Aqueous photochemical reactions of chloride, bromide, and iodide ions in a diode-array spectrophotometer. Autoinhibition in the photolysis of iodide ions

Article abstract of DOI:10.1039/c3dt53255k

The aqueous photoreactions of three halide ions (chloride, bromide and iodide) were studied using a diode array spectrophotometer to drive and detect the process at the same time. The concentration and pH dependences of the halogen formation rates were studied in detail. The experimental data were interpreted by improving earlier models where the cage complex of a halogen atom and an electron has a central role. The triiodide ion was shown to exert a strong inhibiting effect on the reaction sequence leading to its own formation. An assumed chemical reaction between the triiodide ion and the cage complex interpreted the strong autoinhibition effect. It is shown that there is a real danger of unwanted interference from the photoreactions of halide ions when halide salts are used as supporting electrolytes in spectrophotometric experiments using a relatively high intensity UV light source.

Full text of DOI:10.1039/c3dt53255k

Dalton
Transactions
View Article Online
View Journal | View Issue
PAPER
Aqueous photochemical reactions of chloride,
bromide, and iodide ions in a diode-array
spectrophotometer. Autoinhibition in the
photolysis of iodide ions†
Cite this: Dalton Trans., 2014, 43,
4862
József Kalmár, Éva Dóka, Gábor Lente* and István Fábián
The aqueous photoreactions of three halide ions (chloride, bromide and iodide) were studied using a
diode array spectrophotometer to drive and detect the process at the same time. The concentration and
pH dependences of the halogen formation rates were studied in detail. The experimental data were inter-
preted by improving earlier models where the cage complex of a halogen atom and an electron has a
central role. The triiodide ion was shown to exert a strong inhibiting effect on the reaction sequence
leading to its own formation. An assumed chemical reaction between the triiodide ion and the cage
complex interpreted the strong autoinhibition effect. It is shown that there is a real danger of unwanted
interference from the photoreactions of halide ions when halide salts are used as supporting electrolytes
in spectrophotometric experiments using a relatively high intensity UV light source.
Received 18th November 2013,
Accepted 6th January 2014
DOI: 10.1039/c3dt53255k
www.rsc.org/dalton
factor.9 Similarly, the photochemical formation of bromine
atoms from bromide ions on the surface of snow is understood
Introduction
The photochemistry of halide ions was reasonably well charac- to play a significant role in self-cleaning procedures in the
terized in the 1960s.1–8 It was shown that excited states have atmosphere.10,11 In addition, the photoinduced formation of
very short lifetimes, and a cage complex formed between the carcinogenic bromate ions in seawater or simply in the pres-
halide atom and an electron was assumed to be an intermedi- ence of a lower amount of chloride ions is a major concern
ate of the process. Overall, the photoreactions led to the for- for the water industry.12,13
mation of halogen and hydrogen gas.1–8 The dependence of
Diode array spectrophotometers are becoming increasingly
the quantum efficiencies on the pH was interpreted by assum- popular for a large number of different applications. Their
ing that the cage complex can react with hydrogen ions, in main advantage over traditional scanning spectrophotometers
addition to recombination with the original halide ion. The is that they can record an entire UV-vis spectrum in a very
photochemical reactions of the halides are not routinely con- short time. However, this comes with a cost: the sample is illu-
sidered as significant processes in nature or possible side reac- minated with a light beam that is sufficiently intense to trigger
tions in non-photochemical studies. Yet experiments reported photochemical reactions. There are several documented cases
in this paper will show that photochemical changes may very when neglecting this effect led to grossly erroneous con-
well occur under normal conditions and may seriously corrupt clusions, usually misidentifying a photochemical reaction as a
the interpretation of experimental data. This observation is thermal one.14–16 In a recent case, it was shown that even the
very important as halide salts, especially metal chlorides, are light beam in a stopped-flow instrument exerted a measurable
very frequently used as (supposedly) inert electrolytes to influence on the reaction of a highly absorbing manganese(III)
provide constant ionic strength in kinetic and equilibrium porphyrin complex.17 The photochemical effect of the light of
studies. Photochemical formation of chlorine from the chlor- diode array spectrophotometers is usually unwanted as the
ide ion in acidic atmospheric aerosols is a process of great instrument is intended to be used as a detector of a process
importance, and UV photolysis is certainly a contributing without influencing it. However, a number of examples have
shown that carefully planned studies can actually use this
effect for a highly quantitative study of photoreactions.16,18–26
These experiments sometimes afford mechanistic information
not available from classical photochemical experiments. The
limitation of the experiments is that the light source is necess-
University of Debrecen, Department of Inorganic and Analytical Chemistry,
Debrecen 10, H-4010, P.O.B. 21, Hungary. E-mail: lenteg@science.unideb.hu;
Fax: +36 52 518-660; Tel: +36 52 512-900/22373
†Electronic supplementary information (ESI) available: Additional figures
arily polychromatic, although its spectral properties can be
referred to in the main text. See DOI: 10.1039/c3dt53255k
4862 | Dalton Trans., 2014, 43, 4862–4870
This journal is © The Royal Society of Chemistry 2014

View Article Online
Dalton Transactions
Paper
changed within certain limits using suitable filters or external described in earlier publications.20,27 The reaction mixtures
light sources such as LEDs.26–28
were kept homogeneous during the photochemical experi-
This work will demonstrate how the light of a commercial ments by constant stirring with a 3 mm magnetic stirring rod
diode-array spectrophotometer can induce the formation of using the built-in magnetic stirrer of the cell compartment of
chlorine, bromine and iodine in the solutions of these halide the HP-8543 instrument. The geometry of the experimental
ions. Furthermore, by improving a methodology reported in setup was carefully tested and it was made sure that the stir-
earlier literature,18,19,25,26 we will also show how this un- ring rod never blocked the light beam. The light source was
expected effect can be utilized to obtain detailed information calibrated by ferrioxalate actinometry30,31 as well as reprodu-
about the nature of the photochemical processes. Evidence cing observations on the known photoreaction of 2,6-dichloro-
will also be shown of a strong autoinhibitory feature in the 1,4-benzoquinone.18 Nanosecond flash photolysis measure-
aqueous photoreaction of the iodide ion.
ments were carried out in an LKS.60 nanosecond transient
absorption spectrometer manufactured by Applied Photo-
physics Ltd. The instrument is equipped with a Quantel
Brilliant Nd:YAG laser (repetition frequency 10 Hz). The fourth
harmonic generator operating at 266 nm and Q-switching were
used in the present study (nominal energy: 40 mJ per pulse,
pulse length: 6 ns, beam diameter: 6 mm). The analyzing light
source was a 150 W ozone free xenon arc lamp (OSRAM 150W/
CR OFR) and its light beam entered a 1 × 1 cm fluorescence
quartz cuvette with cross-beam excitation. The analyzing light
beam hit close to the front face of the cuvette, to maximize the
detected concentration of transient species. A programmable
f/3.4 grating monochromator with a symmetrical Czerny–
Turner optical configuration was combined with a R928 photo-
multiplier for the transient signal detection at different wave-
lengths. Data points were collected by an Agilent Infiniium
digital storage oscilloscope (model number DSO8064A) with a
maximum sampling speed of 4 GSa s−1 (0.25 ns between data
points), bandwidth of 600 MHz, and output impedance of
50 Ω. Transient absorption kinetic curves were registered with
the control software of the LFP instrument, Pro-Data LKS
(version number 1.1.0.5), and visualized using Pro-Data Viewer
(version number 4.2.5). Pro-Data Viewer has a built-in curve
fitting application that was used in first order fitting. Other
fitting procedures were carried out using the software
Scientist32 and Matlab.33
Experimental
Materials
Perchloric acid (Spektrum 3D) was used as a strong acid to set
the pH. The ionic strength was set with NaClO4, which was
prepared as described earlier.29 All other chemicals, most
importantly hydroiodic acid (Sigma-Aldrich, unstabilized,
99.99%), hydrobromic acid (Fluka, alt.) and hydrochloric acid
(Scharlau, ar.), used in this study were of analytical reagent
grade and purchased from commercial sources. It should be
emphasized that the use of unstabilized hydroiodic acid is
imperative as the most common stabilized product contains
hypophosphorous acid which reacts with iodine readily. Pot-
assium iodide (Reanal), potassium bromide (Reanal) and pot-
assium chloride (Spektrum 3D) were also used as halide ion
sources. Special problems were encountered with the purity of
chlorides (see Results and discussion): the best chloride
source in this respect (i.e. the one that showed the smallest
interferences) was found to be high-purity concentrated hydro-
chloric acid, which was used in most of the experiments
reported here. Doubly deionized and ultrafiltered water from a
Millipore Q system was used to prepare the stock solutions
and samples.
Instrumentation and computation
UV-vis spectra were recorded on an HP-8543 diode-array or a
Perkin Elmer Lambda 2S scanning spectrophotometer. Con-
stant temperature at 25.0 0.1 °C was maintained with a built-
Results
Basic observations
in Hewlett-Packard 89090A Peltier thermostat or an external When a solution of a hydrogen halide is kept in a diode-array
circulating thermal bath. In some experiments, 0.01 mm path spectrophotometer, the formation of halogens or trihalide
length cuvettes were used. This posed a special problem of ions is detectable. No similar changes are observed if the same
path length reproducibility, which was solved by developing a solution is kept in a scanning spectrophotometer, which
spectrophotometric method not requiring the exact knowledge clearly proves the purely photochemical origin of the chemical
of path length (see Results and discussion). A YSI 5100 Dis- reactions.16,25 A series of time-resolved spectra detected during
solved Oxygen Meter and a YSI Model 5239 probe with YSI such a process in a pure HCl solution is shown in Fig. 1.
5906 membrane cap were used for measuring the concen- Fig. S1 in ESI† shows the corresponding kinetic trace at
tration of dissolved oxygen.
(Medicor, Hungary, 500 W) was used as a UV light source in in HBr and HI solutions under similar conditions.
the photochemical experiments using the oxygen electrode. The photochemical nature of the observed process can be
A high-power quartz lamp 230 nm. Fig. S2 and S3 in ESI† show spectral changes detected
Quantitative kinetic measurements on the photochemical reac- further demonstrated by the use of filters, which can be
tions were performed with the HP-8543 diode-array instrument inserted into the light beam of the diode-array spectrophoto-
in a stoppered standard quartz cuvette (optical path length: meter. Fig. 2 shows such a pair of conclusive experiments in a
1.000 cm) using the method and general operating protocols solution of HBr. The filter in this case was a 1 mm quartz
This journal is © The Royal Society of Chemistry 2014
Dalton Trans., 2014, 43, 4862–4870 | 4863

View Article Online
Paper
Dalton Transactions
photochemical reaction proceeded in this case as evidenced by
the absorbance change. In curve b, the only difference relative
to experiment a was that the filter was placed between the light
source and the sample. Therefore, light that could be absorbed
by Br− did not reach the sample in this case. As a consequence,
no reaction occurred. These observations demonstrate that the
light of the diode-array spectrophotometer in fact drives the
photolysis of halogen ions.
Although earlier measurements clearly showed the for-
mation of hydrogen in the aqueous photochemical reactions
of halide ions,1–8 some additional efforts were made in this
work to show that dissolved oxygen does not play a significant
role in the overall reaction. Decreasing the initial concen-
tration of dissolved oxygen by bubbling with argon before
experiments did not influence any of the experimental
observations. Furthermore, in a few trial experiments, the
concentration of dissolved oxygen was monitored during
illumination. The result of such an experiment is shown in
Fig. S5 in ESI.† No evidence of any loss of dissolved oxygen
attributable to the photoreaction was seen.
Fig. 1 Spectral changes in a HCl solution upon continuous UV illumina-
tion. [HCl] = 5.056 M, T = 25.0 °C.
Determination of trihalide ion stability constants. The for-
mation equilibrium constants of trihalide ions (eqn (1)) have
been reported in previous studies.34,35
½Xꢀ3 ꢁ
X2 þ Xꢀ ¼ Xꢀ3 KX
¼
ð1Þ
½Xꢀꢁ½X2ꢁ
For Br− and I−, we determined these constants by spectro-
photometric titration in this work (Table 1); these values
compare well with literature data, especially if the differences
in the experimental conditions are taken into account.34,35 For
Cl−, KCl is too low to be determined by the same method as
the solubilities of chloride salts are insufficient to obtain high
enough free chloride ion concentrations. This fact led us to
conclude that KCl < 1 M−1. Consequently, a literature value
(0.194 M−1) was used, which was obtained with a different
method34 and did not contradict our estimate for the higher
Fig. 2 Kinetic curves during the photolysis of a HBr solution with the
filter after (a) or before (b) the cell. [HBr] = 0.322 M, T = 25.0 °C.
cuvette filled with pure chloroform. The absorption of this estimate. The same method also yielded the molar absorption
filter overlaps with the absorption of a bromide ion (both
absorb significantly below 250 nm) in such a way that it
coefficients of trihalide ions needed in the calculation. The
triiodide ion, however, presented a somewhat unexpected
almost completely shuts out light absorbed by Br− if it is problem. To evaluate the HI photolysis data on a time scale of
inserted before the sample (Fig. S4 in ESI†). The main absorb-
hours, it was necessary to determine the molar absorption
coefficients of I3 in the entire spectral region from 200 to
−
ing product of the photochemical reaction, the tribromide ion,
can be detected even with the use of the filter because Br3− has 500 nm. This is not a trivial problem because I3− can only exist
at high free I− concentrations, which makes the molar absorp-
absorption above 250 nm. In trace a of Fig. 2, the filter was
inserted into the light beam following the cell. Thus the intense
unfiltered light went through the sample first and the filter only iodide ion absorbs strongly (200–260 nm). Our solution to this
tion coefficient determination difficult in the region where the
blocked some of the light from reaching the detector. The
problem relied on the use of cuvettes with very short path
Table 1 Equilibrium and spectral data for trihalide ions. T = 25.0 °C, I = 1.0 M
−
−
−
Cl3
Br3
I3
KX (M−1
)
—
9.7 1.9
1935
(6.7 0.2) × 102
74035
0.19434
λmax (nm)
230
270
288
352
ε (λmax) (M−1 cm−1
)
(1.50 0.02) × 104
(4.4 0.4) × 104
(2.03 0.02) × 104
(1.33 0.02) × 104
4864 | Dalton Trans., 2014, 43, 4862–4870
This journal is © The Royal Society of Chemistry 2014

View Article Online
Dalton Transactions
Paper
lengths (0.01 mm). Unfortunately, the reproducibility of the
path length between successive experiments is not satisfactory
for these special cuvettes: our experiments showed that the
path length may change within a factor of two. One possible
reason is that small dust particles that cannot be completely
washed away from the surface of the cuvette prevent the two
quartz sheets from perfectly fitting onto each other every time.
This problem was overcome by selecting a reference wave-
length where only I3− absorbs and the molar absorption coeffi-
cient is measurable independently of the irreproducible path
lengths. Using the absorbance measured at this wavelength,
spectra can be normalized and the method can be made inde-
pendent of the path length. A simple mathematical derivation
based on Beer’s law gives the following formula:
Fig. 3 Reaction rate as a function of illumination ratio for the photolysis
ð2Þ of HBr solutions. [HBr] = 0.500 M, T = 25.0 °C.
½Iꢀꢁ
εI ðλÞ
ꢀ
ꢀ
AðλÞ
εI ðλÞ
3
¼
ꢂ
þ
AðrefÞ εI ðrefÞ ½I3ꢀꢁ εI ðrefÞ
ꢀ
ꢀ
3
3
ꢀ
In this equation, every variable except εI ðλÞ is known or
3
ratio, which clearly implies that the reaction does not progress
any further in the absence of light. Therefore, it is unlikely to
be a chain reaction.
The chloride ion presented particular problems related to
purity in this study. Kinetic curves detected during the photo-
lysis of Cl− featured an initial region where the rate of
product formation slightly accelerated (Fig. S1 in ESI† shows
measured directly, so the unknown value can be computed.
The spectra of I− and I3 are shown in Fig. S6 in ESI.† Even
−
with this special technique, reliable molar absorption coeffi-
cients could not be obtained in a fairly narrow region from 240
to 264 nm primarily because I3 does not absorb significant
−
amounts of light relative to other species present in the solu-
tion under any conditions. For Fig. S6,† values in this region
were extrapolated by assuming that the overall spectrum of I3
was a combination of Gaussian curves. It should also be men-
tioned that these values do not affect the calculations for the
photochemical reactions in HI solutions in any way because
the absorption of the triiodide ion is insignificantly low at
these wavelengths.
Quantitative observations. Quantitative information on the
photochemical reactions was obtained by the initial rate
method. The main advantage of this method is that the initial
change in the absorbance is free from any possible effect the
product might cause. The initial rate was calculated from the
initial slope of absorbance (A) change measured at a wave-
such an effect). This is most probably due to the fact that the
−
starting materials contain uncontrollably low levels of organic
impurities, which consume some of the Cl2 formed at the
beginning of the process. This explanation was consistent with
the fact that the length of the initial, slightly accelerating
phase varied depending on the source of chemicals, and could
be reduced by re-crystallizing solid starting materials or
decreasing the initial concentrations. The example shown in
Fig. S1 in ESI† is the worst in this respect because the chloride
concentration is very high there (ca. 5 M). In these cases, the
maximum rate was used instead of the literal initial rate for
quantitative purposes. Similar impurities might also be
present in Br− and I− sources, but they understandably consti-
tute less of a problem because the product halogens probably
react with organic contaminations at a much lower rate.
−
length where only X3 absorbs and using the molar absorptiv-
ities (ε) of the trihalide ions.
ꢀ
ꢁ
The initial rate of the photochemical reaction depended on
the pH. In all cases, the rate increased with decreasing pH. To
display the three studied halide ions in a single graph, the
rates were normalized before plotting them in Fig. 4. The
experimental curves could be fitted to the following equation:
d½X2ꢁ dAðλÞ
1
1
v ¼
¼
ꢂ
ꢂ
þ 1
ð3Þ
dt
dt
εX ðλÞ
½XꢀꢁKX
ꢀ
3
Continuous and interrupted illumination methods19,25
have both been used in this study. In the latter method,
periods of illumination are regularly interrupted by dark
periods of various lengths by closing the shutter of the spectro-
photometer. The spectrophotometer can only measure the
absorbance during illuminated periods, and the initial rate
vi½Hþꢁ þ vf a
v ¼
ð4Þ
½Hþꢁ þ a
The determined values of parameters vi, vf and a have been
measured under these circumstances is the average rate occur- presented in Table 2 for the three studied halide ions.
ring during illuminated and dark periods. The dependence of Hydrogen iodide photolysis on a long time scale. The
the reaction rate on the illumination ratio (which is defined as photoreaction of hydroiodic acid could be conveniently fol-
the length of illumination time relative to the overall time lowed on somewhat longer time scales. This was not possible
passed) may yield information on reactions occurring in the in the case of Cl− and Br− because the halogen product is
dark period.19 Such dependence is shown for a HBr solution much more volatile in these cases and eqn (1) is much less
in Fig. 3. The rate is directly proportional to the illumination shifted toward the formation of the trihalide ion. The kinetic
This journal is © The Royal Society of Chemistry 2014
Dalton Trans., 2014, 43, 4862–4870 | 4865

View Article Online
Paper
Dalton Transactions
Fig. 4 Normalized reaction rate as a function of pH for the photolysis
of hydrogen halide solutions [Cl−] = 1.000 M, [Br−] = 0.500 M, [I−] =
1.00 mM, T = 25.0 °C.
Fig. 6 Rate as a function of the (relative) number of photons absorbed
by iodide ions during the photolysis of a HI solution. [HI] = 2.00 mM,
T = 25.0 °C. Dotted line: expected line for the inner filter effect only.
Table 2 Kinetic parameters for the photoreactions of halide ions.
experiments, the photon flux actually absorbed by iodide ions
was calculated using the known absorption characteristics of
I− and I3−, and the emission properties of the light source
using the following equation:
T = 25.0 °C, I = 1.0 M, [Cl−] = 1.000 M, [Br−] = 0.500 M, [I−] = 1.00 × 10−3
M
Cl−
Br−
I−
vi (M s−1
vf (M s−1
a (M)
)
)
(2.8 0.6) × 10−7 (1.84 0.03) × 10−8 (2.22 0.05) × 10−4
ð
λmax
(2.0 0.3) × 10−8 (7 1) × 10−10
(9 1) × 10−4
(3.9 0.2) × 10−5
(4.9 0.4) × 10−5
½I ꢁεI ðλÞ
ꢀ
ꢀ
Nrel
¼
ꢂ ΦPλ ꢂ ð1 ꢀ 10ꢀAðλÞÞdλ ð5Þ
4
1
ꢀ
½I ꢁεI ðλÞ þ ½I3ꢀꢁεI ðλÞ
ꢀ
ꢀ
λmin
3
In this equation, ΦPλ is the spectral photon flux. In the
actual calculations, the integration in eqn (5) is replaced by
summation with a base interval of 1 nm, which is set by the
spectral resolution of the diode array spectrophotometer.
Fig. 6 shows the detected reaction rate as a function of the
number of absorbed photons by the iodide ion. Instead of the
expected direct proportionality, a sharp drop is seen. There-
fore, the drop in rate cannot be simply caused by the inner
filter effect of the product triiodide ion. It follows that I3− must
exert some sort of strong chemical inhibitory effect on the
photoreaction.
For a more insightful analysis of the data, the normalized
rate (vadj) of reaction was defined during the iodide photoreac-
tion. This normalizes the reaction rate to the photon flux
absorbed by the iodide ion specifically, and therefore takes the
inner filter effect into account (vadj = v/Nrel). Fig. 7 displays the
normalized rate as a function of triiodide ion concentration.
The data collected at constant hydrogen ion concentration
could be approximated by the following function:
curve of a long experiment with HI is shown in Fig. 5. It is
seen that the rate of I3 formation decreases gradually. This is
−
not as easily interpreted as could be thought at first sight. The
drop in rate does not occur because of the depletion of the
reactants, as the conversion is only 2% at the end of the time
scale depicted in Fig. 5 (after two hours). An alternative expla-
nation could be the inner filter effect of the absorbing product
I3−. In an independent series of experiments using attenuating
filters inserted into the light beam, it was made sure that the
initial rate is directly proportional to the photon flux36
absorbed by iodide ions (Fig. S7 in ESI†). In the long-term
x
vcor
¼
ð6Þ
y þ ½I3ꢀꢁ
The parameters x = (4.87 0.03) × 10−13 M2 s−1 and y = (2.2
0.1) × 10−6 M have been determined in this fitting procedure.
Their physical meaning will be explained in the discussion.
Laser flash photolysis study of the reactions of iodine atoms in
water
As the halogen atoms are very likely to play an important role
during the photolysis of halide ions,1–8 several measurements
Fig. 5 Kinetic curves detected during the photolysis of a HI solution on
a long time scale. [HI] = 2.00 mM, T = 25.0 °C.
4866 | Dalton Trans., 2014, 43, 4862–4870
This journal is © The Royal Society of Chemistry 2014

View Article Online
Dalton Transactions
Paper
Fig. 8 Typical kinetic curve measured in a laser flash photolysis experi-
ment on a split time base. Markers: experimental points. Lines: double
exponential fit (short time scale), second-order fit (long time scale). [I−]
= 0.50 mM, [S2O82−] = 0.10 M, pH = 7 (unbuffered), T = 25.0 °C.
Fig. 7 Normalized reaction rate as a function of the triiodide ion con-
centration during the photolysis of a HI solution. [HI] = 2.00 mM,
T = 25.0 °C. Straight line: best fit to eqn (6).
were carried out by laser flash photolysis to validate some
assumptions about their reactions. As the results obtained
during the long time scale measurements of hydrogen iodide
clearly imply that the product triiodide ion seriously interferes
with the photochemical reactions, a method not dependent on
the direct photolysis of halide ions was considered. It was
found that the reactions of iodine atoms could be favorably
studied in experiments where sulfate ion radicals were used to
prepare iodine atoms. A sulfate ion radical was generated
through the photolysis of a peroxodisulfate ion initiated by a
laser pulse at 266 nm:
2ꢀ
ꢀ
S2O8 ! 2SO4
ð7Þ
The sulfate ion radical then reacts with an iodide ion to
produce an iodine atom:
Fig. 9 Pseudo-first order rate constants determined in the double
exponential fits as a function of iodide ion concentration from the laser
flash photolysis experiments. [S2O82−] = 0.10 M, pH = 7 (unbuffered),
T = 25.0 °C.
SO4 þ Iꢀ ! SO42ꢀ þ I•
ð8Þ
ꢀ•
A sample kinetic curve detected at 380 nm is shown in
Fig. 8. Iodine atoms give iodine molecule ions in a subsequent
reaction with excess iodide ions in a reversible process:
order rate constants is shown in Fig. 9. It is seen that the
higher pseudo-first order rate constants depend linearly on the
concentration of iodide ions with a slope of k+ = (1.19 0.05) ×
1010 M−1 s−1 and intercept of k− = (6.8 0.3) × 106 s−1. The
lower pseudo-first order rate constants, on the other hand, are
directly proportional to the concentration of iodide ions with a
slope of k+ = (5.74 0.07) × 109 M−1 s−1. It is well known that
reaction (9) is reversible,38 and the ratio of the slope and inter-
cept k+/k− = (1.74 0.15) × 103 M−1 is in agreement with earlier
estimates of the equilibrium constant of the process obtained
in direct photolysis studies of iodide ions.37–43 Therefore, this
double exponential curve must represent the somewhat
perhaps unintuitive, but theoretically well understood, case
when the larger pseudo-first order rate constant belongs to the
second process in a reaction sequence.44
ꢀ•
I• þ Iꢀ ¼ I2
ð9Þ
The product of this reaction, the iodine molecule ion (I2−•),
is known to absorb strongly at 380 nm.37,38 The production of
−•
I2 is the reason for the increase of absorbance up to 1 ms in
Fig. 8. Finally, iodine molecule ions or iodine atoms recombine
to form iodine (I2) or triiodide ions (I3−). This is in agreement
with the observations on the longer time scale shown in Fig. 8.
As the iodide ion was used in large excess during the experi-
ments, its concentration could be considered constant in a
single experiment. The kinetic curve on the short time scale
(up to 1 μs) in Fig. 8 clearly cannot be fitted to a single expo-
nential expression, though. This is because reactions (8) and
(9) have similar time scales. Experimental curves were fitted
successfully with double exponential functions. The iodide ion
concentration dependence of the corresponding pseudo-first
In the case of chloride and bromide ions, the reactions of
sulfate ion radicals with the halide ions were too slow
This journal is © The Royal Society of Chemistry 2014
Dalton Trans., 2014, 43, 4862–4870 | 4867

View Article Online
Paper
Dalton Transactions
halide ion (see eqn (3) by substituting zero as the trihalide ion
concentration). Then this cage complex can undergo either
secondary recombination or chemical reaction with the solvent
water or hydrogen ion to produce a halogen atom and a hydro-
gen atom:
ðXeꢀÞ ! Xꢀ
ð11Þ
v ¼ krec½ðX ꢃ eꢀÞꢁ
ðX ꢃ eꢀÞ þ H2O ! X þ H þ OHꢀ
ð12Þ
v ¼ kwat½ðX ꢃ eꢀÞꢁ
ðXeꢀÞ þ Hþ ! X þ H
ð13Þ
v ¼ khyd½ðX ꢃ eꢀÞꢁ½Hþꢁ
Fig. 10 Spectral changes in a sulfur(IV) solution acidified by HCl in a
diode array spectrophotometer. ΔA(H2SO4): absorbance change in the
photooxidation of sulfur(IV) measured in sulfuric acid. [HCl] = 0.0479 M,
[S(IV)] = 2.02 M, [O2] = 0.2 mM reaction time: 280 s.
Hydrogen atoms and halogen atoms can then engage in
further reactions, eventually resulting in the formation of
hydrogen gas and elementary halogen, the latter in equili-
brium with the trihalide ion in the presence of excess halide:
compared to the formation of the halogen anions (X2−•); there-
fore no information on reactions analogous to the iodide case
could be obtained.
2H ! … ! H2
2X ! … ! X2
ð14Þ
ð15Þ
Effect of hydrochloric acid in the sulfur(IV) system
No rate expressions are shown for reactions (14) and (15) as
they are after the rate determining step. Eqn (10)–(15) essen-
tially give the scheme that was proposed in earlier studies.1–8
It is possible to devise more elaborate series of reactions
instead of eqn (15):
Kinetic phenomena in the uncatalyzed autoxidation of
sulfur(IV) were used to demonstrate how neglecting the
aqueous photochemistry of halide ions may lead to erroneous
conclusions. The autoxidation of sulfur(IV) does not proceed
in the absence of
a catalyst and light even at high
ꢀ•
X • þ Xꢀ Ð X2
ð16Þ
ð17Þ
temperatures.45–47 UV-light, however, drives the reaction.20 In
the earlier study of this process,20 sulfuric acid and perchloric
acid were used to set the acidity. Fig. 10 shows an experiment
in hydrochloric acid. The loss of sulfur(IV) is much faster in
hydrochloric acid than the expectations based on the results in
perchloric and sulfuric acids20 (which is shown by a small
arrow in Fig. 10). The product of the photoreaction of chloride
ions, chlorine, is known to oxidize sulfur(IV) quite rapidly.48
Therefore, two parallel photochemical reactions proceed in
hydrochloric acid, and both lead to the same overall reaction
because of subsequent thermally activated reactions.
2X2 ! Xꢀ þ X3
ꢀ•
ꢀ
These assumptions are in agreement with the chemical
knowledge about the properties of halogen atoms.37–43
Although reactions (16) and (17) occur after the rate determin-
ing step in the photochemical process, their existence was con-
firmed and rate constants measured by a laser flash photolysis
study in the case of the iodide ion.
Kinetic derivations from this scheme with a steady-state
assumption for species (X·e−) show that the initial overall rate
of halogen formation, as defined in eqn (3), can be given as
ð
khyd½Hþꢁ þ kwat
λmax
Discussion
v ¼
q
ΦPλ ꢂ ð1 ꢀ 10ꢀAðλÞÞdλ ð18Þ
khyd½Hþꢁ þ kwat þ krec
λmin
Our experimental data can be interpreted by improving the
scheme proposed earlier by direct photochemical measure-
ments in aqueous solutions containing halide ions.1–8 In the
first step, a halide ion forms a halogen atom–electron solvent
cage complex upon illumination:
This equation is parametrically equivalent to the experi-
mentally found expression (4) with the following connections
between experimentally determined and theoretical para-
meters:
ð
X
ꢀ þ hν ! ðX ꢃ eꢀÞ
λmax
ð
λmax
vi ¼ q
ΦPλ ꢂ ð1 ꢀ 10ꢀAðλÞÞdλ
ð19Þ
ð20Þ
ð10Þ
ΦPλ ꢂ ð1 ꢀ 10ꢀAðλÞÞdλ
λmin
v ¼ q
λmin
ð
λmax
kwat
kwat þ krec
Here, q is the quantum yield of the process, whereas the
remaining term gives the photon flux actually absorbed by the
vf ¼
q
ΦPλ ꢂ ð1 ꢀ 10ꢀAðλÞÞdλ
λmin
4868 | Dalton Trans., 2014, 43, 4862–4870
This journal is © The Royal Society of Chemistry 2014

View Article Online
Dalton Transactions
Paper
kwat þ krec
Acknowledgements
ð21Þ
a ¼
khyd
The authors thank the Hungarian Science Foundation for
financial support under grant no. K77936 and NK 105156. The
research was supported by the EU and co-financed by the Euro-
pean Social Fund under the project ENVIKUT (TÁMOP-4.2.2.
A-11/1/KONV-2012-0043). The Hungarian Academy of Sciences
also supported these studies through funding from Research
Group on Homogeneous Catalysis and Reaction Mechanisms.
Éva Dóka’s participation in this research was supported by the
European Union and the State of Hungary, co-financed by the
European Social Fund in the framework of TÁMOP-4.2.4.A/
2-11/1-2012-0001 ‘National Excellence Program’. Some experi-
mental help was kindly provided by Dr Ildikó Kerezsi, whose
assistance is gratefully acknowledged.
However, as already pointed out, the observations on the
photolysis of hydrogen iodide solutions and the dependence
of the overall rate on the concentration of the triiodide ion
shown in eqn (6) strongly support some interaction between
the triiodide ion and one of the reactive intermediates in the
system. We found that the simplest way to interpret this effect
is assuming the interaction of a triiodide ion with the cage
complex formed as the product of reaction (10):
ðI ꢃ eꢀÞ þ I3 ! I2 þ 2Iꢀ
ꢀ
ð22Þ
v ¼ ktri½ðI ꢃ eꢀÞꢁ½I3ꢀꢁ
This additional steps and the inner filter effect of the tri-
iodide ion give rise to a more complicated rate law:
ð
khyd½Hþꢁ þ kwat
λmax
ꢀ
ꢀ
½I ꢁεI ðλÞ
v ¼
q
khyd½Hþꢁ þ kwat þ krec þ ktri½I3ꢀꢁ
ꢀ
½I ꢁεI ðλÞ þ ½I3ꢀꢁεI ðλÞ
References
ꢀ
ꢀ
λmin
3
ꢂ ΦPλ ꢂ ð1 ꢀ 10ꢀAðλÞÞdλ
1 J. Jortner, M. Ottolenghi and G. Stein, J. Phys. Chem., 1961,
65, 1232–1238.
2 E. Hayon, J. Phys. Chem., 1961, 65, 1937–1940.
3 J. Jortner, R. Levine, M. Ottolenghi and G. Stein, J. Phys.
Chem., 1961, 65, 2029–2036.
ð23Þ
This theoretically derived rate expression is in full agree-
ment with the experimental observations summarized in
eqn (4) and (6).
4 J. Jortner, M. Ottolenghi and G. Stein, J. Phys. Chem., 1962,
66, 2029–2036.
5 J. Jortner, M. Ottolenghi and G. Stein, J. Phys. Chem., 1962,
66, 2037–2042.
6 J. Jortner, M. Ottolenghi and G. Stein, J. Phys. Chem., 1962,
66, 2042–2045.
Conclusion
The results clearly prove that the photoreactions of halide ions
can be initiated by the light of a diode array spectrophoto-
meter. The rates of halogen and trihalide ion formation can be
determined and their dependence on the concentrations of
reactants and light intensity can serve as a basis for postulat-
ing a mechanism. The systematic study of the photoreaction of
7 J. Jortner, M. Ottolenghi and G. Stein, J. Phys. Chem., 1963,
67, 1271–1274.
8 J. Jortner, M. Ottolenghi and G. Stein, J. Phys. Chem., 1964,
68, 247–255.
9 J. M. Roberts, H. D. Osthoff, S. S. Brown and
A. R. Ravishankara, Science, 2008, 321, 1059–1059.
the iodide ion has revealed that the triiodide ion has a specific 10 V. Potery, M. Fárník, P. Slavíček, U. Buck and V. V. Kresin,
(auto)inhibiting effect on the process. This inhibiting effect J. Chem. Phys., 2007, 126, 071101.
can be interpreted by a single reaction occurring between the 11 K. A. Pratt, K. D. Custard, P. B. Shepson, T. A. Douglas,
triiodide ion and the aqueous cage complex of the iodine atom
and an electron. It is also clear that these processes can be a
source of unwanted interference in cases when halide ions
D. Pöhler, S. General, J. Zielcke, W. R. Simpson, U. Platt,
D. J. Tanner, L. G. Huey, M. Carlsen and B. H. Stirm, Nat.
Geosci., 2013, 6, 351–356.
are used as a background electrolyte in spectrophotometric 12 D. L. Macalady, J. H. Carpenter and C. A. Moore, Science,
studies using a diode array spectrophotometer, as proven by 1977, 195, 1335–1337.
the example of the sulfur(IV)–oxygen system in this work, 13 R. P. Schwarzenbach, B. I. Escher, K. Fenner,
where using aqueous sodium chloride as a medium greatly
changes the observations. It should be emphasized that the
T. B. Hofstetter, C. A. Johnson, U. von Gunten and
B. Wehrli, Science, 2006, 313, 1072–1077.
possible interference is independent of the monitoring wave- 14 D. M. Stanbury and J. N. Figlar, Coord. Chem. Rev., 1999,
length range, and it depends on the light source of the spectro- 187, 223–232.
photometer. For the chloride ion, any illumination below 15 A. P. Oliveira and R. B. Faria, J. Am. Chem. Soc., 2005, 127,
235 nm could cause such an effect, whereas the same cutoff 18022–18023.
value is about 300 nm for both bromide and iodide ions. The 16 M. Galajda, G. Lente and I. Fábián, J. Am. Chem. Soc., 2007,
exact details depend on the particular concentrations used in 129, 7738–7739.
the experiments; detailed calculations or – even better – direct 17 J. Kalmár, B. Biri, G. Lente, I. Bányai, A. Budimir, M. Biruš,
experiments are needed to rule out such possible
interferences.
I. Batinić-Haberlee and I. Fábián, Dalton Trans., 2012, 41,
11875–11884.
This journal is © The Royal Society of Chemistry 2014
Dalton Trans., 2014, 43, 4862–4870 | 4869

View Article Online
Paper
Dalton Transactions
18 G. Lente and J. H. Espenson, J. Photochem. Photobiol., A, 33 MatLab for Windows, version 4.2c1, The Mathworks, Inc.,
2004, 163, 249–258. Natick, MA, USA, 1994.
19 I. Kerezsi, G. Lente and I. Fábián, J. Am. Chem. Soc., 2005, 34 M. S. Sherill and E. F. J. Izard, J. Am. Chem. Soc., 1931, 53,
127, 4785–4793. 1667–1674.
20 I. Kerezsi, G. Lente and I. Fábián, Dalton Trans., 2006, 955– 35 I. Lengyel, J. Li, K. Kustin and I. Epstein, J. Am. Chem. Soc.,
960. 1996, 118, 3708–3719 and references therein.
21 I. Kerezsi, G. Lente and I. Fábián, Inorg. Chem., 2007, 46, 36 S. E. Braslavsky, Pure Appl. Chem., 2007, 79, 293–465.
4230–4238.
37 L. I. Grossweiner and M. S. Matheson, J. Chem. Phys., 1955,
23, 2443–2444.
22 G. Lente, J. Kalmár, Z. Baranyai, A. Kun, I. Kék, D. Bajusz,
M. Takács, L. Veres and I. Fábián, Inorg. Chem., 2009, 48, 38 L. I. Grossweiner and M. S. Matheson, J. Phys. Chem., 1957,
1763–1773. 61, 1089–1095.
23 A. Filáry and A. K. Horváth, Phys. Chem. Chem. Phys., 2010, 39 J. H. Baxendalpe, P. L. T. Bevan and D. A. Stott, Trans.
12, 6742–6749. Faraday Soc., 1968, 64, 2389–2397.
24 T. Zidki, H. Cohen and D. Meyerstein, Phys. Chem. Chem. 40 P. F. de Violet, R. Bonneau and J. Joussot-Dubien, Chem.
Phys., 2010, 12, 12862–12867. Phys. Lett., 1973, 19, 251–253.
25 I. Fábián and G. Lente, Pure Appl. Chem., 2010, 82, 1957– 41 P. F. de Violet, R. Bonneau and S. R. Logan, J. Phys. Chem.,
1973. 1974, 78, 1698–1701.
26 M. Gombár, É. Józsa, M. Braun and K. Ősz, Photochem. 42 P. Neta, R. E. Huie and A. B. Ross, J. Phys. Chem. Ref. Data,
Photobiol. Sci., 2012, 11, 1592–1595. 1988, 17, 1027–1284.
27 O. Shvydkiv, A. Yavorskyy, K. Nolan, A. Youssef, E. Riguet, 43 J. M. Gardner, M. Abrahamsson, B. H. Farnum and
N. Hoffmann and M. Oelgemöller, Photochem. Photobiol.
Sci., 2010, 9, 1601–1603.
28 U. Megerle, R. Lechner, B. König and E. Riedle, Photochem.
Photobiol. Sci., 2010, 9, 1400–1406.
G. J. Meyer, J. Am. Chem. Soc., 2009, 131, 16206–16214.
44 J. H. Espenson, in Chemical Kinetics and Reaction
Mechanisms, McGraw-Hill, New York, 2nd edn, 1995,
pp. 75–76.
29 G. Gordon and P. H. Tewari, J. Phys. Chem., 1966, 70, 200– 45 C. Brandt and R. van Eldik, Chem. Rev., 1995, 95, 119–190
204. and references therein.
30 C. G. Hatchard and C. A. Parker, Proc. R. Soc. London, Ser. 46 C. Brandt, I. Fábián and R. van Eldik, Inorg. Chem., 1994,
A, 1956, 235, 518–536. 33, 687–701 and references therein.
31 T. Lehóczki, É. Józsa and K. Ősz, J. Photochem. Photobiol., A, 47 G. Lente and I. Fábián, Inorg. Chem., 2004, 43, 4019–
2013, 251, 63–68. 4025.
32 SCIENTIST, version 2.0, Micromath Software, Salt Lake City, 48 K. D. Fogelman, D. M. Walker and D. W. Margerum, Inorg.
UT, USA, 1995.
Chem., 1989, 28, 986–993.
4870 | Dalton Trans., 2014, 43, 4862–4870
This journal is © The Royal Society of Chemistry 2014