Reactions of carbon dioxide radical anion with substituted benzenes

Article abstract of DOI:10.1002/poc.365

We report mechanistic studies on the reactivity of CO₂^·- radicals towards substituted benzenes, XPh, with X = NO₂, COOH, COH, CH₃, H, Cl and HO. Vacuum ultraviolet (VUV, 172 nm) irradiation of aqueous solutions containing formate was used as a method of CO₂^·- radical generation for the study of the reaction products, which were determined by gas chromatography-mass spectrometry and high-performance liquid chromatography. The nature of some of the reaction intermediates was studied by laser flash photolysis of aqueous solutions of peroxodisulfate in the presence of formate. The observed products and intermediates of reaction are discussed and rationalized by a dual reactivity of the CO₂^·- radical anion with substituted benzenes: a one-electron reduction and radical addition to the aromatic ring. For substituted benzenes with X = NO₂, COOH and COH, the observed products support an electron transfer from CO₂^·- to the substrates as the primary reaction channel, while the reactions with toluene follow the channel of radical addition. Benzene and chlorobenzene react with CO₂^·- by both routes, electron transfer and radical addition. Our experimental results indicate that H abstraction by the CO₂^·- radical is negligible under the chosen experimental conditions. Copyright

Full text of DOI:10.1002/poc.365

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY
J. Phys. Org. Chem. 2001; 14: 300–309
DOI:10.1002/poc.365
Reactions of carbon dioxide radical anion with substituted
benzenes
Janina A. Rosso,1 Sonia G. Bertolotti,2 Andre M. Braun,3 Daniel O. MaÂrtire1* and MoÂnica C. Gonzalez1*
1Instituto de Investigaciones FisicoquõÂmicas Teo ricas y Aplicadas ꢀINIFTA), Facultad de Ciencias Exactas, Universidad Nacional de La Plata,
Casilla de Correo 16, Sucursal 4, ꢀ1900) La Plata, Argentina
2Universidad Nacional de RÂõo Cuarto, Co rdoba, Argentina
3Lehrstuhl fuÈ r Umweltmesstechnik, UniversitaÈ t Karlsruhe, D-76128 Karlsruhe, Germany
Received 09 November 2000; revised 26 December 2000; accepted 03 January 2001
ABSTRACT: We report mechanistic studies on the reactivity of CO2.À radicals towards substituted benzenes, XPh,
with X = NO2, COOH, COH, CH3, H, Cl and HO. Vacuum ultraviolet (VUV, 172 nm) irradiation of aqueous solutions
containing formate was used as a method of CO2.À radical generation for the study of the reaction products, which
were determined by gas chromatography–mass spectrometry and high-performance liquid chromatography. The
nature of some of the reaction intermediates was studied by laser flash photolysis of aqueous solutions of
peroxodisulfate in the presence of formate. The observed products and intermediates of reaction are discussed and
rationalized by a dual reactivity of the CO2.À radical anion with substituted benzenes: a one-electron reduction and
radical addition to the aromatic ring. For substituted benzenes with X = NO2, COOH and COH, the observed products
.À
support an electron transfer from CO2 to the substrates as the primary reaction channel, while the reactions with
toluene follow the channel of radical addition. Benzene and chlorobenzene react with CO2.À by both routes, electron
transfer and radical addition. Our experimental results indicate that H abstraction by the CO2.À radical is negligible
under the chosen experimental conditions. Copyright  2001 John Wiley & Sons, Ltd.
KEYWORDS: carbon dioxide radical anion; vacuum ultraviolet irradiation; electron transfer; radical addition;
phenol; toluene; benzene; chlorobenzene; benzaldehyde; benzoic acid
INTRODUCTION
The exact determination of the reduction potential, E°,
.À
of the couple CO2/CO2 was difficult because of the
Carbon dioxide radical anions, CO2.À, are commonly
used in aqueous chemistry as a reducing agent for
metalloporphyrins or as intermediate in the formation of
irreversibility of the electron transfer process due to the
change in structure from linear to bent.12 Results of
different approximations and experimental methods
agree in a value of E° (CO2/CO2.À) ꢀ À1.9 V.12,13
Despite the fact that the CO2.À radical is characterized by
.À
superoxide anion. CO2 has been reported to undergo
efficient electron transfer reactions with methyl violo-
gen,1 quinones,2 alkyl halides,3 fumarates,4 nitro and
nitrosobenzenes5,6 and chlorinated benzaldehydes.7 With
nitrobenzenes and chlorinated benzaldehydes, electron
attachment occurs on the nitro and aldehyde groups,
.À
a more negative reduction potential than O2 radicals
[E°(O2/O2.À) ꢀ À0.16 V], in agreement with its efficient
reaction with oxygen,2 the reaction of CO2.À with sulfur
radical cation complexes14 is slower by a factor of 2.6.
The trend observed in reactivity is in line with theoretical
calculations15 predicting a high internal reorganization
.À
respectively. CO2 radicals have also been reported to
add to some unsaturated compounds such as acrylamide8
and pyridin-3-ol.9 Efficient hydrogen abstraction from
mercaptobenzenes have also been reported.10,11
energy (ꢀ = 447 kJ molÀ1) for the couple CO2/CO2
.À
than for the couple O2/O2 (ꢀ = 88.6 kJ molÀ1).
.À
The estimated value for the reduction potential of the
*Correspondence to: D. O. Ma´rtire or M. C. Gonzalez, Instituto de
‡
À
couple CO2.À, H /HCO2 is 1.07 V at pH 7.12 The fact
Investigaciones Fisicoqu´ımicas Teo´ricas
y Aplicadas (INIFTA),
.À
that CO2 is also able to oxidize hydrogencarbonate to
Facultad de Ciencias Exactas, Universidad Nacional de La Plata,
Casilla de Correo 16, Sucursal 4, (1900) La Plata, Argentina.
