Cobalt-catalyzed, room-temperature addition of aromatic imines to alkynes via directed C-H bond activation

Article abstract of DOI:10.1021/ja2047073

A quaternary catalytic system consisting of a cobalt salt, a triarylphosphine ligand, a Grignard reagent, and pyridine has been developed for chelation-assisted C-H bond activation of an aromatic imine, followed by insertion of an unactivated internal alkyne that occurs at ambient temperature. The reaction not only tolerates potentially senstitive functional groups (e.g., Cl, Br, CN, and tertiary amide), but also displays a unique regioselectivity. Thus, the presence of substituents such as methoxy, halogen, and cyano groups at the meta-position of the imino group led to selective C-C bond formation at the more sterically hindered ortho positions. Under acidic conditions, the hydroarylation products of dialkyl- and alkylarylacetylenes underwent cyclization to afford benzofulvene derivatives, while those of diarylacetylenes afforded the corresponding ketones in moderate to good yields. A mechanistic investigation into the reaction with the aid of deuterium-labeling experiments and kinetic analysis has indicated that oxidative addition of the ortho C-H bond is the rate-limiting step of the reaction. The kinetic analysis has also shed light on the complexity of the quaternary catalytic system.

Full text of DOI:10.1021/ja2047073

ARTICLE
pubs.acs.org/JACS
Cobalt-Catalyzed, Room-Temperature Addition of Aromatic Imines to
Alkynes via Directed CꢀH Bond Activation
Pin-Sheng Lee, Takeshi Fujita,† and Naohiko Yoshikai*
Division of Chemistry and Biological Chemistry, School of Physical and Mathematical Sciences, Nanyang Technological University,
Singapore 637371, Singapore
S Supporting Information
b
ABSTRACT: A quaternary catalytic system consisting of a cobalt salt, a
triarylphosphine ligand, a Grignard reagent, and pyridine has been devel-
oped for chelation-assisted CꢀH bond activation of an aromatic imine,
followed by insertion of an unactivated internal alkyne that occurs at ambient
temperature. The reaction not only tolerates potentially senstitive functional
groups (e.g., Cl, Br, CN, and tertiary amide), but also displays a unique regioselectivity. Thus, the presence of substituents such as
methoxy, halogen, and cyano groups at the meta-position of the imino group led to selective CꢀC bond formation at the more
sterically hindered ortho positions. Under acidic conditions, the hydroarylation products of dialkyl- and alkylarylacetylenes
underwent cyclization to afford benzofulvene derivatives, while those of diarylacetylenes afforded the corresponding ketones in
moderate to good yields. A mechanistic investigation into the reaction with the aid of deuterium-labeling experiments and kinetic
analysis has indicated that oxidative addition of the ortho CꢀH bond is the rate-limiting step of the reaction. The kinetic analysis has
also shed light on the complexity of the quaternary catalytic system.
’ INTRODUCTION
very recently.4c,d For example, in 1994, Kisch and co-workers
reported that the ortho CꢀH bonds of azobenzenes undergo anti-
addition to diphenylacetylene in the presence of CoH(N2)-
(PPh3)3 or CoH3(PPh3)3.4c Unfortunately, this reaction has a
limited scope, and the mechanism of the anti-addition remains
ambiguous.
The reacitivity of cobalt toward an aromatic CꢀH bond has
also been studied from a stoichiometric point of view.5,6 In
particular, cyclometalation reactions of Co(CH3)(PMe3)4 with a
series of aromatic substrates such as aryl ketones, imines,
phosphines, and thioketones, which have been explored by Klein
and co-workers, are noteworthy.6,7 While these cyclometalation
reactions are potential elementary processes for catalytic ortho
CꢀH bond functionalization, such a catalytic turnover has yet to
be achieved.
Recently, we reported that a ternary catalytic system, consisit-
ing of CoBr2, PMePh2, and MeMgCl, promotes the syn-addition
reaction of a 2-phenylpyridine derivative to an unactivated
internal alkyne to afford a trisubstituted olefin in good yield
with a high stereoselectivity (eq 1).8 This reaction has
demonstrated the feasibility of chelation-assisted CꢀH addi-
tion to a CꢀC multiple bond using a first-row transition metal
catalyst9 and has aroused renewed interest in cobalt-catalyzed
CꢀH bond functionalization.4aꢀd,10,11 Nevertheless, the co-
balt catalysis required as harsh conditions (100 °C) as those
usually required for analogous hydroarylation reactions to alkynes
using second- or third-row transition metal catalysts.12ꢀ15
Chelation-assisted CꢀH bond cleavage by a transition
metal complex has been established as a versatile strategy
for regioselective functionalization of ubiquitous but unreac-
tive CꢀH bonds since the discovery of ruthenium-catalyzed
ortho-alkylation of aromatic ketones with olefins by Murai and
co-workers.1,2 Regardless of the mechanism of the CꢀH bond
cleavage, the concept of chelation assistance has proven to be
broadly applicable, leading to the development of an extre-
mely wide variety of CꢀC and Cꢀheteroatom bond-forming
reactions through CꢀH bond functionalization. While sec-
ond-row transition metals such as ruthenium, rhodium, and
palladium have played a major role in this type of transforma-
tion, it is desirable to develop cost-effective, alternative
processes using naturally abundant first-row transition metals.
Furthermore, exploration of catalysis using these metals
would offer opportunities to find unique reactivities and
selectivities.3,4
The potential utility of cobalt, the first-row transition metal of
group 9, in chelation-assisted CꢀH bond functionalization has
been known since the mid 1950s, when Murahashi reported that
the reaction of an aldimine in the presence of Co2(CO)8 and
high-pressure carbon monoxide (100ꢀ200 atm) at high tem-
peratures (>200 °C) led to CO insertion into the ortho CꢀH
bond of the aldimine, affording an isoindolinone as the
product.4a Murahashi also reported the ortho CꢀH functio-
nalization of an azobenzene with carbon monoxide under
cobalt catalysis, which afforded an indazolone or a quinazoline-
dione, depending on the reaction temperature.4b However, these
seminal findings have only been followed by a few examples of ortho
CꢀH functionalization reactions catalyzed by cobalt complexes until
Received: May 23, 2011
Published: September 28, 2011
r
2011 American Chemical Society
17283
dx.doi.org/10.1021/ja2047073 J. Am. Chem. Soc. 2011, 133, 17283–17295
|

Journal of the American Chemical Society
ARTICLE
Scheme 1. Cobalt-Catalyzed Addition of Aromatic Imines to
Alkynes
Table 2. Effects of Grignard Reagent and Cobalt Sourcea
entry
CoXn
CoBr2
RMgX
yield (%)b
1
2
tBuCH2MgBr
MeMgCl
43
0
CoBr2
CoBr2
CoBr2
CoBr2
CoBr2
CoBr2
CoCl2
CoI2
3
nBuMgBr
5
4
iPrMgBr
35
9
5
tBuMgBr
6
Me3SiCH2MgCl
PhMgBr
16
5
7
Table 1. Screening of Ligandsa
8
tBuCH2MgBr
tBuCH2MgBr
tBuCH2MgBr
tBuCH2MgBr
30
1
9
10
11
Co(acac)3
Co(acac)2
27
3
a The reaction was performed on a 0.3 mmol scale. b Determined by GC/
GCꢀMS analysis using n-tridecane as an internal standard. The imine
E/Z ratio was not determined.
The reaction occurs readily at 20 °C, which is a remarkably mild
condition as compared to that for related transformations
reported thus far.2,12ꢀ15,17ꢀ19 Because of the mild reaction
conditions, the reaction tolerates a variety of imines and alkynes,
including those bearing potentially sensitive functional groups
(e.g., Cl, Br, CN, and tertiary amide). Furthermore, the reaction
exhibits unique regioselectivity. Thus, in the presence of a
substituent such as methoxy, halogen, and cyano groups at the
meta position, CꢀH bond cleavage takes place preferentially at
the more sterically hindered ortho position. Upon hydrolysis, the
hydroarylation products are converted to either benzofulvene
derivatives14c,17g,20,21 or ketones, depending on the nature of the
alkyne substituents. The present hydroarylation reaction is likely
to involve three major elementary steps, that is, chelation-assisted
oxidative addition of the ortho CꢀH bond, insertion of the alkyne
into the CoꢀH bond, and reductive elimination of the diorga-
nocobalt species. A mechanistic investigation with the aid of
deuterium-labeling experiments and kinetic analysis has indi-
cated that the CꢀH bond cleavage is the rate-determining step
and provided insight into how each of the reactants is involved in
the catalytic cycle. The complexity of the quaternary catalytic
system was also addressed by the kinetic analysis, which revealed
the dependence of the catalytic activity on each of the catalytic
components.
entry
ligand
PMePh2
RMgX
yield (%)b
1
2
MeMgCl
1
4
PPh3
tBuCH2MgBr
tBuCH2MgBr
tBuCH2MgBr
tBuCH2MgBr
tBuCH2MgBr
tBuCH2MgBr
tBuCH2MgBr
tBuCH2MgBr
tBuCH2MgBr
tBuCH2MgBr
tBuCH2MgBr
tBuCH2MgBr
3
P(4-MeOC6H4)3
P(4-Me2NC6H4)3
P(3-ClC6H4)3
P(4-ClC6H4)3
P(3-FC6H4)3
P(4-FC6H4)3
P(4-CF3C6H4)3
P(3,5-(CF3)2C6H3)3
P(C6F5)3
2
4
8
5
43
17
23
35
15
0
6
7
8
9
10
11
12
13
0
P(2-furyl)3
2
PCy3
0
a The reaction was performed on a 0.3 mmol scale. b Determined by GC/
GCꢀMS analysis using n-tridecane as an internal standard. The imine
E/Z ratio was not determined.
’ RESULTS AND DISCUSSION
Reaction Development. We chose the imine 1a derived from
acetophenone and p-anisidine and 4-octyne 2a as the model
substrates and carried out extensive screening of the reaction
conditions. Table 1 summarizes the screening data during the
early stage. The reaction of 1a (0.3 mmol) and 2a (0.45 mmol,
1.5 equiv) was performed at 40 °C by using CoBr2 (10 mol %),
a ligand (10ꢀ20 mol %), and a Grignard reagent (100 mol %)
In this Article, we report on a new cobalt-based quaternary
catalytic system consisting of CoBr2, P(3-ClC6H4)3, neopentyl-
magnesium bromide, and pyridine that efficiently promotes an
addition reaction of an aromatic imine to an internal alkyne
through chelation-assisted CꢀH bond activation (Scheme 1).16
17284
dx.doi.org/10.1021/ja2047073 |J. Am. Chem. Soc. 2011, 133, 17283–17295

