Formation of cyanogen chloride from the reaction of monochloramine with formaldehyde

Article abstract of DOI:10.1021/es990153q

Methanediol dehydrates to give formaldehyde, which reacts rapidly and reversibly with monochloramine to form N-chloroaminomethanol. Under drinking water conditions, N-chloroaminomethanol undergoes a relatively slow decomposition that eventually leads to the formation of cyanogen chloride (ClCN) in apparently stoichiometric amounts. The following reaction sequence is proposed: CH₂(OH)₂ ? CH₂O + H₂O; CH₂O + NH₂Cl ? CH₂(OH)NHCl; CH₂(OH)NHCl → CH₂NCl + H₂O; CH₂NCl → HCl + HCN; CN^- + NH₂Cl + H⁺ → ClCN + NH₃. These reactions were studied at 25.0°C and an ionic strength of 0.10 M (NaClO₄). Stopped-flow photometry was used to monitor rapid, reversible reactions, and photometry was used to study relatively slow decomposition reactions. Equilibrium and rate constants for the addition of formaldehyde to monochloramine were (6.6 ± 1.5) x 10⁵ M^-1 and (2.8 ± 0.1) x 10⁴ M^-1 s^-1, respectively. The dehydration of N- chloroaminomethanol was catalyzed by both H⁺ and OH^-, with respective rate constants of 277 ± 7 and 26.9 ± 5.6 M^-1 s^-1. Under characteristic drinking water conditions, the decay of N-chloroaminomethanol is the rate- limiting step. N-Chloromethanimine, formed by the dehydration of N- chloroaminomethanol, had a decomposition rate constant of (6.65 ± 0.06) x 10^-4 s^-1. At the relatively high methanediol concentrations used in this study, the intermediary N-chlorodimethanolamine was formed by the rapid and reversible reaction of N-chloroaminomethanol with formaldehyde. N- Chlorodimethanolamine then decayed relatively slowly. The following reaction sequence is proposed: CH₂(OH)NHCl + CH₂O ? {CH₂(OH)}₂NCl; {CH₂(OH)}₂NCl → CH₂NCl + CH₂O + H₂O. The equilibrium and rate constants for the addition of formaldehyde to N-chloroaminomethanol were (9.5 ± 2.5) x 10⁴ M^-1 and (3.6 ± 0.1) x 10³ M^-1 s^-1, respectively. The decomposition of N-chlorodimethanolamine was catalyzed by OH^-, with a rate constant of 19.2 ± 3.7 M^-1 s^-1. N-Chlorodimethanolamine would not be present under typical drinking water treatment conditions. Methanediol dehydrates to give formaldehyde, which reacts rapidly and reversibly with monochloramine to form N-chloroaminomethanol. Under drinking water conditions, N-chloroaminomethanol undergoes a relatively slow decomposition that eventually leads to the formation of cyanogen chloride (ClCN) in apparently stoichiometric amounts. The following reaction sequence is proposed: CH₂(OH)₂qqCH₂O+H₂O; CH₂O+NH₂ClqqCH₂(OH)NHCl; CH₂(OH)NHCl→CH₂NCl+H₂O; CH₂NCl→HCl+HCN; CN^-+NH₂Cl+H⁺→ClCN+NH₃. These reactions were studied at 25.0 °C and an ionic strength of 0.10 M (NaClO₄). Stopped-flow photometry was used to monitor rapid, reversible reactions, and photometry was used to study relatively slow decomposition reactions. Equilibrium and rate constants for the addition of formaldehyde to monochloramine were (6.6±1.5)×10⁵ M^-1 and (2.8±0.1)×10⁴ M^-1 s^-1, respectively. The dehydration of N-chloroaminomethanol was catalyzed by both H⁺ and OH^-, with respective rate constants of 277±7 and 26.9±5.6 M^-1 s^-1. Under characteristic drinking water conditions, the decay of N-chloroaminomethanol is the rate-limiting step. N-Chloromethanimine, formed by the dehydration of N-chloroaminomethanol, had a decomposition rate constant of (6.65±0.06)×10^-4 s^-1. At the relatively high methanediol concentrations used in this study, the intermediary N-chlorodimethanolamine was formed by the rapid and reversible reaction of N-chloroaminomethanol with formaldehyde. N-Chlorodimethanolamine then decayed relatively slowly. The following reaction sequence is proposed: CH₂(OH)NHCl+CH₂Oqq{CH₂(OH)}₂NCl; {CH₂(OH)}₂NCl→CH₂NCl+CH₂O+H₂O. The equilibrium and rate constants for the addition of formaldehyde to N-chloroaminomethanol were (9.5±2.5)×10⁴ M^-1 and (3.6±0.1)×10³ M^-1 s^-1, respectively. The decomposition of N-chlorodimethanolamine was catalyzed by OH^-, with a rate constant of 19.2±3.7 M^-1 s^-1. N-Chlorodimethanolamine would not be present under typical drinking water treatment conditions.

Full text of DOI:10.1021/es990153q

Environ. Sci. Technol. 1999, 33, 4239-4249
Introduction
Formation of Cyanogen Chloride
Cyanogen chloride (ClCN) is a toxic disinfection byproduct
(DBP) commonly observed in chloraminated drinking water
(1). The formation of ClCN in natural waters and the
corresponding role of water quality parameters are not well
understood. This lack of information prompted the U.S.
Environmental Protection Agency to include ClCN in the
Information Collection Rule (ICR) (2). The ICR required
quarterly monitoring of ClCN by public water systems using
chloramines and serving populations of at least 100 000 with
surface water or 50 000 with groundwater (2). The World
Health Organization recently proposed a guideline value of
70 µg/ L for all cyanide species (including cyanogen com-
pounds) in drinking water (3).
from the Reaction of
Monochloramine with Formaldehyde
E R I K J . P E D E R S E N I I I , †
E D W A R D T . U R B A N S K Y, ‡
B E N I T O J . M A R I N˜ A S , * , § A N D
D A L E W . M A R G E R U M * , ‡
School of Civil Engineering and Department of Chemistry,
Purdue University, West Lafayette, Indiana 47907, and
Department of Civil and Environmental Engineering,
University of Illinois, Urbana, Illinois 61801
Preliminary testing (4, 5) has shown that ClCN forms when
monochloramine and formaldehyde are reacted at concen-
trations representative of drinking water treatment. Form-
aldehyde, also a common DBP, is formed during ozonation
or chlorination of waters containing natural organic matter
(6-8) and during through reaction of glycine with monochlor-
amine (9). Monochloramine is commonly considered as a
secondary disinfectant following ozonation.
In the present work, we show that formaldehyde reacts
rapidly and reversibly with monochloramine to form N-
chloroaminomethanol, CH2(OH)NHCl, a carbinolamine or
hemiaminal. This type of nucleophilic addition reaction is
known to occur by attack of the lone electron pair of primary
amines on the electrophilic carbonyl carbon of aldehydes
(10-15). CH2(OH)NHCl decomposes relatively slowly to form
N-chloromethanimine, H2CdNCl, which can rapidly lose HCl
to yield hydrocyanic acid. Cyanide ion and HCN are known
to react rapidly with monochloramine to produce ClCN (16).
In cases where monochloramine is present in excess of total
formaldehyde (e.g., drinking water conditions), the conver-
sion of formaldehyde to ClCN appears to be stoichiometric
after accounting for ClCN decomposition by hydrolysis (4,
5). Although there are other possible formation pathways for
ClCN in chloraminated drinking water (17-20), the reaction
series from CH2O to CH2(OH)NHCl to CH2NCl to HCN to
ClCN would appear to be a major pathway, particularly when
pre-ozonation is employed.
To study the kinetics of the formation and subsequent
decomposition of N-chloroaminomethanol, the concentra-
tion ratios of the reactants were varied from a slight excess
of monochloramine to a large excess of formaldehyde. Under
conditions of excess formaldehyde, a second rapid, reversible
reaction of formaldehyde occurs with N-chloroaminometha-
nol to form N-chlorodimethanolamine, {CH2(OH)}2NCl,
which also undergoes relatively slow decay. N-Chlorodimeth-
anolamine would not be present at significant levels under
drinking water treatment conditions. The formation and
decomposition of this compound, however, affected reaction
rates under experimental conditions used in this study and
is therefore of interest. The major objectives of this inves-
tigation were to study the rapid formation reactions and also
the relatively slow decay sequences of both N-chloroamino-
methanol and N-chlorodimethanolamine.
