A curve-crossing model for oxidative decarboxylation. Kinetics of anilino carboxylate fragmentations

Article abstract of DOI:10.1021/jp040536x

The kinetics of oxidative decarboxylation of a series of anilino carboxylates have been measured. These compounds find use in the recently described two-electron sensitization process for silver halide photography, and control of their fragmentation kinet

Full text of DOI:10.1021/jp040536x

J. Phys. Chem. A 2004, 108, 10949-10956
10949
A Curve-Crossing Model for Oxidative Decarboxylation. Kinetics of Anilino Carboxylate
Fragmentations
Ian R. Gould*
Department of Chemistry and Biochemistry, Arizona State UniVersity, Tempe, Arizona 85287-1604
Jerome R. Lenhard and Samir Farid
Research Laboratories, Eastman Kodak Company, Rochester, New York 14650-2109
ReceiVed: August 6, 2004
The kinetics of oxidative decarboxylation of a series of anilino carboxylates have been measured. These
compounds find use in the recently described two-electron sensitization process for silver halide photography,
and control of their fragmentation kinetics is essential to be useful in this application. The measured rate
constants vary over 5 orders of magnitude. The decarboxylation rate constants decrease substantially with
increasing solvent polarity, an effect that results in misleading Arrhenius analysis of temperature effect
measurements. The rate constants also depend on the product radical stability and the oxidation potential of
the aniline precursor. A valence bond curve-crossing model is proposed that provides a combined picture of
this and the related decarboxylation of acyloxy radicals. The model is able to explain the substitution and
solvent polarity effects. Consistent with this model, manipulation of solvation via changes in substituents
(hydrophilic vs hydrophobic) around the reaction site is shown to strongly influence the decarboxylation rate
constants. In this way, the decarboxylation rate constants can be altered by nearly 3 orders of magnitude at
constant solvent polarity and precursor oxidation potential.
Introduction
donor, X-Y•+.6 Because of their wide importance (see above),
a number of studies of the kinetics of oxidative decarboxylation
and the related acyloxy radical decarboxylation have appeared,
from which a number of broad generalizations can be made.10
The rate of acyloxy radical decarboxylation depends on the
stability of the radical formed in the fragmentation process and
can occur with very high rate constants (>1010 s-1).10c,j The
rate of radical cation fragmentation tends to be slower and varies
considerably depending upon the structure of the aryl group
and radical stability. The relationship between the acyloxy
radical and radical cation decarboxylation processes has been
little discussed, except by Saveant in his work delineating the
mechanism of the Kolbe reaction,11 and in some cases the
difficulty in assigning the structure of the intermediate to the
radical cation or acyloxy radical has been pointed-out.12
Decarboxylation reactions of organic radical cations have
been widely studied in mechanistic and synthetic chemistry1
and have been implicated in several important biological
processes (reaction A, Scheme 1).2 A closely related process is
the electrochemical Kolbe reaction, where oxidation of (usually)
an aliphatic carboxylate leads to an acyloxy radical that
decarboxylates to form an alkyl radical (reaction B, Scheme
1).3 Also related is the decarboxylation reaction of acyloxy
radicals (e.g., reaction C, Scheme 1), which plays a central role
in the chemistry of peroxides.4
Oxidative decarboxylation also represents an example of a
photoinduced electron transfer/fragmentation reaction, which
have proved to be the most technologically important processes
in electron-transfer photochemistry.5 In this regard, we recently
reported a new technological application of electron transfer/
fragmentation in silver halide photography, as part of a novel
sensitization process called two-electron sensitization (TES).6,7
The central idea behind TES is that the oxidized form of the
photographic sensitizing dye (formed upon light exposure and
electron transfer to silver halide8) further oxidizes a fragmentable
electron donor (X-Y) to form X-Y•+. Cleavage of X-Y•+
yields a radical X•, which is designed to be a powerful reducing
agent capable of injecting another electron into the conduction
band of silver halide. In this way, two electrons are transferred
per absorbed photon, thus doubling the photographic efficiency.
Decarboxylation of X-Y•+ proved to be one of the most useful
of the fragmentation reactions in the TES process, providing
the motivation for the present work.6,9
Despite this prior work, some confusion remains regarding
the quantification of the factors that controlled the rate constants
for fragmentation. For example, some of the most detailed
previous work on the kinetics of radical cation decarboxylation
is reported in ref 10n and p. In these papers, however, the radical
cation fragmentation rate constants were reported to be both
strongly10n and also weakly10p dependent upon the oxidation
potential of the aryl group. Although decarboxylation of acyloxy
radicals is clearly faster when more stable radicals are formed,
it was reported that this was not generally the case for radical
cation decarboxylation.10p Solvent effects had been studied in
very few cases,10e and virtually no reports of the magnitudes of
the reaction barriers are available. Thus, we initiated a quantita-
tive study of the factors that control the rate constants for
oxidative decarboxylation and sought a qualitative theoretical
description that would accommodate all of the important
observations and factors that control the rates. The reactions
An important issue in the TES process is control over the
rate constant for fragmentation (decarboxylation) of the oxidized
* Corresponding author. E-mail: igould@asu.edu.
10.1021/jp040536x CCC: $27.50 © 2004 American Chemical Society
Published on Web 11/20/2004

10950 J. Phys. Chem. A, Vol. 108, No. 50, 2004
Gould et al.
SCHEME 1
SCHEME 2
that we studied are best described as radical cation decarboxy-
lations, but the simple theoretical description results in a
combined picture of the three types of reaction shown in Scheme
1. Our goal was a comprehensive description of the factors
controlling the rate of reaction, so that suitable TES donors could
be designed.
Results and Discussion
To be useful in a TES application, the radicals formed upon
decarboxylation must be strongly reducing. We have shown
previously that R-amino radicals are appropriately reducing,
which means that most useful TES donors are amines.6 We have
found bis-carboxylates of aniline derivatives of the general
formula 1 (Scheme 2) to be particularly useful because of their
high thermal stability and the ease of synthesis of several
derivatives.
The effect of reaction exothermicity on the decarboxylation
rate constant, k-CO , was explored by varying the ring substit-
Figure 1. Transient absorption spectra of the radical cations of
anilinodicarboxylates 1c and 1d, at room temperature in (9) 4:1
acetonitrile/water and (O) 100% acetonitrile.
