Reduction of diesters of 1,2-diols. Regioselective C-O bond cleavage of the anionic forms

Article abstract of DOI:10.1021/jo060535w

The electrochemical reduction of benzoate diesters of glycols has been studied in acetonitrile and N,N-dimethylformamide as solvents. The reductions occur in two closely spaced one-electron steps, and it was found that the dianion diradicals decompose by one of two routes, depending on the substituents on the ethylene moiety: cleavage of two benzoates to produce alkene or formation of benzil by way of a postulated cyclic intermediate to produce also the dianion of the diol. These correspond to cleavage of the R-OC(O)Ar bonds and the RO-C(O)Ar bonds, respectively. When the radical formed by the former cleavage is a primary or secondary radical, the reaction is too slow to compete with the latter reaction that produces benzil. However, when that radical is either tertiary or benzylic, the former cleavage reaction is fast and no benzil is detected. The dianions of p-cyano- and p-nitrobenzoate esters are rather stable on the voltammetric time scale. However, the addition of lithium ions results in detectable formation of 4,4′-dicyanobenzil from four different p-cyanobenzoate diesters.

Full text of DOI:10.1021/jo060535w

Reduction of Diesters of 1,2-Diols. Regioselective C-O Bond
Cleavage of the Anionic Forms
Norma A. Mac´ıas-Ruvalcaba, Cheryl L. Moy, Zi-Rong Zheng,† and Dennis H. Evans*
Department of Chemistry, UniVersity of Arizona, Tucson, Arizona 85721
dheVans@email.arizona.edu
ReceiVed March 10, 2006
The electrochemical reduction of benzoate diesters of glycols has been studied in acetonitrile and N,N-
dimethylformamide as solvents. The reductions occur in two closely spaced one-electron steps, and it
was found that the dianion diradicals decompose by one of two routes, depending on the substituents on
the ethylene moiety: cleavage of two benzoates to produce alkene or formation of benzil by way of a
postulated cyclic intermediate to produce also the dianion of the diol. These correspond to cleavage of
the R-OC(O)Ar bonds and the RO-C(O)Ar bonds, respectively. When the radical formed by the former
cleavage is a primary or secondary radical, the reaction is too slow to compete with the latter reaction
that produces benzil. However, when that radical is either tertiary or benzylic, the former cleavage reaction
is fast and no benzil is detected. The dianions of p-cyano- and p-nitrobenzoate esters are rather stable on
the voltammetric time scale. However, the addition of lithium ions results in detectable formation of
4,4′-dicyanobenzil from four different p-cyanobenzoate diesters.
Introduction
the addition of magnesium triflate to tetrabutylammonium
tetrafluoroborate electrolyte in dimethylformamide.4 In that
work, it was proposed that ion pairing with magnesium favors
dimerization of the anion radicals of methyl benzoate to produce
a sigma-bonded dimeric dianion that decomposes to benzil and
methoxide.4 Coupling is also observed in the electrochemical
reduction of phenyl benzoate.5
Less attention has been directed to the electrochemical
reduction of diesters. Kashimura et al.6 examined the prepara-
tive-scale reductions of symmetrical and asymmetrical diesters
of ethylene glycol in tetrahydrofuran with LiClO4 electrolyte.
Good yields of symmetrical and unsymmetrical 1,2-diketones
were obtained, including a yield of 71% benzil from compound
1 of the present paper. The objective of the present work was
to study the mechanism of the decomposition reactions of the
anion radicals and dianions formed by reduction of some
aromatic diesters of 1,2-diols, 1-14 (Chart 1).
