Chemistry Letters 2001
625
CH CN as solvent (Entry 4). It was clearly recognized that
3
reaction efficiency of the metal-catalyzed reaction was depend-
ent on the ligand. Hence the reactions were conducted in the
presence of several types of ligands (Entries 5–9). Among the
five types tested, the best result was recorded when the reaction
was carried out using 10 mol% of dichloro-bisoxazoline ligand
(
(
iPr-Cl box) and desired dimer 2 was obtained in 78% yield
Entry 8). However, no enantioselectivity was observed even
2
when an optically active ligand was employed for these reac-
tions. Dimer 2 was obtained with only 7% ee when the reaction
was carried out in the presence of optically active iPr-Me box
2
ligand (Entry 7). From these results, it seems that the dimeriza-
tion reaction takes place without influence of the ligand via the
radical cation pathway as hypothesized initially in Scheme 1,
though the ligand employed obviously affects the Fe –Fe
oxidation reduction system.
In addition, a mixture (59:41) of anti-(1α,2β,3α,4β)-2 and
syn-(1α,2β,3β,4α)-2 was obtained in 15% yield and the starting
material was recovered in 73% yield when our reaction was car-
ried out at low temperature and stopped at 10 min (eq 2). This
result was the same as that reported by the dimerization of trans-
the counter anion from perchlorate to a safer anion species, will
make it even more beneficial.
3
+
2+
The authors are grateful to the SC-NMR Laboratory of
Okayama University for the NMR measurements.
References and Notes
1
2
A review see. P. Laszlo, Acc. Chem. Res., 19, 121 (1986).
H. Ohara, K. Kudo, T. Itoh, M. Nakamura, and E.
Nakamura, Heterocycles, 52, 505 (2000).
anethole 1 mediated by tris(p-bromophenyl)aminium hexa-
chloroantimoate.3g GC analysis revealed that syn-dimer 2 was
3
A review see. a) N. L. Bauld, Tetrahedron, 45, 5307
(
1989). Photodimerization: b) K. Mizuno, C. Pac, and H.
completely isomerized to anti-dimer 2 with forming of a small
amount of trans-anethole 1 when an acetonitrile solution of mix-
ture of anti-2 and syn-2 (59:4 1) was treated with 10 mol% of
Fe(ClO ) at rt for 6 h. It therefore is concluded that perfect
Sakurai, Chem. Lett., 1973, 309. c) C. Pac, K. Mizuno, T.
Sugioka, and H. Sakurai, Chem. Lett., 1973, 187. d) C. Pac,
T. Sugioka, K. Mizuno, and H. Sakurai, Bull. Chem. Soc.
Jpn., 46, 238 (1973). e) F. D. Lewis and M. Kojima, J. Am.
Chem. Soc., 110, 8660 (1988). D F. D. Lewis and M.
Kojima. J. Am. Chem. Soc., 110, 8664 (1988). Tris(p-
bromophenyl)aminium hexachloroantimonate: g) N. L.
Bauld and R. Pabon, J. Am. Chem. Soc., 105, 633 (1983).
a) F. A. Bell, R. A. Crellin, and A. Ledwith, J. Chem. Soc.,
D, 1969, 251. b) D. W. Reynolds, K. T. Lorenz, H-S.
Chiou, D. J. Bellville, R. A. Pabon, and N. L. Bauld, J. Am.
Chem. Soc., 109, 4960 (1987).
4
3
stereoselectivity of the present reaction was attributable to the
formation of thermodynamically stable dimer anti-2 during the
equilibrium between the cycloadduct 2 and the substrate 1 which
3g
was produced by the cycloreversion of dimer 2.
4
5
It was proposed that [2+2]-cycloaddition proceeds through
an electron transfer (ET) pathway: a) for a review see: A.
Ledwith, Acc. Chem. Res., 5, 133 (1972). b) for the most
recent example see: G. D. Reddy and O. Wiest, J. Org.
Chem., 64, 2860 (1999).
3
e 1
6
7
Some spectroscopic properties of (1α,2β,3α,4β)-2 : H
NMR (200 MHz, CDCl ): δ 7.13 (d, 4H, J = 8.7 Hz), 6.82
3
(
2
d, 4H, J = 8.7 Hz), 3.77 (s, 6H), 2.80 (m, 2H), 1.82 (m,
H), 1.18 (d, 6H, J = 6.0 Hz); C NMR (50 MHz, CDCl3)
1
3
It was well recognized that efficiency of the catalyst was
dependent on the proper supporting materials. We next investi-
gated these supporting materials; neither celite nor silica gel
accelerated or improved the chemical yield (Entries 1 and 2,
Table 2). We finally found that the alumina supported
Fe(ClO ) catalyst gave excellent results, though requiring a
δ 157.9, 135.9, 127.7, 113.7, 55.2 52.4, 43.2, 18.9.
Catalyst was prepared as follows: neutral alumina (714 mg,
Aldrich) was added to a solution of Fe(ClO ) ·nH O (286
4 3
2
mg) in freshly distilled CH CN (0.5 mL) and the solvent
3
was removed at 60 °C under reduced pressure using a
rotary evaporator. The mixture was dried under vacuum
and the resulting uniformly orange powder was used for
the reaction. The catalyst should be handled with care
because it has been reported that iron(III) perchlorate salt
detonates with violence when subjected to mechanical or
thermal shock. However, our catalyst is believed to be safe
due to the support by neutral alumina. The content of
Fe(III) ion was determined by iodometry titration in aque-
ous solution.
4
3
7
long reaction time. The desired dimeric compound 2 was
obtained in the highest yield of 92% using only 3 mol% of the
catalyst (Entry 3 in Table 2). To the best of our knowledge, this
is the least amount ever recorded of catalyst used for this type
of reaction.
In conclusion, we demonstrated that alumina supported
Fe(ClO ) catalyzed cyclodimerization of trans-anethole 1 to
4
3
afford (1α,2β,3α,4β)-1,2-bis(4-methoxyphenyl)-3,4-dimethyl-
cyclobutane (2)3e,g in excellent yield. Further investigation of
the scope and limitations of this reaction, especially switching