Transfer of amido groups from isolated rhodium(I) amides to alkenes and vinylarenes

Article abstract of DOI:10.1021/ja052473h

The reaction of monomeric and dimeric rhodium(l) amido complexes with unactivated olefins to generate imines is reported. Transamination of {(PEt ₃)₂RhN(SiMePh₂)₂} (1a) or its -N(SiMe₃)₂ analogue 1b with p-toluidine gave the dimeric [(PEt₃)₂Rh(M-NHAr)]₂ (Ar = p-tolyl) (2a) in 80% isolated yield. Reaction of 2a with PEt₃ generated the monomeric (PEt₃)₃Rh(NHAr) (Ar = p-tolyl) (3a). PEt ₃-ligated arylamides 2a and 3a reacted with styrene to transfer the amido group to the olefin and to form the ketimine Ph(Me)-C=N(p-tol) (4a) in 48-95% yields. The dinuclear amido hydride (PEt₃)₄Rh ₂(μ-NHAr)(μ-H) (Ar = p-tolyl) (5a) was formed from reaction of 2a in 95% yield, and a mixture of this dimeric species and the (PEt ₃)_nRhH complexes with n = 3 and 4 was formed from reaction of 3a in a combined 75% yield. Propene reacted with 2a to give Me ₂C=N(p-tol) (4b) and 5a in 90 and 57% yields. Propene also reacted with 3a to give 4b and 5a in 65 and 94% yields. Analogues of 2a and 3a with varied electronic properties also reacted with styrene to form the corresponding imines, and moderately faster rates were observed for reactions of electron-rich arylamides. Kinetic studies of the reaction of 3a with styrene were most consistent with formation of the imine by migratory insertion of olefin into the rhodium-amide bond to generate an aminoalkyl intermediate that undergoes β-hydrogen elimination to generate a rhodium hydride and an enamine that tautomerizes to the imine.

Full text of DOI:10.1021/ja052473h

Published on Web 08/09/2005
Transfer of Amido Groups from Isolated Rhodium(I) Amides
to Alkenes and Vinylarenes
Pinjing Zhao, Christopher Krug, and John F. Hartwig*
Contribution from the Department of Chemistry, Yale UniVersity,
P.O. Box 208107 New HaVen, Connecticut 06520-8107
Received April 15, 2005; E-mail: john.hartwig@yale.edu
Abstract: The reaction of monomeric and dimeric rhodium(I) amido complexes with unactivated olefins to
generate imines is reported. Transamination of {(PEt3)2RhN(SiMePh2)2} (1a) or its -N(SiMe3)2 analogue
1b with p-toluidine gave the dimeric [(PEt3)2Rh(µ-NHAr)]2 (Ar ) p-tolyl) (2a) in 80% isolated yield. Reaction
of 2a with PEt3 generated the monomeric (PEt3)3Rh(NHAr) (Ar ) p-tolyl) (3a). PEt3-ligated arylamides 2a
and 3a reacted with styrene to transfer the amido group to the olefin and to form the ketimine Ph(Me)-
CdN(p-tol) (4a) in 48-95% yields. The dinuclear amido hydride (PEt3)4Rh2(µ-NHAr)(µ-H) (Ar ) p-tolyl)
(5a) was formed from reaction of 2a in 95% yield, and a mixture of this dimeric species and the (PEt3)nRhH
complexes with n ) 3 and 4 was formed from reaction of 3a in a combined 75% yield. Propene reacted
with 2a to give Me2CdN(p-tol) (4b) and 5a in 90 and 57% yields. Propene also reacted with 3a to give 4b
and 5a in 65 and 94% yields. Analogues of 2a and 3a with varied electronic properties also reacted with
styrene to form the corresponding imines, and moderately faster rates were observed for reactions of
electron-rich arylamides. Kinetic studies of the reaction of 3a with styrene were most consistent with formation
of the imine by migratory insertion of olefin into the rhodium-amide bond to generate an aminoalkyl
intermediate that undergoes â-hydrogen elimination to generate a rhodium hydride and an enamine that
tautomerizes to the imine.
Introduction
at the M-N bond of isolated late-transition-metal amides.