Esters of aromatic acids are sufficiently good electron
acceptors that formation of anion radicals by electrochemical
reduction, though difficult, is achievable within the potential
window available in various dipolar aprotic solvents. The anion
radicals undergo decomposition by two general paths, cleavage
-
to give R• and ArCO2 and coupling to produce 1,2-diketones
and RO- (Scheme 1). For alkyl benzoates in the presence of
SCHEME 1
tetraalkylammonium counterions, the anion radicals generally
cleave to form the alkyl radical and benzoate. The rates of these
reactions are similar for the formation of primary and secondary
alkyl radicals, larger for tertiary radicals, and largest for benzylic
radicals.1,2 The coupling reaction (right side of Scheme 1) is an
electrochemical analogue of the acyloin condensation.3 It has
been shown that coupling is promoted for alkyl benzoates by
(3) Carey, F. A.; Sundberg, R. J. AdVanced Organic Chemistry, 4th
edition; Kluwer Academic/Plenum: New York, 2001; Part B, pp 305-
306.
(4) Pletcher, D.; Slavin, L. J. Chem. Soc., Perkin Trans. 2 1995, 2005-
2012.
(5) Seeber, R.; Magno, F.; Bontempelli, G.; Mazzocchin, G. A. J.
Electroanal. Chem. 1976, 72, 219-228.
(6) Kashimura, S.; Murai, Y.; Washika, C.; Yoshihara, D.; Kataoka, Y.;
Murase, H.; Shono, T. Tetrahedron Lett. 1997, 38, 6717-6720.
† Present affiliation: MedImmune, Inc., Gaithersburg, MD.
(1) Wagenknecht, J. H.; Goodin, R. D.; Kinlen, P. J.; Woodard, F. E. J.
Electrochem. Soc. 1984, 131, 1559-1565.
(2) Masnovi, J. J. Am. Chem. Soc. 1989, 111, 9081-9089.
10.1021/jo060535w CCC: $33.50 © 2006 American Chemical Society
Published on Web 06/01/2006
J. Org. Chem. 2006, 71, 4829-4834
4829

Mac´ıas-Ruvalcaba et al.
CHART 1
whose transition state resembles 15, but we will discuss our
results in terms of intermediates such as 15. The very basic
ethylene glycol dianion, 17, is expected to be protonated by
constituents of the medium.
It is possible that the anion radical also decomposes. Such
reactions are difficult to detect because one cannot generate the
anion radical exclusively, owing to the disproportionation equi-
librium that exists in solution near the electrode, 2 DiEst•-
h
DiEst + DiEst2-, where DiEst indicates the diester species.
However, when the potential was reversed about halfway up
the cathodic peak (where the amount of dianion will be smaller
than when the scan is extended through the peak), the peaks
for the oxidation of benzil anions were almost absent, supporting
the notion that the reactions of the anion radical are either slow
or do not produce benzil.
Results and Discussion
Qualitative Studies in Acetonitrile. All compounds were
investigated in acetonitrile to determine the general features of
the electrode reactions. A voltammogram for 1 at 3.00 V/s is
shown in Figure 1. The height of the cathodic peak corresponds
Compounds 2 and 3 feature strong electron-withdrawing
groups that have two effects: the compounds are much more
easily reduced than 1 and the dianion diradicals are stabilized
with respect to the cleavage reaction. No benzil is detected in
the voltammograms. Each features a two-electron reduction peak
composed of two closely spaced one-electron reactions, and the
size of the oxidation peak is consistent with the anion radicals
and dianions being stable on the voltammetric time scale.
Compounds 4 and 5 behave similarly to 1 in that the anodic
peak associated with the reduction peak is small, indicating
decomposition of the dianion diradical, and benzil is detected
on the return sweep. In the case of 5 (from the dl pair of the
diol), in which both methyl groups can occupy equatorial sites
in the postulated cyclic intermediate (18), the size of the peaks
for the oxidation of benzil dianion and the loss of reversibility
of the main reduction peak suggest that the rate of decomposition
is comparable to that of the dianion diradical of 1. However,
with 4 (from the meso diol), one methyl group must be axial in
the cyclic intermediate (19), and in this case, it is observed that
the peaks for the oxidation of benzil dianion are smaller than
those observed for compounds 1 and 5, suggesting that the rate
of decomposition of the dianion of 4 is smaller than that of 1
and 5. This finding is consistent with a less favorable intermedi-
ate for 4 as a result of the 1,3-diaxial interaction between methyl
and oxygen.
