EFFECTS OF PHOSPHATE-BASED CORROSION INHIBITORS …...ii effects of phosphate-based corrosion...

EFFECTS OF PHOSPHATE-BASED CORROSION

INHIBITORS ON DISINFECTANT STABILITY AND

HAA/NDMA FORMATION WHEN IN CONTACT WITH

COPPER, IRON AND LEAD

by

Hong Zhang

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 Hong Zhang 2012

ii

EFFECTS OF PHOSPHATE-BASED CORROSION

INHIBITORS ON DISINFECTANT STABILITY AND

HAA/NDMA FORMATION WHEN IN CONTACT WITH

COPPER, IRON AND LEAD

Hong Zhang

Doctor of Philosophy, 2012

Graduate Department of Civil Engineering

University of Toronto

ABSTRACT

This research examined the impacts of water quality, phosphate-based corrosion

inhibitors and pipe wall exposure on free chlorine (HOCl)/chloramine (NH2Cl) degradation and

haloacetic acid (HAA)/N-nitrosodimethylamine (NDMA) formation in simulated distribution

system water mains and household plumbing at bench-scale and pilot scale.

In bench-scale bottle tests, the reactivity of fresh/pre-corroded pipe materials with

HOCl/NH2Cl in decreasing order was: ductile iron, copper, lead. The addition of phosphate-

based corrosion inhibitors generally increased HOCl/NH2Cl degradation for fresh iron coupons,

but decreased HOCl/NH2Cl decay only for fresh copper coupons. Generally, these corrosion

inhibitors did not impact HAA formation.

Copper corrosion products, including Cu(II), Cu2O, CuO and Cu2(OH)2CO3, catalyzed

HAA and NDMA formation. For HAAs, copper catalysis increased with increasing pH from 6.6

to 8.6 and/or increasing concentrations of these copper corrosion products. Interactions of copper

with natural organic matter (NOM), likely by complexation, and the subsequent increase in the

reactivity of NOM were proposed to be the primary reason for the increased HAA formation.

iii

NDMA formation increased with increasing Cu(II) concentrations, DMA concentrations,

alkalinity and hardness but was inhibited by the presence of NOM. The transformation of NH2Cl

to dichloramine (NHCl2) and complexation of copper with DMA were proposed to be involved

in elevating the formation of NDMA at pH 7.0.

Finally, in pilot-scale modified pipe loop tests, copper catalysis of NDMA formation was

confirmed, especially under laminar flow conditions, and iron was shown to possibly catalyze

NDMA formation under turbulent conditions. Orthophosphate increased the catalytic effects of

iron but decreased copper catalysis on NDMA formation by either modifying the properties of

the iron-associated suspended particles or reducing the dissolved metal concentrations.

Orthophosphate increased chloramine decay when in contact with iron, likely by promoting

nitrite formation, but orthophosphate decreased chloramine decay for copper and lead by

reducing the availability of metal corrosion products.

iv

ACKNOWLEDGMENTS

This research was supported financially by the Canadian Water Network, and the Natural

Sciences and Engineering Research Council of Canada.

I would like to thank my supervisor, Professor Susan Andrews, for her advice and

guidance on my research work. I would especially like to thank her for her patience and

encouragement throughout my studies, especially during my paper and thesis writing. I am

grateful for my committee members Professor Robert Andrews, Professor Brent Sleep, Professor

Ron Hofmann and Professor Michele Prevost for offering their suggestions. I also very much

appreciate Tim Walton from Municipal Region of Waterloo for providing me with many water

samples during my research study and generously assisting in the built-up of my pilot-scale pipe

loops. Help from Ian Douglas from City of Ottawa was also much appreciated. I would also

thank Srebri Petrov from the Department of Chemistry, University of Toronto and Peter

Broderson from the Department of Chemical Engineering, University of Toronto for performing

metal surface analyses and providing valuable suggestions for interpreting data. Special regards

are extended to Russell D’Souza in the Civil Engineering Environmental labs for his

professionalism and to Jennifer Lee and the student colleagues in Drinking Water Research

Group for their kind friendship.

Finally, I give special thanks to my husband (Jim) and my parents for their love and

encouragement.

v

TABLE OF CONTENTS

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

ACKNOWLEDGMENTS ............................................................................................................. iv

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

LIST OF ACRONYMS ...................................................................................................................x

LIST OF TABLES ....................................................................................................................... xiv

LIST OF FIGURES ..................................................................................................................... xvi

Introduction .................................................................................................................................1 1

1.1 Background ........................................................................................................................1

1.2 Objectives ...........................................................................................................................3

1.3 Thesis Organization ............................................................................................................5

1.4 References ..........................................................................................................................6

Literature Review ........................................................................................................................9 2

2.1 Disinfectants .......................................................................................................................9

Chemistry of Chlor(am)ination ................................................................................9 2.1.1

Disinfectant Residual Reactions in Distribution Systems .....................................11 2.1.2

Kinetic Models for Disinfectant Residual Decay in Distribution Systems............13 2.1.3

2.2 Disinfection Byproducts ...................................................................................................16

Haloacetic Acids ....................................................................................................16 2.2.1

N-nitrosodimethylamine ........................................................................................16 2.2.2

Disinfection By-Products in the Distribution System ............................................18 2.2.3

2.3 Corrosion ..........................................................................................................................20

Metal Dissolution and Precipitation in the Distribution System ...........................20 2.3.1

vi

Factors Affecting Metal Levels at the Tap ............................................................21 2.3.2

Corrosion Control ..................................................................................................23 2.3.3

2.4 Nitrification ......................................................................................................................24

2.5 Summary of Research Gaps .............................................................................................26

2.6 References ........................................................................................................................27

Free Chlorine Degradation and HAA Formation in Two Water Matrices in Contact with 3

Three Metal Materials ...............................................................................................................40

3.1 Introduction ......................................................................................................................42

3.2 Materials and Methods .....................................................................................................44

Reagents and Materials ..........................................................................................44 3.2.1

Experimental Procedures .......................................................................................45 3.2.2

3.3 Results and Discussion .....................................................................................................47

Free Chlorine Degradation .....................................................................................47 3.3.1

HAA Formation .....................................................................................................55 3.3.2

HAA Speciation .....................................................................................................62 3.3.3

3.4 Summary ..........................................................................................................................66

3.5 References ........................................................................................................................68

A Comparison of Iron, Copper and Lead Corrosion in Simulated Distribution Systems .........72 4

4.1 Introduction ......................................................................................................................74

4.2 Materials and Methods .....................................................................................................75

Reagents and Materials ..........................................................................................75 4.2.1

Experimental Procedures .......................................................................................77 4.2.2

Coupon Surface Analysis .......................................................................................78 4.2.3

4.3 Results and Discussion .....................................................................................................79

Metal Release and Phosphate-based Corrosion Inhibitors.....................................79 4.3.1

XPS Results for Corroded Coupons in Mannheim Water .....................................84 4.3.2

vii

Water Quality and Disinfectant Type ....................................................................89 4.3.3

4.4 Summary ..........................................................................................................................92

4.5 References ........................................................................................................................94

Catalytic Impacts of Copper Corrosion Products on Chlorine Decay and HAA Formation 5

in Simulated Distribution Systems ............................................................................................99

5.1 Introduction ....................................................................................................................101

5.2 Materials and Methods ...................................................................................................103

Reagents and Materials ........................................................................................103 5.2.1

Experimental Procedures .....................................................................................104 5.2.2

5.3 Results and Discussion ...................................................................................................105

Chlorine Decay ....................................................................................................106 5.3.1

HAA Formation and Speciation...........................................................................111 5.3.2

HAA Speciation ...................................................................................................115 5.3.3

5.4 Summary ........................................................................................................................117

5.5 References ......................................................................................................................118

Factors Affecting Copper Catalysis of NDMA Formation from DMA in Simulated 6

Premise Plumbing ...................................................................................................................122

6.1 Introduction ....................................................................................................................124

6.2 Materials and methods ....................................................................................................126

Chemicals and Materials ......................................................................................126 6.2.1


6.3 Results and discussion ....................................................................................................128

Effect of Copper Concentrations .........................................................................128 6.3.1

Effects of DMA Concentration and Cu-NH2Cl Interactions ...............................129 6.3.2

Cu-DMA Complexation.......................................................................................131 6.3.3

Effect of Alkalinity ..............................................................................................132 6.3.4

Effect of NOM .....................................................................................................134 6.3.5

viii

Effect of Hardness................................................................................................135 6.3.6

Effect of pH and the Role of NHCl2 in NDMA Formation .................................136 6.3.7

6.4 Summary ........................................................................................................................138

6.5 References ......................................................................................................................139

Effects of Pipe Materials, Orthophosphate, and Flow Conditions on Chloramine Decay 7

and NDMA Formation in Modified Pipe Loops .....................................................................144

7.1 Introduction ....................................................................................................................146

7.2 Materials and Methods ...................................................................................................147

Reagents and Materials ........................................................................................147 7.2.1

Modified Pipe Loops............................................................................................148 7.2.2


7.3 Results and Discussion ...................................................................................................151

Metal and Nitrogen Species Concentrations ........................................................151 7.3.1

Chloramine decay ................................................................................................154 7.3.2

NDMA formation.................................................................................................158 7.3.3

7.4 Summary ........................................................................................................................162

7.5 References ......................................................................................................................164

Conclusions .............................................................................................................................168 8

Practical Implications and Suggestions for Future Research ..................................................170 9

Appendices ..............................................................................................................................173 10

10.1 Authorisations to Include Copyright Material in Thesis ................................................173

10.2 QA/QC Protocols ...........................................................................................................177

GC/MS .................................................................................................................178 10.2.1

GC/ECD ...............................................................................................................180 10.2.2

Flame Atomic Absorption Spectrometry .............................................................185 10.2.3

Ion Chromatography ............................................................................................187 10.2.4

ix

10.3 Free Chlorine 24-hour Residuals through the Duration of Metal Coupon

Conditioning ...................................................................................................................190

10.4 HAA Speciation for Fresh Metal Coupons in Britannia Water ......................................191

10.5 XRD Analysis for Fresh Iron Coupons ..........................................................................192

10.6 Metal Release Kinetics and Results for Metal Surface Analysis ...................................193

10.7 NDMA Formation in the Presence of Orthophosphate in Modified Pipe Loops ...........197

10.8 Effects of Corrosion Inhibitors and the Extent of Metal Corrosion on

Monochloramine Degradation and NDMA Formation ..................................................198

Materials and Methods .........................................................................................199 10.8.1

Monochloramine Degradation .............................................................................202 10.8.2

NDMA Formation from DMA for Fresh Coupons ..............................................207 10.8.3

References ............................................................................................................212 10.8.4

10.9 Degradation Potential of Iron, Lead and Their Corrosion Products on HAA9 ...............216

Materials and Methods .........................................................................................217 10.9.1

HAA9 Degradation by Corroded Iron Coupons ...................................................218 10.9.2

HAA Degradation by Lead ..................................................................................223 10.9.3

References ............................................................................................................227 10.9.4

x

LIST OF ACRONYMS

AOB Ammonia-oxidizing bacteria

ACS Chemical purity designation as defined and published by the

American Chemical Society Committee on Analytical

Reagents

ANOVA Analysis of variance

APHA American Public Health Association

AWWA American Water Works Association

BCAA Bromochloroacetic acid

BDCAA Bromodichloracetic acid

BDOC Bioeliminable dissolved organic carbon

CDBAA Chlorodibromoacetic acid

CDPH California Department of Public Health

DBAA Dibromoacetic acid

DBP Disinfection byproducts

DCAA Dichloroacetic acid

D/DBPR Disinfectant/ Disinfection Byproduct Rule

DMA Dimethylamine

ECD Electron capture detector

FAAS Flame atomic absorption spectrometry

xi

GC/MS Gas chromatography/mass spectrometry

HAA Haloacetic acid

HOCl Hypochlorous acid

HPLC High-performance liquid chromatography

HRT Hydraulic retention time

H2SO4 Sulphuric acid

ID Inner diameter

LCR Lead and Copper Rule

LSD Least Significant Difference (Fisher’s)

MAC Maximum Acceptable Concentration

MCAA Monochloroacetic acid

MBAA Monobromoacetic acid

MCL Maximum Contaminant Level

MDL Method detection limit

MOE Ministry of the Environment (Ontario)

MS-FP Material-specific formation potential

MS-SDS Material-specific simulated distribution system

MWTP Mannheim Water Treatment Plant

NA Not available

NaOH Sodium hydroxide

xii

NCl3 Trichloramine or nitrogen trichloride

NDEA N-nitrosodiethylamine

NDMA N-nitrosodimethylamine

NDPA N-nitrosodi-n-propylamine

NDPhA N-nitrosodiphenylamine

NH2Cl Monochloramine

NHCl2 Dichloramine

NMOR N-nitrosomorpholine

NOB Nitrite-oxidizing bacteria

NOM Natural organic matter

NPIP N-nitrosopiperidine

NPYR N-nitrosopyrollidine

PACl Polyaluminum chloride

PVC Polyvinyl chloride

QA/QC Quality Assurance and Quality Control

SDS Simulated distribution system

SUVA UVA254/TOC

TBAA Tribromoacetic acid

TCAA Trichloroacetic acid

THM Trihalomethane

xiii

TOC Total organic carbon

UDMH Unsymmetrical dimethylhydrazine

USEPA United States Environmental Protection Agency

UV Ultraviolet

UVA254 Ultraviolet absorbance at a wavelength of 254nm

WEF Water Environment Foundation

XRD X-ray diffraction

XPS X-ray photoelectron spectroscopy

xiv

LIST OF TABLES

Table 2-1 Chlorine reactions in drinking water treatment ............................................................ 10

Table 2-2 Primary reactions of monochloramine ......................................................................... 11

Table 2-3 Free chlorine and chloramine model coefficients ........................................................ 15

Table 2-4 Relevant equilibrium reactions for copper in carbonated-bearing water ..................... 21

Table 3-1 Summary of water quality parameters for post-filtration water ................................... 45

Table 3-2 Summary of analytical methods ................................................................................... 46

Table 3-3 Chlorine decay constants for fresh metal coupons in the presence and absence of

corrosion inhibitors with Mannheim Water (n=4) ....................................................... 49

Table 3-4 Comparison of HOCl overall decay constants (h-1) for fresh coupons between two

water matrices (n=4) ..................................................................................................... 51

Table 3-5 Comparison of HOCl overall decay constants for corroded coupons between

Mannheim Water and Britannia Water (n=4) ............................................................... 54

Table 4-1 Summary of water quality parameters for two post-filtration water sources ............... 77

Table 4-2 Summary of analytical methods ................................................................................... 78

Table 5-1Water quality parameters for post-filtration water from MWTP ................................ 104

Table 7-1 Summary of water quality parameters for the influent of the pipe loops ................... 147

Table 7-2 Summary of design parameters for pipe loops ........................................................... 149

Table 7-3 Significance of the effects of flow conditions on chloramine decay determined by the

LSD test (95% confidence level) ................................................................................ 155

Table 7-4 Significance of the effects of orthophosphate on chloramine decay determined by the

LSD test (a confidence level of 95%) ........................................................................ 156

xv

Table 10-1 GC/MS method – Method detection limits for NDMA (n=8, 99% confidence level)

.................................................................................................................................... 178

Table 10-2 GC-ECD method – Method detection limits for HAA9 (n=8, 99% confidence level)

.................................................................................................................................... 181

Table 10-3 FAAS method – Method detection limits for iron, copper and lead ........................ 186

Table 10-4 IC method – Method detection limits for nitrite, nitrate, chloride, bromide, sulphate

and phosphate ............................................................................................................. 189

Table 10-5 Water quality parameters for post-filtration water ................................................... 200

Table 10-6 Summary of analytical methods ............................................................................... 201

Table 10-7 Comparison of NH2Cl decay constants for fresh coupons in two water matrices (n=4)

.................................................................................................................................... 202

Table 10-8 Comparison of NH2Cl overall decay constants for corroded coupons between

Mannheim Water and Britannia Water (n=4) ............................................................. 207

Table 10-9 Summary of the results for the single contrasts using the LSD test ......................... 209

xvi

LIST OF FIGURES

Figure 3-1 Free chlorine decay for fresh metal coupons with time in the presence and absence of

corrosion inhibitors in one set of experiments, n=2; (a) iron, (b) copper, (c) lead. ..... 48

Figure 3-2 Free chlorine overall decay constants for corroded coupons with Mannheim Water.

Initial free chlorine concentration 12.3 mg/L, error bars indicate standard deviation

(n=4) ............................................................................................................................. 52

Figure 3-3 Copper release kinetics for corroded coupons in Mannheim Water and Britannia

Water; error bars indicate the measured maximum and minimum values (n=2) ......... 55

Figure 3-4 Lead release kinetics for corroded coupons in Mannheim Water and Britannia Water;

error bars indicate the measured maximum and minimum values (n=2) ..................... 55

Figure 3-5 HAA6 formations with time in the presence and absence of corrosion inhibitors for

fresh metal coupons with Mannheim Water in one set of experiments, error bars

indicate the measured maximum and minimum values (n=2) ..................................... 56


fresh metal coupons with Britannia Water in one set of experiments, error bars indicate

the measured maximum and minimum values (n=2) ................................................... 59

Figure 3-7 HAA6 formation at 48 hours in the presence and absence of corrosion inhibitors for

corroded metal coupons with Mannheim Water and Britannia Water in one set of

experiments, error bars indicate the measured maximum and minimum values (n=2) 60

Figure 3-8 HAA formation and free chlorine demand for bulk water in the absence of corrosion

inhibitors ....................................................................................................................... 61

Figure 3-9 HAA formation and free chlorine demand for copper in the absence of corrosion

inhibitors ....................................................................................................................... 62

xvii

Figure 3-10 HAA speciation with time in the presence of 1 mg/L orthophosphate for fresh metal

coupons with Mannheim Water in one set of experiments, error bars indicate the

measured maximum and minimum values (n=2) ......................................................... 63

Figure 3-11 HAA speciation with time in the presence of 1 mg/L orthophosphate for corroded

metal coupons with Mannheim Water in one set of experiments, error bars indicate the



metal coupons with Britannia Water in one set of experiments, error bars indicate the


Figure 4-1 Kinetics of metal release from fresh metal coupons in the presence and absence of

corrosion inhibitors with HOCl in Mannheim Water; disinfectant concentrations, 12.3

mg/L; error bars indicate the measured maximum and minimum values (n=2) .......... 80

Figure 4-2 Comparison of XRD patterns for iron powders scratched from the surface of oxidized

iron coupons in the absence and presence of corrosion inhibitors ............................... 81

Figure 4-3 Comparison of XRD patterns for copper coupons in the absence and presence of

corrosion inhibitors. ...................................................................................................... 82

Figure 4-4 Comparison of XRD patterns for lead coupons in the absence and presence of 1 mg/L

orthophosphate ............................................................................................................. 82

Figure 4-5 Metal concentrations after 24 hours for corroded coupons in the absence/presence of

orthophosphate with HOCl and NH2Cl in Mannheim Water, initial disinfectant

concentrations 5.5 mg/L; error bars indicate the measured maximum and minimum

values (n=2) .................................................................................................................. 84

Figure 4-6 Comparison of elemental distribution for copper, lead and iron coupons in the

absence and presence of orthophosphate with NH2Cl ................................................. 86

Figure 4-7 Corrosion products of Fe, Cu and Pb as well as their relative distribution in the scales

...................................................................................................................................... 87

xviii

Figure 4-8 Comparison of iron concentrations at 24 hours in the presence of orthophosphate for

fresh and corroded coupons under free chlorine and chloramine, error bars indicate the


Figure 5-1 Chlorine degradation at different pH values in the absence and presence of 1 mg/L Cu

(II) for MWTP water; initial HOCl =10 mg/L; triplicate ........................................... 106

Figure 5-2 Chlorine decay for synthetic water in the presence and absence of 1 mg/L Cu(II) and

NOM; initial HOCl =4.2 mg/L; pH 8.3; duplicate ..................................................... 108

Figure 5-3 Chlorine degradation in the presence of different concentrations of Cu(II) for MWTP

water; initial HOCl =10 mg/L; triplicate; pH 8.3 ....................................................... 109

Figure 5-4 Pseudo-first-order decay rates of free chlorine for MWTP water containing dissolved

Cu(II) and solid copper corrosion products ................................................................ 110

Figure 5-5 HAA9 at 100 hours at different pH values in the absence and presence of 1 mg/L

Cu(II) for MWTP water, initial HOCl =10 mg/L (error bars represent standard

deviation of triplicate tests) ........................................................................................ 112

Figure 5-6 HAA9 formation in the presence of Cu(II) with varying concentrations at pH 8.3 for

MWTP water (error bars represent standard deviation of triplicate tests) ................. 113

Figure 5-7 HAA9 at a reaction time of 32 hours in the presence of copper corrosion solids at pH

8.3 for MWTP water (error bars represent standard deviation of triplicate tests) ...... 114

Figure 5-8 HAA9 formation rates for MWTP water containing dissolved Cu(II) without solids

and solid copper corrosion products ........................................................................... 115

Figure 5-9 Effect of pH on HAA speciation in the absence and presence of 1 mg/L Cu(II) at

reaction time of 100 hours for MWTP water (error bars represent standard deviation of

triplicate tests) ............................................................................................................ 116

Figure 6-1 NDMA formation with increasing copper concentrations from added CuSO4; pH 7.0,

initial NH2Cl 2.3±0.1 mg/L, 11.2 µg/L DMA spiked, 24 hours, error bars indicate the

measured maximum and minimum values (n=2) ....................................................... 129

xix

Figure 6-2 NDMA formation kinetics in the absence and presence of copper; pH 7, Cu(II) 1

mg/L spiked, initial NH2Cl 2.3±0.1 mg/L, 11.2 µg/L DMA, error bars indicate the


Figure 6-3 NDMA formation with DMA concentrations; pH 7.0, initial NH2Cl 2.3±0.1 mg/L,

Cu(II) 1 mg/L spiked, 24 hours, d error bars indicate the measured maximum and

minimum values (n=2) ............................................................................................... 130

Figure 6-4 Chloramine speciation in the absence and presence of copper at pH 6.7 and 7, Milli-Q

water ........................................................................................................................... 131

Figure 6-5 Dissolved copper concentrations with varying DMA; filtered by 0.2 µm Nylaflo®

Nylon membrane filter paper, pH 7, Cu(II) 1 mg/L spiked, 6 hours, error bars indicate

the measured maximum and minimum values (n=2) ................................................. 132

Figure 6-6 NDMA formation with increasing alkalinity and copper speciation as a function of

alkalinity (determined by MINEQL+ version 4.5); pH 7.0, initial NH2Cl 2.3±0.1

mg/L, Cu(II) 1 mg/L spiked, 11.2 µg/L DMA, 24 hours, duplicate ......................... 133

Figure 6-7 NDMA, NH2Cl residual and dissolved copper with increasing SR-NOM

concentrations; pH 7.0, initial NH2Cl 2.3±0.1 mg/L, Cu(II) 1 mg/L spiked, 11.2 µg/L

DMA, 24 hours, error bars indicate the measured maximum and minimum values

(n=2) ........................................................................................................................... 134

Figure 6-8 Variation of NDMA formation and dissolved copper concentrations with increasing

hardness; pH 7.0, Cu(II) 1 mg/L spiked, initial NH2Cl 2.3±0.1 mg/L, 11.2 µg/L DMA

spiked, SR-NOM 4.1 mg/L as TOC, 24 hours, error bars indicate the measured

maximum and minimum values (n=2) ....................................................................... 136

Figure 6-9 NDMA formation at four pH levels in the absence and presence of 1 mg/L Cu(II) in

Milli-Q water; 24 hour, NH2Cl 2.3±0.1 mg/L, 11.2 µg/L DMA, error bars indicate the


Figure 7-1 Module configuration (adapted from Cantor, 2009) ................................................. 148

Figure 7-2 Schematic of the modified pipe loop (not to scale)................................................... 149

xx

Figure 7-3 Metal concentrations in the pipes of iron, copper and lead in the absence and presence

of 1 mg/L orthophosphate under two flow conditions ............................................... 152

Figure 7-4 Variations of nitrate, nitrite and ammonia in the lead loop....................................... 153

Figure 7-5 Pseudo-first-order chloramine decay constants for the four pipe loops under different

flow conditions, initial chloramine 1.3~1.6 mg/L as Cl2, error bars indicate the


Figure 7-6 Box and Whisker plots for nitrite concentrations before and after the addition of

orthophosphate overlaid with pseudo-first-order chloramine decay rate constants,

laminar flow (n=5), turbulent flow (n=6). The top and bottom of the box represent the

75th and 25th percentile, respectively, while the whiskers represent the maximum and

minimum values ......................................................................................................... 157

Figure 7-7 NDMA formation in four pipe loops in the absence of orthophosphate under two flow

conditions ................................................................................................................... 159

Figure 10-1 Example of GC/MS calibration curve for NDMA (NDMA 0~200 ng/L, d6-NDMA

50 ng/L) ...................................................................................................................... 178

Figure 10-2 GC/MS method for NDMA- spike recovery chart .................................................. 179

Figure 10-3 Example of GC-ECD calibration curves for HAA9 (HAA9 0~100 µg/L, 2,3,5,6-

TFBA 100 µg/L) ......................................................................................................... 180


TFBA 100 µg/L), continued ....................................................................................... 181

Figure 10-5 GC-ECD method for MCAA- spike recovery chart ............................................... 182

Figure 10-6 GC-ECD method for MBAA- spike recovery chart ............................................... 182

Figure 10-7 GC-ECD method for DCAA- spike recovery chart ................................................ 183

Figure 10-8 GC-ECD method for TCAA- spike recovery chart................................................. 183

xxi

Figure 10-9 GC-ECD method for DBAA- spike recovery chart ................................................ 184

Figure 10-10 GC-ECD method for BCAA- spike recovery chart .............................................. 184

Figure 10-11 Example of FAAS calibration curves for iron, copper and lead ........................... 185

Figure 10-12 Example of IC calibration curves for chloride, sulphate and bromide ................. 187

Figure 10-13 Example of IC calibration curves for nitrite, nitrate and phosphate ..................... 188

Figure 10-14 HAA speciation with time in the presence 1 mg/L orthophosphate for fresh metal

coupons with Britannia Water in one set of experiments, error bars indicate the


Figure 10-15 Comparison of XRD patterns for oxidized iron coupon in the presence of

polyphosphate-orthophosphate blends and polished iron coupon as control. ............ 192


corrosion inhibitors with NH2Cl in Mannheim Water; disinfectant concentrations, 12.3

mg/L; error bars indicate the measured maximum and minimum values (n=2) ........ 193

Figure 10-17 Kinetics of metal release from fresh metal coupons with time in the presence and

absence of corrosion inhibitors in Britannia Water; disinfectant concentrations, 12.3

mg/L; error bars indicate the measured maximum and minimum values (n=2) ........ 194

Figure 10-18 Kinetics of metal release from corroded coupons in the absence/presence of

orthophosphate in Britannia Water, initial disinfectant concentrations 5.5 mg/L; error

bars indicate the measured maximum and minimum values (n=2) ............................ 195

Figure 10-19 Comparison of elemental distribution for iron, copper and lead coupons in the

absence and presence of orthophosphate with HOCl ................................................. 196

Figure 10-20 NDMA formation in four pipe loops in the presence of orthophosphate under two

flow conditions ........................................................................................................... 197

Figure 10-21 NH2Cl overall decay constants for fresh coupons with Mannheim Water; NH2Cl

12.3 mg/L, error bars indicate the measured maximum and minimum values (n=2) 203

xxii

Figure 10-22 Released metal concentrations at 24 hours for fresh metal coupons in Mannheim

Water; NH2Cl 12.3 mg/L; error bars indicate the measured maximum and minimum

values (n=2) ................................................................................................................ 204

Figure 10-23 Released iron and copper concentrations at 24 hours in the absence and presence of

corrosion inhibitors for fresh metal coupons in Britannia Water; NH2Cl 12.3 mg/L;

error bars indicate the measured maximum and minimum values (n=2) ................... 206

Figure 10-24 NDMA formation from DMA as a result interactions of metal coupons, corrosion

inhibitors and NH2Cl. Initial NH2Cl 14.5±0.5 mg/L, DMA 1 µM/L, pH 8.3, in Milli-

Q, error bars indicate standard deviation (n=3) .......................................................... 208

Figure 10-25 Total iron concentrations and NDMA formation for iron coupons in the absence

and presence of corrosion inhibitors, error bars indicate standard deviation (n=3) ... 210

Figure 10-26 Cu(II) concentrations and NDMA formation for copper coupons in the absence and

presence of corrosion inhibitors, error bars indicate standard deviation (n=3) .......... 211

Figure 10-27 Reduction of HAA9 by corroded iron coupons in the absence and presence of

different corrosion inhibitors. Milli-Q water, error bars indicate the measured


Figure 10-28 HAA degradation by corroded iron coupons treated with poly/ortho blends, Milli-Q

water, error bars indicate the measured maximum and minimum values (n=2) ........ 219

Figure 10-29 Reduction of MCAA by corroded iron coupons in the absence and presence of

different corrosion inhibitors. Milli-Q water, error bars indicate the measured



poly/ortho-phosphate blends and polished iron coupon as control. Red: iron coupon as

control; Black: non-scratched iron coupon with poly/ortho-phosphate blends .......... 220

Figure 10-31 HAA speciation degradation by corroded iron coupons treated with poly/ortho

blends, error bars indicate the measured maximum and minimum values (n=2)....... 221

xxiii

Figure 10-32 Reduction of HAA9 by fresh lead coupons, Milli-Q water, pH 8.3, error bars

indicate the measured maximum and minimum values (n=2) ................................... 224

Figure 10-33 Degradation of MCAA by fresh lead coupons, Milli-Q water, pH 8.3, error bars

indicate the measured maximum and minimum values (n=2) ................................... 224

Figure 10-34 Degradation of single HAA species by fresh lead coupons, Milli-Q water, pH 8.3,

error bars indicate the measured maximum and minimum values (n=2) ................... 225

Figure 10-35 Degradation of HAA species in the presence of 1g/L lead corrosion products. Milli-

Q water, pH 8.3, error bars indicate the measured maximum and minimum values

(n=2) ........................................................................................................................... 227

1

Introduction 1

1.1 Background

Free or combined chlorine is commonly applied as a disinfectant residual to prevent

water quality deterioration in distribution systems beyond the treatment plant and to ensure the

delivery of safe and clean water to consumers. However, there are drawbacks to chemical

disinfection. It increases the risks of disinfection by-product (DBP) formation when these

secondary disinfectants react with natural organic matter present in the water. During

chlorination, trihalomethanes (THMs) and haloacetic acids (HAA9) are the dominant classes of

chlorinated DBPs, whereas combined chlorine potentially increases the formation of

nitrosamines (Choi et al., 2002; Najm and Trussell, 2001). This thesis focuses on HAAs and

nitrosamines because THM regulations are already well-established, while HAA and nitrosamine

regulations have been under recent development. In addition, switching from free chlorine to

combined chlorine to decrease HAAs (and THMs) has been observed to increase nitrosamines.

Therefore, nitrosamine formation might be an unintended consequence of meeting new HAA

regulations.

These DBPs have potential health risks. For example, the toxicological properties of

HAAs vary with the specific compound. Dichloroacetic acid (DCAA) is considered to be a

probable carcinogen, and trichloroacetic acid (TCAA) is regarded as a possible carcinogen to

humans (Health Canada, 2008). The Stage 2 Disinfectant/ Disinfection Byproduct Rule

(D/DBPR) regulates the Maximum Contaminant Level (MCL) of HAA5 (monochloroacetic acid,

DCAA, TCAA, monobromoacetic acid, and dibromoacetic acid) at 60 µg/L (USEPA, 2006). The

Guidelines for Canadian Drinking Water Quality established the Maximum Acceptable

Concentration (MAC) for HAA5 in drinking water at 80 µg/L based on a locational running

annual average of a minimum of quarterly samples taken from the distribution system (Health

Canada, 2008). N-nitrosodimethylamine (NDMA) is classified as a probable human carcinogen,

and guidelines to limit its concentration in drinking water vary. Health Canada has established

the guideline for NDMA in drinking water at a MAC of 40 ng/L (Health Canada, 2011). The

Ontario Ministry of the Environment, Canada, has set a MAC for NDMA of 9 ng/L (MOE,

2

2002). The California Department of Public Health has set a notification level of 10 ng/L for

NDMA (CDPH, 2009).

As potential oxidants, secondary disinfectants may also increase metal corrosion rates for

distribution pipes and household plumbing materials. Although confounded by other factors such

as pH, alkalinity and pipe age, free chlorine is generally more aggressive than combined chlorine

in metal corrosion due to its higher oxidation potential. Increased iron and copper corrosion rates

have been observed during chlorination (Boulay and Edwards, 2001; Cantor et al., 2003;

MacQuarrie et al., 1997; Rahman et al., 2007). However, increased lead leaching has been linked

to the switch of the disinfectant residual from free chlorine to monochloramine (Renner, 2004).

Despite some disadvantages associated with combined chlorine, the application of combined

chlorine will receive increasing attention due to the implementation of Stage 2

Disinfectant/Disinfection Byproduct Rule (D/DBPR) (USEPA, 2006) and the need to decrease

chlorinated DBPs.

To date, extensive studies have investigated disinfectant decay and corrosion processes

for lead, copper and iron individually in simulated and/or full distribution systems where free

chlorine and chloramine are used (including but not limited to Rossman et al., 2001; Haas et al.,

2002; Hallam et al., 2002; Al-Jasser, 2007; Lytle and Schock, 2005; McNeill and Edwards,

2001; Sarin et al., 2003; Xiao et al., 2007). These studies have examined the roles of physical

and chemical compositions of water, biofilms, and hydrodynamics on disinfectant residual

stability and corrosion. The addition of phosphate-based corrosion inhibitors as a corrosion

control strategy has been widely employed by many water systems to achieve compliance with

the USEPA’s Lead and Copper Rule (LCR) (USEPA, 2008). These corrosion inhibitors may

interfere with metal solubility and speciation, thereby making interactions between corrosion

products, disinfectant residual and DBP formation more complicated. To the author’s

knowledge, to date, no research has been conducted to investigate the interactive impacts of

disinfectant residual, metal corrosion and corrosion inhibitors on DBP formation (especially for

HAAs or nitrosamines such as NDMA). Furthermore, there have been no reports that compare

the behavior of disinfectant degradation and DBP formation for all three metal materials in the

same water matrix to develop a more comprehensive understanding about the impacts of metal

materials and their interactions with corrosion inhibitors on disinfectant decay and DBP

formation.

3

Therefore, several questions that could be asked include:

Are there differences in disinfectant degradation and HAA/NDMA formation for

different pipe materials treated with different phosphate-based corrosion

inhibitors?

Do metal age and water quality affect the efficacy of corrosion inhibitors and

disinfectant decay as well as HAA/NDMA formation?

How do flow conditions potentially affect secondary disinfectant stability and

NDMA formation?

What are the impacts of metal corrosion products on the fate of HAA/NDMA in

distribution systems?

An improved understanding of these interactions is needed as a basis for the development of

effective mitigation strategies.

1.2 Objectives

The overall objective of this research was to evaluate the impacts that water quality

characteristics (e.g. pH and alkalinity), phosphate-based corrosion inhibitors, pipe wall exposure

and flow conditions may have on disinfectant residual stability and the formation of disinfection

byproducts (e.g., haloacetic acids and nitrosamines) in distribution systems. Regulatory

guidelines for HAA5 and N-nitrosodimethylamine (NDMA) have recently been added to the

“Guidelines for Canadian Drinking Water Quality” (Health Canada, 2008 and 2011). This

information, therefore, has increased the awareness of water utilities and households to minimize

HAAs and NDMA in their water systems. Given that the regulations for THMs have been

already well-established in Canada, their formation in distribution systems is, therefore, beyond

the scope of this research. Results of this research will advance the understanding of the complex

physical, chemical and, to some extent, biological reactions which potentially cause disinfectant

degradation and HAA/NDMA formation in distribution system water mains and household

plumbing. The specific objectives of this thesis were to:

1) Investigate disinfectant degradation and HAA/NDMA formation in different water matrices

under the influence of metal corrosion and corrosion inhibitors. No relevant information has

been found in the literature regarding the impacts of corrosion inhibitors and/or their

4

interactions with metal surface on disinfectant degradation and HAA/NDMA formation in

newly installed or the aged pipe systems. Therefore, there is a need to investigate the

impacts of metal corrosion and corrosion preventive strategies on: a) disinfectant residual

stability and b) the fate of HAA/NDMA in distribution mains and household plumbing.

2) Examine the impact of disinfectant types, corrosion inhibitors and water quality on metal

release kinetics using material-specific simulated distribution system tests. Metal corrosion

is a primary reason for disinfectant residual degradation in distribution systems. To date,

most studies investigating metal corrosion processes have been conducted by performing

either pilot-scale pipe rig (loop) tests or full-scale distribution system tests. Impacts of

corrosion inhibitors, disinfectant types and water quality on metal corrosion may have been

confounded by hydrodynamics and microorganisms, and thus cannot be fully elucidated.

Therefore, the investigation of factors influencing metal corrosion under controlled

experimental conditions is important for evaluating the impacts of metal types, corrosion

inhibitors and water quality on disinfectant degradation and HAA/NDMA formation.

3) Evaluate the catalytic potential of copper and its corrosion products during HAA9 and

NDMA formation. The investigation of copper catalysis during HAA9 and NDMA formation

is essential as it may affect the fate of HAA9 and NDMA in household plumbing where

copper pipes and alloys are widely used. Limited research to date has been performed to

look into copper catalytic potential, the factors affecting copper catalysis (including pH, the

presence of organic matter, and copper concentrations), and the likely mechanisms of copper

catalysis during HAA9 and NDMA formation.

4) Examine the effects of pipe materials and orthophosphate on chloramine decay and NDMA

formation under different flow conditions using modified pipe loops. Information about

chloramine degradation kinetics and NDMA formation under the influence of pipe materials,

corrosion inhibitors, nitrite formation, and hydrodynamic conditions is limited and

inconsistent. This demonstrates a need to identify the factors which may play a key role in

maintaining chloramine residual and the formation of NDMA in pipes of different metallic

materials.

5

1.3 Thesis Organization

This thesis is composed of a general introduction (Chapter 1), a brief literature review

(Chapter 2), and five results chapters. Early experiments to survey the possible impacts of free

chlorine on HAA formation and corrosion are summarized and separated into Chapters 3 and 4.

Chapter 3 provides detailed discussion about the effects of corrosion inhibitors and their

interactions with metal surfaces on free chlorine degradation and HAA formation. Chapter 4

examines the metal release kinetics of iron, copper and lead through the use of these bench-scale

material-specific simulated distribution system tests and material-specific formation potential

tests. Copper catalysis of HAA and NDMA formation, under controlled bench-scale

experimental conditions, are reported in Chapters 5 and 6, respectively. Chapter 7 looks further

into some NDMA formation issues and discusses the results of pilot-scale tests regarding the

impacts of pipe materials, orthophosphate and also flow conditions on chloramine decay and

NDMA formation. Additional tests to investigate monochloramine degradation and NDMA

formation as well as the potential of iron, lead and their corrosion products to degrade HAA

results were initiated but not completed to the same extent as the other chapters, and thus the

results are presented in the Appendices 10.8 and 10.9, respectively.

In general, each results chapter is a separate manuscript that has been submitted for peer-

reviewed journal publication. At the time of writing, some contents from Chapters 3, 4 and 5

have been published as:

Zhang, H., Andrews, S. A. (2012) Effects of Phosphate-based Corrosion

Inhibitors on the Kinetics of Chlorine Degradation and HAA Formation in

Contact with Three Metal Materials. Canadian Journal of Civil Engineering, 39,

44-54.

Zhang, H., Andrews, S.A. (2012) Catalysis of Copper Corrosion Products on

Chlorine Decay and HAA Formation in Simulated Distribution Systems. Water

Research, 46 (8), 2665-2673.

The contents from Chapters 6 and 7 have been submitted as:

Zhang, H., Andrews, S.A. Factors Affecting Copper Catalysis during NDMA

Formation from DMA in Simulated Premise Plumbing. Water Research.

6

Zhang, H., Andrews, S.A. Effects of Pipe Materials, Orthophosphate, and Flow

Conditions on Chloramine Decay and NDMA Formation in Modified Pipe Loops.

Journal of Water Supply: Research and Technology – Aqua.

In addition, a separate Methods and Materials chapter is not included in this thesis

because details of the methods and materials employed in each experiment have been explained

in the relevant results chapters/papers. Similarly, while a brief literature review is included as

Chapter 2 to provide an overall context for the thesis, the papers in the results chapters contain

additional references as appropriate. QA/QC protocols followed in all experiments are

summarized in Chapter 10 (Appendices 10.2).

All the work and writing presented in this thesis were done by the author, with review

and editing by Professor Susan Andrews. Metal coupon surface analyses (Chapters 4 and 7) were

performed by Srebri Petrov from the Department of Chemistry, University of Toronto and Peter

Broderson from the Department of Chemical Engineering, University of Toronto.

1.4 References

Al-Jasser, A.O. (2007) Chlorine decay in drinking-water transmission and distribution systems:

Pipe service age effect. Water Research, 41(2), 387-396.

Boulay, N., and Edwards, M. (2001) Role of temperature, chlorine, and organic matter in copper

corrosion by-product release in soft water. Water Research, 35(3), 683-690.

California Department of Public Health (CDPH). California Drinking Water: NDMA and Other

Nitrosamines - Drinking Water Issues. http://www.cdph.ca.gov/certlic/drinkingwater/

Pages/NDMA.aspx (accessed September 15, 2011).

Cantor, A. F., Park, J. K., and Vaiyavatjamai, P. (2003) Effect of chlorine on corrosion in

drinking water systems. Journal American Water Works Association, 95(5), 112-123.

Choi, J., Duirk, S. E., and Valentine, R. L. (2002) Mechanistic studies of N-

nitrosodimethylamine (NDMA) formation in chlorinated drinking water. Journal of

Environmental Monitoring, 4(2), 249-252.

http://www.cdph.ca.gov/certlic/drinkingwater/%20Pages/


7

Haas, C.N., Gupta, M., Chitluru, R. and Burlingame, G. (2002) Chlorine demand in disinfecting

water mains. Journal of American Water Works Association, 94(1), 97.

Hallam, N.B., West, J.R., Forster, C.F., Powell, J.C. and Spencer, I. (2002) The decay of chlorine

associated with the pipe wall in water distribution systems. Water Research, 36(14),

3479-3488.

Health Canada (2008) Guidelines for Canadian Drinking Water Quality: Guideline Technical

Document — Haloacetic Acids., Water, Air and Climate Change Bureau, Healthy

Environments and Consumer Safety Branch, Health Canada, Ottawa, Ontario.

Health Canada (2011). Guidelines for Canadian Drinking Water Quality: Guideline Technical

Document- N-Nitrosodimethylamine (NDMA). Water, Air and Climate Change Bureau,

Healthy Environments and Consumer Safety Branch, Health Canada, Ottawa, Ontario.

(Catalogue No H128-1/11-662E).

Lytle, D. A., Sarin, P., and Snoeyink, V. L. (2005) The effect of chloride and orthophosphate on

the release of iron from a cast iron pipe section. Journal of Water Supply Research and

Technology-Aqua, 54(5), 267-281.

McNeill, L.S. and Edwards, M. (2001) Iron pipe corrosion in distribution systems. Journal of

American Water Works Association, 93(7), 88-100.

MacQuarrie, D. M., Mavinic, D. S., and Neden, D. G. (1997) Greater Vancouver Water District

drinking water corrosion inhibitor testing. Canadian Journal of Civil Engineering, 24(1),

34-52.

Ministry of the Environment (MOE). Safe Drinking Water Act 2002. Ontario Regulation 169/03,

Schedule 2. http://www.e-laws.gov.on.ca/html/regs/english/elaws_regs_030169_e.htm

(accessed September 15, 2011)

Najm, I., and Trussell, R. R. (2001) NDMA formation in water and wastewater. Journal


http://www.e-laws.gov.on.ca/html/regs/english/elaws_regs_030169_e.htm

8

Rahman, S., McDonald, B. C., and Gagnon, G. A. (2007) Impact of secondary disinfectants on

copper corrosion under stagnation conditions. Journal of Environmental Engineering-

Asce, 133(2), 180-185.

Renner, R. (2004) Plumbing the depths of DC's drinking water crisis. Environmental Science &

Technology, 38(12), 224A-227A.

Rossman, L. A., Brown, R. A., Singer, P. C., and Nuckols, J. R. (2001) DBP formation kinetics

in a simulated distribution system. Water Research, 35(14), 3483-3489.

Sarin, P., Clement, J.A., Snoeyink, V.L. and Kriven, W.W. (2003) Iron release from corroded

unlined cast-iron pipe. Journal American Water Works Association, 95(11), 85.

USEPA (2006) 40 CFR Parts 9, 141, and 142 National Primary Drinking Water Regulations:

Stage 2 Disinfectants and Disinfection Byproducts Rule. Federal Register 71(2), 387-493.

USEPA (2008) Lead and Copper Rule: A Revised Quick Reference Guide. http://water.epa.gov/

lawsregs/rulesregs/sdwa/lcr/upload/LeadandCopperQuickReferenceGuide_2008.pdf

(accessed October, 2011).

Xiao, W.Z., Hong, S.K., Tang, Z.J. and Taylor, J.S. (2007) Effects of blending on total copper

release in distribution systems. Journal American Water Works Association, 99(1), 78-88

http://water.epa.gov/%20lawsregs/rulesregs/sdwa/lcr/


9

Literature Review 2

To date, considerable efforts have been made to address the protection of water quality in

distribution systems, mainly by optimizing in-plant treatment processes. Strategies to stabilize

disinfectant residual while simultaneously controlling DBP formation include water quality

adjustment and DBP precursor removal. Fewer studies have been conducted on the fate of DBPs,

especially haloacetic acids (HAA9) and nitrosamines, in distribution systems, under the

interactive impacts of disinfectant residual and the internal pipe environment. This literature

review begins with a brief review of disinfectant chemistry and its application in distribution

systems. The mechanisms responsible for HAA9 and nitrosamine (especially N-

nitrosodimethylamine, NDMA) formation and influential factors associated with water quality

and distribution system characteristics are summarized. Finally, other aspects of water quality

deterioration in distribution systems resulting from disinfectant application are reviewed

(corrosion and nitrification).

2.1 Disinfectants

Chemistry of Chlor(am)ination 2.1.1

Free and combined chlorine are commonly used as secondary disinfectants to reduce the

possible occurrence of biological regrowth. The dominance of free chlorine species (HOCl, OCl-

and Cl2) is dependent on pH. Under the typical pH range in distribution systems, hypochlorous

acid (HOCl) and hypochlorite ion (OCl-) are the main chlorine species. Normally, HOCl is

predominant at pH lower than 7.5 (the pKa of HOCl), whereas OCl- is the main species at higher

pH (> 7.5). Reactions that free chlorine can undergo in drinking water treatment generally

involve oxidation and substitution. The significant reactions of chlorine are summarized in Table

2-1.

10

Table 2-1 Chlorine reactions in drinking water treatment

Reaction type Examples

Ammonia substitution NH3+HOClNH2Cl+H2O

NH2Cl+HOClNHCl2+H2O

Inorganic oxidation

Mn2+

+HOCl+2H2O→MnO(OH)2+3H++Cl

-

2Fe2+

+HOCl+5H2O→2Fe(OH)3+5H++Cl

-

NO2-+HOCl+2H2O→NO3

-+H

++Cl

-

HSO3-+HOCl→SO4

2-+2H

++Cl

-

HS-+HOCl→S+2H

++H2O +Cl

-

Br-+HOCl→HOBr +Cl

-

Organic reaction oxidation RCHO+HOCl→RCOOH +H

++Cl

-

RCOOH+HOCl→R+CO2+ H++Cl

-

Decomposition 3HOCl→3H

++2Cl

-+ClO3

-

2HOCl→2H++2Cl

-+O2

(Source: Dickenson, 2005)

In the presence of ammonia, free chlorine reacts rapidly and in a stepwise manner to form

chloramine compounds, including monochloramine (NH2Cl), dichloramine (NHCl2), and

trichloramine (NCl3). The application of NH2Cl has received increasing attention due to

requirements in the U.S. to comply with the Stage 2 Disinfectant/Disinfection Byproducts Rule

(D/DBPR) and the need to decrease chlorinated DBPs. NHCl2 and NCl3 formation in drinking

water are avoided due to odor and taste problems. The major reactions with regard to the

complex chemistry of chloramines have been summarized in Table 2-2.

Usually, NH2Cl is the dominant species of chloramines. Its rapid formation is observed at

a pH value of approximately pH 8.0 and a Cl2/N mass ratio of less than 5:1, which represents

common conditions for drinking water chloramination. A Cl2/N mass ratio from 3:1 to 5:1,

typically 4:1, is currently used to form NH2Cl to minimize the risks of NHCl2 and DBP

formation while reducing nitrification and biofilm growth associated with the excess of NH3

(USEPA, 1999). NHCl2 is formed at higher Cl2/N ratios (mass ratio >5:1) or at low pH values

(pH <5). Its formation can also be catalyzed by the presence of acetic acid, phosphate, carbonate,

and silicate species (Schreiber and Mitch, 2007; Trofe et al., 1980; Valentine et al., 1988;

Vikesland et al., 2001).

11

Table 2-2 Primary reactions of monochloramine

Reaction Stoichiometry Rate/Equilibrium

Constant (25 ºC) Reaction Stoichiometry

Rate/Equilibrium

Constant (25 ºC)

HOCl+NH3→NH2Cl+H2O k1=1.5X1010

M-1

h-1 b

I+ NHCl2→HOCl+ N2+ HCl k9=1.0 X 108M

-1h

-1

NH2Cl+H2O→HOCl+NH3 k2=7.6 X 10-2

h-1 b

I+ NH2Cl→N2+ H2O +HCl k10=3.0 X 107M

-1h

-1

HOCl+ NH2Cl→NHCl2+ H2O k3=1.0 X 106M

-1h

-1 NH2Cl+H+NH3Cl

+ k11=28M-1

NHCl2+ H2O→HOCl+ NH2Cl k4=2.3X X 103h

-1 NH3Cl++Br

- NH3Br

++Cl

- k12=1.8 X 108M

-1h

-1

NH2Cl + NH2Cl→NHCl2+NH3 ak5=pH dependent NH3Br

+NH2Br+H

+ pKa=6.4

NHCl2+NH3→NH2Cl + NH2Cl k6=2.16 X 108M

-2h

-1 HOCl+Br-HOBr+Cl

- k13=5.6 X 106M

-1h

-1

NH2Cl + NHCl2→N2+3H++3Cl

- k7=55.0M-1

h-1 HOBr+NH3NH2Br+H2O k14=2.7 X 10

11M

-1h

-1

NHCl2+H2O→bI+2HCl k8=4.0 X 10

5M

-1h

-1 HOBr+NH2ClNHBrCl k15=1.0 X 109M

-1h

-1

te.intermedia reactive :I .C 25 at ,hM800k and

,hM100.4k ,hM105.2k :]HCO[k]COH[k]H[kk :Note

b12

3HCO

1233CO2H

127

H33HCO323CO2HH5

a

(Source: Vikesland et al., 2001)

Disinfectant Residual Reactions in Distribution Systems 2.1.2

Both free chlorine and NH2Cl experience temporal and spatial degradation in distribution

systems in the bulk water and at the pipe wall due to chemical and biological reactions.

Therefore, both the bulk water and the pipe wall create a demand for the disinfectant residual.

The bulk water demand comes from dissolved organic matter, ammonia and nitrite, and some

dissolved metallic compounds (e.g. ferrous ions and manganese). NH2Cl may also spontaneously

decay in the absence of other reactants by auto-decomposition, which includes a complex series

of reactions and accounts for approximately half of the NH2Cl loss (Duirk et al., 2005). The

stoichiometry of NH2Cl auto-decomposition can be characterized by the generalized reaction

(Vikesland et al., 1996) of:

3NH2Cl N2 + NH3 +3Cl- + 3H

+ Equation 2-1

Auto-decomposition is impacted by pH and Cl2/N ratio. It is recommended that the pH remain

above 8.3 and the Cl2/N mass ratio be approximately 4:1 to maintain a chloramine residual in

distribution systems (Wilczak et al., 2003a).

The pipe wall demand of the disinfectant residual is caused by biofilms, deposits of

corrosion byproducts, and organic matter adsorbed onto metal oxide or carbonate scales (Lu et

12

al., 1999; Rossman et al., 2001; Vikesland and Valentine, 2002). The reactions of chlorine with

the scales on the inner pipe surface are regarded as the main reason for the loss of secondary

disinfectant residual within distribution networks (DiGiano and Zhang, 2005). Vikesland and

Valentine (2002) have reported that iron corrosion products can catalyze monochloramine

degradation, and the catalytic activities of the oxides can be ranked in the following order:

magnetite > goethite (α-FeO(OH)) > lepidocrocite (-FeO(OH)) hematite (Fe2O3) >

ferrihydrite. Lu et al. (1999) observed that at steady state biofilms chlorine demand increased

with increasing bioeliminable dissolved organic carbon (BDOC) of the water and the

surface/volume ratio of the pipes. However, Hallam et al. (2002) concluded in their research that

the biofilm could slow down chlorine consumption because it prevented chlorine reaching the

pipe surface. Regarding the impacts of organic matter, Rossman et al. (2001) reported that

although the rate of chlorine consumption in the pipe was much greater than in the bottle, the

increased formation of haloacetic acids and trihalomethanes (by 15%) was observed due to the

reactions of chlorine with organic precursor materials associated with deposits on the pipe wall.

Substantial research to date either in the field or in the laboratory has concluded that the

rates of disinfectant residual degradation in distribution systems increase with increasing

temperature, initial disinfectant concentration, total organic matter (TOC), the presence of

corrosion product and residence time (Brereton and Mavinic, 2002; Lu et al., 1999; Vikesland

and Valentine, 2002; Wable et al., 1991). Generally, NH2Cl is much less reactive with corrosion

products compared to free chlorine, and thus the demand exerted by the corrosion products

occurs more slowly than with chlorine (Valentine et al, 2000). However, in chloraminated

distribution systems, accelerated chloramine decay has been observed in systems with high

levels of nitrification (Wolfe et al., 1990; Woolschlager et al., 2001). It is primarily due to the

redox reaction between monochloramine and nitrite, and the reaction can be described as:

NH2Cl + NO2- + H2O NO3

- + NH3 + HCl Equation 2-2

Based on the stoichiometric relationship between nitrite and NH2Cl as shown in Equation 2-2,

each mg/L of NO2--N will consume as much as 5 mg/L as Cl2 of residual (Zhang and Edwards,

2009).

13

Kinetic Models for Disinfectant Residual Decay in Distribution Systems 2.1.3

Disinfectant residuals in distribution systems are consumed by soluble components

within the bulk liquid phase (bulk decay) and materials associated with the pipe wall such as the

pipe surface, corrosion products and the biofilm (wall decay). Therefore, the overall disinfectant

degradation in distribution systems can be expressed as:

Equation 2-3

Some researchers have postulated that chlorine decay models could be based on first-order or

parallel first-order kinetics for bulk reactions, and, either zero-order or first-order kinetics for

pipe wall reactions (Digiano and Zhang, 2005; Kiéné et al., 1998; Wable et al. 1991;

Vasconcelos et al., 1997).

For chlorine decay in the bulk flow, Kiéné et al. (1998) described the kinetic equation as:

CKdt

dCb

Bulk

Equation 2-4

where Kb =bulk decay constant; C =chlorine concentration in the bulk. The kinetic constants for

reactions are dependent on the water quality, specifically on temperature and organic content of

the water (Kiéné et al., 1998). Therefore, Kb can be described as:

T

b

b eTOCaK

Equation 2-5

where a =1.8X106L/mg·h; b =6050; T =temperature; TOC represents the organic content.

For disinfectant loss as a result of pipe wall consumption, a film resistance model with a

first-order rate equation has been proposed by Rossman et al. (1994) as:

Equation 2-6

WallBulkTotal dt

dC

dt

dC

dt

dC

W1W

Wall

CKV

S

dt

dC

14

where S = surface area; V = pipe volume; KW1 = wall reaction rate constant; CW = disinfectant

concentration at the pipe wall. According to this model (Rossman et al., 1994), mass transfer of

chlorine towards the pipe wall is proportional to the difference in chlorine concentration between

the bulk liquid and the pipe wall. If it is assumed that chlorine reacts at the pipe wall and there is

no accumulation of chlorine at the wall, the rate of mass transfer of chlorine would be equal to

the rate of chlorine decay at the pipe wall. Therefore, the flux of chlorine to the wall can be

written as:

Equation 2-7

where KF = mass transfer coefficient; C = disinfectant concentration in the bulk. Solving

Equation 2-7 for CW and substituting it into Equation 2-6 gives:

Equation 2-8

where KW = overall wall decay constant combining the mass transfer constant, KF, and the wall

reaction constant KW1.

Rossman et al. (1994) also provided a relationship between the mass transfer coefficient,

KF (Equation 2-7) and the dimensionless Sherwood number (Sh) as:

Equation 2-9

Equation 2-10

3. 0.0 8(

)

1+0.0 (

)

2 3 for Re 2,300 Equation 2-11

Equation 2-12

Equation 2-13

)CC(KCK WFW1W

CK)V

S(C

)KK(

KK)

V

S(

dt

dCW

F1W

F1W

Wall

p

hFd

DSK

2,300Refor ;Re023.0 333.083.0 ScSh

v

ud pRe

D

vSc

15

where Sh = Sherwood number; dp = pipe diameter; Re = Reynolds number; Sc = Schmidt

number; L= pipe length; D = molecular diffusivity of chlorine in water; u = flow velocity in pipe;

v = kinematic viscosity of water. Based on Equations 2-9 to 2-13, a positive relationship between

disinfectant decay rates and flow velocity would be expected. Flow velocity can influence

turbulence, radial diffusion and boundary layer thickness by affecting the Reynolds and

Sherwood numbers. This helps to explain why some studies have shown that the rates of free

chlorine and chloramine decay can increase with rising flow velocities in cast iron, unlined

ductile iron and even PVC pipes (Digiano and Zhang, 2005; Hallam et al., 2002; Mutoti et al.,

2007; Westbrook and Digiano, 2009). In addition to the mass transfer rate, some researchers

such as Mutoti et al. (2007) have attributed some of the increase in disinfectant decay at high

velocity to the increase in the release rate of corrosion products from the pipe surface.

Kb and KW as shown in Equations 2-4 and 2-8 will exhibit different values for the various

pipe materials due to the different pipe wall reaction kinetics. Table 2-3 summarizes some of

these coefficients for free chlorine and chloramine in contact with different pipe materials.

Table 2-3 Free chlorine and chloramine model coefficients

Material Parameter Free chlorine value Parameter Chloramine value

PVC Kb 0.048 h

-1 Kb 0.020 h

-1

KW 0.007 h-1

KW 0.007 h-1

Lined cast

iron

Kb 0.084 h-1

Kb 0.017 h-1

KW 0.007 h-1

KW 0.011 h-1

Unlined cast

iron

Kb 0.172 h-1

Kb 0.147 h-1

KW 0.063 h-1

KW 0.084 h-1

(Source: Taylor et al., 2005) Note: both bulk decay for chlorine and chloramine followed parallel first order kinetics,

and wall decay for both species followed first order kinetics.

Cast iron Kb 0.04 h

-1 First order

KW 36.5 mg/m2/h Zero order

Ductile iron Kb 0.04 h

-1 First order

KW 0.07-0.26 h-1

First order

(Source: Digiano and Zhang, 2005)

16

2.2 Disinfection Byproducts

Haloacetic Acids 2.2.1

Haloacetic acids (HAA) form by the chlorination of natural organic matter (NOM). Since

the 1970s, considerable efforts have been made to understand HAA formation mechanisms by

using NOM or well-defined model compound precursors (Kanokkantapong et al., 2006a;

Kanokkantapong et al., 2006b; Morris, 1975; Reckhow et al., 1990; Rook, 1977). Reaction

mechanisms involved in HAA formation generally include oxidation, substitution, addition, and

hydrolysis (Morris, 1975). Among the nine possible chlorinated and/or brominated HAA

compounds, dichloroacetic acid (DCAA) and trichloroacetic acid (TCAA) are the most

commonly reported HAA species. The dominance of these species is dependent on the chlorine

dose, as well as the water’s temperature, pH, and the specific ultraviolet absorbance (SUVA,

defined as the UV absorbance of a water sample at a given wavelength normalized for total

organic carbon concentration). Higher HAA concentrations are favoured at an increased chlorine

dose, temperature and SUVA and/or at low pH. Species that are generally low in concentration

in water include bromochloroacetic acid (BCAA), dibromoacetic acid (DBAA),

monochloroacetic acid (MCAA), monobromoacetic acid (MBAA), bromodichloracetic acid

(BDCAA), chlorodibromoacetic acid (CDBAA) and tribromoacetic acid (TBAA). Toxicological

properties of the different HAA species depend on the extent and type of halogen substitution

(bromine or chlorine). Guidelines for Canadian Drinking Water Quality established the

Maximum Acceptable Concentration (MAC) for HAA5 (MCAA, DCAA, TCAA, MBAA, and

DBAA) in drinking water at 80 µg/L (Health Health Canada, 2008). The Stage 2 D/DBPR

regulated the Maximum Contaminant Level (MCL) of HAA5 at 60 µg/L (USEPA, 2006).

N-nitrosodimethylamine 2.2.2

Although the application of chloramines has effectively reduced the formation of

halogenated DBPs, it may cause some unintended changes in water quality, including increased

formation of nitrosamines (Choi et al., 2002; Najm and Trussell, 2001). N-nitrosodimethylamine

(NDMA) is a suspected human carcinogen. Currently, federal drinking water guidelines for

nitrosamines are under development in Canada and the United States. Health Canada has

established the guideline for NDMA in drinking water at a Maximum Acceptable Concentration

17

(MAC) of 40 ng/L (Health Canada, 2011). The Ontario Ministry of the Environment, Canada,

has set a MAC for NDMA of 9 ng/L (MOE, 2002), and the California Department of Health

Services has set a notification level of 10 ng/L for each of NDMA, N-nitrosodiethylamine

(NDEA) and N-nitrosodi-n-propylamine (NDPA), respectively (CDHS, 2009). Mechanistic

studies of nitrosamines, especially NDMA, have been widely implemented with the

improvement of sensitive analytical techniques (Cheng et al., 2006; Mitch et al., 2003a). Several

reaction pathways have been investigated under controlled conditions as follows.

a. NDMA Formation from the Reaction of Monochloramine and Dimethylamine

Two steps are included in this mechanism: a nucleophilic substitution of dimethylamine

(DMA) by monochloramine (NH2Cl) to form an unsymmetrical dimethylhydrazine (UDMH) and

subsequent oxidation of UDMH intermediate to NDMA (Choi and Valentine, 2002; Mitch et al.,

2003b; Mitch et al., 2005; Mitch and Sedlak, 2002). The reactions are given in Equations (2-14)

and (2-15):

NH2Cl + (CH3)2NH → (CH3)2NNH2 + H+ +Cl

+ Equation 2-14

(CH3)2NNH2 + 2NH2Cl + H2O → (CH3)2NNO + 2NH3 + 2H++ 2Cl

- Equation 2-15

The rate of UDMH formation is slow and increases with pH (Yagil and Anbar, 1962),

whereas the UDMH oxidation occurs nearly instantaneously but with low yields (<1%) (Mitch

and Sedlak, 2002). The overall reaction rate is therefore controlled by the first step, and it

requires a long reaction time for NDMA formation. This may explain the increased NDMA

formation in distribution systems, especially in dead-end regions where long retention times are

expected.

Schreiber and Mitch (2007) proposed two other NDMA formation pathways involving

NHCl2: a relatively slow reaction of NHCl2 with amine precursors in the presence of dissolved

oxygen, and a fast reaction involving reactive breakpoint chlorination intermediates (HNO)

which was produced from the hydrolysis of NHCl2. Normally, NHCl2 can be minimized by

controlling pH (> 8.5) and the Cl2/N mass ratio (< 5:1) (Schreiber and Mitch, 2006). In addition,

UDMH

18

increasing the temperature accelerates the auto-decomposition of NH2Cl, thereby decreasing

NDMA formation (Mitch et al., 2003a).

b. NDMA Formation from Other Precursors

Humic substances in natural organic matter and tertiary amines with dimethylamine

functional groups have been suggested to serve as additional sources for nitrosamine precursors

(Chen and Valentine, 2006; Mitch et al., 2003b; Siddiqui and Atasi, 2001). Chen and Valentine

(2007) reported that hydrophilic fractions of natural organic matter appeared to form more

NDMA than hydrophobic fractions, and basic fractions tended to have a larger NDMA formation

potential than acid fractions. NDMA formation from tertiary amines is actually attributed to the

formation of secondary amines which contain dimethylamine functional groups through a

dealkylation process (Ellis and Soper, 1954). The primary amine (monomethylamine), the

quaternary amine (tetramethylamine), and amino acids or proteins do not form a significant

amount of NDMA after chloramination (Mitch and Sedlak, 2002). Strong-base anion exchange

resin and cationic polyelectrolytes (Epi-DMA and polyDADMAC products) used as flocculation

aids, which contain organic nitrogen, have also been observed to produce NDMA by reactions

with chloramine or hypochlorous acid (Gerecke and Sedlak, 2003; Najm and Trussell, 2001;

Westerhoff and Mash, 2002). Industrial products containing dimethylamine functional groups

including fungicides (e.g. tetramethylthiuram disulfide (thiram)), and drugs (e.g. ranitidine) have

also exhibited NDMA formation potential during disinfection of drinking water (Graham et al.,

1996; Schmidt et al., 2006).

Disinfection By-Products in the Distribution System 2.2.3

The concentration of HAA9 and nitrosamines varies both spatially and temporally within

distribution systems. Factors affecting DBP formation in distribution systems include the

concentration and chemical properties of the precursors, water temperature, pH, bromide ion

concentrations, disinfectant type, dose and residual, as well as contact time (Baribeau et al.,

2001; Baribeau et al., 2006; Cowman and Singer, 1996; Liang and Singer, 2003; Nguyen et al.,

2002; Obolensky and Frey, 2002; Rossman et al., 2001; Singer et al., 2002).

The HAA9 concentration in chlorinated systems either increases or remains

approximately constant as water age increases, while remaining relatively constant in

http://www.sciencedirect.com/science/article/pii/S0043135410006913#bbib38

19

chloraminated distribution systems (Baribeau et al., 2006). In both systems, the observed

decrease in HAA9 concentration following long residence time has been attributed primarily to

biodegradation (Chen and Weisel, 1998; Rodriguez et al., 2004). Research has also been

performed to investigate the impacts of pipe metals (mainly iron and copper) and their corrosion

byproducts on the occurrence of HAA9 in distribution systems. In most cases, less trichloroacetic

acid (TCAA) was present than dichloroacetic acid (DCAA) due to the consumption of

disinfectant residuals by these pipe metals and their corrosion byproducts (Chen and Weisel,

1998; Hassan et al., 2006; Li et al., 2008; Rossman et al., 2001; Tuovinen et al., 1984). It is

noteworthy that high chlorine losses that can be observed in some distribution systems due to

reactions with corrosion products will not necessarily reduce DBP formation rates. The organic

material sorbed to the pipe wall may actually increase DBP levels (Hassan et al., 2006; Korshin

et al., 1997; Rossman et al., 2001).

Direct reactions between pipe materials and HAA9 were reported by Holzalski et al.

(2001). They have observed that trichloro-, tribromo- chlorodibromo-, and bromodichloro- acetic

acid readily reacted with Fe0 via sequential hydrogenolysis (replacement of a halogen by

hydrogen). The reduction of HAA9 by Fe0

may account for some losses of HAA9 observed in

full-scale distribution systems where iron is abundant. However, when the percentage of Fe0 on

the pipe surface is far less than 10%, as might be in cases of old cast iron pipes, the degradation

of DBP by Fe0 may not be important (Zhang et al., 2004).

Limited data is available concerning the formation and fate of nitrosamines in full-scale

distribution systems. While no particular trend in NDMA concentration with increasing water

age has been observed in either chlorinated or chloraminated distribution systems by some

researchers (Baribeau et al., 2006), others have reported that NDMA production increased with

the retention time in distribution systems (Charrois et al., 2007; Wilczak et al., 2003b). Other

nitrosamines, including N-nitrosomorpholine (NMOR), N-nitrosopyrollidine (NPYR), N-

nitrosopiperidine (NPIP) and N-nitrosodiphenylamine (NDPhA), have also been detected in

some distribution systems at extreme ends (Zhao et al., 2006). Thus, residence time tends to

significantly affect the observed concentrations of nitrosamines in distribution systems. Two

studies have investigated the reduction potential of granular iron and nickel-enhanced iron on

NDMA, and they reported that iron can transform NDMA to dimethylamine (DMA) and

ammonium via catalytic hydrogenation (Gui et al., 2000; Odziemkowski et al., 2000). The

20

presence of a small amount of nickel (0.25%) plated onto the iron can greatly enhance the

NDMA transformation rate. However, it is uncertain if the same reaction can take place in

distribution mains where iron pipes are widely present and corroded. Schreiber and Mitch (2006)

have reported that copper may catalyze NDMA formation. However, details concerning factors

that impact copper catalysis and the catalytic potential of copper solid corrosion products on

NDMA formation are not available. Therefore, the role of iron, copper and other pipe metals as

well as their corrosion products on the fate of NDMA and other nitrosamines in distribution

systems needs to be further investigated.

2.3 Corrosion

Corrosion is a process that deteriorates materials through chemical reactions with their

environment. In drinking water distribution systems, the materials may be a metal pipe or fitting,

the cement in a pipe lining or a polyvinyl chloride (PVC) pipe. In recent years, considerable

efforts have been made to understand the mechanisms of metal release from pipes and corrosion

scales, the rates at which they occur under typical disinfectant residual concentrations and

hydraulic conditions, and effective strategies to control corrosion and/or metal release (e.g.

phosphate-based corrosion inhibitor addition).

Metal Dissolution and Precipitation in the Distribution System 2.3.1

Corrosion of metallic materials in distribution systems is an electrochemical process in

which the metallic material is oxidized to a cation at an anodic site according to the reaction of:

M Mn+

+ ne- Equation 2-16

The released electrons travel through the conducting electrolyte (e.g. water) to a site which acts

as the cathode. At the cathode, the most common reaction in distribution systems is the

acceptance of electrons by dissolved oxygen or by a disinfectant residual that is in contact with

the metal. The metal ions are then either released into the drinking water as corrosion products or

they may react with components present in the water (such as OH-, CO3

2-, Cl

-, and SO4

2-) to form

hydroxyl, carbonate, chloride, and sulfate complexes as well as solid phases that may precipitate

on the pipe surface. Table 2-4 summarizes the principal equilibrium reactions of copper in

drinking water.

21

Table 2-4 Relevant equilibrium reactions for copper in carbonated-bearing water

Log K* (25 °C)

A) Direct Oxidation of Copper Metal in Chlorinated Water

HOCl+Cu(s) = CuO(s) + H+ + Cl

- 31.78

HOCl+Cu(s)+H2O= Cu(OH)2(s) + H+ + Cl

- 30.5

HOCl+Cu(s)+HCO3-= CuCO3(s) + H2O + Cl

- 38.41

2HOCl+2Cu(s)+HCO3-= Cu2(OH)2(CO3)(s) + H

+ + 2Cl

- 71.84

B) Dissolution of Solid Phases

CuCO3(s) = Cu2+

+ CO32-

-11.5

Cu2(OH)2(CO3)(s) = 2Cu2+

+ 2OH- + CO3

2- -33.3

Cu(OH)2(s) = Cu2+

+ 2OH- -19.32

CuO+H2O = Cu2+

+ 2OH- -20.35

C) Metal-Ligand Complexation

Cu2+

+ CO32-

= CuCO3(aq) 6.77

Cu2+

+ 2CO32-

= Cu(CO3)22-

(aq) 10.2

Cu2+

+ OH- = CuOH

+(aq) 6.5

Cu2+

+ 2OH- = Cu(OH)2(aq) 11.8

Cu2+

+ Cl- = CuCl

+ 0.2

(Source: Hong and MacAuley, 1998)

Normally, tenorite [CuO], cupric hydroxide [Cu(OH)2], and malachite [Cu2CO3(OH)2]

are dominant solid phase for copper. Iron and lead corrosions in distribution systems are also

complex processes which may involve many equilibrium reactions between dissolution and

precipitation, between soluble complexes and ion pairs, and even redox reactions. The common

species found in the iron pipe passive layers include goethite [α-FeOOH], lepidocrocite [γ-

FeOOH], magnetite [Fe3O4], ferric hydroxide [Fe(OH)3]. As for lead, PbCO3 and

Pb3(CO3)2(OH)2 are primary solid phases in drinking water. When free chlorine is present,

divalent lead can be further oxidized to tetravalent lead dioxide (PbO2) (Switzer et al., 2006).

The dissolution of the dissolved metal from the solid phase is controlled by the solubility of the

solid species, the concentration of ions between the pipe wall and bulk liquid, flow velocity, and

the water quality. The influence of some of these factors will be reviewed in Section 2.3.2.

Factors Affecting Metal Levels at the Tap 2.3.2

pH and alkalinity are the most significant factors influencing metal levels in distribution

systems. pH will influence the solubility of corrosion byproducts, thus affecting metal

22

concentrations at the tap. Normally, lead, copper and iron solubility decreases with elevated pH

(McNeill and Edwards, 2001; Sarin et al., 2003; Schock and Gardels, 1983; Schock, 1989; Xiao

et al., 2007). In the neutral pH range (i.e., pH from 6 to 8), the equilibrium concentration of lead

could vary by a factor of at least 5 to 10 per pH unit (Schock and Wagner, 1985). Solubility

models show that the lowest lead levels occur when pH is around 9.8 (Schock, 1989; Schock and

Gardels, 1983). Examination of copper levels under the USEPA Lead and Copper Rule from the

experience of 361 utilities revealed that the average 90th-percentile copper levels were highest in

waters with pH below 7. and that no utilities with pH above 7.8 exceeded the USEPA’s action

level for copper of 1.3 mg/L (Dodrill and Edwards, 1995). For iron, Sarin et al. (2003) have

reported that increasing pH from 7.6 to 9.5 can reduce the amount of iron released to water (<

0.25 mg/L).

The degree to which alkalinity affects the solubility of iron, copper and lead is different.

Lower iron corrosion rates and iron concentrations in distribution systems have been associated

with higher alkalinity due to the formation of the less soluble siderite species (FeCO3, Ksp =

2.110-11

] (Sarin et al., 2003). However, it has been demonstrated in many laboratories and in

utilities that the release of copper corrosion products is worse at higher alkalinity due to the

formation of soluble cupric bicarbonate and carbonate complexes (Edwards et al., 1996; Knox et

al., 2005; Schock et al., 1995). For lead, the impact of alkalinity is dependent on the form of lead

carbonate present on the pipe surface. In a study by Schock (1990), when PbCO3 (Ksp =7.4

10-14

) was present, increasing alkalinity above 50 mg/L CaCO3 reduced lead solubility in the pH

range of approximately 6 to 7.3. The author also found that when Pb3(CO3)2(OH)2 (Ksp =10-18.8

)

was present, lead solubility would not be significantly affected by alkalinity in this pH range, but

the solubility would decrease if the pH was increased above 7.3.

Limited research has been conducted on the role of disinfectant type on corrosion.

LeChevallier et al., 1990) reported that free chlorine was more corrosive to iron than NH2Cl in

deionized water with extremely low alkalinity and hardness. Cantor et al. (2003) also found an

increased iron corrosion rate in an iron pipe loop in the presence of free chlorine compared to

NH2Cl. Free chlorine residual was also shown to increase copper corrosion at low pH because of

the increased oxidizing strength of HOCl (Boulay and Edwards, 2001; Cantor et al., 2003;

Rahman et al., 2007). Limited information has been reported concerning the effect of chloramine

on copper. Rahman et al. (2007) reported that the application of NH2Cl could reduce the

23

dissolution of copper pipes from 1.26 mg/L in the presence of free chlorine to 0.48 mg/L.

However, ammonia released from chloramine decomposition can be aggressive to copper and

copper alloys by forming stable copper-ammonia complexes (Cu(NH3)22+

, Log K at 25 °C =

7.47) (Boyd et al., 2009; Schock and Lytle, 1995). The high oxidation potential of free chlorine

is sufficient for the formation of lead dioxide [PbO2] (Ksp =10-66

). In contrast, divalent lead

solids, primarily in the form of Pb3(CO3)2(OH)2 (hydrocerussite, Ksp =10-18.8

), are favored under

chloramination conditions (Vasquez et al., 2006). In the fall of 2003 in Washington DC, a high

level of Pb2+

(48000 ppb) was detected in their distribution systems following their shift from

free chlorine to NH2Cl in 2000 (Renner, 2004). It was theorized that free ammonia combined

with a high concentration of nitrate may have synergistically driven the lead corrosion by

interfering with the formation of the passive layer on its metal surface (Edwards and Dudi, 2004;

Uchida and Okuwaki, 1998; Zhang et al., 2009a). The direct reaction of metallic lead with nitrate

can be illustrated with the formula:

NO3- + Pb NO2

- + PbO Equation 2-17

Corrosion Control 2.3.3

Released metal ions and particles may affect aesthetic quality of delivered water and/or

pose potential risk to human health. Therefore, for metal pipes, the drinking water guidelines

have established the maximum acceptable concentration (MAC) of lead, based on its health

effects in children, at 0.010 mg/L. For iron and copper, based on aesthetic considerations, their

MACs are set at 0.3mg/L and 1.0 mg/L, respectively (Health Canada, 2009).

To control metal corrosion, the addition of phosphate-based chemicals has been used as

one effective strategy. Phosphate-based corrosion inhibitors can be dosed as either

orthophosphoric acid, combinations of orthophosphoric acid and zinc orthophosphate,

polyphosphates, or blends of orthophosphoric acid and polyphosphate. The effectiveness of

orthophosphate as a corrosion control measure relies on the development of impervious solid

films to act as a barrier scale between the corroding metal and the water (Demora and Harrison,

1984; Singley, 1994). However, polyphosphates have a different mode of action. Polyphosphates

are strong chelating agents which actually reduce scaling and have been shown to be effective in

sequestrating Fe2+ ions to treat “red water” (Facey and Smith, 199 ; Maddison et al., 2001;

24

Williams, 1990). Several authors have demonstrated that either orthophosphate or polyphosphate

can reduce iron levels and prevent iron corrosion (Facey and Smith, 1995; Lytle and Snoeyink,

2002; Maddison et al., 2001; Sarin et al., 2003; Williams, 1990).

Orthophosphate may reduce copper release in short term by blocking active sites on the

copper surface and thus protecting metallic copper against oxidation (Dartmann et al., 2004).

However, in long term, orthophosphate has been shown to interfere with the natural development

of malachite scales [Cu2(OH)2CO3], leading to an increased copper concentration in the water

(Cantor et al., 2000). As well, the performance of polyphosphate at a practical dosage (about 1

mg/L as P) on copper corrosion control can be unpredictable. It has been shown to either slightly

decrease or increase copper corrosion (Cantor et al., 2000; Colling et al., 1992; Edwards et al.,

2002).

Orthophosphate and polyphosphate have opposite effects on lead solubility. The

application of orthophosphate can decrease lead release through the formation of relatively

insoluble hydroxypyromorphite [Pb5(PO4)3OH] scales (Edwards et al., 1999). A 70% reduction

in lead release by orthophosphate has been reported (Edward and McNeill, 2002). In contrast,

polyphosphate might increase lead leaching into the water by forming lead polyphosphate

complexes (Leroy, 1993; MacQuarrie et al., 1997; Schock and Wagner, 1985).

2.4 Nitrification

It has been reported that nearly two thirds of chloraminated utilities in the United States

experience nitrification episodes (Wilczak et al., 1996). As a result of NH2Cl application as a

secondary disinfectant, ammonia is present in distribution systems and can be converted to

nitrate (NO3-) by ammonia-oxidizing bacteria (AOB) and nitrite-oxidizing bacteria (NOB)

through the formation of the nitrite (NO2-).

Nitrification in a distribution system can cause many potential water quality problems

(Wilczak et al., 1996; Zhang et al., 2009b), including:

the depletion of disinfectant residual;

DBP formation due to mitigation techniques;

the reduction in pH and alkalinity;

25

the formation of nitrite and nitrate.

and the increased autotrophic and heterotrophic microorganism growth.

Although the reduction in pH and alkalinity is not a potential threat to public health, it could

theoretically violate the USEPA Lead and Copper Rule (2008) and cause the dissolution of lead

and copper to unacceptably high levels if these designated water quality parameters fail to be

maintained. The simultaneous occurrence of nitrification and copper corrosion has been

observed within household plumbing systems (USEPA, 2002).

Factors affecting nitrification include pipe materials, phosphate, nutrients, pH,

disinfectant types and residuals, and temperature (Wilczak et al., 1996; Zhang and Edwards,

2009; Zhang et al., 2009a; Zhang and Edwards, 2010b; Zhang et al., 2008). Different pipe

materials can affect nitrification by providing a source of trace nutrients, acting as a toxic metals,

providing a support for attached growth and consuming disinfectant (Zhang et al., 2009b). A low

level of copper (1-10 ppb) from corrosion may facilitate nitrification, whereas an increase in

copper concentration (> 100 ppb) could inhibit nitrification (Zhang and Edwards, 2010b). Lead

materials are capable of producing nitrite and ammonia via electrochemical reactions with nitrate

(Edwards and Dudi, 2004; Uchida and Okuwaki, 1998; Zhang et al., 2009b), and the released

ammonia can support nitrification activity. Therefore, lead surfaces appear to be favored by

nitrifiers rather than other surfaces (Zhang et al., 2008). As such, increased lead corrosion rates

may stimulate nitrification by recycling ammonia from nitrate. Corroded iron surfaces are

favored by nitrifying bacteria because they can provide essential micronutrients for nitrifier

growth (Morton et al., 2005). Iron corrosion can also accelerate chloramine decay and release

ammonia to support the nitrifier population (Odell et al., 1996).

Water treatment practices to control nitrification include altering the chlorine-to-

ammonia dosing ratio and increasing the initial disinfectant dose. During chloramination, the

Cl2/N molar ratio is typically maintained between 0.6 to just below 1.0. Increasing the Cl2/N

ratio has been found to effectively mitigate the potential of nitrification by minimizing the

available ammonia to as close as possible to 1:1 (Lieu et al., 1993). Since free chlorine residual

is more effective in inactivating AOB than chloramine, periodic breakpoint chlorination is often

used to control nitrification in distribution systems (Odell et al., 1996; Wolfe et al., 1990).

Nitrifying bacteria have slow growth rates, and thus grow best with an extended detention time,

26

e.g. in large reservoirs or in dead-end regions of distribution systems. Therefore, operational

activities that shorten water age may be considered to control nitrification, including increased

daily turnover of water in reservoirs/storage tanks, installation of recirculation facilities, and the

use of small diameter pipes (Odell et al., 1996).

2.5 Summary of Research Gaps

The water industry’s move to consider changing the secondary disinfectant from free

chlorine to chloramine can have negative impacts on the delivered drinking water quality,

including unintended nitrosamine formation, nitrification and increased corrosion rates (e.g.

lead). However, the literature information regarding the behavior of HAA9 and nitrosamine

(especially NDMA) concentrations in distribution systems, particularly with respect to the

influence of pipe wall materials, the presence of corrosion inhibitors, flow conditions, and, to

some extent, biofilm formation, is very limited and inconsistent. Therefore, further studies are

needed to understand the fate (formation and/or degradation) of these DBPs in the complex

physiochemical and biological environment of distribution systems. Several questions that could

be asked include:

Are there differences in disinfectant degradation and HAA/NDMA formation for

different pipe materials treated with different phosphate-based corrosion

inhibitors?

Do metal age and water quality affect the efficacy of corrosion inhibitors and

disinfectant decay as well as HAA/NDMA formation?

How do flow conditions potentially affect secondary disinfectant stability and

NDMA formation?

What are the impacts of solid corrosion products on the fate of HAA/NDMA in

distribution systems?

Therefore, the current research was implemented to at least partially answer these questions and

bridge the aforementioned gaps thus aiding the water treatment industry to develop strategies

that minimize water quality degradation.

27

2.6 References

Baribeau, H., Prevost, M., Desjardins, R., and Lafrance, P. (2001) Changes in chlorine and DOX

concentrations in distribution systems. Journal American Water Works Association,

93(12), 102-114.

Baribeau, H., Boulos, L., Hileselassie, H., Crozes, G., Singer, P. C., Nichols, C., Schleisinger, S.

A., Gullick, R. W., Williams, S. L., Williams, R. L., Fountleroy, L., Andrews, S. A., and

Moffat, E. (2006) Formation and decay of disinfection byproducts in the distribution

system. Water Research Foundation and US EPA, Project # 2770, Denver, USA.

Boyd, G. R., Dewis, K. M., Korshin, G. V., Reiber, S. H., Schock, M. R., Sandvig, A. M., and

Giani, R. (2009) Effects of Changing Disinfectants on Lead and Copper Release (vol 100,

pg 75, 2008). Journal American Water Works Association, 101(2), 10-10.



Brereton, J.A. and Mavinic, D.S. (2002) Field and material-specific simulated distribution

system testing as aids to understanding trihalomethane formation in distribution systems.

Canadian Journal of Civil Engineering, 29(1), 17-26.

California Department of Public Health (CDPH) (2009). California Drinking Water: NDMA and

Other Nitrosamines - Drinking Water Issues.

http://www.cdph.ca.gov/certlic/drinkingwater/Pages/NDMA.aspx (accessed September

15, 2011).

Cantor, A. F., Denig-Chakroff, D., Vela, R. R., Oleinik, M. G., and Lynch, D. L. (2000) Use of

polyphosphate in corrosion control. Journal American Water Works Association, 92(2),

95-102.

Cantor, A.F., Park, J.K. and Vaiyavatjamai, P. (2003) Effect of chlorine on corrosion in drinking

water systems. Journal American Water Works Association, 95(5), 112-123.


28

Charrois, J. W. A., Boyd, J. M., Froese, K. L., and Hrudey, S. E. (2007) Occurrence of N-

nitrosamines in Alberta public drinking-water distribution systems. Journal of

Environmental Engineering and Science, 6(1), 103-114.

Chen, Z. and Valentine, R.L. (2006) Modeling the formation of N-nitrosodimethylamine

(NDMA) from the reaction of natural organic matter (NOM) with monochloramine.

Environmental Science & Technology, 40(23), 7290-7297.

Chen, Z. and Valentine, R.L. (2007) Formation of N-nitrosodimethylamine (NDMA) from humic

substances in natural water. Environmental Science & Technology, 41(17), 6059-6065.

Chen, W.J. and Weisel, C.P. (1998) Halogenated DBP concentrations in a distribution system.

Journal American Water Works Association, 90(4), 151-163.

Cheng, R.C., Hwang, C.J., Andrews-Tate, C., Guo, Y.B., Carr, S. and Suffet, I.H. (2006)

Alternative methods for the analysis of NDMA and other nitrosamines in water. Journal


Choi, J. and Valentine, R.L. (2002) A kinetic model of N-nitrosodimethylamine (NDMA)

formation during water chlorination/chloramination. Water Science and Technology,

46(3), 65-71.

Colling, J. H., Croll, B. T., Whincup, P. A. E., and Harward, C. (1992) Plumbosolvency effects

and control in hard waters. Journal of the Institution of water and Environmental

Management, 6(3), 259-268.

Cowman, G.A. and Singer, P.C. (1996) Effect of bromide ion on haloacetic acid speciation

resulting from chlorination and chloramination of aquatic humic substances.


Dartmann, J., Alex, T., Dorsch, T., Schevalje, E., and Johannsen, K. (2004) Influence of

decarbonisation and phosphate dosage on copper corrosion in drinking water systems.

Acta Hydrochimica Et Hydrobiologica, 32(1), 25-32.

Demora, S. J., and Harrison, R. M. (1984) Lead in tap water - contamination and chemistry.

Chemistry in Britain, 20(10), 900-904.

29

Dickenson E. R.V. (2005) Short-term disinfection byproduct formation in a chlorine/chloramine

approach: natural waters and model compounds. PhD Dissertation, University of

Colorado

Digiano, F. A., and Zhang, W. D. (2005) Pipe section reactor to evaluate chlorine-wall reaction.


Dodrill, D.M. and Edwards, M. (1995) Corrosion control on the basis of utility experience.

Journal of American Water Works Association, 87(7): 74–85.

Duirk, S. E., Gombert, B., Croue, J. P., and Valentine, R. L. (2005) Modeling monochloramine

loss in the presence of natural organic matter. Water Research, 39(14), 3418-3431.

Edwards, M., Schock, M.R. and Meyer, T.E. (1996) Alkalinity, pH, and copper corrosion by-

product release. Journal of American Water Works Association, 88(3): 81–94.

Edwards, M., Jacobs, S., and Dodrill, D. (1999) Desktop guidance for mitigating Pb and Cu

corrosion by-products. Journal American Water Works Association, 91(5), 66-77.

Edwards, M., Hidmi, L., and Gladwell, D. (2002) Phosphate inhibition of soluble copper

corrosion by-product release. Corrosion Science, 44(5), 1057-1071.

Edwards, M., and Dudi, A. (2004) Role of chlorine and chloramine in corrosion of lead-bearing

plumbing materials. Journal American Water Works Association, 96(10), 69-81.

Edwards, M., and McNeill, L. S. (2002) Effect of phosphate inhibitors on lead release from

pipes. Journal American Water Works Association, 94(1), 79-90.

Ellis, A.J. and Soper, F.G. (1954) Studies of N-halogeno-compounds. Part VI. The kinetics of

chlorination of tertiary amines. Journal of the Chemical Society, 1750-1755.

Facey, R. M., and Smith, D. W. 1995. Soft, low-temperature water-distribution corrosion:

Yellowknife, NWT. Journal of Cold Regions Engineering, 9(1), 23-40.

Gerecke, A.C. and Sedlak, D.L. (2003) Precursors of N-nitrosodimethylamine in natural waters.


30

Graham, J.E., S.A. Andrews, G.J. Farquhar and O. Meresz, (1996) Thiram as an NDMA

Precursor in Drinking Water Treatment. Proceedings, American Water Works

Association Annual Conference, (Water Research), Toronto, Ontario. pp. 15-27.

Gui, L., Gillham, R.W. and Odziemkowski, M.S. (2000) Reduction of N-nitrosodimethylamine

with granular iron and nickel enhanced iron. 1. Pathways and kinetics. Environmental

Science & Technology, 34(16), 3489-3494.

Hallam, N. B., West, J. R., Forster, C. F., Powell, J. C., and Spencer, I. (2002) The decay of

chlorine associated with the pipe wall in water distribution systems. Water Research,

36(14), 3479-3488.

Hassan, K.Z.A., Bower, K.C. and Miller, C.M. (2006) Iron oxide enhanced chlorine decay and

disinfection by-product formation. Journal of Environmental Engineering-Asce, 132(12),

1609-1616.

Health Canada (2008) Guidelines for Canadian Drinking Water Quality: guideline technical

document — haloacetic acids., Water, Air and Climate Change Bureau, Healthy


Health Canada (2009) Guidelines on Corrosion Control in Drinking Water Distribution Systems.

Federal-Provincial-Territorial Committee on Drinking Water, available from

www.healthcanada.gc.ca.





Hong, P. K. A., and MacAuley, Y. Y. (1998) Corrosion and leaching of copper tubing exposed to

chlorinated drinking water. Water air and Soil Pollution, 108(3-4), 457-471.

Hozalski, R.M., Zhang, L. and Arnold, W.A. (2001) Reduction of haloacetic acids by Fe0:

Implications for treatment and fate. Environmental Science & Technology, 35(11), 2258-

2263.

31

Kiéné, L., Lu, W. and Levi, Y. (1998) Relative importance of the phenomena responsible for

chlorine decay in drinking water distribution systems. Water Science and Technology,

38(6), 219-227.

Korshin, G. V., Li, C. W., and Benjamin, M. M. (1997) Monitoring the properties of natural

organic matter through UV spectroscopy: A consistent theory. Water Research, 31(7),

1787-1795.

LeChevallier, M. W., Lowry, C. D., and Lee, R. G. (1990) Disinfecting Biofilms in a Model

Distribution-System. Journal American Water Works Association, 82(7), 87-99.

Leroy, P. (1993) Lead in drinking water: Origins, solubility, treatment. Aqua (Oxford), 42(4),

233-238.

Li, B., Liu, R.P., Liu, H.J., Gu, J.N. and Qu, J.H. (2008) The formation and distribution of

haloacetic acids 9 in copper pipe during chlorination. Journal of Hazardous Materials,

152(1), 250-258.

Liang, L., and 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.

Lieu, N.I., Wolfe, R.L. and Means, E.G. (1993) Optimizing chloramine disinfection for the

control of nitrification. Journal of American Water Works Association, 85(2), 84-90.

Lu, W., Kiene, L. and Levi, Y. (1999) Chlorine demand of biofilms in water distribution

systems. Water Research, 33(3), 827-835.

Lytle, D. A., and Snoeyink, V. L. (2002) Effect of ortho- and polyphosphates on the properties of

iron particles and suspensions. Journal American Water Works Association, 94(10), 87-

99.

Kanokkantapong, V., Marhaba, T. F., Panyapinyophol, B., and Pavasant, P. (2006a) FTIR

evaluation of functional groups involved in the formation of haloacetic acids during the

chlorination of raw water. Journal of Hazardous Materials, 136(2), 188-196.

32

Kanokkantapong, V., Marhaba, T. F., Pavasant, P., and Panyapinyophol, B. (2006b)

Characterization of haloacetic acid precursors in source water. Journal of Environmental

Management, 80(3), 214-221.

Knox, G., Mavinic, D., Atwater, J. and MacQuarrie, D. (2005) Assessing the impact of corrosion

control measures on tap drinking water of the Greater Vancouver Regional District.




34-52.

Maddison, L. A., Gagnon, G. A., and Eisnor, J. D. (2001) Corrosion control strategies for the

Halifax regional distribution system. Canadian Journal of Civil Engineering, 28(2), 305-

313.

McNeill, L.S. and Edwards, M. (2001) Iron pipe corrosion in distribution systems. Journal of


Ministry of the Environment (MOE). Safe Drinking Water act 2002. Ontario Regulation 169/03,

Schedule 2. http://www.e-laws.gov.on.ca/html/regs/english/elaws_regs_030169_e.htm

(accessed September 15, 2011)

Mitch, W.A. and Sedlak, D.L. (2002) Formation of N-nitrosodimethylamine (NDMA) from

dimethylamine during chlorination. Environmental Science & Technology, 36(4), 588-

595.

Mitch, W.A., Gerecke, A.C. and Sedlak, D.L. (2003a) A N-nitrosodimethylamine (NDMA)

precursor analysis for chlorination of water and wastewater. Water Research, 37(15),

3733-3741.

Mitch, W.A., Sharp, J.O., Trussell, R.R., Valentine, R.L., Alvarez-Cohen, L. and Sedlak, D.L.

(2003b) N-nitrosodimethylamine (NDMA) as a drinking water contaminant: A review.

Environmental Engineering Science, 20(5), 389-404.

http://www.e-laws.gov.on.ca/html/regs/english/elaws_regs_030169_e.htm

33

Mitch, W.A., Oelker, G.L., Hawley, E.L., Deeb, R.A. and Sedlak, D.L. (2005) Minimization of

NDMA formation during chlorine disinfection of municipal wastewater by application of

pre-formed chloramines. Environmental Engineering Science, 22(6), 882-890.

Morris, J. C. (1975) Formation Of Halogenated Organics By Chlorination Of Water Supplies, a

Review. National Service Center for Environmental Publications (NSCEP), USEPA, N.

S. C. f. E. P. (NSCEP), ed., USEPA, Washington D.C.

Morton, S. C., Zhang, Y., and Edwards, M. A. (2005) Implications of nutrient release from iron

metal for microbial regrowth in water distribution systems. Water Research, 39(13),

2883-2892.

Mutoti, G., Dietz, J. D., Arevalo, J., and Taylor, J. S. (2007) Combined chlorine dissipation: Pipe

material, water quality, and hydraulic effects. Journal American Water Works

Association, 99, 96-106.

Najm, I. and Trussell, R.R. (2001) NDMA formation in water and wastewater. Journal


Nguyen, M. L., Baker, L. A., and Westerhoff, P. (2002) DOC and DBP precursors in western US

watersheds and reservoirs. Journal American Water Works Association, 94(5), 98-112.

Obolensky, A., and Frey, M. (2002) Distribution system DBP results and SDS performance. In:

Information Collection Rule Data Analysis. Prepared by McGuire, M.J., McLain, J.L.,

and Obolensky, A. Denver, Colo.: AWWARF and AWWA, 169-198.

Odell, L. H., Kirmeyer, G. J., Wilczak, A., Jacangelo, J. G., Marcinko, J. P., and Wolfe, R. L.

(1996) Controlling nitrification in chloraminated systems. Journal American Water

Works Association, 88(7), 86-98.

Odziemkowski, M.S., Gui, L. and Gillham, R.W. (2000) Reduction of N-nitrosodimethylamine

with granular iron and nickel-enhanced iron. 2. Mechanistic studies. Environmental


34



Asce, 133(2), 180-185.

Reckhow, D. A., Singer, P. C., and Malcolm, R. L. (1990) Chlorination of Humic Materials - by-

Product Formation and Chemical Interpretations. Environmental Science & Technology,

24(11), 1655-1664.



Rodriquez, M.J., Serodes, J.B. and Levallois, P. (2004) Behavior of trihalomethanes and

haloacetic acids in a drinking water distribution system. Water Research, 38(20), 4367-

4382.

Rook, J. J. (1977) Chlorination Reactions of Fulvic Acids in Natural-Waters. Environmental


Rossman, L.A., Clark, R.M. and Grayman, W.M. (1994) Modeling chlorine residuals in drinking

water distribution systems. Journal of Environmental Engineering-ASCE, 120(4), 803-

820.



Sarin, P., Clement, J.A., Snoeyink, V.L. and Kriven, W.W. (2003) Iron release from corroded

unlined cast-iron pipe. Journal American Water Works Association, 95(11), 85.

Schmidt, C.K., Sacher, F., Brauch, H.J. (2006) Strategies for minimizing formation of NDMA

and other nitrosamines during disinfection of drinking water. In: Proceedings of the

AWWA Water Quality Technology Conference, Denvor, C.O., November 5–9, 2006.

Schock, M.R. and Gardels, M.C. (1983) Plumbosolvency reduction by high pH and low

carbonate- solubility relationships. Journal American Water Works Association, 75(2),

87-91.

35

Schock, M.R. (1989) Understanding corrosion control strategies for lead. Journal American

Water Works Association, 81(7), 88-100.

Schock, M.R. (1990) Causes of temporal variability of lead in domestic plumbing systems.

Environmental Monitoring and Assessment, 15(1), 59-82.

Schock, M.R., Lytle, D.A. and Clement, J.A. (1995) Effect of pH, DIC, orthophosphate, and

sulfate on drinking water cuprosolvency. U.S. Environmental Protection Agency,

Cincinnati, OH (Report No, EPA/600/R-95/085).

Schock M, Wagner I. (1985) The corrosion and solubility of lead in drinking water. In: Internal

Corrosion of Water Distribution Systems. Denver: AWWA Research Foundation,

Denver.

Schreiber, I.M. and Mitch, W.A. (2006) Nitrosamine formation pathway revisited: The

importance of chloramine speciation and dissolved oxygen. Environmental Science &

Technology, 40(19), 6007-6014.

Schreiber, I.M. and Mitch, W.A. (2007) Enhanced nitrogenous disinfection byproduct formation

near the breakpoint: Implications for nitrification control. Environmental Science &

Technology, 41(20), 7039-7046.

Siddiqui M, Atasi K (2001) NDMA occurrence and formation – A review. Proceedings

American Water Works Association Annual Conference, Washington, DC.

Singer, P.C.; Winberg, H.S.; Brophy, K.; Liang, L., Roberts, M.; Grisstede, I.; Krasner, S. and

Baribeau, H., Arora, H. and Najm, I. (2002) Relative dominance of haloacitic acids and

trihalomethanes in treated drinking water. American Water Works Association Research

Foundation Report, Denver, CO.

Singley, J. E. (1994) Electrochemical nature of lead contamination. Journal American Water


Switzer, J.A., Rajasekharan, V.V., Boonsalee, S., Kulp, E.A. and Bohannan, E.W. (2006)

Evidence that monochloramine disinfectant could lead to elevated Pb levels in drinking

water. Environmental Science & Technology, 40(10), 3384-3387.

36

Tuovinen, O.H., Mair, D.M. and Banovic, J. (1984) Chlorine demand and trihalomethane

formation by tubercles from cast-iron water mains. Environmental Technology Letters,

5(3), 97-108.

Trofe, T. W., Inman, G. W., and Johnson, J. D. (1980) Kinetics of monochloramine

decomposition in the presence of bromide. Environmental Science & Technology, 14(5),

544-549.

Taylor, J.S, Dietz, D.J., and Randall, A.A. (2005) Effects of blending on distribution systems

water quality. AWWA Research Foundation Report. University of Central Florida,

Orlando

Uchida, M., and Okuwaki, A. (1998) Decomposition of nitrate by in situ buff abrasion of lead

plate. Hydrometallurgy, 49(3), 297-308.

USEPA (1999) Alternative Disinfectants and Oxidation Guidance Manual (EPA 815-R-99-014).

Office of Water, United States Environmental Protection Agency.

USEPA (2002) Nitrification. Office of Ground Water and Drinking Water, Standards and Risk

Management Division. Washington DC. Available at http://www.epa.gov/safewater/disin

-fection/tcr/regulation_revisions.html


Stage 2 Disinfectants and Disinfection Byproducts Rule. Federal Register, 71(2), 387-

493.

USEPA (2008) Lead and Copper Rule: A Revised Quick Reference Guide. http://water.epa.gov/

lawsregs/rulesregs/sdwa/lcr/upload/LeadandCopperQuickReferenceGuide_2008.pdf

(accessed October, 2011).

Valentine, R. L., and Jafvert, C. T. (1988) Genearl acid catalysis of monochloramine

disproportionation. Environmental Science & Technology, 22(6), 691-696.

Valentine, R.L., Vikersland, P.J., Angerman, B.D., Hackett, S.A., Shoup, M., Slattenow, S.

(2000) The role of pipe-water interface in DBP formation and disinfectant loss. American

Water Works Association Research Foundation, Denver, CO.

http://www.epa.gov/safewater/disin%20-fection/tcr/

http://www.epa.gov/safewater/disin%20-fection/tcr/



37

Vikesland, P., Valentine, R. and Ozekin, K. (1996). Application of product studies in the

eludication of chloramine reaction pathways. Water Disinfection and Natural Organic

Matter. R. Miner and G. Amy (eds.). American Chemical Society, Washington DC, pp.

115–125.

Vikesland, P.J. and Valentine, R.L. (2002) Iron Oxide Surface-Catalyzed Oxidation of Ferrous

Iron by Monochloramine: Implications of Oxide Type and Carbonate on Reactivity.


Vikesland, P. J., Ozekin, K., and Valentine, R. L. (2001) Monochloramine decay in model and

distribution system waters. Water Research, 35(7), 1766-1776.

Vasconcelos, J.J., Rossman, L.A., Grayman, W.M., Boulos, P.F. and Clark, R.M. (1997)

Kinetics of chlorine. Journal of American Water Works Association, 89(11), 54.

Vasquez, F. A., Heaviside, R., Tang, Z. J., and Taylor, J. S. (2006) Effect of free chlorine and

chloramines on lead release in a distribution system. Journal American Water Works

Association, 98(2), 144-154.

Wable, O., Dumoutier, N., Duguet, J. P., Jarrige, P. A., Gelas, G., and Depierre, J. F. (1991)

Modeling chlorine concentrations in a network and applications to Paris distribution

network. Proceedings of Water Quality Modeling in Distribution Systems Conference,

American Water Works Association Research Foundation, Denver, CO.

Westbrook, A., and Digiano, F. A. (2009) Rate of chloramine decay at pipe surfaces. Journal


Westerhoff, P. and Mash, H. (2002) Dissolved organic nitrogen in drinking water supplies: a

review. Journal of Water Supply Research and Technology-Aqua, 51(8), 415-448.

Wilczak, A., Jacangelo, J. G., Marcinko, J. P., Odell, L. H., Kirmeyer, G. J., and Wolfe, R. L.

(1996) Occurrence of nitrification in chloraminated distribution systems. Journal


Wilczak, A., Hoover, L.L. and Lai, H.H. (2003a) Effects of treatment changer on chloramine

demand and decay. Journal American Water Works Association, 95(7), 94-106.

38

Wilczak, A., Assadi-Rad, A., Lai, H.H., Hoover, L.L., Smith, J.F., Berger, R., Rodigari, F.,

Beland, J.W., Lazzelle, L.J., Kinicannon, E.G., Baker, H. and Heaney, C.T. (2003b)

Formation of NDMA in chloraminated water coagulated with DADMAC cationic

polymer. Journal American Water Works Association, 95(9), 94-106.

Williams, S. M. (1990) The use of sodium-silicate and sodium polyphosphate to control water-

quality problems. Water Supply, 8(1 and 2), 195-198.

Wolfe, R.L., Lieu, N.I., Izaguirre, G. and Means, E.G. (1990) Ammonia-oxidizing bacteria in a

chloraminated distribution system- seasonal occurrence, distribution, and disinfection

resistance. Applied and Environmental Microbiology, 56(2), 451-462.

Woolschlagar, J., Rittmann, B., Piriou, P., Kiene, L., and Schwartz, B. (2001) Using a

comprehensive model to identify the major mechanisms of chloramine decay in

distribution systems. Water science and technology: water supply, 1 (4),103-110

Xiao, W.Z., Hong, S.K., Tang, Z.J. and Taylor, J.S. (2007) Effects of blending on total copper

release in distribution systems. Journal American Water Works Association, 99(1), 78-88.

Yagil, G., and Anbar, M. (1962) Kinetics of hydrazine formation from chloramine and ammonia.

Journal of the American Chemical Society, 84(10), 1797-1803.

Zhang, L., Arnold, W.A. and Hozalski, R.M. (2004) Kinetics of haloacetic acid reactions with

Fe(0). Environmental Science & Technology, 38(24), 6881-6889.

Zhang, Y., Griffin, A., and Edwards, M. (2008) Nitrification in premise plumbing: Role of

phosphate, pH and pipe corrosion. Environmental Science & Technology, 42(12), 4280-

4284.

Zhang, Y., and Edwards, M. (2009) Accelerated chloramine decay and microbial growth by

nitrification in premise plumbing. Journal American Water Works Association, 101(11),

51-61.

Zhang, Y., Griffin, A., Rahman, M., Camper, A., Baribeau, H., and Edwards, M. (2009a) Lead

Contamination of Potable Water Due to Nitrification. Environmental Science &

Technology, 43(6), 1890-1895.

39

Zhang, Y., Love, N., and Edwards, M. (2009b) Nitrification in Drinking Water Systems. Critical

Reviews in Environmental Science & Technology, 39(3), 153-208.

Zhang, Y., and Edwards, M. (2010b) Nutrients and metals effects on nitrification in drinking


Zhao, Y. Y., Boyd, J., Hrudey, S. E., and Li, X. F. (2006) Characterization of new nitrosamines

in drinking water using liquid chromatography tandem mass spectrometry.


40

Free Chlorine Degradation and HAA Formation in Two 3

Water Matrices in Contact with Three Metal Materials

Experiments that formed the basis for Chapters 3 and 4 compared the kinetics of metal

corrosion, chlorine degradation and HAA formation for iron, copper and lead treated with

different corrosion inhibitors on a “survey” basis. Metal release kinetics data will be discussed in

Chapter 4 due to the quantity of data. Because most of the results from this survey study

regarding each metal material can be explained well by previously proposed theory and/or

mechanisms in the literature, no further mechanistic investigation was explored in these two

chapters. However, efforts to compare the impacts of corrosion inhibitors on free chlorine decay

and HAA formation among three metal materials have not been reported previously, and thus

this information is relatively new to the field. The observations concerning enhanced HAA

formation in the presence of copper also laid the groundwork for further investigations into

copper catalysis as reported in Chapters 5 and 6.

The results of this chapter in Sections 3.3.1.1, 3.3.2.1, 3.3.2.3, and 3.3.3.1 that deal with

chlorine degradation and HAA formation for fresh metal coupons in one type of water

(Mannheim Water) have been published as part of:

Zhang, H., Andrews, S. A. (2012) Effects of Phosphate-based Corrosion Inhibitors on the

Kinetics of Chlorine Degradation and HAA Formation in Contact with Three Metal

Materials. Canadian Journal of Civil Engineering, 39, 44-54.

Data from similar experiments using fresh metal coupons to investigate free chlorine decay and

HAA formation in another water matrix (Britannia Water) and experiments using pre-corroded

coupons for both tested water matrices are included in this chapter for completeness.

Results from this chapter focus on the research gaps “Are there differences in disinfectant

degradation and HAA formation for different pipe materials treated with different phosphate-

based corrosion inhibitors?” and “Do metal age and water quality affect disinfectant stability and

HAA formation?”

41

Abstract

Pipe materials and their corrosion products can significantly affect disinfectant residual

stability. However, with increasing application of corrosion inhibitors in distribution systems for

corrosion control, no relevant information has been found in literature about the impacts of

corrosion inhibitors and/or their interactions with metal surfaces on disinfectant degradation and

disinfection byproduct formation for both newly installed and aged pipe systems. Therefore, this

chapter investigated and compared free chlorine degradation and HAA9 formation in the absence

and presence of orthophosphate and polyphosphate for two water matrices containing fresh or

aged ductile iron, lead and copper coupons. Material-specific formation potential (MS-FP) and

material-specific simulated distribution system (MS-SDS) tests were applied at bench scale.

Regardless of metal age and water types, the reactivity of metal materials with free chlorine

followed the sequence of Fe> Cu> Pb. The addition of phosphate-based corrosion inhibitors

generally increased HOCl degradation for fresh iron coupons, but decreased HOCl decay only

for fresh copper coupons. Generally, these phosphate-based corrosion inhibitors did not impact

HAA formation. For fresh copper coupons in both investigated water matrices, HAA formation

was enhanced in the presence of high levels of copper ions, indicating possible catalytic potential

of copper on HAA formation. In addition, DCAA was the dominant species observed for water

in contact with fresh metal coupons, whereas either MCAA or DCAA formation was favored

over other species in water containing corroded metal coupons.

Keywords: Free chlorine; Haloacetic acids; Iron; Copper; Lead; Corrosion inhibitor

42

3.1 Introduction

To maintain microbial water quality in distribution systems, secondary disinfectants are

applied. Free chlorine is currently the most widely used secondary disinfectant in the distribution

systems due to its low cost and effectiveness. In distribution systems, free chlorine may

experience temporal and spatial degradation due to chemical and biological consumption that

occurs in the bulk water and on the pipe wall. At least 0.2 mg/L of free chlorine residual should

be maintained to reduce the possible occurrence of further contamination (USEPA, 1989).

Factors affecting disinfectant residual stability in distribution systems include temperature,

concentration of organic matter, biofilms, hydraulics, and infrastructure such as pipe materials,

corrosion products, and pipe service age (Lu et al., 1999; Rossman et al., 2001; Brereton and

Mavinic, 2002; Haas et al., 2002; Hallam et al., 2002; Vikesland and Valentine, 2002; Al-Jasser,

2007).

Haloacetic acids (HAA) are formed by the chlorination of organic materials.

Dichloroacetic acid (DCAA) and trichloroacetic acid (TCAA) are the most commonly-reported

HAA species. Other species, generally at lower levels in low-bromide water, include

bromochloroacetic acid (BCAA), dibromoacetic acid (DBAA), monochloroacetic acid (MCAA),

monobromoacetic acid (MBAA), bromodichloracetic acid (BDCAA), chlorodibromoacetic acid

(CDBAA) and tribromoacetic acid (TBAA). The toxicological properties of HAAs depend on

the extent and type of halogen substitution (bromine or chlorine). The Stage 2 Disinfectants and

Disinfection Byproducts Rules D/DBPR regulated the Maximum Contaminant Level (MCL) of

HAA5 (MCAA, DCAA, TCAA, MBAA, and DBAA) at 0 μg L (USEPA, 2006). Guidelines for

Canadian Drinking Water Quality established the Maximum Acceptable Concentration (MAC)

for HAA5 in drinking water at 80 μg L (Health Health Canada, 2008).

Precursors of HAAs are primarily natural organic matter, including humic and fulvic

substances (Kanokkantapong et al., 2006). During chlorination, the formation of DCAA and

TCAA is significant, and the dominance of both species is dependent on chlorine dose,

temperature, pH, and specific ultraviolet absorbance (SUVA) of water. More halogenated species

are favored at increasing chlorine dose, temperature and/or SUVA or at lower pH. In distribution

systems, the concentrations of HAA9 vary both spatially and temporally. HAA9 concentrations in

chlorinated systems generally increase or remain approximately constant as water age increases,

43

while remaining relatively constant in chloraminated distribution systems (Baribeau et al., 2006).

In both systems, observed decreases in HAA9 concentration following long residence times have

been attributed primarily to biodegradation (Chen and Weisel, 1998; Rodriguez et al., 2004).

It is widely accepted that disinfectant residuals in distribution systems increase corrosion

rates, as secondary disinfectants serve as potential oxidants for distribution system pipes and

related plumbing materials. In recent years, considerable efforts have been made to understand

the mechanisms of metal release from pipes and corrosion scales, the rates at which they occur

under the influence of disinfectant residuals and hydraulic conditions. Effective strategies to

control corrosion and/or metal release have also been developed. One such strategy is applying

phosphate-based corrosion inhibitors in the form of either orthophosphoric acid, combinations of

orthophosphoric acid and zinc orthophosphate, polyphosphates, or blends of orthophosphate and

polyphosphate. Generally, orthophosphate forms an impervious solid film as a barrier between

the corroding metal and the water (Demora and Harrison, 1984; Singley, 1994), whereas

polyphosphates are strong chelating agents which have been successfully used for reducing

scaling and sequestering Fe2+ ions to treat “red water” (Facey and Smith, 199 ; Maddison and

Gagnon, 1999; Williams, 1990).

It can be hypothesized that the application of phosphate-based corrosion inhibitors which

significantly affect corrosion rates may, in turn, impact disinfectant degradation rates and the

formation kinetics of DBPs. The purpose of this study was to investigate how pipe materials

(ductile iron, copper and lead) and phosphate-based corrosion inhibitors (orthophosphate and

polyphosphate) influence free chlorine degradation and HAA formation. These pipe materials are

widely present in either water mains (such as ductile iron) or household plumbing (such as

copper), and lead may be leached into drinking water from old lead service lines, soldered joints

and brass plumbing fittings (Health Canada, 2009). Given that no reports have been found that

compare the behavior of free chlorine degradation and HAA formation for all three metal

materials, understanding the stability of free chlorine residual and the fate of HAA compounds

under the influence of these pipe materials and their phosphate-based corrosion preventive

strategies will benefit utilities and households to develop effective strategies to control the

deterioration of water quality in both distribution system water mains and premise plumbing.

44

3.2 Materials and Methods

Reagents and Materials 3.2.1

All chemicals used in this study were ACS grade or higher. The chlorine dosing solution

(approximately 3500 mg/L as Cl2) was prepared by diluting a concentrated solution of sodium

hypochlorite (NaOCl, 6%, VWR) in Milli-Q water. The target disinfectant concentration applied

in the tests was achieved by spiking free chlorine dosing solution into 1 L test water.

Orthophosphate (Na3PO4) and polyphosphate ([Na(PO2)]6) were selected to investigate their

impacts on metal corrosion in this study. They were prepared at a concentration of 500 mg/L as P

and were kept in the dark at 4ºC. The targeted dosages of these corrosion inhibitors were 1 mg/L

as P.

Test coupons of ductile iron, copper and lead were purchased from Metal Samples Co.,

Alabama, US. The size of these coupons is 1 2”3”1 1 ”. For experiments with fresh coupons,

before each batch of experiments, any corrosion products were removed from the coupons by

polishing with 60-grit sandpaper followed by 120-grit sandpaper, and then rinsing with deionized

water and acetone followed by Milli-Q water (Scholze et al., 1994). Polished coupons were

submerged into the test water to simulate contact with new pipes without any impacts from

service age. In addition, to investigate the aged pipe environment, some coupons of each pipe

material were pre-conditioned in tap water in the absence and presence of orthophosphate. The

initial free chlorine concentration for conditioning purposes was 10 mg/L. After each 24 hour

reaction period for the iron coupons and each 48 hour reaction period for the copper and lead

coupons, the free chlorine residual was measured and then the water was replaced with fresh tap

water dosed with 10 mg/L free chlorine. The conditioning process proceeded until the 24 hour

chlorine residual for each combination of metal type and corrosion inhibitors was approximately

constant (< 4.0 mg/L chlorine decay for the iron coupons alone, 6.8 mg/L for the iron coupons

treated with orthophosphate, 2.7 mg/L for all of the copper coupons, and 1.9 mg/L for all of the

lead coupons). The measured 24 hour chlorine residuals for iron, copper and lead are shown in

Appendices Figure 10-14. Orthophosphate was selected since it effectively controlled metal

release for fresh metal coupons in short term, especially for copper and lead.

45

Unchlorinated post-filtration water was collected from Mannheim Water Treatment Plant

(MWTP) and Britannia Water Purification Plant (BWPP), Ontario. Water quality parameters are

listed in Table 3-1.

Table 3-1 Summary of water quality parameters for post-filtration water

Parameters

Values

Mannheim

Water

Britannia

Water

pH 7.5 ± 0.2 6.6±0.5

Alkalinity (mg/L) 187±33 12±2

UV254 (cm-1

) 0.058±0.008 0.065±0.005

TOC (mg/L) 4.5±0.3 3.3±0.2

SUVA (L/mg·cm-1) 0.013±0.002 0.019±0.0002

Bromide (μg L) 65.0±15.5 15.8±0.2

Chloride (mg/L) 84.5±2.5 3.1±0.6

Sulfate (mg/L) 35.0±2.0 23.8±0.8

Cl-:SO4

2- ratio 2.4±0.1 0.13±0.04

Note: TOC, total organic carbon; SUVA, specific ultraviolet absorbance.

Experimental Procedures 3.2.2

“Material-specific” simulated distribution system (MS-SDS) and material-specific

formation potential (MS-FP) tests were applied in this study. Details of MS-SDS procedures

have been described by Brereton and Mavinic (2002). They consist of incubating metal coupons

in water samples under conditions representative of actual field conditions in terms of reaction

time, pH, temperature, and disinfectant application. In MS-SDS tests, not only is disinfectant

interaction with precursors in the bulk water considered, but the influences of distribution system

pipe walls on disinfectant degradation and HAA generation are also examined. The MS-SDS

tests were performed for experiments employing corroded metal coupons, using an initial free

chlorine concentration of approximately 5.5 mg/L to ensure detectable disinfectant residuals (0.2

mg/L) after 24 hours. Experiments with fresh metal coupons employed material-specific

formation potential (MS-FP) tests that were performed similarly to the MS-SDS tests except that

higher concentrations of free chlorine (12.3 mg/L) were applied than would normally be

encountered during typical water treatment to meet the high chlorine demand of fresh metal

coupons and their corrosion products and ensure detectable chlorine residuals after 24 hours as

well.

46

All of the MS-SDS and MS-FP experiments were performed in 1 L amber bottles with

PTFE lined caps and at room temperature (21±2 ºC). The reaction bottles were made chlorine

demand free before use by soaking them in a concentrated sodium hypochlorite solution (~1000

mg/L as Cl2) for at least 24 hours. Thereafter, the bottles were rinsed thoroughly with deionized

water and distilled water, and were heated at 250 °C for at least 4 hours. Metal coupons were

suspended in the bottles using nylon threads. To maximize the contact of water with coupon

surface, a Big Bill Orbital shaker was used (Barnstead International, USA) to maintain a gentle

mixing at a speed of 25 rpm. After each designated reaction time, 50 mL of water were

withdrawn from the test bottles for the analysis of disinfectant residual, HAA and metal

concentrations. Seven to ten time points were employed for each of the kinetic studies. As will

be discussed in Section 4.3.1 (Chapter 4), metal concentrations reached equilibrium after 10

hours, and the reduction of water volume due to sequential sampling did not significantly affect

equilibrated metal concentrations. Separate control tests were also performed to confirm that

there were also no losses of HAA compounds due to volatility during the sequential sampling

events.

Two sets of kinetic experiments were performed to test the reproducibility of the results,

and duplicate tests were conducted in each set of experiments. All of the tests also included

control samples, which were prepared the same as test samples but without metal coupons.

Analytical methods that were employed to examine water quality and analytes of interest are

summarized in Table 3-2.

Table 3-2 Summary of analytical methods

Analyte Unit Instrument /procedure Reference method

TOC mg/L O1-Analytical TOC analyzer SM 5310 C

pH pH meter

UVA254 cm-1

Hewlett Packard 8452A Diode Array UV

spectrophotometer SM 5910B

Alkalinity mg/L Titration SM2320B

Chlorine mg/L Hach DR2700 Spectrophotometer SM 4500-Cl G

Anions µg/L Dionex DX-300 Series Ion

Chromatography System SM 4110 B

Haloacetic acids µg/L Hewlett Packard 5890 Series II Plus Gas

Chromatography SM 6251

Metal (Pb, Cu, and Cu) mg/L Varian SpectrAA.20 SM 3111B

Note: TOC, total organic carbon; SM represents Standard Methods for the Examination of Water and Wastewater

(APHA, AWWA, WEF, 2005).

47

3.3 Results and Discussion

The concentrations of free chlorine, released metal and HAA9 were monitored by

periodically withdrawing and analyzing samples from the test bottles. Using these data, it was

possible to compare free chlorine decay rates and HAA formation among the different

combinations of metals and corrosion inhibitors.

Free Chlorine Degradation 3.3.1

3.3.1.1 Fresh Coupons in Two Water Matrices

Figure 3-1 illustrates the observed free chlorine decay as a function of time for iron,

copper and lead coupons in the presence and absence of corrosion inhibitors in Mannheim

Water. Free chlorine degraded more quickly when in contact with all three pipe materials than

for the control samples without metal coupons, regardless of the type of corrosion inhibitors

present, although the effect was small in the presence of lead. Compared with the presence of

other metals, free chlorine consumption in the presence of iron coupons was much more rapid,

and almost no chlorine residual detected after 24 hours (Figure 3-1a). Therefore, to ensure

detectable chlorine residual at 24 hours (0.2 mg/L), an initial chlorine concentration of 12.3

mg/L was applied in all of the MS-FP bottles, even though it would be higher than observed

during normal applications.

The effects of the various experimental conditions on chlorine decay for copper coupons

can been seen in Figure 3-1b. The presence of corrosion inhibitors significantly reduced free

chlorine degradation rate in comparison with copper coupons in the absence of corrosion

inhibitors. Two possibilities may account for this phenomenon. One was due to rapid

consumption of free chlorine by copper corrosion itself. Without the protection from corrosion

inhibitors, the copper surface was more readily available to react with chlorine. The other

possibility was that dissolved copper ions could either directly enhance free chlorine degradation

to produce oxygen and chloride as final degradation products (OCl- O2+Cl

-; Gray et al., 1977)

or increase the reactivity of DBP precursors (Blatchley et al., 2003; Fu et al., 2009). Both of

these options would lead to faster consumption of free chlorine in the presence of pure copper

alone. No obvious impacts of corrosion inhibitors on free chlorine decay were observed for lead

coupons (Figure 3-1c). They appeared to have similar free chlorine decay behavior in the

48

presence of these corrosion inhibitors compared with metal coupons alone. It should be noted

that error bars, which represent the measured maximum and minimum values, cannot be seen in

Figure 3-1 since they were small (0.0 mg/L ~ to 0.2 mg/L) relative to the measured chlorine

residuals.

Figure 3-1 Free chlorine decay for fresh metal coupons with time in the presence and absence of

corrosion inhibitors in one set of experiments, n=2; (a) iron, (b) copper, (c) lead.

The rates of free chlorine decay were obtained by fitting a pseudo first-order decay

equation to the chlorine data from each experiment. In Ct = C0 exp (-kt), Ct is the free chlorine

concentration (mg/L) at time t, C0 the free chlorine concentration (mg/L) at time zero, and k the

0

3

6

9

12

15

0 10 20 30

Fre

e c

hlo

rin

e (

mg

/L)

Fe

Fe+polyphosphate

Fe+orthophosphate

Control

0

3

6

9

12

15

0 20 40 60 80 100 120

Fre

e c

hlo

rin

e (

mg

/L)

Cu

Cu+polyphosphate

Cu+orthophosphate

Control

b

c

a

0

3

6

9

12

15

0 20 40 60 80 100 120

Fre

e c

hlo

rin

e (

mg

/L)

Time (hours)

Pb

Pb+polyphosphate

Pb+orthophosphate

Control

49

overall first-order decay constant (h-1

) which is the sum of a bulk constant and a wall constant: k

= kb + kw. In this equation for k, kb is the bulk first-order chlorine decay constant (h-1

) and kw the

wall first-order chlorine decay constant (h-1

). The bulk water reaction in the MS-PF tests is the

same as would occur in the control bottles containing no metal coupons, thus the wall decay

constant is the difference between the overall decay constant and the bulk decay constant. Table

3-3 summarizes chlorine decay constants, including overall decay constants (k) and wall decay

constants (kw) for all three metal materials and the different corrosion inhibitors that were tested.

The variability in the decay constants as shown in Table 3-3 were likely attributed to the

different coupon surface conditions between sets of experiments even though all of the corroded

coupons were polished following the same procedures described in Section 3.2.2 before each

new set of experiments.

Table 3-3 Chlorine decay constants for fresh metal coupons in the presence and absence of

corrosion inhibitors with Mannheim Water (n=4)

Material Corrosion inhibitors k (h

-1) kw (h

-1)

Average Stdev Average Stdev

Iron

None 0.1451 0.0006 0.1402 0.0014

Polyphosphate 0.1834 0.0115 0.1789 0.0121

Orthophosphate 0.1668 0.0016 0.1620 0.0025

Copper

None 0.0292 0.0064 0.0239 0.0051

Polyphosphate 0.0169 0.0071 0.0117 0.0056

Orthophosphate 0.0137 0.0052 0.0084 0.0037

Lead

None 0.0078 0.0005 0.0029 0.0013

Polyphosphate 0.0075 0.0006 0.0030 0.0012

Orthophosphate 0.0078 0.0011 0.0029 0.0020

Control

(water)

None 0.0049 0.0008

NA Polyphosphate 0.0045 0.0006

Orthophosphate 0.0049 0.0009

Note: NA-not available.

In order to evaluate the interactive effects of corrosion inhibitors and metal surface type

on chlorine decay in Mannheim Water and to determine whether the treatment factor (the

different corrosion inhibitors) had a significant influence on the response factor (free chlorine

wall decay constants), a single-factor ANOVA test at a confidence level of 95% was applied to

the wall decay constants that were determined for each metal material. When the ANOVA test

signified statistically significant impacts of corrosion inhibitors on free chlorine wall decay, a

Fisher’s Least Significant Difference (LSD) test was applied to further determine if significant

50

differences existed between each pair of treatments at a 95% of confidence level (Montgomery,

2000).

As shown in Table 3-3, the reactivity of three metal materials with free chlorine followed

the sequence in decreasing order: iron, copper, and lead. Iron had wall decay constants two

orders of magnitude higher than for the bulk decay, irrespective of the type of corrosion

inhibitors present, so chlorine decay was predominant on the pipe wall. The ANOVA test

determined that there were significant differences in free chlorine decay among the three

treatments (p-value of 0.02), and the results of the LSD tests indicated that the corrosion

inhibitors significantly increased the free chlorine wall decay.

As was observed with iron, copper consumed free chlorine primarily by wall reactions,

and corrosion inhibitors exhibited significant impacts on free chlorine degradation (p-value of

0.01). However, the extent of the contribution of wall reactions to the overall chlorine decay was

dependent on the type of corrosion inhibitors present. The relative sequence of impact was: no

corrosion inhibitor < polyphosphate < orthophosphate. When LSD tests were used to compare

free chlorine wall decay constants in the presence of each type of corrosion inhibitors with those

for copper coupons alone, the results indicated that corrosion inhibitors significantly decreased

free chlorine degradation. This may have been from the formation of protective scales and/or

decreasing released copper concentrations, thereby making the copper surface less available to

react with free chlorine and/or decreasing copper catalysis.

In contrast to the results for iron and copper, chlorine consumption was dominated by

bulk water decay for lead coupons, for which their bulk water decay coefficients were at least

40% higher than their wall decay constants. Corrosion inhibitors did not exhibit significant

impacts on free chlorine wall decay relative to lead coupons in the absence of corrosion

inhibitors (p-value >0.05).

The relative reactivity of three metal materials with HOCl in Britannia Water had a

similar sequence to that observed in Mannheim Water. Namely, fresh iron coupons were most

reactive with free chlorine among three metal materials followed by copper, and then lead. The

addition of phosphate-based corrosion inhibitors significantly increased chlorine decay rates for

iron coupons, but reduced chlorine decay for copper and lead coupons.

51

The impacts of water quality and corrosion inhibitors on chlorine decay in the two water

matrices were determined using a two-factor ANOVA test at a confidence level of 95%. The

results of the ANOVA test (p-values) are shown in Table 3-4, along with the pseudo-first-order

HOCl decay constants.

Table 3-4 Comparison of HOCl overall decay constants (h-1

) for fresh coupons between two

water matrices (n=4)

Material Corrosion

inhibitors

Mannheim

Water

Britannia

Water

p-value (=5%)

Water

quality

Corrosion

inhibitors

Iron

None 0.1451±0.0006 0.1555±0.0305

0.02 0.001 Polyphosphate 0.1834±0.0115 0.2446±0.0040

Orthophosphate 0.1668±0.0016 0.1671±0.0041

Copper

None 0.0292±0.0064 0.0257±0.0075

0.23 0.01 Polyphosphate 0.0169±0.0071 0.0123±0.0011

Orthophosphate 0.0137±0.0052 0.0093±0.0009

Lead

None 0.0078±0.0005 0.0083±0.0021

0.25 0.27 Polyphosphate 0.0075±0.0006 0.0050±0.0006

Orthophosphate 0.0078±0.0011 0.0066±0.0023

Control

(water)

None 0.0049±0.0008 0.0038±0.0003

0.04 0.52 Polyphosphate 0.0045±0.0006 0.0034±0.0001

Orthophosphate 0.0049±0.0009 0.0041±0.0007

Initial free chlorine concentration: 12.3 mg/L as Cl2

The ANOVA test results suggest that in Britannia Water chlorine decay constants for

fresh iron coupons significantly increased (p-value 0.02) relative to Mannheim Water, while the

bulk water control samples in Britannia Water had slightly decreased chlorine decay constants

(p-value 0.04). Unexpectedly, fresh copper and lead coupons had statistically the same chlorine

degradation rates in each of the two water matrices. Since Britannia Water had lower pH and

alkalinity than Mannheim Water (Table 3-1), Britannia Water would be more corrosive to metal

coupons than Mannheim Water, and thus more rapid free chlorine degradation should have been

observed in Britannia Water than in Mannheim Water. However, this was only observed for

fresh iron coupons and the differences were small. Differences in the chlorine bulk decay in the

two water matrices can be likely attributed to their different TOC levels (Hallam et al., 2003;

Powell et al., 2000). In general, the higher the TOC concentration, the more rapidly free chlorine

will degrade. Therefore, it is consistent that the free chlorine bulk decay rates in Mannheim

Water were statistically higher than those in Britannia Water, regardless of the presence and the

types of corrosion inhibitors.

52

The ANOVA test results in Table 3-4 also indicate that in both water matrices the

presence of corrosion inhibitors significantly increased free chorine decay for fresh iron coupons

(p-value 0.001), while for copper coupons these corrosion inhibitors statistically decelerated

chlorine degradation (p-value 0.01). However, free chlorine degradation did not show any

statistical difference in the absence and presence of these corrosion inhibitors (p-value 0.27) for

solutions with fresh lead coupons and the bulk water control samples. In general, the results of

this two-factor ANOVA test provide new information concerning the effects of two different

water quality and corrosion inhibitors on free chlorine degradation.

3.3.1.2 Corroded Coupons in Two Water Matrices

Figure 3-2 compares the pseudo-first order free chlorine decay constants for three types

of corroded metal coupons in contact with Mannheim Water. The reactivity of these materials

with free chlorine followed the general sequence of iron >copper >lead, although the decay rates

were similar for solutions with corroded copper and lead when there was no corrosion inhibitor

present. The overall decay rates of chlorine in solutions with corroded iron treated with

orthophosphate were 0.21±0.001 h-1

, and were not statistically different from the decay rates in

the absence of orthophosphate (0.15±0.067 h-1

).

Figure 3-2 Free chlorine overall decay constants for corroded coupons with Mannheim Water.

Initial free chlorine concentration 12.3 mg/L, error bars indicate standard deviation (n=4)

For corroded copper coupons alone, the overall decay constant was only 70% higher than

those in the bulk water control, indicating that bulk water reaction was dominant for free chlorine

consumption. In contrast, corroded copper coupons treated with orthophosphate had an overall

free chlorine decay rate 1.8 times higher than the bulk water control, suggesting that the metal

0.00

0.05

0.10

0.15

0.20

0.25

No inhibitor Orthophosphate

1st o

rder

overa

ll decay c

onsta

nt

(h

-1)

Fe Cu Pb Control

53

wall reactions dominated free chlorine degradation. Some studies have revealed that

orthophosphate may interfere with the aging processes of passivating films of copper, e.g., from

soluble Cu(OH)2 to malachite over the long term (Li et al., 2004), thereby releasing copper from

the aged pipes. An increased copper corrosion rate in the presence of orthophosphate was also

observed in this water matrix (from 0.4 mg/L in the absence of orthophosphate to 0.5 mg/L after

the addition of orthophosphate at 24 hours). In addition, copper has been reported to catalyze

free chlorine degradation (Fu et al., 2009b; Gray et al., 1977). Further experiments have also

been conducted in this research to investigate factors affecting copper catalysis on chlorine

degradation (Chapter 5) and the extent of copper catalysis was shown to increase with increasing

copper concentrations (Chapter 5). Therefore, it was not surprising to observe accelerated free

chlorine degradation in the presence of orthophosphate relative to that for corroded copper

coupons alone (single factor ANOVA test, p =0.03).

Corroded lead coupons had comparable overall free chlorine decay constants with the

bulk water control, indicating that bulk water reactions were the primary pathways for chlorine

degradation. There was no significant difference in chlorine decay constants for corroded lead

coupons in the absence or presence of orthophosphate.

Similar to the results observed with Mannheim Water, for corroded metal coupons in

Britannia Water, free chlorine was more reactive with iron, followed by copper and then lead.

Wall reactions dominated over bulk water reactions for corroded iron coupons. However, the

wall consumption of free chlorine was also dominant for copper coupons alone, which was

different from the observations with Mannheim Water. The addition of orthophosphate did not

statistically affect free chlorine decay for corroded iron and copper coupons, but significantly

decreased free chlorine degradation for corroded lead coupons (single factor ANOVA test, p =

0.005).

A two-factor ANOVA test at a confidence level of 95% was also performed to evaluate

the impacts of water quality and the interactive effects of corrosion inhibitors with metal surface

in two water matrices on chlorine decay. The results are shown in Table 3-5. For the three tested

metal materials and bulk water, regardless if orthophosphate was present, free chlorine degraded

more quickly in Mannheim Water than in Britannia Water with all of the p-values below 5%.

This suggests that something specific to the water types (e.g., the relatively higher TOC level in

54

Mannheim Water than in Britannia Water) was the primary reason for the rapid free chlorine

degradation in Mannheim Water.

Table 3-5 Comparison of HOCl overall decay constants for corroded coupons between

Mannheim Water and Britannia Water (n=4)

Material Corrosion

inhibitors

Mannheim

Water

Britannia

Water

p-value (=5%)

Water

quality

Corrosion

inhibitors

Iron None 0.1850±0.0178 0.0953±0.0044

0.0009 0.19 Orthophosphate 0.2064±0.0010 0.1075±0.0242

Copper None 0.0346±0.0056 0.0200±0.0068

0.002 0.08 Orthophosphate 0.0513±0.0028 0.0195±0.0026

Lead None 0.0268±0.0046 0.0156±0.0003

0.009 0.26 Orthophosphate 0.0251±0.0056 0.0081±0.0007

Control

(water)

None 0.0254±0.0007 0.0073±0.0006 0.0002 0.90

Orthophosphate 0.0248±0.0014 0.0070±0.0014

Orthophosphate appeared not to affect free chlorine degradation for the tests performed in

these two tested water matrices since all of the p-values were above 5%. However, by further

comparing chlorine decay constants for corroded copper and lead coupons in the absence and

presence of orthophosphate for each single water matrix (Table 3-5), orthophosphate had a

different impact on free chlorine degradation.

For corroded copper coupons, free chlorine in Mannheim Water degraded more rapidly in the

presence of orthophosphate, but there was no statistical difference in free chlorine degradation in

the absence and presence of orthophosphate for Britannia Water. Figure 3-3 shows copper

release kinetics for corroded copper coupons in two water matrices. In the presence of

orthophosphate, the copper release rate in Mannheim Water was generally increased relative to

that in the absence of orthophosphate. However, orthophosphate slightly decreased the copper

release rate in Britannia Water after 30 hours.

For corroded lead coupons, orthophosphate significantly decreased free chlorine

degradation in Britannia Water (single factor ANOVA test, p = 0.005), but exhibited no

statistical impacts on free chlorine degradation in Mannheim Water. As shown in Figure 3-4,

although orthophosphate effectively reduced lead release in both water matrices, the magnitude

of reduction by orthophosphate was more significant in Britannia Water than that in Mannheim

Water. Therefore, the interactions of orthophosphate and corroded metal surface, which are

55

likely affected by water quality, may play an important role in determining free chlorine

degradation rates for corroded copper and lead coupons in different water matrices.

Figure 3-3 Copper release kinetics for corroded coupons in Mannheim Water and Britannia

Water; error bars indicate the measured maximum and minimum values (n=2)

Figure 3-4 Lead release kinetics for corroded coupons in Mannheim Water and Britannia Water;

error bars indicate the measured maximum and minimum values (n=2)

HAA Formation 3.3.2

3.3.2.1 Fresh Metal Coupons

In this study, 9 compounds of haloacetic acids were analyzed. However, due to relatively

low concentrations of bromodichloro-, chlorodibromo-, and tribromo-acetic acids, only the

concentrations of the other 6 haloacetic acids (HAA6) are reported here. A discussion concerning

the HAA6 speciation will be provided in Section 3.3.3. Figure 3-5 demonstrates formation

kinetics of HAA6 for fresh metal coupons in the absence and presence of corrosion inhibitors in

Mannheim Water. It should be noted that for water samples containing iron coupons, despite no

0.0

0.4

0.8

1.2

0 10 20 30 40 50 60

Cu (

mg/L

)

Time (hours)

Mannheim Water

Cu

Cu+orthophosphate

0.0

0.5

1.0

1.5

2.0

0 10 20 30 40 50 60

Cu (

mg/L

)

Time (hours)

Britannia Water

Cu

Cu+orthophosphate

0

1

2

3

4

0 10 20 30 40 50 60

Pb (

mg/L

)

Time (hours)

Mannheim Water

Pb

Pb+orthophosphate

0

1

2

3

4

0 10 20 30 40 50 60

Pb (

mg/L

)

Time (hours)

Britannia Water

Pb

Pb+orthophosphate

56

chlorine residual after 24 hours, HAA concentrations were monitored to investigate the fate of

HAA6 when they were in contact with iron. As shown in Figure 3-5, the rates of HAA6 formation

for all of the samples were relatively fast in the initial 8 hours, and then tended to level off after

24 hours for iron coupons and after 100 hours for lead coupons and bulk water control. However,


fresh metal coupons with Mannheim Water in one set of experiments, error bars indicate the

measured maximum and minimum values (n=2)

0

40

80

120

160

200

0 30 60 90 120

HA

A 6

(µ

g/L

)

Time (hours)

CuCu+polyphosphateCu+orthophosphateControl

0

40

80

120

160

200

0 30 60 90 120

HA

A 6

(µ

g/L

)

Time (hours)

Pb

Pb+polyphosphate

Pb+orthophosphate

Control

0

40

80

120

160

200

0 30 60 90 120

HA

A 6

(µ

g/L

)

Time (hours)

FeFe+polyphosphateFe+orthophosphateControl

57

no plateaus of HAA formation were observed for copper coupons in the absence and presence of

corrosion inhibitors, and HAA6 continued to form after 100 hours. Compared with bulk water

control, iron coupons produced approximately 50% less HAA6 at similar reaction time. It is

evident since free chlorine was consumed primarily by iron corrosion and thus less free chlorine

was available to react with HAA6 precursors to form HAA6. No significant influence of

corrosion inhibitors on HAA6 formation was observed for fresh iron coupons.

For water containing copper coupons, lower HAA6 production was also observed at each

reaction time compared with bulk water. However, the HAA6 formation rate for copper alone

was obviously higher than for other copper coupons in the presence of different corrosion

inhibitors and even in the bulk water. When the reaction time was 98 hours, 157 µg/L of HAA6

formed for copper alone, significantly higher than that in the bulk water which was only 141

µg/L (student t test, p-value =0.008). Other water containing copper coupons showed faster

HAA6 formation kinetics than for the bulk water after 72 hours as well, even though the absolute

concentrations of HAA6 were still lower than in the bulk water. The fact that copper corrosion

itself consumed some free chlorine implied that less HAA6 should have been formed in the

system. Enhanced formation of HAA compounds in the presence of higher levels of copper ions

indicated that these copper ions could catalyze HAA formation. The mechanism is believed to be

similar to the catalytic effect of copper(II) and copper oxides in THM formation, in which copper

could complex with THM precursor compounds and enhance oxidative decarboxylation and

enolization of the keto-groups (Blatchley et al., 2003; Li et al., 2008). This new observation also

laid the groundwork for further investigation of copper catalysis on HAA formation in Chapter 5.

Catalytic potential of copper corrosion products on HAA formation has been confirmed, and

copper catalysis was dependent on copper concentration, pH and the types of its solid corrosion

products.

In terms of HAA formation for water containing lead coupons in the absence of corrosion

inhibitors, the consumption of free chlorine by lead corrosion resulted in less HAA6 being

produced. Comparable HAA formation was observed for lead coupons treated with 1 mg/L

polyphosphate and 1 mg/L orthophosphate as bulk water control. Further investigation of the

distribution of HAA6 species (Section 3.3.3) shows that the variation of HAA concentrations

among lead coupons treated with different corrosion inhibitors was primarily due to the

difference in MCAA concentrations that were formed. The MCAA formation rate for lead

58

coupons alone was lower than lead coupons treated with polyphosphate and orthophosphate for

72 hours.

Generally, corrosion inhibitors did not affect HAA formation during short reaction

periods (<12 hours) for fresh metal coupons, which has not been reported previously. However,

no conclusion can be made about the long-term effect of corrosion inhibitors on HAA formation

for water containing iron coupons due to fast depletion of free chlorine within 24 hours. Copper

and lead coupons exhibited different HAA formation rates in the presence of different corrosion

inhibitors. The role of these corrosion inhibitors on HAA formation was essentially due to their

effects on the interactions between free chlorine and metals.

In terms of HAA formation in Britannia Water, Figure 3-6 displays HAA6 formation

kinetics for fresh metal coupons in the absence and presence of corrosion inhibitors and in the

bulk water control. Different from the observations in Mannheim Water in which iron coupons

produced much less HAA6 compared with the bulk water control, iron coupons in Britannia

Water had similar HAA formation kinetics to the bulk water control at least for the first 24 hours.

This indicates that the free chlorine concentration, despite its consumption by iron corrosion,

may not be a rate-limiting factor for HAA formation. The difference in some aspect of the water

quality of Britannia Water from Mannheim Water may also be the reason for the observed

different HAA formation kinetics between these two water matrices. Corrosion inhibitors did not

significantly affect HAA formation in this test water as well.

For fresh copper coupons, in the initial 24 hours, HAA formation kinetics were similar to

those observed in the bulk water. After 24 hours and in the absence of corrosion inhibitors,

HAA6 exhibited a faster formation rate compared with other copper coupons in the presence of

corrosion inhibitors and the bulk water control. This indicates that copper catalysis also played

an important role in the enhanced formation of HAA. However, no differences in HAA

formation rates were observed between copper coupons in contact with corrosion inhibitors and

bulk water control in the duration of tests. In Britannia Water, up to 1.1 mg/L Cu(II) was

released from fresh copper coupons after 24 hours, but only 0.7 and 0.4 mg/L Cu(II) was

released from copper coupons when they were in contact with polyphosphate and

orthophosphate, respectively. This suggests that copper catalysis was a function of dissolved

59

copper concentration, and the catalytic impacts of copper at a concentration below 0.7 mg/L may

not be enough to enhance HAA formation.


fresh metal coupons with Britannia Water in one set of experiments, error bars indicate the


For the lead coupons, comparable HAA formation rates were observed in the absence and

presence of corrosion inhibitors relative to the bulk water control. Further investigation of the

0

30

60

90

120

150

0 10 20 30

HA

A6 (

µg

/L)

Time (hours)

Fe

Fe+polyphosphate

Fe+orthophosphate

Control

0

30

60

90

120

150

0 20 40 60 80

HA

A6 (

µg

/L)

Time (hours)

Cu

Cu+polyphosphate

Cu+orthophosphate

Control

0

30

60

90

120

150

0 20 40 60 80

HA

A6 (

µg

/L)

Time (hours)

PbPb+polyphosphatePb+orthophosphateControl

60

distribution of HAA6 species showed that MCAA formation rates for all of the lead coupons

were similar in Britannia Water. Along with the observation that higher MCAA formation rates

corresponded to higher HAA formation rates in Mannheim Water, a general conclusion is made

that different HAA6 formation kinetics for lead coupons in the absence and presence of corrosion

inhibitors was likely attributed to their different MCAA formation kinetics.

3.3.2.2 Corroded Metal Coupons

Figure 3-7 displays HAA6 formation at 48 hours for corroded metal coupons in the

absence and presence of orthophosphate compared with the bulk water control. In Mannheim

Water, at 48 hours, corroded iron coupons had less HAA6 formation (40.7~48.7 µg/L) compared

with bulk water (62.3 µg/L). Due to fast consumption of free chlorine by iron corrosion, it is

reasonable to observe reduced formation of HAA relative to bulk water control. For corroded

copper coupons, there was no significant difference in HAA formation compared with that in

bulk water control. Copper has been shown to catalyze HAA formation for fresh copper coupons

(Section 3.3.2.1), and its catalysis increases nonlinearly with increasing dissolved copper

concentrations, and to become more prominent at longer reaction times (Chapter 5). Therefore,

due to the relatively low concentration of the released copper (< 0.5 mg/L) in the solution,

copper catalysis during HAA formation may not be expected to be significant within 48 hours,

Figure 3-7 HAA6 formation at 48 hours in the presence and absence of corrosion inhibitors for

corroded metal coupons with Mannheim Water and Britannia Water in one set of experiments,


0

30

60

90

120

HA

A6 (

µg/L

)

Mannheim Water Britannia Water

61

which is consistent with the observed results. Also, since free chlorine degraded at a similar rate

in the bulk water control and the water in contact with lead coupons (Section 3.3.1.2), it was not

surprising to observe similar HAA formation kinetics for corroded lead coupons in Mannheim

Water.

In Britannia Water, HAA formation at 48 hours for all of metal coupons and bulk water

control were statistically the same, regardless of the presence of orthophosphate. As discussed in

Section 3.3.1.2, free chlorine degraded faster in the presence of metal coupons, especially when

in contact with corroded iron coupons. Comparable HAA yields between corroded metal

coupons and the bulk water, again, suggests that free chlorine concentration may not be a rate-

limiting factor for HAA formation in Britannia Water.

3.3.2.3 HAA Formation and Chlorine Demand

Figure 3-8 shows the correlation between consumed free chlorine and HAA6 formation

for bulk water control. Strong correlation between disinfectant demand and HAA6 formation (R2

=0.98) confirms that DBP formation is proportional to disinfectant consumption in the bulk

solution as expected (Clark and Sivaganesan, 1998; Gang et al., 2002).

Figure 3-8 HAA formation and free chlorine demand for bulk water in the absence of corrosion

inhibitors

Although free chlorine was consumed partially as a result of metal corrosion, a similar

linear relationship between the overall free chlorine demand and HAA formation was observed

for all of the iron and lead coupons regardless of metal age, water quality, and the presence of

corrosion inhibitors (plots are not provided). This linear relationship was also observed for

R² = 0.98

0

40

80

120

160

200

0 2 4 6

HA

A6 (

µg

/L)

Chlorine demand (mg/L)

62

corroded copper coupons in the absence and presence of orthophosphate in both water matrices.

It is inferred that metal corrosion products consumed free chlorine linearly as well.

However, an exponential relationship between chlorine demand and HAA6 formation was

observed for fresh copper coupons in the absence and presence of different corrosion inhibitors.

Figure 3-9 shows this nonlinear relationship for copper coupons alone in Mannheim Water. In

the first several hours, HAA formation increased linearly with chlorine demand. It can be

hypothesized that a small amount of copper ions initially released into the water did not affect

HAA formation significantly. When more copper ions were released with time, their catalytic

effects on HAA formation became prominent. More HAA formation was observed after 98 hours

in the presence of copper ions (157 µg/L, Point A) compared with the extrapolated value at 98

hours on the assumed linear line established between initial HAA formation and chlorine

demand (approximately 90 µg/L, Point B). Further experiments have been conducted and the

catalysis of copper of HAA formation has been confirmed (Chapter 5).

Figure 3-9 HAA formation and free chlorine demand for copper in the absence of corrosion

inhibitors

HAA Speciation 3.3.3

3.3.3.1 Fresh Coupons

Figure 3-10 displays the speciation of HAA6 in Mannheim Water when in contact with

fresh metal coupons and in bulk water control as 1 mg/L orthophosphate was applied as a

R² = 1.00

0

50

100

150

200

0 3 6 9 12 15

HA

A6 (

µg

/L)

Chlorine demand (mg/L)

Point A

Point B

Assumed linear HAA formation

63

corrosion inhibitor. Similar results were obtained for Britannia Water, shown in Appendices

Figure 10-15. Since DBAA and MBAA concentrations were consistently low, only DCAA,

TCAA, MCAA and BCAA are shown. Generally, DCAA formation was dominating over the

production of TCAA for all metal coupons, which is consistent with other research (Rossman et

al., 2001). As a result of the reactions between free chlorine and metal coupons, less availability

of free chlorine in the solution limited the formation of TCAA (Rossman et al., 2001).

Figure 3-10 HAA speciation with time in the presence of 1 mg/L orthophosphate for fresh metal

coupons with Mannheim Water in one set of experiments, error bars indicate the measured

maximum and minimum values (n=2)

For iron coupons, the concentrations of DCAA, MCAA and BCAA leveled off after 24

hours. It is primarily because free chlorine was depleted within 24 hours, making one reactant for

HAA formation unavailable. On the other hand, TCAA concentration exhibited the decrease

trend with time after 24 hours. As reported by Hozalski et al. (2001), HAA compounds may

experience reductive transformations by Fe0 via sequential hydrogenolysis (replacement of a

halogen by hydrogen), but no degradation of TCAA was observed over 150 hours in the presence

0

15

30

45

60

75

0 20 40 60 80 100 120

HA

A (

µg

/L)

Time (hours)

Fe

DCAA TCAA

MCAA BCAA

0

15

30

45

60

75

0 20 40 60 80 100 120

HA

A (

µg

/L)

Time (hours)

Cu

0

15

30

45

60

75

0 20 40 60 80 100 120

HA

A (

µg

/L)

Time (hours)

Pb

0

15

30

45

60

75

0 20 40 60 80 100 120

HA

A (

µg

/L)

Time (hours)

Control

b

64

of goethite, magnetite and aqueous Fe(II) (Chun et al., 2005). XRD analysis was performed on

the surface of iron coupons for short term corrosion, and Fe0 has been identified (Appendices

Figure 10-16). It indicates that 24 hours was not long enough for iron coupons to be fully

oxidized by free chlorine, and the decrease in TCAA was primarily attributed to the reduction by

Fe0 present on the surface of corroded iron coupons. To fully understand the fate of HAA in the

water in contact with iron, experiments have been conducted to investigate the reduction

potential of iron on HAA under controlled experimental conditions, and results are provided in

Appendices 10.9.

For copper coupons, DCAA formation contributed to more than 60% of HAA6 after 24

hours, whereas TCAA, BCAA and MCAA concentrations were generally below 15 µg/L. For

lead coupons, DCAA formation kinetics was slightly faster than TCAA, and up to 22 µg/L of

MCAA formed after 24 hours. The relative contribution of DCAA, TCAA and MCAA to HAA6

was 40%, 30% and 20%, respectively. Similarly, less than 15 µg/L BCAA was formed for lead

coupons during the test period. Bulk water control had comparable DCAA and TCAA formation

rates. MCAA concentrations in bulk water control were consistently lower than DCAA and

TCAA, but higher than BCAA concentrations which were always below 15 µg/L.

3.3.3.2 Corroded Coupons

The observed HAA speciation after 48 hours of contact with chlorine for corroded

coupons in the presence of orthophosphate with Mannheim Water is shown in Figure 3-11.

MBAA, BCAA and DBAA were all <5 µg/L and so they are not shown. MCAA dominated over

the production of DCAA and TCAA for all the investigated metal coupons and the bulk water

control. The dominance of MCAA over DCAA for corroded metal coupons was different from

the observations for fresh metal coupons in the same water matrix. In the tests employing

corroded metal coupons, 5.5 mg/L free chlorine was applied, but a high concentration of free

chlorine (12.3 mg/L) was dosed in the tests employing fresh metal coupons. Therefore, the

difference in HAA speciation between these two tests was likely because the low concentration

of free chlorine applied for corroded metal coupons limited the formation of highly chlorinated

species, such as DCAA and TCAA (Rossman et al., 2001).

65


metal coupons with Mannheim Water in one set of experiments, error bars indicate the measured


Figure 3-12 demonstrates HAA speciation at 48 hours for corroded coupons in the

presence of orthophosphate with Britannia Water. DCAA formation was observed to dominate

over TCAA and MCAA formation. A similar trend was also observed for fresh metal coupons in

Britannia Water. The dominance of DCAA formation over other species, regardless of the

concentrations of free chlorine, indicates that free chlorine concentration may not affect relative

distribution of HAA species in Britannia Water.


metal coupons with Britannia Water in one set of experiments, error bars indicate the measured


0

10

20

30

40

50

Fe+ortho Cu+ortho Pb+ortho Control

HA

A (

µg

/L)

MCAA DCAA TCAA

0

10

20

30

40

Fe+ortho Cu+ortho Pb+ortho Control

HA

A (

µg

/L)

MCAA DCAA TCAA

66

3.4 Summary

This study compared the kinetics of free chlorine degradation and HAA formation as well

as HAA speciation in two water matrices in contact with three metal materials (either fresh or

aged) commonly found in distribution system water mains (e.g. Fe) and premise plumbing (e.g.

Cu and Pb). MS-FP and MS-SDS tests were conducted on a “survey” basis mainly to help

identify topics/theories for further research (later chapters), and thus no detailed mechanistic

investigation was explored at this point. Most of the results were consistent with those of other

researchers, providing confidence in the testing procedures. New or unexpected results were

either further explored in subsequent tests or simply identified for follow-up by other

researchers.

The following conclusions can be drawn based on the experimental results:

a. In agreement with previous studies, free chlorine degradation generally followed pseudo-

first-order kinetics, regardless of metal age and water quality. The experimental design

uniquely enabled the direct comparison of free chlorine degradation kinetics among three

metal materials. The reactivity of these pipe materials, for both fresh and aged ones, as

investigated in this study was: ductile iron >copper >lead.

b. New observations concerning the impacts of two different water quality on free chlorine

degradation showed that these impacts were dependent on metal type and metal age. The

low pH and alkalinity in Britannia Water led to increased free chlorine decay by

increasing iron corrosion, but only for fresh iron coupons. For all of the corroded metal

coupons, free chlorine degraded consistently more quickly in Mannheim Water than in

Britannia Water, most likely due to the higher TOC level in Mannheim Water.

c. The impacts of phosphate-based corrosion inhibitors on chlorine decay and HAA

formation were newly identified. The addition of these corrosion inhibitors generally

increased HOCl degradation for fresh iron coupons, and they decreased HOCl decay only

for fresh copper coupons. These corrosion inhibitors did not impact HOCl decay for both

fresh and pre-corroded lead coupons. Generally, phosphate-based corrosion inhibitors did

not impact HAA formation, regardless of metal type, metal age and water quality.

d. For fresh copper coupons in both investigated water matrices, HAA formation was

enhanced in the presence of high levels of copper ions, indicating possible catalytic

67

potential of copper on HAA formation. This new observation laid the groundwork for

further investigation of copper catalysis on HAA formation in Chapter 5.

e. Consistent with previous studies, DCAA was the dominant species when water was in

contact with fresh metal coupons. For corroded metal coupons, either MCAA or DCAA

formation was favored.

In addition, although the results from this survey study suggest that metal materials and

their interactions with corrosion inhibitors affected HOCl decay and HAA formation, these

impacts may be site-specific due to their dependence on metal age and water quality. This study

also reveals that the addition of phosphate-based corrosion inhibitors can decrease chlorine

demand, particularly for copper, and thereby allow for the maintenance of high disinfectant

residuals in distribution mains and premise plumbing. Although a higher level of residuals is

desirable to control microbial growth and pathogen amplification in large building such as

hospitals, it may also potentially increase halogenated DBP formation. Therefore, water utilities

may need to decrease their initial chlorine dosage in distribution systems when considering

applying these corrosion inhibitors in their systems.

One should also keep in mind that these tests were conducted for relatively new pipe

materials in bench-scale simulated distribution systems under the static condition. Although pre-

corroded metal coupons were employed in this study, they were aged only for up to one month.

Therefore, all of these results are mainly indicative of what may take place in dead-ends and

during stagnation within plumbing and distribution systems when new pipes are installed and/or

after these new pipes experience only short-term exposure to secondary disinfectants. Any

effects from hydrodynamics, pipe service age and microorganisms should also be taken into

account, and some of these effects were addressed in the pilot-scale experiments that were

performed for Chapter 7. Furthermore, since HAA concentrations may be reduced to meet new

regulations by switching from free chlorine to combined chlorine, but this can increase NDMA

formation, the impacts of copper catalysis on NDMA formation were also examined in

experiments that are summarized in Chapter 6.

68

3.5 References

Al-Jasser, A. O. (2007) Chlorine decay in drinking-water transmission and distribution systems:

Pipe service age effect. Water Research, 41(2), 387-396.

APHA, AWWA, WEF (2005) Standard Methods for the Examination of Water & Wastewater,

21th Edition, Washington D C, USA.





Blatchley, E. R., Margetas, D., and Duggirala, R. (2003) Copper catalysis in chloroform

formation during water chlorination. Water Research, 37(18), 4385-4394.

Brereton, J. A., and Mavinic, D. S. (2002) Field and material-specific simulated distribution



Chen, W.J. and Weisel, C.P. (1998) Halogenated DBP concentrations in a distribution system.


Chun, C. L., Hozalski, R. M., and Arnold, T. A. (2005( Degradation of drinking water

disinfection byproducts by synthetic goethite and magnetite. Environmental Science &

Technology, 39(21), 8525-8532.





Chemistry in Britain 20(10), 900-904.

Facey, R. M., and Smith, D. W. (1995) Soft, low-temperature water-distribution corrosion:


69

Fu, J., Qu, J. H., Liu, R. P., Qiang, Z. M., Liu, H. J., and Zhao, X. (2009b) Cu(II)-catalyzed

THM formation during water chlorination and monochloramination: A comparison study.

Journal of Hazardous Materials, 170(1), 58-65.

Gray, E. T., Taylor, R. W., and Margerum, D. W. (1977) Kinetics and Mechanisms of Copper-

Catalyzed Decomposition of Hypochlorite and Hypobromite - Properties of a Dimeric

Copper(III) Hydroxide Intermediate. Inorganic Chemistry, 16(12), 3047-3055.



36(14), 3479-3488.

Hallam, N. B., Hua, F., West, J. R., Forster, C. F., and Simms, J. (2003) Bulk decay of chlorine

in water distribution systems. Journal of Water Resources Planning and Management-

Asce, 129(1), 78-81.

Haas, C. N., Gupta, M., Chitluru, R., and Burlingame, G. (2002 Chlorine demand in disinfecting

water mains. Journal American Water Works Association, 94(1), 97-102.




Health Canada (2009) Guidance on Controlling Corrosion in Drinking Water Distribution

Systems. Water, Air and Climate Change Bureau, Healthy Environments and Consumer

Safety Branch, Health Canada, Ottawa, Ontario (Catalogue No. H128-1/09-595E).

Hozalski, R. M., Zhang, L., and Arnold, W. A. (2001) Reduction of haloacetic acids by Fe0:

Implications for treatment and fate. Environmental Science & Technology, 35(11), 2258-

2263.

Kanokkantapong, V., Marhaba, T. F., Panyapinyophol, B., and Pavasant, P. (2006) FTIR

evaluation of functional groups involved in the formation of haloacetic acids during the

chlorination of raw water. Journal of Hazardous Materials, 136(2), 188-196.

70


233-238.

Li, B., Liu, R. P., Liu, H. J., Gu, J. N., and Qu, J. H. (2008) The formation and distribution of


152(1), 250-258.

Li, S. Y., Lixiao, N. X., Sun, C., and Wang, L. S. (2004) Influence of organic matter on

orthophosphate corrosion inhibition for copper pipe in soft water. Corrosion Science,

46(1), 137-145.

Lu, W., Kiene, L., and Levi, Y. (1999) Chlorine demand of biofilms in water distribution




34-52.

Maddison, L.A. and Gagnon, G.A. (1999) Evaluating corrosion control strategies for a pilot scale

distribution system. In Proceedings of the 1999 American Water Works Association

Annual Conference, Tampa Bay, Fla. American Water Works Association, Denver, Colo.

McNeill, L. S., and Edwards, M. (2000) Phosphate inhibitors and red water in stagnant iron

pipes. Journal of Environmental Engineering-Asce, 126(12), 1096-1102.

Montgomery, D. C. (2000) Design and Analysis of Experiments, 5th edition, John Wiley &

Sons, New York.

Powell, J. C., Hallam, N. B., West, J. R., Forster, C. F., and Simms, J. (2000) Factors which

control bulk chlorine decay rates. Water Research, 34(1), 117-126.

Rodriguez, M.J., Serodes, J.B. and Levallois, P. (2004) Behavior of trihalomethanes and

haloacetic acids in a drinking water distribution system. Water Research, 38(20), 4367-

4382.

71



Schock M, Wagner I. (1985) The corrosion and solubility of lead in drinking water. In: Internal

Corrosion of Water Distribution Systems. Denver: AWWA Research Foundation,

Denver.

Scholze R.J.; Pontow K.A.; Kanchlbhatla G. and Ray B.T. (1994) Using the CERL pipe loops

system (PLS) to evaluate corrosion inhibitors that can reduce lead in drinking water

(FEAP-TR-EP-94/04). U.S. Army Construction Engineering Research Laboratories,

Champaign, IL.





493.

USEPA (1989) 40 CFR Parts 141 and 142 Drinking water; National Primary Drinking Water

Regulations; Filtration, Disinfection; Turbidity, Giardia lamblia, Viruses, Legionelia, and

Heterotrophic Bacteria; Final Rule. Federal Register, 54 (124), 27486-27541.

Vikesland, P. J., and Valentine, R. L. (2002) Modeling the kinetics of ferrous iron oxidation by

monochloramine. Environmental Science & Technology, 36(4), 662-668.


quality problems. Water Supply, 8 (1 and 2), 195-198.

72

A Comparison of Iron, Copper and Lead Corrosion in 4

Simulated Distribution Systems

Experiments that formed the basis for Chapters 3 and 4 compared the kinetics of metal

corrosion, chlorine degradation and HAA formation for iron, copper and lead treated with

different corrosion inhibitors on a “survey” basis. All of the metal release kinetics data were

collected at the same time as for Chapter 3, but are discussed as a separate chapter herein

because of the quantity of the data. Most of the observations regarding the impacts of corrosion

inhibitors, water quality, and disinfectant type follow the trends that have been reported by other

researchers, and thus no mechanistic investigations were involved in this chapter. Interpretations

concerning the impacts of these parameters on metal corrosion essentially provided valid

explanations for the findings in Chapter 3.

As such, the results in this chapter that deal with metal release kinetics for fresh metal

coupons in one type of water (Mannheim Water, Section 4.3.1) have been published as part of:

Zhang, H., Andrews, S. A. (2012) Effects of Phosphate-based Corrosion Inhibitors on the

Kinetics of Chlorine Degradation and HAA Formation in Contact with Three Metal

Materials. Canadian Journal of Civil Engineering, 39, 44-54.

Data from similar experiments using another water matrix (Britannia Water) and experiments

using pre-corroded coupons for both tested water matrices are included in this chapter for

completeness.

Generally, results from this chapter focus on the research gap "Do metal age and water

quality affect the efficacy of corrosion inhibitors?"

73

Abstract

Pipe materials and their corrosion products may affect the formation and the degradation

of disinfection byproducts in distribution systems. The investigation of factors influencing metal

corrosion under controlled experimental conditions is important for evaluating the impacts of

metal type, corrosion inhibitors and water quality on disinfectant degradation and DBP

formation. This study examined the effects of disinfectant type, metal age, phosphate-based

inhibitors, and water quality on metal corrosion at bench scale. Orthophosphate had no beneficial

effects to control iron corrosion, regardless of water quality, disinfectant type or the age of the

metal surface. However, orthophosphate significantly decreased released copper concentrations,

in particular for fresh coupons during short term exposures, and it also significantly reduced lead

release, irrespective of the age of the metal surface or water quality. In two investigated water

matrices, the water with lower pH and alkalinity generally exhibited higher corrosion potential to

metal coupons. Only for pre-corroded copper coupons, HOCl was consistently more aggressive

than NH2Cl. The results of XPS surface analysis suggest that the effectiveness of orthophosphate

on copper and lead corrosion control was due to facilitated precipitation of calcium. For iron

coupons, CaCO3 precipitation decreased the amount of iron oxides formed on the surface.

Keywords: Ductile iron; Copper; Lead; Corrosion inhibitor; Secondary disinfectant; XPS

74

4.1 Introduction

Corrosion is defined as “the deterioration of a material that results from a reaction with

its environment” (NACE International, 2009). In drinking water distribution systems, iron pipes

are widely used, while in household plumbing systems copper pipes are commonly found.

Although the use of lead in drinking water applications has been eliminated in Canada since

1992, lead may be leached into drinking water from old lead service lines, soldered joints and

brass plumbing fittings (Health Canada, 2009). The established drinking water guideline for lead

based on its health effects in children is 0.010 mg/L. Guidelines for iron and copper based on

aesthetic considerations are ≤0.3 mg/L for iron and ≤1.0 mg/L for copper (Health Canada, 2009).

Metal corrosion in distribution systems is affected by many physical and chemical

factors. Among water quality parameters, pH and alkalinity are the most significant factors

influencing metal levels in distribution systems. It has been consistently reported that the

solubility of lead, copper and iron decreases with elevated pH (McNeill and Edwards, 2001;

Sarin et al., 2003; Schock, 1989; Xiao et al., 2007). Normally, lower iron corrosion rates and

iron concentrations in distribution systems have been associated with higher alkalinity due to the

formation of less soluble siderite (FeCO3) (Sarin et al., 2003). In contrast, copper corrosion

release was increased at higher alkalinity due to the formation of soluble cupric bicarbonate and

carbonate complexes (Edwards et al., 2002; Edwards et al., 1996; Schock and Lytle, 1995). The

degree to which alkalinity affects lead solubility depends on the water pH and the form of lead

carbonate present on the pipe surface (Schock, 1980; Schock, 1990).

Free chlorine (HOCl) has been observed to increase metal corrosion rates due to its high

oxidation potential relative to monochloramine (NH2Cl) (Boulay and Edwards, 2001; Cantor et

al., 2003; LeChevallier et al., 1990a; Rahman et al., 2007; Vasquez et al., 2006). However,

ammonia (a product from chloramines hydrolysis) may increase the corrosivity of water on

copper tubing by forming stable copper-ammonia complexes (Boyd et al., 2008; Schock and

Lytle, 1995).

Addition of phosphate-based corrosion inhibitors is an effective strategy to control

corrosion. The basis for orthophosphate addition as a method of corrosion control relies on the

fact that impervious solid films will be developed as a barrier between the corroding metal and

75

the water (Demora and Harrison, 1984; Singley, 1994), whereas polyphosphates are strong

chelating agents which have been successfully used for reducing scaling and sequestrating Fe2+

ions to treat “red water” (Facey and Smith, 1995; Maddison et al., 2001; Williams, 1990).

Although orthophosphate may reduce copper release in the short term by forming a cupric

phosphate scale, in the long term it may interfere with the natural evolution of scales to form the

least soluble species, malachite, leading to increased copper concentrations in the water (Cantor

et al., 2000). Some studies have also reported detrimental effects of polyphosphate on lead

corrosion control due to the formation of lead polyphosphate complexes (Leroy, 1993;

MacQuarrie et al., 1997).

Although considerable studies have been focused on understanding the effects of

corrosion inhibitors and disinfectant type on metal corrosion, most of these studies were

performed in pilot-scale and full-scale distribution systems with aged metal pipes. In these

studies, the impacts of corrosion inhibitors, disinfectant type and water quality on metal

corrosion may have been somewhat confounded by hydrodynamics and microorganisms.

Therefore, the objective of the current study was to assess the effects of phosphate-based

inhibitors, disinfectant type and water quality on metal corrosion for different metal ages under

controlled experimental conditions at bench scale. HOCl and NH2Cl were applied as secondary

disinfectants, and two water matrices with distinctly different water quality were examined in

regards to their corrosion potential. Metal coupons of ductile iron, copper and lead were

employed. These pipe materials are widely present in either water mains (such as ductile iron) or

household plumbing systems (such as copper), and lead may be leached into drinking water from

old lead service lines, soldered joints and brass plumbing fittings (Health Canada, 2009). Results

of this study will help utilities and households identify possible reasons for increased corrosion

rates to aid in taking effective measures to reduce corrosion in their systems.





hypochlorite (NaOCl, 6%, VWR) in Milli-Q water. The NH2Cl dosing solution was preformed

76

by adding the free chlorine dosing solution to a 50 mmol/L ammonium chloride solution in a

well-stirred 250 mL amber bottle at a Cl2/N molar ratio of 0.8:1. The target disinfectant

concentrations applied in the tests were achieved by spiking the disinfectant dosing solution into

1 L test water. An initial disinfectant concentration of 12.3 mg/L as Cl2 was applied in

experiments with fresh coupons, and an initial disinfectant concentration of 5.5 mg/L as Cl2 was

applied for experiments employing corroded metal coupons. Orthophosphate (Na3PO4) and

polyphosphate ([Na(PO2)]6) were selected as corrosion inhibitors to investigate their impacts on

metal corrosion. They were prepared at a concentration of 500 mg/L as P and were kept in the

dark at 4ºC. The targeted dosage of both corrosion inhibitors was 1 mg/L as P.

Test coupons of ductile iron, copper and lead were purchased from Metal Samples Co.,

Alabama, US. The size of these coupons was 1 2”3”1 1 ”. For experiments with fresh

coupons, before each batch of experiments, any corrosion products were removed from the

coupons with 60-grit sandpaper followed by 120-grit sandpaper, and then rinsed with deionized

water and acetone followed by Milli-Q water (Scholze et al., 1994). Polished coupons were

submerged into the test water to simulate contact with new pipes without any impacts from

service age. In addition, to investigate the aged pipe environment, some coupons of each pipe

material were pre-conditioned in tap water in the absence and presence of orthophosphate. The

initial free chlorine concentration for conditioning purposes was 10 mg/L. After each 24 hour

reaction period for the iron coupons and each 48 hour reaction period for the copper and lead

coupons, the free chlorine residual was measured and then the water was replaced with fresh tap

water dosed with 10 mg/L free chlorine. The conditioning process proceeded until the 24 hour

chlorine residual for each combination of metal type and corrosion inhibitors was approximately

constant (< 4.0 mg/L chlorine decay for the iron coupons alone, 6.8 mg/L for the iron coupons

treated with orthophosphate, 2.7 mg/L for all of the copper coupons, and 1.9 mg/L for all of the

lead coupons). The measured 24 hour chlorine residuals for iron, copper and lead are shown in

Appendices Figure 10-14. Orthophosphate was selected since it can effectively control metal

release for fresh metal coupons in a short time, especially for copper and lead.

Unchlorinated post-filtration water was collected from two water treatment plants in

Ontario (Mannheim Water and Britannia Water). Water quality parameters are listed in Table

4-1.

77

Table 4-1 Summary of water quality parameters for two post-filtration water sources

Parameters Mannheim Water Britannia Water

pH 7.5 ± 0.2 6.6±0.5


UV254 (cm-1

) 0.058±0.008 0.065±0.005

TOC (mg/L) 4.5±0.3 3.3±0.2

SUVA (L/mg·cm-1) 0.013±0.002 0.019±0.0002

Bromide (μg L) 65.0±15.5 15.8±0.2

Chloride (mg/L) 84.5±2.5 3.1±0.6

Sulfate (mg/L) 35.0±2.0 23.8±0.8

Cl-:SO4

2- ratio 2.4±0.1 0.13±0.04


“Material-specific” simulated distribution system (MS-SDS) and material-specific

formation potential (MS-FP) tests were applied in this study. Details of MS-SDS procedures

have been described by Brereton and Mavinic (2002). In experiments with corroded metal

coupons, MS-SDS tests were performed. In experiments with fresh metal coupons, MS-FP tests

were carried out similarly to MS-SDS tests, except that MS-FP tests applied a higher

concentration of initial disinfectant (12.3 mg/L) than would be encountered in distribution

systems to meet the high chlorine demand of the fresh metal coupons and ensure detectable

disinfectant residuals (0.2 mg/L) after 24 hours.

All of the MS-SDS and MS-FP experiments were performed in 1 L amber bottles with

PTFE lined caps and at room temperature (21±2 ºC). The reaction bottles were made chlorine

demand free before use. Metal coupons of single metal species were suspended in the bottles

using nylon threads. The testing of galvanic effects resulting from combinations of different

metal coupons was beyond the scope of this investigation. To maximize the contact of water with

coupon surface, a Big Bill Orbital shaker was used (Barnstead International) to maintain gentle

mixing at a speed of 25 rpm. After each designated reaction time, 50 mL of water were

withdrawn from the test bottles for the analysis of disinfectant residual and metal concentrations.

Seven to ten time points were employed for each of the kinetic studies. As will be discussed in

Section 4.3.1, metal concentrations reached equilibrium after 10 hours, and the reduction of

water volume due to sequential sampling did not significantly affect equilibrated metal

concentrations. In addition, since all of the tests were conducted in closed systems, the impact of

78

oxygen on metal corrosion was considered to be negligible and thus its concentration was not

measured in this study.



control samples, which were prepared the same as test samples but without metal coupons.

Analytical methods employed to examine analytes of interest are summarized in Table 4-2.



TOC mg/L O1-Analytical TOC analyzer SM1 5310 C

pH VWR Scientific Model 8015 pH

meter

UVA254 cm-1

Hewlett Packard 8452A Diode

Array UV spectrophotometer SM 5910B

Alkalinity mg/L Titration SM 2320B

Chlorine mg/L Hach DR2700 Spectrophotometer SM 4500-Cl G

Monochloramine mg/L Hach DR2700 Spectrophotometer Hach method 101712

Anions mg/L Dionex DX-300 Series Ion

Chromatography System SM 4110 B

Metal (Fe, Cu and Pb) mg/L Varian SpectrAA.20 SM 3111 B

Note: 1. SM represents Standard Methods for the Examination of Water and Wastewater APHA et al., 2005. 2. Hach, 2007;

Coupon Surface Analysis 4.2.3

Metal surface analysis, including X-ray diffraction (XRD) and X-ray photoelectron

spectroscopy (XPS), were performed on metal surfaces to identify corrosion products which may

govern the solubility of metal in the solution. For fresh metal coupons, XRD analysis was

performed at the University of Toronto (Department of Chemistry) to identify solid-phase

corrosion products. XRD is a versatile, non-destructive analytical technique that reveals the

mineralogy of the dominant scale solids. For iron coupons, non-uniform corrosion layers were

formed on the coupon surface as a result of short-term reactions, and thus the XRD analysis was

carried out on powders which were scratched from the coupon surface and put on low-

background silicon sample holders (Tang et al., 2006). For lead and copper coupons, the analysis

was performed by placing the samples directly in the diffractometer. A Siemens D5000

Diffractometer System operating at 50 kV/35 mA was used to collect the diffraction patterns. A

high-power, line focus Cu-K-source was used combined with a solid state Kevex detector for

79

elimination of K-lines. The experimental data were collected on a step scan mode (0.02° /1.5

sec) within the most informative range (2-theta degrees). The obtained data were processed by

various Diffrac Plus software packages including Eva 8.0 and Topas v. 2.1.

Surface analysis of corroded metal coupons was performed with XPS at the University of

Toronto (Department of Chemical Engineering). Before XPS analysis, corroded metal coupons

were stored in an anaerobic glove box to remove adsorbed moisture. A Thermo Scientific K-

Alpha XPS system (East Grinstead, UK) was used to scan all of the metal samples. A

monochromatic Al K-Alpha X-ray source was used. The vacuum pressure was approximately

1x10-7

mbar. The survey spectra were acquired with high pass energy (200 eV) and low point

density (1 eV step size). The regional data, used for quantitative evaluation and fitting, was

acquired with a pass energy of 50 eV and a high point density (0.1 eV). The data acquired from

the instrument was processed using the software Avantage (provided by the manufacturer). Any

charging shift produced by the samples was carefully removed by using a 284.6 eV adventitious

carbon (C1s) standard.


Metal Release and Phosphate-based Corrosion Inhibitors 4.3.1

4.3.1.1 Fresh Coupons in Mannheim Water

Iron, copper and lead release kinetics in the absence and presence of corrosion inhibitors

under chlorination in Mannheim Water is illustrated in Figure 4-1. Iron, copper and lead release

kinetics under chloramination exhibited a similar pattern with and without the addition of

corrosion inhibitors, the data for which are shown in Appendices Figure 10-17. Generally, metal

concentrations increased rapidly with increasing reaction time initially, and then more gradually

increased or stabilized after 10 hours. Lower dissolved metal concentrations were observed for

copper and lead coupons when in the presence of the corrosion inhibitors. However, for iron

coupons, phosphate-based corrosion inhibitors did not reduce the amount of iron release under

both chlorination and chloramination, which was also evidenced by McNeill and Edwards (2000

and 2001).

80


corrosion inhibitors with HOCl in Mannheim Water; disinfectant concentrations, 12.3 mg/L;


The X-ray diffraction patterns of iron powders obtained from the scratched surface

substances of freshly oxidized iron coupons in the absence and presence of different corrosion

inhibitors demonstrate almost identical diffraction patterns for these iron coupons (Figure 4-2).

This may explain the similar metal release kinetics for fresh iron coupons regardless of the

presence of corrosion inhibitors. The main component on the surface of three oxidized iron

coupons was identified as poorly crystalline goethite, α-FeOOH.

0

2

4

6

8

10

0 5 10 15 20 25 30F

e (

mg/L

) Time (hours)

HOCl

FeFe+polyphosphateFe+orthophosphate

0

1

2

3

4

0 20 40 60 80 100 120

Cu (

mg/L

)

Time (hours)

Cu

Cu+polyphosphate

Cu+orthophosphate

0

1

2

3

4

0 20 40 60 80 100 120

Pb (

mg/L

)

Time (hours)

Pb

Pb+polyphosphate

Pb+orthophosphate

81

Figure 4-2 Comparison of XRD patterns for iron powders scratched from the surface of oxidized

iron coupons in the absence and presence of corrosion inhibitors

For fresh copper coupons, phosphate-based corrosion inhibitors significantly reduced

copper release for this water matrix during the short-term kinetic studies in comparison with

copper coupons alone (Figure 4-1), which is consistent with previous studies. Orthophosphate at

1 mg/L reduced copper concentration more effectively than polyphosphate, with the copper

concentration generally below 0.52 mg/L. Results of the XRD analysis demonstrate that only

Cu2O was present on the surface of tested coupons. Figure 4-3 illustrates the relative distribution

of Cu2O on the surface of tested coupons in the absence and presence of corrosion inhibitors. A

smaller percentage of Cu2O (2.25%) in the presence of orthophosphate suggested that

orthophosphate could block active sites on the metal surface, and consequently protect metallic

copper against oxidation (Dartmann et al., 2004). In the presence of polyphosphate, relatively

higher concentrations of copper (approximately 1.1 mg/L) were detected compared with the

orthophosphate treated coupons, but these copper concentrations were consistently lower than

those for copper coupons alone (up to 2.5 mg/L).

The reduction of lead release in the presence of 1 mg/L orthophosphate was also

observed for both HOCl and NH2Cl in Mannheim Water. In contrast, polyphosphate significantly

increased lead solubility, especially when HOCl was used (Figure 4-1). Surface analysis of lead

coupons by XRD indicated that the surface of these coupons had very similar compositions

(Figure 4-4). Aside from the large reflections from the lead, small peaks of lead oxide, litharge

(PbO), could be identified. While hydrocerussite, Pb3(CO3)2(H2O)2, was also possibly present,

PbO was the dominant species.

α-FeO(OH)

α-FeO(OH)

α-FeO(OH)

Fe+poly

Fe+ortho

Fe

10 20 30 40 50 60 70 80

2-theta

Inte

nsity

— Reference pattern of α-FeO(OH)

82

Figure 4-3 Comparison of XRD patterns for copper coupons in the absence and presence of

corrosion inhibitors.

Figure 4-4 Comparison of XRD patterns for lead coupons in the absence and presence of 1 mg/L

orthophosphate

For all of the investigated metal coupons, no phosphate-containing solids were identified

on the surface of coupons exposed to the phosphate-spiked water. Since these tests were initiated

by submerging fresh metal coupons into the test water immediately after free chlorine addition, it

is possible that these phosphate-containing solids were poorly crystalized within 100 hour

reaction time so as not to be identified by XRD (Hsu, 1982; Moriarty, 1990). Nevertheless,

phosphate corrosion inhibitors may still form an amorphous or semi-amorphous protective film

15 20 30 40 50 60 2-theta

Pb+ortho Pb

Pb

O

Pb

O

Pb Pb Pb

Pb

3(C

O3) 2

(OH

)

Inte

nsity

— Reference pattern of Pb; — Reference pattern of PbO; — Reference pattern of Pb3(CO3)2(OH)2

Pb

O

Sample Cu, %. Cu2O, % Cu-control 100.00 0 Cu+ortho 97.75 2.25 Cu+poly 97.07 2.93 Cu 94.58 5.42

Cu Cu+poly Cu+ortho Cu+poly/ortho

Cu control

Cu2O

Cu Cu

Cu2O

28 30 40 50 60

2-theta (degree)

— Reference pattern of Cu2O

Inte

nsity

83

and provide a barrier between metal surface and chlorinated water, thereby reducing the released

metal concentrations. The addition of phosphate corrosion inhibitors may also facilitate the

precipitation of already released metal ions, as insoluble/low soluble metal salts (such as copper

phosphate and lead phosphate), thereby causing the measured metal concentrations to be lower in

the presence of phosphate corrosion inhibitors than those in the absence of corrosion inhibitors

(McNeill and Edwards, 2004). In the case of polyphosphate-treated lead coupons, the increase in

lead concentrations relative to lead coupons alone (especially under chlorination) can be

explained such that polyphosphate increased the solubility of already released lead ions by

complexation (Edwards and McNeill, 2002; Leroy, 1993; MacQuarrie et al., 1997).

4.3.1.2 Corroded Coupons in Mannheim Water

Metal release kinetics for corroded iron, copper and lead coupons was also investigated.

The released metal concentrations following 24 hour exposure under chlorination and

chloramination in Mannheim Water are illustrated in Figure 4-5. Irrespective of the type of

disinfectant, orthophosphate at 1 mg/L increased iron release by at least 90% compared with that

in the absence of orthophosphate. This is likely because orthophosphate increased the solubility

of iron by forming iron-phosphate complexes, as evidenced by McNeill and Edwards (2000). For

corroded copper coupons under chlorination, orthophosphate increased copper release by

approximately 24%, which is different from the observations in the experiments employing fresh

coupons (Figure 4-1). However, this agrees with previous reports in which orthophosphate

interfered with the natural evolution of scales, reducing the formation of the least soluble species

(malachite) and thus having an adverse impact on cuprosolvency for aged pipes, especially in

chlorinated distribution systems (Cantor et al., 2000; Schock and Lytle, 1995; Schock and

Sandvig, 2009). Under chloramination, however, orthophosphate was still effective to reduce

copper concentration. In terms of corroded lead coupons, orthophosphate effectively minimized

lead corrosion irrespective of disinfectant type. Lead concentrations in the presence of

orthophosphate were consistently below 0.1 mg/L under both conditions.

The impacts of phosphate-based corrosion inhibitors on iron and lead release in Britannia

Water were similar to the trends observed with Mannheim Water that orthophosphate had no

beneficial effects on iron release but effectively suppressed released lead concentrations for both

84

Figure 4-5 Metal concentrations after 24 hours for corroded coupons in the absence/presence of

orthophosphate with HOCl and NH2Cl in Mannheim Water, initial disinfectant concentrations

5.5 mg/L; error bars indicate the measured maximum and minimum values (n=2)

fresh and corroded coupons regardless of disinfectant type (Appendices, Figure 10-18 and Figure

10-19). Interestingly, orthophosphate also effectively suppressed copper release for both fresh

and corroded copper coupons under chlorination and chloramination in Britannia Water, while in

Mannheim Water orthophosphate increased the copper corrosion rate for corroded copper

coupons under chlorination (Figure 4-5). The different impacts of orthophosphate on corroded

copper corrosion in the two water matrices indicate that some aspect of the water quality affects

the effectiveness of orthophosphate to control copper corrosion. The dependency of

orthophosphate on water quality to control copper corrosion has also been reported by Dartmann

et al. (2004).

XPS Results for Corroded Coupons in Mannheim Water 4.3.2

Surface analysis by X-ray photoelectron spectroscopy (XPS) was performed on corroded

metal coupons after kinetics experiments with Mannheim Water to identify the compositions of

the corrosion products on the surfaces of the metal coupons and to evaluate the impacts of

disinfectant type and orthophosphate on metal corrosion. Survey scans provided an elemental

analysis of each sample, and high resolution scans allowed specific products to be identified.

0.0

0.2

0.4

0.6

0.8

Me

tal co

nce

ntr

atio

n (

mg

/L)

HOCl NH2Cl

85

4.3.2.1 Survey Scans

Through the survey scan, the elements of carbon, calcium and oxygen were detected on

the surface of most of the corroded metal coupons. Phosphorus was detected only on the surface

of coupons exposed to the orthophosphate-spiked water. For corroded copper and lead coupons,

the percentage of calcium increased significantly when orthophosphate was applied, indicating

that the application of orthophosphate facilitated the precipitation of calcium. The similar

observations were also reported by Stone (2008). XPS survey scan results for corroded copper,

iron and lead coupons under chloramination are displayed in Figure 4-6. The XPS survey scans

for three types of corroded metal coupons under chlorination are displayed in Appendices Figure

10-20.

As shown in Figure 4-6, calcium on the surface of corroded copper coupons increased

from 0.5% in the absence of orthophosphate to 2% when orthophosphate was applied. For lead,

6% of calcium was detected for lead coupons exposed to the orthophosphate-spiked water, but

no calcium was found on the surface of lead coupons in the absence of orthophosphate. The

binding energy of the calcium peak (347.2~347.4 eV) indicated that calcium was present

primarily in combination with phosphate. The formation of calcium phosphate may function as

an additional protective layer to decrease the exposure of the copper surface to the disinfectant

residual, leading to lower levels of copper released into the test water (as shown in Figure 4-5).

However, for corroded iron coupons, calcium did not vary significantly before and after the

addition of orthophosphate, indicating that there was no enhanced precipitation of calcium by

phosphate on the surface of the corroded iron coupons. In this case, the detected phosphorous

was likely present primarily in the form of iron-phosphate complexes since phosphate anions

have a high affinity for iron (Persson et al., 1996). As a result, these complexes would increase

the solubility of iron in the water and result in elevated iron concentrations when orthophosphate

was applied (Figure 4-5).

86

Figure 4-6 Comparison of elemental distribution for copper, lead and iron coupons in the

absence and presence of orthophosphate with NH2Cl

4.3.2.2 High-Resolution Scans

Identification of corrosion products were performed through deconvolutions of high-

resolution XPS scans for the corroded iron, copper and lead coupons (Figure 4-7). For corroded

44%

2% 5%

42%

7%

Cu+NH2Cl+orthophosphate

C

Ca

Cu

O

P

52%

0.5 6%

41%

Cu+NH2Cl

C

Ca

Cu

O

58% 17%

25%

Pb+NH2Cl

Pb

C

O

7%

11%

58%

8%

6%

10%

Pb+ortho+NH2Cl

Pb

C

O

Al

Ca

P

21%

56%

0.9

22%

Fe+NH2Cl

Fe

O

Ca

C

17%

51%

1.2

24%

7%

Fe+ortho+NH2Cl

Fe

O

Ca

C

P

87

Figure 4-7 Corrosion products of Fe, Cu and Pb as well as their relative distribution in the scales

iron coupons, magnetite (Fe3O4) and hydrated ferric oxide (FeOOH) were detected in the scales

regardless of the presence of orthophosphate and disinfectant type. Binding energies for Fe3O4

and FeOOH were 710~710.2 eV and 711.4~711.7 eV, respectively. Both compounds account for

similar fractions in the corrosion scales for all four of the investigated conditions. Under

chlorination, more iron was oxidized in the presence of orthophosphate than in the absence of

0

2

4

6

8

0

2

4

6

8

10

HOCl HOCl+ortho NH2Cl NH2Cl+ortho

Calc

ium

(%

)

Fe,

%

Fe3O4 FeOOH Calcium

0

2

4

6

8

10


Cu,

%

Cu2O CuO Cu(OH)2

0

10

20

30

40

50


Pb,

%

Pb PbO2 PbO Pb3(OH)2(CO3)2

88

orthophosphate (4.4% vs. 2.6%), indicating that orthophosphate did not protect the metal surface

against oxidation and thus was not beneficial for iron corrosion control. However, under

chloramination, the addition of orthophosphate decreased the amount of oxidized iron in the

scales. Therefore, although orthophosphate was observed to be detrimental for iron release in the

two water matrices investigated in this study, its effects on solid corrosion scales under HOCl

and NH2Cl were inconclusive.

In addition, it was interesting to observe that iron oxides in the scales of iron coupons

under chlorination were less abundant than those under chloramination, and that the sequence of

the amount of oxidized iron in the corrosion scales for four investigated situations was opposite

to the trend in the amount of calcium in the scales. The binding energy of calcium was

346.6~347eV, indicating that it was present mainly in the form of CaCO3. Therefore, CaCO3

played a significant role in determining the abundance of oxidized iron in the corrosion scales,

which has not been reported previously. It followed that greater amounts of CaCO3 precipitation

lead to smaller amounts of iron oxides formed on the metal surface.

Copper corrosion products on the corroded copper coupon surface included Cu2O

(932.2~932.5 eV), CuO (934.0~934.3 eV), Cu(OH)2 and/or CuCO3 (935.2~935.4 eV). Copper

orthophosphate solids, such as cupric phosphate dihydride [Cu3(PO4)2·2H2O] and cupric

phosphate [Cu3(PO4)2], were also possibly present on the surface of copper coupons exposed to

orthophosphate, but they were not identified conclusively. For all of the investigated conditions,

Cu2O was the most abundant species compared with other copper corrosion products, and thus

was the controlling solid phase for copper solubility. In addition, relatively greater amounts of

oxidized copper were detected for coupons exposed to HOCl than to NH2Cl. For example, in the

absence of orthophosphate, 6.7~9.0 % of copper was present in the scale of copper coupons

under chlorination, but only 3% was present under chloramination. This was primarily because

the high oxidation potential of HOCl accelerated copper corrosion rates (Schock and Lytle,

1995). The addition of orthophosphate for both HOCl and NH2Cl tended to decrease the

oxidation rate of Cu(0) to Cu(I) and Cu(II). All this is in agreement with previous studies in

which orthophosphate was able to form an insoluble film on the metal surface to prevent copper

from oxidation, thereby reducing copper release (Edwards et al., 2002).

89

For corroded lead coupons, PbO (litharge, 139~139.1 eV), Pb3(OH)2(CO3)2

(hydrocerussite, 138.1 eV), PbCO3 (cerussite, 138.4 eV), and PbO2 (plattnerite, 137.1~137.4 eV)

were identified as possible corrosion products. The formation and transformation of these lead

species changed with disinfectant type and the presence/absence of orthophosphate. Generally,

oxidized lead was present in the scales of lead coupons in the presence of orthophosphate but

was less abundant than on coupons in the absence of orthophosphate. Only 3.1% and 3.3% of

lead was in the oxidized form in the presence of orthophosphate under chlorination and

chloramination, respectively, but 25% and 43.7% were oxidized in the absence of

orthophosphate with free chlorine and chloramine, respectively. This was most likely because the

facilitated precipitation of calcium by orthophosphate formed a solid Ca3PO4 film and thus

protected lead against further oxidation, which is also evidenced by lower concentrations of lead

in the presence of orthophosphate in Figure 4-5. In addition, PbO2 was detected as a predominant

oxide for lead coupons without the addition of orthophosphate under chlorination (Figure 4-7).

This is mainly because HOCl is a strong oxidant which can oxidize lead to a high oxidation state

(+4), and orthophosphate inhibits the oxidation of Pb(II) to Pb(IV) by forming an insoluble scale

on the metal surface. It is reasonable to find a small amount of Pb3(OH)2(CO3)2 (hydrocerussite)

in the corrosion scales for lead coupons in contact with HOCl since PbO2 can be formed through

further oxidation of Pb3(OH)2(CO3)2 by HOCl (Kim and Herrera, 2010). For lead coupons

exposed to orthophosphate under chlorination and lead coupons alone under chloramination,

Pb3(OH)2(CO3)2 (hydrocerussite) and/or PbCO3 (cerussite) were dominant species, whereas PbO

was a solid controlling phase for lead coupons in the presence of orthophosphate under

chloramination.

Water Quality and Disinfectant Type 4.3.3

The ability of orthophosphate to mitigate impacts of water quality parameters (pH,

alkalinity and NOM) and disinfectant type (HOCl vs NH2Cl) was tested with fresh and pre-

corroded coupons and two water types. As shown in Table 4-1, Britannia Water had a lower pH

(6.60.5) and a significantly lower level of alkalinity (12 mg/L as CaCO3) than Mannheim

Water, pH and alkalinity of which were 7.5 ± 0.2 and 187 mg/L as CaCO3. Furthermore, HOCl

has a higher oxidation potential than NH2Cl. Therefore, it was hypothesized that Britannia Water

would be more corrosive than Mannheim Water, and more metal ions would be released under

chlorination than under chloramination. To test these hypotheses, the released metal

90

concentrations that were measured after exposure to similar initial concentrations of HOCl and

NH2Cl were compared. The results are displayed in Figure 4-8.

Figure 4-8 Comparison of iron concentrations at 24 hours in the presence of orthophosphate for

fresh and corroded coupons under free chlorine and chloramine, error bars indicate the measured


0

2

4

6

8

10

MannheimWater:fresh

MannheimWater:

corroded

BritanniaWater:fresh

BritanniaWater:

corroded

Iro

n (

mg/L

)

HOCl NH₂Cl

0.0

0.3

0.6

0.9

1.2

MannheimWater:fresh

MannheimWater:

corroded


BritanniaWater:

corroded

Copper

(mg/L

)

HOCl NH₂Cl

0.0

0.2

0.4

0.6

MannheimWater:fresh

MannheimWater:

corroded


BritanniaWater:

corroded

Lead (

mg/L

)

HOCl NH₂Cl

91

For either fresh or corroded metal coupons, generally, there were greater amounts of

metals released into Britannia Water than Mannheim Water. A paired student’s t test at a

confidence level of 95% was applied to all of the metal concentration values in Mannheim Water

and Britannia Water, and its p-value was 0.03, confirming that Britannia Water was statistically

more corrosive to metal coupons than Mannheim Water, likely due to the lower pH and

alkalinity values of Britannia Water.

For iron coupons, as shown in Figure 4-8, the impacts of HOCl and NH2Cl on the

released iron concentrations were inconsistent regarding each type of water and coupon ages. For

fresh iron coupons, NH2Cl increased the released iron concentrations in Mannheim Water but

exhibited no significant difference relative to HOCl in Britannia Water. For corroded iron

coupons, NH2Cl tended to be less aggressive in Mannheim Water but more aggressive in

Britannia Water. It is interesting that Cantor et al. (2003) and LeChevallier et al. (1990b) have

reported an increased iron corrosion rate under chlorination than under chloramination. Perhaps

they employed different water matrix and metal age, thereby leading to a different conclusion

about the impacts of disinfectant type on iron corrosion.

For fresh copper coupons, a higher concentration of copper was released under

chloramination than chlorination in both water matrices. This observation was different from the

trend reported in several studies, in which HOCl was more aggressive towards copper than

chloramine due to its higher oxidation potential (Boulay and Edwards, 2001; Boyd et al., 2008;

Cantor et al., 2003; Rahman et al., 2007). However, few studies have been conducted to compare

the impacts of disinfectants on copper release for new copper pipes. In chloraminated copper

plumbing, ammonia or ammonium ions, a byproduct from chloramine decay, tends to form

stable complexes with copper ions [Cu(I) and Cu(II)] and thus increase the solubility of copper

corrosion solids (Boyd et al., 2008; Schock and Lytle, 1995). Furthermore, the stability of

dissolved Cu(I)-ammonia complexes may hasten corrosion processes by reactions such as

[Cu(NH3)4]2+

+ Cu(s) 2[Cu(NH3)2]+ (Cotton and Wilkinson, 1988), especially in relatively

new copper pipes where a large area of pure copper is exposed to an oxidizing environment. For

corroded copper coupons, as shown in Figure 4-8, copper concentrations resulting from HOCl

exposure were approximately 2 fold higher than those under NH2Cl conditions in both water

matrices. This observation is consistent with the results of most studies in that HOCl is more

aggressive to copper corrosion than NH2Cl.

92

For both fresh and corroded lead coupons in Mannheim Water, as shown in Figure 4-8,

the released lead concentrations after 24 hours of chlorination (0.5 mg/L and 0.07 mg/L for fresh

and corroded coupons, respectively) were significantly higher than those under chloramination

(0.2 mg/L and 0.03 mg/L for fresh and corroded coupons, respectively). In Britannia Water,

however, there was no significant difference in lead concentrations between HOCl and NH2Cl

for fresh lead coupons after 24 hours, but NH2Cl tended to increase lead dissolution for corroded

coupons. Normally, the higher oxidation potential of HOCl promotes the formation of less

soluble lead dioxide (PbO2), whereas divalent lead solids [e.g., Pb3(CO3)2(OH)2] will form under

chloramination. As such, an increased lead level generally has been observed in chloraminated

water (Boyd et al., 2009; Edwards and Dudi, 2004). However, the transition kinetics of Pb(II) to

Pb(IV) depends on initial chlorine concentration, hydrodynamics, and water quality parameters

such as alkalinity and NOM. Only a dense surface layer of PbO2 could dramatically decrease

lead release (Boyd et al., 2008). Therefore, the observed increase in lead concentrations under

HOCl than those exposed to NH2Cl for fresh coupons in two water matrices were possibly

because these coupons had undergone oxidation by HOCl only for 24 hours and PbO2 was not

readily formed during the short-term reaction. Instead, Pb(II)-containing solids were most likely

the controlling solid phase for lead solubility, as was evidenced by the XRD results (Section

4.3.1). For corroded lead coupons in Mannheim Water, due to relatively high levels of alkalinity

and NOM in Mannheim Water, the transformation of Pb(II) to Pb(IV) was likely inhibited.

Therefore, the high HOCl oxidation potential also induced more lead to be released in the

chlorinated Mannheim Water. In contrast, for corroded coupons in Britannia Water, PbO2 was

likely the dominant solid phase under HOCl, and its low solubility determined that the lead

concentrations after a further 24 hour exposure was significantly reduced compared with that

under NH2Cl. Overall, the inconsistent effects of HOCl and NH2Cl on lead corrosion in two

water matrices indicate that water quality and metal age may confound the impacts of

disinfectant type, thereby affecting the compositions of solid controlling phase and the

subsequent dissolution of lead into the water (Edwards and Dudi, 2004; Kim and Herrera, 2010).

4.4 Summary

This chapter summarizes the metal species and concentration data that were obtained

during the experiments described in Chapter 3, presented separately here due to the quantity of

data. In this chapter, the effects of disinfectant type, metal age, water quality and phosphate-

93

based inhibitors on metal corrosion were evaluated. The test coupons of ductile iron, copper and

lead, representing either water main or premise plumbing, were investigated. The following

conclusions can be drawn based on the experimental results:

a. The observations concerning the impacts of orthophosphate on released iron, copper and

lead concentrations were generally consistent with those of other researchers.

Orthophosphate had no impacts on the released iron levels, regardless of metal surface

age, water quality, or disinfectant type. However, orthophosphate significantly reduced

the released copper concentrations, in particular for fresh copper surfaces and short term

exposures. Lead release was significantly reduced in the presence of orthophosphate,

irrespective of the age of the metal surface or water quality involved.

b. Results of XPS surface analysis suggest that the effectiveness of orthophosphate on

copper and lead corrosion control was due to facilitated precipitation of calcium by

orthophosphate to form calcium phosphate, which is consistent with previous studies.

The role of CaCO3 in determining the abundance of oxidized iron on the iron surface was

newly identified. Surface characterization of corroded iron coupons demonstrates that

CaCO3 precipitation decreased the amount of iron oxides formed on the iron surface.

c. In agreement with other research, high levels metal ions of iron, copper and lead were

released in waters with low pH and alkalinity.

d. The experimental design uniquely enabled the direct comparison of released metal

concentrations under chlorination and chloramination. However, the impacts of HOCl

and NH2Cl on the released iron, copper and lead concentrations for both fresh and pre-

corroded metal materials in the two tested water matrices were inconclusive. Only for

pre-corroded copper coupons was HOCl consistently more aggressive than NH2Cl in both

Mannheim and Britannia water.

In addition, since all of the data in this chapter were collected at the same time as those in

Chapter 3, they can also help provide some explanations for the trends in chlorine decay and

HAA formation that were described in Chapter 3. For example, for fresh copper coupons in

Mannheim water, phosphate-based corrosion inhibitors significantly sequestered copper release

compared with copper coupons in the absence of corrosion inhibitors. Due to the protection from

these corrosion inhibitors, the copper surface was also less readily available to react with

94

chlorine, leading to decreased free chlorine degradation rates in comparison with copper coupons

in the absence of corrosion inhibitors.

4.5 References

APHA, A., WEF, (2005) Standard Methods for the Examination of Water & Wastewater. 21th

Edition, Washington D C, USA.




Giani, R. (2008) Effects of changing disinfectants on lead and copper release. Journal










95-102.



Cotton, F. A., and Wilkinson, G. (1988) Advanced Inorganic Chemistry. Fifth Edition. John

Wiley & Sons, Inc., New York.




95









Edwards, M., Schock, M. R., and Meyer, T. E. (1996) Alkalinity, pH, and copper corrosion by-

product release. American Water Works Association Journal, 88(3), 81-94.

Hach (2007) Chloramine (Mono) - Indophenol Method 10171 (DOC316.53.01015), edition 6.

Hach Company, Loveland, Colorado.


Systems (Catalogue No. H128-1/09-595E). Water, Air and Climate Change Bureau,


Hsu, P. H. (1982) Crystallization of Iron(Iii) Phosphate at Room-Temperature. Soil Science

Society of America Journal, 46(5), 928-932.

Kim, E. J., and Herrera, J. E. (2010) Characteristics of Lead Corrosion Scales Formed during

Drinking Water Distribution and Their Potential Influence on the Release of Lead and

Other Contaminants. Environmental Science & Technology, 44(16), 6054-6061.

LeChevallier, M. W., Lowry, C. D., and Lee, R. G. (1990) Disinfecting Biofilms in a Model

Distribution System. American Water Works Association Journal, 82(7), 87-99.


233-238.

96



34-52.



313.



McNeill, L. S., and Edwards, M. (2001) Iron pipe corrosion in distribution systems. American

Water Works Association Journal, 93(7), 88-100.

McNeill, L. S., and Edwards, M. (2004) Importance of Pb and Cu particulate species for

corrosion control. Journal of Environmental Engineering-Asce, 130(2), 136-144.

Moriarty, B. E. (1990) Surface Studies of Corrosion-Inhibitors in Cooling Water-Systems.

Materials Performance, 29(1), 45-48.

NACE International (2009) NACE International glossary of corrosion-related terms. National

Association of Corrosion Engineers International, Houston, TX.

Persson, P., Nilsson, N., and Sjoberg, S. (1996) Structure and bonding of orthophosphate ions at

the iron oxide aqueous interface. Journal of Colloid and Interface Science, 177(1), 263-

275.



Asce, 133(2), 180-185.

Sarin, P., Clement, J. A., Snoeyink, V. L., and Kriven, W. W. (2003) Iron release from corroded

unlined cast-iron pipe. Journal American Water Works Association, 95(11), 85-96.

Schock, M. R. (1980) Response of lead solubility to dissolved carbonate in drinking water.


97

Schock, M. R. (1989) Understanding corrosion control strategies for lead. Journal American


Schock, M. R. (1990) Causes of temporal variability of lead in domestic plumbing systems.


Schock, M. R., and Lytle, D. A. (1995) Effect of pH, DIC, Orthophosphate and Sulfate on

Drinking Water Cuprosolvency. U.S. Environmental Protection Agency, Water Supply

and Water Resources Division, Cincinnati, Ohio 45268, EPA/600/R-95/085.

Schock, M. R., and Sandvig, A. M. (2009) Long-term effects of orthophosphate treatment on

copper concentration. Journal American Water Works Association, 101(7), 71-82.

Scholze R.J.; Pontow K.A.; Kanchlbhatla G. and Ray B.T. (1994) Using the CERL pipe loops

system (PLS) to evaluate corrosion inhibitors that can reduce lead in drinking water

(FEAP-TR-EP-94/04). U.S. Army Construction Engineering Research Laboratories,

Champaign, IL.



Stone, E. D. (2008) Effects of Orthophosphate Corrosion Inhibitor in Blended Water Quality

Environments. PhD Dissertation, University of Central Florida.

Summers, R. S., Hooper, S. M., Shukairy, H. M., Solarik, G., and Owen, D. (1996) Assessing the

DBP yield: Uniform formation conditions. Journal American Water Works Association,

88(6), 80-93.

Tang, Z., Hong, S., Xiao, W., and Taylor, J. (2006) Characteristics of iron corrosion scales

established under blending of ground, surface, and saline waters and their. Corrosion

Science, 48(2), 322-342.



Association, 98(2), 144-154.

98


quality problems. Water Supply, Vol 8, Nos 1 and 2, 195-198.

Xiao, W. Z., Hong, S. K., Tang, Z. J., and Taylor, J. S. (2007) Effects of blending on total copper


99

Catalytic Impacts of Copper Corrosion Products on 5

Chlorine Decay and HAA Formation in Simulated

Distribution Systems

The results in this chapter have been published as:

Zhang, H., Andrews, S.A. (2012) Catalysis of Copper Corrosion Products on Chlorine Decay

and HAA Formation in Simulated Distribution Systems. Water Research, 46 (8), 2665-2673.

Results from this chapter focus on the research gap “What are the impacts of metal

corrosion products on the fate of HAA in distribution systems?”

100

Abstract

This study investigated the effect of copper corrosion products, including Cu(II), Cu2O,

CuO and Cu2(OH)2CO3, on chlorine degradation, HAA formation, and HAA speciation under

controlled experimental conditions. Chlorine decay and HAA formation were significantly

enhanced in the presence of copper with the extent of copper catalysis being affected by the

solution pH and the concentration of copper corrosion products. Accelerated chlorine decay and

increased HAA formation were observed at pH 8.6 in the presence of 1.0 mg/L Cu(II) compared

with that observed at pH 6.6 and pH 7.6. Further investigation of chlorine decay in the presence

of both Suwannee River NOM and Cu(II) indicated that an increased reactivity of NOM via

interacting with dissolved and/or solid surface-associated Cu(II), rather than chlorine auto-

decomposition, was a primary reason for the observed rapid chlorine decay. Copper corrosion

solids [Cu2O, CuO, Cu2(OH)2CO3] exhibited catalytic effects on both chlorine decay and HAA

formation. Contrary to the results observed when in the absence of copper corrosion products,

DCAA formation was consistently predominant over other HAA species in the presence of

copper corrosion products, especially at neutral and high pH. This study improves the

understanding for water utilities and households regarding chlorine residuals and HAA

concentrations in distribution systems, in particular once the water reaches domestic plumbing

where copper is widely used.

Keywords: Copper; Corrosion; Catalysis; Free chlorine; Haloacetic acids

101

5.1 Introduction

To maintain the microbial stability of distributed water, free chlorine is used as a

secondary disinfectant in distribution systems. As chlorine may experience temporal and spatial

degradation due to chemical and biological consumption that occurs in the bulk water and on the

pipe wall, at least 0.2 mg/L of free chlorine residual should be maintained to reduce the possible

occurrence of biological regrowth (USEPA, 1989).

Haloacetic acids (HAAs) are formed primarily by the chlorination of natural organic

matter (NOM). In its simplest terms, NOM can be classified as either humic substances or non-

humic compounds. However, humic substances are a mixture of many molecules, and the major

functional groups which contribute to surface charge and reactivity of humic substances include

carboxyl groups, some phenolic groups, alcohol groups, methoxyl groups, ketones, and

aldehydes (Reckhow et al., 1990). Humic substances in the environment are capable of

interacting with metal ions to form soluble complexes, colloidal substances and/or insoluble

substances (Stevenson, 1994). Alvarez-Puebla et al. (2004a, 2004b) have investigated possible

retention mechanisms of Cu(II), Co(II) and Ni(II) on humic substances and found surface

complexation and electrostatic retention played key roles in the retention of these metals on

humic substances, and Cu(II) had a higher affinity for humic substances compared with Co(II)

and Ni(II). Furthermore, the humic substances studied displayed a great selectivity for different

Cu(II) species, and the selectivity depended on pH.

Since the 1970s, considerable effort has been made to understand HAA formation

mechanisms by using NOM or well-defined model compound precursors (Kanokkantapong et

al., 2006; Morris, 1975; Reckhow et al., 1990). Reaction mechanisms involved in HAA

formation have been found to generally include oxidation, substitution, addition, and hydrolysis

(Morris, 1975). The reaction rates and HAA speciation are dependent on temperature, pH,

chlorine dose and the nature of the organic compounds that contribute to NOM. Although there

are nine HAA species, monochloroacetic acid (MCAA), dichloroacetic acid (DCAA),

trichloroacetic acid (TCAA), monobromoacetic acid (MBAA) and dibromoacetic acid (DBAA)

are commonly detected in drinking water, and DCAA and TCAA are normally dominant over

other species. Due to the toxicological properties of HAAs, the Stage 2

102

Disinfectant/Disinfection Byproduct Rule (D/DBPR) has regulated a Maximum Contaminant

Level (MCL) of HAA5 (MCAA, DCAA, TCAA, MBAA, and DBAA) at 0 μg L (USEPA,

2006). The Guidelines for Canadian Drinking Water Quality have established a Maximum

Acceptable Concentration (MAC) for HAA5 in drinking water at 80 μg L, based on a locational

running annual average of a minimum of quarterly samples taken in the distribution system

(Health Canada, 2008).

Copper corrosion occurs as water containing a disinfectant residual travels through

copper pipes. Copper ions [Cu(II)], cuprite [Cu2O], tenorite [CuO], cupric hydroxide [Cu(OH)2],

and malachite [Cu2CO3(OH)2] are dominant corrosion products of copper in drinking water

systems (Xiao et al., 2007). Copper has no acute toxicity to humans except at high dose (>15

mg/day). An aesthetic objective of ≤1.0 mg/L has been established for copper in drinking water

(Health Canada, 1992).

The catalytic potential of copper in many reaction processes has been widely investigated

(Onuchukwu, 1994; Paidar et al., 1999; Pintar et al., 1997). Other studies have also demonstrated

copper catalysis on free chlorine and monochloramine degradation and on THM formation

(Blatchley et al., 2003; Fu et al., 2009a; Fu et al., 2009b; Li et al., 2007; Li et al., 2008). The

mechanism of Cu-catalyzed THM formation proposed by Blatchley et al. (2003) was that copper

could complex with THM precursor compounds and enhance the oxidative decarboxylation and

enolization of the keto-groups.

Since HAA formation has been found to be concurrent with THM formation, with both

DBP classes having similar organic precursors and some similarities in their formation

mechanisms (Morris, 1975; Reckhow and Singer, 1985), and elevated HAA formation in the

presence of high levels of copper ion was observed in Chapter 3, it is hypothesized that copper

will play a similar catalytic role in HAA formation. Li et al. (2008) have reported elevated

formation of HAAs in copper pipes. However, they did not examine in detail the catalytic

potential of Cu(II) and other corrosion products to affect chlorine degradation and HAA

formation. Furthermore, a regulatory guideline for HAA5 has recently been added to the

“Guidelines for Canadian Drinking Water Quality” (Health Canada, 2008). This information,

therefore, has increased the awareness of water utilities and households to minimize HAAs in

their water systems. As such, the objective of this study was to investigate the effect of Cu(II)

103

and its solid corrosion products on chlorine degradation and HAA formation, including HAA

speciation, by applying different concentrations of copper corrosion products and under different

pH conditions. Since copper is widely used in domestic plumbing systems, understanding the

roles of copper and its corrosion products on the fate of HAAs will benefit utilities and

households to minimize HAAs in their systems by adopting preventive strategies for their

control.





hypochlorite (NaOCl, 6%, VWR) in Milli-Q water. The target chlorine concentration applied in

the tests was achieved by spiking 3.5 mL chlorine dosing solution in 1 L test water. An initial

free chlorine concentration of 10 mg/L as Cl2 was applied to achieve a detectable 24-h residual of

0.2 mg/L as Cl2, which was found to be necessary especially when the effects of Cu2O were

investigated. Phosphate buffer (H2PO4-/HPO4

2-) and borate buffer (H3BO3/NaOH) solutions were

prepared and added to test solutions at 1 mM concentration to control reaction solution pH at

desired levels. Sulfuric acid (H2SO4, 50%) and sodium hydroxide (NaOH, 20%) were also used

for pH adjustment. L-Ascorbic acid (99.0%, Sigma Aldrich) at 200 mg/L was applied to quench

chlorine residuals before HAAs extraction. Over the reaction period, the pH was maintained at

their specified target values (± 0.1).

Unchlorinated post-filtration water was collected from the Mannheim Water Treatment

Plant (MWTP), Ontario, for testing. Water quality parameters are listed in Table 5-1. In the tests

to determine reaction pathways of Cu(II) catalysis on chlorine decay, synthetic water with

similar inorganic chemical composition to the water from the MWTP was prepared. Suwannee

River NOM (International Humic Substances Society) was dosed into synthetic water at the

concentration of 2.4 mg/L and 4.7 mg/L.

104

Table 5-1Water quality parameters for post-filtration water from MWTP

Parameters Parameter values

pH 7.5 ± 0.2

UV254 (cm-1

) 0.045 ± 0.015

TOC (mg/L) 3.9 ± 0.4

SUVA (L/mg·cm-1

) 0.016 ± 0.004

Bromide (μg L) 65.0 ± 15.5

Chloride (mg/L) 84.5 ± 2.5

Sulfate (mg/L) 35.0 ± 2.0

Cl-:SO4

2- ratio 2.4 ± 0.1


All of the experiments were performed in 1 L amber bottles with PTFE-lined caps and at

room temperature (21±2 ºC). The reaction bottles were made chlorine demand free before use.

All of the tests also included control samples, which were prepared in the same way as for test

samples but without the addition of copper corrosion products.

Copper ions were spiked in the form of copper nitrate [Cu(NO3)2] at concentrations of 0,

0.2, 1.0 and 2.0 mg/L of copper to be compatible with the common range of copper

concentrations in distribution systems of 20-2020 µg/L (Health Canada, 1992). Cuprous oxide

(Cu2O), cupric oxide (CuO) and malachite [Cu2CO3(OH)2] were selected as the solid forms of

copper corrosion products of interest due to their prevalence in copper pipes. They were added

into test solutions in the form of powders and mixed with test water thoroughly using a stir bar.

The concentrations investigated in this study were 0, 0.2 and 1.0 g/L, which were comparable to

the dosages applied by Li et al. (2007). After each prescribed reaction time (between 2 and 120

hours), water samples dosed with corrosion products were taken from the reaction bottles and

filtered through 0.2 µm Nylaflo® Nylon membrane filter paper (Pall Corporation) to separate the

solid copper corrosion products from the aqueous solution. Dissolved copper in the filtrate was

measured by flame atomic absorption spectrometry (FAAS). To account for any influence from

filter paper and filtration performance on the concentration of chlorine residual and HAA9,

control samples were also filtered under similar conditions. When the effects of pH on chlorine

decay and HAA formation were investigated (Sections 5.3.1.1 and 5.3.2.1), the solution pH was

controlled with 1 mM phosphate or borate buffer at pH 6.6, 7.6 or 8.6 to represent a possible pH

range in distribution systems. In the tests to investigate the effects of copper concentration

(Sections 5.3.1.3 and 5.3.2.2) and solid copper corrosion products (Sections 5.3.1.4 and 5.3.2.3),

105

and the tests to determine reaction pathways of Cu(II) catalysis on chlorine decay (Section

5.3.1.2), the pH of both natural and synthetic water was controlled at 8.3 ± 0.1 by the addition of

1 mM borate buffer.

In addition, to evaluate the impacts of NOM on copper solubility, copper ions were

spiked into Milli-Q water and MWTP water in the form of Cu(NO3)2 at concentrations of 0, 0.2,

1.0 and 2.0 mg/L of copper, and pH was controlled at 8.3 with 1 mM borate buffer. The copper-

spiked water was then filtered through 0.2 µm Nylaflo® Nylon membrane filter paper, and

copper concentrations in the filtrate were measured by FAAS.

In this study, free chorine decay rates (Section 5.3.1) were estimated by fitting a pseudo-

first-order decay equation to free chlorine data using Microsoft Excel. HAA formation rates

(Section 5.3.2) were determined by fitting HAA concentrations versus reaction time with a

logarithmic function in Microsoft Excel. A one-factor analysis of variance (ANOVA) test at a

confidence level of 95% was applied to determine whether each treatment factor, including pH

and Cu(II) concentrations, had significant impacts on free chlorine decay and HAA formation

(Montgomery, 2000).

Nine species of haloacetic acids (HAA9) were analyzed according to EPA Standard

Method 6251 B with 2,3,5,6-tetrafluorobenzoic acid as a surrogate standard (APHA, AWWA,

WEF, 2005). Measurement of total organic carbon (TOC) concentration was undertaken using a

Model 1030 TOC analyzer (OI Analytical, USA). UV254 was measured using a CE 3055

Reflectance Spectrophotometer (Cecil Instruments, England) at 254 nm. Measurements of free

chlorine were performed using a Hach DR2700 Spectrophotometer (Hach Company, USA) at

530 nm. Soluble copper concentrations in the filtered water samples were analyzed using a

Varian SpectrAA.20 Flame Atomic Absorption Spectrometer (Agilent Technologies, USA) after

being acidified to pH <2.


The concentrations of free chlorine and HAA9 were monitored by periodically

withdrawing and analyzing samples from the test bottles. Therefore, it was possible to compare

free chlorine decay rates and HAA9 formation kinetics among samples with and without spikes

106

of Cu(II) and/or copper corrosion products and evaluate their catalytic potential on free chlorine

degradation and HAA formation.

Chlorine Decay 5.3.1

5.3.1.1 Effects of pH

In the absence of Cu(II), there was no statistical difference in chlorine decay among the

examined three pH values (p-value >0.05). Pseudo-first-order decay rates were estimated to be

0.0018, 0.0022, and 0.0022 h-1

for pH 6.6, pH 7.6 and pH 8.6, respectively. These results are in

agreement with previous findings (Powell et al., 2000). However, when 1 mg/L Cu(II) was spiked,

as shown in Figure 5-1, free chlorine degraded more rapidly at higher pH. Compared with chlorine

decay rates in the absence of Cu(II), no significant effect was observed at pH 6.6 (p-value >0.05),

but chlorine degradation was significantly accelerated at pH 7.6 and 8.6 (p values of 0.02 and

0.003, respectively). Pseudo first-order decay rates in the presence of Cu(II) at pH 7.6 and 8.6 were

0.0033 and 0.0061 h-1

, respectively.

Figure 5-1 Chlorine degradation at different pH values in the absence and presence of 1 mg/L Cu

(II) for MWTP water; initial HOCl =10 mg/L; triplicate

Both MINEQL+ version 4.5 (Environmental Research Software, Hallowell, Maine, USA)

and some simple filtration tests were used to determine if the spiked copper remained dissolved

during these tests or pH manipulations caused some solid species to form. At pH 6.6 and 7.6,

malachite [Cu2(OH)2CO3] was estimated by MINEQL+ to be the primary form of the solids, and

Cu(OH)2 was the dominant solid at pH 8.6. The concentrations of these solid copper species were

estimated to increase with increasing pH, with 76%, 93% and 100% of the total 1 mg/L spiked

0.0

0.2

0.4

0.6

0.8

1.0

0 20 40 60 80 100 120

C/C

0

Time (hours)

pH 6.6-with Cu

pH 7.6-with Cu

pH 8.6-with Cu

107

copper being transformed into solids at pH 6.6, pH 7.6 and pH 8.6, respectively. These results

were confirmed by filtering and measuring dissolved copper concentrations in the filtrate of

samples that had been prepared in Milli-Q water. However, even at pH 8.6, measured dissolved

copper concentrations in the filtered MWTP water accounted for 93% of the spiked total 1 mg/L

copper, leaving only 7% of copper possibly being present in solid form. This significant increase in

copper solubility in natural water, relative to the results by MINEQL simulation for Milli-Q water

support reports that the presence of NOM in natural water facilitates the formation of Cu-NOM

complexes due to its strong binding affinities and thus hindering the formation of solid copper

species (Korshin et al., 1996). Regardless, although copper had been added in its highly soluble

Cu(NO3)2 form, it did not always remain in dissolved form and many samples had both dissolved

and solid copper species present. As such, at this point in the experiments, faster degradation of

chlorine at higher pH in the presence of Cu(II) as observed herein was hypothesized to have been

due to an accelerated reactivity of organic precursors that had interacted with dissolved and/or

solid copper species at high pH (Blatchley et al., 2003; Fu et al., 2009b) and/or the decomposition

of chlorine (as OCl-) to O2 and Cl

- by Cu(II) in alkaline solutions (Gray et al., 1977). Experiments

described in the following sections were performed to investigate these possibilities.

5.3.1.2 Effects of NOM Concentration

The relative contribution of auto-decomposition and consumption (through reacting with

organic matter) to the observed rapid chlorine decay in the presence of Cu(II) at high pH was

examined using synthetic water with similar inorganic chemical composition to the test water

described in Section 5.3.1.1. Chlorine decay kinetics were compared between solutions with and

without the addition of 1 mg/L Cu(II) and Suwannee River NOM.

As shown in Figure 5-2, in the absence of Suwannee River NOM, no statistical difference

in chlorine decay was observed when comparing results obtained with synthetic water and

synthetic water spiked with Cu(II) (p-value >0.05). This indicates that chlorine auto-

decomposition was not affected by copper catalysis within the 25-h test period of this study.

When 2.4 mg/L Suwannee River NOM was spiked, chlorine degraded significantly faster

(k=0.012 h-1

) with a p-value of 0.007 compared with water in the absence of NOM (k=0.001 h-1

)

as a result of the consumption of chlorine by NOM. When 1.0 mg/L Cu(II) was added to NOM-

spiked water, the chlorine decay rate further increased to 0.048 h-1

, significantly higher than that

108

without the addition of Cu(II) (p-value of 0.001). Figure 5-2 also shows that as the NOM

concentration was doubled in the absence and presence of 1.0 mg/L Cu(II), the chlorine decay

rates increased to 0.055 h-1

and 0.584 h-1

, respectively. Almost no residual was detected after 6

hours for water spiked with 4.7 mg/L NOM and 1.0 mg/L Cu(II). These results suggest that the

observed rapid chlorine degradation was primarily attributed to interactions of NOM with Cu(II)

(dissolved and/or solid phases) and subsequent increases in NOM reactivity with free chlorine.

Figure 5-2 Chlorine decay for synthetic water in the presence and absence of 1 mg/L Cu(II) and

NOM; initial HOCl =4.2 mg/L; pH 8.3; duplicate

5.3.1.3 Effects of Copper Concentration

Cu(NO3)2 were spiked into MWTP water at concentrations of 0 mg/L, 0.2 mg/L, 1.0

mg/L, and 2.0 mg/L of copper. Approximately 10 mg/L chlorine was spiked, and the chlorine

residual was monitored for the following 100 hours. Figure 5-3 compares the observed chlorine

degradation kinetics with respect to the different spiked Cu(II) concentrations. The chlorine

degradation rates significantly increased with an increase in the spiked Cu(II) concentrations

from 0 to 1.0 mg/L (p-value of 1.810-10

), but no additional difference in chlorine degradation

rate was observed for water dosed between 1.0 mg/L and 2.0 mg/L Cu(II) (p-value >0.05). The

increased chlorine degradation in the presence of Cu(II) may be explained by copper’s potential

to complex with organic compounds and enhance DBP formation by accelerating the

decarboxylation and enolization steps involved in DBP formation pathways (Blatchley et al.,

2003). Consequently, more chlorine may be involved in electrophilic addition and substitution

reactions with organic precursors. Since chlorine decay and DBP formation are not independent

0.0

0.3

0.6

0.9

1.2

0 5 10 15 20 25

C/C

0

Time (hours)

Synthetic water

Synthetic water+Cu(II)

Synthetic water+2.4 mg/LNOM

Synthetic water+2.4 mg/LNOM+Cu(II)

Synthetic water +4.7mg/L NOM

Synthetic water+4.7 mg/LNOM+Cu(II)

k=0.012 h-1

k=0.048 h-1

k=0.055 h-1

k=0.584 h-1

k=0.001 h-1

109

phenomena, further discussion of the effects of Cu(II) concentration on HAA formation is

provided in Section 5.3.2.2.

Figure 5-3 Chlorine degradation in the presence of different concentrations of Cu(II) for MWTP

water; initial HOCl =10 mg/L; triplicate; pH 8.3

5.3.1.4 Effects of Solid Copper Corrosion Products

CuO, Cu2O and Cu2(OH)2CO3, all commonly identified copper corrosion products, were

studied to identify their effects on chlorine degradation. Compared with control samples,

chlorine degraded more quickly in the presence of these corrosion solids. The observed sequence

of chlorine degradation in the presence of copper corrosion solids was 1.0 g/L Cu2O > 0.2 g/L

Cu2O > 1.0 g/L malachite > 0.2 g/L malachite 1.0 g/L CuO > control (Figure 5-4). The

concentrations of Cu(II) that are shown for solutions containing these solid corrosion products

were the equilibrated concentrations of Cu(II) released from the solids. Released Cu(II)

concentrations that were measured after 20 hours of reaction were at levels of approximately 0.8

mg/L for solutions of both 1.0 g/L Cu2O and 0.2 g/L Cu2O, and approximately 0.2 mg/L for

solutions of 1.0 g/L and 0.2 g/L malachite. These results indicate that the dissolved Cu(II)

reached an approximate equilibrium with the solids during the tests and that the equilibrium

Cu(II) concentrations measured agree with published solubility data for these compounds (Li et

al., 2008; Merkel et al., 2002).

0.0

0.2

0.4

0.6

0.8

1.0

0 20 40 60 80 100 120

C/C

0

Time (hours)

0 mg/L

0.2 mg/L

1.0 mg/L

2.0 mgL

110

Figure 5-4 Pseudo-first-order decay rates of free chlorine for MWTP water containing dissolved

Cu(II) and solid copper corrosion products

Li et al. (2007, 2008) have reported that the effects of CuO and Cu2O on the degradation

of chlorine were actually due to the effect of Cu(II) released from the oxides. To determine if the

released Cu(II) from copper corrosion solids also played a significant role in the observed

difference in chlorine decay kinetics among these solid corrosion products in the present study,

the pseudo-first-order decay constants for free chlorine were compared with those for free

chlorine in contact with only dissolved Cu(II) in the filtered samples spiked with Cu(NO3)2

(Figure 5-4). In all cases, the decay constants were obtained by fitting a pseudo first-order decay

equation to the chlorine data once the released Cu(II) concentration reached equilibrium after

approximately 20 hours reaction time.

Chlorine decay rates were higher in the presence of both solid corrosion products (0.2-

1.0 g/L) and their associated dissolved Cu(II) (0.1-0.8 mg/L) than in the solutions spiked with

only Cu(NO3)2 (up to 2 mg/L of copper). For reactions involving only the dissolved form of

Cu(II), free chlorine decay rates increased approximately linearly with increases in the dissolved

Cu(II) concentrations from 0 mg/L to 1.0 mg/L (0.004 h-1

- 0.0117 h-1

) as shown in the inset plot

in Figure 5-4. In the presence of solid copper corrosion products, free chlorine decay rates were

significantly higher relative to those for solutions with similar aqueous dissolved Cu(II)

concentrations. For example, the decay constant for free chlorine in contact with 1.0 g/L CuO

was 0.0193 h-1

and the released Cu(II) concentration at equilibrium was 0.15 mg/L; whereas a

decay rate of only 0.0053 h-1

would be estimated from the curve at the same concentration of

0.0

0.1

0.2

0.3

0.0 0.2 0.4 0.6 0.8 1.0

Chlo

rine d

ecay c

onsta

nt (h

-1)

Measured dissolved Cu(II) (mg/L)

Cu(II) without solids 0.2 g/L Cu2O

1.0 g/L Cu2O 1.0 g/L CuO

0.2 g/L malachite 1.0 g/L malachite

0.000

0.005

0.010

0.015

0.020

0.0 0.2 0.4 0.6 0.8 1.0

111

dissolved Cu(II) in the absence of corrosion solids. In addition, however, increases in mass

concentrations of solids did not lead to the same magnitudes of increases in free chlorine

degradation rates. Therefore, the quantitative relationship between mass concentrations of copper

corrosion solids and free chlorine degradation needs further investigation. Regardless, the

observed significant increase in the decay constants for free chlorine in the presence of these

solid corrosion products indicates that not only the dissolved copper species, but also surface-

associated Cu(II) from these corrosion products, may be involved in the reactions and accelerate

free chlorine degradation.

These results are in contrast to those reported by Li et al. (2007, 2008) who discovered

that rapid chlorine decay in the presence of CuO and Cu2O was actually due to the effect of

Cu(II) released from the oxides. Therefore, differences in surface area of the solids may explain

some of the differences in the results from these two studies. For reactions mediated by surface-

bound metal, the reaction rate is affected by the mineral surface area and the density of sorbed

metal ions (Chun et al., 2005; Lee et al., 2008). Qualitatively, it follows that the total surface

area and hence the concentrations of reactive sites from 1.0 g/L Cu2O and malachite are larger

than for 0.2 g/L Cu2O and malachite, respectively. As a result, it is not surprising that free

chlorine degraded faster with 1.0 g/L Cu2O and malachite than with 0.2 g/L Cu2O and malachite.

However, due to the lack of information about surface area for the Cu2O, CuO and malachite

investigated in this study, normalization of the measured decay constants and subsequent more

quantitative comparisons of the reactivity of these solid corrosion products with free chlorine

cannot be made.

HAA Formation and Speciation 5.3.2

5.3.2.1 Effects of pH

The effect of pH on Cu(II) catalysis during HAA formation was investigated by

controlling the pH of MWTP water at 6.6, 7.6 and 8.6 and comparing HAA formation between

Cu(II)-spiked and unspiked water samples. Results indicate that copper catalysis during HAA

formation was highly dependent on pH, and the degree of enhancement of HAA formation by

Cu(II) became more apparent at higher pH values. After 100 hours of reaction time, as shown in

Figure 5-5, there was no statistical difference in HAA formation before and after the addition of

112

Cu(II) at pH 6.6 and 7.6 (p values >0.05). However, at pH 8.6, HAA formation significantly

increased with 1 mg/L Cu(II) as compared with the control (p-value of 0.006). Since

complexation of Cu(II) with organic matter is formed via Cu(II) interactions with carbonyl and

hydroxyl groups (Alvarez-Puebla et al., 2004b), and copper was present primarily in the form of

complexes with NOM in MWTP water at high pH (evidenced by 93% of copper being present in

the filtrate after passage through 0.2 µm filter paper), the observed effects of pH on HAA

formation in the presence of Cu(II) suggest that high pH promoted the electrostatic interactions

of dissolved and/or solid surface-associated Cu(II) with organic matter by favoring the ionization

of the aforementioned carbonyl and hydroxyl groups (Blatchley et al., 2003). Accelerated base-

catalyzed enolization and hydrolysis involved in HAA formation from a model precursor has

also been shown to contribute to elevated HAA formation in the presence of Cu(II) (Deborde and

von Gunten, 2008).

Figure 5-5 HAA9 at 100 hours at different pH values in the absence and presence of 1 mg/L

Cu(II) for MWTP water, initial HOCl =10 mg/L (error bars represent standard deviation of

triplicate tests)

5.3.2.2 Effects of Copper Concentration

The catalytic effect of Cu(II) on HAA formation under conditions of varying Cu(II)

concentrations is illustrated in Figure 5-6. HAA9 had similar formation kinetics in the first 30

hours for all of the water samples tested. When reactions proceeded further (t >70 hours), more

HAA compounds were produced for the water spiked with higher concentrations of Cu(II). There

was a significant increase in HAA9 formation with 1.0 mg/L Cu(II) after 100 hours of reaction

compared with HAA9 concentrations formed in control water (p-value of 3.210-6

). However, a

0

50

100

150

200

250

pH 6.6 pH 7.6 pH 8.6

HA

A9 (

µg

/L)

Cu(II) 0 mg/L Cu(II) 1.0 mg/L

113

further increase in the spiked copper concentration from 1.0 mg/L to 2.0 mg/L did not affect

HAA yields or formation rates significantly (p-value >0.05). Although the ability of Cu(II) to

complex with organic matter is essential for copper to act as a catalyst (Bansal et al., 2008), the

Cu(II) binding capacity of humic acid is limited, having been measured in the range of 48 to 160

mg/g of humic acid (Stevenson, 1985). The absence of a difference in HAA formation for the

spiked Cu(II) concentrations between 1.0 mg/L and 2.0 mg/L suggests that the functional groups

or binding sites of the organic molecules in the solution could have been saturated by 1.0 mg/L

of copper, regardless of its speciation. Since the reactivity of NOM with chlorine was enhanced

by the retention of dissolved and/or surface-bound Cu(II) on NOM (Sections 5.3.1.2, 5.3.1.3, and

5.3.1.4), it follows that the reactivity of NOM with chlorine would be increased with increasing

Cu(II) concentration until the active sites on NOM are completely occupied by Cu(II).

Figure 5-6 HAA9 formation in the presence of Cu(II) with varying concentrations at pH 8.3 for

MWTP water (error bars represent standard deviation of triplicate tests)

5.3.2.3 Effects of Solid Copper Corrosion Products

Copper corrosion solids were also observed to catalyze HAA formation at pH 8.3, the

extent of the effect being dependent on their species and mass concentrations. Compared with

the control sample, 1.0 g/L CuO had similar HAA formation, but 0.2 g/L Cu2O significantly

promoted HAA formation (p-value of 0.001) (Figure 5-7). No statistical difference in HAA

formation was observed for waters spiked with 0.2 g/L malachite and the control samples.

However, in the water spiked with 1.0 g/L of malachite, the HAA9 formation also significantly

increased compared with the control sample (p-value of 0.008).

0

30

60

90

120

150

0 20 40 60 80 100 120

HA

A9 (

µg

/L)

Time (hours)

0 mg/L

0.2 mg/L

1.0 mg/L

2.0 mg/L

114

Figure 5-7 HAA9 at a reaction time of 32 hours in the presence of copper corrosion solids at pH

8.3 for MWTP water (error bars represent standard deviation of triplicate tests)

Figure 5-8 enables a comparison of HAA formation rates for solutions spiked with

Cu(NO3)2 (dissolved Cu(II) without solids) and those spiked with solid copper corrosion

products. The concentrations of Cu(II) that are shown in Figure 5-8 for these solid corrosion

products were the measured equilibrated concentrations of Cu(II) released from the solids. As

shown in Figure 5-8, HAA formation rates for solutions spiked with the dissolved form of Cu(II)

increased with increases in the Cu(II) concentrations. The addition of copper corrosion solids

significantly increased HAA formation rates relative to those on the curve for solutions with

similar aqueous Cu(II) concentrations. This suggests that surface-associated Cu(II) from these

corrosion products may also participate in the reactions for HAA formation by interacting with

NOM and hence increasing its reactivity with free chlorine. As a result, free chlorine decay and

HAA formation were accelerated. As reported in Section 5.3.1.4, 0.2 g/L Cu2O and 1.0 g/L

malachite accelerated free chlorine degradation relative to 1.0 g/L CuO and 0.2 g/L malachite,

respectively. Therefore, it is not surprising to observe that 0.2 g/L Cu2O and 1.0 g/L malachite

also had faster HAA formation rates than 1.0 g/L CuO and 0.2 g/L malachite, respectively.

The current observation of increased HAA formation rates supports the theory of Cu-

catalyzed formation of HAA by increasing the reactivity of NOM rather than Cu-induced

chlorine decay (which would result in decreased HAA formation). Since free chlorine

degradation and HAA formation in the presence of these copper corrosion products were

reactions mediated by both aqueous and surface-bound Cu(II), further investigation is needed to

0

40

80

120

160

200

No Cu(II)and solids

CuO 1g/L Cu2O0.2g/L

Malachite0.2 g/L

Malachite1.0 g/L

HA

A9 (

µg

/L)

115

evaluate the reactivity of these solids with free chlorine and their potential to enhance HAA

formation.

Figure 5-8 HAA9 formation rates for MWTP water containing dissolved Cu(II) without solids

and solid copper corrosion products

HAA Speciation 5.3.3

Among the nine possible HAA species in MWTP water, the sum of the concentrations of

MCAA, DCAA and TCAA generally contributed to 65-93% of HAA9 independent of copper

concentration for the reaction periods examined. In addition, the concentrations of BCAA and

BDCAA were significantly higher than for other brominated species, especially at low pH.

Therefore, Figure 5-9 shows a comparison of these five more predominant HAA species

(MCAA, DCAA, TCAA, BCAA and BDCAA) after a reaction time of 100 hours under the

influence of pH in the absence and presence of 1.0 mg/L Cu(II). When Cu(II) was absent, TCAA

was the dominant species at low pH but DCAA was predominant at neutral and high pH. When

1.0 mg/L Cu(II) was present, DCAA formation consistently exceeded MCAA and TCAA

formation at all three pH levels. The MCAA concentration tended to increase with increasing pH

both with and without a spike of 1.0 mg/L Cu(II). BCAA concentrations were the same at all

three pH levels in the absence of Cu(II), but exhibited a slight increase with increasing pH in the

Cu(II)-spiked water. Conversely, BDCAA had a higher concentration at pH 6.6 than at pH 8.6 in

the absence of Cu(II), but significantly decreased and exhibited no significant difference in its

concentrations at all three pH levels in the presence of Cu(II).

10

20

30

40

50

0.0 0.2 0.4 0.6 0.8 1.0HA

A f

orm

ation r

ate

[µ

g·L

-1·ln(h

)-1]

Measured dissolved Cu(II) (mg/L)

Cu(II) without solids 0.2 g/L Cu2O1.0 g/L CuO 0.2 g/L malachite1.0 g/L malachite

116

Figure 5-9 Effect of pH on HAA speciation in the absence and presence of 1 mg/L Cu(II) at

reaction time of 100 hours for MWTP water (error bars represent standard deviation of triplicate

tests)

These observations indicate that chlorination of HAA organic precursors was influenced

by both pH and the presence of Cu(II). The observed lack of significant differences in DCAA

formation yields for the three pH values and the prevalence of TCAA at low pH in the absence of

Cu(II) are in agreement with previous studies (Reckhow and Singer, 1985; Reckhow et al., 1990;

Liang and Singer, 2003).

However, enhanced formation of DCAA with increasing pH in the presence of Cu(II)

suggests that Cu(II) would preferentially complex with DCAA precursors and increase their

reactivity with chlorine. In the presence of Cu(II), although elevated reactivity of NOM by Cu(II)

complexation (especially at high pH) led to more chlorine being involved in reactions with

NOM, the increased reactivity of DCAA precursors by Cu(II) dominated the formation of

DCAA. Furthermore, the shift of chlorine species at high pH from the potent halogenating HOCl

to less reactive OCl- should have resulted in a lesser formation of halogenated species (such as

TCAA). The different behavior of DCAA and TCAA formation under different pH conditions

suggests that their formation mechanisms or precursors may be different, especially in the

investigated water source. However, the similar pattern of pH effects for DCAA and BCAA

formation as well as for TCAA and BDCAA formation indicate that each species belonging to

the same subclass of HAA9 (i.e. dihaloacetic acids and trihaloacetic acids) had similar precursors

and reaction pathways with chlorine. All of these findings are also in agreement with other

studies (Reckhow and Singer, 1985; Liang and Singer, 2003). In terms of MCAA, Reckhow et

al. (1990) and Liang and Singer (2003) stated that pH did not impact MCAA formation.

0

20

40

60

80

100

120

pH 6.6 pH 7.6 pH 8.6 pH 6.6 pH 7.6 pH 8.6

HA

A (

µg

/L)

MCAA DCAA TCAA BCAA BDCAA

0 mg/L Cu(II) 1.0 mg/L Cu(II)

117

However, the increased MCAA formation rate at high pH that was observed in the absence and

presence of Cu(II) in this study may be due to different compositions of NOM in the water tested

in their research and in this study.

In the water samples spiked with copper corrosion solids, DCAA formation was still

consistently predominant over TCAA over the reaction period (figures are not shown). Again,

this suggests that DCAA precursors, rather than MCAA and TCAA precursors, preferentially

interacted with the dissolved copper species and/or surface-associated Cu(II) from these copper

solid corrosion products, thereby suppressing MCAA and TCAA formation. The shift of chlorine

species from HOCl to OCl- at high pH of 8.3 also made the formation of highly chlorinated HAA

species less likely (Cowman and Singer, 1996).

5.4 Summary

This is the first study that investigated catalytic potential of copper corrosion products,

including Cu(II), Cu2O, CuO, and Cu2(OH)2CO3, on chlorine decay and HAA formation. The

impacts of pH and the concentrations of these corrosion products on copper catalysis were

evaluated under controlled experimental conditions at bench scale. The following conclusions

can be drawn based on the experimental results:

a. For the three pH levels investigated, accelerated chlorine decay and HAA formation were

observed at pH 8.6 upon the addition of 1.0 mg/L Cu(II). Further investigation of

chlorine decay pathways in the presence of Cu(II) with synthetic water indicated that the

presence of dissolved and/or solid surface-associated Cu(II) would increase the reactivity

of NOM with chlorine. As a result, chlorine decay was accelerated, likely by reacting

with active Cu(II)-NOM complexes, and HAA formation was enhanced.

b. Free chlorine decayed faster and higher concentrations of HAA formed as the Cu(II)

concentration increased from 0 mg/L to 1.0 mg/L, but an absence of further changes with

increases of Cu(II) from 1.0 mg/L to 2.0 mg/L suggests that the capacity of NOM-Cu

interactions had been reached by that point.

c. Solid copper corrosion products were also observed to catalyze both chlorine decay and

HAA formation.

d. The presence of Cu(II) and its solid corrosion products led to DCAA formation

118

consistently predominating over other HAA species.

Results of this study may provide some implications for distributed water quality in

domestic plumbing systems where copper pipes are primarily installed. These pipe surfaces are

often covered by corrosion solids, and in the aqueous phase dissolved copper ions can also be

sorbed onto corrosion solids. From the present study, interactions of free chlorine and HAA

precursors with copper corrosion products will likely affect the stability of secondary

disinfectants and the fate of HAAs, especially in the premise plumbing of distribution systems.

Understanding the catalytic potential of copper on chlorine degradation and HAA formation will

be of benefit to water utilities and households for the management of their distribution systems

and water quality. Since copper concentration, pH and reaction time are the main factors to

impact the nature and extent of copper catalysis in the present study, considerations for corrosion

preventive strategies are suggested to include pH adjustment, addition of corrosion inhibitors and

periodic flushing to decrease contact time of water with pipe corrosion products to alleviate

HAA formation by copper catalysis.

5.5 References

Alvarez-Puebla, R. A., Valenzuela-Calahorro, C., and Garrido, J. J. (2004a) Cu(II) retention on a

humic substance. Journal of Colloid and Interface Science, 270(1), 47-55.

Alvarez-Puebla, R. A., Valenzuela-Calahorro, C., and Garrido, J. J. (2004b) Retention of Co(II),

Ni(II), and Cu(II) on a purified brown humic acid. Modeling and characterization of the

sorption process. Langmuir, 20(9), 3657-3664.



Bansal, V. K., Kumar, R., Prasad, R., Prasad, S., and Niraj. (2008) Catalytic chemical and

electrochemical wet oxidation of phenol using new copper(II) tetraazamacrocycle

complexes under homogeneous conditions. Journal of Molecular Catalysis a-Chemical,

284(1-2), 69-76.



119

Chun, C. L., Hozalski, R. M., and Arnold, T. A. (2005) Degradation of drinking water


Technology, 39(21), 8525-8532.

Cowman, G. A., and Singer, P. C. (1996) Effect of bromide ion on haloacetic acid speciation



Deborde, M., and 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.

Fu, J., Qu, J., Liu, R., Qiang, Z., Zhao, X., and Liu, H. (2009a) Mechanism of Cu(II)-catalyzed

monochloramine decomposition in aqueous solution. Science of the Total Environment,

407(13), 4105-4109.




Gray, E. T., Taylor, R. W., and Margerum, D. W. (1977) Kinetics and mechanisms of copper-

catalyzed decomposition of hypochlorite and hypobromite - properties of a dimeric

copper(III) hydroxide intermediate. Inorganic Chemistry, 16(12), 3047-3055.

Health Canada (1992) Guidelines for Canadian drinking water quality: guidelines technical

document- copper. Water quality and health bureau, Health Environments and Consumer

Safety Branch. Health Canada, Ottawa, Ontario.

Health Canada (2008) Guidelines for Canadian Drinking Water Quality: guideline technical

document — haloacetic acids., Water, Air and Climate Change Bureau, Healthy


Kanokkantapong, V., Marhaba, T. F., Pavasant, P., and Panyapinyophol, B. (2006)

Characterization of haloacetic acid precursors in source water. Journal of Environmental

Management, 80(3), 214-221.

120

Korshin, G. V.; Perry, S. A. L.; Ferguson, J. F. (1996) Influence of NOM on copper corrosion.

Journal American Water Works Association, 88 (7), 36-47.

Lee, J. Y., Pearson, C. R., Hozalski, R. M., and Arnold, W. A. (2008) Degradation of

trichloronitromethane by iron water main corrosion products. Water Research, 42(8-9),

2043-2050.

Li, B., Qu, J. H., Liu, H. J., and Hu, C. Z. (2007) Effects of copper(II) and copper oxides on

THMs formation in copper pipe. Chemosphere, 68, 2153-2160.



152(1), 250-258.



Technology, 37(13), 2920-2928.

Merkel, T. H., Gross, H. J., Werner, W., Dahlke, T., Reicherter, S., Beuchle, G., and Eberle, S.

H. (2002) Copper corrosion by-product release in long-term stagnation experiments.

Water Research, 36(6), 1547-1555.


Sons, New York.

Morris, J. C. (1975) Formation of halogenated organics by chlorination of water supplies, a

review. National Service Center for Environmental Publications (NSCEP), USEPA,

Washington D.C.

Onuchukwu, A. I. (1994) The effect of metal-doped copper ferrite activity on the catalytic

decomposition of hydrogen-peroxide. Materials Chemistry and Physics, 37(2), 129-131.

Paidar, M., Rousar, I., and Bouzek, K. (1999) Electrochemical removal of nitrate ions in waste

solutions after regeneration of ion exchange columns. Journal of Applied

Electrochemistry, 29(5), 611-617.

121

Pintar, A., Bercic, G., and Levec, J. (1997) Catalytic liquid phase oxidation of aqueous phenol

solutions in a trickle-bed reactor. Chemical Engineering Science, 52(21-22), 4143-4153.

Reckhow, D. A., and Singer, P. C. (1985) Mechanisms of organic halide formation during acid

fulvic-acid chlorination and implications with respect to preozonation. Jolley, R. L. et al.

(editors.). Water Chlorination, Vol. 5. Chemistry, Environmental Impact and Health

Effects. Fifth Conference on Water Chlorination: Environmental Impact and Health

Effects. Lewis Publishers, Inc.: Chelsea, Mich., USA. Illus. Maps, 1229-1258.

Reckhow, D. A., Singer, P. C., and Malcolm, R. L. (1990) Chlorination of humic materials by-

product formation and chemical interpretations. Environmental Science & Technology,

24(11), 1655-1664.



Stevenson F.J. (1985) In Humic substances in soil, sediment and water: geochemistry and

characterization. Edited by Aiken G. R., Mc Knight D.M., Wershaw R. L. and

MacCarthy P. Wiley Interscience, NY.

Stevenson, F.J. (1994) Humus Chemistry: Genesis, Composition and Reactions (2nd

edition)

Wiley, New York.

USEPA (1989) 40 CFR Parts 141 and 142 Drinking water; National Primary Drinking Water

Regulations; Filtration, Disinfection; Turbidity, Giardia lamblia, Viruses, Legionelia, and

Heterotrophic Bacteria; Final Rule. Federal Register 54 (124), 27486-27541.


Stage 2 Disinfectants and Disinfection Byproducts Rule. Federal Register 71(2), 387-

493.

Xiao, W. Z., Hong, S. K., Tang, Z. J., Seal, S., and Taylor, J. S. (2007) Effects of blending on

surface characteristics of copper corrosion products in drinking water distribution

systems. Corrosion Science, 49(2), 449-468.

122

Factors Affecting Copper Catalysis of NDMA Formation 6

from DMA in Simulated Premise Plumbing

The results of this chapter have been submitted for publication as follows:

Zhang, H., Andrews, S.A. Factors Affecting Copper Catalysis of NDMA Formation from

DMA in Simulated Premise Plumbing. Water Research.

Results from this chapter focus on the research gap “What are the impacts of metal

corrosion products on the fate of NDMA in distribution systems?”

123

Abstract

This study investigated the effects of copper, a metal commonly employed in household

plumbing systems, on N-nitrosodimethylamine (NDMA) formation from a known NDMA

precursor, dimethylamine (DMA). Copper-catalyzed NDMA formation increased with increasing

copper concentration, DMA concentration, alkalinity and hardness, but decreased with

increasing natural organic matter (NOM) concentration. pH influenced the speciation of

chloramine and the interactions of copper with DMA. The transformation of monochloramine

(NH2Cl) to dichloramine (NHCl2) and complexation of copper with DMA were proposed to be

involved in elevating the formation of NDMA by copper at pH 7.0. The inhibiting effect of

NOM on copper catalysis could be attributed to the rapid consumption of NH2Cl by NOM and/or

the competitive complexation of NOM with copper to limit the formation of DMA-copper

complexes. Hardness ions, as represented by Ca2+

, also competed with copper for binding sites

on NOM, thereby weakening the inhibitory effect of NOM on NDMA formation. Common

copper corrosion products also participated in these reactions but in different ways. Aqueous

copper released from malachite [Cu2CO3(OH)2] was shown to promote NDMA formation while

NDMA formation decreased in the presence of CuO, possibly due to the adsorption of DMA.

Keywords: Copper; NDMA; NOM; premise plumbing

124

6.1 Introduction

With the implementation of the Stage 2 Disinfectant/Disinfection Byproduct Rule

D/DBPR, combined chlorine has received increasing attention as a secondary disinfectant in

distribution systems. The formation of monochloramine (NH2Cl) generally dominates over other

species of combined chlorine at a pH value of approximately 8 and at a Cl2/N molar ratio of less

than 1:1. Dichloramine (NHCl2) is formed at higher Cl2/N ratios (molar ratio >1:1) or at low pH

values (pH <5), and its formation is strongly catalyzed by the presence of acetic acid, phosphate,

carbonate, and silicate species (Schreiber and Mitch, 2007; Valentine et al., 1988). It is well

established that fewer halogenated disinfection byproducts (DBPs) are formed upon

chloramination than chlorination. However, the application of chloramines also causes some

unintended changes in water quality, including increased formation of nitrosamines (Choi et al.,

2002; Najm and Trussell, 2001) and elevated metal corrosion rates (Boyd et al., 2008).

One of the typical reaction pathways for NDMA formation is from the reaction of NH2Cl

and dimethylamine (DMA). Two steps are included in this mechanism: a nucleophilic

substitution of DMA by NH2Cl to form unsymmetrical dimethylhydrazine (UDMH) and the

subsequent oxidation of this UDMH intermediate to NDMA (Choi and Valentine, 2002a; Mitch

et al., 2003). The rate of UDMH formation is slow and increases with an increase in pH (Kim

and Clevenger, 2007), whereas the UDMH oxidation occurs nearly instantaneously but with low

yields (<1%) (Mitch and Sedlak, 2002). In addition, Schreiber and Mitch (2006, 2007) proposed

two other NDMA formation pathways involving NHCl2: a relatively slow reaction of NHCl2

with amine precursors in the presence of dissolved oxygen, and a fast reaction involving a

reactive breakpoint chlorination intermediate (HNO) which is produced from the hydrolysis of

NHCl2. Normally, NHCl2 can be minimized by controlling pH (> 8.5) and Cl2/N molar ratios

(<1) (Schreiber and Mitch, 2006).

Drinking water guidelines for nitrosamines have been under recent development in

Canada and the United States. Health Canada has established a guideline for NDMA in drinking

water at a Maximum Acceptable Concentration (MAC) of 40 ng/L (Health Canada, 2011). The

Ontario Ministry of the Environment, Canada, has set a standard for NDMA of 9 ng/L (MOE,

2002), and the California Department of Public Health has set notification levels of 10 ng/L for

125

each of NDMA, N-nitrosodiethylamine (NDEA) and N-nitrosodi-n-propylamine (NDPA)

(CDPH, 2009).

Copper is commonly used in pipes and copper alloys that are found in household

plumbing. Copper corrosion products in drinking water systems include copper ions [Cu(II)],

cuprite [Cu2O], tenorite [CuO], cupric hydroxide [Cu(OH)2], and malachite [Cu2CO3(OH)2]

(Xiao et al., 2007). The catalytic potential of various copper species in many reaction processes

has been widely investigated (Onuchukwu, 1994; Pintar et al., 1997; Paidar et al., 1999). Some

studies have demonstrated copper catalysis on free chlorine and NH2Cl degradation as well as

THM and HAA formation (Blatchley et al., 2003; Fu et al., 2009b; Fu et al., 2009c; Li et al.,

2007). Although Schreiber and Mitch (2006) reported elevated NDMA formation in the presence

of copper, details concerning factors that impact copper catalysis during NDMA formation are

not available.

The most likely mechanism of copper catalysis during THM and HAA formation has

been speculated to involve interactions of Cu(II) with natural organic matter (NOM) and a

subsequent increased reactivity of NOM with chlorine. NOM is composed of an extremely

diverse group of materials that include carboxyl groups, some phenolic groups, alcohol groups,

methoxyl groups, ketones, and aldehydes (Reckhow et al., 1990). The binding affinity of NOM

with Cu(II) has been reported to depend on the types of functional groups involved, pH, copper

concentration, and the concentrations of Ca2+

and Mg2+

(Brown et al., 1999; Lu and Allen, 2002;

Rate et al., 1993). For example, according to Lu and Allen (2002), hardness ions (Ca and Mg)

had competitive impacts on Cu-NOM complexation, although the binding affinities of Ca and

Mg are much weaker than that of Cu. They also reported that competitive impacts of hardness

ions on Cu complexation increased with increasing pH and hardness concentration, but these

impacts became less significant at successively higher hardness concentrations.

The objective of this study was to investigate the catalytic potential of Cu(II) on NDMA

formation under controlled experimental conditions by performing simulated distribution system

(SDS) tests. Investigated factors that were theorized to affect copper’s impact on NDMA

formation from DMA included copper concentrations, DMA concentrations, alkalinity, NOM

and hardness. Since copper is widely used in domestic plumbing systems, understanding the

impacts of copper and its corrosion products on the formation of NDMA may help utilities and

126

households to minimize the potential for NDMA contamination in their systems by suggesting

preventive strategies.

6.2 Materials and methods

Chemicals and Materials 6.2.1

All chemicals used in this study were ACS grade or higher. High quality Milli-Q water

produced from a Millipore Milli-Q UV plus ultrapure water system (Millipore, Mississauga, ON)

was used for all water blanks and chemical reagent solutions. Dissolved Cu(II) was added in the

form of CuSO4 (1000 mg/L stock solution). Phosphate buffer (H2PO4-/HPO4

2-) and borate buffer

(H3BO3/NaOH) solutions were prepared and added to test solutions at 1 mM concentration to

control reaction solution pH at desired levels. The impacts of different buffers at a concentration

of 1 mM on NDMA formation were negligible according to Schreiber and Mitch (2006). Sulfuric

acid (H2SO4, 50%) and sodium hydroxide (NaOH, 20%) were also used for pH adjustment. A

sodium carbonate (Na2CO3, 300 mM) solution was used to adjust alkalinity. A calcium chloride

(CaCl2, 1000 mM) solution was prepared for hardness tests. Suwannee River natural organic

matter (SR-NOM), purchased from the International Humic Substances Society, was added to

Milli-Q water to adjust NOM concentration. Dimethylamine (DMA, 2.0 M in methanol) was

purchased from Sigma-Aldrich. In all of the tests except for those evaluating effects of DMA

concentrations, DMA at a concentration of 11.2 µg/L (250 nM) was dosed to ensure the

formation of NDMA at measureable levels. L-Ascorbic acid (≥ 99.0%, Sigma Aldrich) at 200

mg/L was applied to quench total chlorine residuals before NDMA extraction.

The free chlorine dosing solution (approximately 3500 mg/L as Cl2) was made by

diluting a concentrated solution of sodium hypochlorite (NaOCl, 6%, VWR) in Milli-Q water.

The resulting solution was standardized by further dilution with Milli-Q water and then

measurement using the DPD colorimetric method with a Hach DR 2700 spectrophotometer. The

total chlorine dosing solution, containing at least 90% monochloramine (NH2Cl), was preformed

by adding the free chlorine dosing solution to a 2.7 g/L (50 mM) ammonium chloride solution in

a well-stirred 250 mL amber bottle at a Cl2/N molar ratio of 0.8:1.

127


All of the experiments were performed in 1 L amber bottles with PTFE-lined caps and at

room temperature (20±1 ºC). The reaction bottles were made chlorine demand free before use.

In the tests to evaluate effects of copper concentration on NDMA formation, copper

sulphate was spiked to provide copper ion concentrations ranging from 0 to 1.0 mg/L to be

compatible with Drinking water guidelines for copper – maximum acceptable concentration 1.0

mg/L (Health Canada, 2009). Malachite [Cu2CO3(OH)2] and tenorite (CuO) were selected as the

solid forms of copper corrosion products of interest according to the results of MINEQL+

version 4.5 (Environmental Research Software, Hallowell, Maine, USA) simulations. They were

added into test solutions in the form of powders and mixed with test water thoroughly using a stir

bar. After 24 hours, water samples dosed with corrosion products were filtered through 0.2 µm

Nylaflo® Nylon membrane filter paper (Pall Corporation) to separate the solid copper corrosion

products from the aqueous solution. Dissolved copper in the filtrate was measured by flame

atomic absorption spectrometry (FAAS). To account for any influence from filter paper and

filtration performance on NDMA concentrations, control samples without the addition of solid

copper corrosion products were also filtered under similar conditions.

To evaluate the impacts of NOM on copper solubility and to determine if DMA-Cu

complexes played a role in copper catalysis, copper sulphate was spiked in the presence of

different concentrations of NOM and/or DMA. The resulting solutions were equilibrated for 6

hours and then filtered through 0.2 µm Nylaflo® Nylon membrane filter paper, and copper

concentrations in the filtrate were measured by FAAS.

All experiments were performed at pH 7.0±0.1 (buffered by 1 mM H2PO4-/HPO4

2-

solution) in Milli-Q water except when evaluating pH effects, and the copper level was

maintained at 1 mg/L except in the tests where effects of copper concentrations were evaluated.

An initial NH2Cl concentration of 2.3±0.1 mg/L was applied for all of the tests, chosen to be

typical of that which is observed in drinking water entering distribution systems.

The concentration and speciation of chloramine in the absence and presence of copper

were determined according to Schreiber and Mitch (2005, 2006). A CE 3055 Reflectance

Spectrophotometer (Cecil Instruments, England) was used. To ensure the measurability of

128

NH2Cl and dichloramine (NHCl2) via spectrophotometric analysis, a preformed NH2Cl stock

solution at a concentration of 130 mg/L was tested. The solution pH was controlled at pH 6.7 and

7.0, respectively. After 20 hours reaction, NH2Cl and NHCl2 were distinguished and quantified

by monitoring absorbance at their respective λmax (λmax, NH2Cl= 2 , λmax, NHCl2= 295) and solving

simultaneous equations.

NDMA extract preparation and measurement were carried out according to Taguchi et al.

(1994). Isotope-labeled surrogate (d6-NDMA) was used to correct for interferences from

nitrosamine extraction processes and matrix effects. The instrument used for nitrosamine

analysis was a Varian 4000 GC/MS operated in chemical ionization mode.

6.3 Results and discussion

Since alkalinity, NOM, hardness and pH may affect copper speciation and solubility

(Broo et al., 1998; Korshin et al., 1996), the effects of these water quality parameters on NDMA

formation from DMA were all examined in this study. The impacts of DMA concentrations on

NDMA formation and copper solubility as well as its major complexes/precipitates were also

tested.

Effect of Copper Concentrations 6.3.1

Figure 6-1 illustrates the observed NDMA formation when varying the copper sulphate

concentrations (expressed as mg/L total Cu). NDMA formation increased approximately linearly

(31 to 104 ng/L) when spiked Cu(II) concentrations were increased from 0 mg/L to 1.0 mg/L,

suggesting some participation or possible catalytic effect of copper during NDMA formation

from DMA.

Separate NDMA formation kinetics tests also showed that less than half the amount of

NDMA would form in the absence of copper than in the presence of 1.0 mg/L copper (Figure

6-2), providing further support to the theory that copper may enhance or catalyze NDMA

formation.

129

Figure 6-1 NDMA formation with increasing copper concentrations from added CuSO4; pH 7.0,

initial NH2Cl 2.3±0.1 mg/L, 11.2 µg/L DMA spiked, 24 hours, error bars indicate the measured


Figure 6-2 NDMA formation kinetics in the absence and presence of copper; pH 7, Cu(II) 1

mg/L spiked, initial NH2Cl 2.3±0.1 mg/L, 11.2 µg/L DMA, error bars indicate the measured


Effects of DMA Concentration and Cu-NH2Cl Interactions 6.3.2

Figure 6-3 shows a comparison of the impacts of DMA concentrations (varying from 0 to

1800 µg/L) on NDMA formation in the absence and presence of 1.0 mg/L spiked Cu(II).

Although this range of concentrations extends well above the range of any NDMA precursors

that would be expected in typical drinking water sources, the test shows the extent of the effects

of copper presence, even at a relatively low concentration of 1.0 mg/L Cu(II). In the absence of

copper, NDMA formation increased with increasing DMA concentrations, and the yield of

0

30

60

90

120

0 5 10 15 20 25

ND

MA

(ng/L

)

Time (hours)

1 mg/L Cu(II)

0 mg/L Cu(II)

0

30

60

90

120

0.0 0.1 0.2 0.3 0.4 0.5 0.7 1.0

ND

MA

(ng/L

)

Copper from CuSO4 (mg/L)

130

NDMA at 24 hours was 2670 ng/L at a DMA concentration of 1800 µg/L (approximately 0.15

±0.01% yield, which is typical for this precursor). In the presence of copper, NDMA formation

was enhanced by 2 to 4 times compared with that in the absence of copper. These data further

confirmed the catalytic impacts of copper on NDMA formation that were shown in Figures 6-1

and 6-2. The inset plot in Figure 6-3 shows NDMA yields at DMA concentrations ranging from

0 to 11.2 µg/L. Comparable NDMA yields were obtained for the samples that had similar

conditions to those of Figure 6-1 (e.g. 11.2 µg/L DMA and 1 mg/L copper produced 107 ng/L vs.

104 ng/L NDMA), providing further evidence of the excellent reproducibility and reliability of

the results in this study.

Figure 6-3 NDMA formation with DMA concentrations; pH 7.0, initial NH2Cl 2.3±0.1 mg/L,

Cu(II) 1 mg/L spiked, 24 hours, d error bars indicate the measured maximum and minimum

values (n=2)

In addition, as shown in Figure 6-3, the yields of NDMA in the absence of copper were

approaching a maximum of approximately 2600 ng/L at the molar ratio of DMA (1800 µg/L, 40

µM) to NH2Cl (2.4 mg/L, 34 µM) of approximately 1.2:1, which is close to the 1:1 molar ratio

for the maximum NDMA formation observed by Choi and Valentine (2002b). As they proposed,

a further increase in the DMA:NH2Cl ratio would make the amount of NH2Cl available to

oxidize the UDMH intermediate to be rapidly depleted due to chlorine transfer. In the current

research and in the presence of copper, a maximum NDMA yield occurred at a DMA

concentration of approximately 900 µg/L (20 µM) with the DMA:NH2Cl ratio of only 0.6. This

shift of the maximum DMA:NH2Cl ratio in the presence of copper is proposed to be associated

with copper’s interactions with NH2Cl, given that copper has been shown to enhance the

decomposition of NH2Cl to NHCl2 under acidic conditions (Fu et al., 2009a; Fu et al., 2009c)

0

2000

4000

6000

8000

0 400 800 1200 1600 2000

ND

MA

(n

g/L

)

DMA (µg/L)

0 mg/L Cu(II)

1 mg/L Cu(II)

0

50

100

150

0 5 10 15

107 ng/L

131

and NHCl2 has been previously reported to increase NDMA formation (Schreiber and Mitch,

2006 and 2007).

Separate tests utilizing the spectrophotometric method described by Schreiber and Mitch

(2005, 2006) were performed to confirm the proposed shift in the relative amounts of NH2Cl and

NHCl2 by determining the concentration and speciation of chloramine in the absence and

presence of copper. As shown in Figure 6-4, the addition of 1 mg/L copper at pH 7.0

significantly increased NHCl2 formation (by 2 fold) while the NH2Cl concentration was

decreased by 45%. The effect was even further enhanced at a slightly lower pH (pH 6.7), and

will be discussed again in Section 6.3.7. Thus, the catalyzed formation of NHCl2 by copper at pH

7.0 could be one reason for the enhanced NDMA formation observed in Figure 6-3. Furthermore,

the decomposition of NH2Cl by copper limited the amount of NH2Cl which could participate in

the reactions for oxidizing the UDMH intermediate to NDMA, thereby shifting the concentration

of DMA for the maximum formation of NDMA from 1800 µg/L in the absence of copper to 900

µg/L in the presence of copper (Figure 6-3).

Figure 6-4 Chloramine speciation in the absence and presence of copper at pH 6.7 and 7, Milli-Q

water

Cu-DMA Complexation 6.3.3

In addition to the influence of increased NHCl2 on enhancing the formation of NDMA,

the formation of DMA-Cu complexes was also hypothesized to be able to increase the

conversion of DMA to NDMA. The ability of Cu(II) to complex with organic matter has been

0

30

60

90

120

0

20

40

60

80

100

120

6.7 7.0 6.7 7.0

Su

m o

f N

H2C

l a

nd

N

HC

l 2(m

g/L

as C

l 2)

Chlo

rin

e (

mg

/L a

s C

l 2)

pH

NH₂Cl NHCl₂ Sum of NHCl₂ and NH₂Cl

Without copper With copper

132

reported to be essential for copper to act as a catalyst (Bansal et al., 2008). Aside from

carboxylic and phenolic groups, copper has the potential to complex with the lone pair of

electrons on the central nitrogen of DMA (Croue et al., 2003; Tuschall and Brezonik, 1980;

Westerhoff and Mash, 2002). The complexation of copper with organic nitrogen is moderately

stable, with conditional stability constants ranging from 1.6 106 to 1.3 10

7 (Ko and Lee, 2010;

Tuschall and Brezonik, 1980). Therefore, to detect the occurrence of DMA-Cu complexation,

five concentrations of DMA (0, 45, 360, 702 and 3510 µg/L) were dosed in Milli-Q water

containing 1.0 mg/L copper at pH 7.0, and the amount of copper remaining in dissolved form

was measured. If the complexation between the nitrogen of DMA and copper occurred, the

solubility of copper was expected to increase. As shown in Figure 6-5, the filtered copper

concentration was only 0.3 mg/L in the absence of DMA (0 µg/L), and increased to 0.4 and 0.8

mg/L Cu at DMA concentrations of 45 and 360 µg/L, respectively. Further additions of DMA to

702 and 3510 µg/L did not change the filtered copper concentration significantly. Regardless, the

increased solubility of copper for solutions with DMA at concentrations below 360 µg/L strongly

suggests the complexation of the nitrogen in DMA with copper. This complexation can

essentially increase the reactivity of DMA, thereby enhancing NDMA formation.

Figure 6-5 Dissolved copper concentrations with varying DMA; filtered by 0.2 µm Nylaflo®

Nylon membrane filter paper, pH 7, Cu(II) 1 mg/L spiked, 6 hours, error bars indicate the


Effect of Alkalinity 6.3.4

The effect of alkalinity was examined in this study because copper can form carbonate-

containing species which may play a role in copper’s participation in NDMA-forming reactions.

0.0

0.2

0.4

0.6

0.8

1.0

0 45 360 702 3,510

Dis

solv

ed c

opper

(mg/L

)

DMA (µg/L)

133

Figure 6-6 illustrates the yields of NDMA at different alkalinity levels along with the likely

major copper species determined by MINEQL+ version 4.5. For alkalinity values below 50 mg/L

as CaCO3, NDMA concentrations remained approximately constant at 94~100 ng/L. For

alkalinity values greater than 50 mg/L as CaCO3, NDMA formation increased approximately

linearly from 98 to 225 ng/L (R2

=0.9), suggesting a possible promoting effect of alkalinity on

copper catalyzed NDMA formation.

Figure 6-6 NDMA formation with increasing alkalinity and copper speciation as a function of

alkalinity (determined by MINEQL+ version 4.5); pH 7.0, initial NH2Cl 2.3±0.1 mg/L, Cu(II) 1

mg/L spiked, 11.2 µg/L DMA, 24 hours, duplicate

This apparent promotional effect of alkalinity may have been more related to the copper

species present than to the alkalinity itself. According to MINEQL+ simulations, at pH 7.0 the

spiked copper was present mainly in the form of precipitates (> 85%). For alkalinity values

below 50 mg/L as CaCO3, which generally would be much lower than observed in distribution

systems, Cu3(PO4)2 was the primary form of solids due to the addition of phosphate buffer; for

alkalinity values above 50 mg/L as CaCO3, malachite [Cu2(OH)2CO3] was the dominant solid.

The rest of the copper was present in the dissolved form as free Cu(II) and soluble complexes,

and the concentration of these dissolved components increased with increasing alkalinity.

When Cu3(PO4)2 was present at concentrations of between 0.92 and 0.97 mg/L as copper,

again, only at very low alkalinities and so not likely to be seen in distribution systems, NDMA

yields remained constant, indicating that such low alkalinity values and/or Cu3(PO4)2 had no

significant impacts on NDMA formation from DMA. In the presence of malachite, given that

0

50

100

150

200

250

0.0

0.2

0.4

0.6

0.8

1.0

0 20 35 50 100 150 200 300

ND

MA

(n

g/L

)

Cop

pe

r sp

ecia

tio

n (

mg

/L)

Alkalinity (mg/L as CaCO3)

Aqueous copper PrecipitationNDMA

Cu3(PO

4)2 Cu

2(OH)

2CO

3

134

NDMA formation increased with the increase in the corresponding aqueous copper

concentrations, it follows that the copper catalysis of NDMA formation was essentially a

reaction mediated by aqueous copper released from the malachite. Separate experiments were

performed to test NDMA formation and copper release in the presence of pre-dosed malachite.

The yields of NDMA were increased from 24 ng/L in the absence of malachite to 144 and 171

ng/L after dosing 0.1 and 1.0 g/L malachite, respectively, whereas the released copper

concentrations in the presence of 0.1 and 1.0 g/L malachite were 0.2 and 0.6 mg/L, respectively.

These results support the theory that aqueous copper released from malachite could catalyze

NDMA formation.

Effect of NOM 6.3.5

As discussed in Section 6.3.3, copper can complex with DMA and thus increase DMA

reactivity to form NDMA (Figure 6-5). In all natural waters, natural organic matter (NOM) is

ubiquitous and it also has a strong affinity for metal ions (Stevenson, 1994). Therefore, to

demonstrate the competitive impacts of NOM on the interactions between DMA and copper and

the subsequent NDMA formation, a range of concentrations of Suwannee River NOM (SR-

NOM) were dosed into DMA-spiked water in the presence of 1 mg/L copper. NDMA formation

results are shown in Figure 6-7, along with corresponding NH2Cl residual and dissolved copper

concentrations after 24-h reaction time. In the absence of SR-NOM, 111 ng/L NDMA formed

which is compatible to the yields obtained in previous experiments (104 and 107 ng/L). NDMA

Figure 6-7 NDMA, NH2Cl residual and dissolved copper with increasing SR-NOM

concentrations; pH 7.0, initial NH2Cl 2.3±0.1 mg/L, Cu(II) 1 mg/L spiked, 11.2 µg/L DMA, 24

hours, error bars indicate the measured maximum and minimum values (n=2)

0

30

60

90

120

0.0 1.0 2.0 3.0 4.0 5.0

ND

MA

(n

g/L

)

SR-NOM (as mg/L TOC)

0.0

0.2

0.4

0.6

0.8

1.0

0.0

0.5

1.0

1.5

2.0

2.5

0.0 1.0 2.0 3.0 4.0 5.0

SR-NOM (mg/L as TOC)

Dis

so

lve

d C

u(I

I) (

mg

/L)

NH

2C

l re

sid

ua

l (m

g/L

)

NH₂Cl residual

Dissolved Cu(II)

135

concentrations decreased sharply for SR-NOM concentrations up to 1.0 mg/L as TOC, and then

more gradually decreased to 26 ng/L with further increases in SR-NOM concentrations to 4.1

mg/L as TOC. Figure 6-7 also shows that increases in SR-NOM concentrations corresponded to

NH2Cl residual decreases, whereas the dissolved copper concentrations increased.

There are two possible explanations for the observed decreases in NDMA formation.

First, SR-NOM likely formed strong complexes with copper, as indicated by the corresponding

increased dissolved copper concentrations with increasing SR-NOM concentrations. This

complexation, therefore, would likely compete with the interactions of copper with DMA and

subsequently decrease the catalytic effects of copper on NDMA formation. In addition, SR-NOM

also competes with DMA for NH2Cl, as indicated by the decreased NH2Cl residual concentration

upon increased SR-NOM concentrations, and NH2Cl is a rate-limiting factor for NDMA

formation (Chen and Valentine, 2006). For both of these reasons, it was reasonable to have

observed decreased NDMA formation after the addition of SR-NOM.

Effect of Hardness 6.3.6

The impacts of hardness ions on copper catalysis during NDMA formation have also

been investigated in this study given that these ions may compete with copper for the available

binding sites on NOM or DMA. Due to the abundance of hardness ions in most natural waters,

the impacts of these ions on copper catalysis may be an important aspect to consider when

determining NOM (and/or DMA)-Cu complexation. Since calcium is the dominant hardness-

related ion in most natural waters, calcium (Ca2+

) was chosen to be representative of the hardness

ions in this study.

Figure 6-8 displays the NDMA yields and corresponding dissolved copper concentrations

obtained at hardness concentrations varying from 0 to 200 mg/L as CaCO3 in the presence of 4.1

mg/L SR-NOM as TOC. NDMA formation increased with increasing hardness while the

dissolved copper concentrations exhibited the opposite trend. This suggests that Ca2+

can

compete with copper for the available binding sites on SR-NOM, thereby decreasing the amount

of copper complexed with NOM (NOM-Cu complexation was discussed in Section 6.3.5). It

follows that copper-catalyzed NDMA formation would be increased in hard waters due to the

competitive effects of Ca2+

on NOM-Cu complexation, freeing up copper ions to participate in

NDMA formation reactions.

136

Figure 6-8 Variation of NDMA formation and dissolved copper concentrations with increasing

hardness; pH 7.0, Cu(II) 1 mg/L spiked, initial NH2Cl 2.3±0.1 mg/L, 11.2 µg/L DMA spiked,

SR-NOM 4.1 mg/L as TOC, 24 hours, error bars indicate the measured maximum and minimum

values (n=2)

Effect of pH and the Role of NHCl2 in NDMA Formation 6.3.7

This study also investigated the impact of pH on copper catalysis during NDMA

formation since pH can affect the speciation of chloramine and copper as well as the interactions

of copper with DMA and organic matter. Figure 6-9 summarizes the observed NDMA yields and

corresponding pseudo-first-order NH2Cl decay constants for tests performed in the absence and

presence of 1.0 mg/L Cu(II) at pH levels that span a range found in distribution systems. In the

absence of added Cu(II), NDMA formation from DMA increased with increasing pH, which is in

agreement with previous findings (Kim and Clevenger, 2007). When 1 mg/L Cu(II) was present,

NDMA formation was enhanced at pH 6.7, 7.0 and 8.0 by factors of 8.0, 2.1 and 2.3,

respectively. At pH 9.0, however, NDMA formation was only approximately 80% of that formed

in the absence of Cu(II).

The observed increased NDMA formation under slightly acidic or neutral conditions (pH

6.7 and 7.0), as shown in Figure 6-9, has been proposed to be due to a greater formation of

NHCl2, as discussed in Section 6.3.2. As estimated by NH2Cl pseudo-first-order decay constants

at four pH levels (shown in Figure 6-9), regardless of the presence of Cu(II), NH2Cl significantly

decayed at both pH 6.7 and pH 7.0, but remained approximately constant over the test period of

48 hours at pH 8.0 and pH 9.0. It is known that NH2Cl is more stable at pH 8.0 and that it

0.0

0.2

0.4

0.6

0.8

1.0

0

10

20

30

40

50

0 50 100 150 200

Dis

so

lve

d C

u(I

I) (

mg

/L)

ND

MA

(n

g/L

)

Hardness (mg/L as CaCO3)

NDMA Dissolved Cu(II)

137

decomposes to NHCl2 under even slightly acidic conditions (Valentine et al., 1988). In this

study, separate tests using the spectrophotometric method described in Section 6.2.2 confirmed

an enhancement in the formation of NHCl2 by copper under acidic and neutral conditions (Figure

6-4). Therefore, the catalyzed formation of NHCl2 by Cu(II) could be one reason for the

observed enhanced NDMA formation in the presence of Cu(II), especially at pH 6.7.

Figure 6-9 NDMA formation at four pH levels in the absence and presence of 1 mg/L Cu(II) in

Milli-Q water; 24 hour, NH2Cl 2.3±0.1 mg/L, 11.2 µg/L DMA, error bars indicate the measured


Since an enhanced conversion of NH2Cl to NHCl2 by Cu(II) does not occur in basic

solutions, as evidenced in this study and by Fu et al. (2009a, 2009c), the increased NDMA

formation observed at pH 8.0 in the presence of Cu(II) (Figure 6-9) was likely attributed to other

factors. For example, the complexation of copper with the central nitrogen of DMA was

discussed in Section 6.3.3, and Ko and Lee (2010) have reported the increased potential for

nitrogen to chelate Cu(II) under alkaline conditions relative to acidic conditions. Therefore,

increasing the pH from 7.0 to 8.0 was expected to form more DMA-Cu complexes, thereby

increasing the reactivity of DMA with NH2Cl and subsequent NDMA formation. The observed

NDMA formation results support this theory.

However, the results of separate filtration tests (similar to those described in Section

6.3.3) with five concentrations of DMA at pH 8.0 showed that total dissolved copper

concentrations were only increased from 0.02 mg/L in the absence of DMA to 0.05 mg/L at a

DMA concentration of 702 µg/L. This is in contrast to the results at pH 7.0 for which at least 0.8

0.000

0.003

0.006

0.009

0.012

0

100

200

300

400

500

600

6.7 7.0 8.0 9.0

Pseudo 1

st ord

er

decay c

onsta

nt (h

-1)

ND

MA

(n

g/L

)

pH levels

NDMA 0 mg/L Cu(II) NDMA 1 mg/L Cu(II)

NH₂Cl 0 mg/L Cu(II) NH₂Cl 1 mg/L Cu(II)

604 ng/L

138

mg/L copper was dissolved at the same concentration of DMA (Figure 6-5). This insignificant

formation of DMA-Cu complexes at pH 8.0 suggests that the increased formation of NDMA at

pH 8.0 is more likely a solid surface mediated reaction, and that the formation of DMA-Cu

complexes did not contribute significantly to NDMA formation at this slightly basic pH.

At pH 9.0, the yield of NDMA in the presence of Cu(II) decreased by 23% compared

with that in the absence of Cu(II). The filtration tests described in Section 6.3.3 were also

performed at pH 9.0 to identify the likely formation of DMA-Cu complexes, and the filtered

copper concentrations before and after the addition of DMA remained below Method Detection

Limit (MDL) of FAAS (0.009 mg/L). This indicates that the majority of the spiked copper was

present in the solid form (CuO, determined by MINEQL+) so DMA could not have formed

significant amount of soluble complexes with copper at pH 9.0. Since DMA has a pKa of 10.7

and NDMA is a neutral compound, it is likely that either DMA or NDMA exhibited some

adsorption to CuO at pH 9.0. In a confirmatory experiment in which NDMA was formed from

DMA in contact with a suspension of CuO at concentrations of 0.2 and 1.0 g/L, NDMA yields

decreased with increasing amounts of CuO. Further investigation is needed to describe the

adsorption behavior of DMA and/or NDMA onto CuO.

6.4 Summary

This is the first study that demonstrated the catalytic potential of copper to enhance

NDMA formation from DMA. The impacts of copper concentrations, DMA concentrations,

alkalinity, SR-NOM, hardness, and pH on copper catalysis were evaluated under controlled

experimental conditions at bench scale. The following conclusions can be drawn based on the

experimental results:

a. Complexation of copper with DMA and the subsequent increased reactivity of DMA with

NH2Cl were suggested to be essential for copper to catalyze NDMA formation.

b. NDMA formation was inhibited by the presence of SR-NOM likely because SR-NOM

may compete with DMA for NH2Cl and/or may strongly complex with copper, thereby

limiting the interactions of copper with DMA.

c. The addition of hardness ions (Ca2+

) weakened the complexation of SR-NOM with

copper by competing with copper for binding sites on SR-NOM, explaining the increased

139

formation of NDMA with increasing hardness.

d. pH had a complex impact on copper catalysis during NDMA formation, influencing the

speciation of chloramine and the interactions of copper with DMA.

e. Aqueous copper released from malachite [Cu2CO3(OH)2] was shown to promote NDMA

formation while NDMA formation decreased in the presence of CuO, possibly due to the

adsorption of DMA

This study provides information that should be considered in relation to the formation

and fate of NDMA in premise plumbing where copper pipes are primarily installed. This

improved understanding about the factors affecting NDMA formation in copper pipes will be

useful in developing strategies to control copper corrosion and reduce NDMA formation. For

example, given the demonstrated impacts of dissolved copper concentrations and retention time

on NDMA formation, the addition of phosphate-based corrosion inhibitors to control copper

corrosion are suggested. Flushing household lines before using water has been recommended to

minimize exposure to dissolved copper. The present research suggests that periodic flushing to

decrease contact time of the water with the pipe material will also help minimize exposure to

NDMA. However, further investigation is needed to evaluate the impacts of real water matrices,

including the complicated impacts of pH, alkalinity, hardness and NOM combinations with

different water compositions on the speciation and solubility of copper and the subsequent

copper catalysis during NDMA formation.

6.5 References

Bansal, V. K., Kumar, R., Prasad, R., Prasad, S., and Niraj (2008) Catalytic chemical and

electrochemical wet oxidation of phenol using new copper(II) tetraazamacrocycle

complexes under homogeneous conditions. Journal of Molecular Catalysis a-Chemical,

284(1-2), 69-76.






140

Broo, A. E., Berghult, B., and Hedberg, T. (1998) Copper corrosion in water distribution systems

- The influence of natural organic matter (NOM) on the solubility of copper corrosion

products. Corrosion Science, 40(9), 1479-1489.

Brown, G. K., Cabaniss, S. E., MacCarthy, P., and Leenheer, J. A. (1999) Cu(II) binding by a

pH-fractionated fulvic acid. Analytica Chimica Acta, 402(1-2), 183-193.

California Department of Public Health (CDPH) (2009) California Drinking Water: NDMA and

Other Nitrosamines - Drinking Water Issues.,

http://www.cdph.ca.gov/certlic/drinkingwater/Pages/default.aspx, ed.

Chen, Z., and Valentine, R. L. (2006) Modeling the formation of N-nitrosodimethylamine

(NDMA) from the reaction of natural organic matter (NOM) with monochloramine.





Choi, J., and Valentine, R. L. (2002a) A kinetic model of N-nitrosodimethylamine (NDMA)

formation during water chlorination/chloramination. Water Science and Technology,

46(3), 65-71.

Choi, J. H., and Valentine, R. L. (2002b) Formation of N-nitrosodimethylamine (NDMA) from

reaction of monochloramine: a new disinfection by-product. Water Research, 36(4), 817-

824.

Croue, J. P., Benedetti, M. F., Violleau, D., and Leenheer, J. A. (2003) Characterization and

copper binding of humic and nonhumic organic matter isolated from the South Platte

River: Evidence for the presence of nitrogenous binding site. Environmental Science &

Technology, 37(2), 328-336.

Fu, J., Qu, J., Liu, R., Qiang, Z., Zhao, X., and Liu, H. (2009a) Mechanism of Cu(II)-catalyzed

monochloramine decomposition in aqueous solution. Science of the Total Environment,

407(13), 4105-4109.

141




Fu, J., Qu, J. H., Liu, R. P., Zhao, X., and Qiang, Z. M. (2009c) The influence of Cu(II) on the

decay of monochloramine. Chemosphere, 74(2), 181-186.

Health Canada (2009) Guidelines on Corrosion Control in Drinking Water Distribution Systems.

Federal-Provincial-Territorial Committee on Drinking Water, available from

www.healthcanada.gc.ca.


Document- N-Nitrosodimethylamine (NDMA) (Catalogue No H128-1/11-662E). Water,

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

Health Canada, Ottawa, Ontario.

Kim, J., and Clevenger, T. E. (2007) Prediction of N-nitrosodimethylamine (NDMA) formation

as a disinfection by-product. Journal of Hazardous Materials, 145(1-2), 270-276.

Ko, C. K., and Lee, W. G. (2010) Effects of pH variation in aqueous solutions on dissolution of

copper oxide. Surface and Interface Analysis, 42(6-7), 1128-1130.

Korshin, G. V., Perry, S. A. L., and Ferguson, J. F. (1996) Influence of NOM on copper

corrosion. Journal American Water Works Association, 88(7), 36-47.



Lu, Y. F., and Allen, H. E. (2002) Characterization of copper complexation with natural

dissolved organic matter (DOM) - link to acidic moieties of DOM and competition by Ca

and Mg. Water Research, 36(20), 5083-5101.

Ministry of the Environment (MOE) (2002) Safe Drinking Water Act 2002. Ontario Regulation

169/03, Schedule 2. http://www.e-

laws.gov.on.ca/html/regs/english/elaws_regs_030169_e.htm.

142

Mitch, W. A., and Sedlak, D. L. (2002) Formation of N-nitrosodimethylamine (NDMA) from


595.

Mitch, W. A., Sharp, J. O., Trussell, R. R., Valentine, R. L., Alvarez-Cohen, L., and Sedlak, D.

L. (2003) N-nitrosodimethylamine (NDMA) as a drinking water contaminant: A review.

Environmental Engineering Science, 20(5), 389-404.



Onuchukwu, A. I. (1994) The Effect of Metal-Doped Copper Ferrite Activity on the Catalytic

Decomposition of Hydrogen-Peroxide. Materials Chemistry and Physics, 37(2), 129-131.

Paidar, M., Rousar, I., and Bouzek, K. (1999) Electrochemical removal of nitrate ions in waste

solutions after regeneration of ion exchange columns. Journal of Applied

Electrochemistry, 29(5), 611-617.

Pintar, A., Bercic, G., and Levec, J. (1997) Catalytic liquid phase oxidation of aqueous phenol

solutions in a trickle-bed reactor. Chemical Engineering Science, 52(21-22), 4143-4153.

Rate, A. W., McLaren, R. G., and Swift, R. S. (1993) Response of Copper (Ii) Humic-Acid

Dissociation Kinetics to Factors Influencing Complex Stability and Macromolecular

Conformation. Environmental Science & Technology, 27(7), 1408-1414.

Reckhow, D. A., Singer, P. C., and Malcolm, R. L. (1990) Chlorination of Humic Materials - by-

Product Formation and Chemical Interpretations. Environmental Science & Technology,

24(11), 1655-1664.

Schreiber, I. M., and Mitch, W. A. (2005) Influence of the order of reagent addition on NDMA

formation during chloramination. Environmental Science & Technology, 39(10), 3811-

3818.

Schreiber, I. M., and Mitch, W. A. (2006) Nitrosamine formation pathway revisited: The

importance of chloramine speciation and dissolved oxygen. Environmental Science &

Technology, 40(19), 6007-6014.

143

Schreiber, I. M., and Mitch, W. A. (2007) Enhanced nitrogenous disinfection byproduct

formation near the breakpoint: Implications for nitrification control. Environmental


Stevenson, F.J. (1994) Humus Chemistry: Genesis, Composition and Reactions (2nd

edition)

Wiley, New York.

Taguchi, V., Jenkins, S. D. W., Wang, D. T., Palmentier, J., and Reiner, E. J. (1994)

Determination of N-Nitrosodimethylamine by isotope-dilution, high-resolution mass-

spectrometry. Canadian Journal of Applied Spectroscopy, 39(3), 87-93.

Tuschall, J. R., and Brezonik, P. L. (1980) Characterization of Organic Nitrogen in Natural-

Waters - Its Molecular-Size, Protein-Content, and Interactions with Heavy-Metals.

Limnology and Oceanography, 25(3), 495-504.

Valentine, R. L., Jafvert, C. T., and Leung, S. W. (1988) Evaluation of a chloramine

decomposition model incorporating general acid catalysis. Water Research, 22(9), 1147-

1153.

Westerhoff, P., and Mash, H. (2002) Dissolved organic nitrogen in drinking water supplies: a


Xiao, W. Z., Hong, S. K., Tang, Z. J., Seal, S., and Taylor, J. S. (2007) Effects of blending on

surface characteristics of copper corrosion products in drinking water distribution

systems. Corrosion Science, 49(2), 449-468.

144

Effects of Pipe Materials, Orthophosphate, and Flow 7

Conditions on Chloramine Decay and NDMA Formation in

Modified Pipe Loops

The results of this chapter have been submitted for publication as follows:

Zhang, H., Andrews, S.A. Effects of Pipe Materials, Orthophosphate, and Flow Conditions

on Chloramine Decay and NDMA Formation in Modified Pipe Loops. Journal of Water

Supply: Research and Technology – Aqua.

Results from this chapter focus on the research gap “How do flow conditions potentially

affect secondary disinfectant stability and NDMA formation?”

145

Abstract

Secondary disinfectants experience temporal and spatial degradation in distribution

systems, and the application of combined chlorine (mainly monochloramine) can potentially

increase the formation of nitrosamine disinfection byproducts. To date, the literature concerning

the fate of nitrosamines in full-scale water mains and household plumbing is limited and

inconsistent. This study examined the effects of pipe materials (iron, copper, lead and PVC) and

orthophosphate on chloramine decay and NDMA formation under stagnant, laminar and

turbulent flow conditions using modified pipe loops. Generally, turbulent conditions increased

the released metal concentrations compared with laminar conditions regardless of

orthophosphate addition. Orthophosphate did not statistically affect iron corrosion rates, but

effectively reduced released copper concentrations under both laminar and turbulent conditions.

Chloramine presence was a rate-limiting factor for NDMA formation, and its decay rate

generally increased with increasing flow velocity. Orthophosphate increased chloramine decay in

contact with iron by increasing nitrite formation, but decreased chloramine decay in contact with

copper and lead by reducing metal corrosion. Copper consistently catalyzed NDMA formation

from DMA under laminar flow conditions. Iron also catalyzed NDMA formation but only under

turbulent conditions. Orthophosphate increased the catalytic effects of iron by modifying the

properties of the associated suspended particles, but decreased copper catalysis by reducing the

dissolved copper concentrations.

Keywords: Chloramine; NDMA; Iron; Copper; Lead; Modified pipe loop

146

7.1 Introduction

The application of a secondary disinfectant in distribution systems is to maintain the

microbial stability of the treated water. However, secondary disinfectants may experience

temporal and spatial degradation in distribution systems due to chemical and biological reactions

that occur in the bulk water and on the pipe wall. Therefore, the overall disinfectant decay within

a pipe is usually postulated as a pseudo-first-order reaction consisting of parallel reactions

occurring in the bulk flow and on the pipe wall (Rossman et al., 1994). Generally, both bulk

water and pipe wall disinfectant decay rates increase with increasing flow velocity as a result of

an increased mass transfer rate of disinfectant residual to the pipe surface, and/or an increased

release rate of corrosion products from the pipe surface (Digiano and Zhang, 2005; Hallam et al.,

2002; Mutoti et al., 2007; Westbrook and Digiano, 2009).

Recently, combined chlorine has received growing attention by water utilities that are

challenged to comply with the Stage 2 Disinfectant/Disinfection Byproducts Rule D/DBPR. The

advantages of applying combined chlorine as an alternative to free chlorine include a minimized

production of trihalomethances (THMs) and haloacetic acids (HAAs), and a relatively stable

residual in distribution systems. However, combined chlorine (mainly monochloramine) can

potentially increase the formation of nitrosamines (Choi et al., 2002; Najm and Trussell, 2001),

of which N-nitrosodimethylamine (NDMA) is the major species of interest. The occurrence of

nitrification and elevated lead release have also been observed in chloraminated distribution

systems (Vasquez et al., 2006; Wilczak et al., 1996).

Considerable effort has been made to investigate factors affecting disinfection byproduct

(DBP) formation in distribution systems. These factors include the concentration and chemical

properties of the DBP precursors, water temperature, pH, disinfectant type, dose and residual,

and contact time (Rossman et al., 2001; Liang and Singer, 2003; Baribeau et al., 2005). Metal

pipe materials and their corrosion byproducts may also affect the formation and the fate of

halogenated DBPs in distribution systems as has been reported for iron and copper-based pipes

by Li et al. (2007), Zhang and Andrews (2012) and Rossman et al. (2001). However, no relevant

information concerning the impacts of pipe deposits on nitrosamine formation has been reported

at the time of writing, and the literature concerning the fate of nitrosamines in full-scale

distribution systems is varied (Wilczak et al., 2003; Baribeau et al., 2006; Charrois et al., 2007).

147

In addition, given the impacts of flow conditions on chloramine degradation and that chloramine

concentration is a rate-limiting factor for NDMA formation, the impacts of flow conditions on

NDMA formation merits some investigation.

Therefore, the objective of this study was to evaluate the impacts of pipe materials

(ductile iron, PVC, copper and lead) on chloramine stability and NDMA formation in the

absence and presence of orthophosphate as a corrosion inhibitor and, in particular, under

different flow conditions. These pipe materials are widely present in either water mains (such as

ductile iron and PVC) or household plumbing systems (such as copper), and lead can be

commonly found in old lead service lines, soldered joints and brass plumbing fittings. Modified

pipe loops were used in the tests under typical chloramination conditions. Results of this study

improve our understanding about key factors that affect chloramine stability and NDMA

formation in the complex physiochemical and biological environment of distribution systems.



All chemicals used in this study were ACS grade or higher. Orthophosphate (K3PO4) was

the corrosion inhibitor selected for testing. The test water was the chloraminated reservoir

effluent from the Mannheim Water Treatment Plant (MWTP), Ontario. The chloramine

concentration in the test water ranged from 1.3 mg/L to 1.6 mg/L as Cl2, and other water quality

parameters in the test water are listed in Table 7-1.

Table 7-1 Summary of water quality parameters for the influent of the pipe loops

Parameters Values

pH 7.4 ± 0.1

Alkalinity (mg/L) 224±11

UV254 (cm-1

) 0.039±0.007

TOC (mg/L) 2.58 ± 0.25

SUVA (L/mgcm-1

) 0.015±0.003

Ammonia (mg/L as N) 0.17±0.05

Nitrate (mg/L as N) 4.06±0.34

148

Due to the consistently low formation of NDMA in the test water, one NDMA precursor,

dimethylamine (DMA, purchased from Sigma-Aldrich), was spiked to ensure NDMA formation

at a measurable level (> 1 ng/L). A dosing solution containing DMA and/or orthophosphate was

prepared by adding 1 mL 25 g/L (0.56 M) DMA and/or 20 g K3PO4 into 25 L unchloraminated

post-filtration water from the MWTP. The target concentrations of orthophosphate and DMA in

the influent were 1 mg/L as P and 20 µg/L (450 nM), respectively.

Modified Pipe Loops 7.2.2

Dalhousie pipe loops that were designed and manufactured by Dr. Gagnon’s research

group (Gagnon et al., 2008) were modified according to Cantor (2009) to increase their

efficiency and expand the range of operating conditions. Sixteen square test coupons (6.4 6.4

0.16 cm; Metal Samples Co., Alabama, US) were stacked inside 10 cm (ID) Schedule 40 PVC

pipes to create test modules. The module construction is shown in Figure 7-1. The main module

design parameters are summarized in Table 7-2. The lengths of the PVC-, iron-, copper- and

lead-containing modules were determined based on volume-to-surface area ratios commonly

observed in water mains and residential plumbing systems.

Figure 7-1 Module configuration (adapted from Cantor, 2009)

Flange Pipe insertion rack

4” diameter PVC pipe

Internal of module: 8 metal plates stacked on the tongue of the insertion rack, separated by plastic spacers and connected with a PVC bolt

Metal plate identification

149

Table 7-2 Summary of design parameters for pipe loops

V/S*

(mL/cm2)

Module

length (cm)

# of

plates

Equivalent distribution

system pipe diameter

(cm)

PVC 3.9 52 0 16

Iron 4.5 53 16 18

Copper 3.0 26 16 12

Lead 3.2 28 16 13

Note: * Volume-to-surface area ratio;

Volume is calculated as the sum of volumes of the return section and the test module.

The modified pipe loops are illustrated in Figure 7-2. Each pipe loop consisted of 6

components: the test module, the recirculating pump, the return section, the aluminum support

frame, the influent feed pump, and the influent and effluent ports. A PVC flexible hose was used

as the return section. The influent port was located on the return section and was the connection

point for the influent feed pump to provide fresh water to the pipe loop. The effluent port was

located on the module and acted as both a sample collection point and the outflow of the water.

The test water was introduced into the pipe loop by the influent feed pump, the flow rate of

which was adjusted based on the desired retention time and the volume of the pipe loop. The

recirculating pump propelled the water through the pipe loop to provide the desired velocity

(typically 0.3 m/s) within the test section.

Figure 7-2 Schematic of the modified pipe loop (not to scale)

Recirculation pump

Influent feed pump Influent reservoir

Return section

Transition section

Aluminum support frame DMA/PO43-

dosing tank

DM

A/P

O4

3- fe

ed

pu

mp

Effluent port

150


After the pipe loops were modified, they were first flushed with the chloraminated

reservoir effluent for 3 days, and then conditioned for periods of up to 120 days to allow released

metal concentrations to stabilize in the test water matrix. During conditioning, the flow rates in

the four pipe loops were increased stepwise to obtain hydraulic retention times (HRTs) of 12

hours, 6 hours and 2 hours. To eliminate chlorine demand from the newly installed pipes, a high

concentration of free chlorine (390 mg/L) was pumped into the loops at a flow rate of 30~38

mL/min for 3 hours. Due to a persistently high chlorine demand in the copper loop, another two

conditioning treatments (180 mg/L free chlorine, 10 days) were conducted for that loop. After

these treatments, the 6 hour chloramine demand in the PVC, iron and lead loops decreased from

1.1, 1.5 and 1.5 mg/L before treatment to 0.1, 0.5 and 0.5 mg/L, respectively, and the 2 hour

chloramine demand in the copper loop decreased from 1.5 mg/L before treatment to 1.2 mg/L.

Experiments were performed in two phases: first, without the addition of orthophosphate

and then with the addition of 1 mg/L orthophosphate. In each phase, three flow regimes typically

encountered in distribution systems were examined - turbulent, laminar and stagnant. Turbulent

and laminar flow conditions were achieved by turning on and off the recirculating pumps,

resulting in flows with Reynolds numbers (Re) of approximately 30000 and <10, respectively.

HRTs in each pipe loop were varied by adjusting the influent feed pump flow rate. The stagnant

flow condition was achieved by turning off both the influent feed pump and the recirculating

pump.

Water samples from the influent and the effluents of the four pipe loops were collected

three times a week for chloramine, released metal concentrations and NDMA measurement. pH,

TOC, UV254, alkalinity, temperature, phosphate, ammonia, nitrite and nitrate were measured

once a week. To ensure a relatively constant concentration of DMA being dosed into the test

water, the flow rates of the DMA dosing solution and the reservoir effluent were monitored on

each sampling date. In addition, the DMA concentration being dosed into the influent water was

checked before and after the addition of orthophosphate by monitoring the 24 hour NDMA

formation in 1 L amber bottles with the ambient chloramine concentration.

Measurement of total organic carbon (TOC) concentration was undertaken using a Model

1030 TOC analyzer (OI Analytical, USA). UV254 was measured using a CE 3055 Reflectance

151

Spectrophotometer (Cecil Instruments, England) at 254 nm. Measurements of total chlorine and

ammonia were performed using a Hach DR2700 Spectrophotometer (Hach Company, USA) at

530 nm and 425 nm, respectively. Routinely, total chlorine was measured instead of

monochloramine because of greater temperature dependency of the monochloramine test,

however, the total chlorine residuals were shown to consist almost exclusively of

monochloramine. Metal concentrations were analyzed using a Varian SpectrAA.20 Flame

Atomic Absorption Spectrometer (Agilent Technologies, USA) after being acidified to pH <2.

Nitrite, nitrate and phosphate concentrations were measured using a Dionex DX-300 Series Ion

Chromatography System (Thermo Scientific, USA). NDMA extract preparation was carried out

according to Taguchi et al. (1994), and the instrument used for NDMA analysis was a Varian

4000 GC/MS (Agilent Technologies, USA).


Metal and Nitrogen Species Concentrations 7.3.1

Figure 7-3 compares metal concentrations in the absence and presence of orthophosphate

under two flow conditions. Generally, the released metal concentrations were lower under

laminar conditions compared with those under turbulent conditions regardless of orthophosphate

addition, which is consistent with previous studies in that high flow rates can loosen the

corrosion byproducts deposited on the pipe wall and cause more metal to be released from the

pipe surface (Mutoti et al., 2007).

With respect to specific impacts of orthophosphate, as shown in Figure 7-3, the results

varied with both the type of pipe materials employed and the hydraulic conditions present. The

addition of orthophosphate significantly reduced the released copper concentrations under both

laminar and turbulent conditions, as expected (Edwards et al., 2002). Also, as expected,

orthophosphate did not exhibit obviously beneficial effects on iron corrosion control under both

laminar and turbulent conditions, as also evidenced by McNeill and Edwards (2000; 2001).

However, orthophosphate effectively reduced lead release but mainly when water was turbulent.

In the absence of orthophosphate, the released lead concentrations varied from 0.4 to 0.6 mg/L,

but they were below 0.1 mg/L after the addition of orthophosphate. Under laminar conditions,

152

there were no significant differences in lead concentrations before and after the addition of

orthophosphate (varying between 0.1 and 0.2 mg/L).

Figure 7-3 Metal concentrations in the pipes of iron, copper and lead in the absence and presence

of 1 mg/L orthophosphate under two flow conditions

Given that orthophosphate previously has been shown to effectively decrease lead release

(Edwards et al., 1999; Edwards and McNeill, 2002), the observed relative ineffectiveness of

orthophosphate on lead corrosion control under laminar conditions in the current study might be

explained by the relative newness of the module materials (leading to higher than normal lead

release rates) and/or interplays among lead and dissolved species such as ammonia, nitrite and

0.0

0.2

0.4

0.6

0.8

0 3 6 9 12 15

Pb

(m

g/L

)

HRT (hours)

No phosphate spiked

Phosphate spiked

0.0

0.2

0.4

0.6

0.8

0 3 6 9 12 15

Pb

(m

g/L

)

HRT (hours)

No phosphate spiked

Phosphate spiked

0.0

0.2

0.4

0.6

0.8

1.0

0 3 6 9 12 15

Fe

(m

g/L

)

HRT (hours)

No phosphate spiked

Phosphate spiked

0.0

0.2

0.4

0.6

0.8

1.0

0 3 6 9 12 15

Fe

(m

g/L

)

HRT (hours)

No phosphate spiked

Phosphate spiked

0.0

0.5

1.0

1.5

2.0

0 3 6 9 12 15

Cu

(m

g/L

)

HRT (hours)

TURBULENT

No phosphate spiked

Phosphate spiked

0.0

0.5

1.0

1.5

2.0

0 3 6 9 12 15

Cu

(m

g/L

)

HRT (hours)

LAMINAR

No phosphate spiked

Phosphate spiked

153

nitrate (Edwards and Dudi, 2004; Uchida and Okuwaki, 1998; Zhang et al., 2009a) as shown in

Reactions 7-1 and 7-2:

NO3- + Pb NO2

- + PbO Reaction 7-1

NO2- + 3Pb +2H2ONH3 + 3PbO+OH

- Reaction 7-2

According to these reactions, free ammonia from the application of monochloramine combined

with a high concentration of nitrate may have synergistically increased the lead corrosion by

interfering with the formation of the passive PbO layer on its surface. If this phenomenon was

occurring, then according to Reactions 7-1 and 7-2, there should be a strong correlation among

the concentrations of the three nitrogen species. Namely, a decrease in nitrate concentrations

should lead to an increase in both nitrite and ammonia concentrations. Figure 7-4 illustrates the

variations in the nitrogen species concentrations measured in the lead loop. It confirms that

decreases in nitrate concentrations were associated with increases in both nitrite and ammonia

concentrations and provides evidence for the potential occurrence of Reactions 7-1 and 7-2 in the

lead loop. The co-presence of ammonia, nitrite and nitrate in significant concentrations in the

lead loop may also suggest the occurrence of nitrification as a potential source of these ions, but

confirmation of this was beyond the scope of the study and requires further investigation.

Figure 7-4 Variations of nitrate, nitrite and ammonia in the lead loop

0.0

0.1

0.2

0.3

3.0

3.5

4.0

4.5

5.0

27-Apr-11 17-May-11 6-Jun-11 26-Jun-11

Am

mo

nia

/nitrite

(m

g/L

-N)

Nitra

te (

mg

/L-N

Nitrate Nitrite Ammonia

154

Chloramine decay 7.3.2

Generally, chloramine concentrations in the four pipe loops all followed pseudo-first-

order decay kinetics, so decay constants were estimated by fitting a pseudo-first-order decay

equation to the chloramine data using Microsoft Excel.

7.3.2.1 Effects of Orthophosphate and Flow Conditions

Figure 7-5 displays pseudo-first-order chloramine decay constants for the four pipe

materials before and after the addition of orthophosphate under three flow regimes (Reynolds

numbers of Re=30000, Re 10, and Re=0). In addition, for each pipe material, a Fisher’s Least

Significant Difference (LSD) test at a confidence level of 95% was applied to pairs of

chloramine decay constants that were obtained with different treatments to estimate the

significance of the effects of different flow conditions and the presence or absence of

orthophosphate on chloramine decay (Montgomery, 2000). The results of the LSD tests to

determine the effects of flow conditions are summarized in Table 7-3.

Figure 7-5 Pseudo-first-order chloramine decay constants for the four pipe loops under different

flow conditions, initial chloramine 1.3~1.6 mg/L as Cl2, error bars indicate the measured


0.0

0.2

0.4

0.6

0.8

1.0

Re=0 Re<10 Re=30000

De

ca

y c

on

sta

nt (h

-1)

Fe

0.0

0.2

0.4

0.6

0.8

1.0

Re=0 Re<10 Re=30000

Deca

y c

on

sta

nt (h

-1)

Cu

0.0

0.2

0.4

0.6

0.8

1.0

Re=0 Re<10 Re=30000

Deca

y c

on

sta

nt (h

-1)

Pb

0.0

0.2

0.4

0.6

0.8

1.0

Re=0 Re<10 Re=30000

Deca

y c

on

sta

nt (h

-1)

PVC

No orthophosphate Orthophosphate spiked

155

Table 7-3 Significance of the effects of flow conditions on chloramine decay determined by the

LSD test (95% confidence level)

Treatments compared Fe Pb Cu PVC

No

orthophosphate

Re=0 Re<10 ↓

Re<10 Re=30000 ↑ ↓ ↑ ↑

Re=0 Re=30000 ↑ ↑ ↑ ↑

Orthophosphate

spiked

Re=0 Re<10 ↑ ↓ ↓ ↑

Re<10 Re=30000 ↑ ↑ ↑

Re=0 Re=30000 ↑ ↑ ↑

Note: , no significant effect; ↑, increasing decay; ↓, decreasing decay

Generally, chloramine decay constants for all four pipe materials increased with

increasing water flow. This agrees with previous findings that an increase in flow velocity

increases the mass transfer rate of chloramine to the pipe surface and/or increases the release rate

of metal corrosion products from the pipe surface, thereby increasing chloramine consumption

(Rossman et al., 1994; Westbrook and Digiano, 2009). The results for the PVC loop also agree

with a previous finding that bulk chlorine decay rates are dependent on flow velocity (Menaia et

al., 2003).

The decay constants were also influenced by the type of pipe material present, which is

newly identified in this study. As shown in Figure 7-5, there was a marked increase in

chloramine decay in both the lead and copper pipes under turbulent conditions. The rapid loss of

chloramine was observed to occur in the initial 1~2 hours of the tests during which time

approximately 70% of chloramine was degraded.

In the lead loop, the chloramine decay may have been enhanced by the presence of nitrate

in the water. As discussed in Section 7.3.1, the capability of lead to react with nitrate has been

well documented (Uchida and Okuwaki, 1998; Edwards and Dudi, 2004; Zhang et al., 2009a).

Reactions 7-1 and 7-2 can essentially increase nitrite formation and lead corrosion rates. Under

turbulent conditions, up to 0.08 mg/L nitrite as N and 0.4 mg/L Pb (Figures 7-3 and 7-4) were

detected in the lead pipe in the absence of orthophosphate, and they were 40% and 75% higher,

respectively, than those in the presence of orthophosphate. Since chloramine reacts with NO2-

and considering that NO2- is formed in Reaction 7-1, it was reasonable to have observed

accelerated chloramine degradation in the lead loop.

156

The high chloramine demand in the copper pipe under turbulent conditions is consistent

with previous findings that chloramine can be rapidly consumed by copper corrosion via the

following reaction (Nguyen, 2005):

½NH2Cl +H+ +½Cu

0 + ½Cu

2+ Cu

2+ + ½NH4

+ + ½Cl

- Reaction 7-3

Under turbulent conditions, the copper surface can be scoured, exposing fresh copper and

providing a virtually unlimited supply of metallic copper (Cu0) to participate in this reaction. The

observed rapid chloramine decay under these conditions is, therefore, primarily due to the high

chlorine demand from the direct oxidation of fresh copper to cupric species.

To evaluate the significance of the presence or absence of orthophosphate on chloramine

decay, the LSD tests were also applied and the results are summarized in Table 7-4. As shown in

Table 7-4 and also in Figure 7-5, the addition of orthophosphate did not cause statistical changes

in chloramine degradation at Re =0 and Re <10 for both the copper and PVC loops. However,

under turbulent flow (Re =30,000), orthophosphate decreased chloramine degradation in the

copper loop, and the opposite effect was observed in the PVC loop. Surprisingly, orthophosphate

significantly increased chloramine decay in the iron loop, but only at Re =0 and Re <10 for the

lead loop.

Table 7-4 Significance of the effects of orthophosphate on chloramine decay determined

by the LSD test (a confidence level of 95%)

Materials Treatments Re=0 Re<10 Re=30000

Fe

No

orthophosphate

Orthophosphate

spiked

↑ ↑ ↑

Pb ↑ ↑ ↓

Cu ↓

PVC ↑

Note: , no significant effect; ↑, increasing decay; ↓, decreasing decay

For conditions under which orthophosphate significantly increased chloramine decay (in

the iron and lead loops), Figure 7-3 indicates that metal release kinetics were comparable before

and after the addition of orthophosphate. This suggests that, rather than metal corrosion

processes, other factors (e.g. nitrite formation) may have played an important role in accelerating

chloramine decay. Further discussions about this theory are provided in Section 7.3.2.2.

157

7.3.2.2 Chloramine Decay and Nitrite Formation

Nitrite (NO2-) may be present from either or both of abiotic reactions (e.g. Reaction 7-1

when in contact with lead) and from nitrification (Wilczak et al., 1996), and its impacts on

chloramine decay for each type of pipe material present were further investigated in this study.

Although the presence of nitrifiers could not be detected directly in this study, nitrite

concentrations were measured in the effluents of the four pipe loops. These data are plotted in

Figure 7-6 along with the corresponding chloramine pseudo-first-order decay constants for two

tested flow conditions (turbulent and laminar).

Figure 7-6 Box and Whisker plots for nitrite concentrations before and after the addition of

orthophosphate overlaid with pseudo-first-order chloramine decay rate constants, laminar flow

(n=5), turbulent flow (n=6). The top and bottom of the box represent the 75th and 25th

percentile, respectively, while the whiskers represent the maximum and minimum values

Nitrite Decay constants

0.0

0.2

0.4

0.6

0.8

1.0

0.0

0.1

0.2

0.3

Pse

ud

o-f

irst-

ord

er

de

ca

y

co

nsta

nt (h

-1)

Nitri

te (

mg

/La

s N

) LAMINAR

0.0

0.2

0.4

0.6

0.8

1.0

0.0

0.1

0.2

0.3

Pse

ud

o-f

irst-

ord

er

de

ca

y

co

nsta

nt (h

-1)

Nitri

te (

mg

/L a

s N

) TURBULENT

158

In general, a high concentration of nitrite corresponded to a rapid chloramine degradation

constant under both laminar and turbulent conditions, as expected. In addition, under laminar

conditions, the iron and lead loops exposed to orthophosphate had a relatively higher

concentration of nitrite (~ 0.2 mg/L) compared with other pipe materials (mostly below 0.05

mg/L). This was in agreement with previous findings that orthophosphate can enhance

nitrification and, therefore, nitrite formation (Zhang and Edwards, 2010b). However, the

measured nitrite concentration did not always increase with the addition of orthophosphate, as

observed for the iron and lead loops under turbulent flow conditions. Turbulent flow can disturb

the stability of biofilms where nitrifiers could be present, thereby decreasing their activity

(Zhang et al., 2009b). Decreased nitrite formation in the lead loop after the addition of

orthophosphate under turbulent conditions is also likely because the protective layer on the lead

surface decreased the exposure of lead to nitrate in the water (Reaction 7-1).

Nitrite concentrations in the copper loop were consistently lower than in the other loops,

with the maximum concentrations for both flow conditions being below 0.02 mg/L (Figure 7-6).

It is known that high levels of copper (> 0.1 mg/L) are toxic to nitrifiers, thereby inhibiting

nitrification (Zhang et al., 2009b) and the measured copper concentrations in this study ranged

from 0.08 to 1.0 mg/L (Figure 7-3). The lack of nitrite in the copper loop also suggests that

abiotic consumption of chloramine due to copper corrosion (Reaction 7-3) was a likely reason

for the rapid chloramine dissipation under turbulent flow (Figure 7-5).

NDMA formation 7.3.3

In addition to determining the potential effects of pipe materials, orthophosphate and

flow conditions on NDMA formation, the pipe loops were operated at different hydraulic

retention times (HRTs) to evaluate the impacts of reaction time and provide additional

confirmation of the observed results.

7.3.3.1 Effects of Flow Conditions

Figure 7-7 provides a general overview of the results by comparing NDMA

concentrations in the effluent of the four pipe loops in the absence of orthophosphate under

turbulent (Re =30000) and laminar (Re <10) conditions. Results for tests performed with the

addition of 1 mg/L orthophosphate were similar, except that all of the NDMA concentrations

159

were reduced (Appendices Figure 10-21, to be discussed further in later section). Generally,

NDMA formation in each pipe loop increased with increasing HRT, as expected. Under

turbulent conditions, the sequence of NDMA formation potentials in the four pipe loops followed

the order of Cu Pb< Fe< PVC, whereas chloramine decay constants followed the opposite order

(Figure 7-5), which is in agreement with common knowledge that chloramine presence is a rate-

limiting factor for NDMA formation.

Figure 7-7 NDMA formation in four pipe loops in the absence of orthophosphate under two flow

conditions

It was interesting to observe that changing flow conditions changed these relationships.

NDMA concentrations were generally higher under laminar conditions than turbulent conditions

and the pipe materials exhibited even more pronounced but different effects on NDMA

formation. NDMA formation did not always directly correlate with chloramine presence. For

0

40

80

120

160

200

2 6 12

ND

MA

(ng/L

)

HRT (hours)

TURBULENT

0

40

80

120

160

200

2 6 12

ND

MA

(ng/L

)

HRT (hours)

LAMINAR

160

example, although chloramine decay constants in the four pipe loops were statistically similar

under this condition (Figure 7-5), the NDMA formation potentials followed a distinctly different

sequence (an increasing order of Fe, Pb, PVC and Cu, as illustrated in Figure 7-7). In particular,

the measured NDMA concentration in the copper loop tended to plateau at approximately 180

ng/L after 6 hours, but in the PVC loop (the control loop) there was only 65 ng/L NDMA at HRT

of 6 hours and 150 ng/L NDMA present after 12 hours. Thus, the chloramine decay constants

were comparable but NDMA formation was enhanced in the copper loop relative to the PVC

loop, indicating that Cu(II) may have catalyzed NDMA formation under laminar flow conditions.

In addition, for the iron loop, although chloramine degraded more quickly under turbulent

conditions than under laminar flow (Figure 7-5 and Table 7-3), Figure 7-7 shows that NDMA

yields after HRT 6 hours under turbulent conditions were at least 70% higher than those under

laminar conditions. Catalytic effects of iron corrosion products during NDMA formation may,

therefore, be suggested. In general, reactions mediated by surface-bound metal are affected by

the mineral’s surface area and the density of sorbed metal ions (Chun et al., 2005). Under

turbulent conditions, the high velocity of the water may enhance the release of iron corrosion

products from the pipe wall, thereby increasing the surface area of iron corrosion products in

contact with DMA. As a consequence, NDMA formation was likely catalyzed by iron corrosion

products under turbulent flow in the current tests despite the relatively rapid chloramine

degradation. These observations concerning the dependency of copper and iron influences on

flow conditions have not been reported previously.

7.3.3.2 Effects of Orthophosphate

Since orthophosphate’s corrosion preventive effects can generally reduce the impacts of

the tested parameters on corrosion rates and metal concentrations, this study also investigated the

effects of orthophosphate on NDMA formation. However, partway through the experimental

plan, the treatment plant performed some operational changes (increasing polyaluminum

chloride, PACl) to improve their filter performance. Coincidently, NDMA yields in the four pipe

loops significantly decreased. Since the make-up and flow rate of the orthophosphate/DMA

dosing solution and the flow rate of the reservoir effluent remained constant before and after the

operational changes, the decreased NDMA formation was most likely attributed to the change in

quality of the chloraminated reservoir water as a result of the increased PACl addition. Such

161

effects of water matrix on NDMA formation has also been reported by Oya et al. (2008).

Regardless, to account for variations in water quality, the effects of orthophosphate on NDMA

formation were determined after first normalizing the NDMA concentrations in the different

loops relative to those observed in the PVC control loop.

Normalized NDMA concentrations are shown in Figure 7-9. Since the calculated relative

percentage values of NDMA formation in the PVC loop were all 100%, they are not shown

individually, but are indicated by the horizontal line at 100%.

Figure 7-9 NDMA formation relative to the PVC control in the absence and presence of 1 mg/L

orthophosphate under laminar and turbulent flow conditions

Under turbulent conditions, NDMA formation relative to the PVC loop in all of the three

metal loops in the presence of orthophosphate increased compared with that in the absence of

orthophosphate. The greatest increase was observed in the iron loop, in which more NDMA

formed than in the PVC loop after the addition of orthophosphate. Recall that there was higher

0

50

100

150

200

250

300

350

2 6 12 2 6 12

ND

MA

form

ation r

ela

tive t

o the

PV

C c

ontr

ol (%

)

TURBULENT

Fe Cu Pb

No orthophosphate Orthophosphate

0

50

100

150

200

250

300

350

2 6 12 2 6 12

ND

MA

form

ation r

ela

tive t

o the

PV

C c

ontr

ol (%

)

HRT (hours)

LAMINAR

Fe Cu Pb

No orthophosphate Orthophosphate

162

NDMA formation but more rapid chloramine decay in the iron loop under turbulent conditions

than laminar conditions, suggesting the possible catalysis of iron on NDMA formation (Section

7.3.3.1). The additional evidence of increased NDMA formation after the addition of

orthophosphate in the iron loop indicates that orthophosphate can significantly increase these

catalytic effects of iron corrosion products, possibly by affecting the properties of iron particles

and their suspension in the solution (Lytle and Snoeyink, 2002). As for the copper and lead

loops, according to Table 7-4, orthophosphate significantly decreased chloramine decay by

reducing metal corrosion rates. Since chloramine decay is a rate-limiting factor for NDMA

formation, it is also reasonable to have observed the increased formation of NDMA in these two

loops after the addition of orthophosphate.

Under laminar conditions, the main observation is that NDMA formation relative to the

PVC control in the copper loop is significantly higher than 100%, especially in the absence of

orthophosphate. This confirms the trends shown in Figure 7-7 that illustrate copper catalysis of

NDMA formation under laminar conditions. Furthermore, the addition of orthophosphate

decreased the percentage values of NDMA formation relative to the PVC control in the copper

loop for all the HRTs. This suggests orthophosphate may mitigate copper catalysis on NDMA

formation by decreasing the released copper concentrations (Figure 7-3).

7.4 Summary

This is the first study to employ modified pipe loops to determine the effects of pipe

materials, flow conditions and orthophosphate on chloramine decay and NDMA formation from

DMA. Ductile iron, copper, lead and PVC were tested. Three flow regimes encountered in

distribution systems (turbulent, laminar and stagnant) were examined. In general, the pipe

materials, flow conditions and orthophosphate were all observed to influence metal corrosion,

chloramine decay and NDMA formation as follows:

a. In agreement with previous studies, turbulent conditions generally increased the released

metal concentrations compared with laminar conditions regardless of orthophosphate

addition. Orthophosphate had no beneficial effects on iron corrosion control, but

effectively reduced copper concentrations under both laminar and turbulent conditions.

163

b. Chloramine degradation kinetics for four pipe materials were compared. The new

observation was that the impacts of flow conditions and orthophosphate on chloramine

decay were highly dependent on the type of pipe material. Orthophosphate increased

chloramine degradation in the iron loop regardless of flow conditions. Abiotic copper

corrosion was the primary reason for the accelerated chloramine decay under turbulent

condition. For the lead loop, a rapid dissipation of chloramine was primarily due to the

electrochemical reactions between metallic lead with nitrate to form nitrite. The addition

of orthophosphate mitigated the impacts of turbulence on chloramine decay in both the

copper and lead loops.

c. Consistent with previous studies, chloramine concentration was generally a rate-limiting

factor for NDMA formation, especially under turbulent conditions.

d. The impacts of flow conditions on metal catalysis of NDMA formation were newly

identified. Regardless of the presence of orthophosphate, copper catalyzed NDMA

formation from DMA under laminar conditions, but iron catalyzed NDMA formation

only under turbulent conditions. Iron catalysis was increased by the addition of

orthophosphate likely because orthophosphate modified properties of the associated

suspended particles, while orthophosphate decreased copper catalysis likely by reducing

dissolved copper concentrations.

Results of this study help in examining the complex reactions involving metal corrosion,

chloramine degradation and NDMA formation under different flow conditions in distribution

systems. Since it is an initial study to have identified the catalytic impacts of iron and copper on

NDMA formation and their dependency on flow conditions, further study is recommended to

investigate the mechanisms for iron and copper catalysis. Nevertheless, knowing more about

impacts of pipe materials and corrosion inhibitors as well as the possible occurrence of

nitrification on the stability of chloramine residuals and subsequent NDMA formation will be

useful in developing strategies to control metal corrosion and reduce disinfection by-product

formation in distribution systems.

164

7.5 References

Baribeau, H., Krasner, S. W., Chinn, R., and Singer, P. C. (2005) Impact of biomass on the

stability of HAAs and THMs in a simulated distribution system. Journal American Water






Cantor, A.F. (2009) Water Distribution System Monitoring: A Practical Approach for Evaluating

Drinking Water Quality. CRC Press, ISBN: 9781439800522.







Chun, C. L., Hozalski, R. M., and Arnold, T. A. (2005) Degradation of drinking water


Technology, 39(21), 8525-8532.

Digiano, F. A., and Zhang, W. D. (2005) Pipe section reactor to evaluate chlorine-wall reaction.






165





Gagnon, G. A., Baribeau, H., Rutledge, S. O., Dumancic, R., Oehmen, A., Chauret, C., and

Andrews, S. (2008) Disinfectant efficacy in distribution systems: a pilot-scale

assessment. Journal of Water Supply Research and Technology-Aqua, 57(7), 507-518.



36(14), 3479-3488.





Technology, 37(13), 2920-2928.

Lytle, D. A., and Snoeyink, V. L. (2002) Effect of ortho- and polyphosphates on the properties of

iron particles and suspensions. Journal American Water Works Association, 94(10), 87-

99.





Menaia, J., Coelho, S. T., Lopes, A., Fonte, E., and Palma, J. (2003) Dependency of bulk

chlorine decay rates on flow.velocity in water distribution networks. 3rd World Water

Congress: Water Services Management, Operations and Monitoring, 3(1-2), 209-214.

166


Sons, New York.






Nguyen, C.K. (2005) Interactions between copper and chlorine disinfectants: chlorine decay,

chloramine decay, and copper pitting. Master’s thesis, Civil and Environmental

Engineering, Virginia Polytechnic and State University, Blacksburg, Virginia, USA.

Oya, M., Kosaka, K., Asami, M. and Kunikane, S. (2008) Formation of N-nitrosodimethylamine

(NDMA) by ozonation of dyes and related compounds. Chemosphere, 73(11): 1724-

1730.

Rossman, L. A., Clark, R. M., and Grayman, W. M. (1994) Modeling chlorine residual in

drinking-water distribution systems. Journal of Environmental Engineering-Asce, 120(4),

803-820.



Taguchi, V., Jenkins, S. D. W., Wang, D. T., Palmentier, J., and Reiner, E. J. (1994)

Determination of N-Nitrosodimethylamine by isotope-dilution, high-resolution mass-

spectrometry. Canadian Journal of Applied Spectroscopy, 39(3), 87-93.





Association, 98(2), 144-154.

167



Wilczak, A., Assadi-Rad, A., Lai, H. H., Hoover, L. L., Smith, J. F., Berger, R., Rodigari, F.,

Beland, J. W., Lazzelle, L. J., Kinicannon, E. G., Baker, H., and Heaney, C. T. (2003)






Zhang, H., Andrews, S.A. (2012) Catalysis of copper corrosion products on chlorine decay and

HAA formation in simulated distribution systems. Water Research, 46 (8), 2665-2673.



Zhang, Y., Griffin, A., Rahman, M., Camper, A., Baribeau, H., and Edwards, M. (2009a) Lead

Contamination of Potable Water Due to Nitrification. Environmental Science &

Technology, 43(6), 1890-1895.

Zhang, Y., Love, N., and Edwards, M. (2009b) Nitrification in Drinking Water Systems. Critical

Reviews in Environmental Science and Technology, 39(3), 153-208.

168

Conclusions 8

The specific objectives of this thesis outlined in Section 1.2 were addressed in this work and

resulted in some main conclusions. Generally, of the results discussed below, those that are

summarized in Conclusion 1 (from Chapter 3) compared the impacts of corrosion inhibitors and

their interactions with metal surfaces on chlorine degradation and HAA9 formation, and the

results in Conclusions 3 to 5 (from Chapters 5 to 7) in terms of copper catalysis and flow

conditions on HAA/NDMA formation are essentially new contributions to the field. The results

in Conclusion 2 (from Chapter 4) concerning the impacts of disinfectant type, corrosion

inhibitors and water quality on metal release kinetics confirmed those of previous studies.

The main conclusions of this thesis include:

1) Phosphate-based corrosion inhibitors and their interactions with metal materials affected

the kinetics of disinfectant degradation (both HOCl and NH2Cl) and DBP formation

(mainly HAAs) in two types of water matrices, depending on metal type and metal age.

Newly discovered was that HAA formation was enhanced in the presence of high levels of

copper ions from fresh metal coupons in both investigated water matrices, indicating

possible catalytic potential of copper on HAA formation. The addition of phosphate-based

corrosion inhibitors alone generally did not impact HAA formation in both tested water

matrices. Enhanced formation of NDMA from DMA in water containing fresh iron and

copper coupons also indicated the possible catalysis of iron and copper on NDMA

formation, and that the catalysis may increase with the released metal concentrations.

Consistent with previous studies, both HOCl and NH2Cl generally followed pseudo-

first-order kinetics regardless of metal age and water quality. The reactivity of these pipe

materials with HOCl and NH2Cl followed the sequence of decreasing order of ductile iron,

copper and lead. For fresh metal coupons in both tested water matrices, the addition of

phosphate-based corrosion inhibitors significantly increased chlorine decay for iron

coupons, but reduced chlorine decay for copper coupons. For corroded coupons in both

tested water matrices, however, orthophosphate did not statistically affect free chlorine

decay, regardless of metal type.

2) Disinfectant types, corrosion inhibitors and water quality affected metal release kinetics in

simulated distribution systems. Early experiments generally confirmed the results of other

169

researchers and also suggested that orthophosphate increased the level of released iron,

regardless of the age of metal surface, water quality, and disinfectant type. However,

orthophosphate significantly decreased released copper concentrations, in particular for

short term exposures to fresh copper metal. For corroded copper coupons, an increased

copper corrosion rate was observed in the presence of orthophosphate under chlorination

for the water with high pH and alkalinity values. Lead release was significantly reduced in

the presence of orthophosphate, irrespective of the age of metal surface and water quality.

Regardless of orthophosphate addition, high levels of metal ions were released in the water

with low pH and alkalinity, which is consistent with other research.

3) Copper corrosion products, including Cu(II), Cu2O, CuO, and Cu2(OH)2CO3, exhibited

catalytic effects on HOCl decay and HAA formation. Copper catalysis was affected by pH

and the concentration of these corrosion products. The presence of Cu(II) and its solid

corrosion products led to DCAA formation consistently predominating over the formation

of other HAA species. Further investigation of chlorine decay pathways in the presence of

Cu(II) in synthetic water indicated that Cu(II) would interact with NOM, possibly by

complexation, and increase the reactivity of NOM with chlorine. As a result, chlorine

decay was accelerated by reacting with these active Cu(II)-NOM complexes and HAA

formation was enhanced.

4) Copper was shown to catalyze NDMA formation from DMA. NDMA formation from DMA

was increased with increasing Cu(II) concentrations, DMA concentrations, alkalinity and

hardness, but was inhibited by the presence of NOM. The rapid consumption of NH2Cl by

NOM and/or the competitive complexation of NOM with copper were proposed to be

involved in limiting NDMA formation by NOM. pH influenced the speciation of

chloramine and the interactions of copper with DMA. Elevated formation of NDMA at

neutral pH was primarily attributed to the transformation of monochloramine to

dichloramine and complexation of copper with DMA. In addition, aqueous copper released

from malachite [Cu2CO3(OH)2] was shown to promote NDMA formation while the

presence of CuO decreased NDMA formation.

5) Turbulent conditions were shown to increase chloramine decay, but affected NDMA

formation differently, particularly for copper and iron. Orthophosphate increased

chloramine degradation in the iron loop irrespective of flow conditions. Accelerated

chloramine decay was observed in the copper loop under turbulent flow conditions

170

primarily due to abiotic copper corrosion. The addition of orthophosphate effectively

decreased chloramine decay by reducing the released copper concentrations. Rapid

chloramine decay was also observed in the lead loop under turbulent conditions, but the

addition of orthophosphate mitigated the impacts of turbulence on chloramine decay. Bulk

water chlorine reactions in the PVC loop were also increased with increasing flow velocity,

as has also been reported by others.

Chloramine concentration was generally a rate-limiting factor for NDMA

formation, and high chlorine demands due to metal corrosion and/or nitrite formation was

associated with reduced NDMA formation. Pipe materials affected the transformation of

NDMA to DMA: copper consistently exhibited its catalysis on NDMA formation from

DMA under laminar conditions, whereas iron catalyzed NDMA formation only under

turbulent conditions. Orthophosphate was shown to reduce catalytic effects of copper but

appeared to increase iron catalysis by affecting the metal corrosion processes.

Practical Implications and Suggestions for Future Research 9

This research investigated the impacts of metal corrosion and corrosion inhibitors on

disinfectant residual degradation and DBP formation (HAA and NDMA) at both bench- and

pilot-scale. Metal materials that were investigated include ductile iron, copper and lead. Key

factors affecting metal corrosion, disinfectant residual degradation and subsequent DBP

formation were examined, including metal age, water quality, and flow conditions. The catalytic

potential of copper to affect HAA and NDMA formation were evaluated under controlled

experimental conditions at bench-scale, with further experiments concerning NDMA formation

being performed at pilot-scale.

The findings about copper catalysis during HAA and NDMA formation provide

important implications for distributed water quality in domestic plumbing systems where

primarily copper pipes are installed. These pipe surfaces are often covered by corrosion solids.

Interactions of disinfectant residuals and DBP precursors with copper corrosion products will

affect the stability of secondary disinfectants and the fate of DBPs. Understanding the catalytic

potential of copper on disinfectant degradation and HAA/NDMA formation will be of benefit to

water utilities and households for the management of their distribution systems and water

quality. Since copper concentration, pH, reaction time and flow conditions are the main factors

171

impacting the nature and extent of copper catalysis in the present study, considerations for

corrosion preventive strategies are recommended to include: pH adjustment and addition of

corrosion inhibitors at the treatment plant to control copper corrosion, and flushing prior to use in

the home to decrease contact time of water with pipe corrosion products. These steps will help to

alleviate HAA and NDMA formation by copper catalysis and ensure safe and clean water is

delivered to the consumer.

This research helps achieve a better understanding of the complex reactions involving

metal corrosion, disinfectant residual degradation and DBP formation. The results provide

insight into the factors affecting disinfectant residual stability and the fate of DBPs in

distribution system water mains and household plumbing. Although phosphate-based corrosion

inhibitors have received increasing attention as an effective strategy for corrosion control, some

of these corrosion inhibitors have been shown to exhibit either beneficial or detrimental effects

on disinfectant decay and subsequent DBP formation. The effects observed in the present

research have depended on the metal types, metal age, hydrodynamic conditions and water

quality. As well, these potential impacts should be considered site-specific.

In general, results from this research will provide water utilities with insight into

efficiently targeting their water characteristics and coordinating, spatially and temporally, the

critical factors to maintain their distributed water quality. For example, booster chlorination or

chloramination facilities may be used for the pipe systems with high water age or where rapid

disinfectant degradation occurs (e.g. in turbulent conditions). Controlling water chemistry (such

as pH and alkalinity) has been confirmed to be able to minimize metal corrosion, and thus it can

be considered when the addition of phosphate-based corrosion inhibitors is not effective for

corrosion control (e.g. iron). Selection of alternative pipe materials to control metal leaching will

also mitigate aesthetic and public health impacts caused by increased metal concentrations.

Limitations of this research and recommendations for future study include:

1) The impacts of copper catalysis on NDMA formation need to be investigated further in

different water matrices. Although bench-scale experiments were performed to investigate

factors of alkalinity, pH, NOM, and hardness affecting copper catalysis during NDMA

formation from DMA in Milli-Q water under controlled experimental conditions (Chapter

6), in real water, different combinations of levels of these parameters may complicate

172

NDMA formation. For instance, different pH and alkalinity conditions may affect the

speciation and solubility of hardness ions. As well, NOM in different water matrices has

different compositions and binding affinities with copper, which may affect the interactions

between DMA and copper.

2) Both bench- and pilot-scale experiments involving NDMA formation employed DMA as a

model or surrogate NDMA precursor (Chapters 6 and 7). This particular precursor may

only be relevant in worst-case scenarios in distribution systems where wastewater

significantly impacts drinking water sources or intrusion occurs introducing a significant

amount of DMA into the distributed water. However, it is unlikely that there are significant

concentrations of DMA in drinking water. Factors affecting NDMA formation from other

precursors in distribution systems needs to be examined. These precursors may include

tertiary amines containing DMA moieties, amine containing polymers, and humic

substances typically present in drinking water.

3) Pilot-scale pipe loop experiments investigated the impact of flow conditions on total

chlorine decay and NDMA formation (Chapter 7), however, only three flow conditions

were considered (Re=0, Re<10 and Re=30000 representing stagnant, laminar and turbulent

flow in distribution systems, respectively). A wider range of Reynolds numbers commonly

encountered in distribution systems should be studied in the future for a better

understanding of the fate of NDMA in distribution systems.

4) In pilot-scale pipe loop experiments, the enhanced formation of NDMA from DMA by iron

was hypothesized to be due to iron catalysis (Chapter 7), however, further work with iron

could not be accommodated in this thesis. Therefore, the mechanisms related to catalysis of

iron corrosion products on NDMA formation need to be identified and the critical factors

influencing the catalytic potential of these corrosion products need to be evaluated.

173

Appendices 10

10.1 Authorisations to Include Copyright Material in Thesis

NRC RESEARCH PRESS LICENSE

TERMS AND CONDITIONS

Mar 09, 2012

This is a License Agreement between Hong Zhang ("You") and NRC Research Press ("NRC

Research Press") provided by Copyright Clearance Center ("CCC"). The license consists of your

order details, the terms and conditions provided by NRC Research Press, and the payment terms

and conditions.

All payments must be made in full to CCC. For payment instructions, please see information

listed at the bottom of this form.

License Number 2864841496472

License date Mar 09, 2012

Licensed content publisher NRC Research Press

Licensed content publication Canadian Journal of Civil Engineering

Licensed content title Effects of phosphate-based corrosion inhibitors on the kinetics of chlorine

degradation and haloacetic acid formation in contact with three metal materials

Licensed content author Hong Zhang et al.

Licensed content date Jan 1, 2012

Volume number 39

Issue number 1

Type of Use Thesis/Dissertation

Requestor type Author (original work)

Format Electronic

174

Portion Full article

Order reference number

Title of your thesis /dissertation

EFFECTS OF DISINFECTANT CHANGES AND CORROSION CONTROL STRATEGIES

ON DISTRIBUTED WATER QUALITY

Expected completion date Jun 2012

Estimated size (pages) 230

Total 0.00 CAD

Terms and Conditions

General Terms & Conditions

Permission is granted upon the requester's compliance with the following terms and conditions:

1. A credit line will be prominently placed in your product(s) and include: for books the author,

book title, editor, copyright holder, year of publication; for journals the author, title of article,

title of journal, volume number, issue number, and the inclusive pages.

The credit line must include the following wording: "© 2008 Canadian Science Publishing or

its licensors. Reproduced with permission," except when an author of an original article

published in 2009 or later is reproducing his/her own work.

2. The requester warrants that the material shall not be used in any manner that may be

derogatory to the title, content, or authors of the material or to Canadian Science Publishing,

including but not limited to an association with conduct that is fraudulent or otherwise illegal.

3. Permission is granted for the term (for Books/CDs-Shelf Life; for Internet/Intranet-In

perpetuity; for all other forms of print-the life of the title) and purpose specified in your

request. Once term has expired, permission to renew must be made in writing.

4. Permission granted is nonexclusive, and is valid throughout the world in English and the

languages specified in your original request. A new permission must be requested for

revisions of the publication under current consideration.

175

5. Canadian Science Publishing cannot supply the requester with the original artwork or a "clean

copy."

6. If the Canadian Science Publishing material is to be translated, the following lines must be

included: The authors, editors, and Canadian Science Publishing are not responsible for errors

or omissions in translations.

v1.4

If you would like to pay for this license now, please remit this license along with your payment

made payable to "COPYRIGHT CLEARANCE CENTER" otherwise you will be invoiced

within 48 hours of the license date. Payment should be in the form of a check or money order

referencing your account number and this invoice number RLNK500736370.

Once you receive your invoice for this order, you may pay your invoice by credit card.

Please follow instructions provided at that time.

Make Payment To:

Copyright Clearance Center

Dept 001

P.O. Box 843006

Boston, MA 02284-3006

For suggestions or comments regarding this order, contact RightsLink Customer

Support: [email protected] or +1-877-622-5543 (toll free in the US) or

+1-978-646-2777.

Gratis licenses (referencing $0 in the Total field) are free. Please retain this printable license for

your reference. No payment is required.

176

ELSEVIER

Ways to Use Journal Articles Published by Elsevier: A Practical Guide

How authors can reuse their own articles published by Elsevier

General use of articles

May 18, 2012

Authors publishing in Elsevier journals retain wide rights to continue to use their works to

support scientific advancement, teaching and scholarly communication.

An author can, without asking permission, do the following after publication of the author’s

article in an Elsevier-published journal:

• Make copies (print or electronic) of the article for personal use or the author’s own classroom

teaching

• Make copies of the article and distribute them (including via e-mail) to known research

colleagues for their personal use but not for commercial purposes or systematic distribution as

defined on page 3 of this pamphlet

• Present the article at a meeting or conference and distribute copies of the article to attendees

• Allow the author’s employer to use the article in full or in part for other intracompany use (e.g.,

training)

• Retain patent and trademark rights and rights to any process or procedure described in the

article

• Include the article in full or in part in a thesis or dissertation

• Use the article in full or in part in a printed compilation of the author’s works, such as collected

writings and lecture notes

• Use the article in full or in part to prepare other derivative works, including expanding the

article to book-length form, with each such work to include full acknowledgment of the article’s

original publication in the Elsevier journal

• Post, as described on page 3, the article to certain websites or servers

177

10.2 QA/QC Protocols

Various QA/QC measures were undertaken in this research to ensure analytical precision

and accuracy.

1) High purity organic solvents (HPLC grade) were used in analyses. All other chemicals

were at least of analytical grade.

2) All glassware used in experiments were made chlorine demand free before use by

soaking them in a concentrated sodium hypochlorite solution (~1000 mg/L as Cl2) for at

least 24 hours. Thereafter, the bottles were rinsed thoroughly with deionized water and

distilled water, and were heated at 250 °C for at least 4 hours.

3) Surrogate standards were used for HAA9 and nitrosamine analysis to correct any errors

from the extraction and instrumental analysis.

4) Calibration curves for each species of HAA9, NDMA, metals and anions were established

in the range representative of actual sample concentrations. Calibration standards were

prepared and run with each set of samples. Calibration curves for NDMA, HAA9, metals

and anions are provided in Figures 10-1, 10-3, 10-4, 10-11, 10-12 and 10-13.

5) At the beginning of each analytical run, lab water blanks and sample blanks containing

the surrogate standard were injected to condition the instrument and to verify that

interferences were absent.

6) Spike and recovery standards were processed and run with each set of samples to validate

and assess the accuracy of instrumental analysis. Recoveries should range between 70%

and 130% (USEPA, 2003 and 2004). Spike recovery charts for NDMA, MCAA, MBAA,

DCAA, TCAA, DBAA and BCAA are illustrated in Figures 10-2 and 10-5~10-10.

7) Method detection limits (MDLs) of the instrumental methods of analysis for each organic

compound, metal ion and anion of interest were determined by multiplying the standard

deviation of 8 replicates by the Student t value, and are provided in Tables 10-1~10-4.

8) All of the bench-scale and pilot-scale kinetic experiments were conducted, at a minimum,

in two sets with duplicate testing involved in each set in order to test the reproducibility

of the results and to determine the errors associated with sampling, extraction and

measurement for metal ions, disinfectant residual, HAA9, and nitrosamines.

178

GC/MS 10.2.1

Figure 10-1 Example of GC/MS calibration curve for NDMA (NDMA 0~200 ng/L, d6-NDMA

50 ng/L)

Table 10-1 GC/MS method – Method detection limits for NDMA (n=8, 99% confidence level)

Compound Spiked

Concentration

(ng/L)

%

Recovery RSD (%)

MDL

(ng/L)

NDMA 1 ng/L 93 31.8 0.88

y = 55.153x - 5.543 R² = 0.9993

0

50

100

150

200

250

0.0 1.0 2.0 3.0 4.0

ND

MA

co

nce

ntr

atio

n (

ng/L

)

Ratio of peak areas (NDMA/d6-NDMA)

179

Figure 10-2 GC/MS method for NDMA- spike recovery chart

0

20

40

60

80

100

120

140

0 10 20 30 40 50 60

Sp

ike

re

cove

ry (

%)

Sample set

2010 2011

130%

70%

180

GC/ECD 10.2.2


TFBA 100 µg/L)

y = 28.898x + 1.4759 R² = 0.9932

0

20

40

60

80

100

120

0.0 1.0 2.0 3.0 4.0

Concentr

ations (

µg/L

)

Ratio of peak areas

DCAA

y = 10.783x + 0.3458 R² = 0.9967

0

20

40

60

80

100

120

0.0 2.0 4.0 6.0 8.0 10.0

Concentr

ations (

µg/L

)

Ratio of peak areas

TCAA

y = 283.7x + 2.3356 R² = 0.9881

0

20

40

60

80

100

120

0.0 0.1 0.2 0.3 0.4

Concentr

ations (

µg/L

)

Ratio of peak areas

MCAA

y = 33.09x + 2.9221 R² = 0.9911

0

20

40

60

80

100

120

0.0 1.0 2.0 3.0 4.0

Concentr

ations (

µg/L

)

Ratio of peak areas

MBAA

y = 13.485x + 1.7596 R² = 0.9964

0

20

40

60

80

100

120

0.0 2.0 4.0 6.0 8.0

Concentr

ations (

µg/L

)

Ratio of peak areas

BCAA

y = 12.571x + 0.2658 R² = 0.9984

0

20

40

60

80

100

120

0.0 2.0 4.0 6.0 8.0 10.0

Concentr

ations (

µg/L

)

Ratio of peak areas

DBAA

181


TFBA 100 µg/L), continued

Table 10-2 GC-ECD method – Method detection limits for HAA9 (n=8, 99% confidence level)

Compound

Spiked

Concentration

(µg/L)

%

Recovery RSD (%)

MDL

(µg/L)

MCAA 2.0 105 4.1 0.5

MBAA 2.0 96 4.1 0.2

DCAA 2.0 97 4.1 0.2

TCAA 2.0 94 3.7 0.2

BCAA 2.0 94 3.2 0.2

BDCAA 2.0 94 3.3 0.2

DBAA 2.0 99 3.8 0.2

CDBAA 2.0 102 4.5 0.3

TBAA 2.0 81 15.3 0.7

y = 9.7142x - 1.5579 R² = 0.98

0

20

40

60

80

100

120

0.0 2.0 4.0 6.0 8.0 10.0

Concentr

ations (

µg/L

)

Ratio of peak areas

BDCAA

y = 11.721x - 3.0322 R² = 0.9699

0

20

40

60

80

100

120

0.0 2.0 4.0 6.0 8.0 10.0

Concentr

ations (

µg/L

)

Ratio of peak areas

CDBAA

y = 13.781x - 2.7168 R² = 0.9832

0

20

40

60

80

100

120

0.0 2.0 4.0 6.0 8.0

Concentr

ations (

µg/L

)

Ratio of peak areas

TBAA

182

Figure 10-5 GC-ECD method for MCAA- spike recovery chart

Figure 10-6 GC-ECD method for MBAA- spike recovery chart

2009 2010

0

20

40

60

80

100

120

140

0 10 20 30 40

Sp

ike

re

cove

ry (

%)

Sample set

MBAA

2009 2010

130%

70%

0

20

40

60

80

100

120

140

0 10 20 30 40

Sp

ike

re

cove

ry (

%)

Sample set

MCAA

2009 2010

130%

70%

183

Figure 10-7 GC-ECD method for DCAA- spike recovery chart

Figure 10-8 GC-ECD method for TCAA- spike recovery chart

0

20

40

60

80

100

120

140

0 10 20 30 40

Sp

ike

re

cove

ry (

%)

Sample set

DCAA

2009 2010

130%

70%

0

20

40

60

80

100

120

140

0 10 20 30 40

Sp

ike

re

cove

ry (

%)

Sample set

TCAA

2009 2010

130%

70%

184

Figure 10-9 GC-ECD method for DBAA- spike recovery chart

Figure 10-10 GC-ECD method for BCAA- spike recovery chart

0

20

40

60

80

100

120

140

0 10 20 30 40

Sp

ike

re

cove

ry (

%)

Sample set

DBAA

2009 2010

130%

70%

0

20

40

60

80

100

120

140

0 10 20 30 40

Sp

ike

re

cove

ry (

%)

Sample set

BCAA

2009 2010

130%

70%

185

Flame Atomic Absorption Spectrometry 10.2.3

Figure 10-11 Example of FAAS calibration curves for iron, copper and lead

y = 13.813x - 0.4067 R² = 0.9974

0

2

4

6

8

10

0 0.2 0.4 0.6 0.8

Me

tal C

on

ce

ntr

ation

(m

g/L

)

Absorbance

Fe

y = 5.8834x - 0.0437 R² = 0.998

0.0

0.5

1.0

1.5

2.0

2.5

0 0.1 0.2 0.3 0.4

Me

tal C

on

ce

ntr

ation

(m

g/L

)

Absorbance

Cu

y = 18.423x + 0.0167 R² = 0.9911

0.0

1.0

2.0

3.0

4.0

5.0

0.00 0.10 0.20 0.30

Me

tal C

on

ce

ntr

atio

n

(mg/L

)

Absorbance

Pb

186

Table 10-3 FAAS method – Method detection limits for iron, copper and lead

Compound Spiked

Concentration

(mg/L)

%

Recovery RSD (%)

MDL

(mg/L)

Iron 0.2 101 6.7 0.041

Copper 0.1 95 3.3 0.009

Lead 0.1 95 8.3 0.024

187

Ion Chromatography 10.2.4

Figure 10-12 Example of IC calibration curves for chloride, sulphate and bromide

y = 1E-06x + 1.2651 R² = 0.9995

0

20

40

60

80

100

120

0.0E+00 2.0E+07 4.0E+07 6.0E+07 8.0E+07 1.0E+08

Ion

Con

ce

ntr

ation

(m

g/L

)

Area counts

Chloride

y = 0.0035x + 2.6704 R² = 0.9969

0

40

80

120

160

0 10,000 20,000 30,000 40,000 50,000

Ion

Con

ce

ntr

ation

(µ

g/L

)

Area counts

Bromide

y = 2E-06x - 2.0752 R² = 0.9803

0

20

40

60

80

100

0.0E+00 1.0E+07 2.0E+07 3.0E+07 4.0E+07

Ion

Co

nce

ntr

atio

n (

mg/L

)

Area counts

Sulphate

188

Figure 10-13 Example of IC calibration curves for nitrite, nitrate and phosphate

y = 7.5E-07x - 1.2E-02 R² = 1.0E+00

0.0

0.2

0.4

0.6

0.8

1.0

1.2

0.0E+00 5.0E+05 1.0E+06 1.5E+06

Ion

co

nce

ntr

ation

(m

g/L

)

Area counts

Nitrite

y = 5.3E-07x + 1.4E-02 R² = 1.0E+00

0

2

4

6

8

10

12

0.0E+00 5.0E+06 1.0E+07 1.5E+07 2.0E+07

Ion

co

nce

ntr

atio

n (

mg/L

)

Area counts

Nitrate

y = 1.7E-06x + 1.2E-02 R² = 1.0E+00

0.0

0.2

0.4

0.6

0.8

1.0

1.2

0.0E+00 2.0E+05 4.0E+05 6.0E+05 8.0E+05

Ion

co

nce

ntr

ation

(m

g/L

)

Area counts

Phosphate

189

Table 10-4 IC method – Method detection limits for nitrite, nitrate, chloride, bromide, sulphate

and phosphate

Compound Spiked

Concentration

(mg/L)

%

Recovery RSD (%)

MDL

(mg/L)

Nitrite 0.1 95 4.1 0.012

Nitrate 1.0 101 1.9 0.057

Chloride 0.02 91 14 0.007

Bromide 0.02 99 4.2 0.003

Sulphate 1.0 90 10 0.28

Phosphate 0.05 131 10 0.02

References

Domino, M. M., Pepich, B.V., Munch, D.J., Fair, P.S. and Xie. Y. (2003) Method 552.3:

Determination of haloacetic acids and dalapon in drinking water by liquid-liquid

microextraction, derivatization, and gas chromatography with electron capture detection.

Revision 1.0. Technical Support Center, Office of Ground Water and Drinking Water,

U.S. Environmental Protection Agency. Cincinnati, Ohio 45268.

Munch, J. W., and Bassett, M. V. (2004) Method 521: Determination of nitrosamines in drinking

water by solid phase extraction and capillary column gas chromatography with large

volume injection and chemical ionization tandem mass spectrometry (MS/MS). Version

1.0. National Exposure Research Laboratory, Office of Research and Development, U.S.

Environmental Protection Agency, Cincinnati, Ohio 45268.

190

10.3 Free Chlorine 24-hour Residuals through the Duration of Metal

Coupon Conditioning

Figure 10-14 24 hour chlorine residuals monitored through the duration of metal coupon

conditioning

0

3

6

9

12

0 5 10 15 20HO

Cl 24h r

esid

ual (m

g/L

)

Time (d)

Initial dosage Fe Fe+ortho

0

3

6

9

12

0 5 10 15 20HO

Cl 24h r

esid

ual (m

g/L

)

Time (d)

Initial dosage Cu Cu+ortho

0

3

6

9

12

0 5 10 15 20HO

Cl 24h r

esid

ual (m

g/L

)

Time (d)

Initial dosage Pb Pb+ortho

191

10.4 HAA Speciation for Fresh Metal Coupons in Britannia Water

Figure 10-15 HAA speciation with time in the presence 1 mg/L orthophosphate for fresh metal

coupons with Britannia Water in one set of experiments, error bars indicate the measured


0

10

20

30

40

50

0 10 20 30

HA

A (

µg

/L)

Time (hours)

Fe

DCAA

TCAA

MCAA

0

10

20

30

40

50

0 20 40 60 80

HA

A (

µg

/L)

Time (hours)

Cu

0

10

20

30

40

50

0 20 40 60 80

HA

A (

µg

/L)

Time (hours)

Pb

0

10

20

30

40

50

0 20 40 60 80

HA

A (

µg

/L)

Time (hours)

Control

192

10.5 XRD Analysis for Fresh Iron Coupons

Figure 10-16 Comparison of XRD patterns for oxidized iron coupon in the presence of polyphosphate-

orthophosphate blends and polished iron coupon as control.

Po

ly-o

rth

o b

len

ds

10 20 30 40 50 60 70 80

Inte

nsity

Ca

rbo

n

Fe

Fe Fe C

on

tro

l

193

10.6 Metal Release Kinetics and Results for Metal Surface Analysis


corrosion inhibitors with NH2Cl in Mannheim Water; disinfectant concentrations, 12.3 mg/L;


0

1

2

3

0 20 40 60 80 100

Pb (

mg/L

)

Time (hours)

Pb

Pb+polyphosphate

Pb+orthophosphate

0

2

4

6

8

10

0 5 10 15 20 25 30

Fe (

mg/L

)

Time (hours)

NH2Cl

Fe

Fe+polyphosphate

Fe+orthophosphate

0

1

2

2

3

4

0 20 40 60 80 100

Cu (

mg/L

)

Time (hours)

Cu

Cu+polyphosphate

Cu+orthophosphate

194

Figure 10-18 Kinetics of metal release from fresh metal coupons with time in the presence and

absence of corrosion inhibitors in Britannia Water; disinfectant concentrations, 12.3 mg/L; error

bars indicate the measured maximum and minimum values (n=2)

0

4

8

12

16

0 10 20 30

Fe (

mg/L

)

Time (hours)

HOCl

Fe

Fe+poly

Fe+ortho

0

2

4

6

0 20 40 60 80

Cu(m

g/L

)

Time (hours)

Cu

Cu+poly

Cu+ortho

0

2

4

6

8

0 20 40 60 80

Pb (

mg/L

)

Time (hours)

Pb

Pb+poly

Pb+ortho

0

4

8

12

16

0 10 20 30

Fe (

mg/L

)

Time (hours)

NH2Cl

Fe

Fe+poly

Fe+ortho

0

2

3

5

6

0 20 40 60 80

Cu (

mg/L

)

Time (hours)

Cu

Cu+poly

Cu+ortho

0

2

4

6

8

0 20 40 60 80

Pb (

mg/L

)

Time (hours)

Pb

Pb+poly

Pb+ortho

195

Figure 10-19 Kinetics of metal release from corroded coupons in the absence/presence of

orthophosphate in Britannia Water, initial disinfectant concentrations 5.5 mg/L; error bars

indicate the measured maximum and minimum values (n=2)

0

2

4

6

0 10 20 30 40 50 60

Fe (

mg/L

)

Time (hours)

HOCl

Fe Fe+ortho

0.0

0.5

1.0

1.5

2.0

0 10 20 30 40 50 60

Cu (

mg/L

)

Time (hours)

Cu Cu+ortho

0

1

2

3

4

0 10 20 30 40 50 60

Pb (

mg/L

)

Time (hours)

Pb

Pb+ortho

0

2

4

6

0 10 20 30 40

Fe (

mg/L

)

Time (hours)

NH2Cl

Fe

Fe+ortho

0.0

0.5

1.0

1.5

2.0

0 10 20 30 40 50 60

Cu (

mg/L

)

Time (hours)

Cu Cu+ortho

0

1

2

3

4

0 10 20 30 40 50 60

Pb (

mg/L

)

Time (hours)

Pb Pb+ortho

196

Figure 10-20 Comparison of elemental distribution for iron, copper and lead coupons in the

absence and presence of orthophosphate with HOCl

7%

47%

6%

37%

3%

Fe+HOCl

Fe

O

Ca

C

Si 2p

15%

52%

3%

23%

5% 2%

Fe+ortho+HOCl

Fe

O

Ca

C

P

Si 2p

28%

0%

16%

53%

1% 2%

Cu+HOCl

C

Ca

Cu

O

Cl

S

11%

7%

10%

58%

11%

3%

Cu+ortho+HOCl

C

Ca

Cu

O

P

Cl

31%

22%

47%

Pb+HOCl

Pb

C

O

6%

13%

56%

6%

9%

10%

Pb+ortho+HOCl

Pb

C

O

Al

Ca

P

197

10.7 NDMA Formation in the Presence of Orthophosphate in Modified

Pipe Loops

Figure 10-21 NDMA formation in four pipe loops in the presence of orthophosphate under two

flow conditions

0

10

20

30

40

2 6 12

ND

MA

(ng/L

)

HRT (hours)

Turbulent

0

10

20

30

40

2 6 12

ND

MA

(ng/L

)

HRT (hours)

Laminar

198

10.8 Effects of Corrosion Inhibitors and the Extent of Metal Corrosion

on Monochloramine Degradation and NDMA Formation

This research investigated the impacts of corrosion inhibitors and metal corrosion on

monochlormaine decay in two water matrices. However, due to consistently low formation of

NDMA in these two water matrices, the effects of phosphate-based corrosion inhibitors on

NDMA formation kinetics could not be evaluated in the two tested water matrices (Waters A and

B) when in contact with iron, copper and lead. Although not being completed to the same extent

as Chapter 3 and other chapters, some preliminary results concerning NDMA formation from

pre-dosed DMA in Milli-Q water are provided in this section in order to give implications about

the impacts of corrosion inhibitors and metal materials on NDMA formation.

199

The objective of this section was to investigate the effects of pipe materials (ductile iron,

copper and lead) and phosphate-based corrosion inhibitors (orthophosphate and polyphosphate)

on NH2Cl degradation and NDMA formation by performing material-specific formation

potential (MS-FP) and material-specific simulated distribution system (MS-SDS) tests at bench

scale. It is hypothesized that the application of phosphate-based corrosion inhibitors which

significantly affect corrosion rates may, in turn, impact NH2Cl degradation and NDMA

formation.

Materials and Methods 10.8.1

10.8.1.1 Reagents and materials



hypochlorite (NaOCl, 6%, VWR) in Milli-Q water. The NH2Cl dosing solution was then

prepared by adding the chlorine dosing solution to a well-stirred ammonium chloride solution

(700 mg/L as N). The chlorine dosing solution and ammonium chloride solution were combined

at a Cl2/N molar ratio of 0.8:1 to achieve a NH2Cl solution. Then the NH2Cl solution was

equilibrated for at least 30 min before use. The concentration of NH2Cl was then measured by

the indophenol method with a Hach DR 2700 spectrophotometer, and it ranged between 1400

and 1600 mg/L. At least 90% of the added chlorine was converted to NH2Cl. The water in batch

reactors (one-liter capacity amber bottles with PTFE lined caps) was dosed with preformed

NH2Cl.

Due to the consistently low formation of NDMA in two selected water matrices and to

ensure NDMA formation at a measurable level, one NDMA precursor, dimethylamine (DMA),

was spiked into Milli-Q water at a concentration of 1 µg/L to evaluate the impacts of corrosion

inhibitors and metal corrosion on NDMA formation.

The dosing solution of orthophosphate (Na3PO4) or polyphosphate ([Na(PO2)]6) was

prepared at a concentration of 500 mg/L as P and kept in dark at 4ºC. The targeted dosages for

each of corrosion inhibitors in test solution was 1 mg/L as P. Test coupons of ductile iron, copper

and lead were purchased from Metal Samples Co., Alabama, US. The size of these coupons is

1 2”3”1 1 ”. Unchlorinated post-filtration water was collected from two water treatment

200

plants in Ontario (Mannheim Water and Britannia Water). Water quality parameters are listed in

Table 10-5.

Table 10-5 Water quality parameters for post-filtration water

Parameters Values

Mannheim Water Britannia Water

pH 7.5 ± 0.2 6.6±0.5


UV254 (cm-1

) 0.058±0.008 0.065±0.005

TOC (mg/L) 4.5 ± 0.3 3.3±0.2

SUVA(cm-1

·L/mg) 0.013±0.002 0.019±0.0002

Bromide (μg L) 65.0±15.5 15.8±0.2

Chloride (mg/L) 84.5 ± 2.5 3.1±0.6

Sulfate (mg/L) 35.0 ± 2.0 23.8±0.8

Cl-:SO4

2- ratio 2.4 ± 0.1 0.13±0.04

10.8.1.2 Experimental procedures

Experiments were performed with both fresh and pre-corroded metal coupons. Fresh

metal coupons were prepared before each batch of experiments by removing any corrosion

products from the coupons by polishing with 60-grit sandpaper followed by 120-grit sandpaper,

and then rinsing with deionized water and acetone followed by Milli-Q water. These polished

coupons simulated new pipes without any impacts from service age. The corroded metal coupons

were prepared by soaking them in tap water in the absence and presence of orthophosphate for at

least two weeks to allow corrosion products to be built up on the surface (details are provided in

Section 3.2.1). Orthophosphate was selected since it is in common use and it can effectively

control metal release in a short term for fresh metal coupons, especially for copper and lead.

“Material-specific” simulated distribution system or MS-SDS tests incorporated the

influences of the distribution system pipe wall on disinfectant residual stability and DBP

generation. Details of MS-SDS procedures have been described by Brereton and Mavinic (2002).

They consist of incubating metal coupons in water samples under conditions representative of

actual field conditions in terms of reaction time, pH, temperature, and disinfectant application. In

experiments with corroded metal coupons, MS-SDS tests were performed and initial NH2Cl

concentration was approximately 5.5 mg/L. In experiments with fresh metal coupons, material-

specific formation potential (MS-FP) tests were performed similarly to MS-SDS tests, except

that MS-FP tests applied a higher concentration of NH2Cl (12.3 mg/L) than would be

201

encountered during typical water treatment. This was done to meet the high chlorine demand of

fresh metal coupons and their initial corrosion products and ensure detectable NH2Cl residuals

after 24 hours. All of the MS-SDS and MS-FP experiments were performed in 1 L amber bottles

(chlorine-demand free) with PTFE lined caps and at room temperature (21±2 ºC). No pH

adjustment was performed, and thus reactions of metal corrosion and NH2Cl degradation in the

two water matrices proceeded under their respective ambient conditions. All of the metal

coupons were suspended in amber bottles using nylon threads. To maximize the contact of water

with coupon surface, a Big Bill Orbital shaker was used (Barnstead International) to maintain a

gentle mixing at a speed of 25 rpm.



control samples, which were prepared similarly to test samples but without metal coupons. To

account for possible differences in NDMA yields due to the different pH conditions of the water

that were tested, a small number of additional experiments was performed, in which the pH was

controlled at 8.3±0.2 using 1 mM borate buffer in these tests, and the initial NH2Cl concentration

was set 14.5±0.5 mg/L as Cl2. Analytical methods that were employed to examine water quality

and the analytes of interest are summarized in Table 10-6.



TOC mg/L OI-Analytical TOC analyzer SM1 5310 C

pH pH meter

UVA254 cm-1

Hewlett Packard 8452A Diode

Array UV spectrophotometer

SM 5910B

Alkalinity mg/L Titration SM 2320B

NH2Cl mg/L Hach DR2700 Spectrophotometer Hach method 101712

Anions µg/L Dionex DX-300 Series Ion

Chromatography System

SM 4110 B

NDMA ng/L Varian 4000 GC-MS USEPA 5213

Metal (Fe, Cu and Pb) mg/L Varian SpectrAA.20 SM 3111 B

Notes: 1. SM represents Standard Methods for the Examination of Water and Wastewater (APHA, AWWA, WEF, 2005); 2.

Hach, 2007; 3. Munch and Bassett, 2004.

202

Monochloramine Degradation 10.8.2

10.8.2.1 Fresh Coupons in Two Water Matrices

In the two investigated water matrices, NH2Cl degradation generally followed first-order

kinetics. The overall decay rate constants of NH2Cl (k) were obtained by fitting a pseudo-first-

order decay equation to the measured NH2Cl residual concentrations from each experiment. k is

usually considered as the sum of a first-order bulk decay constant (kb) and a first-order wall

decay constant (kw) (Rossman et al., 1994). In the present tests, the bottles had been made

chlorine demand free prior to testing, so the walls of the bottles did not contribute significantly to

the overall chlorine demand and the bulk water reaction in the MS-PF tests could be considered

to be the same as would occur in control samples without metal coupons. Thus in these tests, the

wall decay constant was the difference between the overall decay constant and the bulk decay

constant obtained from the control samples. Table 10-7 summarizes NH2Cl overall decay

constants (k) and the kb values for the three tested metal materials in the absence and presence of

corrosion inhibitors. An overview of the results for Mannheim Water, for example, is provided in

Figure 10-22 to illustrate the effects of metal materials and corrosion inhibitors on NH2Cl

degradation in Mannheim Water (as determined by the difference in bar size relative to their

respective control samples) in addition to showing the general bulk water decay (represented by

the control samples themselves).

Table 10-7 Comparison of NH2Cl decay constants for fresh coupons in two water matrices (n=4)

Material Corrosion

inhibitors Mannheim Water

Britannia

Water

k

Iron

None 0.1298 ±0.0073 0.1460±0.0438

Orthophosphate 0.1100 ±0.0151 0.1802±0.0423

Polyphosphate 0.2092 ±0.1099 0.2844±0.0021

Copper

None 0.0150 ±0.0086 0.0509±0.0098

Orthophosphate 0.0093 ±0.0016 0.0144±0.0003

Polyphosphate 0.0131 ±0.0062 0.0259±0.0054

Lead

None 0.0075 ±0.0012 0.0110±0.0006

Orthophosphate 0.0066 ±0.0003 0.0097±0.0015

Polyphosphate 0.0094 ±0.0037 0.0122±0.0008

kb Water

None 0.0081 ±0.0008 0.0159±0.0018

Orthophosphate 0.0074 ±0.0008 0.0122±0.0001

Polyphosphate 0.0105 ±0.0032 0.0141±0.0003

Initial NH2Cl concentration: 12.3 mg/L as Cl2

203

Figure 10-22 NH2Cl overall decay constants for fresh coupons with Mannheim Water; NH2Cl

12.3 mg/L, error bars indicate the measured maximum and minimum values (n=2)

For each metal material, a single-factor ANOVA test at a confidence level of 95% was

applied to the overall decay constants to determine whether the treatment factor (corrosion

inhibitors) had a significant influence on the response factor (NH2Cl decay constants). When the

ANOVA test signified statistically significant impacts of corrosion inhibitors on NH2Cl decay, a

Fisher’s Least Significant Difference (LSD) test was further applied to determine if significant

differences existed between each pair of treatments at a 95% of confidence level (Montgomery,

2000).

As shown in Table 10-7 and Figure 10-22, kb for Mannheim Water was not significantly

affected by corrosion inhibitors. This is consistent with the previous finding that corrosion

inhibitors did not significantly impact disinfectant bulk water degradation (Zhang and Andrew,

2012). Small variations in kb values also show good QA/QC between experiments. The most

striking feature of the data shown in Figure 10-22 is that iron was more much reactive with

NH2Cl than copper and lead. NH2Cl decay constants in the presence of iron coupons were two

orders of magnitude higher than the bulk water reaction constants irrespective of the types of

corrosion inhibitors present. Therefore, pipe wall reactions dominated NH2Cl consumption in the

presence of iron coupons (i.e., reactions with the coupon surface or with iron released from the

coupons). Corrosion inhibitors did not exhibit significant impacts on NH2Cl decay for iron

coupons (p value >0.05, the ANOVA test).

0.00

0.05

0.10

0.15

No inhibitors Orthophosphate Polyphosphate

1st o

rde

r d

eca

y c

on

sta

nt (h

-1)

Fe Cu Pb Control

0.21±0.08

kb

kw k

204

The NH2Cl overall decay rates for fresh copper coupons were 11~ 80% higher than those

in the bulk water (Figure 10-22), indicating that the contributions of the wall reactions and the

bulk reactions to NH2Cl overall decay were on a similar scale. Results of the single factor

ANOVA tests demonstrate that the addition of phosphate-based corrosion inhibitors did not

significantly impact NH2Cl degradation in the presence of copper coupons (p value >0.05).

Interestingly, the overall NH2Cl decay constants for lead coupons were similar to those for the

bulk water irrespective of the types of corrosion inhibitors. As a result, wall decay constants of

lead coupons were calculated to be negligible. Corrosion inhibitors did not significantly affect

NH2Cl decay for lead coupons as well (p value >0.05).

In order to help differentiate between effects due to reactions with the coupons directly

and effects due to reactions with ions released from the coupons, dissolved metal concentrations

at 24 hours were measured and presented in Figure 10-23. The released metal concentrations

were much higher for the iron coupons than for the lead and copper coupons, suggesting that the

higher NH2Cl decay rate for test solutions containing iron coupons could be attributed to

reactions with iron that was released from the coupons. Comparable iron concentration at 24

hours in the absence and presence of corrosion inhibitors (Figure 10-23) indicates that

phosphate-based corrosion inhibitors had no significant impacts on iron corrosion in Mannheim

Water. The variability in the decay constants for iron coupons in the presence of polyphosphate

Figure 10-23 Released metal concentrations at 24 hours for fresh metal coupons in Mannheim

Water; NH2Cl 12.3 mg/L; error bars indicate the measured maximum and minimum values (n=2)

0

2

4

6

8


Me

tal co

nce

ntr

ation

(m

g/L

)

Fe Cu Pb

205

was large (between 0.13 h-1

and 0.29 h-1

, Figure 10-22), likely due to different polishing

conditions of the iron coupon surface between batches of experiments.

For copper coupons, orthophosphate effectively decreased the released copper

concentration from 1.5 mg/L in the absence of corrosion inhibitors to 0.7 mg/L, and reduced the

released lead concentration from 0.7 mg/L in the absence of corrosion inhibitors to 0.2 mg/L. For

copper and lead coupons in contact with NH2Cl over short period oxidation, Cu2O and divalent

lead solids have been reported as primary corrosion products (Vasquez et al., 2006; Xiao et al.,

2007). As such, the decrease in copper and lead concentrations due to the addition of

orthophosphate can stoichiometrically reduce the consumption of NH2Cl by 0.9 mg/L for copper

and 0.3 mg/L for lead. These two values only account for 4% and 0.8% of the initial NH2Cl

concentration applied (12.3 mg/L) for copper and lead, respectively. As a result, the slight

decrease in NH2Cl consumption due to the reduced copper and lead corrosion rates by

orthophosphate would not be expected to result in a significant difference in NH2Cl decay

relative to that without the addition of orthophosphate. Therefore, it was concluded that released

copper and lead did not contribute significantly to NH2Cl decay in this water matrix.

The relative sequence of reactivity of the three metal materials with NH2Cl in Britannia

Water had a similar pattern to that observed in Mannheim Water (Table 10-7) except that

phosphate-based corrosion inhibitors significantly affected NH2Cl decay for fresh iron and

copper coupons (p values of 0.04 and 0.02, respectively). Results of the LSD tests indicate that

polyphosphate statistically increased NH2Cl decay for iron coupons, while both polyphosphate

and orthophosphate statistically decreased NH2Cl degradation for copper coupons. Figure 10-24

displays dissolved iron and copper concentrations at 24 hours for Britannia Water. For iron

coupons, released iron concentrations were significantly increased after the addition of

phosphate-based corrosion inhibitors likely due to the formation of iron-phosphate complexes

and thus the increased iron solubility (McNeill and Edwards, 2000). Increased NH2Cl

degradation with increases in released iron concentrations may also suggest that the reactions

with released iron from the coupons were a primary reason for the accelerated NH2Cl decay for

iron coupons in Britannia Water. For copper coupons, results of stoichoimetrical calculation

indicate that the significant decrease in NH2Cl decay after the addition of phosphate-based

corrosion inhibitors was primarily due to the effective reduction of released copper

concentrations.

206

Figure 10-24 Released iron and copper concentrations at 24 hours in the absence and presence of

corrosion inhibitors for fresh metal coupons in Britannia Water; NH2Cl 12.3 mg/L; error bars


In addition, as shown in Table 10-7, NH2Cl degradation constants in Britannia Water in

contact with copper and lead as well as in bulk phase were significantly higher than those in

Mannheim Water (p-values 0.003, 0.02 and 0.0009, respectively), but were only statistically

different for fresh iron coupons in two water matrices at a confidence level of 85%. Edwards and

Dudi, 2004) have attributed the relatively low confidence levels in statistical analysis for iron to

low number of samples collected, which may also be applied to this study. Since the pH and

alkalinity values in Britannia Water were lower than those in Mannheim Water (Table 10-5), the

metal coupons were more subject to corrosion in Britannia Water which in turn would induce

more NH2Cl to participate into reactions with metal coupons. The details about the impacts of

water quality on metal corrosion are provided in Chapter 4. The accelerated auto-decomposition

of NH2Cl in the bulk Britannia Water at low pH may also be responsible for the increased NH2Cl

degradation than in Mannheim Water.

10.8.2.2 Corroded Coupons in Two Water Matrices

For corroded metal coupons, NH2Cl degradation also followed first-order kinetics. Table

10-8 summarizes NH2Cl overall decay constants (k) and the kb values for the three tested metal

materials in the absence and presence of orthophosphate. For both water matrices, the reactivity

of three metal materials with NH2Cl followed the sequence of Fe>CuPb. Only for corroded iron

0

3

6

9

12


Me

tal co

nce

ntr

ation

(m

g/L

)

Fe Cu

207

coupons, NH2Cl degradation was dominated by wall reactions, while wall decay constants of

corroded copper and lead coupons were calculated to be negligible. In addition, orthophosphate

significantly increased NH2Cl degradation for the corroded iron coupons (ANOVA tests, p-value

of 0.04) with reasons as explained in Section 10.7.2.1, but exhibited no statistical impacts for the

corroded copper and lead coupons (p-values >0.05).

Table 10-8 Comparison of NH2Cl overall decay constants for corroded coupons between

Mannheim Water and Britannia Water (n=4)

Material Corrosion

inhibitors

Mannheim

Water

Britannia

Water

k

Iron None 0.0495±0.0071 0.0733±0.0158

Orthophosphate 0.0781±0.0037 0.1447±0.0142

Copper None 0.0114±0.0013 0.0186±0.0001

Orthophosphate 0.0065±0.0040 0.0173±0.0011

Lead None 0.0084±0.0011 0.0168±0.0013

Orthophosphate 0.0057±0.0009 0.0156±0.0006

kb Water None 0.0083±0.0010 0.0182±0.0008

Since Britannia Water had lower pH and alkalinity than Mannheim Water (Table 10-5), it

was expected to be more corrosive to metal coupons than Mannheim Water. Therefore, more

NH2Cl should have participated in the metal corrosion reactions, leading to more rapid NH2Cl

degraded in Britannia Water than in Mannheim Water. As shown in Table 10-8, corroded iron

coupons in contact with Britannia Water had k values of at least 48% higher than with Mannheim

Water, and the wall decay constants (kw, the difference between k and kb) at least 34% higher.

Therefore, NH2Cl decay for corroded iron coupons exhibited the expected trend. However, for

copper and lead, although NH2Cl decay constants in Britannia Water were significantly higher

than those in Mannheim Water with all of the p-values less than 0.05, the consumption of NH2Cl

by coupon surfaces (indicated by kw values) was calculated to be negligible. As such, the

increased NH2Cl overall decay in Britannia Water for both copper and lead coupons was

primarily attributed to its significantly increased bulk phase reaction rates (p-value of 0.008),

mostly likely via auto-decomposition.

NDMA Formation from DMA for Fresh Coupons 10.8.3

In addition to NH2Cl degradation, this section also investigated interactive impacts of

metal corrosion and corrosion inhibitors on NDMA formation. As discussed in Section 10.7.2.2,

208

orthophosphate did not significantly affect NH2Cl degradation for corroded copper and lead

coupons. Since NH2Cl concentration is a rate-limiting factor of NDMA formation, it was

postulated that orthophosphate would not affect NDMA formation for the corroded metal

coupons as well. Therefore, only fresh metal coupons were employed in this test to evaluate the

interactive impacts of metal corrosion and corrosion inhibitors on NDMA formation. In addition,

since corrosion inhibitors did not significantly affect NH2Cl degradation in the bulk water, only

NDMA formation in Milli-Q water in the absence of metal coupons and corrosion inhibitors was

considered as a control for comparison purposes.

Figure 10-25 compares NDMA formation at 24 hours for each combination of metal

materials and corrosion inhibitors. NH2Cl residuals at 24 hours are also demonstrated in Figure

10-25. Compared with bulk water control, Milli-Q water in contact with iron coupons had

significantly higher yields of NDMA irrespective of the types of corrosion inhibitors. In

particular, iron coupons exposed to polyphosphate had a NDMA yield at least seven folds as

high as that in bulk water control. NH2Cl residual at 24 hours for iron coupons in the presence of

polyphosphate dramatically decreased to 0.2 mg/L, but there were still 3.1 mg/L and 3.6 mg/L

NH2Cl remaining at 24 hours for iron coupons alone and in the presence of orthophosphate,

respectively.

Figure 10-25 NDMA formation from DMA as a result interactions of metal coupons, corrosion

inhibitors and NH2Cl. Initial NH2Cl 14.5±0.5 mg/L, DMA 1 µM/L, pH 8.3, in Milli-Q, error

bars indicate standard deviation (n=3)

0

4

8

12

16

0

600

1200

1800

2400

3000

Polyphosphate No inhibitors Orthophosphate

NH

2C

l (m

g/L

)

ND

MA

(n

g/L

)

Fe-NDMA Cu-NDMA Pb-NDMAWater-NDMA Fe-NH2Cl Cu-NH2ClPb-NH2Cl Water-NH2Cl

209

NDMA formation at 24 hours for copper coupons alone, and copper coupons treated with

polyphosphate and orthophosphate also increased compared with that in bulk water control.

However, the magnitude of the increase was not as significant as that for iron coupons. The

addition of orthophosphate decreased NH2Cl degradation, and 10.1 mg/L NH2Cl remained after

24 hours. However, there were only 8.2 mg/L NH2Cl left for copper coupons alone and 8.5 mg/L

for copper coupons in the presence of polyphosphate after 24 hours.

Lead coupons had comparable 24-hour NH2Cl residual concentration and NDMA

formation with bulk water control regardless of the presence of corrosion inhibitors. As

discussed in Section 10.7.2.1, NH2Cl wall reactions had little contribution to overall NH2Cl

decay for fresh lead coupons. It follows that the consumption of NH2Cl due to metal surface

reaction may not significantly affect the residual concentration of NH2Cl in reaction with DMA

in aqueous solution. As such, it is expected to observe comparable NDMA formation in the

water containing lead coupons and bulk water control.

To statistically evaluate the interactive effects of metal corrosion and corrosion inhibitors

on NDMA formation, and to determine whether the treatment factor (corrosion inhibitors) had a

significant influence on the response factor (NDMA concentrations), a single-factor ANOVA

test at a confidence level of 95% was applied to the results that were obtained for each metal

material combined with different corrosion inhibitors. When the ANOVA test results signified

significant difference in NDMA formation between metal coupons and bulk water control, each

treatment was further compared with bulk water control by using the LSD tests for single

contrasts at a confidence level of 95% (Montgomery, 2000). Results of the single contrasts using

the LSD test are summarized in Table 10-9.

Table 10-9 Summary of the results for the single contrasts using the LSD test

Comparison Significantly different NDMA formation

No inhibitor Polyphosphate Orthophosphate

Milli-Q Fe Yes Yes Yes

Milli-Q Cu Yes Yes No

Note: a statistically significant at a confidence level of 95%

The ANOVA test results confirmed that NDMA formation for iron and copper coupons

in the absence and presence of corrosion inhibitors was significantly different from that in bulk

water control. The results of single contrasts using the LSD tests for iron coupons indicate that

210

iron coupons significantly increased NDMA formation, regardless of the presence of and/or

types of corrosion inhibitors. According to NH2Cl degradation studies in Section 10.7.2.1, wall

reactions primarily due to iron corrosion dominated NH2Cl decay, leading to less availability of

NH2Cl residual to react with the constituents in the bulk phase. Since NH2Cl concentration is a

rate-limiting factor of NDMA formation, it follows that lower concentrations of NDMA should

have formed in the water in contact with iron coupons than in the bulk water control. The

unexpected elevation in NDMA formation in the water containing iron coupons suggests that

iron and/or iron corrosion products may catalyze the transformation of DMA to NDMA. To

identify the possible roles of iron corrosion products in NDMA formation, NDMA concentration

was further plotted versus released total iron concentration. As shown in Figure 10-26, NDMA

yields increased with increases in released total iron concentrations. It indicates that iron

catalysis during NDMA formation was positively affected by released iron concentration. It can

also be observed in Figure 10-26 that corrosion inhibitors significantly affected NDMA

formation primarily by affecting released iron concentration.

Figure 10-26 Total iron concentrations and NDMA formation for iron coupons in the absence

and presence of corrosion inhibitors, error bars indicate standard deviation (n=3)

In terms of copper coupons, the LSD comparison results indicate that NDMA formation

significantly increased for copper coupons alone and treated with polyphosphate compared with

that in bulk water control. However, there was no statistical difference in NDMA yields between

copper coupons exposed to orthophosphate and bulk water control. It was also observed that

0

2

4

6

8

10

0

600

1200

1800

2400

3000


To

tal iro

n (

mg

/L)

ND

MA

(n

g/L

)

NDMA Total iron

211

NDMA concentration positively increased with increases in released Cu(II) concentration

(Figure 10-27). Therefore, it was hypothesized that copper can catalyze NDMA formation with

the extent of catalysis being affected by copper concentration. Further experiments have been

performed to test this hypothesis (Chapter 6). Results of these experiments confirmed copper

catalysis during NDMA formation, and that copper catalysis nonlinearly increased with

increasing copper concentration. Although 0.5 mg/L copper was released from copper coupons

treated with orthophosphate, its catalytic effect on NDMA formation may not be significant

under current experimental conditions so that comparable NDMA yields were observed for

copper coupons in the presence of orthophosphate and bulk water control.

Figure 10-27 Cu(II) concentrations and NDMA formation for copper coupons in the absence and

presence of corrosion inhibitors, error bars indicate standard deviation (n=3)

This study provides important implications about factors affecting NH2Cl stability and

NDMA formation in distribution mains and household plumbing. Phosphate-based corrosion

inhibitors have been increasingly used for corrosion control. However, these corrosion inhibitors

may exhibit either beneficial or detrimental effects on NH2Cl stability and NDMA formation,

depending on metal types, age and water quality. Therefore, their impacts on NH2Cl decay and

NDMA formation due to the interactions with metal surfaces should be considered site-

specifically.

0.0

0.3

0.6

0.9

1.2

0

200

400

600

800

1000


Cu

(II

) (m

g/L

)

ND

MA

(n

g/L

)

NDMA Cu(II)

212

References 10.8.4










California Department of Public Health (CDPH) (2009). California Drinking Water: NDMA and

Other Nitrosamines - Drinking Water Issues.

http://www.cdph.ca.gov/certlic/drinkingwater/ Pages/NDMA.aspx (accessed September

15, 2011).












Duirk, S.E., Gombert, B., Croue, J.P. and Valentine, R.L. (2005) Modeling NH2Cl loss in the

presence of natural organic matter. Water Research, 39(14), 3418-3431.


213



Facey, R. M., and Smith, D. W. (1995) Soft, low-temperature water-distribution corrosion:


Gerecke, A.C. and Sedlak, D.L. (2003) Precursors of N-nitrosodimethylamine in natural waters.








Lu, W., Kiene, L., and Levi, Y. (1999) Chlorine demand of biofilms in water distribution






313.

Ministry of the Environment (MOE). Safe Drinking Water Act 2002. Ontario Regulation 169/03,

Schedule 2. http://www.search.e-laws.gov.on.ca/en/isysquery/9426015e-057d-4983-

85d4-5e751801b608/2/doc/?search=browseStatutes&context=#hit1 (accessed September

15, 2011).

Mitch, W.A. and Sedlak, D.L. (2002) Formation of N-nitrosodimethylamine (NDMA) from


595.

214


Sons, New York.

Munch, J. W., and Bassett, M. V. (2004) Method521: Determination of nitrosamines in drinking




Environmental Protection Agency, Cincinnati, Ohio 45268, 2004.







Rossman, L. A., Clark, R. M., and Grayman, W. M. (1994) Modeling chlorine residuals in

drinking-water distribution-systems. Journal of Environmental Engineering-Asce, 120(4),

803-820.





Association, 98(2), 144-154.

Vikesland, P., Valentine, R. and Ozekin, K. (1996). Application of product studies in the

eludication of chloramine reaction pathways. Water Disinfection and Natural Organic

Matter. R. Miner and G. Amy (eds.). American Chemical Society, Washington DC, pp.

115–125.

Vikesland, P. J., Ozekin, K., and Valentine, R. L. (2001) NH2Cl decay in model and distribution

system waters. Water Research, 35(7), 1766-1776.

215



Westerhoff, P. and Mash, H. (2002) Dissolved organic nitrogen in drinking water supplies: a


Wilczak, A., Hoover, L. L., and Lai, H. H. (2003a) Effects of treatment changer on chloramine


Wilczak, A., Assadi-Rad, A., Lai, H. H., Hoover, L. L., Smith, J. F., Berger, R., Rodigari, F.,

Beland, J. W., Lazzelle, L. J., Kinicannon, E. G., Baker, H., and Heaney, C. T. (2003b)




quality problems. Water Supply, 8 (1 and 2), 195-198.

Woolschlagar, J., Rittmann, B., Piriou, P., Kiene, L., and Schwartz, B. (2001) Using a

comprehensive model to identify the major mechanisms of chloramine decay in

distribution systems. Water Science & Technology: Water Supply, 1(4), 103-110.


release in distribution systems. Journal American Water Works Association, 99(1), 78-

88.

Zhang, H., Andrews, S. A. (2012) Effects of phosphate-based corrosion inhibitors on the kinetics

of chlorine degradation and HAA formation in contact with three metal materials.

Canadian Journal of Civil Engineering, 39, 44-54.

216

10.9 Degradation Potential of Iron, Lead and Their Corrosion Products

on HAA9

This research investigated the degradation potential of iron and lead on HAA9 in Milli-Q

water under controlled experimental conditions. However, no further experiments have been

conducted to evaluate the impacts of water quality of the two tested water matrices (Waters A

and B) on HAA degradation by iron and copper. Although not being completed to the same

extent as other chapters, some results of this section which deal with iron support the discussions

in Section 3.3.3, Chapter 3. The results that deal with lead degradation potential also help

understand of the fate of HAA compounds in lead pipes.

217

The purpose of this section was to investigate the degradation potential of iron and lead

as well as their corrosion products on HAA9 under controlled experimental conditions. Since iron

is widely used in water mains and lead may still be present in old lead service lines, soldered

joints and brass plumbing fittings, the knowledge about the degradation potential of iron and lead

for HAA9 is important to understand the fate of these HAA compounds in water distribution

mains and domestic plumbing systems.

Materials and Methods 10.9.1

All chemicals used in this study were ACS grade or higher. Test coupons of ductile iron

and lead were purchased from Metal Samples Co., Alabama, US. The size of these coupons was

1 2”3”1 1 ”. For the tests investigating the degradation potential of iron and its corrosion

products on HAAs, iron coupons were conditioned in the absence and presence of different

corrosion inhibitors in tap water for 24 hours to simulate the scenario when pipes have been

corroded over a short period. Initial chlorine concentration for conditioning was 12.3 mg/L.

Corrosion inhibitors that were applied were orthophosphate 1.0 mg/L as P, polyphosphate 1.0

mg/L as P, and orthophosphate-polyphosphate blends (0.5/0.5 mg/L as P). Borate buffer

(H3BO3/NaOH) solutions were also prepared to control the solution pH at desired levels

(8.3±0.2). In the tests investigating the degradation potential of lead, fresh lead coupons were

used to simulate the contact of DBPs with metal surface without any impacts from service age.

All of the tested metal coupons were suspended in 1 L amber bottles using nylon threads. To

maximize the contact of water with coupon surface, a Big Bill Orbital shaker was used

(Barnstead International) to maintain a gentle mixing at a speed of 25 rpm.

The degradation potential of lead corrosion products, including PbO, Pb(OH)2(CO3)2, and

PbO2 (Sigma Aldrich), on HAAs were also investigated. Each solid corrosion product was

spiked into 1 L Milli-Q water in the form of powder at a concentration of 1 g/L, and then mixed

with Milli-Q water thoroughly by a stir bar. After each prescribed reaction time, 50 mL water

was taken from reaction bottles and filtered through 0.2 µm Nylaflo® Nylon membrane filter

paper (Pall Corporation) to separate powdered corrosion products from the aqueous solution.

All of the experiments were performed in duplicate along with control bottles, which

were prepared in a similar manner except that no coupons were soaked or no corrosion products

218

were added. To account for any influence from filter paper and filtration performance on the

concentration of HAA9, control samples for the experiments employing powdered lead corrosion

products were also filtered under similar conditions and its HAA9 concentrations after filtration

were reported as Control in the results section.

In addition, X-ray diffraction (XRD) analysis was applied to the surface of iron coupons

after 24-hour conditioning to characterize the compositions of iron corrosion products after a

short-term reaction and to determine their possible roles during the degradation of HAA9. A

Siemens D5000 Diffractometer System operating at 50 kV/35 mA was used to collect the

diffraction patterns. A high-power, line focus Cu-K-source was used combined with a solid

state Kevex detector for elimination of K-lines. The experimental data were collected on a step

scan mode (0.02° /1.5 sec) within the most informative range (2-theta degrees). The obtained

data were processed by various Diffrac Plus software packages including Eva 8.0 and

Topas v. 2.1. For all of the tested iron coupons, due to the formation of non-uniform corrosion

layers on the coupon surface as a result of short-term reactions, XRD analysis was carried out

on powders which were scratched from the iron coupon surface.

Nine species of haloacetic acid (HAA9) were analyzed according to USEPA Standard

Method 6251 B (APHA et al., 2005). The principle was based on liquid–liquid extraction of

HAA9 with methyl-tert-butyl-ether (MtBE) at an acidic pH followed by diazomethane

derivatization and gas chromatography with electron capture detection (GC/ECD) analysis. The

surrogate standard was 2,3,5,6-tetrafluorobenzoic acid. All of the extracted and derivatized

samples were stored at -10 °C or less, and analyzed within 21 days of extraction. The gas

chromatograph used for this analysis was a Hewlett Packard 5890 Series II Plus Gas

Chromatograph. The GC column was DB1701 (30 m х 0.2 mm х 0.2 μm). The column

temperature programs for HAAs measurement were: hold at 35 °C for 10 min; ramp to 65 °C at

2.5 °C/min; ramp to 85 °C at 10 °C/min; ramp to 205 °C at 20 °C/min and hold for 7 min.

HAA9 Degradation by Corroded Iron Coupons 10.9.2

The degradation potential of iron was investigated by employing pre-corroded iron

coupons, and the variation of HAA9 concentrations with reaction time is shown in

219

Figure 10-28 Compared with the control bottles where no loss of HAA9 was observed,

HAA9 were rapidly degraded in the water containing corroded iron coupons. The HAA9

degradation kinetics followed the sequence of decreasing order: poly/ortho blends,

polyphosphate orthophosphate, and no corrosion inhibitor.

Figure 10-28 Reduction of HAA9 by corroded iron coupons in the absence and presence of

different corrosion inhibitors. Milli-Q water, error bars indicate the measured maximum and

minimum values (n=2)

The degradation of each species of HAA9 except for MCAA in the presence of iron

coupons treated with poly/ortho blends is plotted in Figure 10-29. All of the eight HAA

compounds were readily degraded by corroded iron coupons. DCAA degraded relatively slowly

Figure 10-29 HAA degradation by corroded iron coupons treated with poly/ortho blends, Milli-Q

water, error bars indicate the measured maximum and minimum values (n=2)

0

100

200

300

400

500

0 20 40 60 80 100

HA

A9 (

µg

/L)

Time (hours) Fe Fe+polyFe+ortho Fe+poly/orthoControl

0

10

20

30

40

50

0 20 40 60 80 100

HA

A (

µg

/L)

Time (hours)

MBAA DCAA

TCAA BCAA

DBAA BDCAA

CDBAA TBAA

220

in the earlier 24 hours. For MCAA, as shown in Figure 10-30, its concentrations increased with

increasing reaction time in the initial 24 hours, and then followed a decreasing trend after 24

hours.

Figure 10-30 Reduction of MCAA by corroded iron coupons in the absence and presence of

different corrosion inhibitors. Milli-Q water, error bars indicate the measured maximum and

minimum values (n=2)

Hozalski (2001) has proposed sequential hydrogenolysis to be a primary degradation

mechanism for HAA compounds in contact with Fe0. To identify if Fe

0 was present on the

surface of the pre-conditioned iron coupons, XRD analysis was performed, and the results are

shown in Figure 10-31. A significant amount of Fe0 (represented by black peak at 45 degree) was

present on the corroded iron surface even after 24 hour oxidation. Due to the presence of Fe0 on


poly/ortho-phosphate blends and polished iron coupon as control. Red: iron coupon as control;

Black: non-scratched iron coupon with poly/ortho-phosphate blends

0

20

40

60

80

100

0 20 40 60 80 100

MC

AA

(µ

g/L

)

Time (hours)

Fe Fe+polyFe+ortho Fe+poly/orthoControl

Inte

nsity

10 20 30 40 50 60 70 80

2-theta (degree)

Fe

Fe Fe Fe

+poly

/ort

ho

Fe

contr

ol

221

the surface of freshly oxidized coupons, therefore, it was hypothesized that sequential

hydrogenolysis was the reaction pathways for HAA degradation in the presence of these

corroded iron coupons.

To test this hypothesis, single species degradation experiments were performed regarding

MCAA, DCAA, TCAA, and BCAA due to their wide presence in the chlorinated water. The

results of the degradation experiments for four compounds are collectively illustrated in Figure

10-32. The degradation products, including HAA compounds and anions (e.g. chloride and

bromide) were also analyzed. However, the concentrations of chloride and bromide were below

method detection limits (MDLs) of these anions by ion chromatography, and thus not reported

herein.

Figure 10-32 HAA speciation degradation by corroded iron coupons treated with poly/ortho

blends, error bars indicate the measured maximum and minimum values (n=2)

As shown in Figure 10-32, all of the parent HAA compounds experienced degradation in

the presence of these corroded iron coupons over the reaction period, whereas their

0

15

30

45

60

0 20 40 60 80 100 120

HA

A (

µg

/L)

Time (hours)

MCAA MCAA Control

0

10

20

30

40

0 20 40 60 80 100 120

HA

A (

µg

/L)

Time (hours)

DCAA

MCAA

DCAA control

0

10

20

30

40

0 20 40 60 80 100 120

HA

A (

µg

/L)

Time (hours)

TCAA DCAA

MCAA Control

0

10

20

30

40

0 20 40 60 80 100 120

HA

A (

µg

/L)

Time (hours)

MCAA

BCAA

BCAA control

222

concentrations in the control samples remained constant. For example, TCAA was reduced

within 72 hours with DCAA and MCAA being produced. DCAA concentrations were

consistently below 2 µg/L, and MCAA concentrations were initially increased to 7.6 µg/L at 44

hours and then dropped to 2 µg/L after 120 hours. The plots of TCAA degradation with reaction

time illustrate the destruction of the parent compound, the production of the reaction

intermediates (e.g. DCAA), and the degradation of final products (e.g. MCAA). Therefore, the

results of single species experiments suggest that HAA degradation by corroded iron coupons,

most likely due to the presence of Fe0, also followed the mechanism of sequential

hydrogenolysis. The degradation of these HAA compounds can be illustrated as follows

(Hozalski, 2001):

TCAADCAA

DCAAMCAA

MCAACAA

BCAAMCAA

In particular, during BCAA degradation, MCAA rather than MBAA was produced, indicating

that bromide was favored to be lost rather than chloride in the process of hydrogenolysis.

Constant concentrations of MCAA, DCAA, TCAA and BCAA in the control samples over the

reaction time indicate that sequential hydrogenolysis by iron was the only mechanism

responsible for the loss of HAA.

Generally, the degradation of HAA compounds followed the pseudo-first-order decay

kinetics. The pseudo-first-order degradation constants for four HAA compounds were estimated

by fitting a first-order decay equation to the HAA concentration data in Microsoft Excel. The

degradation rates for MCAA, DCAA, TCAA and BCAA were 0.015, 0.031, 0.047, and 0.055 h-1

,

respectively. Therefore, the reactivity of four HAA compounds with corroded iron coupons

followed a sequence of BCAA >TCAA >DCAA>MCAA, which is in agreement with the results

reported by Zhang et al. (2004). In their study, at similar initial concentrations of four HAA

compounds, the degradation rates of MCAA, DCAA, TCAA and BCAA were 0.001, 0.008,

0.502 and 3.87 h-1

, respectively. Since the pseudo-first-order rate of DBP reduction by iron was

223

significantly affected by pH, the mineral surface and iron surface speciation (Chun et al., 2005),

the difference in the decay constants for four compounds between this study and those reported

in the literature is likely because the reaction pH, iron type and/or iron surface conditions were

different.

In addition, according to the mechanisms of hydrogenolysis, DCAA can be formed as an

intermediate product of TCAA and BDCAA degradation. Therefore, it may initially accumulate

in the solution, and thus exhibit slower degradation compared with other compounds (Figure

10-29). Although MBAA and DBAA were formed as intermediate products as well, their

degradation rate constants were reported at least one order of magnitude higher than that of

DCAA (Zhang et al., 2004). Therefore, these two species experienced faster degradation rather

than accumulating in the system (Figure 10-32). The variation of MCAA concentrations over the

reaction time, as shown in Figure 10-32, indicates that MCAA, as an intermediate product and

due to its slow degradation kinetics, could also accumulate in the solution in the initial 24 hours

(illustrated by the increase of its concentration before 24 hours). After the degradation of other

HAA compounds was completed, the degradation of MCAA dominated its formation reactions

and caused the decrease in MCAA concentrations after 24 hours.

HAA Degradation by Lead 10.9.3

Compared with iron, a standard reduction potential of which being 0.44V at 25°C (Fe

Fe2+

+2e), lead has a lower standard reduction potential of 0.13V at 25°C (Pb Pb2+

+2e).

Although smaller, the positive reduction potential of lead determines that it may degrade HAA

compounds as well. Therefore, this study also investigated the degradation potential of lead for

HAA compounds, and the results are shown in Figure 10-33. Except for MCAA, the

concentrations of other species exhibited a decreasing trend with increasing reaction time.

DCAA had a relatively smaller degradation rate compared with other species, especially in the

initial 15 hours. In contrast, MCAA concentration increased quickly in the initial 10 hours (from

110 µg/L to 231 µg/L), and then tended to plateau afterwards (only 13% increase of MCAA

observed in the following 40 hours).

Single species degradation experiments by fresh lead coupons were also performed for

MCAA, DCAA, TCAA and BCAA. Figure 10-34 displays the variation of MCAA

224

concentrations with reaction time in the presence of fresh lead coupons. Constant concentrations

of MCAA over the test period of 48 hours indicate that lead may not reduce MCAA. While the

degradation of MCAA by iron was observed in Section 10.8.1, the difference in MCAA

degradation between iron and lead was primarily because lead has a weaker reductive activity

than iron.

Figure 10-33 Reduction of HAA9 by fresh lead coupons, Milli-Q water, pH 8.3, error bars


Figure 10-34 Degradation of MCAA by fresh lead coupons, Milli-Q water, pH 8.3, error bars


The degradation of DCAA, TCAA and BCAA in water samples containing fresh lead

coupons is illustrated in Figure 10-35. The pathways of DCAA, TCAA and BCAA degradation

were similar as those observed for corroded iron coupons. In the case of TCAA, its concentration

decreased over the reaction time with concomitant formation of DCAA and MCAA as

degradation products. DCAA concentration increased in the initial 10 hours, and then exhibited

0

60

120

180

240

300

0 10 20 30 40 50

HA

A (

µg

/L)

Time (hours)

TCAA BCAAMBAA MCAADBAA DCAABDCAA CDBAATBAA

0

30

60

90

120

150

0 0 1 2 2 4 8 10 24 33 48

MC

AA

(µ

g/L

)

Time (hours)

225

only 25% reduction in the following 22 hours. MCAA concentrations continuously increased

with increasing reaction time, and plateaued at 32 hours. The decreased concentration of TCAA

and the increased formation of DCAA and MCAA suggest that the degradation of HAA

compounds by lead also followed the mechanism of sequential hydrogenolysis. For BCAA, only

MCAA was detected as a degradation product rather than MBAA. It indicates that bromide was

preferentially removed from HAA compounds over chloride, which is similar to the observations

for corroded iron coupons. Constant concentrations of DCAA, TCAA and BCAA over the

Figure 10-35 Degradation of single HAA species by fresh lead coupons, Milli-Q water, pH 8.3,


0

20

40

60

80

100

120

0 10 20 30 40

HA

A (

µg

/L)

DCAA MCAA

0

20

40

60

80

100

0 10 20 30 40

HA

A (

µg

/L)

TCAA

DCAA

MCAA

0

20

40

60

80

100

120

0 10 20 30 40

HA

A (

µg

/L)

Time (hours)

BCAA MCAA

226

reaction period in the control bottles (plots are not shown) indicate that hydrogenolysis was the

primary mechanism for the destruction of the parent HAA compounds, the production of the

reaction intermediates, and the accumulation of final products in the presence of lead.

The pseudo-first-order degradation rate constants for DCAA, TCAA and BCAA were

estimated by fitting a first-order decay equation to the HAA concentration data in Microsoft

Excel, and they were 0.048, 0.089, and 0.14 h-1

, respectively. As such, the reactivity of three

HAA species with lead followed a sequence in decreasing order of BCAA, TCAA, and DCAA.

The degradation of BCAA is relatively rapid likely because bromide was preferentially removed

relative to chloride by the tested lead coupons. However, due to different metal surface

conditions of iron and lead coupons tested in this study, direct comparisons of their degradation

rates for HAA compounds cannot be made.

The degradation potential of lead corrosion products, including PbO, Pb(OH)2(CO3)2 and

PbO2, on HAA9 were also investigated, and the results regarding MCAA, MBAA, DCAA,

TCAA and BCAA are shown in Figure 10-36. The concentrations of these compounds remained

constant over 72 hours. Similar trends were also observed for the other four HAA compounds. It

indicates that lead corrosion products, in which lead is in an oxidized state, did not have

degradation potential for HAA9, and only elemental lead (Pb0) could react with HAA

compounds.

Results of this study provide an important implication about the fate of HAA compounds

in distribution systems and household plumbing. Besides biodegradation, HAA compounds may

also experience abiotic degradation by metal materials, e.g., iron and lead. However, the loss of

HAA compounds due to the reduction by these pipe materials may only be viable in a pipe

environment where new pipe are installed and are not corroded substantially.

227

Figure 10-36 Degradation of HAA species in the presence of 1g/L lead corrosion products. Milli-

Q water, pH 8.3, error bars indicate the measured maximum and minimum values (n=2)

References 10.9.4

APHA, AWWA, and WEF (2005) Standard Methods for the Examination of Water &

Wastewater. 21th Edition, Washington D C, USA.

0

30

60

90

120

150

0 3 9 24 48 72

HA

A (

µg

/L)

Time (hours)

PbO

MCAA MBAA DCAA TCAA BCAA

0

30

60

90

120

150

0 1 4 9 23 47 72

HA

A (

ug

/L)

Time (hours)

Pb(OH)2(CO3)2


0

30

60

90

120

150

0 1 4 9 23 47 72

HA

A (

ug

/L)

Time (hours)

PbO2


228

Baribeau, H., Krasner, S. W., Chinn, R., and Singer, P. C. (2005) Impact of biomass on the

stability of HAAs and THMs in a simulated distribution system. Journal American Water


Baribeau, H., Prevost, M., Desjardins, R., and Lafrance, P. (2001) Changes in chlorine and DOX

concentrations in distribution systems. Journal American Water Works Association,

93(12), 102-114.
















95-102.






229

Chun, C. L., Hozalski, R. M., and Arnold, T. A. (2005) Degradation ot drinking water


Technology, 39(21), 8525-8532.

Cotton, F. A., and Wilkinson, G. (1988) Advanced Inorganic Chemistry. Fifth Edition. John

Wiley & Sons, Inc., New York.

Cowman, G. A., and Singer, P. C. (1996) Effect of bromide ion on haloacetic acid speciation








DiGiano, F. A., and Zhang, W. (2005a) Pipe section reactor to evaluate chlorine-wall reaction.

American Water Works Association Journal, 97(1), 74-85.

Digiano, F. A., and Zhang, W. D. (2005b) Pipe section reactor to evaluate chlorine-wall reaction.


Duirk, S. E., Gombert, B., Croue, J. P., and Valentine, R. L. (2005) Modeling monochloramine

loss in the presence of natural organic matter. Water Research, 39(14), 3418-3431.







230

Edwards, M., and McNeill, L. S. (2002) Effect of phosphate inhibitors on lead release from pipes.


Edwards, M., Schock, M. R., and Meyer, T. E. (1996) Alkalinity, pH, and copper corrosion by-

product release. American Water Works Association Journal, 88(3), 81-94.

Facey, R. M., and Smith, D. W. (1995) Soft, low - temperature water - distribution corrosion -




Hallam, N. B., Hua, F., West, J. R., Forster, C. F., and Simms, J. (2003) Bulk decay of chlorine

in water distribution systems. Journal of Water Resources Planning and Management-

Asce, 129(1), 78-81.



36(14), 3479-3488.

Hassan, K. Z. A., Bower, K. C., and Miller, C. M. (2006) Iron oxide enhanced chlorine decay

and disinfection by-product formation. Journal of Environmental Engineering-Asce,

132(12), 1609-1616.





Systems (Catalogue No. H128-1/09-595E). Water, Air and Climate Change Bureau,


Hong, P. K. A., and MacAuley, Y. Y. (1998) Corrosion and leaching of copper tubing exposed to

chlorinated drinking water. Water Air and Soil Pollution, 108(3-4), 457-471.

231

Hsu, P. H. (1982) Crystallization of Iron(Iii) Phosphate at Room-Temperature. Soil Science

Society of America Journal, 46(5), 928-932.

Kim, E. J., and Herrera, J. E. (2010) Characteristics of Lead Corrosion Scales Formed during

Drinking Water Distribution and Their Potential Influence on the Release of Lead and

Other Contaminants. Environmental Science & Technology, 44(16), 6054-6061.

Korshin, G. V., Benjamin, M. M., and Sletten, R. S. (1997) Adsorption of natural organic matter

(NOM) on iron oxide: Effects on nom composition and formation of organo-halide

compounds during chlorination. Water Research, 31(7), 1643-1650.

LeChevallier, M. W., Lowry, C. D., and Lee, R. G. (1990a) Disinfecting Biofilms in a Model

Distribution-System. Journal American Water Works Association, 82(7), 87-99.

LeChevallier, M. W., Lowry, C. D., and Lee, R. G. (1990b) Disinfecting Biofilms in a Model

Distribution System. American Water Works Association Journal, 82(7), 87-99.


233-238.



152(1), 250-258.

Li, S. Y., Lixiao, N. X., Sun, C., and Wang, L. S. (2004) Influence of organic matter on

orthophosphate corrosion inhibition for copper pipe in soft water. Corrosion Science,

46(1), 137-145.



Technology, 37(13), 2920-2928.



34-52.

232



313.

McNeill, L. S., and Edwards, M. (2000) Phosphate inhibitors and red water in stagnant iron pipes.

Journal of Environmental Engineering-Asce, 126(12), 1096-1102.



McNeill, L. S., and Edwards, M. (2004) Importance of Pb and Cu particulate species for

corrosion control. Journal of Environmental Engineering-Asce, 130(2), 136-144.

Moriarty, B. E. (1990) Surface Studies of Corrosion-Inhibitors in Cooling Water-Systems.

Materials Performance, 29(1), 45-48.

Morris, J. C. (1975) Formation Of Halogenated Organics By Chlorination Of Water Supplies, a

Review. National Service Center for Environmental Publications (NSCEP), USEPA, N. S.

C. f. E. P. (NSCEP), ed., USEPA, Washington D.C.

Morton, S. C., Zhang, Y., and Edwards, M. A. (2005) Implications of nutrient release from iron

metal for microbial regrowth in water distribution systems. Water Research, 39(13),

2883-2892.

Munch, J. W., and Bassett, M. V. (2004) Method521: Determination of nitrosamines in drinking




Environmental Protection Agency, Cincinnati, Ohio 45268, 2004.




NACE International (2009) NACE International glossary of corrosion-related terms. National

Association of Corrosion Engineers International, Houston, TX.

233



Nguyen, M. L., Baker, L. A., and Westerhoff, P. (2002) DOC and DBP precursors in western US

watersheds and reservoirs. Journal American Water Works Association, 94(5), 98-112.

Odell, L. H., Kirmeyer, G. J., Wilczak, A., Jacangelo, J. G., Marcinko, J. P., and Wolfe, R. L.

(1996) Controlling nitrification in chloraminated systems. Journal American Water


Persson, P., Nilsson, N., and Sjoberg, S. (1996) Structure and bonding of orthophosphate ions at

the iron oxide aqueous interface. Journal of Colloid and Interface Science, 177(1), 263-

275.





Asce, 133(2), 180-185.





Rossman, L. A., Clark, R. M., and Grayman, W. M. (1994) MODELING CHLORINE

RESIDUALS IN DRINKING-WATER DISTRIBUTION-SYSTEMS. Journal of

Environmental Engineering-Asce, 120(4), 803-820.

Sarin, P., Clement, J. A., Snoeyink, V. L., and Kriven, W. W. (2003) Iron release from corroded

unlined cast-iron pipe. Journal American Water Works Association, 95(11), 85-96.

Schock, M. R. (1980) Response of lead solubility to dissolved carbonate in drinking water.


234

Schock, M. R. (1989) Understanding corrosion control strategies for lead. Journal American


Schock, M. R. (1990) Causes of temporal variability of lead in domestic plumbing systems.


Schock, M. R., and Lytle, D. A. (1995) Effect of pH, DIC, Orthophosphate and Sulfate on

Drinking Water Cuprosolvency. U.S. Environmental Protection Agency, Water Supply

and Water Resources Division, Cincinnati, Ohio 45268, EPA/600/R-95/085.

Schock, M. R., and Sandvig, A. M. (2009) Long-term effects of orthophosphate treatment on

copper concentration. Journal American Water Works Association, 101(7), 71-82.



Trofe, T. W., Inman, G. W., and Johnson, J. D. (1980) KINETICS OF MONOCHLORAMINE

DECOMPOSITION IN THE PRESENCE OF BROMIDE. Environmental Science &

Technology, 14(5), 544-549.





493.

Valentine, R. L., Jafvert, C. T., and Leung, S. W. (1988) Evaluation of a chloramine

decomposition model incorporating general acid catalysis. Water Research, 22(9), 1147-

1153.



Association, 98(2), 144-154.

235

Vikesland, P. J., Ozekin, K., and Valentine, R. L. (2001) Monochloramine decay in model and

distribution system waters. Water Research, 35(7), 1766-1776.

Vikesland, P. J., and Valentine, R. L. (2002) Iron oxide surface-catalyzed oxidation of ferrous

iron by monochloramine: Implications of oxide type and carbonate on reactivity.




Wilczak, A., Hoover, L. L., and Lai, H. H. (2003) Effects of treatment changer on chloramine






quality problems. Water Supply, Vol 8, Nos 1 and 2, 195-198.



Zhang, Y., and Edwards, M. (2009a) Accelerated chloramine decay and microbial growth by


51-62.

Zhang, Y., and Edwards, M. (2009b) Accelerated chloramine decay and microbial growth by


51-61.



236

Zhang, Y., Griffin, A., and Edwards, M. (2008) Nitrification in premise plumbing: Role of

phosphate, pH and pipe corrosion. Environmental Science & Technology, 42(12), 4280-

4284.

Zhang, Y., Love, N., and Edwards, M. (2009) Nitrification in Drinking Water Systems. Critical

Reviews in Environmental Science and Technology, 39(3), 153-208.

EFFECTS OF PHOSPHATE-BASED CORROSION INHIBITORS …...ii effects of phosphate-based corrosion...

Documents

Transcript of EFFECTS OF PHOSPHATE-BASED CORROSION INHIBITORS …...ii effects of phosphate-based corrosion...