Chain-amplified photochemical fragmentation of N-alkoxypyridinium salts: Proposed reaction of alkoxyl radicals with pyridine bases to give pyridinyl radicals

Article abstract of DOI:10.1021/jo301975j

Photoinduced electron transfer to N-alkoxypyridiniums, which leads to N-O bond cleavage and alkoxyl radical formation, is highly chain amplified in the presence of a pyridine base such as lutidine. Density functional theory calculations support a mechanism in which the alkoxyl radicals react with lutidine via proton-coupled electron transfer (PCET) to produce lutidinyl radicals (BH^?). A strong electron donor, BH^? is proposed to reduce another alkoxypyridinium cation, leading to chain amplification, with quantum yields approaching 200. Kinetic data and calculations support the formation of a second, stronger reducing agent: a hydrogen-bonded complex between BH^? and another base molecule (BH^??B). Global fitting of the quantum yield data for the reactions of four pyridinium salts (4-phenyl and 4-cyano with N-methoxy and N-ethoxy substituents) led to a consistent set of kinetic parameters. The chain nature of the reaction allowed rate constants to be determined from steady-state kinetics and independently determined chain-termination rate constants. The rate constant of the reaction of CH₃O^? with lutidine to form BH^?, k₁, is ～6 × 10⁶ M^-1 s^-1; that of CH₃CH₂O ^? is ～9 times larger. Reaction of CD₃O ^? showed a deuterium isotope effect of ～6.5. Replacing lutidine by 3-chloropyridine, a weaker base, decreases k₁ by a factor of ～400.

Full text of DOI:10.1021/jo301975j

Article
pubs.acs.org/joc
Chain-Amplified Photochemical Fragmentation of
N‑Alkoxypyridinium Salts: Proposed Reaction of Alkoxyl Radicals
with Pyridine Bases To Give Pyridinyl Radicals
Deepak Shukla,*,† Shashishekar P. Adiga,*,† Wendy G. Ahearn,† Joseph P. Dinnocenzo,*,‡
and Samir Farid*,‡
†Research Laboratories, Eastman Kodak Company, Rochester, New York 14650, United States
‡Department of Chemistry, University of Rochester, Rochester, New York 14627, United States
S
* Supporting Information
ABSTRACT: Photoinduced electron transfer to N-alkoxypyridiniums, which leads to
N−O bond cleavage and alkoxyl radical formation, is highly chain amplified in the
presence of a pyridine base such as lutidine. Density functional theory calculations
support a mechanism in which the alkoxyl radicals react with lutidine via proton-coupled
electron transfer (PCET) to produce lutidinyl radicals (BH•). A strong electron donor,
BH• is proposed to reduce another alkoxypyridinium cation, leading to chain
amplification, with quantum yields approaching 200. Kinetic data and calculations
support the formation of a second, stronger reducing agent: a hydrogen-bonded
complex between BH• and another base molecule (BH•···B). Global fitting of the
quantum yield data for the reactions of four pyridinium salts (4-phenyl and 4-cyano with
N-methoxy and N-ethoxy substituents) led to a consistent set of kinetic parameters. The
chain nature of the reaction allowed rate constants to be determined from steady-state
kinetics and independently determined chain-termination rate constants. The rate constant of the reaction of CH3O• with
lutidine to form BH•, k1, is ∼6 × 106 M−1 s−1; that of CH3CH2O• is ∼9 times larger. Reaction of CD3O• showed a deuterium
isotope effect of ∼6.5. Replacing lutidine by 3-chloropyridine, a weaker base, decreases k1 by a factor of ∼400.
INTRODUCTION
■
Reactions of alkoxyl radicals have received considerable
attention because of their importance in a variety of organic,1
biological,2 and atmospheric3 processes. These radicals undergo
a variety of inter- and intramolecular reactions such as
unimolecular fragmentation,4 hydrogen atom abstraction,5
addition to unsaturated compounds,6 and isomerization.7
Hydrogen atom abstraction reactions by alkoxyl radicals are
of particular interest because of their importance in enzymatic
and biological systems.8 Alkoxyl radicals also play an important
role in atmospheric oxidation of volatile organic compounds,
For example, alkoxyl radicals generated from alkoxypyridinium
where isomerization, fragmentation, and reaction with O2 are
salts have been used to initiate free radical polymerization,
the dominant reactions.3 Intramolecular isomerization of
which extended the photosensitivity of these polymerizations to
light throughout the visible region.11 In addition, an interesting
alkoxyl radicals via 1,5-hydrogen shift to the corresponding
hydroxyalkyl radicals is well-known in solution,7a−c and similar
aspect of electron transfer reactions of the alkoxypyridiniums is
reactions are also proposed to take place in the gas phase.3 The
their ability to lead to chain amplification, as in the oxidation of
alcohols to ketones or aldehydes, eq 2.12 Another dramatic
isomerization of methoxyl radical to hydroxymethyl radical via
1,2-hydrogen shift is thermodynamically favorable but has a
example of chain amplification is described in the present work.
Importantly, the chain nature of the reaction was instrumental
in the determination of kinetic parameters for the methoxyl and
high barrier,9 although the isomerization has been shown to be
catalyzed by water,7d−f alcohols,7g and acid.7h
In this paper we describe the novel reaction of methoxyl and
ethoxyl radicals with pyridine bases. The alkoxyl radicals were
conveniently generated by photoinduced electron transfer to N-
alkoxypyridinium salts, which leads to fast N−O bond cleavage,
eq 1.10 N-Alkoxypyridinium salts undergo a range of photo-
chemical reactions that have been exploited in a variety of ways.
ethoxyl radicals, which otherwise would not have been
Special Issue: Howard Zimmerman Memorial Issue
Received: September 11, 2012
© XXXX American Chemical Society
A
dx.doi.org/10.1021/jo301975j | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
accessible through conventional flash photolysis because of
their lack of a convenient absorption chromophore.
RESULTS AND DISCUSSION
■
The photoreductions of four N-alkoxypyridinium salts (1a−d)
were examined: the substituent in the 4-position was either
phenyl or cyano, and the alkoxy substituent was either methoxy
or ethoxy. No reversible potential can be obtained for these
pyridiniums by using conventional electrochemical methods
due to fast fragmentation of the N−O bond upon one electron
reduction. Nonetheless, estimates of the reduction potentials of
the 4-phenyl and 4-cyano derivatives of −1.0 and −0.5 V vs
SCE in acetonitrile have been previously made11a based on a
comparison between charge transfer absorption bands of the N-
alkoxy and the corresponding N-alkyl pyridinium salts, which
show reversible potentials.13
Electron transfer to N-alkoxypyridinium salts can be induced
by either singlet or triplet excited donors.12a The energetic
requirements are given by eq 3, where (Eexcit)D is the excitation
energy of the sensitizing donor (singlet or triplet, depending on
which state reacts with the pyridinium), (Eox)D is the oxidation
potential of the donor, (Ered)A is the reduction potential of the
pyridinium, and Δ is an energy increment varying from tens of
meV in acetonitrile14 to ca. 0.3 eV in nonpolar media.15 Triplet
sensitizers are more efficient in reducing the pyridinium salts
because energy-wasting return electron transfer within the
geminate pair is spin forbidden. Thioxanthone, which has an
intersystem crossing (ISC) efficiency of 0.88 in acetonitrile,12a
has proven to be effective for these reactions.12a In the current
work 2-chlorothioxanthone (CTX) was used, which has an ISC
efficiency very close to unity16 and still meets the requirements
of eq 3.17 Consistent with this conclusion, triplet CTX is
quenched by 1a with a rate constant of 4.7 × 109 M−1 s−1 and
by 1c, a 0.5 eV more exergonic reaction, with a rate constant of
8.3 × 109 M−1 s−1.
Figure 1. Fragmentation quantum yields of the pyridinium salts (1a−
d), eq 4, in CD3CN as a function of lutidine concentration. The
average concentration of 1 between the start and end of irradiation is
∼0.018 M. 2-Chlorothioxanthone (CTX), 2 mM, was used as
sensitizer; excitation at 405 nm. See text for details regarding the
fitting curves.
(Eexcit)D > (Eox)D − (Ered)A + Δ
(3)
The fragmentation quantum yields for 1a−d induced by
triplet CTX are ∼1 in acetonitrile. Depending on the starting
pyridinium salt, the reaction product is 4-phenyl- or 4-
cyanopyridine. A dramatic increase in the quantum yield (up
to ∼200) was observed in the presence of lutidine. In these
latter cases, the reaction products were the pyridine derived
from the pyridinium salt, formaldehyde or acetaldehyde, and
the protonated lutidinium cation, eq 4. The mass balances for
conversion of the pyridiniums to their pyridine products were
uniformly >95%.
1. Substituents Effect. The dependence of the fragmenta-
tion quantum yields on the concentration of lutidine was
investigated with a starting concentration of the pyridinium
salts of 0.02 M. Irradiations were carried out to ∼20%
conversion. The quantum yields were determined by using
NMR spectroscopy from the ratio of the pyridine product to
the starting pyridinium salt. The estimated error in the
quantum yields is 5−10%. As shown in Figure 1, there is a
sharp increase in the quantum yields with increasing [lutidine]
up to ∼0.2 M, followed by a modest increase at higher
concentrations. The highest attainable quantum yield increases
by a factor of ∼2 upon changing the N-alkoxy group from
methoxy to ethoxy (cf. 1a vs 1b and 1c vs 1d) and also by
changing the 4-substituent from phenyl to cyano (cf. 1a vs 1c
and 1b vs 1d). Thus, there is a ∼4-fold increase in the limiting
quantum yield between 4-phenyl-N-methoxypyridinium (1a,
Φlim ∼50) and 4-cyano-N-ethoxypyridinium (1d, Φlim ∼200).
Increasing the concentration of the pyridinium salts also
leads to an increase in quantum yield, but the effect is much less
pronounced than that of increasing [lutidine]. For example, at
[lutidine] of 0.3 or 0.5 M, the quantum yield for 1b increases
upon increasing its concentration from 0.02 to 0.05 M by only
∼25% and, within the experimental error, stays at that level at
concentrations as high as 0.2 M (see the Supporting
Information). As described below, the differing sensitivities of
the quantum yield to the changes in lutidine and pyridinium
B
dx.doi.org/10.1021/jo301975j | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
concentrations are consistent with different contributions from
chain termination reactions.
2. Deuterium Isotope Effect. The deuterium isotope
effect on the α-hydrogens of the alkoxy group was evaluated by
comparing the reactions of 1a to an analogue in which the
OCH3 group was replaced by OCD3. As shown in Figure 2, as a
Figure 3. Comparison between the fragmentation quantum yields of
1b in the presence of 3-chloropyridine versus lutidine. The reaction
conditions are the same as in Figure 1. See text for details regarding
the fitting curves.
1) react with the added base (B) to give a second reaction
intermediate that is capable of transferring an electron to
another P+, thus propagating the chain reaction. To efficiently
reduce the 4-phenylpyridiniums (1a or 1b), the secondary
intermediate must have an oxidation potential similar to or
more negative than the reduction potential of the pyridiniums
(Ered ≈ −1 V vs SCE).
In principle, the reaction of the methoxyl or ethoxyl radical
with B could lead to several possible products. For example,
proton transfer from the alkoxyl radicals to B would generate
the radical anions of formaldehyde or acetaldehyde, both of
which could reduce P+ to propagate the chain. As discussed in
the next section, these proton transfers are kinetically
incompetent, i.e., the endergonicities for the proton transfers
exceed the activation barriers estimated for reaction of the
alkoxyl radicals with B based on the kinetic scheme used to fit
the quantum yield data. The reaction of methoxyl or ethoxyl
radical with B could alternatively result in their isomerization to
the corresponding α-hydroxy radicals, RCH•−OH (R = H or
CH3). Although these radicals have estimated oxidation
potentials19 less negative than −0.5 V vs SCE, making electron
transfer to the pyridinium cations energetically unfavorable, the
intermediates could react with the pyridine bases to form an
intermediate (RCH•−OH···B) that is capable of energetically
favorable electron transfer to the alkoxypyridiniums.20 As
described in detail below, quantum chemical calculations
instead predict that methoxyl (ethoxyl) radicals react with
lutidine or 3-chloropyridine to give formaldehyde (acetalde-
hyde) and the lutidinyl or 3-chloropyridinyl radical (BH•), eq
5a. To test whether these pyridinyl radicals are capable of
Figure 2. Deuterium isotope effect on the fragmentation quantum
yields (Φ) of 4-phenyl-N-methoxypyridinium (1a) as a function of the
lutidine concentration (top) and of the pyridinium concentration
(bottom). Methoxy group: OCH3 (red, filled circles); OCD3 (blue,
unfilled squares). The fitting curves are based on a deuterium isotope
effect for reaction of the methoxyl radical with lutidine (k1) of 6.5.
function of the concentration of lutidine or of the pyridinium
salt, the deuterated compound exhibited considerably lower
quantum yields. The magnitude of the deuterium isotope effect
of ∼6.5 was determined by globally fitting both sets of data (see
kinetics section).
3. Basicity of the Pyridine. To examine the effect of the
basicity of the pyridine, the reaction of 1b with 3-
chloropyridine (pKa ∼10)18 was compared to that described
above with lutidine (pKa ∼14).18 As shown in Figure 3, the
weaker base is much less effective in propagating the chain
reaction. The lack of leveling off in quantum yield at higher
concentrations of 3-chloropyridine and the fitting curves in
Figure 3 are discussed in the kinetics section.
4. Reaction Intermediates and Kinetic Scheme. As
shown above, the photosensitized fragmentation of N-
alkoxypyridinium cations (P+) proceeds by a chain reaction
mechanism in the presence of pyridine bases. It is plausible to
conclude that alkoxyl radicals generated by reduction of P+ (eq
reducing the pyridinum cations, leading to chain propagation,
their oxidation potentials were estimated from the reduction
potentials of the corresponding protonated bases. The cyclic
C
dx.doi.org/10.1021/jo301975j | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
1
voltammograms of protonated lutidine and 3-chloropyridine
showed irreversible reductions in acetonitrile with peak
potentials at −1.58 and −1.16 V vs SCE, respectively (see
Supporting Information), indicating that reduction of the
pyridinium cations by both pyridinyl radicals (BH•) is
energetically favorable.
Quantum chemical calculations also revealed that the
pyridinyl radicals (BH•) were capable of forming hydrogen-
bonded complexes (BH•···B) with their pyridine precursors, eq
5b, and that these complexes were stronger reducing agents than
BH• alone (see Section 6). Importantly, a consistent set of
kinetic parameters to model the experimental quantum yields
over the entire range of base concentrations could not be
obtained for all four pyridinium cations 1a−d unless the
BH•···B complexes were included in the kinetic analysis.
Scheme 1 shows the full range of kinetic processes needed to
globally fit the combined quantum yield data. In addition to
Φ =
αβ + αγδ
1 − γε
1 −
(6)
where
k1[B]
1
α =
=
1 + a + b[B−1]
k1[B] + k′1[B] + k2
k4[P+]
k4[P+] + k′4[P+] + k5 + k3[B]
β =
1
=
1 + c + d[P+]−1 + e[B][P+]−1
k3[B]
γ =
k4[P+] + k′4[P+] + k5 + k3[B]
e[B][P+]−1
=
1 + c + d[P+]−1 + e[B][P+]−1
Scheme 1
k6[P+]
k6[P+] + k′6[P+] + k7 + k−3
1
δ =
=
1 + f + g[P+]−1 + h[P+]−1
k−3
ε =
k6[P+] + k′6[P+] + k7 + k−3
h[P+]−1
=
1 + f + g[P+]−1 + h[P+]−1
k′1
k1
k2
k1
a =
c =
f =
b =
d =
g =
k′4
k4
k5
k4
k3
k4
e =
h =
initiation and propagation reactions, several chain-termination
steps are included in the kinetic scheme. Two of them, k′1 and
k2, involve the alkoxyl radicals. Competing with the reaction of
RO• with the pyridine bases that leads to the formation of BH•,
k1, there is a minor, chain-terminating hydrogen atom
abstraction from the methyl groups of lutidine, k′1, when this
base is used. The other termination rate constant, k2, is
analogous to that encountered in a related system.12a The k2
step represents the sum of two reactions: deuterium atom
abstraction from the solvent CD3CN and a minor reaction
between RO• and the photosensitizer, 2-chlorothioxanthone
(CTX).12a Competing with the propagation reaction of BH•,
k4, are two possible termination reactions, k′4 and k5. The
corresponding reactions for BH•···B are k6, k′6, and k7. In both
cases k′4 and k′6 represent reactions with trace impurities in P+
and/or minor reactions of the radicals with P+, other than the
chain-propagating electron transfer. As explained in Section 5,
the termination steps k5 and k7 with lutidine as base can
predominantly be attributed to electron transfer to the
sensitizer, CTX.
k′6
k6
k7
k6
k−3
k6
Although there are several fitting parameters (a−h), global
analysis was helpful in limiting their range because some of the
rate constant ratios are constrained by being common among
certain pairs of the four pyridiniums 1a−d. For example, the
rate constant for reaction of the methoxyl radical with lutidine
(k1)MeO does not depend on the starting pyridinium (1a or 1c);
thus, a and b were kept constant for these two compounds. The
same applies for the reaction of the ethoxyl radical (k1)EtO
,
whether from 1b or 1d. In addition, the ratio a/b = k′1/k2 is
expected to be constant for all reactions with lutidine. Because
k′1/k2 is the ratio of rate constants for hydrogen abstraction
from lutidine to the sum of deuterium abstraction from the
solvent and reaction with the sensitizer, it is unlikely that it will
differ between methoxyl and ethoxyl radicals.
The electron transfer rate constants to 4-phenylpyridiniums,
(k4)Ph and (k6)Ph, are assumed to be the same for the N-
methoxy- and N-ethoxy-pyridiniums, 1a and 1b; thus, the ratios
d, e, g, and h were kept constant for these two compounds. The
same applies to (k4)CN and (k6)CN, the corresponding electron
transfer reactions of the 4-cyanopyridiniums, 1c and 1d.
Another constraint derives from the reactant independence of
the equilibration rate constants between BH• and BH•···B, k3
and k−3, which are common for all pyridiniums reacting with a
Based on Scheme 1, the quantum yield for consumption of
the N-alkoxypyridininium salt (P+) can be expressed in terms of
eq 6 (see Supporting Information for derivation). Because the
quantum yield depends only on ratios of rate constants and not
on their absolute values, it was convenient to use fractionation
factors (α−ε), which in turn can be expressed in terms of rate
constants ratios (a−h).
D
dx.doi.org/10.1021/jo301975j | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
single pyridine base. The rate constants k5 and k7, which
represent termination reactions, are also common for all
pyridiniums. Therefore, e/d = k3/k5 and h/g = k−3/k7 should be
the same for all pyridiniums with lutidine as base.21 Values of all
the global-fitting parameters are listed in the Supporting
Information.
minimum activation free energy for the proton transfer reaction.
The experimental activation free energy (ΔG⧧exp) can be
estimated from the rate constant k1 at 295K to be ∼8 kcal/mol.
That ΔG°pt ≫ ΔG⧧ excludes proton transfer for the k1 step
exp
in Scheme 1. As discussed below, computational results provide
independent evidence against a proton transfer mechanism for
the reaction of the alkoxyl radicals with pyridine bases.
5. Rate Constants. Reaction of Alkoxyl Radicals with the
Base, k1. As indicated in the Introduction, a key aspect of this
work is the reaction of alkoxyl radicals with pyridine bases that
is proposed to lead to the reducing radical BH•, k1. Although,
the quantum yield data afford only ratios of rate constants, an
estimate of k1 can be obtained from the ratio b = k2/k1 because
the value of k2 (4.2 × 103 s−1) is available from previous work.22
To ensure internal consistency, we first checked if the kinetic
analyses led to congruent relationships between the independ-
ently determined value of k2 and that of k′1, the other
termination rate constant of the alkoxyl radical. As mentioned
above, the ratio a/b = k′1/k2 was kept constant for all reactions
with lutidine. The best value for this ratio from the global fitting
was 10 M−1, which corresponds to k′1 of 4.2 × 104 M−1 s−1 for
reaction of the alkoxyl radicals with lutidine. For comparison,
the rate constant for hydrogen abstraction from toluene by tert-
butoxyl radical is ∼2 × 105 M−1 s−1.5a Considering there are
two methyl groups in lutidine versus one in toluene leads to a
∼10-fold difference in the rate constants for hydrogen atom
abstraction, or ∼1.3 kcal/mol in activation barrier. However,
hydrogen atom abstraction by tert-butoxyl radical from toluene
is 3 kcal/mol more favorable than hydrogen atom abstraction
by methoxyl (or ethoxyl) radical from lutidine,23 which could
readily account for the lower reactivity of lutidine. This
comparatively good agreement with independent literature data
supports the general reliability of the fitting procedure.
Fitting the kinetic data for methoxypyridiniums 1a and 1c in
Figure 1 gave k2/k1 of 7.2 × 10−4 M. Based on the value of k2
mentioned above, k1 for the reaction of CH3O•/lutidine,
(k1)MeO, is estimated to be 5.8 × 106 M−1 s−1. Fitting the kinetic
data for the ethoxypyridiniums, 1b and 1d, required k2/k1 of 8
× 10−5 M, yielding (k1)EtO of 5.3 × 107 M−1 s−1.
The significant decrease in quantum yield upon replacing the
CH3O group in 1a by CD3O shown in Figure 2 is reflected in a
factor of 6.5 increase in k2/k1 required to fit the data, which can
most plausibly be attributed to a 6.5-fold decrease in k1. This is
clearly a primary isotope effect, consistent with the proposed
hydrogen atom transfer from the methoxyl radical to lutidine.
The precipitous decrease in quantum yield using 3-
chloropyridine instead of lutidine in the reaction of 1b (Figure
3) required a ∼400-fold decrease in k1 to fit the data. As
expected, because of the absence of abstractable hydrogen in
chloropyridine the best fit was obtained with no contribution
from k′1 in this case. The absence of chain termination via k′1
with 3-chloropyridine explains the lack of leveling off in the
quantum yields at high base concentrations (Figure 3). The
leveling off in quantum yield is a dominant feature of the
reactions with lutidine (Figure 1), where the plateau is defined
largely by 1/a or k1/k′1.
Propagation Reactions k4 and k6. Based on the parameters
derived from the data fitting (see the Supporting Information),
(k4)Ph = 0.15 × (k4)CN and (k6)Ph = 0.6 × (k6)CN. Being rate
constants for strongly exergonic electron transfer reactions
(driving force >1 eV), which are likely to proceed over long
distance, both (k4)CN and (k6)CN are expected to be ∼1.2 ×
1010 M−1 s−1.25 This value yields (k4)Ph of ∼2 × 109 M−1 s−1
and (k6)Ph of ∼7 × 109 M−1 s−1, which are reasonable for the
reactions of 1a and 1b that are 0.5 eV less exergonic than those
of 1c and 1d. The larger value of (k6)Ph compared to (k4)Ph can
be attributed to the ∼0.15 eV greater reducing power of
BH•···B compared to that of BH• (see Section 6).
From the ratio d, the pseudo first-order rate constant of the
termination reaction k5 is ∼1.8 × 106 s−1. More than one
reaction may be represented by k5, for example, electron
transfer to residual oxygen, oxidant impurities, and/or the
sensitizer, CTX. The last reaction, however, is most likely the
main contributor. CTX has reduction potential of −1.53 V vs
SCE,17a which allows for energetically favorable electron
transfer from BH• and at [CTX] = 0.002 M can account for
the observed rate constant for k5.26
With 3-chloropyridine as base, there are insufficient quantum
yield data to obtain reliable estimates for k4 and k6. The
reducing power of BH• in this case is certainly less than when
the base is lutidine. Combined with a relatively large driving
force to form the hydrogen-bonded complex, BH•···B, which is
a stronger reducing agent than BH• (see Section 6), it is likely
that chain-propagating electron transfer to the pyridinium
cation 1b proceeds predominantly via k6 in this case.
Equilibrium Constant (k3/k−3). Based on the assumption
that (k4)CN ≈ (k6)CN ≈ 1.2 × 1010 M−1 s−1, the ratios e and h
for these reactions, 0.3 and 0.0072, respectively, yield k3 of 3.6
× 109 M−1 s−1 and k−3 of 8.7 × 107 s−1. From these values (k3/
k−3) is ∼40, i.e., BH•···B is more stable than BH• by 0.094 eV;
the computed stabilization energy is 0.076 eV (see below).
6. Computational Results. As described above, the
quantum yield for the chain fragmentation of N-alkoxypyr-
idinium cations is strongly dependent on the basicity of the
pyridine base. This observation is consistent with the reaction
of the alkoxyl radicals and the bases having some proton-
transfer character. We therefore initiated computational studies
that focused on reaction of the α-C−H bonds of the alkoxyl
radicals with the nitrogen atom of the pyridine bases.
All calculations were done using the hybrid B3LYP density
functional method.27 All open-shell calculations were per-
formed with the unrestricted UB3LYP method. Geometry
optimizations, transition structure searches, and vibrational
analyses were performed with a 6-311++(d,p) basis set. The
polarizable continuum model (PCM)28 was used to model the
reaction solvent, acetonitrile. The Universal Force Field
(UFF)29 radii for all atoms, including hydrogen atoms, were
used in cavity building. All of the calculations were carried out
with the Gaussian 03 suite of programs.30 Vibrational
frequencies were determined by using the analytic Hessian,
calculated for each local minimum and each transition state
structure. In all cases, local minima had positive vibrational
As mentioned above, the rate constant k1 for reaction of
CH3O• with lutidine can be used to evaluate the kinetic
competence of the proton transfer mechanism to form
formaldehyde radical anion and the lutidinium cation. The
pKa of CH3O• in acetonitrile is estimated to be ∼25.24 Based
on the pKa of the lutidinium cation in acetonitrile (∼14),18
proton transfer from CH3O• to lutidine is estimated to be
endergonic by ∼14 kcal/mol (ΔG°pt). This corresponds to the
E
dx.doi.org/10.1021/jo301975j | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
frequencies and transition state structures had only one
imaginary frequency.
Calculations on the reactions of the alkoxyl radicals with the
pyridine bases showed that a weakly bound reactant complex is
formed (see Figure 4), where the α-C−H bond of the alkoxyl
(2.7 and 3.1 kcal/mol, respectively) and significantly higher for
reaction of ethoxyl radical with 3-chloropyridine (5.6 kcal/
mol). The higher calculated barrier for the reaction of ethoxyl
radical with 3-chloropyridine is in agreement with experiment.
The predicted barriers for reaction of methoxyl and ethoxyl
radical with lutidine are reversed relative to the experimental
data, although the energetic differences are relatively small.
Before discussing the detailed nature of the hydrogen atom
transfer reactions, we note that our calculations show that
deprotonation of the alkoxyl radicals by the pyridine bases are
energetically unfavorable relative to the hydrogen atom transfer
reactions. For example, deprotonation of the methoxyl and
ethoxyl radicals by the stronger base, lutidine, are both
predicted to be endothermic relative to the reactant complexes
by ∼10−11 kcal/mol (see Supporting Information). Thus,
deprotonation of the alkoxyl radicals is not expected to
measurably compete with the hydrogen atom transfer reactions.
This conclusion is consistent with the kinetic competence test
described above that similarly excluded a proton transfer
mechanism.
Interestingly, our calculations reveal that the transition states
for hydrogen atom transfer have a significant degree of ionic
character. For example, in the reaction of the methoxyl radical
with lutidine, the Mulliken charges on the formaldehyde
fragment, the lutidine fragment, and the transferring hydrogen
atom are −0.76, +0.21, and +0.55, respectively. Similar results
were found for the other hydrogen atom transfer reactions (see
Supporting Information). These group charges are consistent
with a reaction that has significant proton-transfer character. In
support of this idea, examination of the singly occupied
molecular orbital (SOMO) for the transition state (Figure 5)
Figure 4. Calculated energy differences (in kcal/mol) between
intermediates relative to that of an alkoxyl radical plus a pyridine
base (1). Interaction between the pair (see inserts) leads to a
hydrogen-bonded complex (2). Hydrogen atom transfer from (2)
leads to a pyridinyl radical, hydrogen-bonded to an aldehyde (3).
Return hydrogen atom transfer from (3) leading to a hydrogen-
bonded complex between an α-hydroxy radical and the corresponding
pyridine (4). Energies of the transition states for the conversion of (2)
to (3) and (3) to (4) are marked by TS. Structures and bond lengths
are given in the Supporting Information. Calculational details are given
in the text.
Figure 5. Singly occupied molecular orbital (SOMO) density surface
of the transition state for hydrogen atom transfer from methoxyl
radical to lutidine computed with an isodensity value of 0.12 au.
radical is hydrogen-bonded to the nitrogen atom of the pyridine
base.31 Interestingly, reaction of this complex leads, in all cases,
to transfer of a hydrogen atom rather than a proton from the
alkoxyl radical to the nitrogen of the pyridine base, resulting in
the formation of formaldehyde (or acetaldehyde) and an N-
hydropyridinyl radical. As described above, the oxidation
potentials of both the lutidinyl and the 3-chloropyridinyl
radicals are sufficiently low to be capable of reducing the N-
alkoxypyridinium cations used in this work, thus promoting
chain propagation. As shown in Figure 4, the energy barriers for
the hydrogen atom transfers are predicted to be relatively low
for the reactions of methoxyl and ethoxyl radicals with lutidine
shows that the orbital is localized between the nitrogen of the
lutidine and the methoxyl radical fragment; no occupation of
the π-orbitals for the lutidine is evident, as would be expected
during formation of the lutidinyl radical. The transition states
for reaction of ethoxyl radical with lutidine and 3-chloropyr-
idine showed similar results (see the Supporting Information).
The combined results are consistent with these hydrogen atom
transfers proceeding by a proton-coupled electron transfer
(PCET) mechanism,32 where partial proton transfer precedes
electron transfer in the net hydrogen atom transfer reaction. It
is worth noting that the valence bond curve-crossing model of
F
dx.doi.org/10.1021/jo301975j | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
a
Table 1. Calculated Stabilization Energies
a
Energies (in eV) of hydrogen-bonded complexes of pyridinyl radicals with the corresponding pyridine derivative relative to the separated species
ΔG(•) and the stabilization energies for the corresponding pyridinium derivatives ΔG(+). The increment ΔG(+) − ΔG(•) gives the difference in
oxidation potential of the complex versus that of the free radical, ΔEox.
Shaik and co-workers33 predicts that for reactions with a low-
lying proton transfer state, such as the reactions of alkoxyl
radicals with pyridines studied here, a PCET process is
expected to have a relatively low activation energy.
Table 1, DFT calculations support this hypothesis. The relative
reducing ability of the pyridinyl radicals versus their complexes
were determined by comparing the calculated binding free
energies of the pyridinyl radicals vs the corresponding
pyridinium cations with the pyridine bases. The difference in
the binding free energies is equal to the difference in the
oxidation potentials of the unbound versus bound radicals. For
the lutidine and 3-chloropyridine, the pyridinium cations were
found to be more strongly bound than the pyridinyl radicals by
0.152 and 0.171 eV, respectively, consistent with the base-
bound pyridinyl radicals being stronger reducing agents. These
predictions are consistent with the hypothesis that, at high base
concentration, pyridinyl radical-pyridine complexes are formed,
which are able to more rapidly reduce the N-alkoxypyridinium
cations and promote chain propagation.
7. Concluding Remarks. In the presence of pyridine bases,
photoinduced reductions of N-alkoxypyridinium salts lead to
highly efficient chain reactions resulting in N−O bond cleavage.
A mechanism is proposed wherein alkoxyl radical intermediates
react with the pyridine bases by proton-coupled electron
transfer (PCET) to generate pyridinyl radicals that are capable
of reducing the N-alkoxypyridinium salts, thereby propagating
the chain reaction. The chain nature of the reaction provided an
opportunity to determine rate constants for the key steps of the
chain by combining rate constant ratios derived from the
dependence of the reaction quantum yields on the reactant
concentrations with established rate constants for termination
reactions. Quantum yield data were best fit by a kinetic scheme
in which reduction of the alkoxypyridinium salts occurred by
pyridinyl radicals as well as by hydrogen-bonded complexes
with their corresponding pyridines. Consistent with the PCET
mechanism, the rate constant for the reaction of alkoxyl radicals
with pyridines was found to decrease with decreasing basicity of
the pyridine. Further support for the PCET mechanism was
derived from computational studies, which revealed transition
states for the hydrogen atom transfer reactions that resemble
proton transfer, and where electron transfer to the pyridine
base was found to occur along the reaction coordinate after the
transition state.
Additional calculations revealed that, following the initial
hydrogen atom transfer from the alkoxyl radicals to the pyridine
bases, a subsequent return hydrogen atom transfer can occur
from the pyridinyl radicals to the aldehydes to give the
corresponding pyridines and α-hydroxy alkyl radicals. As shown
in Figure 4, for the reactions of both pyridinyl radicals with
acetaldehye, the return hydrogen atom transfers are predicted
to have relatively large energy barriers (∼11 and ∼15 kcal/
mol). Thus these processes are not expected to compete with
diffusional separation of the pyridinyl radical/aldehyde
complexes that are formed after the initial hydrogen atom
transfer. For the reaction of methoxyl radical with lutidine,
however, the second hydrogen atom transfer is predicted to
have a significantly lower barrier (∼4 kcal/mol), presumably
due to the greater driving force of the reaction and, perhaps,
less steric hindrance in the transition state. Nonetheless, the
lutidinyl radical/formaldehyde pair is predicted to be bound by
only 1.6 kcal/mol (see Supporting Information), thus diffu-
sional separation is again predicted to proceed more rapidly
than the return hydrogen atom transfer.
In summary, the calculational evidence leads to the
prediction that the pyridinyl radicals are the principal chain-
propagating species that reduce the N-alkoxypyridinium cations
for all cases investigated here. We cannot rule out the
possibility, however, that for the N-methoxypyridinium
cation/lutidine reactions, a minor amount of hydroxymethyl
radical may be formed, which leads to chain-propagating
reduction of the pyridinium cation. Finally, we note that,
despite considerable effort, we were unable to locate transition
states wherein the pyridine bases catalyze isomerization of the
alkoxyl radicals to α-hydroxy alkyl radicals in a one-step
reaction.
As described above, a mechanistic model that included
reduction of the N-alkoxypyridinium cations only by the
pyridinyl radicals was unable to fit the experimental quantum
yield data with a consistent set of kinetic parameters for the
different pyridiniums over the full range of base concentrations.
The data suggested that, at high concentration, the pyridine
base played an additional role in promoting chain propagation.
We hypothesized that the pyridinyl radicals might react with
the pyridines to form hydrogen-bonded complexes that were
more strongly reducing than the pyridinyl radicals. As shown in
It is reasonable to expect that the reaction of alkoxyl radicals
with pyridine will be general for other compounds that are
sufficiently basic and capable of accepting a hydrogen atom
through a proton-coupled electron transfer mechanism.
Potential examples include other nitrogen heterocycles, imines,
azo compounds, etc. It is also intriguing to speculate that when
alkoxyl radicals are generated in the proximity of appropriate
G
dx.doi.org/10.1021/jo301975j | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
4-Phenyl-N-d3-methoxypyridinium Hexafluorophosphate. To a
stirred solution of 4-phenylpyridine-N-oxide (7.0 g, 40.46 mmol) in
acetonitrile (dry, 400 mL) was added methyl-d3 iodide (10 g, 68.9
mmol), and the reaction mixture was refluxed for 12 h. After cooling,
the reaction mixture was poured into rapidly stirred diethyl ether (700
mL) to obtain an off-white precipitate. The precipitate was filtered and
then dissolved in water (200 mL) followed by addition of a solution of
potassium hexafluorophosphate in water (7.0 g in 50 mL H2O) to
afford a white precipitate. The crude material was recrystallized twice
from hot methanol. 1H NMR (CD3CN) δ: 8.95−8.91 (m, 2H), 8.34−
8.30 (m, 2H), 7.93−7.89 (m, 2H), 7.71−7.63 (m, 3H). 13C NMR
(CD3CN) δ: 157.2, 141.5, 134.5, 133.5, 130.9, 129.2, 127.2, 118.4.
Instrumentation. Steady-state photolyses were carried out with an
Oriel 200 W medium-pressure Hg lamp. The excitation wavelength
(405 nm) was isolated by passing the lamp output through a Corning
5-58 bandpass filter immersed in water, followed by a 405 nm
interference filter. Generally, 3 mL of an acetonitrile-d3 solution
containing 2-chlorothioxanthone (CTX, 0.002 M), an alkoxypyridi-
nium (1a−1d), and a pyridine base (lutidine or 3-chloropyridine) in a
1 cm × 1 cm quartz cell was purged with a thin stream of argon for 3
min and then irradiated for 1−10 min to achieve ∼20% conversion.
Argon was continuously bubbled through the reaction mixture during
photolysis to purge as well as stir the solution.
DNA and RNA bases they may react in part by PCET resulting
in net hydrogen atom transfer to the bases. If so, this would
represent a nontraditional reaction of alkoxyl radicals in
biology, where hydrogen atom transfer to, not from, alkoxyl
radicals is the generally accepted paradigm.
EXPERIMENTAL SECTION
■
Materials. 4-Phenylpyridine-N-oxide, 4-cyanopyridine-N-oxide,
trimethyloxonium tetrafluoroborate, triethyloxonium tetrafluoroborate,
iodomethane, and iodomethane-d3 (>99% D) were obtained from
Aldrich and used as received. The N-alkoxypyridinium salts were
prepared by minor modification of literature procedures.10a,13a,34 2-
Chlorothioxanthone (Aldrich) was recrystallized from ethanol before
use. 2,6-Lutidine (Aldrich) and 3-chloropyridine (Aldrich) were
purified by passing through a short column of neutral alumina before
use. Acetonitrile-d3 (Cambridge Isotope, 99.6% D) was used as
received.
4-Phenyl-N-methoxypyridinium Hexafluorophosphate (1a). To a
stirred solution of 4-phenylpyridine-N-oxide (5.0 g, 29.24 mmol) in
acetonitrile (dry, 300 mL) was added methyl iodide (12 g, 84.5
mmol), and the reaction mixture was refluxed for 6 h. After cooling,
the reaction mixture was poured into rapidly stirred diethyl ether (700
mL) to obtain an off-white precipitate. The precipitate was filtered and
then dissolved in water (200 mL) followed by addition of a solution of
potassium hexafluorophosphate in water (7.0 g in 50 mL H2O) to
afford a white precipitate. The crude material was recrystallized twice
from hot methanol to give white crystals (8 g, 83%). 1H NMR
(CD3CN) δ: 8.95−8.90 (m, 2H), 8.34−8.30 (m, 2H), 7.93−7.89 (m,
2H), 7.71−7.63 (m, 3H), 4.40 (s,1H). 13C NMR (CD3CN) δ: 157.3,
141.5, 134.5, 130.9, 129.2, 127.2, 118.4, 70.9.
4-Phenyl-N-ethoxypyridinium Hexafluorophosphate (1b). To a
stirred solution of 4-phenylpyridine-N-oxide (5.0 g, 29.24 mmol) in
acetonitrile (dry, 300 mL) was added ethyl iodide (12 g, 76.9 mmol),
and the reaction mixture was refluxed for 12 h. After cooling, the
reaction mixture was poured into rapidly stirred diethyl ether (700
mL) to obtain an off-white precipitate. The precipitate was filtered and
then dissolved in water (200 mL) followed by addition of a solution of
potassium hexafluorophosphate in water (7.0 g in 50 mL H2O) to
afford a white precipitate. The crude material was recrystallized twice
from hot ethanol to obtain white crystals (9 g). 1H NMR (CD3CN) δ:
8.90−8.88 (m, 2H), 8.33−8.29 (m, 2H), 7.93−7.89 (m, 2H), 7.71−
7.63 (m, 3H), 4.66 (q, J = 6.96 Hz, 2H), 1.46 (t, J = 6.96 Hz, 3H). 13C
NMR (CD3CN) δ: 157.3, 142.2, 133.4, 130.9, 129.2, 127.2, 118.4,
81.4, 13.4.
The photon flux was determined by using the photocycloaddition
reaction of phenanthrenequinone to trans-stilbene in benzene as an
actinometer.35 The light intensity was typically 2−5 × 10−9 einstein
s−1. The quantum yield for product formation was determined from
the percent conversion and the light intensity.
1
Product Analysis. After photolysis, the H NMR spectrum of the
photolysate was recorded and the percent conversion of the starting
materials was determined by integration of diagnostic signals of the
pyridinium reactant and the pyridine product (see Supporting
Information). The mass balance was consistently high, >95%.
ASSOCIATED CONTENT
■
S
* Supporting Information
Triplet energy of CTX in acetonitrile; cyclic voltammograms;
fragmentation quantum yields versus pyridinium concentration;
derivation of eq 6; kinetic fitting parameters; energy diagrams
with selected structures for reactions of alkoxyl radicals with
pyridines; singly occupied molecular orbitals of the transition
states for reaction of ethoxyl radical with lutidine and 3-
chloropyridine; calculated energies, expectation values of S2
(<S2>), and geometries; Mulliken population analyses of
transition state structures; transition state imaginary frequencies
and the corresponding vibrational modes; NMR spectra. This
material is available free of charge via the Internet at http://
pubs.acs.org.
4-Cyano-N-methoxypyridinium Hexafluorophosphate (1c). To a
stirred solution of 4-cyanopyridine-N-oxide (3.0 g, 25 mmol) in
acetonitrile (dry, 100 mL) was added trimethyloxonium tetrafluor-
oborate (5g, 33.8 mmol), and the reaction mixture was refluxed for 12
h. After cooling, the reaction mixture was poured into rapidly stirred
diethyl ether (400 mL) to obtain a white precipitate that was filtered
and dried in air. The dried precipitate was dissolved in acetonitrile (50
mL) and then added to a solution of potassium hexafluorophosphate
(5.0 g in 100 mL H2O) to afford a white precipitate that was
AUTHOR INFORMATION
■
Corresponding Author
1
recrystallized twice from hot methanol. H NMR (CD3CN) δ: 9.15−
*E-mail: deepak.shukla@kodak.com; s.p.adiga@kodak.com;
jpd@chem.rochester.edu; farid@chem.rochester.edu.
9.12 (m, 2H), 8.47−8.44 (m, 2H), 4.45 (s, 3H). 13C NMR (CD3CN)
δ: 142.9, 133.8, 128.9, 118.4, 71.1.
Notes
4-Cyano-N-ethoxypyridinium Hexafluorophosphate (1d). To a
stirred solution of 4-cyanopyridine-N-oxide (3.0 g, 25 mmol) in
acetonitrile (dry, 100 mL) was added triethyloxonium tetrafluor-
oborate (5 g, 33.8 mmol), and the reaction mixture was refluxed for 12
h. The reaction mixture was concentrated by partial removal of solvent
(∼75 mL), and then excess of diethyl ether was added into the
reaction mixture to afford a pale yellow solid. The dried precipitate was
dissolved in a minimum amount of acetonitrile (∼25 mL) and then
added to a solution of potassium hexafluorophosphate (7.0 g in 100
mL H2O) to afford a white precipitate that was recrystallized twice
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
Research support was provided by a grant from the National
Science Foundation (CHE-1057615). We thank Mr. Kenny
Baptiste (Kodak) for cyclic voltammetry measurements and Dr.
Jerome Lenhard (Kodak) for electrochemistry discussions.
1
from hot ethanol. H NMR (CD3CN) δ: 9.13−9.09 (m, 2H), 8.47−
REFERENCES
(1) Walling, C. Pure Appl. Chem. 1967, 15, 69.
8.44 (m, 2H), 4.71 (q, J = 6.96 Hz, 2H), 1.47 (t, J = 6.96 Hz, 3H). 13C
NMR (CD3CN) δ: 143.9, 133.9, 128.5, 118.4, 82.2, 13.3.
■
H
dx.doi.org/10.1021/jo301975j | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
(2) (a) Pryor, W. A. In Free Radicals in Biology; Pryor, W. A., Ed.;
Academic Press: New York, 1976; Vol. I, pp 1−49. (b) Erben-Russ,
M.; Michel, C.; Bors, W.; Saran, M. J. Phys. Chem. 1987, 91, 2362.
(3) Orlando, J. J.; Tyndall, G. S.; Wallington, T. J. Chem. Rev. 2003,
103, 4657.
(16) (a) The intersystem crossing quantum yield for CTX of 0.98 in
acetonitrile was measured as described in ref 12a by triplet-triplet
energy transfer from 3CTX* to 1-cyanonaphthalene using the
corresponding reaction of benzophenone as an actinometer. This
value is likely to be more accurate than the recently reported16b
Φisc of
(4) (a) Richardson, W. H.; Yelvington, M. B.; Andrist, A. H.; Ertley,
E. W.; Smith, R. S.; Johnson, T. D. J. Org. Chem. 1973, 38, 4219.
(b) Walling, C.; Clark, R. T. J. Am. Chem. Soc. 1974, 96, 4530.
(c) Kochi, J. K. J. Am. Chem. Soc. 1962, 84, 1193. (d) Zhang, D.;
Zhang, R.; Allen, D. T. J. Chem. Phys. 2003, 118, 1794. (e) Beckwith,
A. L. J.; Hay, B. P. J. Am. Chem. Soc. 1989, 111, 230. (f) Bietti, M.;
Lanzalunga, O.; Salamone, M. J. Org. Chem. 2005, 70, 1417.
(5) (a) Paul, H.; Small, R. D., Jr.; Scaiano, J. C. J. Am. Chem. Soc.
1978, 100, 4520. (b) Das, P. K.; Encinas, M. V.; Steenken, S.; Scaiano,
J. C. J . Am. Chem. Soc. 1981, 103, 4162. (c) Malatesta, V.; Scaiano, J.
0.91, which unlike our approach uses actinometry that requires
accurate knowledge of extinction coefficients. We measured a
fluorescence quantum yield of 0.01 for CTX in acetonitrile, which is
in agreement with the value reported in ref 16b. Thus, within the
experimental error, there is practically no internal conversion from
singlet excited CTX. (b) Santiago, L. E. P.; Garcia, C.; Lhiaubet-Vallet,
V.; Miranda, M. A.; Oyola, R. Photochem. Photobiol. 2011, 87, 611.
(17) (a) Based on its phosphorescence spectrum in acetonitrile at
room temperature, the triplet energy, ET, of CTX is 62.1 kcal/mol
(2.69 eV); details are in the Supporting Information. No well-defined
oxidation wave of CTX could be determined by cyclic voltammetry.
An upper limit for the oxidation potential, Eox, however, can be
estimated from the singlet excitation energy, ES*, and the reduction
potential, Ered, based on the relationship ES* = Eox − Ered + C, where
the increment C, measured for many compounds, ranges between
∼0.1 and 0.4 eV.11a,17b−d From absorption and fluorescence spectra of
CTX in acetonitrile, ES* = 3.09 eV, which agrees with a literature
value.16b Reversible reduction of CTX was observed in acetonitrile,
yielding Ered of −1.53 V vs SCE. Thus, Eox is estimated to be between
1.2 and 1.5 V vs SCE. From eq 3, electron transfer from triplet CTX to
1a and 1b is expected to be exothermic by at least 0.1 eV. (b) Lenhard,
J. R. J. Imaging Sci. 1986, 30, 27. (c) Lenhard, J. R.; Cameron, A. D. J.
Phys. Chem. 1993, 97, 4916. (d) Loutfy, R. O.; Sharp, J. H. Photogr. Sci.
Eng. 1976, 20, 165.
́
C. J. Org. Chem. 1982, 47, 1455. (d) Baignee, A.; Howard, J. A.;
Scaiano, J. C.; Stewart, L. C. J. Am. Chem. Soc. 1983, 105, 6120.
(e) Nazran, A. S.; Griller, D. J. Am. Chem. Soc. 1983, 105, 1970.
(f) Finn, M.; Friedline, R.; Suleman, N. K.; Wohl, C. J.; Tanko, J. M. J.
Am. Chem. Soc. 2004, 126, 7578. (g) Salamone, M.; Giammarioli, I.;
Bietti, M. J. Org. Chem. 2011, 76, 4645. (h) Bietti, M.; Salamone, M.
Org. Lett. 2010, 12, 3654.
(6) (a) Elson, I. H.; Mao, S. W.; Kochi, J. K. J. Am. Chem. Soc. 1975,
97, 335. (b) Wong, P. C.; Griller, D. J. Org. Chem. 1981, 46, 2327.
(c) Wong, P. C.; Griller, D.; Scaiano, J. C. J. Am. Chem. Soc. 1982, 104,
5106. (d) Jones, M. J.; Moad, G.; Rizzardo, E.; Solomon, D. H. J. Org.
Chem. 1989, 54, 1607. (e) Nunes, P. M.; Estac
Agapito, F.; Santos, R. C.; Cabral, B. J. C.; Borges dos Santos, R. M.;
́
io, S. G.; Lopes, G. T.;
Martinho Simoes, J. A. J. Phys. Chem. A 2009, 113, 6524.
̃
(7) (a) Barton, D. H. R.; Beaton, J. M.; Geller, L. E.; Pechet, M. M. J.
Am. Chem. Soc. 1961, 83, 4076. (b) Niki, H.; Maker, P. D.; Savage, C.
M.; Breitenbach, L. P. J. Phys. Chem. 1981, 85, 2698. (c) Horner, J. H.;
Choi, S.-Y.; Newcomb, M. Org. Lett. 2000, 2, 3369. (d) Gilbert, B. C.;
Holmes, R. G. G.; Laue, H. A. H.; Norman, R. O. C. J. Chem. Soc.,
Perkin Trans. 2 1976, 1047. (e) Konya, K. G.; Paul, T.; Lin, S.; Lusztyk,
(18) Augustin-Nowacka, D.; Chmurzynski, L. Anal. Chim. Acta 1999,
̃
381, 215.
(19) (a) Oxidation potentials of α-hydroxy radicals are difficult to
determine reliably because of sensitivity to medium effects. For
example, hydrogen bonding to another alcohol molecule or water
could lower the oxidation potential due to coupled electron/proton
transfer.19b Accurately determined oxidation potentials of aryl-
substituted α-hydroxy radicals were found to be the same or similar
to those of the corresponding alkoxyl radical.19c,d Thus, the oxidation
́
J.; Ingold, K. U. J. Am. Chem. Soc. 2000, 122, 7518. (f) Fernandez-
Ramos, A.; Zgierski, M. Z. J. Phys. Chem. A 2002, 106, 10578.
(g) Elford, P. E.; Roberts, B. P. J. Chem. Soc., Perkin Trans. 2 1996,
2247. (h) Buszek, R. J.; Sinha, A.; Francisco, J. S. J. Am. Chem. Soc.
2011, 133, 2013. (i) Ingold, K. U.; Smeu, M.; DiLabio, G. A. J. Org.
Chem. 2006, 71, 9906. (j) DiLabio, G. A.; Ingold, K. U.; Lin, S.;
Litwinienko, G.; Mozenson, O.; Mulder, P.; Tidwell, T. T. Angew.
Chem., Int. Ed. 2010, 49, 5982. (k) Salamone, M.; Bietti, M.; Calcagni,
A.; Gente, G. Org. Lett. 2009, 11, 2453.
•
potentials of CH2 OMe and MeCH•OEt, −0.24 and −0.45 V vs SCE,
respectively,19e can be used as reasonable estimates for those of the
corresponding alcohols. Consistent with these values, the oxidation
potential of Me2C•OH is −0.61,19c i.e., there are similar increments
between the primary, secondary, and tertiary radicals. (b) Lilie, V. J.;
Beck, G.; Henglein, A. Ber. Bunsen-Ges. Phys. Chem. 1971, 75, 458.
(c) Lund, T.; Wayner, D. D. M.; Jonsson, M.; Larsen, A. G.; Daasbjerg,
K. J. Am. Chem. Soc. 2001, 123, 12590. (d) Workentin, M. S.; Wayner,
D. D. M. Res. Chem. Intermed. 1993, 19, 777. (e) Wayner, D. D. M.;
McPhee, D. J.; Griller, D. J. Am. Chem. Soc. 1988, 110, 132.
(20) Shukla, D; Young, R. H.; Farid, S. J. Phys. Chem. A 2004, 108,
10386.
(8) Dix, T. A.; Aikens, J. Chem. Res. Toxicol. 1993, 6, 2.
(9) (a) Saebø, S.; Radom, L.; Schaefer, H. F., III J. Chem. Phys. 1983,
78, 845. (b) Colwell, S. M. Theor. Chim. Acta 1988, 74, 123.
(c) Adams, G. F.; Bartlett, R. J.; Purvis, G. D. Chem. Phys. Lett. 1982,
87, 311. (d) Curtiss, L. A.; Kock, L. D.; Pople, J. A. J. Chem. Phys.
1991, 95, 4040.
(21) The ratios c = (k′4/k4) and f = ((k′6/k6) are not expected to be
the same for all pyridiniums, as these compounds may have varying
amounts of minor impurities. We allowed for such differences to
obtain the best fit. It should be noted, however, that these ratios are
quite small (∼0 to 0.01), but whenever the quantum yields are very
large, they become very sensitive to any minor differences in
termination reactions. To minimize the number of variables, however,
both ratios were kept constant for each pyridinium.
(10) (a) Lorance, E. D.; Kramer, W. H.; Gould, I. R. J. Am. Chem. Soc.
2002, 124, 15225. (b) Lorance, E. D.; Kramer, W. H.; Gould, I. R. J.
Am. Chem. Soc. 2004, 126, 14071. (c) Lorance, E. D.; Hendrickson, K.;
Gould, I. R. J. Org. Chem. 2005, 70, 2014.
(11) (a) Gould, I. R.; Shukla, D.; Giesen, D.; Farid, S. Helv. Chim.
Acta 2001, 84, 2796. (b) Williams, J. L. R.; Specht, D. P.; Farid, S.
Polym. Eng. Sci. 1983, 23, 1022.
(12) (a) Shukla, D.; Ahearn, W. G.; Farid, S. J. Org. Chem. 2005, 70,
6809. (b) Shukla, D.; Ahearn, W. G.; Farid, S. Photochem. Photobiol.
2006, 82, 146.
(13) (a) Lee, K. Y.; Kochi, J. K. J. Chem. Soc., Perkin Trans. 2 1992,
1011. (b) Howell, J. O.; Goncalves, J. M.; Amatore, C.; Klasinc, L.;
Wightman, R. M.; Kochi, J. K. J. Am. Chem. Soc. 1984, 106, 3968.
(14) (a) Farid, S.; Dinnocenzo, J. P.; Merkel, P. B.; Young, R. H.;
Shukla, D. J. Am. Chem. Soc. 2011, 133, 4791. (b) Farid, S.;
Dinnocenzo, J. P.; Merkel, P. B.; Young, R. H.; Shukla, D.; Guirado, G.
J. Am. Chem. Soc. 2011, 133, 11580.
(22) The rate constant k2 is a composite of deuterium abstraction
from CD3CN and reaction with the donor sensitizer (at 0.002 M).12a
(23) (a) The O−H BDE in tert-butanol is 106.3 kcal/mol, which is
1.7 kcal/mol larger than that in methanol or ethanol (104.6 kcal/
mol).23b DFT calculation yields a 1.3 kcal/mol larger BDE (in
acetonitrile) for the methyl C−H bond in lutidine than in toluene.
(b) Luo, Y.-R. Handbook of Bond Dissociation Energies in Organic
Compounds; CRC Press: Boca Raton, 2003.
(24) (a) The pKa of CH3O• can be estimated from that of •CH2OH
and the free energy difference between the two radicals. The pKa of
•CH2OH in acetonitrile is estimated to be ∼30.19c The free energy
(15) Arnold, B. R; Farid, S.; Goodman, J. L.; Gould, I. R. J. Am.
Chem. Soc. 1996, 118, 5482.
I
dx.doi.org/10.1021/jo301975j | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
difference between CH3O• and •CH2OH can be estimated from their
heats of formation (4.1 and −2 kcal/mol, respectively)24b and the
statistical factor of 3 that favors •CH2OH in the equilibrium. Based on
these thermodynamic quantities, the pKa of CH3O• in acetonitrile is
estimated to be ∼25. (b) Tsang, W. Heats of Formation of Organic
Free Radicals by Kinetic Methods. In Energetics of Organic Free
Radicals; Martinho Simoes, J. A., Greenberg, A., Liebman, J. F., Eds.;
̃
Blackie Academic and Professional: London, 1996; pp 22−58.
(25) (a) At moderate exergonicity bimolecular electron transfer
requires contact interaction. With increasing driving force, however,
long distance electron transfer leading directly to solvent-separated
radical ion pairs dominates.25b As a result, the rate constant
increases,25c and in the case of charge shift reactions it reaches 1.2
× 1010 M−1 s−1.25d (b) Gould, I. R.; Young, R. H.; Moody, R. E.; Farid,
S. J. Phys. Chem. 1991, 95, 2068. (c) Rosspeintner, A.; Kattnig, D. R.;
Angulo, G.; Landgraf, S.; Grampp, G. Chem.Eur. J. 2008, 14, 6213.
(d) Luo, P.; Dinnocenzo, J. P.; Merkel, P. B.; Young, R. H.; Farid, S. J.
Org. Chem. 2012, 77, 1632.
(26) Because of uncertainty about the exact oxidation potential of the
lutidinyl radical, the rate constant for electron transfer from BH• to
CTX (ket) might be less than the diffusion-controlled limit (kdiff ≈ 1 ×
1010 M−1 s−1). If ket ≈ kdiff/10, it would correspond to the measured
pseudo first-order rate constant k5. If ket is between kdiff/10 and kdiff, it
would not necessarily contradict the experimental data: CTX•− could
propagate the chain by reducing the pyridinium and only partially lead
to termination, e.g., via proton transfer from trace amounts of water.
(27) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (b) Stephens, P.
J.; Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J. J. Phys. Chem. 1994,
98, 11623.
(28) Miertus,
̌
S.; Scrocco, E.; Tomasi, J. Chem. Phys. 1981, 55, 117.
(29) Rappe, A. K.; Casewit, C. J.; Colwell, K. S.; Goddard, W. A., III;
́
Skiff, W. M. J. Am. Chem. Soc. 1992, 114, 10024.
(30) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.;
Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.;
Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.;
Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.;
Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao,
O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J.
B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R.
E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.;
Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J.
J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.;
Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman,
J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.;
Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.;
Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.;
Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen,
W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03, Revision
C.02; Gaussian, Inc.: Wallingford, CT, 2004.
(31) (a) Analogous complexes of this type have been recently
proposed in the reaction of benzyloxy radicals with aliphatic
amines.31b−d (b) Salamone, M.; DiLabio, G. A.; Bietti, M. J. Am.
Chem. Soc. 2011, 133, 16625. (c) Salamone, M.; DiLabio, G. A.; Bietti,
M. J. Org. Chem. 2011, 76, 6264. (d) Salamone, M.; Anastasi, G.;
Bietti, M.; DiLabio, G. A. Org. Lett. 2011, 13, 260.
(32) See: Hammes-Schiffer, S.; Stuchebrukhov, A. A. Chem. Rev.
2010, 110, 6939 and references therein.
(33) Lai, W.; Li, C.; Chen., H.; Shaik, S. Angew. Chem., Int. Ed. 2012,
51, 5556.
(34) Katz, R. B.; Mistry, J.; Mitchell, M. B. Synth. Commun. 1989, 19,
317.
(35) Brown-Wensley, K. A.; Mattes, S. L.; Farid, S. J. Am. Chem. Soc.
1978, 100, 4162.
J
dx.doi.org/10.1021/jo301975j | J. Org. Chem. XXXX, XXX, XXX−XXX