Organic Letters
Letter
Scheme 2. Scope of Cobalt-Catalyzed C(sp2)−C(sp3)
subsequent experiments for operational simplicity and
consistency (entry 6). Modest modifications to L1 such as
replacing the hydrogen on each nitrogen with an N-methyl
group (TMCyDA, L2)importantly, a ligand that has been
successfully used in previously reported, cobalt-catalyzed cross-
c
Suzuki−Miyaura Cross Coupling
18
coupling methods or the removal of the cyclohexane
backbone (DMEDA, L3) resulted in diminished yields of 37
and 33%, respectively (entries 4 and 5). Because the yields for
reactions with L2 and L3 are comparable to the yield for CoBr2
alone (31%, entry 2), inefficient complexation may be
responsible for the low yields; however, other effects might
also be operative.
With a suitable ligand identified, additional reaction
parameters were investigated. Relatively strong alkoxide bases
were required to promote cross coupling (entries 7 and 8), as
weaker phenoxide or silanolate derivatives resulted in
elimination with the former and overall poor conversion with
the latter (Table S8). The stoichiometry of the base also
proved important, as increasing the equivalents of KOMe from
1.25 to 1.5 had a deleterious effect on the formation of 3a with
increased elimination (entry 7).
Optimal cross coupling was observed with a slight excess of
aryl boron reagent 1a (1.5 equiv) relative to KOMe (1.25
equiv). This likely reduces the amount of base in solution,
thereby minimizing side products. An evaluation of common
solvents established DMA as optimal (entry 6), likely a result
of its high polarity that solubilizes all of the reaction
components. Ethereal solvents are widely used in other
cobalt-catalyzed cross coupling reactions;10c however, C−Br
reduction 2a′ was a significant side product in both THF and
1,4-dioxane, with 2a′ being the major product in THF (entries
9 and 10). Increasing the amount of the cobalt precursor to 15
mol % resulted in a modest improvement in yield (entries 11
and 12) and no coupling product was observed in the absence
of cobalt (entry 13). Finally, PhB(OH)2 and PhB(Pin), more
commonly encountered aryl boron reagents, afforded only
small amounts of 3a (entries 14 and 15), demonstrating the
unique efficacy of ArB(neo) reagents for transmetalation with
cobalt.
a
1
Yield determined by H NMR spectroscopy using mesitylene as an
b
internal standard. Diastereomeric ratio (dr) of 4:1 measured by
crude H NMR spectroscopy; major diastereomer shown. Yield of
isolated product shown unless otherwise stated.
c
1
With the optimized conditions established, the scope of the
cobalt-catalyzed cross coupling with various nucleophiles 1 and
electrophiles 2 was explored (Scheme 2). Reactivity trends
were determined for both 1 and 2 while keeping the other
coupling partner constant. Coupling products 3b−g derived
from a number of sterically and electronically differentiated
aryl B(neo) reagents were prepared. Specifically, sterically
demanding 2-methylphenyl 3b and 1-naphthyl 3c aromatic
rings were tolerated, along with alkoxy- (3g), phenoxy- (3d),
and trifluoromethyl-substituted (3f) arenes. Branched alkyl
electrophiles were effective partners, providing 3h−j in up to
73% yield. A set of electrophiles of constant chain length
between the site of C−C bond formation and a synthetically
versatile functional group provided products with an ester
(3k), a protected alcohol (3l), and a protected amine (3m),
each in good yield. The scope of this method was expanded to
include diversely functionalized coupling partners into the final
Suzuki products. Highlights include nucleophile- (3q) and
electrophile- (3p and 3s) derived heterocycles, adamantyl-19
(3o) and oxetane-20 (3p) based bioisosteres, and compounds
with functional groups that can be directly used for (−Ac 3n,
−CN 3o, and 3r), or deprotected prior to (−OTBS 3r and
−NBoc 3s), further synthetic manipulation.
The discovery of an effective C(sp2)−C(sp3) cross-coupling
method that is compatible with some common organic
functional groups inspired studies into the mechanism of the
reaction. Observation of the reduction product 2a′, derived
from the electrophile, suggested the intermediacy of radicals
that likely undergo competing H-atom abstraction. The
intermediacy of carbon-centered radicals is well established
for cross-coupling reactions using first-row transition-metal
catalysts including cobalt.10c Such intermediates are often
derived from organohalide halogen atom abstraction or single
electron transfer to some other type of activated C−X bond by
a reduced metal intermediate. A series of radical trap
experiments were performed to determine if radical inter-
mediates are formed in the present system. Cobalt-catalyzed
cross coupling of 4 with 6-bromo-1-hexene 5, a well-
established radical clock,21 produced a 7:1 mixture of cyclized
6a to linear 6b coupling products in 74% overall yield (Scheme
3A). This supports a pathway where the rate of direct radical
cross coupling to give linear product 6b is approximately the
same order of magnitude as the rate of initial radical cyclization
that occurs prior to C−C bond formation for cyclized product
6a.22 Intramolecular migratory insertion of the tethered alkene
of the Co-alkyl intermediate that occurs prior to reductive
C
Org. Lett. XXXX, XXX, XXX−XXX