The mechanism of N-vinylindole formation via tandem imine formation and cycloisomerisation of o-ethynylanilines

Article abstract of DOI:10.1039/b915269e

The reaction of 2-(2-phenylethynyl)aniline with acetone in presence of [IrCp*Cl₂]₂ has previously been found to yield a vinyl indole derivative and not the indole expected to form following a hydroamination reaction. Experimental dat

Full text of DOI:10.1039/b915269e

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
The mechanism of N-vinylindole formation via tandem imine formation and
cycloisomerisation of o-ethynylanilines†
Danielle F. Kennedy,a Ainara Nova,b Anthony C. Willis,c Odile Eisenstein*b and Barbara A. Messerle*a
Received 29th July 2009, Accepted 17th September 2009
First published as an Advance Article on the web 14th October 2009
DOI: 10.1039/b915269e
The reaction of 2-(2-phenylethynyl)aniline with acetone in presence of [IrCp*Cl2]2 has previously been
found to yield a vinyl indole derivative and not the indole expected to form following a hydroamination
reaction. Experimental data, including labelling studies, isolation and solid state structure
determination of a reaction intermediate together with DFT calculations were used to develop a
mechanism for the formation of the vinyl indole. In the mechanism proposed, acetone plays a
significant role in several steps of the reaction path, participating in the fragmentation of the dinuclear
Ir complex and the formation of the reactive form of the catalyst as well as blocking the formation of
the expected hydroamination product by coordination to the Ir catalyst. Coordinated acetone reacts
with the aniline to form an imine derivative, which yields the final product following proton transfer
promoted by acetone. The proposed mechanism is in good agreement with the experimental data.
of amines with alcohols11,12 and the transfer hydrogenation of
Introduction
ketones13 and quinolines.14
Substituted indoles have become important pharmaceutical agents
because of the well documented biological activity exhibited by
N-vinyl indoles have been used as monomers in the produc-
tion of poly(N-vinyl indoles) which found application in the
compounds containing an indole motif.1,2 Indoles are said to have
a “privileged” structure due to their ability to bind with high
affinity to many biological receptors.3 Numerous methods have
been developed for the synthesis of substituted indoles, however,
the efficient synthesis of suitably modified indoles remains a
challenge to the synthetic chemist.3,4 In particular, it remains
difficult to synthesise many simple unsubstituted indoles. Bulky,
electron withdrawing or donating substituents on indole have been
widely used because they have significant directing effects on the
efficiency of synthesis, blocking, activating or deactivating various
positions to substitution.5
Increasingly, metal complexes containing Ir(III) metal cen-
tres are being investigated as catalysts for a variety of trans-
formations. Work by Bergmann et al. with the complexes
Cp*(PMe3)Ir(CH3)(OTf), and TpMe2(PMe3)Ir(CH3)(OTf) illus-
trated the ability of Ir(III) complexes to effect C–H activation.6,7
Crabtree et al. have utilised an Ir(III) hydride complex to achieve
alkyne insertion into a Ir-H, and also catalyse the intramolecular
hydroamination and hydroalkoxylation of alkynes.8,9 [Cp*IrCl2]2
is also effective as a catalyst for a variety of transformations, such
as the hydroborylation of styryl sulfonamides,10 the N-alkylation
production of photorefractive materials.15 There are, however,
relatively few reports of the synthesis of N-vinyl indoles.16-18
Work by Li et al. utilises a gold catalysed synthesis of N-
vinyl indoles via a Au(III) catalysed double hydroamination
of o-alkynylanilines and terminal alkynes.16 Investigation into
the mechanism of this reaction indicated that the first step of
this reaction was intermolecular hydroamination between the
aniline and terminal acetylene to generate the imine, followed
by intramolecular hydroamination.16 Recently we reported the
Ir(III) catalysed synthesis of N-(2-methylvinyl)-2-phenylindole 2
via the cyclisation of 2-(2-phenylethynyl)aniline 1 by the Ir(III)
dimer, [IrCp*Cl2]2, Scheme 1a.19 This reaction was an unexpected
outcome, and it was presumed to be the direct result of either a
hydroamination reaction or a modified hydroamination reaction
incorporating one molecule of the solvent acetone. These results
aSchool of Chemistry, The University of New South Wales, Sydney N.S.W.
2052, Australia. E-mail: b.messerle@unsw.edu.au; Fax: ++61-2-9385-6141
bInstitut Charles Gerhardt, Universite´ Montpellier 2, CNRS 5253, cc 1501,
Place Euge`ne Bataillon, 34095, Montpellier Cedex 5, France. E-mail:
Odile.Eisenstein@univ-montp2.fr
cResearch School of Chemistry, The Australian National University,
Canberra A.C.T. 0200, Australia
† Electronic supplementary information (ESI) available: Computational
results obtained with cationic Ir species. Optimized structures of all
stationary points with their absolute energies calculated in solvent. CCDC
reference number 741292. For ESI and crystallographic data in CIF or
other electronic format see DOI: 10.1039/b915269e
Scheme 1
10296 | Dalton Trans., 2009, 10296–10304
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
contrast with the more commonly observed hydroamination
cyclisation reaction without addition of acetone, forming the
indole product 3 (Scheme 1b), and the alternate Ir(III) catalysed
cyclisation reported recently where a mixture of 2 and 3 was
formed (Scheme 1c).19 This reaction provides a simple method
for the synthesis of the previously unknown vinyl indole 2, and
was discovered as a part of our investigations into new catalysts for
hydroamination. Here we describe our studies into the mechanism
of this reaction.
Deuterium labelling study
To gain further understanding of the catalysed formation of
N-(2-methylvinyl)-2-phenylindole 2, Scheme 1a, the reaction of
[IrCp*Cl2]2 with 1 was performed using acetone-d6 and monitored
via 1H NMR spectroscopy. When 2-(2-phenylethynyl)aniline 1 was
treated with 2.8 mol% [IrCp*Cl2]2 in acetone-d6 at 323 K (Fig. 2),
conversion of 1 to 2 was achieved in ca. 20 hours with 100%
deuteration of the methylvinyl group. This shows that the methyl
vinyl group does originate from the acetone. Complete deuteration
at C3 on the indole ring was also observed. This is consistent
with Pathway II, where the proton at C3 of the indole ring derives
directly from the deuterated acetone reactant. The hydroamination
pathway (Pathway I, Fig. 1) would result in protonation at C3 of
the indole, with the proton derived from N–H of the aniline.
The 1H NMR spectra were examined to detect the presence of
any of the possible intermediates A or B, or free 2-phenylindole
(3) (Fig. 2). No resonances due to the stable 2-phenylindole (3)
or the metal bound indole intermediate A were observed at any
time during the course of the reaction further suggesting that
Pathway I is not operative. Pathway II was not excluded from
consideration, even though the imine intermediate B was not
observed experimentally, as rapid cyclisation would consume B
immediately.
Results and discussion
The catalyst promoting the formation of vinyl indole (2) by reac-
tion of 2-(2-phenylethynyl)aniline (1) with acetone was the Ir(III)
dimer, [IrCp*Cl2]2 (Scheme 1a). When 2-(2-phenylethynyl)aniline
1 was reacted with 2.8 mol% [IrCp*Cl2]2, 95% in acetone,
conversion to vinyl indole 2 was observed in 20 h at 323 K. When
NaBF4 was included with the catalyst a significantly enhanced rate
of reaction (from 20 to 4 h) was observed. Two possible routes for
the formation of compound 2 are shown in Fig. 1 (Pathways I and
II). The first route (Pathway I) involves initial hydroamination
leading to the indole intermediate A followed by the reaction
of A with acetone. In this pathway the metal does not play an
active role in the addition of acetone. The second route (Pathway
II) involves initial reaction of the acetone with the aniline 1 to
generate an imine intermediate B, which then undergoes metal
catalysed cyclisation to generate the vinyl indole product 2. In this
work, we have combined experimental and computational studies
to investigate if one of these pathways is operative or if other
elementary steps should be included.
The reaction of 1 with [IrCp*Cl2]2 and NaBF4 was also
performed using acetone-d6 on a small scale and monitored
via 1H NMR spectroscopy. The aniline, 1, was reacted with
2.4 mol% [IrCp*Cl2]2 and NaBF4 in acetone-d6 at 323 K. The
reaction proceeded to more than 95% conversion in under 4 h,
a significantly faster rate than in the absence of NaBF4. The
1H NMR spectra from this reaction were examined and also
Fig. 1 Possible reaction routes for the generation of 2 from 1 in acetone indicating Pathway I via indole intermediate A and Pathway II via imine
intermediate B.
This journal is
The Royal Society of Chemistry 2009
Dalton Trans., 2009, 10296–10304 | 10297
©

View Article Online
Fig. 2 Stacked plots of the 1H NMR spectra for the catalysed reaction of 1 to 2 by 5.7 mol% [IrCp*Cl2]2 over time. The stars indicate resonances due to
1 and the circles indicate resonances due to 2.
failed to show the presence of 2-phenylindole 3 or any imine
intermediate at any stage of the reaction. The combination of
[IrCp*Cl2]2 and NaBF4 as catalyst gave an identical deuteration
pattern to that observed when [IrCp*Cl2]2 alone was used as
catalyst. This suggests that the addition of NaBF4 accelerates the
reaction without modifying its mechanism.
from the reaction in a low yield. One of the reasons for the
low yield was the significant loss of the vinyl unit from 2
during chromatography, generating 2-phenylindole (3). This was
confirmed by monitoring the decomposition of 2 in slightly acidic
solutions. Acidic decomposition generated 3 quantitatively over a
period of several days at room temperature.
The second product, the yellow solid isolated from the reaction,
was recrystallised from hexane and CH2Cl2 to yield orange crystals
of complex 4 of crystallographic quality. The 1H NMR spectrum
of complex 4 in acetone-d6 showed the presence of one molar
equivalent of bound Cp* per mole of 1, and confirmed that the
aniline 1 retained both of its hydrogens, with the NH2 resonance
integrating to two protons. The resonance due to the NH2 of the Ir
bound aniline 1 is shifted downfield to 5.80 ppm compared to the
NH2 resonance of the free aniline 1 which occurs at 5.11 ppm. 1H-
1H NOESY NMR spectroscopy also showed a clear correlation
between the methyl groups of the Cp* unit and the protons of the
NH2 group of the aniline 1 confirming that a complex had been
isolated which contained 1 bound to the Ir(III) metal centre.
The solid state structure of 4 was obtained using X-ray
crystallography and is shown with thermal ellipsoids at the 30%
probability level, Fig. 3. Selected bond lengths and angles for 4
are presented in Table 1 and crystal structure and refinement data
are given in the experimental section. The crystallographic asym-
metric unit consists of one [IrCp*Cl2(aniline-kN)] molecule and
one dichloromethane molecule of solvation. The crystal contains
only one enantiomer of the molecule. In solution the compound is
Reactions of 2-phenylindole (3)
In order to rule out Pathway I completely, 2-phenylindole 3 was
heated at 323 K with 5 mol% [IrCp*Cl2]2 and NaBF4 in acetone
for 18 hours. At the end of the reaction, only the starting material
was detected and no discernible decomposition was observed.
It is therefore unlikely that the reaction route involves initial
hydroamination followed by addition of acetone (Pathway I).
The role of solvent
In order to investigate the role of the solvent on the mechanism
depicted in Fig. 1, the reaction of [IrCp*Cl2]2 with aniline 1 was
tried using the following solvent mixtures: thf-d8, thf-d8 in the
presence of 1 molar equivalent of H-acetone, and thf-d8 in the
presence of 3.5 molar equivalents of H-acetone.
In the first case, the aniline 1 was reacted in thf-d8 with 2.5 mol%
[IrCp*Cl2]2 and NaBF4 at 323 K and monitored via H NMR
1
spectroscopy. After 18 hours neither conversion to 2-phenylindole
(3) nor decomposition of 1 was observed. The same result was
obtained when 1 was heated at 323 K in thf-d8 with one molar
equivalent of H-acetone in presence of [IrCp*Cl2]2. However, when
the reaction was repeated with the amount of acetone increased
to 3.5 molar equivalents, a very slow conversion of the aniline 1
to a mixture of the vinyl indole product 2 and 2-phenylindole 3
(1:2) was observed after 7 hours with a yield of 48%. Therefore
hydroamination of 1 yielding 3 (first step of Pathway I) is possible,
suggesting that intermediate A is accessible in these conditions.
However, this reaction appears to be blocked when the solvent is
pure acetone. Product 3 is not formed in the absence of acetone,
which seems to indicate that acetone is required to break the Ir
dimer, [IrCp*Cl2]2.
Isolation of [IrCp*Cl2(aniline-jN)] (4)
When the reaction of 1 with acetone was performed with a large
amount of [IrCp*Cl2]2 (5 mol%) and NaBF4 at 323 K, isolation
and characterization of the intermediate [IrCp*Cl2(aniline-kN)]
(4) was possible. The product, vinylindole 2, was also isolated
Fig.
3 Anisotropic displacement ellipsoid plot of [(C10H15)IrCl2-
(C14H11N)], 4, with labeling of selected atoms. Ellipsoids show 30%
probability levels. Hydrogen atoms have been deleted for clarity.
10298 | Dalton Trans., 2009, 10296–10304
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
◦
a
˚
tion of [IrCp*Cl2]2 occurred as red crystals. This shows that 4 is
an active species in acetone and is on the reaction path to yield 2.
Table 1 Selected bond lengths (A) and bond angles ( ) for the inner
coordination sphere of 4
˚
Atomic distance (A)
Ir1–Cl1
Ir1–Cl2
Ir1–N1
Computational studies
2.4157 (8)
2.4179 (8)
2.179 (2)
1.192 (4)
1.433 (4)
Ir1–C15
Ir1–C16
Ir1–C17
Ir1–C18
Ir1–C19
2.172 (3)
2.153 (3)
2.155 (3)
2.158 (3)
2.141 (3)
Thermodynamic considerations. The main difference between
Pathways I and II as they are described earlier in the paper (Fig. 1)
is that, in the former, the reaction starts with the cyclisation of
aniline 1 to give indole 3, while in the latter the reaction starts
with the formation of the imine (5). The thermodynamic values
determined for the transformations of the organic molecules alone
are shown in Scheme 2 for both pathways. The formation of the
final vinylindole (2) via the 2-phenylindole (3) is unlikely because of
the endothermicity of the 3 to 2 transformation (17.7 kcal mol-1).
Therefore the formation of this species should avoid formation of
3 which would act as a trap. This agrees with the experimental data
that showed that in neat acetone and in the presence of catalyst,
3 is not converted to 2 and allows us to disregard Pathway I as a
possible route for the formation of 2.
Pathway II (Fig. 1) starts with the formation of intermediate
B where the imine coordinated to Ir comes from the reaction
of 1 with acetone. The organic reaction of 1 + acetone → 5
+ H2O is moderately endothermic as shown in Scheme 2, but
the strong thermodynamic drive associated with the cyclisation
to give 2 makes the step from 1 to 5 energetically feasible. As
the condensation of an amine and acetone can take place in the
absence of any metal, we did not study the condensation reactions,
and we compare the energy profile for the imine cyclisation with
that obtained for the amine cyclisation.
C7–C8
C1–N1
Bond angle (◦)
Cl1–Ir1–Cl2
Cl1–Ir1–N1
Cl2–Ir1–N1
91.00 (3)
79.31 (7)
81.60 (7)
C6–C7–C8
C7–C8–C9
Ir1–N1–C1
176.5 (4)
178.5 (4)
119.44 (19)
a Estimated standard deviation in the least significant figure are given in
parentheses.
presumably racemic but it appears to have spontaneously resolved
on crystallisation. The solid state structure of 4 is analogous to
the structure previously reported for [IrCp*Cl2(NH2Ph)],20 and
all bond distances are comparable. The only significant difference
between the two complexes is a slightly smaller Ir1–N1–C1 angle
119.44(19)◦ in 4 compared 121.40◦ for [IrCp*Cl2(NH2Ph)].
The fact that substrate 1 is N-bound to the Ir(III) metal centre,
rather than p-bound is consistent with the amine being a stronger
Lewis base than the alkyne. Structural rearrangement would be
necessary for the reaction to proceed to completion.
Reactivity of [IrCp*Cl2(aniline-jN)] (4)
In order to probe if 4 is an intermediate for the formation of 3,
complex 4 was dissolved in CD2Cl2, heated at 323 K and monitored
via 1H NMR spectroscopy. No conversion to 2-phenylindole
3 was observed, indeed no change was observed at all in the
complex over a period of 18 h at 323 K. Coordinated aniline,
as in 4, cannot achieve hydroamination because the alkyne is
not activated towards nucleophilic addition, and also because
the nucleophilicity of the amine has been removed. This is in
agreement with the proposal by Crabtree et al. that coordination
through the nitrogen of ethynylanilines leads to unproductive
binding rather than hydroamination.9 However the formation of
3 when the reaction takes place with 3.5 molar equivalents of
acetone in thf suggests that the change in the coordination from
As has been suggested by the experiments the fragmentation of
the Ir dimer, [IrCp*Cl2]2, is assisted by acetone. This process has
not been studied computationally and the unsaturated CpIrCl2 has
been considered as the reactive form of the catalyst. This fragment
can coordinate acetone, aniline (1) or imine (5).
Cyclisation of the amine vs. the imine. The following notation
has been introduced. All species derived from the amine cyclisation
(i.e. Pathway I, Fig. 1) will be labelled with letter a, while all species
originating from the imine (i.e. Pathway II, Fig. 1) will be labelled
with the letter i.