E-mail: dmartire@volta.ing.unlp.edu.ar; gonzalez@inifta.unlp.edu.ar
Contract/grant sponsor: Fundacio´n Antorchas, Agencia Nacional de
Promocio´n Cientı´fica y Technolo´gica (ANPCyT).
CO3 (k = 2 Â 103 l molÀ1 sÀ1 16a
)
although the reduc-
.À
2À
tion potential for the couple CO3.À/CO3 is E° = 1.5
V,16b is an indication that CO2 radicals may oxidize
.À
Contract/grant sponsor: State of Baden-Wu¨rttemberg.
Contract/grant sponsor: Deutsche Akademische Austauschdieinst
(DAAD).
some classes of substrates by hydrogen abstraction and/or
by an addition–elimination pathway.
Contract/grant sponsor: International Office of BMBF.
Dimerization to oxalate, reaction (R1), seems to be the
Copyright  2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 300–309

REACTIONS OF CO2 RADICAL ANION WITH BENZENES
301
.À
.À
.À
2À
main reaction of CO2 in the absence of added
substrates.5
CO2 radicals. Reaction of CO2 radicals with S2O8
ions,22 reaction (R7), yields again SO4 radicals as
.À
reaction products, and its main effect is that of depleting
formate and peroxodisulfate from the reaction system.
ÁÀ
ÁÀ
2À
CO2 ‡ CO2 ! C2O4
ꢁR1†
ÁÀ
À
2À
ÁÀ
‡
SO4 ‡ HCO2 ! SO4 ‡ CO2 ‡ H
.À
The reaction channels of CO2 radicals in the presence
of organic substrates depend strongly on the particular
substrate and need still to be investigated in detail. In this
paper we report mechanistic studies on the reactivity of
k6 ˆ 1:1  108 l molÀ1 sÀ1
ꢁR6†
ꢁR7†
2À
ÁÀ
ÁÀ
2À
S2O8 ‡ CO2 ! SO4 ‡ SO4 ‡ CO2
k7 ˆ 1  105 l molÀ1 sÀ1
.À
CO2 radicals towards substituted benzenes, XPh, with
X = HO, CH3, H, Cl, COH, COOH and NO2.
CO2.À radicals show absorption in the UV (ꢀmax = 235
nm) with a relatively high extinction coefficient [e (235
nm) = 3000 l molÀ1 cmÀ1], independent of pH between
The reaction products obtained from the reaction of
.À
CO2 with the aromatic substrates with reaction (R8)
were determined by gas chromatography–mass spectro-
metry (GC–MS) and high-performance liquid chroma-
tography (HPLC). The nature of some of the reaction
intermediates was studied by laser flash photolysis.
pH 3 and 13.17 The conjugate acid of CO2.À, i.e., HCO2 ,
.
has a pKa value of 1.4.2
.À
For our investigations, CO2 radicals were generated
in aqueous systems by two different ways:
(i) by reaction of HO. radicals and H atoms with formic
acid (pKa = 3.75) or formate ions, reactions (R2–R5):2
ÁÀ
CO2 ‡ XPh ! products
ꢁR8†
À
ÁÀ
HOÁ ‡ HCO2 ! H2O ‡ CO2
k1 ˆ 3:2  109 l molÀ1 sÀ1
ꢁR2†
ꢁR3†
ꢁR4†
ꢁR5†
RESULTS ANDDISCUSSION
Á
HOÁ ‡ HCO2H ! H2O ‡ HCO2
.À
Generation of CO2 by VUV irradiation of aqueous
solutions requires efficient scavenging by formic acid and
formate ions [reactions (R2–R5)] of HO. and H. which are
primarily formed by photolysis. Under the chosen
k2 ˆ 1:3  108 l molÀ1 sÀ1
HÁ ‡ HCO2 ! H2 ‡ CO2
À
ÁÀ
À
experimental conditions, eaq are efficiently quenched
k3 ˆ 2:1  108 l molÀ1 sÀ1
by H3O . Other reactions of H., HO. and eÀaq, such as
‡
HÁ ‡ HCO2H ! H2 ‡ HCO2
Á
recombination reactions leading to H2 and H2O2,
respectively, and with the substituted benzenes should
be negligible.
k4 ˆ 6:1  107 l molÀ1 sÀ1
The high absorbance of water (kH O, 172 nm
=
2
Vacuum ultraviolet (VUV) photolysis of liquid water
(ꢀexc <185 nm) provides a clean source of HO. radicals
and H atoms and, with much lower efficiency, hydrated
electrons (eaqÀ).18 Formate reacts efficiently with HO.
and H., as shown in reactions (R2–R5). The quantum
yields of hydrated electron formation, f(eaqÀ), are very
low, and in acidic solutions of pH ꢂ4 the concentration of
550 cmÀ1), the quantum yield of water homolysis
(ΦH O, 172 nm = 0.42 Æ 0.04 mol einsteinÀ1 23
)
and the
2
short lifetime of H atoms and HO radicals lead to strong
heterogeneity in the macroscopically homogeneous
reaction system. A volume of primary reactions close
to the surface of the lamp, in which the rates of reactions
(R2–R5) are controlled by the diffusion of formic acid/
formate ions into this layer, and a large volume of
secondary reactions in which slower thermal reactions
dominate, may be differentiated. To avoid undesirable
competing reactions as those already mentioned (see
above), high local concentrations of formic acid/formate
ions in the proximity of the surface of the light source are
needed. Taking into account that HO. radicals and H
atoms react mainly with formate, the minimum average
concentration of formate in the reaction system,
[CO2HÀ]min, which ensures that its concentration on the
surface of the lamp remains above zero, may be estimated
according to Eqn. (1),23 assuming steady-state conditions
in the volume of primary reactions and, hence, involving
equal rates of formate depletion [reactions (R2–R5)] and
diffusion. In a first approximation, formate/formic acid
À
eaq is furtÀher decreased owing to the efficient scaven-
‡
ging of eaq by H3O to yield H atoms (k = 2.4 Â 1010
l molÀ1 sÀ1 19
(see below). Consequently, VUV irradia-
)
tion (ꢀexc = 172 nm) of N2-saturated aqueous solutions
(pH ꢀ 4) containing formic acid/formate ions yields
.À
mainly CO2 radicals and this generation method was
used for experiments under steady-state conditions.
(ii) Sulfate radicals (SO4.À) efficiently react with
formate ions yielding CO2 radicals,20 reaction (R6).
.À
The photolysis of peroxodisulfate, S2O82À, providing a
.À
clean source of SO4 radical anions with high pH-
independent quantum yields,21 laser flash photolysis
(ꢀexc = 266 nm) of Ar-saturated peroxodisulfate solutions
of pH 5 in the presence of formate ions was used as a
convenient method for time-resolved studies involving
Copyright  2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 300–309

302
J. A. ROSSO ET AL.
from saturated aqueous solutions of pH 4 containing
different [HCO2À]0.
À
The depletion rates of toluene observed for [HCO2
]
0
ꢃ0.2 M are coincident, within experimental error. This
observation indicates that under these conditions, the
CO2.À formation rates are independent of [HCO2À]0, and
HO. and H. formed by water homolysis are quantitatively
scavenged by formic acid/formate ions, in agreement
with our previous discussion (see above).
The choice of [HCO2À]0 = 1 M then guarantees that
HO. and H atoms are quantitatively scavenged by
HCO2H/HCO2À, even in the presence of efficient HO.
and H. scavengers and for long irradiation periods, where
even after a 50% consumption of [HCO2À]0 the depletion
rates of toluene are not affected.
Figure 1. Depletion of toluene during VUV irradiation of
saturated aqueous solutions of toluene containing different
analytical concentrations of formate ion: 0 ꢀ*); 0.01 ꢀ*);
0.02 ꢀ~); 0.1 ꢀ~); 0.2 ꢀ&); 0.5 ꢀ&); 1 ꢀ^); and 1.78 M ꢀ+).
The effect of synthetic air bubbling on the concentration of
toluene ꢀ!) is also shown
Owing to the high concentrations of formic acid/
formate ions used, VUV absorption by these substances
may be suspected to compete with that of water. In fact,
formic acid is known to absorb in the gas phase at
wavelengths >180 nm, e.g. "200 nm = 40 l molÀ1 cmÀ1
,
and below 180 nm several Rydberg transitions have been
found leading to an ionization potential of 11.33 eV
(corresponding in energy to light of 109 nm).27–29
Photolysis of formic acid in the gas phase has been
reported to undergo two main reaction channels [e.g.
reactions (R9–R11)], of which reaction (R9) is the most
important process for irradiation wavelengths from 190
to 234 nm.28,29
photolysis may be neglected (see below).
2 Â 103 Â ÈH O Â P0
À
2
‰HCO2
Š
ˆ
ꢁ1†
min
À
DHCO  kH O;172 nm  ln 10
2
2
HCO2H ‡ hꢁ ! HOÁ ‡ HÁCO
HÁCO ! CO ‡ HÁ
ꢁR9a†
ꢁR9b†
ꢁR10†
ꢁR11†
Taking the incident photon rate per unit area
P0 = 8.5 Â 10À8 einstein sÀ1 cmÀ2 18,23
,
and for the
diffusion coefficient, DHCO À, itsÀvalue at infinite dilution
2
(1.45 Â 10À5 cm2 sÀ1),24 [HCO2
]
is estimated to be
min
HCO2H ‡ hꢁ ! HÁ ‡Á CO2H
4 Â 10À3 M. It has been shown in studies using methanol
as a quencher of HO. radicals, that an excess factor
n = [MeOH]0/[MeOH]min >70 is required for a scaven-
ging reaction of HO. radicals.23 If a similar ratio is
expected for formate, the required concentration of
formate ion in the reaction system should be >0.3 M.
In order to check the validity of this estimation, and to
establish the experimental conditions under which the
added organic substrate, XPh, will not react with HO.
radicals and H atoms, a series of experiments were
performed to investigate the depletion rate of XPh as a
function of the analytical initial concentration of formate
[HCO2À]0. Toluene was used as the organic probe, since
ÁÀ
‡
ÁCO2H ! CO2 ‡ H
Photolysis at 253.7 nm of formate/formic acid in
aqueous solutions has been reported to lead to dissocia-
tion producing HO. and H.CO, reaction (R9a). Formyl
radicals, H.CO, efficiently dissociate to CO and H atoms,
reaction (R9b).30 Thus, electronic excitation of formic
.À
acid/formate results in the final formation of CO2 also
involving reactions (R2–R5).
Our experiments show no effect on the decay rate of
toluene when [HCOOÀ]0 is increased to >0.2 M. This
it efficiently reacts with HO. (k = 5.1 Â 109 l molÀ1 sÀ1 25
)
result indicates that similar CO2 concentrations are
.À
and H. (k = 2.6 Â 109 l molÀ1 sÀ1 26
)
and is expected to
formed. If photolysis of formate/formic acid were to
contribute significantly to the formation of CO2.À, light
should be quantitatively absorbed by formic acid already
at [HCO2À]0 = 0.2 M and, consequently, the formate/
formic acid absorption coefficient at 172 nm should be
much higher than 2.7 Â 104 l molÀ1 cmÀ1 owing to the
even lower local concentrations of formate/formic acid
near the surface of the lamp. We therefore conclude that
formate/formic acid photolysis is not a significant process
under our experimental conditions.
show relatively low reactivity towards the electron-rich
.À
CO2 radicals.
In the absence of formate, mainly reactions with HO.
radicals and H atoms deplete toluene. In the presence of
formate, both toluene and formate compete for HO. and
H.. Under conditions of high formate to toluene molar
ratios, the reactive intermediates are quantitatively
scavenged by formate and toluene reacts primarily with
CO2.À. Figure 1 shows the depletion of toluene obtained
Copyright  2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 300–309

REACTIONS OF CO2 RADICAL ANION WITH BENZENES
Reaction products
303
Table 1. Reaction products
Substrate
Nitrobenzene
Benzaldehyde
Benzoic acid
Benzene
Nitrosobenzene, aniline, N-phenylformamide, azoxybenzene, azobenzene
Benzyl alcohol
Benzaldehyde
Benzaldehyde, 1,4-cyclohexadiene, phenol, benzoic acida, biphenyl
2-Methylphenol, 3,3'-dimethylbiphenyl
Phenol, 2-chlorophenol, benzaldehyde
Toluene
Chlorobenzene
Phenol
b
–
a
Determined by HPLC. Identified by its retention time and UV absorption spectrum.
b No products were detected by GC. One unidentified product was detected by HPLC.
In the absence of aromatic substrÀates and for
[HCO2À]0 = 1 M, diminution of [HCO2 ] follows a
zero-order decay of (8 Æ 3) Â 10À5 mol lÀ1 sÀ1 for
experiments made with 400 ml of solution. Conse-
quently, ꢄ3 Â 10À5 mol sÀ1 of formate are consumed
by its reactions with HO. and H atoms near the surface of
the lamp. Taking into account that 2 mol of formate are
consumed per mole of water homolysed and that
Φ
H O = 0.42 mol einsteinÀ1 at 172 nm, P0 is estimated
2
to be (1.7 Æ 0.7) Â 10À7 einstein sÀ1 cmÀ2, as expected
for Xe lamps of this type (see above).
.À
As CO2 radicals are formed from formate/formic
acid due to reactions (R2–R5), ꢄ1.4 Â 10À7 mol cmÀ2
s
À1 of CO2.À radicals are produced near the surface of the
.À
lamp. Depletion of CO2 in the primary volume of
reactions is governed by their second-order recombina-
Scheme 1
Copyright  2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 300–309

304
J. A. ROSSO ET AL.
tion, reaction (R1),31 and by reaction with the organic
matter, reaction (R8), hence severely reducing the steady
.À
concentration of CO2 which under such conditions
cannot diffuse far out of this primary volume. In order to
understand the complex kinetic process taking place in
the primary (irradiated) volume of reactions, a differ-
ential equation system should take into account the
diffusion of CO2.À and of the organic matter to and from
.À
the bulk solution (secondary volume), the rate of CO2
formation and reactions (R1) and (R8). Assuming the
steady-state condition for CO2.À radicals and the organic
matter in the primary volume of reactions, the product
k8 Â [CO2 ]ss is estimated to be of the order of 6 Â 10À2
.À
sÀ1 for 109 >k8 (l molÀ1 À1) >106. The macroscopic
s
depletion rate of aromatic substrates is then controlled by
the rate of reaction (R8) in the primary volume of
reaction and diffusion of the organic substrates from the
bulk solution to the neighbourhood of the lamp surface.
Under such conditions, substrate depletion rates of the
order of 10À6 mol sÀ1 are expected, as observed experi-
mentally for all the substrates investigated, except for
phenol, for which an even lower depletion rate is found.
The reaction products formed after the reaction of
Figure 2. Absorption spectrum of the transient obseÀrv2ed
S2O82À acid solution ꢀpH 5) containing 0.4 M formate ion and
3 mM nitrobenzene
3 ms after laser excitation of an Ar-saturated 2.25 Â 10
M
at ꢄ285 nm, in agreement with the reported spectrum of
the radical anion of nitrobenzene.6 The observed transient
formed by reaction (R12) is therefore assigned to the
radical anion of nitrobenzene.
.À
CO2 with the substituted benzenes were determined
either by GC–MS or HPLC under the described
experimental conditions. Products were analysed for
different irradiation times up to 10 min and for <10%
reactant consumption in order to avoid secondary
reactions of the reaction products. The products formed
from the substrates investigated and properly identified
are listed in Table 1. Aliphatic compounds of low
molecular weight, i.e. containing four carbon atoms or
less, are eluted with the solvent and could not be detected
with our chromatographic set-up.
Formation of nitrosobenzene may be explained by
disproportionation of the radical anion of nitrobenzene to
nitrobenzene and nitrosobenzene,32 reaction (R13).
.À
Nitrosobenzene is known to react with CO2 radical
anions with k = 4 Â 109 l molÀ1 sÀ1 6
,
yielding the
nitrosobenzene radical, reaction (R14), which under our
experimental conditions is present in its acid form
(pKa = 11.76). Further condensation of the acid form of
the nitrosobenzene radical with loss of water yields
azoxybenzene, reaction (R15). Disproportionation of the
conjugate acid of nitrosobenzene radical, reaction (R16),
may yield nitrosobenzene and phenylhydroxylamine. The
latter product was not observed, however, it may have
been further reduced by CO2.À to the intermediate amine
radical, C6H5N.H, which subsequently disproportionates
to phenylhydroxylamine and aniline, reactions (R17) and
(R18), respectively.
NITROBENZENE
Concerning our investigations with nitrobenzene, the
formation of products involving a reduction process
(Table 1), can be explained if an electron transfer from
.À
CO2 to nitrobenzene, reaction (R12), is the primary
reaction (Scheme 1)
Formation of N-phenylformamide at longer irradiation
times may arise, either from addition of CO2 to the
.À
Flash photolysis experiments with Ar-saturated per-
oxodisulfate solutions containing formate and nitroben-
zene showed the formation of a transient species in the
microsecond time range with an absorption maximum of
290–280 nm. Under the conditions of our experiments,
nitrogen atom of the amine radical followed by reduction, or
by a condensation reaction between aniline and formic acid.
Formation of azobenzene, which is observed after
5 min of photolysis, may be explained by condensation of
the amine radical with the nitrosobenzene radical,
reaction (R19), or by the reaction of the amine radical
with nitrosobenzene and further reduction.33
.À
SO4 radicals generated from the photolysis of perox-
odisulfate readily react with formate ions according to
reaction (R6) (with a lifetime tSO .À = 15 ns) yielding
4
CO2.À radical anions. Absorption of CO2.À radicals is not
observed in our experiments owing to the high absorption
.À
of the solutions at ꢀ <250 nm. CO2 radicals further
BENZALDEHYDE
react with nitrobenzene according to reaction (R12)
(tCO .À = 1 ms) leading to the observed transient. Figure 2
shows the transient absorption spectrum with a maximum
For benzaldehyde as a substrate, benzyl alcohol is the
only observed reaction product. Reaction of CO2
2
.À
Copyright  2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 300–309

REACTIONS OF CO2 RADICAL ANION WITH BENZENES
305
Scheme 2
Scheme 3
radicals with benzaldehyde, reaction (R20) (Scheme 2),
was reported to proceed by electron attachment on the
CHO group, yielding the carbonyl radical anion.7
Reversible protonation of the carbonyl radical anion
yields the carbonyl radical showing a pKa of 8.4.7 Further
disproportionation of the carbonyl radical yields benzyl
alcohol and benzaldehyde, reaction (R21).
reaction (R22) (Scheme 3), as already observed with the
hydrated electron.34 Further disproportionation of the
benzoyl radical produces benzoic acid and benzaldehyde
as shown in reaction (R23).
Benzene
The reaction of benzene with CO2.À leads to both reduced
and oxidized products. Formation of 1,4-cyclohexadiene
may be explained by an electron transfer from CO2 to
the aromatic ring leading to the radical anion of benzene,
reaction (R24) (Scheme 4). Subsequent reaction with
water leads to the intermediate formation of cyclohexa-
BENZOIC ACID
.À
.À
The reaction of CO2 with benzoic acid leads to
benzaldehyde. The electron transfer from the CO2
radical to the carboxyl group yields the benzoyl radical,
.À
Scheme 4
Copyright  2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 300–309

306
J. A. ROSSO ET AL.
.À
the reaction of CO2 with benzene, flash photolysis
experiments with Ar-saturated peroxodisulfate solutions
containing formate and benzene were performed. For-
mation of a transient species with an absorption maxi-
mum at ꢄ300 nm was observed, as shown in Fig. 3.
The radical anion of benzene is reported to present an
absorption maximum at 290 nm in aqueous solution and
two absorption maxima at approx. 285 and 385 nm in a
matrix at 77 K.40,41 Since the radical anion of benzene is
known to protonate to H–CHD on a microsecond time-
scale,42 the latter radical may be detected with our laser
equipment. Figure 3 shows the spectrum of the transient
species obtained, along with that of the H–CHD and HO–
CHD radicals of benzene42,43 whose participation in the
overall reaction has already been discussed (see above).
The observed transient shows important contributions of
the H–CHD and HO–CHD radicals; however, absorption
at wavelengths higher than 340 nm cannot entirely be
attributed to these radicals.
Figure 3. Absorption spectrum of the transient obseÀrv2ed
3 ms after laser excitation of an Ar-saturated 2.25 Â 10
M
S2O82À acid solution ꢀpH 5) containing 0.4 M formate ion and
1 Â 10À2 M benzene. The solid and dashed ꢀÐ..Ð) lines
show the literature absorption spectra of the HO±CHD43 and
H±CHD42 radicals of benzene, respectively. The inset shows
the literature spectrum of the H±CHD of benzoic acid.34
The formation of CO2À–CHD has been postulated in
order to account for the generation of benzoic acid as a
reaction product (see above). We expect that the lifetime
of this radical is longer than 50 ms in order to be able to
undergo disproportionation to benzoic acid. The ob-
served absorption at ꢀ >340 nm may in fact be attributed
to the CO2À–CHD radical, because of the agreement of
its absorption spectrum with that of the structurally
related H adduct of benzoic acid showing, both,
absorption maxima at around 350 nm.34 Despite the fact
that Ph–CHD phenylcyclohexadienyl radicals also show
an absorption with a maximum between 330 and 340
nm,39 its contribution to the observed spectrum is
neglected, because the recombination product, phenylcy-
clohexadiene, was not found (see above). Consequently,
the time-resolved experiments with benzene support
both the addition and electron transfer pathways of
reaction.
dienyl radicals, H–CHD, which undergo subsequent
disproportionation to benzene and 1,4-cyclohexadiene,
reaction (R25). A similar reaction sequence was proposed
for the reaction of benzene with hydrated electrons.27
The formation of benzoic acid, benzaldehyde, biphe-
.À
nyl and phenol may be explained by an addition of CO2
to benzene, reaction (R26). Disproportionation of the
adduct, CO2À–CHD, yields benzene with elimination of
formate, and benzoic acid, reaction (R27). Benzaldehyde
would then be formed by the subsequent reduction of
benzoic acid (see above). Alternatively, the CO2À–CHD
adduct may eliminate formate and add water, hence
producing a hydroxycyclohexadienyl radical (HO–
CHD), reaction (R28), in agreement with the observed
reactivity of sulfate and phosphate radical adducts of
aromatic substrates in aqueous solution35 and aÀs e3x5dpected
for the relatively good leaving group HCO2
.
HO–
Toluene
CHD can either disproportionate to yield phenol and
benzene with elimination of water, reaction (R29), or
recombine yielding biphenyl36–38 as shown in reaction
(R30).
.À
The reaction of CO2
with toluene leads to 2-
methylphenol and 3,3'-dimethylbiphenyl. Both products
may be formed by an addition – elimination reaction
channel leading to the HO–CHD radical of toluene,
which may either disproportionate to 2-methylphenol and
toluene with elimination of water or recombine to
dimethylbiphenyl, as observed for benzene in reactions
(R26), (R28), (R29) and (R30).
.À
H-abstraction by CO2 from the aromatic ring could
also lead to the formation of benzoic acid, if the
intermediate phenyl radical were subsequently to react
with CO2.À. However, phenyl radicals are highly reactive
and may alternatively react with benzene (k = 4.5 Â 105
l molÀ1 À1) to yield a phenylcyclohexadienyl radical
s
(Ph–CHD), which upon disproportionation would lead to
biphenyl (observed as a reaction product, see above) and
phenylcyclohexadiene.38,39 The latter product not being
detected in our experiments, we conclude that biphenyl is
formed by an addition mechanism rather than by an H-
abstraction.
Chlorobenzene
.À
The reaction of CO2 with chlorobenzene leads to 2-
chlorophenol, benzaldehyde and phenol. An electron
transfer from CO2 to the aromatic ring is expected to
.À
In order to obtain information on the intermediates of
lead to the dehalogenation of the aromatic substrate, as
Copyright  2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 300–309

REACTIONS OF CO2 RADICAL ANION WITH BENZENES
307
Scheme 5
observed for the reaction of chlorobenzene with the
Phenol
hydrated electrons.44 The dehalogenation of chloroben-
.À
zene yields reactive phenyl radicals, reaction (R31),
The reaction of CO2 with phenol resulted in a
surprisingly small substrate depletion. No products were
found within the detection limits of our chromatographic
means (GC), except for one unidentified product detected
by HPLC.
.À
which upon further reaction with CO2
radicals
generates benzoic acid, reaction (R32) (Scheme 5).
.À
Further reduction of benzoic acid by CO2 leads to the
formation of benzaldehyde, an observed reaction product
(see above).
Formation of chlorophenol may be explained by an
addition–elimination reaction leading to an HO–CHD
adduct of chlorobenzene which disproportionates to
chlorophenol and chlorobenzene, as shown in reactions
(R33) and (R34).
Formation of phenol very likely requires an ipso
addition of CO2.À to chlorobenzene, as shown in reaction
(R35). The ipso position was also reported to be the
preferred site for hydroxyl radical addition to chloro-
benzene.45 The bimolecular reaction of the ipso adduct
CO2ÀCHD with HO–CHD radicals of chlorobenzene is
an alternative reaction channel yielding benzoic acid,
HCl aÀnd chlorophenol, reaction (R36). Elimination of
HCO2 from the ipso adduct, reaction (R37), and
subsequent reaction with HO–CHD radicals leads to the
formation of phenol, HCl and chlorophenol,46 reaction
(R38).
CONCLUSIONS
Products and reaction intermediates observed in the
present study are rationalized by a dual reactivity of the
.À
CO2 radical anion with substituted benzenes, i.e. one-
electron reduction of and radical addition to the aromatic
ring. For substrates with electron-withdrawing substitu-
ents, i.e., with positive s values such as X = NO2, COOH
and COH, the observed products can be explained by an
.À
electron transfer from CO2 to the substrates as the
primary reaction channel. Benzene and chlorobenzene
react by both electron transfer and radical addition. For
toluene, with an electron-donating substituent, products
of radical addition are found. The dual reactivity of
.À
CO2 radical anion has recently been reported for
thymine,47 in agreement with our observations.
Copyright  2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 300–309

308
J. A. ROSSO ET AL.
H abstraction by the CO2.À radical is only supported by
analyses were performed with a temperature program
starting at 80°C and ending at 200°C at a rate of 10°C
minÀ1, and the final temperature being held at 200°C for
5 min. Helium was used as the carrier gas at a flow-rate of
29 cm3 sÀ1. Injection volumes were 20 ml.
Substrate depletion and product formation were also
followed by HPLC with an HP1090 Series II liquid
chromatograph with a diode-array detector. A 10%
methanol–water mixture was used as eluent.
one of the routes proposed for the formation of benzoic
acid in our experiments with benzene. If such a reaction
channel were of importance, formation of substituted
benzoic acids should have been observed for other
substituted benzenes, as the dissociation energy of the
C—H bonds in these compounds is expected to be of the
same order as in benzene. Moreover, formation of
products derived from the benzyl radical was not
observed in our experiments with toluene. We therefore
conclude that the contribution of H-abstraction channel
Formate analysis was performed by ion chromatogra-
phy (Dionex DX 500) with suppressed conductivity
detection. A 250 Â 4 mm i.d. IONPAC AS11 (Dionex)
column and a 50 Â 4 mm i.d. IONPAC AG11 (Dionex)
guard were used. The eluent was 60 mM aqueous NaOH.
.À
by CO2 radical ions as a potential reaction channel of
this intermediate is negligible under our experimental
conditions.
The observed tendency in the dual reactivity of CO2
The flow-rate was 1 ml minÀ1
.
.À
radical anion may be rationalized by the reduction
potential of the substrate. An electron transfer mechan-
ism is observed for those substituted benzenes which are
reduced by the hydrated electron with diffusion-con-
trolled rates.27 For substrates showing lower reactivity
towards the hydrated electron, i.e. which are more
Laser photolysis experiments. Flash photolysis experi-
ments were performed with a Spectron SL400 Nd:YAG
system generating 266 nm pulses (ꢄ8 ns pulse width).
The laser beam was defocused in order to cover the entire
pathlength (1 cm) of the analysing 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 monochromator coupled to a Hama-
matsu R666 PM tube. The signals were acquired and
averaged by a digitising oscilloscope (Hewlett-Packard
Model 54504) and transferred to a computer for further
treatment.
.À
difficult to reduce, CO2 addition is the preferred
reaction channel. These observations are in line with
studies reporting the reduction of nitro compounds by a-
hydroxyalkyl radicals, where the redox potential of the
nitro compound determines the probability of radical
addition or electron transfer.48
2À
The concentrations of S2O8 and formate were 22.5
and 400 mM, respectively, and the concentrations of the
aromatic substrates were 3 mM for nitrobenzene and
10 mM for benzene. Under these reactions, the reactions
EXPERIMENTAL
Distilled water was passed through a Millipore filter. All
the chemicals were of p.a. quality (Merck) and used as
received. All experiments were performed with oxygen-
free argon-saturated solutions.
.À
of the organic substrates with SO4 are negligible.
Acknowledgements
VUV irradiation experiments. For the VUV irradiation
experiments, an excimer lamp emitting at 172 nm
(200 mm long, 35 mm diameter18,23) was used. The lamp
was immersed in a photochemical reactor of annular
geometry and a total volume of 400 ml. The reactor was
immersed in a water-bath at constant temperature.
Aqueous solutions of the substrates in the concentra-
tion range 0.5–23 mM were used. The pH of the solutions
was adjusted to 4 by addition of HClO4. Samples were
taken at constant time intervals and the total volume of
sample withdrawn was always less than 10% of the
volume of the reactor.
This research was supported by the Fundacio´n Antorchas,
Agencia Nacional de Promocio´n Cientı´fica y Tecnolo´gi-
ca, (ANPCyT), Argentina, the State of Baden-Wu¨rttem-
berg, Germany, the Deutsche Akademische Austausch-
dienst (DAAD), Germany, and the International Office
of BMBF, Germany. J.A.R. thanks the Consejo
Nacional de Investigaciones Cient´ıficas y Te´cnicas,
(CONICET), Argentina, for a graduate studentship.
D.O.M. is a research member of the Comisio´n de
Investigaciones Cient´ıficas de la Provincia de Buenos
Aires (CIC), Argentina. S.G.B and M.C.G. are research
members of CONICET.
Chromatographic analyses. Organic substrates were
extracted from the aqueous solutions with a fixed volume
of chloroform and the extracts stored at 4°C in glass vials
with PTFE/silicone septum-lined screw-caps and mini-
mized headspace. Analysis of the extracts was performed
by GC with an HP 6890 gas chromatograph equipped
with an HP5-MS fused-silica capillary column and
coupled to an HP 5973 mass-selective detector. The
REFERENCES
1. Chitose N, Katsummura Y, Domae M, Zuo Z, Murakami T,
LaVerne JA. J. Phys. Chem. A 1999; 103: 4769.
2. Ross AB, Mallard WG, Helman WP, Buxton GV, Huie RE, Neta P.
NDRL-NIST Solution Kinetics Database: Version 3.0. Notre Dame
Radiation Laboratory: Notre Dame, IN, and National Institute of
Copyright  2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 300–309

REACTIONS OF CO2 RADICAL ANION WITH BENZENES
309
Standards and Technology: Gaithersburg, MD, 1998; http://
www.rcdc.nd.edu.
3. Beckwith ALJ, Norman ROC. J. Chem. Soc. B 400 (1969).
4. Hayon E, Simic M. J. Am. Chem. Soc. 1973; 95: 2433.
5. Buxton GV, Sellers RM. J. Chem. Soc., Faraday Trans. 1 1973;
69: 555.
6. (a) Asmus K-D, Wigger A, Henglein A. Ber. Bunsenges. Phys.
Chem. 1966; 70: 966; (b) Asmus K-D, Beck G, Henglein A,
Wigger A. Ber. Bunsenges. Phys. Chem. 1966; 70: 869.
7. Solar S, Getoff N, Holcman J, Sehested K. J. Phys. Chem. 1995;
99: 9425.
8. Chambers KW, Collinson E, Dainton FS. Trans. Faraday Soc.
1970; 66: 142.
25. Roder M, Wojnarovits L, Foldiak G. Radiat. Phys. Chem. 1990;
36: 175.
26. Sauer MC, Bard B. J. Phys. Chem. 1967; 71: 3971.
27. Bell S, Ng TL, Walsh AD. J. Chem. Soc., Faraday Trans. 2 1975;
71: 393.
28. Calvert JC, Pitts Jr JN. Photochemistry. John Wiley & Sons: New
York, 1966; 428.
29. Okabe H. Photochemistry of Small Molecules. John Wiley & Sons:
New York, 1978; 314–315.
30. Adams GE, Hart EJ. J. Am. Chem. Soc., 1962; 84: 3994.
31. (a) Fotjik A, Czapski G, Henglein A. J. Phys. Chem. 1970; 74:
3204; (b) Gennaro A, Isse AA, Save´ant J-M, Severin M-G,
Vianello E. J. Am. Chem. Soc. 1996; 118: 7190.
9. Naik DB, Moorthy PN. J. Chem. Soc., Faraday Trans. 2 1990; 705.
10. Armstrong DA, Sun Q, Triphati GNR, Schuler RH, McKinnon D.
J. Phys. Chem. 1993; 97: 5611.
11. Surdhar PS, Mezyk SP, Armstrong DA. J. Phys. Chem. 1989; 93:
3360.
12. Koppenol WH, Rush JD. J. Phys. Chem. 1987; 91: 4429.
13. Butler J, Henglein A. Radiat. Phys. Chem. 1980; 15: 603.
14. Bonifacic M, Hug GL, Scho¨neich C. J. Phys. Chem. A 2000; 104:
1240.
15. Pacansky J, Wahlgren U, Bagus PS. J. Phys. Chem. 1975; 62:
2740.
32. Piccinini P, Minero C, Vincenti M, Pelizzetti E. J. Chem. Soc.,
Faraday Trans. 1997; 93: 1993.
33. Maldotti A, Andreotti L, Molinari A, Tollari S, Penoni A, Cenini S.
J. Photochem. Photobiol. A: Chem. 2000; 113: 129.
34. Simic M, Hoffman MZ. J. Phys. Chem. 1972; 76: 1398.
35. (a) Rosso JA, Rodr´ıguez Nieto FJ, Gonzalez MC, Ma´rtire DO. J.
Photochem. A: Chem. 1998; 116: 21; (b) Cencione SS, Gonzalez
MC, Ma´rtire DO. J. Chem. Soc., Faraday Trans. 1998; 94: 2933;
(c) Rosso JA, Allegretti PE, Ma´rtire DO, Gonzalez MC. J. Chem.
Soc., Perkin Trans. 2 1999; 205; (d) Steenken S. Top Curr. Chem.
1995; 177: 141.
16. (a) Draganic ZD, Negron-Mendoza A, Sehested K, Vulosevic SI,
Navarro-Gonzalez R, Albarran-Sanchez MG, Draganic IG. Radiat.
Phys. Chem. 1991; 38: 317; (b) Lilie J, Hanrahan RJ, Henglein A.
Radiat. Phys. Chem., 1978; 11: 225.
17. Neta P, Simic M, Hayon E. J. Phys. Chem. 1969; 73: 4207.
18. Gonzalez MC, Braun AM. Res. Chem. Intermed. 1995; 21: 837.
19. Chen R, Avotinsh Y, Freeman GR. Can. J. Chem. 1994; 72: 1083.
20. Wine PH, Tang Y, Thorn RP, Wells JR, Davis DD. J. Geophys.
Res. D 1989; 94: 1085.
36. Kochi JK. In Free Radicals, vol.1, Kochi JK, (ed). John Wiley &
Sons: New York, 1973; 629.
37. Karpel Vel Leitner N, Gombert B, Ben Abdessalem R, Dore´ M.
Chemosphere 1996; 32: 893.
38. Eberhardt M. J. Phys. Chem. 1974; 78: 1795.
39. Scaiano JC, Stewart LC. J. Am. Chem. Soc. 1983; 105: 3069.
40. Chutny B. Nature (London) 1967; 213: 593.
41. (a) Shida T. Phys. Sci. Data 1988; 34: 47; (b) Shida T, Iwata S. J.
Am. Chem. Soc. 1973; 95: 3473.
21. McElroy WJ, Waygood SJ. J. Chem. Soc., Faraday Trans. 1990;
86: 2557.
22. Buxton GV, Salmon GA, Wood ND. In Physico-chemical
Behaviour of Atmospheric Pollutants, Restelli G, Angeletti G
(eds). Kluwer: Dordrecht, 1990; 245.
42. Gordon S, Schmidt KH, Hart EJ. J. Phys. Chem. 1977; 81: 104.
43. Mantaka A, Marketos DG, Stein G. J. Phys. Chem. 1971; 75: 3886.
44. Anbar M, Hart EJ. J. Am. Chem. Soc. 1964; 86: 5633.
45. Jakob L, Hashem TM, Bu¨rki S, Guindy N, Braun AM. J.
Photochem. Photobiol. A: Chem. 1993; 75: 97.
23. Heit G, Neuner A, Saugy P, Braun A. J. Phys. Chem. A 1998; 102,
5551.
24. Lide DR (ed). CRC Handbook of Chemistry and Physics (76th
edn). CRC Press: Boca Raton, FL, 1996.
46. Merga G, Schuchmann H-P, Rao BSM, Von Sonntag C. J. Chem.
Soc., Perkin Trans. 2 1996; 1097.
47. Ito T, Hatta H, Nishimoto S. Int. J. Radiat. Biol., 2000; 76: 683.
48. Wardman P. Int. J. Radiat. Biol. 1975; 28: 585.
Copyright  2001 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 300–309