Recent Advances in Disinfection By-Products

Recent Advances in
Publication Date (Web): August 24, 2015 | doi: 10.1021/bk-2015-1190.fw001
Disinfection By-Products
In Recent Advances in Disinfection By-Products; Xie, et al.;

ACS Symposium Series; American Chemical Society: Washington, DC, 2015.

ACS SYMPOSIUM SERIES 1190
Recent Advances in
Tanju Karanfil, Editor

Clemson University
Anderson, South Carolina
Bill Mitch, Editor

Stanford University
Stanford, California
Paul Westerhoff, Editor

Arizona State University
Tempe, Arizona
Yuefeng Xie, Editor

The Pennsylvania State University
Middletown, Pennsylvania
Sponsored by the
ACS Division of Environmental Chemistry, Inc.
American Chemical Society, Washington, DC
Distributed in print by Oxford University Press

Library of Congress Cataloging-in-Publication Data
Recent advances in disinfection by-products / Tanju Karanfil, editor, Clemson University,

Anderson, South Carolina [and three others] ; sponsored by the ACS Division of
Environmental Chemistry, Inc.
pages cm. -- (ACS symposium series ; 1190)
Includes bibliographical references and index.
ISBN 978-0-8412-3076-7 -- ISBN 978-0-8412-3077-4 1. Disinfection and disinfectants.
I. Karanfil, Tanju. II. American Chemical Society. Division of Environmental Chemistry.
RA765.R43 2015
614.4′8--dc23
The paper used in this publication meets the minimum requirements of American National
Standard for Information Sciences—Permanence of Paper for Printed Library Materials,
ANSI Z39.48n1984.
Copyright © 2015 American Chemical Society
Distributed in print by Oxford University Press
All Rights Reserved. Reprographic copying beyond that permitted by Sections 107 or 108
of the U.S. Copyright Act is allowed for internal use only, provided that a per-chapter fee of
$40.25 plus $0.75 per page is paid to the Copyright Clearance Center, Inc., 222 Rosewood
Drive, Danvers, MA 01923, USA. Republication or reproduction for sale of pages in this
book is permitted only under license from ACS. Direct these and other permission requests
to ACS Copyright Office, Publications Division, 1155 16th Street, N.W., Washington, DC
20036.
The citation of trade names and/or names of manufacturers in this publication is not to be
construed as an endorsement or as approval by ACS of the commercial products or services
referenced herein; nor should the mere reference herein to any drawing, specification,
chemical process, or other data be regarded as a license or as a conveyance of any right
or permission to the holder, reader, or any other person or corporation, to manufacture,
reproduce, use, or sell any patented invention or copyrighted work that may in any way be
related thereto. Registered names, trademarks, etc., used in this publication, even without
specific indication thereof, are not to be considered unprotected by law.
PRINTED IN THE UNITED STATES OF AMERICA

Foreword
The ACS Symposium Series was first published in 1974 to provide a
mechanism for publishing symposia quickly in book form. The purpose of
the series is to publish timely, comprehensive books developed from the ACS
sponsored symposia based on current scientific research. Occasionally, books are
developed from symposia sponsored by other organizations when the topic is of

keen interest to the chemistry audience.
Before agreeing to publish a book, the proposed table of contents is reviewed

for appropriate and comprehensive coverage and for interest to the audience. Some
papers may be excluded to better focus the book; others may be added to provide
comprehensiveness. When appropriate, overview or introductory chapters are
added. Drafts of chapters are peer-reviewed prior to final acceptance or rejection,
and manuscripts are prepared in camera-ready format.
As a rule, only original research papers and original review papers are
included in the volumes. Verbatim reproductions of previous published papers
are not accepted.
ACS Books Department

Preface
This book is part of a continuing series of books based on the chlorination
conferences of the 1970s and 1980s and the more recent American Chemical
Society (ACS) symposia on natural organic matter (NOM) and disinfection
by-products (DBPs). These books have reflected the state-of-the-art in disinfection
Publication Date (Web): August 24, 2015 | doi: 10.1021/bk-2015-1190.pr001
by-products (DBP) research around the world. The current book is based on
a symposium entitled Occurrence, Formation, Health Effects and Control of
Disinfection By-Products in Drinking Water, held at the 248th National Meeting
of the ACS Division of Environmental Chemistry, in San Francisco, California,
on August 10-14, 2014.
The formation, control, and health effects of DBPs in drinking water are
issues of international concern because of the health effects (e.g., bladder
cancer and potential adverse reproductive-development impacts) associated with
exposure to certain DBPs. As a result, many countries, as well as the World Health
Organization, have regulations and/or guidelines on acceptable concentrations of
DBPs in water.
In recent years, DBP research worldwide has focused on determining
the possible adverse health effects of emerging, yet unregulated, DBPs,
specifically halogenated (e.g., iodinated) and non-halogenated nitrogenous (e.g.,
nitrosamines) DBPs. The breadth of DBP research is very broad from source
waters (e.g., wastewater, wildfire, seawater intrusion influences) to treatment
strategies and technologies, followed by distribution system and point of entry
issues (e.g., biofilms, heating, swimming pools), as well as health effects and
analytical method developments. Recent research is helping to understand factors
controlling formation and to develop a cost-effective control of a wide range of
regulated and emerging DBPs. Furthermore, the pace of research on emerging
DBP toxicity has increased and generated diverse findings, with comparative
toxicity and the molecular mechanisms leading to improved understanding of
their toxicity pathways and potential adverse biological effects.
This book represents the latest research efforts to understanding these
important DBP-related issues. The authors of the chapters in this book are a
multidisciplinary group of scientists and engineers, who are conducting studies
in many parts of the world. The chapters in this book address both regulated and
emerging DBPs and are organized under the sections on DBP toxicology and
health effects, modeling of DBP formation, precursors and reactions involving
nitrosamines, and formation of halogenated DBPs. This book will be of interest
to researchers, drinking water utility scientists and engineers, toxicologists,
epidemiologists, and regulators interested in the formation and control of and
exposure to DBPs.
xi
We wish to thank all individuals who served as peer reviewers of our
chapter manuscripts. We also wish to gratefully acknowledge the Water Research
Foundation for their financial assistance in support of the symposium, and the
Association of Environmental Engineering and Science Professors for their
assistance in disseminating information on the symposium.
Tanju Karanfil
Department of Environmental Engineering and Earth Sciences
Clemson University
Anderson, South Carolina 29625
Publication Date (Web): August 24, 2015 | doi: 10.1021/bk-2015-1190.pr001
Bill Mitch
Department of Civil and Environmental Engineering
Stanford University
Stanford, California 94305
Paul Westerhoff
School of Sustainable Engineering and The Built Environment
Arizona State University
Tempe, Arizona 85287-3005
Yuefeng Xie
Civil and Environmental Engineering Programs
Capital College
The Pennsylvania State University
Middletown, Pennsylvania 17057
xii
Chapter 1
Charting a New Path To Resolve the Adverse

Health Effects of DBPs
Michael J. Plewa* and Elizabeth D. Wagner
Publication Date (Web): August 24, 2015 | doi: 10.1021/bk-2015-1190.ch001
Department of Crop Sciences and the Center of Advanced Materials for the
Purification of Water with Systems, Safe Global Water Institute, University
of Illinois at Urbana-Champaign, Urbana, Illinois 61801
*E-mail: [email protected].
The disinfection of drinking water was an outstanding

public health achievement of the 20th century. Disinfection
by-products (DBPs) are the unintended consequence of
reactions between disinfectants with organic and halide
precursors in the source water. Exposure to DBPs is associated
with unfavorable health effects from cancer induction to
adverse pregnancy outcomes, yet the forcing agents responsible
are largely unknown. Of over 600 DBPs identified, only 11
are regulated by the U.S. EPA and in many studies these are
not the most toxic DBPs. Identified DBPs represent less than
half of the total organic halogen in drinking water and only
~80 DBPs have been evaluated by systematic quantitative
toxicological analyses. Epidemiological studies are hampered
by problems with exposure metrics and the suite of DBPs
present in drinking water. Although the last decade experienced
increased interdisciplinary collaborations amongst chemists,
biologists, epidemiologists and engineers, resolving the risks of
DBPs follows a dated paradigm. A new integrated approach is
required to determine the contaminants in source and drinking
waters that increase health risks and to provide the foundation
for novel disinfection practices. We suggest the following areas
in resolving adverse health effects of DBPs: 1) identification
and occurrence of emerging DBPs especially iodinated-DBPs
and nitrogenous-DBPs, 2) quantitative, comparative in vitro
© 2015 American Chemical Society

toxicology of DBPs and DBP classes, 3) a nationwide analysis
on the toxicity of drinking waters, 4) determining the impact
of DBP precursors and mixture effects on water toxicity, 5)
understanding the molecular mechanisms of DBP toxicity, 6)
identification of sensitive subpopulations to DBP toxicity, 7)
employment of analytical chemistry and toxicity information as
an essential part of epidemiological studies, and 8) engineering
research on disinfection based on toxic outcomes.
Introduction
The disinfection of drinking water to reduce waterborne disease was

arguably the greatest public health achievement of the 20th century (1). Chemical
disinfectants inactivate pathogens in source waters; however, an unintended
consequence is their reaction with natural organic matter (NOM), anthropogenic
contaminants and bromide/iodide to form disinfection by-products (DBPs) (2).
Factors including the concentration and type of organic matter, pH, temperature,
disinfectant type and concentration, and contact time affect the formation of DBPs
(3, 4). The most widely employed disinfectants include chlorine, chloramines,
chlorine dioxide, and ozone; each disinfectant generates DBPs with a different
spectra of chemical classes (5, 6). Since their discovery (7, 8) over 600 DBPs
have been chemically characterized (9). This number of DBPs represents only
a fraction of the total organic halogen (TOX) generated in disinfected water.
Of this number, approximately 80 have undergone systematic, quantitative,
comparative toxicological analyses (10–12). By extrapolation from TOX, the
total number of DBPs may reach over 2,000. This may be an analytical chemist’s
dream in the search to identify all possible DBPs. However, it becomes an
analytical biologist’s dilemma to choose a DBP class or individual compound
for comprehensive toxicological analyses. Unfortunately, DBPs represent a
regulatory nightmare; of the hundreds of possibly highly toxic DBPs, the U.S.
EPA currently regulates only 11 agents (10, 13). A large number of DBPs are
cytotoxic, neurotoxic, mutagenic, genotoxic, carcinogenic and teratogenic (10).
Epidemiological research demonstrated low but significant associations between
disinfected drinking water and adverse health effects (10) including cancer of the
bladder (14–16), colon (17, 18) and rectum (19). Some studies report a weak
association with adverse pregnancy outcomes and DBPs; yet, the evidence is
inconclusive (20–23).
Research into the formation, occurrence, toxicity, epidemiology and
engineering processes have resulted in enhanced interest in the adverse impact
of DBPs on the environment and the public health (12, 24–35). Although
the last decade experienced increased interdisciplinary collaborations amongst
chemists, biologists, epidemiologists and engineers, resolving the risks of DBPs
follows a dated paradigm. A new integrated approach is required to determine
the contaminants in source and drinking waters that increase health risks and to
provide the foundation for novel disinfection practices.
4
Charting a New Pathway for Integrated DBP Research
Although there has been an increase in interdisciplinary studies, current
DBP research is in many ways unintegrated. High quality work continues to be
published on the formation and identification of emerging DBPs, comparative
toxicology and the molecular mechanisms of DBP toxicity, exposure assessment,
risk assessment and epidemiology as well as novel engineering processes to
control DBPs in drinking water. Yet the future demands that we address complex
questions regarding DBPs. Some of these questions include the following.
What are the environmental and public health impacts of unidentified DBPs
(~70% TOX)? What are the important DBP classes or individual agents that have
epidemiological significance? Can we predict human health risks if we know the
biological mechanisms of DBP toxicity? Can we integrate research in analytical
chemistry, analytical biology and exposure assessment to enhance the resolution

of epidemiological studies on the adverse health effects of DBPs? Can we use
biological and chemical data to optimize engineering processes for wastewater
reuse and desalination? These questions and indeed other crucial questions
associated with DBPs may be addressed most efficiently by well planned, focused
interdisciplinary science. In this paper we argue for a new pathway of integrated
DBP research that includes the following.
• A nationwide drinking water in vitro toxicity survey.

• Analytical chemical analyses of the DBPs from the surveyed water
samples expressing the highest levels of toxicity.
• Identification of the forcing agents associated with this toxicity.
• Definition of the molecular mechanisms of toxicity for the DBPs
identified as forcing agents.
• Identification of selected human biomarkers associated with the
molecular mechanisms of toxic DBP forcing agents.
• Identification of susceptible human subpopulations based on bio-markers.
• Definition of the modes of DBP exposure.
• Development of focused epidemiologic studies on exposed populations
and integration of data with DBP exposure, toxicity, biomarkers and
adverse health outcomes.
The information derived from this integrated interdisciplinary approach may

form a cornerstone for optimizing engineering processes for water disinfection.
Nationwide Drinking Water in Vitro Toxicity Survey
In 2002 the U.S. EPA published a landmark study commonly referred to as the
Nationwide Occurrence Study (36, 37). In this study, novel quantitative analytical
methods were developed for measuring the priority DBPs; the occurrence of a
wide variety of DBPs in drinking waters throughout the U.S. was determined; the
effect of source water and treatment conditions on DBP formation was measured;
the fate and transport of these DBPs in the distribution system were studied; and
the identification of new DBPs was reported. This study clearly demonstrated
5
that hundreds of DBPs were differentially distributed in drinking waters within
the U.S. However, the Nationwide Occurrence Study did not identify the toxicity
of the water samples. In concert with the data generated by the Nationwide
Occurrence Study, our laboratory developed mammalian cell microplate-based
assays and generated in vitro, systematic, analytical, comparative databases
on the chronic cytotoxicity and genotoxicity of individual DBPs. To date we
have analyzed 87 DBPs (Table I). Their comparative cytotoxicity in Chinese
hamster ovary (CHO) cells (LC50) is presented in Figure 1 (carbon-based DBPs)
and Figure 2 (nitrogen-containing DBPs). Their comparative genotoxicity is
presented in Figure 3. These data provide a direct comparison of the cytotoxicity
and genotoxicity of DBPs within a chemical class and among chemical classes
and allow for detailed structure activity relationships to be studied by integrating
analytical chemistry and analytical biology. The data for Figures 1-3 were
published (12, 24–27, 29–31, 33, 38–43) and represent the largest toxicity
database for DBPs. This work as well as recent bioanalytical research will
provide a foundation to identify forcing agents that contribute to the toxicity of
disinfected water.
Analytical Chemistry and the Identification of the Forcing Agents of

Drinking Water Toxicity
To use these data most effectively, we propose that a nationwide in vitro
toxicity survey of drinking waters be conducted. This study would employ
bioanalytical assays (12, 44) to identify U.S. drinking waters that express
enhanced in vitro toxicity. Source water characteristics such as pH, industrial
pollutants, pharmaceutical contaminants, total organic carbon (TOC), dissolved
organic carbon (DOC), specific UV absorbance, concentrations of I- and Br-,
disinfectant concentration, coagulants (polyDADMAC) and engineering practice
all play a role in the toxicity of finished drinking water (2, 5, 6, 45–54).
Detailed analytical chemical studies for DBPs would be conducted with
drinking waters that induce the highest adverse biological effects. The forcing
agents (DBPs or classes of DBPs) that are associated with the toxicity would be
identified and rank ordered (2).
Molecular Mechanisms of Action of the DBP Forcing Agents Associated

with Enhanced Toxicity of Drinking Water
The regulated and unregulated DBPs that are identified as the forcing agents
for drinking water toxicity would be the primary candidates for investigating
their molecular modes of action. Currently few molecular mechanisms for
specific DBP-induced toxicity have been reported. Although many of the adverse
biological effects involved the induction of oxidative stress, the mechanisms that
generate reactive oxygen species (ROS) follow different pathways and the DBPs
may interact with different cellular targets to induce damage. ROS is the resulting
insult that causes cellular dysfunction and is a phenomenon, not a mechanism per
se. Cytotoxicity induced by bromate involves ROS-mediated mitogen activated
protein kinases (MAPK) activation (55). The halobenzoquinones generate redox
6
reactions producing reactive metabolites that damage biological macromolecules
(56). The primary cellular target molecule for the monohaloacetic acids
(monoHAAs) is glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (57).
Inhibition of GAPDH leads to pyruvate starvation and mitochondrial stress that
generates ROS (58, 59). New research demonstrated that the monohalonitriles,
at non-cytotoxic concentrations, impact the M-phase of the cell cycle and induce
hyperploidy in mammalian cells (35). These examples indicate that the molecular
mechanisms of DBP toxicity involve a diversity of cellular targets and affect a
host of different biological pathways.
Identification of Human Biomarkers Associated with Molecular Mechanisms

of Toxic DBPs
In the context of this paper, a biomarker refers to a measurable indicator of

an adverse biological state or condition induced or influenced by exposure to
a toxic DBP. Biomarkers associated with a cellular mode of toxic action may
provide a vehicle to identify adverse effects in an exposed human population.
One promising approach is toxicogenomic analyses. With the combination of
toxicology and genomics, toxicogenomics represents a powerful interdisciplinary
tool to analyze the modulation of gene expression after exposure to a toxic agent.
Metabolic pathways involved in the cellular responses to toxins can be identified
and provide insight on the underlying biological mechanisms. Toxicogenomic
research has identified a number of genes that demonstrate altered expression
related to DBP exposure. In vitro toxicogenomic analyses of DBPs identified
as forcing agents in drinking water toxicity would have high resolving power to
identify human biomarkers that may be useful in selecting sensitive populations
for epidemiological studies. To provide high resolving power for toxicological
metrics, in vitro quantitative comparative toxicogenomic analyses must be
conducted rather than screening studies (60). These experimental designs will
demand that the genomic studies are conducted at non-cytotoxic levels. If DBP
concentrations induce cell death, the RNA extracted for microarray analysis
would be from dead or dying cells and may not reflect its molecular mechanism
of toxicity. Also many toxicogenomic studies employ human tumor cell lines
because they are easy to grow and yield large amounts of RNA (61). However,
such studies may be difficult to interpret due to the genetic instability of tumor
cell lines and their inherent aberrant gene expression. Because of these concerns
we encourage the use of nontransformed human cells and treatment conditions
with no or low cytotoxicity at equivalent biological effects. In order to generate
quantitative comparative toxicogenomic analyses among a series of DBPs it is
important to generate transcriptome profiles from cells that were exposed to
agents at concentrations normalized to generate equivalent biological responses.
As illustrated in Figure 4, the concentrations of IAA, BAA and CAA that induced
equivalent genomic DNA damage (50%Tail DNA) were 22 µM, 57 µM and
3.4 mM, respectively (62). Employing concentrations that induce equivalent
biological effects allow for the normalization of adverse biological effects and
transcriptome profiles within a comparative experimental design and leads to
more meaningful toxicogenomic information.
7
8
Figure 1. CHO cell chronic cytotoxicity analyses of carbon based DBPs. The abbreviations for the DBPs are listed in Table I. Each DBP is
repre-sented as its LC50 concentration (50% cell density as compared to the negative control).

9
Figure 2. CHO cell chronic cytotoxicity analyses of N-DBPs. The abbreviations for the DBPs are listed in Table I. Each DBP is represented
as its LC50 concentration (50% cell density as compared to the negative control).

10
Figure 3. CHO cell genotoxicity analyses of DBPs. The abbreviations for the DBPs are listed in Table I. Each DBP is represented as its
midpoint of the SCGE genotoxicity concentration-response curve or its 50%Tail DNA metric.

Table I. DBPs Analyzed for CHO Cell Cytotoxicity and Genotoxicity and
Abbreviations
Abbreviation DBP Abbreviation DBP
HAA Haloacetic Acids THM Halomethanes
BAA Bromoacetic acid TIM Triiodomethane
CAA Chloroacetic acid DBIM Dibromoiodomethane
IAA Iodoacetic acid BCIM Bromochloroiodomethane
DBAA Dibromoacetic acid TBM Tribromomethane
DCAA Dichloroacetic acid CDBM Chlorodibromomethane
DIAA Diiodoacetic acid TCM Trichloromethane
BCAA Bromochloroacetic acid BDCM Bromodichloromethane

BIAA Bromoiodoacetic acid DCIM Dichloroiodomethane
CIAA Chloroiodoacetic acid BDIM Bromodiiodomethane
TBAA Tribromoacetic acid CDIM Chlorodiiodomethane
TCAA Trichloroacetic acid
Dibromochloroacetic
DBCAA HAL Haloacetaldehydes
acid
Bromodichloroacetic
BDCAA BAL Bromoacetaldehyde
acid
CAL Chloroacetaldehyde
HA Halo Acids IAL Iodoacetaldehyde
3,3-Dibromo-4-
DB4OPA DBAL Dibromoacetaldehyde
oxopentanoic acid
3,3-Dibromochloro-4-
DBC4OPA DCAL Dichloroacetaldehyde
oxopentanoic acid
2-Iodo-3-bromo-
2I3BPPA BCAL Bromochloroacetaldehyde
propenoic acid
3-Iodo-3-
3I3BPPA TBAL Tribromoacetaldehyde
bromopropenoic acid
3,3-Dibromopropenoic
3,3DBPPA TCAL Trichloroacetaldehyde
acid
2-Iodo-3- Dibromochloroacetalde-
2I3MBDA DBCAL
methylbutenedioic acid hyde
TBPPA Tribromopropenoic acid BDCAL Bromodichloroacetaldehyde
2-Bromobutenedioic
BBTDA
acid
2,3-Dibromopropenoic
2,3DBPPA HNM Halonitromethanes
acid
Continued on next page.
11
Table I. (Continued). DBPs Analyzed for CHO Cell Cytotoxicity and
Genotoxicity and Abbreviations
2-Bromo-3-
2B3MBTDA BNM Bromonitromethane
methylbutenedioic acid
CNM Chloronitromethane
DBNM Dibromonitromethane
HPYL Halophenolics DCNM Dichloronitromethane
BCNM Bromochloronitromethane
4-Hydroxy-3-iodo-1-
4HIPOLA TBNM Tribromonitromethane
phenolic acid
TIPOL 2,4,6-Triiodo-1-phenol TCNM Trichloronitromethane
Bromodichloroni-
BDCNM
tromethane
Dibromochloroni-
HAN Haloacetonitriles DBCNM
tromethane
BAN Bromoacetonitrile
CAN Chloroacetonitrile HAcAm Haloacetamides
IAN Iodoacetonitrile BAcAm Bromoacetamide
DBAN Dibromoacetonitrile CAcAm Chloroacetamide
DCAN Dichloroacetonitrile IAcAm Iodoacetamide
Bromochloroacetoni-
BCAN DBAcAm Dibromoacetamide
trile
TCAN Trichloroacetonitrile DCAcAm Dichloroacetamide
DIAcAm Diiodoacetamide
CN Cyanogen Halides BIAcAm Bromoiodoacetamide
BN Cyanogen Bromide CIAcAm Chloroiodoacetamide
CN Cyanogen Chloride BCAcAm Bromochloroacetamide
IN Cyanogen Iodide TBAcAm Tribromoacetamide
TCAcAm Trichloroacetamide
NA Nitrosamines DBCAcAm Dibromochloroacetamide
N-nitrosodimethy-
NDMA BDCAcAm Bromodichloroacetamide
lamine
NDPhA N-nitrosodiphenylamine N2DCAcAm N,N-Dichloroacetamide
NPIP N-nitrosopiperidine NCAcAm N-Chloroacetamide
NMOR N-nitrosomorpholine
12
Table I. (Continued). DBPs Analyzed for CHO Cell Cytotoxicity and
Genotoxicity and Abbreviations
Other DBPs
Bromate Bromate
3-chloro-4(dichloromethyl)-
MX
5-hydroxy-2[5H]-furanone
4-Hydroxy-3,5-diiodo-1-
4HDINB
nitrobenzene
TBPy Tribromopyrrole
We conducted a series of quantitative gene array studies on the monoHAAs

using quantitative real-time PCR (qRT-PCR) focused on pathways involving
DNA damage and repair as well as pathways involved in responses to
oxidative stress. Examples of pathways uncovered by transcriptome profiles
of nontransformed human small intestine epithelial cells, line FHs 74 Int, after
exposure to non-cytotoxic concentrations of the monoHAAs is presented in Table
II (59, 62). Expression in DNA repair genes were modulated especially those
involved in double strand DNA (dsDNA) damage repair (62, 63). These data are
especially important because there is a strong association between p53, ATM,
Brca1 and Brca2 proteins and dsDNA break repair pathways and tumorigenesis
(64). Recently IAA was reported to induce malignant transformation and tumor
induction (65).
Figure 4. The selection of concentrations of the monoHAAs that generated

equivalent biological effects (genomic DNA damage) in nontransformed human
FHs74 Int cells for use in toxicogenomic studies.
13
Table II. MonoHAA-Induced Transcriptome Profile Pathways
Cell Pathway Monohaloacetic
Acids
BAA CAA IAA
ATM Signaling Pathway ● ● ●
Cell Cycle Control ● ●
Cyclins and Cell Cycle Regulation ●
FC Epsilon RI Signaling Pathway ●
MAPK Signaling Pathway ● ●
p53 Signaling Pathway ● ●
BRCA1, BRCA2 and ATR Mediated ● ● ●

Cancer Susceptibility and dsDNA
Repair Pathways
Nrf2/ARE-Dependent ROS Pathways ● ● ●
PTGS2 (COX2)-Mediated Pathways ● ● ●
MPO, LPO and NOX5 ROS Pathways ● ● ●
GRS/GSR Pathway ● ● ●
Peroxiredoxin Oxidative Stress Pathway ● ●
Toxicogenomic profiles combined with enzyme inhibition studies and ATP

quantification demonstrated the impact of the monoHAAs on energy metabolism
(Table II) (58, 59). These studies uncovered other relevant pathways and potential
human biomarkers for DBP-mediated health effects. These pathways include
the hypohalous acid generating peroxidase enzymes lactoperoxidase (LPO)
and myeloperoxidase (MPO), nicotinamide adenine dinucleotide phosphate
(NADPH)-dependent oxidase 5 (NOX5), and PTGS2 (COX-2) mediated
arachidonic acid metabolism. Many pathways of DBP toxic action involve the
generation of ROS; the up-regulation of the Antioxidant Response Element
(ARE) is an exceedingly sensitive and useful bioanalytical marker (44, 59, 66).
Identification of Susceptible Subpopulations Based on Human Biomarkers

Associated with the Molecular Mechanisms of DBP Toxicity
From the gene array studies and from gene reporter studies suitable
human molecular biomarkers are emerging that may be useful in identifying
subpopulations that may have enhanced sensitivity to DBPs. These biomarker
indicators include GAPDH polymorphisms (67), BRCA1 and BRCA2
polymorphisms (64), p53 polymorphisms (64), elevated COX2 expression (59)
and cyclooxygenase-2 mediated production of PGE2 (prostaglandin E2) and COX2
polymorphisms (68), enhanced ARE expression (69), glutathione-S-transferase
polymorphisms (70, 71), and DNA repair gene polymorphisms (72). These traits
14
within a subpopulation may make individuals more susceptible to the toxic action
of DBPs and lead to adverse health outcomes. In concert with characterizing the
number and concentrations of toxic DBPs as well as the toxicity of the drinking
water, individuals expressing one or more of these polymorphisms may have
heightened sensitivity and greater risk of adverse health effects. The use of these
and other polymorphisms that interface the molecular pathways of toxicity of
DBPs may provide greater resolution to epidemiological studies on DBPs and
health issues.
Definition of DBP Exposure in Selected Waters with High Toxicity
In most epidemiological studies, the DBP metric is the concentration of

trihalomethanes (THMs) in drinking water (73). Few studies have compared

the number or concentrations of other DBPs or DBP classes with the overall
toxicity of water samples. In the Health Impacts of Long-Term Exposure to
Disinfection By-Products in Drinking Water (HIWATE) program, 95 DBPs were
identified among 11 water samples from several cities in Europe. We found a
correlation between the number of DBPs in the water samples and the induction of
cytotoxicity in mammalian cells (74). Both cytotoxicity (Figure 5A) and genomic
DNA damage (Figure 5B) demonstrated a positive association with the relative
DBP concentration for each HIWATE water sample. The concentration of THMs
alone may not be the best or most appropriate metric to quantify DBP exposure
for drinking waters and other more comprehensive metrics are needed (73).
Figure 5. Mammalian cell cytotoxicity (A) and genotoxicity (B) responses as a

function of the relative concentrations of DBPs in 11 drinking water samples.
Quantifying human exposure to DBPs is essential for precise risk assessment

in epidemiological studies. Future studies would be aided by the development
of metrics of exposure that would link the impact of DBPs in individuals with a
15
toxic response. An example of this was a study that reported genotoxic effects of
swimmers in chlorinated pools (75). Enhanced bladder cancer rates were observed
in swimmers exposed to disinfected pool water (76). Other biomarkers for DBP
exposure include urinary trichloroacetic acid measurements in individuals exposed
to DBPs (77–79). In a commentary, Kogevinas stated, “Future studies should
be large, combine ecological and individual information on exposure, evaluate
multiple chemicals and not only THMs, and evaluate multiple routes. In a situation
of relatively low risks, even relatively small improvements in exposure assessment
may make significant differences in the studies. The use of new evidence on
mechanisms of these compounds including use of genotyping to evaluate effect
modification should be promoted (73).”
Integration of Epidemiological Studies with DBP Exposure, Toxicity,

Biomarkers, and Adverse Health Outcomes
Epidemiology is essential to our understanding of the role of DBP exposure

in human disease. These environmental micropollutants pose a vexing problem.
DBPs express a wide diversity of toxic response; however, this toxicity is spread
across many hundreds of individual agents within a complex mixture. Most
of the DBPs that comprise TOX are unknown. Of those that are known, only
a relatively small group has been systematically analyzed for their capacity
to induce adverse biological effects. Epidemiological advances leading to an
understanding of health risks posed by DBPs will only occur with concomitant
advances in exposure science, biomarkers of internal dose, biomarkers of affected
pathways and biomarkers of population variability in susceptibility (80).
Creative approaches to the epidemiology of DBPs demand the development
of studies that identify the forcing agents leading to the toxicity of drinking water
as well as the identification of population segments of increased risk. These
susceptible populations may be identified as susceptible because of demographics,
genetics and lifestyles (1). In general, using a focused epidemiological approach
will improve the quality of evidence. Achieving these improvements will require
systematic, long-term interdisciplinary studies (Figure 6).
Conclusion
DBP exposure is a constant in modern life. Every time you take a drink of
water, every time you prepare food, take a shower or swim in a swimming pool
you are exposed to DBPs. The level of concern of the adverse health effects of
these toxic agents is increasing with the increased level of compromised source
waters used in the generation of drinking water. With the advent of climate
change-mediated drought, increased reliance upon direct or indirect wastewater
recycling or disinfected desalinated water, new emerging DBP classes are entering
the drinking water stream.
16
Figure 6. A new pathway for integrated DBP research.
Biologists, chemists and engineers must form teams to address problems

posed by hazardous DBPs and other micropollutants in water. Systematic,
comparative in vitro toxicology must be integrated as a feed-back information
loop into innovative engineering processes to remove and degrade micropollutants
and disinfect water. The biological mechanisms of DBP toxicity must be included
in epidemiological studies. We must develop systems to prevent unintended toxic
consequences as we move forward in the implementation of new methods to
desalinate, decontaminate, reuse and disinfect water.
References
1. Calderon, R. L. The epidemiology of chemical contaminants of drinking
water. Food Chem.Toxicol. 2000, 38, S13–20.
2. Richardson, S. D.; Postigo, C. Drinking water disinfection by-products. In
Emerging Organic Contaminants and Human Health; Springer: Berlin, 2012;
pp 93−137.
3. Singer, P. C. Control of disinfection by-products in drinking water. J. Environ.
Eng. 1994, 120, 727–744.
4. Hong, H.; Xiong, Y.; Ruan, M.; Liao, F.; Lin, H.; Liang, Y. Factors affecting
THMs, HAAs and HNMs formation of Jin Lan Reservoir water exposed to
chlorine and monochloramine. Sci. Total Environ. 2013, 444, 196–204.
5. Zhang, X.; Echigo, S.; Minear, R. A.; Plewa, M. J., Characterization and
comparison of disinfection by-products of four major disinfectants. In
Natural Organic Matter and Disinfection By-Products: Characterization
and Control in Drinking Water; Barrett, S. E., Krasner, S. W., Amy, G. L.,
Eds.; American Chemical Society: Washington, DC, 2000; pp 299−314.
17
6. Hua, G. H.; Reckhow, D. A. Comparison of disinfection byproduct formation
from chlorine and alternative disinfectants. Water Res. 2007, 41, 1667–1678.
7. Rook, J. J. Formation of haloforms during chlorination of natural waters.
Water Treat. Exam. 1974, 23, 234–243.
8. Bellar, T. A.; Lichtenbert, J. J.; Kroner, R. C. The occurrence of organohalides
in chlorinated drinking waters. J. Am. Water Works Assoc. 1974, 66,
703–706.
9. Richardson, S. D., Disinfection by-products: formation and occurrence in
drinking water. In Encyclopedia of Environmental Health; Nriagu, J. O., Ed.;
Elsevier: Burlington, 2011; Vol. 1, pp 110−136.
10. Richardson, S. D.; Plewa, M. J.; Wagner, E. D.; Schoeny, R.; DeMarini, D.
M. Occurrence, genotoxicity, and carcinogenicity of regulated and emerging
disinfection by-products in drinking water: A review and roadmap for
research. Mutat. Res. 2007, 636, 178–242.

11. Plewa, M. J.; Wagner, E. D. Drinking water disinfection by-products:
comparative mammalian cell cytotoxicity and genotoxicity. In Encyclopedia
of Environmental Health; Nriagu, J. O., Ed.; Elsevier: Burlington, 2011; Vol.
1, pp 806−812.
12. Plewa, M. J.; Wagner, E. D. Mammalian Cell Cytotoxicity and Genotoxicity
of Disinfection By-Products; Water Research Foundation: Denver, CO, 2009;
p 134.
13. U. S. Environmental Protection Agency, National primary drinking water
regulations: Stage 2 disinfectants and disinfection byproducts rule. Fed.
Regist. 2006, 71, 387–493.
14. Costet, N.; Villanueva, C. M.; Jaakkola, J. J.; Kogevinas, M.; Cantor, K.
P.; King, W. D.; Lynch, C. F.; Nieuwenhuijsen, M. J.; Cordier, S. Water
disinfection by-products and bladder cancer: is there a European specificity?
A pooled and meta-analysis of European case-control studies. Occup.
Environ. Med. 2011, 68, 379–385.
15. Villanueva, C. M.; Cantor, K. P.; Cordier, S.; Jaakkola, J. J.; King, W. D.;
Lynch, C. F.; Porru, S.; Kogevinas, M. Disinfection byproducts and bladder
cancer: a pooled analysis. Epidemiology 2004, 15, 357–367.
16. Bove, G. E.; Rogerson, P. A.; Vena, J. E. Case-control study of the effects
of trihalomethanes on urinary bladder cancer risk. Arch. Environ. Occup.
Health 2007, 62, 39–47.
17. King, W. D.; Marrett, L. D.; Woolcott, C. G. Case-control study of colon
and rectal cancers and chlorination by-products in treated water. Cancer
Epidemiol., Biomarkers Prev. 2000, 9, 813–818.
18. Rahman, M. B.; Driscoll, T.; Cowie, C.; Armstrong, B. K. Disinfection by-
products in drinking water and colorectal cancer: a meta-analysis. Int. J.
Epidemiol. 2010, 39, 733–745.
19. Bove, G. E., Jr.; Rogerson, P. A.; Vena, J. E. Case control study of the
geographic variability of exposure to disinfectant byproducts and risk for
rectal cancer. Int. J. Health Geogr. 2007, 6, 18.
20. Bove, F.; Shim, Y.; Zeitz, P. Drinking water contaminants and adverse
pregnancy outcomes: a review. Environ. Health Perspect. 2002, 110 (Suppl
1), 61–74.
18
21. Grellier, J.; Bennett, J.; Patelarou, E.; Smith, R. B.; Toledano, M. B.;
Rushton, L.; Briggs, D. J.; Nieuwenhuijsen, M. J. Exposure to disinfection
by-products, fetal growth, and prematurity: a systematic review and
meta-analysis. Epidemiology 2010, 21, 300–313.
22. Colman, J.; Rice, G. E.; Wright, J. M.; Hunter, E. S., III; Teuschler, L. K.;
Lipscomb, J. C.; Hertzberg, R. C.; Simmons, J. E.; Fransen, M.; Osier, M.;
Narotsky, M. G. Identification of developmentally toxic drinking water
disinfection byproducts and evaluation of data relevant to mode of action.
Toxicol. Appl. Pharmacol. 2011, 254, 100–126.
23. Rivera-Nunez, Z.; Wright, J. M. Association of brominated trihalomethane
and haloacetic acid exposure with fetal growth and preterm delivery in
Massachusetts. J. Occup. Environ. Med. 2013, 55, 1125–1134.
24. Plewa, M. J.; Wagner, E. D.; Jazwierska, P.; Richardson, S. D.; Chen, P. H.;
McKague, A. B. Halonitromethane drinking water disinfection byproducts:

chemical characterization and mammalian cell cytotoxicity and genotoxicity.
Environ. Sci. Technol. 2004, 38, 62–68.
25. Plewa, M. J.; Wagner, E. D.; Richardson, S. D.; Thruston, A. D., Jr.; Woo, Y.
T.; McKague, A. B. Chemical and biological characterization of newly
discovered iodoacid drinking water disinfection byproducts. Environ. Sci.
Technol. 2004, 38, 4713–4722.
26. Muellner, M. G.; Wagner, E. D.; McCalla, K.; Richardson, S. D.; Woo, Y. T.;
Plewa, M. J. Haloacetonitriles vs. regulated haloacetic acids: Are nitrogen
containing DBPs more toxic? Environ. Sci. Technol. 2007, 41, 645–651.
27. Plewa, M. J.; Muellner, M. G.; Richardson, S. D.; Fasano, F.; Buettner, K.
M.; Woo, Y. T.; McKague, A. B.; Wagner, E. D. Occurrence, synthesis
and mammalian cell cytotoxicity and genotoxicity of haloacetamides: an
emerging class of nitrogenous drinking water disinfection by-products.
28. Pals, J.; Clay, J.; Komaki, Y.; Marinas, B. J.; Wagner, E. D.; Plewa, M.
J. Cyanogen halide drinking water disinfection byproducts. Environ. Mol.
Mutagen. 2009, 50, 578.
29. Plewa, M. J.; Simmons, J. E.; Richardson, S. D.; Wagner, E. D. Mammalian
cell cytotoxicity and genotoxicity of the haloacetic acids, a major class of
drinking water disinfection by-products. Environ. Mol. Mutagen. 2010, 51,
871–878.
30. Wagner, E. D.; Hsu, K. M.; Lagunas, A.; Mitch, W. A.; Plewa, M. J.
Comparative genotoxicity of nitrosamine drinking water disinfection
byproducts in Salmonella and mammalian cells. Mutat. Res. 2012, 741,
109–115.
31. Kimura, S. Y.; Komaki, Y.; Plewa, M. J.; Mariñas, B. J. Chloroacetonitrile and
N,2-dichloroacetamide formation from the reaction of chloroacetaldehyde
and monochloramine in water. Environ. Sci. Technol. 2013, 47,
12382–12390.
32. Escobar-Hoyos, L. F.; Hoyos-Giraldo, L. S.; Londono-Velasco, E.;
Reyes-Carvajal, I.; Saavedra-Trujillo, D.; Carvajal-Varona, S.; Sanchez-
Gomez, A.; Wagner, E. D.; Plewa, M. J. Genotoxic and clastogenic effects
19
of monohaloacetic acid drinking water disinfection by-products in primary
human lymphocytes. Water Res. 2013, 47, 3282–3290.
33. Wagner, E. D.; Osiol, J.; Mitch, W. A.; Plewa, M. J. Comparative in vitro
toxicity of nitrosamines and nitramines associated with amine-based carbon
capture and storage. Environ. Sci. Technol. 2014, 48, 8203–8211.
34. Ali, A.; Kurzawa-Zegota, M.; Najafzadeh, M.; Gopalan, R. C.; Plewa, M.
J.; Anderson, D. Effect of drinking water disinfection by-products in human
peripheral blood lymphocytes and sperm. Mutat. Res. 2014, 770, 136–143.
35. Komaki, Y.; Marinas, B. J.; Plewa, M. J. Toxicity of drinking
water disinfection by-products: cell cycle alterations induced by
monohaloacetonitriles. Environ. Sci. Technol. 2014, 48, 11662–11669.
36. Weinberg, H. S.; Krasner, S. W.; Richardson, S. D.; Thruston, A. D., Jr.
The Occurrence of Disinfection By-Products (DBPs) of Health Concern
in Drinking Water: Results of a Nationwide DBP Occurrence Study;

EPA/600/R02/068; U.S. Environmental Protection Agency National
Exposure Research Laboratory: Athens, GA, 2002; pp 1−460.
37. Krasner, S. W.; Weinberg, H. S.; Richardson, S. D.; Pastor, S. J.; Chinn, R.;
Sclimenti, M. J.; Onstad, G. D.; Thruston, A. D., Jr. The occurrence of a
new generation of disinfection by-products. Environ. Sci. Technol. 2006, 40,
7175–7185.
38. Kargalioglu, Y.; McMillan, B. J.; Minear, R. A.; Plewa, M. J., A new
assessment of the cytotoxicity and genotoxicity of drinking water disinfection
by-products. In Natural Organic Matter and Disinfection By-Products:
Characterization and Control in Drinking Water; Barrett, S. E., Krasner, S.
W., Amy, G. L., Eds.; American Chemical Society: Washington, D.C., 2000;
pp 16−27.
39. Plewa, M. J.; Kargalioglu, Y.; Vankerk, D.; Minear, R. A.; Wagner, E. D.
Mammalian cell cytotoxicity and genotoxicity analysis of drinking water
disinfection by-products. Environ. Mol. Mutagen. 2002, 40, 134–142.
40. Richardson, S. D.; Thruston, A. D., Jr.; Rav-Acha, C.; Groisman, L.;
Popilevsky, I.; Juraev, O.; Glezer, V.; McKague, A. B.; Plewa, M. J.;
Wagner, E. D. Tribromopyrrole, brominated acids, and other disinfection
byproducts produced by disinfection of drinking water rich in bromide.
41. Plewa, M. J.; Wagner, E. D.; Muellner, M. G.; Hsu, K. M.; Richardson,
S. D., Comparative mammalian cell toxicity of N-DBPs and C-DBPs.
In Occurrence, Formation, Health Effects and Control of Disinfection
By-Products in Drinking Water; Karanfil, T., Krasner, S. W., Westerhoff, P.,
Xie, Y., Eds.l American Chemical Society: Washington, D.C., 2008; Vol.
995, pp 36−50.
42. Richardson, S. D.; Fasano, F.; Ellington, J. J.; Crumley, F. G.; Buettner, K.
M.; Evans, J. J.; Blount, B. C.; Silva, L. K.; Waite, T. J.; Luther, G. W.;
McKague, A. B.; Miltner, R. J.; Wagner, E. D.; Plewa, M. J. Occurrence
and mammalian cell toxicity of iodinated disinfection byproducts in drinking
water. Environ. Sci. Technol. 2008, 42, 8330–8338.
20
43. Kimura, S. Y. Formation of (Halo)Acetamides and (Halo)Acetonitriles
from the Reaction of Monochloramine and (Halo)Acetaldehydes in Water.
University of Illinois at Urbana-Champaign, Urbana, IL, 2013.
44. Escher, B. I.; van Daele, C.; Dutt, M.; Tang, J. Y.; Altenburger, R. Most
oxidative stress response in water samples comes from unknown chemicals:
the need for effect-based water quality trigger values. Environ. Sci. Technol.
2013, 47, 7002–7011.
45. Yang, Y.; Komaki, Y.; Kimura, S.; Hu, H. Y.; Wagner, E. D.; Marinas, B. J.;
Plewa, M. J. Toxic impact of bromide and iodide on drinking water disinfected
with chlorine or chloramines. Environ. Sci. Technol. 2014, 48, 12362–12369.
46. World Health Organization. Guidelines for Drinking Water Quality, 4th ed.;
United Nations: Geneva, Switzerland, 2011; Vol. 1, p 541.
47. Duirk, S. E.; Lindell, C.; Cornelison, C.; Kormos, J. L.; Ternes, T. A.; Attene-
Ramos, M. S.; Osiol, J.; Wagner, E. D.; Plewa, M. J.; Richardson, S. D.

Formation of toxic iodinated disinfection by-products from compounds used
in medical imaging. Environ. Sci. Technol. 2011, 45, 6845–6854.
48. Postigo, C.; Richardson, S. D. Transformation of pharmaceuticals during
oxidation/disinfection processes in drinking water treatment. J. Hazard.
Mater. 2014, 279, 461–475.
49. Shen, R.; Andrews, S. A. Demonstration of 20 pharmaceuticals and
personal care products (PPCPs) as nitrosamine precursors during chloramine
disinfection. Water Res. 2011, 45, 944–952.
50. Schmidt, C. K.; Brauch, H. J. N,N-dimethylsulfamide as precursor for N-
nitrosodimethylamine (NDMA) formation upon ozonation and its fate during
drinking water treatment. Environ. Sci. Technol. 2008, 42, 6340–6346.
51. Wilczak, A.; Assadi-Rad, A.; Lai, H. H.; Hoover, L. L.; Smith, J. F.;
Berger, R.; Rodigari, F.; Beland, J. W.; Lazzelle, L. J.; Kinicannon, E. G.;
Baker, H.; Heaney, C. T. Formation of NDMA in chloraminated water
coagulated with DADMAC cationic polymer. J. Am. Water Work Assoc.
2003, 95, 94–106.
52. Padhye, L.; Luzinova, Y.; Cho, M.; Mizaikoff, B.; Kim, J. H.; Huang, C. H.
PolyDADMAC and dimethylamine as precursors of N-nitrosodimethylamine
during ozonation: reaction kinetics and mechanisms. Environ. Sci. Technol.
2011, 45, 4353–4359.
53. Kemper, J. M.; Walse, S. S.; Mitch, W. A. Quaternary amines as nitrosamine
precursors: a role for consumer products? Environ. Sci. Technol. 2010, 44,
1224–1231.
54. Park, S.-H.; Piyachaturawat, P.; Taylor, A. E.; Huang, C.-H. Potential
N-nitrosodimethylamine (NDMA) formation from amine-based water
treatment polymers in the reactions with chlorine-based oxidants and
nitrosifying agents. Water Sci. Technol.: Water Supply 2009, 9, 279–288.
55. Zhang, X.; De Silva, D.; Sun, B.; Fisher, J.; Bull, R. J.; Cotruvo, J. A.;
Cummings, B. S. Cellular and molecular mechanisms of bromate-induced
cytotoxicity in human and rat kidney cells. Toxicology 2010, 269, 13–23.
56. Du, H.; Li, J.; Moe, B.; McGuigan, C. F.; Shen, S.; Li, X. F. Cytotoxicity and
oxidative damage induced by halobenzoquinones to T24 bladder cancer cells.
21
57. Pals, J.; Ang, J.; Wagner, E. D.; Plewa, M. J. Biological mechanism for the
toxicity of haloacetic acid drinking water disinfection byproducts. Environ.
Sci. Technol. 2011, 45, 5791–5797.
58. Dad, A.; Jeong, C. H.; Pals, J.; Wagner, E. D.; Plewa, M. J. Pyruvate
remediation of cell stress and genotoxicity induced by haloacetic acid
drinking water disinfection byproducts. Environ. Mol. Mutagen. 2013, 54,
629–637.
59. Pals, J.; Attene-Ramos, M. S.; Xia, M.; Wagner, E. D.; Plewa, M. J. Human
cell toxicogenomic analysis linking reactive oxygen species to the toxicity of
monohaloacetic acid drinking water disinfection byproducts. Environ. Sci.
Technol. 2013, 47, 12514–12523.
60. Collins, F. S.; Gray, G. M.; Bucher, J. R. Toxicology. Transforming
environmental health protection. Science 2008, 319, 906–907.
61. Le Fevre, A. C.; Boitier, E.; Marchandeau, J. P.; Sarasin, A.; Thybaud, V.
Characterization of DNA reactive and non-DNA reactive anticancer drugs by
gene expression profiling. Mutat. Res. 2007, 619, 16–29.
62. Attene-Ramos, M. S.; Wagner, E. D.; Plewa, M. J. Comparative human cell
toxicogenomic analysis of monohaloacetic acid drinking water disinfection
byproducts. Environ. Sci. Technol. 2010, 44, 7206–7212.
63. Muellner, M. G.; Attene-Ramos, M. S.; Hudson, M. E.; Wagner, E. D.;
Plewa, M. J. Human cell toxicogenomic analysis of bromoacetic acid: a
regulated drinking water disinfection by-product. Environ. Mol. Mutagen.
2010, 51, 205–214.
64. Khanna, K. K.; Jackson, S. P. DNA double-strand breaks: signaling, repair
and the cancer connection. Nat. Genet. 2001, 27, 247–254.
65. Wei, X.; Wang, S.; Zheng, W.; Wang, X.; Liu, X.; Jiang, S.; Pi, J.; Zheng, Y.;
He, G.; Qu, W. Drinking water disinfection byproduct iodoacetic acid induces
tumorigenic transformation of NIH3T3 cells. Environ. Sci. Technol. 2013,
47, 5913–5920.
66. Neale, P. A.; Antony, A.; Bartkow, M. E.; Farre, M. J.; Heitz, A.; Kristiana, I.;
Tang, J. Y.; Escher, B. I. Bioanalytical assessment of the formation of
disinfection byproducts in a drinking water treatment plant. Environ. Sci.
Technol. 2012, 46, 10317–10325.
67. Butterfield, D. A.; Hardas, S. S.; Lange, M. L. B. Oxidatively modified
glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and Alzheimer’s
disease: many pathways to neurodegeneration. J. Alzheimers Dis. 2010, 20,
369–393.
68. Pereira, C.; Pimentel-Nunes, P.; Brandao, C.; Moreira-Dias, L.; Medeiros, R.;
Dinis-Ribeiro, M. COX-2 polymorphisms and colorectal cancer risk: a
strategy for chemoprevention. Eur. J. Gastroenterol. Hepatol. 2010, 22,
607–613.
69. Wang, X.; Tomso, D. J.; Chorley, B. N.; Cho, H. Y.; Cheung, V. G.;
Kleeberger, S. R.; Bell, D. A. Identification of polymorphic antioxidant
response elements in the human genome. Hum. Mol. Genet. 2007, 16,
1188–1200.
70. Hayes, J. D.; Strange, R. C. Glutathione S-transferase polymorphisms and
their biological consequences. Pharmacology 2000, 61, 154–166.
22
71. Cantor, K. P.; Villanueva, C. M.; Silverman, D. T.; Figueroa, J. D.; Real, F.
X.; Garcia-Closas, M.; Malats, N.; Chanock, S.; Yeager, M.; Tardon, A.;
Garcia-Closas, R.; Serra, C.; Carrato, A.; Castano-Vinyals, G.; Samanic, C.;
Rothman, N.; Kogevinas, M. Polymorphisms in GSTT1, GSTZ1, and
CYP2E1, disinfection by-products, and risk of bladder cancer in Spain.
Environ. Health Perspect. 2010, 118, 1545–1550.
72. Andrew, A. S.; Karagas, M. R.; Nelson, H. H.; Guarrera, S.; Polidoro, S.;
Gamberini, S.; Sacerdote, C.; Moore, J. H.; Kelsey, K. T.; Demidenko, E.;
Vineis, P.; Matullo, G. DNA repair polymorphisms modify bladder cancer
risk: a multi-factor analytic strategy. Hum. Hered. 2008, 65, 105–118.
73. Kogevinas, M. Epidemiological approaches in the investigation of
environmental causes of cancer: the case of dioxins and water disinfection
by-products. Environ. Health 2011, 10 (Suppl 1), S3.
74. Jeong, C. H.; Wagner, E. D.; Siebert, V. R.; Anduri, S.; Richardson, S. D.;
Daiber, E. J.; McKague, A. B.; Kogevinas, M.; Villanueva, C. M.; Goslan, E.
H.; Luo, W.; Isabelle, L. M.; Pankow, J. F.; Grazuleviciene, R.; Cordier, S.;
Edwards, S. C.; Righi, E.; Nieuwenhuijsen, M. J.; Plewa, M. J. The occurrence
and toxicity of disinfection byproducts in European drinking waters in relation
with the HIWATE epidemiology study. Environ. Sci. Technol. 2012, 46,
12120–12128.
75. Kogevinas, M.; Villanueva, C. M.; Font-Ribera, L.; Liviac, D.;
Bustamante, M.; Espinoza, F.; Nieuwenhuijsen, M. J.; Espinosa, A.;
Fernandez, P.; DeMarini, D. M.; Grimalt, J. O.; Grummt, T.; Marcos, R.
Genotoxic effects in swimmers exposed to disinfection by-products in indoor
swimming pools. Environ. Health Perspect. 2010, 118, 1531–1537.
76. Villanueva, C. M.; Cantor, K. P.; Grimalt, J. O.; Malats, N.; Silverman, D.;
Tardon, A.; Garcia-Closas, R.; Serra, C.; Carrato, A.; Castano-Vinyals, G.;
Marcos, R.; Rothman, N.; Real, F. X.; Dosemeci, M.; Kogevinas, M. Bladder
cancer and exposure to water disinfection by-products through ingestion,
bathing, showering, and swimming in pools. Am. J. Epidemiol. 2007, 165,
148–156.
77. Nieuwenhuijsen, M. J.; Toledano, M. B.; Elliott, P. Uptake of chlorination
disinfection by-products; a review and a discussion of its implications for
exposure assessment in epidemiological studies. J. Exposure Anal. Environ.
Epidemiol. 2000, 10, 586–599.
78. Froese, K. L.; Sinclair, M. I.; Hrudey, S. E. Trichloroacetic acid as a biomarker
of exposure to disinfection by-products in drinking water: a human exposure
trial in Adelaide, Australia. Environ. Health Perspect. 2002, 110, 679–687.
79. Savitz, D. A. Invited commentary: biomarkers of exposure to drinking water
disinfection by-products--are we ready yet? Am. J. Epidemiol. 2012, 175,
276–278.
80. Nachman, K. E.; Fox, M. A.; Sheehnan, M. C.; Burke, T. A.; Rodricks, J. V.;
Woodruff, T. J. Leveraging epidemiology to improve risk assessment. Open
Epidemiol. J. 2011, 4, 3–29.
23
Chapter 2
Analysis, Occurrence, and Toxicity of

Haloacetaldehydes in Drinking Waters:
Iodoacetaldehyde as an Emerging
Disinfection By-Product
Cristina Postigo,*,1 Clara H. Jeong,2 Susan D. Richardson,3

Elizabeth D. Wagner,2 Michael J. Plewa,2 Jane Ellen Simmons,4
and Damià Barceló1,5
1Department of Environmental Chemistry, Institute for Environmental
Assessment and WaterResearch, (IDAEA-CSIC), Carrer Jordi Girona
18-26, 08034, Barcelona, Spain
2Department of Crop Sciences and the Center of Advanced Materials for the
Purification of Water with Systems, Safe Global Water Institute, University

of Illinois at Urbana-Champaign, Urbana, Illinois 61801, United States
3Department of Chemistry and Biochemistry, University of South Carolina,
JM Palms Centre for GSR, 631 Sumter Street,

Columbia, South Carolina 29208, United States
4National Health and Environmental Effects Research Laboratory,
(NHEERL-U.S. EPA), 109 T.W. Alexander Drive,

Research Triangle Park, North Carolina 27709, United States
5Catalan Institute for Water Research (ICRA), Parc Científic i Tecnològic de
la Universitat de Girona, Edifici H2O, Carrer d’Emili Grahit, 101,

17003 Girona, Spain
Chlorinated and brominated haloacetaldehydes (HALs) are

considered the 3rd largest class of disinfection by-products
(DBPs) by weight. The iodinated HAL, iodoacetaldehyde,
has been recently reported as an emerging DBP in finished
drinking waters. Overall, iodinated DBPs, e.g., iodoacetic
acids, iodoacetamides, and iodonitriles, are among the most
genotoxic of all DBPs identified. In this context, this chapter
reviews the analytical methods available to date to determine
HALs in water, and the concentrations at which they are present

in finished drinking waters. Since systematic toxicological
effects have been only investigated for selected chloro- and
bromo- HALs, a comparative study of the genotoxicity and
cytotoxicity of this DBP class to mammalian cells is also
presented.
Keywords: disinfection by-products; genotoxicity;

cytotoxicity; occurrence; drinking waters; halogenated
aldehydes; chloral hydrate; mammalian cells
Introduction
The nature and quantity of the disinfection by-products (DBPs) formed

during water disinfection treatments is determined by the type of disinfectant
agent used, the dose applied and the constituents present in the water among
other factors (1–4). DBP formation, occurrence and health effects have been
matters of scientific concern over the last three decades. More than 600 DBPs
are known to date; however, less than 100 have been studied in depth, in terms
of their occurrence, formation mechanisms, and toxicity. Moreover, 50% of
the halogenated material formed during drinking water disinfection treatments
(e.g., chlorination, chloramination, and ozonation) is still unknown, and so is the
toxicological risk that it may pose to human health (5–7).
Studies on the toxicity of individual DBPs have been directed to investigate
their potential carcinogenicity, genotoxicity, mutagenicity, and cytotoxicity.
Such studies have aided health authorities of different countries in establishing
regulations and guidelines to control the levels of specific DBPs in drinking waters,
e.g., bromate, and trihalomethanes (THMs) in Europe (8) and the USA (9), and
chlorite and haloacetic acids (HAAs) in the USA (9). However, there is still a good
number of DBPs for which some toxicity data are available that are not regulated
(6). This is the case for some halogenated acetaldehydes (HALs), which were also
reported to be the third most abundant and widespread class of DBPs in drinking
water behind THMs and HAAs in a U.S. Nationwide Occurrence Study (5, 10, 11).
In this regard, chloral hydrate (CH), the hydrated form of trichloroacetaldehyde
(TCAL), has been shown to be genotoxic in numerous prokaryotic and eukaryotic
assay systems including mammalian cells, and mutagenic in vivo and in vitro
(12–15). Similar to CH, dichloroacetaldehyde (DCAL) and chloroacetaldehyde
(CAL) were reported to induce mitotic aneuploidy (16, 17). Regarding the
bromine containing HALs, tribromoacetaldehyde (TBAL) was shown to be more
effective than CH in inducing DNA breaks (18), and bromoacetaldehyde (BAL)
was reported to irreversibly bind to DNA and proteins in rat liver microsomes
(19).
Whereas the presence of CH in drinking waters as a DBP has been
investigated since the late 1980s, most of the HALs were not quantitatively
determined until 2006 for the first time, as standard solutions for these compounds
were not commercially available (20). Moreover, data on the occurrence of
iodo-containing HALs, e.g., iodoacetaldehyde (IAL), has been missing until now.
26
Note that iodinated DBPs like iodoacetic acids, iodoacetamides, and iodonitriles,
have been reported to be more toxic than their respective chlorinated and
brominated analogues, and are among the most genotoxic of all DBPs identified
(10, 21, 22).
In this context, the present manuscript reviews the actual analytical methods
available to evaluate the occurrence of chloro-, bromo- and iodo-HALs in water,
and the levels of this DBP class in source and treated waters. Since systematic
toxicological effects have been only investigated for selected chloro- and bromo-
HALs, a comparative study of the chronic genotoxicity and acute cytotoxicity of
the complete set of HALs in mammalian cells is also presented.
Haloacetaldehyde Determination in Water

The chemical structures and main physical-chemical properties of the HALs

investigated in water so far are shown in Figure 1. TCAL is converted to CH in
water, thus, the latter is actually the one monitored in water samples. As it can be
observed, HALs present very different physical chemical properties, and therefore,
they cannot be analyzed at once using only one analytical procedure.
Figure 1. Physical-chemical properties of HALs. (*) Experimental value. Source

data: (1) Physprop database using CAS RN (http://goo.gl/ukAJno). (2) EPI Suite
software using SMILES (http://goo.gl/sSPHpr).
27
Table 1 summarizes the different approaches used to determine HAL
concentrations in water. All of them, as discussed in detail in the next subsections,
follow different analyte extraction protocols. However, in all cases HAL detection
is based on gas chromatography (GC) coupled to either electron capture (ECD)
or mass spectrometry (MS) detection. Special consideration in the present review
has been given to those methodologies covering a large number of HALs (20,
23–27).
Sample Collection and Preservation
Samples were usually collected in pre-cleaned amber glass bottles with

polytetrafluoroethylene (PTFE) screw caps (26, 28) or in PTFE bottles (24),

filled with no headspace to avoid evaporation of the most volatile compounds.
The analysis of these compounds is recommended upon sample reception in the
laboratory because HALs, mainly the brominated species, are very unstable in
disinfected waters, and may degrade to their corresponding trihalomethanes (20,
25, 26), even in the dark and under cool conditions (4ºC). Due to sample transport
logistics, the time between sample collection and analysis may exceed 24 h.
Therefore, sample preservation requires the addition of a dechlorinating agent
and acidification of the sample to a pH below 4.5 to prevent analyte hydrolysis
(20). In this regard, Serrano et al. (26) reported that the acidification of the
sample with sulfuric acid to a pH of 3-3.5 was sufficient to maintain chloro- and
bromo-HAL concentrations constant for 14 days at 4ºC (BAL and IAL were
not included in this study). On the other hand, whereas ammonium chloride
and sodium sulfite react at pH 7 and 8 with bromochloroacetaldehyde (BCAL),
bromodichloroacetaldehyde (BDCAL), dibromochloroacetaldehyde (DBCAL),
CH, and TBAL present in water; sodium arsenite, sodium borohydride, and
ascorbic acid did not show any negative effect on the stability of these HALs
(25). However, only ascorbic acid has been recommended as a dechlorinating
agent, as its decomposition effect has been investigated for the whole spectra of
HALs, and when not in a large excess, it seems to preserve HAL concentrations
in water, even that of IAL, at acidic pH (24).
Sample Extraction
The determination of CH concentrations in water, which is indeed the most

investigated HAL so far, was achieved in most peer-reviewed research with diverse
liquid-liquid extraction (LLE) approaches derived from EPA Method 551.1 for
the determination of chlorination DBPs, chlorinated solvents, and halogenated
pesticides/herbicides in drinking water (29). In all reviewed methods, CH was
extracted from the water sample (10-50 mL), occasionally acidified, with a small
volume of methyl tert-butyl ether (MTBE) (2-3 mL).
28
Table 1. Analytical Methodologies Used for HAL Analysis in Water
Analyte extraction Analyte detection LOD
HALs Sample preservation Reference
(Sample volume) (GC column) (μg/L)
GC-ECD
LLE with MTBE or pentane (50 DB-1 (30 m, 0.25 mm i.d, 1
CH Sodium sulfite (100 mg/L) 0.011** (29)
mL) μm) and Rtx-1301 (30 m, 0.25
mm i.d, 1 μm)
GC-ECD / GC-MS
DCAL, BCAL, DBAL,
AscAc (35 mg/L) SPE (Bond Elut PPL) (36 mL) DB-5 and DB-1 (30 m, 0.32 1-2.5 (for
BDCAL, DBCAL, CH, (23)
pH=3-4 LLE with MTBE (30 mL) * mm i.d, 1 μm) / Rtx 1 (30 m, SPE)**
TBAL
0.25 mm i.d, 1 μm)
DCAL, BCAL, DBAL, LVI-PTV-GC-MS
0.006 -
BDCAL, DBCAL, CH, pH = 3-3.5 MLLE with ethyl acetate (9 mL) HP-5 MS (30 m, 0.25 mm i.d., (26)
0.020
29
TBAL 0.25 μm)

DCAL, BCAL, DBAL, GC-ECD / GC-MS
AscAc (0.1 mL, 0.114 M) LLE with MTBE
BDCAL, DBCAL, CH, DB-5 and DB-1 (30 m, 0.32 0.03-0.05 (20)
pH=4.5 (50 mL) *
TBAL mm i.d, 1 μm)
GC-MS
CAL, BAL, DCAL, AscAc (12.5 mg/L) PFBHA derivatization + LLE
Zebron ZB-5 (30 m, 0.25 mm 0.05-0.25 (24)
BCAL, DBAL pH= 3.5 with Hexane (100 mL)
i.d., 0.25 μm)
GC-MS
BDCAL, DBCAL, CH, AscAc (12.5 mg/L) SPE (Oasis HLB)
Zebron ZB-5 (30 m, 0.25 mm 0.1-0.5 (24)
TBAL pH= 3.5 (100 mL)
i.d., 0.25 μm)
* LLE approach based on EPA 551.1.; ** RL: method reporting limit AscAc: ascorbic acid, ECD: electron capture detection, GC: gas chromatography, LLE:
liquid-liquid extraction, LOD: method limit of detection, LVI-PTV-GC-MS: large volume injection - programmable temperature vaporizer, MLLE: micro
liquid-liquid extraction, MS: mass spectrometry, MTBE: methyl tert-butyl ether, SPE: solid phase extraction.

Several grams of anhydrous Na2SO4 (2-14 g) or NaCl (4-15 g) were usually
added to increase the ionic strength of the solution, and thus, to enhance the
extraction of this polar compound (23, 27, 30–42). Liquid-liquid extraction with
MTBE also proved to be suitable for the extraction of the remaining chloro and
bromo containing tri-HALs, and the di-HALs, DCAL and BCAL (20, 25). A
miniaturized approach of EPA Method 551.1 using 0.2 mL of ethyl acetate as the
extracting solvent was recently developed by Serrano et al. (26) for the extraction
of chloro and bromo containing di-HALs and tri-HALs from water. This fast
and “green” method, e.g. minimal consumption of solvents and salts, provided
slightly lower limits of detection (0.006-0.02 µg/L) than EPA Method 551.1
(0.02-0.03 µg/L) (26).
An approach based on EPA Method 556 (43) was applied for the extraction of
very polar, low molecular weight, and high volatile aldehydes (and ketones), e.g.,
mono- and di-HALs (11, 24), which in their native form, are almost impossible
to extract from water. This method consisted of the derivatization of these
compounds with O-(2,3,4,5,6-pentafluorobenzyl)hydroxylamine (PFBHA) and
subsequent LLE of the oximes formed with hexane. The derivatization reaction is
shown in Scheme 1. This is also the analytical strategy selected for the extraction
of the very polar IAL from water. Limits of detection achieved for mono- and
di-HALs with this approach ranged between 0.05 µg/L and 0.25 µg/L (24).
Scheme 1. Derivatization reaction of PFBHA with mono- and di-HALs

(mono-HALs: X= halogen and Y=H; di-HALs: X,Y = halogen)
Solid phase extraction (SPE) onto a modified styrene divinylbenzene

polymer, i.e., Bond Elut PPL (Agilent) or a hydrophilic-lipophilic balanced
polymer, i.e., Oasis HLB (Waters), was also applied to preconcentrate trihalo and
dihalo species present in water (23, 24). Eluting solvents used were a mixture
of hexane:dichloromethane (23) or MTBE (24). Contrary to the satisfactory
recoveries obtained for TBAL, DCBAL, and BDCAL with both SPE sorbents,
low recoveries were reported for di-HALs and the most polar tri-HAL, CH.
Compared to the aforementioned extraction approaches, slightly higher limits of
detection, always in the low µg/L range, were obtained for the target HALs with
SPE methods, i.e.: 1-2.5 µg/L (23) or 0.1-0.5 µg/L (24).
30
Analyte Detection
Regardless of the extraction method followed, HAL detection was always
achieved by means of GC coupled to either ECD or MS. In most of the ECD
based methods, due to the low specificity of this technique, a secondary column in
addition to the primary analytical column was used for analyte confirmation (20,
23, 29, 33, 35, 44). GC/MS analysis was also used by Koudjonou and Lebel (20)
for analyte confirmation. Electron ionization (EI) was used to acquire the analyte
spectra, and chemical ionization (CI) was used to confirm the molecular weights
of the compounds (20).
Although less sensitive than GC-ECD detection, selective GC/EI-MS
detection was also commonly used in the selected ion monitoring (SIM) mode
for HAL identification and quantification (23–27, 36, 37, 45). Serrano et al (26)
showed that programmable temperature vaporizer (PTV)-based large volume

injection (LVI) that allows injecting sample volumes between 20 and 50 µL is an
effective tool for enhancing limits of detection of HAL analysis with GC/MS by
one to two orders of magnitude over conventional splitless injection techniques
(sample volume injected below 2 µL).
Quantification of individual HALs was usually performed with internal
standard normalized response factors obtained from the analyses of multi-level
fortified HAL-free waters (groundwater or Milli-Q purified water) extracted and
analyzed under identical conditions as the real water samples (20, 23, 24, 26).
Haloacetaldehyde Occurrence in Water

The occurrence of HALs in treated waters (drinking water plant effluent,
disinfected water at different locations of distribution systems, and swimming
pool water) as reported in the peer-reviewed literature has been summarized in
Table 2.
As mentioned previously and as it can be observed in Table 2, CH (the
hydrated form of trichloroacetaldehyde, TCAL) is the most investigated
halogenated aldehyde in water. Not only has its occurrence in drinking waters
been a matter of study, but also its removal through different technologies
(27, 33, 39, 46) and its formation mechanisms (30, 34, 47–51). The highest
CH concentration was determined in swimming pool water (340 µg/L) (26),
whereas maximum levels reported in drinking water surpassed in most cases 10
µg/L, but did not exceed 40 µg/L. The World Health Organization (WHO) has
not considered it necessary to derive a drinking water guideline value for this
compound, as it occurs at concentrations well below the calculated health-based
tolerable daily intake (TDI) value of 100 µg/L (52). The remaining halogenated
aldehydes were usually found in drinking water at maximum levels below 10
µg/L.
The whole spectrum of HALs was recently investigated in seven drinking
water treatment plants (DWTPs) in the USA (24). To date, this is the only study
including BAL and IAL as target analytes (see Table 2). Figure 2 summarizes the
frequency of detection and the concentration range measured for individual HALs
in this study. All investigated HALs (10 total) were detected in all treated waters,
31
except IAL and TBAL, which were present in only 57% and 43% of the samples,
respectively. IAL formation was exclusively observed during chloramination
processes, and its contribution to total HAL concentrations ranged between 6 to
28% (24). Contrary to IAL, the other investigated mono-HALs presented the
lowest overall concentrations of all target analytes.
Table 2. Occurrence of HALs in Disinfected Water

HALs Concentration (μg/L)** Country Reference
CAL 0.2-0.3 ef U.S. (24)
n.d. -2.4 ef / n.d.-2.1 ds U.S. (11)
BAL <0.5-1.3 ef U.S. (24)

IAL n.d.-4.6 ef U.S. (24)
DCAL 0.3-2.0 ef (24)
U.S.
n.d.-14 ef / n.d.-16 ds (5, 11)
1.3-3.5 ef / 1.9 ds (20)
Canada
1-3.3 ef / 1.1-3.6 ds (28)
<0.01-4 ds / 1.8-23 p Spain (26)
DBAL <0.3-3.1 ef U.S. (24)
ef
n.d.-1.7 / n.d ds (20)
Canada
n.d.-1.1 ef / n.d.-1.0 ds (28)
n.d. ds / n.d. p Spain (26)
BCAL <0.3-2.2 ef U.S. (24)
n.d.-4 ef / n.d.-7 ds U.S. (11)
0.1-2.7 ef / 0.4-0.8 ds (20)
Canada
0.6-2.7 ef / 0.4-3.1 ds (28)
CH (TCAL) <0.3-4.1 ef (24)
U.S.
n.d.-16 ef / n.d.-62 ds (11)
<0.1-15.1ef / <0.1-22.5 ds (57)
1.1-12 ef / 2.4-5.8 ds (20)
Canada
1.3-6.9 ef / 1.4-8.9 ds (28)
0.2-12.2 ef+ds (58)
<1-12.1 ds (40)
Spain
1.2-38 ds / 53-340 p (26)
0.3-11 ef /1-12.5 ds Greece (35)
n.d.-5.4 ef Japan (59)
n.d.-34.9 p (37)
Korea
n.d.-33.2 ef (60)
32
Table 2. (Continued). Occurrence of HALs in Disinfected Water
HALs Concentration (μg/L)** Country Reference
2.1*-12.2 ds (42)
China
n.d.-10.4 ef+ds (41)
0.2-19 ds Australia (61)
n.d.-8.6 ds Poland (45)
0.4-8.9 ds Poland (55)
n.d.-12.6 ef (24)
U.S.
n.d.-3 ef / n.d.-2 ds (11)
TBAL n.d.-1.5 ef / n.d ds (20)
Canada
n.d.-2 ef / n.d-1.6 ds (28)

BDCAL 1.1-2.2 ef U.S. (24)
ef
0.3-7.2 / 1.5-2.4 ds (20)
Canada
0.8-6.2 ef / 1-8.3ds (28)
DBCAL <1-2.9 ef U.S. (24)
n.d-11.5 ef / 0.7-0.9 ds (20)
Canada
0.6-9 ef / 0.7-11.6 ds (28)
* median value, ** ef: plant effluent, ds: distribution system, p: swimming pool.
According to different studies, carbonyl compounds such as aldehydes,

which are precursors of halogenated aldehydes, are the most common ozonation
by-products (53, 54). Therefore the application of ozone before chlorination may
increase the HAL formation potential. This was confirmed for CH in several
studies (55, 56). Krasner et al. (56) also observed that tri-HAL concentrations
were lower than di-HAL concentrations in chloramine-based systems, and that the
formation of chloro and bromo containing di-HALs was also enhanced in those
plants using ozone before chloramination. Contrary to what might be expected
from these results, overall HAL levels formed in preozonation based systems
investigated by Jeong et al. (24) were not higher than those generated in DWTPs
that did not apply ozone. Moreover, the highest tri-HAL concentrations were
found in chloraminated water, with TBAL the main contributing HAL species
(44%) to total HAL levels (28.8 µg/L). However, these findings could be biased
by the extraordinarily high concentrations of bromide (0.54 µg/L) present in the
source water that had the highest HAL generation potential (24). If this sample is
excluded in this analysis, the highest HAL concentration was observed for CH in
chlorinated water, whereas the formation potential of CH during chloramination
processes was very low (24). Bromine incorporation in HALs in the samples
investigated by Jeong et al. (24) increased with bromide concentration of the
33
source water, as previously reported by Koudjonou et al. (28). In this regard,
the lowest concentrations, and thus, contributions to total HAL levels, of the
chlorine-based HAL species, CAL, DCAL, TCAL, and BDCAL, were measured
in disinfected waters that originated from the highest bromide content source
waters (24).
Besides haloacetaldehydes, additional halogenated aldehydes, i.e.,
iodobutanal, dichloropropenal, and 4-chloro-2-butenal, were identified in the U.S.
Nationwide DBP Occurrence Study (5, 11). However, their concentrations and
formation mechanisms have not been investigated yet.
Figure 2. a) Frequency of detection and b) concentration range of individual

HALs in finished drinking waters collected in US drinking water treatment plants
(24).
Haloacetaldehyde Toxicity
As reviewed in the Introduction, toxicological studies were previously only
conducted on CAL, BAL, DCAL, CH and TBAL. A recent study examined the
genotoxicity and cytotoxicity of the complete set of the ten HALs, including IAL
(24). These toxicity assays were performed in mammalian cells as described in
the next subsections. Results obtained allowed comparison of the toxic potencies
of individual HALs and of this DBP class with other DBP classes.
34
Toxicity Assays
Chinese Hamster Ovary Cells
Chinese hamster ovary (CHO) cell line AS52, clone 11-4-8 was used for
the toxicity studies (62–64) and was maintained in modified Ham’s F12 medium
(Mediatech, Inc., Manassas, VA) supplemented with 5% heat-inactivated fetal
bovine serum (FBS), 1% L-glutamine and 1% antibiotic-antimycotic solution
(Invitrogen, Carlsbad, CA) at 37°C in a humidified atmosphere of 5% CO2. The
cells exhibit normal morphology, express cell contact inhibition, and grow as a
monolayer without expressing neoplastic foci.
CHO Cell Chronic Cytotoxicity Assay
This bioanalytical assay measures the survivorship of the cells as the

reduction in cell density as a function of the HAL concentration over a period of
72 h. (22, 65). The procedure was previously published (22, 65). In general, for
each HAL concentration, 8 replicates were analyzed and the experiments were
repeated 2-4 times. A concentration-response curve was generated for each HAL,
and a regression analysis was generated for each curve. The LC50 values were
calculated, where the LC50 represents the HAL concentration that induced a 50%
reduction in cell density as compared to the concurrent negative controls.
Single Cell Gel Electrophoresis Assay
The single cell gel electrophoresis (SCGE) or “comet assay” quantitatively

measures genomic DNA damage in individual nuclei (66–68). The detailed
procedure of the microplate methodology used in this study was published
elsewhere (68). The SCGE metric for genomic DNA damage induced by the
HALs was the %Tail DNA value, which is the amount of DNA that migrated
from the nucleus into the microgel (69). For each HAL concentration range where
the cell viability was >70%, a concentration-response curve was generated. A
regression analysis was used to fit the curve, and the concentration inducing a
50% Tail DNA value was calculated.
Comparative Haloacetaldehyde Toxicity
The CHO cell chronic cytotoxicity analyses (72 h exposures) of each

HAL are summarized in Table 3. The mean bootstrap cytotoxicity index (CTI)
(LC50−1)(103) (±SE) values are presented in Figure 3a. An all pairwise ANOVA
test of the CTI values generated a descending rank order of chronic cytotoxicity
as TBAL ≈ CAL > DBAL ≈ BCAL ≈ DBCAL > IAL > BAL ≈ BDCAL > DCAL
> TCAL, as it is shown in Figure 3a.
35
CHO cell acute genotoxicity analyses (4 h exposures) of each HAL are
summarized in Table 3. An all pairwise ANOVA test of the GTI values generated
a descending rank order of genotoxicity of the ten HALs as DBAL > CAL ≈
DBCAL > TBAL ≈ BAL > BDCAL > BCAL ≈ DCAL > IAL, as shown in Figure
3b. TCAL was not genotoxic.
Table 3. Mammalian Cell Cytotoxicity and Genotoxicity of HALs (24)

Lowest Lowest 50%
LC50
HALs Cytotoxic Conc. Genotoxic Tail DNA
(µM) b
(µM) a Conc. (µM) c (µM) d
CAL 0.5 3.5 100 142.8

BAL 8 17.3 200 381.2
Mono- IAL 5 6.0 900 1009
HALs and
Di-HALs DCAL 8 29.3 800 795
DBAL 2 4.7 50 111.3
BCAL 2.5 5.3 500 621.4
CH
375 116 NS NS
(TCAL)
Tri-HALs TBAL 2 3.6 100 340.3

BDCAL 10 20.4 300 470.4
DBCAL 4 5.2 100 143.7
a Lowest cytotoxic concentration was the lowest concentration of the HAL in the
concentration-response curve that induced a statistically significant reduction in cell density
as compared to the concurrent negative controls. b The LC50 value is the concentration
of the HAL, determined from a regression analysis of the data, that induced a cell
density of 50% as compared to the concurrent negative controls. c The lowest genotoxic
concentration was the lowest concentration of the HAL in the concentration-response
curve that induced a statistically significant amount of genomic DNA damage as compared
to the negative control. d The SCGE 50% Tail DNA value is the HAL concentration
determined from a regression analyses of the data that was calculated to induce a 50%
SCGE Tail DNA value.
We compared the CHO cell toxic potencies of the HALs to other DBP
chemical classes with their calculated cytotoxicity and genotoxicity indices.
The cytotoxicity index was determined by calculating the mean LC50 value of
all of the individual compounds of a single class of DBPs. The genotoxicity
index was determined by calculating the mean SCGE genotoxic potency value,
which is defined by the midpoint of the SCGE tail moment or the 50% Tail
DNA values from the individual compounds within a single class of DBPs (22).
Six DBP chemical classes were compared, including the THMs, HAAs, HALs,
halonitromethanes, haloacetonitriles, and haloacetamides. As summarized in
36
Table 4, HALs constitute the second most cytotoxic DBP class, whereas they rank
as the second least genotoxic DBP class.
Figure 3. The mean bootstrap a) CHO cell cytotoxic index values (±SE) and b)
CHO cell genotoxic index values (±SE) of the HALs (24).
Table 4. Mammalian Cell Cytotoxicity Index Values (CTI) and Genotoxicity

Index Values (GTI) of DBP Chemical Classes (24)
CTI GTI
DBP Class
Values Values
Trihalomethanes 2.39 0
Haloacetic acids 39.3 14.8
Haloacetaldehydes 144 2.80
Halonitromethanes 75.0 11.4
Haloacetonitriles 16.0 11.5
Haloacetamides 302 12.3
Conclusions
Different analytical methods have been used to determine HALs in disinfected
waters. Due to their very different physical-chemical properties, the complete
set of nine chloro and bromo HALs and IAL cannot be analyzed by a unique
approach. Regardless of the extraction approach followed, analyte detection
is carried out by GC coupled to ECD or MS detection. The evaluation of the
occurrence of HALs in drinking waters allowed this DBP class to be identified
as the third largest group by weight of DBPs reported. Besides CH (TCAL),
which is currently the most abundant HAL in drinking water, bromo containing
acetaldehydes and IAL may be relevant in chloraminated waters that contain
37
high levels of bromide and iodide. Despite the fact that HAL concentrations
may increase in preozonation-based systems, individual HAL concentrations in
disinfected water strongly depend on the source water quality and the type of
disinfection treatment applied. A systematic quantitative comparative study of
their acute cytotoxicity and chronic genotoxicity to mammalian cells revealed
HALs as the second most cytotoxic class of DBPs after haloacetamides. Taking
all this into account, it is important to determine their potential health risks and
to investigate their formation mechanisms to control their presence in drinking
waters.
Acknowledgments
We would like to thank the drinking water treatment plants for generously
providing us with samples. This work was supported by the EPA STAR Grant
R834867. We appreciate support by the Center of Advanced Materials for
the Purification of Water with Systems (WaterCAMPWS), a National Science
Foundation Science and Technology Center, under Award CTS-0120978. C.J. was
supported by a NIEHS Pre-doctoral Fellowship under Grant No. T32 ES007326.
C.P. acknowledges support from the European Union Seventh Framework
Programme (FP7/2007-2013) under grant agreement n° 274379 (Marie Curie
IOF). This work has been financially supported by the Generalitat de Catalunya
(Consolidated Research Groups “2014 SGR 418 - Water and Soil Quality Unit”
and 2014 SGR 291 - ICRA). This work reflects only the author’s views. The EU
is not liable for any use that may be made of the information contained therein.
This work does not reflect EPA policy.
References
1. Liang, L.; Singer, P. C. Factors influencing the formation and relative
distribution of haloacetic acids and trihalomethanes in drinking water.
2. Hua, G.; Reckhow, D. A. Comparison of disinfection byproduct formation
3. Shah, A. D.; Mitch, W. A. Halonitroalkanes, halonitriles, haloamides, and
N-nitrosamines: A critical review of nitrogenous disinfection byproduct
formation pathways. Environ. Sci. Technol. 2012, 46, 119–131.
4. Zhang, Z.; Echigo, S.; Minear, R. A.; Plewa, M. J., Characterization and
Eds. American Chemical Society: Washington, DC, 2000; pp 299−314.
Sclimenti, M. J.; Onstad, G. D.; Thruston, A. D., Jr. Occurrence of a new
generation of disinfection byproducts. Environ. Sci. Technol. 2006, 40,
7175–7185.
38
6. Richardson, S. D.; Postigo, C., Drinking water disinfection by-products. In
Emerging Organic Contaminants and Humand Health; Barcelo, D., Ed.; The
Handbook of Environmental Chemistry; Springer-Verlag: Berlin Heidelberg,
2012; Vol. 20, pp 93−138.
7. Plewa, M. J.; Wagner, E. D. Drinking water disinfection by-products:
comparative mammalian cell cytotoxicity and genotoxicity. In Encyclopedia
of Environmental Health; Nriagu, J. O., Ed.; Elsevier: Burlington, 2011;
Vol. 1, pp 806-812.
8. Council Directive 98/83/EC of 3 November 1998 on the quality of water
intended for human consumption. Official Journal of the European
Communities L 330/32; http://goo.gl/REht5N (accessed November, 2014).
9. U.S. EPA. Basic information about disinfection byproducts in drinking water:
total trihalomethanes, haloacetic acids, bromate and chlorite; http://goo.gl/
EEVHZ3 (accessed November, 2014).

research. Mutat. Res., Rev. Mutat. Res. 2007, 636, 178–242.
11. Weinberg, H. S.; Krasner, S. W.; Richardson, S. D.; Thruston, A. D. J.
The Occurrence of Disinfection By-Products (DBPs) of Health Concern in
Drinking Water: Results of a Nationawide DBP Occurrence Study; National
Exposure Research Laboratory, Office of Research and Development, U.S.
Environmental Protection Agency: Athens, GA, 2002.
12. Beland, F. A. National Toxicology Program technical report on the
toxicity and metabolism studies of chloral hydrate (CAS No. 302-17-0).
Administered by Gavage to F344/N rats and B6C3F1 mice. Natl. Toxicol.
Program Toxic. Rep. Ser. 1999, 59, A1–A7.
13. Furnus, C. C.; Ulrich, M. A.; Terreros, M. C.; Doulout, F. N. The induction of
aneuploidy in cultured Chinese hamster cells by propionaldehyde and chloral
hydrate. Mutagenesis 1990, 5, 323–326.
14. Allen, J. W.; Collins, B. W.; Evansky, P. A. Spermatid micronucleus analyses
of trichloroethylene and chloral hydrate effects in mice. Mutat. Res. 1994,
323, 81–88.
15. Migliore, L.; Nieri, M. Evaluation of twelve potential aneuploidogenic
chemicals by the in vitro human lymphocyte micronucleus assay. Toxicol.
In Vitro 1991, 5, 325–336.
16. Crebelli, R.; Conti, G.; Conti, L.; Carere, A. Induction of Somatic
Segregation by Halogenated Aliphatic-Hydrocarbons in Aspergillus-
Nidulans. Mutat. Res. 1984, 138, 33–38.
17. Crebelli, R.; Conti, G.; Conti, G.; Carere, A. Chloroacetaldehyde is
a powerful inducer of mitotic aneuploidy in Aspergillus-Nidulans.
Mutagenesis 1990, 5, 165–168.
18. Liviac, D.; Creus, A.; Marcos, R. DNA damage induction by two halogenated
acetaldehydes, byproducts of water disinfection. Water Res. 2010, 44,
2638–2646.
19. Guengerich, F. P.; Mason, P. S.; Stott, W. T.; Fox, T. R.; Watanabe, P. G.
Roles of 2-haloethylene oxides and 2-haloacetaldehydes derived from vinyl
39
bromide and vinyl chloride in irreversible binding to protein and DNA.
Cancer Res. 1981, 41, 4391–4398.
20. Koudjonou, B. K.; LeBel, G. L. Halogenated acetaldehydes: Analysis,
stability and fate in drinking water. Chemosphere 2006, 64, 795–802.
McKague, A. B.; Miltner, R. J.; Wagner, E. D.; Plewa, M. J. Occurrence and
Mammalian Cell Toxicity of Iodinated Disinfection Byproducts in Drinking
Water. Environ. Sci. Technol. 2008, 42, 8330–8338.
of Disinfection By-products. Water Research Foundation: Denver, CO, 2009;
p 134.
23. Chinn, R.; Lee, T.; Krasner, S. W.; Dale, M.; Richardson, S. D.; Pressman,
J. G.; Speth, T. F.; Miltner, R. J.; Simmons, J. E. Solid phase extraction

of 35 DBPs with analysis by GC/ECD and GC/MS. In Proceedings of the
Water Quality Technology Conference; American Water Works Association:
Denver, CO, 2007; pp 1−20, http://goo.gl/U3h3cd (accessed November,
2014).
24. Jeong, C. H.; Postigo, C.; Richardson, S. D.; Simmons, J. E.; Kimura, S. Y.;
Marinas, B. J.; Barcelo, D.; Wagner, E. D.; Plewa, M. The occurrence and
comparative toxicity of haloacetaldehyde disinfection byproducts in drinking
water. Environ. Sci. Technol. 2014 submitted for publication.
25. Kristiana, I.; Lethorn, A.; Joll, C.; Heitz, A. To add or not to add: The use
of quenching agents for the analysis of disinfection by-products in water
samples. Water Res. 2014, 59, 90–98.
26. Serrano, M.; Silva, M.; Gallego, M. Micro liquid–liquid extraction combined
with large-volume injection gas chromatography–mass spectrometry for the
determination of haloacetaldehydes in treated water. J. Chromatogr. A 2011,
1218, 8295–8302.
27. Linge, K. L.; Blythe, J. W.; Busetti, F.; Blair, P.; Rodriguez, C.; Heitz, A.
Formation of halogenated disinfection by-products during microfiltration and
reverse osmosis treatment: Implications for water recycling. Sep. Purif.
Technol. 2013, 104, 221–228.
28. Koudjonou, B.; LeBel, G. L.; Dabeka, L. Formation of halogenated
acetaldehydes, and occurrence in Canadian drinking water. Chemosphere
2008, 72, 875–881.
29. Much, D. J.; Hautman, D. P. U.S. EPA Method 551.1. Methods for the
Determination of Organic Compounds in Drinking Water - Supplement III
- EPA/600/R-95/131; 1995.
30. Bond, T.; Henriet, O.; Goslan, E. H.; Parsons, S. A.; Jefferson, B.
Disinfection byproduct formation and fractionation behavior of natural
organic matter surrogates. Environ. Sci. Technol. 2009, 43, 5982–5989.
31. Bougeard, C. M. M.; Goslan, E. H.; Jefferson, B.; Parsons, S. A. Comparison
of the disinfection by-product formation potential of treated waters exposed
to chlorine and monochloramine. Water Res. 2010, 44, 729–740.
40
32. Cimetiere, N.; De Laat, J. Effects of UV-dechloramination of swimming
pool water on the formation of disinfection by-products: A lab-scale study.
Microchem. J. 2014, 112, 34–41.
33. Doederer, K.; Farré, M. J.; Pidou, M.; Weinberg, H. S.; Gernjak, W. Rejection
of disinfection by-products by RO and NF membranes: Influence of solute
properties and operational parameters. J. Membr. Sci. 2014, 467, 195–205.
34. Fang, J. Y.; Ma, J.; Yang, X.; Shang, C. Formation of carbonaceous and
nitrogenous disinfection by-products from the chlorination of Microcystis
aeruginosa. Water Res. 2010, 44, 1934–1940.
35. Golfinopoulos, S. K.; Nikolaou, A. D. Survey of disinfection by-products in
drinking water in Athens, Greece. Desalination 2005, 176, 13–24.
36. Hansen, K. M. S.; Zortea, R.; Piketty, A.; Vega, S. R.; Andersen, H. R.
Photolytic removal of DBPs by medium pressure UV in swimming pool
water. Sci. Total Environ. 2013, 443, 850–856.

37. Lee, J.; Jun, M.-J.; Lee, M.-H.; Lee, M.-H.; Eom, S.-W.; Zoh, K.-D.
Production of various disinfection byproducts in indoor swimming pool
waters treated with different disinfection methods. Int. J. Hyg. Environ.
Health 2010, 213, 465–474.
38. Nikolaou, A.; Golfinopoulos, S.; Rizzo, L.; Lofrano, G.; Lekkas, T.;
Belgiorno, V. Optimization of analytical methods for the determination of
DBPs: Application to drinking waters from Greece and Italy. Desalination
2005, 176, 25–36.
39. Radjenovic, J.; Farre, M. J.; Mu, Y.; Gernjak, W.; Keller, J. Reductive
electrochemical remediation of emerging and regulated disinfection
byproducts. Water Res. 2012, 46, 1705–1714.
40. Villanueva, C. M.; Castaño-Vinyals, G.; Moreno, V.; Carrasco-Turigas, G.;
Aragonés, N.; Boldo, E.; Ardanaz, E.; Toledo, E.; Altzibar, J. M.; Zaldua, I.;
Azpiroz, L.; Goñi, F.; Tardón, A.; Molina, A. J.; Martín, V.; López-Rojo, C.;
Jiménez-Moleón, J. J.; Capelo, R.; Gómez-Acebo, I.; Peiró, R.; Ripoll, M.;
Gracia-Lavedan, E.; Nieuwenhujsen, M. J.; Rantakokko, P.; Goslan, E. H.;
Pollán, M.; Kogevinas, M. Concentrations and correlations of disinfection
by-products in municipal drinking water from an exposure assessment
perspective. Environ. Res. 2012, 114, 1–11.
41. Wei, J. R.; Ye, B. X.; Wang, W. Y.; Yang, L. S.; Tao, J.; Hang, Z. Y. Spatial
and temporal evaluations of disinfection by-products in drinking water
distribution systems in Beijing, China. Sci. Total Environ. 2010, 408,
4600–4606.
42. Gan, W.; Guo, W.; Mo, J.; He, Y.; Liu, Y.; Liu, W.; Liang, Y.; Yang, X.
The occurrence of disinfection by-products in municipal drinking water in
China’s Pearl River Delta and multipathway cancer risk assessment. Sci.
Total Environ. 2013, 447, 108–115.
43. Munch, J. W.; Munch, D. J.; Winslow, S. D. U.S. EPA Method 556
for the Determination of Carbonyl Compounds in Drinking Water
by Pentafluorobenzlhydroxylamine Derivatization and Capillary Gas
Chromatography with Electron Capture Detection; 1997; http://goo.gl/
QDCtd5 (accessed November 2014).
41
44. Dąbrowska, A.; Świetlik, J.; Nawrocki, J. Formation of aldehydes upon ClO2
45. Włodyka-Bergier, A.; Bergier, T.; Kot, M. Occurrence of volatile organic
chlorination by-products in water distribution system in Krakow (Poland).
Desalin. Water Treat. 2014, 52, 3898–3907.
46. Włodyka-Bergier, A.; Rajca, M.; Bergier, T. Removal of halogenated
by-products precursors in photocatalysis process enhanced with membrane
filtration. Desalin. Water Treat. 2014, 52, 3698–3707.
47. Doederer, K.; Gernjak, W.; Weinberg, H. S.; Farré, M. J. Factors affecting the
formation of disinfection by-products during chlorination and chloramination
of secondary effluent for the production of high quality recycled water. Water
Res. 2014, 48, 218–228.
48. Lyon, B. A.; Dotson, A. D.; Linden, K. G.; Weinberg, H. S. The effect
of inorganic precursors on disinfection byproduct formation during

UVchlorine/chloramine drinking water treatment. Water Res. 2012, 46,
4653–4664.
49. Yang, X.; Guo, W.; Lee, W. Formation of disinfection byproducts upon
chlorine dioxide preoxidation followed by chlorination or chloramination of
natural organic matter. Chemosphere 2013, 91, 1477–1485.
50. Yang, X.; Guo, W.; Zhang, X.; Chen, F.; Ye, T.; Liu, W. Formation of
disinfection by-products after pre-oxidation with chlorine dioxide or ferrate.
Water Res. 2013, 47, 5856–5864.
51. Yang, X.; Peng, J.; Chen, B.; Guo, W.; Liang, Y.; Liu, W.; Liu, L. Effects of
ozone and ozone/peroxide pretreatments on disinfection byproduct formation
during subsequent chlorination and chloramination. J. Hazard. Mater. 2012,
239-240, 348–354.
52. WHO - World Health Organization. Guidelines for Drinking-Water Quality,
4th ed.; 2011; http://goo.gl/fW7qJG (accessed October 2014).
53. Weinberg, H. S.; Glaze, W. H.; Krasner, S. W.; Sclimenti, M. J. Formation
and removal of aldehydes in plants that use ozonation. J. - Am. Water Works
Assoc. 1993, 85, 72–85.
54. Dabrowska, A.; Kasprzyk-Hordern, B.; Nawrocki, J. Aldehydes formation
during water disinfection by ozonation and chlorination process. Global
NEST J. 2005, 7, 61–71.
55. Dabrowska, A.; Nawrocki, J. Controversies about the occurrence of chloral
hydrate in drinking water. Water Res. 2009, 43, 2201–2208.
56. Krasner, S. W.; Mitch, W. A.; Westerhoff, P.; Dotson, A. Formation and
control of emerging C- and N-DBPs in drinking water. J. - Am. Water Works
Assoc. 2012, 104, E582–E595.
57. Williams, D. T.; Lebel, G. L.; Benoit, F. M. Disinfection by-products in
canadian drinking waters. Chemosphere 1997, 34, 299–316.
58. Aranda-Rodriguez, R.; Koudjonou, B.; Jay, B.; Lebel, G. L.; Benoit, F. M.
Disinfection by-products (DBPs) in drinking water from eight systems using
chlorine dioxide. Water Qual. Res. J. Can. 2008, 43, 11–22.
59. Kawamoto, T.; Makihata, N. Distribution of bromine/chlorine-containing
disinfection by-products in tap water from different water sources in the
Hyogo Prefecture. J. Health Sci. 2004, 50, 235–247.
42
60. Lee, K. J.; Kim, B. H.; Hong, J. E.; Pyo, H. S.; Park, S. J.; Lee, D. W. A
study on the distribution of chlorination by-products (CBPs) In treated water
in Korea. Water Res. 2001, 35, 2861–2872.
61. Simpson, K. L.; Hayes, K. P. Drinking water disinfection by-products: An
Australian perspective. Water Res. 1998, 32, 1522–1528.
62. Hsie, A. W.; Brimer, P. A.; Mitchell, T. J.; Gosslee, D. G. The
dose-response relationship for ultraviolet-light-induced mutations at the
hypoxanthine-guanine phosphoribosyltransferase locus in Chinese hamster
ovary cells. Somatic Cell Genet. 1975, 1, 383–389.
63. Tindall, K. R.; Stankowski, L. F., Jr; Machanoff, R.; Hsie, A. W. Detection
of deletion mutations in pSV2gpt-transformed cells. Mol. Cell. Biol. 1984,
4, 1411–1415.
64. Wagner, E. D.; Rayburn, A. L.; Anderson, D.; Plewa, M. J. Analysis of
mutagens with single cell gel electrophoresis, flow cytometry, and forward
mutation assays in an isolated clone of Chinese hamster ovary cells. Environ.
Mol. Mutagen. 1998, 32, 360–368.
66. Rundell, M. S.; Wagner, E. D.; Plewa, M. J. The comet assay: Genotoxic
damage or nuclear fragmentation? Environ. Mol. Mutagen. 2003, 42, 61–67.
67. Tice, R. R.; Agurell, E.; Anderson, D.; Burlinson, B.; Hartmann, A.;
Kobayashi, H.; Miyamae, Y.; Rojas, E.; Ryu, J. C.; Sasaki, Y. F. Single
cell gel/comet assay: Guidelines for in vitro and in vivo genetic toxicology
testing. Environ. Mol. Mutagen. 2000, 35, 206–221.
68. Wagner, E. D.; Plewa, M. J. Microplate-based comet assay. In The Comet
Assay in Toxicology, Dhawan, A., Anderson, D., Eds.; Royal Society of
Chemistry: London, 2008; pp 79−97.
69. Kumaravel, T. S.; Jha, A. N. Reliable Comet assay measurements for
detecting DNA damage induced by ionising radiation and chemicals. Mutat.
Res., Genet. Toxicol. Environ. Mutagen. 2006, 605, 7–16.
43
Chapter 3
Effect of Boiling on Halogenated DBPs

and Their Developmental Toxicity in
Real Tap Waters
Jiaqi Liu, Xiangru Zhang,* and Yu Li
Department of Civil and Environmental Engineering,

The Hong Kong University of Science and Technology,
Clear Water Bay, Kowloon, Hong Kong, China
The use of chlorine for disinfection results in the formation

of halogenated disinfection by-products (DBPs) in tap water.
Evidence has shown that halogenated DBPs may cause chronic
adverse effects on human health, and brominated DBPs
are generally significantly more toxic than their chlorinated
analogues. Previously, the authors’ group found that boiling of
a simulated tap water for 5 min significantly reduced the overall
levels of brominated and chlorinated DBPs, thus reducing the
cytotoxicity of the simulated tap water to mammalian cells.
In this study, we further investigated the effect of boiling on
the level and developmental toxicity of halogenated DBPs
in two “real” tap water samples. With a novel precursor ion
scan approach using electrospray ionization-triple quadrupole
mass spectrometry, the whole pictures of polar brominated and
chlorinated DBPs in both tap water samples without and with
boiling were revealed. After 5 min boiling, the concentrations of
total organic bromine (a collective parameter for all brominated
DBPs) in the two tap water samples decreased by 43.6% and
37.5%, respectively; the concentrations of total organic chlorine
(a collective parameter for all chlorinated DBPs) in the two tap
water samples decreased by 39.0% and 57.1%, respectively; the
developmental toxicity of the two tap water samples decreased
by 53.0% and 57.1%, respectively. This study suggests a simple

way to reduce the adverse health effect of halogenated DBPs in
humans through tap water ingestion.
Introduction
With the purpose of inactivating harmful microorganisms and eradicating
waterborne diseases, disinfection has been applied in drinking water treatment.
Chlorine is a widely used disinfectant for drinking water. Chlorine-disinfected
drinking water is distributed to households via a public water distribution system,
in which a certain level of chlorine residual should be maintained to prevent
regrowth of microorganisms in the water, especially when breaks or cracks
occur in the pipeline. However, chlorinated disinfection by-products (DBPs)
unintentionally form (from the reaction between hypochlorous acid and natural
organic matter in source water) during chlorination in a drinking water treatment
plant and in a water distribution system. Besides, source waters contain bromide
ions resulting from geologic dissolution, seawater intrusion and human activities
(1). During chlorination, hypochlorous acid can oxidize bromide to hypobromous
acid, which subsequently reacts with natural organic matter in source water to
generate brominated DBPs. Brominated DBPs have been reported to exhibit
significantly higher toxicity than their chlorinated analogues (2–6).
Humans are unavoidably exposed to DBPs via water ingestion. It has
been reported that the total cancer risks from the commonly known haloacetic
acids and trihalomethanes mainly result from tap water ingestion (7). In most
Western nations, tap water is directly drunk, whereas in some Asian nations,
people are used to boiling tap water before drinking. Boiling has its own
advantages, including effectively removing chlorine or chloramine residual
and thus eliminating chlorinous taste and odor of the water (8); inactivating
chlorine-resistant pathogenic protozoa Cryptosporidium and Giardia (9); and
reducing some major DBPs (including trihalomethanes, haloacetic acids,
haloketones, haloacetonitriles, haloaldehydes, and halonitromethanes) (10–12).
Recently, a powerful precursor ion scan (PIS) approach with electrospray
ionization-triple quadrupole mass spectrometry coupled with ultra performance
liquid chromatography (UPLC/ESI-tqMS) has been developed for selectively
detecting polar halogenated DBPs (13–19). By using this approach, the authors’
group has studied the effect of boiling on the halogenated DBPs in a simulated
tap water, and found that boiling for 5 min significantly reduced the halogenated
DBPs and cytotoxicity of the water (20). However, a real tap water may be
different from the simulated tap water because a real tap water originates from
a real source water, which may contain a variety of organic and inorganic
components including (micro)pollutants. Accordingly, the purpose of this study
was to determine the effect of boiling on the halogenated DBPs and toxicity in
real tap waters.
Hutchinson’s group developed an in vivo assay using the sensitive
embryo-larval stages of a polychaete, Platynereis dumerilii, which is a
cosmopolitan species, extending from the tropics to cold temperate latitudes
in both hemispheres (21, 22). The authors’ group modified Hutchinson et al.’s
46
method and significantly lowered the relative standard deviation to <5% (5). The
improved bioassay becomes a sensitive metric and has been successfully applied
in determining the comparative developmental toxicity of individual DBPs
(5, 23). In this study, P. dumerilii was employed to evaluate the comparative
developmental toxicity of real tap waters without and with boiling.
Experimental Methods
Water Sampling and Characterization
Two source water samples (source waters 1 and 2) were collected at the inlets
of two drinking water treatment plants. Correspondingly, two tap water samples
that originated from the two drinking water treatment plants were collected at
two faucets on the same day. Tap waters 1 and 2 corresponded to source waters
1 and 2, respectively. The collected water samples were filtered through 0.45 μm
membrane. Dissolved organic carbon (DOC), UV absorbance, pH, and bromide
concentration of each source water sample were measured with a total organic
carbon analyzer (Model TOC-Vcsh, Shimadzu), a UV/visible spectrophotometer
(GE healthcare, Ultrospec 4300 pro), a pH meter, and an ion chromatograph
(Dionex DX500), respectively. Alkalinity in each source water sample and
chlorine residual in each tap water sample were determined according to the
Standard Methods (24). The characteristics of the source water samples were
summarized in Table 1. The chlorine residual levels in tap waters 1 and 2 were
0.75 and 0.63 mg/L as Cl2, respectively. For comparison, the characteristics of
the simulated source water used in the previous study were also listed in Table
1. This simulated source water was disinfected with 5 mg/L NaOCl as Cl2 in
darkness at ~21 ºC for 15 h to generate the simulated tap water. No chlorine
residual was detectable in the simulated tap water (20).
Table 1. Characteristics of Two Real Source Waters and the Simulated

