Z -selective alkene isomerization by high-spin cobalt(II) complexes

Article abstract of DOI:10.1021/ja408238n

The isomerization of simple terminal alkenes to internal isomers with Z-stereochemistry is rare, because the more stable E-isomers are typically formed. We show here that cobalt(II) catalysts supported by bulky β-diketiminate ligands have the appropriate kinetic selectivity to catalyze the isomerization of some simple 1-alkenes specifically to the 2-alkene as the less stable Z-isomer. The catalysis proceeds via an "alkyl" mechanism, with a three-coordinate cobalt(II) alkyl complex as the resting state. β-Hydride elimination and [1,2]-insertion steps are both rapid, as shown by isotopic labeling experiments. A steric model explains the selectivity through a square-planar geometry at cobalt(II) in the transition state for β-hydride elimination. The catalyst works not only with simple alkenes, but also with homoallyl silanes, ketals, and silyl ethers. Isolation of cobalt(I) or cobalt(II) products from reactions with poor substrates suggests that the key catalyst decomposition pathways are bimolecular, and lowering the catalyst concentration often improves the selectivity. In addition to a potentially useful, selective transformation, these studies provide a mechanistic understanding for catalytic alkene isomerization by high-spin cobalt complexes, and demonstrate the effectiveness of steric bulk in controlling the stereoselectivity of alkene formation.

Full text of DOI:10.1021/ja408238n

Article
pubs.acs.org/JACS
Z‑Selective Alkene Isomerization by High-Spin Cobalt(II) Complexes
Chi Chen,†,‡ Thomas R. Dugan,† William W. Brennessel,† Daniel J. Weix,*,† and Patrick L. Holland*,†,‡
†Department of Chemistry, University of Rochester, 120 Trustee Road, Rochester, New York 14627, United States
‡Department of Chemistry, Yale University, 225 Prospect Street, New Haven, Connecticut 06520, United States
S
* Supporting Information
ABSTRACT: The isomerization of simple terminal alkenes to
internal isomers with Z-stereochemistry is rare, because the more
stable E-isomers are typically formed. We show here that cobalt(II)
catalysts supported by bulky β-diketiminate ligands have the
appropriate kinetic selectivity to catalyze the isomerization of
some simple 1-alkenes specifically to the 2-alkene as the less stable
Z-isomer. The catalysis proceeds via an “alkyl” mechanism, with a
three-coordinate cobalt(II) alkyl complex as the resting state. β-
Hydride elimination and [1,2]-insertion steps are both rapid, as
shown by isotopic labeling experiments. A steric model explains the
selectivity through a square-planar geometry at cobalt(II) in the transition state for β-hydride elimination. The catalyst works not
only with simple alkenes, but also with homoallyl silanes, ketals, and silyl ethers. Isolation of cobalt(I) or cobalt(II) products from
reactions with poor substrates suggests that the key catalyst decomposition pathways are bimolecular, and lowering the catalyst
concentration often improves the selectivity. In addition to a potentially useful, selective transformation, these studies provide a
mechanistic understanding for catalytic alkene isomerization by high-spin cobalt complexes, and demonstrate the effectiveness of
steric bulk in controlling the stereoselectivity of alkene formation.
Although the isomerization of terminal alkenes to internal
alkenes is thermodynamically favorable, catalytic reactions often
give mixtures of internal alkenes.12−24 Further, the products are
typically mixtures of the E- and Z-isomers in the ratio expected
from the relative free energies (for simple alkenes, this ratio is
generally greater than 3:1 favoring the E-isomer).12,25−28
Several catalysts are known for selective formation of the E-
isomer with very good selectivity.14,17,18,29−31 On the other
hand, selectivity for the less thermodynamically stable internal
isomer is rare. Some alkene isomerization catalysts form the Z-
isomers at a very early stage, but the mixture quickly converges
to the predominantly E thermodynamic ratio.32 One successful
strategy for overcoming thermodynamically dominated selec-
tivity has been to use a directing group, for example in the
isomerization of 2-allylphenol to its Z-2-isomer with Ru(cot)-
(cod)/PEt3,33 and the isomerization of 1,3-dienes to 2Z,4E-
dienes with CoBr2(dpppMe2).34 A few homogeneous catalysts
selectively produce the E-isomer of vinyl ethers, amines, and
acetates, which is the less stable isomer.14,30,35−37 However,
selectivity for the less-stable Z-isomer of simple olefins remains
an unsolved challenge. We report here the selective isomer-
ization of normal α-olefins to Z-2-olefins through the action of
a sterically demanding cobalt(II) catalyst.
INTRODUCTION
■
Alkenes are key functional groups in many natural and
industrial products, as well as precursors for other types of
organic molecules. Among the methods available for the
formation of alkenes, relatively few allow direct access to the
thermodynamically less-stable Z-isomer, even though Z-alkenes
are of great use in synthesis and are commonly found in natural
products.1 Common ways to access Z-alkenes include the
Wittig olefination,2,3 the Still−Gennari modification to the
Horner−Wadsworth−Emmons olefination,4 the cross-coupling
of Z-vinyl halides or Z-vinyl organometallic reagents,5 reduction
of alkynes over a poisoned catalyst,6 and, recently, Z-selective
olefin metathesis.7−9 An approach that has not yet met with
great success is selective alkene isomerization (Figure 1),10,11
even though isomerization is an atom-economic organic
transformation that has many applications in industry.11
The mechanisms of metal-catalyzed alkene isomerizations
have been studied in great detail.38 The two most common
mechanisms are termed alkyl and allyl mechanisms (Scheme 1).
In the alkyl mechanism (Scheme 1a), sequential β-hydride
Received: August 8, 2013
Figure 1. Approaches to Z-olefins.
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja408238n | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Scheme 1. Mechanisms of Catalytic Alkene Isomerization
Figure 2. Catalysts used in this study. The cobalt(II) catalyst 1 is an
effective and selective catalyst that shows the hallmarks of a
homogeneous catalyst.
they show exceptional selectivity: they are selective for 2-
alkenes, and specifically form the Z-isomer.
RESULTS
■
Synthesis and Characterization of Catalysts. As
previously reported, β-diketiminate-supported cobalt(II) alkyl
complexes can be synthesized by reacting LtBuCoCl46 with
Grignard reagents.43,44 In order to provide a suitable catalyst
with β-hydrogen atoms on the alkyl ligand, we prepared
LtBuCo(n-hexyl) (1) from the corresponding Grignard reagent.
This compound is extremely sensitive to air and to moisture,
and its characterization and catalytic reactions were studied
under rigorously dry conditions under argon or N2. The
molecular structure of 1 was determined by X-ray crystallog-
raphy (Figure 3). The Co−N bond distances and the N−Co−
elimination and [1,2]-insertion cause migration of the metal,
and the oxidation state of the metal is constant. In contrast, the
allyl mechanism (Scheme 1b) involves the reversible oxidative
addition of allylic C−H bonds.
In this work, we describe the isomerization of alkenes using a
paramagnetic cobalt(II) catalyst that has a high-spin electronic
configuration. High-spin paramagnetic complexes present
challenges for study, because these complexes do not follow
the “18-electron rule,” and the NMR spectra of the complexes
contain broadened and shifted resonances. Despite these issues,
a number of groups have discovered interesting and distinctive
reactivity from high-spin complexes.39,40 In a relevant example,
we have shown that three-coordinate β-diketiminate iron(II)
alkyl complexes perform reversible β-H elimination followed by
alkene exchange, which results in transfer hydrogenation and/
or isomerization of alkyl ligands.41,42 These transformations
comprise the first two steps of the alkyl mechanism for alkene
isomerization, but preliminary investigations indicated that the
iron(II) complexes were very slow isomerization catalysts.
