Formation of iodo-trihalomethanes during disinfection and oxidation of iodide-containing waters

Article abstract of DOI:10.1021/es9914590

The formation of iodo-trihalomethanes (I-THMs) such as iodoform (CHI₃) during oxidative treatment of iodide-containing drinking waters can be responsible for taste and odor problems. I-THMs are formed by reactions of hypoiodous acid (HOI) with natural organic matter. HOI is quickly formed from naturally occurring iodide (I^-) by oxidation with ozone, chlorine, or chloramine. The kinetics of reactions of HOI with organic model compounds as well as the resulting CHI₃ formation were measured. Substituted phenols, phenol, and, to a smaller extent, α-methyl carbonyl compounds were found to be reactive toward HOI and: also to yield CHI₃. Resorcinol (m-hydroxyphenol) had the highest yield of CHI₃. The kinetics of I-THM formation were also measured in natural waters which were oxidatively treated with ozone, chlorine, or chloramine. When ozone was used, no I-THMs were detected and ≥ 90% of I^- was transformed to IO₃^-. Chlorine led to the formation of both IO₃^- and I-THMs. With increasing chlorine doses, the CHI₃ formation decreased, whereas IO₃^- formation, as well as the formation of classical THMs such as chloroform, increased. In chloramination processes, I-THMs (especialliiy CHI₃)were the main products. The CHI₃ formation in the oxidation of natural waters increased in the order O₃ < Cl₂ < NH₂Cl. The formation of iodo-trihalomethanes (I-THMs) such as iodoform (CHI₃) during oxidative treatment of iodide-containing drinking waters can be responsible for taste and odor problems. I-THMs are formed by reactions of hypoiodous acid (HOI) with natural organic matter. HOI is quickly formed from naturally occurring iodide (I^-) by oxidation with ozone, chlorine, or chloramine. The kinetics of reactions of HOI with organic model compounds as well as the resulting CHI₃ formation were measured. Substituted phenols, phenol, and, to a smaller extent, α-methyl carbonyl compounds were found to be reactive toward HOI and also to yield CHI₃. Resorcinol (m-hydroxyphenol) had the highest yield of CHI₃. The kinetics of I-THM formation were also measured in natural waters which were oxidatively treated with ozone, chlorine, or chloramine. When ozone was used, no I-THMs were detected and ≥90% of I^- was transformed to IO₃^-. Chlorine led to the formation of both IO₃^- and I-THMs. With increasing chlorine doses, the CHI₃ formation decreased, whereas IO₃^- formation, as well as the formation of classical THMs such as chloroform, increased. In chloramination processes, I-THMs (especially CHI₃) were the main products. The CHI₃ formation in the oxidation of natural waters increased in the order O₃ < Cl₂ < NH₂Cl.

Full text of DOI:10.1021/es9914590

Environ. Sci. Technol. 2000, 34, 2784-2791
waters near the sea coast or under special geological
Formation of Iodo-Trihalomethanes
during Disinfection and Oxidation of
Iodide-Containing Waters
circumstances (11). Therefore, the formation of CHI3 above
its organoleptic threshold concentration from naturally
occurring iodide (I-) is possible during oxidative drinking
water treatment.
I-THMs were detected in several drinking waters. In one
case, a change from chlorination to chloramination led to
the formation of 5 µg/ L CHI3 from an I- concentration of 50
µg/ L (8). In another water, up to 30 µg/ L CHI3 and only low
concentrations of the classical THMs were formed from 200
µg/ L I- when chloramine was utilized (12). When chlorine
was used to treat the same water, only classical THMs were
observed; no CHI3 was detected. Another study reports the
appearance of CHI3 after the chlorination of a water with a
high I- concentration (150-200 µg/ L) (6). The same study
describes the formation of I-THMs in an ozonation process
during failure of the ozonation step (90 µg/ L I-). Another
study showed that I-THMs were formed in several waters
during chlorination although the iodide concentration in
the raw water was orders of magnitude lower than the
bromide concentration (13).
YV E S B I C H S E L A N D U R S V O N G U N T E N *
Swiss Federal Institute for Environmental Science and
Technology EAWAG, CH-8600 Du¨ bendorf, Switzerland
The formation of iodo-trihalomethanes (I-THMs) such as
iodoform (CHI3) during oxidative treatment of iodide-containing
drinking waters can be responsible for taste and odor
problems. I-THMs are formed by reactions of hypoiodous
acid (HOI) with natural organic matter. HOI is quickly
formed from naturally occurring iodide (I-) by oxidation
with ozone, chlorine, or chloramine. The kinetics of reactions
of HOI with organic model compounds as well as the
resulting CHI3 formation were measured. Substituted phenols,
phenol, and, to a smaller extent, R-methyl carbonyl
compounds were found to be reactive toward HOI and
also to yield CHI3. Resorcinol (m-hydroxyphenol) had the
highest yield of CHI3. The kinetics of I-THM formation were
also measured in natural waters which were oxidatively
treated with ozone, chlorine, or chloramine. When ozone
was used, no I-THMs were detected and g90% of I- was
I- is quickly oxidized to hypoiodous acid (HOI) in the
presence of ozone, chlorine, or chloramine (14). The further
reaction of HOI with natural organic matter (NOM) can result
in the formation of I-THMs. In the present study, we
investigated the kinetics of the reaction of HOI with organic
compounds and the formation kinetics of I-THMs in both
natural and model waters. The relative reaction rates of the
oxidation of HOI to IO3- (14), the disproportionation of HOI
(15), and its addition to organic compounds determine the
transformed to IO -. Chlorine led to the formation of
-
3
product distribution between IO3 and iodoorganic com-
both IO - and I-THMs. With increasing chlorine doses,
pounds such as I-THMs. The formation of IO3- is probably
not a toxicological problem because ingested IO3- is quickly
reduced to I- in vivo by glutathione (16).
3
the CHI3 formation decreased, whereas IO - formation, as
3
well as the formation of classical THMs such as
chloroform, increased. In chloramination processes,
I-THMs (especially CHI3) were the main products. The
CHI3 formation in the oxidation of natural waters increased
in the order O < Cl2 < NH Cl.