Journal of the American Chemical Society
ARTICLE
in THF (0.2 M) for 12 h. While the catalytic system for the hydro-
arylation of arylpyridines (i.e., CoBr2/PMePh2/MeMgCl)8 gave
the hydroarylation product 3aa in only 1% yield (entry 1),
throughout several ensuing experiments we found that a combi-
nation of PPh3 and tBuCH2MgBr afforded a slight increase in the
yield of 3aa (entry 2).
Table 3. Effects of Temperature, Concentration, and
Additivesa
Subsequent screening of triarylphosphine ligands displayed an
intriguing trend. While triarylphosphines bearing electron-do-
nating groups at the 4-position did not have a pronounced effect
on the catalytic activity (entries 3 and 4), those bearing electron-
withdrawing chloro, fluoro, or trifluoromethyl substituents at the
3- or 4-position improved the reaction to give 3aa in 15ꢀ43%
yields (entries 5ꢀ9). Among these ligands, P(3-ClC6H4)3
emerged as the most promising (entry 5). Other electron-poor
ligands, such as P(3,5-(CF3)2C6H3)3 and P(C6F5)3, did not
promote the reaction at all (entries 10 and 11), and P(2-furyl)3
exhibited only a poor catalytic activity (entry 12). While PCy3
was an effective ligand for branched hydroarylation of styrenes,10b it
did not promote the present alkenylation reaction at all (entry
13). Bidentate phosphines such as dppe, dppp, and Xantphos
were entirely ineffective (data not shown).
Having identified the promising triarylphosphine ligand, P(3-
ClC6H4)3, we then examined the effects of the Grignard reagent
and the cobalt source (Table 2). However, none of our examina-
tions led to an improvement in the catalytic activity. Methyl and
primary and tertiary alkyl Grignard reagents afforded much lower
yields of 3aa than did tBuCH2MgBr (entries 2, 3, 5, and 6),
except that the use of iPrMgBr resulted in a moderate yield of
35% (entry 4). In addition, PhMgBr poorly promoted the
reaction (entry 7). CoBr2 emerged as the best cobalt source,
while the use of CoCl2 and Co(acac)3 resulted in moderate
activities (entries 8ꢀ11).
entry
temp (°C)
conc (M)
additive (mol %)
yield (%)b
1
2
40
20
50
40
40
40
40
20
0.2
0.2
0.2
0.4
0.2
0.2
0.2
0.4
none
43
53
45
53
54
37
43
87d
none
3
none
4
none
5
pyridine (40)
DMI (40)
DMPU (40)
pyridine (80)
6
7
8c
a The reaction was performed on a 0.3 mmol scale. DMI = N,N-
dimethylimidazolidinone. DMPU = 1,3-dimethyl-3,4,5,6-tetrahydro-
2(1H)-pyrimidinone. b Determined by GC/GCꢀMS analysis using
n-tridecane as an internal standard. The imine E/Z ratio was not
c
determined unless otherwise noted. 5 mol % of CoBr2, 10 mol % of
P(3-ClC6H4)3, 50 mol % of tBuCH2MgBr, 1.3 equiv of 4-octyne
were used. d Isolated yield. Imine E/Z ratio was 84/16 as determined
by 1H NMR. The dialkenylation product was obtained in 8% yield.
underwent addition to 4-octyne smoothly to afford the correspond-
ing hydroarylation products 3baꢀ3la in moderate to good yields.
While the reaction was generally accompanied by E/Z isomeriza-
tion of the CdN bond to some extent (80:20ꢀ90:10), CꢀC and
CꢀH bond formation took place with exclusive syn-stereoselectiv-
ity. Electron-donating (e.g., OMe), neutral, and electron-with-
drawing (e.g., CF3) substituents at the para- or meta-position
were tolerated. The presence of potentially sensitive chloro and
cyano substituents was also tolerated (see 3ea, 3ia, and 3ja). The
catalytic system allowed for a gram-scale reaction. Thus, the
reaction of the 3-methylacetophenone imine on a 6 mmol scale
(1.44 g) took place as efficiently as the reaction on a 0.3 mmol scale,
affording the product 3fa in 92% yield.
The reactions of the meta-substituted acetophenone imines
exhibited intriguing regioselectivities. While the substrates
bearing m-methyl and m-trifluoromethyl groups were selectively
alkenylated at the less hindered ortho position (see 3fa and 3ga,
regioselectivity g94:6), those bearing m-methoxy, m-chloro,
m-cyano, or m-fluoro substituents reacted preferentially at the
position proximal to these groups to afford 3haꢀ3ka in 87%,
66%, 58%, and 92% yields, respectively. While the regio-
isomers of 3ha, 3ja, and 3ka could not be either detected or
characterized in our hands, the regioisomer of 3ia was isolated in
19% yield. While the strong secondary directing effect of the
methoxy and fluoro substituents has been reported previously for
chelation-assisted CꢀH functionalization reactions using ruthe-
nium and iridium catalysts,14c,24 the high regioselectivities ob-
served here for the chloro, cyano, and bromo (vide infra)
substituents are unprecedented.25
Further improvement of the reaction was achieved by modify-
ing reaction conditions including the reaction temperature,
concentration, and additives (Table 3). Lowering the tempera-
ture to 20 °C improved the yield slightly, while increasing the
temperature to 50 °C did not lead to any notable change (entries
2 and 3). Increasing the concentration from 0.2 to 0.4 M also
improved the yield (entry 4). Furthermore, among several Lewis
basic additives examined, pyridine displayed a positive effect on
the reaction (entries 5ꢀ7). We eventually optimized the reaction
by taking the above three observations into consideration. Thus,
the reaction of 1a and 2a (1.3 equiv) in the presence of 5 mol %
CoBr2, 10 mol % P(3-ClC6H4)3, 50 mol % tBuCH2MgBr, and 80
mol % pyridine in THF (0.4 M) at 20 °C afforded the hydro-
arylation product 3aa in 87% isolated yield with exclusive syn-
selectivity and E/Z isomerization of the imine moiety (84:16)
(entry 8).22 The formation of 3aa was accompanied by a small
amount of a dialkenylation product (8%). It is notable that an
addition of the Grignard reagent to the imine CdN bond did not
take place under the optimized conditions. It should also be
noted that only a trace amount of a product due to cyclotrimer-
ization of 2a was observed.23 The amount of the Grignard
reagent used was critical for the activity of the cobalt catalyst,
as in other cases previously reported by our group.8,10 The use of
40 mol % or less of tBuCH2MgBr led to a significant decrease in
the reaction rate, which suggests that the Grignard reagent is not
a mere reducing agent for the cobalt precatalyst (vide infra).
Addition of Aryl Imines to Alkylacetylenes. With the
quaternary catalytic system in hand, we first performed the hydro-
arylation reaction of a variety of aromatic imines and alkylacetylenes
(Chart 1). Imines derived from substituted acetophenones
The imine derived from 3,5-difluoroacetophenone was highly
reactive and afforded a mixture of mono- and dialkenylated
products (ca. 27% and 57% yields, respectively) under the standard
17285
dx.doi.org/10.1021/ja2047073 |J. Am. Chem. Soc. 2011, 133, 17283–17295

