v/trM
United States
Environmental Protection
Agency
Six-Year Review 3 Technical Support
Document for Nitrosamines
-------
Office of Water (4607M)
EPA 810-R-16-009
December 2016
-------
Disclaimer
This document is not a regulation. It is not legally enforceable, and does not confer legal rights
or impose legal obligations on any party, including EPA, states, or the regulated community.
While EPA has made every effort to ensure the accuracy of any references to statutory or
regulatory requirements, the obligations of the interested stakeholders are determined by statutes,
regulations or other legally binding requirements, not this document. In the event of a conflict
between the information in this document and any statute or regulation, this document would not
be controlling.
-------
This page intentionally left blank.
-------
Table of Contents
1 Introduction 1-1
2 Contaminant Background 2-1
2.1 Physical and Chemical Properties 2-1
2.2 Production, Use and Release 2-4
2.2.1 Production and Use 2-4
2.2.2 Environmental Release 2-4
2.3 Formation in Environmental Media 2-9
2.3.1 NDMA 2-9
2.3.2 Other Nitrosamines 2-9
2.4 Environmental Fate and Transport 2-9
2.4.1 Persistence, Bioaccumulation and Toxicity (PBT) Profiler 2-10
2.5 Regulatory and Non-Regulatory Actions for Nitrosamines 2-15
3 Health Effects 3-1
3.1 Health Effects Assessments for Individual Nitrosamine Compounds 3-1
3.1.1 NDBA 3-2
3.1.2 NDEA 3-3
3.1.3 NDMA 3-4
3.1.4 NDPA 3-5
3.1.5 NMEA 3-6
3.1.6 NPYR 3-7
3.1.7 Sensitive Populations 3-8
3.2 Calculation of Health Reference Levels (HRLs) for Nitrosamine Compounds 3-9
4 Analytical Methods 4-1
4.1 Introduction 4-1
4.2 EPA Method 521 4-2
4.2.1 Calculation of EPA Method 521 Nitrosamine MRLs 4-3
4.2.2 Comparison of EPA Method 521 Performance to HRLs 4-4
4.3 Other Drinking Water Methods 4-5
4.4 Other Published Methods for Measurement of Nitrosamines in Aqueous Media 4-5
4.5 Other Methods Used in Research 4-6
5 Occurrence and Exposure in Drinking Water 5-1
5.1 Introduction 5-1
5.2 UCMR 2 Monitoring Program and Dataset 5-2
Six-Year Review 3 i December 2016
Technical Support Document for Nitrosamines
-------
5.2.1 Description of Data Collected Under UCMR 2 5-3
5.2.2 Stratification of UCMR 2 Data for Nitrosamine Group Analysis 5-5
5.2.3 Summary of UCMR 2 Dataset for Nitrosamines 5-7
5.3 Summary of UCMR 2 Occurrence Findings for Nitrosamines 5-12
5.3.1 Rates of Detection 5-12
5.3.2 Detected Concentrations 5-15
5.3.3 Sample Location Analysis 5-19
5.3.4 Population Affected 5-27
5.4 Nitrosamine Co-Occurrence and Aggregate Occurrence in UCMR 2 Sampling 5-33
5.5 Modeling of NDMA Occurrence 5-37
5.5.1 Overview of the Modeling Approach for Generating National Occurrence 5-37
5.5.2 National Occurrence and Exposure Estimates: One or More Detections 5-43
5.5.3 National Occurrence and Exposure Estimates: Mean Concentrations 5-49
5.5.4 National Co-Occurrence of Nitrosamines 5-58
5.6 Summary and Discussion 5-60
6 Formation 6-1
6.1 Introduction 6-1
6.2 Nitrosamine Formation Potential 6-1
6.3 Formation Pathways 6-2
6.3.1 Chloramination Pathway 6-2
6.3.2 Chlorine Enhanced Nitrosation Pathway 6-5
6.3.3 Breakpoint Chlorination 6-6
6.3.4 Other Formation Pathways 6-6
6.4 Precursors: Sources and Characterization 6-8
6.4.1 Natural Precursors 6-9
6.4.2 Anthropogenic Precursors 6-10
6.4.3 Precursor Characterization 6-18
6.5 Key Factors Impacting Formation 6-21
6.5.1 The Impact of Chlorination and Chloramination 6-21
6.5.2 The Impact of Water Quality Parameters 6-22
6.6 Kinetics and Predictive Models 6-24
6.7 Summary 6-24
7 Treatment 7-1
7.1 Introduction 7-1
7.2 Prevention of Nitrosamine Formation 7-1
Six-Year Review 3 ii December 2016
Technical Support Document for Nitrosamines
-------
7.2.1 Precursor Removal 7-2
7.2.2 Modification of Disinfection Practice 7-10
7.3 Nitrosamine Removal 7-11
7.3.1 Enhanced Coagulation 7-12
7.3.2 Adsorption 7-12
7.3.3 Membrane Filtration 7-13
7.3.4 Metal Catalysis 7-14
7.3.5 Sunlight Photolysi s 7-15
7.3.6 UV Photolysis 7-16
7.3.7 Advanced Oxidation Processes (AOP) 7-17
7.3.8 Electrochemical Techniques 7-19
7.3.9 Biological Techniques 7-19
7.4 Summary 7-19
8 References 8-1
Six-Year Review 3 iii December 2016
Technical Support Document for Nitrosamines
-------
Appendices
Appendix A: Non-UCMR 2 Occurrence in Ambient Water and Drinking Water
Appendix B: UCMR 2 Systems and Population Served Predicted To Have Detections Exceeding
Thresholds
Appendix C: National Extrapolation of Facilities and Population Served Predicted To Have
Detections Exceeding Thresholds
Appendix D: National Extrapolation of Systems and Population Served Predicted To Have
Detections Exceeding Thresholds
Appendix E: National Extrapolation of Facilities and Population Served Predicted To Have a
Mean Concentration Exceeding Thresholds
Appendix F: National Extrapolation of Systems and Population Served Predicted To Have a
Mean Concentration Exceeding Thresholds
Appendix G: National Extrapolation of Systems and Population Served Predicted To Have a
Locational Annual Average (LAA) Exceeding Thresholds
Six-Year Review 3
Technical Support Document for Nitrosamines
iv
December 2016
-------
List of Exhibits
Exhibit 2.1: Chemical Structure for Nitrosamines 2-1
Exhibit 2.2: Chemical Structures of NDBA, NDEA, NDMA, NDPA, NMEA, and
NPYR 2-1
Exhibit 2.3: Physical and Chemical Properties of NDBA, NDEA, NDMA, NDPA,
NMEA, and NPYR 2-2
Exhibit 2.4: Environmental Releases (in pounds) of NDBA in the United States,
1998-2010 2-5
Exhibit 2.5: Environmental Releases (in pounds) of NDEA in the United States,
1998-2010 2-6
Exhibit 2.6: Environmental Releases (in pounds) of NDPA in the United States,
1998-2010 2-6
Exhibit 2.7: Summary of Total Releases and Total Surface Water Discharges (in
pounds) of NDBA in 2002, 2004, 2006, 2008 and 2010 2-7
Exhibit 2.8: Summary of Total Releases and Total Surface Water Discharges (in
pounds) of NDEA in 2002, 2004, 2006, 2008 and 2010 2-7
Exhibit 2.9: Summary of Total Releases and Total Surface Water Discharges (in
pounds) of NDPA in 2002, 2004, 2006, 2008 and 2010 2-7
Exhibit 2.10: PBT Profiler Data for Six Nitrosamines 2-10
Exhibit 3.1: Sensitive Populations, as Suggested by Data from Animal Studies 3-8
Exhibit 3.2: EPA-Derived Cancer Risk Values and HRLs for Six Nitrosamines 3-11
Exhibit 4.1: Analytical Methods for SYR3 Nitrosamines 4-2
Exhibit 4.2: Performance Metrics for Six Nitrosamines in EPA Method 521 4-3
Exhibit 4.3: Method Sensitivity Ratios (MSRs) for Nitrosamines 4-4
Exhibit 4.4: Method Performance Metrics for Nitrosamines Using Various Methods
Applicable to Aqueous Matrices1 4-6
Exhibit 5.1: HRLs and MRLs for the Six Nitrosamine Compounds 5-2
Exhibit 5.2: PWS and Population Stratification for UCMR 2 Screening Survey
Monitoring 5-4
Exhibit 5.3: Source Water Type Classification Scheme 5-6
Exhibit 5.4: Disinfectant Classification Scheme 5-7
Exhibit 5.5: Summary of UCMR 2 Dataset for Nitrosamines 5-8
Exhibit 5.6: Number of PWS s in the UCMR 2 Dataset for Nitrosamines by
Disinfectant Type 5-8
Exhibit 5.7: Number of Entry Points in the UCMR 2 Nitrosamines Dataset by
Disinfectant Type 5-9
Exhibit 5.8: Number of Maximum Residence Time Locations in the UCMR 2
Nitrosamines Dataset by Disinfectant Type 5-10
Exhibit 5.9: Number of Samples in the UCMR 2 Dataset for NDMA by Disinfectant
Type 5-10
Exhibit 5.10: Number of Samples at Entry Points in the UCMR 2 Dataset for NDMA
by Disinfectant Type 5-11
Exhibit 5.11: Number of Samples at Maximum Residence Time Locations in the
UCMR 2 Dataset for NDMA by Disinfectant Type 5-12
Exhibit 5.12: Nitrosamine Detection Rates for UCMR 2 Data by PWS Size and
Source Water Type 5-13
Six-Year Review 3
Technical Support Document for Nitrosamines
v
December 2016
-------
Exhibit 5.13: Nitrosamine Detection Rates for UCMR 2 Data by Disinfectant 5-14
Exhibit 5.14: Summary of Nitrosamine Concentrations in Samples with Detections at
All Sampling Locations, by Disinfectant Type 5-16
Exhibit 5.15: Summary of Nitrosamine Concentrations in Samples from EP
Locations, with Detections by Disinfectant Type 5-17
Exhibit 5.16: Summary of Nitrosamine Concentrations in Samples from MR
Locations, by Disinfectant Type 5-18
Exhibit 5.17: Mean Nitrosamine Concentrations (in ng/L) in Samples with
Detections by Disinfectant Type 5-19
Exhibit 5.18: Percentage of PWSs in the UCMR 2 Dataset Detecting Nitrosamines
At Least Once by PWS Size and Source Water Type 5-21
Exhibit 5.19: Percentage of PWSs in the UCMR 2 Dataset Detecting Nitrosamines
At Least Once by Disinfectant Type 5-22
Exhibit 5.20: Percent of Entry Points in the UCMR 2 Dataset Detecting Nitrosamines
At Least Once by PWS Size and Source Water Type 5-23
Exhibit 5.21: Percent of Entry Points in the UCMR 2 Dataset Detecting Nitrosamines
At Least Once by Disinfectant Type 5-24
Exhibit 5.22: Percent of Maximum Residence Time Locations in the UCMR 2
Dataset Detecting Nitrosamines At Least Once by Size and Source Water
Type 5-25
Exhibit 5.23: Percent of Maximum Residence Time Locations in the UCMR 2
Dataset Detecting Nitrosamines At Least Once by Disinfectant 5-26
Exhibit 5.24: Percentage of Population Served by PWSs in the UCMR 2 Dataset
Detecting Nitrosamines At Least Once, by PWS Size and Source Water
Type 5-28
Exhibit 5.25: Percentage of Population Served by PWSs in the UCMR 2 Dataset
Detecting Nitrosamines At Least Once, by Disinfectant Type 5-29
Exhibit 5.26: Percentage of Population at Entry Points in the UCMR 2 Dataset
Detecting Nitrosamines At Least Once, by PWS Size and Source Water
Type 5-29
Exhibit 5.27: Percentage of Population at Entry Points in the UCMR 2 Dataset
Detecting Nitrosamines At Least Once, by Disinfectant Type 5-30
Exhibit 5.28: Percentage of Population at Maximum Residence Time Locations in
the UCMR 2 Dataset Detecting Nitrosamines At Least Once, by Size and
Source Water Type 5-31
Exhibit 5.29: Percentage of Population at Maximum Residence Time Locations in
the UCMR 2 Dataset Detecting Nitrosamines At Least Once, by
Disinfectant Type 5-32
Exhibit 5.30: Co-Occurrence Venn Diagram 5-33
Exhibit 5.31: Aggregate Occurrence of Nitrosamines (UCMR 2 PWSs and
Population Exposed) 5-34
Exhibit 5.32: Co-Occurrence of NDMA with Other Nitrosamines: UCMR 2 PWSs
Affected 5-35
Exhibit 5.33: Co-Occurrence of NDMA with Other Nitrosamines: UCMR 2
Population Exposed 5-35
Exhibit 5.34: UCMR 2 Co-Occurrence Summary Statistics 5-37
Technical Support Document for Nitrosamines
-------
Exhibit 5.35: Percentage of Entry Points Predicted To Have One or More Detections
Exceeding the Threshold and the Associated Population Exposed 5-43
Exhibit 5.36: Percentage of Maximum Residence Time Locations Predicted To Have
One or More Detections Exceeding the Threshold and the Associated
Population Exposed 5-44
Exhibit 5.37: Percentage of the Population Predicted To Be Exposed to One or More
Detections at Levels Above the HRL (0.6 ng/L) at Entry Points 5-45
Exhibit 5.38: Percentage of the Population Predicted To Be Exposed to One or More
Detections at Levels Above the HRL (0.6 ng/L) at Maximum Residence
Time Locations 5-46
Exhibit 5.39: PWSs Predicted To Have One or More Detections Exceeding Various
Thresholds 5-47
Exhibit 5.40: Populations Served by PWSs Predicted To Have One or More
Detections Exceeding Various Thresholds 5-48
Exhibit 5.41: Percentage of the Population Predicted To Be Exposed to One or More
Detections at Levels Above the HRL (0.6 ng/L), Based on PWSs 5-49
Exhibit 5.42: Percentage of Entry Points with the Predicted Mean Concentration
Exceeding the Threshold and the Associated Population Exposed 5-50
Exhibit 5.43: Percentage of Maximum Residence Time Locations with the Predicted
Mean Concentration Exceeding the Threshold and the Associated
Population Exposed 5-50
Exhibit 5.44: Percentage of the Population Predicted To Be Exposed to Mean
Concentrations Greater Than the HRL (0.6 ng/L) at Entry Points 5-52
Exhibit 5.45: Percentage of the Population Predicted To Be Exposed to Mean
Concentrations Greater Than the HRL (0.6 ng/L) at Maximum Residence
Time Locations 5-53
Exhibit 5.46: PWSs with Predicted Mean Concentration Exceeding Various
Thresholds 5-54
Exhibit 5.47: Population Served by PWSs with Predicted Mean Concentration
Exceeding Various Thresholds 5-55
Exhibit 5.48: Percentage of the Population Predicted To Be Exposed to Mean
Concentrations Greater Than the HRL (0.6 ng/L) in PWSs 5-56
Exhibit 5.49: Percentage of PWSs Predicted To Have an LAA Greater Than the
Threshold and the Associated Population Exposed 5-57
Exhibit 5.50: Percentage of the Population Predicted To Be Exposed to LAAs
Greater Than the HRL (0.6 ng/L) in PWSs 5-58
Exhibit 5.51: Co-Occurrence of Nitrosamines (National Extrapolation from
UCMR 2) 5-59
Exhibit 6.1: Mechanism of NDMA Formation via the Chloramination Pathway 6-4
Exhibit 6.2: NDMA Formation via the Enhanced Nitrosation Pathway 6-6
Exhibit 6.3: Studies Correlating Organic Carbon and Organic Nitrogen to NDMA
Formation Potential 6-20
Exhibit 6.4: Number and Percentage of Entry Points and Maximum Residence Time
Locations in UCMR 2 Using Chloramine-Only Disinfection With At
Least One Sample Exceeding the Indicated NDMA Thresholds 6-22
Exhibit 7.1: Oxidation of Nitrosamine Precursors 7-10
Technical Support Document for Nitrosamines
-------
Exhibit 7.2: Summary of Removal Efficiencies of Precursors, NDMA and Other
Nitrosamines 7-21
Six-Year Review 3 viii December 2016
Technical Support Document for Nitrosamines
-------
This page intentionally left blank.
Six-Year Review 3 ix December 2016
Technical Support Document for Nitrosamines
-------
Abbreviations
ADAF
Age Dependent Adjustment Factors
AOP
Advanced Oxidation Process
BD
Berlin-Druckrey
CAS
Chemical Abstracts Service
CCL 3
Third Contaminant Candidate List
CI-MS/MS
Chemical ionization tandem mass spectrometry
CSF
Cancer Slope Factor
CT
Concentration x Time
CWS
Community Water System
CWSS
Community Water Systems Survey
CYP
Cytochrome P450
DBA
Dibutylamine
DBP
Disinfection By-Products
D/DBPR
Disinfectants and Disinfection By-Products Rule
DEA
Diethylamine
DL
Detection Limit
DMA
Dimethylamine
DMAP
4-dimethylaminoantipyrine
DMBzA
Dimethylbenzylamine
DMiPA
Dimethylisopropylamine
DNA
Deoxyribonucleic Acid
DO
Dissolved Oxygen
DOC
Dissolved Organic Carbon
DOM
Dissolved Organic Matter
DON
Dissolved Organic Nitrogen
DPA
Dipropylamine
EP
Entry Point (to the distribution system)
EPA
U.S. Environmental Protection Agency
EPCRA
Emergency Planning and Community Right-to-Know Act
FP
Formation Potential
GAC
Granular Activated Carbon
GC
Gas Chromatography
GWR
Ground Water Rule
GWUDI
Ground Water Under the Direct Influence of Surface Water
HAA
Haloacetic Acid
HPLC
High Performance Liquid Chromatography
HRL
Health Reference Level
HSDB
Hazardous Substances Data Bank
IARC
International Agency for Research on Cancer
LAA
Locational Annual Average
LCMRL
Lowest Concentration Minimum Reporting Level
MDBP
Microbial and Disinfection Byproduct
MDL
Method Detection Limit
MEA
Methyl ethyl amine
MF
Microfiltration
Six-Year Review 3
Technical Support Document for Nitrosamines
x
December 2016
-------
MIEX
Magnetic Ion Exchange Resin
MOA
Mode of Action
MR
Maximum Residence (location in the distribution system)
MRL
Minimum Reporting Limit
MSR
Method Sensitivity Ratio
NDBA
A-Nitroso-di-n-butyl amine
NDEA
A-Nitrosodi ethyl amine
NDMA
A-Nitrosodi methyl amine
NDPA
A-Nitrosodi propyl amine
NDPhA
A-Nitrosodi phenyl amine
NF
Nanofiltration
NMEA
A^-Nitrosom ethyl ethyl amine
NMOR
A-Nitrosomorpholine
NO
Nitric Oxide
NOM
Natural Organic Matter
NPA
n-Propylamine
NPDWR
National Primary Drinking Water Regulation
NPIP
A-Nitrosopi peri dine
NPYR
A-Nitrosopyrrolidine
NTNCWS
Non-Transient Non-Community Water System
PAC
Powdered Activated Carbon
PBT
Persistence, Bioaccumulation and Toxicity
PWS
Public Water System
PYR
Pyrrolidine
RO
Reverse Osmosis
RSD
Relative Standard Deviation
SDWIS
Safe Drinking Water Information System
SM
Standard Methods
SPE
Solid Phase Extraction
SUVA
Specific Ultraviolet Absorbance
SYR
Six-Year Review
SYR3
Third Six-Year Review
THM
Trihalomethane
TMA
Trimethylamine
TOC
Total Organic Carbon
TONO
Total N-nitrosamines
TRI
Toxics Release Inventory
UCMR2
Second Unregulated Contaminant Monitoring Regulation
UDMH
Unsymmetrical Dimethylhydrazine
UF
Ultrafiltration
USEPA
U.S. Environmental Protection Agency
UV
Ultraviolet
UVA
Ultraviolet Absorbance
Six-Year Review 3
Technical Support Document for Nitrosamines
xi
December 2016
-------
1 Introduction
The Safe Drinking Water Act requires the United States Environmental Protection
Agency (EPA) to conduct a periodic review of existing National Primary Drinking Water
Regulations (NPDWRs) and determine which, if any, are candidates for revision. The purpose of
the review, called the Six-Year Review (SYR), is to evaluate current information for each
NPDWR to determine if there is new information on: health effects, treatment technology,
analytical methods, occurrence and exposure, implementation, and/or other factors that provide a
health or technical basis to support a regulatory revision that will improve or strengthen public
health protection.
Under the Third Six-Year Review (SYR3), EPA is reviewing the regulated chemical,
radiological and microbiological contaminants included in previous reviews, as well as the
Microbial and Disinfection By-Products (MDBP) regulations that were promulgated under the
following actions: the Disinfectants and Disinfection By-Products Rules (D/DBPRs), the Surface
Water Treatment Rules, the Ground Water Rule (GWR) and the Filter Backwash Recycling
Rule. The Surface Water Treatment Rules consist of the Surface Water Treatment Rule, the
Interim Enhanced Surface Water Treatment Rule, the Long Term 1 Enhanced Surface Water
Treatment Rule (LT1) and the Long Term 2 Enhanced Surface Water Treatment Rule (LT2).
This is the first time that EPA is reviewing the MDBP rules. For more information about the
SYR of the D/DBPRs, the reader is referred to EPA's Six-Year Review 3 Technical Support
Document for the Disinfectants/Disinfection Byproducts Rules (USEPA, 2016a). Under the
SYR3, EPA also is evaluating unregulated disinfection by-products (DBPs): for example,
nitrosamines and chlorate.
In the Federal Register notice for Preliminary Regulatory Determination 3 (USEPA,
2014a), the Agency stated that "because chlorate and nitrosamines are DBPs that can be
introduced or formed in PWSs partly because of disinfection practices, the Agency believes it is
important to evaluate these unregulated DBPs in the context of the review of the existing DBP
regulations. DBPs need to be evaluated collectively, because the potential exists that the
chemical disinfection used to control a specific DBP could affect the concentrations of other
DBPs. Therefore, the Agency is not making a regulatory determination for chlorate and
nitrosamines at this time."
Due to the limitations of available analytical methods, only six nitrosamines were
included in EPA's Second Unregulated Contaminant Monitoring Rule (UCMR 2) (monitored
using EPA Method 521) and evaluated together as candidates for regulation under the Third
Regulatory Determinations program in 2014. They are: A-nitrosodi-n-butylamine (NDBA), N-
nitrosodiethylamine (NDEA), N-nitrosodimethylamine (NDMA), A'-nitrosodi-n-propylamine
(NDPA), A^-nitrosom ethyl ethyl amine (NMEA) and A-nitrosopyrrolidine (NPYR). Four of these
nitrosamines (NDEA, NDMA, NDPA, and NPYR) were included on EPA's Third Contaminant
Candidate List (CCL 3).1
1 An additional nitrosamine, A -nitrosodiplienv 1 ami nc (NDPhA), was also on the CCL 3, but was not included in
UCMR 2 because of the lack of an adequate analytical method.
Six-Year Review 3
Technical Support Document for Nitrosamines
December 2016
1-1
-------
Nitrosamines are a class of nitrogen-containing organic compounds that share a common
nitrosamino functional group. Since nitrosamines have increasingly attracted attention in both
field and laboratory studies, a considerable amount of information on the formation, fate,
occurrence and health effects of this group of compounds in water has become available. Of the
six nitrosamines covered in this document, NDMA is the focus of discussion in many chapters.
This is because NDMA is the nitrosamine contaminant with the most information available and
with the highest occurrence in drinking water. However, it should be noted that NDMA may
account for only approximately five percent of total nitrosamines in chloraminated drinking
waters, where it tends to occur most (see Section 5.1).
The remainder of this document provides a summary of available information and data
relevant to EPA's understanding of the: contaminant background, health effects, analytical
methods, occurrence and exposure, formation and treatment/control strategies for nitrosamines.
The information cutoff date for the SYR3 was December 2015. That is, information published
after December 2015 was not considered for this document. The Agency recognizes that
scientists and other stakeholders are continuing to investigate and publish information relevant
information subsequent to the cutoff date. While not considered as part of the SYR3, the Agency
anticipates providing consideration for that additional information in subsequent activities.
Six-Year Review 3
Technical Support Document for Nitrosamines
1-2
December 2016
-------
2 Contaminant Background
This chapter presents background information on six unregulated nitrosamine compounds that
EPA is evaluating under the SYR3 program: A'-nitrosodi-n-butyl amine (NDBA), N-
nitrosodiethylamine (NDEA), A'-nitrosodi methyl amine (NDMA), A-nitrosodi-n-propyl amine
(NDPA), A^-nitrosom ethyl ethyl amine (NMEA) and A'-nitrosopyrrolidine (NPYR). The following
topics are covered in the chapter: physical and chemical properties; production, use and release;
formation in environmental media; environmental fate and transport; and regulatory and non-
regulatory actions.
2.1 Physical and Chemical Properties
Nitrosamines share a common structure, illustrated in Exhibit 2.1. In the case of the nitrosamines
discussed in this report, the Ri and R2 substituents/side chains are normal alkyl groups or cyclic
moieties. NDBA, NDEA, NDMA, and NDPA are symmetrical dialkylnitrosamines, NMEA is an
asymmetrical dialkylnitrosamine, and NPYR is a cyclic nitrosamine.
Exhibit 2.1: Chemical Structure for Nitrosamines
D
Ki
V
0
//
N-
-N
/
r<2
Exhibit 2.2 presents the chemical structures of NDBA, NDEA, NDMA, NDPA, NMEA and
NPYR. Physical and chemical properties and other reference information for these nitrosamines
are listed in Exhibit 2.3. NDMA is the smallest molecule among nitrosamines.
Exhibit 2.2: Chemical Structures of NDBA, NDEA, NDMA, NDPA, NMEA, and NPYR
NDBA
NDEA
NDMA
NDPA
NMEA
NPYR
H,C
/
H3c-
H1C'/Vvn'^SXCK,
i
0^
l-LG N
3 N %
1
GHg
n~o
CH,
H-C N ^0
I
CH3
°W
N — N
Source: ChemlDPIus, 2010
Six-Year Review 3 2-1
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 2.3: Physical and Chemical Properties of NDBA, NDEA, NDMA, NDPA, NMEA, and NPYR
Property
NDBA
NDEA
NDMA
NDPA
NMEA
NPYR
Chemical Abstracts
Service (CAS) Registry
Number
924-16-3
55-18-5
62-75-9
621-64-7
10595-95-6
930-55-2
Chemical Formula
C8H18N2O
C4H10N2O
C2H6N2O
C6H14N2O
C3H8N2O
C4H8N2O
Molecular Weight
158.24 g/mol
(HSDB, 2010;
ChemlDPIus, 2010)
102.14 g/mol
(HSDB, 2010;
ChemlDPIus, 2010)
74.0822 g/mol (Lide,
1995-96)
130 g/mol (RAIS,
2009; HSDB, 2010;
ChemlDPIus, 2010)
88.1 g/mol (HSDB,
2010; ChemlDPIus,
2010)
100.12 g/mol
(HSDB, 2010;
ChemlDPIus, 2010)
Color/Physical State
Yellow oil (HSDB,
2010)
Slightly yellow liquid
(HSDB, 2010)
Yellow liquid at 25 deg
C (O'Neil etal., 2001;
Lewis, 1981)
Yellow liquid
(HSDB, 2010)
Yellow liquid
(HSDB, 2010)
Yellow liquid (HSDB,
2010)
Boiling Point
116 deg C at 14 mm
Hg (HSDB, 2010)
175-177 deg C
(HSDB, 2010)
151-153 deg C (O'Neil
etal., 2001)
206 deg C (RAIS,
2009)
163 deg Cat 747
mm Hg (HSDB,
2010)
214 deg Cat 760
mm Hg (HSDB,
2010)
Melting Point
< 25 deg C
(ChemlDPIus, 2010)
< 25 deg C
(ChemlDPIus, 2010)
-50 deg C (WHO, 2008)
-12-6.6 deg C
(estimated)
(ATSDR, 1989a)
-
-
Vapor Density
0.9009 g/mL at 20
deg C/4 deg C
(HSDB, 2010)
0.9422 g/mL at 20
deg C/4 deg C
(HSDB, 2010)
1.0059 g/mL (Bednar et
al., 2009)
0.916 g/mL (HSDB,
2010)
0.9448 g/mL at 18
deg C/4 deg C
(HSDB, 2010)
1.1 g/mL (HSDB,
2010)
Freundlich Adsorption
Coefficient
-
-
6.9E-05
(mg/g)/(mg/L)1/n (Faust
and Aly, 1998)
-
-
-
Vapor Pressure (Pv)
4.69E-2 mm Hg at
25 deg C
(Extrapolated)
(HSDB, 2010)
0.86 mm Hg at 20
deg C (ChemlDPIus,
2010)
360 Pa at 20 deg C
(International
Occupational Safety
and Health Centre,
2001)
0.086 mm Hg
(HSDB, 2010)
1.100 mm Hg at 25
deg C
(ChemlDPIus,
2010)
0.06 mm Hg at 20
deg C (HSDB, 2010;
ChemlDPIus, 2010)
Henry's Law Constant
1.32E-5 atm-m3/mol
at 37 deg C (HSDB,
2010; ChemlDPIus,
2010)
3.63E-06 atm-m3/mol
(HSDB, 2010;
ChemlDPIus, 2010)
1.82E-06 atm-m3/mol at
37 deg C (HSDB, 2010;
ChemlDPIus, 2010)
5.38E-06 atm-
m3/mol at 37 deg C
(ChemlDPIus,
2010; HSDB, 2010)
1.44E-06 atm-
m3/mol at 25 deg C
(ChemlDPIus,
2010)
4.89E-08 atm-m3/mol
at 37 deg C (HSDB,
2010; ChemlDPIus,
2010)
Six-Year Review 3
Technical Support Document for Nitrosamines
2-2
December 2016
-------
Property
NDBA
NDEA
NDMA
NDPA
NMEA
NPYR
Log Kow
2.63 (dimensionless)
(HSDB, 2010;
ChemlDPIus, 2010)
0.48 (dimensionless)
(HSDB, 2010;
ChemlDPIus, 2010)
-0.57 (dimensionless)
(International
Occupational Safety
and Health Centre,
2001; HSDB, 2010;
ChemlDPIus, 2010)
1.36
(dimensionless)
(HSDB, 2010;
ChemlDPIus, 2010)
0.04
(dimensionless)
(HSDB, 2010;
ChemlDPIus, 2010)
-0.19
(dimensionless)
(HSDB, 2010;
ChemlDPIus, 2010)
Koc
642 L/kg, estimated
(HSDB, 2010)
43 L/kg, estimated
(HSDB, 2010)
12 L/kg (HSDB, 2010)
130 L/kg, estimated
(HSDB, 2010)
25 L/kg, estimated
(HSDB, 2010)
19 L/kg, estimated
(HSDB, 2010)
Water Solubility (Csat)
1,270 mg/L at 24
deg C (HSDB, 2010;
ChemlDPIus, 2010)
106,000 mg/L at 24
deg C (ChemlDPIus,
2010)
1,000,000 mg/L at 24
deg C (ChemlDPIus,
2010)
13,000 mg/L (RAIS,
2009; ChemlDPIus,
2010)
300,000 mg/L at 20
deg C
(ChemlDPIus,
2010)
1,000,000 mg/L at
24 deg C
(ChemlDPIus, 2010)
Solubility in Other
Solvents
Organic solvents &
vegetable oils
(HSDB, 2010)
Alcohol, ether,
organic solvents, &
lipids (HSDB, 2010)
Alcohols, ether, organic
solvents, and lipids
(ACGIH, 1991; O'Neil et
al., 2001)
Alcohol, ether, and
other organic
solvents (ATSDR,
1989a)
Organic solvents &
lipids (HSDB,
2010)
Organic solvents &
lipids (HSDB, 2010)
Conversion Factors (at
20 deg C, 1 atm)
1 ppm = 6.58 mg/m3
1 mg/m3 = 0.152
ppm
(calculated at 20 deg
C)
1 ppm = 4.25 mg/m3
1 mg/m3 = 0.235
ppm
(calculated at 20 deg
C)
1 ppm = 3.08 mg/m3; 1
mg/m3 = 0.325 ppm
(Verscheuren, 1996)
1 ppm (v/v) = 5.41
mg/m3; 1 mg/m3 =
0.185 ppm (v/v)
(ATSDR, 1989a)
1 ppm = 3.66
mg/m3
1 mg/m3 = 0.273
ppm
(calculated at 20
deg C)
1 ppm = 4.16 mg/m3;
1 mg/m3 = 0.240
ppm
(calculated at 20 deg
C)
Six-Year Review 3
Technical Support Document for Nitrosamines
2-3
December 2016
-------
2.2 Production, Use and Release
This section presents information about the production, use and release of members of the
nitrosamine group.
2.2.1 Production and Use
Each of the six nitrosamines can be produced as an unintended byproduct of manufacturing
processes that involve the use of nitrite or nitrate and amines. For example, nitrosamines may
form at tanneries; fish processing plants; foundries; and pesticide, dye, rubber, and tire
manufacturing plants. Nitrosamines have been found in tobacco products, cured meat, ham,
bacon, beer, whiskey, fish, cheese, canned fruit, soybean oil, toiletries, household cleaners,
pesticides, rubber baby bottle nipples and pacifiers, and drugs formulated with aminopyrine
(ATSDR 1989b; NTP, 2005b; Fine et al. 1977; Yurchenko and Molder, 2007; Drabik-
Markiewicz et al. 2009; Perez et al. 2008). NDMA is currently used only in research but was
once used in the production of rocket fuel, as a solvent and as a rubber accelerator (an aid to
vulcanization in the manufacturing process). It was also used or proposed for use as an
antioxidant, an additive for lubricants, and a softener for copolymers (ATSDR, 1989b). NDMA
is no longer produced commercially in the United States (HSDB, 2010).
NDEA is produced commercially. It is used as an additive in gasoline and in lubricants, as an
antioxidant and as a stabilizer in plastics, although no data are available on quantities produced
for commerce (HSDB, 2010).
NDBA, NDPA, NMEA and NPYR are not produced commercially but may be produced in small
quantities for research purposes (HSDB, 2010).
No production data on any of the six compounds are available from EPA's Inventory Update
Reporting or Chemical Data Reporting program.
2.2.2 Environmental Release
EPA's Toxics Release Inventory (TRI) and other sources provide information about industrial,
commercial and consumer releases of NDBA, NDEA, NDMA, NDPA, NMEA and NPYR. That
information is summarized in this section. In addition, this section discusses releases from
municipal wastewater.
2.2.2.1 Industrial and Commercial Releases
Toxics Release Inventory (TRI)
EPA established TRI in 1987 in response to Section 313 of the Emergency Planning and
Community Right-to-Know Act (EPCRA). EPCRA Section 313 requires the reporting of annual
information on toxic chemical releases from facilities that meet specific criteria. This reported
information is maintained in a database accessible through TRI Explorer (USEPA, 2012).
Although TRI can provide a general idea of release trends, it has limitations. Not all facilities are
required to report all releases. Facilities are required to report releases, both on-site and off-site,
if they manufacture or process more than 25,000 pounds of a chemical or use more than 10,000
Six-Year Review 3 2-4
Technical Support Document for Nitrosamines
December 2016
-------
pounds. On-site releases are subdivided by environmental media. Reporting requirements have
changed over time (e.g., reporting thresholds have decreased); this creates the potential for
misleading data trends over time (USEPA, 1996a). TRI data are meant to reflect releases and
should not be used to estimate general public exposure to a chemical (USEPA, 2004a). For the
purposes of TRI, "State" counts include the District of Columbia and U.S. territories in addition
to the 50 states.
TRI contains release data for NDBA, NDEA, NDMA and NDPA. No data on NMEA and NPYR
releases are available from TRI. Neither of these nitrosamines has ever been included in the TRI
list of chemicals for which reporting is required.
TRI data for NDMA are reported for the years 1998 through 2000 and 2005 through 2007. The
only non-zero reported releases of NDMA during those years were air emissions reported from
South Carolina in 1998 and California in 1999, in the amounts of 129 pounds and 5 pounds,
respectively (USEPA, 2012). More recently, amine-based carbon dioxide capture systems used
for post-combustion carbon sequestration could have released some nitrosamines into the
environment (Dai et al., 2012; Dai and Mitch, 2015).
Exhibit 2.4 through Exhibit 2.6 summarize TRI data for NDBA, NDEA and NDPA, respectively,
from the years 1998 to 2010. Though there were significant (up to almost 10,000 pounds
annually) on-site releases of NDEA to land in 1999 and 2001, off-site releases dominated the
total releases of NDEA. Off-site releases also dominate reported releases of NDPA and NDBA.
TRI reports on-site air emissions of NDPA in the range of hundreds of pounds annually, and
smaller on-site releases of NDBA in the form of air emissions and surface water discharges.
Exhibit 2.4: Environmental Releases (in pounds) of NDBA in the United States,
1998-2010
Year
On-Site
Air Emissions
On-Site
Surface
Water
Discharges
On-Site
Underground
Injection
On-Site
Releases
to Land
Off-Site
Releases
Total On- &
Off-Site
Releases
1998
-
-
-
-
-
0
1999
2
1
0
0
4
7
2000
5
0
0
0
10
15
2001
5
0
0
0
4,505
4,510
2002
0
0
0
0
500
500
2003
0
0
0
0
255
255
2004
0
0
0
0
5
5
2005
1
0
0
0
500
501
2006
0
0
0
0
500
500
2007
0
0
0
0
500
500
2008
0
0
0
0
185
185
2009
0
0
0
0
398
398
2010
0
0
0
0
383
383
Note:represents releases that fell below minimum reporting thresholds, or non-responses from facilities, or information that was
not required to be reported in a particular year (as reporting forms changed overtime).
Six-Year Review 3 2-5
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 2.5: Environmental Releases (in pounds) of NDEA in the United States,
1998-2010
Year
On-Site
Air Emissions
On-Site
Surface
Water
Discharges
On-Site
Underground
Injection
On-Site
Releases
to Land
Off-Site
Releases
Total On- &
Off-Site
Releases
1998
2
-
0
0
0
2
1999
30
1
0
7,640
4,124
11,795
2000
5
0
0
0
10
15
2001
234
0
0
9,959
7,597
17,790
2002
0
0
0
0
500
500
2003
0
0
0
0
255
255
2004
0
0
0
0
1,000
1,000
2005
0
0
0
0
500
500
2006
0
0
0
0
500
500
2007
0
0
0
0
500
500
2008
0
0
0
0
650
650
2009
0
0
0
0
651
651
2010
0
0
0
0
633
633
Note:represents releases that fell below minimum reporting thresholds, or non-responses from facilities, or information that was
not required to be reported in a particular year (as reporting forms changed overtime).
Exhibit 2.6: Environmental Releases (in pounds) of NDPA in the United States,
1998-2010
Year
On-Site
Air Emissions
On-Site
Surface
Water
Discharges
On-Site
Underground
Injection
On-Site
Releases
to Land
Off-Site
Releases
Total On- &
Off-Site
Releases
1998
879
-
0
0
1,500
2,379
1999
5
-
0
0
-
5
2000
2
-
0
0
0
2
2001
6
0
0
0
505
511
2002
2
0
0
0
255
257
2003
1
0
0
0
257
258
2004
251
0
0
0
255
506
2005
252
0
0
0
503
755
2006
251
0
0
0
500
751
2007
251
0
0
0
500
751
2008
93
0
0
0
330
423
2009
100
0
0
0
328
428
2010
67
0
0
0
315
382
Note:" represents releases that fell below minimum reporting thresholds, or non-responses from facilities, or information that was
not required to be reported in a particular year (as reporting forms changed overtime).
Exhibit 2.7 through Exhibit 2.9 present the TRI total releases and total surface water discharges
of NDBA, NDEA and NDPA, respectively, for the years 2002, 2004, 2006, 2008 and 2010. No
surface water discharges were reported for any of these three contaminants. Reported total
Six-Year Review 3 2-6
Technical Support Document for Nitrosamines
December 2016
-------
releases of these compounds during the years listed came consistently from Ohio (all three
compounds in all years) and Indiana (NDPA in all years).
Exhibit 2.7: Summary of Total Releases and Total Surface Water Discharges (in
pounds) of NDBA in 2002, 2004, 2006, 2008 and 2010
Year
Total Releases
Count of States with
Releases
Total Surface Water
Discharges
Count of States with
Surface Water
Discharges
2002
500
1
0
0
2004
5
1
0
0
2006
500
1
0
0
2008
185
1
0
0
2010
383
1
0
0
Exhibit 2.8: Summary of Total Releases and Total Surface Water Discharges (in
pounds) of NDEA in 2002, 2004, 2006, 2008 and 2010
Year
Total Releases
Count of States with
Releases
Total Surface Water
Discharges
Count of States with
Surface Water
Discharges
2002
500
1
0
0
2004
1,000
1
0
0
2006
500
1
0
0
2008
650
1
0
0
2010
633
1
0
0
Exhibit 2.9: Summary of Total Releases and Total Surface Water Discharges (in
pounds) of NDPA in 2002, 2004, 2006, 2008 and 2010
Year
Total Releases
Count of States with
Releases
Total Surface Water
Discharges
Count of States with
Surface Water
Discharges
2002
257
2
0
0
2004
506
2
0
0
2006
751
2
0
0
2008
423
2
0
0
2010
382
2
0
0
Additional Information on Industrial Releases
NDMA may be present in waste streams from manufacturing facilities where the compound was
inadvertently generated as a byproduct. Such facilities may include amine manufacturing plants,
tanneries, food processing industries, foundries, and manufacturers of rubber and tires, rocket
fuel, pesticides, dyes, soaps, detergents, surfactants, lubricants and copolymers (ATSDR, 1989b;
HSDB, 2010; WHO, 2008). Such releases are typically to surface water, although atmospheric
releases are also a concern (ATSDR, 1989b; WHO, 2008) and ground water contamination was
observed at a rocket fuel manufacturing facility (Fleming et al., 1996). NDMA has been
Six-Year Review 3 2-7
Technical Support Document for Nitrosamines
December 2016
-------
identified as both an impurity and a breakdown product of hydrazine-based rocket fuels (Rajat,
2008).
Industrial releases of NDBA, NDEA and NPYR have been detected in waste streams from
rubber manufacturers and chemical plants. For example, 132 ng/L of NDEA was found in
receiving river waters from a chemical plant (HSDB, 2010). NDBA can be formed as a waste
product and has been detected in ambient air in rubber manufacturing plants and in factories
using metal-working fluids (HSDB, 2010). Concentrations of NDEA ranging from 0.04 to 0.39
|ig/m3 were also found in the air in the passenger compartment of new 1979 cars (HSDB, 2010).
NPYR has been detected in air samples from rubber tire plants and in effluent water from
chemical factories. The average concentration of NPYR in the air of seven rubber manufacturers
was 0.26 |ig/m3, NPYR concentrations detected in effluent water from multiple chemical plants
are reported to have ranged from 0.02 to 0.09 |ig/L (20-90 ng/L) (HSDB, 2010).
Commercial and Consumer Releases
Nitrosamine compounds have been found in tobacco products, cured meats, ham, bacon, beer,
whiskey, fish, cheese, soybean oil, toiletries, household cleaners, pesticides, and rubber baby
bottle nipples and pacifiers (ATSDR, 1989b; NTP, 2011; Fine et al., 1977; Yurchenko and
Molder, 2007; Drabik-Markiewicz et al., 2009; Perez et al., 2008). NPYR can be found in
processed meats and spice premixes (HSDB, 2010). Consumption and disposal of food items,
cigarettes and other products contaminated with NDMA and other nitrosamines may lead to
environmental releases.
NDBA, NDEA, NMEA and NPYR have all been found in tobacco smoke. Higher concentrations
were found in tobacco grown in high-nitrogen soil (HSDB, 2010). NDMA has been detected in
the exhaust of diesel vehicles (Health Canada, 2011). Studies have found that rubber gaskets may
leach NDMA into drinking water distribution systems (Morran et al., 2011; Teefy et al., 2011).
NDPA has been found as a contaminant in substituted dinitrotrifluralin herbicides (HSDB,
2010). NDMA can be present as a contaminant in dimethylamine (DMA)-based pesticides such
as bromacil, benzolin, 2,4-D, dicamba, 2-methyl-4-chlorophenoxyacetic acid, and mecoprop.
Use of such pesticides in agricultural settings, hospitals and homes can lead to environmental
release of NDMA (WHO, 2008). Testing of over 100 samples of DMA-formulated phenoxy acid
herbicides in Canada since 1990 revealed that NDMA was present in 49 samples, with an
average concentration of 0.44 |ig/g and maximum concentration of 2.32 jug/g. Six samples
contained concentrations above 1.0 |ig/g (Health Canada, 2011). On the whole, concentrations of
NDMA in pesticides appears to be decreasing over time (WHO, 2008).
2.2.2.2 Releases from Wastewater Treatment Facilities
NDMA is commonly present in municipal sewage sludge, and has been measured at
concentrations ranging from 0.6 to 45 parts per billion (ppb) (ATSDR, 1989b; HSDB, 2010). To
a lesser extent, NPYR has also been observed in sewage sludge (HSDB, 2010). Investigators
attribute nitrosamine formation in sewage sludge to interaction between alkylamines and nitrite
(ATSDR, 1989b).
Six-Year Review 3 2-8
Technical Support Document for Nitrosamines
December 2016
-------
2.3 Formation in Environmental Media
Nitrosamines may form in air, soil, water, sewage, food, and animal and microbial systems
where precursors (e.g., amines and nitrite) are present. Formation in those media is discussed in
this section. Formation of the nitrosamines as disinfection by-products (DBPs) in drinking water
is discussed in Chapter 6.
2.3.1 NDMA
NDMA can form in small quantities in air, water and soil as a result of biological, chemical and
photochemical processes. The precursor DMA and other precursors occur naturally in the
environment (e.g., in plants, fish, algae, urine and feces), and may be introduced by human
activity as well (e.g., via pesticides) (ATSDR, 1989b).
Biological formation involves the reaction of secondary or tertiary amines with nitrite. The nitrite
can be produced by microbial action from ammonia or nitrate, or it may be of anthropogenic
origin (ATSDR, 1989b). NDMA can also be produced endogenously in humans from the
interaction of nitrates and nitrites with amines in the stomach (Mirvish 1974, 1992; Tricker et al.,
1994). Fristachi and Rice (2007) estimated that the mean endogenous formation for adults was
about 20 |ig/day.
Chemical formation occurs when primary, secondary or tertiary amines react with nitrite. Acidic
conditions are most favorable. DMA has been found to react at night in the atmosphere with
oxides of nitrogen to form NDMA (ATSDR, 1989b).
Photochemical formation of NDMA from DMA and nitrite, unlike chemical formation, occurs
most readily under alkaline conditions (ATSDR, 1989b).
2.3.2 Other Nitrosamines
Limited data are available on the formation of other nitrosamines in environmental media.
NDEA may be formed by nitrate-reducing bacteria. In addition, a study concluded that formation
of NDEA in river water samples at concentrations ranging from 0.13 to 7.02 |ig/L (130-7,020
ng/L) was due to the reaction of nitrite with two different tracer dyes (Rhodamine B and
Rhodamine WT) (HSDB, 2010).
NMEA may form in the atmosphere at night (i.e., while not subject to photodegradation) when
atmospheric amines react with nitrous acid. NMEA can form from methylethyl amine or from
any tertiary amine containing one each of methyl and ethyl functional groups (HSDB, 2010).
Wastewater effluent from industries using amines has led to the formation of nitrosamines in
ocean water and river water (HSDB, 2010).
2.4 Environmental Fate and Transport
This section presents information on the environmental fate and transport of the six nitrosamines.
Information was gathered from EPA's Persistence, Bioaccumulation and Toxicity (PBT) Profiler
and other sources.
Six-Year Review 3 2-9
Technical Support Document for Nitrosamines
December 2016
-------
The nitrosamines are subject to a variety of natural processes when present in soil and water,
including volatilization, photodegradation and microbial degradation. Under some
circumstances, nitrosamines may persist in ambient waters that are used as drinking water
sources, especially ground waters, at levels that could result in contamination of finished
drinking water. However, as indicated by the occurrence monitoring data in Chapter 5 and the
literature review of formation in Chapter 6, the formation of nitrosamines (particularly NDMA)
during certain disinfection processes is understood to be the major source of nitrosamines found
in finished drinking water.
2.4.1 Persistence, Bioaccumulation and Toxicity (PBT) Profiler
EPA developed the PBT Profiler to serve as a screening tool for estimating the percentage of a
contaminant that is predicted to partition to water, air, soil, and sediment in a four-compartment
system as well as the half-life of the contaminant in each medium (biodegradation half-life in the
case of water, soil, and sediment, and half-life based on photochemical reactions with hydroxyl
radicals and ozone in the case of air). Exhibit 2.10 presents the results of PBT Profiler modeling
for NDBA, NDEA, NDMA, NDPA, NMEA and NPYR (PBT Profiler, 2010).
Exhibit 2.10: PBT Profiler Data for Six Nitrosamines
PBT Profiler Data
NDBA
NDEA
NDMA
NDPA
NMEA
NPYR
% partition to water
26%
53%
52%
44%
52%
48%
% partition to soil
72%
46%
44%
55%
46%
51%
% partition to air
1%
2%
4%
1%
2%
1%
% partition to sediment
0%
0%
0%
0%
0%
0%
Half-life in water
15 days
38 days
38 days
38 days
38 days
38 days
Half-life in soil
30 days
75 days
75 days
75 days
75 days
75 days
Half-life in air
0.58 days
0.92 days
6.2 days
0.67 days
1.6 days
1 day
Half-life in sediment
140 days
340 days
340 days
340 days
340 days
340 days
Source: PBT Profiler, 2010
2.4.1.1 Additional Information on Environmental Fate and Transport
NDBA
Based on partitioning coefficients, the Hazardous Substances Data Bank (HSDB) reports that
NDBA will have "low mobility" in soil. The Henry's Law constant for NDBA suggests that
volatilization from moist soil will be an important fate process. The vapor pressure value
indicates that NDBA is not expected to volatilize from dry soils (HSDB, 2010).
Six-Year Review 3 2-10
Technical Support Document for Nitrosamines
December 2016
-------
Partitioning coefficients suggest that NDBA may adsorb to suspended solids and sediment in
water. The Henry's Law constant for NDBA suggests that it is expected to volatilize from water
surfaces. A modeling study suggests NDBA volatilization half-lives of 2.4 days from rivers and
30 days from lakes. Bioconcentration is expected to be moderate for NDBA (HSDB, 2010).
In water, nitrosamines exposed to sunlight photolyze rapidly to amino radicals and nitric oxide.
However, at neutral pH and in the absence of radical scavengers, the amino radicals and nitric
oxide (NO) may recombine. HSDB reports that one study measured NDBA photolysis half-lives
ranging from 16 minutes to 3.6 hours at various pH values. Hydrolysis is not expected to be an
important fate process for NDBA. Reports on the likelihood of nitrosamine biodegradation in
water are mixed (HSDB, 2010).
In the atmosphere, NDBA is expected to exist solely in the vapor phase and to be subject to
degradation by photochemically produced hydroxyl radicals. The estimated atmospheric NDBA
half-life for degradation by photochemically produced hydroxyl radicals is 1.4 hours (HSDB,
2010).
NDEA
Based on partitioning coefficients, HSDB reports that NDEA will have "very high mobility" in
soil. The Henry's Law constant for NDEA suggests that volatilization from moist soil will be an
important fate process. Comparison of vapor pressure values given by HSDB for NDEA and
NMEA suggest that NDEA, like NMEA, may volatilize from dry soils. HSDB reports that 30-80
percent of NDEA applied to a soil surface volatilized within a few hours, while less than 25
percent of NDEA incorporated 7.5 cm below a soil surface volatilized in two days (HSDB,
2010). The half-life for mineralization of NDEA in soil has been measured as slightly longer
than one week (HSDB, 2010).
Partitioning coefficients suggest that NDEA will not adsorb to suspended solids and sediment in
water. The Henry's Law constant for NDEA suggests that it is expected to volatilize from water
surfaces. A model estimates NDEA volatilization half-lives of 10 days from rivers and 78 days
from lakes. Bioconcentration is expected to be low for NDEA (HSDB, 2010).
In water, NDEA exposed to sunlight photolyzes rapidly to amino radicals and nitric oxide.
However, at neutral pH and in the absence of radical scavengers, the amino radicals and NO may
recombine. HSDB reports that one study measured 88.7 percent photolysis of NDEA in solution
after seven hours. NDEA was stable in lake water for 108 days in the absence of light.
Hydrolysis is not expected to be an important fate process for NDEA. One set of authors
identifies three microbes as capable of metabolizing NDEA, but several other studies found no
convincing evidence of NDEA biodegradation (HSDB, 2010).
In the atmosphere, NDEA is expected to exist solely in the vapor phase and to be subject to
degradation by photochemically produced hydroxyl radicals. The estimated atmospheric half-life
for degradation by photochemically produced hydroxyl radicals is 22 hours for NDEA (HSDB,
2010).
Six-Year Review 3 2-11
Technical Support Document for Nitrosamines
December 2016
-------
NDMA
With a vapor pressure of approximately 2.7 mm Hg, significantly higher than the threshold for
partition to the atmospheric particulate phase (10"4 mm Hg), NDMA is expected to be present in
the atmosphere only in the vapor phase (ATSDR, 1989b). NDMA is not expected to persist long
in the atmosphere, being subject to rapid photolysis to DMA in direct sunlight. The half-life for
photolysis has been estimated as 5-30 minutes (ATSDR, 1989b) and 30-60 minutes (WHO,
2008). NDMA is also subject to degradation by hydroxyl radicals, with a half-life of
approximately 6.3 days (HSDB, 2010) or between 25.4 and 254 hours (WHO, 2008). ATSDR
(1989b) concludes that degradation by hydroxyl radicals or ozone would be too slow to be
environmentally significant.
The high water solubility and low soil organic carbon-water partitioning coefficient (K OC )of
NDMA suggest that it is mobile in soils and can enter ground water. High vapor pressure
suggests that NDMA on a dry soil surface will be subject to volatilization (HSDB, 2010; Health
Canada, 2011); the half-life for volatilization has been estimated as 1-2 hours (ATSDR, 1989b).
Photolysis should also be significant at the soil surface. In the subsurface, microbial degradation
may be significant (ATSDR, 1989b). Aerobic conditions may be more favorable to
biodegradation than anaerobic conditions (WHO, 2002). The half-life for NDMA in aerobic soil
was found to be about three weeks under laboratory conditions, with volatilization and
biodegradation dominating (HSDB, 2010). The rate and extent of NDMA removal/degradation
appears to correlate strongly with the amount of organic matter present in sandy loam soils
(HSDB, 2010).
NDMA readily dissolves in water and is unlikely to adsorb to particulates (WHO, 2008). In
surface water, photodegradation and volatilization are the most important fate processes. In the
laboratory, a photodegradation half-life of 79 hours was measured for NDMA in distilled water
(HSDB, 2010). Under environmental conditions, photodegradation may be slowed or inhibited
by suspended solids, organic material, or ice cover (WHO, 2008). Based on the measured
Henry's Law constant of 1.82 x 10"6 atm-m3/mol, modeling indicates an estimated volatilization
half-life of 17 days from a river (1 m deep, flowing 1 m/sec, wind velocity of 3 m/sec) and 130
days from a lake (1 m deep, flowing 0.05 m/sec, wind velocity of 0.5 m/sec) (HSDB, 2010).
Oxidation, hydrolysis, biotransformation and biodegradation are not expected to be significant
fate processes in surface water (WHO, 2008; Health Canada, 2011). When lake water samples
were stored in the dark for 3.5 months at 30 degrees C, no change in NDMA concentration was
observed (HSDB, 2010).
In ground water, where photolysis and volatilization do not occur, microbial transformation may
be the most important fate process (ATSDR, 1989b). Biodegradation has been observed in the
laboratory (HSDB, 2010). Assuming aerobic conditions for biodegradation, Howard et al. (1991)
estimated that the half-life of NDMA in ground water would range from 42 to 360 days.
NDMA may be taken up by organisms in the environment. In an experimental setting, lettuce
and spinach plants were found to be capable of absorbing NDMA from sand, soil and water
(ATSDR, 1989b). Once taken up, NDMA may be subject to biotransformation.
Six-Year Review 3 2-12
Technical Support Document for Nitrosamines
December 2016
-------
NDPA
Some wastewater studies have shown biodegradation of NDPA, although other studies on soil
and sewage have not shown any microbial biodegradation. In one wastewater study, the extent of
biodegradation after seven days of incubation was 27 percent, with subcultures yielding
biodegradation of 37, 47 and 50 percent, respectively. Other sludge studies and lake die-away
studies have yielded zero percent degradation. In a static culture flask biodegradability test based
on the biochemical oxygen demand, slow degradation of NDPA (40-50 percent after 28 days)
was observed under the test conditions established (5 mg yeast/L, 5 mg/L test compound)
(HSDB, 2010).
When applied to the soil surface, NDPA will rapidly volatilize; one study demonstrated a 50
percent loss after six hours (HSDB, 2010). However, when incorporated into the soil, NDPA
volatilization is slower. One experiment showed only six percent volatilization after eight days
(HSDB, 2010). Under anaerobic conditions, NDPA degraded slowly in sandy and silty loam with
over half of the applied NDPA remaining after 60 days. Under aerobic conditions, NDPA
dissipated in these two soils to less than 10 percent and 1 percent, respectively, of the initially
applied amount after 69 days. However, much of the dissipation in the aerobic experiment was
attributed to volatilization (Sacher et al., 2008). The dissipation half-life was 21-40 days in a
field study with sandy and silty clay loam soils (HSDB, 2010). NDPA's Koc value is suggestive
of high mobility in soil (HSDB, 2010), and thus NDPA has the potential to leach into ground
water (ATSDR, 1989a). Under aerobic conditions, the half-life for NDPA in subsurface soil is
14-40 days (ATSDR, 1989a). Under anaerobic conditions, the half-life for degradation is 47-80
days (ATSDR, 1989a). Three strains of nonpathogenic microorganisms (Rhizopus oryzae,
Streptococcus cremoris, and Saccaromyces rouxii) were found to degrade 80, 70, and 50 percent
of NDPA, respectively. NDPA was therefore degraded more rapidly than the other nitrosamines
tested (NDEA and NDMA). In cells pre-cultured with NDPA, the degradation of NDPA was
slightly improved (ca. 10 percent higher yield), suggesting that the degradation might rely on an
inducible enzyme (Sacher et al., 2008).
NDPA will photolyze in surface waters; this appears to be the most significant degradation
pathway (HSDB, 2010). Photolytic degradation in lake water shows a half-life of about 2.5 hours
and 90 percent degradation after 8 hours (HSDB, 2010; ATSDR, 1989a). Otherwise, ATSDR
(1989a) reports that NDPA is not expected to undergo abiotic degradation in natural waters.
With a low Henry's Law constant, NDPA is expected to volatilize slowly from surface waters
(HSDB, 2010). In one study, NDPA did not disappear from samples of lake water incubated at
30 deg C over a period of nearly four months (Sacher et al., 2008). NDPA is not expected to
adsorb to sediment in water (ATSDR, 1989a).
Based on the octanol-water partition coefficient (log Kow), a bioconcentration factor of 6 has
been assigned to NDPA, indicating that accumulation in aquatic species will be low (HSDB,
2010).
In the atmosphere, NDPA is expected to exist solely in vapor form (HSDB, 2010). In the vapor
phase, the half-life of NDPA is estimated to be 16 hours, with degradation occurring by reaction
with photochemically produced hydroxyl radicals (HSDB, 2010).
Six-Year Review 3 2-13
Technical Support Document for Nitrosamines
December 2016
-------
NMEA
Based on partitioning coefficients, HSDB reports that NMEA will have "very high mobility" in
soil. Henry's Law constant for NMEA suggests that volatilization from moist soil will be an
important fate process. The vapor pressure value for NMEA indicates that NMEA may volatilize
from dry soils (HSDB, 2010).
Partitioning coefficients suggest that NMEA will not adsorb to suspended solids and sediment in
water. The Henry's Law constant for NMEA suggests that it is expected to volatilize slowly from
water surfaces. A modeling study suggests NMEA volatilization half-lives of 24 days from rivers
and 180 days from lakes. Bioconcentration is expected to be low for NMEA (HSDB, 2010).
In water, nitrosamines exposed to sunlight photolyze rapidly to amino radicals and nitric oxide.
However, at neutral pH and in the absence of radical scavengers, the amino radicals and NO may
recombine. The photolysis half-life for NMEA in aqueous solution exposed to sunlight is
estimated to be 5.8 minutes. Hydrolysis is not expected to be an important fate process for
NMEA. Reports on the likelihood of nitrosamine biodegradation in water are mixed (HSDB,
2010).
In the atmosphere, NMEA is expected to exist solely in the vapor phase and to be subject to
degradation by photochemically produced hydroxyl radicals. The estimated atmospheric half-life
for degradation by photochemically produced hydroxyl radicals is 1.6 days for NMEA. The
photolysis half-life of NMEA exposed to sunlight in the vapor phase is estimated to be 5.8
minutes (HSDB, 2010).
NPYR
Based on partitioning coefficients, HSDB reports that NPYR will have "high mobility" in soil.
The Henry's Law constant for NPYR suggests that volatilization from moist soil will not be an
important fate process. The vapor pressure value indicates that NPYR will not volatilize from
dry soils (HSDB, 2010).
Partitioning coefficients suggest that NPYR will not adsorb to suspended solids and sediment in
water. The Henry's Law constant for NPYR suggests that NPYR will not volatilize from water
surfaces. Bioconcentration is expected to be low for NPYR (HSDB, 2010).
In water, nitrosamines exposed to sunlight photolyze rapidly to amino radicals and nitric oxide.
However, at neutral pH and in the absence of radical scavengers, the amino radicals and NO may
recombine. In the absence of light, NPYR was found to be stable in neutral or alkaline water at
room temperature for more than 14 days. Hydrolysis is not expected to be an important fate
process for NPYR. Reports on the likelihood of nitrosamine biodegradation in water are mixed
(HSDB, 2010).
In the atmosphere, NPYR is expected to exist solely in the vapor phase and to be subject to
degradation by photochemically produced hydroxyl radicals. The estimated atmospheric half-life
for NPYR degradation by photochemically produced hydroxyl radicals is 20 hours (HSDB,
2010)
Six-Year Review 3 2-14
Technical Support Document for Nitrosamines
December 2016
-------
2.5 Regulatory and Non-Regulatory Actions for Nitrosamines
Several U.S. states and foreign agencies have established regulations or advisories to address
nitrosamine contamination of drinking water.
The World Health Organization issued a guideline value (non-enforceable standard) of 100 ng/L
(0.1 ng/L) for NDMA in drinking water, based on the organization's estimated 10"5 cancer risk
level (WHO, 2008). In 2011, Health Canada adopted a maximum acceptable concentration (an
enforceable standard) of 40 ng/L for NDMA in drinking water, based on the agency's estimated
10"5 cancer risk level (Health Canada, 2011). The Australian Drinking Water Guidelines
(Australia NHMRC, 2013) list a health-based guideline value of 100 ng/L for NDMA.
California has created regulatory and advisory levels for NDEA, NDMA, and NDPA in drinking
water (CalEPA, 2006; CDPH, 2013). For NDMA, California has established a public health goal
(a non-enforceable standard) of 3 ng/L, a notification level (an enforceable standard) of 10 ng/L
and a response level (an enforceable standard) of 300 ng/L. At the response level, the California
Department of Public Health recommends removing a contaminated source from service. For
NDEA, the State has issued a notification level of 10 ng/L and a response level of 100 ng/L. For
NDPA, the State has issued a notification level of 10 ng/L and a response level of 500 ng/L.
California has not established public health goals for NDEA and NDPA.
New Jersey has issued ground water quality criteria of 0.7 ng/L for NDMA and 5 ng/L for
NDPA (NJDEP, 2014). These ground water criteria are considered enforceable, but they may be
limited by the State's practical quantitation levels.
Massachusetts has established a regulatory limit of 10 ng/L for NDMA in drinking water
(Massachusetts DEP, 2004). This limit is based on the practical quantitation limit identified by
California as the concentration of NDMA that most analytical laboratories can detect in drinking
water.
Six-Year Review 3 2-15
Technical Support Document for Nitrosamines
December 2016
-------
3 Health Effects
This chapter presents information on the health effects of six nitrosamine compounds that the
U.S. Environmental Protection Agency (EPA) is evaluating under the third Six-Year Review
(SYR3) program: A'-nitrosodi-n-butyl amine (NDBA), A-nitrosodi ethyl amine (NDEA), N-
nitrosodimethylamine (NDMA), A'-nitrosodi-n-propyl amine (NDPA), N-
nitrosomethylethylamine (NMEA) and A-nitrosopyrrolidine (NPYR).
Section 3.1 presents a summary of health effects for each nitrosamine. Those data are used to
derive health reference levels (HRLs), which are presented in Section 3.2. The HRL is a risk-
derived concentration against which occurrence data from public water systems (PWSs) can be
compared to determine if a nitrosamine occurs with a frequency and at levels of public health
concern.
3.1 Health Effects Assessments for Individual Nitrosamine Compounds
The health endpoints that have been most thoroughly investigated for the nitrosamines are
carcinogenicity and genotoxicity. There are very few studies published regarding the systemic,
reproductive, developmental, neurological, or immunological impacts of nitrosamine exposures.
The majority of the epidemiology studies examined the association of total nitrosamine
exposure, mostly dietary, with the increased risk for a variety of tumor types. Several studies are
specific to the relationship between dietary NDMA exposure and cancer (Knekt et al., 1999;
Larsson et al., 2006; Loh et al., 2011; Jakszyn et al., 2011), and others apply to nitrosamines as a
group (e.g., Jakszyn and Gonzalez, 2006; Larsson et al., 2006). The NDMA-specific studies
identify an increased risk for gastrointestinal and bladder tumors.
In accordance with the most recent Guidelines for Carcinogen Risk Assessment (USEPA, 2005a),
EPA (USEPA, 2014a) categorized the six nitrosamine compounds as likely to be carcinogenic to
humans by a mutagenic mode of action based on sufficient evidence of carcinogenicity in animal
studies (Clapp et al., 1968, 1971; Druckrey et al., 1967; Lijinsky, 1987a, 1987b; Peto et al.,
1991a, 1991b). These studies found tumors in multiple organs (predominately liver, esophageal
and lung) in both sexes and in multiple animal species (e.g., rats, mice and hamsters). All of the
six nitrosamines have been determined to cause cancer through a mutagenic MOA because of
DNA adduct formation leading to errors in DNA replication, altered cell proliferation and,
ultimately, tumors (Diaz Gomez et al., 1986; Goto et al., 1999; Jarabek et al., 2009; Souliotis et
al., 1998). The mutagenic MOA is supported by positive findings from mutagenicity and
genotoxicity in vitro and in vivo studies (Gollapudi et al., 1998; Kushida et al., 2000; Martelli et
al., 1988; Robbiano et al., 1996; Tinwell et al., 1994). The EPA classifications of carcinogenicity
are in agreement with those of other institutions like the International Agency for Research on
Cancer (IARC) and the National Toxicology Program.
Fetuses, newborns and infants may be potentially sensitive to the carcinogenic effects of
nitrosamines due to nitrosamines' mutagenic MOA and evidence of transplacental mutagenicity
(Donovan and Smith, 2008; Althoff et al., 1977). Fetuses, infants and children appear to be at
increased risk for mutagenic changes to deoxyribonucleic acid (DNA) from nitrosamine
exposure, based on animal studies that found younger animals more susceptible to the
development of liver tumors than older animals (Peto et al., 1984; Vesselinovitch et al., 1984;
Gray et al., 1991). In addition, habitual consumers of alcohol may be a sensitive population
Six-Year Review 3 3-1
Technical Support Document for Nitrosamines
December 2016
-------
because alcohol increases the metabolism of nitrosamines via a metabolic pathway that leads to
the formation of mutagenic DNA adducts. Co-exposure to ethanol can exacerbate the cancer
effects of nitrosamines in animal studies (McCoy et al., 1986; Anderson et al., 1993; Kamataki et
al., 2002). There are approximately 5 million people in the U.S. who suffer from alcoholism
(O'Day et al., 1998), and these individuals may be at increased risk if co-exposed to nitrosamines
(Verna et al. 1996; Amelizad et al. 1989). Individuals with genetic defects in DNA repair
enzymes (Hannson, 1992) may also be at increased risk.
3.1.1 NDBA
No short-term human case reports or epidemiologic studies were identified for NDBA. However,
some epidemiology studies tentatively associate ingestion of foods containing high levels of total
nitrosamines, including foods containing NDBA, with gastric and esophageal cancer (Jakszyn
and Gonzalez, 2006; Larsson et al., 2006).
No studies on the subchronic, chronic (non-cancer), reproductive, or developmental effects of
NDBA in animals were identified. Hematuria in mice was observed in one study; however, the
study authors noted that the hematuria was related to the developing bladder tumors in the
affected animals (Bertram and Craig, 1970).
Several chronic animal studies demonstrate the carcinogenic effects of NDBA and its
metabolites. Oral administration of NDBA predominantly resulted in liver, esophageal and
bladder tumors in rats, mice and guinea pigs (Druckrey et al., 1967; Bertram and Craig, 1970;
Lijinsky and Reuber, 1983; Tsuda et al., 1987; Nishikawa et al., 2003). NDBA was tested and
found positive for mutagenicity and genotoxicity in a number of in vitro and in vivo assays
(Prival et al., 1979; Negishi and Hayatsu, 1980; Andrews and Lijinsky, 1980; Brambilla et al.,
1981, 1987; Parodi et al., 1983; Araki et al., 1984; Mochizuki et al., 1986; Langenbach, 1986;
Martelli et al., 1988; Janzowski et al., 1989; Negishi et al., 1991; Shu and Hollenberg, 1996;
Kushida et al., 2000; Fujita and Kamataki, 2001).
EPA classified NDBA as likely to be carcinogenic to humans by a mutagenic mode of action
under the EPA (2005a) Guidelines for Carcinogen Risk Assessment, based on evidence of
carcinogenicity in animal studies (USEPA, 2014a). The cancer slope factor (CSF) for NDBA is
0.4 (mg/kg/day)"1, as derived by linear low-dose extrapolation from the point of departure for the
incidence of liver and esophageal and/or pharyngeal tumors in a lifetime diet study in Berlin-
Druckrey (BD) rats (Druckrey et al., 1967). NDBA was determined to cause cancer through a
mutagenic mode of action (MOA) for all but the bladder tumors. The bladder tumors are linked
to metabolites that lack direct DNA alkylating ability and have an undetermined MOA.
Mutagenicity results for the metabolites linked to the bladder tumors are equivocal (Nagao et al.,
1977; Olajos et al., 1978; Pool et al., 1988; Brendler et al., 1992; Janzowski et al., 1994). No
NDBA-specific data quantifying the increased cancer risk due to early-life exposure were
available. Therefore, based on recommendations of the EPA's Supplemental Guidance for
Assessing Susceptibility from Early-life Exposure to Carcinogens (USEPA, 2005b), Age
Dependent Adjustment Factors (ADAFs) and age-specific exposure factors (USEPA, 2011) were
applied in the evaluation of risk from early-life exposures.
The fetus, newborns and infants may be particularly sensitive to the carcinogenic effects of
NDBA as a consequence of early-life exposures, because NDBA has a mutagenic MOA.
Six-Year Review 3 3-2
Technical Support Document for Nitrosamines
December 2016
-------
Tracheal, nasal, larynx and bronchial tumors were exhibited in offspring of hamsters treated
during pregnancy (Althoff et al., 1977). Lofberg and Tjalve (1986) found thatNDBA is
distributed across the placenta, and becomes activated in the fetus during the late stage of
gestation. Fujii et al. (1977) injected infant mice subcutaneously with NDBA or its metabolites,
and a large proportion of the animals were found to develop lung or liver tumors. A study in
mice suggests that males are potentially more sensitive than females to bladder tumors from
NDBA exposure (Bertram and Craig, 1970). However, other studies in different strains of mice
did not identify an increased sensitivity of males (Irving et al., 1984; He et al., 2012). Among the
six nitrosamines monitored under EPA's Second Unregulated Contaminant Monitoring
Regulation (UCMR 2), NDBA has the strongest link to bladder cancer. Habitual consumers of
alcoholic beverages may also be a sensitive population, based on animal study findings that
demonstrated enhanced effects of the other nitrosamines following co-exposures to ethanol
(Griciute et al., 1982; McCoy et al., 1986; Anderson et al., 1993). No data specific to NDBA
were identified in these studies.
3.1.2 NDEA
No short-term human case reports or epidemiologic studies were identified for NDEA. However,
some epidemiology studies tentatively associate ingestion of foods containing high levels of total
nitrosamines, including foods containing NDEA, with gastric or esophageal cancer (Jakszyn and
Gonzalez, 2006; Larsson et al., 2006).
Limited animal studies of short-term, subchronic, or chronic (non-cancer) effects of NDEA were
identified. The few available studies reported adverse effects on liver and kidney enzymes
(Bansal et al., 2000) and increased levels of cholesterol in the liver (Tang et al., 1992). No
studies on the developmental or reproductive effects of NDEA were identified.
Many chronic animal studies demonstrate the carcinogenic effects of NDEA. Oral administration
of NDEA resulted in liver, esophageal, nasal cavity and kidney tumors in rats; liver, esophageal,
lung and forestomach tumors in mice; tracheal, lung, liver, nasal cavity and bronchial tumors in
Syrian golden hamsters; forestomach, esophageal and liver tumors in Chinese hamsters; liver
tumors in guinea pigs and rabbits; liver and nasal cavity tumors in dogs; and liver tumors in
monkeys (IARC, 1978; NTP, 2011). NDEA was tested and found positive for mutagenicity and
genotoxicity in a numerous in vitro and in vivo assays (McCann et al., 1975; Montesano and
Bartsch, 1976; Dean and Senner, 1977; Kuroki et al., 1977; Yahagi et al., 1977; Amacher and
Paillet, 1983; Quillardet and Hofnung, 1985; Mochizuki et al., 1986; Zeiger et al., 1987; Liu and
Guttenplan, 1992; Yamazaki et al., 1992a,b; Vogel and Nivard, 1993; Goto et al., 1999; Aiub et
al., 2003, 2006; Donovan and Smith, 2008).
EPA classified NDEA as likely to be carcinogenic to humans by a mutagenic mode for action
under the USEPA (2005a) Guidelines for Carcinogen Risk Assessment, based on evidence of
carcinogenicity in animal studies (USEPA, 2014a). The CSF for NDEA is 30 (mg/kg/day)"1, as
derived by linear low-dose extrapolation from the point of departure for the incidence of
combined total liver tumors and esophageal tumors in a study in rats (Peto et al., 1991a, b).
NDEA causes cancer through a mutagenic MO A, but there are no NDEA-specific data
quantifying the increased cancer risk due to early-life exposure (Peto et al., 1984; Gray et al.,
1991). Therefore, based on recommendations of the EPA's Supplemental Guidance for Assessing
Susceptibility from Early-life Exposure to Carcinogens (USEPA, 2005b), ADAFs and age-
Six-Year Review 3 3-3
Technical Support Document for Nitrosamines
December 2016
-------
specific exposure factors (USEPA, 2011) were applied in the evaluation of risk from early-life
exposures.
The fetus, newborns and infants may be particularly sensitive to the carcinogenic effects of
NDEA. In mice and hamsters, NDEA exposure of pregnant animals resulted in an increased
incidence of liver and lung tumors in their offspring (Mohr and Althoff, 1965; Mohr et al., 1966;
Anderson et al., 1989). Another study on NDEA found that younger rats were more susceptible
to the development of liver tumors, with a six-fold increase in tumor onset rates compared to rats
exposed as adults (Peto et al., 1984; Gray et al., 1991). Habitual consumers of alcoholic
beverages may also be considered a sensitive population based on animal studies that have
shown an increase in esophageal tumors (Gibel, 1967; Aze et al., 1993) and liver tumors (Takada
et al., 1986) in rats after administration of NDEA and ethanol, compared to animals administered
NDEA alone.
3.1.3 NDMA
Limited human data are available from NDMA poisoning incidents wherein the following
clinical symptoms have been reported: irritation of eyes, skin irritation, irritation of the
respiratory tract, nausea, vomiting, diarrhea, abdominal cramps, headache, sore throat, cough,
weakness, fever, enlarged liver, jaundice and low platelet count (OSHA, 2006). Reports from
cases of human intentional (poisoning) or accidental ingestion identify the liver as the organ of
greatest toxicological concern (Cooper and Kinbrough, 1980; Pedal et al., 1982).
Epidemiological studies reported mixed results on the association between NDMA and several
types of cancer (gastric, lung and brain cancer). Several studies reported a positive association
between NDMA exposure and gastric cancer (Gonzalez et al., 1994; La Vecchia, 1995; Pobel et
al., 1995), while other studies did not report an association (Risch et al., 1985; Knekt et al., 1999;
Straif et al., 2000; Jakszyn et al., 2006). A positive association was noted between NDMA and
upper digestive tract cancers (Rogers et al., 1995), lung cancer (Goodman et al., 1992; De Stefani
et al., 1996), brain cancer in men (Giles et al., 1994) and childhood brain cancer from gestational
exposure through the maternal diet (Preston-Martin, 1989); no association was reported between
NDMA and brain cancer in women (Giles et al., 1994).
NDMA is acutely toxic to rodents when exposure occurs orally and by inhalation at high
concentrations (ATSDR, 1989b; RTECS, 2002). In studies conducted in rats, the liver was the
primary target organ after short-term exposures, with effects observed after 1- and 30-day
exposures at levels as low as 0.7 and 1 mg/kg/day, respectively (Korsrud et al., 1973; Maduagwu
and Bassir, 1980). According to subchronic and chronic studies, the liver appears to be the
primary target organ in rats, mice and hamsters. Observed hepatotoxic effects include
centrilobular congestion, hepatocyte vacuolization and increased serum levels of liver enzymes
(Takayama and Oota, 1965; Otsuka and Kuwahara, 1971; Anderson et al., 1986; Desjardins et
al., 1992). Limited data from reproductive/developmental studies in rats and mice indicate
reduced fetal body weights and increased fetal mortality (Bhattacharyya, 1965; Napalkov and
Aleksandrov, 1968; Nishie, 1983; Anderson et al., 1989; ATSDR, 1989b). However, there are
many limitations to these publications, including lack of controls and insufficient information on
study design and outcomes.
Six-Year Review 3 3-4
Technical Support Document for Nitrosamines
December 2016
-------
Many chronic animal studies demonstrate the carcinogenic effects of NDMA. Oral
administration of NDMA resulted in liver, lung, nasal cavity, bile duct and kidney tumors in rats;
liver, lung and kidney tumors in mice; liver, nasal cavity and bile duct tumors in hamsters; and
liver and bile duct tumors in rabbits and guinea pigs (Tomatis et al., 1964; IARC, 1978; Lijinsky,
1983; Preussmann and Stewart, 1984; ATSDR, 1989b; Peto et al., 1991a,b; IPCS, 2002;
Terracini et al., 1966, 1967). An increased incidence of tumors was reported in the offspring of
rats and mice administered NDMA during pregnancy (Aleksandrov, 1968; Tomatis, 1973;
Althoff et al., 1977; Anderson et al., 1979). NDMA was tested and found positive for
mutagenicity and genotoxicity in a number of in vitro and in vivo assays (ATSDR, 1989b; IPCS,
2002; CalEPA, 2006; Health Canada, 2010).
EPA classified NDMA as likely to be carcinogenic to humans by a mutagenic mode of action
under the USEPA (2005a) Guidelines for Carcinogen Risk Assessment, based on evidence for
human carcinogenicity in epidemiologic studies and substantial animal data demonstrating
carcinogenicity (USEPA, 2014a). The CSF for NDMA is 21 (mg/kg/day)"1, as derived by linear
low-dose extrapolation from the point of departure for the incidence of liver tumors in a study in
rats (Peto et al., 1991a, b). NDMA causes cancer through a mutagenic MOA but there are no
NDMA-specific data quantifying the increased cancer risk due to early-life exposure (Peto et al.,
1984; Gray et al., 1991). Therefore, based on recommendations of the EPA's Supplemental
Guidance for Assessing Susceptibility from Early-life Exposure to Carcinogens (USEPA,
2005b), ADAFs and age-specific exposure factors (USEPA, 2011) were applied in the evaluation
of risk from early-life exposures.
The fetus, newborns and infants can be particularly sensitive to the carcinogenic effects of
NDMA. Several studies have demonstrated an increase in tumors in the offspring of pregnant
animals exposed to NDMA by oral exposure (Aleksandrov, 1968; Anderson et al., 1979;
Tomatis, 1973; Althoff et al., 1977). Habitual consumers of alcoholic beverages may also be a
sensitive population, based on animal studies showing an increased incidence of lung tumors or
tumors of the nasal cavity in mice exposed to NDMA and ethanol, compared to NDMA alone
(IARC, 2010).
3.1.4 NDPA
No short-term human case reports or epidemiologic studies were identified for NDPA. Some
epidemiology studies that examined ingestion of foods containing high levels of total
nitrosamines report an association between ingestion of foods known to contain nitrosamines and
gastric or esophageal cancer (Jakszyn and Gonzalez, 2006; Larsson et al., 2006); those studies,
however, were not specific to NDPA.
Very limited animal data are available on the short-term, subchronic, or chronic (non-cancer)
effects of NDPA. No reproductive studies are available. One developmental study reported
increased mortality in hamster offspring born to dams administered NDPA by subcutaneous
injection (Althoff et al., 1977).
Many chronic animal studies demonstrate the carcinogenic effects of NDPA. Oral administration
of NDPA to rats and mice resulted in liver, nasal cavity, esophageal, tongue, forestomach and
lung tumors and lymphomas (Druckrey et al., 1967; Lijinsky and Taylor, 1978, 1979; Lijinsky
and Reuber, 1981, 1983; Griciute et al., 1982, Griciute and BarauskaTte, 1989). NDPA was tested
Six-Year Review 3 3-5
Technical Support Document for Nitrosamines
December 2016
-------
and found positive for mutagenicity and genotoxicity in a number of in vitro and in vivo assays
(e.g., Montesano and Bartsch, 1976; Brambilla et al., 1981; Shu and Hollenberg, 1997).
EPA classified NDPA as likely to be carcinogenic to humans by a mutagenic mode of action
under the USEPA (2005a) Guidelines for Carcinogen Risk Assessment, based on evidence of
carcinogenicity in animal studies (USEPA, 2014a). The CSF for NDPA is 2 (mg/kg/day)"1, as
derived by linear low-dose extrapolation from the point of departure for the incidence of liver
tumors in a study in rats (Druckrey et al., 1967). As a result of the findings on NDPA
mutagenicity and the MOA information available for the other dialkylnitrosamines, the cancer
recommendations of the EPA's Supplemental Guidance for Assessing Susceptibility from Early-
life Exposure to Carcinogens (USEPA, 2005b) were applied in the evaluation of risk from early-
life exposures.
The fetus, newborns and infants may be particularly sensitive to the carcinogenic effects of
NDPA, since NDPA has a mutagenic MOA. In hamsters, single subcutaneous injections of
NDPA during gestation resulted in respiratory and digestive tract tumors in both the dams and
their offspring (Althoff et al., 1977; Althoff and Grandjean, 1979). Habitual consumers of
alcoholic beverages may also be considered a sensitive population based on animal studies that
have shown an increase in spinocellular, esophageal and forestomach tumors in mice after
administration of NDPA and ethanol, compared to animals administered NDPA in water
(Griciute et al., 1982).
3.1.5 NMEA
No short-term human case reports or epidemiologic studies are available for NMEA. However,
some epidemiology studies tentatively associate ingestion of foods containing high levels of total
nitrosamines, including foods containing NMEA, with gastric or esophageal cancer (Jakszyn and
Gonzalez, 2006; Larsson et al., 2006). An LD50 value of 90 mg/kg in rats suggests that NMEA
has a moderate potential for acute toxicity when administered orally as a single dose (Druckrey
et al., 1967). No studies on the short-term, subchronic, reproductive, or developmental effects of
NMEA are available. Lung inflammation was noted in one chronic oral study in rats (Lijinsky
and Reuber, 1981). These data were from a single-dose study without a control group and are not
suitable for derivation of a reference value.
Multiple chronic animal studies demonstrate the carcinogenic effects of NMEA. Oral
administration of NMEA resulted in liver, esophageal, renal, lung and nasal tumors in rats
(Druckrey et al., 1967; Lijinsky and Reuber, 1980, 1981; Michejda et al., 1986; Lijinsky et al.,
1987) and liver and nasal mucosa tumors in hamsters (Lijinsky et al., 1987). NMEA was tested
and found positive for mutagenicity and genotoxicity in a number of in vitro and in vivo assays
(Phillipson and Ioannides, 1985; Kerklaan et al., 1986; Brambilla et al., 1987; Vogel and Nivard,
1993; Morita et al., 1997; Fujita and Kamataki, 2001).
EPA classified NMEA as likely to be carcinogenic to humans by a mutagenic mode of action the
USEPA (2005a) Guidelines for Carcinogenic Risk Assessment, based on evidence of
carcinogenicity in animal studies (USEPA, 2014a). The CSF for NMEA is 4 (mg/kg/day)"1, as
derived by linear low-dose extrapolation from the point of departure for the incidence of liver
tumors in a study in rats (Druckrey et al., 1967). NMEA has been determined to cause cancer
through a mutagenic MOA, but no chemical-specific data quantifying the increased cancer risk
Six-Year Review 3 3-6
Technical Support Document for Nitrosamines
December 2016
-------
due to early-life exposure were available. Therefore, based on the recommendations of the
EPA's Supplemental Guidance for Assessing Susceptibility from Early-life Exposure to
Carcinogens (USEPA, 2005b), ADAFs and age-specific exposure factors (USEPA, 2011) were
applied in the evaluation of risk from early-life exposures.
The fetus, newborns and infants may be particularly sensitive to the effects of NMEA due to
early-life exposure because NMEA has a mutagenic MO A, however data demonstrating
sensitivity from early life studies are lacking. Habitual consumers of alcoholic beverages may
also be a sensitive population, based on studies that have shown that exposure of animals to
NMEA and ethanol result in increased tumor incidence compared to animals exposed to NMEA
alone (Anderson et al., 1993; McCoy et al., 1986).
3.1.6 NPYR
No short-term human case reports or epidemiologic studies are available for NPYR. However,
some epidemiology studies tentatively associate ingestion of foods containing high levels of total
nitrosamines with gastric and esophageal cancers (Jakszyn and Gonzalez, 2006; Larsson et al.,
2006).
Limited animal data are available on the short-term, subchronic or chronic (non-cancer) effects
of NPYR. Studies in rats and hamsters suggest that NPYR has low potential for acute toxicity
when administered orally as a single dose (Druckrey et al., 1967; Ketkar et al., 1982). In chronic
toxicity studies in animals, oral administration of NPYR is associated with decreased body
weights and survival rates (Greenblatt and Lijinsky, 1972a,b; Chung et al., 1986). No studies on
the reproductive or developmental effects of NPYR were identified.
Many chronic animal studies for NPYR demonstrate the carcinogenic effects of NPYR
(Greenblatt and Lijinsky 1972a,b; Preussmann et al., 1977; Lijinsky and Reuber, 1981; Ketkar et
al., 1982; Peto et al., 1984; Chung et al., 1986; Berger et al., 1987; Gray et al., 1991). Oral
administration of NPYR in rats and hamsters consistently resulted in liver tumors. NPYR was
tested and found positive for mutagenicity and genotoxicity in a number of in vitro and in vivo
assays (Gilbert et al., 1984; Martelli et al., 1988; Zielenska et al., 1990; Vogel and Nivard, 1993;
Kanki et al., 2005; Wang et al., 2007).
EPA classified NPYR as likely to be carcinogenic to humans by a mutagenic mode of action
under the EPA (2005a) Guidelines for Carcinogen Risk Assessment, based on evidence of
carcinogenicity in animal studies (USEPA, 2014a). The CSF for NPYR is 7.0 (mg/kg/day)"1, as
derived by linear low-dose extrapolation from the point of departure for the incidence of liver
tumors in a well-conducted and well-reported lifetime drinking water study in rats (Peto et al.
1984; Gray et al., 1991). NPYR has been determined to cause cancer through a mutagenic MOA,
but no chemical-specific data quantifying the increased cancer risk due to early-life exposure are
available. In the absence of life-stage specific data the cancer recommendations of the EPA's
Supplemental Guidance for Assessing Susceptibility from Early-life Exposure to Carcinogens
(USEPA, 2005b) were applied in the evaluation of risk from early-life exposures based on the
mutagenicity information and the identification of modified DNA bases in studies of exposed
animals.
No data examining the impact of early life exposure to NPYR were identified. However, NPYR-
DNA adducts were detected in liver and kidney DNA of rat pups nursed by mothers exposed to
Six-Year Review 3 3-7 December 2016
Technical Support Document for Nitrosamines
-------
NPYR, indicating exposure via breast milk with subsequent modification of DNA in early life
(Diaz Gomez et al., 1986). Habitual consumers of alcoholic beverages may also be a sensitive
population, based on studies that have shown that exposure of animals to NPYR and ethanol
result in increased tumor incidence compared to animals exposed to NPYR alone (Anderson et
al., 1993; McCoy et al., 1986).
3.1.7 Sensitive Populations
Fetuses, infants and children can be more susceptible than adults to the mutagenic effects of the
six nitrosamines. Although studies are available that indicate an increased risk to fetuses, infants
and children, the data may be inadequate to support developing an alternative to the agency's
default ADAFs. ADAFs are values recommended by EPA (USEPA, 2005b) to adjust CSFs for
chemicals when chemical-specific data that quantify the increased risk are lacking.
Exposure to alcohol along with nitrosamines has been proposed to lead to enhanced carcinogenic
effects (Chhabra et al., 1995). The International Agency for Research on Cancer (IARC) (2010)
reviewed available studies on the impact of NDEA, NDMA or NDPA combined with ethanol,
and found an increased incidence of esophageal, liver, lung, nasal cavity, or forestomach tumors
in the animals given the mixture, compared to the animals administered nitrosamine without
ethanol. Anderson et al. (1993) found a doubling of lung tumor incidence and a 5.5 fold increase
in tumor multiplicity in mice given a mixture of 10 percent ethanol and 40 ppm NPYR (6.6
mg/kg/day) for four weeks in their drinking water, compared to mice given NPYR alone. IARC
(2010) concluded that alcohol ingestion enhanced the carcinogenic potency of nitrosamine
mixtures via induction of CYP2E1, an enzyme that converts nitrosamines to carcinogenic
intermediates. Although it is clear that chronic alcohol exposure can induce the metabolism of
the nitrosamines and the formation of mutagenic metabolites, adequate information may not be
available to justify any toxicity adjustment factor for this effect in human populations, either for
the individual nitrosamines or the group as a whole. Based on the information presented in the
section above for individual compounds, the availability of data on which nitrosamines may
affect sensitive populations is summarized in Exhibit 3.1.
Exhibit 3.1: Sensitive Populations, as Suggested by Data from Animal Studies
Population
NDBA
NDEA
NDMA
NDPA
NMEA
NPYR
Fetus,
infants,
children
Fetus,
Infants,
Children
Fetus,
Infants,
Children
Fetus,
Infants,
Children
Fetus,
Infants,
Children
No data
Infants,
Children
Chronic
alcohol
consumers
No data
Increased
tumors with
mixture
Increased
tumors with
mixture
Increased
tumors with
mixture
No data
Increased
tumors with
mixture
Other potential sensitive subpopulations include individuals with enhanced expression of the
activating cytochrome P450 (CYP) enzymes, individuals with compromised DNA repair
capacity, individuals undergoing cancer therapy, and individuals with enhanced exposure to
Six-Year Review 3 3-8
Technical Support Document for Nitrosamines
December 2016
-------
chemical genotoxins or radiation by medical or industrial exposure routes. However, EPA did
not identify any epidemiology studies regarding increased risk for these groups.
3.2 Calculation of Health Reference Levels (HRLs) for Nitrosamine Compounds
To evaluate the number of PWSs and PWS-served individuals exposed to nitrosamines in
drinking water, as described in more detail in Chapter 5, EPA screened the UCMR 2 monitoring
data for each chemical against several thresholds. These included a health-based threshold called
the HRL. The HRL is a risk-derived concentration in drinking water against which occurrence
data from PWSs can be compared to determine if a nitrosamine occurs with a frequency and at
levels of public health concern. In the case of chemicals that are known or likely to cause cancer,
the HRL is the concentration in drinking water associated with an increased risk of one excess
cancer case above background among a million exposed persons over a lifetime exposure (i.e.,
estimated lifetime excess cancer risk of one-in-a-million, lx 10"6) (USEPA, 2014a). The HRL is
a benchmark that is set to compare the risks of different chemicals based on drinking water being
the sole route of exposure. It does not integrate added risks associated with other exposure media
(e.g., food, air). It does not consider whether or not analytical methods are able to accurately
measure the concentration at the stated level or the feasibility to treat water in a way that reduces
the concentration to that level.
There are two general approaches to the derivation of an HRL for carcinogens. One approach is
used for chemicals that cause cancer and exhibit a linear response to dose and the other applies to
carcinogens evaluated using a non-linear approach. For those contaminants that are considered to
be likely or probable human carcinogens by a mutagenic or unknown MO A, the agency
calculates a toxicity value that defines the relationship between dose and response (i.e., the CSF).
There are two approaches for the derivation of the HRL for cancer effects depending on whether
or not there is information to support a mutagenic MOA. In the absence of data, the dose
response is assumed to be linear but there is no requirement for application of the Supplemental
Guidance for Assessing Susceptibility from Early-life Exposure to Carcinogens (USEPA,
2005b). Application of the supplemental guidance is required only for chemicals with a
demonstrated mutagenic MOA
(1) MOA: Unknown
In cases where the data on the MOA are lacking, EPA typically uses a default low dose linear
extrapolation to calculate a CSF. The unit risk is the estimated upper-bound excess lifetime
cancer risk from a continuous exposure to a chemical at a concentration of 0.001 mg/L in
drinking water and expressed in units of (|ig/L)"', The exposure estimate assumes an adult body
weight of 70 kg and the 90th percentile adult drinking water intake of 2 L/day.
Unit Risk (iig/L)"1 = CSF x [(DWI x UA)/BW]
where:
CSF = Cancer Slope Factor (mg/kg/day)"1
DWI = Drinking Water Intake for an adult, assumed to be 2 L/day (90th
percentile)
Six-Year Review 3 3-9
Technical Support Document for Nitrosamines
December 2016
-------
UA
BW
= Unit adjustment, from mg to (_ig
= Body Weight for an adult, assumed to be 70 kilograms (kg)
The cancer HRL is the concentration of a contaminant in drinking water corresponding to an
excess estimated lifetime cancer risk of one-in-a-million (lx 10"6), calculated as follows:
HRL (|ig/L) = Risk Level of 10"6 ^ Unit Risk (|ig/L)"'
As noted above, HRLs are not final determinations about the level of a contaminant in drinking
water that must not be exceeded to protect any particular population. Rather, HRLs are risk
derived concentrations against which to evaluate the occurrence data during the regulatory
determination process to determine if contaminants occur at levels of potential public health
concern.
(2) MO A: Mutagenic
If the chemical has a mutagenic MO A, low dose linear extrapolation is used to calculate the CSF
as described in the preceding paragraph. The U.S. EPA's 2005 Guidelines for Carcinogen Risk
Assessment (USEPA, 2005a) requires that the potential increased cancer risk due to early-life
exposure be taken into account for chemicals with a mutagenic MOA. When chemical-specific
data to quantify the increased risk are lacking, ADAFs are applied to estimate age-adjusted unit
risks. The age-adjusted unit risk is determined by using the sum of the unit risks for each of the
three ADAF developmental groups (birth to < 2 yrs; 2 yrs to < 16 yrs; 16 yrs to 70 yrs). The age-
adjusted unit risks include a ten-fold adjustment for early life (birth to < 2 yrs) exposures, a
three-fold adjustment for childhood/adolescent (2 yrs to < 16 yrs) exposures, and no additional
adjustment for exposures later in life (16 yrs to 70 yrs), in conjunction with age-specific drinking
water intake values derived from the U.S. EPA's 2011 Exposure Factors Handbook (USEPA,
2011), and the fraction of a 70 year lifetime applicable to each age period. The increase in risk
during early life results from active tissue growth resulting in limited time for repair of DNA
replication errors. The age-adjusted unit risk is the upper-bound excess lifetime cancer risk
estimated to result from continuous postnatal exposure to a chemical at a concentration of 0.001
mg/L in drinking water.
Age-Adjusted Unit Risk (iig/L)"1 = £(CSF x ADAF x DWI/BWR x UA x F)
where:
CSF = Cancer Slope Factor (mg/kg/day)"1
ADAF = The Age Dependent Adjustment Factor for the age group birth to
two-years (ADAF=10), two years to sixteen years (ADAF=3), and
sixteen to seventy years (ADAF=1)
Six-Year Review 3 3-10
Technical Support Document for Nitrosamines
December 2016
-------
DWI/BWR = Drinking Water Intake Body Weight Ratio (DWI/BWR) expressed
as liters per day per kg body weight for the age-specific group (90th
percentile, consumers only)2
UA = Unit adjustment, from mg to (_ig
F = The fraction of a 70 year lifetime applicable to the age period: 2/70
for birth to two years, 14/70 for two years to sixteen years and
54/70 for sixteen years to seventy years
The cancer HRL is the concentration of a contaminant in drinking water corresponding to an
excess estimated lifetime cancer risk of one-in-a-million (lx 10"6), calculated as follows:
HRL (|ig/L) = Risk Level of 10"6 ^ Age-Adjusted Unit Risk (|ig/L)"'
The six nitrosamines had data available to classify them as known or likely human carcinogens
with a mutagenic MOA. Low-dose linear extrapolations and ADAFs were applied to all four of
the CCL 3 nitrosamines: NDMA, NDPA, NDEA and NPYR, as well as the two non-CCL 3
nitrosamines, NMEA and NDBA.
The HRL for each nitrosamine was derived from the CSF using the age-adjusted unit risk. Since
the nitrosamines were determined to cause cancer by way of a mutagenic MOA, the unit risk is
adjusted for the increased risk associated with early life exposures through the application of
ADAFs and age-specific exposure factors. The HRL concentration for each nitrosamine is
presented in Exhibit 3.2. In some cases, the MRLs of the analytical methods used for UCMR 2
are above the HRL. The available data indicate that each of the six nitrosamines presents a
cancer risk with unit risks that span a range of values. Moreover, when multiple nitrosamines
from this group are present in finished water together their individual cancer risks are additive
(Berger et al., 1987). Therefore, EPA finds that the nitrosamines individually or as a group may
have an adverse effect on the health of persons.
Exhibit 3.2: EPA-Derived Cancer Risk Values and HRLs for Six Nitrosamines
Nitrosamines
Studies Used to Establish a
Slope Factor
Cancer
Slope Factor
(mg/kg-day)"1
Age-
Adjusted
Unit Risk
(M9/L)"1
HRL1
(H9/L)
HRL2
(ng/L)
NDBA
Liver and esophageal tumors in
rats (Druckrey et al., 1967)
0.4
3.0 x10"5
3.0 x 10-2
30
NDEA
Liver and esophageal tumors in
rats (Peto et al., 1991 a,b)
30
2.3 x10"3
4.0 x 10"4
0.4
NDMA
Liver tumors in rats (Peto et al.,
1991a,b)
21
1.6 x10"3
6.0 x 10"4
0.6
2 The drinking water intake values were derived from the data in the U.S. EPA's Exposure Factors Handbook
(USEPA, 2011).
Six-Year Review 3 3-11
Technical Support Document for Nitrosamines
December 2016
-------
Nitrosamines
Studies Used to Establish a
Slope Factor
Cancer
Slope Factor
(mg/kg-day)"1
Age-
Adjusted
Unit Risk
(M9/L)"1
HRL1
(H9/L)
HRL2
(ng/L)
NDPA
Liver and esophageal tumors in
rats (Druckrey et al., 1967)
2
1.5 x 10"4
7.0 x 10"3
7
NMEA
Liver tumors in rats (Druckrey
et al., 1967)
4
3.0 x 10"4
3.0 x 10"3
3
NPYR
Liver tumors in rats (Peto et
al., 1984; Gray et al., 1991)
7
5.3 x 10"4
2.0 x 10"3
2
Note: Source: USEPA, 2014a.
1) The cancer HRL is determined by dividing the population risk level of one in a million (1 x 10"6) by the age-adjusted unit
risk.
2) The nitrosamine HRL values were converted to ng/L by multiplying the |jg/L values by 1000
Six-Year Review 3 3-12
Technical Support Document for Nitrosamines
December 2016
-------
4 Analytical Methods
4.1 Introduction
This chapter presents information on analytical methods for detection and quantitation of
nitrosamines in drinking water and other aqueous media. The focus is on the six nitrosamines
that EPA is evaluating under the third Six-Year Review (SYR3) program: i.e., A'-nitrosodi-n-
butylamine (NDBA), A'-nitrosodi ethyl amine (NDEA), A'-nitrosodi methyl amine (NDMA), N-
nitrosodi-n-propylamine (NDPA), A^-nitrosom ethyl ethyl amine (NMEA), and A'-
nitrosopyrrolidine (NPYR).
Methods capable of detecting nitrosamines in drinking water include one EPA-developed
method and two Standard Methods (SM): EPA Method 521, SM 6450B and SM 6450C. All
three methods include all six nitrosamines of interest on their list of analytes. Section 4.2
describes EPAMethod 521, Version 1.0, 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 (CI-MS/MS) (USEPA, 2004b)
and discusses the sensitivity of the method relative to the six nitrosamines' health reference
levels (HRLs). EPA Method 521 is the method that was approved for use under the second cycle
of the Unregulated Contaminant Monitoring Rule (UCMR 2) to monitor the six nitrosamines. A
brief description of SM 6450B and SM 6450C (SM, 2012a, b) is provided in Section 4.3.
Section 4.4 provides a brief overview of other published analytical methods that may be used to
quantitate various nitrosamines in aqueous matrices (but not specifically drinking water). Section
4.5 discusses several methods recently used in research, including a method for the analysis of
total nitrosamines in disinfected water (Kulshrestha et al., 2010), and methods using various
types of liquid chromatography followed by either photolysis and colorimetric determination
(Lee et al., 2013) or tandem mass spectrometry (Kadmi et al., 2014).
Exhibit 4.1 shows the analytical methods that can potentially be used for quantitation of each
nitrosamine in water.
Six-Year Review 3 4-1
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 4.1: Analytical Methods for SYR3 Nitrosamines
Analyte
Included in
EPA 521?
Included in SM
6450B and SM
6450C?
Other Analytical Methods for Water
Analysis
NDBA
Yes
Yes
EPA 8015C, EPA 8260B w/5031 prep., EPA
8270D
NDEA
Yes
Yes
EPA 8270D
NDMA
Yes
Yes
EPA 1625, EPA 607, EPA 625, EPA 8270D,
SM6410B, USGS 0-3118-83
NDPA
Yes
Yes
EPA 1625, EPA 607, EPA 625, EPA 8270D,
DOE OM100R, SM 6410B, USGS 0-3118-
83
NMEA
Yes
Yes
EPA 8270D
NPYR
Yes
Yes
EPA 8270D
Method Citations:
DOE Method OM100R (USDOE, 1997)
EPA Method 8260B (USEPA, 1996b)
EPA Method 5031 (USEPA, 1996c)
EPA Method 1625 (USEPA, 2001a)
EPA Method 607 (USEPA, 2001b)
EPA Method 625 (USEPA, 2001c)
4.2 EPA Method 521
EPA Method 521 (USEPA, 2004b)
EPA Method 8015C (USEPA, 2007a)
EPA Method 8270D (USEPA, 2007b)
SM 6450B/C (SM, 2012a, b)
SM 6410B (SM, 2012c)
USGS Method 0-3118-83 (USGS, 1983)
EPA Method 521 relies on solid phase extraction (SPE) followed by capillary column gas
chromatography (GC) with large volume injection coupled with chemical ionization tandem
mass spectrometry (CI-MS/MS). In EPA Method 521, Version 1.0 (USEPA, 2004b), a 0.5-L
water sample containing a known concentration of a surrogate analyte is extracted by passing
through a SPE cartridge containing 80-120 mesh coconut charcoal. Contaminants and the
surrogate are eluted from the cartridge with methylene chloride. Following drying, concentration
and the addition of an internal standard, the components are separated, identified and measured
by injection of an aliquot of the extract onto a fused silica capillary column in a GC/MS/MS
system equipped with a large-volume injector and operated in chemical ionization mode.
Contaminants are identified by comparing their product ion mass spectra and retention times to
reference spectra and retention times obtained through analysis of calibration standards measured
under the same conditions as the samples. Analytes and surrogates are quantified by measuring
their product ion responses relative to those of the internal standard. The surrogate analyte is
added to each water sample to evaluate extraction efficiency, while the internal standard is added
to each concentrated extract to evaluate run-to-run and/or day-to-day changes in sensitivity of the
GC/MS/MS instrumentation.
To describe how well the method performs, EPA Method 521 reports EPA-determined values for
the detection limit (DL) and lowest concentration minimum reporting level (LCMRL). The DL
appears in place of the method detection limit (MDL) in the more recent EPA-developed
methods. Over time, drinking water compliance methods have migrated away from requiring
MDL determinations in favor of confirming minimum reporting levels (MRLs). Various
regulatory bodies often still require determination of DLs. As a result, most of the newer
Six-Year Review 3
Technical Support Document for Nitrosamines
4-2
December 2016
-------
drinking water analytical methods incorporate a DL determination that is defined and conducted
exactly like the MDL as described in 40 CFR Part 136, Appendix B.
EPA Method 521 also includes values for the LCMRL for each analyte in the method. The
LCMRL is a single-laboratory reporting level and represents a change from how analytical
methods have typically presented performance data in the past (in the past, the DL was the focus;
in EPA Method 521, the MRL is the focus). The LCMRL is defined as the lowest spiking
concentration such that the probability of spike recovery in the 50 percent to 150 percent range is
at least 99 percent.
The LCMRL serves as a laboratory- and analyte-specific reporting level. Different analysts using
different equipment in different laboratories will not necessarily be able to achieve the LCMRLs
that are published in EPA analytical methods; however, EPA's published LCMRLs are an
indication that low analyte concentrations can be reliably reported.
In conjunction with the LCMRL, EPA has developed a statistically derived MRL that is
determined using raw LCMRL study data and represents an estimate of the lowest concentration
of a contaminant that can be reliably measured by members of a group of experienced drinking
water laboratories. Note that MRLs are not established in a published analytical method; rather,
they are derived from the LCMRLs obtained by laboratories using a specific analytical method in
an LCMRL study. Hence, the nitrosamine MRLs defined for laboratories that participated in the
analysis of drinking water samples under UCMR 2 are not published as part of EPA Method 521,
but they are method-specific (see Section 7.2.1).
Exhibit 4.2 summarizes the DLs, LCMRLs, MRLs, average percent recoveries and percent
relative standard deviation (RSD) results for six nitrosamines in EPA Method 521.
Exhibit 4.2: Performance Metrics for Six Nitrosamines in EPA Method 521
Analyte
DL
(ng/L)
LCMRL
(ng/L)
MRL
(ng/L)
Recovery
Range1
Relative Standard Deviation
(RSD) Range1
NDBA
0.36
1.4
4
79.7%-104%
2.9%-16%
NDEA
0.26
2.1
5
84.6%-95.6%
6.5%-14%
NDMA
0.28
1.6
2
83.7%-94.7%
3.8%-12%
NDPA
0.32
1.2
7
77.1%-97.0%
3.7%-10.2%
NMEA
0.28
1.5
3
81,4%-91.0%
4.5%-9.6%
NPYR
0.35
1.4
2
85.2%-102%
4.0%-12%
Source: USEPA, 2004b (DL, LCMRL, Recovery Range and RSD Range); USEPA, 2005c (MRL)
Note:
1) Recoveries and RSDs measured in reagent water, and in chlorinated drinking water from three sources: ground water,
surface water, and surface water with high total organic carbon levels.
4.2.1 Calculation of EPA Method 521 Nitrosamine MRLs
The procedure EPA used to determine the LCMRLs and MRLs for the nitrosamines in EPA
Method 521 was developed during the years 2003-2005 for eventual use in UCMR 2. The
procedure was described in EPA's Proposed Rule for UCMR 2 (USEPA, 2005c):
"To determine the MRLs listed in today's action, each laboratory that conducted the primary
analytical method development, or second or third laboratory studies, determined LCMRLs as
Six-Year Review 3 4-3
Technical Support Document for Nitrosamines
December 2016
-------
detailed in the statistical protocol. The mean of these LCMRL values was calculated for each
analyte. In cases where data from three or more laboratories were available, three times the
standard deviation of the LCMRLs was added to the mean of the LCMRLs, to establish the
MRL. In cases where data from two laboratories were available, three times the difference of the
LCMRLs was added to the mean of the LCMRLs. In statistical theory (Chebyshev's Inequality),
three standard deviations around the mean incorporates the vast majority (at least 88.9 percent)
of the data points. In the case where there are only two laboratories, the difference serves as a
surrogate for the standard deviation due to the uncertainty in the estimate of the standard
deviation with only two data points."
Since the implementation of UCMR 2, the LCMRL and MRL determination procedures have
been substantially upgraded by incorporating more rigorous statistics. These new procedures
were implemented for UCMR 3. While it may be possible to process the nitrosamine LCMRL
data from UCMR 2 development through the revised procedures for LCMRL and MRL
determination, it is unlikely that lower values for the LCMRLs or MRLs would be obtained. The
low ng/L range may be the practical limit of quantitation for nitrosamines using any analytical
method given difficulties differentiating true analyte signals from the instrument baseline and
background levels of nitrosamines.
4.2.2 Comparison of EPA Method 521 Performance to HRLs
A key factor in evaluating the performance of an analytical method is the comparison of
anticipated reporting levels to concentrations of concern to human health (e.g., HRLs). Reporting
levels can be estimated as approximately 5-10 times the MDL or DL, but in the case of EPA
Method 521, EPA's single-laboratory reporting level (the LCMRL) and a national reporting level
(the MRL) are available. Since the MRL served as a national reporting level for laboratories that
analyzed samples for UCMR 2 (USEPA, 2007c), it is the ideal metric for comparison to HRLs.
For each of the six nitrosamines, a method sensitivity ratio (MSR) was calculated to determine
whether the available analytical method is capable of reliable quantitation at or below the HRL
(see Exhibit 4.3). The MSR is calculated from the following equation:
MSR = HRL (ng/L) / MRL (ng/L)
A favorable MSR is one that is greater than 10. That is, it is preferable that the HRL be at least
10 times greater than the concentration at which data can be reliably reported. This provides a
margin of safety for uncertainty in the HRL and/or method performance (see USEPA, 2009). For
information on the calculation of HRLs, see Chapter 3.
Exhibit 4.3: Method Sensitivity Ratios (MSRs) for Nitrosamines
Analyte
MRL (ng/L)
HRL (ng/L)
MSR
NDBA
4
30
7.5
NDEA
5
0.4
0.08
NDMA
2
0.6
0.3
NDPA
7
7
1
NMEA
3
3
1
NPYR
2
2
1
Six-Year Review 3 4-4
Technical Support Document for Nitrosamines
December 2016
-------
The MSRs based on the HRLs are all less than 10. Hence, EPA Method 521 is not capable of
quantitation at levels at least 10 times below the HRLs. With an MSR of 7.5, NDBA has the
most favorable MSR. For NDPA, NMEA, and NPYR, the MRL is equal to the HRL, leaving no
margin for uncertainty in discerning concentrations at the threshold of health concern. For
NDEA and NDMA, the MSR is less than one, indicating that reliable quantitation is not possible
at the level of the HRL.
4.3 Other Drinking Water Methods
Two additional drinking water methods, SM 6450B and SM 6450C, were identified as applicable
to the nitrosamines. Method 6450B is a solid-phase extraction method that uses a granular
carbonaceous adsorbent resin. With a 500- to 1000-fold concentration factor, it can achieve
MDLs in the range of 0.5 to 2.0 ng/L. Method 6450C is a micro-liquid-liquid extraction method.
This method achieves a 200-fold concentration factor, with MDLs in the 2- to 4-ng/L range (SM,
2012a, b).
While contaminant-specific reporting levels are not available for SM 6450B and SM 6450C,
comparison of MDLs suggests that they may be comparable to, or slightly higher than, reporting
levels for EPA Method 521.
4.4 Other Published Methods for Measurement of Nitrosamines in Aqueous Media
Exhibit 4.4 presents a list of analytical methods that are available for the analysis of nitrosamines
in aqueous matrices but are not specifically intended for drinking water. The exhibit also lists
performance metrics, when available, including DLs, MDLs, Minimum Levels, Lower Limits of
Quantitation and Reporting Levels. For comparison, the DLs and LCMRLs for EPA Method 521
and MDLs for SM 6450B and SM 6450C are also included.
The methods listed in Exhibit 4.4 demonstrate variable coverage of the six nitrosamines. The
aqueous methods that are not specified for drinking water analyses are less sensitive than EPA
Method 521, SM 6450B and SM 6450C, sometimes by several orders of magnitude.
Six-Year Review 3 4-5
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 4.4: Method Performance Metrics for Nitrosamines Using Various Methods
Applicable to Aqueous Matrices1
Method2
Metric3
NDBA
(ng/L)
NDEA
(ng/L)
NDMA
(ng/L)
NDPA
(ng/L)
NMEA
(ng/L)
NPYR
(ng/L)
EPA Method 521
DL
0.36
0.26
0.28
0.32
0.28
0.35
EPA Method 521
LCMRL
1.4
2.1
1.6
1.2
1.5
1.4
SM6450B
MDL
0.71
0.81
0.84
1.08
0.45
0.71
SM 6450C
MDL
2.2
2.5
1.7
2.1
1.7
4.4
EPA 1625
ML
N/A
N/A
50,000
20,000
N/A
N/A
EPA 607
MDL
N/A
N/A
150
460
N/A
N/A
EPA 625
none
N/A
N/A
listed only
listed only
N/A
N/A
EPA 8015C
none
listed only
N/A
N/A
N/A
N/A
N/A
EPA 8260B/5031
MDL
14,000
N/A
N/A
N/A
N/A
N/A
EPA 8270D
LLQ
10,000
20,000
listed only
10,000
listed only
40,000
DOE OM100R
LLQ
N/A
N/A
N/A
5,000
N/A
N/A
SM6410B
none
N/A
N/A
listed only
listed only
N/A
N/A
USGS 0-3118-83
RL
N/A
N/A
5,000
10,000
N/A
N/A
Note:
1) Some published methods do not provide detection or quantitation values for some or all listed analytes. Analytes that are
listed in a published method without performance metrics are designated in the exhibit as "listed only"; analytical
difficulties may be encountered if using the indicated methods for these analytes. "N/A" = not applicable (the contaminant
is not listed in the published method).
2) Method Citations:
DOE Method OM100R (USDOE, 1997)
EPA Method 8260B (USEPA, 1996b)
EPA Method 5031 (USEPA, 1996c)
EPA Method 1625 (USEPA, 2001a)
EPA Method 607 (USEPA, 2001b)
EPA Method 625 (USEPA, 2001c)
3) Detection-based metrics include:
Detection Limit (DL) Method Detection Limit (MDL)
Quantitation-based metrics include:
Minimum Level (ML) Lowest Concentration Minimum Reporting Level (LCMRL)
Reporting Level (RL) Lower Limit of Quantitation (LLQ)
4.5 Other Methods Used in Research
A method for the determination of total A'-nitrosamines (TONO) in disinfected water used for
swimming pools, as well as the source water used to fill the pools, was described in Kulshrestha
et al. (2010). This TONO method is an adaptation and optimization of an assay used for the
determination of nitrite, »Y-nitrosothiols and A'-nitrosamines in biological samples. The goal of the
study was to gather data on whether NDMA and other nitrosamines of interest are dominant or
minor constituents of the TONO measured. The method is summarized here, and its possible
usefulness is discussed.
In this analytical method, disinfected water samples are analyzed for nitrosamines via reduction
by aqueous tri-iodide in glacial (i.e., concentrated, anhydrous) acetic acid. The acidic solution
reduces nitrosamines to nitric oxide (NO) which is measured by chemiluminescence. Major
interferences (i.e., other compounds also reduced to NO) in the reaction may include S-
EPA Method 521 (USEPA, 2004b)
EPA Method 8015C (USEPA, 2007a)
EPA Method 8270D (USEPA, 2007b)
SM 6450B/C (SM, 2012a, b)
SM 6410B (SM, 2012c)
USGS Method 0-3118-83 (USGS, 1983)
Six-Year Review 3 4-6
Technical Support Document for Nitrosamines
December 2016
-------
nitrosothiols and nitrite; these two interferences are eliminated by the sequential addition of
mercuric chloride and sulfanilamide, respectively.
Using NDMA as a model nitrosamine, the authors report an MDL of 0.11 |iM (8,200 ng/L) and a
method reporting limit of 0.3 |iM (22,000 ng/L) (Kulshrestha et al., 2010). To improve method
sensitivity to approach the ng/L-scale concentrations of NDMA and other nitrosamines
anticipated to be found in the subject water samples, a continuous liquid-liquid extraction
procedure was employed using ethyl acetate. This allowed for improved extraction of polar
nitrosamines (which extract more poorly relative to NDMA and other less polar nitrosamines
when using a less polar extraction solvent) and a decreased MDL. The MDL was further
decreased by the implementation of a multi-step extraction procedure in which 400 mL of ethyl
acetate was used over three extractions and then evaporated and blown down to 1 mL. The
resultant MDL for total nitrosamines was 5 ng/L (as NDMA). The results were cross-evaluated
by performing duplicate analysis of the samples using EPA Method 521 to identify NDMA
concentrations and to establish the predominance of NDMA in the TONO analyses. The authors
reported that NDMA represents a range of approximately 3 to 46 percent of the TONO results.
This TONO assay may require further testing and adaptation for establishing the occurrence of
nitrosamines in drinking water. It may have applications as a screening tool in the analysis of N-
nitrosamines.
Other techniques have been proposed recently in the published literature as described below.
These techniques may need further testing and adaptation before they could be employed for
analysis of nitrosamines in finished drinking water samples.
Lee et al. (2013) photolytically converted nitrosamines to nitrite at UV-254 nm. Nitrite was
subsequently converted to a highly colored azo dye via the Griess reaction and detected
spectrophotometrically. SPE was used to pre-concentrate nitrosamines in the samples. The SPE
was conducted using carbon columns, and extraction efficiencies were 45 to 96 percent. The
extracted nitrosamines were then separated by high performance liquid chromatography (HPLC.
In a post-column reactor, the nitrosamines were photolyzed to nitrite, which was detected and
quantitated using the Griess colorimetric determination. The MDLs with SPE ranged from 5.9
ng/L for NDMA to 27.6 ng/L for NDBA. Note that these MDLs are approximately one order of
magnitude greater than MDLs obtained using EPA Method 521.
Kadmi et al., (2014) proposed a method for NMEA that uses SPE followed by HPLC coupled
with tandem mass spectrometry (MS/MS). They reported a linear calibration curve from 0.1 to
100 ng/L. The extraction efficiency was 86 percent and the level of quantitation was reported as
0.8 ng/L. Although this single-laboratory value is less than EPA's LCMRL for NMEA of 1.5
ng/L, it is not clear how comparable the level of quantitation is to EPA's LCMRL.
Six-Year Review 3 4-7
Technical Support Document for Nitrosamines
December 2016
-------
5 Occurrence and Exposure in Drinking Water
5.1 Introduction
Data are available for nitrosamine occurrence in finished drinking water in public water systems
(PWSs) from the nationally representative monitoring completed under the Second Unregulated
Contaminant Monitoring Rule (UCMR 2). UCMR 2 monitoring included monitoring for all six
nitrosamines discussed in this document: A'-ni trosodi-n-butyl amine (NDBA),
A'-ni trosodi ethyl ami ne (NDEA), A'-ni trosodi methyl amine (NDMA), A'-ni trosodi-n-propyl amine
(NDPA), A'-ni trosomethyl ethyl amine (NMEA) and A'-ni trosopyrroli dine (NPYR). EPA consulted
other sources as well to gather additional data on the occurrence of the nitrosamines in ambient
water and drinking water. These data are presented in Appendix A. On the whole, data from
these additional sources support the conclusions reached from the UCMR 2 analyses.
Of the six nitrosamines, NDMA is the only one that had a sufficient number of samples with
detections under UCMR 2 to allow for modeled estimates of occurrence and population exposure
below the detection limit (DL). Nitrosamines detectable with EPA Method 521 (which include
these six nitrosamines, plus A'-ni trosopi peri dine) may constitute only 5-10 percent (as molar
percentage) of total nitrosamines in drinking water (Kulshrestha et al., 2010; Dai and Mitch,
2013; Krasner et al., 2013). Krasner et al. (2013) found that NDMA alone may account for about
five percent of total nitrosamines in chloraminated waters, where it tends to occur most. As
mentioned in Chapter 4 of this document, while some studies have been conducted to quantify
total A'-nitrosamines (TONO) in disinfected water, the analytical methods reported in the
literature may require further adaptation before they can be used to establish the occurrence of
TONO in drinking water. While occurrence and exposure data for all six nitrosamines are
discussed in this chapter, and some indications of national occurrence and co-occurrence are
provided for all six, EPA has focused primarily on using the NDMA UCMR 2 data to
characterize national occurrence and exposure in detail.
Exhibit 5.1 recaps the health reference levels (HRLs) for each of the six nitrosamines discussed
in Chapter 3 and also shows the corresponding UCMR 2 minimum reporting levels (MRLs).
(See Section 4.2.1 for background on the derivation of MRLs). As discussed in Chapter 3, HRLs
are risk-derived concentrations against which contaminant occurrence data from PWSs can be
compared to determine if the contaminant occurs with a frequency and at levels of public health
concern. It is important to note that because the HRLs for nitrosamines are based on carcinogenic
effects, long-term average concentrations are more relevant than intermittent or short-term peak
concentrations for making those comparisons. For four of the six nitrosamines (NDBA, NDPA,
NMEA and NPYR), the MRL values in UCMR 2 are equal to or below their corresponding
HRLs. For two of the nitrosamines (NDMA and NDEA), however, the MRLs are greater than
the corresponding HRL values. Specifically, NDMA has an MRL of 2 ng/L, which is greater
than its HRL of 0.6 ng/L, and NDEA has an MRL of 5 ng/L, which is greater than its HRL of 0.4
ng/L. The significance of this is that one cannot directly observe the number of NDMA and
NDEA samples with concentrations at or above their respective HRLs.
While it is possible to develop estimates of a contaminant's occurrence below its MRL value
using modeling techniques, there must be a substantial number of positive sample results (that is,
at or above the MRL value) to support such modeling efforts. Only NDMA had sufficient
positive samples in the UCMR 2 occurrence dataset to support this type of modeling.
Six-Year Review 3 5-1
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 5.1: HRLs and MRLs for the Six Nitrosamine Compounds
Contaminant
Health Reference
Level (HRL), in ng/L
Minimum Reporting
Level (MRL), in ng/L
NDBA
30
4
NDEA
0.4
5
NDMA
0.6
2
NDPA
7
7
NMEA
3
3
NPYR
2
2
The remainder of this chapter presents a detailed analysis of information from UCMR 2 on the
occurrence of the nitrosamines in drinking water. The following topics are covered:
• Description of the UCMR 2 monitoring program and dataset;
• Occurrence of the nitrosamines, in aggregate and individually, on the basis of simple
detections and HRL exceedances (at the level of individual samples and at the level of
PWSs);
• Co-occurrence analysis for six nitrosamines in the UCMR 2; and
• Parametric modeling of NDMA mean concentrations, including occurrence below the
MRL.
5.2 UCMR 2 Monitoring Program and Dataset
The purpose of EPA's Unregulated Contaminant Monitoring Rules is to collect occurrence data
on contaminants that do not have established health-based national standards under the Safe
Drinking Water Act but are suspected to be present in drinking water. UCMR 2 monitoring,
conducted between January 2008 and December 2010, provided the data for the nitrosamine
occurrence analysis presented in this section. This dataset is available from the agency's website
(https://www.epa.gOv/dwucmr/occurrence-data-unregulated-contaminant-monitoring-rule#2). A
more comprehensive discussion of UCMR 2, and results for all contaminants included in the
survey, are provided in Occurrence Data from the Second Unregulated Contaminant Monitoring
Regulation (UCMR 2) (USEPA, 2014b).
The occurrence analyses that are described in this chapter were based on data collected through
June 2011 and released in July 2011. A relatively small amount of additional data was received
and added to the UCMR 2 dataset after June 2011. The UCMR 2 dataset was not considered
"final" until December 2011. The "final" numbers are presented and analyzed in the UCMR 2
Occurrence Document (USEPA, 2014b).
Section 5.2.1 provides a description of data collected for UCMR 2. Section 5.2.2 presents the
stratification of UCMR 2 data for the purposes of nitrosamines occurrence analysis, and Section
5.2.3 summarizes the dataset of nitrosamine results.
Six-Year Review 3 5-2
Technical Support Document for Nitrosamines
December 2016
-------
5.2.1 Description of Data Collected Under UCMR 2
The UCMR 2 involved two types of occurrence monitoring. Selected community water systems
(CWSs) and non-transient non-community water systems (NTNCWSs) were required to conduct
Assessment Monitoring of 10 chemicals using common laboratory analytical techniques and a
Screening Survey of 15 chemicals by means of analytical techniques that are not commonly used
in drinking water analysis. The six nitrosamines included in UCMR 2 were part of the Screening
Survey and underwent monitoring using EPA Method 521 (described in more detail in Chapter
4). Although other nitrosamines (e.g., A-nitrosomorpholine, A-nitrosopi peri dine) have been
identified in finished water (Mitch et al., 2009), they were not included in the UCMR 2 due to
limitations of this analytical method. Under the UCMR 2, PWSs were required to collect a
sample at each entry point to the distribution system, as well as at the maximum residence time
locations within the distribution system associated with each entry point, and to report the
disinfectant type in use at these locations while the samples were being taken.
For the Screening Survey component, all CWSs and NTNCWSs serving more than 100,000
people ("very large" systems) and a representative sample of 800 CWSs and NTNCWSs serving
100,000 or fewer people (320 "large" systems serving 10,001 to 100,000 people and 480 "small"
systems serving 10,000 or fewer people) were required to participate. Transient non-community
water systems (TNCWSs) and any PWSs that purchase all of their water from other PWSs were
not required to conduct sampling under UCMR 2. (Note: The population served by these
purchasing PWSs is considered in the national occurrence analyses, as discussed below.)
To obtain a nationally representative sample of CWSs and NTNCWSs serving 100,000 or fewer
people, EPA statistically categorized (stratified) the PWSs by their source water type (ground or
surface water) and by the size of the population served, using data from EPA's Safe Drinking
Water Information System (SDWIS). (The SDWIS system source water and population served
data used for the analyses is from July 2005, so that these system inventory data are consistent
with the approximate timeframe of the UCMR 2 sampling.) The initial stratification by source
water type was consistent with SDWIS classification protocol, meaning that PWSs using ground
water under the direct influence of surface water (GWUDI) and those using a mix of ground
water and surface/GWUDI sources were classified as surface water PWSs. EPA used a different
source water type stratification for the nitrosamines analysis, as is explained in Section 5.2.2.
To stratify PWSs by population served, EPA used data representing the total population served
by a PWS. Adjustments were made to populations served by systems to include the populations
served by PWSs that purchase water (since these systems were not required to conduct UCMR 2
monitoring) and to avoid double counting of populations served by some systems. In cases where
one PWS sells water to another PWS (or multiple PWSs), EPA added the population served by
each PWS that purchases water to the population served by the PWS that sells water to calculate
the total population served as represented by UCMR 2 sampling systems. (The PWS that
purchases water was not required to conduct UMCR 2 sampling and the systems that sells water
did conduct UCMR 2 sampling.) For example, if PWS A serving 10,000 people also sells water
to PWS B serving 20,000 people, the total population served by PWS A, the system that
conducted sampling, would be considered 30,000 people. Prior to analyzing UCMR 2 results,
EPA also made another adjustment to the population served data to address potential double
counting in cases where a PWS purchases water from more than one seller (which was the case
for 386 PWSs). If a system purchased water from two or more systems, the population associated
Six-Year Review 3 5-3
Technical Support Document for Nitrosamines
December 2016
-------
with the purchasing system was equally divided across (added to) the population-served values
of the selling systems. This proportionate distribution of purchasing system populations to
selling systems was done because data on the relative quantities of water purchased from each
seller were not available.
Exhibit 5.2 shows PWS stratification for two source water and six population categories,
information on total population served by PWSs in each category based on SDWIS system
inventory data and adjusted population served by each category following the correction for
double counting. (The SDWIS inventory data are from July 2005 so that systems information is
consistent with the dates of UCMR 2 sampling. Some system source water classifications and
other systems properties had changed after the systems were selected for UCMR 2 monitoring.)
See Exhibit 5.5 for the breakdown of system size, source water classifications, etc., for
nitrosamines in the final dataset.
Exhibit 5.2: PWS and Population Stratification for UCMR 2 Screening Survey
Monitoring
Source Water
Type
Size Category
(Population Served)
Number of
PWSs2
Total Population
Served3
Adjusted Population
Served Corrected for
Double Counting4
Ground Water
Small (25-500)
80
11,774
11,774
Ground Water
Small (501-3,300)
80
95,960
95,960
Ground Water
Small (3,301-10,000)
80
440,630
440,630
Ground Water
Large (10,001-50,000)
75
1,680,865
1,679,711
Ground Water
Large (50,001-100,000)
76
5,011,923
4,904,473
Ground Water
Very Large (>100,000)
63
17,363,412
17,269,919
Surface Water1
Small (25-500)
80
18,507
18,507
Surface Water1
Small (501-3,300)
80
149,657
149,494
Surface Water1
Small (3,301-10,000)
80
502,109
492,556
Surface Water1
Large (10,001-50,000)
85
1,888,542
1,804,026
Surface Water1
Large (50,001-100,000)
84
5,922,105
5,640,805
Surface Water1
Very Large (>100,000)
335
129,475,277
124,711,765
Totals
1,198
162,560,761
157,219,620
Note:
1) Includes PWSs using GWUDI and PWSs that use a mix of ground water and surface or GWUDI sources.
2) Includes CWSs and NTNCWSs. A minimum of two PWSs are located in each state.
3) Total population used for PWS stratification for UCMR 2. Based on combined retail and wholesale population as reported in
SDWIS (July 2005).
4) Total population served after adjustments made for those PWSs that purchase water from more than one seller (386 PWSs).
In coordination with the states, EPA assigned a monitoring schedule to each participating PWS.
Surface water PWSs (including GWUDI and mixed PWSs) were required to monitor quarterly
(i.e., every three months) over a 12-month period. Ground water PWSs were required to monitor
twice (at a five- to seven-month interval) over a 12-month period. As noted previously,
monitoring was primarily conducted between January 2008 and December 2010.
Two aspects of the UCMR 2 sampling were unique to the nitrosamines. First, while all UCMR 2
PWSs were required to collect samples at all entry points (EPs) to their distribution systems, the
PWSs sampling for nitrosamines were also required to collect samples at maximum residence
Six-Year Review 3 5-4
Technical Support Document for Nitrosamines
December 2016
-------
(MR) time locations. UCMR 2 also required that the PWSs indicate which MR location(s) was or
were associated with each EP location.
The other aspect unique to nitrosamine monitoring under UCMR 2 was the requirement that the
PWSs also report the type of disinfection in use at each relevant EP when those samples were
collected. (Note that disinfection type at the MR locations was not explicitly reported in UCMR
2, but was inferred from the dataset, as described further in the next section.)
EPA conducted several quality assurance reviews including completeness of the entire UCMR 2
dataset. For more information on quality assurance steps, see the UCMR 2 Occurrence
Document (USEPA, 2014b).
5.2.2 Stratification of UCMR 2 Data for Nitrosamine Group Analysis
As described in the previous section, the UCMR 2 included all very large PWSs (serving more
than 100,000 people) in the Screening Survey and used a stratification approach for selecting a
nationally representative sample of 800 PWSs serving 100,000 or fewer people. The sample of
smaller systems was stratified by source water type (ground or surface water) and PWS size
(based on five population-served categories), with some additional consideration given to
geographic location to ensure minimum representation from each state. To enhance the
representativeness of the UCMR 2 for developing national occurrence and exposure estimates,
EPA further stratified the results to reflect more specific aspects of source water and disinfection
type at the PWS and at EP and MR locations. The approach to those additional stratification
aspects is described in this section.
5.2.2.1 Source Water Type
The UCMR 2 required PWSs to report the source water type (surface water, GWUDI or ground
water) for each EP location for each sampling event. Also, for each EP location and sampling
event, PWSs were required to identify the associated MR locations. EPA used this information to
develop a source water classification scheme for PWSs, EP locations and MR locations as shown
in Exhibit 5.3. Note that in this further stratification, EPA included a source water category of
"mixed" to capture those PWSs, EP locations and MR locations that reported using both ground
water and surface water (or GWUDI) sources during the UCMR 2 monitoring period.
Six-Year Review 3 5-5
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 5.3: Source Water Type Classification Scheme
Source
Water Type
Designation
Rules for Assigning
the Source Water
Type to PWSs
Rules for Assigning
the Source Water Type
to EP Sites
Rules for Assigning
the Source Water
Type to MR Sites1
Rules for Assigning
the Source Water Type
to MR Sites (If No
Associated EP)1
Ground Water
All entry points at the
PWS are GW only
The entry point uses GW
only
All associated entry
points are GW only
Entry points at the PWS
are GW only
Surface
Water
All entry points at the
PWS are SW or
GWUDI only
The entry point uses SW
and/or GWUDI only
All associated entry
points are SW or
GWUDI only
Entry points at the PWS
are SW or GWUDI only
Mixed
Entry points at the
PWS are a mix of GW
and SW and/or
GWUDI
The entry point uses a
mix of GW and SW
and/or GWUDI
Associated entry
points are a mix of GW
and SW and/or
GWUDI
Entry points at the PWS
are a mix of GW and
SW and/or GWUDI
Abbreviations: EP = entry point; GWUDI = ground water under the direct influence of surface water; GW = ground water; MR =
maximum residence time location; SW = surface water
Note:
1) For EP sites, PWSs reported associated MR sites. In some cases, MR sites did not have associated EP sites.
5.2.2.2 Disinfection Type
The UCMR 2 required PWSs to report the type of disinfectant in use at the time of sampling for
each EP location and associated MR location and for each sampling event. If the PWS did not
report a disinfectant type with the sample results, EPA reviewed other data submitted by the
PWS and, as appropriate, contacted the PWS to determine the disinfectant in use for all sample
results.
For the analysis of the nitrosamine results, EPA developed a stratification scheme involving six
categories of disinfectant type, as shown in Exhibit 5.4. The disinfectant of primary interest with
respect to nitrosamine formation is chloramine, used either alone or in conjunction with free
chlorine or some other disinfectant. The six disinfectant types used for this analysis are: (1)
chloramines only, (2) chloramines in combination with another disinfectant, (3) chlorine only,
(4) chlorine in combination with another disinfectant other than chloramines, (5) a disinfectant
other than chlorine or chloramines and (6) no disinfection. Disinfection is required at systems
using surface water. Four small surface water PWSs are listed in the dataset as "ND" (no
disinfection); this is presumably due to incorrect self-reporting.