We have also reported three-coordinate cobalt(II) alkyl
complexes LtBuCoR (R = methyl, cyclohexyl) supported by the
b u l ky l i g a n d 2 , 2 , 6 , 6 - t e t r am e t h y l - 3, 5 - b i s ( 2 , 6-
diisopropylphenylimido)heptyl (LtBu, see Figure 2).43−45
These complexes were characterized by a range of structural
and spectroscopic techniques, showing that they have a high-
spin d7 (S = 3/2) electronic configuration. Here, we report that
the related complex LtBuCo(n-hexyl) mediates efficient alkene
isomerization through an alkyl mechanism, and we evaluate the
scope and mechanism of the reaction as well as mechanisms of
catalyst decomposition. These studies are important because
Figure 3. ORTEP drawing of the molecular structure of 1. Thermal
ellipsoids are shown at the 50% probability level. Hydrogen atoms
have been omitted for clarity. Important bond distances (Å) and
angles (deg): Co−N 1.950(2), 1.960(2), Co−C 2.011(2); N−Co−N
97.69(7), N−Co−C 120.60(9), 141.26(9).
N angle are similar to those in planar three-coordinate β-
diketiminate cobalt(II) complexes like LtBuCoCl and
LtBuCoCH3.43 The Co−C bond distance of 2.011(2) Å is
longer than that in LtBuCoCH3 (1.963(3) Å), probably due to
the added steric hindrance. The 1H NMR spectrum of 1 shows
resonances that are paramagnetically shifted and lie over a
range of more than 300 ppm. Resonances in the β-diketiminate
ligand can be assigned on the basis of integration to the seven
unique proton environments, with chemical shifts similar to
those of LtBuCoCH3.43 Therefore, the solution symmetry is C2v,
indicating that rotation around the Co−C bond is rapid on the
NMR time scale. Additionally, resonances corresponding to
B
dx.doi.org/10.1021/ja408238n | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
protons of the hexyl ligand were evident, corresponding to all
except the two closest to the paramagnetic cobalt(II) center.
Catalysis was also tested using a previously reported high-
spin cobalt(I) complex LtBuCo (2), in which the supporting β-
diketiminate ligand has κ1:η6-binding to the metal (Figure 2).47
Despite the unusual binding mode, this complex is known to
react through pathways that rearrange the supporting ligand
back to the traditional κ2 binding, with coordination of
additional donors.
Isomerization of 1-Alkenes to 2-Alkenes. We began
with isomerization of 1-hexene to 2-hexene in benzene solution,
with mesitylene as an internal standard for gas chromatography
(GC) (Scheme 2). Initial reactions used 1 M 1-hexene and 1%
Table 2. Solvent Dependence of 1-Hexene Isomerization
with Catalyst 1
a
b
b
entry
solvent
yield
E:Z
1
2
3
4
benzene
toluene
THF
88%
83%
84%
86%
1:3.9
1:3.2
1:2.1
1:0.9
nonane
a
Reaction conditions: 1 equiv mesitylene, 0.5 mol % catalyst, 80 °C,
b
12 h. Yield of products and E:Z ratios determined by GC analysis.
Table 3. Variation of Conditions for Isomerization of 1-
a
Hexene with Catalyst 1
T
catalyst loading
(mol %)
[1-hexene]
(mol/L)
b
b
Scheme 2. Cobalt-Catalyzed Isomerization of 1-Hexene
entry (°C)
yield
6%
E:Z
1
2
3
4
5
6
7
8
9
25
45
60
80
80
80
80
80
80
80
0.5
0.5
0.5
0.5
1.0
5.0
0.5
0.5
0.5
0.5
1.0
1.0
1.0
1.0
1.0
1.0
0.1
0.5
1.0
1.0
all Z
1:4.9
1:5.7
1:5.3
1:2.8
1:0.3
1:5.7
1:5.6
1:4.5
1:4.0
11%
31%
88%
86%
90%
47%
72%
86%
84%
loading of cobalt catalyst at 80 °C for 24 h, and the progress of
1
the reactions was monitored by GC and verified by H NMR
spectroscopy. Both catalysts gave regioselective isomerization
to 2-hexenes, and <5% of 3-hexenes were observed. Therefore,
the isomerization does not reach thermodynamic equilibrium.
Interestingly, the selectivity is very different for the two
catalysts (Table 1). Catalysis by 2 gives E:Z ratios near the
c
d
10
a
The reaction solvent is benzene, and the reaction time is 12 h unless
otherwise specified. Yield of products and E:Z ratios determined by
GC analysis. Added 2 drops of Hg. Added 0.025 mol % PPh3.
b
c
d
Table 1. Isomerization of 1-Hexene to E- and Z-2-Hexenes
a
with 1 and 2
temperature, but temperature did not have a significant impact
on the selectivity up to 80 °C (entries 1−4 in Table 3). The
catalyst loading had a significant influence on the Z-selectivity,
with higher selectivity at lower loading (entries 4, 5, and 6 in
Table 3). The concentration of substrate did not affect
selectivity and reactivity significantly (entries 4, 7, and 8 in
Table 3).
Scope: Aliphatic and Aromatic Alkenes. To investigate
the scope of the isomerization reaction catalyzed by LtBuCo(n-
hexyl), we tested other terminal alkenes.
b
b
catalyst
time (h)
yield
E:Z
1
1
1
2
2
2
6
63%
88%
88%
36%
50%
62%
1:6.9
1:4.2
1:2.5
12
24
6
4.1:1
12
24
4.0:1
3.8:1
a
b
1.0 mol % catalyst loading, 80 °C, 1.0 M substrate. Yields of
products and E:Z ratios were determined by GC analysis.
The cobalt-catalyzed isomerization reaction as shown above
was effective for aliphatic alkenes, even for substrates that form
a bulky trisubstituted alkene products. 1-Octene and 1-
dodecene give even better discrimination between E- and Z-
products, but gave products from more extensive isomerization
of the double bond to other internal positions. The yields for
these substrates in Table 4 reflect only the 2-isomers; other
internal isomers could not be distinguished from one another
thermodynamic ratio of 4:1.48 When a drop of elemental
mercury was added to the isomerization reaction catalyzed by 2,
the rate of isomerization decreased significantly. The mercury
effect strongly implicates a heterogeneous Co colloid as the
actual catalyst, which would be generated from degradation of
2.
On the other hand, catalysis by 1 gave impressive selectivity
for the less thermodynamically preferred Z-2-isomer at early
times, and the selectivity slowly degraded at later times (see
below for more detail). The reaction rate and the selectivity of
isomerization catalyzed by 1 were unchanged by the addition of
mercury, suggesting that it remains a homogeneous catalyst
under the reaction conditions.49 Therefore, all subsequent studies
were performed with catalyst 1.
The unusual Z-selectivity of catalyst 1 was dependent on the
solvent, temperature, catalyst loading, and substrate concen-
tration. Tables 2 and 3 show the results obtained when
stopping the reaction after 12 h, a time period that was a good
balance between conversion (which increased over time) and
selectivity (which began to degrade at longer times).
The Z-selectivity was enhanced in aromatic solvents (entries
1−4 in Table 2). The reaction rate was faster at higher
1
by H NMR spectroscopy, including their E:Z ratios. Since
these alkenes reacted more quickly, the amount of over-
isomerization could be minimized by shortening the reaction
time, at minor expense to the conversion (Table 5).
Homoallylsilane could be isomerized to Z-crotylsilane with
high selectivity, showing the tolerance of the reaction to silanes.
However, isomerization of allylsilanes to vinylsilanes gave more
E- than Z-product (Table 4, entries 7 and 8). This loss of Z-
selectivity is attributed to formation of a heterogeneous catalyst,
because the reactivity of allylsilanes decreases significantly upon
addition of a drop of Hg (the Z-selective conversion of
homoallylsilane remains unaffected). The same phenomenon
was observed with dienes: 4,8-dimethylnona-1,7-diene (Table
4, entry 6) gave isomerization of terminal double bonds to
mostly E-product, but activity decreases with addition of Hg.
Thus, under conditions we examined, the homogeneous
C
dx.doi.org/10.1021/ja408238n | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
catalyst always displayed Z-selectivity, but degradation of the
selective catalyst by certain substrates gave products near the
thermodynamic ratio.
Table 4. Isomerization of Other Alkenes
The cobalt-catalyzed alkene isomerization also tolerated silyl-
protected alcohols with branched methyl, protected ketone, and
aromatic alkenes. Alkenes containing an aromatic ring gave low
conversion and poor Z-selectivity under standard conditions,
but it was possible to improve these isomerizations through
manipulation of conditions (discussed below). Terminal dienes
(1,5-hexadiene and 1,7-octadiene) gave no isomerization
products, and the catalyst decomposed. Alcohols rapidly
degraded the catalyst; some silyl-protected alcohols (silyl
ethers) also reacted with the catalyst, through a reaction that
will be discussed in more detail below. However, the presence
of a methyl substituent between the terminal alkene and the
siloxy group was sufficient to enable TBS-protected alcohols to
give excellent yields of Z-2-isomers under standard reaction
conditions, with no impact on the siloxy group.