FIGURE 1. Cyclic voltammogram of 2.08 mM 1 in acetonitrile with
0.10 M Bu4NPF6. Glassy carbon working electrode. 3.00 V/s.
to an overall four-electron process: two-electron reduction of
neutral 1 to a dianion diradical, followed by two-electron
reduction of benzil that is formed in a subsequent cleavage
reaction (see below). The associated anodic peak on the return
scan is very small, indicating rapid decomposition of the dianion.
Two oxidation peaks are seen near -1.9 and -1.6 V, and, as
was demonstrated in separate voltammetric studies of benzil
(16), they correspond to the stepwise oxidation of the dianion
of benzil to its anion radical and neutral forms, respectively.
We propose that the decomposition of the dianion diradical
occurs through cyclic intermediate 15. That this is a reasonable
intermediate is supported by the existence of the O-alkylated
forms (methyl and ethyl) of 15 as the trans isomer shown.7 Also,
in the preparative electrochemical reduction of diesters, com-
pounds were isolated that were thought to be magnesium salts
of dianions analogous to 15.6 The trans isomer is favored by
having the negative charges as far apart as possible, having the
phenyl groups in equatorial positions and possibly by an
anomeric effect of the ring oxygens that could stabilize oxygens
in axial positions. We cannot rule out the possibility that the
formation of benzil is a concerted double cleavage reaction
Analogous behavior is exhibited by 7 and 8 where the amount
of benzil formed is larger with 7 compared to 8. The dianion
of 7 would require a trans-decalin-type intermediate (20), which
is certainly much more favorable than the cis-decalin-type
intermediate (21) from 8. Accordingly, the decomposition rate
of the dianion diradicals of 7 is much greater than that of 8.
In compound 13, derived from 2,3-dimethyl-2,3-butanediol,
a cyclic intermediate would require having two axial methyl
groups that would strongly disfavor this reaction path. For 13,
(7) Calo`, V.; Lopez, L. Synthesis 1984, 774-776.
4830 J. Org. Chem., Vol. 71, No. 13, 2006

Reduction of Diesters of 1,2-Diols
Compound 14, derived from dimethyl R,R-tartrate, behaved
similarly to the hydrobenzoin derivatives, 10 and 11, in the sense
that cleavage of the dianion diradical produced the correspond-
ing olefin and benzoate. This diester was studied in DMF. The
olefin, dimethyl fumarate, was detected on the return sweep of
the voltammogram by the oxidation of its anion radical at about
-1.8 V. The reduction of 14 was completely irreversible,
attesting to the very rapid decomposition of the dianion diradical.
The peak height corresponded to approximately four electrons
per molecule of 14. Dimethyl fumarate, as shown by separate
voltammetric experiments, is reduced to the anion radical near
-1.8 V, and the second step of reduction coincides almost
exactly with the reduction peak for 14. Quantitative interpreta-
tion of the voltammetry was hindered by the complex electro-
chemical behavior of dimethyl fumarate whose anion radical
undergoes dimerization.9
a two-electron reduction peak was observed, and oxidation peaks
for the anions of benzil were completely absent. The dianion
diradical is not stable and does decompose, as evidenced by
the lack of an anodic peak for its oxidation back to the neutral
diester at small scan rates. However, when the scan rate was
increased to 20 V/s, the anodic peak grew in, suggesting that
the decomposition can be easily outrun. In this case, initial
cleavage of benzoate is favored because a tertiary alkyl radical
is formed (Scheme 2). The cleavage of tertiary radicals from
Finally, a qualitative investigation of 1,3-propanediol diben-
zoate, 24, showed two closely spaced reversible one-electron
reduction steps. Both benzoate ester functions are reduced almost
independently, and the resulting dianion diradical has a stability
comparable to the anion radical of other primary alkyl benzoates
such as methyl benzoate.4 Benzil was not detected on the
voltammetric time scale. A seven-membered ring intermediate
would be required for the formation of benzil.
