Iridium-catalyzed, intermolecular hydroamination of unactivated alkenes with indoles

Article abstract of DOI:10.1021/ja412116d

The addition of an N-H bond to an olefin is the most direct route for the synthesis of alkylamines. Currently, intermolecular hydroamination is limited to reactions of a narrow range of reagents containing N-H bonds or activated alkenes, and all the examples of additions to unactivated alkenes require large excesses of alkene. We report intermolecular hydroamination reactions of indoles with unactivated olefins. The reactions occur with as few as 1.5 equiv of olefin to form N-alkylindoles exclusively and in good yield. Characterizations of the catalyst resting state, kinetic data, labeling studies, and computational data imply that the addition occurs by olefin insertion into the Ir-N bond of an N-indolyl complex and that this insertion reaction is faster than insertion of olefin into the Ir-C bond of the isomeric C-2-indolyl complex.

Full text of DOI:10.1021/ja412116d

Article
pubs.acs.org/JACS
Iridium-Catalyzed, Intermolecular Hydroamination of Unactivated
Alkenes with Indoles
Christo S. Sevov,†,‡ Jianrong (Steve) Zhou,‡ and John F. Hartwig*,†,‡
†Division of Chemical Sciences, Lawrence Berkeley National Laboratory, Department of Chemistry, University of California,
Berkeley, California 94720
‡Department of Chemistry, University of Illinois, 600 South Mathews Avenue, Urbana, Illinois 61801, United States
S
* Supporting Information
ABSTRACT: The addition of an N−H bond to an olefin is
the most direct route for the synthesis of alkylamines.
Currently, intermolecular hydroamination is limited to
reactions of a narrow range of reagents containing N−H
bonds or activated alkenes, and all the examples of additions to
unactivated alkenes require large excesses of alkene. We report
intermolecular hydroamination reactions of indoles with unactivated olefins. The reactions occur with as few as 1.5 equiv of olefin
to form N-alkylindoles exclusively and in good yield. Characterizations of the catalyst resting state, kinetic data, labeling studies,
and computational data imply that the addition occurs by olefin insertion into the Ir−N bond of an N-indolyl complex and that
this insertion reaction is faster than insertion of olefin into the Ir−C bond of the isomeric C-2-indolyl complex.
exploit the abundance of alkene starting materials to form
products with complete atom economy, but these reactions do
not occur without a catalyst.13 Metal-catalyzed hydroamination
with indole as the N−H donor has the potential to form N-
alkyl rather than C-alkyl indole products in a single step with
chemo-, regio-, or stereochemical control; however, there are
no examples of additions of indole N−H bonds to unactivated
alkenes.
INTRODUCTION
■
Nitrogen-containing heterocycles constitute many biologically
active and medicinally important compounds.1−3 The indole
core is one of the most prevalent architectures found in these
compounds.4 As a result, extensive effort has been dedicated to
developing new strategies for site-selective functionalization of
the indole ring.5−7 However, participation of the lone pair on
nitrogen in the aromatic system makes indoles more
nucleophilic at C3 than at the nitrogen. Thus, in the absence
of deactivating groups,8 intermolecular alkylation of indoles
selectively at nitrogen remains challenging.9
Classical methodologies for the alkylation of amines involve
substitution reactions of aliphatic electrophiles or reductive
amination of carbonyl compounds.10 However, the scope of
these processes does not include reactions of indoles. Brønsted
or Lewis acid catalyzed formation of N-vinylindoles that could
undergo reduction is often slow, and indole polymerization is
common under these conditions. Furthermore, indoles are
often competitively reduced to indolines under hydridic
conditions.11 Finally, nucleophilic substitution with secondary
alkyl electrophiles leads to hydrodehalogenation or competitive
alkylation at the C3 position of indole (Scheme 1).12
Catalytic hydroamination has been studied by many
researchers,14−16 but hydroaminations of unactivated alkenes
are limited in scope. Most reported hydroaminations are
intramolecular cyclizations of aminoalkenes.17−23 With few
exceptions,24−30 intermolecular hydroamination is limited to
reactions of activated alkenes, such as bicycloalkenes,31−34
vinylarenes,35−38 allenes,39,40 or dienes,41,42 or to reactions of
ethylene.43−45 Hydroaminations of unactivated alkenes are
conducted with solvent quantities or large excesses of the
alkene and are limited to amides,30 cyclic ureas,28 sulfona-
mides,26,27 or cyclopentyl- and benzylamines.29 The prevalence
of indole cores in biologically active molecules renders indoles
an important, yet unexplored, class of substrates for hydro-
amination.4 Furthermore, synthesis of N-alkylindoles by olefin
hydroamination, instead of reaction with an alkyl electrophile,
could occur selectively at the N−H bond over the nucleophilic
C3 position. These reactions would provide valuable
mechanistic data on reactions of N−H bonds of azoles with
organometallic complexes and subsequent addition to olefins.
We report iridium-catalyzed additions of the N−H bond of
indoles to unactivated α-olefins. In contrast to the reactions of
strained alkenes, which form the products from C−H addition
Alternatively, one could envision forming N-alkyl indoles by
hydroamination: the formal addition of an N−H bond across
an unsaturated C−C bond. Such addition reactions would
Scheme 1. Common Challenges to Traditional Alkylation
Protocols of Indoles at Nitrogen
Received: November 27, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja412116d | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Reactions conducted with Ir complexes of Segphos or DM-
Segphos (DM = 3,5-dimethylphenyl) formed products in low
yield at 150 °C (Table 1, entries 1 and 2). Much higher
reactivity was observed when the bulky DTBM analogue of
Segphos was used as ligand (entry 3). The selectivity for 1 over
1-ene was higher when the reaction was conducted at 100 °C
instead of 150 °C, but 1 was formed in only 45% yield, due to
slow rates and low conversion (entry 5 vs 3). Various solvents
were tested to increase the reaction rate at 100 °C, but
reactions in most solvents occurred to lower conversion.
However, an increase in the reaction rates and yields were
observed when the reactions were conducted with added
EtOAc. A plot of the conversion vs the amount of added EtOAc
is given in Figure 1. These data show that this promoting effect
of the ester is the largest with 1 equiv of EtOAc, relative to the
indole (Table 1, entry 8), and smaller at higher concentrations
of EtOAc.
Scheme 2. Ir-Catalyzed C−H vs N−H Bond Addition to
Olefins
of indole (Scheme 2, top),46 the reactions of indole with
unstrained α-olefins formed exclusively the products of N−H
addition (Scheme 2, bottom). These N−H addition reactions
require as few as 1.5 equiv of olefin and form N-alkylindole
products in good yield. Experimental mechanistic studies reveal
information about the effect of ligand on the structure of the
catalyst resting state. A combination of experimental and
computational studies indicates that the barrier to insertion of
alkenes into the Ir−N bond of an indolyl complex is lower than
that for insertion into the Ir−C bond of the tautomer and imply
that the turnover-limiting step is olefin insertion into an Ir−N
bond.
RESULTS AND DISCUSSION
■
Figure 1. Effect of added EtOAc on reaction conversion after 18 h
with 0.9 M initial indole concentration.
Method Development. The reactions of the indole N−H
bond with an α-olefin under a variety of conditions are shown
in Table 1. Previously, we showed that Ir complexes of DTBM-
Segphos (DTBM = 3,5-di-tert-butyl-4-methoxyphenyl) catalyze
the hydroheteroarylation of bicycloalkenes with indoles.46 In
contrast, the combination of [Ir(cod)Cl]2 (cod =1,5-cyclo-
octadiene) and a series of bisphosphine ligands as catalyst led
to the addition of the indole N−H bond to 1-octene in varying
yields. In all cases, the reaction occurred exclusively at the N−H
bond with Markovnikov selectivity to form 1, along with
smaller amounts of oxidative amination product 1-ene. The
reaction occurred with substantial enantiomeric excess (70% at
80 °C, vide infra), a result that rules out hydroamination of the
alkene by a proton-catalyzed process alone. Trace amounts of
octane were detected, suggesting that 1-octene is the terminal
oxidant for the process that forms 1-ene.
The reaction conducted with 1 equiv of added EtOAc
proceeded with a 1.6-fold higher initial rate than the reaction
conducted in the absence of EtOAc. The origin of its positive
effect on reaction rate is unknown at this time; the added
EtOAc was not consumed during the reaction. Nevertheless,
since this protocol involved added EtOAc, the yields of
reactions at 100 °C were comparable to those of reactions
conducted at 150 °C without EtOAc, and smaller amounts of 1-
ene formed.
Scope of Olefin Hydroamination with Indoles. The
scope of addition of the N−H bond of a range of indoles to 1-
octene by the procedure just described is illustrated in Chart 1.
The products from addition of indoles containing benzyl- or
a
Table 1. Reaction Development for the Ir-Catalyzed N−H Bond Addition of Indole to 1-Octene
entry
ligand
x mol % [Ir(cod)Cl]2
temp (°C)
EtOAc (equiv)
% 1
% 1-ene
1/1-ene
1
2
3
4
5
6
7
8
9
(S)-SEGPHOS
1
1
1
2
2
2
2
2
2
4
150
150
150
120
100
100
100
100
100
80
0
0
9
7
5
1.3
3.0
3.2
3.6
4.1
3.8
4.0
4.2
3.9
5.4
(S)-DM-SEGPHOS
15
b
(S)-DTBM-SEGPHOS
(S)-DTBM-SEGPHOS
(S)-DTBM-SEGPHOS
(S)-DTBM-SEGPHOS
(S)-DTBM-SEGPHOS
(S)-DTBM-SEGPHOS
(S)-DTBM-SEGPHOS
(S)-DTBM-SEGPHOS
0
74
72
45
50
52
76
63
65
23
20
11
13
13
18
16
12
0
0
10
5
c
e
1
0.5
1
d
10
a
b
c
d
e
Yields were determined by GC analysis. 20% ee. 64% ee. 48 h reaction time. 70% ee.
B
dx.doi.org/10.1021/ja412116d | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
the terminal alkene to unreactive internal olefin isomers was
observed for reactions of allyl cyclohexane (entry 12, 15%
terminal alkene remains). Good yields of product were
obtained with bulky alkenes by conducting the reaction with
up to 8 equiv of olefin. Reactions conducted with these higher
concentrations of alkene required up to 2 equiv of EtOAc to
achieve an increase in the reaction rate that equals the increase
observed for additions to the less hindered 1-octene (see the
Supporting Information for conditions specific to each
substrate).
Chart 1. Ir-Catalyzed Additions of Indoles to 1-Octene
Ir-catalyzed addition to gaseous alkenes was achieved with
only 4 equiv of ethene, propene, or butene (entries 20−22).
Reactions of indole and ethene proceed to complete
conversion, but they formed 20 in poor yield as a result of
competitive reaction at the C2 position of indole. In contrast,
additions to propene and butene occurred exclusively at
nitrogen.
a
Mechanistic Studies of Olefin Hydroamination with
Indoles. Intermolecular hydroamination of unactivated alkenes
remains challenging, and reactions are generally conducted with
large excesses of olefin. However, little mechanistic data are
available that provides insight into that challenges associated
with additions to this class of olefin. Our studies on the scope
of α-olefin hydroamination with indoles has revealed that
reactions can be conducted with as few as 1.5 equiv of alkene,
and reactions occur at temperatures as low as 80 °C.
Mechanistic study of this mild reaction, relative to other NH
additions to unactivated alkenes, could provide information
about the properties of the catalytic system that promote
hydroamination. Thus, we conducted experimental mechanistic
studies and computational studies on the reported reactions.
Isotope Labeling Experiments. We began our mecha-
nistic investigation by measuring the enantiomeric excess (ee)
with which the addition product is formed. The ee of the
products from a range of Ir-catalyzed additions of indoles to α-
olefins were not exceptional but were substantial (45−74%, see
the Supporting Information for details). In contrast to reactions
of alkenes that are linear or have a tertiary branch point, the ee’s
of the product from the reaction of indoles with tert-
butylpropene were only 4−5% ee (Chart 2: see 1 vs 14 or
15 vs 16).
Reported yields are isolated yields.
acetyl-protected alcohols, esters, and halides were formed in
good yield. No products formed from reactions of indoles with
substituents at the 2- or 7-positions, but electron-donating or
-withdrawing substituents at the 3-position were tolerated
(products 8 and 9). The electron density of the indole ring had
little influence on the final yield, but reactions of electron-poor
indoles proceeded with higher rates than reactions of electron-
rich indoles. The limits on the temperature at which the
reaction can be performed and the amount of olefin that can be
used were tested. The addition of 5-chloroindole to octene
occurred at just 80 °C. Under these conditions, compound 7
was generated in a high 83% yield with 74% ee. The reaction
conducted with 1.5 equiv of 1-octene at 100 °C formed 10 in
68% yield.
The scope of alkene that undergoes the addition reaction is
summarized in Chart 2. Products from additions to olefins
containing bulky substituents near the alkene moiety were
generally formed in lower yields than those from additions to 1-
octene. In these reactions, significant amounts of vinylindole
side products were formed (20% to 30%), and isomerization of
a
Chart 2. Ir-Catalyzed Addition of Idoles to α-Olefins
Scheme 3. Ir-Catalyzed Additions of D-Indole to Olefins
An isotope-labeling experiment was conducted to gain
insight into the mechanism of N−H bond addition across an
alkene, potential side reactions, and the origin of the large
differences in ee for the two classes of alkenes. The addition of
1,3-dideuterioindole to 1-octene formed 1-D with deuterium
located exclusively β- to nitrogen (Scheme 3, right). However,
addition to tert-butylpropene formed 14-D with measurable
amounts of deuterium α- to nitrogen (Scheme 3, left).
Furthermore, a larger amount of oxidative amination product
was generated from tert-butylpropene than from octene.
Incorporation of deuterium α- to nitrogen in the product
from tert-butylpropene likely occurs by hydrogenation of the
vinylindole product by an Ir−D complex.
a
Reported yields are isolated yields.
C
dx.doi.org/10.1021/ja412116d | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
To determine the effect of this reduction of the vinylindole
on stereoselectivity, we determined the absolute configuration
of the products from hydroamination and vinylindole
reduction. The absolute stereochemistry of the major
enantiomer of representative hydroamination product 22 was
determined to be 2S by single-crystal X-ray diffraction of the
crystalline derivative 24 (eq 1). Reductive workup of
i
vinylindole 1-ene with PrOH catalyzed by the Ir complex
remaining in solution converted 1-ene into 1. Quantitative
conversion of 1-ene was observed, but this reduction of 1-ene
formed the opposite enantiomer of 1 than did the hydro-
amination reaction. These data suggest that a significant
amount of 14 may be generated in low ee from unselective
reduction of 14-ene in situ. A similar correlation was observed
during our study on the addition of amides to α-olefins.30 The
reactions of amides produced alkyl amide products with poor
ee, along with equimolar amounts of eneamide products.
Mechanism of C−N Bond Formation. Iridium-catalyzed
C−H and N−H additions to bicycloalkenes have been shown
to occur by olefin insertion into an Ir−C46 or Ir−N47,34,30
bond, respectively. Alternatively, hydroaminations of α-olefins
catalyzed by late-transition-metal complexes have been
proposed to occur by nucleophilic attack on a coordinated
olefin ligand.44,26,28 Moreover, palladium-catalyzed oxidative
amination of alkenes occurs by both syn- and anti-amino-
palladation of the olefin.48−53 Given the different chemo-
selectivity of Ir-catalyzed indole addition to bicycloalkenes (C−
H addition) versus α-olefins (N−H addition), the mechanism
of C−N bond formation in Ir-catalyzed additions to α-olefins
was studied.
To distinguish between the pathways described above, the
geometry of the vinylindole products from reactions with
geometrically defined alkenes was determined. The Ir-catalyzed
reaction was performed with indole and (E)-octene-1-d1
(Figure 2a). After 24 h, a 1:1 ratio of E and Z isomers of 1-
ene had formed. When the reaction was monitored at lower
conversions, (E)-1-ene was observed as the major vinylindole
isomer (Figure 2b). These results suggest that (E)-1-ene is the
product of the initial Ir-catalyzed reaction, while the Z isomer is
formed by isomerization of the E isomer. The formation of (E)-
1-ene as the kinetic product is consistent with a mechanism
involving migratory insertion of (E)-octene-1-d1 into an Ir−N
bond, followed by stereospecific syn-coplanar β-H elimination,
as illustrated in Figure 2c.
Figure 2. (a) Ir-catalyzed addition of indole to (E)-D-octene. (b) Ratio
of (E)-1-ene to (Z)-1-ene vs time. (c) Proposed mechanism for the
stereospecific formation of (E)-1-ene.
further supports a carbon- rather than nitrogen-bound indolyl
group in the catalyst resting state. Correlations between ligand
31P atoms and upfield-shifted vinylic protons were observed in
the 31P−1H HMBC NMR spectrum, indicating that the sixth
ligand of the octahedron is 1-octene. Although this complex
contains a carbon-bound indolyl ligand, heating the solution of
25 to 100 °C for 2 h formed N-alkyl products 1 and 1-ene in a
combined yield of 94%. No product from reaction at the C2
position of indole was observed.
The high solubility of 25 prevented isolation in pure form.
Thus, we conducted similar stoichiometric reactions with more
polar Segphos derivatives that also formed catalysts for the
addition of the indole N−H bond to α-olefins, albeit less active
catalysts than those containing DTBM-Segphos. In contrast to
stoichiometric studies with Ir complexes of DTBM-Segphos,
reaction of indole with Ir complexes of Segphos and DM-
Segphos formed an iridium product (Figure 3a) containing two
hydride ligands (as determined by 1H NMR spectroscopy) and
one indole-derived ligand per two iridium centers. In addition,
the 31P NMR spectrum of the new complex contained four
chemically inequivalent resonances of equal intensity.
Single crystals of the Segphos-ligated version of this complex
26 were obtained, and the structure was determined by X-ray
diffraction (Figure 3b). Complex 26 results from a rare
combination of N−H and C−H oxidative addition of the same
substrate to two different metal centers. Complexes 26 and 27
were prepared in the absence of 1-octene, and their spectra
match those of the iridium species formed in the reactions of
indole with 1-octene catalyzed by Ir complexes of DM-Segphos.
The nuclearity of 25−27 in solution was assessed by NMR
DOSY techniques. The diffusion coefficients of complexes 25−
27 were measured under the standard catalytic reaction
conditions by monitoring the hydride resonances of the metal
complexes. The measured values were compared to the
diffusion coefficient of the known mononuclear complex 28
Characterization of the Resting State Ir Complexes.
The mode by which the indole binds to the catalyst was
determined by synthesis of potential intermediates and was
unexpected. The reaction of DTBM-Segphos, [Ir(cod)Cl]2,
with 1 equiv of indole and 10 equiv of 1-octene cleanly formed
a complex with corresponding NMR signals matching those of
the major species observed during catalytic reactions. This
complex was characterized by NMR spectroscopy in solution
and was shown to be complex 25, the product of C−H
activation at the C2 position of indole (eq 2). An Ir(deuteride)
2
resonance was observed by H NMR spectroscopy when a
similar experiment was conducted with 2-indole-d1, a result that
D
dx.doi.org/10.1021/ja412116d | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
support a turnover-limiting step (TLS) that occurs from a
resting-state containing a bound indolyl group and olefin.
The identity of the TLS was revealed by kinetic isotope
effects (KIEs) and activation parameters. A comparison of the
initial rates for the hydroamination of 1-octene with indole and
1,3-dideuteroindole in separate vessels revealed a KIE of 1.7
(Scheme 3). Although greater than unity, this KIE is likely too
small to result from a turnover-limiting N−H cleavage.19,23
This isotope effect is more consistent with reversible oxidative
addition of the indole N−H bond to the Ir center prior to the
TLS.
Scheme 4 shows a catalytic cycle that is consistent with our
mechanistic data. The assignment of the resting state to be the
Scheme 4. Proposed Catalytic Cycle
Figure 3. (a) Formation of the observed dinuclear catalyst resting state
with Segphos or DM-Segphos as ligand. (b) Crystal structure of 26.
(Figure 4a).30 The diffusion coefficients for 26 and 27 were
smaller than that for the model mononuclear complex 28, while
the diffusion coefficient of 25 was similar to that of 28.
Furthermore, DOSY NMR data for 25 were acquired with
equimolar amounts of 27 (Figure 4b, left) and 28 (Figure 4b,
right) present in solution as internal standards. These data are
consistent with the structures for these three complexes in
solution being those shown in eq 2 and Figure 3.
mononuclear complex 25 is based on the NMR DOSY
experiments and a first-order dependence of the rate on the
conentration of iridium. The deuterium-labeling experiments
indicate that 25 contains a C-bound indolyl ligand, and 31P−1H
HMBC NMR spectroscopy reveals coupling correlations
between resonances of the ligand phosphorus atoms and
resonances of vinylic protons of a metal-bound olefin.
The isotope effect and stereochemical study revealed the
events that occur from the resting state. The KIE of 1.7
suggests that reversible isomerization of C-bound 25 to a less
stable N-bound intermediate 29 occurs prior to the TLS. The
stereochemistry of the vinyl indole product from the reaction of
(E)-octene-1-d1 (Figure 2) implied that insertion of the bound
alkene into the transient N-bound indolyliridium complex
occurs from 29. The alkylindole product then would be
generated by C−H bond-forming reductive elimination from
30; the vinylindole product would be formed by β-H
elimination from the alkyl group of 30.
Further data on the composition of the highest energy
transition state was obtained by conducting an Eyring analysis
of the observed rate constants for additions of 5-fluoroindole to
1-octene at temperatures from 353 to 396 K. These data
revealed the following thermodynamic parameters for the
catalytic process: ΔH⧧ = 30.0 kcal/mol and Δ⧧ = 6 eu (Figure
5). The small ΔS⧧ is consistent with the same composition of
the highest energy transition state as of the resting state. The
highest energy transition state is likely the C−N bond
formation, occurring by olefin insertion into the Ir−N bond
of 29.
Figure 4. (a) Structures, molecular weights, and diffusion coefficients
of 27, 28, and unknown resting state complex 25. (b) NMR DOSY of
the Ir−H region of complexes 25, 27, and 28.
Kinetic Studies of a Representative Catalytic Reac-
tion. Kinetic analysis of the addition of indole to 1-octene
catalyzed by 25 was conducted by the method of initial rates
(to 20% conversion). The reaction was found to be first order
in catalyst and zeroth order in indole. The rate of the reaction
was independent of the concentration of 1-octene under the
standard catalytic conditions but depended positively on the
concentration of alkene below 1.5 M. These measurements
Computational Studies. This reaction system provides
fundamental data on rates of alkene migratory insertion into
Ir−C and Ir−N bonds (Figure 6). The proposed mechanism in
Scheme 4 includes olefin insertion into the Ir−N bond of 29 to
E
dx.doi.org/10.1021/ja412116d | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
basis set and ECP for iridium and the 6-311++g** basis set for
all other atoms. Solvation in mesitylene was modeled with the
IEFPCM.
Similar studies on the energy barriers for alkene insertions
into late transition metal−nitrogen and metal−carbon bonds
have been performed by DFT.54 Calculations of a bisphosphine
amidorhodium(I) and methylrhodium(I) complexes predicted
that the barrier to ethylene insertion into the Rh−N bond of
the amido complex will be lower than that for insertion into the
Rh−C bond of the analogous methyl complex. These
calculations suggested that a major contributing factor to the
difference in rates of alkene insertion stems from the presence
of a lone pair of electrons on the amido ligand. The ground
state of the Rh-amido species is destabilized, relative to the
analogous methyl complex, by repulsive interactions between
the filled metal d orbital and the filled amido p orbital, whereas
the product of insertion is stabilized by a dative Rh−N
interaction. Thus, there is less Rh−N bond cleavage than Rh−
C bond cleavage in the two transition states.
Figure 5. (a) Consumption of 5-fluoroindole vs time at variable
temperatures (b) Eyring analysis of the measured rate constants.
Results from these prior calculations on olefin insertions of
rhodium(amido) and rhodium(alkyl) complexes are consistent
with the lower energy for insertion of the alkene into the Ir−N
indolide bond than for insertion into the Ir−C indolide bond
required by the catalytic cycle in Scheme 4. However, the
2
2
indolides are bound through M−Csp and M−Nsp bonds,
3
3
rather than M−Csp and M−Nsp bonds of alkyl and amido
complexes. Therefore, it is unclear whether this trend in relative
rates would translate to our hydroamination of alkenes with
indoles. Migratory insertion into metal aryl bonds is generally
slower than insertion into metal alkyl bonds, due to the
stronger metal−arene bond.55,56 Thus, we sought to determine
how the change in hybridization at nitrogen and the inclusion
of the electron pair as part of an aromatic system would affect
the barrier to insertion into the Ir−N bond.
All possible diastereomeric transition states of C-bound 25-
Ph⧧ and N-bound 29-Ph⧧ were calculated. The barriers to
propene insertion into the Ir−N bond by 29-Ph⧧ fell within a
range of 26−32 kcal/mol, while barriers to insertion into the
Ir−C bond by 25-Ph⧧ ranged between 28 and 33 kcal/mol.
The lowest computed barrier for propene insertion into the Ir−
N bond of N-bound indolide 29-Ph is 1.4 kcal/mol lower than
the lowest computed barrier for propene insertion into the Ir−
C bond of C-bound indolide 29-Ph. Although the experimental
difference in energies of the system with DTBM-Segphos as
ligand must be higher than 1.4 kcal/mol, the results of
calculations on the truncated system with Segphos as ligand
follow the trend of faster insertion into the Ir−N bond of 29
than into the Ir−C bond of 25. These computed results parallel
the trend for previously studied olefin insertions into amido-
and methylrhodium complexes and support the mechanism
shown in the catalytic cycle of Scheme 4.
Analysis of the optimized geometry of alkyliridium product
30-Ph provides insight into factors that control the energy
barrier for olefin insertion into N-indolide−iridium bonds and
the relationship between this insertion reaction and the
insertion into an amido complex. Interaction between iridium
and nitrogen through the lone pair of electrons on nitrogen is
revealed by a pyramidalization of the nitrogen atom (Figure 7).
The alkyl group on nitrogen projects 40° out of the trigonal
plane. As would be expected from the hybridization at nitrogen
approaching sp3, and weak interaction of the lone pair with the
π-system of the indole ring, the C2−C3 bond of the indolyl
fragment in 30-Ph has more double bond character than that of
Figure 6. Depictions of the transition-state structures for olefin
insertion into the Ir−C and Ir−N bond calculated by density
functional theory. Free energies at 298 K are provided in parentheses.
form the observed hydroamination product, even though C-
bound 25 is more stable than N-bound 29.
To address this issue, we calculated by density functional
theory (DFT) the barriers for migratory insertion reactions of
analogues of 25 and 29 containing bound propene in place of
1-octene. To simplify calculations, transition states of
mononuclear complexes 25-Ph⧧ and 29-Ph⧧ containing the
parent Segphos ligand were modeled, rather than the systems
containing the full DTBM−Segphos ligand. As previously
discussed and illustrated in Figure 3, reactions catalyzed by
iridium complexes of the parent Segphos ligand form a
dinuclear resting state complex 26. Thus, all calculated energies
were normalized to 26.
The calculations were conducted with the Gaussian 09
package. Geometry optimization was conducted with the M06
functionals with the lanl2dz basis set and ECP for iridium and
the 6-31g(d,p) basis set for all other atoms at 298 K. Frequency
analysis was conducted at the same level of theory to verify the
stationary points to be minima or saddle points and to obtain
the thermodynamic energy corrections. Single-point energies
were calculated with the M06 functionals with the lanl2tz(f)
F
dx.doi.org/10.1021/ja412116d | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by the Director, Office of Science, of
the U.S. Department of Energy under Contract No. DE-AC02-
05CH11231 and the Molecular Graphics and Computation
Facility at UC Berkeley supported by the National Science
Foundation (CHE-0840505). We thank Johnson-Matthey for a
generous gift of [Ir(cod)Cl]2, and Takasago for a generous gift
of (S)-DTBM-Segphos. C.S.S. thanks the NSF and the
Springborn family for graduate research fellowships.
Figure 7. Computed geometry of the product from propene migratory
insertion into the Ir−N (30-Ph) by transition state 29-Ph⧧.
unbound indole. Thus, these calculations suggest that the lone
pair on nitrogen is available for stabilizing interaction with
iridium in the transition state, even though the starting ligand is
a planar indolide with an sp2-hybridized nitrogen.
REFERENCES
■
CONCLUSIONS
■
(1) Bagley, M. C.; Dale, J. W.; Merritt, E. A.; Xiong, X. Chem. Rev.
2005, 105, 685.
(2) Kim, J.; Movassaghi, M. Chem. Soc. Rev. 2009, 38, 3035.
(3) Joule, J. A.; Mills, K. Heterocyclic Chemistry; John Wiley & Sons:
New York, 2010.
(4) Sundberg, R. J.; Katritzky, A. R.; Meth-Cohn, O.; Rees, C. S.
Indoles; Elsevier Science: Waltham, MA, 1996.
(5) Bandini, M.; Melloni, A.; Umani-Ronchi, A. Angew. Chem., Int. Ed.
2004, 43, 550.
(6) Bandini, M.; Melloni, A.; Tommasi, S.; Umani-Ronchi, A. Synlett
2005, 1199.
In summary, we have reported a rare example of a catalytic
process for the addition of the N−H bonds of indoles to simple
α-olefins. This strategy for indole alkylation occurs exclusively
at nitrogen. Good product yields were obtained from reactions
conducted with as few as 1.5 equiv of alkene. These data reveal
properties of this catalytic system that promote hydroamination
of α-olefins with indoles.
1. The bulky DTBM-Segphos ligand facilitates turnover-
limiting olefin insertion into the Ir−N bond.51
2. The bulky ligand promotes the persistence of an active
mononuclear catalyst resting state, rather than off-cycle
dinuclear complexes. Complexes of the less bulky
Segphos and DM-Segphos ligands form dinuclear species
26 and 27, respectively.
(7) Bandini, M.; Eichholzer, A. Angew. Chem., Int. Ed. 2009, 48, 9608.
(8) Stanley, L. M.; Hartwig, J. F. Angew. Chem. 2009, 121, 7981.
(9) Cui, H.-L.; Feng, X.; Peng, J.; Lei, J.; Jiang, K.; Chen, Y.-C. Angew.
Chem., Int. Ed. 2009, 48, 5737.
(10) Trost, B. M.; Fleming, I. Comprehensive Organic Synthesis:
Reduction; Elsevier Science & Technology Books: Waltham, MA, 1991.
(11) Robinson, B. Chem. Rev. 1969, 69, 785.
(12) Heaney, H.; Ley, S. V. J. Chem. Soc., Perkin Trans. 1 1973, 499.
(13) Hartwig, J. F. In Organotransition Metal Chemistry: From Bonding
to Catalysis; University Science Books: Sausalito, CA, 2010; Vol. 1, p
700.
3. The IrIII complex 25 containing a hydride, a C-bound
indolyl ligand, and an alkene does not catalyze
isomerization of the terminal alkene to less reactive
internal isomers. However, bisphosphine IrI complexes
do catalyze isomerization; therefore, reactions containing
an IrI resting state tend to occur with a larger amount of
alkene isomerization than the reaction described here.
4. The resting state contains both substrate and olefin
bound to the metal center. The presence of the alkene in
the resting state mitigates the requirement for conduct-
ing the reaction under conditions with neat alkene
because the rate of the reaction is not dependent on the
concentration of alkene.
(14) Muller, T. E.; Hultzsch, K. C.; Yus, M.; Foubelo, F.; Tada, M.
̈
Chem. Rev. 2008, 108, 3795.
(15) Hesp, K. D.; Stradiotto, M. ChemCatChem 2010, 2, 1192.
(16) Hannedouche, J.; Schulz, E. Chem.Eur. J. 2013, 19, 4972.
(17) Hong, S.; Marks, T. J. Acc. Chem. Res. 2004, 37, 673.
(18) Dunne, J. F.; Fulton, D. B.; Ellern, A.; Sadow, A. D. J. Am. Chem.
Soc. 2010, 132, 17680.
(19) Hesp, K. D.; Tobisch, S.; Stradiotto, M. J. Am. Chem. Soc. 2010,
132, 413.
(20) Julian, L. D.; Hartwig, J. F. J. Am. Chem. Soc. 2010, 132, 13813.
(21) Shen, X.; Buchwald, S. L. Angew. Chem., Int. Ed. 2010, 49, 564.
(22) Leitch, D. C.; Platel, R. H.; Schafer, L. L. J. Am. Chem. Soc. 2011,
133, 15453.
(23) Liu, Z.; Yamamichi, H.; Madrahimov, S. T.; Hartwig, J. F. J. Am.
Chem. Soc. 2011, 133, 2772.
(24) Ryu, J.-S.; Li, G. Y.; Marks, T. J. J. Am. Chem. Soc. 2003, 125,
12584.
(25) Brunet, J. J.; Chu, N. C.; Diallo, O. Organometallics 2005, 24,
3104.
(26) Karshtedt, D.; Bell, A. T.; Tilley, T. D. J. Am. Chem. Soc. 2005,
5. The reactions occurring with less formation of vinyl-
indole side product occur in higher ee than those
occurring with more formation of the vinylindole side
product. The vinyl indole is reduced in situ to the
opposite enantiomer of the product than the alkyl indole
formed directly from hydroamination.
Efforts to mitigate oxidative amination processes and increase
the enantioselectivity of the reaction are in progress.
ASSOCIATED CONTENT
* Supporting Information
Experimental procedures and characterization of all new
compounds including NMR spectroscopy data, conditions for
HPLC separations on a chiral stationary phase, kinetic studies,
optimization data, and CIF. This material is available free of
charge via the Internet at http://pubs.acs.org.
■
S
127, 12640.
(27) Zhang, J.; Yang, C.-G.; He, C. J. Am. Chem. Soc. 2006, 128, 1798.
(28) For a seminal example of asymmetric, intermolecular hydro-
amination of unactivated alkenes with gold, see: Zhang, Z.; Lee, S. D.;
Widenhoefer, R. A. J. Am. Chem. Soc. 2009, 131, 5372.
(29) For a seminal example of asymmetric, intermolecular hydro-
amination of unactivated alkenes with lanthanides, see: Reznichenko,
A. L.; Nguyen, H. N.; Hultzsch, K. C. Angew. Chem., Int. Ed. 2010, 49,
8984.
AUTHOR INFORMATION
Corresponding Author
jhartwig@berkeley.edu
■
(30) Sevov, C. S.; Zhou, J.; Hartwig, J. F. J. Am. Chem. Soc. 2012, 134,
11960.
G
dx.doi.org/10.1021/ja412116d | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(31) Brunet, J.-J.; Commenges, G.; Neibecker, D.; Philippot, K. J.
Organomet. Chem. 1994, 469, 221.
(32) Dorta, R.; Egli, P.; Zurcher, F.; Togni, A. J. Am. Chem. Soc. 1997,
̈
119, 10857.
(33) McBee, J. L.; Bell, A. T.; Tilley, T. D. J. Am. Chem. Soc. 2008,
130, 16562.
(34) Zhou, J.; Hartwig, J. F. J. Am. Chem. Soc. 2008, 130, 12220.
(35) Kawatsura, M.; Hartwig, J. F. J. Am. Chem. Soc. 2000, 122, 9546.
(36) Utsunomiya, M.; Hartwig, J. F. J. Am. Chem. Soc. 2003, 125,
14286.
(37) Qian, H.; Widenhoefer, R. A. Org. Lett. 2005, 7, 2635.
(38) Pan, S. G.; Endo, K.; Shibata, T. Org. Lett. 2012, 14, 780.
(39) Nishina, N.; Yamamoto, Y. Synlett 2007, 1767.
(40) Butler, K. L.; Tragni, M.; Widenhoefer, R. A. Angew. Chem., Int.
Ed. 2012, 51, 5175.
(41) Lober, O.; Kawatsura, M.; Hartwig, J. F. J. Am. Chem. Soc. 2001,
̈
123, 4366.
(42) Johns, A. M.; Sakai, N.; Ridder, A.; Hartwig, J. F. J. Am. Chem.
Soc. 2006, 128, 9306.
(43) Brunet, J.-J.; Cadena, M.; Chu, N. C.; Diallo, O.; Jacob, K.;
Mothes, E. Organometallics 2004, 23, 1264.
(44) Wang, X.; Widenhoefer, R. A. Organometallics 2004, 23, 1649.
(45) Cao, P.; Cabrera, J.; Padilla, R.; Serra, D.; Rominger, F.;
Limbach, M. Organometallics 2012, 31, 921.
(46) Sevov, C. S.; Hartwig, J. F. J. Am. Chem. Soc. 2013, 135, 2116.
(47) Casalnuovo, A. L.; Calabrese, J. C.; Milstein, D. J. Am. Chem.
Soc. 1988, 110, 6738.
(48) Liu, G.; Stahl, S. S. J. Am. Chem. Soc. 2007, 129, 6328.
(49) Hanley, P. S.; Markovic,
2010, 132, 6302.
́
D.; Hartwig, J. F. J. Am. Chem. Soc.,
(50) Neukom, J. D.; Perch, N. S.; Wolfe, J. P. J. Am. Chem. Soc. 2010,
132, 6276.
(51) Hanley, P. S.; Hartwig, J. F. J. Am. Chem. Soc. 2011, 133, 15661.
(52) Weinstein, A. B.; Stahl, S. S. Angew. Chem., Int. Ed. 2012, 51,
11505.
(53) Hanley, P. S.; Hartwig, J. F. Angew. Chem., Int. Ed. 2013, 52,
8510.
(54) Tye, J. W.; Hartwig, J. F. J. Am. Chem. Soc. 2009, 131, 14703.
(55) Garrou, P. E.; Heck, R. F. J. Am. Chem. Soc. 1976, 98, 4115.
(56) Hartwig, J. F. In Organotransition Metal Chemistry: From Bonding
to Catalysis; University Science Books: Sausalito, CA, 2010; Vol. 1, p
349.
H
dx.doi.org/10.1021/ja412116d | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX