oxidation of COT was carried out in acetic acid.6 The
products at a carbon anode were found to be a 70:30 mixture
of 5,6-diacetoxybicyclo[4.2.0]octadiene (2a, as a mixture of
diastereomers) and cycloheptatrienyl acylal (3a). The anodic
processes leading to 2a and 3a are clearly complex. We felt
therefore that a reexamination of the anodic behavior of 1
was in order. We chose methanol as a neutral solvent for
this purpose. Because methanol is less acidic than acetic acid,
the occurrence of acid-promoted side reactions was consid-
ered less likely.
The anodic oxidation of COT was carried out at constant
current at a graphite carbon cloth anode in methanol
containing lithium tetrafluoroborate as the supporting elec-
trolyte.7 The cathode was a platinum mesh electrode. The
electrolysis afforded a single major product in 95% yield.
1H and COSY NMR spectra (see Supporting Information),
together with the mass spectral fragmentation pattern, clearly
established the product as the known dimethyl acetal of
7-formylcycloheptatriene (3b).8 A few other substances were
formed in 1% yield or less in the reaction, but none exhibited
a mass spectrum suggestive of 2b. (One was phenylacetal-
dehyde, presumably produced by the presence of a small
amount of adventitious water in the electrolysis medium).8
The probable pathways for production of 2 and 3 are
shown in Scheme 2. The first intermediate is undoubtedly
cyclooctatriene 6 and subsequent orbital symmetry allowed
disrotatory electrocyclization would lead to the bicyclic
product 2 (path A). On the other hand, a 1,2 shift in 5 would
afford first the rearranged cation 7, followed by attack by
solvent to afford 3 (path B).
The failure to produce 2b in methanol presents an
interesting mechanistic problem. This might be accounted
for by a greater driving force for rearrangement of cation
5b to 7b compared to the 5a to 7a conversion because of
the greater electronegativity of the acetoxy group in 7a
compared to methoxy and/or the higher polarity of methanol,
allowing 5b a longer lifetime in which to rearrange. On the
other hand, the greater nucleophilicity of methanol ought to
favor formation of 6b. We have examined this question
computationally.
Computational Results. The most likely geometry for
each structure was first established by a global minimum
energy conformational search using the GMMX subroutine
in PCMODEL 9.0.11 The energies of the neutral species 2a,b,
3a,b, and 6a,b and cations 5a,b and 7a,b were then computed
at the Hartree-Fock 6-31G* level,12 both for the unsolvated
species and with solvation included, using the PCM (polar-
ized continuum method) of Tomasi.13 The computations
(Table 1) indicate that rearrangement of the eight-membered
ring cation 5b to the corresponding seven-membered ring
species 7b is exothermic by 1.8 kcal/mol in the absence of
solvent and by 0.9 kcal/mol in methanol. The transition-state
energy for the rearrangement of 5b to 7b, respectively, was
computed by the synchronous transit-guided quasi-Newton
method of Schlegel,14 as implemented (QST2) in the Gauss-
ian 03 suite of programs.12 The activation energy for
rearrangement of 5b is computed to be 7.0 kcal/mol in
methanol. The low value of the activation energy is unsur-
prising because the structures of the two ions are very
similar: like its chloro analogue,15 5b is a nonplanar
8-homotropylium ion, and therefore the ends of its heptatrie-
nylic system are held closely together (Figure 1). Conversion
of 5b to 7b requires a simple shift of an electron pair with
relatively little motion of the other atoms. Conversion of the
neutral dimethoxycyclooctatriene 6b to acetal 3b is computed
to be exothermic by 14.2 kcal/mol. Electrocyclization of 6b
to 2b is exothermic by 5.4 kcal/mol. Acetal 3b is the most
stable of the three neutral dimethoxy substances in this
system, and the kinetic barrier to rearrangement of 5b to 7b
is quite low. Formation of acetal 3b is therefore favored both
kinetically and thermodynamically in methanol.
Scheme 2. Mechanism of Electrochemical Oxidation of
Cyclooctatetraene in Protic Media
4, the cation radical of 1. Attack by solvent SOH upon 4,
followed by loss of a second electron, should afford the key
intermediate, cation 5. There is a great deal of literature
precedent on the anodic oxidation of alkenes for this
conversion of 1 to 5.9,10 Attack of solvent on 5 to afford the
In dramatic contrast, the conversion of 5a to 7a is
computed to be endothermic by 15.0 kcal in acetic acid, with
an activation energy of 15.3 kcal/mol in that solvent. Acylal
3a is however similar to 3b in that it is the most stable of
the three neutral diacetoxy species; it is more stable than
(6) Eberson, L.; Nyberg, K.; Finkelstein, M.; Peterson, R. C.; Ross, S.
D. J. Org. Chem. 1967, 32, 16.
(7) COT was found to exhibit a single irreversible voltammetric wave
at +1.50 V vs SCE.
(8) Cope, A. C.; Nelson, N. A.; Smith, D. S. J. Am. Chem. Soc. 1954,
76, 1100.
(9) Halas, S.; Okyne, K.; Fry, A. J. Electrochim. Acta 2003, 48, 1837.
(10) Steckhan, E.; Schafer, H. Angew. Chem., Int. Ed. Engl. 1974, 13,
472.
(11) Serena Software, Bloominton, IN.
(12) Frisch, M. J. et al. Gaussian 03, revision B. 04; Gaussian, Inc.:
Pittsburgh, PA, 2003. Full references for Gaussian programs are provided
in the Supporting Information.
(13) (a) Tomasi, J.; Persico, M. Chem. ReV. 1994, 94, 2027. (b) Tomasi,
J.; Mennucci, B.; Cammi, R. Chem. ReV. 2005, 105, 716.
(14) Peng, C.; Schlegel, H. B. Isr. J. Chem. 1994, 33, 449.
(15) (a) Huisgen, R.; Gasteiger, J. Tetrahedron Lett. 1972, 13, 3661. (b)
Gasteiger, J.; Huisgen, R. Tetrahedron Lett. 1972, 13, 3665.
1672
Org. Lett., Vol. 9, No. 9, 2007