G.A., Singer, P.C. 1993. The Impact of Ozonation and Coagulation on Disinfection By-product Formation. Proc. AWWA Annual Conf., San Antonio, Texas, 451-473. 90. Sayato, Y., Nakamuro, K., Ueno, H. 1989. Mutagenicity on Chlorination of Products Formed By Ozonation of Naphthoresorcinol in Water. Mutation Res., 226, 151-155. 91. Peldszus, S., Huck, P.M., Andrews, S.A. 1996. Determination of Short-chain Aliphatic, Oxo- and Hydroxy-acids In Drinking Water at Low Microgram per Litre Concentrations. J. Chromatog. A, 723, 27-34. 92. Cooper, J.C., Zika, R.G., Steinhauer, S. 1985. Bromide-Oxidant Interactions and THM Formation: A Literature Review. J. Am. Water Works Assoc., 77(4), 116-121. 93. Morrison, R.T., Boyd, R.N. Eds., Organic Chemistry 3rd Ed., 1973. Allyn and Bacon Inc., Chapter 26, p853. 94. Westrick, J.A., DeMarco, J., Metz, D.H., Hartman, D.J., Pohlman, R.C. 1998. An Evaluation of the Relative Percentage of Individual THM and HAA6 Species for Cincinatti Water Works and Potential Water Treatment Processes. Proc. AWWA Water Quality Technology Conference, San Diego, California. 95. Andrews, S.A., Huck, P.M. 1993. Identification of Disinfection By-Product Precursors Using Fractionated NOM. Proc. AWWA Water Quality Technology Conference, Miami, Fl. 177-189. 96. Miller, J.W., Uden, P.C. 1983. Characterisation of Non-volatile Aqueous Chlorination Products of Humic Substances. Environ. Sci. Technol., 17, 150-157. 97. Glaze, W.H., Weinberg, H.S., Cavanagh, J.H. 1993. Evaluating the Formation of Brominated DBPs During Ozonation, J. Am. Water Works Assoc., 85(1), 96-103. 98. Wu, W.W., Chadik, P.A. 1998. Effect of Bromide on HAA Formation During Chlorination of Biscayne Aquifer Water. J. Environ. Eng., ASCE, 124, 932-938. 99. Cowman, G.A., Singer, P.C. 1994. Effect of Preozonation on HAA Speciation in Chlorinated Waters Containing Bromide. Proc. AWWA Water Quality Technology Conference, San Francisco, CA. 947-964. 100. Reckhow, D.A., Singer, P.C., Malcolm, R.L. 1990. Chlorination of Humic Materials: Byproduct Formation and Chemical Interpretation. Environ. Sci. Technol., 24, 1655-1664. 101. Becke, C., Maier, D., Sontheimer, H. 1984. Formation of Chloropicrin in Drinking Water. Vom Wasser, 62, 125.
86 CRC for Water Quality and Treatment Research Report 9-2001
102. McKnight, A., Reckhow, D.A. 1992. Reactions of Ozonation By-Products with Chlorine and Chloramines. Proc. AWWA Annual Conf., Denver, Colorado. 103. Hoigne, J., Bader, H. 1988. The Formation of Trichloronitromethane (Chloropicrin) and Chloroform in a Combined Ozonation-Chlorination Treatment of Drinking Water. Water Res., 22, 313-319. 104. Environmental Protection Authority, 2000. Stage 2 Microbial and Disinfection Byproducts Federal Advisory Committee Agreement in Principle. Federal Register, 65:251, p83015. 105. Finch, G.R., Black, E.K., Gyurek, L., Belosevic, M. 1994. Ozone Disinfection of Giardia and Cryptosporidium. AWWA Research Foundation and AWWA, Denver, Colorado. 106. Krasner, S.W., Barrett, S.E., Dale, M.S., Hwang, C.J. 1989. Free Chlorine Versus Monochloramine for Controlling Off-Tastes and Off-Odors. J. Am. Water Works Assoc., 81(2), 86-98. 107. Aieta, E.M., Berg, J.D. 1986. A Review of Chlorine Dioxide in Drinking Water Treatment. J. Am. Water Works Assoc., 78(6), 62-72. 108. Pedersen, E.J.III., Urbansky, E.T., Marifias, B.J., Margerum, D.W. 1999. Formation of Cyanogen Chloride from the Reaction of Monochloramine and Formaldehyde. Environ. Sci. Technol. 33, 4239-4249. 109. Amy, G.L., Lo, T., Davis, M.K. 1991. The Effects of Ozonation and Activated Carbon Adsorption on Trihalomethane Speciation. Water Res., 25, 191-202. 110. Dore, M., Merlet, N., Legube, B., Croue, J-P. 1988. Interactions Between Ozone, Halogens, and Organic Compounds. Ozone Sci. Eng., 10, 153. 111. Zhou, X., Reckhow, D.A., Tobiason, J.E. 1992. Formation and Removal of Aldehydes in Drinking Water Treatment Processes. Proc. AWWA Water Quality Technology Conference, Denver, CO., 291-315. 112. McKnight, A., Reckhow, D.A. 1992. Reactions of Ozonation By-products with Chlorine and Chloramines. Proc. AWWA Annual Conf., Denver CO., 399-402.