SCHEME 2
the anion radicals of monoesters is over 100 times faster than
cleavage to give primary and secondary radicals.1 The first
cleavage is probably followed by the rapid loss of the second
benzoate, producing the alkene, 2,3-dimethyl-2-butene. A
qualitative interpretation is that formation of benzil via a cyclic
intermediate is not possible for the dianion diradical of 13, but
the dianion diradical can nonetheless react quickly, owing to
the formation of a tertiary radical upon loss of benzoate.
Compounds 10 and 11 are derived from the diastereomers
of hydrobenzoin. As with 13, rapid decomposition of the dianion
diradical (22; Scheme 3) occurs, but no benzil is detected.
Quantitative Studies in N,N-Dimethylformamide. The
decomposition reactions of some of the diester dianion diradicals
are rather rapid in acetonitrile, meaning that relatively large scan
rates would be required for the quantitative evaluation of
reaction rates and the overall mechanism. The rates are
significantly lower in N,N-dimethylformamide (DMF), so this
solvent was selected for quantitative studies. Figure 2 shows a
SCHEME 3
However, in this case, the product formed by cleavage of two
benzoates from the dianion diradical, stilbene (23), can be
detected in the voltammetry through its reversible reduction to
the anion radical, which is slightly negative of the main
reduction peak. There is also a second irreversible reduction of
stilbene near the solvent/electrolyte breakdown potential. In this
case, cleavage of the first benzoate leaves behind a benzylic
radical, which would rapidly expel the second benzoate to
produce stilbene. That this overall reaction is rapid is consistent
with the observation2 that cleavages producing benzylic radicals
from anion radicals of alkyl benzoates have the largest rates
measured. Voltammetry does not allow the determination of
the isomeric identity of the stilbene that is formed, because the
reduction potentials of cis- and trans-stilbene are almost
identical.8
FIGURE 2. Cyclic voltammogram of 3.02 mM 1 in DMF with 0.10
M Bu4NPF6. Glassy carbon working electrode. 0.30 V/s. Symbols:
simulation. Curve: background-corrected voltammogram.
voltammogram of 3.02 mM 1, obtained at a glassy carbon
electrode at 0.30 V/s. The fact that the cathodic peak at -2.8
V is a composite of two closely spaced reductions is quite
evident in Figure 2, as is the small anodic peak near -2.6 V.
The peaks for the stepwise oxidation of the benzil dianion appear
at -1.9 (shoulder) and -1.56 V. The very broad and low plateau
at -1.9 V for the oxidation of the benzil dianion was also found
(8) Sioda, R. E.; Cowan, D. O.; Koski, W. S. J. Am. Chem. Soc. 1967,
89, 230-235.
(9) Doherty, A. P.; Scott, K. J. Electroanal. Chem. 1998, 442, 35-40.
J. Org. Chem, Vol. 71, No. 13, 2006 4831

Mac´ıas-Ruvalcaba et al.
TABLE 1. Simulation Parameters for 1, 4, 5, 7, and 8a
parameter
reaction
(unit)
1
4
5
7
8
DiEst + e- h DiEst•-
DiEst•- + e- h DiEst2-
DiEst2- f Bz
E°1 (V)
E°2 (V)
kf,5 (s-1
-2.632 ( 0.006
-2.737 ( 0.007
2.2 ( 0.1
-2.645
-2.756
0.5
-2.654
-2.762
1.3
-2.658 ( 0.002
-2.780 ( 0.002
16 ( 1
-2.645
-2.754
130
)
a Results for 1 are the averages of four data sets obtained at four different concentrations, 0.97, 1.95, 2.08, and 3.02 mM. For 7, two data sets were
obtained at 1.83 and 1.96 mM.
in studies of benzil alone, as was the small peak near -1.2 V.
Also shown in Figure 2 is a simulation (symbols) of the
voltammogram. The reaction scheme used for the simulation is
given by reactions 1-11, along with the equilibrium and kinetic
parameters associated with each reaction.
tion energies, their rate constants were made large. These
reactions produce a state of redox equilibrium in the solution
in the diffusion layer. It is probably not necessary to include
all of them, as only two or three should be sufficient to bring
about equilibrium. We did not attempt to remove this redun-
dancy. It should be mentioned that reasonable fits of simulations
to experiment could be obtained with only reactions 1-5.