In an effort to explain the poor performance of aromatic
alkenes and dienes, we isolated decomposition products from
the attempted isomerizations of allylbenzene and 1,5-hexadiene,
and characterized them using X-ray crystallography.
The X-ray crystal structure of the product from the reaction
of 1 with allylbenzene (Figure 4) shows η2-allylbenzene
coordinated to a cobalt(I) complex. The Co−N bond distances
in LtBuCo(η2-allylbenzene) are similar to Co−N bond distances
in other LtBuCoI complexes.47 The CC bond is neither in the
diketiminate plane nor perpendicular to it, as observed in the
iron complex LMeFe(η2-styrene).50 The CC bond distance is
elongated from 1.34 Å in a typical alkene51 to 1.382(3) Å by
coordination to cobalt(I). This coordinated CC bond is not
as long as those in LMeFe(η2-styrene) (1.401(8) Å)50 and
LMeFe(Ph2CCH2) (1.411 Å),50,52 probably because of
reduced backbonding from the more electronegative cobalt.
The CC−C angle of 122(2)° is slightly larger than the
expected 120° for sp2-hybridized carbon, presumably from
steric effects.
This cobalt(I) complex was synthesized independently by
mixing allylbenzene with LtBuCo, and was fully characterized. It
is high-spin (S = 1) and paramagnetic, but has a relatively sharp,
a
b
5 mol % catalyst, 80 °C. Yield and E:Z ratio determined by NMR
analysis. N/A = not applicable; ND = E,Z-mixture could not be
c
d
resolved. 0.5 mol % catalyst, 80 °C. Other internal products
observed; see Table 5 for details. 5% of 1-ethylcyclohexene. 2 mol %
catalyst, 80 °C. 7% of but-1-en-1-yl trimethylsilane. 9% of tert-
butyldimethyl((2-methylpent-2-en-1-yl)oxy)silane. 1 mol % catalyst,
80 °C. 3% of but-1-en-1-ylbenzene. Trace amount of pent-1-en-1-
ylbenzene observed.
e
f
g
h
i
j
k
Table 5. Isomerization from 1-Octene and 1-Dodecene at
a
Different Reaction Times
reaction
2-
E:Z for 2- other internal
time (h) conversion alkene
alkene
alkenes
1-octene
12
8
93%
89%
95%
89%
67%
68%
37%
58%
1:7.2
1:10
1:5.3
1:6.3
26%
21%
58%
31%
1-octene
Figure 4. ORTEP drawing of LtBuCo(η2-allylbenzene), which came
from attempted isomerization of allylbenzene by catalyst 1. Thermal
ellipsoids shown at the 50% probability level. Hydrogen atoms omitted
for clarity. Important bond distances (Å) and angles (deg): Co−N
1.9386(16), 1.9849(16), Co−C 2.027(2), 2.034(2), C−C 1.382(3);
N−Co−N 96.47(7).
1-dodecene
1-dodecene
12
7
a
0.5 mol % catalyst loading, 80 °C; yields and E:Z ratios determined
1
by H NMR spectroscopy.
D
dx.doi.org/10.1021/ja408238n | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
well-resolved 1H NMR spectrum. The coordinated alkene
causes a reduction of symmetry from C2v to C2, which is
reflected in the larger number of peaks observed in the NMR
Scheme 3. Proposed Bimolecular Mechanism of Catalyst
Decomposition with Aromatic Substrates
1
spectrum (see Experimental Section). The H NMR spectrum
of LtBuCo(η2-allylbenzene) matched the paramagnetic part of
the spectrum for mixtures generated from the catalytic
isomerization of allylbenzene.
An attempted diene isomerization reaction also yielded a
crystal suitable for X-ray crystallography. The crystal structure
showed a dicobalt(I) hexadiene complex with the two CC
bonds of 1,5-hexadiene coordinated to two different cobalt
centers (Figure 5). The CC bond is neither in the β-
position. Consistent with this hypothesis, lowering the substrate
and catalyst concentrations by a factor of 20 gave much better
conversion and selectivity (Table 6).
Table 6. Allylbenzene Isomerization with Different Catalyst
a
Concentrations
b
b
b
entry
[cat.] (mM)
yield
E:Z
starting material recovered
1
2
3
4
50.0
25.0
5.0
17%
24%
44%
1:0.4
1:0.6
1:1.3
1:1.4
77%
65%
50%
trace
2.5
90%
a
b
5 mol % catalyst, 80 °C. Yields and E:Z are determined by NMR
analysis.
Using the modified substrate concentration of 0.05 M, the
isomerization of aromatic substrates gave significantly better
results (Table 7).
Table 7. Isomerization of Aromatic Substrates at Lower
a
Concentration
Figure 5. ORTEP drawing of the molecular structure of (LtBuCo)2(μ-
η2:η2-1,5-hexadiene). Thermal ellipsoids at 50% probability. Hydrogen
atoms omitted for clarity. Important bond distances (Å): Co−N
1.974(5), 1.950(5); N−Co−N 96.9(2).
diketiminate plane nor perpendicular to it, which is similar to
LtBuCo(η2-allylbenzene) and LMeFe(η2-styrene).50 The mole-
cule lies on a crystallographic inversion center; thus, one-half is
unique. The two β-diketiminate planes are parallel to each
other, and the two CC bonds are parallel. The CC bond
distances are elongated from 1.34 Å in a typical alkene51 to
1.394(9) Å, which is similar to that in LtBuCo(η2‑allylbenzene).
As in LtBuCo(η2-allylbenzene), the CCC angle
(125.3(7)°) is larger than expected for a sp2 hybridized carbon.
The isolation of these two products suggests that the active
cobalt(II) catalyst can be reduced to unselective or inactive
cobalt(I) species in the presence of certain substrates. In the
case of allylbenzene, mechanistic insight came from the fact that
propylbenzene was detected in the final reaction mixture by
GC−MS. To explain the transformation from LtBuCo(n-hexyl)
to LtBuCo(alkene), we propose a tentative mechanism
consisting of a bimolecular hydrogen atom transfer followed
by reductive elimination of a C−H bond (Scheme 3). In the
disproportionation, we cannot rule out the involvement of free
radicals, but a pathway with intermolecular hydrogen atom
transfer directly between cobalt centers has more precedents in
intermolecular hydrogen atom transfer reactions from metal
hydrides.53−58 Because this reaction involves two cobalt species
reacting with one another, it predicts that lowering the
concentration of catalyst should lower the rate of the
bimolecular reaction and therefore prevent catalyst decom-
a
b
0.05 M substrate, 5 mol % catalyst, 80 °C. Yields and E:Z
c
determined by NMR analysis. 13% but-1-en-1-ylbenzene observed
(E:Z = 1: 1). 10% of pent-1-en-1-ylbenzene observed (E:Z = 1:1.1).
d
All isomerizations of aromatic substrates are Z-selective,
though the selectivity is lower for the formation of styrene
derivatives. Using substrates for which migration by a single
double bond to an allylic or homoallylic product was desired,
NMR spectra showed small amounts of styrene derivatives
formed from more extensive isomerization.
The mechanism through which dienes degrade the catalyst is
unknown. Interestingly, dilution of the reactions gives no
improvement in the diene reaction, which suggests that the
decomposition of catalyst proceeds through a different
mechanism than that described for the aromatic substrates
above.
E
dx.doi.org/10.1021/ja408238n | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Mechanistic Studies: Cobalt Profile. The cobalt species
present in the catalytic mixture during the isomerization
1
reaction could be monitored by H NMR spectroscopy. These
studies benefit from the paramagnetic shifts in the proton NMR
resonances of the cobalt complexes, which are easily
distinguished from those of diamagnetic substrates and
products. During the reaction, the only cobalt species observed
is LtBuCo(n-hexyl). When LtBuCoEt is used as an alternative
precatalyst for 1-hexene isomerization, LtBuCo(n-hexyl) is still
observed as the paramagnetic cobalt species. This observation
indicates that the cobalt−alkyl complex is the resting state for
the catalytic cycle, and that alkene exchange is facile.