1
2
k (N) to h (CC) has to be possible at least in the presence of
acetone. Therefore, the acetone appears to be also necessary for
the hydroamination process, not only for the fragmentation of the
Ir dimer.
The amine and imine cyclisations start with the coordination
of 1 and 5 to [CpIrCl2], respectively. The unsaturated CpIrCl2
fragment coordinates preferentially the most Lewis basic site of
either 1 or 5. It thus coordinates preferentially to the nitrogen to
1
2
Complex 4 was also heated in acetone-d6 at 323 K and within 3
minutes complete conversion of the metal complex was observed.
An organic product was identified in the reaction mixture as N-
form a k (N) complex rather than to the alkyne to form a h (CC)
complex. Hence, complex a1 (see Scheme 3) is the initial reactant
for the amine cyclisation and it is also used as the reference point
for all energies throughout the remainder of the paper. Species i1
1
(2-methylvinyl)-2-phenylindole (2) by H NMR while precipita-
Scheme 2 Reactions that could account for the formation of vinylindole 2 via Pathway I (left side) and Pathway II (right side) in Fig. 1. Energies are
given in kcal mol-1 and include the solvent (acetone) effect.
This journal is
The Royal Society of Chemistry 2009
Dalton Trans., 2009, 10296–10304 | 10299
©

View Article Online
Scheme 3 Cyclisation mechanism of amine (top) and imine (bottom) including the intramolecular proton transfer. Energies are given relative to a1,
in kcal mol-1 and include the solvent (acetone) effect.
is the reactant for the imine cyclisation. It is 13.4 kcal mol-1 higher
in energy than a1 because the formation of the imine and water
from the amine and acetone is endothermic and because the imine
nitrogen is a weaker Lewis base than the amine nitrogen.
To form the final product, either 2 or 3, a proton has to be
transferred to C3 of the indole. As indicated earlier, the source of
the proton is geometrically far from C3 since the proton should
come either from the nitrogen in the case of Pathway I, Fig. 1,
(via the amine) or from one of the methyl groups of the iminium
group. No direct proton intramolecular transfer from reactant to
product was identified and, in all cases, the only possible routes
found using the calculations were multi-step processes. In the case
of Pathway I, a two-step process is necessary to form 3: the proton
migrates from N to C2 of the indole (from a3 to a4) with an
energy barrier of 35.1 kcal mol-1 followed by a transfer from a4 to
a5 with a rather low energy barrier of 16.4 kcal mol-1. In the case
of the imine, a similar path takes place between i3 and i5, but the
TS energy for the first step has an even higher energy (TSi3-4 =
43.6 kcal mol-1).21
From a1 and i1, the first cyclisation mechanism that we
considered is shown in Scheme 3. The mechanism can be divided in
three steps: 1) change from N coordination to alkyne coordination;
we did not search for a transition state in this step, which is a
simple ligand substitution; 2) intramolecular nucleophilic addition
of N to the activated alkyne to form the 5-membered ring; and 3)
intramolecular proton transfer to replace the Ir-C bond by a C-H
bond of the final products 3 or 2. In the amine cyclisation route the
proton originates from the NH2 group, and in the imine cyclisation
route the proton originates from the CH3 group of acetone. In
the first instance, we selected this intramolecular proton transfer
because it readily provides a suitable explanation for the selective
deuteration at C3 of the indole when using acetone-d6.
These intramolecular proton transfers with high energy barriers
are unlikely to occur. In another hydroamination study reported
recently, it is proposed that the proton transfer is assisted by the
The nucleophilic addition of the nitrogen to the alkyne
(Scheme 3) starts with a change in the coordination of the amine
counterion (OTf-),22 and in this case the BF4 anion is not very
-
1
2
or imine, from k (N) to h (CC). This process is energetically easier
from i1 to i2 (8.4 kcal mol-1), than from a1 to a2 (10.5 kcal mol-1)
because the imine is less basic than the amine. The next step,
which produces the 5-membered ring, is the nucleophilic addition
of a Lewis base to an activated multiple bond. Remarkably, the
energy barrier for the cyclisation step is lower with imine, TSi2-3
(3.7 kcal mol-1) than with amine, TSa2-3 (10.9 kcal mol-1). This is
due to the fact that the N lone pair of a2 is conjugated with the
phenyl group and a 90◦ rotation of the NH2 group around the C-N
bond is needed for the amine to add to the alkyne. Such rotation is
not needed with i2. Despite the fact that the energy barrier for the
cyclisation step is lower in the case of the imine cyclisation route,
the energies of the transition states for amine and imine cyclisation
are close in energy relative to reactants (energy of 21.4 kcal mol-1
for TSa2-3 and of 25.5 kcal mol-1 for TSi2-3), and the energy
of the transition state for the amine cyclisation is slightly lower.
Thus, while the cyclisation of the imine 5 is energetically easier
than that of the amine 1, the energy required to form 5 disfavours
the cyclisation step in Pathway II.
likely to play this role. The species a3 or i3 are rather acidic, and
intermolecular proton transfer to a molecule which can act as a
base such as the amine or the acetone, is more likely to occur.
Although the amine is more basic than acetone, the molar ratio of
acetone: 1 is large, making the acetone a candidate for promoting
the transfer. In addition the stereoselectivity of the proton transfer
from acetone to C3, demonstrated by the use of acetone-d6, can
only be understood if acetone assists the proton transfer. Therefore
we considered this possibility and the energies obtained are shown
in Scheme 4.
The proton transfer from either a3 or i3 to acetone gives an
anionic Ir complex and a cationic protonated acetone which can
form ion pairs. Either for a3 or for i3, we found that the transition
state for forming the ion pair a4ass or i4ass (ass for assisted by the
solvent) is very close in energy to that of the ion pairs themselves.
