Embed Size (px)
Citation preview
POTENTIAL OF PHARMACEUTICALS AND PERSONAL CARE PRODUCTS (PPCPS) AS NITROSAMINE PRECURSORS
DURING DRINKING WATER DISINFECTION
by
Ruqiao Shen
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 Ruqiao Shen (2013)
ii
POTENTIAL OF PHARMACEUTICALS AND PERSONAL CARE PRODUCTS (PPCPS) AS NITROSAMINE PRECURSORS DURING DRINKING WATER DISINFECTION
Ruqiao Shen
Doctor of Philosophy, 2013 Graduate Department of Civil Engineering University of Toronto
ABSTRACT
N-nitrosamines are considered as a group of emerging disinfection byproducts (DBPs) with
potential carcinogenicity at ng/L level. The presence of nitrosamines in drinking water is most
commonly associated with chloramination of amine-based precursors. This research investigates
the potential of amine-based pharmaceuticals and personal care products (PPCPs) as nitrosamine
precursors under practical drinking water disinfection conditions, as well as some critical factors
that may affect the nitrosamine formation via PPCPs.
All of the twenty selected PPCPs were able to form the corresponding nitrosamines upon
chloramine disinfection, and eight of them rendered molar conversions higher than 1 % under
practical disinfection conditions. Ranitidine had the highest N-nitrosodimethylamine (NDMA)
molar conversion among the tested PPCPs.
A three-parameter kinetic model was proposed to describe and predict the NDMA formation
from pharmaceuticals during chloramination in various water matrices. The model accurately
reflected all three significant characteristics of the NDMA formation curve, including an initial
lag phase, followed by a fast increase in NDMA formation, and eventually reaching a plateau.
iii
In lab-grade water, the NDMA formation from pharmaceuticals was affected by the Cl2:NH4-N
mass ratio, pH, and prechlorination. The NDMA formation increased with the Cl2:NH4-N mass
ratio, indicating an enhancement effect of dichloramine. The pH affected both the ultimate
NDMA conversion and the reaction rate. The reaction rate is mainly determined by the level of
non-protonated amine species, and it increased consistently with increasing pH. The ultimate
NDMA conversion is limited by the level of dichloramine, and the maximum NDMA formation
occurred in the pH range of 7 to 8. The application of prechlorination may increase or reduce the
NDMA conversion, depending on the chlorine reactivity towards the amine group and its
surrounding structures.
Water matrix components can slow down the initial NDMA formation from selected
pharmaceuticals most likely due to the formation of natural organic matter (NOM)-
pharmaceutical complexes, while they had less impact on the ultimate NDMA molar conversion.
The application of prechlorination may enhance the initial reaction by destroying the NOM-
pharmaceutical complexes, but prolonged prechlorination may further inhibit the NDMA
formation due to the binding between pharmaceuticals and NOM breakdown products.
iv
ACKNOWLEDGMENTS
This research was financially supported by the Canadian Water Network, the Natural Sciences
and Engineering Research Council of Canada, and the Ontario Research Fund.
I would like to thank my thesis supervisor, Professor Susan Andrews, for her advice,
encouragement and guidance in my research work and professional life over the last five years. I
would also like to thank Professors Bob Andrews and Brent Sleep for being on my supervisory
committee and offering their suggestions.
I must thank Richard Jones at the Ajax Water Treatment Plant and John Armour at the
Peterborough Water Treatment Plant for their assistance in water sampling over the course of my
research work. I would also like to thank Hong Zhang for her great patience in helping me with
the GC-MS. I also want to thank Sabrina Diemert for her help with the LC-OCD related work.
My appreciation also goes to Russell D’Souza for his general assistance in the lab, to my summer
student Song Lim for helping me preparing large number of samples, and to Nicolas Peleato for
driving me to the water treatment plant on a snowing day.
I would like to express my appreciation to Sarah Wilson, who I started my graduate school
together with and has been a dear friend since I first came to Canada. I would also like to thank
Kyla Smith, Heather Wray, Juan Zhang, Jacque-Ann Grant, and Anwar Sadmani for their moral
support inside and outside the lab during my difficult times. And I am very thankful to meet all
the wonderful people in the Drinking Water Research Group and have their general support
along the way.
Last but not least, I would like to thank my dearest parents for their support, encouragement and
understanding. It has been hard to be away from home halfway across the globe, but they have
always been the breath of life to me when I was going through the difficult times. And big thank
you goes to my boyfriend, Xinyuan, for going through the long distance with me together and for
his endless support and love along the way.
v
TABLE OF CONTENTS
ABSTRACT ............................................................................................................................ ii
ACKNOWLEDGMENTS ......................................................................................................... iv
TABLE OF CONTENTS ........................................................................................................... v
LIST OF TABLES .................................................................................................................... xi
LIST OF FIGURES ................................................................................................................. xiii
NOMENCLATURE .............................................................................................................. xviii
Chapter 1 Introduction ........................................................................................................ 1
1.1 Background ......................................................................................................................... 1
1.2 Research Objectives ............................................................................................................ 2
1.3 Associated Journal Publications ......................................................................................... 4
1.4 References ........................................................................................................................... 5
Chapter 2 Literature Review .............................................................................................. 8
2.1 NDMA and Other Nitrosamines ......................................................................................... 8
2.1.1 Background ............................................................................................................. 8
2.1.2 Source of NDMA and Its Precursors ...................................................................... 9
2.1.3 Formation of NDMA in Drinking Water .............................................................. 11
2.1.4 Operating Factors Affecting NDMA Formation ................................................... 14
2.1.5 NDMA Formation Control in Drinking Water ..................................................... 17
2.2 Pharmaceuticals and Personal Care Products (PPCPs) ..................................................... 18
2.2.1 Background ........................................................................................................... 18
2.2.2 Removal of PPCPs in Water Treatment Processes ............................................... 20
vi
2.2.3 Transformation of PPCPs upon Water Treatment Processes ................................ 21
2.3 Formation of Nitrosamines from PPCPs ........................................................................... 22
2.4 Research Questions and Gaps ........................................................................................... 23
2.5 References ......................................................................................................................... 25
Chapter 3 Materials and Methods .................................................................................... 46
3.1. Materials ........................................................................................................................... 46
3.1.1. Selection of Target PPCPs .................................................................................... 46
3.1.2. Preparation of Working Solutions ........................................................................ 49
3.1.3. Water Matrices ...................................................................................................... 50
3.2. Analytical Methods ........................................................................................................... 53
3.2.1 Nitrosamine Analysis (GC-MS) ........................................................................... 53
3.2.2 Quantum Property Calculation for PPCPs ............................................................ 55
3.2.3 Basic Water Quality Measurements and NOM Characterization ......................... 55
3.3 Nitrosamine Formation Protocol ....................................................................................... 58
3.4 QA/QC .............................................................................................................................. 60
3.5 References ......................................................................................................................... 61
Chapter 4 Demonstration of 20 Pharmaceuticals and Personal Care Products (PPCPs) as Nitrosamine Precursors During Chloramine Disinfection ..... 64
Abstract .................................................................................................................................... 65
Keywords ................................................................................................................................. 65
4.1 Introduction ....................................................................................................................... 66
4.2 Materials and Methods ...................................................................................................... 67
4.3 Results and Discussion ..................................................................................................... 69
4.3.1 Nitrosamine-FP under MFP Conditions ............................................................... 69
vii
4.3.2 Nitrosamine-FP vs. Molecular Properties ............................................................. 71
4.3.3 Nitrosamine-FP under SDS Conditions ................................................................ 77
4.3.3.1 Matrix Effect........................................................................................... 78
4.3.3.2 Impact of Initial Pharmaceutical Concentration ..................................... 80
4.3.3.3 Mixture Effect ........................................................................................ 83
4.3.3.4 Impact of Cl2:NH4-N Mass Ratio ........................................................... 85
4.4 Summary ........................................................................................................................... 86
4.5 References ......................................................................................................................... 88
Chapter 5 NDMA Formation From Four Pharmaceuticals: Reaction Kinetics and Water Matrix Effects ...................................................................................... 92
Abstract .................................................................................................................................... 93
Keywords ................................................................................................................................. 93
5.1 Introduction ....................................................................................................................... 94
5.2 Materials and Methods ...................................................................................................... 95
5.3 Results and Discussion ..................................................................................................... 97
5.3.1 Formation Kinetics in MQ Water ......................................................................... 97
5.3.2 Formation Kinetics in Different Water Matrices .................................................. 98
5.3.3 Kinetic Model ..................................................................................................... 103
5.4 Summary ......................................................................................................................... 110
5.5 References ....................................................................................................................... 111
Chapter 6 Formation of NDMA From Ranitidine and Sumatriptan: the Role of pH .. 115
Abstract .................................................................................................................................. 115
Keywords ............................................................................................................................... 115
6.1 Introduction ..................................................................................................................... 116
viii
6.2 Materials and Methods .................................................................................................... 118
6.3 Results and Discussion ................................................................................................... 119
6.3.1 Impact of pH on the NDMA Formation Kinetics ............................................... 119
6.3.2 Impact of Low pH on the NDMA-FP from Ranitidine ....................................... 128
6.4 Summary ......................................................................................................................... 129
6.5 References ....................................................................................................................... 130
Chapter 7 NDMA Formation From Amine-Based Pharmaceuticals: Impact From Prechlorination and Water Matrix ............................................................. 135
Abstract .................................................................................................................................. 136
Keywords ............................................................................................................................... 136
7.1 Introduction ..................................................................................................................... 137
7.2 Materials and Methods .................................................................................................... 139
7.3 Results and Discussion ................................................................................................... 142
7.3.1 Prechlorination Impacts in MQ Water ................................................................ 142
7.3.1.1 24 hr NDMA-FP upon Prechlorination ................................................ 142
7.3.1.2 NDMA Formation Kinetics upon Prechlorination ............................... 144
7.3.2 Prechlorination Impacts in Real Water Matrices ................................................ 146
7.3.3 NOM-Pharmaceutical-Cl2 Interactions ............................................................... 150
7.3.3.1 NOM-Pharmaceutical Interactions ....................................................... 150
7.3.3.2 Cl2-Pharmaceutical Interactions ........................................................... 151
7.3.3.3 Cl2-NOM Interactions........................................................................... 154
7.4 Summary ......................................................................................................................... 157
7.5 References ....................................................................................................................... 158
ix
Chapter 8 Conclusions and Recommendations for Future Research ......................... 165
8.1 Discussion of Major Themes from the Thesis as a Whole ............................................. 165
8.1.1 Nitrosamine Formation in Lab-Grade Water ...................................................... 165
8.1.2 NDMA Formation in Water Matrices Containing NOM .................................... 167
8.1.3 Critical Factors Affecting NDMA Formation from Pharmaceuticals ................. 168
8.1.4 Implications for NDMA Formation Control ....................................................... 170
8.2 Conclusions ..................................................................................................................... 172
8.3 Recommendations for Future Research .......................................................................... 174
8.4 References ....................................................................................................................... 177
Appendix 1. Nitrosamine Analysis: QA/QC (Chapter 3).................................................... 179
Appendix 2. Nitrosamine Formation from Water Matrices (NOM) (Chapter 3) ................ 183
Appendix 3. Potential NDMA Contamination in MQ Water (Chapter 3) .......................... 187
Appendix 4. Matrix Effect on the 24 hr NDMA-FP from Selected Pharmaceuticals – Statistical Analysis (Chapter 4) ...................................................................... 189
Appendix 5. Pharmaceutical Concentration Effect on the 24 hr NDMA-FP from Selected Pharmaceuticals – Statistical Analysis (Chapter 4) ........................................ 191
Appendix 6. Pharmaceutical Mixture Effect on the 24 hr NDMA-FP from Selected Pharmaceuticals – Statistical Analysis (Chapter 4) ........................................ 194
Appendix 7. Impact of Cl2:NH4-N Mass Ratio on the 24 hr NDMA-FP from Selected Pharmaceuticals – Statistical Analysis (Chapter 4) ........................................ 195
x
Appendix 8. NDMA Formation Kinetics: Reproducibility (Chapter 5) ............................. 196
Appendix 9. Summary of the Estimated Kinetic Model Parameters under Different Treatment Conditions and the Overall Kinetic Model Verification (Chapter 5-7) .................................................................................................................. 198
Appendix 10. Preliminary Chlorine/Chloramine Demand Tests to Determine the Initial NaClO and NH4Cl Dosage for the Sequential Disinfection Experiments (Chapter 7) ...................................................................................................... 207
Appendix 11. Prechlorination Impact on the 24 hr NDMA-FP from Selected Pharmaceuticals – ANOVA and Tukey’s HSD Analysis (Chapter 7) ........... 209
Appendix 12. Additional NDMA Formation Experiments with Ranitidine in Otonabee River Water: The Possible Impact of Cl2:TOC Ratio? (Chapter 7) ............... 211
Appendix 13. Prechlorination Impact on the NDMA Formation Kinetics from Ranitidine and Sumatriptan (Chapter 7) – ANOVA and Tukey’s HSD Analysis ........... 213
Appendix 14. Preliminary LC-MS results (Chapter 7) .......................................................... 215
Appendix 15. Additional LC-OCD Data for Chapter 7 ......................................................... 219
xi
LIST OF TABLES
Table 3.1. PPCPs investigated ...................................................................................................... 47
Table 3.2. Physico-chemical properties of selected PPCPs .......................................................... 48
Table 3.3. Buffer solution recipes (Source: Sigma-Aldrich buffer reference centre) ................... 50
Table 3.4. Basic water quality parameters for the selected water matrices (n = 5-10) ................. 51
Table 3.5. The nitrosamine extraction and concentration method outline .................................... 53
Table 3.6. The GC-MS analysis of nitrosamines method outline ................................................. 54
Table 3.7. GC-MS method: Method detection limits for nitrosamines (in Lake Ontario water) . 54
Table 3.8. Hydrophilic NOM fractions from LC-OCD analysis (summarized from Huber et al.,
2011) ..................................................................................................................................... 57
Table 3.9. LC-OCD: Analytical variance for selected water matrices (mg/L; pH adjusted to 7.0 ±
0.1) ........................................................................................................................................ 57
Table 3.10. General nitrosamine formation method outline ......................................................... 58
Table 3.11. General nitrosamine-FP experimental conditions (MQ) ........................................... 59
Table 3.12. The chloramine dosage applied in each water matrix under the SDS conditions ...... 59
Table 4.1. Basic water quality parameters (n = 5-10) and the chloramine dosage applied in each
matrix under the simulated distribution system (SDS) conditions ....................................... 67
Table 4.2. Comparisons with literature (Common conditions: pH = 7.0 ± 0.1, NH2Cl = 28.4
mg/L, PPCPs = 25 nM, room temperature) .......................................................................... 70
Table 4.3. The nitrosamine molar conversion from selected PPCPs vs. relevant molecular
properties............................................................................................................................... 72
Table 4.4. Example of the mixture effect with and without ranitidine ......................................... 84
xii
Table 5.1. Basic water quality parameters (n = 5-10) and the chloramine dosage applied in each
matrix under the simulated distribution system (SDS) conditions ....................................... 96
Table 5.2. LC-OCD results for the selected water matrices (Unit: mg/L carbon) ........................ 96
Table 5.3. Kinetic model parameter estimation and model verification ..................................... 104
Table 5.4. Correlation coefficients (R2) between the kinetic parameters and water quality
parameters ........................................................................................................................... 106
Table 6.1. Estimated kinetic model parameters under different pH conditions (MQ water, 25 nM
of pharmaceuticals, preformed chloramine = 2.5 ± 0.2 mg/L) ........................................... 121
Table 7.1. Water matrix source and basic water quality measurements (n = 5-10) .................... 141
Table 7.2. LC-OCD results for the selected water matrices (Unit: mg/L carbon) ...................... 141
Table 7.3. Estimated kinetic model parameters for the NDMA formation from ranitidine and
sumatriptan upon sequential chlorine and chloramine disinfection .................................... 145
Table 7.4. Comparison of chlorine reactivity towards ranitidine and sumatriptan ..................... 153
xiii
LIST OF FIGURES
Figure 1.1. Research motivation ..................................................................................................... 2
Figure 2.1. NDMA formation pathways upon drinking water disinfection (figure reproduced
from Shah and Mitch, 2012) ...................................................................................... 14
Figure 2.2. Distribution diagram for chloramine species with pH (Source: Palin, 1950) ............ 15
Figure 2.3. Theoretical breakpoint curve (Source: USEPA, 1999) .............................................. 16
Figure 3.1. Structures of selected PPCPs for this research ........................................................... 46
Figure 3.2. Structures of DMA, NDMA, DEA, and NDEA ......................................................... 47
Figure 3.3. Structures of d6-ranitidine and d6-sumatriptan .......................................................... 49
Figure 3.4. LC-OCD: Hydrophilic vs. Hydrophobic DOC distribution for selected water matrices
.................................................................................................................................... 52
Figure 3.5. LC-OCD: Hydrophilic NOM fractions of selected water matrices ............................ 52
Figure 3.6. The Lewatit® AF 5 activated carbon beads for nitrosamine extraction ..................... 53
Figure 3.7. A typical LC-OCD chromatograph for surface water (Lake Ontario) ....................... 56
Figure 4.1. Nitrosamine-FPs for selected PPCPs under the MFP conditions (Initial concentration
of individual PPCPs = 25 nM; 28.4 mg/L of NH2Cl; error bars represent the
variability due to multiple formation potential tests (n = 3)) ..................................... 70
Figure 4.2. Relationships between the nitrosamine molar conversions from selected PPCPs
(PPCPs = 25 nM, MFP, MQ) and the relevant molecular properties ........................ 73
Figure 4.3. The molecular electrostatic potential map for furan and thiazole ring ....................... 74
Figure 4.4. The molecular electrostatic potential map for the four H1-antihistamines ................ 76
xiv
Figure 4.5. Comparison of NDMA-FPs for selected PPCPs between the MFP and the SDS
conditions (Initial concentration of individual PPCPs = 25 nM; NH2Cl MFP = 28.4
mg/L; NH2Cl SDS = 2.5 mg/L; error bars represent the variability due to multiple
formation potential tests (n = 3))................................................................................ 78
Figure 4.6. NDMA-FPs for selected PPCPs: the matrix effect (SDS conditions; initial
concentration of individual PPCPs = 25 nM; error bars represent the variability due
to multiple tests (n = 3)) ............................................................................................. 79
Figure 4.7. NDMA-FPs and molar conversions for selected PPCPs at different initial
concentrations (SDS conditions; tap water; error bars represent the variability due to
multiple formation potential tests (n = 3)) ................................................................. 81
Figure 4.8. NDMA-FPs for sum of single PPCPs vs. PPCP-mixture (SDS conditions; eight-
pharmaceutical mixture: ranitidine, doxylamine, sumatriptan, chlorphenamine,
nizatidine, diltiazem, carbinoxamine, and tetracycline; seven-pharmaceutical
mixture: ranitidine excluded from the above eight; error bars represent the variability
due to multiple formation potential tests (n = 3); “*” indicates significant difference
between two bars, t-test, 95 % confidence level) ....................................................... 83
Figure 4.9. Impact of (a) Cl2:NH4-N mass ratio or (b) the fraction of NHCl2 on the NDMA
formation via selected pharmaceuticals (SDS conditions; tap water; eight-
pharmaceutical mixture: ranitidine, doxylamine, sumatriptan, chlorphenamine,
nizatidine, diltiazem, carbinoxamine, and tetracycline; 5 nM or 25 nM of each
pharmaceutical in the mixture; error bars represent the variability due to multiple
formation potential tests (n = 3))................................................................................ 85
Figure 5.1. NDMA molar conversion over time for chlorphenamine, doxylamine, ranitidine, and
sumatriptan in MQ water (SDS conditions; error bars represent the maximum and
minimum values in the formation potential tests under the same conditions (n = 2,
chlorphenamine and doxylamine), and represent the variability due to multiple
formation potential tests (n = 3, ranitidine and sumatriptan)) .................................... 97
xv
Figure 5.2. NDMA molar conversion over time for chlorphenamine, doxylamine, ranitidine, and
sumatriptan in different water matrices (SDS conditions; error bars represent the
maximum and minimum values in the formation potential tests under the same
conditions (n = 2, chlorphenamine and doxylamine), and represent the variability due
to multiple tests (n = 3, ranitidine and sumatriptan)) ................................................. 99
Figure 5.3. NDMA formation kinetics for ranitidine (25 nM) in pre-chloraminated Lake Ontario
water and Toronto tap water (SDS conditions; error bars represent the variability due
to multiple formation potential tests (n = 3)) ........................................................... 102
Figure 5.4. Linear correlation between (a) Lag and TOC; (b) Lag and SUVA; (c) k and TOC; and
(d) k and SUVA for four pharmaceuticals (SDS conditions; pharmaceutical
concentration = 5 and 25 nM; error bars represent the 95 % confidence interval for
the estimated model parameters) .............................................................................. 106
Figure 5.5. Linear correlation between the model-predicted and the independently measured
NDMA molar conversion at 24 hr for chlorphenamine, doxylamine, ranitidine, and
all four pharmaceuticals together (SDS conditions; data from four matrices (MQ,
Tap, LW, and RW) and two pharmaceutical concentration levels (5 and 25 nM); error
bars represent the standard deviation from multiple formation potential tests (n = 3))
.................................................................................................................................. 109
Figure 6.1. NDMA formation kinetics from ranitidine and sumatriptan upon chloramination
under different pH conditions (MQ, 25 nM of pharmaceutical, preformed chloramine
= 2.5 ± 0.2 mg/L; error bars represent the variability due to multiple formation
potential tests (n = 3)) .............................................................................................. 120
Figure 6.2. Potential correlation between the initial lag phase (Lag), percentage of non-
protonated pharmaceuticals, and pH (MQ, 25 nM of pharmaceuticals, preformed
chloramine = 2.5 ± 0.2 mg/L; error bars in the top two figures represent the 95 %
confidence interval of the model parameter) ........................................................... 122
Figure 6.3. Potential correlation between the reaction rate constant (k), percentage of non-
protonated pharmaceuticals, and pH (MQ, 25 nM of pharmaceuticals, preformed
xvi
chloramine = 2.5 ± 0.2 mg/L; error bars in the top two figures represent the 95 %
confidence interval of the model parameter) ........................................................... 123
Figure 6.4. Monochloramine decay and dichloramine increase under different pH (MQ,
preformed chloramine = 2.5 ± 0.2 mg/L) ................................................................. 124
Figure 6.5. The ultimate NDMA molar conversion, initial lag phase, dichloramine percentage,
and pharmaceutical species under different pH conditions (MQ, 25 nM of
pharmaceuticals, preformed chloramine = 2.5 ± 0.2 mg/L; error bars for the NDMA
molar conversion represent the variability due to multiple formation potential tests (n
= 3); error bars for the initial lag phase represent the 95 % confidence interval for the
estimated model parameter) ..................................................................................... 126
Figure 6.6. Generalized impact of precursor species and chloramine species on the NDMA
formation .................................................................................................................. 127
Figure 6.7. NDMA molar conversion from ranitidine (25 nM) upon chloramination under
different pH conditions (MQ water, preformed chloramine = 2.5 ± 0.2 mg/L,
incubation time = 72 hr; error bars represent the variability due to multiple formation
potential tests (n = 3)) .............................................................................................. 128
Figure 7.1. 24 hr NDMA-FP from eight pharmaceuticals upon sequential chlorination and
chloramination disinfection (MQ water; error bars represent the variability due to
multiple tests (n = 3)) ............................................................................................... 143
Figure 7.2. NDMA formation kinetics from ranitidine and sumatriptan upon sequential chlorine
and chloramine disinfection in MQ water (Error bars represent the maximum and
minimum values under the same reaction conditions (n = 2)) ................................. 144
Figure 7.3. NDMA formation kinetics from ranitidine upon sequential chlorine and chloramine
disinfection in Lake Ontario and Otonabee River water (Error bars represent the
variability due to multiple formation potential tests (n = 3)) ................................... 146
Figure 7.4. NDMA formation kinetics from sumatriptan upon sequential chlorine and chloramine
disinfection in Lake Ontario and Otonabee River water (Error bars represent the
variability due to multiple formation potential tests (n = 3)) ................................... 148
xvii
Figure 7.5. Comparison of the kinetic model parameters upon sequential chlorine and chloramine
disinfection in three water matrices (Error bars represent the 95 % confidence
interval) .................................................................................................................... 149
Figure 7.6. Possible chlorine attack site and predicted fragmentation pattern of ranitidine ....... 152
Figure 7.7. Possible chlorine attack site(s) and predicted chlorination product of sumatriptan . 153
Figure 7.8. Change of major NOM fractions upon sequential chlorination and chloramination in
Lake Ontario water ................................................................................................... 155
Figure 7.9. Change of major NOM fractions upon sequential chlorination and chloramination in
Otonabee River water ............................................................................................... 156
Figure 7.10. Potential interactions in between NOM, pharmaceuticals, and free chlorine ........ 156
Figure 8.1. Critical factors that affect the NDMA formation from amine-based precursors ...... 170
xviii
NOMENCLATURE
% Percent ºC Degree(s) Celsius Amp Ampere; electric current unit amu Atomic mass unit ANOVA Analysis of variance AOP Advanced Oxidation Process APHA American Public Health Association AVE Average AWWA American Water Works Association bar Unit of pressure; equivalent to 100 kilopascals C Carbon C18 Carbon 18 C measured Measured concentration of target analyte C spike Spiked concentration of target analyte CCL3 Contaminant candidate list 3 CDPH California Department of Public Health CI Chemical ionization Cl Chlorine Cl-UDMH Chlorinated unsymmetrical dimethylhydrazine ClO2 Chlorine dioxide Cl2 Chlorine Cl2:NH4-N The mass ratio of chlorine to ammonia nitrogen when preparing the
chloramine solution Cl2/TOC The ratio of the applied chlorine or chloramine dosage to the total organic
carbon concentration of the respective water matrix cm Centimeter(s) CT Concentration × time d6-NDMA Deuterated NDMA d6-Ranitidine Deuterated ranitidine d6-Sumatriptan Deuterated sumatriptan Da Dalton
xix
DBP Disinfection by-product DCM Dichloromethane DEA Diethylamine DEET N,N-diethyltoluamide DI Deionized DMA Dimethylamine DOC Dissolved organic carbon EDG Electron-donating group EfOM Effluent organic matter EPA Environmental Protection Agency ESI Electrospray ionization eV Electron volt(s) EWG Electron-withdrawing group FIX Fluidized ion exchange FP Formation potential g Gram(s) GAC Granular activated carbon GC-MS Gas Chromatography - Mass Spectrometry H2O2 Hydrogen peroxide HAAs Haloacetic acids HCl Hydrochloric acid HFBA Heptafluorobutyric acid HIN Molecular file format hr Hour(s) HS Humic substances; humics HSD (Tukey’s) Honestly significant difference ID Inner diameter IRIS Integrated Risk Information System k Parameter in the kinetic model; the pseudo-first order rate constant KH2PO4 Potassium dihydrogen phosphate K2HPO4 Dipotassium phosphate L Liter(s) Lag Parameter in the kinetic model; the time required to achieve 50 % of the
xx
ultimate molar conversion LC-MS Liquid Chromatography - Mass Spectrometry LC-OCD Liquid Chromatography - Organic Carbon Detection LDPE Low-density polyethylene Log D The log of the distribution coefficient Log P The log of the partition coefficient LMW acids Low molecular weight acids LMW neutrals Low molecular weight neutrals LW Lake Ontario water m Meter(s) m/h Meter(s) per hour mm Millimeter(s) MassDEP Massachusetts Department of Environmental Protection MDL(s) Method detection limit(s) MEP Molecular electrostatic potential MFP Modified formation potential mg Milligram(s) mg/L Milligram(s) per liter
µ Micro
µg Microgram(s)
µg/L Microgram(s) per liter
min Minute(s) mL Milliliter(s) mL/min Milliliter(s) per minute
µL Microliter(s)
µm Micrometer(s)
MOE Ontario Ministry of the Environment MOL Molecular file format MQ Milli-Q® water MW Molecular weight N Nitrogen n Number of measurements NA Not applicable/available
xxi
NaClO Sodium hypochlorite NaHCO3 Sodium bicarbonate NaOH Sodium hydroxide Na2CO3 Sodium carbonate Na2HPO4 Disodium hydrogen phosphate Na2S2O3 Sodium thiosulfate NDMA N-nitrosodimethylamine NDEA N-nitrosodiethylamine NH2Cl Monochloramine NHCl2 Dichloramine NH4Cl Ammonia chloride ng Nanogram(s) ng/L Nanogram(s) per liter nM Nanomole(s) nm Nanometer(s) NOM Natural organic matter O Oxygen O3 Ozone OD Outer diameter OEHHA Office of Environmental Health Hazard Assessment ·OH Hydroxyl radicals OND Organic nitrogen detector p Statistical significance of correlation pH The negative log of the hydrogen ion concentration; -log {H+} pKa The negative log of the acid dissociation constant; -log Ka polyDADMAC Polydiallyldimethylammonium chloride PPCPs Pharmaceuticals and personal care products psi Pounds per square inch PTV Programmed temperature vaporizer QA/QC Quality assurance/ Quality control QSAR Quantitative structure-activity relationship R2 Statistical coefficient of determination RO Reverse osmosis
xxii
rpm Revolutions per minute RSD Relative standard deviation RW Otonabee River water S Sulfur s Second(s)
θ Parameter in the kinetic model; the ultimate NDMA molar conversion
SDS Simulated distribution system STD Standard deviation SUVA Specific ultraviolet absorbance; UV254/TOC t Reaction time Tap Toronto tap water THMs Trihalomethanes TiO2 Titanium dioxide TOC Total organic carbon TRW Treated Otonabee River water, after filtration and prior to chlorination U Parameter in the kinetic model; the ultimate NDMA-FP UDMH Unsymmetrical dimethylhydrazine USEPA United States Environmental Protection Agency UV Ultraviolet UV(A)254 Ultraviolet absorbance at a wavelength of 254nm V Volt(s) vs. Versus WHO World Health Organization WTP Water treatment plant WWTP Wastewater treatment plant
1
Chapter 1 Introduction
1.1 Background
N-nitrosodimethylamine (NDMA) is a member of N-nitrosamines which are considered as a
group of emerging disinfection byproducts (DBPs) with potential carcinogenicity at ng/L levels
(EPA IRIS, 1993). The presence of nitrosamines in finished drinking water has been commonly
associated with chloramine disinfection. With more utilities switching from free chlorine to
chloramine as the secondary disinfectant, NDMA has been detected in many chloraminated
drinking water systems across the world (Asami et al., 2009; Blute et al., 2010; Charrois et al.,
2004, 2007; Russell et al., 2012). A number of research efforts have been invested in identifying
potential NDMA precursors relevant to drinking water. Dimethylamine (DMA) and natural
organic matter (NOM) are the most well-studied NDMA precursors (Chen and Valentine, 2007;
Dotson et al., 2007; Gerecke and Sedlak, 2003; Mitch and Sedlak, 2004), but the typically low
yields cannot always explain all the NDMA precursors detected in natural waters. Much higher
level of NDMA formation has been associated with wastewater-impacted surface water
(Krasner, 2009; Schreiber and Mitch, 2006b; Shah et al., 2012), indicating the contribution from
anthropogenic compounds, but few specific compounds have been identified.
Concurrently, pharmaceuticals and personal care products (PPCPs) have gained much attention
in recent years because of the worldwide increasing consumption of these substances and their
frequent detection in the aquatic and terrestrial environment at ng/L to lower µg/L levels. PPCPs
are usually poorly removed at water treatment plants (WTPs) because conventional technologies
are not specifically designed to remove them at trace levels. Therefore, they might pass through
the treatment train and come into contact with the disinfectant. Currently, there are a number of
researches focusing on removing PPCPs using advanced technologies, but their transformation
upon different treatment processes remains largely unknown, especially during the drinking
water disinfection process. Theoretically, any tertiary amines containing DMA groups could be
potential NDMA precursors, while there are a lot of amine-based PPCPs with DMA in their
structures. It is therefore very likely that PPCPs may contribute to the formation of NDMA upon
chloramine disinfection. Although the presence of PPCPs in drinking water sources has been
2
suggested to be inconsequential due to their low concentrations, it might be linked to potential
adverse health effects via the transformation into carcinogenic DBPs such as NDMA in finished
drinking water, and therefore requires more research.
Figure 1.1. Research motivation
The formation of NDMA is affected by several factors such as pH (Krasner et al., 2010; Mitch
and Sedlak, 2002a, b), Cl2:NH3-N mass ratio (Krasner et al., 2010; Mitch et al., 2005; Schreiber
and Mitch, 2005, 2006a), and the application of pre-oxidation (Chen and Valentine, 2008;
Charrois and Hrudey, 2007; Mitch et al., 2010; Shah et al., 2012). Moreover, reaction time is
also significant since further NDMA formation has been observed in distribution systems with
increasing water age (Barrett et al., 2003; Charrois and Hrudey, 2007). Knowledge on the
chemistry and impact factors is vital in predicting and controlling the NDMA formation.
However, most of the findings to-date have been based on the model precursors DMA and
NOM, while little is known for PPCP-based precursors. Specifically, PPCPs are usually present
at trace levels and thus their concentrations are several orders of magnitude lower than the
disinfectants, while studies using DMA and NOM typically employed comparable
concentrations of precursors and disinfectants; therefore findings obtained from DMA and NOM
may not always apply to PPCP-based precursors. As such, there is a need to further investigate
the critical factors that may affect the NDMA formation and reaction kinetics from PPCPs.
1.2 Research Objectives
The overall objective of this research was to investigate the transformation of selected PPCPs
under practical drinking water disinfection conditions, with the major focus on their potentials to
form nitrosamines and critical factors that may affect the nitrosamine formation via PPCPs.
3
Some of the findings may also apply to other nitrosamine precursors with similar structures, and
thus can provide some useful information for utilities to plan for effective NDMA control
strategies. In addition, this could be of particular concern for water reuse processes where much
higher concentrations of pharmaceuticals might be subjected to chloramination.
The specific objectives were to:
1. Determine the nitrosamine formation potential from amine-based PPCPs upon
chloramination. Chapter 4 demonstrates the formation of nitrosamines from 20 amine-
based PPCPs under a high chloramine dosage and long incubation time scenario, as well
as a practical disinfection scenario. It also looks into the impact from pharmaceutical
concentration, pharmaceutical mixture, as well as the Cl2:NH3-N mass ratio.
2. Examine the NDMA formation kinetics from selected PPCPs upon chloramination. To-
date, the limited studies on NDMA formation via pharmaceuticals have been mostly
conducted in lab-grade water. However, the presence of organic matter in real water
matrices may compete with PPCPs for chloramine or interact with the PPCPs, thus may
have an impact on their conversion into NDMA. Chapter 5 discusses the NDMA
formation kinetics from three amine-based pharmaceuticals in both lab-grade water and
natural water samples, and proposes a statistical model to describe and predict the
NDMA formation from selected pharmaceuticals in various water matrices.
3. Investigate the role of pH in NDMA formation from selected PPCPs. The formation of
NDMA is mainly determined by the speciation of chloramines and the precursor amine
groups, both of which are highly dependent on pH. The impact of pH on NDMA
formation has been studied for the model precursor DMA and NOM, but little is known
for amine-based PPCPs. Specifically, there is lack of study on the NDMA formation
kinetics under different pH. Chapter 6 investigates the impact of pH on the formation of
NDMA, especially the reaction kinetics, from two amine-based pharmaceuticals under
drinking water relevant conditions.
4
4. Investigate the impact of prechlorination on the NDMA formation from selected PPCPs.
One option to control the NDMA formation is to remove the precursors via pre-
oxidation, and prechlorination is among the most effective options in reducing NDMA
formation. However, most of the findings are based on the single-precursor scenario
using the model precursor DMA and NOM, while few studies considered the potential
interactions between water matrix components and the target precursors when
investigating the prechlorination impact. Specifically, little is known for the behavior of
amine-based PPCPs. Chapter 7 demonstrates the impact of prechlorination on the
NDMA formation kinetics from several amine-based pharmaceuticals, and compares
the prechlorination impact with and without the presence of NOM. In addition, Chapter
7 discusses the potential interactions in between NOM, pharmaceuticals, and free
chlorine, as well as how these interactions could affect the NDMA formation from
selected pharmaceuticals.
1.3 Associated Journal Publications
Part of Chapter 4 was previously published as “Shen, R., Andrews, S.A., 2011a. Demonstration
of 20 pharmaceuticals and personal care products (PPCPs) as nitrosamine precursors during
chloramine disinfection. Water Res. 45, 944-652”.
Part of Chapter 5 was previously published as “Shen, R., Andrews, S.A., 2011b. NDMA
formation kinetics from three pharmaceuticals in four water matrices. Water Res. 45, 5687-
5694”.
Part of Chapter 6 was published as “Shen, R., Andrews, S.A., 2013a. Formation of NDMA from
ranitidine and sumatriptan: the role of pH. Water Res. 47, 802-810”.
Part of Chapter 7 has been submitted to Water Research (under review) as “Shen, R., Andrews,
S.A., 2013b. NDMA formation from amine-based pharmaceuticals: impact from prechlorination
and water matrix”.
5
The first three manuscripts have been reproduced in this thesis with permission of the publisher.
The fourth manuscript is still under review, and copyright permission will be obtained once the
publication is finalized.
1.4 References
Asami, M., Oya, M., Kosaka, K., 2009. A nationwide survey of NDMA in raw and drinking
water in Japen. Sci. Total. Environ. 407, 3540-3545
Barrett, S., Hwang, C., Guo, Y., Andrews, S.A., Valentine, R., 2003. Occurrence of NDMA in
drinking water: a North American survey, 2001-2002. In: Proceedings of the American Water
Works Association's Annual Conference, Anaheim, CA, June 15-19, 2003
Blute, N., Russell, C., Chowdhury, Z., Wu, X., Via, S., 2010. Nitrosamine occurrence in the U.S.
– analysis and interpretation of UCMR2 data. In: Proceedings of the AWWA Water Quality
Technology Conference, Savannah, GA, November 14-18, 2010
Charrois, J.W.A., Arend, M.W., Froese, K.L., Hrudey, S.E., 2004. Detecting N-nitrosamines in
drinking water at Nanogram per liter levels using ammonia positive chemical ionization. Env.
Sci. Technol. 38, 4835-4841
Charrois, J.W.A., Boyd, J.M., Froese, K.L., Hrudey, S.E., 2007. Occurrence of N-nitrosamines in
Alberta public drinking-water distribution systems. J. Environ. Eng. Sci. 6, 103-114
Charrois, J.W.A., Hrudey, S.E., 2007. Breakpoint chlorination and free-chlorine contact time:
implications for drinking water N-nitrosodimethylamine concentrations. Water Res. 41, 674-682
Chen, Z., Valentine, R.L., 2007. Formation of N-Nitrosodimethylamine (NDMA) from humic
substances in natural water. Environ. Sci. Technol. 41, 6059-6065
6
Chen, Z., Valentine, R.L., 2008. The influence of the pre-oxidation of natural organic matter on
the formation of N-Nitrosodimethylamine (NDMA). Environ. Sci. Technol. 42, 5062-5067
Dotson, A., Westerhoff, P., Krasner, S.W., 2007. Nitrosamine formation from natural organic
matter isolates and sunlight photolysis of nitrosamines. In: Proceedings of AWWA Annual
Conference and Exposition. Toronto, ON, Canada, June 24-28, 2007
EPA Integrated Risk Information System (IRIS), 1993. N-Nitrosodimethylamine; CASRN 62-
75-9. www.epa.gov/iris/subst/0045.htm
Gerecke, A.C., Sedlak, D.L., 2003. Precursors of N- Nitrosodimethylamine in natural waters.
Environ. Sci. Technol. 37, 1331-1336
Krasner, S.W., 2009. The formation and control of emerging disinfection by-products of health
concern. Phil Trans. R Soc. A 367, 4077-4095
Krasner, S.W., Dale, M.S., Lee, C.F.T., Garcia, E.A., Wong, T.M., Mitch, W., Von Gunten, U.,
2010. Difference in reactivity and chemistry of NDMA precursors from treated wastewater and
from polyamine polymers. In: Proceedings of the AWWA Water Quality Technology
Conference, Savannah, GA, November 14-18, 2010
Mitch, W.A., Oelker, G.L., Hawley, E.L., Deeb, R.A., Sedlak, D.L., 2005. Minimization of
NDMA formation during chlorine disinfection of municipal wastewater by application of pre-
formed chloramines. Environ. Eng. Sci. 22 (6), 882-890.
Mitch, W.A., Krasner, S.W., Lee, C.F.T., Wong, T.M., 2010. Tradeoffs in DBP formation during
chloramination following pre-oxidation for nitrosamine control. In: Proceedings of the AWWA
Water Quality Technology Conference, Savannah, GA, November 14-18, 2010
Mitch, W.A., Sedlak, D.L., 2002a. Factors controlling nitrosamine formation during wastewater
chlorination. Wa. Sci. Technol. 2 (3), 191-198
7
Mitch, W.A., Sedlak, D.L., 2002b. Formation of N- Nitrosodimethylamine (NDMA) from
dimethylamine during chlorination. Environ. Sci. Technol. 36, 588-595
Mitch, W.A., Sedlak, D.L., 2004. Characterization and fate of N-nitrosodimethylamine
precursors in municipal wastewater treatment plants. Environ. Sci. Technol. 38, 1445-1454
Russell, C.G., Blute, N.K., Via, S., Wu, X., Chowdhury, Z., 2012. Nationwide assessment of
nitrosamine occurrence and trends. J. Am. Water Works Assoc. 104, 205-217
Schreiber, I.M., Mitch, W.A., 2005. Influence of the order of reagent addition on NDMA
formation during chloramination. Environ. Sci. Technol. 39, 3811-3818
Schreiber, I.M., Mitch, W.A., 2006a. Nitrosamine formation pathway revisited: the importance
of chloramines speciation and dissolved oxygen. Environ. Sci. Technol. 40, 6007-6014
Schreiber, I.M., Mitch, W.A., 2006b. Occurrence and fate of nitrosamines and nitrosamine
precursor in wastewater-impacted surface waters using boron as a conservative tracer. Environ.
Sci. Technol. 40, 3203-3210
Shah, A.D., Krasner, S.W., Lee, C.F.T., von Gunten, U., Mitch, W.A., 2012. Trade-offs in
disinfection byproduct formation associated with precursor preoxidation for control of N-
Nitrosodimethylamine formation. Environ. Sci. Technol. 46, 4809-4818
8
Chapter 2 Literature Review
This chapter mainly reviews the literature published up to the early stage of this research, and
has been updated to include more recent publications in this field during the course of this work.
2.1 NDMA and Other Nitrosamines
2.1.1 Background
N-nitrosodimethylamine (NDMA) is a member of N-nitrosamines found in food, beer, cured
meats, rubber products, tobacco smokes and more recently, drinking water. There is a growing
concern regarding the health effects associated with exposure to nitrosamines because of their
potential carcinogenicity (EPA IRIS, 1993), with the risk of two to three orders of magnitude
higher than “traditional” halogenated disinfection byproducts (DBPs) currently regulated in
drinking water (e.g., trihalomethanes (THMs) and haloacetic acids (HAAs)). According to the
USEPA integrated risk information service (IRIS) database, a drinking water concentration of
0.7 ng/L for NDMA or 0.2 ng/L for NDEA (N-nitrosodiethylamine) will result in a 10-6 lifetime
cancer risk; in contrast, a drinking water concentration of 6 µg/L for chloroform (a THM) is
required to result in the same level of risk.
In November 1989, lower µg/L levels of NDMA were discovered in the aquifer which supplied
water to the town of Elmira, Ontario, Canada; this was attributed to contamination from a nearby
chemical plant (Jobb et al., 1994). Later in 1998, elevated levels of NDMA were detected in
drinking water wells near two rocket testing facilities in Sacramento County, CA, U.S., which
used unsymmetrical dimethylhydrazine (UDMH)-based rocket fuels (CDPH, 2011). These
NDMA incidents have driven the local governments to take actions. NDMA is currently
regulated in drinking water in several states and provinces across North America, including
Ontario (9 ng/L (0.12 nM); MOE, 2003), Massachusetts (10 ng/L (0.13 nM); MassDEP, 2004),
and California (10 ng/L (0.13 nM); OEHHA, 2006). The World Health Organization (WHO)
suggests a 100 ng/L (1.35 nM) guideline concentration for NDMA (WHO, 2008). USEPA also
9
placed it on the drinking water contaminant candidate list 3 (CCL3) together with other four
nitrosamines (USEPA, 2009). More recently, Health Canada has proposed a drinking water
maximum acceptable concentration for NDMA at 40 ng/L (0.54 nM; Health Canada, 2010), and
in Australia NDMA is included in the Australian Drinking Water Guideline at 100 ng/L (1.35
nM) since 2011 (NHMRC, NRMMC, 2011).
The presence of NDMA in treated drinking water has been commonly associated with
chloramine disinfection. With more utilities switching from free chlorine to chloramine, NDMA
has been detected in many drinking water systems. Charrois et al. (2004) reported the NDMA
concentration of 67 ng/L in finished drinking water and further increased to 180 ng/L within the
distribution system at one location in Alberta, Canada. A later survey of 20 Alberta municipal
drinking water distribution systems revealed the occurrence of NDMA up to 100 ng/L as well as
other two nitrosamines (Charrois et al., 2007). The national survey conducted by USEPA
showed positive NDMA detection in 35 % of the chloraminated drinking water systems (Blute et
al., 2010). Most of the time the NDMA detected in chloraminated systems was below 10 ng/L;
only a small portion of samples showed higher concentrations (Russell et al., 2012).
2.1.2 Source of NDMA and Its Precursors
NDMA is widely used in industrial products, such as rubber, rocket fuel solvents, gasoline and
lubricant additives, plasticizer, and pesticide in the inhibition of nitrification in soil. As a result,
NDMA may be present in discharges of such industries and thus in raw sewage and treated
wastewater effluents. For example, median concentration of 46 ng/L of NDMA in secondary
wastewater effluents was reported in several sewage treatment plants before disinfection in
California (Sedlack et al., 2005). And chlorination of secondary wastewater effluent typically
results in the formation of NDMA at concentrations between 20 and 100 ng/L (Mitch and Sedlak,
2002a), which will then impact the downstream receiving water body. However, NDMA can be
readily degraded in the aquatic environment through photolysis (Dotson et al., 2007) and
biotransformation (Fournier et al., 2006; Yang et al., 2005), and thus the NDMA detected in
drinking water is more likely formed during the disinfection process.
10
Dimethylamine (DMA) has been used a lot as the model precursor of NDMA. DMA is present in
food and can be liberated during digestion in human bodies. Due to its excretion and widespread
industrial uses, it is ubiquitous in natural waters. However, its concentration and low NDMA
yield can only account for a small portion of NDMA formed upon chloramination (Gerecke and
Sedlak, 2003). Other aliphatic tertiary amines containing DMA functional groups could also
contribute to NDMA formation. For example, DMAI (3- (Dimethylaminomethyl) indole) and
DMAP (4-Dimethylaminoantipyrine) from malted barley are known to be the responsible tertiary
amine alkaloids for the formation of NDMA in beer (Mangino and Scanlan, 1985; Poocharoen
et al., 1992), and they also showed the highest conversions rates upon ozonation disinfection
among the model tertiary amines (Lee et al., 2007). Natural organic matter (NOM) has also been
demonstrated to form NDMA upon chloramine disinfection, depending on the water source and
the property of NOM in each water matrix. Humic fraction of NOM has been associated with the
formation of NDMA (Chen and Valentine, 2007), and hydrophilic fraction of NOM and isolates
with a lower carbon to nitrogen ratio (C: N) typically formed high concentrations of NDMA
(Dotson et al., 2007). Krasner et al. (2008) have further suggested that non-humic substances in
NOM may also be a significant source of NDMA precursors. However, chloramination of these
precursors typically gave low yields of NDMA which cannot always explain all the precursors
detected in natural waters.
Higher level of NDMA formation has been associated with wastewater-impacted surface water
(Krasner, 2009; Schreiber and Mitch, 2006b; Shah et al., 2012). It was suggested that the
NDMA precursors in municipal wastewater effluent usually consisted of compounds other than
DMA that are capable of passing through 3000 Daltons (Da) cutoff ultrafilters (Mitch and
Sedlak, 2004), and half of the precursors were hydrophobic and associated with colloids or
macromolecules (Krauss et al., 2010). The NDMA formation has also been associated with low
molecular weight organic nitrogen-containing species in wastewater effluent (Pehlivanoglu-
Mantas and Sedlak, 2006), and several model nitrogenous organic compounds with different
dissolved organic nitrogen (DON) characteristics were studied in terms of their nitrosamine
formation potential (Chang et al., 2011). Tertiary alkylamines could be important NDMA
precursors associated with EfOM (wastewater effluent organic matter), and model tertiary
amines were demonstrated to form NDMA upon chlorination and chloramination (Le Roux et al.,
11
2012; Mitch and Schreiber, 2008). Moreover, soluble microbial products (SMPs) generated from
a biofilter was found to be an important source of NDMA precursor, with high reactivity at
higher pH and lower temperature (Krasner et al., 2010).
NDMA formation from EfOM also indicates the possible contribution from anthropogenic
compounds. For instance, some amine-based polymers and resins used in water and wastewater
treatment plant could serve as sources of NDMA precursors (Kohut and Andrews, 2003; Mitch
and Sedlak, 2004; Najm and Trussell, 2001; Park et al., 2009; Siddiqui and Atasi, 2004; Wilczak
et al., 2003). In addition, there are some other industrial products containing DMA functional
groups being reported as potential NDMA precursors, such as the fungicide thiram (Graham et
al., 1995), the metal chelator dimethyldithiocarbamate (Weissmahr and Sedlak, 2000), the
herbicide diuron (Chen and Young, 2008), and dimethylsulfamide (DMS), a degradation product
of the fungicide tolyfluanide (Schmidt and Brauch, 2008). More recently, studies have linked
the NDMA formation to ingredients in pharmaceuticals and personal care products (Kemper et
al., 2010; Le Roux et al., 2011a, b).
2.1.3 Formation of NDMA in Drinking Water
NDMA can be formed via nitrosation mechanism at acidic environment, which involves the
formation of nitrosyl cation (NO+) or similar nitrogen-containing species (e.g., N2O4) during
acidification of nitrite, followed by the formation of NDMA from the reaction between the
nitrosyl cation and an amine (Choi and Valentine, 2003) (Figure 2.1, C). This mechanism is
mainly observed in food processing where nitrite is used to prevent bacteria growth. In drinking
water, the rate of nitrosation would be slow at neutral and basic pH, though some studies
reported that the reaction may still occur at circumneutral pH with the existence of catalysts such
as CO2 and carbonyl compounds (Lv et al., 2007, 2009). Usually this mechanism is of less
importance in drinking water due to the generally low nitrite concentrations; however, it may
cause a concern if breakpoint chlorination is conducted to achieve a significant free chlorine
residual in the presence of nitrite, where enhanced nitrosamine formation may occur through the
reaction with nitrite and hypochlorite (Schreiber and Mitch, 2007).
12
During drinking water treatment process, NDMA is most commonly formed via the slow
reaction between chloramines and amine-based precursors. The UDMH pathway was initially
proposed by Choi and Valentine (2002a) and Mitch and Sedlak (2002b), involving a reaction
between monochloramine and DMA to form UDMH and its subsequent oxidation to NDMA by
chloramines (Figure 2.1, A). The rate of UDMH formation via this mechanism was found to be
pH-dependent, with the maximum rate at neutral and slightly basic pH. However, the NDMA
concentration formed during chloramination of DMA was found at least 2 orders of magnitude
higher than chloramination of equivalent concentrations of UDMH, the proposed intermediate;
and the rate constant for UDMH formation from chloramination of DMA, determined via direct
measurement of UDMH (0.081 M-1s-1, Yagil and Anbar, 1962), was 2 orders of magnitude lower
than that proposed by Choi and Valentine (2002a) (6.4 M-1s-1) which modeled the NDMA
formation from NH2Cl and DMA. It was observed that the occurrence of NHCl2 significantly
enhanced NDMA formation, regardless of its relatively minor fraction in total chloramines
(Mitch et al., 2005). Therefore, Schreiber and Mitch (2006a) revised the NDMA formation
pathway which involves a reaction between DMA and dichloramine to form chlorinated UDMH
(UDMH-Cl) that is further oxidized by dissolved oxygen to form NDMA (Figure 2.1, B). The
revised pathway was demonstrated to accurately predict the NDMA formation upon
chloramination. Even with preformed monochloramine, the UDMH-Cl pathway was able to
explain almost all the NDMA formation from the traces of dichloramine formed via
monochloramine disproportionation.