Six-Year Review 3 5-6
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 5.4: Disinfectant Classification Scheme
Disinfectant
Designation
Rules for
Assigning
Disinfectant
Designation to
PWSs
Rules for Assigning
Disinfectant Designation
to EP Sites
Rules for Assigning
Disinfectant
Designation to MR
Sites1
Rules for Assigning
Disinfectant
Designation to MR Sites
(If No Associated EP)1
CA Only
All entry points with
disinfection use
chloramine (some
may be ND)
The entry point uses only
CA when disinfecting (but
may be ND at some times)
At least one associated
entry point has CA,
none have CL or OT
(some may be ND)
All entry points from the
PWS have CA, none
have CL or OT (some
may be ND)
CA with CL/OT
At least one entry
point uses CA, one
or more others use
chlorine or other
disinfectants
The entry point uses CA
for at least one sampling
period, but also uses CL
or OT for at least one
other sampling period
At least one associated
entry point has CA, one
or more have CL or OT
(and some may be ND)
At least one entry point
from the PWS has CA,
one or more have CL or
OT (and some may be
ND)
CLOnly
All entry points with
disinfection use
chlorine (some may
be ND)
The entry point uses only
CL when disinfecting (but
may be ND in some
sampling periods)
At least one associated
entry point has CL, none
have CA or OT (some
may be ND)
All entry points from the
PWS have CL, none
have CA or OT (some
may be ND)
CL with OT
At least one entry
uses CL and one or
more entry points
use OT (and some
others may be ND)
The entry point uses CL
for at least one sampling
period, but also uses OT
for at least one other
sampling period
At least one associated
entry point has CL, one
or more have OT (and
some may be ND)
At least one entry point
from the PWS has CL,
one or more have OT
(and some may be ND)
OT Only
All entry points with
disinfection use OT
(some may be ND)
The entry point uses OT
for at least one sampling
period, but never CA or
CL (some periods may be
ND)
At least one associated
entry point has OT,
none have CA or CL
(some may be ND)
At least one entry point
from the PWS has OT,
none have CA or CL
(some may be ND)
No Disinfection
(ND)2
All entry points are
ND
There is no disinfection at
the entry point for any
sampling periods
No disinfection at any of
the associated entry
points for any sampling
periods
All entry points from the
PWS are ND only
Abbreviations: CL = chlorine; CA = chloramine; OT = all other types of disinfectant (e.g., chlorine dioxide or ozone); EP = entry
point; MR = maximum residence time location; ND = no disinfectant
Note:
1) For EP sites, PWSs reported associated MR sites. In some cases, MR sites did not have associated EP sites.
2) Disinfection is required for surface water. Four small surface water PWSs are listed in the dataset as "ND." This is
presumably due to incorrect self-reporting.
5.2.3 Summary of UCMR 2 Dataset for Nitrosamines
Exhibit 5.5 presents the total number of PWSs, EP sample locations and MR sample locations in
the UCMR 2 dataset for nitrosamines, reflecting the addition of mixed sources to the source
water stratification. In summary, the dataset includes monitoring results for 1,198 PWSs, 4,666
EP locations and 2,397 MR locations. Exhibit 5.5 also presents the population served by the
UCMR 2 PWSs in the nitrosamine dataset. Note that the PWSs in the UCMR 2 Screening
Survey included both CWSs and NTNCWSs. CWSs predominate. Most analyses presented in
this chapter use data from both CWSs and NTNCWSs. Modeled national extrapolations on
NDMA occurrence, however (presented and described in Section 5.5), apply to CWSs only.
Six-Year Review 3 5-7
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 5.5: Summary of UCMR 2 Dataset for Nitrosamines
PWS Size1
Source Water
Type
Total Number
of PWSs
Total Population
Served
Total Number of
EPs
Total Number
of MRs
Small
All PWSs
480
1,208,921
677
503
Small
Surface Water
222
573,691
232
226
Small
Ground Water
240
548,364
399
238
Small
Mixed Water
18
86,866
46
39
Large
All PWSs
320
14,029,015
1,124
544
Large
Surface Water
120
5,187,934
153
145
Large
Ground Water
163
7,021,968
770
311
Large
Mixed Water
37
1,819,112
201
88
Very Large
All PWSs
398
141,981,684
2,865
1,350
Very Large
Surface Water
218
89,237,751
400
367
Very Large
Ground Water
72
18,472,835
1,042
490
Very Large
Mixed Water
108
34,271,098
1,423
493
Totals
1,198
157,219,620
4,666
2,397
Note: Dataset includes CWSs and NTNCWSs. Population served based on 2005 SDWIS data.
1) Small = serving < 10,000; Large = serving 10,001 - 100,000; Very Large = serving > 100,000.
Exhibit 5.6 through Exhibit 5.8 show the number of PWSs, EP locations and MR locations,
respectively, in the UCMR 2 nitrosamines dataset, broken out by PWS size, source water type
and disinfection type. Most PWSs and sampling locations are classified as using free chlorine
only, with the proportion of PWSs using only free chlorine higher in small PWSs compared to
large and very large PWSs. Exhibit 5.6 shows that 283 PWSs (approximately 24 percent) in the
dataset use chloramines alone or with another disinfectant.
Exhibit 5.6: Number of PWSs in the UCMR 2 Dataset for Nitrosamines by
Disinfectant Type
PWS Size1
Source Water
Type
CA Only:
Number of
PWSs
CA with
CL/OT:
Number of
PWSs
CL Only:
Number
of PWSs
CL with OT:
Number of
PWSs
OT Only:
Number
of PWSs
ND Only:
Number
of PWSs
Small
All PWSs
20
25
378
1
1
55
Small
Surface Water
14
14
188
1
1
4
Small
Ground Water
3
3
183
0
0
51
Small
Mixed Water
3
8
7
0
0
0
Large
All PWSs
55
9
240
8
5
3
Large
Surface Water
30
2
85
1
2
0
Large
Ground Water
21
1
130
5
3
3
Large
Mixed Water
4
6
25
2
0
0
Very Large
All PWSs
129
45
207
13
3
1
Very Large
Surface Water
84
17
110
6
1
0
Very Large
Ground Water
17
4
45
4
1
1
Very Large
Mixed Water
28
24
52
3
1
0
Six-Year Review 3 5-8
Technical Support Document for Nitrosamines
December 2016
-------
PWS Size1
Source Water
Type
CA Only:
Number of
PWSs
CA with
CL/OT:
Number of
PWSs
CL Only:
Number
of PWSs
CL with OT:
Number of
PWSs
OT Only:
Number
of PWSs
ND Only:
Number
of PWSs
Totals
204
79
825
22
9
59
Abbreviations: CA = chloramine; CL = chlorine; OT = all other types of disinfectant (e.g., chlorine dioxide); ND = no disinfectant.
Note: See Section 5.2.2 for assumptions used to classify PWSs by source water and disinfectant type.
1) Small = serving < 10,000; Large = serving 10,001 - 100,000; Very Large = serving > 100,000.
Exhibit 5.7: Number of Entry Points in the UCMR 2 Nitrosamines Dataset by
Disinfectant Type
PWS Size1
Source Water
Type
CA Only:
Number
of EPs
CA with
CL/OT:
Number of
EPs
CLOnly:
Number
of EPs
CL with
OT:
Number
of EPs
OT Only:
Number
of EPs
ND Only:
Number
of EPs
Small
All PWSs
31
27
537
1
1
80
Small
Surface Water
15
14
197
1
1
4
Small
Ground Water
4
5
314
0
0
76
Small
Mixed Water
12
8
26
0
0
0
Large
All PWSs
91
12
950
12
8
51
Large
Surface Water
44
2
104
1
2
0
Large
Ground Water
31
1
679
7
6
46
Large
Mixed Water
16
9
167
4
0
5
Very Large
All PWSs
450
56
2,243
25
9
82
Very Large
Surface Water
142
25
225
5
3
0
Very Large
Ground Water
68
8
942
12
3
9
Very Large
Mixed Water
240
23
1,076
8
3
73
Totals
572
95
3,730
38
18
213
Abbreviations: CA = chloramine; CL = chlorine; OT = all other types of disinfectant (e.g., chlorine dioxide); ND = no disinfectant.
Note: See Section 5.2.2 for assumptions used to classify PWSs by source water and disinfectant type.
1) Small = serving < 10,000; Large = serving 10,001 - 100,000; Very Large = serving > 100,000.
Six-Year Review 3 5-9
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 5.8: Number of Maximum Residence Time Locations in the UCMR 2
Nitrosamines Dataset by Disinfectant Type
PWS Size1
Source Water
Type
CA Only:
Number of
MRs
CA with
CL/OT:
Number of
MRs
CLOnly:
Number of
MRs
CL with
OT:
Number
of MRs
OT Only:
Number
of MRs
ND Only:
Number of
MRs
Small
All PWSs
31
25
428
1
1
17
Small
Surface Water
15
14
191
1
1
4
Small
Ground Water
4
3
218
0
0
13
Small
Mixed Water
12
8
19
0
0
0
Large
All PWSs
75
12
438
11
7
1
Large
Surface Water
39
2
101
1
2
0
Large
Ground Water
28
1
268
8
5
1
Large
Mixed Water
8
9
69
2
0
0
Very Large
All PWSs
288
52
987
14
6
3
Very Large
Surface Water
131
19
210
4
3
0
Very Large
Ground Water
32
5
448
3
2
0
Very Large
Mixed Water
125
28
329
7
1
3
Totals
394
89
1,853
26
14
21
Abbreviations: CA = chloramine; CL = chlorine; OT = all other types of disinfectant (e.g., chlorine dioxide); ND = no disinfectant.
Note: See Section 5.2.2 for assumptions used to classify PWSs by source water and disinfectant type.
1) Small = serving < 10,000; Large = serving 10,001 - 100,000; Very Large = serving > 100,000.
Exhibit 5.9 through Exhibit 5.11 show the number of samples, EP location samples and MR
location samples, respectively, for NDMA (as a representative Screening Survey analyte) in the
UCMR 2 nitrosamines dataset, broken out by PWS size, source water type and disinfection type.
The numbers for other Screening Survey contaminants (including the other nitrosamines) vary
only slightly from the NDMA numbers. Most samples are from sampling locations associated
with the use of chlorine only.
Exhibit 5.9: Number of Samples in the UCMR 2 Dataset for NDMA by Disinfectant
Type
PWS Size1
Source Water
Type
CA Only:
Number of
Samples
CA with
CL/OT:
Number of
Samples
CL Only:
Number of
Samples
CL with OT:
Number of
Samples
OT Only:
Number of
Samples
ND Only:
Number of
Samples
Small
All PWSs
169
2,650
203
8
195
8
Small
Surface Water
106
1,503
116
8
28
8
Small
Ground Water
16
1,038
16
0
167
0
Small
Mixed Water
47
109
71
0
0
0
Large
All PWSs
80
3,220
524
60
95
36
Large
Surface Water
16
784
322
8
0
16
Large
Ground Water
4
1,826
116
30
86
20
Six-Year Review 3 5-10
Technical Support Document for Nitrosamines
December 2016
-------
PWS Size1
Source Water
Type
CA Only:
Number of
Samples
CA with
CL/OT:
Number of
Samples
CL Only:
Number of
Samples
CL with OT:
Number of
Samples
OT Only:
Number of
Samples
ND Only:
Number of
Samples
Large
Mixed Water
60
610
86
22
9
0
Very Large
All PWSs
387
7,679
2,401
116
164
45
Very Large
Surface Water
175
1,617
1,068
36
0
24
Very Large
Ground Water
28
2,750
204
30
18
9
Very Large
Mixed Water
184
3,312
1,129
50
146
12
Totals
636
13,549
3,128
184
454
89
Abbreviations: CA = chloramine; CL = chlorine; OT = all other types of disinfectant (e.g., chlorine dioxide); ND = no disinfectant.
Note: See Section 5.2.2 for assumptions used to classify sampling locations by source water and disinfectant type. The number and
distribution of samples in the dataset vary slightly from contaminant to contaminant. The numbers shown here represent NDMA.
1) Small = serving < 10,000; Large = serving 10,001 - 100,000; Very Large = serving > 100,000.
Exhibit 5.10: Number of Samples at Entry Points in the UCMR 2 Dataset for NDMA
by Disinfectant Type
PWS Size1
Source Water
Type
CA Only:
Number of
EP
Samples
CA with
CL/OT:
Number of
EP
Samples
CLOnly:
Number of
EP
Samples
CL with
OT:
Number of
EP
Samples
OT Only:
Number of
EP
Samples
ND Only:
Number of
EP
Samples
Small
All PWSs
88
1,451
102
4
158
4
Small
Surface Water
54
769
59
4
15
4
Small
Ground Water
10
618
8
0
143
0
Small
Mixed Water
24
64
35
0
0
0
Large
All PWSs
34
2,090
285
32
94
19
Large
Surface Water
8
401
170
4
0
8
Large
Ground Water
2
1,300
60
14
85
11
Large
Mixed Water
24
389
55
14
9
0
Very Large
All PWSs
192
4,919
1,397
70
158
25
Very Large
Surface Water
99
846
552
20
0
12
Very Large
Ground Water
16
1,863
136
24
18
5
Very Large
Mixed Water
77
2,210
709
26
140
8
Totals
314
8,460
1,784
106
410
48
Abbreviations: CA = chloramine; CL = chlorine; OT = all other types of disinfectant (e.g., chlorine dioxide); ND = no disinfectant.
Note: See Section 5.2.2 for assumptions used to classify sampling locations by source water and disinfectant type.
The number and distribution of samples in the dataset vary slightly from contaminant to contaminant. The numbers shown here
represent NDMA.
1) Small = serving < 10,000; Large = serving 10,001 - 100,000; Very Large = serving > 100,000.
Six-Year Review 3 5-11
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 5.11: Number of Samples at Maximum Residence Time Locations in the
UCMR 2 Dataset for NDMA by Disinfectant Type
PWS Size1
Source Water
Type
CA Only:
Number of
MR
Samples
CA with
CL/OT:
Number of
MR Samples
CL Only:
Number
of MR
Samples
CL with OT:
Number of
MR
Samples
OT Only:
Number of
MR
Samples
ND Only:
Number
of MR
Samples
Small
All PWSs
81
1,199
101
4
37
4
Small
Surface Water
52
734
57
4
13
4
Small
Ground Water
6
420
8
0
24
0
Small
Mixed Water
23
45
36
0
0
0
Large
All PWSs
46
1,130
239
28
1
17
Large
Surface Water
8
383
152
4
0
8
Large
Ground Water
2
526
56
16
1
9
Large
Mixed Water
36
221
31
8
0
0
Very Large
All PWSs
195
2,760
1,004
46
6
20
Very Large
Surface Water
76
771
516
16
0
12
Very Large
Ground Water
12
887
68
6
0
4
Very Large
Mixed Water
107
1,102
420
24
6
4
Totals
322
5,089
1,344
78
44
41
Abbreviations: CA = chloramine; CL = chlorine; OT = all other types of disinfectant (e.g., chlorine dioxide); ND = no disinfectant.
Note: See Section 5.2.2 for assumptions used to classify sampling locations by source water and disinfectant type.
1) Small = serving < 10,000; Large = serving 10,001 - 100,000; Very Large = serving > 100,000.
The number and distribution of samples in the dataset vary slightly from contaminant to
contaminant. The numbers shown here represent NDMA.
5.3 Summary of UCMR 2 Occurrence Findings for Nitrosamines
A total of 18,040 samples were collected and analyzed for all six nitrosamines at the 1,198 PWSs
included in UCMR 2. Of these, 10,792 samples (59.8 percent) were collected from 398 very
large PWSs serving more than 100,000 people, 4,015 samples (22.3 percent) from 320 large
PWSs serving between 10,001 and 100,000 people, and 3,233 samples (17.9 percent) from 480
small PWSs serving 10,000 people or fewer.
Section 5.3.1, below, provides an overview of detection rates for contaminants in the nitrosamine
group. Section 5.3.2 discusses the range of observed concentrations. Section 5.3.3 breaks down
the data at the level of the PWS and the sampling location (EP versus MR point in the
distribution system). Section 5.3.4 provides an analysis of the estimated populations exposed to
contamination at PWSs participating in the UCMR 2 survey.
5.3.1 Rates of Detection
Exhibit 5.12 and Exhibit 5.13 show the detection rates (i.e., the percentage of samples in which a
nitrosamine was measured at or above its MRL) for each of these six compounds and for the
group as a whole. Results are broken down by PWS size and source water type in Exhibit 5.12
and by disinfectant type in Exhibit 5.13.
Six-Year Review 3 5-12
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 5.12: Nitrosamine Detection Rates for UCMR 2 Data by PWS Size and Source Water Type
PWS Size1
Source Water
Type
All Six
Nitrosamines:
% Sample
Detections
NDBA:
% Sample
Detections
NDEA:
% Sample
Detections
NDMA:
% Sample
Detections
NDPA:
% Sample
Detections
NMEA:
% Sample
Detections
NPYR:
% Sample
Detections
Small
Surface Water
13.8%
(245 of 1,769)
0%
(0 of 1,769)
0%
(Oof 1,769)
13.8%
(245 of 1,769)
0%
(Oof 1,769)
0.2%
(3 of 1,769)
0.6%
(11 of 1,769)
Small
Ground Water
1.5%
(18 of 1,237)
0.2%
(3 of 1,237)
0.1%
(1 of 1,237)
1.1%
(14 of 1,237)
0%
(Oof 1,237)
0%
(Oof 1,237)
0%
(Oof 1,237)
Small
Mixed Water
35.7%
(81 of 227)
0%
(0 of 227)
0%
(0 of 227)
35.7%
(81 of 227)
0%
(0 of 227)
0%
(0 of 227)
0%
(0 of 227)
Large
Surface Water
16.3%
(187 of 1,146)
0%
(0 of 1,146)
0.1%
(1 of 1,146)
16%
(183 of 1,146)
0%
(Oof 1,146)
0%
(Oof 1,146)
0.5%
(6 of 1,146)
Large
Ground Water
2.4%
(51 of 2,084)
0%
(0 of 2,082)
0.1%
(3 of 2,070)
2.3%
(48 of 2,082)
0%
(0 of 2,082)
0%
(0 of 2,082)
0%
(0 of 2,082)
Large
Mixed Water
16.3%
(128 of 787)
0%
(0 of 787)
0.6%
(5 of 787)
15.5%
(122 of 787)
0%
(0 of 787)
0%
(0 of 787)
0.1%
(1 of 787)
Very Large
Surface Water
18.3%
(536 of 2,924)
0%
(0 of 2,921)
0.4%
(11 of 2,923)
17.7%
(518 of 2,920)
0%
(0 of 2,922)
0%
(0 of 2,921)
0.5%
(15 of 2,921)
Very Large
Ground Water
2.2%
(67 of 3,041)
0.1%
(3 of 3,039)
0.2%
(7 of 3,041)
1.9%
(57 of 3,039)
0%
(0 of 3,041)
0%
(0 of 3,039)
0.03%
(1 of 3,039)
Very Large
Mixed Water
12.3%
(594 of 4,838)
0.06%
(3 of 4,835)
0.4%
(18 of 4,838)
11.9%
(573 of 4,833)
0%
(0 of 4,838)
0%
(0 of 4,835)
0.1%
(7 of 4,835)
All
Surface Water
16.6%
(968 of 5,839)
0%
(0 of 5,836)
0.2%
(12 of 5,838)
16.2%
(946 of 5,835)
0%
(0 of 5,837)
0.1%
(3 of 5,836)
0.5%
(32 of 5,836)
All
Ground Water
2.1%
(136 of 6,362)
0.09%
(6 of 6,358)
0.2%
(11 of 6,348)
1.9%
(119 of 6,358)
0%
(0 of 6,360)
0%
(0 of 6,358)
0.02%
(1 of 6,358)
All
Mixed Water
13.7%
(803 of 5,852)
0.05%
(3 of 5,849)
0.4%
(23 of 5,852)
13.3%
(776 of 5,847)
0%
(0 of 5,852)
0%
(0 of 5,849)
0.1%
(8 of 5,849)
Note: See Exhibit 5.1 for the MRLs for each nitrosamine.
1) Small = serving < 10,000; Large = serving 10,001 - 100,000; Very Large = serving > 100,000.
Six-Year Review 3
Technical Support Document for Nitrosamines
5-13
December 2016
-------
Exhibit 5.13: Nitrosamine Detection Rates for UCMR 2 Data by Disinfectant
Disinfectant
All Six
Nitrosamines:
% Sample
Detections
NDBA:
% Sample
Detections
NDEA:
% Sample
Detections
NDMA:
% Sample
Detections
NDPA:
% Sample
Detections
NMEA:
% Sample
Detections
NPYR:
% Sample
Detections
Any chloramine
34.5%
(1,301 of 3,768)
0%
(0 of 3,765)
0.3%
(11 of 3,767)
34.1%
(1,284 of 3,764)
0%
(0 of 3,767)
0%
(0 of 3,765)
0.7%
(25 of 3,765)
Chlorine or
other
4.3%
(593 of 13,831)
0.07%
(9 of 13,824)
0.2%
(29 of 13,817)
4.0%
(549 of 13,822)
0%
(0 of 13,828)
0.02%
(3 of 13,824)
0.1%
(16 of 13,824)
No disinfection
2.9%
(13 of 454)
0%
(0 of 454)
1.3%
(6 of 454)
1.8%
(8 of 454)
0%
(0 of 454)
0%
(0 of 454)
0%
(0 of 454)
All Disinfectant
Categories
10.6%
(1,907 of 18,053)
0.05%
(9 of 18,043)
0.3%
(46 of 18,038)
10.2%
(1,841 of 18,040)
0%
(0 of 49)
0.02%
(3 of 18,043)
0.2%
(41 of 18,043)
Note: See Exhibit 5.1 for the MRLs for each nitrosamine.
Six-Year Review 3
Technical Support Document for Nitrosamines
5-14
December 2016
-------
Among the six nitrosamines, NDMA was by far the most frequently detected. Across all PWS
types, NDMA was detected in 10.2 percent (1,841 of 18,040) of the samples. By comparison, the
detection rates for the other five nitrosamines were between 0 percent and 0.3 percent (46 of
18,038). NDMA accounted for most of the 10.6 percent (1,907 of 18,053) of samples reported to
have one or more nitrosamine detection.
Exhibit 5.12 shows that the NDMA detection rate was significantly higher in samples from
PWSs using surface water (16.2 percent; 946 of 5,835) or mixed water sources (13.3 percent;
776 of 5,847) than in samples from PWSs using ground water only (1.9 percent; 119 of 6,358).
This pattern for NDMA of higher frequency of detections in surface water and mixed water PWS
samples was similar for all three PWS size categories. (Note that the markedly higher frequency
of 35.7 percent [81 of 227] in samples from the small mixed water PWSs may be due to the
relatively small number of PWSs, and thus samples, in that category.)
The NDMA data in Exhibit 5.13 show a clear pattern of occurrence associated with the PWS's
residual disinfection type. As expected (see Chapter 6), NDMA detection rates are much higher
in samples from PWSs using chloramines (34.1 percent; 1,284 of 3,764) than those from PWSs
using free chlorine or other disinfectants (4.0 percent; 549 of 13,822) or samples from non-
disinfecting PWSs (1.8 percent; 8 of 454). Significantly lower detection rates of NDMA in non-
disinfected samples than in disinfected samples indicate that NDMA is formed as a result of
disinfection practices, rather than occurring in source water.
Because of the extremely low sample detection rates for the other five nitrosamines, it is not
possible to discern patterns as clearly as those seen for NDMA. Caution should be exercised in
drawing any conclusions regarding patterns of occurrence in different types of PWSs without
additional information, possibly including (though not limited to) information about source water
quality, treatment processes and disinfection operations.
5.3.2 Detected Concentrations
Exhibit 5.14 shows for each of the six nitrosamines the minimum, 5th percentile, median, 95th
percentile and maximum concentration for the subset of samples with detections at or above the
MRL. Exhibit 5.15 and Exhibit 5.16 break these results down by sample location (EP and MR,
respectively). The median concentrations of NDEA and NDBA for all samples are higher than
the median concentrations of NDMA, NMEA and NPYR; this reflects at least in part the higher
MRLs of NDEA and NDBA. Among the nitrosamines, NDEA's 95th percentile concentration is
highest (followed by NDMA), and NDMA's maximum concentration is by far the highest
(followed by NDEA). Across disinfection types there is not much variation in median
concentration values, but the highest concentrations were generally found in chloraminating
PWSs.
Six-Year Review 3 5-15
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 5.14: Summary of Nitrosamine Concentrations in Samples with Detections
at All Sampling Locations, by Disinfectant Type
Disinfectant Type
Summary
Statistic
NDBA
NDEA
NDMA
NDPA
NMEA
NPYR
All
# Detects
9
46
1,841
0
3
41
All
Min (ng/L)
4
5
2
N/A
3.6
2.1
All
5%tile (ng/L)
4.2
5.2
2.1
N/A
3.7
2.1
All
Median (ng/L)
6.7
7.1
4.1
N/A
4.5
3.9
All
95%tile (ng/L)
16.4
46.8
26.2
N/A
4.9
14.4
All
Max (ng/L)
20.6
100.0
630.0
N/A
4.9
23.8
Any chloramine
# Detects
0
11
1,284
0
0
25
Any chloramine
Min (ng/L)
N/A
5.7
2
N/A
N/A
2.1
Any chloramine
5%tile (ng/L)
N/A
5.9
2.1
N/A
N/A
2.1
Any chloramine
Median (ng/L)
N/A
13.0
4.2
N/A
N/A
3.7
Any chloramine
95%tile (ng/L)
N/A
92.5
26.6
N/A
N/A
12.7
Any chloramine
Max (ng/L)
N/A
100.0
630.0
N/A
N/A
17.2
Chlorine or other
# Detects
9
29
549
0
3
16
Chlorine or other
Min (ng/L)
4
5
2
N/A
3.6
2.1
Chlorine or other
5%tile (ng/L)
4.2
5.1
2.1
N/A
3.7
2.2
Chlorine or other
Median (ng/L)
6.7
7.0
4.0
N/A
4.5
5.1
Chlorine or other
95%tile (ng/L)
16.4
32.6
26.1
N/A
4.9
13.2
Chlorine or other
Max (ng/L)
20.6
50.0
84.6
N/A
4.9
23.8
No disinfection
# Detects
0
6
8
0
0
0
No disinfection
Min (ng/L)
N/A
6
2
N/A
N/A
N/A
No disinfection
5%tile (ng/L)
N/A
6.2
2.1
N/A
N/A
N/A
No disinfection
Median (ng/L)
N/A
6.9
3.4
N/A
N/A
N/A
No disinfection
95%tile (ng/L)
N/A
8.3
7.4
N/A
N/A
N/A
No disinfection
Max (ng/L)
N/A
8.4
8.0
N/A
N/A
N/A
Note: This table shows statistics for samples in the UCMR 2 dataset for nitrosamines with concentrations at or above the MRL only.
See Exhibit 5.1 for the MRLs for each nitrosamine.
Six-Year Review 3 5-16
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 5.15: Summary of Nitrosamine Concentrations in Samples from EP
Locations, with Detections by Disinfectant Type
Disinfectant Type
Summary
Statistic
NDBA
NDEA
NDMA
NDPA
NMEA
NPYR
All
# Detects
8
28
736
0
0
17
All
Min (ng/L)
4
5
2
N/A
N/A
2.1
All
5%tile (ng/L)
4.1
5.3
2.1
N/A
N/A
2.1
All
Median (ng/L)
6.5
7.0
4.2
N/A
N/A
3.0
All
95%tile (ng/L)
16.9
26.6
27.7
N/A
N/A
7.0
All
Max (ng/L)
20.6
50.0
470.0
N/A
N/A
7.6
Any chloramine
# Detects
0
5
503
0
0
9
Any chloramine
Min (ng/L)
N/A
5.7
2
N/A
N/A
2.1
Any chloramine
5%tile (ng/L)
N/A
5.8
2.1
N/A
N/A
2.1
Any chloramine
Median (ng/L)
N/A
6.6
4.6
N/A
N/A
3.0
Any chloramine
95%tile (ng/L)
N/A
27.2
27.9
N/A
N/A
5.8
Any chloramine
Max (ng/L)
N/A
28.0
470.0
N/A
N/A
6.1
Chlorine or other
# Detects
8
18
227
0
0
8
Chlorine or other
Min (ng/L)
4
5
2
N/A
N/A
2.2
Chlorine or other
5%tile (ng/L)
4.1
5.3
2.1
N/A
N/A
2.2
Chlorine or other
Median (ng/L)
6.5
7.5
3.6
N/A
N/A
4.5
Chlorine or other
95%tile (ng/L)
16.9
26.2
26.4
N/A
N/A
7.3
Chlorine or other
Max (ng/L)
20.6
50.0
61.7
N/A
N/A
7.6
No disinfection
# Detects
0
5
6
0
0
0
No disinfection
Min (ng/L)
N/A
6
2
N/A
N/A
N/A
No disinfection
5%tile (ng/L)
N/A
6.1
2.3
N/A
N/A
N/A
No disinfection
Median (ng/L)
N/A
7.1
3.4
N/A
N/A
N/A
No disinfection
95%tile (ng/L)
N/A
8.3
7.6
N/A
N/A
N/A
No disinfection
Max (ng/L)
N/A
8.4
8.0
N/A
N/A
N/A
Note: This table shows statistics for samples in the UCMR 2 dataset for nitrosamines with concentrations at or above the MRL only.
See Exhibit 5.1 for the MRLs for each nitrosamine.
Six-Year Review 3 5-17
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 5.16: Summary of Nitrosamine Concentrations in Samples from MR
Locations, by Disinfectant Type
Disinfectant Type
Summary
Statistic
NDBA
NDEA
NDMA
NDPA
NMEA
NPYR
All
# Detects
1
18
1,105
0
3
24
All
Min (ng/L)
9
5
2
N/A
3.6
2.1
All
5%tile (ng/L)
9.3
5.2
2.1
N/A
3.7
2.2
All
Median (ng/L)
9.3
7.6
4.1
N/A
4.5
4.6
All
95%tile (ng/L)
9.3
87.3
25.0
N/A
4.9
16.8
All
Max (ng/L)
9.3
100.0
630.0
N/A
4.9
23.8
Any chloramine
# Detects
0
6
781
0
0
16
Any chloramine
Min (ng/L)
N/A
6.0
2
N/A
N/A
2.1
Any chloramine
5%tile (ng/L)
N/A
6.2
2.1
N/A
N/A
2.4
Any chloramine
Median (ng/L)
N/A
23.5
4.1
N/A
N/A
4.3
Any chloramine
95%tile (ng/L)
N/A
96.3
24.0
N/A
N/A
15.1
Any chloramine
Max (ng/L)
N/A
100.0
630.0
N/A
N/A
17.2
Chlorine or other
# Detects
1
11
322
0
3
8
Chlorine or other
Min (ng/L)
9
5
2
N/A
3.6
2.1
Chlorine or other
5%tile (ng/L)
9.3
5.1
2.1
N/A
3.7
2.2
Chlorine or other
Median (ng/L)
9.3
7.0
4.2
N/A
4.5
5.2
Chlorine or other
95%tile (ng/L)
9.3
31.5
25.4
N/A
4.9
18.9
Chlorine or other
Max (ng/L)
9.3
37.0
84.6
N/A
4.9
23.8
No disinfection
# Detects
0
1
2
0
0
0
No disinfection
Min (ng/L)
N/A
6.6
2
N/A
N/A
N/A
No disinfection
5%tile (ng/L)
N/A
6.6
2.1
N/A
N/A
N/A
No disinfection
Median (ng/L)
N/A
6.6
2.9
N/A
N/A
N/A
No disinfection
95%tile (ng/L)
N/A
6.6
3.7
N/A
N/A
N/A
No disinfection
Max (ng/L)
N/A
6.6
3.8
N/A
N/A
N/A
Note: This table shows statistics for samples in the UCMR 2 dataset for nitrosamines with concentrations at or above the MRL
only. See Exhibit 5.1 for the MRLs for each nitrosamine.
Exhibit 5.17 shows the mean concentrations among detections for each nitrosamine. The mean
concentrations vary and have a wider range of values than the median concentrations shown in
Exhibit 5.14. For example, while the median concentrations of NDEA and NDBA are
approximately 7 ng/L (see Exhibit 5.14), the mean concentration of NDEA is significantly higher
at 15.3 ng/L, almost twice that of the mean concentration of NDBA, 8.4 ng/L. For NDEA and (to
a lesser extent) NDMA, mean concentrations vary considerably across the three disinfection
categories presented. Exhibit 5.14 and Exhibit 5.17 together show that NDMA concentrations are
significantly higher in samples from disinfecting PWSs than in samples from non-disinfecting
PWSs. This observation, along with the higher frequency of detection at disinfecting PWSs
(Exhibit 5.13), provides further evidence that NDMA occurs significantly as a byproduct of
disinfection. NDEA too has higher mean and maximum concentrations at disinfecting PWSs,
though Exhibit 5.13 shows a higher frequency of detection in samples from non-disinfecting
Six-Year Review 3 5-18
Technical Support Document for Nitrosamines
December 2016
-------
PWSs. NMEA, NDBA and NPYR were not detected at all in samples from non-disinfecting
PWSs.
Exhibit 5.17: Mean Nitrosamine Concentrations (in ng/L) in Samples with
Detections by Disinfectant Type
Sampling
Points
Disinfectant
Type
Mean of
NDBA
Mean of
NDEA
Mean of
NDMA
Mean of
NDPA
Mean of
NMEA
Mean of
NPYR
All
All samples
8.4
15.3
8.8
N/A
4.3
5.2
All
Any chloramine
N/A
28.7
9.2
N/A
N/A
4.8
All
Chlorine or other
8.4
11.9
8.0
N/A
4.3
5.9
All
No disinfection
N/A
7.2
4.0
N/A
N/A
N/A
EP locations
All samples
8.2
11.0
8.5
N/A
N/A
3.9
EP locations
Any chloramine
N/A
14.1
9.2
N/A
N/A
3.3
EP locations
Chlorine or other
8.2
11.1
7.3
N/A
N/A
4.6
EP locations
No disinfection
N/A
7.3
4.3
N/A
N/A
N/A
MR locations
All samples
9.3
22.0
9.0
N/A
4.3
6.1
MR locations
Any chloramine
N/A
40.8
9.2
N/A
N/A
5.6
MR locations
Chlorine or other
9.3
13.2
8.6
N/A
4.3
7.2
MR locations
No disinfection
N/A
6.6
2.9
N/A
N/A
N/A
Note: This table shows statistics for samples in the UCMR 2 dataset for nitrosamines with concentrations at or above the MRL only.
See Exhibit 5.1 for the MRLs for each nitrosamine.
5.3.3 Sample Location Analysis
The preceding sections presented a sample-level analysis of UCMR 2 results. This section
describes the frequency of detection of nitrosamines expressed as a percentage of PWSs, EP
locations and MR locations.
Exhibit 5.18 through Exhibit 5.23 show the percentage of PWSs, EP locations and MR locations
with detections of nitrosamines at least once during the UCMR 2 sampling period. Results are
organized by PWS size, source water type and disinfectant type. It is important to note that the
UCMR 2 survey represents a combination of samples and census, and therefore summary
statistics from the survey should not be interpreted as a simple surrogate for national occurrence.
Modeled national occurrence and exposure estimates based on the UCMR 2 data for NDMA,
taking into account appropriate weighting of these results with respect to the national distribution
of PWSs with various source water types, sizes and disinfection practice characteristics, are
presented in Section 5.5.
All three analyses (for PWSs, EP locations and MR locations) show similar trends, with higher
detection rates at surface water and mixed water systems than at ground water systems, and
higher detection rates at PWSs using chloramine disinfection than at PWSs with other
disinfection practices. For example, Exhibit 5.19 shows that 73.5 percent (208 of 283) of PWSs
using chloramines detected NDMA at least once, with only 13.3 percent (114 of 856) of PWSs
using other disinfectants detecting NDMA.
Six-Year Review 3 5-19
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 5.20, Exhibit 5.21, Exhibit 5.22 and Exhibit 5.23 show that, compared to EPs, MR
locations have higher NDMA detection rates for all PWS sizes, source water types and
disinfectant types. This observation is especially significant for chloraminating PWSs: 31.2
percent (208 of 667) of chloraminating PWSs detected NDMA at EPs, but 61.9 percent (299 of
483) detected NDMA at MR locations, indicating that formation of NDMA continues in the
distribution system. This conclusion is consistent with findings from the literature presented in
Chapter 6. In all three analyses, NDBA, NDPA and NMEA are detected at very low levels.
Six-Year Review 3 5-20
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 5.18: Percentage of PWSs in the UCMR 2 Dataset Detecting Nitrosamines At Least Once by PWS Size and
Source Water Type
PWS Size1
Source Water
Type
Nitrosamine Group:
% PWSs with Detects
NDBA:
% PWSs with
Detects
NDEA:
% PWSs with
Detects
NDMA:
% PWSs with
Detects
NDPA:
% PWSs with
Detects
NMEA:
% PWSs with
Detects
NPYR:
% PWSs with
Detects
Small
Surface Water
27.5%
(61 of 222)
0%
(0 of 222)
0%
(0 of 222)
27.5%
(61 of 222)
0%
(0 of 222)
1.4%
(3 of 222)
1.4%
(3 of 222)
Small
Ground Water
5.4%
(13 of 240)
0.8%
(2 of 240)
0.4%
(1 of 240)
4.2%
(10 of 240)
0%
(0 of 240)
0%
(0 of 240)
0%
(0 of 240)
Small
Mixed Water
72.2%
(13 of 18)
0%
(Oof 18)
0%
(Oof 18)
72.2%
(13 of 18)
0%
(Oof 18)
0%
(0 of 18)
0%
(Oof 18)
Large
Surface Water
36.7%
(44 of 120)
0%
(0 of 120)
0.8%
(1 of 120)
35%
(42 of 120)
0%
(0 of 120)
0%
(Oof 120)
4.2%
(5 of 120)
Large
Ground Water
9.2%
(15 of 163)
0%
(0 of 163)
0.6%
(1 of 163)
9.2%
(15 of 163)
0%
(0 of 163)
0%
(Oof 163)
0%
(0 of 163)
Large
Mixed Water
43.2%
(16 of 37)
0%
(0 of 37)
5.4%
(2 of 37)
37.8%
(14 of 37)
0%
(0 of 37)
0%
(0 of 37)
2.7%
(1 of 37)
Very Large
Surface Water
44.5%
(97 of 218)
0%
(0 of 218)
2.3%
(5 of 218)
42.7%
(93 of 218)
0%
(0 of 218)
0%
(0 of 218)
3.2%
(7 of 218)
Very Large
Ground Water
30.6%
(22 of 72)
1.4%
(1 of 72)
7%
(5 of 72)
25%
(18 of 72)
0%
(0 of 72)
0%
(0 of 72)
1.4%
(1 of 72)
Very Large
Mixed Water
57.4%
(62 of 108)
1.9%
(2 of 108)
10.2%
(11 of 108)
53.7%
(58 of 108)
0%
(0 of 108)
0%
(Oof 108)
3.7%
(4 of 108)
All
Surface Water
36.1%
(202 of 560)
0%
(0 of 560)
1.1%
(6 of 560)
35%
(196 of 560)
0%
(0 of 560)
0.5%
(3 of 560)
2.7%
(15 of 560)
All
Ground Water
10.5%
(50 of 475)
0.6%
(3 of 475)
1.5%
(7 of 475)
9.1%
(43 of 475)
0%
(0 of 475)
0%
(0 of 475)
0.2%
(1 of 475)
All
Mixed Water
55.8%
(91 of 163)
1.2%
(2 of 163)
8%
(13 of 163)
52.2%
(85 of 163)
0%
(0 of 163)
0%
(0 of 163)
3.1%
(5 of 163)
Note: See Exhibit 5.1 for the MRLs for each nitrosamine.
1) Small = serving < 10,000; Large = serving 10,001 - 100,000; Very Large = serving > 100,000.
Six-Year Review 3
Technical Support Document for Nitrosamines
5-21
December 2016
-------
Exhibit 5.19: Percentage of PWSs in the UCMR 2 Dataset Detecting Nitrosamines At Least Once by Disinfectant Type
Disinfectant
Nitrosamine Group:
% PWSs with
Detects
NDBA:
% PWSs with
Detects
NDEA:
% PWSs with
Detects
NDMA:
% PWSs with
Detects
NDPA:
% PWSs with
Detects
NMEA:
% PWSs with
Detects
NPYR:
% PWSs with
Detects
Any chloramine
75.3%
(213 of 283)
0%
(0 of 283)
3.5%
(10 of 283)
73.5%
(208 of 283)
0%
(0 of 283)
0%
(0 of 283)
3.5%
(10 of 283)
Chlorine or other
15%
(128 of 856)
0.6%
(5 of 856)
1.9%
(16 of 856)
13.3%
(114 of 856)
0%
(0 of 856)
0.4%
(3 of 856)
1.3%
(11 of 856)
No disinfection
3.4%
(2 of 59)
0%
(0 of 59)
0%
(0 of 59)
3.4%
(2 of 59)
0%
(0 of 59)
0%
(0 of 59)
0%
(0 of 59)
All Disinfectant
Categories
28.6%
(343 of 1,198)
0.4%
(5 of 1,198)
2.2%
(26 of 1,198)
27.0%
(324 of 1,198)
0%
(0 of 1,198)
0.3%
(3 of 1,198)
1.8%
(21 of 1,198)
Note: See Exhibit 5.1 for the MRLs for each nitrosamine.
Six-Year Review 3
Technical Support Document for Nitrosamines
5-22
December 2016
-------
Exhibit 5.20: Percent of Entry Points in the UCMR 2 Dataset Detecting Nitrosamines At Least Once by PWS Size and
Source Water Type
PWS Size1
Source Water
Type
Nitrosamine Group:
% EPs with Detects
NDBA: % EPs
with Detects
NDEA: % EPs
with Detects
NDMA: % EPs
with Detects
NDPA: % EPs
with Detects
NMEA: % EPs
with Detects
NPYR: % EPs
with Detects
Small
Surface Water
19.4%
(45 of 232)
0%
(0 of 232)
0%
(0 of 232)
19.4%
(45 of 232)
0%
(0 of 232)
0%
(0 of 232)
0.9%
(2 of 232)
Small
Ground Water
2.3%
(9 of 399)
0.8%
(3 of 399)
0.3%
(1 of 399)
1.3%
(5 of 399)
0%
(0 of 399)
0%
(0 of 399)
0%
(0 of 399)
Small
Mixed Water
39.1%
(18 of 46)
0%
(0 of 46)
0%
(0 of 46)
39.1%
(18 of 46)
0%
(0 of 46)
0%
(0 of 46)
0%
(0 of 46)
Large
Surface Water
22.2%
(34 of 153)
0%
(0 of 153)
0%
(0 of 153)
20.9%
(32 of 153)
0%
(Oof 153)
0%
(0 of 153)
2.61%
(4 of 153)
Large
Ground Water
2.7%
(21 of 770)
0%
(0 of 770)
0.3%
(2 of 770)
2.5%
(19 of 770)
0%
(0 of 770)
0%
(0 of 770)
0%
(0 of 770)
Large
Mixed Water
19.4%
(39 of 201)
0%
(0 of 201)
2%
(4 of 201)
17.4%
(35 of 201)
0%
(0 of 201)
0%
(0 of 201)
0%
(0 of 201)
Very Large
Surface Water
19.8%
(79 of 400)
0%
(0 of 400)
1%
(4 of 400)
18.3%
(73 of 400)
0%
(0 of 400)
0%
(0 of 400)
1.5%
(6 of 400)
Very Large
Ground Water
2.6%
(27 of 1,042)
0.2%
(2 of 1,042)
0.3%
(3 of 1,042)
2.2%
(23 of 1,042)
0%
(Oof 1,042)
0%
(0 of 1,042)
0%
(0 of 1,042)
Very Large
Mixed Water
8.9%
(126 of 1,424)
0.2%
(3 of 1,423)
0.8%
(11 of 1,424)
8%
(114 of 1,423)
0%
(Oof 1,424)
0%
(0 of 1,423)
0.2%
(3 of 1,423)
All
Surface Water
20.1%
(158 of 785)
0%
(0 of 785)
0.5%
(4 of 785)
19.1%
(150 of 785)
0%
(0 of 785)
0%
(0 of 785)
1.5%
(12 of 785)
All
Ground Water
2.6%
(57 of 2,211)
0.2%
(5 of 2,211)
0.3%
(6 of 2,211)
2.1%
(47 of 2,211)
0%
(0 of 2,211)
0%
(0 of 2,211)
0%
(0 of 2,211)
All
Mixed Water
11%
(183 of 1,671)
0.2%
(3 of 1,670)
0.9%
(15 of 1,671)
10%
(167 of 1,670)
0%
(0 of 1,671)
0%
(0 of 1,670)
0.2%
(3 of 1,670)
Note: See Exhibit 5.1 for the MRLs for each nitrosamine.
1) Small = serving < 10,000; Large = serving 10,001 - 100,000; Very Large = serving > 100,000.
Six-Year Review 3
Technical Support Document for Nitrosamines
5-23
December 2016
-------
Exhibit 5.21: Percent of Entry Points in the UCMR 2 Dataset Detecting Nitrosamines At Least Once by Disinfectant Type
Disinfectant
Nitrosamine Group:
% EPs with Detects
NDBA: % EPs
with Detects
NDEA: % EPs
with Detects
NDMA: % EPs
with Detects
NDPA: % EPs
with Detects
NMEA: % EPs
with Detects
NPYR: % EPs
with Detects
Any chloramine
32.2%
(215 of 667)
0%
(0 of 667)
0.8%
(5 of 667)
31.2%
(208 of 667)
0%
(0 of 667)
0%
(0 of 667)
1.1%
(7 of 667)
Chlorine or other
4.5%
(172 of 3,787)
0.2%
(8 of 3,786)
0.4%
(15 of 3,787)
4%
(150 of 3,786)
0%
(0 of 3,787)
0%
(0 of 3,786)
0.2%
(8 of 3,786)
No disinfection
5.2%
(11 of 213)
0%
(0 of 213)
2.4%
(5 of 213)
2.8%
(6 of 213)
0%
(0 of 213)
0%
(0 of 213)
0%
(0 of 213)
All Disinfectant
Categories
8.5%
(398 of 4,667)
0.2%
(8 of 4,666)
0.5%
(25 of 4,667)
7.8%
(364 of 4,666)
0%
(0 of 4,667)
0%
(0 of 4,666)
0.3%
(15 of 4,666)
Note: See Exhibit 5.1 for the MRLs for each nitrosamine.
Six-Year Review 3
Technical Support Document for Nitrosamines
5-24
December 2016
-------
Exhibit 5.22: Percent of Maximum Residence Time Locations in the UCMR 2 Dataset Detecting Nitrosamines At Least
Once by Size and Source Water Type
PWS Size1
Source Water
Type
Nitrosamine
Group: % MRs with
Detects
NDBA: % MRs
with Detects
NDEA: % MRs
with Detects
NDMA: % MRs
with Detects
NDPA: % MRs
with Detects
NMEA: % MRs
with Detects
NPYR: % MRs
with Detects
Small
Surface Water
24.8%
(56 of 226)
0%
(0 of 226)
0%
(0 of 226)
24.8%
(56 of 226)
0%
(0 of 226)
1.3%
(3 of 226)
1.3%
(3 of 226)
Small
Ground Water
2.9%
(7 of 238)
0%
(0 of 238)
0%
(0 of 238)
2.9%
(7 of 238)
0%
(0 of 238)
0%
(0 of 238)
0%
(0 of 238)
Small
Mixed Water
48.7%
(19 of 39)
0%
(0 of 39)
0%
(0 of 39)
48.7%
(19 of 39)
0%
(0 of 39)
0%
(0 of 39)
0%
(0 of 39)
Large
Surface Water
31.7%
(46 of 145)
0%
(0 of 145)
0.7%
(1 of 145)
31%
(45 of 145)
0%
(0 of 145)
0%
(Oof 145)
1.4%
(2 of 145)
Large
Ground Water
4.8%
(15 of 311)
0%
(0 of 311)
0.3%
(1 of 311)
4.5%
(14 of 311)
0%
(0 of 311)
0%
(0 of 311)
0%
(0 of 311)
Large
Mixed Water
34.1%
(30 of 88)
0%
(0 of 88)
0%
(0 of 88)
33%
(29 of 88)
0%
(0 of 88)
0%
(0 of 88)
1.1%
(1 of 88)
Very Large
Surface Water
35.7%
(131 of 367)
0%
(0 of 367)
1.4%
(5 of 367)
34.1%
(125 of 367)
0%
(0 of 367)
0%
(0 of 367)
2.2%
(8 of 367)
Very Large
Ground Water
5.7%
(28 of 490)
0.2%
(1 of 493)
0.8%
(4 of 493)
4.5%
(22 of 493)
0%
(0 of 493)
0%
(0 of 493)
0.2%
(1 of 493)
Very Large
Mixed Water
30.6%
(151 of 493)
0%
(0 of 493)
1.2%
(6 of 493)
29.2%
(144 of 493)
0%
(0 of 493)
0%
(0 of 493)
0.8%
(4 of 493)
All
Surface Water
31.6%
(233 of 738)
0%
(0 of 738)
0.8%
(6 of 738)
30.6%
(226 of 738)
0%
(0 of 738)
0.4%
(3 of 738)
1.8%
(13 of 738)
All
Ground Water
4.8%
(50 of 1,039)
0.1%
(1 of 1,039)
0.5%
(5 of 1,039)
4.1%
(43 of 1,039)
0%
(0 of 1,039)
0%
(0 of 1,039)
0.1%
(1 of 1,039)
All
Mixed Water
32.3%
(200 of 620)
0%
(0 of 620)
1%
(6 of 620)
31%
(192 of 620)
0%
(0 of 620)
0%
(0 of 620)
0.8%
(5 of 620)
Note: See Exhibit 5.1 for the MRLs for each nitrosamine.
1) Small = serving < 10,000; Large = serving 10,001 - 100,000; Very Large = serving > 100,000.
Six-Year Review 3
Technical Support Document for Nitrosamines
5-25
December 2016
-------
Exhibit 5.23: Percent of Maximum Residence Time Locations in the UCMR 2 Dataset Detecting Nitrosamines At Least
Once by Disinfectant
Disinfectant
Nitrosamine Group:
% MRs with Detects
NDBA: % MRs
with Detects
NDEA: % MRs
with Detects
NDMA: % MRs
with Detects
NDPA: % MRs
with Detects
NMEA: % MRs
with Detects
NPYR: % MRs
with Detects
Any chloramine
63.4%
(306 of 483)
0%
(0 of 483)
1.2%
(6 of 483)
61.9%
(299 of 483)
0%
(0 of 483)
0%
(0 of 483)
2.3%
(11 of 483)
Chlorine or other
9.2%
(175 of 1,893)
0.1%
(1 of 1,893)
0.5%
(10 of 1,893)
8.5%
(160 of 1,893)
0%
(Oof 1,893)
0.2%
(3 of 1,893)
0.4%
(8 of 1,893)
No disinfection
9.5%
(2 of 21)
0%
(0 of 21)
4.8%
(1 of 21)
9.52%
(2 of 21)
0%
(0 of 21)
0%
(0 of 21)
0%
(0 of 21)
All Disinfectant
Categories
20.2%
(483 of 2,397)
0.04%
(1 of 2,397)
0.7%
(17 of 2,397)
19.2%
(461 of 2,397)
0%
(0 of 2,397)
0.1%
(3 of 2,397)
0.8%
(19 of 2,397)
Note: See Exhibit 5.1 for the MRLs for each nitrosamine.
Six-Year Review 3
Technical Support Document for Nitrosamines
5-26
December 2016
-------
5.3.4 Population Affected
Exhibit 5.24 through Exhibit 5.29 show the percentage of the population served by PWSs, EP
locations and MR locations, respectively, for which at least one detection was reported during
the UCMR 2 sampling period. Results are organized by PWS size, source water type and
disinfectant type. It is important to note that the UCMR 2 monitoring results represent a
combination of sample (statistically representative samples of small and large systems) and
census (a census of very large systems), and therefore summary statistics from the survey should
not be interpreted as a simple surrogate for national occurrence. Modeled national occurrence
and exposure estimates based on the UCMR 2 data for NDMA, taking into account appropriate
weighting of these results with respect to the national distribution of PWSs with various source
water types, sizes and disinfection practice characteristics, are presented in Section 5.5.
All three analyses (for PWSs, EP locations and MR locations) show similar trends: Rates of
nitrosamine exposure are higher for populations served by surface water and mixed water
systems than for populations served by ground water systems; rates of exposure are also higher
for populations served by systems employing chloramine disinfection than for populations
served by systems in other disinfection categories. For example, Exhibit 5.25 shows that 79.1
percent (52 million of 65 million) of the population served by PWSs using chloramines was
served by PWS where NDMA was detected at least once, while only 15.0 percent (14 million of
92 million) of the population served by PWSs using other disinfectants was served by PWSs
detecting NDMA. Exhibit 5.26 through Exhibit 5.29 show that more people are served by water
from MR locations with NDMA detections than are served by water from EP locations with
NDMA detections, and that this is true across all PWS sizes, source water types and disinfectant
types. This observation is especially significant for populations served by chloraminating PWSs,
where 27.4 percent (8.7 million of 32 million) of the population was served by PWSs detecting
NDMA at EPs, but 64.4 percent (17 million of 26 million) of the population was served by
PWSs that detected NDMA at the MR locations, indicating that formation of NDMA continues
in the distribution system. This conclusion is consistent with the literature findings presented in
Chapter 6. In all three analyses, the percentage of the population served by PWSs detecting
NDBA, NDPA and NMEA was very low.
Six-Year Review 3 5-27
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 5.24: Percentage of Population Served by PWSs in the UCMR 2 Dataset Detecting Nitrosamines At Least Once,
by PWS Size and Source Water Type
PWS Size1
Source Water
Type
Nitrosamine
Group: % Pop.
Served by PWSs
with Detects
NDBA: % Pop.
Served by
PWSs with
Detects
NDEA: % Pop.
Served by
PWSs with
Detects
NDMA: % Pop.
Served by
PWSs with
Detects
NDPA: % Pop.
Served by
PWSs with
Detects
NMEA: % Pop.
Served by
PWSs with
Detects
NPYR: % Pop.
Served by
PWSs with
Detects
Small
Surface Water
33.6%
(193K of 574K)
0%
(0 of 574K)
0%
(0 of 574K)
33.6%
(193K of 574K)
0%
(0 of 574K)
0.8%
(4.5K, of 574K)
2.5%
(15K of 574K)
Small
Ground Water
8.6%
(47K of 548K)
2.6%
(14K of 548K)
1.2%
(6.6K of 548K)
4.8%
(26K of 548K)
0%
(0 of 548K)
0%
(0 of 548K))
0%
(0 of 548K)
Small
Mixed Water
76.8%
(68K of 87K)
0%
(0 of 87K)
0%
(0 of 87K)
76.8%
(67K of 87K)
0%
(0 of 87K)
0%
(0 of 87K)
0%
(0 of 87K)
Large
Surface Water
38.1%
(2.0M of 5.2M)
0%
(0 of 5.2M)
0.3%
(14K of 5.2M)
37.1%
(2.0M of 5.2M)
0%
(0 of 5.2M)
0%
(0 of 5.2M)
3.3%
(169K of 5.2M)
Large
Ground Water
11.9%
(837K of 7M)
0%
(0 of 7M)
1.3%
(94K of 7M)
11.9%
(837K of 7M)
0%
(0 of 7M)
0%
(0 of 7M)
0%
(0 of 7M)
Large
Mixed Water
47.2%
(859K of 1,8M)
0%
(Oof 1.8M)
7.4%
(135K of 1.8M)
42.3%
(770K of 1,8M)
0%
(Oof 1.8M)
0%
(Oof 1.8M)
1.1%
(20K of 1,8M)
Very Large
Surface Water
46.6%
(42M of 89M)
0%
(0 of 89M)
4.6%
(4.1 M of 89M)
44.3%
(40M of 89M)
0%
(0 of 89M
0%
(0 of 89M
2.6%
(2.3 M of 89M)
Very Large
Ground Water
36.02%
(6.7M of 19M)
5.7%
(1.1M of 19M)
3.6%
(669K of 19M)
32.7%
(6.1M of 19M)
0%
(Oof 18M)
0%
(Oof 18M)
0.8%
(148K of 19M)
Very Large
Mixed Water
60.6%
(21M of 34M)
1.8%
(626K of 34M)
18.2%
(6.2M of 34M)
46.5%
(16M of 34M)
0%
(0 of 34M)
0%
(0 of 34M)
14%
(0 of 34M)
All
Surface Water
46.1%
(44M of 95M)
0%
(0 of 95M)
4.3%
(4.1 M of 95 M)
43.8%
(42M of 95M)
0%
(0 of 95M)
0.005%
(4.5K of 95M)
2.6%
(2.5M of 95M)
All
Ground Water
29%
(7.5M of 26M)
4.07%
(1.1M of 26M)
3%
(769K of 26M)
26.5%
(6.9M of 26M)
0%
(0 of 26M)
0%
(0 of 26M)
0.6%
(148K of 26M)
All
Mixed Water
59.9%
(22 M of 36 M)
1.7%
(626K of 36M)
17.6%
(6.4M of 36M)
46.4%
(17M of 36M)
0%
(0 of 36M)
0%
(0 of 36 M)
13.3%
(4.8M of 36M)
Note: See Exhibit 5.1 for the MRLs for each nitrosamine.
1) Small = serving < 10,000; Large = serving 10,001 - 100,000; Very Large = serving > 100,000.
Six-Year Review 3
Technical Support Document for Nitrosamines
5-28
December 2016
-------
Exhibit 5.25: Percentage of Population Served by PWSs in the UCMR 2 Dataset Detecting Nitrosamines At Least Once,
by Disinfectant Type
Disinfectant
Nitrosamine
Group: % Pop.
Served by PWSs
with Detects
NDBA: % Pop.
Served by PWSs
with Detects
NDEA: % Pop.
Served by PWSs
with Detects
NDMA: % Pop.
Served by PWSs
with Detects
NDPA: % Pop.
Served by PWSs
with Detects
NMEA: % Pop.
Served by PWSs
with Detects
NPYR: % Pop.
Served by PWSs
with Detects
Any chloramine
80.9%
(53M of 65M)
0%
(0 of 65M)
6.3%
(4.1M of 65M)
79.1%
(52M of 65M)
0%
(0 of 65M)
0%
(0 of 65M)
4%
(2.6M of 65M)
Chlorine or
other
22.7%
(20M of 92M)
1.8%
(1.7M of 92M)
7.8%
(7.2 M of 92M)
15.0%
(14M of 92M)
0%
(0 of 92M)
0.005%
(4.5K of 92M)
5.3%
(4.9M of 92M)
No disinfection
2.7%
(8.7K of 318K)
0%
(0 of 317K)
0%
(0 of 317K)
2.7%
(8.7K of 317K)
0%
(0 of 317K)
0%
(0 of 317K)
0%
(0 of 317K)
All Disinfectant
Categories
46.4%
(73 M of 157 M)
1.1%
(1.7M of 157M)
7.1%
(11M of 157M)
41.5%
(65 M of 157 M)
0%
(0 of 157M)
0.003%
(4.5K of 157M)
4.7%
(7.4M of 157M)
Note: See Exhibit 5.1 for the MRLs for each nitrosamine.
Exhibit 5.26: Percentage of Population at Entry Points in the UCMR 2 Dataset Detecting Nitrosamines At Least Once, by
PWS Size and Source Water Type
PWS Size1
Source Water
Type
Nitrosamine
Group: % Pop.
Served at EPs
with Detects
NDBA: % Pop.
Served at EPs
with Detects
NDEA: % Pop.
Served at EPs
with Detects
NDMA: % Pop.
Served at EPs
with Detects
NDPA: % Pop.
Served at EPs
with Detects
NMEA: % Pop.
Served at EPs
with Detects
NPYR: % Pop.
Served at EPs
with Detects
Small
Surface Water
25.7%
(74K of 289K)
0%
(0 of 289K)
0%
(0 of 289K)
25.7%
(74K of 289K)
0%
(0 of 289K)
0%
(0 of 289K)
1.4%
(3.9K of 289K)
Small
Ground Water
2.7%
(9.3K of 340K)
0.7%
(2.5K of 340K)
0.2%
(822 of 340K)
1.6%
(5.9K of 340K)
0%
(0 of 340K)
0%
(0 of 340K)
0%
(0 of 340K)
Small
Mixed Water
41.8%
(20K of 49K)
0%
(0 of 49K)
0%
(0 of 49K)
41.4%
(20K of 49K)
0%
(0 of 49K)
0%
(0 of 49K)
0%
(0 of 49K)
Large
Surface Water
25.6%
(676K of 2.6M)
0%
(0 of 2.6M)
0%
(0 of 2.6M)
24.7%
(650K of 2.6M)
0%
(0 of 2.6M)
0%
(0 of 2.6M)
2.9%
(77K of 2.6M)
Large
Ground Water
3.6%
(168K of 4.7M)
0%
(0 of 4.7M)
0.1%
(4.5K of 4.7M)
3.5%
(163K of 4.7M)
0%
(0 of 4.7M)
0%
(0 of 4.7M)
0%
(0 of 4.7M)
Large
Mixed Water
17.4%
(209K of 1,2M)
0%
(0 of 1.2M)
1.3%
(16K of 1,2M)
16.03%
(193K of 1.2M)
0%
(Oof 1.2M)
0%
(0 of 1.2M)
0%
(0 of 1.2M)
Six-Year Review 3
Technical Support Document for Nitrosamines
5-29
December 2016
-------
PWS Size1
Source Water
Type
Nitrosamine
Group: % Pop.
Served at EPs
with Detects
NDBA: % Pop.
Served at EPs
with Detects
NDEA: % Pop.
Served at EPs
with Detects
NDMA: % Pop.
Served at EPs
with Detects
NDPA: % Pop.
Served at EPs
with Detects
NMEA: % Pop.
Served at EPs
with Detects
NPYR: % Pop.
Served at EPs
with Detects
Very Large
Surface Water
18.9%
(8.8M of 47M)
0%
(0 of 47M)
2.1%
(957K of 47M)
16.5%
(7.7M of 47M)
0%
(0 of 47M)
0%
(0 of 47M)
0.9%
(422K of 47M)
Very Large
Ground Water
3.4%
(401K of 12M)
0.04%
(4.4K of 12M)
0.7%
(78K of 12M)
2.6%
(320K of 12M)
0%
(0 of 12M)
0%
(Oof 12M)
0%
(Oof 12M)
Very Large
Mixed Water
12.6%
(2.7M of 22M)
0.03%
(5.6K of 22M)
1.8%
(388K of 22M)
10.6%
(2.3M of 22M)
0%
(0 of 22M)
0%
(0 of 22M)
1.5%
(324K of 22M)
All
Surface Water
19.3%
(9.6M of 50M)
0%
(0 of 50 M)
1.9%
(957K of 50M)
17%
(8.4M of 50M)
0%
(0 of 50M)
0%
(0 of 50M)
1.01%
(503K of 50M)
All
Ground Water
3.5%
(578K of 17M)
0.04%
(6.9K of 17M)
0.5%
(83K of 17M)
3%
(490K of 17M)
0%
(0 of 17M)
0%
(0 of 17M)
0%
(0 of 17M)
All
Mixed Water
12.9%
(2.9M of 23M)
0.02%
(5.6K of 23M)
1.8%
(404K of 23 M)
11%
(2.5M of 23M)
0%
(0 of 23M)
0%
(0 of 23M)
1.4%
(324K of 23M)
Note: See Exhibit 5.1 for the MRLs for each nitrosamine.
1) Small = serving < 10,000; Large = serving 10,001 - 100,000; Very Large = serving > 100,000.
Exhibit 5.27: Percentage of Population at Entry Points in the UCMR 2 Dataset Detecting Nitrosamines At Least Once, by
Disinfectant Type
Disinfectant
Nitrosamine
Group: % Pop.
Served at EPs
with Detects
NDBA: % Pop.
Served at EPs
with Detects
NDEA: % Pop.
Served at EPs
with Detects
NDMA: % Pop.
Served at EPs
with Detects
NDPA: % Pop.
Served at EPs
with Detects
NMEA: % Pop.
Served at EPs
with Detects
NPYR: % Pop.
Served at EPs
with Detects
Any chloramine
28.6%
(9.1 M of 32M)
0%
(0 of 32M)
0.5%
(146K of 32M)
27.4%
(8.7M of 32M)
0%
(0 of 32M)
0%
(0 of 32M)
0.8%
(268K of 32M)
Chlorine or other
7.04%
(4M of 56M)
0.02%
(13K of 56M)
2.3%
(1.3M of 56M)
4.7%
(2.7M of 56M)
0%
(0 of 56M)
0%
(0 of 56M)
1%
(560K of 56M)
No disinfection
2.9%
(28K of 966K)
0%
(0 of 966K)
1.4%
(14K of 966K)
1.5%
(14K of 966K)
0%
(0 of 966K)
0%
(0 of 966K)
0%
(0 of 966K)
All Disinfectant
Categories
14.7%
(13M of 89M)
0.01%
(13K of 89M)
1.6%
(1.4M of 89M)
12.8%
(11M of 89M)
0%
(0 of 89 M)
0%
(0 of 89 M)
0.9%
(827K of 89M)
Note: See Exhibit 5.1 for the MRLs for each nitrosamine.
Six-Year Review 3
Technical Support Document for Nitrosamines
5-30
December 2016
-------
Exhibit 5.28: Percentage of Population at Maximum Residence Time Locations in the UCMR 2 Dataset Detecting
Nitrosamines At Least Once, by Size and Source Water Type
PWS Size1
Source Water
Type
Nitrosamine
Group: % Pop.
Served at MRs
with Detects
NDBA: % Pop.
Served at MRs
with Detects
NDEA: % Pop.
Served at MRs
with Detects
NDMA: % Pop.
Served at MRs
with Detects
NDPA: % Pop.
Served at MRs
with Detects
NMEA: % Pop.
Served at MRs
with Detects
NPYR: % Pop.
Served at MRs
with Detects
Small
Surface Water
29.9%
(85K of 285K)
0%
(0 of 285K)
0%
(0 of 285K)
30%
(85K of 285K)
0%
(0 of 285K)
0.8%
(2.2K of 285K)
2.6%
(7.3K of 285K)
Small
Ground Water
2.4%
(5K of 209K)
0%
(0 of 209K)
0%
(0 of 209K)
2.4%
(5K of 209K)
0%
(0 of 209K)
0%
(0 of 209K)
0%
(0 of 209K)
Small
Mixed Water
43.9%
(17K of 38K)
0%
(0 of 38K)
0%
(0 of 38K)
43.9%
(17K of 38K)
0%
(0 of 38K)
0%
(0 of 38K)
0%
(0 of 38K)
Large
Surface Water
33.9%
(864K of 2.6M)
0%
(0 of 2.6M)
0.3%
(7K of 2.6M)
33.2%
(845K of 2.6M)
0%
(0 of 2.6M)
0%
(0 of 2.6M)
1.02%
(26K of 2.6M)
Large
Ground Water
7.5%
(178K of 2.4M)
0%
(0 of 2.4M)
0.09%
(2.2K of 2.4M)
7.4%
(176K of 2.4M)
0%
(0 of 2.4M)
0%
(0 of 2.4M)
0%
(0 of 2.4M)
Large
Mixed Water
35.5%
(219K of 617K)
0%
(0 of 617K)
0%
(0 of 617K)
34.6%
(214K of 617K)
0%
(0 of 617K)
0%
(0 of 617K)
0.8%
(5K of 617K)
Very Large
Surface Water
37.5%
(16M of 43M)
0%
(0 of 43M)
3.3%
(1.4M of 43M)
34.6%
(15M of 43M)
0%
(0 of 43M)
0%
(0 of 43M)
1.4%
(605K of 43M)
Very Large
Ground Water
19%
(1.3M of 6.8M)
0.03%
(2.2K of 6.8M)
1.06%
(72K of 6.8M)
17.2%
(1.2M of 6.8M)
0%
(0 of 6.8M)
0%
(0 of 6.8M)
0.7%
(49K of 6.8M)
Very Large
Mixed Water
28.8%
(3.7M of 13M)
0%
(0 of 13M)
0.5%
(59K of 13M)
27.8%
(3.6M of 13M)
0%
(Oof 13M)
0%
(0 of 13M)
0.7%
(90Kof 13M)
All
Surface Water
37.2%
(17M of 45M)
0%
(0 of 45M)
3.2%
(1.4M of 45M)
34.5%
(16M of 45M)
0%
(0 of 45 M)
0%
(0 of 45M)
1.4%
(638K of 45M)
All
Ground Water
15.7%
(1.5M of 9.4M)
0.02%
(2.2K of 9.4M)
0.8%
(75K of 9.4M)
14.4%
(1.4M of 9.4M)
0%
(0 of 9.4M)
0%
(0 of 9.4M)
0.5%
(49K of 9.4M)
All
Mixed Water
29.2%
(3.9M of 14M)
0%
(0 of 14M)
0.4%
(59K of 14M)
28.2%
(3.8M of 14M)
0%
(0 of 14M)
0%
(0 of 14M)
0.7%
(95K of 14M)
Note: See Exhibit 5.1 for the MRLs for each nitrosamine.