Journal of the American Chemical Society
ARTICLE
Chart 1. Addition of Aryl Imines to Alkylacetylenesa
a The reaction was performed on a 0.3 mmol scale. The yields refer to the isolated products. The E/Z ratio of the imine moiety is shown in parentheses.
Unless otherwise noted, the stereochemistry of the olefin moiety was exclusively E. b The reaction time was 24 h. c The regioselectivity was 96:4. d The
regioselectivity was 94:6. e Only the regioisomer shown was isolated, and the presence of the other regioisomer could not be confirmed. f E/Z ratio of the
olefin moiety was 89:11. g Combined yield of 3ia (66%) and its regioisomer (19%). h 2.6 equiv of 4-octyne was used. i The regioselectivity was 57:43.
jCꢀC bond formation took place with regioselectivity of >99:1. k The regioselectivity of CꢀH bond cleavage was 89:11. l The regioselectivity was 97:3.
m rr = Regioisomeric ratio as estimated by 1H NMR. See the Supporting Information for the imine E/Z ratio.
conditions (i.e., 1.5 equiv of 4-octyne). When the amount of
4-octyne was increased to 2.6 equiv, the dialkenylation product
3la was obtained in near quantitative yield. The high reactivity of
this substrate would originate from the strong effect of the ortho-
fluorine atoms directing CꢀH bond cleavage.26 The imine derived
from 2-acetonaphthone also participated in the reaction to afford
3ma in a good yield, while the regioselectivity was only modest
(57:43). Heteroarenes such as thiophene and N-methylpyrrole
bearing the imino group at the 2-position failed to participate in the
reaction, while the 3-position of benzofuran could be alkenylated in
30% yield (see 3na).
3ac, and 3fc in moderate to good yields. In the latter two
cases, the aryl group was exclusively introduced to the less
hindered acetylenic carbon atom. 1-Phenyl-1-propyne and its
derivatives also participated in the reaction to afford 3adꢀ
3af in modest yields, where the CꢀC bond formation took
place highly regioselectively (g97:3) at the acetylenic car-
bon proximal to the methyl group. In contrast, only modest
regioselectivity (ca. 3:2) was observed for the reaction of
1-phenyl-1-butyne (see 3ag). Note that heteroaryl alkynes
such as 1-(2-thienyl)-1-propyne, 1-(2-furyl)-1-propyne, and
1-(2-pyridyl)-1-propyne as well as terminal alkynes such as
1-octyne and phenylacetylene failed to participate in the
reaction.
Aliphatic alkynes such as 1,4-bis(trimethylsilyl)-2-butyne and
4-methyl-2-pentyne afforded hydroarylation products 3ab,
17286
dx.doi.org/10.1021/ja2047073 |J. Am. Chem. Soc. 2011, 133, 17283–17295

Journal of the American Chemical Society
ARTICLE
Scheme 2. Transformation of 3fa under Hydrolysis
Conditions
Scheme 4
Scheme 3
Chart 2. Benzofulvene Derivativesa
Conversion of Hydroarylation Products to Benzofulvenes.
When the hydroarylation product 3fa was subjected to acidic
conditions (3 M HCl, 18 h), a mixture of benzofulvenes 4fa
(a mixture of E and Z isomers) and 4fa0 was obtained in 81% yield
with a ratio of 73:25:2 (Scheme 2).20,27 The ratio of the isomers
was fairly reproducible from run to run, the error being less than
5%. When quenched within a period of 1 h, the reaction afforded
an aminoindanol 5fa (dr > 95:5) and a ketone 6fa (E/Z = 91/9)
in 52% and 20% yields, respectively.28 Furthermore, milder
hydrolysis conditions (HOAc/NaOAc buffer) effected the con-
version of 3fa to the aminoindanol 5fa in 92% yield with a high
diastereoselectivity (>95:5).29
The mechanism of the diastereoselective formation of the
aminoindanol 5fa may be rationalized as illustrated in Scheme 3.
Thus, the cyclization process is triggered by protonation of the
imine, followed by intramolecular nucleophilic attack of the
olefin moiety on the iminium moiety. The relative stereochem-
istry of the C1 and C2 positions of the final product 5fa is
controlled by the transition structure in this step (see Scheme 3).
A water molecule attacks the resulting benzylic cation inter-
mediate selectively from the sterically less hindered side to afford
the aminoindanol 5fa.
a Unless otherwise noted, the reaction was performed at room tempera-
ture in 3 M HCl/THF for 18 h. b The ratio of the isomers is shown in
c
d
parentheses. The reaction was performed at 50 °C for 8 h. The
reaction time was 72 h. e The reaction was performed at 50 °C for 9 h.
the product mixture was different between the two cases. While
5fa afforded only the regioisomer 4fa in 87% yield with a
stereoselectivity of 76:24 (Scheme 4a), 6fa afforded both 4fa
and 4fa0 in 98% yield with a ratio of 14:50:36 (Scheme 4b).
Again, the ratios of the benzofulvene isomers obtained by the two
reactions were reproducible (error < 5%). Thus, the product
composition of the former case was much closer to that observed
Upon exposure to 3 M HCl, both the isolated aminoindanol
5fa and the ketone 6fa afforded a mixture of the benzofulvene
derivatives 4fa and 4fa0 (Scheme 4). However, the constitution of
17287
dx.doi.org/10.1021/ja2047073 |J. Am. Chem. Soc. 2011, 133, 17283–17295

Journal of the American Chemical Society
ARTICLE
Chart 3. Addition of Aryl Imines to Diarylacetylenesa
a The reaction was performed on a 0.3 mmol scale. The yields refer to isolated products. The E/Z ratio is shown in parentheses. b NR'2 = morpholino.
Combined yield of 6ph (44%) and its cyclization (i.e., indenol-type) product (13%). c The regioselectivity was >99:1 according to 1H NMR analysis of
the crude product. d Combined yield of 6je (59%) and its regioisomer (17%). e Hydrolysis was performed at 55 °C for 3 h. f Concentration was 0.2 M.
g Catalyst loading was doubled. h The major isomer among three is shown. The ratio of the isomers was 63:23:14 as determined by lH NMR. i Combined
yield of 6ao0 (37%) and its dehydration (i.e., benzofulvene-type) product (7%). jConcentration was 0.3 M. k Obtained from the reaction of 6ai with
tetrabutylammonium fluoride (1 M in THF) at 60 °C for 12 h.
in the direct transformation to the benzofulvenes (73:25:2, see
Scheme 2). It should also be noted that the conversion of 5fa to
4fa required much shorter time (2 h) than the conversion of 6fa
to 4fa/4fa0 did (14 h) as monitored by TLC. Therefore, we
consider that the direct conversion of the imine 3fa to 4fa/4fa0 in
3 M HCl involves the aminoindanol 5fa and the ketone 6fa as the
major and minor intermediates, respectively. The conversion of
5fa to 4fa would involve acid-mediated sequential deamination
and dehydration processes. On the other hand, because of the
reversibility of the imine-to-ketone hydrolysis, the conversion of
6fa to 4fa/4fa0 could occur either directly (cf., Scheme 4b) or
indirectly via regeneration of 3fa and its conversion to 5fa
(Scheme 4c). We consider that the latter indirect pathway is
more important, because if only the former pathway is operative,
the isomer ratios obtained from 3fa and 5fa should have been
much different.
Some of the other hydroarylation products were also subjected
to the acidic conditions to afford the corresponding benzoful-
venes (Chart 2).30 The reaction was found to be sensitive to the
electronic property of the substituent. Among the products
shown in Chart 1, those bearing electron-donating or neutral
substituents at the ortho- or para-position to the olefin moiety
underwent the reaction smoothly to afford the corresponding
benzofulvenes in moderate to good yields. In contrast, when an
electron-withdrawing substituent was present at the para- or
meta-position (e.g., 3ca), cyclization of the alkenylated product
to the corresponding aminoindanol was observed, while its
conversion to the benzofulvene was sluggish. Note that the
17288
dx.doi.org/10.1021/ja2047073 |J. Am. Chem. Soc. 2011, 133, 17283–17295