Casalnuovo reported that an iridium(III) arylamido complex
Transfer of an alkyl group from a transition-metal complex
to an olefin to form a new C-C bond, typically by migratory
insertion, is a common organometallic reaction. Related transfers
of an amido group from an isolated amido complex to an olefin
to form a new C-N bond are rare. A few catalytic reactions
have been reported in which insertion of an olefin into a metal
amide is a likely step.1-4 Although strong evidence for insertion
has been presented in some cases, the precursors to the insertion
process and the immediate products from this step were not
observed directly in these processes. More common are the
reactions of late-transition-metal olefin complexes with free
amines to generate aminoalkyl products.5-8
inserts norborene, but reactions with olefins that lack the strain
of norbornene did not occur.10 Trogler reported the reaction of
a platinum arylamide with acrylonitrile, but reactions with
olefins that lack the polarity of acrylonitrile were not reported.11
Two examples of the insertion of unactivated alkynes into late
transition-metal amides have been reported recently, one into a
molybdenum amide12 and one into a ruthenium amide.13 The
reaction of a palladium amide with the activated dimethylacety-
lene dicarboxylate has also been described.14 However, reactions
with alkenes have not. We report the transfer of an amido group
of a series of isolated rhodium(I) amido complexes to alkenes
and vinylarenes. These reactions generate imines in a new
carbon-nitrogen bond-forming process.
Intramolecular insertion of an olefin into a lanthanide amide
has been observed directly as part of the mechanism of
cyclizations by hydroamination,9 but only activated olefins, such
as norbornene and acrylonitrile, have displayed any reactivity
Results and Discussion
1. Preparation of Rhodium Amido Complexes. A. Syn-
thesis of Dinuclear Amido Complexes. Isolable dimeric
rhodium arylamido complexes ligated by PEt3 were prepared
by the route in Scheme 1. The three-coordinate complex {Rh-
(PEt3)2[N(SiMePh2)2]} (1a) served as an isolable precursor to
(1) Tsutsui, H.; Narasaka, K. Chem. Lett. 1999, 45.
(2) Helaja, J.; Gottlich, R. Chem. Commun. 2002, 720.
(3) Ney, J. E.; Wolfe, J. P. Angew. Chem., Int. Ed. 2004, 43, 3605.
(4) Ney, J. E.; Wolfe, J. P. J. Am. Chem. Soc. 2005, 127, 8644.
(5) Collman, J. P.; Hegedus, L. S.; Norton, J. R.; Finke, R. G. In Principles
and Applications of Organotransition Metal Chemistry; University Science
Books, 1987; p 409.
(6) Ba¨ckvall, J.-E.; Nordberg, R. E.; Zetterberg, K.; Åkermark, B. Organo-
metallics 1983, 2, 1625.
(10) Casalnuovo, A. L.; Calabrese, J. C.; Milstein, D. J. Am. Chem. Soc. 1988,
110, 6738.
(7) Pryadun, R.; Sukumaran, D.; Bogadi, R.; Atwood, J. D. J. Am. Chem. Soc.
2004, 126, 12414.
(11) Cowan, R. L.; Trogler, W. C. Organometallics 1987, 6, 2451.
(12) Katayev, E.; Li, Y. H.; Odom, A. L. Chem. Commun. 2002, 838.
(13) Takemoto, S.; Kobayashi, T.; Matsuzaka, H. J. Am. Chem. Soc. 2004, 126,
10802.
(8) Panunzi, A.; DeRenzi, A.; Palumbo, R.; Paiaro, G. J. Am. Chem. Soc. 1969,
91, 3879.
(9) Gagne, M. R.; Stern, C. L.; Marks, T. J. J. Am. Chem. Soc. 1992, 114,
275.
(14) Villanueva, L. A.; Abboud, K. A.; Boncella, J. M. Organometallics 1992,
11, 2963.
9
12066
J. AM. CHEM. SOC. 2005, 127, 12066-12073
10.1021/ja052473h CCC: $30.25 © 2005 American Chemical Society

Transfer of Amido Groups from Isolated Rhodium(I)
A R T I C L E S
Scheme 1
Scheme 2
consisted of resonances corresponding to a 2:1 ratio of ethyl
groups of the two types of inequivalent phosphines. A single
signal for an N-H proton was also observed, along with a single
methyl group and a single set of aromatic resonances for a tolyl
group. The 31P NMR spectrum of this product consisted of a
2:1 ratio of a doublet of doublets and a triplet of doublets with
141 and 153 Hz couplings to rhodium.
the desired arylamido complexes. The related {Rh(PEt3)2-
[N(SiMe3)2]} (1b) generated in situ was also a convenient
precursor to the dinuclear arylamido complexes. These com-
plexes are both related to {Rh(PPh3)2[N(SiMe3)2]}, which was
reported many years ago by Lappert.15 Because complex 1a is
more crystalline than 1b, purification is simpler, and this
complex was isolated in 65% yield in pure form. The mono-
nuclearity of these silylamides was demonstrated by X-ray
diffraction of 1a.