Methanediol dehydrates to give formaldehyde, which
reacts rapidly and reversibly with monochloramine to
form N-chloroaminomethanol. Under drinking water
conditions, N-chloroaminomethanol undergoes a relatively
slow decomposition that eventually leads to the formation
of cyanogen chloride (ClCN) in apparently stoichiometric
amounts. The following reaction sequence is proposed:
CH (OH)2 / CH O + H O; CH O + NH Cl / CH (OH)NHCl;
2
2
2
2
2
2
CH (OH)NHCl f CH NCl + H O; CH NCl f HCl + HCN;
2
2
2
2
CN- + NH Cl + H+ f ClCN + NH . These reactions were
2
3
studied at 25.0 °C and an ionic strength of 0.10 M (NaClO ).
4
Stopped-flow photometry was used to monitor rapid,
reversible reactions, and photometry was used to study
relatively slow decomposition reactions. Equilibrium and rate
constants for the addition of formaldehyde to monochloram-
ine were (6.6 ( 1.5) × 105 M-1 and (2.8 ( 0.1) × 104 M-1
s-1, respectively. The dehydration of N-chloroaminomethanol
was catalyzed by both H+ and OH-, with respective rate
constants of 277 ( 7 and 26.9 ( 5.6 M-1 s-1. Under
characteristic drinking water conditions, the decay of
N-chloroaminomethanol is the rate-limiting step. N-Chlo-
romethanimine, formed by the dehydration of N-chloro-
aminomethanol, had a decomposition rate constant of
(6.65 ( 0.06) × 10-4 s-1. At the relatively high methanediol
concentrations used in this study, the intermediary
N-chlorodimethanolamine was formed by the rapid and
reversible reaction of N-chloroaminomethanol with
formaldehyde. N-Chlorodimethanolamine then decayed
relatively slowly. The following reaction sequence is
proposed: CH (OH)NHCl + CH O / {CH (OH)}2NCl;
2
2
2
{CH (OH)}2NCl f CH NCl + CH O + H O. The equilibrium
2
2
2
2
and rate constants for the addition of formaldehyde to
N-chloroaminomethanol were (9.5 ( 2.5) × 104 M-1 and
(3.6 ( 0.1) × 103 M-1 s-1, respectively. The decomposition
of N-chlorodimethanolamine was catalyzed by OH-, with
a rate constant of 19.2 ( 3.7 M-1 s-1. N-Chlorodimeth-
anolamine would not be present under typical drinking water
treatment conditions.
Experimental Section
* Corresponding author phone: (217)333-6961 (B.J.M.) or (765)-
494-5268 (D.W.M); fax: (217)333-6968 (B.J.M.) or (765)494-0239
(D.W.M.); e-mail: marinas@uiuc.edu (B.J.M.) or margerum@
chem.purdue.edu (D.W.M.).
Reagents and Solutions. Formaldehyde solutions were
prepared by dilution of a 37% stock solution also containing
approximately 10% methanol. The stock solution was stand-
ardized gravimetrically by reacting with 5,5-dimethyl-1,3-
cyclohexanedione and forming an insoluble adduct (21).
Paraformaldehyde initially present in the stock solution was
† School of Civil Engineering, Purdue University.
‡ Department of Chemistry, Purdue University.
§
Department of Civil and Environmental Engineering, University
of Illinois.
9
10.1021/es990153q CCC: $18.00
Published on Web 10/19/1999
1999 Am erican Chem ical Society
VOL. 33, NO. 23, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 4 2 3 9

allowed to depolymerize for several hours following dilution
(22, 23). Low concentrations of methanol (final concentra-
tions of <0.25%) and trace levels of formic acid (a decom-
position product of formaldehyde) (24) did not affect reaction
rates or interfere with any of the reactants.
A Hewlett-Packard series II 5890 gas chromatograph (GC)
equipped with a 5972 series mass selective (MS) detector
was used to identify intermediates and products in methylene
chloride extracts of reaction solutions. The temperature of
a split/ splitless injector was held at a constant value of 100
°C. A 60 m long Supelco VOCOL capillary column with 0.25
mm i.d. was used. The column oven was held at 35 °C for
6 min following injection, then increased at 2 °C/ min to 70
°C, held constant for 30 s, increased at 20 °C/ min to 120 °C,
and held at that temperature for 5 min. The flow rate of the
carrier gas, helium, was constant at 0.48 mL/ min. The de-
tector scanned for eluting compounds in the range of 10-
300 amu.
Monochloramine solutions were prepared by adding
sodium hypochlorite solution dropwise to a rapidly stirred
buffered solution of ammonium chloride until achieving a
N/ Cl molar ratio of 1.1. The pH of these stock solutions was
adjusted to a value of 9 or greater in order to minimize
disproportionation of monochloramine (25). Ammonium
chloride solutions were prepared by dissolving the salt in
water, and sodium hypochlorite solutions were prepared by
dilution of a commercially available stock solution, which
was standardized spectrophotometrically. Monochlor-
amine solutions were also standardized spectrophoto-
A Varian 3300 GC equipped with a 63Ni pulsed electron
capture detector (ECD) was used in the analysis of ClCN in
gas headspace samples. GC components included an on-
column injector, and a 30 m long, 0.53 mm i.d. Supelco
VOCOL fused silica capillary column. The temperatures of
the injector, column, and detector were constant at 100, 35,
and 290 °C, respectively. Nitrogen was used as carrier and
makeup gas at respective flow rates of 6.1 and 26
mL/ min.
An Orion model SA 720 digital pH meter equipped with
a Corning combination electrode was used for pH measure-
ments. The electrodes were calibrated so that the measured
pH values could be converted to p[H+] (p[H+] ) -log[H+]).
This was accomplished by monitoring the electrode response
during a titration of standard NaOH with perchloric acid at
an ionic strength of 0.1 M and then correlating pH measure-
ments to the known p[H+] values. For the slow decay
reactions, measurements were generally taken immediately
after the reaction started as well as in the middle of the
reaction to confirm that the pH remained constant.
Gas Chrom atography Sam ple Preparation. Headspace
samples analyzed by GC-ECD were prepared by first filling
a 17-mL vial approximately half-full with the solution of
interest. The vial was sealed liquid-tight with Teflon tape,
shaken vigorously for approximately 10 s, and allowed to
equilibrate for 1-2 min. A 450-µL headspace gas sample was
then captured with a gas-tight syringe equipped with gas-
tight valve and injected into the GC-ECD.
Samples analyzed by GC-MS were prepared by first filling
a 17-mL glass vial with screw cap and Teflon septa with 7 mL
of the sample. One milliliter of methylene chloride was added
to the vial and shaken vigorously for approximately 10 s. A
1.5 ( 0.1 µL methylene chloride sample was withdrawn with
a syringe (after rapid equilibration) and injected into the
GC-MS.
Experim ental Matrix. A summary of experimental sets
performed for this study is presented in Table 1. Experiments
were designed to assess the role of p[H+], initial formaldehyde
concentration, buffer type, and buffer concentration in the
kinetics of the initial formaldehyde addition reactions (sets
AP-1-9) and decomposition kinetics of intermediaries (sets
DP-1-7, DB-1-8, and DC-1).
metrically prior to use (λmax ) 243 nm, ꢀ243 ) 461 M-1 cm-1
(36).
)
Stock ClCN solutions were prepared gravimetrically by
discharging a small amount of pressurized ClCN gas over a
tared flask, which contained water acidified to a pH < 3 (to
minimize ClCN hydrolysis). A small cylinder containing
approximately 300 g of >99% pure ClCN (Solkatronic
Chemicals, Inc., Fairfield, NJ) was used for this purpose. Once
the weight stabilized, the flask was stoppered and the so-
lution was mixed. Other stock solutions were prepared by
dissolution of appropriate amounts of reagent-grade chemi-
cals.
Instrum ental Methods. A Durrum model 110 stopped-
flow spectrophotometer interfaced to a computer with a
MetraByte Dash-16 A/ D converter was used to study rapid
addition reactions. The absorbance was monitored at 242
nm. This wavelength gave optimum resolution for changes
in absorbance during the conversion of reactants to products.
All absorbance values were converted so that they cor-
responded to a 1 cm cell path length. The drive syringes as
well as the observation cell were thermostated at 25.0 ( 0.1
°C. For a given set of conditions, 4-10 replicate tests were
performed. Initial absorbance (A ), equilibrium absorbance
0
(A ), and observed rate constant (kobs) were estimated for
e
each test by nonlinear least-squares fitting of absorbance
data to rate expressions presented in the following section.
Values reported for these parameters were the arithmetic
average of each replicate set.
The rates of the slower decomposition reactions were
studied by monitoring absorbances at 250 nm with a Perkin-
Elmer Lambda 9 UV-Vis spectrophotometer interfaced to a
computer. In most cases, absorbance measurements were
recorded at this wavelength at a given time interval (time
drive). Repetitive scans of a given wavelength range were
also performed at a specified time interval to observe spectral
changes. Reactions were initiated by two methods. For
relatively slow reactions (e.g., p[H+] 6-8, 25 °C), an aliquot
of monochloramine was added to a rapidly stirred formal-
dehyde solution of equal volume. Under conditions where
the reaction was expected to be faster, a T-mixer was used
to provide rapid mixing of the reactants. The cell holder of
the spectrophotometer was thermostated at 25.0 ( 0.1 °C for
all experiments. The Teflon-stoppered quartz cells used were
essentially gastight and headspace-free.
Membrane introduction mass spectrometry (MIMS) analy-
ses were performed to provide evidence for the formation of
the two proposed N-chloro species. The apparatus consisted
of a Finnigan TSQ 700 mass spectrometer equipped with an
alkylated silicone membrane in the probe. The sample was
circulated through the membrane at a flow rate of 1.0 mL/
min, and the probe temperature was either ambient or 66
°C. The intensity of ions with mass-to-charge ratio (m/ z) in
the range of 60-300 amu produced by electron impact
ionization was recorded.
Results and Discussion
Reaction Sequence. A sequence (eqs 1 and 2) analogous to
that observed between formaldehyde and ammonia (27) is
proposed for the reaction of formaldehyde with monochlor-
amine:
CH2O + NH2Cl { k } CH2(OH)NHCl
(1)
(2)
2
k-2
CH2(OH)NHCl + CH2O { k } {CH2(OH)}2NCl
3
k-3
Both reactions are reversible and occur over a millisecond-
second time scale for conditions used in this study. N-
9
4 2 4 0 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 23, 1999

TABLE 1. Summary of Experiments Performed To Study Formaldehyde Addition (Sets AP-1-9) and Intermediates Decomposition
(Sets DP-1-7, DB-1-8, and DC-1) Reactionsa
exp set
p[H+]
[NH Cl]0 (M)
[CH O]T,0 (M)
type
buffer concn (M)
2
2
AP-1
AP-2
AP-3
AP-4b
AP-5
AP-6
AP-7
AP-8
AP-9
DP-1
DP-2
DP-3
DP-4
DP-5
DP-6
DP-7
DB-1
DB-2
DB-3
DB-4
DB-5
DB-6
DB-7
DB-8
DC-1
6.95
6.96
6.96
7.34
6.69-7.73
6.70
6.89
6.91
0.000965
0.000957
0.000961
0.00102
0.00103
0.00102
0.000486-0.00110
0.000365-0.00142
0.000243-0.00122
0.000779
0.000768
0.000771
0.0512-0.273
0.0338-0.203
0.203
0.0473
0.0473
0.0473
0.0270-0.101
0.0473
0.0338
0.101
phosphate
phosphate
phosphate
phosphate
phosphate
phosphate
phosphate
phosphate
phosphate
phosphate
phosphate
phosphate
phosphate
phosphate
phosphate
phosphate
borate
borate
borate
borate
borate
borate
borate
borate
carbonate
0.015
0.015
0.015
0.0140-0.0385
0.015
0.009-0.027
0.012
0.018
0.0100-0.0225
0.020
0.020
0.020
0.000472
0.000495
0.000488
0.000493
0.000488
0.000483
0.000488
0.000494
0.000491
0.000491
0.000488
0.000494
0.000486
0.000488
0.000487
0.000487
0.000497
0.000497
0.000487
7.19
2.82-5.18
5.59-6.03
5.63-6.15
5.83-5.86
5.91-5.93
7.45-7.48
7.42-7.46
8.36-8.65
8.64
9.05
8.17
8.89
8.88
0.020
0.014-0.035
0.011-0.021
0.011-0.021
0.020
0.020
0.020
0.0473
0.0270-0.101
0.0203-0.101
0.0270-0.101
0.0338
0.101
0.0338
0.020
0.015-0.030
0.015-0.030
0.012-0.030
0.012-0.030
0.020
8.44
8.42
9.23-10.23
0.101
0.0473
a
b
Tem perature ) 25 °C, ionic strength ) 0.1 M unless otherwise indicated. Ionic strength ) 0.2 M to allow for a broader total buffer concentration
range.
Chloroaminomethanol and N-chlorodimethanolamine then
decay relatively slowly to form N-chloromethanimine (eqs
3 and 4) as previously reported (28, 29):
M-1 s-1, 5.1 × 10-3 s-1, 0.088 M-1 s-1, 0.39 M-1 s-1, and 1600
M-1 s-1, respectively, at 25 °C and an ionic strength (µ) of 0.2
M (32). The equilibrium constant for eq 8 at 25 °C in the
presence of phosphate is (33)
k4
CH2(OH)NHCl
9
8 CH2NCl + H2O
(3)
(4)
[CH2O]
K1 )
) 6.31 × 10-4 + 5.1 × 10-4[H2PO4-] +
21.6 × 10-4[HPO42-] (9)
k5
[CH2(OH)2]
{CH2(OH)}2NCl
98 CH2O + CH2NCl + H2O
N-Chloromethanimine decomposes to hydrogen cyanide (28,
30) and hydrogen chloride:
Formaldehyde also tends to polymerize at relatively high
concentrations. The predominant polymeric species at the
highest concentrations investigated in this study ([CH2O]T )
0.27 M) were the dimer and trimer, which form according
to
k6
CH2NCl
9
8 HCN + HCl
(5)
The hydrogen cyanide dissociates and reaches equilibrium
with cyanide ion:
2HO(CH2O)H {K } HO(CH2O)2H + H2O
(10)
di
Ka
HCN y\
z CN- + H+
(6)
HO(CH2O)2H + HO(CH2O)H {Ktri} HO(CH2O)3H + H2O
with pKa ) 9.01 at 25 °C and µ ) 0.1 M (31). Finally, cyanide
ion reacts with monochloramine to form cyanogen chloride
and ammonia via the general acid-catalyzed reaction (16):
(11)
Kdi and K have reported values of 0.083 and 0.12 M-1
,
tri
respectively (34). Methanediol concentrations were calculated
after taking eqs 10 and 11 into consideration for all cases in
which polymer formation was appreciable ([CH2O]T g 40
mM).
k7
CN- + NH2Cl + H+ 98 ClCN + NH3
(7)
with a rate constant of 4.32 × 1010 M-2 s-1 at µ ) 1.0 M
(NaClO4) and 25 °C (16).
Formaldehyde forms a stable gem-diol, methanediol, in
aqueous solution:
UV Spectral Evidence for N-Chloroam inom ethanol and
N-Chlorodim ethanolam ine. For the rapid and reversible
formaldehyde addition reactions (eqs 1 and 2), there are three
species of interest that absorb in the UV region: NH2Cl, CH2-
(OH)NHCl, and {CH2(OH)}2NCl. These molecules absorb due
to n-π* electronic transition at the N-Cl bond. Formalde-
hyde, methanediol, and their polymeric species do not absorb
in this region. Figure 1 presents equilibrium spectra corre-
sponding to several solutions prepared with initial total
formaldehyde concentrations in the range of 0-0.11 M and
a constant initial monochloramine concentration of 0.457
mM. All solutions were buffered with phosphate at a total
CH2(OH)2 { k } CH2O + H2O
(8)
1
k-1
The equilibrium in eq 8 is shifted almost entirely to the left,
i.e., most of the total formaldehyde present in the form of
methanediol in dilute aqueous solutions (24). The dehydra-
tion reaction undergoes general acid-base catalysis; rate
constants for H+, H2O, H2PO4-, HPO42-, and OH- were 2.7
9
VOL. 33, NO. 23, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 4 2 4 1

FIGURE 1. Equilibrium spectra of solutions resulting from the
reaction of 0.457 mM monochloramine with formaldehyde at initial
concentrations ranging from 0 to 0.114 M. Additional experimental
FIGURE 2. Experimental and predicted stopped-flow traces resulting
from the reaction between monochloramine ([NH Cl]0 ) 0.961 (
2
0.004 mM) and formaldehyde ([CH O]T,0 ) (O) 0.243, (0) 0.608, (4)
2
conditions: p[H+] ) 7.84 ( 0.10, [PO ] ) 0.010 M, 25.0 ( 0.1 °C,
0.972, and (3) 1.22 mM). Additional experimental conditions: p[H+]
) 6.96 ( 0.03, [PO ] ) 0.015 M, 25.0 ( 0.1 °C, µ ) 0.1 M (NaClO )
4 T
µ ) 0.1 M (NaClO ), reaction time ) 3-5 min.
4
4 T
4
(set AP-3, Table 1).
concentration [PO4]T ≈ [H2PO4-] + [HPO42-]) ) 0.01 M, p[H+]
) 7.84 ( 0.10, µ ) 0.10 M, and 25.0 ( 0.1 °C. The curves
corresponding to relatively low methanediol concentrations
(0-4.5 mM) had isosbestic points (filled circles in Figure 1)
consistent with an equilibrium between two absorbing
compounds, proposed to be NH2Cl and CH2(OH)NHCl. The
spectra of solutions with greater total formaldehyde con-
centrations deviated from these isosbestic points, indicating
the presence of at least a third absorbing species; this is
proposed to be {CH2(OH)}2NCl. New isosbestic points were
reached (open circles in Figure 1) at methanediol concentra-
tions of 32-110 mM, consistent with an equilibrium between
CH2(OH)NHCl and {CH2(OH)}2NCl. The spectral shift pre-
sented in Figure 1 was similar to that reported for the
NH2Cl-CH3NHCl-(CH3)2NCl system (35).
Mass Spectral Evidence of Interm ediates and Products.
A monochloramine solution, prepared by adding 7.5 mM
hypochlorite to 12.5 mM ammonium chloride, was reacted
with 100 mM total formaldehyde. The reaction solution was
analyzed by MIMS within 2 h. Strong signals for ions with
m/ z values of 82 and 84 amu at the expected isotopic ratio
for chlorine revealed the presence of CH2(OH)NHCl. This
experiment also provided evidence for the existence of
{CH2(OH)}2NCl, with weaker signals at 111 and approximately
113.
Another test was performed by reacting 123 mM total
formaldehyde with 20.4 mM monochloramine at pH 8.
Methylene chloride extracts were analyzed by GC-MS after
elapsed times of 8, 127, 229, and 331 min. N-Chlorometh-
animine was identified as a rapidly forming intermediate,
which decayed slowly and was not present at significant levels
by the end of the test. ClCN was also identified; the increase
in its concentration corresponded qualitatively to the loss
of N-chloromethanimine. The observation of significant
amounts of N-chloromethanimine was somewhat unex-
pected; at pH 8, this species should not accumulate to
appreciable levels in the reaction solution. A possible
explanation for its presence is that N-chloroaminomethanol
undergoes thermally induced dehydration inside the GC-
MS injection port (100 °C). Another possibility is that
N-chloroaminomethanol dehydrates in methylene chloride.
Identification of ClCN by GC-ECD. The formation of ClCN
was also confirmed in an experiment at p[H+] of 5.6.
Additional conditions included [NH2Cl]0 ) 0.492 mM,
[CH2O]T,0 ) 47.3 mM, and [PO4]T ) 49 mM. Headspace GC-
ECD analysis revealed the presence of a large peak at an
elution time identical to that of ClCN.
Reaction between Form aldehyde and Monochloram ine.
The reaction of formaldehyde with monochloramine (eq 1)
was relatively fast. Kinetic analysis was complicated some-
what by the subsequent reaction of formaldehyde with
N-chloroaminomethanol (eq 2), which was also rapid though
somewhat slower. This interference was minimized by
performing stopped-flow experiments at relatively low
[CH2O]T,0; concentrations ranged from 0.243 to 1.42 mM (sets
AP-1-3, Table 1). Representative results obtained for ex-
perimental set AP-3 are presented in Figure 2 in terms of
decreasing absorbance at 242 nm with increasing reaction
time.
The low reactant dosages used to isolate eq 1 resulted in
relatively small changes in species distribution and ab-
sorbance. Kinetic data obtained under these conditions can
be analyzed by the chemical relaxation, or concentration
jump, method (36). The kinetic system is given by
A {kk } B + H2O
(12)
(13)
1
-1
B + C {kk } D
2
-2
where A ) CH2(OH)2, B ) CH2O, C ) NH2Cl, and D ) CH2-
(OH)NHCl. Corresponding changes in concentrations of the
species involved are
dδ
A ) -k1δA + k-1δB
(14)
dt
B ) k1δA - (k-1 + k2[C]e)δB - k2[B]eδc + k-2δD (15)
dδ
dt
dδc -dδD
)
) -k2[C]eδB - k2[B]eδC + k-2δD (16)
dt
dt
where the subscript e denotes equilibrium, and δ is the
9
4 2 4 2 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 23, 1999

TABLE 2. Summary of Rate and Equilibrium Constants Determined for the Reactions of Formaldehyde with Monochloramine and
N-Chloroaminomethanol, the Decomposition Reactions of N-Chloroaminomethanol, N-Chlorodimethanolamine, and
N-Chloromethanimine, and the Equilibrium Constant for Cyclic Borate Ester Formationa
reaction
CH2O + NH2Cl a CH2(OH)NHCl
constant
0
k2 ) (2.8 ( 0.1) × 104 M-1 s-1
HPO4
k2
) (7.5 ( 0.6) × 105 M-2 s-1
K2 ) (6.6 ( 1.5) × 105 M-1
0
CH2O +CH2(OH)NHCl a {CH2(OH)}2NCl
k3 ) (3.6 ( 0.1) × 103 M-1 s-1
HPO4
k3
) (8.2 ( 0.7) × 104 M-2 s-1
K3 ) (9.5 ( 2.5) × 104 M-1
CH2(OH)NHCl f CH2NCl + H2O
k4H ) 277 ( 7 M-1 s-1
k4OH ) 26.9 ( 5.6 M-1 s-1
k5OH ) 19.2 ( 3.7 M-1 s-1
{CH2(OH)}2NCl f CH2O + CH2NCl + H2O
CH2NCl f HCN + HCl
0
k6 ) (6.60 ( 0.20) × 10-4 s-1
k6OH ) (1.47 ( 0.19) × 106 M-1 s-1
KB ) (1.42 ( 0.46) × 10-7
B(OH)3 + {CH2(OH)}2NCl a ClN(CH2O)2B(OH)2- + H2O + H+
Conditions: µ ) 0.10 M (NaClO4), 25 °C.
a
difference between equilibrium and actual concentrations
(e.g., δA ) ([A]e - [A])). Equations 15 and 16 were obtained
after neglecting a relatively small nonlinear term with the
form k2δBδC. Under the experimental conditions used, B can
be regarded as a low-level, steady-state species (36). The
steady-state concentration jump (δB,ss) can be estimated by
setting eq 15 equal to zero:
k1δA - k2[B]eδC + k-2δD
δΒ,ss
)
(17)
k-1 + k2[C]e
Substitution of eq 17 and mass balance conditions δB ) -δA
- δD and δC ) -δD into eq 16 gives
dδD
dt
(k1 - k2[B]e - k-2
k1 + k-1 + k2[C]e
)
) -
k2[C]e + k2[B]e + k-2 δD
)
(
)
-kobsδD (18)
where kobs is the observed first-order relaxation rate constant,
and [C]e and [B]e are
k-1k-2
k1k2
1
2
FIGURE 3. Observed and fitted apparent first-order rate constants
[C]e ) [C]0 - [A]0 -
+
(
corresponding to relaxation experiments as a function of [CH -
2
(OH)2]0. Experimental conditions: [CH O]T,0 )0.243-1.42 mM, [NH Cl]0
2
2
2
k-1k-2
k1k2
k-1k-2
k1k2
) 0.961 ( 0.004 mM, p[H+] ) 6.96 ( 0.03, [PO ] ) 0.015 M, 25.0 (
[C]0 - [A]0 -
+ 4
[C]0 (19)
4 T
)
(
x
)
0.1 °C, µ ) 0.1 M (NaClO ) (sets AP-1-3, Table 1).
4
k
[B]e ) 1 ([A]0 - [C]0 + [C]e)
(20)
TABLE 3. Molar Absorptivities of Relevant Compounds
k-1
species
wavelength (nm)
E (M-1 cm-1
)
Observed absorbances, A, can be related to δD with
NH2Cl
242
250
242
250
242
250
250
250
459
429
(A - A ) ) l(ꢀC - ꢀD)δD
(21)
e
CH2(OH)NHCl
326 ( 7
345 ( 14
226 ( 15
280 ( 5
186 ( 1
313 ( 2
where A is the equilibrium absorbance, ꢀC and ꢀD are the
molar absorptivities for C and D, respectively, and l is the
light pathway length in the spectrophotometer cell. Equation
21 can be used to substitute δD for (A - A ) in eq 18, and the
resulting equation can be integrated to obtain
e
{CH2(OH)}2NCl
CH2NCl
ClN(CH2O)2B(OH)2
-
e
equilibrium constant, K2 ) k2/ k-2, are presented in Table 2.
obst
(A - A ) ) (A0 - A )e-k
(22)
The average K2 value, the equilibrium absorbance A of each
e
e
e
trace, and the molar absorptivity for monochloramine (Table
3) were used to obtain the molar absorptivity for N-
chloroaminomethanol at the experimental wavelength of 242
nm (Table 3).
The occurrence of general base catalysis was assessed by
performing experiments at [PO4]T ) 9-38.5 mM and p[H+]
) 6.69-7.73 (sets AP-4-6, Table 1). Equilibrium constants
used for H2PO4- and H2O were pKa ) 6.72 and pKw ) 13.78
both at 25 °C and µ ) 0.1 M (31). The relatively narrow p[H+]
where A is the initial absorbance.
0
Values of kobs, A , and A were determined by fitting
0
e
experimental absorbance data to eq 22 by nonlinear least-
squares regression. Resulting average kobs values were plotted
in Figure 3 against corresponding initial methanediol con-
centrations. Rate constants k2 and k-2 were then obtained
by fitting kobs values to the expression inside the parenthesis
of eq 18. Resulting values for k2 and k-2 and corresponding
9
VOL. 33, NO. 23, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 4 2 4 3

FIGURE 4. Effect of HPO 2- concentration on k2 for experimental
4
FIGURE 5. Experimental and predicted stopped-flow traces resulting
sets performed at (O) p[H+] ) 6.70 ( 0.01, [PO ] ) 0.009-0.027 M,
4 T
from the reaction between monochloramine ([NH Cl]0 ) 0.472 mM)
µ ) 0.1 M; (0) p[H+] ) 7.34 ( 0.02, [PO ] ) 0.0140-0.0385 M, µ
2
4 T
and formaldehyde ([CH (OH)2]0 ) (O) 50.8, (0) 100, and (4) 198 mM)
) 0.2 M; (4) p[H+] ) 6.69-7.73, [PO ] ) 0.015 M, µ ) 0.1 M (sets
2
4 T
(set AP-7, Table 1).
AP-4-6, Table 1) and k3 for an experimental set performed at (b)
p[H+] ) 7.19 ( 0.04, [PO ] ) 0.0100-0.0225 M, µ ) 0.1 M (NaClO );
4 T
4
The effect of methanediol concentration on k3 under these
pseudo-first-order conditions was investigated with experi-
mental sets AP-7 and AP-8 (Table 1). Typical traces obtained
are presented in Figure 5. Changes in absorbance were
monitored for overall reaction times of at least three times
that required for preequilibration of eq 1 (i.e., nearly complete
disappearance of monochloramine). Data corresponding to
reaction times after preequilibration of eq 1, or after the first
one-third of the overall duration of each experiment, could
be represented with
[CH O]T,0 ) 0.203 M, [NH Cl]0 ) 0.488 ( 0.020 mM, 25.0 ( 0.1 °C (set
2
2
AP-9, Table 1).
range and low buffer concentrations were selected to
minimize the acid-base-catalyzed decomposition of N-
chloroaminomethanol, a reaction discussed subsequently.
Equation 22 was used to analyze the data for each trace, with
k-2 ) k2/ K2 and K2 equal to the average value determined
from the relaxation experiments (Table 2). Data analysis
revealed that HPO42- was the only base with an appreciable
catalytic effect under the conditions investigated. The buffer
dependence of k2 (see Figure 4) could therefore be repre-
sented with
dδD -dδ
)
E ) -(k3[B]0 + k-3)δD ) -kobsδD (25)
dt
dt
HPO4
2-
k2 ) k20 + k2
[HPO4
]
(23)
where [B]0 is the constant concentration of formaldehyde,
and kobs is the apparent first-order rate constant. Equation
25 could now be converted to an expression with the form
of eq 22. Average kobs values, obtained by nonlinear least-
squares regression, are presented in Figure 6 as a function
of corresponding methanediol concentration. The linear
dependence of kobs on [CH2(OH)2]0 for each set was consistent
with eq 25. Differences between the two lines revealed that
base catalysis was affecting the kinetics of eq 2. Rate constants
k3 and k-3 were estimated for each linear plot, and the
corresponding average K3 value is presented in Table 2.
The buffer dependence of k3 was investigated with
experimental set AP-9 (Table 1). Parallel to the finding for eq
Base catalysis constants estimated by linear regression of
the k2 data shown in Figure 4 are presented in Table 2. The
HPO
4
reverse rate constants, k-20 and k-2
, can be calculated by
dividing the respective forward rate constants by K2. Within
the relatively narrow p[H+] range and at the reaction times
studied, the slow, acid-base-catalyzed decomposition of
N-chloroaminomethanol did not interfere with the deter-
mination of k2 and k-2
.
The equilibrium and rate constants presented in Table 2
for eq 1 were used to predict all absorbance traces performed
at relatively low total formaldehyde concentrations. As
illustrated in Figure 2, good agreement was generally observed
between experimental and predicted traces.
Reaction between Form aldehyde and N-Chloroam ino-
m ethanol. The reaction of formaldehyde with N-chloro-
aminomethanol (eq 2) was analyzed by performing tests with
methanediol in large excess compared to monochloramine.
Under these conditions and after N-chloroaminomethanol
reached peak concentration, eqs 1 and 8 can be assumed to
be in rapid preequilibrium as compared to eq 2. The system
could thus be represented with eqs 12 and 13 at preequi-
librium as
2-
1, HPO4 was the only base found to affect the kinetics of
eq 2 under the range of experimental conditions investigated.
The buffer dependence of k3 is depicted in Figure 4. The
slope and intercept of this plot were used to estimate the
0
HPO
4
corresponding k3 and k3
values (Table 2) with an
expression of the form of eq 23. The reverse rate constants,
0
HPO
4
k-3 and k-3
, can be calculated by dividing the corre-
sponding forward rate constants by K3.
The molar absorptivity for N-chlorodimethanolamine at
242 nm (Table 3) was estimated using A and K3 (Table 2)
e
values and the N-chloroaminomethanol molar absorptivity
(Table 3).
The kinetic and equilibrium constants presented in Table
2 were applied to predict complete stopped-flow photometry
traces at relatively high [CH2O]T,0. A kinetic model including
rate expressions for the five species involved in eqs 1, 2, and
B + D {kk } a E
(24)
3
-3
with B ) CH2O, D ) CH2(OH)NHCl, and E ) {CH2(OH)}2-
NCl.
9
4 2 4 4 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 23, 1999

8 was solved by a finite difference approach. As illustrated
in Figure 5, the model represented the experimental data
adequately.
Decom position of N-Chloroam inom ethanol and N-
Chlorodim ethanolam ine. The formaldehyde addition reac-
tions (eqs 1 and 2) practically reached equilibrium prior to
a relatively slower decrease in absorbance resulting from the
decompositions of N-chloroaminomethanol and N-chloro-
dimethanolamine. Four of the species involved in eqs 3-7
absorb light at 250 nm: monochloramine, N-chloroamino-
methanol, N-chlorodimethanolamine, and N-chlorometh-
animine. It is subsequently demonstrated that, with the
exception of tests performed at p[H+] < 5.5, the rates of eqs
3 and 4 were much slower than that of eq 5. Accordingly, the
overall decay rate for N-chloro species could be expressed
as
dC
-
T,NCl ) k4[CH2(OH)NHCl] + k5[{CH2(OH)}2NCl]
dt
(26)
with
FIGURE 6. Observed and predicted apparent first-order rate
constants corresponding to two experimental sets performed at
CT,NCl ) [NH2Cl] + [CH2(OH)NHCl] + [{CH2(OH)}2NCl]
relatively high [CH O]T,0 and different buffer concentrations. Ex-
2
(27)
perimental conditions: (b) [CH O]T,0 ) 51.2-273 mM, p[H+] ) 6.89
2
( 0.06, [PO ] ) 0.012 M, [NH Cl]0 ) 0.472 mM; (9) [CH O]T,0
)
4 T
2
2
Equation 26 is valid only if the reaction of cyanide ion with
monochloramine (eq 7) is much slower than eqs 3 and 4.
The validity of this assumption is addressed in a later section.
The concentration of each N-chloro species can be
expressed as a fraction of CT,NCl by combining the equilibrium
expressions for eqs 1 and 2 with eq 27:
33.8-203 mM, p[H+] ) 6.91 ( 0.02, [PO ] ) 0.018 M, [NH Cl]0 )
4 T
2
0.495 mM; 25.0 ( 0.1 °C, and µ ) 0.1 M (NaClO ) for both sets (sets
4
AP-7 and AP-8, Table 1).
The initial decay rates for traces corresponding to most
experimental sets were first order:
dA
dt
[NH2Cl]
-
) kobsA
(33)
) R0 )
CT,NCl
1
where A is absorbance and kobs is the observed first-order
rate constant.
Substitution of eq 32 into eq 33 gives an expression of the
form of eq 31 with
(28)
1 + K1K2[CH2(OH)2] + K12K2K3[CH2(OH)2]2
[CH2(OH)NHCl]
) R1 )
CT,NCl
kobs ) R1(k4 + k5K3K1[CH2(OH)2])
(34)
K1K2[CH2(OH)2]
(29)
Values for equilibrium constants K1, K2, and K3 (Table 2) can
be applied to calculate the fractions of N-chloro compounds
for the range of experimental conditions investigated with
phosphate buffer. The estimated fractions ranged from R0 )
0.033, R1 ) 0.369, and R2 ) 0.597 at [CH2O]T,0 ) 0.0270 M to
R0 ) 0.003, R1 ) 0.143, and R2 ) 0.854 at [CH2O]T,0 ) 0.101
M.
1 + K1K2[CH2(OH)2] + K12K2K3[CH2(OH)2]2
[{CH2(OH)}2NCl]
) R2 )
CT,NCl
K12K2K3[CH2(OH)2]2
(30)
1 + K1K2[CH2(OH)2] + K12K2K3[CH2(OH)2]2
Rate Constants. Absorbance traces obtained for set DC-1
(Table 1) are shown in Figure 7. First-order reactions (eq 33)
were observed in all four tests. Similarly, the initial decay
rates for traces corresponding to most other experimental
sets were also first order. A tendency of the apparent rate to
deviate from initial first order, which was more pronounced
near neutral p[H+], and 3.5 e p[H+] e 5.5 will be addressed
subsequently. kobs values were estimated from the initial slope
of absorbance traces for all experiments performed at
[NH2Cl]0 ) 0.5 mM, [CH2O]T,0 ) 47.3 mM, and buffer
concentration of 20 mM (phosphate, borate, or carbonate).
Initial apparent first-order decay rates obtained at 2.8 <p[H+]
< 3.5 and 5.5 < p[H+] < 10.2 are plotted against the
corresponding p[H+] in Figure 8. As discussed previously,
the rate of absorbance decay had a tendency to deviate from
first order as reactions proceeded. The main reason for this
deviation at p[H+] > 5 was that the rate of monochloramine
decomposition according to eq 7 became comparable to those
for eqs 1-4 as the concentration of cyanide resulting from
eqs 5 and 6 increased with time. The initial reaction time
Substitution of eqs 29 and 30 into eq 26 results in the following
rate expression:
dC
-
T,NCl ) R1(k4 + k5K3K1[CH2(OH)2])CT,NCl (31)
dt
The total absorbance of the solution can be expressed as
a function of CT,NCl
:
A ) (R0ꢀ0 + R1ꢀ1 + R2ꢀ2)CT,NCl
(32)
ꢀ0, ꢀ1, and ꢀ2 are the molar absorptivities of monochloramine,
N-chloroaminomethanol, and N-chlorodimethanolamine at
250 nm. Molar absorptivities (Table 3) were estimated by
fitting the initial absorbance of all traces performed with
phosphate buffer at p[H+] > 5.5 with eq 32.
9
VOL. 33, NO. 23, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 4 2 4 5

FIGURE 9. Effect of methanediol concentration on the initial apparent
first-order rate constant for tests performed at (O) [CH O]T,0 ) 47-
2
100 mM, total phosphate buffer concentration of 20 mM, p[H+] )
FIGURE 7. Experimental traces observed at [NH Cl]0 ) 0.5 mM,
2
5.8, (NaClO ) (set DP-4, Table 1), and (2) [CH O]T,0 ) 20-100 mM,
4
2
[CH O]T,0 ) 47 mM, p[H+] ) 9.2-10.2, carbonate buffer concentration
total borate buffer concentration of 20 mM, p[H+] ) 9.05 (set DB-3,
2
of 0.02 M, 25.0 ( 0.1 °C, and µ) 0.1 M (NaClO ) (set DC-1, Table
4
Table 1).
1).
SCHEME 1
with slope k5K3K1 and intercept k4. As depicted in Figure 9
for set DP-4 data (Table 1), eq 3 was responsible for most of
the decomposition of N-chloro species at p[H+] ) 5.8. On
the basis of the criteria defined previously for interference
by eq 7, first-order rate constants could only be determined
for tests at high formaldehyde concentrations within this
set. The results are consistent with acid-catalyzed dehydration
of N-chloroaminomethanol to N-chloromethanimine. As
shown in Scheme 1, the hydroxyl group of N-chloroamino-
methanol is susceptible to protonation. Acid-catalyzed
dehydration of N-chlorodimethanolamine would not be
expected because the nitrogen atom has no proton to lose.
Also, Antelo and co-workers performed a study on the
decomposition rate of a similar compound, N-chlorodieth-
anolamine, and found that base catalysis and not acid
catalysis was the predominant decay pathway (37). Conse-
quently, eq 33 simplifies to kobs ) R1k4 for traces run at 5.5
< p[H+] < 6.5.
FIGURE 8. Effect of pH on the initial apparent first-order rate constant
for tests performed at [NH Cl]0 ) 0.5, [CH O]T,0 ) 47.3 mM, total
2
2
buffer concentration of 20 mM, 25.0 ( 0.1 °C and µ ) 0.1 M (NaClO ).
4
Continuous regression lines correspond to the phosphate data at
p[H+] ) 5.5-6.5 and carbonate data at p[H+] ) 9.2-10.2. Dashed
line is a representation of k6 based on k60 and k6OH values in Table
2.
Two acids, H2PO4- and H3O+, were checked for contribu-
tion to k4 within this p[H+] range. Analysis of the data for
experimental set DP-5 (Table 1) revealed that the effect of
range for which the rate remained nearly first order was
estimated for each trace by setting the criteria that the rate
of eq 7 be generally less than 5% of the overall apparent rate.
A trial-and-error approach was followed for this purpose using
the rate constants reported by Schurter et al. (16) for eq 7.
Inspection of Figure 8 reveals the occurrence of both acid
and base catalysis for p[H+] > 5.5. kobs reached an apparent
constant value at p[H+] ≈ 3 due to a shift in limiting reactant
from N-chloroaminomethanol and N-chlorodimethanol-
amine to N-chloromethanimine as discussed in a subsequent
section. First-order kinetics were not observed at 3.5 < p[H+]
< 5.5.
-
H2PO4 was negligible. Consequently, the overall rate con-
stant k4 can be expressed as
k4 ) k4H[H+]
(35)
Values of kobs obtained from trace data from sets DP-2-5
that met the criteria of lack of interference by eq 7 were used
to determine the contribution of H+ to the overall acid
catalysis of eq 3. Resulting kobs/ R1 values are plotted against
H
5.5 < p[H+] < 6.5. According to eq 34, a plot of kobs/ R1
values against [CH2(OH)2] ≈ [CH2(OH)2]0 should be linear
[H+] in Figure 10. The rate constant k4 (Table 2) was
estimated from the slope of the regression line in Figure 10.
9
4 2 4 6 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 23, 1999

8 <p[H+]<9. The decomposition of N-chloro compounds
in this p[H+] range was investigated with borate-buffered
solutions (Table 1). In general, experimental kobs values
tended to be lower than expected. This deviation was more
pronounced at relatively high values of buffer concentration,
p[H+], and [CH2O]T,0. Boric acid, known to form cyclic borate
esters with 1,2-diols (38), probably reacted with N-chloro-
dimethanolamine. The following reaction is proposed to
account for this effect:
B(OH)3 + {CH2(OH)}2NCl {K
B
\}
ClN{CH2O}2B(OH)2- + H2O + H+ (36)
with KB as the reaction equilibrium constant. The total
concentration of N-chloro compounds is
CT,NCl ) [NH2Cl] + [CH2(OH)NHCl] +
[{CH2(OH)}2NCl] + [ClN{CH2O}2B(OH)2-] (37)
and the reaction rate expression is
FIGURE 10. Effect of H+ concentration on the initial apparent first-
dC
-
T,NCl ) â1(k4 + k5K3K1[CH2(OH)2])CT,NCl ) kobsCT,NCl
order rate constant for tests performed at [NH Cl]0 ) 0.5 mM, [CH O]T,0
2
2
dt
) 47 mM, p[H+] ) 5.8-6.2, total phosphate buffer concentration of
(38)
20 mM, 25.0 ( 0.1 °C, and µ ) 0.1 M (NaClO ) (sets DP-2-5, Table
4
1).
with â1 substituting R1 (see eq 34) as the new fraction for
N-chloroaminomethanol:
SCHEME 2
â1 ) K1K2[CH2(OH)2]/ {1 + K1K2[CH2(OH)2] +
K12K2K3[CH2(OH)2]2(1 + KB[B(OH)3]/ [H]+)} (39)
Equationss 38 and 39 are based on the assumptions that eq
36 is reversible and in rapid equilibrium with respect to eqs
3 and 4. Also, notice that according to eq 39, â1 would decrease
with increasing values of both p[H+] and [CH2O]T,0. This effect
would be consistent with the observed trend in experimental
kobs values as a function of both p[H+] (Figure 8) and [CH2O]T,0
.
Experimental kobs values for data from sets DB-1-8 (Table
1) were fitted with eqs 38 and 39. For this purpose, k4 and
k5 were assumed to include the following components:
SCHEME 3
k4 ) k4H[H+} + k4OH[OH-]
k5 ) k5OH[OH-]
(40)
(41)
H
with k4 being the value determined previously (Table 2).
Equations 40 and 41 were based on the assumption that the
borate/ boric acid buffer had a negligible catalytic effect on
kobs over the range of experimental conditions tested.
Estimated fitting parameters KB, k4OH, and k5OH are presented
in Table 2. The dissociation constant used for boric acid was
pKa ) 8.97 at 25 °C and µ ) 0.1 M (31). On the basis of KB
) 1.42 × 10-7, the fraction of total N-chloro compounds
occurring as the proposed cyclic borate ester are estimated
to range from 18% at p[H+] ) 8.17 and [CH2O]T,0 ) 0.0270
M to 55% at p[H+] ) 9.05 and [CH2O]T,0 ) 0.101 M.
Results for experimental set DB-3 (Table 1) are plotted in
Figure 9 according to eq 38. The increase in both slope and
intercept of the linear regressions with increasing p[H+]
confirmed that both k4 and k5 included base catalysis
components in accordance with eqs 40 and 41. N-Chloro-
aminomethanol and N-chlorodimethanolamine are both
expected to undergo base catalysis since proton removal
should occur in each case (Schemes 2 and 3).
in which â0, â1, â2, and â3 are the molar fractions of
monochloramine, N-chloroaminomethanol, N-chlorodimeth-
anolamine, and cyclic borate ester; ꢀ0, ꢀ1, ꢀ2, and ꢀ3 are the
corresponding molar absorptivities at 250 nm. The molar
absorptivity of the cyclic borate ester (Table 3) was estimated
by fitting the initial absorbance of all traces performed with
borate buffer with eq 42 using the ꢀ0, ꢀ1, and ꢀ2 values
previously determined for phosphate buffer solutions (Table
3).
The total absorbance of the solution can be expressed as
p[H+] > 9 and p[H+] ) 7.4. The validity of k4 and k5
OH
OH
a function of CT,NCl
:
values estimated with borate buffer data (Table 2) was
checked by comparing to experimental results obtained for
set DC-1 with carbonate buffer and for set DP-7 with
A ) (â0ꢀ0 + â1ꢀ1 + â2ꢀ2 + â3ꢀ3)CT,NCl
(42)
9
VOL. 33, NO. 23, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 4 2 4 7

The decomposition of N-chloromethanimine (eq 5) is
proposed as the rate-limiting reaction at p[H+] < 3. The
presence of N-chloromethanimine was supported by a shift
in the spectrum of the reaction solutions at this pH range.
The analysis of experimental traces at 3.0 < p[H+] < 5.5
required taking into consideration the simultaneous de-
composition of N-chloroaminomethane and N-chlorodimeth-
anolamine (eqs 3 and 4) and the formation and decompo-
sition of N-chloromethanimine (eq 5) or
d[CH2NCl]
) R1k4(CT,NCl - [CH2NCl]) - k6[CH2NCl]
dt
(43)
dC
-
T,NCl ) k6[CH2NCl]
(44)
dt
where k6 is the rate constant for N-chloromethanimine
decomposition, R1 and k4 are defined by eqs 29 and 40, and
the total concentration of N-chloro compounds is given by
CT,NCl ) [NH2Cl] + [CH2(OH)NHCl] +
[{CH2(OH)}2NCl] + [CH2NCl] (45)
FIGURE 11. Comparison of experimental and predicted initial
apparent first-order rate constants for tests performed at [NH Cl]0
2
) 0.5 mM, [CH O]T,0 ) 47 mM, p[H+] ) 9.2-10.2, carbonate buffer
2
Integration of eqs 43 and 44 resulted in the following
analytical solutions:
concentration of 0.02 M, 25.0 ( 0.1 °C, and µ ) 0.1 M (NaClO ) (set
4
DC-1, Table 1). Solid line corresponds to predicted values and is
not a regression.
CT,NCl ) a1eλ t + a2eλ t
(46)
(47)
1
2
1
[CH2NCl] ) - (a1λ1eλ t +a2λ2eλ
)
t
2
1
k6
with
λ2
a1 )
a2 )
CT,NCl,o
(48)
(49)
λ2 - λ1
λ1
CT,NCl,o
λ1 - λ2
-(R k + k ) ( (R k + k )2 - 4R k k
x
1
4
6
1
4
6
1
4
6
λ1,λ2 )
(50)
2
Finally, total solution absorbance can be expressed as a
function of CT,NCl and [CH2NCl]:
A ) (R0ꢀ0 + R1ꢀ1 + R2ꢀ2)(CT,NCl - [CH2NCl]) +
ꢀ4[CH2NCl] (51)
FIGURE 12. Comparison of experimental and predicted absorbance
where ꢀ0, ꢀ1, ꢀ2, and ꢀ4 are the molar absorptivities of
monochloramine, N-chloroaminomethanol, N-chlorodimeth-
anolamine, and N-chloromethanimine at 250 nm. The molar
absorptivity for N-chloromethanimine, estimated by fitting
the initial absorbance of the trace performed with phosphate
buffer at p[H+] ) 2.8 (set DP-1) with eq 51, is also presented
in Table 3.
Equations 46-51, with the k4 value from Table 2, were
used to fit traces for experimental set DP-1. A trial and error
analysis revealed that eq 5 was base-catalyzed according to
traces at [NH Cl]0 ) 0.5 mM, [CH O]T,0 ) 47 mM, p[H+] ) 2.8-5.2,
2
2
phosphate buffer concentration of 0.02 M, 25.0 ( 0.1 °C, and µ )
0.1 M (NaClO ) (set DP-1, Table 1).
4
phosphate buffer (Table 1). The comparison for set DC-1 is
depicted in Figure 11. Very good agreement was found
between predicted and experimental values. It should be
noted that the line presented in the figure is not a regression
of the data but rather that predicted with eqs 34, 40, and 41
H
OH
OH
using the k4 and k5 values in Table 2.
The apparent rate constant estimated with eqs 34, 40,
and 41 using k4H, k4OH, and k5 values in Table 2 for the
k6 ) k60 + k6OH[OH-]
(52)
OH
conditions of experimental set DP-7, kobs ) 1.08 × 10-5 s-1
,
0
OH
was within 10% of the average initial rate constant value kobs
) 0.99 × 10-5 s-1 obtained for the experimental traces. These
results also confirmed the assumptions made in eqs 40 and
41 that water-assisted rate terms were negligible as compared
to the acid and base catalysis terms.
where k6 and k6 are the rate constants for water and
0
hydroxide ion catalysis. Values of k6 and k6OH, calculated
from fitting of the experimental traces, are presented in Table
2.
Good agreement was found between fitted curves and
experimental traces as depicted in Figure 12. The deviation
observed at initial reaction times might have resulted from
p[H+] < 5.5. Experimental results observed for set DP-1
(Table 1) were consistent with a shift in rate-limiting reaction.
9
4 2 4 8 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 23, 1999

(12) Lowry, T. H.; Richardson, K. S. Mechanism and Theory in Organic
Chemistry, 3rd ed.; Harper & Row: New York, 1987; pp 661-
709.
SCHEME 4
(13) March, J. Advanced Organic Chemistry, 4th ed.; Wiley: New
York, 1992; pp 796-800.
(14) Sollenberger, P. Y.; Martin, R. B. In The Chemistry of the Amino
Group; Patai, S., Ed.; Interscience: New York, 1968; pp 367-
380.
(15) Wade, L. G., Jr. Organic Chemistry; Prentice Hall: Englewood
Cliffs, NJ, 1987; pp 804-806.
(16) Schurter, L. M.; Bachelor, P. P.; Margerum, D. W. Environ. Sci.
Technol. 1995, 29, 1127-1134.
(17) Ohya, T.; Kanno, S. Chemosphere 1989, 19, 1835-1842
(18) Hirose, Y.; Maeda, N.; Ohya, T.; Nojima, K.; Kanno, S. Chemo-
sphere 1988, 17, 865-873.
neglecting the contribution by eq 7 to the overall rate of
N-chloro compounds decomposition.
(19) Maeda, N.; Ohya, T.; Nojima, K.; Kanno, S. Chemosphere 1987,
16, 2249-2258.
(20) Hirose, Y.; Maeda, N.; Ohya, T.; Nojima, K.; Kanno, S. Chemo-
sphere 1989, 18, 2101-2108.
The proposed mechanism for base-catalyzed decomposi-
tion of N-chloromethanimine is depicted in Scheme 4. Proton
removal is required for the formation of HCN. The rates of
water and hydroxide assisted decay pathways are significantly
higher for N-chloromethanimine than those for the decay of
N-chloroaminomethanol and N-chlorodimethanolamine
(see Figure 8). Consequently, N-chloromethanimine should
not accumulate to significant levels at p[H+] > 5.5.
(21) Sawicki, E.; Sawicki, C. R. Aldehydes-Photometric Analysis;
Academic: New York, 1975; p 210.
(22) Hasse, H.; Maurer, G. Ind. Eng. Chem. Res. 1991, 30, 2195-
2200.
(23) Hahnenstein, I.; Albert, M.; Hasse, H.; Kreiter, C. G.; Maurer, G.
Ind. Eng. Chem. Res. 1995, 34, 440-450.
(24) Walker, J. F. Formaldehyde, 3rd ed.; Reinhold: New York, 1964;
pp 54-107.
(25) Gray, E. T., Jr.; Margerum, D. W.; Huffman, R. P. In Organometals
and Organometalloids, Occurrence and Fate in the Environment;
Brinckman, F. E., Bellama, J. M., Eds.; ACS Symposium Series
82; American Chemical Society: Washington, DC, 1978; pp 264-
277.
(26) Kumar, K.; Day, R. A.; Margerum, D. W. Inorg. Chem. 1986, 25,
4344-4350.
(27) Kawasaki, A.; Ogata, Y. Mem. Fac. Eng., Nagoya Univ. 1967, 19,
1-45.
Acknowledgments
This work was supported by Purdue University, U.S. Depart-
ment of Education Graduate Assistantships in Areas of
National Needs Program, and National Science Foundation
Grant CHE 9622683. We wish to thank N. Srinivasan and R.
Graham Cooks for their assistance with MIMS analyses.
Literature Cited
(1) Krasner, S. W.; McGuire, M. J.; Jacangelo, J. G.; Patania, N. L.;
Reagan, K. M.; Aieta, E. M. J. Am. Water Works Assoc. 1989, 80,
41-53.
(2) U.S. EPA. Fed. Regist. 1996, 61, 24353-24388.
(3) World Health Organization. Guidelines for Drinking Water
Quality, 2nd ed.; World Health Organization: Geneva, 1993;
Vol. I.
(4) Pedersen, E. J., III; Marin˜ as, B. J. Proceedings of the 1995 American
Water Works Association Annual Conference, Water Research,
Anaheim, CA, June 18-22, 1995; American Water Works
Association: Denver, 1995; pp 359-369.
(5) Pedersen, E. J., III; Urbansky, E. T.; Marin˜ as, B. J.; Margerum,
D. W. Preprints of Papers Presented at the 210th American
Chemical Society National Meeting, Division of Environmental
Chemistry, Chicago, IL, August 20-24, 1995; American Chemical
Society: Washington, DC, 1995; Vol. 35, No. 2, pp 710-713.
(6) Weinberg, H. S.; Glaze, W. H.; Krasner, S. W.; Sclimenti, M. J.
J. Am. Water Works Assoc. 1993, 85, 72-85.
(7) Glaze, W. H.; Koga, M.; Cancilla, D. Environ. Sci. Technol. 1989,
23, 838-847.
(8) Richardson. In Encyclopedia of Environmental Analysis and
Remediation; Meyers, R., Ed.; Wiley: New York, 1998; p 1398.
(9) Hand, V. C.; Snyder, M. P.; Margerum, D. W. J. Am. Chem. Soc.
1983, 105, 4022-4025.
(28) Cross, C. F.; Bevan, E. J.; Bacon, W. J. Chem. Soc. Trans. 1910,
97, 2404-2406.
(29) Lindsay, M.; Soper, F. G. Proc. Chem. Soc., London 1946, 791-
792.
(30) Le Cloirec, C.; Martin, G. Water Chlorination; Jolley, R. L., Bull,
R. J., Davis, W. P., Katz, S., Roberts, M. H., Jr., Jacobs, V. A., Eds.;
Lewis Publishers: Chelsea, MI, 1985; pp 821-834.
(31) Smith, R. M.; Martell, A. E. Critical Stability Constants. Volume
4: Inorganic Complexes; Plenum Press: New York, 1976.
(32) Bell, R. P.; Evans, P. G. Proc. R. Soc., Ser. A 1966, 291, 297-323.
(33) Los, J. M.; Roeleveld, L. F.; Wetsema, B. J. C. J. Electroanal. Chem.
1977, 75, 819-837.
(34) Skell, P.; Suhr, H. Chem. Ber. 1961, 94, 3317-3327.
(35) Ferriol, M.; Gazet, J.; Rizk-Ouaini R. Anal. Chim. Acta 1990, 231,
161-163.
(36) Bernasconi, C. F. Relaxation Kinetics; Academic: New York,
1976; pp 3-10, 20-31.
(37) Antelo, J. M.; Arce, F.; Casal, D.; Rodriguez, P.; Varela, A.
Tetrahedron 1989, 45 (12), 3955-3966.
(38) Shriver, D. F.; Atkins, P. W.; Langford, C. H. Inorganic Chemistry;
W. H. Freeman and Company: New York, 1990; p 336.
(10) Cocivera, M.; Fyfe, C. A.; Effio, A.; Vaish, S. P.; Chen, H. E. J. Am.
Chem. Soc. 1976, 98, 1573-1578.
(11) Jencks, W. P. In Physical Organic Chemistry, Vol. 2; Cohen, S.
G., Streitwieser, A., Jr., Taft, R. W., Eds.; Interscience: New York,
1964; pp 63-122.
Received for review February 10, 1999. Revised manuscript
received September 4, 1999. Accepted September 10, 1999.
ES990153Q
9
VOL. 33, NO. 23, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 4 2 4 9