Journal of the American Chemical Society
ARTICLE
cyclization reaction of 3ab to 4ab was accompanied by the loss of
one of the two Me3Si groups, which can be ascribed to desilyla-
tion occurring during nucleophilic attack of the allylsilane moiety
to the iminium moiety.
Scheme 5. Deuterium-Labeling Experimentsa
Addition to Diarylacetylenes. The present catalytic system is
also applicable to the reaction of aromatic imines with dipheny-
lacetylene 2h (Chart 3). In contrast to the cases of alkylacety-
lenes, acidic treatment of the addition product did not cause a
cyclization reaction, but cleanly hydrolyzed the imine moiety to
afford the corresponding ketone within 1 h, as a mixture of
olefinic E- and Z-isomers (typically 9:1). We confirmed that the
minor Z-isomer formed during the catalytic reaction but not after
quenching, while the mechanism of its formation is not clear at
present. The scope of the applicable imines was as wide as that for
the reaction with 4-octyne (see Chart 1). Note that the imines
derived from propiophenone and tetralone reacted as smoothly
as the acetophenone derivatives did (see 6rh and 6sh). The
reaction tolerated the presence of chloro, bromo, cyano, and
tertiary amide substituents (see 6eh, 6ih, 6jh, and 6ohꢀ6qh).
The low yields obtained for the bromo-substituted substrates
(6oh and 6qh) were because of the sluggishness of the reaction.
Thus, unreacted starting material largely remained, and hydro-
debromination only took place to a small extent (<10% ob-
served). Note that no alkenylation took place at the ortho
position of the amide functional group (see 6ph).
As we observed in the reactions of 4-octyne, methoxy, halogen
(F, Cl, and Br), and cyano groups at the meta-position served as
secondary directing groups for the cobalt catalyst, leading to
selective introduction of the alkenyl group onto the more
hindered ortho positions (see 6hhꢀ6kh and 6qh). The reaction
of 3,5-difluoroacetophenone imine with diphenylacetylene ex-
clusively afforded the monoalkenylation product 6lh. This was in
stark contrast to the reaction with 4-octyne, which preferentially
afforded the dialkenylation product (see 3la in Chart 1). This
difference may be ascribed to the greater steric bulk of the 1,2-
diphenylvinyl group than the 4-octenyl group, which should have
prevented the second alkenylation.
Symmetric diarylacetylenes bearing 3-methyl, 3-fluoro, 4-chloro,
3-bromo, and 4-bromo substituents reacted with imine 1a to
afford the corresponding hydroarylation products 6aiꢀ6am,
respectively, albeit in modest yields.31 On the other hand, diary-
lacetylenes having 4-methoxy and 4-trimethylsilyl groups afforded
indenol derivatives 6ao0 and 6ap0, respectively, presumably be-
cause these substituents increased the nucleophilicity of the
olefinic moieties of the initially formed alkenylation products.
An unsymmetrical diarylacetylene substituted with 4-methoxy and
4-fluoro groups underwent CꢀC bond formation preferentially at
the acetylenic carbon promixal to the fluorophenyl group, afford-
ing the adduct 6an as the major isomer, whose regiochemistry was
confirmed by 2D-NMR (HMQC and HMBC) analysis. Finally,
the reaction of imine 1a and phenyltrimethylsilylacetylene resulted
in exclusive CꢀC bond formation at the acetylenic carbon atom
proximal to the phenyl group, affording the product 6aq in 41%
yield. In addition, deprotection of the silyl group of 6aq was
achieved using tetrabutylammonium fluoride to afford the 1,1-
diarylethene 6ar in a moderate yield.
a Standard conditions: CoBr2 (0.02 M), P(3-ClC6H4)3 (0.04 M),
tBuCH2MgBr (0.2 M), pyridine (0.32 M), THF, 20 °C.
without any noticeable H/D crossover (Scheme 5a). This result
clearly demonstrates that the aryl group and the olefinic hydro-
gen atom in the product molecule come from the same reactant
molecule, which excludes a deprotonation mechanism for the
ortho CꢀH bond cleavage and conforms to a mechanism
involving oxidative addition of the CꢀH bond. Second, a
competitive reaction of an equimolar mixture of 1a and 1a-d5
with diphenylacetylene was quenched at the early stage to yield
the hydroarylation products 6ah/6ah-d5 in 22% total yield with
the ratio of 81:19 (Scheme 5b). The kinetic isotope effect (KIE)
value of 4.3 determined for the competitive reaction suggests that
either aromatic CꢀH bond cleavage or vinylic CꢀH bond
formation is the first irreversible step of the reaction.32,33 Third,
we measured initial rates of individual reactions of 1a and 1a-d5
with diphenylacetylene under the standard reaction conditions
(Scheme 5c, see the Supporting Information for details). The
average initial rates of product formation were 1.7 ((0.1) ꢁ
10ꢀ4 and 3.3 ((0.2) ꢁ 10ꢀ5 M sꢀ1 for 1a and 1a-d5, respectively.
Thus, a large KIE value of 5.2 ( 0.4, the magnitude of which was
close to the one observed for the competitive reaction (4.3,
Scheme 5b), was determined. This suggests that, under the
Deuterium Labeling Experiments and Kinetic Isotope
Effects. To probe the reaction mechanism, we initially per-
formed a series of experiments using deuterium-labeled imine
1a-d5 (Scheme 5). First, the reaction of an equimolar mixture of
1a-d5 and the nondeuterated imine 1b with diphenylacetylene
afforded 6ah-d5 and 6bh in 26% and 56% yields, respectively,
17289
dx.doi.org/10.1021/ja2047073 |J. Am. Chem. Soc. 2011, 133, 17283–17295

Journal of the American Chemical Society
ARTICLE
Figure 1. Initial reaction rate against the imine concentration [1a]0.
Reaction conditions: 1a (0.05ꢀ0.6 M), 2h (0.6 M), CoBr2 (0.02 M),
P(3-ClC6H4)3 (0.04 M), tBuCH2MgBr (0.2 M), pyridine (0.32 M),
THF, 20 °C, 10ꢀ15 min.
Figure 2. Initial reaction rate against alkyne concentration ((a) diphe-
nylacetylene, [2h]0; (b) 4-octyne, [2a]0). Reaction conditions: 1a (0.4
M), 2h or 2a (0.1ꢀ0.8 M), CoBr2 (0.02 M), P(3-ClC6H4)3 (0.04 M),
tBuCH2MgBr (0.2 M), pyridine (0.32 M), THF, 20 °C, 10ꢀ15 min.
standard conditions, the rate-limiting step involves either CꢀH
bond cleavage or CꢀH bond formation.
Rate Dependence on Imine and Alkyne. We then studied
the dependence of the initial reaction rate on each of the
reactants (i.e., imine 1a and alkyne 2h).16 First, we measured
the initial rate (Δ[3ah]/Δt) against the initial concentra-
tion of the imine 1a ([1a]0) ranging from 0.05 to 0.6 M
and observed a clear positive correlation between [1a]0 and
Δ[3ah]/Δt (Figure 1a). No obvious saturation kinetics was
observed at high concentrations up to 0.6 M. Logarithm plot
and linear fitting of these data gave a slope of 1.05, indicating a
first-order dependence of the initial rate on the imine con-
centration (Figure 1b). This, together with the results of the
KIE experiments (Scheme 5b,c), suggests that CꢀH bond
cleavage is the rate-limiting step of the reaction at least in the
early stage of the reaction under the standard conditions, where a
sufficient amount of diphenylacetylene is present (0.6 M).
In contrast to the above first-order relationship, when a similar
analysis was performed for diphenylacetylene 2h (0.1ꢀ0.8 M),
the initial rate exhibited a saturation behavior at high concentra-
tions (Figure 2a). Thus, the reaction rate showed a positive
response to [2h]0 at low concentrations (0.1ꢀ0.15 M) but
reached a plateau at the concentration of 0.4 M. Similarly, the
initial rate of the reaction of 1a (0.4 M) and 4-octyne 2a (0.1ꢀ
0.8 M) saturated when the concentration of 2a became as high as
0.6 M (Figure 2b). Furthermore, the initial reaction rates of
diphenylacetylene and 4-octyne in the saturation range were sig-
nificantly different, the former (ca. 1.5 ꢁ 10ꢀ4 M sꢀ1) being
approximately 3 times faster than the latter (ca. 5 ꢁ 10ꢀ5 M sꢀ1).
Notably, the difference of the two alkynes was even more pro-
nounced in a competitive reaction, where 1a reacted exclusively with
diphenylacetylene (Scheme 6).
Scheme 6. Alkyne Competition Experimenta
a Standard conditions: CoBr2 (0.02 M), P(3-ClC6H4)3 (0.04 M),
tBuCH2MgBr (0.2 M), pyridine (0.32 M), THF, 20 °C.
Combining the above pieces of information, we suggest two
possible catalytic cycles A and B as illustrated in Scheme 7. The
catalytic cycle A (Scheme 7a) involves a low-valent cobalt species
I generated by the reduction of the cobalt(II) precatalyst with the
Grignard reagent (vide infra) and its complexation with the
alkyne (II) prior to the rate-limiting step, that is, chelation-
assisted oxidative addition of the ortho CꢀH bond (II to III).6,7
The CꢀH bond cleavage is followed by syn-insertion of the
coordinated alkyne into the CoꢀH bond (III to IV)5,34ꢀ36 and
reductive elimination to afford the product and regenerate the
cobalt species I. The alternative catalytic cycle B (Scheme 7b) is
initiated by reversible CꢀH bond activation of the imine with the
species I, where the forward reaction is slower than the reverse
reaction, affording a cobaltacycle II0 as a transient intermediate.
The reaction proceeds in a forward direction via interception of
the intermediate II0 by the alkyne (II0 to III) followed by the
alkyne insertion and reductive elimination steps.
17290
dx.doi.org/10.1021/ja2047073 |J. Am. Chem. Soc. 2011, 133, 17283–17295

Journal of the American Chemical Society
ARTICLE
Scheme 7. Proposed Catalytic Cycles
Figure 3. Initial reaction rate against the catalyst concentration. Reaction
conditions: 1a (0.4 M), 2f (0.6 M), CoBr2 (0.008ꢀ0.032 M), THF, 20 °C,
10ꢀ15 min. CoBr2:P(3-ClC6H4)3:tBuCH2MgBr:pyridine = 1:2:10:16 (9);
CoBr2:P(3-ClC6H4)3:tBuCH2MgBr = 1:2:10, pyridine (0.32 M) (b).
during the alkyne insertion step.36a Note that the same sense of
regioselectivity has been reported for related hydroarylation and
hydroalkenylation reactions of internal alkynes by our group8,10a
and others,9,12ꢀ14,38 while the difference of the regioselectivities
for 1-phenyl-1-propyne and 1-phenyl-1-butyne (see 3ac and 3ag
in Chart 1) indicates the sensitivity of the cobalt catalyst to subtle
steric change. Not only steric but also electronic factors appear
to play important roles in the insertion step. Thus, polarization of
the putative cobalt hydride intermediate (i.e., Co(δ+)ꢀH(δꢀ))
may account for the regioselectivity observed in the reaction
of 1-fluoro-4-[(4-methoxyphenyl)ethynyl]benzene (see 6an in
Chart 3), where the CꢀH bond formation took place preferen-
tially on the acetylenic carbon that is proximal to the methox-
yphenyl group and is likely to be more electron-poor due to
inductive and resonance effects.39
Rate Dependence on Catalyst and Catalyst Components.
To shed light on the complexity of the quaternary catalytic
system, we extended the kinetic analysis to investigate the initial
rate dependence on the catalyst loading as well as on each of the
catalyst components. First, we measured the initial rate for the
reaction of 1a (0.4 M) and 2h (0.6 M) against the catalyst
concentration ranging from 0.008 to 0.032 M (2ꢀ8 mol %,
Figure 3). When the analysis was performed with the constant
ratio of the catalyst components (CoBr2:P(3-ClC6H4)3:
tBuCH2MgBr:pyridine = 1:2:10:16) as employed for the stan-
dard conditions (5 mol % CoBr2, 10 mol % P(3-ClC6H4)3, 50
mol % tBuCH2MgBr, 80 mol % pyridine), we observed a steady
increase of the initial rate with the increase of the catalyst loading
(9 in Figure 3a), while logarithm plot and linear fitting of the data
indicated a deviation from the first-order dependence (slope =
0.58, Figure 3b). In contrast, a more drastic influence of the
Both of the proposed catalytic cycles account for the important
kinetic features, (1) rate-limiting CꢀH activation of the imine in
the presence of a sufficient amount of the alkyne and (2)
saturation of the reaction rate with increase in the alkyne
concentration. At present, we cannot distinguish these two
possibilities because we have not been able to identify the nature
of the reactive species and the resting state of the catalytic
reaction (vide infra). We would rather prefer the catalytic cycle
A because of the large difference in the reaction rates of
diphenylacetylene and 4-octyne (cf., Figure 2). Thus, we spec-
ulate that the alkyne is intimately involved in the CꢀH activation
step to explicitly influence its activation barrier.37
The regioselectivities of the CꢀH bond cleavage and the CꢀC
bond formation observed in the study of the substrate scope
(Charts 1 and 3) warrant discussion in relation to the above
mechanistic proposal. The selective CꢀH bond cleavage at the
more hindered ortho positions of the meta-methoxy, halogen, and
cyano-substituted imines (see 3haꢀ3ka in Chart 1 and 6hfꢀ6kf
and 6qf in Chart 3) may be ascribed to the ability of electron-
withdrawing substituents (typically F) to strengthen a metalꢀ
carbon bond at the ortho position,26 or to the coordination of
lone-pair electrons or π-electrons to the metal center in the
CꢀH oxidative addition process.24a,25 The selective CꢀC bond
formation at the less hindered acetylenic carbon (see 3acꢀ3af
in Chart 1 and 6aq in Chart 3) may be rationalized in terms of
the preference of the cobalt center to avoid steric repulsion
17291
dx.doi.org/10.1021/ja2047073 |J. Am. Chem. Soc. 2011, 133, 17283–17295

Journal of the American Chemical Society
ARTICLE
Scheme 8. Stoichiometric Experiments
absence (Figure 4a). The reaction rate increased with the increase
of the phosphine loading from 2.5 to 10 mol % (0.01ꢀ0.04 M,
0.5ꢀ2 equiv to CoBr2) and reached the maximum with the
loading of 10 mol %. Further increase of the phosphine loading to
15ꢀ20 mol % (0.06ꢀ0.08 M, 3ꢀ4 equiv to CoBr2) caused only a
slight inhibitory effect on the reaction. For the Grignard reagent,
we determined a “threshold” loading of 20 mol % (0.08 M), with
which no reaction took place (Figure 4b). Above this threshold,
the reaction rate steadily increased with the Grignard loading of
30ꢀ50 mol % (0.12ꢀ0.20 M), and saturated with 50 mol % and
above. Note that the threshold loading (20 mol %) is higher than
required to reduce 5 mol % of CoBr2 to Co(0), that is, 10 mol %.
Pyridine is not an indispensable catalyst component, because the
reaction took place even in its absence (Figure 4c). However, the
addition of pyridine (0.08ꢀ0.32 M, 20ꢀ80 mol %) clearly
resulted in rate enhancement. The reaction rate was at its highest
(3.0 ꢁ 10ꢀ4 M sꢀ1) with 20 mol % (0.08 M) of pyridine (ca.
9 times faster than in its absence, 3.5 ꢁ 10ꢀ5 Msꢀ1). Higherloading
of pyridine (40ꢀ80 mol %, 0.16ꢀ0.32 M) slowed the reaction,
which was, however, still much faster than the pyridine-free
reaction. The inhibitoryeffect caused bya large amount ofpyridine
is consistent with the kinetic behavior observed in Figure 3a, while
we did not notice it during the reaction optimization.
While the above kinetic analysis revealed how each of the
catalyst components impacts the catalytic activity, the exact
nature of the active cobalt species I and other proposed inter-
mediates (Scheme 7) remains elusive. Attempts to gain informa-
tion from stoichiometric experiments have thus far been met
with only limited success (Scheme 8): First, a reaction of CoBr2
(0.1 mmol), P(3-ClC6H4)3 (0.2 mmol), and tBuCH2MgBr
(0.3 mmol) at 20 °C quantitatively afforded 2,2,5,5-tetramethyl-
hexane (0.104 mmol) as a result of oxidative homocoupling of
the Grignard reagent (Scheme 8a), indicating that the cobalt(II)
salt was reduced to cobalt(0). Second, a stoichiometric reaction
of 1a, CoBr2 (1 equiv), P(3-ClC6H4)3 (2 equiv), tBuCH2MgBr
(5 equiv), and pyridine (1 equiv) at 0 °C for 30 min followed by
quenching with D2O afforded a mixture of 1a with partial ortho-
deuteration, an ortho-neopentylated imine,11c and other uniden-
tified products.40 This may suggest the presence of an inter-
mediate bearing a C(ortho)ꢀCo bond, because a control
experiment in the absence of the cobalt salt expectedly resulted
in no ortho-deuteration.
Figure 4. Initial reaction rates against catalyst components ((a) P(3-
ClC6H4)3, (b) tBuCH2MgBr, (c) pyridine). Reaction conditions: 1a
(0.4 M), 2h (0.6 M), CoBr2 (0.02 M), P(3-ClC6H4)3 (0ꢀ0.08 M for
(a); 0.04 M for (b) and (c)), tBuCH2MgBr (0.08ꢀ0.32 M for (b); 0.2 M
for (a) and (c)), pyridine (0ꢀ0.32 M for (c); 0.32 M for (a) and (b)),
THF, 20 °C, 10ꢀ15 min.
catalyst loading resulted from the analysis under similar
conditions except that a fixed amount of pyridine (80 mol %,
0.32 M) was used (b in Figure 3a). Thus, as compared to the
reaction under the standard conditions (5 mol %), the reaction
became exceedingly faster (more than twice) with 8 mol %
catalyst loading while it almost stopped with 2 mol % loading.
These data suggest that an excessive amount of pyridine is
detrimental to the reaction (vide infra).
Given the anomalous kinetic behavior observed in the above
experiments, we became interested in the influence of each
catalyst component on the catalytic activity. Thus, the initial rate
was measured for the reaction of 1a (0.4 M) and 2h (0.6 M)
under the standard conditions (5 mol % CoBr2, 10 mol % P(3-
ClC6H4)3, 50 mol % tBuCH2MgBr, 80 mol % pyridine) except
for a change in the amount of either P(3-ClC6H4)3 (0ꢀ20 mol
%), tBuCH2MgBr (20ꢀ80 mol %), or pyridine (0ꢀ80 mol %)
(Figure 4aꢀc). The phosphine ligand P(3-ClC6H4)3 is evidently
essential for the reaction, because no reaction took place in its
At this stage, we put forth some speculations on the nature of
the catalytically active species and the role of each of the catalyst
components: (1) The active species should bear one or two
molecules of the phosphine ligand because no reaction took
place without the ligand and because the reaction rate reached
the maximum with 10 mol % (2 equiv to Co) of the phosphine
loading (Figure 4a). (2) The Grignard reagentmay not onlyreduce
17292
dx.doi.org/10.1021/ja2047073 |J. Am. Chem. Soc. 2011, 133, 17283–17295