Because of the ease of removing the volatile hexamethyl-
disilazine from the reaction mixture, most of the dinuclear
rhodium amido complexes in this work were prepared by the
reaction of arylamines with crude hexamethyldisilazide 1b. In
some cases, we generated 1b in situ from [(PEt3)2Rh(µ-Cl)]2
and LiN(SiMe3)2, and to the solution of this complex we added
the arylamine. In other cases, we isolated 1b, partially purified
it by drying in vacuo to a thick, dark-purple gel, and to a solution
of this material added the arylamine to generate the rhodium
amide. For example, addition of p-toluidine to 1a generated the
dinuclear toluidide complex 2a in 77% isolated yield.
This complex and the analogous monomeric p-anisidide,
o-anisidide, p-trifluoroemthylanilide, and 2,6-diisopropylanilide
complexes 3b-e were most conveniently prepared by a one-
pot sequence. This sequence involved the addition of 6 equiv
of PEt3 to [(COE)2Rh(µ-Cl)]2, which generated 2 equiv of [Rh-
(PEt3)3Cl]. To this complex was added LiN(SiMe3)2 and then
the arylamine. Reaction of the phosphine-ligated rhodium
chloride occurred only after addition of the arylamine; no
reaction was observed between [Rh(PEt3)3Cl] and LiN(SiMe3)2
at room temperature. This procedure afforded the corresponding
rhodium arylamido complexes in 44-68% isolated yields.
We also generated secondary arylamido complexes by this
route. Reactions of [(COE)2Rh(µ-Cl)]2 with 6 equiv of PEt3 and
2 equiv of LiN(SiMe3)2 per dimer, followed by 2 equiv of di-
4-anisylamine, generated the trisphosphine complex 3f in 78%
yield. The same procedure with N-methylaniline afforded
N-methylanilide 3g in 42% yield. The product from addition
of N-methylaniline to 1b without added PEt3 was unstable.
C. Solid-State Structures of the Silylamido and Dimeric
Arylamido Complexes. The ORTEP diagrams of rhodium
amides 1a and 2e are provided in Figure 1. In the solid state,
1a adopted a Y-shaped geometry at rhodium that contrasts with
the T-shaped geometry of most monomeric, three-coordinate
d8 transition-metal complexes,17,18 including that of a recently
isolated three-coordinate Pd(II) diarylamido complex.19 The
difference in geometry may be attributed to both steric and
electronic factors. The steric bulk of the silylamido substituents
favor a Y-shaped geometry, and the preference for Y-shape
geometry of Ni(I) amido complexes was similarly attributed to
the steric bulk of the amido group.20,21 In addition, a Y-shaped
geometry allows for overlap of the nitrogen electron lone pair
with a d orbital in the trigonal plane.22
Alternatively, the arylamido complexes were prepared in one
pot from [(COE)2Rh(µ-Cl)]2. For example, reaction of the
rhodium olefin complex with PEt3, followed by LiN(SiMe3)2
and then p-toluidine, generated the dimeric amido complex 2a
in 80% yield. The p-anisidide (2b), o-anisidide (2c), and
p-trifluoromethylanilide (2d) analogues of toluidide 2a were
generated cleanly by analogous one-pot procedures and isolated
in 55-72% yields. The dinuclear arylamido complexes 2a-d
were mixtures of anti and syn isomers in a ratio of 5:1 to 20:1.
