Application of the UV/Chlorine Advanced Oxidation …...advice on UV technology since my Master’s study. I appreciate Prof. Susan Andrewsalso for being on my supervisory committee

Application of the UV/Chlorine Advanced Oxidation Process

for Drinking Water Treatment

by

Ding Wang

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy

Graduate Department of Civil Engineering

University of Toronto

© Copyright by Ding Wang 2015

ii

Application of the UV/Chlorine Advanced Oxidation Process

for Drinking Water Treatment Ding Wang

Doctor of Philosophy, 2015

Graduate Department of Civil Engineering

University of Toronto

ABSTRACT

This research investigated the feasibility of a novel advanced oxidation process (AOP)

using ultraviolet light combined with free chlorine (UV/chlorine) in drinking water treatment.

A bench-scale study using a medium pressure UV collimated beam apparatus showed that

the UV/chlorine process was more efficient than the UV combined with hydrogen peroxide

(UV/H2O2) AOP for the destruction of trichloroethylene (TCE) at pH 5 in a laboratory prepared

water, but was less efficient than the latter at pH 7.5 and 10. A Matlab® mathematical model

made accurate predictions of the observed experimental rates of TCE decay. The model

predicted that increasing concentrations of hydroxyl radical scavengers in the treated water

would tend to raise the pH at which UV/chlorine would remain competitive relative to

UV/H2O2.

Full-scale experiments at the City of Cornwall Water Purification Plant (Ontario, Canada)

and pilot-scale tests in a Rayox® batch UV reactor using water from the Keswick Water

Treatment Plant (Ontario, Canada) demonstrated comparable performance of UV/chlorine AOP

to UV/H2O2 for geosmin, 2-methylisoborneol (MIB), and caffeine destruction.

iii

Organic and inorganic disinfection by-products (DBPs) were also monitored in the full-

and pilot-scale tests. Minimal trihalomethane and haloacetic acid formation was observed across

the UV reactor, while dichloroacetonitrile and bromochloroacetonitrile were produced rapidly,

although overall concentrations were below 6 µg L–1. Adsorbable organic halide was formed

rapidly (up to 70 µg Cl L–1) in water that had not been prechlorinated, while little formation was

observed in previously chlorinated water. Chlorate and bromate were formed, equivalent to

approximately 2–17% and 0.01–0.05% of the photolyzed chlorine, respectively, while no

perchlorate or chlorite formation was observed. In addition, the 24 h organic DBP formation

potential was increased by UV/chlorine pretreatment to an extent that was similar to that

observed when the water was pretreated with UV/H2O2.

iv

ACKOWLEDGEMENTS

I would like to express my sincere gratitude to my supervisor, Professor Ron Hofmann, for

his suggestions, guidance, encouragement, understanding, patience, and support throughout my

research. I would like to thank Prof. James R. Bolton for his continuously invaluable help and

advice on UV technology since my Master’s study. I also appreciate Prof. Susan Andrews for

being on my supervisory committee and providing excellent suggestions in my research. I am

grateful to Prof. Robert Andrews for his encouragement.

I also appreciate Leigh McDermott from Stantec Consulting Ltd., Owen O'Keefe, Daniel

Drouin, and Morris McCormick from Cornwall Water Purification Plant, Dr. Vasile Furdui from

the Ontario Ministry of Environment and Climate Change, Dr. Hong Zhang, Dr. A.H.M. Anwar

Sadmani, Zhen (Jim) Wang, and Jiafan (Kevin) Yang from Drinking Water Research Group for

their great help in my research.

Last but not least, I would like to give my special thanks to my wife and parents for their

endless support and encouragement.

This work was financially supported by Natural Sciences and Engineering Research

Council of Canada (NSERC) through the Engage Grant program and the Industrial Research

Chair program, and by Stantec Consulting Ltd. and the Centre for Control of Emerging

Contaminants (CCEC).

v

TABLE OF CONTENTS

ABSTRACT.............................................................................................................................. ii

ACKOWLEDGEMENTS ....................................................................................................... iv

TABLE OF CONTENTS ..........................................................................................................v

LIST OF TABLES .................................................................................................................. ix

LIST OF FIGURES ..................................................................................................................x

LIST OF ACRONYMS .......................................................................................................... xii

1. INTRODUCTION .............................................................................................................1

References ..............................................................................................................................2

2. LITERATURE REVIEW .................................................................................................5

2.1 Advanced Oxidation Processes (AOPs) ..........................................................................5

2.1.1 Theory of Advanced Oxidation Processes (AOPs) ......................................................5

2.1.2 Miscellaneous Methods for AOPs ...............................................................................7

2.2 UV Combined with Chlorine as an AOP ........................................................................7

2.2.1 Fundamental Chemistry of Aqueous Free Chlorine .....................................................7

2.2.2 Fundamental Chemistry of the UV/Chlorine AOP..................................................... 10

2.3 Formation of Disinfection By-Products (DBPs) in UV/Chlorine ................................. 17

2.3.1 Chlorinated DBPs ..................................................................................................... 17

2.3.2 DBP Formation Kinetics ........................................................................................... 19

2.3.3 DBP Formation by UV and AOPs ............................................................................. 22

2.4 Summary of Literature Review .................................................................................... 25

References ............................................................................................................................ 26

3. MEDIUM PRESSURE UV COMBINED WITH CHLORINE ADVANCED

OXIDATION FOR TRICHLOROETHYLENE DESTRUCTION IN A MODEL

WATER ................................................................................................................................... 36

Abstract ............................................................................................................................... 36

3.1 Introduction ................................................................................................................... 36

3.2 Materials and Methods .................................................................................................. 40

3.2.1 Reagents and Materials ............................................................................................. 40

3.2.2 UV Exposure and Irradiance Measurements .............................................................. 40

vi

3.2.3 Analytical Methods ................................................................................................... 41

3.3 Results and Discussion .................................................................................................. 41

3.3.1 Molar Absorption Coefficients of TCE, Active Chlorine, Peroxide and Hydroxide

Species .............................................................................................................................. 41

3.3.2 Quantum Yields of Active Chlorine and Hydrogen Peroxide Photolysis.................... 42

3.3.3 TCE Decay Rates by UV Alone, and the UV/Chlorine and the UV/H2O2 AOPs ........ 44

3.3.4 Mathematical Modeling of the TCE Decay ............................................................... 51

3.3.5 Comment on Active Chlorine Reaction with •OH ..................................................... 52

3.4 Conclusions .................................................................................................................... 54

Acknowledgements .............................................................................................................. 55

References ............................................................................................................................ 55

4. FULL-SCALE COMPARISON OF ULTRAVIOLET/CHLORINE ADVANCED

OXIDATION TO ULTRAVIOLET/HYDROGEN PEROXIDE FOR TASTE AND

ODOUR CONTROL IN DRINKING WATER TREATMENT .......................................... 60

Abstract ............................................................................................................................... 60

4.1 Introduction ................................................................................................................... 60

4.2 Material and Methods ................................................................................................... 62

4.2.1 Reagents and Materials ............................................................................................. 62

4.2.2 Experimental Facilities and Procedures ..................................................................... 62

4.2.3 Sample Analysis ....................................................................................................... 64

4.3 Results and Discussion .................................................................................................. 65

4.3.2 Free Chlorine Decay ................................................................................................. 65

4.3.3 Geosmin and MIB Decay .......................................................................................... 66

4.3.4 Caffeine Decay ......................................................................................................... 70

4.3.5 Electrical Energy per Order (EEO) ............................................................................. 71

4.3.6 Comment on Chlorine Radical (•Cl) and Disinfection By-Product (DBP)

Formation during Chlorine Photolysis ............................................................................... 73

4.4 Conclusions .................................................................................................................... 74


References ............................................................................................................................ 74

vii

5. FORMATION OF DISINFECTION BY-PRODUCTS IN THE

ULTRAVIOLET/CHLORINE ADVANCED OXIDATION PROCESS .............................. 78

Abstract ............................................................................................................................... 78

5.1 Introduction ................................................................................................................... 78

5.2 Material and Methods ................................................................................................... 80

5.2.1 Experimental Procedures .......................................................................................... 80

5.2.2 Analytical Methods ................................................................................................... 83

5.3 Results ............................................................................................................................ 84

5.3.1 THMs ....................................................................................................................... 89

5.3.2 HAAs ....................................................................................................................... 89

5.3.3 HANs, HKs, and CP ................................................................................................. 90

5.3.4 AOX ......................................................................................................................... 90

5.3.5 Inorganic DBPs: ClO4–, ClO3

–, ClO2–, and BrO3

– ...................................................... 92

5.4 Discussion....................................................................................................................... 94

5.4.1 Rapid DBP Formation within the UV/Chlorine Reactor ............................................ 94

5.4.2 Impact of UV/Chlorine on 24 Hour DBP-FP ............................................................. 95

5.4.3 Role of the Chlorine Radical (•Cl) ............................................................................ 95

5.5 Conclusions .................................................................................................................... 96


References ............................................................................................................................ 97

6. SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS ................................. 102

6.1 Summary and Conclusions .......................................................................................... 102

6.2 Recommendations for Future Work ........................................................................... 103

APPENDICES ....................................................................................................................... 105

A. Pulse Radiolysis Analysis for Determination of Rate Constant of Free Chlorine

with Hydroxyl Radical ...................................................................................................... 105

B. Determination of the Fluence Rate of the MP Lamp in the Collimated Beam

Apparatus .......................................................................................................................... 108

C. Example of Matlab® Codes ......................................................................................... 110

C.1 Matlab® Codes for Simulation of Trichloroethylene Decay by the UV/Chlorine

AOP at 11 mg L–1 and pH 5 ............................................................................................. 110

viii

C.2 Matlab® Codes for Simulation of Trichloroethylene Decay by the UV/Chlorine

AOP at 11 mg L–1 and pH from 5 to 10............................................................................ 112

D. Estimation of OH Radical Concentration Using an Excel Spreadsheet ..................... 113

E. UV Dose Estimation Using UVCalc® Version 2B ........................................................ 115

F. Absorption Spectra of Geosmin, MIB, and Caffeine ................................................... 116

G. Sample Analysis ............................................................................................................ 118

G.1 Geosmin and MIB..................................................................................................... 118

G.2 Caffeine .................................................................................................................... 120

G.3 Trihalomethanes (THMs), Haloacetonitriles (HANs), Haloketones (HKs),

Chloropicrin (CP), and Trichloroethylene (TCE) ............................................................. 125

G.4 Haloacetic Acids (HAAs) ......................................................................................... 129

G.5 Chlorate .................................................................................................................... 134

H. Quality Assurance / Quality Control (QA/QC) ........................................................... 135

I. Experimental Data ......................................................................................................... 140

ix

LIST OF TABLES

Table 2.1 Application of AOPs in different areas and for different contaminants ....................... 6

Table 2.2 Quantum yields of OCl– decay and photochemical product formation .......................16

Table 2.3 Halogenated by-products in drinking water treatment ...............................................20

Table 3.1 Fluence-based rate constants (10–6 m2 J–1) for active chlorine and peroxide

photolysis .................................................................................................................44

Table 3.2 Comparison of reported quantum yields of active chlorine photolysis .......................45

Table 3.3 TCE Fluence-based decay rate constants (10–4 cm2 mJ– 1) (excluding evaporation)

by UV alone, and the UV/chlorine and the UV/H2O2 AOPs from experimental

and model results .....................................................................................................48

Table 3.4 Calculated hydroxyl radical concentrations (10–13 M) in TCE solutions treated by

the UV/chlorine and the UV/H2O2 AOPs at various pH values .................................50

Table 3.5 Reaction mechanisms of TCE decay by UV alone, the UV/chlorine and the

UV/H2O2 AOPs ........................................................................................................53

Table 4.1 Post-filtration water quality parameters for full- and pilot-scale tests ........................63

Table 4.2 Full- and pilot-scale EEO values (kWh m–3 order–1) for geosmin, MIB, and

caffeine removal .......................................................................................................72

Table 5.1 Full- and pilot-scale post-filtration water quality parameters .....................................80

Table 5.2 Monitored organic and inorganic DBPs ....................................................................83

x

LIST OF FIGURES

Figure 2.1 Percentage distribution of HOCl as a function of pH ................................................ 9

Figure 2.2 Molar absorption coefficients of HOCl, OCl–, and NH2Cl .......................................11

Figure 3.1 Absorption spectra of TCE, chlorine, peroxide and hydroxide species .....................38

Figure 3.2 Relative spectral emittance of the MP lamp in this research .....................................38

Figure 3.3 Rates of active chlorine (a) and hydrogen peroxide (b) photolysis at various pH

values. Error bars represent the standard deviations of triplicate runs. Straight

lines represent the linear regression. ........................................................................43

Figure 3.4 Experimental TCE decay rates by UV alone, the UV/chlorine and the UV/H2O2

AOPs at various pH values. Error bars represent the standard deviations of

triplicate runs. Straight lines represent the linear regression.....................................47

Figure 3.5 Solution pH at which the UV/chlorine and the UV/H2O2 AOPs are equally

efficient as a function of TOC concentration ...........................................................54

Figure 4.1 Full-scale Trojan UVSwift reactor (left) and pilot-scale Rayox® reactor (right) .......63

Figure 4.2 Percentage of free chlorine photolysis by UV exposure. Error bars represent the

values of experimental duplicates. ...........................................................................66

Figure 4.3 Geosmin (top plot) and MIB (bottom plot) decay in the 1st full-scale test. Error

bars represent the values of experimental duplicates. ...............................................67

Figure 4.4 Geosmin decay in the 2nd full-scale test. Error bars represent the values of

experimental duplicates. ..........................................................................................69

Figure 4.5 Caffeine decay in the 1st full-scale (top plot) and pilot-scale (bottom plot) tests.

Error bars represent the values of experimental duplicates.......................................71

Figure 5.1 Full-scale Trojan UVSwift reactor (left) and pilot-scale Rayox® reactor (right) .......81

Figure 5.2 THM formation in full- and pilot-scale tests. Plots on the left show THMs after

various treatment processes for short reaction time (30–60 s contact time). Plots

on the right show THM formation potentials (free chlorine dose: 6.5 mg L–1 for

24 h) in the water pretreated by selected processes shown on the x-axis. Error

bars represent the values of the duplicates measured. ..............................................85

xi

Figure 5.3 HAA formation in full- and pilot-scale tests. Plots on the left show HAAs after


on the right show HAA formation potentials (free chlorine dose: 6.5 mg L–1 for



Figure 5.4 HAN formation in full- and pilot-scale tests. Plots on the left show HANs after


on the right show HAN formation potentials (free chlorine dose: 6.5 mg L–1 for



Figure 5.5 AOX formation in full- and pilot-scale tests. Plots on the left show AOX after


on the right show AOX formation potentials (free chlorine dose: 6.5 mg L–1 for



Figure 5.6 Formation of ClO3– relative to free chlorine photodecomposition in the full- and

pilot-scale experiments. Error bars represent the values of the duplicates

measured. Low, medium, and high represent free chlorine doses of 2, 6, and 10

mg L–1, respectively. ...............................................................................................93

xii

LIST OF ACRONYMS

AOP Advanced oxidation process

BCAA Bromochloroacetic acid

BCAN Bromochloroacetonitrile

BDCAA Bromodichloroacetic acid

BDCM Bromodichloromethane

Br– Bromide ion

BrO3– Bromate ion

Ca(OCl)2 Calcium hypochlorite

CDBAA Chlorodibromoacetic acid

CDBM Chlorodibromomethane

CI Chemical ionization

•Cl Chlorine radical

Cl– Chloride ion

Cl2 Chlorine molecule

ClO2– Chlorite ion

ClO3– Chlorate ion

ClO4– Perchlorate ion

CO32– Carbonate ion

CP Chloropicrin

DBAA Dibromoaceic acid

DBAN Dibromoacetonitrile

DBP Disinfection by-product

DCAA Dichloroacetic acid

DCAN Dichloroacetonitrile

DCP 1,1-Dichloropropanone

EEO Electrical energy per order

EI Electron ionization

FP Formation potential

xiii

GC-ECD Gas chromatography-electron capture detector

GC-MS Gas chromatography-ion-trap mass spectrometry

HAA Haloacetic acid

HAN Haloacetonitrile

HCO3– Bicarbonate ion

H2CO3 Carbonic acid

HK Haloketone

H2O2 Hydrogen peroxide

HOBr Hypobromous acid

HOCl Hypochlorous acid

HOI Hypoiodous acid

H2SO4 Sulphuric acid

HS-SPME Headspace solid phase micro-extraction

IC Ion chromatograph

IC-MS/MS Ion chromatograph tandem mass spectrometer

LP Low pressure

MAC Maximum acceptable concentration

MBAA Monobromoacetic acid

MCAA Monochloroacetic acid

MCL Maximum contaminant level

MDL Method detection limit

MIB 2-Methylisoborneol

MP Medium pressure

MTBE Methyl tertiary-butyl ether

NaOCl Sodium hypochlorite

NaOH Sodium hydroxide

Na2SO3 Sodium sulphite

NB Nitrobenzene

NDMA N-nitrosodimethylamine

NDRL Notre Dame Radiation Laboratory

NH4Cl Ammonium chloride

xiv

(NH4)2SO4 Ammonium sulphate

OCl– Hypochlorite ion

•OH Hydroxyl radical

OH– Hydroxide ion

pCBA para-Chlorobenzoic acid

SPE Solid-phase extraction

TBAA Tribromoacetic acid

TBAN Tribromoacetonitrile

TBM Bromoform/tribromomethane

TCAA Trichloroacetic acid

TCAN Trichloroacetonitrile

TCE Trichloroethylene

TCM Chloroform/trichloromethane

TCP 1,1,1-Tichloropropanone

THM Trihalomethane

TOC Total organic carbon

TOX Total organic halides

UV Ultraviolet

VOC Volatile organic compound

VUV Vacuum-UV

WHO World Health Organization

1

1. INTRODUCTION

Free chlorine (hypochlorous acid, HOCl, and hypochlorite ion, OCl–) exposed to UV light

at wavelengths ranging from 200 to 400 nm (UV/chlorine) is a potential advanced oxidation

process (AOP). An AOP refers to a water or wastewater treatment process that produces and

utilizes hydroxyl radicals (•OH) to destroy contaminants that are not readily oxidized by

conventional oxidants (Nowell and Hoigné, 1992a; Watts and Linden, 2007; Parsons, 2004;

Gültekin and Ince, 2007; Suty et al., 2004). •OH has been proven to be a very strong oxidizing

agent that reacts with many reductants extremely rapidly (Svrcek and Smith, 2004;

Tchobanoglous et al., 2003). Since HOCl is a weak acid, with a pKa of 7.54 at 25°C (Deborde

and von Gunten, 2008), free chlorine contains both HOCl and OCl– in the normal pH range of

drinking water (Health Canada, 2012). Although both HOCl and OCl– photolysis can produce

•OH, the concentrations of •OH generated by HOCl and OCl– at the same molar concentration

are expected to be different based on their different molar absorption coefficients, quantum

yields of •OH formation, and •OH scavenging efficiencies (Feng et al., 2007; Watts and Linden,

2007). This in turn leads to a UV/chlorine efficiency that is sensitive to pH (Watts et al., 2007).

Compared to other commonly used AOPs, such as UV combined with hydrogen peroxide

(UV/H2O2), the UV/chlorine AOP is still novel and not fully explored. Previous studies

indicated that it was potentially an alternative to the UV/H2O2 AOP, since HOCl/OCl– absorbs

UV photons more efficiently than H2O2 when using typical low pressure (LP) or medium

pressure (MP) UV lamps, and produces •OH relatively efficiently under some conditions (Watts

and Linden, 2007; Watts et al., 2007; Feng et al., 2007; Nowell and Hoigné, 1992b; Chan et al.,

2012). However, the mechanisms of chlorine photolysis that are probably associated with a

series of chain reactions with the formation of many intermediates and products (Buxton and

Subhani, 1972a, 1972b, and 1972c) are not completely understood. The quantum yields of

chlorine photolysis and hydroxyl radical formation, which are decisive factors in UV/chlorine

efficiency, are inconsistent among published research papers (Buxton and Subhani, 1972b; Feng

et al., 2007; Watts and Linden, 2007; Jin et al., 2011; Chan et al., 2012). The effectiveness of

this process in drinking water treatment has been only investigated preliminarily and

incompletely.

2

In addition, the application of chlorine can lead to disinfection by-products (DBPs).

Chlorine doses used for the UV/chlorine process (5–10 mg L–1 as free chlorine) are generally

higher than doses normally used for disinfection (0.2–2 mg L–1), but the chlorine contact time in

a practical UV/chlorine AOP can be much shorter than that in disinfection (e.g. seconds/minutes

compared to hours) (Watts et al., 2007; Wang et al., 2011; AWWA, 1999; USEPA, 1999; Sadiq

and Rodriguez, 2004). Under such unique chlorination conditions, DBP formation during the

UV/chlorine treatment could be different from that in chlorine disinfection, but has not been

investigated.

This research therefore (1) explored the mechanisms of chlorine photolysis, (2)

systematically evaluated the UV/chlorine efficiency in the removal of selected contaminants

(trichloroethylene, taste and odour compounds, and caffeine) under different conditions

(chlorine doses and pH) and at different experimental scales (bench-, pilot-, and full-scales), and

(3) investigated the DBP formation during the UV/chlorine process. At each step, UV/chlorine

was compared to UV/H2O2 as a reference. The relevant results are shown in Chapters 3 to 5.

The analytical details and raw data have been included in Appendices G to I.

References American Water Works Association (AWWA), 1999. Water Quality and Treatment: A

Handbook of Community Water Supplies, 5th Edition. McGraw-Hill, Inc., New York, USA.

Buxton, G.V., Subhani, M.S., 1972a. Radiation chemistry and photochemistry of oxychlorine

ions. Part 1.–Radiolysis of aqueous solutions of hypochlorite and chlorite ions. Journal of the

Chemical Society, Faraday Transactions 1: Physical Chemistry in Condensed Phases, 68,

947–957.

Buxton, G.V., Subhani, M.S., 1972b. Radiation chemistry and photochemistry of oxychlorine

ions. Part 2.–Photodecomposition of aqueous solutions of hypochlorite ions. Journal of the


958–969.

Buxton, G.V., Subhani, M.S., 1972c. Radiation chemistry and photochemistry of oxychlorine

ions. Part 3.–Photodecomposition of aqueous solutions of chlorite ions. Journal of the


970–977.

http://www.rsc.org.myaccess.library.utoronto.ca/Publishing/Journals/F1/article.asp?doi=F19726800947






3

Chan, P.Y., Gamal El-Din, M., Bolton, J.R., 2012. A solar-driven UV/chlorine advanced

oxidation process. Water Research, 46(17), 5672–5682.

Deborde, M., von Gunten, U., 2008. Reactions of chlorine with inorganic and organic

compounds during water treatment–kinetics and mechanisms: a critical review. Water

Research, 42(1–2), 13–51.

Feng, Y., Smith, D.W., Bolton, J.R., 2007. Photolysis of aqueous free chlorine species (HOCl

and OCl–) with 254 nm ultraviolet light. Journal of Environmental Engineering and Science,

6(3), 277–284.

Gültekin, I., Ince, N.H., 2007. Synthetic endocrine disruptors in the environment and water

remediation by advanced oxidation processes. Journal of Environmental Management, 85(4),

816–832.

Health Canada, 2012. Guidelines for Canadian Drinking Water Quality – Summery Table.

Water, Air and Climate Change Bureau, Healthy Environments and Consumer Safety Branch,

Health Canada, Ottawa, Ontario, Canada.

Jin, J., Gamal El-Din, M., Bolton, J.R., 2011. Assessment of the UV/chlorine process as an

advanced oxidation process. Water Research, 45(4), 1890–1896.

Parsons, S., 2004. Advanced Oxidation Processes for Water and Wastewater treatment. IWA

Publishing, London, UK.

Nowell, L.H., Hoigné, J., 1992a. Photolysis of aqueous chlorine at sunlight and ultraviolet

wavelengths–I. Degradation rates. Water Research, 26(5), 593–598.

Nowell, L.H., Hoigné, J., 1992b. Photolysis of aqueous chlorine at sunlight and ultraviolet

wavelengths–II. Hydroxyl radical production. Water Research, 26(5), 599–605.

Sadiq, R., Rodriguez, M.J., 2004. Disinfection by-products (DBPs) in drinking water and

predictive models for their occurrence: a review. Science of the Total Environment, 321(1–3),

21–46.

Suty, H., De Traversay, C., Cost, M., 2004. Applications of advanced oxidation processes:

present and future. Water Science and Technology, 49(4), 227–233.

Svrcek, C., Smith, D.W., 2004. Cyanobacteria toxins and the current state of knowledge on

water treatment options: a review. Journal of Environmental Engineering and Science, 3(3),

155–185.

4

Tchobanoglous, G., Burton, F.L., Stensel, H.D., 2003. Wastewater Engineering, Treatment and

Reuse, 4th edition. The McGraw-Hill Companies, Inc., New York.

United States Environmental Protection Agency (USEPA), 1999. Alternative disinfectants and

oxidants guidance manual. Available from <http://www.epa.gov/ogwdw000/mdbp/alternative

_disinfectants_guidance.pdf > (accessed 07.31.14).

Wang, D., Walton, T., McDermott, L., Hofmann, R., 2011. Control of TCE using UV combined

with hydrogen peroxide or chlorine. Proceedings of IOA-IUVA 2011 World Congress &

Exhibition, Paris, France.

Watts, M.J., Linden, K.G., 2007. Chlorine photolysis and subsequent OH radical production

during UV treatment of chlorinated water. Water Research, 41(13), 2871–2878.

Watts, M.J., Rosenfeldt, E.J., Linden, K.G., 2007. Comparative OH radical oxidation using UV-

Cl2 and UV-H2O2 processes. Journal of Water Supply: Research and Technology–AQUA,

56(8), 469–477.

5

2. LITERATURE REVIEW

2.1 Advanced Oxidation Processes (AOPs)

2.1.1 Theory of Advanced Oxidation Processes (AOPs)

Advanced oxidation processes (AOPs) involve the generation and application of hydroxyl

radicals (•OH), which have a very high oxidation potential of 2.80 V (Parsons, 2004). AOPs are

used in water and wastewater treatment to destroy refractory compounds that cannot be oxidized

by conventional oxidants such as oxygen, ozone, and chlorine (AWWA, 1999; Bolton, 2010;

Gültekin and Ince, 2007; Parsons, 2004; Pera-Titus et al., 2004; Svrcek and Smith, 2004;

Tchobanoglous et al., 2003). AOPs have been proven to be very effective in treating a variety of

dissolved organic contaminants in water, since the short-lived •OH can oxidize almost all

reducing materials including a large group of organic chemicals at very high reaction rates (in

the range of 108 to 1010 M–1 s–1) at normal pressure and temperature through non-selective

pathways (Rosenfeldt et al., 2006; Bolton, 2010; Gültekin and Ince, 2007; Parsons, 2004;

Tchobanoglous et al., 2003). AOPs can be applied in many areas to treat a broad range of

aqueous contaminants, as shown in Table 2.1. The reaction types between hydroxyl radicals and

target compounds are commonly radical addition, hydrogen abstraction, and/or electron transfer

(Tchobanoglous et al., 2003, Wang et al., 2006). After a series of oxidation reactions, the target

compounds convert to less complex intermediate products, or are eventually mineralized to CO2

with sufficient contact time and an enough hydroxyl radical concentration (Bolton, 2010;

Gültekin and Ince, 2007). However, since the life of hydroxyl radicals is very short (in the order

of nanoseconds) and their concentration in water is fairly low (for example, approximately 10–13

to 10–12 M generated by hydrogen peroxide (H2O2) photolysis at concentrations of 2–10 mg L–1)

(Mamane et al., 2007; Roots and Okada, 1975), it is not practicable to use AOPs for the removal

of a large amount of contaminants from water. AOPs are thus usually used to treat trace

amounts of contaminants (Tchobanoglous et al., 2003). One important advantage of AOPs over

other oxidants is that AOPs are environmental-friendly without transfer of pollutants from one

phase to another or production of a large amount of sludge (Gültekin and Ince, 2007;

Tchobanoglous et al., 2003).

6

Table 2.1 Application of AOPs in different areas and for different contaminants

(modified from Parsons, 2004)

Different areas for AOPs

Groundwater Surface water

Industrial wastewater Municipal wastewater

Odour and volatile organic compounds (VOCs) Industrial sludge

Municipal sludge Water recycling

Leachates Swimming pool

Disinfection Ultra purification

Different contaminants treated by AOPs

Amino acids Methyl tertiary-butyl ether (MTBE)

Antibiotics Arsenic

Chromium Pesticides

Coliforms Escherichia coli

Cryptosporidium Disinfection by-products (DBPs)

Paper mill effluent Distillery wastewater

Landfill leachate Drug residues

Glass fibre wastewater Taste and odour causing compounds

Hospital wastewater Grey water

Insecticides Rubber process wastewater

Kraft leaching wastewater Chemical specialities wastewater

Natural organic matter (NOM) Humic materials

Nickel plating wastewater Oilfield wastewater

Cyanide Olive mill wastewater

Parasites Municipal wastewater treatment plant effluent

Phenolic wastewater Urine

Printing wastewater Seed corn wastes

Endocrine disruptors Spent caustic

Pharmaceutical compounds Cyanotoxins

N-nitrosodimethylamine (NDMA) Textile industry wastewater

Chlorophenols Nitrophenols

7

2.1.2 Miscellaneous Methods for AOPs

AOPs can be realized through many methods, some of which are exemplified as shown

below (Gültekin and Ince, 2007; Parsons, 2004; Pera-Titus et al., 2004; Tchobanoglous et al.,

2003). The most common AOPs are the combinations of hydrogen peroxide, ozone, and/or UV

light (Tezcanli-Güyer and Ince, 2004).

1. Photochemical processes

• Vacuum-UV (VUV) photolysis

• UV + H2O2

• UV + O3

• UV + H2O2 + O3

• UV + ultrasound

• Photo-Fenton (H2O2 + Fe2+/ Fe3+ + UV/visible light)

• Photocatalysis (UV + TiO2)

2. Non-photochemical processes

• Ozonation (O3)

• O3 + H2O2

• Fenton (H2O2 + Fe2+)

• O3 + Fe2+/ Fe3+

• Ultrasound

• O3 + ultrasound

2.2 UV Combined with Chlorine as an AOP

2.2.1 Fundamental Chemistry of Aqueous Free Chlorine

Chlorine is a traditional and widely used oxidant and disinfectant in water and wastewater

treatment. It is available in gaseous form (Cl2), concentrated aqueous solution (sodium

hypochlorite, NaOCl) and solid form [calcium hypochlorite, Ca(OCl)2] (AWWA, 1999;

Koivunen and Heinonen-Tanski, 2005; Kuo and Smith, 1996; Sadiq and Rodriguez, 2004). The

use of chlorine is relatively economical compared to other methods (Sadiq and Rodriguez,

2004). However, since chlorination can lead to the formation of regulated and/or harmful

8

disinfection by-products (DBPs), such as trihalomethanes (THMs) and haloacetic acids (HAAs),

and chlorine has also been proven to be challenged by some resistant pathogens (e.g. Giardia

lamblia cysts and Cryptosporidium parvum oocysts), alternative oxidation and disinfection

methods (e.g. chlorine dioxide, ozone, UV, and their combinations) have been investigated and

applied (Benabbou et al., 2007; Hijnen et al., 2006; Jung et al., 2008; Koivunen and Heinonen-

Tanski, 2005).

When Cl2 gas is added to water, a hydrolysis reaction takes place to form hypochlorous

acid (HOCl), which may further dissociate to hypochlorite ion (OCl–), shown in Equations [2.1]

and [2.2] (AWWA, 1999; Tchobanoglous et al., 2003).

Cl2 (aq) + H2O ↔ HOCl + H+ + Cl– [2.1]

Equilibrium constant at 25°C: K = 2

[HOCl][H ][Cl ][Cl (aq)]

+ −

= 4.5 × 10–4 M2

Forward reaction rate constant at 20°C: kforward = 11.0 s–1 (Eigen and Kustin, 1962)

Reverse reaction rate constant at 20°C: kreverse = 1.80 × 104 M–1 s–1 (Eigen and Kustin, 1962)

HOCl ↔ H+ + OCl– [2.2]

Equilibrium constant at 25°C: K = Ka = [H ][OCl ][HOCl]

+ −

= 2.9 × 10–8 M, thus pKa = 7.54 (Deborde

and von Gunten, 2008; Morris, 1966).

According to K in Equation [2.1], a large fraction of molecular chlorine can dissolve into

water to produce HOCl/OCl– except at a very low pH and/or a high chloride concentration.

When NaOCl and Ca(OCl)2 are added to water, HOCl/OCl– may be produced following the

reactions [2.3], [2.4], and [2.2].

NaOCl → Na+ + OCl– [2.3]

Ca(OCl)2 → Ca2+ + 2 OCl– [2.4]

9

The total quantity of HOCl and OCl– is referred to as free available chlorine. The relative

proportions of these chlorine species are dependent on the solution temperature and pH. Figure

2.1 illustrates the pH effect on the percentage distribution of HOCl concentration at 25°C in a

closed aqueous system containing 0.05 M free chlorine, determined using Equations [2.1] and

[2.2]. Since the acceptable range of pH in drinking water is approximately 6.5 to 8.5, both HOCl

and OCl– may be predominant (USEPA, 2009). However, the oxidation and disinfection

efficiencies of HOCl can be about 100 times higher than OCl– (AWWA, 1999; Edstrom

Industries, 2003). As a result, low pH is preferred for chlorine disinfection (USEPA, 1999).

However, pH values should not be kept very low, since at pH lower than 4, HOCl can convert to

dissolved Cl2 gas, which is harmful to structures (Edstrom Industries, 2003). In addition, the

auto-decomposition of free chlorine, with a maximum decomposition rate at pH of

approximately 6.7, results in the formation of chlorite (ClO2–) and chlorate (ClO3

–) (Adam et al.,

1992; Chapin, 1934). These undesirable species, usually as the main disinfection DBPs of

chlorine dioxide, induce oxidative damage in human red blood cells. The World Health

Organization (WHO) provisional guideline values for both are 0.7 mg L–1 (WHO, 2005),

compared to the maximum acceptable concentrations (MACs) of 1 mg L–1 for both established

by Health Canada (2012). In the Stage 1 Disinfectants/Disinfection By-products (D/DBP) Rule

issued by USEPA, the maximum contaminant level (MCL) of chlorite is 1.0 mg L–1, with no

information available for chlorate (USEPA, 1999 and 2009).

Figure 2.1 Percentage distribution of HOCl as a function of pH

10

2.2.2 Fundamental Chemistry of the UV/Chlorine AOP

Since conventional chlorine disinfection has been found to be challenged by resistant

protozoa, especially Cryptosporidium parvum, alternative disinfectants, such as ozone, chlorine

dioxide, and UV, and different disinfection strategies, such as sequential treatment with two or

more disinfectants, have been applied (Betancourt and Rose, 2004; Venczel et al., 1997). The

incorporation of UV and chlorine is a multiple-barrier disinfection scenario where chlorination

can be applied before, during, and/or after UV disinfection (Watts and Linden, 2007). This

process enhances the inactivation of Cryptosporidium due to the efficacy of UV and also

maintains residual disinfection effect in distribution systems (Betancourt and Rose, 2004; Feng

et al., 2007). However, free and combined chlorine were observed to degrade when exposed to

UV light (Feng et al., 2007; Watts and Linden, 2007). For example, free chlorine is not stable in

the presence of sunlight in treated water and swimming pool. Nowell and Hoigné (1992a) found

that HOCl and OCl– exposed to sunlight followed the first-order degradation kinetics with the

decay rate constants of 2 × 10–4 s–1 and 1.2 × 10–3 s –1, respectively. A higher chlorine dose

should be added to swimming pool to maintain an appropriate chlorine residual during sunny

days (Nowell and Hoigné, 1992a).

Many studies focused on the photolysis of gaseous HOCl in the stratosphere, since the

photochemical products of HOCl in the atmosphere may play a role in catalyzed destruction of

stratospheric ozone layer (Guo, 1993; Molina and Molina, 1978; Molina et al., 1980; Tanaka et

al., 1998; Vogt and Schindler, 1992). Photolysis of aqueous chlorine may be very different from

that in the atmosphere, but has been explored in very few studies. Chlorine photolysis by UV

light is revealed to be complex, which involves a series of chain reactions with the formation of

many intermediates and products, including •OH and chlorine radicals (•Cl) (Nowell and

Hoigné, 1992a; Watts and Linden, 2007). The molar absorption coefficients of HOCl, OCl–, and

NH2Cl (Figure 2.2) are wavelength dependent, according to the results of Watts and Linden

(2007). This is consistent with Morris (1966) and Feng et al. (2007), who also indicated that the

absorption spectrum peaked at 236 nm for HOCl and 292 nm for OCl–.

Buxton and Subhani (1972a, 1972b and 1972c) explored the photochemical chain reactions

of aqueous hypochlorite and chlorite ions. In alkaline solution, in which OCl– is predominant, a

series of OCl– photodecomposition reactions at 254, 313, and 365 nm were proposed by Buxton

and Subhani (1972b), shown in Equations [2.10] to [2.33]. The final products at 365 nm are O2,

11

Cl–, ClO2–, and ClO3

–, while the yield of ClO2– is not observed at the other two lower

wavelengths, since ClO2– is completely depleted at wavelengths lower than 320 nm shown in

Equations [2.31] and [2.33] (Buxton and Subhani, 1972b). The production of ClO2– and ClO3

–

may need to be considered due to their adverse effect on human health.

Figure 2.2 Molar absorption coefficients of HOCl, OCl–, and NH2Cl

OCl– + hν → Cl– + O(3P) (ground state triplet oxygen atom) [2.10]

OCl– + hν → •Cl + O– [2.11]

OCl– + hν → Cl– + O(1D) (excited singlet oxygen atom) at λ < 320 nm [2.12]

O(3P) + OCl– → ClO2– [2.13]

O(3P) + OCl– → Cl– + O2 [2.14]

O(3P) + ClO2– → ClO3

– [2.15]

O(3P) + ClO2– → Cl– + O2 + O(3P) [2.16]

O– + H2O ↔ •OH + OH– [2.17]

Forward reaction rate constant at 20–25°C for [2.17]: kforward ≈ 108 s–1 (Buxton, 1969)

Reverse reaction rate constant at 20–25°C for [2.17]: kreverse ≈ 1010 M–1 s–1 (Buxton, 1969)

•OH + OCl– → ClO• + OH– [2.18]

0

50

100

150

200

250

300

350

400

450

200 225 250 275 300 325 350 375 400

Mol

ar a

bsor

ptio

n co

effic

ient

(M

–1cm

–1)

Wavelength (nm)

HOCl

OCl-

NH2ClOCl–

NH2Cl

12

Reaction rate constant: k = (9.0 ± 0.5) × 109 M–1 s–1 (Buxton and Subhani, 1972a) or

8 × 109 M–1 s–1 (Nowell and Hoigné, 1992a; Watts and Linden, 2007)

•OH + ClO2– → ClO2 + OH– [2.19]

Reaction rate constant: k = (6.3 ± 0.5) × 109 M–1 s–1 (Buxton and Subhani, 1972a)

O– + OCl– → ClO• + O2– [2.20]

O– + ClO2– → ClO2 + O2– [2.21]

•Cl + OCl– → ClO• + Cl– [2.22]

Reaction rate constant for [2.22]: k = 8.2 × 109 M–1 s–1 (NIST Database, 2002)

2 ClO• ↔ Cl2O2 [2.23]

Cl2O2 + H2O → ClO2– + OCl– + 2 H+ [2.24]

Cl2O2 + H2O → O2 + Cl– + OCl– + 2 H+ [2.25]

Cl2O2 + ClO2– + H2O → ClO3

– + 2 OCl– + 2 H+ [2.26]

O(1D) + H2O → H2O2 [2.27]

H2O2 + OCl– → O2 + Cl– + H2O [2.28]

If the ClO2– concentration is high, such as in a ClO2

– solution, this species can also absorb

photons and produce photochemical products (Buxton and Subhani, 1972c).

ClO2– + hν → OCl– + O(3P) at λ > 320 nm [2.29]

ClO2– + hν → ClO• + O– [2.30]

ClO2– + hν → OCl– + O(1D) at λ < 320 nm [2.31]

ClO2– + hν → (ClO2

–)* [2.32]

(ClO2–)* + ClO2

– → ClO2 + OCl– + O– [2.33]

13

Oppenländer (2003) indicated that Cl– reacts with hydroxyl radicals described in the following

equations:

•OH + Cl– → •ClOH– [2.34]

Reaction rate constant for [2.34]: k = 4.3 × 109 M–1 s–1 (Oppenländer, 2003)

•ClOH– + H+ → •Cl + H2O [2.35]

Meanwhile, Cl– also reacts with the generated •Cl very rapidly, shown in Equation [2.36], to

produce •Cl2–, which is much less reactive than •Cl and •OH (Buxton et al., 1998). The further

reactions of •Cl2– with target contaminants are negligible (Buxton et al., 1998).

•Cl + Cl– ↔ •Cl2– [2.36]

Forward reaction rate constant: kforward = 8.5 × 109 M–1 s–1 (Buxton et al., 1998)

Reverse reaction rate constant: kreverse = 6.0 × 104 M–1 s–1 (Buxton et al., 1998)

No past study systematically elaborated the photodecomposition process of HOCl in acidic

chlorine solutions. The primary photodecomposition of HOCl exposed to UV light is shown in

Equation [2.37] (Feng et al., 2007; Oliver and Carey, 1977; Thomsen et al., 2001; Watts and

Linden, 2007). Thomsen et al. (2001) indicated that the dissociation of HOCl molecule is very

fast, less than 1 picosecond, while the recombination is slower, taking approximately 50

picoseconds.

HOCl + hv → •OH + •Cl [2.37]

The proof of this reaction is that some organic compounds that do not react with HOCl

efficiently in dark solution, such as ethanol, n-butanol, and benzoic acid, can degrade largely in

the presence of sunlight probably via two chain processes shown in Equations [2.38] to [2.41]

(Feng et al., 2007; Kobayashi and Okuda, 1972; Nowell and Hoigné, 1992b; Oliver and Carey,

14

1977). However, it should be noted that chlorinated products may be produced, which may

generate undesirable chlorinated DBPs. For example, based on the results of Oliver and Carey

(1977), organic chlorine accounts for 16.5% by weight of total photochemical products of an n-

butanol solution containing 0.01 M HOCl at pH 4 after 30 min exposure of 350 nm UV light.

(1) •OH chain reactions:

•OH + RH → R• + H2O [2.38]

R• + HOCl → RCl + •OH [2.39]

(2) •Cl chain reactions:

•Cl + RH → R• + HCl [2.40]

R• + HOCl → ROH + •Cl [2.41]

Since HOCl is the conjugate acid of OCl–, photochemical chain reactions of HOCl are

probably analogous to those of OCl–, which have been shown above. Feng et al. (2007) have

shown some possible chain reactions related to HOCl in Equations [2.42] and [2.43]. However,

based on their results, Feng et al. (2007) also proposed that the chain reactions of OCl–

photolysis may not be as important as those for HOCl.

•OH + HOCl → H2O + ClO• [2.42]

Reaction rate constant: k = 8.46 × 104 M–1 s–1 (Watts and Linden, 2007)

•Cl + HOCl → HCl + ClO• [2.43]

Reaction rate constant: k = 3 × 109 M–1 s–1 (NIST Database, 2002)

Feng et al. (2007) also indicated that the termination of HOCl chain reactions may involve:

15

•Cl + ClO• + H2O → 2 HOCl [2.44]

2 ClO• + H2O → HCl + HClO3 [2.45]

The quantum yields of HOCl/OCl– decay and photochemical product formation have been

evaluated. The quantum yield (Φ) is used to measure the photon efficiency in a photochemical

reaction. It is defined as the number of moles of reactant consumed or product produced per

einstein (1 einstein = one mole of photons = 6.022 x 1023 photons) of photons absorbed (Bolton,

2010). Buxton and Subhani (1972b) determined the quantum yields of OCl– decay and its

photochemical product formation at 365, 313, and 254 nm in OCl– solutions at concentrations of

10–4 to 10–3 M, which are shown in Table 2.2. Feng et al. (2007) reported the quantum yields of

1.0 ± 0.1 and 0.9 ± 0.1 for HOCl and OCl– decay, respectively, at 254 nm at initial

concentrations of less than 70 mg L–1 as free chlorine, while these values as determined by

Watts and Linden (2007) were 1.5 and 1.3 for HOCl and OCl–, respectively, at 254 nm at initial

concentrations of approximately 1–4 mg L–1 as free chlorine. The quantum yields higher than 1

are probably because the chain reactions that are initiated by the generated •OH and •Cl lead to

the further destruction of HOCl and OCl– (Feng et al., 2007; Watts and Linden, 2007). For

example, Feng et al. (2007) proposed that the linear increase of quantum yield of HOCl

photolysis with the increase of HOCl concentration from 71 to 1,350 mg L–1 as free chlorine

was caused by the promotion of the HOCl chain reaction at higher concentrations. However, the

detailed process of HOCl chain reactions has not yet been identified. In addition, Feng et al.

(2007) also found that the OCl– quantum yield was independent of its concentration ranging

from 3.5 to 640 mg L–1 as free chlorine. They therefore postulated no chain reaction initiated by

OCl– photoproducts, which conflicts with Buxton and Subhani (1972a, 1972b and 1972c).

The photolysis of free chlorine solutions is influenced by multiple factors. Based on Table

2.2, the OCl– quantum yield increases with the reduction of the wavelength. A higher production

ratio of •OH generated by HOCl at 254 nm was observed when compared to sunlight irradiation

at higher wavelengths (Nowell and Hoigné, 1992b). Watts and Linden (2007) found a much

higher quantum yield for HOCl decay (3.7) when exposed to a medium pressure (MP) mercury

UV lamp (a polychromatic UV source emitting 200 to 400 nm UV light), than the value of 1.5

when exposed to a low pressure (LP) mercury UV lamp (a monochromatic UV source only

emitting 254 nm UV light in the range of 200 to 400 nm). In addition, free chlorine photolysis is

16

also impacted by pH and different water matrices. Nowell and Hoigné (1992a) found that the

half-life of free chlorine species was longer at lower pH. Both Kobayashi and Okuda (1972) and

Watts and Linden (2007) reported that the water matrix was an impact factor for free chlorine

photolysis. This may be because the different compounds in the water play various roles in the

promotion or inhibition of chlorine photolysis chain reactions, and/or the different photon

absorption efficiencies of these compounds results in attenuation of photon irradiation on free

chlorine molecules to different extents (Kobayashi and Okuda, 1972; Watts and Linden, 2007).

Table 2.2 Quantum yields of OCl– decay and photochemical product formation

365 nm 313 nm 254 nm

OCl– 0.60 ± 0.02 0.39 ± 0.01 0.85 ± 0.02 ClO2

– 0.160 ± 0.005 0 0

O2 0.04 ± 0.02 0.069 ± 0.005 0.200 ± 0.005 ClO3

– 0.08 ± 0.02 0.08 ± 0.02 0.15 ± 0.02

Cl– 0.36 ± 0.03 0.27 ± 0.02 0.70 ± 0.03

The photolysis of OCl– was observed to be less efficient in the destruction of nitrobenzene

(NB) than HOCl photolysis under parallel conditions (Watts et al., 2007), which implies that the

concentration of •OH produced by UV/OCl– is lower than UV/HOCl. Nowell and Hoigné

(1992b) found that the •OH yield factor for OCl– photolysis was only 0.1 (i.e., 1 mole of OCl–

consumed can produce 0.1 mole of •OH), while this value for HOCl was 0.7 in sunlight and 0.9

in 254 nm UV light. However, less difference in the quantum yields of •OH formation by HOCl

and OCl– photolysis was reported by other studies, such as 1.40 and 0.28, respectively, reported

by Watts and Linden (2007) and Watts et al. (2007), or 0.46 and 0.61, respectively, reported by

Jin et al. (2011) and Chan et al.(2012). As a result, the lower efficiency of UV/OCl– relative to

UV/HOCl may be more attributed to the much stronger •OH scavenging efficiency of OCl– than

that of HOCl (shown in Equations [2.18] and [2.42]). For example, Watts et al. (2007) found

that almost 100% of the available •OH was consumed by NB at pH 5, while this value was only

less than 10% at pH 7. The other fraction of available •OH was consumed by OCl–. This,

therefore, implies that UV/HOCl is probably a more promising AOP than UV/OCl– (Watts and

Linden, 2007). Unfortunately, few studies have focused on this topic. Using •OH probe

17

compounds, i.e., para-chlorobenzoic acid (pCBA) and NB, Watts and Linden (2007) found that

the steady-state •OH concentrations were 5.7 × 10–13 M and 1.7 × 10–12 M generated by 1 and 3

mg L–1 free chlorine solutions, respectively, under LP UV exposure at pH 4. This •OH

concentration level is comparable to that generated by UV/H2O2 (Mamane et al., 2007). Watts

and Linden (2007) also proposed the potential use of UV/HOCl under mildly acidic conditions

as an alternative to UV/H2O2 AOP, based on their findings that HOCl produces •OH more

efficiently than does H2O2, but the •OH scavenging rate is lower than the latter. Watts et al.

(2007) evaluated the efficiency of UV/HOCl for treatment of NB, and found that the first-order

decay of NB increased following this order of treatment: UV/chlorine (at pH 7) < UV/H2O2 (at

pH 7) < UV/chlorine (at pH 6) < UV/chlorine (at pH 5). Additionally, in terms of the energy

demand, the efficiency of UV/HOCl AOP at pH 5 is also higher than UV/H2O2 and UV/chlorine

at higher pH values (Watts et al., 2007).

The photolysis of monochloramine exposed to LP and MP lamps was also investigated by

Watts and Linden (2007) and Zhang et al. (2015). Although NH2Cl has much higher molar

absorption coefficients at low UV wavelengths than HOCl and OCl– (based on Figure 2.2), the

quantum yields of NH2Cl photolysis was observed to be minimal, implying that it is not a

promising chemical that can be used as an AOP.

2.3 Formation of Disinfection By-Products (DBPs) in UV/Chlorine

2.3.1 Chlorinated DBPs

Free chlorine is the most commonly used primary disinfectant in North America to

inactivate pathogens and limit waterborne diseases, but it also leads to the formation of

chlorination DBPs, which can be regulated and/or may be suspected of being associated with

adverse health effects such as toxicity, carcinogenicity, mutagenicity, and/or genotoxicity

(USEPA, 1999; AWWA, 1999; Nieuwenhuijsen et al., 2000; Matilainen and Sillanpää, 2010).

THMs were the first group of DBPs discovered as a result of drinking water chlorination in

1974 (USEPA, 1999). At present, more than 700 halogenated DBPs have been reported,

whereas in many cases more than 50% of total organic halides (TOX) formed by chlorination is

still unknown (USEPA, 1999; Richardson, 2003; USEPA, 2002; Hansen et al., 2012; Matilainen

and Sillanpää, 2010). Most of the identified DBPs can be produced by the reactions between

18

free chlorine, hypobromous acid (HOBr) (from the oxidation of bromide by HOCl or ozone,

shown in Equations [2.46] and [2.47]) or hypoiodous acid (HOI) (mainly present in

chloramination, because of instability in the presence of free chlorine) and natural organic

matter (NOM), such as humic and fulvic acids (Glezer et al., 1999; Nieuwenhuijsen et al., 2000;

Kristiana et al., 2009; Richardson, 2003; Westerhoff et al., 2004). The reaction types between

halogens and NOM are mainly oxidation (i.e., breakage of carbon-carbon double bonds of NOM

molecules) and substitution (i.e., replacement of functional groups by halogen molecules)

(Westerhoff et al., 2004). Therefore, there are generally two approaches to reduce DBP

formation in water treatment: (1) using a non-chlorine-based disinfectant or minimizing chlorine

doses for primary disinfection, (2) reducing the amount of NOM before chlorination using

chemical and physical processes, such as coagulation, flocculation, and filtration (Chin and

Bérubé, 2005).

Br– + HOCl → HOBr + Cl– [2.46]

Reaction rate constant: k = 2950 M–1 s–1 (Westerhoff et al., 2004)

Br– + O3 → OBr– + O2 [2.47]

Reaction rate constant: k = 160 M–1 s–1 (Westerhoff et al., 2004)

Amongst many halogenated DBPs, THMs and HAAs occur most consistently and are

frequently associated with the highest concentrations during chlorination (AWWA, 1999;

Glezer et al., 1999; Nieuwenhuijsen et al., 2000; Tchobanoglous et al., 2003). THMs include

chloroform, bromodichloromethane, chlorodibromomethane, and bromoform (when not

considering the iodinated species of THMs). Compared with the other components, chloroform

is most prevalent (AWWA, 1999). HAAs, which are the second most common DBPs, consist of

9 chlorinated/brominated compounds: monochloro-, monobromo-, dichloro-, dibromo-,

bromochloro-, trichloro-, tribromo-, bromodichloro-, chlorodibromoacetic acids. Dichloro- and

trichloroacetic acids are the most prevalent HAAs (AWWA, 1999; Boorman et al., 1999;

USEPA, 2002). In addition, haloacetonitriles (HANs), haloketones (HKs), and chloropicrin (CP)

19

are also produced during chlorination with the amounts typically less than THMs and HAAs

(Glezer et al., 1999; Boorman et al., 1999; USEPA, 1999; Yang et al., 2007). HANs, HKs, and

CP have not been regulated yet; however, they have been included in the USEPA Information

Collection Rule for periodically monitoring requirements because of their potentially

carcinogenic and/or mutagenic effects on human health (Yang et al., 2007; USEPA, 1996;

Plewa et al., 2008; Muellner et al., 2007). Futhermore, inorganic DBPs, including perchlorate

(ClO4–), chlorate (ClO3

–), chlorite (ClO2–), and bromate (BrO3

–), may be also produced during

free chlorine photolysis (Buxton and Subhani, 1972a, 1972b, and 1972c; von Gunten and

Hoigné, 1994; Kang et al., 2006). These compounds have been considered for their health

concerns (Richardson et al., 2007; Korn et al., 2002; York et al., 2001). Table 2.3 enumerates

some typical halogenated by-produces, their median concentrations in treated drinking water in

the United States, and the USEPA’s MCLs, WHO’s guideline values, and/or Health Canada’s

MACs (Boorman et al., 1999; USEPA, 2009; WHO, 2011; Health Canada, 2012).

2.3.2 DBP Formation Kinetics

DBP formation by chlorination depends on a number of factors, including the type and

concentration of organic precursors, free chlorine concentration, bromide concentration, pH,

temperature, etc. (Tchobanoglous et al., 2003; USEPA, 1999; Yang et al., 2007). NOM is the

primary precursor of organic DBPs, which principally consists of humic material (hydrophobic)

and fulvic material (hydrophilic), including proteins, lipids, carbohydrates, carboxylic acids,

amino acids, and hydrocarbons (USEPA, 1999; Westerhoff et al., 2004). However, accurate

determination of NOM molecular propensities is very difficult because NOM is a heterogeneous

mixture of compounds. For example, the molecular weights of humic acid can vary from one

thousand to several hundred thousand grams per mole (Cameron et al., 1972).

In general, increasing pH promotes the formation of THMs but reduces the formation of

HAAs, HANs and HKs, while pH effects on different components might be different

(Pourmoghaddas and Stevens, 1995; Yang et al., 2007). For example, the formation of the four

components of THMs is enhanced at higher pH but to different extents (Peters et al., 1980;

Hansen et al., 2012; Liang and Singer, 2003); increasing pH has a minimal effect on the

formation of monohaloacetic acid and dihaloacetic acid, but may significant decrease

trihaloacetic acid formation (Liang and Singer, 2003) or may not (Hansen et al., 2012); and

20

dichloroacetonitrile, trichloroacetonitrile, and 1,1,1-trichloro-2-propanone formation decreases

with the increase of pH (Yang et al., 2007; Glezer et al., 1999; Hansen et al., 2012).

Table 2.3 Halogenated by-products in drinking water treatment

By-products U.S. Median

concentrations (μg L–1)

USEPA MCLs (mg L–1)

WHO (Provisional)

Guideline values (mg L–1)

Health Canada MACs (mg L–1)

THMs 0.08 (total THMs) 0.1 (total THMs) Chloroform 25 0.3 Bromodichloromethane 9.5 0.06 Chlorodibromomethane 1.6 0.1 Bromoform < 0.2 0.1 HAAs 0.06 (HAA5) 0.08 (total HAAs) Dichloroacetic acid 15 0.05 Trichloroacetic acid 11 0.2 Bromochloroacetic acid 2.69 Monochloroacetic acid 1.3 0.02 Dibromoaceic acid < 0.5 Monobromoacetic acid < 0.5 0.02 Tribromoacetic acid Bromodichloroacetic acid Chlorodibromoacetic acid Haloacetonitriles (HANs) Dichloroacetonitrile 2.1 0.02 Bromoacetonitrile 0.7 Bromochloroacetonitrile 0.6 Dibromoacetonitrile < 0.5 0.07 Trichloroacetonitrile < 0.02 Tribromoacetonitrile Haloketones (HKs) 1,1,1-Tichloropropanone 1.0 1,1-Dichloropropanone 0.4 1,3-Dichloropropanone Others Chlorate 161 0.7 1 Chlorite 1.0 0.7 1 Bromate 0.01 0.01 0.01 Perchlorate Chloral hydrate 2.1 Chloropicrin 0.4 MX 0.005 Cyanogen chloride 0.62 Cyanogen bromide Halonitriles 0.4 N-nitrosodimethylamine (NDMA) 0.0001 0.00004 2,4,6-Trichlorophenol 0.2 0.005 Formaldehyde 2-Chlorophenol

21

The mechanisms of DBP formation from NOM are still uncertain because of the high

heterogeneity and complexity of NOM components (Iriarte-Velasco et al., 2006). Among

various types of DBPs, THMs have been investigated most thoroughly (Gallard and von

Gunten, 2002). THM formation can generally be divided into two stages: a rapid initial reaction

stage followed by a slow reaction stage (Boccelli et al., 2003; Gallard and von Gunten, 2002;

Westerhoff et al., 2004). The THMs formed in the first stage account for approximately 15–30%

of the total THMs (Gallard and von Gunten, 2002). The two reaction stages probably result from

different reaction rates between chlorine and THM precursors, or different sites in the NOM

structure with different reactivities (Gallard and von Gunten, 2002; Greca and Fabbricino,

2008). For example, Gallard and von Gunten (2002) indicated that resorcinol-type structures,

which present 15–30% of THM precursors, and readily enolizable compounds, such as β-

diketones and β-ketoacids, might play a role in the fast reacting THM precursors, while other

phenolic compounds might be the slowly reacting THM precursors. Adin et al. (1991) proposed

that the two stage THM formation arose from a multi-step process, consisting of rapid reactions

between chlorine and precursors to produce chlorinated intermediates and the subsequently

slower reactions to produce THMs and other products.

The intersection point between the rapid and slow reaction stages is not well-defined. Sohn

et al. (2004) indicated that this point occurred at about 5 h of reaction time, which is generally

consistent with Iriarte-Velasco et al. (2006) and Gallard and von Gunten (2002), who both found

that the fast formation of THM from chlorination occurred in the first 3 h of reaction. In

contrast, Westerhoff et al. (2004) reported that the rapid initial reaction only lasted 1–5 min. The

reason for the discrepancy is unknown.

Gallard and von Gunten (2002) indicated that THM formation at the slow stage exhibited

second-order reaction kinetics with rate constants between 0.01 and 0.03 M–1 s–1, which was

controlled by the reactions between free chlorine and NOM. Iriarte-Velasco et al. (2006)

reported that the second-order rate constants of chloroform formation at the fast stage from the

chlorination of humic and fulvic acid were 1.47 and 1.43 M–1 s–1, respectively, at 20°C and pH

7.0, while the corresponding values for slow reaction were 0.24 and 0.129 M–1 s–1. Westerhoff et

al. (2004) investigated free chlorine consumption by NOM, and reported that the rate constant of

the initial rapid chlorine consumption was 50–500 M–1 s–1 within the first 1–5 min followed by

0.7–5 M–1 s–1 in the subsequent slower chlorine consumption stage. However, no past study

22

focused on DBP formation kinetics within the first one minute after addition of free chlorine,

which may be the scenario when the UV/chlorine AOP is applied.

Clark (1998) proposed that the formation rate of THMs was proportional to the

consumption rate of chlorine during the slow reaction stage. This is in agreement with Gang et

al. (2003), who pointed out that 31 to 42 μg THM was formed per mg Cl2 consumed. Gallard

and von Gunten (2002) also confirmed this linear relationship, and found 0.029 mol CHCl3

formed per mol Cl2 consumed (equivalent to 48.8 μg CHCl3 formed per mg Cl2 consumed). This

implies that during the slow reaction stage only a small fraction (less than 10%) of the

consumed chlorine is involved in the final formation of THMs.

2.3.3 DBP Formation by UV and AOPs

UV exposure alone cannot produce halogenated DBPs in the water without halogenating

agents such as chlorine and chloramine, while disinfection doses of UV (e.g., ≤ 200 mJ cm–2 for

4-log inactivation of viruses, based on USEPA Long Term 2 Enhanced Surface Water

Treatment Rule, LT2ESWTR) can even slightly decrease some DBPs, such as CDBM, TBM,

DBAA, DBAN (Lyon et al., 2012; Malley Jr. et al., 1995; USEPA, 2006). When extended UV

exposure (such as hours) with high UV doses (e.g., thousands of mJ cm–2) is applied, most of

common DBPs, such as THMs, HAAs, HANs, HKs, CP, as well as adsorbable organic halides

(AOX) and NOM, undergo slow photodecomposition processes (Hansen et al., 2013; Li and

Blatchley III, 2007; Weng et al., 2012; Kulovaara et al., 1996; Corin et al., 1996; Deng et al.,

2014; Höfl et al., 1997). Generally, bromated DBPs are more susceptible to UV exposure than

their chlorinated counterparts (Jo, 2008; Hansen et al., 2013).

Variable effects of UV treatment on DBP formation during secondary chlorination have

been reported in the literature. Certain studies have shown that THM, HAA, and AOX

formation during subsequent chlorination was not significantly affected by UV exposure

(Malley Jr. et al., 1995; Lyon et al., 2010; Lyon et al., 2012; Kashinkunti et al., 2004; Reckhow

et al., 2010; Liu et al., 2002a), while some other studies have shown contradictions or different

UV impacts on other DBPs. For example, Liu et al. (2006) and Liu et al. (2012) observed

significant increases of THM, HAA, and AOX formation during chlorination for 1 h, and 1, 3,

and 7 days, because of LP or MP UV pretreatment at a dose of 60 mJ cm–2. Lyon et al. (2012)

found that UV irradiation at disinfection doses did not significantly affect THM and HAA

23

formation during subsequent chlorination, but did increase CP when using MP UV rather than

LP UV. It was explained that MP UV, not LP UV, promotes nitrate photolysis to generate

reactive nitrogen species, which can react with NOM to produce CP, as also reported by

Reckhow et al. (2010) and Shah et al. (2011).

The enhancement effect of UV at an AOP dose on DBP formation during subsequent

chlorination is likely to be greater than when UV is applied at a disinfection dose. Lyon et al.

(2012) observed that MP UV pretreatment at a dose of 1,000 mJ cm–2 could increase TCM

formation by 30–40% during subsequent 24 h chlorination, while the impact of LP UV at such

doses on THM and HAA formation was low. MP UV did not increase HAA formation. This is

partially consistent with research carried out by Dotson et al. (2010), who found that MP UV

pretreatment increased THM formation to an extent much higher than that for LP UV, while

HAA formation was decreased by LP UV, but increased by MP UV. This UV effect was not

able to be reproduced by the similar work performed by Liu et al. (2002a), who found both

THM and HAA formation during secondary chlorination was not significantly increased by UV

pretreatment at doses of up to 5,000 mJ cm–2, while THM formation was slightly decreased.

Toor and Mohseni (2007) did not detect any impact of UV at a UV does of 2,500 mJ cm–2 on

THM formation during subsequent chlorination. The reason for this discrepancy is uncertain,

but reflects the complicated mechanisms of DBP formation by chlorination in water pretreated

by UV exposure. UV exposure is proposed to affect NOM structures and properties, and

subsequently to change its reactivity toward chlorine (Lyon et al., 2012; Liu et al., 2012).

During UV exposure, high-molecular-weight NOM components probably follow complicated

photochemical reactions to decompose to low-molecular-weight compounds with more carboxyl

and carbonyl carbon atoms (Kulovaara et al., 1996; Corin et al., 1996).

When UV-based AOPs are applied, •OH can eliminate undesired organic pollutants, as

well as various DBPs (Glauner et al., 2005; Höfl et al., 1997; Tang and Tassos, 1997; NIST

Database, 2002; Lifongo et al., 2004; Jo, 2008). •OH can also simultaneously react with NOM

resulting in the mineralization of NOM under extreme conditions (e.g., a UV/H2O2 AOP at a

UV dose of 3,500–5,000 mJ cm–2 with a H2O2 dose of 20–100 mg L–1), or changes in NOM

structure and reactivity in mild AOP treatment (e.g., a UV/H2O2 AOP at a UV dose in the order

of 1,000 mJ cm–2). Changes to NOM can lead to DBP reduction or enhancement depending on

the specific characteristics of the changes during subsequent chlorination (Dotson et al., 2010;

24

Toor and Mohseni, 2007; Pisarenko et al., 2013; Glauner et al., 2005; Goslan et al., 2006;

Sarathy et al., 2011; Rezaee et al., 2014; Matilainen and Sillanpää, 2010). For example, Toor

and Mohseni (2007) found that after the treatment of UV/H2O2 AOP at an H2O2 concentration

of 23 mg L–1 and a UV dose of ~3,500 mJ cm–2, THM and HAA formation was decreased by

~70% and ~40%, respectively, during secondary chlorination for 3 days, which is well

consistent with the work carried out by Lamsal et al. (2011), who showed a similar reduction in

THM and HAA formation during 24 h chlorination preceded by a UV/H2O2 AOP at a UV dose

of ~1,100 mJ cm–2 and 23 mg L–1 H2O2 concentration. Liu et al. (2002b) also reported reduction

in THM and HAA formation by 20–50% and 40–60%, respectively, during secondary

chlorination for 24 h after UV/H2O2 pretreatment at 100 mg L–1 initial H2O2 dose and 5,000 mJ

cm–2 UV dose. In contrast, Dotson et al. (2010) observed an approximately 10–20% increase of

THM and HAA formation during 24 h chlorination attributed to the UV/H2O2 pretreatment at a

LP or MP UV dose of 1,000 mJ cm–2 and a H2O2 concentration of 10 mg L–1, whereas AOX

formation either increased or decreased under different conditions. Likewise, Kleiser and

Frimmel (2000) reported that 50 min UV/H2O2 pretreatment at an initial H2O2 dose of 8 mg L–1

(UV dose unavailable) led to increases in THM and AOX formation by 20% and 5%,

respectively, during subsequent 48 h chlorination, while extended treatment of UV/H2O2 for up

to 1050 min resulted in the reduction of THM and AOX formation by 75% and 71%,

respectively. Bond et al. (2009) indicated that incomplete mineralization of NOM by the

UV/H2O2 process increased DCAA formation in subsequent chlorination, which was verified by

Toor and Mohseni (2007).

DBP formation in the UV/chlorine AOP is probably more complicated than that in other

AOPs, because the formation of Cl radicals (•Cl) by free chlorine photolysis may react with

NOM to produce DBPs, and inorganic DBPs, such as ClO4–, ClO3


–

(Pisarenko et al., 2013; Deng et al., 2014; Buxton and Subhani, 1972a, 1972b, and 1972c; von

Gunten and Hoigné, 1994; Kang et al., 2006). The role of •Cl in the UV/chlorine AOP is not

well established. It has been considered to be negligible in some of studies (Watts and Linden,

2007; Nowell and Hoigné, 1992b). In contrast, Fang et al. (2014) reported evidence that •Cl

reacts with benzoic acid, suggesting that if NOM were to contain structurally similar

components, chlorinated DBPs could possibly be formed.

25

To date, only a limited number of studies have investigated DBP formation by the

UV/chlorine AOP. Pisarenko et al. (2013) found HAA concentration was increased by ~50% by

UV/chlorine at 10 mg L–1 free chlorine and 3,900 mJ cm–2 LP UV dose (2 h exposure),

compared to a dark control with only chlorination, while THM concentration did not change,

and AOX was either increased or decreased under different conditions. Weng et al. (2012)

observed the promotion of DCAN formation by UV/chlorine at 3 mg L–1 chlorine and 120–360

mJ cm–2 LP UV dose (10–30 min exposure). Deng et al. (2014) reported the significant

promotion of CP formation by UV/chlorine for a 5 min exposure (chlorine doses: 4–10 mg L–1,

LP UV doses: 600–1,600 mJ cm–2). UV/chlorine has also been shown to affect subsequent DBP

formation upon chlorination. Liu et al. (2006) found UV/chlorine at 7 mg L–1 chlorine

concentration and 60 mJ cm–2 LP or MP UV dose led to increases in THMs and HAAs during

the subsequent 3 day chlorination, compared with those without pretreatment. Shah et al. (2011)

also found increased CP and DCAN formation in subsequent chlorination when pretreated with

UV/chlorine at UV doses ranging from 60 to 1,500 mJ cm–2 and a chlorine dose of 7 mg L–1.

2.4 Summary of Literature Review The literature review in this chapter discussed the following key points:

• AOPs are very effective in treating a variety of dissolved organic contaminants at trace levels.

• UV/chlorine is a potential AOP, since the photolysis of free chlorine in the UV range

produces hydroxyl radicals. However, chlorine photolysis may involve a complicated series

of chain reactions, which is not completely understood. Furthermore, UV/chlorine efficiency

is probably pH dependent, since the quantum yields of •OH formation and the •OH

scavenging efficiencies for both HOCl and OCl– were observed to be different.

• Organic and inorganic DBPs may be produced in the UV/chlorine process. Since higher

chlorine doses but much shorter contact times are probably applied in the UV/chlorine AOP

compared to conventional chlorine disinfection, DBP formation by UV/chlorine may be

different from that observed when applying chlorine conventionally, and has not been

explored.

26

References Adam, L.C., Fábián, I., Suzuki, K., Gordon, G., 1992. Hypochlorous acid decomposition in the

pH 5–8 region. Inorganic Chemistry, 31(17), 3534–3541.

Adin, A., Katzhendler, J., Alkaslassy, D., Rav-Acha, C., 1991. Trihalomethane formation in

chlorinated drinking water: a kinetic model. Water Research, 25(7), 797–805.

American Water Works Association (AWWA), 1999. Water Quality and Treatment: A

Handbook of Community Water Supplies, 5th Edition. McGraw-Hill, Inc., New York.

Benabbou, A.K., Derriche, Z., Felix, C., Lejeune, P., Guillard, C., 2007. Photocatalytic

inactivation of Escherischia coli Effect of concentration of TiO2 and microorganism, nature,

and intensity of UV irradiation. Applied Catalysis B: Environmental, 76(3–4), 257–263.

Betancourt, W.Q., Rose, J. B., 2004. Drinking water treatment processes for removal of

Cryptosporidium and Giardia. Veterinary Parasitology, 126(1–2), 219–234.

Boccelli, D.L., Tryby, M.E., Uber, J.G., Summers, R.S., 2003. A reactive species model for

chlorine decay and THM formation under dechlorination conditions. Water Research, 37(11),

2654–2666.

Bolton, J. R., 2010. Ultraviolet Applications Handbook, 3rd edition. Bolton Photosciences Inc.

Published by ICC Lifelong Learn Inc., Edmonton, AB, Canada.

Boorman, G.A., Dellarco, V., Dunnick, J.K., Chapin, R.E., Hunter, S., Hauchman, F., Gardner,

H., Cox, M., Sills, R.C., 1999. Drinking water disinfection byproducts: review and approach

to toxicity evaluation. Environmental Health Perspective, 107(Supplement 1), 207–217.

Bond, T., Goslan, E.H., Jefferson, B., Roddick, F., Fan, L., Parsons, S.A., 2009. Chemical and

biological oxidation of NOM surrogates and effect on HAA formation. Water Research,

43(10), 2615–2622.

Buxton, G.V., 1969. Pulse radiolysis of aqueous solutions. Some rates of reaction of OH and O–

and pH dependence of the yield of O3–. Transactions of the Faraday Society, 65, 2150–2158.

Buxton, G.V., Bydder, M., Salmon G.A., 1998. Reactivity of chlorine atoms in aqueous solution.

Part 1. The equilibrium Cl• + Cl– ↔ Cl2•–. Journal of the Chemical Society, Faraday

Transactions, 94(5), 653–657.





27


947–957.




958–969.




970–977.

Cameron, R.S., Thornton, B.K., Swift, R.S., Posner, A.M., 1972. Molecular weight and shape of

humic acid from sedimentation and diffusion measurements on fractionated extracts. Journal

of Soil Science, 23(4), 394–408.

Chin, A., Bérubé, P.R., 2005. Removal of disinfection by-products precursors with ozone-UV


Chapin, R.M., 1934. The effect of hydrogen-ion concentration on the decomposition of

hypohalites. Journal of the American Chemical Society, 56(11), 2211–2215.

Corin, N., Backlund, P., Kulovaara, M., 1996. Degradation products formed during UV-

irradiation of humic waters. Chemosphere, 33(2), 245–255.



Research, 42(1–2), 13–51.

Deng, L., Huang, C.-H., Wang, Y.-L., 2014. Effects of combined UV and chlorine treatment on

the formation of trichloronitromethane from amine precursors. Environmental Science &

Technology, 48(5), 2697–2705.

Dotson, A.D., Keen, V.S., Metz, D., Linden, K.G., 2010. UV/H2O2 treatment of drinking water

increases post-filtration DBP formation. Water Research, 44(12), 3703–3713.

Edstrom Industries, Inc., 2003. Forms of chlorine in water. Available from

<http://www.edstrom.com/Resources.cfm?doc_id=164> (accessed 07.10.2009).

Eigen, M., Kustin, K., 1962. The kinetics of halogen hydrolysis. Journal of the American

Chemical Society, 84(8), 1355–1361.





28

Fang, J., Fu, Y., Shang, C., 2014. The roles of reactive species in micropollutant degradation in

the UV/free chlorine system. Environmental Science & Technology, 48(3), 1859–1868.



6(3), 277–284.

Gallard, H., von Gunten, U., 2002. Chlorination of natural organic matter: kinetics of

chlorination and of THM formation. Water Research, 36(1), 65–74.

Gang, D.C., Clevenger, T.E., Banerji, S.K., 2003. Relationship of chlorine decay and THMs

formation to NOM size. Journal of Hazardous Materials, 96(1), 1–12.

Glauner, T., Kunz, F., Zwiener, C., Frimmel, F.H., 2005. Elimination of swimming pool water

disinfection by-products with advanced oxidation processes (AOPs). Acta Hydrochimica et

Hydrobiologica, 33(6), 585–594.

Glezer, V., Harris, B., Tal, N., Iosefzon, B., Lev, O., 1999. Hydrolysis of haloacetonitriles:

linear free energy relationship, kinetics and products. Water Research, 33(8), 1938–1948.

Goslan, E.H., Gurses, F., Banks, J., Parsons, S.A., 2006. An investigation into reservoir NOM

reduction by UV photolysis and advanced oxidation processes. Chemosphere, 65(7), 1113–

1119.

Greca, G.D., Fabbricino, M., 2008. DBP formation in drinking water: kinetics and linear

modelling. Water Science & Technology: Water Supply, 8(2), 161–166.

Guo, H., 1993. Time-dependent quantum dynamical study of the photodissociation of HOCl.

The Journal of Physical Chemistry, 97(11), 2602–2608.



816–832.

Hansen, K.M.S., Willach, S., Antoniou, M.G., Mosbæk, H., Albrechtsen, H.-J., Andersen, H.R.,

2012. Effect of pH on the formation of disinfection byproducts in swimming pool water – Is

less THM better? Water Research, 46(19), 6399–6409.

Hansen, K.M.S., Zortea, R., Piketty, A., Vega, S.R., Andersen, H.R., 2013. Photolytic removal

of DBPs by medium pressure UV in swimming pool water. Science of the Total Environment

443, 850–856.

http://www.sciencedirect.com/science/article/pii/S0043135412006513

29




Hijnen, W.A.M., Beerendonk, E.F., Medema, G.J., 2006. Inactivation credit of UV radiation for

viruses, bacteria and protozoan (oo)cysts in water: A review. Water Research, 40(1), 3–22.

Höfl, C., Sigl, G., Specht, O., Wurdack, I., Wabner, D., 1997. Oxidative degradation of AOX

and COD by different advanced oxidation processes: a comparative study with two samples

of a pharmaceutical wastewater. Water Science and Technology, 35(4), 257–264.

Iriarte-Velasco, U., Álvarez-Uriarte, J.I., González-Velasco, J.R., 2006. Kinetics of chloroform

formation from humic and fulvic acid chlorination. Journal of Environmental Science and

Health, Part A: Toxic/Hazardous Substances & Environmental Engineering, 41(8), 1495–

1508.

Jo, C.H., 2008. Oxidation of disinfection byproducts and algae-related odorants by UV/H2O2,

Ph.D. dissertation, Virginia Polytechnic Institute and State University, Blacksburg, Virginia,

USA.

Jung, Y.J., Oh, B.S., Kang, J-W., 2008. Synergistic effect of sequential or combined use of

ozone and UV radiation for the disinfection of Bacillus subtilis spores. Water Research, 42(6–

7), 1613–1621.

Kang, N., Anderson, T.A., Jackson, W.A., 2006. Photochemical formation of perchlorate from

aqueous oxychlorine anions. Analytica Chimica Acta, 567(1), 48–56.

Kashinkunti, R.D., Linden, K.G., Shin, G.-A., Metz, D.H., Sobsey, M.D., Moran, M.C.,

Samuelson, A.M., 2004. Investigating multibarrier inactivation for Cincinnati – UV, by-

products, and biostability. Journal of American Water Works Association, 96(6), 114–127.

Kleiser, G., Frimmel, F.H., 2000. Removal of precursors for disinfection by-products (DBPs) –

differences between ozone- and OH-radical-induced oxidation. The Science of the Total

Environment, 256(1), 1–9.

Kobayashi, T., Okuda, T., 1972. A continuous method of monitoring water quality by chlorine

consumption under ultraviolet radiation. Water Research, 6(2), 197–209.

Koivunen, J., Heinonen-Tanski, H., 2005. Inactivation of enteric microorganisms with chemical

disinfection, UV irradiation and combined chemical/UV treatments. Water Research, 39(8),

1519–1526.

30

Korn, C., Andrews, R.C., Escobar, M.D., 2002. Development of chlorine dioxide-related by-

product models for drinking water treatment. Water Research, 36(1), 330–342.

Kristiana, I., Gallard, H., Joll, C., Croué, J.-P., 2009. The formation of halogen-specific TOX

from chlorination and chloramination of natural organic matter isolates. Water Research,

43(17), 4177–4186.

Kulovaara, M., Corin, N., Backlund, P., Tervo, J., 1996. Impact of UV254-radiation on aquatic

humic substances. Chemosphere, 33(5), 783–790.

Kuo, J.-F., Smith, S.O., 1996. Disinfection. Water Environment Research, 68(4), 503–510.

Lamsal, R., Walsh, M.E., Gagnon, G.A., 2011. Comparison of advanced oxidation processes for

the removal of natural organic matter. Water Research, 45(10), 3263–3269.

Li, J., Blatchley III, E.R., 2007. Combined application of UV radiation and chlorine:

implications with respect to DBP formation and destruction in recreational water applications.

Proceedings of the Water Environment Federation Annual Technical Exhibition and

Conference 2007, Disinfection, San Diego, California, USA, 128–133.

Liang, L., Singer, P.C., 2003. Factors influencing the formation and relative distribution of

haloacetic acids and trihalomethanes in drinking water. Environmental Science & Technology,

37(13), 2920–2928.

Lifongo, L.L., Bowden, D.J., Brimblecombe, P., 2004. Photodegradation of haloacetic acids in

water. Chemosphere, 55(3), 467–476.

Liu, W., Andrews, S.A., Bolton, J.R., Linden, K.G., Sharpless, C., Stefan, M., 2002a.

Comparison of disinfection byproduct (DBP) formation from different UV technologies at

bench scale. Water Science and Technology: Water Supply, 2(5–6), 515–521.

Liu, W., Andrews, S.A., Sharpless, C., Stefan, M., Linden, K.G., Bolton, J.R., 2002b. Bench-

scale investigation into comparative evaluation of DBP formation from different UV/H2O2

technologies. Proceedings of 2002 AWWA Water Quality Technology Conference, Seattle,

WA, USA.

Liu, W., Cheung, L.-M., Yang, X., Shang, C., 2006. THM, HAA and CNCl formation from UV

irradiation and chlor(am)ination of selected organic waters. Water Research, 40(10), 2033–

2043.

31

Liu, W., Zhang, Z., Yang, X., Xu, Y., Liang, Y., 2012. Effects of UV irradiation and

UV/chlorine co-exposure on natural organic matter in water. Science of the Total

Environment, 414, 576–584.

Lyon, B., Weinberg, H., Dotson, A., Linden, K., 2010. Impact of UV-chlorine/chloramine

drinking water treatment on DBP formation in bromide- and nitrate-containing waters.

Proceedings of 2010 Water Quality Technology Conference and Exposition, Savannah,

Georgia, USA.

Lyon, B.A., Dotson, A.D., Linden, K.G., Weinberg, H.S., 2012. The effect of inorganic

precursors on disinfection byproduct formation during UV-chlorine/chloramine drinking

water treatment. Water Research, 46(15), 4653–4664.

Malley Jr., J.P., Shaw, J.P., Ropp, J.R., 1995. Evaluation of by-products produced by treatment

of groundwaters with ultraviolet irradiation. American Water Works Association Research

Foundation, Denver, CO., USA

Mamane, H., Shemer, H., Linden, K.G., 2007. Inactivation of E. coli, B. subtilis spores, and

MS2, T4, and T7 phage using UV/H2O2 advanced oxidation. Journal of Hazardous Materials,

146(3), 479–486.

Matilainen, A., Sillanpää, M., 2010. Removal of natural organic matter from drinking water by

advanced oxidation processes. Chemosphere, 80(4), 351–365.

Molina, M.J., Ishiwata, T., Molina, L.T., 1980. Production of OH from photolysis of HOCl at

307–309 nm. The Journal of Physical Chemistry, 84(8), 821–826.

Molina, L.T., Molina, M.J., 1978. Ultraviolet spectrum of HOCl. The Journal of Physical

Chemistry, 82(22), 2410–2414.

Morris, J.C., 1966. The acid ionization constant of HOCl from 5 to 35°. The Journal of Physical

Chemistry, 70(12), 3798–3805.

Muellner, M.G., Wagner, E.D., McCalla, K., Richardson, S.D., Woo, Y.-T., Plewa, M.J., 2007.

Haloacetonitriles vs. regulated haloacetic acids: are nitrogen-containing DBPs more toxic?

Environmental Science & Technology, 41(2), 645–651.

Nieuwenhuijsen, M.J., Toledano, M.B., Eaton, N.E., Fawell, J., Elliott, P., 2000. Chlorination

disinfection byproducts in water and their association with adverse reproductive outcomes: a

review. Occupational and Environmental Medicine, 57(2), 73–85.

32

NIST Database, 2002. NDRL/NIST chemical kinetics database on the web. Available from

<http://kinetics.nist.gov/solution> (accessed 07.10.2009).





Oliver, B.G., Carey, J.H., 1977. Photochemical production of chlorinated organic in aqueous

solutions containing chlorine. Environmental Science & Technology, 11(9), 893–895.

Oppenländer, T., 2003. Photochemical Purification of Water and Air Advanced Oxidation

Processes (AOPs): Principles, Reaction Mechanisms, Reactor Concepts. Wiley–VCH Verlag

GmbH & Co. KGaA, Weinheim.

Parsons, S., 2004. Advanced Oxidation Processes for Water and Wastewater Treatment. IWA

Publishing, London, UK.

Pera-Titus, M., García-Molina, V., Baños, M.A., Giménez, J., Esplugas, S., 2004. Degradation

of chlorophenols by means of advanced oxidation processes: a general review. Applied

Catalysis B: Environmental, 47(4), 219–256.

Peters, C.J., Young, R.J., Perry, R., 1980. Factors influencing the formation of haloforms in the

chlorination of humic materials. Environmental Science & Technology, 14(11), 1391–1395.

Pisarenko, A.N., Stanford, B.D., Snyder, S.A., Rivera, S.B., Boal, A.K., 2013.Investigation of

the use of chlorine based advanced oxidation in surface water: oxidation of natural organic

matter and formation of disinfection byproducts. Journal of Advanced Oxidation

Technologies, 16(1), 137–150.

Plewa, M.J., Wagner, E.D., Muellner, M.G., Hsu, K.-M., Richardson, S.D., 2008. Comparative

mammalian cell toxicity of N-DBPs and C-DBPs. In: Disinfection By-Products in Drinking

Water: Occurrence, Formation, Health Effects, and Control. American Chemistry Society,

Washington, D.C., USA, 36–50.

Pourmoghaddas, H., Stevens, A.A., 1995. Relationship between trihalomethanes and haloacetic

acids with total organic halogen during chlorination. Water Research, 29(9), 2059–2062.

Reckhow, D.A., Linden, K.G., Kim, J., Shemer, H., Makdissy, G., 2010. Effect of UV treatment

on DBP formation. Journal of American Water Works Association, 102(6), 100–113.

33

Rezaee, R., Maleki, A., Jafari, A., Mazloomi, S., Zandsalimi, Y., Mahvi, A.H., 2014.

Application of response surface methodology for optimization of natural organic matter

degradation by UV/H2O2 advanced oxidation process. Journal of Environmental Health

Science & Engineering, 12(67), 1–8.

Richardson, S.D., 2003. Disinfection by-products and other emerging contaminants in drinking

water. Trends in Analytical Chemistry, 22(10), 666–684.

Roots, R., Okada, M., 1975. Estimation of life times and differences of radicals involved in X-

ray-induced DNA strand breaks or killing of mammalian cells. Radiation Research, 64(2),

306–320.

Rosenfeldt, E.J., Linden, K.G., Canonica, S., von Gunten, U., 2006. Comparison of the

efficiency of •OH radical formation during ozonation and the advanced oxidation processes

O3/H2O2 and UV/H2O2. Water Research, 40(19), 3695–3704.

Sadiq, R., Rodriguez, M.J., 2004. Disinfection by-products (DBPs) in drinking water and

predictive models for their occurrence: a review. Science of the Total Environment, 321(1–3),

21–46.

Sarathy, S.R., Stefan, M.I., Royce, A., Mohseni, M., 2011. Pilot-scale UV/H2O2 advanced

oxidation process for surface water treatment and downstream biological treatment: effects on

natural organic matter characteristics and DBP formation potential. Environmental

Technology, 32(15), 1709–1718.

Shah, A.D., Dotson, A.D., Linden, K.G., Mitch, W.A., 2011. Impact of UV disinfection

combined with chlorination/chloramination on the formation of halonitromethanes and

haloacetonitriles in drinking water. Environmental Science & Technology, 45(8), 3657–3664.

Sohn, J., Amy, G., Cho, J., Lee, Y., Yoon, Y., 2004. Disinfection decay and disinfection by-

products formation model development: chlorination and ozonation by-products. Water

Research, 38(10), 2461–2478.

Svrcek, C., Smith, D.W., 2004. Cyanobacteria toxins and the current state of knowledge on

water treatment options: a review. Journal of Environmental Engineering and Science, 3(3),

155–185.

Tanaka, Y., Kawasaki, M., Matsumi, Y., Fujiwara, H., Ishiwata, T., Rogers, L.J., Dixon, R.N.,

Ashfold, M.N.R., 1998. The ultraviolet photodissociation of Cl2O at 235 nm and of HOCl at

235 and 266 nm. Journal of Chemical Physics, 109(4), 1315–1323.

34

Tang, W.Z., Tassos, S., 1997. Oxidation kinetics and mechanisms of trihalomethanes by

Fenton’s reagent. Water Research, 31(5), 1117–1125.

Tchobanoglous, G., Burton, F.L., Stensel, H.D., 2003. Wastewater Engineering, Treatment and

Reuse, 4th edition. The McGraw-Hill Companies, Inc., New York.

Tezcanli-Güyer, G., Ince, N.H., 2004. Individual and combined effects of ultrasound, ozone and

UV irradiation: a case study with textile dyes. Ultrasonics, 42(1–9), 603–609.

Thomsen, C.L., Madsen, D., Poulsen, J.A., Thøgersen, J., Jensen, S.J.K., Keiding, S.R., 2001.

Femtosecond photolysis of aqueous HOCl. Journal of Chemical Physics, 115(20), 9361–9369.

Toor, R., Mohseni, M., 2007. UV-H2O2 based AOP and its integration with biological activated

carbon treatment for DBP reduction in drinking water. Chemosphere, 66(11), 2087–2095.

United States Environmental Protection Agency (USEPA), 1996. ICR manual for bench- and

pilot-scale treatment studies. Available from

<http://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey= 20001VOG.txt> (accessed 08.20.2014).

United States Environmental Protection Agency (USEPA), 1999. Alternative Disinfectants and

Oxidants Guidance Manual. Available from

<http://www.epa.gov/safewater/mdbp/mdbptg.html#disinfect> (accessed 07.10.2009).

United States Environmental Protection Agency (USEPA), 2002. The occurrence of disinfection

by-products (DBPs) of health concern in drinking water: results of a nationwide DBP

occurrence study. Available from

<http://epa.gov/athens/publications/reports/EPA_600_R02_068.pdf> (accessed 07.10.2009).

United States Environmental Protection Agency (USEPA), 2006. Ultraviolet Disinfection

Guidance Manual for the Final Long Term 2 Enhanced Surface Water Treatment Rule.

Washington, DC.

United States Environmental Protection Agency (USEPA), 2009. National Primary Drinking

Water Regulations. Available from <http://water.epa.gov/drink/contaminants/index.cfm#List>

(accessed 07.10.2009).

Venczel, L.V., Arrowood, M., Hurd, M., Sobsey, M., 1997. Inactivation of Cryptosporidium

parvum oosysts and Clostridium perfringens spores by a mixed-oxidant disinfectant and by

free chlorine. Applied and Environmental Microbiology, 63(4), 1598–1601.

Vogt, R., Schindler, R.N., 1992. Product channels in the photolysis of HOCl. Journal of

Photochemistry and Photobiology A: Chemistry, 66(2), 133–140.

35

von Gunten, U., Hoigné, J., 1994. Bromate formation during ozonation of bromide-containing

waters: interaction of ozone and hydroxyl radical reactions. Environmental Science &

Technology, 28(7), 1234–1242.

Wang, L.K., Hung, Y.-T., Shammas, N.K., 2006. Handbook of Environmental Engineering,

Volume 4: Advanced Physicochemical Treatment Processes. The Humana Press Inc., Totowa,

NJ, USA.





56(8), 469–477.

Weng, S., Li, J., Blatchley III, E.R., 2012. Effects of UV254 irradiation on residual chlorine and

DBPs in chlorination of model organic-N precursors in swimming pools. Water Research,

46(8), 2674–2682.

Westerhoff, P., Chao, P., Mash, H., 2004. Reactivity of natural organic matter with aqueous

chlorine and bromine. Water Research, 38(6), 1502–1513.

Zhang, X., Li, W., Blatchley III, E.R., Wang, X., Ren, P., 2015. UV/chlorine process for

ammonia removal and disinfection by-product reduction: Comparison with chlorination.

Water Research, 68(1), 804–811.

36

3. MEDIUM PRESSURE UV COMBINED WITH CHLORINE

ADVANCED OXIDATION FOR TRICHLOROETHYLENE

DESTRUCTION IN A MODEL WATER

Most of this chapter has been previously published as:

Wang, D., Bolton J.R., Hofmann, R., 2012. Medium pressure UV combined with chlorine advanced oxidation

for trichloroethylene destruction in a model water. Water Research, 46(15), 4677–4686.

Reproduction of the paper has been permitted by the publisher.

Abstract The effectiveness of ultraviolet (UV) combined with chlorine as a novel advanced

oxidation process (AOP) for drinking water treatment was evaluated in a bench scale study by

comparing the rate of trichloroethylene (TCE) decay when using UV/chlorine to the rates of

decay by UV alone and UV/hydrogen peroxide (H2O2) at various pH values. The UV/chlorine

process was more efficient than the UV/H2O2 process at pH 5, but in the neutral and alkaline pH

range, the UV/H2O2 process became more efficient. The pH effect was probably controlled by

the increasing concentration of OCl– at higher pH values. A mechanistic kinetic model of the

UV/chlorine treatment of TCE showed good agreement with the experimental data.

3.1 Introduction Advanced oxidation processes (AOPs) involve the generation of the hydroxyl radical

(•OH), a very strong and non-selective chemical oxidant, that can destroy organic contaminants

in water and wastewater which are not readily oxidized by conventional oxidants (Parsons,

2004; Gültekin and Ince, 2007; Suty et al., 2004). Ultraviolet (UV)-based AOPs are becoming

more common in water treatment, and typically involve the addition of hydrogen peroxide

(H2O2) prior to the UV reactor (the UV/H2O2 AOP), with the photolysis of H2O2 forming •OH

(Parsons, 2004).

Ideally, a compound selected for photolysis in a UV-based AOP to generate •OH should be

able to absorb UV light efficiently and then produce a high yield of •OH. Although H2O2

photolysis gives an •OH formation quantum yield (Φ) of 1.11 ± 0.07 (Goldstein et al., 2007),

37

H2O2 absorbs UV only weakly. For example, the molar absorption coefficient of H2O2 at 254

nm, the emission peak of a low-pressure (LP) mercury UV lamp as a monochromatic light

source, is only ~19 M–1 cm–1 (Stefan et al., 1996; Baxendale and Wilson, 1957).

It may be possible to use the photolysis of aqueous active chlorine (i.e., hypochlorous acid,

HOCl, and/or hypochlorite ion, OCl–) as an alternative UV-based AOP, since active chlorine

species absorb UV photons and produce •OH relatively efficiently under some conditions

(Watts and Linden, 2007; Watts et al., 2007). HOCl and OCl– absorb UV light more efficiently

than H2O2 when using typical LP lamps or medium-pressure (MP) UV lamps (a polychromatic

light source emitting UV from 200 to 400 nm), according to their absorption spectra shown in

Figure 3.1. In general, when exposed to a LP lamp, HOCl and OCl– both absorb ~3.3 times

more UV photons than H2O2 at the same molar concentration. In the case of the MP lamp used

in this study, with the relative spectral emittance shown in Figure 3.2, HOCl and OCl– absorb

UV light 2.3 and 10.7 times more efficiently than H2O2, respectively, under equal molar

conditions, because of the high absorbance of OCl– at 290–320 nm. Compared to the well-

established mechanisms of H2O2 photolysis, however, the mechanisms of aqueous HOCl/OCl–

photolysis have not been fully explored. Evidence suggests that active chlorine

photodecomposition involves a series of chain reactions with the formation of many

intermediates and products (Buxton and Subhani, 1972a, 1972b and 1972c). Both •OH and

chlorine radicals (•Cl) are produced from photolysis at wavelengths less than 400 nm via

Equations [3.1] – [3.3] (Nowell and Hoigné, 1992a, 1992b; Oliver and Carey, 1977; Feng et al.,

2007; Watts and Linden, 2007).

HOCl + hv → •OH + •Cl [3.1]

OCl– + hν → •Cl + •O– [3.2]

•O– + H2O ↔ •OH + OH– [3.3]

Compared to •OH, •Cl was observed to play a negligible role in the oxidation of organic probes

(e.g., nitrobenzene and 1-chlorobutane), based on work performed by Nowell and Hoigné

(1992b) and Watts and Linden (2007).

The values reported in the literature for the quantum yields of •OH formation by HOCl and

OCl– photolysis vary. Watts and Linden (2007) and Watts et al. (2007) reported quantum yields

38

of 1.4 and 0.28 for HOCl and OCl–, respectively, while Jin et al., (2011) and Chan et al. (2012)

reported values of 0.46 and 0.61. The reasons for these discrepancies are not clear.

Figure 3.1 Absorption spectra of TCE, chlorine, peroxide and hydroxide species

Figure 3.2 Relative spectral emittance of the MP lamp in this research

0

200

400

600

800

1000

200 250 300 350 400

Wavelength (nm)

Mol

ar a

bsor

ptio

n co

effic

ient

s of

H

OC

l, O

Cl– , H

2O2,

HO

2– and

OH

–

(M–1

cm

–1)

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

Mol

ar a

bsor

ptio

n co

effic

ient

of

TCE

(M–1

cm

–1)

HO2–

OH–

HOCl

OCl–

H2O2

TCE

0

5

10

15

20

200 250 300 350 400

Wavelength (nm)

Rel

ativ

e sp

ectra

l em

ittan

ce

39

While HOCl and OCl– photolysis can produce •OH, they are also •OH scavengers as

shown in Equations [3.4] and [3.5], with reaction rate constants of 8.46 × 104 (Watts and

Linden, 2007) and 9.0 × 109 M–1 s–1 (Buxton and Subhani, 1972a), respectively.

•OH + HOCl → H2O + ClO• k = 8.46 × 104 M–1 s–1 [3.4]

•OH + OCl– → ClO• + OH– k = 9.0 × 109 M–1 s–1 [3.5]

Although both HOCl and OCl– photolysis can produce •OH, the concentrations of •OH

generated by the photolysis of HOCl and OCl– for equimolar concentrations are different,

because of their different molar absorption coefficients (shown in Figure 3.1), the different

quantum yields of •OH formation, and the different •OH scavenging efficiency for HOCl and

OCl–. Therefore, the efficiencies of HOCl and OCl– photolysis used as AOPs may not be the

same. Since the components of aqueous active chlorine are pH dependent [(pKa for HOCl at

25°C is 7.54 according to Deborde and von Gunten (2008)], the effectiveness of the

UV/chlorine AOP is therefore predicted to be sensitive to pH.

While there are still many unknowns about the specifics of active chlorine photolysis,

enough information is available to suggest that it is relatively efficient at absorbing UV and

produces •OH on photolysis. The goal of this study was to confirm the theoretical effectiveness

of the UV/chlorine AOP in the context of drinking water treatment, using the destruction of

trichloroethylene (TCE) as an example.

TCE (ClHC=CCl2, molecular weight: 131.39 g mol–1) is susceptible to •OH oxidation with

a reaction rate constant of 2.4 × 109 M–1 s–1 (Li et al., 2007; Li et al., 2004). The Middleton

Water Supply System in Waterloo (Ontario, Canada) has already explored the application of the

UV/chlorine process at full scale to remove TCE from groundwater (Wang et al., 2011). This

study is intended to verify further its feasibility and to explore the UV/chlorine treatment from a

fundamental perspective under controlled laboratory conditions. Bench-scale experiments using

a MP collimated beam apparatus and a mathematical modeling analysis were both performed to

evaluate the effectiveness of the UV/chlorine AOP, comparing it to the more conventional

UV/H2O2 AOP under parallel conditions.

40

3.2 Materials and Methods

3.2.1 Reagents and Materials

An appropriate volume of TCE (≥99.5%, A.C.S. grade, Sigma-Aldrich) was diluted in

Milli-Q® water to prepare the working solutions to be exposed at a TCE concentration of

approximately 145 μg L–1 (1.10 × 10–6 M). Active chlorine and hydrogen peroxide solutions

were prepared from sodium hypochlorite (NaOCl) solution (10–15 wt. %, reagent grade, Sigma-

Aldrich) and H2O2 solutions (50 wt. %, Sigma-Aldrich), respectively. The concentrations of

active chlorine and H2O2 in the working solutions were both approximately 0.15 mM. Other

compounds used in this research were all analytical reagent grade. Milli-Q® water was used in

all experiments and analytical determinations. The Milli-Q® water contained a low

concentration of total organic carbon (TOC) (approximately 0.1 mg L–1 as C), which was

included in the modeling studies.

3.2.2 UV Exposure and Irradiance Measurements

A 1 kW MP mercury UV lamp (Heraeus Noblelight GmbH, Germany) installed in a

collimated beam apparatus (Model: PS1-1-120, Calgon Carbon Corporation) was used to expose

15 mL TCE samples with/without the addition of active chlorine or H2O2 contained in Pyrex®

Petri dishes (inner diameter: 4.9 cm). The samples were buffered at pH 5, 7.5 or 10 using 5 mM

phosphate and/or borate buffers. The exact pH was adjusted by adding NaOH and/or H2SO4.

According to the pKa of HOCl at 25°C, at pH 5, 7.5 and 10, the active chlorine components

consisted of 99.7% HOCl, 52.3% HOCl + 47.7% OCl–, and 99.6% OCl–, respectively. An

appropriate length of exposure time was performed to deliver the desired fluence into each

sample. The incident irradiance (mW cm–2) of samples from 200 to 345 nm was calculated from

the difference measured by ferrioxalate actinometry with and without a 3 mm-thick 345 nm

long-pass filter placed in the light path, according to the procedures described by Bolton et al.

(2009) and Sharpless and Linden (2003). The different quantum yields of Fe2+ formation in the

ferrioxalate system at the various wavelengths were also considered (Goldstein and Rabani,

2008). After obtaining the relative spectral emittance of the MP lamp (shown in Figure 3.2), the

fluence rate from 200 to 400 nm was calculated proportionally (shown in Appendix B). The

41

correction factors, such as Petri factor, water factor, divergence factor and reflection factor,

were also applied as per the procedures described by Bolton and Linden (2003).

3.2.3 Analytical Methods

The active chlorine (HOCl and/or OCl–) concentrations in working solutions were

determined by the DPD colorimetric method (APHA et al., 2005), using a HACH®

spectrophotometer (Model: DR/2500, HACH). H2O2 concentrations were analyzed with the

triiodide method (Klassen et al., 1994). Gas chromatography-electron capture detector (GC-

ECD) (Model: HP 5890 Series II, Hewlett-Packard) was applied for measurement of TCE,

following the USEPA method 551.1 (USEPA, 2008). The method detection limit of TCE was 1

μg L–1. Details for the analytical method are shown in Appendix G. Chlorine and H2O2 residuals

were quenched by sodium sulphite (Na2SO3) before TCE analysis. The relative spectral

emittance of the MP lamp was measured using a calibrated spectroradiometer (Model:

USB4000-UV-VIS, Ocean Optics) with a fiber-optic cable (Model: QP200-2-SR-BX, Ocean

Optics) and a cosine corrector (Model: CC-3-UV, Ocean Optics). In addition, a Cecil UV/vis

spectrophotometer (Model: CE3055, Cecil Instruments) and a PerkinElmer UV/vis

spectrophotometer (Model: Lambda 25, PerkinElmer) were applied for measuring solution

absorbances at a single wavelength and in the entire UV band, respectively.

3.3 Results and Discussion

3.3.1 Molar Absorption Coefficients of TCE, Active Chlorine, Peroxide and Hydroxide

Species

As part of this work, mathematical models of the UV/chlorine and the UV/H2O2 AOPs

were built to interpret experimental results. These models required measurement of the molar

absorption coefficients of TCE, HOCl (at pH 5.0), OCl– (at pH 10.0) and H2O2 (at pH ~5.5) at

wavelengths ranging from 200 to 400 nm, shown in Figure 3.1. The maximum molar absorption

coefficients of HOCl and OCl– were 98 and 359 M–1 cm–1, peaking at 235 and 292 nm,

respectively. These values are very close to those measured by Feng et al. (2007), who found

HOCl and OCl– peak molar absorption coefficients of 101 and 365 M–1 cm–1 at 236 and 292 nm,

respectively. The molar absorption coefficient of HO2– was also required because H2O2

disassociates in strong alkaline solutions (Equation [3.6]) (Czapski and Bielski, 1963).

42

H2O2 ↔ H+ + HO2– pKa = 11.8 [3.6]

Unlike other compounds in this study, however, the absorbance of HO2– could not be measured

directly, because at pH near 13.8, where ~100% H2O2 converts to HO2–, the absorbance of OH–

at the wavelengths <220 nm is too high to allow measurement of the HO2– absorbance

accurately (e.g., OH– absorbance at 200 nm >1.9, according to Figure 3.1). As a result, in order

to determine the molar absorption coefficient of HO2–, the absorbance of a mixture of H2O2 and

HO2– containing a low concentration of OH– (1 mM) was first measured, followed by accurately

measuring the total concentration of peroxide species and the pH. Using the pKa of H2O2, the

respective concentrations of H2O2 and HO2– were subsequently calculated. After the molar

absorption coefficients of H2O2 and OH– were obtained, the corresponding values (also shown

in Figure 3.1) for HO2– were calculated. Although the values may not be as accurate as those

measured directly, they are generally reliable, compared with the data at several selected

wavelengths reported by Baxendale and Wilson (1957). For example, the molar absorption

coefficients of HO2– at 254, 270, and 290 nm were calculated to be 269, 145, 31 M–1 cm–1,

respectively, compared with 229, 122 and 45 M–1 cm–1 at the corresponding wavelengths

directly measured by Baxendale and Wilson (1957).

3.3.2 Quantum Yields of Active Chlorine and Hydrogen Peroxide Photolysis

When performing the exposure experiments for TCE decay by the UV/chlorine and the

UV/H2O2 AOPs using the collimated beam apparatus, the fluence-based decay rate s for active

chlorine and H2O2 (Figure 3.3 and Table 3.1) first had to be determined to estimate the quantum

yields of active chlorine and H2O2 photolysis, following methods reported by Bolton and Stefan

(2002) and Stefan and Bolton (2005). Details for chlorine and H2O2 photolysis are included in

Appendix I. It is noted that these observed quantum yields are specific to the water matrix, and

in particular the presence of •OH scavengers. The •OH scavengers may either inhibit the

photodegradation of active chorine or hydrogen peroxide, since the scavengers consume •OH

and diminish the reactions between •OH and active chlorine components or hydrogen peroxide;

or in contrast, may enhance their photolysis, since the products of scavengers may further react

with active chlorine components and/or hydrogen peroxide. For example, Feng et al. (2010) and

43

Jin et al., (2011) found that in the presence of TOC and methanol, the photodegradation of

HOCl is enhanced and thus increases the observed photolysis quantum yield.

Figure 3.3 Rates of active chlorine (a) and hydrogen peroxide (b) photolysis at various pH

values. Error bars represent the standard deviations of triplicate runs. Straight lines

represent the linear regression.

The observed quantum yields determined here for HOCl, OCl– and H2O2 photolysis in the

water matrices studied are 1.06 ± 0.01, 0.89 ± 0.02 and 0.76 ± 0.01, respectively. The chlorine

quantum yields match closely the values reported by Feng et al. (2007), but are considerably

-1.2

-1

-0.8

-0.6

-0.4

-0.2

0

0 5000 10000 15000 20000

Fluence (J m–2)

ln([a

ctiv

e C

l]/[a

ctiv

e C

l] 0) pH 5 pH 7.5 pH 10

(a)

-0.12

-0.1

-0.08

-0.06

-0.04

-0.02

0

0 5000 10000 15000 20000

Fluence (J m–2)

ln([H

2O2]

/[H2O

2]0)

pH 5 pH 7.5 pH 10

(b)

44

different from some of the other reported quantum yields (Table 3.2). Likewise, the quantum

yield for hydrogen peroxide photolysis in this study (0.76) is lower than the published value of

~1.0 (Hunt and Taube, 1952; Baxendale and Wilson, 1957; Volman and Chen, 1959; Goldstein

et al. 2007). The difference of the H2O2 photolysis quantum yields between this study and the

literature may arise from the presence of TCE in this study, compared to organic-free water used

in the other studies. Like chlorine species, the photolysis of H2O2 may involve chain reactions

(Crowell et al., 2004; Luňák and Sedlák, 1992). The •OH formed by H2O2 photolysis may react

with H2O2, resulting in an increase in the observed quantum yield of H2O2 photolysis.

Therefore, the presence of •OH scavengers (such as TCE) may induce a lower observed

photolysis quantum yield. Indeed, results from other trials in this study (not reported) indicated

that the quantum yield of H2O2 photolysis increased by ~10% in the absence of TCE. In

addition, the previous studies reporting the quantum yield of H2O2 photolysis were also

conducted exclusively using monochromatic LP lamps, as opposed to the polychromatic MP

lamp used in this study. The calculated quantum yield of photolysis would therefore differ if

they were to vary with wavelength over the emission spectrum of the MP lamp. Further research

is needed under controlled conditions (e.g. well-defined •OH scavenging, single wavelengths,

etc.) to confirm the quantum yields of active chlorine or hydrogen peroxide photolysis in the

absence of water matrix and lamp effects.

Table 3.1 Fluence-based rate constants (10–6 m2 J–1) for active chlorine and peroxide

photolysis

pH 5 pH 7.5 pH 10

Chlorine species 15.0 ± 0.1 33.7 ± 0.4 58.7 ± 1.5 H2O2 species 4.74 ± 0.09 4.77 ± 0.04 5.67 ± 0.02

3.3.3 TCE Decay Rates by UV Alone, and the UV/Chlorine and the UV/H2O2 AOPs

TCE decay rates were measured in the presence of UV alone, as well as in the UV/chlorine

and the UV/H2O2 AOPs. TCE is sensitive to both UV exposure and •OH oxidation, and is also

volatile (Li et al., 2004; Li et al., 2007; Peng and Wan, 1997). The measured decrease of TCE

concentration therefore arises from the sum of evaporative loss, direct photolysis by UV

45

exposure, and •OH oxidation loss (if AOPs are applied). In a dilute aqueous solution, the change

of a solute concentration arising from evaporation is first-order (Tinsley, 2004). Since the rate of

TCE photolysis and the rate of TCE decay by •OH oxidation also follow first-order kinetics

(assuming steady-state •OH concentration), the total rate of TCE loss follows first-order

kinetics, as described in Equation [3.7] (Schwarzenbach et al., 2003; Rosenfeldt and Linden,

2004; Rosenfeldt, 2005).

Table 3.2 Comparison of reported quantum yields of active chlorine photolysis

Chlorine species

Quantum yield

Experimental conditions Reference Initial chlorine

concentration wavelengt

h (nm) pH

OCl– 0.60 ± 0.02 7 mM 365 12.1 Buxton and Subhani (1972b)



HOCl 1.0 ± 0.1 < 2mM 254 5 Feng et al. (2007)

OCl– 0.9 ± 0.1 < 2 mM 254 10 Feng et al. (2007)

HOCl 1.5 0.014 – 0.056 mM 254 4 Watts and Linden (2007)

OCl– 1.3 0.014 – 0.056 mM 254 10 Watts and Linden (2007)

HOCl 3.7 0.014 – 0.056 mM MP lamp 4 Watts and Linden (2007)

OCl– 1.7 0.014 – 0.056 mM MP lamp 10 Watts and Linden (2007)

HOCl 1.0 ± 0.1 1.41 mM LP lamp 5 Jin et al. (2011)

OCl– 1.15 ± 0.08 1.41 mM LP lamp 10 Jin et al. (2011)

OCl– 0.87 ± 0.01 0 – 4.23 mM 303 10 Chan et al. (2012)

HOCl 1.06 ± 0.01 0.15 mM MP lamp 5 This work

OCl– 0.89 ± 0.02 0.15 mM MP lamp 10 This work

evaporation UV OH0

ln ( )C k k k tC •= − + + [3.7]

where, C is the final TCE concentration (M) after the exposure, C0 is the initial TCE

concentration (M) before the exposure, kevaporation is the first-order evaporative rate constant (s–1)

of TCE, kUV is the first-order photolysis rate constant (s–1) of TCE due to UV exposure, k•OH is

the first-order decay rate constant (s–1) of TCE due to •OH oxidation, and t is the exposure time

46

(s). kevaporation was determined from the reduction of TCE concentration of a 15 mL sample (with

no addition of active chlorine or H2O2) that was placed in the same Petri dish used for

experiments and stirred for 5 min (the longest time in exposure experiments). According to 10

replicates, 25.0 ± 0.51% of initial TCE concentration was lost due to evaporation. Therefore, the

evaporative rate was 9.6 × 10–4 s–1. After subtracting the evaporation loss from the total loss of

TCE, the TCE destruction rate arising from photolysis and •OH oxidation is calculated using

Equation [3.8]. Details for TCE decay after subtracting the evaporation loss are shown in

Appendix I.

evaporation UV OH0

ln ( )C k t k k tC •+ = − + [3.8]

Since the fluence (mJ cm–2) delivered into the sample is the product of the exposure time (s) and

the fluence rate (mW cm–2) in the same wavelength range, Equation [3.8] is equivalent to:

evaporation UV OH0

ln ( )C k t k k FC •′ ′+ = − + [3.9]

where F is the fluence (mJ cm–2) from 200 to 400 nm, and UVk′ and OHk•′ are fluence-based rate

constants (cm2 mJ–1) (Bolton and Stefan, 2002; Stefan and Bolton, 2005), while the unit for

kevaporation is still s–1. The fluence rate is considered to be constant during the exposure, since in

low absorbance solutions the decrease of solute concentrations impacts the fluence rate only

slightly. Therefore, evaporation0

ln C k tC

+ is a linear function of F with a slope of ( )UV OHk k•′ ′− + . The

experimental results are shown in Figure 3.4. The TCE decay rate constants with the standard

deviations of triplicates after subtracting out evaporation under different conditions are

summarized in Table 3.3. The corresponding values calculated from numerical models are also

shown in Table 3.3, and will be discussed later.

47

Figure 3.4 Experimental TCE decay rates by UV alone, the UV/chlorine and the UV/H2O2

AOPs at various pH values. Error bars represent the standard deviations of triplicate

runs. Straight lines represent the linear regression.

-3

-2

-1

0

0 500 1000 1500 2000

Fluence (mJ cm–2)

ln(C

/C0)

+ k

evap

orat

ion

t

UV alone UV+Chlorine UV+H2O2

pH 5

UV+H2O2

-3

-2

-1

0

0 500 1000 1500 2000

Fluence (mJ cm–2)

ln(C

/C0)

+ k

evap

orat

ion

t

UV alone UV+chlorine UV+H2O2

pH 7.5

UV+H2O2

-3

-2

-1

0

0 500 1000 1500 2000

Fluence (mJ cm–2)

ln(C

/C0)

+ k

evap

orat

iont

UV alone UV+Chlorine UV+H2O2

pH 10

UV+H2O2

48

Table 3.3 TCE Fluence-based decay rate constants (10–4 cm2 mJ– 1) (excluding evaporation)

by UV alone, and the UV/chlorine and the UV/H2O2 AOPs from experimental and model

results

pH 5 pH 7.5 pH 10

Experimental Model Experimental Model Experimental Model

UV alone 7.23 ± 0.25 5.30 6.03 ± 0.05 5.30 5.25 ± 0.03 5.30

UV/chlorine AOP 80.9 ± 1.0 54.4 8.15 ± 0.03 7.89 7.33 ± 0.20 7.68

UV/H2O2 AOP 34.7 ± 1.1 35.3 32.3 ± 0.8 35.2 9.69 ± 0.09 11.8

According to Figure 3.4 and Table 3.3, TCE is photolyzed by UV exposure at a fairly high

photolysis rate (i.e., a fairly high kUV). For example, more than 60% of TCE is destroyed at a

fluence of approximately 1,900 mJ cm–2 from 200 to 400 nm. Based on the rates of TCE

photolysis (in the case of UV alone) in Table 3.3, the observed quantum yields of TCE

photolysis at pH 5, 7.5 and 10 were calculated to be 0.64 ± 0.02, 0.53 ± 0.04 and 0.46 ± 0.03,

respectively, following the methods discussed by Bolton and Stefan (2002) and Stefan and

Bolton (2005). This is in general agreement with Li et al. (2004), who indicated that TCE is

fairly sensitive to UV exposure, comprising several photochemical pathways with an overall

quantum yield of 0.354 at a neutral pH. Interestingly, however, as the pH increases, the TCE

photolysis rate constant declines slightly. The reason is not clear. It may be because in an

alkaline solution more OH– is available to scavenge •Cl as shown in Equation [3.10], which is

an intermediate produced by TCE photolysis and promotes the destruction of TCE (Kläning and

Wolff, 1985; Li et al., 2004).

•Cl + OH– → •ClOH– k = 1.8 × 1010 M–1 s–1 [3.10]

From Figure 3.4 and Table 3.3, the UV/chlorine and the UV/H2O2 AOPs were found to be

more efficient at destroying TCE than UV alone because of the additional loss of TCE by •OH

oxidation. When comparing the TCE decay rates by the UV/chlorine and the UV/H2O2 AOPs at

the same pH, the UV/chlorine AOP is observed to be 2.3 times more efficient for destruction of

TCE than the UV/H2O2 AOP at pH 5, while it is less efficient than the latter at pH 7.5 and 10.

49

Moreover, it is obvious that the efficiencies of both AOPs become lower at a higher pH because

the TCE decay rates by both AOPs decrease with increasing pH. In particular, the efficiency of

the UV/chlorine AOP decreases significantly faster than that of the UV/H2O2 AOP.

The different efficiencies of the UV/chorine versus UV/H2O2 processes are largely a

reflection of different steady-state concentrations of hydroxyl radicals, [•OH]ss, with a higher

[•OH]ss giving a faster rate of TCE decay. [•OH]ss can be calculated using Equation [3.11]

(Rosenfeldt and Linden, 2004).

k•OH = k•OH-TCE × [•OH]ss [3.11]

where, k•OH is the first-order decay rate constant (s–1) of TCE due to •OH oxidation, k•OH-TCE is

the second-order rate constant (M–1 s–1) between •OH and TCE, k•OH-TCE = 2.4 × 109 M–1 s–1, and

[•OH]ss is the steady-state concentration (M) of •OH. k•OH can be determined experimentally

using values from Table 3.3 by subtracting the reaction rate constants of TCE photolysis, kUV,

from the overall reaction rate constants of TCE decay observed in the UV/chorine or UV/H2O2

experiments. It is noted that the k•OH calculated this way has units of cm2 mJ–1, which is different

from the units (s–1) used in Equation [3.11]. Therefore, in order to fit Equation [3.11], k•OH must

be converted from cm2 mJ–1 to s–1 by multiplying the value of k•OH by the average fluence rate

(mW cm–2) in the solution. Table 3.4 presents the calculated [•OH]ss in the UV/chlorine and the

UV/H2O2 AOP tests at different pHs, based on Table 3.3 and Equation [3.11]. In addition to

Equation [3.11], assuming that the quantum yield of •OH formation is independent of

wavelength, [•OH]ss can be theoretically determined from Equation [3.12], as described by

Schwarzenbach et al. (2003).

( )OH

ssOH-

( ) [1 10 ][Ox]

[ OH][ ]

a zp

jj

Ea z

k j

λλ

λ λ

λ εΦ

−

•

•

−⋅ ⋅

• =⋅

∑∑

[3.12]

where, Φ•OH is the quantum yield of •OH formation by chlorine or hydrogen peroxide

photolysis, Ep(λ) is the incident photon irradiance (10–3 einstein cm–2 s–1) at wavelength λ, ελ is

50

the molar absorption coefficient (M–1 cm–1) of the chlorine or hydrogen peroxide species at

wavelength λ, aλ is the decadic absorption coefficient (cm–1) of the TCE solution at wavelength

λ, z is the solution depth (cm), [Ox] is the concentration (M) of the oxidant, i.e., chlorine or

hydrogen peroxide species, k•OH-j is the second-order reaction rate constants (M– 1 s– 1) between

•OH and a scavenger, and [j] is the concentration (M) of the corresponding •OH scavenger j.

Table 3.4 Calculated hydroxyl radical concentrations (10–13 M) in TCE solutions treated

by the UV/chlorine and the UV/H2O2 AOPs at various pH values

pH 5 pH 7.5 pH 10

UV/chlorine AOP 197 ± 2.8 5.58 ± 0.16 5.48 ± 0.54 UV/H2O2 AOP 73.3 ± 3.0 70.1 ± 2.1 11.8 ± 0.2

According to Equation [3.12] and the calculated [•OH]ss for the UV/chlorine AOP at pH 5

and 10 and the UV/H2O2 AOP at pH 5 shown in Table 3.4, the observed quantum yields of •OH

formation by the photolysis of HOCl, OCl– and H2O2 are determined to be 0.79 ± 0.01, 1.18 ±

0.12 and 1.15 ± 0.05, respectively. The value for H2O2 (1.15 ± 0.05) determined in this study is

very close to the literature (1.11 ± 0.07) (Goldstein et al., 2007), while the values for HOCl and

OCl– (0.79 ± 0.01 and 1.18 ± 0.12, respectively) are quite different from other published data:

e.g., 1.40 and 0.28, respectively (Watts and Linden, 2007; Watts et al. 2007), or 0.46 (Jin et al.,

2011) and 0.61 (Chan et al., 2012). This difference could arise from the different experimental

conditions in these studies, such as active chlorine concentrations, •OH scavenging potentials,

UV lamp types, etc.

According to Equation [3.12], the variation of [•OH]ss with pH arises theoretically from the

difference in Φ•OH, ελ and/or k•OH-j[j]. For the UV/chlorine AOP, the components of chlorine

species at pH 5, 7.5 and 10 are very different (pKa = 7.54). Although OCl– absorbs more photons

than HOCl and the Φ•OH for OCl– is higher than that for HOCl, the much higher reported

reaction rate of OCl– with •OH than that for HOCl (reaction rate coefficient: 9.0 × 109 M–1s–1 for

OCl– vs. 8.46 × 104 M–1s–1 for HOCl) makes the UV/chlorine AOP less efficient with increasing

pH. For the UV/H2O2 AOP, although the quantum yield of •OH formation from HO2–

photolysis is unknown, it is likely similar to that of H2O2. Therefore, the lower efficiency of the

UV/H2O2 AOP at a higher pH may also arise principally from the much higher (~278 times)

51

•OH scavenging efficiency of HO2– than H2O2 as shown in Equations [3.13] and [3.14] (Stefan

et al., 1996).

H2O2 + •OH → HO2• + H2O k = 2.7 × 107 M–1 s–1 [3.13]

HO2– + •OH → •O2

– + H2O k = 7.5 × 109 M–1 s–1 [3.14]

In addition, because of a relatively high pKa for H2O2 disassociation equilibrium constant

compared to that of HOCl (pKa = 11.8 for H2O2 vs. 7.54 for HOCl), only a small fraction (1.6%)

of H2O2 is converted to HO2– at pH 10. The result of these factors is that the efficiency of the

UV/chlorine AOP decreases more quickly with increasing pH than that of the UV/H2O2 AOP.

3.3.4 Mathematical Modeling of the TCE Decay

Numerical models were built with Matlab® to simulate TCE decay by UV alone, as well as

by the UV/chlorine and the UV/H2O2 AOPs. The TCE decay rate arising from UV exposure was

first predicted based on the quantum yields of TCE photolysis through different pathways

discussed by Li et al. (2004). The TCE destruction rates by the UV/chlorine and the UV/H2O2

AOPs were subsequently simulated according to Equations [3.8], [3.11] and [3.12]. The

parameters required in the models (shown in Table 3.5), such as the reaction rate constants of

•OH with the scavengers, the quantum yield of H2O2 photolysis and the quantum yield of •OH

formation by H2O2 photolysis, were obtained from the published literature. However, because of

the discrepancy in the literature (Watts and Linden, 2007; Watts et al. 2007; Jin et al., 2011;

Chan et al., 2012) for the quantum yields of HOCl and OCl– photolysis, and the quantum yields

of •OH formation by their photolysis, the parameters determined by this study were used in the

models. The •OH scavenging of total organic carbon (TOC) was also included, since

approximately 0.1 mg L–1 as carbon of TOC is present in the Milli-Q® water. Radical

scavenging by inorganic carbon and buffers was ignored, as it was calculated that they would

contribute to less than 2% of the •OH scavenging potential. Examples of Matlab® codes are

shown in Appendix C. A simpler and more approximate estimation can be made using a

Microsoft Excel spreadsheet, shown in Appendix D.

The TCE decay rate constants by UV alone, UV/chlorine, and UV/H2O2 simulated by the

models are shown in Table 3.3. Compared to the experimental results in Table 3.3, the models

52

generally give good predictions, suggesting that the reactions with their parameters described in

Table 3.5 are generally accurate.

One useful application of the model is to predict the solution pH at which the UV/chlorine

and the UV/H2O2 AOPs are equally efficient: that is, the same [•OH]ss exists in both scenarios.

Based on Equation [3.12], [•OH]ss depends on the scavenging potential in solutions. Assuming

that the TOC is the only •OH scavenger other than 0.15 mM active chlorine and the hydrogen

peroxide species, the impact of pH and TOC on the relative efficiency of the UV/chlorine AOP

versus the UV/H2O2 AOP is shown in Figure 3.5. It is evident that in pure waters, such as those

used for these experiments, with minimal background •OH scavenging, the UV/H2O2 AOP is

more efficient than the UV/chlorine AOP at all but very low pH values (i.e. less than 5.3). If the

water contains more •OH scavengers, the UV/chlorine AOP becomes more competitive at a

higher pH. For example, it is predicted that the UV/chlorine AOP will be equally efficient as the

UV/H2O2 AOP at pH values higher than 7.0 provided that the TOC is approximately 5.0 mg/L

or above.

3.3.5 Comment on Active Chlorine Reaction with •OH

An important element in modeling the UV/chlorine AOP is the reported reaction rate

constant between •OH and HOCl or OCl–, with the reaction reportedly 5 orders of magnitude

faster with OCl– than with HOCl. This was used to explain the observed reduced UV/chlorine

AOP efficiency at higher pH: the OCl– became a dominant •OH scavenger. Feng et al. (2007),

however, reported that the quantum yield of HOCl photolysis increased with increasing HOCl

concentration. This suggests a chain reaction that is initiated by the reaction between •OH and

HOCl, which would likely only occur if there were a fast reaction between •OH and HOCl.

Feng et al. (2007) also reported that the quantum yield of OCl– photolysis was not a function of

OCl– concentration, implying a possible slow reaction between •OH and OCl–. These results

contradict the earlier research. It is therefore clear that more fundamental research is needed to

clarify the mechanism of active chlorine photolysis. Appendix A shows a preliminary trial using

pulse radiolysis analysis to determine the rate constant between HOCl/OCl– and •OH. The

results successfully verified the rate constant of OCl– with •OH reported by Buxton and Subhani

(1972a), but cannot be used to determine the rate constant between HOCl and •OH.

53

Table 3.5 Reaction mechanisms of TCE decay by UV alone, the UV/chlorine and the

UV/H2O2 AOPs

No. Reaction Rate constant Reference

UV alone

3.15 TCE + hν → ClHC=C•Cl + Cl• ΦTCE,1 = 0.13 Li et al. (2004)

3.16 TCE + hν → ClHC(OH)CHCl2 ΦTCE,2 = 0.1 Li et al. (2004)

3.17 TCE + hν → HC≡CCl + Cl2 ΦTCE,3 = 0.032 Li et al. (2004)

3.18 TCE + hν → ClC≡CCl + HCl ΦTCE,4 = 0.092 Li et al. (2004)

3.19 TCE + Cl• → Cl2HC-C•Cl2 4.88 × 1010 M–1 s–1 Li et al. (2004)

UV/chlorine AOP Reactions [3.15] – [3.19], and 3.20 OCl– + H2O ↔ HOCl + OH– kforward = 1.8 × 103 s–1

kreverse = 3 × 109 M–1s–1 Fogelman et al. (1989)

3.1 HOCl + hv → •OH + •Cl ΦHOCla = 1.06

Φ•OHb = 0.79

This work

3.21 OCl– + hν → •OH + other products ΦOCl–c = 0.89 Φ•OH

b = 1.18 This work

3.4 •OH + HOCl → H2O + ClO• 8.46 × 104 M–1 s–1 Watts and Linden (2007)

3.5 •OH + OCl– → ClO• + OH– 9.0 × 109 M–1 s–1 Buxton and Subhani (1972a)

3.22 TOC + •OH → products 3 × 108 M–1 s–1 Westerhoff et al. (1999)

3.23 TCE + •OH → ClCH(OH)-C•Cl2 2.4 × 109 M–1 s–1 Li et al. (2007)

UV/H2O2 AOP Reactions [3.15] – [3.19], [3.22] – [3.23], and 3.6 H2O2 ↔ H+ + HO2

– kforward = 0.126 s–1

kreverse = 5 × 1010 M–1s–1 Song (1996)

3.24 H2O2 + hv → 2•OH ΦH2O2d = 1.0

Φ•OHb = 1.11

Stefan et al. (1996) Goldstein et al. (2007)

3.13 H2O2 + •OH → HO2• + H2O 2.7 × 107 M–1 s–1 Stefan et al. (1996)

3.14 HO2– + •OH → •O2

– + H2O 7.5 × 109 M–1 s–1 Stefan et al. (1996)

aΦHOCl is the quantum yield of HOCl photolysis bΦ•OH is the quantum yield of •OH formation cΦOCl

– is the quantum yield of OCl– photolysis dΦH2O2 is the quantum yield of H2O2 photolysis

54

Figure 3.5 Solution pH at which the UV/chlorine and the UV/H2O2 AOPs are equally

efficient as a function of TOC concentration

3.4 Conclusions TCE can be eliminated by direct UV photolysis; however, the decay rate increases when

using UV as part of an advanced oxidation process, since TCE is rapidly oxidized by •OH.

According to our experiments, the UV/chlorine AOP is more efficient than the UV/H2O2 AOP

at pH 5 in the model systems explored in this study (relatively free of other •OH scavengers),

but it loses efficiency with increasing pH. The numerical modeling analysis is generally

consistent with the experimental results in terms of the TCE decay rates by UV alone and the

UV/chlorine and the UV/H2O2 AOPs. Interestingly, in water containing more •OH scavengers

(i.e. more ‘realistic’ waters), the model predicts that the UV/chlorine AOP will become more

competitive relative to the UV/H2O2 AOP, with all other factors being equal, as the scavenger

concentration increases. However, this is admittedly a simplification. Factors such as the

reaction of chlorine with organic matter (with hydrogen peroxide typically being much less

reactive) would complicate such predictions. More research is required.

The authors also wish to emphasize that very little work has been done to assess potential

chlorination by-product formation in the UV/chlorine AOP. Some limited research suggests that

approximately 17–30% (by mass) of chlorine that is photolyzed by UV might yield chlorate

(Buxton and Subhani, 1972b; Feng et al., 2010). The formation of organochlorine species under

5.0

5.5

6.0

6.5

7.0

7.5

0 1 2 3 4 5

TOC concentration (mg L–1)

Equa

lly e

ffici

ent p

H s

55

the unique conditions of the UV/chlorine AOP system (i.e., high chlorine dose but very low

contact time) is also largely unexplored. Research on this issue is recommended.

Acknowledgements This work was partially funded by the NSERC Engage Grant program, and by Stantec

Consulting Ltd. The technical and logistical assistance of Leigh McDermott of Stantec

Consulting Ltd., and Tim Walton of the Region of Waterloo, is also gratefully acknowledged.

References APHA, AWWA, WEF, 2005. Standard Methods for the Examination of Water & Wastewater,

21st edition. 4500-Cl G. DPD colorimetric method, 4-67. American Public Health Association,

Washington, D.C., USA.

Baxendale, J.H., Wilson, J.A., 1957. The photolysis of hydrogen peroxide at high light

intensities. Transactions of the Faraday Society, 53, 344–356.

Bolton, J.R., Linden, K.G., 2003. Standardization of methods for fluence (UV dose)

determination in bench-scale UV experiments. Journal of Environmental Engineering, 129(3),

209–215.

Bolton, J.R., Stefan, M.I., 2002. Fundamental photochemical approach to the concepts of

fluence (UV dose) and electrical energy efficiency in photochemical degradation reactions.

Research on Chemical Intermediates, 28(7–9), 857–870.

Bolton, J.R., Stefan, M.I., Shaw, P.-S., Lykke, K.R., 2009. Determination of the quantum yield

of the ferrioxalate and KI/KIO3 actinometers and a method for the calibration of radiometer

detectors. CDROM Proceedings 5th UV World Congress, Amsterdam, The Netherlands.




947–957.




958–969.





56




970–977.



Crowell, R.A, Lian, R., Sauer Jr., M.C., Oulianov, D.A., Shkrob, I.A., 2004. Geminate

recombination of hydroxyl radicals generated in 200 nm photodisassociation of aqueous

hydrogen peroxide. Chemical Physics Letters, 383(5–6), 481–485.

Czapski, G., Bielski, B.H.J., 1963. The formation and decay of H2O3 and HO2 in electron-

irradiated aqueous solutions. The Journal of Physical Chemistry, 67(10), 2180–2184.



Research, 42(1–2), 13–51.



6(3), 277–284.

Feng, Y., Smith, D.W., Bolton, J.R., 2010. A potential new method for determination of the

fluence (UV dose) delivered in UV reactors involving the photodegradation of free chlorine.

Water Environment Research, 82(4), 328–334.

Fogelman, K.D., Walker, D.M., Margerum, D.W., 1989. Non-metal redox kinetics: hypochlorite

and hypochlorous acid reactions with sulfite. Inorganic Chemistry, 28(6), 986–993.

Goldstein, S., Aschengrau, D., Diamant, Y., Rabani, J., 2007. Photolysis of aqueous H2O2:

quantum yield and applications for polychromatic UV actinometry in photoreactions.


Goldstein, S., Rabani, J., 2008. The ferrioxalate and iodide-iodate actinometers in the UV region.

Journal of Photochemistry and Photobiology A: Chemistry, 193(1), 50–55.



816–832.



57

Hunt, J.P., Taube, H., 1952. The photochemical decomposition of hydrogen peroxide. Quantum

yields, tracer and fractionation effects. Journal of the American Chemical Society, 74(23),

5999–6002.



Kläning, U.K., Wolff, T., 1985. Laser flash photolysis of HClO, ClO–, HBrO, and BrO– in

aqueous solution. Reaction of Cl- and Br-atoms. Berichte der Bunsen-Gesellschaft für

Physikalische Chemie, 89(3), 243–245.

Klassen, N.V., Marchington, D., McGowan H.C.E., 1994. H2O2 determination by the I3– method

and by KMnO4 titration. Analytical Chemistry, 66(18), 2921–2925.

Li, K., Stefan, M.I., Crittenden, J.C., 2004. UV photolysis of trichloroethylene: product study

and kinetic modelling. Environmental Science & Technology, 38(24), 6685–6693.

Li, K., Stefan M.I., Crittenden, J.C., 2007. Trichloroethylene degradation by UV/H2O2 advanced

oxidation process: product study and kinetic modelling. Environmental Science &

Technology, 41(5), 1696–1703.

Luňák, S., Sedlák, P., 1992. Photoinitiated reactions of hydrogen peroxide in the liquid phase.

Journal of Photochemistry and Photobiology A: Chemistry, 68(1), 1–33.





Oliver, B.G., Carey, J.H., 1977. Photochemical production of chlorinated organics in aqueous


Parsons, S., 2004. Advanced Oxidation Processes for Water and Wastewater Treatment, 1–6.

IWA Publishing, London, UK.

Peng, J., Wan, A., 1997. Measurement of Henry’s constants of high-volatility organic

compounds using a headspace autosampler. Environmental Science & Technology, 31(10),

2998–3003.

Rosenfeldt, E.J., Linden, K.G., 2004. Degradation of endocrine disrupting chemicals bisphenol

A, ethinyl estradiol, and estradiol during UV photolysis and advanced oxidation processes.


http://www.sciencedirect.com/science/journal/10106030

58

Rosenfeldt, E.J., Melcher, B., Linden, K.G., 2005. UV and UV/H2O2 treatment of

methylisoborneol (MIB) and geosmin in water. Journal of Water Supply: Research and

Technology–AQUA, 54(7), 423–434.

Schwarzenbach, R.P., Gschwend, P.M., Imboden, D. M., 2003. Environmental Organic

Chemistry, 2nd edition, 611–686. John Wiley & Sons, Inc., Hoboken, New Jersey, USA.

Sharpless, C.M., Linden, K.G., 2003. Experimental and model comparisons of low- and

medium-pressure Hg lamps for the direct and H2O2 assisted UV photodegradation of N-

nitrosodimethylamine in simulated drinking water. Environmental Science & Technology,

37(9), 1933–1940.

Song, R., 1996. Ozone-bromide-NOM interactions in water treatment, Ph.D. thesis. University

of Illinois at Urbana-Champaign, IL, USA.

Stefan, M.I., Bolton J.R., 2005. Fundamental approach to the fluence-based kinetic and

electrical energy efficiency parameters in photochemical degradation reactions:

polychromatic light. Journal of Environmental Engineering and Science, 4(S1), S13–S18.

Stefan, M.I., Hoy, A.R., Bolton, J.R., 1996. Kinetics and mechanism of the degradation and

mineralization of acetone in dilute aqueous solution sensitized by the UV photolysis of

hydrogen peroxide. Environmental Science & Technology, 30(7), 2382–2390.

Suty, H., De Traversay, C., Cost, M., 2004. Applications of advanced oxidation processes:

present and future. Water Science and Technology, 49(4), 227–233.

Tinsley, I.J., 2004. Chemical Concepts in Pollutant Behavior, 2nd edition, 123–148. John Wiley

& Sons, Inc. Hoboken, New Jersey, USA.

United States Environmental Protection Agency (USEPA), 2008. Method 551.1. Determination

of chlorination disinfection byproducts, chlorinated solvents, and halogenated

pesticides/herbicides in drinking water by liquid-liquid extraction and gas chromatography

with electron-capture detection, Revision, 1.0. Available from

<http://www.epa.gov/sam/pdfs/EPA-551.1.pdf> (accessed 11.09.11).

Volman, D.H., Chen, J.C., 1959. The photochemical decomposition of hydrogen peroxide in

aqueous solutions of allyl alcohol at 2537 Å. Journal of the American Chemical Society,

81(16), 4141–4144.

http://www.epa.gov/sam/pdfs/EPA-551.1.pdf

59

Wang, D., Walton, T., McDermott, L., Hofmann, R., 2011. The effectiveness of UV+chlorine

treatment of TCE in drinking water. Proceedings of the 2011 IOA and IUVA World Congress

& Exhibition, Paris, France.





56(8), 469–477.

Westerhoff, P., Aiken, G., Amy, G, Debroux, J., 1999. Relationships between the structure of

natural organic matter and its reactivity towards molecular ozone and hydroxyl radicals.

Water Research, 33(10), 2265–2276.

60

4. FULL-SCALE COMPARISON OF

ULTRAVIOLET/CHLORINE ADVANCED OXIDATION TO

ULTRAVIOLET/HYDROGEN PEROXIDE FOR TASTE AND

ODOUR CONTROL IN DRINKING WATER

TREATMENT

This chapter has been submitted for publication as follows:

Wang, D., Bolton J.R., Andrews, S.A., Hofmann, R. UV/chlorine control of drinking water taste and odour at

pilot and full-scale and the potential for caffeine as an experimental surrogate. Water Research.

Abstract Advanced oxidation processes (AOPs) can be used to destroy taste and odour-causing

compounds in drinking water. This work investigated pilot- and full-scale performance of the

ultraviolet/chlorine AOP for the destruction of geosmin, 2-methylisoborneol (MIB) and caffeine

(as a surrogate) in two different surface waters. The efficiency of UV/chlorine at pH 7.5 and 8.5

was comparable to that of UV/hydrogen peroxide (UV/H2O2) under parallel conditions, and was

superior at pH 6.5. Caffeine was found to be a suitable surrogate for geosmin and MIB, and

could be used in future research as a more economical alternative to geosmin or MIB spiking.

4.1 Introduction The previous chapter discussed UV/chlorine efficiency for TCE decay using a bench-scale

UV collimated beam apparatus. It was found that UV/chlorine is a promising process that could

be an alternative to UV/H2O2. A mathematical reaction kinetic model demonstrated similar

promising results for UV/chlorine. To ultimately implement UV/chlorine at full-scale, however,

a greater degree of confidence would be required. To this end, the performance of UV/chlorine

was assessed using pilot-scale and full-scale testing.

Geosmin and 2-methylisoborneol (MIB) can be destroyed during drinking water treatment

by advanced oxidation processes (AOPs), which generate hydroxyl radicals (•OH). The

reactions of geosmin and MIB with •OH are rapid, with rate constants of 7.8 × 109 and 5.1 × 109

M–1 s–1, respectively (Peter and von Gunten, 2007; Rosenfeldt et al., 2005). While ozone-based

61

AOPs tend to be the most common in the drinking water industry, UV-based AOPs—usually

UV/H2O2—are becoming more popular. UV/H2O2 is used at the City of Cornwall Water

Purification Plant (Ontario, Canada) for the control of seasonal taste and odour events that

usually occur in late summer (McCormick et al., 2013). UV/H2O2, while effective, can create

operational problems. Most of the applied H2O2 survives UV exposure, and the residual

therefore needs to be quenched because it will otherwise create a strong chlorine demand that

would interfere with secondary disinfection. At Cornwall, the residual H2O2 is quenched with a

stoichiometric excess of chlorine, but operationally this is challenging because of an inability to

measure or accurately predict the H2O2 residual, issues with different residual H2O2 in the

effluent from multiple UV reactors, H2O2 handling challenges, and other factors.

The UV/chlorine AOP may be an alternative to UV/H2O2 for water and wastewater

treatment. It has been investigated to treat contaminants such as para-chlorobenzoic acid,

benzoic acid, nitrobenzene, phenol, maleic acid, and trichloroethylene in lab prepared water

(Watts et al., 2007; Jin et al., 2011; Fang et al., 2014; Zhao et al., 2011; Wang et al., 2012),

emerging contaminants and taste and odour compounds in drinking water (Zhang et al., 2014;

Sichel et al., 2011; Watts et al., 2012), and naphthenic acids and fluorophore organic

compounds in oil sands wastewater (Chan et al., 2012; Shu et al., 2014). For a drinking water

treatment plant, operations could be relatively simple and cost-effective compared to UV/H2O2,

especially if the chlorine dose is selected to provide adequate photolysis for the control of taste

and odour compounds, with remaining chlorine serving as a secondary disinfectant (Watts et al.,

2012). Research on UV/chlorine, however, is still largely at the theoretical level or laboratory-

scale. In this study, the effectiveness of a medium pressure (MP) UV/chlorine AOP for taste and

odour control was investigated at the Cornwall Water Purification Plant through full-scale trials,

with a comparison of UV/chlorine to UV/H2O2 under parallel conditions. Caffeine as a potential

surrogate for geosmin and MIB was also evaluated. A pilot-scale study using caffeine

destruction in a Rayox® batch reactor was conducted to further investigate UV/chlorine

efficiency in water from a second source.

62

4.2 Material and Methods

4.2.1 Reagents and Materials

Geosmin and MIB (>95% purity) from Dalton Chemical Laboratories were dissolved

together in Milli-Q® water to make a stock solution at a concentration of approximately 80 mg

L–1 for full-scale tests. Caffeine stock solutions at concentrations of 10 g L–1 and 1 g L–1 for the

full- and pilot-scale tests, respectively, were prepared from purchased caffeine (ReagentPlus®

grade, Sigma-Aldrich). Deuterated d3-geosmin (99 atom % D, Sigma-Aldrich) and d3-caffeine

(99 atom % D, CDN Isotopes) were used as internal standards for quantification of

geosmin/MIB and caffeine, respectively. Sodium hypochlorite solution (NaOCl) (10–15 wt. %,

reagent grade, Sigma-Aldrich) and H2O2 solution (50 wt. %, Sigma-Aldrich) were used in the

pilot-scale tests. Industrial grade NaOCl solution (12.5 wt. %, NSF 60 certified, Olin Chlor

Alkali) and H2O2 solution (35 wt. %, NSF 60 certified, Arkema Inc.) were used in the Cornwall

full-scale tests. Sulphuric acid (H2SO4, 95–98%, A.C.S. grade, Sigma-Aldrich) and sodium

hydroxide (NaOH, ≥97.0%, A.C.S. grade, Sigma-Aldrich) were freshly diluted to appropriate

concentrations for pH adjustment in both full- and pilot-scale tests. Other compounds used in

experiments and sample analyses were all analytical reagent grade or higher. Milli-Q® water

was used in all experiments and analytical determinations.

4.2.2 Experimental Facilities and Procedures

Cornwall Full-Scale Tests

Full-scale experiments were carried out at the Cornwall Water Purification Plant in early

summer (May) and late summer (September, when taste and odour events typically occur), and

are referred to as the 1st and 2nd full-scale tests in the following text. One of the MP UV reactors

(Model: UVSwift 8L24, TrojanUV) was isolated for the experiments (shown in Figure 4.1). UV

power was set at maximum output to perform the UV/chlorine and UV/H2O2 AOPs, which

delivered a UV dose of 2,000 ± 150 mJ cm–2 from 200 to 400 nm for an exposure of 7.2 seconds

at a water flowrate of 100 L s–1, as estimated based on the free chlorine photolysis (Wang et al.,

2012). The water was drawn from the St. Lawrence River and had been treated by conventional

treatment (prechlorination, alum coagulation, flocculation, settling, and conventional

sand/anthracite filtration). Water quality parameters are summarized in Table 4.1.

63

Figure 4.1 Full-scale Trojan UVSwift reactor (left) and pilot-scale Rayox® reactor (right)

Table 4.1 Post-filtration water quality parameters for full- and pilot-scale tests

Test Source pH Turbidity

(NTU)

Alkalinity (mg CaCO3

L–1)

Total organic carbon (TOC)

(mg C L–1)

Nitrate (NO3

–) (mg L–1)

Spiked target chemicals

Cornwall 1st full-scale

St. Lawrence

River 7.9 0.02 92 1.5 1.2

400 ng L–1 geosmin 400 ng L–1 MIB

20 μg L–1 caffeine

Cornwall 2nd full-

scale

St. Lawrence

River 8.1 0.03 88 1.8 – Ambient geosmin

and MIB

Rayox® Pilot-scale

Lake Simcoe 7.5 0.2 123 3.5 0.67 20 μg L–1 caffeine

In the 1st full-scale test, approximately 400 ng L–1 geosmin and MIB, and 20 μg L–1

caffeine were spiked into the water flow upstream of the UV reactor, along with a chlorine dose

of 2, 6, or 10 mg L–1 as free chlorine, or an H2O2 dose of 1.0, 2.9, or 4.8 mg L–1 (equimolar

concentrations as chlorine) and pH adjustment to 6.5, 7.5, or 8.5. Additional trials in the absence

of chlorine and H2O2 were carried out to evaluate the stability of geosmin, MIB, and caffeine to

UV exposure alone. Preliminary tests (not shown) indicated that caffeine, geosmin and MIB

64

were all stable in the presence of 40 mg L–1 free chlorine or 10 mg L–1 H2O2 for at least 20

minutes, with the decay less than 0.5%.

The 2nd full-scale test of UV/chlorine at Cornwall was more limited than the first, with no

geosmin, MIB, or caffeine spiked, with the performance of UV/chlorine (only) monitored for

controlling the existing 18 ng L–1 of geosmin in the incoming water (MIB was below detection

limits). The main purpose of this second test was to monitor by-product formation (reported

elsewhere), but it was also used as another opportunity to validate the UV/chlorine performance

under more limited conditions. Chlorine doses of 2, 6, or 10 mg L–1 at pH 6.5, 7.5, and 8.5 were

applied.

Rayox® Pilot-Scale Test

A 40 L Rayox® completely-mixed batch reactor (Model: PS1-1-120, Calgon Carbon

Corporation), shown in Figure 4.1, equipped with a 1 kW MP UV lamp (Heraeus Noblelight

GmbH, Germany) was used in the pilot-scale experiments to evaluate UV/chlorine efficiency,

using spiked caffeine as a performance indicator. Water was collected post-filtration from the

Keswick Water Treatment Plant (Ontario, Canada), which draws water from Lake Simcoe.

Treatment at the Keswick plant is similar to that at Cornwall, except that no prechlorination is

used. The water contained approximately twice the total organic carbon (TOC) as the St.

Lawrence River water (Table 4.1). The experimental conditions in the Rayox® reactor were

similar to those applied in the full-scale tests. Caffeine at 20 μg L–1 was spiked into 40 L water

and treated by UV alone, UV/chlorine (2, 6, or 10 mg L–1) or UV/H2O2 (1.0, 2.9, or 4.8 mg L–1)

at pH 6.5, 7.5, or 8.5. The UV exposure time in the Rayox® reactor was 40 s, which was

predicted to deliver a UV dose (200–400 nm) of 1,820 ± 110 mJ cm– 2, using UVCalc®

software version 2B (from Bolton Photosciences Inc.). The UV dose varied slightly with

different chlorine/H2O2 doses and pH values.

4.2.3 Sample Analysis

Free chlorine was determined using a HACH® spectrophotometer (Model: DR/2500,

HACH), according to the Standard Methods 4500-Cl G (APHA et al., 2005). A Cecil UV/vis

spectrophotometer (Model: CE3055, Cecil Instruments) was used to determine H2O2

concentrations based on the triiodide method described by Klassen et al. (1994). An Agilent

8453 UV/vis photodiode array spectrophotometer (Model: G1103A, Angilent Technologies)

65

was used to determine solution absorbances in the entire UV range. Geosmin and MIB samples

spiked with 100 ng L–1 d3-geosmin were extracted and concentrated using the headspace solid

phase micro-extraction (HS-SPME) method and quantified using a Varian 3800 gas

chromatography coupled with a Varian 4000 ion-trap mass spectrometry (GC-MS) in the

electron ionization (EI) mode, according to Standard Methods 6040D (APHA et al., 2012).

Caffeine samples with 1 μg L–1 d3-caffeine added were extracted and concentrated using solid-

phase extraction (SPE) cartridges and analyzed using the same GC-MS, but in the positive ion

chemical ionization (CI) mode, based on the method described by Verenitch et al. (2006).

Method detection limits (MDLs) for geosmin, MIB, and caffeine were 2, 9, and 31 ng L–1,

respectively. Details for the analytical methods are shown in Appendix G.

4.3 Results and Discussion

4.3.2 Free Chlorine Decay

One of the potential advantages of UV/chlorine over UV/H2O2 is the predicted greater

photolysis of chlorine across the UV reactor than H2O2, minimizing the need to quench excess

oxidant. For both the full-scale and pilot-scale tests, the chlorine concentrations decreased by

approximately 40–80% across the UV reactors at doses (200 nm to 400 nm) of 1,800–2,000 mJ

cm–2 (Figure 4.2). Raw data are summarized in Appendix I. H2O2 decay in parallel tests was at

most approximately 5% (data not shown). Chlorine was observed to undergo greater photolysis

at a higher pH, with about 1.5–2 times the total decay at pH 8.5 compared to pH 6.5. This can be

explained by the higher OCl– concentration relative to HOCl at the higher pH (pKa of HOCl =

7.54 at 25 °C). OCl– absorbs MP UV light about 4.5 times more than HOCl (Wang et al., 2012),

and photolyzes at approximately the same rate as HOCl (quantum yield of 0.9, compared to 1.0

for HOCl photolysis) (Feng et al., 2007). It was also observed that chlorine photodecomposition

was faster at a lower initial dose, especially at pH 8.5. It is assumed that this is due to a higher

average UV dose at a lower chlorine concentration, because the UV dose is a function of the

chlorine concentration (chlorine blocks transmission of the UV light into the water). For

example, using UVCalc® 2B software, the UV dose for 40 s exposure in the Rayox® reactor at

an initial chlorine dose of 2 mg L–1 at pH 8.5 was predicted to be approximately 13% higher

than that at a dose of 10 mg L–1. It should be noted that the UV reactors used at the Cornwall

66

plant were an early version of an advanced oxidation system, and are likely undersized relative

to a modern installation. Similarly, the UV dose applied in the pilot-scale reactor was also

relatively “low”, to remain consistent with the Cornwall tests. As such, it is likely that a much

greater amount of chlorine photolysis would occur in a current UV-AOP reactor. The authors

have observed greater than 90% chlorine photolysis across a UV reactor used for

trichloroethylene destruction in a groundwater treatment system (Wang et al., 2011).

Figure 4.2 Percentage of free chlorine photolysis by UV exposure. Error bars represent the

values of experimental duplicates.

4.3.3 Geosmin and MIB Decay

The destruction of geosmin and MIB due to UV photolysis alone, UV/chlorine, or

UV/H2O2 for the 1st full-scale test is shown in Figure 4.3. Approximately 20% and 10% of

spiked geosmin and MIB, respectively, were destroyed by UV alone (~2,000 mJ cm–2). This is

generally consistent with Rosenfeldt et al. (2005), who reported that MP UV exposure at 2,000

mJ cm–2 led to about 35–40% destruction of geosmin and MIB. When an AOP was applied,

geosmin and MIB destruction was increased due to the additional •OH oxidation. UV/chlorine

and UV/H2O2 led to similar amounts (less than 10% in most cases) of geosmin and MIB

destruction at pH 7.5 and 8.5. At pH 6.5, however, UV/chlorine was substantially superior to

UV/H2O2, resulting in 10–25% more destruction for all applied doses.

67

Figure 4.3 Geosmin (top plot) and MIB (bottom plot) decay in the 1st full-scale test.

Error bars represent the values of experimental duplicates.

The superior UV/chlorine performance at lower pH has been reported by Watts and Linden

(2007), Watts et al. (2007), and Wang et al. (2012), and it is probably because of the conversion

of OCl– to HOCl at the lower pH. HOCl absorbs MP UV light about 2.3 times more efficiently

than H2O2 (Wang et al. 2012) and produces •OH at a similar efficiency (quantum yield of •OH

formation for HOCl: 0.85 vs. 1.11 for H2O2) (Nowell and Hoigné, J., 1992; Goldstein et al.,

2007), but also reacts with •OH (i.e. scavenges) more slowly than H2O2 (rate constant with •OH

for HOCl: 8.46 × 104 M–1 s–1 vs. 2.7 × 107 M–1 s–1 for H2O2) (Watts and Linden, 2007; Stefan et

al., 1996). This all leads to a higher efficiency of UV/HOCl than that of UV/H2O2. In contrast,

68

the much higher reaction (scavenging) rate of OCl– with •OH than either HOCl or H2O2 (rate

constant: 9.0 × 109 M–1 s–1) (Buxton and Subhani, 1972) more than offsets the benefit of OCl–′s

stronger UV absorption. In previous studies it has been proposed that UV/chlorine would be

superior to UV/H2O2 at pH <6 in pure water (i.e. laboratory grade water, containing no •OH

scavengers). It was also proposed that the pH at which UV/chlorine remained competitive

relative to UV/H2O2 would increase with increasing •OH scavenger concentration. The

Cornwall water had an •OH scavenging potential 2–17 times higher than the three previous

studies as estimated using the total organic carbon (TOC) and alkalinity concentrations reported

in Table 4.1. As such, this result tends to corroborate the theory that UV/chlorine remains

competitive with UV/H2O2 at higher pH—up to pH 8.5 in this case—with sufficient •OH

scavengers present.

Similar to UV/chlorine, a pH effect on UV/H2O2 efficiency was also observed, shown in

Figure 4.3. However, unlike UV/chlorine, whose efficiency is largely dependent on the

differential •OH scavenging of HOCl and OCl–, UV/H2O2 efficiency is probably changed by the

bicarbonate/carbonate equilibrium. With the increase of pH from 6.5 to 8.5, carbonic acid

(H2CO3) and bicarbonate (HCO3–) present in the water convert to carbonate (CO3

2–) (pKa of

H2CO3 = 6.37, pKa of HCO3– = 10.36) (Fanghänel et al, 1996; Kolthoff and Bosch, 1928). CO3

2–

is a much stronger •OH scavenger than H2CO3 and HCO3–. The rate constant of •OH with CO3

2–

is 3.9 × 108 M–1 s–1, compared to 8.5 × 106 M–1 s–1 for HCO3– and negligible for H2CO3 (Buxton

et al., 1988; Liao et al., 2001). In the presence of an alkalinity of 92 mg CaCO3 L–1 in the

Cornwall water, the H2CO3/HCO3–/CO3

2– species was calculated to increase the •OH

scavenging potential of the H2O2 solution by approximately 12%. However, this is still less than

the 40% decrease of the rates of geosmin and MIB decay due to UV/H2O2 when pH increased

from 6.5 to 8.5, as shown in Figure 4.3. This implies that there might be other •OH scavengers

present in the water that could also increase the •OH scavenging potential of the solution at a

higher pH. In addition, the effect of carbonate species was also present in the UV/chlorine

treated water. However, since OCl– is a much stronger •OH scavenger than carbonate species,

the contribution of carbonate species to the increase of the chlorine solution scavenging

potential was estimated to be minimal, compared to OCl– (theoretically less than 1%).

Since the geosmin and MIB decay in the UV/chlorine and UV/H2O2 processes consisted of

the direct photolysis by UV exposure and the oxidation by •OH, the difference in their

http://link.springer.com/search?facet-author=%22Th.+Fangh%C3%A4nel%22

69

concentrations after treatment of UV and an AOP reflects the net destruction by •OH only. In

theory, the rate of decay of a compound reacting with •OH is a function of the second-order rate

constant with •OH and the concentration of •OH. The ratio of the rates of geosmin and MIB

decay by •OH in the same solution is thus directly proportional to the respective reaction rate

constants with •OH. As shown in Figure 4.3, the rate of MIB decay by •OH was observed to be

approximately 50–90% of that for geosmin under parallel conditions. This is generally

consistent with the difference in their rate constants with •OH—the rate constant for MIB is

approximately 65% of that for geosmin (Peter and von Gunten, 2007).

In the 2nd full-scale test, only UV/chlorine performance was evaluated, and only in terms of

the destruction of the 18 ng L–1 geosmin already present in the water. The trend in geosmin

destruction (Figure 4.4) was similar to the first testing campaign. UV/chlorine was observed to

be more efficient at lower pH, but substantial geosmin destruction was still accomplished at pH

7.5 and 8.5. The natural pH at Cornwall is approximately 8, and the data suggest that chlorine

doses in the order of 6 mg L-1 could reduce the incoming geosmin of 18 ng L–1 to below the

common detection threshold of 7 to 10 ng/L at that pH.

Figure 4.4 Geosmin decay in the 2nd full-scale test. Error bars represent the values of

experimental duplicates.

70

4.3.4 Caffeine Decay

The feasibility of using caffeine as a surrogate for geosmin and MIB to estimate the

efficiency of UV/chlorine for taste and odour control in future research was investigated in the

1st full-scale test. Spiking caffeine simultaneously with geosmin and MIB also gave more

confidence in the performance of UV/chlorine treatment. Caffeine was subsequently used in the

pilot-scale test to further evaluate the UV/chlorine efficiency in a second water matrix (Lake

Simcoe) that contained approximately twice the TOC as at Cornwall, and therefore a presumed

higher scavenging potential. The full-scale results (top plot of Figure 4.5) show that caffeine is

similarly photosensitive as geosmin and MIB, with a UV dose (alone) of ~2,000 mJ cm–2

leading to a 10–15% decrease, which was approximately the same as the destruction of geosmin

(20%) and MIB (10%).

Past studies have reported the rapid reaction between caffeine and •OH, with an average

rate constant of 5.0 × 109 M–1 s–1 (Kesavan and Powers, 1985; Shi et al., 1991; Devasagayam et

al., 1996; Brezová et al., 2009; and León-Carmona and Galano, 2011). The rate constant is in

the same order of magnitude as those for geosmin (7.8 × 109 M–1 s–1), and is almost identical to

MIB (5.1 × 109 M–1 s–1). This means in theory that the rate of caffeine decay by •OH should be

approximately the same as that of MIB, but 36% less than that of geosmin. The results shown in

Figure 4.5 generally reflected the theory, which illustrated that the experimental rate of caffeine

decay by •OH was approximately 90% and 70% (on average) those observed for MIB and

geosmin, respectively. After including UV photolysis, caffeine destruction by UV/chlorine and

UV/H2O2 was on average equivalent to 95% and 67% of MIB and geosmin decay, respectively.

Caffeine, therefore, is an almost perfect surrogate to estimate the MIB destruction by

UV/chlorine or UV/H2O2 treatment, and that geosmin decay can be expected to be

approximately 35% greater. Based on the caffeine destruction in the Lake Simcoe water in the

pilot-scale test (bottom plot of Figure 4.5), which was found to be similar to that in the Cornwall

water, UV/chlorine was predicted to have similar performance for geosmin and MIB destruction

in both waters.

71

Figure 4.5 Caffeine decay in the 1st full-scale (top plot) and pilot-scale (bottom plot) tests.

Error bars represent the values of experimental duplicates.

4.3.5 Electrical Energy per Order (EEO)

Electrical energy per order (EEO) (kWh m–3 order–1) is the electrical energy in kilowatt

hours (kWh) required to degrade a target contaminant by one order of magnitude (90%) in 1 m3

of water (Bolton et al., 2001). The equations for full-scale and pilot-scale EEO are given in

Equations [4.1] and [4.2]:

72

EO lg( / )i f

PEF C C

= for full-scale reactor [4.1]

EO1000lg( / )i f

PtEV C C

= for pilot-scale Rayox® batch reactor [4.2]

where P is the electrical power input (kW) into the UV reactor, F is the flowrate (m3 h–1), Ci and

Cf are the initial and final concentrations (M) of the target contaminant before and after the

AOP, t is the UV exposure time (h) in the batch reactor, and V is the volume of treated water (L)

in the batch reactor. In this study, P was 83.5 kW for the full-scale test and 1.8 kW for the pilot-

scale test; F was 360 m3 h–1 converted from 100 L s–1; t was 0.011 h converted from 40 s; and V

was 40 L. The calculated EEO values for geosmin, MIB, and caffeine are given in Table 4.2.

Table 4.2 Full- and pilot-scale EEO values (kWh m–3 order–1) for geosmin, MIB, and

caffeine removal

Treatment pH 6.5 pH 7.5 pH 8.5

Geosmin MIB Caffeine Geosmin MIB Caffeine Geosmin MIB Caffeine

1st full-scale test

UV/chlorine 2 mg L–1 0.26 0.35 0.43 0.49 0.68 0.65 0.65 1.0 0.87

UV/chlorine 6 mg L–1 0.22 0.29 0.26 0.39 0.51 0.47 0.59 0.81 0.67

UV/chlorine 10 mg L–1 0.16 0.22 0.21 0.28 0.34 0.42 0.40 0.58 0.54

UV/H2O2 1.0 mg L–1 0.47 0.66 0.94 0.60 0.90 1.2 1.0 2.2 1.9

UV/H2O2 2.9 mg L–1 0.30 0.46 0.49 0.37 0.53 0.69 0.51 0.73 0.87

UV/H2O2 4.8 mg L–1 0.23 0.31 0.36 0.29 0.34 0.41 0.36 0.52 0.52

Pilot-scale test

UV/chlorine 2 mg L–1 0.98 1.5 2.1

UV/chlorine 6 mg L–1 0.53 1.0 1.4

UV/chlorine 10 mg L–1 0.39 0.88 1.1

UV/H2O2 1.0 mg L–1 2.1 2.7 3.1

UV/H2O2 2.9 mg L–1 1.2 1.3 1.7

UV/H2O2 4.8 mg L–1 0.80 0.91 1.1

EEO reflects the relative efficiency of an AOP. Comparing the EEO values for each

compound in the full-scale test summarized in Table 4.2, it was found that EEO values for

73

UV/chlorine were lower than those for UV/H2O2 in most cases, which demonstrates the superior

efficiency of UV/chlorine. The only case where UV/chlorine was less efficient (higher EEO) than

UV/ H2O2 occurred at the highest pH (8.5) with the highest oxidant dose (chlorine: 10 mg L–1

and H2O2: 4.8 mg L–1). Similarly, in the pilot-scale test using a different water matrix, the EEO

values for UV/chlorine were all lower than or the same as those for UV/H2O2. This tends to

verify the theory proposed by Wang et al. (2012) that UV/chlorine becomes more competitive

relative to UV/H2O2 with more •OH scavengers, because the levels of two major scavengers,

TOC and alkalinity, in the Lake Simcoe water were approximately 2 and 1.3 times higher than

those of the Cornwall water.

4.3.6 Comment on Chlorine Radical (•Cl) and Disinfection By-Product (DBP) Formation

during Chlorine Photolysis

Chlorine radicals (•Cl) are known to be generated during chlorine photolysis (Oliver and

Carey, 1977; Nowell and Hoigné, J., 1992), however, their role in the destruction of a target

compound is uncertain. Nowell and Hoigné (1992) and Watts and Linden (2007) considered •Cl

effectiveness to be negligible compared to •OH for the destruction of 1-chlorobutane and para-

chlorobenzoic acid, while Fang et al. (2014) reported that •Cl contributed to the majority of

benzoic acid decay during UV/chlorine treatment. This may be due to a reported high selectivity

of •Cl to different compounds. For example, •Cl reacts with benzoic acid faster than •OH (rate

constant with •Cl: 1.8 × 1010 M–1 s–1 vs. 5.9 × 109 M–1 s–1 with •OH), while it reacts with 1-

chlorobutane more slowly than •OH (rate constant with •Cl: 1.3 × 108 M–1 s–1 vs. 3.0 × 109 M–1

s–1 with •OH) and the reaction with nitrobenzene is negligible (Fang et al., 2014, Nowell and

Hoigné, J., 1992; Bell et al., 1981). The competition between •Cl and •OH dictates the impact of

•Cl in the destruction of the compound. In this study, the role of •Cl in geosmin, MIB, and

caffeine destruction is unknown due to the lack of information in the literature for the reaction

rates between •Cl and these compounds.

The concentration of chlorine used in practice for the UV/chlorine AOP may be in the

order of 5–10 mg L–1, compared to the more traditional 0.2–2 mg L–1 used for chlorine

disinfection. This might cause an increased formation of chlorination (disinfection) by-products.

Tending to mitigate this effect, however, is the very short (i.e. seconds) chlorine contact time

during the AOP process during which much of the higher chlorine dose undergoes photolysis.

74

Nevertheless, the formation of by-products during UV/chlorine treatment is an issue that

warrants careful study, but is beyond the scope of this paper.

4.4 Conclusions The UV/chlorine AOP was found to be equally or more efficient than the UV/H2O2 AOP in

the destruction of geosmin, MIB, and caffeine for similar oxidant (molar) doses, except at the

highest pH (8.5) and the highest oxidant dose in the Cornwall water. Caffeine was found to be

an appropriate surrogate for geosmin and MIB. These results are likely to be water-specific, but

were generally consistent for the two waters tested in this study, containing different amounts of

TOC and carbonate alkalinity (TOC: 1.5 vs. 3.5 mg C L–1, alkalinity: 92 vs. 123 mg CaCO3

L–1). Given the presumed operational advantages of UV/chlorine over UV/H2O2, this suggests

merit in further investigating this treatment technology. One key issue to be explored before

full-scale implementation is the potential for by-product formation.

Acknowledgements This work was funded by the Natural Sciences and Engineering Research Council of

Canada through the Industrial Research Chair program. The authors express their appreciation

to the staff at the Cornwall Water Purification Plant, including Owen O'Keefe, Daniel Drouin,

and Morris McCormick, for their help with the full-scale experiments. The help of Zhen (Jim)

Wang, Jiafan (Kevin) Yang, Hong Zhang and A.H.M. Anwar Sadmani is also gratefully

acknowledged.


21st edition. American Public Health Association, American Water Works Association, and

Water Environment Federation, Washington, D.C., USA.

APHA, AWWA, WEF, 2012. Standard Methods for the Examination of Water & Wastewater,

22nd edition. American Public Health Association, American Water Works Association, and


75

Bell, T.N., Perkins, K.A., Perkins, P.G., 1981. Chlorination of paraffin hydrocarbons. 3.

chlorination of n-butane, 1-chlorobutane, and 2-chlorobutane. The Journal of Physical

Chemistry, 82(2), 160–165.

Bolton, J.R., Bircher, K.G., Tumas, W., Tolman, C.A., 2001. Figures-of-merit for the technical

development and application of advanced oxidation technologies for both electric- and solar-

driven systems. Pure and Applied Chemistry, 73(4), 627–637.

Brezová, V., Šlebodová, A., Staško, A., 2009. Coffee as a source of antioxidants: an EPR study.

Food Chemistry, 114(3), 859–868.

Buxton, G.V., Greenstock, C.L., Helman, W.P., Ross, A.B., 1988. Critical review of rate

constants for reactions of hydrated electrons, hydrogen atoms and hydroxyl radicals (•OH/•O–)

in aqueous solution. Journal of Physical and Chemical Reference Data, 17(2), 513–886.

Buxton, G.V., Subhani, M.S., 1972. Radiation chemistry and photochemistry of oxychlorine



947–957.



Devasagayam, T.P.A., Kamat, J.P., Mohan, H., Kesavan, P.C., 1996. Caffeine as an antioxidant:

inhibition of lipid peroxidation induced by reactive oxygen species. Biochimica et Biophysica

Acta, 1282(1), 63–70.

Fanghänel, T., Neck, V., Kim, J.I., 1996. The ion product of H2O, dissociation constants of

H2CO3 and Pitzer parameters in the system Na+/H+/OH–/HCO3–/CO3

2–/ClO4–/H2O at 25°C.

Journal of Solution Chemistry, 25(4), 327–343.





6(3), 277–284.



http://scitation.aip.org/content/aip/journal/jpcrd



http://link.springer.com/search?facet-author=%22Th.+Fangh%C3%A4nel%22

76

Kesavan, P.C., Powers, E.L., 1985. Differential modification of oxic and anoxic components of

radiation damage in Bacillus megaterium spores by caffeine. International Journal of

Radiation Biology, 48(2), 223–233.



Kolthoff, I.M., and Bosch, W., 1928. The influence of neutral salts on acid-salt equilibria: III.

The second dissociation constant of carbonic acid and the influence of salts on the activity of

the hydrogen ions in a bicarbonate-carbonate mixture. Recueil des Travaux Chimiques des

Pays-Bas, 47(10), 819–825.

León-Carmona, J.R., Galano, A., 2011. Is caffeine a good scavenger of oxygenated free radicals?

The Journal of Physical Chemistry B, 115(15), 4538–4546.

Liao, C.-H., Kang, S.-F., Wu, F.-A., 2001. Hydroxyl radical scavenging role of chloride and

bicarbonate ions in the H2O2/UV process. Chemosphere, 44(5), 1193–1200.

McCormick, M., O’Keefe, O., Drouin, D., 2013. The transition from GAC to UV/H2O2: taste &

odour control at the Cornwall Water Purification Plant. In: 2013 OWWA Treatment

Committee’s Spring Seminar, Vaughan, Ontario, Canada.

Nowell, L.H., Hoigné, J., 1992. Photolysis of aqueous chlorine at sunlight and ultraviolet


Oliver, B.G., Carey, J.H., 1977. Photochemical production of chlorinated organic in aqueous


Peter, A., von Gunten, U., 2007. Oxidation kinetics of selected taste and odor compounds during

ozonation of drinking water. Environmental Science & Technology, 41(2), 626–631.

Rodenfeldt, E.J., Melcher, B., Linden, K.G., 2005. UV and UV/H2O2 treatment of

methylisoborneol (MIB) and geosmin in water. Journal of Water Supply: Research and

Technology– AQUA, 54(7), 423–434.

Shi, X., Dalal, N.S., Jain, A.C., 1991. Antioxidant behavious of caffeine: efficient scavenging of

hydroxul radicals. Food and Chemical Toxicology, 29(1), 1–6.

Shu, Z., Li, C., Belosevic, M., Bolton, J.R., Gamal El-Din, M., 2014. Application of a solar

UV/chlorine advanced oxidation process to oil sands process-affected water remediation.


http://en.wikipedia.org/wiki/International_Journal_of_Radiation_Biology

http://en.wikipedia.org/wiki/International_Journal_of_Radiation_Biology

http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)0165-0513

http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)0165-0513

77

Sichel., C., Garcia, C., Andre, K., 2011. Feasibility studies: UV/chlorine advanced oxidation

treatment for the removal of emerging contaminants. Water Research, 45(19), 6371–6380.

Verenitch, S.S., Lowe, C.J., Mazumder, A., 2006. Determination of acidic drugs and caffeine in

municipal wastewaters and receiving waters by gas chromatography-ion trap tandem mass

spectrometry. Journal of Chromatography A, 1116(1–2), 193–203.

Wang, D., Bolton J.R., Hofmann, R., 2012. Medium pressure UV combined with chlorine

advanced oxidation for trichloroethylene destruction in a model water. Water Research,

46(15), 4677–4686.




Watts, M.J., Hofmann, R., Rosenfeldt, E., 2012. Low-pressure UV/Cl2 for advanced oxidation

of taste and odor. Journal of American Water Works Association, 104(1), E58–E65.





56(8), 469–477.

Zhang, X., Li, W., Gong, X., Fan, W., Ren, P., 2014. Optimization of the UV/chlorine process

for ammonia removal and disinfection by-products reduction. Desalination and Water

Treatment, 52, published online.

Zhao, Q., Shang, C., Zhang, X., Ding, G., Yang, X., 2011. Formation of halogenated organic

byproducts during medium-pressure UV and chlorine coexposure of model compounds,

NOM and bromide. Water Research, 45(19), 6545–6554.

78

5. FORMATION OF DISINFECTION BY-PRODUCTS IN THE

ULTRAVIOLET/CHLORINE ADVANCED OXIDATION

PROCESS

This chapter has been submitted for publication as follows:

Wang, D., Bolton J.R., Andrews, S.A., Hofmann, R. Formation of disinfection by-products in the

ultraviolet/chlorine advanced oxidation process. Science of the Total Environment.

Abstract Disinfection by-product (DBP) formation may be a concern when applying ultraviolet light

and free chlorine (UV/chlorine) as an advanced oxidation process (AOP) for drinking water

treatment because of the typically large chlorine doses (e.g. 5–10 mg L–1 as free chlorine)

relative to normal chlorination. A potential mitigating factor is the low chlorine contact times

for AOP treatment (e.g. seconds). Full-scale and pilot-scale test results showed minimal

trihalomethane (THM) and haloacetic acid (HAA) formation during UV/chlorine treatment,

while dichloroacetonitrile (DCAN) and bromochloroacetonitrile (BCAN) were produced

rapidly. Adsorbable organic halide (AOX) formation was significant in application of the

UV/chlorine process in water that had not been previously chlorinated, while little additional

formation was observed in prechlorinated water. Chlorine photolysis led to chlorate and bromate

formation, equivalent to approximately 2–17% and 0.01–0.05% of the photolyzed chlorine,

respectively. No perchlorate or chlorite formation was observed. During simulated secondary

disinfection of the AOP-treated water, the DBP formation potential for THMs, HAAs, HANs,

and AOX was observed to increase approximately to the same extent as was observed for

pretreatment using UV combined with hydrogen peroxide (UV/H2O2) AOP.

5.1 Introduction In Chapters 3 and 4, it was demonstrated that the UV/chlorine AOP has comparable

efficiency to UV/H2O2 for the destruction of organic contaminants in water treatment under

certain conditions. One of the concerns associated with the UV/chlorine treatment is the need

for relatively high chlorine doses relative to conventional drinking water chlorination (e.g. 5–10

79

mg L–1 vs. 0.2–2 mg L–1 as free chlorine for disinfection) (Wang et al., 2012; Watts et al., 2007).

This might tend to promote the formation of chlorinated by-products (conventionally referred to

as disinfection by-products, DBPs). A possible mitigating factor is that chlorine contact times

during UV/chlorine treatment might be in the order of seconds because of the rapid and almost

complete chlorine photolysis expected in the UV reactor (Wang et al., 2011), in contrast to the

hours of contact time that are normally associated with DBP formation in chlorine disinfection.

The formation of DBPs associated with chlorination has been widely studied, but little

information is available on the unique high dose/short contact time conditions that would occur

during UV/chlorine treatment. UV/chlorine photolysis also leads to the formation of the chlorine

radical (•Cl), which may react with natural organic matter (NOM) to form chlorinated DBPs.

This has not been experimentally confirmed, although a study reported by Fang et al. (2014)

indicated that •Cl played an important role in the destruction of benzoic acid in a UV/chlorine

AOP under laboratory conditions. Chlorine photolysis may also directly produce inorganic

DBPs of possible health concern, such as perchlorate (ClO4–), chlorate (ClO3

–), and chlorite

(ClO2–) (Buxton and Subhani, 1972; Feng et al., 2010; Kang et al., 2006), as well as bromate

(BrO3–) in the presence of bromide (Br–) (von Gunten and Hoigné, 1994).

The potential impact of UV/chlorine treatment on DBPs includes not only DBPs that may

form during the treatment, but also the possible effect on organic precursors that may

subsequently react with chlorine on secondary chlorination. Previous studies on the impact of

UV/H2O2 pretreatment on subsequent THM and HAA formation have suggested that such an

effect is AOP-dose specific, with low and moderate AOP doses (e.g. UV in the order of 1,000

mJ cm–2) leading to enhanced THM/HAA formation on subsequent chlorination (Dotson et al.,

2010; Kleiser and Frimmel, 2000), while THM and HAA formation following higher AOP

doses (UV of 3,500–5,000 mJ cm–2) have been observed to be decreased (Toor and Mohseni,

2007; Liu et al., 2002).

There have been only a limited number of studies to date on DBP formation following

UV/chlorine treatment, most of which have reported some formation of organic chlorinated

DBPs during the AOP and/or in subsequent secondary chlorination (Liu et al., 2006; Pisarenko

et al., 2013; Weng et al., 2012; Deng et al., 2014; Shah et al., 2011). None of these studies have

simulated the conditions that would be expected to occur in a plant, that is high free chlorine

doses (5–10 mg L–1 as free chlorine), high UV doses (>1,000 mJ cm–2), but short reaction times

80

(<1 minute). This research addresses such conditions, as well as the impact of the AOP

treatment on subsequent secondary chlorination DBP formation. The work was conducted at

full-scale using a medium pressure (MP) UV reactor with a treatment capacity of 8.6 MLD, as

well as using a pilot-scale batch MP UV reactor with a volume of 40 L. The impact of

UV/chlorine treatment on DBPs was compared to parallel treatment using the more

conventional UV/H2O2 AOP.

5.2 Material and Methods

5.2.1 Experimental Procedures

Full-Scale Experiments

Full-scale experiments were conducted at the City of Cornwall Water Purification Plant

(Ontario, Canada). The plant draws water from the St. Lawrence River, which is treated by

prechlorination, alum coagulation, flocculation, settling, conventional sand/anthracite filtration,

and UV disinfection. The water quality after the post-filtration stage is summarized in Table 5.1.

Following filtration, Trojan MP UVSwift 8L24 UV reactors are used for primary disinfection as

well as for UV/H2O2 advanced oxidation during periods of taste and odour problems, which

historically occur in late summer. One UV unit was isolated for this study, and operated with its

100 L s–1 flow going to waste (shown in Figure 5.1).

Table 5.1 Full- and pilot-scale post-filtration water quality parameters

Test Cornwall - early April Cornwall - late May Pilot-scale

Source St. Lawrence River St. Lawrence River Lake Simcoe

Temperature (°C) 3 12 7–11

pH 8.1 7.9 7.5

Turbidity (NTU) 0.04 0.02 0.2 Alkalinity (mg CaCO3 L–1) 85 92 123

TOC (mg C L–1) 1.6 1.5 3.5 Absorbance at 254 nm (cm–1) 0.02 0.02 0.04

Chlorine residual (mg L–1 as free chlorine)

0.3 0.1 0

81

Figure 5.1 Full-scale Trojan UVSwift reactor (left) and pilot-scale Rayox® reactor (right)

The full-scale study involved treating the flow with chlorine alone at three doses (2, 6, and

10 mg L–1 as free chlorine), UV alone, UV/chlorine (2, 6, and 10 mg L–1), UV/H2O2 (1.0, 2.9,

and 4.8 mg L–1 H2O2, the same molar concentrations as chlorine), and at three pH levels (6.5,

7.5, and 8.5). Sodium hypochlorite (NaOCl) stock solution (12.5 wt. %, NSF 60 certified, Olin

Chlor Alkali) or H2O2 stock solution (35 wt. %, NSF 60 certified, Arkema Inc.), provided by the

Cornwall plant, was injected between the filters and the UV reactors. The travel time between

chemical injection immediately downstream of the filter and the UV reactor effluent was

determined by tracer tests to be approximately 30 s at the flowrate of 100 L s–1. The UV dose

from 200 to 400 nm for all treatment conditions was estimated to be approximately 1,800 ± 100

mJ cm–2, based on the method of Wang et al. (2012). DBP formation by UV/chlorine was

compared to that by UV/H2O2 under parallel conditions. pH adjustment was achieved by the

addition of 10% (w/w) sulphuric acid (H2SO4) or 1.5% (w/w) sodium hydroxide (NaOH)

through injection ports beside the chlorine or H2O2 injection port. Once the desired pH value

was achieved, it was found to be very stable in all trials.

The DBPs that were monitored are listed in Table 5.2. DBP samples were collected

downstream of the UV reactor with simultaneous quenching of residual free chlorine (i) using

200 mg L–1 sodium sulphite for THMs (Farré et al., 2011), HAAs (Plummer and Edzwald,

82

2001), and AOX (followed by acidification to pH <2) (Crebelli et al., 2005), (ii) using 50 mg

L–1 H2O2 for HANs, HKs, and CP (Shams El Din and Mohammed, 1998) with addition of

phosphate buffer to pH 4.8–5.5 (USEPA, 1995), and (iii) using 50 mg L–1 ethylenediamine for

inorganic DBPs (USEPA, 1997). In addition to the measurement of these DBPs immediately

after the UV reactor representing 30 s of chlorine contact time, the water downstream of the

reactor that had been treated with UV alone and UV/chlorine at the maximum chlorine dose (i.e.

10 mg L–1) was dosed with chlorine to reach 6.5 mg L–1 as free chlorine (an arbitrarily ‘high’

amount), and then subjected to DBP formation potential (DBP-FP) tests for 24 h at room

temperature, according to a modified uniform formation condition test (Summers et al., 1996).

DBP-FP following UV/H2O2 treatment at 4.8 mg L–1 H2O2 was also tested, but 0.2 mg L–1

catalase from bovine liver (powder, 2,000–5,000 units/mg protein, Sigma-Aldrich) was added to

samples to quench the residual H2O2 before dosing with 6.5 mg L–1 chlorine (Liu et al., 2003).

The purpose of these formation potential tests was to observe the impacts of the UV,

UV/chlorine, and UV/H2O2 pretreatments on subsequent chlorination DBP formation.

Pilot-Scale Experiments

Pilot-scale experiments were carried out in a 40 L Rayox® completely-mixed batch reactor

(Model: PS1-1-120, Calgon Carbon Corporation) equipped with a 1 kW MP lamp (Heraeus

Noblelight GmbH, Germany) (shown in Figure 5.1) to simulate the same UV and chemical

oxidant doses as applied at Cornwall, and the same pH conditions. In this test, however, water

was collected post-filter from the Keswick Water Treatment Plant (Ontario, Canada). This plant

treats water from Lake Simcoe, a water with approximately twice the total organic carbon

(TOC) as the St. Lawrence River (Table 5.1), by conventional treatment processes similar to

those used at Cornwall except that prechlorination is not employed. All experimental and

analytical methods were similar to those used at Cornwall, except for tests of DBP formation by

chlorine alone, which were conducted by chlorinating 500 mL water for 60 s in a beaker with

the same chlorine doses and pH values as those used in the Rayox® reactor. The UV exposure

time in the Rayox® reactor was 40 s, which was found to result in approximately the same

percentage of chlorine decay as was observed at Cornwall. UVCalc® software version 2B (from

Bolton Photosciences Inc.) predicted that the UV dose for 40 s in the Rayox reactor was 1,820 ±

110 mJ cm– 2 (200–400 nm), varying slightly with different chlorine or H2O2 doses and pH

values. DBP-FPs using an initial chlorine concentration of 6.5 mg L–1 were tested after the

83

treatment of UV alone, UV/chlorine at 10 mg L–1, and UV/H2O2 at 4.8 mg L–1, as well as with

no treatment, at the same three pH levels (6.5, 7.5, and 8.5) employed in the full-scale tests at

Cornwall.

Table 5.2 Monitored organic and inorganic DBPs

Group Components Trihalomethanes (THMs) Chloroform/trichloromethane (TCM) Bromodichloromethane (BDCM) Chlorodibromomethane (CDBM) Bromoform/tribromomethane (TBM) Haloacetic acids (HAAs) Monochloroacetic acid (MCAA) Dichloroacetic acid (DCAA) Trichloroacetic acid (TCAA) Monobromoacetic acid (MBAA) Dibromoacetic acid (DBAA) Bromochloroacetic acid (BCAA) Bromodichloroacetic acid (BDCAA) Chlorodibromoacetic acid (CDBAA) Tribromoacetic acid (TBAA) Haloacetonitriles (HANs) Bromochloroacetonitrile (BCAN) Dibromoacetonitrile (DBAN) Dichloroacetonitrile (DCAN) Trichloroacetonitrile (TCAN) Haloketones (HKs) 1,1-Dichloro-2-propanone (DCP) 1,1,1-Trichloro-2-propanone (TCP) Chloropicrin (CP) Adsorbable organic halides (AOX) Inorganic DBPs Chlorite (ClO2

–) Chlorate (ClO3

–) Perchlorate (ClO4

–) Bromate (BrO3

–)

5.2.2 Analytical Methods

Free chlorine was measured using a HACH® spectrophotometer (Model: DR/2500, HACH)

and the DPD method (APHA et al., 2005). A Cecil UV/vis spectrophotometer (Model: CE3055,

Cecil Instruments) was used to determine H2O2 concentrations based on the triiodide method

(Klassen et al., 1994). THMs, HANs, HKs, and CP were extracted and analyzed using gas

chromatography-electron capture detection (GC-ECD, Model: HP 5890 Series II, Hewlett-

Packard), according to USEPA Method 551.1 (USEPA, 1995). HAAs samples were extracted

and methylated using APHA Method 6251 B (APHA et al., 2005) and analyzed using the same

84

GC-ECD equipment. AOX was determined using an AOX analyzer (Model: Xplorer, TE

Instruments) based on APHA Method 5320 (APHA et al., 2005). ClO3– samples at expected

concentrations higher than 50 μg L–1 were analyzed using an ion chromatograph (IC, Model:

Dionex ICS-5000+ analytical system, Thermo Scientific) based on USEPA Method 300.1

(USEPA, 1997). ClO2–, ClO4

–, and BrO3–, as well as ClO3

– at concentrations lower than 50 μg

L–1 were analyzed by the Ministry of Environment and Climate Change of Ontario, Canada,

using an ion chromatograph tandem mass spectrometer (IC-MS/MS, Model: Dionex 2500,

Thermo Scientific) following the method described by Furdui and Tomassini (2010). The

method detection limits (MDLs) for THMs, HAAs, HANs, HKs, and CP ranged from 0.2 to 1.2

μg L–1. The MDL for ClO3– using IC was 7 μg L–1, while the MDLs of ClO2

–, ClO3–, ClO4

–, and

BrO3– using IC-MS/MS ranged from 0.03 to 0.05 μg L–1. Details for the analytical methods are

shown in Appendix G.

5.3 Results The data are presented in the same format for all classes of organic DBPs, so Figure 5.2

(THMs) will be explained in detail to assist the reader. The figure is divided into two principal

sections: short-term THM formation during the 30–60 s of travel time across the UV reactor (the

left side), and THM formation potential (THM-FP) following UV/AOP after being dosed with

6.5 mg L–1 chlorine for 24 h (the right side). The treatment conditions (from left to right) include

reporting any initial DBPs prior to UV/AOP treatment (from prechlorinated water, for example),

the application of chlorine alone in the absence of UV, the application of UV alone at a dose of

approximately 1,800 mJ cm– 2, UV/chlorine treatment at three chlorine concentrations, and then

UV/H2O2 at the same three molar H2O2 concentrations as for chlorine in the UV/chlorine tests.

The right section reports the formation potentials in the water without any pretreatment as a

control, followed by those in the water pretreated by UV, UV/chlorine, and UV/H2O2 at the

highest chemical oxidant dose (i.e. UV with 10 mg L–1 chlorine or UV with 4.8 mg L–1 H2O2).

Each treatment scenario was repeated at three pH values, as shown by the triplets of shaded

bars. Results from all waters tested are shown in the same figure to facilitate comparisons. All

DBPs were investigated in the full-scale test carried out in early April at Cornwall and the pilot-

scale test. HAAs, AOX, and inorganic DBPs were also investigated in the other full-scale test

carried out in late May. All raw data are summarized in Appendix I.

85

Figure 5.2 THM formation in full- and pilot-scale tests. Plots on the left show THMs after various treatment processes for

short reaction time (30–60 s contact time). Plots on the right show THM formation potentials (free chlorine dose: 6.5 mg L–1

for 24 h) in the water pretreated by selected processes shown on the x-axis. Error bars represent the values of the duplicates

measured.

.

THM

s (μ

g L–1

) TH

M-F

P (μ

g L–1

)

pH 6.5 pH 7.5 pH 8.5

Lake Simcoe (pilot-scale)

St. Lawrence River (full-scale test in April)

86

HA

As

(μg

L–1)

HA

A-F

P (μ

g L–1

)

pH 6.5 pH 7.5 pH 8.5

Figure 5.3 HAA formation in full- and pilot-scale tests. Plots on the left show HAAs after various treatment processes for short

reaction time (30–60 s contact time). Plots on the right show HAA formation potentials (free chlorine dose: 6.5 mg L–1 for 24 h) in

the water pretreated by selected processes shown on the x-axis. Error bars represent the values of the duplicates measured.


St. Lawrence River (full-scale test in May)


87

HA

Ns

(μg

L–1)

HA

N-F

P (μ

g L–1

)

pH 6.5 pH 7.5 pH 8.5

Figure 5.4 HAN formation in full- and pilot-scale tests. Plots on the left show HANs after various treatment processes for

short reaction time (30–60 s contact time). Plots on the right show HAN formation potentials (free chlorine dose: 6.5 mg L–1

for 24 h) in the water pretreated by selected processes shown on the x-axis. Error bars represent the values of the duplicates

measured.



88

AO

X (μ

g C

l L–1

) AO

X-FP

(μg

Cl L

–1)

pH 6.5 pH 7.5 pH 8.5

Figure 5.5 AOX formation in full- and pilot-scale tests. Plots on the left show AOX after various treatment processes for short

reaction time (30–60 s contact time). Plots on the right show AOX formation potentials (free chlorine dose: 6.5 mg L–1 for 24 h) in

the water pretreated by selected processes shown on the x-axis. Error bars represent the values of the duplicates measured.


St. Lawrence River (full-scale test in May)


89

5.3.1 THMs

There was no observed short-term THM formation across the UV reactor (30 s contact

time) during the full-scale UV/chlorine or chlorine alone trial at Cornwall, with reactor effluent

concentrations similar to the influent concentrations of about 18 µg L–1 (Figure 5.2). For the

Lake Simcoe pilot system, there was a small formation by chlorine alone of approximately 8 µg

L–1 THMs during the 60 s of contact time at the highest 10 mg L–1 chlorine dose, but the

presence of UV had no observable impact. THMs consisted primarily of TCM (~50%) and

BDCM (~40%) due to the low Br– in these waters (2–3 µg L–1). It is unknown whether the more

brominated species would form more quickly during these short reaction times.

UV and both AOP pretreatments did appear to have a significant impact on THM-FP. UV

alone at a dose of approximately 1,800 mJ cm –2 increased the 24 h THM formation by 20–30%

for both Cornwall and Lake Simcoe waters compared to a non-pretreated control, suggesting

that UV exposure at that dose can create THM precursors. Under AOP conditions, THM-FP

increased even more by 30–110% compared to the controls. Both UV/chlorine and UV/H2O2

generally had the same effect, except at pH 6.5 where UV/chlorine led to more THMs than

UV/H2O2 (increased by 90–110%, compared to the controls). Previous modelling by the authors

suggests that UV/chlorine at pH 6.5 is more effective at producing •OH than UV/H2O2 in a

similar water matrix at that pH (Wang et al., 2012).

5.3.2 HAAs

The predominant HAAs were DCAA and TCAA, contributing approximately 60% of the

total detected HAAs. As shown in Figure 5.3, only a small amount of HAAs was formed by

chlorine alone during the 30–60 s of contact time in either of the full- or pilot-scale water.

However, the UV/chlorine AOP led to HAA formation of up to 13 µg L–1 at the highest chlorine

dose (10 mg L–1) at pH 6.5, with the result more pronounced for Lake Simcoe water than for

Cornwall water (which had been prechlorinated).

Pretreatment by UV alone was observed to increase the 24 h HAA formation potential by

10–25% for Cornwall water compared to a non-pretreated control, but led to a decrease in HAA-

FP by 10–30% for Lake Simcoe water. Similar to THMs, pretreatment by both UV/chlorine and

UV/H2O2 AOPs resulted in a higher 24 h HAA formation when compared to the controls (by

90

40–110% in Cornwall water and 20–90% in Lake Simcoe water). As with the THMs, the effect

was associated with the higher anticipated AOP efficiency, with UV/chlorine at pH 6.5 leading

to the greatest increase in HAA-FP in all waters.

5.3.3 HANs, HKs, and CP

No short-term HK or CP formation was observed across the UV reactor under any of the

conditions. Of the four HANs monitored, only DCAN and BCAN were detected following the

30–60 s contact time, with significantly more formed during UV/chlorine treatment (0.4–2.9 µg

L–1 in Cornwall water and 1.6–5.2 µg L–1 in Lake Simcoe water) than when using chlorine alone

(0–0.3 µg L–1 in Cornwall water and 0.6–0.8 µg L–1 in Lake Simcoe water), as shown in Figure

5.4. The total concentrations of DCAN and BCAN after 30–60 s UV/chlorine treatment could

reach more than 50% and 100% of the 24 h HAN formation potentials in the non-pretreated

controls for Lake Simcoe and Cornwall waters, respectively, although overall concentrations

were still quite low at below 6 µg L–1. The highest concentrations were observed at pH 6.5

relative to the other pH values, which is consistent with a possible role of radicals in their

formation during UV/chlorine treatment, since the higher formation coincided with greater

radical generation at lower pH. This effect, however, may be confounded with other evidence

that HAN formation is promoted at lower pH even in the absence of AOP treatment (Yang et al.,

2007; Glezer et al., 1999; Hansen et al., 2012).

UV pretreatment increased the 24 h HAN formation by 80–100% in Cornwall water and

20–30% in Lake Simcoe water, compared to a non-pretreated control. The formation was

enhanced more significantly by UV/chlorine and UV/H2O2 AOPs, which led to increases of

110–260% and 50–220% in Cornwall and Lake Simcoe waters, respectively. Similar to THMs

and HAAs, the highest formation occurred with UV/chlorine pretreatment at pH 6.5.

5.3.4 AOX

AOX (unit: µg Cl L–1) is a collective parameter to indicate the total halogenated DBPs that

can be adsorbed by activated carbon. Prechlorination at the Cornwall plant resulted in AOX

concentrations in the order of 40 µg Cl L–1 to 80 µg Cl L–1 (April and May, respectively) in the

post-filtration water, as shown in Figure 5.5. Subsequent addition of up to 10 mg L–1 free

chlorine for AOP treatment led to negligible increases in AOX during the 30 s travel across the

91

UV reactor. The situation was quite different for the Lake Simcoe water treated at pilot scale.

This water had not been prechlorinated, so presumably the organic matter would be expected to

be more reactive to chlorine. This was not readily apparent when chlorine alone was added, with

only a small AOX formation of approximately 15 µg Cl L–1 measured during the 60 s of

chlorine contact time. However, when UV/chlorine was applied, AOX formation was increased

in most cases to approximately 20 to 70 µg Cl L–1. Previous research suggests that advanced

oxidation can increase the THM, HAA, and HAN formation potentials, presumably by altering

the organic matter to make it more readily reactive with chlorine to form such by-products (Toor

and Mohseni, 2007; Pisarenko et al., 2013; Glauner et al., 2005). The AOX results suggest that

this alteration can produce organic precursors that can react with chlorine very quickly—within

60 s—to form AOX species, but that the species formed during this time do not include

appreciable THMs or HAAs (but some DCAN and BCAN). Again, the maximum AOX

formation during the 60 s of UV/chlorine contact time occurred at pH 6.5 (70 µg Cl L–1), where

it is predicted that maximum •OH formation occurs.

Pretreatment by AOPs generally led to much smaller differences in 24 h AOX formation

potential (AOX-FP) relative to the non-pretreated controls than were observed for the THMs,

HAAs, or HANs under parallel conditions. For example, the largest increase in 24 h AOX-FP

was when pretreating with UV/chlorine at pH 6.5, with a 30–60% increase compared to the non-

pretreated controls in all waters. This is compared to 90–110% for THMs and HAAs, and 220–

260% for HANs. This also suggests that UV/chlorine treatment changes the nature of the AOX

precursors, such that they react more quickly (as observed during the 60 s chlorine contact time

in the pilot-scale test discussed earlier), but that there is only a relatively small increase in the

total precursor concentration when given a long enough contact time (24 h) to form AOX. In

other words, AOP pretreatment makes the AOX precursors react more quickly but leads to a

smaller increase in the total precursor material relative to THMs and HAAs, under the

conditions tested.

Of the total AOX-FP without any pre-UV or AOP treatment, the unidentified fraction

consistently made up about 40–50% for both Cornwall and Lake Simcoe waters (data not

shown). This is consistent with much previous research (e.g. Pourmoghaddas and Stevens, 1995;

Dotson et al., 2010). When the samples were then exposed to UV photolysis, the unidentified

fraction of the AOX-FP was reduced by 2–15%. Pretreatment by UV/chlorine or UV/H2O2 led

92

to an additional 2–15% reduction in unidentified DBPs. This suggests that UV and AOPs

convert the high-molecular-weight AOX precursors to low-molecular-weight precursors that

lead to the formation of identifiable DBPs. Based on the data shown previously, the increase in

identifiable DBPs exhibits a disproportionately higher increase in HAN-FP (100–260% increase

relative to no pretreatment), compared to THM-FP and HAA-FP (30–110% increases relative to

no pretreatment for both).

5.3.5 Inorganic DBPs: ClO4–, ClO3


–

No ClO4–, ClO3

–, ClO2–, or BrO3

– was formed by the application of H2O2 alone, or by UV

alone or UV/H2O2. The free chlorine solution contained a trace level of ClO2– approximately

equivalent to 0.2% of the final free chlorine concentration, as well as lower amounts of ClO4–,

and BrO3– (0.001–0.01% and 0.01–0.03%, respectively). ClO3

– was present in the free chlorine

at a much higher concentration, ranging from 1–15% depending on the chlorine source. ClO3– in

hypochlorite solutions arises from auto-decomposition of free chlorine during manufacture,

shipment, and storage (Stanford et al., 2011).

The pre-existing ClO2– concentration decreased by more than 99% on exposure to

UV/chlorine treatment at 10 mg L–1. However, the UV/chlorine AOP was observed to have no

effect on the measured ClO4–, with the pre-existing ClO4

– in the free chlorine solution remaining

unchanged. The minimal formation of ClO4– and ClO2

– is consistent with reports by Buxton and

Subhani (1972) and Feng et al. (2010).

BrO3– formation in the order of 0.1 µg L–1 to 2 µg L–1 was observed on UV/chlorine

treatment, with the higher formation occurring at the lower pH. Chlorine photolysis was

reported to produce •OH and ozone (Forsyth et al., 2013), which both play important roles in

BrO3– formation from Br– (Hofmann and Andrews, 2006). BrO3

– normally is formed in an

ozonated water preferentially at higher pH (Song et al. 1996), so this trend likely reflects the

greater •OH formation by UV/chlorine at lower pH. While this bromate formation is reasonably

small, the pre-existing bromate in the hypochlorite solution available at the Cornwall plant (full-

scale test in April) added approximately 3 µg L–1 to the water when the chlorine was dosed at 10

mg L–1. Therefore, the possible formation of BrO3– during the application of UV/chlorine needs

to be considered in view of bromate limits that are often in the order of 10 µg L–1 (e.g. Health

Canada, 2012).

93

ClO3– was found to be a major product of chlorine photolysis. Approximately 2–17% (by

mass) of the photolyzed free chlorine was converted to ClO3–. There is a possible trend that can

be observed in Figure 5.6 whereby the percentage formation of chlorate from photolyzed

chlorine may increase with both increasing chlorine dose and pH; however, there was not

sufficient data generated to determine this trend with statistical confidence. Buxton and Subhani

(1972) observed that the percentage of photolyzed free chlorine being converting to ClO3– was

9% when exposed to monochromatic UV light at 365 or 313 nm. They also found that the

percentage increased to 17% at 254 nm, a result that was substantiated by Feng et al. (2010),

who found 20–30% of the photolyzed free chlorine became ClO3– at 254 nm. It is noted that MP

UV was used in this research, different from the studies carried out by Buxton and Subhani

(1972) and Feng et al. (2010).

Figure 5.6 Formation of ClO3

– relative to free chlorine photodecomposition in the full- and

pilot-scale experiments. Error bars represent the values of the duplicates measured. Low,

medium, and high represent free chlorine doses of 2, 6, and 10 mg L–1, respectively.

Some jurisdictions have limits on allowable ClO3– in drinking water. In Canada, a national

guideline of 1 mg L–1 exists (Health Canada, 2012). Assuming 17% conversion of chlorine to

chlorate on photolysis, the maximum chlorine concentration that can undergo photolysis for a

UV/chlorine AOP would be 5.8 mg L–1, assuming that no additional chlorate is added through

94

free chlorine solution contamination. In this work, not all chlorine that was applied was

photolyzed. For example, at the highest chlorine dose of 10 mg L–1, approximately up to 6 mg

L–1 of free chlorine underwent photolysis, however the UV reactors used for the full-scale

testing in this study were early AOP models and may not apply as high a UV dose as would be

used today. As such, it would be expected that more complete chlorine photolysis would be

experienced across modern reactors.

5.4 Discussion

5.4.1 Rapid DBP Formation within the UV/Chlorine Reactor

To the authors’ knowledge, this is the first work to explore DBP formation during the very

short chlorine contact times (30–60 s) but relatively high chlorine doses (2–10 mg L–1)

associated with a UV/chlorine AOP. Previous work that explored DBP formation from

UV/chlorine used longer contact times. For example, Pisarenko et al. (2013) compared THM

and HAA formation when exposing a natural water to either UV/chlorine or chlorine alone, for

2 h. Exposure to UV/chlorine led to increases in both THM and HAA formation of up to 27 µg

L–1 relative to chlorination alone. Similarly, Weng et al. (2012) observed significantly greater

formation of DCAN from its precursors, L-histidine and L-arginine, when applying UV/chlorine

over 10–30 min of exposure compared to chlorine alone.

In this work, THM and HAA formation by UV/chlorine during the 30–60 s contact time

were always observed to be below 14 µg L–1, and usually much lower. DCAN and BCAN

formation was quick, with up to 6 µg L–1 formation within 30–60 s, approximately equivalent to

the 24 h formation potential. AOX formation in water that had previously been chlorinated was

minimal. For the water that had not been prechlorinated, little AOX was formed because of the

chlorine exposure alone (15 µg Cl L–1), but UV/chlorine produced up to 70 µg Cl L–1. This

suggests that the formation of AOX within the short contact time of a UV reactor is enhanced in

part by UV photolysis and/or radical oxidation.

While the formation of organic DBPs during the short exposure time was relatively small,

chlorate formation was significant, at up to 17% (by mass) of the free chlorine that underwent

photolysis. These results suggest that chlorate may be a limiting DBP associated with

UV/chlorine treatment for waters similar to those tested in this study. The formation of chlorate

95

is likely to be manageable through careful chlorine dosing. More research is needed, however,

to identify factors that might increase the conversion of chlorine to chlorate during UV/chlorine

treatment.

5.4.2 Impact of UV/Chlorine on 24 Hour DBP-FP

Previous studies have reported the impact of AOP pretreatment on subsequent DBP-FP, but

usually associated with the UV/H2O2 AOP. In the previous work, low AOP doses (e.g., UV in

the order of 1,000 mJ cm–2) often caused increases in THM/HAA-FP (Dotson et al., 2010;

Kleiser and Frimmel, 2000; Liu et al., 2006; Shah et al., 2011; Liu et al., 2012), but higher AOP

doses (UV of 3,500–5,000 mJ cm–2) led to reductions in THM/HAA-FP (Toor and Mohseni,

2007; Liu et al., 2002). Our tests applied AOP doses that were in the ‘low to moderate’ range

where often THM/HAA-FP reportedly increased. Consistent with this previous work, the results

showed that THM, HAA, HAN, and AOX formation potentials were all increased by the

UV/chlorine and UV/H2O2 AOPs. In general, UV/chlorine and UV/H2O2 when applied at molar

equivalent concentrations were found to have similar effects on the subsequent 24 h DBP-FP,

except that the greatest increase in FP was consistently associated with UV/chlorine at pH 6.5,

where the greatest formation of •OH is predicted to occur (Wang et al., 2012).

In terms of the individual components of THMs and HAAs, increases of TCM-FP and

BDCM-FP by UV/chlorine and UV/H2O2 were observed to be higher than the increase of

CDBM-FP. For example, at pH 7.5 with the highest chlorine and H2O2 doses (10 and 4.8 mg

L–1, respectively) TCM-FP and BDCM-FP increased by 30–60% and 60–80%, respectively,

while less than a 10% increase in CDBM-FP was found. Difference in DCAA-FP from TCAA-

FP and BCAA-FP were also observed. Under the same condition of the previous example,

DCAA-FP increasd by approximately 50–100% because of the AOP pretreatment, while

TCAA-FP and BCAA-FP increased by less than 20% in most cases. Further investigation

should be conducted to explore the impact on individual DBP species in more detail.

5.4.3 Role of the Chlorine Radical (•Cl)

•Cl and •OH are simultaneously produced on chlorine photolysis, but it is difficult to

distinguish the role of each one in terms of DBP formation. In general, •Cl is a strong but

selective oxidant. It is scavenged quickly (8.5 × 109 M–1 s–1) by chloride (Cl–) that is usually

96

present in free chlorine solutions to form •Cl2–, which in turn has negligible reactivity with

NOM (Buxton et al., 2000; Jayson et al., 1973; Nagarajan and Fessenden, 1985). The role of •Cl

in forming organic DBPs might therefore depend on its competitive reactivity with NOM

compared to chloride. Unfortunately, no reaction rate with NOM has been reported. As a

hypothetical example, if it is assumed that •Cl reacts with NOM at the same rate as with

benzene (1.8 × 1010 M–1 s–1; Fang et al., 2014), then, with the same amount of NOM and

chloride present in the Cornwall and Lake Simcoe waters, it might be predicted that about 13-

14% of the •Cl might react with the NOM, suggesting to some that the role of •Cl in organic

DBP formation would be non-negligible. More research in this area is needed.

5.5 Conclusions The UV/chlorine AOP process is attractive because of its operational simplicity and its

comparable efficiency to the UV/H2O2 process under some conditions. The most significant

potential drawback, however, is likely to be concerns about DBP formation. This research

suggests that under the conditions tested, the formation of organic DBPs within the UV reactor

could be characterized as ‘low’ relative to regulatory limits and current DBP concentrations at

the sites from which water was collected. There was evidence, however, that AOP treatment

formed DBP precursors, leading to higher organic DBP concentrations in formation potential

tests. This effect was similar for both UV/chlorine and UV/H2O2 treatment. However, since the

formation potential tests are designed to accentuate DBP formation by applying unrealistically

high chlorine concentrations, the magnitude of any increases in DBP formation in the

distribution system following AOP treatment is therefore likely to be much lower than was

observed in the formation potential tests in this study.

The most likely limiting factor for UV/chlorine treatment would be the formation of

chlorate, for waters similar to those studied in this work. The maximum conversion of chlorine

to chlorate that was observed was 17% (by mass).

Acknowledgements The authors express their appreciation to the staff at the Cornwall Water Purification Plant,

including Owen O'Keefe, Daniel Drouin, and Morris McCormick, for their help with the full-

scale experiments. The authors also would like to thank Dr. Vasile Furdui from the Ontario

97

Ministry of Environment and Climate Change for his generous help in analyzing inorganic DBP

samples.

This work was funded by the Natural Sciences and Engineering Research Council of

Canada through the Industrial Research Chair program.


21st edition. American Public Health Association, American Water Works Association, and


Buxton, G.V., Bydder, M., Salmon, G.A., Williams, J.E., 2000. The reactivity of chlorine atoms

in aqueous solution. Part III. The reaction of Cl• with solutes. Physical Chemistry Chemical

Physics, 2, 237–245.

Buxton, G.V., Subhani, M.S., 1972. Radiation chemistry and photochemistry of oxychlorine



958–969.

Crebelli, R., Conti, L., Monarca, S., Feretti, D., Zerbini, I., Zani, C., Veschetti, E., Cutilli, D.,

Ottaviani, M., 2005. Genotoxicity of the disinfection by-products resulting from peracetic

acid- or hypochlorite-disinfected sewage wastewater. Water Research, 39(6), 1105–1113.

Deng, L., Huang, C.-H., Wang, Y.-L., 2014. Effects of combined UV and chlorine treatment on

the formation of trichloronitromethane from amine precursors. Environmental Science &

Technology, 48(5), 2697–2705.

Dotson, A.D., Keen, V.S., Metz, D., Linden, K.G., 2010. UV/H2O2 treatment of drinking water

increases post-filtration DBP formation. Water Research, 44(12), 3703–3713.



Farré, M.J., Reungoat, J., Argaud, F.X., Rattier, M., Keller, J., Gernjak, W., 2011. Fate of N-

nitrosodimethylamine, trihalomethane and haloacetic acid precursors in tertiary treatment

including biofiltration. Water Research, 45(17), 5695–5704.

http://en.wikipedia.org/wiki/Physical_Chemistry_Chemical_Physics

http://en.wikipedia.org/wiki/Physical_Chemistry_Chemical_Physics



98

Feng, Y., Simith, D.W., Bolton, J.R., 2010. A potential new method for determination of the

fluence (UV dose) delivered in UV reactors involving the photodegradation of free chlorine.

Water Environment Research, 82(4), 328–334.

Forsyth, J.E., Zhou, P., Mao, Q., Asato, S.S., Meschke, J.S., Dodd, M.C., 2013. Enhanced

inactivation of Bacillus subtilis spores during solar photolysis of free available chlorine.


Furdui, V.I., Tomassini, F., 2010. Trends and sources of perchlorate in arctic snow.


Glauner, T., Kunz, F., Zwiener, C., Frimmel, F.H., 2005. Elimination of swimming pool water

disinfection by-products with advanced oxidation processes (AOPs). Acta Hydrochimica et

Hydrobiologica, 33(6), 585–594.

Glezer, V., Harris, B., Tal, N., Iosefzon, B., Lev, O., 1999. Hydrolysis of haloacetonitriles:

linear free energy relationship, kinetics and products. Water Research, 33(8), 1938–1948.

Hansen, K.M.S., Willach, S., Antoniou, M.G., Mosbæk, H., Albrechtsen, H.-J., Andersen, H.R.,

2012. Effect of pH on the formation of disinfection byproducts in swimming pool water – Is

less THM better? Water Research, 46(19), 6399–6409.




Hofmann, R., Andrews, R.C., 2006. Impact of H2O2 and (bi)carbonate alkalinity on ammonia’s

inhibition of bromate formation. Water Research, 40(18), 3343–3348.

Jayson, G.G., Parsons, B.J., Swallow, A.J., 1973. Some simple, highly reactive, inorganic

chlorine derivatives in aqueous solution. Their formation using pulse of radiation and their

role in the mechanism of the Fricke dosimeter. Journal of the Chemistry Society, Faraday

Transaction 1: Physical Chemistry in Condensed Phases, 69, 1597–1607.

Kang, N., Anderson, T.A., Jackson, W.A., 2006. Photochemical formation of perchlorate from

aqueous oxychlorine anions. Analytica Chimica Acta, 567(1), 48–56.



http://www.sciencedirect.com/science/article/pii/S0043135412006513

99

Kleiser, G., Frimmel, F.H., 2000. Removal of precursors for disinfection by-products (DBPs) –

differences between ozone- and OH-radical-induced oxidation. The Science of the Total

Environment, 256(1), 1–9.

Liu, W., Andrews, S.A., Bolton, J.R., Linden, K.G., Sharpless, C., Stefan, M., 2002.

Comparison of disinfection byproduct (DBP) formation from different UV technologies at

bench scale. Water Science and Technology: Water Supply, 2(5–6), 515–521.

Liu, W., Andrews, S.A., Stefan, M.I., Bolton, J.R., 2003. Optimal methods for quenching H2O2

residuals prior to UFC testing. Water Research, 37(15), 3697–3703.

Liu, W., Cheung, L.-M., Yang, X., Shang, C., 2006. THM, HAA and CNCl formation from UV

irradiation and chlor(am)ination of selected organic waters. Water Research, 40(10), 2033–

2043.

Liu, W., Zhang, Z., Yang, X., Xu, Y., Liang, Y., 2012. Effects of UV irradiation and

UV/chlorine co-exposure on natural organic matter in water. Science of the Total

Environment, 414, 576–584.

Nagarajan, V., Fessenden, R.W., 1985. Flash photolysis of transient radicals. 1. X2– with X = Cl,

Br, I., and SCN. The Journal of Physical Chemistry, 89(11), 2330–2335.

Pisarenko, A.N., Stanford, B.D., Snyder, S.A., Rivera, S.B., Boal, A.K., 2013.Investigation of

the use of chlorine based advanced oxidation in surface water: oxidation of natural organic

matter and formation of disinfection byproducts. Journal of Advanced Oxidation

Technologies, 16(1), 137–150.

Plummer, J.D., Edzwald, J.K., 2001. Effect of ozone on algae as precursors for trihalomethane

and haloacetic acid production. Environmental Science & Technology, 35(18), 3661–3668.

Pourmoghaddas, H., Stevens, A.A., 1995. Relationship between trihalomethanes and haloacetic

acids with total organic halogen during chlorination. Water Research 29 (9), 2059–2062.

Shah, A.D., Dotson, A.D., Linden, K.G., Mitch, W.A., 2011. Impact of UV disinfection

combined with chlorination/chloramination on the formation of halonitromethanes and

haloacetonitriles in drinking water. Environmental Science & Technology, 45(8), 3657–3664.

Shams El Din, A.M., Mohammed, R.A., 1998. Kinetics of the reaction between hydrogen

peroxide and hypochlorite. Desalination, 115(2), 145–153.

100

Song, R., Donohoe, C., Minear, R., Westerhoff, P., Ozekin, K., Amy, G., 1996. Empirical

modeling of bromate formation during ozonation of bromide-containing waters. Water

Research, 30(5), 1161–1168.

Stanford, B.D., Pisarenko, A.N., Snyder, S.A., Gordon, G., 2011. Perchlorate, bromate, and

chlorate in hypochlorite solutions: guidelines for utilities. Journal of American Water Works

Association, 103(6), 71–83.

Summers, R.S., Hooper, S.M., Shukairy, H.M., Solarik, G., Owen, D., 1996. Assessing DBP

yield: uniform formation conditions. Journal of American Water Works Association, 88(6),

80–93.

Toor, R., Mohseni, M., 2007. UV-H2O2 based AOP and its integration with biological activated

carbon treatment for DBP reduction in drinking water. Chemosphere, 66(11), 2087–2095.

United States Environmental Protection Agency (USEPA), 1995. Method 551.1. Determination

of chlorination disinfection byproducts, chlorinated solvents, and halogenated

pesticides/herbicides in drinking water by liquid-liquid extraction and gas chromatography

with electron-capture detection, Revision, 1.0. Available from

<http://www.epa.gov/sam/pdfs/EPA-551.1.pdf> (accessed 11.09.11).

United States Environmental Protection Agency (USEPA), 1997. Determination of inorganic

anions in drinking water by ion chromatography. Available from

<http://water.epa.gov/scitech/drinkingwater/labcert/upload/met300.pdf > (accessed 09.09.14).

Von Gunten, U., Hoigné, J., 1994. Bromate formation during ozonation of bromide-containing

waters: interaction of ozone and hydroxyl radical reactions. Environmental Science &

Technology, 28(7), 1234–1242.

Wang, D., Bolton J.R., Hofmann, R., 2012. Medium pressure UV combined with chlorine

advanced oxidation for trichloroethylene destruction in a model water. Water Research,

46(15), 4677–4686.







101



56(8), 469–477.

Weng, S., Li, J., Blatchley III, E.R., 2012. Effects of UV254 irradiation on residual chlorine and

DBPs in chlorination of model organic-N precursors in swimming pools. Water Research,

46(8), 2674–2682.

Yang, X., Chang, C., Westerhoff, P., 2007. Factors affecting formation of haloacetonitriles,

haloketones, chloropicrin and cyanogen halides during chloramination. Water Research, 41(6),

1193–1200.

102

6. SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS

6.1 Summary and Conclusions Chlorine photolysis by UV exposure is a novel AOP, which probably involves simpler

operation than UV/H2O2, but has not been widely investigated in the literature. This work has

explored the mechanisms of chlorine photolysis through bench-scale experiments and numerical

simulation (Chapter 3). Based on the results for HOCl and OCl– photolysis at the same initial

molar concentration (0.15 mM), although OCl– absorbed MP UV light approximately 4.5 times

higher than HOCl, and the quantum yields of photolysis and •OH formation for both HOCl and

OCl– were quite close to each other, the •OH concentration generated by HOCl was

approximately 40 times higher than that generated by OCl–. This implies that OCl– is a much

stronger •OH scavenger than HOCl.

Bench-scale TCE destruction (Chapter 3), pilot-scale caffeine destruction, and full-scale

geosmin, MIB, and caffeine destruction (Chapter 4) all showed that the efficiency of the

UV/chlorine AOP was generally comparable to that of the UV/H2O2 AOP at pH 7.5 and 8.5, but

was superior at pH 6.5. This reflects the impact of the prevalence of HOCl relative to OCl– at

the lower pH, with its reported lower •OH scavenger potential.

The formation of various DBPs by the UV/chlorine AOP was investigated in full- and

pilot-scale experiments in two post-filtration waters (Chapter 5) under practical operating

conditions (chlorine doses: 2–10 mg L–1, contact time: <1 minute). THM and HAA formation

was observed to be minimal, while fast formation of DCAN and BCAN was found. Fast

formation of AOX was observed in the water without prechlorination, while insignificant

additional formation was found in the prechlorinated water. Among the monitored inorganic

DBPs, including chlorate, bromate, perchlorate, and chlorite, only chlorate and bromate were

formed during chlorine photolysis, equivalent to approximately 2–17% and 0.01–0.05% of

photolyzed chlorine, respectively. Additionally, DBP formation potential for THMs, HAAs,

HANs, and AOX was observed to increase following UV/chlorine pretreatment, approximately

to the same extent as was observed following UV/H2O2 pretreatment.

In summary, UV/chlorine is a promising AOP that may be an alternative to the UV/H2O2

AOP for the destruction of organic contaminants and the control of taste and odour issues in

103

drinking water treatment. Under the tested conditions, UV/chlorine produces low concentrations

of organic DBPs within the UV reactors relative to the regulatory limits. The process does

appear to form DBP precursors that would consequently increase DBP formation during

secondary chlorination, but the effect is similar to that of UV/H2O2 treatment. Chlorate is a

significant photoproduct of chlorine photolysis. This is the single factor that may possibly limit

UV/chlorine applications.

6.2 Recommendations for Future Work A number of areas were identified that deserve further investigation, including:

• Investigation of rate constant of free chlorine with •OH. The reaction rate constants

between HOCl/OCl– and •OH were obtained from limited studies, and need to be verified.

Pulse radiolysis analysis is often used to determine reaction rate constants involving •OH,

and it was successfully used in this work to verify the rate constant between OCl– and •OH

by measuring the change of the •OCl absorbance (shown in Appendix A). •OCl is a product

of the reaction between OCl– and •OH. However, the value between HOCl and •OH could

not be measured, because there was no significant increase of •OCl absorbance. A novel

approach is therefore required.

• Investigation of quantum yields of chlorine photolysis and •OH formation without the

impact of chain reactions. Chlorine photolysis has been found to involve chain reactions

that may influence the apparent quantum yields of chlorine photolysis and •OH formation.

It is thus important to determine the quantum yields accurately in the absence of the chain

reactions, such that accurate kinetic models of chlorine photolysis can be built. A chain

reaction inhibitor may be applied for this purpose. However, the inhibitor needs to be

selected carefully. For example, it should be a strong •OH scavenger, but stable in the

presence of UV or chlorine individually. It should not absorb UV light significantly, and its

products with •OH must not react with free chlorine.

• Investigation of UV/chlorine efficiency in a wide range of water quality. UV/chlorine

efficiency was evaluated in 3 types of water in this study, which could all be considered to

have low- to moderately-low •OH scavenging potential (TOC levels less than 3.5 mg L–1

and alkalinity less than 150 mg L–1 as CaCO3). Since UV/chlorine is probably more

competitive with UV/H2O2 in a water with a higher •OH scavenging potential, it is

104

important to investigate UV/chlorine performance over a wide range of water quality, and

in particular in waters with a higher TOC level where not only higher scavenging potential

exists but where the TOC may exert a significant chlorine demand.

• Further investigation of DBP formation. Potential DBP formation is likely the most

significant concern with UV/chlorine treatment, and while the results of this study are

promising in this area, only two natural water matrices were evaluated. More waters should

be tested, such as a water containing more DBP precursors and/or bromide (which converts

to hypobromous acid in a chlorinated water to produce brominated organic DBPs). The

formation of chlorate may possibly be a limiting factor for UV/chlorine treatment, and the

influence of water quality and treatment variables on its formation are unknown.

• Investigation of UV/chlorine using low-pressure lamps. This research was carried out

exclusively using MP UV lamps. Since LP UV lamps are also commonly used in practice,

and have a very different spectral emittance from that of MP lamps, it is necessary to

investigate the UV/chlorine mechanisms, efficiency, and DBP formation with the

application of LP lamps.

105

APPENDICES

A. Pulse Radiolysis Analysis for Determination of Rate Constant of Free

Chlorine with Hydroxyl Radical Pulse radiolysis was performed at the Notre Dame Radiation Laboratory (NDRL) for

determination of the rate constants between free chlorine (HOCl and OCl–) and •OH, according

to the methods described by Buxton and Subhani (1972)1, Ulanski and von Sonntag (2000)2,

and Westerhoff et al. (2007)3. pH values were adjusted to 5 and 10, respectively, to obtain

generally pure HOCl and OCl– solutions. A single electron pulse was applied to impinge a free

chlorine solution and to make the solution excited and ionized instantly. Three transients

were produced during this process (with the relative abundance): •OH (45%), H (10%), and eaq–

(45%). Since the solutions were saturated with N2O, eaq– was converted to •OH promptly.

Therefore, it is considered that after electron pulse within several nanoseconds 90% of transients

were •OH and 10% were H atoms that are less reactive than •OH. The produced •OH then

reacted with chlorine and produced •OCl, which was monitored by a spectrophotometer at 280

nm in very short time intervals (2.5 × 10–7 s). By measuring the absorbance changes versus time

at different initial chlorine concentrations, the second order rate constant between chlorine and

•OH can be determined.

Figure A.1 shows the raw data and the regression for pulse radiolysis analysis for rate

constant between OCl– and •OH at pH 10. The y axis is the solution absorbance at 280 nm,

which was primarily contributed by •OCl. Therefore, it can be considered to be the absorbance

of •OCl solely. The x axis shows the time scale after the pulse was performed.

The main reaction between OCl– and •OH is: OCl– + •OH → •OCl + OH–. This is an

irreversible reaction, thus

1 Buxton, G.V., Subhani, M.S., 1972. Radiation chemistry and photochemistry of oxychlorine ions. Part 1.–Radiolysis of aqueous solutions of hypochlorite and chlorite ions. Journal of the Chemical Society, Faraday Transactions 1: Physical Chemistry in Condensed Phases, 68, 947–957. 2 Ulanski, P., von Sonntag, C., 2000. OH-radical-induced chain scission of chitosan in the absence and presence of dioxygen. Journal of the Chemical Society, Perkin Transactions 2, 2000, 2022–2028. 3 Westerhoff, P., Mezyk, S.P., Cooper, W.J., Minakata, D., 2007. Electron pulse radiolysis determination of hydroxyl radical rate constants with Suwannee River fulvic acid and other dissolved organic matter isolates. Environmental Science & Technology, 41(13), 4640–4646.



106

Figure A.1 Pulse radiolysis results for rate constant between OCl– with •OH

[ OCl] [ OH][OCl ]d kdt

−•= • [A.1]

Since the concentration of OCl– was much higher than that of •OH, the change of OCl–

concentration was minimal. Therefore, [OCl–] is considered to be constant. •OH was generated

immediately (at t = 0) after the pulse radiolysis was applied, and it was gradually consumed by

OCl– until depleted completely when the curves reached plateaus shown in Figure A.1. When

•OH was just exhausted, •OCl reached the maximum concentration. Then it started to decrease,

because •OCl could not be generated any more, but underwent decay. Assuming ' [OCl ]k k −= ,

[ OCl] '[ OH]d kdt•

= • [A.2]

On the other hand, [ OH] [ OH][OCl ]d kdt

−•= − • , or

[ OH] '[ OH]d kdt•

= − • [A.3]

Integrating Equation [A.3] gives

0

[ OH]ln '[ OH]

k t•= −

• [A.4]

If k’ values at different [OCl–] are known, k can be determined by plotting k’ versus [OCl–].

Therefore, the purpose was to determine k’ from the raw data shown in Figure A.1. k’ can be

calculated using the data fitting method discussed below.

107

Supposing •OH was consumed by OCl– only, besides the Equations [A.2] and [A.3], there

was a relationship between [•OCl] and [•OH]:

[•OH] = [•OH]0 – [•OCl] [A.5]

which means that the concentration of •OH at time t was equal to the initial concentration of

•OH at time 0 minus the concentration of •OCl at time t, because 1 mole of •OCl was produced

by 1 mole of •OH consumed.

Substituting Equation [A.5] to Equation [A.2] yields 0[ OCl] '([ OH] [ OCl])d k

dt•

= • − • . After

integration, '

0[ OCl] [ OH] (1 )k te−• = • − [A.6]

The absorbance of •OCl (shown in y-axis of Figure A.1) also had the same relationship

with t as [•OCl]. Since the initial absorbance of the solution before pulse radiolysis was not

zero, the relationship between •OCl absorbance and t can be expressed as: '

0k ty y Ae−= + [A.7]

where, y is the absorbance of •OCl at time t, k’ is the observed first order rate constant in

Equations [A.2] and [A.3], y0 and A are constants.

According to the least square regression method, k’ as well as y0 and A was solved using

the “Solver” function in Excel. Since ' [OCl ]k k −= , k was then determined, which is the slope of

k’ as a function of [OCl–]. k was calculated to be 7.17 × 109 M–1 s–1, which is close to 9.0 × 109

M–1 s–1 determined by Buxton and Subhani (1972).

The raw data for pulse radiolysis analysis for the rate constant between HOCl and •OH at

pH 5 are shown in Figure A.2. It shows that there was no significant trend of •OCl absorbance

increase, which means the formation of •OCl was not obvious. There are several probable

reasons:

• The product of HOCl with •OH was not •OCl, thus the product cannot be detected at 280 nm.

However, since HOCl is the conjugate acid of OCl–, •OCl is likely the main product.

• The reaction rate between HOCl and •OH is slow compared with that between OCl– with

•OH. Therefore, the principal portion of •OH was consumed by Cl–, according to Equation

[2.34], and, thus, •OCl concentration was too low to detect.

A different approach to measure the rate constant between HOCl and •OH is therefore required.

108

Figure A.2 Pulse radiolysis results for rate constant between HOCl with •OH

B. Determination of the Fluence Rate of the MP Lamp in the Collimated

Beam Apparatus A ferrioxalate actinometer method was used to determine the fluence rate of the MP lamp

in the collimated beam apparatus, following the procedures described by Bolton et al. (2009)4.

The actinometer with a volume of 10 mL was contained in a Petri dish (diameter = 4.9 cm),

with/without the coverage of a 345 nm long-pass filter. An opaque aluminum cap (diameter =

10 cm) with a circular hole (diameter = 1.50 cm, area = 1.77 cm2) was also placed at the centre

of the Petri dish, so that the determined fluence rate reflected the value at the centre of the Petri

dish without the impact of a Petri factor. After the lamp exposure for 3 min in a dark room, the

rate of Fe2+ generation (d[Fe2+]/dt) due to 200–345 nm UV exposure can be determined by

subtracting the d[Fe2+]/dt in the presence of the filter from the d[Fe2+]/dt in the absence of the

filter, according to Sharpless and Linden (2003)5. The fluence rate (200–400 nm) was then

calculated using the spreadsheet shown in Table B.1. The determined value was then used to

calculate the average fluence rate in a specific trichloroethylene solution, using the Bolton

spreadsheet (Bolton, 2002)6.

4 Bolton, J.R., Stefan, M.I., Shaw, P.-S., Lykke, K.R., 2009. Determination of the quantum yield of the ferrioxalate and KI/KIO3 actinometers and a method for the calibration of radiometer detectors. CDROM Proceedings 5th UV World Congress, Amsterdam, The Netherlands. 5 Sharpless, C.M., Linden, K.G., 2003. Experimental and model comparisons of low- and medium-pressure Hg lamps for the direct and H2O2 assisted UV photodegradation of N-nitrosodimethylamine in simulated drinking water. Environmental Science & Technology, 37(9), 1933–1940. 6 Bolton, J.R., 2002. Germicidal fluence (UV dose) calculation for a medium pressure UV lamp. Available at: http://www.iuva.org or from J.R. Bolton ([email protected]). Accessed on Sep. 24, 2009.

109

Table B.1 Calculation of 200–400 nm fluence rate from the actinometry and spectroradiometer results A B C D E F G H I J K L

Wavelength (nm)

Photon energya

(J einstein–1)

Assumed spectral irradiance from the spectroradiometerb

(µW cm–2 nm–1)

Assumed spectral photon irradiancec

(einstein cm–2 s–1 nm–1)

∆λd (nm)

Reflection factor

from air to watere

Quantum yield of

Fe2+ formationf

Assumed d[Fe2+]/dt

200–345 nmg

Actual d[Fe2+]/dt

200–345 nmh

Correction factori

Actual spectral photon

irradiancej (einstein cm–2 s–1)

Spectral irradiancek (mW cm–2)

199.81 200.03 5.98E+05 0 0 0.22 0.969 1.49 0 2.60E-06 2.22E-06 0 0.00E+00 200.24 5.97E+05 0 0 0.21 0.970 1.49 0 0 0.00E+00 200.46 5.97E+05 0 0 0.22 0.970 1.49 0 0 0.00E+00 200.67 5.96E+05 0 0 0.21 0.970 1.49 0 0 0.00E+00 200.89 5.95E+05 0 0 0.22 0.970 1.49 0 0 0.00E+00

⁞ ⁞ ⁞ ⁞ ⁞ ⁞ ⁞ ⁞ ⁞ ⁞ 253.84 4.71E+05 19.0 4.03E-05 0.22 0.975 1.40 2.14E-03 1.97E-11 9.29E-03 254.05 4.71E+05 15.8 3.37E-05 0.21 0.975 1.40 1.70E-03 1.57E-11 7.39E-03 254.26 4.70E+05 13.3 2.82E-05 0.21 0.975 1.40 1.43E-03 1.32E-11 6.19E-03 254.47 4.70E+05 11.0 2.34E-05 0.21 0.975 1.40 1.19E-03 1.09E-11 5.14E-03 254.69 4.70E+05 9.74 2.07E-05 0.22 0.975 1.40 1.10E-03 1.01E-11 4.76E-03 254.9 4.69E+05 9.80 2.09E-05 0.21 0.975 1.40 1.06E-03 9.74E-12 4.57E-03

⁞ ⁞ ⁞ ⁞ ⁞ ⁞ ⁞ ⁞ ⁞ ⁞ 398.57 3.00E+05 2.43 8.11E-06 0.20 0.979 1.25 2.08E-03 3.60E-12 1.08E-03 398.78 3.00E+05 2.47 8.22E-06 0.21 0.979 1.25 2.21E-03 3.83E-12 1.15E-03 398.98 3.00E+05 2.41 8.02E-06 0.20 0.979 1.25 2.06E-03 3.56E-12 1.07E-03 399.19 3.00E+05 2.41 8.03E-06 0.21 0.979 1.25 2.16E-03 3.74E-12 1.12E-03 399.40 3.00E+05 2.38 7.95E-06 0.21 0.979 1.25 2.14E-03 3.71E-12 1.11E-03 399.60 2.99E+05 2.43 8.10E-06 0.20 0.979 1.25 2.08E-03 3.60E-12 1.08E-03 399.81 2.99E+05 2.43 8.12E-06 0.21 0.979 1.25 2.18E-03 3.79E-12 1.13E-03

Total 1.17E+00 (200– 345 nm) 7.36

(200–400 nm)

aPhoton energy (J einstein–1) = 6.6216 × 10–34 ×2.9979 × 108 × 6.02214 × 1023 / (Column A × 10–9), according to Bolton (2001)7 bColumn C was from the data determined by the spectroradiometer (Model: USB4000-UV-VIS, Ocean Optics) cColumn D = Column C / Column B d∆λ represents the difference of the two adjacent wavelengths in Column A eColumn F is calculated using the Schiebener’s model described by Huiber (1997)8 fData were obtained from Goldstein and Rabani (2008)9 gColumn H = Column G × Column D × Column E × 1.77 × Column F / (10 / 1000) hThis result was from the ferrioxalate actinometry measurements iCorrection factor = Actual d[Fe2+]/dt from 200–345 nm / the sum of Column H from 200–345 nm jColumn K = correction factor × Column D × Column E kColumn L = Column K × Column B × 1000

7 Bolton, J. R., 2001. Ultraviolet Applications Handbook. Bolton Photosciences Inc. 628 Cheriton Cres. NW, Edmonton, AB, Canada. 8 Huibers, P.D.T., 1997. Models for the wavelength dependent of the index of refraction of water. Applied Optics, 36(16), 3785–3787. 9 Goldstein, S., Rabani, J., 2008. The ferrioxalate and iodide-iodate actinometers in the UV region. Journal of Photochemistry and Photobiology A: Chemistry, 193(1), 50–55.

110

C. Example of Matlab® Codes

C.1 Matlab® Codes for Simulation of Trichloroethylene Decay by the UV/Chlorine AOP at

11 mg L–1 and pH 5

function dydt = TCE_UV_Cl2(t,y); % y(1): OH radical, y(2): HOCl, y(3): OCl–, y(4): DOC, y(5): TCE, y(6): HCO3

–, y(7):CO32–, y(8): Cl radical formed by TCE photolysis

phiHOCl_OH=0.79; phiOCl_OH=1.18; %data from my TCE experiments % phiHOCl_OH and phiOCl_OH mean quantum yields of OH radical formation by HOCl and OCl– photolysis, respectively phiHOCl=1.06, phiOCl=0.89; %data from my TCE experiments % phiHOCl and phiOCl mean quantum yields of HOCl and OCl– photolysis, respectively Ep0=[0.00E+00 0.00E+00 0.00E+00 0.00E+00 0.00E+00 0.00E+00 4.74E-12 4.43E-12 2.87E-12 3.88E-12 3.33E-12 3.17E-12 4.02E-12 3.48E-12 3.45E-12 4.61E-12 4.89E-12 4.33E-12 5.74E-12 6.38E-12 5.66E-12 7.72E-12 8.44E-12 7.23E-12 9.48E-12 1.03E-11 8.47E-12 1.04E-11 1.04E-11 1.10E-11 8.92E-12 9.96E-12 9.23E-12 7.40E-12 1.03E-11 9.63E-12 5.89E-12 9.10E-12 9.14E-12 7.28E-12 8.54E-12 4.80E-12 3.31E-12 2.77E-12 4.26E-12 4.86E-12 5.23E-12 1.59E-11 2.14E-11 8.87E-12 5.22E-12 6.80E-12 1.47E-11 1.22E-11 8.10E-12 1.17E-11 1.73E-11 2.84E-11 2.90E-11 2.53E-11 1.73E-11 1.73E-11 1.43E-11 1.24E-11 2.90E-11 4.54E-11 2.40E-11 7.70E-12 9.40E-12 1.49E-11 1.17E-11 8.78E-12 5.88E-12 5.39E-12 6.42E-12 1.23E-11 8.62E-12 4.06E-12 5.20E-12 1.49E-11 2.93E-11 1.52E-11 7.04E-12 4.34E-12 3.94E-12 3.40E-12 4.11E-12 3.67E-12 9.11E-12 1.42E-11 1.09E-11 4.85E-12 7.86E-12 4.93E-12 3.98E-12 7.26E-12 4.11E-11 2.85E-11 1.45E-11 5.72E-12 7.14E-12 2.84E-11 7.53E-11 3.77E-11 8.42E-12 3.91E-12 4.09E-12 3.71E-12 3.59E-12 2.95E-12 4.07E-12 1.41E-11 1.07E-10 1.01E-10 5.88E-11 1.45E-11 6.71E-12 3.95E-12 4.22E-12 3.72E-12 3.42E-12 3.27E-12 2.46E-12 2.93E-12 2.89E-12 2.80E-12 2.15E-12 2.65E-12 2.52E-12 2.54E-12 2.62E-12 2.15E-12 2.79E-12 1.42E-11 2.74E-11 1.05E-11 2.76E-12 2.35E-12 2.39E-12 2.51E-12 1.77E-12 2.05E-12 2.07E-12 2.28E-12 2.65E-12 2.12E-12 2.51E-12 2.43E-12 2.22E-12 2.12E-12 2.32E-12 1.89E-12 2.17E-12 2.46E-12 3.57E-12 3.04E-12 2.55E-12 3.07E-12 2.96E-12 3.11E-12 3.11E-12 2.62E-12 4.05E-12 8.74E-12 1.70E-10 3.77E-10 1.99E-10 4.68E-11 1.79E-11 8.85E-12 7.77E-12 6.04E-12 4.49E-12 3.09E-12 3.58E-12 3.64E-12 3.13E-12 3.04E-12 3.34E-12 2.79E-12 3.00E-12 2.85E-12 2.95E-12 2.55E-12 2.40E-12 2.02E-12 2.66E-12 2.38E-12 2.43E-12 3.45E-12 7.50E-12 5.28E-12 2.29E-12 2.33E-12 2.31E-12 2.32E-12 2.36E-12 2.33E-12 2.48E-12 2.02E-12]; %Ep0 is a matrix, which represents the spectral photon irradiance at each wavelength (200–400 nm) at the centre of the Petri dish, and which corresponds to a total irradiance of 1 mW cm–2 from 200 to 400 nm Ep1=7.10*Ep0; %Ep1 is the actual photon irradiance at the centre of the Petri dish used in TCE experiments E1=[8022 7769 7440 7109 6805 6541 6306 6089 5885 5675 5449 5179 4876 4546 4212 3892 3613 3374 3165 2972 2785 2601 2412 2223 2040 1858 1682 1512 1352 1198 1055 922 801 689 588 501 423 354 293 242 198 160 128 102 80 61 47 36 27 19 13 8 3 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02]; %E1 is the spectral molar absorption coefficient of TCE at each wavelength (200 to 400 nm) E2=[91 77 66 58 53 49 46 44 43 42 42 43 44 46 48 50 52 55 58 61 64 68 71 74 77 80 83 86 89 91 93 95 96 97 98 98 98 98 97 96 95 94 92 90 87 85 82 79 76 73 70 67 64 61 58 55 53 50 47 45 43 41 39 37 36 35 33 32 31 31 30 30 29 29 29 28 28 28 28 28 29 29 29 29 29 29 29 30 30 30 30 30 30 30 29 29 29 29 28 28 28 27 27 26 26 25 25 24 24 23 22 22 21 20 20 19 18 17 17 16 15 15 14 13 13 12 12 11 10 10 9 9 8 8 7 7 7 6 6 5 5 5 4 4 4 4 3 3 3 3 3 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02 1.00E-02]; %E2 is the spectral molar absorption coefficient of HOCl at each wavelength (200 to 400 nm) E3=[362 325 294 267 242 220 199 180 162 144 128 113 99 86 74 63 53 45 38 32 26 22 18 15 13 11 10 9 8 7 7 7 7 7 8 8 9 10 11 12 14 15 17 19 21 24 26 30 33 37 41 45 50 55 60 67 73 80 87 95 103 111 120 130 139 149 159 169 180 190 201 212 223 234 245 255 266 276 286 295 304 312 320 327 334 339 345 349 352 355 357 358 359 359 357 355 353 350 346 342 337 331 325 318 311 304 296 288 280 271 262 253 244 235 226 217 208 199 190 181 173 164 156 148 140 133 126 118 112 105 99 93 87 82 77 72 67 62 58 54 50 47 43 40 37 34 32 29 27 25 23 21 20 18 17 15 14 13 12 11 10 9 8 8 7 6 6 5 5 5 4 4 4 3 3 3 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1.00E-02]; %E3 is the spectral molar absorption coefficient of OCl– at each wavelength (200 to 400 nm) A=E1*y(5)+E2*y(2)+E3*y(3);

111

%A is the solution absorption coefficient at each wavelength (200 to 400 nm) RF=[0.969457168 0.969647411 0.969831759 0.970010483 0.970183836 0.970352058 0.970515374 0.970673997 0.970828125 0.970977949 0.971123644 0.971265381 0.971403318 0.971537605 0.971668386 0.971795796 0.971919963 0.972041009 0.972159049 0.972274195 0.97238655 0.972496214 0.972603282 0.972707845 0.972809989 0.972909796 0.973007345 0.97310271 0.973195965 0.973287177 0.973376412 0.973463732 0.973549199 0.97363287 0.973714799 0.973795041 0.973873646 0.973950662 0.974026138 0.974100117 0.974172643 0.974243758 0.974313502 0.974381913 0.974449029 0.974514885 0.974579516 0.974642954 0.974705233 0.974766382 0.974826432 0.974885411 0.974943347 0.975000267 0.975056196 0.97511116 0.975165183 0.975218288 0.975270497 0.975321834 0.975372318 0.97542197 0.97547081 0.975518857 0.975566131 0.975612648 0.975658426 0.975703483 0.975747834 0.975791496 0.975834484 0.975876813 0.975918497 0.975959552 0.975999989 0.976039824 0.976079068 0.976117735 0.976155836 0.976193383 0.976230389 0.976266864 0.976302819 0.976338264 0.976373211 0.976407669 0.976441648 0.976475158 0.976508208 0.976540806 0.976572963 0.976604685 0.976635983 0.976666863 0.976697334 0.976727404 0.97675708 0.97678637 0.976815281 0.976843819 0.976871992 0.976899806 0.976927268 0.976954385 0.976981162 0.977007605 0.977033721 0.977059515 0.977084993 0.977110161 0.977135024 0.977159586 0.977183854 0.977207833 0.977231527 0.97725494 0.977278079 0.977300947 0.97732355 0.97734589 0.977367973 0.977389804 0.977411385 0.977432722 0.977453817 0.977474676 0.977495301 0.977515697 0.977535867 0.977555815 0.977575544 0.977595057 0.977614359 0.977633452 0.977652339 0.977671024 0.97768951 0.9777078 0.977725896 0.977743802 0.97776152 0.977779054 0.977796406 0.977813579 0.977830575 0.977847397 0.977864048 0.97788053 0.977896845 0.977912996 0.977928985 0.977944815 0.977960488 0.977976006 0.977991371 0.978006586 0.978021652 0.978036571 0.978051347 0.978065979 0.978080472 0.978094826 0.978109044 0.978123126 0.978137076 0.978150895 0.978164584 0.978178146 0.978191582 0.978204894 0.978218083 0.978231151 0.9782441 0.978256931 0.978269646 0.978282246 0.978294733 0.978307108 0.978319373 0.978331528 0.978343577 0.978355519 0.978367356 0.97837909 0.978390721 0.978402252 0.978413683 0.978425015 0.978436251 0.97844739 0.978458434 0.978469385 0.978480244 0.978491011 0.978501688 0.978512275 0.978522775 0.978533188 0.978543514 0.978553756]; %RF is the reflection factor at each wavelength (200 to 400 nm) WF=(1-10.^(-A*0.80))./(2.3026*A*0.8); %WF is the water factor at each wavelength (200 to 400 nm) Ep=Ep1.*RF.*WF*0.9814*0.945; %Ep is the average spectral photon irradiance at each wavelength (200 to 400 nm) through the TCE solution. 0.9814 is the divergence factor, 0.945 is the Petri factor sumEp=sum(sum(Ep)); %sumEp is the total photon irradiance at 200–400 nm through the TCE solution R1=((Ep./WF).*E1.*(1-10.^(-A*0.80)))./(A*0.80)*1000; %R1 is the specific rate of photon absorption by TCE at each wavelength (200 to 400 nm) R2=((Ep./WF).*E2.*(1-10.^(-A*0.80)))./(A*0.80)*1000; %R2 is the specific rate of photon absorption by HOCl at each wavelength (200 to 400 nm) R3=((Ep./WF).*E3.*(1-10.^(-A*0.80)))./(A*0.80)*1000; %R3 is the specific rate of photon absorption by OCl– at each wavelength (200 to 400 nm) sumR1=sum(sum(R1)); sumR2=sum(sum(R2)); sumR3=sum(sum(R3)); %sumR means the summation of specific rate of photon absorption from 200 to 400 nm OHFormation=phiHOCl_OH*sumR2*y(2)+phiOCl_OH*sumR3*y(3); % OHFormation means OH radical formation rate HOClPhotolysis=phiHOCl*sumR2*y(2); OClPhotolysis=phiOCl*sumR3*y(3); % HOClPhotolysis and OClPhotolysis mean HOCl and OCl– photolysis rates, respectively % Assume HOClPhotolysis and OClPhotolysis are positive numbers kforward=1.8e3, kbackward=3e9; % for HOCl hydrolysis kw=1.9055e-14; %[OH–] = kw/10^-pH kforward2=2.2, kbackward2=5e10; % for HCO3

– hydrolysis kHOCl_OH=8.46e4; kOCl_OH=9e9; kDOC_OH=3e8; kTCE_OH=2.4e9; kHCO3_OH=8.5e6; kCO3_OH=3.9e8; phiTCE=0.354; phiTCE1=0.13; %data from Li et al. (2004) % phiTCE means the total quantum yield of TCE photolysis; phiTCE1 means the quantum yield of TCE photolysis in Reaction 1 (see Table 3.5) TCEPhotolysis=phiTCE*sumR1*y(5); ClFormation = phiTCE1*sumR1*y(5); % TCEPhotolysis means TCE photolysis rate, ClFormation means the Cl radical formation rate % Assume TCEPhotolysis is a positive number kTCE_Cl=4.88e10; dydt=[OHFormation-kHOCl_OH*y(2)*y(1)-kOCl_OH*y(3)*y(1)-kDOC_OH*y(1)*y(4)-kTCE_OH*y(1)*y(5)-kHCO3_OH*y(1)*y(6)-kCO3_OH*y(1)*y(7); kforward*y(3)-kbackward*y(2)*kw/(10^-pH)-HOClPhotolysis-kHOCl_OH*y(2)*y(1); -kforward*y(3)+kbackward*y(2)*kw/(10^-pH)-OClPhotolysis-kOCl_OH*y(3)*y(1); -kDOC_OH*y(4)*y(1); -kTCE_OH*y(5)*y(1)-TCEPhotolysis-kTCE_Cl*y(5)*y(8); -kforward2*y(6)+kbackward2*y(7)*10^-pH-kHCO3_OH*y(6)*y(1); kforward2*y(6)-kbackward2*y(7)*10^-pH-kCO3_OH*y(7)*y(1); ClFormation-kTCE_Cl*y(5)*y(8)]; end close all clear all

112

clc pH=5; y0 = [0,1.55e-4/(1+10^(pH-7.52)),1.55e-4-1.55e-4/(1+10^(pH-7.52)),8.33e-6,1.10e-6,0,0,0]; tspan = [0:1:40]; [t,y]=ode15s('TCE_UV_Cl2',tspan,y0);

C.2 Matlab® Codes for Simulation of Trichloroethylene Decay by the UV/Chlorine AOP at

11 mg L–1 and pH from 5 to 10

function dydt = TCE_MP_Cl2_pHX(t,y,pH); % y(1): OH radical, y(2): HOCl, y(3): OCl–, y(4): DOC, y(5): TCE, y(6): HCO3

–, y(7):CO32–, y(8): Cl radical formed by TCE photolysis

phiHOCl_OH=1.4; phiOCl_OH=0.28; %data from Watts and Linden (2007) % phiHOCl_OH and phiOCl_OH mean quantum yields of OH radical formation by HOCl and OCl– photolysis, respectively phiHOCl=3.7, phiOCl=1.7; %data from Watts and Linden (2007) % phiHOCl and phiOCl mean quantum yields of HOCl and OCl– photolysis, respectively Ep0=[1.10E-11 1.11E-11 8.89E-12 1.08E-11 1.09E-11 8.86E-12 1.13E-11 1.10E-11 8.74E-12 1.10E-11 1.09E-11 8.87E-12 1.11E-11 1.10E-11 8.78E-12 1.13E-11 1.10E-11 8.92E-12 1.12E-11 1.11E-11 8.99E-12 1.12E-11 1.18E-11 9.12E-12 1.15E-11 1.17E-11 9.66E-12 1.18E-11 1.19E-11 1.18E-11 9.68E-12 1.20E-11 1.20E-11 9.56E-12 1.22E-11 1.23E-11 9.49E-12 1.26E-11 1.22E-11 9.92E-12 1.23E-11 1.18E-11 1.16E-11 9.50E-12 1.19E-11 1.20E-11 9.80E-12 1.41E-11 1.65E-11 1.05E-11 1.20E-11 1.22E-11 1.48E-11 1.23E-11 1.23E-11 1.34E-11 1.32E-11 1.90E-11 1.93E-11 1.87E-11 1.40E-11 1.61E-11 1.52E-11 1.29E-11 1.95E-11 2.93E-11 1.83E-11 1.13E-11 1.40E-11 1.66E-11 1.31E-11 1.39E-11 1.28E-11 1.29E-11 1.08E-11 1.67E-11 1.43E-11 1.02E-11 1.28E-11 1.66E-11 2.62E-11 1.51E-11 1.38E-11 1.27E-11 1.24E-11 1.04E-11 1.29E-11 1.25E-11 1.51E-11 1.78E-11 1.61E-11 1.29E-11 1.60E-11 1.13E-11 1.30E-11 1.35E-11 3.77E-11 2.72E-11 1.84E-11 1.40E-11 1.52E-11 2.41E-11 6.76E-11 3.26E-11 1.54E-11 1.10E-11 1.33E-11 1.31E-11 1.33E-11 1.06E-11 1.35E-11 1.78E-11 9.68E-11 9.51E-11 5.13E-11 2.05E-11 1.60E-11 1.17E-11 1.41E-11 1.36E-11 1.38E-11 1.34E-11 1.05E-11 1.32E-11 1.32E-11 1.31E-11 1.03E-11 1.32E-11 1.29E-11 1.31E-11 1.31E-11 1.06E-11 1.34E-11 2.38E-11 4.06E-11 2.02E-11 1.09E-11 1.32E-11 1.29E-11 1.35E-11 1.04E-11 1.29E-11 1.28E-11 1.27E-11 1.30E-11 1.04E-11 1.28E-11 1.29E-11 1.27E-11 1.27E-11 1.24E-11]; %Ep0 is a matrix, which represents the spectral photon irradiance at each wavelength (200–400 nm) at the centre of the Petri dish, and which corresponds to a total irradiance of 1 mW cm–2 from 200 to 400 nm Ep=5.11*Ep0; %Ep1 is the actual photon irradiance at the centre of the Petri dish used for the model E1=[7.20E+03 6.90E+03 6.60E+03 6.30E+03 6.10E+03 5.80E+03 5.70E+03 5.50E+03 5.30E+03 5.10E+03 4.80E+03 4.60E+03 4.30E+03 4.00E+03 3.70E+03 3.50E+03 3.20E+03 3.00E+03 2.80E+03 2.70E+03 2.50E+03 2.30E+03 2.10E+03 2.00E+03 1.85E+03 1.70E+03 1.55E+03 1.35E+03 1.20E+03 1.10E+03 1.00E+03 9.00E+02 8.00E+02 7.00E+02 6.00E+02 5.00E+02 4.00E+02 3.50E+02 3.00E+02 2.50E+02 2.00E+02 1.75E+02 1.50E+02 1.25E+02 1.00E+02 8.00E+01 6.00E+01 4.00E+01 2.00E+01 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100 1.00E-100]; %E1 is the spectral molar absorption coefficient of TCE at each wavelength (200 to 400 nm) E2=[91 77 66 58 53 49 46 44 43 42 42 43 44 46 48 50 52 55 58 61 64 68 71 74 77 80 83 86 89 91 93 95 96 97 98 98 98 98 97 96 95 94 92 90 87 85 82 79 76 73 70 67 64 61 58 55 53 50 47 45 43 41 39 37 36 35 33 32 31 31 30 30 29 29 29 28 28 28 28 28 29 29 29 29 29 29 29 30 30 30 30 30 30 30 29 29 29 29 28 28 28 27 27 26 26 25 25 24 24 23 22 22 21 20 20 19 18 17 17 16 15 15 14 13 13 12 12 11 10 10 9 9 8 8 7 7 7 6 6 5 5 5 4 4 4 4 3 3 3 3 3]; %E2 is the spectral molar absorption coefficient of HOCl at each wavelength (200 to 400 nm) E3=[362 325 294 267 242 220 199 180 162 144 128 113 99 86 74 63 53 45 38 32 26 22 18 15 13 11 10 9 8 7 7 7 7 7 8 8 9 10 11 12 14 15 17 19 21 24 26 30 33 37 41 45 50 55 60 67 73 80 87 95 103 111 120 130 139 149 159 169 180 190 201 212 223 234 245 255 266 276 286 295 304 312 320 327 334 339 345 349 352 355 357 358 359 359 357 355 353 350 346 342 337 331 325 318 311 304 296 288 280 271 262 253 244 235 226 217 208 199 190 181 173 164 156 148 140 133 126 118 112 105 99 93 87 82 77 72 67 62 58 54 50 47 43 40 37 34 32 29 27 25 23]; %E3 is the spectral molar absorption coefficient of OCl– at each wavelength (200 to 400 nm) A=E1*y(5)+E2*y(2)+E3*y(3); %A is the solution absorption coefficient at each wavelength (200 to 400 nm) R1=(Ep.*E1.*(1-10.^(-A*3.45)))./(A*3.45)*1000; %R1 is the specific rate of photon absorption by TCE at each wavelength (200 to 400 nm) R2=(Ep.*E2.*(1-10.^(-A*3.45)))./(A*3.45)*1000; %R2 is the specific rate of photon absorption by HOCl at each wavelength (200 to 400 nm) R3=(Ep.*E3.*(1-10.^(-A*3.45)))./(A*3.45)*1000;

113

%R3 is the specific rate of photon absorption by OCl– at each wavelength (200 to 400 nm) sumR1=sum(sum(R1)); sumR2=sum(sum(R2)); sumR3=sum(sum(R3)); %sumR means the summation of specific rate of photon absorption from 200 to 400 nm OHFormation=phiHOCl_OH*sumR2*y(2)+phiOCl_OH*sumR3*y(3); % OHFormation means OH radical formation rate HOClPhotolysis=phiHOCl*sumR2*y(2); OClPhotolysis=phiOCl*sumR3*y(3); % HOClPhotolysis and OClPhotolysis mean HOCl and OCl– photolysis rates, respectively % Assume HOClPhotolysis and OClPhotolysis are positive numbers kforward=1.8e3, kbackward=3e9; % for HOCl hydrolysis kw=1.9055e-14; ; %[OH–] = kw/10^-pH kforward2=2.2, kbackward2=5e10; % for HCO3

– hydrolysis kHOCl_OH=8.46e4; kOCl_OH=8e9; kDOC_OH=3e8; kTCE_OH=2.4e9; kHCO3_OH=8.5e6; kCO3_OH=3.9e8; phiTCE=0.354; phiTCE1=0.13; %data from Li et al. (2004) % phiTCE means the total quantum yield of TCE photolysis; phiTCE1 means the quantum yield of TCE photolysis in Reaction 1 (see Table 3.5) TCEPhotolysis=phiTCE*sumR1*y(5); ClFormation = phiTCE1*sumR1*y(5); % TCEPhotolysis means TCE photolysis rate, ClFormation means the Cl radical formation rate % Assume TCEPhotolysis is a positive number kTCE_Cl=4.88e10; dydt=[OHFormation-kHOCl_OH*y(2)*y(1)-kOCl_OH*y(3)*y(1)-kDOC_OH*y(1)*y(4)-kTCE_OH*y(1)*y(5)-kHCO3_OH*y(1)*y(6)-kCO3_OH*y(1)*y(7); kforward*y(3)-kbackward*y(2)*kw/(10^-pH)-HOClPhotolysis-kHOCl_OH*y(2)*y(1); -kforward*y(3)+kbackward*y(2)*kw/(10^-pH)-OClPhotolysis-kOCl_OH*y(3)*y(1); -kDOC_OH*y(4)*y(1); -kTCE_OH*y(5)*y(1)-TCEPhotolysis-kTCE_Cl*y(5)*y(8); -kforward2*y(6)+kbackward2*y(7)*10^-pH-kHCO3_OH*y(6)*y(1); kforward2*y(6)-kbackward2*y(7)*10^-pH-kCO3_OH*y(7)*y(1); ClFormation-kTCE_Cl*y(5)*y(8)]; end close all clear all clc tspan = [0:.1:50]; pH = 4.9; for i=1:51, %Set up a loop for different pH. i changes from 1 to 51, which makes pH changes from 5 to 10 pH=pH+0.1; y0 = [0,1.27e-4/(1+10^(pH-7.52)),1.27e-4-1.27e-4/(1+10^(pH-7.52)),5.42e-5,3.81e-8,2.88e-3/(1+10^(pH-10.36)),2.88e-3-2.88e-3/(1+10^(pH-10.36)),0]; s(i)=pH; %pH is represented by s(i) [t,y]=ode15s(@(t,y)TCE_MP_Cl2_pHX(t,y,pH),tspan,y0); %@(t,y) means ode15s solves the function dydt only considering t and y as variables, other values, like pH, are considered as parameters. Z(i)=y(20,1); %In the matrix of [y1(t1), y2(t1), y3(t1), y4(t1) % y1(t2), y2(t2), y3(t2), y4(t2) % y1(t3), y2(t3), y3(t3), y4(t3) % ...........................................] % y(20,1) means y(1) (i.e. OH radical) at 20th time (0.1s *20 = 2s) end plot(s,Z) %plot the relationship between pH and y(1) at 20th time

D. Estimation of OH Radical Concentration Using an Excel Spreadsheet Besides the Matlab® calculation, a Microsoft Excel spreadsheet can be used to estimate the

•OH concentration in the UV/chlorine and UV/H2O2 AOPs. An example for the spreadsheet is

briefly shown in Table D.1. The steady-state •OH concentration was assumed in the calculation.

114

Table D.1 Estimation of the •OH concentration generated by the UV/chlorine AOP at 11 mg L–1 and pH 5

Wavelength λ (nm)

Incident photon irradiance Ep(λ)a

(einstein cm–2 s–1)

Reflection Factorb

Water Factorc

Incident photon irradiance (adjusted by factors) (einstein cm–2 s–1)d

Cl2 molar absorption coefficient

ε (M–1 cm–1)

Solution absorption coefficient a (cm–1)e

Specific rate of photon absorption

(einstein mol–1 s–1)

[•OH]ss (M)

HOCl OCl– HOClf OCl–g Spectralh Totali A B C D E F G H I J K L

200 0 0.969 0.979 0 91 362 0.0230 0.00E+00 0.00E+00 0.00E+00 9.3E-13 201 0 0.970 0.981 0 77 325 0.0206 0.00E+00 0.00E+00 0.00E+00 202 0 0.970 0.983 0 66 294 0.0186 0.00E+00 0.00E+00 0.00E+00 203 0 0.970 0.985 0 58 267 0.0170 0.00E+00 0.00E+00 0.00E+00 204 0 0.970 0.986 0 53 242 0.0158 0.00E+00 0.00E+00 0.00E+00 205 0 0.970 0.987 0 49 220 0.0148 0.00E+00 0.00E+00 0.00E+00 206 3.36E-11 0.971 0.987 2.99E-11 46 199 0.0141 3.15E-06 1.37E-05 3.21E-15 207 3.15E-11 0.971 0.988 2.80E-11 44 180 0.0136 2.83E-06 1.16E-05 2.88E-15 208 2.04E-11 0.971 0.988 1.82E-11 43 162 0.0132 1.79E-06 6.77E-06 1.82E-15

⁞ ⁞ ⁞ ⁞ ⁞ ⁞ ⁞ ⁞ ⁞ ⁞ ⁞ 254 5.75E-11 0.975 0.992 5.16E-11 58 60 0.0090 6.92E-06 7.18E-06 6.95E-15 255 8.28E-11 0.975 0.992 7.43E-11 55 67 0.0086 9.47E-06 1.14E-05 9.52E-15 256 1.23E-10 0.975 0.993 1.10E-10 53 73 0.0082 1.33E-05 1.85E-05 1.34E-14 257 2.02E-10 0.975 0.993 1.81E-10 50 80 0.0078 2.09E-05 3.33E-05 2.10E-14

⁞ ⁞ ⁞ ⁞ ⁞ ⁞ ⁞ ⁞ ⁞ ⁞ ⁞ 395 1.65E-11 0.979 1.000 1.50E-11 0.01 0.63 1.85E-06 3.45E-10 2.19E-08 4.43E-19 396 1.68E-11 0.979 1.000 1.52E-11 0.01 0.62 1.84E-06 3.50E-10 2.16E-08 4.48E-19 397 1.65E-11 0.979 1.000 1.50E-11 0.01 0.55 1.81E-06 3.46E-10 1.89E-08 4.31E-19 398 1.76E-11 0.979 1.000 1.60E-11 0.01 0.53 1.80E-06 3.68E-10 1.94E-08 4.56E-19 399 1.43E-11 0.979 1.000 1.30E-11 0.01 0.01 1.56E-06 2.99E-10 2.99E-10 3.00E-19

[HOCl] 1.54E-04 M [OCl–] 4.66E-07 M

ΦOH, HOCl 0.79 mole einstein–1 ΦOH, OCl- 1.18 mole einstein–1 kHOCl, ·OH 8.46E+04 M–1 s–1 kOCl-, ·OH 9.00E+09 M–1 s–1

aData of Column B are from Column K of Table B.1 bData of Column C are from Column F of Table B.1 cWater factor = (1 – 10–al) / [ln(10)al], based on Bolton and Linden (2003)10, where a is the decadic absorption coefficient of the solution (Column G); l is the solution depth, l = 0.795 cm for 15 mL solution contained in a Petri dish with a diameter of 4.9 cm. dColumn E = Column B × Column C × Column D × 0.9814 (divergence factor) × 0.945 (Petri factor) eColumn H = [HOCl] × Column F + [OCl–] × Column G fColumn I = Column E / Column D × Column F × (1 – 10–al) / (al) × 1000 gColumn J = Column E / Column D × Column G × (1 – 10–al) / (al) × 1000 hColumn K = (ΦOH, HOCl × [HOCl] × Column I + ΦOH, OCl- × [OCl–] × Column J) / (kHOCl, ·OH × [HOCl] + kOCl-, ·OH × [OCl–]) iThe total [[•OH]ss from 200 to 400 nm is the sum of Column K

10 Bolton, J.R., Linden, K.G., 2002. Standardization of methods for fluence (UVdose) determination in bench-scale UV experiments. Journal of Environmental Engineering, 129(3), 209–215.

115

E. UV Dose Estimation Using UVCalc® Version 2B UVClac® version 2B purchased from Bolton Photosciences Inc. was used to calculate the

UV doses delivered by the 1 kW MP UV lamp to the 40 L Lake Simcoe post-filtration water in

the Rayox® reactor for 40 s exposure. An example of the software interface associated with the

input of parameters is shown in the following screen images (Figure E.1). Since various

chlorine/H2O2 doses and pH values were applied, which changed the solution absorbances,

calculated UV doses were different in different water matires, which are summerized in Table

E.1.

116

Figure E.1 UVcalc® 2B interface and parameter selection

Table E.1 Calculated UV doses (mJ cm–2) in the Rayox® reactor

pH 6.5 pH 7.5 pH 8.5 UV/chlorine 2 mg L–1 1885 1847 1774 UV/chlorine 6 mg L–1 1813 1751 1681 UV/chlorine 10 mg L–1 1795 1663 1572 UV/H2O2 1.0 mg L–1 1881 1862 1839 UV/H2O2 2.9 mg L–1 1862 1855 1840 UV/H2O2 4.8 mg L–1 2071 1887 1835

F. Absorption Spectra of Geosmin, MIB, and Caffeine The molar absorption coefficients of geosmin, MIB, and caffeine at wavelengths from 200

to 400 nm are shown in Figure F.1. UV absorption is an important parameter that allows the

photodecomposition to take place. However, the determined molar absorption coefficients of

117

geosmin and MIB are different from those reported by Jo et al. (2011)11, Rosenfeldt et al.

(2005)12, and Kutschera et al. (2009)13, which also conflict with each other. The reason is

unknown, but may be due to the interference of methanol in the UV absorption of the water

sample (shown in Figure F.1), which was the solvent used for purchased geosmin and MIB

solutions in these previous studies. For example, the typically commercial geosmin and MIB

standard solutions are at a concentration of 100 mg L–1 in methanol. This means the methanol

molar concentration is at least 40,000 times higher than that of geosmin or MIB. Although

geosmin and MIB molar absorption are approximately 5,000 times higher than methanol, the

much higher difference in their concentrations leads to the methanol absorbance being a

significant interference in the measurement for geosmin and MIB. In this work, geosmin and

MIB solutions prepared by dissolving 50 mg pure geosmin and MIB in 500 mL Milli-Q® water

were used to simplify the solution matrices.

Figure F.1 Absorption spectra of geosmin, MIB, and methanol (amplified by 1000 times)

(left y-axis), and absorption spectrum of caffeine (right y-axis), with typical emission

spectrum of MP UV lamp

11 Jo, C.H., Dietrich, A.M., Tanko, J.M., 2011. Simultaneous degradation of disinfection byproducts and earthy-musty odorants by the UV/H2O2 advanced oxidation process. Water Research, 45(8), 2507–2516. 12 Rosenfeldt, E.J., Melcher, B., Linden, K.G., 2005. UV and UV/H2O2 treatment of methylisoborneol (MIB) and geosmin in water. Journal of Water Supply: Research and Technology– AQUA, 54(7), 423–434. 13 Kutschera, K., Börnick, H., Worch, E., 2009. Photoinitiated oxidation of geosmin and 2-methylisoborneol by irradiation with 254 nm and 185 nm UV light. Water Research, 43(8), 2224–2232.

118

G. Sample Analysis

G.1 Geosmin and MIB

Geosmin and 2-methylisoborneol (MIB) were extracted and concentrated from aqueous

samples by headspace solid phase micro-extraction (HS-SPME) and quantified using gas

chromatography-mass spectrometry (GC-MS), according to the Standard Methods 6040D

(APHA et al., 2012)14. An internal standard, d3-Geosmin, at a concentration of 100 ng L–1 was

spiked in each sample. The internal standards were used to monitor retention time, relative

response, and quantify analytes in the sample. The analysis was carried out with a Varian® 3800

Gas Chromatograph with a Varian® Ion-trap Mass Spectrometer Detector, using electron impact

(EI) ionization and an autosampler.

Samples were collected in 23 mL amber vials with Teflon®-lined septa screw caps

(headspace-free), followed by the addition of 166 μL of 25g L–1 sodium azide (NaN3) (Reagent

Plus® grade, ≥99.5%) as a preservative. Sample preparation involved first adding by pipette 10

mL Milli-Q® water for a calibration standard or a blank, or 10 mL of a sample into a 20 mL

clear vial (Supelco, Bellefonte, PA) that contained 3.5 g of reagent grade sodium chloride

(NaCl). For the calibration standard, an appropriate volume of geosmin and MIB stock solution

at 100 μg L–1 was then spiked. Internal standard (d3-Geosmin) was then added to the sample by

dispensing 100 μL of a stock solution at 10 μg L–1 in methanol to achieve a 100 ng L–1

concentration in the sample. The vial was capped with a Teflon®-lined septum magnetic crimp

cap (Supelco, Bellefonte, PA), which is manufactured for use with the autosampler. The vial

was then placed into a sample tray, where the autosampler took the sample vial and delivered it

to the spinning box. The temperature of the spinning box was preset to 65°C ± 1°C at a rotation

speed of 500 rpm. The vial was placed in the spinning box for 5 minutes to dissolve the salt. The

needle (23 gauge) containing a 1 cm long SPME fiber (Supelco, Bellefonte, PA) was then

inserted into the vial through the septum, and the fiber was extended into the vial’s headspace

for exactly 30 minutes. At the end of the contact time, the fiber was then retracted back into the

SPME holder. The needle was then inserted directly into the GC-MS, the fiber was extended

14 APHA, AWWA, WEF, 2012. Standard Methods for the Examination of Water & Wastewater, 22nd edition. American Public Health Association, American Water Works Association, and Water Environment Federation, Washington, D.C., USA.

119

into the GC/MS injection port, and the GC-MS run was started. After five minutes of

desorption, the fiber was retracted back into the SPME holder, a new sample was put into the

spinning box, and a new sample extraction process was started.

The GC-MS operating conditions are shown in Table G.1. The quantitative ions (m/z) for

geosmin, MIB, and d3-Geosmin are 112, 95, and 115, with retention times of 11.1, 9.0, and 11.1

min, respectively.

Calibration and Method Detection Limit

An example for geosmin and MIB calibration is shown in Table G.2. and Figure G.1.

Method detection limits (MDLs) for geosmin and MIB were determined to be 1.7 and 9.0 ng

L–1, respectively, based on the 8 replications at spiked concentrations of 10 ng L–1 for geosmin,

and 50 ng L–1 for MIB (shown in Table G.2).

Table G.1 GC-MS operating conditions for geosmin and MIB analysis

Parameter Description Column VF-5MS capillary column (30 m × 0.25 mm ID, 0.25

µm film thickness) Carrier gas Helium at 1 mL min–1 at 25°C Injection method Temperature: 250°C

Desorbing time: 5 min Mode: Splitless for first 2 minutes, split after 2 minutes with split ratio of 50 Split Valve: Open after 2 min, flow at 50 mL min–1 Injection Volume: 1 µL at normal speed

Autosampler method Syringe: SPME Fiber Supelco Divinylbenzene/Carboxen/Polydimethylsiloxane (DVB/CAR/PDMS), df 50/30 μm, needle size 24 ga Agitator Temperature: 65.0°C Pre-incubation time: 5 min Extraction agitation speed: 400 rpm Extraction time: 30 min

GC temperature program Initial: start from 40°C, hold for 2 min Ramp: increase to 250°C at 15°C min–1 Equilibration: hold at 250°C for 7 min

MS conditions Scan Mode: SIS (Single Ion Selection) Ionization Type: EI Emission current: 30 µAmps Scan average: 3 microscans (0.89 s per scan) Multiplier Offset: 150 volts

Total run time 23.00 min per run

120

Table G.2 Typical calibration standards for geosmin and MIB MIB Geosmin

Concentration (ng L–1) Peak area d3-Geosmin

area Ratio Concentration (ng L–1) Peak area d3-Geosmin

area Ratio

0 0 2.81E+06 0.00E+00 0 159560 2.36E+06 6.77E-02 10 40888 3.18E+06 1.29E-02 30 563147 2.78E+06 2.02E-01 50 260845 3.51E+06 7.44E-02 50 672080 2.15E+06 3.12E-01

100 615533 3.46E+06 1.78E-01 80 993562 2.08E+06 4.77E-01 100 586764 3.54E+06 1.66E-01 100 1.20E+06 2.34E+06 5.14E-01 300 2.06E+06 3.60E+06 5.73E-01 300 4.28E+06 2.71E+06 1.58E+00 500 3.39E+06 3.43E+06 9.88E-01 500 4.89E+06 1.93E+06 2.53E+00 800 5.24E+06 3.15E+06 1.67E+00 700 8.72E+06 2.53E+06 3.44E+00 1000 7.51E+06 3.33E+06 2.26E+00 1000 6.92E+06 3.27E+06 2.12E+00 1200 8.89E+06 3.40E+06 2.62E+00

MDLs MDLs 50 256476 3.34E+06 7.67E-02 10 158478 1.31E+06 1.21E-01 50 230848 3.43E+06 6.73E-02 10 176778 1.46E+06 1.21E-01 50 241160 3.71E+06 6.51E-02 10 125828 1.07E+06 1.17E-01 50 254573 3.46E+06 7.35E-02 10 187364 1.64E+06 1.15E-01 50 258391 3.41E+06 7.58E-02 10 201311 1.74E+06 1.16E-01 50 239846 3.37E+06 7.13E-02 10 198185 1.74E+06 1.14E-01 50 218141 3.42E+06 6.38E-02 10 185367 1.64E+06 1.13E-01 50 277794 3.44E+06 8.07E-02 10 221789 1.85E+06 1.20E-01

Recovery 89% Recovery 124% MDL (ng L–1) 9.0 MDL (ng L–1) 1.7

Figure G.1 Example for MIB and geosmin calibration curves

(internal standard: 100 ng L–1)

G.2 Caffeine

Caffeine samples were analyzed using a Varian 3800 gas chromatography (equipped with a

DP 1701 column, 30 m × 0.25 mm × 0.25 µm) coupled with a Varian 4000 ion-trap mass

spectrometry (GC-MS) in the positive ion chemical ionization (CI) mode, based on the method

121

described by Verenitch et al. (2006)15. The quantitative ions (m/z) for caffeine and d3-caffeine

are 195, and 198, respectively, with retention time of 17.1 min for both.

Table G.3 GC-MS operating conditions for caffeine analysis

Parameter Description Column DP 1701 capillary column (30 m × 0.25 mm × 0.25 µm) Carrier gas Helium at 1 mL min–1 at 25°C Injection method Temperature: 250°C

Mode: Splitless for first 0.5 min, split after 0.5 min with split ratio of 20 Split Valve: Open after 0.5 min, flow at 20 mL min–1 Injection Volume: 2 µL at normal speed

GC temperature program 100°C for 4 min 25°C min–1 temperature ramp to 180°C 180°C for 5 min 10°C min–1 temperature ramp to 240°C 240°C for 4 min

MS conditions Filament delay: 12 min Scan Mode: full Ionization Type: CI, reagent: methanol Emission current: 10 µAmps Scan average: 3 microscans Multiplier Offset: 300 volts

Total run time 22.20 min per run

Samples were collected in 1 L pre-cleaned amber bottles with TeflonTM-lined caps, stored

in the refrigerator (4° C), extracted within 28 days of the sampling date, and analyzed within 7

days thereafter. Samples or Milli-Q® water (for calibration standards) with volumes of 500 mL

were spiked with 1 µg L–1 d3-caffeine (as the internal standard) by adding 0.5 mL of the internal

standard stock solution at a concentration of 1 mg L–1. Samples were then extracted using

Waters Oasis hydrophilic-lipophilic balance (HLB) solid phase extraction (SPE) cartridges

(WAT106202) placed on a vacuum manifold (Visiprep, Supelco Inc.). Two blank samples

consisting of 500 mL of the water matrix examined in the experiments were spiked with internal

standard and processed alongside samples. Each of extracted samples into the cartridges was

then eluted by 6 mL methanol/MTBE mixture (1:9 v./v.). After that, 1.5 mL of each elute was 15 Verenitch, S.S., Lowe, C.J., Mazumder, A., 2006. Determination of acidic drugs and caffeine in municipal wastewaters and receiving waters by gas chromatography-ion trap tandem mass spectrometry. Journal of Chromatography A, 1116(1–2), 193–203.

122

blown-down to dryness using nitrogen, reconstituted to a volume of 150 µL in chloroform, and

then loaded onto the autosampler of the GC-MS with the operating conditions summarized in

Table G.3.

Sample extraction

• Samples with volumes of at least 600 mL are collected in 1 L amber glass bottles with

TeflonTM-lined caps. Prior to extraction, samples are refrigerated (4°C) and stored in the

dark to avoid photo-decomposition.

• Using a volumetric flask, transfer 500 mL samples to clean 1 L bottles. For calibration

standards, transfer 500 mL Milli-Q® water.

• Spike 1 µg L–1 d3-caffeine to each sample by adding 0.5 mL of the 1 mg L–1 stock solution.

• Place the appropriate number of SPE cartridges on the Visiprep manifold.

• Condition the SPE cartridges on the manifold using the following procedure. Flow should

be set at approximately 1 drop every two seconds.

• Slowly aspirate approximately 3 mL MTBE through each SPE cartridge and do not allow

the cartridges to go dry.

• Slowly aspirate approximately 3 mL methanol through each SPE cartridge and do not allow

the cartridges to go dry.

• Slowly aspirate approximately 3 mL Milli-Q® water through each SPE cartridge and do not

allow the cartridges to go dry.

• Close the valve on each cartridge.

• Attach a pre-rinsed Teflon adaptor/Teflon tube to each SPE cartridge.

• Place the free end of each Teflon adaptor/Teflon tube in a separate bottle, making sure the

tube reaches the bottom of the sample bottle. Label each SPE cartridge.

• Open the valve on each cartridge and apply vacuum (approximately -10 to -15 inHg) to the

Visiprep manifold. Flow rates through the SPE cartridges should be approximately 5–10

mL min–1. Close the value when the bottle is empty. Make sure that cartridge is not dry.

• After all samples passed completely through the cartridges, rinse each sample bottle with 30

mL of Milli-Q® water. Rinses are allowed to flow through the cartridges at approximately

5–10 mL min–1. Close the value when the bottle is empty. Make sure that cartridge is not

dry.

• Wash the cartridges with 2 ml of 25% methanol in Milli-Q® water (v./v.).

123

• Vacuum dry each cartridge by applying vacuum (approximately -10 to -15 inHg) for 2

minutes. Once all the water has been aspirated, turn OFF the manifold vacuum and remove

the Teflon adaptor/Teflon tubes from the SPE cartridges.

• Place the Teflon adaptor/Teflon tubes on Kimwipes. Remove the SPE cartridges from the

Visiprep manifold and place them on clean Kimwipes. Make sure the cartridges are labelled

properly.

• Using cotton swabs and/or Kimwipes, dry the inside of the SPE cartridges.

• Remove the Visiprep manifold cover. Dry any excess water from the underside of the

manifold cover and place the cover on clean Kimwipes.

• Allocate one 15 mL polypropylene centrifuge tube for each SPE cartridge.

• Label each 15 mL centrifuge tube with the appropriate sample identification number and

place the tube in the proper slots of the Visiprep collection rack.

• Place the Visiprep collection rack in the vacuum manifold and reseat the vacuum manifold

cover with the SPE cartridges. Check to ensure that the manifold cover exhaust tubes are

aligned with the 15 mL polypropylene centrifuge tubes.

• Add 1 mL of methanol to each SPE cartridge. Turn ON the manifold vacuum and let

methanol slowly reach the bottom of the cartridge. Turn OFF vacuum and close the valves,

soak the cartridges with methanol for 2 minutes.

• Add 6 mL methanol/MTBE mixture (1:9 v./v.) to each SPE cartridge.

• Open manifold valves and elute by gravitational force until dry. Collect the eluent in the

corresponding centrifuge tube. NOTE: Some cartridges may require slight vacuum when

beginning the elution to initiate a constant flow of methanol.

• Once all solvent appears to be eluted, turn ON the manifold vacuum for approximately 2

minutes to aspirate remaining solvent in the cartridge.

• Turn OFF the manifold vacuum. Lift the manifold cover and remove the collection rack

containing the polypropylene centrifuge tubes with the sample extracts from the manifold.

• Using a calibrated pipettor, transfer 1.5 mL of each final extract to separate 1.8 mL clean

GC vials.

• Evaporate extracts in vials to dryness using a gentle stream of nitrogen (pressure < 2 psi).

Low heat at 40°C can be applied to expidite evaporation.

124

• Reconstitute each vial with 150 μL of chloroform and cap the vial with a TeflonTM lined

septum screw cap. Rinse the walls of the vial thoroughly by gently tumbling the vial.

Transfer the concentrated extract into a vial insert and recap.

• Store in freezer (-15°C) until required for GC-MS analysis.


An example for caffeine calibration is shown in Table G.4 and Figure G.2. Method

detection limit was determined to be 31 ng L–1, based on the 8 replications at a spiked

concentration of 200 ng L–1.

Figure G.2 Example for caffeine calibration curves (internal standard: 1 µg L–1)

Table G.4 Typical calibration standards for caffeine Concentration (µg L–1) Peak area d3-Caffeine area Ratio

0 75771 503662 0.150 0.2 1.27E+06 4.17E+06 0.303 0.4 2.27E+06 4.83E+06 0.470 1 1.84E+06 1.71E+06 1.074 5 1.20E+07 2.26E+06 5.292 10 4.78E+06 471955 10.124 10 2.64E+07 2.64E+06 10.004 10 4.34E+07 4.18E+06 10.385 30 1.42E+08 4.79E+06 29.522 50 6.05E+06 108358 55.815

MDLs 0.2 230625 790957 0.292 0.2 248069 787763 0.315 0.2 565086 1.90E+06 0.297 0.2 269516 887596 0.304 0.2 915212 3.16E+06 0.290 0.2 625972 2.24E+06 0.279 0.2 1.04E+06 3.46E+06 0.301 0.2 64686 213594 0.303

Recovery 101% MDL (ng L–1) 31

125

G.3 Trihalomethanes (THMs), Haloacetonitriles (HANs), Haloketones (HKs),

Chloropicrin (CP), and Trichloroethylene (TCE)

Trihalomethanes (THMs), haloacetonitriles (HANs), haloketones (HKs), chloropicrin (CP),

and trichloroethylene (TCE) were analyzed by liquid-liquid extraction and gas chromatography

with electron-capture detection (GC-ECD), according to USEPA Method 551.1 (USEPA,

1995)16. A 25 mL sample or Milli-Q® water for calibration, expect TCE, was spiked with ~50

μg L–1 1,2-dibromopropane (1,2-DBP, as the internal standard) in methanol and then extracted

with 4 mL of MTBE. For the TCE samples, 10 mL aliquot of each sample was spiked with ~10

μg L–1 1,2-DBP and extracted followed the same way as that for the other compounds. One μL

of the extract was then injected into a Hewlett Packard 5890 Series II Plus GC equipped with a

DB 5.625 fused silica capillary column (30 m × 0.25 mm ID, with 0.25 µm film thickness), and

subsequently separated and detected by an ECD. The operating conditions are shown in Table

G.5.

Table G.5 GC operating conditions for THMs, HANs, HKs, CP, and TCE

Parameter Description System HP 5890 Series II Plus Column DB 5.625 capillary column Injector temperature Splitless 200°C

15 sec purge activation time Detector temperature 300°C Temperature program 35°C for 10.0 min

4°C min–1 temperature ramp to 60°C 60°C for 1.0 min 20°C min–1 temperature ramp to 110°C 110°C for 4.5 min 30°C min–1 temperature ramp to 240°C 110°C for 2.0 min

Carrier gas Helium Flow rate 24.8 cm sec–1 at 150°C Total run time 27.6 min

16 United States Environmental Protection Agency (USEPA), 1995. Method 551.1. Determination of chlorination disinfection byproducts, chlorinated solvents, and halogenated pesticides/herbicides in drinking water by liquid-liquid extraction and gas chromatography with electron-capture detection, Revision, 1.0. Available from <http://www.epa.gov/sam/pdfs/EPA-551.1.pdf> (accessed 11.09.11).


126

Sample storage

• TCE samples collected in the bench-scale experiments and DBP samples collected in the

pilot-scale experiments were not required to store, since immediately extraction to the

MTBE phase after sample collection was carried out. Sample storage was conducted in the

full-scale tests. Samples were collected in 40 mL vials (headspace-free) preceded by the

addition of 0.8 g dry salt of the buffer. Chlorine quenching agent, hydrogen peroxide

(H2O2) solution, at a concentration of 100 mg mL–1 with a volume of 20 μL was added to

the sample immediately after the collection. H2O2 concentration in the samples was

approximately 50 mg L–1. Prepare the buffer by adding 1% dibasic sodium phosphate

(Na2HPO4) to 99% monobasic potassium phosphate (KH2PO4) by weight (example: 2 g

Na2HPO4 and 198 g KH2 PO4 to yield a total weight of 200 g). The phosphate buffer is used

to lower the sample matrix pH to 4.8 to 5.5 in order to inhibit base catalyzed degradation of

the haloacetonitriles.

• Store samples in the dark at 4ºC for up to 14 days.

• Before extracting samples, remove samples from refrigerator and place them at room

temperature for 30 min.

Sample extraction

• Transfer 25 mL of a sample (or Milli-Q® water for a calibration standard or a blank) into a

clean 40 mL vial. For the calibration standard, subsequently spike an appropriate volume of

a stock solution at 10 μg mL–1 (in methanol for THMs and TCE, or in acetone for HANs,

HKs, and CP).

• Add 24 µL of the internal standard (1,2-DBP) stock solution at 50 µg mL–1 in methanol.

• Add 2 tsp of sodium sulphate (Na2SO4) using scoop to increase extraction efficiency.

• Add 4 mL of MTBE extraction solvent using a bottle-top dispenser and cap the vial using a

cap with a Teflon®-lined silicon septa.

• Shake the sample vial for approximately 30 seconds and place it on counter on its side.

Repeat and complete for all samples, blanks and standards before proceeding.

• Place all the vials upright in a rack. Let samples stand for 20 minutes for phase separation.

• Remove 2 mL from the MTBE layer, without disturbing the water layer, using a Pasteur

pipette and place it in a 1.8 mL clean GC vial without headspace. The GC vial was

127

previously added with oven baked Na2SO4 with the amount that just covers the bottom of

the GC vial. Use a clean pipette for each sample.

• Place samples in the freezer. If not analyzing immediately, samples may be kept in the

freezer for up to 21 days. Before loading onto the GC, check the appearance of Na2SO4 at

the bottom of each GC vial. If the salt lumps, this means water is present in the sample. In

this case, transfer the top portion of the sample into a new GC vial with oven baked

Na2SO4.

Table G.6 Typical calibration standards for trichloroethylene Concentration (µg L–1) Peak area 1,2-DBP area Ratio

0 0 6.0804 0 6 1.434 5.5508 0.259 10 2.220 5.4167 0.410 20 3.8856 5.0337 0.772 30 5.7235 5.2746 1.085 50 9.4544 4.772 1.981 70 9.3315 3.5521 2.627 100 17.9739 4.0345 4.456 149 32.2459 5.2264 6.170 198 43.54 5.1045 8.530

MDLs 6 1.4345 5.5508 0.258 6 1.4615 5.2567 0.278 6 1.5296 5.4308 0.282 6 1.4266 5.3629 0.266 6 1.4026 5.3185 0.264 6 1.2492 5.0891 0.245 6 1.4199 5.4948 0.258 6 1.5341 5.4693 0.280

Recovery 113% MDL (µg L–1) 1.1

Table G.7 Typical calibration standards for THMs, HANs, HKs, and CP Conc.

(µg L–1) Peak area

1,2-DBP TCM BDCM CDBM TBM TCAN DCAN DCP CP BCAN TCP DBAN 0 47.976 2.693 0 0 0 0 0 0 0 0 n.a. n.a. 2 46.809 3.002 3.951 4.218 2.768 2.342 12.156 1.785 17.353 17.868 20.956 21.467 4 51.616 4.649 8.543 8.977 5.460 5.614 26.566 4.012 17.337 17.860 20.946 21.446 6 51.843 6.087 14.262 14.724 8.420 10.961 41.971 7.122 17.323 17.855 20.937 21.435 8 50.924 7.668 21.803 21.784 11.729 14.079 52.950 9.862 17.328 17.866 20.952 21.447 10 50.237 9.226 29.167 29.020 15.152 19.022 64.208 12.496 17.322 17.861 20.945 21.437 20 51.344 16.330 57.109 58.870 29.460 47.274 146.325 29.592 17.303 17.854 20.939 21.425 30 55.145 23.078 90.602 90.236 42.685 77.828 212.409 47.322 17.297 17.857 20.943 21.424

MDLs 2 47.046 3.359 4.175 4.209 2.639 2.497 12.126 1.737 17.35 17.861 20.951 21.454 2 47.151 3.163 4.488 4.581 2.844 2.590 12.473 2.002 17.337 17.852 20.938 21.441 2 45.857 3.035 4.333 4.467 2.852 2.493 12.419 1.9004 17.339 17.854 20.938 21.444 2 48.282 3.185 4.654 4.750 2.954 2.754 12.836 2.044 17.343 17.856 20.939 21.448 2 42.078 3.401 4.326 4.371 2.706 2.891 13.054 1.994 17.353 17.868 20.954 21.463 2 47.311 3.196 4.547 4.712 2.925 2.632 12.881 1.985 17.341 17.854 20.938 21.442 2 54.280 3.498 5.135 5.272 3.252 2.959 14.049 2.296 17.348 17.867 20.958 21.458 2 48.620 3.509 5.002 5.131 3.190 2.851 13.758 2.205 17.335 17.852 20.932 21.439

Recovery 95% 91% 92% 106% 81% 105% 89% 82% 95% 82% 77% MDL

(µg L–1) 1.2 0.3 0.3 0.3 0.4 0.4 0.4 0.5 0.5 0.4 0.5

128


Examples for the calibration of the above-mentioned compounds are shown in Tables G.6

and G.7, and illustrated in Figure G.3. Method detection limits are also shown in these two

tables.

129

Figure G.3 Example for calibration curves for TCE, THMs, HANs, HKs, and CP

(internal standards: 10 µg L–1 for TCE, and 50 µg L–1 for others)

G.4 Haloacetic Acids (HAAs)

Nine haloacetic acids (HAA9) were analyzed by liquid-liquid extraction and gas

chromatography with electron-capture detection (GC-ECD), according to Standard Method

130

6251 B (APHA et al., 2005)17. A 25 mL sample (or Milli-Q® water for calibration) spiked with

1600 μg L–1 2,3,4,5-tetrafluorobenzoic acid (TFBA, as the internal standard) in MTBE was

acidified with 2.8 mL concentrated sulfuric acid (95–98% v./v.) and extracted with 5 mL of

MTBE. A Hewlett Packard 5890 Series II Plus gas chromatography (GC) equipped with a DB

5.625 fused silica capillary column (30 m × 0.25 mm ID, with 0.25 µm film thickness) was then

used to separate the analytes, followed by a electron-capture detector (ECD). The GC operating

conditions are shown in Table G.8.

Table G.8 GC operating conditions for HAAs

Parameter Description System HP 5890 Series II Plus Column DB 5.625 capillary column Injector temperature 200°C Detector temperature 300°C Temperature program 35°C for 10.0 min

2.5°C min–1 temperature ramp to 65°C 10°C min–1 temperature ramp to 85°C 20°C min–1 temperature ramp to 205°C 205°C for 7.0 min

Carrier gas Helium 30cm s–1 Makeup gas: 5% CH4 + 95% Ar at a flowrate of 23.1 mL min–1 Total run time 37 min

Sample storage

• Collect samples in 40 mL amber vials quenched with 8 mg sodium sulfite powder. Ensure

that samples are headspace free.

• Store dechlorinated sample at 4°C, but for no more than 9 days. Sample extracts can be held

in a freezer at -11°C for 21 days.

• Before extracting samples, remove samples from refrigerator and place them at room

temperature for 30 min.

Sample extraction

17 APHA, AWWA, WEF, 2005. Standard Methods for the Examination of Water & Wastewater, 21st edition. American Public Health Association, American Water Works Association, and Water Environment Federation, Washington, D.C., USA.

131

• Transfer 25 mL of a sample (or Milli-Q® water for a calibration standard or a blank) into a

clean 40 mL vial. For the calibration standard, subsequently spike an appropriate volume of

a stock solution at 20 μg mL–1 in MTBE.

• Add 20 µL of the internal standard (TFBA) stock solution at 2000 µg mL–1 in MTBE.

• Add 2 tsp of sodium sulphate (Na2SO4) using scoop to increase extraction efficiency.

• Use bottle-top dispensers to add 2.8 mL concentrated sulphuric acid and 5 mL of MTBE

extraction solvent in turn, and cap the vial using a cap with a Teflon®-lined silicon septa.

• Shake the sample vial for approximately 30 seconds and place it in a rack for 20 minutes for

phase separation. Repeat and complete for all samples, blanks and standards before

proceeding.

• Transfer ~1.5 mL of the extract from MTBE layer to a 1.8 mL clean GC vial, without

disturbing the water layer, using a Pasteur pipette. The GC vial was previously added with

oven baked Na2SO4 with the amount that just covers the bottom of the GC vial. Use a clean

pipette for each sample.

• Place samples in the freezer. If not analyzing immediately, samples may be kept in the

freezer for up to 21 days. Sample need to be derivatized overnight using 150 µL

diazomethane before loading onto the GC. When loading samples on the GC, check the

appearance of Na2SO4 at the bottom of each GC vial. If the salt lumps, this means water is

present in the sample. In this case, transfer the top portion of the sample into a new GC vial

with oven baked Na2SO4.

Diazomethane Production

• Set up MNNG diazomethane generation apparatus (shown in Figure G.4) in a beaker filled

with crushed ice and water.

• Add ~0.5 cm of Diazald (N-methyl-N-nitroso-p-toluenesulfonamide) to the inner tube of the

apparatus using the large end of a Pasteur pipette.

• Add methanol to the inner tube until the bulb of the inner tube is half full. Secure cap and

septum

• Add 2.6 mL of MTBE to the outer tube of the apparatus and place the apparatus in the ice

bath.

132

• Place O-ring in glass joint, position inside tube firmly on top and secure clamp. Ensure that

vapour exit hole is located on opposite side of clamp and rest clamp on spout of beaker. The

seal must be very tight to ensure maximum CH2N2 generation and recovery.

• Add 600 µL of 20% NaOH solution (100 g of NaOH in 500 mL Milli-Q® water) dropwise

to inner tube with gas tight syringe (1 drop per 5 seconds). When NaOH is initially being

added, there is a slight delay before the Diazald reacts violently so be sure to add dropwise.

Aim drops straight down into Diazald in bottom of inner tube, avoiding tube surface and

vapour exit hole. Leave syringe in place after all NaOH has been added. Removal of the

syringe will leave a hole in the septum from where CH2N2 may escape.

• Allow CH2N2 to form for 30–45 min in ice bath. MTBE will become yellow when CH2N2 is

formed.

• Transfer CH2N2 in MTBE to 15 mL vials using specially flamed Pasteur pipette and store

vials in explosion-proof freezer. The solution can be used within 2–4 weeks if possible.

• Rinse inner tube and NaOH syringe several times with Milli-Q® water.

• Rinse inner and outer tube with methanol and MTBE until glassware is clean. Put glassware

in oven at ~100°C until dry.

Figure G.4 MNNG diazomethane generation apparatus


An example for the calibration of HAAs is shown in Table G.9 and Figure G.5. Method

detection limits are also shown in the table.

133

Table G.9 Typical calibration standards for HAAs Concentration

(µg L–1) Peak area

TFBA MCAA MBAA DCAA TCAA BCAA DBAA BDCAA CDBAA TBAA 0 35.8296 0 0 0 0 0 0 0 0 0 2 36.5602 0.2383 2.0356 3.6594 7.98 5.7279 5.6102 1.587 0.7932 0.426 4 36.2235 0.488 3.8329 7.0241 14.1269 10.7051 10.7964 3.4509 1.2331 0.8273 6 35.4795 0.6946 5.6331 9.7997 20.5488 15.7549 16.1369 6.6554 2.556 1.33 8 35.2702 0.9442 7.4464 12.7828 28.4983 21.6905 22.5879 11.2472 4.7431 2.0227 10 36.6199 1.3107 10.017 16.4279 39.962 28.8607 30.3848 14.6404 7.5273 3.4555 20 35.4867 2.657 19.1228 29.9972 70.7059 54.2047 57.0819 39.5848 19.3598 8.7111 40 35.6578 5.1763 35.8799 54.7879 102.1244 53.0131 18.8538

MDLs 2 35.4688 0.2424 2.4333 3.9949 9.563 6.7062 6.7666 3.2167 1.469 0.2951 2 34.9655 0.2325 2.4021 4.0265 9.3444 6.5562 6.5297 3.1634 1.4594 0.3062 2 35.1744 0.2296 2.4171 3.9915 10.0212 7.4291 6.5622 3.2067 1.519 0.4432 2 35.576 0.2413 2.3725 4.0296 9.4425 6.6518 6.5133 3.2298 1.5012 0.4943 2 36.9077 0.2492 2.4766 4.1631 9.8079 6.7175 6.7302 4.2271 1.6079 0.4639 2 37.0982 0.2385 2.5654 4.1934 9.9635 6.7559 6.7187 4.2566 1.624 0.4707 2 36.6272 0.2408 2.4111 4.1871 9.7929 6.8322 6.7494 4.2643 1.6394 0.4965 2 37.8648 0.2315 2.2217 3.8122 8.5484 6.1489 5.9556 2.6197 0.9235 0.5516

Recovery 96% 124% 123% 127% 121% 113% 131% 119% 75% MDL (µg L–1) 0.2 0.4 0.3 0.5 0.5 0.4 1.3 1.2 0.9

134

Figure G.5 Example for calibration curves for HAAs (internal standards: 1600 μg L–1)

G.5 Chlorate

Chlorate (ClO3–) samples at expected concentrations higher than 50 μg L–1 were analyzed

using an ion chromatograph (IC, Model: Dionex ICS-5000+ analytical system, Thermo

135

Scientific) based on USEPA Method 300.1 (USEPA, 1997) 18 . The calibration of HAAs is

shown in Table G.10 and Figure G.6. Method detection limit was 7.2 μg L–1.

Table G.10 Calibration standards for chlorate NaClO3 concentration

(μg L–1) ClO3

– concentration (μg L–1)

Retention time (min) Peak area Regression

(μg L–1) 0 0 0 15.61 20 15.7 10.813 0.0075 32.23 50 39.2 10.92 0.0155 49.95 100 78.4 10.957 0.0335 89.82 200 156.8 10.983 0.0676 165.36 500 392.0 10.977 0.1696 391.30

1000 783.9 10.973 0.3278 741.73 2000 1567.8 10.943 0.6798 1521.45 5000 3919.6 10.887 1.7743 3945.89 MDL

50 39.2 11.003 0.0115 27.20 50 39.2 11.003 0.0117 27.66 50 39.2 11 0.0116 27.43 50 39.2 11.013 0.0139 32.74 50 39.2 11.007 0.0132 31.12 50 39.2 11.007 0.013 30.66 50 39.2 11.003 0.0145 34.12 50 39.2 11.017 0.013 30.66 50 39.2 11.017 0.0137 32.28

Recovery 78% MDL (µg L–1) 7.2

Figure G.6 Calibration curve for chlorate

H. Quality Assurance / Quality Control (QA/QC) Various QA/QC measures were undertaken to ensure analytical precision and accuracy for

all analytes, including:

• All chemicals used were analytical or high grade.

18 United States Environmental Protection Agency (USEPA), 1997. Determination of inorganic anions in drinking water by ion chromatography. Available from <http://water.epa.gov/scitech/drinkingwater/labcert/upload/met300.pdf> (accessed 09.09.14).

136

• All glassware to come in contact with samples were cleaned, by first rinsing them

thoroughly three times with hot tap water, twice with demineralized water, and twice with

distilled water, and drying them in the oven overnight at 250°C.

• The glassware to come in contact with organic solvents were cleaned, by first rinsing them

thoroughly three times with acetone, three times with MTBE, three times with methanol,

and three times with Milli-Q® water, followed by heating them in the oven overnight at

250°C. Volumetric flasks were not dried in the oven.

• Seal and store glassware in clean drawers free of all potential contamination. The lids of

glass containers were also washed with hot tap water three times and with distilled water

twice, and subsequently dried inverted at the room temperature.

• The bottles used in disinfection by-product formation potential test were prepared chlorine-

demand-free, by washing them with hot tap water and soaking them in a concentrated

sodium hypochlorite solution (~1000 mg L–1) for at least 24 h. The bottles were thereafter

rinsed thoroughly with distilled water three times and dried in the oven.

• Internal standards were used for analyses of geosmin, MIB, caffeine, and organic DBPs.

• Calibration curves for all analytes were made, by spiking known concentrations of the

analytes to Milli-Q® water. Calibration standards were prepared and run with each set of

samples. At least 7 standards were prepared, with the range of concentrations covering the

expected concentrations in the samples. The typical calibration curves are shown in

Appendix G.

• Along with each set of samples, blanks were prepared using Milli-Q® water and spiked with

internal standards to verify the interference was absent.

• Check standards at certain concentrations in the range of expected sample concentrations

were prepared at the same time and in the same way as the samples, and were analysed

every 10 samples. These check standards were then plotted on quality control charts. Figure

H.1 is an example shown the quality control charts of DBPs. The quality control charts are

used to validate and assess the accuracy of instrumental analysis. Recalibration is required

if any of the following occurs: (where M is the historical mean and SD is the historical

standard deviation). The historical mean and historical standard deviation are based on 8

standards prepared individually and analyzed consecutively.

(1) 2 consecutive measurements are outside the control limits of M ± 3 SD

137

(2) 3 out of 4 consecutive measurements are outside the warning limits of M ± 2 SD

(3) 5 out of 6 consecutive measurements are outside of M ± SD

(4) 5 out of 6 consecutive measurements follow an increasing or decreasing trend

(5) 7 consecutive measurements are above M or 7 consecutive measurements are below M

• Method detection limit (MDL) of each analyte was determined by multiplying the standard

deviation of 8 replicates near the anticipated detection limit by the Student t value (i.e.,

2.998).

• All the experiments were conducted at least in duplicate.

Quality control charts for organic DBPs are shown below.

138

139

Figure H.1 Quality control charts for organic DBPs. Lines above and below the mean (M)

represent ±1, 2 and 3 standard deviations (SD) of the samples from the mean.

140

I. Experimental Data Table I.1 Chlorine and H2O2 photolysis measured in the bench-scale trichloroethylene study (Chapter 3)

UV/chlorine pH 5 UV/chlorine pH 10 UV/chlorine pH 7.5

Fluence (J m–2)

ln(C/C0) for Cl2

Std dev Fluence

(J m–2) ln(C/C0) for Cl2

Std dev Fluence

(J m–2) ln(C/C0) for Cl2

Std dev

0 0 0 0 0 0 0 0 0 3848 -0.04 0.0056 3835 -0.22 0.008 3836 -0.10 0.020 7696 -0.11 0.0060 7669 -0.44 0.010 7672 -0.25 0.023

11543 -0.15 0.0062 11504 -0.65 0.015 9590 -0.31 0.014 15391 -0.21 0.0133 15338 -0.89 0.011 11507 -0.38 0.029 19239 -0.30 0.0072 19173 -1.14 0.041 15343 -0.52 0.025

19179 -0.63 0.009 UV/H2O2 pH 5 UV/H2O2 pH 10 UV/H2O2 pH 7.5

Fluence (J m–2)

ln(C/C0) for H2O2

Std dev Fluence

(J m–2) ln(C/C0) for H2O2

Std dev Fluence

(J m–2) ln(C/C0) for H2O2

Std dev

0 0 0 0 0.000 0 0 0.0000 0 1536 -0.0082 0.0051 3839 -0.025 0.0033 1537 -0.0070 0.0016 3073 -0.0139 0.0039 7679 -0.046 0.0034 3074 -0.0151 0.0026 4609 -0.0214 0.0040 11518 -0.065 0.0041 4612 -0.0236 0.0026 6146 -0.0279 0.0032 15358 -0.086 0.0021 6149 -0.0322 0.0048 7682 -0.0397 0.0025 19197 -0.112 0.0012 7686 -0.0366 0.0032

13444 -0.0608 0.0045 13450 -0.0615 0.0028 19206 -0.0925 0.0034 19215 -0.0935 0.0007

141

Table I.2 Trichloroethylene decay by the bench-scale UV/chlorine and UV/H2O2 AOPs

UV only pH 5 UV only pH 10 UV only pH 7.5 Fluence

(mJ cm–2) ln(C/C0) for TCE Std dev

Fluence (mJ cm–2)

ln(C/C0) for TCE Std dev

Fluence (mJ cm–2)


0 0 0 0 0 0 0 0 0 385 -0.30 0.057 385 -0.21 0.045 385 -0.26 0.097 770 -0.56 0.053 770 -0.43 0.083 770 -0.49 0.053 1155 -0.88 0.071 1155 -0.64 0.028 1154 -0.70 0.063 1540 -1.13 0.022 1540 -0.82 0.045 1539 -0.95 0.023 1925 -1.39 0.047 1925 -1.00 0.053 1924 -1.17 0.023

UV/chlorine pH 5 UV/chlorine pH 10 UV/chlorine pH 7.5

Fluence (mJ cm–2)


Fluence (mJ cm–2)


Fluence (mJ cm–2)


0 0 0 0 0.00 0 0 -0.02 0.034 51 -0.42 0.062 383 -0.30 0.056 384 -0.32 0.027 103 -0.75 0.066 767 -0.54 0.075 767 -0.63 0.056 154 -1.17 0.084 1150 -0.84 0.078 959 -0.77 0.031 205 -1.62 0.052 1534 -1.16 0.059 1151 -0.92 0.083 257 -2.10 0.034 1917 -1.39 0.032 1534 -1.31 0.045

1918 -1.55 0.036 UV/H2O2 pH 5 UV/H2O2 pH 10 UV/H2O2 pH 7.5

Fluence (mJ cm–2)


Fluence (mJ cm–2)


Fluence (mJ cm–2)


0 0.00 0 0 0 0 0 0 0 154 -0.49 0.029 384 -0.44 0.037 154 -0.54 0.158 307 -1.03 0.015 768 -0.78 0.060 307 -1.01 0.120 461 -1.52 0.092 1152 -1.18 0.050 461 -1.52 0.084 615 -2.15 0.064 1536 -1.52 0.066 615 -2.06 0.079 768 -2.64 0.079 1920 -1.87 0.027 769 -2.46 0.055

142

Table I.3 Raw data of chlorine, H2O2, geosmin, MIB, and caffeine destruction in the full- and pilot-scale studies Process UV alone UV/chlorine UV/H2O2 Dose 1800–2100 mJ cm–2 2 mg L–1 6 mg L–1 10 mg L–1 1.0 mg L–1 2.9 mg L–1 4.8 mg L–1 pH 6.5 7.5 8.5 6.5 7.5 8.5 6.5 7.5 8.5 6.5 7.5 8.5 6.5 7.5 8.5 6.5 7.5 8.5 6.5 7.5 8.5 Cornwall full-scale in late May, 2013 Cl2/H2O2 initial (mg L–1) 2.0 6.2 10.4 0.89 2.81 4.75 Cl2/H2O2 final (mg L–1) 1.2 0.87 0.40 4.0 3.4 1.9 6.4 5.5 4.4 0.88 0.67 0.78 2.58 2.57 2.48 4.45 4.57 5.05 Geosmin initial (ng L–1) 352 Geosmin final (ng L–1) 260 273 277 48 118 155 30 89 141 13 52 93 114 144 211 59 84 123 35 56 80 MIB initial (ng L–1) 406 MIB final (ng L–1) 357 367 363 88 185 241 65 144 210 35 83 162 180 255 320 127 148 195 71 85 146 Caffeine initial (µg L–1) 22.09 Caffeine final (µg L–1) 20.21 19.29 18.53 6.43 9.66 11.96 2.88 7.15 9.94 1.66 6.21 8.19 12.49 14.24 16.76 7.38 10.23 11.98 5.04 6.09 7.91 Cornwall full-scale in early September, 2013 Cl2/H2O2 initial (mg L–1) 2.0 6.3 10.7 Cl2/H2O2 final (mg L–1) 1.0 0.70 0.34 3.5 2.1 1.5 6.6 5.3 4.7 Geosmin initial (ng L–1) 18 Geosmin final (ng L–1) 16 15 15 5 7 11 2 6 10 0 5 9 Cornwall full-scale in early April, 2014 Cl2/H2O2 initial (mg L–1) 2.2 6.0 9.8 0.91 2.76 5.07 Cl2/H2O2 final (mg L–1) 1.3 1.1 0.73 4.2 3.5 2.4 6.9 5.4 4.9 0.87 1.15 0.83 2.65 2.81 2.45 4.29 4.90 5.05 Rayox® pilot-scale Cl2/H2O2 initial (mg L–1) 2.0 2.1 2.2 6.5 6.4 6.5 10.8 10.8 11.1 0.87 0.91 0.92 2.56 2.74 2.77 4.52 4.36 4.65 Cl2/H2O2 final (mg L–1) 0.88 0.53 0.36 3.8 2.6 2.0 6.8 5.3 4.4 0.84 0.87 0.90 2.48 2.65 2.62 4.33 4.13 4.42 Caffeine initial (µg L–1) 18.35 Caffeine final (µg L–1) 15.78 15.95 15.92 5.65 8.53 10.46 2.13 5.78 8.26 0.99 4.96 6.61 10.70 12.00 12.69 7.20 7.63 9.16 4.39 5.19 6.69

143

Table I.4 Raw data of organic DBP formation (µg L–1) in the full-scale test in late May, 2013 MCAA MBAA DCAA TCAA BCAA DBAA BDCAA CDBAA TBAA TCAN DCAN BCAN DBAN AOX Initial 1.3 0 7.4 3.3 1.4 0 1.8 0 0 0 1.2 1.2 0 83 pH 7.5 Cl2 10 mg L–1 1.0 0 6.3 3.2 1.3 0 1.8 0 0 0 2.4 1.5 0 79 pH 7.5 Cl2 6 mg L–1 0.9 0 7.3 3.6 1.4 0 1.9 0 0 0 2.0 1.6 0 71 pH 7.5 Cl2 2 mg L–1 1.2 0 6.7 3.3 1.3 0 1.7 0 0 0 1.7 1.6 0 71 pH 7.5 UV 1.1 0 6.6 3.2 1.1 0 1.1 0 0 0 1.3 1.3 0 65 pH 7.5 UV/Cl2 10 mg L–1 1.7 0 9.3 3.9 1.3 0 1.4 0 0 0 2.9 3.2 0 77 pH 7.5 UV/Cl2 6 mg L–1 1.6 0 8.8 3.7 1.2 0 1.2 0 0 0 2.7 3.3 0 74 pH 7.5 UV/Cl2 2 mg L–1 1.4 0 7.7 3.4 1.1 0 1.3 0 0 0 2.3 2.7 0 76 pH 7.5 UV/H2O2 4.8 mg L–1 1.3 0 7.1 3.8 1.1 0 1.5 0 0 0 1.4 1.5 0 63 pH 7.5 UV/H2O2 2.9 mg L–1 1.8 0 9.6 3.8 1.3 0 0.7 0 0 0 1.4 1.5 0 69 pH 7.5 UV/H2O2 1.0 mg L–1 1.7 0 8.4 3.7 1.3 0 0.8 0 0 0 1.6 1.5 0 75 pH 6.5 Cl2 10 mg L–1 0.5 0 8.8 3.6 1.3 0 1.4 0 0 0 2.3 1.7 0 78 pH 6.5 Cl2 6 mg L–1 0.6 0 7.5 3.3 1.3 0 1.4 0 0 0 2.0 1.6 0 75 pH 6.5 Cl2 2 mg L–1 0.4 0 7.0 3.1 1.2 0 1.4 0 0 0 1.6 1.5 0 72 pH 6.5 UV 0.6 0 7.4 3.0 1.0 0 0.9 0 0 0 1.6 1.4 0 80 pH 6.5 UV/Cl2 10 mg L–1 2.0 0 11.7 4.0 1.1 0 1.2 0 0 0 4.4 5.3 0 89 pH 6.5 UV/Cl2 6 mg L–1 1.4 0 10.3 3.7 1.1 0 0.9 0 0 0 3.6 4.2 0 85 pH 6.5 UV/Cl2 2 mg L–1 1.4 0 10.0 3.5 1.1 0 1.0 0 0 0 2.9 3.2 0 82 pH 6.5 UV/H2O2 4.8 mg L–1 1.7 0 8.1 3.7 1.3 0 1.6 0 0 0 1.6 1.5 0 69 pH 6.5 UV/H2O2 2.9 mg L–1 1.1 0 8.0 3.5 1.1 0 1.3 0 0 0 1.5 1.5 0 69 pH 6.5 UV/H2O2 1.0 mg L–1 1.0 0 8.4 3.3 1.1 0 1.1 0 0 0 1.6 1.5 0 81 pH 8.5 Cl2 10 mg L–1 0.7 0 6.3 3.8 1.3 0 1.5 0 0 0 2.3 1.6 0 62 pH 8.5 Cl2 6 mg L–1 0.9 0 5.7 3.3 1.3 0 1.4 0 0 0 2.0 1.5 0 62 pH 8.5 Cl2 2 mg L–1 1.1 0 5.6 3.2 1.3 0 1.5 0 0 0 1.6 1.5 0 64 pH 8.5 UV 1.1 0 5.9 3.0 1.1 0 1.0 0 0 0 1.6 1.4 0 69 pH 8.5 UV/Cl2 10 mg L–1 1.4 0 6.4 3.3 1.2 0 1.2 0 0 0 2.1 2.4 0 66 pH 8.5 UV/Cl2 6 mg L–1 1.3 0 6.5 3.5 1.2 0 0.9 0 0 0 1.9 2.3 0 65 pH 8.5 UV/Cl2 2 mg L–1 1.5 0 7.1 3.1 1.1 0 0.5 0 0 0 1.3 1.8 0 70 pH 8.5 UV/H2O2 4.8 mg L–1 1.4 0 6.4 3.6 1.1 0 1.3 0 0 0 1.3 1.3 0 56 pH 8.5 UV/H2O2 2.9 mg L–1 1.3 0 6.2 3.6 1.1 0 1.2 0 0 0 1.4 1.4 0 62 pH 8.5 UV/H2O2 1.0 mg L–1 1.1 0 5.9 3.4 1.1 0 1.1 0 0 0 1.6 1.5 0 64 24 h formation potential pH 7.5 Initial 2.1 2.6 16.5 11.1 3.3 1.9 5.1 0 0 0 1.9 2.3 0 265 pH 7.5 H2O2 4.8 mg L–1 2.1 2.0 13.2 8.6 2.4 1.5 3.8 0 0 0 1.6 2.0 0 242 pH 7.5 UV 4.3 2.8 19.2 10.9 2.9 2.2 4.3 0 0 0 2.4 5.0 0 289 pH 7.5 UV/Cl2 10 mg L–1 3.3 3.2 29.5 13.6 3.5 2.6 6.4 0 0 0 2.7 7.9 0 305 pH 7.5 UV/H2O2 4.8 mg L–1 4.3 3.0 32.2 15.6 3.1 2.8 6.5 0 0 0 2.6 8.5 0 303 pH 6.5 Initial 2.0 2.9 15.8 12.4 3.7 1.8 5.7 0 0 0 2.0 2.3 0 278 pH 6.5 H2O2 4.8 mg L–1 1.9 1.8 11.6 6.3 2.4 1.3 3.4 0 0 0 1.5 1.6 0 214 pH 6.5 UV 5.0 3.5 21.9 12.7 3.1 2.1 6.2 0 0 0 3.2 4.2 0 284 pH 6.5 UV/Cl2 10 mg L–1 4.2 3.4 46.2 20.4 4.2 2.7 10.0 0 0 0 4.3 11.0 0 364 pH 6.5 UV/H2O2 4.8 mg L–1 3.5 3.2 34.4 17.5 3.8 2.6 9.4 0 0 0 3.4 8.4 0 319 pH 8.5 Initial 2.1 3.1 15.2 10.3 3.4 2.4 4.7 0 0 0 1.7 2.2 0 287 pH 8.5 H2O2 4.8 mg L–1 1.9 2.3 14.2 10.4 2.6 1.8 3.4 0 0 0 1.7 2.0 0 260 pH 8.5 UV 4.9 3.6 20.9 11.2 3.0 3.4 4.0 0 0 0 2.0 4.2 0 286 pH 8.5 UV/Cl2 10 mg L–1 4.1 3.9 26.2 13.5 3.6 3.6 6.6 0 0 0 2.0 4.9 0 287 pH 8.5 UV/H2O2 4.8 mg L–1 4.4 3.3 27.3 16.8 3.5 3.7 6.5 0 0 0 2.2 6.2 0 335

144

Table I.5 Raw data of organic DBP formation (µg L–1) in the full-scale test in early April, 2014 TCM BDCM CDBM TBM MCAA MBAA DCAA TCAA BCAA DBAA BDCAA CDBAA TBAA TCAN DCAN BCAN DBAN AOX Initial 9.2 6.5 2.0 0 0 0 4.1 2.8 1.6 0.4 0.9 0 0 0 0.6 0.7 0 46 pH 7.5 Cl2 10 mg L–1 9.2 6.6 2.0 0 0 0 4.7 3.0 1.6 0.4 1.2 0 0 0 0.6 0.7 0 41 pH 6.5 Cl2 10 mg L–1 8.7 6.8 2.2 0 0 0 5.0 3.0 1.6 0.4 0.9 0 0 0 0.7 0.8 0 44 pH 8.5 Cl2 10 mg L–1 10.1 5.1 2.1 0 0 0 4.0 2.7 1.5 0.4 1.0 0 0 0 0.6 0.8 0 46 pH 7.5 Cl2 6 mg L–1 9.2 6.2 1.9 0 0 0 4.8 2.8 1.6 0.4 0.7 0 0 0 0.5 0.7 0 46 pH 6.5 Cl2 6 mg L–1 8.7 6.6 2.0 0 0 0 5.2 2.9 1.6 0.4 1.4 0 0 0 0.6 0.7 0 41 pH 8.5 Cl2 6 mg L–1 10.2 6.7 2.1 0 0 0 4.2 2.9 1.6 0.4 1.4 0 0 0 0.5 0.8 0 43 pH 7.5 Cl2 2 mg L–1 9.3 6.3 2.0 0 0 0 4.4 2.9 1.5 0.4 1.4 0 0 0 0.7 0.8 0 47 pH 6.5 Cl2 2 mg L–1 8.8 6.5 2.1 0 0 0 5.5 2.9 1.6 0.4 1.5 0 0 0 0.6 0.7 0 50 pH 8.5 Cl2 2 mg L–1 10.4 6.8 2.1 0 0 0 4.1 2.8 1.6 0.4 1.2 0 0 0 0.6 0.7 0 45 pH 7.5 UV 9.2 6.7 1.4 0 0 0 5.3 3.4 1.5 0.3 1.0 0 0 0 0.8 1.0 0 45 pH 6.5 UV 7.8 6.6 1.5 0 0 0 5.5 3.3 1.4 0.3 0.9 0 0 0 0.9 1.0 0 45 pH 8.5 UV 9.2 6.5 1.5 0 0 0 4.5 3.0 1.4 0.3 1.0 0 0 0 0.7 0.9 0 39 pH 7.5 UV/Cl2 10 mg L–1 8.9 6.5 1.5 0 0 0 6.1 3.6 1.5 0.3 1.4 0 0 0 1.2 1.5 0 50 pH 6.5 UV/Cl2 10 mg L–1 8.1 6.1 1.5 0 0 0 7.6 3.9 1.4 0.3 1.2 0 0 0 1.7 2.4 0 57 pH 8.5 UV/Cl2 10 mg L–1 8.8 4.6 1.4 0 0 0 4.6 3.2 1.5 0.3 1.2 0 0 0 0.7 1.0 0 33 pH 7.5 UV/Cl2 6 mg L–1 8.6 6.3 1.4 0 0 0 6.5 3.6 1.5 0.3 1.5 0 0 0 1.1 1.4 0 48 pH 6.5 UV/Cl2 6 mg L–1 7.3 5.9 1.3 0 0 0 7.6 3.9 1.6 0.3 1.4 0 0 0 1.5 2.0 0 52 pH 8.5 UV/Cl2 6 mg L–1 9.0 5.8 1.5 0 0 0 5.1 3.4 1.6 0.3 1.1 0 0 0 0.7 0.9 0 39 pH 7.5 UV/Cl2 2 mg L–1 8.4 6.1 1.4 0 0 0 5.6 3.3 1.5 0.3 1.0 0 0 0 1.0 1.2 0 45 pH 6.5 UV/Cl2 2 mg L–1 7.8 6.1 1.4 0 0 0 6.3 3.4 1.5 0.3 0.8 0 0 0 1.2 1.4 0 49 pH 8.5 UV/Cl2 2 mg L–1 9.2 6.1 1.5 0 0 0 4.5 3.0 1.5 0.3 0.9 0 0 0 0.7 0.9 0 37 pH 7.5 UV/H2O2 4.8 mg L–1 8.1 5.8 1.3 0 0 0 4.8 2.9 1.4 0.3 1.0 0 0 0 0.7 0.7 0 38 pH 6.5 UV/H2O2 4.8 mg L–1 7.9 5.9 1.4 0 0 0 5.4 3.0 1.4 0.3 0.8 0 0 0 0.9 0.7 0 37 pH 8.5 UV/H2O2 4.8 mg L–1 9.3 6.2 1.7 0 0 0 4.4 3.0 1.4 0.3 1.0 0 0 0 0.6 0.7 0 34 pH 7.5 UV/H2O2 2.9 mg L–1 7.9 6.0 1.5 0 0 0 4.8 2.9 1.4 0.3 0.9 0 0 0 1.1 0.7 0 37 pH 6.5 UV/H2O2 2.9 mg L–1 7.6 6.0 1.5 0 0 0 5.5 2.9 1.4 0.3 0.9 0 0 0 1.0 0.7 0 43 pH 8.5 UV/H2O2 2.9 mg L–1 9.0 6.1 1.4 0 0 0 4.4 3.0 1.4 0.3 1.2 0 0 0 0.8 0.7 0 36 pH 7.5 UV/H2O2 1.0 mg L–1 8.6 5.9 1.3 0 0 0 4.8 2.9 1.4 0.3 0.8 0 0 0 0.9 0.8 0 42 pH 6.5 UV/H2O2 1.0 mg L–1 7.7 5.5 1.1 0 0 0 5.0 2.8 1.3 0.2 0.7 0 0 0 1.0 0.8 0 44 pH 8.5 UV/H2O2 1.0 mg L–1 9.4 6.1 1.4 0 0 0 4.1 2.8 1.4 0.3 0.9 0 0 0 0.7 0.7 0 34 24 h formation potential pH 7.5 Initial 16.1 13.2 3.4 0 0.5 0.1 12.2 12.5 3.6 0.8 3.7 0.7 0 0 1.6 2.0 0 100 pH 6.5 Initial 13.5 11.5 3.0 0 0.7 0.2 14.5 13.2 4.4 0.9 4.6 0.8 0 0 1.9 1.9 0 99 pH 8.5 Initial 19.8 16.0 4.0 0 0 0.2 11.2 10.4 3.3 0.8 3.1 0.9 0 0 1.3 1.8 0 94 pH 7.5 UV 19.2 17.7 2.6 0 1.0 0.3 18.6 14.0 3.9 0.7 3.1 1.2 0 0 2.1 4.4 0 94 pH 6.5 UV 17.1 17.0 2.3 0 0.9 0.2 18.1 14.4 4.3 0.9 3.7 0 0 0 2.8 3.8 0 87 pH 8.5 UV 24.3 21.4 3.8 0 0.8 0.3 13.2 11.5 3.6 0.6 3.2 1.7 0 0 1.9 4.4 0 88 pH 7.5 UV/Cl2 10 mg L–1 21.1 22.8 3.3 0 0.8 0.3 25.3 14.6 4.4 1.0 6.5 1.3 0 0 2.6 6.4 0 100 pH 6.5 UV/Cl2 10 mg L–1 23.4 24.4 4.1 0 0.9 0.5 35.8 24.5 5.6 1.7 9.0 0 0 0 4.4 9.3 0 101 pH 8.5 UV/Cl2 10 mg L–1 24.4 23.2 4.9 0 0.7 0.3 15.1 13.8 4.0 0.8 4.5 3.0 0 0 2.0 4.7 0 98 pH 7.5 H2O2 4.8 mg L–1 15.1 12.3 2.7 0 0.5 0.2 10.4 12.0 3.3 0.6 3.2 0.6 0 0 1.3 1.8 0 119 pH 6.5 H2O2 4.8 mg L–1 13.0 11.5 2.5 0 0.6 0.3 11.1 13.5 3.6 0.6 4.6 0.8 0 0 1.6 1.9 0 156 pH 8.5 H2O2 4.8 mg L–1 17.0 13.7 3.8 0 0 0.2 9.9 11.1 2.9 0.5 2.9 0.5 0 0 1.1 1.6 0 103 pH 7.5 UV/H2O2 4.8 mg L–1 25.6 23.4 3.2 0 0.6 0.4 30.0 19.0 4.4 1.3 6.8 1.7 0 0 2.0 7.9 0 127 pH 6.5 UV/H2O2 4.8 mg L–1 22.1 20.7 3.0 0 0.8 0.3 32.5 22.5 5.0 1.4 7.4 0.9 0 0 2.9 7.8 0 135 pH 8.5 UV/H2O2 4.8 mg L–1 31.1 22.6 4.9 0 0.7 0.4 20.2 17.0 4.6 1.1 6.2 1.8 0 0 2.2 5.2 0 119

145

Table I.6 Raw data of organic DBP formation (µg L–1) in the pilot-scale test TCM BDCM CDBM TBM MCAA MBAA DCAA TCAA BCAA DBAA BDCAA CDBAA TBAA TCAN DCAN BCAN DBAN AOX Initial for UV/H2O2 2.4 0.8 0.3 0 0 0 0.4 0.2 0.1 0 0 0 0 0 0 0 0 21 Initial for UV only 2.2 0.7 0.4 0 0 0 0.4 0.2 0.1 0 0 0 0 0 0 0 0 24 Initial for UV/Cl2 1.9 0.7 0.2 0 0 0 0.5 0.2 0.2 0 0 0 0 0 0 0 0 24 Initial for Cl2 only 2.1 0.8 0.3 0 0 0 0.4 0.3 0.1 0 0 0 0 0 0 0 0 33 pH 6.5 Cl2 10 mg L–1 5.4 3.2 1.1 0 0 0 1.4 1.3 0.2 0 0 0 0 0 0.8 0 0 47 pH 7.5 Cl2 10 mg L–1 5.7 4.5 1.3 0 0 0 1.3 1.2 0.2 0 0 0 0 0 0.7 0 0 38 pH 8.5 Cl2 10 mg L–1 4.8 3.7 1.0 0 0 0 1.1 0.9 0.2 0 0 0 0 0 0.6 0 0 38 pH 6.5 Cl2 6 mg L–1 4.3 2.9 1.0 0 0 0 1.2 1.0 0.2 0 0 0 0 0 0.6 0 0 36 pH 7.5 Cl2 6 mg L–1 5.2 4.3 1.3 0 0 0 1.2 1.1 0.3 0 0 0 0 0 0.6 0 0 38 pH 8.5 Cl2 6 mg L–1 4.5 3.4 1.0 0 0 0 1.1 0.8 0.3 0 0 0 0 0 0 0 0 37 pH 6.5 Cl2 2 mg L–1 2.3 1.8 0.7 0 0 0 0.9 0.7 0.2 0 0 0 0 0 0 0 0 35 pH 7.5 Cl2 2 mg L–1 3.0 2.4 0.8 0 0 0 1.0 0.8 0.2 0 0 0 0 0 0 0 0 42 pH 8.5 Cl2 2 mg L–1 2.8 1.6 0.4 0 0 0 0.7 0.5 0.2 0 0 0 0 0 0 0 0 34 pH 6.5 UV 2.5 0.6 0.3 0 0 0 0.4 0.2 0.1 0 0 0 0 0 0 0 0 18 pH 7.5 UV 2.3 0.7 0.3 0 0 0 0.4 0.2 0.1 0 0 0 0 0 0 0 0 19 pH 8.5 UV 2.2 0.8 0.4 0 0 0 0.4 0.2 0.1 0 0 0 0 0 0 0 0 17 pH 6.5 UV/Cl2 10 mg L–1 4.8 3.8 0.7 0 0 0.5 7.9 3.9 0.7 0.0 1.6 0 0 0 2.2 3.1 0 92 pH 7.5 UV/Cl2 10 mg L–1 4.7 3.7 1.0 0 0 0.2 5.0 2.2 0.6 0.0 0.9 0 0 0 1.3 1.9 0 71 pH 8.5 UV/Cl2 10 mg L–1 4.7 2.1 0.4 0 0 0.0 3.6 1.2 0.3 0.0 0.0 0 0 0 0.8 1.1 0 42 pH 6.5 UV/Cl2 6 mg L–1 4.4 3.6 0.8 0 0 0.6 6.8 3.1 0.8 0.0 1.4 0 0 0 1.7 2.4 0 66 pH 7.5 UV/Cl2 6 mg L–1 4.5 3.4 0.9 0 0 0.3 5.4 2.2 0.7 0.0 0.9 0 0 0 1.2 1.8 0 48 pH 8.5 UV/Cl2 6 mg L–1 3.8 1.9 0.4 0 0 0.0 3.4 1.1 0.4 0.0 0.0 0 0 0 0.8 0.9 0 34 pH 6.5 UV/Cl2 2 mg L–1 3.6 2.5 0.7 0 0 0.4 4.6 2.0 0.6 0.0 0.9 0 0 0 1.2 1.5 0 52 pH 7.5 UV/Cl2 2 mg L–1 4.5 3.2 0.9 0 0 0.3 4.8 1.7 0.6 0.0 0.0 0 0 0 1.1 1.5 0 38 pH 8.5 UV/Cl2 2 mg L–1 3.6 1.5 0.4 0 0 0.0 3.0 0.8 0.3 0.0 0.0 0 0 0 0.8 0.8 0 26 pH 6.5 UV/H2O2 4.8 mg L–1 2.4 0.7 0.3 0 0 0.0 0.5 0.2 0.1 0.0 0.0 0 0 0 0 0 0 17 pH 7.5 UV/H2O2 4.8 mg L–1 2.3 0.7 0.4 0 0 0.0 0.4 0.2 0.1 0.0 0.0 0 0 0 0 0 0 18 pH 8.5 UV/H2O2 4.8 mg L–1 2.2 0.7 0.3 0 0 0.0 0.5 0.2 0.1 0.0 0.0 0 0 0 0 0 0 21 pH 6.5 UV/H2O2 2.9 mg L–1 2.2 0.7 0.3 0 0 0.0 0.5 0.2 0.1 0.0 0.0 0 0 0 0 0 0 20 pH 7.5 UV/H2O2 2.9 mg L–1 2.2 0.7 0.3 0 0 0.0 0.4 0.2 0.1 0.0 0.0 0 0 0 0 0 0 16 pH 8.5 UV/H2O2 2.9 mg L–1 2.2 0.6 0.2 0 0 0.0 0.4 0.2 0.1 0.0 0.0 0 0 0 0 0 0 12 pH 6.5 UV/H2O2 1.0 mg L–1 2.4 0.7 0.3 0 0 0.0 0.4 0.2 0.1 0.0 0.0 0 0 0 0 0 0 14 pH 7.5 UV/H2O2 1.0 mg L–1 2.1 0.7 0.3 0 0 0.0 0.4 0.2 0.1 0.0 0.0 0 0 0 0 0 0 14 pH 8.5 UV/H2O2 1.0 mg L–1 2.4 0.8 0.4 0 0 0.0 0.5 0.2 0.1 0.0 0.0 0 0 0 0 0 0 14 24 h formation potential pH 6.5 Initial 24.0 21.7 2.6 0 0 0.7 25.4 26.1 4.0 0.4 12.4 3.2 0 0 3.7 5.1 0.2 193 pH 7.5 Initial 32.7 29.2 3.8 0 0 0.5 24.3 26.6 3.6 0.3 10.4 2.1 0 0 3.1 4.8 0.2 190 pH 8.5 Initial 45.7 37.1 6.1 0 0 0.7 19.1 19.1 4.4 0.8 7.4 3.4 0 0 1.7 0.7 0.3 190 pH 6.5 UV 25.2 28.6 2.9 0 0 0.6 28.7 21.0 3.4 0.4 8.4 2.0 0 0 3.4 8.4 0.2 196 pH 7.5 UV 37.4 37.2 4.3 0 0 0.7 26.9 19.1 3.2 0.4 8.6 2.8 0 0 2.8 6.9 0.3 19 pH 8.5 UV 52.4 46.0 6.8 0 0 0.7 17.4 14.8 3.4 0.7 5.3 2.7 0 0 1.8 1.1 0.4 204 pH 6.5 UV/Cl2 10 mg L–1 40.7 55.6 3.0 0 0 1.1 74.9 40.5 5.6 0.3 13.1 2.5 0 0 5.7 23.4 0.2 288 pH 7.5 UV/Cl2 10 mg L–1 42.3 54.1 4.2 0 0 0.7 52.3 28.3 5.4 0.5 11.8 3.2 0 0 3.5 13.2 0.2 227 pH 8.5 UV/Cl2 10 mg L–1 54.0 55.3 8.1 0 0 0.5 28.7 20.9 5.0 1.0 7.2 3.5 0 0 1.9 1.9 0.4 214 pH 6.5 UV/H2O2 4.8 mg L–1 31.8 41.2 2.8 0 0 0.5 50.1 29.8 4.8 0.4 11.4 2.4 0 0 3.6 15.4 0.2 249 pH 7.5 UV/H2O2 4.8 mg L–1 43.0 48.0 4.0 0 0 0.7 46.3 27.6 4.4 0.5 11.1 2.9 0 0 2.8 12.1 0.2 233 pH 8.5 UV/H2O2 4.8 mg L–1 58.7 54.7 6.1 0 0 0.8 29.7 22.3 4.6 0.7 5.0 2.2 0 0 1.9 2.0 0.3 231

146

Table I.7 Raw data of inorganic DBP formation (µg L–1) in the full- and pilot-scale tests Process UV alone UV/chlorine UV/H2O2 Dose 1800–2100 mJ cm–2 2 mg L–1 6 mg L–1 10 mg L–1 1.0 mg L–1 2.9 mg L–1 4.8 mg L–1 pH 6.5 7.5 8.5 6.5 7.5 8.5 6.5 7.5 8.5 6.5 7.5 8.5 6.5 7.5 8.5 6.5 7.5 8.5 6.5 7.5 8.5 Cornwall full-scale in late May, 2013 Chlorate initial 1.87 Chlorate final 1.71 2.44 2.20 1.65 1.90 1.65 1.65 1.76 1.68 1.96 1.84 1.66 Chlorite initial 0 Chlorite final 0 0.15 0.29 0.12 0.26 0.41 0.16 0.29 0.34 0.16 0.22 0.23 Perchlorate initial 0.05 Perchlorate final 0.04 0.05 0.06 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 Bromate initial 0.05 Bromate final 0 0.05 0 0 0 0 0 0 0 0 0.05 0 Cornwall full-scale in early September, 2013 Chlorate initial 2.63 44.0 43.5 40.2 132 124 126 219 214 231 Chlorate final 2.37 2.32 2.74 90.1 132 175 330 608 730 759 1153 1323 Chlorite initial 0 2.44 4.09 3.67 10.9 11.5 11.9 20.1 21.4 19.1 Chlorite final 0 0.14 0.28 0.79 1.30 1.90 0.25 0.12 0.26 0.15 0.15 0.16 Perchlorate initial 0.06 0.08 0.08 0.07 0.10 0.10 0.10 0.13 0.13 0.15 Perchlorate final 0.06 0.06 0.06 0.08 0.08 0.07 0.10 0.10 0.10 0.14 0.13 0.13 Bromate initial 0 0.30 0.28 0.28 0.96 0.87 0.93 1.67 1.52 1.56 Bromate final 0 0 0 0.48 0.46 0.40 2.00 1.84 1.33 3.88 2.88 2.05 Cornwall full-scale in early April, 2014 Chlorate initial 0.06 276 283 270 872 857 876 1542 1532 1506 0.06 Chlorate final 0.18 00.49 0.37 291 309 337 1045 1122 1183 1731 1865 2145 0.04 0.06 0.09 0.04 0.07 0.08 0.04 0.20 0.08 Chlorite initial 0 2.93 4.30 4.24 7.73 11.9 13.7 11.2 16.8 21.3 0 Chlorite final 0 0.67 2.11 0.93 1.19 2.47 0.36 0 0 0 0 0 0 0.30 1.04 0 0.36 0.87 0 0.18 0.32 Perchlorate initial 0.08 0.20 0.21 0.20 0.50 0.50 0.52 0.79 0.77 1.11 0.08 Perchlorate final 0.07 0.07 0.07 0.21 0.21 0.20 0.50 0.51 0.50 0.80 0.87 0.84 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 Bromate initial 0.02 0.57 0.64 0.60 1.91 1.87 1.97 3.17 3.07 2.96 0.02 Bromate final 0.04 0.05 0.03 0.72 0.70 0.67 2.74 2.58 2.22 4.94 4.70 3.78 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.02 0.01 Rayox® pilot-scale Chlorate initial 0.02 284 899 1577 0.02 Chlorate final 0.03 0 0 416 450 504 1233 1403 1555 2169 2517 2718 0.04 0.01 0.01 0.01 00.01 0 0.03 0 0 Chlorite initial 0 3.92 9.92 16.8 0 Chlorite final 0 0 0.04 0.61 0.79 1.78 0.07 0.06 0.22 0.07 0.06 0.07 0 0.04 0.09 0 0 0.04 0 0 0.09 Perchlorate initial 0.01 0.09 0.31 0.54 0.01 Perchlorate final 0.01 0.01 0.02 0.13 0.13 0.13 0.36 0.37 0.36 0.68 0.60 0.60 0.01 0.01 0.02 0.02 0.02 0.03 0.01 0.01 0.02 Bromate initial 0.04 0.24 0.65 1.04 0.04 Bromate final 0.05 0.05 0.04 0.38 0.44 0.33 1.69 1.77 1.04 3.08 3.05 1.61 0.04 0.05 0.04 0.05 0.05 0.05 0.05 0.04 0.05

Documents

Application of the UV/Chlorine Advanced Oxidation …...advice on UV technology since my Master’s study. I appreciate Prof. Susan Andrewsalso for being on my supervisory committee