Journal of the American Chemical Society
ARTICLE
cobalt(II) to cobalt(0) but also give rise to an organocobalt(0)ate
species as the catalytically active species, because the alkyl group of
the Grignard reagent significantly influences the catalytic activity
(Table 2) and because an excess amount of the Grignard reagent
is necessary to achieve catalytic turnover (Figure 4b). Note that
organocobalt(0)ate species have been proposed as reactive
species for several cobalt-catalyzed reactions that take place
in the presence ofexcess Grignard reagents.41 (3) Pyridine appears
to serve as a coligand for the cobalt catalyst, while the origin of
its rate enhancement remains unclear. The rate inhibition with
higher loading of pyridine (Figures 3a and 4c) may be ascribed to
competitive binding of pyridine and the imine substrate to the
catalyst.
’ REFERENCES
(1) Murai, S.; Kakiuchi, F.; Sekine, S.; Tanaka, Y.; Kamatani, A.;
Sonoda, M.; Chatani, N. Nature 1993, 36, 529–531.
(2) For recent reviews, see: (a) Wencel-Delord, J.; Dr€oge, T.; Liu, F.;
Glorius, F. Chem. Soc. Rev. 2011, 40, 4740–4761. (b) Colby, D. A.;
Bergman, R. G.; Ellman, J. A. Chem. Rev. 2010, 110, 624–655. (c) Lyons,
T. W.; Sanford, M. S. Chem. Rev. 2010, 110, 1147–1169. (d) Sun, C.-L.;
Li, B.-J.; Shi, Z.-J. Chem. Commun. 2010, 46, 677–685. (e) Ackermann,
L.; Vicente, R.; Kapdi, A. R. Angew. Chem., Int. Ed. 2009, 48, 9792–9827.
(f) Chen, X.; Engle, K. M.; Wang, D.-H.; Yu, J.-Q. Angew. Chem., Int. Ed.
2009, 48, 5094–5115. (g) Kakiuchi, F.; Kochi, T. Synthesis 2008, 3013–
3039. (h) Alberico, D.; Scott, M. E.; Lautens, M. Chem. Rev. 2007, 107,
174–238.
(3) (a) Kulkarni, A. A.; Daugulis, O. Synthesis 2009, 4087–4109.
(b) Nakamura, E.; Yoshikai, N. J. Org. Chem. 2010, 75, 6061–6067.
(c) Sun, C.-L.; Li, B.-J.; Shi, Z.-J. Chem. Rev. 2011, 111, 1293–1314.
(d) Nakamura, E.; Sato, K. Nat. Mater. 2011, 10, 158–161.
’ CONCLUSION
We have reported in this Article that a quaternary catalytic
system consisting of CoBr2, P(3-ClC6H4)3, tBuCH2MgBr, and
pyridine promotes hydroarylation of an internal alkyne with an
aromatic imine through chelation-assisted CꢀH bond cleavage.
The reaction takes place smoothly at room temperature to afford
the alkenylated imine in moderate to good yield with good regio-
and stereoselectivities. The hydroarylation products were, de-
pending on the substituents of the alkynes, transformed to the
corresponding benzofulvene or ketone derivatives under acidic
conditions. Deuterium-labeling experiments and kinetic analysis
suggested plausible mechanistic scenarios that involve oxidative
addition of the ortho CꢀH bond of the imine as the rate-limiting
step, while the role of the alkyne in the catalytic cycle remains
debatable and warrants further investigation. The kinetic analysis
also shed some light on the high level of complexity of the
quaternary catalytic system. The secondary directing effect of
ether, halogen, and cyano groups on the CꢀH bond cleavage
step is also an intriguing subject of further mechanistic studies
and may offer unique synthetic approaches to polysubstituted
aromatic compounds. On the basis of the present study as well as
recent studies of our group and others,8,10,11 we envision that a
wider variety of productive catalytic systems involving cobalt-
mediated CꢀH bond activation could be developed.
(4) For examples of chelation-assisted CꢀH functionalization reac-
tions using first-row transition metal catalysts, see: (a) Murahashi, S.
J. Am. Chem. Soc. 1955, 77, 6403. (b) Murahashi, S.; Horiie, S. J. Am.
Chem. Soc. 1956, 78, 4816. (c) Halbritter, G.; Knoch, F.; Wolski, A.;
Kisch, H. Angew. Chem., Int. Ed. Engl. 1994, 33, 1603–1605. (d) Funk,
J. K.; Yennawar, H.; Sen, A. Helv. Chim. Acta 2006, 89, 1687–1695.
(e) Chen, X.; Hao, X.-S.; Goodhue, C. E.; Yu, J.-Q. J. Am. Chem. Soc.
2006, 128, 6790–6791. (f) Kuninobu, Y.; Nishina, Y.; Takeuchi, T.;
Takai, K. Angew. Chem., Int. Ed. 2007, 46, 6518–6520. (g) Norinder, J.;
Matsumoto, A.; Yoshikai, N.; Nakamura, E. J. Am. Chem. Soc. 2008,
130, 5858–5859. (h) Yoshikai, N.; Matsumoto, A.; Norinder, J.;
Nakamura, E. Angew. Chem., Int. Ed. 2009, 48, 2925–2928. (i) Yoshikai,
N.; Matsumoto, A.; Norinder, J.; Nakamura, E. Synlett 2010, 313–316. (j)
Ilies, L.; Asako, S.; Nakamura, E. J. Am. Chem. Soc. 2011, 133, 7672–7675.
(5) Lenges, C. P.; Brookhart, M.; Grant, B. E. J. Organomet. Chem.
1997, 528, 199–203.
(6) (a) Klein, H.-F.; Beck, R.; Flo€rke, U.; Haupt, H.-J. Eur. J. Inorg.
Chem. 2002, 3305–3312. (b) Klein, H.-F.; Beck, R.; Fl€orke, U.; Haupt,
H.-J. Eur. J. Inorg. Chem. 2003, 1380–1387. (c) Klein, H.-F.; Camadanli,
S.; Beck, R.; Leudel, D.; Fl€orke, U. Angew. Chem., Int. Ed. 2005,
44, 975–977. (d) Beck, R.; Sun, H.; Li, X.; Camadanli, S.; Klein, H.-F.
Eur. J. Inorg. Chem. 2008, 3253–3257. (e) Beck, R.; Frey, M.; Camadanli,
S.; Klein, H.-F. Dalton Trans. 2008, 4981–4983. (f) Camadanli, S.; Beck,
R.; Fl€orke, U.; Klein, H.-F. Dalton Trans. 2008, 5701–5704. (g) Wang,
A.; Sun, H.; Li, X. Organometallics 2008, 27, 5434–5437.
(7) For a recent review on cyclometalation, see: Albrecht, M. Chem.
Rev. 2010, 110, 576–623.
’ ASSOCIATED CONTENT
(8) Gao, K.; Lee, P.-S.; Fujita, T.; Yoshikai, N. J. Am. Chem. Soc.
2010, 132, 12249–12251.
S
Supporting Information. Experimental procedures and
b
characterization of new compounds. This material is available
free of charge via the Internet at http://pubs.acs.org.
(9) For CꢀH addition reactions to unsaturated compounds using
first-row transition metal catalysts without chelation assistance, see:
(a) Lenges, C. P.; Brookhart, M. J. Am. Chem. Soc. 1997, 119, 3165–
3166. (b) Lenges, C. P.; White, P. S.; Brookhart, M. J. Am. Chem. Soc.
1998, 120, 6965–6979. (c) Nakao, Y.; Kanyiva, K. S.; Oda, S.; Hiyama,
T. J. Am. Chem. Soc. 2006, 128, 8146–8147. (d) Nakao, Y.; Kanyiva,
K. S.; Hiyama, T. J. Am. Chem. Soc. 2008, 130, 2448–2449. (e) Nakao, Y.;
Kashihara, N.; Kanyiva, K. S.; Hiyama, T. J. Am. Chem. Soc. 2008,
130, 16170–16171. (f) Nakao, Y.; Idei, H.; Kanyiva, K. S.; Hiyama, T.
J. Am. Chem. Soc. 2009, 131, 5070–5071. (g) Nakao, Y.; Idei, H.;
Kanyiva, K. S.; Hiyama, T. J. Am. Chem. Soc. 2009, 131, 15996–15997.
(h) Mukai, T.; Hirano, K.; Satoh, T.; Miura, M. J. Org. Chem. 2009,
74, 6410–6413. (i) Tsai, C.-C.; Shih, W.-C.; Fang, C.-H.; Li, C.-Y.;
Ong, T.-G.; Yap, G. P. A. J. Am. Chem. Soc. 2010, 132, 11887–11889.
(j) Nakao, Y.; Yamada, Y.; Kashihara, N.; Hiyama, T. J. Am. Chem. Soc.
2010, 132, 13666–13668.
’ AUTHOR INFORMATION
Corresponding Author
nyoshikai@ntu.edu.sg
Present Addresses
†Department of Chemistry, Graduate School of Pure and
Applied Sciences, University of Tsukuba, Tsukuba, Ibaraki
305-8571, Japan.
’ ACKNOWLEDGMENT
We thank the Singapore National Research Foundation (NRF-
RF2009-05 to N.Y.) and Nanyang Technological University for
generous financial support, and Takeshi Yamakawa for technical
assistance. We appreciate the valuable comments and suggestions
provided by the reviewers during the reviewing process.
(10) (a) Ding, Z.; Yoshikai, N. Org. Lett. 2010, 12, 4180–4183.
(b) Gao, K.; Yoshikai, N. J. Am. Chem. Soc. 2011, 133, 400–402. (c)
Yoshikai, N. Synlett 2011, 1047–1051. (d) Gao, K.; Yoshikai, N. Angew.
Chem., Int. Ed. 2011, 50, 6888–6892. (e) Ding, Z.; Yoshikai, N. Synthesis
2011, 2561–2566.
17293
dx.doi.org/10.1021/ja2047073 |J. Am. Chem. Soc. 2011, 133, 17283–17295

