In a recent report, Nakamura7 circumvented this problem
by using a diarylzinc prepared by transmetalation from the
corresponding aryl Grignard reagent.
Table 1. Cobalt-Catalyzed Cross-Coupling of Phenylmagnesium
Bromide with 2-Bromo- or 2-Iodobutane
However, 1.5 equivalents of diarylzinc are required
since only one of the two aryl groups of Ar2Zn is
transferred. Therefore, it is necessary to use 3 equivalents
of the starting Grignard reagent. To improve the procedure,
Nakamura proposed to use an organozinc compound
ArZnCH2SiMe3 prepared by reacting ZnCl2 successively
with ArMgCl and Me3SiCH2MgCl. Indeed, this unsym-
metrical diorganozinc reagent selectively transfers the aryl
group. Unfortunately, it is less reactive than the corre-
sponding diarylzinc, and 2 equivalents are necessary to
obtain satisfactory yields. Moreover, Me3SiCH2MgCl is
a very expensive material. Therefore, these procedures are
not very attractive for large scale applications since at
least 2 equivalents of the starting aryl Grignard reagent
have to be used.
From both an economical and an environmental point
of view, cobalt salts are not as convenient as iron or
manganese8 salts. However, they compare favorably to
palladium or nickel salts and deserve to be considered
when iron salts are not efficient. In 2006, Oshima9
described the cobalt-catalyzed alkylation of aromatic
Grignard reagents. A mixture CoCl2/N,N,N′,N′-tetram-
ethyl-1,2-cyclohexanediamine is used as a catalytic sys-
tem. As a rule, yields are excellent, and various primary
and secondary alkyl iodides were coupled successfully
under mild conditions. The procedure has been success-
fully applied to the synthesis of AH13205, an EP2-receptor
agonist that lowers intraocular pressure. Nevertheless, it
should be noted that the use of nonactivated alkyl
bromides in place of the corresponding iodides results in
lower yields (73% instead of 93%, Table 1, entries 4 and
6).
a The yield was determinated by GC with pentadecane as an internal
standard.
The results obtained by reacting 2-iodo- and 2-bromobu-
tane with phenylmagnesium bromide under cobalt catalysis
are summarized in Table 1.
With cobalt chloride as a catalyst, 2-phenylbutane 3d was
formed in only 10% yield from 2-iodobutane (entry 1). In
agreement with the results published by Oshima,9 the
presence of N,N,N′,N′-tetramethyl-1,2-cyclohexanediamine
allows an excellent yield (93%, entry 4). Unfortunately, from
2-bromobutane, the yield is significantly lower (73% instead
of 93%, entries 4 and 6).
According to our experience in the case of the alkylation
of aromatic Grignard reagents under iron catalysis,2a the use
of a metal acetylacetonate instead of the corresponding
chloride as a catalyst can give better yields of coupling
product. Thus, we tried to use cobalt acetylacetonate in
place of cobalt chloride. In the absence of ligand, yields
obtained from both salts are very similar (entries 1 and
2). Surprisingly, we discovered that by using the complex
Co(acac)3/TMEDA (1:1) formed in situ, the yield jumps
from 14% to 94% (entries 2 and 3). This was unexpected
since Oshima had shown that, with cobalt chloride as a
catalyst, TMEDA is clearly less efficient than N,N,N′,N′-
tetramethyl-1,2-cyclohexanediamine (Table 2, entry 9 and
note d).9 This result is very interesting since commercial
TMEDA is a very simple and inexpensive starting material
compared to N,N,N′,N′-tetramethyl-1,2-cyclohexanedi-
amine.
In light of the previous considerations and because of
our background in this field,10 we tried to prepare ester 2
from bromoester 1 via a cobalt-catalyzed coupling reac-
tion. Therefore, we decided to reinvestigate this coupling
reaction to improve the yield in the case of nonactivated
alkyl bromides.
(4) For exhaustive reviews on iron-mediated coupling reactions
see: (a) Fu¨rstner, A.; Martin, R. Chem. Lett. 2005, 34, 624. (b) Bolm, C.;
Legros, J.; Le Paih, J.; Zani, L. Chem. ReV. 2004, 104, 6217.
(5) Palladium- or nickel-catalyzed alkylation reactions generally suffer
from ꢀ-elimination processes from the intermediate alkyl-metal species
formed during the reaction, thus limiting the scope of the possible
halogenoalkanes used as a substrate.
(6) Several interesting exemples of reaction between ArMgX and primary
functionalized alkyl halides in the presence of [Fe(C2H4)4][Li(tmeda)]2 as
a catalyst were recently described by Fu¨rstner.4a However, the yields are
significantly lower when secondary functionalized unactivated alkyl bro-
mides are used.
(7) Nakamura, M.; Ito, S.; Matsuo, K.; Nakamura, E. Synlett 2005, 1794.
(8) Manganese salts are also very attractive from both economical and
environmental points of view. For Mn-catalyzed cross-coupling reactions,
see: (a) Cahiez, G.; Duplais, C.; Buendia, J. Chem. ReV. 2008, in press. (b)
Cahiez, G.; Gager, O.; Lecomte, F. Org. Lett. 2008, ASAP.
(9) Tsuji, T.; Yorimitsu, H.; Oshima, K. J. Am. Chem. Soc. 2006, 128,
1886. For reviews on Co-catalyzed coupling reactions, see: (a) Yorimitsu,
H.; Oshima, K. Pure Appl. Chem. 2006, 78, 441. (b) Shinokubo, H.; Oshima,
K. Eur. J. Org. Chem. 2004, 2081. (c) Iqbal, J.; Mukhopadhyay, M.; Mandal,
A. K. Synlett 1997, 876.
Moreover, we were pleased to note that the new catalytic
system allows similar yields from 2-iodo- and 2-bromobu-
278
Org. Lett., Vol. 11, No. 2, 2009