Communication
be found in the literature, showing that 1,3-enynes facilitate
[2+2] cycloaddition with alkenes under cobalt catalysis.[7a,8]
To elucidate our working hypothesis, we first examined the
reactions of two different types of alkenes with a 1,3-enyne.
Based on the reaction conditions developed by the groups of
Hilt and Cheng,[6] a CoBr2/dppp/Zn/ZnI2 (ratio 1:1:4:4) catalyst
system was used. As expected, styrene (1) reacted with 2-
methyl-1-hexen-3-yne (2) in the presence of the cobalt catalyst
to give desired cyclobutene 3 in 80% yield [Eq. (1)].
CꢀC reductive elimination. Then, C undergoes reductive elimi-
nation to give allene 5.
According to the preliminary results and mechanistic hy-
pothesis, alkenes with no allylic hydrogen atoms have the po-
tential to undergo [2+2] cycloaddition with 1,3-enynes. Sever-
al optimizations of reaction conditions were investigated
based on Equation (1) (for details, see the Supporting Informa-
tion).[13] An equimolar, or a slightly excessive, amount of 1 was
enough to obtain 3 in a sufficient yield. A screening of ligands
revealed that only dppp was suitable for catalytic conversion
into the product. The reaction with dppe, rac-BINAP, or BIHEP
gave cyclobutene 3 in approximately 10% yield.[14] The use of
other mono- or bidentate ligands was unsuccessful. Regarding
cobalt(II) salts, CoBr2, CoI2, and Co(OAc)2 were applicable to
this [2+2] cycloaddition, whereas CoCl2 was less effective.
After further optimizations, we selected the reaction conditions
as follows: CoBr2/dppp/Zn/ZnI2 (1:1:1:2) as the catalyst, 2 mol%
catalyst loading, and CH2Cl2 as the solvent.
Next, the [2+2] cycloaddition of styrene derivatives with
enyne 2 was examined (Scheme 4). Under the optimized con-
ditions, cyclobutene 3 was obtained in 79% isolated yield. The
reaction of para-methylstyrene and para-methoxystyrene gave
6 and 7 in 83 and 80% yield, respectively. Halogen atoms (F,
Cl, and Br) at the para-position were compatible with this [2+
2] cycloaddition, and provided the corresponding cyclobutenes
8–10 in high yields. These halogen substituents, Br and Cl sub-
stituents in particular, can be further functionalized by cross-
coupling reactions. Among the meta-substituted styrenes em-
ployed, the reaction of meta-methylstyrene was comparable to
the corresponding para-substituted example, giving 11 in 83%
yield. On the other hand, electron-withdrawing groups, such
as CF3 and F groups, decreased the reactivity to afford 12 and
13 in 23 and 43% yields, respectively. ortho-Tolyl-substituted
14 was obtained in only 8% yield, probably owing to steric
hindrance. However, neither ortho-methoxy- nor ortho-chloro-
substituents diminished the reactivity and provided com-
pounds 15 (91%) and 16 (51%), respectively. In these cases,
the reactions might be facilitated by chelation to the cobalt
center through an oxygen- or chlorine atom. para-Cyanostyr-
ene and 2-vinylpyridne did not react at all, probably owing to
the deactivation of the catalyst by the strong coordination of
the nitrogen atom. Although the yields were more modest, vi-
nylboronate and vinylsilane reacted to provide cyclobutenes
17 and 18, which can be utilized for further transformations,
such as cross-coupling, allylic substitution, and oxidation. The
scope of 1,3-enynes was also investigated. The reaction of
enynes bearing longer alkyl chains, and chloro- or silyloxy sub-
stituents afforded the corresponding cyclobutenes 19, 20, and
21 in moderate to high yields. With 5 mol% catalyst loading,
1,5-dien-3-yne also reacted to give 1,2-diisopropenylcyclobu-
tene 22 in 49% yield. 1,3-Enynes bearing a naked vinyl group
were also applied to this reaction, affording 23 (50%) and 24
(34%) in the presence of 5 mol% catalyst loading. However,
In contrast to the nickel-catalyzed reaction, the cobalt-cata-
lyzed [2+2] cycloaddition proceeded even at room tempera-
ture and 3 was obtained as a single regioisomer. On the other
hand, the reaction of 1-decene (4) with 2 gave no cyclobutene
under the same conditions; instead, tetrasubstituted allene 5
was obtained as the major product in 76% yield, accompanied
by a small amount of unidentified isomers [Eq. (2)].[10] Although
the cobalt-catalyzed cross-dimerization of alkenes with alkynes
gives (E,E)-dienes,[6] the configuration of the alkenyl moiety of
5 was Z.
The difference in the reactivity of alkenes 1 and 4, under
cobalt catalysis, can be rationalized as described in Scheme 3.
In both cases, the reaction is initiated by the oxidative cycliza-
tion of an alkene with a 1,3-enyne at the cobalt(I) center, form-
ing a h3-butadienyl cobaltacycle (A!B).[11,12] In the reaction of
styrene (1, R=Ph), endo-cyclic b-H elimination could possibly
be suppressed by the h3-butadienyl coordination, and reduc-
tive elimination could occur to give cyclobutene 3. In the case
of 1-decene (4, R=C8H17), however, exo-cyclic b-H elimination
from B takes place to form CoꢀH species C in preference to
a
competing homodimerization of enynes was also ob-
served.[15] Although the yield of the product was not improved,
the slow addition of the enyne through a syringe drive sup-
pressed the formation of the enyne dimer in the case of 23. As
Scheme 3. Proposed mechanism.
&
&
Chem. Eur. J. 2014, 20, 1 – 6
2
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!