Journal of the American Chemical Society
ARTICLE
(11) (a) Chen, Q.; Ilies, L.; Nakamura, E. J. Am. Chem. Soc. 2011,
133, 428–429. (b) Ilies, L.; Chen, Q.; Zeng, X.; Nakamura, E. J. Am.
Chem. Soc. 2011, 133, 5221–5223. (c) Li, B.; Wu, Z.-H.; Gu, Y.-F.; Sun,
C.-L.; Wang, B.-Q.; Shi, Z.-J. Angew. Chem., Int. Ed. 2011, 50,
1109–1113. (d) Kim, J. Y.; Cho, S. H.; Joseph, J.; Chang, S. Angew.
Chem., Int. Ed. 2010, 49, 9899–9903. (e) Chen, Q.; Ilies, L.; Yoshikai, N.;
Nakamura, E. Org. Lett. 2011, 13, 3232–3234.
(12) Rh catalysis: (a) D€urr, U.; Kisch, H. Synlett 1997, 1335–1341.
(b) Lim, Y.-G.; Lee, K.-H.; Koo, B. T.; Kang, J.-B. Tetrahedron Lett. 2001,
42, 7609–7612. (c) Lim, S.-G.; Jun, H. L.; Moon, C. W.; Hong, J.-B.; Jun,
C.-H. Org. Lett. 2003, 5, 2759–2761. (d) Cheng, K.; Yao, B.; Zhao, J.;
Zhang, Y. Org. Lett. 2008, 10, 5309–5312. (e) Katagiri, T.; Mukai, T.;
Satoh, T.; Hirano, K.; Miura, M. Chem. Lett. 2009, 38, 118–119.
(f) Schipper, D. J.; Hutchinson, M.; Fagnou, K. J. Am. Chem. Soc. 2010,
132, 6910–6911.
(13) Ru catalysis: (a) Kakiuchi, F.; Yamamoto, Y.; Chatani, N.;
Murai, S. Chem. Lett. 1995, 681–682. (b) Kakiuchi, F.; Sato, T.;
Tsujimoto, T.; Yamauchi, M.; Chatani, N.; Murai, S. Chem. Lett.
1998, 1053–1054. (c) Kakiuchi, F.; Uetsuhara, T.; Tanaka, Y.; Chatani,
N.; Murai, S. J. Mol. Catal. A: Chem. 2002, 182ꢀ183, 511–514.
(14) Ir catalysis: (a) Satoh, T.; Nishinaka, Y.; Miura, M.; Nomura, M.
Chem. Lett. 1999, 615–616. (b) Tsuchikama, K.; Kasagawa, M.; Hashimoto,
Y.; Endo, K.; Shibata, T. J. Organomet. Chem. 2008, 693, 3939–3942.
(c) Tsuchikama, K.; Kasagawa, M.; Endo, K.; Shibata, T. Synlett 2010,
97–100.
(21) Benzofulvene derivatives have been reported to serve as
monomers for facile spontaneous polymerization to afford polymers
of unique chemical/physical properties: Cappelli, A.; Galeazzi, S.;
Giuliani, G.; Anzini, M.; Donati, A.; Zetta, L.; Mendichi, R.; Aggravi, M.;
Giorgi, G.; Paccagnini, E.; Vomero, S. Macromolecules 2007, 40,
3005ꢀ3014 and references cited therein.
(22) The E-stereochemistry of the olefin moiety of the major isomer
was confirmed by NOESY experiments. Upon reduction with NaBH4,
the major and minor isomers converged into a single product, which
confirmed that the formation of the minor isomer was due to E/Z
isomerization of the imine moiety rather than that of the olefin moiety
(see the Supporting Information).
(23) For cyclotrimerization of alkynes catalyzed by low-valent cobalt
complexes, see: (a) Chopade, P. R.; Louie, J. Adv. Synth. Catal. 2006,
348, 2307–2327. (b) Kohta, S.; Brahmachary, E.; Lahiri, K. Eur. J. Org.
Chem. 2005, 4741–4767. (c) Saito, S.; Yamamoto, Y. Chem. Rev. 2000,
100, 2901–2915. (d) Vollhardt, K. P. C. Angew. Chem., Int. Ed. Engl.
1984, 23, 539–556.
(24) (a) Sonoda, M.; Kakiuchi, F.; Chatani, N.; Murai, S. Bull. Chem.
Soc. Jpn. 1997, 70, 3117–3128. (b) Martinez, R.; Genet, J.-P.; Darses, S.
Chem. Commun. 2008, 3855–3857. (c) Martinez, R.; Simon, M.-O.;
Chevalier, R.; Pautigny, C.; Genet, J.-P.; Darses, S. J. Am. Chem. Soc.
2009, 131, 7887–7895. (d) Ackermann, L.; Vicente, R.; Potukuchi,
H. K.; Pirovano, V. Org. Lett. 2010, 12, 5032–5035. (e) Ackermann, L.;
Hofmann, N.; Vicente, R. Org. Lett. 2011, 13, 1875–1877.
(25) A cyano group was reported to serve by itself as a directing
group for CꢀH bond functionalization: Kakiuchi, F.; Sonoda, M.;
Tsujimoto, T.; Chatani, N.; Murai, S. Chem. Lett. 1999, 1083–1084.
(26) (a) Clot, E.; Mꢀegret, C.; Eisenstein, O.; Perutz, R. N. J. Am.
Chem. Soc. 2009, 131, 7817–7827. (b) Clot, E.; Eisenstein, O.; Jasim, N.;
Macgregor, S. A.; McGrady, J. E.; Perutz, R. N. Acc. Chem. Res. 2011, 44,
333ꢀ348 and references cited therein.
(15) Re catalysis: Kuninobu, Y.; Kikuchi, K.; Tokunaga, Y.; Takai, K.
Tetrahedron 2008, 64, 5974–5981.
(16) A small portion of this work has been communicated in ref 8.
(17) Rh(III)-catalyzed ortho-alkenylation followed by (oxidative)
cyclization has been extensively studied by MiuraꢀSatoh, Fagnou, and
others. For a review and recent examples, see: (a) Satoh, T.; Miura, M.
Chem.-Eur. J. 2010, 16, 11212–11222. (b) Morimoto, K.; Hirano, K.;
Satoh, T.; Miura, M. Org. Lett. 2010, 12, 2068–2071. (c) Guimond, N.;
Gouliaras, C.; Fagnou, K. J. Am. Chem. Soc. 2010, 132, 6908–6909.
(d) Hyster, T. K.; Rovis, T. J. Am. Chem. Soc. 2010, 132, 10565–10569.
(e) Too, P. C.; Wang, Y.-F.; Chiba, S. Org. Lett. 2010, 12, 5688–5691.
(f) Stuart, D.; Alsabeh, P.; Kuhn, M.; Fagnou, K. J. Am. Chem. Soc. 2010,
132, 18326–18339. (g) Patureau, F. W.; Besset, T.; Kuhl, N.; Glorius, F.
J. Am. Chem. Soc. 2011, 133, 2154–2156. (h) Zhang, X.; Chen, D.; Zhao,
M.; Zhao, J.; Jia, A.; Li, X. Adv. Synth. Catal. 2011, 353, 719–723.
(i) Muralirajan, K.; Parthasarathy, K.; Cheng, C.-H. Angew. Chem., Int.
Ed. 2011, 50, 4169–4172. (j) Wang, Y.-F.; Toh, K. K.; Lee, J.-Y.; Chiba,
S. Angew. Chem., Int. Ed. 2011, 50, 6049–6053. (k) Hyster, T. K.; Rovis,
T. Chem. Sci. 2011, 2, 1606–1610.
(27) The stereochemistry of the CdC double bond of 4fa and
related benzofulvene derivatives could not be unambiguously deter-
mined by NOESY experiments.
(28) The reaction also afforded a mixture of intractable amino
compounds. A mixture of aminoindenene derivatives shown below
was suggested by GCꢀMS and 1H NMR analysis.
(18) Recatalyzed ortho-alkenylation of aromatic aldimines followed
by cyclization to afford indene derivatives has been reported by
Kuninobu, Takai, and co-workers: (a) Kuninobu, Y.; Kawata, A.; Takai,
K. J. Am. Chem. Soc. 2005, 127, 13498–13499. (b) Kuninobu, Y.;
Tokunaga, Y.; Kawata, A.; Takai, K. J. Am. Chem. Soc. 2006, 128,
202–209.
(19) For ortho CꢀH functionalization of imines except for alkenyla-
tion with alkynes, see ref 2 (and references cited therein) and the
following recent examples: (a) Tredwell, M. J.; Gulias, M.; Bremeyer,
N. G.; Johanson, C. C. C.; Collins, B. S. L.; Gaunt, M. J. Angew. Chem.,
Int. Ed. 2011, 50, 1076–1079. (b) Kuninobu, Y.; Nakahara, T.; Yu, P.;
Takai, K. J. Organomet. Chem. 2011, 696, 348–351. (c) Kim, M.; Kwak,
J.; Chang, S. Angew. Chem., Int. Ed. 2009, 48, 8935–8939. (d) Kuninobu,
Y.; Nishina, Y.; Okaguchi, K.; Shouno, M.; Takai, K. Bull. Chem. Soc. Jpn.
2008, 81, 1393–1401.
(20) For recent reports on the synthesis of benzofulvenes, see refs
14c 17g, and the following papers: (a) Bryan, C. S.; Lautens, M. Org. Lett.
2010, 12, 2754–2757. (b) Ye, S.; Yang, X.; Wu, J. Chem. Commun. 2010,
46, 2950–2952. (c) Ye, S.; Gao, K.; Zhou, H.; Yang, X.; Wu, J. Chem.
Commun. 2009, 5406–5408. (d) Schmittel, M.; Vavilala, C. J. Org. Chem.
2005, 70, 4865–4868. (e) Kovalenko, S. V.; Peabody, S.; Manoharan,
M.; Clark, R. J.; Alabugin, I. V. Org. Lett. 2004, 6, 2457–2460. (f) Bekele,
T.; Christian, C. F.; Lipton, M. A.; Singleton, D. A. J. Am. Chem. Soc.
2005, 127, 9216–9223.
(29) For an example of biological activities of aminoindanol deri-
vatives, see: Wu, D.; Pontillo, J.; Ching, B.; Hudson, S.; Gao, Y.; Fleck, B. A.;
Gogas, K.; Wade, W. S. Bioorg. Med. Chem. Lett. 2008, 18, 4224–4227.
(30) Upon concentration and standing at ambient temperature, the
benzofulvene derivatives underwent polymerization relatively easily (cf.,
ref 21).
(31) Diheteroaryl alkynes such as bis(2-thienyl)acetylene failed to
participate in the reaction.
(32) The observation of the primary H/D KIE for the competitive
reaction is consistent with our previous study on the reaction of
2-phenylpyridine (ref 8, KIE = 2.1), but is in stark contrast to the
observation of Shi and co-workers on the cobalt-catalyzed direct
arylation of benzo[h]quinoline (KIE = 1.04, ref 11c).
(33) (a) Vo, L. K.; Singleton, D. A. Org. Lett. 2004, 6, 2469–2472.
(b) Yoshikai, N.; Matsuda, H.; Nakamura, E. J. Am. Chem. Soc. 2008,
130, 15258–15259.
(34) Otsuka, S.; Nakamura, A. Adv. Organomet. Chem. 1976, 14,
245–283.
(35) (a) Schrauzer, G. N.; Windgassen, R. J. J. Am. Chem. Soc. 1967,
89, 1999–2007. (b) Yamazaki, H.; Hagihara, N. J. Organomet. Chem.
1970, 21, 431–443.
17294
dx.doi.org/10.1021/ja2047073 |J. Am. Chem. Soc. 2011, 133, 17283–17295

