Kinetic study of the reactions of chlorine atoms and Cl2·- radical anions in aqueous : Solutions. 1. Reaction with benzene

Article abstract of DOI:10.1021/jp9929768

The photolysis of Na_S2O₈ aqueous solutions containing Cl^- ions is a clean technique for kinetic studies of the species Cl^./Cl₂^.- in the presence and absence of added aromatic substrates. Laser and conventional flash-photolysis methods were used to study the aqueous phase reactions of chlorine atoms and Cl₂^.- (340 nm) radical ions in the presence and absence of benzene. A mechanism that considers the decay of Cl₂^.- in aqueous solutions with chloride ion concentrations from 1 x 10^-4 to 0.6 M, total radical (Cl^. + Cl₂^.-) concentrations at (0.1-1.5) x 10^-5 M, and 2.5-3 pH was proposed. Kinetic computer simulations supported interpretation of the experimental data. The rate constants 6 x 10⁹/M-sec ≤ k ≤ 1.2 x 10¹⁰/M-sec and < 1 x 10⁵/M-sec were determined for the reactions of Cl^. and Cl₂^.-, respectively, in the aqueous phase. The organic radicals produced from these reactions showed an absorption band with maximum at 300 nm that was assigned to a Cl-cyclohexadienyl radical (Cl-CHD). The kinetic analysis of the traces supported a reversible reaction between Cl-CHD and O₂. A reaction mechanism leading to the formation of chlorobenzene was proposed.

Full text of DOI:10.1021/jp9929768

J. Phys. Chem. A 2000, 104, 3117-3125
3117
•-
Kinetic Study of the Reactions of Chlorine Atoms and Cl2 Radical Anions in Aqueous
Solutions. 1. Reaction with Benzene
Mar´ıa L. Alegre, Mariana Gerone´s, Janina A. Rosso, Sonia G. Bertolotti,† Andre´ M. Braun,‡
Daniel O. Ma´rtire,* and Mo´nica C. Gonzalez*
Instituto de InVestigaciones Fisicoqu´ımicas Teo´ricas y Aplicadas (INIFTA), Facultad de Ciencias Exactas,
UniVersidad Nacional de La Plata, Casilla de Correo 16, Sucursal 4, (1900) La Plata, Argentina
ReceiVed: August 23, 1999; In Final Form: December 16, 1999
The photolysis of Na2S2O8 aqueous solutions containing Cl- ions is a clean method for kinetic studies of the
•-
species Cl•/ Cl2 in the absence and presence of added aromatic substrates. Laser and conventional flash-
•-
photolysis techniques were employed to investigate the aqueous phase reactions of chlorine atoms and Cl2
(340 nm) radical ions in the presence and absence of benzene. A mechanism is proposed which accounts for
•-
the decay of Cl2 in aqueous solutions containing chloride ion concentrations in the range 1 × 10-4 to 0.6
M, total radical (Cl• + Cl2•-) concentrations in the range (0.1-1.5) × 10-5 M, and pH in the range 2.5-3.0.
Interpretation of the experimental data is supported by kinetic computer simulations. The rate constants 6 ×
109 M-1 s-1 e k e 1.2 × 1010 M-1 s-1 and < 1 × 105 M-1 s-1 were determined for the reactions of Cl• and
•-
Cl2 with benzene, respectively, in the aqueous phase. The organic radicals produced from these reactions
exhibit an absorption band with maximum at 300 nm, which was assigned to a Cl-cyclohexadienyl radical
(Cl-CHD). The kinetic analysis of the traces supports a reversible reaction between O2 and Cl-CHD. A
reaction mechanism leading to the formation of chlorobenzene is proposed.
Introduction
as shown in eq 1. The detailed mechanism is still under
discussion,2,7 because it cannot account for the observed
Chloride is one of the most abundant anions in the tropo-
spheric aqueous phase. In marine clouds, typical concentrations
are in the order of 0.5 mM and higher concentrations can be
found in aerosols.1 HO• and SO4•- radicals, also present in cloud
droplets and aerosols, have been reported to react with chloride
experimental dependence of the Cl2•- decay on Cl- concentra-
tion. Therefore, aqueous phase reactions of Cl• and Cl2•- need
further investigations before Cl• reactions with constituents of
the troposphere may be safely neglected.
•-
•-
Cl• + Cl- a Cl2
ions yielding chlorine atoms, which are in equilibrium with Cl2
radical ions.2,3 It is now apparent that reactions involving
nitrogen oxide species and NaCl yielding chlorine atoms may
also occur.1c Therefore, oxidation of chloride may represent an
important sink of strong oxidants in tropospheric multiphase
+H2O
V
-Cl-
Cl• + H2O a HOClH•-+ H+ a HO• + Cl•
(1)
•-
systems, depending on the final fate of the Cl• atoms and Cl2
radicals. In the marine scenario, the importance of the chemistry
of the halogen atoms was suggested to be comparable to that
of the HO• radical.
Cl• and Cl2•- radicals are strong oxidants (E(Cl2•-/2Cl-) ) 2.0
V and E(Cl•/Cl-) ) 2.4 V vs NHE)8,9 and may react with
organic compounds by electron transfer, addition to double
bonds and H-abstraction, with rates depending on the difference
of the redox potentials.10
Reported rate constants of the reactions of Cl2•- indicate that
this radical ion may have a lifetime of the order of fractions of
milliseconds in tropospheric aqueous phase of low pH and high
Cl- concentrations.4 Therefore, reactions with organic and
inorganic constituents of tropospheric droplets and aerosols may
be important sinks for Cl2•-. In contrast, due to the fast Cl•
reaction with water, a lifetime of less than 5 µs was estimated
for chlorine atoms under tropospheric conditions. Under such
conditions, reactions of Cl• atoms with organic and inorganic
compounds in the tropospheric aqueous phase should be
negligible.5,6 However, the reactions of Cl• and Cl2•- with water
were reported to involve a series of not well-established
equilibria leading to the final formation of hydroxyl radicals,
Aromatic compounds are important constituents of automotive
gasoline and contribute to the formation of ozone and secondary
organic aerosols. Our understanding of the atmospheric chem-
istry of aromatic compounds is incomplete, and assessments of
the environmental impact of the release of such species are
uncertain.11 The determination of the reactivity of Cl• and Cl2
•-
radicals toward aromatic compounds is of environmental
importance, as these reactions may lead to the formation of
undesired chlorinated aromatic derivatives or else promote their
oxidation.
The reaction of chlorine atom with benzene in organic
solvents has been interpreted to proceed via one or both of the
intermediates (Chart 1) to explain the experimental results:12,13
a π complex and the 6-chlorocyclohexadienyl radical (Cl-
CHD). The relation between these two species is not well-
established, as related investigations present difficulties in
* To whom correspondence should be addressed. E-mail: dmartire@
volta.ing.unlp.edu.ar. gonzalez@inifta.unlp.edu.ar. Fax: 54 221 425 4642.
† Universidad Nacional de R´ıo Cuarto, Co´rdoba, Argentina.
‡ Lehrstuhl fu¨r Umweltmesstechnik, Universita¨t Karlsruhe, D-76128
Karlsruhe, Germany.
10.1021/jp9929768 CCC: $19.00 © 2000 American Chemical Society
Published on Web 03/21/2000

3118 J. Phys. Chem. A, Vol. 104, No. 14, 2000
Alegre et al.
•-
TABLE 1: Manifold of Reactions of Cl• and Cl2 in
CHART 1
Aqueous Solutions
k/M-1 s-1 a
•-
S2O82- + hν f 2 SO4
(1)
(2)
SO4•- + Cl- f SO42- + Cl•
log k ) 8.43 +
1.018I1/2 b
Cl• + Cl- f Cl2
8.5 × 109 d
(3)
(4)
(5)
•-
Cl2•- f Cl• + Cl-
6.0 × 104 s-1 b
log k ) 8.8 + 1.6I1/2
(I1/2 + 1)c
extracting information from systems equilibrating at rates which
compete with bimolecular reactions.
Cl2•- + Cl2•- f Cl2 + 2Cl-
/
An equilibrium between Cl•, C6H6, and the intermediates was
Cl2•- + HO- f ClOH•- + Cl-
Cl2•- + H2O f (HOClH•) + Cl-
Cl• + H2O f (HOClH)•
4 × 106 i
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
(15)
(16)
(17)
(18)
(19)
suggested, with an equilibrium constant <103 M-1 13,14
Absorp-
.
1300 s-1 b,d
tion bands centered at 490 and 320 nm were observed, the first
exhibiting a lifetime of 1 to 10 µs, was attributed to the π
complex.12,14 On the basis of these observations, a mechanism
was suggested involving the conversion of the π complex to
the Cl-CHD radical.14 Because both absorbances remained
independent of reaction temperature, other authors favored the
existence of only the π complex intermediate.12 However,
investigations on the role of the Cl•-benzene complex in
enhancing the selectivity of alkane photochlorinations favor the
formation of the Cl-CHD radical. The hypothesis of a Cl-
CHD intermediate is also supported by the formation of
chlorobenzene as a reaction product.13a Moreover, attempts to
correlate the energy of the UV and visible absorption bands of
a series of Cl•/ arene complexes with the vertical ionization
potentials and basicities of the arenes investigated failed to
indicate a σ or π nature of the complex.15
2.5 × 105 s-1 b,d
(5 ( 2) × 104 s-1 k
1.0 × 108 s-1 b
3.0 × 1010 e
6.1 × 109 s-1 e
4.3 × 109 e
(HOClH)• f Cl• + H2O
(HOClH)• f ClOH•- + H+
ClOH•- + H+ f (HOClH)•
ClOH•- f HO• + Cl-
HO• + Cl- f ClOH•-
(HOClH)• + Cl- f Cl2•- + H2O
Cl2 + H2O f ClOH + Cl- + H+
(8 ( 2) × 109 l
11.0 s-1 b
-
Cl2 + Cl- a Cl3
K ) 0.18 M-1 h
SO4•- + S2O82- f SO42- + S2O8
SO42- + HO• f SO4•- + HO-
S2O82- + HO• f
1.2 × 105 g
.-
1.0 × 106 f
<106 f
a Unless otherwise indicated. b Reference 2. c This work, in agree-
ment with ref 2 at zero ionic strength. d Reference 7. e Reference 3.
f Reference 32. g Reference 26. h Reference 25. i Reference 4. k This
work. The program sensitivity to these rate constants is of the order of
10%. However, a larger dispersion is observed from the values required
for the simulation of different experiments. l This work, in agreement
with the value proposed in ref 2.
The gas-phase reaction between Cl• atoms and benzene
proceeds via H-atom abstraction and adduct formation.16 The
equilibrium between Cl• atoms, benzene, and the intermediate
strongly shifted toward the reactants, precludes the detection
of the intermediate, for which a chlorocyclohexadienyl radical
structure was proposed.
the present study, we may assume that equilibrium conditions
are attained for reactions 3 and 4 (vide infra). Taking a value
of K3,4 ) 1.4 × 105 M-1 for the stability constant,7 a ratio
[Cl2•-]/[Cl•] > 10 is expected for experiments with [Cl-] g 1
× 10-4 M. Consequently, any contribution of chlorine atoms
to the observed absorption traces can be neglected. Under the
conditions of the time-resolved experiments, a transient species
with an absorption maximum at λ ∼ 340 nm, whose spectrum
agrees with that of Cl2•- radical ions is observed.19 The transient
decay follows a complex kinetics strongly depending on [Cl-],
as shown in Figure 1.
Despite the long history of the reaction of Cl• atoms with
benzene, the nature of the main reaction channels is still not
resolved. Moreover, this reaction has apparently never been
investigated in the aqueous phase.
In this paper, we report a kinetic and mechanistic study on
the aqueous phase reactions of Cl2•- and Cl• radicals and their
reaction with benzene. This work is intended to contribute to
the quantification of the aqueous phase atmospheric chemistry
The transient absorbance profiles at a given wavelength, A(λ,
t), may be well fitted over more than three lifetimes with
simultaneous first and second-order processes, according to eq
2.
and to a better evaluation of the importance of Cl2 and Cl•
•-
radicals in multiphase tropospheric chemistry.1
Photolysis of S2O82- (λexc < 300 nm), reaction 1 in Table 1,
is a clean source of sulfate radical ions with high pH-
independent quantum yields.17 In the presence of Cl- ions,
reaction 2 has been reported to yield Cl• atoms, which reversibly
ae-at
c(λ) - b(λ)e-at
A(λ,t) )
+ d(λ)
(2)
•-
react with Cl- to yield Cl2 radical ions, reactions 3 and 4.2
The reactions of Cl• and Cl2 are studied by laser and
•-
conventional flash-photolysis as well as by steady-state pho-
where all the terms in the equation are highly sensitive to the
chloride ion concentration. The constant a is independent of
the detection wavelength λ while all other terms are wavelength
dependent. Terms a and b(λ) are related to the rate constants of
the first and second-order decay processes, respectively. They
are, within experimental error, independent of the presence of
molecular oxygen and on the concentration of S2O82-. Terms
c(λ) and d(λ) are related to the concentration of Cl2•-, formed
immediately after excitation and to the absorption of a longer
lived species formed after Cl2 depletion, respectively. Both
c(λ) and d(λ) are highly sensitive to the irradiance and, in
general, d(λ) , c(λ).
The relative contribution of the first- and second-order decay
processes strongly depends on [Cl-] and on the irradiance. In
experiments with [Cl-] > 0.2 M, the dichloride radical ion was
observed to decay mainly by a second-order process on the
2-
tolysis of S2O8 solutions containing variable concentrations
of Cl- in the presence and absence of benzene.
Results and Discussion
Cl• and Cl2 Reactions in Aqueous Solutions in the
•-
Absence of Benzene. Photolysis of aqueous peroxodisulfate
solutions in the pH range of 2-3 showed formation of a transient
species in the wavelength range from 300 to 550 nm whose
decay rate and spectrum are in agreement with that reported in
•-
•-
the literature for SO4 radical ions in a wider pH range.17,18
During irradiation experiments with chloride ion concentra-
tions [Cl-] in the range of 1 × 10-4 M to 0.6 M, SO4•-, radicals
are readily depleted due to the high efficiency of reaction 2
yielding chlorine atoms. Reaction of Cl• with Cl- ions reversibly
•-
yields Cl2 radical ions, reactions 3 and 4. At [Cl-] used in

•-
Reactions of Chlorine Atoms and Cl2 Radical Anions
J. Phys. Chem. A, Vol. 104, No. 14, 2000 3119
Figure 2. Dependence of the logarithm of the recombination rate
constant k5 on the ionic strength. Corresponding linear regression
parameters are ordinate ) 8.8, slope ) 1.6, r2 ) 0.94. Dotted curves
show the 99% confidence interval. Inset: First-order rate coefficient
for the loss of Cl2 as a function of [Cl-].
•-
A mechanism involving reactions 3 and 4 and those of Cl•
and Cl2 with water, shown in eq 1, has been proposed for
Figure 1. Absorption profiles at λanal ) 340 nm obtained by
conventional flash photolysis of 5 × 10-3 M K2S2O8 solutions in the
presence of different [Cl-]: (a) 0.1 M, (b) 1 × 10-3 M and (c) 1 ×
10-4 M, and (d) 0.6 M. Traces e and f correspond to absorption profiles
•-
explaining the apparent first-order component of the Cl2•- decay
and its dependence on [H+] and [Cl-].2,3,6,7 Any contribution
of a reaction of Cl• and Cl2•- with peroxodisulfate is neglected,
as no dependence of the first-order decay component on
peroxodisulfate concentration is observed, in agreement with
literature reports.7b According to the mechanism shown in eq
following laser flash photolysis (λexc ) 266 nm) of 2.25 × 10-2
M
K2S2O8 solutions in the presence of (e) 0.098 M Cl- and λanal ) 340
nm; (f) 0 M Cl- and λanal ) 450 nm. Solid curves represent computer
simulations (refer to text).
millisecond time scale, as shown in Figure 1d for experiments
•-
1, reaction of Cl• and Cl2 with water involves a series of
with [Cl-] ) 0.6 M. In these experiments, almost no first-order
equilibria leading to the final formation of chloride ions and
HO• radicals. It may therefore be expected that further reactions
removing HO• radicals would severely affect the equilibrium
condition in eq 1, and the first-order decay component should
depend on the rate of such reactions. When HO• radicals are
completely removed, no equilibrium is established and reactions
component is observed within experimental error, even for very
•-
low initial concentrations of Cl2
.
The second-order component may be associated with the
bimolecular decay of Cl2•- radical ions yielding Cl2 and Cl-,2,20
reaction 5 in Table 1. Term b(λ) is related to the rate constant
k5 by the expression b(λ) ) 2k5/lꢀ(Cl2•-)λ, where the optical
path length l is 20 cm for conventional flash and 1 cm for laser
experiments. The high sensitivity of b(λ) to [Cl-] may be
attributed to a real dependence on the ionic strength. The
dependence of k5 on the ionic strength (I), depicted in Figure
2, is consistent with the Debye-Hu¨ckel equation for a reaction
between two single charged negative ions, eq 3.
of Cl• and Cl2 with water follow simple first-order kinetics,
•-
as observed by Buxton et al.7 Possible reactions removing HO•
radicals in our reaction system are recombination and reactions
2-
2-
with S2O8 and SO4 ions, reactions 18 and 19 in Table 1,
respectively.
The kinetic analysis being rather complex, a computer
program based on the numerical resolution of the differential
equations system by the third-order Runge Kutta method was
used in order to simulate the decay of the transient traces.23a
Stepsizes of the order of 10-10-10-9 s (depending on the
experimental conditions) were checked for convergence and
fixed throughout the complete simulation of a decay trace. The
ability of the program to solve the system of differential
equations was checked with a stiff solver method.23b The
program considers the pulsed excitation as a delta function,
x
A I
log k5 ) log k05 + 2
(3)
x
1 +
I
where ko stands for the rate constant at zero ionic strength,
5
and A is a function of the temperature and physical constants
of the solvent. Though this equation is strictly valid only at
very low ionic strength, it holds reasonably well up to ionic
strength values of 0.5 M, as also observed for the reactions of
•-
•-
producing SO4 radicals. The concentration of SO4 present
immediately after excitation [SO4•-]o was taken as an input
parameter, estimated from experiments under identical experi-
mental conditions, but in the absence of Cl-, as shown in Figure
1 for experiments e and f. Typical [SO4•-]o values were of the
order of (2-3) × 10-5 M and (2-3) × 10-6 M for experiments
performed with conventional and laser flash techniques, respec-
tively.
•-
SO4 with Cl-.21,22 Although the observed value A ) 0.8 is
higher than the accepted value of 0.55 for water at 25 °C,
0
extrapolation to infinite dilution leads to a rate constant k5
)
6.1 × 108 M-1 s-1. The data reported in ref 2 for the ionic
strength dependence of k5 fall in the straight line shown in Figure
2.
•-
The first-order rate coefficient of the Cl2 decay markedly
increases for [Cl-] < 10-3 M, a lower limit being found for
[Cl-] > 0.1 M, as illustrated in Figure 2 (inset). The observed
tendency corroborates with the results reported by McElroy.2
The program incorporates a set of well-established reactions
involving the species present in the irradiated system, reactions
2-8 and 15-19, recombination of HO• radicals to yield H2O2,

3120 J. Phys. Chem. A, Vol. 104, No. 14, 2000
Alegre et al.
•-
TABLE 2: Manifold of Reactions of Cl• and Cl2 to Be
Considered in the Presence of Benzene
k/M-1 s-1
Cl2•- + Bz f Cl-
organic radicals
Cl• + Bz f
+
<1 × 105 a
6 × 109 ek21 e1.2 × 1010 b
(20)
(21)
(22)
(23)
(24)
organic radicals (R•)
Cl2•- + R• f
products
c
Bz + HO• f
HCHD
7.8 × 109 d
3 × 109 e
Bz + SO4•- f
HCHD
R• + R• f
(1.4 ( 0.4) × 109 c
(25)
(-21)
R• f Cl• + Bz
<104 s-1 c
a This work. b This work. The upper limit is estimated for k22 < 1
× 109 M-1 s-1, and the lower limit for k22 > 5 × 109 M-1 s-1, see text.
c See text. d Reference 32. e Reference 26.
both for laser and conventional flash photolysis experiments
with 1 × 10-4 M < [Cl-] < 0.6 M, as shown by the solid
traces in Figure 1. We may therefore deduce that the set of
reactions depicted in Table 1 may account for the experimental
•-
Cl2 decay. According to the proposed mechanism, the
Figure 3. Linear dependence of the first-order decay component (kapp
)
of Cl2•- on [Bz] for experiments with (a) [Cl-] ) 0.1 M , (b) 0.16 M,
(c) 0.24 M, (d) 0.4 M, (e) and 0.5 M, respectively. Inset: Signals
obtained at λanal ) 340 nm by conventional flash photolysis experiments
with 5 × 10-3 M K2S2O8 and 0.1 M NaCl solutions in the absence
(upper trace) and in the presence of 2.4 × 10-3 M benzene (lower
trace). The solid lines represent computer simulations (refer to text).
observed final absorption of a long-lived species is related to
the formation of Cl2 and Cl3- (reaction 5). Chlorine hydrolysis
to ClOH and HCl in aqueous solutions (τ ≈ 100 ms, reaction
•-
15) is too slow to affect the Cl2 decay.
•-
Dependence of the Cl2 Radical Ion Decay Kinetics on
the Presence of Benzene. Experiments with Benzene Concen-
trations ([Bz]) < 5 × 10-3 M. Photolysis of air-saturated
aqueous solutions of peroxodisulfate solutions containing
chloride ions in the range 0.1 M < [Cl-] < 0.6 M and [Bz] <
5 × 10-3 M, lead to the formation of transient species with an
absorption maximum at 340 nm. The transient spectrum agrees
and further reactions involving the latter two species.23c When
considerable dispersion in the rate constants is reported, only
those values supported by detailed experimental methods were
considered, as depicted in Table 1. The computer program does
not assume a priori equilibrium conditions for any of the
reversible reactions, as such an operation would limit our
analysis to particular sets of equilibrium conditions.
•-
with that of the Cl2 radical ion obtained in the absence of
benzene and is, therefore, assigned to this species.
For given [Cl-], faster decay rates of Cl2 are observed by
•-
Formation of intermediate (HOClH)• in the reactions of water
with Cl2 and Cl• (reactions 7 and 8, respectively) and the
•-
addition of increasing [Bz], as shown in Figure 3 (inset). The
experimental absorption traces at a given wavelength of analysis,
A(λ), could be fitted according to eq 4. The wavelength-
independent constant g depends linearly on [Bz] (Figure 3) and
is related to the first-order decay rate of Cl2•-. The remaining
absorption (term h(λ)) may be associated with a longer lived
species formed after Cl2 depletion. h(λ) is less than 10% of
the initial absorption of Cl2 (preexponential factor f(λ)) and
could not be accurately determined from these fittings.
reverse reaction 9, with k9 e 103 s-1 was proposed by McElroy2
in order to account for the experimentally observed dependence
of the Cl2•- decay on pH. Deprotonation of (HOClH)• radicals
to ClOH•- (reaction 10), its reverse reaction 11, and the
subsequent reversible decomposition of ClOH•- to OH• and Cl-
ions (reactions 12 and 13) are in agreement with the reactions
proposed by Jayson for the formation of Cl2•- by HO• and Cl-.3
The diffusion-controlled reaction between (HOClH)• and Cl-,
reaction 14, was also proposed in ref 2 in order to account for
•-
•-
A(λ) ) f(λ)e-gt + h(λ)
(4)
•-
the observed rate of Cl2 formation in pulse radiolysis
experiments of acidic NaCl solutions.3 Though not well
documented, this latter reaction has been taken into account in
this work since, otherwise, simulated profiles differed consider-
ably from experimental results. Simulations showed that the
recombination of Cl• atoms is not significant under our
experimental conditions, and consequently, this reaction has
been neglected.
Cl• and Cl2•- radicals are strong oxidants able to react with many
organic compounds. The effect of benzene on the decay rate of
•-
Cl2 can be understood, if reactions 20 and 21 (Table 2)
efficiently compete with the decay reactions of Cl2 and Cl•
•-
depicted in Table 1.
The total concentration of the organic radicals R• formed from
-
•-
Cl2•-, Cl•, and Cl2/Cl3 are the main species showing
the reaction of benzene with Cl• and Cl2 cannot exceed
•-
absorption at 340 nm. Simulated concentration profiles for Cl2
,
[SO4•-]o, therefore, [R•] e 3 × 10-5 M (vide supra). Conse-
Cl•, and Cl2/Cl3 were converted into the corresponding
quently, a diffusion-controlled reaction between R• and Cl2
,
-
•-
absorbance profiles taking ꢀ(Cl2•-) ) 9600 M-1 cm-1 7,19
(Cl•) ) 3800 M-1 cm-1 24 ꢀ(Cl2) ) (68 ( 5) M-1 cm-1 25
and
ꢀ(Cl3-) ) (178 ( 5) M-1 cm-1 25
The simulations show that
,
ꢀ-
reaction 22, may accelerate the decay kinetics of Cl2 radical
ions, but in a first approximation, this contribution of reaction
22 will not be considered.
•-
,
,
.
Cl• contribution to the absorption profiles at 340 nm is
negligible, in agreement with our previous assumptions.
For k9 ) 5 × 104 s-1 and k14 ) 8 × 109 s-1 M-1, excellent
agreement between experimental and simulated data is obtained,
The efficient removal of HO• radicals in the presence of
benzene, reaction 23, does not allow the establishment of
•-
equilibrium conditions in eq1, and reactions of Cl2 and Cl•
with water may be considered as simple first-order reactions.

•-
Reactions of Chlorine Atoms and Cl2 Radical Anions
J. Phys. Chem. A, Vol. 104, No. 14, 2000 3121
Figure 4. Plot of the slopes (S) of the straight lines shown in Figure
3 vs [Cl-]-1. The dotted line shows the 99% confidence interval for a
linear regression.
TABLE 3: Intercepts of the Plots of kapp vs Benzene
Concentration Obtained for Different Values of [Cl-]
[Cl-]/M
intercept/s-1
Figure 5. Absorption spectrum of the organic transients obtained (O)
20 µs and (b) 400 µs after irradiation with Nd:YAG laser and
conventional flash lamp, respectively, of nitrogen-saturated 5 × 10-3
M solutions of K2S2O8 and 0.1 M of NaCl in the presence of benzene
(>1 × 10-2 M). Inset: Transient absorption profiles at 300 nm obtained
0.10
0.12
0.16
0.24
0.40
0.50
1050 ( 160
1275 ( 120
1055 ( 180
1130 ( 40
1065 ( 40
1110 ( 30
from (a) N2 and (b) air-saturated solutions containing 5 × 10-3
M
K2S2O8, 0.1 M of NaCl, and 2 × 10-2 M of benzene. The solid lines
represent computer simulations (refer to text).
If reactions 3, 4, 7, 8, 20, and 21 are the main reactions
depleting Cl2•-, which is assumed in equilibrium with Cl•, and
considering K3,4[Cl-] > 1, the apparent first-order decay rate
constant of Cl2•-, kapp, is given by eq 5.
experimental profiles, as shown by the solid traces in Figure 3
(inset). Reactions of Cl• atoms with the organic radicals R. show
a negligible effect on the simulated decays, even assuming
diffusion controlled rate constants, and were therefore not taken
into account.
k8
k21
kapp ) k7 +
+
+ k20 [Bz] (5)
K3,4 [Cl-]
K3,4 [Cl-]
A value k22 g 5 × 109 M-1 -1
s
may be realistic, since reported
[
]
•-
•
•
reactions between Cl2 and radicals such as ClO2 and HO2
are of that order of magnitude.26 The value k21 ) 6 × 109
In fact, plots of kapp vs [Bz] (Figure 3) fit to eq 5. However, for
M-1s-1 is in complete agreement with that reported for the
the Cl- concentration range used in these experiments, reaction
•-
•-
reaction in benzene as solvent.12 A very low reactivity of Cl2
8 does not contribute to the first-order decay component of Cl2
as (k8/K3,4 [Cl-]) , k7. The intercept values for the straight
lines as those shown in Figure 3 are depicted in Table 3. As,
within the experimental error, these values are independent of
the chloride ion concentration, the average value 1115 ( 140
s-1 38 is a good estimation for k7, in agreement with previous
reported values.7 The slopes of the curves as those shown in
Figure 3 depend on [Cl-]-1 as expected from eq 5 and shown
in Figure 4.
toward benzene is also in agreement with the reported rate
constants for its reactions with most organic substrates.5,7,27
Experiments with [Bz] > 10-2 M. For irradiation experiments
with [Cl-] ) 0.1 M and [Bz] g 1 × 10-2 M, reaction 21 is
very efficient and Cl2•-/Cl• radicals are readily scavenged. Under
these experimental conditions, a lifetime τ < 60 µs is expected
•-
for Cl2 and observed in laser flash photolysis experiments.
At longer times of analysis, formation of transient species
absorbing in the wavelength region from 290 to 360 nm with a
maximum at approximately 300-310 nm was observed. The
transient spectrum observed 20 µs after the laser shot agrees,
within experimental error, with that observed in conventional
flash experiments at times up to fractions of milliseconds, as
shown in Figure 5. Blank experiments performed with benzene
solutions under identical experimental conditions but in the
From the slope of the linear function in Figure 4, the
bimolecular rate constant k21 ) (1.2 ( 0.6) × 1010 M-1 s-1 is
obtained. The observed negligible value for the intercept yields
k20 < 1 × 105 M-1 s-1 within the limits of experimental error.
To evaluate the contribution of reaction 22 to the concentra-
tion profiles of Cl2•-, a detailed kinetic analysis was performed
with the aid of computer simulations. For this purpose, the set
of reactions depicted in Table 1 along with reactions 20-24 in
Table 2, was used to fit typical experiments, as that shown in
Figure 3 (inset).
2-
absence of S2O8 showed no signal, indicating that any
contribution of benzene photolysis is negligible and that the
observed transients are mainly formed after reaction 21.
Simulated and experimental profiles show excellent agree-
ment when the k20 and k21 values retrieved from the kinetic
The absorption coefficient of the transient is estimated from
laser experiments performed with N2-saturated solutions in the
presence and absence of benzene. Under otherwise identical
experimental conditions, it may be assumed that the same
amount of Cl2 is formed in all the experiments and mainly
consumed by reaction 21. The estimated absorption coefficient
analysis already described are used, and for k22 < 109 M-1 s-1
.
Despite that k20 < 105 M-1s-1 from the simplified analysis, the
•-
simulation program shows that reaction 20 does not contribute
to the Cl2 decay already for k20 e 106 M-1s-1. For higher
•-
input values of k22, simulated decays are faster than experimental
ones. Assuming k22 > 5 × 109 M-1s-1, a value k21 ) 6 × 109
M-1 s-1 is required for an agreement between simulated and
would be a lower limit value if this latter condition does not
•-
apply. Taking ꢀ(Cl2•-, 340 nm) ) 9600 M-1cm-1 7,19
the Cl2
,
concentration formed in a typical experiment as shown in Figure

3122 J. Phys. Chem. A, Vol. 104, No. 14, 2000
Alegre et al.
1e is ≈(1.45 ( 0.1) × 10-6 M (also in complete agreement
with [SO4•-]0, vide supra). Thus, a value of ꢀ(300 nm) ) 2800
( 500 M-1 cm-1 is obtained for the organic transient.
The experimental traces obtained at a given wavelength of
analysis λ from conventional flash photolysis experiments with
nitrogen saturated solutions (Figure 5 inset) may be well fitted
to a second-order law. The second-order rate constants are
insensitive to [S2O82-] and [Bz] in the range of [Bz] g 1 ×
10-2 M and may be associated with the recombination of the
organic radicals (reaction 25). For the optical path length l )
11 cm and ꢀ(300 nm) ) 2800 ( 500 M-1 cm-1, the rate constant
2k25 ) (1.4 ( 0.4) × 109 M-1 s-1 is calculated.
A transient absorbing at 490 nm with a lifetime of 1-10 µs,
which has been attributed to the π complex,12,14 is not observed
in our laser experiments. We may therefore conclude that a π
complex between benzene and Cl• atoms is either not formed
or decomposes in aqueous solutions in less than 50 ns to yield
a secondary organic radical with an absorption maximum at
approximately 300 nm.
Hydroxycyclohexadienyl radicals (HCHD) are known to
present a single absorption maximum at 300-310 nm28,29 with
an absorption coefficient g2000 M-1 cm-1. The bimolecular
decay rates 2k25/ꢀmax are in the range from (1-5) × 105 s-1
cm-1 for a wide number of substituted HCHD.28-33 The
observed transient showing absorption and kinetic properties
very similar to those of HCHD is assumed to represent the Cl-
CHD radical, in agreement with the postulated structure of the
corresponding σ complex. In fact, the absorption spectra of CHD
radicals of 3-chlorotoluene formed upon addition of H atoms
and those formed upon addition of HO• radicals are very
similar.28c
Figure 6. Depletion of (b) [Bz] and evolution of (2) [chlorobenzene]
and 2,4- hexadiene-aldehyde (4, arbitrary units) vs irradiation time for
preparative experiments performed with air saturated aqueous solutions
containing 1 × 10-4 M of K2S2O8, 0.5 M of NaCl, and 5 × 10-3 M of
benzene. Depletion of (O) [Bz] in blank experiments performed in the
absence of K2S2O8.
104 M-1 31b to 0.5 × 104 M-1 31a
.
The value of K26 found for
the Cl-CHD radical is on the order of the one reported for the
HCHD radical of chlorobenzene.
To test the presence of reaction 26, the experimental traces
were simulated. An adequate set of reactions includes equilib-
rium 26, recombination of Cl-CHD (reaction 25), recombina-
tion of PR radicals, and a first-order decay of PR.30 This
simulation shows that the decay of Cl-CHD radicals in air-
and oxygen-saturated solutions is not sensitive to the input
values concerning the recombination of PR radicals. Due to
equilibrium (26), the depletion rate of Cl-CHD is controlled
by the first-order decay of PR under our experimental conditions
for oxygen- and air-saturated solutions. A good agreement
between experimental and simulated profiles (Figure 5 (inset))
is obtained for k ) 1.4 × 104 s-1, where k is the mono-molecular
decay rate constant of PR which is of the order reported for the
decomposition of many alkyl peroxyl radicals.34
HCHD radicals reversibly react with molecular oxygen
yielding peroxyl radicals (PR), reaction 26, which typically
exhibit broad absorption only at λ < 300 nm.30,31
HCHD + O2 a PR
(26)
Experiments performed under identical conditions but in the
presence of molecular oxygen show the same transient spectrum,
although with lower absorbance and faster depletion rates
(Figure 5 (inset)). Experimental traces obtained for air-saturated
solutions could be well-fitted to a first-order law. The high
absorption of the solutions below 280 nm precluded time-
resolved detection at shorter wavelengths, and, consequently,
PR formation cannot be observed. However, the lower concen-
tration of Cl-CHD and its faster decay kinetics in the presence
of molecular oxygen may be rationalized by the existence of a
reversible reaction between Cl-CHD radicals and O2, similar
to reaction 26. An equilibrium constant K26 ) 1100 ( 500 M-1
is estimated from conventional flash experiments performed with
N2, O2, and air-saturated solutions under otherwise identical
experimental conditions, assuming that identical initial concen-
trations of radicals are formed in all the experiments and that
equilibrium conditions for reaction 26 are established.
Reported data concerning the equilibrium constants K26 of
several HCHD radicals show that the equilibrium shifts away
from the peroxyl radicals with increasing electron-withdrawing
power of the substituent, indicating that a considerable electron-
density at open-shell carbon is required for the formation of a
C-O2 bond. Cl being a stronger electron-withdrawing group
than HO, it may be expected that the reaction of O2 with Cl-
CHD radicals is less favorable than that with the parent HCHD
radical. In fact, the presence of a Cl substituent in the HCHD
radical of benzene lowers the equilibrium constant from 2.6 ×
Product Analysis. Identification of the products formed after
the reactions of Cl2•-/ Cl• with benzene in air-saturated solutions
was performed with photolysis experiments using immersion-
type reactors. The photolysis of air-saturated aqueous solutions
of benzene (2 × 10-2 M) yielding oxidation products, such as
phenol, catechol, dihydrocatechol, and benzoquinone, required
experimental conditions designed to minimize benzene pho-
tolysis. High [Cl-]/[S2O82-] ratios avoid oxidation of benzene
by sulfate radicals, reaction 24.
Irradiation of aqueous solutions containing [Bz] ) 5 × 10-3
M, [S2O82-] ) 1 × 10-4 M, and [Cl-] ) 0.5 M showed
depletion of benzene and the simultaneous formation of two
main products identified as chlorobenzene and 2,4- hexadiene-
aldehyde by GC/MS (Figure 6). Experiments with increasing
concentrations of benzene showed higher yields of chloroben-
zene. Diminution of the concentration of dissolved molecular
oxygen is also observed.
Because chlorobenzene represented less than 10% of the
consumed benzene, we deduce that oxidation yielding aliphatic
species must be the major route of the reaction of benzene with
Cl• in air saturated aqueous solutions. Aliphatic C1 to C4
products could not be analyzed and identified, because they elute
with the solvent under the given analytical conditions.

•-
Reactions of Chlorine Atoms and Cl2 Radical Anions
J. Phys. Chem. A, Vol. 104, No. 14, 2000 3123
Conclusions
SCHEME 1
The understanding of the complex aqueous phase chemistry
of the Cl2•-/ Cl• couple is relevance to discerning the fate of
oxidizing radicals in cloudwater droplets containing significant
concentrations of chloride ions and is a prerequisite for a detailed
investigation of their reactivity toward organic substrates. With
the reaction sequences depicted in Table 1, purely second-order,
largely first-order and mixed first- and second-order decays of
Cl2•-, observed for different [Cl2•-] to [Cl•] ratios, may be
simulated. According to the proposed mechanism, reactions of
Cl2 and Cl• with water (reactions 7 and 8, respectively)
•-
efficiently lead to the formation of HO• radicals for [Cl-] <
10-2 M, where the condition v14 < v10 applies. Therefore, in
marine clouds containing [Cl-] ≈ 0.5 mM, reactions 7 and 8
will lead to an increasing oxidation capacity of the aqueous
medium. Under conditions of [Cl-] ≈ 0.5 mM and low radical
(-21)). A possible explanation for the apparent mechanistic
differences is that Cl-CHD is always formed in each phase
but it does not significantly dissociate on the time scale of our
observations. Under the experimental conditions attained in our
•-
concentrations, Cl2 lifetime is expected to be of the order of
•-
time-resolved experiments, Cl-CHD scavenging by Cl2
fractions of milliseconds, and Cl2•- may thus react with organic
constituents of the aqueous troposphere. Chlorine atoms con-
centrations are less important. However, due to the reversibility
of reaction 3 and their high reactivity (vide infra), Cl• reactions
with organic substrates may constitute an important sink for
(reaction 2) may be sufficiently fast to prevent dissociation, thus
V22 > V-21 and reactions 21 and -21 are not equilibrated. Taking
k22 > 5 × 109 M-1 s-1 and [Cl2•-] ≈ 10-6 M, k-21 < 104 s-1
is estimated. From the gas-phase data16 it can be estimated that
reaction (-21) would have a half-life of ca. 0.2 µs which is
significantly shorter than that estimated in aqueous solutions.
These observations indicate that the role of the aqueous medium
is that of stabilizing the Cl-CHD radicals. A similar behavior
was reported for the HCHD radical lifetimes in the gas and
aqueous phase.36
•-
Cl2 radical ions, if organic constituents are present in
concentrations higher than 10-5 M.
We found that Cl• atoms are much more reactive toward
benzene (6 × 109 M-1 s-1 e k21 e 1.2 × 1010 M-1 s-1) than
Cl2•- radical ions (k20 < 1 × 105 M-1 s-1). This observation is
in agreement with studies involving aliphatic organic compounds
reported by Gilbert10 and Buxton.19 Reaction of Cl• atoms with
benzene proceeds via C6H6-Cl adduct formation, probably best
represented as a Cl-CHD radical. We were unable to observe
any stable π adduct between Cl• atoms and C6H6. Reported
calculations on such a π complex estimate very low bond
energies (<12 kJ),16 suggesting that, unless highly stabilized
by the solvent, it cannot survive more than a few picoseconds
at room temperature and, consequently, is too short-lived to play
any role in a sequence of chemical reactions.
The rate constant for H abstraction in the gas phase is 7.8 ×
104 M-1 s-1 16
. Assuming a similar value in aqueous media, the
competition of this reaction with the efficient Cl-CHD forma-
tion is of no significance, as reaction -21 is also irrelevant.
Reaction of Cl-CHD with O2 has been proposed to yield
chlorobenzene and HO2 radicals, reaction 27 in Scheme 1.13a
•
However, chlorobenzene is found to be a minor product (molar
yield < 10%) in preparative experiments performed with air-
saturated solutions. This result is an indication that the reaction
of Cl-CHD with O2 proceeds via more than one reaction
channel, as already observed in the gas phase.16 In analogy to
the reactions leading to aliphatic oxidation products from
hydroxycyclohexadienylperoxyl radicals,31b reaction 28 may
compete with HO2• elimination. At the present state of knowl-
edge, the formation of 2, 4-hexadiene-aldehyde from Cl-CHD
radicals cannot be explained. Subsequent reactions of the peroxyl
radical formed by the addition of molecular oxygen to Cl-
CHD seem to be rather complex and have not yet been
investigated in detail.16
Cl-CHD recombination, reaction 25, leads to chlorobenzene
formation and may contribute to the decay of Cl-CHD radicals
in air-saturated solutions (vide supra). This reaction is certainly
less significant in steady-state experiments, where only very
low radical concentrations are obtained. Biphenyl products were
not detected (detection limit: 0.1 µM).
H atom abstraction from the aromatic ring yielding reactive
phenyl radicals would be another possible reaction pathway.
Phenyl radicals readily react (half-life <3µs in air saturated
solutions) with molecular oxygen to yield aryl peroxyl radicals
absorbing in the spectral domain of 400-500 nm.35 Such
peroxyl radicals decay by bimolecular recombination with rate
constants <107 M-1 s-1 35b
and consequently, could be detected
,
with our experimental setup. The absence of traces absorbing
in the wavelength range of 400-500 nm does not support an
important participation of such reactions.
Like the hydroxyl radical, the reaction of Cl• with aromatics
involves addition to the ring rather than hydrogen abstraction.36
Halogenation and radiation-induced hydroxylation of substituted
benzenes have been suggested to be similar in nature, as both
reacting species are electrophiles. The Hammett F values
reported for aromatic chlorination and hydroxylation (-0.56 and
-0.52, respectively) as well as the relative distribution of radical
addition are comparable.28c Formation of π complexes between
the electrophilic HO• and the π electrons of aromatic systems
with diffusion-controlled rates, irrespective of the substituents
at the aromatic ring, has been proposed as a precursor of a σ
complex shown as HCHD radicals.36 Rearrangement from the
π to σ complex takes place in the nanosecond time range.
The significant difference between gas and aqueous phase
reactions of Cl atoms and benzene is that in our experiments in
aqueous solutions there is no evidence for H-atom abstraction
from the aromatic ring nor for dissociation of Cl-CHD (reaction
Experimental Section
Materials. Potassium peroxodisulfate (Riedel de Hae¨n),
sodium chloride (Merck p.a.), and benzene (Merck p.a.) were
used as received. Distilled water was obtained from a Millipore
system (>18 Ω cm-1, <20 ppb of organic carbon).
Time-Resolved Experiments. Flash-photolysis experiments
were carried out using conventional equipment (Xenon Co.
model 720C) with modified optics and electronics (ref 23 and
references therein). The emission of the flash lamps was filtered
with a saturated solution of benzene in water, to minimize

3124 J. Phys. Chem. A, Vol. 104, No. 14, 2000
Alegre et al.
TABLE 4: Mass Spectra and Probable Structures of
Detected Products
concentration of dissolved oxygen in the samples was deter-
mined with a specific oxygen electrode (Orion 97-0899).
Acknowledgment. We thank Dr. G. Carrillo Leroux from
the Faculty of Engineering, University of Sao Paulo (Brazil),
for doing the simulations with the stiff solver program. This
research was supported by Fundacio´n Antorchas, Consejo
Nacional de Investigaciones Cient´ıficas y Te´cnicas, Argentina
(CONICET), Comisio´n de Investigaciones Cient´ıficas de la
Provincia de Buenos Aires (CIC), and the Deutsche Akade-
mische Austauschdienst (DAAD). J.A.R. thanks CONICET for
a graduate studentship. D.O.M. is a research member of CIC,
S.G.B and M.C.G. are research members of CONICET.
References and Notes
(1) (a) Walcek, C.; Yuan, H. H.; Stockwell, W. R. Atmos. EnViron.
1997, 31, 1221. (b) Jacob, D. J. J. Geophys. Res. 1986, 91, 9807. (c) Abbatt,
J. P. D.; Waschewsky, G. C. G. J. Phys. Chem. 1998, 102, 3719.
(2) McElroy, W. J. J. Phys. Chem. 1990, 94, 2435.
(3) Jayson, G. G.; Parsons, B. J.; Swallow, A. J. J. Chem. Soc., Faraday
Trans. 1973, 69, 1597.
(4) Jacobi, H. W.; Herrmann, H.; Zellner, R. Ber. Bunsen-Ges. Chem.
1997, 101, 1909.
(5) Hasegawa, K.; Neta, P. J. Phys. Chem. 1978, 82, 854.
(6) Wagner, I.; Kartha¨user, J.; Strehlow, H. Ber. Bunsen-Ges. Chem.
1986, 90, 861.
(7) (a) Buxton, G. V.; Bydder, M.; Salmon, G. A. J. Chem. Soc.,
Faraday Trans. 1998, 94, 653. (b) Buxton, G. V.; Bydder, M.; Salmon, G.
A. Phys. Chem. Chem. Phys. 1999, 1, 269.
(8) (a) Beitz, T.; Bechmann, W.; Mitzner, R. J. Phys. Chem. 1998,
102, 6766. (b) de Violet, P. F. ReV. Chem. Intermed. 1981, 4, 121.
(9) Steenken, S. In Free Radicals: Chemistry, Pathology and Medicine;
Rice-Evans, C., Ed.; Dormandy T. Richelieu Press: London, 1988; p 51.
(10) Gilbert, B. C.; Stell, J. K.; Peet, W. J.; Radford, K. J. Chem. Soc.,
Faraday Trans. 1 1988, 84, 3319.
(11) Platz, J.; Nielsen, O. J.; Wallington, T. J.; Ball, J. C.; Hurley, M.
D.; Strasccia, A. M.; Schneider, W. F.; Sehested, J. J. Phys. Chem. 1998,
102, 7964.
(12) Bunce, N. J.; Ingold, K. U.; Landers, J. P.; Lusztyk, J.; Scaiano, J.
C. J. Am. Chem. Soc. 1985, 107, 5464.
(13) (a) Skell, P. S.; Baxter, H. N.; Tanko, J. M.; Chebolu, V. J. Am.
Chem. Soc. 1986, 108, 6300. (b) Chatauneuf, J. E. Chem. Commun. 1998,
2099.
(14) (a) Bu¨hler, R. E. Radiat. Res. ReV. 1972, 4, 233 and references
therein. (b) Bu¨hler, R. E.; Bossy, J. M. Int. J. Radiat. Phys. Chem. 1974,
6, 95.
(15) Raner, K. D.; Lusztyk, J.; Ingold, K. U. J. Phys. Chem. 1989, 93,
564.
(16) Sokolov, O.; Hurley, M. D.; Wallington, T. J.; Kaiser, E. W.; Platz,
J.; Nielsen, O. J.; Berho, F.; Rayez, M. T.; Lesclaux, R. J. Phys. Chem.
1998, 102, 10671.
(17) McElroy, W. J.; Waygood, S. J. J. Chem. Soc., Faraday Trans.
1990, 86, 2557.
(18) Choure, S. C.; Bamatraf, M. M. M.; Rao, B. S. M.; Das, R.; Mohan,
H.; Mittal, J. P. J. Phys. Chem. A 1997, 101, 9837. Hug, G. L. Nat. Stand.
Ref. Data Ser. (U.S. Nat. Bur. Stand.) 1981, 69.
benzene photolysis in the reaction system. The optical path
length of the reaction cell l is either 11 or 20 cm. Laser
experiments were performed with a Spectron SL400 Nd:YAG
system generating 266 nm pulses (∼8 ns pulse width). The laser
beam was defocalized in order to cover the entire path length
(1 cm) of the analyzing beam produced by a 150 W Xe Lamp.
The experiments were performed with a quartz cell in a 90°
geometry. The detection system comprised a PTI monochro-
mator coupled to a Hamamatzu R666 PM tube. The signal was
acquired by a digitizing scope (Hewlett-Packard 54504), where
it was averaged and transferred to a computer.
Preparative Experiments. A cylindrical low-pressure Hg
lamp (Heraeus, MNNI 35/20, Germany) emitting at 254 nm
and presenting much lower emissions at λ > 312 nm was used.
The photochemical reactor was of annular geometry (volume:
400 mL) adapted for the lamp which was immersed within a
quartz tube. The annular optical path in the reactor was of the
order of 1 cm. The whole reactor was immersed in a thermostat
controlling the temperature at 25 ( 1 °C.
2-
Methods. Unless otherwise indicated, typical S2O8 con-
centrations were 5.0 × 10-3 M for conventional flash photolysis
and 2.0 × 10-2 M for laser experiments. The pH of the
peroxodisulfate solutions was approximately 3.0 to 2.05 due to
the acid content incorporated with the K2S2O8, containing water
and acid as impurities.37 Addition of buffers was avoided, since
their components may react with SO4•- radical ions.34 The ionic
strength of the solutions was within the range of 0.005-0.2 M.
Solutions of benzene were prepared by dilution of a saturated
aqueous solution at 25 °C. For experiments performed in the
absence of molecular oxygen, the benzene-saturated aqueous
solutions and the water used for dilution were bubbled with N2
or Ar. The concentrations of benzene and chloride ions used in
all experiments fulfilled the condition that reaction of benzene
with sulfate radicals can be neglected.
(19) Adams, D. J.; Barlow, S.; Buxton, G. V.; Malone, T.; Salmon, G.
A. J. Chem. Soc., Faraday Trans. 1995, 91, 3303.
(20) (a) Langmuir, M. E.; Hayon, E. J. J. Phys. Chem. 1967, 71, 3803.
(b) Wu, D.; Wong, D.; di Bartolo, B. J. Photochem. 1980, 14, 303.
(21) Huie, R. E. In AdVanced Series in Physical Chemistry; Barker, J.
R., Ed.; World Scientific: New Jersey, 1995; vol. 3.
(22) Huie, R. E.; Clifton, C. L. J. Phys. Chem. 1990, 94, 8561.
(23) (a) Gonzalez, M. C.; Braun, A. M. Res. Chem. Intermed. 1995,
21, 837. (b) Hindmarsh, A. C. ACM SIGNUM Newsletter 1980, 15, 11-
19. (c) Ma´rtire, D. O.; Gonzalez, M. C. Int. J. Chem. Kinet. 1997, 29, 589.
(24) Nagarajan, V.; Fessenden, R. W. J. Phys. Chem. 1985, 89, 2330.
(25) Zimmerman, G.; Strong, F. C. J. Am. Chem. Soc. 1956, 79, 2063.
(26) Neta, P.; Huie, R. E.; Ross, A. B. J. Phys. Chem. Ref. Data 1988,
17, 1146.
(27) Jacobi, H. W.; Wicktor, F.; Herrmann, H.; Zellner, R. Int. J. Chem.
Kinet. 1999, 31, 169.
(28) (a) Steenken, S. In Free Radicals in Synthesis and Biology. NATO
ASI Series C 260; Minisci, F., Ed.; Kluwer Academic Publishers: Dordrecht,
1989; p 220. (b) Merga, G.; Rao, B. S. M.; Mohan, H.; Mittal, J. P. J.
Phys. Chem. 1994, 98, 9158. (c) Mohan, H.; Mudaliar, M.; Aravindakumar,
Organic products were extracted from the aqueous solutions
with a fixed volume of chloroform and the extracts stored in
glass vials with PTFE/silicone septum-lined screw caps and
minimized headspace. Analysis of the extracts was performed
by gas chromatography with a HP 6890 chromatograph equipped
with a fused silica HP5-MS GC capillary column and coupled
to an HP 5973 mass selective detector. The analysis was
performed by using a temperature program starting at 80 °C
and ending at 200 °C at a rate of 10 °C/min and held at 200 °C
for 5 min. Helium was used as carrier gas with a flow rate of
29 cm3/s. Injection volumes were of 20 µL.
Mass spectra and probable structures of detected products
are shown in Table 4. The pH of the samples was periodically
controlled with a Methrom-Herisau pH meter model E512. The

•-
Reactions of Chlorine Atoms and Cl2 Radical Anions
J. Phys. Chem. A, Vol. 104, No. 14, 2000 3125
C. T.; Rao, B. S.; Mittal, J. P. J. Chem. Soc., Perkin Trans. 2 1991, 2,
1387.
(29) (a) Land, E. J.; Ebert, M. Trans. Faraday Soc. 1967, 63, 1181. (b)
Draper, P. B.; Fox, M. A.; Pelizetti, E.; Serpone, N. J. Phys. Chem. 1989,
93, 1938.
(30) Cencione, S. S.; Gonzalez, M. C.; Ma´rtire, D. O. J. Chem. Soc.,
Faraday Trans. 1998, 94, 2933.
(33) Mohan, H.; Mittal, J. P. J. Chem. Soc., Faraday Trans. 1995, 91,
2121.
(34) Neta, P.; Huie, R. E.; Ross, A. B. J. Phys. Chem. Ref. Data 1990,
19, 413.
(35) (a) Sommeling, P. M.; Mulder, P.; Louw, R.; Avila, D. V.; Lusztyk,
J.; Ingold, K. U. J. Phys. Chem. 1993, 97, 8361. (b) Alfasi, Z. B.; Maguet,
S.; Neta, P. J. Phys. Chem. 1994, 98, 8019.
(31) (a) Fang, X.; Pan, X.; Rahmann, A.; Schuchmann, H. P.; von
Sonntag, C. Chem. Eur. J. 1995, I, 423. (b) von Sonntag, C.; Schuchmann,
H. P. Angew. Chem., Int. Ed. Engl. 1991, 30, 1229.
(32) Buxton, G. V.; Greenstock, C. L.; Helman, W. P.; Ross, A. B. J.
Phys. Chem. Ref. Data 1988, 17, 513.
(36) Ashton, L.; Buxton, G. V.; Stuart, C. J. Chem. Soc., Faraday Trans.
1995, 91, 1631.
(37) Riedel de Ha¨en (Germany). Personal communication. 1998.
(38) The error bars for the values reported here are in all cases considered
for a 99% confidence level.