Source Water
Source water 1 2 Simulateda
DOC (mg/L as C) 2.68 2.06 3.00
UV254 absorbance (1/cm) 0.032 0.027 0.16
SUVAb (L/mg-m) 1.20 1.31 5.17
pH 6.94 7.22 8.50
Alkalinity (mg/L as CaCO3) 25.1 22.1 90.0
Bromide (µg/L) 26.5 16.7 2000
a The simulated source water was prepared with ultrapure water containing 3 mg/L
Suwannee River humic acid as C, 90 mg/L NaHCO3 as CaCO3, and 2 mg/L KBr as Br−
(20). b Specific UV absorbance.
47
Chemicals and Seawater
All chemical solutions were prepared from chemicals of reagent grade or

above. 3,5-Dibromo-4-hydroxybenzaldehyde was purchased from Alfa Aesar.
Bromobutenedioic acid, bromoacetic acid, dibromoacetic acid, bromochloroacetic
acid, dichloroacetic acid, dibromochloroacetic acid, bromodichloroacetic
acid, trichloroacetic acid, methyl tert-butyl ether (MtBE), and acetonitrile
(HPLC-grade) were purchased from Sigma-Aldrich. Seawater was collected
locally. Prior to use, it was filtered through 0.45 μm membrane, autoclaved at 121
ºC for 20 min. After cooled to ambient temperature, the seawater was aerated for
15 min.
Pretreatment of Tap Water Samples with/without Boiling
Our previous study showed that boiling for 5 min obviously reduced the
halogenated DBPs and toxicity of the simulated tap water (20). Accordingly,
for either tap water 1 or tap water 2, two parallel samples (3.4 L each) were
heated to boiling (at 100 ºC) in open 5 L glass beakers, and kept boiling for 5
min. Afterwards, both boiled water samples were cooled to ambient temperature
quickly in an ice bath (to minimize the variation of the sample components during
cooling). Then, both samples were brought to 3.4 L by adding ultrapure water,
and the chlorine residual in each sample was measured (24). After boiling, no
chlorine residual was detectable, which was consistent with Zhang’s results (8).
One of the parallel samples was divided into two aliquots (300 mL and
3000 mL). The 300 mL aliquot was subjected to total organic halogen (TOX)
measurement, and the 3000 mL aliquot was pretreated per Zhang et al.’s procedure
(25). In brief, it was further divided into three 1000 mL aliquots, and each aliquot
was first acidified to pH 0.5 with 70% (v/v) aqueous sulfuric acid, followed by the
addition of sodium sulfate to saturation. Then, the acidified sample was extracted
with 100 mL MtBE, and 70 mL of the organic layer was transferred to a rotary
evaporator and concentrated to 0.5 mL. The 0.5 mL MtBE solutions obtained
from three aliquots were combined together (totally ~1.5 mL) and mixed with
20 mL acetonitrile. The mixture was concentrated to 1 mL, and stored at 4 °C.
Prior to UPLC/ESI-tqMS measurement, the acetonitrile solution was diluted to
2 mL with ultrapure water, and filtered with 0.45 μm membrane. For the other
parallel sample, 3000 mL was pretreated following the procedure above, except
that after evaporation, the 0.5 mL MtBE solutions obtained from three aliquots
were combined together and dried with nitrogen sparging. The solid obtained
was stored at 4 °C and dissolved in seawater 1 h prior to the toxicity test as a
toxicity test stock solution. This stock solution was diluted by seawater to various
concentration factors (in comparison with the original water sample).
For comparison, two 3.4 L parallel aliquots of each tap water sample without
boiling were pretreated following the procedure above, except that the chlorine
residual in the 300 mL aliquot for TOX measurement was quenched with NaAsO2
at 105% of the stoichiometric amount of the chlorine residual (26).
48
TOX Measurement
TOX was measured according to Standard Method 5320B (24), except that
an off-line ion chromatograph was used as a halide separator and detector for the
measurement of total organic chlorine (TOCl) and total organic bromine (TOBr)
(26–28). Total organic iodine (TOI) was not included in this study because the
TOI levels in the two tap waters were very low (about 1.3 µg/L as I) (29). Each
300 mL tap water sample without or with boiling was adjusted to pH 2 with
nitric acid. A 3-channel adsorption module (Mitsubishi Chemical Analytech)
was used for activated carbon adsorption. After adsorption, the activated carbon
columns were rinsed with 10 mL of 5000 mg/L KNO3 as NO3− to remove
inorganic halides and were subsequently subjected to pyrolysis at 1000 °C with
an AQF-100 automatic quick furnace (Mitsubishi Chemical Analytech). The
hydrogen halide and halogen gases produced from pyrolysis were absorbed in a 5
mL 0.003% H2O2 solution, which contained 2 mg/L KH2PO4 as PO43− serving as
an internal standard to estimate the volume variations introduced by the GA-100
gas absorption unit (Mitsubishi Chemical Analytech). Thus, TOCl and TOBr
in the original water sample were converted and transferred into the absorption
solution in the form of Cl− and Br−, respectively. One mL of the absorption
solution was analyzed by an ICS-3000 ion chromatograph (Dionex) with an
IonPac analytical column (AS19, 4×250 mm) and a guard column (AG19, 4×50
mm). A KOH eluent was generated by the EGC KOH cartridge (Dionex) at a flow
rate of 1 mL/min. Chloride and bromide were separated with an isocratic eluent
of 10 mM KOH from 0 to 10 min followed by a linear gradient eluent of 10-45
mM KOH from 10 to 25 min. The concentrations of the halides were quantified
with a conductivity detector. The practical quantitation limits for TOCl and TOBr
were 0.002 mg/L as Cl and 0.002 mg/L as Br, respectively (26). Each sample was
subjected to triplicate TOX analyses.
(UPLC/)ESI-tqMS Analyses
An ESI-tqMS (Waters Acquity) was employed to analyze the pretreated

water samples. The operation parameters were set according to previous studies
(14, 20): ESI negative mode; capillary voltage 2.8 kV; cone voltage 30 V for
chlorinated DBPs, and 15 V for brominated DBPs; desolvation temperature 350
ºC; source temperature 120 ºC; cone gas flow 50 L/h; desolvation gas flow 650
L/h; collision energy 30 eV for chlorinated DBPs, and 20 eV for brominated
DBPs; Argon collision gas flow 0.25 mL/min; mass resolution 15 (1-unit
resolution). By setting PISs m/z 79/81 (or m/z 35/37), all electrospray-ionizable
brominated (or chlorinated) DBPs can be detected. Multi-channel analysis mode
was used for data collection for all PISs (with a scan time of 0.3 s, run durations
of 5 min for PISs m/z 79/81 and 7 min for PISs m/z 35/37), by which precursor
ion intensities were substantially enhanced by accumulating multiple scans (1000
scans for PISs m/z 79/81 and 1400 scans for PISs m/z 35/37), and intensity
fluctuation in a single scan was effectively eliminated.
A UPLC (Waters) was coupled with the ESI-tqMS for preseparation. Seven
μL of a pretreated water sample was injected into the UPLC with an HSS T3
49
column (Waters, 100 × 2.1 mm, 1.8 μm particle size). The eluent was kept
at 0.50 mL/min and composed of water and acetonitrile. The composition of
water/acetonitrile (v/v) changed linearly from 90/10 to 10/90 in the first 8 min,
and returned within 0.1 min to 90/10, which was maintained for 1.9 min for
re-equilibration. In UPLC/ESI-tqMS analyses, the desolvation gas flow was 800
L/h; the desolvation temperature was 400 °C; and the other operation parameters
of ESI-tqMS were set as aforementioned. For a halogenated DBP detected by
PISs, its structure was proposed based on the retention time in UPLC/ESI-tqMS
multiple reaction monitoring (MRM) scan, the intensity ratio of selected MRM
mass transition, and the fragment ions in product ion scans at the corresponding
retention time. For a tentatively proposed structure of a detected DBP, the
corresponding standard compound was purchased to confirm the proposed
structure.
Developmental Toxicity Bioassay with P. dumerilii

Stock cultural conditions of P. dumerilii were maintained per the procedure
by Hutchinson’s group (21, 22), and the bioassay was conducted according to
previous studies (5, 23). Briefly, the polychaete was cultivated in the seawater
at 19 °C, and fed on a diet composed of chopped spinach, fish flakes and algae.
Breeding of the polychaete was controlled at a daily illumination regime of 16 h
light (at approximately 300 lux) and 8 h absolute darkness. The developmental
toxicity tests were conducted with the polychaete embryos. To collect the
embryos, sexually mature males and females were transferred to culturing bottles
and allowed to spawn naturally in 50 mL seawater. As fertilization was confirmed,
the developing embryos were transferred to a culturing tank containing 500 mL
seawater. After 12 h fertilization, embryos were transferred to a 1.6 cm tissue
culture testplate that contained seawater (as control) or test sample (with a total
volume of 2.0 mL). The embryos were allowed to develop for another 12 h. By
24 h post-fertilization, normally developed embryos were expected to achieve the
first larval (trochophore) stage. The numbers of normally developed and total
embryos in the seawater control or test sample were counted with the aid of an
inverted microscope, and the normal development percentage was calculated.
For either tap water without/with boiling, a preliminary test was conducted
to determine the critical concentrated range, in which the highest and lowest
concentration factors induced 100% abnormal development and a significant
amount of abnormal development (the normal development percentage at the
concentration factor was 5% less than that of the seawater control sample),
respectively (5, 23). Then, a series of test samples (with different concentration
factors within the critical concentrated range) were prepared by diluting the
toxicity test stock solution with seawater, and another bioassay was conducted.
Afterwards, a concentration factor-response curve was obtained for the water
sample, and the EC50 value (the concentration factor at which the normal
development percentage was 50% of that in the seawater control) was calculated
via regression analysis using the software SigmaPlot 12 (Systat Software Inc.,
San Jose, CA). For all the water samples, duplicate developmental toxicity tests
were conducted.
50
Results and Discussion
Reduction of TOX during Boiling
As shown in Figure 1, the concentrations of TOCl and TOBr in the two tap
waters tested substantially reduced after 5 min boiling. In tap water 1, TOCl and
TOBr were reduced by 39.0% and 43.6%, respectively; in tap water 2, TOCl
and TOBr were reduced by 51.7% and 37.5%, respectively. The previous study
(20) demonstrated that the TOX lost during boiling was caused by volatilization
to air and conversion to inorganic halides. The reduction percentages of TOCl
and TOBr by 5 min boiling in the simulated tap water (61.1% and 62.8%,
respectively) (20) were higher than those in the two real tap waters, mainly
because there was no chlorine residual in the simulated tap water. In the two real
tap water samples, the chlorine residual might keep reacting with organic matter
and bromide ions to form chlorinated and brominated DBPs during heating.
However, chlorine residual decreases quickly with the increase of temperature,
and it almost completely decays at 100 °C (8). Besides, the formed DBPs might
decompose during the subsequent 5 min boiling at 100 °C (20). Accordingly, the
“apparent” reduction of TOX after boiling was the combined result of formation,
volatilization and decomposition.
Figure 1. TOCl and TOBr concentrations in (a) tap water 1 and (b) tap water 2
without boiling and with 5 min boiling. Each datum presents the average and
standard deviations of triplicate measurements.
Decomposition of Polar Brominated and Chlorinated DBPs during Boiling

Figure 2 summarizes the ESI-tqMS PIS spectra of m/z 79 of the two
tap waters without and with 5 min boiling. For the brominated DBPs
formed in the simulated tap water (20), nine of them (seven confirmed DBPs
including bromoacetic acid, bromodichloroacetic acid, bromochloroacetic
acid, bromobutenedioic acid, dibromochloroacetic acid, dibromoacetic acid
and 3,5-dibromo-4-hydroxybenzaldehyde, and two proposed DBPs including
bromopropenoic acid and dibromomethylbutenedioic acid) were also detected in
the two real tap waters without boiling.
51
Figure 3 summarizes the ESI-tqMS PIS spectra of m/z 35 of two tap waters
without and with 5 min boiling. The simulated source water contained 2 mg/L
bromide, whereas the two real source waters contained significantly lower
concentrations of bromide (Table 1). The decrease of bromide concentration
in a source water could shift the DBPs (generated during chlorination) from
fully brominated species to bromochloro-mixed species, and further to fully
chlorinated species (14). Accordingly, the halogenated DBPs detected in the real
tap waters were somewhat different from those in the simulated tap water. First,
for dihaloacetic acids, dibromo- and bromochloro-acetic acids were detected
in the simulated and real tap waters, but dibromoacetic acid showed a higher
level in the simulated tap water than in real tap waters, and dichloroacetic acid
was detected in both real tap waters. Second, for trihaloacetic acids, tribromo-,
dibromochloro- and bromodichloro-acetic acids were detected in the simulated
tap water, while tribromoacetic acid was not detected and trichloroacetic acid
was detected in both real tap waters. Third, 2,4,6-tribromophenol, which has
been confirmed in the simulated tap water, was not detected in either real tap
water. However, ion clusters m/z 283/285/287 (with the isotopic abundance
ratio of 1:2:1) and m/z 239/241/243 (with the isotopic abundance ratio of 3:4:1)
were detected in PISs m/z 35 of both real tap waters, and they were proposed
to be 2,4,6-dibromochlorophenol and 2,4,6-bromodichlorophenol, respectively.
Fourth, ion cluster m/z 285/287 was detected in PIS m/z 79 of the simulated
tap water, and it was proposed to be dibromomethylbutenedioic acid. This ion
cluster was only detected in tap water 1, and ion clusters m/z 241/243 (with the
isotopic abundance ratio of 3:1 in PIS m/z 79) and m/z 197/199 (with the isotopic
abundance ratio of 3:1 in PIS m/z 35) were detected in both real tap waters.
These two ion clusters were proposed to be bromochloromethylbutenedioic
acid and dichloromethylbutenedioic acid. Additionally, some saturated aliphatic
DBPs (which did not form in the simulated tap water) were detected in
both real tap waters, including bromopropanoic acid, bromobutanedioic acid,
bromochlorobutanedioic acid and dichlorobutanedioic acid.
Figure 4 shows the effect of boiling on the concentrations of the halogenated
DBPs in the two real tap waters, based on the peak areas of the corresponding
ion clusters in the UPLC/ESI-tqMS MRM chromatograms. More than
half of the listed DBPs decomposed during boiling, including trihaloacetic
acids (i.e., dibromochloro-, bromodichloro-, and trichloro-acetic acids),
butenedioic acids (i.e., bromochloro-, dibromomethyl-, bromochloromethyl-,
dichloromethyl-, trichloromethyl-butenedioic acids), butanedioic acids (i.e.,
bromochloro- and dichloro-butanedioic acids), and 2,4,6-trihalophenols
(i.e., 2,4,6-dibromochlorophenol and 2,4,6-bromodichlorophenol). The
decomposition mechanisms of trihaloacetic acids, dihalobutenedioic acids,
dihalomethylbutenedioic acids, and trihalomethylbutenedioic acids included
decarboxylation and hydrolysis (20).
52
Figure 2. ESI-tqMS PIS spectra of m/z 79 of (a) tap water 1 without boiling, (b)
tap water 1 with 5 min boiling, (c) tap water 2 without boiling, and (d) tap water
2 with 5 min boiling. The y-axes are on the same scale. Above the spectra shows
the corresponding structures, of which bromoacetic acid, bromodichloroacetic
acid, bromochloroacetic acid, bromobutenedioic acid, dibromochloroacetic acid,
dibromoacetic acid and 3,5-dibromo-4-hydroxybenzaldehyde were confirmed
with the standard compounds, while others were proposed structures.
53
Figure 3. ESI-tqMS PIS spectra of m/z 35 of (a) tap water 1 without boiling, (b)
tap water 1 with 5 min boiling, (c) tap water 2 without boiling, and (d) tap water
2 with 5 min boiling. The y-axes are on the same scale. Above the spectra shows
the corresponding structures, of which dichloroacetic acid and trichloroacetic
acid were confirmed with the standard compounds, while others were proposed
structures.
54
Figure 4. Effect of boiling on the concentrations of the confirmed or proposed

halogenated DBPs in the tap waters. For each DBP, its peak area in the
UPLC/ESI-tqMS MRM chromatogram of tap water 1 without boiling was defined
as 100%, and its peak areas in the UPLC/ESI-tqMS MRM chromatograms of
other samples were expressed as the percentages of the peak area of tap water
1 without boiling.
55
Boiling also resulted in accumulation of a few DBPs, including bromoacetic
acid, dihaloacetic acids, bromobutenedioic acid, bromobutanedioic acid, and
bromopropanoic acid. For some DBPs, boiling showed different impacts
in the two real tap waters: The levels of bromo- and chloro-propenoic
acids increased in tap water 1, but decreased in tap water 2; the level of
3,5-dibromo-4-hydroxybenzaldehyde decreased in tap water 1, but increased in
tap water 2. Monohalopropenoic acids and monohalobutenedioic acids were
found to decompose during boiling in the simulated tap water (20). However, in
the real tap waters, their levels increased, possibly due to the continued reaction
between chlorine residual and the organic matter and bromide ions during heating.
It was also reported that some aromatic DBPs in the simulated tap water increased
during boiling because of the degradation of their precursors (20). Tap water
2 might also contain the precursors of 3,5-dibromo-4-hydroxybenzaldehyde,
leading to the increase of its level after boiling. Tap water 1 might not contain
the precursors of 3,5-dibromo-4-hydroxybenzaldehyde, and both tap waters 1
and 2 might not contain the precursors of 2,4,6-dibromochlorophenol and 2,4,6-
bromodichlorophenol. Besides, the decomposition of 2,4,6-dibromochlorophenol
and 2,4,6-bromodichlorophenol caused the additional formation of dibromo-,
bromochloro- and dichloro-acetic acids, following the mechanism proposed by
Zhai and Zhang (13).
Total ion intensity (TII) in the ESI-tqMS PIS spectrum of m/z 79 (or m/z 35)
can be approximately proportional to the total quantity of polar brominated (or
chlorinated) DBPs in a water sample. It was defined as the sum of ion intensities
from m/z 100 to 500 in the PIS spectrum of m/z 79 (or from m/z 50 to 400 in the
PIS spectrum of m/z 35). After boiling, for tap water 1, the TII values in the PIS
spectra of m/z 79 and m/z 35 decreased by 28.0% and 51.3%, respectively; for tap
water 2, the TII values in the PIS spectra of m/z 79 and m/z 35 decreased by 17.4%
and 36.4%, respectively.
Detoxification of Real Tap Water by Boiling
Since the levels of TOX and polar halogenated DBPs in the two tap waters
were substantially reduced by 5 min boiling, the toxicity of the boiled tap waters
was expected to be lower. Figure 5 shows the concentration factor-response
curves of the developmental toxicity (against the polychaete P. dumerilii) of the
two tap waters without and with boiling. A lower normal development percentage
indicates a higher toxicity. Through regression analysis of the concentration
factor-response curve, the EC50 value for each tap water without/with boiling
was calculated (Table 2). A lower EC50 value represents a higher developmental
toxicity potential (5, 23). Considering the EC50 values, the developmental toxicity
of tap waters 1 and 2 was decreased by 53.0% and 57.1%, respectively, after
boiling. These indicate that boiling is an effective “detoxification” method for
real tap waters.
56
Figure 5. Concentration factor-response curves of the developmental toxicity
against the polychaete P. dumerilii of (a) tap water 1 without boiling and with 5
min boiling, and (b) tap water 2 without boiling and with 5 min boiling. Each
datum presents the mean of duplicate measurements and the difference between
the mean and the measured value.
Table 2. Developmental Toxicity against the Polychaete P. dumerilii of Two

Real Tap Waters without and with 5 min Boiling (n = 2)
Tap water Concentration factor range (×-fold) EC50 (×-fold) R2 a
1, without boiling 90-250 129 0.98
1, with boiling 75-540 274 0.99
2, without boiling 50-240 123 0.98
2, with boiling 50-400 286 0.99
a The regression coefficient of the concentration factor-response curve for each water
sample.
Conclusions
The results of TOX measurements and (UPLC/)ESI-tqMS analyses obtained
in this study demonstrated that boiling for 5 min substantially reduced the
brominated and chlorinated DBPs in real tap waters. Accordingly, boiling is an
effective method for reducing human exposure to halogenated DBPs through tap
water ingestion. From the data of developmental toxicity against the polychaete
P. dumerilii, 5 min boiling significantly decreased the toxicity of tap waters.
To further investigate the effect of boiling on “detoxification” of tap water,
additional toxicity tests may be conducted with different bioassays, and relevant
epidemiological studies may also be needed.
57
Acknowledgments
This research was supported by grants from the Research Grants Council of
Hong Kong, China (projects 622412, 622913 and FSGRF12EG60). The authors
thank Jingyi Jiang for her assistance in the DOC measurement, Dave Ho for
his daily maintenance of the TOX analyzer, and Adriaan W. C. Dorresteijn (at
the Johannes Gutenberg-Universität Mainz, Germany) for providing parental P.
dumerilii.
References
1. Amy, G. L.; Siddiqui, M. S. Strategies to Control Bromate and Bromide;
AWWA Research Foundation: Denver, CO, 1999; pp 7−8.

disinfection by-products in drinking water: a review and roadmap for
research. Mutat. Res., Rev. Mutat. Res. 2007, 636, 178–242.
3. Pals, J. A.; Attene-Ramos, M. S.; Xia, M.; Wagner, E. D.; Plewa, M. J.
Human cell toxicogenomic analysis linking reactive oxygen species to the
toxicity of monohaloacetic acid drinking water disinfection byproducts.
4. Dad, A.; Jeong, C. H.; Pals, J. A.; Wagner, E. D.; Plewa, M. J. Pyruvate
remediation of cell stress and genotoxicity induced by haloacetic acid
629–637.
5. Yang, M.; Zhang, X. Comparative developmental toxicity of new aromatic
halogenated DBPs in a chlorinated saline sewage effluent to the marine
polychaete Platynereis dumerilii. Environ. Sci. Technol. 2013, 47,
10868–10876.
6. Liu, J.; Zhang, X. Comparative toxicity of new halophenolic DBPs in
chlorinated saline wastewater effluents against a marine alga: Halophenolic
DBPs are generally more toxic than haloaliphatic ones. Water Res. 2014,
65, 64–72.
7. Chowdhury, S.; Rodriguez, M. J.; Serodes, J. Model development for
predicting changes in DBP exposure concentrations during indoor handling
of tap water. Sci. Total Environ. 2010, 408, 4733–4743.
8. Zhang, A. L. Removal of chlorine residual in tap water by boiling or adding
ascorbic acid. Int. J. Eng. Res. Appl. 2013, 3, 1647–1651.
9. Carey, C. M.; Lee, H.; Trevors, J. T. Biology, persistence and detection of
Cryptosporidium parvum and Cryptosporidium hominis oocyst. Water Res.
2004, 38, 818–862.
10. Krasner, S. W.; Wright, J. M. The effect of boiling water on disinfection by-
product exposure. Water Res. 2005, 39, 855–864.
11. Wu, W.; Benjamin, M. M.; Korshin, G. V. Effects of thermal treatment on
halogenated disinfection by-products in drinking water. Water Res. 2001,
35, 3545–3550.
58
12. Zhang, X. L.; Yang, H. W.; Wang, X. M.; Fu, J.; Xie, Y. F. Formation
of disinfection by-products: Effect of temperature and kinetic modeling.
Chemosphere 2013, 90, 634–639.
13. Zhai, H.; Zhang, X. Formation and decomposition of new and unknown
polar brominated disinfection byproducts during chlorination. Environ. Sci.
Technol. 2011, 45, 2194–2201.
14. Pan, Y.; Zhang, X. Four groups of new aromatic halogenated disinfection
byproducts: Effect of bromide concentration on their formation and
speciation in chlorinated drinking water. Environ. Sci. Technol. 2013, 47,
1265–1273.
15. Zhai, H.; Zhang, X.; Zhu, X.; Liu, J.; Ji, M. Formation of brominated
disinfection byproducts during chloramination of drinking water: New polar
species and overall kinetics. Environ. Sci. Technol. 2014, 48, 2579–2588.
16. Zhai, H. Y.; Zhang, X. A new method for differentiating adducts of common
drinking water DBPs from higher molecular weight DBPs in electrospray
ionization-mass spectrometry analysis. Water Res. 2009, 43, 2093–2100.
17. Xiao, F.; Zhang, X.; Zhai, H.; Yang, M.; Lo, I. M. C. Effects of enhanced
coagulation on polar halogenated disinfection byproducts in drinking water.
Sep. Purif. Technol. 2010, 76, 26–32.
18. Xiao, F.; Zhang, X.; Zhai, H.; Lo, I. M. C.; Tipoe, G. L.; Yang, M.; Pan, Y.;
Chen, G. H. New halogenated disinfection byproducts in swimming pool
water and their permeability across skin. Environ. Sci. Technol. 2012, 46,
7112–7119.
19. Ding, G.; Zhang, X.; Yang, M.; Pan, Y. Formation of new brominated
disinfection byproducts during chlorination of saline sewage effluents.
Water Res. 2013, 47, 2710–2718.
20. Pan, Y.; Zhang, X.; Wagner, E. D.; Osiol, J.; Plewa, M. J. Boiling of
simulated tap water: effect on polar brominated disinfection byproducts,
halogen speciation, and cytotoxicity. Environ. Sci. Technol. 2014, 48,
149–156.
21. Hutchinson, T. H.; Jha, A. N.; Dixon, D. R. The polychaete, Platynereis
dumerilii (Audouin and Milne-Edwards): A new species for assessing the
hazardous potential of chemicals in the marine environment. Ecotoxicol.
Environ. Saf. 1995, 31, 271–281.
22. Jha, A. N.; Hutchinson, T. H.; Mackay, J. M.; Elliott, B. M.; Dixon, D.
R. Evaluation of the genotoxicity of municipal sewage effluent using the
marine worm Platynereis dumerilii (Polychaeta: Nereidae). Mutat Res.,
Genet. Toxicol. Environ. Mutagen. 1997, 391, 179–188.
23. Yang, M.; Zhang, X. Halopyrroles: A new group of highly toxic DBPs
formed in chlorinated saline wastewater. Environ. Sci. Technol. 2014, 48,
11846–11852.
24. APHA, AWWA, WEF. Standard Methods for the Examination of Water and
Wastewater, 22th ed.; Washington, DC, 2012.
25. Zhang, X.; Talley, J. W.; Boggess, B.; Ding, G.; Birdsell, D. Fast selective
detection of polar brominated disinfection byproducts in drinking water using
precursor ion scans. Environ. Sci. Technol. 2008, 42, 6598–6603.
59
26. Liu, J.; Zhang, X. Effect of quenching time and quenching agent dose on
total organic halogen measurement. Int. J. Environ. Anal. Chem. 2013, 93,
1146–1158.
27. Hua, G.; Reckhow, D. A. Determination of TOCl, TOBr and TOI in drinking
water by pyrolysis and off-line ion chromatography. Anal. Bioanal. Chem.
2006, 384, 495–504.
28. Li, Y.; Zhang, X.; Shang, C. Effect of reductive property of activated
carbon on total organic halogen analysis. Environ. Sci. Technol. 2010, 44,
2105–2111.
29. Gong, T.; Zhang, X. Determination of iodide, iodate and organo-iodine in
waters with a new total organic iodine measurement approach. Water Res.
2013, 47, 6660–6669.
60
Chapter 4
Bromination and Chlorination of NOM: New

Modeling Approaches and Mechanistic Insights
Paolo Roccaro,*,1 Federico G. A. Vagliasindi,1 and Gregory V. Korshin2
1Department of Civil Engineering and Architecture, University of Catania,

Viale A. Doria 6, 95125 Catania, Italy
2Department of Civil and Environmental Engineering,
University of Washington, Seattle, Washington 98195-2700

This study investigated formation of trihalomethanes (THMs),

haloacetic acids (HAAs), haloacetonitriles (HANs) generated in
two chlorinated surface waters. DBPs formation reactions and
concurrent NOM transformations were examined based on the
kinetic analysis of DBP concentrations and also via differential
spectroscopy that quantifies the extent of NOM halogenation.
The evolution of differential absorbance (ΔA) was correlated
with both DBPs yields and speciation. The modeling of DBP
formation using ΔA data showed the existence of compound-
specific proportionality coefficients that change predictably as
function of Br- concentration but do not depend on chlorine
concentration or reaction time. The presented approach was
employed to develop a DBP formation model that is based on the
kinetics of ΔA changes combined with the bromide-depended
DBP yield coefficients.
Background
Disinfection of drinking water by chlorine has dramatically reduced the
transmission of potentially fatal waterborne diseases. The main issues that
affect current chlorination practices include the limited efficiency of chlorine
against some pathogens and formation of toxic disinfection by-products (DBPs)
(1) formed as a result of reactions of chlorine and other halogen species with
natural organic matter (NOM) to produce chlorinated, brominated and, in much

smaller levels, iodinated DBPs (2–4). Even though trihalomethanes (THMs) and
haloacetic acids (HAAs) occur at the highest concentrations (up several hundred
mg/L) in chlorinated water (5) and are the most regulated DBPs worldwide (6),
more than 500 other individual DBPs have been identified (7).
DBP formation and speciation are affected by concentration and reactivity of
NOM, levels of bromide, iodine and ammonia, chlorine dose, pH, temperature and
reaction time (2, 4, 8–10). The formation of brominated DBPs is of particular
concern because these compounds are much more toxic than their chlorinated
analogues (5, 11, 12). The concentration of bromide in natural waters is usually
considerably lower than 1 mg/L (10) but in some case Br- concentrations up to 4
mg/L have been detected (13, 14).
Numerous alternative DBP formation models that use either statistical and
mechanistic approaches have been presented in prior research (e.g., (15–17)).
Due to the complexities of both NOM chemistry and halogenations processes

per se, most efforts in DBP modelling have focused on the development of
empirical/statistical models that predict DBP concentrations using numerous
alternative expressions and a set of fitting parameters applied to represent a
specific set of DBP data (18). While multi-parameter statistical fitting employed
to develop these models can be very useful for a specific set of DBP generation
conditions, they need to be recalibrated to be applicable to any water quality
dissimilar from the dataset used to develop such models. Statistical fitting per se
also does not provide any insights on DBP formation mechanisms whose inclusion
in DBP modelling allows in principle for more concise and clear approaches.
Fundamentally, mechanistic approach to DBPs formation modelling
incorporates clear assumptions that reflect the nature of NOM halogenation
(e.g., formation of various intermediates, branching kinetic pathways), equilibria
describing major aspects of the aquatic chemistry of halogen species, and a system
of differential equations that represent the consumption of halogen species and
reactive sites in NOM, generation and breakdown of halogenated intermediates
and cleavage of individual DBP species from them. A number of such models
employ the assumption that chlorine and bromine atoms are incorporated in the
NOM substrate through several sequential steps of bromine and chlorine reactions
with the reactive NOM sites (9, 19–21). In these models, relative yields of
chlorinated and brominated products originating from the intermediates formed at
each node of halogen incorporation are defined by the dimensionless ratios of the
intrinsic kinetic rates of reaction of the corresponding intermediate with chlorine
and bromine species (22, 23).
This approach can be combined with and enhanced by unambiguous
measurements of the extent of engagement of the reactive sites in NOM by
halogen species. This can be done using the principle of differential absorbance
that quantifies the consumption of NOM reactive sites in halogenation reactions.
Applications of this principle has allowed generating very strong relationships
between concentrations of a wide range of individual or classes of DBPs and, on
the other hands, intensities of differential absorbance measured at any desired
wavelength (which frequently has been at 272 nm; the differential absorbance at
this wavelength is accordingly denoted as ΔA272) (24). These correlations have
64
been shown to be independent vs. chlorine doses, reaction times and temperature
(25) but affected by pH or bromide concentration (26).
In this study we pursued the development of a combined approach that
incorporates a kinetic model to represent changes of ΔA272 values as a function
of time and bromide concentration while concentrations of individual DBPs are
obtained in this approach by incorporating DBP vs. ΔA272 relationships. This
approach was applied to the formation of three highly important classes of DBPs
(THMs, HAAs and haloacetonitriles, HANs) generated for a highly varying
reaction times and bromide levels.
Experimental
Chlorination experiments were conducted using Lake Washington (LW) water

(Seattle, USA) and Ancipa water (Sicily, Italy). Main water quality parameters of
LW and Ancipa waters are shown in Table 1.
All chemicals were ACS reagent grade or better. Solvents used in extractions
were high-purity grade. Reagent water was obtained from a Millipore Super-Q
Plus water system. Chlorine stock solution was prepared by dilution of a reagent
grade sodium hypochlorite solution (5% available chlorine) with Milli-Q water.
Table 1. Main Water Quality Parameters of Lake Washington Water and

Raw, Treated, and Fractionated Ancipa Water
Water DOC (mg/L) SUVA254 (L mg-1 m-1) Br- (mg/L)
LW 3.0 2.20 0.02
Ancipa 2.9 2.89 0.05
Chlorination was carried out with free chlorine at pH 7.0 in the presence of
0.03 mol/L phosphate buffer in headspace-free 1.6 L PTFE sampling bags, which
were used to prevent the loss of volatile DPBs when samples were taken at different
reaction times (10 minutes to 7 days). Selected experiments were carried out at
varying bromide concentrations (from its background levels to 2 mg/L) at chlorine
dose of 1.5 mg Cl2 per mg DOC. Chlorinated samples were analyzed for DBPs only
if chlorine residual was found. Requisite amount of Na2SO3 or NH4Cl were used
to quench residual chlorine. Chlorine concentrations were determined using the
DPD colorimetric method. Absorbance spectra were obtained with a Perkin-Elmer
Lambda 18 spectrophotometer using 5 cm quartz cells. All reported spectra were
normalized to the cell length of 1 cm. TOC was analyzed using an O.I. Analytical
1010 or a Shimadzu VCSH organic carbon analyzer. Concentrations of THMs,
HANs and HAAs were determined using standard analytical procedures (EPA
methods 551.1 and 552.2) and a Perkin-Elmer AutoSystem gas chromatograph
equipped with an electron capture detector. Other aspects of these analyses are
described in our previous publications (27, 28).
65
Kinetics of DBP Formation in Chlorinated Water
In agreement with the data of prior research (e.g., (2, 22, 25, 27)),
concentrations of THMs, HAAs and dihaloacetonitriles (DHANs) increased
gradually with time at given chlorine dose. Higher chlorine doses or temperatures
caused higher levels of these DBPs species (25). For instance, Figure 1 shows
the formation kinetics of CHCl3, CHCl2Br and CHClBr2 for chlorinated Ancipa
water. Similar behaviour was observed for the other investigated individual DBPs
formed in chlorinated Ancipa and LW water.
Figure 1. Kinetics of chloroform (TCM), dichlorobromomethane (DCM) and

chlorodibromomethane formation for chlorinated Ancipa water.
Modelling the Kinetics of Changes of Differential Absorbance Using

Dimensionless Ratios of Bromination/Chlorination Reaction Rates
In accord with the results of prior studies, bromide concentrations strongly

affected the yields and speciation of THMs, HAAs and DHANs (9, 19, 20, 22,
23) favouring the formation of bromine-containing DBP and faster rates of their
release at increased bromine levels.
While the presence of bromide did not appreciably affect the shape of the
differential spectra, changes of the differential absorbance vs. time were faster at
increasing bromide concentrations, especially for reaction times below four hours,
as shown in Figure 2 for the kinetic profiles measured at a background bromide
concentration (0.04 mg/L) and in the case of addition of 1 mg/L bromide. At higher
reactions time, ΔA272 values tended to plateau and reach a level common for all
examined conditions.
66
The formal description of the development of the differential absorbance vs.
time and effects of bromine on it was pursued based on the three-site model similar
to that used to model concentrations at varying bromine levels (22). Formally,
the changes of differential absorbance at 272 nm vs. time were described by the
following function:
Figure 2. Kinetic profiles of the differential absorbance of chlorinated LK water

in the presence of 0.04 and 1.0 mg/L bromide.
67
A more detailed analysis of this function was presented in our preceding
publication (22). Suffice it to mention here that values correspond to
the changes of absorbance associated with the consumption of the operationally
defined very fast, fast and slow NOM reactive sites while correspond to
the reaction rates of each site with chlorine species. Dimensionless values
quantify the relative preference of bromination over chlorination for each site.
Fitting the experimental ΔA272 vs. time profiles using the above model
showed that the very fast, fast and slow sites contribute to ca. 21%, 34% and 45%,
respectively, of the overall change of absorbance observed at highest reaction
times and bromide concentrations. The relative preferences of the sites to the
bromination pathway were similar, with the values of 5.1, 5.1 and 6.0 for
the very fast, fast and slow sites, respectively. These values are comparable with
those reported in prior research in formal descriptions of the speciation of THMs
and HAAs (9, 19, 20, 22, 23).
The use of the above approach allowed achieving precise modelling of the
changes of differential absorbance from the entire range of reactions (from 10
minutes to 3 days) and bromine concentrations (from 0.04 to 2.04 mg/L). This
allowed using this model for predicting concentrations of individual DBPs species
once ΔA272 values have been measured experimentally or calculated using the
model. This aspect of our approach is discussed in the sections that follow.
Modelling DBPs Formation and Speciation via Differential Absorbance

As was mentioned above, prior studies have established the existence of very
strong albeit non-linear correlations between differential absorbance (ΔA272) and
DBPs formation (24, 27). For individual DBP species, these correlations are
independent of chlorine dose, reaction times, and temperature and they can be
well represented by a power function (25). On the other hand, parameters of the
fitting power function depend on the pH and bromide concentration.
Following these observation, the concentrations of DBPs were modelled in
this study using ΔA272 as master parameter and the following expression:
In the above expression, DBPi stands for the concentration of any target
individual DBP compound, ki is the proportionality coefficient specific to this
species while x is the parameter of the fitting power function. As will be explained
below, the parameters x was determined to have the same value for all examined
DBPs.
Eq.2 can be useful for modelling concentrations of any target DBP (DBPi)
because it allows simplifying the number of water quality parameters that need to
be taken into account. For a constant pH, Eq.2 can be simplified as:
68
Fitting the experimental DBPs and ΔA272 values with Eq.3 at varying bromide
concentrations showed that a single value of x=1.7 can be used to fit the data for
all the investigated DBPs and waters. This is demonstrated in Figure 3-6 for data
sets showing these correlations for trichloroacetic acid (TCAA) and tribromoacetic
acid (TBAA) formed in Lk. Washington water at bromide concentrations 0.1 and
1.0 mg/L. This result allows reducing Eq. 3 to the following simple form:
Similarly to the results for Lk. Washington water, DBP concentrations found
in chlorinated Ancipa water could be well fitted by using Eq. 4, as shown in Figure
1 for the formation kinetics of CHCl3, CHCl2Br and CHClBr2.
When represented vs. the concentration of bromide, values of the ki
proportionality coefficients obtained via the fitting of the experimental DBP
and ΔA272 data exhibited a type of functional dependence similar to that of the
dimensionless speciation coefficients for THMs, HAAs and HANs reported in
prior studies (9, 19, 22, 23). This is demonstrated in Figure 7 and 8 that present the
ki coefficients for the trihalogenated (THAA) and dihalogenated HAAs (DHAA),
respectively, formed in chlorinated LW water. As a result, the proportionality
coefficients, ki, are speciation coefficients which can be determined by the
correlations between DBPs and differential absorbance, alternatively to the formal
modeling, which employs dimensionless ratios of bromination/chlorination
reaction rates.
The ki coefficients have also been found to be strongly correlated with the
logarithms of bromide concentrations (Figure 9-11). The fitting shown in these
figures was achieved using second-order polynomial curves. In the case of DBPs
containing only chlorine or bromine, the fitting predicted a gradual decrease or
increases, respectively, of the ki coefficients while for mixed compounds the
behavior of these coefficients was non-monotonic (Figure 9-11).
The use of logarithms of bromide concentration and correlating them with the
ki coefficients appears to be a reasonably straightforward procedure that is expected
to be applicable to any water quality. Further studies will demonstrate the extent
of effects of NOM site-specificity on these correlations of NOM as well as on the
absolute values of ki coefficients.
69
Figure 3. Correlations between concentration of trichloroacetic acid (TCAA)

and differential absorbance at 272 nm for chlorinated LW water. Br=0.1 mg/L.
Figure 4. Correlations between concentrations of tribromoacetic acid (TBAA)

70
Figure 5. Correlations between trichloroacetic acid (TCAA) and differential

absorbance at 272 nm for chlorinated LW water. Br=1.0 mg/L.
Figure 6. Correlations between concentrations of tribromoacetic acid (TBAA)

71
Figure 7. Effects of bromide concentration on the proportionality coefficients (ki)

for trichloroacetic (TCAA), bromodichloacetic (BDCAA), dibromochloroacetic
(DBCAA) and (TBAA) tribromoacetic acids formed in LW water.
Figure 8. Effects of bromide concentration on the proportionality coefficients

(ki) for dichloroacetic acid (DCAA), bromochloacetic acid (BCAA) and
dibromoacetic acid (DBAA) formed in chlorinated LW water.
72
Figure 9. Correlations between proportionality coefficients (ki) for dichloroacetic

(DCAA) or trichloroacetic (TCAA) acids and logarithms of bromide
concentrations. Data for chlorinated LW water.
Figure 10. Correlations between proportionality coefficients (ki) for

dibromoacetic (DBAA) or tribromoacetic (TBAA) acids and logarithms of
bromide concentration. Chlorinated LW water.
73
Figure 11. Correlations between proportionality coefficients (ki) for

bromodichloacetic (BDCAA), dibromochloroacetic (DBCAA) and
bromochloacetic acids (BCAA) and logarithms of bromide concentrations.
Chlorinated LW water.
Conclusions
This study examined the formation of THM, HAA and HAN species and
concurrent NOM transformations reactions based on the kinetic analysis of DBPs
data and also via differential spectroscopy. The approach treated the differential
absorbance (ΔAλ) as master parameter that can be used to predicts concentrations
of any selected individual DBP compounds. The kinetics of ΔAλ changes was
determined to be sensitive to the presence of bromide. It also reflected the
fact of a faster bromination of NOM reactive sites, compared with the rates
of chlorination. Comparison of ΔAλ values and concurrently measured DBP
concentrations showed that DBPs levels were correlated with ΔAλ by a uniformly
applicable power function having a 1.7 power coefficient but this function also
contains a proportionality coefficient which is compound specific and varies
with the bromide level. The proportionality coefficients were found to behave
similarly to the formal speciation coefficients reported in prior research and are
strongly correlated with logarithms of the bromide concentrations.
The joint interpretation of the kinetic DBP and differential absorbance data
confirmed the applicability of the DBPs speciation model. The results indicate that
both the absolute levels and speciation of several classes or bromine-containing
DBPs can be interpreted and modeled based on the presented approach. Further
research is needed to expand the modeling of DBPs formation to a wider range
of surface waters as well as to examine its performance for chloraminated waters
containing bromide and/or iodide.
74
Acknowledgments
This study was partially supported by the Water Research Foundation
(Project #2597), the United States EPA/Cadmus (grant 069-UW-1), and the Italian
Ministry of Instruction, University, and Research (MIUR), through the Research
Projects of National Interest - PRIN 2009 (grant 20092MES7A_002). Paolo
Roccaro and Gregory Korshin acknowledge the U.S.-Italy Fulbright Commission
for supporting their research in the U.S. and in Italy, respectively. Views expressed
in this paper do not necessarily reflect those of the funding agencies.
References
1. Sedlak, D. L.; von Gunten, U. The Chlorine Dilemma. Science 2011, 331,
42–43.
2. Reckhow, D. A.; Singer, P. C.; Malcolm, R. L.; Reckhow, D. A.; Singer, P. C.;
Malcolm, R. L. 1990. Chlorination of humic materials: byproduct formation
and chemical interpretations. Environ. Sci. Technol. 1990, 24, 1655–1664.
3. Richardson, S. D.; Thruston, A. D., Jr.; Rav-Acha, C.; Groisman, L.;
4. Hua, G.; Reckhow, D. A.; Kim, J. Effect of bromide and iodide ions on
the formation and speciation of disinfection byproducts during chlorination.
disinfection byproducts in drinking water: a review and roadmap for
research. Mutation Res. 2007, 636, 178–242.
6. Roccaro, P.; Mancini, G.; Vagliasindi, F. G. A. Water intended for human
consumption – part I: compliance with European water quality standards.
Desalination 2005, 176, 1–11.
7. Krasner, S. W.; Weinberg, H S.; Richardson, S. D.; Pastor, S. J.; Chinn, R.;
Sclimenti, M. J.; Onstad, G. D.; Thruston, A. D. Occurrence of a new
7175–7185.
8. Leenheer, J. A.; Crouè, J.-P. Characterizing dissolved aquatic organic
matter, comprehensive approach to preparative isolation and fractionation
of dissolved organic carbon from natural waters and waste waters. Environ.
Sci. Technol. 2003, 37, 18A–26A.
9. Cowman, G. A.; Singer, P. C. Effect of bromide ion on haloacetic acid
speciation resulting from chlorination and chloramination of aquatic humic
substances. Environ. Sci. Technol. 1996, 30, 16–24.
10. Obolensky, A.; Singer, P. C. Halogen substitution patterns among
disinfection byproducts in the information collection rule database. Environ.
Sci. Technol. 2005, 39, 2719–2730.
75
11. Zhang, L.; Xu, L.; Zeng, Q.; Zhang, S.-H.; Xie, H.; Liu, A.-L.; Lu, W.-
Q. Comparison of DNA damage in human-derived hepatoma line (HepG2)
exposed to the fifteen drinking water disinfection byproducts using the single
cell gel electrophoresis assay. Mutat. Res. 2012, 741, 89–94.
12. Yang, M.; Zhang, X. Comparative developmental toxicity of new aromatic
halogenated dbps in a chlorinated saline sewage effluent to the marine
polychaete Platynereis dumerilii. Environ. Sci. Technol. 2013, 47,
10868–10876.
13. Heller-Grossman, L.; Idin, A.; Relis, B. L.; Rebhun, M. Formation
of cyanogen bromide and other volatile DBPs in the disinfection of
bromide-rich lake water. Environ. Sci. Technol. 1999, 33, 932–937.
14. Magazinovic, R. S.; Nicholson, B. C.; Mulcahy, D. E.; Davey, D. E. Bromide
levels in natural waters: its relationship to levels of both chloride and total
dissolved solids and the implications for water treatment. Chemosphere

2004, 57, 329–335.
15. Francis, R. A.; Vanbriesen, J. M.; Small, M. J. Bayesian statistical modeling
of disinfection byproduct (DBP) bromine incorporation in the ICR database.
16. Chowdhury, S.; Champagne, P.; McLellan, P. J. Models for predicting
disinfection byproduct (DBP) formation in drinking waters: a chronological
review. Sci. Total Environ. 2009, 407, 4189–206.
17. Heeb, M. B.; Criquet, J.; Zimmermann-Steffens, S. G.; von Gunten, U.
Oxidative treatment of bromide-containing waters: formation of bromine
and its reactions with inorganic and organic compounds - a critical review.
Water Res. 2014, 48, 15–42.
18. Sadiq, R.; Rodriguez, M. Disinfection byproducts (DBPs) in drinking water
and predictive models for their occurrences: a review. Sci. Total Environ.
2004, 321, 21–46.
19. Nokes, C. J.; Fenton, E.; Randall, C. J. Modeling the formation of brominated
trihalomethanes in chlorinated drinking waters. Water Res. 1999, 33,
3557–3568.
20. Chang, E. E.; Lin, Y. P.; Chiang, P. C. Effects of bromide on the formation of
THMs and HAAs. Chemosphere 2001, 43, 1029–1034.
byproducts: effect of bromide concentration on their formation and
1265–1273.
22. Roccaro, P.; Chang, H.-S.; Vagliasindi, F. G. A.; Korshin, G. V. Modeling
bromide effects on yields and speciation of dihaloacetonitriles formed in
chlorinated drinking water. Water Res. 2013, 47, 5995–6006.
23. Roccaro, P.; Korshin, G. V.; Cook, D.; Chow, C. W. K.; Drikas, M. Effects
of pH on the speciation coefficients in models of bromide influence on the
formation of trihalomethanes and haloacetic acids. Water Res. 2014, 62,
117–126.
24. Korshin, G. V.; Li, C.-W.; Benjamin, M. M. The decrease of UV absorbance
as an indicator of TOX formation. Water Res. 1997, 31, 946–949.
76
25. Roccaro, P.; Chang, H.-S.; Vagliasindi, F. G. A.; Korshin, G. V. Differential
absorbance study of effects of temperature on chlorine consumption and
formation of disinfection byproducts in chlorinated water. Water Res. 2008,
42, 1879–1888.
26. Roccaro, P.; Vagliasindi, F. G. A. Differential versus absolute UV absorbance
approaches in studying NOM reactivity in DBPs formation: comparison and
applicability. Water Res. 2009, 43, 744–750.
27. Korshin, G. V.; Wu, W. W.; Benjamin, M. M.; Hemingway, O. Correlations
between differential absorbance and the formation of individual DBP species.
Water Res. 2002, 36, 3273–3282.
28. Roccaro, P.; Vagliasindi, F. G. A.; Korshin, G. V. Quantifying the formation
of nitrogen-containing disinfection by-products in chlorinated water using
absorbance and fluorescence indexes. Water Sci. Technol. 2011, 63, 40–44.
77
Chapter 5
Modeling NDMA Formation

Kinetics During Chloramination of Model
Compounds and Surface Waters Impacted
by Wastewater Discharges
Jinwei Zhang,1 David Hanigan,2 Ruqiao Shen,3 Susan Andrews,3

Pierre Herckes,1 and Paul Westerhoff*,2
1Department of Chemistry and Biochemistry, Arizona State University,
Tempe, Arizona 85287-1604, United States
2School of Sustainable Engineering and the Built Environment,
Arizona State University, Tempe, Arizona 85287-3005, United States

3Department of Civil Engineering, University of Toronto,
Toronto, Ontario M5S 1A4, Canada

*E-mail: [email protected]. Telephone: 1-480-965-2885.
N-nitrosodimethylamine (NDMA) is a chloramination

disinfection by-product (DBP). Despite extensive research on
NDMA occurrence and evaluation of formation potential from
select precursors and in different matrices, less information
exists on the kinetics of NDMA formation. A kinetic model
is developed and parameterizered using the limited amount of
previously published kinetic data involving addition of model
compounds to pH buffered ultrapure water and/or surface
waters containing organic matter. A simple second-order kinetic
model was developed using a NDMA precursor concentration,
based upon a measured maximum NDMA concentration as the
initial precursor concentration, and measured monochloramine
concentrations. The second order model was able to fit NDMA
formation kinetics well by determining the best-fit apparent
rate constant (kapp), suggesting the model represents a common
rate limiting step and/or an intermediate product. An associated
reaction may exist during the transformation of a wide variety of
precursors into NDMA. Values for kapp ranged by approximately
one-order of magnitude and were lower in the presence of

natural organic matter (NOM), which may suggest that more
reactive chloramine species or NDMA intermediate products
are scavenged by NOM. Overall, the modeling suggests lines of
future research to help elucidate and predict NDMA formation
mechanisms such that strategies to control its formation could
be applied during drinking water treatment.
Introduction
Over the past decades, N-nitrosamines (NAs), which form as disinfection
by-products during chlorination, in particular chloramination, have emerged as
a great concern because of an increasing use of chloramines by water utilities for

residual disinfection to meet trihalomethane (THM) and haloacetic acid (HAA)
regulations (1) . NAs were included in the Unregulated Contaminant Monitoring
Rule 2 (UCMR2) and then listed on the third Contaminant Candidate List (CCL3)
(2). The USEPA classifies NAs as probable human carcinogens in drinking water
at low ng/L levels associated with a 10-6 lifetime cancer risk. California’s Office
of Environmental Health Hazard Assessment (OEHHA) set a notification level of
10ng/L for NDMA (3). Because of their high carcinogenic potential, the USEPA
may soon make a regulatory determination for NAs.
N-nitrosodimethylamine (NDMA) is the most commonly investigated
nitrosamine. Most studies have found that NDMA is rather associated with
chloramination than with chlorination (4–6). Systems with high plant effluent
NDMA (i.e., >50 ng/L) typically use chloramines as the primary rather than
secondary disinfectant (6), reflecting the potential for precursor deactivation by
strong pre-oxidants such as chlorine. Due to the long time-scale for nitrosamine
formation, plants with long in-plant chloramine contact times (e.g., 12–18 hr)
tend to have higher NDMA concentrations in the plant effluent than those with
short (e.g., 0.5–2 hr) contact times (7). NDMA concentrations tend to increase
throughout chloraminated distribution systems (7–11).
Primary amines can be nitrosated to yield primary NAs, however the latter
are unstable and decay nearly instantaneously (12). Secondary amines can
be transformed to their stable NA analogues and most research focused on
NDMA formation from dimethylamine (DMA) (13–17). However, some studies
indicated that the presence of DMA was not sufficient to explain the NDMA
formed in surface water and wastewaters (18, 19). Tertiary amines may also
serve as nitrosamine precursors. For example, trimethylamine (TMA) would
decay instantly upon chlorination or chloramination to release DMA and then
quantitatively form NDMA (20). Mechanistic studies indicate that yields of
NDMA from chloramination of most secondary and tertiary amines are very low
(~0–3%) but recently a subset of tertiary amines were found to have substantially
higher NDMA yields (up to 90%) compared to previously studied precursors
(21–23). In particular, ranitidine, a widely found amine-based pharmaceutical,
forms NDMA at yields higher than 80% (21, 23). It was suggested that the electron
donating group (furan ring) in ranitidine favors the reaction with chloramine
80
leading to higher NDMA yields. The differences of NDMA formation yields
among tertiary amines revealed the importance of tertiary amine structures (N
bond leaving group) on the NDMA formation. The higher yields also suggest that
tertiary amines can form nitrosamines without proceeding through a secondary
amine intermediate (24).
Figure 1. NDMA formation pathways as proposed in the literature: (1) Choi

and Valentine, 2002; Mitch and Sedlak, 2002. (2) Schreiber and Mitch, 2006.
(3) Selbes et al. 2013.
Several NDMA formation mechanisms during chloramination had been

proposed (Figure 1). Mechanistic studies using DMA as the model precursor
found that unprotonated DMA undergoes nucleophilic substitution with
monochloramine, yielding the unsymmetrical dimethylhydrazine (UDMH)
intermediate. This UDMH is then oxidized by monochloramine to NDMA (13,
25). The importance of chloramine speciation and presence of dissolved oxygen
was later discovered and a second mechanism proposed ((2) in Figure 1). Here,
dichloramine reacts to yield NDMA via the formation of the chlorinated UDMH
intermediate (UDMH-Cl) and this UDMH-Cl is then oxidized by oxygen to
produce NDMA (15). Based upon competition reaction kinetics, it has been
suggested that the monochloramine mechanism (1) is negligible compared to
dichloramine (2) pathway. As the molar yield of NDMA from DMA is low
(i.e., <5%) compared to that from other compounds (e.g., ranitidine), it was
suspected that a third pathway, other than through DMA, exists. Compounds such
81
as ranitidine may follow a different series of reactions involving nucleophilic
attack of the amine group on chloramines ((3) in figure 1) (26). Through this
pathway, NDMA formed from ranitidine was more sensitive to monochloramine
(21). Suspected NDMA precursors with electron withdrawing groups may
preferentially react with monochloramine while compounds with electron
donating groups react preferentially with dichloramine (24, 27). Many suspected
precursors could react with both monochloramine and dichloramine. Therefore,
NDMA formation is likely a combination of reactions between NDMA precursors
and both chloramine species.
Despite extensive research on yields of NDMA from model compounds and
in surface water or wastewater (19, 20, 23), less information exists related to
NDMA formation kinetics in the latter water sources. NDMA formation from
DMA has been initially modeled through UDMH and Cl-UDMH pathways
((1) and (2) in figure 1) and the latter successfully predicted NDMA formation
over a range of conditions (15). A statistical model was proposed to predict
NDMA formation kinetics from pharmaceutical compounds such as ranitidine
in various water matrices (23). In addition, natural organic matter (NOM) was
also investigated as a source of precursors in a NDMA formation model (28).
No attempt has been ever made to model NDMA formation in wastewater or
wastewater-impacted waters. The challenge for a model development lies in the
multitude of precursors and precursor types present and hence the possibility of
having various mechanisms, with distinct kinetics occurring simultaneously.
The aim of this work is to formulate and evaluate a NDMA formation kinetics
model that is applicable to different water sources and model compounds. A
simple second-order kinetic reaction model was developed for NDMA formation.
In this model, the precursor concentration, quantified as the maximum NDMA
concentration formed during chloramination under specific water conditions, and
chloramine decay were used to predict transformation of precursors to NDMA.
The model was parameterized using NDMA formation data from literature data
on NOM and model precursor compounds (i.e. ranitidine). The optimization of
the kinetic model parameters and the resulting model performance are discussed.
Model Description
NDMA Formation Model
As detailed in the introduction, three different pathways for NDMA formation
have been proposed (Figure 1). The original concept of the model was that
different types of precursors may proceed along different mechanistic pathways
with different chloramine species and the reactions may potentially involve
oxygen to produce NDMA. Such a model would classify NDMA precursors
(nM quantities) as having either higher or lower yields and presumably different
reaction rates with different chloramine species. In a first step, using data obtained
from our research of NDMA formation in wastewater, we developed a simple
second-order model for the reaction of NDMA precursors (P) in the presence of
monochloramine (NH2Cl):
82
where [NH2Cl], [NHCl2] = monochloramine or dichloramine concentration at time
t
[P] = NDMA precursor concentration at time t
kmono, kdi = rate constant of NDMA formation from monochloramine or
dichloramine
kapp = second order rate constant of NDMA formation in our model
α = ratio of dichloramine and monochloramine
Table 1 shows the decomposition reactions for chloramines in water

(29). The relationship between monochloramine and dichloramine derives
from reactions 3, 4 and 5. Dichloramine forms via an acid catalyzed reaction
from two monochloramines, then dichloramine reacts instantaneously with
monochloramine, producing nitrogen gas. In this assumption, dichloramine
occurs at a low concentration and at steady state. Thus, the dichloramine
concentrations can be represented as a proportion (α) of monochloramine.
Equation 1 shows the formation kinetics of NDMA as a combination of both
chloramines reactions. kmono and kdi were the rate constants for mono- and
di-chloramine reactions respectively. A new apparent rate constant kapp was
then used in Equation 1c as the best fit second-order rate constant (M-1s-1) in
terms of monochloramine and NDMA precursor. Measured monochloramine was
used in the model to represent all chloramines since it is the dominant source of
chloramines in neutral and basic conditions, and is easier to measure compared
to dichloramine.
Equation 1c simplifies NDMA formation kinetics in different water matrices
despite there being various reactions pathways of chloramines and probably
multiple groups of precursor species with highly variable yields (0.017% for
TMA, 90% for ranitidine) and influences of various water quality parameters
on chloramine decay in water. Currently there is no method to predict NDMA
formation kinetics, and most research has relied upon simulated distributed
system (SDS) tests (NH2Cl~2.5mg/L)) or formation potential (FP) tests
(NH2Cl>100mg/L). Even model compounds such as DMA showed variable
NDMA yields (21, 30). Organic compound concentrations were thought to be
associated with NDMA formation from NOM. However, there was no correlation
between NDMA formation (or precursor concentration) and UV254 or DOC in
natural waters or in wastewaters (31). P0 is the maximum NDMA formation
which is measurable in NDMA kinetic tests. Together with the measurement
of monochloramine, the model is designed to simulate the NDMA formation
processes. The concentration of NDMA at any time ([NDMA(t)]) would be
83
related to the maximum amount of NDMA precursors available for reaction in an
experiment ( [P]0 = [NDMAmax] ) as follows:
where [P] NDMA precursor concentration at time t

The measured concentration of NH2Cl or an estimated simulated
monochloramine decomposition is used in the model (28, 32). NDMA formation
data at different time points were fitted to the model using Kintecus (33) and the
rate constant, kapp, was optimized with the highest correlation coefficient between
experimental values and model fitting.
Table 1. Chloramine Decomposition Kinetics and Associated Rate Constants

(29)
Reaction Rate Constant
1 HOCl + NH3→ NH2Cl + H2O k1= 1.0×1010M-1h-1
2 NH2Cl + H2O → HOCl + NH3 k2=0.1h-1
3 HOCl + NH2Cl → NHCl2 + H2O k3=1.26×106 M-1h-1
4 NHCl2 + H2O→HOCl + NH2Cl k4=2.3×10-3 h-1
5 NH2Cl + NH2Cl → NHCl2 + NH3 kd*
6 NHCl2 + NH3 → NH2Cl + NH2Cl k6=2.0×108 M-1h-1
7 NH2Cl + NHCl2 →N2 + 3H+ + 3Cl- k7=55.0 M-1h-1
8 NHCl2 + H2O → NOH + 2HCl k8=6.0×105 M-1h-1
9 NOH + NHCl2 → N2 + H2O + HCl k9=1.0×108 M-1h-1
10 NOH + NH2Cl → N2 + H2O +HCl k10=3.0×107 M-1h-1
11 NH4+ → NH3 + H+ pKa=9.3
12 H2CO3 →HCO3- + H+ pKa=6.3
13 HCO3- → CO32- + H+ pKa=10.3
*kd= kH[H+] +kH2CO3[H2CO3] + KHCO3[HCO3-]; kH=2.5×107 M-2h-1; kHCO3= 800 M-2h-1;
kH2CO3=40000 M-2h-1.
Monochloramine Degradation
Monochloramine decays over time due to hydrolysis and reaction with

organic and inorganic species (Table 1). Autodecomposition or hydrolysis
forms other oxidants such as free chlorine or dichloramine, which would react
with DOC. Another pathway is, in the presence of NOM, monochloramine
reacting directly with organic matter. The rate of monochloramine loss in both
84
pathways was found to be dependent on experimental conditions such as initial
monochloramine concentration, DOC or pH (34). It was difficult to accurately
simulate monochloramine decomposition with detailed models published in
previous research because various parameters (i.e. fast and slow reactive fractions
of NOM) need to be reoptimized since they are specific to reaction conditions.
A simple first order degradation model with respect to monochloramine is
used to simulate the decay of monochloramine in the presence of NDMA
precursors (Equation 3). The rate constant is specific to experiment conditions
and representative for similar water qualities in general.

Modeling of NDMA Formation in NOM
Chen et al. modeled NDMA formation from reactions between NOM and
monochloramine (28). They suggested that NOM in surface waters could be
oxidized involving two types of reactive sites of monochloramine loss. A portion
of the dissolved organic carbon (DOC) could react with monochloramine rapidly
within hours and the other part had a much slower reaction with free chlorine
(over days). However, their model had five parameters (i.e. fast and slow reactive
fractions of NOM) to be optimized which would vary between water sources.
Here we fitted their experimental data with our model. The monochloramine
decay was modeled as a first order reaction in Equation 2 and the monochloramine
decay rate constants are shown in Table 2. The decay rate of monochloramine
increased when pH was reduced from 9 to 7 since at low pH monochloramine
would decay to dichloramine more easily (Reaction 5, Table 1).
Figure 2. Model prediction of NDMA formation and monochloramine decay

at various pH in surface water. (Symbols: observation data, lines: model
predictions. Data from Chen et al. 2006 (28))
85
Then the NDMA formation data was fitted with our second order model
(Figure 2). NDMA concentration at 120 h was used as NDMAmax (P0) although
NDMA formation had not reached its maximum at this time. The model fits
NDMA formation over time well at pH 7, 8 or 9. However, at pH 6 when
dichloramine is more prevalent chloramine species, the model was not able to
simulate the NDMA formation because our assumption that dichloramine is
proportional to monochloramine is not valid at pH 6. NDMA formation yields
appear to increase with decreasing pH from 9 to 7 due to presence of more
dichloramine at lower pH. The optimized rate constant kapp varied within a factor
of three for different pH conditions, however, no obvious correlation was found
between pH values and kapp.
Table 2. Optimized NDMA Formation Rate Constant and Monochloramine

Decomposition Rate Constants in NOM under Various Reaction Conditions.
(Data from Chen et al. 2006 (28)). R2 Is Correlation Coefficient between
Model and Observation. Notes: Experiments Were Conducted at pH 7 with
Variable Cl/N ratios (and Ammonia Concentrations): a 0.7 (0.07 mM NH3); b
0.3 (0.17 mM NH3); c 0.10 (0.5 mM NH3)
pH kapp M-1s-1 (R2) kdecay (h-1)
9 0.25 (0.96) 3.4E-03
8 0.09 (0.96) 4.0E-03
7 0.15a (0.95) 9.7E-03
0.12b (0.95) 4.1E-03
0.10c (0.84) 2.5E-03
Figure 3. Model prediction of NDMA formation and monochloramine decay

at various ammonia concentrations (Cl/N ratio) in surface water. (Symbols:
observation data, lines: model predictions, pH=7. Data from Chen et al. 2006
(28))
86
The NDMA formation data in surface waters with various Cl/N ratios in
Chen’s work was also fitted with our model. Both monochloramine decay
and NDMA formation were well predicted (Figure 3). Results showed that a
decreasing Cl/N ratio reduced monochloramine decay rates, NDMA formation
rates and NDMA yields. This was the result of dichloramine formation being
favored at high Cl/N ratios (low ammonia concentrations) according to reaction
3 in Table 1 With the monochloramine degradation measurements shown in
Chen’s work (28), the model successfully predicted NDMA formation over time
in surface waters from pH 7 to 9 and in conditions of Cl/N ratio from 0.1 to 0.7.
Modeling of NDMA Formation from Model Compounds

In addition to NOM, model compounds such as DMA and ranitidine were

also studied in NDMA formation kinetics experiments. Selbes reported NDMA
formation from various amine precursors (35). However, NDMA formation
kinetics were not discussed. Here we apply our model to the NDMA kinetics
data of five precursor compounds from Selbes’ work (35). DMA and four
tertiary amines (trimethyalmine (TMA), dimethylisopropylamine (DMiPA),
dimethylbenzylamine (DMBzA) and ranitidine) were selected. These five
precursors were chloraminated by Selbes in three reaction conditions: in distilled
deionized water at pH 7.5, a formation potential (FP) test condition with
1.4mM monochloramine, and simulated distribution system (SDS) with 0.04mM
monochloramine, and SDS condition with 100 mg/L ammonia to minimize
dichloramine formation according to Reaction 3 in Table 1.
Monochloramine decay data was not published in Selbes’ work. The
monochloramine model simulation in NOM in Chen’s work showed that pH has
a more significant effect on monochloramine decay than DOC concentrations
or Cl/N ratio. A monochloramine decay rate of 0.007 h-1, between the rate
constants at pH 7 (0.01 h-1) and pH 8 (0.004 h-1) in Chen’s work, was used
to predict monochloramine decomposition in the deionized water. Figure 4
shows NDMA formation data obtained from Selbes’ work and model fitting for
precursors under various conditions and the resulting rate constants are shown
in the figure and listed in Table 3. The model successfully fitted the NDMA
formation kinetics for all five chemicals in most tests. However, in SDS condition
with excess ammonia, our model overestimates NDMA formation in the first
half reaction time and underestimates NDMA in the second half period for
ranitidine and DMBzA. It was the same with DMiPA in the SDS condition test.
In previous studies it was found that ranitidine and DMBzA were more sensitive
to monochloramine and DMiPA was only sensitive to dichloramine (27). In
NDMA formation from tertiary amines such as ranitidine, an intermediate could
form by nucleophilic substitution between ranitidine and monochloramine (26,
27). A possible explanation could be, in conditions with high monochloramine
concentrations in FP tests, this intermediate degrades quickly to form NDMA.
Therefore, the presence of intermediates could ‘delay’ the NDMA formation and
make our model overestimate NDMA formation.
87
88
Figure 4. NDMA formation from model precursor compound data (Symbols),

model fitting (Lines) in FP tests (NH2Cl 1.4mM, pH=7.5), SDS conditions
(NH2Cl 0.04mM, pH=7.5) and SDS conditions with excess ammonia. (NDMA
data from Selbes, 2014 (35))
Although kapp× [NH2Cl] was greater in the FP tests, indicating a faster NDMA
formation in FP conditions than in SDS conditions, the optimized rate constants
(kapp) in the FP tests were less than under SDS conditions, by an order of magnitude
for model compounds. Compared to amine precursor concentrations (200nM) in
Selbes’ work, the monochloramine concentrations (1.4mM) were more than 5000
times the precursor concentrations in the FP test. Thus, in the FP test, there is a
substantial excess of monochloramine present that reduced the rate constant. Rate
constants kapp of ranitidine, DMBzA and DMA are similar and that of DMiPA
is much smaller. The similar kapp between ranitidine and DMA was unexpected
89
since DMA and ranitidine have very different yields (<3% and 80% respectively)
and different proposed NDMA formation mechanisms. In SDS conditions with
excess ammonia, optimized kapp for both ranitidine and DMBzA are less than those
under SDS conditions, suggesting dichloramine reactions may be important. The
dichloramine reaction rate constant, which is the difference between kapp between
two conditions, was found to be greater than the monochloramine rate constant.
Table 3. Optimized Rate Constant kapp for Model Compounds under Various
Reaction Conditions. (NDMA data from Selbes, 2014 (35))
Compounds kapp M-1s-1 (R2)
NDMA Yields FP test SDS SDS+ excess

in FP test % ammonia
Ranitidine ~80 0.09 (0.99) 1.2 (0.92) 0.18 (0.97)
DMBzA ~80 - 0.71 (0.92) 0.12 (0.87)
DMA 1.1 0.06 (0.98) 0.77 (0.96) -
TMA 1.7 0.02 (0.97) - -
DMiPA 83 0.01 (0.98) 0.1 (0.96) -
Modeling of NDMA Formation of Pharmaceutical Compounds in the

Presence of NOM
Our model was also applied to NDMA formation by amine precursors in

the presence of NOM in natural waters. Shen et al. investigated the NDMA
formation kinetics of pharmaceutical compounds in three water matrices (MQ
water: TOC=0mg/L, Lake: TOC=2mg/L, River: TOC=6mg/L) under simulated
distribution system condition (SDS, NH2Cl=0.035mM or 0.07mM) (23). Four
pharmaceutical compounds, sumatriptan, chlorphenamine, doxylamine and
ranitidine were used in the kinetic experiments. Measured monochloramine decay
concentrations over time in Shen’s work were used to model monochloramine
decomposition. Model fitting and optimized rate constants are shown in Figure
5 and Table 4. The model in equation 1c simulates NDMA formation from
ranitidine in MQ water with a high correlation coefficient (R2>0.93) and yields
an optimized rate constant of 1.3 M-1s-1. This is similar to the rate constant (1.2
M-1s-1) in Selbes’ work under similar reaction conditions (DDW water, SDS,
NH2Cl=0.04mM) and demonstrates that our model is applicable to predict NDMA
formation from ranitidine and monochloramine under similar reaction conditions.
However, for all four pharmaceutical precursors, the model overestimates the
NDMA formation in the first half experiment time and underestimated NDMA in
the second half. This is possibly due to the formation of an intermediate that delays
the NDMA formation, as suggested previously.
90
Figure 5. NDMA formation from pharmaceutical compounds under SDS

condition (MQ water, NH2Cl=0.035mM, pH=7), data (symbols) and model fitting
(lines). (NDMA data from Shen et al., 2011 (23))
Table 4. Optimized Rate Constant kapp for Model Compounds in Different

Water Matrices with Varied TOC. (NDMA Data from Shen et al. 2011 (23))
Compounds NDMA kapp M-1s-1 (R2)
Yields in
SDS MQ % SDS 2mg/L TOC 6mg/L TOC
ranitidine ~90 1.3 (0.93) 0.82 0.25

sumatriptan ~2 0.28 (0.86) 0.24 0.08
chlorphenamine 3 0.73 (0.90) - 0.22
doxylamine 10 0.28 (0.84) - 0.13
In Shen’s work on NDMA formation from pharmaceutical compounds in

natural waters, there was an initial lag-time when NDMA formation did not
increase for all four pharmaceutical compounds in lake and river waters. The
lag-time was longer at higher TOC concentrations. However, the lag-time could
not be simulated with our proposed model. In Shen’s work it was suggested
that there was NOM-pharmaceutical binding that inhibited the initial NDMA
91
formation (23). In addition, dichloramine has a lower electron density on nitrogen
due to two chlorine atoms so it would be a more preferable species to interact
with negatively charged NOM. Thus in the presence of NOM, dichloramine could
more easily react with NOM compared to monochloramine and hence NDMA
formation from dichloramine would be suppressed by NOM.
To better understand how NOM affects NDMA formation rates, our proposed
model was adjusted to fit the kinetics data. As there is no NDMA formed during
the lag-time, we start the model simulation at the end of the NDMA formation
lag. The lag time was removed and the kinetics data was refitted with our model.
The model fits the experimental data better for all four pharmaceutical compounds
(Figure 6) and the optimized rate constants were listed in Table 3. The optimized
second-order rate constants kapp were found to decrease with increasing TOC
concentrations. This can be explained by a competition between NOM and model
compounds to react with mono- and/or dichloramine. In the river water samples
(TOC=6mg/L), the optimized rate constants kapp for the pharmaceutical model
compounds were in the range of 0.08 to 0.25M-1s-1 (Table 4) and were comparable
to the rate constants (0.02 to 0.09 M-1s-1) of NDMA formation from NOM in
surface waters with similar TOC concentration (TOC=3.4mg/L) in Chen’s work
we discussed previously. By using the lag model, we were able to compare the
rate constant of NDMA formation at different DOC concentrations and reveal
how the DOC impacted the NDMA formation kinetics.
Figure 6. NDMA formation of amine precursors in river water (SDS, pH=7,

TOC=6mg/L) and modeling of NDMA formation without lag-time, data (symbols)
and model fitting (lines). (NDMA data from Shen et al., 2011 (23))
92
Summary and Conclusions
A model was developed and parameterized based on NDMA formation
trends observed in the literature. Results show that a second-order kinetic rate
model simulates NDMA formation from NOM and model precursors over a
range of reaction conditions. A narrow range of rate constants were obtained and
appear to indicate a similar rate-limiting reaction step among different precursors
and water sources. Though the model was not intended to reveal the relative
importance of mono- or dichloramine reactions in NDMA formation kinetics,
it is applicable to describe how water conditions such as DOC and pH affect
NDMA formation kinetics. In real water utilities, the model could be used to
predict NDMA formation by measuring the NDMAmax and monochloramine.
The proposed model needs to be further evaluated using a wider range of
reaction conditions. For example, the importance of dissolved oxygen needs be

investigated since it was found to be an important factor in NDMA formation
from different amines such as DMA and ranitidine. NDMA kinetics test could be
performed in other water matrices such as wastewater effluents. In addition, further
detailed characterization of more precursors and their kinetics studies could help
better understand NDMA formation in different water sources and a more precise
model could be developed.
Acknowledgments
This research was supported by the Water Research Foundation (Project 4499,
managed by Djanette Khiari), American Water Works Association Abel Wolman
Fellowship, Water Environment Federation Canham Scholarship, and the Arizona
State University Fulton School of Engineering Dean’s Fellowship.
References
1. Krasner, S. W.; Mitch, W.; McCurry, D. L.; Westerhoff, P.; Hanigan, D.
Formation, Precursors, Control, and Occurrence of Nitrosamines in Drinking
Water: A Review. Water Res. 2013, 47, 4433–4450.
2. EPA, U. Contaminant Candidate List 3 (CCL3); 2009, http://water.epa.gov/
scitech/drinkingwater/dws/ccl/ccl3.cfm (accessed September 26th, 2014).
3. OEHHA. Public Health Goal for Chemicals in Drinking Water:
N-Nitrosodimethylamine; California Environmental Protection Agency:
Sacramento, CA, 2006.
4. Mitch, W.; Kraner, S.; Westerhoff, P.; Dotson, A. Occurence and Formation
of Nitrogenous Disinfection By-products. Water Research Foundation
Report; Water Research Foundation: Denver, CO, 2009.
5. Boyd, J. M.; Hrudey, S. E.; Richardson, S. D.; Li, X. F. Solid-phase extraction
and high-performance liquid chromatography mass spectrometry analysis of
nitrosamines in treated drinking water and wastewater. Trends Anal. Chem.
2011, 30 (9), 1410–1421.
93
6. Russell, C. G.; Blute, N. K.; Via, S.; Wu, X.; Chowdhury, Z.; More, R.
Nationwide assessment of nitrosamine occurrence and trends. J. AWWA
2012, 104 (3), 57–58.
7. Krasner, S. W.; Westerhoff, P.; Mitch, W. A.; Skadsen, J. Case Studies on
Nitrosamine Formation and Control at Full-Scale Drinking Water Treatment
Plants. Presented at 2012 AWWA Wat. Qual. Technol. Conf., Toronto,
Ontario, 2012.
8. Valentine, R. L.; Choi, J.; Chen, Z.; Barrett, S. E.; Hwang, C. J.; Guo, Y.;
Wehner, M.; Fitzsimmons, S.; Andrews, S. A.; Werker, A. G.; Brubacher,
C.; Kohut, K. D., Factors affecting the formation of NDMA in water and
occurrence. AWWA Research Foundation: Denver, CO, 2005.
9. Krasner, S. W.; Westerhoff, P.; Chen, B.; Rittmann, B. E.; Amy, G.
Occurrence of disinfection byproducts in united states wastewater treatment
plant effluents. Environ. Sci. Technol. 2009, 43 (21), 8320–8325.

10. Liew, D.; Linge, K. L.; Joll, C.; Heitz, A.; Trolio, R.; Breckler, L.;
Henderson, R.; Charrois, J. W. A.; Formation of Nitrogenous Disinfection
By-Products (N-DBPs) in Raw and Treated Drinking Water. In Ozwater ’11,
May 9−11 2011, Australian Water Association: Adelaide,2011.
11. Krasner, S. W.; Sclimenti, M. J.; Lee, C. F. T. Occurrence and Formation of
NDMA in Consecutive Systems in a Major Metropolitan Area. Proc. 2012
AWWA Water Quality Technology Conference; AWWA: Denver, CO, 2012.
12. Ridd, J. H. Nitrosation, Diazotisation, and Deamination. Q. Rev. Chem. Soc.
1961, 15 (4), 418–441.
13. Mitch, W. A.; Sedlak, D. L. Formation of N-nitrosodimethylamine (NDMA)
from dimethylamine during chlorination. Environ. Sci. Technol. 2002, 36
(4), 588–595.
14. Choi, J.; Valentine, R. L. Formation of N-nitrosodimethylamine (NDMA)
from reaction of monochloramine: a new disinfection by-product. Water Res.
2002, 36 (4), 817–824.
15. Schreiber, I. M.; Mitch, W. A. Nitrosamine formation pathway revisited: The
importance of chloramine speciation and dissolved oxygen. Environ. Sci.
Technol. 2006, 40 (19), 6007–6014.
16. Shah, A.; Mitch, W. A. Halonitroalkanes, halonitriles, haloamides and N-
nitrosamines: A critical review of nitrogenous disinfection byproduct (N-
DBP) formation pathways. Environ. Sci. Technol. 2012, 46 (1), 119–131.
17. Schreiber, I. M.; Mitch, W. A. Occurrence and fate of nitrosamines and
nitrosamine precursors in wastewater-impacted surface waters using boron
as a conservative tracer. Environ. Sci. Technol. 2006, 40 (10), 3203–3210.
18. Gerecke, A. C.; Sedlak, D. L. Precursors of N-nitrosodimethylamine in
natural waters. Environ. Sci. Technol. 2003, 37 (7), 1331–1336.
19. Mitch, W. A.; Sedlak, D. L. Characterization and fate of
N-nitrosodimethylamine precursors in municipal wastewater treatment
plants. Environ. Sci. Technol. 2004, 38 (5), 1445–1454.
20. Mitch, W. A.; Schreiber, I. M. Degradation of tertiary alkylamines during
chlorination/chloramination: Implications for formation of aldehydes,
nitriles, halonitroalkanes, and nitrosamines. Environ. Sci. Technol. 2008,
42 (13), 4811–4817.
94
21. Le Roux, J.; Gallard, H.; Croue, J. P. Chloramination of nitrogenous
contaminants (pharmaceuticals and pesticides): NDMA and halogenated
DBPs formation. Water Res. 2011, 45 (10), 3164–3174.
disinfection. Water Res. 2011, 45 (2), 944–952.
23. Shen, R. Q.; Andrews, S. A. NDMA formation kinetics from three
pharmaceuticals in four water matrices. Water Res. 2011, 45 (17),
5687–5694.
24. Selbes, M.; Kim, D.; Ates, N.; Karanfil, T. The Roles Of Tertiary Amine
Structure, Background Organic Matter And Chloramine Species On NDMA
Formation. Water Res. 2013, 47 (2), 945–953.25.
from reaction of monochloramine: a new disinfection by-product. Water Res.

2002, 36 (4), 817–824.
26. Le Roux, J.; Gallard, H.; Croue, J. P.; Papot, S.; Deborde, M. NDMA
Formation by Chloramination of Ranitidine: Kinetics and Mechanism.
Environ. Sci. Technol. 2012, 46 (20), 11095–11103.
27. Liu, Y. D.; Selbes, M.; Zeng, C.; Zhong, R.; Karanfil, T. Formation
Mechanism of NDMA from Ranitidine, Trimethylamine, and Other Tertiary
Amines during Chloramination: A Computational Study. Environ. Sci.
Technol. 2014, 48 (15), 8653–8663.
28. Chen, Z.; Valentine, R. L. Modeling the formation of N-
nitrosodimethylamine (NDMA) from the reaction of natural organic matter
(NOM) with monochloramine. Environ. Sci. Technol. 2006, 40 (23),
7290–7297.
29. Ozekin, K.; Valentine, R. L.; Vikesland, P. J. Modeling the decomposition
of disinfecting residuals of chloramine. Water Disinfection and Natural
Organic Matter: Characterization and Control; Minear, R. A., Amy, G. L.,
Eds.; American Chemical Society: Washington, DC, 1996; Vol. 649, pp
115−125.
30. Mitch, W.; Krasner, S. W.; Westerhoff, P.; Dotson, A. Occurrence and
Formation of Nitrogenous Disinfection By-Products (Final Report, Project
#3014); AwwaRF/USEPA: Denver, CO, 2009; p 145.
31. Chen, B. Y.; Westerhoff, P. Predicting disinfection by-product formation
potential in water. Water Res. 2010, 44 (13), 3755–3762.
32. Vikesland, P. J.; Ozekin, K.; Valentine, R. L. Monochloramine decay in
model and distribution system waters. Water Res. 2001, 35 (7), 1766–1776.
33. Ianni, J. C., Kintecus. Windows Version 2.80, 2002, www.kintecus.com,
(accessed September 10, 2014).
34. Duirk, S. E.; Gombert, B.; Croué, J.-P.; Valentine, R. L. Modeling
monochloramine loss in the presence of natural organic matter. Water Res.
2005, 39 (14), 3418–3431.
35. Selbes, M. The effects of amine structure, chloramine species and oxidation
strategies on the formation of N-nitrosodimethylamine. Ph.D. Thesis,
Clemson University, Clemson, South Carolina, 2014.
95
Chapter 6
Meta-Analysis of Trihalomethane Formation

Models and Application to Bromide Intrusion
Treavor H. Boyer*
Department of Environmental Engineering Sciences, Engineering School of

Sustainable Infrastructure & Environment (ESSIE), University of Florida,
P.O. Box 116450, Gainesville, Florida 32611-6450
This research evaluated previously published trihalomethane

(THM) formation models for their statistical robustness to
be applied outside of their original calibration data set, and
be used as a predictive tool at the water utility scale. All
models predicted THM4 (i.e., the sum of the four chlorine- and
bromine-containing THM species) based on the chlorination of
natural waters and were developed using different combinations
of precursor types (i.e., organic carbon concentration,
UV-absorbing substances, and bromide concentration) and
chlorination conditions (i.e., chlorine dose, pH, temperature,
and time) as explanatory variables. All models were log (base
10) transformed into a common format, and were evaluated
using a nationally representative water quality and THM4
formation data set based on the statistical metrics standard
error, mean absolute percentage error, R2, and adjusted R2. The
most robust log10(THM4) formation models had standard error
equal to 0.226–0.262 and adjusted R2 equal to 0.696–0.783.
The THM4 formation models that included bromide as an
explanatory variable tended to under predict THM4 formation
as a function of increasing bromide concentration. Overall, the
results of this research show that several previously published
THM4 formation models were developed with the appropriate
explanatory variables and calibrated with a sufficiently broad
data set such that the models may give a reasonable prediction
of THM4 formation, depending on the level of accuracy desired,
for chlorinated water conditions separate from the original

calibration data set. However, the majority of THM4 formation
models were not sensitive to changing bromide concentration,
and as such, under predicted THM4 formation at bromide
concentrations greater than 500 µg/L.
1. Introduction
Numerous empirical models have been developed to predict the formation
of disinfection by-products (DBPs) (1–5). The majority of the models are on the
formation of THM4 (i.e., the sum of the four chlorine- and bromine-containing
trihalomethanes) following chlorination of natural waters. The THM4 and
other DBP formation models are typically developed using multiple regression
techniques based on DBP formation considering key explanatory variables. The

most relevant and frequently used explanatory variables include dissolved organic
carbon (DOC) or total organic carbon (TOC) concentration, UV absorbance
(typically at 254 nm; UV254), bromide ion (Br−) concentration, chlorine dose
(Cl2), reaction time (t), pH, and temperature (T) (6). For instance, many of
the models are empirical multiple-parameter power function models of the
form THM4 = a(DOC)b(UV254)c(Br−)d(Cl2)e(t)f(T)g where a, b, c, d, e, f, and
g are constants. THM4 and other DBP formation models take the form of
multiple-parameter power function models because this type of model is able to
empirically describe the DBP formation trends that have been experimentally
observed based on changes in precursor concentration and chlorination conditions.
The multiple-parameter power function model is linearized by log transformation
to the form log(THM4) = log(a) + b·log(DOC) + c·log(UV254) + d·log(Br−)
+ e·log(Cl2) + f·log(t) + g·log(T). The linearized equation is solved using
multiple regression techniques by which values for the independent variable and
explanatory variables are input into the model, and the constants (i.e., regression
coefficients) are estimated such that the error between model predictions and
known values are minimized (4). The input values for multiple regression use
a variety of water quality and chlorination data sets, some of which include a
wide range of values for the explanatory variables while others are more limited.
For example, although bromide is a key precursor in DBP formation, many
chlorine-based THM and haloacetic acid formation models have been developed
for low bromide waters and therefore lack bromide as an explanatory variable (7,
8).
The usefulness of previously published DBP formation models depends on
the desired accuracy and the temporal and spatial scale of interest. For example,
at the molecular level, DBP formation models can provide insight on DBP
formation mechanisms (9), where the predictive accuracy of the model may be less
important relative to understanding the reaction pathway. As another example,
DBP formation models can be used at the utility scale to assess changes in water
quality or treatment modifications, and the consequent implications of these
changes on DBP formation (3, 10, 11). In this case, the predictive accuracy (or
predictive capability) of the models is the important consideration. For instance,
Lu et al. developed a site-specific THM4 formation model to evaluate water
98
utility compliance with the U.S. EPA Disinfectants/Disinfection By-Products
Rule under different water quality and treatment conditions (3). Many water
utilities, however, do not have the resources to develop DBP formation models
that are specific to their water supply and treatment processes. As such, there is a
need for previously published DBP formation models to be generally applicable
at the water utility scale to aid in daily operation and long term planning. This is
especially true for small water systems and water systems that are experiencing
an increased frequency of extreme events or changes in water quality. The
usefulness, however, of applying previously published DBP formation models at
the water utility scale will depend on the robustness of the existing models. In
particular, can DBP formation models developed and calibrated locally be applied
globally?
Accordingly, the goal of the research described in this chapter was to evaluate
previously published DBP formation models for their statistical robustness to be

applied separately from their original calibration data set so as to be used at the
water utility scale as a predictive tool. The specific objectives of the work were
to (1) evaluate the predictive capability of previously published chlorine-based
THM4 formation models using a nationally representative data set on THM4
formation, and (2) evaluate the sensitivity and application of the THM4 formation
models for water sources that experience an increase in bromide concentration.
This research focused on THM4 formation by chlorine because it constitutes
that largest number of previously published DBP formation models, and because
THM4 formation and speciation during chlorination is strongly influenced by
bromide concentration.
2. Methods
Previously published chlorine-based THM4 formation models were collected
from peer-reviewed journal articles and reports as described in detail elsewhere
(6). The models evaluated in this research are listed in Table 1. The models
follow the multiple-parameter power function format or log-transformed power
function format. The model number is the same as used in previous work (6).
For subsequent analysis, the THM4 formation models were divided into two
categories; those that included bromide as an explanatory variable and those that
did not include bromide as an explanatory variable. The predictive capability of
the THM4 formation models was assessed by log (base 10) transformation of the
power function models to a common format, i.e., log y = a + b1log x1 + bplog xp,
where y is the predicted THM4 concentration (in µg/L), x1 to xp are explanatory
variables, a and b1 to bp are regression coefficients, and p is the number of
explanatory variables. The power function models were log transformed to
linearize all models, which was the original form of the models used in the
multiple regression analysis, and to allow for graphical display of all THM4 data,
which ranged from 1 to 1000 µg/L. For each model, the predicted log10(THM4)
was compared with the measured log10(THM4) using the following statistical
metrics: linear coefficient of determination or R-squared (R2), adjusted R-squared
(adj-R2), standard error (SE), and mean absolute percentage error (MAPE).
99
Table 1. Summary of THM4 Formation Models Evaluated in this Work
No.a Modelb Ref
2 THM4 = 0.82(pH – 2.8)TOC(Cl2)0.25(t)0.36 (15)

5c,d THM4 = 7.21(TOC)0.004(UV254)0.534(Cl2 – 7.6NH3)0.224(t)0.255(Br– + 1)2.01(T)0.480(pH – 2.6)0.719 (16)
6 THM4 = 12.7254(TOC)0.291(t)0.2706(Cl2)–0.0719 (17)
7 THM4 = 108.8(TOC)0.2466(t)0.2956(UV254)0.9919(Cl2)0.126 (17)
8 THM4 = 131.7492(t)0.2931(UV254)1.075(Cl2)0.1064 (17)
9 THM4 = 14.6(pH – 3.8)1.01(Cl2)0.206(UV254)0.849(t)0.306 (18)
10 THM4 = 0.0412(DOC)1.10(Cl2)0.152(Br–)0.068(T)0.61(pH)1.60(t)0.260 (5)
11 THM4 = 0.044(DOC)1.030(t)0.262(pH)1.149(Cl2)0.277(T)0.968 (19)
100
12 THM4 = 1.392(DOC)1.092(pH)0.531(T)0.255 (19)

15 THM4 = 10-1.385(DOC)1.098(Cl2)0.152(Br–)0.068(T)0.609(pH)1.601(t)0.263 (4)
16 THM4 = 0.42(UV254)0.482(Cl2)0.339(Br–)0.023(T)0.617(pH)1.609(t)0.261 (4)
17e THM4 = 0.283(DOC·UV254)0.421(Cl2)0.145(Br–)0.041(T)0.614(pH)1.606(t)0.261 (4)
18 THM4 = 0.00253(DOC)1.22(Cl2)0.442(T)1.34(pH)1.75(t)0.34 (4)
19e THM4 = 0.012(DOC·UV254)0.47(Cl2)0.48(T)1.10(pH)2.38(t)0.35 (4)
20f THM4 = 3.296(DOC)0.801(Cl2)0.261(Br–)0.223(t)0.264 (4)
21 f THM4 = 75.7(UV254)0.593(Cl2)0.332(Br–)0.0603(t)0.264 (4)
22 f THM4 = 23.9(DOC·UV254)0.403(Cl2)0.225(Br–)0.141(t)0.264 (4)
24 THM4 = 4.527(t)0.127(Cl2)0.595(TOC)0.596(Br–)0.103(pH)0.66 (20)

No.a Modelb Ref
28 THM4 = 0.0001(Cl2)3.14(pH)1.56(TOC)0.69(t)0.175 (1)
29 THM4 = 0.0707(TOC + 3.2)1.314(pH – 4.0)1.496(Cl2 – 2.5)–0.197(T + 10)0.724 (11)

30 log10(THM4) = 1.078 + 0.398log10(TOC) + 0.158log10(T) + 0.702log10(Cl2) (21)
31e THM4 = 10–1.375(t)0.258(Cl2/DOC)0.194(pH)1.695(T)0.507(Br–)0.218 (22)
32 e THM4 = 10–0.038(Cl2)0.654(pH)1.322(t)0.174(SUVA)0.712 (8)
35d,g ln(THM4) = 6.11 + 0.211ln(t) + 1.64ln(Br–) + 0.34ln(Cl2) – 0.80ln(T) (23)
38 d THM4 = 1147(DOC)0.00(UV254)0.83(Br– + 1)0.27 (24)
a Model numbers from Ged et al. (6) Models are shown in format as originally presented, i.e., power function model or log-transformed, linearized model.
Statistical analysis was conducted on log transformed (base 10) models. b Units: THM4 (µg/L), Br− (µg/L) unless noted otherwise, DOC (mg/L), TOC
101
(mg/L), UV254 (1/cm), pH (–), T (° C), t (h), Cl2 (mg/L), SUVA (L/mg·m). c NH3 term neglected. d Br− (mg/L). e Combined terms separated following
logarithm rules. f Not corrected for pH to maintain consistent format with other log transformed models. g All terms divided by ln(10) to convert model to
log base 10.

The statistical metrics used in this work provide different goodness of fit
measurements for the predicted log10(THM4) data from the different models with
respect to the y = x line. For instance, R2 and adjusted R2 measure how close the
predicted data conform to a linear relationship with the measured data. In addition,
the adjusted R2 accounts for the number of explanatory variables in the model. The
higher the value for R2 and adjusted R2, the better the predicted data conforms to
a linear relationship with the experimental data, with R2 and adjusted R2 having
a maximum value of 1. R2 was calculated using the RSQ function in Microsoft
Excel as RSQ(log10(THM4i) predicted, log10(THM4i) measured). Adjusted R2
was calculated from R2 as 1 – (1 – R2)((n – 1)/(n – p – 1)) where n is the sample
size and p is the number of explanatory variables in the model. The SE is the
standard error of the estimate and measures the accuracy of the prediction, i.e., the
average distance between the predicted value and the y = x line. The SE has the
same units as the response variable. The smaller the value of SE, the higher the
accuracy of the prediction. SE was calculated in Microsoft Excel using the sum of
squares of differences function as SQRT(SUMXMY2(log10(THM4i) measured,
log10(THM4i) predicted)/n). The MAPE is a measure of the accuracy of the
predictions, and has the formula ∑abs[(log10(THM4i) measured – log10(THM4i)
predicted)/log10(THM4i) measured]/n where the summation is over all data and
the final value is expressed as a percent. The MAPE formula was calculated in
Microsoft Excel. The smaller the value of MAPE, the more accurate the prediction
by the model.
The measured THM4 data and corresponding water quality and chlorination
conditions came from a national study of bromide in U.S. drinking water
supplies (12). The subset of THM4 formation models that included bromide
as an explanatory variable were evaluated for their sensitivity to changing
bromide concentration. The bromide-containing THM4 formation models were
evaluated using a data set that measured THM4 formation for increasing bromide
concentration with all other water quality and chlorination conditions constant
(13). Finally, the bromide-containing THM4 formation models were used to
simulate the change in THM4 concentration that would occur for different
scenarios of seawater intrusion based on a bromide to total dissolved solids (TDS)
mass ratio of 0.0019134 for standard seawater (14).
3. Results and Discussion

3.1. Predictive Capability of THM4 Formation Models
Figure 1 shows individual plots of each THM4 formation model in terms
of log10(THM4) predicted versus log10(THM4) measured. The y = x line is
included in each plot to illustrate perfect agreement between model predictions
and measured data, which would give R2 = 1, SE = 0, and MAPE = 0%. Many
of the THM4 formation models evaluated in this work give predictions that are
tightly scattered along the y = x line, e.g., models 5, 9, 16, 17, 19, 21, and 22
have MAPE equal to 30 to 38% (see Table 2), thereby illustrating relatively
accurate predictions. This is a somewhat surprising result considering that the
102
THM4 formation models were originally calibrated using their own unique data
sets. This suggests that many of the models were developed to include the correct
collection of explanatory variables (i.e., precursor types and concentrations,
and chlorination conditions) and were calibrated using data sets in which the
explanatory variables covered a range of reasonable values. For instance,
the nationally representative data set that was used to evaluate the predictive
capability of the THM4 formation models had minimum, maximum, and average
values of the explanatory variables as follows: DOC (0.11, 18, 3.0 mg/L), UV254
(0.003, 0.533, 0.078 1/cm), Br− (2.5, 446, 76.6 µg/L), Cl2 (0.34, 54, 9.0 mg/L),
pH (constant 7), T (constant 20°C), and t (constant 96 h) (12). The DOC range
in the national data set covers almost the entire range of DOC that would be
expected during water treatment (25). As discussed in the next section, however,
the bromide range of < 500 µg/L is limited to freshwater with very low saline
water inputs (13). The predictive capability of several of the models, e.g., 8 and
12, could be improved by increasing the constant while keeping the same value
for the regression coefficients for the explanatory variables. This represents a
hybrid approach in which a water utility could slightly modify an existing THM4
formation model using historic water quality and THM4 formation data but not
redo the multiple regression analysis to develop an entirely new model.
Table 2 provides a summary of the goodness of fit statistics that quantify the
predictive capability of the log (base 10) transformed THM4 formation models.
The models with the lowest SE were models 21 (SE = 0.226), 22 (SE = 0.230),
16 (SE = 0.232), 30 (SE = 0.234)), 17 (SE = 0.235), 5 (SE = .241), and 9 (SE =
0.262). It is important to recognize that the SE values are for log10(THM4), i.e.,
log base 10 scale, such that a small value in SE can translate to a large variation
in THM4 concentration depending on the magnitude of log10(THM4). Table 3
illustrates the accuracy of the model estimates for log10(THM4) = 1.6 (40 µg/
L) and log10(THM4) = 1.9 (80 µg/L). For example, for a log10(THM4) model
prediction of 1.6, the model prediction ± one SE is 24 to 67 µg/L after anti-log
transformation. However, for a log10(THM4) model prediction of 1.9, the model
prediction ± one SE is 48 to 135 µg/L after anti-log transformation, which is
too wide of a range to be useful for regulatory compliance. Among the models
with the lowest SE discussed above, model 9 had the lowest MAPE (30%) and
model 22 had the highest adjusted R2. Models 5, 16, 17, 21, and 22 included
Br−, UV254, Cl2, and t as common explanatory variables. These four explanatory
variables represent the minimum set of explanatory variables needed to accurately
predict THM4 formation. For instance, Br− and UV254 represent both inorganic
and organic precursor material in THM4 formation, and Cl2 and t dictate the extent
of the chlorination reaction.
103
104
105
106
Figure 1. Log10(THM4) measured versus log10(THM4) predicted. Model

number given in legend. Measured THM4 is 145 unique samples.
107
Table 2. Summary of Goodness of Fit Statistics for Log (Base 10)
Transformed THM4 Formation Models.
Modela SE MAPE R2 adj-R2 p Bromide
2 0.346 36% 0.759 0.752 4 no
5 0.241 38% 0.749 0.736 7 yes
6 0.500 56% 0.759 0.754 3 no
7 0.563 31% 0.726 0.719 4 no
8 0.669 36% 0.671 0.664 3 no
9 0.262 30% 0.704 0.696 4 no
10 0.294 34% 0.766 0.756 6 yes
11 0.272 34% 0.759 0.750 5 no

12 0.767 51% 0.759 0.754 3 no
15 0.291 34% 0.766 0.756 6 yes
16 0.232 37% 0.765 0.755 6 yes
17 0.235 33% 0.790 0.779 7 yes
18 0.365 35% 0.759 0.750 5 no
19 0.299 35% 0.788 0.779 6 no
20 0.285 35% 0.747 0.739 4 yes
21 0.226 34% 0.759 0.752 4 yes
22 0.230 31% 0.791 0.783 5 yes
24 0.375 50% 0.767 0.758 5 yes
28 2.028 150% 0.759 0.752 4 no
29 0.646 28% 0.535 0.518 4 no
30 0.234 37% 0.759 0.754 3 no
31 0.559 59% 0.046 0.005 6 yes
32 0.370 50% 0.610 0.595 5 no
35 2.235 132% 0.113 0.087 4 yes
38 0.273 36% 0.646 0.641 2 yes
a N = 145 for all models except model 29, N = 118.
108
Table 3. Illustration of Predictive Accuracy of Model 21 (SE = 0.226) and Model 9 (SE = 0.262).
THM4, log10(THM4) SE log10(THM4) Anti- log10(THM4) Anti-
µg/L – SE log(log10(THM4) + SE log(log10(THM4)
– SE), µg/L + SE), µg/L

40 1.6 0.226 1.38 24 1.83 67
80 1.9 0.226 1.68 48 2.13 135
40 1.6 0.262 1.34 22 1.86 73
80 1.9 0.262 1.64 44 2.16 146
109

3.2. Application of THM4 Formation Models to Bromide Intrusion
The concentration of bromide in the aquatic environment can range from

less than 25 µg/L in freshwater to 67 mg/L in seawater (26). Brackish water,
such as seawater-impacted groundwater and surface water, can have bromide
concentrations on the order of 1 mg/L and higher (13, 27). The concern pertaining
to THM4 formation is the increase in bromide that occurs with seawater intrusion
into fresh groundwater and surface water, and the consequent increase in
brominated THMs upon chlorination (13). Hence, THM4 formation models
would be particularly useful if the models could accurately predict changes in
THM4 concentration due to changing bromide concentration. Figure 2 evaluates
the predictive capability of the THM4 formation models that included bromide
as an explanatory variable. Figure 2a shows log10(THM4) predicted versus
log10(THM4) measured. The measured THM4 data were from experiments

in which seawater was added to fresh groundwater and then chlorinated under
uniform formation conditions (13). The increase in bromide from seawater
resulted in a dramatic increase in measured THM4, and a shift from predominantly
chloroform in low bromide water (< 100 µg/L Br−) to predominantly bromoform
in high bromide water (> 500 µg/L Br−) (13). The seawater represented 2% v/v or
less of the freshwater–seawater mixture, which allowed for all other water quality
characteristics to remain approximately constant. The predicted log10(THM4)
was calculated by inputting the water quality characteristics and chlorination
conditions of the freshwater–seawater mixture samples into the THM4 models.
The results in Figure 2a show a linear relationship between log10(THM4)
predicted and log10(THM4) measured with R2 values equal to 0.918–0.984.
However, most of the data do not conform to the y = x line, which indicates
poor agreement between the model predictions and experimental measurements.
For example, SE values for all models (excluding model 5) ranged from 0.228
to 1.264. Predicted data using model 5 showed very good agreement with the
experimental measurements based on SE, with SE equal to 0.092. The main
difference between model 5 and the other models that included bromide as an
explanatory variable was the higher value of the regression coefficient for Br− in
model 5 than the other models (see Table 1). This made model 5 more responsive
to changes in bromide concentration.
Figure 2b shows the THM4 concentration, i.e., the anti-log of log10(THM4)
predicted, as a function of increasing bromide concentration. The data in Figure
2b are anti-log transformed to better see the sensitivity of the predicted THM4
concentration as a function of bromide concentration. The THM4 formation
models show several important trends with respect to increasing bromide
concentration. Foremost, many of the models were not sensitive to changing
bromide concentration, and as such, under predicted THM4 formation with
increasing bromide concentration. The low sensitivity of the models to changing
bromide concentration was anticipated based on the small value of the regression
coefficient for Br− in many of the models. Models 5 and 35 showed the greatest
sensitivity in terms of increasing THM4 concentration with increasing bromide
concentration, and these models also had the largest values for the bromide
regression coefficient among the models that included bromide as an explanatory
110
variable (see Table 1). Model 5 was the only model that could relatively accurately
predict THM4 formation at low and high bromide concentrations, which was
indicated in Figure 2a and by the low SE of 0.092. However, despite model 5
being the most accurate model, the predicted data showed a different trend than
the measured data in terms of increasing THM4 concentration with increasing
bromide. Model 35 predicted increasing THM4 concentration with increasing
bromide concentration; however, the predictions were not accurate as shown in
Figure 2a and by the high SE of 1.264. In summary, at bromide concentration in
the range of 50–100 µg/L, there were models that both under predicted and over
predicted THM4 formation (see Figure 2b). This was because other factors are
more important than bromide on THM4 formation at low bromide concentrations,
such as the concentration and character of DOC. At bromide concentration greater
than 500 µg/L, all models under predicted THM4 formation. Thus, there is a clear
need for THM4 formation models that are calibrated using water quality data sets
that contain bromide in the range of 500 µg/L to great than 1 mg/L.
Given that model 5, among the models considered, showed the best agreement
between predicted THM4 concentration and measured THM4 concentration
for the condition of increasing bromide concentration, it was used to predict
THM4 formation under a more generalized scenario of seawater intrusion. In
particular, Figure 3 shows THM4 formation as a function of increasing TDS
where the bromide concentration was calculated from the bromide-to-TDS ratio
for standard seawater (14). Model 5 includes TOC, UV254, Br−, Cl2, t, pH, and T
as explanatory variables; the value used for each variable is given in the caption
to Figure 3. The variables represent reasonable values for finished drinking water,
e.g., 1 mg/L TOC and Cl2-to-TOC mass ratio of 2 (25). Figure 3 shows that
THM4 increases steadily with increasing TDS, which is due to the increasing
bromide concentration as shown in Figure 2b. An interesting data point in Figure
3 is 500 mg/L TDS, which is equivalent to THM4 formation of 144 µg/L. Thus, at
the secondary standard for TDS, THM4 formation exceeds its primary maximum
contaminant level. This is significant because secondary drinking water standards
are not health-based, whereas primary drinking water standards are in place to
protect public health (28). Hence, the model predictions in Figure 3 suggest that
seawater intrusion, and in particular the co-transport of bromide with chloride,
will be problematic in terms of increasing the formation of THM4 well before
the water would become unpalatable due to high TDS. Furthermore, laboratory
studies have shown that high chloride concentrations can catalyze the formation
of highly reactive bromine species during chlorination, which further increases
the formation of brominated DBPs (29). Overall, the results in Figure 3 illustrate
a new and unique challenge for drinking water treatment because there are a
limited number of processes that can effectively remove bromide and seawater
intrusion is expected to increase in many coastal areas due over pumping, changes
in recharge, and sea-level rise (30, 31).
111
Figure 2. (a) Log10(THM4) measured versus log10(THM4) predicted. Model

number given in legend. Measured THM4 is 6 unique samples. (b) Measured and
predicted THM4 as a function of increasing bromide concentration. Measured
THM4 data from Ged et al. (13) Model number given in legend. Experimental
conditions and model inputs: DOC = 1.4 mg/L, UV254 = 0.037 1/cm, pH 8, T =
20 °C, chlorine dose = 2.7 mg/L as Cl2, time = 24 h, bromide = variable.
112
Figure 3. THM4 predicted by model 5 as a function of increasing TDS. Model

inputs: DOC = 1.0 mg/L, SUVA254 = 2.5 L/mg·m, UV254 = 0.025 1/cm, pH
8, T = 20 °C, chlorine dose = 2.0 mg/L as Cl2, time = 24 h, Br− (mg/L) =
0.0019134×TDS (mg/L) from Millero et al. (14)
4. Conclusions
Several previously published THM4 formation models were determined
to be moderately robust in their model development and calibration such that
the models could be applied at the water utility scale using water quality and
chlorination conditions separate from the original calibration data set. For
example, the anti-log transformation of log10(THM4) ± SE for log10(40) = 1.6
± 0.226 gives an accuracy of 24 to 67 µg/L for a target THM4 concentration
of 40 µg/L. This example used the highest level of accuracy that was obtained
in this work, so there is the potential for some of these models to be used
at the water utility scale. In particular, the more statistically robust THM4
formation models determined in this work are expected to be useful for small
water systems and water systems experiencing frequent changes in water quality,
such as variable TOC concentrations. With minimal to no modification, water
utilities can use these models to predict THM4 and help inform short-term
operation and long-term planning. One caveat to the THM4 formation models
evaluated in this work was the low sensitivity of many of the models to increasing
bromide concentration. Hence, previously published THM4 formation models
should be used with caution if the application of the model is predicting THM4
concentration under conditions of variable bromide concentration, such as the
case of seawater intrusion. An important side result of the bromide sensitivity
analysis was the illustration that THM4 concentration could reach problematic
113
levels (> 100 µg/L) when the TDS was at or below its secondary standard of
500 mg/L. This highlights a new challenge for drinking water treatment whereby
seawater intrusion, with its drivers in urbanization and global climate change, can
alter the water matrix and dramatically increase the formation of DBPs, especially
brominated DBPs.
Acknowledgments
This publication is based upon work supported by the University of Florida
Research Opportunity Seed Fund award: Florida as a Laboratory for Global
Urbanization, Sea Level Rise, and Future Health Risks of Drinking Water Sources.
References
1. Chowdhury, S.; Champagne, P.; McLellan, P. J. Models for predicting
disinfection byproduct (DBP) formation in drinking waters: A chronological
review. Sci. Total Environ. 2009, 407, 4189–4206.
2. Chowdhury, S.; Champagne, P.; McLellan, P. J. Investigating effects of
bromide ions on trihalomethanes and developing model for predicting
bromodichloromethane in drinking water. Water Res. 2010, 44, 2349–2359.
3. Lu, O.; Krasner, S. W.; Liang, S. Modeling approach to treatability analyses
of an existing treatment plant. J. Am. Water Work Assoc. 2011, 103, 103–117.
4. Sohn, J.; Amy, G.; Cho, J.; Lee, Y.; Yoon, Y. Disinfectant decay and
disinfection by-products formation model development: chlorination and
ozonation by-products. Water Res. 2004, 38, 2461–2478.
5. Amy, G. L.; Siddiqui, M.; Ozekin, K.; Zhu, H. W.; Wang, C. Empirical
based models for predicting chlorination and ozonation byproducts:
haloacetic acids, chloral hydrate, and bromate. Report CX 819579.; U.S.
Environmental Protection Agency: Washington, D.C., 1998.
6. Ged, E. C.; Chadik, P. A.; Boyer, T. H. Predictive capability of chlorination
disinfection byproducts models. J. Environ. Manage. 2015, 149, 253–262.
7. Nikolaou, A. D.; Golfinopoulos, S. K.; Arhonditsis, G. B.; Kolovoyiannis, V.;
Lekkas, T. D. Modeling the formation of chlorination by-products in river
waters with different quality. Chemosphere 2004, 55, 409–420.
8. Uyak, V.; Ozdemir, K.; Toroz, I. Multiple linear regression modeling of
disinfection by-products formation in Istanbul drinking water reservoirs.
Sci. Total Environ. 2007, 378, 269–280.
9. Gallard, H.; von Gunten, U. Chlorination of natural organic matter: kinetics
of chlorination and of THM formation. Water Res. 2002, 36, 65–74.
10. Black, B. D.; Harrington, G. W.; Singer, P. C. Reducing cancer risks by
improving organic carbon removal. J. Am. Water Work Assoc. 1996, 88,
40–52.
11. Uyak, V.; Toroz, I.; Meric, S. Monitoring and modeling of trihalomethanes
(THMs) for a water treatment plant in Istanbul. Desalination 2005, 176,
91–101.
114
12. Amy, G.; Siddiqui, M.; Zhai, W.; DeBroux, J.; Odem, W., Survey of Bromide
in Drinking Water and Impacts on DBP Formation. AWWA Research
Foundation: Denver, CO, 1993.
13. Ged, E. C.; Boyer, T. H. Effect of seawater intrusion on formation
of bromine-containing trihalomethanes and haloacetic acids during
chlorination. Desalination 2014, 345, 85–93.
14. Millero, F. J.; Feistel, R.; Wright, D. G.; McDougall, T. J. The composition of
Standard Seawater and the definition of the Reference-Composition Salinity
Scale. Deep Sea Res., Part I 2008, 55, 50–72.
15. Urano, K.; Wada, H.; Takemasa, T. Empirical rate equation for
trihalomethane formation with chlorination of humic substances in water.
Water Res. 1983, 17, 1797–1802.
16. Malcolm-Pirnie Bay-Delta Water Quality Modeling. Report No. 15-041;
Prepared for the Metropolitan Water District of Southern California: White

Plains, NY, December 1993.
17. Chang, E. E.; Chao, S. H.; Chiang, P. C.; Lee, J. F. Effects of chlorination on
THMs formation in raw water. Toxicol. Environ. Chem. 1996, 56, 211–225.
18. Rathbun, R. E. Regression equations for disinfection by-products for the
Mississippi, Ohio and Missouri rivers. Sci. Total Environ. 1996, 191,
235–244.
19. Rodriguez, M. J.; Serodes, J.; Morin, M. Estimation of water utility
compliance with trihalomethane regulations using a modelling approach. J.
Water Supply: Res. Technol.--AQUA 2000, 49, 57–73.
20. Al-Omari, A.; Fayyad, M.; Qader, A. Modeling trihalomethane formation
for Jabal Amman water supply in Jordan. Environ. Model Assess. 2004, 9,
245–252.
21. Toroz, I.; Uyak, V. Seasonal variations of trihalomethanes (THMs) in water
distribution networks of Istanbul City. Desalination 2005, 176, 127–141.
22. Hong, H. C.; Liang, Y.; Han, B. P.; Mazumder, A.; Wong, M. H. Modeling
of trihalomethane (THM) formation via chlorination of the water from
Dongjiang River (source water for Hong Kong’s drinking water). Sci. Total
Environ. 2007, 385, 48–54.
23. Semerjian, L.; Dennis, J.; Ayoub, G. Modeling the formation of
trihalomethanes in drinking waters of Lebanon. Environ. Monit. Assess.
2009, 149, 429–436.
potential in water. Water Res. 2010, 44, 3755–3762.
25. Obolensky, A.; Singer, P. C.; Shukairy, H. M. Information collection rule
data evaluation and analysis to support impacts on disinfection by-product
formation. J. Environ. Eng.-ASCE 2007, 133, 53–63.
26. Hem, J. D., Study and interpretation of the chemical characteristics of
natural water. United State Geological Survey Water-Supply Paper 2254
1985, Alexandria, VA.
27. Magazinovic, R. S.; Nicholson, B. C.; Mulcahy, D. E.; Davey, D. E. Bromide
levels in natural waters: its relationship to levels of both chloride and total
dissolved solids and the implications for water treatment. Chemosphere
2004, 57, 329–335.
115
28. USEPA, Drinking Water Contaminants: National Primary Drinking Water
Regulations [website]. http://water.epa.gov/drink/contaminants/index.cfm:
accessed 09/23/2013, 2013.
29. Sivey, J. D.; Arey, J. S.; Tentscher, P. R.; Roberts, A. L. Reactivity of
BrCl, Br2, BrOCl, Br2O, and HOBr Toward Dimethenamid in Solutions
of Bromide + Aqueous Free Chlorine. Environ. Sci. Technol. 2013, 47,
1330–1338.
30. Watson, K.; Farré, M. J.; Knight, N. Strategies for the removal of halides from
drinking water sources, and their applicability in disinfection by-product
minimisation: A critical review. J. Environ. Manage. 2012, 110, 276–298.
31. Werner, A. D.; Bakker, M.; Post, V. E. A.; Vandenbohede, A.; Lu, C.; Ataie-
Ashtiani, B.; Simmons, C. T.; Barry, D. A. Seawater intrusion processes,
investigation and management: Recent advances and future challenges. Adv.
Water Resour. 2013, 51, 3–26.
116
Chapter 7
Nitrosamine Precursors and Wastewater

Indicators in Discharges in the
Sacramento-San Joaquin Delta
Chih-Fen Tiffany Lee,* Stuart W. Krasner, Michael J. Sclimenti,

Matthew Prescott, and Yingbo C. Guo
Water Quality Section, Metropolitan Water District of Southern California,

700 Moreno Avenue, La Verne, California 91750-3303
Wastewater treatment plant (WWTP) discharges in

the Sacramento-San Joaquin Delta were shown to be
a source of N-nitrosodimethylamine (NDMA) and
N-nitrosomorpholine (NMOR), as well as precursors for
NDMA and N-nitrosopyrridine (NPYR). NDMA and NPYR
were disinfection by-products, whereas NMOR was a
wastewater contaminant. Nitrosamines were not found
downstream of the WWTPs due to dilution with river water
and/or sunlight photolysis, whereas NDMA precursor loadings
in the river downstream of the WWTPs were at higher levels
than at the upstream sites. The anticonvulsant primidone
and the artificial sweetener sucralose were found to be good
wastewater indicators and the percentage of treated wastewater
in the Sacramento River (2-3%) and the San Joaquin River
(2-14%) was determined by these. Moreover, the increase in
NDMA precursor loadings in downstream sites was consistent
with the wastewater volumetric contributions to river flow.
Introduction
The Sacramento-San Joaquin Delta (Delta) is at the nexus of a California
statewide water system, which provides a critical water supply (State Project Water
[SPW]) to 23 million people in California. The Delta is at the confluence of the
Sacramento and San Joaquin Rivers (Figure 1). The Delta empties into the Pacific

Ocean via an estuary to the San Francisco Bay. Due to tidal activity, especially
during droughts, there can be tidal actions in the San Joaquin River. There are
nine wastewater treatment plants (WWTPs) discharging from 0.3 to 181 million
gallons per day (MGD) into the Delta (1). The City of Stockton and Sacramento
Regional operate the two largest WWTPs in the Delta. They discharge into the
San Joaquin and Sacramento River, respectively (Figure 1).
Wastewater discharges (effluent organic matter [EfOM]) contain emerging
chemical contaminants, such as nitrosamines and their precursors, and
pharmaceuticals and personal care products (PPCPs). EfOM is known to be
one of the major sources of N-nitrosodimethylamine (NDMA) precursors (2).
NDMA is a disinfection by-product (DBP) preferentially formed by chloramines
(3). N-Nitrosomorpholine (NMOR) is also commonly detected in wastewater
discharges, where it appeared to be a wastewater contaminant and not a DBP per
se (4).
Nitrosamines are known to be probable human carcinogens, and they have
gained a lot of attention from health and regulatory agencies in the United
States (U.S.) and internationally (5). Although there is no federal regulation for
nitrosamines in drinking water, the California State Water Resources Control
Board Division of Drinking Water (DDW) set a notification level of 10 ng/L each
for NDMA and two other nitrosamines (6). The objective of this study was to
evaluate the presence and fate of NDMA, other nitrosamines, their precursors,
and selected PPCPs in the Delta, including seasonal and year-to-year variability.
Figure 1. Map of Sacramento –San Joaquin Delta (triangles = WWTPs).
120
Experimental
Overview of Study
The sampling sites in this study included locations upstream and downstream
of the Stockton and Sacramento Regional WWTPs, and WWTP effluent or river
water immediately adjacent to the discharge. Table 1 lists the sampling locations.
There were a total of eight sampling events: both watersheds were sampled in six
of them, covering summer, fall, and winter seasons over a two-year period; only the
San Joaquin River and the Stockton WWTP were sampled in two additional events.
The samples were analyzed for eight nitrosamines, including NDMA, NMOR, and
N-nitrosopyrridine (NPYR). Certain PPCPs found in wastewater can be used as
conservative indicators of wastewater impacts in drinking water supplies. The
indicators used in this project were primidone (an anticonvulsant drug) (7, 8)
and sucralose (an artificial sweetener - Splenda®) (9). The percentage of treated
wastewater in the rivers was calculated as the ratio of the concentrations of these
indicators in the rivers against their concentrations in the WWTP effluents. In
addition, modeling of volumetric contributions of the WWTP discharges to the
river flow was estimated using a simulation model.
Table 1. Sampling Locations in the Study

Sampling Location
San Joaquin River Watershed Sacramento River Watershed
Upstream of Stockton WWTP Upstream of Sacramento Regional WWTP
Stockton WWTP filter effluent Sacramento Regional WWTP outfall (grab
sample in Sacramento River by discharge)
Stockton WWTP plant effluent
Downstream of Stockton WWTP Downstream of Sacramento Regional WWTP
Study Sites
Stockton WWTP
The Stockton WWTP was a tertiary treatment plant. The advanced treatment
units included oxidation ponds, wetlands, nitrifying biotowers, dissolved air
floatation (DAF), and mix media filters. The effluent was chlorinated and
ammonia was added (if needed; ammonia was present or not, depending on how
well the biotowers nitrified the water to form chloramines). The tertiary processes
provided additional treatment to clean up the wastewater, including removal of
ammonia, algae, and a final filtering through sand, gravel, and anthracite coal to
remove more organic material and fine suspended particles.
121
Sacramento Regional WWTP
The Sacramento Regional WWTP was a secondary treatment plant.