Journal of the American Chemical Society
ARTICLE
(36) (a) Foo, T.; Bergman, R. G. Organometallics 1992,
11, 1811–1819. (b) Bassetti, M.; Casellato, P.; Gamasa, M. P.; Gimeno,
J.; Gonzꢀalez-Bernardo, C.; Martín-Vaca, B. Organometallics 1997,
16, 5470–5477. (c) Bassetti, M.; Marini, S.; Díaz, J.; Gamasa, M. P.;
ꢀ
Gimeno, J.; Rodríguez-Alvarez, Y.; García-Granda, S. Organometallics
2002, 21, 4815–4822.
(37) We also feel it reasonable to assume that the alkyne serves as a
good π-accepting ligand for the low-valent cobalt species I and are
tempted to interpret the result of the alkyne competition experiment
(Scheme 6) in terms of better π-accepting ability of diphenylacetylene
than of 4-octyne.
(38) (a) Colby, D. A.; Bergman, R. G.; Ellman, J. A. J. Am. Chem. Soc.
2008, 130, 3645–3651. (b) Parthasarathy, K.; Jeganmohan, M.; Cheng,
C.-H. Org. Lett. 2008, 10, 325–328.
(39) NPA charges of the acetylenic carbon atoms of 1-fluoro-4-[(4-
methoxyphenyl)ethynyl]benzenewere calculatedto be+0.012 (proximal
to 4-methoxyphenyl) and ꢀ0.008 (proximal to 4-fluorophenyl) at the
B3LYP/6-31G* level.
(40) A prolonged reaction time (3 h) resulted in deprotonation of
the α-position of the imine, as indicated from deuterium incorporation
into the α-position.
(41) (a) Ohmiya, H.; Wakabayashi, K.; Yorimitsu, H.; Oshima, K.
Tetrahedron 2006, 62, 2207–2213. (b) Wakabayashi, K.; Yorimitsu, H.;
Oshima, K. J. Am. Chem. Soc. 2001, 123, 5374–5375. (c) Czaplik, W. M.;
Mayer, M.; Jacobi von Wangelin, A. Synlett 2009, 2931–2934.
(d) Kauffmann, T. Angew. Chem., Int. Ed. 1996, 35, 386–403.
17295
dx.doi.org/10.1021/ja2047073 |J. Am. Chem. Soc. 2011, 133, 17283–17295