Likewise the energy of the transition state for transferring the
proton from the protonated acetone to C3 is almost at the same
energy as the ion pair. The calculations show that a4ass is only
6.0 kcal mol-1 above a3 and that i4ass is only 17.6 kcal mol-1
10300 | Dalton Trans., 2009, 10296–10304
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
Scheme 5 Possible reactants for the cyclisation of amine or formation
of imine. Energies are given relative to a1, in kcal mol-1 and include the
solvent (acetone) effect.
conditions with equal quantities of the amine 1 and acetone. This
is not a realistic model of the experimental conditions; the large
excess of acetone, used as solvent, should displace the equilibrium
in favor of o1, and thus, the concentration of o1 could dominate
over a2 and compete with a1. So, o1 should be considered as a
possible intermediate in the formation of the imine. Calculations
show that this result does not change on considering other possible
complexes in which one of the chlorides has been replaced by a
neutral ligand such as acetone (details in the ESI†).
We then considered an iridium assisted formation of the imine
using o1 and the amine as the starting reactants, as depicted
in Scheme 6. The first step is the nucleophilic addition of the
amine 1 to the activated acetone. This step, with an energy barrier
of 10 kcal mol-1 (TSo1-2), gives intermediate o2 with an acidic
proton bound to N. From this species two proton migrations from
nitrogen to oxygen giving finally H2O and coordinated imine (i1
or A in Fig. 1) can occur. These proton transfers can take place
through intermediates o2 and o3 with energies under 20 kcal mol-1.
The proton transfer between o2, o3, o4 and the loss of water have
not been studied in further detail. Such steps are well known,
and probably take place with the assistance of acetone with low
energy barriers as shown in the case of the amine proton transfer
(TSa3-4ass).
Scheme 4 Proton transfer of amine (top) and imine (bottom) assisted by
acetone. Energies are given relative to a1, in kcal mol-1 and include the
solvent (acetone) effect.
above i3. These two ion pairs thus are significantly lower in energy
than the transition states TSa3-4 or TSi3-4. Consequently, an
intermolecular proton transfer assisted by acetone is preferred
over an intramolecular proton transfer. This role of the acetone
can explain why the experimental complex 4 (modelled by a1) does
not give hydroamination in pure CD2Cl2 solvent and requires at
least 3.5 molar equivalents of acetone.
Looking at the whole path, it appears that the highest transition
state on Pathway I for the formation of the amine 3 (TSa2-3 =
21.4 kcal mol-1) is slightly lower than the highest transition state
for the formation of the vinyl imine 2 on Pathway II (TSi3-4ass =
27.1 kcal mol-1). Thus at least a mixture of 2 and 3 should be
observed in the reaction of 1 with [IrCp*Cl2]2 as does happen
when reaction of 1 with [IrCp*Cl2]2 takes place in thf-d6 with
3.5 molar equivalents of acetone. However, since only 2 is observed
when acetone is the solvent, we have re-considered the pathway for
formation of imine, or intermediate A in Pathway II, and search
for a possible rationalization.
Using these results, we propose that when the reaction takes
place in pure acetone the iridium fragment coordinates preferen-
2
tially to acetone rather than the amine by h coordination. As a
consequence, the amine reacts preferentially with the coordinated
acetone to form the imine and not with the iridium fragment to
form the non-vinylic indole, 3. In other words the large excess
of acetone blocks the regular hydroamination process and only
allows the formation of 2.
Mechanism for the imine formation catalyzed by CpIrCl2. The
formation of the imine comes from the condensation of the amine
with acetone with loss of water. The electrophilicity of acetone is
increased following coordination of the oxygen to a Lewis acid,
which in this reaction can be the Ir catalyst. As has been described
in the experimental section above, acetone is required to dissociate
the dinuclear [Cp*IrCl2]2 and to form the mononuclear acetone
adduct [Cp*IrCl2(acetone)], modelled by o1 (Scheme 5). Species
o1 is 8.9 kcal mol-1 less stable than a1 and 1.6 kcal mol-1 more stable
than a2. This energy pattern is obtained considering stoichiometric
Proposed mechanism for the formation of 2. From the com-
putational studies and the comparison with experimental data, a
detailed catalytic cycle, slightly different from Pathway II proposed
in Fig. 1, is shown in Scheme 7. The first step is the fragmentation
of [Cp*IrCl2]2 by acetone. Two Lewis bases can compete with
Scheme 6 Mechanism proposed for the imine formation assisted by [CpIrCl2] complex. Energies are given relative to a1, in kcal mol-1 and include the
solvent (acetone) effect.
This journal is
The Royal Society of Chemistry 2009
Dalton Trans., 2009, 10296–10304 | 10301
©

View Article Online
seems that this amount of acetone is enough to give fragmentation,
formation of imine and proton transfer leading 2 and 3, but it
is not enough to completely avoid the formation of a2, which
subsequently leads to the formation of 3.
Conclusions
This combined experimental computational study of the un-
usual formation of vinyl indole by the reaction of 2-(2-
phenylethynyl)aniline with acetone in presence of [IrCp*Cl2)2
suggests a mechanistic route, in which acetone plays a key role
in several steps of the multistep pathways. It participates in the
fragmentation of the dinuclear Ir complex and the formation
of the reactive form of the catalyst. It blocks the formation
of the expected hydroamination product by coordination to
the Ir catalyst. Coordinated acetone reacts with the aniline to
form an imine derivative, which yields the final product after
proton transfer helps by acetone itself. This mechanistic proposal
accounts for all experimental data carried out to analyze this
mechanistic issue.
Experimental
Scheme 7 Catalytic cycle proposed for the reaction of 1 with acetone to
give 2, catalyzed by [Cp*IrCl2]2.