Complexes 2a-d were characterized by spectroscopic meth-
ods and elemental analysis. 1H and 31P NMR spectra of isolated
2a contained resonances from the anti and syn isomers in a 6:1
1
ratio. The H NMR spectrum consisted of two sets of methyl
and methylene groups of equivalent phosphines, along with two
sets of tolyl methyl groups. The 31P NMR spectrum consisted
of two sets of doublets with 176 and 171 Hz couplings to
rhodium. X-ray diffraction data on the PPhEt2 analogue of 2a,
[(PPhEt2)2Rh(µ-NHAr)]2 (Ar ) p-tolyl) (2e), revealed a single
anti isomer. Dissolution of the crystalline material in C6D6 at
room temperature afforded a mixture of anti and syn isomers
in a 1.4:1 ratio. Therefore, the identity of the major isomer of
the isomeric mixture of 2a-d in solution is not definitive.
B. Synthesis of Mononuclear Amido Complexes. Mono-
meric arylamides were prepared by the synthetic route in
Scheme 2. Reaction of dimeric amido complex 2a with PEt3
formed the corresponding monomeric 3a containing three
Dimeric arylamide 2e exists as a single isomer in the solid
state. The Rh2N2 core is planar, as defined by crystallographic
(16) The anilide complex has been generated in solution, but never isolated,
Brunet, J. J.; Commenges, G.; Neibecker, D.; Philippot, K.; Rosenberg, L.
Inorg. Chem. 1994, 33, 6373.
(17) Knobler, C. B.; Marder, T. B.; Mizusawa, E. A.; Teller, R. G.; Long, J.
A.; Behnken, P. E.; Hawthorne, M. F. J. Am. Chem. Soc. 1984, 106, 2990.
(18) Stambuli, J. P.; Incarvito, C. D.; Buhl, M.; Hartwig, J. F. J. Am. Chem.
Soc. 2004, 126, 1184.
(19) Yamashita, M.; Hartwig, J. F. J. Am. Chem. Soc. 2004, 126, 5344.
(20) Bradley, D. C.; Hursthouse, M. B.; Smallwood, R. J.; Welch, A. J. J. Chem.
Soc., Chem. Commun. 1972, 872.
(21) Eller, P. G.; Bradley, D. C.; Hursthouse, M. B.; Meek, D. W. Coord. Chem.
ReV. 1977, 24, 1.
(22) Riehl, J. F.; Jean, Y.; Eisenstein, O.; Pelissier, M. Organometallics 1992,
11, 729.
1
phosphine ligands.16 The H NMR spectrum of this product
(15) Cetinkaya, B.; Lappert, M. F.; Torroni, S. Chem. Commun. 1979, 843.
9
J. AM. CHEM. SOC. VOL. 127, NO. 34, 2005 12067

A R T I C L E S
Zhao et al.
Figure 1. ORTEP diagram of 1a (left) and 2e (right). Only hydrogens bonded to the amido hydrogens are shown for clarity.
symmetry. The rhodium centers adopt slightly distorted square
planar geometries; the sum of the angles at each rhodium center
totaled 363°. The two p-tolyl substituents on the amido moieties
are anti to each other. This anti planar geometry contrasts with
recently reported folded structures of dinuclear Rh(I) arylamido
complexes containing diolefin or nitrile ligands.23 It is difficult
to predict whether the core of a dinuclear d8 amido complex
will be planar or puckered because of a complicated combination
of different structural factors. Nevertheless, it has been suggested
that bulky substituents on the terminal ligands as well as on
the bridging amido groups could promote planar geometries by
minimizing steric repulsion.24,25
2. Reactions of Amido Complexes with Olefins. A. Reac-
tions of Monomeric Arylamido Complexes with Vinylarenes.
Both the monomeric and dimeric primary amido complexes
reacted with olefins to generate products from a combination
of C-N bond formation and elimination of imine. Monomeric
toluidide complex 3a reacted with an excess of styrene in the
absence of added PEt3 to produce the ketimine Ph(Me)CdN(p-
tol) (4a) and the dimeric hydride (PEt3)4Rh2(µ-NHAr)(µ-H) (5a)
in greater than 95% yields after 7 h at 85 °C (eq 1). Reaction
of 3a with styrene in the presence of added PEt3 at 105 °C also
generated ketimine 4a, although more slowly and in yields (44-
61%) that depended on the concentration of added PEt3. In the
presence of added PEt3, a mixture of the dinuclear hydride 5a
and the known monomeric hydrides26 (PEt3)4RhH and (PEt3)3-
RhH was generated in a total yield of roughly 75%.
methods with internal standards. The imine was identified by
comparison of the NMR and GC mass spectra to those of
material prepared independently and fully characterized. The
amido hydride complex 5a was isolated from the reactions and
was fully characterized, including a confirmation of connectivity
by X-ray diffraction.
We investigated the steric and electronic effects on the process
by conducting the reaction with varied arylamido and vinylarene
reagents (eq 2). The monomeric p-anisidide (3b), o-anisidide
(3c), and 4-trifluoromethylanilide (3d) complexes reacted with
styrene to form the analogous imines in 56, >95, and 50%
yields, respectively. The rates of the reactions of anilides 3b
and 3c with styrene were comparable to the rates of the reactions
of the toluidide complex 3a with styrene. The p-toluidide 3a
reacted with 20 equiv of styrene at 95 °C with a half life of
roughly 15 min, while the p-anisidide 3b and o-anisidide 3c
reacted under the same conditions with half-lives of roughly
10 and 20 min, respectively. Thus, the more electron-rich amide
reacted slightly faster, and the mono ortho-substituted amide
reacted slightly slower, but the differences in rates were small.
Reactions of the monomeric 4-trifluoromethylanilide 3d were
much slower than those of the toluidide or anisidide complexes.
The half life for reaction of 3d was about 2 h at 105 °C. This
slower rate is consistent with the trend of faster rates for
reactions of the more electron-rich amides. However, the
difference in rate between reaction of p-toluidide 3a and
4-trifluoromethylanilide 3d is much larger than the difference
in rates between reaction of 3a and p-anisidide 3b. We have
not been able to obtain structural data on 3d or fully purify this
complex as a solid because attempts to crystallize this complex
3d simply led to dissociation of phosphine and formation of
crystalline samples of the dinuclear complex 2d. The structure
of 3d may be important to understand the large difference in
The yields of the rhodium and imine products were deter-
mined by a combination of NMR and gas chromatography (GC)
(23) Tejel, C.; Ciriano, M. A.; Bordonaba, M.; Lopez, J. A.; Lahoz, F. J.; Oro,
L. A. Chem.-Eur. J. 2002, 8, 3128.
(24) Ruiz, J.; Martinez, M. T.; Vicente, C.; Garcia, G.; Lopez, G.; Chaloner,
A.; Hitchcock, P. B. Organometallics 1993, 12, 4321 and the references
therein.
(26) Yoshida, T.; Thorn, D. L.; Okano, T.; Otsuka, S.; Ibers, J. A. J. Am. Ceram.
Soc. 1980, 102, 6451.
(25) Aullon, G.; Lledos, A.; Alvarez, S. Inorg. Chem. 2000, 39, 906.
9
12068 J. AM. CHEM. SOC. VOL. 127, NO. 34, 2005

Transfer of Amido Groups from Isolated Rhodium(I)
A R T I C L E S
reactivity because the NMR spectra of this complex contained
some unexpected features. Four distinct aryl resonances were
observed in the 1H NMR spectrum, six distinct aryl resonances
were observed in the 13C NMR spectrum, and one aromatic
hydrogen resonated upfield (6.3 ppm in C6D6) of the typical
aromatic region. Although the C-H coupling constant for this
upfield hydrogen was similar in value to those of the other
aromatic hydrogens, we suggest, albeit tentatively, that the large
difference in activity is due to stabilization of the ground-state
structure of 3d by a weak agostic C-H or η2 arene interaction.17
The more hindered 2,6-diisopropylanilide complex 3e also
reacted with styrene, and this complex was much more reactive
than the other amido complexes. At 60 °C, this complex
generated the corresponding imine product in 52% yield, along
with free amine in 32% yield. Heating of 3e in the absence of
olefin at 60 °C generated the amine in 26% yield. Thus, the
lower yield from reaction of this amido complex results from
faster competitive decomposition to form free amine by a
pathway that has not yet been identified. Although the yield of
the reaction of this complex with styrene was lower than the
yields of reactions of the less hindered amides, the reaction was
the fastest. The half life of the reaction of this complex with
styrene was about 10 min at 60 °C vs 10-30 min at 95 °C for
reactions of the other arylamido complexes.
C. Reactions of Dinuclear Arylamido Complexes with
Vinylarenes and Alkenes. Dimeric p-toluidide complex 2a
underwent similar reactions with olefins, as shown in eqs 5 and
6. Complex 2a reacted with styrene (30 equiv) in C6D6 at 95
°C to form the ketimine 4a in 72% yield and the amido hydride
dimer 5a in 75% yield after 1.5 h. The p-anisidide and
o-anisidide analogues 2b and 2c also reacted with styrene under
these conditions to form the corresponding ketimines in 80 and
62% yields, respectively. Dimeric trifluoromethylanilide 2d was
less reactive than the more electron-rich arylamides, and required
10 h at 95 °C to consume the amido complex and form the
ketimine (60% yield).
Dimeric 2a also reacted with propene. Reaction with an
excess of propene formed the ketimine Me2CdN(p-tol) (4b)
(90%) and amido hydride 5a (57%) after 16 h at 95 °C (Scheme
3). The rhodium allyl complex [(PEt3)2Rh(η3-C3H5)] was also
formed from the reaction with propene in roughly 30% yield.
The identity of the allyl complex was confirmed by independent
synthesis from [(PEt3)2Rh(µ-Cl)]2 and allyl Grignard.
The effect of the electronic properties of the vinylarenes on
the amido transfer process was evaluated by conducting
reactions with 4-vinylanisole and both 4- and 3-trifluorometh-
ylstyrene (eq 3). The monomeric toluidide 3a reacted with 20
equiv of 4-vinylanisole, 4-trifluoromethylstyrene, and 3-trifluo-
romethylstyrene to afford the corresponding imines in 64, 51,
and 56% yields. The reaction with 4-trifluoromethylstyrene was
faster than the reaction with 4-vinylanisole but by a factor of
only about 2. Moreover the reaction of 3a with a 1:1 mixture
of 4-trifluoromethylstyrene and 4-vinylanisole favored formation
of the product from reaction of trifluoromethylstyrene by a factor
of about 2. Thus, the reaction is slightly faster with the more
electron-poor vinylarenes, but the electronic effects are not large.
D. Reactions of Monomeric Diarylamido and N-Alkylary-
lamido Complexes with Vinylarenes. The amido complexes
generated from secondary amines were less reactive toward
olefins. Nevertheless, the diarylamide (PEt3)3Rh(NAr2) (Ar )
p-anisyl, 3f) did react with styrene to transfer the amido group
and form enamine (eq 7). Free diarylamine HN(p-anisyl)2 was
the major product (ca 80% yield), but reaction with 200 equiv
of styrene led to the formation of enamine (Ph)[N(p-anisyl)2]Cd
CH2 in 10-15% yield. In contrast, heating of the N-methyla-
nilide 3g with vinylarenes only led to â-hydrogen elimination
without participation of the olefin. This reaction generated
the amidine PhNdC(H)N(Me)Ph as the organic product
(eq 8).27
B. Reactions of Monomeric Arylamido Complexes with
Alkenes. Monomeric toluidide complex 3a also reacted with
alkenes (eq 4). The reaction of 3a with 15 equiv of 1-hexene
generated a mixture of syn and anti imines (Me)(n-Bu)CdN-
(p-tol) (4c) in a total yield of 64%, along with rhodium hydride
5a in 90% yield. Reaction with 30 equiv of propene generated
the imine (Me)2CdN(p-tol) (4b) in 65% yield and 5a in 94%
yield. Complex 3a did not react with internal alkenes, such as
cyclohexene, and did not react with unstrained 1,1-disubstituted
alkenes, such as isobutylene.
3. Studies of the Mechanism of the Amido Transfer to
Olefins. A. Potential Formation of the Imine by a Rhodium-
Catalyzed Process. Small amounts of free amine are generated
(27) Hartwig, J. F. J. Am. Chem. Soc. 1996, 118, 7010.
9
J. AM. CHEM. SOC. VOL. 127, NO. 34, 2005 12069

A R T I C L E S
Scheme 3
Zhao et al.
olefin31,32 and subsequent reductive elimination from metalla-
cycle 13 to form the aminoalkyl intermediate.
