A common carbanion intermediate in the recombination and proton-catalysed disproportionation of the carboxyl radical anion, CO2.-, in aqueous solution

Article abstract of DOI:10.1002/1521-3765(20010216)7:4<796::AID-CHEM796>3.0.CO;2-J

The carboxyl radical anion, CO₂^.-, was produced by the reactions of OH radicals with either CO or formic acid in aqueous solution. The pK_a(^.CO₂H) was determined by pulse radiolysis with conductometric detection at pH≈2.3. The bimolecular decay rate constant of CO₂^.- (2k≈1.4 × 10⁹ dm³mol^-1s^-1) was found to be independent of pH in the range 3-8 at constant ionic strength. The yields of the products of the bimolecular decay of the carboxyl radicals, CO₂ and the oxalate anion were found to depend strongly on the pH of the solution with an inflection point at pH 3.8. This pH dependence is explained by assuming a head-to-tail recombination of the CO₂^.- radicals followed by either rearrangement to oxalate or a protonation of the adduct, which subsequently leads to the formation of CO₂ and formate. The recombination of CO₂^.- to give oxalate directly is estimated to have a contribution of <25%. Wiley-VCH Verlag GmbH, 2001.

Full text of DOI:10.1002/1521-3765(20010216)7:4<796::AID-CHEM796>3.0.CO;2-J

FULL PAPER
A Common Carbanion Intermediate in the Recombination and
Proton-Catalysed Disproportionation of the Carboxyl Radical Anion,
.
CO2 , in Aqueous Solution
Roman Flyunt,[a, b] Man Nien Schuchmann,[a] and Clemens von Sonntag*[a]
Abstract: The carboxyl radical anion,
CO2 , was produced by the reactions of
constant ionic strength. The yields of the
products of the bimolecular decay of the
carboxyl radicals, CO2 and the oxalate
anion were found to depend strongly on
the pH of the solution with an inflection
point at pH 3.8. This pH dependence is
explained by assuming a head-to-tail
recombination of the CO2 radicals
.
.
OH radicals with either CO or formic
acid in aqueous solution. The
followed by either rearrangement to
oxalate or a protonation of the adduct,
which subsequently leads to the forma-
tion of CO2 and formate. The recombi-
.
pKa( CO2H) was determined by pulse
radiolysis with conductometric detec-
tion at pH ꢀ 2.3. The bimolecular decay
Keywords: carbanions
´ dimeriza-
.
.
rate constant of CO2
(2k ꢀ 1.4 Â
nation of CO2 to give oxalate directly
tion ´ formic acid ´ pulse radiolysis
´ radicals
109 dm3 mol 1 s ) was found to be inde-
is estimated to have a contribution of
<25%.
1
pendent of pH in the range 3 ± 8 at
Introduction
Experimental Section
.
Aqueous solutions of the substrates (concentrations typically 1 Â
The carboxyl radical, CO2 , is one of the most widely used
3
10 3 moldm for sodium formate) were prepared in Milli-Q-filtered
species in free-radical studies because of its excellent reducing
(Millipore) water and saturated with N2O. The pH of the solutions was
adjusted to the desired value with NaOH or HClO4. g-Radiolyses were
carried out in a panoramic 60Co source at a dose rate of 0.16 Gys 1 to total
doses of up to 150 Gy (i.e., the conversion of the substrate remained below
10%; the dose of 1 Gy equals 1 J of radiation energy deposited in a
.
properties [E(CO2/CO2 ) ˆ 1.9 V][1] and its facile genera-
tion in aqueous solutions by ionising radiation.[2] For example,
this radical may be produced by the reactions of hydroxyl
radicals and H atoms with formic acid/formate ion, of
hydroxyl radicals with carbon monoxide or of solvated
electrons with carbon dioxide.[3, 4] In neutral to basic solutions,
oxalate was found to be the main product of the bimolecular
3
substrate mass of 1 kg, here 1 dm of solution).
Oxalic acid was analysed by ion chromatography (Dionex DX-2010i) on an
AS14 column (4 Â 250 mm with a 4 Â 50 mm AG14 pre-column, ASRS-
ULTRA suppressor). The eluent was an aqueous solution of Na2CO3 and
.
[4, 5]
3
NaHCO3 (1.8 Â 10 3 moldm and 1.7 Â 10 3 moldm 3, respectively) at a
decay of CO2
.
However, in acidic solutions CO2 is
flow rate of 1 mLmin 1. The retention time was 13.6 min.
formed instead.[6, 7] In the present work we have reinvesti-
.
CO2 was scrubbed from irradiated samples with argon as the carrier gas and
measured by GC on a 3.2 m Porapak-9 capillary column.
gated the reactions of CO2 by means of pulse radiolysis and
product analysis with the aim of finding an explanation for the
surprising pH dependence of product formation upon its
bimolecular decay.
Pulse radiolysis was carried out with a 2.8 MeV Van-de-Graaff accelerator
delivering electron pulses of 0.4 ms duration. Intermediates were monitored
by optical detection. The pulse radiolysis set-up has been described
3
previously.[8] For dosimetry, a N2O-saturated 10 2 moldm thiocyanate
.
solution was used for optical detection, with G  e((SCN)2 ) ˆ 4.8 Â
1
10 4 m2 J at 480 nm[9] (although more recently a 10% higher value for
G Â e has been redetermined,[10] for consistency with our previous
[a] Prof. Dr. C. von Sonntag, Dr. R. Flyunt, Dr. M. N. Schuchmann
Max-Planck-Institut für Strahlenchemie
Stiftstrasse 34 ± 36, P.O. Box 101365
measurements the lower value was used).
45470 Mülheim an der Ruhr (Germany)
Fax : (‡49)208-306-3951
Results and Discussion
E-mail: vonsonntag@mpi-muelheim.mpg.de
[b] Dr. R. Flyunt
The radical-generating system: Hydroxyl radicals are gener-
ated during the radiolysis of water [reaction (1)]. The
Permanent address:
Institute of Physico-Chemistry
National Academy of Science of the Ukraine
Naukova Street 3a, 79053 Lꢀviv (Ukraine)
ionising
.
.
‡
H2O
eaq, OH, H , H , H2O2, H2
(1)
!
radiation
796
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0704-0796 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 4

796±799
radiation-chemical yields (G values) of the primary radicals
In the light of the present pKa value of ꢀ2.3, its previous
determinations deserve a comment. The data reported by
Buxton and Sellers[12] have been remeasured, and the
inflection point near pH 1.4 was found to have been an
artefact (G. V. Buxton, private communication). Jeevarajan
et al.[13] stated that their data could be also consistent with a
higher pKa value if their measured changes of the 13C
hyperfine constants were the result of the protonation of
.
.
7
7
are G( OH) ꢀ G(eaq† ˆ 2.8  10 , G(H ) ˆ 0.6  10
and
1
G(H2O2) ꢀ 0.7 Â 10 7 molJ . N2O is used to convert the
.
solvated electron into OH [reaction (2)]. In acidic solutions,
.
eaq ‡ N2O ‡ H2O
!
OH ‡ OH ‡ N2
(2)
protons also compete with N2O for the solvated electron
[reaction (3)]. The OH radical and the H atom will react with
.
.
.
CO2H, and the 13C hyperfine constants of CO2 and CO2H
.
‡
eaq ‡ H
!
H
(3)
were identical. Our pKa value of 2.3 is supported by the
(empirical) Pauling rule of pKa values of oxoacids of the
general form XOp(OH)q.[14] This rule indicates that the pKa
value of an oxoacid is largely determined by the number of
oxygen atoms at the central atom and that the latter has only
moderate influence (pKa ˆ 8 5p). Thus all mono-oxo acids
should have pKa values of ꢀ3.0. Within a certain range, this is
indeed the case: CO(OH)2 (3.6), HC(O)OH (3.75), OClOH
(2.0), ONOH (3.15) OP(OH)3 (2.2) OAs(OH)3 (2.25),
OSe(OH)2 (2.46) and OTe(OH)2 (2.48). As to the discrep-
ancies of our data with those reported by Fojtik et al.[11], there
is no obvious explanation, since both are based on the same
system and obtained with the same technique.
formic acetic/formate anion [pKa(HCO2H) ˆ 3.75] to yield the
.
.
carboxyl radical or its anion, CO2H/CO2 [reaction (4)].
.
.
.
.
OH (H ) ‡ HCO2H/HCO2 ! H2O (H2) ‡ CO2H/CO2
(4)
.
Alternatively, the reaction of OH with carbon monoxide can
.
.
be used to generate CO2H/CO2 [reaction (5)].[11] At a
formate concentration of 1 Â 10 3 moldm , the radiation-
3
.
chemical yield, G value, is G(CO2 ) ꢀ 6.2 Â 10 7 molJ , and
1
a material balance can be based on this value.
.
.
CO ‡ OH
!
CO2H
(5)
(6)
We have redetermined the rate of the bimolecular decay of
.
.
‡
CO2H > CO2 ‡ H
.
CO2 by monitoring the absorption decay at l ˆ 250 nm at a
3
constant ionic strength of 0.5 moldm , and found that this
.
The CO2H radical is an acid [equilibrium reaction (6)].
There is considerable uncertainty as to its pKa value. Fojtik
et al.[11] have produced this intermediate by the reaction of
rate constant practically does not vary over the pH range 3 ± 9
1
(Figure 2, 2k ꢀ 1.4 Â 109 dm3 mol 1 s ). This agrees fairly well
with the reported values of 1.5 Â 109,[4] 9.0 Â 108,[11] 1.0 Â 109[3]
.
OH with CO [reaction (5)], and from pulse radiolysis with
1 [12]
and 7.6 Â 108 dm3 mol 1 s .
The latter two values were
conductometric detection they concluded that the pKa value is
ꢀ3.9. By following the optical absorption change of the
obtained at zero ionic strength. These authors also found no
change in the bimolecular decay rate constant of CO2 within
.
.
.
CO2 / CO2H radical at l ˆ 250 nm in a pulsed irradiated
solution of formic acid/formate ion as a function of pH, Buxton
and Sellers[12] obtained a pKa value of 1.4. An even lower value
of 0.2 has been determined by Jeevarajan et al.[13] by measur-
ing the change in the 13C hyperfine constant as a function of pH.
We have repeated the experiments by Fojtik et al.[11] and
arrived at a value of pKa ꢀ 2.3 (Figure 1), which is significantly
lower than that given by Fojtik et al., but still considerably
higher than those of the other studies.
a pH range of 3 ± 7.
It is generally believed that CO2 only reacts by a
recombination at carbon to give oxalate [reaction (7)].[4, 5, 7]
.
.
2CO2
! (CO2
)
2
(7)
Indeed, when product studies were carried out in neutral
and alkaline aqueous solutions, oxalate was practically the
Figure 2. Pulse radiolysis of N2O-saturated aqueous solutions of formic
3
3
acid (10 3 moldm
) in the presence of 0.5 moldm NaClO4. Rate
Figure 1. Pulse radiolysis of N2O/CO (4:1 v/v)-saturated aqueous solu-
constant of the bimolecular decay as a function of pH. Inset A: Decay of
~
tions. Conductivity changes as a function of pH ( : original experimental
the absorption at l ˆ 270 nm as a function of time, pH 4.6, 26 Gy per pulse.
.
*
data; : experimental data corrected for the loss of OH radicals by
Inset B: The inverse of the first half-life of absorption decay of CO2 at
.
&
competition of protons with N2O for the solvated electrons; ÐÐ: pKa curve
based on pKa ˆ 2.3).
l ˆ 270 nm as a function of the CO2 concentration at pH 4 ( ) and pH 5
~
).
(
Chem. Eur. J. 2001, 7, No. 4
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0704-0797 $ 17.50+.50/0
797

C. von Sonntag et al.
FULL PAPER
Figure 3. g-Radiolysis of formate (1 Â 10 3 moldm 3) in argon-saturated
~
*
solution. G values of oxalate ( ) and CO2 ( ) as a function of pH. The sum
Scheme 1.
^
(
) of these two products is also plotted as function of pH. Some oxalate
measurements were also carried out in N2O-saturated solutions.
.
aspect of this mechanism is that the CO2 radicals react
1
only
product
with
G(oxalate) ˆ 3.5  10 7 molJ
mainly (>90%) by a head-to-tail recombination, and that the
rearrangement reaction (12) competes with the H -assisted
disproportionation reactions (13) ± (16) (Scheme 1).
The CO2 radical can be represented by two mesomeric
forms, one with the spin at carbon and one with the spin at
oxygen. We now suggest that the CO2 radicals react with one
.
.
[0.5G( OH‡H )].[4] We observed that G(oxalate) ˆ 2.7 Â
‡
10 7 molJ at pH 4.6, and the oxalate yield drops as the pH
1
is decreased (Figure 3). At pH 3 its yield is negligible (Fig-
ure 3). In this system, one has to recall that when N2O is used
to convert eaq into OH [reaction (2)] a chain reaction takes
place to yield CO2 as a chain product [reaction (8)].[15] Hence,
for most of the experiments, the system was saturated with
.
.
.
another by
a head-to-tail recombination [reaction (11),
Scheme 1]. The resulting carbanion may rearrange into
oxalate [reaction (12)]. When protonated [reactions (13)
and (14), followed by reaction (15)], however, the mixed
anhydride decomposes into carbon dioxide and formic acid
[reaction (16)]. Although the latter product cannot be deter-
mined under our experimental conditions (formic acid was
.
argon, and eaq was converted into H according to reaction (3)
.
.
CO2 ‡ N2O ‡ H2O ! CO2 ‡ N2 ‡ OH ‡ OH
(8)
We also found that the yield of CO2, in contrst to that of
oxalic acid, increased with increasing proton concentration
(Figure 3), in excellent agreement with a previous report.[7] As
shown in Figure 3, these two yield curves cross at the
inflection point of pH 3.8.
.
.
used as the source for CO2 ), it is inferred that two CO2
radicals give rise to either one molecule of oxalate or one
molecule of CO2 plus one molecule of formic acid. The sum of
G(oxalate)‡G(CO2) should be constant over the whole pH
.
Disproportionation between two HCOO radicals [reac-
.
.
range and equal to 3.1  10 7 molJ [0.5G( OH‡H )]. This
1
tion (9))] is generally assumed to be responsible for the
formation of CO2 in acidic solutions of formic acid (for a
review see ref. [16]). Similarly, the cross-termination between
material balance is indeed obtained (Figure 3).
Protonation at a heteroatom is fast (between 5 Â 109 and
5 Â 1010 dm3 mol 1 s ); however, protonation at carbon is
1
.
.
CO2H and CO2 could give the same products [reaction (10)]
typically slower.[17] Thus, reaction (13) should be faster than
reaction (14) (Scheme 1), but once protonated at oxygen
[reaction (13)] subsequent intramolecular protonation at
carbon [reaction (15)], possibly assisted by a molecule of
water, may speed-up the protonation at carbon. If the rate
.
.
.
CO2H ‡ CO2H ! CO2 ‡ HCO2H
(9)
.
CO2H ‡ CO2
! CO2 ‡ HCO2
(10)
However, our results show that the above simple mecha-
constant for the relevant protonation reaction is taken as
1
nism [reactions (7), (9) and (10)] as generally proposed, is
inadequate to explain the observations in the pH range of
Figure 3. In order for the displacement of the pH for the cross-
over in yields of CO2 and oxalate from the radical pKa of 2.3
(cf. Figure 1) to 3.8 (Figure 3) to be explained, the rate
constants of reactions (9) and (10) would have to be about
18 times higher than 2k7. The bimolecular rate constant of
1 Â 1010 dm3 mol 1 s
and the inflection point at pH 3.8
‡
3
([H ] ˆ 1.6  10 4 moldm ), then the competing reac-
1
tion (12) must have a rate constant of ꢀ1.6 Â 106 s . A slower
rate of protonation would require a longer lifetime of the
postulated carbanion [reaction (11)].
This mechanistic scheme depicted by reactions (11) ± (16)
(Scheme 1) requires that the oxalate yield vanishes in a
sigmoidal fashion at low pH. The solid lines in Figure 3 were
calculated on the basis of the competition between reac-
tion (12) versus reaction (13)/(14) with an inflection point at
.
1
CO2 , 2k7 ꢀ 1.4 Â 109 dm3 mol 1 s (Figure 2, the slight var-
iation over the pH range shown is within experimental error),
is already close to diffusion controlled. Moreover, a value of
.
2k9 ˆ 1.7  109 dm3 mol 1 s determined at pH 0 has been
pH 3.8. Taking our pKa value of 2.3 (cf. Figure 1), the CO2H
1
.
reported.[12]
radical concentration at pH ꢀ 3.3 is ꢀ10% of the total CO2
/
.
To account for our results we propose the mechanism
represented by reactions (11) ± (16) in Scheme 1. The major
CO2H radical concentration. In this pH range, the potential
contribution of CO2H to product formation [cf. reactions (9)
.
798
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0704-0798 $ 17.50+.50/0
Chem. Eur. J. 2001, 7, No. 4

Carbanion Intermediates
796±799
[1] P. Wardman, J. Phys. Chem. Ref. Data 1989, 18, 1637 ± 1755.
[2] C. von Sonntag, The Chemical Basis of Radiation Biology, Taylor and
Francis, London, 1987.
and (10)] will be of the same magnitude. This would explain
the fact that the experimental oxalate data are somewhat
lower than the calculated curve at pH < 3.5.
One expects that reaction (7) occurs in competition with
reaction (11) (Scheme 1). However, if there were a substantial
contribution from reaction (7) (>25%) then the oxalate yield
at low pH should be noticeably higher than the calculated
curve.
[3] J. P. Keene, Y. Raef, A. J. Swallow, in Pulse Radiolysis (Eds.: M. Ebert,
J. P. Keene, A. J. Swallow, H. J. Baxendale), Academic Press, New
York, 1965, pp. 99 ± 106.
[4] P. Neta, M. Simic, E. Hayon, J. Phys. Chem. 1969, 73, 4207 ± 4213.
[5] T. J. Hardwick, Radiat. Res. 1960, 12, 5 ± 12.
[6] H. Fricke, E. J. Hart, H. P. Smith, J. Chem Phys. 1938, 6, 229 ± 240.
[7] E. J. Hart, J. K. Thomas, S. Gordon, Radiat. Res. Suppl. 1964, 4, 74 ± 88.
[8] C. von Sonntag, H.-P. Schuchmann, Methods Enzymol. 1994, 233, 3 ±
20.
[9] R. H. Schuler, L. K. Patterson, E. Janata, J. Phys. Chem. 1980, 84,
2088 ± 2089.
[10] G. V. Buxton, C. R. Stuart, J. Chem. Soc. Faraday Trans. 1995, 91,
279 ± 281.
[11] A. Fojtik, G. Czapski, A. Henglein, J. Phys. Chem. 1970, 74, 3204 ±
3208.
[12] G. V. Buxton, R. M. Sellers, J. Chem. Soc. Faraday Trans. 1 1973, 69,
555 ± 559.
[13] A. S. Jeevarajan, I. Carmichael, R. W. Fessenden, J. Phys. Chem. 1990,
94, 1372 ± 1376.
[14] D. F. Shriver, P. W. Atkins, C. H. Langford, Inorganic Chemistry, 2nd
ed., Oxford University Press, Oxford, 1994, pp. 193 ± 196.
[15] M. I. Al-Sheikhly, H.-P. Schuchmann, C. von Sonntag, Int. J. Radiat.
Biol. 1985, 47, 457 ± 462.
In water, unusual disproportionation reactions are not
uncommon. For example at pH 12, the hydroxymethyl radical,
which at these conditions is fully deprotonated (pKa ˆ 10.7)
.
and is present as CH2O , only disproportionates into form-
aldehyde and methanol, despite the fact that no b-hydrogen is
available for a straightforward disproportionation reaction.[18]
Furthermore, the radicals derived from acetic acid
.
.
ˆ
( CH2CO2H $ CH2 C(O )OH), which show structural sim-
.
ilarities to the CO2 radicals (spin density at oxygen),
undergo disproportionation, despite the absence of b-hydro-
gens, in addition to the expected recombination at carbon
(formation of succinic acid). It was concluded that a labile
head-to-tail dimer must be the intermediate on the route to
disproportionation.[19] Similarly, head-to-tail dimers play a
major role in the free-radical chemistry of malonic acid.[19]
[16] A. O. Allen, in The Radiation Chemistry of Water and Aqueous
Solutions, Van Nostrand, Princeton, New Jersey, 1961, pp. 117 ± 140.
[17] M. Eigen, W. Kruse, G. Maass, L. De Maeyer, Progr. Reaction Kinetics
1964, 2, 285 ± 318.
[18] W.-F. Wang, M. N. Schuchmann, V. Bachler, H.-P. Schuchmann, C.
von Sonntag, J. Phys. Chem. 1996, 100, 15843 ± 15847.
[19] W.-F. Wang, M. N. Schuchmann, H.-P. Schuchmann, C. von Sonntag,
Chem. Eur. J. 2001, 7, 791 ± 795.
Acknowledgements
Â
We would like to thank Drs. G. Merenyi, G. V. Buxton and I. Carmichael
for valuable discussions.
Received: July 14, 2000 [F2599]
Chem. Eur. J. 2001, 7, No. 4
ꢁ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
0947-6539/01/0704-0799 $ 17.50+.50/0
799