87 CRC for Water Quality and Treatment Research Report 9-2001
APPENDIX I
ADDITIONAL METHODS DEVELOPMENT - ALDEHYDES
I.1 Standard solutions
O-(2,3,4,5,6-Pentafluorobenzyl)hydroxylamine hydrochloride (PFBHA-HCl) was prepared immediately before use gravimetrically as a 10mg/mL aqueous solution. Primary stock standard solutions (5g/L) were prepared gravimetrically for each aldehyde by injecting the liquid analytes directly into the bulk of the methanol and determining exact weight after addition. Formaldehyde (37% solution in water), glyoxal (40% solution in water) and methyl glyoxal (40% solution in water) were prepared in water. Due to the high volatility of acetaldehyde, the neat solution was kept in a freezer at all times and the measurement syringe placed in a freezer for approximately 10 minutes before preparing this standard.
The stock solutions were stored at 4oC and prepared fresh every three months. Secondary stocks (40mg/L) and a standard mixture (1mg/L) were prepared in water by volumetrically diluting the primary stocks and were freshly prepared every month. Five calibration standards were prepared ranging in concentrations from 2, 5, 10, 50 and 100µg/L respectively for each analyte. Table I.1 summarises the retention times for each oxime derivative. The GC operating conditions are shown in Table I.2.
Table I.1: GC retention data for oxime derivatives
# Analyte* LOQ (µg/L) RTDB-1 RTDB-1701 (min.) (min.) 1,2-Dibromopropane - 8.59 5.12 Formaldehyde 1.0 14.49 9.75 E-Acetaldehyde 0.5 18.24 12.89 Z-Acetaldehyde 0.5 18.62 13.16 E-Propionaldehyde 0.5 21.55 15.65 Z-Propionaldehyde 0.5 21.85 15.90 E-Butyraldehyde 0.5 24.89 18.56 Z-Butyraldehyde 0.5 25.12 18.82 E-Heptaldehyde 0.5 34.70 27.34 Z-Heptaldehyde 0.5 34.79 27.50 Benzaldehyde 0.5 38.33 32.10 E-Glyoxal 0.5 44.96 39.14 Z-Glyoxal 0.5 45.20 39.24 Methyl Glyoxal 0.5 45.75 39.92 * The aldehydes listed above are the oxime isomers formed upon derivatisation # LOQ: Limit of quantitation RT: Retention times in minutes.
88 CRC for Water Quality and Treatment Research Report 9-2001
Table I.2: Gas chromatographic conditions for the determination of low molecular weight aldehydes.
Parameter Value Column A Type DB-1 (J&W Scientific) Film thickness 1 micron Internal Diameter 0.25 mm Length 30 m Column B Type DB-1701 (J&W Scientific) Film thickness 0.25 micron Internal Diameter 0.25 mm Length 30 m Initial column temperature 50oC for 1 min. Temperature program 1 4oC/min. to 220oC Temperature program 2 20oC/min. to 275oC with 4min. hold time Injector Varian 1078 temperature programmable Temperature 200oC Splitless injection Split valve opened at 0.6 min. Injection volume 1.0 µL Detector Type Electron capture, 63Ni Temperature 300oC Gases Carrier gas Hydrogen (99.9999% purity) Carrier flow rate 2.0 mL/min Detector makeup gas Nitrogen (99.999% purity) Detector makeup flow rate 28 ± 2 mL/min Autosampler pneumatics Compressed air (industrial grade)
89 CRC for Water Quality and Treatment Research Report 9-2001
APPENDIX II ADDITIONAL METHODS DEVELOPMENT – KETOACIDS
II.1 Standard solutions PFBHA-HCl was prepared as indicated previously in Appendix I. Primary stock solutions (5g/L) were prepared gravimetrically in methanol. Glyoxylic acid is supplied as a 50% by weight aqueous solution and was prepared by injecting the analyte directly into the bulk of the methanol and determining exact weight by difference after addition. Diluted solutions and calibration standards were prepared and stored as described in appendix I.1. Diazomethane (CH2N2) was generated by the base-catalysed decomposition of N-Nitroso-N-methyl-4-toluenesulfonamide. Diazomethane, once generated, was transferred by bubbling nitrogen through an outlet tube to a crimp-capped glass bottle containing MTBE. The diazomethane solution (100mL) was teflon sealed and stored in the freezer.
Table II.1: GC retention data for oxime derivatives
# Analyte* LOQ (µg/L) RTDB-1 (min.) RTDB-1701 (min.) 1,2-Dibromopropane - 8.82 4.76 Glyoxylic acid 0.5 20.58 17.62 Pyruvic acid 0.5 21.26 17.99 2-Ketobutyric acid 0.5 21.97 18.55 Ketomalonic acid 0.5 23.83 21.32 Oxalacetic acid Qualitativea 24.79 22.15 * The ketoacids listed above are the methylated oximes formed upon derivatisation # LOQ: Limit of quantitation a oxalacetic acid was qualitatively monitored
Table II.2: Gas chromatographic temperature program for the determination of low molecular weight ketoacids.
Parameter Value Initial column temperature 50oC for 1 min. Temperature program 1 5oC/min. to 100oC Temperature program 2 10oC/min. to 220oC Temperature program 3 20oC/min. to 280oC with a 3min. hold time Injection volume 1.0 µL
II.2 Preparation of Diazomethane in Methyl tert-Butyl Ether (MTBE).
O 3 N CH - + N O K OOS OOS KOH
CH3OCH2CH2OH + CH2N2 B
CH3 CH3 A
90 CRC for Water Quality and Treatment Research Report 9-2001
Warning: Diazomethane is a very toxic yellow gas. Concentrated solutions may explode violently, especially if impurities are present. Ground glass apparatus and rough glass surfaces should not be used. Diazomethane is an insidious poison and extreme care must be taken with safety glasses, gloves and a well ventilated hood is absolutely necessary.
II.2.1 Diazomethane (B) is generated by the base-catalysed decomposition of N-Nitroso-N methyl-4- toluenesulfonamide (pTSN, A).
II.2.2 Approximately 8 mL of pTSN in methoxyethanol (20% w/w of pTSN in methoxyethanol) was slowly added from a dropping funnel to a flask containing 20 mL of a KOH methanolic solution (10% w/w) at 40oC under nitrogen. The stirred reaction mixture becomes opaque yellow when diazomethane is generated. Diazomethane, once generated, is transferred by bubbling nitrogen through an outlet tube to a glass bottle containing 100 mL of pure MTBE with the tube located below the surface of the solvent. The solution is teflon sealed and stored in the freezer.
91 CRC for Water Quality and Treatment Research Report 9-2001
APPENDIX III MASS SPECTRA AND IDENTIFICATION OF CARBOXYLIC ACIDS
Figure III.1: Mass spectrum of pentafluorophenylformate.
Figure III.2: Mass spectrum of pentafluorophenylacetate.
92 CRC for Water Quality and Treatment Research Report 9-2001
Figure III.3: Mass spectrum of pentafluorophenylpropionate.
Figure III.4: Mass spectrum of pentafluorophenylbenzoate.
93 CRC for Water Quality and Treatment Research Report 9-2001
APPENDIX IV
IV.1 RAW DATA – OZONATION BY-PRODUCT FORMATION - ALDEHYDES
Myponga water Hope Valley water Tod water Raw 0.5:1 1:1 2:1 3:1 5:1 raw 0.5:1 1:1 2:1 3:1 5:1 raw 0.5:1 1:1 2:1 3:1 5:1 Formaldehyde 3.0 4.6 6.1 118 141 167 6.7 12 27 47 62 79 14 10 26 50 60 57 Acetaldehyde 1.7 2.6 3.2 28 27 26 4.4 4.6 24 21 26 25 12 12 14 25 30 28 Propionaldehy 1.7 2.0 2.4 5.9 4.7 3.7 2.9 3.3 4.5 5.2 5.0 4.3 4.3 5.4 5.3 6.7 7.5 6.4 de Butyraldehyde 3.3 3.2 3.6 4.5 3.7 3.6 7.2 7.0 7.2 7.5 7.6 7.5 9.4 11 10 11 11 10 Heptaldehyde <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 <1 Benzaldehyde <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 <0.5 0.8 0.6 0.7 Glyoxal 0.5 11 22 47 52 61 <0.5 3.5 9.0 16 22 28 0.5 2.2 4.5 14 17 18 Methyl glyoxal 0.5 7.2 15 31 42 50 0.8 3.2 7.8 14 16 17 0.6 3.6 5.3 12 12 13
NB: All values in µg/L.
94 CRC for Water Quality and Treatment Research Report 9-2001
IV.2 RAW DATA – OZONATION BY-PRODUCT FORMATION - KETOACIDS
Myponga water Hope Valley water Tod water Raw 0.5:1 1:1 2:1 3:1 5:1 raw 0.5:1 1:1 2:1 3:1 5:1 raw 0.5:1 1:1 2:1 3:1 5:1 Glyoxylic 3.7 60 80 120 170 206 0.8 14 32 55 91 120 2.1 3.1 4.7 87 91 51 Pyruvic 2.0 47 54 135 243 399 2.8 7.7 44 82 138 210 3.8 6.1 7.3 88 51 13 2-ketobutyric <0.5 1.0 3.0 4.0 7.0 10 <0.5 <0.5 1.2 2.6 3.2 5.2 <0.5 <0.5 1.0 3 3.6 2.3 Ketomalonic 2.0 47 54 114 124 144 0.8 4.7 7.6 14 29 54 0.8 3.0 4.0 13 10 7.8
NB: All values in µg/L
95 CRC for Water Quality and Treatment Research Report 9-2001
APPENDIX V
V.1 RAW DATA – CHLORINATION BY-PRODUCT FORMATION - TRIHALOMETHANES
Myponga water Hope Valley water Tod water Raw 0.5:1 1:1 2:1 3:1 5:1 raw 0.5:1 1:1 2:1 3:1 5:1 raw 0.5:1 1:1 2:1 3:1 5:1
CHCl3 443 361 309 240 215 180 109 86 70 55 58 48 7.8 5.6 5.1 3.4 3.9 3.9
CHCl2Br 181 180 170 161 137 114 90 75 73 61 61 42 75 47 36 26 29 29
CHClBr2 42 62 72 70 64 46 43 47 52 41 37 20 373 270 185 127 134 124
CHBr3 2 4 6 7 7 3 3.9 5.5 7.7 6.4 5.6 2.6 534 463 375 304 336 388
NB: All values in µg/L.
V.2 RAW DATA – CHLORINATION BY-PRODUCT FORMATION – CHLOROACETIC ACIDS
Myponga water Hope Valley water Tod water Raw 0.5:1 1:1 2:1 3:1 5:1 raw 0.5:1 1:1 2:1 3:1 5:1 raw 0.5:1 1:1 2:1 3:1 5:1 MCAA 49 46 45 46 45 43 10 11 10 9.9 9.8 8.4 4 <2.5 <2.5 <2.5 <2.5 <2.5 DCAA 230 226 208 198 198 190 90 85 80 69 70 70 15 12 11 10 12 12 TCAA 226 178 132 84 64 42 127 85 65 46 37 31 5 3.2 2.5 1.5 1.2 1.6
NB: All values in µg/L.
96 The Cooperative Research Centre for Water Quality and Treatment is an unincorporated joint venture between:
• ACTEW Corporation • Australian Water Quality Centre • Australian Water Services Pty Ltd • Brisbane City Council CRC for Water Quality and • Centre for Appropriate Treatment Technology Inc Private Mail Bag 3 • City West Water Limited Salisbury SOUTH AUSTRALIA 5108 • CSIRO Tel: (08) 8259 0351 • Curtin University of Technology Fax: (08) 8259 0228 • Department of Human Services E-mail: [email protected] Victoria Web: www.waterquality.crc.org.au • Griffith University • Melbourne Water Corporation • Monash University • Orica Australia Pty Ltd • Power and Water Corporation • Queensland Health Pathology & Scientific Services • RMIT University • South Australian Water Corporation • South East Water Ltd • Sydney Catchment Authority • Sydney Water Corporation • The University of Adelaide • The University of New South Wales The Cooperative Research Centre (CRC) for Water Quality and Treatment is Australia’s national drinking water research • The University of Queensland centre. An unincorporated joint venture between 29 different • United Water International Pty Ltd organisations from the Australian water industry, major universities, CSIRO, and local and state governments, the CRC • University of South Australia combines expertise in water quality and public health. • University of Technology, Sydney
The CRC for Water Quality and Treatment is established • Water Corporation Researchand supported under the Federal Government’s Report Cooperative • Water Services Association Research Centres Program. of Australia • Yarra Valley Water Ltd