1) Small = serving < 10,000; Large = serving 10,001 - 100,000; Very Large = serving > 100,000.
Six-Year Review 3
Technical Support Document for Nitrosamines
5-31
December 2016
-------
Exhibit 5.29: Percentage of Population at Maximum Residence Time Locations in the UCMR 2 Dataset Detecting
Nitrosamines At Least Once, by Disinfectant Type
Disinfectant
Nitrosamine Group:
% Pop. Served at
MRs with Detects
NDBA: % Pop.
Served at MRs
with Detects
NDEA: % Pop.
Served at MRs
with Detects
NDMA: % Pop.
Served at MRs
with Detects
NDPA: % Pop.
Served at MRs
with Detects
NMEA: % Pop.
Served at MRs
with Detects
NPYR: % Pop.
Served at MRs
with Detects
Any chloramine
66.2%
(17M of 26M)
0%
(0 of 26M)
2.4%
(634K of 26M)
64.4%
(17M of 26M)
0%
(0 of 26M)
0%
(0 of 26M)
2.2%
(569K of 26M)
Chlorine or other
11.7%
(5M of 42M)
0.01%
(2.2K of 42M)
2.2%
(926K of 42M)
9.3%
(4M of 42M)
0%
(0 of 42M)
0.01%
(2.2K of 42M)
0.5%
(214K of 42M)
No disinfection
36.03%
(5.5K of 15K)
0%
(Oof 15K)
18.02%
(2.8K of 15K)
36.03%
(5.5K of 15K)
0%
(0 of 15K)
0%
(Oof 15K)
0%
(Oof 15K)
All Disinfectant
Categories
32.7%
(22M of 68M)
0.003%
(2.2K of 68M)
2.29%
(1.6M of 68M)
30.5%
(21M of 68M)
0%
(0 of 68M)
0.003%
(2.2K of 68M)
1.15%
(780K of 68M)
Note: See Exhibit 5.1 for the MRLs for each nitrosamine.
Six-Year Review 3
Technical Support Document for Nitrosamines
5-32
December 2016
-------
5.4 Nitrosamine Co-Occurrence and Aggregate Occurrence in UCMR 2 Sampling
Although NDMA occurs most frequently in finished water among six nitrosamines, the UCMR 2
data also indicate that other nitrosamines can co-occur with NDMA and with each other in
PWSs. This section provides an analysis of nitrosamine co-occurrence as well as aggregate
occurrence in UCMR 2 monitoring. In this analysis, "co-occurrence" means the detection of two
or more analytes (of the six in UCMR 2) at a single PWS during the course of monitoring, not
necessarily detection in the same sample. Co-occurrence results should be interpreted with
caution since MRLs vary from contaminant to contaminant. "Aggregate occurrence" is the
occurrence of any one or more members of the nitrosamine group.
Exhibit 5.30 is a Venn diagram that provides a visual overview of nitrosamine co-occurrence
within the UCMR 2 dataset. The diagram shows which nitrosamines co-occur and the number of
PWSs at which they occur and co-occur. Of the 343 PWSs with nitrosamine detections, 34 have
two or more co-occurring nitrosamines. NDMA co-occurred with every other nitrosamine
(except NDPA, which was not detected at all). Two PWSs detected three nitrosamines; in each
case, NDMA and NPYR were two of the three. There is only one instance of co-occurrence not
involving NDMA: one PWS detected NDEA and NPYR but not NDMA. Overall, the UCMR 2
data show that the nitrosamines have significant co-occurrence, mostly involving NDMA.
Exhibit 5.30: Co-Occurrence Venn Diagram
PWSs witn
detections of
NDPA 0
detect!oris of
NOMA =324
PWSs with
PWSs with
detections of
NDEA =26
PWSs with
detectionsof
NPYR = 21
PWSs with
detections of
N M E A ^ 3
PWS% with no drk-c^onsof any nitrosaivnr.'. 855
Six-Year Review 3 5-33
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 5.31 provides an overview of aggregate occurrence of the nitrosamines. A total of 28.6
percent of PWSs participating in the UCMR 2 Screening Survey, serving 46.4 percent of the
customers of participating PWSs, had detections of one or more nitrosamine. The higher
percentage for population than for PWSs reflects the generally higher rates of detection at larger
PWSs. Ground water PWSs generally had lower rates of detection than surface water and mixed
water PWSs. It is important to note that because the UCMR 2 survey represents a combination of
samples and census, summary statistics from the survey should not be interpreted as a simple
surrogate for national occurrence. Modeled national occurrence and exposure estimates based on
the UCMR 2 data for NDMA, taking into account appropriate weighting of these results with
respect to the national distribution of PWSs with various source water types, sizes and
disinfection practice characteristics, are presented in Section 5.5.
Exhibit 5.31: Aggregate Occurrence of Nitrosamines (UCMR 2 PWSs and
Population Exposed)
PWS Size1
Source Water
Type
UCMR 2 PWSs with At
Least One Detection of Any
Nitrosamine (Percent)
Population Served by
UCMR 2 PWSs with At
Least One Detection of
Any Nitrosamine (Percent)
Small
All Types
87 (18.1%)
306,211 (25.3%)
Small
Surface Water
61 (27.5%)
192,568 (33.6%)
Small
Ground Water
13 (5.4%)
46,926 (8.6%)
Small
Mixed Water
13 (72.2%)
66,717 (76.8%)
Large
All Types
75 (23.4%)
3,673,223 (26.2%)
Large
Surface Water
44 (36.7%)
1,978,092(38.1%)
Large
Ground Water
15 (9.2%)
836,633 (11.9%)
Large
Mixed Water
16 (43.2%)
858,499 (47.2%)
Very Large
All Types
181 (45.5%)
69,024,919 (48.6%)
Very Large
Surface Water
97 (44.5%)
41,614,960 (46.6%)
Very Large
Ground Water
22 (30.6%)
6,654,815(36.0%)
Very Large
Mixed Water
62 (57.4%)
20,755,144 (60.6%)
Total
343 (28.6%)
73,004,353 (46.4%)
Note: This table presents detection-based analysis of results from participating UCMR 2 systems. Nitrosamines detected under the
UCMR 2 are NDBA, NDEA, NDMA, NMEA and NPYR. NDPA was not detected by any PWS. See Exhibit 5.1 for the MRLs for each
nitrosamine.
1) Small = serving < 10,000; Large = serving 10,001 - 100,000; Very Large = serving > 100,000.
Exhibit 5.32 shows the significance of NDMA in nitrosamine co-occurrence in the UCMR 2
dataset and presents co-occurrence results by PWS size and source water type. The first two
columns show PWSs with detections of NDMA only (count and percentage). These PWSs are
the 291 depicted in the non-overlapping portion of the NDMA circle in the Venn diagram above.
The second set of columns shows PWSs with detections of other nitrosamines but not NDMA.
These are the PWSs that lie in the portions of the NDEA, NPYR and NDBA circles that are
outside the NDMA circle in the Venn diagram. The third set of columns shows PWSs that
detected NDMA and one or more other nitrosamine. These are the PWSs in the areas of overlap
between the NDMA circle and the other four circles in the Venn diagram. The final set of
columns shows the aggregate results. The table shows that occurrence rates are higher in surface
water and mixed water PWSs than in PWSs served only by ground water within each of the size
Six-Year Review 3 5-34
Technical Support Document for Nitrosamines
December 2016
-------
categories, and that for both surface water and ground water PWSs, occurrence rates increase
with increasing PWS size category.
Exhibit 5.32: Co-Occurrence of NDMA with Other Nitrosamines: UCMR 2 PWSs
Affected
PWS Size1
Source Water
Type
UCMR2 PWSs
with At Least
One Detection
of NDMA Only
(Percent)
UCMR 2 PWSs
with At Least
One Detection of
Other
Nitrosamines
Only (Percent)
UCMR 2 PWSs with
At Least One
Detection of NDMA
and Another
Nitrosamine
(Percent)
UCMR 2 PWSs
with At Least
One Detection
of Any
Nitrosamine
(Percent)
Small
All PWSs
78 (16.3%)
3 (0.6%)
6(1.3%)
87 (18.1%)
Small
Surface Water
55 (24.8%)
0 (0.0%)
6 (2.7%)
61 (27.5%)
Small
Ground Water
10 (4.2%)
3 (1.3%)
0 (0.0%)
13 (5.4%)
Small
Mixed Water
13 (72.2%)
0 (0.0%)
0 (0.0%)
13 (72.2%)
Large
All PWSs
66 (20.6%)
4(1.3%)
5(1.6%)
75 (23.4%)
Large
Surface Water
39 (32.5%)
2 (1.7%)
3 (2.5%)
44 (36.7%)
Large
Ground Water
14 (8.6%)
0 (0.0%)
1 (0.6%)
15 (9.2%)
Large
Mixed Water
13 (35.1%)
2 (5.4%)
1 (2.7%)
16 (43.2%)
Very Large
All PWSs
147 (36.9%)
12 (3.0%)
22 (5.5%)
181 (45.5%)
Very Large
Surface Water
85 (39.0%)
4 (1.8%)
8 (3.7%)
97 (44.5%)
Very Large
Ground Water
15 (20.8%)
4 (5.6%)
3 (4.2%)
22 (30.6%)
Very Large
Mixed Water
47 (43.5%)
4 (3.7%)
11 (10.2%)
62 (57.4%)
Total
291 (24.3%)
19 (1.6%)
33 (2.8%)
343 (28.6%)
Note: Represents detection-based analysis of results from participating UCMR 2 systems. "Other nitrosamines" detected under the
UCMR 2 are NDBA, NDEA, NMEA and NPYR. NDPA was not detected by any PWS. See Exhibit 5.1 for the MRLs for each
nitrosamine.
1) Small = serving < 10,000; Large = serving 10,001 - 100,000; Very Large = serving > 100,000.
Exhibit 5.33 presents corresponding information about the population exposed to NDMA only,
nitrosamines other than NDMA, and both NDMA and other nitrosamines, at PWSs participating
in the UCMR 2 Screening Survey. The information presented in this exhibit is consistent with
the trends observed in Exhibit 5.32. The proportion of the population served by PWSs with
detections is often higher than the proportion of PWSs with detections because the larger PWSs
(those serving larger populations) have more detections.
Exhibit 5.33: Co-Occurrence of NDMA with Other Nitrosamines: UCMR 2
Population Exposed
PWS Size1
Source Water
Type
Population
Served by
UCMR2 PWSs
with At Least
One Detection
of NDMA Only
(Percent)
Population
Served by UCMR
2 PWSs with At
Least One
Detection of
Other
Nitrosamines
Only (Percent)
Population Served
by UCMR 2 PWSs
with At Least One
Detection of NDMA
and Another
Nitrosamine
(Percent)
Population
Served by
UCMR 2 PWSs
with At Least
One Detection
of Any
Nitrosamine
(Percent)
Small
All PWSs
266,292
(22.0%)
20,867
(1.7%)
19,052
(1.6%)
306,211
(25.3%)
Small
Surface Water
173,516
(30.2%)
0
(0.0%)
19,052
(3.3%)
192,568
(33.6%)
Six-Year Review 3 5-35
Technical Support Document for Nitrosamines
December 2016
-------
PWS Size1
Source Water
Type
Population
Served by
UCMR2 PWSs
with At Least
One Detection
of NDMA Only
(Percent)
Population
Served by UCMR
2 PWSs with At
Least One
Detection of
Other
Nitrosamines
Only (Percent)
Population Served
by UCMR 2 PWSs
with At Least One
Detection of NDMA
and Another
Nitrosamine
(Percent)
Population
Served by
UCMR 2 PWSs
with At Least
One Detection
of Any
Nitrosamine
(Percent)
Small
Ground Water
26,059
(4.8%)
20,867
(3.8%)
0
(0.0%)
46,926
(8.6%)
Small
Mixed Water
66,717
(76.8%)
0
(0.0%)
0
(0.0%)
66,717
(76.8%)
Large
All PWSs
3,256,392
(23.2%)
141,093
(1.0%)
275,738
(2.0%)
3,673,223
(26.2%)
Large
Surface Water
1,809,584
(34.9%)
52,137
(1.0%)
116,371
(2.2%)
1,978,092
(38.1%)
Large
Ground Water
742,973
(10.6%)
0
(0.0%)
93,660
(1.3%)
836,633
(11.9%)
Large
Mixed Water
703,836
(38.7%)
88,956
(4.9%)
65,707
(3.6%)
858,499
(47.2%)
Very Large
All PWSs
53,138,701
(37.4%)
7,530,629
(5.3%)
8,355,589
(5.9%)
69,024,919
(48.6%)
Very Large
Surface Water
35,243,460
(39.5%)
2,109,340
(2.4%)
4,262,160
(4.8%)
41,614,960
(46.6%)
Very Large
Ground Water
4,792,295
(25.9%)
609,237
(3.3%)
1,253,283
(6.8%)
6,654,815
(36.0%)
Very Large
Mixed Water
13,102,946
(38.2%)
4,812,052
(14.0%)
2,840,146
(8.3%)
20,755,144
(60.6%)
Total
56,661,385
(36.0%)
7,692,589
(4.9%)
8,650,379
(5.5%)
73,004,353
(46.4%)
Note: Represents detection-based analysis of results from participating UCMR 2 systems. "Other nitrosamines" detected under the
UCMR 2 are NDBA, NDEA, NMEA and NPYR. NDPA was not detected by any PWS. See Exhibit 5.1 for the MRLs for each
nitrosamine.
1) Small = serving < 10,000; Large = serving 10,001 - 100,000; Very Large = serving > 100,000.
As shown in Exhibit 5.30 above, only two PWSs detected more than two nitrosamines during the
UCMR 2 Screening Survey. One was a large surface water PWS and the other was a very large
mixed water PWS. Together these two PWSs serve 182,540 customers.
EPA also evaluated the UCMR 2 dataset to determine the occurrence of multiple nitrosamines in
the same sample. Results show that 18 PWSs (1.5 percent) had at least one sample with exactly 2
nitrosamines present, and 1 PWS (0.1 percent) had at least one sample with exactly 3
nitrosamines present. No PWSs had any samples with more than three nitrosamines present.
Exhibit 5.34 provides some additional summary statistics about nitrosamine co-occurrence in the
UCMR 2 dataset. The average frequency of detections of any nitrosamine monitored under
UCMR 2 is 5.7 detections per PWS (among those PWSs with detections). NDMA represents the
majority (95 percent; 1,841 of 1,940) of detections. There is significant co-occurrence between
the nitrosamines. In most cases of co-occurrence at PWSs, NDMA is one of the co-occurring
nitrosamines. The third column in Exhibit 5.34 shows that 63.5 percent (33 of 52) of the PWSs
that detected any non-NDMA nitrosamine also detected NDMA. The last column shows that 9.9
percent (34 of 343) of PWSs that detected any nitrosamine had co-occurrence of two or more
nitrosamines.
Six-Year Review 3 5-36
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 5.34: UCMR 2 Co-Occurrence Summary Statistics
Chemical
Average Number of
Detections per PWS1
Co-Occurrence with
NDMA at the PWS Level2
Co-Occurrence with Any
Other Nitrosamine at the
PWS Level3
NDBA
1.8 (9/5)
3/5 (60%)
3/5 (60%)
NDEA
1.8 (46/26)
17/26 (65.4%)
18/26 (69.2%)
NDMA
5.7 (1,841/324)
NA
33/324 (10.2%)
NDPA
0
0
0
NMEA
1 (3/3)
3/3 (100%)
3/3 (100%)
NPYR
2 (41/21)
12/21 (57.1%)
13/21 (61.9%)
Any nitrosamine
5.7 (1,940/343)
33/52 (63.5%)
34/343 (9.9%)
Note:
1) The average number of detections per PWS at PWSs with at least one detection. (The number of detections and the
number of PWSs are shown in parenthesis.)
2) The number and percent of PWSs with detections that also had NDMA present at any time during the monitoring period.
3) The number and percent of PWSs with detections that had another nitrosamine present at any time during the monitoring
period.
5.5 Modeling of NDMA Occurrence
As noted earlier, of the six nitrosamines NDMA is the only one with a sufficient number of
positive samples from the UCMR 2 data to support detailed extrapolations of occurrence to the
national level. Since the MRL of NDMA (2 ng/L) is substantially lower than the HRL (0.6 ng/L),
EPA used modeling methodologies, described below, to develop a more fully rounded picture of
NDMA occurrence and estimate national NDMA occurrence and exposure at the HRL along
with other thresholds.
Section 5.5.1 summarizes the modeling approach used to develop national occurrence and
exposure estimates for NDMA. Sections 5.5.2, 5.5.3 and 5.5.3.3 present the modeled national
occurrence and exposure estimates for NDMA, on the basis of detection at a PWS or sampling
point, mean concentration at a PWS or sampling point, and locational annual average (LAA) at a
sampling point, respectively. Since NDMA is a carcinogen, the mean concentration at a PWS is
an appropriate benchmark for evaluating health outcomes based on chronic exposure. EPA
modeled the LAA as well as the mean to see how well a hypothetical monitoring regime, similar
to that used under the Stage 2 D/DBPRs, would change the estimated number of systems and
population with exceedance of a threshold.
5.5.1 Overview of the Modeling Approach for Generating National Occurrence
The modeling approach developed to estimate national occurrence for NDMA was designed to
accomplish a number of objectives. The key objectives and how the modeling approach
addresses them are as follows:
• Estimation of occurrence and exposure for NDMA above any specified concentration
(threshold) of interest. The MRL of 2 ng/L for NDMA means that evaluations of
occurrence based on the UCMR 2 dataset are limited to observable occurrence at or
above that concentration. Describing occurrence and exposure above thresholds of health
interest—most notably, the HRL of 0.6 ng/L—requires modeling that uses probability
Six-Year Review 3 5-37
Technical Support Document for Nitrosamines
December 2016
-------
distributions to characterize estimated occurrence along a continuum of below-MRL
concentrations.
• Estimation of occurrence and exposure for NDMA both on the basis of individual
detections exceeding thresholds of interest (based on maximum concentrations,
approximating acute exposure) and on the basis of long-term average concentrations
exceeding thresholds of interest (based on mean concentrations, approximating chronic
exposure). Detection at concentrations above a given threshold indicates occurrence and
exposure, while a long-term average above a given threshold provides a more meaningful
indication of the public health risk at PWSs in the case of nitrosamines, because the
health end-point, the basis of the HRL, is carcinogenicity. Modeling based on continuous
occurrence probability distributions can produce estimates of both of those occurrence
measures. To produce these estimates, the expected contaminant concentration is first
computed for each sample given what is known about the sampling point where the
sample was collected, including the treatment used (if any) at the sampling point at that
time. The results of these sample level computations are then appropriately aggregated to
the sampling point level and then to the PWS level. In the presentation of results below,
the two approaches are labeled as "detection-based modeling" and "mean-based
modeling."
• Estimation of occurrence and exposure for NDMA in a manner that properly reflects the
expected influences of source water type, PWS size, PWS type, sample location (i.e., EP
or MR) and disinfection practices. The UCMR 2 dataset indicates that NDMA
occurrence at PWSs is influenced by source, size, PWS type, sample location, and
disinfection type. To capture these influences in the modeling, EPA estimated occurrence
probability distributions separately for EP locations and MR locations and parameterized
the distributions using data that were stratified to reflect PWS source and size
characteristics. EPA incorporated "fixed effects" into the model to reflect the influence of
different PWS types, types of source water (separating strictly ground water from strictly
surface water and strictly surface water from mixtures of ground and surface water) and
disinfection practices.
• Characterization of the uncertainty in the estimation of occurrence and exposure for
NDMA. Recognizing that the UCMR 2 dataset has inherent limitations with respect to the
proportion of PWSs sampled in the "small" and "large" size categories, the short length
of the sampling period (one year), the limited number of samples taken during that year
(generally two at each sampling location for ground water systems and four for surface
water systems) and others, any national occurrence and exposure estimates derived from
those data will necessarily have some uncertainty associated with them. The Bayesian
modeling approach, used to obtain the parameters for the occurrence probability
distributions, is intended to explicitly reflect the uncertainty in those parameter estimates,
and that uncertainty is carried forward in applying those occurrence distributions to
making the national estimates. Specifically, estimates produced by the Bayesian
modeling approach are presented as a range, the 90 percent credible interval, defined as
the range into which there is a 90 percent chance that the actual value falls. There is a 5
percent chance that the actual value is above the highest value in the range, and a 5
percent chance that the actual value is below the lowest value in the range.
Six-Year Review 3 5-38
Technical Support Document for Nitrosamines
December 2016
-------
The following two subsections present a summary of the key methods and assumptions and
describe how these results were used to generate the national estimates reflecting the various
source water types, sizes, and disinfection practices at PWSs, EPs and MRs.
5.5.1.1 Generating the Parametric Occurrence Probability Distributions
EPA used a Bayesian methodology to obtain NDMA occurrence parameter estimates. This
approach for estimating occurrence parameters has been used previously to support drinking
water contaminant occurrence and exposure efforts, including Cryptosporidium occurrence
analysis for the Long-Term 2 Enhanced Surface Water Treatment Rule, occurrence analyses in
support of the first Six-Year Review (SYR) of National Primary Drinking Water Regulations
(NPDWRs) (USEPA, 2003) and an analysis of perchlorate occurrence using data from UCMR 1
(USEPA, 2010). The Bayesian modeling conducted for NDMA was based largely on the
approach used in the peer-reviewed analysis of perchlorate; however, a number of modifications
to the specifics of that approach were necessary to accommodate features of the UCMR 2 dataset
specific to NDMA.
The occurrence distributions for NDMA were estimated at the EP and MR sampling point levels.
An assumption was made that the log of NDMA concentration is normally distributed at each EP
or MR sampling point; that is, observations for multiple samples taken at a given sample point
over time would reflect a lognormal distribution of values. A lognormal probability distribution
is specified by a parameter mu reflecting the central tendency (i.e., mean of log concentration,
also known as logmean) and a parameter sigma reflecting dispersion (i.e., standard deviation of
log concentration) around that central tendency.
The central tendency parameters were estimated to reflect the following observable differences
among sampling points:
• Sample point type (EP or MR),
• Size of the PWS in which the sample point is located (i.e., small, large or very large),
• Source water type at that sample point when the sample was collected,
• Disinfectant used at that sample point when the sample was collected, and
• PWS type (i.e., CWS or NTNCWS)
Four types of variance parameters were built into the model (and because EPs were estimated
separately from MRs, those four types of variance parameters were estimated separately for EPs
and for MRs). The first variance parameter reflects the temporal variability among sample
concentrations within a sample point. The other three variance parameters reflect heterogeneity
at the sampling point, PWS and strata levels. Specifically, these address the variability between
sampling point logmeans within a particular PWS, the variability between PWS logmeans within
a stratum and the variability between strata logmeans within the nation.
Each PWS was assigned to one of nine strata. These included the three major PWS size
categories (very large, large and small) that were used in the UCMR 2 sampling scheme for
Six-Year Review 3 5-39
Technical Support Document for Nitrosamines
December 2016
-------
nitrosamines as discussed in Section 5.2.1 and the three PWS source water categories of surface,
ground and mixed water as described in Section 5.2.2.
Fixed effects were incorporated into the model to adjust the central tendency results for the
sampling locations in the nine strata to account for the following influences:
• Two PWS types (CWS and NTNCWS);
• Six disinfection categories at the sample point (CA only, CA with CL/OT, CL Only, CL
with OT, OT Only or ND Only); and
• For the stratum of PWSs designated as mixed source water locations, a fixed effect was
incorporated at the sampling point level to reflect whether that specific location was a
ground water, surface water or mixed water location.
Note that for NDMA occurrence modeling results presented in this document, the PWS type was
set to CWS only. EPA decided to focus exclusively on CWSs because there were too few
NTNCWSs in the UCMR 2 dataset for an NTNCWS-focused analysis. In addition, because
NTNCWSs do not have distribution systems like CWSs do, it was determined that NTNCWS
results could not simply be extrapolated from CWS data.
The Bayesian estimation process begins with a description of the uncertainty on each
parameter's value, prior to observing the UCMR 2 data, with a statistical distribution. In this
case, the prior distributions used are "flat" (i.e., uninformative, capturing the complete
uncertainty that is implicitly assumed in classical statistical methods). The estimation process
then updates those priors with the information learned about the parameters from the UCMR 2
data to produce posterior distributions of uncertainty over the parameters' values. The technical
implementation of this estimation process uses Markov Chain Monte Carlo, which generates a
sequence of random draws that converge to the joint posterior distribution. Each drawn set of
parameters, or "realization," represents a consistent estimate, with the central clustering of those
realizations representing the best estimate and dispersion around that central cluster representing
the uncertainty in that best estimate.
5.5.1.2 Generating the National Occurrence Estimates from the Probability Distributions
For generating the NDMA national occurrence and exposure estimates, the Bayesian process
generated over 500 realizations (exactly 534 realizations) of occurrence probability distributions
for each of the 4,666 EP and 2,397 MR locations. Each of those 534 distributions at a given
location reflects central tendency and temporal variability among samples taken at those
locations that is consistent with the observations in UCMR 2. The differences among those 534
realizations for each location reflect the uncertainty in the estimate of the central tendency and
temporal variability parameter estimates.
To address concerns that 534 realizations of the occurrence probability distribution might not be
enough, the model was later run again with 1,000 realizations, as discussed below in the
subsection on "Model Validation." The results (central tendencies and 90 percent credible
intervals) from the second effort were substantially the same as the initial results presented in
this report.
Six-Year Review 3 5-40
Technical Support Document for Nitrosamines
December 2016
-------
Although the UCMR 2 dataset includes all of the very large PWSs serving more than 100,000
people, it included only a sample of the PWSs serving 100,000 or fewer people. It was, therefore,
necessary to "scale up" the UCMR 2 results from that sample to the national level to obtain the
overall national occurrence and exposure estimates.
To accomplish this, the following basic assumptions were used:
• To maintain consistency with the UCMR 2 census of very large PWSs, EPA used the
2005 SDWIS inventory as the basis for the total number of PWSs and population served
by those PWSs for the small and large PWSs using surface water or ground water. (The
2005 SDWIS data served as the basis for identifying and selecting PWSs for UCMR 2).
• Because SDWIS classifies mixed PWSs as surface water PWSs, EPA used information
from the 2006 Community Water Systems Survey (CWSS) to estimate the number of
PWSs nationally in the small and large size groups that are purely surface water and the
number that are mixed water PWSs.
• Although UCMR 2 required PWSs to report information on the type of disinfection used
at the sampling locations when the NDMA samples were taken, the selection of the
sample of 800 PWSs for the small and large size PWSs was not specifically designed to
be nationally representative of those disinfection practices among these PWSs. In the
absence of information to the contrary, EPA assumed that the proportion of UCMR 2
PWSs and sampling locations reporting the various types of disinfection was a reasonable
approximation of the proportion using those disinfection methods nationally. A
comparison with data from the 2006 CWSS for the two major types of practices—
chlorination and chloramination—showed similar proportions.
• EPA also assumed that the numbers of EPs and MRs per PWS for the small and large
PWSs in the UCMR 2 dataset were representative of those fractions nationally, and again
a comparison with the 2006 CWSS data indicated that assumption to be reasonable.
• Lastly, because no data are available on the proximity of sampling locations to
customers' taps, it was assumed, for this analysis, that each sampling location was
equally representative of water delivered to customers. Customers were assigned in equal
measure to each sampling location. For example, if a PWS had two EPs and one MR,
one-third of its population was assigned to each EP and one-third to the MR.
Using the occurrence distributions at EP and MR locations from the modeling described above,
EPA obtained occurrence and exposure estimates for the census of very large PWSs in UCMR 2
and combined these with the estimates for the small and large UCMR 2 PWSs that were "scaled
up" to the national level using the above assumptions. The estimates for the various categories of
PWSs in the various strata were combined to obtain the overall estimates of national occurrence
and exposure.
Three forms of national occurrence and exposure estimates were generated and are presented in
the sections that follow:
• One or more detections at PWSs, EPs and MRs exceeding various thresholds;
Six-Year Review 3 5-41
Technical Support Document for Nitrosamines
December 2016
-------
• PWS, EP and MR mean concentrations exceeding various thresholds; and
• PWS LAAs exceeding various thresholds
The national estimates of occurrence and exposure for the above measures are presented as the
percentage of PWSs and associated populations exceeding each of three "threshold"
concentrations of interest:
• 0.6 ng/L (the HRL and 10"6 cancer risk value),
• 2 ng/L (the MRL), and
• 6 ng/L (the 10"5 cancer risk value)
In the tables presenting the results, the column labeled "mean" reflects the central estimate
within the full range of uncertainty reflected by the 534 realizations of occurrence distributions
as described above, and these are accompanied by the lower and upper 90 percent credible
intervals for those estimates, capturing the uncertainty around the mean or central estimates.
Note that the percentages exceeding the 2 ng/L MRL. in the national estimates shown below will
differ from those shown previously for the UCMR 2 dataset alone. For example, Exhibit 5.19
indicates that 27 percent of PWSs in the UCMR 2 Screening Survey detected NDMA in one or
more samples. However, in Exhibit 5.39, based on the modeling and the "scaling up" of UCMR
2 results to the national level, only 10.3 percent of all PWSs nationally are expected to have
NDMA present in one or more samples above the MRL value of 2 ng/L (assuming that the
number of samples taken per system is similar to that taken in UCMR 2). This divergence is not
unexpected. It reflects the fact that UCMR 2 included a census of larger systems, which had
higher occurrence rates, and only a statistical sample of smaller systems, which had lower
occurrence rates. Because there are many more small systems nationally than were included in
the sample, the extrapolated national occurrence of NDMA would be expected to be much lower
than the simple aggregate of UCMR 2 observations.
5.5.1.3 Model Validation
Model validation efforts included the following:
• Using the model to reproduce results from the analysis of UCMR 2 PWSs with one or
more detections presented above (Section 5.3.3). A comparison of the output of the
model with UCMR 2 observed findings is presented in Appendix B. The comparison of
the modeled results and the observed results for one or more detections above various
thresholds showed a high degree of concordance in the aggregate and for specific subsets
of systems, based on size, source and disinfection practice.
• Running the model again, with 1,000 iterations rather than 534. The output values (mean
and 90 percent credible interval values for national occurrence and exposure) for the
larger number of iterations did not differ significantly from those obtained with fewer
iterations.
Six-Year Review 3 5-42
Technical Support Document for Nitrosamines
December 2016
-------
5.5.2 National Occurrence and Exposure Estimates: One or More Detections
As described in the previous section, EPA used modeling to estimate the percentage of sampling
locations and PWSs with one or more detections of NDMA exceeding a given threshold value,
along with the associated population served by those locations and PWSs. Results are presented
first for EP and MR locations, then for PWSs as a whole.
5.5.2.1 Entry Points and Maximum Residence Locations
Exhibit 5.35 and Exhibit 5.36 summarize results of the detection-based modeling analysis for EP
and MR locations, respectively. They show the mean expected values and the 90 percent credible
intervals for 3 size categories and for 3 threshold values for NDMA: 0.6 ng/L (the HRL), 2 ng/L
(the MRL) and 6 ng/L. See Appendix C for results for other thresholds and for predictions of the
total population exposed (extrapolated from the UCMR 2 population). The analysis based on EPs
gives lower bound estimates of exposure based on NDMA concentrations in drinking water
leaving the treatment plant. The analysis based on MRs gives estimates of higher exposure, as
the highest concentrations of NDMA would be expected at the MR time location, since NDMA
forms over time.
Exhibit 5.35: Percentage of Entry Points Predicted To Have One or More
Detections Exceeding the Threshold and the Associated Population Exposed
Total
National
Total
National
Threshold
(ng/L)
Percentage of EPs
Predicted To Have Any
Percentage of Population
Served by EPs Predicted
PWS Size1
Inventory:
PWS
Count
Inventory:
Population
Served
Detections Exceeding
Threshold: Expected
Value (90% CI)
To Have Any Detections
Exceeding Threshold:
Expected Value (90% CI)
Small
41,962
41,154,840
0.6
11.0%
(7.0%-16.1%)
16.5%
(10.8%-23.4%)
Small
41,962
41,154,840
2
4.0%
(2.7% - 5.6%)
6.8%
(4.5% - 9.9%)
Small
41,962
41,154,840
6
1.5%
(1.1%-2.1%)
2.5%
(1.7%-3.8%)
Large
2,812
84,318,927
0.6
20.6%
(16.2%-25.9%)
25.0%
(19.3%-31.9%)
Large
2,812
84,318,927
2
9.0%
(7.3%-11.0%)
11.5%
(9.2%-14.6%)
Large
2,812
84,318,927
6
3.1%
(2.3% - 4.0%)
4.6%
(3.5% - 5.8%)
Very Large
394
141,407,211
0.6
21.0%
(17.6%-24.9%)
31.4%
(25.0%-39.5%)
Very Large
394
141,407,211
2
7.7%
(6.7% - 8.9%)
13.5%
(11.2%-16.5%)
Very Large
394
141,407,211
6
2.6%
(2.2% - 3.0%)
4.9%
(4.1%-5.9%)
All
45,168
266,880,978
0.6
12.7%
(8.6%-17.7%)
26.8%
(20.8%-34.3%)
All
45,168
266,880,978
2
4.8%
(3.5% - 6.5%)
11.7%
(9.5%-14.8%)
All
45,168
266,880,978
6
1.8%
(1.3%-2.4%)
4.4%
(3.5% - 5.5%)
Source: Appendix C.
Abbreviation: CI = credible interval; EP = entry point
Note: Result of detection-based modeling of NDMA occurrence. Assumes that the number of samples taken per PWS is comparable
to the number taken for UCMR 2 sampling.
1) Small = serving < 10,000; Large = serving 10,001 - 100,000; Very Large = serving > 100,000.
Six-Year Review 3 5-43
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 5.36: Percentage of Maximum Residence Time Locations Predicted To
Have One or More Detections Exceeding the Threshold and the Associated
Population Exposed
Total
National
Total
National
Threshold
(ng/L)
Percentage of MRs
Predicted To Have Any
Percentage of Population
Served by MRs Predicted
PWS Size1
Inventory:
PWS
Count
Inventory:
Population
Served
Detections Exceeding
Threshold: Expected
Value (90% CI)
To Have Any Detections
Exceeding Threshold:
Expected Value (90% CI)
Small
41,962
41,154,840
0.6
20.7%
(13.8%-28.5%)
27.9%
(19.7%-36.8%)
Small
41,962
41,154,840
2
6.4%
(4.5% - 9.0%)
11.0%
(8.1%-15.1%)
Small
41,962
41,154,840
6
2.1%
(1.5%-2.9%)
4.3%
(3.2% - 5.8%)
Large
2,812
84,318,927
0.6
40.9%
(33.4% - 49.0%)
45.3%
(36.7%-54.8%)
Large
2,812
84,318,927
2
18.9%
(15.7%-22.8%)
22.8%
(19.2%-27.5%)
Large
2,812
84,318,927
6
7.3%
(5.8% - 8.9%)
10.3%
(8.3%-12.6%)
Very Large
394
141,407,211
0.6
42.9%
(37.6% - 48.8%)
55.6%
(47.6% - 64.4%)
Very Large
394
141,407,211
2
21.8%
(19.9%-24.2%)
31.9%
(28.3%-36.7%)
Very Large
394
141,407,211
6
9.2%
(8.2%-10.2%)
13.0%
(11.1%-15.2%)
All
45,168
266,880,978
0.6
23.4%
(16.5%-31.2%)
48.7%
(40.6% - 57.8%)
All
45,168
266,880,978
2
8.1%
(6.1%-10.8%)
26.3%
(22.9%-31.0%)
All
45,168
266,880,978
6
2.9%
(2.1%-3.7%)
11.0%
(9.2%-13.1%)
Source: Appendix C.
Abbreviation: CI = credible interval; MR = maximum residence time location
Note: This exhibit shows the results of detection-based modeling of NDMA occurrence. It is assumed that the number of samples
taken per PWS is comparable to the number taken for UCMR 2 sampling.
1) Small = serving < 10,000; Large = serving 10,001 - 100,000; Very Large = serving > 100,000.
Exhibit 5.35 shows that between 8.6 and 17.7 percent of EPs are predicted to exceed the HRL
(0.6 ng/L) at least once. Results are slightly higher for very large PWSs compared to large and
small PWSs, although the difference is small compared to the confidence bounds. The associated
percentage of the population served by EP locations predicted to exceed the HRL at least once is
between 21.8 and 34.3 percent. The percentage of population served is higher than the
percentage of EPs because UCMR 2 detections are somewhat higher in the larger PWSs (with
more customers served) within each size category.
As expected for a disinfection byproduct, the percentage of locations with one or more detections
is greater for MR than for EP locations. As shown in Exhibit 5.36, between 16.5 and 31.2 percent
of MR locations are expected to exceed the HRL (0.6 ng/L) at least once. Very large PWSs
exhibit higher predicted detection rates than the large or small PWSs, which is expected due to
their larger distribution systems and thus, potentially higher MR times.
Exhibit 5.37 and Exhibit 5.38 compare the predicted percentage of population exposed to at least
one sample with an NDMA concentration greater than the HRL at EPs and MRs, respectively,
for each combination of source water and disinfectant type. As is consistent with the literature
Six-Year Review 3 5-44
Technical Support Document for Nitrosamines
December 2016
-------
and proposed formation mechanisms discussed in Chapter 6, predictions are greater for
chloramines than for other disinfectant types. For EP sites, the percentage of the population
exposed is approximately three times greater for PWSs using chloramines than other
disinfectants. The percentage of population exposed is also generally greater for surface water
and mixed water PWSs than for ground water PWSs.
Exhibit 5.37: Percentage of the Population Predicted To Be Exposed to One or
More Detections at Levels Above the HRL (0.6 ng/L) at Entry Points
Mixed
CA
CA + CL/OT
CL
CL+OT
OT
ND
¦ GW
49.4%
22.9%
6.2%
3.5%
13.3%
2.9%
¦ SW
70.9%
46.6%
19.1%
9.6%
28.0%
19.7%
Mixed
63.9%
45.0%
11.5%
34.4%
29.9%
0.0%
Source: Appendix C.
Abbreviations: HRL = health reference level; CA = chloramines; CL = chlorine; ND = no disinfectant; OT = other.
Note: This exhibit shows the results of detection-based modeling of NDMA occurrence.
Six-Year Review 3 5-45
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 5.38: Percentage of the Population Predicted To Be Exposed to One or
More Detections at Levels Above the HRL (0.6 ng/L) at Maximum Residence Time
Locations
Mixed
ND
CA
CA + CL/OT
CL
CL+OT
OT
ND
¦ GW
78.4%
46.6%
12.5%
11.9%
26.2%
11.1%
¦ SW
93.4%
84.3%
36.4%
28.9%
57.0%
42.2%
Mixed
87.1%
85.0%
35.8%
52.9%
99.7%
63.0%
Source: Appendix C.
Abbreviations: HRL = health reference level; CA = chloramines; CL = chlorine; ND = no disinfectant; OT = other.
Note: This exhibit shows the results of detection-based modeling of NDMA occurrence.
5.5.2.2 PWSs
EPA also used modeling to predict the percentage of PWSs with at least one NDMA detection at
levels above various thresholds. The analysis aggregates results from all sampling locations
within the PWS—in other words, if one sampling location has a detection but the rest do not, the
PWS is still counted as having one or more NDMA detections. To determine the associated
population exposed, however, EPA used only the population served by sampling locations within
the PWS with one or more detections. As noted earlier in this chapter, specific information on
population served per sampling location was not available from the UCMR 2 dataset; thus, EPA
assumed that each sampling location serves an equal proportion of the population within the
PWS.
Exhibit 5.39 displays the results by size category for this analysis. It shows the mean expected
values and the 90 percent credible interval for three size categories and for three threshold
values: 0.6 ng/L (the HRL), 2 ng/L (the MEL) and 6 ng/L. Exhibit 5.40 shows predicted total
population exposed at those thresholds. See Appendix D for results at other thresholds.
Six-Year Review 3 5-46
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 5.39: PWSs Predicted To Have One or More Detections Exceeding Various
Thresholds
PWS Size1
Total
National
Inventory:
PWS
Count
Total
National
Inventory:
Population
Served
Threshold
(ng/L)
PWSs Predicted To Have
Any Detections
Exceeding Threshold:
Expected Value (90% CI)
Percentage of PWSs
Predicted To Have Any
Detections Exceeding
Threshold: Expected
Value (90% CI)
Small
41,962
41,154,840
0.6
11,309
(8,426- 14,366)
27.0%
(20.1%-34.2%)
Small
41,962
41,154,840
2
3,673
(2,687-4,811)
8.8%
(6.4%-11.5%)
Small
41,962
41,154,840
6
1,276
(959- 1,696)
3.0%
(2.3%-4.0%)
Large
2,812
84,318,927
0.6
1,616
(1,347- 1,876)
57.5%
(47.9%-66.7%)
Large
2,812
84,318,927
2
790
(662 - 950)
28.1%
(23.5%-33.8%)
Large
2,812
84,318,927
6
353
(295-419)
12.6%
(10.5%-14.9%)
Very Large
394
141,407,211
0.6
296
(268 - 324)
74.5%
(67.4%-81.3%)
Very Large
394
141,407,211
2
182
(164-203)
45.7%
(41.3%-50.9%)
Very Large
394
141,407,211
6
89
(80-99)
22.4%
(20.1%-24.9%)
All
45,168
266,880,978
0.6
13,221
(10,041 - 16,565)
29.3%
(22.2% - 36.7%)
All
45,168
266,880,978
2
4,644
(3,513-5,963)
10.3%
(7.8%-13.2%)
All
45,168
266,880,978
6
1,719
(1,334-2,214)
3.8%
(3.0%-4.9%)
Source: Appendix D.
Abbreviation: CI = credible interval
Note: This exhibit shows the results of detection-based modeling of NDMA occurrence. It is assumed that the number of samples
taken per PWS is comparable to the number taken for UCMR 2 sampling.
1) Small = serving < 10,000; Large = serving 10,001 - 100,000; Very Large = serving > 100,000.
Six-Year Review 3 5-47
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 5.40: Populations Served by PWSs Predicted To Have One or More
Detections Exceeding Various Thresholds
PWS Size1
Total
National
Inventory:
PWS
Count
Total
National
Inventory:
Population
Served
Threshold
(ng/L)
Population Served by PWSs
Predicted To Have Any
Detections Exceeding
Threshold: Expected Value
(90% CI)
Percentage of Population
Served by PWSs
Predicted To Have Any
Detections Exceeding
Threshold: Expected
Value (90% CI)
Small
41,962
41,154,840
0.6
15,690,723
(12,017,905 - 19,464,953)
38.1%
(29.2% - 47.3%)
Small
41,962
41,154,840
2
6,728,174
(4,731,066-8,860,238)
16.4%
(11.5%-21.5%)
Small
41,962
41,154,840
6
2,808,022
(2,121,357-3,881,905)
6.8%
(5.2% - 9.4%)
Large
2,812
84,318,927
0.6
51,521,139
(42,815,800 -59,615,539)
61.1%
(50.8%-70.7%)
Large
2,812
84,318,927
2
26,391,618
(22,285,126-31,541,136)
31.3%
(26.4% - 37.4%)
Large
2,812
84,318,927
6
12,399,960
(10,343,152 - 14,790,641)
14.7%
(12.3%-17.5%)
Very Large
394
141,407,211
0.6
110,803,964
(95,110,205- 123,563,993)
78.0%
(67.0%-87.0%)
Very Large
394
141,407,211
2
68,223,812
(60,037,674 -80,915,126)
48.1%
(42.3%-57.0%)
Very Large
394
141,407,211
6
33,492,275
(28,213,900 -39,584,290)
23.6%
(19.9%-27.9%)
All
45,168
266,880,978
0.6
178,015,826
(149,943,910-202,644,485)
66.6%
(56.1%-75.8%)
All
45,168
266,880,978
2
101,343,604
(87,053,866- 121,316,500)
37.9%
(32.6% - 45.4%)
All
45,168
266,880,978
6
48,700,256
(40,678,409 -58,256,836)
18.2%
(15.2%-21.8%)
Source: Appendix D.
Abbreviation: CI = credible interval
Note: This exhibit shows the results of detection-based modeling of NDMA occurrence. It is assumed that the number of samples
taken per PWS is comparable to the number taken for UCMR 2 sampling.
1) Small = serving < 10,000; Large = serving 10,001 - 100,000; Very Large = serving > 100,000.
As shown in Exhibit 5.39, the predicted percentage of PWSs having one or more NDMA
detections above the HRL (0.6 ng/L) is between 22.2 and 36.7 percent. Results are significantly
higher for very large PWSs (74.5 percent) than for large and small PWSs (57.5 and 27.0 percent,
respectively). This outcome reflects the fact that very large PWSs have more sample locations
and thus have more opportunities to observe an NDMA concentration above the HRL. It also
reflects the longer residence time in very large PWSs' distribution systems.
Exhibit 5.41 compares the predicted population exposed to at least one sample with levels
greater than the HRL (0.6 ng/L) for different source water and disinfectant types. Results for
PWSs follow a similar pattern to results for EP and MR location analysis: exposure is highest in
PWSs with surface water or mixed sources and those using chloramines. Note that in Exhibit
5.41 there is no predicted exposure in the "mixed water ND" category, while the same is not true
in Exhibit 5.38. This is a consequence of the way PWSs, EPs, and MRs are classified, as
described in Section 5.2.2. No PWSs in the data set are classified as "mixed water ND."
Six-Year Review 3 5-48
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 5.41: Percentage of the Population Predicted To Be Exposed to One or
More Detections at Levels Above the HRL (0.6 ng/L), Based on PWSs
Mixed
ND
CA
CA + CL/OT
CL
CL+OT
OT
ND
¦ GW
90.2%
75.4%
48.7%
47.5%
40.3%
18.0%
¦ SW
98.3%
92.4%
62.3%
64.0%
68,7%
54.5%
Mixed
98.3%
96.3%
73.4%
64.2%
99.8%
0.0%
Source: Appendix D.
Abbreviations: HRL = health reference level; CA = chloramines; CL = chlorine; ND = no disinfectant; OT = other.
Note: This exhibit shows the results of detection-based modeling of NDMA occurrence.
5.5.3 National Occurrence and Exposure Estimates: Mean Concentrations
In addition to predicting the percentage of sampling locations and PWSs with one or more
detections over given thresholds, EPA used modeling to predict the percentage of sampling
locations and PWSs with a mean NDMA concentration over the same thresholds. The mean
NDMA concentration is the annual average value calculated by the model. For the EP and MR
analyses, it is the average annual value at each location based on four quarterly samples for
surface water and mixed PWSs and two semiannual samples for ground water PWSs (i.e., the
LAA). For the PWS analysis, it is the average of all annual average values at all EP and MR
sample locations within the PWS (i.e., PWS annual average). Results are presented first for EP
and MR locations, then for PWSs as a whole.
5.5.3.1 Entry Points and Maximum Residence Locations
Exhibit 5.42 and Exhibit 5.43 summarize results of the mean-based modeling analysis for EP and
MR locations, respectively. They show the mean expected value concentrations and the 90
percent credible intervals for three size categories and for three threshold values. See Appendix
E for results for other thresholds and for predictions of the total population exposed (extrapolated
from the UCMR 2 population). Similar to the modeled analysis of detections, the analysis of
mean concentrations at EPs gives a lower bound estimate of exposure based on NDMA
concentrations in finished water entering the distribution system. The analysis based on MRs
Six-Year Review 3 5-49
Technical Support Document for Nitrosamines
December 2016
-------
gives an upper bound estimate of exposure at locations that are expected to represent higher
NDMA concentrations.
Exhibit 5.42: Percentage of Entry Points with the Predicted Mean Concentration
Exceeding the Threshold and the Associated Population Exposed
PWS Size1
Total
National
Inventory:
PWS Count
Total
National
Inventory:
Population
Served
Threshold
(ng/L)
Percentage of EPs
with Mean
Exceeding
Threshold:
Expected Value
(90% CI)
Percentage of
Population Served by
EPs with Mean
Exceeding Threshold:
Expected Value (90% CI)
Small
41,962
41,154,840
0.6
17.3%
(12.9% -22.9%)
22.7%
(16.5%-30.2%)
Small
41,962
41,154,840
2
7.1%
(5.0%-9.1%)
10.8%
(7.6%-14.3%)
Small
41,962
41,154,840
6
3.0%
(2.1%-4.3%)
4.6%
(3.0% - 7.2%)
Large
2,812
84,318,927
0.6
14.1%
(10.3%-18.9%)
19.7%
(14.4%-26.6%)
Large
2,812
84,318,927
2
5.1%
(3.6% - 6.9%)
7.7%
(5.4%-10.6%)
Large
2,812
84,318,927
6
1.4%
(0.7% - 2.2%)
2.5%
(1.3%-4.0%)
Very Large
394
141,407,211
0.6
15.7%
(12.6%-19.3%)
23.6%
(18.1%-31.3%)
Very Large
394
141,407,211
2
4.7%
(3.8% - 5.9%)
8.4%
(6.4%-10.6%)
Very Large
394
141,407,211
6
1.2%
(0.8%-1.6%)
2.5%
(1.7%-3.5%)
All
45,168
266,880,978
0.6
15.5%
(12.1%-19.7%)
23.2%
(17.8%-30.9%)
All
45,168
266,880,978
2
5.2%
(3.9% - 6.6%)
8.3%
(6.3%-10.6%)
All
45,168
266,880,978
6
1.5%
(0.9%-2.1%)
2.5%
(1.7%-3.6%)
Source: Appendix E.
Abbreviation: CI = credible interval; EP = entry point
Note: This exhibit shows the results of mean-based modeling of NDMA occurrence.
1) Small = serving < 10,000; Large = serving 10,001 - 100,000; Very Large = serving > 100,000.
Exhibit 5.43: Percentage of Maximum Residence Time Locations with the
Predicted Mean Concentration Exceeding the Threshold and the Associated
Population Exposed
PWS Size1
Total
National
Inventory:
PWS Count
Total
National
Inventory:
Population
Served
Threshold
(ng/L)
Percentage of MRs
with Mean Exceeding
Threshold: Expected
Value (90% CI)
Percentage of
Population Served by
MRs with Mean
Exceeding Threshold:
Expected Value (90% CI)
Small
41,962
41,154,840
0.6
29.5%
(23.3%-37.0%)
34.5%
(26.5%-43.9%)
Small
41,962
41,154,840
2
11.8%
(9.3%-14.7%)
15.9%
(12.1%-20.0%)
Small
41,962
41,154,840
6
5.1%
(3.8% - 6.8%)
7.7%
(5.6%-10.8%)
Large
2,812
84,318,927
0.6
28.6%
(22.1%-36.2%)
36.5%
(28.2% - 46.4%)
Large
2,812
84,318,927
2
11.0%
(8.3%-14.2%)
16.2%
(12.4%-20.9%)
Six-Year Review 3 5-50
Technical Support Document for Nitrosamines
December 2016
-------
PWS Size1
Total
National
Inventory:
PWS Count
Total
National
Inventory:
Population
Served
Threshold
(ng/L)
Percentage of MRs
with Mean Exceeding
Threshold: Expected
Value (90% CI)
Percentage of
Population Served by
MRs with Mean
Exceeding Threshold:
Expected Value (90% CI)
Large
2,812
84,318,927
6
3.6%
(2.2% - 5.5%)
6.1%
(3.5%-8.9%)
Very Large
394
141,407,211
0.6
34.9%
(29.7%-41.0%)
46.6%
(38.4% - 56.6%)
Very Large
394
141,407,211
2
16.5%
(14.4%-18.9%)
23.8%
(19.9%-28.8%)
Very Large
394
141,407,211
6
4.7%
(3.7% - 5.9%)
6.0%
(4.4%-9.1%)
All
45,168
266,880,978
0.6
32.3%
(26.6%-39.1%)
45.7%
(37.5%-55.7%)
All
45,168
266,880,978
2
14.2%
(11.9%-16.9%)
23.1%
(19.3%-28.1%)
All
45,168
266,880,978
6
4.5%
(3.4% - 6.0%)
6.0%
(4.4%-9.1%)
Source: Appendix E.
Abbreviation: CI = credible interval; MR = maximum residence time location
Note: This exhibit shows the results of mean-based modeling of NDMA occurrence.
1) Small = serving < 10,000; Large = serving 10,001 - 100,000; Very Large = serving > 100,000.
Exhibit 5.42 and Exhibit 5.43 show that between 12.1 and 19.7 percent of EP locations and
between 26.6 and 39.1 percent of MR locations are predicted to have mean NDMA
concentrations greater than the HRL (0.6 ng/L). These results are lower than the detection-based
modeled predictions shown in Section 5.5.2 (i.e., Exhibit 5.35 and Exhibit 5.36), which is
expected because individual samples greater than the HRL may be averaged with lower values to
result in a mean below the HRL for the present modeled analysis, whereas they would be
counted as above the HRL for the detection-based modeled analysis.
Exhibit 5.44 and Exhibit 5.45 show the results of mean-based modeling by disinfectant type and
source water type for EP and MR locations, respectively. Similar to the results of detection-based
modeling, results of mean-based modeling are highest in PWSs using chloramines alone and are
also significantly higher in PWSs using chloramines in combination with chlorine. Other
disinfectants, including chlorine alone, chlorine with other disinfectants, other disinfectants alone
and no disinfection show lower exposures to mean concentrations above the MRL, especially at
EP locations. It is interesting to note that for the MR analysis, a high percentage of the
population is predicted to be exposed to mean NDMA concentrations over the HRL for mixed
source waters using other or no disinfectants. This result may be influenced by the small number
of PWSs in this category.
Six-Year Review 3 5-51
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 5.44: Percentage of the Population Predicted To Be Exposed to Mean
Concentrations Greater Than the HRL (0.6 ng/L) at Entry Points
100%
80%
60%
40%
20%
0%
Mixed
CA
CA +
CL/OT
sw
CL+ OT
GW
OT
T3
03
l/>
9- 80%
o
"+¦»
J2
3
Q.
O
CL
-------
Exhibit 5.45: Percentage of the Population Predicted To Be Exposed to Mean
Concentrations Greater Than the HRL (0.6 ng/L) at Maximum Residence Time
Locations
Mixed
CA
CA + CL/OT
CL
CL + OT
OT
ND
¦ GW
77.9%
40.3%
7.7%
7.4%
19.9%
10,5%
¦ sw
89.5%
74.7%
22,3%
15.1%
38.9%
28,7%
Mixed
83.4%,
76-9%
26.3%
42.2%
100.0%
63.8%
Source: Appendix E.
Abbreviations: HRL = health reference level; CA = chloramines; CL = chlorine; ND = no disinfectant; OT = other.
Note: This exhibit shows the results of mean-based modeling of NDMA occurrence.
5.5.3.2 PWSs
EPA also used modeling to predict the number and percentage of PWSs with a PWS-wide
average NDMA concentration greater than each of the three thresholds. Exhibit 5.46 shows the
expected mean values and the 90 percent credible interval for three size categories and for three
threshold values. Exhibit 5.47 shows predicted total population exposed at those thresholds. See
Appendix F for results at other thresholds.
Results show that between 4.8 and 11.2 percent of PWSs are predicted to have average NDMA
concentrations greater than the HRL (0.6 ng/L).
Six-Year Review 3 5-53
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 5.46: PWSs with Predicted Mean Concentration Exceeding Various
Thresholds
PWS Size1
Total
National
Inventory:
PWS
Count
Total
National
Inventory:
Population
Served
Threshold
(ng/L)
PWSs with Predicted
Mean Exceeding
Threshold: Expected
Value (90% CI)
Percentage of PWSs
with Predicted Mean
Exceeding Threshold:
Expected Value (90% CI)
Small
41,962
41,154,840
0.6
2,705
(1,669-4,207)
6.5%
(4.0%-10.0%)
Small
41,962
41,154,840
2
1,056
(546 -1,591)
2.5%
(1.3%-3.8%)
Small
41,962
41,154,840
6
415
(340-483)
1.0%
(0.8%-1.2%)
Large
2,812
84,318,927
0.6
541
(408-695)
19.2%
(14.5%-24.7%)
Large
2,812
84,318,927
2
222
(150-296)
7.9%
(5.3%-10.5%)
Large
2,812
84,318,927
6
76
(58- 120)
2.7%
(2.1%-4.3%)
Very Large
394
141,407,211
0.6
121
(101 - 145)
30.5%
(25.4% - 36.4%)
Very Large
394
141,407,211
2
46
(37-55)
11.7%
(9.3%-13.8%)
Very Large
394
141,407,211
6
14
(10-17)
3.5%
(2.5%-4.3%)
All
45,168
266,880,978
0.6
3,367
(2,178-5,048)
7.5%
(4.8%-11.2%)
All
45,168
266,880,978
2
1,324
(733 - 1,942)
2.9%
(1.6%-4.3%)
All
45,168
266,880,978
6
505
(408-620)
1.1%
(0.9%-1.4%)
Source: Appendix F.
Abbreviation: CI = credible interval
Note: This exhibit shows the results of mean-based modeling of NDMA occurrence.
1) Small = serving < 10,000; Large = serving 10,001 - 100,000; Very Large = serving > 100,000.
Six-Year Review 3 5-54
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 5.47: Population Served by PWSs with Predicted Mean Concentration
Exceeding Various Thresholds
PWS Size1
Total
National
Inventory:
PWS Count
Total National
Inventory:
Population
Served
Threshold
(ng/L)
Population Served by
PWSs with Predicted Mean
Exceeding Threshold:
Expected Value (90% CI)
Percentage of Population
Served by PWSs with
Predicted Mean
Exceeding Threshold:
Expected Value (90% CI)
Small
41,962
41,154,840
0.6
4,885,569
(3,172,960-7,631,299)
11.9%
(7.7% -18.5%)
Small
41,962
41,154,840
2
1,995,203
(1,092,084-2,846,984)
4.9%
(2.7% - 6.9%)
Small
41,962
41,154,840
6
752,480
(481,741 - 1,045,792)
1.8%
(1.2%-2.5%)
Large
2,812
84,318,927
0.6
17,731,575
(14,201,885-22,622,389)
21.0%
(16.8%-26.8%)
Large
2,812
84,318,927
2
7,866,335
(5,235,856- 10,817,716)
9.3%
(6.2%-12.8%)
Large
2,812
84,318,927
6
2,252,768
(1,757,426-4,420,072)
2.7%
(2.1%-5.2%)
Very Large
394
141,407,211
0.6
42,734,267
(32,742,467-54,906,687)
30.1%
(23.1%-38.7%)
Very Large
394
141,407,211
2
14,892,932
(12,187,487- 18,796,017)
10.5%
(8.6% -13.2%)
Very Large
394
141,407,211
6
3,168,706
(2,231,047-3,966,754)
2.2%
(1.6%-2.8%)
All
45,168
266,880,978
0.6
65,351,410
(50,117,312 -85,160,375)
24.4%
(18.7%-31.8%)
All
45,168
266,880,978
2
24,754,470
(18,515,427-32,460,717)
9.3%
(6.9%-12.1%)
All
45,168
266,880,978
6
6,173,954
(4,470,214-9,432,618)
2.3%
(1.7%-3.5%)
Source: Appendix F.
Abbreviation: CI = credible interval
Note: This exhibit shows the results of mean-based modeling of NDMA occurrence.
1) Small = serving < 10,000; Large = serving 10,001 - 100,000; Very Large = serving > 100,000.
Exhibit 5.48 shows trends with respect to source water and disinfection type based on mean-
based modeling for PWSs. With limited exceptions (namely, mixed water PWSs using chlorine
in combination with other disinfectants), PWSs using non-chloramine disinfectants are predicted
to have less than 15 percent of their served population exposed to mean concentrations greater
than the HRL. Nearly 74 percent of customers served by surface water PWSs using chloramines
are predicted to be exposed to average NDMA concentrations above the HRL. The percentage of
the population exposed for ground water PWSs is significantly lower than that for surface water
and mixed water PWSs in all disinfectant categories.
Six-Year Review 3 5-55
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 5.48: Percentage of the Population Predicted To Be Exposed to Mean
Concentrations Greater Than the HRL (0.6 ng/L) in PWSs
Mixed
ND
CA
CA + CL/OT
CL
CL+OT
OT
ND
¦ GW
45.0%
10.5%
1.3%
0.2%
2.9%
1.2%
¦ SW
73.9%
39.7%
11.6%
1.1%
14.2%
7.2%
Mixed
70.6%
46.2%
5.3%
37.5%
11.8%
0.0%
Source: Appendix F.
Abbreviations: HRL = health reference level; CA = chloramines; CL = chlorine; ND = no disinfectant; OT = other.
Note: This exhibit shows the results of mean-based modeling of NDMA occurrence.
5.5.3.3 National Occurrence and Exposure Estimates: Locational Annual Average
EPA modeled the LAA of NDMA at each UCMR 2 monitoring location. (Note that these LAAs
are not necessarily the same as the LRAAs identified for the Stage 2 D/DBPR.) EPA conducted
the LAA analysis at the PWS level, meaning that if one location in a PWS exceeded the
threshold based on the LAA, the entire PWS was counted. The associated population exposed to
LAAs above the threshold, however, was based on the population served by individual sampling
locations. As discussed earlier in this chapter, specific information on population served per
sampling location was not available from the UCMR 2 dataset; thus, EPA assumed that an equal
proportion of the population served could be assigned to each sampling location within the PWS.
It is important to note that because the UCMR 2 dataset contains only one year of data, the LAA
is equivalent to the annual average value at each location.
Exhibit 5.49 shows the expected mean percentages and the 90 percent credible interval for 3 size
categories and for 3 threshold values for the LAA analysis. See Appendix G for results for other
thresholds and for associated predicted total population exposed (extrapolated from the UCMR 2
population). Results show that between 15.1 and 29.8 percent of PWSs are predicted to exceed
the HRL (0.6 ng/L) at one or more sampling locations, based on the LAA. Predicted percentages
are significantly higher for very large PWSs compared to large PWSs and for large PWSs
compared to small PWSs. This may be due to the higher number of sampling locations in larger
PWSs and thus the greater chance for an LAA to have an occasional high concentration. Note
Six-Year Review 3 5-56
Technical Support Document for Nitrosamines
December 2016
-------
that the percentage of PWSs predicted to have an LAA above the HRL and the associated
population served are roughly equivalent to the sum of the predicted values for EP and MR
locations based on the annual mean (see Section 5.5.3).
Exhibit 5.49: Percentage of PWSs Predicted To Have an LAA Greater Than the
Threshold and the Associated Population Exposed
PWS Size1
Total
National
Inventory:
PWS
Count
Total
National
Inventory:
Population
Served
Threshold
(ng/L)
Percentage of PWSs
Predicted To Have LAA
Exceeding Threshold:
Expected Value
(90% CI)
Percentage of Population
Served by PWSs
Predicted To Have LAA
Exceeding Threshold:
Expected Value
(90% CI)
Small
41,962
41,154,840
0.6
20.2%
(13.4%-27.7%)
16.5%
(11.4%-22.9%)
Small
41,962
41,154,840
2
5.4%
(3.3% - 8.3%)
6.0%
(3.9% - 9.0%)
Small
41,962
41,154,840
6
1.8%
(1.3%-3.0%)
2.1%
1.5%-3.3%)
Large
2,812
84,318,927
0.6
45.2%
(34.8%-56.5%)
26.3%
(21.2%-32.5%)
Large
2,812
84,318,927
2
18.6%
(14.0%-23.9%)
11.1%
(9.0%-13.8%)
Large
2,812
84,318,927
6
7.0%
(4.8%-11.4%)
3.9%
(2.4% - 5.6%)
Very Large
394
141,407,211
0.6
62.2%
(53.3%-70.9%)
33.7%
(28.5%-40.6%)
Very Large
394
141,407,211
2
33.1%
(28.4% - 38.4%)
15.1%
(12.7%-17.8%)
Very Large
394
141,407,211
6
10.5%
(7.3%-14.1%)
4.0%
(3.1%-5.7%)
All
45,168
266,880,978
0.6
22.1%
(15.1%-29.8%)
28.7%
(23.5%-35.3%)
All
45,168
266,880,978
2
6.5%
(4.2% - 9.5%)
12.5%
(10.2%-15.2%)
All
45,168
266,880,978
6
2.2%
(1.5%-3.6%)
3.7%
(2.6% - 5.3%)
Source: Appendix G.
Abbreviation: CI = credible interval; LAA = locational annual average.
Note: This exhibit shows the results of mean-based modeling of NDMA occurrence.
1) Small = serving < 10,000; Large = serving 10,001 - 100,000; Very Large = serving > 100,000.
Exhibit 5.50 compares population exposed for different source water and disinfection types. As
with previous analyses, results are higher in surface water and mixed water PWSs compared to
ground water PWSs, as well as for PWSs using chloramines compared to other disinfectants. The
predicted percentage of the population exposed to LAAs greater than the HRL ranges from
nearly 71 percent for surface water PWSs using chloramines to 3 percent for ground water PWSs
using other disinfectants with chlorine or not disinfecting.
Six-Year Review 3 5-57
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 5.50: Percentage of the Population Predicted To Be Exposed to LAAs
Greater Than the HRL (0.6 ng/L) in PWSs
Mixed
CA
CA + CL/OT
CL
CL+OT
OT
ND
¦ GW
49,4%
22.9%
6.2%
3.5%
13.3%
2.9%
¦ SW
70.9%
46,6%
19.1%
9.6%
28.0%
19.7%
Mixed
63.9%
45.0%
11.5%
34.4%
29.9%
0.0%
Source: Appendix G.
Abbreviations: HRL = health reference level; CA = chloramines; CL = chlorine; LAA = locational annual average; ND = no
disinfectant; OT = other.
Note: This exhibit shows the results of mean-based modeling of NDMA occurrence.
5.5.4 National Co-Occurrence of Nitrosamines
Exhibit 5.51 shows co-occurrence results within the nitrosamine group, based on exposure at or
above the MRLs (see Section 5.4, in particular Exhibit 5.32 and Exhibit 5.33), extrapolated from
the UCMR 2 study population to the national population. These national extrapolations are
produced by multiplying occurrence rates from the study by national baseline inventory numbers
for PWSs in various size and source water categories. Since UCMR 2 monitoring included a
census of very large systems, no extrapolation was necessary in that size category. Note that the
simple extrapolation methodology used here produces different results for NDMA than the
modeling process described in Section 5.5. As noted in earlier discussions of co-occurrence,
these results should be interpreted with caution since MRLs vary from contaminant to
contaminant.
In summary, an estimated 4,591 PWSs serving a population of just over 101 million people are
potentially exposed to detectable levels of nitrosamines (NDMA or other) in finished water. An
estimated 123 PWSs serving a population of approximately 10.5 million people are potentially
exposed to more than one nitrosamine.