The NDMA formation from tertiary amines and NOM were proposed to proceed via two steps,
involving a dealkylation reaction with chloramines to form DMA (the rate-limiting step) and the
subsequent oxidation to NDMA (Chen and Valentine, 2006; Mitch and Sedlak, 2004). However,
Lee et al. (2007) observed a higher NDMA formation from tertiary amines than DMA in their
experiment, and suggested the possibility of an alternative pathway for the direct formation of
NDMA from tertiary amines not involving DMA production, but no further details have been
investigated.
The occurrence of bromide ion was found to catalyze the NDMA formation from secondary
amines upon chloramination (Choi and Valentine, 2002b). Bromide is frequently a trace
13
component in drinking water, which can be oxidized by free chlorine or chloramine to form
bromamines. They have similar chemical property with chloramines but with higher reactivity,
thus could enhance the NDMA formation. However, bromide was found to have an inhibitory
effect on NDMA formation during chloramination from tertiary amines, providing indirect
evidence that the NDMA formation from tertiary amines might follow a different pathway from
the secondary alkylamines (Chen et al., 2010). A recent study further demonstrates that the
formation of NDMA from some less reactive precursors was inhibited by bromide possibly due
to the competitive reactions leading to the production of brominated DBPs (Le Roux et al.,
2012). Moreover, Padhye et al. (2010) have discovered that some activated carbon (AC) may
promote the transformation of secondary amines to nitrosamines under ambient aerobic
conditions and in water samples, but the AC-catalyzed NDMA yield from DMA was very low
(molar conversion < 0.01 %).
Several alternative disinfectants are also found to produce NDMA. For example, NDMA was
formed during ozonation of DMA (Andrzejewski et al., 2008) and amine-based dyes (Oya et al.,
2008); specifically, high yield of NDMA was observed by ozonation of DMS, a degradation
product of the fungicide tolyfluanide (Schmidt and Brauch, 2008), and the formation was
catalyzed by the presence of bromide (von Gunten et al., 2010) (Figure 2.1, D). Free chlorine can
also produce NDMA through reaction with secondary amines but at a slower rate of
approximately an order of magnitude lower than monochloramine (Mitch and Sedlak, 2002b).
The NDMA formation pathways relevant upon drinking water disinfection are summarized in
Figure 2.1 (reproduced from Shah and Mitch, 2012).
14
Figure 2.1. NDMA formation pathways upon drinking water disinfection (figure reproduced from Shah
and Mitch, 2012)
2.1.4 Operating Factors Affecting NDMA Formation
The role of dichloramine in NDMA formation has implied the significance of pH which
determines the major chloramine species (Figure 2.2). In literature, the maximum NDMA
formation was in general observed at neutral or slightly basic pH. For example, the maximum
NDMA formation from DMA was observed between pH 7 and 8 (Kim and Clevenger, 2007;
Mitch and Sedlak, 2002b), and that from secondary wastewater effluent was found to occur at a
circumneutral pH (Mitch and Sedlak, 2002a). Krasner et al. (2010) reported that the NDMA
formation from polyDADMAC (an amine-based polymer, widely used as a coagulant) –
impacted water was higher at pH 8 than pH 9. And the maximum NDMA formation from
ranitidine was observed at pH 7.9 (Le Roux et al., 2011a).
15
Figure 2.2. Distribution diagram for chloramine species with pH (Source: Palin, 1950)
Another factor that affects the chloramine species is the Cl2:NH4-N mass ratio, which determines
the dominant chloramine species along with pH (6.5 – 8.5) typically encountered in drinking
water disinfection (Figure 2.3). Monochloramine is predominately formed when the applied ratio
is less than 5:1; dichloramine starts to form as the ratio increases, yielding a mixture of
monochloramine and dichloramine; breakpoint reaction occurs when the ratio is above 7.6:1,
resulting in the formation of free chlorine and nitrogen trichloride. Krasner et al. (2012) have
demonstrated that waters impacted by polyDADMAC and ranitidine had peak NDMA formation
at Cl2:NH4-N mass ratios of 5.1:1 or less, while waters impacted by treated wastewater had peak
or substantial NDMA formation at higher Cl2:NH4-N mass ratios approaching breakpoint,
indicating a stronger role of dichloramine in the NDMA formation from wastewater-impacted
waters. Another study compared the impact of Cl2:NH4-N mass ratio on wastewater-impacted
raw source water and treated water, finding that the peak NDMA formation for treated water was
to the left of the maximum (i.e., Cl2:NH4-N mass ratio of 5:1), while to the right of the maximum
for the raw water (Charrois and Hrudey, 2007), confirming the significance of dichloramine in
the case of wastewater-impacted waters.
16
Figure 2.3. Theoretical breakpoint curve (Source: USEPA, 1999)
Temperature does not always affect the NDMA formation, depending on the type of precursors
and other disinfection conditions. For example, the effect of temperature was found to be
minimal for the NDMA conversion from thiram (Graham et al., 1995). For polyDADMAC-
impacted waters, temperature did not affect the NDMA formation at pH 8 or 9, but more NDMA
was formed in warmer water at pH 7 (Krasner et al., 2012). Surprisingly, water impacted by
SMPs had higher apparent NDMA formation at lower temperature (Krasner et al., 2012);
however, in this study prechlorination was applied prior to chloramination, as such the higher
NDMA formation at lower temperature might be accounted for by the lower chlorine reactivity
at destroying the NDMA precursors.
In addition, the NDMA formation was found to increase with the chloramine dosage applied for
DMA (Kim and Clevenger, 2007; Mitch and Sedlak, 2002b) and treated wastewater effluents
(Farre et al., 2011b). The contact time between chloramine and precursors is also an important
parameter. Studies found that NDMA could be further formed in the distribution systems with
increasing water age (Barrett et al., 2003; Charrois and Hrudey, 2007); experiments with treated
wastewater effluent also shown that the NDMA formation after 24 hr of chloramine contact time
was almost 10 times of that formed after 2 hr (Farre et al., 2011b).
17
2.1.5 NDMA Formation Control in Drinking Water
NDMA is highly water soluble with low Henry’s Law constant and hydrophilic with low Log
Kow (OEHHA, 2006), thus it is hard to be removed by volatilization or air stripping, and it sorbs
poorly to soil, activated carbon, and other sorbents. It is also proved to be poorly removed via
reverse osmosis (RO) membranes as a small and uncharged molecule (Plumlee et al., 2008).
Currently, ultra-violet (UV) photolysis is the most commonly applied aqueous NDMA treatment
method. Several studies have demonstrated the effective elimination of NDMA using low-
pressure UV lamp (Sharpless and Linden, 2003; Stefan and Bolton, 2002), medium-pressure UV
lamp (Lee et al., 2005a, b; Plumlee et al., 2008; Sharpless and Linden, 2003), and simulated
natural sunlight (Plumlee and Reinhard, 2007). However, the effectiveness of UV photolysis
would be significantly reduced by high turbidity, color, and other water constituents due to the
competition for photons. Moreover, NDMA reformation might become a problem for UV-treated
water, since some precursors are resistant to UV, such as DMA, the major degradation product.
Application of advanced oxidation processes (AOPs) can effectively improve the destruction of
NDMA and its precursors, and thus inhibit its reformation in post-treatment storages and
distribution systems. The most common UV-AOPs for NDMA treatment are UV/H2O2 (Liang et
al., 2003; Plumlee et al., 2008; Sharpless and Linden, 2003; Swaim et al., 2008) and HV/TiO2
(Lee and Choi, 2005), proceeding via oxidation with the hydroxyl radicals (·OH) or with
reducing hydrated electrons and hydrogen atoms (Mezyk et al., 2004, 2006).
Although the UV-based treatment is effective, it requires the UV dosage approximately 250
times higher than that applied for disinfection purpose, and thus is expensive and energy
intensive. Some relatively less expensive technologies were developed to remove NDMA, such
as granular 0Fe (Gui et al., 2000; Odziemkowski et al., 2000) and H2-based reduction with metal
catalysts (Davie et al., 2006; Frierdich et al., 2008) or membrane biofilm reactor (Chung et al.,
2008). But those processes are most in the research stage and performances in larger scale
applications are still to be determined.
A more fundamental option for the control of NDMA in drinking water is the prevention of
NDMA formation by removing its precursors or selecting proper disinfection strategy. For
18
example, RO was able to remove the dissolved NDMA precursors including tertiary amines
(Mitch and Sedlak, 2004); an application of ultrafiltration (UF)-RO membrane process was
demonstrated to effectively remove more than 98.5 ± 0.5 % of NDMA precursors associated
with treated wastewater (Farre et al., 2011b); biological treatment was found effective at
removing DMA (Pehlivanoglu and Sedlak, 2006); a more recent study has demonstrated the
adsorption of NDMA precursors from blends of river water and treated wastewater effluent by
powdered and granular activated carbon (Hanigan et al., 2012). Moreover, nitrification of
wastewater to completely remove ammonia prior to hypochlorite addition was also found to
reduce NDMA formation significantly (Mitch et al., 2003). Specifically, the application of
preoxidation prior to chloramination was reported to effectively reduce the NDMA formation
from DMA and NOM, including prechlorination (Chen and Valentine, 2008; Charrois and
Hrudey, 2007; Mitch et al., 2010), O3 and O3/H2O2 (Chen and Valentine, 2008; Pisarenko et al.,
2012), ClO2 (Lee et al., 2007), ferrate (Lee et al., 2008), and KMnO4 (Chen and Valentine,
2008). Among all the options, ozone and chlorine were found to be the most cost-effective in
reducing NDMA formation (Shah et al., 2012). Especially chlorine is commonly used as the
primary disinfectant, and in practice chloramine is usually formed in situ by adding free chlorine
and ammonia separately. As discussed by Schreiber and Mitch (2005), adding Cl2 to well-mixed
ammonia may enhance NDMA formation by promoting local NHCl2 formation due to the high
Cl2: N mass ratio at the point of Cl2 addition; while the addition of free chlorine prior to
ammonia promotes chlorinated amine precursors and minimizes NHCl2 formation, both of which
inhibit the NDMA formation. The application of preformed monochloramine was also shown to
significantly reduce the NDMA formation (Farre et al., 2011a; Mitch et al., 2005). Therefore,
the proper control of the order of reagent addition, degree of mixing, as well as the free chlorine
contact time could help minimize the NDMA formation.
2.2 Pharmaceuticals and Personal Care Products (PPCPs)
2.2.1 Background
Pharmaceuticals and personal care products (PPCPs) are a group of compounds including
pharmaceutical drugs, cosmetic ingredients, food supplements, and ingredients in other personal
19
care products such as shampoos and lotions, as well as their respective metabolites and
transformation products. During recent years, a wide variety of PPCPs have been detected in the
aquatic environment all over the world with low concentrations (ng/L ~ lower µg/L), including
surface water (Calamari et al., 2003; Conley et al., 2008; Jasim et al., 2006; Kasprzyk-Hordern
et al., 2008; Kim et al., 2007; Kolpin et al., 2002, 2004; Metcalfe et al., 2003b; Peng et al.,
2008; ; Zuccato et al., 2005), groundwater (Godfrey et al., 2007), treated wastewater (Esplugas
et al., 2007; Hirsch et al., 1999; Lindqvist et al., 2005; Lishman et al., 2006; Metcalfe et al.,
2003a; Miao et al., 2004; Servos et al., 2005; Terzic et al., 2008) and treated drinking water (
Servos et al., 2007; Stackelberg et al., 2004, 2007; Ternes et al., 2002).
PPCPs are continuously introduced into the aquatic environment through different pathways.
Pharmaceuticals are excreted via feces and urine and thus are present in the domestic water; the
disposal of expired or unused medicines and other personal care products via toilets is a further
source. After passing through sewage treatment plants, the residues enter receiving waters.
Meanwhile, some PPCPs are not effectively degraded in wastewater treatment plants and
accumulated in biosolids that could subsequently be disposed of on land. The biosolids land
application could be an important source for PPCPs in soil environment (Gobel et al., 2005;
Radjenovic et al., 2009; Xia et al., 2005), and thus the terrestrial runoffs from agricultural fields
following land application of municipal biosolids (Sabourin et al., 2009; Topp et al., 2008) or
transport to adjacent surface waters via tile drainage systems (Edwards et al., 2009; Lapen et al.,
2008) could serve as highly dispersed routes of PPCPs introduction to the aquatic environment.
PPCPs are usually developed with the intention to perform a biological effect; therefore they
have many of the necessary properties to bioaccumulate and provoke effects in the aquatic
environment. Although their target effects in humans are well tested, there is limited knowledge
about their unintended effects towards non-target organisms, especially the chronic health effects
caused by the continual and long-term accumulation of trace level substances. Currently, PPCPs
are not subject to any regulated monitoring protocol, and there is no clear evidence of immediate
public health impacts. However, the unknown long-term impacts cannot be ignored, and the
removal of PPCPs during wastewater and drinking water treatment processes is desirable and of
significant concern.
20
2.2.2 Removal of PPCPs in Water Treatment Processes
Most PPCPs are refractory and hard to be effectively removed by conventional treatment
approaches. Chemical coagulation/flocculation generally can only achieve limited removal of
PPCPs (Stackelberg et al., 2007; Ternes et al., 2002; Westerhoff et al., 2005). Very few
pharmaceuticals can be partly reduced during biodegradation process (Bundy et al., 2007;
Lishman et al., 2006; Nakada et al., 2006). Disinfection (i.e., chlorine, chloramine, and ClO2)
applied in drinking water treatment mostly acts as a partial barrier for PPCPs and usually reacts
selectively with certain compounds (Huber et al., 2005b; Stackelberg et al., 2007; Westerhoff et
al., 2005), and the intensity used for UV disinfection purposes is often too low to induce
transformation in pharmaceuticals (Vieno et al., 2007).
Some advanced treatment technologies are proved to be more efficient. PPCPs with high
hydrophobicity are easy to be removed by both granular activated carbon (GAC) adsorption
(Ternes et al., 2002; Vieno et al., 2007; Yoon et al., 2003) and membrane (Nghiem et al., 2004;
Yoon et al., 2004, 2007) processes, but both technologies are highly energy and material
intensive and require the disposal of wastes. Ozonation was proved to effectively degrade a great
number of PPCPs (Hua et al., 2006; Jasim et al., 2006; Nakada et al., 2007; Vieno et al., 2007),
but the generally low mineralization might result in generation of toxic intermediates and by-
products (Andreozzi et al., 2002; Dantas et al., 2007; Esplugas et al., 2007; Rivas et al., 2009).
Moreover, ozone is a selective oxidant and thus its reactivity is product-specific. Generally,
ozone is reactive with non-aromatic double bonds, amines, thioether sulphurs, and activated
aromatic rings.
AOPs are proved capable of overcoming the difficulties encountered during ozonation, such as
the improved mineralization rate and less pH-dependent kinetics. Ozone, hydroxyl radicals (·OH)
and photolysis can occur simultaneously or competitively when applying AOPs, depending on
the combination, water matrix, and pH. Compounds which are not susceptible to ozone could be
better degraded due to the non-selective radical reactions. Studies have demonstrated the
successful application of AOPs to remove a wider range of PPCPs, including O3/H2O2 (Huber et
al., 2003; Westerhoff et al., 2005; Zwiener and Frimmel, 2000), UV/H2O2 (Andreozzi et al.,
2003a, b; Linden et al., 2007; Kim et al., 2009; Rosenfeldt et al., 2007; Vogna et al., 2002,
21
2004a, b), UV/TiO2 (Abellan et al., 2007; Addamo et al., 2005; Doll and Frimmel, 2004, 2005a,
b; Ohko et al., 2002; Mendez-Arriaga et al., 2008), and photo-assisted Fenton (Feng et al., 2005;
Molinari et al., 2007; Perez-Estradal et al., 2005; Ravina et al., 2002; Shemer et al., 2006).
However, the rate of radical formation depends greatly on the water matrix, especially the
alkalinity and total organic carbon (TOC). The matrix constituents may inhibit the degradation of
PPCPs through oxidant competitions and/or radical scavenging mechanisms (Pereira et al.,
2007a, b; Vogna et al., 2004b), while they may also enhance the degradation by increasing the
radical formation (Huber et al., 2003; von Gunten, 2003).
2.2.3 Transformation of PPCPs upon Water Treatment Processes
Under most circumstances, the removal of PPCPs during the treatment processes does not
necessarily mean complete mineralization. Instead, the PPCPs are transformed into a number of
unknown intermediates, and their transformation upon different treatment processes remains
largely unknown. One major concern is that in some cases the transformation products may pose
higher toxicity than the parent compounds. For example, the photo-transformation products of
naproxen were found to be more toxic than the parent compound (Isidori et al., 2005). The UV-
AOP treatment of bezafibrate and carbarmazepine may also lead to the formation of more toxic
intermediates (Andreozzi et al., 2002, 2004; Dantas et al., 2007). Therefore, the application of
advanced treatment technologies to remove PPCPs needs to be evaluated carefully.
In particular, there is very limited data regarding the transformation of PPCPs during disinfection
process. A few studies investigated the degradation of PPCPs upon chlorination. Compared with
ozone and AOPs, chlorine is a weaker oxidant and typically causes small modifications in the
parent compounds. In general, chlorine tends to react selectively with amines, reduced sulfur
moieties or activated aromatic amines (Deborde and von Gunten, 2008). Pinkston and Sedlak
(2004) reported that amine-containing pharmaceuticals reacted rapidly with hypochlorous acid to
form chlorinated amines, such as indometacine, acetaminophen, propranolol, naproxen, and
gemfibrozil; while the reaction rates with chloramines were significantly slower for these
compounds and no significant transformation was observed. Krkosek et al. (2008) have identified
18 and 2 chlorinated byproducts for naproxen and genfibrozil, respectively. Buth (2009) has
22
identified four major products of cimetidine upon chlorination and proposed the reaction
pathways, and two of the chlorinated products were estimated to have lower predicted no-effect
concentrations than cimetidine. However, studies on the transformation of PPCPs upon
chloramine disinfection or post-chloramination following preoxidation are very limited.
2.3 Formation of Nitrosamines from PPCPs
PPCPs are usually poorly removed at water treatment plants because conventional treatment
technologies are not specifically designed to remove them at trace level, and thus might pass
through the treatment train and come into contact with the disinfectant. Theoretically, any
compounds that contain the amine group in their structures may form the corresponding
nitrosamines upon chloramination, while there are a great number of PPCPs with amine groups,
such as the histamine antagonists and some macrolide antibiotics. Some early studies have
reported the nitrosation of amine drugs in stomachs of tested animals to form nitrosamines, and a
lot of these drugs are no longer used in most countries because of this (Andrews et al., 1980;
Lijinksy and Taylor, 1977; Mergens et al., 1979). Pharmaceuticals first came into attention as
potential NDMA precursors in drinking water when ranitidine was demonstrated to convert into
NDMA at high conversion rate during chloramination (Schmidt et al., 2006). As such, it is very
likely that other amine-based PPCPs might also contribute to the formation of nitrosamines
during drinking water disinfection.
Moreover, PPCPs in the aquatic environment may undergo biotic and/or abiotic transformation.
However, as long as the amine groups remain in the structure, the transformation products still
retain the nitrosamine formation potential. For example, ranitidine was transformed into two
major environmental metabolites under simulated solar irradiation in water, both of which still
contain the DMA functional group in their structures (Isidori et al., 2009). The photo-
degradation products of diphenhydramine, a commonly used histamine antagonist, were also
found to retain the DMA group (Chen et al., 2009). In addition, chloramine is typically used as a
secondary disinfectant following the primary disinfection, usually a preoxidation process such as
chlorine, UV, or ozone. The preoxidation may partially modify the PPCPs, yielding
transformation products that may or may not react with the subsequent chloramine to form
23
nitrosamines. In worse scenarios, some treatment plants may turn on the UV/H2O2, increase
ozone dosage, or apply other AOP systems in order to overcome periodical problems such as
taste and odor. These stronger oxidation processes could further destruct the PPCPs and even
release the amine groups that could directly contribute to the formation of nitrosamines upon
post-chloramination. For instances, tetracycline was found to easily decompose under UV/TiO2
treatment and release DMA (Addamo et al., 2005), and deamination reaction was observed for
macrolide antibiotic clarithromycin upon ozonation (Lange et al., 2006). Therefore, it is
suggested that amine-based pharmaceuticals and their breakdown products might be part of the
NDMA precursor pool (Krasner, 2009).
As such, although the presence of PPCPs in drinking water sources has been suggested to be
inconsequential due to their low concentrations, it might be linked to potential adverse health
effects via the transformation into carcinogenetic DBPs such as nitrosamines in finished drinking
water.
2.4 Research Questions and Gaps
Currently, most of the studies on NDMA and the factors affecting NDMA formation used the
model precursor DMA and NOM, while they cannot always explain all the NDMA formation
detected based on their yields, suggesting that there are other as yet unknown precursors. The
association of NDMA formation with EfOM has indicated the anthropogenic sources in the
“precursor pool”, but very few specific compounds have been identified and studied separately.
Limited literature has indicated the possibility of amine-based PPCPs as potential nitrosamine
precursors, thus it is worth further investigating the level of nitrosamine conversion from amine-
based PPCPs upon drinking water disinfection. Specifically, most research employed DMA at a
relatively high concentration because of its low NDMA yield, or studied with NOM at the
concentration comparable with the chloramine dosage applied, while PPCPs are typically present
at concentrations several orders of magnitude lower than the applied chloramine dosage (i.e.,
ng/L or lower µgL vs. mg/L); therefore the findings obtained using DMA and NOM may not
always apply to PPCPs. Moreover, previous studies on NDMA mostly considered single-
precursor scenario, i.e., using selected model compound or NOM as the only precursor or
24
precursor pool, but the possible interactions between targeted precursors and the water matrix
components have not been considered when evaluating the NDMA formation. As such, research
is needed to investigate the NDMA formation from PPCPs with the presence of water matrix
components especially NOM.
In addition, knowledge about the reactivity and chemistry is essential in predicting the NDMA
formation from selected precursors, while data regarding the reaction kinetics are largely lacking,
especially in real water matrices. In the literature, some kinetic models have been developed for
the prediction of NDMA formation from DMA (Choi and Valentine, 2002a; Kim and Clevenger,
2007), NOM (Chen and Valentine, 2006), and EfOM (Chen and Westerhoff, 2010); however,
these models were based on comparable concentrations of precursors and chloramines, thus may
not apply to PPCPs which are usually present at much lower concentrations relative to the
chloramine concentrations in real samples. Krasner et al. (2010) have monitored the NDMA
formation over time from ranitidine under different conditions, but only conducted limited
experiments in deionized water. Therefore, more kinetic study is needed to look into the NDMA
formation kinetics from amine-based PPCPs in real water matrices.
Moreover, it is important to understand the factors that may affect the NDMA formation in order
to effectively control its level in treated drinking water. The literature has suggested that
different precursors may respond differently to certain operational parameters, such as the case
that the peak NDMA formation occurred at different Cl2:NH3-N mass ratios from
polyDADMAC-impacted waters and wastewater-impacted waters (Krasner et al., 2012).
Another example is the preoxidation of precursors for NDMA formation control. Studies have
demonstrated that oxidation of precursors did not necessarily lead to the reduction in NDMA
formation. Insufficient prechlorination may increase NDMA formation from NOM (Chen and
Valentine, 2006, 2008; Shah et al., 2012); advanced oxidation like UV/ H2O2 pretreatment
increased the NDMA formation from polyDADMAC, possibly due to the release of amine
monomers (Harvey, 2009); the oxidation products of pharmaceutical tramadol by UV and UV/
H2O2 in general have higher NDMA formation potentials than the parent compound (Radjenovic
et al., 2012). Right now, information on PPCP-based NDMA precursors is largely lacking, and
thus further research is needed to investigate the factors that can affect the NDMA formation
25
from amine-based PPCPs, such as pH, Cl2:NH3-N mass ratio, as well as the application of
preoxidation. Additionally, most research that examine the impact factors only measured the
NDMA formation potential after a long contact time, while very little is known in terms of the
reaction kinetics under different reaction conditions. Knowledge on the reaction kinetics can help
predicting the NDMA concentration along the distribution systems and help determining the best
disinfection strategy at the water treatment plants.
2.5 References
Abellan, M. N., Bayarri, B., Gimenez, J., Costa, J., 2007. Photocatalytic degradation of
sulfamethoxazole in aqueous suspension of TiO2. Appl. Catal. B Environ. 74, 233-241
Addamo, M., Augugliaro, V., di Paola, A., Garcia-Lopez, E., Loddo, V., Marci, G., Palmisano,
L., 2005. Removal of drugs in aqueous systems by photoassisted degradation. J. Appl.
Electrochem. 35, 765-774
Andreozzi, R., Marotta, R., Pinto, G., Pollio. A., 2002. Carbamazepine in water: persistence in
the environment, ozonation treatment and preliminary assessment on algal toxicity. Water Res.
36, 2869-2877
Andreozzi, R., Caprio, V., Marotta, R., Radovnikovic, A., 2003a. Ozonation and H2O2/UV
treatment of clofibric acid in water: a kinetic investigation. J. Hazard. Mater. 103,223-246
Andreozzi, R., Caprio, V., Marotta, R., Vogna, D., 2003b. Paracetamol oxidation from aqueous
solutions by means of ozonation and H2O2/UV system. Water Res. 37, 993-1004
Andreozzi, R., Campanella, L., Fraysse, B., Garric, J., Gonnella, A., Giudice, R.L., Marotta, R.,
Pinto, G., Pollio, A., 2004. Effects of advanced oxidation processes (AOPs) on the toxicity of a
mixture of pharmaceuticals. Water Sci. Technol. 50, 23-28
Andrews, A.W., Fornwald, J.A., Lijinsky, W., 1980. Nitrosation and mutagenicity of some
amine drugs. Toxicol. Appl. Pharmacol.52, 237-244
26
Andrzejewski, P., Kasprzyk-Hordern, B., Nawrocki, J., 2008. N-nitrosodimethylamine (NDMA)
formation during ozonation of dimethylamine-containing waters. Water Res. 42, 863-870
Barrett, S., Hwang, C., Guo, Y., Andrews, S.A., Valentine, R., 2003. Occurrence of NDMA in
drinking water: a North American survey, 2001-2002. In: Proceedings of the American Water
Works Association's Annual Conference, Anaheim, CA, June 15-19, 2003
Blute, N., Russell, C., Chowdhury, Z., Wu, X., Via, S., 2010. Nitrosamine occurrence in the U.S.
– analysis and interpretation of UCMR2 data. In: Proceedings of the AWWA Water Quality
Technology Conference, Savannah, GA, November 14-18, 2010
Bundy, M.M., Doucette, W.J., McNeill, L., Ericson, J.F., 2007. Removal of pharmaceuticals and
related compounds by a bench-scale drinking water treatment system. J. Water Supply Res. T.
56, 105-115
Buth, J.M., 2009. Transformation of pharmaceutical and personal care product pollutants by
wastewater chlorine disinfection and aqueous photolysis. PhD thesis. University of Minnesota,
Minnesota
Calamari, D., Zuccato, E., Castiglioni, S., Bagnati, R., Fanelli, R., 2003. Strategic survey of
therapeutic drugs in the Rivers Po and Lambro in Northern Italy. Environ. Sci. Technol. 37,
1241-1248
California Department of Public Health (CDPH), 2011. A brief history of NDMA findings in
drinking water. (Last update: January 4, 2011). Retrieved from
http://www.cdph.ca.gov/certlic/drinkingwater/Pages/NDMAhistory.aspx (accessed June 26,
2012)
Chang, H., Chen, C., Wang, G., 2011. Identification of potential nitrogenous organic precursors
for C-, N-DBPs and characterization of their DBPs formation. Water Res. 45, 3753-3764
27
Charrois, J.W.A., Arend, M.W., Froese, K.L., Hrudey, S.E., 2004. Detecting N-Nitrosamines in
drinking water at nanogram per liter levels using ammonia positive chemical ionization. Environ.
Sci. Technol. 38, 4835-4841
Charrois, J.W.A., Boyd, J.M., Froese, K.L., Hrudey, S.E., 2007. Occurrence of N-nitrosamines in
Alberta public drinking-water distribution systems. J. Environ. Eng. Sci. 6, 103-114
Charrois, J.W.A., Hrudey, S.E., 2007. Breakpoint chlorination and free-chlorine contact time:
implications for drinking water N-nitrosodimethylamine concentrations. Water Res. 41, 674-682
Chen, B., Westerhoff, P., 2010. Predicting disinfection by-product formation potential in water.
Water Res. 44, 3755-3762
Chen, Y., Hu, C., Hu, X., Qu, J., 2009. Indirect photodegradation of amine drugs in aqueous
solution under simulated sunlight. Environ. Sci. Technol. 43, 2760-2765
Chen, Z., Valentine, R.L., 2006. Modeling the formation of N-Nitrosodimethylamine (NDMA)
from the reaction of Natural Organic Matter (NOM) with monochloramine. Environ. Sci.
Technol. 40, 7290-7297
Chen, Z., Valentine, R.L., 2007. Formation of N-Nitrosodimethylamine (NDMA) from humic
substances in natural water. Environ. Sci. Technol. 41, 6059-6065
Chen, Z., Valentine, R.L., 2008. The influence of the pre-oxidation of natural organic matter on
the formation of N-Nitrosodimethylamine (NDMA). Environ. Sci. Technol. 42, 5062-5067
Chen, Z., Yang, L., Zhai, X., Zhao, S., Li, A., Shen, J., 2010. N-nitrosamine formation during
chlorination/chloramination of bromide-containing water. Wa. Sci. Technol. 10, 462-471
Chen, W.H., Young, T.M., 2008. NDMA formation during chlorination and chloramination of
aqueous diuron solutions. Environ. Sci. Technol. 42, 1072-1077
28
Choi, J.H., Valentine, R.L., 2002a. A kinetic model of N-nitrosodimethylamine (NDMA)
formation during water chlorination/chloramination. Water Sci. Technol. 46, 65-71
Choi, J.H., Valentine, R.L., 2002b. Formation of N-nitrosodimethylamine (NDMA) from
reaction of monochloramine: a new disinfection by-product. Water Res. 36, 817-824
Choi, J.H., Valentine, R.L., 2003. N-Nitrosodimethylamine Formation by Free-Chlorine-
Enhanced Nitrosation of Dimethylamine. Environ. Sci. Technol. 37, 4871-4876
Chung, J., Ahn, C.H., Chen, Z., Rittman, B.E., 2008. Bio-reduction of N- nitrosodimethylamine
(NDMA) using a hydrogen-based membrane biofilm reactor. Chemosphere 70, 516-520
Conley, J.M., Symes, S.J., Schorr, M.S., Richards, S.M., 2008. Spatial and temporal analysis of
pharmaceutical concentrations in the upper Tennessee River basin. Chemosphere 73, 1178-1187
Dantas, R.F., Canterino, M., Marotta, R., Sans, C., Esplugas, S., Andreozzi, R., 2007.
Bezafibrate removal by means of ozonation: Primary intermediates, kinetics, and toxicity
assessment. Water Res. 41, 2525-2532
Davie, M.G., Reinhard, M., Sharpley, J.R., 2006. Metal-catalyzed reduction of N-
Nitrosodimethylamine with hydrogen in water. Environ. Sci. Technol. 40, 7329-7335
Deborde, M., von Gunten, U., 2008. Reactions of chlorine with inorganic and organic
compounds during water treatment – kinetics and mechanisms: a critical review. Water Res. 42,
13-51
Doll, T.E., Frimmel, F.H., 2004. Kinetic study of Photocatalytic degradation of carbamazepine,
clofibric acid, iomeprol and iopromide assisted by different TiO2 materials – determination of
intermediates and reaction pathways. Water Res. 38, 955-964
29
Doll, T.E., Frimmel, F.H., 2005a. Cross-Flow Microfiltration with Periodical Back-Washing for
Photocatalytic Degradation of Pharmaceutical and Diagnostic Residues - Evaluation of the Long-
Term Stability of the Photocatalytic Activity of TiO2. Water Res. 39, 847-854
Doll, T.E., Frimmel, F.H., 2005b. Photocatalytic degradation of carbamazepine, clofibric acid
and iomeprol with P25 and Hombikat UV100 in the presence of natural organic matter (NOM)
and other organic water constituents. Water Res. 39, 403-411
Dotson, A., Westerhoff, P., Krasner, S.W., 2007. Nitrosamine Formation from Natural Organic
Matter Isolates and Sunlight Photolysis of Nitrosamines. In: Proceedings of the AWWA Annual
Conference and Exposition, Toronto, ON, June 24-28, 2007
Edwards, M., Topp, E., Metcalfe, C.D., Li, H., Gottschall, N., Bolton, P., Curnoe, W., Payne, M.,
Beck, A., Kleywegt, S., Lapen, D.R., 2009. Pharmaceutical and personal care products in tile
drainage following surface spreading and injection of dewatered municipal biosolids to an
agricultural field. Sci. Total Environ. 407, 4220-4230
EPA Integrated Risk Information System (IRIS), 1993. N-Nitrosodimethylamine; CASRN 62-
75-9. Retrieved from www.epa.gov/iris/subst/0045.htm (accessed June 26, 2012)
Esplugas, S., Bila, D.M., Krause, L.G.T., Dezotti, M., 2007. Ozonation and advanced oxidation
technologies to remove endocrine disrupting chemicals (EDCs) and pharmaceuticals and
personal care products (PPCPs) in water effluents (Review Article, Jan. 2000 ~ May 2007). J.
Hazard. Mater. 149, 631-642
Farre, M.J., Doderer, K., Hearn, L., Poussade, Y., Keller, J., Gernjak, W., 2011a. Understanding
the operational parameters affecting NDMA formation at advanced water treatment plants. J.
Hazard. Mater. 185, 1575-1581
Farre, M.J., Keller, J., Holling, N., Poussade, Y., Gernjak, W., 2011b. Occurrence of N-
nitrosodimethylamine precursors in wastewater treatment plant effluent and their fate during
ultrafiltration-reverse osmosis membrane treatment. Wa. Sci. Technol. 63, 605-612
30
Feng, X., Ding, S., Tu, J., Wu, F., Deng, N., 2005. Degradation of estrone in aqueous solution by
photo-Fenton system. Sci. Total Environ. 345, 229-237
Fournier, D., Hawari, J., Streger, S.H., McClay, K., Hatzinger, P.B., 2006. Biotransformation of
N- nitrosodimethylamine by Pseudomonas mendocina KR1. Appl. Environ. Microb. 72, 6693-
6698
Frierdich, A.J., Sharpley, J.R., Strathmann, T.J., 2008. Rapid reduction of N-nitrosamine
disinfection byproducts in water with hydrogen and porous nickel catalysts. Environ. Sci.
Technol. 42, 262-269
Gerecke, A.C., Sedlak, D.L., 2003. Precursors of N- Nitrosodimethylamine in natural waters.
Environ. Sci. Technol. 37, 1331-1336
Gobel, A., Thomsen, A., McArdell, C.S., Joss, A., Giger, W., 2005. Occurrence and sorption
behavior of sulfonamides, macrolides, and trimethoprim in activated sludge treatment. Environ.
Sci. Technol. 39, 3981-3989
Godfrey, E., Woessner, W.W., Benotti, M., 2007. Pharmaceuticals in On-Site Sewage Effluent
and Ground Water, Western Montana. Ground Water 45 (3), 263-271
Graham, J.E., Andrews, S.A., Farquhar, G.J., Meresz, O., 1995. Factors affecting NDMA
formation during drinking water treatment. In: Proceedings of the AWWA Water Quality
Technology Conference, New Orleans, LA, November 12-16, 1995
Gui, L., Gillham, R.W., Odziemkowski, M.S., 2000. Reduction of N- nitrosodimethylamine with
granular iron and nickel-enhanced iron. 1. Pathways and Kinetics. Environ. Sci. Technol. 34,
3489-3494
Hanigan, D., Zhang, J., Herckes, P., Krasner, S.W., Chen, C., Westerhoff, P., 2012. Adsorption
of N-Nitrosodimethylamine precursors by powdered and granular activated carbon. Environ. Sci.
Technol., doi: 10.1021/es302922w
31
Harvey, M. 2009. The impacts of UV direct photolysis and UV/H2O2 advanced oxidation
processes on the formation of nitrosamines and organic chloramines from subsequent
chlor(am)ination. Master’s thesis, University of Toronto, Ontario
Health Canada, 2010. Guideline Technical Document on N-nitrosodimethylamine (NDMA) in
drinking water for public comment. Retrieved from http://www.hc-sc.gc.ca/ewh-
semt/consult/_2010/ndma/draft-ebauche-eng.php#a3 (accessed June 26, 2012)
Hirsch, R., Ternes, T.A., Haberer, K., Kratz, K., 1999. Occurrence of antibiotics in the aquatic
environment. Sci. Total Environ. 225, 109-118
Hua, W., Bennett, E.R., Letcher, R.J., 2006. Ozone treatment and the depletion of detectable
pharmaceuticals and atrazine herbicide in drinking water sourced from the upper Detroit River,
Ontario, Canada. Water Res. 40, 2259-2266
Huber, M.M., Canonica, S., Park, G., von Gunten, U., 2003. Oxidation of Pharmaceuticals
during Ozonation and Advanced Oxidation Processes. Environ. Sci. Technol. 37, 1016-1024
Huber, M.M., Gobel, A., Joss, A., Hermann, N., Loffler, D., McArdell, C.S., Ried, A., Siegrist,
H., Ternes, T.A., von Gunten, U., 2005a. Oxidation of pharmaceuticals during ozonation of
municipal wastewater effluents: a pilot study. Environ. Sci. Technol. 39, 4290-4299
Huber, M.M., Korhonen, S., Ternes, T.A., von Gunten, U., 2005b. Oxidation of pharmaceuticals
during water treatment with chlorine dioxide. Water Res. 39, 3607-3617
Isidori, M., Lavorgna, M., Nardelli, A., Parrella, A., Previtera, L., Rubino, M., 2005. Ecotoxicity
of naproxen and its phototransformation products. Sci. Total Environ. 348, 93-101
Isidori, M., Parrella, A., Pistillo, P., Temussi, F., 2009. Effects of ranitidine and its
photoderivatives in the aquatic environment. Environ. Int. 35, 821-825
32
Jasim, S.Y., Irabelli, A., Yang, P., Ahmed, S., Schweitzer, L., 2006. Presence of pharmaceuticals
and pesticides in Detroit River water and the effect of ozone on removal. Ozone Sci. Eng. 28,
415-423
Jobb, D.B., Hunsinger, R.B., Meresz, O., Taguchi, V., 1994. Removal of N-
nitrosodimethylamine from Ohsweken (Six Nations) water supply final report. Ontario Ministry
of Environment and Energy, Toronto, ON
Kasprzyk-Hordern, B., Dinsdale, R.M., Guwy, A.J., 2008. The occurrence of pharmaceuticals,
personal care products, endocrine disruptors and illicit drugs in surface water in South Wales,
UK. Water Res. 42, 3498-3518
Kemper, J.M., Walse, S.S., Mitch, W.A., 2010. Quaternary amines as nitrosamine precursors: a
role for consumer products? Environ. Sci. Technol. 44, 1224-1231
Kim, S.D., Cho, J., Kim, I.S., Vanderford, B.J., Snyder, S.A., 2007. Occurrence and removal of
pharmaceuticals and endocrine disruptors in South Korean surface, drinking, and waste waters.
Water Res. 41, 1013-1021
Kim, J., Clevenger, T.E., 2007. Prediction of N-nitrosodimethylamine (NDMA) formation as a
disinfection by-product. J. Hazard. Mater. 145, 270-276
Kim, I., Yamashita, N., Tanaka, H., 2009. Performance of UV and UV/H2O2 processes for the
removal of pharmaceuticals detected in secondary effluent of a sewage treatment plant in Japan.
J. Hazard. Mater. 166, 1134-1140
Kohut, K.D., Andrews, S.A., 2003. Polyelectrolyte age and N-nitrosodimethylamine formation
in drinking water treatment. Water Qual. Res. J. Can. 38 (4), 719-735
Kolpin, D.W., Furlong, E.T., Meyer, M.T., Thurman, E.M., Zaugg, S.D., Barber, L.B., Buxton
H.T., 2002. Pharmaceuticals, Hormones, and Other Organic Wasterwater Contaminants in U.S.
Streams, 1999 - 2000: A National Reconnaissance. Environ. Sci. Technol. 36, 1202-1211
33
Kolpin, D.W., Skopec, M., Meyer, M.T., Furlong, E.T., Zaugg, S.D., 2004. Urban contribution
of pharmaceuticals and other organic wastewater contaminants to streams during differing flow
conditions. Sci. Total Environ. 328, 119-130
Krasner, S.W., Garcia, E.A., Dale, M.S., Labernik, S.M., Yun, T.I., 2008. Source and removal of
NDMA precursors. In: Proceedings of the AWWA Annual Conference and Exposition, Atlanta,
GA, June 8-12, 2008
Krasner, S.W., 2009. The formation and control of emerging disinfection by-products of health
concern. Phil Trans. R Soc. A 367, 4077-4095
Krasner, S.W., Dale, M.S., Lee, C.F.T., Garcia, E.A., Wong, T.M., Mitch, W., Von Gunten, U.,
2010. Difference in reactivity and chemistry of NDMA precursors from treated wastewater and
from polyamine polymers. In: Proceedings of the AWWA Water Quality Technology
Conference, Savannah, GA, November 14-18, 2010
Krasner, S.W., Lee, C.F.T., Mitch, W.A., von Gunten, U., 2012. Development of a bench-scale
test to predict the formation of nitrosamines. Denver, CO, Water Research Foundation Report
Krauss, M., Longree, P., van Houtte, E., Cauwenberghs, J., Hollender, J., 2010. Assessing the
fate of nitrosamine precursor in wastewater treatment by physicochemical fractionation. Environ.
Sci. Technol. 44, 7871-7877
Krkosek, W.H., Koziar, S.A., Peldszus, S., Huck, P.M., Gagnon, G.A., 2008. Formation and fate
of chlorinated pharmaceutically active compounds during wastewater disinfection. In:
Proceedings of the AWWA Annual Conference and Exposition, Atlanta, GA, June 8-12, 2008
Lange, F., Cornelissen, S., Kubac, D., Sein, M.M., von Sonntag, J., Hannich, C.B., Golloch, A.,
Heipieper, H.J., Moder, M., von Sonntag, C. 2006. Degradation of macrolide antibiotics by
ozone: A mechanistic case study with clarithromycin. Chemosphere 65, 17-23
34
Lapen, D.R., Topp, E., Metcalfe, C.D., Li, H., Edwards, M., Gottschall, N., Bolton, P., Curnoe,
W., Payne, M., Beck, A., 2008. Pharmaceutical and personal care products in tile drainage
following land application of municipal biosolids. Sci. Total Environ. 399, 50-65
Le Roux, J., Gallard, H., Croue, J., 2011a. Chloramination of nitrogenous contaminants
(pharmaceuticals and pesticides): NDMA and halogenated DBPs formation. Water Res. 45,
3164-3174
Le Roux, J., Gallard, H., Croue, J., 2011b. Formation of NDMA by chloramination of
nitrogenous contaminants: potential role of bromide and dissolved oxygen. In: Proceedings of
the AWWA Water Quality Technology Conference, Phoenix, AZ, November 13-17, 2011
Le Roux, J., Gallard, H., Croue, J., 2012. Formation of NDMA and halogenated DBPs by
chloramination of tertiary amines: The influence of bromide ion. Environ. Sci. Technol. 46,
1581-1589
Lee, C., Choi, W., 2005. Photocatalytic degradation of N- nitrosodimethylamine: mechanism,
product distribution, and TiO2 surface modification. Environ. Sci. Technol. 39, 6800-6807
Lee, C., Choi, W., Kim, Y.G., Yoon, J., 2005a. UV photolytic mechanism of N-
nitrosodimethylamine in water: dual pathways to methylamine versus dimethylamine. Environ.
Sci. Technol. 39, 2101-2106
Lee, C., Choi, W., Yoon, J., 2005b. UV photolytic mechanism of N-nitrosodimethylamine in
water: roles of dissolved oxygen and solution pH. Environ. Sci. Technol. 39, 9702-9709
Lee, C., Schimidt, C., Yoon, J., von Gunten, U., 2007. Oxidation of N-nitrosodimethylamine
(NDMA) precursors with ozone and chlorine dioxide: kinetics and effect on NDMA formation
potential. Environ. Sci. Technol. 41, 2056-2063
35
Lee, C., Lee, Y., Schmidt, C., Yoon, J., Von Gunten, U., 2008. Oxidation of suspected N-
nitrosodimethylamine (NDMA) precursors by ferrate (VI): Kinetics and effect on the NDMA
formation potential of natural waters. Water Res. 42, 433-441
Liang, S., Min, J.H., Davis, M.K., Green, J.F., Remer, D.S., 2003. Use of pulsed-UV processes
to destroy NDMA. J. Am. Water Works Ass. 95, 121-131
Lijinsky, W., Taylor, H.W., 1977. Feeding tests in rats on mixtures of nitrite with secondary and
tertiary amines of environmental importance. Food Chem. Toxicol. 15, 269-274
Linden, K.G., Rosenfeldt, E.J., Kullman, S.W., 2007. UV/H2O2 degradation of endocrine-
disrupting chemicals in water evaluated via toxicity assays. Water Sci. Technol. 55, 313-319
Lindqvist, N., Tuhkanen, T., Kronberg, L., 2005. Occurence of acidic pharmaceuticals in raw
and treated sewages and in receiving waters. Water Res. 39, 2219-2228
Lishman, L., Smyth, S.A., Sarafin, K., Kleywegt, S., Toito, J., Peart, T., Lee, B., Servos, M.,
Beland, M., Seto, P., 2006. Occurrence and reductions of pharmaceuticals and personal care
products and estrogens by municipal wastewater treatment plants in Ontario, Canada. Sci. Total
Environ. 367, 544-558
Lv, C.L., Liu, Y.D., Zhong, R., Wang, Y., 2007. Theoretical studies on the formation of N-
nitrosodimethylamine. J. Mol. Struc. –THEOCHEM 802, 1-6
Lv, C.L., Liu, Y.D., Zhong, R., 2009. Theoretical investigation of N-nitrosodimethylamine
formation from dimethylamine nitrosation catalyzed by carbonyl compounds. J. Phys. Chem.
113, 713-718
Mangino, M.M., Scanlan, R.A., 1985. Nitrosation of the alkaloids hordenine and gramine,
potential precursors of N-nitrosodimethylamine in barley malt. J. Agr. Food Chem. 33, 699-705
36
Massachusetts Department of Environmental Protection (MassDEP), 2004. Current Regulatory
Limit: n-Nitrosodimethylamine (NDMA), CASRN 62759. Retrieved from
http://www.mass.gov/dep/water/drinking/standards/ndma.htm (accessed June 26, 2012)
Mendez-Arriaga, F., Esplugas, S., Gimenez, J., 2008. Photocatalytic degradation of non-steroidal
anti-inflammatory drugs with TiO2 and simulated solar irradiation. Water Res. 42, 585-594
Mergens, W.J., Vane, F.M., Tannenbaum, S.R., Green, L., Skipper, P.L., 1979. In vitro
nitrosation of methapyrilene. J. Pharm. Sci. 68, 827-832
Metcalfe, C.D., Koenig, B.G., Bennie, D.T., Servos, M., Ternes, T.A., Hirsch, R., 2003a.
Occurrence of neutral and acidic drugs in the effluents of Canadian sewage treatment plants.
Environ. Toxicol. Chem. 22, 2872-2880
Metcalfe, C.D., Miao, X.S., Koenig, B.G., Struger, J., 2003b. Distribution of acidic and neutral
drugs in surface waters near sewage treatment plants in the lower Great Lakes, Canada. Environ.
Toxicol. Chem. 22 (12), 2881-2889
Mezyk, S.P., Cooper, W.J., Madden, K.P., Bartels, D.M., 2004. Free radical destruction of N-
nitrosodimethylamine in water. Environ. Sci. Technol. 38, 3161-3167
Mezyk, S.P., Ewing, D.B., Kiddle, J.J., Madden, K.P., 2006. Kinetics and mechanisms of the
reactions of hydroxyl radicals and hydrated electrons with nitrosamines and nitramines in water.
J. Phys. Chem. A. 110, 4732-4737
Miao, X.S., Bishay, X.S., Chen, M., Metcalfe, C.D., 2004. Occurrence of antimicrobials in the
final effluents of wastewater treatment plants in Canada. Environ. Sci. Technol. 38, 3533-3541
Mitch, W.A., Schreiber, I.M., 2008. Degradation of tertiary alkylamines during
chlorination/chloramination: implications for formation of aldehydes, nitriles, halonitroalkanes,
and nitrosamines. Environ. Sci. Technol. 42, 4811-4817
37
Mitch, W.A., Sedlak, D.L., 2002a. Factors controlling nitrosamine formation during wastewater
chlorination. Wa. Sci. Technol. 2 (3), 191-198
Mitch, W.A., Sedlak, D.L., 2002b. Formation of N- Nitrosodimethylamine (NDMA) from
dimethylamine during chlorination. Environ. Sci. Technol. 36, 588-595
Mitch, W.A., Sedlak, D.L., 2004. Characterization and fate of N-nitrosodimethylamine
precursors in municipal wastewater treatment plants. Environ. Sci. Technol. 38, 1445-1454
Mitch, W.A., Sharp, J.O., Trussell, R.R., Valentine, R.L., Alvarez-Cohen, L., Sedlak, D.L., 2003.
N-Nitrosodimethylamine (NDMA) as a Drinking Water Contaminant: A Review. Environ. Eng.
Sci. 20, 389-404
Mitch, W.A., Oelker, G.L., Hawley, E.L., Deeb, R.A., Sedlak, D.L., 2005. Minimization of
NDMA formation during chlorine disinfection of municipal wastewater by application of pre-
formed chloramines. Environ. Eng. Sci. 22 (6), 882-890
Mitch, W.A., Krasner, S.W., Lee, C.F.T., Wong, T.M., 2010. Tradeoffs in DBP formation during
chloramination following pre-oxidation for nitrosamine control. In: Proceedings of the AWWA
Water Quality Technology Conference, Savannah, GA, November 14-18, 2010
MOE, 2003. Ontario regulation 268/03 made under the safe drinking water act, 2002. Retrieved
from http://www.e-laws.gov.on.ca/html/source/regs/english/2003/elaws_src_regs_r03268_e.htm
(accessed June 26, 2012)
Molinari, R., Argurio, P., Peorio, T., Bonaddio, F., 2007. Photo Assisted Fenton in a Batch and a
Membrane Reactor for Degradation of Drugs in Water. Separ. Sci. Technol. 42, 1597-1611
Najm, I., Trussell, R.R., 2001. NDMA formation in water and wastewater. J. Am. Water Works
Assoc. 93 (2), 92-99
38
Nakata, N., Tanishima, T., Shinohara, H., Kiri, K., Takada, H., 2006. Pharmaceutical chemicals
and endocrine disrupters in municipal wastewater in Tokyo and their removal during activated
sludge treatment. Water Res. 40, 3297-3303
Nakata, N., Shinohara, H., Murata, A., Kiri, K., Managaki, S., Sato, S., Takada, H., 2007.
Removal of selected pharmaceuticals and personal care products (PPCPs) and endocrine-
disrupting chemicals (EDCs) during sand filtration and ozonation at a municipal sewage
treatment plant. Water Res. 41, 4373-4382
Nghiem, L.D., Manis, A., Soldenhoff, K., Schafer, A.I., 2004. Estrogenic hormone removal from
wastewater using NF/RO membranes. J. Membrane Sci. 242, 37-45
NHMRC, NRMMC, 2011. Australian Drinking Water Guidelines Paper 6 National Water
Quality Management Strategy. National Health and Medical Research Council, National
Resource Management Ministerial Council, Commonwealth of Australia, Canberra.