We assume that other LtBuCo(alkyl) complexes are present in
low concentrations during catalysis; however, they must be
thermodynamically less stable and rearrange rapidly to
LtBuCo(n-hexyl). We tested this idea by treating LtBuCoCl
with 3-hexylmagnesium bromide at room temperature, which
immediately gave the isomerized product LtBuCo(n-hexyl). This
experiment shows that the terminal alkyl isomer is more stable,
as we have described in detail for the related iron(II)
complexes,42 that the chain walking is rapid even at room
temperature, and that even alkyl complexes that might give
more stable 3-hexenes through β-hydride elimination do not
prefer to do so.59
These observations most strongly implicate the alkyl
mechanism for isomerization, rather than the allyl mechanism.
Several additional pieces of evidence supporting this idea will
be presented below.
Mechanistic Studies: Alkene Profile. Monitoring the
substrate and product concentrations during the reaction was
most accurately accomplished using gas chromatography. The
reaction profiles for isomerization of 1-hexene under two near-
optimal conditions are shown in Figure 6. In both experiments,
1-hexene was consumed quickly in the first 10 h, and then, its
concentration remained unchanged; the concentration of E-2-
hexene increased slowly throughout the reaction time. The
concentration of Z-2-hexene reached its apex at around 10 h,
and then decreased. This degradation of selectivity was more
rapid using higher catalyst loading, suggesting that there could
be bimolecular decomposition of the selective catalyst over the
course of the reaction.
These data suggest a model in which all three of these
alkenes are in reversible equilibrium with a common steady-
state intermediate (see Supporting Information). However,
kinetic fits were underdetermined, and did not yield reliable
rate constants. Thus, we were able to use the kinetic data only
qualitatively.
Despite the difficulties with quantitative modeling, several
lessons may be learned from the product profile. First, the
isomerization reaction is reversible, because at very long times
the ratio of isomers trends toward the thermodynamic ratio.48
However, the ratio never reaches the thermodynamic limit,
suggesting that the catalyst degrades slowly over time. Second,
there are two concurrent processes: rapid isomerization of 1-
hexene to Z-2-hexene, and a slower isomerization between Z-2-
hexene and E-2-hexene that degrades the initially high Z-
selectivity. This suggests that later in the reaction when the
concentration of 1-hexene is lower, E/Z isomerization
dominates over double bond shifting and thus lowers the
selectivity. In order to test this idea, E-2-hexene and Z-2-hexene
were mixed with catalyst 1 separately under standard
conditions. Each isomerized toward the thermodynamic ratio
very slowly, but Z-2-hexene reacted more rapidly than E-2-
Figure 6. Concentration profiles of 1-hexene isomerization by 1. The
lines do not represent reliable kinetic models, but help to guide the
eye.
hexene. This rate sequence suggests that the Z isomer is more
easily incorporated into the catalytic cycle. The principle of
microscopic reversibility suggests that the Z-isomer would also
be released more easily by a catalytic intermediate, which agrees
with the Z selectivity of the catalyst (more detailed reasons for
the selectivity will be discussed in detail below).
We hypothesized that the undesired E/Z isomerization could
be slowed by adding an alkene that binds more weakly than 1-
hexene but more strongly than Z-2-hexene, thus preventing
binding and isomerization of Z-2-hexene. Table 8 shows that
Table 8. 1-Hexene Isomerization Reaction with Additives
time
(h)
a
a
condition
yield
E:Z
0.5 mol % cat., 80 °C, 1.0 M 1-hexene
24
48
24
48
88%
93%
86%
89%
1:4.0
1:1.8
1:4.6
1:2.4
0.5 mol % cat., 5.0 mol % 3,3-dimethyl-1-butene, 80
°C, 1.0 M 1-hexene
a
Yield of products and E:Z ratios determined by GC analysis.
added 3,3-dimethyl-1-butene to 1-hexene reaction does slow
the E/Z isomerization and maintains Z-selectivity for a longer
time. Addition of 3,3-dimethyl-1-butene did not significantly
improve the performance of aromatic substrates.
Mechanistic Studies: Isotope Labeling. The isomer-
ization of aromatic substrates produced some products in which
the CC double migrates more than one bond. In order to
provide a clearer picture of this “chain walking”,29,60−62 we
studied the isomerization of (1-13C)-1-hexene, shown in
Scheme 4. We monitored the reaction using 1H NMR
F
dx.doi.org/10.1021/ja408238n | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
spectroscopy after 0−2 h and after 24 h, and the results are
shown in Table 9. At the early time points, the 13C label was
proximal to the olefin, while after 24 h heating, the 13C label
was evenly distributed between the proximal and distal
Scheme 5. Independent Synthesis of Cobalt Siloxide
Complex
1
positions. H NMR analysis also showed that (6-13C) and
(1-13C)-2-hexene have nearly the same E/Z ratio at 24 h.
Scheme 4. Isomerization of (1-13C)-1-Hexene with Catalyst
1
Table 9. Concentrations of Different Species after Placing
(1-13C)-1-Hexene under Standard Isomerization Conditions
a
with Catalyst 1
1.0 h
2.0 h
24 h
Figure 7. ORTEP drawing of the molecular structure of LtBuCoOTBS.
Thermal ellipsoids at the 50% probability level. Hydrogen atoms
omitted for clarity. Important bond distances (Å) and angles (deg):
Co−N 1.9277(9), 1.9285(9), Co−O 1.798(4); N−Co−N 97.74(4),
Co−O−Si 162.6(3).
(1-13C)-1-hexene
(6-13C)-1-hexene
(1-13C)-2-hexene
(6-13C)-2-hexene
64%
9%
44%
11%
29%
11%
1%
b
b
1%
15%
5%
47%
46%
b
a
5 mol % catalyst, 80 °C; yields determined by NMR analysis. Due to
limitations of the analysis used, up to 5% of unlabeled 1-hexene or 2-
hexene derived from exchange with the hexyl group on the catalyst
may be included in these values.
The X-ray crystal structure of LtBuCoOTBS is shown in
Figure 7. The Co−N bond distances in LtBuCoOTBS are
similar to the Co−N bond distances in other LtBuCoII
complexes.43 The Co−O bond distance of 1.799(1) Å is
similar to the Fe−O distances in LMeFeOtBu (1.78(1) Å)63 and
LtBuFeOtBu (1.786(3) Å).64 The N−Co−N angle is 97.74(4)°,
and the N−Co−O angles are 133.2(1)° and 129.0(1)°; the sum
of 359.9° indicates that the cobalt center is planar. Like other
cobalt(II) complexes mentioned above, the 1H NMR spectrum
shows paramagnetically shifted resonances consistent with C2v
symmetry; this indicates that the TBS group rotates around the
C2 axis rapidly on the NMR time scale.
Observation of 13C label on both ends of olefins at 0.5 h,
before extensive product formation, indicates that chain walking
is faster than release of the isomerized product. However, the
preference for C1-labeled 2-hexene at early times shows that
the rates of chain walking and dissociation of alkene are similar.
Further, it shows that product formation is tied to the kinetics
of product release, since the E/Z ratio is the same for both
isotopologues after 24 h of heating. It is particularly interesting
that end-to-end chain walking occurs with 2-hexenes as major
products over more stable free 3-hexenes in the product
mixture: thus, elimination of 3-hexenes must be kinetically
limited by a high barrier.
We attribute the formation of LtBuCoOTBS to chain walking
followed by β-alkoxide elimination; note that chain walking was
also implied by the 13C labeling experiments above. We also
considered the possibility that the cobalt-OTBS species could
result from direct reaction with the C−O bond by a cobalt
hydride through a four-membered transition state, but reactions
with tert-butyldimethyl(octyloxy)silane and tert-butyldimethyl-
(3-phenylpropoxy)silane, which lack an alkene, did not form a
siloxide product. Thus, the elimination could be blocked by
preventing cobalt center from approaching OTBS.