2
uents R1 and R2. Variations in these substituents result in
corresponding variations in the oxidation potential of the aniline
derivative, and in turn this changes the reaction exothermicity
(see ref 13 and further below). The effect of the stability of the
product radical was examined by changing the side chain
substituent R′. In addition to substituent effects, we also
examined the effect of solvent and of temperature on the
decarboxylation rate constants of the radical cations.
Figure 1. The radical cation absorptions are somewhat red-
shifted compared to the aniline radical cations from which they
are derived.14 As illustrated in the figure, the spectra also exhibit
a dependence on the water content of the medium. The
carboxylates are hydrogen-bonded in water, and in its absence,
an intramolecular interaction with the radical cation anilino
moiety may be responsible for the observed changes in the
absorption spectra. Only a small amount of water is required to
effect the observed spectral changes, presumably due to efficient
hydration of the carboxylates, and the spectra in a 19:1
acetonitrile-water mixture are close to those in a 4:1 mixture.15
Rate constants for decarboxylation of the different derivatives
of 1•+ were obtained in conventional nanosecond transient
absorption experiments, by monitoring the decay in absorbance
of the radical cations as a function of time, as described in detail
previously for a closely related system.9b In all cases the
absorption decays were well-described by exponential kinetics
Determination of Rate Constants
The radical cations of the aniline derivatives, 1•+, were
generated and characterized as described previously, from the
reaction of excited 9,10-dicyanoanthracene (DCA) with a
cosensitizer, biphenyl (BP), which leads to formation of free
DCA•- and BP•+ in a high yield.6a,9b Secondary electron transfer
from 1 to BP•+ yields 1•+. Experiments were performed in nitrile
solvents and in acetonitrile-water mixtures, and the radical
cations were characterized in these solvents as described
previously.6a,9b Representative absorption spectra are shown in

Kinetics of Anilino Carboxylate Fragmentations
J. Phys. Chem. A, Vol. 108, No. 50, 2004 10951
TABLE 1: Rate Constants (at Room Temperature),
Oxidation Potentials, and Apparent Arrhenius Activation
Parameters for Decarboxylation of Aniline Radical Cations
in 4:1 Acetonitrile/Water
Figure 2. Plot of the logarithms of the rate constant for decarboxylation
of the radical cations of anilinodicarboxylates as a function of the
oxidation potential in 4:1 acetonitrile/water at room temperature for
(0) aniline diacetates, 1a-1d, and (b) aniline acetate-propionates, 1e-
1h.
thermodynamically meaningful oxidation potentials via the well-
known eq 1a.17b Here, a is equal to νF/RT, where F is the
Faraday constant, ν is the scan rate in V/s, and R and T have
their usual meanings. At 22 °C and a scan rate of 0.1 V/s, eq
1a reduces to eq 1b.
Eox ) Ep + (RT/2F){ln(k-CO /a) - 1.56}
(1a)
(1b)
a Peak potentials measured at a carbon electrode at a scan rate of
0.1 V/s in 4:1 acetonitrile/water containing 0.1 M LiClO4. b Oxidation
potentials derived from the peak potentials and the decarboxylation
rate constants according to eqs 1. c The decarboxylation rate constant
of the radical cation of compound 1e is too high to measure accurately.
2
Eox ) Ep + 0.0127 {ln(k-CO /3.93) - 1.56}
2
Dependence of Decarboxylation Rate Constants on
Reaction Energetics
over more than two lifetimes. Competing second-order recom-
bination with the DCA•- was ensured to be negligible by using
low pulsed laser energies, resulting in low concentrations of
radical ions in solution. The reactions were assigned to
decarboxylation based on the facts that the products were
reducing R-amino radicals.6 Deprotonation at the R-methylene
is a possible competing reaction; however, we and others have
shown elsewhere that this process occurs with pseudo-first-order
rate constants that are orders of magnitude lower than those
observed here.16 Most of the experiments were performed in
4:1 acetonitrile-water. This solvent mixture was initially
selected both to ensure solublization of all of the amine
carboxylates and to provide an appropriate model environment
for the mainly aqueous photographic environment.6 Monitoring
wavelengths around 530-630 nm were generally found to be
useful for measuring the decay kinetics of the radical cations,
Figure 1.
Summarized in Table 1 are the rate constants for decarboxyl-
ation measured at 22 °C in 4:1 acetonitrile-water. The results
of temperature-dependent experiments included in Table 1 in
the form of apparent activation energies and the logarithms of
the preexponentials obtained from least-squares linear fits to
Arrhenius plots.
Also given in Table 1 are the oxidation potentials of the
donors, Eox. The method by which these were obtained is
described in detail elsewhere.17a Cyclic voltammetry of com-
pounds with reactive radical cation such as these gives irrevers-
ible oxidation waves, from which only peak potentials were
determined, Ep. Combination of such peak potentials with the
measured decarboxylation rate constants, however, can provide
As mentioned above, the dependence of the decarboxylation
rate constants on the reaction energetics was examined by
varying (a) the oxidation potential of the aniline derivatives, 1,
(b) the stability of the resultant radicals, 1r, and (c) the medium
polarity.
As illustrated in Figure 2, the rate constants are found to
depend strongly on the oxidation potential for both the radical
cations of the diacetate derivatives (R′ ) H) and the acetate-
propionate derivatives (R′ ) Me). In both cases, the rate
constants decrease by roughly 4 orders of magnitude as the
oxidation potential decreases by ∼0.32 V, which compares
reasonably well with the data of Candeias et al..10n In addition,
all rate constants of the derivatives with R′ ) Me are displaced
by ∼0.8 on the logarithmic scale above those with R′ ) H
(Figure 2), i.e., are higher by an average factor of ∼6.
Shown in Figure 3 are some typical Arrhenius plots. The
slopes and intercepts of these plots are given as “apparent”
activation energies and preexponential factors in Table 1.
Although the plots appear linear, the activation parameters are
unusual in that the preexponential factors are all significantly
greater than 1013 s-1, which is much larger than normally
observed for simple unimolecular reactions.18 In addition, there
appears to be a relationship between the magnitude of the
apparent activation energies and the preexponential factors; see
Table 1. The most reasonable explanation for the high preex-
ponential factors is that although the Arrhenius plots appear to
be linear oVer the experimental temperature range, they are
actually curved, with the slope increasing with decreasing
reciprocal temperature. The reactions with the higher temper-

10952 J. Phys. Chem. A, Vol. 108, No. 50, 2004
Gould et al.
Figure 3. Arrhenius plots for decarboxylation of the radical cations
of 1b, 1f, 1g, and 1h in 4:1 acetonitrile/water. The straight lines
represent best linear fits to the data; see Table 1 for the parameters.
Figure 5. Arrhenius plot for decarboxylation of the radical cation of
anilio diacetate 1d in (O) acetonitrile, (0) propionitrile, and (2)
acetonitrile/propionitrile solvent mixtures that give a dielectric constant
of 33.1 at all temperatures measured. The Arrhenius parameters in the
solvent mixture are log A ) 11.8 and Ea ) 5.9 kcal/mol.
The dielectric constant (ꢀ) of acetonitrile is 36 at room
temperature but decreases to 33.1 at 40 °C.20 Propionitrile,
although a lower polarity solvent (at room temperature, ꢀ )
27), also has a dielectric constant of 33.1 °C at -4 °C, the same
as acetonitrile at 40 °C.20 This allows an experiment to be
performed, in which the temperature is varied, but ꢀ is kept
constant. Rate constants were measured at different temperatures
in acetonitrile-propionitrile mixtures of varying proportions
such that the bulk dielectric constant remained at 33.1. The
results for the aniline derivative 1d are shown in Figure 5,
together with the corresponding data for experiments performed
in pure acetonitrile and in pure propionitrile. Similar to the
results in the 4:1 acetonitrile-water mixtures mentioned above,
the apparent preexponential factors for the reactions in pure
acetonitrile and in pure propionitrile are again in the range ∼1015
to 1016 s-1. However, the acetonitrile-propionitrile mixed-
solvent experiment in which ꢀ was maintained constant gave a
preexponential factor (log A ) 11.8) much closer to the expected
range for a simple unimolecular reaction.18 Also the apparent
activation barriers of 11.9 and 9.5 kcal/mol obtained from the
reactions in pure acetonitrile and in pure propionitrile are now
replaced by an activation barrier of only 5.9 kcal/mol in the
constant ꢀ experiment. The obvious conclusion, then, is that
the Arrhenius plots in the pure solvents and in the 4:1
acetonitrile-water mixtures are unusual as a result of the
extreme sensitivity of the present reactions to solvent polarity,
and when this is corrected for, reasonable Arrhenius behavior
is restored.
Figure 4. Plots of the logarithms of the rate constants for decarboxyl-
ation of the radical cations of anilino dicarboxylates (b) 1b, (9) 1c,
and (2) 1d, showing a dramatic decrease in rate constant with added
water.
ature dependencies extrapolate to larger preexponential factors
because the experimental data are on the “lower” part of the
curve.
These experiments were performed in a 4:1 acetonitrile-water
mixture. For the donors 1b, 1c, and 1d, the decarboxylation
rate constants were also determined as a function of the water
content in the solvent mixture, Figure 4. The dependence of
the rate constants on the amount of added water is dramatic.
For the radical cations of 1c and 1d, the decarboxylation rate
constants decrease by ∼3 orders of magnitude with the addition
of 10% water. For 1b the decrease is somewhat less severe but
still substantial. Part of the water effect is presumably due to
the carboxylate hydration effect that results in the spectral
changes observed in Figure 1. However, continued addition of
water continues to decrease the reaction rate constant after the
spectra have stopped changing, implying an additional effect
of medium polarity.
Decarboxylation Mechanism
For an aryl-substituted carboxylate, the product of oxidation
is expected to be an acyloxy radical, unless the oxidation
potential of the aryl moiety is less than that of carboxylate
(estimated to be ∼1.9 V vs SCE21), as in the present cases. The
products of cleavage of the radical cation of an anilino
carboxylate are a neutral R-amino radical and carbon dioxide,
Scheme 2. The reaction thus involves charge annihilation; i.e.,
at some stage in the course of the reaction a transfer of charge
must take place. This is one way in which the current reactions
differ from decarboxylation of an acyloxy radical (reaction C,
Scheme 1). Charge neutralization could occur by intramolecular
electron transfer from the carboxylate group (-CO2-) to the
anilino radical cation (An•+-) to yield the acyloxy radical
This effect might explain the unusual Arrhenius plots
discussed above. It is known that the bulk polarity of solvents
decreases with increasing temperature.19 Thus, as the temper-
ature is raised, an additional increase in the decarboxylation
rate constant could be due to the decrease in medium polarity,
resulting in overall nonlinear Arrhenius plots. It was important,
therefore, to test whether the dependence of the solvent dielectric
constant was a significant factor for the present reactions.

Kinetics of Anilino Carboxylate Fragmentations
J. Phys. Chem. A, Vol. 108, No. 50, 2004 10953
spin character). A state in which the electrons in the C-C bond
are decoupled in a triplet configuration is repulsive with respect
to bond stretching (r2, Figure 6).
Furthermore, one of these unpaired electrons can now undergo
a bonding interaction with the unpaired electron on oxygen, and
the ground state of the products is formed. The acyloxy radical
state r1 and r2 are charge-transfer excited states of the ground z
state. In addition to decoupling the electrons, a spin flip of the
odd electron is required to maintain doublet character in r2,
which thus represents a double-excitation of the r1 state.
Cleavage of either the radical cation or the acyloxy radical
states thus involves an electronic barrier due to an avoided
crossing with r2. The matrix element that splits the r1 and r2
curves, Vrr, is an energy transfer type25 and is likely to be large,
and thus the barrier to reaction in the acyloxy radical is
small.10c,10j,24 The matrix element that splits the z and (now
mixed) r curves, Vzr, is an electron-transfer type.26 Vzr may be
somewhat smaller than Vrr;27 however being intramolecular it
is still expected to be substantial.
Figure 6. Qualitative valence bond energy curves for the (z, red)
zwitterionic radical cation/carboxylate state, (r1, blue) the acyloxy
radical state, and (r2, black) C-C bond decoupled acyloxy radical state,
which are involved in the oxidative cleavage of anilino carboxylates.
The r1 and r2 states experience avoided crossing characterized by a
matrix element Vrr, and the z state experiences avoided crossing with
the mixed r state, Vzr.
This reaction model can explain all of the observed effects
on the decarboxylation rate constant of the anilino carboxylate
radical cations. The model is redrawn in Figure 7 to illustrate
the lowest energy adiabatic reaction surface formed as a result
of the avoided crossings for three important cases.
(reaction a, eq 2), followed by homolytic bond cleavage to form
the products (reaction b, eq 2). Under these conditions, the
The reaction activation energy, Ea, is given by the energy
required to reach the crossing point of the z and r curves minus
the matrix element Vzr (Figure 6). It is clear that the energy
required to form the acyloxy radical from the anilino radical
cation, ∆, is much larger than Ea, as observed. The decarboxyl-
ation rate constants decrease with increasing electron-donating
ability of the substituents on the benzene ring, i.e., with
decreasing oxidation potential of the anilino moiety (Table 1).
Such substituents stabilize (i.e., lower the energy of) the
zwitterionic state, which increases the energy difference ∆
(energy diagram B relative to A, Figure 7). This increase in
vertical displacement of the curves causes the energy required
to reach the curve crossing point to be larger for the more
stabilized donor radical cation, thus increasing the reaction
activation barrier and decreasing the rate constant. This is in
agreement with the data given in Table 1.
Stabilization of the zwitterionic state as a result of decreasing
oxidation potential also lowers the exothermicity of the reaction,
∆′. The diagram thus provides a connection between the kinetics
of the reaction and the predictions from thermodynamic cycles,
i.e., that decreasing oxidation potential is related to a decrease
in reaction exothermicity.13
Figure 7 also provides an explanation for the effect of the
substituent R′ (H vs Me, i.e., N-acetate vs N-propionate) on the
decarboxylation rate constant. The decarboxylation rate con-
stants are higher for the propionate derivatives 1e-1h, compared
to the corresponding acetate derivatives 1a-1d. The decar-
boxylation product of a propionate derivative is a secondary
radical, which is more stable than the corresponding primary
radical from the reaction of an acetate derivative (i.e., larger
exothermicity, ∆′).
intramolecular electron transfer would presumably become the
rate-determining step, since decarboxylation of acyloxy radicals
is so rapid. A minimum activation energy for this reaction is
given by the difference in oxidation potential between the aniline
and carboxylate. Even though the carboxylate potential is not
known accurately, it is clear that this cannot be the mechanism,
since, as mentioned above, the activation barrier for decarboxyl-
ation of the radical cation of 1d in nitrile solvents is ∼6 kcal/
mol, whereas the energy difference in oxidation potentials is
estimated to be >30 kcal/mol.21 Thus, an alternative mechanism
to the “stepwise” processes (reactions a and b, eq 2), i.e., a
“direct path” (reaction c, eq 2), is required to explain the data.
The dependence on oxidation potential is expected, based on
the simple thermodynamic cycle argument that a lower bond
dissociation energy results when the oxidation potential of the
precursor is decreased;13 however, we sought a more detailed
model in terms of the important orbital and state correlations
for the reactions that could qualitatively account for all of the
structural observations. We have found the valence bond (VB)
approach of Shaik and Pross to be useful.22-24 Shown in Figure
6 is an energy curve as a function of C-C bond length for the
radical cation (zwitterionic) state of an anilino carboxylate, z.
Stretching the C-C bond raises the energy of the spin-paired
electrons in the bond. In the VB formalism, the electrons are
paired even at large C-C distances, and a cation and a radical
anion of carbon dioxide is formed. Corresponding bond stretch-
ing in the acyloxy radical state, r1, again raises the energy of
the spin-paired electrons in the C-C bond while maintaining
their pairing, preventing bonding with the odd electron on the
carboxylate oxygen. An excited state of carbon dioxide is
formed, in a singlet configuration (to maintain overall doublet
Increasing the radical stability lowers the energy of all of
the bond breaking correlations in Figure 6, resulting in a lower
energy for curve crossing (energy diagram C relative to A,
Figure 7), a smaller activation energy, and a larger reaction rate
constant.
The mechanism also explains the solvent effects. More polar
solvents (for example, acetonitrile versus propionitrile) stablize
the zwitterionic state, decreasing the rate (energy diagram B
relative to A, Figure 7). This again will increase the energy

10954 J. Phys. Chem. A, Vol. 108, No. 50, 2004
Gould et al.
Figure 7. Adiabatic reaction energy curves for decarboxlyation of anilino carboxlyate radical cations. Curve A shows that the activation energy
for reaction via avoided crossing with the radical states, Ea, is smaller than the energy difference between the radical cation and acyloxy radical
states, ∆ (defined in Figure 6). Curve B shows the increase in activation energy and decrease in reaction exothermicity, ∆′, in more polar solvents
and for lower oxidation potential aniline moieties as a result of stabilization of the zwitterionic state. Curve C shows the decrease in activation
energy and increase in reaction exothermicity as a result of formation of a more stable radical product.
SCHEME 3: N-H (2-4) and N-butyl (5-7) Aniline
Derivatives; Decarboxylation Rate Constants (in s-1) of
Their Radical Cations in at 22 °C; Peak Potentials (in
Parentheses) from Irreversible Oxidations (V vs SCE at
Scan Rate of 0.1 V/s), Also in 4:1 Acetonitrile/Water; and
the Oxidation Potentials Calculated According to Eqs 1
(V vs SCE)
gap ∆, causing an increase in the activation energy and a
decrease in the rate constant (Figure 5). It is also well-known
that the oxidation potential of the carboxylate anion (and most
anions) is much higher in water than in acetonitrile due to
stronger solvation and H-bonding.21b,28 This provides an expla-
nation of the dramatic water effect on decarboxylation rate
constant (Figure 4), although the water must also be influencing
the structure of the reactive radical cations, as observed by the
changes in absorption spectra (Figure 1).
The model of Figure 6 has other interesting features. The
reactions of the radical cations occur by stretching the C-C
bond to the point of crossing with the radical state, at which
point mixing causes electron transfer and a decrease in energy
toward the ground state products. The reaction thus proceeds
via only one transition state, is concerted, and fits Saveant’s
definition of, and provides an intramolecular example of, a
dissociative electron-transfer reaction.11,23,29 Dissociative electron-
transfer reactions in turn are an example of bond-coupled
electron-transfer reactions that are becoming increasingly
recognized as common processes.30,31 A final point is that
electron-transfer reactions (in the adiabatic limit) are understood
as examples of radiationless transitions whose rates depend on
the value of an electronic coupling matrix element (squared),26
whereas this and other bond-coupled electron-transfer reactions
represent examples of chemical processes whose rates are
determined by the magnitude of an electron-transfer matrix
element.
Control of the Decarboxylation Rate Constant
The combined experimental results and the theoretical model
clearly illustrate that solvation effects (either via bulk polarity
or specific solvation) can have a very strong effect on the
decarboxylation rate constant. This suggested a method of
controlling the rate constant without changing the bulk medium,
which is clearly not possible in the TES application, and without
changing the oxidation potential of the aniline, by varying the
solvation at the reaction site.
In the acetate/propionate derivatives (1e-1h) discussed
above, decarboxylation occurs on the propionate group to yield
a secondary radical. The molecular solvation in the vicinity of
this reactive group was altered by substitution with other
hydrophobic and hydrophilic groups, while maintaining decar-
boxylation to give an equivalent secondary radical. In one set
of compounds, the acetate group on the nitrogen was replaced
by a simple hydrogen, as in compounds 2-4 of Scheme 3. In
another set, the acetate was replaced by a hydrophobic butyl
group, as in compounds 5-7 of Scheme 3. A third set of
compounds were obtained by replacement of the propionate on
nitrogen with the presumably more hydrophilic succinate group,
compounds 4 and 7. To determine whether any effects of the
fragmentation rate constant were genuinely due to changes in
solvation, adjustments for corresponding changes in oxidation
potential were required. The oxidation potentials of these new
compounds were determined as described above from the peak
potentials of irreversible cyclic voltammetry experiments and
the corresponding decarboxylation rate constants, according to
eqs 1.
The rate constant data are summarized in Scheme 3 and are
illustrated graphically in Figure 8. From these, a number of
conclusions can be drawn. Replacement of the acetate by
hydrogen (compounds 2 and 3) does not significantly affect the
rate of decarboxylation, as these points fall on the same curve
for compounds 1e-1h. This is presumably a consequence of
balancing solvation effects. The acetate is obviously larger than

Kinetics of Anilino Carboxylate Fragmentations
J. Phys. Chem. A, Vol. 108, No. 50, 2004 10955
was refluxed under nitrogen for 2 days. The mixture was cooled
and filtered, and the filterate was dissolved in dichloromethane
and extracted with aqueous sodium bicarbonate solution and
then with water. The solution was dried over anhydrous sodium
sulfate and filtered, and the solvent was distilled off. The residue
was distilled (108-115 °C at 0.05 mm) to give 31.9 g (72%)
of ethyl 2-phenylaminopropionate as colorless crystals.
To a solution of 24.6 g of this propionate ester and 31.9 g of
ethyl bromoacetate in 200 mL of acetonitrile 4.2 g potassium
iodide and 35 g anhydrous potassium carbonate were added,
and the mixture was refluxed under nitrogen for 6 days. The
mixture was cooled and filtered, and the filtrate was taken into
dichlormethane and washed with aqueous sodium bicarbonate
and with water. The residue after stripping off the solvent was
distilled (138-142 °C at 0.028 mm) to yield 20.8 g (50%) of
ethyl 2-(ethoxycarbonylmethylphenylamino)propionate (the di-
ethyl ester analogue of 1a).
A solution of 11.4 g of this diester and 3.3 g of sodium
hydroxide in a mixture of 10 mL of water, 10 mL of ethanol,
and 14 mL of tetrahydrofuran was refluxed for 15 h and cooled,
and the precipitated salt was collected and recrystallized from
ethanol to give 7.5 g (69%) of 1a, as colorless crystals. 1H NMR
(D2O): δ 1.42 (d, 3), 3.78 (d, 1), 4.05 (d, 1), 4.15 (q, 1), 6.58
(m, 2), 6.70 (m, 1), 7.25 (m, 2).
All other materials were available from previous work.6,9b
The aniline radical cations were formed using the DCA/
biphenyl cosensitization system in acetonitrile.32 Using DCA/
biphenyl, the aniline radical cations were formed with constant
and high yield, and we have used this method for studying the
kinetics of a wide variety of radical cation reactions using
transient absorption spectroscopy.6,9,30 The experimental details
for determining the rate constants for fragmentation of aniline
radical cations using this method have been described in detail
in a previous publication.9b The observed transient species were
assigned to radical cations, since their formation occurred in
concert with decay of the precursor biphenyl radical cation, their
lifetimes did not vary with dissolved oxygen, and their
fragmentation kinetics varied with structure as expected. As
discussed above, the actual absorption maxima of the radical
cations varied somewhat with reaction conditions (Figure 1) and
were not determined in all cases. In addition, the wavelength
chosen for observation of the decay kinetics was not necessarily
the maximum but that which gave the most easily analyzed
signal, which was typically between 530 and 630 nm. In all
cases the decays of the radical cations fit well to pseudo-first-
order kinetics over at least two lifetimes.
Figure 8. Plots of the logarithms of rate constants for decarboxylation
of radical cations vs oxidation potentials illustrating the effect of
micropolarity. The closed circles represent the N-acetate-N-propionate
derivatives 1e-1h, shown in Figure 2. The open circles represent N-H
derivatives, 2-4, and the open squares, N-butyl derivatives, 5-7. In
compounds 4 and 7, the propionate group is replaced by a succinate
(Scheme 3).
hydrogen, reducing solvent interaction at the reaction site for
simple steric reasons. However, acetate is also much more
hydrophilic, which may increase the effective water concentra-
tion at the reactive site. On the other hand, replacement of
acetate or hydrogen with a butyl group (compounds 5 and 6)
leads to an increase in the decarboxylation rate constant by ca.
2 orders of magnitude, after correcting for oxidation potential,
Figure 8. Presumably the hydrophobic group decreases the
effective water concentration at the reactive site. Increasing the
number of hydrophilic groups by replacing the propionate by a
succinate (compare compound 4 to 2 and 3 and also 7 to 5 and
6) decreases the rate constant for fragmentation by nearly an
order of magnitude, Figure 8.
These data clearly demonstrate that varying the hydrophobic/
hydrophilic nature of substituent groups can be used to control
specific solvation, and thus the rate constant for fragmentation
over almost 3 orders of magnitude, at constant oxidation
potential.
Summary
Decarboxylation of the radical cations of aryl carboxylates
is related to decarboxylation of acyloxy radicals via a valence
bond curve-crossing model. Both reactions involve an avoided
crossing and an electronic barrier to reaction. The rate constants
of the radical cation reactions depend strongly upon the
oxidation potential of the amine, the stability of the radical
product, and, most importantly, solvent polarity. Making use
of this information, we have been able to design a series of
molecules in which the decarboxylation rate constants, adjusted
for differences in oxidation potential, could be dramatically
varied by changing the molecular micropolarity.
The details of the electrochemical measurements will be given
elsewhere.17a
Acknowledgment. The work at Arizona State University is
supported by a grant from the National Science Foundation,
CHE-0213445, which is gratefully acknowledged. We would
like to thank M. Ezenyilimba, S. Godleski, and P. Zielinski
(Eastman Kodak Company) for the synthesis of the anilino
carboxylates. We acknowledge useful discussions related to the
TES chemistry with A. Muenter, S. Godleski, and P. Zielinski
(Eastman Kodak Company). We are very grateful to J. Dinno-
cenzo (University of Rochester) for extensive discussions
regarding the curve-crossing mechanism.
Experimental Section
The solvents were spectrograde and used as received. Water
was deionized and distilled. The anilino-dicarboxylates 1a-1h
and 2-7 were prepared by alkylation of the anilines with the
appropriate bromo-reagent as described in detail in ref 7b. An
illustrative example is reproduced here for compound 1a.
Synthesis of 1a. A suspension of 21.4 g of aniline, 50 g of
ethyl 2-bromopropionate, 4.6 g of potassium iodide, and 60 g
of anhydrous potassium carbonate in 300 mL of acetonitrile
References and Notes
(1) For example, see: (a) Anderson, J. M.; Kochi, J. K. J. Am. Chem.
Soc. 1970, 92, 2450. (b) Davidson, R. S.; Steiner, P. R. J. Chem. Soc.,
Perkin Trans. 2 1972, 1557. (c) Meiggs, T. O.; Grossweiner, L. I.; Miller,
S. I. J. Am. Chem. Soc. 1972, 94, 7981. (d) Trahanovsky, W. S.; Cramer,

10956 J. Phys. Chem. A, Vol. 108, No. 50, 2004
Gould et al.
C. J.; Brixius, D. W. J. Am. Chem. Soc. 1974, 96, 1077. (e) Davidson, R.
S.; Orton, S. P. J. Chem. Soc., Chem. Commun. 1974, 209. (f) Dessau, R.
M.; Heiba, E. I. J. Org. Chem. 1975, 40, 3647. (g) Battacharyya, S. M.;
Das, P. K. J. Chem. Soc., Faraday Trans. 1984, 80, 1107. (h) DeCosta, D.
P.; Pincock, J. A. J. Am. Chem. Soc. 1989, 111, 8948. (i) Jeon, Y. T.; Lee,
C. P.; Mariano, P. M. J. Am. Chem. Soc. 1991, 113, 8863. (j) Budac, D.;
Wan, P. J. Photochem. Photobiol., A 1992, 67, 135. (k) d’Alessandro, N.;
Albini, A.; Mariano, P. M. J. Org. Chem. 1993, 58, 937. (l) Wang, D.;
Schuchmann, H. P.; von Sonntag, C. Z. Naturforsch. B 1993, 48, 761. (m)
Nevill, S. M.; Pincock, J. A. Can. J. Chem. 1997, 75, 232. (n) Fleming, S.
A.; Pincock, J. A. Mol. Supramol. Photochem. 1999, 3, 211. (o) Xu, M.;
Wan, P. Chem. Commun. 2000, 21, 2147. (p) Morrison, J.; Osthoff, H.;
Wan, P. Photochem. Photobiol. Sci. 2002, 1, 384.
(2) For example, see: (a) Richard, J.; Job, D. Eur. J. Biochem. 1974,
44, 359. (b) Monig, J.; Chapman, R.; Asmus, K.-D. J. Phys. Chem. 1985,
89, 3139. (c) Cleland, W. W. Acc. Chem. Res. 1992, 25, 862. (d) Komuro,
M.; Nagatsu, Y.; Higuchi, T.; Hirobe, M. Tetrahedron Lett. 1992, 33, 4949.
(e) Candeias, L. P.; Folkes, L. K.; Dennis, M. F.; Patel, K. B.; Everett, S.
A.; Stratford, M. R. L.; Wardman, P. J. Phys. Chem. 1994, 98, 10131. (f)
Bobrowski, K.; Hug, G. L.; Marciniak, B.; Kozubek, H. J. Phys. Chem.
1994, 98, 537. (g) Candeias, L. P.; Folkes, L. K.; Porssa, M.; Parrick, J.;
Wardman, P. Biochemistry 1996, 35, 102. (h) Bonifacic, M.; Stefanic, I.;
Hug, G. L.; Armstrong, D. A.; Asmus, K.-D. J. Am. Chem. Soc. 1998, 120,
9930.
(3) (a) Eberson, L. In The Chemistry of Carboxylic Acids; Patai, S.,
Ed.; Wiley-Interscience: New York, 1969. (b) Utley, J. H. P. Techniques
in Electroorganic Synthesis, Vol. 5; Wiley: New York, 1974. (c) Eberson,
L.; Utley, J. H. P. In Organic Electrochemistry; Baizer, M. M., Lund, H.,
Eds.; Wiley: New York, 1983. (d) Schafer, H. J. Angew. Chem., Int. Ed.
Engl. 1981, 20, 911.
(4) (a) Koenig, T. In Free Radicals, Vol. 1; Kochi, J, K., Ed.; Wiley:
New York, 1973. (b) Kochi, J. K. Free Radicals 1973, 2, 665. (c) Wentrup,
C. ReactiVe Molecules; Wiley: New York, 1984. (d) Howard, J. A. Chem.
Peroxides 1983, 235.
Pratt, A.; Albini, A.; Long, C. J. Am. Chem. Soc. 1998, 120, 284. (p) Su,
Z.; Mariano, P. M.; Falvey, D. E.; Yoon, U. C.; Oh, S. W. J. Am. Chem.
Soc. 1998, 120, 10676.
(11) Andrieux, C. P.; Gonzalez, F.; Saveant, J.-M. J. Electroanal. Chem.
2001, 498, 171.
(12) Baciocchi, E.; Bietti, M. J. Chem. Soc., Perkin Trans. 2 2002, 720.
(13) (a) Popielarz, R.; Arnold, D. R. J. Am. Chem. Soc. 1990, 112, 3068.
(b) Maslak, P.; Vallombroso, T. M.; Chapman, W. H., Jr.; Narvaez, J. N.
Angew. Chem., Int. Ed. Engl. 1994, 33, 73. (c) Gaillard, E. R.; Whitten, D.
G. Acc. Chem. Res. 1996, 29, 292.
(14) Shida, T. Electronic Absorption Spectra of Radical Ions; Elsevier:
Amsterdam, 1988.
(15) It is assumed that the carboxylate groups are deprotonated under
the present conditions of 20% water in acetonitrile. Oxidative decarboxy-
lation is significantly slower for carboxylic acids than for the corresponding
carboxylates.12 The carboxylates would obviously be fully deprotonated in
water, and since the acidity of carboxylic acids decreases with decreasing
solvent polarity, the reaction rate might be expected to decrease with
decreasing water content due to protonation of the carboxylates. In fact the
opposite is observed (Figure 4). If one of the two carboxylates were
protonated, different rates of reaction would be observed if the proton were
on the propionate versus the acetate groups in compounds 1e-1h, and such
differential kinetics were not observed.
(16) (a) Zhang, X.; Yeh, S.-R.; Hong, S.; Freccero, M.; Albini, A.;
Falvey, D. E.; Mariano, P. S. J. Am. Chem. Soc. 1994, 116, 4211. (b)
Dombrowski, G.; Dinnocenzo, J. P.; Goodman, J. L.; Gould, I. R.; Farid,
S. Manuscript in preparation.
(17) (a) Gould, I. R.; Wosinska, Z.; Lenhard, J. L.; Farid, S. Manuscript
in preparation. (b) Nicholson, R. S.; Shain, I. Anal. Chem. 1964, 36, 706.
(18) Moore, J. W.; Pearson, R. G. Kinetics and Mechanism, 3rd ed.;
Wiley: New York, 1981.
(19) Smyth, C. P. Dielectric behaVior and structure; dielectric constant
and loss, dipole moment and molecular structure; McGraw-Hill: New York,
1955.
(5) (a) Crivello, J. V. AdV. Polym. Sci. 1984, 62, 1. (b) Reiser, A.
PhotoreactiVe Polymers, the Science and Technology of Resists; Wiley:
New York, 1989. (c) Saeva, F. D Top. Curr. Chem. 1990, 156, 59. (d)
DeVoe, R. J.; Olofson, P. M.; Sahyun, M. R. V. AdV. Photochem. 1992,
17, 313. (e) Saeva, F. D. AdV. Electron-Transfer Chem. 1994, 4, 1. (f)
Paczkowski, J.; Neckers, D. C. In Electron Transfer in Chemistry, Volume
5; Balzani, V., Ed.; Wiley-VCH: New York, 2001; Vol. 5, p 516.
(6) (a) Gould, I. R.; Lenhard, J. R.; Muenter, A. A.; Godleski, S. A.;
Farid, S. J. Am. Chem. Soc. 2000, 122, 11934. (b) Gould, I. R.; Lenhard,
J. R.; Muenter, A. A.; Godleski, S. A.; Farid, S. Pure Appl. Chem. 2001,
73, 455.
(7) (a) Farid, S. Y.; Lenhard, J. R.; Chen, C. H.; Muenter, A. A.; Gould,
I. R.; Godleski, S. A.; Zielinski, P. A. U.S. Patent 5,747,235, May 5, 1998.
(b) Farid, S. Y.; Lenhard, J. R.; Chen, C. H.; Muenter, A. A.; Gould, I. R.;
Godleski, S. A.; Zielinski, P. A.; Weidner, C. H. U.S. Patent 5,747,236,
May 5, 1998. (c) Gould, I. R.; Farid, S.; Godleski, S. A.; Lenhard, J. R.;
Muenter, A. A.; Zielinski, P. A. U.S. Patent 5,994,051, Nov. 30, 1999. (d)
Farid, S. Y.; Lenhard, J. R.; Chen, C. H.; Muenter, A. A.; Gould, I. R.;
Godleski, S. A.; Zielinski, P. A.; Weidner, C. H. U.S. Patent 6,010,841,
Jan. 4, 2000. (e) Adin, A.; Looker, J. J.; Farid, S. Y.; Gould, I. R.; Godleski,
S. A.; Lenhard, J. R.; Muenter, A. A.; Vishwakarma, L. C.; Zielinski, P.
A. U.S. Patent 6,054,260, Apr. 25, 2000. (f) Farid, S. Y.; Gould, I. R.;
Godleski, S. A.; Lenhard, J. R.; Muenter, A. A.; Zielinski, P. A.; Weidner,
C. H. U.S. Patent 6,153,371, Nov. 28, 2000. (g) Adin, A.; Looker, J. J.;
Farid, S. Y.; Gould, I. R.; Godleski, S. A.; Lenhard, J. R.; Muenter, A. A.;
Vishwakarma, L. C.; Zielinski, P. A. U.S. Patent 6,306,570, Oct. 23, 2001.
(8) Fyson, J. R.; Twist, P. J.; Gould, I. R. In Electron Transfer
Processes in SilVer Halide Photography; Balzani, V., Ed.; Electron Transfer
in Chemistry. Vol. 5; Wiley-VCH: New York, 2001; p 285.
(20) CRC Handbook of Chemistry and Physics, 84th ed.; Lide, D. R.,
Ed.; CRC Press: Boca Raton, FL, 2003.
(21) (a) The oxidation potential of the carboxylate anion depends strongly
upon conditions and is not easy to determine accurately; however, in aqueous
acetonitrile it is estimated to be ∼1.9 V vs SCE.21b (b) Eberson, L. E.
Electron Transfer Reactions in Organic Chemistry; Springer-Verlag: New
York, 1988.
(22) (a) Shaik, S. S. J. Am. Chem. Soc. 1981, 103, 3692. (b) Pross, A,
“Theoretical Physical Principles of Organic Reactivity”; Wiley: New York,
1995. (c) Shaik, S., Shurki, A. Angew. Chem., Int. Ed. Engl. 1999, 38, 586.
(23) Saveant has described a curve-crossing model for fragmentation
of the radical canions of aryl and benzyl halides in several publications,
See, for example: (a) Saveant, J.-M. J. Phys. Chem. 1992, 96, 3716. (b)
Saveant, J.-M. J. Phys. Chem. 1994, 98, 3716. (c) Pause, L.; Robert, M.;
Saveant, J.-M. J. Am. Chem. Soc. 2000, 122, 9829. (d) Pause, L.; Robert,
M.; Saveant, J.-M. J. Am. Chem. Soc. 2000, 122, 9829. (e) Costentin, C.;
Robert, M.; Saveant, J.-M. J. Am. Chem. Soc. 2003, 125, 105.
(24) For a full state correlation treatment of decarboxylation of acyloxy
radicals, see: Pacansky, J.; Brown, D. W. J. Phys. Chem. 1983, 87, 1553.
(25) Butler, L. J. Annu. ReV. Phys. Chem. 1998, 49, 125.
(26) (a) Hopfield, J. J. Proc. Natl. Acad. Sci. U.S.A. 1974, 71, 3640. (b)
Marcus, R. A. J. Chem. Phys. 1984, 81, 4494. (c) Ulstrup, J.; Jortner, J. J.
Chem. Phys. 1975, 63, 4358.
(27) Closs, G. L.; Johnson, M. D.; Miller, J. R.; Piotrowiak, P. J. Am.
Chem. Soc. 1989, 111, 3751.
(28) Isaacs, N. Physical Organic Chemistry, 2nd ed.; Longman: U.K.,
1987.
(29) (a) Andrieux, C. P.; Merz, A.; Saveant J.-M. J. Am. Chem. Soc.
1985, 107, 6097. (b) Costentin, C.; Hapiot, P.; Medebielle, M.; Saveant,
J.-M. J. Am. Chem. Soc. 1999, 121, 4451. (c) Pause, L.; Robert, M.; Saveant,
J.-M. J. Am. Chem. Soc. 1999, 121, 7158. (d) Costentin, C.; Robert, M.;
Saveant, J.-M. J. Phys. Chem. 2000, 104, 7492.
(9) (a) Oxidative desilylation reactions related to TES have been
described recently.9b (b) Gould, I. R.; Godleski, S. A., Zielinski, P. A.;
Farid, S. Can. J. Chem. 2003, 81, 777.
(30) (a) Karki, S. B.; Dinnocenzo, J. P.; Farid, S.; Goodman, J. L.; Gould,
I. R.; Zona, T. A. J. Am. Chem. Soc. 1997, 119, 431. (b) Cuppoletti, A.;
Dinnocenzo, J. P.; Goodman, J. L.; Gould, I. R. J. Phys. Chem. A 1999
103, 11253. (c) Wang, Y.; Luttrull, D. K.; Dinnocenzo, J. P.; Goodman, J.
L.; Farid, S.; Gould, I. R. Photochem. Photobiol. Sci. 2003, 11, 1169.
(31) Proton-coupled electron-transfer reactions represent special class
of bond-coupled electron-transfer reactions: (a) Biczok, L.; Linschitz, H.
J. Phys. Chem. 1995, 99, 1843. (b) Biczok, L.; Gupta, N.; Linschitz, H. J.
Am. Chem. Soc. 1997, 119, 12601. (c) Cukier, R. I.; Nocera, D. G. Annu.
ReV. Phys. Chem. 1998, 49, 337. (d) Sjoedin, M.; Styring, S.; Kermark, B.;
Sun, L.; Hammarstroem, L. J. Am. Chem. Soc. 2000, 122, 3932. (e)
Georgievskii, Y.; Stuchebrukhov, A. A. J. Chem. Phys. 2000, 113, 10438.
(f) Cukier, R. I. J. Phys. Chem. B 2002, 106, 1746. (g) Weatherly, S. C.;
Yang, I. V.; Armistead, P. A.; Thorp, H. H. J. Phys. Chem. B 2003, 107,
372.
(10) (a) Moenig, J.; Chapman, R.; Asmus, K. D. J. Phys. Chem. 1985,
89, 3139. (b) Yamauchi, S.; Hirota, N.; Takahara, S.; Sakuragi, H.;
Tokumaru, K. J. Am. Chem. Soc. 1985, 107, 5021. (c) Falvey, D. E.;
Schuster, G. B. J. Am. Chem. Soc. 1986, 108, 7419. (d) Chateauneuf, J.;
Lusztyk, J.; Ingold, K. U. J. Am. Chem. Soc. 1987, 109, 897. (e)
Chateauneuf, J.; Lusztyk, J.; Ingold, K. U. J. Am. Chem. Soc. 1988, 119,
2886. (f) Steenken, S.; Warren, C. J.; Gilbert, B. C. J. Chem. Soc., Perkin
Trans. 2 1990, 335. (g) Hilborn, J. W.; Pincock, J. A. J. Am. Chem. Soc.
1991, 113, 2683. (h) Prasad, D. R.; Hoffman, M. Z.; Mulazzani, Q. C.;
Rodgers, M. A. J. J. Photochem. Photobiol. A 1992, 67, 135. (i) Jovanovic,
S. V.; Steenken, S. J. Phys. Chem. 1992, 96, 6674. (j) Hilborn, J. W.;
Pincock, J. A. Can. J. Chem. 1992, 70, 992. (k) Candeias, L. P.; Folkes, L.
K.; Porssa, M.; Parrick, J.; Wardman, P. Free Radical Res. 1995, 23, 403.
(l) Bockman, T. M.; Hubig, S. M.; Kochi, J. K. J. Org. Chem. 1997, 62,
2210. (m) Steenken, S.; O’Neill, P.; Schulte-Frohlinde, D. J. Phys. Chem.
1977, 81, 26. (n) Mehta, L. K.; Porssa, M.; Parrick, J.; Candeias, L. P.;
Wardman, P. J. Chem. Soc., Perkin Trans. 2 1997, 1487. (o) Freccero, M.;
(32) Gould, I. R.; Ege, D.; Moser, J. E.; Farid, S. J. Am. Chem. Soc.
1990, 112, 4290.