Odziemkowski, M.S., Gui, L., Gillham, R.W., 2000. Reduction of N- nitrosodimethylamine with
granular iron and nickel-enhanced iron. 2. Mechanistic studies. Environ. Sci. Technol. 34, 3495-
3500
Office of Environmental Health Hazard Assessment (OEHHA), 2006. Public Health Goal for N-
nitrosodimethylamine and cadmium in drinking water. Retrieved from
http://www.oehha.org/water/phg/cadndma122206.html (accessed June 26, 2012)
Ohko, Y., Iuchi, K.I., Niwa, C., Tatsuma, T., Nakashima, T., Iguchi, T., Kubota, Y., Fujishima,
A., 2002. 17β-Estradiol degradation by TiO2 photocatalysis as a means of reducing estrogenic
activity. Environ. Sci. Technol. 36, 4175-4181
Oya, M., Kosaka, K., Asami, M., Kunikane, S., 2008. Formation of N-nitrosodimethylamine
(NDMA) by ozonation of dyes and related compounds. Chemosphere 73, 1724-1730
39
Radjenovic, J., Farre, M.J., Gernjak, W., 2012. Effect of UV and UV/H2O2 in the presence of
chloramines on NDMA formation potential of tramadol. Environ. Sci. Technol. 46, 8356-8364
Russell, C.G., Blute, N.K., Via, S., Wu, X., Chowdhury, Z., 2012. Nationwide assessment of
nitrosamine occurrence and trends. J. Am. Water Works Assoc. 104, 205-217
Padhye, L., Wang, P., Karanfil, T., Huang, C.H., 2010. Unexpected role of activated carbon in
promoting transformation of secondary amines to N-nitrosamines. Environ. Sci. Technol. 44,
4161-4168
Palin, A., 1950. A study of the chloro derivatives of ammonia. Water Water Eng. 54, 248-258
Park, S.H., Wei, S., Mizaikoff, B., Taylor, A.E., Favero, C., Huang, C.H., 2009. Degradation of
amine-based water treatment polymers during chloramination as N-nitrosodimethylamine
(NDMA) precursors. Environ. Sci. Technol. 43, 1360-1366
Pehlivanoglu-Mantas, E., Sedlak, S.L., 2006. Wastewater-derived dissolved organic nitrogen:
analytical methods, characterization, and effects – a review. Crit. Rev. Env. Sci. Tec. 36, 261-
285
Peng, X., Yu, Y., Tang, C., Tan, J., Huang, Q., Wang, Z., 2008. Occurence of steroid estrogens,
endocrine-disrupting phenols, and acid pharmaceutical residues in urban riverine water of the
Pearl River Delta, South China. Sci. Total Environ. 397, 158-166
Pereira, V.J., Linden, K.G., Weinberg, H.S., 2007a. Evaluation of UV irradiation for photolytic
and oxidative degradation of pharmaceutical compounds in water. Water Res. 41, 4413-4423
Pereira, V.J., Weinberg, H.S., Linden, K.G., Singer, P.C., 2007b. UV degradaiont kinetics and
modeling of pharmaceutical compounds in laboratory grade and surface water via direct and
indirect photolysis at 254 nm. Environ. Sci. Technol. 41, 1682-1688
40
Perez-Estrada, L.A., Malato, S., Gemjak, W., Aguera, A., Thurman, E.M., Ferrer, I., Fernandez-
Alba, A.R., 2005. Photo-Fenton degradation of diclofenac: identification of main intermediates
and degradation pathway. Environ. Sci. Technol. 39, 8300-8306
Pinkston, K., Sedlak, D.L., 2004. Transformation of aromatic ether- and amine-containing
pharmaceuticals during chlorine disinfection. Environ. Sci. Technol. 38, 4019-4025
Pisarenko, A.N., Stanford, B.D., Yan, D., Gerrity, D., Snyder, S.A., 2012. Effects of ozone and
ozone/peroxide on trace organic contaminants and NDMA in drinking water and water reuse
applications. Water Res. 46, 316-326
Plumlee, M.H., Reinhard, M., 2007. Photochemical attenuation of N- nitrosodimethylamine
(NDMA) and other nitrosamines in surface water. Environ. Sci. Technol. 41, 6170-6176
Plumlee, M.H., Lopez-Mesas, M., Heidlberger, A., Ishida, K.P., Reinhard, M., 2008. N-
nitrosodimethylamine (NDMA) removal by reverse osmosis and UV treatment and analysis via
LC-MS/MS. Water Res. 42, 347-355
Poocharoen, B., Barbour, J.F., Libbey, L.M., Scanlan, R.A., 1992. Precursors of N-
nitrosodimethylamine in malted barley. 1. Determination of hordenine and gramine. J. Agr. Food
Chem. 40, 2216-2221
Radjenovic, J., Petrovic, J.M., Barcelo, D., 2009. Fate and distribution of pharmaceuticals in
wastewater and sewage sludge of the conventional activated sludge (CAS) and advanced
membrane bioreactor (MBR) treatment. Water Res. 43, 831-841
Ravina, M., Campanella, L., Kiwi, J., 2002. Accelerated mineralization of the drug Diclofenac
via Fenton reactions in a concentric photo-reactor. Water Res. 36, 3553-3560
Rivas, J., Gimeno, O., Encinas, A., Beltran, F., 2009. Ozonation of the pharmaceutical
compound ranitidine: Reactivity and kinetic aspects. Chemosphere 76, 651-656
41
Rosenfeldt, E.J., Chen, P.J., Kullman, S., Linden, K.G., 2007. Destruction of estrogenic activity
in water using UV advanced oxidation. Sci. Total Environ. 377, 105-113
Sabourin, L., Beck, A., Duenk, P.W., Kleywegt, S., Lapen, D.R., Li, H., Metcalfe, C.D., Payne,
M., Topp, E., 2009. Runoff of pharmaceuticals and personal care products following application
of dewatered municipal biosolids to an agricultural field. Sci. Total Environ. 407, 4596-4604
Schmidt, C.K., Brauch, H.J., 2008. N,N-Dimethylsulfamide as precursor for N-
Nitrosodimethylamine (NDMA) formation upon ozonation and its fate during drinking water
treatment. Environ. Sci. Technol. 42, 6340-6346
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
Schreiber, I.M., Mitch, W.A., 2005. Influence of the order of reagent addition on NDMA
formation during chloramination. Environ. Sci. Technol. 39, 3811-3818
Schreiber, I.M., Mitch, W.A., 2006a. Nitrosamine formation pathway revisited: the importance
of chloramines speciation and dissolved oxygen. Environ. Sci. Technol. 40, 6007-6014
Schreiber, I.M., Mitch, W.A., 2006b. Occurrence and fate of nitrosamines and nitrosamine
precursor in wastewater-impacted surface waters using boron as a conservative tracer. Environ.
Sci. Technol. 40, 3203-3210
Schreiber, I.M., Mitch, W.A., 2007. Enhanced nitrogenous disinfection byproduct formation near
the breakpoint: implications for nitrification control. Environ. Sci. 41, 7039-7046
Sedlak, D.L., Deeb, R.A., Hawley, E.L., Mitch, W.A., Durbin, T.D., Mowbray, S., Carr, S.,
2005. Sources and fate of nitrosodimethylamine and its precursors in municipal wastewater
treatment plants. Water Environ. Res. 77, 32-39
42
Servos, M.R., Bennie, D.T., Bumison, B.K., Jurkovic, A., McInnis, R., Neheli, T., Schnell, A.,
Seto, P., Smyth, S.A., Ternes, T.A., 2005. Distribution of estrogens, 17-estradiol and estrone,
in Canadian municipal wastewater treatment plants. Sci. Total Environ. 336, 155-170
Servos, M.R., Smith, M., McInnis, R., Burnison, B.K., Lee, B.H., Seto, P., Backus, S., 2007. The
presence of selected pharmaceuticals and the antimicrobial triclosan in drinking water in Ontario,
Canada. Water Qual. Res. J. Can. 42 (2), 130-137
Shah, A.D., Mitch, W.A., 2012. Halonitroalkanes, halonitriles, haloamides, and N-nitrosamines:
a critical review of nitrogenous disinfection byproduct formation pathways. Environ. Sci.
Technol. 46, 119-131
Shah, A.D., Krasner, S.W., Lee, C.F.T., von Gunten, U., Mitch, W.A., 2012. Trade-offs in
disinfection byproduct formation associated with precursor preoxidation for control of N-
Nitrosodimethylamine formation. Environ. Sci. Technol. 46, 4809-4818
Sharpless, C.M., Linden, K.G., 2003. Experimental and model comparisons of low- and
medium- pressure Hg lamps for the direct and H2O2 assisted UV photodegradation of N-
nitrosodimethylamine in simulated drinking water. Environ. Sci. Technol. 37, 1933-1940
Shemer, H., Kunukcu, Y.K., Linden, K.G., 2006. Degradation of the pharmaceutical
Metronidazole via UV, Fenton and photo-Fenton processes. Chemosphere 63, 269-276
Siddiqui, M., Atazi, K.Z., 2004. N-nitrosodimethylamine: A disinfectant byproduct and its
occurrence in wastewater. Water Environ. Res. 76, 316-326
Stackelberg, P.E., Furlong, E.T., Meyer, M.T., Zaugg, S.D., Henderson, A.K., Reissman, D.B.,
2004. Persistence of pharmaceutical compounds and other organic wastewater contaminants in a
conventional drinking-water-treatment plant. Sci. Total Environ. 329, 99-113
43
Stackelberg, P.E., Gibs, J., Furlong, E.T., Meyer, M.T., Zaugg, S.D., Lippincott, R.L., 2007.
Efficiency of conventional drinking-water-treatment processes in removal of pharmaceuticals
and other organic compounds. Sci. Total Environ. 377, 255-272
Stefan, M.I., Bolton, J.R., 2002. UV direct photolysis of N- nitrosodimethylamine (NDMA):
Kinetic and Product study. Helv. Chim. Acta 85, 1416-1426
Swaim, P., Royce, A., Smith, T., Maloney, T., Ehlen, D., Carter, E., 2008. Effectiveness of UV
advanced oxidation for destruction of micro-pollutants. Ozone-Sci. Eng. 30, 34-42
Ternes,s T.A., Meisenheiner, M., McDowell, D., Sacher, F., Brauch, H.J., Haist-Gulde, B.,
Preuss, G., Wilme, U., Zulei-Seibert, N., 2002. Removal of Pharmaceuticals during Drinking
Water Treatment. Environ. Sci. Technol. 36, 3855-3863
Terzic, S., Senta, I., Ahel, M., Gros, M., Petrovic, M., Barcelo, D., Muller, J., Knepper, T.,
Marti, I., Ventura, F., Jovancic, P., Jabucar, D., 2008. Occurence and fate of emerging
wastewater contaminants in Western Balkan Region. Sci. Total Environ. 399, 66-77
Topp, E., Monteiro, S.C., Beck, A., Coelgo, B.B., Boxall, A.B.A., Duenk, P.W., Kleywegt, S.,
Lapen, D.R., Payne, M., Sabourin, L., Li, H., Metcalfe, C.D., 2008. Runoff of pharmaceuticals
and personal care products following application of biosolids to an agricultural field. Sci. Total
Environ. 396, 52-59
USEPA, 1999. Alternative Disinfectants and Oxidants Guidance Manual, Chapter 6:
Chloramines.
USEPA, 2009. Contaminant Candidate List 3 (CCL3). Retrieved from
http://water.epa.gov/scitech/drinkingwater/dws/ccl/ccl3.cfm (accessed June 26, 2012)
Vieno, N.M., Harkki, H., Tuhkanen, T., Kronberg, L., 2007. Occurrence of pharmaceuticals in
river water and their elimination in a pilot-scale drinking water treatment plant. Environ. Sci.
Technol. 41, 5077-5084
44
Vogna, D., Marotta, R., Napolitano, A., d’Ischia, M., 2002. Advanced oxidation chemistry of
paracetamol: UV/H2O2-induced hydroxylation/degradation pathways and 15N-aided inventory of
nitrogenous breakdown products. J. Org. Chem. 67, 6143-6151
Vogna, D., Marotta, R., Napolitano, A., Andreozzi, R., d’Ischia, M., 2004a. Advanced oxidation
of the pharmaceutical drug diclofenac with UV/H2O2 and ozone. Water Res. 38, 414-422
Vogna, D., Marotta, R., Andreozzi, R., Napolitano, A., d’Ischia, M., 2004b. Kinetic and
chemical assessment of the UV/H2O2 treatment of antiepileptic drug carbamazepine.
Chemosphere 54, 497-505
Von Gunten, U., 2003. Ozonation of drinking water: Part I. Oxidation kinetics and product
formation. Water Res. 37, 1443-1467
Von Gunten, U., Salhi, E., Schmidt, C.K., Arnold, W.A., 2010. Kinetics and mechanisms of N-
nitrosodimethylamine formation upon ozonation of N, N-dimethylsulfamide-containing waters:
bromide catalysis. Environ. Sci. Technol. 44, 5762-5768
Weissmahr, K.W., Sedlak, D.L., 2000. Effect of metal complexation on the degradation of
dithiocarbamate fungicides. Environ. Toxicol. Chem. 19, 820-826
Westerhoff, P., Yoon, Y., Snyder, S., Wert, E., 2005. Fate of endocrine-disruptor,
pharmaceutical, and personal care product chemicals during simulated drinking water treatment
processes. Environ. Sci. Technol. 39, 6649-6663
Wilczak, A., Assadi-Rad, A., Lai, H.H., Hoover, L.L., Smith, J.F., Berger, R., Rodigari, F.,
Beland, J.W., Lazzelle, L.J., Kincannon, E.G., Baker, H., Heaney, C.T., 2003. Formation of
NDMA in chloraminated water coagulated with DADMAC cationic polymer. J. Am. Water
Works Assoc. 95 (9), 94-106
World Health Organization (WHO), 2008. Guidelines for Drinking-water Quality. Retrieved
from http://www.who.int/water_sanitation_health/dwq/fulltext.pdf (accessed June 26, 2012)
45
Xia, K., Bhandari, A., Das, K., Pillar, G., 2005. Occurrence and Fate of Pharmaceuticals and
Personal Care Products (PPCPs) in Biosolids. J. Environ. Qual. 34, 91-104
Yagil, G., Anbar, M., 1962. Kinetics of hydrazine formation from chloramine and ammonia. J.
Am. Chem. Soc. 84, 1797
Yang, W.C., Gan, J., Liu, W.P., Green, R., 2005. Degradation of N-Nitrosodimethylamine
(NDMA) in landscape soils. J. Environ. Qual. 34, 336-341
Yoon, Y., Westerhoff, P., Snyder, S.A., Esparza, M., 2003. HPLC-fluorescence detection and
adsorption of bisphenol A, 17β-estradiol, and 17α-ethynyl estradiol on powdered activated
carbon. Water Res. 37, 3530-3537
Yoon, Y., Westerhoff, P., Snyder, S.A., 2004. Removal of 17β estradiol and fluoranthene by
nanofiltration and ultrafiltration. J. Environ. Eng. 130, 1460-1467
Yoon, Y., Westerhoff, P., Snyder, S.A., Wert, E.C., Yoon, J., 2007. Removal of endocrine
disrupting compounds and pharmaceuticals by nanofiltration and ultrafiltration membranes.
Desalination 202, 16-23
Zuccato, E., Castiglioni, S., Fanelli, R., 2005. Identification of the pharmaceuticals for human
use contaminating the Italian aquatic environment. J. Hazard. Mater. 122, 205-209
Zwiener, C., Frimmel, F.H., 2000. Oxidative treatment of pharmaceuticals in water. Water
Resour. 34, 1881-1885
46
Chapter 3 Materials and Methods
3.1. Materials
3.1.1. Selection of Target PPCPs
Chemical structures of twenty compounds selected for this research are summarized in Figure
3.1, including nineteen pharmaceuticals and one personal care product (i.e., N, N-
diethyltoluamide (DEET)).
Figure 3.1. Structures of selected PPCPs for this research
47
The twenty PPCPs were selected based on three criteria. First, they contain DMA or
diethylamine (DEA) functional groups which are typical precursors of NDMA and N-
nitrosodiethylamine (NDEA), respectively (Figure 3.2). Among the twenty PPCPs, there are
eighteen NDMA precursors and two NDEA precursors (i.e., DEET and lidocaine).
Figure 3.2. Structures of DMA, NDMA, DEA, and NDEA
Second, most of the selected PPCPs are on the list of Top 50 dispensed medications in Canada
(IMS Health, 2009, 2011) and/or Top 200 pharmaceutical drugs in the US by retail sales (Drugs
Information Online, 2010) or by prescription dispensed (RxList, 2012). The general use and
prevalence in the North American market for the selected PPCPs are summarized in Table 3.1.
Table 3.1. PPCPs investigated
Compound Use Market prevalence Lidocaine Anesthetic Ranked 36th /200 in US, 2010 (sales) Azithromycin Antibiotics Ranked 15th /200 in US, 2012 (prescription) Clarithromycin Antibiotics Erythromycin Antibiotics Roxithromycin Antibiotics Tetracycline Antibiotics Amitriptyline Antidepressant Ranked 19th /50 in Canada, 2008 (prescription);
Ranked 120th / 200 in US, 2012 (prescription) Escitalopram Antidepressant Ranked 37th /50 in Canada, 2010 (prescription);
Ranked 12th /200 in US, 2010 (sales) Venlafaxine Antidepressant Ranked 7th /50 in Canada, 2010 (prescription);
Ranked 19th /200 in US, 2010 (sales) Metformin Anti-diabetic Ranked 32nd /50 in Canada, 2010 (prescription)
Ranked 19th /200 in US, 2012 (prescription) Sumatriptan Anti-migraine Ranked 70th /200 in US, 2008 (sales) Diltiazem Calcium channel blocker Ranked 191st /200 in US, 2012 (prescription) Carbinoxamine H1-antihistamines (Anticholinergic) Chlorphenamine H1-antihistamines (Cold and allergy relief) Diphenhydramine H1-antihistamines (Sleep aids) Doxylamine H1-antihistamines (Cold and allergy relief) Nizatidine H2-antihistamines (Antacid) Ranitidine H2-antihistamines (Antacid) Ranked 188th /200 in US, 2012 (prescription) DEET Insect repellent Tramadol Pain reliever Ranked 33rd / 200 in US, 2012 (prescription)
48
Lastly, the detection of some selected PPCPs in drinking water or source waters has been
reported in the literature. Very high concentrations of tramadol (up to 7731 ng/L) were reported
in two rivers in the UK (Kasprzy-Horden et al., 2008). The five antibiotics were detected in
surface waters in the US and Europe with concentrations up to several hundred ng/L. Most other
compounds were detected at lower concentrations, including amitriptyline (21 ng/L), DEET (60-
130 ng/L), diltiazem (21-106 ng/L), metformin (110 ng/L), and ranitidine (1.3-73 ng/L)
(Calamari et al., 2003; Kasprzy-Horden et al., 2008; Kolpin et al., 2002, 2004; Zuccato et al.,
2005). DEET has also been detected in finished drinking water with a maximum concentration of
66 ng/L, according to the US Geological Survey (Stackerlberg et al., 2004). Although the four
H1-antihistamines and nizatidine are neither on the top prescription list nor reported being
detected in surface waters, they were selected because of their structural similarity with
ranitidine, the first pharmaceutical which drew attention as a potential NDMA precursor with a
high conversion rate upon chloramination (Schmidt et al., 2006).
The basic physico-chemical properties of the selected PPCPs were obtained from ChemAxon
(ChemSpider), including the molecular weight (MW), Log P, and pKa of the amine group (Table
3.2). All of the selected compounds are hydrophilic (Log P < 5).
Table 3.2. Physico-chemical properties of selected PPCPs
Compound MW Log P pKa (amine) Compound MW Log P pKa (amine) Lidocaine 234.3 2.84 7.75 Sumatriptan 295.4 0.62 9.63 Azithromycin 749.0 2.44 8.91 Diltiazem 414.5 2.73 8.18 Clarithromycin 748.0 3.24 8.38 Carbinoxamine 290.8 3.27 8.87 Erythromycin 734.0 2.60 8.38 Chlorphenamine 274.8 3.58 9.47 Roxithromycin 837.1 3.00 9.08 Diphenhydramine 255.4 3.65 8.87 Tetracycline 444.5 -3.55 8.24 Doxylamine 270.4 2.96 8.87 Amitriptyline 277.4 4.81 9.76 Nizatidine 331.5 0.76 6.83 Escitalopram 324.4 3.76 9.78 Ranitidine 314.4 0.98 8.20 Venlafaxine 277.4 2.74 8.91 DEET 191.3 2.50 NA* Metformin 129.2 -1.36 NA* Tramadol 263.4 2.45 9.23
* NA: The pKa value of the amine group is not available; the amine group is always 100 % non-protonated.
49
3.1.2. Preparation of Working Solutions
Stock solutions of PPCPs were prepared in methanol and stored at 4 °C until use. NDMA and
NDEA (reagent grade) were used as standards, and deuterated NDMA (d6-NDMA, 98 atom %D)
was used as the internal standard for both compounds. All the above chemicals were purchased
from Sigma-Aldrich Canada (Oakville, Ontario). In addition, d6-ranitidine and d6-sumatriptan
(99 atom %D, Figure 3.3) were purchased from Synfine Research (Richmond Hill, Ontario) and
used as internal standards for ranitidine and sumatriptan, respectively.
Figure 3.3. Structures of d6-ranitidine and d6-sumatriptan
The chlorine stock solution (10000-15000 mg/L as Cl2) was prepared by making a 1 in 10
dilution of 10-15 % sodium hypochlorite (NaClO; Sigma-Aldrich Canada, Oakville, Ontario)
with Milli-Q® (MQ) water (MilliPore, Etobicoke, Ontario) and stored at 4 °C; the chlorine stock
solution was discarded if the concentration dropped by 15 %. Monochloramine dosing solution
(preformed) was prepared fresh daily by mixing an ammonium chloride solution (1.35 g NH4Cl
dissolved into 1 L MQ water) and a chlorine stock solution at a Cl2:NH4-N mass ratio of 4.2:1
(unless specified otherwise), and equilibrating for at least one hour before use. The
monochloramine concentration in general accounted for more than 90 % of the total chlorine in
the preformed dosing solution. The actual concentrations of chlorine stock solution,
monochloramine dosing solution, and the final test solutions were determined using DPD
colorimetry (DR2010 HACH-Kit, Hach Canada, Mississauga, Ontario) which measures the
concentration of total chlorine, free chlorine, and monochloramine as Cl2 (mg/L). L-ascorbic
acid was used to quench chlorine or chloramine (approximately 200-300 mg per 1 L water
sample) and stop the reaction following each test period. Buffer stock solutions (0.2 M) were
prepared in MQ water, as summarized in Table 3.3. The NH4Cl, L-ascorbic acid, and the
50
chemicals used to prepare the buffer solutions were all purchased from Sigma-Aldrich Canada
(Oakville, Ontario).
Table 3.3. Buffer solution recipes (Source: Sigma-Aldrich buffer reference centre)
pH Concentration Recipe 6.0 0.2 M 21.05 g/L KH2PO4 and 3.75 g/L Na2HPO4 7.0 0.2 M 12.4 g/L KH2PO4 and 15.6 g/L Na2HPO4 8.0 0.2 M 1.27 g/L KH2PO4 and 28.86 g/L Na2HPO4 9.0 0.2 M 1.06 g/L Na2CO3 and 7.56 g/L NaHCO3
3.1.3. Water Matrices
Various water matrices were used including lab-grade water (MQ), Lake Ontario water (LW),
Toronto tap water (Tap), Otonabee River water (RW), and treated Otonabee River water prior to
chlorination (TRW). Lake Ontario is the major source of drinking water for people living in
Southern Ontario, including the city of Toronto. The raw, untreated Lake Ontario water was
collected from influent to the Ajax Water Treatment Plant (Ajax, Ontario) in June 2010 and
September 2011. The tap water was collected from the tap in the Drinking Water Research
Group lab at the University of Toronto (Toronto, Ontario); the samples were collected after
letting the water running for at least 10 min. Otonabee River is the source of drinking water for
the city of Peterborough. The raw, untreated Otonabee River water was collected from influent to
the Peterborough Water Treatment Plant (Peterborough, Ontario) in October 2010 and January
2012. For the latter sampling, treated water but prior to final chlorination was also collected at
the water treatment plant in parallel with the raw water. The overall process train included
prechlorination (only turned on for water temperatures greater than 12 °C for the control of zebra
mussels; at the time the treated water was collected, prechlorination was not performed),
coagulation/flocculation/sedimentation, filtration, and final chlorine disinfection.
The basic water quality parameters are summarized in Table 3.4. The methods for measuring
these basic parameters are presented in Section 3.2.3. These water quality parameters were
measured every week during the sample storage period (no more than two months), and the
results in Table 3.4 are the average and standard deviation of 5 to 10 measurements.
51
Table 3.4. Basic water quality parameters for the selected water matrices (n = 5-10)
Water source Sampling date
TOC (mg/L) pH Alkalinity
(mg/L as CaCO3) UV254 (cm-1)
SUVA (L/mg·m)
Milli-Q® NA* 0.0 7.5 ± 0.1 1.8 ± 0.3 0.000 0.0 Lake Ontario, raw water
June 2010 2.3 ± 0.2 8.0 ± 0.1 94.6 ± 1.5 0.024 ± 0.002 1.1 ± 0.1 Sept. 2011 2.3 ± 0.1 8.2 ± 0.1 95.4 ± 1.8 0.023 ± 0.002 1.0 ± 0.1
Toronto tap water NA* 2.1 ± 0.1 7.1 ± 0.1 88.5 ± 2.8 0.021 ± 0.001 1.0 ± 0.1 Otonabee River, raw water
Oct. 2010 6.2 ± 0.5 7.8 ± 0.1 86.5 ± 1.7 0.143 ± 0.003 2.3 ± 0.2 Jan. 2012 6.2 ± 0.1 7.8 ± 0.1 94.9 ± 2.9 0.162 ± 0.002 2.6 ± 0.1
Otonabee River, treated water Jan. 2012 3.4 ± 0.1 7.2 ± 0.1 77.9 ± 0.1 0.067 ± 0.001 1.9 ± 0.1
* Not applicable
Excluding the lab-grade MQ water, the other four matrices represent a range of water sources
and characteristics. For example, there were untreated water samples (LW and RW) versus
treated water samples (Tap and TRW); for the treated samples, there were chloraminated
samples (Tap) versus non-chlorinated samples (TRW), As a result, the total organic carbon
(TOC) and specific UV absorbance (SUVA) also had a distribution of low (LW and Tap),
medium (TRW), and high (RW) values. The alkalinity was not strikingly different across the
selected matrices. Although pH varied from 7 to 8, buffer solution was used to control the
reaction pH through all the experiments and thus the variation in pH in the water was of less
significance.
Although TOC and SUVA are the most commonly used parameters to characterize the natural
organic matter (NOM) in different water matrices, these bulk water quality measurements do not
provide any information on the composition of NOM. In order to better characterize the water
matrix, NOM components of all the selected matrices were analyzed using liquid
chromatography – organic carbon detection (LC-OCD; Figure 3.4 and Figure 3.5). Details about
the analytical method and the separated NOM fractions are presented in Section 3.2.3. MQ water
was used as the reagent blank for the instrument and the dissolved organic carbon (DOC) level in
MQ was usually below 0.1 mg/L with an approximately equal mix of hydrophilic and
hydrophobic components (Figure 3.4). For the other four matrices selected, the majority of the
NOM was hydrophilic (more than 80 %, Figure 3.4). The distribution of hydrophilic NOM sub-
fractions is compared in Figure 3.5. For the selected matrices, humic substances (HS) was
52
always dominant, followed by building blocks; the two fractions combined accounted for more
than 70 % of the total hydrophilic DOC. The low-molecular-weight (LMW) fractions (acids and
neutrals) accounted for 15-20 % of the total hydrophilic DOC. The waters from the lake source
(LW and TAP) contained a higher portion of biopolymers (13-17 %) than the waters from the
river source (RW and TRW, 5-6 %).
Figure 3.4. LC-OCD: Hydrophilic vs. Hydrophobic DOC distribution for selected water matrices
Figure 3.5. LC-OCD: Hydrophilic NOM fractions of selected water matrices
53
3.2. Analytical Methods
3.2.1 Nitrosamine Analysis (GC-MS)
The procedure for extraction and concentration of nitrosamines (NDMA and NDEA) in water
samples was adopted from that reported by Taguchi et al. (1994). In general, the nitrosamines
were extracted with Lewatit® AF 5 mesh size 20-50 activated carbon beads (Figure 3.6),
desorbed with dichloromethane (DCM) and analyzed by Gas Chromatography-Mass
Spectrometry (GC-MS). The detailed extraction/concentration procedure and the GC-MS method
are outlined in Table 3.5 and Table 3.6, respectively. Method detection limits (MDLs) for
NDMA and NDEA were determined in Lake Ontario water (a more conservative estimation than
in MQ water) based on the EPA standard method (USEPA, 1986) and are provided in Table 3.7.
Figure 3.6. The Lewatit® AF 5 activated carbon beads for nitrosamine extraction
Table 3.5. The nitrosamine extraction and concentration method outline
Extraction: An aliquot of 500 mL sample was transferred to a clean 1 L amber bottle with an LDPE (low-density polyethylene) cap. After the addition of internal standard d6-NDMA (50 ng/L in 500 mL sample) and 200 mg of Lewatit® AF 5 beads (conditioned at 320 °C for 3 hr before use), the bottle was swirled at 250 rpm for 1 hr on an orbit shaker (Thermolyne Bigger Bill M49235, Barnstead International, Asheville, N.C., USA). The beads were then collected by filtration, air dried for 20 min in the fume hood (or until the beads could move freely without sticking to each other), transferred to a 2.0 mL auto-sampler vial, and further air dried for at least 1 hr.
54
Concentration: 500 µL of dichloromethane (DCM, 99.9 %) was then added directly into the 2.0 mL auto-sampler vial to extract nitrosamines from the beads. The cap for the vial was Teflon-lined and contained no rubber. The concentrated samples in DCM were ready to be analyzed on GC-MS. Sample storage: If the extraction could not be performed immediately, the chloramine residual in the samples (if any) was quenched with L-ascorbic acid (approximately 200 - 300 mg per 1 L water sample). Then 500 mL of the sample was measured and the internal standard was added. Samples at this condition were stored at 4 °C for no more than 48 hr prior to the extraction. Extracted nitrosamines in DCM could be stored in amber vials at -15 °C or less for up to 28 days after sample extraction (Munch and Bassett, 2004).
Table 3.6. The GC-MS analysis of nitrosamines method outline
The extracted nitrosamine samples were analyzed via a Varian 3800 GC coupled with a Varian 4000 ion trap MS and CombiPAL autosampler (Agilent Technologies, Mississauga, ON).
GC conditions:
The injector was fitted with a Carbofrit liner (Chromatographic Specialties; 3.4 mm ID and 5.0 mm OD; 54 mm length) and a programmed temperature vaporizer (PTV), and a DB 1701 column (30 m × 0.25 mm × 0.25 µm) was employed.
8 µL of sample was injected into the GC through the PTV inlet, with the initial temperature of 25 °C held for 0.8 min, increased by 200 °C/min to 240 °C, and held for 24 min. Coolant was enabled at 200 °C after the run to bring the injector back to the initial temperature. Column flow was 1.2 mL/min, with a pressure pulse of 20 psi held for 4 min.
Oven temperature was initially held at 35 °C for 5.5 min, increased by 15 °C/min to 155 °C, and further increased by 40 °C/min to 240 °C which was held for 10 min.
MS conditions:
Chemical ionization (CI) was applied with methanol as the reagent liquid. CI parameters were as follows: 3 µScan; emission current of 50 µAmps; electron multiplier offset of +300 V.
Filament delay was 8.2 min. NDMA and d6-NDMA were both eluted at a retention time of 8.6 min, with indicating ions monitored at 75 and 81 amu, respectively. NDEA was eluted at a retention time of 10.5 min with the major indicating ion monitored at 103 amu.
Table 3.7. GC-MS method: Method detection limits for nitrosamines (in Lake Ontario water)
Compound 1 2 3 4 5 6 7 AVE (ng/L) STD (ng/L) Recovery MDL (ng/L) NDMA 2.1 2.0 2.5 2.3 2.1 1.9 2.1 2.1 0.20 107 % 0.6 NDEA 1.7 1.7 1.8 1.5 1.9 1.8 1.9 1.8 0.14 88 % 0.4
55
3.2.2 Quantum Property Calculation for PPCPs
The molecular modeling software HyperChem® (Version 8.0, Hypercube Inc., Gainesville,
Florida) was used to generate the molecular electrostatic potential (MEP) map and to calculate
the atomic partial charge distribution for all the selected PPCPs discussed in Section 3.1.1.
HyperChem® required the input of a molecular structure in the format of an HIN file, which was
converted from an MOL file for each respective compound (obtained from ChemAxon
(Chemspider)) using OpenBabel® (Version 2.3.1, GNU General Public License). The MEP map
of each compound was generated via a single-point calculation after geometry optimization
(molecular mechanics, MM+ mode). The RM1 semi-empirical model was selected to perform
the single-point calculation, because it has been demonstrated to accurately calculate electrostatic
potentials (Puranen et al., 2010) and to improve the estimation of charge distribution especially
for nitro compounds (Rocha et al., 2006). The atomic partial charge was computed via the QSAR
(quantitative structure-activity relationship) property module.
3.2.3 Basic Water Quality Measurements and NOM Characterization
The pH was determined using a pH meter (Model 8015, VWR Scientific Inc., Mississauga,
Ontario). Alkalinity was measured based on an end-point titration, according to Standard Method
2320B (APHA, 2005). The total organic carbon content (TOC) was analyzed with an Aurora
1030 TOC analyzer (O.I. Analytical, College Station, Texas). The ultraviolet absorbance at 254
nm (UV254) was determined using a CE3055 Reflectance Spectrophotometer (Cecil Instruments
Ltd., Cambridge, England); it indicates the presence of UV absorbing structures such as
aromatics or conjugated double bonds. The specific UV absorbance (SUVA) was calculated by
normalizing the UV254 to the TOC; SUVA represents the aromaticity of a water sample that is
independent of its total organic matter level.
NOM components in each water were analyzed using the LC-OCD analyzer at the University of
Waterloo (Waterloo, Ontario). In addition to the raw water samples, several chlorine- and/or
chloramine- treated samples (lake and river; blank control samples) were collected along with
the disinfection experiments to investigate the potential change of NOM upon treatment. The
chlorine residual was quenched with sodium thiosulfate (Na2S2O3) at a dosage of twice the
56
stoichiometric ratio to the concentration of chlorine (personal communication, University of
Waterloo); the samples were then filtered through 0.45 µm membrane filter paper (Supor®-450
PES membrane filter, non-sterile, Pall Corporation) and shipped to University of Waterloo.
Samples were stored at 4°C for no more than one week prior to analysis.
The LC-OCD uses a size-exclusion column to separate the hydrophilic NOM fractions. The
hydrophobic matter does not elute within the limited measuring time of 120 min, and thus was
calculated as “DOC minus hydrophilic DOC”. A typical LC-OCD chromatograph for surface
water is shown in Figure 3.7. A portion of the sample is bypassed the column for the DOC
analysis and is shown as the “bypass” peak. The major hydrophilic NOM fractions are eluted in
the order of biopolymers, humics, building blocks, LMW acids, and LMW neutrals. Further
details about the LC-OCD system are described by Huber et al. (2011). Proprietary software was
used for data acquisition and processing (ChromCalc, DOC-LABOR, Karlsruhe, Germany).
Because the full separation of individual peaks cannot be achieved, the boundary of each peak
needs to be assigned arbitrarily prior to data processing and it is acknowledged that this arbitrary
assignment of peaks may introduce some errors in the quantification of individual fractions. A
detailed summary of the five major NOM fractions is provided in Table 3.8.
Figure 3.7. A typical LC-OCD chromatograph for surface water (Lake Ontario)
57
Table 3.8. Hydrophilic NOM fractions from LC-OCD analysis (summarized from Huber et al., 2011)
Fractions Elution order
Molecular weight
Quantification accuracy Other notes Possible
composition
Biopolymers 1st High Good Degrading overtime (likely converted into L MW fractions)
Polysaccharides; Proteins
Humic substances (HS)
2nd High Good Long-terms exposure to oxygen (aging) or sunlight can convert HS into building blocks
Humic acids; Fulvic acids
Building blocks 3rd Medium Moderate HS-like material
LMW acids 4th Low Less good The elution method restricted to saturated monoprotic acids
Monoprotic, saturated acids; LMW-HS
LMW neutrals 5th Low Poor Small and non-ionic
Alcohols; aldehydes; ketones
The MDLs for this method have not been provided by the analytical lab at University of
Waterloo. A rough estimation of the analytical variance was attained using limited data by
calculating the means and standard deviations of the raw water samples from the same source
over the storage period of 15 to 30 days (Table 3.9). If the change in NOM fractions upon the
disinfection treatment was smaller than the analytical variance, it was then considered to be not a
significant difference. Some fractions tended to degrade over time (i.e., biopolymers and humics)
into smaller molecules (LMW acids and neutrals), and thus it was difficult to differentiate if the
variance was due to the auto-degradation or the analytical variance. However, given that
biopolymers have the highest molecular weight among the five fractions, it was unlikely that
other fractions could degrade into biopolymers, and thus any increase in biopolymers observed
was more likely attributed to analytical variance.
Table 3.9. LC-OCD: Analytical variance for selected water matrices (mg/L; pH adjusted to 7.0 ± 0.1)
Fractions Lake Ontario (n = 2) Otonabee River (n = 3) Biopolymers 0.34 ± 0.03 0.26 ± 0.12
Humics 0.90 ± 0.03 3.39 ± 0.15
Building blocks 0.46 ± 0.02 0.93 ± 0.04
LMW acids 0.10 ± 0.01 0.17 ± 0.03
LMW neutrals 0.19 ± 0.01 0.50 ± 0.09
Hydrophilic DOC 1.99 ± 0.07 5.26 ± 0.36
DOC 2.22 ± 0.01 5.78 ± 0.41
58
3.3 Nitrosamine Formation Protocol
Nitrosamine formation experiments were conducted in 1 L amber bottles with LDPE (low-
density polyethylene) caps under modified formation potential (MFP) conditions or simulated
distribution system (SDS) conditions. The modified nitrosamine formation protocol was adapted
from standard operating procedures for disinfection byproduct (DBP) yields under uniform
formation conditions (Summers et al., 1996), and a general outline of the procedure is
summarized in Table 3.10. Formation potential (FP) tests usually apply high doses of
disinfectants to predict the ultimate formation potential; while the SDS tests (Koch et al., 1991)
simulate the conditions common to water treatment plants and distribution systems. The two sets
of general experimental conditions applied in MQ water are summarized in Table 3.11, only
differing in the concentrations of chloramine (preformed) applied: 28.4 mg/L for the MFP tests
was adopted from Schmidt et al. (2006), while 2.5 mg/L for the SDS tests met the requirement
for chloraminated distribution systems allowed by the Ontario Drinking Water Quality Standards
(MOE, 2006).
Table 3.10. General nitrosamine formation method outline
Glassware preparation:
Bottles were washed in the dishwasher (equipped with acid wash), rinsed three times with distilled water and baked at 350 °C in the oven for at least 4 hr prior to use. Bottles were made chlorine-demand-free by adding approximately 1 mL of 6% bleach per 1L of container volume, filling the bottles headspace free with deionized (DI) water and leaving in the dark overnight. The bottles were then rinsed thoroughly with distilled water and allowed to air dry.
Water sample preparation:
Bottles were filled three quarters full with a selected water matrix and dosed with PPCP stock solution(s) (volume calculated based on the target concentration). Unless otherwise specified, reactions were conducted at pH 7.0 ± 0.1, controlled by adding 2 mL/L of the pH 7.0 phosphate buffer. A calculated volume of preformed chloramine was dosed to achieve the target concentration. Bottles were then filled headspace-free with the selected water, capped, and incubated at room temperature for the designated contact time. Along with each set of experiments, blank samples were prepared according to the same procedure, using 1 L aliquot of the same water matrix without dosing any PPCPs.
Stop the reaction:
Reactions were halted by the addition of excess L-ascorbic acid powder (approximately 200 - 300 mg per 1 L water sample).
59
Table 3.11. General nitrosamine-FP experimental conditions (MQ)
Experimental conditions MFP SDS pH 7.0 ± 0.1 7.0 ± 0.1 Temperature 21 °C 21 °C Incubation time 24 hr 24 hr Cl2:NH3-N mass ratio 4.2 :1 4.2 :1 Chloramine (preformed) dosage * 28.4 ± 0.2 mg/L 2.5 ± 0.2 mg/L
* This is the dosage in the form of total chlorine; the volume of the preformed chloramine dosing solution required to achieve this concentration is also calculated based on the total chlorine measurements.
When performing the nitrosamine formation experiments in real water matrices, SDS conditions
were applied with a chloramine dosage at 2.5 ± 0.2 mg/L after satisfying the 24 hr chloramine
demand for each matrix. Preliminary chloramine demand tests were performed on each matrix
sample (pH adjusted to 7.0 ± 0.1) to determine the initial chloramine dosage. The chloramine
demand tests were conducted in 1 L chlorine-demand-free amber bottles at room temperature (21
°C), by measuring the chloramine decay after 24 hr of incubation. For the waters sampled in
2010, the chloramine decay was monitored for two initial concentrations (10 and 20 mg/L), and
the average decay was determined as the chloramine-demand. For the waters sampled in 2011 to
2012, the chloramine-demand was determined as the average chloramine decay for three initial
concentrations (5, 7.5, and 10 mg/L); this adjustment was made because the applied chloramine
dosage was below 10 mg/L and thus the modified protocol was thought to better reflect the
chloramine demand. The actual chloramine dosage applied in each matrix is summarized in
Table 3.12. The different chloramine dosages applied for the two samplings of Otonabee River
raw water were very likely due to the slight difference in the chloramine-demand tests. In
addition, there was some chloramine residual (< 1.0 mg/L) in tap water samples, for which the
chloramine was topped up to 2.5 mg/L.
Table 3.12. The chloramine dosage applied in each water matrix under the SDS conditions
Water matrix Sampling date Chloramine dosage (mg/L as total chlorine) Milli-Q® NA* 2.5 ± 0.2 Lake Ontario, raw water
June 2010 2.5 ± 0.2 Sept. 2011 2.5 ± 0.2
Toronto tap water NA* 2.5 ± 0.2 Otonabee River, raw water
Oct. 2010 4.4 ± 0.3 Jan. 2012 3.5 ± 0.2
Otonabee River, treated water Jan. 2012 3.8 ± 0.2 * Not applicable
60
The MFP conditions were only applied in some of the experiments that were conducted for
Chapter 4. SDS conditions were applied in most other experiments from Chapter 4 to Chapter 7,
with selected single parameters (i.e., Cl2:NH4-N mass ratio, reaction time, pH, non-preformed
chloramine) being modified for the purpose of the respective studies. Specific reaction
conditions used in experiments for Chapter 4 to Chapter 7 are presented in each chapter.
3.4 QA/QC
Nitrosamine concentrations were determined through isotope dilution using a 6-level calibration
curve (Appendix 1, Table A1.1 and Figure A1.1). Calibration standards were prepared along with
every batch of experiments using the respective water matrix. Further QA/QC data are provided
in Appendix 1. The nitrosamines formed from the water matrix itself were determined using
blank control samples (i.e., the respective water matrix without dosing any PPCPs) subjected to
the same treatment and extraction process, and were subtracted from the total amount of
nitrosamines formed in the samples (i.e., the respective water matrix dosed with selected
PPCPs). Along with every batch of tests, the blank samples were prepared in triplicate for the
NDMA-FP experiments (i.e., NDMA-FP at 24 hr); one blank sample was prepared at each time
point for the NDMA formation kinetic experiments. A detailed summary of the nitrosamine-FP
results from all of the selected water matrices during the course of this work is provided in
Appendix 2.
61
3.5 References
American Public Health Association (APHA), American Water Works Association (AWWA),
Water Environment Federation (WEF), 2005. Standard Methods for the Examination of Water
and Wastewater, 21st Ed. Editors Eaton, A. D., Clesceri, L. S., Rice, E. W., Greenberg, A. E.,
Washington, DC
Calamari, D., Zuccato, E., Castiglioni, S., Bagnati, R., Fanelli, R., 2003. Strategic survey of
therapeutic drugs in the Rivers Po and Lambro in Northern Italy. Environ. Sci. Technol. 37,
1241-1248
Chemspider, the free chemical database. Retrieved from http://www.chemspider.com/
(compounds calculated on July 10th, 2012)
Drugs Information Online. Top 200 pharmaceutical drugs by retail sales in 2010. Retrieved from
http://www.drugs.com/top200.html (accessed July 6, 2012)
Huber, S.A., Balz, A., Abert, M., Pronk, W., 2011. Characterisation of aquatic humic and non-
humic matter with size-exclusion chromatography – organic carbon detection – organic nitrogen
detection (LC-OCD-OND). Water Res. 45, 879-885
IMS Health, 2009. Pharmaceutical trends: Top 50 dispensed medications in Canada, 2008.
Retrieved from
http://www.imshealth.com/deployedfiles/imshealth/Global/Americas/North%20America/Canada
/StaticFile/Trends03_En09.pdf (accessed July 6, 2012)
IMS Health, 2011. Pharmaceutical trends: Top 50 dispensed medications in Canada, 2010.
Retrieved from
http://www.imshealth.com/deployedfiles/ims/Global/North%20America/Canada/Home%20Page
%20Content/Pharma%20Trends/Top50Dispensed_En_11.pdf (accessed July 6, 2012)
62
Kasprzyk-Hordern, B., Dinsdale, R.M., Guwy, A.J., 2008. The occurrence of pharmaceuticals,
personal care products, endocrine disruptors and illicit drugs in surface water in South Wales,
UK. Water Res. 42, 3498-3518
Koch, B., Krasner, S.W., Sclimenti, M.J., Schimpff, W.K., 1991. Predicting the formation of
DBPs by the simulated distribution system. J. Am. Water Works Assoc. 83, 62-70
Kolpin, D.W., Furlong, E.T., Meyer, M.T., Thurman, E.M., Zaugg, S.D., Barber, L.B., Buxton
H.T., 2002. Pharmaceuticals, Hormones, and Other Organic Wasterwater Contaminants in U.S.
Streams, 1999 - 2000: A National Reconnaissance. Environ. Sci. Technol. 36, 1202-1211
Kolpin, D.W., Skopec, M., Meyer, M.T., Furlong, E.T., Zaugg, S.D., 2004. Urban contribution
of pharmaceuticals and other organic wastewater contaminants to streams during differing flow
conditions. Sci. Total Environ. 328, 119-130
MOE, 2006. Procedure for Disinfection of Drinking Water in Ontario. Retrieved from
http://www.ene.gov.on.ca/environment/en/resources/STD01_076415.html (accessed July 9,
2012)
Munch, J.W., Bassett, M.V., 2004. USEPA 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, Cincinnati, O.H., September 2004. EPA 600-R-05-054
Puranen, J.S., Vainio, M.J., Johnson, M.S., 2010. Accurate conformation-dependent molecular
electrostatic potentials for high-throughput in silico drug discovery. J. Comput. Chem. 31, 1722-
1732
Rocha, G.B., Freire, R.O., Simas, A.M., Stewart, J.J.P., 2006. RM1: A reparameterization of
AM1 for H, C, N, O, P, S, F, Cl, Br, and I. J. Comput. Chem. 27, 1101-1111
63
RxList, 2012. Top 200 drugs by prescription dispensed in US. Retrieved from
http://www.rxlist.com/script/main/hp.asp (accessed July 6, 2012)
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
Sigma-Aldrich buffer reference centre. Retrieved from http://www.sigmaaldrich.com/life-
science/core-bioreagents/biological-buffers/learning-center/buffer-reference-center.html
(accessed on July 10th, 2012)
Stackelberg, P.E., Furlong, E.T., Meyer, M.T., Zaugg, S.D., Henderson, A.K., Reissman, D.B.,
2004. Persistence of pharmaceutical compounds and other organic wastewater contaminants in a
conventional drinking-water-treatment plant. Sci. Total Environ. 329, 99-113
Summers, R.S., Hooper, S.M., Shukairy H.M., Solarik, G., Owen, D., 1996. Assessing DBP
yield: uniform formation conditions. J. Am. Water Works Assoc. 88, 80-93
Taguchi, V., Jenkins, S.D.W., Wang, D.T., Palmentier, J.P.F.P., Reiner, E.J., 1994.
Determination of N-nitrosodimethylamine by isotope dilution, high-resolution mass
spectrometry. Can. J. Appl. Spectroscopy 39, 87-93
USEPA, 1986. Definition and procedure for the determination of the method detection limit. 40
CFR Part 136, Appendix B, revision 1.11, updated on June 30th, 1986.
Zuccato, E., Castiglioni, S., Fanelli, R., 2005. Identification of the pharmaceuticals for human
use contaminating the Italian aquatic environment. J. Hazard. Mater. 122, 205-209
64
Chapter 4 Demonstration of 20 Pharmaceuticals and Personal Care Products
(PPCPs) as Nitrosamine Precursors During Chloramine Disinfection
Most of this chapter has been previously published as:
Shen, R., Andrews, S.A., 2011a. Demonstration of 20 pharmaceuticals and personal care
products (PPCPs) as nitrosamine precursors during chloramine disinfection. Water Research,
45, 944-652.
The already-published results include Section 4.3.1 that investigates the nitrosamine formation
from all selected PPCPs when using modified formation potential (MFP) conditions; the part of
Section 4.3.3 that investigates the NDMA formation from selected pharmaceuticals under the
simulated distribution system (SDS) conditions; factors that may affect the reaction, including
the initial pharmaceutical concentration (part of Section 4.3.3.2), the effect of having mixtures of
PPCPs (part of Section 4.3.3.3), and the effect of the Cl2:NH4-N mass ratio (Section 4.3.3.4).
Results discussed in the published paper are only based on experiments performed in lab-grade
Milli-Q® (MQ) water and Toronto tap water. Data from similar experiments that were performed
in Lake Ontario water and Otonabee River water have been added to their respective sections in
this chapter for completeness. The potential for matrix effects is discussed in Section 4.3.3.1.
Moreover, the published paper includes a general discussion that begins to explain the different
nitrosamine molar conversions based on the structural differences of selected PPCPs. In this
chapter, a preliminary quantitative analysis is included to further correlate the nitrosamine molar
conversion with certain molecular properties of the selected PPCPs (Section 4.3.2).
65
Abstract
The worldwide detection of pharmaceuticals and personal care products (PPCPs) in the aquatic
environment and drinking water has been a cause for concern in recent years. The possibility for
concurrent formation of nitrosamine DBPs (disinfection by-products) during chloramine
disinfection has become another significant concern for delivered drinking water quality because
of their potent carcinogenicity. This chapter demonstrates that a group of PPCPs containing
amine groups can serve as nitrosamine precursors during chloramine disinfection, and the level
of nitrosamine conversion is significantly affected by the molecular structures and properties
surrounding the amine groups. Molar yields higher than 1 % are observed for eight
pharmaceuticals, with ranitidine showing the strongest potential to form N-nitrosodimethylamine
(NDMA). The molar conversion increases with the Cl2:NH4-N mass ratio, suggesting an
enhancement effect from dichloramine. The nitrosamine formation is inhibited by the presence
of natural organic matter (NOM) in water matrices, but further kinetic experiments are needed to
investigate the role of NOM in the nitrosamine formation from PPCP-based precursors.
Although the trace level of PPCPs in the environment suggests that they may not account for the
majority of nitrosamine precursors during the disinfection process, this study demonstrates a
connection between the transformation of PPCPs and the formation of nitrosamines during
chloramine disinfection. This both expands the pool of potential nitrosamine precursors, and
provides a possible link between the presence of trace levels of certain PPCPs in drinking water
sources and potential adverse health effects.
Keywords
PPCPs Nitrosamine NDMA Chloramine Disinfection Precursor
66
4.1 Introduction
The worldwide detection of pharmaceuticals and personal care products (PPCPs) in the aquatic
environment and drinking water has been a cause for concern in recent years. PPCPs are usually
poorly removed at water treatment plants because conventional treatment technologies are not
specifically designed to remove them at trace level, and thus might pass through the treatment
train and come into contact with the disinfectant. Currently, most research projects have been
focused on the removal of PPCPs using advanced technologies, but data in terms of their
transformation during these treatment processes are largely lacking. In particular, there is quite
limited information on the transformation of PPCPs upon drinking water disinfection.
Concurrently, formation of nitrosamine disinfection by-products (DBPs) during chloramine
disinfection has become another significant concern for delivered drinking water quality because
of their potent carcinogenicity (EPA IRIS, 1993). A considerable amount of research has been
conducted to investigate the potential precursors of nitrosamines, especially N-
nitrosodimethylamine (NDMA). The most well-known NDMA precursors related to water and
wastewater treatment include dimethylamine (DMA; Bond and Templeton, 2011; Mitch et al.,
2003), tertiary and quaternary amines containing DMA groups (Kemper et al., 2010; Lee et al.,
2007), natural organic matter (NOM) (Chen and Valentine, 2007; Dotson et al., 2007; Gerecke
and Sedlak, 2003; Krasner et al., 2008; Mitch and Sedlak, 2004), polyelectrolytes and resins
used in water and wastewater treatment plants (Kohut and Andrews, 2003; Mitch and Sedlak,
2004; Najm and Trussell, 2001; Wilczak et al., 2003), and some agriculturally relevant
fungicides and herbicides (Chen and Young, 2008; Graham et al., 1995; Schmidt and Brauch,
2008). However, current research regarding the potential precursors cannot always account for
all the nitrosamines detected based on their yields during drinking water disinfection, indicating
the possibility of other as yet unknown precursors. Ranitidine, one of the most prescribed drugs
for the treatment of stomach and intestine ulcers, has been demonstrated to render a high
conversion rate to NDMA upon chloramination (Schmidt et al., 2006). Some early studies have
also reported the formation of nitrosamines via amine drugs in the stomach (Lijinsky and Taylor,
1977; Andrews et al., 1980). Therefore, it is possible that drugs with amine groups might
contribute to the formation of nitrosamines during drinking water disinfection.
67
This chapter demonstrates the transformation of twenty selected PPCPs to form nitrosamines.
PPCPs containing DMA or diethylamine (DEA) in their structures were selected as potential
precursors for NDMA and N-Nitrosodiethylamine (NDEA), respectively. Selection of target
compounds was also based on their prevalence in the North American pharmaceutical markets
and/or their frequent detection in the environment (see details in Section 3.1.1).
4.2 Materials and Methods
Experiments were conducted in four different water matrices, including lab-grade Milli-Q® (MQ)
water (Millipore, Etobicoke, Ontario), Toronto tap water (Tap), Lake Ontario water (Lake), and
Otonabee River water (River). The raw, untreated Lake Ontario water and Otonabee River water
were collected from the influent of two water treatment plants in June and October 2010,
respectively. The water sources and basic water quality parameters are summarized in Table 4.1.
Table 4.1. Basic water quality parameters (n = 5-10) and the chloramine dosage applied in each matrix
under the simulated distribution system (SDS) conditions
Water source
TOC (mg/L) pH Alkalinity
(mg/L as CaCO3) UV254 (cm-1)
SUVA (L/mg·m)
Chloramine dosage (mg/L)
Milli-Q® 0 7.5 ± 0.1 1.8 ± 0.3 0.000 0.0 2.5 ± 0.2 Toronto tap water 2.1 ± 0.1 7.1 ± 0.1 88.5 ± 2.8 0.021 ± 0.001 1.0 ± 0.1 2.5 ± 0.2
Lake Ontario 2.3 ± 0.2 8.0 ± 0.1 94.6 ± 1.5 0.024 ± 0.002 1.1 ± 0.1 2.5 ± 0.2
Otonabee River 6.2 ± 0.5 7.8 ± 0.1 86.5 ± 1.7 0.143 ± 0.003 2.3 ± 0.2 4.4 ± 0.3
Nineteen pharmaceutical compounds and one personal care product (i.e., N, N-diethyltoluamide
(DEET)) were tested in terms of their potential to form nitrosamines during chloramination. The
structures and physico-chemical properties of the selected PPCPs are presented in Section 3.1.1.
Moreover, HyperChem® (Version 8.0, Hypercube Inc., Gainesville, Florida) was used to
generate the molecular electrostatic potential (MEP) map and to calculate the atomic partial
charge distribution for selected PPCPs (see Section 3.2.2).
68
Nitrosamine formation potential (FP) experiments were conducted in 1 L amber bottles with
LDPE (low-density polyethylene) caps. The nitrosamine formation protocol was described in
detail in Section 3.3. The FP tests were performed in triplicate to determine the extent of
nitrosamine conversion from individual compound or a mixture of PPCPs, and controls were
prepared in triplicate using the respective water matrix without spiking any PPCPs. Two sets of
experimental conditions were applied in the FP experiments, only differing in the concentrations
of chloramine applied: 28.4 mg/L for the modified FP (MFP) tests was adopted from Schmidt et
al. (2006), while 2.5 mg/L for the simulated distribution system (SDS) tests met the requirement
for chloraminated distribution system allowed by the Ontario Drinking Water Quality Standards
(MOE, 2006). When performing the nitrosamine-FP experiments in real water matrices, the SDS
conditions were applied with a chloramine dosage at 2.5 ± 0.2 mg/L after satisfying the 24 hr
chloramine demand for each matrix. For tap water that had chloramine residual (< 1.0 mg/L), the
chloramine was topped up to the designated concentration. The actually chloramine dosage
applied in each matrix under the SDS conditions is summarized in the last column of Table 4.1.
The pH was controlled at 7.0 ± 0.1 by adding 2 mL/L of the pH 7.0 phosphate buffer solution in
all the experiments. After incubating at room temperature (21 °C) for 24 hr, the reactions were
halted by the addition of excess ascorbic acid powder (approximately 200 - 300 mg per 1 L water
sample). Nitrosamines in water samples were then extracted with Lewatit® AF 5 mesh size 20-50
activated carbon, desorbed with dichloromethane (DCM) and analyzed by Gas Chromatography-
Mass Spectrometry (GC-MS). Quantification of nitrosamines was attained through isotope
dilution using d6-NDMA as the internal standard. Further details about the nitrosamine
extraction, concentration, analysis, and quantification methods are provided in Section 3.2.1. In
this chapter, error bars in all graphs demonstrate the variability due to multiple formation
potential tests (n = 3) under the same reaction conditions.
Preformed monochloramine was used in all of the experiments in this chapter. If not specified,
the monochloramine dosing solution was prepared fresh daily by mixing an ammonium chloride
solution and a chlorine stock solution at Cl2:NH4-N mass ratio of 4.2:1, and equilibrating for at
least one hour before use. Only in the study to evaluate the impact of the Cl2:NH4-N mass ratio,
the chloramine dosing solution was prepared at a series of different Cl2:NH4-N mass ratios
(3.0:1, 4.2:1, 5.1:1, and 6.3:1).
69
4.3 Results and Discussion
4.3.1 Nitrosamine-FP under MFP Conditions
Nitrosamine-FP upon chloramination was determined for all twenty selected PPCPs under the
MFP conditions, in both MQ and tap water. Results are summarized in Figure 4.1. The
nitrosamine molar conversion was calculated as ������������� �� ����������� ������ (��)������������� �� ������ ���� (��)
. Among the
tested PPCPs, eight pharmaceuticals showed molar conversions higher than 1 % (i.e., 18.5 ng/L
of NDMA or 25.5 ng/L NDEA formed from 25 nM of individual PPCPs). Ranitidine rendered
the highest conversion (89.9 % ~ 94.2 %), followed by doxylamine (8.0 % ~ 9.7 %), sumatriptan
(6.1 %), chlorphenamine (5.2 % ~ 5.5 %), nizatidine (4.5 % ~ 4.8 %), diltiazem (2.1 % ~ 2.6 %),
carbinoxamine (1.0 % ~ 1.4 %) and then tetracycline (0.8 % ~ 1.2 %). In both types of water,
the nitrosamine-FP varied generally within ± 25 % for most compounds, with somewhat higher
variability observed for limited tests in MQ water with the four macrolide antibiotics
(azithromycin, clarithromycin, erythromycin, and roxithromycin; ± 30 ~ 60 %). Additional MFP
tests that were performed with ranitidine indicated that the overall NDMA-FP from ranitidine
varied within ± 5 % (n = 9).
The results were generally in good agreement with a previous study by Schmidt et al. (2006), as
summarized in Table 4.2. Ranitidine gave a much higher yield of NDMA in the present study
than reported in the literature, even with a shorter reaction time. However, there was not enough
information available on the characteristics of the drinking water matrix used in the literature,
and thus it was difficult to further compare the results and explain the discrepancy. Further
studies would be needed to determine the potential impact from various water matrices relevant
to drinking water.
70
Figure 4.1. Nitrosamine-FPs for selected PPCPs under the MFP conditions (Initial concentration of individual PPCPs = 25 nM; 28.4 mg/L of NH2Cl; error bars represent the variability due to multiple
formation potential tests (n = 3))
Table 4.2. Comparisons with literature (Common conditions: pH = 7.0 ± 0.1, NH2Cl = 28.4 mg/L, PPCPs = 25 nM, room temperature)
Compound
Schmidt et al., 2006 (Drinking water, 7d)
Present data (Milli-Q® water, 24h)
Present data (Tap water, 24h)
NDMA (ng/L)
Molar conversion
NDMA (ng/L)
Molar conversion
NDMA (ng/L)
Molar conversion
Ranitidine 1200 62.9 % 1665 ± 6 89.9 ± 0.3 % 1744 ± 82 94.2 ± 4.4 %
Nizatidine 91 4.9 % 88.0 ± 1.3 4.8 ± 0.1 % 82.7 ± 6.9 4.5 ± 0.4 %
Tetracycline 23 1.2 % 23.0 ± 1.6 1.2 ± 0.1 % 14.9 ± 1.0 0.8 ± 0.1 %
71
4.3.2 Nitrosamine-FP vs. Molecular Properties
The selected PPCPs can be described as tertiary amines containing DMA or DEA functional
groups. Mitch and Schreiber (2008) have discussed the degradation pathways for model tertiary
amines to form nitrosamines during chloramination, involving a chlorine transfer from the
chloramine to the nitrogen atom in the tertiary amines as the rate-limiting step. When the
nitrosamine formation reaction can be an electrophilic attack on the N-atom of the DMA group,
the nitrosamine conversion of a precursor is then dependent on the electrostatic potential of the
amine group. Specifically, a portion of a molecule that has a negative electrostatic potential will
be susceptible to electrophilic attack. Moreover, the pKa of the amine group is also important
because it determines the amine species at a given pH, and non-protonated amines would favor
the electrophilic reaction.
Table 4.3 summarizes the nitrosamine molar conversion from all 20 selected PPCPs under the
MFP conditions in MQ water, along with the relevant molecular properties which would be
expected to affect the nitrosamine conversion. The eight pharmaceuticals with NDMA molar
conversions higher than 1 % are highlighted in bold. The pKa of the amine group was obtained
from ChemAxon (ChemSpider), and the percentage of non-protonated amine at pH 7.0 was
calculated based on the conventional pKa-pH relationship (fraction of non-protonated amine
= �����(������)). The quantum properties calculated using Hyperchem® included the minimum
electrostatic potential (Min. EP) of the molecule, the atomic partial charge of the N-atom on the
DMA (DEA) group, as well as the charge on atoms in the surrounding structures (up to the third
atom adjacent to the amine group). The atomic partial charge reflects the polarity and
electronegativity of atoms within the molecule. While Hyperchem® does not give the
electrostatic potential of each atom, it generates a molecular electrostatic potential (MEP) map
based on the atomic partial charges. For the majority of the selected PPCPs, the N-atom on the
amine group was usually close to the location of the minimum electrostatic potential, therefore
the Min. EP value was used to compare the electronegativity of the amine groups on a relative
scale.
72
Table 4.3. The nitrosamine molar conversion from selected PPCPs vs. relevant molecular properties
Category Compound Nitrosamine molar conversion
Minimal electrostatic potential (eV)
Charge of N-amine
Charge of adjacent atoms Structure adjacent
to amine pKa (amine)
% of non-protonated amine at pH 7.0 1st 2nd 3rd
H2-antihistamines
Ranitidine 89.9% -0.157 -0.27 0.10 0.13 -0.36 Furan 8.20 5.9% Nizatidine 4.8% -0.147 -0.27 0.10 0.14 -0.21 Thiazole 6.83 59.7%
H1-antihistamines
Carbinoxamine 1.0% -0.147 -0.27 0.11 0.16 -0.35 8.87 1.3% Chlorphenamine 5.5% -0.215 -0.27 0.07 0.02 0.02 9.47 0.3% Diphenhydramine 0.3% -0.066 -0.27 0.11 0.16 -0.36 8.87 1.3% Doxylamine 9.7% -0.141 -0.27 0.11 0.16 -0.36 8.87 1.3%
Macrolide antibiotics
Azithromycin 0.2% -0.248 -0.27 0.10 0.22 0.05 Macro-cyclic hydrocarbon
8.91 1.2% Clarithromycin 0.1% -0.227 -0.27 0.10 0.22 0.05 8.38 4.0% Erythromycin 0.3% -0.190 -0.27 0.10 0.22 0.05 8.38 4.0% Roxithromycin 0.1% NAa -0.27 0.10 0.22 0.05 9.08 0.8%
Others
Amitriptyline 0.3% -0.012 -0.27 0.07 0.02 0.00 9.76 0.2% Diltiazem 2.6% -0.156 -0.27 0.09 0.09 -0.25 Benzothiazepine 8.18 6.2% Escitalopram 0.4% -0.139 0.00 0.05 -0.06 -0.02 9.78 0.2% Metformin 0.1% -0.122 -0.24 0.21 -0.21 -0.15 C=N bond NA b 100.0% Sumatriptan 6.1% -0.216 -0.27 0.07 0.02 0.02 Indole 9.63 0.2% Tramadol 0.2% -0.130 -0.27 0.07 0.04 0.13 9.23 0.6% Tetracycline 1.2% -0.200 -0.27 0.09 0.17 0.05 8.24 5.4% Venlafaxine 0.2% -0.146 -0.27 0.07 0.04 0.00 8.91 1.2%
NDEA precursors
DEET 0.2% -0.166 -0.25 0.21 -0.21 0.05 Carbonyl NA b 100.0% Lidocaine 0.2% -0.284 -0.27 0.13 0.24 -0.20 7.75 15.1%
a Not available. Hyperchem® was not able to generate the molecular electrostatic potential map for roxithromycin under the selected model (RM1). The semi-empirical model has limitations when the molecule is too big or has complicated steric structure. bThe pKa value of the amine group is not available; the amine group is always 100 % non-protonated.
73
In general, the N-atom on the DMA (DEA) group has a negative atomic partial charge which is
associated with a negative electrostatic potential, suggesting that it will be susceptible to
electrophilic attack. Escitalopram is the only PPCP on which the N-atom on the DMA has a zero
partial charge, thus potentially explaining its very small NDMA molar conversion. Other than
that, for the twenty PPCPs there was no significant difference in the partial charge of the N-atom
as well as the 1st and 2nd atom adjacent to it. However, for the eight pharmaceuticals showing
high NDMA formation, either the 3rd atom adjacent to the N-atom had a high negative partial
charge, or the minimum electrostatic potential had a more negative value, as shown in Figure 4.2
as blue circles.
Figure 4.2. Relationships between the nitrosamine molar conversions from selected PPCPs (PPCPs = 25 nM, MFP, MQ) and the relevant molecular properties
Moreover, because it is an electrophilic substitution on the amine group, the nature of moieties
close to the DMA or DEA group can influence the reaction rate and thus affect the molar
conversions. Generally, an electron donating group close to the DMA or DEA group can increase
the electron density on the nitrogen atom and thus help attract chlorine; while an electron
withdrawing group can decrease the electron density and slow down the reaction. In this study,
the eight pharmaceuticals showing high NDMA-FPs all have the DMA group bound to an
electron-rich moiety. For example, ranitidine, nizatidine, diltiazem, and sumatriptan all have an
aromatic ring system adjacent to the DMA group that likely increases the electron density on the
74
N-atom. Amitriptyline and escitalopram have a long alkyl carbon chain between the DMA group
and the bulky aromatic system, and this weakens their electron donating effect. In contrast,
carbonyls or double bonds adjacent to the DMA group may inhibit NDMA conversion because
of their electron-withdrawing effect, such as occurs with the two NDEA precursors DEET and
lidocaine. A similar structure is observed for metformin, which has the DMA group bound to an
electron-withdrawing biguanide. The exceptions indicated in Figure 4.2 (the dotted circle) are the
three macrolide antibiotics (azithromycin, clarithromycin, and erythromycin). Although the
minimum electrostatic potential for them was more negative, the moieties with the minimum
electrostatic potential were not close to the DMA group, and their low NDMA conversion was
also likely due to the steric hindrance from the bulky cyclic hydrocarbons. Similarly, tramadol
and venlafaxine also have complicated steric structures close to the DMA group, which may
hinder its reaction towards chloramines.
It is worth noting that ranitidine and nizatidine have very similar structures, both having the
DMA group bound to the C2 site of a five-element heterocyclic ring, but ranitidine has a much
higher molar conversion than nizatidine. Figure 4.3 compares the MEP maps between the furan
and thiazole ring. The map is color coded; blue indicates positive electrostatic potential, and red
indicates negative electrostatic potential. The furan ring of ranitidine has a symmetric structure
and the C2 site is strongly electrophilic due to the electron-donating effect of the oxygen
heteroatom. However, the C2 site on the thiazole ring of nizatidine is a slightly nucleophilic site
because of the combined effects from the nitrogen and the sulfur atoms. This may explain the
much higher NDMA conversion from ranitidine than that from nizatidine.
Figure 4.3. The molecular electrostatic potential map for furan and thiazole ring
75
A similar structural analysis was performed for the four structurally similar H1-antihistamines
(carbinoxamine, chlorphenamine, diphenhydramine, and doxylamine) to explain their relative
potentials to form nitrosamines. They all have an electron-rich bulky aromatic system in their
structures, but the distance between the aromatic structure and the DMA group is farther than
that of ranitidine, resulting in the overall lower molar conversions. Among these four
pharmaceuticals, the NDMA molar conversion follows the order of doxylamine >
chlorphenamine > carbinoxamine > diphenhydramine. The molecular electrostatic potential
maps for the four compounds are compared in Figure 4.4. Diphenhydramine has the least
negative minimal electrostatic potential (Min.EP) value (-0.066 eV), indicating that the amine
group of it is the least reactive one for an electrophilic attack, thus it had the lowest NDMA
molar conversion. Compared with diphenhydramine, carbinoxamine has a pyridine structure and
one chlorine atom on the benzene ring; in particular the chlorine on the ortho position of the
aromatic ring could increase the electron density of the ring via resonance, and the net result is a
more electronegative DMA group for carbinoxamine, thus more prone to the electrophilic attack.
Chlorphenamine has a very similar structure to carbinoxamine but has a higher NDMA molar
conversion; the only structural difference is that chlorphenamine does not have the oxygen atom
on carbinoxamine which may reduce the electron density of the DMA group due to its inductive
electron-withdrawing effect. Doxylamine has an extra methyl group which is an inductive
electron donating group that increases the electron density of the DMA group. Moreover,
chlorphenamine has a more negative Min. EP value (-0.215 eV) than doxylamine (Min. EP = -
0.141 eV), but it also has less non-protonated amine at pH 7.0 (0.3 %) compared with
doxylamine (1.3 %), therefore the overall NDMA molar conversion from chlorphenamine under
the given reaction conditions was lower than that from doxylamine. Similar structural analysis of
other amine-based chemicals may provide insight into their relative potentials to form
nitrosamines via electrophilic reaction.
76
Figure 4.4. The molecular electrostatic potential map for the four H1-antihistamines
77
4.3.3 Nitrosamine-FP under SDS Conditions
Although it was clear from the MFP experiments reported in Section 4.3.1 that nitrosamines
could be formed from PPCPs when high concentrations of NH2Cl (28.4 mg/L) were present,
SDS conditions are usually applied to mimic practical disinfection conditions common to water
treatment plants and distribution systems. Therefore, Nitrosamine-FP upon chloramination under
SDS conditions (2.5 mg/L NH2Cl) was determined for the eight pharmaceuticals which rendered
an NDMA molar conversion higher than 1 % under MFP conditions. The results are summarized
and compared with the MFP results in Figure 4.5. No significant difference was observed
regarding the NDMA molar conversion between the two sets of conditions (p-value = 0.071
(MQ) and 0.357 (Tap), paired t-test, 95 % confidence level). Under either condition, the amount
of chloramine was in large excess relative to that of the pharmaceuticals (mg/L vs. µg/L),
suggesting that the reaction is not limited by the availability of chloramine and that essentially
complete reactions could be achieved under the SDS conditions. It also indicates that the eight
pharmaceuticals are capable of forming NDMA under the practical chloramine disinfection
conditions.
78
Figure 4.5. Comparison of NDMA-FPs for selected PPCPs between the MFP and the SDS conditions (Initial concentration of individual PPCPs = 25 nM; NH2Cl MFP = 28.4 mg/L; NH2Cl SDS = 2.5 mg/L; error
bars represent the variability due to multiple formation potential tests (n = 3))
4.3.3.1 Matrix Effect
The NDMA-FP for these eight pharmaceuticals was further determined in untreated Lake
Ontario water and Otonabee River water, as compared in Figure 4.6. In general, the NDMA
formed from selected pharmaceuticals after 24 hr followed the order of MQ water ≈ tap water >
lake water > river water. Differences between pairs of matrices were compared using the paired
t-test at 95 % confidence level (Appendix 4, Table A4.2). The matrix effect on the 24 hr NDMA-
79
FP was determined to be significant for all of the selected pharmaceuticals (p-value < 0.0001,
one-way ANOVA). Further details about the statistical analysis are presented in Appendix 4.
Figure 4.6. NDMA-FPs for selected PPCPs: the matrix effect (SDS conditions; initial concentration of
individual PPCPs = 25 nM; error bars represent the variability due to multiple tests (n = 3))
Matrix effects are usually associated with the matrix components, especially NOM, which may
affect NDMA formation in two ways: competition for chloramine and direct interaction with
PPCPs. The competition with pharmaceuticals for chloramine was considered to be minimal due
to the small observed chloramine decay (< 0.5 mg/L within 24 hr) and the large excess of
chloramine relative to the pharmaceuticals (mg/L vs. lower µg/L). On the other hand, NOM may
also interact with pharmaceuticals and thus reduce their conversion into NDMA or slow down
the reaction; thus further kinetic experiments were needed to investigate the role of NOM in the
NDMA formation from selected pharmaceuticals. The potential impact of water matrix on the
NDMA formation kinetics will be further discussed in Chapter 5.
80
4.3.3.2 Impact of Initial Pharmaceutical Concentration
Since reaction yields can sometimes be affected by the initial concentrations of the reactants,
SDS tests for the eight target pharmaceuticals were also performed using a series of initial PPCP
concentrations in various matrices (Figure 4.7). When evaluating the individual pharmaceuticals
in different matrices, the NDMA molar conversion varied as the initial concentration changed,
but no common pattern was observed. Even for the same pharmaceutical, the impact of initial
concentration varied among different water matrices (Appendix 5, Table A5.2). When evaluating
all the pharmaceuticals tested in one matrix as a whole (paired t-test, 95 % confidence level),
however, the impact of initial pharmaceutical concentration was determined to be not significant
(Appendix 5, Table A5.4). Under all of the concentrations tested (from 280 ng/L
(chlorphenamine and doxylamine) to 22 µg/L (tetracycline)), chloramine was always in large
excess. Thus, it was not surprising that the NDMA molar conversion from pharmaceuticals was
relatively independent of the initial pharmaceutical concentration.
The real environmental concentrations for PPCPs are typically lower than the concentrations
used in this work. For example, ranitidine has been detected in surface waters at various
locations with concentrations in the tens of ng/L range (Kasprzyk-Hordern et al., 2008; Kolpin et
al., 2002, 2004; Zuccato et al., 2005). Diltiazem has been detected in some US and UK streams
with concentrations up to several hundred ng/L (Kasprzyk-Hordern et al., 2008; Kolpin et al.,
2002, 2004). Tetracycline has been frequently detected in many US and Canadian sites with
concentrations up to 300 ng/L (Kolpin et al., 2002, 2004; Miao et al., 2004). Because of the
largely excess chloramine, the NDMA molar conversions determined in this work should still
apply to the pharmaceuticals at their environmental concentrations. In the present study, the
lowest concentration tested was 100 ng/L for ranitidine, nizatidine, chlorphenamine, and
doxylamine. Ranitidine showed the strongest potential to form NDMA, with molar conversions
higher than 77 % at all of the concentrations tested. Even at 100 ng/L (or 0.32 nM), ranitidine
can form 18.2 ± 1.2 ng/L (or 0.25 ± 0.02 nM) of NDMA, which is beyond the current Ontario
regulation of 9 ng/L (MOE, 2003) and the California & Massachusetts regulation of 10 ng/L
(MassDEP, 2004; OEHHA, 2006). NDMA formed via the other three compounds at 100 ng/L
were low (1-2 ng/L), but still within the quantifiable range.
81
Figure 4.7. NDMA-FPs and molar conversions for selected PPCPs at different initial concentrations (SDS conditions; tap water; error bars represent the variability due to multiple formation potential tests (n = 3))
82
In addition, the experiments at different initial pharmaceutical concentrations further confirmed
the matrix effect results. The paired t-test for data obtained at 1 nM and 5 nM of initial
pharmaceutical concentration suggest that the NDMA-FP after 24 hr followed the order of MQ
water ≈ tap water > lake water > river water (Appendix 4, Table A4.2), in good agreement with
the results obtained at 25 nM of pharmaceuticals.
Furthermore, it is worth noting that to better evaluate the NDMA-FP of pharmaceuticals, it will
be necessary to take into consideration all the PPCP-derived species containing the DMA groups
that may enter the drinking water treatment scheme. Pharmaceutical substances usually undergo
metabolism within the human body and thus are excreted as a mixture of parent compounds
together with the metabolites; also some pharmaceuticals are subjected to transformations in the
environment, resulting in the formation of different transformation products. However, as long as
the DMA functional groups are components of the metabolites and/or transformation products,
they may still contribute to the formation of NDMA when reacting with chloramines. Take
ranitidine as an example, earlier pharmacokinetics and pharmacodynamics studies have indicated
that 30 ~ 70 % of ranitidine is excreted as the parent form (Jjemba, 2006), and its major
metabolites in human body include N-oxide, S-oxide, and desmethylranitidine (Carey et al.,
1981). In the aquatic environment, ranitidine is transformed into two major products under solar
irradiation (Isidori et al., 2009). The S-oxide metabolite and both the solar irradiation products
maintain the DMA groups in their structures. Moreover, removal of parent ranitidine in
conventional wastewater treatment plants has been reported to vary between 0 to 89 % in
different seasons (Castiglioni et al., 2006), but no data is available in terms of the removal of its
major metabolites. Currently, most occurrence studies have been only focused on the parent
compounds but have not detected or reported the likely substantial amounts of their metabolites
and transformation products. As a result, even though the concentration of the parent compound
in the environment is low (ng/L level), altogether with its metabolites and transformation
products, the overall nitrosamine formation potential may still be high and should be taken into
consideration.
83
4.3.3.3 Mixture Effect
In real environments, PPCPs are usually present in the form of mixtures rather than as single
compounds. Even though the majority of the selected PPCPs form low levels of NDMA, added
together they may still pose a concern in terms of the overall formation of nitrosamines.
In order to examine the potential effect of mixtures on the formation of NDMA via PPCPs, the
eight pharmaceuticals were prepared in a mixture and subjected to chloramination under the SDS
conditions. Although a slight antagonistic effect was observed in the mixture of pharmaceuticals
in MQ water, the NDMA-FP was reduced by less than 10-15 % compared with the sum of
NDMA concentrations produced from single compounds at the same concentration. However,
the antagonistic effect was determined to be not significant for most of the tests performed in real
water matrices (Figure 4.8 (a); Appendix 6). As demonstrated in Section 4.3.3.1, the NDMA-FP
from individual pharmaceuticals in real water samples could be inhibited due to the presence of
NOM; therefore, the slight reduction in NDMA formation from the mixture of pharmaceuticals
which was observed in MQ water may not always be observed in real water matrices.
Figure 4.8. NDMA-FPs for sum of single PPCPs vs. PPCP-mixture (SDS conditions; eight-
pharmaceutical mixture: ranitidine, doxylamine, sumatriptan, chlorphenamine, nizatidine, diltiazem, carbinoxamine, and tetracycline; seven-pharmaceutical mixture: ranitidine excluded from the above
eight; error bars represent the variability due to multiple formation potential tests (n = 3); “*” indicates significant difference between two bars, t-test, 95 % confidence level)
84
Moreover, because ranitidine had an NDMA molar conversion one order of magnitude higher
than the rest of the other pharmaceuticals and thus contributes to the majority of the overall
NDMA-FP (more than 80 %), the matrix effect on ranitidine alone can significantly affect the
overall mixture effect. An example is presented in Table 4.4 using the data from 5 nM of PPCPs
in MQ and the river water. Theoretically, if there is no significant mixture effect, the reduction in
NDMA-FP due to the matrix effect should be the same for the sum of eight pharmaceuticals and
the eight-pharmaceutical mixture. However, the river water had a profound inhibition effect on
the NDMA formation from ranitidine alone (i.e., 4.26 nM in MQ water and 2.57 nM in river
water, 40 % reduction), and the reduction due to the matrix effect for the eight-pharmaceutical
mixture (41 %) was lower than the reduction for the sum of eight pharmaceuticals (47 %), i.e.,
the mixture effect in the river water was undermined by the matrix effect on ranitidine. In
contrast, when ranitidine was excluded, the reduction due to the matrix effect for the seven-
pharmaceutical mixture (84 %) was higher than the reduction for the sum of seven
pharmaceuticals (70 %), indicating that the matrix effect alone cannot explain all the NDMA-FP
reduction in the mixture, and the reduction was a combined impact from both the matrix and
mixture. As shown in Figure 4.8 (b), the NDMA-FP reduced in the seven-pharmaceutical
mixture was determined to be significant in both MQ and the river water when ranitidine was
excluded (Appendix 6). These results suggest that the mixture effect could be complicated by the
presence of NOM or certain precursors with exceptionally high NDMA formation potentials.
Table 4.4. Example of the mixture effect with and without ranitidine
(PPCP = 5 nM) NDMA-FP (nM) MQ River Reduction due to matrix effect
Ranitidine included
Sum of 8 PPCPs 5.50 2.94 47 %
8-PPCP mixture 4.91 2.91 41 %
Reduction due to mixture effect 11 % 1 %
Significant mixture effect? Yes No
Ranitidine excluded
Sum of 7 PPCPs 1.24 0.37 70 %
7-PPCP mixture 0.96 0.15 84 %
Reduction due to mixture effect 23 % 60 %
Significant mixture effect? Yes Yes
85
4.3.3.4 Impact of Cl2:NH4-N Mass Ratio
Chlorine to ammonia nitrogen (Cl2:NH4-N) mass ratio is an important factor for chloramine
disinfection. It can determine the dominant chloramine species along with pH (6.5 ~ 8.5)
typically encountered in drinking water disinfection (USEPA, 1999). Monochloramine is
predominately formed when the applied mass ratio is less than 5:1; dichloramine starts to form as
the ratio increases, yielding a mixture of monochloramine and dichloramine; breakpoint reaction
occurs when the mass ratio is above 7.6:1, resulting in the formation of free chlorine and
nitrogen trichloride.
Mitch et al. (2005) have reported that the occurrence of dichloramine can significantly enhance
the NDMA formation via tertiary amines, regardless of its relatively minor fraction. In the
present study, the potential impact from the Cl2:NH4-N mass ratio was studied by exposing a
mixture of eight pharmaceuticals to chloramines prepared at different Cl2:NH4-N mass ratios. It
was observed that the NDMA molar conversion increased as the ratio increased from 3:1 to
6.3:1, corresponding to an increasing fraction of dichloramine from approximately 10 % to 40 %
in the dosed samples (Figure 4.9).
Figure 4.9. Impact of (a) Cl2:NH4-N mass ratio or (b) the fraction of NHCl2 on the NDMA formation via selected pharmaceuticals (SDS conditions; tap water; eight-pharmaceutical mixture: ranitidine,
doxylamine, sumatriptan, chlorphenamine, nizatidine, diltiazem, carbinoxamine, and tetracycline; 5 nM or 25 nM of each pharmaceutical in the mixture; error bars represent the variability due to multiple
formation potential tests (n = 3))
86
For utilities applying chloramine disinfection, it is generally recommended to maintain the
Cl2:NH4-N mass ratio near to but below 5:1 to achieve the required residual and to avoid
breakpoint reactions (USEPA, 1999). However, it is difficult to maintain a stable operating ratio,
and a slight shift of the ratio above 5:1 might cause a spike of NDMA formation. Furthermore,
monochloramine undergoes disproportionation to form some dichloramine over a period of a day
or so (USEPA, 1999). Therefore, the chloramine residual towards the further end of the
distribution system very likely includes a portion of dichloramine. If there are any potential
nitrosamine precursors present in the finished drinking water, such as trace-level PPCPs, the
prolonged formation of nitrosamines in the distribution system may cause a concern especially
when water needs to be delivered over a long distance.
4.4 Summary
All of the 20 selected PPCPs were able to form corresponding nitrosamines upon chloramine
disinfection. Eight pharmaceuticals rendered molar conversions higher than 1 %, showing
the potential to form NDMA under practical disinfection conditions. Ranitidine showed a
particularly strong potential to form NDMA, even at environmentally relevant
concentrations. The molar yields of NDMA via ranitidine (0.1 ~ 15.7 µg/L) were higher than
77 %.
Some quantum molecular properties (i.e., the electrostatic potential map, the atomic partial
charges) of the selected PPCPs were compared with their relative potentials to form
nitrosamines. This preliminary structural analysis still well supported the observed
nitrosamine formation results.
The NDMA-FP at 24 hr in different water matrices followed the order of MQ ≈ Tap > Lake
> River, suggesting the possible inhibition effect due to the presence of NOM. Kinetic
studies are required to further investigate the role of NOM in the NDMA formation from
PPCPs.
87
At the concentration range tested in this work, the NDMA-FP was relatively independent of
the initial pharmaceutical concentration. Although the majority of these compounds gave
yields of less than 1 % molar conversion, when added together they may still contribute
significantly to the formation of nitrosamines during chloramine disinfection.
NDMA molar conversion increased with the Cl2:NH4-N mass ratio, indicating an
enhancement effect of dichloramine on the formation of NDMA via selected PPCPs. This
may cause potential concern in the distribution system.
Overall, results from the present study have suggested that PPCPs with substituted amine groups
can serve as potential nitrosamine precursors during chloramine disinfection. Due to their trace
level in source waters, it is not likely that PPCPs will account for the majority of nitrosamine
precursors in drinking water. However, this study proves the possible connection between the
transformation of PPCPs and the formation of nitrosamines during chloramination process.
Further research would be needed to determine the possible impact from different water
matrices. Kinetic studies are also required to investigate the possible reaction mechanisms
involved. Moreover, metabolites and transformation products of some PPCPs may also pose the
potential to form nitrosamines, thus the overall nitrosamine formation potential of PPCPs should
consider the parent compounds, their metabolites, as well as the possible transformation
products.
88
4.5 References
Andrews, A.W., Fornwald, J.A., Lijinsky, W., 1980. Nitrosation and mutagenicity of some
amine drugs. Toxicol. Appl. Pharmacol.52, 237-244
Bond, T., Templeton, M.R., 2011. Nitrosamine formation from the oxidation of secondary
amines. Wa. Sci. Technol. 11 (3), 259-265
Carey, P.F., Martin, L.E., Owen, P.E., 1981. Determination of ranitidine and its metabolites in
human urine by reversed-phase ion-pair high-performance liquid chromatography. J.
Chromatogr. 225, 161-168
Castiglioni, S., Bagnati, R., Fanelli, R., Pomati, F., Calamari, D., Zuccato, E., 2006. Removal of
pharmaceuticals in sewage treatment plants in Italy. Environ. Sci. Technol. 40, 357-363
Chemspider, the free chemical database. Retrieved from http://www.chemspider.com/
(compounds calculated on July 10th, 2012)
Chen, W.H., Young, T.M., 2008. NDMA formation during chlorination and chloramination of
aqueous diuron solutions. Environ. Sci. Technol. 42, 1072-1077
Chen, Z., Valentine, R.L., 2007. Formation of N-Nitrosodimethylamine (NDMA) from humic
substances in natural water. Environ. Sci. Technol. 41, 6059-6065
Dotson, A., Westerhoff, P., Krasner, S.W., 2007. Nitrosamine Formation from Natural Organic
Matter Isolates and Sunlight Photolysis of Nitrosamines. In: Proceedings of the AWWA Annual
Conference and Exposition, Toronto, ON, June 24-28, 2007
EPA Integrated Risk Information System (IRIS), 1993. N-Nitrosodimethylamine; CASRN 62-
75-9. Retrieved from www.epa.gov/iris/subst/0045.htm (accessed June 26, 2012)
Gerecke, A.C., Sedlak, D.L., 2003. Precursors of N- Nitrosodimethylamine in natural waters.
Environ. Sci. Technol. 37, 1331-1336
89
Graham, J.E., Andrews, S.A., Farquhar, G.J., Meresz, O., 1995. Factors affecting NDMA
formation during drinking water treatment. In: Proceedings of the AWWA Water Quality
Technology Conference, New Orleans, LA, November 12-16, 1995
Isidori, M., Parrella, A., Pistillo, P., Temussi, F., 2009. Effects of ranitidine and its
photoderivatives in the aquatic environment. Environ. Int. 35, 821-825
Jjemba, P.K., 2006. Excretion and ecotoxicity of pharmaceutical and personal care products in
the environment. Ecotox. Environ. Safe 63, 113-130
Kasprzyk-Hordern, B., Dinsdale, R.M., Guwy, A.J., 2008. The occurrence of pharmaceuticals,
personal care products, endocrine disruptors and illicit drugs in surface water in South Wales,
UK. Water Res. 42, 3498-3518
Kemper, J.M., Walse, S.S., Mitch, W.A., 2010. Quaternary amines as nitrosamine precursors: a
role for consumer products? Environ. Sci. Technol. 44, 1224-1231
Kohut, K.D., Andrews, S.A., 2003. Polyelectrolyte age and N-nitrosodimethylamine formation
in drinking water treatment. Water Qual. Res. J. Can. 38 (4), 719-735
Kolpin, D.W., Furlong, E.T., Meyer, M.T., Thurman, E.M., Zaugg, S.D., Barber, L.B., Buxton
H.T., 2002. Pharmaceuticals, Hormones, and Other Organic Wasterwater Contaminants in U.S.
Streams, 1999 - 2000: A National Reconnaissance. Environ. Sci. Technol. 36, 1202-1211
Kolpin, D.W., Skopec, M., Meyer, M.T., Furlong, E.T., Zaugg, S.D., 2004. Urban contribution
of pharmaceuticals and other organic wastewater contaminants to streams during differing flow
conditions. Sci. Total Environ. 328, 119-130
Krasner, S.W., Garcia, E.A., Dale, M.S., Labernik, S.M., Yun, T.I., 2008. Source and removal of
NDMA precursors. In: Proceedings of the AWWA Annual Conference and Exposition, Atlanta,
GA, June 8-12, 2008
90
Lee, C., Schimidt, C., Yoon, J., von Gunten, U., 2007. Oxidation of N-nitrosodimethylamine
(NDMA) precursors with ozone and chlorine dioxide: kinetics and effect on NDMA formation
potential. Environ. Sci. Technol. 41, 2056-2063
Lijinsky, W., Taylor, H.W., 1977. Feeding tests in rats on mixtures of nitrite with secondary and
tertiary amines of environmental importance. Food Chem. Toxicol. 15, 269-274
Massachusetts Department of Environmental Protection (MassDEP), 2004. Current Regulatory
Limit: n-Nitrosodimethylamine (NDMA), CASRN 62759. Retrieved from
http://www.mass.gov/dep/water/drinking/standards/ndma.htm (accessed June 26, 2012)
Miao, X.S., Bishay, X.S., Chen, M., Metcalfe, C.D., 2004. Occurrence of antimicrobials in the
final effluents of wastewater treatment plants in Canada. Environ. Sci. Technol. 38, 3533-3541
Mitch, W.A., Oelker, G.L., Hawley, E.L., Deeb, R.A., Sedlak, D.L., 2005. Minimization of
NDMA formation during chlorine disinfection of municipal wastewater by application of pre-
formed chloramines. Environ. Eng. Sci. 22 (6), 882-890
Mitch, W.A., Schreiber, I.M., 2008. Degradation of tertiary alkylamines during
chlorination/chloramination: implications for formation of aldehydes, nitriles, halonitroalkanes,
and nitrosamines. Environ. Sci. Technol. 42, 4811-4817
Mitch, W.A., Sedlak, D.L., 2004. Characterization and fate of N-nitrosodimethylamine
precursors in municipal wastewater treatment plants. Environ. Sci. Technol. 38, 1445-1454
Mitch, W.A., Sharp, J.O., Trussell, R.R., Valentine, R.L., Alvarez-Cohen, L., Sedlak, D.L., 2003.
N-Nitrosodimethylamine (NDMA) as a Drinking Water Contaminant: A Review. Environ. Eng.
Sci. 20, 389-404
MOE, 2003. Ontario regulation 268/03 made under the safe drinking water act, 2002. Retrieved
from http://www.e-laws.gov.on.ca/html/source/regs/english/2003/elaws_src_regs_r03268_e.htm
(accessed June 26, 2012)
91
MOE, 2006. Procedure for Disinfection of Drinking Water in Ontario. Retrieved from
http://www.ene.gov.on.ca/environment/en/resources/STD01_076415.html (accessed July 9,
2012)
Najm, I., Trussell, R.R., 2001. NDMA formation in water and wastewater. J. Am. Water Works
Assoc. 93 (2), 92-99
Office of Environmental Health Hazard Assessment (OEHHA), 2006. Public Health Goal for N-
nitrosodimethylamine and cadmium in drinking water. Retrieved from
http://www.oehha.org/water/phg/cadndma122206.html (accessed June 26, 2012)
Schmidt, C.K., Brauch, H.J., 2008. N,N-Dimethylsulfamide as precursor for N-
Nitrosodimethylamine (NDMA) formation upon ozonation and its fate during drinking water
treatment. Environ. Sci. Technol. 42, 6340-6346
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
USEPA, 1999. Alternative Disinfectants and Oxidants Guidance Manual, Chapter 6:
Chloramines.
Wilczak, A., Assadi-Rad, A., Lai, H.H., Hoover, L.L., Smith, J.F., Berger, R., Rodigari, F.,
Beland, J.W., Lazzelle, L.J., Kincannon, E.G., Baker, H., Heaney, C.T., 2003. Formation of
NDMA in chloraminated water coagulated with DADMAC cationic polymer. J. Am. Water
Works Assoc. 95 (9), 94-106
Zuccato, E., Castiglioni, S., Fanelli, R., 2005. Identification of the pharmaceuticals for human
use contaminating the Italian aquatic environment. J. Hazard. Mater. 122, 205-209
92
Chapter 5 NDMA Formation From Four Pharmaceuticals: Reaction Kinetics and
Water Matrix Effects
Most of this chapter has been previously published as:
Shen, R., Andrews, S.A., 2011b. NDMA formation kinetics from three pharmaceuticals in
four water matrices. Water Research, 45, 5687-5694.
Experiments that formed the basis of the published paper were performed for three
pharmaceuticals (chlorphenamine, doxylamine, and ranitidine) in four water matrices (lab-grade
Milli-Q® (MQ) water, Toronto tap water, Lake Ontario water and Otonabee River water). These
include the parts of Section 5.3.1 that investigate NDMA formation kinetics from the three
pharmaceuticals in MQ water; parts of Section 5.3.2 that investigate NDMA formation kinetics
in real water matrices and proposes the NOM-pharmaceutical binding theory that explains the
longer initial lag phase in NDMA formation in the presence of NOM; parts of Section 5.3.3 that
discuss the kinetic model development and verification.
Data that have been added to this chapter include results from one additional pharmaceutical
(sumatriptan) and one more water matrix (the treated Otonabee River water). Results published
in the paper were reported for one set of samples only to maximize the clarity of the findings
within the word limit of the journal. In this chapter, the NDMA formation kinetics from
ranitidine was repeated in Lake Ontario water and Otonabee River water to confirm its
reproducibility (Appendix 8). Results in Sections 5.3.1, 5.3.2, and 5.3.3 have been updated to
include all of the experimental data for all four pharmaceuticals tested in the five water matrices.
In addition, the correlation between the kinetic model parameters and the water quality
measurements (in Section 5.3.3) was updated to include the liquid chromatography – organic
carbon detection (LC-OCD) analysis results.
93
Abstract
N-nitrosodimethylamine (NDMA) is an emergent disinfection by-product (DBP) that has been
widely detected in many drinking water systems and commonly associated with the chloramine
disinfection process. Some amine-based pharmaceuticals have been demonstrated to form
NDMA during chloramination, but studies regarding the reaction kinetics are largely lacking.
This chapter investigates the NDMA formation kinetics from chlorphenamine, doxylamine,
ranitidine and sumatriptan under practical chloramine disinfection conditions. The formation
profile was monitored in both lab-grade water and real water matrices, and a statistical model is
proposed to describe and predict the NDMA formation from selected pharmaceuticals in various
water matrices. The results indicate the significant impact of water matrix components and
reaction time on the NDMA formation from selected pharmaceuticals, and provide fresh insights
on the estimation of ultimate NDMA formation potential from pharmaceutical-based precursors.
Keywords
NDMA Chlorphenamine Doxylamine Ranitidine Sumatriptan Chloramination Kinetics
94
5.1 Introduction
N-nitrosodimethylamine (NDMA) is an emerging disinfection by-product (DBP) that has been
widely detected in many drinking water systems and commonly associated with the chloramine
disinfection process. There is growing concern regarding the health effects associated with
exposure to nitrosamines because of their potential carcinogenicity (EPA IRIS, 1993). A number
of research efforts have been invested in identifying potential NDMA precursors relevant to
drinking water. Theoretically, any amine compounds containing dimethylamine (DMA) groups
may react with chloramine to form NDMA. Typical precursors found in source water include
some tertiary and quaternary amines (Kemper et al., 2010; Mitch et al., 2003; Mitch and
Schreiber, 2008), and fractions of natural organic matter (NOM) (Chen and Valentine, 2007;
Dotson et al., 2007; Gerecke and Sedlak, 2003; Mitch and Sedlak, 2004). Some chemicals used
in water treatment processes may also contribute to NDMA formation, such as certain amine-
based polymers and anion exchange resins (Kohut and Andrews, 2003; Mitch and Sedlak, 2004;
Najm and Trussell, 2001; Wilczak et al., 2003). More recently, dimethylsulfamide, a degradation
product of the fungicide tolyfluanide, was newly identified as an NDMA precursor during
ozonation (Schmidt and Brauch, 2008). Pharmaceuticals first came to attention as potential
NDMA precursors when ranitidine was demonstrated to convert into NDMA at a high
conversion rate during chloramination (Schmidt et al., 2006). Krasner (2009) has suggested that
amine-based pharmaceuticals and their breakdown products might be part of the NDMA
precursor pool in wastewater effluent organic matter (EfOM).
Up to now, studies on NDMA formation via pharmaceuticals have been mostly conducted using
lab-grade water. Specifically, data regarding the reaction kinetics in real water matrices are
largely lacking. Krasner et al. (2010) have investigated the NDMA formation over time from
ranitidine under different pH and temperature, but only conducted the experiments in deionized
water. Due to the lack of knowledge about the reactivity and chemistry, it is difficult to predict
NDMA formation from pharmaceuticals using traditional kinetic models. In the literature, some
kinetic models have been developed for the prediction of NDMA formation from DMA (Choi
and Valentine, 2002; Kim and Clevenger, 2007), NOM (Chen and Valentine, 2006), and EfOM
(Chen and Westerhoff, 2010); however, these models use comparable concentrations of
precursors and chloramines, and thus might not apply to pharmaceuticals which are usually
95
present at trace levels in the source water and are at much lower concentrations relative to
chloramine concentrations in real samples.
In Chapter 4 the author has demonstrated the formation of nitrosamines from twenty amine-
based pharmaceuticals and personal care products (PPCPs) upon chloramine disinfection. This
Chapter investigates the NDMA formation kinetics from four amine-based pharmaceuticals in
five different water matrices, and proposes a statistical model to describe and predict the NDMA
molar conversion from selected pharmaceuticals during chloramination. The four
pharmaceuticals chosen for study (chlorphenamine, doxylamine, ranitidine, and sumatriptan) had
the highest NDMA molar conversions among the twenty PPCPs tested in Chapter 4.
5.2 Materials and Methods
NDMA formation experiments were conducted at bench scale using 1 L amber bottles with
LDPE (low-density polyethylene) caps, and the formation protocol was described in detail in
Section 3.3. NDMA formation from each pharmaceutical was monitored for up to 144 hr based
on preliminary results, depending on the compound and water matrix. At each time point
samples were prepared in duplicate or triplicate, along with one blank control using the
respective water matrix to account for the potential background interference. Error bars in the
kinetic graphs represent the maximum and minimum values in the formation potential tests under
the same reaction conditions (n = 2), or the variability (standard deviation) due to multiple
formation potential tests under the same reaction conditions (n = 3).
Experiments were conducted in five water matrices, including lab-grade Milli-Q® water (MQ,
MilliPore, Etobicoke, Ontario), Lake Ontario water (LW), Toronto tap water (Tap), Otonabee
River water (RW), and treated Otonabee River water prior to chlorination (TRW). The water
sources and basic water quality parameters are summarized in Table 5.1. A quantitative analysis
of the NOM in each water matrix was performed using the liquid chromatography – organic
carbon detection (LC-OCD). The LC-OCD results for the selected water matrices are
summarized in Table 5.2, and details concerning the LC-OCD analysis were presented in Section
3.2.3.
96
Table 5.1. Basic water quality parameters (n = 5-10) and the chloramine dosage applied in each matrix under the simulated distribution system (SDS) conditions
Water source Sampling date
TOC (mg/L) pH Alkalinity
(mg/L) UV254 (cm-1)
SUVA (L/mg·m)
Chloramine dosage (mg/L)
Milli-Q® (MQ) NA* 0 7.5 ± 0.1 1.8 ± 0.3 0.000 0.0 2.5 ± 0.2
Toronto tap water (Tap)
NA* 2.1 ± 0.1 7.1 ± 0.1 88.5 ± 2.8 0.021 ± 0.001 1.0 ± 0.1 2.5 ± 0.2
Lake Ontario (LW)
June 2010 2.3 ± 0.2 8.0 ± 0.1 94.6 ± 1.5 0.024 ± 0.002 1.1 ± 0.1 2.5 ± 0.2 Sept. 2011 2.3 ± 0.1 8.2 ± 0.1 95.4 ± 1.8 0.023 ± 0.002 1.0 ± 0.1 2.5 ± 0.2
Otonabee River (RW)
Oct. 2010 6.2 ± 0.5 7.8 ± 0.1 86.5 ± 1.7 0.143 ± 0.003 2.3 ± 0.2 4.4 ± 0.3 Jan. 2012 6.2 ± 0.1 7.8 ± 0.1 94.9 ± 2.9 0.162 ± 0.002 2.6 ± 0.1 3.5 ± 0.2
Treated Otonabee River (TRW)
Jan. 2012 3.4 ± 0.1 7.2 ± 0.1 77.9 ± 0.1 0.067 ± 0.001 1.9 ± 0.1 3.8 ± 0.2
* Not applicable
Table 5.2. LC-OCD results for the selected water matrices (Unit: mg/L carbon)
Water source DOC Hydrophobic
DOC Hydrophilic
DOC
Hydrophilic DOC fractions
Biopolymers Humics Building blocks
LMW neutrals
LMW acids
MQ 0.07 0.03 0.04 0.01 0.000 0.003 0.02 0.004 Tap 1.91 0.40 1.51 0.19 0.64 0.40 0.18 0.10 LW 2.20 0.20 2.00 0.34 0.90 0.46 0.19 0.10 RW 5.90 0.20 5.70 0.34 3.82 0.72 0.62 0.20
TRW 3.40 0.40 3.00 0.14 1.55 0.69 0.41 0.19
The raw, untreated lake and river water samples were taken from the influents of two drinking
water treatment plants; the treated river water (prior to chlorination) was collected in parallel
with the raw river water during the sampling in January 2012. The experiments were carried out
under the simulated distribution system (SDS) conditions (pH = 7.0 ± 0.1; 21 °C; chloramine
dosage = 2.5 ± 0.2 mg/L after satisfying 24hr chloramine demand) using preformed
monochloramine (Cl2:NH4-N mass ratio = 4.2:1), except that instead of limiting the incubation
time to 24 hr, it varied up to 144 hr depending on the pharmaceuticals and water matrices. The
actual chloramine dosage applied in each matrix is also summarized in Table 5.1 (see details in
Section 3.3). Further details about the NDMA extraction, concentration, analysis, and
quantification methods are provided in Section 3.2.1.
97
5.3 Results and Discussion
5.3.1 Formation Kinetics in MQ Water
Kinetic experiments in MQ water were conducted for all four pharmaceuticals at two
concentration levels (5 and 25 nM), as shown in Figure 5.1. The markers in the figure are the
measured NDMA molar conversion values, and the lines are model-estimated results. Details
about the model development and estimation will be discussed in Section 5.3.3.
Figure 5.1. NDMA molar conversion over time for chlorphenamine, doxylamine, ranitidine, and sumatriptan in MQ water (SDS conditions; error bars represent the maximum and minimum values in the
formation potential tests under the same conditions (n = 2, chlorphenamine and doxylamine), and represent the variability due to multiple formation potential tests (n = 3, ranitidine and sumatriptan))
98
NDMA formation via the four pharmaceuticals followed similar pattern over time. Generally, an
initial lag period was observed, followed by a fast increase in NDMA concentration; the molar
conversion then gradually levelled off and eventually reached a plateau (maximum molar
conversion). Moreover, the formation kinetic behavior was observed to be relatively independent
of the initial pharmaceutical concentration, except that NDMA formation from doxylamine in
MQ water showed a more significant difference after 24 hr than did the other pharmaceuticals.
Given the large excess of chloramine relative to the pharmaceuticals (mg/L vs. lower µg/L),
availability of chloramine was not a limiting factor at the concentration range of pharmaceuticals
tested. These results also support observations made in Chapter 4 where the NDMA molar
conversion at 24 hr from the selected pharmaceuticals was found to be independent of their
initial concentrations.
5.3.2 Formation Kinetics in Different Water Matrices
Kinetic experiments were also performed using real water samples dosed with selected
pharmaceuticals. The same shape of NDMA formation curve was observed in real water matrices
as in MQ water except that the initial lag phase was longer, especially for tests performed using
lake and river water (Figure 5.2). Similarly to Figure 5.1, the markers are the measured NDMA
molar conversion values, and the lines are the model-estimated results that will be discussed in
Section 5.3.3. Although the NDMA formation kinetics was shown to be unique to each water
matrix tested, the kinetic behavior was relatively independent of the initial pharmaceutical
concentration within a given water matrix, further confirming that observation in MQ water.
99
Figure 5.2. NDMA molar conversion over time for chlorphenamine, doxylamine, ranitidine, and sumatriptan in different water matrices (SDS conditions; error bars represent the maximum and minimum
values in the formation potential tests under the same conditions (n = 2, chlorphenamine and doxylamine), and represent the variability due to multiple tests (n = 3, ranitidine and sumatriptan))
The different NDMA formation profiles were likely influenced by the water matrix components,
rather than by added reagents, since the pH of the water samples was controlled with a phosphate
buffer and the same chloramine dosage was applied to all samples. Both bromide and NOM
have been shown to influence NDMA formation, with bromide being reported to either catalyze
NDMA formation (Le Roux et al., 2012; Mitch et al., 2003; Valentine et al., 2005) or have an
inhibitory effect (Chen et al., 2010; Le Roux et al., 2012;). The bromide levels in the water
sources selected for this work are typically low (Lake Ontario, ~ 40 µg/L (Comerton et al.,
100
2006); Otonabee River, < 11.4 µg/L (Woodbeck, 2007)), while bromide concentrations applied
were much higher (mg/L) for studies that have reported these effects, so the differences in the
observed NDMA formation profiles were thought to be due to some aspect of the NOM. Since
bromide is in higher concentration and so may be more of a concern in coastal waters due to
saltwater intrusion, the potential impact from bromide was considered to be outside the scope of
the present tests but would be of interest for future study on NDMA formation in coastal waters.
NOM may affect NDMA formation in two ways: competition for chloramine and direct
interaction with PPCPs. The influence of NOM’s competition for chloramine was considered to
be minimal due to the small observed chloramine decay (less than 50 % chloramine decay during
the course of the kinetic experiments, see Appendix 2, Figure A2.2) and the large excess of
chloramine relative to the pharmaceuticals (mg/L vs. lower µg/L) at the end of the kinetic
experiment. On the other hand, NOM may interact with the pharmaceuticals and then inhibit
their reaction to form NDMA, and it is these interactions that were thought to better explain the
observed results. Previous studies have demonstrated that aromatic amines undergo reversible
covalent binding with carbonyls and quinones in soil humic substances in the environment
(Parris, 1980; Thorn et al., 1996; Weber et al., 1996). For example, Weber et al. (1996) have
observed significant binding at a concentration of 466 µg/L of aniline and 250 mg/L of humic
substances, i.e., 1.86 µg/mg sorption ratio; while in this work, 6 ~ 8 µg/L of selected
pharmaceuticals were dosed into 2 ~ 6 mg/L of TOC (the majority of which was humic
substances, see details in Section 3.1.3). Therefore, the potential sorption ratio was relatively
comparable with the study by Weber et al (1996). Considering the trace level of pharmaceuticals
relative to the amount of NOM in natural water samples, a similar scenario with the sorption of
anilines onto soil/sediment organic matter could exist. Therefore, it is possible that certain
fractions or functional groups in NOM may interact with these amine-based pharmaceuticals and
thus hinder their initial contact with chloramine species. As the binding is reversible and
chloramine is in large excess, eventually the NDMA conversion from pharmaceuticals can still
reach the maximum level given enough reaction time.
Currently, although no direct spectroscopic evidence exists for the NOM-pharmaceutical binding
in the aqueous phase, this theory is indirectly supported by some literature investigating the
101
removal of pharmaceuticals during coagulation/flocculation processes, where the removal of
pharmaceuticals was likely due to the sorption or electrostatic attraction onto particulate organic
matter and co-removed through the settling process (Ballard and Mackay, 2005; Diemert, 2012;
Stackelberg et al., 2007; Vieno et al., 2006; Westerhoff et al., 2005). Stackelberg et al. (2007)
also detected the target pharmaceuticals in the dried solids of settled sludge. In addition, de
Ridder et al. (2011) observed enhanced removal of some positively charged pharmaceuticals
using granular activated carbon preloaded with NOM. They attributed the enhancement to the
electrostatic attraction since the surface of NOM is usually negatively charged due to the
abundance of carboxyl groups. In the current study, the selected amine pharmaceuticals are
positively charged at neutral pH, therefore this possible electrostatic attraction may also lead to
the formation of NOM-pharmaceutical complexes.
The proposed NOM-pharmaceutical binding theory can well explain the initial lag phase
observed in the kinetic study, thus theoretically the length of the initial lag phase should be
proportional to the amount and/or type of NOM in the water samples. NOM components can be
at least partially described by the samples’ TOC and SUVA values. It was observed that water
with higher TOC and SUVA levels tended to have a longer initial lag phase; all four
pharmaceuticals exhibited their longest initial lag period in river water samples. However, the
treated river water had higher TOC and SUVA values than the lake water, yet ranitidine showed
a slightly shorter initial lag phase in the treated river water (Figure 5.2). According to the
literature, binding usually occurs between aromatic amines and some specific functional groups
in the humic substances; therefore, it is the amount of these functional groups that directly affects
the binding, rather than the amount of NOM. The treated river water and lake water came from
different sources, thus they could have different compositions of surface functional groups. In
addition, the treatment processes at the water treatment plant, especially the
coagulation/flocculation process, not only remove NOM, but can also change its surface
characteristics (e.g., steric structure, charge distribution, etc.). This suggests that some specific
NOM fractions or moieties might be more relevant than would be indicated by simple bulk
measurements of water quality, such as TOC and SUVA.
102
Moreover, although tap and lake water samples had similar TOC and SUVA values, ranitidine
showed a longer initial lag phase in lake water samples (Figure 5.2). This is most likely because
of the presence of chloramine residual in the tap water (before the pharmaceuticals were dosed)
which may have reacted with the “binding sites” on the NOM surface and thus reduced the
binding potential. In order to confirm this theory, the lake water was dosed with 1.0 mg/L
preformed monochloramine (consistent with general chloramine residuals in Toronto tap water)
24 hr before the experiments to mimic the tap water. On the day of the tests, ranitidine was dosed
into the pre-chloraminated lake water samples, and then the chloramine dosage was boosted to
2.5 mg/L. The results were compared in Figure 5.3. Considering that Lake Ontario water is the
source of Toronto tap water, the NDMA formation kinetic profiles for the pre-chloraminated lake
water and the tap water were in good agreement with each other.
Figure 5.3. NDMA formation kinetics for ranitidine (25 nM) in pre-chloraminated Lake Ontario water and Toronto tap water (SDS conditions; error bars represent the variability due to multiple formation
potential tests (n = 3))
103
5.3.3 Kinetic Model
The NDMA formation curves for the pharmaceuticals in this study have a sigmoidal shape that
resembles the typical shape of dose-response curves. A standard dose-response curve can be
defined by a four-parameter logistic function,
𝑦 = 𝑏 + 𝑎 − 𝑏
1 + 10�×(���)
where y is the response caused by a certain dose of pharmaceuticals (x); a and b are the
maximum and baseline response, respectively; c is the slope of the curve; and d is the dose which
provokes a response halfway between the baseline and maximum (Motulsky and Christopoulos,
2003). Accordingly, the following model was proposed to describe the reaction kinetics for
NDMA formation from selected pharmaceuticals,
𝑌 = 𝜃
1 + 10�×(�����)
where Y is the NDMA molar conversion at given reaction time (t); θ is the ultimate NDMA
molar conversion, i.e., the maximum molar conversion obtained at the plateau during kinetic
testing; k is the pseudo-first order reaction rate constant; Lag is the time required to achieve 50 %
of the ultimate molar conversion, and thus is associated with the length of initial lag phase
observed. Comparing this model with the four-parameter logistic function, the parameter b was
set to zero because any possible NDMA in the background and any potential NDMA formed
from the matrix components were accounted for by the blank control samples. It is noted that the
proposed model does not pass through the point of (0, 0), although, once background NDMA has
been subtracted, there should be zero molar conversion at the beginning. Since the dose-response
model is based on log (drug dose), it is always positive on the x-axis, and thus does not go
through the point of (0, 0). Therefore, the proposed model was arbitrarily set to be:
𝑌 = � 0 (𝑡 = 0)
𝜃1 + 10�×(�����) (𝑡 > 0)
�
The formation curve was fitted using GraphPad Prism 5® software, and the estimated model
parameters for each compound in different matrices are summarized in the Parameter Estimation
section of Table 5.3.
104
Table 5.3. Kinetic model parameter estimation and model verification
Compound Concentration Matrix Parameter Estimation Model Verification
θ a Lag (hr) a k (hr-1) a R2 Model-predicted conversion @ 24 hr
Measured conversion @ 24 hr b
Chlorphenamine 5 nM MQ 0.027 (0.003) 7.0 (1.3) 0.143 (0.050) 0.974 2.7% 2.9 % (0.2 %) TAP 0.023 (0.003) 10.8 (3.0) 0.065 (0.024) 0.953 2.0% 2.0 % (0.8 %) RW 0.037 (0.002) 38.4 (3.4) 0.039 (0.007) 0.990 0.8% 1.0 % (0.1 %) 25 nM MQ 0.033 (0.002) 8.7 (0.6) 0.234 (0.079) 0.994 3.3% 1.8 % (0.1 %) TAP 0.030 (0.002) 19.4 (1.8) 0.070 (0.015) 0.988 2.0% 1.5 % (0.1 %) RW 0.043 (0.002) 39.9 (1.9) 0.056 (0.009) 0.996 0.5% 0.5 % (0.02 %) Doxylamine 5 nM MQ 0.062 (0.005) 16.6 (2.2) 0.068 (0.017) 0.975 4.7% 3.8 % (0.1 %) TAP 0.068 (0.006) 48.8 (5.3) 0.027 (0.005) 0.993 1.2% 2.5 % (0.2 %) RW 0.059 (0.003) 67.0 (4.3) 0.026 (0.006) 0.990 0.4% 1.1 % (0.03 %) 25 nM MQ 0.106 (0.007) 22.0 (2.0) 0.069 (0.018) 0.985 6.1% 4.2 % (0.1 %) TAP 0.092 (0.006) 46.1 (4.1) 0.030 (0.005) 0.992 1.6% 3.2 % (0.3 %) RW 0.060 (0.002) 51.1 (1.8) 0.051 (0.011) 0.996 0.2% 0.5 % (0.04 %) Ranitidine 5 nM MQ 0.912 (0.045) 6.3 (0.5) 0.225 (0.043) 0.992 91.2% 85.2 % (0.8 %) TAP 0.902 (0.045) 6.7 (0.8) 0.169 (0.045) 0.986 90.1% 83.4 % (8.1 %) LW 0.729 (0.032) 13.3 (0.7) 0.251 (0.057) 0.997 72.7% 64.1 % (3.6 %) RW 0.822 (0.056) 20.8 (1.8) 0.086 (0.026) 0.984 53.6% 51.4 % (4.9 %) 25 nM MQ 0.906 (0.045) 4.6 (0.5) 0.313 (0.084) 0.991 90.6% 82.7 % (2.4 %) TAP 0.847 (0.039) 6.5 (0.7) 0.177 (0.043) 0.988 84.6% 88.4 % (5.9 %) LW + NH2Cl c 0.828 (0.029) 7.0 (0.4) 0.190 (0.030) 0.993 NA d NA LW 0.824 (0.025) 12.5 (0.7) 0.182 (0.028) 0.993 81.7 % 70.1 % (4.8 %) TRW 0.895 (0.012) 9.2 (0.3) 0.250 (0.040) 0.998 NA NA RW 0.841 (0.044) 21.9 (1.4) 0.083 (0.021) 0.990 50.4% 43.2 % (7.1 %) Sumatriptan 5 nM MQ 0.023 (0.002) 22.8 (2.7) 0.080 (0.040) 0.966 1.3 % 1.3 % (0.2 %) 25 nM MQ 0.028 (0.001) 27.0 (2.4) 0.060 (0.010) 0.992 1.1 % 2.0 % (0.1 %) LW 0.022 (0.001) 31.8 (2.3) 0.060 (0.010) 0.991 0.6 % 1.6 % (0.1 %) TRW 0.025 (0.001) 32.5 (2.5) 0.040 (0.010) 0.988 NA NA RW 0.024 (0.002) 66.8 (4.0) 0.030 (0.010) 0.988 NA NA a Numbers in these brackets represent the 95 % confidence interval of each model parameter. b Numbers in these brackets represent the standard deviation from multiple tests (n = 3). c Pre-chloraminated lake water mimicked the tap water scenario. d Not available. Independent 24 hr NDMA-FP experiments were not performed.
105
The proposed model fit the experimental data very well, with correlation coefficients (R2) higher
than 0.95 in all cases, and predicted accurately all three phases of the NDMA formation curve, as
shown previously in Figure 5.1 and Figure 5.2. It is worth noting that the model requires data
capturing all three phases of the NDMA formation curve in order to acquire reliable model
parameters. For datasets lacking the plateau data, the model will arbitrarily assume the last point
as the plateau. In this study, the NDMA formation curve for ranitidine (5 nM) in LW samples did
not achieve the plateau within the 24hr of the experiment (Figure 5.2). Therefore, the calculated
θ (0.729 ± 0.032) may underestimate the ultimate NDMA molar conversion.
The estimated model parameters each well reflected the different aspects of the NDMA
formation profiles that were observed in the different water matrices. Generally, the matrix had a
minor impact on the ultimate NDMA molar conversion for chlorphenamine (θ = 3.2 ± 0.7 %;
mean ± standard deviation of θ in all the matrices), doxylamine (θ = 7.5 ± 2.0 %), ranitidine (θ =
85.1 ± 5.6 %), and sumatriptan ((θ = 2.4 ± 0.2 %). Instead, the matrix components had a more
profound impact on the initial lag phase (Lag) and the pseudo-first order rate constant (k). As
summarized in Figure 5.4, the Lag value is positively correlated with both TOC and SUVA
values for all four pharmaceuticals; k value is negatively correlated with TOC and SUVA values
for chlorphenamine, ranitidine and sumatriptan, but not well related for doxylamine. The
estimated model parameters and these correlations support the theory that water matrix
components can affect NDMA formation from selected pharmaceuticals by inhibiting the initial
reaction with chloramine and/or slowing down subsequent reactions. In addition, the potential
correlation was also assessed between the model parameters (Lag and k) and the NOM fractions
analysed by LC-OCD. The correlation coefficients (R2) for all of the fractions are summarized in
Table 5.4 and compared with the coefficients for TOC and SUVA. No specific fraction
correlated particularly well with the model parameters across the board, and there was no
significant improvement on the R2 for the NOM fractions compared with those for TOC and
SUVA. This further confirms that the matrix effect on the NDMA formation kinetics is not
affected by the amount of NOM alone, and the amount of certain moieties or functional groups
might be more directly relevant.
106
Figure 5.4. Linear correlation between (a) Lag and TOC; (b) Lag and SUVA; (c) k and TOC; and (d) k and SUVA for four pharmaceuticals (SDS conditions; pharmaceutical concentration = 5 and 25 nM; error
bars represent the 95 % confidence interval for the estimated model parameters)
Table 5.4. Correlation coefficients (R2) between the kinetic parameters and water quality parameters
Parameter Compound TOC SUVA Biopolymers Humics Building blocks
LMW neutrals
LMW acids
Lag Chlorphenamine 0.952 0.917 0.851 0.960 0.839 0.964 0.901 Doxylamine 0.768 0.834 0.889 0.626 0.895 0.702 0.852 Ranitidine 0.829 0.676 0.652 0.875 0.570 0.774 0.593 Sumatriptan 0.841 0.629 0.494 0.939 0.490 0.781 0.534
k Chlorphenamine 0.594 0.673 0.749 0.445 0.759 0.522 0.696 Doxylamine 0.286 0.375 0.481 0.153 0.497 0.217 0.404 Ranitidine 0.675 0.523 0.434 0.653 0.443 0.579 0.460 Sumatriptan 0.770 0.741 0.338 0.769 0.638 0.799 0.701
107
Chen and Westerhoff (2010) recently found NDMA-FP very difficult to predict based upon bulk
water quality measurements such as TOC or UVA254. Results from the current study have
indicated that knowledge of the reaction kinetics is essential in the prediction of NDMA
formation. While typical water quality measurements like TOC and SUVA can have significant
impact on the reaction kinetics, they are not directly associated with the ultimate NDMA molar
conversion, and thus are not appropriate for directly predicting NDMA-FP using empirical
models. Moreover, knowledge of the reaction time employed is specifically crucial for
compounds that react slowly with chloramine, such as doxylamine and sumatriptan. Recently,
more and more utilities have shown interest in conducting NDMA-FP tests. Because there is lack
of standard protocol at the moment, many are considering adopting typical disinfection by-
product formation potential tests (Summers et al., 1996) and have applied a 24 hr incubation time
from a practical viewpoint; however, the NDMA formed after 24 hr from some compounds may
only represent a small portion of their ultimate formation potential, especially in real water
matrices where the initial reaction could be significantly inhibited. For example, the NDMA
molar conversion at 24 hr for doxylamine in TAP and RW samples only accounted for less than
20 % of its ultimate NDMA molar conversion, and the NDMA molar conversion at 24 hr for
sumatriptan in RW samples only accounted for 5.5 % of its ultimate NDMA molar conversion.
Recently, Krasner et al. (2012) have developed a bench-scale test to predict the nitrosamine
formation, and they have applied a chloramination contact time of 3 days. Among the four
pharmaceuticals tested in this work, 3 days of incubation time is good enough to predict the
NDMA formation from most of them, except that the NDMA molar conversion for doxylamine
and sumatriptan in RW samples after 3 days only accounted for about 60 % of their ultimate
conversion. The results have suggested that typical bench-scale NDMA-FP tests may
underestimate the ultimate NDMA-FP for some slow-reacting precursors, and the proper contact
time required is highly case-specific, depending on the precursor properties as well as the water
matrix. For water systems with higher water age, prolonged NDMA formation in the outreaches
of the distribution system might be a potential risk and should be taken into consideration.
The model was verified by comparing the 24hr NDMA-FP predicted using the estimated model
with that measured from independent 24 hr formation potential tests, as summarized in the last
two columns of Table 5.3 (Model Verification). Linear regression was applied between the
108
measured and predicted NDMA molar conversion for each compound individually (except
sumatriptan which did not have enough data) and for the four compounds all together (Figure
5.5). This goodness of fit test has suggested that there is significant correlation between the
measured and model-predicted molar conversion (p-value < 0.05, F- test, 95 % confidence level),
except for chlorphenamine (p-value = 0.0546), although even this correlation was determined to
be significant at 90 % confidence level. Following the F-test, the Student’s t-test (95 %
confidence level) was conducted to determine whether the slope of each regression line differed
significantly from 1.0 (i.e., p-value < 0.05). The slopes in Figure 5.5 were reported as “the best
fit value ± standard error” or as 1.0 if the t-test suggests so. The t-test has indicated that the
slopes of the regression line for chlorphenamine and doxylamine were much lower than 1.0 (p-
value of 0.0071 and < 0.0001, respectively), while the slope for ranitidine was no different from
1.0 (p-value of 0.2075). The slope for all four pharmaceuticals together was slightly lower than
1.0 (p-value < 0.0001); however, the model predicted molar conversion in general was within the
95 % confidence interval of the measured values.
109
Figure 5.5. Linear correlation between the model-predicted and the independently measured NDMA molar conversion at 24 hr for chlorphenamine, doxylamine, ranitidine, and all four pharmaceuticals
together (SDS conditions; data from four matrices (MQ, Tap, LW, and RW) and two pharmaceutical concentration levels (5 and 25 nM); error bars represent the standard deviation from multiple formation
potential tests (n = 3))
The kinetic model will be applied and further verified under different reaction conditions in
Chapter 6 (under different pH values) and Chapter 7 (upon sequential chlorination and
chloramination). All of the estimated kinetic model parameters under different reaction
conditions are summarized in Appendix 9. A more comprehensive model verification using all of
the experimental data is also presented in Appendix 9, along with all related statistical analyses.
110
5.4 Summary
NDMA formation kinetics from chlorphenamine, doxylamine, ranitidine and sumatriptan during
chloramination was determined in five water matrices. The NDMA conversion over time
followed a general three-phase formation curve: an initial lag phase was observed, followed by a
fast increase in NDMA formation, and eventually a plateau was reached that represented the
ultimate NDMA molar conversion. The NDMA formation profile was relatively independent of
the initial pharmaceutical concentration in the same matrix. Water matrix components affected
the NDMA conversion rates, most likely by inhibiting their initial contact with chloramine and
slowing down the reaction, while they had less impact on the ultimate NDMA molar conversion.
A three-parameter kinetic model was proposed to describe the NDMA formation over time
during chloramination. The model accurately reflected all the three significant characteristics of
the NDMA formation curve, and was able to predict the NDMA molar conversion from the
selected pharmaceuticals to within the 95 % confidence interval of the measured values. The
model needs to be further verified using different potential precursors, water matrices, and
reaction conditions. Bulk water quality measurements such as TOC and SUVA were found to
correlate better with model parameters Lag and k than with the ultimate NDMA molar
conversion (θ). Some correlation was also observed between these model parameters (Lag and k)
and the NOM fractions analyzed via LC-OCD, but there was no significant improvement on the
correlation coefficient for the NOM fractions compared with those for TOC and SUVA. These
results have indicated that interactions between the pharmaceuticals and NOM that might affect
NDMA formation are not limited to those based on the general organic character or aromatic
nature of either substance; the amount of certain moieties or functional groups on the NOM
surface might be more directly relevant.
Knowledge about the formation kinetics is essential in the prediction of NDMA formation from
pharmaceuticals. Short-term NDMA-FP tests (≤ 24 hr), although practical, may underestimate
the contribution of certain slow-reacting precursors. Thus, prolonged tests (perhaps > 4 days for
substances examined in this study) are required to determine the ultimate NDMA-FP, especially
in distribution systems with long water age.
111
5.5 References
Ballard, B.D., Mackay, A.A., 2005. Estimating the removal of anthropogenic organic chemicals
from raw drinking water by coagulation flocculation. J. Environ. Eng. 131, 108-118
Chen, B., Westerhoff, P., 2010. Predicting disinfection by-product formation potential in water.
Water Res. 44, 3755-3762
Chen, Z., Valentine, R.L., 2006. Modeling the formation of N-Nitrosodimethylamine (NDMA)
from the reaction of Natural Organic Matter (NOM) with monochloramine. Environ. Sci.
Technol. 40, 7290-7297
Chen, Z., Valentine, R.L., 2007. Formation of N-Nitrosodimethylamine (NDMA) from humic
substances in natural water. Environ. Sci. Technol. 41, 6059-6065
Chen, Z., Yang, L., Zhai, X., Zhao S., Li, A., Shen, J., 2010. N-nitrosamine formation during
chlorination/chloramination of bromide-containing water. Water Sci. Technol.: Water Supply 10,
462-471
Choi, J.H., Valentine, R.L., 2002. A kinetic model of N-nitrosodimethylamine (NDMA)
formation during water chlorination/chloramination. Water Sci. Technol. 46, 65-71
Comerton, A.M., Andrews, R.C., Bagley, D.M., 2006. Impact of blending reuse and lake water
on treated water quality. J. Environ. Eng. Sci. 5: 359-363
de Ridder, D.J., Verliefde, A.R.D., Heijman, S.G.J., Verberk, Q.J.C., Rietveld, L.C., van der Aa,
L.T.J., Amy, G.L., van Dijk, J.C., 2011. Influence of natural organic matter on equilibrium
adsorption of neutral and charged pharmaceuticals onto activated carbon. Wat. Sci. Technol. 63,
416 – 423
Diemert, S.A., 2012. The impact of coagulation on endocrine disrupting compounds,
pharmaceutically active compounds and natural organic matter. Master’s thesis, Univesity of
Toronto, Ontario
112
Dotson, A., Westerhoff, P., Krasner, S.W., 2007. Nitrosamine Formation from Natural Organic
Matter Isolates and Sunlight Photolysis of Nitrosamines. In: Proceedings of the AWWA Annual
Conference and Exposition, Toronto, ON, June 24-28, 2007
EPA Integrated Risk Information System (IRIS), 1993. N-Nitrosodimethylamine; CASRN 62-
75-9. Retrieved from www.epa.gov/iris/subst/0045.htm (accessed June 26, 2012)
Gerecke, A.C., Sedlak, D.L., 2003. Precursors of N- Nitrosodimethylamine in natural waters.
Environ. Sci. Technol. 37, 1331-1336
Le Roux, J., Gallard, H., Croue, J., 2012. Formation of NDMA and halogenated DBPs by
chloramination of tertiary amines: The influence of bromide ion. Environ. Sci. Technol. 46,
1581-1589
Kemper, J.M., Walse, S.S., Mitch, W.A., 2010. Quaternary amines as nitrosamine precursors: a
role for consumer products? Environ. Sci. Technol. 44, 1224-1231
Kim, J., Clevenger, T.E., 2007. Prediction of N-nitrosodimethylamine (NDMA) formation as a
disinfection by-product. J. Hazard. Mater. 145, 270-276
Kohut, K.D., Andrews, S.A., 2003. Polyelectrolyte age and N-nitrosodimethylamine formation
in drinking water treatment. Water Qual. Res. J. Can. 38 (4), 719-735
Krasner, S.W., 2009. The formation and control of emerging disinfection by-products of health
concern. Phil Trans. R Soc. A 367, 4077-4095
Krasner, S.W., Dale, M.S., Lee, C.F.T., Garcia, E.A., Wong, T.M., Mitch, W., Von Gunten, U.,
2010. Difference in reactivity and chemistry of NDMA precursors from treated wastewater and
from polyamine polymers. In: Proceedings of the AWWA Water Quality Technology
Conference, Savannah, GA, November 14-18, 2010
113
Krasner, S.W., Lee, C.F.T., Mitch, W.A., von Gunten, U., 2012. Development of a bench-scale
test to predict the formation of nitrosamines. Denver, CO, Water Research Foundation Report
Mitch, W.A., Schreiber, I.M., 2008. Degradation of tertiary alkylamines during
chlorination/chloramination: implications for formation of aldehydes, nitriles, halonitroalkanes,
and nitrosamines. Environ. Sci. Technol. 42, 4811-4817
Mitch, W.A., Sedlak, D.L., 2004. Characterization and fate of N-nitrosodimethylamine
precursors in municipal wastewater treatment plants. Environ. Sci. Technol. 38, 1445-1454
Mitch, W.A., Sharp, J.O., Trussell, R.R., Valentine, R.L., Alvarez-Cohen, L., Sedlak, D.L., 2003.
N-Nitrosodimethylamine (NDMA) as a Drinking Water Contaminant: A Review. Environ. Eng.
Sci. 20, 389-404
Motulsky, H.J., Christopoulos, A., 2003. Fitting models to biological data using linear and
nonlinear regression. A practical guide to curve fitting. GraphPad Software Inc., San Diego, CA
Najm, I., Trussell, R.R., 2001. NDMA formation in water and wastewater. J. Am. Water Works
Assoc. 93 (2), 92-99
Parris, G.E., 1980. Covalent binding of aromatic amines to humates. 1. reactions with carbonyls
and quinones. Environ. Sci. Technol. 14, 1099-1106
Schmidt, C.K., Brauch, H.J., 2008. N,N-Dimethylsulfamide as precursor for N-
Nitrosodimethylamine (NDMA) formation upon ozonation and its fate during drinking water
treatment. Environ. Sci. Technol. 42, 6340-6346
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
114
Stackelberg, P.E., Gibs, J., Furlong, E.T., Meyer, M.T., Zaugg, S.D., Lippincott, R.L., 2007.
Efficiency of conventional drinking-water-treatment processes in removal of pharmaceuticals
and other organic compounds. Sci. Total Environ. 377, 255-272
Summers, R.S., Hooper, S.M., Shukairy H.M., Solarik, G., Owen, D., 1996. Assessing DBP
yield: uniform formation conditions. J. Am. Water Works Assoc. 88, 80-93
Thorn, K.A., Pettigrew, P.J., Goldenberg, W.S., Weber, E.J., 1996. Covalent binding of aniline
to humic substances. 2. 15N NMR studies of nucleophilic addition reactions. Environ. Sci.
Technol. 30, 2764-2775
Valentine, R.L., Choi, J., Chen, Z., Barrett, S.E., Hwang, C., Guo, Y.C., Wehner, M.,
Fitzsimmons, S., Andrews, S.A., Werker, A.G., Brubacher, C., Kohut, K., 2005. Factors
affecting the formation of NDMA in water and occurrence, Denver, CO, pp. 81-90
Vieno, N., Tuhkanen, T., Kronberg, L., 2006. Removal of pharmaceuticals in drinking water
treatment: effect of chemical coagulation. Environ. Technol. 27, 183 – 192
Weber, E.J., Spidle, D.L., Thorn, K.A., 1996. Covalent binding of aniline to humic substances. 1.
kinetic studies. Environ. Sci. Technol. 30, 2755-2763
Westerhoff, P., Yoon, Y., Snyder, S., Wert E., 2005. Fate of endocrine-disruptor,
pharmaceutical, and personal care product chemicals during simulated drinking water treatment
processes. Environ. Sci. Technol. 39, 6649-6663
Wilczak, A., Assadi-Rad, A., Lai, H.H., Hoover, L.L., Smith, J.F., Berger, R., Rodigari, F.,
Beland, J.W., Lazzelle, L.J., Kincannon, E.G., Baker, H., Heaney, C.T., 2003. Formation of
NDMA in chloraminated water coagulated with DADMAC cationic polymer. J. Am. Water
Works Assoc. 95 (9), 94-106
Woodbeck, M. 2007. Evaluating the potential genotoxicity of disinfected source waters. Master’s
thesis, University of Toronto, Ontario
115
Chapter 6 Formation of NDMA From Ranitidine and Sumatriptan: the Role of pH
This chapter has been published as:
Shen, R., Andrews, S.A., 2012. Formation of NDMA from ranitidine and sumatriptan: the
role of pH. Water Research 47, 802-810.
Results from additional experiments which investigate the impact of low pH on the NDMA
formation from ranitidine are added in Section 6.3.2.
Abstract
N-Nitrosodimethylamine (NDMA) is an emerging disinfection by-product (DBP) which can be
formed via the chloramination of amine-based precursors. The formation of NDMA is mainly
determined by the speciation of chloramines and the precursor amine groups, both of which are
highly dependent on pH. The impact of pH on NDMA formation has been studied for the model
precursor dimethylamine (DMA) and natural organic matter (NOM), but little is known for
amine-based pharmaceuticals which have been newly identified as a group of potential NDMA
precursors, especially in waters impacted by treated wastewater effluents. This chapter
investigates the role of pH in the formation of NDMA from two amine-based pharmaceuticals,
ranitidine and sumatriptan, under drinking water relevant conditions. The results indicate that pH
affects both the ultimate NDMA formation as well as the reaction kinetics. The maximum
NDMA formation typically occurs in the pH range of 7 to 8. At lower pH, the reaction is limited
due to the lack of non-protonated amines. At higher pH, although the initial reaction is enhanced
by the increasing amount of non-protonated amines, the ultimate NDMA formation is limited
because of the lack of dichloramine.
Keywords
NDMA Chloramination pH Ranitidine Sumatriptan
116
6.1 Introduction
N-nitrosodimethylamine (NDMA) is considered as an emerging disinfection by-product (DBP)
and has been commonly associated with chloramine disinfection during drinking water
treatment. There has been a surge of interest in NDMA in drinking water because of its potential
carcinogenicity at ng/L levels (EPA IRIS, 1993). With more utilities switching from free chlorine
to chloramine as the secondary disinfectant, NDMA has been detected in many chloraminated
drinking water systems across North America (Blute et al., 2010; Charrois et al., 2007; Russell
et al., 2012). Most of the time the reported NDMA concentrations have been below 10 ng/L;
only a small portion of samples showed higher concentrations (Russell et al., 2012). At the time
of writing, NDMA has been mainly regulated at regional levels. In Canada, the Ontario Ministry
of the Environment (MOE) has established a maximum acceptable level of 9 ng/L (0.12 nM) for
NDMA in drinking water (MOE, 2003), and Health Canada has proposed a maximum acceptable
concentration for NDMA of 40 ng/L (0.54 nM) in drinking water (Health Canada, 2010). In the
US, the California Department of Health Services and Massachusetts Department of
Environmental Protection have implemented an NDMA drinking water regulatory limit of 10
ng/L (0.13 nM; MassDEP, 2004; OEHHA, 2006). NDMA is on the drinking water contaminant
candidate list 3 (CCL3) together with four other nitrosamines (USEPA, 2009). In Australia,
NDMA has been included in the Australian Drinking Water Guideline at 100 ng/L (1.35 nM)
since 2011 (NHMRC, NRMMC, 2011).
Dimethylamine (DMA) is the most studied model precursor of NDMA (Bond and Templeton,
2011; Mitch et al., 2003). Although DMA is ubiquitous in natural waters, its concentration and
low NDMA yield cannot account for all the NDMA precursors detected in natural water samples
(Gerecke and Sedlak, 2003). Other identified precursors include natural organic matter (NOM)
(Chen and Valentine, 2007; Dotson et al., 2007; Gerecke and Sedlak, 2003; Mitch and Sedlak,
2004) and tertiary amines (Le Roux et al., 2012; Mitch and Schreiber, 2008), but NDMA yields
from the chloramination of these precursors are generally low. A much higher level of NDMA
precursors has been associated with waters impacted by wastewater effluent organic matter
(EfOM) (Krasner, 2009; Schreiber and Mitch, 2006; Shah et al., 2012), suggesting
anthropogenic origins. Possible sources include constituents in personal care products (Kemper
et al., 2010), amine-based pharmaceuticals and pesticides (Le Roux et al., 2011), as well as some
117
amine-based polymers and resins used in water treatment plants (Kohut and Andrews, 2003;
Mitch and Sedlak, 2004; Najm and Trussell, 2001; Wilczak et al., 2003). Moreover,
dimethylsulfamide (a degradation product of the fungicide tolyfluanide) has been identified as an
NDMA precursor specifically associated with ozonation (Schmidt and Brauch, 2008).
In order to effectively control the NDMA level in treated drinking water, it is important to
understand the factors that may affect its formation. The pH plays a significant role in NDMA
formation upon chloramination. In the literature, the impact of pH was mainly studied using the
model precursor DMA, for which the maximum NDMA formation was observed between pH 7
and 8 (Mitch and Sedlak, 2002b). Several studies also looked into the pH impact on NDMA
formation from treated wastewater (Mitch and Sedlak, 2002a; Krasner et al., 2010), polyamine
polymers (Krasner et al., 2010), as well as the pharmaceutical ranitidine (Le Roux et al., 2011).
All of these studies focused on the impact of pH on the NDMA formation potential (FP);
however, little is known about how the reaction kinetics might be affected, and other studies
have found that NDMA could be further formed in distribution systems as water age increases
(Barrett et al., 2003; Charrois and Hrudey, 2007). In water treatment plants, pH is often boosted
to 8 to 8.5 before water enters the distribution system for the purpose of corrosion control.
Therefore, knowledge concerning the role of pH is significant in the control of NDMA formation
via chloramination.
This chapter investigates the impact of pH on the NDMA formation from two amine-based
pharmaceuticals, ranitidine and sumatriptan. They were selected because of their relatively high
NDMA-FP among the twenty pharmaceuticals and personal care products (PPCPs) tested in
Chapter 4; and ranitidine and sumatriptan each has a pKa value that respectively represents the
lower and higher end of the range of pKa values for the twenty PPCPs. The NDMA formation
kinetics and conversion yields from ranitidine and sumatriptan were studied at a pH range of 6 to
9, which is relevant to typical drinking water disinfection and distribution conditions. Since
ranitidine has the highest NDMA conversion rate among the currently known precursors, and it
is among the top 200 pharmaceuticals currently being consumed (Drugs Information Online,
2010), it was also of interest to see if there should be any concern for people who take ranitidine
pills with tap water that contains a chloramine residual. Therefore, some additional tests were
118
performed with ranitidine to determine its NDMA conversion at a pH as low as 2 to include the
extremely acidic condition that simulates the human stomach. Preformed chloramines were used
in this chapter in order to better understand the pH impacts on NDMA formation kinetics from
the selected pharmaceuticals independent of other relevant factors. It is recognized that, in
practice, it is quite common to apply free chlorine first, followed by the addition of ammonia to
form chloramine on-site, and that prechlorination has been reported to oxidize NDMA precursors
and thus reduce the NDMA formation from DMA and NOM (Bond and Templeton, 2011; Chen
and Valentine, 2008; Charrois and Hrudey, 2007; Mitch et al., 2010). However, since the impact
of pH will be more complicated when prechlorination is applied, in part because pH determines
the free chlorine speciation and thus can affect the effectiveness of prechlorination as well as
subsequent chloramination, this chapter employed only preformed chloramines.
6.2 Materials and Methods
The experiments were performed in lab-grade Milli-Q® (MQ; Millipore, Etobicoke, Ontario)
water, and the experimental conditions were modified based on the simulated distribution system
(SDS) conditions as described in Section 3.1.4. Most of the conditions were maintained the same
(i.e., 21 ± 1 °C; Cl2: NH4-N mass ratio = 4.2:1; preformed chloramine dosage = 2.5 ± 0.2 mg/L),
except that the pH was controlled by the addition of 2 mL/L of the appropriate corresponding
buffer solutions (Chapter 3, Table 3.3). Intermediate pH values were achieved by the dropwise
addition of 25 % HCl or NaOH after the buffer solution was added. The chloramine
concentrations were determined using DPD colorimetry (DR2010 HACH-Kit, Hach Canada,
Mississauga, Ontario) which measures the concentration of total chlorine, free chlorine, and
monochloramine as Cl2 (mg/L). Dichloramine concentration was calculated as “total chlorine −
monochloramine − free chlorine”.
All the NDMA formation experiments were performed in 1 L amber bottles with LDPE (low-
density polyethylene) caps. The selected pharmaceutical and target dosage of preformed
chloramine were added into the pH adjusted MQ water, and the samples were incubated at room
temperature for designated contact time, followed by the addition of excess ascorbic acid powder
to quench the chloramine residual (see details in Section 3.1.4). Further details about the NDMA
119
extraction, concentration, analysis, and quantification methods are provided in Section 3.2.1. For
the kinetic experiments, the NDMA formation was monitored up to 72 hr and 96 hr for ranitidine
and sumatriptan, respectively. At each time point, samples (i.e., MQ water dosed with individual
pharmaceuticals) were prepared in triplicate, along with one blank control (i.e., MQ water
without dosing any pharmaceuticals). For the extra NDMA-FP tests for ranitidine, samples
(dosed with ranitidine) and blanks (MQ) were both prepared in triplicate and incubated for 72 hr.
Error bars in all the graphs demonstrate the variability due to multiple formation potential tests
(n = 3) under the same reaction conditions.
6.3 Results and Discussion
6.3.1 Impact of pH on the NDMA Formation Kinetics
The impact of pH within the drinking water relevant range (i.e., pH 6 to 9) on NDMA formation
kinetics from chloramination of ranitidine and sumatriptan is summarized in Figure 6.1. Markers
in the figures represent the actually measured NDMA molar conversion and the lines represent
the model-predicted conversion. The three-parameter kinetic model used was of the form:
𝑌 = ������×(�����) (Eq.6.1)
where Y is the NDMA molar conversion at given reaction time (t); θ is the ultimate NDMA
molar conversion; Lag is the time required to achieve 50 % of the ultimate molar conversion and
thus associated with the length of the initial lag phase; k is the pseudo-first order reaction rate
constant. Verification of the kinetic model under different pH conditions is presented in
Appendix 9. Further details about the model and additional verification data have been discussed
in Chapter 5.
120
Figure 6.1. NDMA formation kinetics from ranitidine and sumatriptan upon chloramination under
different pH conditions (MQ, 25 nM of pharmaceutical, preformed chloramine = 2.5 ± 0.2 mg/L; error bars represent the variability due to multiple formation potential tests (n = 3))
121
The estimated model parameters are summarized in Table 6.1. The model fit the experimental
data very well, with correlation coefficients (R2) higher than 0.90 in most cases. Similar trends
were observed for both pharmaceuticals. In general there was a maximum NDMA conversion at
a medium pH (pH 7 for ranitidine and pH 8 for sumatriptan, Figure 6.1); but the initial lag phase
(Lag) was reduced and the reaction rate constant (k) was increased as the pH was increased,
indicating that faster reactions were occurring at higher pH values. As shown in Figure 6.2 and
Figure 6.3, the initial lag phase (Lag, estimated model parameter) for both pharmaceuticals
consistently decreased with increases in the pH and the percentage of non-protonated
pharmaceuticals (calculated based on the conventional relationship between pH and pKa);
accordingly, the reaction rate constant (k, estimated model parameter) consistently increased
with increasing pH and the percentage of non-protonated pharmaceuticals.
Table 6.1. Estimated kinetic model parameters under different pH conditions (MQ water, 25 nM of pharmaceuticals, preformed chloramine = 2.5 ± 0.2 mg/L)
Compound pH θ Lag (hr) k (hr-1) R2 Ranitidine 6.0 0.571 (0.023) a 28.5 (1.9) 0.06 (0.01) 0.992
6.5 0.636 (0.032) 12.9 (1.1) 0.14 (0.03) 0.980
7.0 0.906 (0.045) 4.6 (0.5) 0.31 (0.08) 0.991
7.5 0.778 (0.024) 4.5 (0.3) 0.31 (0.05) 0.993
8.0 0.628 (0.022) 1.7 (0.2) 0.66 (0.15) 0.983
9.0 0.504 (0.020) 0.6 (0.1) 1.63 (0.45) 0.955
Sumatriptan 6.0 0.022 (0.001) 36.6 (3.8) 0.04 (0.01) 0.976
7.0 0.028 (0.001) 27.0 (2.4) 0.06 (0.01) 0.992
8.0 0.042 (0.002) 16.0 (1.7) 0.07 (0.01) 0.974
8.5 0.030 (0.002) 7.8 (1.8) 0.11 (0.04) 0.930
9.0 0.024 (0.002) 6.1 (1.9) 0.11 (0.05) 0.889 a Numbers in the brackets represent the 95 % confidence interval of each model parameter.
122
Figure 6.2. Potential correlation between the initial lag phase (Lag), percentage of non-protonated
pharmaceuticals, and pH (MQ, 25 nM of pharmaceuticals, preformed chloramine = 2.5 ± 0.2 mg/L; error bars in the top two figures represent the 95 % confidence interval of the model parameter)
123
Figure 6.3. Potential correlation between the reaction rate constant (k), percentage of non-protonated
pharmaceuticals, and pH (MQ, 25 nM of pharmaceuticals, preformed chloramine = 2.5 ± 0.2 mg/L; error bars in the top two figures represent the 95 % confidence interval of the model parameter)
Generally, there are two major factors determining the rate and yield of NDMA formation: the
form of the precursor species (i.e., protonated amines or not, pKa-pH relevant) and the
chloramine species present (i.e., monochloramine or dichloramine). If the NDMA formation
reaction can be described as involving an electrophilic attack on the nitrogen atom of the amine
group (Mitch and Schreiber, 2008), then non-protonated amines would be favored in the
reaction. Ranitidine has a pKa of 8.2 and sumatriptan has a pKa of 9.6 (for the tertiary amine
group; calculated using ChemAxon (ChemSpider)), thus the non-protonated forms of these
amines would be present at more significant concentrations at more basic pH, which could
explain the lower NDMA conversion at acidic pH as well as the faster initiation of the reaction
observed at the higher pH values tested. In addition to the direct effect of pH on the tertiary
124
amine groups of these molecules, indirect effects of proton interactions at other sites in the
molecules were also considered. However, the pKa values of other groups in these molecules are
too extreme (pKa of 3.0 and 11.2 for the primary amine group on ranitidine and sumatriptan,
respectively, calculated using ChemAxon (ChemSpider)) relative to the approximately neutral
conditions observed in most source waters, thus proton interactions at these sites is not relevant
in natural waters. At the same time, dichloramine also significantly enhances NDMA formation
compared with monochloramine (Mitch et al., 2005; Schreiber and Mitch, 2005, 2006). As
expected, monochloramine was observed to be dominant at neutral and basic pH, while more
dichloramine was formed at acidic pH (Figure 6.4). Therefore, although the non-protonated
amine is dominant at higher pH, the ultimate NDMA conversion is expected to be less due to the
lack of dichloramine.
Figure 6.4. Monochloramine decay and dichloramine increase under different pH (MQ, preformed
chloramine = 2.5 ± 0.2 mg/L)
In addition, the oxidation of these amine-based pharmaceuticals may involve an acid release step
that would inhibit NDMA formation at lower pH. For example, one might consider that the first
step of the NDMA formation reaction from DMA, which is also the rate-limiting step, involves a
release of H+ (Eq. 6.2 and Eq. 6.3; Mitch and Sedlak, 2002b; Schreiber and Mitch, 2006).
Although these two equations are for when DMA is the precursor, it has been suggested that
NDMA formation from tertiary amines may proceed via a dealkylation reaction with
chloramines to first form DMA and then subsequently oxidize DMA to NDMA; and that the
dealkylation reaction itself also involves a release of H+ (Mitch and Sedlak, 2004). If a similar
125
acid release step is applicable to the reactions in the current research, then the reaction between
amine-based pharmaceuticals and chloramines would be expected to be inhibited at lower pH,
which could lead to the slower NDMA reaction rates (k) and much longer initial lag phases (Lag)
observed at acidic pH, as observed in this study (Table 6.1).
NH(CH3)2 + NH2Cl → N(CH3)2-NH2 + H+ + Cl- (Eq. 6.2)
NH(CH3)2 + NHCl2 → N(CH3)2-NHCl + H+ + Cl- (Eq. 6.3)
Figure 6.5 is a comprehensive presentation of the pH effect, summarizing the relationships in
between the pharmaceutical species, chloramine species and pH, as well as how they affect the
NDMA formation from selected pharmaceuticals. The maximum NDMA molar conversion
occurred at pH 7 for ranitidine and pH 8 for sumatriptan, which are approximately 1.2 and 1.6
units lower than their respective pKa values. The percentage of non-protonated amines increases
with increasing pH (as expected, based on the mathematical relationship between pH and pKa),
corresponding with shorter initial lag phase at higher pH values; however, the lower NDMA
conversion at basic pH was associated with a lack of dichloramine. Although more dichloramine
was available at acidic pH, the NDMA conversion was limited due to the low concentration of
non-protonated amine groups.
126
Figure 6.5. The ultimate NDMA molar conversion, initial lag phase, dichloramine percentage, and pharmaceutical species under different pH conditions (MQ, 25 nM of pharmaceuticals, preformed
chloramine = 2.5 ± 0.2 mg/L; error bars for the NDMA molar conversion represent the variability due to multiple formation potential tests (n = 3); error bars for the initial lag phase represent the 95 % confidence
interval for the estimated model parameter)
From these competing issues, the expected effects of pH on NDMA yields may be summarized
as shown in Figure 6.6. At lower pH, there is more dichloramine but less non-protonated amines;
at higher pH, there is more non-protonated amine but less dichloramine. Therefore, there will be
an optimum pH in the middle of the range that maximizes the NDMA conversion, which is what
has been observed for both ranitidine and sumatriptan.
127
Figure 6.6. Generalized impact of precursor species and chloramine species on the NDMA formation
In summary, the maximum NDMA formation usually occurs at a pH which is approximately 1.2
to 1.6 units lower than the pKa of the precursor amine group. This is only drawn based on results
from two pharmaceuticals with pKa from 8 to 10, and needs to be further verified with other
precursors. However the optimum pH for NDMA formation might not exceed 8, because
dichloramine decays rapidly at pH above 8 (Jafvert and Valentine, 1992). So far, the maximum
rate of NDMA formation reported in literature was in general within the range of pH 7 to 8 (Le
Roux et al., 2011; Mitch and Sedlak, 2002a, b). At lower pH, the NDMA conversion from
ranitidine and sumatriptan is inhibited because of the protonated amine groups as well as the H+-
limited reaction. At higher pH, the ultimate conversion is limited by the lack of dichloramine, but
the NDMA formation is faster because of more non-protonated amines. The enhanced reaction
rate with increasing pH has never been reported, which may also play a significant role in
controlling NDMA formation. For example, utilities may add silicate to boost the water pH to
8.0 to 8.5 before it enters the distribution system for the purpose of corrosion control (Health
Canada, 2009; Thompson et al., 1997). From the perspective of NDMA formation control, the
basic pH might help reduce the maximum NDMA concentration in the farther ends of the
distribution system; however, there might be more areas experiencing a medium NDMA
concentration given the much faster reaction at higher pH values. Therefore, depending on the
size of the distribution system and the location of households (i.e., the associated water age), the
128
optimum pH should consider the corrosion control, the DBP formation, and most importantly the
disinfection performance itself.
6.3.2 Impact of Low pH on the NDMA-FP from Ranitidine
Ranitidine has a very high NDMA conversion rate compared with most other known precursors.
Since it is mainly used as an antacid for the treatment of stomach ulcers, it was of interest to
know if NDMA can be formed in human stomach by taking the ranitidine pills with
chloraminated drinking water. Additional NDMA-FP tests were performed with ranitidine over a
wide range of pH values in MQ water. The results are shown in Figure 6.7. In an extremely
acidic environment (pH ≤ 3, simulating the human stomach), no NDMA was detected after 72 hr
of incubation. Thus it is unlikely that a significant amount of NDMA would be formed from
ranitidine in human stomach, at least not via the chloramination mechanism. Much higher
NDMA conversions occurred in the pH range of 6 to 9 (the maximum conversion observed at pH
7), which are more relevant to the conditions for drinking water treatment and distribution. These
results are generally in agreement with the results described above and with the findings by Le
Roux et al. (2011), where the maximum NDMA formation from ranitidine was observed at pH
7.9 (3 µM ranitidine and 2.5 mM NH2Cl, 5 d contact time, pH 4 to 10).
Figure 6.7. NDMA molar conversion from ranitidine (25 nM) upon chloramination under different pH
conditions (MQ water, preformed chloramine = 2.5 ± 0.2 mg/L, incubation time = 72 hr; error bars represent the variability due to multiple formation potential tests (n = 3))
129
6.4 Summary
The pH affects both the ultimate NDMA conversion from selected pharmaceuticals as well as the
reaction rate. For the amine-based pharmaceuticals tested, the maximum NDMA formation
occurs at pH approximately 1.2 to 1.6 units lower than the pKa of the precursor amine group, and
most likely within the pH range of 7 to 8. While the ultimate NDMA conversion is limited by the
level of dichloramine, the reaction rate is mainly determined by the amount of non-protonated
amines. Results from this chapter were obtained from experiments using preformed chloramines
in lab-grade water. Further study is needed to investigate the impact of pH on NDMA formation
from amine-based pharmaceuticals in different water matrices and upon different disinfection
strategies.
Although the impact of pH has been studied before for other NDMA precursors, very limited
information is available in terms of the impact on the formation kinetics which is essential in the
prediction of NDMA formation. Elevating pH in the distribution system may help to reduce the
ultimate NDMA formation at the farther ends of the distribution system, but it may also enhance
the initial reaction in sections with shorter water age, thus causing more areas to experience a
moderate NDMA concentration. The optimum pH should consider the size of the distribution
system, the location of households and their associated water age, and most importantly the
disinfection performance itself.
130
6.5 References
Barrett, S., Hwang, C., Guo, Y., Andrews, S.A., Valentine, R., 2003. Occurrence of NDMA in
drinking water: a North American survey, 2001-2002. In: Proceedings of the American Water
Works Association's Annual Conference, Anaheim, CA, June 15-19, 2003
Blute, N., Russell, C., Chowdhury, Z., Wu, X., Via, S., 2010. Nitrosamine occurrence in the U.S.
– analysis and interpretation of UCMR2 data. In: Proceedings of the AWWA Water Quality
Technology Conference, Savannah, GA, November 14-18, 2010
Bond, T., Templeton, M.R., 2011. Nitrosamine formation from the oxidation of secondary
amines. Wa. Sci. Technol. 11 (3), 259-265
Charrois, J.W.A., Boyd, J.M., Froese, K.L., Hrudey, S.E., 2007. Occurrence of N-nitrosamines in
Alberta public drinking-water distribution systems. J. Environ. Eng. Sci. 6, 103-114
Charrois, J.W.A., Hrudey, S.E., 2007. Breakpoint chlorination and free-chlorine contact time:
implications for drinking water N-nitrosodimethylamine concentrations. Water Res. 41, 674-682
Chemspider, the free chemical database. Retrieved from http://www.chemspider.com/
(compounds calculated on July 10th, 2012)
Chen, Z., Valentine, R.L., 2007. Formation of N-Nitrosodimethylamine (NDMA) from humic
substances in natural water. Environ. Sci. Technol. 41, 6059-6065
Chen, Z., Valentine, R.L., 2008. The influence of the pre-oxidation of natural organic matter on
the formation of N-Nitrosodimethylamine (NDMA). Environ. Sci. Technol. 42, 5062-5067
Dotson, A., Westerhoff, P., Krasner, S.W., 2007. Nitrosamine formation from natural organic
matter isolates and sunlight photolysis of nitrosamines. In: Proceedings of AWWA Annual
Conference and Exposition. Toronto, ON, Canada, June 24-28, 2007
131
Drugs Information Online. Top 200 pharmaceutical drugs by retail sales in 2010. Retrieved from
http://www.drugs.com/top200.html (accessed July 6, 2012)
EPA Integrated Risk Information System (IRIS), 1993. N-Nitrosodimethylamine; CASRN 62-
75-9. www.epa.gov/iris/subst/0045.htm
Gerecke, A.C., Sedlak, D.L., 2003. Precursors of N- Nitrosodimethylamine in natural waters.
Environ. Sci. Technol. 37, 1331-1336
Health Canada, 2009. Guidance on Controlling Corrosion in Drinking Water Distribution
Systems. Section B.4.3.2. http://www.hc-sc.gc.ca/ewh-semt/pubs/water-eau/corrosion/index-
eng.php#a1
Health Canada, 2010. Guideline Technical Document on N-nitrosodimethylamine (NDMA) in
drinking water for public comment. http://www.hc-sc.gc.ca/ewh-semt/consult/_2010/ndma/draft-
ebauche-eng.php#a3
Jafvert, C.T., Valentine, R.L., 1992. Reaction scheme for the chlorination of ammoniacal water.
Environ. Sci. Technol. 26, 577 - 586
Kemper, J.M., Walse, S.S., Mitch, W.A., 2010. Quaternary amines as nitrosamine precursors: a
role for consumer products? Environ. Sci. Technol. 44, 1224-1231
Kohut, K.D., Andrews, S.A., 2003. Polyelectrolyte age and N-nitrosodimethylamine formation
in drinking water treatment. Water Qual. Res. J. Can. 38 (4), 719-735
Krasner, S.W., 2009. The formation and control of emerging disinfection by-products of health
concern. Phil Trans. R Soc. A 367, 4077-4095
Krasner, S.W., Dale, M.S., Lee, C.F.T., Garcia, E.A., Wong, T.M., Mitch, W., Von Gunten, U.,
2010. Difference in reactivity and chemistry of NDMA precursors from treated wastewater and
132
from polyamine polymers. In: Proceedings of the AWWA Water Quality Technology
Conference, Savannah, GA, November 14-18, 2010
Le Roux, J., Gallard, H., Croue, J., 2011. Chloramination of nitrogenous contaminants
(pharmaceuticals and pesticides): NDMA and halogenated DBPs formation. Water Res. 45,
3164-3174
Le Roux, J., Gallard, H., Croue, J., 2012. Formation of NDMA and halogenated DBPs by
chloramination of tertiary amines: The influence of bromide ion. Environ. Sci. Technol. 46,
1581-1589
Massachusetts Department of Environmental Protection (MassDEP), 2004. Current Regulatory
Limit: n-Nitrosodimethylamine (NDMA), CASRN 62759.
http://www.mass.gov/dep/water/drinking/standards/ndma.htm
Mitch, W.A., Krasner, S.W., Lee, C.F.T., Wong, T.M., 2010. Tradeoffs in DBP formation during
chloramination following pre-oxidation for nitrosamine control. In: Proceedings of the AWWA
Water Quality Technology Conference, Savannah, GA, November 14-18, 2010
Mitch, W.A., Oelker, G.L., Hawley, E.L., Deeb, R.A., Sedlak, D.L., 2005. Minimization of
NDMA formation during chlorine disinfection of municipal wastewater by application of pre-
formed chloramines. Environ. Eng. Sci. 22 (6), 882-890.
Mitch, W.A., Schreiber, I.M., 2008. Degradation of tertiary alkylamines during
chlorination/chloramination: implications for formation of aldehydes, nitriles, halonitroalkanes,
and nitrosamines. Environ. Sci. Technol. 42, 4811-4817
Mitch, W.A., Sedlak, D.L., 2002a. Factors controlling nitrosamine formation during wastewater
chlorination. Wa. Sci. Technol. 2 (3), 191-198
Mitch, W.A., Sedlak, D.L., 2002b. Formation of N- Nitrosodimethylamine (NDMA) from
dimethylamine during chlorination. Environ. Sci. Technol. 36, 588-595
133
Mitch, W.A., Sedlak, D.L., 2004. Characterization and fate of N-nitrosodimethylamine
precursors in municipal wastewater treatment plants. Environ. Sci. Technol. 38, 1445-1454
Mitch, W.A., Sharp, J.O., Trussell, R.R., Valentine, R.L., Alvarez-Cohen, L., Sedlak, D.L., 2003.
N-nitrosodimethylamine (NDMA) as a drinking water contaminant: a review. Environ. Eng. Sic.
20, 389-404
MOE, 2003. Ontario regulation 268/03 made under the safe drinking water act, 2002.
http://www.e-laws.gov.on.ca/html/source/regs/english/2003/elaws_src_regs_r03268_e.htm
Najm, I., Trussell, R.R., 2001. NDMA formation in water and wastewater. J. Am. Water Works
Assoc. 93 (2), 92-99
NHMRC, NRMMC, 2011. Australian Drinking Water Guidelines Paper 6 National Water
Quality Management Strategy. National Health and Medical Research Council, National
Resource Management Ministerial Council, Commonwealth of Australia, Canberra.
Office of Environmental Health Hazard Assessment (OEHHA), 2006. Public Health Goal for N-
nitrosodimethylamine and cadmium in drinking water.
http://www.oehha.org/water/phg/cadndma122206.html
Russell, C.G., Blute, N.K., Via, S., Wu, X., Chowdhury, Z., 2012. Nationwide assessment of
nitrosamine occurrence and trends. J. Am. Water Works Assoc. 104, 205-217
Schmidt, C.K., Brauch, H.J., 2008. N,N-Dimethylsulfamide as precursor for N-
Nitrosodimethylamine (NDMA) formation upon ozonation and its fate during drinking water
treatment. Environ. Sci. Technol. 42, 6340-6346
Schreiber, I.M., Mitch, W.A., 2005. Influence of the order of reagent addition on NDMA
formation during chloramination. Environ. Sci. Technol. 39, 3811-3818
134
Schreiber, I.M., Mitch, W.A., 2006. Occurrence and fate of nitrosamines and nitrosamine
precursor in wastewater-impacted surface waters using boron as a conservative tracer. Environ.
Sci. Technol. 40, 3203-3210
Shah, A.D., Krasner, S.W., Lee, C.F.T., von Gunten, U., Mitch, W.A., 2012. Trade-offs in
disinfection byproduct formation associated with precursor preoxidation for control of N-
Nitrosodimethylamine formation. Environ. Sci. Technol. 46, 4809-4818
Sigma-Aldrich buffer reference centre. Retrieved from http://www.sigmaaldrich.com/life-
science/core-bioreagents/biological-buffers/learning-center/buffer-reference-center.html
(accessed on July 10th, 2012)
Thompson, J.L., Scheetz, B.E., Schock, M.R., Lytle, D.A., Delaney, P.J., 1997. Sodium silicate
corrosion inhibitors: issues of effectiveness and mechanism. In: Proceedings of the AWWA
Water Quality Technology Conference, Denver, CO, November 9-12, 1997
USEPA, 2009. Contaminant Candidate List 3 (CCL3).
http://water.epa.gov/scitech/drinkingwater/dws/ccl/ccl3.cfm
Wilczak, A., Assadi-Rad, A., Lai, H.H., Hoover, L.L., Smith, J.F., Berger, R., Rodigari, F.,
Beland, J.W., Lazzelle, L.J., Kincannon, E.G., Baker, H., Heaney, C.T., 2003. Formation of
NDMA in chloraminated water coagulated with DADMAC cationic polymer. J. Am. Water
Works Assoc. 95 (9), 94-106
135
Chapter 7 NDMA Formation From Amine-Based Pharmaceuticals: Impact From
Prechlorination and Water Matrix
This chapter has been submitted to Water Research as:
Shen, R., Andrews, S.A., 2012b. NDMA formation from amine-based pharmaceuticals –
impact from prechlorination and water matrix. Water Research (under review)
Additional data that are presented in Appendices include the preliminary chlorine/chloramine
demand tests to determine the initial NaClO and NH4Cl dosage for the sequential disinfection
experiments (Appendix 10), the relevant statistical analyses of the experimental data (Appendix
11) and associated estimated kinetic model parameters (Appendix 13), the experiments to
evaluate the potential impact of Cl2:TOC ratio (Appendix 12), some preliminary LC-MS results
(Appendix 14), and some additional LC-OCD data (Appendix 15).
136
Abstract
The presence of N-nitrosodimethylamine (NDMA) in drinking water is most commonly
associated with the chloramination of amine-based precursors. One option to control the NDMA
formation is to remove the precursors via pre-oxidation, and prechlorination is among the most
effective options in reducing NDMA formation. However, most of the findings to-date are based
on single-precursor scenarios using the model precursor dimethylamine (DMA) and natural
organic matter (NOM), while few studies have considered the potential interactions between
water matrix components and the target precursors when investigating the prechlorination
impact. Specifically, little is known for the behavior of amine-based pharmaceuticals which have
recently been reported to contribute to NDMA formation upon chloramination. This chapter
demonstrates that prechlorination can affect both the ultimate NDMA conversion and the
reaction kinetics from selected pharmaceuticals, and the nature and extent of the impact was
compound-specific and matrix-specific. In the absence of NOM, the NDMA formation from
most pharmaceuticals was reduced upon prechlorination, except for sumatriptan which showed a
consistent increase in NDMA formation with increasing free chlorine contact time. In the
presence of NOM, prechlorination was shown to enhance initial reactions by reducing the
binding between NOM and pharmaceuticals, but prolonged prechlorination broke down NOM
into smaller products which could then form new bonds with pharmaceuticals and thus inhibit
their further conversion into NDMA.
Keywords
NDMA Prechlorination Ranitidine Sumatriptan NOM
137
7.1 Introduction
The presence of N-nitrosodimethylamine (NDMA) in drinking water has been commonly
associated with the disinfection process, especially chloramination. NDMA is a highly
mutagenic compound and a potential human carcinogen, with a 10-6 lifetime cancer risk
associated with a drinking water concentration of 0.7 ng/L (EPA IRIS, 1993). The exposure to
NDMA through drinking water has become a concern especially for utilities that apply
chloramine as the secondary disinfectant. Health Canada has recently proposed a maximum
acceptable concentration for NDMA of 40 ng/L (0.54 nM) in drinking water (Health Canada,
2010). USEPA also placed it on the drinking water contaminant candidate list 3 (CCL3) together
with four other nitrosamines (USEPA, 2009). NDMA is currently regulated in drinking water in
several provinces and states across North America, including Ontario (9 ng/L (0.12 nM); MOE,
2003), Massachusetts (10 ng/L(0.13 nM); MassDEP, 2004), and California (10 ng/L (0.13 nM);
OEHHA, 2006). NDMA is also included in the Australian Drinking Water Guideline at 100 ng/L
(1.35 nM) since 2011 (NHMRC, NRMMC, 2011).
During drinking water treatment processes, NDMA is most commonly formed via the slow
reaction between chloramines (especially dichloramine) and amine-based precursors (Schreiber
and Mitch, 2006a). NDMA is also formed through a nitrosation mechanism at acidic pH (Choi
and Valentine, 2003), but this mechanism is of less importance in drinking water due to the
generally low nitrite concentrations and the neutral/basic pH. Moreover, ozonation can also lead
to the formation of NDMA (Andrzejewski et al., 2008; Oya et al., 2008); an especially high yield
was observed for dimethylsulfamide, a degradation product of the fungicide tolyfluanide
(Schmidt and Brauch, 2008).
Early studies on NDMA typically used the model precursor dimethylamine (DMA; Gerecke and
Sedlak, 2003) and natural organic matter (NOM; Chen and Valentine, 2007; Dotson et al., 2007;
Gerecke and Sedlak, 2003; Mitch and Sedlak, 2004), however chloramination of these precursors
typically gave low yields of NDMA and thus cannot always account for all of the precursors
present in natural waters. More recently, higher levels of NDMA formation have been associated
with wastewater-impacted surface water (Krasner, 2009; Schreiber and Mitch, 2006b; Shah et
al., 2012), indicating the contribution from anthropogenic compounds. Several studies have
138
linked NDMA formation to quaternary amines used in personal care products (Kemper et al.,
2010), pharmaceuticals and pesticides (Le Roux et al., 2011, 2012a, b), as well as some amine-
based polymers and resins (Kohut and Andrews, 2003; Mitch and Sedlak, 2004; Najm and
Trussell, 2001; Wilczak et al., 2003). Mitch and Schreiber (2008) have proposed that the NDMA
formation from tertiary amines proceeds via a chlorine transfer reaction to release the DMA
which is then subsequently oxidized to NDMA. However, NDMA formed from DMA alone
cannot explain the high yields from some tertiary amines such as ranitidine. A recent study by Le
Roux et al. (2012b) has proposed an alternative mechanism for NDMA formation from
chloramination of ranitidine, which involves a direct substitution on the DMA group of ranitidine
and can well explain its high yield of NDMA.
In practice, chloramine is usually used as the secondary disinfectant following primary
disinfection (e.g., Cl2, UV, and O3). Moreover, instead of using preformed monochloramine,
most utilities that perform chloramination typically apply free chlorine first, followed by the
addition of ammonia to form chloramine on site. Compared with chloramine alone, the
application of pre-oxidation may modify or destroy the precursors and release transformation
products that may or may not react with the subsequent secondary disinfectant.
Generally, the application of preoxidation prior to chloramination has been reported to reduce the
NDMA formation from DMA and NOM. Preoxidation processes have included prechlorination
(Chen and Valentine, 2008; Charrois and Hrudey, 2007; Mitch et al., 2010), O3 and O3/H2O2
(Chen and Valentine, 2008; Pisarenko et al., 2012), ClO2 (Lee et al., 2007), ferrate (Lee et al.,
2008), and KMnO4 (Chen and Valentine, 2008). Among all the options, ozone and chlorine were
found to be most effective in reducing NDMA formation (Shah et al., 2012). However, oxidation
of precursors does not necessarily lead to the reduction in NDMA formation. In some cases,
prechlorination may increase NDMA formation at lower exposures due to insufficient oxidation
(Chen and Valentine, 2008; Shah et al., 2012). It has also been observed that UV and UV/ H2O2
pretreatment have increased the NDMA formation from amine-based polymers (e.g.,
polyDADMAC, epi-DMA), possibly due to the increased degradation of the polymers releasing
more NDMA precursors (Harvey, 2009). More recently, Radjenovic et al. (2012) have reported
that the oxidation products of the pharmaceutical tramadol by UV and UV/ H2O2 have higher
139
NDMA formation potentials (FPs) than the parent compound. Thus, the impact of preoxidation
on the formation of NDMA requires further investigation.
Currently, most of the findings regarding the preoxidation impact have been based on DMA and
NOM. Several studies used treated wastewater as a “precursor pool”, but few specific
compounds have been studied separately, especially pharmaceutical-based precursors. Moreover,
very little information is available in terms of how the preoxidation process might affect the
NDMA formation kinetics. This chapter investigates the impact of prechlorination on the NDMA
formation from eight selected pharmaceuticals, especially the impact on their reaction kinetics.
The eight pharmaceuticals were selected because of their relatively high NDMA-FPs (molar
conversions > 1 %) among the twenty pharmaceuticals and personal care products (PPCPs)
tested in Chapter 4. This chapter also compares the prechlorination impact with and without the
presence of NOM, and looks into how the interactions in between NOM, pharmaceuticals, and
free chlorine could affect the NDMA formation from selected pharmaceuticals. Relatively high
pharmaceutical concentrations (6.8 – 11.1 µg/L) compared with their expected environmental
levels were applied in order to be able to measure the differences in NDMA formation under
different disinfection conditions. However, it has been demonstrated in Chapter 5 that the
NDMA formation kinetics in a real water matrix was relatively independent of the initial
pharmaceutical concentration because chloramine was present in large excess relative to the
pharmaceutical concentrations; therefore similar reaction kinetics are expected for selected
pharmaceuticals at their environmental levels. Findings from this chapter could be of particular
concern for water reuse processes where much higher concentrations of pharmaceuticals might
be subjected to chloramination.
7.2 Materials and Methods
The pharmaceuticals (25 nM of each) were dosed into selected water matrices (raw, not filtered)
individually and subjected to different disinfection strategies. The chloramination experiments
(preformed chloramine) were carried out under the simulated distribution system (SDS)
conditions as described in Chapter 3 for their respective water matrices (see details in Section
3.3). The sequential disinfection experiments (prechlorination followed by chloramination)
140
employed modified SDS conditions, where a sodium hypochlorite (NaClO) solution was first
added, followed by the addition of ammonia chloride (NH4Cl) to form chloramine after a range
of target free chlorine contact times (0.5 to 120 min). The resulting chloramine concentration at
the point of NH4Cl addition was the same as that which was applied in the preformed
chloramination experiments for each matrix (i.e., 2.5 ± 0.2 mg/L plus the 24hr chloramine
demand for each matrix). The initial free chlorine dosage and the NH4Cl dosage for each matrix
were determined via preliminary chlorine/chloramine demand tests (Appendix 10). All of the
other conditions remained the same as the SDS conditions (i.e., 21 ± 1 °C; pH = 7.0 ± 0.1; Cl2: N
mass ratio = 4.2:1). Further details concerning the preparation of stock solutions, working
solutions, and the NDMA analysis have been described in Chapter 3 (see details in Sections 3.1.2
and 3.2.1).
NDMA-FPs and kinetics evaluations were both conducted under the (modified) SDS conditions
described above. The 24 hr NDMA-FP was determined for all eight pharmaceuticals in lab-grade
MQ (Milli-Q®, MilliPore, Etobicoke, Ontario) water, where samples (dosed with
pharmaceuticals) and blanks (water matrix only) were both prepared in triplicate. The NDMA
formation kinetics was monitored for ranitidine and sumatriptan, where samples were prepared
in triplicate along with one blank control (using the respective water matrix) at each time point.
Error bars in all graphs in this chapter demonstrate the variability due to multiple formation
potential tests (n = 3 unless otherwise specified) under the same reaction conditions.
The kinetic experiments were performed in three water matrices; the basic water quality
parameters for each source are summarized in Table 7.1. Lake Ontario water and Otonabee River
water samples were taken from the influent of two drinking water treatment plants in September
2011 and January 2012, respectively. The methods for measuring these basic water quality
parameters have been described in Section 3.2.3. Use of raw water instead of partially-treated
water as the experimental matrix provided the study with the maximum amount of unmodified
NOM for the current investigations. Future studies could employ partially-treated water to
account for the removal or modification of some NOM from various treatment processes or
process trains.
141
Table 7.1. Water matrix source and basic water quality measurements (n = 5-10)
Water matrix Water source TOC
(mg/L) pH Alkalinity (mg/L)
UV254 (cm-1)
SUVA (L/mg⋅m))
Milli-Q®
(MQ) Millipore; Etobicoke,
Ontario 0.0 7.5 ± 0.1 1.8 ± 0.3 0.000 0.0
Lake Lake Ontario Ajax, Ontario 2.3 ± 0.1 8.2 ± 0.1 95.4 ± 1.8 0.023 ± 0.002 1.0 ± 0.1
River Otonabee River, Peterborough, Ontario 6.2 ± 0.1 7.8 ± 0.1 94.9 ± 2.9 0.162 ± 0.002 2.6 ± 0.1
In order to better characterize the water matrix, NOM components in each raw water source were
analyzed using size-exclusion liquid chromatography - organic carbon detection (LC-OCD) at
the University of Waterloo (ON, Canada). The LC-OCD results for the selected water matrices
are summarized in Table 7.2, and details concerning the LC-OCD analysis are presented in
Section 3.2.3. In addition, several chlorine- and/or chloramine-treated water samples (lake and
river; from the blank controls) were collected along with the kinetic experiments to investigate
the potential change of NOM upon the disinfection treatment. The chlorine/chloramine residual
was quenched with sodium thiosulfate (Na2S2O3) at the dosage of twice the stoichiometric ratio
of chlorine (personal communication, University of Waterloo); the samples were then filtered
through a 0.45 µm filter paper and shipped to University of Waterloo for LC-OCD analysis.
Samples were stored at 4°C for no more than one week prior to the analysis.
Table 7.2. LC-OCD results for the selected water matrices (Unit: mg/L carbon)
Water matrix DOC Hydrophobic
DOC Hydrophilic
DOC
Hydrophilic DOC fractions
Biopolymers Humics Building blocks
LMW neutrals
LMW acids
MQ 0.07 0.03 0.04 0.01 0.000 0.003 0.02 0.004 Lake 2.20 0.20 2.00 0.34 0.90 0.46 0.19 0.10 River 5.90 0.20 5.70 0.34 3.82 0.72 0.62 0.20
142
7.3 Results and Discussion
7.3.1 Prechlorination Impacts in MQ Water
7.3.1.1 24 hr NDMA-FP upon Prechlorination
To begin investigating the impacts of prechlorination, 24 hr NDMA-FPs following sequential
chlorine and chloramine disinfection were determined for eight selected pharmaceuticals in MQ
water (Figure 7.1). The 0 min pre-Cl condition represents the use of preformed chloramine. In
general, the NDMA formation from most of the pharmaceuticals (except for sumatriptan and
diltiazem) decreased with increasing prechlorination contact time (p-value < 0.0001 except for
doxylamine (p-value = 0.0034); ANOVA analysis, 95 % confidence level). (Detailed
calculations for the statistical analyses are summarized in Appendix 11.) However, the rate and
degree of reduction in NDMA formation varied among the different compounds. For example,
ranitidine and nizatidine responded very quickly to chlorine, with a 50 % reduction in NDMA
formation achieved by 0.5 min and 3 min of prechlorination, respectively. For tetracycline, a
similar amount of reduction required 30 min of prechlorination. For the three structurally similar
H1-antihistamines (carbinoxamine, chlorphenamine, and doxylamine), no significant reduction
was observed until the chlorine contact time increased to 60 - 120 min. Sumatriptan was the
only pharmaceutical showing consistently increasing NDMA formation as the prechlorination
contact time increased (p-value < 0.0001). For diltiazem, a very short chlorine contact time (30 s)
seemed to reduce the NDMA formation, but further increases in chlorine contact time increased
the NDMA formation to levels approaching those observed when preformed monochloramine
was used.
143
Figure 7.1. 24 hr NDMA-FP from eight pharmaceuticals upon sequential chlorination and chloramination
disinfection (MQ water; error bars represent the variability due to multiple tests (n = 3))
These initial tests showed that not all of the NDMA formation potential of the target compounds
was destroyed by prechlorination. Further testing would focus on ranitidine and sumatriptan as
representatives of pharmaceuticals that showed either decreases or increases in NDMA
formation, respectively, following prechlorination.
144
7.3.1.2 NDMA Formation Kinetics upon Prechlorination
In order to further investigate the impact of prechlorination on NDMA reaction kinetics, the
NDMA formation was monitored over time for ranitidine and sumatriptan following sequential
chlorination and chloramination (Figure 7.2). Markers in the figures represent the measured
NDMA molar conversion values and the lines represent model-predictions. The three-parameter
kinetic model used was of the form:
𝑌 = 𝜃
1 + 10�×(�����)
where Y is the NDMA molar conversion at given reaction time (t); θ is the ultimate NDMA
molar conversion; Lag is the time required to achieve 50 % of the ultimate molar conversion and
thus associated with the length of the initial lag phase; k is the pseudo-first order reaction rate
constant. Further details about the model have been discussed in Chapter 5. The 0 min
prechlorination time represents the use of preformed chloramine.
Figure 7.2. NDMA formation kinetics from ranitidine and sumatriptan upon sequential chlorine and
chloramine disinfection in MQ water (Error bars represent the maximum and minimum values under the same reaction conditions (n = 2))
The formation kinetics results were in good agreement with the 24 hr formation potential results
(Figure 7.1). NDMA formation from ranitidine was significantly reduced upon prechlorination
even with a short free chlorine contact time at 3 min, and no further reduction was achieved as
the free chlorine contact time increased. In contrast, the ultimate NDMA conversion from
sumatriptan increased with free chlorine contact time, and the initial lag phase was reduced upon
prechlorination. The estimated model parameters for both compounds are shown in Table 7.3.
145
Table 7.3. Estimated kinetic model parameters for the NDMA formation from ranitidine and sumatriptan upon sequential chlorine and chloramine disinfection
Compound Matrix Pre-Cl contact time (min) θ Lag (hr) k (hr-1) R2
Ranitidine MQ 0 a 0.906 (0.045) b 4.6 (0.5) 0.31 (0.08) 0.991
3 0.492 (0.048) 4.7 (1.4) 0.19 (0.10) 0.943
30 0.399 (0.037) 2.9 (0.7) 0.38 (0.20) 0.953
60 0.460 (0.041) 2.9 (0.6) 0.39 (0.20) 0.956
120 0.479 (0.057) 5.4 (1.5) 0.16 (0.08) 0.949
Lake 0 0.824 (0.025) 12.5 (0.7) 0.18 (0.03) 0.993
3 0.468 (0.029) 4.5 (0.8) 0.22 (0.08) 0.963
30 0.529 (0.023) 3.6 (0.4) 0.33 (0.08) 0.976
60 0.460 (0.024) 2.9 (0.3) 0.49 (0.12) 0.975
120 0.474 (0.018) 2.9 (0.2) 0.48 (0.09) 0.986
River 0 0.863 (0.018) 31.0 (1.1) 0.07 (0.01) 0.998
3 0.662 (0.024) 4.1 (0.4) 0.34 (0.08) 0.985
10 0.125 (0.005) 28.6 (1.9) 0.06 (0.01) 0.991
30 0.028 (0.003) 19.3 (2.3) 0.07 (0.02) 0.957
120 0.034 (0.125) 56.8 (85.0) 0.03 (0.02) 0.939
Sumatriptan MQ 0 0.028 (0.001) 27.0 (2.4) 0.06 (0.01) 0.992
3 0.032 (0.001) 10.4 (1.0) 0.11 (0.02) 0.991
30 0.052 (0.006) 17.5 (3.9) 0.05 (0.02) 0.947
60 0.060 (0.005) 12.4 (2.2) 0.08 (0.02) 0.971
120 0.058 (0.005) 9.9 (2.3) 0.10 (0.04) 0.954
Lake 0 0.022 (0.001) 31.8 (2.3) 0.06 (0.01) 0.991
3 0.013 (0.001) 15.3 (1.4) 0.10 (0.02) 0.976
30 0.043 (0.004) 9.9 (2.1) 0.09 (0.03) 0.935
60 0.029 (0.002) 13.1 (2.1) 0.09 (0.03) 0.947
120 0.022 (0.001) 13.3 (1.8) 0.08 (0.02) 0.962
River 0 0.024 (0.002) 66.8 (4.0) 0.03 (0.01) 0.988
3 0.023 (0.001) 24.9 (1.7) 0.06 (0.01) 0.989
30 0.037 (0.003) 33.2 (4.9) 0.03 (0.01) 0.969
60 0.032 (0.002) 37.4 (4.7) 0.03 (0.01) 0.974
120 0.015 (0.001) 33.9 (4.9) 0.03 (0.01) 0.975
a 0 minute of prechlorination represents the use of preformed chloramine. b Numbers in the brackets represent the 95 % confidence interval of each model parameter.
146
7.3.2 Prechlorination Impacts in Real Water Matrices
In Chapter 5, water matrix components were observed to slow down the initial NDMA formation
from pharmaceuticals (i.e., much longer initial lag phase with the presence of NOM), while they
had less impact on the ultimate NDMA molar conversion. In this chapter, the impact of
prechlorination on the NDMA formation kinetics was also investigated in natural water samples
dosed with selected pharmaceuticals. Some studies have reported the possible impact of bromide
ion on NDMA formation (Chen et al., 2010; Le Roux et al., 2012a; Valentine et al., 2005);
however, the bromide levels in the waters selected for this work (Lake Ontario, ~ 40 µg/L
(Comerton et al., 2006); Otonabee River, < 11.4 µg/L (Woodbeck, 2007)) are much lower
compared with the concentrations reported to be significant in literature (mg/L). Therefore the
role of bromide ion in the NDMA formation from selected pharmaceuticals is not investigated in
this study. The results for ranitidine and sumatriptan in water from Lake Ontario and the
Otonabee River are summarized in Figure 7.3 and Figure 7.4, respectively. Similar to Figure 7.2,
markers represent the measured NDMA molar conversion values and the lines represent model-
predicted conversions. The estimated model parameters in different water matrices are included
in Table 7.3 for easier comparison with similar data determined using MQ water.
Figure 7.3. NDMA formation kinetics from ranitidine upon sequential chlorine and chloramine disinfection in Lake Ontario and Otonabee River water (Error bars represent the variability due to
multiple formation potential tests (n = 3))
147
For ranitidine, prechlorination impacts observed in the lake water (Figure 7.3) showed similar
trends to those that were observed in MQ water (Figure 7.2); the ultimate NDMA formation was
significantly reduced upon prechlorination, and increases in free chlorine contact time did not
cause further reductions in NDMA formation. Moreover, the initial lag period in the lake water
was significantly shortened by more than half upon prechlorination, and the estimated model
parameters in MQ and lake water were comparable with each other when prechlorination was
applied at the same contact time (Table 7.3). Previously, a NOM-pharmaceutical binding
hypothesis was proposed to explain the initial lag phase observed in real water matrices (see
details in Section 5.3.2). The results here suggest that chlorination may destroy some of the
binding between NOM and ranitidine and thus enhance the initial reaction between ranitidine
and chloramine to form NDMA.
However, the results in the river water (Figure 7.3) were quite different for ranitidine. Upon 3
min prechlorination, the ultimate NDMA molar conversion was reduced by 23 % and the initial
lag period was shortened from 31 hr to 4 hr. When the chlorine contact time further increased to
10 min, the NDMA conversion was significantly inhibited to 12.5 %, and very little NDMA
(molar conversion less than 3 %) was detected at longer chlorine contact times. It is not likely
that prolonged chlorination simply destroyed the NDMA precursor (ranitidine in this case)
because similar results would have been obtained in MQ and the lake water. Considering the
NOM-ranitidine binding theory, the results suggest that prolonged chlorination might further
break down the NOM to form smaller products (this was confirmed by LC-OCD results and will
be discussed in Section 7.3.3.3) which may rebind with ranitidine and thus inhibit the NDMA
formation. However, it is still unclear why this “rebinding” was not observed in the lake water.
Since the Cl2:TOC ratio applied was higher in the lake water compared with that in the river
water, initially it was hypothesized that the higher Cl2:TOC ratio applied in the lake water may
have helped prevent the “rebinding”. However, when the Cl2:TOC ratio in the river water was
raised to the same level as applied in the lake water, a similar trend was observed for ranitidine
in the river water (see details in Appendix 12), indicating that the Cl2:TOC ratio was not the
reason for the different results observed in the river water. Another possible explanation could be
the much higher TOC level in the river water, yielding more NOM-breakdown products that can
rebind with ranitidine. Additional LC-OCD results (Appendix 15, Figure A15.4) confirmed that
148
there were small NOM molecules (i.e. building blocks and the LMW fractions) in the river water
due to the breakdown of larger NOM fractions (i.e., biopolymers and humics) upon sequential
chlorination and chloramination, which might be responsible for the “rebinding”.
Figure 7.4. NDMA formation kinetics from sumatriptan upon sequential chlorine and chloramine disinfection in Lake Ontario and Otonabee River water (Error bars represent the variability due to
multiple formation potential tests (n = 3))
Sumatriptan’s response to prechlorination in natural water matrices (Figure 7.4) was quite
different compared with that observed in MQ water (Figure 7.2), except for the common trend in
all of the matrices of a reduced initial lag phase upon prechlorination. However, unlike the
consistently increasing NDMA conversion from sumatriptan in MQ water, there was a maximum
NDMA molar conversion observed upon 30 min of prechlorination. Further increases in chlorine
contact time reduced the ultimate molar conversion. This phenomenon might also be explained
by the proposed rebinding theory that prolonged chlorination could breakdown NOM molecules
to smaller products and thus enhance their rebinding with sumatriptan, as will be discussed
further in Section 7.3.3.3.
In summary, prechlorination was shown to affect both the ultimate NDMA conversion (θ) and
the reaction kinetics (Lag and k). As compared in Figure 7.5, the impact was both compound-
specific and matrix-specific, thus requiring case-by-case investigation. (A detailed statistical
analysis of the estimated model parameters is summarized in Appendix 13.) Currently, most
studies have found that prechlorination with enough CT (concentration × contact time) has, in
149
general, been able to reduce the NDMA formation. However, only single-precursor scenarios
have been considered in most studies, i.e., NOM or DMA as the only precursor. This work
suggests that the potential interactions between NOM and other precursors could also affect
NDMA formation. Moreover, an enhanced NDMA formation rate upon prechlorination has not
been reported previously, which is a new contribution to the field.
Figure 7.5. Comparison of the kinetic model parameters upon sequential chlorine and chloramine
disinfection in three water matrices (Error bars represent the 95 % confidence interval)
150
7.3.3 NOM-Pharmaceutical-Cl2 Interactions
Both free chlorine and NOM have been shown to individually affect the conversion of selected
pharmaceuticals into NDMA. Since systems involving free chlorine, NOM, and the target
pharmaceuticals are more complicated, it is important to investigate the interactions in between
them to further understand the impact from matrix components as well as prechlorination.
7.3.3.1 NOM-Pharmaceutical Interactions
The potential for different types of interactions between NOM and selected pharmaceuticals have
been discussed in Chapter 5 (Section 5.3.2). The formation of NOM-pharmaceutical complexes
is most likely due to the electrostatic attraction between the negatively charged NOM surface
(De Ridder et al., 2011) and the positively charged amines, and/or due to the covalent binding
between aromatic amines and specific functional groups in humic substances such as carbonyls
and quinines (Parris, 1980; Thorn et al., 1996; Weber et al., 1996). Although there is no direct
evidence to prove the binding in aqueous phase, the NOM-pharmaceutical binding hypothesis
has been suggested by other researchers to explain the enhanced pharmaceutical removal during
treatment processes like coagulation and membrane filtration (Ballard and Mackay, 2005;
Diemert, 2012; Stackelberg et al., 2007; Vieno et al., 2006; Westerhoff et al., 2005). Without the
application of prechlorination, the matrix components were found to slow down the initial
NDMA formation (i.e., longer initial lag phases were associated with higher TOC and SUVA),
while they had less impact on the ultimate NDMA molar conversion (Figure 7.5, pre-Cl = 0
min).
Currently, there is no direct spectroscopic evidence proving the existence of the NOM-
pharmaceutical complexes in the aqueous phase. Some preliminary Liquid Chromatography-
Mass Spectrometry (LC-MS) analysis was performed in this work, intended to enable
comparisons of the “free” pharmaceutical concentrations with and without the presence of NOM.
However, the results were inconclusive due to limitations in the analytical method (see details in
Appendix 14).
151
7.3.3.2 Cl2-Pharmaceutical Interactions
Results from this work showed that prechlorination can either increase or decrease NDMA
formation from selected pharmaceuticals. Because the NDMA formation can be described as an
electrophilic reaction on the nitrogen atom of the amine group (Mitch and Schreiber, 2008),
theoretically chlorination could enhance the reaction if it enhances the electron density on the
amine group, and vice versa. Therefore, it is important to evaluate the different reactivity of
selected pharmaceuticals towards free chlorine in order to determine how the final NDMA
conversion will be affected by prechlorination. Moreover, the DMA group is not the only
functional group on the pharmaceuticals that could react with chlorine. If the groups close to the
DMA group are modified upon chlorination, they may also impact the electron density of the
amine group, and thus affect the NDMA conversion. Among the eight pharmaceuticals being
tested, ranitidine and sumatriptan had relatively high NDMA conversions but showed opposite
trends when being subjected to prechlorination. Thus the following discussion mainly focuses on
these two pharmaceuticals.
Deborde and von Gunten (2008) have reviewed the kinetics and mechanisms for the chlorination
of organic compounds. Generally, rate constants for the reaction of chlorine with sulphur-
moieties are typically 1~2 orders of magnitude higher than with amines. According to this
review, the most potential attack site of chlorine on ranitidine would be the sulphur, forming
ranitidine sulfoxide (S=O). Then the ranitidine sulfoxide may go through C-S fragmentation
upon further chlorination, resulting in the formation of smaller molecules (Figure 7.6). The C-S
fragmentation upon chlorination was confirmed by Le Roux et al. (2012b) for ranitidine, and the
breakdown product (5-(dimethylaminomethyl)-2-furanmethanol (DFUR)) is still a potential
NDMA precursor, with NDMA molar conversion being reported from 50 % to 80 % upon
preformed chloramination (Schmidt et al., 2006; Le Roux et al., 2012a). However, the DMA
group on DFUR may now become the primary chlorine attack site. Abia et al. (1998) also
proposed a reaction mechanism for the chlorination of tertiary amines, including an elementary
step in which a positive charge is developed on the N-atom via a chlorine transfer. This supports
the results of a previous study which has suggested that prechlorination can significantly reduce
the NDMA formation from DMA due to the partially formed Cl-DMA that inhibits the
electrophilic attack on the amine group (Schreiber and Mitch, 2005). Le Roux et al. (2012b) have
152
also suggested that the chlorine attack on the amine group could reduce the NDMA conversion
from ranitidine because it limited the direct substitution on the DMA group.
Figure 7.6. Possible chlorine attack site and predicted fragmentation pattern of ranitidine
Unlike ranitidine, the NDMA conversion from sumatriptan was enhanced upon prechlorination.
As indicated in Figure 7.7, the sulphur moiety of sumatriptan is already fully oxidized (-SO2),
therefore the proposed C-S fragmentation for ranitidine may not occur. Moreover, the DMA
group is almost completely protonated at pH 7 (pKa = 9.6; 99.8 % protonated), while the chlorine
reaction is only significant with neutral amines (reaction rate constants in the range of 101 – 102
M-1s-1 at pH 7 (Deborde and von Gunten, 2008)). As a result, the DMA group may not be the
primary chlorine attack site of sumatriptan at the given reaction conditions. In addition, Xu et al.
(2001) have suggested that the C2 position of indole can be activated for oxidation when the C3
site has an R group. In the case of chlorination, the C2 position could be substituted with
chlorine (Gilow and Burton, 1981), and the reaction occurred very fast, with a conservative
estimation of the reaction rate constants to be at least in the range of 101 – 102 M-1s-1 at pH 7 (Lin
and Carlson, 1984). The substituted chlorine on the C2 position may then increase the electron
density at C3 position due to the conjugation effect between the double bound and the lone
electron pair on the Cl (Figure 7.7), thus enhancing NDMA formation from sumatriptan.
Although this reaction is plausible from the literature cited above, it would benefit from mass
spectral confirmation of the intermediate species which, unfortunately, was not possible as part
of the current study.
153
Figure 7.7. Possible chlorine attack site(s) and predicted chlorination product of sumatriptan
In summary, two major factors should be considered when evaluating the effect of
prechlorination on the NDMA conversion from amine-based pharmaceuticals: i) if the DMA
group on these pharmaceuticals is the primary chlorine attack site; ii) if the DMA group is
protonated (pH-pKa relevant). Table 7.4 compares the differences in the chlorine reactivity
towards ranitidine and sumatriptan.
Table 7.4. Comparison of chlorine reactivity towards ranitidine and sumatriptan
Ranitidine Sumatriptan
Molecular structure
DMA group protonated on the parent compound at given pH (7.0 ± 0.1)? ~ 94 % ~ 100 %
Primary chlorine attack site(s) of the parent
compound (indicated by “ ”) Sulphur Secondary amine
C2 site of the indole ring Possible fragmentation upon chlorination? Yes No DMA group protonated on the breakdown product(s) at given pH (7.0 ± 0.1)? ~ 92% /
Primary chlorine attack site(s) of the breakdown product(s) DMA /
Electron density of DMA group increased? No Yes
Ultimate NDMA molar conversion upon prechlorination Reduced Enhanced
154
Usually chlorine would cause a small modification on the parent compound rather than a
complete oxidation. The modification may change the pharmaceutical reactivity towards the
subsequent chloramine, but it could also change their binding potential with NOM which would
affect the availability of the pharmaceuticals as well. For example, if chlorination increases the
electron density on the amine group of sumatriptan (Figure 7.7), it will not only enhance the
NDMA conversion from sumatriptan (i.e., favoring the electrophilic attack), but also reduce the
binding with the negatively charged NOM surface thus improving the initial reaction. The latter
might also contribute to the much shortened initial lag phase for sumatriptan upon
prechlorination in lake and river water (Figure 7.4).
7.3.3.3 Cl2-NOM Interactions
Chlorine can also react with NOM and modify its surface properties, which would then affect the
binding potential between NOM and pharmaceuticals. The potential interactions between
chlorine and NOM were investigated via LC-OCD analysis. Water samples were collected along
with the chloramination and sequential disinfection experiments (from the blank control samples
in each set of kinetic experiments) to investigate the potential change of NOM components upon
the treatment.
For the raw water samples used in this study (lake and river water), the majority of NOM was
hydrophilic (more than 90 % of the total DOC), and the major fractions were humic substances
(HS) and building blocks (accounting for 70 % to 80 % of the total hydrophilic NOM). In
general, DOC was not removed by chlorination or chloramination at the applied reaction
conditions, but there were changes in the major NOM fractions (i.e., HS and building blocks).
Typically, degradation of HS was observed to lead to the formation of building blocks, and
further breakdown of building blocks would lead to the increase in LMW fractions (data for
other NOM fractions are presented in Appendix 15). At the employed reaction conditions, free
chlorine itself did not change any NOM fractions significantly; however, it facilitated the
transformation of HS into building blocks when NH4Cl was added subsequently to form
chloramine. For example, following 120 min prechlorination, the HS in lake water started to
155
degrade after 2 hr of NH4Cl addition. However, without prechlorination, the reduction in HS
occurred only after 4 hr of chloramination (Figure 7.8).
Figure 7.8. Change of major NOM fractions upon sequential chlorination and chloramination in Lake Ontario water
This trend was further confirmed in the river water (Figure 7.9). The application of
prechlorination significantly enhanced the degradation of HS, and the reduction was improved as
the chlorine contact time increased. Compared with preformed chloramine (i.e., pre-Cl = 0 min),
there were more building blocks formed when prechlorination was applied; as chloramination
contact time increased, the building blocks first increased (from the breakdown of HS) and then
decreased subsequently (further breakdown to LMW fractions). Specifically, the highest increase
in building blocks occurred at 30 min of prechlorination followed by chloramination; this is in
agreement with the kinetic experiments for sumatriptan in lake and river water, where the
maximum NDMA molar conversion was also observed upon 30 min of prechlorination (Figure
7.4). It was proposed that prolonged prechlorination may destroy some NOM fractions and the
breakdown products may rebind with the pharmaceutical and thus inhibit its conversion into
NDMA. The LC-OCD results suggest that prolonged chlorine contact time does lead to early
breakdown of building blocks, yielding more LMW molecules which may rebind with
pharmaceuticals and inhibit their further conversion into NDMA.
156
Figure 7.9. Change of major NOM fractions upon sequential chlorination and chloramination in Otonabee River water
Figure 7.10 summarizes the possible interactions in between NOM, pharmaceuticals, and free
chlorine. Without the application of prechlorination, the presence of NOM usually does not
affect the ultimate NDMA conversion, but will affect the reaction rate due to the formation of
NOM-pharmaceutical complexes. However, free chlorine could affect both the NDMA
conversion and the reaction rate by altering the properties of the pharmaceuticals and/or NOM,
as well as the binding potential between the two.
Figure 7.10. Potential interactions in between NOM, pharmaceuticals, and free chlorine
157
7.4 Summary
Previous research that looked into prechlorination effects mostly considered the single-precursor
scenarios, while very few took into consideration the interactions between matrix components
and the precursors. Moreover, little is known in terms of the prechlorination impact on the
NDMA formation kinetics. This work demonstrates that prechlorination affects both the ultimate
NDMA conversion and the reaction rate, and investigates how the interactions in between NOM,
pharmaceuticals, and free chlorine could affect the NDMA formation.
The prechlorination impact is compound-specific and matrix-specific, requiring case-by-case
investigation. In general, prechlorination tends to reduce the NDMA conversion from ranitidine
while it increases that from sumatriptan. In the presence of NOM, prechlorination may enhance
the initial reaction by breaking the binding between NOM and pharmaceuticals, but prolonged
prechlorination may also break down NOM into smaller products which may rebind with
pharmaceuticals and thus inhibit its further conversion into NDMA. In addition, an enhanced
NDMA formation rate upon prechlorination has not been reported previously.
In most cases, it would be expected that applying sufficient prechlorination may help reducing
the maximum NDMA concentration at the farther end of the distribution system. However, for
certain precursors like the amine-based pharmaceuticals studied in this work, short
prechlorination sometimes might break the binding between NOM and precursors and thus
enhance the initial reaction, resulting higher level of NDMA within the sections with shorter
water age. Therefore, a good NDMA control strategy would require knowledge about the source
water, the precursor property, the treatment applied at the plant, the size of the distribution
system, as well as the location of households and water age.
158
7.5 References
Abia, L., Armesto, X.L., Canle, L.M., Garcia, M.V., Santaballa, J.A., 1998. Oxidation of
aliphatic amines by aqueous chlorine. Tetradedron 54: 521-530
Andrzejewski, P., Kasprzyk-Hordern, B., Nawrocki, J., 2008. N-nitrosodimethylamine (NDMA)
formation during ozonation of dimethylamine-containing waters. Water Res. 42, 863-870
Ballard, B.D., Mackay, A.A., 2005. Estimating the removal of anthropogenic organic chemicals
from raw drinking water by coagulation flocculation. J. Environ. Eng. 131, 108-118
Charrois, J.W.A., Hrudey, S.E., 2007. Breakpoint chlorination and free-chlorine contact time:
implications for drinking water N-nitrosodimethylamine concentrations. Water Res. 41, 674-682
Chen, Z., Valentine, R.L., 2007. Formation of N-Nitrosodimethylamine (NDMA) from humic
substances in natural water. Environ. Sci. Technol. 41, 6059-6065
Chen, Z., Valentine, R.L., 2008. The influence of the pre-oxidation of natural organic matter on
the formation of N-Nitrosodimethylamine (NDMA). Environ. Sci. Technol. 42, 5062-5067
Chen, Z., Yang, L., Zhai, X., Zhao S., Li, A., Shen, J., 2010. N-nitrosamine formation during
chlorination/chloramination of bromide-containing water. Water Sci. Technol.: Water Supply 10,
462-471
Choi, J.H., Valentine, R.L., 2003. N-Nitrosodimethylamine Formation by Free-Chlorine-
Enhanced Nitrosation of Dimethylamine. Environ. Sci. Technol. 37, 4871-4876
Comerton, A.M., Andrews, R.C., Bagley, D.M., 2006. Impact of blending reuse and lake water
on treated water quality. J. Environ. Eng. Sci. 5: 359-363
De Ridder, D.J., Verliefde, A.R.D., Heijman, S.G.J., Verberk, Q.J.C., Rietveld, L.C., van der Aa,
L.T.J., Amy, G.L., van Dijk, J.C., 2011. Influence of natural organic matter on equilibrium
159
adsorption of neutral and charged pharmaceuticals onto activated carbon. Wat. Sci. Technol. 63,
416 – 423
Deborde, M., von Gunten, U., 2008. Reactions of chlorine with inorganic and organic
compounds during water treatment – kinetics and mechanisms: a critical review. Water Res. 42,
13-51
Diemert, S.A., 2012. The impact of coagulation on endocrine disrupting compounds,
pharmaceutically active compounds and natural organic matter. Master’s thesis, Univesity of
Toronto, Ontario
Dotson, A., Westerhoff, P., Krasner, S.W., 2007. Nitrosamine formation from natural organic
matter isolates and sunlight photolysis of nitrosamines. In: Proceedings of AWWA Annual
Conference and Exposition. Toronto, ON, Canada, June 24-28, 2007
EPA Integrated Risk Information System (IRIS), 1993. N-Nitrosodimethylamine; CASRN 62-
75-9. www.epa.gov/iris/subst/0045.htm
Gerecke, A.C., Sedlak, D.L., 2003. Precursors of N- Nitrosodimethylamine in natural waters.
Environ. Sci. Technol. 37, 1331-1336
Gilow, H.M., Burton, D.E., 1981. Bromination and chlorination of pyrrole and some reactive 1-
substituted pyrroles. J. Org. Chem. 46: 2221 – 2225
Harvey, M. 2009. The impacts of UV direct photolysis and UV/H2O2 advanced oxidation
processes on the formation of nitrosamines and organic chloramines from subsequent
chlor(am)ination. Master’s thesis, Univesity of Toronto, Ontario
Health Canada, 2010. Guideline Technical Document on N-nitrosodimethylamine (NDMA) in
drinking water for public comment. http://www.hc-sc.gc.ca/ewh-semt/consult/_2010/ndma/draft-
ebauche-eng.php#a3
160
Huber, S.A., Balz, A., Abert, M., Pronk, W., 2011. Characterisation of aquatic humic and non-
humic matter with size-exclusion chromatography – organic carbon detection – organic nitrogen
detection (LC-OCD-OND). Water Res. 45, 879-885
Kemper, J.M., Walse, S.S., Mitch, W.A., 2010. Quaternary amines as nitrosamine precursors: a
role for consumer products? Environ. Sci. Technol. 44, 1224-1231
Kohut, K.D., Andrews, S.A., 2003. Polyelectrolyte age and N-nitrosodimethylamine formation
in drinking water treatment. Water Qual. Res. J. Can. 38 (4), 719-735
Krasner, S.W., 2009. The formation and control of emerging disinfection by-products of health
concern. Phil Trans. R Soc. A 367, 4077-4095
Le Roux, J., Gallard, H., Croue, J., 2011. Chloramination of nitrogenous contaminants
(pharmaceuticals and pesticides): NDMA and halogenated DBPs formation. Water Res. 45,
3164-3174
Le Roux, J., Gallard, H., Croue, J., 2012a. Formation of NDMA and halogenated DBPs by
chloramination of tertiary amines: the influence of bromide ion. Environ. Sci. Technol. 46,
1581-1589
Le Roux, J., Gallard, H., Croue, J., Papot, S., Deborde, M., 2012b. NDMA formation by
chloramination of ranitidine: kinetics and mechanism. Environ. Sci. Technol. 46, 11095-11103
Lee, C., Schimidt, C., Yoon, J., Von Gunten, U., 2007. Oxidation of N-Nitrosodimethylamine
(NDMA) precursors with ozone and chlorine dioxide: kinetics and effect on NDMA formation
potential. Environ. Sci. Technol. 41, 2056-2063
Lee, C., Lee, Y., Schmidt, C., Yoon, J., Von Gunten, U., 2008. Oxidation of suspected N-
nitrosodimethylamine (NDMA) precursors by ferrate (VI): Kinetics and effect on the NDMA
formation potential of natural waters. Water Res. 42, 433-441
161
Lin, S., Carlson, R.M., 1984. Susceptibility of environmentally important heterocycles to
chemical disinfection: reactions with aqueous chlorine, chlorine dioxide, and chloramine.
Environ. Sci. Technol. 18, 743-748
Massachusetts Department of Environmental Protection (MassDEP), 2004. Current Regulatory
Limit: n-Nitrosodimethylamine (NDMA), CASRN 62759.
http://www.mass.gov/dep/water/drinking/standards/ndma.htm
Mitch, W.A., Krasner, S.W., Lee, C.F.T., Wong, T.M., 2010. Tradeoffs in DBP formation during
chloramination following pre-oxidation for nitrosamine control. In: Proceedings of the AWWA
Water Quality Technology Conference, Savannah, GA, November 14-18, 2010
Mitch, W.A., Schreiber, I.M., 2008. Degradation of tertiary alkylamines during
chlorination/chloramination: implications for formation of aldehydes, nitriles, halonitroalkanes,
and nitrosamines. Environ. Sci. Technol. 42, 4811-4817
Mitch, W.A., Sedlak, D.L., 2004. Characterization and fate of N-nitrosodimethylamine
precursors in municipal wastewater treatment plants. Environ. Sci. Technol. 38, 1445-1454
MOE, 2003. Ontario regulation 268/03 made under the safe drinking water act, 2002.
http://www.e-laws.gov.on.ca/html/source/regs/english/2003/elaws_src_regs_r03268_e.htm
Najm, I., Trussell, R.R., 2001. NDMA formation in water and wastewater. J. Am. Water Works
Assoc. 93 (2), 92-99
NHMRC, NRMMC, 2011. Australian Drinking Water Guidelines Paper 6 National Water
Quality Management Strategy. National Health and Medical Research Council, National
Resource Management Ministerial Council, Commonwealth of Australia, Canberra.
Office of Environmental Health Hazard Assessment (OEHHA), 2006. Public Health Goal for N-
nitrosodimethylamine and cadmium in drinking water.
http://www.oehha.org/water/phg/cadndma122206.html
162
Oya, M., Kosaka, K., Asami, M., Kunikane, S., 2008. Formation of N-nitrosodimethylamine
(NDMA) by ozonation of dyes and related compounds. Chemosphere 73, 1724-1730
Parris, G.E., 1980. Covalent binding of aromatic amines to humates. 1. reactions with carbonyls
and quinones. Environ. Sci. Technol. 14, 1099-1106
Pisarenko, A.N., Stanford, B.D., Yan, D., Gerrity, D., Snyder, S.A., 2012. Effects of ozone and
ozone/peroxide on trace organic contaminants and NDMA in drinking water and water reuse
applications. Water Res. 46, 316-326
Radjenovic, J., Farre, M.J., Gernjak, W., 2012. Effect of UV and UV/H2O2 in the presence of
chloramines on NDMA formation potential of tramadol. Environ. Sci. Technol. 46, 8356-8364
Schmidt, C.K., Brauch, H.J., 2008. N,N-Dimethylsulfamide as precursor for N-Nitrosodimethylamine (NDMA) formation upon ozonation and its fate during drinking water treatment. Environ. Sci. Technol. 42, 6340-6346
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.
Schreiber, I.M., Mitch, W.A., 2005. Influence of the order of reagent addition on NDMA
formation during chloramination. Environ. Sci. Technol. 39, 3811-3818
Schreiber, I.M., Mitch, W.A., 2006a. Nitrosamine formation pathway revisited: the importance
of chloramines speciation and dissolved oxygen. Environ. Sci. Technol. 40, 6007-6014
Schreiber, I.M., Mitch, W.A., 2006b. Occurrence and fate of nitrosamines and nitrosamine
precursor in wastewater-impacted surface waters using boron as a conservative tracer. Environ.
Sci. Technol. 40, 3203-3210
163
Shah, A.D., Krasner, S.W., Lee, C.F.T., von Gunten, U., Mitch, W.A., 2012. Trade-offs in
disinfection byproduct formation associated with precursor preoxidation for control of N-
Nitrosodimethylamine formation. Environ. Sci. Technol. 46, 4809-4818
Stackelberg, P.E., Gibs, J., Furlong, E.T., Meyer, M.T., Zaugg, S.D., Lippincott, R.L., 2007.
Efficiency of conventional drinking-water-treatment processes in removal of pharmaceuticals
and other organic compounds. Sci. Total Environ. 377, 255-272
Thorn, K.A., Pettigrew, P.J., Goldenberg, W.S., Weber, E.J., 1996. Covalent binding of aniline
to humic substances. 2. 15N NMR studies of nucleophilic addition reactions. Environ. Sci.
Technol. 30, 2764-2775
USEPA, 2009. Contaminant Candidate List 3 (CCL3).
http://water.epa.gov/scitech/drinkingwater/dws/ccl/ccl3.cfm
Valentine, R.L., Choi, J., Chen, Z., Barrett, S.E., Hwang, C., Guo, Y.C., Wehner, M.,
Fitzsimmons, S., Andrews, S.A., Werker, A.G., Brubacher, C., Kohut, K., 2005. Factors
affecting the formation of NDMA in water and occurrence, Denver, CO, pp. 81-90
Vieno, N., Tuhkanen, T., Kronberg, L., 2006. Removal of pharmaceuticals in drinking water
treatment: effect of chemical coagulation. Environ. Technol. 27, 183 – 192
Weber, E.J., Spidle, D.L., Thorn, K.A., 1996. Covalent binding of aniline to humic substances. 1.
kinetic studies. Environ. Sci. Technol. 30, 2755-2763
Westerhoff, P., Yoon, Y., Snyder, S., Wert E., 2005. Fate of endocrine-disruptor,
pharmaceutical, and personal care product chemicals during simulated drinking water treatment
processes. Environ. Sci. Technol. 39, 6649-6663
Wilczak, A., Assadi-Rad, A., Lai, H.H., Hoover, L.L., Smith, J.F., Berger, R., Rodigari, F.,
Beland, J.W., Lazzelle, L.J., Kincannon, E.G., Baker, H., Heaney, C.T., 2003. Formation of
164
NDMA in chloraminated water coagulated with DADMAC cationic polymer. J. Am. Water
Works Assoc. 95 (9), 94-106
Woodbeck, M. 2007. Evaluating the potential genotoxicity of disinfected source waters. Master’s
thesis, University of Toronto, Ontario
Xu, X., Bartlett, M.G., Stewart, J.T., 2001. Determination of degradation products of sumatriptan
succinate using LC-MS and LC-MS-MS. J. Pharmaceut. Biomed. 26: 367-377
165
Chapter 8 Conclusions and Recommendations for Future Research
This study set out to investigate the potential for amine-based pharmaceuticals and personal care
products (PPCPs) to form nitrosamines upon drinking water disinfection, and has demonstrated
the nitrosamine formation kinetics from selected PPCPs in various water matrices under different
reaction conditions. Currently, most of the findings in the literature are based on tests in lab-
grade water, while few studies consider the potential interactions between natural organic matter
(NOM) and the target precursors which may also affect N-Nitrosodimethylamine (NDMA)
formation. The present work has investigated the role of NOM in NDMA formation from
selected pharmaceuticals. In addition, the general literature on nitrosamines is mainly based on
the model precursor dimethylamine (DMA) and NOM under the scenario of comparable
concentrations of precursors and chloramines. As such, findings for DMA and NOM may not
necessarily apply to PPCP-based precursors which are usually present at much lower
concentrations relative to chloramine concentrations in real water samples. This study has thus
evaluated some critical factors that affect nitrosamine formation from PPCPs by including
considerations for NOM interactions as well as reactions at low PPCP concentrations and typical
chloramine dosages.
8.1 Discussion of Major Themes from the Thesis as a Whole
8.1.1 Nitrosamine Formation in Lab-Grade Water
All of the twenty selected PPCPs were demonstrated to form the corresponding nitrosamines
upon chloramination, and eight of them had molar conversions higher than 1 %. The selected
PPCPs are essentially tertiary amines, and NDMA formation from tertiary amines has been
proposed to involve a release of DMA and the subsequent oxidation of DMA to NDMA (Mitch
and Sedlak, 2004). However, the molar yield of NDMA from DMA was reported to be lower
than 1 % (Mitch et al., 2003), which cannot explain all of the NDMA formation from the eight
pharmaceuticals with molar conversions higher than 1 %. This is in agreement with the
observations by Lee et al. (2007) for a few other tertiary amines, suggesting that the NDMA
166
formation from amine-based pharmaceuticals may undergo a direct formation pathway that does
not involve the release of DMA. For example, if NDMA formation can be described as an
electrophilic attack on the N-atom of an amine group (Mitch and Schreiber, 2008), then high
NDMA-formation potentials (FPs) from the eight pharmaceuticals might be explained by an
enhanced electron density on the N-atoms of the DMA group that have electron-rich moieties
adjacent to them.
When the pharmaceutical concentrations applied in this work (1.4-22.2 µg/L) were 1-2 orders of
magnitude higher than their environmental levels, the NDMA concentrations formed were also
1-2 orders of magnitude higher than the currently regulated NDMA level in drinking water.
When lower pharmaceutical concentrations (100-400 ng/L) were applied, the NDMA formation
varied from 1.0 up to 61.6 ng/L, with some samples exceeding the currently regulated level.
Moreover, because chloramine was present in large excess relative to the PPCP concentrations,
the NDMA molar conversion was relatively independent of the initial pharmaceutical
concentration, and only small antagonistic effects were observed (less than 10-15%) for mixtures
of pharmaceuticals. This indicates that similar molar conversions would be expected for
individual pharmaceuticals at environmental concentrations. Although each of them may only
form NDMA at the sub-ng/L level, added together they may still pose a concern in terms of the
overall NDMA formation potential.
Currently, PPCPs are not subject to any regulated monitoring protocol, and their presence in
drinking water sources at trace levels has been suggested to be inconsequential. However, this
conclusion was drawn based on the toxicity study of the PPCPs themselves, and did not consider
the potential adverse effects associated with their transformation into carcinogenic DBPs upon
drinking water treatment, as demonstrated by this work. Moreover, the formation of nitrosamines
from amine-based PPCPs could be of a particular concern in water reuse processes which
involve the chloramination of much higher concentrations of PPCPs.
167
8.1.2 NDMA Formation in Water Matrices Containing NOM
NDMA formation over time from amine-based pharmaceuticals followed a three-phase
formation curve, including an initial lag phase, a fast increase in NDMA formation, and
eventually a plateau representing the ultimate NDMA molar conversion. The three-phase
formation curve was observed for all selected pharmaceuticals in all of the water matrices tested.
Compared with lab-grade MQ water, the presence of NOM in real water samples was found to
slow down the initial reaction, resulting in a longer initial lag phase, but it had no significant
impact on the ultimate NDMA molar conversion given enough reaction time. It has been
hypothesized that amine-based pharmaceuticals may bind to NOM due to electrostatic attraction
between the positively charged amines and the negatively charged NOM surface (de Ridder et
al., 2011) and/or covalent binding between aromatic amines and specific functional groups on
the NOM surface (Parris, 1980; Thorn et al., 1996; Weber et al., 1996). The formation of NOM-
pharmaceutical complexes could inhibit the initial contact between pharmaceuticals and the
chloramines, thus slowing down the initial reaction.
Although NOM had a profound impact on the NDMA formation rate, the initial lag phase and
the pseudo-first order rate constant were not always well correlated with the bulk water quality
measurements of TOC and SUVA, or any specific NOM fractions quantified via LC-OCD
analysis. For untreated water samples (i.e., raw water from Lake Ontario and Otonabee River),
longer initial lag phase was associated with higher TOC and SUVA values (Chapter 7, Figure
7.5, pre-Cl = 0 min). However, this trend was not followed when including the treated water
samples (i.e., treated Otonabee River water and the Toronto tap water), possibly because the
treatment processes at the water treatment plant changed the surface characteristics of the NOM
(e.g., steric structure, charge distribution, etc.) and thus altered the binding potential between
NOM and pharmaceuticals (Chapter 5, Section 5.3.2). Therefore, the matrix effect is likely not
only affected by the amount of NOM but also the amount of certain moieties or functional
groups as well as the surface property.
Most bench-scale NDMA-FP tests have applied a 24 hr incubation time from a practical
viewpoint, and also for better comparisons with other DBPs measured according to typical DBP
formation potential protocols (e.g., Summers et al., 1996). The delayed NDMA formation due to
168
the presence of NOM indicates that short-term NDMA tests may underestimate the ultimate
NDMA-FP especially for the slow-reacting precursors. It is recommended to perform a
preliminary kinetic test in order to determine the proper incubation time for a specific water
matrix, or to perform the NDMA-FP tests according to the water age of the respective
distribution system of interest.
8.1.3 Critical Factors Affecting NDMA Formation from Pharmaceuticals
Basically, NDMA formation is determined by both the chloramine species present and the
precursor amine species present. Under typical drinking water disinfection conditions, the
chloramine speciation is determined by the Cl2:NH4-N mass ratio along with the pH. NDMA
formation from selected pharmaceuticals was found to increase with the Cl2:NH4-N mass ratio,
suggesting an enhancement effect of dichloramine. Also dichloramine is mainly formed at acidic
pH, and thus NDMA conversion is generally limited at basic pH due to the lack of dichloramine.
On the other hand, the amine speciation is mainly determined by the pH and the conventional
relationship between pH and pKa of the amine group, and the NDMA formation reaction is
favored by the non-protonated amines which increase with increasing pH. Therefore, the reaction
rate increases with pH despite the reduced ultimate NDMA conversion at higher pH, and the
NDMA conversion is limited at acidic pH due to the lack of non-protonated amines. Because of
these competing issues, the maximum NDMA formation from amine-based pharmaceuticals is
expected to occur at a medium pH range of approximately 7-8.
Prechlorination has been shown to destroy or transform NDMA precursors and thus reduce
NDMA formation upon chloramination downstream. However, this study demonstrates that
prechlorination does not always reduce the NDMA formation from amine-based pharmaceuticals
and the efficiency of prechlorination-related precursor removal is significantly affected by the
presence of NOM. Prechlorination typically causes a minor modification on the pharmaceuticals
and changes their reactivity towards chloramine. Depending on the structure of the respective
compound, especially the structures surrounding the amine group, this modification may either
increase or decrease its potential to form NDMA. The transformation products of the amine-
based precursors may still contribute to the overall NDMA formation as long as the amine group
169
is present in the resulting structure. For example, a recent study by Radjenovic et al. (2012)
showed that the oxidation products of the pharmaceutical tramadol by UV and UV/ H2O2
treatment have higher NDMA FPs than the parent compound. Transformation of precursors may
also happen in the environment and/or in other treatment processes, therefore the overall NDMA
formation should consider both the parent compound and its transformation products.
The presence of NOM can slow down the initial NDMA formation from pharmaceuticals via
NOM-pharmaceutical associations, but the application of prechlorination may destroy some of
this binding between NOM and pharmaceuticals and actually enhance the initial reaction. This
was observed in this study when shorter chlorine contact times were applied and/or in water with
lower TOC. Prechlorination can also break down NOM into smaller molecules which may form
new bonds with the pharmaceuticals and inhibit further reaction, especially for longer chlorine
contact times and/or in waters with higher TOC.
To date, most nitrosamine studies have been conducted using lab-grade water spiked with
individual precursors in the absence of NOM, while the present work has demonstrated the
importance of evaluating the impact of various factors in the presence of NOM. Also while
results from this work have been obtained from selected amine-based PPCPs, they may also
apply to other amine-based precursors with similar structures. Figure 8.1 below is a
comprehensive summary of the critical factors that affect NDMA formation as investigated in
this study.
170
Figure 8.1. Critical factors that affect the NDMA formation from amine-based precursors
8.1.4 Implications for NDMA Formation Control
NDMA formation involves a number of variables and is highly case-specific, and minimizing
NDMA formation requires the removal of potential precursors and/or the proper control of
disinfection conditions.
Preoxidation has been shown to be effective in removing certain NDMA precursors, but the
overall effectiveness for a mixture of precursors needs to be evaluated by preliminary tests.
Moreover, the applied dosage and contact time need to be selected with caution. As
demonstrated in this work, short prechlorination contact times may actually enhance the initial
NDMA formation. Prolonged prechlorination could inhibit the further reaction, but the
chlorination of NOM also results in the formation of other regulated DBPs such as
trihalomethanes (THMs) and haloacetic acids (HAAs), posing a significant risk tradeoff.
The proposed NOM-pharmaceutical binding theory suggests an alternative to removing potential
NDMA precursors. Like the amine-based pharmaceuticals selected in this study, most amines are
positively charged in the aquatic environment while the NOM surface is typically negatively
charged; therefore, perhaps the NDMA precursors may be co-removed with NOM via the
171
formation of NOM-complexes. For example, several studies have reported the removal of certain
pharmaceuticals upon coagulation, likely due to the sorption onto particulate organic matter and
co-removed through the settling process (Ballard and Mackay, 2005; Diemert, 2012;
Stackelberg et al., 2007; Vieno et al., 2006; Westerhoff et al., 2005). As such, a good NDMA
control strategy should also consider optimizing the treatment processes prior to the disinfection,
such as coagulation. Moreover, improving the removal of NOM would allow longer contact time
if prechlorination is applied downstream to further remove the NDMA precursors.
On the other hand, the key to optimizing chloramine disinfection is to minimize the formation of
dichloramine, which is mainly determined by the Cl2:NH4-N mass ratio and the pH. For pH
values typically encountered in drinking water treatment plants (pH 6.5-8.5), the amount of
dichloramine formed will be minimal (pH above 8.0). While more dichloramine may be formed
at slightly more acidic pH values, the NDMA formation would also be limited due to the lack of
non-protonated amines. Therefore, it is more important to properly control the Cl2:NH4-N mass
ratio to be close to but less than 5:1. As well, preformed monochloramine is rarely used in
practice and chloramine is usually formed on-site by the sequential addition of free chlorine and
ammonia. Adding ammonia following free chlorine has been recommended because this order of
reagent addition has been found to minimize dichloramine formation by reducing localized
Cl2:NH4-N mass ratios upon ammonia addition (Schreiber and Mitch, 2005). Similarly,
improving mixing upon ammonia addition may also help minimize dichloramine formation.
However, the short free chlorine contact time might enhance some initial NDMA formation by
breaking some NOM-amine complexes. Therefore, it is necessary to control the Cl2:NH4-N mass
ratio, the order of reagent addition, as well as the free chlorine contact time.
Once the treated water leaves the water treatment plant, it is hard to control what happens in the
distribution system, and since there is a delayed reaction due to the presence of NOM, the
concentration of NDMA in the distribution system could be much higher than the level measured
at the water treatment plant. Moreover, utilities may boost the water pH to 8.0-8.5 in the
distribution system for the purpose of corrosion control. As demonstrated in this study, despite
observing an overall reduced NDMA formation at higher pH, the speed at which the formation
occurred increased consistently with increasing pH. Therefore, elevating the pH in the
172
distribution system may help reduce the ultimate NDMA formation at the farther ends of the
distribution system, but it may result in more areas with shorter water age to experience
moderate NDMA levels due to the enhanced initial reaction. Even for the areas with longer water
age, enhanced NDMA formation might still occur due to the disproportionation of
monochloramine to form dichloramine over time. Therefore, the optimum pH for the respective
distribution system should take into consideration the size of the distribution system along with
the locations of consumers and their associated water age.
In summary, in order to develop a proper NDMA formation control strategy, it is important to
know the characteristics of the precursors and the source water, the treatment applied at the
plant, the size of the distribution system and the consumers’ locations within it. Preliminary tests
are required for specific water systems using their respective water sources. At the same time, it
is necessary to consider the risk tradeoffs between NDMA and other DBPs, to maintain the
distribution system stability, and most importantly to ensure that the primary purpose for the
disinfection is met – to provide protection from disease-causing organisms or “pathogens”.
8.2 Conclusions
The specific research objectives outlined in Section 1.2 were addressed in this work and led to
the following main conclusions:
1. Amine-based PPCPs are potential nitrosamine precursors. Molar yields higher than 1 %
were observed for eight of the selected pharmaceuticals under practical chloramine
disinfection conditions, with ranitidine showing the strongest potential to form NDMA.
The NDMA formation increased with the Cl2:NH4-N mass ratio, suggesting that
dichloramine is relevant to the formation of NDMA from these amine-based
pharmaceuticals. The NDMA molar conversion was relatively independent of the initial
pharmaceutical concentration at the range tested in this work, and only a slight
antagonistic effect was observed for the NDMA formation from a mixture of
pharmaceuticals. It is therefore expected that a mixture of amine-based PPCPs at their
environmental concentration levels could still form a substantial amount of nitrosamines.
173
2. NDMA formation from the tested amine-based pharmaceuticals follows a general three-
phase formation curve: an initial lag phase, followed by a fast increase in NDMA
formation, and eventually a plateau that represents the ultimate NDMA molar conversion.
A three-parameter kinetic model has been proposed to describe the NDMA formation
kinetics, and the model accurately reflects all three phases of the NDMA formation curve.
The model has been verified for four pharmaceuticals in five water matrices under
different pH conditions upon chloramination, with and without the application of
prechlorination. The presence of NOM has a profound impact on the NDMA conversion
rates from these pharmaceuticals; it slows down the initial reaction, most likely via the
formation of NOM-pharmaceutical complexes, and has less impact on the ultimate
NDMA molar conversion.
3. The pH affects both the ultimate NDMA conversion from the tested pharmaceuticals and
the reaction rate. The maximum NDMA formation occurs at a medium pH range, most
likely within the range of 7 to 8; but the reaction rate consistently increases as pH
increases. At lower pH values, the NDMA formation is inhibited because of the lack of
non-protonated amines, and also because the reaction to form NDMA involves a release
of H+ and thus is inhibited at acidic pH. At higher pH values, the initial reaction is
enhanced due to the increasing amount of non-protonated amines, but the ultimate
NDMA conversion is limited because of the lack of dichloramine.
4. The impact of prechlorination on NDMA formation from these pharmaceuticals is
significantly affected by the presence of NOM. Generally, prechlorination affects both
the ultimate NDMA conversion and the reaction rate by altering the property of
pharmaceuticals and NOM, as well as the binding potential between the two. In theory,
prechlorination could reduce the NDMA formation upon subsequent chloramination if it
reduces the electron density on the precursor amine groups, and vice versa. With the
presence of NOM, prechlorination at shorter contact times may enhance the initial
NDMA formation by destroying the NOM-pharmaceutical complexes, but prolonged
prechlorination may further inhibit the reaction due to the formation of bonds between the
pharmaceuticals and NOM breakdown products.
174
8.3 Recommendations for Future Research
It is anticipated that drinking water regulations for nitrosamines will be decided in the next three
to five years throughout North America, so utilities that have nitrosamine concerns will need
guidance to help them comply with these new regulations. Currently, research is taking place to
develop a standard bench-scale test to predict the nitrosamine formation for utilities, but what
happens next if the bench-scale test indicates high nitrosamine-FP for their source water? As
each plant uses different source water and applies different treatment technologies, it is important
to know as much as possible about nitrosamine formation, and more research is needed to fill
gaps ranging from precursor properties to the critical factors that may affect formation.
Although this work has demonstrated that amine-based PPCPs can form the corresponding
nitrosamines upon chloramination (Chapter 4), it is not likely that they will contribute to the
majority of nitrosamine precursors in drinking water due to their low concentrations in the
environment, unless the source water is heavily affected by a treated wastewater effluent or
contains pharmaceuticals with a very strong NDMA formation potential such as ranitidine.
However, the fact that the PPCPs are subjected to transformations is not considered in the
monitoring of these compounds, even though it is likely that substantial amounts of the
transformation products can also contribute to nitrosamine formation as long as they still contain
the amine group. As such, much work is needed in the field of analytical chemistry to identify
and quantify the transformation products of PPCPs. Moreover, the present work has
demonstrated some structural similarities among the pharmaceuticals with high NDMA-FPs, and
correlated the nitrosamine conversion with certain quantum properties of the molecules (Chapter
4). Due to the limited number of PPCPs studied, a more definite quantitative correlation between
the molecular structure and the nitrosamine molar conversion could not be established; however,
it indicates the potential application of quantitative structure-activity relationship (QSAR)
models to predict NDMA conversions based on the molecular properties. A large dataset is
required to run the QSAR model in order to acquire a meaningful relationship, therefore it is
necessary to test a variety of organic compounds containing amine groups. It is noted that most
QSAR models use molecular descriptors that describe the average property of the entire
molecule, while in this case descriptors would be needed for specific atoms in the molecule. For
example, the electron density or electrostatic potential of the N-atom on the amine group will
175
likely determine the NDMA conversion; the electrostatic property of the atoms adjacent to the
amine group will also affect the NDMA conversion, as suggested in this work.
Results from this work (Chapter 5) suggest that the ultimate NDMA conversion was mainly
determined by the pharmaceutical properties, but the reaction kinetics (i.e., the initial lag phase
and the pseudo-first order rate constant) were significantly influenced by the presence of NOM
which delayed the NDMA formation most likely via the formation of NOM-pharmaceutical
complexes. As per the binding theory, the impact of NOM on the reaction kinetics is expected to
be determined by the surface property of NOM (e.g., the charge distribution, the steric structure)
as well as the amount of specific functional groups that are responsible for the binding. However,
the bulk water quality measurements that are commonly used to describe NOM (i.e., TOC and
SUVA) and the LC-OCD analysis (which separates NOM based on the molecular size) do not
provide such information. In fact, the NOM-pharmaceutical complexation theory has been
suggested to explain the enhanced pharmaceutical removal during other treatment processes such
as coagulation and membrane filtration; however, there is no direct evidence to prove this
binding in aqueous phase due to the analytical challenge to measure the trace-level
pharmaceuticals and the limited information on the structure and properties of NOM. It would be
of great value to establish a comprehensive method to better characterize NOM, to identify the
moieties responsible for the complexation, and possibly to quantify the “binding potential”. For
example, zeta-potential measures the charge character of colloids and particles. It has been
linked to coagulation/flocculation performance (Sharp et al., 2005, 2006) where NOM is
removed via a charge-related process, the formation of flocs between negatively charged NOM
and positively charged coagulants. Similarly, zeta-potential measurements might be related to the
binding between negatively charged NOM and positively charged amines. In addition, future
research should also consider the potential effects of inorganic components such as cations on
the NOM-pharmaceutical association and their effects on NDMA formation. This information
would also be useful for the optimization of other treatment processes prior to disinfection, such
as coagulation, to maximize the removal of DBP precursors along with NOM.
This work has also investigated the role of pH in the NDMA formation from pharmaceuticals
(Chapter 6), but the conclusions were drawn based on two compounds in lab-grade water upon
176
preformed monochloramine disinfection. Further study is needed to investigate the pH effect in
different water matrices and upon different disinfection strategies With the presence of NOM
and application of prechlorination, the impact of pH will become more complicated. For
example, pH not only determines the chloramine speciation but also affects free chlorine
speciation. Since HOCl is usually more reactive than OCl-, pH can affect the effectiveness of
prechlorination and thus the subsequently precursor reactivity towards chloramine to form
NDMA. Moreover, pH determines the amine speciation and thus alters the degree of electrostatic
attraction between NOM and the amine groups.
Preoxidation has been reported to reduce the NDMA formation from the model precursor DMA
and NOM. Results from this work (Chapter 7) have suggested that prechlorination can either
increase or decrease NDMA formation from amine-based pharmaceuticals depending on the
property of the compound, and the effectiveness is influenced by the matrix components. As
demonstrated in this work, for those pharmaceuticals that the NDMA conversion was reduced
upon prechlorination, the NDMA-FP was only partly reduced, suggesting that the chlorination
transformation products can still contribute to the formation of NDMA; however, the NDMA
formation was significantly inhibited in the presence of NOM possibly due to binding between
the pharmaceuticals and the NOM breakdown products (the case of ranitidine in the Otonabee
River water). In the case of sumatriptan of which the NDMA conversion increased upon
prechlorination in the absence of NOM, the NDMA formation was reduced in the presence of
NOM upon prolonger prechlorination. Chlorination usually would cause a small modification on
the parent compound rather than a complete oxidation, but there are other technologies that are
currently used in water treatment plant as primary disinfection such as UV, ozone, and/or
advanced oxidation processes (AOPs; e.g., UV/H2O2, O3/H2O2). Most of the PPCPs selected in
this work have structures sensitive to UV (i.e., conjugated systems) and ozone (i.e., unsaturated
bonds and aromatic rings substituted by electron donating groups), and AOPs generate hydroxyl
radicals that are strong, non-selective oxidants. Therefore, it is necessary to further investigate
the impact of these stronger preoxidants on the nitrosamine formation from amine-based
precursors, taking into consideration the possible interference from NOM. In addition, different
disinfectants would cause the formation of different conventional or emerging DBPs, thus the
tradeoffs between the formation of nitrosamines and other DBPs need to be evaluated as well.
177
8.4 References
Ballard, B.D., Mackay, A.A., 2005. Estimating the removal of anthropogenic organic chemicals
from raw drinking water by coagulation flocculation. J. Environ. Eng. 131, 108-118
de Ridder, D.J., Verliefde, A.R.D., Heijman, S.G.J., Verberk, Q.J.C., Rietveld, L.C., van der Aa,
L.T.J., Amy, G.L., van Dijk, J.C., 2011. Influence of natural organic matter on equilibrium
adsorption of neutral and charged pharmaceuticals onto activated carbon. Wat. Sci. Technol. 63,
416 – 423
Diemert, S.A., 2012. The impact of coagulation on endocrine disrupting compounds,
pharmaceutically active compounds and natural organic matter. Master’s thesis, Univesity of
Toronto, Ontario
Lee, C., Schimidt, C., Yoon, J., von Gunten, U., 2007. Oxidation of N-nitrosodimethylamine
(NDMA) precursors with ozone and chlorine dioxide: kinetics and effect on NDMA formation
potential. Environ. Sci. Technol. 41, 2056-2063
Mitch, W.A., Schreiber, I.M., 2008. Degradation of tertiary alkylamines during
chlorination/chloramination: implications for formation of aldehydes, nitriles, halonitroalkanes,
and nitrosamines. Environ. Sci. Technol. 42, 4811-4817
Mitch, W.A., Sedlak, D.L., 2004. Characterization and fate of N-nitrosodimethylamine
precursors in municipal wastewater treatment plants. Environ. Sci. Technol. 38, 1445-1454
Mitch, W.A., Sharp, J.O., Trussell, R.R., Valentine, R.L., Alvarez-Cohen, L., Sedlak, D.L., 2003.
N-Nitrosodimethylamine (NDMA) as a Drinking Water Contaminant: A Review. Environ. Eng.
Sci. 20, 389-404
Parris, G.E., 1980. Covalent binding of aromatic amines to humates. 1. reactions with carbonyls
and quinones. Environ. Sci. Technol. 14, 1099-1106
Radjenovic, J., Farre, M.J., Gernjak, W., 2012. Effect of UV and UV/H2O2 in the presence of
chloramines on NDMA formation potential of tramadol. Environ. Sci. Technol. 46, 8356-8364
178
Schreiber, I.M., Mitch, W.A., 2005. Influence of the order of reagent addition on NDMA
formation during chloramination. Environ. Sci. Technol. 39, 3811-3818
Sharp, E.L., Banks, J., Billica, J.A., Gertig, K.R., Henderson, R., Parsons, S.A., Wilson, D.,
Jefferson, B., 2005. Application of zeta potential measurements for coagulation control: pilot-
plant experiences from UK and US waters with elevated organics. Wa. Sci. Technol. 5 (5), 49-56
Sharp, E.L., Jarvis, P., Parsons, S.A., Jefferson, B., 2006. The impact of zeta potential o nthe
physical properties of ferric-NOM flocs. Environ. Sci. Technol. 40, 3934-3940
Stackelberg, P.E., Gibs, J., Furlong, E.T., Meyer, M.T., Zaugg, S.D., Lippincott, R.L., 2007.
Efficiency of conventional drinking-water-treatment processes in removal of pharmaceuticals
and other organic compounds. Sci. Total Environ. 377, 255-272
Summers, R.S., Hooper, S.M., Shukairy H.M., Solarik, G., Owen, D., 1996. Assessing DBP
yield: uniform formation conditions. J. Am. Water Works Assoc. 88, 80-93
Thorn, K.A., Pettigrew, P.J., Goldenberg, W.S., Weber, E.J., 1996. Covalent binding of aniline
to humic substances. 2. 15N NMR studies of nucleophilic addition reactions. Environ. Sci.
Technol. 30, 2764-2775
Vieno, N., Tuhkanen, T., Kronberg, L., 2006. Removal of pharmaceuticals in drinking water
treatment: effect of chemical coagulation. Environ. Technol. 27, 183 – 192
Weber, E.J., Spidle, D.L., Thorn, K.A., 1996. Covalent binding of aniline to humic substances. 1.
kinetic studies. Environ. Sci. Technol. 30, 2755-2763
Westerhoff, P., Yoon, Y., Snyder, S., Wert E., 2005. Fate of endocrine-disruptor,
pharmaceutical, and personal care product chemicals during simulated drinking water treatment
processes. Environ. Sci. Technol. 39, 6649-6663
179
Appendix 1. Nitrosamine Analysis: QA/QC (Chapter 3)
Quantification of nitrosamines was attained through isotope dilution using a 6-level calibration
curve. Deuterated NDMA (d6-NDMA, 50 ng/L) was used as the internal standard for both
NDMA and NDEA, and was added into all the calibration standards and water samples prior to
extraction. The calibration standards were subjected to the same extraction process as water
samples in order to account for the recovery. A typical calibration curve for NDMA (0 – 1000
ng/L) is provided in Table A1.1 and Figure A1.1.
Table A1.1 Typical calibration standards for NDMA
Calibration level
Concentration (ng/L)
Area count Area ratio (NDMA/ d6-NDMA) 50 ng/L d6-NDMA NDMA
1 0 6.94E+05 6.68E+04 0.1 2 50 5.86E+05 5.88E+05 1.0 3 100 5.35E+05 1.13E+06 2.1 4 200 6.67E+05 2.48E+06 3.7 5 500 6.60E+05 6.22E+06 9.4 6 1000 7.45E+05 1.43E+07 19.2
Figure A1.1. Typical calibration curve for NDMA
180
A calibration curve was prepared fresh together with each set of experiments using a given
designated water matrix sample. The use of internal standard was meant to minimize impacts
from matrix effects. As compared in Table A1.2, the slope of the NDMA calibration curve was
generally a constant (0.02) despite being prepared in different water matrices. A larger variance
in the slope was observed for NDEA, mainly because no deuterated NDEA standard was used
and d6-NDMA was used as the internal standard for NDEA. The slope of all the calibration
curves for NDMA during the course of this work is summarized in Figure A1.2. NDEA was only
measured in a limited number of samples and thus is not included in the figure.
Table A1.2. Average slope of the nitrosamine calibration curve prepared in different water matrices
Compound Slope MQ Tap Lake River (raw) River (treated)
NDMA Mean 0.019 0.018 0.019 0.018 0.019 STD 0.002 0.002 0.001 0.002 0.0004 RSD 11.4% 9.2% 5.6% 9.8% 1.9%
NDEA Mean 0.014 0.013 0.025 NA* NA STD 0.002 NA 0.005 NA NA RSD 13.2% NA 21.1% NA NA
* Not available
Figure A1.2. Slope of the NDMA calibration curve prepared in different water matrices
181
In addition, each calibration curve can be treated as a standard addition plot to determine the
background nitrosamine level in the water matrix that is used to prepare the calibration curve.
Using the sample calibration curve in Figure A1.1 as an example, the background NDMA
concentration in this water matrix should be the absolute value of the x-intercept, which is the y-
intercept/slope = 0.0438/0.0191 = 2.3 (ng/L). The background NDMA level in all the selected
matrices during the course of this work is summarized in Figure A1.3. Generally, there was no
NDMA detected in all the lake, river, and treated river water samples, which was expected since
they were not in contact with any chloramines. Very low concentrations (below 3 ng/L) were
detected occasionally in the tap water. The NDMA level in MQ water was usually low (below 3
ng/L) when the machine worked properly. However, when the ion-exchange cartridge was close
to its life span or the UV lamp was malfunctioning, spikes of NDMA were observed in the
“pure” MQ water. Further details about the source of NDMA precursors in the MQ machine and
strategies to cope with the NDMA contamination in MQ water are discussed in Appendix 3. The
background NDEA level in all the selected matrices was in general below the detection limit.
Figure A1.3. Background NDMA level in the selected water matrices
182
Given that quantification was attained via isotope dilution, the extraction recovery was already
auto-corrected. Thus, a control chart was not prepared in the format of a typical “spike and
recovery” check. Instead, we took the area ratio of the 100 ng/L (C spiked) sample from each
calibration curve, recalculated the concentration of this sample (C measured) using the slope and y-
intercept of the respective calibration curve, and calculated the “recovery” as “C measured / C
spiked”. Using the sample calibration curve in Table A1.1 as an example, the area ratio of the 100
ng/L sample is 2.1, thus C measured of the sample is calculated using the slope (0.0191) and y-
intercept (0.0438) as (2.1 – 0.0438)/0.0191 = 108 (ng/L), and the “recovery” is calculated as C
measured / C spiked = 108/100 × 100 % = 108 %. The control chart for NDMA during the course of
this work is summarized in Figure A1.4. The dotted lines at 70 % and 130 % represent
acceptable limits as reported by the USEPA (1986). The recovery for 100 ng/L NDEA sample
varied between 70 % and 140 % and was similarly acceptable.
Figure A1.4. GC-MS method control chart for 100 ng/L NDMA
Reference
USEPA, 1986. Definition and procedure for the determination of the method detection limit. 40
CFR Part 136, Appendix B, revision 1.11, updated on June 30th, 1986.
183
Appendix 2. Nitrosamine Formation from Water Matrices (NOM) (Chapter 3)
In this work, PPCPs were dosed into selected water matrices to determine their nitrosamine-FPs;
however, NOM in the waters is also a potential nitrosamine precursor. The nitrosamines formed
from the water matrix itself were accounted for using blank control samples (i.e., the respective
water matrix without dosing any PPCPs). Since the NDEA formed from water matrices selected
for this study was generally below the detection limit, the following discussion will focus on
NDMA only.
Along with every batch of NDMA-FP experiments (i.e., NDMA-FP at 24 hr), the blank samples
were prepared in triplicate. Figure A2.1 are the normal probability plots for the NDMA-FP (24
hr) of the water matrix components under the SDS conditions during the course of this work.
Figure A2.1. Normal probability plots: 24 hr NDMA-FP of the selected water matrices (SDS conditions;
error bars represent the variability due to multiple formation potential tests (n = 3))
Along with all the NDMA formation kinetic experiments, one blank sample was prepared at each
time point. The chloramine decay in each matrix as well as the NDMA formation from water
matrix itself (NOM) was monitored. The chloramine decay in general followed pseudo-first
order kinetics (Figure A2.2). Tap water contained residual and was already “saturated” with
chloramine, thus the chloramine decay rate (i.e., the slope) in tap water was similar with that in
MQ water. Lake water and treated river water also had comparable chloramine decay rates,
184
possibly because of similar Cl2:TOC ratio applied (around 1.1 for both matrices). And despite the
different chloramine dosage applied in the two batches of river water, the chloramine decay rates
in this matrix were similar.
Figure A2.2. Chloramine decay in different water matrices (pH = 7.0 ± 0.1; numbers in the brackets are
the applied chloramine dosage in each matrix.)
The NDMA formation kinetics from selected water matrices (NOM) upon chloramination
(preformed) are compared in Figure A2.3. For each set of experiments, only one blank sample
was prepared at each time point, therefore the error bars in the figure represent the standard
deviation from multiple batches (n = 2 for the treated river water, and n ≥ 5 for the other two
matrices). The data were fitted with the negative exponential equation (𝑌 = 𝑈(1 − 𝑒���)), where
Y is the NDMA formed (ng/L) at given reaction time (t); U (ng/L) is the ultimate NDMA-FP of
the respective water matrix; k is the pseudo-first order reaction rate constant. The estimated
equation parameters are summarized in Table A2.1, along with the relevant reaction conditions
and water quality. From the limited data obtained in this work, the NDMA formation from NOM
could not be directly associated with TOC or any specific NOM fractions (Table A2.1). Because
the concentration of NDMA formed is so small (ng/L), it was difficult to correlate it
quantitatively with TOC measurements that were several orders of magnitude higher.
185
Figure A2.3. NDMA formation kinetics from water matrices upon chloramine disinfection (SDS
conditions; error bars represent the variability due to multiple formation potential tests (n = 2 for the treated river water and n ≥ 5 for the other two matrices))
Table A2.1. The NDMA-FP of water matrices (NOM) vs. reaction conditions and water quality Matrix Lake River Treated river Ultimate NDMA-FP (U; ng/L) 8.2 (2.3) a 9.3 (2.0) a 20.1 (1.9) a
Pseudo-first order rate constant (k; hr-1) 0.05 (0.03) a 0.04 (0.03) a 0.03 (0.01) a Cl2 (mg/L) /TOC (mg/L) 1.1 0.6 1.1
NDMA/TOC (ng/mg) 2.6 1.5 5.9
LC-OCD NOM fractions (mg/L) (pH adjusted to 7.0 ± 0.1)
Biopolymers 0.34 0.26 0.18 Humic substances 0.90 3.39 1.44 Building blocks 0.46 0.93 0.74 LMW neutrals 0.19 0.50 0.46 LMW acids 0.10 0.17 0.17 Biopolymer-N b 0.03 0.03 0.04 Humics-N b 0.06 0.19 0.12
a Numbers in the brackets are the 95 % confidence intervals. b The LC-OCD is also equipped with an organic nitrogen detector (OND).
Interestingly, the NDMA-FP of treated river water was the highest among the three matrices,
even though it had a medium TOC level. Compared with the raw river water, it was expected
that some of the precursors would be removed during the treatment process and thus reduce the
NDMA-FP; yet the opposite trend was observed. Some polymers used in water treatment plants
could contribute to NDMA formation, but no such chemicals are used at the plant where the
samples were collected. One possible explanation would be the higher Cl2:TOC ratio applied in
the treated river water, which led to faster chloramine consumption (confirmed in Figure A2.2).
186
It could also be the treatment process at the water treatment plant (i.e.,
coagulation/flocculation/filtration) that changed the NOM surface characteristics (e.g., steric
structure), making it more prone to chloramine and thus forming more NDMA.
In addition, the NDMA formation kinetics from NOM was also monitored upon sequential
chlorine + chloramine disinfection in Lake Ontario water and Otonabee River water (Figure
A2.4. The data were also fitted with the negative exponential equation, and the estimated
equation parameters are compared in Figure A2.5. In general, the application of prechlorination
reduced the NDMA formation from NOM, and the ultimate NDMA-FP tended to decrease as
prechlorination contact time increased. This is in good agreement with literature.
Figure A2.4. NDMA formation kinetics from water matrices upon sequential chlorine + chloramine
disinfection (SDS; error bars represent standard deviation from two batches of tests (n = 2))
Figure A2.5. NDMA formation from water matrices (NOM) upon sequential chlorine + chloramine disinfection: the estimated model parameters (SDS; error bars are the 95 % confidence intervals.)
187
Appendix 3. Potential NDMA Contamination in MQ Water (Chapter 3)
The NDMA level in MQ water was usually low (below 3 ng/L) when the machine worked
properly. When the ion-exchange cartridge was close to its life span or the UV lamp was
malfunctioning, spike of NDMA was observed in the pure MQ water, but the NDMA level
would typically fall down to normal level after the replacement of the ion-exchange cartridge or
the UV lamp. However, a potential NDMA contamination in MQ water occurred in early 2012,
with elevating concentrations of NDMA (up to 200 ng/L, Figure A1.3) detected in pure MQ
water while the new cartridge had only been installed for less than three months (the average life
span of the ion-exchange cartridge in our lab is 10 to 12 months). Moreover, the NDMA level in
MQ water decreased over time as samples were collected (Figure A3.1), indicating that NDMA
might have been accumulated in the MQ machine and was being washed out over time.
Figure A3.1. NDMA concentration in MQ water as a function of sampling time
It was verified that the NDMA in the MQ water was not from the feed water. NDMA
concentrations in the feed water (distilled water) and in freshly made MQ water (paired
sampling) were determined to be 3.1 and 28.9 ng/L, respectively, indicating that the NDMA was
from the MQ machine instead of the feed water. The most likely source of the NDMA precursors
in MQ machine is the ion-exchange resin which may release amine-based impurities; whereas
188
the feed water (distilled water) was made from tap water and thus contained some chlorine
residual (in the form of chloramine since Toronto tap water is chloraminated). Therefore, when
the machine was off, NDMA might be formed and accumulated in the cartridge. The level of
NDMA in pure MQ water varied from day to day, depending on the level of chloramine residual
in the feed water (which also varies due to different storage times in the reservoir; the chloramine
residual may dissipate over time, while fresh distilled water may carry a high residual), as well as
the usage of MQ water (i.e., the accumulated NDMA might be already washed off in the first
several batches of water collection, thus the NDMA level in later collected samples might be
lower). Flushing the system before sample collection might help; however, it was too wasteful to
let the machine flush for 30 min to reach an acceptably low background NDMA level (in the
case of Figure A3.1).
Because of this potential contamination, experiments in MQ water can be challenging if the
background NDMA level is different in each sample, making the blank-subtraction meaningless.
Therefore, when MQ water was needed for the NDMA formation experiments, the water was
always collected beforehand in a large container to equalize the background NDMA level, and
then distributed into 1 L amber bottles for further experiments. This procedure ensured an
identical background NDMA concentration in all the blanks and samples, thus a more reliable
blank subtraction.
189
Appendix 4. Matrix Effect on the 24 hr NDMA-FP from Selected Pharmaceuticals – Statistical Analysis (Chapter 4)
Paired t-test
Paired t-tests can be used to compare two sets of data when the data points are “paired” between
the two sets. The test determines whether the mean of differences between each pair of
observations is 0, and thus requires equal number of data points in each data set for the
comparison. If 𝐷� represents the mean of differences between every two observations, the
hypotheses are:
𝐻�:𝐷� = 0; 𝐻�:𝐷� ≠ 0.
The test statistic is t with (n-1) degree of freedom. If the p-value associated with t is smaller than
0.05, there is evidence to reject the null hypothesis and conclude that the two sets of data are not
equivalent.
In order to compare the NDMA-FP among different water matrices, all replicates of each
compound were paired for each pair of matrices, and the paired t-test was performed to
determine whether the two sets of data were significantly different. In addition, the test can
estimate the 95 % confidence interval for the mean of differences, which can tell whether one set
of data is consistently larger than the other. An example for comparing two matrices is provided
in Table A4.1; the same analysis was performed between each pair of matrices.
Table A4.1. Calculation example for the paired t-test Compound Matrix 1 Matrix 2 Difference
A a11 a21 a11 – a21 a12 a22 a12 – a22 a13 a23 a13 – a23
B b11 b21 b11 – b21 b12 b22 b12 – b22 b13 b23 b13 – b23
… … …
H h11 h21 h11 – h21 h12 h22 h12 – h22 h13 h23 h13 – h23
Mean 𝐷�
190
The paired t-test results are summarized in Table A4.2. The results suggested that the NDMA-FP
at 24 hr for selected pharmaceuticals generally followed the order of MQ ≈ TAP > Lake > River.
Table A4.2. Matrix effect: summary of the paired t-test results (95 % confidence level)
Matrix p-value Significant difference?
(p < 0.05) Notes
1 nM 5 nM 25 nM
MQ vs. Tap NA* 0.592 0.505 NO MQ ≈ TAP
MQ vs. Lake NA 0.009 0.029 YES MQ > Lake
MQ vs. River NA 0.008 0.017 YES MQ > River
Tap vs. Lake 0.020 0.017 0.044 YES TAP > Lake
Tap vs. River 0.028 0.011 0.023 YES TAP > River
Lake vs. River 0.036 0.008 0.016 YES Lake > River
* Not available
One-way ANOVA and Tukey’s HSD test
A one-way ANOVA test was performed for each pharmaceutical, followed by a Tukey’s HSD
(honestly significant difference) test. The one-way ANOVA (one variable: water matrix) test is
to determine if the matrix effect was significant, while the Tukey’s HSD test is to determine how
significant the difference was among the different matrices. All of the analyses were performed
using the GraphPad Prism® ANOVA tool.
Table A4.3. Matrix effect: summary of the one-way ANOVA and Tukey’s HSD results (95 % confidence level)
(25 nM) One-way ANOVA Tukey’s multiple comparison (Significant? b)
p-value Significant? a MQ vs. Tap
MQ vs. Lake
MQ vs. River
Tap vs. Lake
Tap vs. River
Lake vs. River
Ranitidine < 0.0001 Yes Yes Yes Yes Yes Yes Yes Doxylamine < 0.0001 Yes Yes No Yes Yes Yes Yes Sumatriptan < 0.0001 Yes Yes Yes No Yes Yes No Chlorphenamine < 0.0001 Yes Yes No Yes Yes Yes Yes Nizatidine < 0.0001 Yes No Yes Yes No Yes Yes Diltiazem < 0.0001 Yes No Yes Yes Yes Yes Yes Carbinoxamine < 0.0001 Yes Yes No Yes Yes Yes Yes Tetracycline < 0.0001 Yes Yes No Yes Yes Yes Yes
a. The means (NDMA-FPs in different water matrices) are significantly different if p < 0.05; b. The difference between two means (NDMA-FP in every two matrices) is significantly different if p < 0.05.
191
Appendix 5. Pharmaceutical Concentration Effect on the 24 hr NDMA-FP from Selected Pharmaceuticals – Statistical Analysis (Chapter 4)
One-way ANOVA and Tukey’s HSD test
A one-way ANOVA test was performed for each pharmaceutical in every water matrix to
determine if the initial pharmaceutical concentration effect was significant; a Tukey’s HSD test
was followed to determine how significant the difference was among the different
pharmaceutical concentrations. All the analysis was performed using the GraphPad Prism®
ANOVA tool.
An example is shown in Table A5.1 for the analysis of data obtained in the tap water (four
concentration levels tested). The same analysis was performed in the other three water matrices
as well, and summarized in Table A5.2. Noted that only two concentration levels were tested in
MQ water (5 nM and 25 nM), and three concentration levels were tested in the lake and river
water (1 nM, 5 nM, and 25 nM).
Table A5.1. Pharmaceutical concentration effect: summary of the one-way ANOVA and Tukey’s HSD results from data obtained in the tap water (95 % confidence level)
Tap water One-way ANOVA Tukey’s multiple comparison (Significant? b)
p-value Significant? a 1 nM vs
5 nM 1 nM vs 25 nM
1 nM vs 50 nM
5 nM vs 25 nM
5 nM vs 50 nM
25 nM vs 50 nM
Ranitidine 0.0820 No No No No No No No Doxylamine 0.0011 Yes No Yes Yes No Yes No
Sumatriptan 0.0543 No No No Yes No No No Chlorphenamine 0.0869 No No No No No No No Nizatidine 0.0211 Yes No No Yes No Yes No Diltiazem 0.004 Yes No Yes No Yes No No Carbinoxamine 0.0006 Yes No No Yes No Yes Yes Tetracycline 0.0001 Yes Yes Yes Yes No Yes Yes
a. The means (NDMA-FPs at different pharmaceutical concentrations) are significantly different if p < 0.05; b. The difference between two means (NDMA-FPs at every two initial pharmaceutical concentrations) is
significantly different if p < 0.05.
192
Table A5.2. Summary of the one-way ANOVA results for individual pharmaceuticals in various water matrices: If the pharmaceutical concentration effect is significant? (95 % confidence level)
MQ Tap Lake River Ranitidine No No No Yes Doxylamine Yes Yes Yes Yes Sumatriptan Yes No Yes Yes Chlorphenamine Yes No Yes Yes Nizatidine Yes Yes Yes No Diltiazem No Yes No No Carbinoxamine No Yes No Yes Tetracycline No Yes No Yes
Paired t-test
This paired t-test compared two columns of data to determine if they were significantly different.
In this case, every column included the NDMA-FP (molar conversion) at one concentration in
one matrix from eight selected pharmaceuticals (eight values in total). For each matrix, the
number of columns was equal to the number of concentration levels tested. A calculation
example is presented in Table A5.3 for the data obtained in the tap water. The same analyses
were performed in the other three water matrices using the available data, and summarized in
Table A5.4.
Table A5.3. Calculation example of the paired t-test for data obtained in the tap water (95 % confidence level)
NDMA molar conversion
1 nM 5 nM 25 nM 50 nM
Paired t-test p-value Significant
difference? (p < 0.05) Ranitidine 83.2% 89.0% 86.7% 76.9% 1 nM vs 5 nM 0.3647 No Doxylamine 10.3% 9.2% 7.3% 6.0% 1 nM vs 25 nM 0.9806 No Sumatriptan 2.2% 3.0% 2.8% 3.5% 1 nM vs 50 nM 0.1537 No Chlorphenamine 1.0% 1.3% 1.5% 1.8% 5 nM vs 25 nM 0.0510 No Nizatidine 5.8% 6.0% 5.4% 4.5% 5 nM vs 50 nM 0.1819 No Diltiazem 2.4% 2.6% 1.8% 2.1% 25 nM vs 50 nM 0.2734 No Carbinoxamine 1.3% 1.4% 1.1% 0.7% Tetracycline 1.4% 1.0% 0.9% 0.5%
193
Table A5.4. Summary of the paired t-test results in various water matrices: If the NDMA-FPs at different initial pharmaceutical concentrations are significantly different? (95 % confidence level)
* Not available
Although the one-way ANOVA results varied for individual pharmaceuticals in different
matrices, the paired t-test results indicate that, in general, there was no significant difference in
the NDMA-FPs from selected pharmaceuticals in a given matrix at different initial
pharmaceutical concentrations.
MQ Tap Lake River
1 nM vs 5 nM NA* No No No 1 nM vs 25 nM NA No No No 1 nM vs 50 nM NA No NA NA 5 nM vs 25 nM No No No No 5 nM vs 50 nM NA No NA NA 25 nM vs 50 nM NA No NA NA
194
Appendix 6. Pharmaceutical Mixture Effect on the 24 hr NDMA-FP from Selected Pharmaceuticals – Statistical Analysis (Chapter 4)
The student’s t-test was performed to determine if two means (i.e., the NDMA formed from sum
of single pharmaceuticals vs. the NDMA formed from a mixture of pharmaceuticals) were
significantly different. The results are summarized in Table A6.1. A calculation example is
provided below as well.
Table A6.1. Pharmaceutical mixture effect: summary of the t-test results (95 % confidence level)
Matrix Pharmaceutical Concentration (nM)
NDMA-FP (nM) a Reduction
t-test Sum of Eight Eight-mixture p-value Significant? b
MQ 5 5.50 ± 0.02 4.91 ± 0.06 11% < 0.0001 Yes 25 25.80 ± 0.21 24.53 ± 0.37 5% 0.0067 Yes
Tap 5 5.55 ± 0.04 4.97 ± 0.38 10% 0.0582 No 25 29.85 ± 0.34 28.84 ± 1.02 3% 0.1791 No
Lake 1 0.90 ± 0.03 0.79 ± 0.03 13% 0.0109 Yes 5 3.72 ± 0.06 3.63 ± 0.32 2% 0.6571 No 25 22.47 ± 0.42 22.28 ± 0.65 1% 0.6925 No
River 1 0.49 ± 0.02 0.43 ± 0.04 12% 0.0808 No 5 2.94 ± 0.09 2.91 ± 0.11 1% 0.7332 No 25 12.29 ± 0.63 10.22 ± 0.50 17% 0.0112 Yes
Matrix Pharmaceutical Concentration (nM)
NDMA-FP (nM) a Reduction
t-test Sum of Seven Seven-mixture p-value Significant? b
MQ 5 1.24 ± 0.02 0.96 ± 0.02 23% < 0.0001 Yes 25 5.12 ± 0.04 4.70 ± 0.04 8% 0.0002 Yes
River 1 0.13 ± 0.004 0.04 ± 0.01 69% < 0.0001 Yes 5 0.37 ± 0.01 0.15 ± 0.01 60% < 0.0001 Yes 25 1.50 ± 0.01 0.98 ± 0.06 35% 0.0001 Yes
a. NDMA-FP: mean ± standard deviation (n = 3); b. The reduction is significant if p < 0.05.
T-test calculation example:
MQ, 5 nM, X�1=5.50, X�2=4.91, 𝑆� = 0.02, 𝑆� = 0.06, n = 3, df = 2n – 2 = 4;
𝑆���� = ���
(𝑆�� + 𝑆��) = ���
(0.02� + 0.06�) = 0.045 , 𝑡 = X�1�X�2
��������
= 5.50-4.91
�.������
= 16.16
The p-value was calculated using the function TDIST (t, df, tail) in Excel:
p-value = TDIST (16.16, 4, 2) = 8.6 × 10-5
195
Appendix 7. Impact of Cl2:NH4-N Mass Ratio on the 24 hr NDMA-FP from Selected Pharmaceuticals – Statistical Analysis (Chapter 4)
A one-way ANOVA test was performed to determine if the Cl2:NH4-N mass ratio significantly
affected the NDMA-FP from the pharmaceutical mixture; a Tukey’s HSD test was followed to
determine how significant the difference was among the different ratios. All of these analyses
were performed using the GraphPad Prism® ANOVA tool.
Table A7.1. Impact of Cl2:NH4-N mass ratio: summary of the one-way ANOVA and Tukey’s HSD results
5 nM PPCP-mixture 25 nM PPCP-mixture
One-way ANOVA p-value 0.0002 < 0.0001 Significant? a Yes Yes
Tukey’s multiple
comparison (Significant? b)
3:1 vs 4.2:1 No Yes 3:1 vs 5.1:1 Yes Yes 3:1 vs 6.3:1 Yes Yes 4.2:1 vs 5.1:1 No Yes 4.2:1 vs 6.3:1 Yes Yes 5.1:1 vs 6.3:1 Yes Yes
a The means (NDMA-FPs at different Cl2:NH4-N ratios) are significantly different if p < 0.05; b The difference between two means (NDMA-FPs between every two Cl2:NH4-N ratios) is significantly
different if p < 0.05.
196
Appendix 8. NDMA Formation Kinetics: Reproducibility (Chapter 5)
The reproducibility of the NDMA formation kinetics was checked by repeating the tests (for 25
nM ranitidine) in the same source water sampled on different dates, as shown in Figure A8.1.
Figure A8.1. NDMA formation kinetics from ranitidine: the reproducibility check (25 nM, SDS; error bars represent the variability due to multiple formation potential tests (n = 3))
The estimated kinetic model parameters are compared in Table A8.1. In Lake Ontario water, the
calculated θ for the June 2010 batch of tests was lower than that for the October 2011 batch of
tests. This is because the NDMA formation curve in the June 2010 test did not achieve a plateau
within the 24 hr experiment, and the model arbitrarily assumed the last point as the plateau,
resulting in an underestimation of the ultimate NDMA molar conversion. In Otonabee River
water, the ultimate NDMA molar conversion (θ) was comparable in both tests, but the
September 2010 test gave a shorter initial lag phase (Lag) and slightly faster reaction rate (k).
This might be explained by the higher chloramine dosage applied in the September 2010 test (4.4
± 0.3 mg/L) than in the January 2010 test (3.5 ± 0.2 mg/L) (see details in Section 3.3).
Previously, a similar trend was observed for the NDMA formation kinetics from ranitidine in
MQ water under the modified formation potential (MFP) and simulated distribution system
(SDS) conditions (Figure A8.2). The much higher chloramine dosage did not increase the
197
ultimate NDMA molar conversion, but it did enhance the reaction rate and reduce the initial lag
phase significantly (Table A8.2).
Table A8.1. Estimated kinetic model parameters for ranitidine: the reproducibility check in Lake Ontario water and Otonabee River water (25 nM, SDS)
Matrix Sampling date Parameter Estimation θ a Lag (hr) a k (hr-1) a R2
Lake Ontario June 2010 0.769 (0.018) 13.1 (0.4) 0.252 (0.028) 0.999 October 2011 0.824 (0.025) 12.5 (0.7) 0.182 (0.028) 0.993 Otonabee River September 2010 0.841 (0.044) 21.9 (1.4) 0.083 (0.021) 0.990 January 2012 0.863 (0.018) 31.0 (1.1) 0.068 (0.007) 0.998 a. Numbers in the brackets represent the 95 % confidence interval of each model parameter;
Figure A8.2. NDMA formation kinetics from ranitidine (25 nM, MQ): MFP vs. SDS conditions (Error
bars represent the variability due to multiple formation potential tests (n = 3))
Table A8.2. Estimated kinetic model parameters for ranitidine (25 nM, MQ): MFP vs. SDS conditions
Reaction conditions Parameter Estimation
θ a Lag (hr) a k (hr-1) a R2
MFP (NH2Cl = 28.4 mg/L) 0.892 (0.015) 1.3 (0.1) 1.218 (0.129) 0.996 SDS (NH2Cl = 2.5 mg/L) 0.906 (0.045) 4.6 (0.5) 0.313 (0.084) 0.991
a. Numbers in the brackets represent the 95 % confidence interval of each model parameter.
198
Appendix 9. Summary of the Estimated Kinetic Model Parameters under Different Treatment Conditions and the Overall Kinetic Model Verification (Chapter 5-7)
A three-parameter kinetic model has been proposed to describe the NDMA formation kinetics
from selected pharmaceuticals. The model is of the form:
𝑌 = 𝜃
1 + 10�×(�����)
Y: The NDMA molar conversion at given reaction time (t);
θ : The ultimate NDMA molar conversion;
Lag: The length of initial lag phase observed;
k: The pseudo-first order reaction rate constant.
In this work, the kinetic model has been applied to data obtained from different pharmaceuticals
in various water matrices under different treatment conditions. The scenarios for which results
will be summarized here include the testing of four pharmaceuticals in five water matrices upon
preformed chloramination (Chapter 5), ranitidine in MQ water upon preformed chloramination
under different pH values (Chapter 6), as well as ranitidine and sumatriptan in three water
matrices upon sequential chlorination and chloramination (Chapter 7). Some data are not
included in the chapters, such as those obtained from repeated tests under the same reaction
conditions, or from similar tests using different concentrations of the pharmaceuticals. Therefore,
a full summary of all of the estimated kinetic model parameters under different treatment
conditions is presented in this appendix. The acronyms used for the matrices include MQ (milli-
Q® water), TAP (Toronto tap water), LW (Lake Ontario water), RW (Otonabee River water), and
TRW (treated Otonabee River water prior to chlorination).
199
Table A9.1. Summary of the estimated kinetic model parameters and model verification: matrix effect (pH = 7.0 ± 0.1; preformed NH2Cl)
Compound Conc. (nM) Matrix Parameter Estimation Model Verification
θ a Lag (hr) a k (hr-1) a R2 Model-predicted conversion @ 24 hr
Measured conversion @ 24 hr b
Chlorphenamine 5 MQ 0.027 (0.003) 7.0 (1.3) 0.143 (0.050) 0.974 2.7% 2.9 % (0.2 %) TAP 0.023 (0.003) 10.8 (3.0) 0.065 (0.024) 0.953 2.0% 2.0 % (0.8 %) RW 0.037 (0.002) 38.4 (3.4) 0.039 (0.007) 0.990 0.8% 1.0 % (0.1 %) 25 MQ 0.033 (0.002) 8.7 (0.6) 0.234 (0.079) 0.994 3.3% 1.8 % (0.1 %) TAP 0.030 (0.002) 19.4 (1.8) 0.070 (0.015) 0.988 2.0% 1.5 % (0.1 %) RW 0.043 (0.002) 39.9 (1.9) 0.056 (0.009) 0.996 0.5% 0.5 % (0.02 %)
Doxylamine 5 MQ 0.062 (0.005) 16.6 (2.2) 0.068 (0.017) 0.975 4.7% 3.8 % (0.1 %) TAP 0.068 (0.006) 48.8 (5.3) 0.027 (0.005) 0.993 1.2% 2.5 % (0.2 %) RW 0.059 (0.003) 67.0 (4.3) 0.026 (0.006) 0.990 0.4% 1.1 % (0.03 %) 25 MQ 0.106 (0.007) 22.0 (2.0) 0.069 (0.018) 0.985 6.1% 4.2 % (0.1 %) TAP 0.092 (0.006) 46.1 (4.1) 0.030 (0.005) 0.992 1.6% 3.2 % (0.3 %) RW 0.060 (0.002) 51.1 (1.8) 0.051 (0.011) 0.996 0.2% 0.5 % (0.04 %)
Ranitidine 5 MQ 0.912 (0.045) 6.3 (0.5) 0.225 (0.043) 0.992 91.2% 85.2 % (0.8 %) TAP 0.902 (0.045) 6.7 (0.8) 0.169 (0.045) 0.986 90.1% 83.4 % (8.1 %) LW 0.729 (0.032) 13.3 (0.7) 0.251 (0.057) 0.997 72.7% 64.1 % (3.6 %) RW 0.822 (0.056) 20.8 (1.8) 0.086 (0.026) 0.984 53.6% 51.4 % (4.9 %) 25 MQ 0.906 (0.045) 4.6 (0.5) 0.313 (0.084) 0.991 90.6% 82.7 % (2.4 %) TAP 0.847 (0.039) 6.5 (0.7) 0.177 (0.043) 0.988 84.6% 88.4 % (5.9 %) LW + NH2Cl c 0.828 (0.029) 7.0 (0.4) 0.190 (0.030) 0.993 NA d NA LW 0.769 (0.018) 13.1 (0.4) 0.252 (0.028) 0.999 72.7 % 64.1 % (3.6 %) LW 0.824 (0.025) 12.5 (0.7) 0.182 (0.028) 0.993 81.7 % 70.1 % (4.8 %) TRW 0.895 (0.012) 9.2 (0.3) 0.250 (0.040) 0.998 NA NA RW 0.863 (0.018) 31.0 (1.1) 0.068 (0.007) 0.998 53.8 % 51.4 % (4.9 %) RW 0.841 (0.044) 21.9 (1.4) 0.083 (0.021) 0.990 50.4% 43.2 % (7.1 %)
Sumatriptan 5 MQ 0.023 (0.002) 22.8 (2.7) 0.080 (0.040) 0.966 1.3 % 1.3 % (0.2 %) 25 MQ 0.028 (0.001) 27.0 (2.4) 0.060 (0.010) 0.992 1.1 % 2.0 % (0.1 %) LW 0.022 (0.001) 31.8 (2.3) 0.060 (0.010) 0.991 0.6 % 1.6 % (0.1 %) TRW 0.025 (0.001) 32.5 (2.5) 0.040 (0.010) 0.988 NA NA RW 0.024 (0.002) 66.8 (4.0) 0.030 (0.010) 0.988 NA NA a Numbers in these brackets represent the 95 % confidence interval of each model parameter. b Numbers in these brackets represent the standard deviation from multiple tests (n = 3). c Pre-chloraminated lake water mimicked the tap water scenario. d Not available. Independent 24 hr NDMA-FP experiments were not performed.
200
Table A9.2. Summary of the estimated kinetic model parameters and model verification: the pH effect (preformed NH2Cl)
Compound Conc. (nM) Matrix Parameter Estimation Model Verification
pH θ a Lag (hr) a k (hr-1) a R2 Model-predicted conversion @ 72 hr
Measured conversion @ 72 hr b
Ranitidine 25 MQ 6.0 0.571 (0.023) 28.5 (1.9) 0.06 (0.01) 0.992 56.9 % 55.8 % (3.0 %) 25 MQ 6.5 0.636 (0.032) 12.9 (1.1) 0.14 (0.03) 0.980 NA c NA 25 MQ 7.0 0.906 (0.045) 4.6 (0.5) 0.31 (0.08) 0.991 90.6 % 82.7 % (2.4 %) 25 MQ 7.5 0.778 (0.024) 4.5 (0.3) 0.31 (0.05) 0.993 NA NA 25 MQ 8.0 0.628 (0.022) 1.7 (0.2) 0.66 (0.15) 0.983 62.8 % 66.4 % (2.1 %) 25 MQ 9.0 0.504 (0.020) 0.6 (0.1) 1.63 (0.45) 0.955 50.4 % 48.6 % (2.8 %) 5 MQ 10.0 0.260 (0.042) 0.9 (0.7) 2.12 (8.96) 0.846 26.0 % 37.7 % (1.9 %) 25 LW 7.0 0.824 (0.025) 12.5 (0.7) 0.18 (0.03) 0.993 NA NA 25 LW 8.0 0.783 (0.108) 5.7 (0.2) 0.27 (0.20) 0.993 NA NA Sumatriptan 25 MQ 6.0 0.022 (0.001) 36.6 (3.8) 0.04 (0.01) 0.976 NA NA 25 MQ 7.0 0.028 (0.001) 27.0 (2.4) 0.06 (0.01) 0.992 NA NA 25 MQ 8.0 0.042 (0.002) 16.0 (1.7) 0.07 (0.01) 0.974 NA NA 25 MQ 8.5 0.030 (0.002) 7.8 (1.8) 0.11 (0.04) 0.930 NA NA 25 MQ 9.0 0.024 (0.002) 6.1 (1.9) 0.11 (0.05) 0.889 NA NA a Numbers in these brackets represent the 95 % confidence interval of each model parameter. b Numbers in these brackets represent the standard deviation from multiple tests (n = 3). c Not available. Independent 72 hr NDMA-FP experiments were not performed.
201
Table A9.3. Summary of the estimated kinetic model parameters and model verification: the prechlorination effect (pH = 7.0 ± 0.1)
Compound Conc. (nM) Matrix Parameter Estimation Model Verification
Pre-Cl (min) θ a Lag (hr) a k (hr-1) a R2 Model-predicted
conversion @ 24 hr Measured conversion
@ 24 hr b Ranitidine 5 MQ 0 c 0.912 (0.045) 6.3 (0.5) 0.23 (0.04) 0.992 91.1 % 85.2 % (0.8 %) MQ 0.5 0.528 (0.061) 2.8 (0.9) 0.30 (0.18) 0.926 52.8 % 52.0 % (1.0 %) MQ 3 0.534 (0.044) 4.8 (1.2) 0.19 (0.09) 0.959 53.4 % 50.6 % (1.4 %) MQ 30 0.396 (0.024) 2.5 (0.4) 0.51 (0.20) 0.978 39.6 % 45.4 % (3.6 %) MQ 60 0.466 (0.013) 2.6 (0.2) 0.70 (0.14) 0.996 46.6 % 52.7 % (1.5 %) MQ 120 0.431 (0.031) 3.2 (0.6) 0.32 (0.12) 0.973 43.1 % 44.6 % (1.8 %) 25 MQ 0 0.906 (0.045) 4.6 (0.5) 0.31 (0.08) 0.991 90.3 % 82.7 % (2.4 %) MQ 0.5 0.434 (0.046) 2.2 (0.6) 0.51 (0.37) 0.932 43.4 % 49.8 % (5.4 %) MQ 3 0.492 (0.048) 4.7 (1.4) 0.19 (0.10) 0.943 49.2 % 52.0 % (1.0 %) MQ 30 0.399 (0.037) 2.9 (0.7) 0.38 (0.20) 0.953 39.9 % 46.9 % (3.5 %) MQ 60 0.460 (0.041) 2.9 (0.6) 0.39 (0.20) 0.956 46.0 % 49.7 % (3.3 %) MQ 120 0.479 (0.057) 5.4 (1.5) 0.16 (0.08) 0.949 47.9 % 48.2 % (4.4 %) 25 LW 0 0.824 (0.025) 12.5 (0.7) 0.18 (0.03) 0.993 NA d NA LW 3 0.468 (0.029) 4.5 (0.8) 0.22 (0.08) 0.963 NA NA LW 30 0.529 (0.023) 3.6 (0.4) 0.33 (0.08) 0.976 NA NA LW 60 0.460 (0.024) 2.9 (0.3) 0.49 (0.12) 0.975 NA NA LW 120 0.474 (0.018) 2.9 (0.2) 0.48 (0.09) 0.986 NA NA 25 RW 0 0.863 (0.018) 31.0 (1.1) 0.07 (0.01) 0.998 NA NA RW 3 0.662 (0.024) 4.1 (0.4) 0.34 (0.08) 0.985 NA NA RW 30 0.125 (0.005) 28.6 (1.9) 0.06 (0.01) 0.991 NA NA RW 60 0.028 (0.003) 19.3 (2.3) 0.07 (0.02) 0.957 NA NA RW 120 0.034 (0.125) 56.8 (85.0) 0.03 (0.02) 0.939 NA NA Sumatriptan 5 MQ 0 0.023 (0.002) 22.8 (2.7) 0.08 (0.04) 0.966 1.3 % 1.3 % (0.2 %) MQ 0.5 0.034 (0.004) 17.1 (3.9) 0.06 (0.03) 0.941 2.4 % 2.6 % (0.3 %) MQ 3 0.032 (0.003) 12.1 (2.3) 0.08 (0.03) 0.965 2.9 % 2.8 % (0.2 %) MQ 30 0.053 (0.006) 16.4 (3.1) 0.07 (0.03) 0.951 4.1 % 4.9 % (0.4 %) MQ 60 0.062 (0.006) 12.2 (2.8) 0.07 (0.03) 0.951 5.4 % 5.1 % (0.8 %) MQ 120 0.057 (0.006) 8.0 (2.6) 0.11 (0.07) 0.923 5.6 % 7.8 % (0.2 %) 25 MQ 0 0.028 (0.001) 27.0 (2.4) 0.06 (0.01) 0.992 1.1 % 2.0 % (0.1 %) MQ 0.5 0.037 (0.003) 14.2 (2.5) 0.07 (0.02) 0.967 3.0 % 3.1 % (0.4 %) MQ 3 0.032 (0.001) 10.4 (1.0) 0.11 (0.02) 0.991 3.1 % 3.4 % (0.1 %) MQ 30 0.052 (0.006) 17.5 (3.9) 0.05 (0.02) 0.947 3.6 % 5.2 % (0.2 %)
202
Compound Conc. (nM) Matrix Parameter Estimation Model Verification
Pre-Cl (min) θ a Lag (hr) a k (hr-1) a R2 Model-predicted
conversion @ 24 hr Measured conversion
@ 24 hr b Sumatriptan 25 MQ 60 0.060 (0.005) 12.4 (2.2) 0.08 (0.02) 0.971 5.3 % 5.5 % (0.7 %) MQ 120 0.058 (0.005) 9.9 (2.3) 0.10 (0.04) 0.954 5.5 % 6.7 % (1.3 %) 25 LW 0 0.022 (0.001) 31.8 (2.3) 0.06 (0.01) 0.991 NA NA LW 3 0.013 (0.001) 15.3 (1.4) 0.10 (0.02) 0.976 NA NA LW 30 0.043 (0.004) 9.9 (2.1) 0.09 (0.03) 0.935 NA NA LW 60 0.029 (0.002) 13.1 (2.1) 0.09 (0.03) 0.947 NA NA LW 120 0.022 (0.001) 13.3 (1.8) 0.08 (0.02) 0.962 NA NA 25 RW 0 0.024 (0.002) 66.8 (4.0) 0.03 (0.01) 0.988 NA NA RW 3 0.023 (0.001) 24.9 (1.7) 0.06 (0.01) 0.989 NA NA RW 30 0.037 (0.003) 33.2 (4.9) 0.03 (0.01) 0.969 NA NA RW 60 0.032 (0.002) 37.4 (4.7) 0.03 (0.01) 0.974 NA NA RW 120 0.015 (0.001) 33.9 (4.9) 0.03 (0.01) 0.975 NA NA a Numbers in these brackets represent the 95 % confidence interval of each model parameter. b Numbers in these brackets represent the standard deviation from multiple tests (n = 3). c 0 minute of prechlorination represents the use of preformed chloramine. d Not available. Independent 24 hr NDMA-FP experiments were not performed.
203
The kinetic model was verified by comparing the NDMA-FP predicted using the estimated
model with that measured from independent formation potential tests (the last two columns in
Table A9.1 to Table A9.3). All of the statistical analysis was performed using the GraphPad
Prism® regression analysis tool. First, linear regression was applied to the measured and
predicted NDMA molar conversion values, and the F-test was performed to determine if the
slope of the regression line differed significantly from zero (i.e., p-value < 0.05, 95 %
confidence level), indicating if the correlation was significant. Following the F-test, the
Student’s t-test was conducted to determine whether the slope of each regression line differed
significantly from 1.0. The t-test compared the fits between the constrained model (slope = 1.0;
null hypothesis) and the slope-unconstrained model (alternative hypothesis). If the p-value was
less than 0.05 (95% confidence level), the null hypothesis was rejected, and the best-fit slope
was different from 1.0 (the best-fit slope > 1 indicates that the model tends to underestimate the
NDMA-FP, and vise versa).
The model was first verified using all of the experimental data obtained in this work, and then
further verified with subgroups. All of the statistical analysis results are summarized in Table
A9.4. The linear regression was performed for all of the experimental data (Figure A9.1), for
each pharmaceutical individually (Figure A9.2, including data in all experiments for each
pharmaceutical, while Figure 5.5 only includes data from Chapter 5, Table 5.3), for the pH
effect (Figure A9.3), as well as for the chloramination experiments with and without
prechlorination (Figure A9.4).
204
Table A9.4. Summary of the F-test and Student’s t-test results for the kinetic model verification (95 % confidence level)
F-test t-test
p-value Significant correlation?
p-value Preferred model Best-fit slope R2
Overall < 0.0001 Yes < 0.0001 Unconstrained 0.942 ± 0.012 0.977 Conversion < 10 % < 0.0001 Yes 0.1417 Constrained 1.000 0.739 Sort by compound Chlorphenamine 0.0546 * Not quite * 0.0071 Unconstrained 0.594 ± 0.132 0.561 Doxylamine 0.0223 Yes < 0.0001 Unconstrained 0.535 ± 0.075 0.759 Ranitidine < 0.0001 Yes < 0.0001 Unconstrained 0.761 ± 0.033 0.884 Sumatriptan < 0.0001 Yes 0.338 Constrained 1.000 0.829 pH effect 0.0035 Yes < 0.0001 Unconstrained 0.725 ± 0.049 0.943 Sort by treatment Preformed chloramine < 0.0001 Yes < 0.0001 Unconstrained 0.934 ± 0.012 0.989 Sequential chlorination + chloramination < 0.0001 Yes 0.0264 Unconstrained 1.043 ± 0.019 0.982
* The correlation was determined to be significant at 90 % confidence level.
Figure A9.1. Model verification: all the experimental data (Error bars represent the standard deviation from multiple formation potential tests (n = 3))
205
Figure A9.2. Model verification: individual pharmaceuticals (Error bars represent the standard deviation from multiple formation potential tests (n = 3))
206
Figure A9.3. Model verification: pH effect (Ranitidine, 25 nM, MQ water; error bars represent the standard deviation from multiple formation potential tests (n = 3))
Figure A9.4. Model verification: preformed chloramination vs. sequential chlorination and chloramination (Error bars represent the standard deviation from multiple formation potential tests (n = 3))
207
Appendix 10. Preliminary Chlorine/Chloramine Demand Tests to Determine the Initial NaClO and NH4Cl Dosage for the Sequential Disinfection Experiments
(Chapter 7)
Step 1: Determine the 24 hr chloramine demand (D24) using preformed NH2Cl
(Water pH controlled within 7.0 ± 0.1 using a phosphate buffer)
Table A10.1. The 24 hr chloramine demand (D24) for Lake Ontario water
Target initial dosage (mg/L) Actual initial conc. (mg/L) 24 hr conc. (mg/L) D24 (mg/L) Total NH2Cl Total NH2Cl (total-Cl)
2.5 2.3 2.3 2.3 1.9 0.0 5.0 4.7 4.8 4.6 3.5 0.1
10.0 9.5 9.4 9.2 8.2 0.3 Mean 0.2
Table A10.2. The 24 hr chloramine demand (D24) for Otonabee River water
Target initial dosage (mg/L) Actual initial conc. (mg/L) 24 hr conc. (mg/L) D24 (mg/L) Total NH2Cl Total NH2Cl (total-Cl)
5 5.0 4.6 4.2 3.6 0.8 7.5 7.6 7.2 6.6 5.3 1.0 10 10.0 9.3 8.7 6.9 1.3
Mean 1.0
Step 2: Determine the free chlorine (NaClO) demand (dti) for the respective chlorine
contact time
(Water pH controlled within 7.0 ± 0.1 using a phosphate buffer)
Table A10.3. The free chlorine demand (dti) for Lake Ontario water
Time (min)
Measure total-Cl conc. (mg/L)
Calculating the Cl2-demand (mg/L) at each time point: [2]
Average Cl2-demand dti (mg/L)
3 [1] 2.05 4.40 8.65 0 0 0 0.0 30 2.05 4.15 8.50 0.00 0.25 0.15 0.1 60 1.90 4.05 8.45 0.15 0.35 0.20 0.2
120 1.85 4.05 8.05 0.20 0.35 0.60 0.4
Table A10.4. The free chlorine demand (dti) for Otonabee River water
Time (min)
Measure total-Cl conc. (mg/L)
Calculating the Cl2-demand (mg/L) at each time point: [2]
Average Cl2-demand dti (mg/L)
3 [1] 3.30 5.70 7.45 0 0 0 0 30 2.65 4.75 6.40 0.65 0.95 1.05 0.9 60 2.25 4.25 6.25 1.05 1.45 1.20 1.2
120 1.85 3.75 5.83 1.45 1.95 1.63 1.7
208
[1] Practically it’s difficult to measure the actual initial concentration. Usually the samples have to be equilibrated
for a few minutes to ensure well mixing of reagents. Considering the short time period from 0 to 3 min, it was
assumed that the measured concentration at 3 min was approximately the initial concentration.
[2] The Cl2-demand at 3 min was assumed to be 0, and the following demands were calculated using the
concentration at 3 min as the initial concentration.
NH2Cl concentration was also measured in the above samples but the concentration was very
low (<0.1 mg/L or below detection limit), suggesting very low NH4+ level in the selected water
matrices.
Step 3: Calculate the dosing volume (V-NH4Cl and V-NaClO)
Target total-Cl dosage at the point of NH4Cl addition = 2.5 + D24 (mg/L), thus the NH4Cl
dosage was calculated based on the concentration of NaClO at the point of NH4Cl addition and
the Cl2:NH4-N mass ratio of 4.2:1.
Table A10.5. The NH4Cl dosage following prechlorination
Matrix [NaClO] @ NH4Cl addition (2.5 + D24)
[NH4Cl] dosage V-NH4Cl (stock solution of 1.35 g/L)
Lake Ontario 2.7 mg/L 2.5 mg/L 1.8 mL/L
Otonabee River 3.5 mg/L 3.2 mg/L 2.4 mL/L
For the prechlorination experiment, the initial NaClO dosage = 2.5 + D24 + dti (mg/L) for the
respective chlorine contact time, and the NaClO dosing volume was then calculated based on
the stock solution concentration (measured daily).
Table A10.6. The initial NaClO dosage for the respective chlorine contact time
Initial [NaClO] dosage (2.5 + D24 + dti) Pre-Cl contact time
(min) 3 30 60 120
Lake Ontario 2.7 2.8 2.9 3.1
Otonabee River 3.5 4.4 4.7 5.2
209
Appendix 11. Prechlorination Impact on the 24 hr NDMA-FP from Selected Pharmaceuticals – ANOVA and Tukey’s HSD Analysis (Chapter 7)
A one-way ANOVA test was performed for each pharmaceutical, followed by a Tukey’s HSD
(honestly significant difference) test. The one-way ANOVA (one treatment: prechlorination;
six-level: contact time from 0 to 120 min) test is to determine if the prechlorination impact was
significant, while the Tukey’s HSD test is to determine how significant the difference was
among the different contact times applied. All of these analyses were performed using the
GraphPad Prism® ANOVA tool.
Table A11.1. Prechlorination impact on the 24 hr NDMA-FP from selected pharmaceuticals: summary of the one-way ANOVA and Tukey’s HSD results (95 % confidence level)
One-way ANOVA Ranitidine Nizatidine Diltiazem Sumatriptan P value < 0.0001 < 0.0001 < 0.0001 < 0.0001 Are means significant different? (P < 0.05) Yes Yes Yes Yes
Number of groups 6 6 6 6 F 49.06 28.52 32.02 32.95 R2 0.9534 0.9224 0.9303 0.9321 Tukey's Multiple Comparison Test Significant? (P < 0.05) 0 min vs 0.5 min Yes No Yes No 0 min vs 3 min Yes Yes Yes No 0 min vs 30 min Yes Yes Yes Yes 0 min vs 60 min Yes Yes No Yes 0 min vs 120 min Yes Yes No Yes 0.5 min vs 3 min No Yes Yes No 0.5 min vs 30 min No Yes Yes Yes 0.5 min vs 60 min No Yes Yes Yes 0.5 min vs 120 min No Yes Yes Yes 3 min vs 30 min No No No Yes 3 min vs 60 min No No Yes Yes 3 min vs 120 min No No Yes Yes 30 min vs 60 min No No Yes No 30 min vs 120 min No Yes Yes Yes 60 min vs 120 min No No No Yes
210
(Cont’d)
One-way ANOVA Carbinoxamine Chlorphenamine Doxylamine Tetracycline P value < 0.0001 0.0001 0.0034 < 0.0001 Are means significant different? (P < 0.05) Yes Yes Yes Yes
Number of groups 6 6 6 6 F 38.75 14.45 6.683 33.95 R2 0.9417 0.8576 0.7358 0.934 Tukey's Multiple Comparison Test Significant? (P < 0.05)
0 min vs 0.5 min No No No No 0 min vs 3 min No No No No 0 min vs 30 min No No No Yes 0 min vs 60 min Yes No No Yes 0 min vs 120 min Yes Yes No Yes 0.5 min vs 3 min No No No No 0.5 min vs 30 min No No No Yes 0.5 min vs 60 min Yes No No Yes 0.5 min vs 120 min Yes Yes Yes Yes 3 min vs 30 min No No No Yes 3 min vs 60 min Yes No No Yes 3 min vs 120 min Yes Yes Yes Yes 30 min vs 60 min Yes No No No 30 min vs 120 min Yes Yes Yes No 60 min vs 120 min Yes Yes Yes No
211
Appendix 12. Additional NDMA Formation Experiments with Ranitidine in Otonabee River Water: The Possible Impact of Cl2:TOC Ratio? (Chapter 7)
As demonstrated in Chapter 7, Section 7.3.2, the NDMA formation from ranitidine upon
prechlorination showed different trends in Lake Ontario water and in the Otonabee River water.
Although the NDMA formation was reduced upon prechlorination in both matrices, the extent
of reduction was quite different. In Lake Ontario water, there was no further reduction as
chlorine contact time increased from 3 min to 120 min, and the initial lag phase was
significantly shortened. It is proposed that prechlorination could break the binding between
NOM and ranitidine and thus enhance the initial reaction to form NDMA. In Otonabee River
water, less NDMA reduction (compared with that in the lake water upon the same chlorine
contact time) was observed upon 3 min prechlorination, but the NDMA formation was
significantly inhibited as chlorine contact time increased to 30 and 120 min. The NDMA
formation from ranitidine (25 nM) was marginal (less than 50 ng/L after 48 hr of
chloramination) considering the typically high NDMA conversion from ranitidine. Considering
the NOM-ranitidine binding theory, it is suggested that prolonged chlorination may further
break down NOM to form smaller products which may rebind with ranitidine and inhibit its
further conversion into NDMA.
However, it is unclear why this “rebinding” was not observed in Lake Ontario water at longer
chlorine contact time. It is noticed that the Cl2:TOC ratio (at the point of NH4Cl addition)
applied was higher in the lake water (1.1) compared with that in the river water (0.6). In order to
evaluate the potential impact of Cl2:TOC ratio, some additional tests were performed for
ranitidine in the river water at a Cl2:TOC ratio of 1.1, the same as what was applied in the lake
water. Following the prechlorination, the NDMA formation from ranitidine was measured after
4 hr and 48 hr of chloramination. The results are summarized in Table A12.1.
Table A12.1. NDMA formation from ranitidine (25 nM) upon sequential chlorination and chloramination in Otonabee River water at different Cl2:TOC ratios (pH = 7.0 ± 0.1)
Chloramination time 4 hr 48 hr Pre-Cl contact time (min) Cl2:TOC = 0.6 Cl2:TOC = 1.1 Cl2:TOC = 0.6 Cl2:TOC = 1.1
3 32.4 % (4.2 %) * 39.9 % (1.1 %) 65.1 % (2.0 %) 58.6 % (1.7 %) 30 0.4 % (0.03 %) 0.2 % (0.02 %) 2.8 % (0.3 %) 1.4 % (0.2 %)
120 0.12 % (0.01 %) 0.1 % (0.003 %) 1.2 % (0.01 %) 1.0 % (0.2 %) * Numbers in the brackets represent the variability due to multiple formation potential tests (n = 3)
212
If the higher Cl2:TOC ratio applied in Lake Ontario water helps prevent the rebinding, it is then
expected that similar phenomenon would be observed in Otonabee River water when the
Cl2:TOC ratio applied is raised to the same level. However, the above results indicate that the
Cl2:TOC ratio is not the reason for the different results observed in the river water. The same
trend was observed when different Cl2:TOC ratios were applied; the NDMA formation from
ranitidine was significantly inhibited in the river water at longer chlorine contact time. The
results were also compared with those obtained in the lake water under the same Cl2:TOC ratio
(Figure A12.1). Therefore, the different results observed in these two matrices for ranitidine are
more likely due to the differences in the matrices themselves rather than the chlorine/chloramine
dosages applied.
Figure A12.1. NDMA formation from ranitidine upon sequential chlorination and chloramination in two matrices (Cl2:TOC = 1.1, ranitidine = 25 nM, pH = 7.0 ± 0.1; error bars represent the variability due to
multiple formation potential tests (n = 3))
213
Appendix 13. Prechlorination Impact on the NDMA Formation Kinetics from Ranitidine and Sumatriptan (Chapter 7) – ANOVA and Tukey’s HSD Analysis
The ANOVA and Tukey’s HSD analysis were performed on the estimated kinetic model
parameters for ranitidine and sumatriptan upon sequential chlorination and chloramination in
each water matrix (data in Appendix 9, Table A9.3). For each pharmaceutical in each water
matrix, the one-way ANOVA test is to determine if the prechlorination impact was significant
on the respective model parameter, while the Tukey’s HSD test is to determine how significant
the difference was among the different contact times applied. All the analysis was performed
using the GraphPad Prism® ANOVA tool.
Table A13.1. Prechlorination impact on the NDMA formation kinetics from ranitidine: summary of the one-way ANOVA and Tukey’s HSD results (95 % confidence level)
Model parameter θ Lag k Matrix MQ Lake River MQ Lake River MQ Lake River One-way ANOVA P value
< 0.0001
< 0.0001
< 0.0001
< 0.0001
< 0.0001 0.4047 0.0082 0.0005
< 0.0001
Are means significantly different? (P < 0.05) Yes Yes Yes Yes Yes No Yes Yes Yes
F 90.04 204.3 201.4 12.4 309.7 1.109 4.434 13.17 42.71 R2 0.988 0.988 0.988 0.919 0.992 0.307 0.803 0.841 0.945
Tukey's Multiple Comparison Test Significant? P < 0.05?
0 min vs 3 min Yes Yes Yes No Yes No No No Yes 0 min vs 30 min Yes Yes Yes No Yes No No No No 0 min vs 60 min Yes Yes NA* No Yes NA No Yes NA 0 min vs 120 min Yes Yes Yes No Yes No No Yes No
3 min vs 30 min No Yes Yes No No No No No Yes 3 min vs 60 min No No NA No Yes NA No Yes NA 3 min vs 120 min No No Yes No Yes No No Yes Yes
30 min vs 60 min No Yes NA No No NA No No NA 30 min vs 120 min No Yes No Yes No No No No No
60 min vs 120 min No No NA Yes No NA No No NA * Not available
214
Table A13.2. Prechlorination impact on the NDMA formation kinetics from sumatriptan: summary of the one-way ANOVA and Tukey’s HSD results (95 % confidence level)
Model parameter θ Lag k Matrix MQ Lake River MQ Lake River MQ Lake River One-way ANOVA P value <
0.0001 <
0.0001 <
0.0001 <
0.0001 <
0.0001 <
0.0001 0.1755 0.1301 0.0024 Are means significantly different? (P < 0.05) Yes Yes Yes Yes Yes Yes No No Yes Number of groups 12 5 5 12 5 5 12 5 5 F 45.78 103.9 90.88 19.12 83.35 62.75 1.748 2.3 9 R2 0.977 0.977 0.973 0.946 0.971 0.962 0.616 0.479 0.782
Tukey's Multiple Comparison Test Significant? P < 0.05?
0 min vs 3 min No Yes No Yes Yes Yes No No Yes 0 min vs 30 min Yes Yes Yes Yes Yes Yes No No No 0 min vs 60 min Yes Yes Yes Yes Yes Yes No No No 0 min vs 120 min Yes No Yes Yes Yes Yes No No No
3 min vs 30 min Yes Yes Yes No Yes No No No Yes
3 min vs 60 min Yes Yes Yes No No Yes No No Yes 3 min vs 120 min Yes Yes Yes No No No No No Yes
30 min vs 60 min No Yes Yes No No No No No No
30 min vs 120 min No Yes Yes Yes No No No No No
60 min vs 120 min No Yes Yes No No No No No No
215
Appendix 14. Preliminary LC-MS results (Chapter 7)
Some preliminary Liquid Chromatography-Mass Spectrometry (LC-MS) analysis was
performed in this work, intent to compare the “free” pharmaceutical concentrations with and
without the presence of natural organic matter (NOM). It is hypothesized that the “free”
pharmaceutical concentration in the presence of NOM would be lower than that in the absence
of NOM due to the formation of NOM-pharmaceutical complexes.
Pharmaceutical Analysis
An LC-MS analytical method was developed for ranitidine and sumatriptan. Because the
pharmaceutical concentrations analyzed on the LC-MS were generally at lower µg/L levels,
water samples (dosed with each single analyte and its respective internal standard) could be
directly injected into the LC-MS without extraction. The LC-MS method is outlined in Table
A14.1. MDLs for ranitidine and sumatriptan were determined in Lake Ontario water based on
the EPA standard method (USEPA, 1986) and are provided in Table A14.2. Given that no
extraction was performed, the MDLs are essentially the instrumental detection limits in this
matrix. Individual pharmaceutical concentrations were estimated via correlation with standards
using a 6-level calibration curve prepared using the respective water matrix. Calibration
standards were included in every analyzed sample set, and were prepared the day before the LC-
MS analysis to allow enough equilibration time between the dosed pharmaceuticals and the
water matrix.
Table A14.1. The LC-MS analysis of pharmaceuticals method outline
The pharmaceutical samples were analyzed via a Varian 212 LC coupled with a Varian 500 MS and a ProStar 410 autosampler (Agilent Technologies, Mississauga, ON).
LC conditions:
The LC was equipped with a Metaguard Pursuit guard column (2.0 mm) and Pursuit XRs Ultra 2.8 C18 analytical column (100 x 2.0 mm, maximum pressured 1000 bar). A full-loop (100 µL) of sample was injected via the autosampler injection port. The mobile phases used were 0.03% HFBA (heptafluorobutyric acid) in LC-MS grade water (A; purchased from Sigma-Aldrich Canada) and acetonitrile (B).
216
Mobile phase gradient: Time (min) %B (Flow rate: 200 µL/min) 00:00 48 05:00 48 07:00 100 09:00 100
10:00 48 20:00 48
MS conditions:
The MS was equipped with an electrospray ionization (ESI) source and operated at full-scan mode for each single analyte. Ranitidine and d6-ranitidine were both eluted at a retention time of 2.2 min, with indicating ions monitored at 315.2 and 321.3 amu, respectively; sumatriptan and d6-sumatriptan were both eluted at a retention time of 2.1 min, with indicating ions monitored at 296.2 and 302.2 amu, respectively. Other MS conditions are as follows:
Ranitidine Sumatriptan Capillary voltage (V) 43 51 RF loading (%) 79 76 Needle voltage (V) 5000 5550
Table A14.2. LC-MS method: Method detection limits for pharmaceuticals (in Lake Ontario water)
Compound 1 2 3 4 5 6 7 AVE (µg/L) STD (µg/L) Recovery MDL (µg/L) Ranitidine 2.7 2.9 3.1 2.6 2.6 3.0 2.8 2.8 0.20 112.4 % 0.6
Sumatriptan 2.9 2.4 3.0 2.6 2.5 2.4 2.5 2.6 0.22 104.3 % 0.7
Experimental Protocol
In order to test the above hypothesis, the calibration curve for ranitidine and sumatriptan was
prepared in three water matrices, including the lab-grade MQ (Milli-Q®; Millipore, Etobicoke,
Ontario) water, Lake Ontario water, and NOM-removed Lake Ontario water. The NOM-
removed Lake Ontario water was obtained by passing the water through a fluidized ion
exchange (FIX) resin column at a flow velocity of 10 m/h (approximately 500 mL/min). The
transparent Perspex column (2.25 m, internal diameter of 6.3 cm, Piedmont Plastics Inc.,
Scarborough, ON, Canada) was packed with approximately 1.9 L of anion exchange resin
(Purolite® PPA860S resin, mean particle diameter of 740 ± 110 µm, The Purolite Company,
Bala Cynwyd, PA, USA). The FIX treatment removed about 75 % of the TOC from Lake
Ontario water. The basic water quality parameters for the three matrices are summarized in
Table A14.3. Although the pH varied from 6.8 to 8.0, buffer solution was used to control the pH
217
within 7.0 ± 0.1 when preparing the calibration samples. The calibration curves were prepared
both with and without the addition of isotope-labeled internal standard (IS).
Table A14.3. Basic water quality parameters for the three water matrices
Matrix TOC (mg/L) pH Alkalinity
(mg/L as CaCO3) UV254 (cm-1)
SUVA (L/mg·m)
Milli-Q® (MQ) 0.0 7.5 ± 0.1 1.8 ± 0.3 0.000 0.0 Lake Ontario water 2.3 ± 0.2 8.0 ± 0.1 94.6 ± 1.5 0.024 ± 0.002 1.1 ± 0.1 NOM-removed Lake Ontario water 0.6 ± 0.05 6.8 ± 0.1 15.0 ± 1.0 0.004 ± 0.001 0.7 ± 0.1
Results and Discussion
Theoretically, if the pharmaceuticals bind with NOM, the “free” pharmaceutical concentration
will be lower and thus the peak area count of the chromatograph will be smaller. The calibration
curve (without the addition of internal standard) was plotted as concentration vs. area count,
thus the slope of the calibration curve in the presence of NOM should be steeper than that in the
absence of NOM. The results are compared in Figure A14.1. As expected, the slope of
calibration curve prepared in Lake Ontario water had the highest value for both ranitidine and
sumatriptan, while those prepared in MQ water and the NOM-removed Lake Ontario water had
relatively less steep slopes of comparable values.
Figure A14.1. Calibration curves for ranitidine and sumatriptan prepared in three water matrices without
the addition of internal standard
218
However, it cannot be concluded that the lower area counts for the pharmaceuticals in the lake
water was only due to the NOM-pharmaceutical binding, because there could be signal
suppression/enhancement caused by the matrix components as well. The addition of isotope-
labeled IS was able to minimize the matrix impact on signal suppression/enhancement (i.e., the
slopes of calibration curves prepared in different water matrices were relatively comparable with
each other, as compared in Figure A14.2), but it cannot separate the matrix impact from the
NOM-pharmaceutical binding. Theoretically, the isotope-labeled IS would behave similarly
with the regular pharmaceuticals and thus should also bind with NOM. Therefore, with the
addition of IS, it can only determine the “total” pharmaceutical concentration (free and bound)
instead of the “free” pharmaceutical concentration.
Figure A14.2. Calibration curves for ranitidine and sumatriptan prepared in three water matrices with the addition of internal standard
Reference
USEPA, 1986. Definition and procedure for the determination of the method detection limit. 40
CFR Part 136, Appendix B, revision 1.11, updated on June 30th, 1986.
219
Appendix 15. Additional LC-OCD Data for Chapter 7
In Chapter 7, Section 7.3.3.3, the potential change of NOM components upon the
chloramination and sequential disinfection treatment was investigated. Only results for the
major NOM fractions (i.e., humics and building blocks) were shown in the chapter; additional
data for the other fractions are summarized in Figure A15.1.
. Figure A15.1. Change of other NOM fractions upon sequential chlorination and chloramination in Lake
Ontario water and Otonabee River water
220
Figure A15.2 is a comprehensive demonstration of the change of all the NOM fractions along
the sequential chlorination (120 min) and chloramination disinfection for both water matrices.
Figure A15.2. Change of NOM fractions upon sequential chlorination (120 min) and chloramination in
Lake Ontario water and Otonabee River water
221
In addition, in Chapter 7, Section 7.3.2, it is suggested that the higher TOC level in the river
water might be responsible for the “rebinding” between ranitidine and the NOM breakdown
products which was not observed in the lake water with a lower TOC. The LC-OCD results
indicate that upon the sequential disinfection, humics was first degraded into building blocks,
and further breakdown of building blocks would lead to the increase in LMW fractions. There is
no significant difference in the distribution of NOM fractions between the two matrices,
although the river water has a higher portion of humics while the lake water has a higher portion
of biopolymers (Figure A15.3). However, considering the much higher TOC level in the river
water, the absolute amount (in mg/L organic carbon) that changed upon the treatment process
can be much bigger than the lake water.
Figure A15.3. The distribution of hydrophilic NOM fractions for two water matrices
Figure A15.4 compares the absolute amount of each NOM fraction changed (in mg/L organic
carbon) along the sequential chlorination (120 min) and chloramination treatment in the two
matrices. In general, biopolymers and humics were both degraded, but the reduced amount was
much bigger in the river water. In the lake water, the reduction in building blocks led to the
increase in LMW fractions; while in river water, the building blocks increased first (due to the
breakdown of humics) and then degraded into LMW fractions. The increase in building blocks
and LMW fractions were also bigger in the river water than that in the lake water. As such, there
are more NOM breakdown products in the river water due to the breakdown of large NOM
fractions, which may be responsible for the “rebinding” with ranitidine and thus the further
inhibited NDMA conversion as chlorine contact time increased.
222
Figure A15.4. Comparison of the change of NOM fractions upon sequential chlorination (120 min) and chloramination in Lake Ontario water and Otonabee River water