General procedures
Reagents were purchased from Aldrich and used as received unless
otherwise stated. All solvents were pre-dried and distilled under
an atmosphere of argon. Acetone and acetone-d6 were dried
with CaSO4 and distilled under argon prior to use. Methanol
the empty coordination site of CpIrCl2; the phenyl amine is
the strongest base and it gives complex a1 (model of 4), with
1
amine k (N) coordinated, but acetone, which is in large excess,
˚
may favor formation of the acetone complex o1. The latter
complex has an electrophilically active carbon, which can react
with the amine. After several proton shuffles and loss of water, an
was dried over 4 A molecular sieves and distilled from CaH2.
Thf-d8 was dried over sodium/benzophenone and distilled prior
to use. 1,1,2,2-Tetrachloroethane was dried with CaSO4 and
distilled under argon. 1,2,3,4,5-Pentamethylcyclopentadiene was
purchased from Lancaster and used without further purification.
Iridium(III) chloride hydrate was obtained from Precious Metals
Online (PMO) and used without further purification. [IrCp*Cl2]2
was synthesised according to the literature.23 1H NMR spectra
were recorded on a Bruker DMX500 spectrometer. All spectra
were recorded at 323 K unless otherwise specified. 1H NMR
chemical shifts were referenced internally to residual solvent
resonances. MS was performed by Biomedical Mass Spectrometry
Facility at the University of New South Wales, Sydney and X-Ray
Crystallographic analysis was performed by Dr Anthony Willis at
the Research School of Chemistry, Australian National University,
Canberra.
1
imine i1 is formed. This imine can easily exchange from k -N to
2
2
h -CC coordination (i1 to i2). The h -CC coordinated imine has
an activated alkyne and reactive N center which results in the
cyclisation with an accessible energy barrier and the formation of
a vinyl indole ring metallated at the C3 position. The resulting
compound i3 is neutral but is both a strong acid at the methyl
of the iminium group and a strong base through the metallated
C3 carbon. Intramolecular proton transfers are found to have
prohibitively high energy barriers. Loss of proton to the solvent
(acetone) and subsequent protonation would be preferred. This
will lead to the product i5 with a favorable energy of reaction and
account for the selective deuteration at C3 of the indole 2. The
liberated CpIrCl2 species is coordinated by acetone, once again
forming the catalyst and continuing with the catalytic cycle.
This new catalytic cycle shows the importance of acetone in the
formation of 2. Thus, acetone participates in the fragmentation of
complex [Cp*IrCl2]2. By coordination of Ir, it blocks the formation
X-ray crystallography
Diffraction images were measured at 200 K on a Nonius Kap-
2
of the h (CC) bonded aniline, a2. Acetone is one of the reactants
paCCD diffractometer (MoKa, graphite monochromator, l =
24
˚
in the imine formation and it also participates in the proton
transfer that takes place in the imine cyclisation. This accounts
for the experimental results observed when different amounts of
acetone are used. When 1 reacts with [Cp*IrCl2]2 in thf-d8 there
is no reaction because acetone is required for the Ir complex
fragmentation. However, 1 does not react with [Cp*IrCl2(aniline-
kN)] (4) in CD2Cl2 probably because acetone is also needed for
the proton transfer. The reaction of 1 with [Cp*IrCl2]2 and 3.5
molar equivalents of acetone gives 2 and 3 in a 1:2 molar ratio. It
0.71073 A) and data extracted using the DENZO package. The
structure was solved by direct methods (SIR92),25 and refined
using the CRYSTALS program package.26 The displacement
ellipsoid diagram was prepared with use of ORTEP-II.27
Crystal data: orange block, 0.26
¥ 0.11 ¥ 0.09 mm,
C24H26Cl2IrN.CH2Cl2, Mr = 676.53, orthorhombic, space group
˚
P212121, a = 8.8981 (1), b = 11.8148 (2), c = 24.1539 (3) A, V =
3
-3
-1
˚
2539.28 (6) A , Z = 4, Dx = 1.770 g cm , m = 56.92 cm . 29393
reflections measured, 2qmax = 55◦, h = -11 → 11, k = -15 →15,
10302 | Dalton Trans., 2009, 10296–10304
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
l = -31 → 31, analytical absorption correction applied: Tmin
0.367, Tmax = 0.661, 5831 independent reflections, Rint = 0.028.
=
compounds. The compound 4 was isolated as a beige solid (143 mg,
0.7 mmol, 16% Yield). Despite complete conversion to the vinyl
indole 2 significant amounts of 2-phenylindole 3 were also isolated
due to decomposition during chromatography. The compound 2
was found to be unstable in acidic solutions losing the vinyl group
generating 2-phenylindole 3 which was identified by comparison
of the 1H NMR data obtained to that of a true sample of
2-phenylindole. The characterisation of this compound has been
reported previously.19
Full matrix least squares refinement on F2, Chebychev polyno-
mial weighting scheme,28,29 R(5566 reflections with I > 2s(I)) =
0.0163, wR(all data) = 0.0302, S = 0.9414, Flack parameter =
-3
˚
-0.019 (4), Drmax = 0.79, Drmin = -1.00 e A .
General procedure for the catalysed cyclisation of 1
The cyclisation of 1 to give 2, as catalysed by [IrCp*Cl2]2, was
performed on a small scale in an NMR tube fitted with Young’s
concentric Teflon valves. The substrate and catalyst were weighed
and the solvent was vacuum transferred into the tube. The
tube was placed under argon and a known amount of 1,1,2,2-
tetrachloroethane (ca. 50 mg) was added as an internal standard.
All reactions were performed with approximately 5 mol% catalyst
loading (w.r.t. Ir) at 50 ◦C by heating the tube within the probe of
the NMR spectrometer. The temperature within the magnet was
calibrated using ethylene glycol. 1H NMR spectra were recorded
and the conversion to N-(2-methylvinyl)-2-phenylindole 2 was
determined by integration of the product resonance (at 7.82 ppm)
relative to the internal standard (at 6.55 ppm). Conversion of
> 95% is taken as the time where no remaining substrate
resonances were evident.
Computational details
DFT calculations were carried out using the hybrid B3PW91
functional.30 Two different basis set, I and II, were used. Basis set
I (ECP-adapted SDDALL31 with a set of polarization functions
for Ir32 and Cl33 and the all-electron 6-31G(d,p)34 for N, O, C,
and H) was used to optimize the geometries, and basis set II (the
same as I but using the all-electron 311+G(d,p)35 for N, O, C, and
H) to refine the energies of stationary points through single-point
calculations. Geometry optimizations were carried out without
any geometrical constraints. The nature of all stationary points
was confirmed by an analytical calculation of frequencies. Each
transition state was relaxed toward reactant and product using
the analytical vibrational data to confirm its nature. The effect of
the acetone solvent (e = 20.7), was evaluated using the continuum
IEFPCM model.36 All energies given in the text include the solvent
effect and were obtained with basis set II. All calculations were
carried out with the Gaussian03 package.37 The geometries of
intermediates and transition states having no remarkable features
are not presented in the paper. These structures are available in
the ESI.†
Reaction of 2-phenylindole 3 and acetone in the presence of
[IrCp*Cl2]2/NaBF4
2-Phenylindole 3 (103.5 mg, 0.522 mmol), [IrCp*Cl2]2 (10.1 mg,
0.0126 mmol) and NaBF4 (8.4 mg, 0.0766 mmol) were dissolved
in acetone (5 mL) and heated at 50 ◦C for 18 hours under an
atmosphere of argon. The solvent was removed in vacuo and
analysis of the crude reaction mixture by 1H NMR spectroscopy
in acetone-d6 showed no conversion or decomposition of the
2-phenylindole 3.
We considered that the dinuclear Ir complex splits into two
monomers which was modelled by CpIrCl2, where Cp = C5H5
and we did not analyze the formation of CpIrCl2. All organic
reactants were modelled in full.
Isolation of [IrCl2(aniline-jN)Cp*] 4
Acknowledgements
Under argon 1 (828 mg, 4.28 mmol), [IrCp*Cl2]2 (209 mg,
0.262 mmol) and NaBF4 (34 mg, 0.31 mmol) were dissolved in
10 mL acetone. The mixture was heated at reflux at 55 ◦C for
18 h. The solution was cooled and the reaction mixture filtered.
The yellow solid obtained was isolated and recrystallised by slow
diffusion of pentane into a concentrated CH2Cl2 solution. After
one week orange crystals were obtained (45.3 mg, 0.0771 mmol,
29%). The compound was identified to be the substrate bound
to the Ir metal centre generating the complex [Cp*IrCl2(2-
(2-phenylethynyl)aniline)]. The orange crystals obtained were
suitable for X-ray crystallographic analysis. C24H26Cl2IrN·CH2Cl2
(676.54): calcd. C, 44.38; H, 4.17; N, 2.07; found: C, 45.25; H,
4.32; N, 2.35 %. MS (ES+) m/z: 593 (M+, 10%), 556 (M–Cl, 80%),
520 (M–2Cl, 100%). 1H NMR (CD2Cl2, 300 MHz, 298 K): d 7.72
(m,1H, H6), 7.52 (m, 1H, H3), 7.46-7.31 (m, 6H, H5 + Ph), 7.15
(m, 1H, H4), 5.84 (br s, 2H, NH2), 1.38 (s, 15H, Cp*) ppm. 13C
NMR (CD2Cl2, 75 MHz, 298 K): d 145.0 (C2), 132.4 (C6), 132.0
(C4), 130.2 (C5), 129.6 (Ph), 129.1 (Ph), 125.3 (C4), 122.8 (i-Ph),
We gratefully acknowledge financial support from The University
of New South Wales and The Australian Research Council.
We also thank the Australian government for an Australian
Postgraduate Award (D.F.K.). BAM and OE thank the RSC
for international grants for authors. A.N. thanks the Spanish
MICINN for a MEC postdoctoral fellowship. The authors thank
S.A. Macgregor (Heriot Watt University, UK) and D. St Claire
Black (UNSW, Australia) for useful discussions.
References
1 Indoles, ed. R. J. Sundberg, Academic Press, London, 1996.
2 G. W. Gribble, Comprehensive Heterocyclic Chemistry II, ed. A. R. Ka-
tritzky, C. W. Ress, E. F. V. Scriven and C. W. Bird, Pergamon Press,
Oxford, 1996, Vol. 2, p. 207.
3 D. A. Horton, G. T. Bourne and M. L. Smythe, Chem. Rev., 2003, 103,
893–930 and references therein.
4 G. R. Humphrey and J. T. Kuethe, Chem. Rev., 2006, 106, 2875–2911.
5 R.J. Sundberg, Pyrroles and Their Benzoderivatives: Synthesis and Ap-
plications. In Comprehensive Heterocyclic Chemistry, A. R. Katritzky
and C. W. Rees, Eds., Pergamon, Oxford, U.K., 1984, Vol. 4, pp 313-
376.
6 P. Burger and R. G. Bergman, J. Am. Chem. Soc., 1993, 115, 10462–
10463.
≡
≡
120.6 (Ph), 115.4 (C1), 98.0 (C C), 86.1 (Cp) 86.2 (C C), 9.1
(Cp–Me) ppm.
The filtrate was purified on a small silica column using
Et2O: Light Petroleum (1:1) as eluent to remove remaining Ir
This journal is
The Royal Society of Chemistry 2009
Dalton Trans., 2009, 10296–10304 | 10303
©

View Article Online
7 D. M. Tellers and R. G. Bergman, Organometallics, 2001, 20, 4819–
4832.
8 X. Li, T. Vogel, C. D. Incarvito and R. H. Crabtree, Organometallics,
2005, 24, 62–76.
28 D. J. Watkin, Acta Crystallogr., Sect. A: Found. Crystallogr., 1994, 50,
411–437.
29 E. Prince, Mathematical Techniques in Crystallography and Materials
Science, Springer-Verlag, New York, 1982.
9 X. Li, A.R. Chianese, T. Vogel and R. H. Crabtree, Org. Lett., 2005, 7,
5437–5440.
10 N. A. Wynberg, L. J. Leger, M. L. Conrad, C. M. Vogels, A. Decken,
S. J. Duffy and S. A. Wescott, Can. J. Chem., 2005, 83, 661–667.
11 K-I. Fujita, Z. Li, N. Ozeki and R. Yamaguchi, Tetrahedron Lett., 2003,
44, 2687–2690.
30 (a) A. D. Becke, J. Chem. Phys., 1993, 98, 5648; (b) J. P. Perdew
and Y. Wang, Phys. Rev. B: Condens. Matter Mater. Phys., 1992, 45,
13244.
31 (a) D. Andrae, U. Hau¨ssermann, M. Dolg, H. Stoll and H. Preuss,
Theor. Chim. Acta, 1990, 77, 123; (b) A. Bergner, M. Dolg, W. Ku¨chle,
H. Stoll and H. Preuss, Mol. Phys., 1993, 80, 1431.
12 D. Balcells, A. Nova, E. Clot, D. Gnanamgari, R. H. Crabtree and O.
Eisenstein, Organometallics, 2008, 27, 2529–2535.
32 A. W. Ehlers, M. Bo¨hme, S. Dapprich, A. Gobbi, A. Ho¨llwarth, V.
Jonas, K. F. Ko¨hler, R. Stegmann, A. Veldkamp and G. Frenking,
Chem. Phys. Lett., 1993, 208, 111.
33 A. Ho¨llwarth, H. Bo¨hme, S. Dapprich, A. W. Ehlers, A. Gobbi, V.
Jonas, K. F. Ko¨hler, R. Stegmann, A. Veldkamp and G. Frenking,
Chem. Phys. Lett., 1993, 208, 237.
13 M. Furegati and A. J. Rippert, Tetrahedron: Asymmetry, 2005, 16,
3947–3950.
14 K-I. Fujita, C. Kitatsuji, S. Furukawa and R. Yamaguchi, Tetrahedron
Lett., 2004, 45, 3215–3217.
15 F. Brustolin, V. Castelvetro, F. Ciardelli, G. Ruggeri and A. Colligiani,
J. Polym. Sci., Part A: Polym. Chem., 2001, 39, 253–262.
16 Y. Zhang, J. P. Donahue and C.-J. Li, Org. Lett., 2007, 9, 627–630.
17 M. Movassaghi and A. E. Ondrus, J. Org. Chem., 2005, 70, 8638–8641.
18 A. Y. Lebedev, V.V. Izmer, D. N. Kazyul’kin, I. P. Beletskaya and A. Z.
Voskoboynikov, Org. Lett., 2002, 4, 623–626.
19 D. F. Kennedy, B. A. Messerle and S. L. Rumble, New J. Chem., 2009,
33, 818–824.
20 D. L. Davies, O. Al-Duaij, J. Fawcett, M. Giardiello, S. T. Hilton and
D. R. Russell, Dalton Trans., 2003, 4132–4138.
21 Similar results were obtained with other systems such as cationic species
were one chloride is replaced by acetone [CpIrCl(acetone)]+ (see ESI†).
22 G. Kovacs, G. Ujaque and A. Lledos, J. Am. Chem. Soc., 2008, 130,
853–864.
23 C. White, A. Yates and P. M. Maitlis, Inorg. Synth., 1992, 29, 228.
24 Z. Otwinowski and W. Minor (1996). Processing of X-ray Diffraction
Data Collected in Oscillation Mode. Methods Enzymol. Eds C. W. Carter
and R. M. Sweet, Academic Press. 1997, 276, 307-326.
25 A. Altomare, G. Cascarano, C. Giacovazzo, A. Guagliardi, M. C. Burla,
G. Polidori, G. and M. Camalli, J. Appl. Crystallogr., 1994, 27, 435–435.
26 P. W. Betteridge, J. R. Carruthers, R. I. Cooper, K. Prout and D. J.
Watkin, J. Appl. Crystallogr., 2003, 36, 1487–1487.
34 P. C. Hariharan and J. A. Pople, Theor. Chim. Acta, 1973, 28, 213.
35 M. J. Frisch, J. A. Pople and J. S. Binkley, J. Chem. Phys., 1984, 80,
3265.
36 (a) M. T. Cance`s, B. Mennucci and J. Tomasi, J. Chem. Phys., 1997,
107, 3032; (b) M. Cossi, V. Barone, B. Mennucci and J. Tomasi, Chem.
Phys. Lett., 1998, 286, 253; (c) B. Mennucci and J. Tomasi, J. Chem.
Phys., 1997, 106, 5151.
37 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A.
Robb, J. R. Cheeseman, J. A. Montgomery, Jr., T. Vreven, K. N.
Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J. Tomasi, V. Barone,
B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G. A. Petersson, H.
Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa,
M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X.
Li, J. E. Knox, H. P. Hratchian, J. B. Cross, V. Bakken, C. Adamo,
J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin,
R. Cammi, C. Pomelli, J. Ochterski, P. Y. Ayala, K. Morokuma, G. A.
Voth, P. Salvador, J. J. Dannenberg, V. G. Zakrzewski, S. Dapprich,
A. D. Daniels, M. C. Strain, O. Farkas, D. K. Malick, A. D. Rabuck,
K. Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul,
S. Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko, P.
Piskorz, I. Komaromi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-
Laham, C. Y. Peng, A. Nanayakkara, M. Challacombe, P. M. W.
Gill, B. G. Johnson, W. Chen, M. W. Wong, C. Gonzalez and J. A.
Pople, GAUSSIAN 03 (Revision E.1), Gaussian, Inc., Wallingford, CT,
2004.
27 C. K. Johnson, ORTEP–II, A Fortran Thermal-Ellipsoid Plot Program,
Report ORNL-5138, Oak Ridge National Laboratory, Oak Ridge,
Tennessee, USA, 1976.
10304 | Dalton Trans., 2009, 10296–10304
This journal is
The Royal Society of Chemistry 2009
©