To describe the reactivity of HOI toward organic com-
pounds and oxidants, its pH-dependent speciation has to be
considered. HOI (pKa ) 10.4) can be deprotonated to
hypoiodite (OI-) (14, 15). At a low pH, H2OI+ (pKa ) 1.4 (
0.3 (17, 18)) can be formed by protonation of HOI.
3
2
Materials and Methods
Introduction
With the exception of pinacolone, which was redistilled, all
About 25 years ago, the formation of chlorine- and bromine-
containing trihalomethanes (THMs: chloroform (CHCl3),
bromodichloromethane (CHBrCl2), dibromochloromethane
(CHBr2Cl), and bromoform (CHBr3)) during drinking water
chlorination was discovered (1-3). Their toxicity led to
stringent regulations, e.g., in the European Union (4) and in
the United States (5). In this study, these compounds will be
referred to as “classical THMs”.
chemicals were used without further purification and were
of the highest purity grade. HOI was freshly produced through
oxidation of I- by HOCl from a NaOCl stock solution (Aldrich)
of approximately 0.68 M prior to experiments because it is
not stable in water (disproportionation). The exact [OCl-]
-
could be determined in excess of I- as I3 (ꢀ288
) 38 200
nm
M-1 cm-1) (19). The pH measurements were carried out with
a Ross electrode (ATI Orion, Boston, MA) and a Metrohm
632 pH-meter (Metrohm, Herisau, Switzerland) which was
calibrated with standard buffer solutions (Merck). Spectro-
photometric measurements were performed on an Uvikon
940 spectrophotometer (Kontron Instruments, Eching, Ger-
many).
In iodide-containing waters, the formation of six ad-
ditional THMs, the iodo-trihalomethanes (I-THMs), is pos-
sible through incorporation of one or more iodine atoms
into a THM. The resulting compounds are iodoform (CHI3),
chlorodiiodomethane (CHClI2), bromodiiodomethane (CH-
BrI2), dichloroiodomethane (CHCl2I), dibromoiodomethane
(CHBr2I), and bromochloroiodomethane (CHBrClI). In the
late 1980s I-THMs (especially CHI3) were found to be
responsible for the occurrence of bad taste and odor in
drinking waters (6-8). The organoleptic threshold concen-
tration of CHI3 lies between 0.03 and 1 µg/ L, which is the
lowest value of all of the I-THMs (9, 10). Total iodine
concentrations in water resources are usually in the range
of 0.5-20 µg/ L but can exceed 50 µg/ L in certain ground-
Consum ption of HOI by Organic Model Com pounds.
The kinetics of the consumption of HOI by organic model
compounds were measured in double-distilled water in 250-
mL batch reactors (for slow kinetics) and in a continuous-
flow apparatus (for fast kinetics) under pseudo-first-order
conditions ([model compound] . [HOI]) at 25 ( 2 °C. HOI
was prepared by mixing I- and OCl- in a volumetric flask
and diluting it to a concentration of 1-4 µM. Between pH
3 and 11, pH was controlled by 0.5-5 mM of phosphate
buffer. For pH < 3, only sulfuric acid was used, whereas for
pH > 11, only NaOH was used. More experimental details
are presented later in Tables1-3.
* Corresponding author phone: +41 1 823 52 70; fax: +41 1 823
52 10; e-mail: vongunten@eawag.ch.
9
2 7 8 4 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 13, 2000
10.1021/es9914590 CCC: $19.00
2000 Am erican Chem ical Society
Published on Web 06/07/2000

TABLE 1. Rate Constants for the Reaction of Phenols with HOI
observed rate constants (M-1 s-1
k3
)
phosphate
(mM)
phenol
(µM)
[HOI](t ) 0)
(µM)
compound
pK
a
pH range
nc
k2
k4
p-m ethoxyphenol 10.20a 3.1-4.7
1
10-20
10
10-500
10-200
10-200
10-1000
1.1-3.5
0.8-2.6
1.7-5.0
1.2-2.3
0.8-5.0
1.1-2.3
11
8
27
14
16
18
(5 ( 1) × 104
(5 ( 2) × 104
(3 ( 1) × 103
(4 ( 1) × 103
(2 ( 1) × 102
< 100
(7 ( 2) × 108
p-cresol
10.26a 2.5-4.9
9.99a 3.1-8.6
9.43a 1.0-7.7
9.20a 0.5-8.2
7.86b 1.9-8.4
0-1
2.5-5
0-1
0-1
0-1
(4.0 ( 1.5) × 104 (3 ( 1) × 102
(7 ( 3) × 108
phenol
(1.0 ( 0.3) × 102
< 5
(2 ( 1) × 106
p-chlorophenol
p-iodophenol
p-cyanophenol
(1.6 ( 0.5) × 105
(1.5 ( 0.8) × 105
(4 ( 2) × 102
(20 ( 8)
(1.5 ( 0.5)
a
b
c
Reference 33. This work. n ) num ber of kinetic experim ents ([HOI] vs t).
The continuous-flow apparatus consisted of three me-
chanically driven glass syringes (Dosimat 665, Metrohm,
Herisau, Switzerland) which pumped a solution of HOI and
phosphate, a solution of the model compound, and an iodide
solution (0.45 M KI), respectively. All three solutions were
pumped with the same flux of 5 mL/ min. The HOI solution
and the model compound solution were mixed in a mixing
tee. Thereafter, they passed a capillary tube of well-defined
volume. The I- solution was added through a second mixing
tee at the end of the capillary tube (observation point). This
experimental setup was similar to the one applied in ref 20.
I- stopped the reaction and allowed at the same time
methanolic solutions at -18 °C. Standard aqueous solutions
of 10 µg/ L I-THMs in ambered flasks were stable for at least
two weeks at room temperature (stable peak areas) in the
pH range 6-9.
IO3 was analyzed with ion chromatography and post-
column reaction with UV-vis detection with a detection limit
of 0.1 µg/ L (0.6 nM) in natural waters (19).
Form ation of I-THMs and IO3- in Natural Waters. Two
natural waters (from the Seine River, Paris, France, and Lake
Zurich, Zurich, Switzerland) were spiked with varying
amounts of I- and oxidatively treated with chloramine,
chlorine, or ozone at 25 ( 2 °C. The dissolved organic carbon
(DOC) concentration was 1.3 mg/ L in the lake water and 3.5
mg/ L in the river water, [Br-] was 30 µg/ L in the river water
and 15 µg/ L in the lake water, and [ammonia] was below 6
µg N/ L in both waters. The pH was controlled by addition
of 10 mM B(OH)3 and adjusted to 8.0 by NaOH (1 M, 10.8
M). The solutions were transferred to headspace free bottles
immediately after the addition of the oxidant. The following
parameters were analyzed during the reaction time (24-35
h): HOI, I-THMs, IO3-, classical THMs, oxidant, and pH.
-
detection of the residual HOI in a 5-cm flow-through
-
photometric cell by formation of I3 (ꢀ351
) 25 700 M-1
nm
cm-1 (19)) as shown in eq 1 (kHOI+I ) 5 × 109 M-1 s-1 (21),
-
9
kI +I ) 6.2 × 10 M-1 s-1(22)). The second absorption maxi-
-
2
mum at 351 nm was chosen in these experiments because
of interferences of organic model compounds at 288 nm.
HOI + 2 I- + H+ ) I3- + H2O
(1)
The reaction time of the model compound with HOI was
defined by the volume of the capillary tube and the flux of
the mixed solution. It varied from 0.8 to 2.7 s. The pH was
measured at the observation point.
Because of the interference of oxidants (NH2Cl, Cl2, and
O3), it was not possible to analyze HOI by the photometric
method as described above. HOI was quenched with phenols
and analyzed as iodophenol by HPLC. An aliquot of 20 µM
phenol was added to an aliquot of the reaction solution to
transform HOI to iodophenol. Phenol reacts quickly and
quantitatively with HOI (see below). Both p-iodophenol and
o-iodophenol were formed in this reaction. Because io-
dophenols react about 10 times slower with HOI than phenols
do (see Table 1) and because phenol was present in high
excess, it can be assumed that no di- or triiodophenols were
formed. The two iodophenols were quantified relative to a
p-iodophenol standard solution by assuming that both
iodophenols had the same sensitivity. The [HOI] was
calculated from their sum. The HPLC separation was done
on a Nucleosil 103-5 C18 column (Macherey-Nagel, Du¨ ren,
Germany) with an eluent consisting of 65% methanol, 34.9%
water, and 0.1% acetic acid (retention times 5.5 min for
o-iodophenol and 6.5 min for p-iodophenol). UV detection
at 231 nm yielded a detection limit of 0.3 µg I/ L.
The decrease of HOI in batch experiments was measured
by placing 25-mL aliquots of the reaction solution into a
10-cm photometric cell together with 5 mL of 1 M KI. I-
stopped the reaction immediately according to eq 1. Detec-
tion at 351 nm resulted in a detection limit for HOI of 0.4 µM.
Form ation of CHI3 in Model Solutions. The kinetics of
the formation of CHI3 in model solutions were investigated
in Nanopure water (Barnstead B-pure system) in batch
reactors at 25 ( 2 °C. The pH was controlled by 1 mM
phosphate buffer (pH ) 7.0) and 1 mM borate buffer (pH )
9.0). The initial concentrations were 1 µM for the model
compound and 6 µM for HOI. This excess of HOI allowed
iodine addition over several steps which eventually led to
CHI3. In addition, a residual HOI concentration could be
guaranteed for the duration of the experiment (32 h). HOI,
CHI3, and IO3- were analyzed in these experiments (HOI by
photometry).
Analysis of the I-THMs was performed with liquid-liquid
extraction into methyl tert-butyl ether (Fluka, “for residue
analysis”, g99.8%, Buchs, Switzerland) and by gas chromo-
tography/ electron-capture detection (GC/ ECD) according
to ref 23. Chromatographic separation was performed on a
DB-5 column (30 m × 0.25 mm i.d., 0.25 µm film thickness,
J&W Scientific, Folsom, CA). The temperature program was
as follows: 35 °C (5 min), heating rate 2 °C/ min to 45 °C (0
min), heating rate 15 °C/ min to 220 °C (3 min). The following
compounds were detected by this method: CHCl2I (limit of
detection (LOD) 1.3 µg/ L (S/ N ) 3)), CHBrClI (LOD 0.9 µg/
L), CHBr2I (LOD 0.6 µg/ L), CHClI2 (LOD 0.3 µg/ L), CHBrI2
(LOD 0.07 µg/ L), and CHI3 (LOD 0.1 µg/ L). With the exception
of CHI3, I-THM standards are not commercially available.
For this study, standards were supplied by F. Ventura, Societat
General d’Aigu¨ es de Barcelona (AGBAR), and stored as
The four classical THMs (CHCl3, CHBrCl2, CHBr2Cl, and
CHBr3) were analyzed by headspace GC/ ECD. Detection
limits (S/ N ) 3) were 0.06 µg/ L for CHBrCl2, 0.1 µg/ L for
CHBr2Cl, 0.3 µg/ L for CHCl3, and 0.5 µg/ L for CHBr3.
Oxidant concentrations were determined by the indigo
method (for ozone) (24) or the ABTS method (for chlorine
and chloramine) (25).
Results and Discussion
Model Com pounds: Consum ption of HOI and Form ation
of CHI3. Several organic model compounds have been tested
for their reactivity toward HOI and formation of CHI3 in these
reactions. The two main classes which have been investigated
are phenols and R-methyl carbonyl compounds. These
compounds are present in substructures of NOM and are
known to react with halogens (26).
9
VOL. 34, NO. 13, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 2 7 8 5

FIGURE 1. pH dependence of the observed second-order rate constants of the reaction of HOI with some phenols. Symbols, experimental
results; lines, results calculated according to eqs 2-4. Conditions: [phenol] ) 10-1000 µM, [HOI] ) 0.8-5.0 µM.
Consumption of HOI by Phenols. The kinetics of the
reactions of HOI with six para-substituted phenols were
investigated under pseudo-first-order conditions ([phenol]
. [HOI]) in a batch and in a continuous-flow system.
Experimental details (pH, buffer concentrations, concentra-
tions of phenols, initial HOI concentrations, and numbers
of experiments) and the calculated rate constants are shown
in Table 1. The kinetics were always first-order in phenol
and first-order in HOI. The rate constants were determined
by a least-squares fit for eqs 2-4.
is the major species at pH ) 5, the reaction HOI + phenolate
dominates the system because of the much higher reaction
rate constant. This behavior results in a 10-fold increase of
the overall rate constant with every pH unit until the pKa of
the phenol is reached. All determined rate constants k2, k3,
and k4 are shown in Table 1. They cover a range of almost
9 orders of magnitude (from 1.5 to 7 × 108 M-1 s-1). A rep-
resentation of the measured rate constants with respect to
the Hammett coefficients, σp, of the substituents shows good
linearity (27). This indicates that the electron density on the
aromatic system, which is strongly influenced by the sub-
stituent, is a key factor for the reaction rate of the phenols
with HOI. This subject is discussed in further detail elsewhere
(28).
k2
H2OI+ + (substituted)phenol
98 products
(2)
k3
HOI + (substituted)phenol
9
8 products
(3)
(4)
The reactions of HOI with substituted phenols are
significantly faster than the corresponding reactions of HOCl.
p-Cresol reacts 2000-20 000 times faster with HOI than with
HOCl (k3HOCl ) 0.14 M-1 s-1; k4HOCl ) 30 000 M-1 s-1) (29). The
difference between HOCl and HOI is smaller for the more
electrophilic phenols such as p-cyanophenol (factor 4-60:
k4
HOI + (substituted)phenolate
98 products
Figure 1 depicts the pH-dependence of the observed
second-order rate constants for a highly reactive phenol (p-
methoxyphenol), a moderately reactive phenol (p-iodophe-
nol), and a less reactive phenol (p-cyanophenol). At low pH
values, the reaction H2OI+ + phenol is dominant (eq 2). The
results for p-iodophenol confirm the pKa value of 1.4 ( 0.3
for H2OI+ (see Figure 1) (17, 18). At pH < 3.5, the observed
second-order rate constant increases with decreasing pH
due to the shift of HOI to H2OI+. The H2OI+ is expected to
have a higher reactivity as a result of its higher electrophilic
character. This increase stops at pH < 1, where [H2OI+] does
not further increase. However, the rate constant continues
to increase at pH < 1 for p-cyanophenol. Other unknown
reactions might be responsible for this behavior. At higher
pH values (3.5-5), the observed second-order rate constants
go through a minimum. In this pH range, reaction 3 involving
HOI and the phenol becomes important for p-cresol, phenol,
p-iodophenol, and p-cyanophenol. For p-chlorophenol and
p-methoxyphenol, only reactions 2 and 4 contributed to the
overall reaction. Reaction 3 was much slower than reactions
2 and 4. It is therefore only possible to indicate a maximum
value for k3. At high pH values (pH > 5), the overall reaction
is dominated by the phenolate (reaction 4). Although phenol
) 0.025 M-1 s-1; k4
) 90 M-1 s-1) (29). HOI can
HOCl
HOCl
k3
therefore compete with HOCl for the reaction with phenols
even if HOCl is present in high excess relative to HOI.
Formation of CHI3 from Phenols. The formation of CHI3
was measured at pH values of 7 and 9 with an initial
concentration of 6 µM HOI and 1 µM phenol or resorcinol.
Both compounds resulted in CHI3 formation but resorcinol
yielded much more CHI3 (normalized to HOI exposure) than
phenol did. For resorcinol, 4-7% of the C atoms were
incorporated into CHI3 after 6 h. This corresponds to 0.25-
0.4 mol of CHI3 per mol of resorcinol. The HOI consumption
of both compounds occurred mainly in the first 2 h and
amounted to 2 mol of HOI per mol of phenol and 5 mol of
HOI per mol of resorcinol.
To be able to compare the results of the model solutions
with natural waters (see below), the CHI3 yield (q) with respect
to the exposure to HOI (∫[HOI] dt) and the initial [DOC] (in
mol/ L C) was calculated according to eq 5. This yield is a
measure for [CHI3] which was formed as a result of a certain
HOI exposure on a certain [DOC].
9
2 7 8 6 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 13, 2000

together with the fitted pH dependence of the observed first-
order rate constant (lines). Acetaldehyde is the most reactive
of these compounds, whereas pinacolone is the least reactive.
[CHI3]
q )
or
[HOI] dt[DOC]
∫
The iodination rate increases with increasing pH because
of enhanced OH- catalysis. For a high enough pH, the
enolization becomes faster than the addition of HOI to the
enol. In this case, the HOI addition to the enol becomes the
rate-determining step. This results in a first-order behavior
in HOI. For acetone, this was observed at pH > 9.1. From the
dependence of the observed second-order rate constant of
HOI plus acetone on the pH (9.1 < pH < 11.6), it can be
concluded that two reactions significantly contribute to the
overall decrease of HOI, HOI plus enol and OI- plus enolate
(eq 8).
[CHI3]
[DOC]
) q [HOI]dt (5)
∫
Figure 2 shows a representation of the experimental data
expressed in terms of eq 5. CHI3 yield, q, can be read as the
slope of a linear fit of the data. For phenol (triangles) this
value is lower than 0.002 M-1 s-1 at pH ) 7. At pH ) 9, no
CHI3 formation could be detected from phenol. For resorcinol
(circles) q reaches a maximum value of 1.1 M-1 s-1 during
the first 6 h (∫[HOI] dt < 0.05 Ms). This steep increase
corresponds to a CHI3 formation of, e.g., 15 nM h-1or 6 µg
L-1 h-1 in a solution of 10 µg L-1 of HOI and 1 mg L-1 of
resorcinol.
Consumption of HOI by R-Methyl Carbonyl Compounds.
The consumption of HOI by R-methyl carbonyl compounds
was investigated in batch reactors under pseudo-first-order
conditions. Since 1822, the reaction of iodine with these
compounds has been known as the “iodoform reaction” (30).
A reaction scheme is presented in eq 6.
d[HOI]
-
) k5[HOI][acetone] + k6[OI-][OH-][acetone]
dt
(8)
Formation of CHI3 from R-Methyl Carbonyl Compounds.
The formation of CHI3 was measured at pH ) 7 and 9 with
initial concentrations of 6 µM HOI and 1 µM acetone or
methylglyoxal (CH3COCHO). At pH ) 7, acetone yielded no
CHI3 and methylglyoxal resulted in a small CHI3 formation.
At pH ) 9, higher CHI3 concentrations were found. In Figure
2, the experimental data are plotted according to eq 5. At pH
) 9, the CHI3 formation yield, q, is in the range of 0.05-0.1
M-1 s-1 for both acetone (squares) and methylglyoxal
(diamonds). At pH ) 7, q is below 0.001 M-1 s-1 for acetone
and about 0.005 M-1 s-1 for methylglyoxal. For acetone, CHI3
formation is about 25% of the value which was expected if
HOI consumption was rate-determining. The observed
difference could be due to a retardation of the overall reaction
by the final hydrolysis of triiodoacetone. This gives an
indication that the hydrolysis is the rate-determining step in
CHI3 formation from acetone and HOI. Another experiment
which was performed in excess of acetone over HOI showed
that the second iodination step of acetone is about 30 times
faster than the first iodination step at pH ) 9. The third
iodination step is at least 30 times faster than the second
step (28).
According to eq 6, HOI does not react directly with the
carbonyl compound but only with the enol. The formation
of the enol (enolization) can be catalyzed by several species:
H+, H2O, and OH-. The halogenation of an R-methyl carbonyl
compound proceeds three times until the methyl group has
been transformed into a triiodomethyl group. The tri-
iodocarbonyl compound then hydrolyzes with the formation
of CHI3 and the corresponding acid (RCOOH).
For the first halogenation step, either the enolization or
the addition of HOI to the enol is rate-determining. If the
enolization is rate-determining, the overall reaction rate
depends only on the carbonyl concentration and the pH,
but not on [HOI], and the rate law becomes zero-order in
HOI (eq 7). It was demonstrated previously that under certain
conditions this was the case because the rate constants of
iodination and bromination reactions were equal (31). This
shows that the halogen does not participate in the rate-
determining step under the corresponding conditions.
At the pH values investigated in this study, OH- and H2O
were the main catalysts for the enolization. The rate of HOI
consumption can be formulated as
Consumption of HOI by Other Organic Compounds.
Besides phenols and carbonyl compounds, other organic
model compounds (allyl alcohol, glucosamine, glycine,
toluene, and oxalate) were tested for their reactivity toward
HOI. Table 3 shows that all of these compounds are
unreactive. Rate constants were below 1 M-1 s-1 or even
below 0.1 M-1 s-1
.
-
Consum ption of HOI and Form ation of THMs and IO3
during Disinfection/Oxidation of Natural Waters. The
-
kinetics of HOI consumption and the formation of IO3
,
I-THMs, and classical THMs were measured during the first
24 h of the oxidation of two natural waters (Lake Zurich,
Switzerland, and Seine River, France) spiked with I-.
Figure 4 shows the formation of IO3-, CHI3, and total I in
I-THMs together with the decrease of HOI and chlorine during
chlorination of the lake water (pH ) 8.0, 0.5 mg/ L Cl2, 400
nM/ 50 µg/ L I-). For the calculation of the parameter I in
I-THMs, the concentrations of the individual I-THMs were
multiplied by the number of iodine atoms they contain. With
this value the percentage of iodine which was incorporated
d[HOI]
dt
-
) kH O[ketone] + kOH-[OH-][ketone] (7)
2
into the I-THMs can be compared to the percentage of iodine
Experimental results for the reaction of HOI with acetone
(at pH <8.3), pinacolone, and acetaldehyde could be modeled
with eq 7. Acetate was also investigated but no reaction could
be detected. The calculated rate constants are presented in
-
incorporated into IO3
.
More than 90% of HOI had disappeared after 30 min.
However, the I-THMs, especially CHI3, were not released
immediately. Half of their final concentration was reached
only after 2 h. As speculated above, the hydrolysis of the
trihalomethyl group could be the rate-determining process
in I-THM formation. Fifty percent of the final IO3- concen-
Table 2. In the literature, a value of 0.52 M-1 s-1 (kOH ) is given
-
for the hydroxide-catalyzed enolization of acetone (32). This
compares quite well with our findings (kOH ) 0.25 M-1 s-1).
-
In Figure 3, the experimental data (symbols) are shown
9
VOL. 34, NO. 13, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 2 7 8 7

FIGURE 2. Formation of CHI3 from model compounds and in natural waters (data transformed according to eq 5). Model compounds:
resorcinol (circles), acetone (open squares), methylglyoxal (open diamonds), and phenol (open triangles). Conditions: pH ) 7 (dashed
lines) and pH ) 9 (solid lines). Inset: Natural waters from Lake Zurich, 1 mg/L NH Cl (full triangles) and 0.5 mg/L NH Cl (full diamonds);
2
2
Seine river, 0.5 mg/L NH Cl (full squares). Conditions: pH ) 8.
2
TABLE 2. Rate Constants for the Reaction of r-Methyl Carbonyl Compounds with HOI
compound
phosphate
(mM)
r-methyl carbonyl
[HOI](t ) 0)
(µM)
CH COR
R
pH range
6.8-8.6
4.8-8.9
8.3-9.9
6.0-8.3
9.1-11.6
(mM)
na
10
14
5
observed rate constants
3
acetaldehyde
pinacolone
acetate
H
1-2
1-2
1
0.1-10
1-100
2.2-2.9
2.0-4.0
2.0-2.3
1.7-5.5
1.3-3.8
kH ) (9 ( 1) × 10-8 s-1
;
2O
-1
kOH ) (0.90 ( 0.04) M s-1
-
tert-butyl
O-
k
) (4 ( 2) × 10-9 s-1
;
H2
O
-
kOH ) (0.037 ( 0.002) M s-1
-1
100-1000
31-64
k
< 2 × 10-10 s-1
;
H2
O
-
-6
kOH < 3 × 10 M-1 s-1
acetone
CH3
1
9
k
k
) (1.3 ( 0.4) × 10-8 s-1
;
H2
O
-
) (0.25 ( 0.03) M s-1
-1
OH
acetone
CH3
0-1
2.0-8.4
9
k5 ) (2.5 ( 1.0) M-1 s-1
;
k6 ) (3 ( 1) × 103 M-2 s-1
a
n ) num ber of kinetic experim ents ([HOI] vs t).
tration was reached after 1 h. The formation of all iodine-
containing disinfection byproducts was finished after 10 h.
In additional experiments in lake and river water, some
of the reaction parameters (oxidant, oxidant dose, and [I-])
were changed. Figure 5 shows the final concentrations (24
h after the oxidant addition) of IO3-, CHI3, the total I in
I-THMs, and the total concentrations of the six I-THMs and
the four classical THMs. The highest CHI3 formation (16 nM
CHI3) was observed in the lake water when using chloramine
(Figure 5, experiments C and D) where 47 nM I, or 12% of
the initial I- (400 nM), was incorporated into CHI3. The
compounds CHI3, CHClI2, CHCl2I, and CHBrClI were found
in several experiments, but neither CHBr2I nor CHBrI2 was
ever detected. The concentration of Br- in the investigated
waters was too low to lead to a significant incorporation of
bromine into the THMs. CHBrClI was the only bromine-
containing I-THM which was detected, with a maximum
concentration of 24 nM in the chlorinated river water
(experiment G).
differences in the formation of iodine-containing disinfection
byproducts when comparing these two oxidants. First,
-
chloramine is not capable of oxidizing HOI to IO3 but
-
chlorine is capable of this (14). This explains the higher IO3
formation in chlorination processes compared to chlorami-
nation processes (Figure 5; experiments B, D). As a conse-
quence, HOI has a longer lifetime in chloramination processes
which leads to an increased reaction time with NOM. Second,
the competition between chlorine and iodine for incorpora-
tion into the THMs is much higher in the presence of chlorine
than in the presence of chloramine. This shifts the product
distribution from CHI3 in chloramination processes to the
mixed I-THMs (e.g., CHCl2I) in chlorination processes. In
the investigated lake water, chloramination resulted in 16
nM CHI3 and 34 nM mixed I-THMs (experiment C). Under
the same conditions, chlorine induced the formation of 3
nM CHI3 and 120 nM mixed I-THMs (experiment A). When
ozone was applied (experiment E), IO3- was the only iodine-
containing byproduct detected. However, I- was not quan-
-
Byproduct Formation during Chloramination, Chlorina-
tion, and Ozonation. The smaller chemical reactivity of
chloramine compared to that of chlorine leads to two general
titatively transformed to IO3 despite the fast oxidation of
HOI by O3 (t1/ 2 < 1 s) (14). Organic compounds which are
highly reactive toward HOI successfully competed with O3.
9
2 7 8 8 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 13, 2000

FIGURE 3. Observed first-order rate constants for the iodination of three r-methyl carbonyl compounds. Symbols, experimental data; lines,
fits according to eq 7.
TABLE 3. Rate Constants for the Reaction of Model Compounds with HOI
phosphate
(mM)
model
compounds (mM)
[HOI](t ) 0)
(µM)
observed
rate constants
compound
pK
a
pH range
nc
allyl alcohol
glucosam ine
glycine
7.2-8.2
5.5-9.2
6.1-9.0
6.0-8.1
4.1-7.3
0.5-1
0.5-10
3-50
2-1000
10-200
2-4
1.5-2.0
1.8-2.5
0.6-2.5
1.7-2.2
1.0-2.9
6
11
10
7
kobs e 1 M-1 s-1
kobs < 1 M-1 s-1
kobs < 0.1 M-1 s-1
kobs < 0.1 M-1 s-1
kobs < 1 M-1 s-1
7.58a
1
9.78b
0-1
1
1
oxalate
1.25, 4.27b
toluene
5
a
b
c
This work. Reference 33. n ) num ber of kinetic experim ents ([HOI] vs t).
Such competitors could be phenolic compounds which have
a higher apparent rate constant with HOI than O3 has.
Chlorine Dose. For a chlorine dose of 0.5 mg/ L, 4%
(experiment F) and 26% (experiment A) of the iodine was
threshold value and the sum of the classical THMs is below
the current drinking water standards (4, 5).
I- Concentration. At [I-] ) 10 µg/ L, formation of mixed
I-THMs but no CHI3 formation was observed. The competi-
tion between HOI and HOCl was shifted in favor of HOCl
under these conditions. Therefore, a potential problem with
CHI3 must be expected only for [I-] . 10 µg/ L.
-
transformed to IO3 in the river water and the lake water,
respectively. With an increased chlorine dose of 1 mg/ L, 43%
(experiment G) and 75% (experiment B) of the total iodine
-
was found as IO3-. In the river water (higher [DOC]), the IO3
Incorporation of Iodine and Chlorine into THMs. Because
of the high excess of HOCl (7 µM/ 0.5 mg/ L Cl2) over HOI (0.4
µM) in experiment F (∫[HOCl] dt : ∫[HOI] dt ) 15), iodine
has to react much faster with THM precursors than chlorine
does in order to produce I-THMs. This was obviously the
case because approximately the same amounts of iodine and
chlorine were incorporated into THMs (both classical and
I-THMs) in this experiment. If HOI reacted 15 times faster
than HOCl did with the THM precursor groups, the HOCl
excess would be compensated and a chlorine and iodine
incorporation in equal amounts (as observed) would be
expected. As shown above, the reaction of HOI with phenols
can be 4-20 000 times faster than the corresponding reaction
of HOCl. Therefore, such functionalities might be responsible
for our observations.
yield was lower than that in the lake water. Due to the higher
[DOC], the reaction of HOI with the NOM was more important
and chlorine was less stable in the river water. As a result of
this, less HOI was oxidized to IO3-. The distribution among
the I-THMs was also affected by the chlorine dose. In the
river water, 8 nM CHI3 and 54 nM mixed I-THMs were
detected for 0.5 mg/ L Cl2 (experiment F). At a higher chlorine
dose of 1 mg/ L, [CHI3] decreased to 1 nM and the mixed
I-THMs increased to 110 nM (experiment G). In the lake
water, 0.5 mg/ L Cl2 resulted in 3 nM CHI3 and 120 nM mixed
I-THMs (experiment A), whereas for 1 mg/ L chlorine
(experiment B), no CHI3 but 46 nM mixed I-THMs were
formed. The lower yield of mixed I-THMs at the higher
-
chlorine dose in the lake water is a result of the high IO3
formation (75% of total I). For the prevention of medicinal
taste and odor caused by CHI3, an increased chlorine dose
can be a reasonable solution. A drawback of that is the
enhanced formation of the classical THMs. The increase of
the chlorine dose from 0.5 to 1 mg/ L led to an increase in
the sum of the classical THMs from 45 nM (5.8 µg/ L) to 210
nM (29 µg/ L) in the river water (experiments F, G). However,
the overall result is still positive because CHI3 is below the
Apparent Rate Constant HOI + NOM. In the experiments
with chloramine as the oxidant, HOI only reacted with the
NOM but was not further oxidized by NH2Cl. In both the lake
and the river waters, the half-life of HOI was about 5 h. A rate
constant for the reaction of HOI with NOM was fitted for
both first-order and zero-order behaviors in HOI. On a
C-atom base for the NOM, the rate constants amount to 0.4
M-1 s-1 or 6 × 10-6 s-1 in the lake water and 0.1 M-1 s-1 or
9
VOL. 34, NO. 13, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 2 7 8 9

FIGURE 4. Consumption of HOI and chlorine together with the formation of IO -, CHI3, and the sum of I in I-THMs. Conditions: initially
3
400 nM I- (50 µg/L), 15 µg/L Br-, 1.3 mg/L DOC, 0.5 mg/L Cl2, pH ) 8.0.
FIGURE 5. Formation of disinfection byproducts in the oxidation of natural waters (24 h after addition of the oxidant). Conditions: pH )
8.0, I- ) 400 nM (50 µg/L). I in THMs (column 3) is the concentration of iodine in THMs (concentration of I-THMs multiplied by the number
of iodine atoms they contain). Column 4 is the total concentration of I-THMs. The concentrations of the classical THMs were not measured
in experiments B-E.
9 × 10-6 s-1 in the river water. These values describe the
reaction of HOI with NOM1 which leads to iodoorganic
compounds such as I-THMs (eq 9).
However, this reaction has no influence on the kinetics of
the HOI decrease because the formed I- is quickly reoxidized
to HOI by chloramine. The above-mentioned rate constants
can be compared with the rate constants of HOI with model
compounds at pH ) 8 (on a C-atom base: phenols 3 × 101-6
× 105 M-1 s-1, carbonyl compounds 6 × 10-9-5 × 10-7 M-1
s-1). A comparison of rate constants from natural waters and
model compounds shows that phenolic entities (but not
HOI + NOM1 ) I - NOM
(9)
HOI + NOM2 ) I- + NOM2,ox
(10)
According to eq 10, HOI can also be reduced to I- by NOM2.
9
2 7 9 0 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 13, 2000

Acknowledgments
TABLE 4. Semiquantitative Assessment of the Sinks of Iodine
during Disinfection: Influence of Disinfectant
Financial support from Suez Lyonnaise des Eaux for Y.B. is
acknowledged. We are grateful to F. Ventura, Societat General
d’Aigu¨ es de Barcelona (AGBAR), who supplied the standards
of the mixed I-THMs, to E. Salhi for her technical support,
and to H. Gallard for the review of the manuscript.
products
disinfectant
iodoform
other I-THMs
iodate
ozone
chlorine
chloram ine
-
+
+++
-
++
++
+++
++
-
Literature Cited
(1) Bellar, T. A.; Lichtenberg, J. J.; Kroner, R. C. J.-Am. Water Works
carbonyl entities) as part of the NOM could account for the
experimental results.
Assoc. 1974, 66 (12), 703-706.
(2) Kleopfer, R. D.; Fairless, B. J. Environ. Sci. Technol. 1972, 6 (12),
1036-1037.
Yield of CHI3. CHI3 formation yield, q, according to eq 5,
was also calculated for the chlorination of the natural waters
(see Figure 2, inset). During chloramination processes, CHI3
formation faced less competition from the formation of mixed
I-THMs. In the lake water, q reached a value of 0.12 M-1 s-1
for 0.5 mg/ L NH2Cl (diamonds in Figure 2, inset) and 0.24
M-1 s-1 for 1.0 mg/ L NH2Cl (triangles in Figure 2, inset).
From the investigated model compounds, only resorcinol
yielded enough CHI3 to account for the results from the lake
water. For the river water (squares in inset Figure 2), q was
0.002-0.035 M-1 s-1. The highest values were found at the
beginning of the experiment. The observed q values could
be explained principally by methylglyoxal or acetone, but
only if they accounted for more than 35% of the NOM.
Because this percentage is too high for NOM, it is more likely
that phenolic groups in the NOM account for the observed
CHI3 formation in the river water. Here again, phenolic groups
as part of the NOM could explain the experimental results
whereas carbonyl compounds could not.
(3) Rook, J. J. Water Treat. Exam. 1974, 23 (2), 234-243.
(4) Amtsblatt der Europa¨ischen Gemeinschaften 1998; European
Union, Nov 3, 1998. (Gazette of the European Union).
(5) USEPA (United States Environmental Protection Agency).
Federal Register Part IV, 1998.
(6) Bruchet, A.; N’Guyen, K.; Mallevialle, J.; Anselme, C. Proceedings
of the 1989 AWWA Seminar on Identification and Treatment of
Taste and Odor Compounds, Los Angeles, CA, 1989; pp 125-
141.
(7) Gittelman, T. S.; Yohe, T. L. Proceedings of the 1989 AWWA
Seminar on Identification and Treatment of Taste and Odor
Compounds, Los Angeles, CA, 1989; pp 105-141.
(8) Hansson, R. C.; Henderson, M. J.; Jack, P.; Taylor, R. D. Water
Res. 1987, 21 (10), 1265-1271.
(9) Khiari, D. Distribution Generated Taste and Odor Phenomena;
American Water Works Association Research Foundation:
Denver, CO (report in progress).
(10) Suez Lyonnaise des Eaux. Conditions de Formation et Destruction
des Trihalome´thanes Iode´s Responsables des Gouˆ ts et Odeurs
Pharmaceutiques; Internal Report; Lyonnaise des Eaux: Le Pecq,
France, 1993.
(11) Fuge, R.; Johnson, C. C. Environ. Geochem. Health 1986, 8 (2),
31-54.
(12) Karpel Vel Leitner, N.; Vessella, J.; Dore´, M.; Legube, B. Environ.
Sci. Technol. 1998, 32 (11), 1680-1685.
(13) Symons, J. M.; Xia, R.; Speitel, G. E., Jr.; Diehl, A. C.; Hwang, C.
J.; Krasner, S. W.; Barrett, S. E. Report on Factors Affecting
Disinfection By-Product Formation During Chlorination. Ameri-
can Water Works Association Research Foundation: Denver,
CO, 1998; Project No. 803.
(14) Bichsel, Y.; von Gunten, U. Environ. Sci. Technol. 1999, 33 (22),
4040-4045.
(15) Bichsel, Y.; von Gunten, U. Water Res. 2000, in press.
(16) Taurog, A.; Howells, E. M.; Nachimson, H. I. J. Biol. Chem. 1966,
241 (20), 4686-4693.
(17) Bell, R. P.; Gelles, E. J. Chem. Soc. 1951, 2734-2740.
(18) Burger, J. D.; Liebhafsky, H. A. Anal. Chem. 1973, 45 (3), 600-
602.
(19) Bichsel, Y.; von Gunten, U. Anal. Chem. 1999, 71 (1), 34-38.
(20) Hunt, N. K.; Marin˜ as, B. J. Water Res. 1997, 31 (6), 1355-1362.
(21) Eigen, M.; Kustin, K. J. Am. Chem. Soc. 1962, 84, 1355-1361.
(22) Turner, D. H.; Flynn, G. W.; Sutin, N.; Beitz, J. V. J. Am. Chem.
Soc. 1972, 94 (5), 1554-1559.
(23) Cancho, B.; Ventura, F.; Galceran, M. T. J. Chromatogr., A 1999,
841, 197-206.
(24) Bader, H.; Hoigne´, J. Water Res. 1981, 15, 449-456.
(25) Pinkernell, U.; Nowack, B.; Gallard, H.; von Gunten, U. Water
Res., in press.
(26) Vollhardt, K. P. C.; Schore, N. E. Organic Chemistry; W. H.
Freeman: New York, 1994.
(27) Johnson, C. D. The Hammett Equation; Cambridge University
Press: Cambridge, U.K., 1980.
Operational Considerations for Drinking Water Treatment.
The formation of CHI3 during oxidative drinking water
treatment is influenced by parameters such as the stability
of the oxidant, I- concentration, type and concentration of
NOM, ammonia concentration, temperature, pH, and others.
Cases of taste and odor problems related to CHI3 formation
in drinking water treatment practice occurred commonly in
chloramination processes and rarely in chlorination pro-
cesses, but never in well-operated ozonation processes (6-
8). Even though we only performed experiments in two
natural waters with a few variable parameters, the observa-
tions from the literature compare well with the results we
obtained in these oxidation experiments. The main results
are summarized semiquantitatively in Table 4. Chloramine
is the oxidant which most likely results in CHI3 formation.
During chloramination, reactions competitive with the
-
formation of CHI3, such as those for the formation of IO3
or mixed I-THMs, are not important. When free chlorine
(ammonia-free water) is used, part of HOI can be oxidized
to IO3-. This reaction which prevents CHI3 formation is
especially important for high chlorine doses and stabilities.
When taste and odor problems are faced in chlorination
processes, an increase of the chlorine dose is a possible
solution. However, this may also lead to an increased
formation of undesired classical THMs such as chloroform.
Another competition reaction which prevents CHI3 formation
in the presence of chlorine is the formation of mixed I-THMs
such as CHCl2I. One factor in this context is the concentration
and reactivity of the NOM. If HOI is consumed quickly by
NOM which is connected to a low chlorine stability, only
minor quantities of IO3- are formed. For less reactive NOM,
HOI and chlorine persist longer and, in turn, high amounts
(28) Bichsel, Y. Behavior of Iodine Species in Oxidative Processes
During Drinking Water Treatment. Thesis, No. 13429, Swiss
Federal Institute of Technology, Zurich, Switzerland, 2000.
(29) Gallard, H., personal communication.
(30) Fuson, R. C.; Bull, B. A. Chem. Rev. 1934, 15, 275-309.
(31) Dubois, J. E.; Toullec, J. Tetrahedron Lett. 1971, 37, 3373-3376.
(32) Bell, R. P.; Longuet-Higgins, H. C. J. Chem. Soc. 1946, 636-638.
(33) Dean, J. A. Lange’s Handbook of Chemistry; McGraw-Hill: New
York, 1985.
-
-
of IO3 can be expected. In general, the amount of IO3
formation is the most important factor for the final fate of
iodine. When ozone was used, CHI3 formation was not
observed in the drinking water treatment practice or in our
experiments. Overall, the risk of production of CHI3-related
taste and odor problems increases in the order ozone <
chlorine < chloramine.
Received for review December 31, 1999. Revised manuscript
received March 30, 2000. Accepted April 12, 2000.
ES9914590
9
VOL. 34, NO. 13, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 2 7 9 1