However, inclusion of the solution-phase electron-transfer
reactions improved the agreement noticeably.
DiEst + e- h DiEst•-
DiEst•- + e- h DiEst2-
Bz + e- h Bz•-
E°1, ks,1, R1
E°2, ks,2, R2
(1)
(2)
(3)
(4)
(5)
(6)
The adequacy of this reaction scheme was tested thoroughly
with 1. Experiments were performed at four concentrations
between 1 and 3 mM and at scan rates from 0.10 to 10.0 V/s.
At each concentration, the fits were optimized by varying all
simulation parameters to find the best agreement with experi-
ment for all scan rates. The simulation shown in Figure 2 used
this optimized parameter set. All simulation parameters are
tabulated in the Supporting Information.
E°3, ks,3, R3
Bz•- + e- h Bz2-
E°4, ks,4, R4
DiEst2- f Bz
kf,5
kf,6
Bz2- f product
The most important parameter controlling the various peak
heights in the simulated voltammograms is kf,5. The average
value found for the four concentrations was 2.2 ( 0.1 s-1. When
2 DiEst•- h DiEst2- + DiEst
Bz + DiEst2- h Bz•- + DiEst•-
Bz•- + DiEst2- h Bz2- + DiEst•-
Bz + DiEst•- h Bz•- + DiEst
Bz•- + DiEst•- h Bz2- + DiEst
K7, kf,7, kb,7 (7)
K8, kf,8, kb,8 (8)
k
f,5 is set equal to zero, the cathodic peak is too small and there
are no anodic peaks associated with benzil. When the value of
f,5 is too large, the cathodic peak exceeds the experimental value
k
and the benzil peaks are too large. The reason that the cathodic
peak grows as kf,5 is increased is that reaction 5 produces benzil,
which is more easily reduced than the diester, so it is reduced
at the potential of the diester reduction, giving more current.
The fact that the growth of the cathodic peak and the growth of
the peaks for benzil is well accounted for by the simulations is
good support for the stoichiometry inherent in the reaction
scheme. This suggests that alternate reactions, such as cleavage
to give benzoate and ethylene, are unimportant. Also, the fact
that nearly identical parameters provide fits for concentrations
from 1 to 3 mM, while assuming that reaction 5 is first order,
argues against any second-order reactions playing an important
role.
K9, kf,9, kb,9 (9)
K10, kf,10, kb,10 (10)
K11, kf,11, kb,11 (11)
Here DiEst denotes the various diester species and Bz denotes
benzil. DiEst2- is the dianion diradical. E°j, ks,j and Rj are the
standard potential, the standard electron-transfer rate constant,
and the electron-transfer coefficient of electrode reaction j. Kk,
kf,k and kb,k are the equilibrium constant, the forward rate
constant, and the backward rate constant of chemical reaction
k. Reactions 1-4 are the electrode reactions of the diester and
benzil. The key chemical step is reaction 5, which produces
benzil from the dianion diradical of the diester. It is considered
to be completely irreversible. As indicated earlier, the reaction
may proceed through a cyclic intermediate dianion, but that
detail is omitted and reaction 5 is simply written as an
irreversible reaction that is first-order in the dianion diradical.
The other product, the ethylene glycol dianion, was not included
in the simulation. Reaction 6 results in an irreversible loss of
some of the dianion of benzil. Inclusion of a small value for
kf,6 improved the fits of simulation to the data. Studies of the
reduction of benzil itself confirmed that the dianion is not
perfectly stable under the reaction conditions.
The standard potentials of reactions 1-2 are well defined by
the simulation, and the values are summarized in Table 1 along
with kf,5. (Tables of all simulation parameters are included in
the Supporting Information). Data for other compounds that
produce benzil on reduction (4, 5, 7, and 8) are also included
in Table 1, and examples of fits of simulation to experimental
voltammograms for 4 and 7 are presented in Figures 3 and 4.
It is interesting to note in Table 1 that the two standard potentials
are close to one another, E°1 - E°2 ranging from 0.10 to 0.12
V. This indicates that the two identical ester functions are
sufficiently separated to be reduced almost independently of
one another. The values of the potentials are remarkably constant
among the five compounds, indicating that the nature of the
substituents on the ethylene group has little effect on the electron
affinity of the ester groups. Also note that the relative values
of kf,5 are consistent with the qualitative results in acetonitrile
that were presented and discussed earlier.
Reactions 7-11 are electron-transfer reactions occurring
among the six solution-phase redox species. The equilibrium
constants of these reactions are dictated by the values of the
standard potentials used for reactions 1-4. As these are simple
electron-transfer reactions that do not involve large reorganiza-
4832 J. Org. Chem., Vol. 71, No. 13, 2006

Reduction of Diesters of 1,2-Diols
FIGURE 5. Cyclic voltammogram of 1.91 mM 10 at 0.30 V/s. Other
conditions as in Figure 2.
FIGURE 3. Cyclic voltammogram of 2.06 mM 4 at 0.50 V/s. Other
conditions as in Figure 2.
TABLE 2. Standard Potentials from Simulations of
Voltammograms of 10 and Stilbene
parameter
reaction
(unit)
10
stilbene
DiEst + e- h DiEst•-
DiEst•- + e- h DiEst2-
Stb + e- h Stb•-
E°1 (V)
E°2 (V)
E°3 (V)
-2.638
-2.589
-2.654
-2.647
stoichiometry (1:1, dianion/Bz or dianion/Stb), and both Bz and
Stb take part in stepwise two-electron reductions.
In Figure 5, the one-electron reduction of stilbene follows
shortly after the two-electron reduction of diester 10. The small
anodic peak seen at about -2.6 V is due to the oxidation of the
anion radical of stilbene. The second irreversible reduction of
stilbene appears at about -3.1 V. The fact that the height of
this peak is well accounted for by the simulation is good
evidence that the decomposition of the dianion proceeds
exclusively to stilbene. Consistent with this conclusion is the
observation that no peaks for benzil are seen. The rate constant
for the decomposition of the dianion, kf,5, is very large, with a
lower limit of about 1000 s-1. We did not attempt to evaluate
this overall rate constant. It should be noted that the rate constant
for cleavage of benzylic radicals from anion radicals of alkyl
benzoates is of the order of 105 s-1 in ethanol.2
The same set of parameters was used to fit scan rates from
0.10 to 10.0 V/s and these are listed in the Supporting
Information. The standard potentials for the two one-electron
reductions of 10 along with that of the first step of reduction of
stilbene are given in Table 2. Included are the results for stilbene
itself. The values of E°1 and E°2 for 10 are very similar,
indicating a very weak interaction between the two ester groups
being reduced.
Cyanobenzoates 2, 6, 9, and 12 were also studied in DMF.
As indicated earlier, the dianion diradicals are quite stable, and
the voltammograms could be readily simulated including only
reactions 1 and 2 with kf,5 ) 0, except for the lowest scan rates,
where some loss of the dianion was indicated. Here we
examined the effect of added metal ions. It was found that
replacement of 0.10 M tetrabutylammonium ion by 0.10 M
sodium ion in the electrolyte had no effect. However, small
amounts of lithium ion, added as lithium perchlorate, resulted
in the formation of 4,4′-dicyanobenzil, and kf,5 could be
evaluated. Selected simulation parameters are included in Table
3, and a listing of all parameters is in the Supporting Informa-
tion.
FIGURE 4. Cyclic voltammogram of 2.03 mM 7 at 1.00 V/s. Other
conditions as in Figure 2.
For these compounds (1, 4, 5, 7, and 8), the competing
reaction of cleavage of two benzoates leaving the alkene is too
slow to affect the results. As shown earlier,1,2 cleavage of
benzoate to produce primary or secondary radicals is slow and
occurs at about the same rate. Cleavage to produce tertiary
radicals is faster, and the production of benzylic radicals is still
faster. For the present compounds, cleavage of the first benzoate
from the dianion diradical would give a primary (1) or secondary
(4, 5, 8, and 9) radical, a reaction which is apparently too slow
to compete with the formation of benzil. Earlier, it was observed
that the dianion diradical of 13 reacted rapidly and did not form
any detectable benzil. For this compound, cleavage of the first
benzoate leaves a tertiary radical, and this reaction is fast enough
to compete successfully with benzil formation, which, in
addition, would require a congested cyclic intermediate.
Let us turn now to compounds 10 and 11, which form stilbene
upon reduction. Here, cleavage of the first benzoate from the
dianion diradical produced a benzylic radical, and this reaction
is fast enough to compete successfully with the formation of
benzil, which again would require a congested cyclic intermedi-
ate. The stilbene that is formed is reversibly reduced to its anion
radical at potentials slightly negative of the reduction peak of
the diester. A peak corresponding to the reduction of the stilbene
anion radical is also observed. Figure 5 shows a voltammogram
for 1.91 mM 10 at 0.30 V/s along with a simulation according
to reactions 1-11. In these reactions, the benzil species (Bz,
Bz•-, and Bz2-) were replaced by the corresponding stilbene
species (Stb, Stb•-, and Stb2-). No other changes were neces-
sary, as each of the two different types of cleavage reactions
produces electroactive products, Bz or Stb, with the same
As mentioned earlier, the cyanobenzoates are about 0.6 V
more easily reduced than the corresponding benzoates. The
J. Org. Chem, Vol. 71, No. 13, 2006 4833

Mac´ıas-Ruvalcaba et al.
TABLE 3. Simulation Parameters for Cyanobenzoate Esters, 2, 6, 9, and 12a
compound
(concentration/mM)
concentration
of LiClO4/ mM
E°1
(V)
E°2
(V)
E°3
(V)
E°4
(V)
kf,5
(s-1
)
0
4
8
0
4
8
0
4
8
0
4
8
-2.016
-2.016
-2.016
-2.037
-2.039
-2.039
-2.040
-2.040
-2.039
-1.986
-1.987
-1.987
-2.092
-2.092
-2.092
-2.124
-2.123
-2.123
-2.134
-2.131
-2.127
-2.078
-2.076
-2.076
0
2 (1.91)
6 (2.01)
9 (2.00)
12 (1.94)
-1.290
-1.300
-1.322
-1.307
0.55
1.16
0
0.80
3.00
0
0.74
2.03
0
1.43
6.95
-1.266
-1.304
-1.325
-1.317
-1.266
-1.332
-1.325
-1.320
-1.363
-1.291
-1.291
-1.278
a Each entry is the parameter value used for simulations from 0.100 to 10.0 V/s.
the corresponding diol with two moles of benzoyl chloride,
4-cyanobenzoyl chloride, or 4-nitrobenzoyl chloride in ether,
dichloromethane, or chloroform at room temperature using pyridine
as the catalyst, according to standard methodology.11 For com-
pounds 10-12, the temperature of the reaction was increased to
50 °C. Diesters 5 and 8 were synthesized by a similar methodology
using dichloromethane as solvent at room temperature and using
4-pyrrolidinopyridine/Et3N as the catalyst system.12,13 For diester
14, the reaction was carried out in toluene at 140 °C using Et3N as
the catalyst.14
Electrochemical Cells, Electrodes, and Instrumentation. These
were as described earlier.10 The working electrode was a 0.3-cm
diameter glassy carbon electrode whose area was determined to be
0.0814 cm2. The reference electrode was a silver wire immersed
in 0.10 M Bu4NPF6/0.010 M AgNO3 in acetonitrile. The potential
of this reference electrode was periodically measured versus the
reversible ferrocene/ferrocenium potential, and all potentials re-
ported in this work are with respect to ferrocene. The temperature
was maintained at 298 K.
standard potentials, E°1 and E°2, are scarcely affected by the
low concentrations of lithium perchlorate used here. However,
the rate of formation of 4,4′-dicyanobenzil is greatly increased.