Mechanistic Studies: Products from Reactions with
Protected Alcohols. We serendipitously learned more about
the “chain walking” through attempted isomerization of two
substrates with protected alcohols (Chart 1). Under standard
Chart 1. Terminal Alkenes Containing Protected Alcohols
The above concept suggests that siloxy groups might be
tolerated if chain walking were slowed. This was achieved
through adding steric hindrance. For example, a ketal (Table 4,
entry 12) is tolerated, perhaps since its C−O bonds are on a
hindered quaternary center. Tolerance of OTBS groups can be
achieved by simply introducing a methyl group between the
alkene and the siloxy group, as shown in Table 4, entry 11. Each
of these observations is fully consistent with the proposed β-
alkoxide elimination mechanism.
isomerization conditions with catalyst 1, the paramagnetic
products of both reactions were the same, as judged by H
1
NMR spectroscopy. This suggested that the cobalt products
were siloxide complexes. The siloxide complex was synthesized
independently from LtBuCoCl and KOTBS (Scheme 5), and its
1H NMR spectrum matched the paramagnetic peaks at the end
of attempted isomerization reactions with TBS-protected
alcohols.
DISCUSSION
■
Isomerization Mechanism. We find that the isomerization
of terminal alkenes to 2-alkene isomers can be catalyzed by
both the cobalt(II) complex LtBuCo(n-hexyl) (1) and the
cobalt(I) complex LtBuCo (2). Since these high-spin catalysts
G
dx.doi.org/10.1021/ja408238n | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
potentially problematic 2-alkene complex shown at the bottom
of the cycle.
Scheme 6. Proposed Mechanism of 1-Hexene Isomerization
The cobalt-catalyzed transformation of terminal alkenes to 2-
alkenes, without forming observable stable 3-alkenes, initially
suggested that chain walking from the terminal alkene might be
limited to one step followed by elimination, as seen in other
metal-hydride catalyzed alkene isomerizations.68 However, this
hypothesis was dispelled by the observation that 13C labeling
on one end of the initial alkene led to products with 13C label
on both ends, but similar E/Z ratios. This result shows that
both the selectivity and stereoselectivity are determined by the
elimination step.
Z-Selectivity in Alkene Isomerization by LtBuCo(n-
hexyl). The catalytic isomerization of terminal alkenes by 1
does not reach equilibrium. The order in thermodynamic
stability of hexene isomers is the following: E-3-hexene (ΔH°f =
−86.1 kJ/mol), E-2-hexene (ΔH°f = −85.5 kJ/mol), Z-2-
hexene (ΔH°f = −83.9 kJ/mol), Z-3-hexene (ΔH°f = −78.9 kJ/
mol), 1-hexene (ΔH°f = −74.2 kJ/mol).48 These heats of
formation are close enough that one would expect to see a
mixture of isomers at equilibrium, and predominantly E-
stereochemistry. The ability of LtBuCo(n-hexyl) to selectively
produce the 2-Z-hexene is thus kinetically controlled.
In a few examples in the literature, there was a preference for
the less thermodynamically stable Z isomer. In most cases, this
was brought about by directing groups.33,34,68,69 In a few cases,
there was slight Z-selectivity at the beginning of the
reaction,32,69 but it degraded rapidly. Most previous cases of
Z-selectivity were interpreted as arising from the allyl
mechanism, with the exception of ReBrH(NO)(dppfc),
where the initially formed 1:1 mixture of E- and Z-isomers
eventually isomerized to predominantly E-2-hexene.68 In this
context, it is interesting that LtBuCo(n-hexyl) catalyzes
isomerizations with Z-selectivities of up to 12:1 without
metal-binding functional groups. The Z-selectivity of the
LtBuCo(n-hexyl)-catalyzed isomerization degrades at longer
times, because the thermodynamic preference for E-isomers
inevitably asserts itself and because the catalyst degrades to a
product that is active for isomerization but not selective. Thus,
choosing the appropriate reaction time is important: longer
times give greater conversion, but in some cases this comes at
the expense of lower stereoselectivity and/or regioselectivity in
the products.
have electrons in M−L σ-antibonding orbitals, one might be
concerned that they could suffer metal−ligand dissociation,
giving a catalyst that is actually a soluble or insoluble metal
cluster or nanoparticle. In the case of catalyst 2, this fear
appears to have been realized, because addition of Hg
eliminates activity. In contrast, with catalyst 1, addition of Hg
or PPh3 has no significant effect on reactivity and selectivity,
which supports the idea that isomerization of alkenes by catalyst 1
is homogeneous.65,66 For this reason, only catalysis by 1 was
studied in detail.
As mentioned in the Introduction, there are two predom-
inant mechanisms for alkene isomerization in the literature,
which are distinguished by the active intermediates with an M−
C bond. In this system, observation of an alkylcobalt(II) species
as the resting form, the lack of a significant induction period for
catalysis by a cobalt-alkyl complex, the ability of an ethylcobalt-
(II) complex to convert to a hexylcobalt(II) complex with 1-
hexene, and the rapid isomerization of a 3-hexyl complex to a 1-
hexyl complex serve as strong evidence in support of the alkyl
mechanism (Scheme 1a). Isomerization between double bond
positions, and between alkyl isomers, is explained through chain
walking through reversible β-hydride elimination from the alkyl
complex.
The idea that the alkyl mechanism is facile is also supported
by our recently reported DFT computations on truncated
diketiminate-bound cobalt(II) alkyl complexes, which indicated
a barrier of only 13 kcal/mol for β-hydride elimination and 1.2
kcal/mol for alkene insertion into a transient cobalt-hydride
species.67 Though these barriers are expected to be higher with
the bulky ligand described here, the computations support the
presence of a feasible pathway for the key β-hydride elimination
and alkene insertion steps.
We propose the mechanism shown in Scheme 6 for
isomerization of 1-hexene to Z-2-hexene. The off-cycle 1-
hexyl complex is the spectroscopically observed resting state.
Note that chain walking occurs before alkene exchange, which
is needed to explain the 13C labeling results. The high Z-
selectivity at early times shows that the selectivity-determining
β-hydride elimination preferentially gives the Z-alkene hydride
complex. At later times, 1-hexenes are depleted, slowing the
alkene exchange, and competitive pathways (E−Z isomer-
ization) can degrade the selectivity. Added 3,3-dimethylbutene
can displace the 2-alkenes (blue) but not the productive 1-
alkene (purple), which reduces the concentration of the
Steric Model For Selectivity. In the isomerization
mechanism in Scheme 6, the proposed selectivity-determining
step is β-hydride elimination. Although LtBuCo(hexyl) has an S
3
=
/
ground state, previous computational studies67 have
2
shown that the lowest barrier for β-hydride elimination is on
1
the doublet (S = /2) potential energy surface that arises from
rapid spin crossover. Interestingly, these DFT studies found
that the transition state has a square-planar geometry, treating
the β-hydride as one of the ligands.
This transition-state geometry is the key to rationalizing the
Z-selectivity of the catalyst. Due to the choice between two β-
hydrogens, there are two distinct rotamers of the transition
state for β-hydride elimination from LtBuCo(2-hexyl) (Figure
8). These different rotamers determine whether the E- or Z-
alkene will be formed in the β-hydride elimination. Hindrance
is lessened in the Z-rotamer because the incipient bound alkene
can “roll” in order to place both alkene substituents between
the bulky isopropyl groups of the supporting ligand (Figure 8,
upper left). On the other hand, the E-isomer cannot reach a
conformation that achieves the key pseudo-square-planar
H
dx.doi.org/10.1021/ja408238n | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
selective isomerization of homoallylsilanes and allylsilanes to E-
crotylsilanes and E-vinylsilanes is known.75
The isomerization does not tolerate weakly acidic functional
groups such as alcohols and ketones, though protected versions
of these substrates (silyl esters and ketals) can be transformed
with excellent Z-selectivity. Though linear siloxy substituents
cause catalyst degradation through chain walking and β-siloxy
elimination, OTBS elimination is not observed when there is
one or more methyl groups in the chain between the alkene
and the protected alcohol. Because functionalized substrates are
much more common than linear substrates, it is likely that most
TBS-protected alcohols can be used as substrates.
Initial experiments with aromatic alkenes gave low
conversions, probably because of a bimolecular reaction.76
Lowering the catalyst concentration restored the Z-selectivity
to a level comparable to that of the aliphatic substrates, except
for styrenes. In the long run, we anticipate that modification of
the catalyst will ameliorate some of the incompatibilities, and
selectivity can be improved by using the steric model of
selectivity in Figure 8.
Figure 8. Steric model explaining the regioselectivity and stereo-
selectivity of alkene isomerization by the bulky cobalt catalyst. In the
transition state leading to E-2-hexene (upper right), one of the alkyl
substituents suffers a steric clash with the isopropyl groups. The
transition state leading to 3-hexenes (bottom) is constrained by the
aryl groups to give an unfavorable syn-pentane-like conformation.
CONCLUSIONS
geometry at Co without suffering severe steric interactions
between the methyl substituent and the diketiminate ligand
(Figure 8, upper right). This also explains why the diketiminate
supporting ligand is especially effective: this ligand electroni-
cally enforces a square-planar transition state, while sterically
constraining substrate substituents to a narrow cleft between
the isopropyl groups. The unusual selectivity for 2-alkenes over
3-alkenes is also rationalized by the bulky steric environment of
the diketiminate-cobalt catalyst. The key transition state leading
to 3-hexene (Figure 8, bottom) has two ethyl groups that are
conformationally constrained by the bulky ligands, in a way that
forces the ethyl groups together in a manner reminiscent of the
well-known syn-pentane interaction.70 In accordance with a
steric explanation for the selectivity, preliminary experiments
with less-bulky diketiminate ligands show neither regioselectiv-
ity nor stereoselectivity.
The mechanism proposed in Scheme 6 requires that the
unfavorable transition states in Figure 8 are not so high as to be
completely inaccessible, because chain walking requires β-
hydride elimination through an olefin hydride complex of the 3-
hexene. However, the alkene exchange step is likely to have
similar steric constraints. Future computational studies will test
this steric model quantitatively. However, the qualitative model
in Figure 8 is an initial guide for further catalyst development.
Scope of Catalytic Z-Selective Isomerization. Though
linear substrates might be expected to give low selectivity due
to the lack of functional groups, the regioselectivity of the
cobalt-catalyzed reaction is very good for isomerization of 1-
hexene to Z-2-hexene. In this case, trace 3-hexenes are observed
only after the reaction is continued for 24 h or longer. Limiting
the reaction time to 12 h gives an E/Z selectivity of 1:4 to 1:5,
and selectivity is best when the catalyst loading is kept low
(presumably avoiding bimolecular catalyst decomposition).
With longer linear alkenes, the E/Z selectivity is also very
good, though increasing the chain length lowers the
regioselectivity. This can be controlled to some extent by
shortening the reaction time, though the conversion is lower.
The cobalt-catalyzed reaction provides a route to function-
alized 2-alkenes. For example, a homoallylsilane was converted
specifically into the Z-crotylsilane with more than 90%
selectivity. Crotylsilanes are synthetically useful for allylations
of carbonyls,71 acetals,72 and enones.73,74 A cobalt catalyst for
■
The β-diketiminate cobalt(II) complex 1 catalyzes the isomer-
ization of terminal alkenes selectively to 2-alkenes with Z-
stereochemistry. Mechanistic studies on 1-hexene isomerization
show that the cobalt(II) catalyst walks rapidly along the hexyl
group, yet it selectively eliminates the Z-isomer of 2-hexene
rather than other isomers of lower energy. The selectivity can
be explained by a steric model, which should inform the future
design of new catalysts with improved Z-selectivity.
The selectivity for the transformation of 1-hexene is
particularly notable considering that there are no directing
groups on the substrate. Other terminal alkenes can be used as
well, including synthetically useful silanes, ketals, and protected
alcohols. Deleterious reactions with certain polyenes and silyl-
protected alcohols can be controlled by modifying the catalyst
concentration and substrate structure. Most importantly, the
models for the mechanism and selectivity lay the groundwork
for rational design of additional catalysts for selective alkene
isomerization with greater robustness and reaction scope.
EXPERIMENTAL SECTION
■
General Considerations. All reagents were purchased from
commercial sources and were dried over activated alumina and then
stored over 3 Å molecular sieves. Glassware was dried at 150 °C
overnight, and Celite was dried overnight at 200 °C under vacuum.
Pentane, hexane, benzene, diethyl ether, and toluene were purified by
passage through activated alumina and Q5 columns from Glass
Contour Co. Deuterated solvents were purchased from Cambridge
Isotope Laboratories, Inc. and dried over activated alumina.
NMR spectra were recorded on Bruker Avance 400 or 500 NMR
spectrometers, at roughly 295 K. All resonances in the 1H NMR
spectra are referenced to residual protiated benzene (δ 7.16 ppm) or
chloroform (δ 7.26 ppm). Resonances were singlets unless otherwise
noted. The NMR probe temperature was calibrated using either
ethylene glycol or methanol.77,78 IR data were recorded on a Shimadzu
FTIR spectrophotometer (FTIR-8400S) using a KBr pellet. UV−vis
spectra were recorded on a Cary 50 spectrophotometer using airtight
cuvettes with a 1 mm optical path length. Solution magnetic
susceptibilities were determined by the Evans method.79,80 Elemental
analyses were obtained from the CENTC Elemental Analysis Facility
at the University of Rochester. Further chromatographic details are
given in the Supporting Information.
L
tBuCo(n-hexyl). Under argon atmosphere, hexylmagnesium bro-
mide (0.84 mL, 2.0 M in THF, 1.7 mmol) was added dropwise to a
I
dx.doi.org/10.1021/ja408238n | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
solution of LtBuCoCl43 (989 mg, 1.66 mmol) in dry toluene (80 mL).
The suspension was stirred at room temperature for 1 h, and the color
changed to dark red-brown. Volatile materials were removed under
vacuum. The solid was extracted with pentane (80 mL), filtered
through Celite, and concentrated to 4.0 mL. Hexamethyldisiloxane
(4.0 mL) was added, and the solution was cooled to −40 °C to give
red-brown crystals (407 mg, 74%). 1H NMR (C6D6, 25 °C) δ −120.0
(12H, iPr methyl), −52.9 (1H, α-H), −50.9 (4H), −28.3 (2H), 3.8
(12H, iPr methyl), 4.2 (2H), 9.8 (2H), 11.3 (2H), 18.0 (2H), 27.0
(18H, tBu), 73.3 (4H), 242.0 (2H). The resonances that integrate for
General Procedure for Alkene Isomerization. Alkenes were
degassed, and dried using 4 Å molecular sieves; multiple treatments
were generally necessary to make the alkenes dry enough to prevent
catalyst degradation. To a resealable NMR tube was added 1-alkene
(0.50 mmol), mesitylene (69.5 μL, 0.50 mmol), LtBuCo(n-hexyl)
(0.025 mmol, 5.0% catalyst loading), and C6D6 (0.50 mL). The
1
reaction mixture was heated at 80 °C and monitored by H NMR
spectroscopy. After 24 h, the reaction mixture was opened to air and
filtered through a short silica pad (1.5 cm) in a disposable pipet to
1
remove Co-containing species. The filtrate was analyzed by H NMR
i
spectroscopy and GC/MS. Details are in the Supporting Information.
4H can be assigned as Pr-methine or aryl m-H. The resonances that
integrate for 2H can be assigned as aryl p-H or methylene in n-hexyl
group. μeff (Evans, C6D6, 299.2 K) = 4.9(1) μB. UV−vis (pentane):
278 nm (ε = 14.6 mM−1 cm−1), 332 nm (ε = 12.9 mM−1 cm−1), 468
nm (ε = 1.2 mM−1 cm−1), 736 nm (ε = 0.1 mM−1 cm−1). Anal. Calcd
for C41H66N2Co: C, 76.24; H, 10.32; N, 4.34. Found: C, 76.07; H,
10.52; N, 4.12. IR (KBr, cm−1): 3055 (w), 2958 (vs), 2924 (s), 2870
(m), 1508 (m), 1431 (m), 1383 (s), 1364 (vs), 1317 (m), 1251 (w),
1221 (w), 1155 (w), 1097 (m), 1055 (m), 1030 (w), 933 (w), 845
(w), 802 (w), 777 (m), 756 (w), 711 (w).
ASSOCIATED CONTENT
* Supporting Information
Additional kinetic, spectroscopic, and crystallographic data
(CIF), and details on the synthesis of organic compounds and
catalytic reactions. This material is available free of charge via
the Internet at http://pubs.acs.org.
■
S
L
tBuCo(η2-allylbenzene). Under argon atmosphere, allylbenzene
AUTHOR INFORMATION
■
(225 μL, 0.46 mmol) was added to a solution of LtBuCo47 (261 mg,
0.46 mmol) in dry diethyl ether (5.0 mL). The mixture was stirred at
room temperature for 0.5 h, and the color changed to dark red-brown.
Volatile materials were removed under vacuum. The solid was
extracted with pentane (10 mL), filtered through Celite, and
concentrated to 2.0 mL. Hexamethyldisiloxane (2.0 mL) was added,
and the solution was cooled to −40 °C to give red-brown crystals (256
mg, 81%). 1H NMR (500 MHz, C6D6) δ −101.5 (2H), −87.8 (6H, iPr
methyl), −47.2 (6H, iPr methyl), −27.4 (2H), −21.4 (2H), −17.4
(6H, iPr methyl), −11.8 (1H), −4.1 (6H, iPr methyl), −2.2 (2H), 11.9
(2H), 18.5 (1H), 22.0 (18H, backbone tBu), 44.2 (2H), 57.7 (1H),
227.7 (1H). The resonances that integrate for 2H can be assigned as
iPr methine, aryl m-H, aryl m-H (allyl benzene), aryl o-H (allyl
benzene), or benzyl-H (allyl benzene). The resonances that integrate
for 1H can be assigned as aryl p-H or α-H. The alkene protons are
presumably broadened into the baseline by the nearby paramagnetic
metal. μeff (Evans, C6D6, 295.8 K) = 3.8(1) μB. UV−vis (pentane): 550
nm (ε = 0.47 mM−1 cm−1). Anal. Calcd for C41H66N2Co: C, 77.68; H,
9.26; N, 4.21. Found: C, 77.87; H, 9.67; N, 4.04. IR (KBr, cm−1): 3057
(w), 3024 (w), 2959 (vs), 2927 (s), 2868 (m), 2385 (w), 1913 (w),
1855 (w), 1797 (w), 1541 (m), 1506 (s), 1463 (w), 1435 (m), 1388
(vs), 1359 (s), 1317 (s), 1286 (w), 1252 (w), 1217 (m), 1188 (m),
1155 (w), 1132 (w), 1095 (m), 1055 (w), 1028 (w), 934 (w), 885
(w), 806 (w), 773 (m), 740 (m), 694 (m), 574 (w), 534 (m, br).
Corresponding Authors
weix@chem.rochester.edu
patrick.holland@yale.edu
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
Funding was provided by the U.S. Department of Energy,
Office of Basic Energy Sciences, Grant DE-FG02-09ER16089.
Analytical data were obtained from the CENTC Elemental
Analysis Facility at the University of Rochester, funded by NSF
Grant CHE-0650456.
REFERENCES
■
(1) Siau, W.-Y.; Zhang, Y.; Zhao, Y. Top. Curr. Chem. 2012, 327, 33.
(2) Vedejs, E.; Peterson, M. J. Top. Stereochem. 1994, 21, 1.
(3) Bergelson, L. D.; Shemyakin, M. M. Tetrahedron 1963, 19, 149.
(4) Still, W. C.; Gennari, C. Tetrahedron Lett. 1983, 24, 4405.
(5) Hegedus, L. S.; Soderberg, B. C. G. Transition Metals in the
̈
Synthesis of Complex Organic Molecules, 3rd ed.; University Science
Books: Sausalito, CA, 2009.
(6) Lindlar, H.; Dubuis, R. Org. Synth. 1966, 46, 89.
(7) Katz, T. J.; McGinnis, J.; Altus, C. J. Am. Chem. Soc. 1976, 98,
606.
(8) Ibrahem, I.; Yu, M.; Schrock, R. R.; Hoveyda, A. H. J. Am. Chem.
Soc. 2009, 131, 3844.
L
tBuCo(OTBS). Under an argon atmosphere, a solution of tert-
butyl(dimethyl)silanol (21 mg, 0.16 mmol) in ether (2.0 mL) was
added dropwise to a suspension of KH (6.4 mg, 0.16 mmol) in ether
(2.0 mL). The reaction mixture was stirred at room temperature for 1
h and filtered through Celite into a solution of LtBuCoCl43 (92 mg,
0.15 mmol) in dry ether (2.0 mL). The suspension was stirred at room
temperature for 1 h, and the color changed to green. Volatile materials
were removed under vacuum. The solid was extracted with pentane
(10 mL) and filtered through Celite to give a green solution, which
was concentrated to 2.0 mL. Hexamethyldisiloxane (2.0 mL) was
added, and the solution was cooled to −40 °C to give green crystals
(9) Endo, K.; Grubbs, R. H. J. Am. Chem. Soc. 2011, 133, 8525.
(10) Herrmann, W. A.; Prinz, M. In Applied Homogeneous Catalysis
with Organometallic Compounds; Cornils, B., Herrmann, W. A., Eds.;
Wiley: New York, 2002; Vol. 3, p 111.
(11) van Leeuwen, P. W. N. M. Homogeneous Catalysis; Kluwer:
Dordrecht, 2004.
(12) Hubert, A. J.; Reimlinger, H. Synthesis 1969, 97.
(13) Hubert, A. J.; Reimlinger, H. Synthesis 1970, 405.
(14) Baudry, D.; Ephritikhine, M.; Felkin, H. J. Chem. Soc., Chem.
Commun. 1978, 694.
(15) Biswas, S.; Huang, Z.; Choliy, Y.; Wang, D. Y.; Brookhart, M.;
Krogh-Jespersen, K.; Goldman, A. S. J. Am. Chem. Soc. 2012, 134,
13276.
1
(47.9 mg, 45%). H NMR (500 MHz, C6D6) δ −89.0 (1H, α-H),
−71.5 (12H, iPr methyl), −34.9 (4H) −31.8 (2H, aryl p-H), −0.6
(12H, iPr methyl), 12.8 (9H, OTBS, tBu), 17.0 (6H, OTBS, methyl),
21.1 (18H, backbone tBu), 56.0 (4H). The resonances that integrate
for 4H can be assigned as iPr-methine or aryl m-H. μeff (Evans, C6D6,
295.8 K) = 4.8(1) μB. UV−vis (pentane): 585 nm (sh, ε = 0.16 mM−1
cm−1), 605 nm (ε = 0.16 mM−1 cm−1), 645 nm (ε = 0.16 mM−1
cm−1). Anal. Calcd for C41H67N2OSiCo: C, 71.26; H, 9.77; N, 4.05.
Found: C, 71.48; H, 10.17; N, 4.00. IR (KBr, cm−1): 3061 (m), 3014
(m), 2960 (vs), 2928 (s), 2872 (s), 2385 (w), 1923 (w), 1860 (w),
1800 (w), 1535 (m), 1506 (s), 1464 (m), 1435 (m), 1375 (s), 1360
(vs), 1317 (s), 1246 (m), 1217 (w), 1196 (m), 1136 (w), 1098 (m),
1057 (w), 982 (vs), 827 (m), 806 (w), 766 (m), 669 (m), 530 (m, br).
(16) Canovese, L.; Visentin, F. Inorg. Chim. Acta 2010, 363, 2375.
(17) Scarso, A.; Colladon, M.; Sgarbossa, P.; Santo, C.; Michelin, R.
A.; Strukul, G. Organometallics 2010, 29, 1487.
(18) Donohoe, T. J.; O’Riordan, T. J. C.; Rosa, C. P. Angew. Chem.,
Int. Ed. 2009, 48, 1014.
(19) Sivaramakrishna, A.; Mushonga, P.; Rogers, I. L.; Zheng, F.;
Haines, R. J.; Nordlander, E.; Moss, J. R. Polyhedron 2008, 27, 1911.
J
dx.doi.org/10.1021/ja408238n | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(20) Price, C. C.; Snyder, W. H. J. Am. Chem. Soc. 1961, 83, 1773.
(21) Fu, G. C. In Modern Rhodium-Catalyzed Organic Reactions;
Evans, P. A., Ed.; Wiley-VCH: Weinheim, 2005; p 79.
(22) Wille, A.; Tomm, S.; Frauenrath, H. Synthesis 1998, 305.