Secondary treatment included a conventional aerobic biological treatment process
and secondary clarification. The effluent was poorly nitrified. Chlorine was
added, which should have reacted with the ammonia to form choramines.
Analytical Methods
Nitrosamines
Standard method 6450B was used to measure eight nitrosamines:

NDMA, N-nitrosomethylethylamine (NMEA), N-nitrosodiethylamine (NDEA),

N-nitrosodi-n-proplyamine(NDPA), N-nitrosodi-butylamine (NDBA), NPYR,
N-nitrosopiperidine (NPIP), and NMOR. Nitrosamines were extracted and
concentrated using solid-phase extraction (SPE) with Ambersorb 572 resin
and were analyzed using gas chromatography/mass spectrometry (GC/MS)
with chemical ionization (10). Nitrosamine concentrations were determined by
comparison of the area ratio of a unique product ion to one of the isotopically
labeled internal standards (i.e., d6-NDMA, d14-NDPA, and 15N2-NDEA) against
calibration curves made from method (SPE) standards. The minimum reporting
level (MRL) for each nitrosamine was 2 ng/L.
Nitrosamine Precursors
Nitrosamine precursors were determined using formation potential (FP) tests

(11, 12). Nitrosamine FP tests were conducted in the presence of chloramines.
Filtered (0.45-µm) samples were chloraminated on a reactivity basis (based on the
amount of organic carbon) and held for 3 days at pH 8 at 25°C. The chlorine dose
was based on the level of total organic carbon (TOC) (i.e., Cl2 = 3 x TOC, weight
basis), and the amount of ammonia was calculated based on a chlorine-to-nitrogen
(Cl2/N) weight ratio of 3:1. In the tests, ammonia was added before chlorine in
order to form chloramines with no free chlorine contact time. Note, if ammonia
was present in the sample, this was factored into how much more was added in the
FP test. For poorly nitrified wastewaters, no ammonia was added in the FP test,
and the resultant Cl2/N weight ratio was typically in the range of 1:1 to 2:1.
PPCPs
PPCPs were concentrated using a SPE workstation with HLB cartridges

(Waters, Milford, MA) and were analyzed using liquid chromatography/tandem
MS (LC/MS/MS) with a 2.0-mm C18 reversed-phase high pressure LC (HPLC)
column (13). For primidone, the MS was operated under an electrospray positive
ion mode. PPCPs were identified by matching both the retention time and the
122
MS/MS transition in the samples with those in authentic standards. An isotope
dilution technique was used for quantitation (14). The MRL for primidone was
2 ng/L.
Sucralose
Filtered (0.45-μm) water samples were injected directly onto the C18 HPLC
column equipped with a 2.0-mm C18 guard column. The MS was operated under
an electrospray negative ion mode. Sucralose was identified as were other PPCPs.
An isotope dilution technique was used for quantitation with d6-sucralose. The
MRL was 0.2 µg/L.
Hydraulic Flow Modeling
The California Department of Water Resources (DWR) used the Delta

Simulation Model II (DSM2) to study the complex hydraulic system in the
Delta (15). DSM2 was a mathematical model for dynamic simulation of
one-dimensional hydrodynamics, water quality, and particle tracking in a network
of riverine or estuarine channels.

N-Nitrosamines and Their Precursors in the Stockton WWTP
Studies have shown that WWTPs may remove N-nitrosamines present
in the wastewater (16). Studies in the U.S. showed that chloramination of
treated wastewater can form N-nitrosamines above the background level (4).
Moreover, studies in the U.S. showed that N-nitrosamine precursors were present
in secondary or tertiary treated wastewater (2). In the Stockton WWTP, NDMA,
NPYR and NMOR were detected in FP samples. Figure 2 shows the impact of the
tertiary treatment process on N-nitrosamine FP. NDMA FP was over 1000 ng/L
after secondary treatment; however, the concentration dropped significantly after
the tertiary treatment began and remained fairly constant throughout the whole
process. Similar trends were observed for NPYR FP and NMOR FP.
In addition to measuring the N-nitrosamine FP, the amount of N-nitrosamines
formed after chloramine addition at the WWTP was determined. Figure 3 shows
there was no NDMA detected at or above the MRL in the WWTP filter effluent;
however, the formation of NDMA spiked up at the WWTP effluent (interquartile
range ~4-82ng/L, median ~18 ng/L) due to chloramine usage. Unlike NDMA,
NMOR was present before chloramine addition, and its occurrence stayed the same
after chloramine addition, as NMOR was a wastewater contaminant and not a DBP
(4). Although NMOR was detected in some of the FP tests, its concentration in
the FP test was often no different than that of the water before FP testing. Neither
NDMA nor NMOR were detected downstream of the WWTP. This was likely due
to dilution in the river water and/or sunlight photolysis (17).
123
Figure 2. Impact of tertiary treatment on N-nitrosamine FP at the Stockton

WWTP.
Figure 3. NDMA and NMOR in the San Joaquin River and at the Stockton
WWTP (bottom and top of the box = 25th to 75th percentile, line through the box
= median, bottom and top of the whiskers = 10th to 90th percentile).
124
Wastewater Indicators in WWTPs and Estimated Wastewater Effluent
Figure 4 shows the concentrations of the wastewater indicators in the effluent

of the Stockton WWTP. The interquartile range (25th to 75th percentile) of
sucralose concentrations was 24 to 31 µg/L, which compared well with the typical
range detected in other U.S. WWTP effluents of 20-30 µg/L (median = 27 µg/L)
(9). The interquartile range of primidone concentrations, 109 to 136 ng/L, also
fell within the typical range of other U.S. WWTPs of 100-200 ng/L (7).
An effluent sample site was not available from the Sacramento Regional
WWTP, so the sampling location was a depth sample in the Sacramento River
near the discharge pipe (outfall); all samples were grab samples. The effluent
sample was diluted to varying extents by river flow. The percent effluent in the
sample was estimated based on either the sucralose or primidone concentration,
assuming a similar level at the Sacramento Regional WWTP as that detected at

the Stockton WWTP (Figure 4). These percentages were similar to estimates
based on the ammonia level in the sample and the ammonia concentration in
discharge records. The N-nitrosamine FP data from the Sacramento Regional
WWTP discharge sample were corrected for dilution to provide an estimate of the
levels likely present in the WWTP effluent as part of the effort to follow the fate
of the precursors in the Sacramento River (Figure 5).
Figure 4. Concentrations of the wastewater indicators in the effluent of the

Stockton WWTP.
125
Figure 5. Estimated percent WWTP effluent in Sacramento Regional WWTP

outfall samples based on sucralose and primidone.
Impact of WWTP Effluents in the Sacramento and San Joaquin Rivers
Figures 6 and 7 show the occurrence of NDMA precursors in the Sacramento

and San Joaquin Rivers upstream and downstream of the WWTPs, and in the
WWTP effluents during the six main sampling events. The percent effluent
in the Sacramento Regional WWTP outfall sample was estimated based on
primidone and then the NDMA FP in the WWTP effluent was back calculated.
The San Joaquin WWTP effluent had 206-358 ng/L (median = 218 ng/L) and
the Sacramento Regional WWTP effluent was estimated to have had 227-812
ng/L (median = 508 ng/L). In the Sacramento River, the NDMA precursor levels
were higher downstream of the WWTP (9-22 ng/L [median = 17 ng/L]) than
upstream (4-9 ng/L [median = 5 ng/L]), due to the wastewater discharge from the
Sacramento Regional WWTP. In the San Joaquin River, in most cases, higher
NDMA precursor levels were found downstream of the WWTP (e.g., 12-48
ng/L [median = 40 ng/L]) than upstream (e.g., 7-38 ng/L [median = 16 ng/L]).
However, in some cases, the upstream location either had similar or higher levels
of NDMA precursors than downstream. These events were due to “reverse river
flow,” where the direction of the natural flow was reversed due to tidal action (the
effect of the tidal flow into the Delta and the surrounding estuaries on the flow of
the river). Sucralose and primidone had similar trends as the NDMA FP in the
San Joaquin River during reverse river flow sampling events.
126
Figure 6. NDMA precursors in the Sacramento River and estimated in the

Sacramento Regional WWTP effluent (based on primidone).
Table 2 shows an example of the impact of the Stockton WWTP discharge

on the San Joaquin River. There was 35 µg/L of sucralose and 132 ng/L of
primidone in the WWTP effluent. Higher levels of primidone (9.1 ng/L increase)
and sucralose (2.7 µg/L increase) were observed at the downstream sample
site due to the wastewater discharge from the Stockton WWTP. Based on these
wastewater indicators, the downstream site was ~7-8% higher in EfOM than
the upstream location (e.g., (5.3 µg/L of sucralose downstream – 2.6 µg/L of
sucralose upstream)/35 µg/L of sucralose in the WWTP effluent)). In terms of the
NDMA FP, there was a higher level downstream of the WWTP (20 ng/L increase).
Figure 7. NDMA precursors in the San Joaquin River and in the Stockton WWTP
effluent (reverse flows during the November 2011 and July 2012 sample events).
127
Table 2. Concentrations of the Wastewater Indicators and NDMA FP in the
San Joaquin River and in the Stockton WWTP Effluent (November 2012)
NDMA FP Primidone Sucralose
ng/L ng/L µg/L
Upstream 25 9.5 2.6
WWTP 283 132 34.9
Downstream 45 18.6 5.3
Estimated % Effluent 10 % 7% 8%
Based on the wastewater indicators, it was predicted that the NDMA FP

should have increased by ~20-23 ng/L, consistent with the observed increase.
Note, there was a smaller WWTP upstream of the Stockton WWTP, which
contributed to the presence of these contaminants in the upstream sample location.
However, in this study, the focus was on the incremental contribution of the
Stockton WWTP to the precursor loading in the river.
Figure 8 shows the relationship of the wastewater indicators to NDMA FP
in both river/WWTP systems. On a central tendency basis, the slope of the trend
line was steeper for the Sacramento River/WWTP system. For example, 10 µg/L
of sucralose corresponded to ~90 ng/L of NDMA FP in the San Joaquin River,
whereas the same amount of sucralose corresponded to a higher level of NDMA
FP (>150 ng/L) in the Sacramento River. This was probably due (in part) to the
different treatment processes at the Sacramento Regional and Stockton WWTPs;
the Sacramento Regional WWTP only had secondary treatment, whereas the
Stockton WWTP had tertiary treatment. Tertiary treatment can remove NDMA
precursors; however, tertiary treatment should have little effect on removing these
wastewater indicators.
Percentage of River Flow that Was Wastewater-Impacted
DSM2 can calculate stages, flows, velocities, transport of individual particles,

and mass transport processes for conservative and non-conservative constituents,
including salts, water temperature, dissolved oxygen, and dissolved organic
carbon. DMS2 was used to determine the volumetric fraction of WWTP effluent
at certain locations, downstream of the WWTPs.
128
Figure 8. Relationship of the wastewater indicators to NDMA FP in the

Sacramento and San Joaquin Rivers and at the corresponding WWTPs
(Sacramento Regional WWTP outfall data plotted as measured, without
correction for dilution of the WWTP effluent in the river).
129
Figure 9 shows the percent wastewater effluent in these two rivers as estimated
by the model. The data from the model was a daily average of the river flow,
whereas the data from our sampling events were based on grab samples. In the San
Joaquin River, the percent WWTP effluent in the downstream location fluctuated
from 2 to 14%, whereas the Sacramento River percentage remained steady at 2-4%.
This was due to the higher river flow of the Sacramento River. The average annual
river flow of Sacramento River was 23,469 cubic foot per second (cfs), whereas
in the San Joaquin River it was 4,460 cfs.
Figure 9. Estimated percentage of WWTP effluent in downstream San Joaquin

and Sacramento River sample sites (volume basis) from hydraulic flow modeling.
130
Figure 10. Estimated percent of WWTP effluent in downstream Sacramento River

and San Joaquin River sample sites based on the wastewater indicators
Figure 10 shows that both wastewater indicators had good agreement in the
estimation of percent WWTP effluent in the downstream locations in these two
rivers. Moreover, the estimations based on wastewater indicators matched well
with the hydraulic flow-calculated contributions.
Conclusions
Effluents from the two largest WWTPs in the Delta were shown to be an
important source of NDMA precursors in a major source of drinking water for
23 million Californians. NDMA precursor loadings in the river were detected
downstream of the WWTPs at higher levels than at the upstream sites, where the
increase was consistent with the percent of river flow due to the WWTP discharge.
However, due to river dilution and/or solar photolysis, the nitrosamines formed or
present at the WWTP were not detected at or above the MRLs downstream of these
WWTPs. The anticonvulsant primidone and the artificial sweetener sucralose were
found to be good wastewater indicators, and the percentage of treated wastewater
in the rivers was determined based on these indicators. In addition, the volumetric
wastewater contribution to river flow was determined using a hydraulic simulation
model, which matched the percent effluent estimations based on the wastewater
131
indicators. A good relationship was found between the level of NDMA precursors
in the river/WWTP systems and the wastewater indicators, where the slope of
the trend line was steeper for the river system with the WWTP with secondary
treatment as compared to the system with the tertiary treatment process.
Acknowledgments
The authors gratefully acknowledge support from the staff of the California
DWR and the City of Stockton, with special thanks going to Carol DiGiorgio
(DWR) and Laura Lazzelle (City of Stockton).
References
1. Guo, Y. C.; Krasner, S. W.; Fitzsimons, S.; Woodside, G.; Yamachika, N.
Source, Fate, and Transport of Endocrine Disruptors, Pharmaceuticals, and
Personal Care Products in Drinking Water Sources in California; National
Water Research Institute: Fountain Valley, Ca., May 2010, pp 14−21
2. Krasner, S. W.; Westerhoff, P.; Chen, B.; Rittmann, B. E.; Nam, S.-N.;
Amy, G. Impact of Wastewater Treatment Processes on Organic Carbon,
Organic Nitrogen, and DBP Precursors in Effluent Organic Matter. Environ.
Sci. Technol. 2009, 43 (8), 2911–2918.
3. Mitch, W. A.; Sharp, J. O.; Trussell, R. P.; Valentine, R. L.; Alvarez-
Cohen, L.; Sedlak, D. L. N-Nitrosodimethylamine as a Drinking Water
Contaminant: A review. Environ. Eng. Sci. 2003, 20 (5), 389–404.
Occurrence of Disinfection Byproducts in United States Wastewater
Treatment Plant Effluents. Environ. Sci. Technol. 2009, 43 (21), 8320–8325.
5. Krasner, S. W.; Mitch, W. A.; McCurry, D. L.; Hanigan, D.; Westerhoff, P.
Formation, Precursors, Control, and Occurrence of Nitrosamines in Drinking
Water: A review. Wat. Res. 2013, 47 (13), 4434–4446.
6. California Environmental Protection Agency State Water Resources
Control Board. NDMA and Other Nitrosamines - Drinking Water
Issues;http://www.waterboards.ca.gov/drinking_water/certlic/
drinkingwater/NDMA.shtml [cited November 1, 2014].
7. Krasner, S. W.; Pastor, S. J.; Garcia, E. A. Measurement of the
Pharmaceutical Primidone as a Conservative Tracer of Wastewater
Influences in Drinking Water Supplies. In Proceedings of 2006 American
Water Works Association (AWWA) Water Quality Technology Conference
(WQTC); AWWA: Denver, CO, 2006.
8. Guo, Y. C.; Krasner, S. W. Occurrence of Primidone, Carbamazepine,
Caffeine, and Precursors for N-Nitrosodimethylamine in Drinking-Water
Sources Impacted by Wastewater. J. Am. Water Resour. Assoc. 2009, 45,
58–67.
9. Oppenheimer, J.; Eaton, A.; Badruzzaman, M.; Haghani, A.; Jacangelo, J.
Occurrence and Suitability of Sucralose as an Indicator Compound of
132
Wastewater Loading to Surface Waters in Urbanized Region. Water Res.
2011, 45, 4019–4027.
10. APHA, AWWA, and WEF. Standard Methods for the Examination of Water
and Wastewater, Method 6450, Nitrosamines; Washington DC, 2008.
11. Krasner, S. W.; Sclimenti, M. J.; Guo, Y. C.; Hwang, C. J.; Westerhoff, P.
Development of DBP and Nitrosamine Formation Potential Tests for Treated
Wastewater, Reclaimed Water, and Drinking Water. In Proceedings of 2004
American Water Works Association (AWWA) Water Quality Technology
Conference (WQTC); AWWA: Denver, CO, 2004.
12. Krasner, S. W.; Sclimenti, M. J.; Mitch, W. A.; Westerhoff, P.; Dotson,
A. Using Formation Potential tests to Elucidate the Reactivity of DBP
precursors with Chlorine versus with Chloramines. Proceedings of 2007
American Water Works Association (AWWA) Water Quality Technology
Conference (WQTC); AWWA: Denver, CO, November 18−22, 2007.

13. Vanderford, B. J.; Pearson, R. A.; Rexing, D. J.; Snyder, S. A. Analysis of
Endocrine Disruptors, Pharmaceuticals, and Personal Care Products in Water
Using Liquid Chromatography/Tandem Mass Spectrometry. Anal. Chem.
2003, 75, 6265–6274.
14. Vanderford, B. J.; Snyder, S. A. Analysis of Pharmaceuticals in Water
by Isotope Dilution Liquid Chromatography/Tandem Mass Spectrometry.
15. Department of Water Resources Bay-Delta. Delta Simulation Model II --
DSM2;http://baydeltaoffice.water.ca.gov/modeling/deltamodeling/models/
dsm2/dsm2.cfm [cited November 1, 2014]
16. Krauss, M.; Longrée, P.; Dorusch, F.; Ort, C.; Hollender, J. Occurrence and
Removal of N-nitrosamines in Wastewater Treatment Plant. Water Res.
2009, 43 (17), 4381–4391.
17. Chen, B.; Lee, W.; Westeroff, P.; Krasner, S. W.; Herckes, P. Solar Photolysis
Kinetics of Disinfection Byproducts. Water Res. 2011, 44 (11), 3401–3409.
133
Chapter 8
National Occurrence of
N-Nitrosodimethylamine (NDMA)
An Investigation of 38 Australian Drinking Water Supplies
Deborah Liew,1 Julie Culbert,2,3 Kathryn Linge,*,1 Maria José Farré,4,7

Nicole Knight,5 Jim Morran,2 David Halliwell,6 Gayle Newcombe,2
and Jeffrey W. A. Charrois1
1Curtin Water Quality Research Centre, Curtin University, GPO Box U1987
Perth, Western Australia 6845, Australia
2Australian Water Quality Centre, GPO Box 1751, Adelaide,
South Australia 5001, Australia

3School of Agriculture, Food and Wine, The University of Adelaide, PMB 1,
Glen Osmond, South Australia, 5064, Australia

4Advanced Water Management Centre, University of Queensland,
Brisbane, Queensland 4072, Australia

5Smart Water Research Centre, School of Environment, Griffith University,
Southport, Queensland 4222, Australia

6Water Research Australia Limited, Docklands, Victoria 3008 Australia
7Present address: Catalan Institute for Water Research, Scientific and
Technological Park of the University of Girona, Girona, Spain

To date, limited exposure data have been published on

N-nitrosodimethylamine (NDMA) occurrence in Australian
drinking water. Here, we present a comprehensive analysis of
data from the largest survey of NDMA in Australian drinking
water, with a total of 211 samples, from 38 drinking water
treatment plants across five states and one territory. Samples
were collected at the treatment plants after disinfection as well
as in the distribution systems. Only nine water treatment plants
reported NDMA above 5 ng/L. NDMA was detected in more
than half of the samples collected (57 of 87) from these plants,
but all detections were below the Australian Drinking Water

Guideline (100 ng/L). In agreement with other studies, NDMA
was more frequently detected in chloraminated waters than
chlorinated waters. The dosing of certain coagulant aids, and
addition of ammonia prior to chlorine during chloramination
were key factors in the formation of NDMA. Overall the results
show that the drivers for NDMA formation in Australia are
similar to those found in other studies worldwide.
Introduction
The formation of N-nitrosodimethylamine (NDMA) in disinfected water
supplies has attracted worldwide attention, given it has been classified as a potent
carcinogen (1). Health guidelines for NDMA adopted by agencies around the
world range from between 9 and 100 ng/L (2–6). NDMA has been detected in both
chlorinated and chloraminated drinking waters, although higher concentrations
have generally been associated with systems using chloramine as the disinfectant
(7–9). While reported NDMA concentrations in drinking water supplies have
tended to be lower than 10 ng/L (10), higher concentrations (>100 ng/L) have
been observed with increasing residence time in the distribution system (11–13).
Some studies have also reported the presence of NDMA in some raw waters (i.e.
without disinfection) (12, 14). Factors that impact NDMA formation in drinking
water include source water quality and the presence of precursors (15–17), choice
of disinfectant (8, 12, 14), operational parameters such as pH or disinfectant dose
(16), use of ion-exchange resins or certain nitrogen-containing coagulants (7, 18),
and the residence time in distribution systems (11–13).
While some monitoring programs have been established by state water
authorities or water utilities, there are limited published data on NDMA
occurrences in Australian drinking water. This paper consolidates results of
NDMA occurrence surveys collected across five states and one territory in
Australia between 2007-2013, representing the most comprehensive analysis
of NDMA concentrations in Australian drinking water supplies. Survey data
were analysed to determine the impact of water treatment processes on NDMA
formation, and to identify key factors in the formation of NDMA in Australian
drinking water treatment plants (DWTPs).
Methods
Sample Sites
A total of 211 samples were collected from 38 DWTPs around Australia (five
states and one territory) between 2007 and 2013, with the majority (85%) collected
between 2009 and 2011. The samples were collected as part of individual
studies or monitoring programs undertaken by research organisations across
Australia, and data were subsequently combined for this study. The DWTPs
studied included ten from Victoria, three from New South Wales, two from South
Australia, all studied by the Australian Water Quality Centre (AWQC), ten from
136
Western Australia and three from the Northern Territory, studied by the Curtin
Water Quality Research Centre (CWQRC), and ten from Queensland, studied
by the Smart Water Research Centre (SWRC) at Griffith University, and the
Advanced Water Management Centre (AWMC) at the University of Queensland.
While some studies did test for other N-nitrosamines in addition to NDMA, there
were significant variations in analytical method and limits of reporting. Therefore
this paper discusses NDMA occurrence only as it was the only N-nitrosamine
measured in all studies.
In terms of disinfection, 18 of the 38 DWTPs tested used chlorine and 14
DWTPs used chloramine. Six plants used mixed disinfection treatment trains;
four plants used pre-ozonation followed by chlorination or chloramination
(DWTPs 29, 31, 37 and 38) and two plants used UV disinfection in addition
to chlorine or chloramine disinfection for additional microbial protection
(DWTPs 14 and 24). Figure 1 shows the general classification of these DWTPs
according to pre-treatment process before disinfection. Disinfection (chlorination
or chloramination) was the sole treatment process for eight DWTPs (21%),
while four plants (11%) utilised filtration (e.g., granulated activated carbon or
membrane filtration) or sedimentation without coagulation before disinfection.
One disinfection-only plant (DWTP 15) that utilised chloramine is unusual
because the raw water reservoir includes pre-treated groundwater, surface
water and desalinated water, and both chloramination and chlorination are used
in the large distribution system to maintain residual. All other DWTPs (26,
68%) utilised coagulation for treatment, with 15 plants following a relatively
conventional treatment process (coagulation, flocculation, sedimentation,
filtration and disinfection, with pH adjustment and pre-chlorination as required).
Four plants (DWTPs 2, 3, 23, and 30) utilised coagulation and direct filtration
(e.g. membranes or dissolved air flotation and filtration) instead of clarification
by sedimentation. Seven DWTPs had atypical aspects to their pre-treament,
including incorporation of desalinated water (DWTPs 5 and 36), use of ion
exchange resins for NOM removal (DWTP 21), and ozonation throughout the
treatment process (DWTPs 29, 31, 37 and 38). The target chlorine or chloramine
residuals leaving each DWTP were generally between 1 and 3 mg/L, although
three chloramination plants (DWTP 14, 15 and 26) servicing large distribution
systems had treatment plant target residuals of between 3 and 4.5 mg/L.
In addition to sampling, operational data were collected for each plant,
including description of the treatment regime, the use and dose of primary
coagulants and coagulant aids, disinfection dose(s), pH, target disinfectant
residual leaving the plant, and system detention time at the sampling point. In
the case of chloramination, details of the order and time between chlorine and
ammonia addition were also obtained. Where possible, operational parameters
were obtained for the exact dates when samples were collected, although for
some DWTPs or systems little or no data were available and collection of
operational data were challenging for systems where the water authorities
providing distribution system water samples were not responsible for initial water
treatment. Physicochemical characteristics such as pH, total dissolved solids
(TDS), temperature, dissolved organic carbon (DOC) and disinfectant residuals
were also recorded for raw and treated water samples, where possible.
137
Figure 1. General classification of DWTPs surveyed in this study according to

treatment type (percentages reported as percent of total number of drinking
water treatment plants (DWTPs).
Sampling and NDMA Analysis
Table 1. Details of Sampling and Analytical Methods for NDMA for Each
Laboratory
Laboratory AWQC CWQRC QHFSS
Preservation Sodium Ascorbic acid Sodium
Agent thiosulphate 20 mg/L thiosulphate
pentahydrate 80-100 mg/L
120 mg/L
Sample volume 500 mL 1000 mL 1000 mL
Blanks Tap water treated Ultrapure water and Ultrapure water
with UV light quenching agent
SPE cartridge 35-50 mesh LiChrolut® EN and Coconut charcoal
resin activated charcoal Carboxen™ 572
Ionisation mode Positive CI with Positive CI with Positive CI with
(GC-MS) isobutane ammonia ammonia
138
Samples were collected in glass amber bottles or aluminium foil-covered
polyethylene bottles containing a quenching agent, kept cool and in the dark in
an ice box during transport to the laboratory, and then refrigerated in the dark at
4°C until extraction. All laboratories used laboratory blank samples to quantify
contamination introduced through the analytical process. In addition, CWQRC
also included trip and field blanks in each sampling event to determine if there
was any contamination through the sampling process, storage and transport.
Trip blanks remained unopened until analysis, and field blanks were opened at
each sampling location. Samples were then analysed for NDMA at one of three
laboratories, AWQC, CWQRC or Queensland Health Forensic and Scientific
Services (QHFSS). Analysis was carried out using solid-phase extraction (SPE)
followed by gas chromatography-mass spectrometry (GC-MS), with minor
variations between laboratories. Details of small variations between the sampling
protocol and analytical method used by each laboratory are listed in Table 1. The
concentration of NDMA was quantified against a deuterated internal standard
(isotope dilution procedure) and the limits of reporting (LOR) ranged between 1
ng/L and 5 ng/L.
All three laboratories undertook two rounds of inter-laboratory comparison
testing as part of the NDMA survey undertaken by the AWQC and funded by Water
Quality Research Australia (WQRA, now known as Water Research Australia
Ltd). During each round, duplicate samples containing between 10 and 130 ng/L
NDMA were sent to each independent laboratory for analysis. Results obtained
from CWQRC and QHFSS compared well with those of the AWQC. With the
exception of one sample, all results were within 13% relative standard deviation
(RSD) (19).
Data Analysis
All data, excluding blanks and replicates, were pooled into a common dataset.
The normality of the dataset was tested by calculation of the Kolmogorov-Smirnov
Statistic and the Shapiro-Wilk Statistic (SPSS Statistics v20) and the data were
found to have a poor fit to the normal distribution curve. Therefore median was
chosen for data analysis rather than average because it is less affected by outliers
and because it is more suited for data that does not follow a normal distribution.
As the LOR for NDMA analyses varied for each of the testing laboratories, the
highest LOR was chosen as the limit for all samples. Therefore if a result was < 5
ng/L then it was allocated a value of < 5 ng/L (i.e. zero for calculations), and this
will influence percentage detection calculations and other statistics.
Calculation of median concentrations incorporated all data points including
samples reported to be below LOR, and these were assumed to be equal to the
LOR for the purposes of that calculation. While this conservative approach will
overestimate the actual median concentration of chemicals reported below LOR
in more than 50% of samples, it was deemed appropriate given our primary goal
of assessing the safety of treated drinking water. As most samples (>80%) did
not contain NDMA above the LOR, average and median concentrations were also
calculated solely based on samples in which NDMA ≥ 5 ng/L.
139
Overview of Occurrence Data
A total of 211 post-disinfection samples were collected and analysed in this
study. Of these, 95 samples were collected at, or within 2km of, the treatment
plant and 116 were collected in the distribution system. The Australian Drinking
Water Guidelines recommend maintaining a chlorine or chloramine residual
greater than 0.5 mg/L for control of Naegleria fowleri (1) and therefore we
expected disinfectant residual throughout the distribution system. Disinfectant
residual data collected in 27 of the 38 DWTP confirmed this, although the residual
was not always > 0.5 mg/L.
Figure 2. Distribution of NDMA concentrations in all water samples. There was

a higher frequency of detection and higher concentrations in chloraminated
samples than in chlorinated samples.
Figure 2 shows that the majority (80%, n = 169) of all samples contained
NDMA concentrations below 5 ng/L, while 12% (n = 25) contained NDMA
concentrations between 5 and 10 ng/L. As the majority of all disinfected samples
measured NDMA below 5 ng/L, the median concentration of all disinfected
samples was also < 5 ng/L. Of the 211 disinfected samples, 126 were from
chloraminated systems while the remaining 85 were from chlorinated systems.
Figure 2 shows that only 5% of samples (n = 4) from chlorinated systems
contained NDMA above 5 ng/L. On the other hand, 30% (n = 38) of samples
from chloraminated systems were measured with NDMA levels of 5 ng/L or
higher. Chloraminated samples also recorded a much higher maximum NDMA
concentration (74 ng/L) than the chlorinated samples (14 ng/L).
140
Further data analysis demonstrated that all NDMA detects ≥ 5 ng/L were from
nine specific DWTPs, where 48% of all samples (n = 88) contained NDMA > 5
ng/L (Table 2). There was a wide variation in observed NDMA levels, with two
DWTPs (4 and 14) exhibiting significantly higher levels of NDMA (up to 46 ng/L
and 74 ng/L, respectively) than the other plants. The high NDMA concentrations
measured in DWTP 4 and 14 have previously been studied, and were attributed
to factors that include poor source water quality, the method of chloramination
and DWTP management practices (20). Despite these results, all detections were
below the Australian Drinking Water Guideline of 100 ng/L for NDMA.
As shown in Table 2, where NDMA was present at ≥ 5 ng/L, the median
concentration for distribution system samples was generally higher than for
treatment plant samples, with the exception of DWTP 4. In addition, a higher
percentage of distribution system samples contained NDMA compared to the
treatment plant samples (30% versus 8%). This is consistent with results of other
surveys suggesting that NDMA can continue to form and increase with time in
the distribution system (11–13).
Table 2. Summary of the Percentage Detections, Median and Maximum

NDMA Concentrations Observed for Each of the Nine Dwtps That Detected
NDMA in One or More Samples. TP = Treatment Plant; DS = Distribution
System. Results Are Determined Using the Maximum LOR for These Nine
DWTPs (3 ng/L).
% detect
Plant No. and Median Median of Median of
≥ 3 ng/L
Disinfection (maximum) TP samples DS samples
(total no.
type (ng/L) (ng/L) (ng/L)
samples)
2 Chloramine 67% (9) 3 (12) n.a. 3
3 Chloramine 67% (6) 4 (5) 4 4
4 Chloramine 100% (6) 53 (74) 74 39
5 Chloramine 75% (4) 7 (21) <3 15
9 Chlorine 44% (9) < 3 (9) n.a. <3
10 Chlorine 33% (9) < 3 (14) n.a. <3
13 Chloramine 100% (9) 7 (10) 4 8
14 Chloramine 100% (9) 19 (46) 15 23
15 Chloramine 50% (26) < 3 (10) <3 4
n.a. = not available; no TP samples were analysed at these plants.
141
Influence of Operational Parameters on NDMA Formation in Chloraminated
Systems
While there was insufficient data to investigate NDMA formation in

chlorinated DWTPs due to the limited number of detections, the dataset was
further evaluated to assess the factors influencing NDMA formation in the 16
chloraminated systems.
Table 3. NDMA Occurrence and Corresponding Factors (Y = Yes, N =

No) in Chloramination Plants. Plants in Which NDMA Was Detected Are
Bolded. n.d. = not detected, n.a. = data not available
Polymer
Raw Water
Median used as Addition of
Plant No. DOC
NDMA ng/L coagulant ammonia first
>10 mg/L
aids
2 3 Y N n.a.
3 4 Y Y N
4 53 N Y Y
5 7 Y Y Y
6 n.d. N n.a. n.a.
7 n.d. N n.a. n.a.
8 n.d. N n.a. n.a.
12 n.d. Y N N
13 7 Y N Y
14 19 Y Y Y
15 3 N Y N
25 n.d. N N n.a.
26 n.d. N N n.a.
27 n.d. N N n.a.
36 n.d. Y N n.a.
38 n.d. N N n.a.
Relationships between operational parameters and NDMA concentrations

in chloraminated waters were initially assessed by calculation of Spearmans’s
correlation coefficients (SPSS Statistics v20). However, it was found
that correlations were heavily skewed by the significantly higher NDMA
concentrations reported in DWTP 4. Furthermore, the dataset was quite limited
as only seven of the 16 chloraminating DWTPs had detections of NDMA in their
waters. Instead, the dataset was qualitatively evaluated with focus on impact
142
of chloramination practice (order of reagent addition and contact time), use of
nitrogen-containing coagulants and coagulant aids, as well as raw water chemistry
on NDMA formation. Table 3 presents a summary of the key factors and NDMA
concentrations in the 16 DWTPs using chloramination. The influence of these
factors on formation of NDMA are discussed in detail below.
Chloramination Practice
The order of chlorine and ammonia addition during chloramination has

previously been reported to have a significant effect on NDMA formation (18,
21–23). Contact with free chlorine prior to ammonia addition can reduce NDMA
formation, because chlorinated NDMA precursors tend to be less reactive (22–24).
In contrast, the addition of chlorine after ammonia has been shown to enhance
NDMA formation from dimethylamine, a model NDMA precursor, by up to an
order of magnitude (22). This increase was attributed to the localised formation
of dichloramine at the point of chlorine addition, due to high localised Cl:N ratio.
NDMA formation from dichloramine is higher than from monochloramine.
Table 4. Distribution of NDMA Concentrations in Chloraminated Samples

Based on Order of Addition of Chlorine and Ammonia
Order of % detect ≥ 5 ng/L Median Maximum
addition (total no. of samples) (ng/L) (ng/L)
Chlorine first 19% (54) < 5 12
Ammonia first 55% (51) 6 74
The disinfection practice at eight of the 16 chloraminating DWTPs in this

study was to chlorinate prior to adding ammonia, while five DWTPs added
ammonia prior to chlorine. There was no chloramination information available
for the remaining three plants (21 samples). Typically, the time between addition
of ammonia and chlorine ranged from seconds to a few minutes but DWTPs
that added chlorine first employed a longer contact time (2.5-30 hours). Table 4
compares NDMA concentrations for DWTPs that use the two different reagent
addition scenarios. Only two of the eight DWTPs that practise chlorine addition
first, had NDMA detected post-disinfection, while all five plants that practised
ammonia addition first had NDMA detected in one or more of their samples.
Higher NDMA concentrations were also found in waters that were treated with
ammonia first compared to those that were treated with chlorine first. The two
DWTPs with the highest average NDMA concentrations (DWTP 4 and 14) both
added ammonia before chlorine.
143
Coagulation and Coagulation Aids
Most of the DWTPs in this study (25 out of 38) used coagulation as part of
their water treatment process. Alum was the most commonly used coagulant,
but other coagulants such as ferric chloride and Magnasol 489, a mixture of
polyaluminium chlorohydrate and polydiallyldimethyl ammonium chloride
(polyDADMAC), were also used. The primary coagulant dose varied greatly,
ranging from 5 mg/L to 240 mg/L. There was no correlation observed between
the primary coagulant dose and NDMA concentration in the treated water. In
some instances where no coagulant was used, NDMA was still observed in the
final product water suggesting the presence of precursors in the raw water.
Fifteen of the DWTPs also employed various coagulant aids, which varied in
dose, ranging from 0.005 mg/L to 4.7 mg/L. Even though higher coagulant doses
can potentially remove more NDMA precursors, the use of certain coagulant aids
can potentially add NDMA precursors to the water. For example, polyDADMAC,
a nitrogen-based polymeric coagulant has received attention for its reported
propensity to form NDMA during chloramination (18, 25). PolyDADMAC was
used in some water treatment plants in South Australia and Queensland, including
some of the DWTPs presented here.
While there was no clear evidence to suggest that the generic use of coagulant
aids resulted in higher level of NDMA in the disinfected water, when coagulant
aids were utilised, NDMA was only observed when polyDADMAC and/or
Magnafloc LT22 was used in chloraminating DWTPs. NDMA was not detected
in DWTPs that used other coagulant aids (e.g. Flopam AN 910 PWG, Magnafloc
LT20, Magnafloc LT21, Magnafloc LT25, Nalco 44560, Nalco 3482, Magnafloc
LT425, SNF AN 905). Among the six DWTPs utilizing PolyDADMAC, NDMA
was detected in two of the three chloraminating plants between 7-46 ng/L,
but was not detected in any of the three plants using chlorine disinfection.
PolyDADMAC consists of quaternary amines with dimethylamine-functional
groups and it has been proposed that NDMA formation can occur from both
the dimethylamine-moieties (25, 26) or the quaternary amines themselves (27),
which act as NDMA precursors. In this study higher NDMA concentrations
were observed in plants using polyDADMAC, when chloramination was the
disinfection process and not chlorination. This suggests that polyDADMAC
may have an influence in forming NDMA in DWTPs, but only when chloramine
disinfection is used.
Only four DWTPs utilised Magnafloc LT22 as the exclusive coagulant
aid, all of which practiced chloramination, and all of which detected NDMA
at least once. Magnafloc LT22 is a polyacrylamide-based polymer and there
have been no studies to-date relating NDMA formation to its use. However,
certain polyacrylamides used in water treatment have been reported to form
NDMA also through the dimethylamine-moieties in the polymer structure (25,
26). While the exact composition of Magnafloc LT22 was unknown, it is possible
that its structure contains dimethylamine-moieties which can explain the NDMA
formation observed in these plants.
In order to assess the contribution of these nitrogen-based polymers towards
NDMA formation during water treatment, NDMA formation potentials (FP)
144
were determined from different commercially available coagulant aids used
in Australia. Coagulant aids (10 mg/L) were dosed with a high concentration
of monochloramine and analysed for NDMA after seven days of contact time,
based on the procedure by Mitch et al. (28). Of the five coagulant aids tested,
only polyDADMAC and Magnafloc LT22 formed > 10 ng/L NDMA (Table 5).
PolyDADMAC showed an extremely high NDMA FP of 15,000 ng/L, several
orders of magnitude higher than any of the other coagulant aids. While Magnafloc
LT22 had a high NDMA FP (160 ng/L), formation was much lower than that
from polyDADMAC. While both polymers may have the potential to form
NDMA during chloramination, the extent of their impact will depend on specific
water treatment conditions. Doses of chloramine and coagulant aids used in the
DWTPs are much lower than the conditions of the FP procedure. Therefore the
contribution of Magnafloc LT22 to NDMA formation may not be significant; on
the other hand, the use of polyDADMAC could still have significant potential for
NDMA formation under DWTP conditions due to its high FP.
Table 5. NDMA Formation Potentials of Various Coagulant Aids. Coagulant

Aids (10 mg/L, Buffered to pH 7) were Dosed with 2mM Monochloramine
(140 mg/L Cl2) and Analysed for NDMA after Seven Days Contact Time,
Based on the Procedure by Mitch et al. (28). Samples (50 mL) Were Analysed
by Liquid-Liquid Extraction Followed by GC-MS with Ammonia PCI.
Sample NDMA FP (ng/L)
Blank (Ultrapure water) < 2
Flopam AN 910 PWG 6
Magnafloc LT25 6
Magnafloc 800 HP 6
Magnafloc LT22 160
PolyDADMAC 15,000
A previous study of NDMA formation from polyDADMAC compared the

role of the bulk polymer with degraded components or impurities of the polymer
solution, by performing NDMA FP experiments on polyDADMAC before and
after dialysis (29). The dialysis retentate was found to have a much higher NDMA
FP than the dialysis permeate, suggesting that most of the observed NDMA
formation was due to interaction of monochloramine with the retained polymer,
rather than with small degraded molecules or impurities. In another study, it was
reported that polyDADMAC remains as a contaminant in treated water when
it is used as a sole coagulant, but not when it is used in conjunction with alum
(30). In the DWTPs surveyed here, polyDADMAC was typically used with alum,
however, these findings suggest that even higher concentrations of NDMA can be
expected if polyDADMAC is used as the primary coagulant.
145
Raw Water Chemistry
Raw water parameters, including pH, TDS, temperature, DOC and turbidity,
were evaluated for relationships with NDMA concentrations, by calculation of
Spearmans’s correlation coefficients (R). There were no correlations observed
between NDMA concentrations and any raw water parameters, except for DOC.
As shown in Table 3, it appears that higher NDMA concentrations (> 5 ng/L)
were found in plants with higher raw water DOC (> 10 mg/L). Raw water DOC
showed a significant positive correlation with average NDMA concentrations (R
= 0.9, p <0.01, data not shown) However, the relationship was heavily skewed
by the significantly higher NDMA concentrations reported in DWTP 4, and there
was no significant correlation observed when data from DWTP 4 was removed.
Furthermore, six of the eight DWTPs practised some form of conventional
treatment and/or filtration before disinfection, and therefore some DOC removal is
expected before disinfection. Raw water DOC therefore cannot be directly related
to NDMA formation in these cases. Another limitation is that the concentrations
of NDMA and its precursors are typically orders of magnitude lower than DOC
(ng/L compared with mg/L), and may not be directly comparable.
DWTP 4, which did not use any form of coagulation, exhibited the highest
NDMA concentrations. A previous study (20) has also identified that DWTP
4 had poor source water quality, and the treatment process (microfiltration)
was not effective in removing DOC. However, laboratory tests using DWTP 4
source water demonstrated that, even with alum coagulation, NDMA precursors
were not effectively removed compared with total DOC (20). This is consistent
with previous findings that NDMA precursors are usually low molecular weight
hydrophilic compounds (31), and are therefore unlikely to be effectively removed
through conventional treatment processes.
There are clearly limitations in assessing NDMA formation according to bulk
parameters such as DOC and organic nitrogen. Rather, the nature of natural organic
matter in the source water is more important for determining NDMA and and can
vary for each water source. Measurements of NDMA precursors (i.e. NDMA-FP)
may be a more useful indication of the potential of a source water to produce
NDMA under certain water treatment conditions.
Conclusions
The concentration of NDMA in Australian drinking water was variable, but
all detections were below the Australian Drinking Water Guideline of 100 ng/L.
As expected, chloraminating plants were found to have higher concentrations,
and more frequent detections of NDMA than chlorinating plants. The order
of reagent addition during chloramination had a significant impact on NDMA
formation, where the addition of ammonia before chlorine typically led to higher
NDMA concentrations, as compared to addition of chlorine first. Under certain
conditions, some nitrogen-containing coagulant aids may provide precursors for
NDMA formation upon chloramination. Employing free chlorine contact time
before ammonia addition can help to control NDMA formation in chloraminating
systems. These findings are the first demonstration of the influence of operational
146
factors on NDMA formation in Australian DWTPs at the full-scale, and overall
they are consistent with previously reported findings from both laboratory studies
and other drinking water systems worldwide.
Acknowledgments
We acknowledge the following people for their technical assistance: Andrew
Chan and Geoff Chidlow of CWQRC; Con Kapralos, Jules Leach and Najwa
Slyman of AWQC; Neil Holling of QHFSS; Claire McInnes from Water Research
Australia. We thank the following people for their sampling assistance: Liza
Breckler, Ralph Henderson and Fern Burgess of Water Corporation (WA), Deb
Gale of Seqwater and other Seqwater staff. Thanks also goes to Adam Holman for
his study on water treatment polymers, and Glendon Shaw and Kalinda Watson
for research support at SWRC. CWQRC research was funded by the Australian
Research Council (LP110100548), Water Research Australia Ltd and the Water
Corporation of Western Australia (WCWA). AWQC research was part of Water
Research Australia Project 1018-09 and thanks the participating water authorities
. SWRC research was funded by the Urban Water Security Research Alliance.
AWMC acknowledges the tailored collaboration between Seqwater and Water
Research Foundation, and would like to thank Djanette Khiari, manager of Water
Research Foundation Project #4484, and the project advisory committee for their
support. Dr Maria Jose Farre also acknowledges the European Commission for
funding project 623711 under the FP7-PEOPLE-2013-IIF - Marie Curie Action:
“International Incoming Fellowships”.
References
1. IARC, Some N-Nitroso compounds. IARC Monographs on the Evaluation
of Carcinogenic Risks to Humans; International Agency for Research on
Cancer, World Health Organisation: Lyon, France, 1978; Vol. 17, p 365.
2. Boyd, J. M.; Charrois, J. W. A.; Hofmann, R.; Hrudey, S. E. NDMA and other
N-nitrosamines – health risk assessment and management. In Disinfection
By-Products – Relevance to Human Health; Hrudey, S., Charrois, J. W. A.,
Eds.; IWA Publishing: London, 2012; pp 125−144.
3. Ontario Regulation 169/03, Schedule 2, Chemical Standards, Ontario Safe
Drinking Water Act; Ontario Ministry of Environment: Toronto, 2002.
4. Guidelines for Canadian Drinking Water Quality – Summary Table; Water,
Air and Climate Change Bureau, Healthy Environments and Consumer
Safety Branch, Health Canada: Ottawa, Ontario, 2012.
5. Guidance on the Water Supply (Water Quality) Regulations 2000 specific
to N-nitrosodimethylamine (NDMA) concentrations in drinking water;
Drinking Water Inspectorate: London, U.K., 2008.
6. Australian Drinking Water Guidelines 6; National Health and Medical
Research Council, National Resource Management Ministerial Council:
Canberra, Australia, 2011.
147
7. Najm, I.; Trussell, R. R. NDMA Formation in Water and Wastewater. J. Am.
Water Works Assoc. 2001, 93, 92–99.
8. Valentine, R. L.; Choi, J.; Chen, Z.; Barrett, E.; Hwang, C. Factors Affecting
the Formation of NDMA in Water and Occurrence; AWWA Research
Foundation: Denver, CO, 2005.
9. Liew, D.; Linge, K. L.; Heitz, A.; Joll, C. A. Nitrogenous DBPs in drinking
water: toxicity, regulation, analysis, occurrence and control. In Disinfection
By-Products – Relevance to Human Health; Hrudey, S., Charrois, J. W. A.,
Eds.; IWA Publishing: London, 2012; pp 83−124.
10. Mitch, W. A.; Sharp, J. O.; Trussell, R. R.; Valentine, R. L.; Alvarez-
Cohen, L.; Sedlak, D. L. N-nitrosodimethylamine (NDMA) as a drinking
water contaminant: a review. Environ. Eng. Sci. 2003, 20, 389–404.
11. Zhao, Y. Y.; Boyd, J.; Hrudey, S. E.; Li, X. F. Characterization of New
Nitrosamines in Drinking Water Using Liquid Chromatography Tandem

Mass Spectrometry. Environ. Sci. Technol. 2006, 40, 7636–7641.
12. Charrois, J. W. A.; Boyd, J. M.; Froese, K. L.; Hrudey, S. E. Occurrence
of N-nitrosamines in Alberta public drinking-water distribution systems. J.
Environ. Eng. Sci. 2007, 6, 103–114.
13. Charrois, J. W. A.; Arend, M. W.; Froese, K. L.; Hrudey, S. E. Detecting
N-Nitrosamines in Drinking Water at Nanogram Per Liter Levels Using
Ammonia Positive Chemical Ionization. Environ. Sci. Technol. 2004, 38,
4835–4841.
14. Zhao, Y. Y.; Boyd, J. M.; Woodbeck, M.; Andrews, R. C.; Qin, F.; Hrudey, S.
E.; Li, X. F. Formation of N-nitrosamines from Eleven Disinfection
Treatments of Seven Different Surface Waters. Environ. Sci. Technol. 2008,
42, 4857–4862.
15. Asami, M.; Oya, M.; Kosaka, K. A nationwide survey of NDMA in raw and
drinking water in Japan. Sci. Total. Environ. 2009, 407, 3540–3545.
16. Brisson, I. J.; Levallois, P.; Tremblay, H.; Serodes, J.; Deblois, C.;
Charrois, J.; Taguchi, V.; Boyd, J.; Li, X. F.; Rodriguez, M. J. Spatial and
temporal occurrence of N-nitrosamines in seven drinking water supply
systems. Environ. Monit. Assess. 2013, 185, 7693–7708.
17. Gerecke, A. C.; Sedlak, D. L. Precursors of N-Nitrosodimethylamine in
Natural Waters. Environ. Sci. Technol. 2003, 37, 1331–1336.
Berger, R.; Rodigari, F.; Beland, J. W.; Lazzelle, L. J.; Kinicannon, E. G.;
coagulated with DADMAC cationic polymer. J. Am. Water Works Assoc.
2003, 95, 94–106.
19. Newcombe, G.; Morran, J.; Culbert, J.; Slyman, N.; Leach, J.; Kapralos, C.
Occurrence and management of NDMA and other nitrogenous disinfection
by-products in Australian drinking and recycled waters. WQRA Project 1018
Milestone Report No 5. Water Quality Research Australia: Adelaide SA,
2012.
20. Culbert, J.; Morran, J.; Slyman, N.; Smith, B.; Newcombe, G. Occurrence
and Management of Nitrosamines in Australian Drinking and Recycled
148
Water. OzWater’12 Conference Proceedings, Sydney, Australia, May 8−10,
2012.
21. Mitch, W. A.; Oelker, G. L.; Hawley, E. L.; Deeb, R. A.; Sedlak, D. L.
Minimization of NDMA formation during chlorine disinfection of municipal
wastewater by application of pre-formed chloramines. Environ. Eng. Sci.
2005, 22, 882–890.
22. Schreiber, I. M.; Mitch, W. A. Influence of the Order of Reagent Addition on
NDMA Formation during Chloramination. Environ. Sci. Technol. 2005, 39,
3811–3818.
23. Charrois, J. W. A.; Hrudey, S. E. Breakpoint chlorination and free-chlorine
contact time: Implications for drinking water N-nitrosodimethylamine
concentrations. Water Res. 2007, 41, 674–682.
24. Mitch, W. A.; Sedlak, D. L. Formation of N-Nitrosodimethylamine (NDMA)
from Dimethylamine during Chlorination. Environ. Sci. Technol. 2002, 36,

588–595.
25. Park, S. H.; Wei, S.; Mizaikoff, B.; Taylor, A. E.; Favero, C.; Huang, C.
H. Degradation of Amine-Based Water Treatment Polymers during
Chloramination as N-Nitrosodimethylamine (NDMA) Precursors. Environ.
Sci. Technol. 2009, 43, 1360–1366.
26. Mitch, W. A.; Sedlak, D. L. Characterization and Fate of
N-Nitrosodimethylamine Precursors in Municipal Wastewater Treatment
Plants. Environ. Sci. Technol. 2004, 38, 1445–1454.
27. Kemper, J. M.; Walse, S. S.; Mitch, W. A. Quaternary Amines as Nitrosamine
Precursors: A Role for Consumer Products? Environ. Sci. Technol. 2010,
44, 1224–1231.
28. Mitch, W. A.; Gerecke, A. C.; Sedlak, D. L. A N-nitrosodimethylamine
(NDMA) precursor analysis for chlorination of water and wastewater. Water
Res. 2003, 37, 3733–3741.
29. Holman, A. Formation of N-Nitrosamines During Drinking Water Treatment.
Chemistry Honors Dissertation, Curtin University, Perth, Western Australia,
2012.
30. Becker, N. S. C.; Bennett, D. M.; Bolto, B. A.; Dixon, D. R.; Eldridge, R. J.;
Le, N. P.; Rye, C. S. Detection of polyelectrolytes at trace levels in water by
fluorescent tagging. React. Funct. Polym. 2004, 60, 183–193.
31. Pehlivanoglu-Mantas, E.; Sedlak, D. L. Measurement of dissolved organic
nitrogen forms in wastewater effluents: Concentrations, size distribution and
NDMA formation potential. Water Res. 2008, 42, 3890–3898.
149
Chapter 9
The Role of Pre-Oxidation in Controlling

NDMA Formation: A Review
Meric Selbes,1 Malcolm Glenn,2 and Tanju Karanfil*,3
1Hazen and Sawyer, Environmental Engineers and Scientists,

Fairfax, Virginia 22030, U.S.A.
2Alliance Consulting Engineers, Columbia, South Carolina 29202, U.S.A.
3Department of Environmental Engineering and Earth Sciences,
coulombslemson University, Anderson, South Carolina 29625, U.S.A.

N-nitrosodimethylamine (NDMA), a probable human

carcinogen, is a disinfection by-product (DBP) that has been
detected in chloraminated drinking water systems. Options
for NDMA control include the physical removal of either
precursors or precursor deactivation prior to chloramination.
This review summarizes some of the recent findings related
to the control of NDMA formation with commonly used
oxidants in drinking water treatment. Each oxidant (chlorine,
chlorine dioxide, ozone) has the potential to reduce NDMA
formation with varying degree of efficiencies depending on the
precursors in source waters. In general, ozone and chlorine
are effective oxidants for controlling NDMA precursors, likely
due to their high reaction rate constants with amines. In some
cases, the oxidants themselves have been shown to increase
the NDMA formation. The selection of pre-oxidant type, dose
and application location for NDMA control would site-specific.
Utilities intend on using pre-oxidation as a strategy for NDMA
control should also the impacts on the formation of regulated
DBPs and other unintended consequences that can be associated
with those oxidants.

Introduction
N-nitrosodimethylamine (NDMA), a probable human carcinogen, is a
disinfection by-product (DBP) that has been detected in chloraminated drinking
water systems. Due to their adverse health effects, methods to control the
formation of nitrosamines during water treatment have gained increased attention
during the past decade. These methods mainly involve either the removal of
NDMA precursors or their transformation into less reactive forms. Since most of
the nitrosamine precursors have been linked low molecular weight hydrophilic
compounds (1), their removal from water is challenging. Pre-oxidation (e.g.,
using chlorine, ozone, or chlorine dioxide) prior to chloramination can be a viable
approach for some water utilities to control the NDMA levels. However, the use
of oxidants should be optimized for a given source water and water treatment
process configuration since these oxidants are also associated with the formation
of both regulated DBPs and NDMA (without chloramination). The objective
of this chapter is to provide a literature review on the oxidant applications for
controlling NDMA formation during drinking water treatment.
The Role of Oxidants/Disinfectants in NDMA Formation

Several pathways have been proposed for the formation of nitrosamines.
In drinking water, formation of nitrosamines with chloramines is practically
most important pathway. It has been shown that mono- and dichloramines can
lead to formation nitrosamines from DMA, tertiary or quaternary amines with
DMA moieties, natural organic matter, polyelectrolytes, ion-exchange resins,
fungicides, pesticides, herbicides, pharmaceuticals, personal care products, and
wastewater effluent impacted waters. In addition to these chloramines (2–11),.
The molecular structures of precursors mentioned in this review are given in Table
1. Besides chloramine, some other oxidants/disinfectants may lead to NDMA
formation (Figure 1).
Chlorine
The first pathway that has been suggested to form NDMA is through
nitrosation of nitrogen-containing compounds. Nitrosation involves introducing
a nitroso group (-NO) into an organic compound causing the formation of
nitroso compounds. By N-nitrosation of nitrogen-containing organic compounds,
nitrosamines are formed. Possible nitrosating agents include nitrous acid (HNO2),
nitrogen oxides (N2O3, N2O4), and some others (4). For example, the nitrosation
reaction induced by nitrite is as follows: “Under acidic conditions, nitrite (NO2-)
is transformed to nitrous acid (HNO2), which is not stable in aqueous solution
but decomposes either to the nitrosyl cation (NO+) or to dinitrogen trioxide
(N2O3). Both of which are capable of reacting with nitrogen-containing organic
compounds to form nitrosamines (4).
152
Table 1. Some of the Structures of NDMA Precursors
153
Figure 1. NDMA formation and the effect of pre-oxidation.
Choi and Valentine (12) proposed that NDMA is formed by nitrosation

of dimethylamine (DMA) with nitrite that was catalyzed by free chlorine.
Specifically, they attributed NDMA formation to the formation of a dinitrogen
tetraoxide (N2O4) intermediate, which is the product of oxidation of nitrite by free
chlorine. This intermediate is a very effective nitrosating agent with its formation
more favorable at neutral pH, which is typical in water treatment facilities, in
contrast to the formation of N2O3, which occurs at low pH. Although the reaction
is rapid, the yields are approximately two orders of magnitude lower than that of
the chloramination pathway (9). Due to reliance on nitrite for NDMA formation,
this pathway has been suggested for NDMA formation in chlorinated wastewater
effluents and natural-bodied, recreational waters (13, 14).
Chlorine Dioxide
Andrzejewski and Nawrocki (15) have shown that chlorine dioxide reaction
with DMA led to the formation of NDMA. The highest observed yield was 0.2%
molar conversion under very high chlorine dioxide concentrations (i.e., 36 mg/L)
at pH 8.0. Compared to chloramination or other oxidants, however, this much
yield is expected to be negligible under drinking water treatment conditions.
Ozone
Ozonation of DMA forms NDMA but yields generally are <0.02% at neutral
pH (16). The yield of reaction was somewhat higher than 0.02% yet still below
0.4% molar conversion with the increasing pH (16). In another study of NDMA
154
formation, Yang et al. (17) used a nitrosation pathway to generate the formation
of NDMA from DMA at pH level of 3.4. They also observed, through an
unknown pathway, the formation of NDMA during ozonation at a level of pH
greater than 7.0. Subsequent studies of NDMA formation during ozonation
showed that 1,1-Dimethyl Hydrazine (UDMH), daminozide and semicarbazide,
both with UDMH-like functional groups, formed NDMA at yields >50% (5, 18,
19). The ozonation of N,N-Dimethylsulfamide (DMS), a transformation product
of the fungicide tolylfluanide, was observed to form NDMA at a 52% yield
(18). Lastly, the ozonation of polyDADMAC, a polymer used in water treatment
plants also formed NDMA (20). These results clearly indicate that the ozonation
of polyDADMAC can release the DMA moiety while concurrently forming
hydroxylamines. The simultaneous reaction of these two products could form
UDMH, at which point the formed UDMHs would be converted to NDMA in the
presence of ozone.
Permanganate
Andrzejewski and Nawrocki (21) have shown that potassium permanganate

reaction with DMA led to the formation of NDMA. The highest observed yield
was 0.04% molar conversion under very high permanganate concentrations (i.e.,
700 mg/L) and alkali pH conditions (pH>8.35). Compared to the chloramination
or other oxidants, this much lower yield is expected to be negligible under actual
drinking water treatment conditions.
UV Irradiation
Ultraviolet (UV) irradiation of nitrite at <400 nm could form reactive

nitrogen species (22, 23), and those reactive species can react with secondary
amines present in waters. NDMA formation yields were reported less than
0.02% from selected secondary amines with UV-A at 300-400 nm wavelengths
with the highest yields observed at alkali conditions (pH ≈10.0) (22). Similar
findings have also been reported by Soltermann et al. (23). UV-C treatment at
254 nm wavelength (dose of 350-850 mJ/cm2) of chlorinated secondary amines
in the presence of monochloramine increased nitrosamine concentrations at
swimming pools (23). However, this NDMA formation mechanism is unlikely to
be important for drinking waters due to the low prevalence of secondary amines.
Furthermore, UV leads to breakage of N-N bond destroying the NDMA molecule
(24). Additional discussion about NDMA’s removal by UV can be found in the
next section.
The Use of Pre-Oxidants for NDMA Control

The oxidants described below, which are commonly used in drinking water
treatment, have been evaluated in different studies for transforming nitrosamine
precursors.
155
Chlorine
The high reaction kinetics of both chlorine and amines (i.e., NDMA
precursors) makes pre-chlorination an effective strategy for minimizing NDMA
formation (25). However, the use of chlorine also leads to the formation
of halogenated DBPs, of which the trihalomethanes (THMs) and haloacetic
acids (HAAs) are regulated, which may limit the application of chlorine as a
pre-oxidant.
The chlorination of wastewaters with ammonia can also cause the formation
of chloramines, which can react with certain forms of organic nitrogen to produce
NDMA. This phenomenon in turn causes the formation of NDMA in secondary
wastewater effluents (2). Therefore, prior to the application of chlorine for
controlling NDMA formation, the ammonia levels in the water should be analyzed
because high levels of ammonia can hinder the inactivation of NDMA precursors.
Lauer (26) observed this phenomenon while investigating the use of oxidants
for controlling the discharge of NDMA precursors in wastewaters. Presence of
ammonia higher than 0.5 mg/L in wastewater samples was sufficient to negate the
chlorine at doses 5-10 mg/L. Only after the breakpoint chlorination, removal of
NDMA precursors was observed which was achieved at 20 mg/L chlorine dose.
However, the wastewater samples that contained ammonia levels lower than 0.5
mg/L, even 5 mg/L of chlorine dose achieved the removal of NDMA precursors
by 25 to 50% in two minutes of contact time. Higher NDMA precursors removal
(70-95%) with increasing chlorine doses and contact times (26). The distinct
effect of ammonia, however, precluded the mention of studies in this review that
concern unintended chloramination during the pre-chlorination process.
Charrois and Hrudrey (27) investigated two full-scale water treatment plants
as well as some bench-scale experiments with samples collected from those
treatment plants. Their results have showed that 2 mg/L free-chlorine contact of
two hours followed by an additional 2 mg/L chloramine, resulted in lower NDMA
levels (16±3.5 ng/L) compared to no free-chlorine contact time (51±8.3 ng/L). In
these bench-scale experiments, they found that a higher dose of free chlorine (4
mg/L) applied for two hours followed by a 4 mg/L chloramine dose, resulted in
even lower NDMA formation (3±0.7 ng/L). In samples collected from surface
water and the Iowa River (Total Organic Carbon [TOC] = 3.4 mg/L), Chen and
Valentine (28) found that NDMA formation decreased with the application of
free chlorine. The samples were exposed to 0.08 mM of free chlorine at pH 7.0,
which reduced NDMA formation from ~32 ng/L to ~12 ng/L within 5 minutes of
contact time and reached the lowest NDMA formation potential (FP) of ~6 ng/L
after 120 minutes. Furthermore, the amount formed decreased with an increase
pre-chlorination dosage for a fixed free chlorine contact time. One hour contact
time with chlorine doses of 0.035, 0.05, 0.06, and 0.08 mM reduced NDMA
FP to ~26, ~18, ~10 and ~6 ng/L, respectively, from the original NDMA FP of
~33 ng/L (with no chlorine exposure). In their studies, they determined that an
increase in both the chlorine dose and contact time decreased the formation of
NDMA during subsequent chloramination.
In another study, grab samples were collected upstream of any oxidant
addition for 10 source waters (29). Following pre-chlorination, samples were
156
chloraminated for 72 hours under uniform formation conditions (UFC) and then
analyzed for NDMA formation. In general their findings have shown that chlorine
can reduce the NDMA formation by more than 50% within a concentration×time
(CT) of <70 mg×min/L. For instance, pre-chlorination of one of the samples
(pH=8.3, TOC=5.0 mg/L, Br-=170 µg/L, NH4+=0.06 mg/L) reduced NDMA
formation from 25 to 5 ng/L by a CT of 37 mg×min/L (29). They also determined
that the pre-oxidation with chlorine at a low exposure followed by chloramination
increased NDMA formation for three wastewater-impacted source waters. In
all three samples, the NDMA formation increased significantly at the lowest
exposure, representing a 3 minute contact time with chlorine before ammonia
addition, but declined at higher exposures (29). Because all three sample waters
contained some level of nitrite, it was hypothesized that NDMA formation was
perhaps due to the reaction of chlorine with nitrite that formed N2O4 (29), a
nitrosating agent linked to the NDMA formation. It was also possible that the
increase in NDMA at low exposures was because the chlorination converted
NDMA precursors to more potent forms, an effect observed by Chen and
Valentine (28). At higher exposures, however NDMA formation declined, likely
because of the deactivation of these more potent NDMA precursors (29).
In separate study of samples at a water facility, Uzun et al. also reported
a decrease in NDMA formation in chlorine treated waters (30). In the utilities
they surveyed even small doses of chlorine (i.e., 0.5 mg/L) used for maintenance
purposes were found to decrease NDMA FP by an average of 10%. Furthermore,
the application of chlorine prior to the addition of ammonia in clearwells resulted in
a removal rate of 10 to 65%. The minimum and maximum chlorination CT in these
clearwells was 101 mg×min/L and 504 mg×min/L, respectively. However, no
strong correlation was observed between the chlorine CT and NDMA FP reduction
at the full-scale plants surveyed.
Several mechanistic studies have also been undertaken to elucidate the
pre-oxidation of NDMA precursors. These studies can be classified based
on the group of amines investigated: (i) Secondary Amines: The NDMA
formation from the secondary amine, namely DMA, decreased from 1.6% to
1.1% within 5 minutes (CT of 15 mg×min/L) of contact time with chlorine.
This increased chlorine contact time (CT of 180 mg×min/L) did not result in
a further decrease in NDMA formation (19). Previous studies also indicated
the high reactivity of DMA with chlorine (kapp at pH 7 ≈ 104 M-1 s-1), which
caused the formation of chlorinated-DMA (23, 25, 31). Therefore, formed
chlorinated-DMAs could remain in the solution and still form NDMA during
sequential chloramination. (ii) Tertiary Amines: Research performed by Shen and
Andrews (32) examined the impact of pre-chlorination on NDMA formation from
tertiary amines (amine-based pharmaceuticals). Here, eight pharmaceuticals (i.e.
ranitidine, nizatidine, tetracycline, doxylamine, chlorphenamine, carbinoxamine,
diltiazem, and sumatriptan) were examined under contact times of 0.5 to 120
minutes with chlorine ([Cl2]0 = 2.5 mg/L). Their results showed no change in
the rate of NDMA formation from diltiazem upon pre-chlorination. Surprisingly,
however, the chlorination of sumatriptan resulted in a three-fold increase of the
initial yield. Except for those two precursors, the NDMA formation from other
pharmaceuticals decreased by 10 to 90% with an increases in the pre-chlorination
157
contact time (32). In their investigation of the interactions between chlorine
and tertiary amines, Selbes et al. (19) examined the pre-chlorination at different
contact times (5 to 60 minutes) with chlorine ([Cl2]0 = 3 mg/L) at pH 7.5 from
a wide array of NDMA precursors. After chlorination, with the exception of
compounds in both the carbonyl and sulfonyl groups, an overall decrease in the
NDMA formation rate was observed, which was influenced by the selected tertiary
amines. The NDMA yield of the former compounds decreased to approximately
half of the initial yield during pre-chlorination within 5 to 15 minutes, resulting
in a corresponding CT of 15 to 45 mg×min/L (19). (iii) Quaternary Amines:
There are mixed findings for the pre-chlorination of quaternary amines, and the
reasons for those differences remain unclear. Chlorination had no distinct effect
on the NDMA yields from quaternary amines (polyDADMAC, polyamine and
polyACRYL), because the positive charge on the nitrogen atoms of the polymers
hindered the nucleophilic substitution (9, 19). A separate analysis however, did
indicate that the pre-chlorination of polyDADMAC and polyAMINE did decrease
NDMA formation within 10 minutes of contact time (33). The decrease was
attributed to the structural changes of either the polymers or the DMAs, which
made them less reactive to further chloramination.
Overall, the effectiveness of pre-oxidation with chlorine is highly dependent
on the type of precursors, a finding supported by mechanistic studies. There are
also precursors that exhibit no reactivity with chlorine. Indeed, chlorination has
caused an increase in NDMA formation, which occurred i) in the presence of nitrite
from chlorine-triggered nitrosation pathways, or ii) or in the presence of ammonia-
forming chloramines that then directly created NDMAs.
Chlorine Dioxide
Chlorine dioxide is sometimes used in lieu of chlorine in water treatment
processes as it produces fewer chlorinated organic substances, does not form
THMs, and significantly suppresses the formation of HAAs. A potential drawback,
however, associated with the use of chlorine dioxide is that approximately 70%
of the compound is reduced to chlorite (ClO2-), which is a regulated DBP in
the United States. Chlorate (ClO3-) may also form, which while not currently
regulated, is associated with adverse health effects at elevated concentrations (34).
Lee et al. (3) also used chlorine dioxide to investigate the effect of
pre-oxidation on the NDMA FPs in natural waters from the Rhein, Neckar, and
Pfinz rivers in central Europe and in Lake Greifensee, a small lake near Zurich,
Switzerland. In experiments performed under typical drinking water treatment
conditions an initial dose of 2.1 mg/L of chlorine dioxide at pH 7.0 were added
to the samples. Subsequent results showed that the residual chlorine dioxide was
quenched after 5 minutes. The samples were then chloraminated for NDMA FP
analyses, which showed an NDMA FP removal rate ranging from 32 to 94%
depending upon the water sample from a given river. In another study, grab
samples were collected upstream of any oxidant addition for several source waters
(29). Following pre-oxidation, samples were chloraminated for 72 hours under
UFC conditions. After 72 hours, the chloramine residual was quenched, and
the DBPs were analyzed. The results showed that a pre-oxidation with chlorine
158
dioxide exposure (mg×min/L) resulted in a reduction in NDMA formation in
some samples by an average of 50% (29). Pre-oxidation with chlorine dioxide
reduced the NDMA formation to approximately 7-9 ng/L at the highest exposure.
These results, however, were not concurrent with all source waters sampled
in the study. Pre-oxidation with chlorine dioxide increased NDMA formation
by an average of 50%. The water samples with an elevated level of NDMAs
were affected by wastewaters, polyDADMAC polymers, and anion exchange
resin-related precursors, making it likely that pretreatment with chlorine dioxide
converted the precursors in these source waters to more potent forms (29).
In another study, 12 water sources (9 surface water, 1 groundwater, and 2
wastewater effluents) were exposed to pre-oxidation with chlorine dioxide at a
dose of 1.0 mg/L at pH 6.0 to 7.0 (35). After chlorine dioxide pretreatment,
reductions in NDMA FP were observed in most of the samples (with a top range of
~33%). Although and average reduction of 58% was observed in the wastewaters
under study, in other samples NDMA FP either remained relatively constant or
increased (with a top range of ~140%).
In a recent study to determine the effect of pre-oxidation, Uzun et al. (36)
conducted experiments using chlorine dioxide on NDMA FPs in pristine surface
waters blended with wastewaters at different facilities. While they found that
chlorine dioxide both increased and decreased in the levels of NDMA FP,
chlorine dioxide oxidation was more effective in NDMA formation in wastewater
impacted sources than non-impacted sources. Also, the percentage of the NDMA
precursor removed using chlorine dioxide increased with an increase in the
level of wastewater (36). They also observed a rapid oxidation of chlorine
dioxide oxidation (<10 min. contact time), which resulted in a maximum level
of NDMA FP removal in all waters in their study. Lauer (26) investigated the
effect of pre-oxidation in five wastewater samples collected prior to disinfection
and exposed them to 2.5, 5, and 10 mg/L at contact times of 2-60 minutes.
Chlorine dioxide was observed to be the most efficient in removing NDMA
precursors with the highest removal (~90%) at a dose of 10 mg/L. At a dose
of 2.5 and 5 mg/L, however, the removal rate was much greater, ranging from
19% to 90% depending on the dose, contact time and background organic matter
concentrations. These lower reductions at the lower doses was perhaps due to a
competition for the chlorine dioxide between the background organic matter and
the NDMA precursors (26).
There are also model compound studies in the literature that can provide
insight regarding these mixed results regarding chlorine dioxide pre-oxidation in
natural water samples. For example, in their 2007 study to determine the change
rate of NDMA FP in eight model precursors (i.e., DMA, trimethylamine) to 0.5
or 1.0 mM of chlorine dioxide at pH 7.0, they observed no distinct difference
in the DMA in the NDMA FP. However, pre-oxidation with chlorine dioxide of
the other tertiary amines did reduce the presence of NDMA FP from 41 to 63%
(3). In another comprehensive mechanistic study recently conducted by Selbes et
al. (19), they examined the effectiveness of pre-oxidation with chlorine dioxide
([ClO2]0 = 1 mg/L) at different contact times (5 to 30 minutes) and in a pH range
of 7.5 for 15 NDMA precursors, including secondary amines, tertiary amines
and quaternary amines. They observed no obvious reactivity of the DMA to the
159
chlorine dioxide. The exposure of chlorine dioxide to a high level of NDMA
yielded (>5%) precursors, which in turn decreased the formation of NDMA
formation. However, upon contact with chlorine dioxide, tertiary amines (e.g.
methylene blue) with a low NDMA formation (<1%) increased the level of the
NDMA FP. This increase was attributed to a release of the DMA moiety from
the parent compound, which transformed the moiety into a more potent form.
These findings suggest that the origin of the various deactivation efficiencies
of precursors may be attributed to one single major product, namely DMA.
Consequently, the application of chlorine dioxide as a pre-oxidant to control
NDMA formation could be effective in source waters with high levels of NDMA
yielding precursors (i.e., >5%). For source waters containing either DMA or
lower NDMA yielding precursors than DMA as major precursors (i.e., <1%),
however, the application would be redundant (19). These findings are supported
by studies from both Selbes et al. (19) and Park et al. (33), who reported
that the pre-oxidation of quaternary amines (polyDADMAC, polyAMINE, and
polyACRYL) with chlorine dioxide did not affect the overall NDMA yields.
These findings indicate that pre-oxidation with chlorine dioxide can either
increase or decrease NDMA formation. It is more likely that the chlorine dioxide
treatment would be beneficial for utilities if the water does have a wastewater
content. The use of chlorine dioxide may be detrimental, however, for those
utilities with a low level of reactivity precursors in their source waters.
Ozone
Ozone is a strong oxidizing agent that can oxidize many organic and inorganic
compounds in water. While ozone does not lead to the formation of halogenated
DBPs, it can lead to the formation bromate (BrO3-), which is a regulated DBP in
the United States.
Previous studies with ozone have provided some promising results to
reduce NDMA formation, despite some observations reporting enhanced NDMA
formation (3, 19, 27–29, 37, 38). The overall body of research here shows ozone
an effective oxidant for reducing NDMA formation, likely because of the high
reaction rate constants with amines, especially in their deprotonated forms (25).
For example, ozone reduced NDMA formation by over 50% within a very short
contact time (i.e., CT ≤0.5 mg×min/L) (3, 28, 29), with only a few cases of
ozonation actually causing the formation of NDMA (18, 39).
In their study to determine pre-oxidation in the Rhein, Neckar, and Pfinz rivers
and in Lake Greifensee, Lee et al. also investigated the effect of pre-ozonation
on the NDMA FPs in those waters (3). Oxidation treatments were performed
under typical drinking water treatment conditions with 1 and 2 mg/L of ozone
at pH 7.0. The NDMA FPs of the treated natural waters was then measured
after a complete depletion of the ozone, with a removal rate ranging from 32 to
94% depending on the natural water. There was however, no distinct difference
between two ozone doses. In another study to determine the effects of treated
wastewaters, polyDADMAC polymers, and anion exchange resin on ten different
natural systems, Shah et al. (29) observed that ozone similarly reduced NDMA
formation by as much as 50% and at exposures as low as 0.4 mg×min/L.
160
Zhao et al. (40) used ozone as an oxidant to control the formation of NDMA
in two source waters, with the contact time set as the length of time required
to achieve the desired dose (CT = 10 mg×min/L). In both instances, higher
concentrations of NDMA were recorded compare to the untreated source waters
used as the control. The DMA pathway reported by Andrzejewski et al. (16) was
found to be cause of this NDMA formation during ozonation. In a similar study
supporting these finds, Zhao et al. (40) also investigated ozonation followed by
chlorination in two separate water bodies, and observed higher concentrations of
NDMA over the ozone treatment alone.
These results suggested that the ozone oxidation of natural organic matter
may release precursors into the source waters that can react with chlorine to
form NDMA. Although Zhao (40) examined ozonation only and then ozonation
followed by chlorination, the increasing trend involving the use of ozonation
closely agrees with more recent analysis. In one German study detailing how
elevated NDMA formation potentials might affect ozonation in areas of the
country under intense agricultural cultivation, DMS, was detected in ground and
surface waters at concentrations ranging from 100-1000 ng/L and 50-90 ng/L,
respectively (5). Batch ozonation experiments in ultrapure buffered water, surface
water, and tap water found that 52% of the DMS, created from the degradation
of the tolylfluanide fungicide, was converted to NDMA (5). The researchers
also identified the following precursors that exhibited high NDMA conversions
during ozonation: N,N-dimethyl-N′-(4-methylphenyl)-sulfamide, tolylfluanid,
daminozid. The authors also found that a 54% molar conversion rate of the
NDMA yield for an initial concentration of DMS (5). Similarly, von Gunten et
al. (18) examined how bromide, from DMS, might affect NDMA formation.
Although the initial ozonation of DMS in samples of phosphate-buffered ultrapure
water (pH = 7.0) yielded a low rate of NDMA conversions (<3%), the addition of
a few µg/L of bromide substantially enhanced the NDMA formation during the
ozonation process. Notably, this addition of bromide to levels of 15 – 20 µg/L,
which are typical for drinking waters, greatly enhanced the maximum NDMA
yield to approximately 50%.
A similar study, but in Japan, was undertaken to determine the association
of NDMA formation (2.2-10 ng/L) and ozonation in finished drinking waters
from the Yodo River (37). Relatively high concentrations of NDMA were
found after ozonation of water samples taken from wastewater treatment plants
located upstream of the water intake points of the drinking water treatment plants
in the Yodo River basin, and the NDMA formation was associated with the
anti-yellowing agents used by the industries.
In the United States, ozone induced NDMA formation in drinking water
treatment plants have not been observed so far. In 10 water samples, ozone
reduced NDMA formation by over 50% within a very short contact time (i.e.,
CT ≤0.5 mg×min/L) (29). Recently, Lauer (26) investigated the effect of
pre-ozonation in five wastewater samples collected prior to disinfection and
exposed them to 2.5, 5, and 10 mg/L at contact times of 2, 5, and 10 minutes.
Ozone was highly effective at removing NDMA precursors, and reacted very
quickly. A higher than 90% decrease in NDMA FP was observed with a 2.5
mg/L ozone dose and 2 minutes of contact time, with little to no further removal
161
observed after 2 minutes or higher doses of ozone. However, ozonation can lead
to the formation of NDMA in some wastewater discharges (41, 42). In selected
wastewaters the net direct formation of NDMA of 6-33 ng/L was observed after
ozonation by Pisarenko et al. (41). Similar findings, reported by Gerrity et al.
(42) indicated that ozonation may lead to the direct formation of both NDMA
and some other nitrosamines (i.e., N-nitrosomorpholine). A recent survey also
indicates that the de facto use of wastewater in drinking water treatments is
increasing in the US (43). In the light of these studies, ozone triggered NDMA
formation may well occur in US drinking water treatment plants impacted by
wastewaters in the near-future.
A number of mechanistic studies connected ozonation to the significant
removal of NDMA precursors at low CT values (<1 mg×min/L) (3, 19). Other
studies, however, indicate the occurrence of NDMA formation during ozonation,
specifically from model compounds with N,N-dimethylamino functions (including

secondary, tertiary and quaternary amines) (19, 20, 33, 38, 39, 44).
This NDMA formation can occur through the liberation of the DMA moiety
that reacts with ozone to form NDMA; and (ii) through direct a reaction of the
precursor to form NDMA. In their investigation of the pre-oxidation at different
contact times (1 to 10 minutes) with ozone ([O3]0 = 3 mg/L), Selbes et al. (19)
found that NDMA forming compounds had the highest NDMA yields within 1-2
min, and that the overall NDMA formation decreased with an increase in contact
time.
Ozone decomposes spontaneously in water though a complex mechanism
that involves the generation of hydroxyl free radicals (45), which can also
decompose NDMA (19, 46). In experiments both with and without hydroxyl
radical scavengers, Lv et al. (46) identified the transformation products using
mass spectrometry. Results indicated that hydroxyl radicals did indeed decompose
NDMA into DMA and the following nitrogenous compounds: methylamine,
nitromethane, and ammonia. Recently, Selbes et al. (19) has reported that while
ozone may stimulate NDMA formation from some precursors, the simultaneous
production of hydroxyl radicals would work against this effect.
The overall premise, based upon these findings is that ozone has been shown as
an effective method for removing NDMA precursors. It may also, however, cause
NDMA formation without subsequent chloramination. Hydroxyl radicals formed
from the decomposition of ozone would decompose the formed NDMAs, reducing
the overall levels at longer contact times. Indeed, case studies in Germany and
Japan have shown that ozonation may increase NDMA formation, which was
attributed to anthropogenic contaminants. Such cases have yet to be reported in
any US water treatment operation, however.
Other Oxidants/Disinfectants
Other oxidants/disinfectants, such as potassium permanganate (KMnO4),
hydrogen peroxide (H2O2), ferrate, and UV irradiation, are all viable disinfectant
alternatives in drinking water treatment processes. A brief overview of these
oxidants/disinfectants and certain treatment strategies are elucidated in this
section.
162
Chen and Valentine (28) examined the influence of three pre-oxidation
strategies (sun simulation, hydrogen peroxide and potassium permanganate)
on the NDMA formation in the Iowa River to determine any correlation in
specific UV absorbance (SUVA) caused by oxidation. They observed in sample
concentrates (TOC = 3.4 mg/L, pH = 7.0, [NH2Cl]0 = 0.05 mM), a decrease in
NDMA formation for both oxidants. The application of 10 mg/L permanganate
for a contact time of one hour reduced NDMA formation by approximately 50%
compared to samples not subjected to pre-oxidation. Although a notable decrease,
it was achieved with permanganate doses much greater than used in typical
drinking water treatment strategies. Overall, the standard dosage of permanganate
in drinking water treatment protocols neither contributed to NDMA formation
nor did it destroy NDMA precursors.
Ferrate was also the subject of study regarding its use as a pre-oxidant
(in the form of potassium ferrate (K2FeO4)) on natural waters and in

synthetic buffered solutions containing NDMA precursors. Mechanistic
studies with model precursors, specifically DMA and trimethylamine, in
which ferrate (1.0 mM) at pH 7.0 was used to determine the NDMA
FP of precursors (0.1 mM) both with and without pre-oxidation (47).
The pre-oxidation with Fe(VI) for 10 seconds reduced the fast-reacting
compounds by 89–99% (dimethyldithiocarbamate, dimethylaminobenzene,
3-(dimethylaminomethyl)indole, 4-dimethylaminoantipyrine), unlike the
slow-reacting compounds (DMA, trimethylamine, dimethylethanolamine,
dimethylformamide), which did not. There was however a high (95%) NDMA
FP rate of removal for most precursors for all samples subjected to pre-oxidation
for 30 minutes. In this same study, the effectiveness of ferrate on natural water
samples from the Rhein, Neckar, and Pfinz rivers in Germany and Central Europe
was also investigated. A high and a low dose (21 and 1.1 mg/L as Fe) were used
in determining the effectiveness of ferrate in reducing NDMA precursors and
formation. At the higher dose of 21 mg/L, the NDMA FPs of the natural waters
was significantly reduced with an efficiency of 46 – 84%. However, there was no
significant removal of NDMA FP (5 – 25%) for the dose of 1.1 mg/L. In another
study, 12 water sources (9 surface waters, 1 groundwater, and 2 wastewater
effluents) were exposed to pre-oxidation with ferrate at a dose of 20.0 mg/L as Fe
at pH 6.0 to 7.0 (35). The findings of that study showed that that is was possible
to remove nearly 80% of the NDMA precursors after the depletion of the ferrate.
Although no change in NDMA formation was reported in three surface waters, in
several samples the authors reported an increase in NDMA formation of 60%.
In their Iowa River study, Chen and Valentine (28) also tested the viability
of using hydrogen peroxide as a pre-oxidant. Using a concentrate of 3.0 mg/L
of hydrogen peroxide, they reduced NDMA formation by approximately 50%
compared pre-oxidation levels. They also determined that an elevation in e
peroxide concentration to 30.0 mg/L increased the removal of NDMA precursors
to as high as 70%. For both doses of peroxide the minimum NDMA formation
occurred 40 minutes after the introduction of hydrogen peroxide; no further
increases were observed after that time, even after 120 minutes (28). The coupling
of hydrogen peroxide with ozone was expected to increase the removal percentage
of NDMA due to the reaction of hydroxyl radicals with either precursors. This
163
reaction would in turn either decrease precursor reactivity or destroy any formed
NDMAs during ozonation. Indeed, a recent study conducted by Pisarenko et al.
(41), reported that the use of ozone/peroxide was 20% more effective in removing
NDMA precursors than with ozone alone.
Factors Influencing Effectiveness of Pre-oxidants

As the background water chemistry is likely to affect the efficiency of
the oxidants described in the previous section, the influential factors of pH,
temperature, background organics and background inorganics are next described.
pH
Reactions between amines and oxidants are highly sensitive to pH changes

(25), which in turn increases the overall NDMA formation from amine precursors.
Indeed, most of the amines have high pKa values (i.e., DMA’s pKa is 10.6), and it is
well-known that chlorine reacts with deprotonated NDMA precursors. Therefore,
the mid-point of pKa’s of two reactants during chlorination will represent an
optimum pH for the inactivation of NDMA precursors since both species may
coexist at the highest concentrations (25). Similarly, chlorine dioxide and ozone
also reacts more rapidly with deprotonated amines (25). Since these two oxidants
lack pKa’s, the reaction is only limited by the availability of deprotonated amines,
possibly indicating the importance of the oxidation pH in NDMA deactivation.
These mechanistic findings in relevance to NDMA precursors have been reported
by Selbes et al. (19). Specially, by switching the pre-oxidation pH from 5.5 to 8.5
(for chlorine, chlorine dioxide, and ozone), they increased, by ten-fold, the level
of NDMA precursor N,N-dimethyliosopropylamine’s removed within the same
CT of an oxidant. Furthermore, a recent investigation of the pre-oxidation with
chlorine dioxide showed the importance of pH in surface waters and wastewater
impacted waters (36). In that study, an increase in the pre-oxidation pH from
6.0 to 7.8 increased the removal of the NDMA precursor removal by 20 to 40%.
These findings indicate the importance of pH during pre-oxidation of NDMA
precursors.
Temperature
It is well-know that the reactivity (i.e., reaction kinetics) of pre-oxidants is

dependent on the temperature. A study conducted by Krasner et al. (48) showed
that chlorine was more effective in destroying NDMA precursors at 25°C than
at 5°C. This finding indicates that the efficiency of precursor removal by pre-
oxidation may change throughout the year depending on the season. Therefore,
utilities may need to develop different strategies control their NDMA formation
especially during cold months.
164
Background Ions
As discussed previously, nitrite also reacts with chlorine to form NDMA
without chloramination (14, 26, 29). The presence of ammonia can also cause the
formation of chloramines which would decrease the efficiency of pre-oxidation
result in the direct formation of NDMA (2, 26). The presence of background
ammonia may also shift the chloramine speciation which would influence the
overall NDMA formation (10). Bromide can also enhance the formation of
NDMA during chloramination at concentrations higher than 400μg/L (8, 29).
Furthermore, ozone can also react with bromide and act as a catalyst for the
formation of NDMA from DMS at low concentrations (15-20 μg/L) (5, 18).
Background Organics
Background organics can compete for chloramine species and decrease the
NDMA formation from some NDMA precursors (10). Furthermore, it is expected
that the presence of background organic matter would create competition with
NDMA precursors for the pre-oxidants. Consequently, this competition would
affect the overall efficiency and the kinetics of NDMA precursor inactivation. This
impact can be clearly seen during the pre-oxidation of surface waters spiked with
NDMA precursors in which the removal of NDMA precursors have been reported
to be lower compared to lab grade water (3, 32). This hindrance was also observed
during pre-oxidation with chlorine dioxide in waters (30) and wastewaters (26).
However, there has been no systematic investigation to determine the effect of
organic matter, on NDMA formation especially during pre-oxidation of NDMA
precursors.
The Use of UV for NDMA Control

The use of UV treatment at doses less than 100 mJ/cm2 has yielded a minimal
effect on NDMA precursors in both surface waters and wastewaters, with a
10 to 20% decrease in NDMA formation at advanced oxidation doses (<2000
mJ/cm2) (26, 29). In another study, the effects of oxidation of treatment water by
various simulated sunlight (250 W/m2) up to 60% removal of NDMA precursors
was achieved within 2 hours (28). Both studies suggest the poor efficacy of
UV treatment in mitigating NDMA precursors under typical drinking water
disinfection conditions. While UV can also cause NDMA formation (22, 23), this
has yet to occur in a full-scale water treatment operation.
Although UV treatment is seemingly ineffective for NDMA precursor control,
it is most effective for destroying NDMA that are present. NDMA removal by
sunlight has been shown to ranges from 12% to 65% within 1.5 hours and almost
complete NDMA removal was observed after 3.0 hours (49). Moreover, low
pressure UV produces light in a narrow band of 254 nm wavelength which is
between the NDMA absorption bands, and medium pressure UV lamps produce
wavelength in the range of 200 to 270 nm which also covers the NDMA adsorption
bands. NDMA has an adsorption peak at 228-nm wavelength and strongly absorbs
UV light from 185 to 275 nm. It also weakly absorbs from 330 to 400 nm (24).
165
Therefore, UV light between the absorption bands will lead to the breakage of N-N
bond destroying the NDMA molecules (24). Due to the low (nanograms per liter)
concentrations of NDMA desired in the distribution system, UV exposure must be
operated at higher doses (i.e., 600-800 mJ/cm2) than typically used for disinfection
(i.e., 100 mJ/cm2) (24). In a study, 90% reduction of NDMA was achieved at
1000 mJ/cm² of UV exposure, approximately 10 times higher than that required
for virus removal. NDMA reduction using UV irradiation is technically feasible
but expensive (50, 51). The principal by-products of UV photolysis of NDMA are
DMA and nitrite (50, 52). The released DMA’s could react with residual chlorine
or chloramine and continue to form NDMA. However, the yield of this reaction
is relatively low (≤3.0%). Consequently, UV exposure (i.e., in clearwells) is an
effective strategy for controlling the NDMA levels in distribution systems.
Conclusions
As a review of the studies in this chapter clearly show, the use of pre-oxidation
prior to disinfection with chloramines is a promising NDMA control strategy
for some utilities. However, each oxidant has its own associated regulated DBP
formation and has a varying degree of efficiency in reducing NDMA formation.
Key findings from the literature are summarized in Table 2.
Although an increase of the CT of chlorine prior to chloramination did
decrease the amount of NDMA formation, utilities employing this method of
precursor control must be cautious, as regulated carbonaceous-DBPs (i.e., THMs
and HAAs) may result. In studies of several wastewater sources, pre-chlorination
caused an increase in NDMA formation due to the nitrosation pathway facilitated
by the presence of nitrite. Even though ozonation is very effective in destroying
NDMA precursors at low CT values, ozonation still leads to an increase in NDMA
via direct formation from specific precursors (i.e., amides). Although chlorine
dioxide is a potential mechanism for controlling NDMA formation, it may
increase the overall NDMA formation-like ozone depending on the precursors
in the source water. Both chlorine dioxide and ozone may convert precursors
into more potent forms (by liberating the DMA moiety). Thus, there is no single
oxidant that is effective for controlling NDMA in all water types, but rather
their effectiveness is highly dependent upon the characteristics of the existing
precursors in source waters. Furthermore, each of those oxidants is associated
with a regulated DBP formation (THMs, HAAs, chlorite, bromate), which results
in a trade-off between the destruction of NDMA precursors and other DBPs
formation. The other oxidants/disinfectants including permanganate, ferrate,
peroxide and UV, did, however exhibit a negligible effect on NDMA precursors.
Of these, UV was found quite effective in the destruction of formed NDMAs.
The selection of pre-oxidant type, dose and application location for NDMA
control will be site-specific. Managers in water utilities that intend to use pre-
oxidation as a strategy for NDMA control should also assess the effects on the
formation of regulated DBPs as well as other unintended consequences inherent
with the use of oxidants.
166
Table 2. Summary of Key Findings from the Literature
Pre-Oxidants/ Disinfectants Effect on NDMA Precursors Other
Reaction with
NDMA Formation DBPs of
NDMA
Source Waters Model Compounds Concern

Direct NDMA · Chlorine can reduce the NDMA ·High reactivity with DMA to
formation has been formation by more than 50% within form chlorinated DMAs (19, 23,
reported during a concentration×time (CT) of <70 25, 31).
chlorination in mg×min/L (27, 29, 30). Even at ·Effectiveness depends on the
the presence of low doses (0.5 mg/L) used for tertiary amines structure. May
THMs
Chlorine nitrosating agents maintenance purposes 10% NDMA lead to increases, decreases NA
HAAs
(4) but unlikely to be FP removal can be achieved (30). or have no effect in NDMA
significant compared · May lead to increased NDMA formation (19, 32).
to chloramination formation in the presence of ·No change or some decreases
in drinking water nitrite which may be a concern in in NDMA FP from quaternary
167
treatment conditions. wastewaters impacted sources (29). amines (19).

Direct NDMA
· No reactivity with DMA (3, 19).
formation has been
· NDMA precursors can be removed · Effectiveness depends on the
reported (15) but
in the range of 19-90% (3, 26, 36). tertiary amines structure. May
unlikely to be Chlorite
Chlorine Dioxide · May lead to increased NDMA lead to increases or decreases in NA
significant compared Chlorate
formation by an average of 50-140% NDMA formation (3, 19).
to chloramination
(29, 35, 36). · No reactivity with quaternary
in drinking water
amines (19, 33).
treatment conditions.

Table 2. (Continued). Summary of Key Findings from the Literature
Reaction with
Source Waters Model Compounds NDMA

Concern
· Ozonation can remove NDMA Ozone does
Direct NDMA · Can reduce NDMA formation by
precursors at low CT values (<1 not react
formation has been 32-94% within a very short CT (≤2
mg×min/L) (3, 19). with NDMA;
reported and it can mg×min/L) in both drinking waters
· NDMA can be formed during however, •OH
be comparable to and wastewaters (3, 26, 28, 29).
Ozone ozonation of precursors with formed during Bromate
chloramination · Some NDMA formation has been
N,N-dimethylamino functions ozonation can
in drinking water observed upon ozonation in surface
(including secondary, tertiary and decompose
treatment conditions waters (37) and wastewater effluents
quaternary amines) (19, 20, 38, NDMA (19,
(18, 19, 37, 39). (41, 42).
39, 44). 46).
168
Direct NDMA
formation has been
reported (21) but · Not effective at doses relevant to
Potassium unlikely to be drinking water treatment (28).
NA NA NA
Permanganate significant compared · Can remove ~50% of NDMA FP
to chloramination at high doses (10 mg/L) (28).
in drinking water

Reaction with
Source Waters Model Compounds NDMA
Concern
· Not effective at doses relevant to

· Effectiveness depends on the
drinking water treatment (1 mg/L as
tertiary amines structure. May
Fe) (35).
Potassium Ferrate NA decrease NDMA formation or NA NA
· Can remove 46-84% of NDMA FP
have no effect on the precursor
at high doses (21 mg/L as Fe) (35,
(47).
47).
UV leads to
Direct NDMA
· Relatively ineffective over breakage of
formation has been
exposures relevant to disinfection N-N bond (24).
reported (22, 23)
(<100 mJ/cm2) (26, 29). Moreover, •OH
but unlikely to be
UV Irradiation · 10-30% reduction of NDMA NA formed with UV NA
significant compared
169
formation for UV fluence relevant to (if applicable)

to chloramination
advanced oxidation (>1000 mJ/cm2) can decompose
in drinking water
(26, 29). NDMA (19,
46).
NA: Information not available.

Acknowledgments
This review and some of the work presented in this review was supported, in
part, by a research grant from the National Science Foundation (CBET 106657).
However, the manuscript has not been subjected to peer and policy review of the
agency and therefore does not necessarily reflect its views.
References
1. Chen, Z.; Valentine, R. L. Environ. Sci. Technol. 2007, 41, 6059–6065.
2. Mitch, W. A.; Sedlak, D. L. Environ. Sci. Technol. 2002, 36, 588–595.
3. Lee, C.; Schmidt, C.; Yoon, J.; von Gunten, U. Environ. Sci. Technol. 2007,
41, 2056–2063.
4. Sacher, F.; Schmidt, C. K.; Lee, C.; Von Gunten, U. Strategies for minimizing
nitrosamine formation during disinfection; Water Research Foundation
Report: Denver, CO, 2008.
5. Schmidt, C. K.; Brauch, H. J. Environ. Sci. Technol. 2008, 42, 6340–6346.
6. Mitch, W. A.; Krasner, S. W.; Westerhoff, P.; Dotson, A. 2009. Occurrence
and formation of nitrogenous disinfection by-products; Water Research
Foundation Report: Denver, CO, 2009.
7. Shen, R.; Andrews, S. A. Water Res. 2011, 45, 944–952.
8. Le Roux, J.; Gallard, H.; Croué, J.-P.; Papot, S.; Deborde, M. Environ. Sci.
Technol. 2012, 46, 11095–11103.
Water Res. 2013, 47, 4433–4450.
10. Selbes, M.; Kim, D.; Ates, N.; Karanfil, T. Water Res. 2013, 47, 945–953.
11. Liu, Y. D.; Selbes, M.; Zeng, C.; Zhong, R.; Karanfil, T. Environ. Sci.
Technol. 2014, 48, 8653–8663.
12. Choi, J.; Valentine, R. L. Environ. Sci. Technol. 2003, 37, 4871–4876.
13. Walse, S. S.; Mitch, W. A. Environ. Sci. Technol. 2008, 42, 1032–1037.
14. Shah, A. D.; Mitch, W. A. Environ. Sci. Technol. 2012, 46, 119–131.
15. Andrzejewski, P.; Nawrocki, J. Water Sci. Technol. 2007, 56, 125–131.
16. Andrzejewski, P.; Kasprzyk-Hordern, B.; Nawrocki, J. Water Res. 2008, 42,
863–870.
17. Yang, L.; Chen, Z. L.; Shen, J. M.; Xu, Z. Z.; Liang, H.; Tian, J. Y.; Ben, Y.;
Zhai, X.; Shi, W. X.; Li, G. B. Environ. Sci. Technol. 2009, 43, 5481–5487.
18. von Gunten, U.; Salhi, E.; Schmidt, C. K.; Arnold, W. A. Environ. Sci.
Technol. 2010, 44, 5762–5768.
19. Selbes, M.; Kim, D.; Karanfil, T. Water Res. 2014, 66C, 169–179.
20. Padhye, L.; Luzinova, Y.; Cho, M.; Mizaikoff, B.; Kim, J.-H.; Huang, C.-H.
21. Andrzejewski, P.; Nawrocki, J. Water Res. 2009, 43, 1219–28.
22. Lee, C.; Yoon, J. J. Photochem. Photobiol. A 2007, 189, 128–134.
23. Soltermann, F.; Lee, M.; Canonica, S.; von Gunten, U. Water Res. 2013, 47,
79–90.
170
24. Sedlak, D.; Kavanaugh, M. Removal and Destruction of NDMA and NDMA
precursors during wastewater treatment; Water Reuse Foundation Report:
Alexandria, VA, 2006.
25. Lee, Y.; von Gunten, U. Environ. Sci. Technol. 2010, 44, 555–566.
26. Lauer, J. P. The use of oxidants for NDMA precursor deactivation in
wastewater treatment. Master’s Thesis, Clemson University, Clemson,
South Carolina, 2014.
27. Charrois, J. W. A.; Hrudey, S. E. Water Res. 2007, 41, 674–682.
28. Chen, Z.; Valentine, R. L. Environ. Sci. Technol. 2008, 42, 5062–5067.
29. Shah, A. D.; Krasner, S. W.; Lee, C. F. T.; von Gunten, U.; Mitch, W. A.
30. Uzun, H.; Kim, D.; Ates, N.; Karanfil, T. In Proceedings of AWWA Annual
Conference and Exposition, Dallas, TX, 2012.
31. Deborde, M.; von Gunten, U. Water Res. 2008, 42, 13–51.
32. Shen, R.; Andrews, S. A. Water Res. 2013, 47, 2446–2457.
33. Park, S. H.; Padhye, L. P.; Wang, P.; Cho, M.; Kim, J. H.; Huang, C. H. J.
Hazard. Mater. 2015, 282, 133–140.
34. Alam, M.; Cantwell, R.; Hofmann, R.; Andrews, R.; Rand, J.; Gagnon, G.;
VanderMarck, M.; Moffat, E.; Andrews, S. J. Environ. Eng. 2008, 134,
478–485.
35. Yang, X.; Guo, W.; Zhang, X.; Chen, F.; Ye, T.; Liu, W. Water Res. 2013, 47,
5856–5864.
36. Uzun, H.; Kim, D.; Karanfil, T. In Proceedings of AWWA Water Quality and
Technology Conference, New Orleans, LA, 2014.
37. Kosaka, K.; Asami, M.; Konno, Y.; Oya, M.; Kunikane, S. Environ. Sci.
Technol. 2009, 43, 5236–5241.
38. Oya, M.; Kosaka, K.; Asami, M.; Kunikane, S. Chemosphere 2008, 73,
1724–1730.
39. Asami, M.; Oya, M.; Kosaka, K. Sci. Total Environ. 2009, 407, 3540–3545.
40. Zhao, Y. Y.; Boyd, J. M.; Woodbeck, M.; Andrews, R. C.; Qin, F.; Hrudey, S.
E.; Li, X. F. Environ. Sci. Technol. 2008, 42, 4857–4862.
41. Pisarenko, A. N.; Stanford, B. D.; Yan, D.; Gerrity, D.; Snyder, S. A. Water
Res. 2012, 46, 316–326.
42. Gerrity, D.; Pisarenko, A. N.; Marti, E.; Trenholm, R. A.; Gerringer, F.;
Reungoat, J.; Dickenson, E. Water Res. 2014, 72, 251–261.
43. Rice, J.; Wutich, A.; Westerhoff, P. Environ. Sci. Technol. 2013, 47,
11099–11105.
44. Marti, E. J.; Pisarenko, A. N.; Peller, J. R.; Dickenson, E. R. V. Water Res.
2014, 72, 262–270.
45. Hoigné, J.; Bader, H. Water Res. 1983, 17, 173–183.
46. Lv, J.; Li, Y.; Song, Y. Water Res. 2013, 47, 4993–5002.
47. Lee, C.; Lee, Y.; Schmidt, C.; Yoon, J.; Von Gunten, U. Water Res. 2008, 42,
433–441.
48. Krasner, S. W.; Lee, C. F. T.; Liang, S.; Mitch, W.; von Gunten, U.;
Westerhoff, P. In Proceedings of the AWWA Annual Conference and
Exposition, Dallas, TX, 2012.
171
49. Soroushian, F.; Shen, Y.; Patel, M.; Wehner, M. In Proceedings of AWWA
Annual Conference and Exposition, Washington, DC, 2001.
Cohen, L.; Sedlak, D. L. Environ. Eng. Sci. 2003, 20, 389–404.
51. Guidelines for Canadian Drinking Water Quality: Guideline Technical
Document — N-Nitrosodimethylamine; Health Canada, Ottawa, ON, 2011.
52. Bolton, J. R.; Stefan, M. I. In Proceedings of AWWA Water Quality and
Technology Conference, Salt Lake City, UT, 2000.
172
Chapter 10
pH Effect on Nitrosamine Precursor Removal

by Activated Carbon Adsorption
Chao Chen,*,1,2,3 David Hanigan,4 Xiaobin Liao,1,5 Jun Wang,1
Xiaojian Zhang,1 I. H. (Mel) Suffet,3 Stuart W. Krasner,6
and Paul Westerhoff4

1School of Environment, Tsinghua University, Beijing 100084, China
2State Key Joint Laboratory of Environment Simulation and Pollution
Control (SKLESPC), Beijing 100084, China
3Department of Environmental Health Sciences, School of Public Health,
University of California at Los Angeles,

Los Angeles, California 90095, U.S.A.
4School of Sustainable Engineering and the Built Environment,
Arizona State University, Tempe, Arizona 85287-3005, U.S.A.

5College of Civil Engineering, Huaqiao University, Xiamen 361021, China
6Metropolitan Water District of Southern California, 700 Moreno Avenue,
La Verne, California 91750, U.S.A.

Powdered activated carbon (PAC) adsorbs certain nitrosamine

(NA) precursors from water. We hypothesized that varying
pH levels influence the extent of NA precursor removal. For
two model NA precursors (ranitidine and chloropheniramine)
and ambient NA precursors present in a treated wastewater
(WW), removal increased on PAC at higher pH. However,
the removal of NDMA formation potential (FP) from an
aquaculture-impacted lake water by PAC declined at higher
pH. The PAC adsorption results agreed with experimental
findings from the Polarity Rapid Assessment Method (PRAM)
for characterizing WW-derived NDMA precursors. The
differences in NDMA precursor removal by PAC adsorption
at varying pH between different waters was likely due to the
different charactersitcs of NDMA precursors in the two water
samples.

1. Nitrosamine Precursor Removal by Activated Carbon
Adsorption
Several nitrosamines (NAs) are disinfection by-products, found in both
wastewater (WW) and drinking water all over the world (1, 2). The present
N‐nitrosodimethylamine (NDMA) regulatory and notification levels in Ontario,
Canada (9 ng/L), Massachusetts and California, U.S.A. (10 ng/L) present major
challenges for drinking water treatment plants, especially for those with high NA
formation potential (FP). A major source of NDMA precursors is from treated
WW discharges (2). Some of these precursors can be attributed to the presence of
pharmaceuticals and personal care products (PPCPs) (3).
Conventional coagulation, sedimentation, and filtration processes result
in very limited removal of nitrosamine precursors (4). The basic strategies to
control nitrosamine formation in water include precursor removal with powdered

or granular activated carbon (PAC, GAC) (5, 6), destruction of precursors with
pre-oxidation (6, 7), and/or optimization of the chloramination conditions (8).
AC adsorption is a practical process to remove non-polar or weakly polar
organic contaminants of various molecular weights (MW), including taste- and
odor-causing compounds, natural organic matter (NOM), endocrine disrupting
chemicals (EDCs), PPCPs, pesticides, and precursors for halogenated disinfection
by-products (DBPs). As PAC or GAC are considered for meeting multiple water
quality objectives, it is important to understand factors influencing AC adsorption
of nitrosamine precursors.
NA precursor removal by AC was recently studied due to its possible future
regulation. Hanigan et al. demonstrated the potential of GAC and PAC to adsorb
NDMA precursors from blends of river water and secondary effluent from a
wastewater treatment plant (WWTP) (5). They suggested that “further research
is needed to better understand NDMA precursor chemical properties”. Liao et
al. reported that GAC was effective at removing several NA precursors (e.g.,
median of 88% for NDMA FP and 83% for N-nitrosodiethylamine [NDEA FP])
in a source water in China highly impacted by aquaculture (6). However, these
investigations presented the removal efficacy or treatability of NA precursors by
AC adsorption without discussion of the influence of key water quality factors.
Moreover, AC adsorption did not always result in satisfactory removal
of NDMA FP under certain conditions. According to Krasner et al. (9)
and Hanigan et al. (10), the addition of the amine-containing polymer
polydiallyldimethylammonium chloride (polyDADMAC) contributed NDMA FP
to the water that was recalcitrant to AC adsorption . Such an amine-containing
polymer is of high charge density. Another unusual performance of AC adsorption
occurred at one drinking water treatment plant (DWTP) in the U.S. where the
water was lime softened at pH 11 and PAC was applied (9). The NDMA FP
removal by AC was less than 20%. However, in subsequent sample events,
NDMA FP removal was 42-58%. During the first sample event, the source water
was not as wastewater-impacted as it was in subsequent events.
A previous study showed that the basic structure of NA precursors in
secondary and tertiary effluent from a WWTP included a protonated amine
functional group and a generally non-polar moiety (11). We hypothesized that
174
the potential for the amine functional group to be protonated could account
for sensitivity to pH, whereas the non-polar moiety was responsible for strong
adsorption by AC.
The efficiency of AC adsorption is affected by the interactions between the
AC, the adsorbate, and the solvent. Thus, the influencing factors for NA precursor
adsorption include AC surface characteristics, precursors properties, and the pH
of the solvent (water). Therefore, we conducted experiments with different types
of AC, different types of NA precursors, and at different pH levels in water.
Calgon WPH PAC (10) was used in this research in the laboratory at Arizona
State University and the pulverized Calgon F400 GAC was used in the laboratory
at Tsinghua University. The two ACs presented similar properties according to the
tests reported in Table 1.
Jar tests were conducted for AC adsorption experiments. The jar test was
conducted at 100 rpm. After the addition of certain amounts of PAC, the timer was
started. After certain spans of adsorption time, the jar test was stopped. The PAC
was removed from the solution by glass-fiber filtration. The residual concentration
of known NDMA precursors or NDMA FP and other water quality indices of the
water samples were measured.
Table 1. Properties of Activated Carbons Used in This Study

Micro- Meso- Macro-
Methyl- Pore
Carbon Iodine BET pore pore pore
ene blue volume
name number (m2/g) volume volume volume
number (mL/g)
(mL/g) (mL/g) (mL/g)
WPH 800 1185 0.840 0.412 0.329 0.099
F400 1071 225 1077 0.667 0.380 0.210 0.077
Note: Micropore is defined as having a diameter of 0 to 2 nm; mesopore and macropore
from 2 to 50 nm and >50 nm, respectively. BET = Brunauer–Emmett–Teller theory specific
surface area.
2. pH Effect on NDMA FP Removal by PAC Adsorption in

Different Water Samples
Two water samples were used to investigate the effect of pH. One was a
blended river/wastewater obtained by mixing secondary effluent of a WWTP (20%
volume) and river water (80% volume), both from central Arizona, U.S.A. The
other was taken from an aquaculture-impacted lake in the Yangtze River Delta,
China. The water quality of these samples is presented in Table 2.
175
Table 2. Basic Water Quality of Water Samples
DOC UV DON NDMA FP NDEA FP
Water Tested
(mg/L) (cm-1) (mg N/L) (ng/L) (ng/L)
Blended (80/20)
river/wastewater 3.7 0.073 0.69 195 ND
(Arizona)
Aquacuture-
impacted lake
4.8 0.106 0.33 198 106
water (Yangtze
River Delta)
Note: DOC = Dissolved organic carbon, DON = dissolved organic nitrogen, UV =
ultraviolet absorbance at 254 nm. DON = TDN - NH3-N - NO3-N - NO2-N, where TDN is
total dissolved nitrogen. NDMA and NDEA FP were measured by the method described
in Mitch et al. (12). ND = Not detected.
2.1. Tests with the Blended Wastewater
Tests were conducted at pH 8.2; which was a common drinking water

pH level; at 9.5 and 11, which represent the range of pH levels used for lime
softening (calcium and magnesium removal, respectively); and at an acidic pH of
3, albeit lower than that used in drinking water. PAC at 8 mg/L was evaluated as
a representative dose. In addition, at some pH levels, one or two other PAC doses
were evaluated (3 and 20 mg/L).
Figure 1a shows that PAC effectively removed wastewater-derived NDMA
precursors under alkaline conditions. However, removal was greatly impaired
under acidic conditions for the one PAC dose (8 mg/L) evaluated.
With a PAC dose of 8 mg/L, the removal of NDMA FP at pH levels of 3.0,
8.2, 9.5, and 11.2 was 29, 54, 58, and 52%, respectively. At pH 8.2 and 11.2, the
removal increased with increasing PAC dose for the three doses evaluated. For
example, at pH 11.2, the removal with PAC doses of 3, 8, and 20 mg/L was 25, 52,
and 71%, respectively. Note, NDMA FP removal did not increase proportionately
with increasing PAC dose. It is likely that only a certain portion of the precursors
were adsorbable, so there was a limit to how much the NDMA FP could be lowered
with PAC. Note, the negative removal value at pH 8.2 and PAC dose of 3 mg/L
was probably due to variability in NDMA FP determination and NDMA analysis
(allowed relative standard derivation for analsysis was ±20% according to USEPA
method 521 (13).
The adsorption capacity under different doses was calculated to clarify the
influence of pH, as shown in Figure 1b. The adsorption isotherm points of PAC,
i.e., Qe at near neutral (pH 8.2) and alkaline (pH 9.5 and 11.2) pH were similar,
whereas at the acidic pH of 3, the adsorption decreased over 50%. The adsorption
capacity at pH 3.0 with Ce =138 ng/L was 7.1 ng NDMA FP/mg PAC. The capacity
at pH 11 was 16.3 ng NDMA FP/mg PAC, according to the simulated Freundlich
isotherm: Qe = 0.193*Ce0.900.
176
The efficacy for removing bulk organic matter (i.e., DOC and UV254) was
typically less than that of the nitrosamine precursors (Figures 1c-1d). Hanigan
et al. (5) observed the same phenomenon, which suggests that the nitrosamine
precursors were trace organic contaminants (e.g., anthropogenic chemicals) that
were better adsorbed than the bulk NOM. More importantly, the influence of pH
on bulk organic matter removal was opposite that of the nitrosamine precursors.
Acidic conditions tended to favor the adsorption of bulk organic matter due to
hydrogenation of acid functional groups on the NOM.
Figure 1. Removal of NDMA FP and bulk organic matter in blended

river/wastewater with WPH PAC at varying pH levels and PAC doses (Note:
Initial NDMA FP = 195 ng/L, adsorption time = 1 h, temperature = 11°C.
Qe is the adsorption capacity of NDMA FP with PAC. Ce is the equilibrium
concentration in the water after adsorption. Triplicate tests were conducted
for part of the dataset. The average value is shown with the relative percent
difference as error bars.)
177
2.2. Tests with an Aquaculture-Impacted Lake Water
Figure 2 shows the removal of NDMA and NDEA precursors in an

aquaculture-impacted lake water at acidic, neutral, and alkaline pH levels.
Figure 2. Removal of nitrosamine precursors and bulk organic matter in an

aquaculture-impacted lake water with PAC at varying pH levels (Note: initial
NDMA FP = 198 ng/L, initial NDEA FP = 106 ng/L; PAC dose = 20 mg/L,
adsorption time = 24 h, temperature = 20°C. Pulverized Calgon F400 GAC
was used in this test.)
At pH 3, PAC removed 83% of the NDMA precursors and 89% of the NDEA
precursors. At pH 7 or 8, PAC removed ~60% of the NDMA precursors and
~75% of the NDEA precursors. At pH 11, PAC removed only 34% of the NDMA
precursors and 51% of the NDEA precursors. Contrary to the trend shown in
Section 2.1, nitrosamine precursor removal in this aquaculture-impacted lake
water gradually decreased as pH increased. This indicates that the behavior of
NA precursor with PAC adsorption can be quite site-specific.
Removal of bulk organic matter from the water samples was similar to that
in Section 2.1. The DOC, UV254, and DON removal gradually decreased as pH
increased, which followed the same trend of the nitrosamine precursors in this lake
water. In almost all cases, NA precursors were removed by PAC more effectively
than the bulk organic matter. In general, the removal trend was as follows: NDEA
FP > NDMA FP > UV254 > DON ~ DOC.
The adsorption capacities of the NDMA FP in the two waters was very
different. The capacity in the blended river/wastewater at pH 8 was 13.2 ng
NDMA FP/mg PAC with Ce = 89.4 ng/L (Figure 1b) and the capacity for the
aquaculture-impacted lake water NDMA precursors was 3.6 ng NDMA FP/mg
PAC with the same Ce at the same pH (28% of the capacity for the blended
river/wastewater precursors). The difference was likely due to the different
charactersitcs of NDMA precursors in the two water samples.
178
3. pH Effect on Removal of Model NDMA Precursors with
PAC Adsorption
The experiments above indicate that the removal of nitrosamine precursors
in real waters with PAC adsorption is complex and site-specific. Therefore, two
model NDMA precursors that are PPCPs, ranitidine and chloropheniramine (3),
were used to simplify the investigation of the fate of NA precursors during the
PAC adsorption process.
Figure 3 shows the removal of the model NDMA precursors with PAC at
varying pH levels and with different PAC doses (model compound concentrations
= 0.01 mM each).
Figure 3. Removal of model NDMA precursors with WPH PAC at varying pH

levels and with different doses (model compound concentrations = 0.01 mM
each).
Ranitidine removal with PAC adsorption increased with increasing pH when

8 mg/L of PAC was used. The removal ranged from ~30 to ~70% for this dose.
Alternatively, when 20 mg/L of PAC was used, pH (for the two values tested [6.4,
11.1]) had no impact. The removal was almost 100%, so the use of a high dose of
PAC overshadowed the impact of pH. The behavior of chlorpheniramine followed
the same trend (somewhat) as that of ranitidine for a PAC dose of 8 mg/L, but it
removed 10-30% less than the ranitidine under the same conditions. In addition,
the removal at pH 3 was the same as that at pH 6.5. For a PAC dose of 20 mg/L,
there was substantially less removal at pH 6.5 than at pH 11.1.
The results with ranitidine and chlorpheniramine provide further evidence of
the pH effect on PAC adsorption of the nitrosamine precursors, where increasing
pH increased the removal of these model precursors. In a wastewater-impacted
drinking water, there can be a wide range of anthropogenic chemicals, which can
be impacted by pH to different extents.
179
4. pH Effect Study with Model NDMA Precursors via PRAM
Approach
The polarity rapid assessment method (PRAM) uses solid-phase extraction

(SPE) cartridges to isolate and fractionate organic matter with different polarities
or charges (14). Reverse-phase (C18) and strong cation exchange (SCX) SPE
cartridges (Extract Clean SPE kit catalog #210100, Alltech Associates, Deerfield,
IL, U.S.A.) have been used to elucidate NDMA precursors in wastewater (11) and
were used in this study to investigate non-polar and ionic interactions of NDMA
precursors with PAC.
Ranitidine or chlorpheniramine solutions at varying pH levels were fed into
the C18 or SCX cartridge and 40 mL of the effluent was collected. The removal
of ranitidine by C18 or SCX is shown in Figure 4, as well as the distribution
of different species (protonated [RAN2+, RAN+] or de-protonated [RAN]) of

ranitidine at varying pH levels. Those for chlorpheniramine (CA) are illustrated
in Figure 5.
Figure 4. Chemaxon modeled speciation of rantidine and its experimental

removal by C18 or SCX at varying pH levels. (Note: Duplicate tests were
conducted. The average values are shown as points and the relative percent
difference were less than 3%. Ranitidine concentration = 0.01 mM.)
It is clear that the adsorption capacity of C18 for ranitidine or chlorpheniramine

was quickly expended at pH 3, only ~20% removal or less was observed.
Chlorpheniramine was better removed (~80%) than ranitidine (~30%) at pH 7.
The capacity for each was almost 100% at pH 11: The C18 cartridges retained
nearly all of its adsorption capacity for the two chemicals, even after a 40-mL
filtration.
180
At pH 11, both model NDMA precursors quickly broke through the SCX
cartridge, with low adsorption affinity. The capacity of each was ~90-100% at
pH 3 and 7: The SCX cartridges retained almost all of the chemicals at acidic or
neutral pH levels, even after a 40-mL filtration.
In Figures 4 and 5, the removal of the two chemicals by C18 matched well
with the neutral species distribution of ranitidine or chlorpheniramine at varying
pH levels. C18 also adsorbed the monovalent cation species of chlorpheniramine
(CA+). In addition, the removal of the two model compounds by SCX coincided
with the presence of the two cationic species.
Figure 5. Chemaxon modeled speciation of chlorpheniramine and its

experimental removal by C18 or SCX at varying pH levels. (Note: Duplicate tests
were conducted. The average values are shown as points and the relative percent
difference were less than 3%. Chlorpheniramine concentrations = 0.01 mM.)
The two model NDMA precursors both have amine functional groups that are
protonated at low pH. For example, the pKa values of ranitidine are 2.7 and 8.2,
respectively. When the pH was greater than 10, the molecule was deprotonated
or neutral and did not sorb via ion exchange. When the pH was less than 6, the
molecule was protonated and thus amenable to ion exchange.
The adsorption capacities of the two chemicals at pH 7 were different.
Ranitidine was still recalcitrant to C18 adsorption at pH 7, whereas
chlorpheniramine was retained better by the C18 cartridge. The difference could
be attributed to the pKa of the two chemicals, which are 9.2 (chlorpheniramine)
and 8.2 (ranitidine).
In general, the neutral species were more easily adsorbed by the C18 cartridge,
but were recalcitrant to SCX. The positive charged species were sorbed to the SCX
media, but were excluded from C18 sorption.
181
5. Mechanism of pH Effect on NA Precursor Removal with
PAC in Natural Waters
Figure 6 illustrates a scheme around our hypothesis that pH influences
the removal of NA precursors on PAC. The NA precursors likely have at least
one dialkylamine functional group, which is a potential target for oxidant
attack and nitrosamine formation. In addition to the basic secondary amines,
such as dimethylamine and diethylamine, other NA precursors have longer
carbon chains and potentially contain functional groups such as carboxyl and
hydroxyl. Therefore, we give a model structure formula for NDMA precursors
with dimethylamine, followed by a carbon chain (Rm) containing carboxyl and
hydroxyl groups, as shown in Figure 6.
Both van der Waals forces and electrostatic interaction influence the
adsorption of organic contaminants on PAC. Van der Waals forces are favored for
nonpolar moieties. The electrostatic interactions are stronger than van der Waals
and, thus, will dominate for charged groups. The pH effect on NA precursor
removal by PAC can be explained by their net effect.
The majority of the fractions of aquatic NOM has carboxyl and hydroxyl
functional groups, bringing about the acidity of the NOM. These functional groups
are negatively charged at neutral or alkaline conditions, but become protonated at
lower pH levels (i.e., below the pKa values common to carboxylic and hydroxyl
groups). On the PAC side, high pH also makes the carboxyl and hydroxyl groups
on the carbon surface depronate or become negatively charged, which excludes the
organic matter. Other major fractions of NOM behave as net neutral molecules,
and a very small amount of the NOM behave as basic molecules. The nonpolar
adsorption by PAC will decline as pH increases, as shown in Figure 1a-1d (blended
river/wastewater) and Figure 2 (aquaculture-impacted lake water).
Figure 6. Possible mechanism of the pH effect on nitrosamine precursor removal

by PAC adsorption. (Note: Bold dashed line indicates the interaction between
PAC and a neutral moiety, light dashed line with a cross indicates the repulsion
between PAC and a charged moiety.)
182
However, NA precursors present positive charges, due to the amine functional
group, which is quite unique among the generally negative bulk organic matter.
Thus, they have different behavior during PAC adsorption at different pH levels
than with the bulk organic matter. Alkaline conditions will deprotonate the amine
group and favor nonpolar adsorption on PAC. Alternatively, acidic or neutral
conditions will increase the precursor’s positive charge and weaken the PAC
adsorption. This can occur if the amine functional group dominates, particularly
in small molecules.
The hypothesis illustrated in Figure 6 neglects the net charge of PAC. The
pulverized F400 GAC had a point of zero charge of pH 9.5, which means the net
charge of this AC was negative at neutral and acidic pH levels. These negatively
charged functional groups can sorb the positively charged NA precursors at neutral
and acidic pH levels. This partially explains the removal of NA FP by WPH PAC
in Figures 1 and 3, although the removal was quite limited.

However, the functional groups on PAC surfaces are limited. The majority
of PAC adsorption sites are nonpolar, graphite or amorphous carbon. Thus, the
charged (positive or negative) chemicals will be recalcitrant to the nonpolar
carbon, as also demostrated by the C18 media, which does not contain ionized
functional groups (Figure 4).
6. Conclusion
The following conclusions can be drawn according to the results of this
investigation.
(1) Experiments with different water samples and model NDMA precursors
indicated that removal of NDMA precursors by PAC adsorption is pH
dependent and site-specific. The removal of two model precursors
(ranitidine and chloropheniramine) and the NDMA FP in a blended
river/wastwater with PAC adsorption increased with higher pH. However,
the NDMA FP in an aquaculture-impacted lake water had declining
NDMA FP reduction with PAC adsorption with increasing pH.
(2) PRAM indicated that the pH effect on PAC adsorption of NDMA
precursors is potentially attributable to the protonization or charge of the
precursors. Alkaline conditions deprotonate amine functional groups
on model precursors and favor adsorption onto PAC, whereas acidic
or neutral conditions result in a precursor with positive charge, which
reduces the affinity for PAC.
Acknowledgments
The testing of the aquaculture-impacted lake water was supported by the
National Natural Science Foundation of China (Grant No. 21477059 and
51290284) and Tsinghua University Initiative Scientific Research Program (Grant
No. 20131089247). The testing of the blended river/wastewater was supported
183
by the Water Research Foundation, Denver, Colo., U.S.A. (project 4370, under
the management of Djanette Khiari).
We highly appreciate Rita Chang from UCLA and Yueying Ouyang from
Tsinghua University, who volunteered to be the laboratory interns and contributed
much to the laboratory work of PRAM.
References
Occurrence of disinfection byproducts in United States wastewater treatment
plant effluents. Environ. Sci. Technol. 2009, 43 (21), 8320–8325.
Formation, precursors, control, and occurrence of nitrosamines in drinking

water: A Review. Wat. Res. 2013, 47 (13), 4433.
disinfection. Wat. Res. 2011, 45 (2), 944–952.
4. Wang, C. K.; Zhang, X. J.; Wang, J.; Liu, S. M.; Chen, C.; Xie, Y. F. Effects
of organic fractions on the formation and control of N-nitrosamine precursors
during conventional drinking water treatment processes. Sci. Total Environ.
2013, 449, 295–301.
5. Hanigan, D.; Zhang, J.; Herckes, P.; Krasner, S.; Chen, C.; Westerhoff, P.
Adsorption of N-nitrosodimethylamine precursors by powdered and granular
activated carbon. Environ. Sci. Technol. 2012, 46, 12630–12639.
6. Liao, X.; Wang, C. K.; Wang, J.; Zhang, X. J.; Chen, C.; Krasner, S. W.;
Suffet, I. H. Nitrosamine precursor and DOM control in a wastewater-
impacted drinking water. J. Am. Water Works Assoc. 2014, 106, E307–E318.
7. Shah, A. D.; Krasner, S. W.; Lee, C. F. T.; von Gunten, U.; Mitch, W. A.
Tradeoffs in disinfection byproduct formation associated with precursor pre-
oxidation for control of N‐nitrosodimethylamine formation. Environ. Sci.
Technol. 2012, 46, 4809–4818.
8. Schreiber, I. M.; Mitch, W. A. Influence of the order of reagent addition on
NDMA formation during chloramination. Environ. Sci. Technol. 2005, 39,
3811–3818.
9. Krasner, S. W.; Shirkhani, R.; Westerhoff, P.; Hanigan, D.; Mitch, W.
A.; McCurry, D. L.; Chen, C.; Skadsen, J.; Von Gunten, U. Controlling
the formation of nitrosamines during water treatment; Water Research
Foundation Final Report; Water Research Foundation: Denver, CO, 2015.
10. Hanigan, D.; Zhang, J.; Herckes, P.; Zhu, E.; Krasner, S. W.; Westerhoff, P.
Contribution and removal of watershed and cationic polymer N-
nitrosodimethylamine precursors. J. Am. Water Works Assoc. 2015, 107 (3),
E152–E196.
11. Chen, C.; Leavey, S.; Krasner, S. W.; Suffet, I. H. Applying polarity rapid
assessment method and ultrafiltration to characterize NDMA precursors in
wastewater effluents. Wat. Res. 2014, 57, 115–126.
184
12. Mitch, W. A.; Gerecke, A. C.; Sedlak, D. L. A N-nitrosodimethylamine
(NDMA) precursor analysis for chlorination of water and wastewater. Wat.
Res. 2003, 37, 3733–3741.
13. USEPA. Method 521: Determination of nitrosamines in drinking water by
solid phase extraction and capillary column gas chromatography with large
volume injection and chemical ionization tandem mass spectrometry (MS/
MS); EPA/600/R-05/054; National Exposure Research Laboratory, Office of
Research and Development, USEPA: Cincinnati, Ohio, 2004.
14. Rosario-Ortiz, F. L.; Snyder, S.; Suffet, I. H. Characterization of the polarity
of natural organic matter under ambient conditions by the polarity rapid
assessment method (PRAM). Environ. Sci. Technol. 2007, 41, 4895–4900.
185
Chapter 11
Formation of DBPs: State of the Science

Susan D. Richardson*,1 and Cristina Postigo2
1Department of Chemistry and Biochemistry,
University of South Carolina, JM Palms Center,

631 Sumter Street, Columbia, South Carolina 29208
2Department of Environmental Chemistry,
Institute for Environmental Assessment and Water Research,

(IDAEA-CSIC), Carrer Jordi Girona 18-26, 08034 Barcelona, Spain
Drinking water disinfection by-products (DBPs) are primarily

formed by the reaction of disinfectants with natural organic
matter (NOM) and bromide or iodide. Precursors can also
involve pollutants, such as pesticides, pharmaceuticals,
antibacterial agents, estrogens, textile dyes, bisphenol A,
parabens, surfactants, and algal toxins. DBPs are also
formed in swimming pool water. Concerns arise due to
adverse human health impacts, including bladder cancer,
miscarriage, and birth defects. Emerging, unregulated DBPs
include halonitromethanes, iodo-trihalomethanes, iodo-acids,
haloamides, halofuranones, haloacetonitriles, nitrosamines,
and halobenzoquinones. Many of these unregulated DBPs
have been shown to be more genotoxic or cytotoxic than those
currently regulated. This chapter discusses these issues, along
with precursors and mechanisms for their formation.

Introduction
Drinking water disinfection by-products (DBPs) are an unintended
consequence of using disinfectants to kill harmful pathogens in potable water.
They are primarily formed by the reaction of disinfectants with natural organic
matter (NOM) and bromide or iodide (1). But, they can also be formed from
pollutants, such as pesticides, pharmaceuticals, antibacterial agents, estrogens,
textile dyes, bisphenol A, parabens, surfactants, and algal toxins (Table 1) (2, 3).
Popular disinfectants for drinking water include chlorine, chloramines, ozone,
chlorine dioxide, and UV.
While 11 DBPs are regulated in the United States (4), more than 600 have
been identified (1). While this seems like a great number, more than 50% of the
halogenated material formed during chlorination is still unknown, and so is the
toxicological risk that it poses to human health (3, 5).

Concerns exist due to adverse human health impacts, including bladder
cancer, miscarriage, and birth defects, which have been seen in human
epidemiologic studies (1, 6–11). It is not certain that regulations are adequately
controlling for these human health effects, as none of the regulated DBPs
cause the primary cancer observed (i.e., bladder cancer) in animals. Therefore,
there is intense research in emerging, unregulated DBPs. These include
halonitromethanes, iodo-trihalomethanes, iodo-acids, haloamides, halofuranones,
haloacetonitriles, haloacetaldehydes, nitrosamines, and halobenzoquinones.
Many of these unregulated DBPs have been shown to be more genotoxic or
cytotoxic than those currently regulated (1, 12–14). For example, iodoacetic acid
is the most genotoxic DBP identified to-date, and it is 2x more genotoxic than
bromoacetic acid (1), which is regulated, but rarely detected in drinking water.
Iodoacetic acid was also recently shown to be tumorigenic in mice (15).
Of the more than 600 DBPs that have been identified (Table 2),
nitrogen-containing DBPs (N-DBPs) are becoming a new focus because they
are generally more toxic than DBPs that do not contain nitrogen. For instance,
many nitrosamines are known to be carcinogens (1). These compounds were on
the U.S. Environmental Protection Agency’s (EPA’s) Unregulated Contaminant
Monitoring Rule (16), and are currently being considered for regulation by the
U.S. EPA.
DBPs are also formed in disinfected swimming pool water (17–21).
Swimming pool DBPs can be similar to those observed in drinking water, but
they can also be quite different, due to the additional human precursors that can
be present in pools (e.g., urine, sweat, hair, sunscreens, lotions, personal care
products etc.) (21, 22). For example, a common DBP in chlorinated swimming
pools is trichloramine, which is formed by the reaction of urea (from urine
or sweat) and chlorine (23). Trichloramine is produced in the water, but is
quickly transported to the air phase due to its high Henry’s Law constant. It is
suspected as the causal agent in the increased asthma that has been observed in
epidemiologic studies of elite swimmers (24, 25). In addition to asthma, one
study has also shown increased incidence of bladder cancer with heavy exposure
from swimming pools (26), and there are also genotoxic effects that have been
observed in swimmers (27)
190
Table 1. Examples of DBPs Generated by Water Pollutants
191
Table 1. (Continued). Examples of DBPs Generated by Water Pollutants
192
Table 2. Examples of DBP Classes Identified (with N-DBPs Highlighted)
In addition to cancer, adverse birth outcomes, and asthma, there are some
newer concerns concerning severe skin rashes, and respiratory and digestive issues
surrounding chloraminated drinking water. To-date, there has not been a controlled
scientific study to investigate these newer adverse effects. Chloramination has
become increasingly popular in the U.S., as drinking water utilities have struggled
to meet the tightened DBP regulations (4). Chloramination is also popular in other
countries, including the UK and Australia. Switching from chlorine to chloramines
can result in ~90% reduction in the levels of regulated trihalomethanes (THMs)
and haloacetic acids (HAAs), and it is also beneficial to utilities for maintaining
disinfection in the distribution system.
The nature and quantity of DBPs formed depends on the type of disinfectant,
dose, and the type of organic matter or other constituents present in the water (1,
28–31). Formation mechanisms for several DBPs and DBP classes have been
investigated, including iodo-DBPs, halonitromethanes, nitrosamines, haloamides,
halopyrroles, and halobenzoquinones (Figures 1-8). The state of the science for
their formation follows.
Iodo-DBPs
Iodo-DBPs identified to-date are shown in Figure 1. They include iodo-THMs
(dichloroiodomethane, bromochloroiodomethane, dibromoiodomethane,
chlorodiiodomethane, bromodiiodomethane, and iodoform); iodo-acids
(iodoacetic acid, bromoiodoacetic acid, chloroiodoacetic acid, diiodoacetic
acid, (Z)-3-bromo-3-iodopropenoic acid, (E)-3-bromo-3-iodopropenoic
acid, and (E)-2-iodo-3-methylbutenedioic acid) (5, 32–35); iodo-amides
(bromoiodoacetamide and chloroiodoacetamide) (1, 13, 36); and the recently
reported iodoacetaldehyde (37, 38).
193
Iodo-THMs
Iodo-THMs were the first iodo-DBPs to be discovered, with a long history

since the mid-1970s when dichloroiodomethane was measured in chlorinated tap
water and was referred to as the “5th trihalomethane” (39). Iodo-THMs have since
been measured in drinking waters treated with chlorination or chloramination (5,
31, 40–44), with the highest levels observed in chloraminated water (up to 15 µg/L
individually) (5). In fact, iodo-THMs can be formed at levels comparable to the
regulated THMs (THM4). In the U.S. Nationwide Occurrence Study, one location
showed iodo-THMs at 81% of the THM4 levels in a chloraminated drinking water
(5, 44). A recent study also revealed that chlorine dioxide can form iodoform (45).
In addition, point-of-use treatment with iodine can produce iodo-THMs (46). Of
the three iodine treatments investigated (iodine tincture, iodine tablets, and the
Lifestraw®), the highest levels were formed with iodine tincture treatments (46).
Finally, new research has revealed that hydraulic fracturing (HF) wastewater can
produce iodo-THMs when treated with chlorine or monochloramine (47). This is
likely due to the high levels of iodide present in HF wastewater (47).
Figure 1. Formation mechanisms for emerging iodo-DBPs (NOM: natural

organic matter, I-THMs: iodo-trihalomethanes, I-HAAs: iodo-acetic acids,
I-HAL: iodoacetaldehyde, I-AcAms: iodo-acetamides) (34, 46, 48–51).
194
Iodo-acids
Iodo-acids, which are the most genotoxic of the iodo-DBPs (34, 35), were
first identified as part of a U.S. Nationwide Occurrence Study (5, 35, 44).
Levels up to 1.7 µg/L were reported in a 23 city survey of chloraminated and
chlorinated drinking water from the U.S. and Canada (34). Chlorine dioxide
was also reported to form iodoacetic acid when reacted with source waters (45).
Iodo-acids were also tentatively identified in simulated drinking waters treated
with chlorine, monochloramine, and chlorine-chloramine (32). The authors
used ultra-performance liquid chromatography (UPLC)/electrospray ionization
(ESI)-tandem mass spectrometry (MS/MS) with precursor ion scanning for the
m/z 127 ion of iodine to broadly detect iodo-DBPs formed in these reactions.
This research revealed the presence of iodoacetic acid, chloroiodoacetic acid,
(E)- and (Z)-iodobutenedioic acid, 4-iodobenzoic acid, 3-iodophthalic acid,

2,4-diiodobenzoic acid, 5,6-diiodosalicylic acid, and 5,6-diiodo-3-ethylsalicylic
acid (32). In addition, iodoacetic acid and chloroacetic acid can form when
chlorinated tap water is allowed to react with iodized table salt (containing
potassium iodide) or with potassium iodide itself (33).
The rank order for genotoxicity is iodoacetic acid >> diiodoacetic
acid > bromoiodoacetic acid > (E)-2-iodo-3-methylbutenedioic acid >
(E)-3-bromo-3-iodopropenoic acid > (E)-3-bromo-2-iodopropenoic acid.
Iodoacetic acid is also teratogenic, producing developmental effects (neural tube
closures) in mouse embryos, at levels (nM) similar to levels that induce DNA
damage in mammalian cells (52, 53). As mentioned earlier, iodoacetic acid was
also recently shown to be tumorigenic in mice (15).
Iodo-amides
Haloamides are formed primarily by chlorine or chloramine, and they

were quantified for the first time in the U.S. Nationwide Occurrence Study (5,
44). Iodoacetamides—bromoiodoacetamide and chloroiodoacetamide—were
subsequently identified in drinking water treated with chloramines or chlorine.
Bromoiodoacetamide was initially found in chloraminated drinking water
from several cities in the U.S. (13), and later, both bromoiodoacetamide and
chloroiodoacetamide were found in chloraminated and chlorinated drinking water
from three provinces in China (36). Haloacetamides can form by the hydrolysis
of the corresponding haloacetonitriles (54, 55), and new research also shows that
they can be formed by an independent pathway and that they are more favored
with chloramination vs. chlorination (56).
Both of these iodoacetamides are highly cytotoxic and genotoxic in
mammalian cells (13). As a class, haloamides are the most cytotoxic of all DBP
classes measured to-date, and they are the second-most genotoxic DBP class,
very close behind the halonitriles.
195
Iodo-DBP Formation Mechanisms
In all cases studied to-date, chloramination increases the formation of all of

these iodo-DBPs. In practice, drinking water treatment plants can add pre-formed
monochloramine (reaction of chlorine and ammonia), but they will generally add
chlorine first and wait for a certain amount of time (free chlorine contact time)
before the ammonia is added, so that a higher level of microbial inactivation is
achieved. Most data shows increased formation of iodo-DBPs with lower free
chlorine contact time (higher monochloramine contact time) (5, 29, 32, 34, 57,
58). This is consistent with a mechanism that Bichsel and von Gunten proposed
for the formation of iodo-DBPs, based on controlled laboratory experiments of
iodo-THMs, which involves competing mechanisms to form iodate and organic
iodo-DBPs (59, 60). Reaction of aqueous chlorine with iodide initially forms
hypoiodous acid (HOI), which can then react further with chlorine to form
iodite and iodate (Figure 2). These reactions are much faster than the competing
reactions to form organic iodo-DBPs (e.g., iodo-THMs and iodo-acids), such that
chlorination favors the formation of iodate over organic iodo-DBPs. However, the
reactions of monochloramine with HOI to form iodite and iodate are much slower
than the corresponding reactions of chlorine with HOI, such that monochloramine
favors the formation of organic iodo-DBPs over iodate.
Figure 2. Mechanism of iodo-DBP formation with chlorine and chloramines

(59, 60).
196
New research shows that ozone pretreatment at lower pH can be used to
minimize iodo-DBP (and bromate) formation by selectively oxidizing iodide to
iodate (61). Ozone could also be used to oxidize iodo-THMs that might already
be present in the water.
Iodide salt is believed to be the major source of iodine in the formation of
iodo-DBPs. But, new research has revealed that compounds used for medical
imaging, i.e., iodinated X-ray contrast media (ICM) can also be a source of iodine
(48, 49, 62). ICM are excreted within ~24 h after medical imaging, and they are
stable during wastewater treatment, which has resulted in levels up to 100 µg/L
in rivers and creeks (63) and up to 2.7 µg/L in drinking water reservoirs (49).
These ICM compounds have a triiodobenzene core structure with 3 amide side
chains, and research shows that chlorine or chloramine can react with ICM to form
iodo-THMs and to a lesser extent, iodo-acids. NOM and pH substantially affect
the formation, and OCl- is believed to be the reacting species. In addition, new
controlled laboratory studies indicate that iodo-THMs are favored at low chlorine
doses, but are suppressed at higher doses (62).
To-date, iopamidol appears to be the most reactive, with much less formation
of iodo-DBPs from other ICM investigated (e.g., iopromide, iohexol, iomeprol,
diatrizoate, hiztodenz, and iodixanol) (48, 49, 62). New research using liquid
chromatography (LC)-high resolution MS/MS and nuclear magnetic resonance
(NMR) spectroscopy is revealing the initial points of reaction on the iopamidol
structure, along with the initial high molecular weight DBPs formed (Figure
3) (48). The proposed reactions involve cleavage of one of the side chains,
substitution of chlorine for iodine on the benzene ring, amide hydrolysis, cleavage
of the other side chains, and oxidation of NH2 to NO2 (Figure 3). Structures for
19 high molecular weight DBPs were deduced in the reaction pathway.
Figure 3. Proposed reaction of iopamidol with chlorine and monochloramine

(48).
197
Nitrosamines
Nitrosamines were discovered to be DBPs in 2002 (64, 65) and have been of
significant interest ever since because several, including N-nitrosodimethylamine
(NDMA), are known carcinogens (1). NDMA was initially discovered in
chlorinated drinking waters from Ontario, Canada (66), and was subsequently
found in other locations (64, 65, 67). The detection of NDMA in drinking
water is largely due to improved analytical techniques that allow its
determination at low ng/L concentrations. Following the discovery of NDMA,
other nitrosamines were identified as DBPs, including N-nitrosopiperidine,
N-nitrosodiphenylamine, N-nitrosopyrrolidine, and N-nitrosomorpholine (68,
69). N-Nitrosodiphenylamine is thermally unstable and requires LC-MS/MS
for its analysis (68). A recently developed total nitrosamine (TONO) assay
indicates that the nitrosamines identified to-date only represent 5-10% of the
total nitrosamines formed in drinking water and recreational waters (70, 71).
The identity of these other nitrosamines is currently unknown. Interestingly,
algal-derived organic matter was an insignificant precursor for EPA Method
521 nitrosamines during chloramination, but a potent precursor for other,
uncharacterized N-nitrosamines, as measured using the TONO assay. In 2014,
tobacco-specific nitrosamines —4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone
and 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol— were also discovered to be
chloramination DBPs (72).
NDMA is regulated in California at 10 ng/L (73) and Ontario, Canada at 9
ng/L (74). A Canadian national drinking water guideline has also recently been
established, which limits NDMA to 40 ng/L in drinking water (75), and the U.S.
EPA is considering its regulation in the United States. NDMA was included in
the U.S. EPA’s second Unregulated Contaminants Monitoring Rule (UCMR-2),
along with 5 other nitrosamines (N-nitrosodiethylamine, N-nitrosodibutylamine,
N-nitrosodipropylamine, N-nitrosomethylethylamine, and N-nitrosopyrrolidine).
National occurrence data are currently available (16). These new data revealed a
maximum level of 530 ng/L for NDMA in chloraminated drinking water, which
surpasses the previous maximum (180 ng/L) observed in Canadian chloraminated
drinking water (68). In addition, NDMA and 4 other nitrosamines are also on the
U.S. EPA’s Contaminant Candidate List (CCL-3), a priority list of drinking water
contaminants (76).
Nitrosamine Formation Mechanisms
NDMA is generally found at highest levels in chloraminated drinking

water. Early research indicated that the nitrogen in monochloramine (NH2Cl)
was incorporated into the structure of the NDMA by-product (64); subsequent
research revealed that dichloramine was actually the primary reactant (77).
Dichloramine always coexists (at lower levels) with monochloramine under
typical chloramination conditions. Nitrite can also serve as a nitrosamine
precursor in chlorination reactions (78, 79). In these reactions, a dinitrogen
tetraoxide (N2O4) intermediate is believed to form, which can nitrosate or nitrate
amines.
198
Chlorination can also form NDMA when nitrogen precursors are present
(e.g., natural ammonia in the source water or nitrogen-containing coagulants
or ion-exchange resins used in the water treatment process) (80–83) (Figure
4). Consumer products, including shampoos, laundry detergents, dish washing
liquids, and fabric softeners, can also be precursors in the formation of
nitrosamines (84). Surprisingly, quaternary amine polymers appear to be more
reactive than the monomers (84); this phenomenon was also observed in previous
studies of diallyldimethylammonium chloride (DADMAC) polymers used as
coagulants in drinking water treatment (83, 85, 86). These findings are important
because nitrosamine formation is often attributed to lower order amine impurities,
but these results clearly show that quaternary amine polymers can also form
NDMA.
Amino acids and hydrophilic/low molecular weight dissolved organic
nitrogen can also serve as nitrosamine precursors (87), as can amine-based

pharmaceuticals (88). Diphenylamine was also shown to be a key precursor
in the formation of N-nitrosodiphenylamine from chloramines (89, 90). The
Lifestraw®, a point-of-use device that uses I3-complexed resins with an activated
carbon filter, can also produce NDMA, but levels rapidly decline to low levels (4
ng/L) after the first few flushes of water (46). Krasner et al. recently published
an excellent review on the formation, precursors, control, and occurrence of
nitrosamines in drinking water (91).
Figure 4. Formation mechanisms for NDMA and other nitrosamines (NOM:

natural organic matter, DADMAC: diallyldimethylammonium chloride) (30,
82–84, 88, 92–95).
199
NDMA (and other nitrosamines) can dramatically increase in concentration
in distribution systems (relative to finished water at the drinking water treatment
plant). For example, an initial level of 67 ng/L in drinking water treatment
plant effluent was shown to increase to 180 ng/L in the distribution system (68).
As a result, measurements taken at water treatment plants may substantially
underestimate the public’s exposure to this carcinogen.
While generally attributed to the use of chloramines or chlorine, NDMA was
recently identified in ozonated drinking water from Germany (92). A fungicide
which contains a dimethylamine group —tolylfluanide— was discovered to be
the precursor in its formation. Subsequent research also revealed that trace levels
of bromide can catalyze its formation (96). Dithiocarbamate pesticides can also
react with NH2Cl, O3, Cl2, or ClO2 to form NDMA (94). In these reactions,
the dithiocarbamate forms an amine by hydrolysis/oxidation, which then forms
NDMA (94).
Finally for the tobacco-specific nitrosamines just identified in 2014, tobacco
alkaloids, including nicotine, nornicotine, and anabasine were determined to be
precursors to their formation in chloraminated drinking water (72).
Haloamides
Haloamides are formed by both chlorine and chloramine, (5, 13, 44,
97, 98), but preferentially by monochloramine (56). The mechanism can
involve the hydrolysis of the corresponding haloacetonitriles (54, 55), or
reaction of monochloramine with organic nitrogen precursors (56) (Figure
5). Experiments involving 15N-labeled monochloramine indicated initial
rapid formation of both dichloroacetamide and dichloroacetonitrile, where
the nitrogen originated from organic nitrogen precursors. However, slower
formation occurs by pathways involving chloramine incorporation into organic
precursors. In addition, experiments with asparagine as a model precursor also
suggested that dichloroacetamide can be formed without a dichloroacetonitrile
intermediate, and humic materials were found to be more potent precursors for
dichloroacetamide formation, while wastewater effluents and algal substances
were more potent precursors for dichloroacetonitrile formation (56). Therefore,
there are independent mechanisms involved in the formation of haloacetamides,
beyond the hydrolysis of the haloacetonitriles.
Several amino acids were also recently found to be precursors of haloamides,
with aspartic acid, histidine, tyrosine, tryptophan, glutamine, asparagine, and
phenylalanine reacting with chlorine to form dichloroacetamide (99).
Halonitromethanes
Nine chloro/bromo halonitromethanes have been identified to date (Figure
6). Chloropicrin (trichloronitromethane) is the most commonly measured
example in this class, but it has not been a concern for toxicity in drinking
water. Brominated nitromethanes, however, have shown significant toxicity
(100) and have been found in drinking water up to 3 μg/L individually
200
(5, 44, 100–102). Bromonitromethanes are more cytotoxic and genotoxic
than most DBPs currently regulated (100). Dibromonitromethane is more
than an order of magnitude more genotoxic to mammalian cells than MX
(3-chloro-4-(dichloromethyl)-5-hydroxy-2(5H)-furanone, a carcinogenic DBP),
and is more genotoxic than all of the regulated DBPs, except for monobromoacetic
acid. Other brominated forms are also potent in this assay. Halonitromethanes
are also mutagenic in the Salmonella bacterial cell assay (103), with mutagenic
potencies greater than that of the regulated THMs (104). The halonitromethanes
are also at least 10x more cytotoxic than the THMs, and the greater cytotoxic and
mutagenic activities of the halonitromethanes was indicated to be likely due to
the greater intrinsic reactivity conferred by the nitro group (104).
Figure 5. Formation mechanisms for haloamides (30).
Bromonitromethanes are substantially increased in formation with the use

of pre-ozonation before post-chlorination or chloramination (5, 44). Laboratory-
scale formation studies indicate that nitrite may also play a role in the formation
of the nitro group in these DBPs (105). Research indicates that hydrophilic NOM
is a more important precursor than hydrophobic or transphilic NOM (106).
Tribromonitromethane (bromopicrin) and other trihalonitromethanes (which
include bromodichloro- and chlorodibromonitromethane) require particular
analytical conditions for their analysis. These compounds are thermally unstable
and decompose under commonly used injection port temperatures during gas
chromatography (GC)-electron capture detection (ECD) or GC/mass spectrometry
(MS) analysis (97).
201
Figure 6. Formation mechanisms for halonitromethanes (30, 105).
Halopyrroles
A halogenated pyrrole —2,3,5-tribromopyrrole (Figure 7)— was first
reported as a DBP in 2003 (107). It was found in finished drinking water
from Israel that was treated with pre-chlorination followed by treatment with
combined chlorine dioxide-chlorine or chlorine dioxide-chloramine. The source
water contained exceptionally high bromide levels (approximately 2 ppm). This
identification resulted from the first study of chlorine dioxide DBPs formed
under high bromide/iodide conditions. Bromide levels in U.S. source waters
generally range up to a maximum of approximately 0.5 ppm, and to-date, this
tribromopyrrole has not been identified in drinking waters from the United States.
Tribromopyrrole is 8x more cytotoxic than dibromoacetic acid (a regulated
DBP) and has about the same genotoxic potency as MX (107), which is also
an animal carcinogen (109). Formation studies of different NOM precursors
revealed that tribromopyrrole forms primarily from humic acid (vs. fulvic acid),
and this isolated humic acid contained a greater contribution of nitrogen in its
chemical structure than the isolated fulvic acid. It is interesting to note that a
soil humic model proposed by Schulten and Schnitzer (110) includes a pyrrole
group in its structure (110), but tribromopyrrole represented the first time a
halopyrrole was reported as a DBP. In none of the samplings from this research
202
was tribromopyrrole found in pre-chlorinated waters (with chlorine treatment
only). Thus, the combination of chlorine dioxide and chlorine (or chloramines)
may be necessary for its formation. It is also possible that chloramination alone
may also be important for its formation.
Figure 7. Formation mechanisms for halopyrroles (107, 108).
New research has revealed that halopyrroles can form as DBPs in chlorinated
saline wastewater effluents (108). Tri- and tetra-halopyrroles were identified,
including brominated, chlorinated, and iodinated analogues (Figure 7). In
mechanistic experiments, chlorophyllin (which contains pyrrole units in its
structure) was found to be an important precursor in their formation. The
chlorinated saline wastewaters were also found to be developmentally toxic to
marine polychaetes (108).
Halobenzoquinones
Halobenzoquinones (HBQs) are a new class of DBP recently identified
(Figure 8). Four HBQs —2,6-dichlorobenzoquinone, 2,6-dibromobenzoquinone,
2,6-dichloro-3-methylbenzoquinone, and 2,3,6-trichlorobenzoquinone— were
initially identified in drinking waters treated with chlorine, chloramines, chlorine-
chloramines, ozone-chloramines, and chloramines-UV (111, 112). Levels ranged
up to 275 ng/L. 2,6-Dichlorobenzoquinone, 2,3,6-trichloro-1,4-benzoquinone,
203
2,3-dibromo-5,6-dimethyl-1,4-benzoquinone, and 2,6-dibromo-1,4-benzoquinone
were also formed in chlorinated swimming pools (113). And, subsequent work
has also revealed the formation of hydroxylated HBQs with UV treatment (114).
Quantitative structure-toxicity relationship (QSTR) analysis had predicted that
haloquinones are highly toxic and may be formed during drinking water treatment.
The chronic lowest observed adverse effect levels (LOAELs) of haloquinones are
predicted to be in the low µg/kg body weight per day range, which is 1000x lower
than most regulated DBPs, except bromate.
Separate controlled laboratory studies using phenol as a precursor
demonstrated that chlorination produced the highest levels of
2,6-dichlorobenzoquinone, while preozonation increased the formation of
2,6-dibromobenzoquinone in the presence of bromide. UV filters and other
aromatic compounds found in lotions and sunscreens were also determined to be
precursors of the HBQs found in swimming pools (113).
Figure 8. Formation mechanisms for halobenzoquinones (112, 113, 115).
Conclusions
In conclusion, the nature and quantity of DBPs formed depends on the type
of disinfectant, dose, and the type of organic matter or other constituents present
in the water. Due to the significant occurrence and toxicity of many emerging
204
DBPs, it is important to understand their formation so that treatment methods
can be developed to minimize them in drinking water. For example, formation
of several of these DBPs is enhanced by chloramination, which is a popular
disinfection alternative that many treatment plants have switched to in order to
minimize regulated THMs and HAAs. However, due to the significant toxicity
associated with several chloraminated DBPs, such as NDMA, iodoacetic acid,
and haloamides, it may be wise to investigate other treatment technologies,
such as granular activated carbon (GAC) or membranes, which could be used
in combination with chlorine or chlorine dioxide to potentially minimize both
emerging chloramination DBPs and regulated DBPs. However, research is
needed to ensure that these treatment changes do not cause other unintended
consequences, such as increased brominated DBP formation. In this regard, when
exploring new treatment methods, it is wise to combine chemistry and toxicology.
While the U.S. EPA is considering regulation of NDMA and other

nitrosamines, it is not known whether other emerging DBPs will be seriously
considered for regulation. A major stumbling block for new regulations is the
lack of in vivo toxicology data and national occurrence data. These are two
key criteria that the U.S. EPA uses to determine whether or not to regulate new
contaminants in drinking water. Because nitrosamines were contaminants of
concern for many decades prior (before they were known to be DBPs), a wealth
of in vivo toxicity data already existed for them, which allowed regulators to
have the data necessary to place them on priority lists, such as the CCL-3 and the
UCMR-2. Once contaminants are on these lists, EPA Methods can be created for
them and national data can be collected (through the UCMR). However, in recent
years, there is reduced emphasis on collecting in vivo data, due to the expense
and effort involved, and it is unlikely that the U.S. EPA would regulate new
DBPs without in vivo data. Thus, there is a tremendous need for in vivo data to
determine whether compounds like iodoacetic acid and dibromonitromethane are
carcinogens or teratogens.
Another limitation lies in the fact that we are still currently missing more than
50% of the DBPs with our current analytical techniques. LC-MS/MS is helping
to reveal a few more DBP classes, such as halobenzoquinones and new bromo-
and iodo-DBPs, and it can measure a broader spectrum of polar chemicals, where
GC/MS is more limited. In addition, the use of halogen-specific total organic
halogen (TOX), which allows chlorinated, brominated, and iodinated DBP species
to be measured as total organic chlorine (TOCl), total organic bromine (TOBr),
and total organic iodine (TOI), is helping to provide new information. As research
continues, it will be important to determine whether TOBr and TOI correlate
with adverse health impacts, and if they do, then possibly new cost-effective
technologies can be developed to remove Br- and I- from source waters.
While many challenges still exist, there has been a renaissance of research in
DBPs, with increased efforts in several countries throughout the world. Drinking
water safety remains an important goal, and it is hopeful that the remaining health
issues surrounding drinking water will be solved.
205
Acknowledgments
C.P. acknowledges support from the European Union Seventh Framework
Programme (FP7/2007-2013) under grant agreement no. 274379 (Marie Curie
IOF). This work has been financially supported by the Generalitat de Catalunya
(Consolidated Research Groups “2014 SGR 418 - Water and Soil Quality Unit”
and 2014 SGR 291 - ICRA). This work reflects only the author’s views. The EU
is not liable for any use that may be made of the information contained therein.
References


2. Postigo, C.; Richardson, S. D. Transformation of pharmaceuticals during
oxidation/disinfection processes in drinking water treatment. J. Hazard.
Mater. 2014, 279, 461–475.
3. Richardson, S. D.; Postigo, C. Drinking water disinfection by-products.
In Emerging Organic Contaminants and Human Health - The Handbook
of Environmental Chemistry; Barcelo, D., Ed.; Springer-Verlag Berlin
Heidelberg: Germany, 2012; Vol. 20, pp 93−138.
4. U.S. Environmental Protection Agency (U.S. EPA). Drinking water
contaminants; http://goo.gl/90OBMZ (accessed November, 2014).
7175–7185.
pregnancy outcomes: A review. Environ. Health Perspect. 2002, 110,
61–74.
Garcia-Closas, R.; Serra, C.; Carrato, A.; Castaño-Vinyals, G.; Samanic, C.;
Rothman, N.; Kogevinas, M. Polymorphisms in GSTT1, GSTZ1, AND
CYP2E1, disinfection by-products, and risk of bladder cancer in Spain.
Environ. Health Perspect. 2010, 118, 1545–1550.
8. Nieuwenhuijsen, M. J.; Toledano, M. B.; Eaton, N. E.; Fawell, J.; Elliott, P.
Chlorination disinfection byproducts in water and their association with
adverse reproductive outcomes: A review. Occup. Environ. Med. 2000, 57,
73–85.
9. Savitz, D. A.; Singer, P. C.; Hartmann, K. E.; Herring, A. H.; Weinberg,
H. S.; Makarushka, C.; Hoffman, C.; Chan, R.; Maclehose, R. Drinking
Water Disinfection By-Products and Pregnancy Outcome; AWWA Research
Foundation: Denver, CO, 2005; p 212.
206
10. Villanueva, C. M.; Cantor, K. P.; Cordier, S.; Jaakkola, J. J. K.; King, W. D.;
Lynch, C. F.; Porru, S.; Kogevinas, M. Disinfection byproducts and bladder
cancer: A pooled analysis. Epidemiology 2004, 15, 357–367.
11. Waller, K.; Swan, S. H.; DeLorenze, G.; Hopkins, B. Trihalomethanes in
drinking water and spontaneous abortion. Epidemiology 1998, 9, 134–140.
of Disinfection By-Products; Water Research Foundation: Denver, CO, 2009;
pp 134.
13. Plewa, M. J.; Muellner, M. G.; Richardson, S. D.; Fasano, F.; Buettner, K.
M.; Woo, Y. T.; McKague, A. B.; Wagner, E. D. Occurrence, synthesis,
and mammalian cell cytotoxicity and genotoxicity of haloacetamides: An
emerging class of nitrogenous drinking water disinfection byproducts.

871–878.
15. Wei, X.; Wang, S.; Zheng, W.; Wang, X.; Liu, X.; Jiang, S.; Pi, J.; Zheng, Y.;
He, G.; Qu, W. Drinking water disinfection byproduct iodoacetic acid induces
tumorigenic transformation of NIH3T3 cells. Environ. Sci. Technol. 2013,
47, 5913–5920.
16. U.S. Environmental Protection Agency (U.S. EPA). Unregulated
Contaminant Monitoring Rule 2 - UCMR 2; http://goo.gl/EQD8bI (accessed
November, 2014).
17. Li, J.; Blatchley, E. R., III. Volatile disinfection byproduct formation
resulting from chlorination of organic - Nitrogen precursors in swimming
pools. Environ. Sci. Technol. 2007, 41, 6732–6739.
18. Zwiener, C.; Richardson, S. D.; DeMarini, D. M.; Grummt, T.; Glauner, T.;
Frimmel, F. H. Drowning in disinfection byproducts? Assessing swimming
pool water. Environ. Sci. Technol. 2007, 41, 363–372.
19. Richardson, S. D.; DeMarini, D. M.; Kogevinas, M.; Fernandez, P.;
Marco, E.; Lourencetti, C.; Ballesté, C.; Heederik, D.; Meliefste, K.;
McKague, A. B.; Marcos, R.; Font-Ribera, L.; Grimalt, J. O.; Villanueva, C.
M. What’s in the pool? a comprehensive identification of disinfection
by-products and assessment of mutagenicity of chlorinated and brominated
swimming pool water. Environ. Health Perspect. 2010, 118, 1523–1530.
20. Weaver, W. A.; Li, J.; Wen, Y.; Johnston, J.; Blatchley, M. R.; Blatchley,
III, E. R. Volatile disinfection by-product analysis from chlorinated indoor
swimming pools. Water Res. 2009, 43, 3308–3318.
21. Chowdhury, S.; Al-hooshani, K.; Karanfil, T. Disinfection byproducts in
swimming pool: Occurrences, implications and future needs. Water Res.
2014, 53, 68–109.
22. Kim, H.; Shim, J.; Lee, S. Formation of disinfection by-products in
chlorinated swimming pool water. Chemosphere 2002, 46, 123–130.
23. Schmalz, C.; Frimmel, F. H.; Zwiener, C. Trichloramine in swimming pools
- Formation and mass transfer. Water Res. 2011, 45, 2681–2690.
207
24. Bougault, V.; Boulet, L. P. Airway dysfunction in swimmers. Brit. J. Sport.
Med. 2012, 46, 402–406.
25. Florentin, A.; Hautemaniere, A.; Hartemann, P. Health effects of disinfection
by-products in chlorinated swimming pools. Int. J. Hyg. Environ. Health.
2011, 214, 461–469.
Tardon, A.; Garcia-Closas, R.; Serra, C.; Carrato, A.; Castaño-Vinyals, G.;
bathing, showering and swimming in pools. Am. J. Epidemiol. 2007, 165,
148–156.

Genotoxic effects in swimmers exposed to disinfection by-products in
indoor swimming pools. Environ. Health Perspect. 2010, 118, 1531–1537.
28. Liang, L.; Singer, P. C. Factors influencing the formation and relative
distribution of haloacetic acids and trihalomethanes in drinking water.
29. Hua, G.; Reckhow, D. A. Comparison of disinfection byproduct formation
31. Zhang, Z.; Echigo, S.; Minear, R. A.; Plewa, M. J. Characterization and
Eds.; American Chemical Society: Washington, DC, 2000; pp 299−314.
32. Ding, G.; Zhang, X. A picture of polar iodinated disinfection byproducts
in drinking water by (UPLC/)ESI-tqMS. Environ. Sci. Technol. 2009, 43,
9287–9293.
33. Becalski, A.; Lau, B. P. Y.; Schrader, T. J.; Seaman, S. W.; Sun, W. F.
Formation of iodoacetic acids during cooking: Interaction of iodized table
salt with chlorinated drinking water. Food Addit. Contam. 2006, 23,
957–962.
McKague, A. B.; Miltner, R. J.; Wagner, E. D.; Plewa, M. J. Occurrence
and mammalian cell toxicity of iodinated disinfection byproducts in drinking
35. Plewa, M. J.; Wagner, E. D.; Richardson, S. D.; Thruston, Jr., A. D.; Woo, Y.
T.; McKague, A. B. Chemical and biological characterization of newly
discovered iodoacid drinking water disinfection byproducts. Environ. Sci.
Technol. 2004, 38, 4713–4722.
36. Chu, W.; Gao, N.; Yin, D.; Krasner, S. W.; Templeton, M. R. Trace
determination of 13 haloacetamides in drinking water using liquid
208
chromatography triple quadrupole mass spectrometry with atmospheric
pressure chemical ionization. J. Chromatogr. A 2012, 1235, 178–181.
37. Jeong, C. H.; Postigo, C.; Richardson, S. D.; Simmons, J. E.; Kimura, S. Y.;
Marinas, B. J.; Barcelo, D.; Wagner, E. D.; Plewa, M. The occurrence and
comparative toxicity of haloacetaldehyde disinfection byproducts in drinking
water. Environ. Sci. Technol. 2014 Submitted for publication.
38. Postigo, C.; Jeong, C. H.; Richardson, S. D.; Wagner, E. D.; Plewa,
M.; Simmons, J. E.; Barcelo, D., Analysis, occurrence and toxicity of
haloacetaldehydes in drinking waters: iodoacetaldehyde as an emerging
disinfection byproduct. In Occurrence, Formation, Health Effects, and
Control of Disinfection By-Products; Karanfil, T., Mitch, W., Westerhoff,
P., Yuefeng, X., Eds.; ACS Symposium Series 1190; American Chemical
Society: Washington, DC, 2014..
39. Brass, H. J.; Feige, M. A.; Halloran, T.; Mello, J. W.; Munch, D.; Thomas,
R. F. The national organic monitoring survey: samplings and analysis for
purgeable organic compounds. In Drinking Water Quality Enhancement
Through Source Protection; Pojasek, R. B., Ed.; Ann Arbor Science: Ann
Arbor, MI, 1977; pp 393−416.
40. Cancho, B.; Ventura, F.; Galceran, M.; Diaz, A.; Ricart, S. Determination,
synthesis and survey of iodinated trihalomethanes in water treatment
processes. Water Res. 2000, 34, 3380–3390.
41. Wei, X.; Chen, X.; Wang, X.; Zheng, W.; Zhang, D.; Tian, D.; Jiang, S.;
Ong, C. N.; He, G.; Qu, W. Occurrence of regulated and emerging iodinated
DBPs in the Shanghai drinking water. PLoS One 2013, 8, 1–7.
42. Criquet, J.; Allard, S.; Salhi, E.; Joll, C. A.; Heitz, A.; von Gunten, U. Iodate
and iodo-trihalomethane formation during chlorination of iodide-containing
waters: Role of bromide. Environ. Sci. Technol. 2012, 46, 7350–7357.
43. Krasner, S. W.; McGuire, M. J.; Jacangelo, J. G.; Patania, N. L.; Reagan, K.
M.; Aieta, E. M. The occurrence of disinfection byproducts in United States
drinking water. J. Am. Water Works Assoc. 1989, 81, 41–53.
44. Weinberg, H. S.; Krasner, S. W.; Richardson, S. D.; Thruston, A. D., Jr.
The Occurrence of Disinfection By-Products (DBPs) of Health Concern in
Drinking Water: Results of a Nationwide DBP Occurrence Study; National
Exposure Research Laboratory, Office of Research and Development, U.S.
Environmental Protection Agency: Athens, GA, 2002.
45. Ye, T.; Xu, B.; Lin, Y. L.; Hu, C. Y.; Lin, L.; Zhang, T. Y.; Gao, N. Y.
Formation of iodinated disinfection by-products during oxidation of iodide-
containing waters with chlorine dioxide. Water Res. 2013, 47, 3006–3014.
46. Smith, E. M.; Plewa, M. J.; Lindell, C. L.; Richardson, S. D.; Mitch, W.
A. Comparison of byproduct formation in waters treated with chlorine and
iodine: Relevance to point-of-use treatment. Environ. Sci. Technol. 2010,
44, 8446–8452.
47. Parker, K. M.; Zeng, T.; Harkness, J.; Vengosh, A.; Mitch, W. A. Enhanced
formation of disinfection byproducts in shale gas wastewater-impacted
drinking water supplies. Environ. Sci. Technol. 2014, 48, 11161–11169.
209
48. Wendel, F. M.; Lütke Eversloh, C.; Machek, E. J.; Duirk, S. E.; Plewa, M.
J.; Richardson, S. D.; Ternes, T. A. Transformation of iopamidol during
chlorination. Environ. Sci. Technol. 2014, 48, 12689–12697.
49. Duirk, S. E.; Lindell, C.; Cornelison, C. C.; Kormos, J.; Ternes, T. A.;
Attene-Ramos, M.; Osiol, J.; Wagner, E. D.; Plewa, M. J.; Richardson, S.
D. Formation of toxic iodinated disinfection by-products from compounds
used in medical imaging. Environ. Sci. Technol. 2011, 45, 6845–6854.
50. Gong, T.; Zhang, X. Detection, identification and formation of new iodinated
disinfection byproducts in chlorinated saline wastewater effluents. Water
Res. 2015, 68, 77–86.
51. Zhang, T. Y.; Xu, B.; Hu, C. Y.; Lin, Y. L.; Lin, L.; Ye, T.; Tian, F. X. A
comparison of iodinated trihalomethane formation from chlorine, chlorine
dioxide and potassium permanganate oxidation processes. Water Res. 2015,
68, 394–403.
52. Hunter, I. E. S.; Rogers, E. H.; Schmid, J. E.; Richard, A. Comparative effects
of haloacetic acids in whole embryo culture. Teratology 1996, 54, 57–64.
53. Hunter, E. S. I.; Tugman, J. A. Inhibitors of glycolytic metabolism affect
neurulation-stage mouse conceptuses in vitro. Teratology 1995, 52, 317–323.
54. Glezer, V.; Harris, B.; Tal, N.; Iosefzon, B.; Lev, O. Hydrolysis of
haloacetonitriles: Linear free energy relationship. Kinetics and products.
Water Res. 1999, 33, 1938–1948.
55. Reckhow, D. A.; MacNeill, A. L.; Platt, T. L.; MacNeill, A. L.; McClellan, J.
N. Formation and degradation of dichloroacetonitrile in drinking waters. J.
Water Supply: Res. Technol. −AQUA 2001, 50, 1–13.
56. Huang, H.; Wu, Q. Y.; Hu, H. Y.; Mitch, W. A. Dichloroacetonitrile
and dichloroacetamide can form independently during chlorination and
chloramination of drinking waters, model organic matters, and wastewater
effluents. Environ. Sci. Technol. 2012, 46, 10624–10631.
57. Jones, D. B.; Saglam, A.; Triger, A.; Song, H.; Karanfil, T. I-THM formation
and speciation: Preformed monochloramine versus prechlorination followed
by ammonia addition. Environ. Sci. Technol. 2011, 45, 10429–10437.
58. Jones, D. B.; Song, H.; Karanfil, T. The effects of selected preoxidation
strategies on I-THM formation and speciation. Water Res. 2012, 46,
5491–5498.
59. Bichsel, Y.; von Gunten, U. Oxidation of iodide and hypoiodous acid in the
disinfection of natural waters. Environ. Sci. Technol. 1999, 33, 4040–4045.
60. Bichsel, Y.; von Gunten, U. Formation of iodo-trihalomethanes during
disinfection and oxidation of iodide-containing waters. Environ. Sci.
Technol. 2000, 34, 2784–2791.
61. Allard, S.; Nottle, C. E.; Chan, A.; Joll, C.; von Gunten, U. Ozonation
of iodide-containing waters: Selective oxidation of iodide to iodate with
simultaneous minimization of bromate and I-THMs. Water Res. 2013, 47,
1953–1960.
62. Ye, T.; Xu, B.; Wang, Z.; Zhang, T. Y.; Hu, C. Y.; Lin, L.; Xia, S. J.; Gao, N.
Y. Comparison of iodinated trihalomethanes formation during aqueous
chlor(am)ination of different iodinated X-ray contrast media compounds in
the presence of natural organic matter. Water Res. 2014, 66, 390–398.
210
63. Ternes, T. A.; Hirsch, R. Occurrence and behavior of X-ray contrast media in
sewage facilities and the aquatic environment. Environ. Sci. Technol. 2000,
34, 2741–2748.
from reaction of monochloramine: A new disinfection by-product. Water
Res. 2002, 36, 817–824.
65. Mitch, W. A.; Sedlak, D. L. Formation of N-nitrosodimethylamine (NDMA)
from dimethylamine during chlorination. Environ. Sci. Technol. 2002, 36,
588–595.
66. Jobb, D. B.; Hunsinger, R.; Meresz, O.; Taguchi, V. Y. A study of the
occurrence and inhibition of formation of N-nitrosodimethylamne (NDMA)
in the Ohsweken water supply. In Proceedings of the Fifth National
Conference on Drinking Water; September 13–15, 1992, American Water
Works Association (AWWA), Winnipeg, Canada, 1993; pp 241–252.

67. Boyd, J. M.; Hrudey, S. E.; Li, X. F.; Richardson, S. D. Solid-phase extraction
and high-performance liquid chromatography mass spectrometry analysis of
nitrosamines in treated drinking water and wastewater. TrAC, Trends Anal.
Chem. 2011, 30, 1410–1421.
68. Charrois, J. W. A.; Arend, M. W.; Froese, K. L.; Hrudey, S. E. Detecting
N-nitrosamines in drinking water at nanogram per liter levels using ammonia
positive chemical ionization. Environ. Sci. Technol. 2004, 38, 4835–4841.
69. Zhao, Y. Y.; Boyd, J.; Hrudey, S. E.; Li, X. F. Characterization of new
nitrosamines in drinking water using liquid chromatography tandem mass
spectrometry. Environ. Sci. Technol. 2006, 40, 7636–7641.
70. Kulshrestha, P.; McKinstry, K. C.; Fernandez, B. O.; Feelisch, M.; Mitch, W.
A. Application of an optimized total N-nitrosamine (TONO) assay to pools:
Placing N-nitrosodimethylamine (NDMA) determinations into perspective.
71. Dai, N.; Mitch, W. A. Relative importance of N-nitrosodimethylamine
compared to total N-nitrosamines in drinking waters. Environ. Sci. Technol.
2013, 47, 3648–3656.
72. Wu, M.; Qian, Y.; Boyd, J. M.; Leavey, S.; Hrudey, S. E.; Krasner, S.
W.; Li, X. F. Identification of tobacco-specific nitrosamines as disinfection
byproducts in chloraminated water. Environ. Sci. Technol. 2014, 48,
1828–1834.
73. California Department of Health - Drinking Water Program. Drinking
Water Notification Levels and Response Levels: An Overview. http://goo.gl/
XvsT74 (accessed November, 2014).
74. Ontario Regulation 169/03. Ontario Drinking Water Quality Standards. Safe
Drinking Water Act, 2002. http://goo.gl/6OTlkl (accessed November 2014).
75. Health Canada. Guidelines for Canadian Drinking Water Quality: Guideline
Technical Document - N-Nitrosodimethylamine (NDMA); Water, Air and
Climate Change Bureau, Healthy Environments and Consumer Safety
Branch, Health Canada, Ottawa, Ontario (Catalogue No H128-1/11-662E)
2011 http://goo.gl/yAUstU (accessed November, 2014).
76. U.S. Environmental Protection Agency (U.S. EPA) Contaminant Candidate
List 3 - CCL 3. http://goo.gl/6M90rO (accessed November, 2014).
211
77. Schreiber, I. M.; Mitch, W. A. Nitrosamine formation pathway revisited: The
importance of chloramine speciation and dissolved oxygen. Environ. Sci.
Technol. 2006, 40, 6007–6014.
78. Choi, J.; Valentine, R. L. N-Nitrosodimethylamine formation by
free-chlorine-enhanced nitrosation of dimethylamine. Environ. Sci. Technol.
2003, 37, 4871–4876.
79. Schreiber, I. M.; Mitch, W. A. Enhanced nitrogenous disinfection byproduct
formation near the breakpoint: Implications for nitrification control. Environ.
Sci. Technol. 2007, 41, 7039–7046.
80. Mitch, W. A.; Gerecke, A. C.; Sedlak, D. L. A N-Nitrosodimethylamine
(NDMA) precursor analysis for chlorination of water and wastewater. Water
Res. 2003, 37, 3733–3741.
Cohen, L.; Sedlak, D. L. N-nitrosodimethylamine (NDMA) as a drinking

water contaminant: A review. Environ. Eng. Sci. 2003, 20, 389–404.
82. Kemper, J. M.; Westerhoff, P.; Dotson, A.; Mitch, W. A. Nitrosamine,
dimethylnitramine, and chloropicrin formation during strong base
anion-exchange treatment. Environ. Sci. Technol. 2009, 43, 466–472.
Berger, R.; Rodigari, F.; Beland, J. W.; Lazzelle, L. J.; Kincannon, E. G.;
coagulated with DADMAC cationic polymer. J. Am. Water Works Assoc.
2003, 95, 94–106+112.
84. Kemper, J. M.; Walse, S. S.; Mitch, W. A. Quaternary amines as nitrosamine
precursors: A role for consumer products? Environ. Sci. Technol. 2010, 44,
1224–1231.
85. Park, S. H.; Piyachaturawat, P.; Taylor, A. E.; Huang, C. H. Potential
N-nitrosodimethylamine (NDMA) formation from amine-based water
treatment polymers in the reactions with chlorine-based oxidants and
nitrosifying agents. Water Sci. Technol. 2009, 9, 279–288.
86. Park, S. H.; Wei, S.; Mizaikoff, B.; Taylor, A. E.; Favero, C.; Huang, C. H.
Degradation of amine-based water treatment polymers during chloramination
as N-nitrosodimethylamine (NDMA) precursors. Environ. Sci. Technol.
2009, 43, 1360–1366.
87. Pehlivanoglu-Mantas, E.; Sedlak, D. L. Measurement of dissolved organic
nitrogen forms in wastewater effluents: Concentrations, size distribution and
NDMA formation potential. Water Res. 2008, 42, 3890–3898.
89. Zhou, W.; Chen, C.; Lou, L.; Yang, Q.; Zhu, L. Formation potential of
nine nitrosamines from corresponding secondary amines by chloramination.
Chemosphere 2014, 95, 81–87.
90. Zhou, W. J.; Boyd, J. M.; Qin, F.; Hrudey, S. E.; Li, X. F. Formation of
N-nitrosodiphenylamine and two new N-containing disinfection byproducts
from chloramination of water containing diphenylamine. Environ. Sci.
Technol. 2009, 43, 8443–8448.
212
Formation, precursors, control, and occurrence of nitrosamines in drinking
water: A review. Water Res. 2013, 47, 4433–4450.
92. Schmidt, C. K.; Brauch, H. J. N,N-dimethylsulfamide as precursor for N-
nitrosodimethylamine (NDMA) formation upon ozonation and its fate during
drinking water treatment. Environ. Sci. Technol. 2008, 42, 6340–6346.
93. Padhye, L.; Luzinova, Y.; Cho, M.; Mizaikoff, B.; Kim, J. H.;
Huang, C. H. PolyDADMAC and dimethylamine as precursors of
N -nitrosodimethylamine during ozonation: Reaction kinetics and
mechanisms. Environ. Sci. Technol. 2011, 45, 4353–4359.
94. Padhye, L. P.; Kim, J. H.; Huang, C. H. Oxidation of dithiocarbamates to
yield N-nitrosamines by water disinfection oxidants. Water Res. 2013, 47,
725–736.
95. Shen, R.; Andrews, S. A. In Nitrosamine Formation from PPCPs in Three

Water Types During Chloramine Disinfection; Water Quality Technology
Conference and Exposition, 2010; pp 1323−1329.
96. Von Gunten, U.; Salhi, E.; Schmidt, C. K.; Arnold, W. A. Kinetics and
mechanisms of N-nitrosodimethylamine formation upon ozonation of
N,N-dimethylsulfamide-containing waters: Bromide catalysis. Environ. Sci.
Technol. 2010, 44, 5762–5768.
97. Chen, P. H.; Richardson, S. D.; Krasner, S. W.; Majetich, G.; Glish, G.
L. Hydrogen abstraction and decomposition of bromopicrin and other
trihalogenated disinfection byproducts by GC/MS. Environ. Sci. Technol.
2002, 36, 3362–3371.
98. Chu, W.; Gao, N.; Yin, D.; Krasner, S. W. Formation and speciation of nine
haloacetamides, an emerging class of nitrogenous DBPs, during chlorination
or chloramination. J. Hazard. Mater. 2013, 260, 806–812.
99. Chu, W. H.; Gao, N. Y.; Deng, Y.; Krasner, S. W. Precursors of
dichloroacetamide, an emerging nitrogenous DBP formed during
chlorination or chloramination. Environ. Sci. Technol. 2010, 44,
3908–3912.
101. Krasner, S. W.; Chinn, R.; Hwang, C. J.; Barrett, S. Analytical methods for
brominated organic disinfection by-products. In Proceedings of the 1990
American Water Works Association Water Quality Technology Conference;
American Water Works Association: Denver, CO, 1991.
102. Richardson, S. D.; Thruston, A. D., Jr.; Caughran, T. V.; Chen, P. H.;
Collette, T. W.; Floyd, T. L.; Schenck, K. M.; Lykins, B. W.; Sun, G.-R.;
Majetich, G. Identification of new ozone disinfection byproducts in drinking
103. Kundu, B.; Richardson, S. D.; Swartz, P. D.; Matthews, P. P.; Richard, A.
M.; DeMarini, D. M. Mutagenicity in Salmonella of halonitromethanes:
A recently recognized class of disinfection by-products in drinking water.
Mutat. Res. 2004, 562, 39–65.
213
104. Kundu, B.; Richardson, S. D.; Granville, C. A.; Shaughnessy, D. T.;
Hanley, N. M.; Swartz, P. D.; Richard, A. M.; DeMarini, D. M. Comparative
mutagenicity of halomethanes and halonitromethanes in Salmonella TA100:
Structure-activity analysis and mutation spectra. Mutat. Res. 2004, 554,
335–350.
105. Choi, J.; Richardson, S. D. Formation of halonitromethanes in drinking water.
In Proceedings of the International Workshop on Optimizing the Design and
Interpretation of Epidemiologic Studies to Consider Alternative Disinfectants
of Drinking Water; U.S. EPA: Raleigh, NC, 2005.
106. Hu, J.; Song, H.; Addison, J. W.; Karanfil, T. Halonitromethane formation
potentials in drinking waters. Water Res. 2010, 44, 105–114.
107. Richardson, S. D.; Thruston, Jr., A. D.; Rav-Acha, C.; Groisman, L.;

108. Yang, M.; Zhang, X. Halopyrroles: A new group of highly toxic disinfection
byproducts formed in chlorinated saline wastewater. Environ. Sci. Technol.
2014, 48, 11846–11852.
109. McDonald, T. A.; Komulainen, H. Carcinogenicity of the chlorination
disinfection by-product MX. J. Environ. Sci. Health - C 2005, 23, 163–214.
110. Schulten, H. R.; Schnitzer, M. A state of the art structural concept for humic
substances. Naturwissenschaften 1993, 80, 29–30.
111. Qin, F.; Zhao, Y. Y.; Zhao, Y.; Boyd, J. M.; Zhou, W.; Li, X. F. A toxic
disinfection by-product, 2,6-dichloro-1,4-benzoquinone, identified in
drinking water. Angew. Chem., Int. Ed. 2010, 49, 790–792.
112. Zhao, Y.; Anichina, J.; Lu, X.; Bull, R. J.; Krasner, S. W.; Hrudey, S. E.; Li, X.
F. Occurrence and formation of chloro- and bromo-benzoquinones during
drinking water disinfection. Water Res. 2012, 46, 4351–4360.
113. Wang, W.; Qian, Y.; Boyd, J. M.; Wu, M.; Hrudey, S. E.; Li, X. F.
Halobenzoquinones in swimming pool waters and their formation from
personal care products. Environ. Sci. Technol. 2013, 47, 3275–3282.
114. Qian, Y.; Wang, W.; Boyd, J. M.; Wu, M.; Hrudey, S. E.; Li, X. F. UV-induced
transformation of four halobenzoquinones in drinking water. Environ. Sci.
Technol. 2013, 47, 4426–4433.
115. Wang, W.; Qian, Y.; Li, J.; Moe, B.; Huang, R.; Zhang, H.; Hrudey, S.
E.; Li, X. F. Analytical and toxicity characterization of halo-hydroxyl-
benzoquinones as stable halobenzoquinone disinfection byproducts in
treated water. Anal. Chem. 2014, 86, 4982–4988.
214
Chapter 12
Carbonaceous and Nitrogenous

Disinfecion By-Product Formation
Potentials of Amino Acids
Meric Selbes,1 Junhong Shan,2 S. Sule Kaplan Bekaroglu,3

and Tanju Karanfil*,4
Fairfax, Virginia 22030, USA
2Public Utilities Board, Singapore 608576, Singapore
3Water Institute, Suleyman Demirel University, Isparta 32260, Turkey
Clemson University, Anderson, South Carolina 29625, U.S.A.

Nine amino acids (AAs) were examined under different

oxidation conditions to determine their formation potential for
regulated carbonaceous DBPs (trihalomethanes (THMs) and
haloacetic acids (HAAs)) and selected emerging nitrogenous
DBPs (haloacetonitriles (HANs), halonitromethanes (HNMs)
and nitrosamines). Aspartic acid and histidine exhibited
very high dihalogenated-HANs and HAAs formation during
chlorination, but their reactivity significantly decreased after
ozonation. Glycine followed by lysine yieled the highest level
of HNMs and THMs formation during ozonation-chlorination,
but they did not have measurable yields for either classes of
DBPs during chlorination. All other AAs showed very low
carbonaceous and nitrogenous DBP yields. The nitrosamine
yields of all nine AAs were very low or below the minimum
reporting levels during chloramination, ozonation, and
ozonation-chloramination. These results indicated that
the presence of AAs in natural waters can result in some
contributions to certain halogenated DBPs depending on the
oxidation conditions. However, they do not appear to be an
important contributor to nitrosamines formation.

Introduction
Recent research has shown that emerging nitrogenous disinfection
by-products (N-DBPs) exhibit orders of magnitude higher cyto- and geno-toxicity
than the regulated carbonaceous DBPs (C-DBPs) (1). Therefore, it should not
be surprising to see N-DBP regulated in the near future. Recent research has
shown that nitrogen-rich organic materials in natural waters play an important
role in the formation of N-DBPs (2–4). However, the important precursors and
the formation mechanisms of N-DPBs still remain largely unknown.
Amino acids (AAs) have been found in fresh waters in a wide range of
concentrations from 5 to 2000 µg/L, in either free or combined peptides, nucleic
acids, purines, pyrimidines, and proteins (5, 6). Thurman (7) reported that the
total AAs, which were the sum of the free and combined AAs, accounted for
2.6% of the dissolved organic carbon (DOC) and 35% of the dissolved organic
nitrogen (DON) in some lakes. Hagedorn and colleagues (8) observed in their
studies of catchment runoff a 20% to greater than 75% of the DON of the total
AAs observed. Elevated amino acid levels were also found in the presence of
algae blooms, the degradation of which is a major cause of increased AA levels
in natural waters (7, 9–11). In a recent survey of sixteen water treatment plants
in the United States, the total AA concentrations, identified as glycine, glutamic
acid, alanine, aspartic acid, leucine, proline and serine, constituted an average
15% of the DON in the source waters (2, 7, 12–14).
The presence of AAs in raw and treated waters requires substantial amounts
of chlorine for treatment (15, 16). The relative chlorine reactivity of the AAs
depends on the side chain groups attached to the α-carbon. Studies conducted for
purposes of reacting AAs with chlorine clearly indicated the formation of various
classes of DBPs including haloacetaldehydes, haloacetonitriles (HANs), cyanogen
chloride, trihalomethanes (THMs) and haloacetic acids (HAAs) (15, 17, 18). A
recent study showed that AAs with activated aromatic structures (e.g. tryptophan,
tyrosine) were potent precursors of THMs and HAAs (17). Additionally, absent
the reactive ring structure, aspartic acid and asparagines produced high levels of
HAAs.
Investigations conducted on the reaction(s) of ozone with AAs determined
that the side chains of AAs were responsible for the high ozone reactivity in
polypeptide structures (19). Structures that possess side groups such as amino
nitrogen or activated aromatic ring were identified as the most reactive AAs.
This AA ozonation leads to the formation of aldehydes such as formaldehyde,
acetaldehyde, glyoxal and glyoxal derivatives.
Unlike the carbonaceous DBPs (C-DBPs), very little is known regarding
how N-DBPs form from amino acids, especially for halonitromethanes (HNMs)
and N-nitrosamines. HNM constitutes one group of N-DBPs that exhibits high
degrees of cyto- and geno-toxicity (1, 21). Drinking waters and wastewater
effluents under chloramination conditions are a particularly relevant medium in
which nitrosamine, especially N-nitrosodimethylamine (NDMA), can form (20).
Nitrosamines, which have been classified as a probable human carcinogen by
the United States Environmental Protection Agency, can pose important health
risks even at ng/L concentrations. As a result, the Canada Ontario Ministry of
216
the Environmental and Energy established a maximum allowable concentration
of 9 ng/L for NDMA, and the California Department of Health Service set an
interim action level of 10 ng/L. Although nitrosamines are not currently regulated
in the United States, NDMA and four other nitrosamines [N-nitrosodiethylamine
(NDEA), N-nitroso-di-n-propylamine (NDPA), N-nitrosodiphenylamine
(NDPhA), N-nitrosopyrrolidine (NPYR)] are on the USEPA’s Contaminant
Candidate List-3, and are monitored under the Unregulated Contaminant
Monitoring Rule 2.
Recent studies on the formation potentials of HNMs in natural waters found
that HNM yields increased with decreasing DOC/DON ratios (i.e., increasing
organic nitrogen content per organic carbon in water) during ozonation followed
by chlorination (3). The hydrophilic components of dissolved NOM, especially
nitrogenous organic compounds, are important precursors of HNMs in the NOM
pool (2, 3). Only a limited number of studies, however, have been conducted to
elucidate trichloronitromethane (TCNM) formation by AAs during chlorination,
using glycine (4, 22), tyrosine (4), tryptophan and aspartic acid (23). Similarly,
very little is known about of the formation of NDMAs caused by AAs, aside
from that undertaken by Mitch et al. (4). Specifically, they observed nitrosamine
formation (NDMA, NMEA and NDEA) in aspartic acid, proline and histidine
after chloramination (24); and NDMA formation of <2ng/L in glycine after
chloramination (4).
Produced during chlorination HANs consist of an acetate derivative group,
known as a cyano (-CN) group, and up to three halogens replacing the hydrogen
atoms. Dichloroacetonitrile (DCAN), the most prevalent of the HANs, was
detected at 0.3-8.1 µg/L in 10 chlorinated drinking waters in southern Ontario
(25), and at median and maximum concentrations of 1 and 12 µg/L, in 12
wastewater treatment plans in the US (26). In both studies, DCAN levels were
10% of the molar concentration for THMs. DCAN formation also occurred when
amino acid moieties were used to during the chlorination of AAs, polypeptides,
and hydrophobic substances (27). Moreover, aspartic acid (16, 23) and alanine
(28) used in chlorination has caused the formation of HANs from AAs.
Since AAs constitute an important fraction of organic nitrogen pool in natural
waters, and previous studies have mainly focused on the formation regulated C-
DBPs (e.g., THM and/or HAA) from AAs mostly for chlorination, the objective of
this study was to concurrently investigate formation potential of both regulated C-
DBPs (THMs and HAAs) and selected N-DBPs (HANs, HNMs and nitrosamines)
from AAs ozonation, chloramination and chlorination conditions. The results were
also used to provide additional insight to the DBP formation mechanisms from
AAs.
Materials and Methods

Amino Acids
Nine AAs (alanine, aspartic acid, cysteine, glutamic acid, glycine, lysine,
histidine, proline and serine) were carefully selected based on charge property,
polarity and hydropobicity. They were classified into four groups depending upon
217
their acidity and polarity: acidic, basic, polar, and nonpolar. The physicochemical
characteristics and structures of the selected AAs are listed in Table 1.
Table 1. AAs Selected for This Study

218
Amino acids were purchased from a certified vendor (Sigma-Aldrich) and a
500 mg/L stock solution of each AA was prepared in deionized distilled water. For
the formation potential (FP) tests, 1 mg/L (and 10 mg /L in selected experiments)
AA sample solutions were prepared with dilution from the main stock. These
solutions were buffered at pH 6.0 or 8.0 using 4 mM sodium bicarbonate and
adjusted with 1M HCl or NaOH. The AA concentrations (1 or 10 mg/L) in the
diluted solutions were confirmed using a TOC analyzer (Shimadzu Corp., USA).
Although these AA concentrations are higher than their typical occurrence levels
in fresh waters, they were intentionally selected at these high levels to magnify
and better examine the DBP formation; an approach that has been used in previous
studies (4, 29).
Formation Potential Tests
The solutions of chlorine stock (500–2,000 mg/L) were prepared by

diluting sodium hypochlorite (~5% available free chlorine). The preformed
monochloramine stock solutions (1,000 mg/L) were generated by mixing the
sodium hypochlorite and ammonium sulfate solutions at a Cl2-to-N mass ratio
of 4:1 at pH of 9. The ozone stock solutions (28–32 mg/L) were produced
using an ozone generator (Griffin Technics, GTC–1B) fed with ultra-high-purity
oxygen gas. These oxidants were used to determine the precursor levels of each
class of DBPs under their favorable disinfection conditions. Chlorination and
ozonation-chlorination were used in the FP tests to examine the formation of the
following halogenated DBPs: THMs, HAAs, HANs and HNMs. Chloramination,
ozonation and ozonation-chloramination were applied for the nitrosamine FP
tests, which were conducted in the presence of an excess disinfectant. The
formula developed by Krasner and co-workers (30) (Equation 1 and 2) was used
to determine the chlorine dosage.
Based on the Equation 1, typical chlorine doses were around 11-12 mg/L for 1
mg/L AA concentration. Ozonation was carried out by adding ozone stock solution
to the samples approximately at 4 mg/L (Equation 3). Following ozone addition,
the samples were mixed on a stir plate for 5 min. After 5 minutes, they were also
exposed same doses of chlorine. Prior to chlorination presence of residual ozone
confirmed and thus, ozone was not a limiting factor during ozonation period. Since
10 mg/L AA concentration was used for nitrosamine FP test conditions, initial
chloramine dose was 10-15 mg/L. Although AAs have high oxidant demands
(i.e., chlorine, ozone), the levels of oxidants were sufficient to maintain a residual
concentration at the end of the reaction time.
219
Table 2. Analytical Methods and Minimum Reporting Levels
Parameter Unit Measurement Method Instrument MRL or Accuracya
DOCb mg/L SM 5310B TOC-VCHS & TNM-1, Shimadzu Corp., Japan 0.1
DN (=DON)c mg/L High Temperature Combustion TOC-VCHS & TNM-1, Shimadzu Corp., Japan 0.1
pH - SM 4500-H+ 420A, Orion Corp., USA ±0.01d
THMs & HANse µg/L US EPA Method 551.1 6890 GC-ECD, Agilent, USA 1.0
HAAsf µg/L SM 6251 Bg 6890 GC- ECD, Agilent, USA 1.0
HNMsg µg/L US EPA Method 551.1 6850 GC-ECD, Agilent, USA 0.7
Nitrosaminesh ng/L US EPA 521 Varian GC/MS/MS 3.0
FAC mg/L SM 4500-Cl F NA 0.1
220
a MRLs were determined in the lab in previous works (31). Accuracy as reported by the manufacturer. b Reagent grade potassium hydrogen phthalate
was used to prepare external standards. c Reagent grade potassium nitrate was used to prepare external standards. Since there was no inorganic nitrogen
present, measured DN values were equal to DON. d Accuracy (pH units). e THMs and HANs were extracted by liquid-liquid extraction with methyl-
tert butyl ether (MtBE) and analyzed by GC-μECD. f HAAs were extracted by liquid-liquid extraction with MtBE, derivatized with diazomethane and
analyzed by GC-μECD. g HNM were extracted by liquid-liquid extraction with MtBE and analyzed by GC-μECD. h Nitrosamines were extracted by solid
phase extraction, eluted with dichloromethane and analyzed by GC-MS. DOC (Dissolved Organic Carbon), DN (Dissolved Nitrogen), DON (Dissolved
Organic Nitrogen), THM (Trihalomethanes), HAN (Haloacetonitriles), HAA (Haloacetic acids), HNM (halonitromethanes), SM (Standard Methods), EPA
(Environmental Protection Agency), GC (Gas Chromatography), MS (Mass Spectrometer), FAC (Free Available Chlorine), NA (Not Applicable).

Experiments for THM, HAA, HAN and HNM FP tests were conducted in
125-ml amber glass bottles without headspace at room temperature (~22°C) in the
dark for one day. For the nitrosamines FP test, 1-L amber glass bottles were used
due to higher sample volume required for extraction, and the reaction time was
seven days.
Analytical Methods
The analytical methods used in the study and their respective minimum
reporting levels (MRLs) are provided in Table 2. Either standard methods (SM)
or USEPA methods were used in the DBP analyses with minor modifications.
The detailed information about these methods can be found elsewhere (31). For
DOC, DN and DBP analyses, samples were analyzed in duplicates. The error
bars in all the graphs show the variability due to multiple analysis (n=2) under
the same conditions.

The FP results have been discussed individually for each class of DBP,
specifically for the experiments conducted at a pH of 8.0. An overall evaluation
of the pH level (6.0 vs. 8.0) is provided at the end.
Nitrogenous-DBPs
Halonitromethanes (HNMs)
The formation of all HNM species from nine amino acids after chlorination
was below MRL, which is in agreement with the reports in the literature. Merlet
et al. (22) reported <0.01% mole/mole yield rate (<0.7 µg/mg-C) of TCNM from
glycine after chlorination at pH 7.1. A recent report by Mitch et al. (4) also
showed low yields of TCNM from glycine and tyrosine during chlorination FP
tests (0.8 µg/L from 64 µM glycine (4.8mg/L)). No TCNM formation was found
from chlorinated tryptophan and aspartic acid (23).
Ozonation followed by chlorination was more favorable than chlorination
for HNM formation, as previously observed in natural waters (3, 26, 32–35).
Furthermore, trihalogenated HNMs (i.e., TCNM) was the major species detected
which is consistent with the literature (34). The ozonation of amines in aqueous
solutions can lead to oxidation at both the amine and on the side chain of
the molecule (19, 36). The higher TCNM formation caused by AAs during
pre-ozonation may be due to the oxidation of the amine group by the strong
oxidant ozone to form a nitro group (22, 37). The nitro compound will, then, be
further substituted to form nitromethanes and react with chlorine. TCNM was the
only HNM species formed in our study, with glycine showing remarkably high
TCNM formation (223.4 µg/L), followed by lysine and proline (14.6 and 17.2
µg/L). We also detected a small amount of TCNM in glutamic acid (0.8 µg/L)
and serine (0.9 µg/L), while for the remainder of the AAs examined, the HNM
221
concentrations were either below MRL or were not detected. A comparison
of our AA result are plotted and normalized in Figure 1 based on their DOC
(µg-TCNM/mg-DOC). The levels of TCNM during ozonation-chlorination were
comparable with our previously reported findings (32).
These findings were used to provide additional insight to the formation
mechanism of TCNM from glycine. The ozonation of glycine is known to
form nitrate, form, formic acid and carbon dioxide as by-products (38). Under
certain conditions, the formation of formic acid can be relatively high (>0.05
mM formed from 1.5 mM O3 vs. 1.1 mM glycine, in the presence of 0.2 mM
hydrogenocarbonate ions) (29), suggesting the possible rupture of the C-C
bond and the subsequent formation of nitromethanes. It has been hypothesized
that the rupture of the C-C bond may occur with the formation of a strongly
reducing α-aminoradical in an unprotonated amino group; this rupture may
occur when the electron abstraction from the amine nitrogen is then re-arranged
with decarboxylation (39). It is also hypothesized that the reaction follows the
oxidation carboxyl from ozone to peroxide (37), followed by decarboxylation
(29, 39), which in turn causes the oxidation of the amine group to the nitro group
(22, 37). A halogenation reaction with chlorine, which causes TCNM formation,
is the result.
The greater TCNM yield from lysine after ozonation-chlorination was
attributed to the production of glycine and glycine-like intermediates formed
during ozonation. The long chain structure of α-C of lysine makes the breakage
of the C-C bond easier compared to other amino acids examined in this study.
Figure 1. TCNM formation potential results from amino acids during

ozonation-chlorination at pH 8.0 (DOC = 0.3 ~ 0.5 mg/L, DON = 0.04 ~ 0.30
mg/L). TCNM formation during chlorination were below MRL and thus, were not
included in this figure.
222
Haloacetonitriles (HANs)
In this study, we also observed aspartic acid to form high levels of DCAN
(549.2 µg/mg-C) during chlorination at a pH of 8.0, with histidine the second most
reactive AA, forming 173.7 µg/mg-C (seen Figure 2). The high level of DCAN
FP of aspartic acid has been validated in two previous studies: that of Trehy et
al. (16) in which they used 158 µg/mg-C at pH of 6.4; and that of Bond et al.
(23) in which they used 0.06 mol/mol 130 µg/mg-C) at pH of 7. Both studies
concur with a study by Ram et al. in which the chlorination of AAs, polypeptides,
and hydrophobic substances with amino acid moieties were used to create DCAN
(27). Although Chu et al. (28) reported the formation of non-detectable HANs
(DCAN and TCAN) in 0.1 mM of alanine after chlorination, in our study we
observed a DCAN formation level of only 3.9 µg/mg-C DCAN after chlorination.
We hypothesize that the different experimental conditions, such as pH or oxidant

dosage is perhaps the reason for this greater alanine yield. We also observed the
formation of 6.1 µg/mg-C chloroacetonitrile (CAN).
Figure 2. DCAN formation potential results from amino acids during chlorination
and ozonation-chlorination at pH 8.0 (DOC = 0.3 ~ 0.5 mg/L, DON = 0.04 ~
0.30 mg/L).
In their chlorination studies of free and combined amino acids, Hureiki et

al. (15) also determined that the chlorination of AAs can cause nitrile formation.
Specifically, after chlorination they observed a substantial formation of DCAN
from histidine, but not for alanine, suggesting that the DCAN is not formed from
the side chain of histidine. If the substitution reaction for chlorine to replace
the side chain of the heterocyclic ring in histidine occurs, an intermediate with
a similar structure to alanine should form. To produce the DCAN, a ring-opening
reaction must occur during the chlorination of histidine. This reaction is possible
223
because the activated aromatic ring was the essential reactive site for tyrosine with
its high chlorine demand and high TCM formation (15). For pre-ozonation, the
oxidation with ozone likely results in the nitro compounds formation and thus,
reduces the formation of nitriles. This oxidation can explain the dramatic decrease
in the DCAN formation compared to chlorination (e.g. from 549.2 to 89.2 µg/mg-
C for aspartic acid and from 173.7 to 8.2 µg/mg-C for histidine).
Figure 3. Proposed reaction mechanism of DCAN formation from aspartic acid

during chlorination. (Figure Adopted from Hureiki et al. (15)).
Nitrosamines
Initial nitrosamines FP test for three selected AAs at 1.0 mg/L concentration
did not produce measurable nitrosamines. To further confirm the results, it was
decided to magnify the initial concentration of AAs during the experiments by
increasing to 10.0 mg/L. Three different oxidation scenarios, monochloramination,
ozonation and ozonation-chloramination were tested for all nine AAs. Ozonation
was also examined independently because of the recently reported NDMA
formation after ozonation in both the lab (40, 41) and in full scale ozonation
plants (42, 43).
224
Despite this increase to 10 mg/L, NDMA and NDBA FP concentrations
remained very low at levels of 5 ng/L NDBA from lysine after chloramination
(Table 3). Mitch and co-workers reported non-detectable nitrosamine (NDMA,
NMEA and NDEA) formation from aspartic acid, proline and histidine during
chloramination (24); and NDMA formation of <2 ng/L from glycine and tyrosine
during chloramination (4).
Table 3. Nitrosamine FPs of AAs Tested in This Study

Ozone-
DOC DON Chloramine Ozone
Chloramine
Amino Acids* (mg (mg
C/L) N/L) NDBA NDBA NPYR NPYR

(ng/L) (ng/L) (ng/L) (ng/L)
Alanine 4.0 1.6 <MRL <MRL <MRL <MRL
Aspartic Acid 3.6 1.1 <MRL <MRL <MRL <MRL
Cysteine 3.0 1.2 <MRL <MRL <MRL <MRL
Glutamic Acid 4.1 1.0 <MRL <MRL <MRL <MRL
Glycine 3.2 1.9 <MRL <MRL <MRL <MRL
Histidine 4.6 2.7 <MRL <MRL <MRL <MRL
Lysine 4.9 1.9 5 9 <MRL <MRL
Proline 5.2 1.2 <MRL 3 4 4
Serine 3.4 1.3 <MRL 3 <MRL <MRL
*Initial concentration of amino acids were 10 mg/L. Other nitrosamines including
NDMA, NDEA, NMEA, NMOR, NDPA, and NPIP were not detected.
Principally all amines with the exception of primary amines (i.e., amino
acids) have the potential to form nitrosamines (44). During chloramination,
only lysine was observed to form NDBA (5 ng/L). This may be due to reaction
of butyl chains of two lysine’s with chloramines. Since ozone is a stronger
oxidant, it is likely to form intermediates that are more prone to reaction (45).
Thus, during ozonation-chloramination NDBA yield from lysine increased to 9
ng/L. In addition, during ozonation-chloramination, NDBA formation of 3 ng/L
from proline and serine was also observed. Proline also lead to formation of
4 ng/L NPYR during both ozonation and ozonation-chloramination conditions.
Other nitrosamines were undetected. NPYR is formed from proline because
of its structure; once ozonation causes the decarboxylation of proline followed
by nitrosation, the nitrogen of the nitrosating agent is likely sourced from the
oxidation of the nitrogen atom in the proline’s ring. The direct formation of
nitrosamines during ozonation hasl also occurred with other precursors (45). The
authors observed very low nitrosamine yields of AAs during the FP tests (<0.01%
molar conversion). Considering the occurrence concentrations of total AAs in
225
natural waters, it is unlikely that AAs will play a role in the formation of NDMA
and other nitrosamines during chloramination, ozonation, and ozonation that is
followed by chloramination.
Carbonaceous-DBPs
Trihalomethanes (THMs)
Trichloromethane (TCM) was the only THM species observed during both
chlorination and ozonation-chlorination (Figure 4), with the selected AAs in this
study forming TCMs within the range of <MRL-13.5 µg/L after chlorination.
These values are consistent with that previously reported by the authors (32).
Ozonation can increase in TCM formation. Bond et al. (23) reported that 0.02
mol TCM formed from 1 mol aspartic acid after chlorination (i.e., 49.7 µg/mg-C),
which is relatively comparable to the yield observed in this study, 18.8 µg/mg-
C. Chu et al. (28) reported a yield rate of 0.23% (i.e., 5.7 µg/mg-C) for TCM
formation from 0.1 mM alanine after chlorination, while no TCM formation from
alanine was measurable during chlorination in this study. Hong et al. (17) studied
chlorination of all 20 amino acids at pH 7 and found low THM formation (<4.9
µg/mg-C), except for tryptophan and tyrosine. The results of our study showed
that pre-ozonation prior to chlorination had a drastic effect on TCM formation
from glycine and lysine which was similar to HNMs. TCMs formation of glycine
and lysine increased from <MRL to 83.4 and 33.9 µg/L, respectively. Other AA’s
TCM formations during ozonation-chlorination were comparable and were within
the range of 6.7-11.8 µg/L.
Reaction sequences have been developed for the molecular ozone attack on
amines (46), which suggests that amides and formylamine are caused by i) the
breakdown of the carbon-nitrogen bond, ii) the formation of inorganic nitrogen
and hydroxylamine, oxime or amine oxide, and iii) by the oxidation of the
carbon.. All these reactions were shown to involve decarboxylation. Specifically,
a study of glycine (29) determined that, for AAs, the intermediate created from
the addition of ozone either released oxygen to create monohydroxylamine or
caused carbon oxidation via a probable ozonide intermediate. Also the presence
of oxamic and oxalic acids indicates that these reactions can occur without
decarboxylation. Based on these findings reported in the literature, ozone
should first launch an electrophilic attack on nitrogen and followed by two steps
of oxidation on N with or without decarboxylation to form hydroxylamines
(R-NHOH) and hypothetical hydronitrous acid, N(OH)2•. Then oxime will be
formed through dehydration and further oxidized to form aldehyde which has
been previously reported in the literature (29). Since aldehyde could be generated
by the hydrolysis of dihalogenated compound, the backward reaction could lead
the formaldehyde to TCM with further halogenation during chlorination.
A significant amount of THM formation in lysine after ozonation-chlorination
(Figure 4), prompted a comparison of the molar yields of DBP/amino acid of TCM
and TCNM from glycine and lysine under ozonation-chlorination. The yields for
glycine was 5.24% mole TCM and 10.20% mole TCNM per mole of glycine,
respectively, and 4.15% mole TCM and only 1.29% of TCNM for lysine. These
226
results suggested that the mechanism for TCM formation from lysine may not
follow the pathway proposed for TCNM. It is possible that intermediates other
than glycine may form. In their study of the formation trends of THM and HNM
Hu et al. (34) also observed a divergence in formation pathways. The findings of
this study would explain these previously observed trends.
Figure 4. TCM formation potential results from amino acids during chlorination
and ozonation-chlorination at pH 8.0 (DOC = 0.3 ~ 0.5 mg/L, DON = 0.04 ~
0.30 mg/L).
Haloacetic Acids (HAAs)
The most reactive AA that caused DCAA formation was aspartic acid, which
yielded a formation of 992.5 µg/mg-C during chlorination (Figure 5), followed
by histidine, which yielded a formation of 336.3 µg/mg-C. These results were
validated by the earlier study of Hong et al. (17), who also confirmed that of the
20 amino acids studied, aspartic acid and histidine caused the highest DCAA
formation during chlorination at pH 7.0. Also since the DCAN produces DCAA
after hydrolysis, these results were also consistent with HAN findings (15, 16, 47).
Aspartic acid is also known to function as a reactive DCAA precursor, forming
0.26 mol/mol (693 µg/mg-C) after chlorination at pH 7.0 (23). Its reactivity
is thought to result from the formation of a β-keto acid intermediate upon
chlorination, a moiety known to have a high DCAA FP (15, 48). In terms of HAA
speciation, aspartic acid and histidine (basic and hydrophilic AAs) formed very
high concentrations of DCAA, whereas the formation of TCAA was significantly
lower. This finding may be partially the result of the hydrophobicity of the
227
surrogate which was reported by Bond et al. (23). In that study, it was determined
that the hydrophilic-acid or transphilic-acid surrogates (i.e., aspartic acid) formed
predominantly DCAA during chlorination. In contrast, the hydrophobic-acid and
hydrophobic-neutral surrogates (i.e., tryptophan and tyrosine) formed mainly
TCAA.
Figure 5. DCAA and TCAA formation potential results from amino acids during
chlorination and ozonation-chlorination at pH 8.0 (DOC = 0.3 ~ 0.5 mg/L, DON
= 0.04 ~ 0.30 mg/L).
228
Table 4. DBP FPs of AAs Tested in This Study at pH 6.0 and 8.0
TCNM DCAN TCM DCAA TCAA
Amino DOC DON (µg/L) (µg/L) (µg/L) (µg/L) (µg/L)
Acid* (mg/L) (mg/L) pH pH

pH 6.0 pH 8.0 pH 6.0 pH 8.0 pH 6.0 pH 8.0 pH 6.0 pH 8.0
6.0 8.0
Alanine 0.40 0.16 <MRL <MRL <MRL 1.6 <MRL <MRL <MRL <MRL <MRL <MRL
Aspartic
0.36 0.11 <MRL <MRL 211.3 197.7 <MRL 6.8 106.5 357.3 1.0 1.3
Acid
Cysteine 0.30 0.12 <MRL <MRL <MRL <MRL 1.4 <MRL <MRL 1.0 <MRL <MRL
Glutamic
0.41 0.10 <MRL <MRL <MRL <MRL <MRL 1.3 2.4 3.8 1.5 3.2
Chlorina- Acid
tion
229
Glycine 0.32 0.19 <MRL <MRL <MRL <MRL <MRL <MRL <MRL <MRL <MRL <MRL
Histidine 0.46 0.27 <MRL <MRL 73.2 79.9 2.1 13.5 42.4 154.7 13.6 20.9
Lysine 0.49 0.19 <MRL <MRL <MRL <MRL 2.5 <MRL <MRL 1.9 <MRL 1.4
Proline 0.52 0.12 <MRL <MRL <MRL 1.5 <MRL 1.3 <MRL 5.9 <MRL 2.1
Serine 0.34 0.13 <MRL <MRL <MRL <MRL <MRL <MRL <MRL 3.5 <MRL 1.4

Table 4. (Continued). DBP FPs of AAs Tested in This Study at pH 6.0 and 8.0
TCNM DCAN TCM DCAA TCAA
Amino DOC DON (µg/L) (µg/L) (µg/L) (µg/L) (µg/L)
Acid* (mg/L) (mg/L) pH pH

pH 6.0 pH 8.0 pH 6.0 pH 8.0 pH 6.0 pH 8.0 pH 6.0 pH 8.0
6.0 8.0
Alanine 0.40 0.16 <MRL <MRL <MRL 1.1 <MRL 8.4 <MRL <MRL <MRL <MRL
Aspartic
0.36 0.11 <MRL <MRL 183.7 32.1 <MRL 9.8 41.6 128.4 6.0 11.4
Acid
Cysteine 0.30 0.12 <MRL <MRL <MRL <MRL <MRL 6.7 <MRL <MRL <MRL <MRL
Ozona- Glutamic
0.41 0.10 <MRL 0.8 2.4 <MRL 1.5 11.8 7.4 27.8 1.9 8.8
tion- Acid
Chlori-
230
Glycine 0.32 0.19 25.9 223.4 <MRL 2.5 7.4 83.4 <MRL <MRL <MRL <MRL
nation
Histidine 0.46 0.27 <MRL <MRL 6.4 3.8 1.1 6.9 12.7 33.3 3.3 8.5
Lysine 0.49 0.19 12.0 14.6 1.5 <MRL 10.5 33.9 12.9 9.8 1.5 3.1
Proline 0.52 0.12 16.4 17.3 <MRL <MRL 5.0 11.5 7.2 16.6 1.2 2.5
Serine 0.34 0.13 <MRL 0.9 <MRL <MRL 2.0 8.4 1.3 1.5 1.1 <MRL
*Initial concentration of AA’s were 1mg/L. DOC and DON values are calculated based AA concentration and AA formulas. MRL: Minimum Reporting
Limit.

Pre-ozonation dramatically decreased the DCAA formation from aspartic acid
and histidine, from 992.5 µg/mg-C and 336.3 µg/mg-C to 356.8 µg/mg-C and
72.3 µg/mg-C, respectively (Figure 5). These finding were consistent with DCAN
findings that are explained in the next section. No obvious effect of pre-ozonation
was observed on DCAA formation from either the AAs or TCAA formations under
study.
The Effect of pH on the Formation of DBPs
The results show that the THMs and HAAs formation increased with
increasing pH (Table 4). It has been previously reported in the drinking water
literature that THM formation decreases with a decrease in pH, while this
decrease in pH leads to an increases in HAAs formation (49). However, in

this study, decreasing trend was observed for both THMs and HAAs formation
with the decreases in pH which were consisted with the previous findings (32).
Furthermore, the decrease in pH from 8.0 to 6.0 also decreased the yields of
HNMs and HANs. The only exception was observed with the formation of DCAN
from aspartic acid. It has been also reported that decreases in pH decreased
formation of HNMs in natural waters (34, 50).
Also the protonated states of the AAs, which reduced their reactivity also
decreased the DBP yields, indicating that alkali conditions facilitated the formation
of DBPs for the selected AAs. The extent of this increase was compound-specific,
however. Moreover, the distinct DBP yield differences observed in our study
highlights the sensitivity of AAs to pH and also explains the differences observed
in the literature. Overall, the results indicate that reducing pH from 8.0 to 6.0
would result in distinct decreases in the formation of THMs, HAAs, HANs, and
HNMs because the AAs are highly sensitive to changes in pH.
Conclusions
The C-DBP and N-DBP formation potentials of AAs tested in this study
were summarized in Table 4. Of the nine AAs analyzed, aspartic acid, glycine
and histidine showed the highest reactivity, which contributed to the formation
of N-DBPs and C-DBPs. Aspartic acid also exhibited the highest DCAN and
DCAA formation during both chlorination and ozonation-chlorination. Under the
same conditions, the highest HNM and THM formation was formed from glycine,
although it did not yield HNM and THM during chlorination. HNM formation
from glycine under ozonation-chlorination condition was remarkably higher than
all AAs tested. Further, although a substantial amount of TCM, DCAA, TCAA
and DCAN was formed with the introduction of histidine during chlorination,
the reactivity was lower for chlorination after ozonation. All other AAs showed
very low C-DBP and N-DBP yields, with nitrosamine yields from all nine AAs
during chlorination either being very low or below the minimum reporting levels.
Furthermore, the DBP yield of the selected AAs was higher at pH 8.0 than 6.0
with the only exception of DCAN formation from aspartic acid.
231
Aspartic acid and glycine are the most commonly detected AAs in natural
waters. Although the total AA concentrations in natural waters were low, their
elevated concentrations during seasonal events (e.g., algae blooms, die-off, run
off) and the presence of aspartic acid and glycine type structures in the organic
nitrogen molecules in natural waters caused a partial generation of certain C-DBPs
and N-DBPs depending on the oxidation conditions. The results from this study
indicate that AAs are an unlikely contributor to NDMA formation because of the
very low yields generated during chloramination.
Acknowledgments
This work was partly supported by a research grant from the National Science
Foundation (CBET 106657). However, the manuscript has not been subjected to
the peer and policy review of the agency and therefore does not necessarily reflect
its views.
References
1. Plewa, M. J.; Wagner, E.; Muellner, M.; Hsu, K.; Richardson, S. ACS Symp.
Ser. 2008, 995, 36–50.
2. Dotson, A.; Westerhoff, P.; Krasner, S. W. Water Sci. Technol. 2009, 60,
135–143.
3. Hu, J.; Song, H.; Addison, J. W.; Karanfil, T. Water Res. 2010, 44, 105–114.
4. Mitch, W. A.; Krasner, S. W.; Westerhoff, P.; Dotson, A. Occurrence
and formation of nitrogenous disinfection by-products; Water Research
Foundation Report: Denver, CO, 2009.
5. Ram, N. M.; Morris, J. C. Environ. Int. 1980, 4, 397–405.
6. Rice, R. G.; Gomez-Taylor, M. Environ. Health Perspect. 1986, 69, 31–44.
7. Thurman, E. M. Organic Geochemistry of Natural Waters; Martinus Nijhoff/
Dr W. Junk Publishers: Dordrecht, Netherlands, 1985.
8. Hagedorn, F.; Schleppi, P.; Waldner, P.; Fluhler, H. Biogeochemistry 2000,
50, 137–161.
9. Meon, B.; Kirchman, D. L. Mar. Chem. 2001, 75, 185–199.
10. Sellner, K. G.; Nealley, E. W. Mar. Chem. 1997, 56, 193–200.
11. Jørgensen, N. O. G. Limnol. Oceanogr. 1987, 32, 97–111.
12. Dotson, A.; Westerhoff, P. J. Am. Water Works Assoc. 2009, 101, 101–115.
13. Chinn, R.; Barrett, S. E. ACS Symp. Ser. 2000, 761, 96–108.
14. Münster, U. Environ. Int. 1999, 25, 209–224.
15. Hureiki, L.; Croue, J. P.; Legube, B. Water Res. 1994, 28, 2521–2531.
16. Trehy, M. L.; Yost, R. A.; Miles, C. J. Environ. Sci. Technol. 1986, 20,
1117–1122.
17. Hong, H. C.; Wong, M. H.; Liang, Y. Arch. Environ. Contam. Toxicol. 2009,
56, 638–645.
18. Na, C.; Olson, T. M. Environ. Sci. Technol. 2006, 40, 1469–1477.
19. Hureiki, L.; Croue, J. P.; Legube, B.; Dore, M. Ozone Sci. Eng. 1998, 20,
381–402.
232
20. Sacher, F.; Schmidt, C. K.; Lee, C.; Von Gunten, U. Strategies for minimizing
nitrosamine formation during disinfection; Water Research Foundation
Report: Denver, CO, 2008.
M. Mutat. Res., Rev. Mutat. Res. 2007, 636, 178–242.
22. Merlet, N.; Thibaud, H.; Dore, M. Sci. Total Environ. 1985, 47, 223–228.
23. Bond, T.; Henriet, O.; Goslan, E. H.; Parsons, S. A.; Jefferson, B. Environ.
Sci. Technol. 2009, 43, 5982–5989.
24. Mitch, W. A.; Sedlak, D. L. Water Sci. Technol.: Water Supply 2002, 2,
191–198.
25. Oliver, B. G. Environ. Sci. Technol. 1983, 17, 80–83.
Sclimenti, M. J.; Onstad, G. D.; Thruston, A. D. Environ. Sci. Technol. 2006,
40, 7175–7185.
27. Ram, N. M. Environ. Int. 1985, 11, 441–451.
28. Chu, W. H.; Gao, N. Y.; Deng, Y.; Dong, B. Z. Chemosphere 2009, 77,
1346–1351.
29. Berger, P.; Leitner, N. K. V.; Dore, M.; Legube, B. Water Res. 1999, 33,
433–441.
30. Krasner, S. W.; Westerhoff, P.; Chen, B.; Rittmann, B. E.; Nam, S. N.;
Amy, G. Environ. Sci. Technol. 2009, 43, 2911–2918.
31. Gan, X.; Kim, D.; Karanfil, T. J. Am. Water Works Assoc. 2013, 105,
E195–E206.
32. Shan, J.; Hu, J.; Kaplan-Berkaroglu, S. S.; Song, H.; Karanfil, T.
Chemosphere 2012, 86, 323–328.
33. Song, H.; Addison, J. W.; Hu, J.; Karanfil, T. Chemosphere 2010, 79,
174–179.
34. Hu, J.; Song, H.; Karanfil, T. Environ. Sci. Technol. 2010, 44, 794–799.
35. Hoigné, J.; Bader, H. Water Res. 1988, 22, 313–319.
36. Bailey, P. S.; Mitchard, D. A.; Khashab, A. I. Y. J. Org. Chem. 1968, 33,
2675–2680.
37. Le Lacheur, R. M.; Gaze, W. H. Environ. Sci. Technol. 1996, 30, 1072–1080.
38. Duguet, J. P.; Jaulin, L.; Aurelle, Y.; Roques, H.; LeÂgeron, J. P. Ozone Sci.
Eng. 1980, 2, 105–122.
39. Moenig, J.; Chapman, R.; Asmus, K. D. J. Phys. Chem. 1985, 89,
3139–3144.
40. Andrzejewski, P.; Kasprzyk-Hordern, B.; Nawrocki, J. Water Res. 2008, 42,
863–870.
41. Yang, L.; Chen, Z. L.; Shen, J. M.; Xu, Z. Z.; Liang, H.; Tian, J. Y.; Ben, Y.;
Zhai, X.; Shi, W. X.; Li, G. B. Environ. Sci. Technol. 2009, 43, 5481–5487.
42. Asami, M.; Oya, M.; Kosaka, K. Sci. Total Environ. 2009, 407, 3540–3545.
43. Planas, C.; Palacios, O.; Ventura, F.; Rivera, J.; Caixach, J. Talanta 2008, 76,
906–913.
45. Selbes, M.; Kim, D.; Karanfil, T. Water Res. 2014, 66C, 169–179.
46. Bailey, P. S. Ozonation of nucleophiles. In Ozonation in Organic Chemistry;
Wasserman H. H., Eds.; Academic Press Inc.: New York, 1982.
233
47. Glezer, V.; Harris, B.; Tal, N.; Iosefzon, B.; Lev, O. Water Res. 1999, 33,
1938–1948.
48. Dickenson, E. R. V.; Summers, R. S.; Croué, J.-P.; Gallard, H. Environ. Sci.
Technol. 2008, 42, 3226–3233.
49. Hua, G.; Reckhow, D. A. J. Am. Water Works Assoc. 2008, 100, 82–89.
50. Joo, S. H.; Mitch, W. A. Environ. Sci. Technol. 2007, 41, 1288–1296.
234
Chapter 13
Formation of Disinfection By-Products

from Bacterial Disinfection
T. W. Ng,1 B. Li,2 A. T. Chow,2 and P. K. Wong*,1
1The School of Life Science, The Chinese University of Hong Kong,

Hong Kong, China
2The Belle W. Baruch Institute of Coastal Ecology and Forest Science,
Clemson University, Clemson, South Carolina 29631, U.S.A.

*E-mail: [email protected]. Tel.: +852 3943 6383. Fax: +852 2603 5767.
Controlling the disinfection by-product (DBP) formation, while

balancing the risk of bacterial contamination in finished water,
requires an in depth understanding in both the chemistry of
DBPs and the kinetic and mechanism of bacterial disinfection.
Although numerous studies have examined different strategies
on the reduction of DBP formation in finished waters, their
main focuses are on natural organic matter (NOM) in source
waters. Few studies evaluated the contributions of bacterial
organic matter (BOM) on DBP formation. Breaking down of
bacterial cells, releasing BOM into water, is an essential process
in water disinfection. Prevalence of bacterial contamination in
source waters and occurrences of biofilms in water treatment
units and distribution systems are frequently reported. The
contributions of organic matter from bacterial and microbial
sources in drinking water treatment cannot be ignored. In this
study, we compared the contributions of bacterial cells and
humic acid (HA), under controlled laboratory environments,
on the DBP formation from chlorination. We also evaluated
the effects of HA, pH, and chlorine dosage on both disinfection
efficiency and DBP formation. The results showed that bacterial
cells are also potential DBP precursor. Presence of HA not
only increased the DBP formation, but also decreased the
disinfection efficiency of chlorination. pH and chlorine dosage
also affect the disinfection efficiency and DBP formation of
chlorination. High dosage of chlorine would increase both the

disinfection efficient and DBP formation, while low pH increase
the disinfection efficient with a lower DBP formation. The
results suggest that high disinfection efficiency with low DBP
formation of chlorination can be achieved by optimization of
operation conditions. Based on our knowledge, this pioneering
work is the first study to investigate the relationship between
bacteria and DBP contaminations in water treatment processes
at the same time.
1. Introduction
Adding disinfectants in drinking water treatment is a common practice
to protect us from infectious diseases. In spite of its advantage to inactivate

waterborne pathogens, disinfectants, such as chlorine and chloramines react with
constituents in waters to form a variety of disinfection by-products (DBPs), such
as trihalomethanes (THMs), haloacetic acids (HAAs), chloral hydrate (CHD),
haloacetonitriles (HANs), and N-nitrosodimethylamine (NDMA) (1, 2). Since
THMs was first identified in finished drinking water in 1976 (3), more than 600
DBPs have been identified but which only account for 30% of total halogenated
DBPs in chlorinated waters (4). Importantly, most of identified DBPs are
cytotoxic and genotoxic (5–7) and are rapidly absorbed into our bodies through
showering, hand washing, drinking, and even breathing (8–11). Consuming water
bearing DBPs could cause bladder cancer, rectal cancer, adverse birth outcomes,
and other adverse health related issues (12–14). Noticeably, both disinfection
efficacy and DBP formation are increased with the dosage of disinfectants
(15)(Figure 1). Therefore, it is a great challenge to balance the risk between
microbial and DBP contaminations for safe drinking water.
Figure 1. Disinfection / Disinfection by-product risk trade-offs. (modified from

Lykins et al. (15))
236
Conventionally, natural organic matter (NOM) such as humic substance
is considered as the major precursor of DBPs (16, 17). Bacterial organic
matter (BOM) is rarely studied although the major function of disinfection is
to inactivate bacterial pathogens. Bacterium can be an important source of
organic matter and DBP precursors in potable water treatment systems because
prevalence of bacterial contamination in source waters (18, 19) and biofilms in
water treatment unit and distribution systems (20, 21) are frequently reported
worldwide. Extracellular polymeric substance (EPS) and many biochemical
compounds released from bacteria and microorganisms such as polysaccharides,
protein, and nucleic acid are commonly found in water treatment systems (22).
Many of these biomolecules have been considered reactive precursors in forming
THMs and other DBPs during chlorination (23–25). Notably, the formation of
DBPs is unavoidable in bacterial disinfection because bacterium is composed
of a variety of organic compounds which can release into water after cell lysis.
This BOM can react with excess disinfectants in water to form DBPs. Moreover,
BOM is relatively enriched in nitrogen than other sources of NOM in natural
waters (26, 27). Therefore, BOM could be an important precursor in forming
nitrogenous DBPs (N-DBPs) such as HANs and NDMA, which could be more
hazardous compared to currently regulated DBPs (i.e. THMs and HAAs) in
finished drinking waters (6, 28, 29).
In this study, we hypothesize that BOM is a DBP precursor and it can
react with chlorine to form a variety of DBP during bacterial disinfection. We
conducted a series of controlled laboratory experiments to evaluate the effect
of water pH, chlorine dosage, and concentrations of NOM on the disinfection
efficiency and DBP formation. The objective of this study is to illustrate the
relationship between bacteria and DBP contaminations in chlorination processes,
as demonstrated in Figure 1. Results of the study could be useful for water
engineers and microbiologists to define the end point of bacterial disinfection and
provide referable information to balance the risk between microbial control and
production of DBPs in finished waters.
2. Experimental
2.1. Chemicals
All the stock solutions and supplies including pipette tips and filter papers for
disinfection experiment were autoclaved (at 121°C for 20 min) to avoid culture
contamination. All glassware were soaked in 5% DECON for 24 h, washed with
Milli-Q water (Millipore, USA), and placed in a furnace at 500°C for 5 h to
eliminate organic contaminants. All the working solutions were conducted in the
phosphate buffered saline (PBS) prepared by suspending 8.01 g of NaCl, 0.20 g
of KCl , 1.78 g of Na2HPO4•2H2O, 0.27 g of KH2PO4 into 1 L Milli-Q water.
Suwannee River Humic Acid (HA) was purchased from International Humic
Substances Society. A stock solution of 100 mg/L of HA was prepared in 0.01
M KCl. The stock solution was filtered through 0.45 μm membrane filter and
kept in clean and glass-topped amber bottle in a refrigerator set at 4°C. The stock
solution was diluted with Milli-Q water to 2 or 4 mg/L of HA for the disinfection
237
studies. The 2 or 4 mg/L of HA solution was equivalent to 0.92 or 1.84 mg/L of
dissolved organic carbon (DOC), respectively, based on our measurements using
Shimadzu TOC analyzer.
2.2. Bacterial Cultures
Seven bacteria commonly found in soil and water, including three gram
positive strains (Bacillus cereus, Bacillus subtilis, and Staphylococcus sciuri)
and four gram negative strains (Acinetobacter junii, Aeromonas hydrophila,
Escherichia coli, and Shigella sonnei), were examined in this study. After a
pre-incubation test to determine the time span of the stationary phase for each
bacterial species, seven strains of bacteria were inoculated individually into
nutrient broth (10% of the instructed dosage) and grown overnight (16 - 20 h)
to reach the stationary phase. Aliquots of bacterial cultures were centrifuged
at 11,600 x g for 5 min, and bacterial cells were collected and washed three
times with saline solution (0.9% of NaCl) to remove the nutrient broth. The
cell pellets were re-suspended in deionized water in Duran bottles to obtain a
cell density in the range of 104–107 colony forming units per milliliter (cfu/mL)
for the chlorination test. Total organic carbon of cell pellets were tested in
SSM-5000A solid sample combustion unit connected to Shimadzu TOC analyzer.
The relationships between total organic carbon and bacterial count of the seven
bacterial strains are summarized in Table 1. To study the effects of pH, chlorine
dosage, and NOM levels, E. coli was inoculated in a low level and a high level
for each parameter as summarized in Table 2.
Table 1. Regressions between Bacterial Total Organic Carbon Bacterial

Cell Count
Bacterial species Relationship n R2
Acinetobacter junii y = 4.9351x – 37.377 15 0.9970
Aeromonas Hydrophila y = 5.0291x – 45.483 15 0.9713
Escherichia coli y = 5.7935x – 52.225 15 0.9954
Shigella sonnei y = 3.5049x – 31.618 15 0.9493
Bacillus cereus y = 2.3581x – 8.448 15 0.9775
Bacillus substilis y = 5.2644x – 50.018 15 0.9856
Staphylococcus sciuri y = 2.2655x – 18.632 15 0.9886
y = TOC (mg/L), x = bacterial cell density (cfu/mL).
238
Table 2. A Low Level and a High Level of pH, Chlorine Dosage, and NOM
Were Tested in This Study
Low Level High Level
pH 5 8
Chlorine Dosage (mg/L) 0.5 1.5
Humic Acid (mg/L) 2 4
2.3. Chlorine Inactivation

Free chlorine (HOCl) stock solution (3,000 mg/L of Cl2) was prepared by
diluting ≥4% sodium hypochlorite (NaOCl) (Aldrich) and buffered to pH to 8.0 ±
0.2. For the effects of pH and chlorine dosage on DBP formation and disinfection
efficiency, working solutions of 0.5 and 1.5 mg/L of HOCl were prepared by
diluting stocking solutions with PBS solution. Their pH was adjusted to 5.0 or
8.0 ± 0.2 using 2 mM of sodium hydroxide or 1 mM of sulfuric acid. The HOCl
concentration was tested daily with a HACH chlorine pocket colorimeter (Hach
Co., Loveland, CO) in accordance with Standard Method 4500-Cl B.
Chlorination was conducted in 60-mL glass bottles. The bottles were all
covered in order to prevent chlorine degradation from light. Bottles were placed
in a shaker with a constant vibration frequency to prevent cell settling. For the
effects of pH, chlorine dosage, and NOM levels, the final concentration of E. coli
was adjusted to 1.5×107 cfu/mL. Bacteria and/or HA spiked into the solution and
mixed continuously at room temperature. An aliquot of 0.5 mL of treated mixture
was sampled at 5, 20, 40, and 80 min during the reaction. The mixture was then
quickly diluted with sterilized saline solution. Each 0.1 mL of original or diluted
sample was then immediately placed on the sterilized and dry nutrient agar. Agar
plates were incubated at 37°C for 24 h. The bacteria cell density at each time
interval was computed by the colonies counted on the agar plates.
2.4. Fluorescence Spectroscopy
A fluorescence spectrometer (F-7000, Hitachi, Japan) was used to obtain the

three-dimensional emission-excitation matrix (EEM) spectra of the samples with
excitation wavelengths in the range of 200-500 nm in 5-nm intervals and with
emission wavelengths in the range of 290-500 nm in 5-nm intervals. The emission
and excitation slits were fixed at 5 nm and the scanning speed was 12,000 nm/min.
The second-order Raleigh scattering in the EEM measurements was eliminated
using a 290 nm emission cutoff filter. The instrument was corrected for excitation
and emission according to the manufacturer’s suggested protocols. Variations in
fluorescence intensity were calibrated using both pure water and quinine sulfate
solution after the measurements of every 5 samples. The fluorescence intensity
values were standardized in quinine sulfate units (QSU) (30).
239
2.5. DBP Analysis
The DBPs analyzed including four THMs (chloroform (TCM),

bromodichloromethane (BDCM), dibromochloromethane (DBCM)
and bromoform (TBM)), four HANs (trichloroacetonitrile (TCAN),
dichloroacetonitrile (DCAN), bromochloroacetonitrile (BCAN) and
dibromoacetonitrile (DBAN)), dichloropropanone (DCP) and chloral hydrate
(CHD). All DBP standards were purchased from Sigma Aldrich and prepared
before experiments for calibration. All water samples were quenched with
sodium thiosulfate (Na2S2O3) and then treated with methyl-tertiary-butyl-ether
(MTBE) (Fluka, Switzerland) and sodium sulfate (Na2SO4) for liquid/liquid
extraction of DBPs from aqueous solution according to U.S. EPA Method 551.1
(31). Samples were kept in refrigerator at 4 °C and analyzed within 10 days.
Gas chromatography and electron capture detection (GC-ECD) was used to

quantify the DBP concentrations. GC column (A - 0.25 mm ID x 30 m fused
silica capillary) was used and the temperature program was slightly modified as
follows: held at 30°C for 7 min then increased to 100°C at 20°C/min and held for
3min, and then increased to 260°C at 30°C/min and held for 2 min. The injector
and detector temperature were set at 250°C.
3. Results
3.1. DBP Formation During Bacteria Inactivation
The results of DBP formation from disinfecting seven strains of bacteria

were summarized in Figure 2. Similar to other NOM, THM was the dominant
species. The specific THM formation ranged from 6.1 to 37.6 μg-THM/mg-C
with an average of 22.4 μg-THM/mg-C. The specific HAN formation ranged 3.1
to 16.3 μg-HAN/mg-C with an average of 6.1 μg-HAN/mg-C. The specific CHD
formation ranged 0.6 to 1.8 μg-CHD/mg-C with an average of 1.1 μg-CHD/mg-C.
Results confirmed that DBPs could be formed directly from bacteria during water
chlorination even no other NOM existed in water. Fluorescence EEM of water
containing E. coli before and after the chlorination was shown in Figure 3. Before
chlorination, only two peaks at Ex = 280 nm / Em = 340 nm and Ex = 220 nm /
Em = 340 nm, occurring in microbial-by-product and aromatic protein regions,
respectively, were observed (Figure 3a). After a low dosage of chlorine treatment
(i.e. 0.5 mg/L), the intensities of the two peaks were decreased and signals in
fulvic acid-like region at Ex = 230 nm / Em = 420 nm were observed (Figure
3b). After a high dosage of chlorine treatment (i.e. 1.5 mg/L), a peak at fulvic
acid-like region was produced (Figure 3c). Results demonstrated a transition of
BOM during water chlorination.
240
Figure 2. The specific DBP formation of halogenation of different bacteria.

The sign in the parenthesis indicates the bacteria belong to Gram positive (+)
or Gram negative (-).
3.2. Effects of NOM on Bacterial Inactivation and DBP Formation

To demonstrate the inhibition effect of NOM on disinfection processes,
chlorination of E. coli solution with two concentrations of HA and two chlorine
dosages were examined. The results were summarized in Figure 4. In all cases,
bacterial inactivation was the most efficient when no HA existed in water.
Existence of HA in water consumed chlorine, lowering the disinfection efficiency.
Consider low chlorine dosage at 0.5 mg/L of Cl2 at pH 8 as an example. The
cell density slightly dropped from 7.5 to 6.5 log cfu/ml in first 5 min with the
presence of 4 mg/L of HA in water and only achieved 2-log reduction after 80
minutes. Results were significantly lower than the 0 and 2 mg/L of HA treatments,
which reached almost 4-log reduction after 80 min. Bacterial inactivation by
241
chlorination generally occurred in a relatively short time after chlorine was added.
Significant decreases in cell density generally occurred in the first 5 min of the
disinfection in water solutions with or without HA. No obvious change in cell
density was observed in later reaction time.
Figure 3. (a) Fluorescence EEM of water solution containing E. coli: (a) before
disinfection treatment, (b) after 80-min treatment with 0.5 mg/L Cl2 at pH 8,
and (c) after 80-min treatment with 1.5 mg/L Cl2 at pH 8. Quenching solution
Na2S2O3 was added after 80-min treatment before fluorescence measurement
was taken.
Both HA and BOM can react with chlorine in forming DBPs. The formations
of DBPs were generally increased with the chlorine dosage (Figures 5). For pure
E.coli culture solution with 1.5×107 cfu/mL at pH 8, 1.1 and 5.7 μg/L of THMs
were produced at 0.5 and 1.5 mg/L of chlorine, respectively. For the pure NOM
solutions with 2 and 4 mg/L of HA at pH 8, 5.0 – 5.2 µg/L and 33.4 – 40.1 µg/
L of THMs were produced at 0.5 and 1.5 mg/L of chlorine, respectively. Other
types of DBPs were also produced but their levels were in ng/L range. The water
solutions containing both E. coli and HA generated less THMs than pure E.coli or
pure HA water solutions at pH 8. Fluorescence EEM demonstrated that different
242
characteristics of BOM and HA. A strong humic acid-like peak was observed in
Ex = 250 nm / Em 450 nm in pure HA solution (Figure 6a) but a strong microbial
by-product-like peak was observed in Ex = 280 nm / Em = 250 nm when bacteria
was added into the solution (Figure 6b).
Figure 4. Effects of humic acid and pH on E. coli cell density in water solution
during 80-min chlorination.
3.3. Effects of pH and Dosage on Bacterial Inactivation and DBP Formation
In general, acidic condition could enhance the effectiveness of bacterial

inactivation. As shown in Figure 4, the cell viability of E. coli after 80 minutes
was significantly lower at pH of 5 (1.5 to 3 log cfu/mL) than those at pH of 8
(2.0 to 5.5 log cfu/mL) in the same chlorine dosage. It is well known that THM
formation is favored at alkaline pH (32). THM formations from pure bacterial cell
culture were higher in pH 8 (1.7 to 5.7 µg-THM/L) than those in pH 5 (0.8-1.1
µg-THM/L). Similarly, THM formation of pure HA solution in pH 8 (5.0 to 40.1
µg/L) was higher than those in pH 5 (2.0 to 16.2 µg/L). Moreover, the yields of
DBP were generally increased with chlorine dosage (32, 33). THM formation in
0.5 mg/L of Cl2 was in the range of 0.8 to 5.3 µg/L and THM formation in 1.5
mg/L of Cl2 ranged from 1.1 to 40.1 µg/L.
243
Figure 5. Effects of humic acid and pH on DBP formation in water solution

after 80-min chlorination.
Figure 6. Fluorescence EEM of water solution after chlorination with initial 1.5
mg/L of Cl2 at pH=8 for 80 min (a) 4 mg/L Humic acid only (b) 4 mg/L Humic
acid + 107 cfu/mL E.coli K-12. Quenching solution Na2S2O3 was added after
80-min treatment before fluorescence measurement was taken.
244
4. Discussions
4.1. Microbial Cells and Cellular Components as DBPs Precursor
Many studies had examined biological molecules such as amino acids,
nucleic acids, carbohydrates on DBP formation and concluded that they are
reactive DBP precursors (23–25, 34, 35). Generally specific THM formation of
these biomolecules usually low than 10 μg-THM/mg-C except few exceptions
such as tryptophan and tyrosine which can be over 100 μg-THM/mg-C (34, 35).
Differences from other previous studies, we examined DBP formation directly
disinfecting bacterial cells. Our results showed the specific THM formation from
bacterial cells ranged from 6.1 to 37.6 μg/mg-C and specific HAN formation from
3.6 to 16.3 μg/mg-C. These numbers are generally comparable to those monomers
or standard chemicals studies as discussed above. The results confirmed that that
these bacterial components can also be the DBP precursor as long as halogenating
reagent exists. We have also demonstrated that there is a positive correlation
between DBP formation and the log-reduction of E. coli by both chlorination and
chloramination (33). The results further confirmed that the DOM released from
the breakdown of bacteria can contribute to DBP formation.
In addition to the source waters, bacteria existed as biofilm are commonly
observed within water delivery systems (36). Disinfectants could react with
extracellular polymeric substance (EPS) released from biofilm in forming DBPs
(37). Not only do the quantity and quality of EPS affect the yield of DBPs, but
also they determine the speciation of DBP formation. For example, the surrogate
protein of EPS favors the formation of HAA formation while the surrogate lipid
favors THM formation. In addition, the bacterial phenotypes and the materials
on which biofilm attached also have a great impact on the bacterial derived DBP
formation. Our research group demonstrated that planktonic cells formed 7-11
times greater THMs per carbon that those from biofilm Furthermore, biofilms
on the polyvinyl chloride consumed comparable chlorine but produced more
THMs than those on galvanized zinc surface (33). All these results indicated the
important of BOM in the DBP formation in the drinking water supply.
There is no obvious difference on the bacteria type, such as Gram staining,
size or source of the bacteria, on the DBP formation of chlorination. However,
as mentioned, the quality of cell components (i.e. EPS) would have great impact
on the DBP formation. Therefore, the DBP formation potential of bacteria maybe
mainly affect by the quality of cell components instead of the thickness of cell
wall. Therefore, no obvious trend is observed in this study.
4.2. Natural Organic Matter versus Bacterial Organic Matter

Our study demonstrated that the specific THM formation of bacterial cell
ranged from 6.1 to 37.6 μg-THM/mg-C, which is relatively less reactive than the
conventional DBP precursors such as humic and fulvic acids, which generally
ranged from 40 to 140 to μg-THM/mg-C (38). The bacterial concentration
in source water is generally around 104-105 cells/mL (39). Consider the
specific THM Formation = 22.4 μg-THM/mg-C, specific HAN formation = 6.1
μg-HAN/mg-C, and specific CHD formation = 1.1 μg-CHD/mg-C as shown in
245
the section 3.1, an estimated DBP formation contributed from bacterial cells in
drinking water chlorination could range from 10-100 ng-THM/L, 3-30 ng-HAN/L,
and 0.5-5 ng-CHD/L. According to the results of humic acid effect (Figure 5),
if source water contains 2-4 mg/L of DOM such as HA, DBP formation should
be in the range of 2-45 µg-THM/L, 0.1-2.2 μg-HAN/L, and 0-2.6 μg-CHD/L.
This calculation demonstrated that the DBP formation from bacterial cells is
considerably lower than that from the NOM. However, the DBP formation
from BOM could be important in some special occasions such as wastewater
treatments or water disinfection in medical facilities. In wastewater effluent, the
bacterial cells concentration can be even up to 107-108 cells/mL (39) and the
THM formation can be as high as 10-100 μg-THM/L, 3-30 μg-HAN/L, and 0.5-5
μg-CHD/L. Any level of chemical contaminants in hospital water supply could
increase risk to the patients who are sensitive to any chemical agents. Notably,
biological molecules such as amino acid and nucleic acid are N-rich. With a lower
C/N ratio, BOM could have a higher reactivity in forming more toxic N-DBPs
when comparing with other sources of NOM (34).
The addition of HA would inhibit the disinfection efficiency as HA
competed for chlorine during the process, resulting in a lower effective chlorine
concentration for bacterial inactivation in water solutions. Chlorine could react
with either E. coli or HA to form DBPs. However, E. coli and HA do not exhibit
an additive effect in DBP formation when they co-existed in the water solutions,
although both of them are DBP precursors. In some cases, the existence of E. coli
in HA solution even resulted a lower DBP formation than the HA solution alone
(Figure 5). A simple competitive reaction model (i.e. E. coli and HA compete
for free chlorine) may not explain the phenomenon when there was excess
chlorine in water. This observation could also attribute to N-rich organic matter
from bacterial cells. N-rich BOM could affect the residual chlorine in water
throughout forming organic chlorine and breakpoint reaction, depending on the
Cl-to-N ratio (40). In fact, studies had demonstrated that organic N could reduce
THM formation during water chlorination (41). These experiments demonstrated
“competition” relationships of HA and bacteria during water chlorination. HA
reduces bacterial inactivation efficiency by consuming reactive chlorine whereas
BOM from lysed cells changes the chlorine chemistry in water and reduces THM
formation from HA. The combined effects showed less disinfection efficiency
and less DBP formation than HA or E. coli existed alone in water. In natural
source waters, other organic matters such as fulvic acid also exist but HA is
generally larger in size than fulvic acid and it should have a greater shielding
effect protecting bacteria from disinfection.
4.3. Alternative Disinfectants on Bacterial Inactivation and DBP Formation

As chlorination produces toxic and carcinogenic DBPs, scientists begin
to find ways to decrease the DBP formation with maintaining high efficiency
in disinfection processes. Use of other alternative disinfectants is one of the
promising approaches to achieve the goal. For example, the disinfection by
TiO2/UV, a type of advance oxidation processes (AOPs), shows much lower DBP
formation in term of type and quantity (42). However, the bacterial inactivation of
246
these AOPs is relatively less efficient than the conventional disinfection methods
such as chlorination and chloramination (43). Our recent study comparatively
investigated the DBP formation and disinfection efficiency of four disinfectants
(chlorination, chloramination, photo-Fenton reaction and TiO2/UV photocatalytic
disinfection) under different pH (5 and 8) and different dosages (43). Among
different disinfectants, chlorination had the fastest bacterial inactivation
following with chloramination, whereas photo-Fenton, and TiO2/UV treatments
required a relatively longer contact times to achieve the same inactivation
efficacy. Nevertheless, chlorination had the highest DBP formation, followed by
chloramination, and the DBP formation from photo-Fenton reaction and TiO2/UV
photocatalytic disinfection were all below the detection limits (1 ng/L). Although
DBP formation can be greatly reduced using AOPs (photo-Fenton reaction or
TiO2/UV photocatalytic disinfection), longer reaction time could significantly
increase the operation costs. More important, the recovery and the impacts of
nanoparticles are still uncertain.
5. Conclusions
The results of this study confirmed our hypothesis that BOM is a DBP
precursor and it can react with chlorine to form a variety of DBP during
bacterial disinfection. Since bacteria in source water are ubiquitous, their relative
contribution to DBP formation in water disinfection has been overlooked. The
DBP formation from inactivating bacterial cells is greatly affected by different
factors, such as types and dosages of disinfectants, water pH, physiology and
phenotypes of bacteria, etc. Results of this study also suggested that a lower
DBP formation and better inactivation efficiency could be achieved in the acidic
condition.
Acknowledgments
This work was funded by the Research Grant Council of Hong Kong SAR
Government (CUHK 477610). P.K. Wong was also supported by the CAS/SAFEA
International Partnership Program for Creative Research Teams, Chinese
Academy of Sciences, China. This research is also based upon work supported
by NIFA/USDA under the project number SC-1700489 and SCN-2013-02784,
and Joint Fire Science Program 14-1-06-19. This manuscript is a technical
contribution no. 6349 of the Clemson University Experimental Station.
References
1. Krasner, S. W. Chemistry of disinfection by-product formation. In Formation
and Control of Disinfection By-Products in Drinking Water; Singer, P. C., ed.;
American Water Works Association: Denver, 1999.; pp. 27-52.
2. Xie, Y. Disinfection Byproducts in Drinking Water: Formation, Analysis, and
Control; Lewis Publishers: Boca Raton, 2004.
247
3. Rook, J. J. Haloforms in drinking-water. J. Am. Water Works Assoc. 1976,
68, 168–172.
7175–7185.
Chemical characterization and mammalian cell cytotoxicity and genotoxicity.
7. Cemeli, E.; Wanger, E. D.; Anderson, D.; Richardson, S. D.; Plewa, M.

J. Modulation of the cytotoxicity and genotoxicity of the drinking water
disinfection byproduct iodoacetic acid by suppressors of oxidative stress.
8. Erdinger, L.; Kuhn, K. P.; Kirsch, F.; Feldhues, R.; Frobel, T.; Nohynek, B.;
Gabrio, T. Pathways of trihalomethane uptake in swimming pools. Int. J.
Hyg. Environ. Health 2004, 207, 571–575.
9. Miles, A. M.; Singer, P. C.; Ashley, D. L.; Lynberg, M. C.; Mendola, P.;
Langlois, P. H.; Nuckols, J. R. Comparison of trihalomethanes in tap water
and blood. Environ. Sci. Technol. 2002, 36, 1692–1698.
10. Xu, X.; Weisel, C. P. Dermal uptake of chloroform and haloketones during
bathing. J. Expo. Anal. Environ. Epidemiol. 2005, 15, 289–296.
11. Zwiener, C.; Richardson, S. D.; De Marini, D. M.; Grummt, T.; Glauner, T.;
pregnancy outcomes: A review. Environ. Health Persp. 2002, 110, 61–74.
13. Bull, R. J.; Krasner, S. W.; Daniel, P. A.; Bull, R. D. Health Effects and
Occurrence of Disinfection By-Products; AWWA Research Foundation and
American Water Works Association: Denver, CO, 2001.
14. Hinckley, A. F.; Bachand, A. M.; Reif, J. S. Late pregnancy exposures to
disinfection by-products and growth-related birth outcomes. Environ. Health
Perspect. 2005, 113, 1808–1813.
15. Lykins, B. W., Jr.; Goodrich, J. A.; Clark, R. M.; Harrison, J. Point-of-use
treatment. Point-of-use/point-of-entry treatment of drinking water. Water
Supply 1994, 12 (1/2), SS4/1–SS4/5.
16. Leenheer, J. A.; Croue, J. P. Characterizing aquatic dissolved organic matter.
Environ. Sci. Technol. 2003, 37, 18a–26a.
17. Zsolnay, A. 2003. Dissolved organic matter: artefacts, definitions, and
functions. Geoderma 2003, 113, 187–209.
18. Davis, K.; Anderson, M. A.; Yates, M. V. Distribution of indicator bacteria
in Canyon Lake, California. Water Res. 2005, 39, 1277–1288.
19. Ruecker, N. J.; Braithwaite, S. L.; Topp, E.; Edge, T.; Lapen, D. R.;
Wilkes, G.; Robertson, W.; Medeiros, D.; Sensen, C. W.; Neumann, N.
248
F. 2007. Tracking host sources of Cryptosporidium spp. in raw water for
improved health risk assessment. Appl. Environ. Microbiol. 2007, 73,
3945–3957.
20. Lehtola, M. J.; Torvinen, E.; Kusnetsov, J.; Pitkanen, T.; Maunula, L.; von
Bonsdorff, C. H.; Martikainen, P. J.; Wilks, S. A.; Keevil, C. W.; Miettinen, I.
T. Survival of Mycobacterium avium, Legionella pneumophila, Escherichia
coli, and caliciviruses in drinking water-associated biofilms grown under
high-shear turbulent flow. Appl. Environ. Microbiol. 2007, 73, 2854–2859.
21. September, S. M.; Els, F. A.; Venter, S. N.; Brozel, V. S. Prevalence of
bacterial pathogens in biofilms of drinking water distribution systems. J.
Water Health 2007, 5, 219–227.
22. Doumeche, B; Galas, L.; Vaudry, H.; Di Martino, P. Membrane foulants
characterization in a drinking water production unit. Food Bioprod. Process.
2007, 85 (C1), 42–48.

23. Li, J.; Blatchley, E. R., III Volatile disinfection byproduct formation
resulting from chlorination of organic-nitrogen precursors in swimming
24. Hong, H. C.; Mazumder, A.; Wong, M. H.; Liang, Y. Yield of trihalomethanes
and haloacetic acids upon chlorinating algal cells, and its prediction via algal
cellular biochemical composition. Water Res. 2008, 42, 4941–4948.
25. Navalon, S.; Alvaro, S.; Garcia, H. Carbohydrates as trihalomethanes
precursors. Influence of pH and the presence of Cl and Br on trihalomethane
formation potential. Water Res. 2008, 42, 3990–4000.
26. Stevenson, F. J.; Cole, M. A. Cycles of Soil: Carbon, Nitrogen, Phosphorus,
Sulfur, Micronutrients, 2nd ed.; John Wiley & Sons, Inc.: New York, 1999.
27. Tate, R. Soil Microbiology; John Wiley: New York, 2000.
28. Moudgal, C. J.; Lipscomb, J. C.; Bruce, R. M. Potential health effects of
drinking water disinfection byproducts using a quantitative structure toxicity
relationship. Toxicology 2000, 147, 109–131.
29. U.S. EPA. N-Nitrosodimethylamine CASRN 62-75-9, Integrated Risk
Information Service (IRIS) Substance File; U.S. EPA: Washington, DC,
1997.
30. Yamashita, Y.; Tanoue, E. Chemical characterization of protein-like
fluorophores in DOM in relation to aromatic amino acids. Marine Chem.
2003, 82, 255–271.
31. U.S. EPA. Method 551.1. Determination of Chlorination Disinfection By-
Products, Chlorinated Solvents, and Halogenated Pesticides/Herbicides in
Drinking Water by Liquid–Liquid Extraction and Gas Chromatograph with
Electron-Capture Detection (Revision 1.0); Munch, D. J., Hautman, D. P.,
Eds.; U.S. EPA: Cincinnati, 1995.
32. Xie, Y. Disinfection Byproducts in Drinking Water: Formation, Analysis, and
Control; Lewis Publishers: Baco Raton, 2003.
33. Wang, J. J.; Liu, X.; Ng, T. W.; Xiao, J. W.; Chow, A. T.; Wong, P. K.
Disinfection byproduct formation from chlorination of pure bacterial cells
and pipeline biofilms. Water Res. 2013, 47, 2701–2709.
249
34. Wang, J. J.; Ng, T. W.; Zhang, Q.; Yang, X. B.; Dahlgren, R. A.; Chow, A.
T.; Wong, P. K. Reactivity of C1 and C2 organohalogens formation – from
plant litter to bacteria. Biogeosciences 2012, 9, 3721–3727.
35. Hong, H. C.; Wong, M. H.; Liang, Y. Amino acids as precursors of
trihalomethane and haloacetic acid formation during chlorination. Arch.
Environ. Contam. Toxicol. 2009, 56, 638–645.
36. Williams, M. M.; Domingo, J. W. S.; Meckes, M. C.; Kelty, C. A.; Rochon, H.
S. Phylogenetic diversity of drinking water bacteria in a distribution system
simulator. J. Appl. Microbiol. 2004, 96, 954–964.
37. Wang, Z.; Kim, J.; Seo, Y. Influence of bacterial extracellular polymeric
substances on the formation of carbonaceous and nitrogenous disinfection
byproducts. Environ. Sci. Technol. 2012, 46, 1136–11369.
38. Chow, A. T.; Gao, S.; Dahlgren, R. A. Physical and chemical fractionation of
dissolved organic matter and trihalomethane precursors: A review. J. Water

Supply: Res. Technol. – AQUA 2005, 54, 475–507.
39. Van Nevel, S.; De Roy, K.; Boon, N. Bacterial invasion potential in water
is determined by nutrient availability and the indigenous community. FEMS
Microbiol. Ecol. 2013, 85, 593–603.
40. Shang, C; Gong, W. L.; Blatchley, E. R. Breakpoint chemistry and volatile
byproduct formation resulting from chlorination of model organic-N
compounds. Environ. Sci. Technol. 2000, 34, 1721–1728.
41. Yang, X.; Shang, C. Chlorination byproduct formation in the presence of
humic acid, model nitrogen organic compounds, ammonia, and bromide.
42. Richardson, S. D.; Thruston, A. D., Jr; Collette, T. W. Identification of TiO2/
UV disinfection byproducts in drinking water. Environ. Sci. Technol. 1996,
30, 3327–3334.
43. Li, B. Comparisons of Bacterial Inactivation Efficiency and Disinfection
Byproduct Formation from Four Different Disinfection Processes. Ph.D.
Thesis, Clemson University, South Carolina, August 2014.
250
Chapter 14
Catalysis of DBP-Precursor Bromination by

Halides and Hypochlorous Acid
John D. Sivey,* Mark A. Bickley, and Daniel A. Victor
Department of Chemistry and

Urban Environmental Biogeochemistry Laboratory,
Towson University, 8000 York Road, Towson, Maryland 21252, United States
When bromide-containing waters are disinfected with free

chlorine, bromide can be converted into free bromine
(principally HOBr). Although often overlooked, chloride and
bromide are capable of catalyzing bromination reactions of
organic compounds by promoting the formation of BrCl and
Br2, respectively. Similarly, HOCl can catalyze bromination
reactions by enhancing concentrations of BrOCl. As
brominating agents, BrCl, Br2, BrOCl (and the related species
Br2O) have been shown to be several orders of magnitude more
inherently reactive than HOBr toward several non-ionizable
aromatic compounds. Under conditions representative of
chlorinated drinking water, BrCl and BrOCl can contribute
more than HOBr to overall bromination rates of modestly
nucleophilic aromatic compounds. This chapter reviews the
literature regarding these reactive brominating agents and
discusses their potential influence on bromination rates of
disinfection by-product precursors.
Introduction
In 1902, for the first time in the United States, chlorine was added to drinking
water (DW) to protect the public from water-borne diseases (1). Since then,
countless lives have been saved by this potent microbicidal agent. Not until
the 1970s, however, were the unintended consequences of DW chlorination
discovered (2, 3). In addition to inactivating bacteria and viruses, chlorine can

also react with natural organic matter (4–7) to form potentially toxic disinfection
by-products (DBPs) (8).
In water with low nitrogen levels, chlorine consists primarily of hypochlorous
acid (HOCl, pKa = 7.58 at 20 °C) (9) and ClO-; the sum of such species is termed
free chlorine. In addition to reactions with organic DBP precursors, free chlorine
can rapidly oxidize bromide (eq 1) (10).
HOCl-mediated oxidation of bromide (eq 1) has a large equilibrium constant

(1.5 x 105, 25 °C, µ = 0) (11) and will proceed “to completion” in the presence
of excess HOCl. The forward-reaction rate associated with eq 1 is proportional to
[HOCl] (k1+ = forward-reaction rate constant = 1.55 x 103 M-1 s-1, 25 °C, µ = 1 M)
(10) and is therefore anticipated to decrease with increasing pH.

As with HOCl, the presence of HOBr in DW is concurrently beneficial and
problematic. Like its chlorinated analogue, HOBr is both a potent microbicide
(1, 12–14) and a reactive electrophile capable of forming organobromine
compounds in DW (1, 5, 15–22), wastewater (23, 24), and recreational waters
(e.g., pools and spas) (25–27). Some of these organobromine compounds are
DBPs whose concentrations are regulated in drinking water (28–31) due to their
potential cytotoxic (32, 33) and genotoxic effects (34–36). Brominated DBPs are
generally more toxic than their chlorinated analogs (34). Therefore, strategies
for minimizing their formation are highly desirable. Such DBP-minimization
strategies necessitate a comprehensive understanding of bromine chemistry.
The purpose of this chapter is to review recent discoveries in the literature
pertaining to the chemistry of free bromine, particularly as it relates to the
bromination of aromatic DBP precursors. We begin with a brief overview of
bromide occurrence in natural and engineered aquatic systems and then discuss
the chemistry of free bromine generated when bromide-containing source waters
undergo disinfection (e.g., with free chlorine). The bulk of our discussion,
however, is devoted to the often-overlooked ability of halides (chloride and
bromide) and HOCl to catalyze bromination reactions. We conclude with an
analysis of how these catalytic reactions can influence DBP precursor bromination
under model DW treatment conditions.
Natural and Anthropogenic Sources of Bromide
Bromide is a ubiquitous constituent of natural waters (Table 1), with median

concentrations ranging from low µg/L in precipitation (37) to 65 mg/L in seawater
(38). Bromide concentrations in groundwater range from low µg/L to low mg/L
(39, 40). Brackish-water or seawater intrusion can, however, increase bromide
concentrations in aquifers near the coasts (41). Bromide levels measured in DW
(39) and wastewater (23) influents (50–250 µg/L) align with the upper levels
measured in fresh surface waters. Even higher levels have been reported for
surface waters impacted by produced water from hydraulic fracturing operations
(Table 1) (42–44).
252
Table 1. Bromide Concentrations in Natural Waters, Drinking Water, and
Wastewaters
[Br–], (µg/L) Ref
Natural Waters
median maximum
Precipitation ~6 12 (37)
U.S. groundwater (potable waters only) 16 58 (40)
U.S. freshwater lakes 23 322 (39)
U.S. rivers 63 426 (39)
U.S. groundwater (includes non-potable
62 2.7 x 103 (39)
waters)
Seawater 6.5 x 104 not reported (38)

Surface water impacted by produced water not
7.5 x 104 (43)
from hydraulic fracturing reported
Reported Ranges of
Drinking Water and Wastewater Ref
[Br–], (µg/L)
U.S. drinking water influent 50–200 (39)
Municipal wastewater influent 100–250 (23)
Chemistry of Free Bromine
The oxidation state of bromine is +I in species such as HOBr (pKa = 8.70 at

20 °C) (45) and BrO-; the sum of all Br(+I) species is referred to as free bromine.
The direct application of free bromine as a disinfectant of DW is uncommon due
to the increased costs, increased DBP formation, and difficulty in maintaining a
disinfectant residual throughout the distribution system relative to free chlorine
(1). Free bromine is, however, employed as a disinfectant in some recreational
waters (e.g., spas) (1). In waters disinfected with ozone (O3), free bromine can
form via the incomplete oxidation of bromide (46–48). Bromide can also be
oxidized into free bromine in solutions dosed with chloramines (49–51).
Conventional wisdom in the literature generally assumes HOBr is the
predominant brominating agent in solutions of free bromine (52). Under this
assumption, bromination rates can be described via eq 2:
where kHOBr is a second-order rate constant (M-1 s-1), [HOBr] is the concentration
of HOBr (mol/L), [Org-H] is the concentration of the parental organic compound
(mol/L), and [Org-Br] is the concentration of the brominated product (mol/L).
When HOBr is present in large excess, kHOBr can be calculated by eq 3:
253
where kobs is a pseudo-first-order rate constant (s-1). Apparent rate constants (kapp)
are also commonly reported for bromination reactions:
where [Br(I)]tot denotes the total concentration of free bromine. Values of kapp
typically vary as a function of pH (e.g., due to the weak-acid character of HOBr).
That BrO- possesses a net negative charge and requires expulsion of a weakly-
labile leaving group (O2-) suggests HOBr will be a more reactive brominating agent
than BrO- (53).
The kinetic models represented by eqs 3 and 4 do not account for the
potential influence of bromination catalysts. A catalyst is “a substance that affects

the rate of a reaction but emerges from the process unchanged (54).” Below,
we describe the chemistry by which chloride, bromide, and HOCl can catalyze
bromination reactions. Each of these reagents can promote the formation of
reactive brominating agents via eq 5:
where HX = HCl, HBr, or HOCl. BrX ostensibly represents a more inherently

reactive brominating agent relative to HOBr (55). BrX can participate in bromine
substitution reactions with organic compounds via eq 6:
Because HX emerges from this series of reactions (eqs 5 and 6) unchanged,

HX species can be viewed as catalysts of bromination. Although the examples
herein focus on bromination of aromatic compounds, the chemistry described
below conceivably also applies to other organic nucleophiles (e.g., alkenes,
ketones, and amines) present in DW and wastewater.
Catalysis by Chloride
Although commonly overlooked, bromination rates of organic compounds
can depend on the concentration of chloride ions (53, 55–57). For example,
a recent investigation involving the herbicide dimethenamid (a substituted
thiophene) revealed that rates of bromination in solutions of free bromine
increased linearly with the concentration of added chloride (Figure 1) (53). The
influence of chloride on rates of dimethenamid bromination (represented by the
slopes in Figure 1) increased with decreasing pH. These findings are consistent
with the participation of BrCl as a brominating agent, whose concentration is
proportional to the concentration of chloride (eq 7):
254
Although an additional free bromine species (BrCl2-) can also form in the
presence of chloride (58), the formal negative charge on BrCl2- is anticipated to
render this species significantly less electrophilic compared to neutral free bromine
species (e.g., BrCl and HOBr) (55).
Figure 1. Pseudo-first-order rate constants of dimethenamid bromination as

a function of chloride concentration. Arrow denotes site of bromination on
dimethenamid. Error estimates denote 95% confidence intervals. Conditions:
free chlorine dose = 0.43 mM = 30 mg/L as Cl2, [Br-]o = 0.10 mM = 8.0 mg/L,
[dimethenamid]o = 9 µM, [NaNO3] + [NaCl] = 0.1 M, [borate buffer] = 10 mM,
T = 20.0 °C. Adapted from reference (53). Copyright 2013, ACS.
The greater electronegativity of chlorine imparts a partial positive charge on

bromine in BrCl. Consequently, BrCl is anticipated to function as a brominating
agent. During the course of aromatic substitutions involving BrCl as the
electrophile, chloride is released as a leaving group from BrCl (Scheme 1).
Accordingly, chloride can be viewed as a catalyst of bromination reactions.
Chloride catalysis via BrCl formation has also been quantified for bromination of
p-xylene (56, 57) and anisole (55).
Scheme 1. Proposed Mechanism of Aromatic Compound Bromination by Free

Bromine Species (X = Cl, Br, OCl, OBr, or OH) Using Anisole as a Model
Compound. From Reference (55). Copyright 2015, ACS.
255
BrCl has an inherent reactivity 3.6 x 106- and 1.4 x 107-times greater than
HOBr toward dimethenamid (53) and the para position of anisole, respectively
(55). As such, the potency of chloride as a bromination catalyst can be substantial.
The median chloride concentration in raw DW is approximately 10 mg/L (59).
Treatment processes (e.g., coagulation) can increase the chloride concentration
of DW. For example, addition of FeCl3 at 2 mM (~300 mg/L) as a coagulant
concomitantly increases the chloride concentration by ~200 mg/L. This additional
chloride is sufficient to increase bromination rates of anisole and dimethenamid
by more than a factor of 11 (Figure 2).
Figure 2. Catalysis of bromination of anisole (to give 4-bromoanisole (55)) and

of dimethenamid (53) as a function of chloride concentration. Bromination rates
are normalized to values calculated at [Cl-] = 10 mg/L. Conditions: pH = 7.00,
free chlorine dose = 2.0 mg/L as Cl2,[Br-]o = 0.10 mg/L, T = 20.0 °C.
In addition to catalyzing bromination reactions, chloride can similarly
catalyze chlorination reactions of several organic compounds (including p-xylene
(56, 57), the antimicrobial agent trimethoprim (60), polycyclic aromatic
hydrocarbons (61), dimethenamid (62), phenolic compounds (63, 64), and
aromatic ethers (65)) by promoting the formation of aqueous molecular chlorine
(Cl2).
256
Catalysis by Bromide
Unoxidized bromide can coexist in solutions of free bromine, including
when (1) free chlorine is added substoichiometrically relative to bromide, (2)
bromide is incompletely oxidized by ozone, or (3) free bromine is added directly
to aqueous solutions containing bromide (e.g., in some pools and spas). Under
these conditions, unoxidized bromide can enhance rates of organic compound

bromination by promoting the formation of molecular bromine (eq 9):
When Br2 serves as a brominating agent via transfer of Br(I) to an organic

compound, bromide is released as a leaving group from Br2. As such, bromide
can function as a catalyst of bromination reactions involving Br2.
Additional free bromine species (i.e., Br2Cl- and Br3–) can also form in the
presence of excess bromide (58, 68). The formal negative charge on these species
suggests that they will be significantly less reactive as brominating agents relative
to neutral free bromine species (e.g., BrCl and HOBr) (55).
The ability of (unoxidized) bromide to increase bromination rates has
been demonstrated for the sequential bromination of anisole to give mono- and
dibrominated anisoles (55). Figure 3 demonstrates the catalytic effect of bromide
on reactions of anisole with free bromine. Bromide catalysis of bromination
has also been reported for other aromatic compounds, including p-xylene (56),
phenols (69), and dimethenamid (53).
The ratio of second-order rate constants for Br2 and HOBr (i.e., kBr2/kHOBr)
is a means of quantifying the catalytic efficacy of bromide during bromination
reactions involving free bromine. For bromination of dimethenamid (53) and
anisole (para position) (55) at 20 °C, kBr2/kHOBr = 6.6 x 104 and 3.2 x 105,
respectively. The greater catalytic potency of bromide toward bromination of
anisole (relative to the more nucleophilic thiophene ring of dimethenamid) is
consistent with the reactivity-selectivity principle, which predicts an increase
in selectivity toward more reactive (brominating) agents as the reactivity of a
substrate (here, an aromatic compound) decreases (70). The increased catalytic
potency of bromide toward bromination of anisole relative to dimethenamid is
also evident in Figure 4, which depicts normalized bromination rates as a function
of the excess bromide concentration under conditions representative of DW
treatment.
257
Figure 3. Effects of excess bromide (calculated as [Br-]xs = [NaBr]o – [free

chlorine]o) on pseudo-first-order rate constants of para bromination of anisole
in solutions of free bromine. Arrow denotes site of bromination. Error bars
represent 95% confidence intervals. Conditions: pH = 7.31, [bicarbonate buffer]
= 20 mM, [free chlorine]o = 119 µM, [Br-]o = 197-588 µM, [anisole]o = 10
µM, [NaNO3] = 98 mM, no added NaCl, T = 20.0 °C. Adapted from reference
(55). Copyright 2015, ACS.
Catalysis by Hypochlorous Acid

As with chloride and bromide, HOCl can also serve as a catalyst of
bromination reactions. For example, bromination rates of dimethenamid in model
laboratory reactors were previously shown to increase as the dose of free chlorine
increased (53). The authors rationalized this finding by invoking BrOCl as a
reactive brominating agent:
258
As with BrCl, the partial positive charge on the bromine atom of BrOCl and
the greater leaving group ability of ClO- (relative to BrO-) suggest this mixed
halogen will function as a brominating agent (53). As HOCl is a product of
aromatic bromination reactions involving BrOCl (53, 55), HOCl can be viewed as
a bromination catalyst. Catalysis of bromination by HOCl has also been reported
for reactions of anisole in solutions prepared by adding excess free chlorine to
sodium bromide (Figure 5) (55).
Figure 4. Catalysis of bromination of anisole (to give 4-bromoanisole (55)) and

of dimethenamid (53) as a function of excess bromide concentration (calculated
as [Br-]xs = [NaBr]o – [free chlorine]o). Bromination rates are normalized to
values calculated at [Br-]xs = 0.1 mg/L. Conditions: pH = 7.00, [free chlorine]o
= 1.2 µM, [Cl-] = 0.3 mM = 11 mg/L, T = 20.0 °C.
259
Figure 5. Influence of free chlorine concentration on pseudo-first-order rate

constants of para bromination of anisole in solutions of free bromine. Arrow
denotes site of bromination. Error estimates denote 95% confidence intervals.
Conditions: pH 8.49, [Br-]o = 230 µM, [NaNO3] = 93 µM, [Cl-] = 3.9 mM,
[anisole]o = 19 µM, and T = 20.0 °C. Adapted from reference (55). Copyright
2015, ACS.
The influence of free chlorine dose on bromination rates of anisole and

dimethenamid are shown in Figure 6 for solution conditions typical of DW
chlorination. A ten-fold increase in free chlorine dose (from 1 to 10 mg/L as Cl2)
results in a predicted 2.6- and 1.8-fold increase in bromination rates of anisole
(para position) and dimethenamid, respectively.
In addition to promoting the formation of BrOCl, residual free chlorine may
also increase observed rates of bromination in natural waters by facilitating re-
oxidation of bromide (e.g., formed via reduction of free bromine by endogenous
reductants) (71).
260
Figure 6. Catalysis of bromination of anisole (to give 4-bromoanisole (55))

and of dimethenamid (53) as a function of initial free chlorine concentration.
Conditions: pH 7.00, [Br-]o = 0.1 mg/L = 1.25 µM, [Cl-] = 11 mg/L = 0.3 mM,
T = 20.0 °C.
In addition to catalyzing bromination reactions, HOCl can also enhance

chlorination rates of organic compounds (including anisole (72), p-xylene (56),
biphenyl (73), trans-2-butenoic acid (74), dimethenamid (62), and aromatic
ethers (65)) by promoting the formation of Cl2O (eq 11).
Br2O as a Putative Brominating Agent

Bromination reactions of organic compounds in solutions of free bromine are
often assumed to be first-order in [Br(I)]tot. This assumption was recently tested
for bromination reactions of dimethenamid (53). Sivey et al. (53) discovered that
bromination rates of dimethenamid had reaction orders in [HOBr] ranging from
near 1.0 to approximately 1.7 (Figure 7).
261
Figure 7. Reaction order in total free bromine concentration, [Br(I)],

as a function of pH for reactions of dimethenamid with free bromine.
Conditions:[dimethenamid]o = 8 µM, [Br-]o = (1.0 – 4.0) x 10-4 M, initial free
chlorine concentration = 430 µM, [borate buffer] = 10 mM, T = 20.0 °C.
Adapted from reference (53). Copyright 2013, ACS.
Sivey et al. (53) suggested that the participation of Br2O (eq 12) as a
brominating agent could account for the measured reaction rates exhibiting a
greater-than-first-order dependence on total free bromine.
Eq 12 indicates [Br2O] is proportional to [HOBr]2. Unlike Br2O, all

other brominating agents discussed herein (HOBr, BrCl, Br2, and BrOCl) have
concentrations that scale linearly with HOBr. Accordingly, Br2O is the only
examined free bromine species anticipated to produce reaction orders in [Br(I)]tot
greater than 1.0. Br2O is 7100- and 4100-times more inherently reactive than is
HOBr in reactions with dimethenamid (53) and anisole (55), respectively.
A More Complete View of Free Bromine Speciation

Although HOBr and BrO- are typically the most abundant constituents of
free bromine (Figure 8), previous reports (53, 55–57) indicate the generally
less-abundant free bromine species (BrCl, Br2, BrOCl, and Br2O) are sufficiently
reactive as to influence overall bromination rates of modestly nucleophilic
aromatic compounds.
262
Figure 8. Speciation of free bromine for a model water containing bromide (160
µg/L) that undergoes disinfection by (A) free chlorine and (B) a sufficient amount
of ozone to convert 95% of initial bromide into free bromine, Br(I), and assuming
higher oxidation products of bromine (e.g., bromate) are not formed. Data from
reference (55) and sources cited therein.
The extent to which these often-overlooked brominating agents influence

bromination rates under conditions representative of drinking water treatment
can be calculated using second-order bromination rate constants reported for
dimethenamid (53) and anisole (Figure 9) (55). As shown in Figure 9A, HOBr
contributes <20% to the total bromination rate of anisole (to give 4-bromoanisole)
at pH 7.0. The contribution of HOBr to the total bromination rate of dimethenamid
at pH 7.0 is approximately 40% (Figure 9B). For both aromatic compounds, the
influence of HOBr decreases with decreasing pH due to the increasing importance
of BrCl. In solutions containing higher concentrations of bromide and/or chloride
(e.g., wastewater (23) and DW impacted by seawater intrusion (77)) and in waters
receiving larger doses of free chlorine, the importance of brominating agents
other than HOBr is anticipated to increase.
263
Figure 9. Contributions of brominating agents to overall bromination rate

of (A) anisole (yielding 4-bromoanisole) (55) and (B) dimethenamid (adapted
from reference (53), copyright 2013, ACS), both under typical drinking water
chlorination conditions (free chlorine dose = 2 mg/L as Cl2, [Cl-] = 11 mg/L,
[Br-]o = 0.1 mg/L, T = 20.0 °C).
Overall, recent literature reports (53, 55, 56) suggest kinetic models that do not
account for the less abundant free bromine species (BrCl, Br2, Br2O, and BrOCl)
are unable to satisfactorily explain the effects of solution chemistry conditions
on measured bromination rates of modestly nucleophilic aromatic compounds.
Accordingly, caution should be exerted when interpreting published rate constants
(e.g., kapp and kHOBr) that include (often untested) assumptions about the effects
of halides and hypochlorous acid on rates of bromination. The extent to which
catalysis by halides (via formation of BrCl and Br2) and hypochlorous acid (via
formation of BrOCl) influences formation rates of regulated DBPs merits future
research. Such catalysis might be particularly influential, for example, during
reactions involving the “slowly reacting fraction” of trihalomethane precursors
(78).
References
1. Black & Veatch Corporation. White’s Handbook of Chlorination and
Alternative Disinfectants, 5th ed.; Wiley: Hoboken, NJ, 2010.
2. Rook, J. Formation of haloforms during the chlorination of natural water.
Water Treat. Exam. 1974, 23, 234–243.
3. Rook, J. Haloforms in drinking water. J. Amer. Water Works Assoc. 1976,
68, 168–172.
4. Norwood, D.; Johnson, J.; Christman, R.; Hass, J.; Bobenrieth, M. Reactions
of chlorine with selected aromatic models of aquatic humic acid. Environ.
Sci. Technol. 1980, 14, 187–190.
264
5. Boyce, S. D.; Hornig, J. F. Reaction pathways of trihalomethane formation
from the halogenation of dihydroxyaromatic model compounds for humic
acid. Environ. Sci. Technol. 1983, 17, 202–211.
6. de Leer, E.; Damste, J.; Erkelens, C.; de Galan, L. Identification of
intermediates leading to chloroform and C-4 diacids in the chlorination of
humic acid. Environ. Sci. Technol. 1985, 19, 512–522.
7. de Leer, E.; Erkelens, C. Chloroform production from model compounds of
aquatic humic material. The role of pentachlororesorcinol as an intermediate.
Sci. Total Environ. 1985, 47, 211–216.
8. Sedlak, D.; von Gunten, U. The chlorine dilemma. Science 2011, 331, 42–43.
9. Morris, J. C. The acid ionization constant of HOCl from 5 to 35°. J. Phys.
Chem. 1966, 70, 3798–3805.
10. Kumar, K.; Margerum, D. W. Kinetics and mechanism of general-acid-
assisted oxidation of bromide by hypochlorite and hypochlorous acid. Inorg.

Chem. 1987, 26, 2706–2711.
11. Bard, A.; Pearson, R.; Jordan, J. Standard Potentials in Aqueous Solution;
Marcel Dekker: New York, 1985.
12. Hewitt, J. H. The bacterial quality of whirlpools: A study of water in
commercially operated bromine disinfected spa pools. Environ. Health
1991, 99, 324–326.
13. Ao, X.; Li, X. J.; Li, H.; Liang, H. S. Observation on efficacy of bromine-
containing disinfectant in on-the-spot disinfection of swimming pool water.
Chin. J. Disinfect. 2007, 6, 024.
14. Das, T. K. Disinfection. In Kirk-Othmer Encyclopedia of Chemical
Technology; Wiley: New York, 2002.
15. Ichihashi, K.; Teranishi, K.; Ichimura, A. Brominated-trihalomethane
formation from phenolic derivatives as a model of humic materials by the
reaction with hypochlorite and hypobromite ions. Chem. Lett. 1999, 9,
957–958.
16. Acero, J. L.; Piriou, P.; von Gunten, U. Kinetics and mechanisms of formation
of bromophenols during drinking water chlorination: Assessment of taste and
odor development. Water Res. 2005, 39, 2979–2993.
17. Crittenden, J. C.; Trussell, R. R.; Hand, D. W.; Howe, K. J.; Tchobanoglous,
G. Water Treatment Principles and Design; Wiley: Hoboken, NJ, 2005.
18. Westerhoff, P.; Chao, P.; Mash, H. Reactivity of natural organic matter with
aqueous chlorine and bromine. Water Res. 2004, 38, 1502–1513.
19. Zhai, H.; Zhang, X. Formation and decomposition of new and unknown
polar brominated disinfection byproducts during chlorination. Environ. Sci.
Technol. 2011, 45, 2194–2201.
20. Zhao, Y.; Anichina, J.; Lu, X.; Bull, R. J.; Krasner, S. W.; Hrudey, S. E.;
Li, X.-F. Occurrence and formation of chloro- and bromo-benzoquinones
during drinking water disinfection. Water Res. 2012, 46, 4351–4360.
byproducts: Effect of bromide concentration on their formation and
1265–1273.
265
22. Gonsior, M.; Schmitt-Kopplin, P.; Stavklint, H.; Richardson, S. D.;
Hertkorn, N.; Bastviken, D. Changes in dissolved organic matter during the
treatment processes of a drinking water plant in Sweden and formation of
previously unknown disinfection byproducts. Environ. Sci. Technol. 2014,
48, 12714–12722.
23. Tchobanoglous, G.; Burton, F. L.; Stensel, H. D. Wastewater Engineering:
Treatment and Reuse; McGraw-Hill: Boston, 2003.
24. Sun, Y.; Wu, Q.; Hu, H.; Tian, J. Effect of bromide on the formation of
disinfection by-products during wastewater chlorination. Water Res. 2009,
43, 2391–2398.
26. LaKind, J. S.; Richardson, S. D.; Blount, B. C. The good, the bad, and the
volatile: Can we have both healthy pools and healthy people? Environ. Sci.
Technol. 2010, 44, 3205–3210.
27. Xiao, F.; Zhang, X.; Zhai, H.; Lo, I. M. C.; Tipoe, G. L.; Yang, M.; Pan, Y.;
Chen, G. New halogenated disinfection byproducts in swimming pool
water and their permeability across skin. Environ. Sci. Technol. 2012, 46,
7112–7119.
28. Amy, G. L.; Chadik, P. A.; King, P. H.; Cooper, W. J. Chlorine utilization
during trihalomethane formation in the presence of ammonia and bromide.
29. Hua, G.; Reckhow, D. A.; Kim, J. Effect of bromide and iodide ions on
the formation and speciation of disinfection byproducts during chlorination.
30. Hu, J.; Song, H.; Karanfil, T. Comparative analysis of halonitromethane and
trihalomethane formation and speciation in drinking water: The effects of
disinfectants, pH, bromide, and nitrite. Environ. Sci. Technol. 2010, 44,
794–799.
31. Shan, J.; Hu, J.; Kaplan-Bekaroglu, S. S.; Song, H.; Karanfil, T. The effects of
pH, bromide and nitrite on halonitromethane and trihalomethane formation
from amino acids and amino sugars. Chemosphere 2012, 86, 323–328.
871–878.
33. Plewa, M. J.; Wagner, E. D.; Mitch, W. A. Comparative mammalian cell
cytotoxicity of water concentrates from disinfected recreational pools.
35. Liviac, D.; Wagner, E. D.; Mitch, W. A.; Altonji, M. J.; Plewa, M. J.
Genotoxicity of water concentrates from recreational pools after various
disinfection methods. Environ. Sci. Technol. 2010, 44, 3527–3532.
266
36. Hansen, K. M. S.; Willach, S.; Antoniou, M. G.; Mosbak, H.;
Albrechtsen, H.-J.; Andersen, H. R. Effect of pH on the formation of
disinfection byproducts in swimming pool water - Is less THM better?
Water Res. 2012, 46, 6399–6409.
37. Lundstrom, U.; Olin, A. Bromide concentration in Swedish precipitation,
surface and ground waters. Water Res. 1986, 20, 751–756.
38. Stumm, W.; Morgan, J. Aquatic Chemistry: Chemical Equilibria and Rates
in Natural Waters, 3rd ed.; John Wiley & Sons, Inc.: New York, 1996.
39. Westerhoff, P.; Siddiqui, M.; Debroux, J.; Zhai, W.; Ozekin, K.; Amy,
G. Nation-wide bromide occurrence and bromate formation potential
in drinking water supplies. In Critical Issues in Water and Wastewater
Treatment: Proceedings of the 1994 National Conference on Environmental
Engineering; Ryan, J. N., Edwards, M., Eds.; American Society of Civil
Engineers: New York, 1994; pp 670−677.

40. Davis, S.; Fabryka-Martin, J.; Wolfsberg, L. Variations in bromide in potable
ground water in the United States. Groundwater 2004, 42, 902–909.
41. Andreasen, D.; Fleck, W. Use of bromide:chloride ratios to differentiate
potential sources of chloride in a shallow, unconfined aquifer affected by
brackish-water intrusion. Hydrogeol. J. 1997, 5, 17–26.
42. Wilson, J. M.; Van Briesen, J. M. Source water changes and energy extraction
activities in the Monongahela River, 2009–2012. Environ. Sci. Technol.
2013, 47, 12575–12582.
43. Hladik, M. L.; Focazio, M.; Engle, M. Discharges of produced waters
from oil and gas extraction via wastewater treatment plants are sources of
disinfection by-products to receiving streams. Sci. Total Environ. 2014,
466-467, 1085–1093.
44. Harkness, J. S.; Dwyer, G. S.; Warner, N. R.; Parker, K. M.; Mitch, W. A.;
Vengosh, A. Iodide, bromide, and ammonium in hydraulic fracturing and oil
and gas wastewaters: Environmental implications. Environ. Sci. Technol.
2015, 49, 1955–1963.
45. Eigen, M.; Kustin, K. The kinetics of halogen hydrolysis. J. Am. Chem. Soc.
1962, 84, 1355–1361.
46. Haag, W. R.; Hoigne, J. Ozonation of bromide-containing waters: kinetics of
formation of hypobromous acid and bromate. Environ. Sci. Technol. 1983,
17, 261–267.
47. Pinkernell, U.; von Gunten, U. Bromate minimization during ozonation:
Mechanistic considerations. Environ. Sci. Technol. 2001, 35, 2525–2531.
48. Bonacquisti, T. P. A drinking water utility’s perspective on bromide, bromate,
and ozonation. Toxicology 2006, 221, 145–148.
49. Trofe, T. W.; Inman, G. W.; Johnson, J. D. Kinetics of monochloramine
decomposition in the presence of bromide. Environ. Sci. Technol. 1980,
14, 544–549.
50. Bousher, A.; Brimblecombe, P.; Midgley, D. Kinetics of reactions in
solutions containing monochloramine and bromide. Water Res. 1989, 23,
1049–1058.
51. Gazda, M.; Dejarme, L. E.; Choudhury, T. K.; Cooks, R. G.; Margerum, D.
W. Mass-spectrometric evidence for the formation of bromochloramine and
267
N-bromo-N-chloromethylamine in aqueous solution. Environ. Sci. Technol.
1993, 27, 557–561.
52. Heeb, M. B.; Criquet, J.; Zimmermann-Steffens, S.; von Gunten, U. Bromine
production during oxidative water treatment of bromide-containing waters
and its reactions with inorganic and organic compounds – A critical review.
Water Res. 2014, 48, 15–42.
53. Sivey, J. D.; Arey, J. S.; Tentscher, P. R.; Roberts, A. L. Reactivity of BrCl,
Br2, BrOCl, Br2O, and HOBr toward dimethenamid in solutions of bromide
+ aqueous free chlorine. Environ. Sci. Technol. 2013, 47, 1330–1338.
54. Fogler, H. Elements of Chemical Reaction Engineering, 4th ed.; Prentice
Hall: Upper Saddle River, NJ, 2006.
55. Sivey, J. D.; Bickley, M. A.; Victor, D. A. Contributions of BrCl, Br2,
BrOCl, Br2O, and HOBr to regiospecific bromination rates of anisole and
bromoanisoles in aqueous solution. Environ. Sci. Technol. 2015, 49,

4937–4945.
56. Voudrias, E. A.; Reinhard, M. Reactivities of hypochlorous and
hypobromous acid, chlorine monoxide, hypobromous acidium ion, chlorine,
bromine, and bromine chloride in electrophilic aromatic substitution
reactions with p-xylene in water. Environ. Sci. Technol. 1988, 22,
1049–1056.
57. Voudrias, E. A.; Reinhard, M. A kinetic model for the halogenation of p-
xylene in aqueous HOCl solutions containing Cl– and Br–. Environ. Sci.
Technol. 1988, 22, 1056–1062.
58. Liu, Q.; Margerum, D. W. Equilibrium and kinetics of bromine chloride
hydrolysis. Environ. Sci. Technol. 2001, 35, 1127–1133.
59. Davis, S. N.; DeWiest, R. J. M. Hydrogeology; Wiley: New York, 1966.
60. Dodd, M. C.; Huang, C.-H. Aqueous chlorination of the antibacterial agent
trimethoprim: Reaction kinetics and pathways. Water Res. 2007, 41,
647–655.
61. Georgi, A.; Reichl, A.; Trommler, U.; Kopinke, F. Influence of sorption
to dissolved humic substances on transformation reactions of hydrophobic
organic compounds in water. I. Chlorination of PAHs. Environ. Sci. Technol.
2007, 41, 7003–7009.
62. Sivey, J. D.; McCullough, C. E.; Roberts, A. L. Chlorine monoxide (Cl2O)
and molecular chlorine (Cl2) as active chlorinating agents in reaction of
dimethenamid with aqueous free chlorine. Environ. Sci. Technol. 2010, 44,
3357–3362.
63. Deborde, M.; Rabouan, S.; Gallard, H.; Legube, B. Aqueous chlorination
kinetics of some endocrine disruptors. Environ. Sci. Technol. 2004, 38,
5577–5583.
64. Curtis, M. P.; Hicks, A. J.; Neidigh, J. W. Kinetics of 3-chlorotyrosine
formation and loss due to hypochlorous acid and chloramines. Chem. Res.
Toxicol. 2011, 24, 418–428.
65. Sivey, J. D.; Roberts, A. L. Assessing the reactivity of free chlorine
constituents Cl2, Cl2O, and HOCl toward aromatic ethers. Environ. Sci.
Technol. 2012, 46, 2141–2147.
268
66. Wang, T. X.; Margerum, D. W. Kinetics of reversible chlorine hydrolysis:
temperature dependence and general-acid/base-assisted mechanisms. Inorg.
Chem. 1994, 33, 1050–1055.
67. Liebhafsky, H. A. The equilibrium constant of the bromine hydrolysis and
its variation with temperature. J. Am. Chem. Soc. 1934, 56, 1500–1505.
68. Wang, T. X.; Kelley, M. D.; Cooper, J. N.; Beckwith, R. C.; Margerum, D.
W. Equilibrium, kinetic, and UV-spectral characteristics of aqueous bromine
chloride, bromine, and chlorine species. Inorg. Chem. 1994, 33, 5872–5878.
69. Tee, O.; Paventi, M.; Bennett, J. Kinetics and mechanism of the bromination
of phenols and phenoxide ions in aqueous solution. Diffusion-controlled
rates. J. Am. Chem. Soc. 1989, 111, 2233–2240.
70. Anslyn, E. V.; Dougherty, D. A. Modern Physical Organic Chemistry;
University Science: Sausalito, CA, 2005.
71. Allard, S.; Fouche, L.; Dick, J.; Heitz, A.; von Gunten, U. Oxidation of
manganese(II) during chlorination: Role of bromide. Environ. Sci. Technol.
2013, 47, 8716–8723.
72. Swain, C. G.; Crist, D. R. Mechanisms of chlorination by hypochlorous acid.
The last of chlorinium ion, Cl+. J. Am. Chem. Soc. 1972, 94, 3195–3200.
73. Snider, E. H.; Alley, F. C. Kinetics of the chlorination of biphenyl under
conditions of waste treatment processes. Environ. Sci. Technol. 1979, 13,
1244–1248.
74. Craw, D.; Israel, G. The kinetics of chlorohydrin formation. Part III. The
reaction between hypochlorous acid and crotonic acid at constant pH. J.
Chem. Soc. 1952, 550–553.
75. Roth, W. A. Zur Thermochemie des Chlors und der unterchlorigen Säure. Z.
Phys. Chem., Abt. A 1929, 145, 289–297.
76. Reinhard, M.; Redden, G. D.; Voudrias, E. A. The hydrolysis constant of
chlorine monoxide and bromine chloride in water. In Water Chlorination:
Environmental Impact and Health Effects; Jolley, R. L., Ed.; Ann Arbor
Science: Ann Arbor, MI, 1990; pp 859−870.
77. Ged, E. C.; Boyer, T. H. Effect of seawater intrusion on formation
of bromine-containing trihalomethanes and haloacetic acids during
chlorination. Desalination 2014, 345, 85–93.
78. Gallard, H.; von Gunten, U. Chlorination of natural organic matter: kinetics
of chlorination and of THM formation. Water Res. 2002, 36, 65–74.
269
Chapter 15
Prescribed Fire Alters Dissolved Organic

Matter and Disinfection By-Product Precursors
in Forested Watersheds - Part I.
A Controlled Laboratory Study
Hamed Majidzadeh,1,2,3 Jun-Jian Wang,2,3 and Alex T. Chow*,2

1School of Forestry and Wildlife Science, Auburn University,
Auburn, Alabama 36849
2Baruch Institute of Coastal Ecology & Forest Science, Clemson University,
Georgetown, South Carolina 29442

3H.M. and J.W. contributed equally.
Detritus material in forested watersheds is the major terrestrial

source of dissolved organic matter (DOM) and disinfection
by-product (DBP) precursors in source waters. Forest fire
reduces the thickness of detritus layer and changes foliar litters
into pyrogenic organic matter (PyOM) on the forest floor,
resulting in different quantity and quality of DOM exported
from forested watersheds. Many studies have examined DBP
precursors exported from forested watersheds; however, DOM
leaching from PyOM, or dissolved black carbon (BC), could
have different reactivity in DBP formation compared to the
DOM leaching from unburned detritus layer. Using controlled
laboratory burning in this study, characteristics of foliar litters
before and after burn were compared. Quality and quantity of
water extractable organic matter (WEOM) from raw and burned
foliar litters commonly found in the southeastern United States,
including baldcypress (Taxodium distichum), boxelder (Acer
negundo), longleaf pine (Pinus palustris), pop ash (Fraxinus
caroliniana), sweetgum (Liquidambar styraciflua), and water
tupelo (Nyssa aquatica) was compared, and evaluated for
their disinfection by-product yield and specific DBP formation
using a uniform formation condition test. Laboratory analysis

using pyrolysis gas chromatography mass spectrometry showed
that foliar litters after burn contain significantly less lignins,
nitrogenous compounds, and polysaccharides because of
loss of organic matter. However, both the formula numbers
and relative abundance of PAHs increased, which can be a
potential health concern. Water extractable organic carbon
(WEOC) and total nitrogen (WETN) decreased after burn and in
particular raw material had less humic acid and fulvic acid-like
compounds. PyOM aromaticity may be either enhanced or
lowered by prescribed fire and highly depending upon the plant
species. The specific formation of chloroform (a prevalent DBP
species) decreased by 12%-60% for all plant species after burn.
However, dissolved BC was more reactive in the formation
of nitrogenous disinfection by-products (N-DBPs), and the

specific haloacetonitriles formation increased after burn for all
species.
Introduction
Forests are critically important to the supply of clean drinking water in the
United States. National forests and grasslands, which represent 30% of the forest
area in the U.S., provide drinking water for over 60 million people (1). The
number of people served by all forests and grasslands in the U.S. are far greater
than this number. Hot temperatures and longer dry periods due to climate change
will increase the likelihood of larger and more severe wildfires (2). Forest fire
rapidly modifies the chemical composition of the detritus layer on the forest
floor, converting lignin and polysaccharide rich and relatively degradable carbon
pools to polycyclic aromatic and charcoal rich and recalcitrant black carbon
(BC) (3). Dissolved organic matter (DOM) leaching from pyrogenic organic
matter (PyOM), or dissolved BC, could have different treatability and reactivity
in disinfection by-product (DBP) formation compared to the DOM leaching from
unburned detritus layer. In contrast to wildfire, prescribed fire is low-intensity
burning of accumulated detritus materials to reduce the fuel amount on forest floor
to prevent catastrophic wildfires (4). It is often conducted when soil is moderately
moist to reduce the intensity, and it brings benefits to decrease the invasive species
and release different nutrients from plant materials (5). While forested ecosystems
have been identified as persistent sinks for atmospheric carbon (6), prescribed
burns can affect forest’s capacity to sequester carbon and carbon quality and
quantity in soil organic matter pools (7, 8). For example, forest fire converts plant
biomass into forms of BC such as polycyclic aromatic hydrocarbons (PAHs) that
are resistant to microbial attack and persist in the soil for thousands of years (9).
These new compounds are the products of incomplete combustion of biomass,
which usually had lower H/C and O/C ratios (10). Hockaday, et al. (2007) (11)
found that export of BC from soils via dissolution and biological transformation
could constitute an important C loss mechanism. In fact, charred plant materials
272
can cause accelerated breakdown of simple carbohydrates (12) and enhance
the loss of forest humus (13). As detritus in forest floor is one major source of
terrestrial DOM (14), export of these newly formed and modified compounds
is of great concern because they may alter the water quality by changing DOM
quantity and quality.
The well-documented toxicological properties of DBPs suggested that
exposure to DBPs in disinfected water may increase the risk of cancers such
as bladder and colorectal cancers (15, 16). As DOM is an important precursor
reacting with disinfectants during water treatment to yield carbonaceous (e.g.,
trihalomethanes; THMs) and nitrogenous disinfection by-products (N-DBPs)
(e.g., haloacetonitriles; HANs) (17–19) increasing studies are focusing on the
sources of DOM and the associated chlorine reactivity (20, 21). Litters have
been identified as a significant watershed source of DOM contributing to the
downstream water supply and DBP precursors (14, 22–24). Different plant
species can have a diverse amount and composition of organic compounds such
as polysaccharides, lignin, tannin, and aliphatic biopolymers (25). Thus, variation
in plant species can affect the DOM composition and may affect the budget and
speciation of DBPs in water treatment process. Burning of diverse plant materials
would affect the quality and quantity of DOM and the subsequent DBP formation
as well, which, however, has not been well studied.
In order to understand the effects of low intensity prescribed forest fire
on the productions of DOM and DBP precursors in forested watersheds, we
conducted a controlled laboratory study (Part I) and a controlled field study (Part
II) independently. In this chapter we present Part I, the laboratory study where
we compared characteristics of water extractable organic matter (WEOM) from
raw and laboratory-controlled burned foliar litters of six plant species commonly
found in the southeastern United States, including baldcypress (Taxodium
distichum), boxelder (Acer negundo), longleaf pine (Pinus palustris), pop ash
(Fraxinus caroliniana), sweetgum (Liquidambar styraciflua), and water tupelo
(Nyssa aquatica), and evaluated their DBP formation using a uniform formation
condition test. Results of this study would illustrate the productions of DOM
and DBP precursors among different tree species under prescribed fire practices.
Results of this controlled experiment would also be used to compare to the real
prescribed fire practices (Part II), as described in the other chapter of Part II.
Material and Methods

Litter Collection and Black Carbon Preparation
Fresh litter samples in Congaree National Park (SC, U.S.A.) were collected
using five 0.5 m × 0.5 m litterfall traps with 1 mm mesh fiberglass screen bottoms
during the winter of 2010. The Congaree River drains a large portion of upstate
South Carolina, and thus, can flood rapidly during heavy rain events (26). The
Congaree National Park has 10 flooding events per year on average, with mean
annual precipitation of 1220 mm (27). The foliar litter of six common tree
species in Congaree National Park, including baldcypress (Taxodium distichum),
273
sweetgum (Liquidambar styraciflua), water tupelo (Nyssa aquatica), pop ash
(Fraxinus caroliniana), longleaf pine (Pinus palustris), and boxelder (Acer
negundo) were separated from the mixed litterfall. After collection, all litter
samples were oven-dried at 50 °C for 48 hr. To simulate the burning, 10.00 g of
dried litter material was placed in a pre-heated furnace at 340 °C for 4 minutes and
then cooled down gradually in the muffle furnace for three hours. Samples that
had flame were not put out manually during the whole process. The temperature
of 340 °C has been reported to be average soil temperature in Ichauway Reserve
during a prescribed fire on 1994 (28). After cooling, samples were weighed again
to determine the weight lost. The burning for each plant species was conducted
in triplicate. To reduce the heterogeneity of the sample, all unburned and burned
litter materials were grounded and all passed through a 2-mm copper screen (20).
Chemical Characteristics of Original and Burned Litters
Pyrolysis gas chromatography mass spectrometry (Py-GC/MS) was used to

analyze the most common three litters (longleaf pine, pop ash, and sweetgum)
before and after burn following the same method described in Song et al. (2010)
(29). Py-GC/MS results are presented as the relative abundance of specific
compound by dividing the peak area of the identified compound with the total
area of all compounds. The major identified compounds were categorized into
seven groups: aromatics, lignin, lipid, PAHs, phenolics, polysaccharide, and
nitrogen. The detail compounds under each group were listed in Table A1.
Characterization of DOM from Litter
Water extractable organic matter (WEOM) was obtained by mixing one

gram of unburned or burned litter with 200 mL Milli-Q water shaking at room
temperature for 2 hr at 210 rpm. Each water extract was filtered through
pre-washed 0.45 μm polyethersulfone membrane Supor-450 (Pall Corporation),
and the filtrate was refrigerated at 4 °C through completion of analysis. Dissolved
organic carbon (DOC) and dissolved total nitrogen (DTN) were measured by
Shimadzu TOC-V CHS analyzer. The extractability of litter material was defined
as the mass of extractable organic carbon per gram original (unburned) or
burned litter. The post-burn water extractable organic carbon (WEOC) yield
from original litter (mg-C/g-original-litter) was calculated as the extractability of
burned litter multiplied by the mass remaining of burned material compared to
original material.
The UV-VIS absorption (UVA) spectroscopy between 200 and 700 nm was
conducted using Shimadzu UV-1800 in room temperature (25±1 °C) in 1 cm
quartz cell. Specific ultraviolet absorbance at 254 nm (SUVA) was determined
by normalizing UVA with DOC concentration and was reported in units of liter
per milligram carbon per meter (L mg-1 m-1) (30). The E2:E3 ratio has been
calculated as the absorbance at 250 nm divided by the absorbance at 365 nm (31).
SUVA and E2:E3 ratios are considered as surrogates for aromatic carbon and
molecular weight of DOM, respectively.
274
Fluorescence emission and excitation matrices (EEMs) were obtained using
a Shimadzu RF5301 with 5 nm slit in excitation and 5 nm slit in emission. EEMs
were corrected and analyzed by a numerical method, the fluorescence regional
integration based on Simpson’s Rule (FRI-SR) as described in Zhou et al. (2013)
(32). The fluorescence regional integration (FRI) is a quantitative technique that
integrates the volume beneath an EEM. Five regions in EEMs were operationally
defined using consistent excitation (ex) and emission (em) wavelength boundaries
based on fluorescence of model compounds, and DOM fractions: I: Aromatic
Protein I (ex: 220 nm-250 nm, em: 280 nm-330 nm), II: Aromatic Protein II (ex:
220 nm-250 nm, em: 330 nm-380 nm), III: Fulvic acid-like (ex: 220 nm-250
nm, em: 380 nm-550 nm), IV: Soluble microbial by-product-like (250 nm < ex
< 400 nm, em: 280 nm-380 nm), and V: Humic acid-like (250 nm < ex < 400
nm, em: 380 nm-550 nm). Volume integrated under the EEM within region I is
normalized to the projected excitation-emission area and resulted in a normalized

percent fluorescence response (PI,n) (32, 33).
Disinfection By-Products Formation

The uniform formation condition (34) was the chlorination procedure used
in this study and three replicates were analyzed for each litter extract. Briefly,
the uniform formation condition is the condition that yields 1±0.4 mg L−1 free
chlorine residual at pH 8±0.2 after 24 hr of incubation in the dark at 20 °C after
chlorination with freshly prepared NaOCl/H3BO3 dosing solution. Chlorine
demand was calculated as the difference between chlorine dose and residual
chlorine concentration after 24 h. Chlorinated samples were analyzed using
headspace GC-ECD (35). Four trihalomethanes (chloroform, dichlorobromo-
methane, dibromochloromethane, and bromoform) and four haloacetonitriles
(trichloroacetonitrile, dichloroacetonitrile, bromochloroacetonitrile, and
dibromoacetonitrile) were quantified with the minimum reporting levels at
approximately 0.5 μg/L. Specific DBP formation, defined as DBP concentration
normalized to the initial DOC concentration, was used to evaluate the reactivity
of DOC in forming DBPs and was expressed in molar units, as mmol/mol-C (20).
Statistical Analyses
All statistical analyses conducted in “R” version 2.15.0 (Copyright (C) 2012
The R Foundation for Statistical Computing). Relationships were considered to
be significant at P < 0.05. Pearson’s correlation was used to examine correlation
between different parameters.

Chemical Characteristics of Litter Materials
In the spectra from Py-GCMS, both intensity and number of peaks decreased
in burned litter compared to original litter, which is mainly due to the loss
275
of most organic matter during burning. The lignins, nitrogenous compounds,
and polysaccharides all decreased for all three plant species, which can be
explained by the degradation and volatilization at high temperatures (8). It
has been reported that distillation of volatiles and loss of organic carbon start
at temperatures between 100 and 200 °C (36, 37). Between 130 and 190 °C
lignin and hemicelluloses begin to degrade, and at temperatures above 250 °C,
polyaromatic structure start to form, whereas above 300 °C PAHs can be detected
as well in plant char derived from cellulose, pectin, lignin and chlorogenic acid
(38–40). This data matches our results considering significant decrease in lignin
and polysaccharide as depicted in Figure 1. Some new peaks were observed in
burned litters as well. The results indicate increases in both the abundance of PAH
types (7 types of PAHs before burn to 15 types after burn) and relative abundance
of PAHs (2.2% before burn and 7.7% after burn). This result is in agreement with
the well-documented PAH formation in burning process (41). The percentage

of aromatic compounds also increased from 15.2% to 32.3%, suggesting the
formation of new aromatic structure in the organic matter. Production of these
carcinogenic PAHs (42) from forest fires and prescribed burns is of great health
concern to firefighters, forest workers and nearby rural communities including
several Native American tribes (43, 44). The characteristics and mutagenic
activities of biomass-derived PAHs from forest fires have been described,
principally on the particulate matter emission to the atmosphere (45, 46).
Effect of Prescribed Fire on Carbon and Nitrogen Quantity

Comparing litters of different plant species before burning, sweetgum litter
was the largest producer of water extractable organic carbon (WEOC), generating
84.62 mg-DOC per gram of litter (mg-DOC/g-original-litter), followed by longleaf
pine, tupelo, baldcypress and pop ash (53.49 to 31.66 mg-DOC/g-original-litter)
(Table 1). After burning, the DOC leached from burned litter decreased by 83% on
average and ranged from 20.72 (longleaf pine) to 1.78 mg-DOC/g-original-litter
(pop ash) (Table 1). Higher WEOC leaching from original litters was expected
because they contained a greater portion of readily soluble and degradable organic
C, and the burning processes caused a substantial loss of organic matter (47, 48).
Thermal treatments such as prescribed fire remove preferentially external oxygen
groups yielding materials with comparatively reduced solubility and colloidal
properties (49, 50).
Water extractable total nitrogen (WETN) was also higher before burn, ranging
from 0.36 (longleaf pine) to 1.65 mg-N/g-original-litter (boxelder) (average = 0.69
mg-N/g-original-litter), while after burn WETN ranged 0.05 (ash) to 0.87 mg-N/
g-original-litter (boxelder) (average = 0.25 mg-N/g-original-litter). WETN after
burn decreased mainly due to volatilizing of nitrogen. At 200 °C nitrogen starts
to volatilize and above 500 °C half of the nitrogen in organic matter is lost to the
atmosphere (8). The carbon to nitrogen ratio (C/N) was also calculated before and
after burn. As expected, the C/N decreases after burn for all the specious except
for longleaf pine although both C and N quantity was decreased for this specious.
A positive relation between litter’s weight lost after burn and initial litter C/N was
observed (r2 = 0.69, p = 0.04).
276
Figure 1. Chemical composition of original (white bars) and burned (gray bars)
litters as determined from pyrolysis gas chromatography mass spectrometry
(Py-GC/MS).
Effect of Fire on Carbon Quality

UV-VIS Spectrometry
SUVA, an indicator of the degree of aromaticity, increased significantly after

burn for baldcypress, boxelder, longleaf pine, pop ash, and sweetgum except for
water tupelo (Table 2). E2/E3 decrease for baldcypress, longleaf pine, pop ash,
and sweet gum, but not for water tupelo and boxelder. Vergnoux et al. (2011)
(44) reported that after burn WEOM exhibited both a higher aromaticity due
to the increase of the hydrophilic aromaticity and lower humification and less
high weighted compounds by the contribution of transphilic and hydrophobic
compunds. However, our results revealed that the WEOM aromaticity of litter
materials may be either enhanced or lowered by prescribed fire highly depending
on the plant species. There were strong correlations between DOC and UVA
both before and after burn. However, different SUVA suggested that their slopes
were different before and after burn (Table 2), indicating a chemical change in
277
DOM occurred. While there is no correlation between DOC and SUVA (r2 =
0.06, p = 0.6) before burn, a strong correlation can be seen after burn (r2 = 0.79,
p = 0.01) (Table 3 & Table 4). DOC from different litters appears to have more
complex composition before burn in comparison to after burn, which reveals the
homogenization of biomass substrate in black carbon, which is in accordance
with a previous study (44) A stronger correlation between DOC and UVA after
burn might be the result of homogenization as well (Table 3 & Table 4).
Table 1. Weight of Litters and Chemical Characteristics of Water Extracts

of Litters before and after Burns (Means of Triplicates Are Present). One
Gram of Raw Litter or Burned Litter Was Mixed with 200 mL Milli-Q
Water and Shaken at Room Temperature for 2 Hours at 210 rpm.
Sample Weight Pre- Post- Pre- Post- Pre- Post-
loss burn burn burn burn burn burn
WEOCa WEOC WETNa WETN C/N C/N
Bald- 70.3% 35.8 6.7 0.72 0.35 57.7 21.6
cypress
Boxelder 68.6% 47.1 16.0 1.65 0.87 33.4 21.6
Longleaf 48.1% 62.2 20.7 0.36 0.07 200.3 348.8
pine
Pop ash 65.8% 31.7 1.8 0.61 0.05 60.6 39.1
Sweet- 54.7% 84.6 3.8 0.48 0.04 206.1 103.4
gum
Water 64.3% 53.5 3.4 0.34 0.11 183.7 34.6
tupelo
a WEOC and WETN are both in mg /g-original-litter.
Fluorescence Spectrometry
A substantial loss of organic matter was observed in region I of the EEM

(Aromatic protein I). The percent fluorescence response in region I (PI,n) ranged
from 29% to 43% before burn, and decreased by 45%-87% after burn for all species
(Table 2 & Figure 2). No significant changes in region II response was observed
after burn except for longleaf pine. Region III and V percentages, which are related
to fulvic acid-like and humic acid-like DOM, both have increased dramatically
after burn (Table 2 & Figure 2). Sum of regions III and V showed an increase of
31%-76% after burn.
278
The more recalcitrant and aromatic structures formed by fires are derived from
the alteration of carbohydrates, lipids, alkylated macromolecules, and peptides
(51, 52). Consequently, some of these newly formed structures, initially non-
humified, can become extractable like the humic- and fulvic- like fractions (46,
53). The FRI data reveals the difference in DOM composition before and after
burning which is in accordance with our SUVA and E2:E3 data. While before
burn region I exhibited high fluorescence response after burn region III and V had
the highest responses except for longleaf pine.
Table 2. Optical Properties of Water Extracts of Litters before and after

Burns. Fluorescence Regional Integration (FRI) Percentage Distribution of
the Water Samples, SUVA, and E2/E3 Ratio.
UV-Vis properties Percent fluorescence response (%)
Species SUVA E2/E3 I II III IV V
(L/mg/m)
Before burn
Baldcypress 1.32 6.19 30.52 29.49 12.53 18.39 9.07
Boxelder 1.14 4.05 29.31 31.13 15.75 15.66 8.16
Longleaf pine 1.01 9.24 43.08 29.21 7.11 16.65 3.94
Pop ash 1.55 4.92 39.16 28.29 11.86 15.20 5.49
Sweetgum 1.93 5.28 38.19 33.10 9.11 15.36 4.24
Water tupelo 2.56 4.89 37.05 29.95 13.55 12.37 7.09
After burn
Baldcypress 2.09 5.42 12.00 30.84 29.11 14.56 13.49
Boxelder 2.07 5.56 10.83 32.16 29.8 14.63 12.58
Longleaf pine 1.21 8.66 23.27 43.48 11.73 17.14 4.39
Pop ash 2.50 4.31 4.92 26.48 36.50 11.88 20.22
Sweetgum 2.32 5.27 7.41 26.17 37.44 11.87 17.11
Water tupelo 2.27 5.73 6.14 26.77 36.85 12.32 17.92
279
Table 3. Pearson’s Correlation Coefficients (r) between Measured Parameters of Litters before Burn. Significant Correlations
(p < 0.05) Are Highlighted in Bold.
Specific
WEOC TN C/N UVA SUVA E2/E3 I II III IV V THM

formation
TN -0.3
C/N 0.81 -0.74
UVA 0.79 -0.45 0.75
SUVA 0.25 -0.47 0.44 0.78
E2/E3 0.21 -0.51 0.48 -0.2 -0.44
I 0.39 -0.77 0.72 0.28 0.13 0.6
280
II 0.79 0.2 0.34 0.7 0.24 -0.31 -0.21

III -0.55 0.7 -0.7 -0.21 0.17 -0.81 -0.82 -0.02
IV -0.21 0.21 -0.37 -0.6 -0.80 0.44 -0.26 -0.12 -0.25
V -0.63 0.56 -0.71 -0.38 -0.02 -0.48 -0.91 -0.14 0.83 0.17
Specific THM
0.80 -0.35 0.59 0.81 0.41 -0.03 0.1 0.80 -0.38 -0.03 -0.3
formation
Specific HAN
-0.35 0.18 -0.42 -0.67 -0.76 0.45 -0.3 -0.27 -0.16 0.98 0.31 -0.14
formation

Table 4. Pearson’s Correlation Coefficients (r) between Measured Parameters of Litters after Burn. Significant Correlations
(p < 0.05) Are Highlighted in Bold.
Specific
WEOC TN C/N UVA SUVA E2/E3 I II III IV V THM

formation
TN 0.41
C/N 0.67 -0.39
UVA 0.91 0.75 0.31
SUVA -0.89 -0.01 -0.87 -0.63
E2/E3 0.81 -0.12 0.90 0.53 -0.97
I 0.88 0.03 0.85 0.64 -0.99 0.93
281
II 0.92 0.08 0.84 0.68 -0.98 0.92 0.98

III -0.89 -0.04 -0.84 -0.64 0.98 -0.91 -0.99 -1.00
IV 0.92 0.28 0.68 0.77 -0.94 0.84 0.96 0.96 -0.96
V -0.93 -0.17 -0.79 -0.74 0.98 -0.92 -0.99 -0.97 0.97 -0.97
Specific THM
-0.27 0.15 -0.36 -0.09 0.23 -0.15 -0.24 -0.36 0.36 -0.23 0.16
formation
Specific HAN
-0.73 -0.48 -0.32 -0.72 0.72 -0.67 -0.71 -0.69 0.68 -0.82 0.78 -0.28
formation

In other words, raw litter extracts have much less humic acid and fulvic acid-
like compounds in comparison to black carbon extracts. Longleaf pine had the
lowest increase both in region III and V after burn and also the lowest decrease
in region I. It also had the lowest weight loss and DOC decrease, suggesting that
longleaf shows significantly fewer changes both in carbon quality and quantity
after burn in comparison to other studied species. It is in accordance with its
classification as a “fire-resistant” species. Resistance against fire is mainly due
to its moist and dense needles (54).
Figure 2. Fluorescence emission-excitation matrix of longleaf pine and sweetgum

water extracts before and after burn.
Effect of Prescribed Burning on Formation of Disinfection

By-Products
DOC has been implicated as a primary contributor to the formation of
THMs (19), while dissolved organic nitrogen has been identified as a precursor
to both THMs and N-DBPs (e.g., HANs) during drinking water purification
(55). However, little is known about how BC (burned litters) contributes to DBP
formation. As shown in our results, chloroform was the major DBP species after
282
chlorination for all plant species (Figure 3). For the original litters before burn,
the specific chloroform formation followed the order: sweetgum (8.34±1.11
mmol/mol-C) > baldcypress (4.67±0.18 mmol/mol-C) = water tupelo (4.45±0.21
mmol/mol-C) > longleaf pine (3.61±0.17 mmol/mol-C) > boxelder (2.99±0.06
mmol/mol-C) > pop ash (2.59±0.03 mmol/mol-C). The dichlorobromomethane
was the other detectable THM species, with similar specific formation
(0.008-0.011 mmol/mol-C) among all original unburned litters. For the HAN,
dichloroacetonitrile and trichloroacetonitrile were the only two detectable species.
The specific dichloroacetonitrile formation followed: boxelder (0.214±0.003
mmol/mol-C) > pop ash (0.137±0.001 mmol/mol-C) > baldcypress (0.127±0.008
mmol/mol-C) > longleaf pine (0.092±0.011 mmol/mol-C) > sweetgum
(0.083±0.006 mmol/mol-C) > water tupelo (0.066±0.004 mmol/mol-C). The
specific trichloroacetonitrile formation was much lower than dichloroacetonitrile,
which was highest for boxelder (0.007±0.000 mmol/mol-C) and did not show
difference among other plant species (0.003-0.005 mmol/mol-C).
Figure 3. Chlorine reactivity of dissolved organic carbon extracted from litter

materials as indicated by the specific DBP formation.
After burning, both the specific THM formation and specific HAN formation
were significantly altered, which is in agreement with the results of field studies
of prescribed fire (Part II) and wildfire. The specific chloroform formation
decreased by 12%-60% for all plant species (56). However, the specific
dichlorobromomethane formation increased for baldcypress (52%), boxelder
(55%), pop ash (347%), and water tupelo (69%). Different from chloroform,
the specific haloacetonitriles formation increased for all species. While the
specific dichloroacetonitrile formation increased by 30%-1457%, the specific
trichloroacetonitrile formation increased by 7%-73%. There was no correlation
between the percent decrease of specific THM formation and the percent decrease
283
of specific HAN formation, suggesting that the fire’s effects on the THM precursor
and HAN precursor were independent. However, correlation between specific
HAN formation and percent fluorescence response of region IV was significant
for both before and after burn (Table 3). This may provide a fast easy method for
prediction of specific HAN formation both before and after burn.
Multiplying the specific DBP formation with the WEOC, we can estimate
the DBP precursor yield (μg/g-original-litter) from the litter material (Figure
4). Before burn, the litter materials can potentially produce THMs ranging
from 82.3±1.0 μg-THMs/g-original-litter μg-THMs/g-original-litter (pop ash)
to 706.4±93.7 μg-THMs/g-original-litter (sweetgum). After burn, the THM
yield decreased by 70%-98% for all species (Figure 4a). For HANs, these litter
materials had the HAN yield ranging from 3.72±0.24 μg-HANs/g-original-litter
(water tupelo) to 10.40±0.14 μg-HANs/g-original-litter (boxelder). As the
specific HAN formation has increased after burn, there was a much less decrease
in HAN yield compared to THM yield after burn. The HAN yield dropped by
29%-69%, for all species (Figure 4b).
Figure 4. The disinfection by-product yields as in trihalomethane (A) and

haloacetonitriles (B) before and after burn at 340 °C. The yield calculations were
based on the original unburned biomass.
The results demonstrated the effects of low intensity burning (i.e. 340 °C) on
changes in quantity and quality of DBP precursors on the forest floor. Because
fire can largely consume organic matter in detritus layer, WEOM (i.e. WEOC
and WETN) decreased after burn. Such consumption of litter materials is likely
an overriding effect to reduce the DBP precursors, which resulted in the overall
reduction in DBP yield from litter materials (Figure 4). Chlorine reactivity of
the burned litter in forming HANs is much higher than that of original litter.
As the HANs have been documented to be much more toxic than the common
carbonaceous DBPs such as THMs and haloacetic acids, attention should be
placed on the nitrogenous DBPs in water treatment of fire-affected DOM.
The study revealed the effect of vegetation cover on the potential DBP
export. We found that among the major plant species within the watershed, the
sweetgum has a much larger content of DBP precursor because its litter has
high WEOM that is also very reactive in forming DBPs (ranking 1st for THM
284
yield and the 2nd for HAN yield; Figure 4). In contrast, the pop ash is a species
with much less contribution to the watershed export of DBP precursor, with
only 12% THM formation and 61% HAN formation relative to the sweetgum.
The response of DBP precursors in litter materials is highly dependent upon
plant species as shown in Figure 4. Therefore, the knowledge acquired in one
watershed about the fire effect on the DBP export may not be directly applicable
to estimate the DBP export in another watershed that has quite different dominant
plant species. The controlled lab study here is designed for elucidating the
effects of simulated burning on the DBP precursor in litter materials without
any interference from other environmental factors such as mixes of vegetations,
fuel moistures, decomposition stages, etc. To evaluate a real suitation on DBP
precursor production from a burnt forest, our research team conducted a controlled
burning experiment and natural exposure of WEOM in a natural forest at Hobcaw
Bonary, Georgetown SC. Results and findings of the controlled field study, as the
Part II of this prescribed fire study, are discussed in the following chapter.
Acknowledgments
Jun-Jian Wang received financial support from the China Scholarship Council
(CSC[2011]3010). Portions of this study were also supported by NIFA/USDA
under project number SC1700489 and 2014-67019-21615, Joint Fire Science
Program 14-1-06-19, and NSF Rapid Grant 1264579 as presented in technical
contribution number 6359 of the Clemson University Experiment Station.
Appendix
Table A1. The Categorization of Pyrogenic Products from Results of
Pyrolysis Gas Chromatography Mass Spectrometry: Ar = aromatics; Lg =
lignins; Lp = lipids; Ps = polysaccharides; N = nitrogens; PAHs = polycyclic
aromatic hydrocarbons; Ph = phenolics.
Ty-
# Compound # Compound Type
pe
1 Phenol, 2-methoxy- Lg 77 Heptadecane Lp
2 Phenol,2-methoxy-4-methyl- Lg 78 1-Eicosene Lp
3 Phenol,4-methoxy-3-methyl- Lg 79 1-Nonene Lp
4 1,2-Benzenediol,3-methoxy- Lg 80 Tridecane Lp
5 Phenol,4-ethyl-2-methoxy- Lg 81 Heptadecane Lp
6 4-Hydroxy-3-methoxystyrene Lg 82 Toluene Ar
7 Phenol,3-methoxy-5-methyl- Lg 83 Ethylbenzene Ar
8 Phenol, 2,6-dimethoxy- Lg 84 P-Xylene Ar
Phenol,2-methoxy-4-(1-
9 Lg 85 P-Xylene Ar
propenyl)-
285
Table A1. (Continued). The Categorization of Pyrogenic Products from
Results of Pyrolysis Gas Chromatography Mass Spectrometry: Ar =
aromatics; Lg = lignins; Lp = lipids; Ps = polysaccharides; N = nitrogens;
PAHs = polycyclic aromatic hydrocarbons; Ph = phenolics.
Ty-
pe
10 3,4-Dimethoxyphenol Lg 86 Benzene, Popyle- Ar
11 Vanillin Lg 87 4-Ethyltoluene Ar
12 Eugenol Lg 88 1,2,4-Trimethylbenzene Ar
Phenol,2-methoxy-4-(1E)-1-
13 Lg 89 2-Ethyltoluene Ar
propenyl-
4-methoxy-3-hydroxyacetophe-
14 Lg 90 2-Methylstyrene Ar
none
3-methoxy-2,4,6-trimethyl- Benzene,1-methyl-2-(1-
15 Lg 91 Ar
phenol methylethyl)-
2-Propanone,1-(4-hydroxy-3-
16 Lg 92 Benzene, 2-propynyl Ar
methoxyphenyl)-
Phenol,2,6-dimethoxy-4-(2- Benzene,1-methyl-4-(1-
17 Lg 93 Ar
propenyl)- methylethenyl)-
3,5-Dimethoxy-4-
18 Lg 94 Benzene, 1-butynyl- Ar
hydroxybenzaldehyde
Phenol,2,6-dimethoxy-4-(2- Benzene,(1-methylene-2-
19 Lg 95 Ar
propenyl)- propenyl)
20 4-Ethyl-2-methoxyphenol Lg 96 Resorcinol Ar
1-ethyl-2,4,5-trimethyl-
21 1,2,4-Trimethoxybenzene Lg 97 Ar
benzene
Benzene, 1,3,5-trimethyl-
22 Phenol, 3,5-dimethoxy- Lg 98 Ar
2-(1,2-propadienyl)-
Benzene, (1-ethynyl-2-
23 Vanillyl ethyl ether Lg 99 Ar
methyl-1-propenyl)-
24 2,5-Dimethylfuran Ps 100 Mesitylene Ar
25 2(5H)-Furanone Ps 101 Benzene Ar
26 Furfural Ps 102 Benzene, 1-propynyl- Ar
Cyclopentene,1-(1-
27 Ps 103 Cinnamaldehyde Ar
methylethyl)-
28 2-Cyclopenten-1-one,2-methyl- Ps 104 Mesitaldehyde Ar
29 Furfuryl alcohol Ps 105 1-Indanone Ar
2-Furancarboxaldehyde,5-
30 Ps 106 Benzene, octyl- Ar
methyl-
286
Ty-
pe
31 2(5H)-Furanone,3-methyl- Ps 107 Phenol Ph
32 Furan, 2,4-dimethyl- Ps 108 2,5-Dimethylphenol Ph
33 2(3H)-Furanone,5-methyl- Ps 109 Phenol,3,5-dimethyl- Ph
34 2,5-Dimethyl furane Ps 110 Phenol,2-ethyl-5-methyl- Ph
35 Furan, 2-ethenyl- Ps 111 Phenol,4-ethyl-3-methyl- Ph
2H-Pyran-2-one,
36 Ps 112 Hydroquinone Ph
4-ethenyltetrahydro-
1,2-Benzenediol,4-
37 2H-Pyran-2-one Ps 113 Ph
methyl-
38 2-Cyclopenten-1-one,2-methyl- Ps 114 1,2-Benzenediol,4-ethyl- Ph
Phenol,4-(2-propen-1-
39 2-Acetylfuran Ps 115 Ph
yl)-
1,4-Benzene-
40 2(3H)-Furanone Ps 116 Ph
diol,dimethyl-
2H-Pyran-2-one,
41 Ps 117 1,2-Benzenediol,4-ethyl- Ph
5,6-dihydro-3,5,5-trimethyl-
42 Furfuryl alcohol Ps 118 Pyrogallol Ph
43 3-Methyl-2-cyclopenten-1-one Ps 119 Methylparaben Ph
2-Cyclopenten-1-one,2,3- 1,4-Benzenediol,2,3,5,6-
44 Ps 120 Ph
dimethyl- tetramethyl-
1,2-Benzenediol,4-
45 2-Cyclopentenone Ps 121 Ph
methyl-
46 Pyridine N 122 1,3-Benzenediol,4-ethyl- Ph
47 Benzonitrile N 123 Pyrogallol Ph
1,4-Benzenediol,2,3,5-
48 Oxazole,2-ethyl-4,5-dihydro- N 124 Ph
trimethyl-
49 1H-Pyrrole, 1-methyl- N 125 m-Tolualdehyde Ph
50 N-Methyl pyrrole N 126 3-Ethylphenol Ph
51 Pyrrole N 127 Phenol,2-methyl- Ph
2-Oxocyclohexanepropiononi-
52 N 128 Phenol,4-methyl- Ph
trile
54 3-Pyrazolidinone, 1,4-dimethyl- N 129 2,3-Dimethylphenol Ph
287
Ty-
pe
55 Perillartine N 130 2,3-Dimethylphenol Ph
56 3-Methylpyrazole N 131 3,5-Dimethylphenol Ph
57 Benzyl cyanide N 132 2,5-Dimethylphenol Ph
58 3-Methylindole N 133 3-Ethylphenol Ph
Benzeneethanamine,4-ethyl-2,5-
59 N 134 Pyrocatechol Ph
dimethoxy-a-methyl-
1,2-Benzenediol,3-
60 Pyridine, 3-methyl- N 135 Ph
methyl-
61 3-Methylpyrrole N 136 4-Methylcatechol Ph
Pyrazine, 2,3 dimethyl-5-(1-
62 N 137 4-Ethylresorcinol Ph
Propenyl)-
Benzeneethanamine, 1,4-Benzenediol,2,3,5-
63 N 138 Ph
2,5-dimethoxy- trimethyl-
64 Pyrazine, tetraethyl- N 139 Naphthalene PAHs
6-Hydroxy-1,2,3,4-
65 tetrahydroisoquinoline-1- N 140 2-Methylnaphthalene PAHs
carboxylic acid
1H-Indole, 4-(3-methyl-2-
66 N 141 2-Methylnaphthalene PAHs
butenyl)-
Naphthalene,1,2-
67 Undecane Lp 142 PAHs
dimethyl-
68 Tetradecane Lp 143 2,6-Dimethylnaphthalene PAHs
Naphthalene,1,6,7-
69 Nonadecane Lp 144 PAHs
trimethyl-
Naphthalene,1,6,7-
70 1-Hexacosene Lp 145 PAHs
trimethyl-
71 Heptadecane Lp 146 1-Methylnaphthalene PAHs
72 1-Eicosene Lp 147 Biphenyl PAHs
73 1-Hexacosene Lp 148 Naphthalene, 1-ethyl- PAHs
74 3-Eicosyne Lp 149 2-Naphthol PAHs
75 1-Docosene Lp 150 Dibenzofuran PAHs
76 1-Nonadecene Lp 151 Carbaryl PAHs
152 Fluorene PAHs
288
References
1. USDA Guidance for implementation of federal wildland fire management
policy; U.S. Department of Agriculture and US Department of the Interior:
Washington, DC, 2009.
2. Williams, A. P.; Allen, C. D.; Millar, C. I.; Swetnam, T. W.; Michaelsen, J.;
Still, C. J.; Leavitt, S. W. Forest responses to increasing aridity and warmth
in the southwestern United States. Proc. Natl. Acad. Sci. U.S.A. 2010, 107,
21289–21294.
3. Vila-Escale, M.; Vegas-Vilarrubia, T.; Prat, N. Release of polycyclic aromatic
compounds into a Mediterranean creek (Catalonia, NE Spain) after a forest
fire. Water Res. 2007, 41, 2171–2179.
4. Certini, G. Effects of fire on properties of forest soils: A review. Oecologia
2005, 143, 1–10.

5. Battle, J.; Golladay, S. W. Prescribed fire’s impact on water quality of
depressional wetlands in southwestern Georgia. Am. Midl. Nat. 2003, 150,
15–25.
6. Richter, D. D.; Markewitz, D.; Trumbore, S. E.; Wells, C. G. Rapid
accumulation and turnover of soil carbon in a re-establishing forest. Nature
1999, 400, 56–58.
7. Gonzalez-Perez, J. A.; Gonzalez-Vila, F. J.; Almendros, G.; Knicker, H.
The effect of fire on soil organic matter - a review. Environ Int. 2004, 30,
855–870.
8. Knicker, H. How does fire affect the nature and stability of soil organic
nitrogen and carbon? A review. Biogeochemistry 2007, 85, 91–118.
9. Harden, J. W.; Trumbore, S. E.; Stocks, B. J.; Hirsch, A.; Gower, S. T.;
O’Neill, K. P.; Kasischke, E. S. The role of fire in the boreal carbon budget.
Global Change Biol. 2000, 6, 174–184.
10. Hammes, K.; Schmidt, M. W. I.; Smernik, R. J.; Currie, L. A.; Ball, W. P.;
Nguyen, T. H.; Louchouarn, P.; Houel, S.; Gustafsson, O.; Elmquist, M.;
Cornelissen, G.; Skjemstad, J. O.; Masiello, C. A.; Song, J.; Peng, P.;
Mitra, S.; Dunn, J. C.; Hatcher, P. G.; Hockaday, W. C.; Smith, D. M.;
Hartkopf-Froeder, C.; Boehmer, A.; Luer, B.; Huebert, B. J.; Amelung, W.;
Brodowski, S.; Huang, L.; Zhang, W.; Gschwend, P. M.; Flores-Cervantes, D.
X.; Largeau, C.; Rouzaud, J. N.; Rumpel, C.; Guggenberger, G.; Kaiser, K.;
Rodionov, A.; Gonzalez-Vila, F. J.; Gonzalez-Perez, J. A.; de la Rosa, J. M.;
Manning, D. A. C.; Lopez-Capel, E.; Ding, L. Comparison of quantification
methods to measure fire-derived (black/elemental) carbon in soils and
sediments using reference materials from soil, water, sediment and the
atmosphere. Global Biogeochem. Cycles 2007, 21, 1–18.
11. Hockaday, W. C.; Grannas, A. M.; Kim, S.; Hatcher, P. G. The transformation
and mobility of charcoal in a fire-impacted watershed. Geochim.
Cosmochim. Acta 2007, 71, 3432–3445.
12. Hamer, U.; Marschner, B.; Brodowski, S.; Amelung, W. Interactive priming
of black carbon and glucose mineralisation. Org. Geochem. 2004, 35,
823–830.
289
13. Wardle, D. A.; Nilsson, M. C.; Zackrisson, O. Fire-derived charcoal causes
loss of forest humus. Science 2008, 320, 629–629.
14. Chow, A. T.; Lee, S. T.; O’Geen, A. T.; Orozco, T.; Beaudette, D.; Wong, P.
K.; Hernes, P. J.; Tate, K. W.; Dahlgren, R. A. Litter contributions to
dissolved organic matter and disinfection byproduct precursors in California
oak woodland watersheds. J. Environ. Qual. 2009, 38, 2334–2343.
research. Mutat. Res. Rev. Mutat. Res. 2007, 636, 178–242.
16. Wang, J. J.; Chow, A. T.; Sweeney, J. M.; Mazet, J. A. Trihalomethanes in
marine mammal aquaria: Occurrences, sources, and health risks. Water Res.
2014, 59, 219–228.
17. Sedlak, D. L.; von Gunten, U. The chlorine dilemma. Science 2011, 331,
42–43.
19. Karanfil, T.; Krasner, S. W.; Westerhoff, P.; Xie, Y. F. Disinfection
By-Products in Drinking Water: Occurrence, Formation, Health Effects,
and Control; Oxford University Press: New York, 2008; Vol. 995.
20. Chow, A. T.; O’Geen, A. T.; Dahlgren, R. A.; Diaz, F. J.; Wong, K. H.;
Wong, P. K. Reactivity of litter leachates from California oak woodlands
in the formation of disinfection by-products. J. Environ. Qual. 2011, 40,
1607–1616.
21. Wang, J. J.; Ng, T. W.; Zhang, Q.; Yang, X. B.; Dahlgren, R. A.; Chow, A.
T.; Wong, P. K. Technical note: Reactivity of C1 and C2 organohalogens
formation – from plant litter to bacteria. Biogeosciences 2012, 9, 3721–3727.
22. Beggs, K. M. H.; Summers, R. S. Character and Chlorine Reactivity of
dissolved organic matter from a mountain pine beetle impacted watershed.
23. Pellerin, B. A.; Hernes, P. J.; Saraceno, J.; Spencer, R. G. M.;
Bergamaschi, B. A. Microbial degradation of plant leachate alters lignin
phenols and trihalomethane precursors. J. Environ. Qual. 2010, 39,
946–954.
24. Reckhow, D. A.; Rees, P. L. S.; Bryan, D. Watershed sources of disinfection
byproduct precursors. Water Supply 2004, 4, 61–69.
25. Hernes, P. J.; Hedges, J. I. Tannin signatures of barks, needles, leaves, cones,
and wood at the molecular level. Geochim. Cosmochim. Acta 2004, 68,
1293–1307.
26. Friebel, B. A. Home range and habitat use of feral hogs (Sus scrofa) in
Congaree National Park. Clemson University, 2007.
27. Wohl, E.; Polvi, L. E.; Cadol, D. Wood distribution along streams draining
old-growth floodplain forests in Congaree National Park, South Carolina,
USA. Geomorphology 2011, 126, 108–120.
290
28. Battle, J.; Golladay, S. W. Prescribed fire’s impact on water quality of
depressional wetlands in southwestern Georgia. Am. Midl. Nat. 2003, 150,
15–25.
29. Song, J. Z.; Peng, P. A. Characterisation of black carbon materials by
pyrolysis-gas chromatography-mass spectrometry. J. Anal. Appl. Pyrol.
2010, 87, 129–137.
30. Chow, A. T.; Dahlgren, R. A.; Zhang, Q.; Wong, P. K. Relationships between
specific ultraviolet absorbance and trihalomethane precursors of different
carbon sources. J. Water Supply: Res. Technol. −AQUA 2008, 57, 471–480.
31. Dalrymple, R. M.; Carfagno, A. K.; Sharpless, C. M. Correlations between
dissolved organic matter optical properties and quantum yields of singlet
oxygen and hydrogen peroxide. Environ. Sci. Technol. 2010, 44,
5824–5829.
32. Zhou, J.; Wang, J. J.; Baudon, A.; Chow, A. T. Improved fluorescence
excitation-emission matrix regional integration to quantify spectra for
fluorescent dissolved organic matter. J. Environ. Qual. 2013, 42, 925–930.
33. Chen, W.; Westerhoff, P.; Leenheer, J. A.; Booksh, K. Fluorescence
excitation - emission matrix regional integration to quantify spectra for
dissolved organic matter. Environ. Sci. Technol. 2003, 37, 5701–5710.
34. Summers, R. S.; Hooper, S. M.; Shukairy, H. M.; Solarik, G.; Owen, D.
Assessing the DBP yield: Uniform formation conditions. J. Am. Water Works
Ass. 1996, 88, 80–93.
35. Cho, D.-H.; Kong, S.-H.; Oh, S.-G. Analysis of trihalomethanes in drinking
water using headspace-SPME technique with gas chromatography. Water
Res. 2003, 37, 402–408.
36. Kang, B. T.; Sajjapongse, A. Effect of Heating on Properties of Some Soils
from Southern Nigeria and Growth of Rice. Plant Soil 1980, 55, 85–95.
37. Giovannini, G.; Lucchesi, S. Modifications induced in soil physico-chemical
parameters by experimental fires at different intensities. Soil Sci. 1997, 162,
479–486.
38. Sharma, R. K.; Wooten, J. B.; Baliga, V. L.; Lin, X. H.; Chan, W. G.;
Hajaligol, M. R. Characterization of chars from pyrolysis of lignin. Fuel
2004, 83, 1469–1482.
39. Sharma, R. K.; Hajaligol, M. R. Effect of pyrolysis conditions on the
formation of polycyclic aromatic hydrocarbons (PAHs) from polyphenolic
compounds. J. Anal. Appl. Pyrol. 2003, 66, 123–144.
40. Oanh, N. T. K.; Ly, B. T.; Tipayarom, D.; Manandhar, B. R.; Prapat, P.;
Simpson, C. D.; Liu, L. J. S. Characterization of particulate matter emission
from open burning of rice straw. Atmos. Environ. 2011, 45, 493–502.
41. Kim, E. J.; Oh, J. E.; Chang, Y. S. Effects of forest fire on the level and
distribution of PCDD/Fs and PAHs in soil. Sci. Total Environ. 2003, 311,
177–189.
42. Smith, H. G.; Sheridan, G. J.; Lane, P. N. J.; Nyman, P.; Haydon, S. Wildfire
effects on water quality in forest catchments: A review with implications for
water supply. J. Hydrol. 2011, 396, 170–192.
291
43. Robinson, M. S.; Anthony, T. R.; Littau, S. R.; Herckes, P.; Nelson, X.;
Poplin, G. S.; Burgess, J. L. Occupational PAH exposures during prescribed
pile burns. Ann. Occup. Hyg. 2008, 52, 497–508.
44. Naeher, L. P.; Brauer, M.; Lipsett, M.; Zelikoff, J. T.; Simpson, C. D.;
Koenig, J. Q.; Smith, K. R. Woodsmoke health effects: A review. Inhal.
Toxicol. 2007, 19, 67–106.
45. Schmidl, C.; Bauer, H.; Dattler, A.; Hitzenberger, R.; Weissenboeck, G.;
Marr, I. L.; Puxbaum, H. Chemical characterisation of particle emissions
from burning leaves. Atmos. Environ. 2008, 42, 9070–9079.
46. Oanh, N. T. K.; Ly, B. T.; Tipayarom, D.; Manandhar, B. R.; Prapat, P.;
Simpson, C. D.; Liu, L. J. S. Characterization of particulate matter emission
from open burning of rice straw. Atmos. Environ. 2011, 45, 493–502.
47. Vergnoux, A.; Di Rocco, R.; Domeizel, M.; Guiliano, M.; Doumenq, P.;
Theraulaz, F. Effects of forest fires on water extractable organic matter

and humic substances from Mediterranean soils: UV-vis and fluorescence
spectroscopy approaches. Geoderma 2011, 160, 434–443.
48. Hongve, D. Production of dissolved organic carbon in forested catchments.
J. Hydrol. 1999, 224, 91–99.
49. Almendros, G.; Gonzalezvila, F. J.; Martin, F. Fire-induced transformation
of soil organic-matter from an oak forest - an experimental approach to the
effects of fire on humic substances. Soil Sci. 1990, 149, 158–168.
50. Almendros, G.; Gonzalezvila, F. J.; Martin, F.; Frund, R.; Ludemann, H. D.
Solid-state NMR-studies of fire-induced changes in the structure of humic
substances. Sci. Total. Environ. 1992, 118, 63–74.
51. Baldock, J. A.; Smernik, R. J. Chemical composition and bioavailability of
thermally, altered Pinus resinosa (Red Pine) wood. Org. Geochem. 2002,
33, 1093–1109.
52. Almendros, G.; Knicker, H.; Gonzalez-Vila, F. J. Rearrangement of carbon
and nitrogen forms in peat after progressive thermal oxidation as determined
by solid-state C-13 and N-15-NMR spectroscopy. Org. Geochem. 2003, 34,
1559–1568.
53. Vergnoux, A.; Dupuy, N.; Guiliano, M.; Vennetier, M.; Theraulaz, F.;
Doumenq, P. Fire impact on forest soils evaluated using near-infrared
spectroscopy and multivariate calibration. Talanta 2009, 80, 39–47.
54. Outcalt, K. W. The longleaf pine ecosystem of the South. Native Plants J.
2000, 1, 42–53.
55. Westerhoff, P.; Mash, H. Dissolved organic nitrogen in drinking water
supplies: a review. J. Water Supply: Res. Technol. − AQUA 2002, 51,
415–448.
56. Wang, J. J.; Dahlgren, R. A.; Ersan, M. S.; Karanfil, T.; Chow, A. T. Wildfire
alters terrestrial precursors of disinfection byproducts in drinking water.
Environ. Sci. Technol. 2015 Submitted for publication.
292
Chapter 16
Prescribed Fire Alters Dissolved

Organic Matter and Disinfection By-Product
Precursor in Forested Watersheds – Part II.
A Controlled Field Study
Kuo-Pei Tsai,1 Mary-Frances Rogers,1 Alex T. Chow,*,1

and Francisco Diaz2
1Baruch Institute of Coastal Ecology and Forest Science,
Clemson University, Georgetown, South Carolina 29442, United States
2Department of Soil Science and Geology, University of La Laguna,
Canary Islands, Spain

Forest detritus material is one of the major terrestrial sources

of dissolved organic matter (DOM) in source waters. There is
a health concern on DOM because it reacts with disinfectants
to form a variety of potentially carcinogenic disinfection
by-products (DBPs) during drinking water treatments.
Prescribed fire is a common forest management practice in
Southeastern US to reduce the risks of wildfire and beetle
infestation. However, this forest management practice
alters the composition and quantity of detritus materials on
forest floor, changing the DOM and DBP precursors exports
from those forested watersheds. In this book chapter, we
discussed a prescribed fire study conducted in three 20m x
20m experimental plots in Hobcaw Barony, Georgetown, South
Carolina. Litter and duff mixtures of field samples before
and after prescribed fire were collected for water extraction
experiments in the laboratory. In addition, water extracts
were further exposed to sunlight for 15 days in order to
understand the biogeochemical processes on the degradability
of DOM and DBP precursors. The concentrations of water
extractable organic carbon (WEOC) and water extractable total
nitrogen (WETN) from unburned detritus (60.3 g-WEOC/m2

and 1.9 g-WETN/m2) were significantly greater than those
from burned detritus (11.4 g-WEOC/m2 and 0.5 g-WETN/m2).
Importantly, the yield of DBP was significantly reduced from
3651 mg-THM/m2 to 484 mg-THM/m2 after fire. There
was no change in specific DBP formation in both burn and
unburned samples after sunlight exposure although significant
decreases in ultraviolet absorbance and fluorescence intensity
were observed. This field study showed that prescribed fire
could significantly decrease the production of dissolved organic
carbon (DOC) and dissolved total nitrogen (DTN) per unit area
in forest floor and potentially decrease the DBP formation in
water supply.
Introduction
It is well knownthat natural organic matter (NOM) is the major precursor
of disinfection by-products (DBPs) during the chlorination process in water
treatment systems (1, 2). It is estimated that NOM constitutes 50-90% of dissolved
organic carbon (DOC) in freshwater systems (3). Reducing the import of DOC
into natural waters used as drinking water sources can potentially minimize the
formation of DBP during the chlorination process. Water in forested watersheds
is a critical source of drinking water. Approximately 180 million people in
over 68,000 communities in the United States rely on forested lands to capture
and filter their drinking water (4). Large scale of natural disturbance in forest
could substantially influence water characteristics and consequently impact the
treatability downstream.
Several studies have reported negative effects of wildfire on the water quality
of forested watersheds (5–7). Prescribed fire treatment is usually implemented
as a low-cost-effective tool to reduce hazardous fuels in the forest and risks of
unwanted wildfires as well as recyclying nutrients back to the soil. Richter and
colleagues in their watershed scale study (8) indicated that periodic prescribed fire
within the Francis Marion National Forest in South Carolina is not likely to have
appreciable effects on the quality of stream rivers. Also, Arkle and Pilliod (9)
found that prescribed fire had no effects on stream habitats in Payette National
Forest, Idaho. Caldwell and colleagues (10) reported that carbon-nitrogen ratios
did not differ between streams affected by prescribed fire treatment and reference
streams within a grassland system on the Valles Caldera National Preserve, New
Mexico. It is noted that nutrients were measured quantitatively in those researches
to assess the effects of prescribed fire on the aquatic systems; however, little is
known regarding the effects of prescribed fire on the quantity and quality of DOM
exported from forest fuels.
Dead and down woody materials, litter, grasses, herbaceous plant materials,
and short shrubs are not only forest fuels but are also parts of terrestrial sources
of DOM in forested watersheds. According to Chow and colleagues (11), the
DOC extracted from California oak woodland litter is an important precursor in
294
forming DBPs and the reactivity of litter and leaf leachate is dependent on types
and decomposition stages of vegetation. When prescribed fire occurs, those forest
litter materials could be chemically oxidized and transformed and consequently
alter the quantity and quality of DOM exported from the forested watersheds.
In order to study effects of prescribed fire on water quality of forested
watersheds, it is important to understand the characteristics of DOC leaching
from burned litter and duff materials. Moreover, DOM in natural water may
undergo various chemical, physical, and biological reactions during water
conveyance in streams before entering water treatment facilities. Photochemical
reactions of DOM such as solar irradiation may cause oxidative degradation of
DOM and result in smaller and more labile organic carbon moieties (12, 13).
In fact, exposure of DOM to natural sunlight could alter the characteristics of
carbon exported from foliar litters and consequently affect DBP formation in
water disinfection processes (14, 15). The objectives of this field study were
to understand the influence of prescribed fire on DOM and DBP precursor
productions using detritus mixture burned in the field instead of using single
vegetation species burned in a temperature controlled oven, as in Part I of the
study. In addition, the leachates extracted from burned and unburned detritus
were further incubated under dark and light conditions for 15 days in order to
determine the influences of biogeochemical processes on the characteristics of
DOM and DBP precursors.

Prescribed Burn in a Managed Forest
The controlled field study was conducted at Hobcaw Barony, a 71-km2

wildlife refuges near Georgetown, South Carolina. The field site represents a
typical coastal plain forest with loblolly pine (Pinus taeda L.) and longleaf pine
(Pinus palustris P.). Within a mixed hardwood-pine stand that was scheduled to
undergo a growing season prescribed fire, three adjacent 20×20m experimental
plots (latitude 33°21′14 dpm, longitude 79°12′19″) were established for the
study. A field investigation on vegetation and fuel was conducted a month prior
to the prescribed burn. Down woody materials were evaluated using Brown’s
Planar Intersect Method (16, 17) to obtain average litter and duff depths for
each plot. Nine measurements were conducted in each plot. In addition, prior
to the execution of the growing season prescribed fire, a representative sample
containing detritus material from each 20m x 20m plot was collected in 17”x9”
quadrants with care as to avoid as much mineral soil collection as possible.
Samples were placed into labeled brown paper bags that were oven-dried and
stored for later analysis. On May 29, 2014, a prescribed fire classified as a
periodic growing season burn was conducted using a drip torch and a backing
fire technique. Characteristics of the prescribed fire were visually observed and
recorded according to the National Park Service Fire Monitoring Handbook (18).
The fire was naturally extinguished after approximately 3 hours.
295
Characterization of DOM and DBP Precursors
An hour prior to ignition and immediately after fire extintion, destructive

surface detritus materials samples including the mixture of litter and decomposed
duff without mineral soil were randomly collected from each plot (one composite
sample from 3 subsamples per plot). Sealed plastic bags were used to transport
the samples to the lab where they were weighted and placed into a drying oven at
50°C. After 72h samples were re-weighed and moisture content calculated.
Dried unburned and burned forest detritus were grounded and sieved through
a 2 mm screen and fraction < 2 mm was used for analysis. To simulate rainwater
flush on forest detritus and to collect the leachate extracted, 20 g of sieved
material was mixed with 200 ml Milli-Q water in a 250 ml Erlenmeyer flask and
placed on an orbital shaker operated at 250 rpm for 2h. Milli-Q water was used
because it generally extracts a greater amount of DOC comparing with other

water solutions containing salt (18), and also minimize any contamination of
carbon source. Extracts were filtered using 0.45 µm polyethersulfone membrane
filters (Supor-450, Pall Gelman Science) previously flushed three times with
Milli-Q water. Ultraviolet absorbance at 254 nm (UVA254), fluorescence, water
extractableorganic carbon (WEOC) and water extractable total nitrogen (WETN)
were measured in the filtered extracts. The ultraviolet absorbance scanning from
200 to 700 nm was measured using a Shimadzu UV-1800 in room temperature.
WEOC and WETN were measured using a Shimadzu TOC-VCSH/CSN analyzer.
Specific ultraviolet absorbance at 254 nm (SUVA254) was defined as carbon
normalized UVA254 (SUVA254 = UVA254/WEOC). Fluorescence scans were
conducted using a Shimadzu spectrofluorescence RF5301 with a 5-nm slit in
excitation and a 5-nm slit in emission. In excitation-emission matrix (EEM)
fluorescence spectroscopy, the matrixes were delineated and operationally defined
into five regions using consistent excitation (ex) and emission (em) wavelength
boundaries based on fluorescence of freshwaters (19): Region I, aromatic protein
I (ex: 200-250 nm; em: 280-330 nm); Region II, aromatic protein II (ex: 200-250
nm; em: 330-380 nm); Region III, fulvic acid-like (ex: 200-250 nm; em: 380-550
nm); Region IV, soluble microbial by-product-like (250 nm < ex <400 nm;
em: 280-380 nm); and Region V, humic acid-like (250 nm < ex < 400 nm; em:
380-550 nm).
The analysis of DBPs formation was based on uniform formation condition
(UFC) with sodium hypochlorite solution (Sigma-Aldrich) as a chlorination
reagent as described in Part I. UFC was selected because its reaction condition
is more closely related to conventional water treatment practice and the DBP
formation in the UFC test is similar to those in finished water. DBPs including
four trihalomethanes (THMs) (trichloromethane, dichlorobromomethane,
dibromochloromethane, tribromomethane), four haloacetonitriles (HANs)
(trichloro-acetonitrile, dichloro-acetonitrile, bromochloro-acetonitrile,
dibromo-acetonitrile), and chloral hydrate (CHD) were measured according to
USEPA method 551.1 using a gas chromatograph with a 63Ni electron capture
detector (Agilent 7890). The specific DBPs formation (i.e. specific THM
formation, specific HAN formation, and specific CHD formation) were calculated
by normalizing THM, HAN, and CHD concentrations with WEOC concentration
296
of water extracts. The measurement of specific DBP formation represents the
reactivity of carbon in forming DBP during chlorination.
Potential changes in organic matter chemical composition due to fire effects
were assessed using a pyrolyzer (CDS5500) connected to a GC–MS system
Agilent 7890 equipped with a fused silica capillary column HP5MS (30 m×250
μm×0.25 μm inner diameter). Approximately 2 mg of unburned and burned
samples were placed in tiny platinum capsules and pyrolysis was set to 700ºC.
Pyrolysis products were identified according to GC retention times, mass spectra
with reference to the Wiley/NIST libraries and published mass spectra of pyrolysis
products (20). Major identified compounds were classified as: aromatic (Ar),
lignin (Lg), lipid (Lp), polycyclid aromatic hydrocarbon (PAH), phenolic (Ph),
polysaccharide (Ps), and nitrogen compounds (N). Individual identified organic
compounds are included in the list shown in Part I of this Chapter.
Dark and Light Incubations

To simulate effects of biogeochemical processes on DOM and DBP precursors
during water convenyance from watershed to treatment facilities, 120 ml of diluted
water extracts from burned and unburned samples were placed into 200 ml custom-
made quartz tubes and exposed to sunlight. For each sample were used 2quartz
tubes, wrapping one with aluminum foil for a dark incubation and using the other
onefor sunlight incubation. All quartz tubes were placed in a wide-open area in
the Baruch Institute(about 1 km from the burned experimental plots), for 15 days
(October 7th to October 22ed, 2014). According to the National Estuarine Research
Reserve System sampling station from North Inlet Winyah Bay, SC (21), daily
maximum photosynthetically active radiation (PAR) values ranged from 1200 to
1600 mmol/m2/15 min (except for 400 mmol/m2/15 min on October 12th) and
temperature ranged from 12 to 30 ˚C during the entire experiment. After 15 days
incubation, samples were characterized for DOC, DTN, UVA, fluorescence EEM,
and DBP formation as described above.
Statistical Analysis
A paired t-test was used to determine differences in characteristics
betweenwater extracts from unburned and burned samples. The level of
significance was set to P < 0.05. All data were analyzed using SPSS.

Controlled Field Burns
A field investigation on vegetation and fuel was conducted a month prior
to the prescribed burn. The depth of litter and duff layers prior to fire in the
experimental plots was 3.6 ± 2.3 cm and 1.1 ± 0.2 cm, respectively. Fuel moisture
content was 11.6 ± 5.7% and top soil moisture content was 20.4 ± 13.1%. At
the start of the prescribed fire, cloud cover was 30% and the wind speed was 8
km/h out of the north as reported on the National Weather Service’s website (22).
297
Precipitation was zero and the air temperature was 30˚C with 63.5% of relative
humidity as obtained from the HOBO U30-Wi-Fi station at the Baruch Institute
(latitude 33°21’39, longitude 79°13’31). Approximately, the average flame height
ranged from 30 to 60 cm, and flame length and depth ranged from 15 to 30 cm.
The area after the prescribed burn had a mosaic burn pattern and was categorized
at light burn severity (23) with the percent burned area of the three plots visually
determined as 35 ± 5%. Mass of detritus before burn was 3.5 ± 0.7 kg/m2, being
reduced to 1.1 ± 0.4 kg/m2 after fire. Images of detritus material prior to, during,
and following the prescribed fire are shown in Figure 1.
Figure 1. Images prior to (a), during (b), and following (c) the prescribed fire in
Hobcaw Barony, SC.
Chemical Composition of Detritus Materials
Py-GCMS spectra from unburned and burned detritus materials are shown
in Figure 2. A total of 53 and 46 peaks were identified in the unburned and
burned sample respectively. Relative intensity of peaks usually decreased in
burned detritus compared to original unburned detritus pointing to the loss of a
high percentage of organic matter during burning. In addition, some of peaks
such as trimethoxybenzene (38.8 min) and dimethoxyacetophenone (43.6 min)
were missing in the burned samples (Figure 2). Meanwhile, an increase of peak
intensity for compounds such as naphthalene (27.0 min) and biphenyl (36.0
min) were observed in burned samples, although those peaks were very small
(Figure 2). Overall, Lg and Lp were decreased from 55 to 41% and 2 to < 1%
respectively, whereas Ar, Ph, and PAH increased from 12 to 18%, 24 to 30%,
and from 0.6 to 3%, respectively after fire. No obvious changes in N compounds
and Ps were observed, staying at about 3 and 4% respectively. Importantly,
the changes in relatively abundance of these fractions matched well with the
controlled laboratory burns (Part I), suggesting mixed and individual species may
not affect the results.
298
Figure 2. Py-GCMS Total Ion Chromatograms and relative abundance (pie

charts; percentage) of different types of pyrogenic products from unburned
and burned forest detritus material; aromatics (Ar); lignins (Lg); lipids (Lp);
polysaccharides (Ps); nitrogens (N); polycyclic aromatic hydrocarbons (PAH);
phenolics (Ph).
Yields of DOM and DBP Precursors
Results showed a significant reduction (p = 0.01) in WEOC after fire,

decreasing from 1,703 to 925 mg/L. On a dried mass basis, WEOC was decreased
from 17 to 9.3 mg-WEOC/g-detritus. In contrast, there was no significant changes
in WETN, C:N ratio, and SUVA between unburned and burned extracts (Table 1).
299
Table 1. WEOC, WETN and DBPs Formation of Forest Detritus Samples
before and after Burn; Average ± Standard Deviation; n = 3
Before burn After burn P value
Water Extractable Organic Matter (1:10 litter to water extraction)
WEOC (mg/L) 1,703 ± 271 925 ± 373 0.01
WETN (mg/L) 55 ± 15 43 ± 17 0.29
C:N (mol: mol) 38 ± 17 25 ± 0 0.15
SUVA 2.03 ± 0.16 1.98 ± 0.47 0.44
Disinfection By-Product Formation
THM (mmol/mol) 6.05 ± 1.15 4.11 ± 1.57 0.16
HAN (mmol/mol) 0.14 ± 0.05 0.22 ± 0.07 0.11

CHD (mmol/mol) 0.30 ± 0.05 0.29 ± 0.03 0.42
Fluorescence EEM of extracts of unburned and burned detritus materials with

the same dilution factor are shown in Figure 3. The fluorescence intensities of the
burned extracts were generally lower than that of unburned extracts, especially
in fulvic acid-like (region III) and aromatic protein-like regions (regions I and
II). However, the relative abundance of the five regions were no different when
integrating with the fluorescence regional integration (FRI) method (data shown
in Figure 5). Although changes in fluorescence regions were observed in Part
I, the area after the prescribed burn had a mosaic burn pattern with the percent
burned area of the three plots visually determined as 35 ± 5%. In fact, significant
portions of detritus materials, especially the lower layer of the detritus malterials,
were partically burned in this low intensity fire. DOM from this material may have
influences on the overall characteristcs of DOM in extracts.
Figure 3. Fluroresence EEM of water extracts from (a) unburned and (b) burned
detritus materials.
300
In terms of DBP formation, a lower reactivity in forming THM but a higher
reactivity in forming HANs were observed after fire although those differences
were statistically no significant. The reactivity of CHD was similar before and
after burn. Therefore all the analyses showed that low intensity prescribed fire
only reduced the quantity of WEOC but the chemistry, in terms of C:N ratio,
optical properties including both UV/VIS and fluorescence measurements, and
DBP formation, was no changed.
Biogeochemical Processes on DOM and DBP Precursors

Incubation experiment showed that quantity and quality of DOM in
detritus leachate could be altered through biogeochemical process during water
conveyance. Regardless of the fire effects, WEOC and WETN after both dark and
sunlight incubations were lower than in the original samples (Figure 4), suggesting
DOM in these extracts was highly degradable. These results were concomitant
with the results showing that WEOC and WETN in burned detritus leachate
were lower than in unburned detritus leachate (Figures 4a and 4b). Consider
burned samples as an example, there were 76 and 59% decreases in WEOC and
WETN respectively, in dark incubation, suggesting microbial processes possibly
consumed these degradable organic matters. For the sunlight treatment, decreases
of WEOC and WETN were even higher (78 and 72% respectively).
Figure 4. WEOC (a), WETN (b), C:N ratio (c), and SUVA (d) of the leachate
extracted from unburned and burned detritus under 15-days dark and sunlight
incubation; average ± standard deviation; n = 3.
301
The C:N ratios of unburned and burned extracts after 15-days dark incubation
were significantly lower than the values of original samples and sunlight
incubation (Figure 4c), with an average of 22 and 14, respectively. On the other
hands, the SUVA of dark incubation were highest among other treatments (Figure
4d). The SUVA values of unburn and burned extracts increased to 3.01 and 2.73
L/mg/m, respectively. Results might suggest microbial processes dominated in
dark incubation and consumed degradable organic carbon, causing an increase
in the proportion of relatively recalcitrant aromatic carbon in water. Lower C:N
ratios and SUVA were observed after sunlight incubation when comparing to
original samples (Figures 4c and 4d). In addition, an obvious decrease of humic
acid-like fraction (Region V) in both unburned and burned extracts was observed
after 15-day sunlight exposure (Figure 5). It has been demonstrated that sunlight
could cause photobleaching of DOM is an important role in DOM characteristics
in nature waters (24, 25).
Figure 5. Percentage disturbtions of the five EEM regions of water extracts

before and after prescribed fire and sunlight treatments. The five regions are:
I (aromatic protein-like), II (aromatic protein-like II), III (Fulvic acid-like), IV
(Microbial by-product-like), and V (Humic Acid-like).
Regardless of unburned and burned detritus leachate, results showed no

significant changes on the THM formation after 15-day incubation (Figure 6a).
SUVA254 has been used as a measurement of DOC propensity to form THM and
is generally proportional to specific THM formation (26, 27). However, this
302
study found that although sunlight exposure significantly decreased SUVA254
values, the reactivity of DOC in forming THM was not significantly changed.
Results also showed no significant changes on HAN formation after incubation,
which suggested that the reactivity of DOC to form HAN was not sensitive to
biogeochemical processes in such short time frame (Figure 6b). In contrast, an
increase of CHD formation ocurred after sunlight exposure as shown in Figure
6c, which is consistent with the previous studies (15).
Figure 6. Specific formation potential of THMs (a), HANs (b), and CHD (c) of
the leachate extracted from unburned and burned detritus under 15-days dark
and sunlight incubation; average ± standard deviation; n = 3.
Implications
Quantities of forest detritus, WEOC, WETN, and DBPs expressed per unit
area in unburned and burned forest floor were summarized in Table 2. Prescribed
fire consumed and oxidized detritus materials, reducing significantly (p = 0.009)
the biomass per unit area (~ 65% reduction). In addition, WEOC and WETN
productions per unit biomass also reduced. Unburned detritus yielded 17 ± 2
g-WEOC/kg-detritus and 0.6 ± 0.2 g-WETN/kg-detritus, whereas burned detritus
produced 9 ± 4 g-WEOC/kg-detritus and 0.4 ± 0.1 g-WETN/kg-detritus. Therefore
there was a 81 and 74% reductions on WEOC and WETN per unit area, also
resulting in a significant decrease in DBP yield per unit area. As summarized
in Table 1, THM, HAN, and CHD yields were decreased in 87, 71, and 81%,
respectively.
303
Table 2. Yields of Detritus, WEOC, WETN, and DBPs before and after
Prescribed Burn in the Experimental Plots
Before burn After burn Reduction (%) P value
Detritus (kg/m2) 3.5 ± 0.7 1.1 ± 0.4 65 0.009
WEOC (g/m2) 60.3 ± 20.8 11.4 ± 8.5 81 0.01
WETN (g/m2) 1.9 ± 0.6 0.5 ± 0.4 74 0.06
THM (mg/m2) 3651 ± 682 484 ± 385 87 0.001
HAN (mg/m2) 79 ± 6 23 ± 12 71 0.01
CHD (mg/m2) 238 ± 56 46 ± 36 81 0.01
Both the controlled laboratory in Part I and controlled field studies in Part II
(this chapter) demonstrated that prescribed forest fire could reduce DOM and DBP
precursors in the detritus layers of forested watersheds in the southeastern US.
Although our controlled laboratory study suggested DOM from burned materials
could have a higher reactivity in forming N-DBPs during chlorination, significant
loss in biomass and a reduction of extractability still result an overall reduction of
DBP yield per unit area in a forested watershed. The results suggest that prescribed
fire could be an effective watershed management controlling DBP precursors in
source water.
Acknowledgments
Portions of this study were also supported by NIFA/USDA under project
number SC1700489 and SCN-2013-02784, Joint Fire Science Program
14-1-06-19, and NSF Rapid grant 1264579 as presented in technical contribution
number 6358 of the Clemson University Experiment Station.
References
1. Beggs, K. M. H.; Summers, R. S. Character and Chlorine Reactivity of
Dissolved Organic Matter from a Mountain Pine Beetle Impacted Watershed.
2. Gopal, K.; Tripathy, S. S.; Bersillon, J. L.; Dubey, S. P. Chlorination
byproducts, their toxicodynamics and removal from drinking water. J.
Hazard. Mater. 2007, 140, 1–6.
3. Uyguner-Demirel, C. S.; Bekbolet, M. Significance of analytical parameters
for the understanding of natural organic matter in relation to photocatalytic
oxidation. Chemosphere 2011, 84, 1009–1031.
4. United States Department of Agriculture. Some Facts About Water;
http:// www.fs.fed.us/managing-land/national-forests-grasslands/water-facts
(accessed Dec 14, 2014).
304
5. Brass, J. A.; Ambrosia, V. G.; Riggan, P. J.; Sebesta, P. D. Consequences
of fire on aquatic nitrate and phosphate dynamics in Yellowstone National
Park. In Ecological Implications of Fire in Greater Yellowstone; Greenlee,
J. M., Ed.; International Association of Wildland Fire: Fairfield, WA, 1996,
pp 53−57.
6. Riggan, P. J.; Lockwood, R. N.; Jacks, P. M.; Colver, C. G.; Weirich, F.;
DeBano, L. F.; Brass, J. A. Effects of fire severity on nitrate mobilization in
watersheds subject to chronic atmospheric deposition. Environ. Sci. Technol.
1994, 28, 369–375.
7. Emelko, M. B.; Silins, U.; Bladon, K. D.; Stone, M. Implications of
land disturbance on drinking water treatability in a changing climate:
Demonstrating the need for “source water supply and protection” strategies.
Water Res. 2011, 45, 461–472.
8. Richter, D. D.; Ralston, C. W.; Harms, W. R. Prescribed fire: effects on water

and forest nutrient cycling. Science 1982, 215, 661–663.
9. Arkle, R. S.; Pilliod, D. S. Prescribed fires as ecological surrogates for
wildfires: A stream and riparian perspective. For. Ecol. Manage. 2010,
893–903.
10. Caldwell, C. A.; Jacobi, G. Z.; Anderson, M. C.; Parmenter, R. R.;
McGann, J.; Gould, W. R.; DuBey, R.; Jacobi, M. D. Prescribed-fire effects
on an aquatic community of a southwest montane grassland system. N. Am.
J. Fish Manage. 2013, 33, 1049–1062.
11. Chow, A. T.; Lee, S. T.; O’Geen, A. T.; Orozco, T.; Beaudette, D.; Wong, P.
K.; Hernes, P. J.; Tate, K. W.; Dahlgren, R. A. Litter contributions to
dissolved organic matter and disinfection byproduct precursors in California
oak woodland watersheds. J. Environ. Qual. 2009, 38, 2334–2343.
12. Sharpless, C. M.; Blough, N. V. The importance of charge-transfer
interactions in determining chromophoric dissolved organic matter (CDOM)
optical and photochemical properties. Environ. Sci.: Processes Impacts
2014, 16, 654–671.
13. Méndez-Díaz, J. D.; Shimabuku, K. K.; Ma, J.; Enumah, Z. O.; Pignatello, J.
J.; Mitch, W. A.; Dodd, M. C. Sunlight-Driven Photochemical Halogenation
of Dissolved Organic Matter in Seawater: A Natural Abiotic Source of
Organobromine and Organoiodine. Environ. Sci. Technol. 2014, 48,
7418–7427.
14. Chow, A. T.; Dahlgren, R. A.; Xhang, Q.; Wong, P. K. Relationships between
specific ultraviolet absorbance and trihalomethane precursors of different
carbon sources. J. Water Supply: Res. Technol.−AQUA 2008, 57, 471–480.
15. Chow, A. T.; Diaz, F. J.; Wong, K. H.; O’Geen, A. T.; Dahlgren, R. A.;
Wong, P. K. Photochemical and bacterial transformations of disinfection
byproduct precursors in water. J. Environ. Qual. 2013, 42, 1589–1595.
16. Brown, J. K. Handbook for inventorying downed woody material; Gen. Tech.
Rep. INT-16; U.S. Department of Agriculture, Forest Service, Intermountain
Forest and Range Experiment Station: Ogden, UT, 1974; pp 1−24.
17. USDI National Park Service. Fire Monitoring Handbook; Fire Management
Program Center, National Interagency Fire Center: Boise, ID. 2003, 274.
305
18. Chow, A. T.; Tanji, K. K.; Gao, S. Production of dissolved organic carbon
(DOC) and trihalomethane (THM) precursor from peat soils. Water Res.
2003, 37, 4475–4485.
19. Zhou, J.; Wang, J. J.; Baudon, A.; Chow, A. T. Improved fluorescence
excitation-emission matrix regional integration to quantify spectra for
fluorescent dissolved organic matter. J. Environ. Qual. 2013, 49, 925–930.
20. Song, J. Z.; Peng, P. A. Characterisation of black carbon materials by
pyrolysis-gas chromatography-mass spectrometry. J. Anal. Appl. Pyrolysis
2010, 87, 129–137.
21. National Estuarine Research Reserve System Home Page; http://
cdmo.baruch.sc.edu/get/export.cfm (accessed Dec 14, 2014).
22. National Weather Service Home Page; http://www.weather.gov (accessed
Dec 14, 2014).
23. Keeley, J. E. Fire intensity, fire severity and burn severity: a brief review and
suggested usage. Int. J. Wildland Fire 2009, 18, 116–126.
24. Brinkmann, T.; Sartorius, D.; Frimmel, F. H. Photobleaching of humic rich
dissolved organic matter. Aquat. Sci. 2003, 65, 415–424.
25. Larson, J. H.; Frost, P. C.; Lodge, D. M.; Lamberti, G. A. Photodegradation
of dissolved organic matter in forested streams of the northern Great Lakes
region. J. N. Am. Benthol. Soc. 2007, 26, 416–425.
26. Chow, A. T.; Gao, S.; Dahlgren, R. A. Physical and chemical fraction of
dissolved organic matter and trihalomethane precursors: a review. J. Water
Supply: Res. Technol.−AQUA 2005, 54, 475–507.
27. Xie, Y. Disinfection Byproducts in Drinking Water: Formation, Analysis and
Control; Lewis Publishers: Boca Raton, FL. 2004.
306
Chapter 17
Control of Halogenated N-DBP

Precursors Using Traditional and Advanced
Drinking Water Treatment Processes:
A Pilot-Scale Study in China’s Lake Taihu
Wenhai Chu,* Naiyun Gao, Yang Deng, and Xin Li
State Key Laboratory of Pollution Control and Resources Reuse,

College of Environmental Science and Engineering, Tongji University,
Shanghai, 200092, China
*E-mail: [email protected]; [email protected].
Tel: +86 21 65982691. Fax: +86 21 65986313.
Haloacetamides (HAcAms), an emerging class of nitrogen-

based disinfection by-products (N-DBPs), have been frequently
identified in drinking waters. However, there is a limited
understanding on the performance of different treatment
technologies in the control of HAcAms. The objective
of this study was to evaluate the potential of traditional
and advanced treatment technologies, including three
pre-treatment processes (i.e., powdered activated carbon [PAC]
adsorption, KMnO4 oxidation, and biological contact oxidation
[BCO]), two combined conventional treatment methods (i.e.
coagulation - inclined plate sedimentation [IPS]-filtration,
and coagulation-dissolved air flotation [DAF]-filtration), and
an advanced processes (i.e. integrated ozone and biological
activated carbon [O3-BAC] treatment), for removing the
precursors of HAcAms while minimizing the formation of
other typical N-DBPs in water. Among the three pre-treatment
processes, PAC adsorption could effectively remove the
precursors of chloroform (CF) (42.7%), dichloroacetonitrile
(DCAN) (28.6%), dichloroacetamide (DCAcAm) (27.2%) and
trichloronitromethane (TCNM) (35.7%), advantageous over
KMnO4 oxidation and/or BCO process. In contrast, the removal
efficiency of dissolved organic carbon (DOC) by the BCO

process (76.5%) was superior to that by PAC adsorption (69.9%)
and KMnO4 oxidation (61.4%). However, BCO increased
the dissolved organic nitrogen (DON) concentration, thereby
leading to the formation of more N-DBPs during the subsequent
chlorination. Soluble microbial products including numerous
DON compounds produced as a result of the BCO treatment
were observed to play an essential role in the DCAcAm
formation. Between the conventional processes, the removal of
algae, DON, DOC and UV254 by the coagulation-DAF-filtration
was better than the coagulation-IPS-filtration. On the average,
the former achieved the removal of 53% DOC, 53% DON and
31% UV254, while the latter only removed 47% DOC, 31% DON
and 27% UV254. Additionally, the coagulation-IPS-filtration
removed less low molecular weight organic molecules than

the coagulation-DAF-filtration process. Consequently, the
concentrations of CF, DCAcAm and DCAN formed from the
coagulation-DAF-filtration treated water reached 13, 1.5 and
4.7 μg/L during chlorination, respectively, which were lower
than those from chlorination of the coagulation-IPS-filtration
treated water (17 μg/L CF, 2.9 μg/L DCAcAm and 6.3 μg/L
DCAN). Among the advanced treatment processes, O3-BAC
significantly improved the removal of turbidity, DOC, UV254,
NH+4-N, and DON by 98–99%, 58–72%, 31–53%, 16–93%
and 35–74%, respectively, and enhanced the removal efficiency
of the DBP precursors. However, this option was almost
ineffective in removing TCNM and DCAcAm precursors.
Ozonation alone could not substantially reduce the DCAcAm
formation, and increased the TCNM formation potential (FP).
However, it chemically altered the molecular structures of the
precursors and increased the biodegradability of N-containing
organic compounds. Consequently, the subsequent BAC
filtration dramatically reduced the formation of the both TCNM
and DCAcAm, thus highlighting a synergistic effect of O3 and
BAC. Additionally, O3-BAC was effective at controlling the
formation of total organic halogen, which is recognized as an
indicator of the formation of unidentified DBPs. Of note, more
N-DBP precursors entered into the post-BAC water without
pre-ozonation, leading to the formation of more N-DBPs
during chlorination, compared with a control group with the
pre-ozonation was continuously operated. Moreover, higher
DBP FP was observed in the effluent of the BAC filter without
pre-ozonation than the FP in the influent of the BAC filter.
Therefore, while the intermittent operation of pre-ozonation
may have cost and other operational benefits (bromate control),
these may be outweighed against the increased N-DBP
formation and potential N-DBP associated health risks.
308
1. Introduction
Water resource shortages and growing water demands have spurred utilities to
exploit the source waters impacted by wastewater effluents and/or algal blooming.
Such source waters are typically characterized by a higher level of dissolved
organic nitrogen (DON). During chlorination or chloramination, the N-containing
organic matters in water can react with the disinfectant to form halogenated
nitrogenous disinfection by-products (N-DBPs), such as haloacetamides
(HAcAms), halonitromethanes (HNMs) and haloacetonitriles (HANs) (1), which
represent an emerging concern due to their cytotoxicity and genotoxicity (2–4).
In particular, HAcAms, an emerging class of halogenated N-DBPs that have been
found in tap waters (5), exhibited much higher genotoxicity and cytotoxicity
than many carbonaceous DBPs (C-DBPs) (e.g., trihalomethanes [THMs] and
haloacetic acids [HAAs]) (6).

N-DBPs are generally present in drinking water at lower concentrations than
THMs and HAAs. One of the main reasons is that dissolved organic nitrogen
DON, believed to be the main precursor of N-DBPs, is typically less than dissolved
organic carbon (DOC), the important precursor of C-DBPs, in source water (e.g.,
THMs and HAAs) (7–12). However, algal blooms fed by industrial wastewater
effluent and fertilizer runoff can cause a dramatic increase in DON in source waters
(8, 13) and N-DBPs during chlorination. Therefore, special attention needs to be
paid to the control of N-DBPs in the algal-rich source waters.
An effective strategy to control the formation of DBPs is to remove
their precursors prior to chlorination. Generally, HAcAm and some other
DBP precursors are characterized by low molecular weight (MW), low
hydrophobicity, and poor removal efficiencies in a conventional DWTP using
coagulation-sedimentation-filtration (14, 15), especially in polluted source waters
characterized with high concentrations of algae [including both prokaryotic
microalgae (blue-green algae) and other eukaryotic algae], NH+4-N and
organic matter content (16). Recently, certain pre-treatment processes (e.g.,
powdered activated carbon [PAC] adsorption, KMnO4 oxidation, and biological
contact oxidation [BCO]), conventional processes (coagulation-inclined
plate sedimentation [IPS]-filtration, and coagulation-dissolved air flotation
[DAF]-filtration), and an advanced processes (e.g., integrated ozone and
biological activated carbon [O3–BAC] treatment), are increasingly applied to
improve the removal of algae, NH+4-N and organic matters in China. However,
the information regarding the performance of these treatment processes on the
control of N-DBPs is highly limited.
The objective of this chapter was to evaluate the formation of halogenated
N-DBPs and the removal of their precursors in algae-rich waters during the
following water treatments, including three pre-treatment processes (i.e.,
PAC adsorption, KMnO4 oxidation, and BCO treatment), two conventional
processes (i.e., coagulation-IPS-filtration, and coagulation- DAF-filtration), and
an advanced processes (i.e., O3–BAC treatment). Results are of importance
for optimization of the treatment design and improvement of water treatment
efficiencies in the treatment facilities with DON-rich waters.
309
2. Pilot-Plant Process Flow
The water samples used was collected from Lake Taihu, the third largest
freshwater lake in China. Lake Taihu is a major potable water source of many
cities and towns in Eastern China including Shanghai, Suzhou, and Wuxi. In
recent years, the combined effects of nutrient enrichment and industrial pollution
frequently degrade the water quality of Lake Taihu, and cause a fouling smell in
the finished water from conventional treatment facilities. This pushes local WTPs
to employ pre-treatment processes and O3-BAC advanced process for further
improvement of water quality. To test the additional treatments, a pilot plant (1
m3/h) was established in a conventional drinking water treatment plant (DWTP)
near Lake Taihu. A flowchart of the pilot plant process is shown in Figure 1.
Figure 1. Flowchart of the pilotscale treatment process.
2.1. Pretreatment
(i) PAC adsorption: 20 mg/L PAC was dosed to a raw water tank from a
hydrated PAC tank by a metering pump. The PAC was mixed in the raw water tank,
and was continuously transported to a coagulation tank (three-stage flocculation
plant). (ii) KMnO4 oxidation: 1.0 mg/L KMnO4 was added to the raw water tank.
The dosages of PAC and KMnO4 were selected based on our previous results of
jar tests and economic considerations. (iii) BCO: The BCO reactor was composed
of three flow-through columns (3.0 m length, 0.38 m diameter, 0.15 m elevation
difference and 2 h residence time), as shown in Figure 2. In these columns, the filler
vinylon silk was fixed firmly on plastic rings, and distributed evenly in the water,
providing the sites for biological films. The raw water and air were pumped into
the column bottom (gas-water ratio = 1:2), and the treated water overflowed from
the top of column. Before the BCO treatment, one month was required to allow
formation of the biomembrane by seeding Lake Taihu raw water. After biofilm
colonization, plentiful brown biological films were found in the filler, and the
removal rates of NH4+-N-N and DOC in the BCO reactors ranged within 75–85%
and 30–40%, respectively. The influent to the BCO reactor was from the raw water
tank and the effluent of the BCO reactor was transported to the ensuing coagulation
tank (three-stage flocculation plant) continuously (17).
310
Figure 2. The pilot plant of BCO process (1- Influent pipe; 2- Effluent pipe; Air
inflow pipe; 4- Discharge pipe; 5- Filler; 6- Water and air distributing device).
2.2. Conventional Process (Coagulation-IPS-Filtration versus

Coagulation-DAF-Filtration)
The coagulation process was performed in a three-stage flocculation plant (1.2
m length, 0.4 m width, and 0.5 m height). The average velocity gradient (G) values
in the first, second and third stages were set as 45, 24 and 11 s-1, respectively, to
match the operational practice requirements of the local WTP. The IPS process was
carried out in an inclined sedimentation tank (1.5 m length, 0.7 m width, and 1.4
m height). The DAF system consisted of a DAF tank with 0.7 m diameter and 1.0
m height, a dissolved air vessel and an air compressor. A scum board was installed
on the top of DAF tank to remove supernatant solids. Air was introduced into the
dissolved air vessel by an air compressor intermittently. The operating pressure
was 0.24–0.34 MPa. Part of the DAF effluent was recycled continuously back to
the inlet of the DAF tank at a recycle ratio of 10–15%. Filtration was achieved by
a filter column with a diameter of 0.3 m. The filter medium was quartz sand with
a grain diameter of 0.7–1.0 mm. The depth of the filter material was 0.9 m and the
filtration rate was 7 m/h. The details of the pilot scale conventional treatment was
described in detail elsewhere (18).
2.3. Advanced Treatment Process (O3–BAC)

Ozonation was carried out in an ozone contact chamber (4.5 m height, and
0.4 m diameter) operated in a concurrent flow mode. Ozone was generated from
air using an ozone generator (HMY-F, Huangming, China) and continuously
introduced into the water in the form of small bubbles through a porous titanium
plate. Dissolved ozone dosage was 1.5–2.0 mg/L, and the contact time was
30 min. Subsequent to ozonation, the water was fed into a granular activated
carbon (GAC) filter (4.0 m height, and 0.3 m diameter) to perform the BAC
treatment tests. The empty bed contact time and filtration rate of the BAC column
311
were maintained at 15 min and 7.0 m/h, respectively. The bed depth, effective
size, uniformity coefficient, iodine value, methylene blue value, and density
of the filled GAC were 1.6 m, 0.55–0.75 mm, <1.9, 1000 mg/g, 266 mg/g and
450 g/L, respectively. Results from the BAC treatment tests with and without
pre-ozonation were compared (19).
2.4. Disinfection Process
In most cases, disinfection was carried out by adding 5 mg/L of chlorine (as
Cl2) to the intake of a clean water tank using a metering pump. The clean water
tank was mixed intensively by mechanical agitation. The disinfectant residuals
were quenched with ascorbic acid, and glacial acetic acid was injected to achieve
pH 5.0 ± 0.2 for the subsequent N-DBPs analysis (17–19). The DBP formation
potential (FP) test was conducted with 48 h of sample collection and involved
chlorination for a further 24 h using free chlorine, as described in detail in previous
papers (14).
3. Pretreatment Processes
3.1. Water Parameters
There were not any significant difference (p=0.42, 0.27, 0.55, corresponding
to BCO, PAC adsorption, KMnO4 oxidation, respectively) on the removal of
turbidity among the three pre-treatment processes coupled with subsequent
conventional treatment process. The removal effect of turbidity in the filtered
water by the three different pretreatment processes (<0.5 NTU) was all better
than that by conventional treatment process alone (approx 0.8 NTU). Moreover,
the algae removal by the three different pretreatment processes was also all
significantly better than the conventional treatment process alone; the PAC
adsorption, KMnO4 oxidation and BCO processes increased the average algal
removal efficiency of the subsequent conventional treatment process from 90% to
97.5%, 98.2% and 99.6%, respectively. These findings confirmed that the three
pretreatment process flows were operated as designed and expected. Additionally,
compared with the PAC adsorption (<10%) and KMnO4 oxidation (<20%), the
BCO pretreatment process achieved a dramatically higher overall removal rate
of NH4+-N (> 60%).
From Table 1, PAC adsorption and KMnO4 oxidation enhanced the average
DON removal rate of the conventional water treatment process from 34.0% to
45.6% and 42.6% (Table 1). However, the BCO pretreatment process increased
the DON concentration of the raw water by 5.1%, likely because the BCO
process produced numerous metabolic by-products in the form of nitrogenous
organic matters. It is known that a range of soluble microbial products (SMPs)
are produced by microorganisms in the BCO process and can be found in BCO
effluent (20). SMP is a pool of complex organic matters including many DON
compounds such as proteins, nucleic acids and amino acids (21–23).
312
Table 1. DON Removal in Different Treatment Trains
DON SUVA
Treatment DON accumulative removal rate (%)
(mg/L) (L/(mg·m))
trains
a b c d a b c d a b c d
A 0.53 - 0.41 0.35 0.0 - 22.6 34.0 2.46 - 2.23 2.09
B 0.57 0.49 0.37 0.31 0.0 14.0 35.1 45.6 2.45 1.75 1.45 1.59
C 0.61 0.52 0.39 0.35 0.0 14.8 36.1 42.6 2.48 1.94 1.98 1.66
D 0.59 0.62 0.53 0.51 0.0 -5.1 10.2 13.6 2.29 1.96 1.82 1.68
Sampling point: a: Original Lake Taihu raw water; b: After pretreatment; c: After coagulation-sedimentation; d: After filtration. Treatment trains: A:
Conventional water treatment process only; B: PAC adsorption + conventional water treatment process; C: KMnO4 oxidation + conventional water treatment
process; D: BCO + conventional water treatment process.
313

Figure 3. Fluorescence EEM spectra of Lake Taihu water. (TLW DOC = 4.2
mg/L, pH = 7.0, T = 23 °C) A: Original raw water; B: After BCO pretreatment;
C: After coagulation-sedimentation; D: After filtration.
EEM spectra were used to characterize the Lake Taihu water natural organic
matters (NOM), as shown in Figure 3. According to the fluorescence regional
integration (FRI) method developed by Chen and colleagues (21), the EEM spectra
were operationally summarized into five regions: five regions in the EEM spectra
based on the fluorescence of model compounds: aromatic proteins (regions I and
II, λex < 250 nm, λem < 380 nm), fulvic acid-like (region III, λex < 250 nm, λem >
380 nm), SMP-like, including tyrosine-, tryptophan-, and protein-like components
(region IV, λex > 250 nm, λem < 380 nm), and humic acid-like (region V, λex >
250 nm, λem > 380 nm) regions.
As shown in Figure 3A, NOM in raw water had its intensest peak (496 mV)
and second highest peak (214 mV) in the humic region. However, when the raw
water was pretreated by the BCO process, the NOM showed the intensest peak
(804 mV) at Ex/Em of 275/320 nm, which fell within the SMP-like region (λex
> 250 nm, λem < 380 nm), and the second-highest peak (771 mV) at Ex/Em of
240/335 nm within the aromatic protein region (λex < 250 nm, λem < 380 nm)
(Figure 3B). This confirmed previous speculation that SMP containing numerous
DON compounds possibly weakens the DON removal performance of the BCO
process. Moreover, SMPs cannot be removed completely by conventional
treatment processes (coagulation-sedimentation-filtration) as shown in Figure
3(C) and (D).
314
3.2. DBPs
CF and DCAN were the most abundant species in most of the water samples
(i.e. there were relatively few brominated analogues) (24). In addition, TCAN and
TCAcAm concentrations in most of the water samples were below our detection
limits. Therefore, the study mainly focused on the chlorinated DBPs (CF,
DCAN, DCAcAm, and TCNM). The concentrations of CF, DCAN, DCAcAm
and TCNM measured after different pretreatment processes (PAC adsorption,
KMnO4 oxidation and BCO) and subsequent conventional treatment processes
are summarized in Figure 4.
Figure 4. The concentrations of CF, DCAN, DCAcAm and TCNM in chlorinated

water. A: Conventional water treatment process only; B: PAC adsorption +
conventional water treatment process; C: KMnO4 oxidation + conventional
water treatment process; D: BCO + conventional water treatment process. The
error bars represent the standard deviation of replicate measurements (n = 3).
As shown in Figure 4, CF in water treated by conventional treatment

processes only reached its maximum value of 16.7 μg/L, which was substantially
greater than that obtained by different pretreatment and subsequent conventional
treatment processes. This finding indicated that the pretreatment processes
could reduce the formation of CF effectively. It is worthwhile to note that mean
N-DBP concentrations (DCAN: 8.6 μg/L, DCAcAm: 5.6 μg/L, TCNM: 1.6
μg/L) in chlorinated water treated by BCO process were significantly greater
than from those of the water treated by conventional treatment processes alone
(DCAN: 6.0 μg/L, DCAcAm: 2.6 μg/L, TCNM: 1.2 μg/L), and pretreated by
PAC adsorption (DCAN: 3.6 μg/L, DCAcAm: 2.0 μg/L, TCNM: 0.2 μg/L)
or by KMnO4 oxidation (DCAN: 4.9 μg/L, DCAcAm: 2.5 μg/L, TCNM: 0.2
μg/L). Additionally, less CF and N-DBPs (DCAN, DCAcAm and TCNM)
were formed in chlorinated water post-PAC. There are two possible reasons for
315
this finding: (i) PAC adsorption can remove these DBP precursors effectively.
(ii) PAC can enhance the performance of subsequent conventional treatment
process (coagulation-sedimentation-filtration) on removing C-DBP and N-DBP
precursors. In order to evaluate the impact of these pretreatment processes, the
DBP FPs in different stages of the water treatment process were investigated.
3.3. DBP FPs
The FPs of CF, DCAN, DCAcAm and TCNM in different stages of the
water treatment trains were investigated (Figure 5). From Figure 5 (A), the
mean FP removal rate of CF (65.7%) by conventional water treatment processes
alone was higher than that of DCAN (50.2%), DCAcAm (24.7%) and TCNM
(57.0%), implying that the precursors of DCAN, DCAcAm and TCNM were
more recalcitrant than CF precursors by conventional water treatment processes
alone, especially for DCAcAm precursors.
Figure 5. The FPs of CF, DCAN, DCAcAm and TCNM in different stages of
the water treatment process (DBP FPs of original raw water were defined as
100%). A: Conventional water treatment process only; B: PAC adsorption +
conventional water treatment process; C: KMnO4 oxidation + conventional water
treatment process; D: BCO + conventional water treatment process. a: Original
raw water; b After BCO pretreatment; c: After coagulation-sedimentation; d:
After filtration. The error bars represent the standard deviation of replicate
measurements (n = 3).
316
From Figure 5 (B) and (C), it can be seen that PAC adsorption can effectively
remove the precursors of CF (42.7%), DCAN (28.6%), DCAcAm (27.2%)
and TCNM (35.7%), which were greater than the mean removal rate of the
precursors of CF (14.0%), DCAN (24.5%), DCAcAm (19.8%) and TCNM
(15.6%) pre-treated by KMnO4 oxidation. As shown in Table 1, the SUVA of raw
water declined to below 2 L/mg·m after PAC adsorption and/or KMnO4 oxidation.
At SUVA < 2 L/mg·m, the organic matter was mostly comprised of non-humics
and was of low hydrophobicity (25), indicating that humics and hydrophobic
organic matter were absorbed by PAC or oxidized by KMnO4. This suggests that
these humics and hydrophobic organic matters, which probably were a significant
part of the THM precursors (26).
From Figure 5 (D), the mean FPs of N-DBPs (DCAN: 120%; DCAcAm:
116%; TCNM: 104%) of pre-treated water by the BCO process were noticeably
higher than that (100%) of the raw water before the BCO process. As discussed
above, numerous metabolic by-products produced by the BCO process raised the
DON level, believed to be the main N-DBP precursor, which probably caused
the observed increase of N-DBP FPs. Figure 6 displays good linear relationships
between the DON concentration and the three N-DBPs (DCAN: R2 = 0.82; TCNM:
R2 = 0.69; DCAcAm: R2 = 0.83) in the DBP FP tests of BCO pretreatment and
conventional treatment processes. This suggests that DON plays an important
role in the formation of DCAN, DCAcAm and TCNM, and the presence of DON
in source waters may act as a reasonable surrogate indicator of N-DBP FPs. In
addition, from Figure 6 (B), a linear relationship was observed between SMP peak
intensity and DCAcAm (R2 = 0.94), implying that the SMP produced in the BCO
process may be an important class of HAcAm precursors and caused the higher
concentrations and FPs of DCAcAm.
Figure 6. The relationships between N-DBP FPs and DON concentration, and
DCAcAm FP and fluorescence EEM SMP peak intensity in a typical run in
November 2008 (DBP FPs of original raw water were defined as 100%).
317
4. Conventional Treatment (Coagulation-IPS-Filtration versus
Coagulation-DAF-Filtration)
4.1. Precursors
Two conventional processes (coagulation-IPS-filtration vs coagulation-DAF-

filtration) were compared, without regard to pre-treatment and advanced treatment
processes. There were no observable differences on the removal of turbidity by the
two conventional treatment processes. The turbidity and DOC of the filtered water
through the IPS or DAF were both less than 1.0 NTU and 2.5 mg/L, respectively,
indicating that the two process flows were operated as designed. Additionally, we
found the UV254 and DOC in the filtered water post-IPS were higher than that post-
DAF (Table 2), and the average removal efficiencies of UV254 (27%) and DOC
(47%) in the filtered water post-IPS were lower than post-DAF (UV254: 31%;
DOC: 53%). This phenomenon was probably linked to the SUVA of the water
(25). At SUVA < 2, the organic matter in the raw water was mostly comprised of
non-humics, and had low hydrophobicity and low MW, resulting in a poor DOC
removal. When SUVA in the raw water was close to 2, the similar organic matter
characteristics (low hydrophobicity and low molecular weight) in the raw water
likely caused a poor DOC removal by the coagulation-IPS-filtration (26).
Figure 7. MW distribution of raw water and filtered water after the two different
coagulation processes (IPS and DAF).
318
Table 2. DOC, DON and UV254 in the Raw Water
DOC UV SUVA DON DON/DOC
Data (mg L-1) (cm -1) (L/(mg·m)) (mg L-1) mg/mg
(d) Raw Raw Raw Raw Raw

IPS DAF IPS DAF
water water water water water
1st 6.07 2.48 2.24 0.155 0.112 0.109 2.55 0.69 8.7
3rd 3.25 1.85 1.64 0.062 0.045 0.043 1.91 0.67 4.9
6th 5.29 3.26 2.85 0.097 0.073 0.068 1.83 0.79 6.7
13rd 4.69 2.45 2.23 0.101 0.079 0.072 2.15 0.83 5.7
17th 5.23 2.48 2.39 0.105 0.078 0.072 2.01 0.55 9.5
22nd 5.38 2.46 2.37 0.118 0.077 0.075 2.19 0.62 8.7
319
28th 4.82 3.07 2.53 0.085 0.065 0.061 1.76 0.73 6.6

The algae removal by the DAF was prominently better than IPS. This is
probably because most of the algae are low-density and suspended in the upper
water column and DAF is more effective than IPS in removing such suspended
material. In addition, the higher removal efficiencies of DON was achieved in
filtered water post-DAF probably because the dissolved nitrogenous organic
compounds with low hydrophobicity and low MW were removed less well by the
IPS process. The average removal efficiencies of DON by the two processes (IPS
and DAF) were 31% and 53%, respectively.
Table 3. Guidelines on the Nature of NOM and Expected DOC Removals at

Different SUVA Values

SUVA DOC removal by coagulation-
Composition
(L mg-1m-1) sedimentation-filtration
Mostly aquatic humics,
≥4 High hydrophobicity Good DOC removal
High molecular weight (MW)
Mixture of Aquatic Humics
and other NOM,
2-4 Mixture of hydrophobic and DOC removals should be fair to Good
hydrophilic
Mixture of MWs
Mostly Non-humics
<2 Low hydrophobicity Poor DOC removal
Low MW
As shown in Figure 7, the MW distribution of raw water revealed that organic

matters in the raw water mainly consisted of low MW (MW < 5000 Da), which
was in accordance with the reduction from SUVA in Table 3 (25). In addition, the
polydispersity (Mw/Mn) value close to 1 means that the constituents in organic
matters mostly made up of organic substances with identical or at least similar
origin and/or properties (27). From Figure 7, Mw/Mn values of Peak A and Peak
B are 1.2 and 21, respectively, implying coagulation–IPS-filtration performed less
well at removing the low MW organics (Peak A: 1 < LogMW < 2) with similar
origin and/or properties.
4.2. C-DBPs
The CF concentrations measured after chlorination following the two algae

removal processes (IPS and DAF) are presented in Figure 8. CF in chlorinated
water post-IPS and post-DAF reached their maximum values of 17 and 13 ug/L,
respectively. More CF was formed after IPS than after DAF, thereby suggesting
320
that DAF removed THM precursors more effectively than IPS. As shown in Table
2, less DOC, the most important precursor indicator of C-DBPs (e.g. CF), was
removed by coagulation–IPS-filtration than by coagulation–DAF-filtration. In
addition, algae also represented an important THM precursor (8, 28, 29). Less
THM was formed during chlorination of post-DAF samples than post-IPS may be
another reason for the improved CF precursors removal of the DAF over IPS.
Figure 8. Chloroform concentration in chlorinated water after the two different

coagulation processes (IPS and DAF). Error bars represent the standard
deviation of replicate measurements (n = 3).
4.3. N-DBPs
The N-DBP concentrations in the water after chlorination following the
two different processes (IPS and DAF) are summarized in Figure 9. DCAcAm,
DCAN and TCNM ranged within 1.5–2.9, 3.5–6.3 and 0.5–1.0 ug/L, respectively,
in the water treated by coagulation–IPS-filtration–chlorination. Additionally,
in the water treated by coagulation–DAF-filtration–chlorination, DCAcAm,
DCAN and TCNM varied within 0.6–1.5, 2.7–4.7 and 0.4–0.9 ug/L, respectively.
Because the better removal in the entire MW range was achieved by the
coagulation-IPS-filtration process, fewer precursors remained to form further
DCAcAm and DCAN. The TCNM concentrations after the two processes were
similar, probably because the TCNM formation was influenced by DIN (30)
instead of DON.
In addition, the average concentrations of DCAcAm (2.3 ug/L) and DCAN
(5.0 ug/L) in the water treated by coagulation–IPS-filtration–chlorination were
higher than the average levels reported in a US N-DBP occurrence study (5). A low
DOC/DON usually indicates that microorganism (e.g. algae, bacteria) products are
a major fraction of the NOM (autochthonous) (8, 31, 32). In a survey of US plants
that were impacted by algae and/or treated wastewater (33), the average DOC/
DON ratio (13 mg/mg) of the surveyed waters confirmed that these sources were
321
influenced by elevated levels of DON compared to a previous US study (34), where
the average DOC/DON ratio in a broader array of surface waters was 18 mg/mg. In
the studied raw water (Table 2), DOC/DON was 4.9 to 9.5 mg/mg, with an average
value of 7.3 mg/mg, indicating the Lake Taihu raw water was relatively organic
nitrogen rich, which possibly caused relative higher DCAcAm and DCAN yields.
We also found a linear relationship (R2 = 0.83) between DCAcAm and DCAN.
This is probably because DCAcAm is a main hydrolysis product of DCAN, and
the two N-DBPs share similar precursors and/or mechanisms of formation (7, 35).
Figure 9. N-DBP concentrations in chlorinated water after the two different

coagulation processes (IPS and DAF). Error bars represent the standard
deviation of replicate measurements (n = 3).
5. Advanced Treatment (O3–BAC)

5.1. Precursors
The advanced treatment process (O3-BAC) were studied after the

conventional processes, without regard to pre-treatment and advanced treatment
processes. As shown in Figure 10, the mean removal rates of turbidity, DOC,
and UV254 reached 98.4%, 56.7%, and 30.9% after the conventional process
(coagulation-sedimentation-filtration) and increased to 99.3%, 72.2%, and 52.6%
after the O3-BAC process, respectively. The final levels of turbidity, DOC, and
UV254 in the effluent after the combined process decreased to below 0.4 NTU, 2.0
mg/L, and 0.05 cm-1, confirming that the combined process flows were operated
as designed and expected.
322
Figure 10. Removal rates of water quality parameters (Figure 10a: turbidity,
DOC, and UV254; Figure 10b: DON and DIN) in different stages of the water
treatment process. (From Figure1, Sampling 1, 2, 3, 4 and 5 represented the
water samples collected from the raw water; after coagulation-sedimentation,
filtration, O3 and BAC processes, respectively.)
The conventional treatment process achieved a poor removal of NH4+-N

relative to turbidity, DOC and UV254. However, the O3-BAC process significantly
improved the removal of NH4+-N from 15.7% to 92.6%, which was in agreement
with other studies (36–39). The BAC process removed 70% of NH4+-N. After the
conventional process (coagulation-sedimentation-filtration), the average removal
rates of NO2--N and NO3--N reached 85.6% and 53.7%, and further increased to
100% and decreased to 45% in the ozonated water, respectively. This is probably
because part of NO2--N was oxidized to NO3--N by the ozone. In addition, the
removal rate of NO3--N decreased by 61.6% in the effluent of the BAC process,
likely due to biological nitrification in the BAC bed. Although more NO3--N
was formed in the BAC effluent, the measured concentrations of NO3--N were
always lower than 2.0 mg/L that is far below the maximum contaminant level (10
mg/L) of Chinese National Standards of Drinking Water Quality and the USEPA
National Primary Drinking Water Regulations for NO3--N.
323
Figure 11. Removal of DON by the BAC process with and without pre-ozonation.
(The removal rates were calculated based on the DON concentrations of
raw waters; The error bars represent the standard deviation of replicate
measurements (n = 3).)
The average removal rate of DON by the conventional process reached only
35.1%, while the post combined O3-BAC process increased its removal efficiency
to 73.7%. In addition, we compared the difference in the DON removal by
BAC with and without pre-ozonation (Figure 11). It was found that only 21.3%
of DON was removed by the BAC process without pre-ozonation, which was
lower than that achieved with pre-ozonation (28.1%). This is likely because
ozonation degrades large molecules of DON compounds into smaller ones and
hence increases the biodegradability of DON.
5.2. DBPs
From Figure 12, the removal rates of the FPs for CF and DCAN both gradually
increased throughout the treatment train. The coagulation-sedimentation-filtration
process removed 52.7% and 49.9% of the precursors for CF and DCAN,
respectively, and O3-BAC combined process further enhanced the precursor
removal efficiency of CF and DCAN to 85.0% and 80.4%, separately. The effect
of O3-BAC is in agreement with the previous observations that ozonation and
BAC substantially reduce the formation of CF and appear not to be significantly
affected by source water quality (36–38). Additionally, the conventional process
(coagulation-sedimentation-filtration) and advanced treatment process (O3-BAC)
both had higher removal rates for the FPs of CF and DCAN than of TCNM and
DCAcAm. A plausible reason is that dissolved humic material, including a major
fraction of the total organic matter in freshwaters (40), is the major precursor
of CF and DCAN but makes little contribution to the formation of TCNM and
DCAcAm (7, 41). Also, hydrophilic organic matters with low-MW trend to form
more TCNM and DCAcAm, and it is more difficult to remove these substances
than humic material with high-MW (16, 42).
324
Figure 12. Removal rates of the FPs of CF, DCAN, TCNM and DCAcAm
in different stages of the water treatment process. (Sampling 1, 2, 3,
4 and 5 represented the water samples collected from raw water; after
coagulation-sedimentation, filtration, O3 and BAC processes, respectively, as
shown in Figure 1, The removal rates were calculated based on the DBP FPs
of raw waters; The error bars represent the standard deviation of replicate
Figure 13. Removal of TCNMFP by the BAC process with and without
pre-ozonation. (The error bars represent the standard deviation of replicate
325
From Figure 13, it is would be noted that the conventional process could
not effectively remove the precursors of TCNM, and the mean removal rates of
TCNMFP decreased from 18% in filtrated water to -187.5% in ozonated water,
but increased to 49.6% after the subsequent BAC process. After ozonation, 46%
of DON was reduced relative to the raw water (Figure 10), but TCNMFP increased
from 2.7 to 7.4 ug/L. This was probably because the ozone-induced oxidation
increased the content of highly reactive specific organic compounds with high
HNMFP (43), even though the overall DON was reduced. The performance of
the BAC process in terms of controlling TCNM formation either with or without
pre-ozonation was also examined (Figure 13). The TCNMFP removal efficiency
accomplished by BAC without ozonation pre-treatment was only 38%, less than
50% achieved by the combined O3–BAC. Thus, although ozone increased the
TCNMFP, it enhanced the biodegradability of the organics responsible for HNMFP
to facilitate their removal in the following BAC treatment.
Figure 14. UV/vis optical spectrums of water samples in different stages of the
water treatment process. (From Figure 1, Sampling 1, 3, 4 and 5 represent the
raw water; after coagulation-sedimentation-filtration, after O3 and after the
BAC processes, respectively.)
Figure 12 also provided the important information in the removal of

DCAcAm. The subsequent BAC process significantly improved the removal of
the precursors of DCAcAm (60% removal). UV/Vis (Figure 14) and EEM (Figure
15) spectra were used to characterize the NOM. As seen in Figure 14, at any
particular wavelength within the range of 220–300 nm, the UV absorbance of the
326
effluent from the BAC column was significantly lower than that of raw, filtered,
and ozonated waters, indicating that BAC is an effective process to remove a
range of organic compounds over a broad characteristic UV absorbance range.
Figure 15. Representative fluorescence EEM spectra of DOM in raw water.
The fluorescence index (FI) can be calculated as the ratio of fluorescence

intensities at emission 450 and 500 nm, at excitation 370 in the EEM data. Higher
and lower FI values reflect organic contents of an autochthonous (microbial) origin
and an allochthonous (terrestrial) origin, respectively (16, 44). From Figure 15,
the FI value after BAC significantly increased from 1.7 to 8.7, probably because
the allochthonous organics were metabolized to autochthonous organics as a result
of microbial activities in BAC. According to the FRI method developed by Chen
et al. (21), the NOM of the raw water (Figure 15a), filtered water (Figure 15b)
and ozonated water (Figure 15c) all had two intense peaks in the SMPs (λex > 250
nm, λex < 380 nm) and aromatic proteins (APs)-like regions (λex < 250 nm, λex <
380 nm), and the fluorescence intensities of two peaks in the raw (SMP: 988.6 mV
and AP: 988.2 mV), filtered (SMP: 985.0 mV and AP: 983.7 mV) and ozonated
water (SMP: 982.5 mV and AP: 981.8 mV) were similar. Thus, the conventional
treatment process and ozonation were ineffective at removing the SMP-like and
AP-like substances.
The NOM in the BAC effluent showed an AP-like peak at relatively low
fluorescence intensities (239.5 mV), and the SMP peak disappeared. Previous
studies (14, 17) reported that SMP-like and AP-like substances, containing
327
elevated organic nitrogen, were the most important precursors of DCAcAm. The
results for the removal of DCAcAmFP (Figure 12) and the variation in SMP-like
and AP-like peak intensities in EEM spectra (Figure 13) suggest that the BAC
process is able to reduce the precursors of DCAcAm.
It should be noted that the FI and FRI methods both provide only a qualitative
and/or semi-quantitative analysis of the data, but they cannot separate the
complex measured signal into its individual underlying fluorescent phenomena
with specific excitation. In order to further understand the characteristics
of DBP precursors with the EEM spectra, and ascertain the performance of
conventional and advanced treatment processes on removal of the DOMs with
different characteristics, an improved EEM analysis model will be adopted in the
future experiments. For example, parallel factor analysis is a valuable tool for
characterizing and quantifying changes in DOM fluorescence, and enabling the
tracing of different fractions in the natural environment (45).
5.3. Total Organic Halogen (TOX)
Korshin et al. (46) reported that a decrease in UV absorbance at 272 nm

caused by the chlorination of DOM correlates linearly with the amount of TOX
formed, under a wide range of water quality conditions and reaction times.
In the present study, therefore, the decrease of UV272 (DUV272) before and
after the DBPFP test was assessed and the removal efficiencies (DUV272) in
different stages of the treatment process were used to suggest the removal of
TOXFP. Removal efficiencies of TOX (DUV272) by the conventional process
(coagulation–sedimentation–filtration), ozonation, and BAC filtration were 44%,
45% and 72%, respectively, suggesting that the BAC process had a potential to
control halogenated DBPs, including unidentified DBPs.
5.4. Effect of Terminating Pre-Ozonation
Two O3-BAC processes were operated in parallel (Figure 16) and initially
achieved the same removal of DOC and DON and the same concentrations of
C- and N-DBPs upon subsequent chlorination. Thereafter, ozone dosing was
terminated in the second day after backwashing for one of the two O3-BAC
processes, which is represented as ‘T-O3-BAC’ in the study. And the other
O3-BAC process was operated still with continuous ozonation before BAC
filtration, which is represented as ‘C-O3-BAC’. The performance of C-O3-BAC
and T-O3-BAC was compared in terms of water quality in the effluents of sand
filtration and BAC filtration in the second, third and fourth day after BAC
backwashing. The removal rate of each parameter by C-O3-BAC or T-O3-BAC
process is presented by equation (1).
CA— Concentration detected after sand filtration

CB— Concentration detected after BAC filtration
328
Figure 16. The parallel O3-BAC pilot plant process flows.
Figure 17. The removal of DOC, DON, C-DBPs and N-DBPs. (a and b Show
the removal by the C–O3–BAC and T–O3–BAC in first backwash cycle after
terminating ozone dosing, respectively; c and d show the removal by the
C–O3–BAC and T–O3–BAC in the second backwash cycle after terminating ozone
dosing, respectively; e and f show the removal by the C–O3–BAC and T–O3–BAC
in the fifth backwash cycle after terminating ozone dosing, respectively. The
removal is the average value of the samples collected three times on the second,
third and fourth day after BAC backwashing. The error bars represent the
standard deviation of three removal values.)
329
Figure 18. The relationship between DOC removal and C-DBP (CF, DCAA, and
TCAA) formation potential (FP) removal by the BAC process.
As shown in Figure17, the formation yields of C-DBPs upon chlorination of

the sand filter effluent, C-O3-BAC filter and T-O3-BAC filter were measured. The
reduction in C-DBP formation yields by the C-O3-BAC and T-O3-BAC filtration
relative to sand filtration was calculated by equation (1). C-O3-BAC achieved
better reduction of C-DBPs (CF, DCAA and TCAA) than T-O3-BAC (p < 0.05).
The reduction in the C-DBP formation yields closely mirrored the reduction
in DOC concentrations, as expected, since DOC is a known C-DBP precursor
indicator measurement (47, 48). The linear correlation between DOC removal
and C-DBP removal was observed in both the C-O3-BAC and T-O3-BAC waters
(Figure 18).
The formation yields of N-DBPs upon chlorination of the sand filter effluent,
C–O3–BAC filter and T–O3–BAC filter were measured, as shown in Figure 17.
The reduction in N-DBP formation yields by the C–O3–BAC and T–O3–BAC
was calculated by Eq. (1). C–O3–BAC filtration achieved a stable removal
of DCAN (63.4–69.7%), TCAN (88.1–89.2%), DCNM (65.5–69.5%), TCNM
(51.7–56.9%), DCAcAm (68.2–71.6%) and TCAcAm (82.1–88.5%) from the
first to fifth backwash cycles (p values of all selected N-DBPs were higher than
0.05). Some N-DBP concentrations increased in the T–O3–BAC filter after the
initial backwash cycle, which were 21.3%, 19.6%, 33.5% and 53.7% higher than
330
that in the chlorinated effluent of the sand filter for DCAN, TCAN, DCAcAm and
TCAcAm (Figure 17), respectively. The increased N-DBP concentrations suggest
that new N-DBP precursors were produced into the water in the T–O3–BAC case.
Also, the concentrations of DCNM and TCNM were only reduced by 36.3% and
29.5% compared to the chlorinated effluent water of the sand filter, respectively.
Figure 19. The relationship between DON removal and N-DBP (HANs
[DCAN and TCAN], HNMs [DCNM and TCNM], and HAcAms [DCAcAm and
TCAcAm]) formation potential (FP) removal by the two BAC processes.
As shown in Figure17, C-O3-BAC presented a stable removal of DOC

(29.7%-32.6%) and DON (33.0%-36.7%) from the first to fifth backwash cycles
(p values of DON and DOC was both higher than 0.05). T-O3-BAC resulted
in lower DOC and DON removal than C-O3-BAC. Of note, T-O3-BAC process
significantly increased the DON concentration of the sand-filtered water by 23.9%
(from 0.46 to 0.57 mg/L) in the first backwash cycle. After that, the T-O3-BAC
partially recovered the ability to remove DON from the second to fifth backwash
331
cycles. The increase in N-DBP formation yields also mirrored the increase in
DON concentrations. As in the case of the relationship between DOC and C-DBP
concentrations, the linear correlation between DON and N-DBP concentrations
existed in both the C-O3-BAC and T-O3-BAC waters (Figure 19). However,
the degree of linear correlation between DON and N-DBP concentrations (R2
< 0.90, in Figure 19) was poorer than that between DOC and C-DBPs (R2 >
0.95, in Figure 19), suggesting that not only the DON concentration but also the
characteristics of the DON (e.g., SMPs) in the post-BAC water influences the
N-DBP yield (14).
Figure 20. MW distribution of DOM treated by sand filtration, ozonation, and

BAC filtration in the first backwash cycle. (SF: treated by sand filtration; O3:
treated by ozonation; C–O3–BAC: treated by C–O3–BAC; T–O3–BAC: treated by
T–O3–BAC.)
It is well known that ozonation preferentially transforms higher molecular

weight compounds into smaller molecules (Figure 20) and thereby increases the
biodegradable organic matters (49) that are readily removed by a subsequent
BAC process (19, 42) When ozone dosing was suddenly terminated, part of the
microorganisms in the BAC might subsequently shift from exogenous respiration
to endogenous respiration due to the reduction in biodegradable organic matters
(39). Consequently, this resulted in the release of numerous soluble microbial
products (SMPs) (50). It is known that a range of SMPs are produced by
microorganisms in the BAC process and can be detected in BAC effluent (20, 51,
52). SMPs comprise a wide range of high and low molecular weight compounds
including many DON compounds such as proteins, polysaccharide and amino
acids (21, 53, 54), which can be divided into growth-related utilization associated
332
products (UAPs) and non-growth-related biomass associated products (BAPs).
The formation of UAPs and BAPs is directly related to the cell growth rate
and the cell endogenous respiration of the microorganisms, respectively (51).
As shown in Figure 20, we investigated the MW distribution of the DOM in
effluents of sand filtration, C-O3-BAC filtration and T-O3-BAC filtration, and
found that C-O3-BAC filtration slightly altered the MW distribution of DOM due
to the ozonation oxidation. However, T-O3-BAC filtration substantially increased
the percentage of high-MW (>10 kDa) and low-MW (< 1 kDa) compared to
C-O3-BAC filtration (p < 0.05), which was in accordance with the deduction that
SMPs were released in T-O3-BAC filtration, considering that most of the SMPs
consisted of the low-MW (<1 kDa, e.g., amino acids) and high-MW (>10 kDa,
e.g., proteins, polysaccharide) compounds (52, 55).
Figure 21. Fluorescence EEM spectra of Lake Taihu water treated by sand
filtration, ozonation, and BAC filtration in the first backwash cycle. (SF: treated
by sand filtration; O3: treated by ozonation; C-O3-BAC: treated by C-O3-BAC;
T-O3-BAC: treated by T-O3-BAC.)
EEM spectra were used to characterize the DOM in effluents of sand filtration,
C-O3-BAC filtration and T-O3-BAC filtration, as shown in Figure 21. As shown,
DOM in the sand-filtered water (Figure 21 A), ozonated water (Figure 21B), and
BAC-filtered water (Figure 21C and D) all had two intense peaks in the SMP
region (λex > 250 nm, λem < 380 nm) and aromatic protein (AP)-like region (λex <
250 nm, λem < 380 nm), and the fluorescence intensities of the two peaks (SMP
and AP) in sand-filtered water was reduced by C-O3-BAC filtration from 553
333
and 362 a.u. to 246 and 77 a.u.. Previous studies (14) reported that SMP-like
and AP-like substances, containing elevated organic nitrogen, were the important
precursors of N-DBPs (HANs and HAcAms). The results for the removal of
HAN and HAcAm formation yields (Figure 17) and the variation in SMP-like
and AP-like peak intensities in EEM spectra (Figure 21 A, B and C) suggest that
the C-O3-BAC process is able to reduce the precursors of these N-DBPs, which
supports earlier evidence of this (19). However, from Figure 21D, it was noted that
the fluorescence intensities of SMP and AP peaks in the effluent of T-O3-BAC
filtration actually increased to 649 and 357 mV, which mirrored the increase of
HAN and HAcAm concentrations.
5. Conclusion
The application of pretreatment processes in water treatment can improve

the removal of algae, NH4+-N and organic matter, and reduce the formation of
C-DBPs and certain N-DBPs. However, the BCO pretreatment process could also
increase DON and result in an increase of the concentrations and FPs of some
N-DBPs. KMnO4 can enhance the performance of conventional water treatment
processes in removing DBP precursors, although it removes less precursors than
PAC adsorption. Hydrophilic SMP produced from the BCO process plays a key
role in DCAcAm formation. An additional benefit of improving the removal of
hydrophilic nitrogenous organics by pretreatment is the reduced formation of
certain NDBPs such as HAcAms.
The transition from coagulation–IPS-filtration to coagulation–DAF-filtration
has the potential of removing DBP precursor indicators (e.g. DOC, DON and
algae) and reducing the formation of C-DBPs and certain halogenated N-DBPs.
Coagulation–DAF-filtration process removed, on the average, 53%, 53% and
31% of DOC, DON and UV254, which were better than 47%, 31% and 27% of
that in coagulation–IPS-filtration process. Moreover, less low MW organics were
removed by coagulation–IPS-filtration than coagulation–DAF-filtration. This
study showed that some N-DBPs (DCAcAm and DCAN) clearly exhibited higher
formation in high-DON and algae-rich waters. Therefore, an additional benefit
of improving DON and algae control is the reduction in the formation of certain
N-DBPs. Additionally, the effect of DON may be a more important factor than
DOC in the future study on risk assessment and concentration prediction of these
N-DBPs in the water with high DON loadings.
The O3-BAC integrated process is a very promising technology to control
the formation of certain halogenated C-DBPs (THMs) and N-DBPs (HANs,
HNMs and HAcAms). Ozonation increased the FPs of TCNM substantially, but
also enhanced the biodegradability of TCNM precursors and further improved
the removal of TCNM precursors by BAC filtration. Ozonation was inefficient
in removing the precursors of DCAcAm, however, the BAC process greatly
improved the removal of DCAcAm precursors including SMP-like and AP-like
substances. The O3-BAC combined process should therefore be considered as an
option to address regulated and emerging DBPs. However, in some cases, ozone
was intermittently dosed (versus continuously) before BAC filtration in DWTPs
334
in order to save costs by water treatment facilities, control the bromate formation
or in emergency situations. This study has showed that a sudden termination
of ozonation cannot effectively remove N-DBP (HAN and HAcAm) precursors
even after only the first backwash cycle, and then leads to substantial formation
of HANs and HAcAms during subsequent chlorination. Moreover, after the first
backwash cycle, BAC filtration alone performed poorly for all measured C-DBPs
and N-DBPs compared to C-O3–BAC.
Acknowledgments
The authors gratefully acknowledge the National Natural Science Foundation
of China (51108327, 51378366), and the National Major Science and Technology
Project of China (2012ZX07403-001,004). The authors also thank their colleagues

for their support on the sampling surveys and DBPs analysis.
Abbreviations
Bromochloroacetonitrile –BCAN; Biological contact oxidation
–BCO; Carbonaceous disinfection by-products –C-DBPs; Chloroform
–CF; Dibromoacetonitrile –DBAN; Disinfection by-products –DBPs;
Dichloroacetamide –DCAcAm; Dichloroacetonitrile –DCAN; Dissolved
inorganic nitrogen –DIN; Dissolved organic carbon –DOC; Dissolved organic
nitrogen –DON; Drinking water treatment plant –DWTP; Excitation-emission
matrix –EEM; Gas chromatograph/mass spectrometry –GC/MS; Fluorescence
regional integration –FRI; Haloacetic acids –HAAs; Haloacetamides –HAcAms;
Haloacetonitriles –HANs; Halonitromethanes –HNMs; Nitrogenous disinfection
by-products –N-DBPs; N-nitrosodimethylamine –NDMA; Natural organic
matter –NOM; Powdered activated carbon –PAC; Soluble microbial product
–SMP; Specific ultraviolet absorption –SUVA; Trichloroacetamide –TCAcAm;
Trichloroacetonitrile –TCAN; Trichloronitromethane –TCNM; Trihalomethanes
–THMs; Total dissolved nitrogen –TDN; Total organic halogen –TOX.
References
research. Mutation Res. 2007, 636 (1-3), 178–242.
3. Muellner, M. G.; Wagner, E. D.; McCalla, K.; Richardson, S. D.; Woo, Y. T.;
Plewa, M. J. Haloacetonitriles vs. regulated haloacetic acids: Are nitrogen-
containing DBPs more toxic? Environ. Sci. Technol. 2007, 41, 645–651.
335
4. Richardson, S. D.; Ternes, T. A. Water analysis: Emerging contaminants and
current issues. Anal. Chem. 2014, 86, 2813–2848.
7175–7185.
6. Plewa, M. J.; Wagner, E. D.; Muellner, M. G.; Hsu, K. M.; Richardson,
S. D. Comparative mammalian cell toxicity of N-DBPs and C-DBPs.
In Occurrence, Formation, Health Effects and Control of Disinfection
by-Products in Drinking Water; Karanfil, T., Krasner, S. W., Xie, Y., Eds.;
ACS Symposium Series 995; American Chemical Society: Washington,
DC, 2008; pp 36−50.
7. Reckhow, D. A.; Platt, T. L.; MacNeill, A. L.; McClellan, J. N. Formation
and degradation of dichloroacetonitrile in drinking waters. J. Water Supply:

Res. Technol. AQUA. 2001, 50 (1), 1–13.
8. Westerhoff, P.; Mash, H. Dissolved organic nitrogen in drinking water
supplies: a review. J. Water Supply: Res. Technol. −AQUA 2002, 51,
415–448.
9. Joo, S. H.; Mitch, W. A. Nitrile, aldehyde, and halonitroalkane formation
during chlorination/chloramination of primary amines. Environ. Sci.
Technol. 2007, 41, 1288–1296.
10. Chu, W. H.; Gao, N. Y.; Deng, Y. Stability of newfound nitrogenous
disinfection byproducts haloacetamides in drinking water. Chin. J. Org.
Chem. 2009, 29, 1569–1574.
potential in water. Water Res. 2010, 44, 3755–3762.
12. Chiang, P. C.; Chang, E. E.; Chuang, C. C.; Liang, C. H.; Huang, C. P.
Evaluating and elucidating the formation of nitrogen-contained disinfection
by-products during pre-ozonation and chlorination. Chemosphere 2010, 80,
327–333.
14. Chu, W. H.; Gao, N. Y.; Deng, Y.; Krasner, S. W. Precursors of
dichloroacetamide, an emerging nitrogenous DBP formed during
chlorination or chloramination. Environ. Sci. Technol. 2010, 44,
3908–3912.
15. Hou, Y. K.; Chu, W. H.; Ma, M. Carbonaceous and nitrogenous disinfection
by-products formation in the surface and ground water treatment plants using
Yellow River as water source. J. Environ. Sci. 2012, 24, 1204–1209.
16. Chu, W. H.; Gao, N. Y.; Yang, D.; Templeton, M. R.; Yin, D. Q.
Formation of nitrogenous disinfection by-products from pre-chloramination.
Chemosphere 2011a, 85, 1187–1191.
17. Chu, W. H.; Gao, N. Y.; Yang, D.; Templeton, M. R.; Yin, D. Q. Impacts
of drinking water pretreatments on the formation of nitrogenous disinfection
by-products. Bioresour. Technol. 2011b, 102, 11161–11166.
336
18. Chu, W. H.; Gao, N. Y.; Templeton, M. R.; Yin, D. Q. Comparison of
inclined plate sedimentation and dissolved air flotation for the minimisation
of subsequent nitrogenous disinfection by-product formation. Chemosphere
2011c, 83, 647–651.
19. Chu, W. H.; Gao, N. Y.; Yin, D. Q.; Deng, Y.; Templeton, M. R. Ozone-
biological activated carbon integrated treatment for removal of precursors of
halogenated nitrogenous disinfection by-products. Chemosphere 2012, 86,
1087–1091.
20. Barker, D. J.; Stuckey, D. C. A review of soluble microbial products (SMP)
in wastewater treatment systems. Water Res. 1999, 33, 3063–3082.
21. Chen, W.; Westerhoff, P.; Leenheer, J. A.; Booksh, K. Fluorescence
excitation−emission matrix regional integration to quantify spectra for
dissolved organic matter. Environ. Sci. Technol. 2003, 37, 5701–5710.
22. Jiang, T. Characterization and modelling of soluble microbial products in

membrane bioreactors. Ph.D. Thesis, Ghent Universtiy, Belgium, 2007.
23. Zhou, W. L.; Wu, B. T.; She, Q. H.; Chi, L.; Zhang, Z. J. Investigation
of soluble microbial products in a full-scale UASB reactor running at low
organic loading rate. Bioresour. Technol. 2009, 100, 3471–3476.
24. Chu, W. H.; Gao, N. Y.; Yin, D. Q.; Krasner, S. W. Formation and speciation
of nine haloacetamides, an emerging class of nitrogenous DBPs, during
chlorination or chloramination. J. Hazard. Mater. 2013, 260, 806–812.
25. Edzward, J. K.; Tobiason, J. E. Enhanced coagulation: US requirements and
a broader view. Water Sci. Technol. 1999, 40, 63–70.
26. Roe, J.; Baker, A.; Bridgeman, J. Relating organic matter character to
trihalomethanes formation potential: a data mining approach. Water Sci.
Technol.: Water Supply 2008, 8, 717–723.
27. Nam, S. N.; Krasner, S. W.; Amy, G. L. Differentiating effluent organic
matter (EfOM) from natural organic matter (NOM): Impact of EfOM on
drinking water sources. In 6th International Symposium on Advanced
Environmental Monitoring; Kim, Y. J., Platt, U., Eds.; Heidelberg, Germany,
2006; pp 259–270.
28. Hoehn, R. C.; Dietrich, A. M.; Farmer, W. S.; Orr, M. P.; Lee, R. G.; Aieta, E.
M.; Wood, D. W.; Gordon, G. Household odors associated with the use of
chlorine dioxide. J. Am. Water Works Assoc. 1990, 82 (4), 166–172.
29. Huang, J.; Graham, N.; Templeton, M. R.; Zhang, Y.; Collins, C.;
Nieuwenhuijsen, M. A comparison of the role of two blue-green algae in
THM and HAA formation. Water Res. 2009, 43, 3009–3018.
30. Thibaud, H.; De Laat, J.; Merlet, N.; Dore, M. Chloropicrin formation
in aqueous solution: effect of nitrites on precursors formation during the
oxidation of organic compounds. Water Res. 1987, 21, 813–821.
31. Aiken, G.; Cotsaris, E. Soil and hydrology – their effect on NOM. J. Am.
Water Works Assoc. 1995, 87 (1), 36–45.
32. Lee, W. Occurrence, Molecular weight and treatability of dissolved organic
nitrogen. Ph.D. Dissertation, Arizona State University, Tempe, AZ, 2005.
33. Mitch, W. A.; Krasner, S. W.; Westerhoff, P.; Dotson, A. Occurrence and
formation of nitrogenous disinfection by-products; American Water Works
Association Research Foundation: Denver, Colorado, 2009.
337
34. Lee, W. P.; Westerhoff, P.; Esparza-Soto, M. Occurrence and removal of
dissolved organic nitrogen in US water treatment plants. J. Am. Water Works
Assoc. 2006, 98 (10), 102–110.
35. Pressman, J. G.; Richardson, S. D.; Speth, T. F.; Miltner, R. J.; Narotsky, M.
G.; Hunter, E. S.; Rice, G. E.; Teuschler, L. E.; McDonald, A.; Parvez, S.;
Krasner, S. K.; Weinberg, H. S.; McKague, A. B.; Parrett, C. J.; Bodin, N.;
Chinn, R.; Lee, C. F. T.; Simmons, J. E. Concentration, chlorination and
chemical analysis of drinking water disinfection byproduct mixtures health
effects research: US EPA’s four lab study. Environ. Sci. Technol. 2010, 44,
7184–7192.
36. Karnik, B. S.; Davies, S. H.; Baumann, M. J.; Masten, S. J. The effects
of combined ozonation and filtration on disinfection by-product formation.
Water Res. 2005, 39, 2839–2850.
37. Chiang, P. C.; Chang, E. E.; Chang, P. C.; Huang, C. P. Effects of

pre-ozonation on the removal of THM precursors by coagulation. Sci. Total.
Environ. 2009, 407, 5735–5742.
38. Yan, M. Q.; Wang, D. S.; Ma, X. N.; Ni, J. R.; Zhang, H. S. THMs precursor
removal by an integrated process of ozonation and biological granular
activated carbon for typical Northern China water. Sep. Purif. Technol.
2010, 72, 263–268.
39. Yapsakli, K.; Mertoglu, B.; Cecen, F. Identification of nitrifiers and
nitrification performance in drinking water biological activated carbon
(BAC) filtration. Process Biochem. 2010, 45, 1543–1549.
40. Lee, S. K.; Freitag, D.; Steinberg, C.; Kettrup, A.; Kim, Y. H. Effects of
dissolved humic materials on acute toxicity of some organic chemicals to
aquatic organisms. Water Res. 1993, 27, 199–204.
41. Rook, J. J. Formation of haloforms during chlorination of natural waters.
Water Treat. Exam. 1974, 23, 234–239.
42. Xu, B.; Gao, N. Y.; Sun, X. F.; Xia, S. J.; Simonnot, M. O.; Causserand, C.;
Rui, M.; Wu, H. H. Characteristics of organic material in Huangpu River
and treatability with the O3-BAC process. Sep. Purif. Technol. 2007, 57,
348–355.
43. Hu, J.; Jesse, H. S.; Addison, W.; Karanfil, T. Halonitromethane formation
potentials in drinking waters. Water Res. 2010, 44, 105–114.
44. McKnight, D. M.; Boyer, E. W.; Westerhoff, P. K.; Doran, P. T.; Kulbe, T.;
Andersen, D. T. Spectrofluorometric characterization of dissolved organic
matter for indication of precursor organic material and aromaticity. Limnol.
Oceanogr. 2001, 46, 38–48.
45. Jørgensen, L.; Stedmon, C. A.; Kragh, T.; Markager, S.; Middelboe, M.;
Søndergaard, M. Global trends in the fluorescence characteristics and
distribution of marine dissolved organic matter. Mar. Chem. 2011, 126,
139–148.
46. Korshin, G. V.; Li, C. W.; Benjamin, M. M. The decrease of UV absorbance
as an indicator of TOX formation. Water Res. 1997, 31, 946–949.
47. Chellam, S.; Krasner, S. W. Disinfection byproduct relationships and
speciation in chlorinated nanofiltered waters. Environ. Sci. Technol. 2001,
35, 3988–3999.
338
48. Xue, C. H.; Wang, Q.; Chu, W. H.; Templeton, M. R. The impact of changes
in source water quality on trihalomethane and haloacetonitrile formation in
chlorinated drinking water. Chemosphere 2014, 117, 251–255.
49. Santos, D. C.; Silva, L.; Albuquerque, A.; Simões, A.; Gomes, A. C.
Biodegradability enhancement and detoxification of cork processing
wastewater molecular size fractions by ozone. Bioresour. Technol. 2013,
147, 143–151.
50. Schiener, P.; Nachaiyasit, S.; Stuckey, D. C. Production of soluble
microbial products (SMPs) in an anaerobic baffled reactor: compositin,
biodegradability and the effect of process parameters. Environ. Technol.
1998, 19, 391–400.
51. Carlson, K. H.; Amy, G. L. The importance of soluble microbial products
(SMPs) in biological drinking water treatment. Water Res. 2000, 34,
1386–1396.
52. Kunacheva, C.; Stuckey, D. C. Analytical methods for soluble microbial
products (SMP) and extracellular polymers (ECP) in wastewater treatment
systems: A review. Water Res. 2014, 61, 1–18.
53. Rittmann, B. E.; Bae, W.; Namkung, E.; Lu, C. J. A critical evaluation of
soluble microbial products in biological processes. Water Sci. Technol. 1987,
19, 517–528.
54. Potvin, C. M.; Zhou, H. D. Interference by the activated sludge matrix on the
analysis of soluble microbial products in wastewater. Chemosphere 2011, 85,
1139–1145.
55. Boero, V. J.; Bowers, A. R.; Eckenfelder, W. W., Jr Molecular weight
distribution of soluble microbial products in biological systems. Water Sci.
Technol. 1996, 34, 241–248.
339
Chapter 18
Variability of Non-Regulated Disinfection

By-Products in Distribution Systems:
Impact of the Storage Tank
Christelle Legay,1 Patrick Levallois,2

Rocio Aranda-Rodriguez,3 Luda Dabeka,3 Joan Hnatiw,3
and Manuel J. Rodriguez*,1
1Centrede Recherche en Aménagement et Développement, Université Laval,
Pavillon Félix-Antoine-Savard, 2325 rue des Bibliothèques,
Québec City, QC G1V 0A6, Canada
2Direction de la Santé Environnementale et de la Toxicologie,
Institut National de Santé Publique du Québec, 945 avenue Wolfe,

Québec City, QC G1V 5B3, Canada
3Exposure and Biomonitoring Division, Environmental Health Science
Bureau, Health Canada, EHC, Tunney’s Pasture 0800C, bktOttawa, ON

K1A OK9, Canada
In this chapter, the spatial variability of non-regulated

disinfection by-products (DBPs) in drinking water –
haloacetonitriles (HANs), haloacetaldehydes (HAs),
haloketones (HKs), chloropicrin (CP) and cyanogen chloride
(CNCl) – was investigated in four distribution systems in
the Québec City area. Attention was directed to the impact,
on DBP levels, of water that is re-chlorinated in distribution
system storage tanks. The results show that when there was a
storage tank as part of the distribution system, flow through
this led to an increase in targeted DBP levels except for CNCl
for which levels were systematically lowest at sites located
after the storage tank. However, for most DBP families, the
variability within the storage tank was different from month to
month. Moreover, the impact on DBP levels of water flowing
through the storage tank differed between the four systems
under study.

This study highlights that, as for regulated DBPs
(trihalomethanes-THMs and haloacetic acids-HAAs), the
selection of the sampling locations for non-regulated DBP
monitoring represents a considerable challenge.
Introduction
Disinfection by-products (DBPs) in drinking water are known to vary within
distribution systems. Several factors may influence DBP variability, such as raw
water source characteristics, the treatment applied and some distribution system
characteristics (1–3). The variability of DBP levels within systems makes it
difficult to monitor these compounds.
Past studies have investigated the spatial variability of DBPs, but focused
mainly on two prevalent families (trihalomethanes-THMs and haloacetic
acids-HAAs) presently regulated in most industrialized countries. However,
other non-regulated DBP families occurring at low levels in drinking water
are considered as potentially of greater health concern than THMs and HAAs
(4). These include, among others, haloacetonitriles (HANs), haloacetaldehydes
(HAs), haloketones (HKs), and halonitromethanes (HNMs). A number of
studies have focused specifically on the spatial variability of these DBPs within
distribution systems where water is disinfected with chlorine (5–10). In Williams
et al. (5) and Shin et al. (6), the spatial variability of DBPs within several systems
was investigated through comparisons with the levels measured in finished water
at the water treatment plant (WTP) and at only one site located in the system (in
the middle or at the extremity according to the study). In Golfinopoulos et al.
(7), finished water and a large number (eight) of sites were characterized within
one distribution system. However, the variation of DBP levels between these
sites was not really investigated. Koudjonou et al. (8) investigated the spatial
variability of DBP levels within systems by sampling the finished water and water
from three sites spatially distributed within each system, yet only one HA species
was investigated in this study. Guilherme and Rodriguez (9) considered three
sampling sites located in systems under study to investigate the spatial variability
of three DBP families. However, the presence of DBPs in finished water was
not measured. Moreover, it is important to note that in the above studies, the
impact on DBP levels of water flowing through a distribution system storage
tank (defined in this paper as the residence time added by the storage tank) was
not investigated. Mercier-Shanks et al. (10) focused on the spatial variability of
DBPs by sampling various sites spatially dispersed in one distribution system
(which is also studied in this chapter), including two sites located after a storage
tank with re-chlorination. In this study (10), the impact of water flowing through
the storage tank was only briefly discussed, only one system was investigated and
DBP families under study were limited (HANs, HKs and one HNM species).
In this chapter, the spatial variability within systems of non-regulated DBPs
(HANs, HAs, HKs, one HNM species and cyanogen chloride) was investigated
for four distribution systems where water was disinfected with chlorine. The
focus was on the impact, on these DBPs, of water flowing through distribution
342
system storage tanks (where re-chlorination was applied). The variation of this
impact according to season was also investigated. Moreover, the evolution within
distribution systems of the relationship between the levels of regulated (THMs)
and non-regulated DBPs was estimated. To conclude, strategies were discussed
to control and survey these non-regulated DBPs within distribution systems for
monitoring or exposure assessment studies.
Material and Methods

Case under Study
This study was carried out in four distribution systems located in the greater
Québec City area (Province of Quebec, Canada) supplying approximately 350,000
inhabitants. The main characteristics of these systems are presented in Table 1.
These systems are supplied by surface water and all apply chlorine as their primary
or secondary disinfectant. However, the type of surface water source (e.g., lake,
river), type and efficiency of treatment applied before secondary disinfection
(e.g., presence of ozonation or not), and distribution system characteristics (e.g.,
size, hydraulic regime, pipe characteristics) differ between the systems. Each
distribution system includes at least one storage tank with a re-chlorination point.
However, the re-chlorination strategy applied (injection location, dose) and water
residence time in the storage tank vary between the systems (Table 1).
The region under study is subject to important climatic variations during the
year, with mean daily temperatures of ambient air ranging from -16.8°C to +
24.2°C (11), and different lengths of seasons (i.e., long winters and relatively short
summers). These climatic variations can result in great temporal variations in raw
water quality.
Water Sampling
Bimonthly sampling campaigns were conducted between August 2006 and

December 2007 (for logistical issues, data from the February 2007 campaign were
not available). During these campaigns, four water samples were taken in each
distribution system: at the finished water outlet the WTP and at three sites located
within the distribution system. In order to investigate the impact of water flowing
through the storage tank (that includes a re-chlorination step) at least one of these
sites was not supplied by the storage tank and at least one was supplied by the
storage tank.
The DBP measurements included the analysis of HAs, HANs, HKs,
chloropicrin (HNM species and noted CP), cyanogen chloride (CNCl) and THMs
(Table 2). Free residual chlorine (FRC) concentration, temperature (T), pH,
turbidity and ultraviolet absorbance at 254 nm (UV254) were also measured for
each sample.
343
Table 1. Description of the Four Distribution Systems under Study
Systems Water source Main treatment process Storage tank characteristics
Capacity (m3) Estimated water residence Location of

time re-chlorination
A St. Lawrence Riverb Pre-chlorination; 4,550 ~ 1 day Outlet
PCTa; Post-ozonation;
Post-chlorination
B St. Lawrence Riverb Pre-chlorination; PCTa; 2,275 ~ 1.5 days Outlet
Post-chlorination
C Chaudière River Flocculation; Sedimentation; 8,500 ~ 4 days Outlet
Inter-chlorination;
Filtration; Post-ozonation;
344
Post-chlorination
Dc St. Charles Lake PCTa; Post-ozonation; 130,000 ~ 3 to 5 days Entrance and outlet
Post-chlorination
aComplete physical-chemical treatment including sieving, coagulation-flocculation, sedimentation and filtration; b The location of the raw water intake is
different for systems A and B; c This system was studied previously by Mercier-Shanks et al. (10)

In each sampling location (public building, private residence or stores),
samples were taken at the faucet of the restroom (except for the finished water
samples taken at the WTP). Cold tap water was allowed to flow for approximately
five minutes to obtain water from the distribution system and not stagnant water
from the building pipes. Duplicate samples were collected in 60ml glass vials
containing 1 mL of buffer solution (pH 4.1). In the field, 0.2 mL ascorbic acid
solution (0.114M), used as quenching agent, were added to each bottle just before
the sampling. The samples were stored at 4°C until the time of analysis. FRC and
T were measured in situ at the same time as the DBP sample collection. 250 mL
plastic bottles were used to collect and transport samples for laboratory analysis
of the other parameters.
Analytical Procedure
As presented in Table 2, seven HA species (the sum represents HA7), four
HAN species (the sum represents HAN4), two HK species (the sum represents
HK2), CP, CNCl and four THM species (the sum represents THM4) were
analyzed. The analysis of HANs, HAs, HKs, CP, CNCl and THMs was conducted
in the Health Canada laboratory and described previously (12–14). Briefly,
liquid-liquid extraction was performed with MTBE containing internal standards.
The extracts were analyzed using a Varian 3800 gas chromatograph equipped
with dual electron capture detectors. Two chromatographic columns were used: a
DB-5 column (30 m x 0.32 mm id; film thickness 1 m) as the primary column, and
a DB-1 column (30 m x 0.32 mm id; film thickness 1 m) for confirmation (15).
The method detection limit (MDL) associated with each individual DBP under
study is presented in Table 2. Measurements below the MDL were considered as
equal to zero.
Measurements of FRC were conducted using the DPD titrimetric method
(Standard method 4500-Cl-F) with a DR-890 colorimeter from Hach. Water pH
was measured with a Denver instrument AP15 pH/mV/FET meter. Turbidity
was analyzed with a Hach 2100N Turbimeter. UV254 results were obtained by
UV/visible spectrometry at 254nm (Hach DR5000) with 5 cm optical path quartz
cells.
Data Analysis
In order to investigate the spatial variability of non-regulated DBP levels in
the systems under study and, specifically, the impact of water flowing through the
storage tank with re-chlorination, sampling sites were divided into three categories
according to their location in each system. The first category included the finished
water, for which samples were taken at the WTP (noted as “FW” in the tables
and figures). The second category included the sampling sites located within
the distribution system and directly supplied by the WTP, thus not supplied by
a storage tank (noted as “Sup/WTP” in the tables and figures). The third category
included only the sites supplied by a storage tank within the distribution system
(noted as “Sup/S.tank” in the tables and figures). Depending on the system, the
second and third categories might include one or two sampling sites.
345
Table 2. DBPs under Study
DBPs Abbreviations MDL (µg/L)
HA7 (sum of the 7 HAs)
Dichloroacetaldehyde DCA 0.07
Trichloroacetaldehyde TCA 0.06
Bromochloroacetaldehyde BCA 0.05
Dibromoacetaldehyde DBA 0.05
Bromodichloroacetaldehyde BDCA 0.04
Chlorodibromoacetaldehyde CDBA 0.06
Tribromoacetaldehyde TBA 0.10
HAN4 (sum of the 4 HANs)

Trichloroacetonitrile TCAN 0.04
Dichloroacetonitrile DCAN 0.05
Bromochloroacetonitrile BCAN 0.06
Dibromoacetonitrile DBAN 0.05
HK2 (sum of the 2 HKs)
1,1-dichloro-2-propanone DCP 0.07
1,1,1-trichloro-2-propanone TCP 0.06
Chloropicrin CP 0.04
Cyanogen chloride CNCl 0.07
THM4 (sum of the 4 THMs)
Chloroform TCM 0.41
Bromodichloromethane BDCM 0.10
Dibromochloromethane DBCM 0.08
Bromoform TBM 0.09
MDL represents the method detection limit.
As previously mentioned, this chapter focused mainly on the impact of water

flowing through a storage tank with re-chlorination on non-regulated DBPs.
THMs were considered in order to determine the evolution within systems of the
relationship between regulated and non-regulated DBPs. With this aim in mind,
correlation analyses (Spearman rank correlation coefficients-rs) were conducted
(using SPSS Version 22.0) between THM and non-regulated DBP levels and for
each category of sampling sites previously described (i.e., FW, Sup/WTP and
Sup/S.Tank).
346
DBP Occurrence in the Area under Study
The finished water quality data (at the WTP) for the four distribution systems
during the study period are presented in Table 3.
DBP levels measured in each distribution system during the period under
study are presented in Table 4. Irrespective of the system under study, the most
abundant DBP family among those analyzed was HAs. As previously observed by
Koudjonou et al. (8), statistically significant higher HA levels (significance level
ρ<0.05) were found in the three systems using ozonation in the treatment train
(i.e., systems A, C and D). Among HAs, TCA (also known as chloral hydrate)
was found to be the main species and represented, on average, 58% of HA7. TCA
levels up to 24.65 µg/L were found in the distribution systems under study. DCA
and BDCA were also important HA species with a mean contribution to HA7 of
26% and 11%, respectively. BCA and CDBA were detected at very low levels
(with maximum values of 0.44 µg/L and 0.66 µg/L, respectively), while TBA and
BDA were not detected in these systems.
Table 3. Average Characteristics of Finished Water at the WTP for the

Four Systems during the Study Period
Systems T FRC pH Turbidity UV254
(°C) (mg/L) (NTU) (cm-1)
A 13.1(9.4) 1.02(0.18) 7.39(0.24) 0.10(0.03) 0.024(0.006)
B 13.0(9.3) 1.05(0.23) 7.25(0.12) 0.14(0.07) 0.033(0.005)
C 11.6(9.3) 1.37(0.50) 7.53(0.26) 0.24(0.12) 0.028(0.007)
D 13.4(8.3) 1.36(0.12) 7.65(0.13) 0.32(0.11) 0.027(0.007)
Standard deviation values are shown into parenthesis. For each system and parameter n=8
except for UV254 which n=7.
HKs were the second major DBP family detected among those evaluated
(Table 4) with HK2 levels ranging from 0.07 to 16.10 µg/L. TCP represented
60% to 100% of HK2 with levels varying between 0.07 and 15.02 µg/L. DCP was
measured at lower levels with a maximum value of 3.41 µg/L. The same order of
magnitude for HK levels was observed in the past for system D (10).
347
Table 4. Mean DBP Levels (µg/L) Found in the Four Systems during the Study Period
Systems FW HA7 Sup/ FW HAN4 Sup/ FW HK2 Sup/
Sup/ S.Tank Sup/ S.Tank Sup/ S.Tank
WTP WTP WTP

A 6.18 9.23 15.09 1.88 2.61 3.44 3.53 4.85 6.77
B 2.61 4.12 6.97 1.15 1.80 3.25 2.13 2.71 3.52
C 8.68 12.0 13.31 1.40 1.96 2.27 7.17 8.40 8.52
D 2.17 7.81 11.59 0.38 1.43 1.86 1.10 3.19 3.98
Systems FW CP Sup/ FW CNCl Sup/
Sup/ S.Tank Sup/ S.Tank
WTP WTP
348
A 0.13 0.20 0.50 1.62 1.38 0.26

B 0.08 0.09 0.12 0.57 0.79 0.35
C 0.29 0.44 0.53 1.79 0.64 0.24
D 0.04 0.21 0.36 0.75 0.14 0.05
FW represents the finished water at the WTP. Sup/WTP represents the average of the sampling sites directly supplied by the WTP. Sup/S.Tank represents the
average of the sampling sites supplied by the storage tank. For each system and parameter n=32.

HANs were detected in more than 99% of samples although they occurred at
lower levels than HAs and HKs. Measured HAN4 levels (above MDL) ranged,
during the period under study, from 0.07 (system D) to 5.58 µg/L (system A). HAN
levels were in the same order of magnitude relative to those previously measured
in system D (10). DCAN was the main HAN species with a mean contribution to
HAN4 of 92%. With a maximum value of 5.38 µg/L, DCAN levels measured in the
study area were well below the World Health Organization provisional guideline
value of 20 µg/L for this compound (16). The rest of HAN4 was composed mainly
of BCAN with levels not exceeding 0.83 µg/L. In fact, TCAN and DBAN were
rarely detected.
CNCl and CP are detected in more than 93% of samples, but at very low levels.
CP levels (above MDL) varied, in the four distribution systems investigated during
the period under study, between 0.04 and 1.43 µg/L. Relatively similar levels were
found in the past for system D (10). For CNCl, levels ranging from 0.07 to 3.84
µg/L were measured during the study period.
Spatial Variability of Non-Regulated DBP Levels within Distribution

Systems
The mean DBP levels found for each category of sampling sites in each
distribution system are presented in Table 4. As observed in this table, the
mean HA7 level tended to increase between the WTP and sites not supplied by
the storage tank (i.e., directly supplied by the WTP) with an increasing factor
ranging from 1.4 to 3.6 according to the system. However, this increase was only
statistically significant (ρ<0.05) for system D. Previous studies also observed an
increase of TCA (the main HA species) levels with long water residence times
(5, 8). Water flowing through the storage tank (including re-chlorination) also
involved an increase of HA7 levels compared to levels measured at the WTP (not
statistically significant at ρ<0.05 for system C) from 1.5 to 5.3 times according
to the system. This increase was due mainly to the water residence time in pipes
during water transit until the storage tank, the water residence time within the
storage tank (extension of DBP formation reactions) and the chlorine addition at
the outlet of the latter (and the entrance for system D). Differences in the mean
HA7 level between sites not supplied by the storage tank and sites supplied by
the storage tank were only statistically significant for systems A and B. For these
two systems, the mean FRC concentrations measured at sites supplied by the
storage tank were lower than those observed for systems C and D (Table 5). It is
important to note that the information about chlorine dose at the storage tank was
not available for the majority of the systems under study.
As observed for HA7, the mean HK2 levels also varied within the systems
(Table 4). In systems A, B and C, the mean HK2 level measured at sites not
supplied by the storage tank was from 1.2 to 1.4 times higher (only statistically
significant at level ρ<0.05 for system A) than the mean level measured at the WTP.
In the case of system D, the mean HK2 level measured at sites not supplied by the
storage tank was 2.9 times higher than the level measured at the WTP (statistically
significant at level ρ<0.05). Water flowing through the storage tank also impacted
HK2 with a mean level after the storage tank being 1.2 to 3.6 times higher (not
349
statistically significant at level ρ<0.05 for system C) than the level measured at the
WTP. The system A was the only system for which the HK2 mean level measured
after the storage tank was significantly (level ρ<0.05) different from the level
measured at sites not supplied by the storage tank. Figure 2 demonstrates that the
pattern of spatial variability of HK2 is not representative of the pattern observed
for each individual HK species. In fact, as observed in previous studies (5–7, 10),
the mean DCP level tended to decrease after water flowing through the storage
tank (only statistically significant at level ρ<0.05 for systems B and C). These
results were probably due to the decomposition of DCP after the consumption of
the majority of HK precursors into other compounds (10, 17). The mean TCP
level increased between the WTP and sites not supplied by the storage tank (only
statistically significant at level ρ<0.05 for systems A and D) and also between WTP
and sites located after the storage tank (not statistically significant at level ρ<0.05
for system C) (Figure 2).
Table 5. Spatial Distribution of FRC Concentrations in the Four Systems

during the Study Period
Systems FW Sup/WTP Sup/S.Tank
A 1.02(0.80-1.32) 0.52(0.08-0.87) 0.20(0.04-0.42)
B 1.05(0.74-1.28) 0.71(0.27-1.47) 0.56(0.05-1.66)
C 1.37(0.61-1.92) 0.73(0.28-1.34) 0.73(0.00-1.74)
D 1.36(1.13-1.46) 0.73(0.22-1.21) 0.91(0.36-1.38)
FW represents the finished water at the WTP. Sup/WTP represents the average of the
sampling sites directly supplied by the WTP. Sup/S.Tank represents the average of the
sampling sites supplied by the storage tank. The range of FRC values measured in each
system during the study period is shown into parenthesis.
The mean HAN4 level tended to increase within systems at sites not supplied
by the storage tank (Table 4) with levels being 1.4 and 3.8 times higher than
the level measured at the WTP. However, this increase was only statistically
significant (ρ<0.05) for system D. The increase of HAN levels related to the
water residence time was observed previously in system D (10) and in other
studies (5, 6). The water flowing through the storage tank resulted in a mean
HAN4 level of, depending on the system, 1.6 to 4.9 times higher than the level
measured at the WTP (not statistically significant at level ρ<0.05 for system
350
C). Mercier-Shanks et al. (10) highlighted an increase of HAN4 levels in water
flowing through the storage tank, but found a decrease of TCAN levels during
summer for system D. Differences in the mean HAN4 levels measured at sites
not supplied by the storage tank and those located after the storage tank were
only statistically significant (ρ<0.05) for system B. The absence of ozonation at
WTP in system B could partially explain the higher mean HAN4 level compared
to the other systems (18).
Figure 1. Spatial distribution of individual HK levels in the four systems under

study. Error bars represent standard deviation values. FW represents the finished
water at the WTP. Sup/WTP represents the average of the sampling sites directly
supplied by the WTP. Sup/S.Tank represents the average of the sampling sites
supplied by the storage tank.
351
As observed in previous studies (5, 6, 10), the mean CP level increased with
the water residence time (Table 4). The spatial variability was only statistically
significant (ρ<0.05) for systems A and D. Again, the absence of ozonation in
system B might partially explain the low formation and spatial variations of CP
within this system (19). For systems A and D, the impact of water flowing through
the storage tank on CP levels was relatively high with an increase of the mean level
(compared to the measured level at the WTP) of 9.0 and 3.8 times, respectively. It
is important to remember that CP was measured at very low levels in the systems
under study.
Contrary to the other DBP families, the mean level of CNCl decreased within
systems (Table 4) due to its decomposition (20). For system B, the mean CNCl
level appeared to increase at sites not supplied by the storage tank (not statistically
significant at level ρ<0.05) and to decrease (statistically significant at level ρ<0.05)
at site supplied by the storage tank. In the other systems, a decrease of the mean
CNCl level was observed at sites not supplied by the storage tank (not statistically
significant at level ρ<0.05 for system A). This increase was relatively important
for systems C and D. Irrespective of the system, the mean CNCl level was lowest
(statistically significant at level ρ<0.05) after the storage tank.
Temporal Variability of the Impact of Water Flowing through the Storage

Tank
As previously discussed, DBP levels varied within the distribution systems
under study and their fate was affected by the presence of a storage tank (with
a re-chlorination step). The influence of the sampling period (i.e., months) on
the impact of water flowing through the storage tank was investigated more
specifically in this section. It is important to remember that the data from the
February campaign were not available.
Regardless of the sampling month, the HA7 levels measured after the storage
tank were higher than the levels measured at the WTP (Figure 2) except for August
2006 in system C where one of the two sites located after the storage tank had
levels below this measured at WTP. However, this increase of HA7 levels in water
that has flowed through the storage tank also varied throughout the year and was
generally higher in summer and fall (in terms of change in level between WTP
and after the storage tank and not in terms of impact factor). These results could
partially be associated with the increase of reaction kinetics in warmer waters (21,
22) and the seasonal variation of the nature and quantity of organic precursors
(21). In the case of system C (Figure 2), the HA7 level measured at one of the
two sites located after the storage tank was lower than what was measured at the
WTP during August 2006. It was also lower than the levels measured at the site
not supplied by the storage tank during October and December 2006, and August
2007. This decrease of HA7 levels was most likely associated, in part, with the
absence of FRC at this site (FRC concentration was below 0.05 mg/L for five of
the eight campaigns) might be due to a long residence time after the storage tank
for water. In fact, Koudjonou et al. (8) highlighted the potential transformation of
TCA to chloroform in warm waters associated with long residence times.
352
353
Figure 2. Temporal variations of spatial distribution of HA7 levels for systems A, B, C and D. Sup/WTP-1 and Sup/WTP-2 represent the
individual sampling sites directly supplied by the WTP. Sup/S.Tank-1 and Sup/S.Tank-2 represent those supplied by the storage tank.

354
Figure 3. Temporal variations of spatial distribution of HK2 levels for systems A, B, C and D. Sup/WTP-1 and Sup/WTP-2 represent the

355
Figure 4. Temporal variations of spatial distribution of HAN4 levels for systems A, B, C and D. Sup/WTP-1 and Sup/WTP-2 represent the

356
Figure 5. Temporal variations of spatial distribution of CP levels for systems A, B, C and D. Sup/WTP-1 and Sup/WTP-2 represent the

357
Figure 6. Temporal variations of spatial distribution of CNCl levels for systems A, B, C and D. Sup/WTP-1 and Sup/WTP-2 represent the

The temporal variations of the impact of water flowing through the storage
tank on the HK2 level are illustrated in Figure 3. For systems A, C and D, the
increase of the HK2 level between the WTP and sites located after the storage tank
was lower in August, in terms of level and impact factor than the other months,
even with a decrease of the HK2 level for one of the two sites of system C (for
which the FRC concentration was generally very low). Moreover, for systems A,
C (in this case, for the second site located after the storage tank) and D, the HK2
levels measured at a the sampling point after the storage tank were similar to or
below those measured at some sites not supplied by the storage tank during several
periods of the year. These results might be due mainly to the degradation of TCP
(presents at higher levels than DCP in the study area) at longer water residence
times (associated with flow through the storage tank) in water warm conditions
(10, 17). In the case of system B, water flowing through the storage tank resulted
systematically in an increase of HK2 level compared to levels measured at the

WTP.
Increases in HAN4 levels associated with water flowing through the storage
tank were observed during all sampling campaigns except in August 2006 for
system C (Figure 4). For this system, a decrease was observed at one of the two
sites located after the storage tank. HAN4 levels close to or below those measured
at some sites not supplied by the storage tank were observed after the storage tank
for systems A, C and D during August months. Mercier-Shanks et al. (10) did
not report a decrease of DCAN and BCAN (representing the main HAN species
detected in the systems under study) after the storage tank. However, the likelihood
of dihaloacetonitriles being degraded with increased time and at higher pH was
reported previously (2, 23). For system B, the levels measured at sites located
after the storage tank were systematically the highest (Figure 4).
The impact of water flowing through the storage tank on CP levels also varied
between sampling campaigns in the distribution systems investigated (Figure 5).
CP levels measured after the storage tank were higher than those measured at the
WTP except in systems B and C (in the latter, for only one of two sites located
after the storage tank) during summer. In general, it is relatively difficult to define
a temporal pattern of variability for the impact of water flowing through the storage
tank on CP.
Regardless of the sampling campaign, CNCl levels measured after water
flowing through the storage tank were lower than those measured at the WTP
(Figure 6).
Relationship between Regulated and Non-Regulated DBPs within Systems

Correlations between THMs and non-regulated DBPs obtained for each
category of sampling sites (FW, Sup/WTP and Sup/S.Tank) are presented in
Table 6. Results show that in the finished water at the WTP, THM4 were very
highly correlated (positively) with HAN4. THM4 were also strongly correlated
with HA7 and HK2 and more moderately with CP. For sites directly supplied by
the WTP, the correlations between these non-regulated DBPs and THM4 were
relatively similar (Table 6). In the case of sites supplied by the storage tank,
results show that the correlations between THM4 and, HA7, HAN4, HK2 and CP
358
were lower than those observed for finished water and for sites directly supplied
by the WTP. This decrease in correlations was especially pronounced for HAN4
and HK2. Whatever the location within the system, THM4 and CNCl were not
correlated (not statistically significant at level ρ<0.05).
Previous studies have produced both comparable and contrasting correlations
between THM4 and non-regulated DBPs (10, 24–26). However, in these studies
the spatial variability of correlations associated with the presence of a storage tank
was not investigated.
Table 6. Spearman Rank Correlation Coefficients (rs) between THM4 and

Non-Regulated DBP Levels within Systems
Location HA7 HAN4 HK2 CP CNCl

FW 0.81** 0.98** 0.83** 0.67** -0.12
Sup/WTP 0.79** 0.91** 0.83** 0.57** -0.05
Sup/S.Tank 0.70** 0.62** 0.53** 0.51** -0.19
FW represents the finished water at the WTP. Sup/WTP represents the sampling sites
directly supplied by the WTP. Sup/S.Tank represents the sampling sites supplied by the
storage tank. Spearman rank correlation coefficients were calculated with DBP data
obtained for the four systems and for the entire period under study. ** Correlation
statistically significant at level ρ<0.01.
Control and Monitoring of Non-Regulated DBP Levels
This study highlights that non-regulated DBPs varied significantly within the
investigated systems. Except for CNCl, higher levels were generally measured
at sites located after the storage tank where chlorine was added. However, the
impact of water flowing through the storage tank, with a re-chlorination step, on
the level of the investigated DBP differed during the year and between systems.
This might be explained by the fact that several factors influence the formation
and degradation of DBPs after water flowing through a storage tank. One possible
factor is the water quality supplying the storage tank which, in turn, depends
on the quality of treated water from the WTP (e.g., pH, amount and nature of
natural organic matter, residual chlorine) and on the distribution conditions (e.g.,
water residence time, age and materials of pipes, presence of biofilm). The
hydraulic characteristics of the storage tank (on which the water residence time
is dependent) and the re-chlorination factors – chlorine dose and the injection
location in the storage tank (at the entrance or/and at the outlet) – also affect DBP
formation and degradation within and downstream of the storage tank. Storage
tank management strategies may include optimization to minimize DBP levels
within the distribution system. For this, further investigations focusing more
specifically on the study of storage tank impact on DBPs should be conducted
in order to compare water quality between the storage tank entrance and points
359
located immediately before and after re-chlorination at different periods of the
year and with variable hydraulic regimes. Since storage tank characteristics
and management differ from one distribution system to another, generalizing
(for all systems) the impact of water flowing through a storage tank, where
re-chlorination is done, on DBPs is not recommended.
As observed for regulated DBPs (e.g., THMs and HAAs), the high spatial
variability of non-regulated DBP levels in distribution systems results in
difficulties in selecting sampling locations for regulatory monitoring, managing
issues and exposure assessment purposes. Except for CNCl, sites located after the
storage tank may be selected for future regulatory monitoring of the DBPs studied
in this chapter. Moreover, results suggest that sites where FRC concentration is
very low (particularly in summer) do not necessarily represent the best locations
for these regulatory survey purposes. In the case of exposure assessment studies,
results presented in this chapter demonstrate that one sampling site for a system
is probably not sufficient to estimate the public’s exposure to these DBPs. This is
exacerbated by the fact that DBP levels vary, according to the season and system,
between sites supplied by the same water infrastructure (i.e., directly by the WTP
or by the storage tank). Results also showed that THM levels could be used as
a potential indicator of some non-regulated DBPs. However, the efficiency of
THMs as indicator for other DBPs tended to decrease after water flowing through
the storage tank where re-chlorination was applied.
Acknowledgments
The authors extend their gratitude to the Drinking Water Research Chair
of Université Laval (Quebec City, Canada) and its partners (NSERC, Cities of
Québec, Levis and Saint-Jérôme, Avensys Solutions, SNC-Lavalin, Association
pour la protection de l’environnement du lac Saint-Charles et des Marais du Nord
(APEL)). The authors are also grateful to Sabrina Simard, Sylvie Leduc, Sonia
Poulin and Annick Dion-Fortier for their laboratory technical support and to all
those who participated in the collection of samples.
References
1. Rodriguez, M. J.; Serodes, J.; Levallois, P. Behavior of trihalomethanes and
haloacetic acids in a drinking water distribution system. Water Res. 2004,
38, 4367–4382.
2. Hua, G. H.; Reckhow, D. A. DBP formation during chlorination and
chloramination: Effect of reaction time, pH, dosage, and temperature. J. –
Am. Water Works Assoc. 2008, 100, 82–95.
3. Krasner, S. W. The formation and control of emerging disinfection
by-products of health concern. Philos. Trans. R. Soc., A 2009, 367,
4077–95.
M. Occurence, genotoxicity, and carcinogenicity of regulated and emerging
360
5. Williams, D. T.; LeBel, G. L.; Benoit, F. M. Disinfection by-products in
Canadian drinking water. Chemosphere 1997, 34, 299–316.
6. Shin, D.; Chung, Y.; Choi, Y.; Kim, J.; Park, Y.; Kum, H. Assessment of
disinfection by-products in drinking water in Korea. J. Exposure Anal.
Environ. Epidemiol. 1999, 9, 192–199.
7. Golfinopoulos, S. K.; Nikolaou, A. D.; Lekkas, T. D. The occurrence of
disinfection by-products in the drinking water of Athens, Greece. Environ.
Sci. Pollut. Res. 2003, 10, 368–372.
2008, 72, 875–881.
9. Guilherme, S.; Rodriguez, M. J. Occurrence of regulated and non-regulated

disinfection by-products in small drinking water systems. Chemosphere
2014, 117, 425–432.
10. Mercier-Shanks, C.; Sérodes, J-B.; Rodriguez, M. J. Spatio-temporal
variability of non-regulated disinfection by-products within a drinking water
distribution network. Water Res. 2013, 47, 3231–3243.
11. Environnement Canada. Normales et moyennes climatiques
au Canada 1971-2000. Service météorologique du Canada;
http://climate.weather.gc.ca/climate_normals/index_f.html (date accessed:
11 December 2013)
12. LeBel, G. L.; Williams, D. T. Assessment of a method, optimized for
cyanogen chloride, for the analysis of Method 551 target DBP compounds.
In Proc. – Water Qual. Technol. Conf. Expo., Boston, MA, 1996.
13. LeBel, G. L.; Benoit, F. M. Chloral hydrate in Canadian drinking water. In
Proc. – Water Qual. Technol. Conf. Expo., Salt Lake City, UT, 2000.
14. Koudjonou, B. K.; Lebel, G. L. Halogenated acetaldehydes: Analysis,
15. Aranda-Rodriguez, R.; Koudjonou, B.; Jay, B.; Lebel, G. L.; Benoit, F. M.
Disinfection by-products (DBPs) in drinking water from eight systems using
chlorine dioxide. Water Qual. Res. J. Can. 2008, 43, 11–22.
16. World Health Organization (WHO). Guidelines for drinking-water quality,
4th ed.; World Health Organization: Geneva, Switzerland, 2011.
17. Nikolaou, A. D.; Lekkas, T. D.; Kostopoulou, M. N.; Golfinopoulos, S.
K. Investigation of the behaviour of haloketones in water samples.
Chemosphere 2001, 44, 907–912.
18. Hua, G.; Reckhow, D. A. Effect of pre-ozonation on the formation and
speciation of DBPs. Water Res. 2013, 47, 4322–30.
19. Hu, J.; Song, H.; Karanfil, T. Comparative analysis of halonitromethane and
trihalomethane formation and speciation in drinking water: the effects of
disinfectants, pH, bromide, and nitrite. Environ. Sci. Technol. 2010, 44,
794–799.
20. Na, C. Z.; Olson, T. M. Stability of cyanogen chloride in the presence of free
chlorine and monochloramine. Environ. Sci. Technol. 2004, 38, 6037–6043.
361
21. Singer, P. Control of disinfection by-products in drinking water. J Environ.
Eng. 1994, 120, 727–744.
22. Hong, H.; Xiong, Y.; Ruan, M.; Liao, F.; Lin, H.; Liang, Y. Factors affecting
THMs, HAAs and HNMs formation of Jin Lan Reservoir water exposed to
chlorine and monochloramine. Sci. Total Environ. 2013, 444, 196–204.
23. Reckhow, D. A.; Platt, T. L.; MacNeill, A. L.; McClellan, J. N. Formation
and degradation of dichloroacetonitrile in drinking waters. J. Water Supply:
Res. Technol.−AQUA 2001, 50, 1–13.
24. Villanueva, C. M.; Castano-Vinyals, G.; Moreno, V.; Carrasco-Turigas, G.;
Aragones, N.; Boldo, E.; Ardanaz, E.; Toledo, E.; Altzibar, J. M.; Zaldua, I.;
Azpiroz, L.; Goni, F.; Tardon, A.; Molina, A. J.; Martin, V.; Lopez-Rojo, C.;
Jimenez-Moleon, J. J.; Capelo, R.; Gomez-Acebo, I.; Peiro, R.; Ripoll, M.;
Gracia-Lavedan, E.; Nieuwenhujsen, M. J.; Rantakokko, P.; Goslan, E. H.;
Pollan, M.; Kogevinas, M. Concentrations and correlations of disinfection

by-products in municipal drinking water from an exposure assessment
perspective. Environ. Res. 2012, 114, 1–1.1.
25. Gan, W.; Guo, W.; Mo, J.; He, Y.; Liu, Y.; Liu, W.; Liang, Y.; Yang, X.
The occurrence of disinfection by-products in municipal drinking water in
China’s Pearl River Delta and a multipathway cancer risk assessment. Sci.
Total Environ. 2013, 447, 108–115.
26. Wei, J. R.; Ye, B. X.; Wang, W. Y.; Yang, L. S.; Tao, J.; Hang, Z. Y. Spatial
and temporal evaluations of disinfection by-products in drinking water
distribution systems in Beijing, China. Sci. Total Environ. 2010, 408,
4600–4606.
362
Chapter 19
Chloral Hydrate Control by Point-of-Use

and Household Appliances
Baiyang Chen,* Xiaoqi Guo, Zhong Tang, and Wenbiao Jin
Harbin Institute of Technology Shenzhen Graduate School,

Shenzhen Key Laboratory of Water Resource Utilization and Environmental
Pollution Control, Shenzhen, China, 518055
*E-mail: [email protected]. Phone: 86-134-80727605.
Unlike other pollutants occurring in raw water, disinfection

by-product (DBP) is usually produced at the end point of the
drinking water treatment plant (DWTP) and, once formed,
it cannot be readily removed by engineering processes. As
a result, treatments of existing DBP by point-of-use and/or
household appliances have become the last line of defense in
alleviating the impact of DBP for general family. In this study,
we evaluated the effectiveness of several residential options,
including reverse osmosis (RO) and granular activated carbon
(GAC) cartridges, microwave oven, boiler, and ultrasonic
cleaner, on the removal of chloral hydrate (CH) under various
operating (e.g., power, stirring speed) and environmental (e.g.,
pH, initial concentration) conditions. The results indicate that
heating by either boiler or microwave oven can reduce CH
from tap water significantly (>90%) under automatic switch-off
conditions. The degree of removal by heating was always
greater in tap water than in ultrapure water, implying that certain
compounds or residual chlorine in tap may have accelerated
the CH transformation process, while CH removal in ultrapure
water is mainly controlled by thermal hydrolysis. In contrast,
volatilization by stirring or sonication exhibited little capacity
(<5%) to remove CH. RO cartridge eliminated >90% of CH
regardless of operating pressure, initial CH concentration, pH,
and type of water, proving it as a robust tool in dealing with
drinking water issues. Cartridges with GAC showed some
potentials (45~90%) for CH removal, which were much greater

than cartridges made by PP cotton and carbon block filter
(<5%); however, the adsorption ability of GAC can be highly
compromised by rapid flowrate and limited retention time,
which suggests that GAC adsorption is not the major contributor
to CH removal in a commercial water purifier. Overall, the data
have proven the effectiveness of many POU and household
appliances on CH removal, and may help consumers to relieve
DBP concerns in case of emergency.
Introduction
Chloral hydrate (CH) is the third most prevalent disinfection by-product

(DBP) detected in water, after trihalomethane (THM) and haloacetic acid (HAA).
Its occurrence has been widely reported around the world at levels up to 100
μg/L, but mostly below 10 μg/L (1–3). Due to its potential toxicity related to
DNA damage (4), carcinogenicity (5) and a series of health risks (6), CH has
been regulated by the Chinese government (GB5749-2006) and is proposed to
be regulated in Austria (7) and the World Health Organization guideline (8).
The maximum contaminant level (MCL) is set as 10μg/L in China and by the
WHO, and proposed to be 20 μg/L in Austria. While accidents involving CH
are not frequently reported, around 10% of finished waters from US drinking
water treatment plants (DWTPs) have CH levels greater than 10μg/L under the
Information Collection Rule (9), suggesting that CH may pose a risk in some
regions and thus deserves more attention in future DBP studies.
In early studies, the control methods to remove CH from drinking water
mostly focused on the removal of DBP precursors (8, 10–12), but not existing
DBPs (i.e., preformed DBPs) (13). Although some post-formation methods
have been evaluated, such as γ-irradiation (14), solar photolysis (15, 16),
synthetic goethite and magnetite (17), and electrochemical reduction (18), they
are somewhat difficult to apply in distribution systems or at the point of use
(POU). Only a few studies have examined the effectiveness of household devices,
such as boiling pot, on CH controls (19, 20); however, the boiling time in those
experiments often took more than 1 minute to achieve the desired result, which is
perhaps impractical because most commercial boilers are now equipped with an
automatic switch-off function, meaning that the boiling effect may last for only
seconds under normal operating conditions.
Of all drinking water DBP control strategies, treating DBP after its formation
by POU and household facilities has never been the favorite option. Rather,
removing DBP precursors from raw water before disinfection (12) and/or altering
disinfectant agents and operating conditions (8) has often been applied in practice
to meet the regulatory limit. The preference of pre-formation measures is likely
related to the cost and efficiency advantages of DWTP facilities. However,
the ability of POU and residential appliances to control DBP should not be
ignored or deemed unimportant, as it may help consumers reduce the risk of
DBPs themselves. Moreover, in epidemiology studies, such information may
364
help determine the actual intake levels of people with different drinking habits,
and accordingly enable better understanding of the causes of DBP exposure and
toxicity.
In light of the abovementioned need and knowledge gap, this study aims to
evaluate the effectiveness of several POU and household appliances in removing
CH under a series of operational and environmental conditions. The POU systems
include a water purifier system composed of a series of cartridges: polypropylene
(PP) cotton, granular activated carbon (GAC), carbon block filter (usually
identified as CTO), and reverse osmosis (RO) in sequence. The household
appliances include a boiler, a microwave oven, a mixer, and an ultrasonic cleaner.
Some factors that may affect the CH removal efficiency were examined, including
the initial concentration of CH, types of water (ultrapure vs. tap vs. lake), pH,
stirring speed, power, capping conditions, etc. The roles of mechanisms on
CH removal, such as volatilization and hydrolysis, were discussed to provide a

fundamental understanding of these treatment processes.

Samples and Chemicals
The CH used in this study was purchased as powder (AR grade, >99% in
purity) instead of dissolved in solvent to avoid potential interference of the solvent
(21). Besides CH, the methyl-tert-butyl ether (MTBE) used for CH extraction
and phosphorous buffer used for pH control were purchased from Aladdin Inc. at
analytical grade (>99.9% in purity). Prior to experiments, stock CH solution was
prepared by ultrapure water at a concentration of 2g/L and stored in a freezer at
4°C temperature if not used immediately. In all tests, control experiments were
conducted.
The water samples were obtained from the city of Shenzhen, China, including
lake water taken from Xili Lake, a drinking water source of the city; tap water from
the laboratory; and ultrapure water produced on site by a Millipore water generator
(Direct-Q3) with UV sterilizer. The lake and tap samples were filtered through a
0.45 µm glass fiber filter (Xingya, Co., Ltd) before use, and some water qualities
are listed in Table 1. The selection of raw water for tests was intended to examine
the potential influences of NOM and ions on CH treatability, although in practice it
is impossible to have CH in raw water. Similarly, a comparison of ultrapure water
and tap water may help better understand the CH treatment differences between
ideal and real conditions.
Analytical Methods
CH was detected by a GC equipped with a ECD detector (GC-9720, Fuli,
China) according to a modified EPA method 551.1, with the method detection
limit (MDL) of 0.35 µg/L. Briefly, CH extraction was carried out for a 25ml
sample by applying 5g of NaCl and 3ml of MTBE, shaking for 1 minute on a
lab-dancer (IKA, Gemany), and then letting it stand for 20 minutes. The GC
column was initially set to be 60 °C for 3 minutes and then ramped up to 150°C
365
with a 20°C/min rate, and maintained the temperature for 1 minute; the flowrate of
carrier gas (99.999% nitrogen in purity) was 2ml/min, and the temperature of ECD
detector was set at 280°C. Residual chlorine was analyzed by the DPD method
using a spectrophotometer (Hach 3900, USA) according to EPA 330.5. Chloride
was analyzed by an ion chromatography instrument (IC2010, Tosoh Inc., Japan)
with a MDL of 1 µg/L. Conductivity was recorded using an electrode (SX-650,
San-Xin Instrumentation, Inc., China), and pH was analyzed using an electrode
meter (pH100, Extech Instruments Corporation, USA). Dissolved organic carbon
(DOC) was assessed by a TOC analyzer (TOC-LCPH, Shimadzu, Japan) according
to the EPA 415.3 method.
Table 1. Characteristics of Waters in This Study

COND (μs) TDS (mg/L) TOC (mg/L)
Ultrapure Water 0.056 <0.26 <0.1
Tap Water 82 58.0 1.230
Lake Water 94.5 65.7 1.627
Apparatus
A series of domestic appliances were used in this study for CH removal

assessment. In general, they are commercially available, domestic devices
that can be readily purchased from most shopping malls in China. The boiler
(BQ-150GA, Shenyi electronics, Inc., China) has an automatic switch-off function
commonly used by owners, and it takes around 4 minutes and 10 seconds to reach
100°C from ambient temperature. The selected stirring speeds (1500 and 3000
RPM) of the mixer (CJ78-1, Jintan automatics, Inc. China) simulates a blending
process of water and flour, but it is not a grinding machine. Although similar
types of electronics are available in the market, such as a juicer and an automatic
soybean milk machine, they are often operated under capped conditions and
less likely to volatilize organic pollutants, and hence not selected in this study.
The POU water purifier system (50GPD, Hercon Inc., China) consists of four
major treatment units installed in sequence: PP cotton, GAC, CTO, and RO. For
evaluation purpose, only one cartridge was used at a time and the other three
cartridges were taken out to distinguish the efficiencies of each type of cartridge.
Experiments using a microwave oven (P70OF20CL, Galanz Inc., China, with a
frequency of 2450 MHz and adjustable power) and ultrasonic cleaner (KQ2200,
Kunshan Ultrasonic Instrument Co., Ltd, 40 kHz, with various power inputs
of 30W/50W/150W) were conducted in both capped and uncapped conditions.
These two appliances are not designed to treat water, but if necessary, they may
be employed to deal with polluted waters. To ensure the quality result, all tests
were conducted in duplicate.
366
CH Treatment by Reverse Osmosis
RO is known to be a powerful tool for eliminating many types of pollutants,

mainly because of the size difference between its membrane and target compounds.
A recently study evaluated the rejection efficiency of RO on CH, and found that the
removal degree of CH can reach almost 100%, regardless water temperature (23 -
35°C) (13). The tests were conducted at an operating pressure of 0.35 MPa, with
the initial CH concentration at 50 μg/L. The spiked CH was initially dissolved
in methanol solvent (13). Because the membrane and operating conditions are
different from a household RO unit, we have done the following experiments to
supplement above knowledge.
Firstly, considering the importance of pressure, we compared the

transmembrane fluxes and CH removal potentials for a range of operating
pressures (Table 2). With the increase of pressure, water recovery rate increased
and wasted retentate decreased, while the removal efficiency (>90%) of CH
appeared independent of the operating pressure (Figure 1a). So, by balancing the
concerns of energy and efficiency, we selected 0.6 MPa in the subsequent tests.
From a practical point of view, the tap pressure in a family house is normally
less than 0.5 MPa and therefore cannot provide enough pressure to generate
a sufficient amount of water within a limited time; however, a commercial,
residential RO system is usually accompanied by a small-scale booster pump
to provide extra pressure, so consumers often do not need to worry about the
pressure issue unless the membrane becomes severely fouled.
Table 2. Effect of Operating Pressure on RO Flowrate

Pressure(M- Permeate rate Retentate Recove- Overall flowrate
Pa) (ml/s) rate(ml/s) ry (%) (ml/s)
0.4 2 20 9% 22
0.6 3 9 25% 12
0.8 4 4 50% 8
The acid-dissociation constant (pKa) of CH is predicted to be 9.51 by

ChemAxon (13), meaning that CH will exist in ionic form at a pH value greater
than 9.51, and consequently, it may exhibit different treatability by the RO
process. However, this study exhibited no differences among waters at pH of
10.0 (Figure 1b), indicating that RO treatment of CH is mainly reliant upon the
pressure-driven ability of the membrane rather than interactions of charged solute
and membrane. However, since this test was conducted on ultrapure waters
fortified with CH, the potential formation of precipitative solid by metals (e.g.,
calcium and magnesium) in real water at high pH condition may not be well
reflected.
367
368
Figure 1. Effects of a) operating pressure, b) pH, c) initial concentration, and d)
water type on CH removal by a domestic RO cartridge (working condition: 0.6
MPa, Co=20 µg/L, pH=7, ultrapure water).
The influence of initial concentration had not been evaluated before, despite
the fact that CH may exist in a wide range of concentration in DWTPs. In this study
we compared the effect of two initial concentrations on CH removal. The results
showed that there were no noticeable CH concentration differences in permeates,
indicating that RO efficiency is independent of initial CH concentration (Figure
1c).
The presence of interfering compounds in tap and lake water, however, may
impose an adverse effect on CH removal. Figure 1d shows that CH removal was
lowered to 83% in lake water and such efficiency is less than those in ultrapure
(95%) and tap water (94%), suggesting that a competing phenomenon from
natural organic matter (NOM) may occur in raw water and treatment of NOM,
e.g., coagulation, is necessary in practice for RO to achieve better performance.
CH Treatment by Adsorptive Materials
CH has a medium level of octonal-water partition coefficient (Kow =0.99) so

that it exhibits some levels of hydrophobicity in water. As a result, adsorption of
CH by adsorptive materials, especially GAC, had been proven feasible by several
early studies. However, the effectiveness of GAC on CH appeared inconsistent in
the literature: some achieved complete removal of CH (22), and others observed
varying levels of removal as a function of materials (23).
In this study, we evaluated three types of adsorption materials, i.e., GAC, PP
cotton, and CTO, on CH removal (Figure 2). The results indicate that PP cotton
and CTO have little capacity for CH control for a range of flowrates, suggesting
that both materials have no size-excluding or adsorptive effects on CH. The major
purpose of their uses in a commercial water purifier is probably to resist particles
in the water. CTO is named after carbon for taste and odor, and the materials
consist of activated carbon and polymers; however, the adsorptive capacity of
activated carbon concreted in polymer seems to be minimal in practice and it is
therefore useless for DBP control. In contrast, GAC demonstrated a considerable
level of treatment ability (44~94%) for CH, with the efficiency increasing with the
decrease of flowrate of water (Figure 2a). The study of extremely low flowrate
(20ml/min, corresponding to a retention time of 16 minute) agreed well with an
earlier study that applied GAC in a gravity filter (23), indicating that the GAC
material is capable of removing most CH if sufficient contact time is allowed.
However in reality, for the purpose of treating an adequate amount of water, the
capability of GAC can be highly compromised.
369
Figure 2. The effects of flowrate (a), adsorption material, and water type (b) on
CH removal by domestic adsorptive cartridges (normal conditions: 1.5L/min,
Co=20 µg/L, pH=7).
Figure 2b shows the CH removal efficiency by virgin adsorbents at different

waters. The treatment levels were independent of the type of water. With a
combination of GAC and RO, the water purifier reduced an influent of 20 μg/L
CH to an effluent of ~1 μg/L (Figure 3), which is far below the regulatory MCL
(i.e., 10 μg/L), and therefore safeguards the water quality well with considerable
confidence. Overall, the use of commercial water purifiers equipped with an RO
unit is highly recommended in reducing the risks of CH from drinking water.
370
Figure 3. Removal of CH by RO and GAC alone or in combination.
CH Treatment by Boiler and Microwave Oven
Boiling water with pot or kettle has been a traditional way in many countries
to inactivate pathogens or to make coffee and tea for daily life. Its roles and
mechanisms in DBP removal, however, are not yet well-known although it has
been proven robust in eliminating many types of DBPs (19, 20). For example,
the results of Wu et al. reported non-detectable CH level after 1-min boiling of
many types of waters originating from distilled water and tap water and in the
presence of hydrophobic acid and bromide (19), indicating a complete removal of
CH by the boiling process. Additionally, the results of Krasner et al. confirmed
such efficiency in real samples, either chlorinated or chloraminated, with >97% of
CH removals. These studies, however, did not consider the automatic switch-off
function of boilers commonly used by the public. Since a boiler stops heating
immediately, which often takes seconds, a more practical evaluation of the heating
process (i.e., <1 min) appears necessary. In order to avoid the interference of
concurrent formation of DBPs (19) during the transformation process of CH, this
test evaluated the boiling effect on CH removal from both ultrapure and tap water.
Figure 4 shows the effectiveness of a boiler on the CH removals for two
types of water. Notably, the efficiency of CH removal in tap water (with residual
chlorine of 0.15 ppm) is significantly higher than that in ultrapure water (no
371
chlorine residual). Such phenomena were also observed in the microwave oven
heating process (Figure 5a). Along with an increasing ratio of tap water to
ultrapure water, the CH removal increased from 21% to 99%. In contrast, because
the addition of sodium chloride (at 1000ppm) added no benefit to CH removal, the
salinity of water is invalidated as an important factor in CH removal during the
heating process (Figure 5a). Early literature suggested that residual chlorine and
chloramine may facilitate the degradation of halogenated aldehydes (24) in water;
and the presence of bromide can facilitate removal of CH (19), more in-depth
study appears necessary to clarify this issue. In this context, we evaluated the
effectiveness of free chlorine on CH degradation at ambient room temperature
(Figure 6). Enhanced degradation of CH was observed with the increase of
residual chlorine concentration. Combined with the information reported from a
recent study that free chlorine can be more rapidly consumed in hot water than in
cold water (25), we postulate that CH is likely subject to expedited degradation

by chlorine under elevated temperatures.
Figure 4. Removal of CH by boiling of spiked waters (Co=20 µg/L, half-full

volume in kettle, 4’10” duration, and automatic switch-off).
372
Figure 5. The effects of water type (a), pH (b), and initial concentration (c) on
CH treatment by a domestic microwave oven (Co=20 µg/L, headspace open).
373
Figure 6. Effect of residual chlorine on CH degradation.
In addition to water type, CH removal during a heating process is highly

related to the water pH but unrelated to initial concentration (Figure 5b and Figure
5c). The mechanism of pH effect is perhaps attributable to the base-catalyzed
hydrolysis (26). As for volatilization, a comparison of capped and uncapped tests
showed no preference in head-space open condition (Figure 7), indicating that the
noticeable CH removal (~20%) is perhaps unrelated to volatilization of CH but a
result of thermo-assisted hydrolysi. Or, because chloral hydrate has a boiling point
of 98°C, it is likely that it was evaporated when the water temperature exceeds that
point.
CH Treatment by Stirring and Ultrasonication
According to a US EPA modeling program called EPIsuite (version 4.11), CH

(CAS No. 302-17-0, 165.4 g/mol) has a Henry’s law constant (HLC) of 5.71E-
009 atm-m3/mole, while its dehydrated form trichloroacetaldehyde (TCA, CAS
No. 75-87-6, 147.4g/mol) has a HLC of 2.97E-006 atm-m3/mole. The difference
in HLC indicates that CH is a non-volatile compound whereas TCA is a semi-
volatile organic compound that may be eliminated by gas bubbling or air-stripping
processes. In this study, we have found little removal of CH under a variety of
stirring conditions (Figure 8a), which were designed to facilitate the volatilization
effect. Similarly, while vortex was observed in the experiments, the concentrations
of CH remained unchanged after the processes.
374
Figure 7. CH (a) and temperature (b) changes during microwave heating process.
375
Figure 8. Effects of stirring speed (a, Co=10 mg/L, half-full volume [125ml] in
conical flask, uncapped) and ultrasonication (b, Co=20 µg/L, 40kHz, 100ml
water, uncapped) on CH removal from water.
Sonolysis has been found to remove volatile DBP compounds, such as

trihalomethanes, successfully (27–29); and its application on CH was evaluated
about four decades ago (30). These studies indicated that the performance of
ultrasound technology is influenced by multiple factors including ultrasonic
frequency, intensity, pH, initial concentration, and the presence of oxygen,
catalysts or inhibitors -- mainly due to the formation of free radicals. However,
376
although the ultrasonic devices used in this experiment have comparable power
inputs (0.2~0.4 w/cm2) with those in the literature (0.1~1.4w/cm2) (30) and
similar ultrasound frequency (40 kHz) as early study (29 kHz) (30), our tests
showed almost no effectiveness on CH removal (Figure 8b). The exact reason
for this is unclear, but may be attributable to the low initial CH concentration,
lack of oxygen and catalysts, or inadequacy of conversion from electricity to
ultrasonic power. What is certain is that ultrasound waves are incapable of either
breaking the chemical bonds of CH directly or stripping CH away from water
indirectly under the conditions of this study. Further study may evaluate more
systematically the treatability and cost-effectiveness of sonolysis in the control of
CH and other DBPs.
Conclusions
1. RO process can eliminate CH in a wide range of operating pressures,

initial concentrations, pH, and water types, and serves as a convenient
and robust tool for controlling CH at home.
2. Adsorption of CH by GAC exhibited some capacity (44~94%) for CH
control, but in practice such capacity can be highly compromised by the
short contact time. In contrast, adsorption of CH by PP cotton and CTO
exhibited no effect on CH control.
3. Heating water to the boiling point by either boiler or microwave oven is
very efficient in controlling CH, especially for CH in tap water (>90%),
but not as high for CH in ultrapure waters (~20%). The underlying
mechanism is perhaps attributable to the presence of residual chlorine,
but not the salinity of the water. The reduction of CH in ultrapure water
may be due to the thermal hydrolysis effect.
4. CH is a non-volatile compound and, as a result, stirring or ultrasonication
technology are unable to lower its concentration to any great extent
(<5%).
Overall, RO treatment and heating are two efficient and convenient POU
methods for CH control by the public in daily life.
Acknowledgments
The study is financially supported by the National Natural Science Foundation
of China (51278144), the Shenzhen Science & Technology R&D Funding
(JCYJ20120613150442560), and the Funding for Returned Oversea Chinese from
Shenzhen (KQCX20130627094615414). Thanks to the editors for organizing
this book, and to reviewers for providing valuable comments and advice.
377
References
1. Gan, W. H.; Guo, W. H.; Mo, J. M.; He, Y. S.; Liu, Y. J.; Liu, W.; Liang, Y.
M.; Yang, X. The occurrence of disinfection by-products in municipal
drinking water in China’s Pearl River Delta and a multipathway cancer risk
assessment. Sci. Total Environ. 2013, 447, 108–115.
7175–7185.
2008, 72, 875–881.
4. Zhang, L.; Xu, L.; Zeng, Q.; Zhang, S. H.; Xie, H.; Liu, A. L.; Lu, W.
Q. Comparison of DNA damage in human-derived hepatoma line (HepG2)
exposed to the fifteen drinking water disinfection byproducts using the single
cell gel electrophoresis assay. Mutat. Res. 2012, 741, 89–94.
5. Daniel, F. B.; Deangelo, A. B.; Stober, J. A.; Olson, G. R.; Page, N. P.
Hepatocarcinogenicity of Chloral Hydrate, 2-Chloroacetaldehyde, and
Dichloroacetic Acid in The Male B6c3f1 Mouse. Fundam. Appl. Toxicol.
1992, 19, 159–168.
6. Benson, R. Toxicological Review of Chloral Hydrate (EPA/635/R-00/006) in
Support of Summary Information on the Integrated Risk Information System
(IRIS); U.S. Environmental Protection Agency: Washington DC, August
2000.
7. Simpson, K. L.; Hayes, K. P. Drinking water disinfection by-products: an
Australian perspective. Water Res. 1998, 32, 1522–1528.
8. Barrott, L. Chloral hydrate: Formation and removal by drinking water
treatment. J. Water Supply: Res. Technol. −AQUA 2004, 53, 381–390.
9. Wang, X.; Yang, H.; Xie, Y. Lessons of USA Disinfection Byproduct
Database to Chinese Drinking Water Standard - Chloral Hydrate
Concentration and Compliance Analysis (in Chinese). Water Wastewater
Eng. 2014, 13–19.
10. Singer, P. C. Control of Disinfection By-Products in Drinking-Water. J.
Environ. Eng. -ASCE 1994, 120, 727–744.
11. Dąbrowska, A.; Nawrocki, J. Controversies about the occurrence of chloral
hydrate in drinking water. Water Res. 2009, 43, 2201–2208.
12. Bond, T.; Goslan, E. H.; Parsons, S. A.; Jefferson, B. Treatment of
disinfection by-product precursors. Environ. Technol. 2011, 32, 1–25.
13. Doederer, K.; Farre, M. J.; Pidou, M.; Weinberg, H. S.; Gernjak, W.
Rejection of disinfection by-products by RO and NF membranes: Influence
of solute properties and operational parameters. J. Membrane Sci. 2014,
467, 195–205.
14. Woods, R. J.; Akhtar, S. Radiation-Induced Dechlorination of Chloral
Hydrate and 1,1,1-Trichloro-2,2-Bis(P-Chlorophenyl) Ethane (Ddt). J.
Agric. Food. Chem. 1974, 22, 1132–1133.
378
15. Chen, B.; Lee, W.; Westerhoff, P. K.; Krasner, S. W.; Herckes, P. Solar
photolysis kinetics of disinfection byproducts. Water Res. 2010, 44,
3401–3409.
16. Lekkas, T. D.; Nikolaou, A. D. Degradation of disinfection byproducts in
drinking water. Environ. Eng. Sci. 2004, 21, 493–506.
17. Chun, C. L.; Hozalski, R. M.; Arnold, W. A. Degradation of drinking water
disinfection byproducts by synthetic goethite and magnetite. Environ. Sci.
Technol. 2005, 39, 8525–8532.
18. Radjenovic, J.; Farre, M. J.; Mu, Y.; Gernjak, W.; Keller, J. Reductive
electrochemical remediation of emerging and regulated disinfection
byproducts. Water Res. 2012, 46, 1705–1714.
19. Wu, W. W.; Benjamin, M. M.; Korshin, G. V. Effects of thermal treatment
on halogenated disinfection by-products in drinking water. Water Res. 2001,
35, 3545–3550.
20. Krasner, S. W.; Wright, J. M. The effect of boiling water on disinfection by-
product exposure. Water Res. 2005, 39, 855–864.
21. Koudjonou, B. K.; LeBel, G. L. Halogenated acetaldehydes: analysis,
22. Lykins, B. W., Jr; Goodrich, J. A.; Koffskey, W. E.; Griese, M. H.; Water, E.;
Utility, S. Controlling disinfection by-products with alternative disinfectants;
USEPA: Cincinnati, OH, June 1991.
23. Canada, H. Guidance on Chloral Hydrate in Drinking Water (H128-1/08-
551E); Water, Air and Climate Change Bureau Healthy Environments and
Consumer Safety Branch, Health Canada: Ottawa, Ontario, 2008.
24. Xie, Y. Disinfection byproducts in drinking water: Formation, analysis, and
control; CRC Press: Boca Raton, FL, 2003.
25. Liu, B.; Reckhow, D. A. Disparity in disinfection byproducts concentration
between hot and cold tap water. Water Res. 2015, 70, 196–204.
26. Chen, B. Hydrolytic Stabilities of Halogenated Disinfection Byproducts:
Review and Rate Constant Quantitative Structure–Property Relationship
Analysis. Environ. Eng. Sci. 2011, 28, 385–394.
27. Guo, Z.; Gu, C.; Zheng, Z.; Feng, R.; Jiang, F.; Gao, G.; Zheng, Y.
Sonodegradation of halomethane mixtures in chlorinated drinking water.
Ultrason. Sonochem. 2006, 13, 487–492.
28. Shemer, H.; Narkis, N. Trihalomethanes aqueous solutions sono-oxidation.
Water Res. 2005, 39, 2704–2710.
29. Kim, I.-K.; Huang, C.-P. Sonochemical process for the removal of DBPs and
precursor in water. J. Environ. Eng. Manage. 2007, 17, 39.
30. Sakai, Y.; Sadaoka, Y.; Takamaru, Y. Decomposition of chloral hydrate in
aqueous solution by the action of ultrasound. J. Phys. Chem. 1977, 81,
509–511.
379
Chapter 20
Disinfection By-Products in Swimming Pool

Water: Formation, Modeling, and Control
Hao L. Tang,1 Ricky J. Ristau II,2 and Yuefeng F. Xie*,2
1Department of Water Engineering and Science, College of Civil

Engineering, Hunan University, Changsha, Hunan 410082, China
2Environmental Engineering Programs, The Pennsylvania State University,
777 West Harrisburg Pike, Middletown, Pennsylvania 17057, U.S.A.

Disinfection is practiced to minimize the occurrence and growth

of microbial pathogens and thus ensure the hygienic safety
of swimming pool water. However, potentially hazardous
disinfection by-products (DBPs) are formed from the reaction
between disinfectants and DBP precursors. This chapter
discusses the formation of these DBPs in swimming pool water
and modeling approaches for quantification of their levels given
estimates of anthropogenic contaminant input in swimming
pools. Promising DBP control technologies are also discussed.
Overview of the three key aspects related to swimming pool
water help readers to better understand the profile of DBPs and
their control in swimming pool water.
Introduction
Swimming, as a physical activity for people of all ages, promotes fitness
and has many advantages over land-based activities. It is generally considered
as a health-enhancing and relatively injury-free activity; it can be a good aerobic
exercise, which contributes to flexibility and muscle strength, and can have
positive social aspects. Swimming pools can be found in different kinds and sizes
in public areas, hotels and spas, or at private residential homes. In the U.S., more
than 368 million pool visits occur each year. In Germany, around 250-300 million
pool visits occur each year, averaging 3 visits per capita. In the UK, one-third of
the children and around 36% of adults (>15 years of age) visit swimming pools at

least once a week, and 55% of children (5-9 years of age) use pools at least once a
month. In the U.S., there are approximately 250,000 public pools and 10,000,000
residential pools. The highest numbers of existing in-ground and above-ground
pools in Europe are found in France (773,000) and Germany (625,000), followed
by UK with 155,000, and Italy with 94,000 (1).
However, after water leaves a public water system and enters a pool, its quality
is no longer regulated as drinking water in the U.S. and many other countries,
and in the U.S. the Safe Drinking Water Act makes no stipulation that the tap
water meets additional requirements for use in pools or spas. Instead, pools are
managed at the local level, usually by state or local public health departments, and
sometimes by pool managers or lifeguards. As the pools are filled with swimmers,
they contain hundreds to thousands of low levels of chemicals originating from
natural and anthropogenic inputs, which complicate the water chemistry. The
undesirable substances (e.g., skin, hair, bodily excretions such as sweat, urine,
and saliva; pathogens; and personal care products such as lotions and sunscreens)
introduced by swimmers also complicate the toxicity and expose swimmers to
potential biological and chemical health risks (2, 3).
To conserve the positive aspects of aquatic activities, it is necessary to
disinfect swimming pool water to prevent outbreaks of infectious illnesses (4,
5). To date, chlorination is still the most common measure for inactivation of
microbial pathogens. Chlorine is added to swimming pools as chlorine gas,
calcium/sodium hypochlorite, or through electrolytic generation of sodium
hypochlorite. Regardless of the method of application, once the chlorine
is in the water it forms hypochlorous acid which dissociates into hydrogen
and hypochlorite ions. The sum of the concentrations of hypochlorous acid,
hypochlorite ion and aqueous chlorine is referred as free chlorine residual.
Chlorine’s popularity is not only due to lower cost, but also to its higher
oxidizing potential, which provides a minimum level of chlorine residual to
inhibit microbial recontamination. Other disinfectants may be used, but less is
known about their ability to inactivate pathogens (6). For effective disinfection,
adequate disinfectant levels are required. In Canada, regulations in Quebec
require free chlorine in swimming pools to be between 0.8 and 2.0 mg/L, while
the required minimum level for British Columbia varies with pH: 0.5 mg/L
for pH 7.4-7.8, and 1.0 mg/L for pH of greater than 7.8 (7). In the U.S., a
survey of twenty-three indoor pools showed that the median chlorine residual
was 3 mg/L (8). In practice, chlorine residuals of 2-4 mg/L are generally
recommended for chlorinated pools (9). However, the reaction of chlorine and
chloramines with organic and inorganic precursors, such as natural organic
matter (NOM), anthropogenic inputs, bromide, etc, may cause the formation of
undesirable disinfection by-products (DBPs). The types and concentrations of
DBPs depend on several factors, including the type and amount of disinfectant
used, characteristics of the swimming pool and pool water and users’ hygiene (1).
These DBPs can be inhaled or ingested by swimmers during swimming, bathing
and showering, or absorbed dermally (10), resulting in negative effects on human
health (11, 12). DBPs were observed to produce cancer in animal models and to
have other toxic consequences, such as reproductive and developmental effects;
in particular, users of chlorinated pools may have an elevated risk for bladder
382
cancer (13) and an increased frequency of micronuclei in peripheral blood cells
(a marker of DNA damage) (14). These published epidemiology studies provide
evidences that disinfected pool water might be genotoxic and carcinogenic to
humans. Given the popularity of swimming that comes with the increased risk of
exposure to pathogenic microorganisms, disinfectants and DBPs, we are only in
the early stages of understanding swimming pool chemistry, human exposures,
and potential health risks (6). Regulations must also ensure that efforts to reduce
DBPs do not result in water that is impaired due to microbial contamination:
This is the microbe-DBP balancing act of minimizing risks while maximizing
beneficial effects.
Research has specifically focused on two classes of DBPs from chlorine-based
disinfection: trihalomethanes (THMs) and haloacetic acids (HAAs) (15–19).
THMs, including chloroform, bromodichloromethane, dibromochloromethane,
and bromoform, are the best known and most intensively investigated class of
DBPs and they are regulated in chlorinated drinking water in the U.S. with a
maximum contaminant level at 80 µg/L. Reports on THMs in swimming pools
first appeared in 1980s, and they showed a tendency of wide variations in the
reported data. Fantuzzi et al. (20) found the THM levels ranged from 18 to 71
µg/L in five indoor swimming pools in Italy. Chu and Nieuwenhuijsen (21) found
the THM levels averaged at 133 µg/L in 44 indoor swimming pools in the UK.
HAAs, commonly referred as HAA6 including monochloroacetic acid (MCAA),
monobromoacetic acid (MBAA), dichloroacetic acid (DCAA), trichloroacetic
acid (TCAA), bromochloroacetic acid (BCAA), and dibromoacetic acid (DBAA),
are also regulated for drinking water in the U.S., with a maximum contaminant
level at 60 µg/L. There is limited information on HAA levels in swimming pools.
Tang (22) reported the HAA level stabilized at approximately 1650 µg/L in an
indoor swimming pool in the U.S.. Simard et al. (23) analyzed 15 indoor and 39
outdoor municipal public pools in Canada and found the HAAs were higher in
outdoor pools (808 vs 413 µg/L). Wang et al. (24) analyzed 5 outdoor pools in
the U.S. and 9 indoor pools in China and found that the HAA levels in the pools
in the U.S. averaged at 1440 µg/L while in China the levels averaged at 117 µg/L
due to lower chlorine residuals.
Some nitrogen-containing DBPs that were previously reported in drinking
water (25) are also identified in swimming pool water (26), due to the presence of
nitrogen sources, such as urine or sweat. Haloacetonitriles (HANs) were the most
significant toxic compounds detected by Hansen et al. (27), though they were
found at much lower concentrations compared to THMs and HAAs. However, it is
not known how much the HANs contribute to the overall mixture toxicity. The pH
plays an important role on DBP classes. For decreasing pH, THM concentrations
are observed to decrease, while HAN concentrations increased. Therefore
caution is warranted if the pH level in swimming pool water is lowered for THM
control, since the the concentration of more toxic HANs appear to increase.
Besides HANs, many other DBPs, including dichloromethylamine (CH3NCl2),
cyanogen chloride (CNCl), trichloronitromethane (TCNM), chloral hydrate
(CH), haloketones (HKs), N-nitrosodimethylamine (NDMA), chloramines
and cyanogen bromide (CNBr) have also been reported in swimming pools
(28, 29), though at much lower concentrations compared to the predominant
383
THM and HAA species. To date, their complete chemical and toxicological
characterizations are not available.
In this chapter, we examine three key aspects related to swimming pool water.
Formation
DBP precursors lead to the formation of DBPs upon chlorination. Research
on DBP precursors in swimming pool water is of keen interest by scientists (30,
31). One option to ensure the safety of public drinking water systems is to improve
the quality of source water through pollution prevention and precursor removal.
However, a source reduction approach is challenging for swimming pools, because
the swimmers themselves are the largest source of organic/inorganic precursors
for DBP formation. The incidental anthropogenic contaminant release as a result
of human excreta such as urine and sweat, either accidentally or on purpose, is

difficult to quantify (32, 33). And its impact on DBP formation in swimming pools
is unclear. In this section, we discussed our research on investigating the formation
of DBPs and the contributions from various precursors in a sample swimming pool.
Modeling
Models for DBP levels in swimming pool water are needed, aiming at
identifying the significance of diverse operational and water quality parameters
controlling the formation of DBPs or at investigating the kinetics for their
formation. It may consist of empirical and mechanistic relationships of water
quality and operational parameters with the prevailing levels of DBPs at various
stages of a same single pool. In this section, we discussed our research on
developing predictive models for DBPs based on data obtained from field and
laboratory-scaled studies.
Control
Swimming pool water treatment in general includes flocculation, sand
filtration, and subsequent disinfection with chlorine. The continuous chlorination
and input of anthropogenic releases by bathers in combination with recirculation
of the pool water may lead to an accumulation of DBPs in the water. Better
technologies are needed to solve this issue. In this section, we discussed our
research on potential technologies on DBP control from swimming pool water.
By giving an overview of our research efforts on the three topics in this
chapter, readers can better understand the profile of DBPs and their precursors in
swimming pool water.

Reagents
All chemicals and standard solutions, unless noted otherwise, were analytical
grade purchased from Sigma-Aldrich.
384
Analysis of Dissolved Organic Carbon (DOC), Ammonia Nitrogen (NH3-N),
UV Absorbance at 254 nm (UV254), and Chlorine
DOC was measured with a total organic carbon (TOC) analyzer (O.I.
Analytical Model 1010, Maryland, USA). NH3-N was measured using a
ammonia-selective electrode method according to Standard Method 4500-NH3
(34). UV254 was measured using an Agilent 8453 UV-Visible Spectrophotometer
(Agilent technologies, California, USA) with a 10 mm quartz cuvette.
Free chlorine and total chlorine were measured immediately using the
N,N-diethyl-p-phenylenediamine (DPD) method with a DR/890 portable
colorimeter (Hach Company, Colorado, USA).
Analysis of THMs and HANs

Table 1 shows a list of investigated compounds. Residual chlorine in THM

and HAN samples was quenched by adding 5 mg sodium sulfite to 40 mL
borosilicate glass vials before they were filled head-space-free with the samples.
The samples were stored at 4°C until analysis by a gas chromatograph (GC) (HP
6890 Series GC system, Hewlett Packard) with mass spectrometer (5973 Mass
selective detector, Hewlett Packard). A DB-1 capillary column (30 m × 0.32
mm i.d., 1.0 µm film thickness) was used. The temperature ramping program
was as follows: Initial at 35°C for 22 minutes, ramp to 145°C at 10°C/min
and hold for 2 minutes. This method was also used for the detection of chloral
hydrate (CH), trichloronitromethane (TCNM), dichloropropanone (DCP), and
trichloropropanone (TCP). For more information refer to EPA 551.1 method.
Analysis of HAAs
For the analysis of HAAs, a modified version of the EPA 552.3 method was
used. Sulfuric acid, sodium sulfate, surrogate standard (2-bromobutanoic acid)
and methyl-tert-butyl ether (MtBE) was added to the samples which then were
hand shaked for two minutes to extract HAAs. The MtBE phase was transferred to
a test tube and acidified methanol was added. The samples were placed in a water
bath at 50°C for 2 h to methylate the HAAs and then back-extracted with sodium
sulfate solution to remove excess methanol. The MtBE phase was transferred to
a GC vial and analyzed by a GC (HP 6890 Series GC system, Hewlett Packard)
equipped with a DB-1701 capillary column (30 m × 0.25 mm i.d., 0.25 µm film
thickness). The temperature ramping program was as follows: Initial at 35°C for
10 minutes, ramp to 75°C at 5°C/min, hold for 16 minutes, ramp to 200°C at
25°C/min and hold for 5 minutes.
DBP Formation Potential (DBPFP) test

A DBPFP test was started by applying an initial chlorine dose to the phosphate
buffered samples, which had been pre-diluted to accommodate 3-day free chlorine
demands (35). The incubation process took 3 days at 25°C in the dark. The DPD
method was used to measure the free chlorine and total chlorine. Samples were
385
then quenched to eliminate chlorine residuals and analyzed for DBPs based on the
above protocols. It is noted that five kinds of anthropogenic contaminants, i.e.
sweat, saliva, skin wash, hair wash, and urine, were also sampled for the DBPFP
tests. The anthropogenic inputs were obtained from a 26-yr-old college male, with
sampling protocols approved by the Pennsylvania State University Institutional
Review Board and a written informed consent obtained from the participant.
Table 1. List of Investigated Disinfection By-Products

Compound Abbreviation Chemical formula
THMs Chloroform TCM CHCl3
Bromodichloromethane BDCM CHBrCl2

Dibromochloromethane DBCM CHBr2Cl
Bromoform TBM CHBr3
HAAs Chloroacetic acid MCAA CH2ClCOOH
Bromoacetic acid MBAA CH2BrCOOH
Dichloroacetic acid DCAA CHCl2COOH
Trichloroacetic acid TCAA CCl3COOH
Bromochloroacetic acid BCAA CHBrClCOOH
Dibromoacetic acid DBAA CHBr2COOH
HANs Dichloroacetonitrile DCAN CHCl2CN
Trichloroacetonitrile TCAN CCl3CN
Bromochloroacetonitrile BCAN CHBrClCN
Dibromoacetonitrile DBAN CHBr2CN
Others Dichloropropanone DCP CHCl2COCH3
Trichloropropanone TCP CCl3COCH3
Trichloronitromethane TCNM CCl3NO2
Chloral hydrate CH C2H3Cl3O2
Formation
The Swimming Pool under Investigation
The DBP formation in a swimming pool was investigated. The pool, with a
volume of 440 m3, was located in Capitol Union Building at The Pennsylvania
State University, Middletown, Pennsylvania, USA. The pool was completely
drained, and refilled with fresh water in the second week of June in 2009. Since
then, the pool water was periodically sampled and analyzed for DBPs during a
386
1.5-yr investigation period. During the operation of the swimming pool, water
was filtered and chlorinated through recirculation of the pool water. Some water
was lost during back-washing of the filters and by evaporation, therefore fresh
water was added to maintain a constant water level at a rate of approximately 1.8
m3 weekly. Some DBPs might be adsorbed to suspended solids in the pool and
removed with the filtration; however, for this study that process and the loss of
water through evaporation were neglected, since they account for a negligible
fraction of the pool water. The pool was open year-around to students, faculty and
staff of the university, members, and guests, i.e. walk-ins of the local community.
Table 2 shows the year-around user statistics of the swimming pool. These data
show that the pool accepted 1441 users per month or 48 users per day on average.
As the typical pool hours were from 9 am to 9 pm in work days except national
holidays, the pool accepted an average of 4 users per hour.
Table 2. Year-around User Statistics of the Swimming Pool

Total visits
Members Students Faculty/Staff Guests* per month
January 398 255 91 1579 2323
Feburary 353 291 81 1006 1732
March 422 312 70 366 1536
April 346 275 65 665 1351
May 360 140 83 432 1015
June 307 141 78 388 1114
July 315 105 122 892 1434
August 290 164 111 224 789
September 396 231 80 651 1358
October 356 243 133 1148 1879
November 326 169 98 1326 1955
December 325 135 60 896 1416
*Guests include the pool users of swim lessons, swim teams, water fitness classes, and
walk-ins of local community.
DBP Levels in the Swimming Pool
Figure 1 shows the profiles of THM and HAA concentrations in the

swimming pool following the water change over a 1.5-yr period. After the pool
water was completely drained and refilled with fresh water, the THM and HAA
concentrations were 81 and 87 µg/L, respectively. These two numbers were
similar to those of finished water (80 µg/L for THMs and 60 µg/L for HAAs)
387
from a typical surface water treatment plant. As the water age increased and
more swimmers were admitted, the THM levels were relatively stable, being
averaged at 62 µg/L with a standard deviation of 35 µg/L. However, the HAA
concentrations substantially increased. Two months following water change, 940
µg/L HAAs were observed in the pool. During the 6th and 8th months following
water change, the HAA level appeared to reach a plateau and fluctuated around
1600 µg/L. From the 8th to 12th month, the HAA level consistently declined to 780
µg/L, and after the 12th month, the HAA concentration steadily increased again to
the 1600 µg/L level when the 1.5-yr investigation period was close to the end. It
is noticeable that the decline in HAA concentrations from January to June could
be attributed to the decrease of the number of pool users during the same period,
as the monthly number of pool users decreased from 2223 in January to 1114 in
June. Figure 2 also shows that a positive correlation between the measured HAA
levels in the pool and the numbers of pool users, and the R2 value was 0.70.
Figure 1. THM and HAA levels in an indoor swimming pool after water change.
388
Figure 2. Correlation between the measured HAAs and the numbers of pool users.
The predominant species of HAAs, THMs, HANs, and HKs in the swimming
pool water were DCAA and TCAA, TCM, DCAN, and TCP, respectively. Table 3
shows their concentrations in the pool water during the 1.5-yr investigation period
since water change. DCAA was found to be in a higher concentration than TCAA.
Because DCAA and TCAA are believed to be produced as a result of different
mechanistic pathways (36), the ratio of DCAA/TCAA implies that the formation
of DCAA is favoured in the swimming pool water, possibly due to the presence
of more precursors for DCAA in pool water. HANs and HKs in swimming pools
were investigated by fewer studies than those of THMs and HAAs. Lee et al. (37)
reported DCAN in the chlorinated pools as 3.9 ± 3.3 µg/L, while TCAN was not
reported. In this research, the DCAN in the swimming pool water was 1.3 ± 0.7 µg/
L. The HAN concentrations were much lower than THM and HAA concentrations,
possibly due to their decomposition to form HAAs at pH conditions higher than
7.0 (27). As the DBP formation is also attributable to precursors for DBPs, the
profiles of individual species need to be reviewed to explore the contribution from
their precursors.
DBP Precursors in the Swimming Pool

The widely accepted precursors for DBP formation in drinking water is the
natural organic matter (NOM). However, this phenomenon is not completely
applicable to swimming pools. Researchers have associated anthropogenic
inputs, i.e. sweat and urine, by swimmers with main sources of DBP precursors
(38–41). It was estimated that a swimmer introduces 25-80 mL of urine and
200-1000 mL of sweat into swimming pool water each swim session (42, 43).
As anthropogenic contaminants are continuously brought in by swimmers and
pools are continuously exposed to disinfectant, the disinfection process resembles
389
a DBPFP test, which maintains a sufficient free chlorine residual and a long
reaction time to convert precursors to DBPs. To further explore their impact on
DBP formation in swimming pools, five kinds of anthropogenic contaminants,
i.e. sweat, saliva, skin wash, hair wash, and urine, were analyzed for DBPFPs
Table 3. Concentrations of Individual Species and the Ratio of DCAA/TCAA

Month DCAA/
DCAA TCAA TCM DCAN TCP
No. TCAA*
1 51 18 87 0.0 0.0 2.8
2 276 65 93 0.8 2.3 4.2
3 636 186 74 1.6 2.1 3.4

4 750 201 181 0.9 1.5 3.7
5 975 290 43 2.3 2.0 3.4
6 1100 326 40 2.7 2.5 3.4
7 1095 366 48 1.6 2.5 3.0
8 1119 381 43 1.8 1.8 2.9
9 1074 489 50 1.5 1.5 2.2
10 942 444 55 1.0 1.3 2.1
11 715 400 50 2.2 1.1 1.8
12 493 362 51 1.4 1.9 1.4
13 504 249 52 0.8 1.6 2.0
14 849 321 43 0.6 1.2 2.6
15 861 345 61 0.4 1.6 2.5
17 1167 480 47 1.6 2.1 2.4
19 1116 504 34 1.3 2.9 2.2
*Data in this ratio column are unitless. Data in the DCAA, TCAA, TCM, DCAN, and TCP
columns are concentrations represented. in µg/L.
Table 4 shows the DBPFPs of the anthropogenic input samples. All the
anthropogenic contaminants were found to contribute to the DBP formation,
which was consistent with the results reported by Kim et al. (39). The HAA
yields of chlorinated sweat, saliva, skin wash, hair wash, and urine samples were
300, 390, 380, 160, and 160 µg/mg C, respectively, while the THM yields of the
corresponding samples were 67, 130, 96, 50, and 74 µg/mg C, respectively. The
relative abundance of HAAs and THMs was presented by the ratio of HAA/THM.
These anthropogenic contaminants contributed to more HAAs than THMs, with
the HAA/THM ratio in the range of 2.1 and 4.5. It is noted that urine had the
390
lowest HAA/THM ratio of the 5 contaminants. For the 3-day Cl2 demand per
DOC, Kanan and Karanfil (8) reported 2-8 mg/L Cl2 demand for 1 mg DOC of
filling water NOM and 17-25 mg/L Cl2 demand for 1 mg DOC of body fluid
analogue. A slight higher number, 21-40 mg/L Cl2 demand for 1 mg DOC of real
anthropogenic contaminants, was found in this research. Table 4 also presents the
yields of HAAs and THMs by normalizing HAAFP and THMFP based on DOC,
to compare different anthropogenic contaminants regardless of their dilution
ratios. Sweat was observed to have an HAA yield of 300 µg/mg C. Urine and
hair were observed to have relatively low HAA yields (160 µg/mg C) compared
to three other anthropogenic contaminants. The THM yields of sweat and urine
were 67 and 74 µg/mg C, respectively. These DBP yield data will be used in the
later modeling section.
Table 4. DBPFPs and DBP Yields of Anthropogenic Input Samples

Skin Hair
Sweat1 Saliva1 Urine2
wash1 wash1
3-day Cl2 demand
[mg Cl2/L] 170 170 210 230 410
THMFP
[µg/L] 430 1010 870 560 740
HAAFP
[µg/L] 1930 3050 3420 1750 1560
3-day Cl2 demand
per DOC
[mg Cl2/mg C] 27 22 23 21 41
THM yield
[µg/mg C] 67 130 96 50 74
HAA yield
[µg/mg C] 300 390 380 160 160
HAA/THM 4.5 3.0 3.9 3.1 2.1
1A dilution ratio of 1:25 was applied to accommodate free Cl2 demand in DBPFP test. 2
A dilution ratio of 1:33 was applied to accommodate free Cl2 demand in DBPFP test. Data
presented are average values of duplicate sets.
Interpretation of DBP Profile in Swimming Pools
Swimming pool water is found to have high DBP levels, especially in forms
of HAAs. THMs, however, did not show an increase trend with water age but were
relatively constant in the pool water. We have summarized four factors that lead
to this specific DBP profile in swimming pools:
391
Firstly, the long water retention time in swimming pools is special. As the
pool water is continuously exposed to disinfectants, all DBP precursors have a
potential to be driven to DBPs. This tends to result in high DBP levels.
Secondly, continuous introduction of anthropogenic contaminants leads to
high DBP levels. With an R2 of 0.70, a positive correlation between the HAA
levels and the numbers of pool users exists (Figure 2).
Thirdly, discrepancies in the chemical loss rates could lead to the high HAA
but low THM profile. Benoit and Jackson (44) observed stable THM levels in 25
whirlpool spas affected by heat, aeration, and agitation. Although these factors,
i.e. the degree of water agitation by pool users, were difficult to measure, the
THM loss rate appeared to be faster than the HAA loss rate because of much higher
Henry’s law constants for THMs. The loss rate includes the effects of evaporation,
adsorption, degradation, and other incurred processes.
Finally, the anthropogenic inputs of swimming pools consist of more HAA

precursors than THM precursors, as data of this research presents (Table 4). It
is obvious that reduction of anthropogenic contaminants will lead to reduction of
DBPs. This research demonstrated the actual effect of anthropogenic contaminants
on the level of DBPs.
Modeling
Model Development
In recent years, a particular interest has been grown on the development of

models to estimate the formation and fate of DBPs. Several predictive models have
been reported (45). These models used different types of explanatory variables for
variety of applications in drinking water industry. The DBP profile in a swimming
pool system can be conceptualized based on mass balance and decay kinetics (46,
47). For chlorination of pool water, a comprehensive model for DBPs should
consider variation of DBP precursors, i.e. sweat and urine from anthropogenic
inputs. Efforts should be focused on the gaps between specific models and their
actual applications.
The concentration data used to model DBP formation in swimming pool water
were obtained from the study by Tang (22) where samples were collected from an
indoor pool once a month over a 1.5-yr period. The DBP model was developed
based on four assumptions: (1) All DBP precursors from anthropogenic inputs
have been converted to DBPs under the chlorination scenarios of swimming pools;
(2) There is no water change during the entire modeling period and the amount
of make-up water is negligible, because a 1.8 m3 weekly make-up water versus
the pool volume of 440 m3 makes the influence of dilution within the range of
variability for DBP analysis. (3) The DBP loss due to evaporation, degradation,
dilution, and adsorption follows first-order kinetics (48); and (4) The DBP loss
rates maintain constant since water temperature of the pool is stable. Equation
1 shows the mass balance of DBPs in swimming pools. It can be reformed to
represent the computational algorithm on the basis of the contribution from daily
anthropogenic inputs to DBPs (Equation 2 and 3).
392
Where
DBPi = Accumulated DBP at Day i, µg DBP/L;

DBPnew,i = New DBP In due to the anthropogenic input at Day i, µg DBP/L;
DBPi-1 = Accumulated DBP at the day before Day i, µg DBP/L;
DBPi-1 × e-k = Remaining DBP after 1-day first-order decrease, µg DBP/L;
k = DBP loss coefficient, day-1.
Where
Ni = Number of pool users at Day i, persons;
YS = DBP yield for sweat, µg DBP/mg C;
CS = Sweat carbon released per person, mg C/person;
YU = DBP yield for urine, µg DBP/mg C;
CU = Urine carbon released per person, mg C/person;
V = Volume of the swimming pool, Liter.
It is noted that CS, CU, and k were difficult to measure and show
variances from one swimming pool to another. They need to be calibrated by
computer-based regression programs. Shuffled complex evolution (SCE), a
global optimization method developed by Duan et al. (49), was used to calibrate
these difficult-to-measure parameters of the swimming pool. The calibration
process used 48 (approximately 70%) of data sets (12 months × 4 classes of
DBPs) based on statistical requirements. The CS, CU, and k were assigned an
initial value, respectively, to start the calibration. If the values were perfect, there
would be an approximate prediction but not necessarily an exact match. To take
care of small discrepancies, small adjustments were made to these parameters
until the predicted data matches the measured ones. The SCE algorithm ensured
that the root mean squared value between the predicted data and the measured
ones was the lowest. The remaining 30% of data sets were used for validation.
The coefficient of determination (R2) was used to verify the regressions between
the predicted DBP data the measured ones in the pool water.
393
Model Performance
The obtained parameters of the DBP model for the swimming pool after
calibration were as follows:
CS , Sweat carbon released per person = 1000 mg C/person;

CU , Urine carbon released per person = 480 mg C/person;
kHAA , HAA loss coefficient = 0.035 day-1;
kTHM , THM loss coefficient = 0.23 day-1;
Figure 3 presents the calibration and validation results of comparing the

predicted and measured data without distinguishing different DBP classes. The
45° line the figure depicted the predicted DBP data that were precisely equal to
the measured ones. As shown in Figure 3a, with an R2 of 0.97, the predicted
data showed a good fit to the measured ones. The validation process used the
remaining 30% of data sets, and was still able to show a good fit to the measured
data, with an R2 of 0.96 (Figure 3b). Therefore, the model calibration led to sound
results for the coefficients of the swimming pool system.
Figure 4 presents the predicted profiles of HAAs and THMs in comparison to
the actual ones. The patterns were developed as a result of a combined effect of
DBP formation and loss rates as described in the previous section. The predicted
HAA curve captured the rising trends and descending trends very well. The
predicted THM pattern also gave a quantitatively correct range. Because of the
high loss rates, THM variations with water age were not as significant as those of
HAAs. An extraordinary THM observation of 181 µg/L at the 104th day could be
attributed to a peak hourly loading of swimmers during the sampling.
Figure 3. Calibration and validation results. (a) calibration based on 70% data;
(b) validation based on 30% data.
394
Figure 4. Predicted DBP profiles as a function of water age. (a) HAAs; (b) THMs.
Model Implications
Computer-based modeling is a useful tool for a number of tasks performed
by water and wastewater professionals. Because of the high R2 values of both
calibration and validation, the DBP model developed based on mass balance
and first-order kinetics is competent to predict DBPs in swimming pools. The
model basically attributes the formation of DBPs to the continuous introduction
of anthropogenic contaminants or the number of pool users.
Anthropogenic inputs represent the major cause of high DBP levels in
swimming pool water. The model suggests 1000 mg carbon from sweat per
person and 480 mg carbon from urine per person were released to the pool
every day. According to Judd and Black (46), body fluids had a carbon content
of approximately 6000 mg/L. The model implies that each pool user released
approximately 170 mL sweat and 80 mL urine to the pool. These values were
in the range reported by Gunkel and Jessen (42) and Erdinger et al. (43), who
estimated that a swimmer introduces in water 25-80 mL/h of urine and 200-1000
mL/h of sweat. The model also presents an approach to estimate the DBP loss
coefficients. It is believed that water agitation affects the DBP levels in swimming
pool water. However, the degree of water agitation is difficult to measure.
According to the model, the HAA loss coefficient was 0.035 per day while the
THM loss coefficient was 0.23 per day in the investigated pool. Lower coefficient
implies longer half life and thus higher concentration.
395
Control
Control Factors
DBPs in drinking water are typically controlled by limiting both the amount
of disinfectants and the amount of NOM. In swimming pool water, the DBP
formation may be limited by three factors: disinfectant residual, the amount of
organic material in the water, and the recirculation of the pool water (50).
The amount of chlorine entering the pool water can be easily controlled, as
the chlorine, in forms of granular powder or tablets, can be added by hand or
by a mechanical dosing system. Proper dosing ensures chlorine residual kept
at a minimum and still effectively eradicate pathogenic microorganisms, in the
meanwhile keeps DBP concentrations at low levels. This is the microbe-DBP
balancing act of minimizing risks while maximizing beneficial effects.
The amount of organic material can be controlled by reducing the input of

DBP precursors (NOM and anthropogenic input). The anthropogenic input can be
reduced considerably by the behavior of swimmers before and during swimming.
For example, showering and using toilet facilities, washing off sunscreen lotions,
and applying water-tight diapers can reduce the anthropogenic load and help to
reduce DBP formation. Keuten et al. (51) found that the duration of shower has
the largest influence on the removal of contaminants and most of contaminants
can be removed within a 60-second shower. Although a showering protocol is
an important first step to inform pool users about their personal influence on the
pool water quality, showering protocols alone might not be enough. The aquatic
staff also needs to be aware of the necessity of pre-swim showering. And finally,
supportive policies (e.g., maintaining clean, well-stock bathroom facilities that not
only encourage showering but also toileting) should be established, implemented,
and enforced.
The recirculation of the pool water leads to an accumulation of DBPs in the
water. From the technological point of view, the treatment efficiency to reduce
DBP precursors and already-formed DBPs can be improved by more efficient
filtration and oxidation. This includes tight membrane (e.g. nano or reverse
osmosis membrane) filtration, carbon filtration, biological filtration, and advanced
oxidation (52). Removal of the low-molecular-weight fraction would be advised
because it contains a large part of the toxicity according to Zwiener et al. (1).
Their toxicity data also suggest keeping bromide and iodide concentrations in
pool water low and not using bromine because brominated and iodinated DBPs
have been found to be even more toxic than chlorinated DBPs. It is also beneficial
to increase air circulation in indoor pool settings to reduce the levels of volatile
DBPs.
Filtration
Typical swimming pool water treatment includes sand filtration and
disinfection with chlorine. Hansen et al. (27) reported a few swimming pools
in Denmark use drum filters based on a weaved microsieve (cutoff 10 or 20
µm) which removes the particles from the pool water fast by washing the filter.
Although most swimming pools remove particles through a filtration unit, the
396
filtration units are not designed to remove organic matter. Control of DBPs
and their organic precursors using adsorbent, i.e. GAC, could be an effective
approach.
Two types of GAC were investigated in bench-scale filters: virgin GAC
supplied by Waterlink (USA) and biologically activated carbon (BAC) collected
at a local drinking water treatment facility. The schematic setup of the GAC
filters is depicted in Figure 5. Swimming pool water was directly added to the
filter columns loaded with 40 mL of GAC and the flow rates were adjusted to 2
mL/min to have an empty bed contact time (EBCT) of 20 minutes.
Figure 5. Schematic setup of bench-scale GAC filters.
The bench-scale GAC filter was operated for 400 hours. Results showed
that TCAA was completely removed during the beginning of the experiment and
the effluent concentration of DCAA steadily increased. Before the GAC column
reached saturation, biological degradation of DCAA was observed after 128 hours
of operation. This is shown in Figure 6 at the highest peak of the curve. At this
point, the total HAA was 566 µg/L, and there was 48% reduction in total HAAs, a
9% reduction in DCAA and 99% reduction in TCAA. The results imply that GAC
was effective in removing TCAA while the adsorption capacity for DCAA was
relatively low.
The bench-scale BAC filter was initially operated for 2 hours filtering
deionized water. Then it started filtering swimming pool water for 480 hours
at the same settings with the GAC filter. Since BAC has almost no adsorption
capacity, it became saturated almost immediately for TCAA (Figure 7). There
was approximately 90% removal of DCAA throughout the entire experiment.
Opposite to what was observed in the bench-scale GAC filter, this occurred
because the bacteria capable of degrading DCAA were already established on the
BAC. The TCAA degrading bacteria, however, took approximately 30 hours to
397
be established, and then the TCAA concentration in the effluent rapidly decreased
throughout the experiment. In the end of the experiment, approximately 70% of
TCAA was removed.
Figure 6. HAA removal from swimming pool water by a bench-scale GAC filter.
Figure 7. HAA removal from swimming pool water by a bench-scale BAC filter.
These experiments show that GAC is able to adsorb TCAA and BAC can
remove DCAA efficiently at the beginning. Once TCAA degrading bacteria
become established, BAC can efficiently remove HAAs from swimming pool
water and outperforms GAC. The study also implies that, given enough time,
bacteria capable of degrading HAAs will develop on the surface of GAC,
converting it to BAC and reducing the need to replace the GAC. A disadvantage
associated with both studies is that the filter effluent had zero chlorine residual.
This may lead to problems when used on an actual swimming pool, because the
398
pool water needs to be continuously chlorinated to keep a residual typically in
a range of 2-4 mg/L. Elevated costs for the use of additional chlorine could be
expected.
Besides adsorption and biodegradation during the filtration process for DBP
control in swimming pool water, other filtration technologies, i.e. membrane,
zero-valent iron, etc, are also being investigated by scientists. Cost-effective HAA
removal technologies for swimming pool water represent a promising direction of
future research.
Conclusions
In this chapter, three key aspects related to swimming pool water were
investigated:
Formation
We measured DBP formation in a swimming pool over a 1.5-yr period since
water change. We found the HAA levels could go very high in the swimming pool,
while THM levels remained relatively constant. By exploring the DBPFPs from
anthropogenic contaminants, we found the anthropogenic inputs to swimming
pools contributed to high HAAs and significantly impacted certain DBP species
in the water.
Modeling
The DBP profile in swimming pools can be modeled and predicted. The
DBP model was developed using the DBP data of the investigated swimming
pool and a computer-based derivatization method based on mass balance and
first-order kinetics. The predictive DBP model was competent to provide
quantitatively acceptable DBP data. The model attributes the formation of DBPs
to the continuous introduction of anthropogenic contaminants and the number of
pool users, which helps the interpretation of the special DBP profile in swimming
pool water.
Control
Of the three limiting factors to DBP formation in swimming pools, reducing
the input of anthropogenic contaminants is critical. With the purpose of reducing
DBP precursors and already-formed DBPs, filtration technologies using GAC and
BAC were investigated by bench-scale experiments. GAC adsorption and BAC
biodegradation were both found to be effective options on certain HAA species
removal.
Future research efforts should be focused on three aspects: (1) monitoring
and assessment of other emerging/remerging DBPs in swimming pools; (2) the
potential health impact of swimming pool exposure to high level HAAs; and (3)
cost-effective HAA removal technologies for swimming pools.
399
References
2. Florentin, A.; Hautemaniere, A.; Hartemann, P. Health effects of disinfection
by-products in chlorinated swimming pools. Int. J. Hyg. Environ. Health
2011, 214, 461–469.
3. De Laat, J.; Feng, W.; Freyfer, D. A.; Dossier-Berne, F. Concentration levels
of urea in swimming pool water and reactivity of chlorine with urea. Water
Res. 2011, 45, 1139–1146.
4. Friedman, M. S.; Roels, T.; Koehler, J. E.; Feldman, L.; Bibb, W. F.;
Blake, P. Escherichia coli O157:H7 outbreak associated with an improperly
chlorinated swimming pool. Clin. Infect. Dis. 1999, 29, 298–303.

5. Leoni, E.; Legnani, P.; Mucci, M. T.; Pirani, R. Prevalence of mycobacteria
in a swimming pool environment. J. Appl. Microbiol. 1999, 87, 683–688.
6. LaKind, J. S.; Richardson, S. D.; Blount, B. C. The good, the bad, and the
volatile: can we have both healthy pools and healthy people? Environ. Sci.
Technol. 2010, 44, 3205–3210.
7. Dyck, R.; Sadiq, R.; Rodriguez, M. J.; Simard, S.; Tardif, R. Trihalomethane
exposures in indoor swimming pools: a level III fugacity model. Water Res.
2011, 45, 5084–5098.
8. Kanan, A.; Karanfil, T. Formation of disinfection by-products in swimming
pool water: The contribution from filling water natural organic matter and
swimmer body fluids. Water Res. 2011, 45, 926–932.
9. Weaver, W. A.; Li, J.; Wen, Y.; Johnston, J.; Blatchley, M. R.; Blatchley, E.
R., III Volatile disinfection by-product analysis from chlorinated indoor
10. Xu, X.; Weisel, C. P. Dermal uptake of chloroform and haloketones during
bathing. J. Exposure Anal. Environ. Epidemiol. 2005, 15, 289–296.
11. Villaneuva, C. M.; Cantor, K. P.; Grimalt, J. O.; Malats, N.; Silverman, D.;
Marcos, R.; Rothman, N.; Real, F. X.; Dosemeci, M.; Kogevina, M. Bladder
bathing, showering, and swimming in pools. Am. J. Epidemiol. 2007, 165,
148–156.
12. Richardson, S. D.; DeMarini, D. M.; Kogevinas, M. M.; Fernandez, P.;
Marco, E.; Lourencetii, C.; Balleste, C.; Heederik, D.; Meliefste, K.;
McKaque, A. B.; Marcos, R.; Font-Ribera, L.; Grimalt, J. O.; Villanueva, C.
M. What’s in the pool? A comprehensive identification of disinfection
by-products and assessment of mutagenicity of chlorinated and brominated
swimming pool water. Environ. Health Perspect. 2010, 118, 1523–1530.
Garcia-Closas, R.; Serra, C.; Carrato, A.; Castano-Vinyals, G.; Samanic, C.;
Rothman, N.; Kogevinas, M. Polymorphisms M. Polymorphisms in GSTT1,
400
GSTZ1, and CYP2E1, disinfection by-products, and risk of bladder cancer
in Spain. Environ. Health Perspect. 2010, 118, 1545–1550.
Genotoxic effects in swimmers exposed to disinfection by-products in
swimming pools. Environ. Health Perspect. 2010, 118, 1531–1537.
15. Bessonneau, V.; Derbez, M.; Clement, M.; Thomas, O. Determinants of
chlorination by-products in indoor swimming pools. Int. J. Hyg. Environ.
Health 2011, 215, 76–85.
16. Cardador, M. J.; Gallego, M. Haloacetic acids in swimming pools: swimmer
and worker exposure. Environ. Sci. Technol. 2011, 45, 5783–5790.
17. Catto, C.; Sabrina, S.; Ginette, C. T.; Manuel, R.; Robert, T. Occurrence and
spatial and temporal variations of disinfection by-products in the water and

air of two indoor swimming pools. Int. J. Environ. Res. Publ. Health 2012,
9, 2562–2586.
18. Parinet, J.; Tabaries, S.; Coulomb, B.; Vassalo, L.; Boudenne, J. Exposure
levels to brominated compounds in seawater swimming pools treated with
chlorine. Water Res. 2012, 46, 828–836.
19. Yeh, R. Y. L.; Farre, M. J.; Stalter, D.; Tang, J. Y. M.; Molendijk, J.;
Escher, B. I. Bioanalytical and chemical evaluation of disinfection
by-products in swimming pool water. Water Res. 2014, 59, 172–184.
20. Fantuzzi, G.; Righi, E.; Predieri, G.; Ceppelli, G.; Gobba, F.; Aggazzotti, G.
Occupational exposure to trihalomethanes in indoor swimming pools. Sci.
Total Environ. 2001, 264, 257–265.
21. Chu, H.; Nieuwenhuijsen, M. J. Distribution and determinants of
trihalomethane concentrations in indoor swimming pools. Occup. Environ.
Med. 2002, 59, 243–247.
22. Tang, H. Disinfection byproduct precursors from wastewater organics:
Formation potential and influence of biological treatment processes. Ph.D.
Thesis, the Pennsylvania State University, University Park, PA, 2011.
23. Simard, S.; Tardif, R.; Rodriguez, M. J. Variability of chlorination by-product
occurrence in water of indoor and outdoor swimming pools. Water Res.
2013, 47, 1763–1772.
24. Wang, X.; Leal, M I G.; Zhang, X.; Yang, H.; Xie, Y. Haloacetic acids in
swimming pool and spa water in the United States and China. Front. Environ.
Sci. Eng. 2014, 8, 820–824.
25. Dotson, A.; Westerhoff, P.; Krasner, S. W. Nitrogen enriched dissolved
organic matter (DOM) isolates and their affinity to form emerging
disinfection by-products. Water Sci. Technol. 2009, 60, 135–143.
26. Walse, S. S.; Mitch, W. A. Nitrosamine carcinogens also swim in chlorinated
27. Hansen, K.; Willach, S.; Mosbak, H.; Andersen, H. R. Particles in swimming
pool filters – Does pH determine the DBP formation? Chemosphere 2012,
87, 241–247.
– Formation and mass transfer. Water Res. 2011, 45, 2681–2690.
401
29. Chowdhury, S.; Alhooshani, K.; Karanfil, T. Disinfection byproducts in
swimming pool: occurrences, implications and future needs. Water Res.
2014, 53, 68–109.
30. Bond, T.; Goslan, E. H.; Parsons, S. A.; Jefferson, B. A critical review of
trihalomethane and haloacetic acid formation from natural organic matter
surrogates. Environmental Technology Reviews 2012, 1, 93–113.
31. Farre, M. J.; Day, S.; Neale, P. A.; Stalter, D.; Tang, J. Y. M.; Escher, B.
I. Bioanalytical and chemical assessment of the disinfection by-product
formation potential: Role of organic matter. Water Res. 2013, 47,
5409–5421.
32. Weng, S.; Blatchley, E. R., III Disinfection by-product dynamics in a
chlorinated, indoor swimming pool under conditions of heavy use: National
swimming competition. Water Res. 2011, 45, 5241–5248.
33. Keuten, M. G. A.; Peters, M. C. F. M.; Daanen, H. A. M.; de Kreuk, M.

K.; Rietveld, L. C.; van Dijk, J. C. Quantification of continual anthropogenic
pollutants released in swimming pools. Water Res. 2014, 53, 259–270.
34. APHA. Standard Method for the Examination of Water and Wastewater, 20th
ed.; American Water Works Association: 1998.
35. Tang, H. L.; Chen, Y.-C.; Xie, Y. F. Quantification of disinfection by-product
formation potential in wastewater. Proceedings of IWA Micropol &
Ecohazard, July 11−13, 2011, Sydney, Australia.
36. Reckhow, D. A.; Singer, P. C.; Malcolm, R. L. Chlorination of humic
materials – by-product formation and chemical interpretations. Environ.
Sci. Technol. 1990, 24, 1655–1664.
37. Lee, J.; Jun, M. J.; Lee, M. H.; Eom, S. W.; Zoh, K. D. Production of
various disinfection byproducts in indoor swimming pool waters treated
with different disinfection methods. Int. J. Hyg. Environ. Health 2010, 213,
465–474.
38. Judd, S.; Jeffrey, J. A. Trihalomethane formation during swimming pool
water disinfection using hypobromous and hypochlorous acids. Water Res.
1995, 29, 1203–1206.
40. Judd, S. J.; Bullock, G. The fate of chlorine and organic materials in
swimming pools. Chemosphere 2003, 51, 869–879.
41. Li, J.; Blatchley, E. R., III Volatile disinfection byproduct formation
resulting from chlorination of organic-nitrogen precursors in swimming
42. Gunkel, K.; Jessen, H. J. The problem of urea in bathing water. Z. Gesamte
Hyg. 1988, 34, 248–250.
43. Erdinger, L.; Kirsch, F.; Sonntag, H. G. Potassium as an indicator
of anthropogenic contamination of swimming pool water. Zbl. Hyg.
Umweltmed. 1997, 200, 297–308.
44. Benoit, F. M.; Jackson, R. Trihalomethane formation in whirlpool spas.
Water Res. 1987, 21, 353–357.
402
45. Sadiq, R.; Rodriguez, M. J. Disinfection by-products (DBPs) in drinking
water and predictive models for their occurrence: a review. Sci. Total
Environ. 2004, 321, 21–46.
46. Judd, S. J.; Black, S. Disinfection byproduct formation in swimming pool
waters: A simple mass balance. Water Res. 2000, 34, 1611–1619.
47. Sohn, J.; Gary, A.; Cho, J.; Lee, Y.; Yoon, Y. Disinfectant decay and
disinfection by-products formation model development: chlorination and
ozonation by-products. Water Res. 2004, 38, 2461–2478.
48. Aggazzotti, G.; Fantuzzi, G.; Righi, E.; Predieri, G. Environmental
and biological monitoring of chloroform in indoor swimming pools. J.
Chromatogr. A 1995, 710, 181–190.
49. Duan, Q. Y.; Gupta, V. K.; Sorooshian, S. Shuffled complex evolution
approach for effective and efficient global minimization. J. Optimiz. Theory
App. 1993, 76, 501–521.

50. Glauner, T.; Waldmann, P.; Frimmel, F. H.; Zwiener, C. Swimming pool
water – fractionation and genotoxicological characterization of organic
constituents. Water Res. 2005, 39, 4494–4502.
51. Keuten, M. G. A.; Schets, F. M.; Schijven, J. F.; Verberk, J. Q. J. C.; van
Dijk, J. C. Definition and quantification of initial anthropogenic pollutant
release in swimming pools. Water Res. 2012, 46, 3682–3692.
52. Glauner, T.; Kunz, F.; Zwiener, C.; Frimmel, F. Elimination of swimming
pool water disinfection by-products with advanced oxidation processes
(AOPs). Acta Hydrochim. Hydrobiol. 2005, 33, 585–594.
403
Chapter 21
Occurrence and Formation of Disinfection

By-Products in Indoor U.S. Swimming Pools
Amer Kanan,1 Meric Selbes,2 and Tanju Karanfil*,3
1Departmentof Environmental and Earth Sciences, Al-Quds University,

Jerusalem, Palestine
Fairfax, Virginia 22030, U.S.A.

Clemson University, Anderson, South Carolina 29625, U.S.A.

Chlorination is commonly used to prevent the spreading of

waterborne infectious diseases in swimming pools. This
required disinfection practice also results in the formation of
undesirable disinfection by-products (DBPs) from the reactions
of chlorine with the organic matter (released by swimmers
or present in the pool filling water) and inorganics (i.e.,
bromide). The main objective of this research was to improve
our understanding of the occurrence and formation of different
classes of DBPs (trihalomethanes [THMs], haloacetic acids
[HAAs], halonitromethanes [HNMs], haloacetonitriles [HANs],
and nitrosamines) in indoor swimming pools operational
conditions in the U.S.. The results showed that the DBPs
in the investigated 23 swimming pools were far higher than
the drinking water regulation values in the U.S. Average
THMs, HAAs, HANs, HNMs, and N-nitrosodimethylamine
concentrations were 80 µg/L, 1541 µg/L, 19 µg/L, 5.4 µg/L,
and 27 ng/L, respectively. An increase in organic matter
released by the swimmers and bromide (from the filling water
or electrochemical generation of chlorine) levels in the water
increased the overall formation of DBPs. Increases in free
available chlorine, pH, and water temperature were shown to

enhance the formation of THMs and HAAs. These favorable
conditions lead to rapid formation (i.e. 3-6 hours) of THMs and
HAAs under swimming pool conditions.
Introduction
The average swimmer releases several grams of body fluids (BFs) and
excretions during an average swim (1–3). The disinfection of this organic
and inorganic matter from swimming pool water is essential for deactivating
the pathogenic microorganisms that can thrive on the released organic matter.
Swimming pool water is continuously circulated, filtered, and disinfected to
maintain clear and biologically safe (4). Chlorine is commonly used in swimming
pools and is added continuously to maintain free available chlorine (FAC) to
prevent microbial growth (5). While deactivating microorganisms, the reaction
of chlorine with the organic (e.g., hair, sweat, dead cells, saliva, cosmetics,
dust and urine) and inorganic (e.g., urea, nitrates, nitrites and free ammonia)
matter can form a wide array of DBPs including trihalomethanes (THMs),
haloacetic acids (HAAs), haloacetonitriles (HANs), haloacids, halodiacids,
iodo-THMs, haloaldehydes, halonitriles, haloketones (HKs), halonitromethanes
(HNMs), haloamides, haloalcohols, nitrosamines, combined available chlorine,
and 3-chloro-4-(dichloromethyl)-5-hydroxy-2(5H)-furanone (MX) and MX
homologues, etc. (6–11). Furthermore, bromide is introduced to when the pool
is being filled and/or as impurities when chlorine is generated electrochemically
from sodium chloride. Chlorine oxidizes bromide to HOBr that can react with
organic matter to produce brominated DBPs. Although there are currently no
federal DBP regulations in swimming pool waters in the U.S., some countries are
including THMs within the swimming pool regulations (12, 13).
The overall formation of DBPs is correlated to the amount of precursors such
as organic matter. It is also well known that temperature, pH, and chlorine dose
influence the formation of DBPs. Therefore, the formation of DBPs in pool waters
would be affected by the operation of the swimming pools. The pool operational
requirements and regulations in the U.S. are variable among different states
and enforced separately by each state department of health and environment.
Generally in the U.S., FAC should be maintained in the range of 1 to 5 mg/L in
the pool water to prevent microbial growth (5, 8, 14). The pH of swimming pool
water is continuously adjusted mainly to assure an acceptable level of disinfection
efficiency, comfortable water for swimmers, and elimination of the damage to
the swimming pool structures. Typically, the pH of the pool water is maintained
between 7.2 and 7.8 range (15). Swimming pool water temperature is maintained
constant usually within the range of 26 to 40˚C depending on pool type and usage.
Finally, the turnover period for the water in swimming pools typically ranges
from 4 to 8 hours (3, 5, 16). From a DBP formation perspective, the reaction
period corresponds to the time available for chlorine to react with precursors and
form DBPs before they are removed through the pool treatment processes. It is
well-known that DBPs including THMs and HAAs have fast formation kinetics
406
at the levels of chlorine dose in drinking waters (17, 18). It is postulated that
DBPs will form very rapidly in swimming pools due to much higher free chlorine
residual concentrations than in drinking waters.
Although intensive work has been devoted to investigate the formation of
regulated THMs and HAAs in drinking waters, there is relatively little known
about the formation and occurrence of DBPs in swimming pools. Moreover,
the effects of swimming pool operational parameters on the formation and
speciation of DBPs in swimming pools have not been systematically investigated.
Understanding the effect of operational parameters is essential for developing
approaches to control the formation of DBPs in pools and reduce exposure of
swimmers and swimming pool attendants to DBPs. Furthermore, to date there
is no information regarding the DBP formation kinetics under swimming pool
conditions.
The objectives of this study were to (i) investigate the occurrence of regulated
DBPs (THMs and HAAs), as well as emerging DBPs including HNMs, HANs, and
nitrosamines in 23 indoor swimming pools, (ii) examine the effects of different
swimming pool operational parameters such as FAC in pool water, pH, organic
carbon, temperature, bromide concentration on the formation and speciation of
THMs and HAAs and (iii) determine how fast THMs and HAAs are produced
under swimming pool conditions.

Samples Collection
Indoor swimming pools (n=23) located in South Carolina, Georgia and North
Carolina were sampled to investigate the occurrence of five classes of DBPs.
Furthermore, three of the swimming pools were monitored over a nine-month
period. Grab samples were collected from 30 cm below the water surface of the
pool. FAC, pH, and temperature of the swimming pool were obtained from the
facilities. A 125-mL amber bottle was used to collect total organic carbon (TOC)
and total nitrogen (TN) samples without any preservatives. A 125-mL amber
glass was used for the halogenated DBPs analysis (THMs, HAAs, HNMs, HANs),
and ammonium chloride was used to quench the free chlorine in these samples.
For nitrosamine analysis, a 1L sample was collected in an amber glass bottle and
quenched by sodium thiosulfate. All samples were transferred immediately to the
lab using an ice box for DBPs analyses, which was conducted either on the same
day of sampling or stored at 4°C for the following day.
Synthetic Swimming Pool Waters

Two synthetic pool waters were prepared using two different filling waters
and a body fluids analog (BFA) recipe developed by Goeres and co-workers
(19) to simulate the body fluids that are introduced in swimming pool water by
swimmers (9). The filling waters were collected from Myrtle Beach (MB) and
407
Startex-Jackson-Wellford-Duncan (SJWD) drinking water treatment plants in
South Carolina prior to the final addition of the disinfectant. These two filling
waters served as two different background matrices for the synthetic swimming
pool waters. The selected waters had distinct differences in their aromaticity of
the organic components determined by their specific ultraviolet absorbances at
254 nm wavelength (SUVA254). The MB water had a high aromaticity with a
SUVA254 value of 4-6 mg/L-m, whereas the SJWD had a lower level of aromatic
organics with a SUVA254 of 2-3 mg/L-m. Both waters were diluted to 1 mg/L
TOC to provide the same amount of background organic carbon in the synthetic
solution. For every experiment 1 mg/L TOC was from the fill water and the
remaining TOC was provided by the BFA. Here and after the two synthetic pool
waters will be referred to as BFA-MB and BFA-SJWD.
DBP FP and Kinetics Tests
THM and HAA formation potential (FP) tests were conducted in the presence
of an excess disinfectant. Chlorine stock solutions (500–2,000 mg/L) were
prepared by diluting sodium hypochlorite (~5% available free chlorine). The
chlorination FP tests were performed in 125 mL amber bottles filled headspace
free with the sample and capped with Teflon-lined PTFE caps. After mixing for
five minutes with magnetic stir bars, the bottles were stored headspace free in a
water bath for 5 days at 26 or 40°C. For the kinetic experiments an initial dose
of 100 mg/L chlorine was used. Each solution was chlorinated, and DBPs were
determined at different reaction times during 5 days (e.g., 0.5, 1, 3, 5, 7, 10, 15,
24, 48, 72, 96, 120 hours).
Analytical Methods
Analytical methods used in the study and their respective minimum reporting
levels (MRLs) are provided in Table 1. Either standard methods (SM) or USEPA
methods were used for DBP analyses with minor modifications. The detailed
information about these methods can be found elsewhere (20–23). For TOC,
TN and DBP analyses, samples were analyzed in duplicates. Error bars in all
the graphs show the variability due to multiple analysis (n=2) under the same
conditions.
408
Table 1. Analytical Methods and Minimum Reporting Levels
Parameter Unit Measurement Method Instrument MRL or Accuracya
TOCb mg/L SM 5310B TOC-VCHS & TNM-1, Shimadzu Corp., Japan 0.1
TNc mg/L High Temperature Combustion TOC-VCHS & TNM-1, Shimadzu Corp., Japan 0.1
UV Absorbanced cm-1 SM 5910 DU 640, Beckman Inst. Inc., USA ±0.005d
Bromide μg/L US EPA Method 300 ED 40, Dionex Corp., USA 10
pH - SM 4500-H+ 420A, Orion Corp., USA ±0.01e
THMs & HANsf μg/L US EPA Method 551.1 6890 GC-ECD, Agilent, USA 1.0
HAAsg μg/L SM 6251 Bg 6890 GC- ECD, Agilent, USA 1.0
HNMsh μg/L US EPA Method 551.1 6850 GC-ECD, Agilent, USA 0.7
409
Nitrosaminesi ng/L US EPA 521 Varian GC/MS/MS 3.0

FAC mg/L SM 4500-Cl F NA 0.1
a MRLs were determined in the lab in previous works (24). Accuracy as reported by the manufacturer. b Reagent grade potassium hydrogen phthalate
was used to prepare external standards. c Reagent grade potassium nitrate was used to prepare external standards. d At wavelengths of 254 nm using a
1- or 5-cm cell. Photo-metric accuracy (absorbance units). e Accuracy (pH units). f THMs and HANs were extracted by liquid-liquid extraction with
methyl-tert butyl ether (MtBE) and analyzed by GC-μECD. g HAAs were extracted by liquid-liquid extraction with MtBE, derivatized with diazomethane
and analyzed by GC-μECD. h HNM were extracted by liquid-liquid extraction with MtBE and analyzed by GC-μECD. i Nitrosamines were extracted by
solid phase extraction, eluted with dichloromethane and analyzed by GC-MS. THM (Trihalomethanes), HAN (Haloacetonitriles), HAA (Haloacetic acids),
HNM (halonitromethanes), SM (Standard Methods), EPA (Environmental Protection Agency), GC (Gas Chromatography), MS (Mass Spectrometer), FAC
(Free Available Chlorine), NA (Not Applicable).

Water Characteristics of the Selected Indoor Swimming Pools (n=23)
All of the indoor swimming pools selected in this study used chlorine as
disinfectant. Some of the sampled pools used calcium hypochlorite directly,
whereas others generated chlorine in situ electrochemically using sodium
chloride. Most of the pools were operated at an approximate temperature of
26°C and some at 34°C. Except for a single pool that used groundwater from a
local well, all the pools used the local water distribution system for both fill and
make-up water. Of the 23 pools under study, sand-filter filtration was used to
treat water, and micro-filter filtration used for the other two. The water samples
characteristics of all pools under study are shown in Table 2. The FAC in the
pools was between <0.1 and 4.0 mg/L, and the pH, TOC and TN ranged between
7.2 to 7.8, 3 to 23.6 mg/L, and 0.8 to 12.3 mg/L, respectively. The TOC was
higher than TN except for four of the swimming pools in this study.
Occurrence of DBPs in Indoor Swimming Pools (n=23)
The occurrence ranges of the five classes of DBPs in the selected twenty three
indoor pools are shown in Figure 1. The concentration of each DBP class, along
with their speciation for each swimming pool is summarized in Table 3 and Table
4. In general, carbonaceous DBPs concentrations were higher than nitrogenous
DBPs. In this survey, the HAAs were the highest measured class of DBPs,
followed by the THMs>HANs>HNMs>nitrosamines. The median, maximum,
and minimum HAAs were 960, 9005, and 172 µg/L, respectively. Trichloroacetic
acid (TCAA) and dichloroacetic acid (DCAA) were the dominant species among
the measured HAAs. Bromochloroacetic acid (BCAA) and bromodichloroacetic
acid (BDCAA) were also detected in all samples although at much lower levels
compared to DCAA and TCAA. In some of the samples dibromoacetic acid
(DBAA) and bromoacetic acid (BAA) were detected at concentrations <6 and <25
µg/L, respectively. These exceptionally (compared to drinking water samples)
high values of HAAs are attributed to (i) nonvolatile and soluble nature of these
compounds resulting in accumulate in the pool waters provided that pool water
dilution is not common in U.S., and (ii).human body fluids that have been shown
to have a high formation potential of HAAs (7, 9).
The median, maximum, and minimum THMs measured in the pools were 63,
213 and 26 µg/L, respectively, with chloroform the major THM determined in all
pools. Although the brominated THM species were at low levels in most cases,
there were relatively high in pools that use sodium chloride to generate chlorine
electrochemically and in the pool that was filled with groundwater. In 43% of the
pools, the total THMs was equal to or higher than 80 µg/L (MCL in U.S. drinking
water), while in all the pools, the measured THM was greater than 20 µg/L (the
maximum allowed in some European countries for swimming pools).
410
Table 2. Water Characteristics of the Selected Indoor Swimming Pools
Tem- Dis-
FAC
Swimming per- infe- Filter TOC TN
pH (mg-
Pool Code ature cta- Type (mg C/L) (mg N/L)
/L)
(°C) nt
S1 27 7.5 3.0 Cl2 Sand 4.3 1.4
S2C 18 7.4 3.0 Cl2 Sand 4.1 1.8
S2W 31 7.4 3.0 Cl2 Sand 11.3 4.7
S3S 29 7.4 1.4 Cl2 Microfilter 10.3 4.5
S3D 28 7.3 3.1 Cl2 Microfilter 5.5 2.8
S4C 27 7.4 4.0 Cl2* Sand 7.1 9.1
S4W 30 7.4 3.0 Cl2* Sand 6.5 3.2

S5 27 7.4 2.0 Cl2 Sand 3.0 1.2
S6L 29 7.6 2.0 Cl2* Sand 7.9 4.8
S6T 34 7.4 3.0 Cl2* Sand 7.7 2.5
S7 27 7.2 1.0 Cl2 Sand 7.8 4.3
S10 31 7.4 3.0 Cl2 Sand 8.4 5.2
S11L 28 7.2 2.0 Cl2 Sand 5.2 3.4
S11D 30 7.2 3.0 Cl2 Sand 5.3 3.6
S12L 27 7.5 2.7 Cl2 Sand 3.7 10.2
S12I 30 7.5 2.4 Cl2 Sand 4.3 7.4
S13 29 7.6 1.0 Cl2 Sand 3.9 1.2
S14 27 7.8 3.0 Cl2 Sand 13.8 12.3
S15L 27 7.7 2.8 Cl2* Sand 3.8 0.8
S15T 32 7.8 1.2 Cl2* Sand 7.5 1.2
S16 27 7.2 <0.1 Cl2 Sand 7.5 2.2
S17L 28 7.6 3.5 Cl2* Sand 7.3 4.1
S17T 33 7.4 3.5 Cl2* Sand 23.6 4.4
Median 28 7.4 3.0 7.1 3.6
Maximum 34 7.8 4.0 23.6 12.3
Minimum 18 7.2 <0.0 3.0 0.8
*Electrochemically generated chlorine. FAC (free available chlorine), TOC (total organic
carbon), TN (total nitrogen).
411
Figure 1. Box and Whisker plots of the occurrence of (A) THMs & HANs, (B)
HNMs, (C) HAAs, and (D) NDMA in indoor swimming pools (n=23).
The median, maximum and minimum HANs were 16, 53, and 5 µg/L,
respectively. The major haloacetonitrile quantified was dichloroacetonitrile
(DCAN) among the six measured HANs. In few pools, brominated acetonitriles
were detected at low levels (1-3 µg/L) except in the pool which was filled
with groundwater. In general, it was observed that the pool waters with high
levels of HAAs had high levels of HANs. Furthermore, when brominated
412
HAA species were high, the brominated HANs species also increased. Since
HANs hydrolyze to HAAs, it is expected to have a positive correlation between
these two classes of DBPs (25). Among the nine HNM species, only three
HNM species [trichloronitromethane (TCNM), bromonitromethane (BNM), and
bromochloronitromethane (BCNM)] were detected in the collected swimming
pool samples. Total HNMs were between 1.4 and 13.3 µg/L.
Among the nitrosamines analyzable using the current method (22), NDMA
was the only species detected in the selected 23 indoor swimming pool water
samples, with NDMA concentrations ranging between 2 and 83 ng/L with a
median of 17 ng/L. These levels of NDMA are somewhat higher than the drinking
water levels, which was consistent with the literature (6), and also indicating
that some NDMA precursors are released with body fluids and excretions.
Nitrosamines are mainly formed during chloramination of amines (26). Although
swimming pools are mostly chlorinated, the ammonia coming from urea would
inevitably lead to the formation of chloramines in the pools. Therefore, it is
not surprising to see some NDMA formation in swimming pools. Furthermore,
there was no correlation between NDMA and the other classes of DBPs, which
clearly indicates that their precursors differ from carbonaceous DBPs precursors,
a finding consistent with the literature (27, 28). Although the nitrogenous DBPs
(HANs, HNMs, NDMA) were detected at lower levels than carbonaceous DBPs,
they were still significant regarding their cyto- and geno-toxicity (29).
DBPs Measured in the Selected Indoor Pools (n=3) for 9-Months
Three indoor swimming pools were further monitored for a 9-month sampling
period. The operational conditions, water quality and DBPs occurrences in these
pools are summarized in Table 5. The average FAC fell within the range of 3
mg/L, which is well within the state guidelines in place regarding adequate pool
chlorination. Although the TN varied over a narrower range, between 2.0 and 5.0
mg/L in the three pools under study, the temperature and pH remained relatively
constant and the TOC fell within a wide range of 5.3 to 25.4 mg/L.
During the nine-month monitoring period for these three pools, a narrow
variation in the DBPs occurrence was observed in each (Table 5). The THM
concentrations in all of them ranged from 29 to 259 µg/L with an average of 108
µg/L. The highest THMs measured were in the warmest pool (S17T), which also
had the highest TOC measurements. Of the THM levels measured in the three
pools, chloroform was the most predominant (data not shown). The authors also
found higher levels of measured BDCM in pools S17L and S17T because sodium
chloride was used in both to generate chlorine electrochemically. The range of the
BDCM in these three pools was 0 to 29 µg/L. The HAAs in all three pools was
also higher than the THMs, which fell within a range between 667 to 11695 µg/L,
with an average of 2820 µg/L DCAA. The TCAA were the dominant species,
and BDCAA was the major brominated HAA species. The maximum amount
of BDCAA was 151 µg/L and the minimum was 9 µg/L with an average of 57
µg/L. The levels of all of the dibrominated species were much lower (<20 µg/L),
however.
413
Table 3. Carbonaceous-DBPs Occurrence in Indoor Swimming Pools Water (n=23)
Swimming Trihalomethanes (μg/L) Haloacetic Acids (μg/L)

Pool Code
TCM BDCM DBCM TBM Total BAA DCAA BCAA TCAA DBAA BDCAA DBCAA Total
S1 41 1 <MRL ND 42 <MRL 100 1 87 <MRL 8 1 198
S2C 49 1 <MRL ND 51 <MRL 314 3 224 <MRL 13 1 555
S2W 119 3 <MRL ND 122 1 896 14 718 <MRL 36 <MRL 1665
S3S 77 2 <MRL ND 80 <MRL 712 5 1537 <MRL 21 <MRL 2276
S3D 62 1 <MRL ND 63 <MRL 937 7 776 <MRL 16 1 1738
S4C 37 1 <MRL ND 38 <MRL 688 10 789 1 27 3 1518
S4W 49 3 <MRL ND 53 1 81 4 241 <MRL 60 5 392
414
S5 26 1 <MRL ND 28 <MRL 82 3 141 <MRL 22 2 251

S6L 72 3 <MRL ND 76 1 928 30 552 2 48 3 1563
S6T 72 12 4 1 90 1 1033 36 351 4 36 7 1468
S7 113 1 <MRL ND 114 <MRL 1139 5 418 <MRL 22 <MRL 1585
S10 89 2 <MRL ND 91 <MRL 1297 4 416 <MRL 13 <MRL 1731
S11L 29 2 <MRL ND 31 <MRL 123 2 76 <MRL 15 <MRL 216
S11D 111 2 <MRL ND 113 <MRL 360 3 149 <MRL 11 <MRL 523
S12L 25 1 <MRL ND 26 <MRL 162 1 95 <MRL 8 3 269
S12I 32 1 <MRL ND 34 <MRL 242 2 216 <MRL 22 2 484
S13 74 28 10 1 114 1 73 13 100 2 55 15 260

Swimming Trihalomethanes (μg/L) Haloacetic Acids (μg/L)
Pool Code TCM BDCM DBCM TBM Total BAA DCAA BCAA TCAA DBAA BDCAA DBCAA Total
S14 47 3 <MRL ND 50 <MRL 52 2 92 <MRL 25 2 172

S15L 38 6 1 ND 45 2 115 21 102 5 47 14 306
S15T 121 5 1 ND 127 2 690 31 183 6 45 2 960
S16 61 2 <MRL ND 63 <MRL 549 5 568 <MRL 20 1 1143
S17L 82 11 2 ND 95 5 504 106 288 25 110 32 1070
S17T 207 6 <MRL ND 213 4 6787 176 1925 16 93 4 9005
Median 63 960
Maximum 213 9005
415
Minimum 26 172
TCM (Trichloromethane), BDCM (bromodichloromethane), DBCM (dibromochloromethane), TBM (tribromomethane), CAA (chloroacetic acid), BAA
(bromoacetic acid), DCAA (dichloroacetic acid), BCAA (bromochloroacetic acid), TCAA (trichloroacetic acid), DBAA (dibromoacetic acid), DBCAA
(dibromoacetic acid), TBAA (tribromoacetic acid), MRL (Minimum Reporting Level), ND (Not Detected).

Table 4. Nitrogenous-DBPs occurrence in indoor swimming pools water (n=23).
Haloacetonitriles (μg/L) Halonitromethanes (μg/L)
Swimming Nitrosamines (ng/L)
Pool Code CAN TCAN DCAN BAN BCAN DBAN Total TCNM BNM BCNM Total NDMA
S1 1 ND 8 ND ND ND 9 <MRL <MRL 2.4 2.9 30
S2C ND ND 4 ND ND ND 5 0.8 <MRL 7.5 8.3 9
S2W 1 ND 22 ND 1 ND 25 <MRL <MRL 1.3 1.9 83
S3S 1 ND 27 ND 1 ND 30 0.7 <MRL 6.3 7.0 28
S3D 1 ND 7 ND ND ND 8 0.9 <MRL 3.9 4.8 14
S4C 1 ND 15 ND 1 ND 16 <MRL <MRL 2.5 3.0 16
S4W 1 ND 21 ND 2 ND 24 <MRL <MRL 1.1 1.7 18
416
S5 1 ND 14 ND 1 ND 15 <MRL <MRL 3.2 3.7 28

S6L 1 ND 12 ND 1 ND 15 0.8 <MRL 2.6 3.6 9
S6T 1 ND 13 ND 4 2 21 <MRL 2.2 3.8 6.5 60
S7 1 ND 14 ND ND ND 15 2 <MRL 3.4 5.4 9
S10 2 ND 16 ND ND 1 18 2.3 <MRL 11 13.3 3
S11L 1 ND 11 ND ND ND 13 <MRL <MRL 3.6 4.1 12
S11D 1 ND 7 ND ND 1 9 0.9 <MRL 4.6 5.5 17
S12L 1 ND 15 ND ND ND 16 1 <MRL 4.3 5.4 2
S12I 1 ND 17 ND ND ND 18 1.1 <MRL 5.5 6.6 6
S13 1 ND 27 1 13 5 47 <MRL 2.0 4.7 7.3 13

Haloacetonitriles (μg/L) Halonitromethanes (μg/L)
Swimming Nitrosamines (ng/L)
Pool Code CAN TCAN DCAN BAN BCAN DBAN Total TCNM BNM BCNM Total NDMA
S14 1 ND 13 ND 1 ND 15 0.9 <MRL 3.7 4.7 18

S15L 1 ND 7 ND 2 1 11 <MRL 0.7 1.5 2.7 15
S15T 1 ND 4 ND ND ND 6 <MRL <MRL 0.8 1.4 76
S16 1 ND 15 ND ND ND 16 0.8 <MRL 2.6 3.4 53
S17L 1 ND 16 1 3 1 22 0.7 1.6 5.8 8.1 17
S17T 3 1 47 ND 2 ND 53 1.4 1 6.8 9.2 42
Median 16 4.8 17
Maximum 53 13.3 83
417
Minimum 5 1.4 2
CAN (chloroacetonitrile), TCAN (trichloroacetonitrile), DCAN (dichloroacetonitrile), BAN (bromoacetonitrile), BCAN (bromochloroacetonitrile, DBAN
(dibromoacetonitrile), TCNM (trichloronitromethane), BNM (bromonitromethane), BCNM (bromochloronitromethane), NDMA (N-nitrosodimethylamine),
MRL (Minimum Reporting Level), ND (Not Detected).

Table 5. Characteristics of the Three Selected Indoor Swimming Pools Waters Monitored for 9-Months
TOC TN FAC Temp. THMs HAAs
Swimming Pool/ Sampling date pH
mg/L mg/L mg/L °C µg/L µg/L
22/05/2009 5.6 2.1 2.6 28 7.8 60 2039

24/05/2009 5.6 2.1 3.9 28 7.3 52 1648
30/05/2009 6.2 2.1 4.6 28 6.7 53 1725
02/06/2009 6.4 2.1 3.2 28 7.0 56 1691
06/06/2009 6.3 2.1 0.7 26 7.0 48 1708
09/06/2009 6.6 2.1 2.1 27 7.6 40 1674
S16
18/10/2009 5.3 2.0 2.3 28 7.5 81 1804
418
25/10/2009 5.5 2.0 0.6 27 7.5 38 1673

22/12/2009 NM NM NM NM NM NM NM
10/01/2010 6.9 2.4 1.6 27 7.7 29 1479
17/01/2010 7.3 2.4 2.9 27 7.6 54 1005
17/02/2010 7.5 2.2 0.1 27 7.2 63 1143

22/05/2009 6.8 3.4 3 28 7.6 92 NM
24/05/2009 6.1 3.4 3.5 28 7.6 91 1437

30/05/2009 6.5 3.5 3 28 7.6 83 1392
02/06/2009 7.0 3.5 1.5 28 7.6 77 1377
06/06/2009 7.4 3.6 4 28 7.6 77 1434
09/06/2009 7.2 3.6 3 28 7.6 84 1470
S17L
18/10/2009 6.6 4.3 3.5 27 7.2 79 1174
25/10/2009 6.6 4.2 3 28 7.2 79 1196
419
22/12/2009 5.8 4.2 3 27 7.6 64 955

10/01/2010 6.3 4.3 4.5 26 7.5 72 793
17/01/2010 6.4 4.4 3.5 29 7.6 91 667
17/02/2010 7.3 4.1 3 27 7.6 95 1070

Table 5. (Continued). Characteristics of the Three Selected Indoor Swimming Pools Waters Monitored for 9-Months
22/05/2009 14.5 2.1 3 33 7.6 163 NM

24/05/2009 14.8 2.2 2 33 7.6 161 2244
30/05/2009 15.5 2.3 4 33 7.6 199 2397
02/06/2009 16.5 2.1 3 33 7.6 192 2427
06/06/2009 16.9 2.3 4 33 7.6 137 2869
09/06/2009 16.9 2.2 2 33 7.6 160 2884
S17T
18/10/2009 21.7 3.1 3.5 34 7.8 259 9691
420
25/10/2009 20.9 3.2 6 33 7.2 249 9348

22/12/2009 22.3 4.7 3.5 34 7.6 176 11695
10/01/2010 25.2 5.0 3 34 7.6 182 7392
17/01/2010 25.4 5.0 3.5 32 7.4 188 6291
17/02/2010 23.6 4.4 3 33 7.5 213 9005
FAC (free available chlorine), TOC (total organic carbon), TN (total nitrogen), THM (Trihalomethanes), HAA (Haloacetic acids), NM (Not measured).

The Effects of Swimming Pool Operational Conditions on DBPs
The swimming pool operational parameters that affect both the THM and
HAA formation is presented in Figure 2. An overall comparison of the results
between the filling background waters (MB and SJWD) showed that formation of
THMs and HAAs was approximately 10% higher in BFA-MB than in BFA-SJWD
water. Although both synthetic pool waters had the same TOC concentration (1
mg/L from the filling water and 5 mg TOC from BFA), there was a slightly higher
rate of THM and HAA formation in the MB than in the SJWD water, which was
due to the aromatic nature of the MB source water. Specifically, the MB water
exhibited a somewhat higher SUVA254 (~2.0 L/mg-m) compared to the SJWD
water (~1.7 L/mg-m), indicating it’s slightly higher aromatic character over SJWD.
An analysis of each parameter, except the filling waters is provided below.
Effect of FAC
FAC levels 1, 3, and 5 mg/L were tested in this study. It was determined than
an increase in FAC concentrations resulted in a slight increase in THM formation:
5% for BFA-MB and 14% for BFA-SJWD (Figure 2). There was, however, a
distinct increase in HAA formation observed, which was caused when the chlorine
dose was increased from 1 mg/L to 5 mg/L. Specifically, the HAA formation
increased 25% and 27%, respectively for the BFA-MB and BFA-SJWD (Figure
2).
This dependence was attributed to the reaction(s) of chlorine with albumin in
BFA. An increase in dosage in turn increased the rate of albumin decomposition
(hydrolysis), which in turn generated more HAA precursors (free amino acids).
The lesser effect of the chlorine concentrations on the THM than HAA was due to
the effect of the chlorine dose on the formation of different DBPs. After satisfying
the demand, a linear relationship has been previously established between the
demand for chlorine demand and THM formation (30), and a weak correlation
between the chlorine levels and chloroform formation (31). Consistent with these
observations, the results indicate that once the chlorine demand was satisfied, there
was no effect of the chlorine dose on the THM formation.
Effect of Temperature
To examine the effect of temperature on DBP formation, the two synthetic

swimming pool waters were also chlorinated at 40°C, the highest possible
temperature allowable in therapeutic swimming pools, hot tubs, and whirlpool
spas. Although this increase in temperature from 26°C to 40°C nearly doubled
the formation of THMs in synthetic pool waters, there was only a 60% increase
in HAA formation (Figure 2). These results clear indicate that exposure to DBPs
in hot tubs or any other type of elevated temperature indoor swimming pools will
421
be higher. The increase in DBP formation with temperature can be explained
by reaction of un-reacted DBPs precursors with chlorine and increase in the
reactivity of chlorine with precursors (35, 36).
Figure 2. The effect of operational parameters on DBPs formation in swimming

pools waters. BFA-MB, BFA-SJWD
422
Effect of pH
The results show that the formation of THMs and HAAs increased with an
increase in pH (Figure 2). It is also known that THM formation increases with an
increase in pH, while the trend is the opposite for HAA formation (32). However,
in this study, an increase in pH resulted in both an increase in both THM and
HAA formation. The opposite behavior that was observed for pH dependence
of HAAs in synthetic swimming water solution was once again attributed to the
hydrolysis of albumin, one of the main components of the BFA, and its reactivity
with chlorine. The observed pH trends for HAAs also indicated that the filling
water NOM was not a major contributor to HAA formation in synthetic pool waters
because in drinking waters, the HAA concentration decreased with an increase in
pH, as mentioned above. Finally, the results clearly indicate that the reduction of
pH from 8.0 to 6.0 would result in a decrease in both THM and HAA formation by
40-60%. This decrease in pH would also result in a corresponding decrease in the
disinfection efficiency of chlorine, particularly below the pKa of chlorine where
HOCl is a more effective disinfectant agent than OCl- (33, 34).
Effect of Bather Load
Although TOC cannot be controlled directly by the facility, there is a

correlation to the number of bathers and their BFs in a swimming pool. Thus,
pool managers can control the TOC level by controlling the number of bathers
using the pool at a given time. As predicted, the formation of both THMs and
HAAs increased with an increase in the concentration of BFAs. This dependence
on TOC concentration confirms a substantial correlation between the DBPs
formation in swimming pools with the population of swimmers (the main TOC
source).
Effect of Bromide
Bromide was spiked at three levels in addition to the ambient fill waters
concentration. The change in THMs and HAAs speciation with the increasing
bromide concentrations is shown in Figure 3. This increase in bromide
concentrations yielded a corresponding increase in both THM and HAA
concentration, which was expected since the brominated species increased while
the chlorinated species decreased. However, bromide caused a greater increase in
THM than HAA formation. The overall mass of THM concentration increased
by 65-72%, 106-116%, and 162-167% in synthetic pool waters at 100, 300,
and 600 µg/L bromide concentrations, respectively compared to the ambient fill
waters THM formation. However, during these same experiments the increase
in HAA concentration was only within a fairly limited range of 22 and 39%.
Bromine incorporation factor (BIF) “n” is the molar amount of bromine in the
423
halogenated compound (THM or HAA) divided by the molar total of that halogen
(37). That is to say when n = 0 only chlorinated compounds form whereas when
n>0 brominated species begin to appear. Subsequent BIF analysis determined
that the presence of bromide increased the formation of brominated THM over
HAA formation (data not shown).
Specifically, the incorporation factor a synthetic swimming water sample
for a bromide level of 600 µg/L was 1.3 for THMs but only 0.8 for HAAs.
This difference in bromide incorporation is also consistent with higher bromine
incorporation of THMs than HAAs in sample testing (38). These results clearly
demonstrate that the use of water with low bromide levels to either fill swimming
pools and as make up water is critical to reduce DBP formation, especially THMs.
Thus, bromide impurities should be minimized, if possible, when generating
chlorine from sodium chloride. Furthermore, the use of seawater or saline water
as either make-up or filling water should be avoided to reduce the DBPs formation
in swimming pools.
Figure 3. The effect of bromide on the formation and speciation of THMs and
HAAs during chlorination of BFA-MB and BFA-SJWD synthetic pool water.
424
DBPs Formation as Function of Time
DBP formation as function of time are presented in Figure 4 for three synthetic
pool waters: (i) a pool water at 6 mg-C/L (5 mg-C/L from BFA and 1 mg-C/L from
MB as the background filling water), (ii) 5 mg-C/L BFA, and (iii) 1 mg-C/L BFA.
The last two samples were used to investigate the formation kinetics from BFAs
alone and the formation kinetics at two Cl2/TOC ratios. Results indicate that THM
formation (within a range of 53% to 68%) occurred during the first five hours of a
five-day formation cycle in the absence and the presence of bromide for synthetic
pool water (BFA-MB) and BFA (Figure 4).
Figure 4. THM formation fraction during five days (A) without bromide and (B)
with bromide (200 µg/L), HAA formation fraction during five days (C) without
bromide and (D) with bromide (200 µg/L). THMt (THM formed at time t), THM120
(THM formed at 120 hours), HAAt (HAA formed at time t), HAA120 (HAA formed
at 120 hours). ◆ BFA-MB (5+1 mg-C/L), □ BFA (5 mg-C/L), ▵ BFA (1 mg-C/L)
This is a faster rate of THMs production at early contact times compared to

previous studies using several chlorination scenarios, which produced only 15-
30% of THMs after at least five hours after treatment (17, 18). Although the fast
425
formation of THMs observed in this study indicates that a significant portion of
THMs formed in the pool prior to treatment, only 15% to 30% of the five-day HAA
formation occurred during the first 5 hours under the same experimental conditions
(Figure 4). This rate was lower as compared to the formation rate of THMs. The
slower rate of HAA formation indicates that there is more opportunity to remove
their precursors through swimming pool water treatment processes during pool
water turnover.
Two BFA solutions at 1 and 5 mg TOC/L showed similar THM formation
patterns indicating very little effect of the Cl2/TOC ratio (100 vs. 20) on the
THM formation rate (Figure 4). Thus, the formation rates observed in this study
are representative of THM formation in swimming pools. There was, however,
some difference between the trends of the BFA solutions at 1 and 5 mg-C/L,
indicating that the decrease in Cl2/TOC ratio from 100 to 20 increased the HAA
formation rate. This difference was attributed to the changes in the formation
rates of di-halogenated HAAs vs. tri-halogenated HAAs (data not shown). Since
there is always a high FAC concentration in the pool water, the HAA formation
rates in swimming pools are likely to be rapid for HAAs. This is especially true
for swimming pools in the U.S. where the dilution of pool water with the filling
water is not regularly practiced and the pool water is not replaced for long time.
Therefore, chlorine demand of filling water is exhausted at the very early period
of operation after filling the pool. Afterwards, the BF components serve as the
primary catalyst for THM and HAA formation.
The presence of NOM (i.e., BFA-MB synthetic pool water) (BFA = 5 mg-C/L)
reduced the formation rate of THMs by ~20% during the first 24 hours as compared
to the absence of NOM. This behavior is due to the higher reactivity of NOM
compared to BFA with chlorine to form THM but which occurs at a slower rate,
as compared to BFA. Concurrently, the presence of NOM (i.e., BFA-MB) did not
affect the formation rate of HAAs, indicating that while the filling water NOM can
affect the THM formation, the formation rate of HAAs remains constant.
There were small differences in the formation rates under ambient bromide
levels for both THMs and HAAs. In the presence of bromide, the difference in
the formation rates of both DBPs diminished. Since halogenation reactions are
faster with bromide than chloride (39–41), higher formation rates observed in this
study is attributed to the formation of brominated DBP species from NOM at high
bromide levels.
The overall formation rate of THMs was higher than that of HAAs in
swimming pools Figure 5. THMs were formed almost instantaneously when
chlorine reacted with BFs, whereas HAA formation occurred at a slower
rate. Practically, these results indicate that within the typical turnover time of
swimming pools most HAA precursors can be removed by the treatment system
whereas THM precursors cannot. Thus, precursors control (i.e., the release of
human BFs) will be critical in reducing the formation of THMs in swimming
pools. The precursor control requires improving the practices and behavior of
swimmers, and implementing more strict hygienic conditions by swimming
pools utilities. Better hygienic conditions include showering immediately before
swimming and not releasing urine intentionally during swimming.
426
Figure 5. ◊ THMs and ▪ HAAs formation during 5-day reaction from BFA.
Conclusions
The DBPs in the 23 swimming pools that were the subject of this study were
far higher than the drinking water regulation values in the U.S., with THM levels
ranging between 26 and 213 µg/L with an average of 80 µg/L. The HAAs ranged
between 173 and 9005 µg/L with an average of 1541 µg/L. HNMs ranged between
1.4 and 13.3 µg/L with an average of 5.4 µg/L. HANs ranged between 5 and 53
µg/L with an average of 19 µg/L. The NDMA ranged between 2 and 83 ng/L with
an average of 26.5 ng/L. The electrochemically generation of chlorine increased
the brominated species of halogenated DBPs (THMs, HAAs, HNMs, HANs).
Furthermore, during the sampling period for nine months, the water quality (TOC,
TN, pH, FAC) and DBPs concentrations (THMs, HAAs) in the swimming pools
remained relatively constant. However, some fluctuation was observed likely due
to time of the year and the specific pool events/activities on the time of sampling.
The formation and speciation of THMs and HAAs were also investigated
under various disinfection and operation conditions typically used in U.S.
swimming pools. Although the increases in free available chlorine, pH, TOC,
water temperature, and bromide levels in the water increased overall DBP
formation, these factors affected the different classes of DBPs at different
427
magnitudes. Higher levels of free available chlorine increased the HAA levels
more than the THMs. The temperature effect was greater on the formation of
THMs than for HAAs whereas contact time increased HAAs more than THMs.
The authors also determined that under swimming pool related conditions,
DPB formation was quite rapid, with an appreciable percentage of the increase
occurring in the first 3-6 hours, which is the typical turnover time for the pool
water. Moreover, THM formation was faster than HAA formation, with 53 to
68% of five-day THMs formed within the first 3-6 hours and 15 to 30% of five-day
HAAs formed during the first six hours. Although it is possible to reduce DBP
formation by the controlling operational parameters (pH, free available chlorine,
bather load, the number of swimmers in a pool in a 24 hour period- or dilution),
these fast formation rates imply that DBP control strategies in swimming pools
should mainly focus on control the DBP precursors at the source (i.e., swimmers).
References
1. Batjer, K.; Cetinkaya, M.; Duszeln, J. V.; Gabel, B.; Lahl, U.; Stachel, B.
Chemosphere 1980, 9, 311–316.
2. Thacker, N. P.; Nitnaware, V. Bull. Environ. Contam. Toxicol. 2003, 71,
633–640.
Frimmel, F. H. Environ. Sci. Technol. 2007, 41, 363–372.
4. Surveillance for waterborne disease and outbreaks associated with
recreational water use and other aquatic facility-associated health events
- United States, 2005-2006; Center for Disease Control and Prevention:
Atlanta, GA, 2008.
5. Guidelines for Safe Recreational Water Environments, volume 2: swimming
pools and similar environments; World Health Organization, 2006.
6. Walse, S.; Mitch, W. Environ. Sci. Technol. 2008, 42, 1032–1037.
7. Kanan, A. Occurrence and Formation of Disinfection Byproducts in Indoor
Swimming Pools Water. Ph.D. Thesis, Clemson University, Clemson, SC,
2010.
8. Richardson, S. D.; DeMarini, D. M.; Kogevinas, M. Environ. Health
Perspect. 2010, 118, 1523–1530.
9. Kanan, A.; Karanfil, T. Water Res. 2011, 45, 926–932.
10. Chowdhury, S.; Al-Hooshani, K.; Karanfil, T. Water Res. 2014, 53, 68–109.
11. Teo, T. L.; Coleman, H. M.; Khan, S. J. Environ. Int. 2015, 76, 16–31.
12. Borgmann-Strahsen, R. Int. Biodeterior. Biodegrad. 2003, 51, 291–297.
13. Barbot, E.; Moulin, P. J. Membrane Sci. 2008, 314, 50–57.
14. Weisel, C. P.; Richardson, S. D.; Nemery, B.; Aggazzotti, G.; Baraldi, E.;
Blatchley, E. R.; Blount, B. C.; Carlsen, K. H.; Eggleston, P. A.;
Frimmel, F. H.; Goodman, M.; Gordon, G.; Grinshpun, S. A.; Heederik, D.;
Kogevinas, M.; LaKind, J. S.; Nieuwenhuijsen, M. J.; Piper, F. C.; Sattar, S.
A. Environ. Health Perspect. 2009, 117, 500–507.
15. Certified Pool-Spa Operator Handbook; National Swimming Pool
Foundation: Colorado Springs, CO, 2006.
428
16. R. 61-51 Public Swimming Pools - Bureau of Water; South Carolina
Department of Health and Environmental Control: Columbia, SC, 2007.
17. Gallard, H.; Gunten, U. Water Res. 2002, 36, 65–74.
18. Nikolaou, A. D.; Lekkas, T. D.; Golfinopoulos, S. K. Chem. Eng. J. 2004,
100, 139–148.
19. Goeres, D. M.; Palys, T.; Sandel, B. B.; Geiger, J. Water Res. 2004, 38,
3103–3109.
20. Method 300.0 - Determination of Inorganic Anions by Ion Chromatography;
U.S. Environmental Protection Agency: Washington, DC, 1993.
21. Method 551.1 - Determination of Chlorination Disinfection Byproducts,
Chlorinated Solvents, and Halogenated Pesticides/Herbicides in Drinking
Water by Liquid–Liquid Extraction and Gas Chromatography With
Electron-Capture Detection, revision 1.0.; U.S. Environmental Protection
Agency: Washington, DC, 1995.

22. Method 521 - Determination of Nitrosamines in Drinking Water by Solid
Phase Extraction and Capillary Column Gas Chromatography with Large
Volume Injection and Chemical Ionization Tandem Mass Spectrometry (MS/
MS), version 1.0.; U.S. Environmental Protection Agency: Washington, DC,
2004.
23. Standard Methods for the Examination of Water and Wastewater, 21st ed.;
APHA, AWWA, and WEF, 2005.
24. Hu, J. Exploring formation and distribution of halonitromethanes in drinking
waters, Ph.D. Thesis, Clemson University, Clemson, SC, 2009.
25. Glezer, V.; Harris, B.; Tal, N.; Iosefzon, B.; Lev, O. Water Res. 1999, 33,
1938–1948.
27. Uzun, H.; Kim, D.; Ates, N.; Karanfil, T. In Proceedings of AWWA Annual
Conference and Exposition; Dallas, TX, 2012.
28. Uzun, H.; Kim, D.; Karanfil, T. Water Res. 2015, 69, 162–172.
29. Plewa, M. J.; Wagner, E.; Muellner, M.; Hsu, K.; Richardson, S. ACS Symp.
Ser. 2008, 995, 36–50.
30. Boccelli, D. L.; Tryby, M. E.; Uber, J G.; Summers, R. S. Water Res. 2003,
37, 2654–2666.
31. Garcia-Villanova, R. J.; Garcia, C.; Cesar; Gomez, J. A.; Garcia, M. P.;
Ardanuy, R. Water Res. 1997, 31, 1405–1413.
32. Hua, G.; Reckhow, D. A. J. Am. Water Works Assoc. 2008, 100, 82–89.
33. Edzwald, J. K. Water Quality and Treatment: A Handbook of Community
Water Supplies, 4th ed.; McGraw-Hill: New York, 1990.
34. White, G. C. Handbook of Chlorination and Alternative Disinfectants, 4th
ed.; Wiley: New York, 1999.
35. Weisel, C. P.; Chen, W. J. Risk Anal. 1994, 14, 101–106.
36. Al-Omari, A.; Fayyad, M.; Qader, A. Environ. Model. Asses. 2004, 9,
245–52.
37. Gould, J. P.; Fitchorn, L. E.; Urheim, E. Formation of brominated
trihalomethanes: Extent and kinetics. In Water Chlorination: Environmental
Impact and Health Effects, Chemistry and Water Treatment; Jolley, R. L.,
Ed.; Ann Arbor Science, Ann Arbor, Michigan, 1983; Vol. 4.
429
38. Kitis, M.; Karanfil, T.; Wigton, A.; Kilduff, J. E. Water Res. 2002, 36,
3834–3848.
39. Westerhoff, P.; Chao, P.; Mash, H. Water Res. 2004, 38, 1502–1513.
40. Adin, A.; Katzhendler, J.; Alkaslassy, D.; Rav-Acha, C. Water Res. 1991,
25, 797–805.
41. Cowman, G. A.; Singer, P. C. Environ. Sci. Technol. 1996, 30, 16–24.
430
Chapter 22
Ion Mobility Spectrometry To Monitor

Trichloramine in Indoor Pool Air
C. Zwiener* and C. Schmalz
Environmental Analytical Chemistry, University of Tuebingen,

Hoelderlinstrasse 12, 72074 Tuebingen, Germany
Trichloramine (TCA) is a volatile, irritant disinfection

by-product formed by reactions of nitrogenous compounds
with chlorine. TCA analysis was performed by a hand-held ion
mobility spectrometer (IMS) with direct intake of gas samples
from indoor air environments. Ionization with a 63Ni beta
emitter produced chloride ions with a characteristic ion mobility
which are completely separated from bromide and other ions by
IMS. The limit of quantification for TCA is 0.1 mg/m3, sufficient
for the TCA guideline value of 0.2 mg/m3 and therefore to use
IMS for monitoring in indoor pool air. The contribution of
further interfering chlorinated disinfection by-products (DBPs)
to the chloride signal from TCA has been estimated to be less
than 15 %. The mobility, the short measurement times of about
20 s and the stability of the calibration curve of the instrument
qualify IMS to determine the spatial and temporal variability of
TCA in indoor pool air. This was demonstrated for the increase
of TCA concentrations dependent on the number and activity
of swimmers. An immediate increase of TCA was detected
after a school class entered the pool, which can be explained
by the sudden increase of the mass transfer coefficient of TCA
from water to air. A further example shows the acquisition of
a spatial profile and the increase of TCA from the inlet to the
outlet region of a non-swimmer area.

Introduction
Swimming is a healthy sport activity which can provide benefits for people
of all ages and physical abilities. Sufficient hygienic and chemical water
quality is a prerequisite to keep the positive aspects of aquatic activities. To
guarantee the hygienic safety of swimming pool water (SPW) disinfection,
mostly with chlorine, is used according to the recommendations from the World
Health Organization (WHO) and to further national standards for swimming
pools in order to inactivate pathogens (1). The chlorine concentrations used
are a compromise between sufficient disinfection and minimized formation of
disinfection by-products. Chlorine levels applied in Germany are in the range of
0.3 and 0.6 mg/L according to the German Industrial Norm DIN 19643. In other
countries up to 3 mg/L chlorine are typically used for pool water disinfection
(United States of America, Australia and other European countries). Public pools
in Germany rely primarily on the treatment scheme, operation, and surveillance
described in DIN 19643. Despite minor differences in threshold values, applied
chemical concentrations, and the requirement for an initial flocculation step, DIN
19643 can serve as a good example of pool water treatment that is similar to that
in much of Western Europe, North America, and Australia (2).
During operation the swimming water is continuously loaded by compounds

which are introduced by bathers and permanently cycled through the treatment.
The constituents of the bather load are from body fluids, skin, hair, and cosmetics.
Further compounds from natural organic matter are introduced to the pool from
filling water. Since there is no efficient elimination of the organic bather load,
during the continuous recirculation and dosing of reactive chlorine, disinfection
by-products are formed by reactions between organic and inorganic water
constituents and chlorine (2). Most prevalent DBPs are trihalomethanes (THMs)
and halogenated acetic acids (HAAs) (3). From nitrogen containing precursors
a variety of nitrogenous DBPs has been found like halogenated acetonitriles
(HAcNs), nitrosamines (e.g. NDMA), and chloramines (2–6). Some DBPs are
carcinogenic, fetotoxic and/or irritant to airways (7). More than 100 DBPs have
been identified in swimming pool waters, many of them were brominated and
nitrogen-containing DBPs. The mutagenicity of pool water has been found to
be in the same range of drinking water (1,200 revertants/L-equivalents in strain
TA100–S9 mix) (3). An increased risk for bladder cancer has been associated
with swimming pool attendance, however with inconclusive evidence (8).
Epidemiological studies showed an association of the attendance of chlorinated
swimming pools and adverse health effects of respiratory functions and on the
risk of developing asthma (9–11). So far no specific DBPs have been identified,
but trichloramine and other volatile DBPs are suspected of contributing to these
effects.
432
Trichloramine in Swimming Pools
Trichloramine (TCA, NCl3) is a volatile, instable disinfection by-product of

pungent odor, which causes the typical indoor pool smell. TCA is the most volatile
of all chloramines and irritates eyes and upper airways. The irritant potency of
TCA is in the same order as chlorine or formaldehyde (12). An inhalation study
for TCA with rats showed acute toxicity with a LC50 of 550 mg/m3 (13) A first
proof of the effects of TCA on humans is from an in vitro exposure study with
human alveolar epithelial lung cells from the cell line A-549 (14). Decreasing
cell viability and an increase in pre-inflammatory biomarkers (cytokines IL-6 and
IL-8) have been found at TCA concentrations above 10 mg/m3.
Therefore exposure of swimmers and pool personnel to TCA should be
minimized. The National Research and Safety Institute for occupational accidents
prevention in France (INRS) and the WHO recommended a reference value

of 0.5 mg/m3 for TCA in air (1). The reference value has been derived from
studies, where symptoms of irritation of eyes, nose and throat have been reported
by swimming instructors and lifeguards in indoor pools at TCA levels from
0.5 mg/m3 (5, 15). Recent monitoring studies from Switzerland and Germany
showed, that the occurrence of TCA in indoor pools are typically in the range
of 0.1 and 0.5 mg/m3 (16, 17) with maximum levels at worst case of more than
18 mg/m3 (18). Under consideration of a precautionary principle, the median
TCA levels in indoor pools and further results from a study in which the lowest
observed no-effect levels for TCA were found at 0.35 mg/m3 Parrat et al. proposed
an occupational exposure limit of 0.3 mg/m3 (17).
Measurement methods for TCA in air are mostly based on reactive air
sampling and have to cope with the instability of TCA. The most applied method
uses reduction of TCA to chloride on glass fiber filters impregnated with As2O3
and subsequent quantification of chloride by ion chromatography (5). Another
reactive method uses an impinger system and the colorimetric chlorine test
reagent N,N-diethyl-p-phenylenediamine (DPD) with potassium iodide (19, 20).
Both methods are not specific and prone to interferences from other halogenated
DBPs and oxidizing air constituents, which can be reduced to chloride or are able
to oxidize the DPD reagent. Another disadvantage is the rather long sampling
time of several hours necessary to obtain sufficiently low detection limits. An
alternative approach for direct TCA measurements in air would be ion mobility
spectrometry.
Ion Mobility Spectrometry
Ion mobility spectrometry (IMS) is an analytical technique for the separation

and detection of gas phase ions at atmospheric pressure and emerged in the
late 1960s (21, 22). IMS is widely applied to detect pollutants, explosives or
illicit drugs by military forces (23), airport and workplace security and it gained
increasing interest as orthogonal separation principle for mass spectrometry (24,
25). The advantages of IMS are its flexibility, speed and low detection limits
paired with its ease of maintenance, the relatively low investment costs and
the availability of handheld devices. More than 10,000 IMS based detectors
433
are therefore used in security checkpoints and more than 50,000 handheld IMS
analyzers are in use for chemical-weapons monitoring by armed forces. Further
application areas for robust and easy-to-use handheld IMS are for example in the
areas of first response and logistics used by police, rescue forces, fire fighters, or
customs and recently in food safety monitoring, clinical diagnostics, forensics
(26) and environmental monitoring (27). In this work we used IMS to monitor the
volatile and instable disinfection by-product TCA in the air of indoor swimming
pools.
The principle of operation includes soft ionization with 63Ni, UV light or
corona discharge to form reactant ion species from the carrier gas, which is
generally air, formation of analyte ions and ion clusters (atmospheric pressure
chemical ionization), electrical injection to a drift region and detection. Ions are
accelerated in an electrostatic field in the drift region against a counter-flow drift
gas and travel at characteristic speeds that are related to the mass, charge, size
and shape of the ions or ion clusters.
The general set-up of an IMS device is shown in Figure 1. The heart of the
IMS is the drift tube which includes an ionization chamber separated by the shutter
grit (21). The tube is built from a stack of metal guard rings which are separated
by thin isolators and attached to a voltage divider so that a smooth increase of
voltages can be applied over the whole string. A steady flow of drift gas (nitrogen
or air) at ambient pressure serves as collision and reactant gas and prevents the
tube from contamination.
Gaseous samples can be introduced directly into the ionization chamber
where an ionization source, e.g. a beta emitter (63Ni) produces reactant ions.
Analytes are ionized by proton-transfer and charge exchange reactions in a kind
of atmospheric chemical ionization mechanism. The type of ionization source
and drift gas is critical for ionization mechanisms and efficiencies, which was
demonstrated for monosubstituted toluenes and anilines for three ionization types
(63Ni beta emitter, corona discharge and photoionization (28). Corona discharge
sources often combine mixed ionization mechanisms and therefore lead to more
complicated response behavior than beta emitters. Photoionization can be used
for direct ionization without reactant ions and therefore lead to much simpler
response behavior and often large ranges of quantitative responses (27).
Ionized analytes are then transferred to the drift tube via an electronic shutter
grid which is periodically triggered. In the drift tube ions experience acceleration
by the electrical field and a number of collisions with the drift gas. This results in
a rather constant drift velocity vd which is proportional to the mobility of an ion
K and the electric field strength E (Equation 1). Analyte ions with different ion
mobilities K due to differences in size, mass and charge experience a separation.
Since the ion mobilities K depend on environmental conditions, generally
normalized mobilities K0 are given, where T is the absolute temperature of the
drift gas, T0 the standard temperature (273 K), P the atmospheric pressure during
the measurement and P0 the standard pressure (760 mm Hg) (Equation 2).
434
Figure 1. Ion mobility spectrometry (IMS) set-up.
Methods
Trichloramine Gas Generation
TCA gas standards were produced continuously in a gas generation unit
as described by Schmalz et al. (14). Shortly, ammonia chloride and sodium
hypochlorite were continuously dosed to a phosphate buffer solution (pH 3) in a
reaction bottle. Nitrogen gas with a flow rate of 100 mL/min was used to strip out
the formed TCA. The gas stream was cleaned in impingers with amidosulfonic
acid solutions and water. The TCA gas stream was splitted into two fractions.
The concentrated gas stream was trapped in cooled iso-octane and the other part
was diluted with synthetic air. The continuous TCA gas generation was controlled
by UV-absorption of the iso-octane solutions every 30 minutes. The diluted gas
stream with a flow rate of up to 7 L/min was adjusted to produce concentrations
in the range of 0.1 and 0.8 mg/m³. The concentration of the diluted gas stream
was determined by the impinger and reactive adsorption method as described
below (TCA analysis). The air humidity was varied by an impinger with water
and controlled with a hygrometer.
Further DBP Gas Standards

The pure substances of trihalomethanes and dichloroacetonitrile
were purchased from Sigma-Aldrich. Aqueous stock solutions of mono-
and dichloramine were synthesized according to literature (29) and
dichloromethylamine according to Cimetiere and De Laat (30). The stock
solutions were analyzed by photometry, for NH2Cl at λ = 254 nm (ε = 388 L
mol-1cm-1), for NHCl2 at λ = 282 nm (ε = 221 L mol-1cm-1) and for CH3NCl2
at λ = 302 nm (ε = 330 L mol 1cm-1).
435
For qualitative and semi quantitative IMS analysis defined amounts (0.5 to
10 µL) of pure THMs were vaporized and diluted in gas-tight bottles and Tedlar
bags filled with nitrogen. Final concentrations in the Tedlar bags were between
0.1 mg/m³ and 1.0 g/m³.
For quantitative analyses diluted aqueous solutions of THMs,
dichloroacetonitrile and synthesized chloramines between 0.01 and 10 mg/L were
prepared (DBP solutions). A total volume of 2.5 L of these DBP solutions was
filled in three stripping vessels (2 L, 0.25 L, 0.25 L) in series. The stripping gas
flow was 64 mL/min and the dilution gas flow was 1932 mL/min at a temperature
of 23 ± 1°C. The system equilibrium was checked with varied volumes of DBP
solutions. Hence the concentration of the DBP gas stream could be calculated
using the Henry constant, the initial aqueous concentrations and the adjusted
gas flows.
Trichloramine Analysis
Photometric Method
The photometric analyses of iso-octane solutions containing the absorbed

TCA were performed by a Cary 50 spectrophotometer (Varian, Germany). The
concentration of TCA was analyzed at three wavelengths (λ = 343 nm, ε = 185
Lmol-1cm-1; λ = 260 nm, ε = 399 Lmol-1cm-1; λ = 225 nm, ε = 5,470 L mol 1cm-1).
Reactive Adsorption Method
Reduction of TCA to chloride on impregnated glass fiber filters was performed

according to literature (5, 16).
Glass fiber filters (d = 50 mm, QM-A, Whatman Schleicher Schuell,
Germany) were impregnated with 0.9 mL of a solution of 40 g/L glycerol, 107
g/L Na2CO3 und 8 g/L As2O3. The flow rate through two impregnated filters in
series was 1 L/min (pump: Gilian 5000, Sensidyne, USA). The sampling time
was 3 hours. After sampling the filters were eluted with water and the chloride
concentrations were determined by ion chromatography after ion exchange with
OnGuard II Na-Cartridges (Dionex, Idstein).
Impinger Method
The reaction of TCA with the reagent N,N-diethyl-phenylenediamine (DPD)

and potassium iodide (KI) was carried out in solution in an impinge (19, 20). DPD
and KI (Chlorine test, Merck, Germany) were dissolved in 50 mL water and filled
into two impingers. The air flow rate through the two impingers in series was 1
L/min. Sampling time was 90 minutes. The red color of the oxidized DPD was
photometrically determined directly after sampling at a wavelength of 510 nm.
436
Ion Mobility Spectrometry
IMS measurements were performed by a hand-held GDA2 spectrometer

(Airsense, Germany) equipped with a 100 MBeq 63Ni-ionisation source. The
drift tube had a length of 6.3 cm and was operated at a temperature between 40
and 45°C and at a voltage between 1.7 and 1.9 kV. The gas inlet was equipped
with a silicone membrane and the total gas flow rate through the system was
regulated by internal pumps at 470 mL/min. One complete IMS-spectrum was
acquired per second, recorded by the software Winmuster GDA Version 1.2.6.0
and stored on a SD card. Generally a stable signal was obtained after about 10
to 20 seconds and the average signal height of at least 40 spectra were used for
quantitative analysis. For field measurements the GDA2 was supplied by a Li-Ion
accumulator. The negative IMS spectra were analyzed for chloride and bromide
ions at a normalized drift time (K0) of 2.7 cm2/(Vs) and 2.5 cm2/(Vs).
Investigation of Indoor Pool Air
Indoor air measurements were performed in an indoor pool with a swimmer

pool (V = 980 m³), a non-swimmer pool (V = 240 m³) and a baby pool (V = 4.7
m³). The total air volume of the hall is 4890 m³. The baby pool is completely
separated. Analyses with the reactive adsorption method (As2O3 filter) and the
DPD/KI method were done at sampling heights of 20 cm above the water surface
and at a distance of 20 cm from the edge of the pool. The analyses with the IMS
were done at different sampling locations and heights at the windows, at fresh air
inlets and outlets and above the surface of the water in the swimming pool hall and
in the engineering rooms and offices.
Results
Trichloramine (TCA) analysis was performed by a hand-held ion mobility
spectrometer (IMS) with direct intake of gas samples from indoor air
environments, gas generation units or from Tedlar bags for spiked gas standards.
A silicone membrane at the IMS inlet protects the IMS analyzer and keeps it at a
constant humidity. A typical measurement cycle consists of the continuous intake
of gaseous sample by a built-in pump and the signal readout after constant signal
intensity was obtained. Single IMS measurements can be performed within 20
seconds and the results are stored on a storage device or can be directly viewed,
if the instrument is calibrated for the analytes. The mobility of the hand-held
instrument and the short response time allow quite flexible applications and
measurements of high spatial and temporal resolution in indoor pool settings.
In contrast to that, the conventional filter method for TCA determination
requires typically three hours of sampling time, further laboratory work for
filter extraction, sample clean-up and ion chromatographic determination of the
formed chloride. Generally, the filter method allows only the determination of
437
averaged concentrations over time periods of several hours, which doesn´t allow
monitoring of the dynamic behavior of TCA concentrations in dependence of
bather activity or of the sampling location in an indoor pool hall.
IMS Signals of Volatile DBPs

The IMS spectra of different halogenated DBPs after ionization by a 63Ni
source generally show two signals, if the compounds contain bromine and
chlorine, and one signal if only chlorinated DBPs have been measured (Figure
2). Further signals can be assigned to reactant ions from ionized gas and water
species. This result shows that from different DBPs primarily only chloride
and bromide ions are produced by electron capture in a dissociative electron
attachment mechanism and therefore only chloride and bromide ions are separated
in the drift tube of the IMS. This process is already known from electron capture
detectors (ECD) and described for halo- and nitro-substituted benzenes (22).
For example halogenated benzenes and alkyl halides only yielded halide ions
by a simple dissociative electron capture, whereas nitrobenzene exhibited an
associative electron capture resulting in negatively charged molecular ions.
Halogenated nitrobenzenes showed both mechanisms associative and dissociative
electron capture.
The normalized ion mobilities for chlorinated compounds are at K0 = 2.7
cm2/(Vs) (chloride ion) and for brominated compounds at K0 = 2.5 cm2/(Vs)
(bromide ion, Figure 2). For bromodichloromethane both signals could be
detected. Bromoform has only one signal for bromide. Other ions in the mobility
spectrum are from reactant gas ions which are formed from air (oxygen) and
water (signals 3 and 4 in Figure 2). The implications of the generalized signal
response of the IMS to different halogenated compounds are that the IMS signal
is not specific for individual DBPs but provides the possibility to measure all
chlorinated DBPs and all brominated DBPs in a kind of sum parameter. The
contribution of single DBPs to the sum parameter strongly depends on their
response factor and individual concentration in the gas sample.
IMS Response and Calibration

Response factors and limits of quantitation (LOQ) have been determined by
IMS individually for some selected DBPs to check for their contribution to the
sum signal based on chloride and bromide. For the selection of DBPs the major
chlorinated compounds like three chloramines and three chlorinated THMs have
been considered, as well as some trace compounds like dichloroacetonitrile and
dichloromethylamine, which are regularly occurring in swimming pool water
(31). From recent monitoring programs there is information available on typical
indoor pool air concentrations of TCA, which are in the range of 0.1 and 0.5
mg/m3 (17, 18). Chloroform concentrations are typically in the same range
between 0.1 and 0.5 mg/m3 (32, 33). Further typical air concentrations have been
estimated from published data for indoor pool air or were calculated from data
of aqueous concentrations and the corresponding Henry´s law coefficients. The
most prominent volatile DBPs are chloroform and TCA. The air concentrations
438
of other brominated and chlorinated THMs and DBPs are by a factor of about ten
to one hundred lower.
Figure 2. IMS spectra (signal intensity vs. drift time) of a blank and volatile
DBPs (TCA, bromodichloromethane and bromoform). Signals are from chloride
(1) and bromide (2) with normalized ion mobilities of K0 = 2.7 cm2/(Vs) and K0
= 2.5 cm2/(Vs), respectively.
The response factors for the selected DBPs range over four orders of
magnitude between 3896 (arbitrary signal units m3/mg) for dichloroacetonitrile
and 0.5 (arbitrary signal units m3/mg) for chloroform (Table 1). The response
factors are mostly dependent on the ionization efficiency during dissociative
electron capture, since the detected ion is in any case chloride for chlorinated
DBPs. Therefore differences in electron capture cross sections and bond energies
of the individual compounds account for the overall response. Interestingly all
chloramines, where the halogen is bound to a nitrogen atom, show a rather high
response. However, the highest value was found for dichloroacetonitrile with
a response factor of almost 4000 (arbitrary signal units m3/mg). Due to a more
than 400fold less response, chloroform contributes to only much less than 1 %
of the chloride signal of TCA in typical indoor swimming pool environments
and can therefore be neglected. The two other THMs (bromodichloromethane,
dibromochloromethane) and dichloroaceto-nitrile also contribute to less than 1 %
to the chloride signal. Also the contribution of chlorine can be neglected due to the
low partitioning of hypochlorous acid to the gas phase and therefore the resulting
low concentrations in the air (31). Only the contribution of dichloromethylamine
is estimated to be in the range of 10 %.
439
Table 1. Response Factors, Limits of Quantitation (LOQ) and Typical Air
Concentrations for Halogenated DBPs in Indoor Swimming Pools Measured
As Chloride and Bromide in Negative Ion Mobility Spectra
Ion DBP Response LOQ Typical air
mobility factor (mg/m³) concentration
K0 (m3/mg) (mg/m3)
(cm2/(Vs))
2.7 (Cl-) NCl3 221 0.1 0.1 - 0.5
2.7 CHCl3 0.5 24 0.1 - 0.5
2.7 CHBrCl2 33 0.4 < 0.02
2.7 CHBr2Cl 21 1.0 < 0.02
2.7 NH2Cl 300 0.1 < 0.004

2.7 NHCl2 180 0.1 < 0.006
2.7 CH3NCl2 906 0.01 0.01-0.07
2.7 CHCl2CN 3896 0.003 < 0.0002
2.5 (Br-) CHBrCl2 46 0.4 < 0.02
2.5 CHBr2Cl 247 0.1 < 0.02
2.5 CHBr3 - 0.2 < 0.02
In a realistic estimate we come to the conclusion that in indoor swimming pool

environments the major contribution to the IMS signal of chlorinated compounds
is from TCA (more than 85 %) with a minor contribution of dichloromethylamine
(about 10 %). Therefore IMS measurements are capable to detect TCA - the
DBP of most concern - together with a further nitrogenous reaction product. For
other environments and air compositions the contribution of interferences has to
be reconsidered separately.
The limit of quantification (LOQ) for the IMS measurement of TCA in air
was determined to be 0.1 mg/m3. The LOQ is therefore sufficient for indoor pool
air monitoring where exposure limits and reference values between 0.3 and 0.5
mg/m3 have to be considered. LOQs of other DBPs are in most cases higher than
the typical air concentrations with the exception of dichloromethylamine.
IMS measurements of TCA in the negative ionization mode between 0.1 and
0.7 mg/m³ show linear and quite robust calibration functions. This is demonstrated
in Figure 3 for calibrations on three days (e.g. on Nov. 8, 9 and 14) for dry
air and humid air with 65 % relative humidity (at 25 °C). Humidity of the gas
sample can affect IMS measurements by formation of water-analyte clusters in
the IMS tube which causes analyte discrimination due to different migration times
of the analyte-water clusters compared to the naked analyte ions. The results show
that humidity doesn´t significantly affect IMS calibration. This may be a result of
the silicone membrane in the instrument inlet, which should guarantee a constant
humidity in the IMS instrument independent of the humidity of the gas sample.
440
Figure 3. Calibration of trichloramine for IMS signals based on chloride (K0

= 2.7 cm2/(Vs)) for 3 different days within a week in dry and humid air (65 %
relative humidity at 25 °C).
Table 2. Figures of Merit for the Linear IMS Calibration Measured at Three
Days for Dry and Humid Air (65 % Relative Humidity at 25 °C)
Day/Humidity (%) Slope Standard error of Intercept
slope
1/0 200 8.6 70
2/0 235 22.1 69
3/0 259 20.2 42
1/65 246 8.8 56
2/65 209 15.0 71
3/65 211 20.2 84
The slopes of the IMS calibration for 3 different days with dry and humid air
show no significant differences, if the two means of the slopes from dry air (mean:
231.3; STD 29.7; n = 3) and from humid air (mean: 222.0; STD 20.8; n = 3)
are subjected to a t-test (Table 2). The interday statistical error is higher than that
between dry and humid gas samples and is also caused by systematic errors of the
production of TCA gas standards. TCA is a rather instable compound which has
to be produced and diluted online to suitable concentration levels for calibration,
which is a rather tedious task. The relative standard errors of the slopes of single
calibration functions are within 4.3 % and 9.4 %, which is a quite acceptable range
for this type of measurement.
441
Temporal Dynamics of Trichloramine Concentrations in Indoor Pool Air
The quick response time of IMS measurements and the mobility of the
instrument allowed us to use it for the investigation of the temporal dynamics of
TCA concentrations at various locations of an indoor pool setting with a large
swimmer hall, which is connected to a non-swimmer pool with low ceiling, and a
separated, confined baby pool area (children area). The measurements were done
for defined time intervals of typically 10 to 15 min.
Figure 4. Temporal dynamics of trichloramine in different compartments of

an indoor pool air in dependence of the number of bathers and their activity.
Temporal separated measurement intervals are presented in grey and black.
Measurement was done at about 20 cm above water surface.
Selected results for the concentrations of TCA in air in dependence of the

number and activity of swimmers are shown in Figure 4. The first samples before
a group of swimmers came in the pool (region A in Figure 4) showed generally
very low TCA concentrations (0.06 mg/m3 in the baby pool, 0.11 mg/m3 in the
non-swimmer and 0.12 mg/m3 in the swimmer area), since no or almost no kids
or swimmers used the pools. This is considered as background levels, which
established overnight. Then a school class with about 45 children attended the
non-swimmer area, many pupils jumped into the water or were splashing around.
This caused a sudden increase of the TCA concentration to about 0.36 mg/m3 in
this area (region B in Figure 4), which can be explained by a sudden increase of
the mass transfer coefficient for TCA from water to air due to the agitation of
the pupils´ activity. It has been shown that the mass transfer coefficient can be
increased from 1.8 x 10-3 g/(h m2) in a quiescent pool to 7.0 x 10-3 g/(h m2) by
swimming activity and even to 12.6 x 10-3 g/(h m2) by increased splashing activity
or by sparging the water in a whirlpool (34). Since the mass transfer is a limiting
factor for an empty pool with smooth water surface, TCA can be built up in the
water overnight and then contribute to a sudden increase of TCA concentration in
442
the air. This example shows that the TCA concentration in indoor pools can be
highly dynamic. For the swimmer pool there is observed also an increase of TCA
concentration with increasing swimmer activity, but it is less pronounced due to
much less swimmers per surface area of the water and due to much higher air space
of the swimming pool hall (10 m) compared to the low ceiling of the non-swimmer
area (about 3.50 m height of the room).
These examples clearly show the necessity to monitor the temporal variability
of TCA concentrations for a more accurate estimation of the exposure of bathers
and personnel in a swimming pool.
Concentration Profiles of TCA in an Indoor Pool

The fast response time of IMS measurements allowed also to measure

concentration profiles of TCA in an indoor pool at different sampling locations.
The assumption was made, that the concentration regime didn´t change during the
short time period of less than 10 min for the complete measurement of the profile.
The results are again from the non-swimmer area with low ceiling and suboptimal
ventilation. The air of the ventilation system is guided on a space diagonal from
the left lower corner to the right upper corner in this non-swimmer area (Figure
5). All measurements were taken at a height of 20 cm above water level. TCA
concentrations show an increasing trend from 0.13 mg/m3 in the inflow region
of air to about 0.26 mg/m3 in the outflow region. This result is clearly a sign of
insufficient ventilation of this area. The air takes TCA up on its way through the
non-swimmer compartment which doubles the TCA concentration. This example
reveals that it is necessary to monitor TCA concentrations at various sampling
locations within one indoor area to get a more complete picture on TCA exposure
of bathers and personnel.
Figure 5. Trichloramine concentration profiles in an indoor non-swimmer pool

area with low ceiling and insufficient air ventilation (measured at 20 cm above
water surface; numbers indicate sampling locations).
443
Conclusions
Ion mobility spectrometry (IMS) revealed as a valuable tool to measure
trichloramine (TCA) concentrations in indoor swimming pool environments.
Since during the ionization in IMS chloramines and other chlorinated DBPs can
contribute to the chloride signal, interfering compounds and their contribution
to the chloride signal have to be considered and found to be less than 15 % in
typical indoor pool air. The short response time of about 20 s, the mobility and
stability of the IMS calibration enable to monitor the highly dynamic behavior in
time and space of TCA and therefore to determine more accurately the exposure
of swimmers and pool personnel.
Acknowledgments
This work was supported by the German Federal Ministry of Education and
Research (BMBF 02WT1090) in the joint project “Health-Related Optimization of
Swimming Pool Water Treatment”. We further thank J. Chrapan and V. Kuemmel
for their contribution to the generation of TCA gas standards and IMS calibration.
References
1. WHO. Guidelines for Safe Recreational Water Environments, Volume 2:
Swimming pools and similar environments; World Health Organization:
Geneva, CH, 2006.
Frimmel, F. H. Drowning in Disinfection Byproducts? Assessing Swimming
Pool Water. Environ. Sci. Technol. 2007, 41, 363–372.
3. Richardson, S. D.; DeMarini, D. M.; Kogevinas, M.; Fernandez, P.;
Marco, E.; Lourencetti, C.; Balleste, C.; Heederik, D.; Meliefste, K.;
McKague, A. B.; Marcos, R.; Font-Ribera, L.; Grimalt, J. O.; Villanueva, C.
M. What’s in the Pool? A Comprehensive Identification of Disinfection
By-products and Assessment of Mutagenicity of Chlorinated and Brominated
Swimming Pool Water. Environ. Health Perspect. 2010, 118, 1523–1530.
4. Walse, S. S.; Mitch, W. A. Nitrosamine carcinogens also swim in chlorinated
5. Hery, M.; Hecht, G.; Gerber, J. M.; Gender, J. C.; Hubert, G.; Rebuffaud, J.
Exposure to Chloramines in the Atmosphere of Indoor Swimming Pools.
Ann. Occup. Hyg. 1995, 39, 427–439.
7. Villanueva, C. M.; Font-Ribera, L. Health impact of disinfection by-products
in swimming pools. Ann. Ist. Super Sanità 2012, 48, 387–396.
Cancer and Exposure to Water Disinfection By-Products through Ingestion,
444
Bathing, Showering, and Swimming in Pools. Am. J. Epidemiol. 2007, 165,
148–156.
9. Nickmilder, M.; Bernard, A. Ecological association between childhood
asthma and availability of indoor chlorinated swimming pools in Europe.
Occup. Environ. Med. 2007, 64, 37–46.
10. Bernard, A.; Carbonnelle, S.; de Burbure, C.; Michel, O.; Nickmilder, M.
Chlorinated Pool Attendance, Atopy, and the Risk of Asthma during
Childhood. Environ. Health Perspect. 2006, 114, 1567–1573.
11. Lévesque, B.; Duchesne, J.-F.; Gingras, S.; Lavoie, R.; Prud’Homme, D.;
Bernard, E.; Boulet, L.-P.; Ernst, P. The determinants of prevalence of health
complaints among young competitive swimmers. Int. Arch. Occup. Environ.
Health 2006, 80, 32–39.
12. Gagnaire, F.; Azim, S.; Bonnet, P.; Hecht, G.; Hery, M. Comparison of the
sensory irritation response in mice to chlorine and nitrogen trichloride. J.

Appl. Toxicol. 1994, 14, 405–409.
13. Barbee, S. J.; Thackara, J. W.; Rinehart, W. E. Acute inhalation toxicology
of nitrogen trichloride. Am. Ind. Hyg. Assoc. J. 1983, 19 (44), 145–146.
14. Schmalz, C.; Wunderlich, H. G.; Heinze, R.; Frimmel, F. H.; Zwiener, C.
Application of an optimized system for the well-defined exposure of human
lung cells to trichloramine and indoor pool air. J. Water Health 2011, 9,
586–596.
15. Massin, N.; Bohadana, A. B.; Wild, P.; Héry, M.; Toamain, J. P.; Hubert, G.
Respiratory symptoms and bronchial responsiveness in lifeguards exposed to
nitrogen trichloride in indoor swimming pools. Occup. Environ. Med. 1998,
55, 258–263.
16. Schmoll, B.; Kellner, R.; Breuer, D.; Buxtrup, M.; Engel, C.; Fliedner, G.;
Franke, U.; Friedrich, C.; Geilenkirchen, A.; van Gelder, R.; Neumann, H.-
D.; Radtke, R.; Richter, D.; Salvadori, U.; Spreckelsen, F.; Stöcker, S.;
Thullner, I.; Weber, B.; Wegscheider, W.; Wimmer, B.; Zirbs, R.
Trichloramin in der Schwimmhallenluft. Archiv Badewes. 2009, 10,
591–611.
17. Parrat, J.; Donze, G.; Iseli, C.; Perret, D.; Tomicic, C.; Schenk, O.
Assessment of Occupational and Public Exposure to Trichloramine in Swiss
Indoor Swimming Pools: A Proposal for an Occupational Exposure Limit.
Ann. Occup. Hyg. 2012, 56, 264–277.
18. Stottmeister, E.; Voigt, K. Trichloramine prevention remains better than cure.
Recreation 2006, 3, 30–33.
19. Weng, S. C.; Weaver, W. A.; Zare Afifi, M.; Blatchley, T. N.; Cramer, J. S.;
Chen, J.; Blatchley, E. R. Dynamics of gas-phase trichloramine (NCl3) in
chlorinated, indoor swimming pool facilities. Indoor Air 2011, 21, 391–399.
20. Predieri, G.; Giacobazzi, P. Determination of nitrogen trichloride (NCl3)
levels in the air of indoor chlorinated swimming pools: an impinger method
proposal. Intern. J. Environ. Anal. Chem. 2012, 92, 645–654.
21. Hill, H. H.; Siems, W. F.; St. Louis, R. H. Ion mobility spectrometry. Anal.
Chem. 1990, 62, 1201A–1209A.
22. Karasek, F. W. Plasma Chromatography. Anal. Chem. 1974, 46,
710A–720A.
445
23. Eiceman, G. A.; Stone, J. A. Peer Reviewed: Ion Mobility Spectrometers in
National Defense. Anal. Chem. 2004, 76, 390A–397A.
24. Cottingham, K. Product Review: Ion mobility spectrometry rediscovered.
Anal. Chem. 2003, 75, 435A–439A.
25. Liuni, P.; Romanov, V.; Binette, M.-J.; Zaknoun, H.; Tam, M.; Pilon, P.;
Hendrikse, J.; Wilson, D. J. Unambiguous Characterization of Analytical
Markers in Complex, Seized Opiate Samples Using an Enhanced Ion
Mobility Trace Detector-Mass Spectrometer. Anal. Chem. 2014, 86,
10772–10779.
26. Armenta, S.; Alcala, M.; Blanco, M. A review of recent, unconventional
applications of ion mobility spectrometry (IMS). Anal. Chim. Acta 2011,
703, 114–123.
27. Márquez-Sillero, I.; Aguilera-Herrador, E.; Cárdenas, S.; Valcárcel, M. Ion-
mobility spectrometry for environmental analysis. Trends Anal. Chem. 2011,

30, 677–690.
28. Borsdorf, H.; Neitsch, K.; Eiceman, G. A.; Stone, J. A. A comparison of the
ion chemistry for mono-substituted toluenes and anilines by three methods
of atmospheric pressure ionization with ion mobility spectrometry. Talanta
2009, 78, 1464–1475.
29. Li, J.; Blatchley, E. R., III. UV Photodegradation of Inorganic Chloramines.
30. Cimetiere, N.; De Laat, J. Henry’s law constant of N,N-
dichloromethylamine: Application to the contamination of the atmosphere
of indoor swimming pools. Chemosphere 2009, 77, 465–470.
31. Weaver, W. A.; Li, J.; Wen, Y.; Johnston, J.; Blatchley, M. R.; Blatchley, E.
R., III Volatile disinfection by-product analysis from chlorinated indoor
32. Aggazzotti, G.; Fantuzzi, G.; Righi, E.; Predieri, G. Environmental
and biological monitoring of chloroform in indoor swimming pools. J.
Chromatogr. A 1995, 710, 181–190.
33. Lévesque, B.; Ayotte, P.; Tardif, R.; Charest-Tardif, G.; Dewailly, E.;
Prud’Homme, D.; Gingras, G.; Allaire, S. Evaluation of the Health Risk
Associated with Exposure to Chloroform in Indoor Swimming Pools. J.
Toxicol. Environ. Health, Part A 2000, 61, 225–243.
– Formation and mass transfer. Water Res. 2011, 45, 2681–2690.
446
Subject Index
A C
Advanced treatment (O3–BAC) Chloral hydrate control

DBPs, 324 CH degradation, effect of residual
DOC removal and C-DBP FP removal, chlorine, 374f
330f CH removal
DON removal and N-DBP, HNMs, and boiling of spiked waters, 372f
HAcAms FP removal, 331f domestic adsorptive cartridges, 370f
effect of terminating pre-ozonation, 328 domestic RO cartridge, 368f
fluorescence EEM spectra of treated effects of stirring speed and
Lake Taihu water, 333f ultrasonication, 376f
Publication Date (Web): August 24, 2015 | doi: 10.1021/bk-2015-1190.ix002
MW distribution, 332f RO and GAC alone or in combination,

parallel O3-BAC pilot plant process 371f
flows, 329f CH treatment
precursors, 322 adsorptive materials, 369
removal of DOC, DON, C-DBPs and boiler and microwave oven, 371
N-DBPs, 329f domestic microwave oven, 373f
removal of DON by BAC process, 324f reverse osmosis, 367
removal of TCNMFP by the BAC stirring and ultrasonication, 374
process, 325f effect of operating pressure on RO
removal rates of FPs, 325f flowrate, 367t
removal rates of water quality materials and methods
parameters, 323f analytical methods, 365
representative fluorescence EEM spectra apparatus, 366
of DOM in raw water, 327f characteristics of waters, 366t
total organic halogen (TOX), 328 samples and chemicals, 365
UV/vis optical spectrums of water microwave heating process, CH and
samples, 326f temperature changes, 375f
Amino acids, 215 point-of-use and household appliances,
analytical methods and minimum 363
reporting levels, 220t Comparative haloacetaldehyde toxicity, 35
carbonaceous-DBPs Control of halogenated N-DBP precursors.
DBP FPs of AAs tested, 229t See Advanced treatment (O3–BAC)
DCAA and TCAA formation potential advanced treatment (O3–BAC)
results, 228f advanced treatment process (O3–BAC),
haloacetic acids (HAAs), 227 311
trihalomethanes (THMs), 226 coagulation-IPS-filtration versus
formation of DBPs, effect of pH, 231 coagulation-DAF-filtration, 311
materials and methods conventional process, 311
amino acids, 217 conventional treatment (coagulation-
analytical methods, 221 IPS-filtration versus coagulation-
formation potential tests, 219 DAF-filtration)
selected amino acids, 218t DBP FPs, 316
nitrogenous-DBPs DBPs, 315
DCAN formation potential results, disinfection process, 312
223f pilot-plant process flow, pretreatment,
haloacetonitriles (HANs), 223 310
halonitromethanes (HNMs), 221 pilotscale treatment process, flowchart,
nitrosamine FPs of AAs tested, 225t 310f
nitrosamines, 224 pretreatment processes
TCNM formation potential results, DON removal in different treatment
222f trains, 313t
453
fluorescence EEM spectra of Lake dimethenamid bromination,
Taihu water, 314f pseudo-first-order rate constants,
water parameters, 312 255f
using traditional and advanced drinking excess bromide, effects on
water treatment processes, 307 pseudo-first-order rate constants,
Controlling NDMA formation, role of 258f
pre-oxidation, 151 free bromine, chemistry, 253
NDMA formation, role of free bromine speciation, more complete
oxidants/disinfectants view, 262
chlorine, 152 free chlorine concentration, influence on
chlorine dioxide, 154 pseudo-first-order rate constants, 260f
ozone, 154 model water containing bromide,
permanganate, 155 speciation of free bromine, 263f
UV irradiation, 155 natural and anthropogenic sources of
NDMA formation and effect of bromide, 252
pre-oxidation, 154f overall bromination rate, contributions

NDMA precursors, structures, 153t of brominating agents, 264f
Conventional treatment (coagulation- reaction order in total free bromine
IPS-filtration versus coagulation-DAF- concentration, 262f
filtration) Disinfection by-products (DBPs), 3
C-DBPs, 320 Disinfection by-products in swimming
coagulation processes pool water
chloroform concentration in control, 384
chlorinated water, 321f bench-scale GAC filters, schematic
N-DBP concentrations in chlorinated setup, 397f
water, 322f factors, 396
different SUVA values, 320t filtration, 396
DOC, DON and UV254 in raw water, formation, 384
319t anthropogenic input samples, DBPFPs
N-DBPs, 321 and DBP yields, 391t
precursors, 318 concentrations of individual species
and ratio of DCAA/TCAA, 390t
DBP levels in swimming pool, 387
DBP precursors in swimming pool,
D 389
indoor swimming pool, THM and
DBP formation models, 64 HAA levels, 388f
DBP-precursor bromination, catalysis interpretation of DBP profile in
aromatic compound bromination, swimming pools, 391
proposed mechanism, 255s measured HAAs and numbers of pool
Br2O as putative brominating agent, 261 users, correlation, 389f
bromination of anisole and swimming pool under investigation,
dimethenamid, catalysis 386
as a function of chloride concentration, year-around user statistics of
256f swimming pool, 387t
as a function of excess bromide HAA removal from swimming pool
concentration, 259f water
as a function of initial free chlorine bench-scale BAC filter, 398f
concentration, 261f bench-scale GAC filter, 398f
catalysis by bromide, 257 list of investigated disinfection
catalysis by chloride, 254 by-products, 386t
catalysis by hypochlorous acid, 258 materials and methods
different sources, bromide ammonia nitrogen (NH3-N), analysis,
concentrations, 253t 385
chlorine, analysis, 385
454
DBP formation potential (DBPFP) unburned, and burned detritus material,
test, 385 fluroresence EEM of water extracts,
dissolved organic carbon (DOC), 300f
analysis, 385 unburned and burned detritus, leachate
HAAs, analysis, 385 extracted, 301f
reagents, 384 specific formation potential, 303f
THMs and HANs, analysis, 385
UV absorbance, analysis, 385
modeling, 384
calibration and validation results, 394f
E
model development, 392
model implications, 395 Effect of fire on carbon quality
model performance, 394 fluorescence emission-excitation matrix,
predicted DBP profiles as function of 282f
water age, 395f fluorescence spectrometry, 278
litters before and after burns
Dissolved organic matter and disinfection

by-product precursors, forest fire effect, optical properties, 279t
271 weight and chemical characteristics,
before and after prescribed burn, yields 278t
of detritus, WEOC, WETN, and Pearson’s correlation coefficients
DBPs, 304t litters after burn, 281t
before and after prescribed fire and litters before burn, 280t
sunlight treatments, EEM regions of UV-VIS spectrometry, 277
water extracts, 302f
DOM and DBP precursors,
biogeochemical processes, 301 F
forested watersheds, 293
implications, 303 Factors influencing effectiveness of
materials and methods pre-oxidants
dark and light incubations, 297 background ions, 165
disinfection by-products formation, background organics, 165
275 pH, 164
DOM and DBP precursors, temperature, 164
characterization, 296 Formation of DBPs, 189
DOM from litter, characterization, DBP classes identified, examples, 193t
274 DBPs generated by water pollutants,
litter collection and black carbon examples, 191t
preparation, 273 haloamides, 200
original and burned litters, chemical formation mechanisms, 201f
characteristics, 274 halobenzoquinones, 203
prescribed burn in managed forest, formation mechanisms, 204f
295 halonitromethanes, 200
statistical analyses, 275 formation mechanisms, 202f
pyrogenic products, categorization, 285t halopyrroles, 202
results and discussion formation mechanisms, 203f
carbon and nitrogen quantity, effect of iodo-DBPs, 193
prescribed fire, 276 iodo-acids, 195
controlled field burns, 297 iodo-amides, 195
detritus materials, chemical iodo-DBP formation mechanisms, 196
composition, 298 iodo-THMs, 194
forest detritus samples before and mechanism of iodo-DBP formation
after burn, 300t with chlorine and chloramines, 196f
litter materials, chemical proposed reaction of iopamidol with
characteristics, 275 chlorine and monochloramine, 197f
yields of DOM and DBP precursors, NDMA and other nitrosamines,
299 formation mechanisms, 199f
455
Formation of disinfection by-products, (UPLC/)ESI-tqMS analyses, 49
effect of prescribed burning, 282 water sampling and characterization,
before and after burn, disinfection 47
by-product yields, 284f results and discussion
dissolved organic carbon extracted from concentrations, effect of boiling, 55f
litter, chlorine reactivity, 283f ESI-tqMS PIS spectra of m/z 35, 54f
Formation of disinfection by-products from ESI-tqMS PIS spectra of m/z 79, 53f
bacterial disinfection, 235 polar brominated and chlorinated
alternative disinfectants, 246 DBPs, decomposition, 51
bacterial inactivation and DBP real tap water, detoxification by
formation, effects of pH and dosage, boiling, 56
243 TOX during boiling, reduction, 51
DBP formation during bacteria
inactivation, 240
effects of humic acid and pH on E. coli
I
cell density, 243f

effects of NOM on bacterial inactivation
and DBP formation, 241 Indoor U.S. swimming pools, disinfection
experimental by-products, 405
bacterial cultures, 238 analytical methods and minimum
chemicals, 237 reporting levels, 409t
chlorine inactivation, 239 carbonaceous-DBPs occurrence, 414t
DBP analysis, 240 DBPs formation as function of time, 425
fluorescence spectroscopy, 239 effects of swimming pool operational
fluorescence EEM of water solution after conditions on DBPs, 422f
chlorination, 244f effect of bather load, 423
fluorescence EEM of water solution effect of bromide, 423
containing E. coli, 242f effect of FAC, 421
microbial cells and cellular components effect of pH, 423
as DBPs precursor, 245 effect of temperature, 421
natural organic matter versus bacterial formation and speciation of THMs and
organic matter, 245 HAAs, effect of bromide, 424f
specific DBP formation of halogenation materials and methods
of different bacteria, 241f analytical methods, 408
DBP FP and kinetics tests, 408
samples collection, 407
synthetic swimming pool waters, 407
H nitrogenous-DBPs occurrence, 416t
occurrence, 410
Haloacetaldehydes, 25 box and whisker plots, 412f
determination in water, 27 selected pools
analyte detection, 31 DBPs measured, 413
analytical methodologies, 29t monitored for 9-months,
derivatization reaction of PFBHA, 30s characteristics, 418t
sample collection, 28 water characteristics, 410, 411t
sample extraction, 28
occurrence in water, 31
disinfected water, 32t
Halogenated DBPs, 45
M
experimental methods
chemicals and seawater, 48 Modeling NDMA formation kinetics
developmental toxicity bioassay with during chloramination, 79
P. dumerilii, 50 model description
tap water samples with/without chloramine decomposition kinetics
boiling, pretreatment, 48 and associated rate constants, 84t
TOX measurement, 49 chloramines, decomposition reactions
in water, 83
456
monochloramine degradation, 84 changes of differential absorbance,
NDMA formation model, 82 modelling kinetics, 66
NDMA formation from model correlations between concentration
compounds, 87 tribromoacetic acid (TBAA) and
optimized rate constant kapp, 90t differential absorbance, 70f, 71f
NDMA formation from model precursor trichloroacetic acid (TCAA) and
compound data, 88f differential absorbance, 70f, 71f
NDMA formation in NOM, 85 correlations between proportionality
rate constant and monochloramine coefficients, 73f
decomposition rate constants, 86t DBP formation in chlorinated water,
NDMA formation of amine precursors kinetics, 66
in river water, 92f differential absorbance of chlorinated
NDMA formation of pharmaceutical LK water, kinetic profiles, 67f
compounds, 90 experimental, 65
optimized rate constant kapp, 91t modelling DBPs formation and
NDMA formation pathways, 81f speciation via differential absorbance,

Monitor trichloramine in indoor pool air 68
blank and volatile DBPs, IMS spectra, NDMA control, use of UV, 165
439f NDMA formation in chloraminated
calibration of trichloramine for IMS systems
signals, 441f operational parameters, influence, 142
disinfection by-products, 432 chloramination practice, 143
further DBP gas standards, 435 coagulation and coagulation aids, 144
halogenated dbps in indoor swimming Nitrosamine
pools, response factors, limits formation mechanisms, 198
of quantitation and typical air precursors and wastewater indicators
concentrations, 440t Nitrosamine precursor removal, 173
ion mobility spectrometry, 431, 433 activated carbon adsorption, 174
ion mobility spectrometry (IMS), set-up, activated carbons, properties, 175t
435f Chemaxon modeled speciation of
linear IMS calibration measured, 441t chlorpheniramine, 181f
results, 437 NA precursor removal with PAC in
IMS response and calibration, 438 natural waters, pH effect mechanism,
volatile DBPs, IMS signals, 438 182
TCA, concentration profiles, 443 pH effect, model NDMA precursors
trichloramine analysis removal
impinger method, 436 PAC adsorption, 179
investigation of indoor pool air, 437 PRAM approach, 180
ion mobility spectrometry, 437 pH effect, NDMA FP removal, 175
photometric method, 436 removal of NDMA FP and bulk
reactive adsorption method, 436 organic matter, 177f
trichloramine concentrations, temporal tests with aquaculture-impacted lake
dynamics, 442 water, 178
trichloramine gas generation, 435 tests with blended wastewater, 176
trichloramine in swimming pools, 433 water samples, basic water quality,
176t
N-Nitrosodimethylamine (NDMA),
national occurrence, 135
N coagulant aids, NDMA formation
potentials, 145t
Natural organic matter (NOM), methods
bromination and chlorination data analysis, 139
background, 63 sample sites, 136
bromide concentration, effects on sampling and NDMA analysis, 138
proportionality coefficients, 72f raw water chemistry, 146
457
results and discussion, occurrence data, drinking water toxicity, forcing agents, 6
overview, 140 human biomarkers, identification, 7
MonoHAA-induced transcriptome
profile pathways, 14t
nationwide drinking water in vitro
P toxicity survey, 5
new pathway, 5, 17f
Precursors and wastewater indicators selected waters, DBP exposure, 15
analytical methods
hydraulic flow modeling, 123
nitrosamine precursors, 122
nitrosamines, 122 T
PPCPs, 122
sucralose, 123 Toxicity, 34
map of Sacramento –San Joaquin Delta, assays
120f Chinese hamster ovary cells, 35
NDMA precursors CHO cell chronic cytotoxicity, 35

Sacramento river, 127f single cell gel electrophoresis, 35
San Joaquin river, 127f mammalian cell cytotoxicity and
percentage of river flow, genotoxicity, 36t
wastewater-impacted, 128 index values, 37t
Sacramento and San Joaquin rivers, Trihalomethane formation models, 97
impact of WWTP effluents, 126 application to bromide intrusion, 110
sampling locations, 121t log (Base 10), goodness of fit statistics,
Stockton WWTP 108t
N-nitrosamine FP, impact of tertiary log10(THM4) measured versus
treatment, 124f log10(THM4) predicted, 104f, 112f
N-nitrosamines and their precursors, methods, 99
123 model 21 and model 9, predictive
study overview, 121 accuracy, 109t
study sites predictive capability, 102
Sacramento Regional WWTP, 122 summary, 100t
Stockton WWTP, 121
wastewater indicators, relationship to
NDMA FP, 129f
wastewater indicators and estimated
U
wastewater effluent, 125
wastewater indicators and NDMA FP, Use of pre-oxidants for NDMA control,
concentrations, 128t 155
WWTP effluent, estimated percentage chlorine, 156
chlorine dioxide, 158
other oxidants/disinfectants
hydrogen peroxide, 163
R potassium ferrate, 163
potassium permanganate, 163
Resolve adverse health effects of DBPs, 3 UV irradiation, 162
action of DBP forcing agents, molecular ozone, 160
mechanisms, 6 secondary amines, 157
CHO cell chronic cytotoxicity analyses tertiary amines, 157
carbon based DBPs, 8f
N-DBPs, 9f
CHO cell cytotoxicity and genotoxicity
and abbreviations, 11t
V
CHO cell genotoxicity analyses, 10f
DBP exposure, toxicity, biomarkers, Variability of non-regulated disinfection
and adverse health outcomes, by-products in distribution systems, 341
epidemiological studies, 15 DBPs under study, 346t
458
four distribution systems non-regulated DBP levels within
description, 344t distribution systems, spatial
finished water at WTP, average variability, 349
characteristics, 347t regulated and non-regulated DBPs
FRC concentrations, spatial within systems, relationship, 358
distribution, 350t spatial distribution, temporal variations
individual HK levels, spatial CNCl levels, 357f
distribution, 351f CP levels, 356f
material and methods HA7 levels, 353f
analytical procedure, 345 HAN4 levels, 355f
case under study, 343 HK2 levels, 354f
data analysis, 345 Spearman rank correlation coefficients,
water sampling, 343 359t
mean DBP levels (μg/L) found in four
systems during study period, 348t
results and discussion

DBP occurrence in area under study,
W
347
impact of water flowing through WWTP effluent, estimated percentage
storage tank, temporal variability, hydraulic flow modeling, 130f
352 wastewater indicators, 131f
non-regulated DBP levels, control and
monitoring, 359
459

Recent Advances in Disinfection By-Products

Uploaded by

Copyright:

Available Formats

Recent Advances in Disinfection By-Products

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Recent Advances in Disinfection By-Products

Uploaded by

Copyright:

Available Formats

Recent Advances in

Publication Date (Web): August 24, 2015 | doi: 10.1021/bk-2015-1190.fw001

In Recent Advances in Disinfection By-Products; Xie, et al.;

In Recent Advances in Disinfection By-Products; Xie, et al.;

Tanju Karanfil, Editor

Bill Mitch, Editor

Paul Westerhoff, Editor

Yuefeng Xie, Editor

American Chemical Society, Washington, DC

Distributed in print by Oxford University Press

In Recent Advances in Disinfection By-Products; Xie, et al.;