Rate constants for reactions between iodine- and chlorine-containing species: A detailed mechanism of the chlorine dioxide/chlorite-iodide reaction

Article abstract of DOI:10.1021/ja953938e

The chlorite-iodide reaction is unusual because it is substrate-inhibited and autocatalytic. Because analytically pure ClO₂^- ion is not easily prepared, it was generated in situ from the rapid reaction between ClO₂ and I^-. The resulting overall reaction is multiphasic, consisting of four separable parts. Sequentially, beginning with mixing, these parts are the (a) chlorine dioxide-iodide, (b) chlorine(III)-iodide, (c) chlorine(III)-iodine, and (d) hypoiodous and iodous acid disproportionation reactions. The overall reaction has been studied experimentally and by computer simulation by breaking it down into a set of kinetically active subsystems and three rapidly established equilibria: protonations of chlorite and HOI and formation of I₃^-. The subsystems whose kinetics and stoichiometries were experimentally measured, remeasured, or which were previously experimentally measured include oxidation of iodine(-1,0,+1,+3) by chlorine(0,+1,+3), oxidation of I^- by HIO₂, and disproportionation of HOI and HIO₂. The final mechanism and rate constants of the overall reaction and of its subsystems were determined by sensitivity analysis and parameter fitting of differential equation systems. Rate constants determined for simpler reactions were fixed in the more complex systems. A 13-step model with the three above-mentioned rapid equilibria fits the overall reaction and all of its subsystems over the range [I^-]₀ < 10^-3 M, [ClO₂^-]₀ < 10^-3 M, [I^-]₀/[ClO₂^-]₀ = 3-5, pH = 1-3.5, and 25°C. The derived model with all experimentally determined rate and equilibrium constants fits both the overall reaction and all of its subsystems within 1% relative accuracy.

Full text of DOI:10.1021/ja953938e

3
708
J. Am. Chem. Soc. 1996, 118, 3708-3719
Rate Constants for Reactions between Iodine- and
Chlorine-Containing Species: A Detailed Mechanism of the
†
Chlorine Dioxide/Chlorite-Iodide Reaction
‡
Istv a´ n Lengyel, Jing Li, Kenneth Kustin,* and Irving R. Epstein
Contribution from the Department of Chemistry and Center for Complex Systems,
Brandeis UniVersity, Waltham, Massachusetts 02254-9110
X
ReceiVed NoVember 27, 1995
Abstract: The chlorite-iodide reaction is unusual because it is substrate-inhibited and autocatalytic. Because
-
analytically pure ClO2 ion is not easily prepared, it was generated in situ from the rapid reaction between ClO2 and
-
I . The resulting overall reaction is multiphasic, consisting of four separable parts. Sequentially, beginning with
mixing, these parts are the (a) chlorine dioxide-iodide, (b) chlorine(III)-iodide, (c) chlorine(III)-iodine, and (d)
hypoiodous and iodous acid disproportionation reactions. The overall reaction has been studied experimentally and
by computer simulation by breaking it down into a set of kinetically active subsystems and three rapidly established
-
equilibria: protonations of chlorite and HOI and formation of I3 . The subsystems whose kinetics and stoichiometries
were experimentally measured, remeasured, or which were previously experimentally measured include oxidation of
-
iodine(-1,0,+1,+3) by chlorine(0,+1,+3), oxidation of I by HIO2, and disproportionation of HOI and HIO2. The
final mechanism and rate constants of the overall reaction and of its subsystems were determined by sensitivity
analysis and parameter fitting of differential equation systems. Rate constants determined for simpler reactions
were fixed in the more complex systems. A 13-step model with the three above-mentioned rapid equilibria fits the
-
-3
-
-3
-
-
overall reaction and all of its subsystems over the range [I ]0 < 10 M, [ClO2 ]0 < 10 M, [I ]0/[ClO2 ]0 ) 3-5,
pH ) 1-3.5, and 25 °C. The derived model with all experimentally determined rate and equilibrium constants fits
both the overall reaction and all of its subsystems within 1% relative accuracy.
10 M s-1,6 in different modeling efforts. Similarly large
8
-1
Introduction
variation is seen in the rate constant found for the HOCl +
Rate constants of reactions between iodine- and chlorine-
containing species are very important in the mechanisms of
chlorite/chlorine dioxide and iodide/iodine/iodate-containing
chemical oscillators. Among the "chlorite-driven" chemical
oscillators one of the most versatile is the chlorite-iodide
reaction, which has become, next to the Belousov-Zhabotinsky
5
3 3,4
8
-1 -1 6
HIO2 reaction: 0, 1 × 10 , and 2 × 10 M s . The
treatment of iodine hydrolysis in the models differs substan-
3,7
tially: the forward reaction is described as a pH-dependent
or a pH-independent process.
4,6
There are also smaller but
1
-
significant variations for the rate constants of the HIO2 + I
+
(
+ H ), HIO2 + HOI, HIO2 disproportionation, HClO2 + HOI,
(BZ) reaction, perhaps the most widely studied reaction in
and HClO2 + HIO2 reactions. Since most of the rate constants
of previous models had not been measured directly, we decided
to reinvestigate the chlorite-iodide reaction, which is at the heart
of chlorite- and iodine-containing oscillators.
Among small molecule inorganic oxidation-reduction reac-
tions, the chlorite-iodide reaction is quite unusual, because it is
both substrate inhibited and autocatalytic. Iodide, which is a
reactant, inhibits the reaction, and iodine, one of the products,
accelerates the reaction. This feature was first recognized by
nonlinear chemical dynamics. The past 15 years has witnessed
publication of more than 200 articles on this reaction. These
studies have been concerned mainly with the system's nonlinear
features, such as oscillations, bistability, stirring and premixing
2
effects, and spatial phenomena. However, only a few of these
articles report detailed mechanistic investigations.
Previous attempts to model these complex dynamical systems
utilized neither rate constant determinations by parameter fitting
to experiments nor direct kinetics measurements of elementary
reactions; essentially these studies were qualitative descriptions
of oscillations in the chlorite-iodide reaction. Not surprisingly,
rate constants vary considerably among the different models.
For example, the rate constant for the reaction between HOCl
8
9,10
Bray and later studied in more quantitative detail.
Kern
9
and Kim established a very accurate empirical rate law valid
up to 90% of iodide conversion to iodine. They proposed that
the reaction is controlled by HOI-generating iodine hydrolysis:
HOI reacts rapidly with chlorous acid followed by even faster
reactions of the intermediates with iodide. This picture is
preserved in the subsequent detailed mechanisms of Epstein and
Kustin and Citri and Epstein. The Citri-Epstein model (C&E),
which is the most widely used model of the reaction, reproduces
many qualitative features of the experiments, but significant
3
3 4
5 5
and HOI was reported to be 0, 2 × 10 , 5 × 10 , and 2 ×
†
Systematic Design of Chemical Oscillators. 94. Part 93: Epstein, I.
6
3
R.; Kustin, K.; Lengyel, I. In Taube Insights: From Electron Transfer
Reactions to Modern Inorganic Chemistry; Isied, S., Ed.; American
Chemical Society: Washington, DC, 1996; In press.
‡
Current address: Chemical Engineering Department, Massachusetts
11,12
differences remain, as shown by Ross and co-workers.
Institute of Technology, 66-250, Cambridge, MA 02139-4307.
X
Abstract published in AdVance ACS Abstracts, April 1, 1996.
(4) R a´ bai, G.; Beck, M. T. Inorg. Chem. 1987, 26, 1195.
(5) Citri, O.; Epstein, I. R. J. Phys. Chem. 1988, 92, 1865.
(6) Epstein, I. R.; Kustin, K. J. Phys. Chem. 1985, 89, 2275.
(7) Noyes, R. M.; Furrow, S. D. J. Am. Chem. Soc. 1982, 104, 45.
(8) Bray, W. C. Z. Physik. Chem. 1906, 54, 741.
(9) Kern, D. M.; Kim, C.-H. J. Am. Chem. Soc. 1965, 87, 5309.
(10) De Meeus, J.; Sigalla, J. J. Chim. Phys. 1966, 63, 453.
(1) Epstein, I. R.; Orb a´ n, M. In Oscillations and TraVeling WaVes in
Chemical Systems; Field, R. J., Burger, M., Eds.; Wiley: New York, 1985;
p 257.
(
2) De Kepper, P.; Boissonade, J.; Epstein, I. R. J. Phys. Chem. 1990,
4, 6525.
3) Citri, O.; Epstein, I. R. J. Phys. Chem. 1987, 91, 6034.
9
(
0
002-7863/96/1518-3708$12.00/0 © 1996 American Chemical Society

Chlorine Dioxide/Chlorite-Iodide Reaction Mechanism
J. Am. Chem. Soc., Vol. 118, No. 15, 1996 3709
(III) was prepared from iodate and iodine in concentrated sulfuric acid.24
In a spectrophotometric cell, a small volume (50-200 µL) of this
solution was injected into a precooled NaOH solution, the concentration
of which was calculated to allow the final solution to reach the desired
pH of the sulfate buffer. The temperature of the solution after dilution
was 25 ( 1 °C.
We believe that there are additional problems with these
models. One shortcoming is that such models attempt to fit
data in regions where incomplete mixing seriously affects
13-19
oscillations,
without explicitly taking into account any such
interference with the chemical kinetics. The best model should
contain a set of rate constants that make it possible to account
for the effects of incomplete mixing using the appropriate
physical mixing model. None of the models proposed to date
Hypochlorous acid was prepared from the reaction of chlorine water
25
and HgO (Fluka). Excess HgO was mixed with water and a CCl
4
4
solution of chlorine. After the solids were filtered off, CCl was
-
includes triiodide ion explicitly. Inclusion of I3 in C&E, for
separated from the water phase containing HOCl, and the remainder
of Cl O in the CCl phase was extracted into water. The HOCl in the
example, would modify the entire dynamics. Another problem
is that HOI, which is the most significant intermediate in the
reaction, becomes protonated around pH 2 (the most frequently
2
4
water phase was further purified by distillation. The distilled HOCl/
Cl O was dissolved in cool water. HOCl obtained this way was free
2
+
from any lower or higher oxidation states of chlorine and from other
ions and was stored at 5 °C.
The ionic strength of the reaction mixtures was fixed at 0.3 M by
used experimental pH). One consequence of H2OI formation
is that the mechanism of iodine hydrolysis changes in going
from pH 3 to pH 1 (see a detailed analysis in ref 20). These
models do not account for such changes and consequently cannot
correctly describe the pH dependence below pH 3.
To obtain a more experimentally based mechanism of the
title reaction we studied the kinetics not only of the chlorite-
iodide reaction but also that of its subsystems, including the
chlorite-iodine, hypochlorous acid-iodine, hypoiodous acid, and
iodous acid disproportionation reactions and fitted the appropri-
ate rate constants of these reactions to the kinetics data. We
therefore obtained a self-consistent set of rate constants for the
overall reaction, which we now present.
adding calculated amounts of NaClO
4
to the buffer solutions. The
with NaOH.
NaClO was obtained by titrating HClO
4
4
Methods. Every reaction was studied by spectrophotometry with
an HP diode array spectrophotometer. In the case of slower reactions
a stirred spectrophotometric cell was used as a reactor, while in the
case of the chlorite-iodine and hypochlorite-iodine reactions, the HP
diode array spectrophotometer was combined with a HiTech stopped-
flow. The dead-time of the stopped-flow apparatus was about 5 ms,
and the maximum collection rate of the diode-array photometer was
1
0 Hz. Since the time scale of these reactions is in the 10 s range,
time resolution errors are negligible. Every reaction was measured
simultaneously at several different wavelengths or in a wavelength
range. Comparisons were made between experiments that used a light
filter which blocked light below 300 nm and those that used no filter.
Within experimental error, no difference was detected. As a further
check on whether light induced a reaction in this system, we varied
the shutter opening time from 0.1 to 1 s at 1 Hz collection rate and
found no detectable effect. At least three replicate runs were averaged
for each concentration set. The spectra of the reaction mixtures were
analyzed and the number of independent absorbing species was
determined by Coleman-Varga-Mastin's method.26 All measurements
Experimental Section
Materials. All chemicals were Fisher reagent grade unless noted
otherwise. Sodium chlorite (80%, Aldrich) was purified as described
21
earlier, and stock solutions were kept in the dark. The chlorite content
of the purified chlorite was 98.8 ( 0.5% by iodometric titrations.
2
2
Chlorine dioxide was prepared as described in the literature. Iodine
stock solutions were prepared by dissolving crystalline iodine in double-
distilled water. The iodide content was checked by measuring the
spectrum of the solution at the experimental pHs. The iodide content
was determined from this measurement, the spectra of iodine and
triiodide, and the triiodide formation equilibrium constant. It was
consistent with the equilibrium concentration of iodide originating from
iodine hydrolysis.
were performed at 25 ( 0.2 °C, except for the HIO
2
disproportionation
where the temperature was 25 ( 1 °C.
In reactors stirred at moderate speeds (300-600 rpm), imperfect
mixing can arise in the chlorite-iodide reaction, which causes irrepro-
ducible results. Fortunately, proper order of addition of reagents can
prevent this occurrence. If the reactor contains a chlorine dioxide
solution to which a lesser amount of iodide ion solution is added, which
we refer to as CI-order, chlorite and iodine are so rapidly produced
that the secondary reaction and mixing rates are competitive. Because
the iodide concentration is so low, there is virtually no inhibition, and
the reaction between chlorite and iodine follows immediately. Depend-
ing on the rate of iodide addition and the average mixing time, different
amounts of iodine will be consumed during mixing. A similar situation
occurs if chlorite replaces chlorine dioxide in the reactor when its
The organic content of “distilled water” reacting with iodine could
produce different iodide concentration levels in acidic media; conse-
quently, special precautions were taken to avoid this interference. For
studying the kinetics of the HOCl + I
2
and HClO
2
2
+ I reactions we
used organic-free, high purity water for the stock solutions and
determined the iodide content of iodine stocks in acidic media. In all
-
cases [I ] was close to that concentration originating solely from iodine
hydrolysis. These iodide concentrations were considered in the
calculations as initial values, although they did not make a significant
difference in the resulting kinetics curves.
-
3
concentration is above 10 M and the pH is 2 or less.
Hypoiodous acid was generated by the instantaneous hydrolysis of
ICl23 dissolved in methyl alcohol. Small amounts (50-200 µL) of ICl
solution were injected into 2.5 mL of sulfate buffer solutions. Iodine-
To avoid this source of error, the following reactor configuration
was used. The reactor contains iodide ion, to which a lesser amount
of chlorine dioxide solution is added (IC-order). In this case, iodide
-
is always in excess, the chlorite-iodide reaction is slow (strong I
11) Stemwedel, J. D.; Ross, J. J. Phys. Chem. 1993, 97, 2863.
inhibition) and the chlorite-iodine reaction does not occur to an
appreciable extent during the mixing process. Reproducible kinetics
curves were obtained with IC-order. At 300 rpm with CI-order, the
time at which all the iodide is consumed (peak of Figure 1) can be as
much as ten times shorter than that found with IC-order. At higher
stirring rates, the CI-order experiments get closer to the IC-order
experiments, and above 1200 rpm they are indistinguishable in the
spectrophotometric cell we used. However, the higher stirring rate
makes the spectrophotometric measurements extremely noisy, because
of cavitation and vortex formation. With the IC-order the reaction is
2
1
851.
(13) Roux, J. C.; Boissonade, J.; De Kepper, P. Phys. Lett. A 1983, 97,
68.
(
14) Boukalouch, M.; Boissonade, J.; De Kepper, P. J. Chim. Phys. Phys.-
Chim. Biol. 1987, 84, 1353.
(
15) Boissonade, J.; De Kepper, P. J. Chem. Phys. 1987, 87, 210.
(16) Menzinger, M.; Boissonade, J.; Boukalouch, M.; De Kepper, P.;
Roux, J. C.; Saadaoui, H. J. Phys. Chem. 1986, 90, 313.
(
(
(
(
(
(
17) Menzinger, M.; Giraudi, A. J. Phys. Chem. 1987, 91, 4391.
18) Ochiai, E. I.; Menzinger, M. J. Phys. Chem. 1990, 94, 8866.
19) Ali, F.; Menzinger, M. J. Phys. Chem. 1991, 95, 6408.
20) Lengyel, I.; Epstein, I. R.; Kustin, K. Inorg. Chem. 1993, 32, 5880.
21) Nagyp a´ l, I.; Epstein, I. R. J. Phys. Chem. 1986, 90, 6285.
22) Lengyel, I.; R a´ bai, G.; Epstein, I. R. J. Am. Chem. Soc. 1990, 112,
(24) Noszticzius, Z.; Noszticzius, E.; Schelly, Z. A. J. Phys. Chem. 1983,
87, 510.
(25) Handbook of PreparatiVe Inorganic Chemistry, 2nd ed.; Brauer,
G., Ed.; Academic Press: New York, 1963; Vol. 1, p 308.
(26) Coleman, J. S.; Varga, L. P.; Mastin, S. H. Inorg. Chem. 1970, 9,
1015.
9
104.
(
23) Wang, Y. L.; Margerum, D. W.; Nagy, J. C. J. Am. Chem. Soc.
1
989, 111, 7838.

3
710 J. Am. Chem. Soc., Vol. 118, No. 15, 1996
Lengyel et al.
Iodate is always one of, but not the only, product of part (c) in
any experiment, even at high iodine excess. At chlorite excess
the stoichiometry is even more complicated, because reaction
between Cl(III) and HOCl to produce ClO2 becomes significant.
The stoichiometry of this reaction also varies with the ratios of
reactants and product chloride. However, for our experiments,
chlorine dioxide formation never occurs in part (c). With ratios
-
-
of [I ]0/[ClO2]0 between 3 and 5, the [ClO2 ]/[I2] ratio in part
c) is always smaller than 0.33; consequently, iodine is in high
(
excess. These conditions prevent chlorine dioxide formation,
resulting in simpler reactions for part (c) and the ensuing part
(d).
Part (d) is a slow disproportionation of HOI and HIO2 to
molecular iodine and iodate. The stoichiometries of the
disproportionations are
Figure 1. The multiphasic kinetics curve of the chlorine dioxide-iodide
-
4
-
-4
reaction. [ClO
2
]
0
) 2.47 × 10 M, [I ]
0
) 8.0 × 10 M, pH
0
)
-
+
5
HOI ) 2I + IO + 2H O + H
(6)
2
-
-
2
3
2
2
.15, [SO
4
] + [HSO
4
] ) 0.05 M, ionic strength ) 0.3 M.
-
+
5
HIO ) I + 3IO + H O + 3H
2 2 3 2
(7)
stirring rate independent if the stirring rate is higher than 300 rpm. We
used the IC-order with stirring rates about 500 rpm.
The different parts of the overall reaction are well separated
from each other (Figure 1). Parts (b), (c) and (d) are the chlorite-
iodide reaction; the reaction between chlorite and iodine and
the parallel hypochlorite-iodine reaction occur in parts (c) and
(d); the disproportionation of iodine(+1,+3) takes place in part
(d).
The relationships between the different reactions are dia-
grammed in eq 8. The direction of an arrow indicates that the
reaction at the base of the arrow plays a determining role in the
reaction at the tip of the arrow.
Experimental Results
Overview. Since analytically pure chlorite is not easily
prepared, in these studies we chose an in situ preparation of
chlorite from the fast reaction between chlorine dioxide and
iodide. By using chlorine dioxide as a source of chlorite, the
purity of chlorite is not a question since chlorine dioxide can
be purified very easily. Also, the initial presence of iodine,
which is an autocatalyst, makes the reaction more reproducible
and faster. Control experiments using pure chlorite, iodine, and
iodide in comparable concentrations give the same kinetics
curves as shown in Figure 1. The first part of the chlorine
-
4
dioxide-iodide reaction is rapid in the 10 M concentration
2
2
range. It is complete in a few seconds and produces chlorite
and iodine according to the following stoichiometry and rate
law:
-
-
2
1
2
3
-
ClO + I ) ClO + I V ) 6.0 × 10 [ClO ][I ] (1)
2
2
2
It is useful to examine a characteristic kinetics curve for the
whole course of the multistep chlorine dioxide-iodide reaction
Although the reverse reactions of iodine hydrolysis are present
in the disproportionation of HOI and HIO2, they do not have a
determining role in the kinetics. However, in the cases of the
chlorite-iodide, chlorite-iodine, and hypochlorite-iodine reac-
tions, the iodine hydrolysis has different controlling roles
discussed below. The unfilled arrow from the HOCl + HClO2/
(Figure 1). The first fast increase of iodine, part (a), is due to
reaction 1. Then, in part (b), the chlorite (chlorous acid)
produced in reaction 1 reacts with the excess iodide to produce
more iodine in an accelerating reaction. The stoichiometry of
this part of the curve is well described by
-
ClO2 reaction indicates that this reaction is important only at
certain experimental conditions and can be completely separated
from the others. Equation 8 will be a guide for investigating
the component processes.
In this section we show experimental results beginning with
the overall chlorine dioxide-iodide reaction, and then proceed
systematically through its subsystems. In the following sections
we present determinations of the rate constants of the component
processes by fitting the parameters of the subsystems and a
discussion of these rate constants and their relations to each
other.
ClO2- + 4I + 4H ) 2I + Cl + 2H O
-
+
-
(2)
2
2
When iodide is consumed, the excess chlorite of reaction 2 reacts
with iodine in a rapid reaction, which is part (c). The higher
the chlorite excess of reaction 2, the more significant the portion
of iodine that is consumed. The stoichiometry of this process
varies with the ratio of chlorite and iodine; it also depends on
their initial concentrations. During part (c), when iodine is in
excess and the chlorine-containing species are Cl(III) and Cl ,
the iodine-containing products of this fast reaction, which is
complete when all chlorite is consumed, are HOI, HIO2 and
-
The Overall Chlorine Dioxide-Iodide Reaction. As we
have seen (Figure 1), the chlorine dioxide-iodide reaction is
complex, consisting of different, comparatively well separated
subprocesses. By analyzing the spectrum of the reaction mixture
-
IO3 . The combination of the next three overall stoichiometric
processes can describe the actual product ratio:
-
+
(Figure 2) after the first fast consumption of chlorine dioxide,
HClO + 2I + 2H O ) 4HOI + Cl + H
(3)
(4)
(5)
2
2
2
we can identify three absorbing species above 300 nm: iodine,
triiodide, and HOI and/or HIO2. The species HOI and HIO2
cannot be separated because of their overlapping spectra and
their small absorbances between 300 and 320 nm. Triiodide
ions are present until iodide is absent in the reaction mixture.
-
+
+
3
HClO + 2I + 2H O ) 4HIO + 3Cl + 3H
2
2
2
2
-
3
-
5
HClO + 2I + 2H O ) 4IO + 5Cl + 9H
2
2
2

Chlorine Dioxide/Chlorite-Iodide Reaction Mechanism
J. Am. Chem. Soc., Vol. 118, No. 15, 1996 3711
Figure 3. Kinetics curves of the chlorine dioxide-iodide reaction: (A)
chlorine dioxide dependence and (B) pH dependence. Dots represents
-
-4
Figure 2. Spectra of the chlorine dioxide-iodide reaction in two
different views showing characteristic absorbances of different species.
experimental points and solid lines are fitted. [I ]
0
) 8.0 × 10 M,
2-
-
[SO
ClO
1.70 × 10 (d) M. pH
4
] + [HSO ] ) 0.05, ionic strength ) 0.3 M in every experiment.
4
4
-
-4
-4
[
]
2 0
) 2.84 × 10 (a), 2.47 × 10 (b,e,f,g), 1.97 × 10 (c) and
) 2.15 (a-d,f), 1.62 (e), and 2.96 (g).
-
4
Iodide ions are consumed by the time the iodine concentration
reaches its maximum, HOI and HIO2 are formed during the
fast iodine consumption, and their disproportionation produces
more iodine. The optimum wavelengths for following the
kinetics are 470 nm (isosbestic point of iodine and triiodide),
0
+
pared with HOI/H2OI . Its reactivity would have to be much
higher than that of HOI to cause any significant kinetic effect.
Because of these considerations, and to have more adequate
kinetics curves for parameter fitting, we reinvestigated this
reaction.
3
50 nm (triiodide maximum absorptivity), and 320 nm (HOI
maximum absorptivity). The kinetics were studied at several
-
4
fixed initial iodide concentrations (8, 4, 2, and 1 × 10 M),
In Figure 4 we show how the spectrum of the reaction mixture
changes in time for excess iodine (A) and chlorite (B). With
iodine in excess, HOI and HIO2 intermediates form in parallel
with iodine consumption. With chlorite in excess, chlorine
dioxide formation can be observed clearly from its characteristic
spectrum.
and the initial chlorine dioxide concentration was changed to
-
reach ratios of [I ]0/[ClO2]0 between 3.0 and 5.0. In this range
of reactant ratios the reaction is easily measurable and very
reproducible. The pH dependence of the reaction was inves-
tigated between pH 1 and 3. In Figure 3 we show some
representative kinetics curves of the reaction with changing
initial iodide concentration (A) and changing pH (B). The time
at which the maximum in the iodine absorbance appears (tmax)
depends critically on the ratio of reactants but changes only
slightly with pH, although the change in hydrogen ion concen-
tration is two orders of magnitude.
Chlorine dioxide is important in the closed system oscillations
of the chlorite-iodide-malonic acid reaction and in Turing
pattern formation when chlorite is in high excess over iodide
2
2
-3
and at high initial reactant concentrations (>10 M). However,
ClO2 formation need not occur under all conditions used in the
chlorine dioxide-iodide reaction or in open system studies of
the chlorite-iodide reaction, which are used in analyzing
oscillations and stirring effects. Modeling studies showed that
the kinetics of the chlorous acid-iodine reaction is independent
of the chlorous acid/chlorite-hypochlorous acid reaction. After
investigating the kinetics and mechanism of the latter reaction
in detail, we found that the mechanism of chlorine dioxide
The Chlorous Acid-Iodine Reaction. This reaction is
responsible for the kinetics of parts (c) and (d) of Figure 1.
The reaction has already been studied,4 particularly at high
chlorite excess. R a´ bai and Beck assigned a key role to ICl;
however, the reaction occurs not only at high initial reactant
concentrations, where ICl may be present, but also at low
concentrations, and their model cannot account for the entire
concentration range. More importantly, they assumed a slow
rate of hydrolysis for ICl, but subsequent experimental studies
,27
-
formation in the HOCl + HClO /ClO reaction is much more
2
2
complex than that of the chlorite-iodide reaction and is well
separated from the latter. Since the mechanisms of the two
reactions are distinct, rate constants derived from models of the
two reactions are not correlated with each other. For clarity
and a better understanding of the chlorite-iodide reaction, we
will report the kinetics and mechanism of the HOCl + HClO2/
2
3
revealed that ICl hydrolysis is several orders of magnitude
faster than assumed by R a´ bai and Beck. Consequently, the
hydrolysis of ICl cannot have a rate-limiting role. In most
experiments its equilibrium concentration is low (<5%) com-
(27) Grant, J. L.; De Kepper, P.; Epstein, I. R.; Kustin, K.; Orb a´ n, M.
-
Inorg. Chem. 1982, 21, 2192.
ClO2 reaction separately.

3
712 J. Am. Chem. Soc., Vol. 118, No. 15, 1996
Lengyel et al.
Figure 5. Kinetics curves of the chlorous acid-iodine reaction with
changing initial chlorite concentrations, short (A) and long (B) time
-
-4
-4
-4
development. [ClO
2
4
]
0
) 2.13 × 10 (a), 3.13 × 10 (b), 5.6 × 10
-4 2-
2 0 4
-
Figure 4. Spectra of the chlorous acid-iodine reaction at (A) iodine
(c), and 8.7 × 10 (d) M; [I
]
) 2.61 × 10 M, pH ) 2.0, [SO
]
-
4
-
-
stoichiometric excess [I
2
]
0
) 2.3 × 10 M and [ClO
2
]
0
) 3.04 ×
+ [HSO
4
] ) 0.05 M and ionic strength ) 0.3 M. Dots represent
-
4
-4
1
×
0
M; (B) chlorite excess, [I
2
]
0
) 1.9 × 10 M and [ClO
2
]
0
) 1.2
experimental points and solid lines are fitted.
10- M. [SO
3
2-
] + [HSO ] ) 0.05 M, ionic strength ) 0.3 M and
-
4
4
pH ) 2.01 in both cases.
0
The Disproportionation of Hypoiodous Acid. Dispropor-
tionation of hypoiodous acid was studied by several groups2
at different conditions, and rate constants between 5 and 250
8-30
Kinetics curves of the chlorous acid-iodine reaction are shown
in Figures 5 and 6, in which the initial chlorite, iodine, and
proton concentrations are varied. The initial rate is independent
of the chlorite concentration (Figure 5A); however, it is iodine-
and pH-dependent (Figure 6 (parts A and B)). Figures 5 and 6
indicate that the hydrolysis of iodine may be the rate limiting
step in these chlorous acid-iodine experiments. The pH
dependence of the reaction is interesting (Figure 6B): the
kinetics becomes autocatalytic with decreasing pH.
-1
-1
M
s
were reported. Because of these differences, which
cannot be explained by differences in experimental conditions,
we carried out experiments to redetermine the kinetics of this
reaction. Figure 9 shows the change in the spectrum of HOI
-
with time. The three absorbing species are HOI, I2, and IO3 .
An isosbestic point at 374 nm, where iodine and hypoiodous
acid absorb, shows that there are no intermediates present at
high concentrations during disproportionation.
The Hypochlorous Acid-Iodine Reaction. The HOCl-I2
reaction was studied in excess HOCl.4 However, this condition
The Disproportionation of Iodous Acid. The kinetics of
this reaction were measured previously in highly acidic media
hardly ever occurs in the chlorite(chlorine dioxide)-iodide
reaction, because HOCl is produced by the reduction of chlorite
by several species, each of which will react with HOCl, too.
To obtain relevant data at stoichiometric excesses of iodine,
we reinvestigated the reaction. Figure 7 shows how the
spectrum of the hypochlorous acid-iodine reaction changes with
excess iodine and hypochlorous acid. At low pH and excess
hypochlorite the reaction is autocatalytic. At lower wavelengths
the production and consumption of HOI and HIO2 can be clearly
seen. On a longer time scale, these species will disproportionate
and produce iodine. Although the kinetics curves of this
reaction are complex, it is simpler than the chlorous acid-iodine
reaction, because all the reactions of HClO2 with iodine-
containing species are missing. Figure 8 shows the kinetics of
this reaction with changing hypochlorite (curves a-d), iodine
2
4,28
and found to be slow and autocatalytic.
Since our reaction
system operates at lower acid concentrations, we reinvestigated
the reaction close to our experimental conditions. Figure 10
shows the spectrum of the reaction mixture as it changes with
time. The formation of iodine is autocatalytic, similar to
Noszticzius et al.’s observations. Several kinetics curves were
measured at different concentrations of HIO2 and different pH
values. Changes in pH between 1 and 2 do not modify the
kinetics significantly. Because of the high sulfuric acid
concentration used in the preparation of HIO2, the experimental
conditions for studying the disproportionation are very limited.
Determination of Rate Constants
To determine the rate constants of likely quasi-elementary
and nonelementary processes we fitted the kinetics curves of
(curves d-g) and hydrogen ion concentrations (curves h-j).
The initial rates are nearly independent of hypochlorite con-
centration but depend on the iodine concentration and pH as in
the chlorite-iodine reaction. At hypochlorous acid excess and
lower pH, the reaction is autocatalytic. The autocatalytic feature
is much more pronounced than in the chlorite-iodine reaction.
(28) Furrow, S. J. Phys. Chem. 1987, 91, 2129.
(
29) Buxton, G. V.; Sellers, R. J. J. Chem. Soc., Faraday Trans. 1 1985,
1, 449.
30) Thomas, T. R.; Pence, D. T.; Hasty, R. A. J. Inorg. Nucl. Chem.
1980, 42, 183.
8
(

Chlorine Dioxide/Chlorite-Iodide Reaction Mechanism
J. Am. Chem. Soc., Vol. 118, No. 15, 1996 3713
Figure 6. Kinetics curves of the chlorous acid-iodine reaction with
changing initial iodine concentration (A) and pH (B). [I
2
-4
]
0
) 3.81 ×
-
-
4
-4
-4
1
1
0
0
(a), 2.86 × 10 (b), 1.84 × 10 (c), 1.36 × 10 (d), 1.05 ×
-4 - -4
2 0
Figure 7. Spectra of the hypochlorous acid-iodine reaction at iodine
4
(e) and 1.96 × 10 (f-j) M; [ClO
]
) 9.0 × 10 (a-e), 1.0
-
4
(
A) and HOCl (B) stoichiometric excess. [I
]
2 0
) 3.2 × 10 M, [HOCl]
0
-
3
×
10 (f-j) M; pH
0
) 1.89 (a-e), 1.19 (f), 1.74 (g), 2.23 (h), 2.50
-
4
-4
)
×
8.03 × 10 M in (A) and [I
2
]
0
) 2.95 × 10 M, [HOCl]
0
) 1.25
2
-
-
(i), 2.82 (j); [SO
4
] + [HSO
4
] ) 0.05 M and ionic strength ) 0.3
-
3
2-
-
4
10 M in (B). pH ) 1.66, [SO
0
4
] + [HSO ] ) 0.05 M and ionic
M. Dots represent experimental points and solid lines are fitted.
strength ) 0.3 M in both (A) and (B).
the above experiments. Except for the hypoiodous acid dis-
proportionation, the parameters of the differential equation
systems derived from possible simpler reactions were fitted.
Sensitivity analysis31 was used to evaluate different reaction
schemes to find the most plausible mechanisms.
new unknown parameters. Consequently, rate constants at the
higher level describe all of the lower level subsystems, the whole
mechanism comprising a self-consistent set of rate constants.
The rate constants so determined depend on the values of
certain equilibrium constants discussed below. For future
modeling, we recommend using the rate constants with the
related equilibrium constants to have a self-consistent set of
parameters that describes all the experiments done at these
experimental conditions.
Rapid Equilibria in the Chlorine Dioxide-Iodide Reaction.
The criterion used to define a rapid equilibrium was a relaxation
time shorter than 50 µs, which is about an order of magnitude
shorter than the relaxation time of the next slowest equilibrium,
the iodine hydrolysis. Foremost among the rapid equilibria is
triiodide formation, since, as was observed by Kern and Kim,9
triiodide ions are not reactive. The reactivity of triiodide in
Altogether, kinetics curves of the various reactions were
measured at more than 60 different experimental conditions,
and 3-5 replicate curves were averaged to improve the signal/
noise ratio. We used 50-250 points from each experimental
curve for fitting, depending on the complexity of the curves.
For the chlorine dioxide-iodide reaction more than 200 points
were used because of its multiphasic shape. Often, the fitting
procedure suggested new experiments at different initial con-
centrations or with simpler subsystems to obtain additional data
from which specific rate constants could be determined with
greater sensitivity and with the least possible correlation with
other parameters. For statistical analysis the goodness of fit
statistics32 of different models were compared. With the
3
4-36
other reactions is in accord with this observation.
With
-
formation of I3 , a significant portion of iodide and iodine is
captured in this unreactive complex. Using the equilibrium
2
exception of the iodous acid disproportionation, where R )
2
33
0
>
.99, R > 0.995. In all other cases, Model Selection Criteria
-
-
-
constant KI - ) [I3 ]/[I2][I ], with total iodide TI- ) [I ] +
3
6.5. The elements of the main diagonal of the correlation
matrix of fitted parameters were smaller than 0.6, and the
-
-
[
I3 ] and total iodine TI ) [I2] + [I3 ] concentrations as
2
variables, the triiodide concentration can be expressed as
-
3
residuals with absolute fitting were smaller than 3 × 10 and
-
2
with orthogonal fitting were smaller than 10 . In the fitting
procedure, we started from the simpler kinetic systems and
proceeded to the more complicated ones, fixing the previously
determined rate constants to make them uncorrelated with the
-
-1
(K + TI
[I ] )
+
3
2
I3-
-
+ TI2)
2
1
1
+
4TI TI2 (9)
-
(
31) Tur a´ nyi, T. Computers Chem. 1990, 14, 253.
2
(
KI3
-
+ T + T
)
x
I
-
I2
(32) Peintler, G. Zita, Version 4.0: A ComprehensiVe Program Package
for Fitting Parameters of Chemical Reaction Mechanisms; Institute of
Physical Chemistry, Attila J o´ zsef University: Szeged, Hungary, 1995.
(34) Bazsa, G.; Beck, M. T. Acta Chim. Hung. 1972, 73, 425.
(35) Adegite, A.; Egboh, H. Inorg. Chim. Acta 1977, 21, 1.
(36) Yiin, B. S.; Margerum, D. W. Inorg. Chem. 1990, 29, 1559.
(
33) MINSQ: Nonlinear Parameter Estimation and Model DeVelopment;
MicroMath Scientific Software: Salt Lake City, 1990.

3
714 J. Am. Chem. Soc., Vol. 118, No. 15, 1996
Lengyel et al.
Figure 8. Kinetics curves of the hypochlorous acid-iodine reaction:
Figure 9. The course of spectra during the disproportionation of
2 0
hypochlorous acid (a-d), iodine (d-g) and pH dependence (h-j). [I ]
-
4
-4
-4
-5
hypoiodous acid: (A) 3D representation and (B) showing the isosbestic
)
3.43 × 10 (a-d), 2.42 × 10 (e), 1.61 × 10 (f), 8.41 × 10
-4 -4 -4
0
-
4
point. [HOI]
0
) 3.76 × 10 M, pH
0
) 1.8, 0.1 M sulfate buffer.
(
g), 1.94 × 10 (h), 1.92 × 10 (i), 1.85 × 10 (j) M; [HOCl]
)
-
4
-4
-3
-3
4
7
.01 × 10 (a), 8.04 × 10 (b), 1.15 × 10 (c), 1.56 × 10 (d-g),
-
4
.38 × 10 (h-j) M; pH
0
) 1.66 (a-g), 1.05 (h), 1.39 (i), 2.82 (j).
2-
-
4
[SO
4
] + [HSO ] ) 0.05 M and ionic strength ) 0.3 M in every
case. Dots represent experimental points and solid lines are fitted.
The free, reactive iodide and iodine concentrations are
-
-
[
I ] ) T
-
- [I ]
(10)
I
3
-
[
I ] ) T - [I ]
2
I
3
2
In the rate equations of the reactions of iodide and iodine we
will refer to these free iodide and iodine concentrations.
The deprotonation constant of chlorous acid KHClO
)
2
-
+
[
ClO2 ][H ]/[HClO2] has been measured at several ionic
3
7
-2
-1
strengths; in our experiments the value is 2 × 10 M .
-
Keeping total chlorite TClO - ) [ClO2 ] + [HClO2] as a variable,
the concentrations of the different forms are
2
2 0
Figure 10. Spectra of the disproportionation of iodous acid. [HIO ]
+
-
4
TClO2
-
[H ]
) 5.1 × 10 M, pH ) 1.42, 0.7 M sulfate buffer.
0
[
HClO ] )
(11)
2
+
ments for studying the disproportionation of HOI show the
existence of this equilibrium by a shift in the spectrum of HOI
to lower wavelengths as pH decreases from 2 to 1. Keeping
KHClO2 + [H ]
TClO2
-
KHClO2
-
+
the total HOI concentration THOI ) [HOI] + [H2OI ] as a
[
ClO ] )
2
+
KHClO2 + [H ]
variable, the concentrations of the two iodine(+1) forms are
THOI[H+]
In the rate equations the two relations of eq 11 will be used to
+
[
H OI ] )
(12)
2
-
+
replace [HClO2] and [ClO2 ].
K
+
+ [H ]
H OI
2
The deprotonation constant of H2OI KH OI ) [HOI][H+]/
+
+
2
38
+
[
H2OI ] was measured by Bell and Gelles, who determined
THOI
K
+
H OI
2
the pH dependence of iodine hydrolysis. Independent experi-
[
HOI] )
ments39 confirmed their results. Spectrophotometric experi-
+
K
H OI+
+ [H ]
2
(
(
(
37) Hong, C. C.; Rapson, W. H. Can. J. Chem. 1968, 46, 2053.
38) Bell, R. P.; Gelles, E. J. Chem. Soc. 1951, 2734.
39) Allen, T. L.; Keefer, R. M. J. Am. Chem. Soc. 1955, 77, 2957.
In the rate equations the two relations of eq 12 will be used to
replace [HOI] and [H2OI ].
+

Chlorine Dioxide/Chlorite-Iodide Reaction Mechanism
J. Am. Chem. Soc., Vol. 118, No. 15, 1996 3715
When the pH of the reaction mixtures is higher than 2.5, the
buffer capacity of 0.05 M sulfate may not be high enough to
keep an approximately constant pH. We therefore tested buffer
concentrations change. From a knowledge of the equilibrium
constant of iodine hydrolysis, the two reverse rate constants (k )
can also be calculated. The rate constants obtained are
r
+
capacity. The value of [H ] did not differ by more than 10%
f
) 1.98 × 10 M s-1
-3
before and after the reaction, due to low reactant concentrations.
Because of this result, we did not include the rapid protonation
of sulfate ions explicitly in modeling, but treated the pH as
constant.
k
I a
2
r
) k /K = 3.67 × 10 M s-1 (16)
f
9
-1
k
I a
I a
I
2
2
2
f
) 5.52 × 10 s-1
-2
Rate Constants for Iodine Hydrolysis. Iodine hydrolysis
k
I b
2
(and its reverse) is one of the most important processes in this
r
f
/K = 3.48 × 10 M s-1
9
-1
k
) k K
I b H OI
+
reaction system. It participates in every subsystem of eq 8 in
which iodine-containing species appear. The overall reaction is
I b
I
2
2
2
2
The rate constant of the proton-inhibited path is very close
-
+
I + H O ) HOI + I + H
(13)
28
-3
2
2
to that found by Furrow (1.8 × 10 ). The rate constant of
the reverse reaction is the same as that found by Eigen and
The kinetics and mechanism of this reaction have already been
4
0
40,41
20
Kustin.
From these rate constants and the equilibrium
studied,
and rate constants have been suggested to describe
experiments in the pH range 3-7. The mechanism of the iodine
hydrolysis is given by the following reaction scheme:
constants of the reactions in the detailed mechanism of iodine
2
0
hydrolysis, the remaining rate constants of the detailed
mechanism can also be determined, assuming diffusion limited
rate constants for protonation processes. Our modeling proce-
dure verified that the detailed and the simplified versions of
iodine hydrolysis give the same kinetics curves for every
subsystem, proving that our simplifying assumptions are justi-
fied. Therefore, we can use a single stoichiometry with a more
complex rate equation for iodine hydrolysis.
Rate Constant for the Disproportionation of HOI. The
disproportionation of hypoiodous acid was studied by several
investigators; the large variation in rate constant cannot be
2
9,30
explained by differences in experimental conditions.
Direct
-
1
-1
measurements by Furrow determined values of 18-25 M
s
This complex equilibrium system can be simplified substan-
with different experimental methods. Fitting experimental
curves similar to those shown in Figure 9 gives a simple second
order dependence on the HOI concentration, using wavelengths
where either iodine or HOI absorbs. Our results agree with
Furrow's studies.
tially for our experiments. (a) The I2 + OH- reaction is not
important at pH < 6. (b) All the protonation equilibria are fast.
-
(
c) I2OH does not have any reactions other than those shown
in eq 14. Thus iodine hydrolysis is considered to occur in two
-
+
parallel pathways: formation of HOI and I through H2OI and
-
2
HOI formation through I2OH . With simplifications (a)-(c)
V ) (22.0 ( 1.5)[HOI]
(17)
the detailed mechanism is resolved into two parallel equilibria
with the following rate equations:
Rate eq 17 gives an accurate description of both the iodine
formation and the HOI consumption. This simple second order
dependence together with the presence of an isosbestic point
suggests that the rate determining step in the production of
iodine is the disproportionation process, and all other reactions
of intermediates are faster. Therefore, in view of the above
rate equation and the kinetics of iodine hydrolysis, there is
another important relationship among the rate constants of
intermediates, since HIO2 does not accumulate in the system.
Either the rate of HIO2 disproportionation or the rate of the
f
+ -1
r
-
V
) k [I ][H ] - k [HOI][I ]
(15)
I a
I a
2
I a
2
2
2
f
r
+
-
V
) k [I ] - k [H OI ][I ]
I b 2 I b 2
2 2
I b
2
The superscripts “f” and “r” refer to forward and reverse rate
constants, respectively. With this scheme we simplify the entire
iodine hydrolysis submechanism without losing significant
-
kinetic detail, provided that I2OH does not undergo any
+
HOI(H2OI ) + HIO2 reaction is significantly faster than the
reactions that are substantially faster than its decay to iodine or
-
+
disproportionation of HOI. Moreover, the HIO2 + I + H
-
to HOI and I . We found no evidence that any such fast
-
reaction must be slower than the HOI + I reaction; otherwise
no second order dependence on HOI concentration could be
observed.
reactions exist with either chlorine-containing species or HIO2.
f
To determine rate constants for the two k in eq 15, we ran
experiments in which iodine hydrolysis is the rate limiting
process: the HClO2 + I2 and HOCl + I2 reactions. As shown
in Figure 5, the initial rates of these reactions are independent
of the chlorite concentration but depend on the iodine concen-
tration (Figure 6A). The pH greatly affects the initial rate
Rate Constants for the Disproportionation of HIO2 and
+
the HOI/H2OI + HIO2 Reactions. We are aware of two
previous investigations aimed at finding rate constants for the
24,28
disproportionation of HIO2.
Perhaps because the preparation
of HIO2 is uncertain, the rate constants differ. Noszticzius et
al. followed the corrosion potential of iodide-selective electrodes
in the system, and observed that the disproportionation process
is autocatalytic. In our experiments, the production of iodine
increases autocatalytically, and there is a parallel autocatalytic
consumption of HIO2, providing evidence that the iodine
production is not delayed, as would be observed for a sequence
of consecutive reactions. This autocatalytic behavior suggests
that the initiating disproportionation reaction is slower than the
(Figure 6B), because it affects the iodine hydrolysis. The
determination of the rate constants in eq 15 is extremely
sensitive to these experiments; consequently, the values obtained
are accurate and uncorrelated. The solid lines are fitted curves
to the experiments. They show excellent agreement in the initial
+
rates in Figures 5, 6, and 8, where the iodine and H
(
(
40) Eigen, M.; Kustin, K. J. Am. Chem. Soc. 1962, 84, 1355.
41) Palmer, D. A.; van Eldik, R. Inorg. Chem. 1986, 23, 928.

3
716 J. Am. Chem. Soc., Vol. 118, No. 15, 1996
Lengyel et al.
Table 1. Mechanism of the Chlorine Dioxide-Iodide Reaction at [I-]
/[ClO
2 0
-] ) 3-5, pH ) 1-3.5, and 25 °Ca
0
-
-
1
3
-
(
(
M1)
M2)
ClO
2
+ I f ClO
2
2
+ /
2
I
2
V
V
V
V
V
V
V
V
V
V
V
V
V
V
1
) 6 × 10 [ClO
2
][I ]
2
-
+
-3
+
9
-
I
2
+ H
O h HOI + I + H
2a) 1.98 × 10 [I
]/[H ] -3.67 × 10 [HOI][I ]
-
2
9
+
-
2b ) 5.52 × 10 [I
2
] -3.48 × 10 [H
2
OI ][I ]
-
+
-
(
(
(
(
(
(
(
(
(
(
(
M3)
M4)
M5)
M6)
M7)
M8)
HClO
HClO
HClO
2
2
2
+ I + H f HOI + HOCl
3
4
5
6
7
8
9
) 7.8[HClO
2
][I ]
7
+ HOI f HIO
2
+ HOCl
) 6.9 × 10 [HClO
2
][HOI]
][HIO
-
+
6
+ HIO
2
f IO
3
+- HOCl + H
) 1.0 × 10 [HClO
2
2
]
-
8
-
HOCl + I f HOI + Cl
) 4.3 × 10 [HOCl][I ]
-
-
+
3
HOCl + HIO
2
f IO
3
+ Cl + 2H
+ I + H h 2HOI
) 1.5 × 10 [HOCl][HIO
2
]
-
+
9
-
+
2
HIO
2HIO
HIO
2
) 1.0 × 10 [HIO
2
][I ][H ] - 22[HOI]
-
+
2
M9)
2
f IO
3
+ HOI + H
2
) 25[HIO ]
+
-
-
+
+
M10)
M11)
M12)
M13)
2
+ H
2
OI f IO
3
+ I + 3H
+ H
10) 110[HIO4
2
][H
2
OI ]
b49
b
-
+
-
+
HOCl + Cl + H h Cl
2
2
O
11) 2.2 × 10 [HOCl][Cl ][H ] - 22[Cl
2
]
-
+
5
Cl
Cl
2
+ I
2
+ 2H
2
O f 2HOI + 2Cl + 2H
O f HIO
12) 1.5 × 10 [Cl
2
2
2
][I ]
][HOI]
b
-
+
6
2
+ HOI + H
2
2
+ 2Cl + 2H
13) 1.0 × 10 [Cl
Rapid Equilibria:
-
+
-
+
-2
(
(
(
M14)
M15)
M16)
HClO
2
h ClO
2
+ H
K
K
K
HClO2 ) [ClO
2
][H ]/[HClO
2
] ) 2.0 × 10
+
+
+
+
+
H
2
OI h HOI + H
+ I h I
, ClO
H OI ) [HOI][H ]/[H
2
OI ] ) 0.034
2
-
-
-
I3 ) [I
-
-
I
2
3
3
]/([I
2
][I ]) ) 740
Variables: TI-,c T 2
,c THOI, HIO
c
, TClO2-,c HOCl, Cl b
- b
I
2
2
2
, Cl
a
b
-3
All concentration units in M, times in s. Reactions M11-M13 are important only at pH < 2.0 and high concentrations of HClO
2
(>10 M)
2 2
or in the HOCl + I reaction. Otherwise they can be neglected together with Cl and Cl as variables. Fits shown in Figure 3
c
-
-
or at high [Cl ]
0
-
+
2 2 2
were produced without considering these reactions and variables. The appropriate concentrations of I , I , HOI, H OI , and HClO can be expressed
using eqs 10-12.
+
HOI/H2OI + HIO2 reaction. The set of reactions M8-M10
and the reverse of M2 in Table 1 provide an accurate description
of the kinetics. Reactions M9 and M10 are rate limiting, and
reactions M8 and the reverse of M2 are fast; consequently, the
kinetics are sensitive only to M9 and M10. The rate constants
for those steps are given in eqs 18 and 19.
Although ICl is more stable toward disproportionation than HOI,
because of its low concentrations and small ratio to the total
iodine(+1) concentration, it will not modify the iodine(+1)
disproportionation rate significantly. This contention is sup-
-
ported by the fact that Cl does not have an effect on the
chlorite-iodide and chlorite-iodine reactions at the low concen-
trations comparable to those produced in the reaction. For these
reasons, we did not include ICl as an independent species.
From experiments performed with the HOCl + I2 reaction,
we can give a good estimate for the rate constant of this direct
reaction. As can be seen (Figure 8, curves a-d), the initial
rate of the reaction is almost independent of the concentration
of HOCl. This observation indicates that the direct reaction
must be slow compared with the reaction pathway regulated
by iodine hydrolysis. Otherwise, with increasing HOCl con-
centration one would observe a significant increase in the initial
-
+
2
2
HIO ) IO + HOI + H V ) (25 ( 8)[HIO2]
(18)
2
3
+
-
3
-
+
HIO + H OI ) IO + I + 3H
2
2
+
V ) (110 ( 15)[HIO ][H OI ] (19)
2
2
Our disproportionation rate constant is higher (about 5 times)
and that for the I(+1) + HIO2 reaction is lower (about half)
than Noszticzius et al. or Furrow’s values, because they did
not consider the protonation of HOI. The rate constant
determination for the reaction between HOI and HIO2 is not
accurate, probably because of the high error of determining
initial concentrations, and the very narrow limit of feasible
experimental conditions. A rate constant below 20 M
for the HOI + HIO2 reaction does not significantly affect the
other experiments, and this step can be neglected.
Rate Constants for HOCl + Iodine-Containing Species.
In this group of reactions we analyze the reactions of HOCl
with I , I2, HOI and HIO2. The reaction of I with HOCl is
fast, complex and involves many intermediates.42 The chlorine
dioxide-iodide, chlorite-iodide and hypochlorous acid-iodine
reactions are not sensitive to the rate constant of this process
rate. Furthermore, the HOCl + I direct reaction should be a
2
pH-independent process. In the real experimental system,
however, we observe a decrease in the initial rate with
decreasing pH (Figure 8B), which also shows that the direct
reaction cannot have a significant contribution. From our
-
1
-1
s
-
1 -1
experiments, the upper limit of this rate constant is 2 M s ,
and with this or lower rate constants the HOCl + I reaction
2
can be neglected.
From a knowledge of the rate of the fast reaction between
-
-
-
HOCl and I and the HOCl-independent initial rate of the HOCl
+
I2 reaction, we obtain a reliable estimate for the upper limit
of the rate constant for the HOCl + HOI reaction: ap-
-1 -1
proximately 20 M s . With higher rate constants the
8
-1 -1
unless it is below 10 M s . Therefore, we used one overall
stoichiometric equation with a simple rate law of the form
description of the pH dependence and the stoichiometry of the
experiments would not be correct. Lower rate constants,
including zero, describe the experiments correctly; consequently,
the reaction between HOCl and HOI can be neglected, like the
HOCl + I2 direct reaction.
The reaction between HOCl and HIO2 is an important
component of the temporal development of the HOCl-I2 reaction
after the initial fast phase. It generates the correct stoichiometric
amount of HOI and HIO2 to produce the expected amount of I2
during the disproportionation of the first two species. This
reaction also has an important effect on the chlorite-iodine
reaction and regulates the level of HOCl available for other
reactions. The best fits for the experimental kinetics curves
can be obtained with the following rate equation.
-
-
8
-
HOCl + I ) HOI + Cl V ) 4.3 × 10 [HOCl][I ] (20)
In the mechanism proposed by Nagy et al.42 one of the
intermediates is ICl. Since the concentration of this species is
quite low in our experiments, and since ICl is a form of iodine-
(
+1), its reactivity would have to be much higher than that of
+
HOI/H2OI to observe any significant kinetic effect. However,
+
-
the rates of the HOI/H2OI + I reactions are already diffusion
limited, therefore, the ICl + I reaction cannot be important.
-
(
42) Nagy, J. C.; Kumar, K.; Margerum, D. W. Inorg. Chem. 1988, 27,
2
773.

Chlorine Dioxide/Chlorite-Iodide Reaction Mechanism
J. Am. Chem. Soc., Vol. 118, No. 15, 1996 3717
-
-
+
3
. Two reactions determine the rate of part (b) in Figure 1.
HOCl + HIO ) IO + Cl + 2H
2
3
-
One of them is the direct reaction between HClO2 and I and
3
V ) (1.5 ( 0.2) × 10 [HOCl][HIO ] (21)
2
the other is the HClO2 + HOI reaction. At low pH, where the
-
[HClO2]/[ClO2 ] ratio is high, and the HOI level from the iodine
Rate Constants for Cl2 + Iodine-Containing Species. Rate
constants for chlorine reacting with iodine-containing com-
pounds can be determined most sensitively from experiments
done with the HOCl + I2 reaction at HOCl excess (Figure 8,
curves c-i). In this case chlorine, which is formed from the
hydrolysis low, the contribution of the direct reaction between
-
HClO2 and I reaction will be high. At high pH, however, since
-
most of the chlorite is in the form of ClO2 and this is not the
reactive form, the direct reaction is slow. On the other hand,
+
lower H concentration shifts the iodine hydrolysis to increase
-
+
HOCl + Cl + H reaction, could have a significant concentra-
tion and contribution to the overall rate.
the level of HOI, and the contribution of the HOI + HClO2
reaction will be significantly higher. These two rate constants
can be determined very sensitively from the experiments done
with the chlorine dioxide-iodide reactions. The stoichiometries
and rate laws are
From our experiments, nothing can be established about the
rate of the direct reaction between chlorine and iodide. This is
certainly a fast reaction; however, even with diffusion controlled
rate constants it does not contribute significantly to the kinetics
of the full system, because iodine and HOI, which are present
at about 6 orders of magnitude higher concentration than iodide,
react rapidly with chlorine.
-
+
HClO + I + H f HOI + HOCl
2
-
V ) (7.8 ( 0.1)[HClO ][I ] (24)
2
Chlorine reacts rapidly with iodine and even faster with
hypoiodous acid. These reactions can explain the autocatalytic
features of the HOCl + I2 reaction at high HOCl excess and
low pH. As we have shown in Figure 7, the intermediates of
the HOCl + I2 reaction are consumed autocatalytically and
rapidly at high acid concentrations and hypochlorous acid
excess. This process is fast because chlorine, the equilibrium
concentration of which is high at higher acidities, reacts rapidly
with the intermediate HOI. At lower acidities the chlorine level
is low and its reactions are less important. The rate laws for
these two reactions are
HClO + HOI f HIO + HOCl
2
2
7
V ) (6.9 ( 0.1) × 10 [HClO ][HOI] (25)
2
The reaction between HClO2 and HIO2 will determine the
kinetics curves of the chlorite-iodine reaction after the first fast
decrease in the iodine concentration (the part that is zeroth-
order in chlorous acid concentration), before the complete
consumption of chlorite. This step produces a slower decrease
in the iodine concentration (Figure 5). Since the prominence
of this reaction accounts for the most substantial difference
between our model of the chlorite-iodide reaction and previous
models, we tried to determine its kinetics by direct measure-
-
+
Cl + I + 2H O f 2HOI + 2Cl + 2H
2
2
2
5
+
ments. Since we have HIO2(IO ) available, which was used
V ) (1.5 ( 0.2) × 10 [Cl ][I ] (22)
2
2
in studying the kinetics of disproportionation of HIO2, we can
use the same experimental procedure for investigating the HClO2
-
+
+
+
HIO2 reaction. The sulfuric acid solution of IO was added
Cl + HOI + H O f HIO + 2Cl + 2H
2
2
2
-
to alkaline ClO2 . After mixing, the change of the spectrum
of the reaction mixture with time was studied. However, the
reaction reached completion during mixing; no changes in the
spectra were observed. In contrast to the HIO2 disproportion-
ation experiments, where iodine formed slowly, no iodine
formed here, and no change in the absorption region of HIO2
was observed after mixing. Knowing the initial concentrations
and mixing time allows us to estimate the lower limit of the
6
V ) (1.0 ( 0.2) × 10 [Cl ][HOI] (23)
2
The reaction between chlorine and iodous acid does not
contribute significantly in our experimental systems, since HOCl
reacts comparatively rapidly with HIO2. Choosing a rate
6
-1 -1
constant smaller than 10 M
s
does not affect the kinetics
of the reactions under consideration. However, higher rate
constants will destroy the correct stoichiometry. From our
experiments the only certainty is that the rate constant should
5
-1 -1
rate constant for this reaction: 3 × 10 M s . The fitted
value to the experiments done with the chlorite-iodine reaction,
together with the stoichiometry is
6
-1 -1
be smaller than 10 M
reaction can be neglected.
s
and that with such values this
Reactions 22 and 23 have minor roles at the initial conditions
used in the chlorine dioxide-iodide reactions; their contributions
are less than 1% even after tmax on the kinetics curves. However,
they have critical roles in the HClO2-I2 and HOCl-I2 reactions
at high initial concentrations or an excess of oxidants.
-
+
HClO + HIO f HOCl + IO + H
2
2
3
6
V ) (1.0 ( 0.3) × 10 [HClO
][HIO ] (26)
2 2
The rate constant of the reaction between HIO2 and I-.
Interestingly, only the most complex subsystems of eq 8, the
chlorite-iodine and chlorine dioxide-iodine reactions, are sensi-
tive to the value of this rate constant. All the other reactions,
even HIO2 disproportionation, require only a rapid reaction with
Rate Constants for HClO2 + Iodine-Containing Species.
The reactions of HClO2 with iodine-containing species are
probably the most important in determining the dynamics of
-
-
the ClO2 + I and HClO2 + I reactions. Iodine does not react
directly with HClO2. This statement can be verified by
inspecting Figure 5, where the chlorite dependence of the
chlorite-iodine reaction is shown. The initial rate is completely
independent of the concentration of chlorite. If there were any
significant direct reaction between HClO2 and I2, the initial rate
would depend on the chlorite concentration. The upper limit
6
-2 -1
+
a rate constant higher than 10
M
s
(H dependent).
However, the more complex systems can tolerate only a narrow
range for this rate constant, because of the fast reaction between
HClO2 and HIO2. If the HIO2 + I reaction were slow, most
of the HIO2 would react with HClO2 rather than with I to
produce iodate in part (b) of Figure 1. This reaction would
destroy the correct stoichiometry and kinetics of the whole
reaction. If the rate of the HIO2 + I + H reaction is faster
than the HClO2 + HIO2 reaction, then HOI will be produced.
Thus, in this model HIO2 is a highly reactive species, reacting
-
-
-
1
-1
for this rate constant is about 0.01 M
ments.
s
from our experi-
-
+
-
The direct reaction between HClO2 and I is slow but has a
determining role in the shape of the kinetics curves of Figure

3
718 J. Am. Chem. Soc., Vol. 118, No. 15, 1996
Lengyel et al.
rapidly with iodide to produce iodine. This step keeps HIO2 in
the reaction cycle and does not allow it to be removed as iodate,
which is an inactive product. Consequently, any uncertainties
in the rate constant of the HClO2 + HIO2 reaction increase the
-
+
uncertainty of the rate constant of the HIO2 + I + H reaction.
The stoichiometry and rate law consistent with our experiments
are
-
+
HIO + I + H f 2HOI
2
9
-
+
V ) (1.0 ( 0.5) × 10 [HIO ][I ][H ] (27)
2
which will ensure that HIO2 remains in the reaction cycle. The
pH dependence of this reaction comes from the pH dependence
of the whole reaction. We tested the possibility of a pH
independent process; however, the stoichiometry of the pH
dependent kinetics curves in the chlorine dioxide-iodide reaction
could not be reproduced.
Discussion
In Table 1 we list the important reactions in the chlorine
dioxide/chlorite + iodide reaction system with their rate
constants. These reactions were necessary and sufficient to
describe the chlorine dioxide-iodide reaction in all of our
experiments and, furthermore, to account for the kinetics of HOI
and HIO2 disproportionation (reactions M2 and M8-M10), the
HOCl + I2 reaction (reactions M2 and M6-M14), and the
HClO2 + I2 reaction at excess iodine (reactions M2 and M4-
-
3
M14) when [HClO2]0/[I2]0 < 2.5 and [HClO2]0 < 2 × 10
M
in the pH range 1-3.5. We did not include several other
reactions which certainly take place in the system but whose
rates are too slow to make observable contributions to the overall
Figure 11. Modeling the open system behavior of the chlorine dioxide-
iodide reaction using reactions M1-M10 and M14-M16 of Table 1.
50
-
0
A) Bifurcation diagram (calculated using CONT ) in the [I ] vs k
0
-
-
-
(reciprocal residence time) plane. The solid line encloses the area of
bistability; the dashed line indicates Hopf bifurcations and encloses
kinetics. These are, e.g., the Cl2 + I and the IO3 + I
reactions.
-
4
+
-2
the range of oscillations. [ClO
M. The experimental bifurcation diagram is published in ref 48. (B)
Dependence of the period of oscillations on k . Filled circles are
experimental points, and the solid line is calculated. [I ]
2
]
0
) 1 × 10 M and [H ] ) 1 × 10
With the exception of the chlorine dioxide-iodide reaction,
all of the reactions considered are two-electron processes; the
model does not include radical processes. Although the ClO2
0
-
-4
0
) 3 × 10
-
-
-4
+
-2
+
I reaction is rapid and first-order in both ClO2 and I , it is
M, [ClO ] ) 1 × 10 M and [H ] ) 1 × 10 M.
2
0
not elementary. The rate limiting step is the direct reaction
-
between ClO2 and I , followed by a faster process; the process
is very fast. Because the HOI level is controlled by iodine
hydrolysis, decreasing iodide concentration shifts reaction 13
to the right, to higher levels of HOI, which makes the reaction
faster as the iodide concentration decreases.
may occur through an inner-sphere mechanism.2
2,43
If fast
radical processes occur, there is no evidence that the radical
intermediates interact with any other species rapidly enough to
have an effect on the overall kinetics. This reaction is well
separated from the others, as illustrated by the observation that
the kinetics of the chlorite-iodide reaction are the same as those
of the chlorine dioxide-iodide reaction if we start from the same
initial concentration that exists after completion of the fast
chlorine dioxide-iodide reaction.
Since there is no, or only a very slow, direct reaction between
HClO and I , the same model holds true for the chlorite-iodine
2
2
reaction. When iodide is present, HIO cannot be oxidized to
2
+
iodate by HClO or H OI , but reaction with iodide will produce
2
2
HOI, and that will be converted to iodine. When iodide is
consumed, nothing prevents HIO from being oxidized to iodate,
2
The two-electron reactions in this model which produce the
kinetic feedback stand in contrast to the mechanisms of many
other, non pH-regulated chemical oscillators. For example, in
the case of the BZ reaction, the autocatalytic feedback process
originates from the one-electron reaction of BrO2 radical with
and iodine starts to decrease very rapidly, converting to HOI/
+
-
H OI , HIO , and IO .
2
2
3
Most of the reactions in the model have mass action kinetics,
which might mean that they are elementary reactions. However,
it is clear that most of these reactions are not elementary; they
go through several intermediates. Some examples of rapid but
not elementary reactions are the hypochlorous acid-iodide
4
4
7,45,46
the catalyst.
The Briggs-Rauscher reaction
and the
permanganate oscillators47 have similar one-electron based
feedback processes. The unique accelerating feature in part (b)
of Figure 1 comes mainly from iodide inhibition. This
conclusion is based on the observations that the direct reaction
2
3
-
reaction, the reaction between HIO and I and chlorine
2
4
0
hydrolysis. In our reaction system, however, the details of
these reactions are not important because of the different time
scales. The reaction between HClO , I , and H is certainly
-
-
+
between HClO2 and I is slow, but the HClO2 + HOI reaction
2
+
not elementary, because its kinetics are [H ] independent. It
(
(
43) Fukutomi, H.; Gordon, G. J. Am. Chem. Soc. 1967, 89, 1362.
44) Noyes, R. M.; Field, R. J.; Thompson, R. C. J. Am. Chem. Soc.
is easy to construct a model consisting only of elementary
-
-
1
971, 93, 7316.
reactions, which would include the protonation of IO or ClO ;
however, this level of detail does not improve the kinetics fit
and is not included in our model. Such a simplification guided
our modeling efforts, for example, when we replaced the detailed
(45) Briggs, T. S.; Rauscher, W. C. J. Chem. Educ. 1973, 50, 496.
(46) De Kepper, P.; Epstein, I. R. J. Am. Chem. Soc. 1982, 104, 49.
(47) Orb a´ n, M.; Lengyel, I.; Epstein, I. R. J. Am. Chem. Soc. 1991, 113,
1
978.

Chlorine Dioxide/Chlorite-Iodide Reaction Mechanism
J. Am. Chem. Soc., Vol. 118, No. 15, 1996 3719
mechanism of iodine hydrolysis with a simplified version that
preserves all the important equilibrium and kinetics features of
the detailed mechanism and used rapidly established equilibrium
constants for the formation of I3 , H2OI and HClO2. The
number of concentration variables in the model decreased
significantly, which benefits applications to more complex
systems.
HOCl + I- reaction, the main disparity lies in the difference
between rate constants for the reactions of HIO2. In the C&E
model, there is no reaction between HClO2 and HIO2, the rate
constant of HIO2 + I + H reaction is only 10 M s , and
the rate of disproportionation is two orders of magnitude faster
-
+
-
+
6
-2 -1
3
-1 -1
(3 × 10
M
s ) than the experimental value. These
differences mean that in part (b) of Figure 1, the HIO2 level is
four orders of magnitude higher, the maximum concentration
of HIO2 at tmax is three times higher and HIO2 is consumed
faster after tmax than in our model. The existence of a reaction
between HClO2 and HIO2 was shown experimentally. Since
this is a fast reaction, the HIO2 + I + H reaction must be
significantly faster than this; otherwise the stoichiometry of part
(b) would not be correct in the model. The central role of HIO2
in the C&E model is shown by the extreme sensitivity of
oscillations in that model to the rate constant for the dispro-
portionation of HIO2, which in the model is two orders of
magnitude higher than the experimental value.
In the Introduction we mentioned some of the discrepancies
between the C&E model and experiment. These are mainly
(1) that it produces homogeneous oscillations where significant
stirring effects occur and (2) that it does not produce oscillations
-
4
-
+
in the chlorine dioxide-iodide reaction at 10
M initial
48
concentration of ClO2. This latter discrepancy is surprising,
since a very simple model,48 based on component processess,
is capable of quantitative description of the open system
behavior, and the Epstein-Kustin model, which lacks the
disproportionation reactions of HOI and HIO2, also gives
oscillations in the chlorine dioxide-iodide reaction. Our model
(
Table 1) does not produce homogeneous oscillations where
The model we have presented (Table 1) uses only experi-
mentally determined rate and equilibrium constants. Further
modeling of the interaction between chlorine- and iodine-
containing species should be carried out, however, to investigate
a wider range of conditions, including the effects of imperfect
mixing.
stirring and mixing effects occur in the chlorite-iodide reaction.
However, it produces homogeneous oscillations at the conditions
observed in the chlorine dioxide-iodide reaction. In Figure 11
we show a two-parameter bifurcation diagram and the oscillation
period Vs flow rate curves. Both are in good agreement with
experiment.
To find the reason for the discrepancy we compared rate
constants of our present model and rate constants used by C&E.
Besides small differences regarding iodine hydrolysis and the
Acknowledgment. We thank G a´ bor Peintler for allowing
us to use his programs Zita (Version 4.0) for parameter fitting
and for analyzing spectra. This work was supported by
Research Grant CHE-9023294 and U.S.-Hungarian Coopera-
tive Grant INT-9322738 from the National Science Foundation
and by the W. M. Keck Foundation.
(48) Lengyel, I.; Li, J.; Epstein, I. R. J. Phys. Chem. 1992, 96, 7032.
(49) Wang, T. X.; Margerum, D. W. Inorg. Chem. 1994, 33, 1050.
(50) Marek, M.; Schreiber, I. Chaotic BehaVior in Deterministic Dis-
sipatiVe Systems; Cambridge Univ. Press: Cambridge, 1991.
JA953938E