This is likely due to ion pairing between the dianion of the
diester and the lithium ions, which should facilitate the formation
of the cyclic intermediate. Interestingly, for compound 12, there
is a change in mechanism compared to compound 10. Both of
these diesters are derived from meso-hydrobenzoin, but 12, in
the presence of lithium ions, produces 4,4′-dicyanobenzil,
whereas 10 undergoes the opposite type of cleavage giving
stilbene.
These results show that the addition of lithium ions to the
electrolyte favors the formation of benzil for these cyanoben-
zoate esters where the dianion diradical is quite stable. In
experiments with 10, whose dianion diradical preferentially
cleaves benzoate to produce stilbene, the addition of 10 mM
LiClO4 caused a transition from stilbene formation to benzil
formation by way of the now favored cyclic intermediate.
Specifically, at 1 V/s, peaks for both stilbene and benzil appear
in the voltammogram, whereas only stilbene is detected in the
absence of lithium ions.
Digital simulations were conducted using DigiElch, version 2.0,
a free software package for the digital simulation of common
electrochemical experiments (http://www.digielch.de).15
Acknowledgment. Support of this research by the National
Science Foundation, Grant CHE 0347471, is gratefully ac-
knowledged. In the summer of 2005, C.L.M. was a participant
in the Collaborative Research in the Chemical Sciences (CRCS)
Program, Research Experiences for Undergraduates (REU),
sponsored by the National Science Foundation (Grant CHE-
0453466) at the University of Arizona.
Conclusions
The dianions of dibenzoate esters of glycols decompose by
two routes giving either benzil and the dianion of the diol or
two benzoates and olefin. The latter reaction occurs exclusively
for cases where cleavage of the first benzoate leaves a tertiary
radical or a benzylic radical. Benzil formation is observed when
cleavage of benzoate would leave a primary or secondary
radical. For these compounds, benzil formation can compete
with alkene formation by cleavage of benzoate. Benzil formation
is postulated to proceed by way of a cyclic intermediate dianion.
When benzoate is replaced by p-cyano- or p-nitrobenzoate, the
corresponding dianion diradicals are rather stable on the
voltammetric time scale. In the case of the cyanobenzoate esters,
slow formation of benzil is observed when 4-8 mM lithium
salts are added to the electrolyte.
Supporting Information Available: Tables of simulation
parameters for 1, 2, 4-10, and 12. This material is available free
of charge via the Internet at http://pubs.acs.org.
JO060535W
(10) Mac´ıas-Ruvalcaba, N. A.; Evans, D. H. J. Phys. Chem. B 2005,
109, 14642-14647.
(11) Reimschuessel, H. K.; Debona, B. T.; Murthy, A. K. S. J. Polym.
Sci., Polym. Chem. Ed. 1979, 17, 3217-39.
(12) Ho¨fle, G.; Steglich, W.; Vorbru¨eggen, H. Angew. Chem., Int. Ed.
Eng. 1978, 17, 569-583.
(13) Wolfe, M. S. Synth. Commun. 1997, 27, 2975-2984.
(14) Caddick, S.; McCarroll, A. J.; Sandham, D. A. Tetrahedron 2001,
57, 6305-6310.
Experimental Section
Chemicals and Reagents. The solvent for electrochemistry was
acetonitrile or N,N-dimethylformamide, and the electrolyte was
tetrabutylammonium hexafluorophosphate. Sources and treatment
of the solvents and electrolyte have been described.10 Diesters 1-4,
6, 7, and 9-13 were synthesized by the reaction of one mole of
(15) (a) Rudolph, M. J. Electroanal. Chem. 2003, 543, 23-29. (b)
Rudolph, M. J. Electroanal. Chem. 2004, 571, 289-307. (c) Rudolph, M.
J. Electroanal. Chem. 2003, 558, 171-176. (d) Rudolph, M. J. Comput.
Chem. 2005, 26, 619-632. (e) Rudolph, M. J. Comput. Chem. 2005, 26,
633-641. (f) Rudolph, M. J. Comput. Chem. 2005, 26, 1193-1204.
4834 J. Org. Chem., Vol. 71, No. 13, 2006