(23) Kuba, S.; Gates, B. C.; Grasselli, R. K.; Knozinger, H. Chem.
Commun. 2001, 321.
(24) Roos, L.; Orchin, M. J. Am. Chem. Soc. 1965, 87, 5502.
(25) Coperet, C. Beilstein J. Org. Chem. 2011, 7, 13.
(26) Krompiec, S.; Kuznik, N.; Penczek, R.; Rzepa, J.; Mrowiec-
Bialon, J. J. Mol. Catal. A: Chem. 2004, 219, 29.
(27) Larsen, C. R.; Grotjahn, D. B. Top. Catal. 2010, 53, 1015.
(28) Gauthier, D.; Lindhardt, A. T.; Olsen, E. P. K.; Overgaard, J.;
Skrydstrup, T. J. Am. Chem. Soc. 2010, 132, 7998.
(29) Grotjahn, D. B.; Larsen, C. R.; Gustafson, J. L.; Nair, R.; Sharma,
A. J. Am. Chem. Soc. 2007, 129, 9592.
(58) Barratt, D. S.; Cole-Hamilton, D. J. J. Chem. Soc., Chem.
Commun. 1985, 1559.
(59) Free 3-hexenes could not be generated quickly and then
completely consumed by the catalyst, because their quantitative
isomerization to other isomers is thermodynamically unfavorable.
(60) Chen, G.; Ma, X. S.; Guan, Z. J. Am. Chem. Soc. 2003, 125, 6697.
(61) Kochi, T.; Hamasaki, T.; Aoyama, Y.; Kawasaki, J.; Kakiuchi, F.
J. Am. Chem. Soc. 2012, 134, 16544.
(62) Guan, Z.; Cotts, P. M.; McCord, E. F.; McLain, S. J. Science
1999, 283, 2059.
(63) Eckert, N. A.; Smith, J. M.; Lachicotte, R. J.; Holland, P. L. Inorg.
Chem. 2004, 43, 3306.
(64) Gibson, V. C.; Marshall, E. L.; Navarro-Llobet, D.; White, A. J.
P.; Williams, D. J. J. Chem. Soc., Dalton Trans. 2002, 4321.
(65) Widegren, J. A.; Finke, R. G. J. Mol. Catal. A: Chem. 2003, 198,
317.
(66) Crabtree, R. H. Chem. Rev. 2011, 112, 1536.
(67) Bellows, S. M.; Cundari, T. R.; Holland, P. L. Organometallics
2013, 32, 4741−4751.
(68) Dudle, B.; Rajesh, K.; Blacque, O.; Berke, H. J. Am. Chem. Soc.
2011, 133, 8168.
(30) Larsen, C. R.; Grotjahn, D. B. J. Am. Chem. Soc. 2012, 134,
10357.
(31) Grotjahn, D. B. Pure Appl. Chem. 2010, 82, 635.
(32) Tuner, M.; Jouanne, J. v.; Brauer, H. D.; Kelm, H. J. Mol. Catal.
1979, 5, 433.
(33) Sato, T.; Komine, N.; Hirano, M.; Komiya, S. Chem. Lett. 1999,
(69) Pines, H.; Schaap, L. A. Adv. Catal. 1960, 12, 117.
(70) Hoffmann, R. W. Angew. Chem., Int. Ed. 2000, 39, 2054.
(71) Denmark, S. E.; Fu, J. Chem. Rev. 2003, 103, 2763.
(72) Hosomi, A.; Ando, M.; Sakurai, H. Chem. Lett. 1986, 15, 365.
(73) Fleming, I.; Barbero, A.; Walter, D. Chem. Rev. 1997, 97, 2063.
(74) Pan, L.-R.; Tokoroyama, T. Chem. Lett. 1990, 19, 1999.
(75) Kobayashi, T.; Yorimitsu, H.; Oshima, K. Chem.Asian J. 2009,
4, 1078.
(76) Tan, E. H. P.; Lloyd-Jones, G. C.; Harvey, J. N.; Lennox, A. J. J.;
Mills, B. M. Angew. Chem., Int. Ed. 2011, 50, 9602.
(77) Ammann, C.; Meier, P.; Merbach, A. E. J. Magn. Reson. 1982, 46,
319.
(78) Kaplan, M. L.; Bovey, F. A.; Cheng, H. N. Anal. Chem. 1975, 47,
1703.
(79) Baker, M. V.; Field, L. D.; Hambley, T. W. Inorg. Chem. 1988,
27, 2872.
(80) Schubert, E. M. J. Chem. Educ. 1992, 69, 62.
441.
(34) Punner, F.; Schmidt, A.; Hilt, G. Angew. Chem., Int. Ed. 2012, 51,
1270.
(35) Ohmura, T.; Yamamoto, Y.; Miyaura, N. Chem. Commun. 1998,
1337.
̈
(36) Tani, K.; Yamagata, T.; Akutagawa, S.; Kumobayashi, H.;
Taketomi, T.; Takaya, H.; Miyashita, A.; Noyori, R.; Otsuka, S. J. Am.
Chem. Soc. 1984, 106, 5208.
(37) Krompiec, S.; Pigulla, M.; Krompiec, M.; Marciniec, B.;
Chadyniak, D. J. Mol. Catal. A: Chem. 2005, 237, 17.
(38) Crabtree, H. R. In The Organometallic Chemistry of the Transition
Metals, 4th ed.; Wiley-Interscience: New York, 2005, p 239.
(39) Poli, R. Chem. Rev. 1996, 96, 2135.
(40) Poli, R. J. Organomet. Chem. 2004, 689, 4291.
(41) Vela, J.; Smith, J. M.; Lachicotte, R. J.; Holland, P. L. Chem.
Commun. 2002, 2886.
(42) Vela, J.; Vaddadi, S.; Cundari, T. R.; Smith, J. M.; Gregory, E. A.;
Lachicotte, R. J.; Flaschenriem, C. J.; Holland, P. L. Organometallics
2004, 23, 5226.
(43) Holland, P. L.; Cundari, T. R.; Perez, L. L.; Eckert, N. A.;
Lachicotte, R. J. J. Am. Chem. Soc. 2002, 124, 14416.
(44) Ding, K.; Brennessel, W. W.; Holland, P. L. J. Am. Chem. Soc.
2009, 131, 10804.
(45) Ding, K.; Dugan, T. R.; Brennessel, W. W.; Bill, E.; Holland, P.
L. Organometallics 2009, 28, 6650.
(46) Ding, K.; Dugan, T. R.; Holland, P. L.; Adhikari, D.; Mindiola,
D. J. Inorg. Synth 2010, 35, 43.
(47) Dugan, T. R.; Sun, X.; Rybak-Akimova, E. V.; Olatunji-Ojo, O.;
Cundari, T. R.; Holland, P. L. J. Am. Chem. Soc. 2011, 133, 12418.
(48) Handbook of Chemistry and Physics, 73rd ed.; Lide, D. R., Ed.;
CRC Press: Boca Raton, FL, 1993.
(49) Crabtree, R. H. Chem. Rev. 2012, 112, 1536.
(50) Yu, Y.; Smith, J. M.; Flaschenriem, C. J.; Holland, P. L. Inorg.
Chem. 2006, 45, 5742.
(51) Cochran, J. C.; Hagen, K. r.; Paulen, G.; Shen, Q.; Tom, S.;
Traetteberg, M.; Wells, C. J. Mol. Struct. 1997, 413, 313.
(52) Bai, G.; Wei, P.; Das, A. K.; Stephan, D. W. Dalton Trans. 2006,
1141.
(53) Trupia, S.; Bitterwolf, T. E.; Geiger, W. E. Organometallics 2012,
31, 6063.
(54) Cabelli, D. E.; Shafiq, F.; Creutz, C.; Bullock, R. M.
Organometallics 2001, 20, 3729.
(55) Berning, D. E.; Noll, B. C.; DuBois, D. L. J. Am. Chem. Soc.
1999, 121, 11432.
(56) Bullock, R. M. Comments Inorg. Chem. 1991, 12, 1.
(57) Estes, D. P.; Norton, J. R.; Jockusch, S.; Sattler, W. J. Am. Chem.
Soc. 2012, 134, 15512.
K
dx.doi.org/10.1021/ja408238n | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX