Transformation of the antibacterial agent sulfamethoxazole in reactions with chlorine: Kinetics, mechanisms, and pathways

Article abstract of DOI:10.1021/es035225z

Sulfamethoxazole (SMX) - a member of the sulfonamide antibacterial class - has beenfrequently detected in municipal wastewater and surface water bodies in recent years. Kinetics, mechanisms, and products of SMX in reactions with free chlorine (HOCl/OCl-) were studied in detail to evaluate the effect of chlorination processes on the fate of sulfonamides in municipal wastewaters and affected drinking waters. Direct reactions of free available chlorine (FAC) with SMX were quite rapid. A half-life of 23 s was measured under pseudo-first-order conditions ([FAC]₀ = 20 μM (1.4 mg/L) and [SMX]₀ = 2 μM) at pH 7 and 25°C in buffered reagent water. In contrast, a half-life of 38 h was determined for reactions with combined chlorine (NH₂Cl, NHCl₂) under similar conditions. Free chlorine reaction rates were first-order in both substrate and oxidant, with specific second-order rate constants of 1.1 × 10³ and 2.4 × 10³ M ^-1 s^-1 for SMX neutral and anionic species, respectively. Investigations with substructure model compounds and identification of reaction products verified that chlorine directly attacks the SMX aniline-nitrogen, resulting in (i) halogenation of the SMX aniline moiety to yield a ring-chlorinated product at sub-stoichiometric FAC concentrations (i.e., [FAC]⁰:[SMX]₀ ≤ 1) or (ii) rupture of the SMX sulfonamide moiety in the presence of stoichiometric excess of FAC to yield 3-amino-5-methylisoxazole, SO₄^2- (via SO₂), and N-chloro-p-benzoquinoneimine. Reaction ii represents an unexpected aromatic amine chlorination mechanism that has not previously been evaluated in great detail. Experiments conducted in wastewater and drinking water matrixes appeared to validate measured reaction kinetics for SMX, indicating that SMX and likely other sulfonamide antibacterials should generally undergo substantial transformation during disinfection of such waters with free chlorine residuals.

Full text of DOI:10.1021/es035225z

Environ. Sci. Technol. 2004, 38, 5607-5615
diseases and infections and in feed additives to promote
Transformation of the Antibacterial
Agent Sulfamethoxazole in
growth rate and weight gain of food animals (1), and continue
to be prescribed in combination with the antibacterial agent
trimethoprim as a major first-line clinical option for treatment
of various bacterial infections (2). Research in recent years
has verified the ubiquity of numerous antibacterial com-
pounds in the aquatic environment (3, 4). Among the target
analytes measured, various sulfonamide antibacterialss
sulfamethoxazole in particularshave been repeatedly de-
tected at concentrations of 70-150 ng/ L in surface waters
and 200-2000 ng/ L in secondary wastewater effluents (3-
7).
Reactions with Chlorine: Kinetics,
Mechanisms, and Pathways
†
M I C H A E L C . D O D D A N D
C H I N G - H U A H U A N G *
School of Civil and Environmental Engineering,
Georgia Institute of Technology, Atlanta, Georgia 30332
The presence of antibacterials in the aquatic environment
is of particular concern because of fears that they may
stimulate dissemination of antibacterial resistance among
native bacterial populations. Though a causal link between
chronic exposure to sub-therapeutic concentrations of
antibacterials and induction of antibacterial resistance has
yet to be proven, singly and multiply antibacterial resistant
bacteria have been detected in municipal wastewater ef-
fluents, sewage-affected surface water systems, and even
drinking water (8-11). The detection in aquatic environ-
mental systems of resistant bacteria and the antibacterial
agents to which they are resistant clearly merits a great deal
of concern regarding the fate of antibacterial compounds in
relevant water treatment processes.
Aerobic biodegradation of sulfonamide antibacterials is
understandably quite limited (12, 13), which provides one
important indication as to why these compounds might
persist within municipal wastewater effluents and surface
water bodies. However, reported results of two more recent
studies indicate that sulfonamides are quite susceptible to
transformation via various chemical oxidation processes
involving application of free chlorine or ozone (though no
transformation products were identified) (14, 15). Observa-
tions from a recent occurrence study appear to support the
relevance of such transformation at field conditions, as
significant losses of sulfonamides were noted during chlorine
disinfection processes in several municipal wastewater
treatment plants located within the United States (7, 16).
Reactions related to sulfonamide chlorination have
previously been investigated only to a limited extent (17).
This prior research was intended to provide an explanation
for adverse toxicological reactions associated with human
use of such antibacterials and did not address reaction
kinetics or investigate reaction mechanisms under conditions
likely to be encountered in municipal wastewater or drinking
water disinfection. The current investigation was undertaken
with the intent of not only quantifying kinetics for reactions
of the sulfonamide antibacterial sulfamethoxazole (referred
to as SMX subsequently) with free available chlorine (FAC)
and combined chlorine (CC) but also identifying associated
transformation mechanisms and degradation products rel-
evant to chlorine-based municipal wastewater and drinking
water disinfection processes.
Sulfamethoxazole (SMX)sa member of the sulfonamide
antibacterial classshas been frequently detected in municipal
wastewater and surface water bodies in recent years.
Kinetics, mechanisms, and products of SMX in reactions
with free chlorine (HOCl/OCl-) were studied in detail to
evaluate the effect of chlorination processes on the fate
of sulfonamides in municipal wastewaters and affected
drinking waters. Direct reactions of free available chlorine
(FAC) with SMX were quite rapid. A half-life of 23 s was
measured under pseudo-first-order conditions ([FAC]0 ) 20
µM (1.4 mg/L) and [SMX]0 ) 2 µM) at pH 7 and 25 °C in
buffered reagent water. In contrast, a half-life of 38 h was
determined for reactions with combined chlorine (NH Cl,
2
NHCl2) under similar conditions. Free chlorine reaction rates
were first-order in both substrate and oxidant, with
specific second-order rate constants of 1.1 × 103 and 2.4
× 103 M-1 s-1 for SMX neutral and anionic species,
respectively. Investigations with substructure model
compounds and identification of reaction products verified
that chlorine directly attacks the SMX aniline-nitrogen,
resulting in (i) halogenation of the SMX aniline moiety to
yield a ring-chlorinated product at sub-stoichiometric FAC
concentrations (i.e., [FAC]0:[SMX]0 e 1) or (ii) rupture of
the SMX sulfonamide moiety in the presence of stoichiometric
excess of FAC to yield 3-amino-5-methylisoxazole, SO 2-
4
(via SO ), and N-chloro-p-benzoquinoneimine. Reaction ii
2
represents an unexpected aromatic amine chlorination
mechanism that has not previously been evaluated in great
detail. Experiments conducted in wastewater and drinking
water matrixes appeared to validate measured reaction
kinetics for SMX, indicating that SMX and likely other
sulfonamide antibacterials should generally undergo
substantial transformation during disinfection of such waters
with free chlorine residuals.
Introduction
As illustrated by SMX (Table 1), sulfonamide compounds
contain two moieties connected to both sides of the
characteristic sulfonamide linkage (-NH-S(O2)-). Among
sulfonamide antibacterials, the aniline moiety in para-
connection to the sulfonyl S is present in all, and variations
in structures typically stem from the other moiety connected
to the sulfonamide N (1). Sulfonamides exhibit wo acid
dissociation constants; one involving protonation of the
aniline N, and the other corresponding to deprotonation of
the sulfonamide NH (18, 19), as shown in Figure 1. A classical
second-order kinetics model (20-22) incorporating specia-
Sulfamethoxazole (structure shown in Table 1) is an im-
portant member of the synthetic sulfonamide antibacterials
or sulfa drugs. Sulfonamides have been used extensively for
many years in human and veterinary medicines to treat
* Corresponding author phone: (404)894-7694; fax: (404)894-8266;
e-mail: ching-hua.huang@ce.gatech.edu.
† Present address: Swiss Federal Institute for Environmental
Science and Technology (EAWAG), Ueberlandstrasse 133, CH-8600
Duebendorf, Switzerland.
9
10.1021/es035225z CCC: $27.50
Published on Web 09/30/2004
2004 Am erican Chem ical Society
VOL. 38, NO. 21, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 5 6 0 7

TABLE 1. Sulfamethoxazole and Associated Substructure Model Compounds
FIGURE 1. Speciation of sulfamethoxazole.
tion of both substrate and free chlorine oxidant was adapted
TABLE 2. Wastewater and Drinking Water Sample
to critically evaluate the effect of pH on SMX/ FAC reaction
rate. Sub-structural model compounds (Table 1) that cor-
respond to either the hypothesized reactive or non-reactive
portions of SMX were used to probe the reactive functional
groups. LC/ MS, GC/ MS, and 1H NMR were utilized in
conjunction with supplementary analytical techniques to
identify reaction products. All of the results were then
combined and analyzed to deduce the reaction mechanisms
and pathways related to transformation of SMX by FAC.
Additional experiments were conducted in real municipal
wastewater and drinking water matrixes to assess field
applicability of observations obtained in the laboratory from
clean water reaction systems.
Characteristics
water quality
data
alkalinity (mg/L,
NH (mg
DOC
3
pH
as CaCO )
NH -N/L) (mg C/L)
3
3
wastewater
drinking water
7.3
6.6
120
13
<0.12a
<0.12a
14.0
1.3
a
Minim um detectable concentration was ∼120 µg NH3-N/L (∼7 ×
10-6 m ol/L).
experiments and ∼1 g/ L FAC stocks for use in wastewater
kinetic experiments and product identification studies. FAC
stock solutions were standardized iodometrically (24). Pre-
formed chloramine stocks were prepared according to Chapin
(25) at pH 4.5, 5, 6, 7, 8, and 9 in 0.1 M acetate (pH e 5), 0.1
M phosphate (6 e pH e 8), and 0.025 M borate (pH 9) buffers
by combining buffered 302 mg/ L solutions of NH4Cl and
buffered 200 mg/ L solutions of FAC at 25 °C in 1:1 proportion
(i.e., at 2:1 [NH4Cl]:[FAC] molar ratio) under completely mixed
conditions. The resulting solutions contained ∼100 mg/ L of
CC. The CC stocks were prepared prior to each experiment,
temporarily stored at <5 °C, and used within 24 h of
generation. Inorganic chloramine concentrations were stan-
dardized using DPD-FAS titrimetry (24).
Materials and Methods
Standards and Reagents. Sulfamethoxazole (SMX) was
obtained from ICN Biomedicals (Irvine, CA). 3,5-Dimeth-
ylisoxazole (DMI), 3-amino-5-methylisoxazole (AMI), 4-ami-
nophenyl methyl sulfone (APMS), and 4-aminophenol were
purchased from Sigma-Aldrich (St. Louis, MO). 4-Methyl-
N-(5-methylisoxazol-3-yl)benzenesulfonamide (MMIB) was
purchased from ASINEX (Moscow, Russia). All commercially
available chemical standards were of at least 98% purity and
were used without further purification. N-Chloro-p-benzo-
quinoneimine (NCBQ) was synthesized from 4-aminophenol
according to the procedure reported by Venuvanalingam et
al. (23). NaOCl was obtained from Fisher Scientific (Pitts-
burgh, PA) at ∼7% purity. All other reagents used (e.g., buffers,
colorimetric agents, reductants, etc.) were obtained from
Fisher Scientific (Pittsburgh, PA) and were of reagent grade
quality or better. All reagent solutions were prepared using
Nanopure purified water (Barnstead, Dubuque, IA).
One hundred milligrams per liter substrate stock solutions
(for use in kinetic studies) were prepared in water, with 10%
methanol as a cosolvent. One gram per liter SMX stock
solutions (for use in LC/ MS product characterization studies)
were prepared using 50% methanol. Stock solutions were
replaced after 2 months of storage at <5 °C. Aqueous sodium
hypochlorite stock solution was diluted to yield ∼100 mg/ L
FAC stocks for use in clean water and drinking water kinetic
Municipal Wastewater and Drinking Water Sam ples.
Reactions of SMX with FAC were studied in wastewater and
drinking water process samples to gauge potential effects of
matrix on the reactions. The procurement and sources of
these water samples are described in the Supporting Infor-
mation (Text S1). Important characteristics of these water
samples are listed in Table 2. These samples represent low-
NH3 process waters that are typically subjected to disinfection
by chlorine oxidants.
Reaction Monitoring by Analysis of Substrate Loss. An
Agilent 1100 series HPLC system equipped with a Zorbax
RX-C18 column (4.6 mm × 250 mm, 5 µm), thermostat (30
°C), and UV diode-array detector was used to monitor loss
of parent compounds and product formation during kinetic
measurements. Detection wavelengths for SMX, DMI, APMS,
and MMIB were set at 275, 210, 210, and 205 nm, respectively.
9
5 6 0 8 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 21, 2004

Depending on the chromatographic behavior of respective
analytes, gradient and isocratic methods were used with
varying ratios of 0.04 M phosphate buffer (containing 1:1
[H3PO4]:[NaH2PO4]) to pure acetonitrile. Sample injection
volumes of 100 µL were used in all cases, and typical
quantification limits for analytes under these conditions were
∼100 nM. Standard curves yielded coefficients of determi-
nation (r 2) greater than 0.99 within the range of experimental
concentrations in all cases. Recoveries of SMX after spiking
into real water samples were 100 ( 5%. Experiments
performed to monitor reactions of SMX in clean water systems
were conducted in triplicate, while all other experiments were
conducted in duplicate. Oxidant controls were included in
all experimental sets to ensure that substrates were stable in
the absence of added FAC or CC.
Clean Water Experiments. Reaction solutions contained
within 30-mL, amber, borosilicate glass bottles were partially
immersed in a 25 °C recirculated water tank and stirred with
Teflon-coated stir bars using a 15-position magnetic stir-
plate. Experiments were conducted at pH values between 4
and 9. Ten millimolar acetate (pH 4, 4.5, and 5), phosphate
(pH 6, 6.5, 7, and 8), and borate (pH 9) buffers were used to
maintain pH in all reagent water experiments. Final pH of
reaction solutions did not vary by more than 0.1 units from
starting values.
Reactions were initiated by adding appropriate volumes
of FAC or CC stock (to achieve at least a 10:1 [oxidant]0:
[substrate]0 ratio) to solutions containing 500 µg/ L of
substrate. One Milliliter samples were subsequently taken at
evenly-spaced time intervals and analyzed by HPLC-UV to
follow loss of substrate. Loss of SMX in the presence of CC
was monitored by HPLC-UV immediately after sampling,
without quenching. Reaction times for CC and SMX were
corrected by taking into account the dead-volume of the
HPLC system to calculate the exact time at which substrates
were separated from the inorganic chloramine oxidants. In
all other cases, 1-mL samples of each reaction solution were
quenched at appropriate time intervals to remove excess
FAC. Sodium thiosulfate was used to quench residual FAC
in experiments with DMI, while samples of MMIB were
analyzed both with and without thiosulfate quenching. A
“soft” quenching technique using NH4Cl was utilized to
consume residual FAC in experiments with SMX and APMS.
Additional details of the latter quenching procedure are
included in the Supporting Information (Text S2). FAC and
CC residual concentrations were measured by DPD colo-
rimetry or DPD-FAS titrimetry (24) at the conclusion of each
kinetic experiment. Oxidant concentrations did not generally
decrease by more than 10% from starting concentrations
during the course of these experiments. Even in cases of
greater chlorine loss, however, pseudo-first-order plots of
substrate loss remained satisfactorily linear (see below).
Municipal Wastewater and Drinking Water Experiments.
Real water matrix experiments were conducted without
modification of solution pH. Twenty milliliter volumes of
each real water sample were spiked with 500 µg/ L of SMX
and dosed with FAC concentrations selected to approximate
those utilized by the plant from which they were derived,
which yielded free chlorine residuals of 1-2 mg/ L. Reactions
were followed in the same manner as for the clean systems
described above, except that oxidant decay was also taken
into account by measuring residual FAC concentrations at
selected intervals during and after corresponding monitoring
periods.
for SMX as error bars in relevant graphs. Additionally, the
relationship between measured k′ and dose of FAC was
app
found to be linear within the range of 10:1 to 25:1 [FAC]0:
[SMX]0 ratios. A log-log plot of [FAC]0 against k′ for SMX
app
yielded a line with a slope close to 1, indicating that this
reaction can be treated as first-order with respect to FAC.
The SMX/ FAC reaction can therefore be described as a
bimolecular, second-order reaction: substrate + FAC f
products.
Product Identification. LC/ MS Analysis of Reaction
Product Mixtures. SMX was added to 0.1 M pH 7 phosphate
buffer to achieve starting concentrations of 100 mg/ L. FAC
solution was subsequently added to initiate reactions at
oxidant:substrate molar ratios ranging from 1:2 to 4:1.
Unquenched sample aliquots were immediately analyzed
by an Agilent 1100 series LC/ MS system equipped with a
Zorbax SB-C18 column (2.1 mm ×150 mm, 5 µm), thermostat
(30 °C), UV diode-array detector, and a single-quadrupole
mass spectrometer. Analyte peaks were resolved using
gradient elutions with varying ratios of 0.2% (v/ v) formic
acid to pure acetonitrile. MS analyses were conducted using
positive mode electrospray ionization (ESI+), over a mass
scan range of 50-1000 m/ z. The mass spectrometer frag-
mentation voltage was typically set to 80 eV, unless additional
fragmentation of molecular ions was desired, in which case
this voltage was set to 120 eV. Spray chamber temperature
and drying gas flow were set to 325 °C and 10 L/ min,
respectively.
Fraction Collection and LC/ MS Confirmation of Products
Observed during Kinetic Studies. Because the HPLC columns
and analytical method parameters used for SMX kinetic
studies (HPLC/ UV/ FLD) and product identification (LC/ MS)
were different, identities of product peaks observed during
kinetic monitoring were investigated by collecting corre-
sponding HPLC fractions and analyzing these by the LC/ MS
methods described above (additional details in the Sup-
porting Information, Text S3).
1H NMR and GC/ MS Analyses of Product SMX2. 1H NMR
and GC/ MS were utilized to provide structural information
for one degradation product of SMX not identifiable by LC/
MS (additional details in the Supporting Information, Text
S4).
Additional Analyses. Sulfate was qualitatively monitored
by precipitation with BaCl2 and subsequent measurement
of turbidity (24).
Results and Discussion
Kinetics of Reactions with Free Available Chlorine. Apparent
pseudo-first-order rate constants (k′app) observed for reac-
tions of SMX with FAC were clearly dependent on solution
pH (Figure 2). This behavior was evaluated by modeling
relative contributions of individual oxidant and antibacterial
substrate species to the overall second-order reaction
described by
d[substrate]tot
) -ka′pp[substrate]tot
)
dt
-k′′ [oxidant]tot[substrate]tot (1)
app
where ka′′pp represents the apparent second-order rate con-
stant for the overall reaction.
Solution pH affects the speciation of both free chlorine
oxidant and antibacterial substrate. HOCl can be modeled
as a typical monoprotic acid, using eq 2 and standard
equilibrium distribution equations for R1(HOCl) and R2(OCl-):
Calculation of Pseudo-First-Order Rate Constants and
Verification of Reaction Orders. Plots of ln([substrate]) versus
time were linear in all kinetic experiments, with r 2 values
ranging from 0.95 to 1.0. Pseudo-first-order rate constants
ka′pp were obtained from the slopes of regression lines of
[oxidant]tot ) [HOCl] + [OCl-] )
Ri[oxidant]tot (2)
∑
such plots. The 95% confidence limits are reported with k′
i)1,2
app
9
VOL. 38, NO. 21, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 5 6 0 9

FIGURE 3. pH dependency of apparent pseudo-first-order rate
constants for reactions of SMX with CC ([SMX]0 ) 2.0 × 10-6 M,
[CC]0 ) 2.0 × 10-5 M (1.4 mg/L), T ) 25 °C).
FIGURE 2. pH dependency of apparent pseudo-first-order rate
constants for reactions of SMX with FAC ([SMX]0 ) 2.0 × 10-6 M,
[FAC]0 ) 2.0 × 10-5 M (1.4 mg/L), T ) 25 °C).
therefore be modeled by
SMX can be modeled as a typical diprotic acid by applying
eq 3 and standard equilibrium distribution equations for R′
1
d[SMX]tot
(cationic SMX), R′ (neutral SMX), and R′ (anionic SMX):
2
3
) (-k12R1R′ - k13R13′)[oxidant]tot[substrate]tot
2
dt
(6)
[substrate]tot
)
[H2A+] + [HA] + [A-] )
R′[substrate]tot (3)
j
∑
The specific rate constants kij for reactions among HOCl
and each individual substrate species were calculated by least-
squares regression of the experimental data. The regression
was performed by using the Solver function in Microsoft
Excel to minimize the square of the difference between
measured and predicted values via optimization of the kij
values. One constraint was imposed on these calculations to
permit determination of specific second-order rate constants
with this approach: that is, substrate species exhibiting a
higher degree of functional group deprotonation were
assumed to be equally or more reactive toward HOCl than
their more fully protonated structural variants (in agreement
with findings from prior investigations of pH-dependent
chlorination reaction rates; 20-22, 26). The k12 (i.e., reaction
of HOCl with neutral SMX) and k13 (i.e., reaction of HOCl
with anionic SMX) values obtained with this procedure are
1.1 × 103 and 2.4 × 103 M-1 s-1, respectively.
j)1,2,3
Incorporation of these oxidant and substrate speciation
relationships into eq 1 yields the general expressions shown
in eqs 4 and 5:
d[substrate]tot
) -
kijRi[oxidant]totR′[substrate]tot
j
∑
i)1,2
dt
j)1,2,3
(4)
(5)
k′′ ) -
kijRiR′
j
app
∑
i)1,2
j)1,2,3
where kij represents the specific second-order rate constant
for reaction of oxidant species i with substrate species j.
Apparent second-order rate constants (k′′ ) can be deter-
As a check on the assumptions made in developing eq 6,
a regression of the experimental data was also performed
with respect to the general model (eq 4). This regression
yielded values close to 0 for the constants kij corresponding
to reactions omitted from the simplified models. Exclusion
of these values did not significantly affect model accuracy.
Application of eq 6susing the calculated k12 and k13 valuess
allows one to successfully model the observed pH depend-
encies of apparent pseudo-first-order rate constants k′app for
SMX as shown in Figure 2. This provides additional support
for the assumptions made in model development.
app
mined using eq 1, where ka′pp (in s-1) ) k′′ [oxidant]tot and
app
ka′′pp (in M-1 s-1) ) k′ / [oxidant]tot
.
app
The general reaction kinetics model represented by eqs
4 and 5 can be simplified by critically evaluating apparent
contributions of each oxidant or substrate species to the rate
constants presented in Figure 2. The reactivity of SMX with
FAC dropped off sharply with decreasing abundance of HOCl,
suggesting that SMX rate constants are dominated by
reactions involving HOCl and that OCl- does not contribute
significantly to the overall reactivity of SMX with FAC. Similar
observationssthat HOCl (rather than OCl-) represents the
primary oxidant species in FAC reactionsscan be found in
the literature (20-22, 26). Reactions with OCl- have thus
been neglected in model calculations. In addition, reaction
of HOCl with the cationic form of SMX has been neglected
on the basis of two assumptions: first, abundance of the
SMX cation is rather low within the pH ranges studied (pKa
for the aromatic amine is 1.7, which is 2.3 pH units lower
than the lowest pH value studied); second, protonation of
the aniline’s primary amino group should at the very least
retard (if not prevent) the reaction between HOCl and SMX
by coordinating the lone-pair electrons associated with this
nitrogen. The kinetics for reaction of SMX with HOCl can
Kinetics of Reactions with Com bined Chlorine. Reactions
of SMX with CC were much slower than with FAC (Figure 3).
Minimum measured half-life was 1 d for reactions of SMX
(pH 6) under pseudo-first-order conditions ([CC]0:[substrate]0
) 10). The 180-d half-life determined for SMX at pH 9
indicates that reactions with monochloramine are extremely
slow and that CC reaction rates observed at lower pH were
dominated by dichloramine. These observations agree with
expected oxidation strengths for each oxidant species, where
HOCl/ OCl- > NHCl2 > NH2Cl (27). This indicates that
reactions of SMX with CC are unlikely to be important in
municipal wastewater or drinking water disinfection pro-
cessesswhere monochloramine typically represents the
primary chloramine disinfectantsunless breakpoint chlo-
rination is practiced, in which case reaction rates should still
be quite slow.
9
5 6 1 0 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 21, 2004

FIGURE 4. Mass spectra for major products of the reaction between SMX and FAC: (a) N-chlorinated SMX [LC/MS spectrum obtained
at 120 eV fragmentation voltage], (b) Product SMX1, and (c) Product SMX2. Spectra a and b were obtained by LC/MS-ESI+, and spectrum
c was obtained by GC/MS-EI. Characteristic LC/MS fragment ions were identified according to Hartig et al. (5); GC/MS fragments were
postulated according to McLafferty and Turecˇek (54).
Reactivity of Substructure Com pounds and Identifica-
tion of Reaction Centers. Reactions of substructure com-
pounds with FAC were examined at pH 4, 7, and 9 using
pseudo-first-order conditions ([FAC]0:[substrate]0 ) 10). DMI
and MMIB were nonreactive with FAC (no measurable
decrease in concentration over 30 to 90 min), while APMS
was quite reactive (ka′pp ) 2.5 × 10-2, 2.2 × 10-3, and 9.8 ×
10-4 s-1 at pH 4, 7, and 9, respectively, in the presence of 20
µM FAC). The reactivity of APMS (which contains the aniline
amino-nitrogen but no sulfonyl amido-nitrogen) and ap-
parent non-reactivity of MMIB (which contains the sulfonyl
amido-nitrogen but no aniline amino-nitrogen) toward FAC
indicate that direct attack of the SMX molecule by FAC occurs
at its aniline moiety rather than at its amido-nitrogen.
Furthermore, the non-reactivity of DMI indicates that the
5-methylisoxazole moiety of SMX does not play an important
role in reactions with free chlorine. The observed pH
dependency of SMX reaction kinetics (where anionic SMX
appeared to be more reactive toward FAC than neutral SMX
(Figure 2)) suggests that, although direct reaction occurs at
SMX’s aniline moiety, this reaction is modulated by proto-
nation equilibria of the SMX amido-nitrogen. Protonation of
the amido-nitrogen in the neutral SMX species should
enhance the electron-withdrawing strength of SMX’s p-
sulfonamide group relative to the corresponding anionic
species, which would in turn reduce the nucleophilicity of
the molecule’s aromatic system by inductive and resonance
effects, thereby reducing its susceptibility to attack by FAC.
Product Identification. Transformation products de-
scribed here pertain only to reactions involving SMX and
FAC. Reactions with CC appeared to be unimportant with
regard to water treatment, as discussed earlier, and so were
not investigated in further detail. Two products observed in
LC/ MS-ESI+ analyses of under-chlorinated SMX/ FAC reac-
tion mixtures ([FAC]0:[SMX]0 < 1)sboth exhibiting molecular
ions of m/ z 288 (with isotopic peaks at 290)sappeared to
represent N-chlorinated and ring-chlorinated SMX molecules
(Figure 4a,b). The former product appeared to be reduced
to SMX upon treatment with Na2S2O3 (Supporting Informa-
tion Figure S1), hence its description as an N-chlorinated
structure. The latter compound (Product SMX1) was detected
only at very low yield, so its behavior could not be investigated
in a similar manner. However, comparison with results from
the prior study of SMX chlorination (17) and consideration
of expected aniline halogenation patterns (28) support the
structural assignments.
A number of additional products were observed in
solutions dosed with stoichiometric excess of FAC at [FAC]0:
[SMX]0 > 1 (Supporting Information Table S1). A prominent,
early-eluting peak detected by LC/ MS was identified as
3-amino-5-methylisoxazole (AMI) (refer to Scheme 1 for
structure) by comparison to a standard of this compound.
Small amounts (as indicated by small MS signal areas) of
apparent dimeric or polymeric products were also detected
by LC/ MS, although only one appeared to be identifiable, as
azosulfamethoxazole (Scheme 1 and Supporting Information
9
VOL. 38, NO. 21, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 5 6 1 1

a
SCHEME 1. Proposed Pathways for Reactions of SMX with FAC
a
b
Structures in brackets are given as probable interm ediates. Estim ated yield obtained from data reported by Uetrecht et al. (17).
Table S1). In addition to AMI and dimeric/ polymeric
products, Product SMX2 was observed by UV detection (λmax
≈ 290 nm) but yielded no signal in positive APCI or ESI modes
of LC/ MS analyses. Product SMX2 was isolated via solid-
phase extraction methods (Supporting Information Text S4)
and analyzed by GC/ MS and 1H NMR.
Proposed Reaction Pathways. Observations pertaining
to apparent ortho-chlorination of the SMX aniline ring in
solutions dosed with [FAC]0:[SMX]0 < 1 are in agreement
with results reported previously by Uetrecht et al. (17). These
investigators suggested that N-chlorinated SMX rearranged
to ortho-chlorinated SMX during analytical time frames of
roughly 1-2 h. Ring-chlorination of anilines has previously
been demonstrated to proceed primarily through intermedi-
ate N-chlorinated aniline structures (28, 34). In the current
study, repeated analyses of unquenched under-chlorinated
reaction solutions showed that a decrease in the N-chlori-
nated SMX peak observed in HPLC chromatograms cor-
responded to an increase in the SMX peak. This suggests
that N-chlorinated SMX also undergoes a back-reaction in
the absence of reducing agentssto yield the parent SMXs
over a time scale of several hours. This is in accord with
results previously obtained by Gassman and Campbell (35),
which indicated that N-chlorinated anilines containing
electron-withdrawing para-substituents undergo solvolytic
decay to yield relatively high proportions (10-30%) of the
parent compounds, in addition to ring-chlorinated products.
The current findings related to generation of NCBQ, AMI,
and SO2 in over-chlorinated SMX solutions point toward an
additional reaction mechanism. Evaluation of SMX/ FAC
product evolution patterns showed that loss of SMX was
accompanied by an initial increase in the concentration of
N-chlorinated SMX (Figure 5). As mentioned in the preceding
paragraph, N-chlorinated SMX decayed slowly over a span
of1 to 2 h in the absence of excess FAC (17). However, product
evolution observations obtained in the current investigation
indicated that N-chlorinated SMX decayed within seconds
in the presence of excess FAC (Figure 5). In the latter case,
decay of N-chlorinated SMX was followed by an increase in
abundance of NCBQ. Quite notably, however, the decrease
in abundance of N-chlorinated SMX did not appear to
correlate directly with the evolution of NCBQ. NCBQ signal
area continued to increase even after all N-chlorinated SMX
had reacted. These observations indicate that the appearance
of NCBQ is likely preceded by formation and decay of an
additional intermediate.
GC/ MS analyses of Product SMX2 (Figure 4c and Sup-
porting Information Table S1) showed a molecular ion of
m/ z 141 (chlorine isotope peak at m/ z 143) and suggested
that the structure could be N-chloro-p-benzoquinoneimine
(NCBQ), which is an intermediate in the Berthelot reaction
for analysis of ammonia (29) and belongs to a family of
compounds commonly applied in the Gibbs reaction for
colorimetric analysis of phenolic compounds in aqueous
solution (30, 31). The UV spectrum and observed half-life
(∼48 h) for decay of a Product SMX2 isolate in aqueous
solution matched those previously reported for NCBQ (29).
In addition, the apparent reduction of Product SMX2 by
sodium thiosulfate and its ability to oxidize potassium iodide
were consistent with such a structure.The 1H NMR spectrum
of Product SMX2 (Supporting Information Figure S3) was
also compatible with the NCBQ structure, where (i) signals
corresponding to the methyl group and heterocyclic proton
of SMX’s isoxazole ring (as observed for pure SMX) were
absent from the spectrum of Product SMX2 and (ii) instead
of two aromatic proton signals (as observed for pure SMX),
the spectrum of Product SMX2 exhibited four proton signals,
presumably due to anisotropic effects created by the presence
of an N-chlorimino group (32). Final confirmation of Product
SMX2 as NCBQ was obtained from comparison of the former
compound’s UV spectrum and HPLC retention time to those
observed for a standard of the latter. Quantification of NCBQ
in HPLC monitoring showed the ultimate yield of this
transformation product to be approximately 85% when FAC
was present in greater than 2-fold molar excess of SMX.
Qualitative measurement of SO42- in solutions dosed with
excess FAC and varying concentrations of SMX indicated
that SO2 is also evolved during the reaction of FAC with SMX
(note that all SO2 would undergo hydration to yield sulfite,
2-
which would be quantitatively oxidized to SO4 in the
presence of excess FAC; 33). Taken together with the products
presented in the preceding paragraphs, the results indicate
that SMX’s sulfonamide structure is disrupted in reactions
of SMX with excess FAC, yielding NCBQ, AMI, and SO2. None
of these products were mentioned in the prior study of SMX
chlorination (17). However, review of that study shows that
reactions were only evaluated in cases for which SMX was
present in substantial molar excess of FAC.
Scully and co-workers (36-38) reported quite extensively
that N,N-dichlorinated amino acidsswhich were generated
from amino acids in the presence of molar excesses of FACs
decayed primarily to N-chlorimine products. Similarly,
synthesis of NCBQ can be achieved by dosing solutions of
4-aminophenol with excess FAC (23, 30). The above reactions
can likely be extrapolated to other aromatic amine com-
pounds such as aniline or SMX. Thus, it seems reasonable
9
5 6 1 2 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 21, 2004

system, resulting in distribution of strong electron-deficiency
to the aromatic ring’s para and ortho positions. Arylnitrenium
species are known to exhibit extremely high rate constants
(∼108-1010 M-1 s-1) for reactions with nucleophiles such as
N3-, Cl-, Br-, and H2O, where nucleophilic attack occurs
preferentially at the carbon atom para to the divalent amino
nitrogen (44, 45). Aqueous reactions of para-substituted
nitrenium ions typically lead to a p-hydroxy intermediate,
which can subsequently rearrange to p-benzoquinoneimine
structures (46, 47). As proposed in Scheme 1, H2O may attack
at the para position of a N-chlorinated SMX nitrenium species
(produced by heterolytic decay of N,N-dichlorinated SMX),
resulting in displacement of the sulfonamide group by water.
This reaction would be expected to yield NCBQ and an
unstable AMI-SO2 complex. The AMI-SO2 complex would
decay readily in the presence of water to yield AMI and SO2
(40). Both AMI and SO2 (after hydration to SO32-) would then
be expected to react subsequently with FAC. Although
products of AMI chlorination were not explicitly investigated,
FIGURE 5. Product evolution for reactions of SMX with FAC ([SMX]0
) 2.0 × 10-6 M, [FAC]0 ) 2.0 × 10-5 M (1.4 mg/L), pH 6.5, T ) 25
2-
2-
SO3 would clearly be oxidized to SO4 (33).
°C). Samples were quenched according to the NH Cl method
4
An alternate explanation for the generation of NCBQ, AMI,
and SO2 might involve direct nucleophilic attack by H2O or
OH- at the para position of N,N-dichlorinated SMX via a
more conventional SN2 substitution. In this case, nucleophilic
addition would produce heterolysis of a N-Cl bond via
intramolecular redistribution of electrons to yield the chlo-
rimine rather than vice versa, as described in the preceding
paragraph. Unfortunately, the similar expected product
distributions of this and the nitrenium mechanism make
explicit confirmation of either a difficult task. However, the
results obtained by Scully and co-workers (36-38) provide
a clue as to which process actually governs the decay of SMX.
Their studies showed that decay of N,N-dichlorinated primary
amino acids was initiated by heterolysis of one N-Cl bond,
which led to generation of the N-chlorimine product. No
products were detected that would have suggested that
hydrolytic attack occurred at electrophilic R-carbons adjacent
to the dichloramine groups prior to N-Cl heterolysis. The
sequence of reaction steps (i.e., N-Cl heterolysis followed
by nucleophilic substitution and chlorimine formation)
depicted in Scheme 1 is in analogy to the findings in these
previous studies (36-38).
Azosulfamethoxazole (Scheme 1) presumably forms via
a coupling reaction involving N,N-dichlorinated SMX (since
N-chlorinated SMX was only observed to react back to SMX
or to rearrange to Product SMX1; 17), similar to polymeri-
zation reactions previously reported to occur during chlo-
rination of aromatic amines (48-50). Whether this reaction
occurs via a radical mechanism (48, 51) or through reactions
involving singlet nitrene species (47) is unclear. However,
the high yields of NCBQ (∼85%) indicate that the dimerization
reaction is not likely very important in the overall SMX/ FAC
reaction scheme. Furthermore, generation of coupling
products is probably even less significant in systems con-
taining lower SMX concentrations typical of environmental
samples.
described in the Supporting Information. A/Amax represents the ratio
of peak area at any given time to maximum measured peak area
at 275 nm.
FIGURE 6. Structures previously found to be susceptible to sulfonyl
group displacement by nucleophilic attack: (a) 4-chloroquinoline
sulfonyl chloride (39) and (b) PNU-15189 (40).
to propose that further chlorination of N-chlorinated SMX
leads to formation of a N,N-dichlorinated SMX intermediate,
which subsequently decays to yield the N-chlorimine NCBQ
as well as AMI and SO2 (Scheme 1). Failure to detect the
N,N-dichlorinated SMX intermediate may have been a result
of typical 2-4 min lag times between sampling and instru-
mental analysis, which could have allowed sufficient time
for its decay prior to detection. The mechanism by which
decay of N,N-dichlorinated SMX could lead to cleavage of
the sulfonamide moiety requires additional explanation, as
discussed below.
Kwart and Body (39) observed facile cleavage of the S-C
bond during chlorination of 4-quinoline sulfonyl chloride
(structure provided in Figure 6a), which they attributed to
production of electron-deficient character in the heterocyclic
quinoline’s nitrogen by attack of chlorine. They suggested
that delocalization of this electrophilic character within the
substrate’s aromatic system led to expulsion of the sulfonyl
group via facilitation of nucleophilic attack at the R-carbon
adjacent to the sulfur atom (para to the heterocyclic nitrogen).
A more recent study showed that cleavage of sulfonamides
could be readily achieved by enzyme-mediated nucleophilic
attack at the R-carbon of the benzenesulfonamide moiety in
compounds possessing very strong electron-withdrawing
substituents para to the sulfonamide group (e.g., the structure
shown in Figure 6b; 40). In each of these studies, the presence
of a strongly electrophilic carbon center R- to the sulfonyl
or sulfonamide groups’ S was cited as essential for the S-C
bond cleavage to occur.
In light of these results, one could deduce that SMX/ FAC
reactions should contribute at least to partial reduction or
elimination of the parent compound’s antibacterial activity
(derived from its antagonistic competition with p-ami-
nobenzoic acid for dihydropteroate synthase enzyme, which
is necessary for bacterial folic acid synthesis; 52). Mechanisms
leading to disruption of the sulfonamide structure (shown
in Scheme 1) could very likely lead to a loss of this activity
from the parent compound, although the resulting products,
particularly NCBQ, might possess higher acute toxicity to
aquatic organisms than the parent substrates. On the other
hand, the effects of aromatic ring-chlorinationsin which the
p-aminobenzenesulfonamide structure is essentially con-
servedsare less clear.
An arylnitrenium cation (Scheme 1, see ref 41 for an
overview)swhich can be generated by heterolysis of aromatic
chloramines’ N-Cl bonds (42, 43)sis compatible with the
requirement stated above and could provide an explanation
for the observations noted here with respect to SMX/ FAC
reactions. Substantial positive character is delocalized from
an arylnitrenium’s cationic nitrogen atom into its aromatic
9
VOL. 38, NO. 21, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 5 6 1 3

h), since the reaction appears to be initiated by attack of a
single HOCl molecule on a single substrate molecule. This
expectation is supported by existing field-scale observations
(7, 16).
However, two cautionary notes should be added to this
discussion. First, reaction pathways for SMX proceed through
a reducible intermediate, which can be converted to the
parent compound upon treatment with strong nucleophiles
such as SO32- or S2O32-. This suggests that in North American
wastewater treatment facilities, at which dechlorination with
reduced sulfur compounds is commonly practiced after
chlorine disinfection, the parent compound could be re-
generated from the N-chlorinated intermediate prior to final
discharge of wastewater effluent into the environment. This
could provide one explanation for the detection of SMX in
North American surface waters (4), for which the main points
of entry are likely those associated with municipal wastewater
discharges. Second, as discussed above, N-chlorinated SMX
was observed to decay either by ring-chlorination (<10%) or
by back-reaction to the parent SMX. This suggests that, in
cases where insufficient free chlorine is added to treated
waters to initiate the sulfonamide rupture mechanism
described herein (Scheme 1), a significant fraction of the
N-chlorinated intermediate may simply react back to SMX.
Such behavior could provide an additional explanation for
the frequent detection of SMX within affected surface water
bodies.
Finally, emphasis must be placed on the fact that the real
water matrixes used in this study were typical of low-NH3
process waters. In each case, the FAC dose applied was at
least 2-fold greater (on a molar basis) than measurable NH3
concentrations, so that a FAC residual was present. Under-
chlorination of process waters containing excess NH3 would
not likely lead to significant loss of SMX, because NH3 should
out-compete SMX for FAC (on the basis of expected reaction
rates for NH3 with FAC at near-neutral pH; 22), and SMX
does not react appreciably with CC.
FIGURE 7. Reaction kinetics of SMX in real waters (T ) 25 °C).
Wastewater: pH 7.3, [SMX]0 ) 2.0 × 10-6 M, [FAC]0 ) 1.6 × 10-4
M (11 mg/L); drinking water: pH 6.6, [SMX]0 ) 2.0 × 10-6 M, [FAC]0
) 2.8 × 10-5 M (2 mg/L).
The findings presented here are particularly important in
two respects. First, the chlorination reaction mechanisms
elucidated for SMX are most likely applicable to other
sulfonamide antibacterials because the reactive p-ami-
nobenzenesulfonamide moiety is common to all members
of this structural family. Second, results of the current study
are quite significant in understanding general reactivities of
aromatic amine-containing compounds toward free chlorine.
Direct electrophilic ring-chlorination has been previously
suggested as the predominant transformation pathway in
aqueous reactions of aromatic amine structures with FAC
(53). However, in the most comprehensive prior investigation
to address aqueous chlorination of anilines, the authors used
a
maximum [FAC]0:[substrate]0 ratio of 1.5 (51). As a
consequence, formation of N,N-dichloroanilines would have
been minimal, so that associated products would have been
generated at relatively low yields and may have escaped
detection. The results of one additional investigation indi-
cated that ring-chlorinated products accounted for only 30%
of total aniline loss (50). Furthermore, the authors of the
latter study suggested production of polymeric substances
such as indoaniline, which forms through a reaction pathway
involving NCBQ (49)sduring extended reaction periods.
Taken into context with the current results, a re-evaluation
of aqueous aniline-chlorine chemistry at higher [FAC]0:
[substrate]0 ratios more typical of water treatment seems
warranted.
Additional investigations will be necessary to more fully
assess implications of the current work, particularly with
respect to the significance of sulfonamide degradation
products. In general, future monitoring studies addressing
the environmental fate of antibacterials should also begin to
take into account the potential presence of degradates
resulting from the chemical transformation mechanisms
presented here, since it is possible that perceived elimination
of parent antibacterials in treated waters might not always
correspond to elimination of their biologically active struc-
tural moieties.
Studies in Real Water Matrixes. Reactions of SMX in real
water samples were modeled by assuming two different
reaction steps: one characterized by rapid decay of FAC
during the first 10 s (due to rapid consumption by highly
reactive substrates such as reduced sulfur compounds,
amines, etc. in the water samples), and one by slower FAC
decay during the remaining reaction time. FAC decay during
the two reaction periods was treated as pseudo-first-order
and was modeled using FAC residual measurements taken
at various time intervals during the experimental monitoring
periods.
Observed oxidation kinetics for SMX corresponded quite
closely to those predicted by the second-order model after
consideration of concurrent FAC decay (Figure 7). Product
evolution patterns in environmental matrix experiments
(observed via HPLC/ UV) were also qualitatively similar to
those observed in clean systems.
Acknowledgments
This study was funded by the Georgia Water Resources
Institute. Financial support for M.C.D. from the U.S. Envi-
ronmental Protection Agency STAR fellowship and the
Georgia Tech Foundation is gratefully acknowledged. The
authors thank Jaehong Kim, James Mulholland, Les Gelbaum,
and Gretchen Onstad for providing helpful suggestions in
discussions of this work as well as Urs von Gunten, Marc
Huber, and Amisha Shah for their helpful comments and
continuous support. In addition, the time and insightful
recommendations of three reviewers are highly appreciated.
Supporting Information Available
Text, figures, and tables addressing (1) real water sample
procurement, source characteristics, and parameter mea-
surements; (2) reaction quenching procedures; (3) supple-
mental means of product identification; and (4) principal
reaction products identified by LC/ MS and 1H NMR. This
material is available free of charge via the Internet at http:/ /
pubs.acs.org.
The rates at which SMX decayed in these real water
samples indicate that substantial conversion of the parent
compound can be expected for residence times typical of
wastewater (5-30 min) and drinking water clearwells (1-24
9
5 6 1 4 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 21, 2004

(29) Harfmann, R. G.; Crouch, S. R. Talanta 1989, 36, 261-269.
(30) Gibbs, H. D. J. Biol. Chem. 1927, 72, 649-664.
(31) Pallagi, I.; Toro´ , A.; Horva´th, G. J. Org. Chem. 1999, 64, 6530-
6540.
(32) Saito` , H.; Nukada, K. Can. J. Chem. 1968, 46, 2989-3000.
(33) Fogelman, K. D.; Walker, D. M.; Margerum, D. W. Inorg. Chem.
1989, 28, 986-993.
(34) Haberfield, P.; Paul, D. F. J. Am. Chem. Soc. 1965, 87, 5502.
(35) Gassman, P. G.; Campbell, G. A. J. Am. Chem. Soc. 1972, 94,
3891-3896.
(36) Nweke, A.; Scully, F. E., Jr. Environ. Sci. Technol. 1989, 23, 989-
994.
(37) Conyers, B.; Scully, F. E., Jr. Environ. Sci. Technol. 1997, 31,
1680-1685.
(38) Fox, T. C.; Keefe, D. J.; Scully, F. E., Jr.; Laikhter, A. Environ. Sci.
Technol. 1997, 31, 1979-1984.
(39) Kwart, H.; Body, W. R. J. Org. Chem. 1964, 30, 1188-1195.
(40) Zhao, Z.; Koeplinger, K. A.; Peterson, T.; Conradi, R. A.; Burton,
P. S.; Suarato, A.; Heinrikson, R. L.; Tomasselli, A. G. Drug Metab.
Dispos. 1999, 27, 992-998.
(41) Novak, M.; Rajagopal, S. Adv. Phys. Org. Chem. 2001, 36, 167-
254.
(42) Gassman, P. G.; Campbell, G. A. J. Am. Chem. Soc. 1971, 93,
2567-2569.
(43) Davidse, P. A.; Kahley, M. J.; McClelland, R. A.; Novak, M. J. Am.
Chem. Soc. 1994, 116, 513-4514.
(44) Fishbein, J. C.; McClelland, R. A. J. Am. Chem. Soc. 1987, 109,
2824-2825.
(45) Fishbein, J. C.; McClelland, R. A. Can. J. Chem. 1996, 74, 1321-
1328.
(46) Novak, M.; Kahley, M. J.; Lin, J.; Kennedy, S. A.; James, T. G. J.
Org. Chem. 1995, 60, 8294-8304.
(47) McClelland, R. A.; Kahley, M. J.; Davidse, P. A.; Hadzialic, G. J.
Am. Chem. Soc. 1996, 118, 4794-4803.
(48) Goldschmidt, S.; Strohmenger, L. Chem. Ber. 1922, 55, 2450-
2470.
(49) Noller, C. R. Chemistry of Organic Compounds, 3rd ed.; W. B.
Saunders Company: Philadelphia, 1965.
Literature Cited
(1) National Research Council. The Use of Drugs in Food Animals;
National Academy Press: Washington, DC, 1999.
(2) Walsh, C. Antibiotics: Actions, Origins, Resistance; ASM Press:
Washington, DC, 2003.
(3) Ternes, T. A. Trends Anal. Chem. 2001, 20, 419-434.
(4) Kolpin, D. W.; Furlong, E. T.; Meyer, M. T.; Thurman, E. M.;
Zaugg, S. D.; Barber, L. B.; Buxton, H. T. Environ. Sci. Technol.
2002, 36, 1202-1211.
(5) Hartig, C.; Storm, T.; Jekel, M. J. Chromatogr., A 1999, 854, 163-
173.
(6) Miao, X.-S.; Bishay, F.; Chen, M.; Metcalfe, C. D. Environ. Sci.
Technol. 2004, 38, 3542-3550.
(7) Renew, J. E.; Huang, C.-H. J. Chromatogr., A 2004, 1042, 113-
121.
(8) Gon˜ i-Urriza, M.; Capdepuy, M.; Arpin, C.; Raymond, N.;
Caumette, P.; Quentin, C. Appl. Environ. Microbiol. 2000, 66,
125-132.
(9) Guardabassi, L.; Lo Fo Wong, D. M. A.; Dalsgaard, A. Water Res.
2002, 36, 1955-1964.
(10) Schwartz, T.; Kohnen, W.; Jansen, B.; Obst, U. FEMS Microbiol.
Ecol. 2003, 43, 325-335.
(11) Armstrong, J. L.; Calomiris, J. J.; Seidler, R. J. Appl. Environ.
Microbiol. 1982, 44, 308-316.
(12) Al-Ahmad, A.; Daschner, F. D.; Ku¨ mmerer, K. Arch. Environ.
Contam. Toxicol. 1999, 37, 158-163.
(13) Ingerslev, F.; Halling-Sørensen, B. Environ. Toxicol. Chem. 2000,
19, 2467-2473.
(14) Adams, C.; Wang, Y.; Loftin, K.; Meyer, M. J. Environ. Eng. 2002,
128, 253-260.
(15) Huber, M. M.; Canonica, S.; Park, G.-Y.; von Gunten, U. Environ.
Sci. Technol. 2003, 37, 1016-1024.
(16) Renew, J. E. M.S. Thesis, School of Civil and Environmental
Engineering, Georgia Institute of Technology, Atlanta, GA, 2003.
(17) Uetrecht, J. P.; Shear, N. H.; Zahid, N. Drug Metab. Dispos. 1993,
21, 830-834.
(18) Lucida, H.; Parkin, J. E.; Sunderland, V. B. Int. J. Pharm. 2000,
202, 47-61.
(19) Pankratov, A. N.; Uchaeva, I. M.; Doronin, S. Y.; Chernova, R.
K. J. Struct. Chem. 2001, 42, 739-746.
(20) Rebenne, L. M.; Gonzalez, A. C.; Olson, T. M. Environ. Sci.
Technol. 1996, 30, 2235-2242.
(21) Gallard, H.; von Gunten, U. Environ. Sci. Technol. 2002, 36,
884-890.
(22) Qiang, Z.; Adams, C. D. Environ. Sci. Technol. 2004, 38, 1435-
1444.
(50) Jenkins, R. L.; Haskins, J. E.; Carmona, L. G.; Baird, R. B. Arch.
Environ. Contam. Toxicol. 1978, 7, 301-315.
(51) Hwang, S.-C.; Larson, R. A.; Snoeyink, V. L. Water Res. 1990, 24,
427-432.
(52) Stratton, C. W. In Antibiotics in Laboratory Medicine; Lorian, V.,
Ed.; Williams and Wilkins: Baltimore, MD, 1996; pp 579-603.
(53) Larson, R. A.; Weber, E. J. Reaction Mechanisms in Environmental
Organic Chemistry, 1st ed.; Lewis Publishers: Boca Raton, FL,
1994.
(23) Venuvanalingam, P.; Chandra Singh, U.; Subbaratnam, N. R.
Spectrochim. Acta 1980, 36A, 103-107.
(54) McLafferty, F. W.; Turecˇek, F. Interpretation of Mass Spectra,
4th ed.; University Science Books: Sausalito, CA, 1993.
(24) Standard Methods for the Examination of Water and Wastewater,
20th ed.; APHA, AWWA, WPCF: Washington, DC, 1998.
(25) Chapin, R. M. J. Org. Chem. 1986, 51, 2112-2117.
(26) Weil, I.; Morris, J. C. J. Am. Chem. Soc. 1949, 71, 1664-1671.
(27) Victorin, K.; Hellstro¨ m, K.-G.; Rylander, R. J. Hyg. 1972, 70, 313-
323.
Received for review November 4, 2003. Revised manuscript
received July 27, 2004. Accepted August 16, 2004.
(28) Gassman, P. G.; Campbell, G. A.; Frederick, R. C. J. Am. Chem.
Soc. 1972, 94, 3884-3891.
ES035225Z
9
VOL. 38, NO. 21, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 5 6 1 5