Six-Year Review 3 5-58
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 5.51: Co-Occurrence of Nitrosamines (National Extrapolation from UCMR 2)
PWS Size1
Source Water
Type
National
Estimate of
PWSs with
At Least One
Detection of
NDMA Only
National
Estimate of
Population
Served by
PWSs with At
Least One
Detection of
NDMA Only
National
Estimate of
PWSs with
At Least One
Detection of
Other
Nitrosamines
(No NDMA)
National
Estimate of
Population
Served by
PWSs with At
Least One
Detection of
Other
Nitrosamines
(No NDMA)
National
Estimate of
PWSs with
Detection of
NDMA and
Another
Nitrosamine
National
Estimate of
Population
Served by
PWSs with
Detection of
NDMA and
Another
Nitrosamine
National
Estimate of
PWSs with
At Least
One
Detection of
Any
Nitrosamine
National
Estimate of
Population
Served by
PWSs with
At Least One
Detection of
Any
Nitrosamine
Small
All PWSs
3,145
6,330,127
482
1,058,317
54
126,180
3,680
7,514,624
Small
Surface Water
492
1,242,334
0
0
54
126,180
546
1,368,514
Small
Ground Water
1,605
2,268,705
482
1,058,317
0
0
2,087
3,327,022
Small
Mixed Water
1,048
2,819,088
0
0
0
0
1,048
2,819,088
Large
All PWSs
627
21,073,473
55
1,584,277
48
2,014,691
729
24,672,442
Large
Surface Water
238
7,823,545
12
266,072
18
576,193
268
8,665,811
Large
Ground Water
111
3,435,597
0
0
8
464,811
119
3,900,408
Large
Mixed Water
278
9,814,331
43
1,318,205
21
973,687
342
12,106,223
Very Large
All PWSs
147
53,138,701
12
7,530,629
22
8,355,589
181
69,024,919
Very Large
Surface Water
85
35,243,460
4
2,109,340
8
4,262,160
97
41,614,960
Very Large
Ground Water
15
4,792,295
4
609,237
3
1,253,283
22
6,654,815
Very Large
Mixed Water
47
13,102,946
4
4,812,052
11
2,840,146
62
20,755,144
Total
3,919
80,542,301
549
10,173,223
123
10,496,460
4,591
101,211,984
Note: Based on UCMR 2 survey results, extrapolated to the national level using information from SDWIS on the population served by PWSs in various size and source water categories. "Other
nitrosamines" detected under the UCMR 2 are NDBA, NDEA, NMEA and NPYR. NDPA was not detected by any PWS. See Exhibit 5.1 for the MRLs for each nitrosamine.
1) Small = serving < 10,000; Large = serving 10,001 - 100,000; Very Large = serving > 100,000.
Six-Year Review 3
Technical Support Document for Nitrosamines
5-59
December 2016
-------
5.6 Summary and Discussion
EPA used data from the UCMR 2 (2008-2010) Screening Survey to evaluate occurrence of the
contaminants in six specific nitrosamines in U.S. drinking water. Under the UCMR 2 Screening
Survey, all very large PWSs and a sample of small and large PWSs were required to collect
samples at each EP to the distribution system and at the MR time locations within the
distribution system associated with each EP with disinfection, for a period of one year.
Monitoring was conducted quarterly at surface water PWSs and semi-annually at ground water
PWSs. PWSs were required to report the type of disinfection in use at the time of sample
collection.
Detections of one or more nitrosamines occurred in 10.6 percent (1,907 of 18,053) of samples
and at 28.6 percent (343 of 1,198) of UCMR 2 PWSs. All nitrosamines except NDPA were
detected during the course of the survey. NDMA predominated, with detections in 10.2 percent
(1,841 of 18,040) of samples and 27.0 percent (324 of 1,198) of PWSs. There was substantial co-
occurrence (i.e., detection of two or more nitrosamines at the same PWS during the year of
monitoring). In all but one instance, NDMA was one of the co-occurring nitrosamines. Detection
rates appear to be lower at ground water PWSs than at surface and mixed water PWSs. Three
points of caution should be borne in mind when interpreting nitrosamine results from UCMR 2.
First, reporting thresholds (MRLs) were not uniform, making contaminant-to-contaminant
comparisons uncertain. Second, MRLs were not always low enough to capture occurrence at
levels of health concern (i.e., the MRLs were above the contaminants' respective HRLs). In the
case of NDMA and NDEA, HRLs were lower than MRLs, and therefore the summary statistics
based on MRLs probably underestimate exposure to those contaminants at levels of health
concern. Third, the UCMR 2 survey of PWSs represents a combination of sample and census,
and therefore summary statistics from the survey should not be interpreted as a simple surrogate
for national occurrence.
Detection rates for NDMA were high enough to support modeling that allowed for
characterization of occurrence at concentrations below the MRL, to overcome the limitation that
its HRL was below its MRL. Modeling suggests that monitoring at the national level (under the
same schedule as UCMR 2), with sufficiently sensitive analytical methods, would show NDMA
occurring at levels of health concern (>HRL) in 22.1 percent to 36.7 percent (90 percent credible
interval range) of the nation's PWSs, affecting approximately 149.9 million to 202.6 million
people. Modeling also predicts that approximately 4.9 percent to 11.1 percent PWSs nationally
(again, a 90 percent credible interval range), serving approximately 50.1 million to 85.2 million
people, would be found to have average NDMA concentrations in excess of the HRL. For
NDMA, which is a carcinogen, the average concentration at a PWS is an appropriate benchmark
for evaluating health outcomes based on chronic exposure.
Since nitrosamines are disinfection by-products (DBPs), one would expect to see variations in
occurrence depending on the disinfectant in use. In particular, higher concentrations and higher
frequencies of detection would be expected at PWSs and at monitoring locations where
chloramines are in use, alone or in combination with other disinfectants. The UCMR 2 data bear
out these correlations. The formation of nitrosamines (including NDMA) is discussed in more
detail in Chapter 6.
UCMR 2 data are subject to limitations including some that are particular to the nitrosamines.
For example, UCMR 2 monitoring occurred between 2008 and 2010, and thus does not capture
Six-Year Review 3 5-60
Technical Support Document for Nitrosamines
December 2016
-------
treatment adjustments that were made by utilities between 2011 and 2015 in response to
LT2SWTR and Stage 2 D/DBPR. The treatment changes under Stage 2 D/DBPR are anticipated
to increase the use of chloramines to reduce the formation of certain other chlorination DBPs
(trihalomethanes [THMs] and haloacetic acids) (USEPA, 2005d), and a possible consequence
will be additional utilities having the increased formation of nitrosamines. A comparison of
disinfectant uses between the period of 2008-2010 (based on UCMR 2 data) and 2013-2015
(based on UCMR 3 data) is presented and discussed in Six Year Review 3 Technical Support
Document for Chlorate (USEPA, 2016b). Such a comparison confirms an increasing trend of
chloramines usage in the nation. Thus, current national occurrence and exposure baselines of
nitrosamines (including NDMA) could be higher than the UCMR 2 data indicate.
Another limitation is that UCMR 2 did not involve monitoring at consecutive systems (PWSs
that purchase 100 percent of their water from other systems and are "downstream" of the system
selling and providing the water). Although the population-served values of participating UCMR
2 wholesale systems (systems selling water to other systems) were adjusted to account for
populations served by purchasing systems, the nitrosamine levels observed in wholesale systems
might not be representative of levels present in the consecutive purchasing systems. A study by
Krasner et al. (2012a) found that NDMA levels increased in distribution systems of consecutive
systems (primarily because of increased residence time). Both of these limitations suggest the
UCMR 2 results may underestimate current, and possibly future, public exposure to nitrosamines
in drinking water.
It is worth bearing in mind that the six nitrosamines discussed in this document are only a subset
of the nitrosamines that may occur in drinking water, many of which have not yet been
identified. Dai and Mitch (2013) developed a TONO assay and applied it to 36 finished water
samples from 11 drinking water treatment plants (including 9 plants using chloramines for
secondary disinfection) and found that NDMA accounted for only around 5 percent of the total
nitrosamines present, and other nitrosamines detectable with EPA Method 521 (including the
other five discussed in this document) accounted for less. Application of the assay to influent
waters indicates that while source waters impaired by algal blooms are not an important source
of NDMA and other nitrosamines detectable with EPA Method 521, they are an important source
of total A-nitrosamines.
Six-Year Review 3 5-61
Technical Support Document for Nitrosamines
December 2016
-------
6 Formation
6.1 Introduction
Although nitrosamines have been found in drinking water sources due to contamination from
chemical manufacturing and the degradation of hydrazine rocket fuel, as discussed in Chapter 2,
their occurrence in source water appears infrequent and at low concentrations when they are
present. However, as described in Chapter 5, data from Second Unregulated Contaminant
Monitoring Regulation (UCMR 2) show that A-nitrosodi methyl amine (NDMA) occurs
frequently in treated drinking water, especially at utilities that use chloramines for disinfection.
(Other nitrosamines have also been detected to a much lesser extent.) These observations are
supported by other studies showing that NDMA formation is elevated when chloramination is
used in surface water where nitrogen-containing organic precursors are present.
This chapter presents information on the mechanisms of formation, sources and types of
precursors, factors impacting formation, formation kinetics, and predictive models. The main
focus of this chapter is NDMA, which is most frequently detected in the UCMR 2, though
information about the other nitrosamines is also included.
Data from the literature on a range of chloraminated natural and wastewater-impacted waters
indicates that the formation potential (FP) for the nitrosamines A-nitrosodi-n-butylamine
(NDBA), A-nitrosodi ethyl amine (NDEA), A'-nitrosodi-n-propyl amine (NDPA), N-
nitrosomethylethylamine (NMEA), A-nitrosopyrrolidine (NPYR) and A'-nitrosopi peri dine
(NPIP) is one to two orders of magnitude lower than that for NDMA (Sacher et al., 2008). Data
from UCMR 2 show that although the ranges of observed concentrations of individual
nitrosamines are similar, NDMA is detected far more frequently than the others. Thus, this
chapter focuses on formation of NDMA.
6.2 Nitrosamine Formation Potential
Many organic nitrogen-containing compounds can contribute to the formation of nitrosamines
(Sacher et al., 2008). Although correlations between dissolved organic nitrogen (DON)
concentration and NDMA precursors have been observed in some studies, they are absent in
others (Lee et al., 2007a; Krasner et al., 2008; Mitch et al., 2009; Xu et al., 2011). The observed
relationships between dissolved organic carbon (DOC) concentration and nitrosamine formation
are also inconsistent (Gerecke and Sedlak, 2003; Hua et al., 2007; Lee et al., 2007a; Krasner et
al., 2008; Mitch et al., 2009; Xu et al., 2011). Dimethylamine (DMA), a known NDMA
precursor that has been widely used in laboratory studies, can be measured at relatively low
concentrations (Mitch et al., 2003a). However, because DMA is not the only precursor
responsible for NDMA formation in drinking water sources, its measurement has limited use in
quantifying NDMA formation. Therefore, to evaluate the presence of nitrosamine precursors in a
variety of water supplies, researchers have developed tests to determine nitrosamine FP.
Nitrosamine FP tests serve as surrogate measures of nitrosamine precursors, and are analogous to
the disinfection by-product (DBP) FP tests used to determine the level of other DBP precursors
in source water.
Nitrosamine FP tests are performed in a variety of ways. For example, Mitch et al. (2003) added
2 mM monochloramine to a water phosphate-buffered at pH 6.8 with a reaction duration of up to
10 days. Krasner et al. (2009), on the other hand, formed chloramines in situ by adding ammonia
Six-Year Review 3 6-1
Technical Support Document for Nitrosamines
December 2016
-------
and then chlorine; samples were held for three days at room temperature at pH 8. Different
experimental protocols, as well as different source waters and different types of precursors, mean
that FP tests may not be directly comparable and should be interpreted with caution.
Although FP tests do not directly predict the concentrations of nitrosamines that will form in
water under field operating conditions, as the tests are typically conducted using extreme
chlorine doses, correlations can be found between FP tests and expected concentrations of
NDMA (Mitch et al., 2003a). The results from FP tests should be carefully evaluated and may
not be comparable across different waters due to varying water quality parameters (i.e., pH,
bromide concentrations, etc.) or treatment (i.e., chlorine vs. chloramine). The FP test, however, is
still a useful tool for estimating precursor concentrations in a drinking water source and
evaluating various nitrosamine removal processes within a single treatment plant (Mitch et al.,
2003a).
Krasner et al. (2011) presented results of a simulated distribution system test to mimic potential
nitrosamine and regulated DBP formation under more realistic disinfection conditions. They
validated simulated distribution system tests against full-scale occurrence for NDMA and
halogenated DBPs (trihalomethanes [THMs] and haloacetic acids [HAAs]). Such a test may help
utilities to determine optimal conditions for controlling a range of DBPs while meeting
disinfection requirements.
6.3 Formation Pathways
As shown by the analysis of UCMR 2 data and other data in Appendix A, nitrosamine
occurrence is generally associated with treated drinking water, rather than source water. Utilities
treating surface waters affected by wastewater flows generally show higher nitrosamine
formation compared to those treating ground water or more pristine surface waters (Padhye et al.,
2010). Nitrosamines have been shown to form under many different disinfection techniques,
including the application of free chlorine (Choi and Valentine, 2003; Schreiber and Mitch, 2007;
Nawrocki and Andrzejewski, 2011; Shah and Mitch, 2012), chlorine dioxide (Andrzejewski and
Nawrocki, 2007; Pozzi et al., 2011), ozone (Sedlak et al., 2005; Andrzejewski and Nawrocki,
2007; Andrzejewski et al., 2008; Schmidt and Brauch, 2008; Yang et al., 2009; Pozzi et al.,
2011), ultraviolet (UV) light (Zhao et al., 2008) and advanced oxidation processes (AOPs) (Zhao
et al., 2008). Krasner et al. (2013) provided a literature review to discuss the formation of
nitrosamines under many of these scenarios. However, data from both the literature (Najm and
Trussell, 2001; Andrzejewski et al., 2008; Krasner et al., 2008; Sacher et al., 2008; Zhao et al.,
2008; Mitch et al., 2009; Krasner et al., 2013) and UCMR 2 indicate that formation via the
chloramination pathway is the primary mechanism of interest for drinking water utilities. This
section presents information on the chloramination pathway first, followed by a discussion on the
chlorine-enhanced nitrosation pathway, the breakpoint chlorination pathway and other minor
formation pathways. Nitrosamines formation due to use of disinfectants other than chlorine or
chloramines is also discussed. In addition to the formation pathways that take place in water as
discussed in this section, NDMA can be produced endogenously in humans via the interaction of
nitrates and nitrites with amines in the stomach (Fristachi and Rice, 2007).
6.3.1 Chloramination Pathway
Findings from a wide range of laboratory and full-scale studies agree with the UCMR 2 findings:
chloramination generally results in higher NDMA formation than other disinfection practices
Six-Year Review 3 6-2
Technical Support Document for Nitrosamines
December 2016
-------
(Najm and Trussell, 2001; Krasner et al., 2008; Sacher et al., 2008; Zhao et al., 2008; Mitch et
al., 2009). This finding is especially relevant because many utilities have been switching from
using chlorine to using alternative disinfectants, including chloramines, to reduce the formation
of regulated THMs and HAAs and to comply with the Stage 2 Disinfectants and Disinfection
By-Products Rule (Stage 2 D/DBPR) (Seidel et al., 2005). Such a change is also discussed in
more detail in the agency's Six-Year Review 3 Technical Support Document for Chlorate
(USEPA, 2016b).
Choi and Valentine (2002) first observed that the addition of monochloramine to a solution of
DMA produced two orders of magnitude more NDMA than the addition of chlorine, and six
times as much as the addition of nitrite. The use of isotopically labeled chloramines indicated
that one of the nitrogens found in NDMA originated from chloramines, indicating that NDMA
was a chloramination DBP (Choi and Valentine, 2002).
Initially, NDMA was thought to form primarily via a monochloramine pathway. The proposed
pathway involved a nucleophilic substitution reaction between DMA and monochloramine,
which creates unsymmetrical dimethylhydrazine (UDMH), which is then oxidized by
chloramines to result in the formation of NDMA (Choi et al., 2002; Mitch and Sedlak, 2002).
However, further research indicated that this mechanism did not fully explain the formation of
NDMA. For example, experiments showed that at least two orders of magnitude higher NDMA
concentrations formed following monochloramine application to DMA than to equivalent
concentrations of UDMH (Schreiber and Mitch, 2006a). Also, studies showed that the presence
of dichloramine significantly increased nitrosamine formation (Mitch et al., 2005; Schreiber and
Mitch, 2005).
Under typical drinking water treatment conditions, monochloramine is the dominant chloramine
species, though dichloramine is also present according to the following equilibrium:
2NH2C1 + H+ <-> NHCh + NH4+
Dichloramine formation from the disproportionation of monochloramine is slow, such that its
formation after the application of preformed monochloramine should be minimal (Schreiber and
Mitch, 2006a). Chloramine speciation is impacted by both pH and the chlorine-to-ammonia
(ChiNEb) molar ratio. At a Ch:NH3 ratio of 1.5 and pH 7, monochloramine and dichloramine are
present in approximately equal molar concentrations. Monochloramine predominates above pH
8.5, while dichloramine predominates below pH 5 (Schreiber and Mitch, 2006a). At Ch:NH3
ratios less than or equal to 1.5, monochloramine is the dominant species. At Ch:NH3 ratios
greater than 1.5, dichloramine is the dominant species.
A revised mechanism was proposed that incorporated the more recent research findings,
including results suggesting that the presence of dichloramine resulted in higher NDMA
formation compared to monochloramine, and that formation of NDMA increased with increased
dissolved oxygen (DO) concentration (Schreiber and Mitch, 2006a). The modified pathway
involves a nucleophilic substitution reaction between DMA (or other secondary amines) and
dichloramine, which creates a chlorinated-UDMH intermediate that is then oxidized by DO
(Exhibit 6.1). The revised pathway has been shown to accurately model NDMA formation over a
wide range of formation conditions (Schreiber and Mitch, 2006a). Schreiber and Mitch (2006a)
also found that other nitrosamines (NDEA, NMEA, A'-nitrosomorpholine [NMOR], NPIP,
NPYR) formed along a parallel pathway from their respective secondary amine precursors. The
Six-Year Review 3 6-3
Technical Support Document for Nitrosamines
December 2016
-------
authors showed that nitrosamine formation is enhanced by higher DO concentrations. The
kinetics of this pathway is such that nitrosamine precursors can react with chloramines over
several days resulting in the continued formation of nitrosamines in the distribution system
(Charrois and Hrudey, 2007; Goslan et al., 2009).
Exhibit 6.1: Mechanism of NDMA Formation via the Chloramination Pathway
C, - H3CN [o] H3<\
\ + ViuOl N—NH > N—N
nh nh-ci / \ / \\
H3C h3C C' h3C °
Source: Adapted from Schreiber and Mitch, 2006a.
In general, all nitrosamines show relatively low molar conversions from their respective amines.
Sacher et al. (2008) determined the following molar percent yields for a range of nitrosamines
from the chloramination of their respective secondary amines (amine concentration of 1,000
ng/L, dosed with 0.4 mM chloramines at pH 7, 20 degrees C for 168 hours):
• NDBA: 0.69%
• NDEA: 0.53%
• NDMA: 0.49%
• NDPA: 0.02%
• NMOR: 0.07%
• NPIP: 1.35%
• NPYR: 1.84%
Selbes et al. (2013) tested 21 different amines for NDMA formation with 100 mg/L chloramine
at pH 7.5. They found that aliphatic amines had low yields: for example, DMA had a yield of 1.2
percent, and dimethylbutylamine had a yield of 0.3 percent. Branched alkyl groups had higher
yields, with dimethylisopropylamine (DMiPA) having a yield of 83.9 percent. They also found
amines with aromatic rings one carbon away from the DMA functional group had higher yields:
for example, dimethylbenzylamine (DMBzA) had a yield of 83.8 percent. The investigators also
tested the formation of NDMA with an addition of extra ammonia to lower the concentration of
dichloramine. They found that some amines reacted preferentially with dichloramine, while
others reacted preferentially with monochloramine. Those amines with electron-withdrawing
groups attached to the nitrogen of the DMA functional group reacted preferentially with
monochloramine. Amines with electron-donating groups attached to the nitrogen of the DMA
functional group preferentially reacted with dichloramine. The investigators proposed that this
was due to the reaction occurring through a nucleophilic attack on the chloramine by the amine.
Nitrosamine formation may also occur via unintended chloramination: for example, when
chlorine is applied to waters containing high concentrations of ammonia. Higher NDMA
Six-Year Review 3 6-4
Technical Support Document for Nitrosamines
December 2016
-------
concentrations were observed following the addition of chlorine to secondary municipal
wastewater effluents than when equivalent doses of preformed monochloramine were added
(Schreiber and Mitch, 2005). This unexpected result may be explained by dichloramine
formation during chlorination of waters with high ammonia content, such that the majority of
nitrosamine formation takes place via the chloramination pathway (Schreiber and Mitch, 2005).
6.3.2 Chlorine Enhanced Nitrosation Pathway
The formation of NDMA from chlorination in the absence of ammonia, but in the presence of
DMA, was described by Choi and Valentine (2003). The authors found that free chlorine added
to a solution of nitrite and DMA created more NDMA than the solutions lacking free chlorine.
The results were explained by the formation of the unstable intermediate dinitrogen tetroxide, a
nitrosating agent. (A nitrosating agent reacts with amines to form nitrosoamines by a nitrosation
reaction.) This mechanism is known as the chlorine enhanced nitrosation pathway and is shown
in Exhibit 6.2. Under comparable conditions, formation of NDMA by enhanced nitrosation has
been shown to be significantly lower than by chloramination (Schreiber and Mitch, 2007).
Nawrocki and Andrzejewski (2011) state that the chlorine enhanced nitrosation reaction is slow
and is not expected to contribute much in NDMA formation. Shah and Mitch (2012) note that the
importance of the enhanced nitrosation pathway in drinking water treatment needs further
clarification for drinking water where nitrite concentrations are typically relatively low.
Liu et al. (2014) proposed a nitrosation pathway involving chloramines and tertiary amines as
well. In their mechanism, the chloramine forms a complex which then reacts by an elimination
reaction, eliminating an ONX agent which is trapped by oxygen to form a nitrosating agent
OONX where X is either hydrogen or chlorine. The nitrosating reagent further reacts to form
NO+, which can react with the amine to form a nitrosamine. The model correctly predicts the
relative yields of many tertiary amines. The model predicts that a strong electron-withdrawing
group on the first carbon will lower NDMA yield, while an electron-donating group will increase
yield. The model did not perform well with all tertiary amines; however, so it is likely multiple
mechanisms are at work.
The enhanced nitrosation pathway may be more important for utilities chlorinating source waters
impacted by nitrified wastewater effluents (Shah and Mitch, 2012). Chen and Young (2009)
observed close to a 10-fold increase in the production of NDMA during the chlorination of
diuron (an herbicide and known NDMA precursor) when 5 mg/L nitrate and/or nitrite was added.
This pathway may also be relevant for utilities chlorinating ground water sources, as nitrate (and
nitrite at relatively lower concentrations) is possibly the most widespread contaminant in ground
water (Nolan et al., 2002). A recent report found persistent widespread contamination of nitrate
in California's ground water resources in heavily agricultural areas, indicating that nitrate
pollution will likely worsen over the next several decades (Harter et al., 2012).
Six-Year Review 3 6-5
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 6.2: NDMA Formation via the Enhanced Nitrosation Pathway
o
HOCI + NO
\ +
N—CI
//
O
o
o"
o
\j—CI + no2"
o
o
o
o
o
c
o
o
o
o
h3c
\l—N + NOj" + H+
/ W
H,C O
Source: Adapted from Choi and Valentine, 2003.
6.3.3 Breakpoint Chlorination
Breakpoint chlorination, the practice of adding enough chlorine to overcome the chlorine
demand, results in a free chlorine residual. Schreiber and Mitch (2007) observed that maximum
nitrosamine formation was observed close to the breakpoint (Ch:NH3 molar ratio of 1.7) during
chloramination of DMA over a range of ChiNEb ratios. Initially, the authors hypothesized that
nitrosamines were forming via a combination of chloramination and chlorine enhanced
nitrosation pathways. However, experimental results indicated that the chlorine enhanced
nitrosation pathway could not account for the rapid formation of nitrosamines observed at the
breakpoint. Two concurrent pathways have been proposed during breakpoint chlorination
(Schreiber and Mitch, 2007). For ChiNEb molar ratios less than or equal to 1.5, nitrosamines
form via the relatively slow chloramination reaction, described in Section 6.3.1. At Ch:NH3
molar ratios greater than 1.5, an additional formation pathway occurs that involves reactive
breakpoint chlorination intermediates and results in the rapid formation of nitrosamines. The
authors propose that reactive breakpoint chlorination intermediates are involved in the direct
nitrosation of DMA, though this pathway requires further verification.
Because treatment facilities typically operate with a sufficient chlorine residual, nitrosamine
formation by this pathway (i.e., breakpoint chlorination with no significant free chlorine
residual) is likely to be of minor importance (Shah and Mitch, 2012).
6.3.4 Other Formation Pathways
Though chlorination and chloramination are the main sources of nitrosamines in drinking water
treatment, other treatment techniques have also been shown to result in the formation of
nitrosamines.
In the absence of a disinfectant, activated carbon has been shown to catalyze the transformation
of NDBA, NDEA, NDMA, NDPA and NMEA from di-//-butylamine (DBA), diethylamine
(DEA), DMA, di-//-propylamine (DPA), and methylethylamine (MEA), respectively (Padhye et
Six-Year Review 3 6-6
Technical Support Document for Nitrosamines
December 2016
-------
al., 2010). Yields ranged from 0.05 to 0.29 percent (calculated by dividing the amount of
nitrosamine formed by the amount of secondary amine adsorbed at pH 7.5) (Padhye et al., 2010).
Nitrosamine formation was higher for experiments using wastewater in comparison to surface
water or deionized water. These findings are relevant not only for drinking water treatment, but
also for analytical methods, as activated carbon is used for nitrosamine extraction (Padhye et al.,
2010).
Some studies have shown that oxidation of DMA may lead to the formation of NDMA,
particularly at higher pH (Nawrocki and Andrzejewski, 2011). The application of strong oxidants
such as ozone, chlorine dioxide, permanganate, UV and hydrogen peroxide to DMA in the
presence and absence of ammonia has been shown to result in NDMA formation (Andrzejewski
and Nawrocki, 2007). A study at 11 wastewater plants using ozone found NDMA formation with
concentrations ranging from less than 10 ng/L to 143 ng/L. NDMA formation was less in
nitrified wastewaters and when the ozone-to-DOC ratio was less than 0.5 (Gerrity et al., 2014). It
is suspected that the reaction between DMA and strong oxidants is direct nitrosation involving
the oxidation of DMA into nitrite and nitrates, though the yields from these reactions are
generally small and pH-dependent (Nawrocki and Andrzejewski, 2011).
Other studies have also documented NDMA formation via ozonation of DMA, but yields were
low and formation only occurred at specific ozone-to-DMA ratios (Andrzejewski et al., 2008;
Yang et al., 2009). Likewise, ozonation of MEA and DEA at a 1-to-l molar ratio formed NMEA
and NDEA, although yields were low and experiments were performed at pH 10.5
(Andrzejewski et al., 2008). Ozonation of UDMH or select compounds with UDMH-like
functional groups results in NDMA yields of greater than 50 percent, however (Schmidt and
Brauch, 2008; Kosaka et al., 2009). Marti et al. (2015) found that reaction of ozone with UDMH,
acetone dimethylhydrazone, 2-furaldehyde dimethylhydrazone, daminozide, tetramethyl-4,4'-
(methylenedi-/;-phenylene)disemicarbazide and a toluene-derived dimethyl semicarbazide in
wastewater all formed NDMA at molar yields of greater than 40 percent. Molar yields in
ultrapure water with and without varying levels of bromide were approximately 40-90 percent of
those in wastewater.
Treatment of DMA with chlorine dioxide or hydrogen peroxide has been shown to form NDMA,
but yields from chlorine dioxide were 0.2 percent and NDMA formation was only observed with
hydrogen peroxide at pH values greater than 10 (Andrzejewski and Nawrocki, 2007). Chlorine
dioxide has also been shown to cause nitrosamine formation in the presence of a number of
pharmaceutical and personal care products but at molar yields ranging from 6.9 x 10"6 to 0.055
percent (Zhang et al., 2014). Exposure of DMA to permanganate can lead to NDMA formation,
but only at permanganate doses much greater than those typically used for water treatment
(Andrzejewski and Nawrocki, 2009). Gan et al. (2015) found that daminozide formed NDMA at
molar yields of up to 5.01 percent in the presence of chlorine dioxide. The proposed formation
mechanism involved oxidation of the compound to form a UDMH intermediate.
Pozzi et al. (2011) observed NDBA, NDEA, NDPA and NMOR concentrations of up to 11.0
ng/L, 30.7 ng/L, 8.1 ng/L, and 83.7 ng/L, respectively, in finished water gathered from seven
surface water plants and two ground water plants that disinfect with ozone and chlorine dioxide.
Zhao et al. (2008) observed elevated NDMA concentrations (compared to untreated water)
following an AOP (namely, hydrogen peroxide (H202)/UV) in the absence of disinfectant
Six-Year Review 3 6-7
Technical Support Document for Nitrosamines
December 2016
-------
addition in some surface waters. Chlorination following AOP also resulted in increased NDMA
concentrations for some surface waters.
Trogolo et al. (2015) found that ozone could react with A'A-dimethylsulfamide, an herbicide
metabolite, to form NDMA. They found that the reaction was effected by hypobromous
acid/hypobromite formed by oxidation of bromide. The mechanism involved oxidation of
bromide to hypobromous acid by ozone followed by bromination of the sulfamoyl nitrogen in the
DMS, deprotonation of the resulting complex, and attack by ozone. Lv et al. (2015) found that
the pharmaceutical chlorpheniramine formed DMA when reacted with ozone. The DMA could
then be oxidized by hydroxyl radicals formed by ozone to form NDMA. The reaction was
dependent on pH, with the reaction inhibited at lower pH.
While the reactions of DMA with oxidants like ozone have been found to have relatively low
yields, it is possible that with sufficiently high concentrations of precursors significant amounts
of NDMA may be formed. It is also possible that other precursors exist that produce higher
yields.
Soltermann et al. (2013) examined the reaction of chlorinated DMA and chloramine when
irradiated with UV light in pool water. They found the reaction could form NDMA. They found
that at some UV doses, the formation of NDMA from chlorinated DMA and chloramines could
offset the destruction of NDMA by UV light. They proposed a reaction pathway involving the
reaction of nitric oxide (NO) with an aminyl radical. They also found that NDEA and NMOR
could be formed from chlorinated DEA and morpholine, respectively, in the presence of
chloramines and UV light. Experiments with NDEA and NMOR suggest that peroxynitrite may
be responsible for some nitrosamine formation under UV light.
Some studies have found oxidation to decrease nitrosamine formation through the oxidation of
precursors to varying degrees (Wilczak et al., 2003; Lee et al., 2007a, 2008; Chen and Valentine,
2008; Mitch et al., 2009; Russell et al., 2012). While these results may seem contradictory, it
should be noted that conditions and experimental setups varied from study to study. Many of the
studies finding NDMA formation through oxidation have only looked at oxidation of DMA as a
model precursor. On the other hand, many of the precursor oxidation studies used more complex
waters with natural precursors or mixtures of precursors. One possible explanation is suggested
by Lee et al. (2007b). They found that while chlorine dioxide reduced NDMA FP, the reduction
in NDMA formation was not always as large as would be suggested by the disappearance of the
NDMA precursor. The study found that oxidation of precursors such as dimethylaminobenzene
resulted in production of DMA, which reacted only slowly with chlorine dioxide (Lee et al.,
2007b). It seems likely that the reactions of strong oxidants with NDMA precursors in natural
waters is a complex reaction that yields differing results depending on the nature and
concentration of precursors and other water quality parameters. Continued research into the
effects of oxidants on NDMA formation may shed more light on the reactions of NDMA
precursors and the formation of NDMA.
6.4 Precursors: Sources and Characterization
This section focuses on the precursors specifically associated with nitrosamine formation. For a
characterization of precursors and their occurrence for a wider range of DBPs, see the Agency's
Six-Year Review 3 Technical Support Document for Disinfectants and Disinfection By-Products
Rules (USEPA, 2016a).
Six-Year Review 3 6-8
Technical Support Document for Nitrosamines
December 2016
-------
NDMA precursors are distinct from the precursors that form chlorinated DBPs, such as THMs
and HAAs, which are regulated under the Stage 1 and Stage 2 D/DBPRs. While naturally
occurring humic substances are widely accepted to be the primary precursors for the formation of
THMs and HAAs, they are not substantial precursors of NDMA (Gerecke and Sedlak, 2003;
Mitch and Sedlak, 2004). Nitrosamine precursors may be naturally occurring or anthropogenic in
origin. Potential NDMA precursors include DMA, organic nitrogen from natural organic matter
(NOM) (Gerecke and Sedlak, 2003; Mitch and Sedlak, 2004; Chen and Valentine, 2007; Dotson
et al., 2009), tertiary and quaternary amines (Mitch et al., 2003a; Mitch and Schreiber, 2008;
Kemper et al., 2010), cationic flocculants (Wilczak et al., 2003; Valentine et al., 2005), and
anionic exchange resins (Najm and Trussell, 2001; Kemper et al., 2010).
Under UCMR 2, NDMA, NMEA and NPYR were detected primarily in surface water PWSs,
NDBA was detected in ground water public water systems (PWSs) but not surface water PWSs,
and NDEA was detected in surface and ground water PWSs at similar frequencies. (NDPA was
not detected in any samples.) Although these observations are based on a very limited number of
detections in the case of several nitrosamines, they suggest that precursors of different
nitrosamine compounds may be associated with different types of source waters. For NDMA,
potential precursor compounds are typically associated with waste-impacted waters, and
evidence suggests that anthropogenic materials have a significant role in formation, though
naturally occurring precursors may also contribute to formation.
6.4.1 Natural Precursors
DMA, which has been used as a model precursor in the majority of laboratory experiments on
NDMA, was initially thought to be a significant precursor of NDMA. DMA is a component of
biological waste of both animal and human origin. Nawrocki and Andrzejewski (2011) report
that few studies describe the occurrence of DMA in natural waters. Available occurrence data
indicate concentrations from 3 |ig/L to over 200 |ig/L (Sacher et al., 1997; Zhao et al., 2003).
Waters affected by wastewater effluent have higher concentrations of DMA than more pristine
waters. Studies by Gerecke and Sedlak (2003) and Mitch and Sedlak (2004) showed that 70
percent of the dissolved NDMA precursors in primary wastewater effluent could be accounted
for by DMA; however, DMA contributed to only 14 percent of the dissolved precursors in
secondary effluent and less than 25 percent of precursors in non-impacted waters. Moreover,
cellular and biological constituents do not serve as significant NDMA precursors (Gerecke and
Sedlak, 2003; Mitch et al., 2003a).
The yields of nitrosamines from DMA and other secondary amines vary, but are relatively low.
Zhou et al. (2014) looked at yields of nitrosamines from the reaction of the secondary amine and
chloramine. They found yields ranging from 0.32 to 2.32 percent. They found that the yield was
the highest for NPYR, followed by A'-nitrosodi phenyl amine (NDPhA), NDMA, NPIP, NMEA,
NMOR, NDEA, NDPA and NDBA.
A number of researchers have shown that chloramination of NOM may result in the formation of
nitrosamines, though not to levels exceeding 10 ng/L (Gerecke and Sedlak, 2003; Chen and
Valentine 2006, 2007). For Iowa River water, the hydrophilic fractions of NOM tended to form
more NDMA than hydrophobic fractions, and basic fractions tended to form more NDMA than
acid fractions when normalized to a carbon basis (Chen and Valentine, 2006). Several studies
found that NDMA precursors are derived primarily from predominantly non-polar molecules that
also contain cationic functional groups (Liao et al., 2015a; Chen et al., 2014). Dotson et al.
Six-Year Review 3 6-9
Technical Support Document for Nitrosamines
December 2016
-------
(2009) reported that nitrosamine precursors were found primarily in the nitrogen-rich colloidal,
hydrophilic neutral and hydrophilic base fractions of organic matter. Chuang et al. (2013) found
NDMA formation to be correlated with (DON) compounds in the hydrophilic and transphilic
fractions of NOM. (The transphilic portion is the portion with a polarity between the
hydrophobic and hydrophilic portions.) Selbes et al. (2013) also found that the transphilic portion
of NOM produced more NDMA than the hydrophobic portion, although when specific
precursors were spiked into the different portions of NOM, in most cases molar yields were
higher in the hydrophobic portions.
DON is a component of NOM and consists of organic nitrogen-containing compounds that pass
through a filter. Total dissolved nitrogen includes DON along with ammonia, nitrate, nitrite and
other inorganic nitrogen compounds. Natural DON sources include autochthonous organic
matter and soluble microbial products. The contribution of DON to total dissolved nitrogen
varies widely in surface waters (27 to 91 percent), as reported by Westerhoff and Mash (2002).
Algally influenced waters have been associated with elevated nitrosamine formation, though, in
comparison, wastewater-impacted source waters are generally more prone to NDMA formation
(Shah and Mitch, 2012). Wang et al. (2015a) found that ash from forest fires formed more than
twice as much NDMA following chloramination as unburnt natural organic matter.
6.4.2 Anthropogenic Precursors
A variety of anthropogenic compounds, including wastewater-derived chemicals in source water
and chemicals added during drinking water treatment, can serve as precursors to nitrosamines.
6.4.2.1 Wastewater Flows
Several studies have demonstrated the accumulation of nitrosamine precursors in surface water, a
phenomenon which has been attributed to the addition of wastewater effluent flows (Schreiber
and Mitch, 2006b; Charrois et al., 2007; Sacher et al., 2008). FP tests of ground water, lakes and
reservoirs indicate maximum NDMA precursor concentrations of 58 ng/L, while wastewaters
were shown to contain up to 1,300 ng/L of nitrosamine precursors (Gerecke and Sedlak, 2003;
Pehlivanoglu-Mantas and Sedlak, 2006). Krasner et al. (2015) found a correlation between
sucralose, which is considered an indicator of wastewater presence, and NDMA FP in some
watersheds. Similarly, Uzun et al. (2015) found that in free-flowing rivers in the southeastern
United States, NDMA FP depended on the ratio of wastewater effluent to river flow. The impact
of wastewater was often found to increase as river flows decreased. Lee et al. (2015) sampled
two wastewater treatment plant effluents as well as upstream and downstream of the two plants
in the Sacramento-San Joaquin Delta. They found median increases in NDMA FP (from the
upstream to downstream locations) of 5 to 17 ng/L at one plant and 16 to 40 ng/L at the other
plant. NDMA and NPYR precursors were found in the plant effluent, as well as preformed
NMOR. No nitrosamines were found in the river water, possibly because of dilution or
photolysis.
High concentrations of NDMA precursors in municipal wastewater effluent may be explained by
the contributions of DMA and related compounds from industrial, domestic and agricultural
sources. Both direct waste flows and surface runoff may contribute to the precursor materials
observed in wastewater flows. In addition to precursors, wastewater effluent may contain
nitrosamines themselves. Nitrosamine concentrations in treated wastewater are affected by
wastewater treatment processes, including nitrification and disinfection (Krasner et al., 2009).
Six-Year Review 3 6-10
Technical Support Document for Nitrosamines
December 2016
-------
Nitrosamines in wastewater effluents are a matter of concern for drinking water PWSs, as
nitrosamines have been found in drinking water influents located downstream from wastewater
plant discharge points (Krasner et al., 2005; Drewes et al., 2006).
Several researchers studied the effects of wastewater treatment on NDMA formation in
wastewater. Qi et al. (2014) found between 300 and 600 ng/L of NDMA in secondary effluent of
a Chinese wastewater plant; no advanced disinfection was in place at that plant. Sgroi et al.
(2015) found that chlorinating wastewater from a wastewater treatment plant used for indirect
potable use formed up to 248 ng/L NDMA. While reverse osmosis (RO) and UV treatment
processes reduced the concentrations of NDMA, up to 16 ng/L was still observed in the plant
effluent. This may have been due to the presence of small nonpolar molecules such as
dimethylformamide, which can pass through RO membranes and form NDMA upon reaction
with chloramines formed by reaction of chlorine with ammonia in the wastewater. Sgroi et al.
(2015) also demonstrated that ozonating wastewater containing chloramines formed more
NDMA than chloramines alone. Similarly, Gerrity et al. (2015) found both NDMA and NMOR
at concentrations up to 89 ng/L for NDMA and 67 ng/L for NMOR in primary wastewater
effluent. They also found ozone-induced formation of nitrosamines ranging from 10 to 143 ng/L
in ten of eleven wastewater treatment plants examined, although biological filtration and/or UV
after ozone was able to control the formed NDMA. Kosaka et al. (2014) found an NDMA
precursor, 1,1,5,5-tetramethylcarbohydrazide, in industrial effluent-containing sewage in the
Yodo River basin in Japan. The precursor reacted with ozone to form NDMA with a molecular
yield of 140 percent. This precursor was found to account for between 42 and 72 percent of all
NDMA in the sewage treatment plant effluent.
Several studies investigated the sources of nitrosamine precursors. Zeng and Mitch (2015)
examined NDMA formation from chloramination or ozonation of various wastewater sources
including shower, sink, toilet and laundry. They found that the laundry contributed 58 percent of
NDMA precursors in chloraminated water and 99 percent of NDMA precursors in ozonated
water. The shower was also a significant source of NDMA precursors in chloraminated water.
Use of ranitidine pharmaceuticals significantly increased NDMA FP of urine although not
enough to become a significant source of NDMA compared to other streams when volume is
taken into account. Yoon and Tanaka (2014) performed laboratory experiments adding amines
from pollutant inventories to wastewater samples from a Japanese wastewater plant. Samples
were reacted with chloramine and ozone and nitrosamine formation was measured. All of the
secondary and tertiary amines tested and many of the primary amines reacted with chloramine to
form NDMA. Ozone produced NDMA from 1,1 dimethylhydrazine, formed NDEA from p-
chloroaniline, and formed NMOR and NDBA from primary amines. Qi et al. (2014) found that
NDMA FP in secondary effluents correlated best to low molecular weight DON fractions. Ma et
al. (2015) examined nitrosamine formation from various wastewater microbial products. They
found that microbial products associated with microbial utilization or decay formed twice as
much NDMA as products associated with biomass or microbial growth.
Industrial Effluents
Though some secondary amines are found in the natural environment, many may also be
industrial in origin. For example, DEA is used in the production of rubber, textiles, resins, dyes
and insecticides and as a flame retardant. DP A, piperidine (precursor to NPIP), and pyrrolidine
(precursor to NPYR) are used in the production of dyes, pesticides, lacquers, and rubber
manufacturing.
Six-Year Review 3 6-11
Technical Support Document for Nitrosamines
December 2016
-------
Industrial effluents may be a significant source of NDMA precursors. Industrial discharges have
been shown to contain up to 82,000 ng/L of NDMA precursors (Deeb et al., 2006). Yoon and
Tanaka (2014) found that 15 amines that are common industrial emissions were precursors for
nitrosamines when oxidized in a Japanese wastewater using either ozone or chloramines.
Parker et al. (2014) investigated the influence of hydraulic fracturing wastewater on NDMA
formation. They diluted wastewater from a domestic wastewater treatment plant and from three
hydraulic fracturing wastewaters with river water from Ohio and Pennsylvania and measured the
formation of nitrosamines. Without either domestic wastewater or water from hydraulic
fracturing wastewater, no NDMA was formed. Addition of 10 percent domestic wastewater led
to formation of about 7 ng/L of NDMA. Adding just 0.1 percent of hydraulic fracturing
wastewater led to a 50 percent increase in NDMA formation. NDMA formation was higher in
waters with higher iodide levels.
Pharmaceuticals and Personal Care Products
Recent studies have shown that a number of pharmaceuticals and personal care products contain
nitrosamine precursors. The pharmaceutical ranitidine (Zantac) has been shown to convert to
NDMA at high yields (over 60 percent) under chloramination (Schmidt et al., 2006; Sacher et al.,
2008; Le Roux et al., 2011; Shen and Andrews, 201 la, 201 lb). Shen and Andrews (201 la)
found that 20 different pharmaceuticals (including ranitidine) tested formed nitrosamines during
chloramination, with eight pharmaceuticals demonstrating NDMA yields greater than 1 percent.
Of the eight, ranitidine was the highest with 89.9 to 94.2 percent yield. The other seven all had
single digit percentage yields.
Nitrosamine formation has also been associated with quaternary amines, which are significant
constituents of consumer products, including shampoos, detergents and fabric softeners (Kemper
et al., 2010). Kemper et al. (2010) chloraminated a range of model compounds and personal care
products to determine their NDMA FP and found that polymeric and benzylated quaternary
amines were stronger precursors than monomeric quaternary alkylamines. An NDMA FP test of
Suave® Shampoo resulted in 0.00005 mass yield of NDMA while cocamidopropyl betaine, an
ingredient found in Suave® Shampoo, resulted in 0.16 percent molar yield of NDMA. Dawn®
detergent exhibited an NDMA FP 25 times higher than Suave® Shampoo (Kemper et al., 2010).
Kemper et al. (2010) showed that the NDMA yields associated with personal care products
indicate that those products could account for 2 to 3 percent of the total NDMA precursor
loading at a wastewater treatment facility. Benzalkonium chloride, a quaternary amine surfactant
found in antimicrobial soaps, gave a 2.1 percent molar yield of NDMA after 10 days of contact
with a high dose of chloramines at pH 7 (Mitch et al., 2009). A likely tertiary amine breakdown
product, DMBzA showed a 300 times higher yield under the same conditions (Mitch et al.,
2009).
Amine-based pharmaceutical and personal care products have also been shown to form
nitrosamines upon disinfection using chlorine or chlorine dioxide, although at much lower molar
yields (Zhang et al., 2014). For example, molar yields for ranitidine with chlorine or chlorine
dioxide were 0.050 and 0.055 percent, respectively, compared to 40.2-90.6 percent for
chloramine. Ranitidine had the highest yield of ten pharmaceutical and personal care products
examined (Zhang et al., 2014).
Six-Year Review 3 6-12
Technical Support Document for Nitrosamines
December 2016
-------
Pesticides and Herbicides
A number of pesticides and herbicides have also been shown to be NDMA precursors, with some
compounds resulting in the formation of NDMA at relatively high yields. DMS, a degradate of
the pesticide tolylfluanid, formed NDMA at 52 percent molar yield upon application of 6 mg/L
ozone. Because DMS is not removed by riverbank filtration, flocculation or activated carbon, it
can contribute significantly to the formation of NDMA in drinking waters. Depending on the
treatment processes and degree of contamination at drinking water facilities, 73 to 100 percent of
DMS can be transformed, resulting in NDMA concentrations ranging from 1.9 to 310 ng/L
(Schmidt and Brauch, 2008).
Other pesticides have been shown to form NDMA upon addition of nitrite, ammonia, and free
chlorine, including thiram, sodium dimethyldithiocarbamate, l,l-dimethyl-3-(4-methoxyphenyl)-
2-thiourea, and l,l,3,3-tetramethyl-2-thiourea. These chemicals produced 5 to 30 times the
NDMA formed from DMA under the same conditions (Valentine et al., 2005). Application of 6
mg/L of ozone for 30 minutes to 2 [j,g/L of the pesticide daminozide led to an 80 percent molar
yield of NDMA (Sedlak et al., 2005). Twenty-two micrograms per liter (22,000 ng/L) of NDMA
formed when 0.086 mM of the pesticide diuron was added to a well-mixed solution of 4.1 mM
ammonia and 3.45 mM free chlorine, while the application of free chlorine alone formed
15.8 [j,g/L (15,800 ng/L) of NDMA. However, diuron is not expected to be a significant source of
NDMA because at reactant concentrations more representative of water treatment operations,
NDMA was not detected upon chloramination of diuron (Chen and Young, 2008, 2009). At the
same time, this formation pathway is important because it can proceed without the addition of
another nitrogen source such as ammonium or nitrate (Chen and Young, 2009). Chen et al.
(2015a) found that chlorine or chloramine could react with five different phenylurea-based
herbicides to produce NDMA or NPYR at molar yields between 0.003 and 0.99 percent.
Dithiocarbamates have been found to form nitrosamines when exposed to chloramine, ozone,
chlorine, or chlorine dioxide (Padhye et al., 2013). Fifty [xM pesticide with 100 [xM disinfectant
yielded between 118 and 302 ng/L nitrosamines. Chloramine and ozone formed more
nitrosamines from dithiocarbamate pesticides than did chlorine or chlorine dioxide. Hydrolysis
of the pesticides to form amines, followed by oxidation of the amines by the disinfectant, likely
contributed to nitrosamine formation.
During the application of fumigants containing dimethyldithiocarbamates for root control to
sewer trunklines in residential and industrial areas, concentrations of NDMA were found to be
2,400 ng/L, and the concentration of NDMA precursors was found to be 89,000 ng/L in the
collection area (Deeb et al., 2006). During the application of metam sodium, a dithiocarbamate
used for root control, NDMA concentrations greater than 2,000 ng/L were detected in residential
trunklines (Sedlak et al., 2005).
Dyes
Ozonation of dyes such as methylene blue, methyl orange, methyl violet B, Auramine, brilliant
green, A', A-di methylaminobenzene and A', A'-dimethyl-p-phenylenediamine yielded NDMA
ranging from 0.0001 to 0.043 percent upon treatment with ozone at a flow rate of 1 mL/min for
15 min at pH 7 in deionized water (Oya et al., 2008).
Anti-yellowing agents 4,4'-hexamethylenebis(l,l-dimethylsemicarbazide) and 1,1,1', 1
tetramethyl-4,4'-(methylenedi-p-phenylene)disemicarbazide contributed to 1.4 percent of the
Six-Year Review 3 6-13
Technical Support Document for Nitrosamines
December 2016
-------
NDMA formation formed from ozonation of a secondary wastewater effluent (Kosaka et al.,
2009).
Nanomaterials
Verdugo et al. (2014) found that carbon nanotubes could act as both a source of NDMA
contamination and a precursor to NDMA formation. They found that carbon nanotube powders
with amine, amide, or other nitrogen-containing polymer groups leached up to 50 ng of NDMA
per mg of nanotube powder before any reactions occurred. These nanotube powders formed
additional NDMA upon disinfection with chlorine, monochloramine, or ozone. Formation yields
of NDMA were approximately equivalent to yields from reactions of chloramine with NOM.
6.4.2.2 Treatment Additions
Polymer Addition
In addition to precursor material from source waters, NDMA formation has also been associated
with flocculation polymers and ion exchange resins applied during the treatment process. A
survey conducted in the United Kingdom indicates that ferric coagulant may be a source of
NDMA (DEFRA, 2008). Numerous studies have shown that cationic polymers, including
polyDADMAC and epichlorohydrin-dimethylamine (epi-DMA), may be sources of NDMA
precursors (Kohut and Andrews, 2003; Valentine et al., 2005; Wilczak et al. 2003; Park et al.,
2009).
PolyDADMAC is a cationic polymer widely used for primary coagulation (Montgomery, 1985).
Several studies have shown correlations between polyDADMAC dose and NDMA formation in
pilot- and full-scale water treatment plants using chloramines, indicating that polyDADMAC
may be a source of NDMA precursors (Wilczak et al., 2003; Valentine et al., 2005; Mitch et al.,
2009). Wilczak et al. (2003) observed enhanced NDMA formation when polyDADMAC was
added prior to chloramination, but not for chlorination. For chloramination, the order of reagent
addition was observed to be important, with the highest NDMA formation observed when
polymer was added to the test water and followed immediately by chloramine formation. The
authors suggest that NDMA formation could be reduced if chloramination occurred after
filtration to allow for polymer removal. They also suggested allowing for a short free chlorine
contact time before ammonia addition to allow for precursor oxidation to reduce NDMA
formation. Mitch et al. (2009) also found that effluent NDMA concentrations at chloramine
treatment plants generally increased with increased polymer dose.
Valentine et al. (2005) found a linear relationship between polyDADMAC dose and NDMA
formation following chlorination, though NDMA concentrations were much less than those
reported by Wilczak et al. (2003) for chloramination. This finding suggests that chlorine may be
more efficient at breaking polymer bonds. However, Valentine et al. (2005) observed greater
NDMA formation when polyDADMAC reacted with chlorine than with chloramine, indicating
that other water quality parameters, such as the presence of nitrite, may impact which
disinfectant contributes to NDMA formation. Experiments showed that polymer dose and the
interaction of nitrite, polymer and chlorine were important to NDMA formation when
polyDADMAC interacted with chlorine, while only polymer dose was found to be important
when polyDADMAC reacted with chloramine (Valentine et al., 2005).
Six-Year Review 3 6-14
Technical Support Document for Nitrosamines
December 2016
-------
Padhye (2010) found in laboratory experiments that doses of 5-10 mg/L of polyDADMAC could
yield up to 800 ng/L of NDMA upon ozonation of the polymer. The NDMA appeared to be
formed as a result of oxidation of the polymer to release DMA, followed by oxidation of the
DMA. Padhye (2010) also found that NDMA formation upon ozonation of polyDADMAC
increased with higher pH.
Park et al. (2014) found that NDMA formation for polyDADMAC was the greatest when the
chlorine-to-ammonia ratio was close to the breakpoint. Later experiments found that a chlorine-
to-ammonia ratio of 1.4 formed the most from polyDADMAC polymers (Park et al., 2015). They
also found that adding chloramine followed by free chlorine gave the highest NDMA formation.
They found that chloramines were the mostly likely oxidant to form NDMA, followed by
chlorine, ozone and chlorine dioxide. Pre-oxidation in some cases was able to lower NDMA
formation by altering the polymer structure to make it less susceptible to reaction. With pre-
oxidation using ozone, however, NDMA formation increased, possibly due to release and
oxidation of DMA.
Other polymers, including polyamine, epi-DMA and epi-DMA with ethylenediamine (epi-DMA-
co-ED), also showed elevated NDMA formation following reactions with chlorine and
chloramines (Valentine et al., 2005; Park et al., 2015). Park et al. (2015) found that a chlorine-to-
ammonia ratio of 1.8 maximized NDMA formation from polyamine polymers. Chlorination of
epi-DMA has been shown to result in greater formation of NDMA than polyDADMAC
(Valentine et al., 2005). NDMA has also been detected during application of free chlorine or
chloramine to polyamines or when tap water containing polyamine was amended with nitrite
(Bolto, 2005; Kohut and Andrews, 2003; Park et al., 2014). No difference in NDMA formation
was seen between chlorine and chloramine reactions with polyamines (Valentine et al., 2005).
Through a series of purification experiments, Park et al. (2009) showed that the polymer itself,
and not any impurities, is responsible for NDMA formation upon chloramination of
polyDADMAC and polyamine polymers. Though free DMA may exist as an impurity in
polymer stocks, Park et al. (2009) found that NDMA formation is more dependent on DMA
released by polymer degradation following reactions with chloramine than by residual DMA
present in the polymer. Ozonation of polyDADMAC has been shown to result in substantial
release of DMA, such that increased NDMA formation may occur if chloramines are added
subsequent to the ozonation process (Huang et al., 2011). Valentine et al. (2005) examined an
experimental version of polyDADMAC reported by the manufacturer to have 20 times less free
DMA than conventional polyDADMAC polymers. Upon chlorination the new polymer did show
lower NDMA production, but the production was about a third of the original polymer, not a
twentieth, as might be expected if free DMA impurities were largely responsible for NDMA
formation (Valentine et al., 2005).
The age of polyDADMAC stocks does not appear to have a substantial impact on NDMA
formation (Kohut and Andrews, 2003; Valentine et al., 2005). Valentine et al. (2005) reported
that the age of polyDADMAC-prepared stock had a significant effect on NDMA over the 50-
hour tested interval; however, the authors concluded that NDMA yields were most likely not
affected by the age of polyDADMAC-prepared stock within the range applicable to coagulant
use, because the stock is used in 48 hours or less. NDMA yields for epi-DMA did not increase
over the entire 50-hour test interval, but did increase within the first five hours of aging, though
NDMA formation reached a plateau as the age further increased (Kohut and Andrews, 2003;
Valentine et al., 2005).
Six-Year Review 3 6-15
Technical Support Document for Nitrosamines
December 2016
-------
In a survey of 100 utilities (including 88 from UCMR 2 and 12 from the Ontario Drinking Water
Surveillance Program), no significant difference was observed in NDMA concentrations between
utilities that used cationic polymers and those that did not (Russell et al., 2012). However, a
smaller study found that among plants employing coagulation there was a 43-82 percent increase
in NDMA in plants using cationic polymers, compared to no increase in a plant not using
cationic polymer (Krasner et al. 2012b). These findings suggest that multiple factors may
influence the possibility of higher NDMA formation from cationic polymers. The results of the
survey by Russell et al. (2012) suggest that the use of chloramines was a more important factor
than polymer use in NDMA formation.
Sgroi et al. (2014) found that polyacrylamide polymers used to treat sludge at wastewater plants
formed NDMA upon reaction with ozone. The polymer entered the wastewater plant through the
recycle stream from the sludge treatment process and was the single largest source of NDMA
precursors in the wastewater treatment plant.
Anion Exchange Resins
Anion exchange resins, which contain quaternary amines, have been found to release NDMA
directly, as well as to release amines that can form NDMA upon reaction with chloramines
(Kemper et al., 2009). Najm and Trussell (2001) examined four strong base anion exchange
resins for NDMA formation. The resins were based on dimethyl-ethanol, trimethyl, triethyl, and
tripropyl functional groups. They soaked the resins in three different unchlorinated waters:
deionized water, ground water and buffered distilled water. They found that the dimethyl-ethanol
and trimethyl functional groups both produced NDMA, with the dimethyl-ethanol quarternary
amine forming the most NDMA. They postulated the triethyl and tripropyl functional groups
may have formed NDEA and NDPA, respectively, but did not test this hypothesis (Najm and
Trussell, 2001).
Kemper et al. (2009) conducted a series of column studies to examine the nitrosamine and
nitrosamine precursors associated with two types of anion exchange resins. In the absence of any
disinfectant, the authors found that NDMA was released by fresh alkylamine (trimethylamine
(TMA) and tributylamine) and dialkylethanolamine anion exchange resins in the range of 2 to 10
ng/L and at up to 20 ng/L following regeneration. In the presence of feedwaters containing 2
mg/L free chlorine or 2 mg/L monochloramine, NDMA concentrations and NDMA precursor
concentrations increased. For alkylamine resins, the NDMA concentration was approximately
300 ng/L when chlorine was applied upstream, and ranged from 10 to 100 ng/L when
monochloramine was applied upstream. For the tributylamine resin, NDBA concentrations were
less than 60 ng/L for both chlorine and monochloramine feedwaters. For the TMA resin, NDMA
precursor concentrations initially spiked to approximately 16,000 and 20,000 ng/L in the
presence of chlorine and chloramine feedwaters, respectively, but then rapidly declined. For the
tributylamine resin, NDMA precursor concentrations were generally low, though NDBA
precursors were observed at levels of 1,700 ng/L and 6,000 ng/L for chlorine and
monochloramine, respectively, before rapidly declining to less than 500 ng/L. For the
dialkylethanolamine resin, NDMA concentrations were similar for feedwaters containing free
chlorine and monochloramine and ranged from 200 to 800 ng/L. NDMA precursor
concentrations were less than 500 ng/L for chloramines and ranged from 1,500 to 3,500 ng/L for
feedwaters containing free chlorine. Nitrosamine concentrations and precursor concentrations
diminished after multiple regeneration cycles, indicating that releases may eventually subside
(Kemper et al., 2009).
Six-Year Review 3 6-16
Technical Support Document for Nitrosamines
December 2016
-------
Flowers and Singer (2013) examined 21 different resins containing trialkylamine or
dimethylethanolamine groups using both batch desorption and laboratory column flow-through
experiments. They found that six resins desorbed nitrosamines in batch experiments where
washed resins were exposed to clean water for an hour. The resins that released nitrosamines
included A530E, A600E, CalRes 2103, PWA2, PWA5 and PWA15. A530E contained
triethylamine and tributylamine groups, A600E contained trimethylamine, CalRes 2103
contained tripropylamine, PWA2 contained tributylamine, PWA5 contained triethylamine and
PWA15 contained trimethylamine. Released nitrosamines included NDMA (from A530E,
A600E, PWA2 and PWA15, at concentrations less than 10 ng/L), NDPA (from CalRes 2103 at a
concentration of about 10 ng/L), NDEA (from PWA15 at a concentration of about 15 ng/L and
A530E at a concentration of about 110 ng/L) and NDBA (from A530E at a concentration of 974
ng/L and PWA2 at a concentration of 592 ng/L). A flow-through column experiment with resin
A300E released 223 ng/L NDMA in the first 10 bed volumes; that level dropped to 23 ng/L
NDMA after 50 bed volumes. It took 240 bed volumes before the NDMA dropped below
detection levels. The experiments with resin A300E also found the release of 1,402 ng/L of
nitrosamine precursors in the initial 10 bed volumes, dropping down to 34 ng/L after 50 bed
volumes. Regeneration of the column led to a temporary increase of 50 ng/L in precursor
concentrations. When similar column experiments were performed with other resins, seven of the
15 resins were found to release quantifiable levels of nitrosamines, which declined over time,
generally washing away after 100 bed volumes. Introducing 0.24 mg/L chlorine or chloramine to
the feedwater increased the concentration of released nitrosamines, with chloramines producing
more nitrosamines than chlorine. For example, resin PWA2 produced no detectable NDMA with
chlorine, but 1,240 ng/L of NDMA with chloramines.
Magnetic ion exchange resin (MIEX) has also been found to form nitrosamines, especially when
used in wastewater with chloramine exposure. Gan et al. (2013) found MIEX resin formed 36
ng/L NDMA when used to treat wastewater in combination with chloramines. If chlorine was
used, NDMA formation was only 10 ng/L. If MIEX alone was used to treat drinking water,
NDMA formation was 5 ng/L.
Use of Rubber Distribution System Components
Studies have also found that NDMA in drinking water may originate from rubber gaskets or
sealing rings within the distribution system (Morran et al., 2011; Teefy et al., 2011). Morran et
al. (2011) did not observe a trend between the amount of NDMA released and sealing ring size
or age from bench-scale studies, and they suggested that the variability may be due to differences
in the manufacturing process and/or in the leaching rates. During a distribution system extension
project in South Australia, elevated NDMA concentrations were observed in a new polyvinyl
chloride pipeline (Morran et al., 2011). NDMA concentrations in the new pipeline exceeded 100
ng/L, in comparison to concentrations of 9 to 34 ng/L for an existing concrete-lined pipe
distributing water from the same source. Laboratory tests revealed that sealing gaskets used in
the pipe were a substantial source of NDMA (-40 ng/L), while neither polyvinyl chloride nor
lubricant released any measureable NDMA (Morran et al., 2011).
Though no changes had occurred in the water treatment process in a California plant, an elevated
NDMA concentration (21 ng/L) was observed in the distribution system downstream from a
temporary storage tank that contained rubber gaskets (Teefy et al., 2011). A bench-scale study
found that NDBA, NDMA and NPIP were present in test water following the exposure of the
gaskets to chloraminated system water for periods of 2 and 14 days. Concentrations of NDMA
Six-Year Review 3 6-17
Technical Support Document for Nitrosamines
December 2016
-------
and NPIP increased over time, while NDBA concentrations were similar for both test periods.
New gasket material released higher concentrations of the measured nitrosamines than gaskets
that had been in use. A comparison study conducted with water in the absence of chloramines
resulted in NDMA concentrations similar to those measured for the chloraminated water,
indicating that the NDMA leached directly from the gaskets, and did not form due to a reaction
with chloramines (Teefy et al., 2011). A follow-up study (Teefy et al., 2014) found that the
nitrosamines were present whether or not chloramines were used and that NDBA and NPIP were
more prevalent in the newer gasket material, while NDMA was prevalent in the used gasket
material. The results of this and other studies discussed previously show that NDMA levels may
increase within the distribution system because of leaching of NDMA from rubber distribution
system components, regardless of the disinfectant used at the treatment facility.
6.4.3 Precursor Characterization
Although specific precursors have not yet been identified, work with model precursors has
shown that tertiary amines with DMA functional groups form NDMA at higher yields than
amides with DMA functional groups (Mitch and Sedlak, 2004; Chen and Young, 2008). Amine-
based nitrosamine precursors tend to be found in low molecular weight hydrophilic neutral or
base fractions, which are poorly removed during conventional water treatment processes
compared to the hydrophobic fractions that contain THM precursors (Dotson and Westerhoff,
2009, Chuang et al., 2013, Wang et al., 2013a). Amines are more reactive with oxidants, such as
free chlorine, than are other major nitrogenous functional groups (Hawkins et al., 2003; Shah and
Mitch, 2012).
Studies on the relationships between NDMA formation and bulk water quality parameters, such
as DOC concentration and DON concentration, have shown inconsistent results, with
relationships appearing to be water-specific. Ultraviolet absorbance (UVA) and fluorescence
have also been studied as possible indicators.
One of the challenges of determining broadly applicable relationships for NDMA formation is
that most studies aggregate data from different water types exhibiting different chemistry and
NOM characteristics. For example, organic matter from algal and other sources may confound
establishment of relationships between NDMA and organic matter of waste origin (Mitch et al.,
2009).
Other properties in addition to DOC, DON and UVA have also been considered. Upon
monitoring total organic carbon (TOC), pH, turbidity and color, Zhao et al. (2008) found no
relationship between these factors and NDMA formation in surface waters. Yang et al. (2015)
used fluorescence excitation emission to correlate NDMA formation to protein-like organic
components.
6.4.3.1 Dissolved Organic Carbon (DOC) Concentration
Though DOC concentration has been used as a surrogate measure for the formation of other
DBPs, the relationship between DOC concentration and nitrosamine formation is unclear.
Gerecke and Sedlak (2003) found a strong linear relationship between DOC concentration and
NDMA FP for Suwanee River NOM (r2 = 0.98), but found only a weak correlation for other
surface waters (r2 = 0.41). Krasner et al. (2008) found a correlation between DOC and NDMA
FP for a wastewater-impacted river. Zhao et al. (2008) did not find a significant relationship
Six-Year Review 3 6-18
Technical Support Document for Nitrosamines
December 2016
-------
between TOC (of which DOC is one component) and NDMA FP in a study of seven surface
waters. Li et al. (2015) reported a linear correlation between DOC and total nitrosamine
concentrations in raw water; however, similar relationships were not established between DOC
and nitrosamine concentrations in finished water or distribution system water.
6.4.3.2 Dissolved Organic Nitrogen (DON) Concentration
Lee et al. (2007b) hypothesized that NDMA formation would increase for organic matter that
was enriched in organic nitrogen. The authors found that NDMA formation increased as the
DOC-to-DON ratio decreased (i.e., increasing nitrogen content of dissolved organic matter, or
DOM). The authors also found that NDMA formation increased as the amino sugar-to-aromatic
ratio of DOM increased.
Drinking water treatment plants with source waters impacted by algae and/or wastewater have
higher DON concentrations than treatment plants with more pristine source waters (Mitch et al.,
2009). Higher DON concentrations observed in algal and wastewater-influenced waters have
also been associated with higher nitrosamine formation (Schreiber and Mitch, 2006b; Krasner et
al., 2008; Chen et al., 2009). However, consistent relationships between DON concentration and
nitrosamine formation have not been determined. The lack of clear relationship is highlighted by
two case studies for wastewater-impacted rivers presented by Krasner et al. (2008). For the South
Platte River, Colorado, a weak relationship between DON concentration and NDMA FP was
found (r2 = 0.49), while a strong linear relationship (r2 = 0.83) was observed for the Santa Cruz
River, Arizona. Mitch et al. (2009) found no correlation between DON concentration and
NDMA FP from wastewater- or algal-impacted raw drinking water. Xu et al. (2011) showed that
DON concentration correlated with NDMA FP for raw water in Shanghai, China (r2 = 0.77).
Wang et al. (2015b) noted an increase in nitrosamine concentrations in post-disinfection drinking
water upon switching from the Luan River to the Yellow River over an 11-month period. Source
water from the Luan River was used from June to October and the source was switched to the
Yellow River for November to April. The authors cite higher levels of precursors in the Yellow
River as a potential cause and note a correlation between DON and the formation of
nitrosamines; however, the effect of the lower water temperatures that would have been present
when the Yellow River was used as source water was not discussed at length. As discussed in
Section 6.5.2, lower water temperatures may be associated with increases in the concentration of
some nitrosamines and decreases in the concentration of other nitrosamines.
6.4.3.3 Ultraviolet Absorbance (UVA)
UVA and specific UVA (SUVA), which is the UV absorbance divided by the DOC
concentration, have both been associated with the formation of other DBPs, though their use as
surrogate measures for nitrosamine formation is less certain. Wastewater, which has been shown
to be rich in nitrosamine precursors, generally has low SUVA values (Krasner et al., 2008). Chen
and Valentine (2007) found an inverse linear relationship with NDMA FP and SUVA at 272 nm
for Iowa River water. However, the authors did not observe a relationship between NDMA FP
and SUVA at 254 nm or 272 nm for NOM fractions, indicating that SUVA may not be useful as
a universal index for NDMA FP. Zhao et al. (2008) found no significant relationship between
UVA at 254 nm and NDMA FP in a study of seven surface waters. Li et al. (2015) reported a
linear correlation between UVA at 254 nm and 272 nm and total nitrosamine concentrations in
raw water; however, similar relationships were not established between UVA at 254 nm and 272
nm and nitrosamine concentrations in finished water or distribution system water. Negative
Six-Year Review 3 6-19
Technical Support Document for Nitrosamines
December 2016
-------
correlations were observed between SUVA at 254 nm and 272 nm and total nitrosamine
concentrations in raw water.
6.4.3.4 Fluorescence
A number of studies have found relationships between DOM fluorescence and regulated DBP
concentrations (Marhaba and Kochar, 2000; Hua et al., 2007; Beggs et al., 2009; Kraus et al.,
2010). Limited information is available regarding relationships between DOM fluorescence and
nitrosamine formation. Lee et al. (2006) found a correlation between DON concentrations and
protein-like (excitation 270-280 nm; emission 300-350 nm) fluorescence intensities (r2 = 0.71).
Hua et al. (2007) also found protein-like fluorescence (excitation 290-310 nm; emission 330-
350 nm) to be associated with NDMA formation. Though no relationship was observed between
DOC concentration and NDMA FP for a diverse group of 53 Missouri lakes, a correlation for a
subset of the lakes with medium to high fluorescence intensity was observed (r2 = 0.63).
Exhibit 6.3 provides a summary of reported correlations between nitrosamine FP and organic
carbon and organic nitrogen.
Exhibit 6.3: Studies Correlating Organic Carbon and Organic Nitrogen to NDMA
Formation Potential
Study
FP Conditions
Correlation
of NDMA
FP to
Organic
Carbon
Correlation
of NDMA
FP to
Organic
Nitrogen
Water Type
Gerecke and
Sedlak, 2003
2 mM chloramines, pH 7, 25 °C, 10
days
Linear (r2 =
0.41)
N/A
Reservoir water, ground
water, and eutrophic lake
Gerecke and
Sedlak, 2003
2 mM chloramines, pH 7, 10 days
Linear (r2 =
0.98)
N/A
Suwannee River NOM
Hua et al., 2007
1 mM chloramines to 50 ml water
sample in a brown bottle; sample
allowed to react in dark for 7 days
at room temperature
None
N/A
Surface waters
Hua et al., 2007
1 mM chloramines to 50 ml water
sample in a brown bottle; sample
allowed to react in dark for 7 days
at room temperature
Linear (r2 =
0.63)
N/A
Lakes with high protein-
like fluorescence intensity
Lee et al.,
2007b
0.634 mM chloramines per mg
DOC, pH 7, 10 days
*
*
Raw drinking water
Krasner et al.,
2008
Chloramines added to 3x weight of
TOC, pH 8, 25 °C, 3 days
N/A
Linear (r2 =
0.47)
Wastewater effluent,
drinking water influent,
ground water, and
surface water
Krasner et al.,
2008
Chloramines added to 3x weight of
TOC, pH 8, 25 °C, 3 days
Linear (r2 =
0.80)
Linear (r2 =
0.83)
Santa Cruz River,
Arizona (wastewater-
impacted)
Krasner et al.,
2008
Chloramines added to 3x weight of
TOC, pH 8, 25 °C, 3 days
N/A
Linear (r2 =
0.49)
South Platte River,
Colorado (wastewater-
impacted)
Mitch et al.,
2009
2 mM chloramines, pH 7, 10 days
N/A
None
Waste- or algal- impacted
raw drinking water
Xu et al., 2011
2 mM chloramines, pH 7, 25 °C, 7
days
N/A
Linear (r2 =
0.77)
Raw water (from river in
China)
* Non-linear correlation (r2 = 0.40) found between DOC/DON ratio and NDMA.
Six-Year Review 3 6-20
Technical Support Document for Nitrosamines
December 2016
-------
6.5 Key Factors Impacting Formation
A range of factors influences the formation of nitrosamines in drinking water. In this section, the
impacts of chloramination/chlorination conditions, along with water quality parameters such as
DO, pH, temperature and bromide, are discussed.
6.5.1 The Impact of Chlorination and Chloramination
Utilities that switch their disinfectant from chlorine to chloramine may see an increase in the
formation of NDMA depending on the availability of precursor material in the source water.
Data from UCMR 2 show that over 34 percent of samples collected from chloraminating PWSs
had NDMA concentrations greater than the MRL, compared to approximately 4 percent of
samples from chlorinated PWSs. In addition to chloraminating PWSs, water recycling facilities
and utilities adding chlorine to ammonia-containing source waters may also see elevated NDMA
formation.
Dichloramine has been shown to contribute to NDMA formation rates at least an order of
magnitude higher than monochloramine (Schreiber and Mitch, 2005, 2006a). Under typical
drinking water treatment conditions, monochloramine is the dominant chloramine species,
though dichloramine is also present. Chloramine speciation is impacted by both pH and the
chlorine-to-ammonia (ChiNFb) molar ratio. For drinking water utilities that practice
chloramination, research has shown that the use of preformed chloramines, formed under
conditions that favor monochloramine (pH > 8.5 and ChiNFb molar ratio «1), helps to
minimize NDMA formation by reducing the formation of dichloramine (Schreiber and Mitch,
2005; Mitch et al., 2005). Dichloramine formation from the disproportionation of
monochloramine is slow, such that dichloramine formation after the application of preformed
monochloramine should be minimal (Schreiber and Mitch, 2006a).
Schreiber and Mitch (2005) found that the order in which reagents are added during
chloramination treatment may also play a role in nitrosamine formation. Based on experiments
with wastewater, the authors proposed that greater NDMA formation occurs in drinking water
when chlorine is added after ammonia for in situ chloramination than when ammonia is added
after chlorine. This occurred even at molar C1:N ratios of less than 1:1, because of localized
dichloramine formation in areas where the molar C1:N ratio exceeds 1:1 at the point of chlorine
addition prior to complete mixing (Schreiber and Mitch, 2005). Also, under the right conditions,
pre-chlorination of NDMA's precursor DMA will result in the formation of chlorinated DMA,
which when exposed to ammonia forms almost an order of magnitude less NDMA than does
DMA (Schreiber and Mitch, 2005).
Krasner et al. (2008) showed that nitrosamine formation increased after chlorination if the
addition of chlorine resulted in the formation of chloramines. For example, at wastewater
treatment plants with a Ch:N ratio of less than 10:1 (by weight), NDMA concentrations ranged
from non-detection to 3,165 ng/L. When the Ch:N ratio exceeded 10:1 (by weight), NDMA
ranged from non-detection to 8.2 ng/L. Other nitrosamines such as NDEA, NDPA and NPYR
also showed increased formation in the presence of chloramines (Krasner et al., 2008).
Longer contact times with chloramines have been associated with increased nitrosamine
concentrations because of the slow kinetics of the reaction (Mitch et al., 2003a). Therefore,
measurements made at the treatment plant may not be representative of the maximum formation,
Six-Year Review 3 6-21
Technical Support Document for Nitrosamines
December 2016
-------
as nitrosamine formation may continue within the distribution system. UCMR 2 data provide
evidence of this continuing reaction, as shown in Exhibit 6.4. The exhibit shows, for maximum
residence (MR) time and entry point (EP) locations sampled in UCMR 2 where only chloramine
disinfection was used, the number and percentage of those locations at which one or more
samples exceeded various concentration thresholds for NDMA. At all concentration thresholds,
the rate of threshold exceedance at MR time locations is approximately twice the rate of
exceedance at the EP locations. Studies have shown that the reaction is typically limited by the
amount of organic precursors, and not the monochloramine concentration (Mitch et al., 2003a).
Exhibit 6.4: Number and Percentage of Entry Points and Maximum Residence
Time Locations in UCMR 2 Using Chloramine-Only Disinfection With At Least One
Sample Exceeding the Indicated NDMA Thresholds
NDMA
Thresholds
(ng/L)
Count of
Entry Points
Percentage of
Entry Points
(n = 572)
Count of
Maximum
Residence Time
Locations
Percentage of
Maximum Residence
Time Locations
(n = 394)
2
171
29.9%
246
62.4%
6
66
11.5%
93
23.6%
10
38
6.6%
50
12.7%
20
16
2.8%
23
5.8%
60
5
0.9%
5
1.3%
100
2
0.3%
3
0.8%
200
1
0.2%
2
0.5%
600
0
0.0%
1
0.3%
Increased chlorine contact time (before conversion to chloramines) has been associated with a
decrease in NDMA formation (Wilczak et al., 2003; Lee et al., 2008; Chen and Valentine, 2008;
Mitch et al., 2009; Russell et al., 2012), though this practice must be approached cautiously due
to the potential for increasing regulated DBP concentrations.
Nitrification control strategies, including altering the Cb:NH3 molar ratio and using breakpoint
chlorination, may result in increased nitrosamine formation. Increasing the Cb:NH3 molar ratio
helps to control nitrification, but results in greater dichloramine and NDMA formation.
Breakpoint chlorination may lead to the rapid formation of NDMA through reactions with
breakpoint chlorination intermediates (Schreiber and Mitch, 2007).
6.5.2 The Impact of Water Quality Parameters
6.5.2.1 Dissolved Oxygen
Schreiber and Mitch (2005) showed that NDMA concentrations increased with increasing DO
concentrations. Le Roux et al. (2011) found similar results when chloraminating the
pharmaceutical ranitidine. In the absence of oxygen, little NDMA was formed, though under
saturated DO conditions, the NDMA molar yield was 54 percent. Padhye et al. (2010) also found
that the presence of oxygen was a critical factor in the activated carbon transformation of
secondary amines, where adsorbed secondary amines exposed to air for longer periods of time
exhibited significantly higher nitrosamine yields.
Six-Year Review 3 6-22
Technical Support Document for Nitrosamines
December 2016
-------
6.5.2.2 pH
Optimum conditions for NDMA formation from UDMH occur at pH 6 to 8 (Mitch and Sedlak,
2002). As use of dichloramines has been associated with high NDMA yields, the pH of
chloramination is also important to consider. Most studies have found that NDMA formation
increases with increasing pH (Mitch and Sedlak, 2002; Valentine et al., 2005; Schreiber and
Mitch, 2006a; Sacher et al., 2008). Shen and Andrews (2013) examined the pH behavior of the
reaction of chloramine with two pharmaceuticals, ranitidine and sumatriptan, and found the
yields of NDMA peaked at pH 7 for ranitidine and pH 8 for sumatriptan, or about 1.2 to 1.6 pH
units below their respective pKa values. The peak conversion was hypothesized to be a result of
the reaction between dichloramine and the deprotonated amine. Because dichloramine is more
prevalent at acid pH conditions and deprotonated amines are more prevalent at higher pH, this
results in a maximum formation at intermediate pH levels.
6.5.2.3 Temperature and Seasonality
Temperature appears to have minimal impact on NDMA formation, though moderate increases
in NDMA concentrations were observed with decreases in temperature (Mitch et al., 2003a).
This result may be due to the reduction in monochloramine contact time with increasing
temperature, as monochloramine is removed more rapidly by auto-decomposition at high
temperatures (Mitch et al., 2003a). Krasner et al. (2010) found that at pH 8 the formation of
NDMA was relatively unaffected by temperature. At pH 9 they found NDMA formation was
greater at lower temperatures. Similar to the conclusions of Mitch et al. (2003a), the authors
propose that the effect can be explained by chloramines being more stable at lower temperatures
and the reaction rate of NDMA precursors being relatively temperature-independent, leading to
more chloramines being available to react with NDMA precursors at lower temperature (Krasner
et al., 2010). Woods et al. (2015) evaluated seasonal/temperature effects on NDMA formation at
a water system in Denver, Colorado. The pH of the water at the water system was reported to be
"generally greater than 8" during the study. A statistically significant negative correlation was
established between water temperature and NDMA concentrations.
Seasonality appears to have a variable impact on nitrosamine formation. Li et al. (2015) also
noted that the maximum concentration of NDMA in the distribution systems of nine drinking
water treatment plants in East China was greater in the winter than in the summer. However, the
opposite was observed for NPYR. Woods and Dickenson (2015) observed no seasonal trend in
NDMA concentrations measured during EPA's monitoring during UCMR 2. Uzun et al. (2015)
found no significant change in mean NDMA concentrations across the seasons at 12 monitoring
sites in the southeastern United States.
6.5.2.4 Bromide Concentrations
Because bromine species are known to be more effective substitution agents than the equivalent
chlorine species (Symons et al., 1993), waters containing bromide may demonstrate increases in
NDMA formation where formation of bromamines is favorable. Although bromamines are more
reactive, they decompose more rapidly than chloramines (Valentine et al., 2005), such that
waters under the influence of 1 mg/L bromide are not expected to exhibit higher concentrations
of NDMA than they would if only chloramines were present. Le Roux et al. (2012) found that
the presence of 1 mM bromide enhanced the formation of NDMA during chloramination of
DMA and DMA-containing compounds, which was thought to be related to the formation of
Six-Year Review 3 6-23
Technical Support Document for Nitrosamines
December 2016
-------
reactive brominated oxidants. Luh and Marinas (2012) performed experiments with 0.2 mM (10
mg/L) chloramine and 0.0005 mM (22.5 |_ig/L) DMA in either the presence or absence of 0.4
mM (32 mg/L) bromide. They performed experiments between pH 6 and 9. They found that
above pH 7 the formation of NDMA increased significantly. At pHs between 6 and 7 the
formation of NDMA decreased in the presence of bromide. They found that at lower pH a
bromochloramine intermediate was formed. At higher pH they were unable to identify the
intermediate responsible for enhanced NDMA formation. Bromamine, tribromide, hypobromite
and hypobromous acid were tested and ruled out. Zha et al. (2014) found that after ozonation and
chlorination of several model DBP precursors, NDMA formation generally decreased as bromide
concentration increased. The pH used by Zha et al. (2014) was not reported.
6.6 Kinetics and Predictive Models
Various attempts have been made to develop models to predict NDMA formation. These models
vary in sophistication ranging from "curve fitting" of relationships between NDMA and bulk
properties to those incorporating rate equations for each elementary step in the hypothesized
reaction mechanism (Choi and Valentine, 2002; Schreiber and Mitch, 2006a; Chen and
Valentine, 2006, 2007; Kim and Clevinger, 2007; Chen and Westerhoff, 2010). The mechanistic
models use DMA as a model precursor and do not take into account other precursors or real
world water matrices. Using Iowa River water, Chen and Valentine (2006, 2007) found that the
amount of NDMA formed could be predicted from the drop in SUVA, although the relationship
was likely dependent on the source water. Chen and Westerhoff (2010) showed that predictive
power law models had a reasonable fit for real world wastewater samples with high NDMA FP
but did not transfer to drinking water sources with lower NDMA FP. The authors found that
addition of an inorganic nitrogen parameter did not substantially improve the predictions (Chen
and Westerhoff, 2010). Unfortunately, the models based on fitting laboratory data to predictive
functions such as those of Chen and Valentine or Chen and Westerhoff tend to be source water-
specific and cannot be applied in a broad way to accurately predict nitrosamine formation at
water utilities or in distribution systems.
6.7 Summary
Data from UCMR 2 indicate that waters disinfected with chloramines have higher and more
frequent detections of NDMA, a finding that is consistent with data from the literature that shows
that chloramination leads to the highest formation of NDMA. UCMR 2 data also indicate that
detection rates of NDMA are about 10 times higher at PWSs that use chloramines than those at
PWSs that use chlorine. Therefore, utilities that switch from chlorine to chloramines may see an
increase in NDMA and other nitrosamines formation. Because formation of nitrosamines
continues in the distribution system, the highest NDMA concentrations generally occur at the
MR locations of surface water PWSs using chloramines. (As noted in Chapter 5, UCMR 2 data
may underestimate NDMA detections due to the exclusion in that study of consecutive PWSs,
which may have very long residence times from the point of treatment in the system from which
they purchase their water.)
Water utilities using surface water under the influence of wastewater are at a higher risk for
nitrosamine formation during chloramination than those making use of water sources that are not
so influenced. Chloramination of wastewater forms an order of magnitude more NDMA than
chloramination of drinking water; thus, the precursors responsible for NDMA formation are
likely to be found in wastewater and in surface waters impacted by wastewater flows. Because
Six-Year Review 3 6-24
Technical Support Document for Nitrosamines
December 2016
-------
DMA and biological material in wastewater are not significant NDMA precursors in secondary
wastewater effluents, anthropogenic materials in wastewater are thought to be the primary source
of precursors. These materials range from pesticides and dyes to personal care products such as
pharmaceuticals, shampoos, and soaps. Additionally, the use of flocculation aides and anion
exchange resins in drinking water and wastewater treatment may also contribute to NDMA
formation and occurrence. Because these materials have a wide range of chemical structures,
they will exhibit varying degrees of removal during conventional drinking water treatment. A
data gap may exist for understanding the extent of the impact from different precursor sources,
such as the percentage of NDMA detections that can be attributed to wastewater discharge,
polymer addition, distribution system materials, etc. Furthermore, because the kinetics of NDMA
formation from these materials reacting with chloramines is on the order of days, NDMA levels
may continue to rise in the distribution system days after these NDMA formation conditions and
processes are initiated.
NDMA precursors have been shown to be oxidized (i.e., removed) by strong oxidants such as
ozone, permanganate and chlorine dioxide. Other studies, however, have shown that DMA can
be oxidized by strong oxidants to form NDMA. A data gap may exist for understanding the
interactions of strong oxidants and NDMA precursors and their effect on ultimate NDMA
formation.
Though data from UCMR 2 and the literature have increased our understanding of nitrosamine
formation, a number of important questions remain. Chloramination has been identified as the
primary pathway leading to NDMA formation, but a data gap may exist for understanding how
particular source water quality and chloramination operating conditions work together to
contribute to NDMA occurrence. Also, as chlorination is the most widely used disinfection
technique, and nitrosamine formation is observed in PWSs that are using chlorination and not
chloramination, further studies focused on the water quality and treatment parameters that lead to
high nitrosamine occurrence in chlorinating treatment systems may be needed. UCMR 2 data has
shown that the highest NDMA levels are found in the distribution system at MR time sampling
points, suggesting that NDMA formation reactions continue in the distribution system. Not
known, however, is the extent to which distribution system effects such as biofilm growth and
nitrification influence NDMA concentrations.
A variety of nitrogen-containing organics originating from a wide range of natural and
anthropogenic sources have been identified as nitrosamine precursors, though it is unknown to
what extent each of these potential sources specifically contributes to overall formation.
Identification of additional precursors and information on co-occurrence of different nitrosamine
precursors coupled with occurrence data on nitrosamines and other relevant water quality
parameters may help to improve understanding of formation mechanisms. Occurrence data can
also help determine whether similar reactions account for all nitrosamine formation or if there
are competing mechanisms that favor production of one nitrosamine over another. Such
precursor occurrence data may also help to predict concentrations of individual nitrosamines
under given water quality and treatment conditions.
Six-Year Review 3 6-25
Technical Support Document for Nitrosamines
December 2016
-------
7 Treatment
7.1 Introduction
This chapter discusses potential strategies for nitrosamine treatment and control. For a discussion
of treatment for a range of disinfection by-products (DBPs), including activities related to
nitrosamines, see the Agency's Six-Year Review 3 Technical Support Document for Disinfectants
and Disinfection By-Products Rules (USEPA, 2016a).
As mentioned in Chapter 5 of this document, nitrosamines are detected more frequently in public
water systems (PWSs) using chloramines as a disinfectant. Chapter 6 discusses how
chloramination of amine-containing precursors appears to be the major pathway for nitrosamine
formation. However, not all chloraminating PWSs have elevated nitrosamine concentrations in
their finished water, which suggests that there are ways to reduce nitrosamine formation during
chloramination. There are several potential points along the drinking water treatment continuum
where strategies can be employed to reduce nitrosamine formation. One strategy is to implement
source water management strategies to reduce the amount of nitrosamine precursors entering the
treatment plant (for example, reducing the influence of wastewater discharges on the source
water used by the PWS). A second strategy is to remove nitrosamine precursors within the
treatment plant itself prior to application of chloramines. A third strategy is to optimize the
chloramination process to reduce nitrosamine formation. The final strategy is to remove or
reduce nitrosamines after they have formed.
Each of the control strategies mentioned above may have other potential beneficial and
detrimental effects on the treatment process and on overall public health protection. Precursor
removal has the added benefit of reducing levels of other DBP precursors, such as precursors of
trihalomethanes (THM) and haloacetic acids (HAA) and lowering chlorine demand (Karanfil et
al., 2008; Templeton and Chen, 2010). However, source water management strategies often
require cooperation of agencies and groups external to the PWS and can be complicated to
implement. Adding treatment to remove nitrosamine precursors or remove nitrosamines after
their formation can be challenging. Altering the chloramination process is a relatively simple
strategy; however, it may result in an increase in the formation of other DBPs.
This chapter presents potential strategies for nitrosamine control in public drinking water. The
chapter is based on literature published up to December 2015. This field is an area of active
research, and new literature is being published that may further inform our understanding of
nitrosamine control. Discussion of wastewater treatment technologies is included as, in some
cases, it may be easier to prevent precursors from entering the drinking water plant through
treatment at the wastewater plant than to undertake treatment at the drinking water plant. Also, in
some cases, wastewater treatment technologies are similar to processes that are used in drinking
water treatment plants. Processes that are used to prevent nitrosamine formation are discussed
first, followed by processes that remove nitrosamines that have already formed. Similar to
Chapter 6, the focus of this chapter is on NDMA, though other nitrosamines are also discussed
when information is available.
7.2 Prevention of Nitrosamine Formation
This section discusses two general prevention strategies for nitrosamine formation prevention:
precursor reduction/removal and modification of disinfection practices. Among the precursor
Six-Year Review 3 7-1
Technical Support Document for Nitrosamines
December 2016
-------
reduction/removal approaches, there are some that can be effectively implemented at wastewater
treatment plants to reduce concentrations of precursors released to source water. However,
mitigating nitrosamine precursors at wastewater treatment plants may not be an option available
to utilities. Wastewater treatment for precursors would require cooperation between the drinking
water and wastewater utilities.
7.2.1 Precursor Removal
As discussed in the previous chapter, numerous nitrosamine precursors may be present in
drinking water sources, including natural and synthetic amines and other nitrogen-containing
compounds. Many nitrosamine precursors have yet to be identified. Therefore, many treatment
studies focusing on precursor removal use either nitrosamine formation potential (FP) or model
precursors such as dimethylamine (DMA) to study removal by treatment processes. Precursor
removal processes include: source management; physical removal processes such as coagulation,
filtration and absorption; and processes which destroy precursors, such as pre-oxidation.
European water utilities may provide good models of precursor removal in action, as they tend to
use practices that minimize chlorine usage (Karanfil et al., 2008). Templeton and Chen (2010)
surveyed seven nitrosamines at selected UK drinking water supply systems employing a variety
of disinfection techniques, such as hypochlorite, chloramination and chlorine gas. The authors
found that NDMA was detected above the minimum detection level only once. The detected
concentration was 1 ng/L; the detection occurred at the sampling point furthest into the
distribution system. This utility treated organic-rich and nitrogen-rich waters and used amine-
based polymers epi-DMA and polyDADMAC, which would appear to put it at risk of much
greater and more frequent nitrosamine concentrations. One possible explanation as to why the
occurrence of NDMA in this study was low relative to NDMA occurrence observed in the
United States is that disinfectant concentrations in North America can be up to 4 mg/L of total
chlorine and are often higher than the average of 0.5 mg/L typically applied in the UK.
Additionally, none of the systems tested in this study were impacted by wastewater effluents
(Templeton and Chen, 2010).
7.2.1.1 Source Water Management
Studies have shown that wastewater discharges are associated with occurrence of nitrosamine
precursors (Schreiber and Mitch, 2006b; Chen et al. 2009). At the South Platte River, Krasner et
al. (2008) showed that dissolved organic nitrogen (DON) increased downstream of wastewater
discharges. Careful consideration of drinking water intake locations relative to wastewater
discharges may reduce human exposure to NDMA through drinking water. Longer distances
between wastewater discharges and drinking water intakes allow natural attenuation processes to
occur, which can reduce levels of nitrosamine precursors.
While altering source water intake practices can be an effective method to control nitrosamine
precursors, implementation can be complicated. Source water management strategies often
require cooperation of outside groups such as wastewater utilities, permitting agencies, and other
stakeholder groups.
Six-Year Review 3 7-2
Technical Support Document for Nitrosamines
December 2016
-------
7.2.1.2 Coagulation
While the combination of coagulation and flocculation is successfully used to remove
chlorination DBP precursors, it is less effective for nitrosamine precursors. Bench-scale jar tests
were performed on German river waters to investigate the effect of coagulation, flocculation and
sedimentation on the fate of nitrosamine precursors. Iron and alum salts added at 5 and 10 mg/L
lowered the nitrosamine FP by less than 10 percent, compared to the FP without coagulation
(Sacher et al., 2008). Lime softening led to only a 12 percent decrease in nitrosamine FP at a
drinking water treatment plant in Ann Arbor, Michigan (Mitch et al., 2009). However, the poor
removal was confounded by the addition of DADMAC polymer, which is an NDMA precursor
and may have offset removal by coagulation (Mitch et al., 2009).
According to Westerhoff et al. (2006), maximum DON removal by coagulation with aluminum
salts was about 40 percent. Coagulation with aluminum salts in conjunction with low dosages of
polyDADMAC improved DON removal by 15 to 20 percent. However, as noted above,
polyDADMAC is a nitrosamine precursor, and although DON was removed, the final effect of
polyDADMAC on nitrosamine removal was not measured. Coagulation preferentially removed
higher molecular weight fractions of DON (Westerhoff et al., 2006). A survey of 16 full-scale
drinking water treatment plants with water impacted by algal blooms or wastewater found an
average removal of 30 percent of DON by coagulation (Dotson and Westerhoff, 2009). Pietsch et
al. (2001) found that aliphatic amines cannot be completely removed using either aluminum or
iron coagulants. They found removal rates of less than 10 percent for most aliphatic amines; the
exceptions were ethanolamine and ethylenediamine, which showed 30 and 45 percent removal,
respectively. Liao et al. (2015b) found that coagulation and sedimentation removed only 18
percent of nitrosamine FP in a Chinese pilot plant. A follow up study (Liao et al., 2015a)
attributed poor removal to the fact that both coagulants and nitrosamine precursors tended to be
cationic.
The use of amine polymers in the coagulation process may obscure benefits achieved through
coagulation. Mitch et al. (2009) examined the removal of DBP precursors through different unit
processes (including coagulation) for surveyed water treatment plants. They found a 21 percent
removal of DON, but a 43 percent increase in NDMA FP, which was most likely due to the
presence of certain polymers. Laboratory experiments with polyDADMAC polymer confirmed
that ozonation of polyDADMAC polymers can produce NDMA (Padhye et al., 2011). Padhye
found that doses of polyDADMAC polymer ranging from 5 to 10 mg/L yielded NDMA
concentrations up to 800 ng/L upon ozonation (Padhye, 2010).
Cornwell et al. (2015) examined a number of natural polymers as potential substitutes for
polyDADMAC. They tested polymers based on corn, potato, tapioca and shellfish (chitosan) on
waters from nine different water treatment plants. None of the tested polymers formed reportable
levels of nitrosamines. For each of the waters tested, at least one of the natural polymers
performed as well or better than polyDADMAC in terms of turbidity removal. Some of the
natural polymers performed worse in terms of headloss. The headloss problems were not present
if the filters were backwashed with chlorinated water, leading to the hypothesis that added
headloss was due to biological activity. Zeng et al. (2014) developed a phosphine based polymer
as a substitute for polyDADMAC polymers. The phosphine polymer did not produce nitrosamine
when reacted with chloramine. The polymer showed the ability to increase dissolved organic
carbon (DOC) removal by 17 to 25 percent when compared to alum or ferric coagulants alone,
Six-Year Review 3 7-3
Technical Support Document for Nitrosamines
December 2016
-------
which was similar to polyDADMAC. Although the polymer shows potential for reducing
nitrosamine formation from water treatment polymers, it is not yet commercially available.
7.2.1.3 Sorption
Powdered activated carbon (PAC) may remove some nitrosamine precursors via sorption
processes. One laboratory study showed that a PAC dose of 5 mg/L lowered the NDMA, N-
nitrosomorpholine (NMOR) and A'-nitrosopyrrolidine (NPYR) FP of river water by a factor of
two (Sacher et al., 2008). Increased doses resulted in modestly increased precursor removal, with
a dose of 100 mg/L of PAC resulting in a lowering of the nitrosamine FP by 83 percent (Sacher
et al., 2008). Westerhoff et al. (2006) found PAC could remove DON. A dose of 5 mg/L
removed 20 percent of DON from surface water and 50 percent from finished drinking water
(Westerhoff et al., 2006). Increasing the dose to 25 mg/L only lowered nitrosamine FP by an
additional 5 to 10 percent (Westerhoff et al., 2006). Hanigan et al. (2012) performed several
adsorption experiments using a secondary wastewater effluent blended with a river water source
to simulate a drinking water source impacted by wastewater. They found removals of NDMA FP
between 70 and 90 percent using doses of PAC ranging from 15 to 210 mg/L over a 4-hour
contact time. In contrast, the PAC decreased DOC by less than 25 percent and UV-254 (a water
quality parameter indicating the presence of UV-absorbing organic matter) by less than 50
percent.
Mixed results have been found for precursor removal by GAC. Hwang et al. (1994) found
limited removal of DMA using GAC filters. Farre et al. (201 la) studied removal of NDMA FP at
a wastewater plant using biologically active GAC. They found 85 percent removal of the NDMA
FP. The differences in removal between these two studies may be that the GAC in the Farre et al.
study was biologically active while the GAC in the Hwang et al. study was not. Hanigan et al.
(2012) performed column studies using GAC and found 50 percent removal of NDMA FP even
after 10,000 bed volumes, even though DOC and total dissolved nitrogen broke through much
earlier. They also examined two full-scale GAC contactors and found NDMA FP removal
ranging from 54 to 84 percent.
Krasner et al. (2015) examined GAC and PAC for nitrosamine precursor removal. GAC and
PAC were effective at removing NDMA precursors and (with a sufficient dose) removed them
more effectively than they removed DOC. PAC also removed NDMA precursors from
polyamine use. PAC and GAC were not, however, effective at removing precursors from
polyDADMAC. Magnetic ion exchange resin (MIEX) was found able to remove NDMA
precursors; however, it also contributed NDMA precursors to the water. Chu et al. (2015) found
that 20 mg/L PAC, plus conventional treatment, removed 37 percent more NDMA precursors
than conventional treatment alone in a Chinese pilot treatment plant. PAC with 1 mg/L
potassium permanganate, plus conventional treatment, was able to eliminate 86 percent more
NDMA precursors than conventional treatment alone.
Wu et al. (2015) compared adsorption of seven secondary and tertiary amines onto zeolites and
PAC using laboratory water. One particular zeolite called mordenite was able to remove over 90
percent of all the amine precursors except for 4-dimethylaminoantipyrine (DMAP) at a dose of
100 mg/L. The zeolites performed best removing small precursors which were positively charged
under the experimental conditions. PAC performed better removing larger, less hydrophilic
precursors such as DMAP. Performance of both the zeolites and the PAC was poorer when a raw
surface water was used to dissolve the precursors instead of laboratory water.
Six-Year Review 3 7-4
Technical Support Document for Nitrosamines
December 2016
-------
Removal of NDMA precursors by activated carbon has been found to be highly pH dependent
and site specific. Chen et al. (2015b) examined water from an Arizona river mixed with 20
percent wastewater effluent and found removal of NDMA precursors increased from 29 percent
at pH 3 to 58 percent at pH 9.5. The opposite trend, however, was found when aquaculture-
impacted Chinese lake water was tested. NDMA precursor removal decreased from 83 percent at
pH 3 to 34 percent at pH 11, while NDEA precursor removal decreased from 89 percent at pH 3
to 51 percent at pH 11.
7.2.1.4 Biological Filtration and Related Techniques
Biological processes at both water and wastewater treatment plants have been shown to remove
nitrosamine precursors. If a drinking water treatment plant uses biological treatment,
downstream filtration and disinfection may be required to prevent sloughed biomass from
colonizing the distribution system.
A study of wastewater treatment plants in California showed that secondary treatment led to only
a small decrease in DON (24 percent) over a 20-day incubation period (Pehlivanoglu-Mantas and
Sedlak, 2006). Krasner et al. (2009) conducted a survey of 23 advanced wastewater treatment
plants in the Midwest and the far western United States. The authors found that partial
nitrification (reduction of dissolved nitrogen in the form of ammonia to 2-10 mg/L) and complete
nitrification (reduction to below 2 mg/L) reduced the level of precursors by approximately 61
percent and 50 percent, respectively. Partial denitrification (nitrate-nitrogen reduction to 5-10
mg/L) and complete denitrification (nitrate-nitrogen reduction to below 5 mg/L) decreased
precursors by approximately 9 percent and 66 percent, respectively, compared to no treatment
(Krasner et al., 2009).
On average, 50 to 70 percent removal of NDMA precursors was demonstrated at four wastewater
treatment facilities using aerated lagoon, advanced biological treatment, conventional secondary
treatment with activated sludge and a membrane bioreactor, respectively (Krasner et al., 2008).
An anaerobic digester at a municipal wastewater treatment plant was able to completely
biodegrade the model precursors pyrrolidine (PYR), diethylamine (DEA), DMA and MEA
(Padhye et al., 2009).
A study of a biologically-active GAC filter (following ozonation) found removals of 99 percent
or more for four pharmaceuticals (known NDMA precursors) at a tertiary treatment facility in
Australia (Farre et al., 201 la). The authors found that pilot-scale biologically active sand filters
did not remove nitrosamine precursors, implying that adsorption followed by biological
degradation was key to the removal process by biologically active GAC.
Biologically-active filtration processes in drinking water treatment have also been shown to
reduce nitrosamine precursor levels. Biologically-active GAC removed 67 percent of NDMA FP
at a drinking water facility in Ann Arbor (Mitch et al., 2009). Krasner et al., (2015) found mixed
results for removal of NDMA precursors by biologically-active filters in water treatment plants.
Out of fourteen samples from biofilters, NDMA formation was lowered in four, remained the
same in one and increased in nine. Liao et al. (2014, 2015b) examined the performance of a
biofilter at a Chinese pilot plant consisting of ozonation, coagulation and sedimentation, further
ozonation, biologically-active GAC filtration and sand filtration. The biologically active GAC
filter removed 59 percent of NDMA precursors, 55 percent of NDEA precursors and more than
70 percent of NPYR precursors (Liao et al., 2015b). Biofiltration was found to remove the lower
Six-Year Review 3 7-5
Technical Support Document for Nitrosamines
December 2016
-------
molecular weight DOC better than higher molecular weight DOC (Liao et al., 2014) and the
cationic portion preferentially (Liao et al., 2015a). Further examining the characteristics of
nitrosamine precursors, this research team found that 70 percent of the NDMA FP was
biodegradable but only 42 percent was adsorbable, while 45 percent of NDEA FP was
biodegradable and 59 percent was adsorbable (Liao et al., 2015b). Liao et al. (2015c) studied the
effects of backwashing on biomass and nitrosamine precursor removal in a biologically-active
filter. The pilot biofilter showed an approximate 50 percent decrease in biomass on the filter
immediately after backwashing. The biomass recovered to near original levels after 2 days.
Nitrosamine removal was initially about 60 percent in the filter. It increased to 80 percent on
backwashing and then returned to 60 percent over time.
During full-scale testing of water facilities, riverbank filtration removed secondary amine
precursors to varying degrees: 76 percent removal for DMA, 66 percent for DEA, 52 percent for
PYR and 80 percent for morpholine were observed. Riverbank filtration decreased NDMA FP by
93 percent and NPYR and NDEA FPs by 50 percent and 66 percent, respectively. Ground water
recharge at a separate utility decreased NDMA FP by 88 percent and NPYR FP by 63 percent
(Sacher et al., 2008). Krasner et al. (2012c) found a 64 percent reduction in NDMA FP at a plant
using riverbank filtration in the United States. In a series of 5-day bench-scale biological sand
filtration trials, NDMA precursor removal ranged from approximately 45 to 80 percent, NPYR
precursor removal ranged from 0 to approximately 85 percent and NMOR precursor removal
ranged from 0 to approximately 50 percent (Krasner et al., 2008).
Liao et al. (2015d) examined the ability of biofilters to degrade DMA specifically. They
extracted a culture from a pilot-scale GAC biofilter and determined the effect of carbon and
nitrogen amendments on degradation and bacterial community composition. By day 5, in the
biofilter culture amended with carbon (in the form of glucose), nearly complete removal of DMA
(initial concentration: 10 mg/L) was observed. Biofilter culture amended with nitrogen (i.e.,
ammonia) did not significantly differ from a culture with no amendments; at day 5, DMA levels
were still approximately 1.5 mg/L, down from 10 mg/L. A rise in ammonia over time in all
inoculated cultures confirmed earlier studies that had shown DMA could be broken down to
ammonia by bioculture. The authors analyzed community composition from all cultures on day 0
and again on day 7 and found significant changes, particularly in the culture amended with
glucose. Brevundimonas was dominant in the initial culture, the culture amended with ammonia,
and the non-amended culture, but was not found in the glucose-amended culture. Acinetobacter
was common in the glucose-amended culture but was not detected in the others. Except for
Bacillus and Pseudomonas, the genera isolated from the cultures were not previously known to
degrade DMA, although they have been shown to biodegrade other organic pollutants.
7.2.1.5 Membrane Filtration
Removal of nitrosamine precursors by membranes depends on the type of filter and the
membrane pore size (microfiltration (MF), ultrafiltration (UF), nanofiltration (NF) and reverse
osmosis (RO)). As expected, because NDMA precursors are associated with low molecular
weight compounds, UF has been found to exhibit negligible removal (Pehlivanoglu-Mantas and
Sedlak, 2008). On the other hand, Deeb et al. (2006) found that the fraction of NDMA precursors
removed during MF ranged from 12 percent to as high as 95 percent at advanced water treatment
facilities. Although media filtration was unsuccessful in removing NDMA precursors at selected
wastewater treatment plants in California, RO demonstrated complete removal (Mitch and
Sedlak, 2004). Steady-state rejections of nitrosamine precursors of concern (viz., di-n-
Six-Year Review 3 7-6
Technical Support Document for Nitrosamines
December 2016
-------
butylamine, DEA, DMA, di-/?-propylamine, MEA and PYR) have been observed to exceed 98
and 99 percent by NF and RO membranes, respectively (Miyashita et al., 2009). Krasner et al.
(2008) also found NPYR and NMOR precursors are well removed by RO membranes.
7.2.1.6 Oxidation of Precursors
NDMA precursors can be degraded by oxidation via chlorine, permanganate, hydrogen peroxide,
ozone, chlorine dioxide and ferrate. Lee et al. (2007b) showed that NDMA precursors such as
natural organic matter (NOM) can be oxidized by ozone and chlorine dioxide. A decrease in
NDMA FP of between 32 and 94 percent was achieved when river waters were dosed with ozone
or chlorine dioxide at concentrations of 1.9 mg/L and 2.7 mg/L, respectively, for a contact time
of 5 minutes (Lee et al., 2007b). With natural waters they found chlorine dioxide and ozone to
achieve similar removals. In laboratory experiments with model precursors, however, they found
that ozone reduced NDMA FP more effectively, as chlorine dioxide produced DMA, which
could react with chloramines to form NDMA (Lee et al., 2007b). For example, they found that
ozone reduced NDMA production by DMA by 95 percent, but chlorine dioxide reduced NDMA
formation by DMA by less than 10 percent (Lee et al., 2007b). Other studies have found slow
oxidation of aliphatic amines by ozone. Only 20 percent decay was observed when 3.5 mg/L of
ozone was added to a solution containing 1.6 mg/L trimethylamine (TMA), 3.2 mg/L DMA and
5.2 mg/L ^-propylamine (NPA) for 100 minutes (Sacher et al., 2008). Temperature and pH were
not specified in the report. Ozonation is more effective for removing cyclic amines, as rate
constants for morpholine and piperazine are two orders of magnitude higher than for aliphatic
amines (Sacher et al., 2008). Ozonation may have the additional benefit of reducing precursors
of other halogenated DBPs such as THMs and HAAs. On the other hand, bromate formation is a
concern with ozone use, and bromate must be monitored when ozone is used as an oxidant.
Liao et al. (2014) examined nitrosamine precursor removal in a Chinese pilot drinking water
plant set up on a drinking water source known to be impacted by wastewater. Drinking water
plants using this water source were known to have detected a wide variety of nitrosamines. The
pilot plant consisted of coagulation, sedimentation, ozone, biologically-active GAC filtration and
sand filtration. They found approximately 20 percent removal of nitrosamine precursors after the
coagulation/sedimentation stage, 50 percent removal after mid-ozonation and 88 percent removal
after GAC filtration. Further experiments found that ozonation preferentially removed nonpolar
fractions of the precursors, removing 50 to 60 percent of this type of precursor (Liao et al.,
2015a). Shah et al. (2012) found that ozone reduced NDMA formation by about 50 percent with
exposure (measured as concentration x time, or CT) of 0.4 mg-min/L and chlorine achieved
similar removal with exposure of about 60 mg-min/L. These exposure values are about 20
percent and 43 percent, respectively, of the CT required for Giardia control.
McCurry et al. (2015) examined the effect of preoxidation on NDMA formation, using 14 source
water samples from 10 drinking water plants. Without pretreatment with oxidants, NDMA
concentrations ranged from 5.6 to 58 ng/L upon chloramination of the water samples. The waters
were treated with each oxidant (ozone, HOC1 and UV) at two different CTs. For HOC1 those
represented 10 and 42 percent of the CT required for 3-log Giardia inactivation. For ozone the
concentrations represented 10 and 50 percent of the CT required for 3-log Giardia inactivation.
For UV (tested using both low-pressure and medium-pressure mercury lamps), fluence rates of
186 mJ/cm2 and 1,000 mJ/cm2 were used. At high doses, ozone, HOC1 and medium-pressure UV
were all fairly effective (median removal of NDMA FP over approximately 80 percent). At the
lower doses ozone was still effective with 78 percent removal, while HOC1 removal efficiency
Six-Year Review 3 7-7
Technical Support Document for Nitrosamines
December 2016
-------
was only 47 percent of the NDMA precursors and medium pressure UV removal efficiency was
only 54 percent of the NDMA precursors. Low-pressure UV performed significantly worse than
the other techniques at both the high and low doses. Reporting on the same set of experiments,
Krasner et al. (2015) indicate that chlorine pre-oxidation was most effective at a pH between 8
and 9. Krasner et al. (2015) also report that although addition of a high dose of polyDADMAC
polymer increased NDMA FP at the bench scale, addition of a low dose of polyDADMAC at
full-scale plants did not appear to affect NDMA FP.
Chen and Valentine (2008) showed that NDMA formation decreased with increasing pre-
chlorination contact time and dose. For example, NDMA concentrations decreased from 30 ng/L
to 10 ng/L with 5.7 mg/L free chlorine contact time for 10 minutes before formation of
chloramines as compared to use of preformed chloramines. Decreases in NDMA formation
ranged from about 17 to 83 percent compared to preformed chloramines depending on chlorine
dose and contact time (Chen and Valentine, 2008). Charrois and Hrudey (2007) also found a
decrease in NDMA formation with increasing free chlorine contact time and dose at water
treatment facilities in Alberta. Reductions of NDMA FP ranged from 68 to 93 percent for a free
chlorine contact time of 2 hours prior to ammonia addition (Charrois and Hrudey, 2007). A
review of several kinetic studies reported that chlorination of aliphatic amines was found to be
considerably faster than ozonation, suggesting that prechlorination may be faster for oxidizing
precursors (Sacher et al., 2008).
Although oxidation of precursors may be effective for nitrosamine control, it may produce other
DBPs. Shah et al. (2012) found that chlorine produced THMs, HAAs and chloral hydrate, while
ozone produced bromate, chloropicrin and chloral hydrate. In this study, doses of pre-oxidants
high enough to remove NDMA precursors did not cause formation of DBPs to concentrations
that exceeded regulatory levels. However, THM formation approached regulatory limits at the
highest chlorine exposure, and bromate formation approached regulatory limits at the highest
ozone exposure (Shah et al., 2012).
River water spiked with 21 mg/L of ferrate reduced between 46 and 84 percent of the NDMA FP
within an hour (Lee et al., 2008). Ferrate, however, has neither been used in full-scale treatment
plants, nor has it been widely studied in this context. Pre-oxidation with ferrate may be difficult
due to the highly unstable nature of potassium and sodium salts (Sharma 2002).
Chen and Valentine (2008) studied precursor oxidation using Iowa River water. Application of
either 10 mg/L of permanganate or 3 mg/L of hydrogen peroxide for one hour reduced NDMA
formation by approximately 50 percent, while exposure to 7 mg/L of ozone reduced NDMA
formation by 75 percent.
Sunlight may also oxidize nitrosamine precursors. Chen and Valentine (2008) found a 25 percent
reduction in NDMA formation when Iowa River water was exposed to simulated sunlight. Mitch
et al. (2003b) found, however, that most nitrosamine precursors were non-reactive with UV light.
It is important to note that some studies have shown that oxidation of DMA may actually lead to
the formation of NDMA, particularly at higher pH (Nawrocki and Andrzejewski, 2011). The
application of strong oxidants such as ozone, chlorine dioxide, permanganate and ferrate (IV) to
DMA in the presence and absence of ammonia have been shown to form NDMA (Andrzejewski
and Nawrocki, 2007). Lee et al. (2007b) found that NDMA FP remaining after pre-oxidation by
ozone or chlorine dioxide could be largely attributed to DMA formation by the oxidants. The
Six-Year Review 3 7-8
Technical Support Document for Nitrosamines
December 2016
-------
DMA can then be oxidized to NDMA by the oxidants. Therefore, oxidation of precursors may
not be a viable option in water with high pH (e.g., >8) and significant DMA concentration.
Studies have also shown that chlorine can react quickly with DMA to form chlorinated DMA
which may remain in solution to later react with chloramine to form NDMA (Selbes et al., 2015).
Lv et al. (2015) found that oxidation of the pharmaceutical chlorpheniramine by ozone formed
DMA, which was then oxidized by ozone to form NDMA. The concentration increased for about
20 minutes, after which it began to decrease as additional ozone contact time oxidized the
NDMA.
Selbes et al. (2014) found complex relationships between oxidation of precursors (using chlorine,
chlorine dioxide and ozone) and NDMA yield. They examined 15 NDMA precursors that
contained a DMA structure and found that most precursors showed a decrease in yield of NDMA
of about half upon exposure to 3 mg/L of chlorine. Water treatment polymers such as
polyDADMAC, polyacryl and polyamine and the pesticide diuron, however, did not show a
reduced yield of NDMA in the presence of chlorine. With some precursors, such as TMA and
ranitidine, the yield of NDMA decreased with increased chlorine contact time, while with others,
such as DMA, increased contact time did not further lower NDMA formation.
Chlorine dioxide was found to significantly decrease the NDMA FP of precursors that had high
NDMA yields such as dimethylisopropylamine (DMiPA), ranitidine and dimethylbenzylamine
(DMBzA). For these precursors, NDMA yields dropped from initial values of around 80 percent
to a yield of 15 percent after a 5-minute contact time and dropped to a yield of 4 percent after a
contact time of 15 minutes. Precursors that had low NDMA yields, such as DMA and TMA,
showed little change in NDMA formation upon exposure to chlorine dioxide. Polymers such as
polyDADMAC showed a slight decrease in NDMA yield of about 5 to 10 percent on exposure to
chlorine dioxide (Selbes et al., 2014). Selbes et al. (2015) conclude that chlorine dioxide would
be more appropriate in waters with high wastewater content than in waters with a low level of
reactivity precursors. Gan et al. (2015) found that pre-oxidation with chlorine dioxide led to a
lower yield of NDMA after chloramination in the case of 10 out of 13 amine precursors. The
counterexamples were the precursors DMA, A', A'-Dimethyl-p-phenylenediamine and daminozide.
For ozone, a slight initial increase of NDMA formation with DMA was noted, followed by a
decrease with increased contact time (Selbes et al., 2014). The authors performed further
experiments to demonstrate that oxidation of DMA by ozone could produce NDMA, but then the
NDMA was oxidized by hydroxyl radicals created from the ozone. Similarly, ozone also
increased NDMA formation initially for precursors such as methylene blue, dimethylaniline,
polyDADMAC and dimethylphenetylamine, but overall NDMA formation decreased with
increased contact time. Ozone also significantly reduced NDMA formation from TMA and high-
NDMA-yielding compounds such as DMiPA, ranitidine and DMBzA. The authors identified
effects of pH on oxidation of precursors as well. They found that the reactions proceeded faster
with deprotonated amines. For chlorine, this resulted in an optimum pH of about 8.5 for pre-
oxidation (about halfway between the pKa for hypochlorous acid and the pKa for the respective
amine). For ozone and chlorine dioxide, a pH above the pKa of the amine generally produced the
greatest reduction in yield of NDMA.
Wang et al. (2015c) documented precursor-specific differences in the effects of oxidants on
NDMA FP. They examined four pharmaceutical precursors (ranitidine, doxylamine, nizatidene
and carbinoaxime) and four oxidants (chlorine, ozone, permanganate and chlorine dioxide).
Ozone was the most effective oxidant. The highest concentration of ozone removed
Six-Year Review 3 7-9
Technical Support Document for Nitrosamines
December 2016
-------
approximately 90 percent of all the pharmaceuticals and approximately 90 percent of NDMA FP
from all the pharmaceuticals. Ozone removed doxylamine and carbinoaxime more effectively
than ratinidine and nizatidine. Chlorine dioxide and chlorine were effective at breaking down the
pharmaceuticals, but did not perform as well at reducing NDMA FP. NDMA FP even increased
when nizatidine was oxidized with chlorine and when carbinoxamine was oxidized with chlorine
dioxide (at certain doses). . While permanganate could oxidize the pharmaceuticals, it was too
slow to be effective in lowering NDMA FP.
Exhibit 7.1 summarizes selected information about the studies cited in this section.
Exhibit 7.1: Oxidation of Nitrosamine Precursors
Oxidant
Study
Precursor
Reduction
in NDMA FP
Potential Issues
Ozone
Lee et al., 2007b
NOM
32-94%
Bromate formation
Ozone
Lee et al., 2007b
DMA
95%
Bromate formation
Ozone
Sacher et al. 2008
DMA, TMA,
NPA
20%
Bromate formation
Ozone
Chen and Valentine, 2008
NOM
75%
Bromate formation
Ozone
Liao et al., 2014
NDMA FP
45%
Bromate formation
Ozone
McCurry et al. 2015
NDMA FP
78+%
Bromate formation
Chlorine Dioxide
Lee et al., 2007b
NOM
32-94%
Potential DMA formation,
chlorite formation
Chlorine Dioxide
Lee et al., 2007b
DMA
<10%
Potential DMA formation,
chlorite formation
Chlorine
Chen and Valentine, 2008
NOM
17-83%
THM and HAA formation
Chlorine
Charrois and Hrudey, 2007
NOM
68-93%
THM and HAA formation
Chlorine
McCurry et al., 2015
NDMA FP
47+%
THM and HAA formation
Ferrate
Lee et al., 2008
NOM
46-84%
Oxidation of DMA
Permanganate
Chen and Valentine, 2008
NOM
50%
Oxidation of DMA
Hydrogen
Peroxide
Chen and Valentine, 2008
NOM
50%
Oxidation of DMA
UV
Chen and Valentine, 2008
NOM
25%
No residual
UV
McCurry et al., 2015
NDMA FP
29+%
No residual
As can be seen from Exhibit 7.1, oxidation of precursors can be effective but effectiveness can
vary significantly depending on precursors, oxidant dose and water quality. Potential issues
associated with use of this approach include considerations related to the potential for formation
of THMs and HAAs, bromate and NDMA.
7.2.2 Modification of Disinfection Practice
As nitrosamines are DBPs, nitrosamine formation can be reduced by changing disinfection
practices. Possible modifications include modification of the chloramination technique or
conversion from chloramination to another form of disinfection.
7.2.2.1 Modification of Chloramination Techniques
Because NDMA is mainly formed from the reaction of amine precursors with dichloramine,
NDMA can be reduced by minimizing dichloramine formation (Mitch et al., 2005; Schreiber and
Six-Year Review 3 7-10
Technical Support Document for Nitrosamines
December 2016
-------
Mitch, 2005). This is done by chloraminating at chlorine to ammonia ratios of much less than 1,
with a pH greater than 8.5. Farre et al. (201 lb) found in bench-scale experiments on wastewater
that using preformed monochloramine instead of adding chlorine to the ammonia in-line reduced
NDMA concentrations from 310 ng/L to 16 ng/L at a dose of 10 mg/L with a 24-hour contact
time. The practice of using low chlorine-to-ammonia ratios may become problematic, however,
as ammonia-oxidizing bacteria may proliferate when high concentrations of ammonia are
present, resulting in nitrification. To control nitrification episodes at low chlorine-to-ammonia
ratios, utilities have used several strategies. Breakpoint chlorination eliminates free ammonia.
Breakpoint chlorination, however, promotes a series of undefined reactions and chemical
intermediates which can also increase NDMA concentrations up to an order of magnitude
(Charrois and Hrudey, 2007; Schreiber and Mitch 2007). Other ways to limit nitrification include
raising pH to above 9, reducing exposure to sunlight, and changing operations to reduce water
age.
Bench-scale experiments at a wastewater treatment plant have also shown that reducing contact
time between disinfectant and wastewater can reduce nitrosamine concentrations (Farre et al.,
201 lb). At a contact time of 8 hours or less, NDMA was not detected when the authors used
preformed chloramine doses up to 15 mg/L. No NDMA was detected using chloramines formed
in-line by addition of chlorine to ammonia or by dichloramine, with disinfectant doses of 4 mg/L
and contact times of 8 hours. Disinfectant concentrations of 10 mg/L did yield 18 ng/L NDMA
for in-line chloramines and 25 ng/L for dichloramines. A 24-hour contact time yielded 16 ng/L
NDMA for a preformed chloramines dose of 10 mg/L, 310 ng/L NDMA for an in-line
chloramines dose of 10 mg/L and 530 ng/L NDMA for a dichloramine dose of 10 mg/L. Of
course, any reduction in contact time would need to be balanced with disinfection requirements
to ensure that disinfection capabilities were not reduced below what is required by regulation
(Farre et al., 2011b).
These studies indicate that using preformed chloramines and limiting contact time may reduce
nitrosamine formation on par with precursor oxidation. However, there are other important
considerations, including maintaining sufficient contact time to achieve disinfection goals,
relative amounts of other DBP formation, and nitrification potential in the distribution system.
7.2.2.2 Conversion from Chloramination to Other Disinfection Practices
Because nitrosamines are found primarily in chloraminated distribution systems, eliminating
chloramination may reduce nitrosamine formation. Switching to a different residual disinfectant,
however, may increase the formation of other DBPs. Free chlorine can result in THM and HAA
formation, ozone can result in bromate formation, and chlorine dioxide can result in chlorite (and
chlorate) formation. Also, the use of free chlorine in ammonia-containing waters may result in
unintended chloramination, such that nitrosamines may still be formed. Also, other oxidants such
as chlorine dioxide and ozone may form NDMA if DMA is present in the source water
(Nawrocki and Andrzejewski, 2011).
7.3 Nitrosamine Removal
Treatment processes placed after the point of disinfectant application can remove or reduce
concentrations of nitrosamines formed as DBPs. Furthermore, treatment processes can remove or
reduce concentrations of nitrosamines that may be present in the drinking water treatment plant
influent.
Six-Year Review 3 7-11
Technical Support Document for Nitrosamines
December 2016
-------
Nitrosamines such as NDMA have low vapor pressures, low Henry's Law Constants and low
octanol-water partitioning coefficients (Mhlongo et al., 2009), meaning they are not highly
volatile and are hydrophilic. Therefore, they are expected to be mobile in water and will tend to
remain in the aqueous phase in contrast to volatilizing to air or sorbing to carbon or soils. From a
treatment standpoint, although some volatilization of nitrosamines may occur, the degree of
volatilization is likely not sufficient to make a treatment process that relies on partitioning from
water to air viable. Enhanced coagulation, adsorption, membranes, metal catalysis, UV,
advanced oxidation and electrochemical techniques can remove nitrosamines and are discussed
briefly in the following sections.
7.3.1 Enhanced Coagulation
Enhanced coagulation uses increased coagulant doses to remove organic compounds from
drinking water. However, conventional alum and ferric coagulants dosed at up to 12 mg/L
achieved minimal NDMA removal (<7 percent) from raw drinking water (Sacher et al., 2008).
Enhanced coagulation would also only remove nitrosamines that entered with raw water or that
were formed before disinfection. Because of its typical placement in a treatment plant (i.e.,
before filtration and final disinfection), enhanced coagulation would not be able to remove
nitrosamines formed during the final disinfection step. Thus, application of enhanced coagulation
is very limited for NDMA removal.
7.3.2 Adsorption
Adsorption can remove nitrosamines to varying degrees. Dosing raw water with 50 mg/L of PAC
for 24 hours resulted in 17 percent NDMA removal (Chung et al., 2009). Although a contact
time of between 1 and 24 hours had little effect, a dose of 200 mg/L of PAC for 60 hours
removed 45 percent of NDMA (Chung et al., 2009). Gunnison et al. (2000) demonstrated that
GAC units have limited effectiveness in removing NDMA from the North Boundary
Containment System near Denver, Colorado, which is designed to treat munitions-contaminated
alluvial aquifers. GAC units in this system decreased NDMA from 350 ng/L to 200 ng/L.
Additionally, results from sequential desorption tests of the soil showed that nearly all adsorbed
NDMA desorbed after one cycle. Since NDMA sorbs poorly onto soil irrespective of soil
properties, riverbank filtration is not likely to be an effective remediation technology for formed
nitrosamines (Mohanty et al., 2006).
Wang et al. (2013b) did laboratory experiments with nanoparticles of activated carbon. They
tested three types of activated carbon (bamboo, charcoal and coconut shell) and found coconut
shell activated carbon had the same or better removal efficiency of the nitrosamine and precursor
compounds than the other carbon types. Removal of 50 percent of most targeted compounds in
lab reagent water was achieved with a typical dosage (from 1 to 20 mg/L) of the activated carbon
nanoparticles and a contact time of 4 hours. Removal rates generally increased with time. For
some nitrosamine compounds, coconut shell-based carbon had much better removal efficiencies
than the other carbon sources. Generally, the removal efficiency was lower in prefiltered natural
river than in reagent water (presumably due to competitive adsorption), while the study's two pH
conditions (6.6 and 8.6) did not significantly affect removal efficiency.
Zhu et al. (2001) attempted to sorb NDMA onto a zeolite, which has a large surface area. The
authors reported higher sorption efficiencies than alumina and silica, but the laboratory
Six-Year Review 3 7-12
Technical Support Document for Nitrosamines
December 2016
-------
conditions used for adsorption were at much higher concentrations than would be found in
typical drinking water plants.
Chen et al. (2015c) adsorbed NDMA onto biochar made from bamboo, rice straw and wood
sawdust. Batch experiments were performed in the laboratory using concentrations of NDMA
ranging from 0.5 to 20 mg/L, much higher than typically present in drinking water. The
researchers found that with a 61.68 percent removal efficiency, biochar from bamboo
manufactured at 500 degrees C was more effective than biochar from wood or rice or from
bamboo prepared at other temperatures. Solution chemistry (e.g., pH, metal ions) did not
significantly affect biochar performance.
Dai et al. (2009) showed that activated carbon made from petroleum coke and coconut shell had
NDMA sorption capacities of 24 mg/g when batch experiments were performed over 24 hours in
deionized water, compared to 17 mg/g demonstrated for zeolites. Modification of the surface by
heat treatment and coating with titanium dioxide nanoparticles increased sorption capacity
Fleming et al. (1996) showed a similar superiority of activated carbon over zeolites. Batch
studies performed with ground water showed 99 percent removal of NDMA using activated
carbon, compared to 15 to 20 percent removal using zeolite, silica or XAD resins. Addition of
silica to zeolite showed little removal of NDMA, but when the combined silica and zeolite were
treated with copper, removal increased to 26 percent. The authors concluded that the best
sorbents for NDMA are carbonaceous resins, such as Ambersorb 572 and 563, both of which
exhibited 99 percent removal of 100 |ig/L NDMA spiked into ground water after 1 hour
(Fleming et al., 1996). Although removal of NDMA by activated carbon has been proven to be
very efficient, it is important to note that the studies presented above were batch experiments,
and other experiments in full-scale facilities have shown less promising results. It is likely that
removal rates will depend on the specific type of adsorbent used, as well as water quality
parameters.
7.3.3 Membrane Filtration
As discussed in Section 7.2.1.5, membrane filtration can remove nitrosamine precursors. It may
be able to remove nitrosamines as well. As membranes are typically placed before final
disinfection in the treatment train, they would not remove nitrosamines formed during final
disinfection. They could, however, remove nitrosamines entering with the raw water or formed
during pre-disinfection with chlorine or another oxidant such as ozone.
Nitrosamine removal by membranes has been studied at several wastewater treatment plants. For
example, an advanced wastewater treatment facility in Southern California using MF and RO
was studied to discern the effects of membrane treatment on NDMA removal. MF was not
effective in removing NDMA. In fact, NDMA increased as a result of the chlorination installed
ahead of the filters in order to prevent membrane fouling. RO, however, exhibited removal
efficiencies of between 24 and 56 percent (Plumlee et al., 2008). Additional experiments
performed with deionized water showed that NDMA was removed by RO in the range of 54 to
70 percent, depending on the membrane used. Nitrosamines of higher molecular weight were
removed more effectively (Steinle-Darling et al., 2007). A strong correlation was found between
the molecular weight of the tested nitrosamine and rejection rates. The authors tested A'-nitroso-
n-dibutylamine (NDBA), NDEA, NDMA, A'-nitrosomethyl ethyl amine (NMEA), N-
nitrosopiperidine (NPIP), A'-nitroso-n-dipropyl amine (NDPA) and NPYR. Nitrosamines
exhibited higher removal in the order: NDBA = NDPA = NPIP > NDEA > NPYR > NMEA >
Six-Year Review 3 7-13
Technical Support Document for Nitrosamines
December 2016
-------
NDMA. When alginate was spiked into the influent to simulate biofouling, a drop in NDMA
rejection from 55 to 38 percent rejection was observed (Steinle-Darling et al., 2007).
Khan and McDonald (2010) studied RO rejection at an advanced wastewater treatment facility in
Australia. NDMA was poorly removed and variable, with 5th and 95th percentile values of 26
percent and 35 percent, respectively. As with other studies, nitrosamines of higher molecular
weight had better removal rates. NDEA was rejected with 5th and 95th percentile values both
around 91 percent. Removal of NDPA was higher, with 5th and 95th percentile values of 97
percent and 98 percent (Khan and McDonald, 2010). Another advanced water treatment facility
in Australia averaged 10 percent removal of NDMA by RO (Poussade et al., 2009). Miyashita et
al. (2009) also found rejection by membranes correlated to molecular weight for both NF and RO
membranes. They performed bench-scale tests with polyamide thin-film composite membranes
in deionized water buffered at pH 7 to study the rejection of NDMA, NDEA, NMEA, NDPA,
NDBA and NPYR. Rejection of nitrosamines by NF membranes varied from 9 to 55 percent.
Rejection by RO membranes ranged from 54 to 97 percent (Miyashita et al., 2009).
Hatzinger et al. (2011) studied NDMA removal in a membrane bioreactor. The bioreactor
combines biological reactions with membrane filtration. The bench-scale study found stable
NDMA removal of 99.95 percent over a 70-day period. Cessation of biological activity by
addition of trichloroethylene resulted in an increase in NDMA in the effluent, indicating that
biological degradation was an important component of the overall removal. Chon et al. (2015)
examined the effectiveness of a laboratory scale system for wastewater reclamation involving a
membrane bioreactor and NF membrane. They found the removal efficiency in the membrane
bioreactor was 73-84 percent for NDMA, 58-61 percent for NPYR, 76-77 percent for NDEA,
57-76 percent for NPIP, 45-57 percent for NMOR, 75-78 percent for NDBA and 53-54 percent
for NDPhA. Removal by NF membranes correlated relatively well with molecular size and was
8-59 percent for NDMA, 38-60 percent for NPYR, 43-75 percent for NDEA, 39-72 percent for
NPIP, 51-74 percent for NMOR, 61-71 percent for NDBA and 69-80 percent for NDPhA.
Fujioka et al. have done several experiments involving removal of NDMA and NMEA by NF
and RO membranes. They found removal highly variable—from 8 to 82 percent for NDMA and
23 to 94 percent for NMEA. For nitrosamines with higher molecular weights, removal was more
consistent, with 90 percent removal for NDPA and NDBA (Fujioka et al. 2013a). For the
nitrosamines with lower molecular weights, removal was correlated with membrane
permeability; an RO membrane designed for boron removal gave the best removal of
nitrosamines, with up to 71 percent removal of NDMA. Their work also showed the importance
of membrane permeability for removal of low molecular weight nitrosamines, showing that
membrane fouling could increase NDMA removal (Fujioka et al., 2013b) and that cleaning the
membranes could actually reduce removal (Fujioka et al., 2014). The same team also examined
the rejection of 7 nitrosamines (including NDMA, NDEA, NMEA, NDPA and NPYR) using
hollow fiber cellulose triacetate RO membranes as a substitute for traditional polyamide
membranes. They found rejection rates ranging from 25 percent (for NDMA) to 78 percent (for
NPIP). Rejection appeared to be influenced by both molecular size and hydrophobicity (Fujioka
et al., 2015).
7.3.4 Metal Catalysis
A number of metal catalysts have been examined for the catalytic reduction of NDMA and other
nitrosamines. Nitrosamine reduction can produce the corresponding amine as a product;
Six-Year Review 3 7-14
Technical Support Document for Nitrosamines
December 2016
-------
therefore, if the reaction does not go to completion, nitrosamines can re-form if the product water
is exposed to an oxidant. Iron and nickel-enhanced iron column flow-through reactors were
found to be capable of destroying NDMA at half-lives of 13 hours and 2 minutes, respectively,
but accumulation of surface oxides caused rapid decreases in rates (Gui et al., 2000;
Odziemkowski et al., 2000). Conversion of nitrosamines to their respective amines was achieved
by porous nickel catalysts under a hydrogen atmosphere. Catalyst concentrated at 500 mg/L in
deoxygenated water buffered at pH 7 under 1 atm of hydrogen pressure gave rate constants of
0.51/min, 0.49/min, 0.42/min, 0.28/min and 0.26/min for NDMA, NDPhA, NDEA, NDPA and
NDBA, respectively. In these experiments, catalysts were sensitive to matrix effects, as nitrate
and sulfide severely inhibited reactivity. Calcium, magnesium, chloride, sulfate, bicarbonate and
NOM decreased rates by a factor of 2.2. However, the use of hydrogen gas and pyrophoric
catalysts at water utilities poses a safety hazard. Additionally, nickel and aluminum leaching may
occur (Frierdich et al., 2007).
Lee et al. (2005a) achieved successful photocatalytic degradation of NDMA by titanium dioxide
(TiCh); however, experiments were performed under conditions far removed from those found at
treatment plants. Davie et al. (2008) found that addition of indium to 5 percent palladium on
aluminum oxide transformed NDMA to DMA and ammonia. An iridium loading of 1 percent
transformed NDMA at a rate of 0.25/hour. Although this technology exhibits favorable kinetics,
the catalyst is prone to sulfide poisoning and may be cost-prohibitive.
Davie et al. (2006) studied a number of metals for potential catalytic properties. Iron, nickel,
nickel-enhanced iron and magnesium resulted in NDMA half-lives of 533 ±218 hours, 8.4 ± 2.2
hours, 107±2 hours and 990 ± 220 hours, respectively, for 10 mg/L of metal. Metal-catalyzed
reduction of NDMA by hydrogen gas resulted in half-lives of 6.0 ± 0.4 hours, 1.0 ± 0.1 hours
and 0 hours for palladium, copper-enhanced palladium and copper, respectively. All experiments
were performed in deionized water. Thus, the feasibility of these techniques at PWSs is unknown
(Davie et al., 2006). However, studies performed using a ground water-soil matrix showed that
nickel and zero-valent iron were ineffective catalysts (Schaefer and Fuller, 2007).
7.3.5 Sunlight Photolysis
Natural sunlight has been found to be effective for nitrosamine removal. When nitrosamines
(NDBA, NDEA, NDMA, NDPA, NMEA, NPYR, NPIP and NMOR) in organic-free water were
exposed to natural sunlight exhibiting intensity ranging from 1150 to 1300 W/m2, 99 percent of
total nitrosamines were photolyzed in one hour (Chen et al., 2010). Removal for the various
nitrosamines corresponded to half-lives ranging from 8 to 10 minutes, with a rate constant of
approximately 4.9/hour. Indoor tests using simulated sunlight at 1325 W/m2 exhibited slightly
faster photolysis with half-lives ranging from 3 to 9 minutes. Nitrosamines with cyclic side
groups (i.e., NMOR, NPYR and NPIP) had the highest rate constants of the nitrosamines,
whereas the three nitrosamines with only methyl and/or ethyl side groups had the lower rate
constants. When nitrosamines were spiked into water containing secondary effluent from a
wastewater facility, first-order rate constants were approximately 50 percent less than those for
nitrosamines spiked into organic-free water, indicating that the presence of NOM may decrease
the performance of UV photolysis (Chen et al., 2010).
Direct photolysis studies were carried out in deionized water using a photosimulator that emitted
light with wavelengths between 290 and 800 nm at 765 watts per square meter (W/m2) (Plumlee
and Reinhard, 2007). The authors calculated half-lives of less than 20 minutes for all
Six-Year Review 3 7-15
Technical Support Document for Nitrosamines
December 2016
-------
nitrosamines tested (NDBA, NDEA, NDMA, NDPA, NMEA, NPYR and NPIP). Methylamine
and nitrite were the dominant products of NDMA degradation. It is possible these products could
be reoxidized to re-form the nitrosamine. Addition of dissolved organic matter (DOM) led to a
decrease in the photodecay rate due to light screening. Further decreases in efficiency are seen
with increasing depth of water. For example, if the depth of the water is reduced to 10 cm from 1
m the half-lives are reduced by a factor of approximately ten (Plumlee and Reinhard, 2007).
Organic-free water buffered at pH 7.2 and amended with 10,000 ng/L NDMA was exposed to
sunlight at an intensity of 1150-1300 W/cm2 (Chen et al., 2010). NDMA was rapidly photolyzed
with a half-life of approximately 8 to 10 minutes. The efficiency of NDMA photolysis was
shown to be sensitive to the water matrix. Experiments done in biologically treated wastewater
effluent and filtered effluent exhibited first-order rate constants 50 percent and 11 percent,
respectively, below those observed in organic-free water (Chen et al., 2010). At the Orange
County Water District of California, shallow sunlit basins with residence times of approximately
1 day resulted in removal of approximately 50 percent of NDMA (Mitch et al., 2003b).
7.3.6 UV Photolysis
Although UV doses required for NDMA degradation are approximately an order of magnitude
higher than those typically used for disinfection purposes, the high efficiency of UV makes it an
attractive option for removal of nitrosamines at drinking water and wastewater utilities (Mitch et
al., 2003b). NDMA exhibits strong absorption bands at 228 and 332 nm, resulting in the
breakdown of the nitrogen-nitrogen bond (Liang et al., 2003). Placement of the UV units after
disinfection may facilitate removal of nitrosamines formed during the disinfection process.
Lee et al. (2005b, 2005c) investigated NDMA photolysis using a low-pressure mercury lamp
emitting at 253.7 nm. Photolysis experiments were performed in phosphate-buffered distilled
water containing 1 mM NDMA. The authors found the efficiency of photolysis to be dependent
on DO; decay rates under oxygen saturation conditions were higher than nitrogen saturation
conditions. This phenomenon only occurred at light irradiation >300 nm. Quantum yields (the
number of molecules reacting per photon) were found to be constant, at 0.31, regardless of pH.
However, product yields of DMA and methylamine were dependent on pH, with methylamine
and nitrate dominating at pH >9.
In an ambient water study, photolysis of NDPA spiked into lake water exhibited fast degradation
independent of pH and initial concentration (Berkowitz, 2008; Sacher et al., 2008). Products
were propylamine and dipropylamine. Nitrosamines such as NDMA and NDEA also converted
to their parent amines upon UV treatment with a high-pressure mercury lamp (Berkowitz, 2008;
Sacher et al., 2008).
Shah et al. (2013) found that UV light at an intensity of 1000 mJ/cm2 achieved 90 percent
removal of nitrosamines from a pilot plant treating amine-based solutions from carbon capture
and storage.
Xu et al. (2008) found that NDEA can be degraded in deionized water by a low-pressure
mercury lamp with an emission at 253.7 nm (1000 |iW/cm2), NDEA degraded to below the
detection limit (DL) in 20 minutes at an initial concentration of 0.01 mM (1.02 mg/L) in
deionized water at pH 6. Reaction rates decreased slightly with increasing pH. To investigate the
effect of NOM on degradation efficiency, humic acid was added at various concentrations.
Six-Year Review 3 7-16
Technical Support Document for Nitrosamines
December 2016
-------
Addition of 2.9 mg/L humic acid decreased the degradation rate by approximately a factor of 2
(from 0.74/min to 0.40/min). Degradation products were methylamine, DMA, DEA and nitrite
(Xu et al., 2008).
Xu et al. (2009a) investigated the photolysis of NPYR and NPIP. Deionized water was spiked
with NPYR or NPIP to achieve a final concentration of 1 |iM (100 |j,g/L and 114 |j,g/L
respectively). Solutions were irradiated with a low-pressure mercury lamp (8 W, 1000 (j,W/cm2).
Ninety-nine percent removal was observed for both nitrosamines after 5 minutes of irradiation.
While pH had little effect on NPYR, it had a large effect on NPIP degradation; when the solution
pH was increased from 3.1 to 10.5, the reaction rate constant decreased by 95 percent. Addition
of NOM had little effect on degradation efficiency for both nitrosamines. NPYR and NPIP
degradation products were similar, as both produced a mixture of aliphatic amines, nitrate and
nitrite.
Genuino et al. (2011) investigated the use of metal photocatalysts to enhance the
photodegradation of NDMA. Their lab-scale experiments found that a platinum-manganese
catalyst and an amorphous manganese oxide resulted in 95 percent and 100 percent removal of
7.4 mg/L NDMA within 3 hours of irradiation.
The intensities required for UV photolysis of nitrosamines are greater than those typically used
for disinfection. Another consideration is that the byproducts of UV photolysis can react with
disinfectants to re-form NDMA. For example, although irradiation of 100 |iM (7.4 mg/L)
NDMA in distilled water by a low-pressure mercury lamp provided 99 percent removal within
20 minutes, subsequent application of 1 mg/L chlorine resulted in formation of 51.8 |ig/L
NDMA (Xu et al., 2009b). Nawrocki and Andrzejewski, (2011) postulate that the re-formation is
a result of DMA and nitrite being formed as products of the photolysis of NDMA.
Using another oxidant with the UV to form an advanced oxidation process (AOP) has been
found to lower NDMA re-formation, probably through oxidation of nitrite and DMA.
Regeneration of NDMA was less than 1 (J,g/L, compared to the initial concentration of 7.4 mg/L,
upon application of UV combined with 6.6 mg/L ozone (Xu et al., 2009b). AOPs, such as
UV/ozone and UV/peroxide, do not enhance NDMA destruction relative to UV alone, as the
decay is primarily driven by photolysis, but rather inhibit its re-formation (Kruithof et al., 2007;
Swaim et al., 2008). Although the chemistry is complex and involves solvated electrons,
hydroxyl radicals and hydrogen atoms, the ultimate goal is to decrease DMA and nitrite
byproducts of UV treatment so as to preclude NDMA re-formation (Liang et al., 2003; Sharpless
and Linden, 2003). However, it has been demonstrated that high doses of peroxide can inhibit
NDMA photolysis by competing for light absorption (Kruithof et al., 2007; Swaim et al., 2008).
7.3.7 Advanced Oxidation Processes (AOP)
AOPs (e.g., UV/ozone, UV/hydrogen peroxide and ozone/hydrogen peroxide) have the common
feature of forming hydroxyl radicals to oxidize contaminants. Their effectiveness in removing
nitrosamines will depend on their placement in the treatment train. The reaction rates of radical
species with nitrosamines are high, such that the use of AOPs for nitrosamine removal may be
feasible.
Hydroxyl radicals and hydrated electrons were introduced into nitrogen-sparged deionized water
by pulse radiolysis in a study conducted by Mezyk et al. (2006). Hydroxyl radical rate constants
Six-Year Review 3 7-17
Technical Support Document for Nitrosamines
December 2016
-------
for NMEA and NDEA were 4.95 x 108 and 6.99 x 108 M/sec, respectively. Reaction of hydroxyl
radicals with nitrosamines led to a hydrogen abstraction from the alkyl group leading to
subsequent nitrosamine degradation. Although reactions with hydrated electrons were faster, the
electron-nitrosamine adduct led to re-formation of the parent nitrosamine (Mezyk et al., 2006).
Xu et al. (2010) compared UV and UV/ozone AOPs for NDEA removal using a low-pressure
mercury lamp. Addition of 0.1 mM (10.2 mg/L) NDEA to deionized water buffered at pH 6 and
containing 6.64 mg/L ozone led to only 10 percent removal of the nitrosamine. NDEA removal
reached 99 percent when exposed to UV radiation over 10 minutes, and a similar removal rate
was seen for combined UV/ozone. While the addition of ozone did not enhance the rate of
NDEA decay, it did affect the product ratio. The yield of DEA and nitrite produced in the
UV/ozone process was less than that in the UV process. This difference ranged from 22 percent
to 54 percent, depending on the pH, with higher difference at lower pH (Xu et al., 2010).
Hydroxyl radicals formed through AOPs react quickly with nitrosamines, with rate constants on
the order of 108to 109 moles per liter per second (Landsman et al., 2007). Nitrosamines with
higher molecular weights have higher rate constants. Reaction rates determined by Landsman et
al. (2007) may be overestimated for drinking water, because experiments were performed in
deionized water containing high concentrations of nitrosamines (1 mM). In real waters, reaction
with NOM and carbonate consumes most hydroxyl radicals (Landsman et al., 2007).
A pilot study (Liang et al., 2003) was performed with Colorado River water and Southern
California ground water to investigate the degradation of NDMA via pulsed-UV processes,
including UV/hydrogen peroxide as one study condition. A pulsed-UV dose of 5.2 kilowatt-
hours per 1,000 gallons resulted in an NDMA pseudo-first-order rate constant of 4.1/min for
ground water and 1.4/min for river water. The discrepancy between the rates suggests
competition for light between NDMA and organic constituents in the water matrix. NDMA
concentrations did not change when UV/hydrogen peroxide was substituted for UV alone.
NDMA concentrations more than doubled, however, when water treated with UV was
chlorinated (Liang et al., 2003). In a separate study, two water treatment plants in Asia exhibited
similar influent NDMA levels, but the plant using peroxide/UV advanced oxidation had lower
effluent NDMA levels (<5 ng/L) than the plant using UV alone (>15 ng/L) (Valentine et al.,
2005). A field study performed at the Bundama advanced water treatment plant in Australia
demonstrated a 1.6 log removal of NDMA by UV/peroxide treatment, with most NDMA values
below 5 ng/L after treatment with a dose of 4-5 mg/L peroxide (Poussade et al., 2009).
In batch experiments performed in deionized water buffered at pH 7, the addition of excess
ozone at concentrations of up to 160 |iM oxidized 13 percent of the initial 74 |j,g/L of NDMA,
whereas 85 percent of NDMA was oxidized by 40 |iM ozone combined with 80 |iM peroxide.
Hydroxyl scavengers, such as carbonate, reduced the rate and percent of NDMA oxidation. For
example, higher oxidant doses were required for the same levels of NDMA oxidation in Lake
Zurich water, with an alkalinity of 2.6 mM as bicarbonate (Lee et al., 2007c). Conventional
ozonation with 160 |iM ozone led to 25 percent NDMA oxidation, whereas the AOP
ozone/peroxide resulted in 55 percent NDMA oxidation. Reaction products consisted of
methylamine and ammonia, with ammonia increasing with hydroxyl radical production.
Although this AOP was effective, its use (as, presumably, is the case with the use of any AOP
involving ozone) may increase bromate such that it exceeds the drinking water maximum
contaminant level of 10 |ig/L. For example, the maximum dose of ozone that can be applied
Six-Year Review 3 7-18
Technical Support Document for Nitrosamines
December 2016
-------
without exceeding 10 |ig/L bromate in Lake Zurich water is approximately 80 |iM, which
corresponds to 39 percent NDMA oxidation (Lee et al., 2007c).
7.3.8 Electrochemical Techniques
Oxidation and reduction of NDMA by boron-doped diamond film electrodes has been studied by
Chaplin et al. (2010a,b). While rates were sufficient to remove up to 97 percent of 37 mg/L of
NDMA, the process was tested on RO concentrates having much higher NDMA concentrations
than typical of drinking water. The rates would be expected to be slower at lower concentrations.
7.3.9 Biological Techniques
Webster et al. (2013) examined the removal of NDMA in a laboratory-scale biological fluidized
bed reactor. The reactor used Rhodococcus ruber bacteria supported on an activated carbon
media with propane as a food source for the bacteria. The reactor treated contaminated ground
water with initial NDMA concentrations of 10 to 20 |_ig/L. With a 20-minute contact time, the
reactor achieved 90 percent removal of NDMA. Increasing the contact time to 30 minutes and
increasing propane and oxygen flow enabled 99 percent removal. Homme and Sharp (2013)
examined laboratory reaction of nitrosamines by the bacteria Rhodococcus jostii grown on
propane and found degradation of NDMA, NDEA, NDPA, NPYR and to a smaller extent,
NMOR. They found the order of reactivity of the different nitrosamines was
NDMA>NDEA>NDPA>NPYR>NMOR. Wang et al. (2015d) examined the effects of
nitrosamine digestion on a biofilter's microbial community. Exposure of the microbial
community to low levels of nitrosamines led to an increase in proteobacteria relative to other
phyla. After assimilation for 10 days, biofilters removed between 8 and 40.7 percent of various
nitrosamine species (including all six that are the subject of this document), with smaller
nitrosamines being removed more efficiently. The research team was able to isolate a species
Rhodococcus cercidiphylli that was able to reduce 5 of 9 nitrosamines studied. In separate trials
using the isolated bacteria, the following removals were obtained after 10 days: NDMA (85.4
percent), NDPA (78.8 percent), NDEA (47.7 percent), NPYR (48.5 percent) and NDBA (38.1
percent).
7.4 Summary
A variety of strategies are available to control and/or reduce nitrosamine concentrations in
drinking water. Each strategy has its advantages and limitations. Exhibit 7.2 provides a summary
of the treatment strategies discussed in this chapter along with ranges of removal efficiencies and
other potential benefits and issues identified for each technology.
Source water management can be an effective strategy for controlling nitrosamines in drinking
water. Longer distances between wastewater discharges and drinking water intakes will allow
natural attenuation processes such as photolysis and biological degradation of organic matter
from wastewater discharges (including nitrosamine precursors) to occur. While source water
management can be an effective way to control nitrosamine formation, it can be a challenging
undertaking because it involves the cooperation of groups outside the control of the drinking
water utility.
Six-Year Review 3 7-19
Technical Support Document for Nitrosamines
December 2016
-------
Removing nitrosamine precursors during drinking water treatment is also a viable and sometimes
advantageous strategy. A significant advantage of tighter precursor control is that less chlorine is
required for primary and secondary disinfection.
Achieving nitrosamine precursor removal through sorption may be difficult because of the
hydrophilicity of NDMA precursors. Sorption can achieve substantial removal of NDMA
precursors under conventional flow-through operating conditions, but it requires large doses of
activated carbon. Customized sorbents may prove more effective but are likely to be more costly.
Current data gaps involve the properties and performance of specific sorbents and results from
full-scale studies of nitrosamine precursor removal.
Biodegradation of NDMA has been shown to occur at widely varying rates depending on site-
specific conditions. Biological reaction often requires adsorption of the compound for the
process to occur efficiently. Also, larger nitrosamines do not degrade as quickly as NDMA, so
this treatment technique may not be effective for all nitrosamines. On the other hand, riverbank
filtration has been shown to provide precursor removal in several cases. Data gaps involve the
effectiveness of biological filtration for reducing nitrosamine formation and the applicability of
riverbank filtration and biologically active filtration for nitrosamine reduction. If a drinking
water treatment plant decides to add biological treatment, downstream filtration and disinfection
may be required to prevent sloughed biomass from colonizing the distribution system.
Six-Year Review 3 7-20
Technical Support Document for Nitrosamines
December 2016
-------
Exhibit 7.2: Summary of Removal Efficiencies of Precursors, NDMA and Other Nitrosamines
Technology
Precursor
Removal
Reference(s)
NDMA Removal
Reference(s)
Other Nitrosamine
Removal
Reference(s)
Other
Potential
Benefits
Issues
Enhanced
Coagulation
10-18% (full-scale
and lab-scale
NDMA FP
removal)
30-45% (full-scale
and lab-scale
DON removal)
Sacher et al., 2008
Mitch et al., 2009
Westerhoff et al.,
2006
Dotson and
Westerhoff, 2009
Liao et al. 2015b
<7%
(lab-scale NDMA
FP removal)
Sacher et al.,
2008
N/A
N/A
Often an
existing
process; Also
removes
THM and
HAA
precursors
Not very effective for
precursor or
nitrosamine removal;
Must be careful with
selection of coagulation
polymers
Sorption
29-90% (using
PAC lab-scale FP
removal)
54-84%
(full-scale GAC
FP removal)
>90% (bench
scale removal of
select amines)
Sacher et al., 2008
Hanigan et al., 2012
Chu et al., 2015
Krasner et al., 2015
Wu et al., 2015
Chen et al., 2015b
AC: 17-99%1 (lab-
scale NDMA
removal)
Zeolite: 15-26%
(bench-scale
NDMA removal)
Carbonaceous
resins: 99%
(bench-scale
NDMA removal)
Activated carbon
nanoparticles (lab-
scale NDMA
removal) 20%
Fleming et al.,
1996
Wang et al.,
2013b
N/A
N/A
Also removes
THM and
HAA
precursors
and other
contaminants
Removal is dependent
on sorbent, CT and
concentration;
Removal is dependent
on the structure of the
compound and on
water quality
parameters;
Sorbent may have to
be regenerated
frequently;
Desorption from GAC
may be problematic
Biological
Processes
67%
(DW full-scale
NDMA FP
removal)
59%, 55%, >70%
(DW pilot scale
removal of NDMA
FP, NDEA FP and
NPYR FP,
respectively)
Mitch et al., 2009
Liao et al., 2015b
90-99%
(laboratory scale
NDMA removal)
85.4% (NDMA
removal
laboratory using
isolated bacteria)
Webster et al.,
2013
Wang et al.
2015d
78.8%,
47.7%,
48.5%,
38.1 % (respective
removal of NDPA,
NDEA, NPYR and
NDBA using
laboratory isolated
bacteria)
Wang et al.,
2015d
Also removes
other A/-DBP
precursors
Limited drinking water
information available;
Rate of degradation is
too slow for drinking
water treatment (Health
Canada 2011)
Six-Year Review 3
Technical Support Document for Nitrosamines
7-21
December 2016
-------
Technology
Precursor
Removal
Reference(s)
NDMA Removal
Reference(s)
Other Nitrosamine
Removal
Reference(s)
Other
Potential
Benefits
Issues
Riverbank
Filtration
50-93%
(full-scale removal
of precursors)
Sacher et al., 2008
N/A
N/A
N/A
N/A
Also removes
other A/-DBP
precursors
Limited drinking water
information available;
Rate of degradation is
too slow for drinking
water treatment (Health
Canada 2011)
Slow Sand
Filtration
23-83%
(lab-scale
precursor
removal)
Krasner et al., 2008
N/A
N/A
N/A
N/A
Good
removal at
full scale
Requires large surface
area
Membrane
Filtration
MF: 12-95%
(removal of
precursors)
RO: 98-99%
Deeb et al., 2006
Miyashita et al.,
2009
MF/UF: negligible
NF: 8-59% (WW
lab scale)
RO: 25 - 84%
(WW lab and full
scale)
Plumlee et al.,
2008, Steinle-
Darling et al.,
2007
Fujioka et al.,
2013a
Chon et al.
2015
Fujioka et al.
2015
23-97% (DW lab
scale)
(NDEA, NMEA,
NDPA, NDBA and
NPYR)
Miyashita et
al., 2009
Fujioka et al.,
2013a
Fujioka et al.,
2015
Chon et al.,
2015
Also removes
THM and
HAA
precursors
and other
contaminants
Results are membrane-
specific;
Typical membrane
placement in process
would not remove
nitrosamines formed
during final disinfection;
To avoid membrane
fouling, additional
chlorine may be added,
which can result in
increased DBP
formation
AOPs
(UV/Hydroge
n Peroxide or
UV/Ozone)
N/A
N/A
25-97 (lab scale
and full scale)
Poussade et
al., 2009
Lee et al.,
2007c
99% (lab scale)
(NDEA)
Xu et al., 2010
May remove
other
contaminants
, prevents
NDMA re-
formation
Can contribute to the
formation of other
DBPs and can be
expensive
Oxidation by
Ozone,
Permanganat
e, or Chlorine
Dioxide
20-94% (lab-
scale NDMA FP
removal)
88% (full-scale
NDMA FP
removal when
combined with
biological GAC
filters)
Lee et al., 2007b
Sacher et al., 2008
Liao et al., 2014
McCurry et al., 2015
Krasner et al., 2015
Wang et al., 2015c
N/A
N/A
N/A
N/A
May remove
other
contaminants
May form other DBPs;
may oxidize DMA to
NDMA
Six-Year Review 3
Technical Support Document for Nitrosamines
7-22
December 2016
-------
Technology
Precursor
Removal
Reference(s)
NDMA Removal
Reference(s)
Other Nitrosamine
Removal
Reference(s)
Other
Potential
Benefits
Issues
Pre-
Oxidation
with Chlorine
17-93%
(lab scale and full
scale)
Charrois and
Hrudey, 2007
Chen and Valentine,
2008
McCurry et al. 2015
Krasner et al., 2015
N/A
N/A
N/A
N/A
Simple to
use;
May remove
other
contaminants
May form chlorination
DBPs (THMs and
HAAs)
Sunlight
Photolysis
25%
(lab-scale NDMA
FP removal)
Chen and Valentine,
2008
50-99% (full scale
and lab scale)
Mitch et al.,
2003b
Chen et al.,
2010
Plumlee and
Reinhard,
2007
99% (lab scale)
(NDBA, NDEA,
NDMA, NDPA,
NMEA, NPYR,
NPIP)
Chen et al.,
2010
Plumlee and
Reinhard,
2007
Effective and
inexpensive
Performance
decreases in presence
of NOM;
Byproducts of
photolysis can react
with chlorine to re-form
nitrosamines;
Longer contact time
needed for natural
sunlight exposures
UV
Photolysis
54+% (MP UV
pilot scale NDMA
FP removal)
29+% (LP UV pilot
scale NDMA FP
removal)
McCurry et al,. 2015
>99%, or 95-
100% (lab scale)
Mitch et al.
(2003b)
Lee et al.
(2005b,c)
Genuino et al.
(2011)
Xu et al.
(2009b)
>99%
(lab scale)(NDEA,
NDPA, NPYR,
NPIP)
Berkowitz,
2008
Sacher et al.,
2008
Xu et al., 2008
Xu et al.
(2009a)
Can provide
additional
disinfection
and may
remove other
contaminants
Requires higher doses;
Byproducts of UV
photolysis can react
with chlorine to re-form
nitrosamines
Note:
1) These Activated Carbon (AC) sorption trials were batch experiments, far removed from normal flow-through operating conditions at water facilities.
Six-Year Review 3
Technical Support Document for Nitrosamines
7-23
December 2016
-------
The small size of NDMA molecules makes MF and UF less effective for nitrosamine control.
RO has been shown to eliminate almost all NDMA precursors. NF and RO can also remove
nitrosamines, with nitrosamines of higher molecular weight exhibiting higher removal rates.
Therefore, not all nitrosamines will be removed with equal effectiveness. NDMA, however, has
the lowest molecular weight of the nitrosamines, so a membrane that can remove it will likely be
adequate for all nitrosamines.
Research has shown that use of pre-oxidants can also be an effective strategy for controlling
nitrosamine formation. Pre-oxidants, however, will form DBPs of their own. Pre-oxidation with
ferrate may be difficult due to the highly unstable nature of potassium and sodium salts. Ferrate
has not been extensively used in drinking water treatment. Data gaps exist regarding the use of
ferrate for nitrosamine control. Permanganate is an easy-to-use pre-oxidant. Chlorine is already
available at most drinking water utilities. However, the pre-oxidation dose and contact time must
be monitored closely so that maximum contaminant levels for other DBPs such as THMs and
MAAs are not exceeded. While pre-oxidation of NDMA precursors could be a practical
treatment technique for utilities to control NDMA formation, waters with high concentrations of
DMA or other precursors may see less net reduction because of the simultaneous formation of
NDMA by oxidation of these precursors. Data gaps exist regarding the optimization of oxidation
processes for nitrosamine control, including examining multiple categories of DBP formation by
oxidation.
Removal of nitrosamines after they have formed is often more difficult than controlling source
water or removing precursors. Generally, nitrosamine removal requires more advanced
technologies. Degradation of nitrosamines by sunlight, however, has been shown to be effective.
It is dependent on water depth, water quality and available sunlight and may not be effective for
all waters. UV photolysis is effective for reducing all nitrosamines and produces few regulated
DBPs and is less dependent on water depth or available sunlight. However, it requires higher
doses. An advantage is that UV photolysis may also be effective in removing other contaminants.
Re-formation of NDMA from disinfection with chlorine or chloramines can be a limiting factor,
although a degree of re-formation generally results in much lower concentrations than the
original nitrosamine concentration. Data gaps exist regarding ways to reduce nitrosamine re-
formation.
Control of nitrosamines can be achieved through a wide array of possible strategies ranging from
traditional technologies to advanced treatment processes. Most treatment technologies discussed
in this chapter are not completely new. Many of these technologies are also discussed in the Six
Year Review 3 Technical Support Document for Disinfectants and Disinfection By-Products
Rules (USEPA, 2016a) and the Simultaneous Compliance Guidance Manual for the Long-Term 2
and Stage 2 DBP Rules (USEPA, 2007d).
Six-Year Review 3 7-24
Technical Support Document for Nitrosamines
December 2016
-------
8 References
Aiub, C.A., L.F.R. Pinto, and I. Felzenszwalb. 2003. A-Nitrosodi ethyl amine mutagenicity at low
concentrations. Toxicology Letters. 145(1): 36-45.
Aiub, C.A., J.L. Mazzei, L.R.F. Pinto, and I. Felzenszwalb. 2006. Evaluation of nitroreductase
and acetyltransferase participation in A;-ni trosodi ethyl am i ne genotoxicity. C hem i co-Bi ol ogi cal
Interactions. 161(2): 146-154.
Aleksandrov, V.A. 1968. Blastomogenic effect of dimethylnitrosamine on pregnant rats and their
offspring. Nature. 218: 280 (as cited in Health Canada, 2010).
Althoff, J. and C. Grandjean. 1979. In vivo studies in Syrian golden hamsters: A transplacental
bioassay of ten nitrosamines. National Cancer Institute Monograph. 51: 251-255.
Althoff, J., P. Pour, C. Grandjean, and S. Marsh. 1977. Transplacental effects of nitrosamines in
Syrian hamsters. Zeitschrift fur Krebsforschung und Klinische Onkologie. 90: 79-86 (as cited in
IPCS, 2002).
Amacher, D.E. and S.C. Paillet. 1983. The activation of procarcinogens to mutagens by cultured
rat hepatocytes in the L5178Y/TK mutation assay. Research/Environmental Mutagenesis and
Related Subjects. 113:77-88 (as cited in ATSDR, 1989).
Amelizad, S., K.E. Appel, M. Schoepke, C.S. Riihl, and F. Oesch. 1989. Enhanced
demethylation and denitrosation of N-nitrosodimethylamine by human liver microsomes for
alcoholics. Cancer Letters. 46(1): 43-49.
American Conference of Governmental Industrial Hygienists (ACGIH). 1991. Documentation of
the TLVand biological exposure indices. 6th ed., vol. 2. Cincinnati, Ohio.
Anderson, L.M., L.J. Priest, and J.M. Budinger. 1979. Lung tumorigenesis in mice after chronic
exposure in early life to a low dose of dimethylnitrosamine. Journal of the National Cancer
Institute. 62(6): 1553-1555 (as cited in CalEPA, 2006).
Anderson, L.M., G.W. Harrington, H.M. Pylypiw, Jr., A. Hagiwara, and P.N. Magee. 1986.
Tissue levels and biological effects of A-nitrosodi methyl amine in mice during chronic low or
high dose exposure with or without ethanol. Drug Metabolism and Disposition. 14(6): 733-739
(as cited in IPCS, 2002; CalEPA, 2006; ATSDR, 1989).
Anderson, L.M., A. Hagiwara, R.M. Kovatch, S. Rehm, and J.M. Rice. 1989. Transplacental
initiation of liver, lung, neurogenic and connective tissue tumors by iV-nitroso compounds.
Toxicological Sciences. 12: 604-620 (as cited in CalEPA, 2006; Health Canada, 2010).
Anderson, L.M., J.P. Carter, C.L. Driver, D.L. Logsdon, R.M. Kovatch, and A. Giner-Sorolla.
1993. Enhancement of tumorigenesis by A-nitrosodi ethyl amine, A-nitrosopyrrolidine and N6-
(methylnitroso)-adenosine by ethanol. Cancer Letters. 68: 61-66.
Andrews, A.W. and W. Lijinsky. 1980. The mutagenicity of 45 nitrosamines in the Salmonella
typhimurium. Teratogenesis, Carcinogenesis, and Mutagenesis. 1(3): 295-303.
Six-Year Review 3 8-1
Technical Support Document for Nitrosamines
December 2016
-------
Andrzejewski, P. and J. Nawrocki. 2007. A-nitrosodi methyl amine formation during treatment
with strong oxidants of dimethylamine containing water. Water Science & Technology. 56(12):
125-131.
Andrzejewski, P. and J. Nawrocki. 2009. A-nitrosodi methyl amine (NDMA) as a product of
potassium permanganate reaction with aqueous solutions of dimethylamine (DMA). Water
Research. 43(5): 1219-1228.
Andrzejewski, P., B. Kasprzyk-Hordern, and J. Nawrocki. 2008. A'-nitrosodi methyl amine
(NDMA) formation during ozonation of dimethylamine-containing waters. Water Research.
42(4-5): 863-870.
Araki, A., M. Muramatsu, and T. Matsushima. 1984. Comparison of mutagenicities of N-
nitrosamines on Salmonella typhimurium TalOO and Escherichia coli Wp2 Uvra/PkmlOl using
rat and hamster liver. Journal of Cancer Research. 75: 8-16.
Agency for Toxic Substances and Disease Registry (ATSDR). 1989a. Toxicological profile for
A'-nitroso-di-n-propyl amine. U.S. Department of Health and Human Services, Public Health
Service. December 1989. Available online at: https://www.atsdr.cdc.gov/ToxProfiles/tpl42.pdf.
ATSDR. 1989b. Toxicological profile for /V-nitrosodimethylamine. U.S. Department of Health
and Human Services, Public Health Service. Available online at:
https://www.atsdr.cdc.gov/toxprofiles/tpl41.pdf.
ATSDR. 2010. Addendum to the toxicological profile for /V-nitroso-di-n-propylamine. U.S.
Department of Health and Human Services. Public Health Service, Agency for Toxic Substances
and Disease Registry. Available online at:
http://www.atsdr.cdc.gov/toxprofiles/n_nitrosodi_n_propylamine_addendum.pdf.
Australia National Health and Medical Research Council (NHMRC). 2013. Australian drinking
water guidelines paper 6 (2011). National water quality management strategy. Version 2.0.
Updated December 2013. Canberra, Australia: National Health and Medical Research Council,
National Resource Management Ministerial Council.
Aze, Y., K. Toyoda, F. Furukawa, K. Mitsumori, and M. Takahashi. 1993. Enhancing effect of
ethanol on esophageal tumor development in rats by initiation of diethylnitrosamine.
Carcinogenesis. 14: 37-40 (as cited in IARC, 2010).
Bansal, A.K., R. Trivedi, G.L. Soni, and D. Bhatnagar. 2000. Hepatic and renal oxidative stress
in acute toxicity of A-nitrosodi ethyl amine in rats. Indian Journal of Experimental Biology. 38(9):
916-920.
Bednar A.J., A.K. Kam, C.R. Mao, D.E. Splichal, D.B. Taggart, and K.H. Coats. 2009.
Determination of low level NDMA in soils. U.S. Army Corps of Engineers. ERDC TN-EQT-09-
01.
Beggs, K.M.H., R.S. Summers, and D.M. McKnight. 2009. Characterizing chlorine oxidation of
dissolved organic matter and disinfection by-product formation with fluorescence spectroscopy
and parallel factor analysis. Journal of Geophysical Research-Biogeosciences. 114: G04001.
Six-Year Review 3 8-2
Technical Support Document for Nitrosamines
December 2016
-------
Berger, M.R., D. Schmahl, and H. Zerban. 1987. Combination experiments with very low doses
of three genotoxic A-nitrosamines with similar organotropic carcinogenicity in rats.
Carcinogenesis. 8(11): 1635-1643.
Berkowitz, B. 2008. Abiotic contaminant transformations in subsurface water. In Contaminant
geochemistry: Transport and fate, pp. 273-293. Berlin: Springer.
Bertram, J.S. and A.W. Craig. 1970. Induction of bladder tumours in mice with
dibutylnitrosamine. British Journal of Cancer. 24: 352-359.
Bhattacharyya, K. 1965. Fetal and neonatal responses to hepatotoxic agents. The Journal of
Pathology and Bacteriology. 90: 151-161 (as cited in ATSDR, 1989).
Bolto, B. 2005. Reaction of chlorine with organic polyelectrolytes in water treatment: A review.
Journal of Water Supply: Research and Technology. 54(8): 531-544.
Brambilla, G., M. Cavanna, A. Pino, and L. Robbiano. 1981. Quantitative correlation among
DNA damaging potency of six A'-nitroso compounds and their potency in inducing tumor growth
and bacterial mutations. Carcinogenesis. 2(5): 425-9.
Brambilla, G., P. Carl, R. Finollo, and L. Sciaba. 1987. Dose-response curves for liver DNA
fragmentation induced in rats by sixteen /V-nitroso compounds as measured by viscometric and
alkaline elution analyses. Cancer Research. 47(13): 3485-3491.
Brendler, S.Y., A. Tompa, K.F. Hutter, R. Preussmann, and B.L. Pool-Zobel. 1992. In vivo and
in vitro genotoxicity of several A-nitrosamines in extrahepatic tissues of the rat. Carcinogenesis.
13(12): 2435-2441.
California Department of Public Health (CDPH). 2013. NDMA and other nitrosamines -
Drinking water issues. Available online at:
http://www.waterboards.ca.gov/drinking_water/certlic/drinkingwater/NDMA.shtml.
California Environmental Protection Agency (CalEPA). 2006. Public health goal forN-
nitrosodimethylamine in drinking water. Office of Environmental Health Hazard Assessment,
Pesticide and Environmental Toxicology Branch. Available online at:
http://oehha.ca.gov/water/phg/pdf/122206NDMAphg.pdf.
Chaplin, B.P., G. Schrader, and J. Farrell. 2010a. Electrochemical destruction of N-
nitrosodimethylamine in reverse osmosis concentrates using boron-doped diamond film
electrodes. Environmental Science & Technology. 44(ll):4264-4269.
Chaplin, B.P., G. Schrader, and J. Farrell. 2010b. Electrochemical oxidation of A-
nitrosodimethylamine with boron-doped diamond film electrodes. Environmental Science &
Technology. 43(21): 8302-8307.
Charrois, J.W.A. and S.E. Hrudey. 2007. Breakpoint chlorination and free-chlorine contact time:
Implications for drinking water/V-nitrosodimethylamine concentrations. Water Research. 41(3):
674-682.
Six-Year Review 3 8-3
Technical Support Document for Nitrosamines
December 2016
-------
Charrois, J.W.A., J.M. Boyd, K.L. Froese, and S.E. Hrudey. 2007. Occurrence of /V-nitrosamines
in Alberta public drinking-water distribution systems. Journal of Environmental Engineering
Science. 6: 103-114
ChemlDPlus. 2010. Toxicology Data Network. Available online at:
http://chem.sis.nlm.nih.gov/chemidplus/. Accessed August 27, 2010.
Chen, Z. and R.L. Valentine. 2006. Modeling the formation of A'-nitrosodi methyl amine (NDMA)
from the reaction of natural organic matter (NOM) with monochloramine. Environmental
Science & Technology. 40(23): 7290-7297.
Chen, Z. and R.L. Valentine. 2007. Formation of A'-nitrosodi methyl amine (NDMA) from humic
substances in natural water. Environmental Science & Technology. 41(17): 6059-6065.
Chen, Z. and R.L. Valentine. 2008. The influence of the pre-oxidization of natural organic matter
on the formation of A'-nitrosodi methyl amine (NDMA). Environmental Science & Technology.
42(14): 5062-5067.
Chen, B. and P. Westerhoff 2010. Predicting disinfection by-product formation potential in
water. Water Research. 44(13): 3755-3762.
Chen, W. and T.M. Young. 2008. NDMA formation during chlorination and chloramination of
aqueous diuron solutions. Environmental Science & Technology. 42(4): 1072-1077.
Chen, W. and T.M. Young. 2009. Influence of nitrogen source on NDMA formation during
chlorination of diuron. Water Research. 43(12): 3047-3056.
Chen, B., S. Nam, P. Westerhoff, S.W. Krasner, and G. Amy. 2009. Fate of effluent organic
matter and DBP precursors in an effluent-dominated river: a case study of wastewater impact on
downstream water quality. Water Research. 43(6): 1755-1765.
Chen, B., W. Lee, P. Westerhoff, S.W. Krasner, and P. Herckes. 2010. Solar photolysis kinetics
of disinfection byproducts. Water Research. 44(11): 3401-3409.
Chen, C., S. Leavey, S.W. Krasner, and I.H. Suffet. 2014. Applying polarity rapid assessment
method and ultrafiltration to characterize NDMA precursors in wastewater effluents. Water
Research. 57: 115-126.
Chen, W., Y. Yang, Y. Wang, C. Li, K. Lin, and J. Lou. 2015a. Effect of molecular
characteristics on the formation of nitrosamines during chloramination of phenylurea herbicides.
Environmental Science Processes and Impacts. 17: 2092-2100.
Chen, C., D. Hanigan, X. Liao, J. Wang, X. Zhang, I.H. Suffet, S.W. Krasner, and P. Westerhoff.
2015b. pH Effect on Nitrosamine Prescursor Removal by Activated Carbon Adsorption. In
Recent Advances in Disinfection By-Products, ed. Karanfil et al, Chapter 10. ACS Symposium
Series.
Chen, C., W. Zhou, and D. Lin. 2015c. Sorption characteristics of n-nitrosodimethylamine onto
biochar from aqueous solution. Bioresource Technology. 179: 359-366.
Six-Year Review 3 8-4
Technical Support Document for Nitrosamines
December 2016
-------
Chhabra, S.K., V.L. Souliotis, J.W. Harbaugh, S.W. Krasnow, A.B. Jones, L.M. Anderson, and
S.A. Kyrtopoulos. 1995. 06-methylguanine DNA adduct formation and modulation by ethanol in
placenta and fetal tissues after exposure of pregnant patas monkeys to N-nitrosodimethylamine.
Cancer Research. 55(24): 6017-6020.
Choi, J. and R.L. Valentine. 2002. Formation of jV-nitrosodimethylamine (NDMA) from reaction
of monochloramine: a new disinfection by-product. Water Research. 36(4): 817-824.
Choi, J. and R.L. Valentine. 2003. A'-nitrosodi methyl amine formation by free-chlorine enhanced
nitrosation of dimethylamine. Environmental Science & Technology. 37(21): 4871-4876.
Choi, J., S.E. Duirk, and R.L. Valentine. 2002. Mechanistic studies of jV-nitrosodimethylamine
(NDMA) formation in chlorinated drinking water. Journal of Environmental Monitoring. 4(2):
249-252.
Chon, K., S.H. Kim, and J. Cho. 2015. Removal of iV-nitrosamines in a membrane bioreactor and
nanofiltration hybrid system for municipal wastewater reclamation: process efficiency and
mechanisms. Bioresource Technology. 170: 490-507.
Chu, W., D. Yao, N. Gao, T. Bond, and M.R. Templeton. 2015. The enhanced removal of
carbanaceous and nitrogenous disinfection by-product precursors using integrated permanganate
oxidation and powdered activated carbon adsorption pretreatment. Chemosphere. 141: 1-6.
Chuang, Y.H., A.Y. Lin, X.H. Wang., and H.H. Tung. 2013. The contribution of dissolved
organic nitrogen and chloramines to nitrogenous disinfection byproduct formation from natural
organic matter. Water Research. 47(3): 1308-1316.
Chung, F.L., T. Tanaka, and S.S. Hecht. 1986. Induction of liver tumors in F344 rats by
crotonaldehyde. Cancer Research. 46(3): 1285-1289.
Chung, J. Y., Y. Yoon, M. Kim, S.B. Lee, H.J. Kim, and C.K. Choi. 2009. Removal of radio N-
nitrosodimethylamine from drinking water by coagulation and powdered activated carbon (PAC)
adsorption. Drinking Water Engineering and Science. 2(2): 49-55.
Clapp, N.K., A.W. Craig, and R.E. Toya, Sr. 1968. Pulmonary and hepatic oncogenesis during
treatment of male RF mice with dimethylnitrosamine. Journal of the National Cancer Institute.
41:1213-1227.
Clapp, N.K., R.L. Tyndall, and J. A. Otten. 1971. Differences in tumor types and organ
susceptibility in BALB/c and Rf mice following dimethylnitrosamine and diethylnitrosamine.
Cancer Research. 31: 196-198.
Cooper, S.W. and R.D. Kimbrough. 1980. Acute dimethylnitrosamine poisoning outbreak.
Journal of Forensic Science. 25(4): 874-882.
Cornwell, D.A., S.W. Krasner, W.A. Mitch, and J.J. Pignatello. 2015. Investigating coagulant aid
alternatives to polyDADMAC polymers. Water Resarch Foundation. Project #4452.
Dai, N and W.A. Mitch 2013. Relative importance of A'-nitrosodi methyl amine compared to total
A'-nitrosamines in drinking water. Environmental Science & Technology. 47: 3648-3656.
Six-Year Review 3 8-5
Technical Support Document for Nitrosamines
December 2016
-------
Dai, N. and W.A. Mitch. 2015. Controlling nitrosamines, nitramines, and amines in amine-based
CO2 Capture Systems with continuous ultraviolet and ozone treatment of washwater.
Environmental Science & Technology. 49(7): 8878-8886
Dai, X., L. Zou, Z. Yan, and M. Millikan. 2009. Adsorption characteristics of N-
nitrosodimethylamine from aqueous solution on surface-modified activated carbons. Journal of
Hazardous Materials. 168(1): 51-56.
Dai, N., A.D. Shah, L. Hu, M.J. Plewa, B. McKague, and W.A. Mitch. 2012. Measurement of
nitrosamine and nitramine formation from NOx reactions with amines during amine-based
carbon dioxide capture for postcombustion carbon sequestration. Environmental Science &
Technology. 46: 9793-9801.
Davie, M.G., M. Reinhard, and J.R. Shapley. 2006. Metal-catalyzed reduction of N-
nitrosodimethylamine with hydrogen in water. Environmental Science & Technology. 40(23):
7329-7335.
Davie, M.G., K. Shih, F.A. Pacheco, J.O. Lickie, and M. Reinhard. 2008. Palladium-indium
catalyzed reduction of /V-nitrosodimethylamine: Indium as a promoter metal. Environmental
Science & Technology. 42(8): 3040-3046.
De Stefani, E., H. Deneo-Pellegrini, J.C. Carzoglio, A. Ronco, and M. Mendilaharsu. 1996.
Dietary nitrosodimethylamine and the risk of lung cancer: a case-control study from Uruguay.
Cancer Epidemiology Biomarkers & Prevention. 5: 679-682 (as cited in IPCS, 2002).
Dean, B.J. and K.R. Senner. 1977. Detection of chemically induced somatic mutations in
Chinese hamsters. Mutation Research/Environmental Mutagenesis and Related Subjects. 46:
403-407 (as cited in Kuroki et al., 1980).
Deeb, R., E. Hawley, D. Sedlak, J. Loveland, and M. Kavanaugh. 2006. Removal and destruction
of NDMA and precursors during wastewater treatment. Proceedings of the Water Environment
Federation. WEFTEC Session 11-20:1468-1477.
Department of Environment Food and Rural Affairs (DEFRA). 2008. NDMA - concentrations in
drinking water and factors affecting its formation, final report, DEFRA 7348. Available online
at: http://dwi.defra.gov.uk/research/completed-research/reports/DWI70_2_210.pdf.
Desjardins, R., M. Fournier, F. Denizeau, and K. Krzstyniak. 1992. Immunosuppression by
chronic exposure to /V-nitrosodimethylamine (NDMA) in mice. Journal of Toxicology and
Environmental Health, Part A Current Issues. 37(3): 351-361.
Diaz Gomez, M.I., D. Tamayo, and J. A. Castro. 1986. Administration of N-
nitrosodimethylamine, A-nitrosopyrrolidine, or N'-nitrosonornicotine to nursing rats: their
interactions with liver and kidney nucleic acids from sucklings. Journal of the National. Cancer
Institute. 76(6): 1133-1136.
Donovan, P. J. and G.T. Smith. 2008. Urethane and A-nitrosodi ethyl amine are mutagenic for the
Syrian hamster fetus. Mutation Research/Genetic Toxicology and Environmental Mutagenesis.
657(2): 160-163.
Six-Year Review 3 8-6
Technical Support Document for Nitrosamines
December 2016
-------
Dotson, A. and P. Westerhoff. 2009. Occurrence and removal of amino acids during drinking
water treatment. Journal American Water Works Association. 101(9): 101-116.
Dotson, A., P. Westerhoff, and S.W. Krasner. 2009. Nitrogen enriched dissolved organic matter
(DOM) isolates and their affinity to form emerging disinfection by-products. Water Science &
Technology. 60(1): 135-143.
Drabik-Markiewicz, G., K. Van den Maagdenberg, E. De May, S. Deprez, T. Kowalska, and H.
Paelinck. 2009. Role of proline and hydroxyproline in A-nitrosamine formation during heating in
cured meat. Meat Science. 81: 479-486.
Drewes, J.E., T. Harfelder, and C. Hoppe. 2006. Sustainable removal of micropollutants through
riverbank filtration (RBF) and aquifer recharge and recovery (ARR). WQTC 2006 Conference
Proceedings.
Druckrey, H., R. Preussmann, S. Ivankovic, and D. Schmahl. 1967. Organotropic carcinogenic
effects of 65 various A-nitroso- compounds on BD rats. Zeitschrift fur Krebsforschung. 69(2):
103-201 (as cited in USEPA, 1987; CalEPA, 2006)
Farre, J.F., J. Reungoat, F.X. Argaud., M. Rattier, J. Keller, and W. Gernjak. 2011a. Fate of A-
nitrosodimethylamine, trihalomethane, and haloacetic acid precursors in tertiary treatment
including biofiltration. Water Research. 45(8): 5695-5704.
Farre, J.F., K. Doderer, L. Hearn, Y. Poussade, J. Keller, and W. Gernjak. 2011b. Understanding
the operational parameters affecting NDMA formation at advanced water treatment plants.
Journal of Hazardous Materials. 185: 1575-1581.
Faust, S.D. and O.M. Aly. 1998. Chemistry of water treatment. 2nd edition. Boca Raton, FL:
CRC Press.
Fine, D.H., R. Ross, D.P. Rounbehler, A. Silvergleid, and L. Song. 1977. Formation in vivo of
volatile A-nitrosamines in many after ingestion of cooked bacon and spinach. Nature. 265: 753-
755.
Fleming, E., J. Pennington, B. Wachob, R. Howe, and D. Hill. 1996. Removal of N-
nitrosodimethylamine from waters using physical-chemical techniques. Journal of Hazardous
Materials. 51(1-3): 151-164.
Flowers, R.C., and P.C. Singer. 2013. Anion exchange resins as a source of nitrosamines and
nitrosamine precursors. Environmental Science & Technology. 47: 7365-7372.
Frierdich, A. J., J.R. Shapley, and T.J. Strathmann. (2007). Rapid reduction of A'-nitrosamine
disinfection byproducts in water with hydrogen and porous nickel catalysts. Environmental
Science & Technology. 42(1): 262-269.
Fristachi, A. and G. Rice. 2007. Estimation of the total daily oral intake of NDMA attributable to
drinking water. Journal of Water and Health 5(3): 341-355.
Fujii, K., S. Odashima, and M. Okada. 1977. Induction of tumours by administration of N-
dibutylnitrosamine and derivatives to infant mice. British Journal of Cancer. 35(5): 610-4.
Six-Year Review 3 8-7
Technical Support Document for Nitrosamines
December 2016
-------
Fujioka, T., S.J. Khan, J.A. McDonald, A. Roux, Y. Poussade, J.E. Drewes, and L.D. Nghiem.
2013a. A'-nitrosamine rejection by nanofiltration and reverse osmosis membranes: The
importance of membrane characteristics. Desalination. 316: 67-75.
Fujioka, T., S.J. Khan, J.A. McDonald, R.K. Henderson, Y. Poussade, J.E. Drewes, and L.D.
Nghiem. 2013b. Effects of membrane fouling on A'-nitrosamine rejection by nanofiltration and
reverse osmosis membranes. Journal of Membrane Science. 427: 311-319.
Fujioka, T., S.J. Khan, J.A. McDonald, A. Roux, Y. Poussade, J.E. Drewes, and L.D. Nghiem.
2014. A'-nitrosamine rejection by reverse osmosis: effects of membrane exposure to chemical
cleaning reagents. Desalination. 343: 60-66.
Fujioka, T., S.J. Khan, J.A. McDonald, and L.D. Nghiem. 2015. Rejection of trace organic
chemicals by a hollow fiber cellulose triacetate reverse osmosis membrane. Desalination. 368:
69-75
Fujita, K and T. Kamataki. 2001. Role of human cytochrome P450 (CYP) in the metabolic
activation of N-alkylnitrosamines: application of genetically engineered Salmonella typhimurium
YG7108 expressing each form of CYP together with human NADPH-cytochrome P450
reductase. Mutation Research/Fundamental and Molecular Mechanisms. 483(1-2): 35-41.
Gan, X., T. Karanfil, S. Sule Kaplan Bekaroglu, and J. Shan. 2013. The control of N-DBP and C-
DBP precursors with MTEX Water Research. 47: 1344-1352.
Gan, W., T. Bond, X. Yang, and P. Westerhoff 2015. Role of chlorine dioxide in A'-
nitrosodimethylamine formation from oxidation of model amines. Environmental Science &
Technology. 49: 11429-11437.
Genuino, H.C., E.C. Njagi, E.M. Benbow, G.E. Hoag, J.B. Collins, and S.L. Suib. 2011.
Enhancement of the photodegradation of A'-nitrosodi methyl amine in water using amorphous and
platinum manganese oxide catalysts. Journal of Photochemistry and Photobiology A: Chemistry.
217: 284-292.
Gerecke, A.C. and D.L. Sedlak. 2003. Precursors of jV-nitrosodimethylamine in natural waters.
Environmental Science & Technology. 37(7): 1331-1336.
Gerrity, D., A.N. Pisarenko, E. Marti, R. A. Trenholm, F. Gerringer, J. Reungoat, and E.
Dickenson. 2014. Nitrosamines in pilot-scale and full-scale wastewater treatment plants with
ozonation. Water Research. In press.
Gerrity, D., A.N. Pisarenko, E. Marti, R.A. Trenholm, F. Gerringer, J. Reungoat, and E.
Dickenson. 2015. Nitrosamines in pilot-scale and full-scale wastewater treatment plants with
ozonation. Water Research. 72: 251-261.
Gibel, W. 1967. Experimental studies on syncarcinogenesis in esophageal carcinoma. Archiv fur
Geschwulstforsch. 30: 181-189 (as cited in IARC, 2010).
Gilbert, P.J., B. Rollmann, F.D. de Valkeneer, J. Rondelet, M. Mercier, and F. Poncelet. 1984.
Quantitative correlation between the metabolism and the mutagenic activity of N-
nitrosopyrrolidine. Toxicology. 29(3): 261-270.
Six-Year Review 3 8-8
Technical Support Document for Nitrosamines
December 2016
-------
Giles, G.G., J.J. McNeil, G. Donnon, C. Webley, M.P. Staples, P.D. Ireland, S.F. Hurley, and M.
Salzberg. 1994. Dietary factors and the risk of glioma in adults: results of a case-control study in
Melbourne, Australia. International Journal of Cancer. 59: 357-362 (as cited in Health Canada,
2010).
Gollapudi, B.B., K.M. Jackson, and W.T. Scott. 1998. Hepatic lacl and ell mutation in
transgenic ([lambda]LIZ) rats treated with dimethylnitrosamine. Mutation Research. 419: 131-
135.
Gonzalez, C.A., E. Riboli, J. Badosa, T. Cardona, S. Pita, J.M. Sanz, M. Torrent, and A. Agudo.
1994. Nutritional factors and gastric cancer in Spain. American Journal of Epidemiology. 139:
466-473 (as cited in IPCS, 2002; Health Canada, 2010).
Goodman, M.T., J.H. Hankin, L.R. Wilkens, and L.N. Kolonel. 1992. High-fat foods and the risk
of lung cancer. Epidemiology. 3: 288-299 (as cited in IPCS, 2002).
Goslan, E.H., S.W. Krasner, M. Bower, S.A. Rocks, P. Holmes, L.S. Levy, and S.A. Parsons.
2009. A comparison of disinfection by-products found in chlorinated and chloraminated drinking
waters in Scotland. Water Research. 43(18): 4698-4706.
Goto, Y., T. Matsuda, andK. Ito. 1999. Mutagenicities of A-nitrosodimethylamine and N-
nitrosodiethylamine in Drosophila and their relationship to the levels of O-alkyl adducts in DNA.
Mutation Research/Fundamental and Molecular Mechanisms of Mutagenesis. 425(1): 125-134.
Gray, R., R. Peto, P. Brantom, and P. Grasso. 1991. Chronic nitrosamine ingestion in 1040
rodents: the effect of the choice of nitrosamines, the species studied and the age of starting
exposure. Cancer Research. 51: 6470-6491 (as cited in Verna et al., 1996).
Greenblatt, M. and W. Lijinsky. 1972a. Nitrosamine studies: Neoplasms of liver and genital
mesothelium in nitrosopyrrolidine-treated MRC rats. Journal of the National Cancer Institute.
48(6): 1687-1696.
Greenblatt, M. and W. Lijinsky. 1972b. Failure to induce tumors in Swiss mice after concurrent
administration of amino acids and sodium nitrite. Journal of the National Cancer Institute. 48(5):
1389-1392.
Griciute, L. and S.V. BarauskaTte. 1989. The carcinogenic activity of A-nitrosodipropylamine
administered perorally to C57BL6I mice. Eksperimentalnaia Onkologgiia. 11(3): 10-11 (as cited
in AT SDR, 2010).
Griciute L., M. Castegnaro, and J.C. Bereziat. 1982. Influence of ethyl alcohol on the
carcinogenic activity of A-nitrosodi-n-propylamine. In N-Nitroso compounds: Occurrence and
Biological Effects, IARC Scientific Publications No. 41, eds. Bartsch, H., M. Castegnaro, I.K.
O'Neill, and M. Okada, pp. 643-648. Lyon: International Agency for Research on Cancer
(IARC) (as cited in IARC, 2010).
Gui, L., R.W. Gillham, and M.S. Odziemkowski. 2000. Reduction of A'-nitrosodimethylamine
with granular iron and nickel-enhanced iron. 1. Pathways and kinetics. Environmental Science &
Technology. 34(16): 3489-3494.
Six-Year Review 3 8-9
Technical Support Document for Nitrosamines
December 2016
-------
Gunnison, D., M.E. Zappi, C. Teeter, J.C. Pennington, and R. Bajpai. 2000. Attenuation
mechanisms of A'-nitrosodi methyl amine at an operating intercept and treat groundwater
remediation system. Journal of Hazardous Materials. 73(2): 179-197.
Guttenplan, J.B. 1987. Structure-activity relationships in oxidation and mutagenicities of N-
nitrosamines. IARC publication onN-nitroso compounds. IARC Sci Publ. 1987(84): 129-31.
Hanigan, D., J. Zhang, P. Herckes, S.W. Krasner, C. Chen, and P. Westerhoff 2012. Adsorption
of A'-nitrosodi methyl amine precursors and granular activated carbon. Environmental Science &
Technology. 46: 12630-12639.
Hannson, J. 1992. Inherited defects in DNA repair and susceptibility to DNA-damaging agents.
Toxicology Letters. 64/65: 141-148.
Harter, T., J.R. Lund, J. Darby, G.E. Fogg, R. Howitt, K.K. Jessoe, G.S. Pettygrove, J.F. Quinn,
J.H. Viers, D.B. Boyle, H.E. Canada, N. DeLaMora, K.N. Dzurella, A. Fryjoff-Hung, A.D.
Hollander, K.L. Honeycutt, M.W. Jenkins, V.B. Jensen, A.M. King, G. Kourakos, D. Liptzin,
E.M. Lopez, M.M. Mayzelle, A. McNally, J. Medellin-Azuara, and T.S. Rosenstock. 2012.
Addressing nitrate in California's drinking water with a focus on Tulare Lake Basin and Salinas
Valley groundwater. Report for the State Water Resources Control Board Report to the
Legislature. Center for Watershed Sciences, University of California, Davis. Available online at:
http: //groundwaternitrate .ucdavis.edu.
Hatzinger, P.B., C. Condee, K.R. McClay., and A.P. Togna. 2011. Removal of N-
nitrosodimethylamine in a propane-fed membrane bioreactor. Water Research. 45: 254-262.
Hawkins, C.L., D.I. Pattison, and M.J. Davies. 2003. Hypochlorite induced oxidation of amino
acids, peptides and proteins. Amino Acids. 25(3-4): 259-274.
Hazardous Substances Data Bank (HSDB). 2003. A'-Nitrosodi-n-propyl amine. Available online
at: http://toxnet.nlm.nih.gov/cgi-bin/sis/htmlgen7HSDB.
HSDB. 2010. Toxicology Data network. Available online at: http://toxnet.nlm.nih.gov/cgi-
bin/sis/htmlgen?HSDB. Accessed on August 27, 2010.
He, Z., W. Kosinska, Z. Lin-Zhao, X.R. Wu, and J.B. Guttenplan. 2012. Tissue-specific
mutagenesis by A'-butyl[A'-(4-hydroxybutyl)nitrosamine as the basis for urothelial
carcinogenesis. Mutation Research/Fundamental and Molecular Mechanisms of Mutagenesis.
742: 92-95.
Health Canada. 2010. A'-nitrosodi methyl amine (NDMA) in drinking water: Document for public
comment.
Health Canada. 2011. Guidelines for Canadian drinking water quality: Guideline technical
document N-nitrosodimethylamine (NDMA). Available online at: http://www.hc-sc.gc.ca/ewh-
semt/pubs/water-eau/ndma/index-eng.php#a2
Homme, C.L., and J.O. Sharp. 2013. Differential microbial transformation of nitrosamines by an
inducible propane monooxygenase. Environmental Science & Technology. 47: 7388-7395.
Six-Year Review 3 8-10
Technical Support Document for Nitrosamines
December 2016
-------
Howard, P.H., R.S. Boethling, W.F. Jarvis, W.M. Meylan, and E.M. Michalenko. 1991.
Handbook of environmental degradation rates. Chelsea, MI: Lewis Publishers (as cited in
Siddiqui and Atasi, 2004).
Hua, B., K. Veum, A. Koirala, J. Jones, T. Clevenger, and B. Deng. 2007. Fluorescence
fingerprints to monitor total trihalomethanes and iV-nitrosodimethylamine formation potentials in
water. Environmental Chemistry Letters. 5(2): 73-77.
Huang, C.H., L.P. Padhye, and S.H. Park. 2011. PolyDADMAC and dimthylamine as precurosrs
of jV-nitrosodimethylamine during ozonation. WQTC Conference Proceedings.
Hwang, Y., T. Matsuo, K. Hanaki, andN. Suzuki. 1994. Removal of odorous compounds in
wastewater by using activated carbon, ozonation, and aerated biofilter. Water Research. 28(11):
2309-2319.
International Agency for Research on Cancer (IARC). 1978. Some /V-nitroso compounds. IARC
Monographs on the Evaluation of Carcinogenic Risks to Humans, vol. 17: 125-175 (as cited in
HSDB, 2003).
IARC. 2010. Alcohol Consumption and Ethyl Carbamate. IARC Monographs on the Evaluation
of Carcinogenic Risks to Humans, vol. 96. Available online at:
http://monographs.iarc.fr/ENG/Monographs/vol96/index.php.
International Occupational Safety and Health Information Centre. 2001. N-
nitrosodimethylamine, CAS 62-75-9. Geneva, Switzerland (March 2001).
International Programme on Chemical Safety (IPCS). 2002. N-Nitrosodimethylamine. Concise
International Chemical Assessment Document 38. Available online at:
http ://www. inchem. org/documents/cicads/cicads/cicad3 8. htm.
Irving, C.C., W.M. Murphy, and D.S. Daniel. 1984. Comparative carcinogenicity of N-butyl-N-
(3-carboxypropyl)-nitrosamine and A'-butyl-A-(4-hydroxybutyl)nitrosamine for the urinary
bladder of (C57BL/6 X DBA/2)F1 mice. Journal of the National Cancer Institute. 73(3): 753-6.
Jakszyn, P. and C. A. Gonzalez. 2006. Nitrosamine and related food intake and gastric and
oesophageal cancer risk: A systematic review of the epidemiological evidence. World Journal of
Gastroenterology. 12(27): 4296-4303.
Jakszyn, P., S. Bingham, G. Pera, A. Agudo, R. Luben, A. Welch, H. Boeing, G. Del Giudice, D.
Palli, C. Saieva, V. Krogh, C. Sacerdote, R. Tumino, S. Panico, G. Berglund, H. Siman, G.
Hallmans, M.J. Sanchez, N. Larranaga, A. Barricarte, M.D. Chirlaque, J.R. Quiros, T.J. Key, N.
Allen, E. Lund, F. Carneiro, J. Linseisen, G. Nagel, K. Overvad, A. Tionneland, A. Olsen, H.B.
Bueno-de-Mesquita, M.O. Ocke, P.H. Peeters, M.E. Numans, F. Clavel-Chapelon, A.
Trichopoulou, C. Fenger, R. Stenling, P. Ferrari, M. Jenab, T. Norat, E. Riboli, and C.A.
Gonzalez. 2006. Endogenous versus exogenous exposure to N-nitroso compounds and gastric
cancer risk in the European Prospective Investigation into Cancer and Nutrition (EPIC-
EURGAST) study. Carcinogenesis. 27(7): 1497-1501.
Six-Year Review 3 8-11
Technical Support Document for Nitrosamines
December 2016
-------
Jakszyn, P., C.A. Gonzalez, and L.L. Lujan-Barroso. 2011. Red meat, dietary nitrosamines and
heme iron and risk of bladder cancer in the European prospective investigation into cancer and
nutrition (EPIC). Cancer Epidemiology Biomarkers and Prevention. 20(3): 555-559.
Janzowski, C., D. Jacob, I. Henn, H. Zankl, B.L. Poole-Zobel, and G. Eisenbrand. 1989.
Investigations on organ-specific metabolism and genotoxic effects of the urinary bladder
carcinogen N-nitrosobutyl-3-carboxypropylamine (BCPN) and its analogs N-
nitrosodibutylamine (NDBA) and N-nitrosobutyl-4-hydroxybutylamine (4-OH-NDBA).
Toxicology. 59(2): 195-209.
Janzowski, C., R. Landsiedel, P. Golzer P, and G. Eisenbrand. 1994. Mitochondrial formation of
beta-oxopropyl metabolites from bladder carcinogenic omega-carboxyalkylnitrosamines.
Chemico-Biological Interactions. 90(1): 23-33.
Jarabek, A.M., L.H. Pottenger, L.S, Andrews, D. Casciano, M.R. Embry, J.H. Kim, R.J. Preston,
M.V. Reddy, R. Schoeny, D. Shuker, J. Skare, J. Swenberg, G.M. Williams, and E. Zeiger. 2009.
Creating context for the use of DNA adduct data in cancer risk assessment: I. Data organization.
Critical Reviews in Toxicology. 39(8): 659-678.
Kadmi, Y., L. Favier, M. Lofti, N. Nasrallah, and D. Wolbert. 2014. A highly sensitive liquid
chromatography-tandem mass spectrometry method for analysis of a toxic wastewater
disinfection by-product, jV-nitrosomethylethylamine. Analytical Methods. 6: 3231.
Kamataki, T., K. Fujita, K. Nakayama, Y. Yamazaki, M. Miyamoto, and N. Ariyoshi. 2002. Role
of human cytochrome P450 (CYP) in the metabolic activation of nitrosamine derivatives:
Application of genetically engineered Salmonella expressing human CYP. Drug Metabolism
Reviews. 34(3): 667-676.
Kanki, K., A. Nishikawa, K. Masumura, T. Umemura, T. Imazawa, Y. Kitamura, T. Nohmi, and
M. Hirose. 2005. In vivo mutational analysis of liver DNA in GPT delta transgenic rats treated
with the hepatocarcinogens iV-nitrosopyrrolidine, 2-amino-3-methylimidazo[4,5-f]quinoline, and
di(2-ethylhexyl)phthalate. Molecular Carcinogenesis. 42(1): 9-17.
Karanfil, T. K., S.W. Krasner, P. Westerhoff, and Y. Xie. 2008. Recent advances in disinfection
by-product formation, occurrence, control, health effects, and regulations. In Disinfection By-
products in Drinking Water. ACS Symposium Series, 995.
Kemper, J.M., P. Westerhoff, A. Dotson, and W.A. Mitch. 2009. Nitrosamine,
dimethylnitramine, and chloropicrin formation during strong base anion-exchange treatment.
Environmental Science & Technology. 43(2): 466-472.
Kemper, J.M., S.S. Walse, and W.A. Mitch. 2010. Quaternary amines as nitrosamine precursors:
A role for consumer products? Environmental Science & Technology. 44(4): 1224-1231.
Kerklaan, P., S. Bouter, J. A. Zijlstra, and G.R. Mohn. 1986. The effect of mixed-function
oxidase and amine oxidase inhibitors on the activation of dialkylnitrosamines and 1 2
dimethylhydrazine to bacterial mutagens in mice. Journal of Cancer Research and Clinical
Oncology. 111(3): 196-202.
Six-Year Review 3 8-12
Technical Support Document for Nitrosamines
December 2016
-------
Ketkar, M.B., P. Schneider, R. Preussmann, Ch. Plass, and U. Mohr. 1982. Carcinogenic effect
of low doses of nitrosopyrrolidine administered in drinking water to Syrian golden hamsters.
Journal of Cancer Research and Clinical Oncology. 104(1-2): 75-79.
Khan, S. J. and J. A. McDonald. (2010). Quantifying human exposure to contaminants for
multiple-barrier water reuse systems. Water Science & Technology. 61(1): 77-83.
Kim, J. and T.E. Clevinger. 2007. Prediction of iV-nitrosodimethylamine (NDMA) formation as a
disinfection by-product. Journal of Hazardous Materials. 145(1-2): 270-276.
Knekt, P., R. Jarvinen, J. Dich, and T. Hakulinen. 1999. Risk of colorectal and other gastro-
intestinal cancers after exposure to nitrate, nitrite and N -nitroso compounds: a follow-up study.
International Journal of Cancer. 80: 852-856 (as cited in IPCS, 2002; CalEPA, 2006).
Kohut, K.D.A. and S. Andrews. 2003. Polyelectrolyte age and A-nitrosodimethylamine
formation in drinking water treatment. Water Quality Research Journal Canada. 38(4): 719-735.
Kosaka, K., M. Asami, Y. Konno, M. Oya, and S. Kunikane. 2009. Identification of
antiyellowing agents as precursors of A'-nitrosodi methyl amine production on ozonation from
sewage treatment plant influent. Environmental Science & Technology. 43(14): 5236-5241.
Korsrud, G.O., H.G. Grice, and T.K. Goodman. 1973. Sensitivity of several serum enzymes for
the detection of thioacetamide, dimethylnitrosamine- and diethanolamine-induced liver damage
in rats. Toxicology and Applied Pharmacology. 26: 299-313.
Kosaka, K., M. Asami, K. Ohkubo, T. Iwamoto, Y. Tanaka, H. Koshino, S. Echigo, and M.
Akiba. 2014. Identification of a new N-nitrosodimethylamine precursor in sewage containing
industrial effluents. Environmental Science & Technology. 48: 11243-11250.
Krasner, S.W., P. Westerhoff, B. Chen, G. Amy, and S.N. Nam. 2005. Contribution of
wastewater to DBP formation: Case study of an effluent-impacted river. WQTC 2005
Conference Proceedings.
Krasner, S.W., P. Westerhoff, B. Chen, G. Amy, S. Nam, Z. Chowdhury, S. Sinha, and B.E.
Rittmann. 2008. Contribution of wastewater to DBP formation. Water Research Foundation,
Project #2948.
Krasner, S.W., P. Westerhoff, B. Chen, B.E. Rittmann, S. Nam, and G. Amy. 2009. Impact of
wastewater treatment processes on organic carbon, organic nitrogen, and DBP precursors in
effluent organic matter. Environmental Science & Technology. 43(8): 2911-2918.
Krasner, S.W., M.S. Dale, C.F.T. Lee, E.A. Garcia, W. Mitch, and U. von Gunten. 2010. Effect
of water quality and operational parameters during chloramination: NDMA versus halogenated
dbps. AWWA Annual Conference & Exposition.
Krasner, S.W., C.F.T. Lee, E.A. Garcia, and W. Mitch. 2011. A simulated distribution system
test to predict the formation of nitrosamines, THMs, and HAAs in drinking water. WQTC
Conference Proceedings.
Six-Year Review 3 8-13
Technical Support Document for Nitrosamines
December 2016
-------
Krasner, S.W, M.J Sclimenti, and C.F.T. Lee. 2012a. Occurrence and formation of NDMA in
consecutive systems in a major metropolitan area. WQTC 2012 Conference Proceedings.
Krasner, S.W., W.A. Mitch, P. Westerhoff, and A. Dotson. 2012b. Formation and control of
emerging C- and N-DBPs in drinking water. Journal of the American Water Works Assoication.
104 (11): 33e34.
Krasner, S.W., P. Westerhoff, W.A. Mitch, and J. Skadsen. 2012c. Case studies on nitrosamine
formation and control at fullscale drinking water treatment plants. WQTC 2012 Conference
Proceedings (as cited in Krasner et al., 2013).
Krasner, S.W., W.A. Mitch, D.L. McCurry, D. Hanigan, and P. Westerhoff. 2013. Formation,
precursors, control, and occurrence of nitrosamines in drinking water: A review. Water Research.
47: 4433-4450.
Krasner, S.W., R. Shirkani, P. Westerhoff, D. Hanigan, W.A. Mitch, D.L. McCurry, C. Chen, J.
Skadsen, and U. von Gunten. 2015. Controlling the formation of nitrosamines during water
treatment. Water Research Foundation, Project #4730.
Kraus, T.E.C., C.A. Anderson, K. Morgenstern, B.D. Downing, B.A. Pellerin, and B.A.
Bergamaschi. 2010. Determining sources of dissolved organic carbon and disinfection byproduct
precursors to the McKenzie River, Oregon. Journal of Environmental Quality. 39(6): 2100-2112.
Kruithof, J.C., P.C. Kamp, and B.J. Martijn. 2007. UV/H2O2 Treatment: A practical solution for
organic contaminant control and primary disinfection. Ozone: Science & Engineering: The
Journal of the International Ozone Association. 29(4): 273-280.
Kulshrestha, P., K.C. McKinstry, B.O. Fernandez, M. Feelisch, and W. Mitch. 2010. Application
of an optimized total A'-nitrosamine (TONO) assay to pools: pi aci ng A'-ni trosodi m ethyl am i ne
(NDMA) into perspective. Environmental Science & Technology. 44(9): 3369-3375.
Kuroki, T., C. Drevon, and R. Montesano. 1977. Microsome-mediated mutagenesis in V79
Chinese hamster cells by various nitrosamines. Cancer Research. 37: 1044-1050.
Kuroki, T., A. Abbondandolo, C. Drevon, E. Huberman, and F. Laval. 1980. Mutagenesis assays
with mammalian cells.. IARC Monographs on the Evaluation of the Carcinogenic Risk of
Chemicals to Humans. Supplement 2: 107-133.
Kushida, H., K. Fujita, A. Suzuki, M. Yamada, T. Endo, T. Nohmi, and T. Kamataki. 2000.
Metabolic activation of A'-alkylnitrosamines in genetically engineered Salmonella typhimurium
expressing CY02E1 or CYP2A6 together with human NADPH-cytochrome P450 reductase.
Carcinogenesis. 21(6): 1227-1232.
La Vecchia, C., B. D'Avanzo, L. Airoldi, C. Braga, and A. Decarli. 1995. Nitrosamine intake
and gastric cancer. European Journal of Cancer Prevention. 4: 469-474 (as cited in IPCS, 2002).
Landsman, N.A., K.L. Swancutt, C.N. Bradford, C.R. Cox, J.J. Kiddle, and S.P. Mezyk. 2007.
Free radical chemistry of advanced oxidation process removal of nitrosamines in water.
Environmental Science & Technology. 41(16): 5818-5823.
Six-Year Review 3 8-14
Technical Support Document for Nitrosamines
December 2016
-------
Langenbach, R. 1986. Mutagenic activity and structure-activity relationships of short-chain
dialkyl A'-nitrosamines in a hamster hepatocyte v79 cell-mediated system. Mutation
Research/Fundamental and Molecular Mechanisms of Mutagenesis. 163(3): 303-11.
Larsson, S.C., L. Bergkvist, and A. Wolk. 2006. Processed meat consumption, dietary
nitrosamines and stomach cancer risk in a cohort of Swedish women. International Journal of
Cancer. 915-919.
Le Roux, J., H. Gallard, and J.P. Croue. 2011. Chloramination of nitrogenous contaminants
(pharmaceuticals and pesticides): NDMA and halogenated DBPs formation. Water Research.
45(10): 3164-3174
Le Roux, J., H. Gallard, and J.P. Croue. 2012. Formation of NDMA and halogenated DBPs by
chloramination of tertiary amines: The influence of bromide ion. Environmental Science &
Technology. 46(3): 1581-1589.
Lee, J., W. Choi, and J. Yoon. 2005a. Photocatalytic degradation of iV-nitrosodimethylamine:
mechanism, product distribution, and TiC>2 surface modification. Environmental Science &
Technology. 39(17): 6800-6807.
Lee, C., W. Choi, and J. Yoon. 2005b. UV photolytic mechanism of iV-nitrosodimethylamine in
water: Roles of dissolved oxygen and solution pH. Environmental Science & Technology.
39(24): 9702-9709.
Lee, C., W. Choi, Y.G. Kim, and J. Yoon. 2005c. UV photolytic mechanism of N-
nitrosodimethylamine in water: Dual pathways to methylamine versus dimethylamine.
Environmental Science & Technology. 39(7): 2101-2106.
Lee, W., P. Westerhoff, and M. Esparza-Soto. 2006. Occurrence and removal of dissolved
organic nitrogen in U.S. water treatment plant. Journal of the American Water Works. 98(10):
102-110.
Lee, W., P. Westerhoff, and J. Croue. 2007a. Dissolved organic nitrogen as a precursor for
chloroform, dichloroacetonitrile, A'-nitrosodimethyl amine, and trichloronitromethane.
Environmental Science & Technology. 41(15): 5485-5490.
Lee, C., C. Schmidt, J. Yoon, andU. von Gunten. 2007b. Oxidation of A'-nitrosodimethylamine
(NDMA) precursors with ozone and chlorine dioxide: Kinetics and effect on NDMA formation
potential. Environmental Science & Technology. 41(6): 2056-2063.
Lee, C., J. Yoon, and U. von Gunten. 2007c. Oxidative degradation of A'-nitrosodimethylamine
by conventional ozonation and the advanced oxidation process ozone/hydrogen peroxide. Water
Research. 41(3): 581-590.
Lee, C., Y. Lee, C. Schmidt, J. Yoon, and U. von Gunten. 2008. Oxidation of suspected N-
nitrosodimethylamine (NDMA) precursors by ferrate (VI) kinetics and effect on NDMA
formation potential of natural waters. Water Research. 42(1-2): 433-441.
Six-Year Review 3 8-15
Technical Support Document for Nitrosamines
December 2016
-------
Lee, M., Y. Lee F. Soltermann, and U. von Gunten. 2013. Analysis of A'-nitrosamines and other
nitro(so) compounds in water by high performance liquid choromatography with post column
UV photolysis/Griess reaction. Water Research. 47: 4893-4903.
Lee, C-F.T., S.W. Krasner, M.J. Sclimenti, M. Prescott, and Y.C. Guo. 2015. Nitrosamine
precursors and wastewater indicators in discharges in the Sacramento-San Joaquin Delta. In
Advances in Disinfection Byproducts, Chapter 7, pp. 119-133. ACS Symposium Series, 1190.
Lewis, R.J. Sr. 1981. Hawley's condensed chemical dictionary. 13th edition. New York, NY:
Van Nostrand Reinhold.
Li, T., D. Yu, Q. Xian, A. Li, and C. Sun. 2015. Variation of levels and distribution of N-
nitrosamines in different seasons in drinking waters of East China. Environmental Science and
Pollution Research. 22: 11792-11800.
Liang, S., H.M. Joon, M.K. Davis, J.F. Green, and D.S. Remer. 2003. Use of pulsed-UV
processes to destroy NDMA. Journal of the American Water Works Association. 95(9): 121-131.
Liao, X., Wang, C., Wang, J., Zhang, X., Chen, C., Krasner, S.W., and I.H. Suffet. 2014.
Nitrosamine precursor and DOM control in an effluent-affected drinking water. Journal of the
American Water Works Association. 106(7): E307-E318.
Liao, X., E. Bei, S. Li, Y Ouyang, J. Wang, C. Chen, X. Zhang, S.W. Krasner, and I.H. Suffet.
2015a. Applying the polarity rapid assessment method to characterize nitrosamine precursors and
to understand their removal by water treatment processes. Water Research. 87: 292-298.
Liao, X., C. Chen, S. Xie, D. Hanigan, J. Wang, X. Zhang, P. Westerhoff, and S.W. Krasner.
2015b. Nitrosamine precursor removal by BAC: A Case study of adsorption versus biotreatment.
Journal of the American Water Works Association. 107(9): E454-E463.
Liao, X., C. Chen, J. Zhang, Y. Dai, X. Zhang, and S. Xie. 2015c. Operational performance,
biomass, and microbial community structure: Impacts of backwashing on drinking water
biofilter. Environmental Science Pollution Research. 22: 546-554.
Liao, X., C. Chen, J. Zhang, and S. Xie. 2015d. Dimethylamine biodegradation by mixed culture
enriched from drinking water biofilter. Chemosphere. 119: 935-940
Lide, D.R. 1995-96. CRC handbook of chemistry and physics. 76th Edition. New York, NY: CRC
Press.
Lijinsky, W. 1983. Species specificity in nitrosamine carcinogenesis. Basic Life Science. 24: 63-
75 (as cited in CalEPA, 2006).
Lijinsky, W. 1987a. Carcinogenicity and mutagenicity of N-nitroso compounds. Molecular
Toxicology. 1(1): 107-119.
Lijinsky, W. 1987b. Structure-activity relations in carcinogenesis
by N-nitroso compounds. Cancer and Metastasis Reviews. 6(3): 301-356
Six-Year Review 3 8-16
Technical Support Document for Nitrosamines
December 2016
-------
Lijinsky, W. and M.D. Reuber. 1980. Carcinogenicity in rats of nitrosomethylethylamines
labeled with deuterium in several positions. Cancer Research. 40(1): 19-21.
Lijinsky, W. and M.D. Reuber. 1981. Comparative carcinogenesis by some aliphatic
nitrosamines in Fischer rats. Cancer Letters. 14(3): 297-302.
Lijinsky, W. and M.D. Reuber. 1983. Carcinogens in Fischer rats by nitrosodipropylamine,
nitrosodibutylamine and nitrosobis(2-oxopropyl)amine given by gavage. Cancer Letters. 19:
207-213.
Lijinsky, W. and H.W. Taylor. 1978. Comparative carcinogenicity of some derivatives of
nitrosodi-n-propylamine in rats. Ecotoxicology and Environmental Safety. 2: 421-426.
Lijinsky, W. and H.W. Taylor. 1979. Carcinogenicity of methylated derivatives of N-
nitrosodiethylamine and related compounds in Sprague-Dawley rats. Journal of the National
Cancer Institute. 62: 407-410 (as cited in USEPA, 1987).
Lijinsky, W., R.M. Kovatch, and C.W. Riggs. 1987. Carcinogenesis by nitrosodialkylamines and
azoxyalkanes given by gavage to rats and hamsters. Cancer Research. 47: 3968-3972.
Liu, Y.X. and J.B. Guttenplan. 1992. Mutational specificities of iV-nitrosamines in a host-
mediated assay: comparison with direct-acting N-nitroso compounds in vitro and an approach to
deducing the nature of ultimate mutagens in vivo. Molecular Carcinogenesis. 6: 232-237.
Liu, Y.D., M. Selbes, C. Zeng, R. Zhong, and T. Karanfil. 2014. Formation mechanism of
NDMA from ranitidine, trimethylamine, and other tertiary amines during chloramination: A
computational study. Environmental Science & Technology. 48: 8653-8663
Lofberg, B. and H. Tjalve. 1986. Perinatal metabolism of N-nitrosodibutylamine in Syrian
golden hamsters. Cancer Lett. 31: 153-161.
Loh, Y.H., P. Jakszyn, R.N. Luben, A.A. Mulligan, P.N. Mitrou and K.T. Khaw. 2011. N-
Nitroso compounds and cancer incidence: The European Prospective Investigation into Cancer
and Nutrition (EPIC)-Norfolk Study. The American Journal of Clinical Nutrition. 93(5): 1053-
61.
Luh, J., and B. Marinas. 2012. Bromide ion effect on iV-nitrosodimethylamine formation by
monochloramine. Environmental Science & Technology. 46(9): 5085-5092.
Lv, J., L. Wang, Y. Song, and Y. Li. 2015. A-nitrosodi methyl amine formation from ozonation of
chlorpheniramine: Influencing factors and transformation mechanism. Journal of Hazardous
Materials. 299: 484-494.
Ma, D., Y. Meng, C. Xia, B. Gao, and Y. Wang. 2015. Fractionation, characterization, and C-,
N-disinfection byproduct formation of soluble microbial products in MBR processes.
Bioresource Technology. 98: 380-387.
Maduagwu, E.N. and O. Bassir. 1980. A comparative assessment of toxic effects of
dimethylnitrosamine in six different species. Toxicology and Applied Pharmacology. 53: 211-
219.
Six-Year Review 3 8-17
Technical Support Document for Nitrosamines
December 2016
-------
Marhaba, T. F. and I.H. Kochar. 2000. Rapid prediction of disinfection by-product formation
potential by fluorescence. Environmental Engineering and Policy. 2(1): 29-36.
Martelli, A., L. Robbiano, G.M. Gazzaniga, and G. Brambilla. 1988. Comparative study of DNA
damage and repair induced by ten A'-nitroso compounds in primary cultures of human and rat
hepatocytes. Cancer Research. 48(15): 4144-52.
Marti, E.J., A.N. Pisarenko, J.R. Peller, and E.R.V. Dickenson. 2015. /V-nitrosodimethylamine
(NDMA) formation from the ozonation of model compounds. Water Research. 72: 262-270.
Massachusetts Department of Environmental Protection (DEP). 2004. Current regulatory limit:
N-nitrosodimethylamine (NDMA). Available online at:
http://www.mass.gov/eea/agencies/massdep/water/drinking/standards/n-nitrosodimethylamine-
ndma.html.
McCann, J., E. Choi, E. Yamasaki, and B.N. Ames. 1975. Detection of carcinogens as mutagens
in the Salmonella/microsome test: assay of 300 chemicals. Proceedings of the National Academy
of Sciences., U.S.A. 72: 5135-5139 (as cited in Aiub et al., 2003).
McCoy, G.D., S.S. Hecht, and K. Furuya. 1986. The effect of chronic ethanol consumption on
the tumorigenicity of A'-nitrosopyrrolidine in male Syrian golden hamsters. Cancer Letters.
33(2): 151-159.
McCurry, D.L., Krasner, S.W., Von Gunten, U. and W.A. Mitch. 2015. Determinants of
disinfectant pretreatment efficacy for nitrosamine control in chloraminated drinking water. Water
Research. 84: 161-170.
Mezyk, S. P., D.B. Ewing, J.J. Kiddle, and K.P. Madden. 2006. Kinetics and mechanisms of the
reactions of hydroxyl radicals and hydrated electrons with nitrosamines and nitramines in water.
The Journal of Physical Chemistry A. 110(14): 4732-4737.
Mhlongo, S.H., B.B. Mamba, and R.W. Krause. 2009. Nitrosamines: A review on their
prevalence as emerging pollutants and potential remediation options. Water SA. 35(5): 735-740.
Michejda, C.J., M.B. Krueger-Koepke, andR.M. Kovatch. 1986. Carcinogenic effects of
sequential administration of two nitrosamines to rats. Cancer Research. 25: 126.
Mirvish, S. S. 1974. Formation of A'-nitroso compounds: Chemistry, kinetics, and in vivo
occurrence. Toxicology and Applied Pharmacology. 31: 325-351.
Mirvish, S. S. 1992. In vivo formation of A'-nitroso compounds: Formation from nitrite and
nitrogen dioxide, and relation to gastric cancer. In Nitrosamines and human cancer, Banbury
Report 12, ed. P.N. Magee, pp. 227-241. New York, NY: Cold Spring Harbor Laboratory, New
York.
Mitch, W.A. and I.M. Schreiber. 2008. Degradation of tertiary alkylamines during
chlorination/chloramination: implications for formation of aldehydes, nitriles, halonitroalkanes,
and nitrosamines. Environmental Science & Technology. 42(13): 4811-4817.
Six-Year Review 3 8-18
Technical Support Document for Nitrosamines
December 2016
-------
Mitch, W.A. andD.L. Sedlak. 2002. Formation of A'-nitrosodimethylamine (NDMA) from
dimethylamine during chlorination. Environmental Science & Technology. 36(4): 588-595.
Mitch, W.A. and D.L. Sedlak. 2004. Characterization and fate of A'-nitrosodimethyl amine
precursors in municipal wastewater treatment plants. Environmental Science & Technology.
38(5): 1445-1454.
Mitch, W.A., A.C. Gerecke, and D.L. Sedlak. 2003a. A iV-nitrosodimethylamine (NDMA)
precursor analysis for chlorination of water and wastewater. Water Research. 37(15): 3733-3741.
Mitch, W.A., J.O. Sharp, R.D Trussell, R.L. Valentine, L. Alvarez-Cohen, and D.L. Sedlak.
2003b. A'-nitrosodimethyl amine (NDMA) as a drinking water contaminant: A review.
Environmental Engineering Science. 20(5): 389-404.
Mitch, W.A., G.L. Oelker, E.L. Hawley, R.A. Deeb, and D.L. Sedlak. 2005. Minimization of
NDMA formation during chlorine disinfection of municipal wastewater by application of pre-
formed chloramines. Environmental Engineering Science. 22(6): 882-890.
Mitch, W.A., S.W. Krasner, P. Westerhoff, and A. Dotson. 2009. Occurrence and formation of
nitrogenous disinfection by-products. Water Research Foundation, Project #3014.
Miyashita, Y., S. Park, H. Hyung, C. Huang, and J. Kim. 2009. Removal of A'-nitrosamines and
their precursors by nanofiltration and reverse osmosis membranes. Journal of Environmental
Engineering. 135(9): 788-795.
Mochizuki, M., T. Anjo, K. Takeda, E. Suzuki, N. Sekiguchi, G.F. Huang, and M. Okada. 1986.
Chemistry and mutagenicity of alpha hydroxy nitrosamines. IARC Sci. PubL. 41: 553-559.
Mohanty, S.K., M. Snehota, and C. Ray. 2006. Transport of selected contaminants in recycled
water through a Hawaii soil. Proceedings of the World Environmental and Water Resources
Congress.
Mohr, U. and J. Althoff. 1965. The transplacental effect of the carcinogen diethylnitrosamine in
the mouse. Zeitschrift fiir Krebsforschung. 67: 152-155. (English translation).
Mohr, U., J. Althoff, and A. Authaler. 1966. Diaplacental effect of the carcinogen
diethylnitrosamine in the golden hamster. Cancer Research. 26: 2349-2352.
Montesano, R. and H. Bartsch. 1976. Mutagenic and carcinogenic A'-nitroso compounds:
possible environmental hazards. Mutation Research/Reviews in Genetic Toxicology. 32: 179-
228 (as cited in Verna et al., 1996).
Montgomery, J.M. 1985. Water treatment principles and design. New York, NY: John Wiley
and Sons.
Morita, T., N. Asano, T. Awogi, Y.F. Sasaki, S. Sato, H. Shimada, S. Sutou, A. Wakata, T.
Sofuni, and M. Hayashi. 1997. Evaluation of the rodent micronucleus assay in the screening of
IARC carcinogens (Groups 1, 2 A and 2B): The summary report of the 6th collaborative study by
CSGMT/JEMS - MOA. Mutation Research/Reviews in Genetic Toxicology. 389: 3-122.
Six-Year Review 3 8-19
Technical Support Document for Nitrosamines
December 2016
-------
Morran, J., M. Whittle, R.B. Fabris, M. Harris, J.S. Leach, G. Newcombe, and M. Drikas. 2011.
Nitrosamines from pipeline materials in drinking water distribution systems. Journal of the
American Water Works Association. 103(10): 76-83.
Nagao, M., E. Suzuki, K. Yasuo, T. Yahagi, and Y. Seino. 1977. Mutagenicity of A'-butyl-A-(4-
hydroxybutyl)nitrosamine, a bladder carcinogen, and related compounds. Cancer Research. 37:
399-407.
Najm, I. and R.R. Trussell. 2001. NDMA formation in water and wastewater. Journal of the
American Water Works Association. 93(2): 92-99.
Napalkov, N.P. and V.A. Aleksandrov. 1968. On the effects of blastomogenic substances during
embryogenesis. Zeitschrift fiir Krebsforschung. 71: 195-205 (as cited in IPCS, 2002).
National Toxicology Program (NTP). 2011. Report on carcinogens. 12th edition. Research
Triangle Park, NC: U.S. Department of Health and Human Services, Public Health Service.
Nawrocki, J. and P. Andrzejewski. 2011. Nitrosamines and water. Journal of Hazardous
Materials. 189(1-2): 1-18.
Negishi, T. and H. Hayatsu. 1980. The pH-dependent response of Salmonella typhimurium
TA100 to mutagenic jV-nitrosamines. Mutation Research/Reviews in Genetic Toxicology. 79(3):
223-230.
Negishi, T., T. Shiotani, K. Fujikawa, and H. Hayatsu. 1991. The genotoxicities of N-
nitrosamines in Drosophila melanogaster in vivo: the correlation of mutagenicity in the wing
spot test with the DNA damages detected by the DNA-repair test. Mutation Research/Reviews in
Genetic Toxicology. 252(2): 119-28.
New Jersey Department of Environmental Protection (NJDEP). 2014. N.J.A.C. 7:9C; Ground
water quality standards. Available online at: http://www.nj.gov/dep/rules/rules/njac7_9c.pdf.
Last amended July 22, 2010.
Nishie, K. 1983. Comparison of the effects of A'-nitrosodi methyl amine on pregnant and
nonpregnant Holtzman rats. Food and Chemical Toxicology. 21(4): 453-462.
Nishikawa, T., E.I. Salim, K. Morimura, M. Kaneko, M. Ogawa, A. Kinoshita, H. Osugi, H.
Kinoshita, and S. Fukushima. 2003. High susceptibility of p53 knockout mice to esophageal and
urinary bladder carcinogenesis induced by N, N-dibutylnitrosamine. Cancer Letters. 194(1): 45-
54.
Nolan, B.T., K.J. Hitt, and B.C. Ruddy. 2002. Probability of nitrate contamination of recently
recharged groundwaters in the conterminous United States. Environmental Science &
Technology. 36(10): 2138-2145.
Occupational Safety and Health Administration (OSHA). 2006. Chemical sampling information:
A'-Nitrosodi methyl amine. Washington, DC: U.S. Department of Labor, Occupational Safety and
Health Administration. Available online at:
http://www.osha.gOv/dts/chemicalsampling/data/CH_258000.html#health.
Six-Year Review 3 8-20
Technical Support Document for Nitrosamines
December 2016
-------
O'Day, R., J. Rench, R. Oen, and A. Castro. 1998. Demographic Distribution of Sensitive
Populations Groups. U.S. Environmental Protection Agency, Office of Water, Office of Science
and Technology, Health and Ecological Criteria Division. Contract No. 68-C6-0036. SRA 557-
05/14.
Odziemkowski, M.S., L. Gui, andR.W. Gillham. 2000. Reduction of A'-nitrosodimethylamine
with granular iron and nickel-enhanced iron. 2. Mechanistic studies. Environmental Science &
Technology. 34(16): 3495-3500.
Olajos, E.J., N. Maverakis, and H.H. Cornish. 1978. Mutagenicity of alkyl-(omega-
hydroxyalkyl) nitrosamines related to dibutylnitrosamine. Mutation Research/Reviews in
Genetic Toxicology. 56(3): 219-23.
O'Neil, M.J., A. Smith, and P.E. Heckelman (eds). 2001. The Merck index: An encyclopedia of
chemicals, drugs, and biologicals. 13th edition. Whitehouse Station, NJ: Merck and Co., Inc.
Otsuka, H. and A. Kuwahara. 1971. Hemangiomatous lesions of mice treated with
nitrosodimethylamine. Gann. 62: 147-156 (as cited in CalEPA, 2006).
Oya, M., K. Kosaka, M. Asami, and S. Kunikane. 2008. Formation of A'-nitrosodimethylamine
(NDMA) by ozonation of dyes and related compounds. Chemosphere. 73(11): 1724-1730.
Padhye, L. 2010. Roles of polydadmacs, dithiocarbamates and activated carbons in formation of
n-nitrosamine contaminants in water. Dissertation, Georgia Institute of Technology School of
Civil and Environmental Engineering.
Padhye, L., U. Tezel, W.A. Mitch, S.G. Pavlostathis, and C. Huang. 2009. Occurrence and fate
of nitrosamines and their precursors in municipal sludge and anaerobic digestion systems.
Environmental Science & Technology. 43(9): 3087-3093.
Padhye, L., P. Wang, T. Karanfil, and C. Huang. 2010. Unexpected role of activated carbon in
promoting transformation of secondary amines to A'-nitrosamines. Environmental Science &
Technology. 44(11): 4161-4168.
Padhye, L., Y. Luzinova, M. Cho, B. Mizaikhoff, J. Kim, and C. Huang. 2011. PolyDADMAC
and dimethylamine as precursors of A'-nitrosodimethylamine during ozonation: Reaction kinetics
and mechanisms. Environmental Science & Technology. 45: 4353-4359.
Padhye, L., Kim, J-H., and C-H. Huang. 2013. Oxidation of dithiocarbamates to yield N-
nitrosamines by water disinfection oxidants. Water Research. 47: 725-736.
Park, S., S. Wei, B. Mizaikoff, A.E. Taylor, C. Favero, and C. Huang. 2009. Degradation of
amine-based water treatment polymers during chloramination as A'-nitrosodimethylamine
(NDMA) precursors. Environmental Science & Technology. 43(5): 1360-1366.
Park, S.H., L.P. Padhye, P. Wang, M. Cho, J.H. Kim, and C.H. Huang. 2014. N-
nitrosodimethylamine (NDMA) formation potential of amine-based water treatment polymers:
Effects of in-situ chloramination, breakpoint chlorination, and pre-oxidation. Journal of
Hazardous Materials. In press.
Six-Year Review 3 8-21
Technical Support Document for Nitrosamines
December 2016
-------
Park, S.H., L.P. Padhye, P. Wang, M. Cho, J-H. Kim, and C-H. Huang. 2015. N-
nitrosodimethylamine (NDMA) formation potential of amine-based water treatment polymers:
Effect of in-situ chloramination, breakpoint chlorination, and pre-oxidation. Journal of
Hazardous Materials. 282: 133-140.
Parker, K., T. Zeng, J. Harkness, A. Vengosh, and W. Mitch. 2014. Enhanced formation of
disinfection byproducts in shale gas wastewater-impacted drinking water supplies.
Environmental Science & Technology. 48: 11161-11169.
Parodi, S., A. Zunino, L. Ottaggio, and L. Santi. 1983. Quantitative correlation between
carcinogenicity and sister chromatid exchange induction in vivo for a group of 11 A-nitroso
derivatives. Journal of Toxicology and Environmental Health, Part A Current Issues. 11(3): 337-
46.
PBT Profiler. 2010. U.S. Environmental Protection Agecy, Office of Chemical Safety and
Pollution Prevention, Environmental Health Analysis Center. Available online at:
http://www.pbtprofiler.net/.
Pedal, I., K. Bresserer, and K. Goerttler. 1982. Fatal nitrosamine poisoning. Archives of
Toxicology. 50(2): 101-112.
Pehlivanoglu-Mantas, E. and D.L. Sedlak. 2006. The fate of wastewater-derived NDMA
precursors in the aquatic environment. Water Research. 40(6): 1287-1293.
Pehlivanoglu-Mantas, E. and D. L. Sedlak 2008. Measurement of dissolved organic nitrogen
forms in wastewater effluents: Concentrations, size distribution and NDMA formation potential.
Water Research. 42(14): 3890-3898.
Perez, D.M., G.G. Alatorre, E.B. Alvarez, E.E. Silva, and J.F.J. Alvarado. 2008. Solid-phase
microextraction of /V-nitrosodimethylamine in beer. Food Chemistry. 107: 1348-1352.
Peto, R., R. Gray, P. Brantom, and P. Grasso. 1984. Nitrosamine carcinogenesis in 5120 rodents:
chronic administration of sixteen different concentrations of NDEA, NDMA, NPYR and NPIP in
the water of 4440 inbred rats, with parallel studies on NDEA alone of the effect of age of starting
(3, 6 or 20 weeks) and of species (rats, mice or hamsters) (as cited in O' Neill et al., 1984).
Peto, R., R. Gray, P. Brantom, and P. Grasso. 1991a. Dose and time relationships for tumor
induction in the liver and esophagus of 4080 inbred rats by chronic ingestion of N-
nitrosodiethylamine or /V-nitrosodimethylamine. Cancer Research. 51(23 Pt 2): 6452-6469.
Peto R, R. Gray, P. Brantom, and P. Grasso. 1991b. Effects on 4080 rats of chronic ingestion of
A'-nitrosodi ethyl amine or A-nitrosodimethylamine: a detailed dose-response study. Cancer
Research. 51(23 Pt2): 6415-6451.
Phillipson, C.E., and C. Ioannides. 1985. Metabolic activation of nitrosamines to mutagens by
various animal species including man. Biochemical Pharmacology. 34(3): 441-442.
Pietsch, J., F. Sacher, W. Schmidt, and H-J Bauch. 2001. Polar nitrogen compounds and their
behavior in the drinking water treatment process. Water Research. 35(15): 3537-3544.
Six-Year Review 3 8-22
Technical Support Document for Nitrosamines
December 2016
-------
Plumlee, M.H. and M. Reinhard. 2007. Photochemical attenuation of A-nitrosodi methyl amine
(NDMA) and other nitrosamines in surface water. Environmental Science & Technology.
41(17): 6170-6176.
Plumlee, M.H., M. Lopez-Mesas, A. Heidlberger, K.P. Ishida, and M. Reinhard. 2008. N-
nitrosodimethylamine (NDMA) removal by reverse osmosis and UV treatment and analysis via
LC-MS/MS. Water Research. 42(1-2): 347-355.
Pobel, D., E. Riboli, J. Cornee, B. Hemon, and M. Guyader. 1995. Nitrosamine, nitrate and
nitrite in relation to gastric cancer: A case-control study in Marseille, France. European Journal
of Epidemiology. 11: 67-73 (as cited in IPCS, 2002).
Pool, B.L., J. Gottfried-Anacker, G. Eisenbrand, and C. Janzowski. 1988. A-Nitroso-butyl-3-
carboxypropylamine (BCPN), a urinary bladder carcinogen, non mutagenic in S. typhimurium.
Mutation Research/Reviews in Genetic Toxicology. 209: 79-83.
Poussade, Y., A. Roux, T. Walker, and V. Zavlanos. 2009. Advanced oxidation for indirect
potable reuse: a practical application in Australia. Water Science & Technology. 60(9): 2419-
2424.
Pozzi, R., P. Bocchini, F. Pinelli, and G.C. Galletti. 2011. Determination of nitrosamines in
water by gas chromatography/chemical/ionization/selective ion trapping mass spectrometry.
Journal of Chromatography A. 1218(14): 1808-1814.
Preston-Martin, S. 1989. Epidemiological studies of perinatal carcinogenesis. IARC Scientific.
Publications. 96: 289-314.
Preussmann, R. and B.W. Stewart. 1984. A-Nitroso carcinogens. In Chemical Carcinogens (2nd
ed), Monograph 182, ed. Searle, C.E., pp. 643-828. Washington, D.C.: American Chemical
Society (as cited in CalEPA, 2006).
Preussmann, R., D. Schmahl, and G. Eisenbrand. 1977. Carcinogenicity of A'-nitrosopyrrolidine:
Dose-response study in rats. Zeitschrift fur Krebsforschung. 90: 161-166.
Prival, M.J., V.D. King, and A.T. Sheldon, Jr. 1979. The mutagenicity of dialkyl nitrosamines in
the Salmonella plate assay. Environmental Mutagenesis. 1(2): 95-104.
Qi, W., L.F. Yee, and H. Jiangyong. 2014. Relationship between organic precursors and N-
nitrosodimethylamine (NDMA) formation in tropical water sources. Journal of Water and
Health. 12(4): 736-746.
Quillardet, P. and M. Hofnung. 1985. The SOS chromotest: a review. Mutation
Research/Reviews in Genetic Toxicology. 297: 235-279 (as cited in Aiub et al., 2003).
Rajat, S. 2008. Kinetics of NDMA (N-nitrosodimethylamine) oxidation by permanganate under a
range of pH and temperature conditions. Dissertation, University of Colorado at Boulder.
Abstract only.
Registry of Toxic Effects of Chemical Substances (RTECS). 2002. Dimethylamine, A'-nitroso.
(as cited in CalEPA, 2006; Health Canada, 2010).
Six-Year Review 3 8-23
Technical Support Document for Nitrosamines
December 2016
-------
Risch, H.A., M. Jain, N.W. Choi, J.G. Fodor, C.J. Pfeiffer, G.R. Howe, L.W. Harrison, K.J.
Craib, and A.B. Miller. 1985. Dietary factors and the incidence of cancer of the stomach.
American Journal of Epidemiology. 122: 947-957 (as cited in CalEPA, 2006).
Risk Assessment Information System (RAIS). 2009. "RAIS Toxicity and Properties." Available
online at: http://rais.ornl.gov/cgi-bin/tools/TOX_search?select=chem. Accessed on August 10,
2009.
Robbiano, L., E. Mereto, C. Corbu, and G. Brambilla. 1996. DNA damage induced by seven N-
nitroso compounds in primary cultures of human and rat kidney cells. Mutation Research.
368(1): 41-47.
Rogers, M.A.M., T.L. Vaughn, S. Davis, and D.B. Thomas. 1995. Consumption of nitrate,
nitrite, and nitrosodimethylamine and the risk of upper aerodigestive tract cancer. Cancer
Epidemiology Biomarkers & Prevention. 4: 29-36 (as cited in CalEPA, 2006).
Russell, C. G., N.K. Blute, S. Via, X. Wu, and Z. Chowdhury. 2012. Nationwide assessment of
nitrosamine occurrence and trends. Journal of the American Water Works Association. 104(3):
E205-E217.
Sacher, F.S., S. Lenz, and H.-J. Brauch. 1997. Analysis of primary and secondary aliphatic
amines in waste water and surface water by gas chromatography-mass spectrometry after
derivatization with 2,4-dinitrofluorobenzene or benzene-sulfonyl chloride. Journal of
Chromatography A. 764(1): 85-93.
Sacher, F. S., C.K. Schmidt, C. Lee, and U. von Gunten. 2008. Strategies for minimizing
nitrosamine formation during disinfection. Water Research Foundation Project, #2729.
Schaefer, C.E., C. Topoleski, and M.E. Fuller. 2007. Effectiveness of zerovalent iron and nickel
catalysts for degrading chlorinated solvents and A'-nitrosodi methyl amine in natural groundwater.
Water Environment Research. 79(1): 57-62.
Schmidt, C. K. and H. Brauch. 2008. A', A'-di m eth y 1 sulfainide as precursor for A'-
nitrosodi methyl amine (NDMA) formation upon ozonation and its fate during drinking water
treatment. Environmental Science & Technology. 42(17): 6340-6346.
Schmidt, C.K., F. Sacher, and H. Brauch. 2006. Strategies for minimizing formation of NDMA
and other nitrosamines during disinfection of drinking water. WQTC 2006 Conference
Proceedings.
Schreiber, I.M. and W.A. Mitch. 2005. Influence of the order of reagent addition on NDMA
formation during chloramination. Environmental Science & Technology. 39(10): 3811-3818.
Schreiber, I.M. and W.A. Mitch. 2006a. Nitrosamine formation pathway revisited: The
importance of chloramine speciation and dissolved oxygen. Environmental Science &
Technology. 40(19): 6007-6014.
Schreiber, I.M. and W.A. Mitch. 2006b. Occurrence and fate of nitrosamines and nitrosamine
precursors in wastewater-impacted surface waters using boron as a conservative tracer.
Environmental Science & Technology. 40(10): 3203-3210.
Six-Year Review 3 8-24
Technical Support Document for Nitrosamines
December 2016
-------
Schreiber, I.M. and W. A. Mitch. 2007. Enhanced nitrogenous disinfection byproduct formation
near the breakpoint: implications for nitrification control. Environmental Science & Technology.
41(20): 7039-7046.
Sedlak, D.L., R.A. Deeb, E.L. Hawley, W.A. Mitch, T.D. Durbin, S. Mowbray, and S. Carr.
2005. Sources and fate of nitrosodimethylamine and its precursors in municipal wastewater
treatment plants. Water Environment Research. 77(1): 32-39.
Seidel, C. J., M.J. McGuire, R.S. Summers, and S. Via. 2005. Have utilities switched to
chloramines? Journal of the American Water Works Association. 97(10): 87-97.
Selbes, M., D. Kim, N. Ates, and T. Karanfil. 2013. The roles of tertiary amine structure,
backround organic matter and chloramine species on NDMA formation. Water Research. 47:
945-953.
Selbes, M., D. Kim, and T. Karanfil. 2014. The effect of pre-oxidation on NDMA formation and
the influence of pH. Water Research. 66: 169-179.
Selbes, M., Glenn M., and T. Karanfil. 2015. The Role of Pre-oxidation in Controlling NDMA
Formation: A Review. In: Recent Advances in Disinfection By-Products, eds. Karanfil, T., B.
Mitch, P. Westerhoff, and Y. Xie, Chapter 9. ACS Symposium Series, 1190.
Sgroi, M., P. Roccaro, G.L. Oelker, and S.A. Snyder. 2014. A-nitrosodi methyl amine formation
upon ozonation and identification of precursors source in a municipal wastewater treatment
plant. Environmental Science & Technology. In press.
Sgroi, M., P. Roccaro, G.L. Oelker, and S.A. Snyder. 2015. A-nitrosodi methyl amine (NDMA)
formation at an indirect potable reuse facility. Water Research. 70: 174-183.
Shah, A.D. and W.A. Mitch. 2012. Halonitroalkanes, halonitriles, haloamides, andiV-
nitrosamines: a critical review of nitrogenous disinfection byproduct formation pathways.
Environmental Science & Technology. 46(1): 119-131.
Shah, A.D., S.W. Krasner, C.F.T. Lee, U. von Gunten, and W.A. Mitch. 2012. Trade-offs in
disinfection byproduct formation associated with precursor pre-oxidation for control of N-
nitrosodimethylamine formation. Environmental Science & Technology. 46(9): 4809-4818.
Shah, A.D., N. Dai, and W.A. Mitch. 2013. Application of ultraviolet, ozone, and advanced
oxidation treatments to washwaters to destroy nitrosamines, amines, and aldehydes formed
during amine based carbon capture. Environmental Science & Technology. 47: 2799-2808.
Sharma, V. K. 2002. Potassium ferrate(VI): an environmentally friendly oxidant. Advances in
Environmental Research. 6: 143-156.
Sharpless, C. M. and K. G. Linden. 2003. Experimental and model comparisons of low- and
medium-pressure Hg lamps for the direct and H2O2 assisted UV photodegradation of N-
nitrosodimethylamine in simulated drinking water. Environmental Science & Technology. 37(9):
1933-1940.
Six-Year Review 3 8-25
Technical Support Document for Nitrosamines
December 2016
-------
Shen, R. and S. Andrews. 2011a. Demonstration of 20 pharmaceuticals and personal care
products (PPCPs) as nitrosamine precursors during chloramine disinfection. Water Research.
45(2): 944-952.
Shen, R. and S. Andrews. 201 lb. NDMA formation kinetics from three pharmaceuticals in four
water matrices. Water Research. 45(17): 5687-5694.
Shen, R. and S. Andrews. 2013. Formation of NDMA from ranitidine and sumatriptan: The role
of pH. Water Research. 47: 802-810.
Shu, L. and P.F. Hollenberg. 1996. Role of cytochrom P450 in DNA damage induced by A'-
nitrosodialkylamines in cultured rat hepatocytes. Carcinogenesis. 17(3): 569-576 (as cited in Shu
and Hollenberg, 1997).
Shu, L. and P.F. Hollenberg. 1997. Alkylation of cellular macromolecules and target specificity
of carcinogenic nitrosodialkylamines: metabolic activation by cytochromes P450 2B1 and 2E1.
Carcinogenesis. 18(4): 801-810.
Standard Methods (SM). 2012a. SM 6450B. Carbonaceous-resin solid-phase extraction GC/MS
method. In Standard methods for the examination of water and wastewater, 22nd edition.
Washington, DC: American Public Health Association.
SM. 2012b. SM 6450C. Micro liquid-liquid extraction GC/MS method. In Standard methods for
the examination of water and wastewater, 22nd edition. Washington, DC: American Public
Health Association.
SM. 2012c. SM 6410B. Extractable base/neutrals and acids. In Standard methods for the
examination of water and wastewater, 22nd edition. Washington, DC: American Public Health
Association.
Soltermann, F., M. Lee, S. Canonica, andU. von Gunten. 2013. Enhanced A'-nitrosamine
formation in pool water by UV irradiation of chlorinated secondary amines in the presence of
monochloramine. Water Research. 47: 79-90.
Souliotis, V.L., J.H. van Delft, M.J. Steenwinkel, R.A. Baan, and S.A. Kyrtopoulos. 1998. DNA
adducts, mutant frequencies and mutation spectra in lambda lacZ transgenic mice treated with N-
nitrosodimethylamine. Carcinogenesis. 19(5): 731-739.
Steinle-Darling, E., M. Zedda, M.H. Plumlee, H.F. Ridgway, and M. Reinhard. 2007. Evaluating
the impacts of membrane type, coating, fouling, chemical properties and water chemistry on
reverse osmosis rejection of seven nitrosoalklyamines, including NDMA. Water Research.
41(17): 3959-3967.
Straif, K., D. Taeger, D. Holthenrich, Y. Sun, M. Bungers, and S.K. Weiland. (2000). Exposure
to N-nitrosamines, carbon black, asbestos, and talc and mortality from stomach, lung, and
laryngeal cancer in a cohort of rubber workers. American Journal of Epidemiology. 152(4): 297-
306 (as cited in CalEPA, 2006).
Six-Year Review 3 8-26
Technical Support Document for Nitrosamines
December 2016
-------
Swaim, P., A. Royce, T. Smith, T. Maloney, D. Ehlen, and B. Carter. 2008. Effectiveness of UV
advanced oxidation for destruction of micro-pollutants. Ozone: Science & Engineering: The
Journal of the International Ozone Association. 30(1): 34-42.
Symons, J.M., S.W. Krasner, L.A. Simms, and M. Sclimenti. 1993. Measurement of THM and
precursor concentrations revisited - the effect of bromide ion. Journal of the American Water
Works Association. 85(1): 51-62.
Takada, A., J. Nei, S. Takase, and Y. Matsuda. 1986. Effects of ethanol on experimental
hepatocarcinogenesis. Hepatology. 6: 65-72 (as cited in IARC, 2010).
Takayama, S. and K. Oota. 1965. Induction of malignant tumors in various strains of mice by
oral administration of iV-nitrosodimethylamine and A'-nitrosodiethylamine. Gann. 56: 189-199
(as cited in AT SDR, 1989).
Tang, H., P.C. Choy, and H. Chen. 1992. Changes in lipid content and composition during the
development of A'-nitrosodiethylamine induced hepatocarcinoma. Molecular and Cellular
Biochemistry. 109(1): 83-87.
Teefy, S., C.C. Chan, W. Wong, and L. Work. 2011. NDMA formation from gaskets used in
water storage tanks. WQTC 2011 Conference Proceedings.
Teefy, S., C.C. Chan, W. Wong, and L. Work. 2014. NDMA formation from gaskets used in
water storage tanks. Journal of the American Water Works Association. 106: 408-418.
Templeton, M. R., and Z. Chen. 2010. NDMA and seven other nitrosamines in selected UK
drinking water supply systems. Journal of Water Supply: Research and Technology-AQUA.
59(4): 277-283.
Terracini, B., G. Palestro, M.R. Gigliardi, G. Montesano, andR. Montesano. 1966.
Carcinogenicity of dimethylnitrosamine in Swiss mice. British Journal of Cancer. 20: 871-876
(as cited in AT SDR, 1989).
Terracini, B., P.N. Magee, and J.M. Barnes. 1967. Hepatic pathology in rats on low dietary
levels of dimethylnitrosamine. British Journal of Cancer. 21: 559-565.
Tinwell, H., P. A. Lefevre, and J. Ashby. 1994. Mutation studies with dimethyl nitrosoamine in
young and old lac I transgenic mice. Mutation Research. 307: 501-508.
Tomatis, L. 1973. Transplacental carcinogenesis. In Modern trends in oncology. Part I, ed.
Raven, R.W. London, England: Butterworths (as cited in CalEPA, 2006).
Tomatis, L, P.N. Magee, and P. Shubik. 1964. Induction of liver tumors in the Syrian golden
hamster by feeding dimethylnitrosamine. Journal of the National Cancer Institute. 33: 341-345
(as cited in CalEPA, 2006).
Tricker, A.R., B. Pfundstein, and R. Preussmann. 1994. Nitrosatable secondary amines:
Exogenous and endogenous exposure and nitrosation in-vivo. In Nitrosamines and relatedN-
nitro compounds: chemistry and biochemistry, eds. Leoppky, R.N. and C.J. Michejda, pp. 93-
101. Washington, DC: American Chemical Society.
Six-Year Review 3 8-27
Technical Support Document for Nitrosamines
December 2016
-------
Trogolo, D., B.K. Mishra, M.B. Heeb, U. von Gunten, and J.S. Arey. 2015. Molecular
mechanism of NDMA formation from A', A'-di m eth y 1 sulfainide during ozonation: Quantum
chemical insights into bromide-catalyzed pathway. Environmental Science & Technology. 49:
4163-4175.
Tsuda H., Y. Mera, K. Seki, T. Aoki, S. Fukushima, and N. Ito. 1987. Induction of tumors in the
liver, urinary bladder, esophagus and forestomach by short-term treatment with different doses of
N,N'-dibutylnitrosamine in rats. Japanese Journal of Cancer Research GANN. 78(3): 227-34.
United States Department of Energy (USDOE). 1997. Method OMIOOR, Analysis of
Semivolatile Organic Compounds Using Capillary Gas Chromatography with Ion Trap Mass
Spectrometric Detection. United States Department of Energy, Oak Ridge National Laboratory.
Available online at:
https://www.nemi.gov/apex/f?p=237:3 8:31572022093 50950: :::P38_METHOD_ID:4795.
United States Environmental Protection Agency (USEPA). 1987. A-Nitrosodi methyl amine. IRIS
(Integrated risk information system). Cincinnati, OH: U.S. EPA, Office of Research and
Development (ORD). Available online at: http://www.epa.gov/iris/subst/0045.htm.
USEPA. 1996a. 1996 Toxics Release Inventory: Public data release report. Chapter 3: Year-to-
Year Comparison of Toxics Release Inventory Data.
USEPA. 1996b. EPA Method 8260B, Volatile Organic Compounds by Gas
Chromatography/mass Spectrometry (GC/MS), SW-846 Online: Test Methods for Evaluating
Solid Waste, Physical/chemical Methods. Available online at:
https://www.epa.gov/sites/production/files/2015-12/documents/8260b.pdf.
USEPA. 1996c. EPA Method 5031, Volatile, Nonpurgeable, Water-soluble Compounds by
Azeotropic Distillation, SW-846. Available online at: https://www.epa.gov/hw-sw846/sw-846-
test-method-5031-volatile-nonpurgeable-water-soluble-compounds-azeotropic.
USEPA. 2001a. EPA Method 1625, Semivolatile Organic Compounds by Isotope Dilution
GCMS. 40 CFR 136, Appendix A.
USEPA. 2001b. EPA Method 607—Nitrosamines. 40 CFR 136, Appendix A.
USEPA. 2001c. EPA Method 625, Base/Neutrals and Acids. 40 CFR 136, Appendix A.
USEPA. 2003. Occurrence Estimation Methodology and Occurrence Findings for the Six-Year
Review of National Primary Drinking Water Regulations. EPA 815-R-03-006. June 2003.
USEPA 2004a. The Toxics Release Inventory (TRI) and Factors to Consider When Using TRI
Data. Available online at: https://www.epa.gov/sites/production/files/2015-
06/documents/factors_to_consider_6.15.15_final.pdf.
USEPA. 2004b. EPA Method 521, Version 1.0, 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 (CI-MS/MS). EPA 600-
R-05-054. September 2004.
Six-Year Review 3 8-28
Technical Support Document for Nitrosamines
December 2016
-------
USEPA. 2005a. Guidelines for Carcinogen Risk Assessment. Risk Assessment Forum,
Washington, DC. EPA 630-P-03-001B. March 2005.
USEPA. 2005b. Supplemental Guidance for Assessing Susceptibility from Early-Life Exposure
to Carcinogens. Risk Assessment Forum, Washington, DC. EPA 630-R-03-003F. March 2005.
USEPA. 2005c. Unregulated Contaminant Monitoring Regulation (UCMR) for Public Water
Systems Revisions; Proposed Rule. 70 FR 49094. August 22, 2005.
USEPA. 2005d. Economic Analysis for the Final Stage 2 Disinfectants and Disinfection
Byproducts Rule. EPA 815-R-05-010. December 2005.
USEPA. 2007a. EPA Method 8015C, Nonhalogenated Organics Using GC/FID, SW-846
Revision 3. Available online: https://www.epa.gov/sites/production/files/2015-
12/documents/8015c.pdf.
USEPA. 2007b. EPA Method 8270D, Semivolatile Organic Compounds by Gas
Chromatography/Mass Spectrometry (GC/MS), SW-846, Revision 4.
USEPA. 2007c. Unregulated Contaminant Monitoring Regulation (UCMR) for Public Water
Systems Revisions; Final Rule. 72 FR 367. January 4, 2007.
USEPA. 2007d. Simultaneous Compliance Guidance Manual for the Long-Term 2 and Stage 2
DBP Rules. EPA 815-R-07-017. March 2007.
USEPA. 2009. Final Contaminant Candidate List 3 Chemicals: Classification of the PCCL to
CCL. Office of Water. EPA 815-R-09-008.
USEPA. 2010. Technical Description of the Statistical Model of Perchlorate Occurrence in
Drinking Water. Peer reviewed draft report.
USEPA. 2011. Age Dependent Adjustment Factor Application. Washington, DC: Office of
Water Policy Document. Office of Science and Technology. Office of Water. Draft.
USEPA. 2012. Toxic Release Inventory (TRI) Explorer: Trends. Available online at:
http://www.epa.gov/triexplorer/trends.htm. Accessed March 9, 2012.
USEPA. 2014a. Announcement of Preliminary Regulatory Determinations for Contaminants on
the Third Drinking Water Contaminant Candidate List. 79 FR 62716. October 20, 2014.
USEPA. 2014b. Occurrence data from the second unregulated contaminant monitoring
regulation (UCMR 2). EPA 815-R-14-004. April 2014
USEPA. 2016a. Six-Year Review 3 Technical Support Document for Disinfectants/Disinfection
Byproducts Rules. EPA 810-R-16-012.
USEPA. 2016b. Six-Year Review 3 Technical Support Document for Chlorate. EPA 810-R-16-
013.
Six-Year Review 3 8-29
Technical Support Document for Nitrosamines
December 2016
-------
United States Geological Survey (USGS). 1983. Method 0-3118-83, Base/Neutral Extractable
Compounds, Total Recoverable, Gas Chromatographic/Mass Spectrometric. Available online at:
https://www.nemi.gov/apex/f?p=237:38:3157202209350950::::P38 METHOD ID:8963 .
Uzun, H., D. Kim, and T. Karanfil. 2015. Seasonal and temporal patterns of NDMA formation
potentials in surface waters. Water Research. 69: 162-172.
Valentine, R.L., J. Choi, Z. Chen, S.E. Barrett, C. Hwang, Y. Guo, M. Wehner, S. Fitzsimmons,
S.A. Andrews, A.G. Werker, C. Brubacher, and K. Kohut. 2005. Factors affecting the formation
of NDMA in water and occurrence. Water Research Foundation, Project # 2678..
Verdugo, E.M., C. Krause, Y. Han, T.E. Mattes, R.L. Valentine, and D.M. Cwiertny. 2014. N-
functionalized carbon nanotubes as a source and precursor of A'-nitrosodi methyl amine:
Implications for environmental fate, transport, and toxicity. Environmental Science &
Technology. 48: 9279-9287.
Verna, L., J. Shysner, and G.M. Williams. 1996. N-nitrosodiethylamine mechanistic data and
risk assessment: Bioactivation, DNA-adduct formation, mutagenicity, and tumor initiation.
Pharmacology and Therapeutics. 71(1-2): 57-81.
Verscheuren, K. 1996. Handbook of Environmental Data on Organic Chemicals. 3rd edition, pp.
829-830. New York, NY: Van Nostrand Reinhold.
Vesselinovitch, S.D., M. Koka, N. Mihailovich, andK.V. Rao. 1984. Carcinogenicity of
diethylnitrosamine in newborn, infant, and adult mice. Journal of Cancer Research and Clinical
Oncology. 108: 60-65.
Vogel, E.W. and M. J.M. Nivard. 1993. Performance of 181 chemicals in a Drosophila assay
predominantly monitoring interchromosomal mitotic recombination. Mutagenesis. 8(1): 57-81.
Wang, M., Y. Lao, G. Cheng, Y. Shi, P.W. Villalta, A. Nishikawa, and S.S. Hecht. 2007.
Analysis of adducts in hepatic DNA of rats treated with /V-nitrosopyrrolidine. Chemical Research
in Toxicology. 20(4): 634-640.
Wang, C., X. Zhang, J. Wang, S. Liu, C. Chen, and Y. Xie. 2013a. Effects of organic fractions
on the formation and control of A'-nitrosamine precursors during conventional drinking water
treatment processes. Science of the Total Environment. 449: 295-301
Wang, C., H. Shi, C.D. Adams, T. Timmons, and Y. Ma. 2013b. Investigation of removal of N-
nitrosamines and their precursors in water treatments using activated carbon nanoparticles.
International Journal of Environmental Technology and Management. 16(1-2): 34-48.
Wang, J., R.A. Dahlgren, M.S. Ersan, T. Karanfil, and A.T. Chow. 2015a. Wildfire altering
terrestrial precursors of disinfection byproducts in forest detritus. Environmental Science &
Technology. 49(10): 5921-5929
Wang, C., S. Liu, J. Wang, X. Zhang, and C. Chen. 2015b. Monthly survey of A'-nitrosamine
yield in a conventional water treatment plant in North China. Journal of Environmental Sciences.
38: 142-149.
Six-Year Review 3 8-30
Technical Support Document for Nitrosamines
December 2016
-------
Wang, X., H. Yang, B. Zhou, X. Wang, and Y. Xie. 2015c. Effect of Oxidation on Amine Based
Pharmaceutical Degradation and N-Nitrosodimethylamine Formation. Water Research. 87: 403-
411.
Wang, W., Y. Guo, Q. Yang, Y. Huang, C. Zhu, J. Fan, and F. Pan. 2015d. Characterization of
the microbial community structure and nitrosamine-reducing isolates in drinking water biofilters.
Science of the Total Environment. 521-522: 219-225.
Webster, T.S., C. Condee, and P.B. Hatzinger. 2013. Ex situ treatment of N-
nitrosodimethylamine (NDMA) in groundwater using a fluidized bed reactor. Water Research.
47 :811-820.
Westerhoff, P. and H. Mash. 2002. Dissolved organic nitrogen in drinking water supplies: A
review. Journal of Water Supply Research and Technology - AQUA. 41(8): 415-448.
Westerhoff, P., W. Lee, J. Croue, H. Gallard, and G. Amy. 2006. Organic nitrogen in drinking
water and reclaimed wastewater. AWWARF. 91116.
Wilczak, A., A. Assadi-Rad, H.H. Lai, L.L. Hoover, J.F. Smith, R. Berger, F. Rodigari, J.W.
Beland, L.J. Lazzelle, E.G. Kincannon, H. Baker, and C.T. Heaney. 2003. Formation of NDMA
in chloraminated water coagulated with DADMAC cationic polymer. Journal of the American
Waterworks Association. 95(9): 94- 106.
Woods, G.C. and E.R.V. Dickenson. 2015. Evaluation of the final UCMR2 database: Nationwide
trends in NDMA. Journal of the American Water Works Association. 107(1): E58-E68.
Woods, G.C., R.A. Trenholm, B. Hale, Z. Campbell, and E.R.V. Dickenson. 2015. Seasonal and
spatial variability of nitrosamines and their precursor sources at a large-scale urban drinking
water system. Science of the Total Environment. 520: 120-126.
World Health Organization (WHO). 2002. Concise international chemical assessment document
38, A'-nitrosodi methyl amine. Geneva, Switzerland: World Health Organization. Available online
at: http://www.inchem.org/documents/cicads/cicads/cicad38.htm.
WHO. 2008. N-Nitrosodimethylamine in drinking water. Background document for development
of WHO guidelines for drinking-water quality. Geneva, Switzerland: World Health
Organization. Available online at:
http://www.who.int/water_sanitation_health/dwq/chemicals/ndma_2add_feb2008.pdf
Wu, Q., H. Shi, Y. Ma, C. Adams, H. Jiang, J. Wang, T. Eichholz, and T. Timmons. 2015.
Removal of N-nitrosamine precursors in drinking water system using adsorption methods.
Separation and Purification Technology. 156: 972-979.
Xu, B., Z. Chen, F. Qi, and L. Yang. 2008. Photodegradation of A'-nitrosodi ethyl amine in water
with UV irradiation. Chinese Science Bulletin. 53(21): 3395-3401.
Xu, B., Z. Chen, F. Qi, J. Ma, and F. Wu. 2009a. Rapid degradation of new disinfection by-
products in drinking water by UV irradiation: A'-nitrosopyrrolidine and A'-nitrosopi peri dine.
Separation and Purification Technology. 69: 126-133.
Six-Year Review 3 8-31
Technical Support Document for Nitrosamines
December 2016
-------
Xu, B., Z. Chen, F. Qi, J. Ma, and F. Wu. 2009b. Inhibiting the regeneration of N-
nitrosodimethylamine in drinking water by UV photolysis combined with ozonation. Journal of
Hazardous Materials. 168(1): 108-114.
Xu, B., Z. Chen, F. Qi, J. Ma, and F. Wu. 2010. Comparison of iV-nitrosodiethylamine
degradation in water by UV irradiation and UV/O3: Efficiency, product and mechanism. Journal
of Hazardous Materials. 179(1-3): 976-982.
Xu, B., T. Ye, D P-. Li, C.-Y. Hu, Y.-L. Lin, S.-J. Xia, F.-X. Tiang, andN.-Y. Gao. 2011.
Measurement of dissolved organic nitrogen in a drinking water treatment plant: Size fraction,
fate, and relation to water quality parameters. Science of the Total Environment. 409(6): 1116-
1122.
Yahagi, T., M. Nagao, Y. Seino, T. Matsushima, and T. Sugimura. 1977. Mutagenicities of A'-
nitrosamines on Salmonella. Mutat. Res. 48: 121-130 (as cited in Aiub et al., 2003).
Yamazaki, H., Y. Inui, C.H. Yun, F.P. Guengerich, and T. Shimada. 1992a. Cytochrome
P4502E1 and 2A6 enzymes as major catalysts for metabolic activation of N-
nitrososdialkylamines and tobacco-related nitrosamines to human liver and microsomes.
Carcinogenesis. 13: 1789-1794 (as cited in Verna et al., 1996).
Yamazaki, H., Y. Oda, Y. Funai, S. Imaoka Y. Inui, F.P. Guengerich, and T. Shimada. 1992b.
Participation of rat liver cytochrome P4502E1 in the activation of A'-nitrosodi methyl amine and
A'-nitrosodi ethyl amine to products genotoxic in an acetyltransferase-overexpressing Salmonella
typhimurium strain (NM2009). Carcinogenesis. 13: 979-985 (as cited in Verna et al., 1996).
Yang, L., Z. Chen, J. Shen, Z. Xu, H. Liang, J. Tian, Y. Ben, X. Zhai, W. Shi, and G. Li. 2009.
Reinvestigation of the nitrosamine-formation mechanism during ozonation. Environmental
Science & Technology. 43(14): 5481-5487.
Yang, L., D. Kim, H. Uzun, T. Karnfil, and J. Hur. 2015. Assessing trihalomethanes (THMs) and
A'-nitrosodi methyl amine formation potentials in drinking water treatment plants using
fluorescence spectroscopy and parallel factor analysis. Chemosphere. 121: 84-91.
Yoon, S., and H. Tanaka. 2014. Formation of A'-nitrosamines by chloramination of amines listed
in pollutant release and transfer registers. Chemosphere. 95: 88-95.
Yurchenko, S. and U. Molder. 2007. The occurrence of volatile A'-nitrosamines in Estonian meat
products. Food Chemistry. 100: 1713-1721.
Zeiger, E., B. Anderson, S. Haworth, T. Lawlor, K. Mortelmans, and W. Speck. 1987.
Salmonella mutagenicity tests. III. Results from the testing of 255 chemicals. Environmental
Mutagenesis. 9: 1-110 (as cited in Aiub et al., 2003).
Zeng. T. and W.A. Mitch. 2015. Contribution of A'-nitrosamines and their precursors to sewage
by greywaters and blackwaters. Environmental Science & Technology. 49: 13158-13167.
Zeng, T., J.J. Pignatello, R.J. Li, and W.A. Mitch. 2014. Synthesis and Application of a
Quaternary Phosphonium Polymer to Avoid N-Nitrosamine Formation. Environmental Science
& Technology. 48: 13392-13401.
Six-Year Review 3 8-32
Technical Support Document for Nitrosamines
December 2016
-------
Zha, X., Y. Liu, Q. Zhang, R. Dai, L. Ying, J. Wu, J. Wang, and L. Ma. 2014. Effects of bromide
and iodide ions on the formation of disinfection byproducts during ozonation and subsequent
chlorination of water containing biological source matters. Environmental Science Pollution
Research. 22: 2714-2723.
Zhang, A., Y. Li, Y. Song, J. Lv, and J. Yang. 2014. Characterization of pharmaceuticals and
personal care products as A-nitrosodi methyl amine precursors during disinfection processes using
free chlorine and chlorine dioxide. Journal of Hazardous Materials. 276: 499-509.
Zhao, Y., L.-S. Cai, Z. Jing, H. Wang, J.X. Yu, and H.-S. Zhang. 2003. Determination of
aliphatic amines using A'-succinimidyl benzoate as a new derivatization reagent in gas
chromatography combined with solid-phase microextraction. Journal of Chromatography A.
1021(1-2): 175-181.
Zhao, Y., J.M. Boyd, M. Woodbeck, R.C. Andrews, F. Qin, S.E. Hrudey, and X. Li. 2008.
Formation of iV-nitrosamines from eleven disinfection treatments of seven different surface
waters. Environmental Science & Technology. 42(13): 4857-4862.
Zhou, W., C. Chen, L. Lou, Q. Yang, and L. Zhu. 2014. Formation potential of nine nitrosamines
from corresponding secondary amines by chloramination. Chemosphere. 95: 81-87.
Zhu, J. H., D. Yan, J.R. Xai, L.L. Ma, and B. Shen. 2001. Attempt to adsorb iV-nitrosamines in
solution by use of zeolites. Chemosphere. 44(5): 949-956.
Zielenska, M., A. Ahmed, andB.W. Glickman. 1990. Mutational specificities of environmental
carcinogens in the Lac I gene of Escherichia coli (III). The cyclic nitrosamine N-
nitrosopyrrolidine is a complex mutagen. Molecular Carcinogenesis. 3(3): 122-125.
Six-Year Review 3 8-33
Technical Support Document for Nitrosamines
December 2016
------- |