-d[3a]/dt ) kobs[3a]
(9)
For Path A
kobsd ) k1k2[olefin]/(k-1[PEt3] + k2)
For Path B
(10)
kobsd ) k1k3[olefin][H2NAr]/(k3[H2NAr] + k-1[PEt3]) (11)
For Path C
-d[3a]/dt ) k1k4[3a]2[olefin]/(k-1[PEt3] + k4[3a]) (12)
For Path D,
if R-elimination is rate-limiting, then
kobsd ) k5
(13)
(14)
if R-elimination is fully reversible, then
kobsd ) K5k6k7[olefin]/(k-6[PEt3] + k7)
The rate equation for Path A is first-order in 3a and olefin,
inverse first-order in added PEt3, and zero-order in added
p-toluidine. The rate equation for Path B involving rate-limiting
attack of aniline on the coordinated olefin is first-order in 3a
and olefin, inverse first-order in added PEt3, and first-order in
p-toluidine. The same reaction orders would be predicted for
Path B with reversible attack of aniline, followed by rate-limiting
proton transfer, or reversible attack of aniline, followed by
reversible proton transfer and rate-limiting dissociation of amine.
The rate equation for Path C predicts that the reaction would
be second-order in 3a because the transition state of the
irreversible step involves two rhodium centers. This rate
equation also predicts that the reaction would be first-order in
olefin, inverse first-order in added phosphine, and zero-order
in added amine. The rate equation for Path D is zero-order in
all reagents if R-elimination to generate the imido species is
rate limiting, but the rate equation is first-order in olefin, inverse
order in added phosphine, and zero-order in amine if the
generation of the imido complex is fully reversible prior to
reaction with the olefin.
The rate equations for all paths are similar if the aminoalkyl
intermediate is generated reversibly and â-hydrogen elimination
is rate-limiting. For this reason, we tested whether the â-hy-
drogen elimination step was rate-limiting by measuring the
isotope effect for reactions of styrene and styrene-d8. The
reactions of styrene and styrene-d8 at 105 °C with 10 equiv of
added PEt3 in C7D8 solvent occurred with rate constants that
were within 10% of each other. Although isotope effects on
â-hydrogen elimination are often small,33-37 the absence of any
detectable isotope effect from these reactions argues strongly
against â-hydrogen elimination as the final irreversible step.
from the reactions of the arylamido complexes with olefins. This
observation and the report by Brunet of slow catalytic formation
of imines from olefins and amines catalyzed by related rhodium
complexes28,29 led us to consider that free p-toluidine generated
during the reactions of 2a or 3a could undergo a reaction with
the olefin catalyzed by the rhodium product. However, reaction
of p-toluidine with styrene in the presence of 5 mol % 5a
generated less than 5% ketimine 4a after 18 h at 95 °C. Thus,
further discussion of the mechanism focuses on stoichiometric
reactions that transfer the amido group to the olefin.
B. Potential Pathways for the Transfer of the Amido
Group. Several potential pathways for the reaction of mono-
meric amide 3a with olefins that involve dissociation of
phosphine are shown in Scheme 3. Pathways involving dis-
sociation of phosphine to form an intermediate olefin complex
6 are considered because the reaction is inhibited by added
phosphine (vide infra). The rate equations that correspond to
Paths A-D are shown in eqs 9-14.
Path A involves insertion of the coordinated olefin on 6 into
the amide group to form aminoalkyl intermediate 7, followed
by â-hydrogen elimination. Tautomerization of the resulting
enamine 8 to the imine and reaction of the rhodium hydride 9
with the starting rhodium amide would generate the final
products. Path B involves external attack of the amine on a
coordinated olefin to form an aminoalkyl intermediate 10 that
would undergo proton transfer and dissociation of the resulting
coordinated amine to generate the same three-coordinate ami-
noalkyl intermediate 7 as is formed by Path A. Path C involves
external attack of the rhodium amide on the coordinated olefin
to generate a dinuclear intermediate 11 that could dissociate to
generate the same three-coordinate aminoalkyl intermediate 7.30
Path D involves R-elimination to form a rhodium imido hydride
intermediate 12 that could undergo a [2 + 2] reaction with the
(31) Polse, J. L.; Andersen, R. A.; Bergman, R. G. J. Am. Chem. Soc. 1998,
120, 13405.
(32) Glueck, D. S. W., Jr.; Hollander, F. J.; Bergman, R. G. J. Am. Chem. Soc.
1991, 113, 2041.
(28) Brunet and co-workers observed a mixture of amine and imine products,
along with products from hydroarylation of olefins.
(29) Brunet, J.-J.; Neibecker, D.; Philippot, K. Tetrahedron Lett. 1993, 34, 3877.
(30) Ritter, J. C. M.; Bergman, R. G. J. Am. Chem. Soc. 1997, 119, 2580.
(33) Evans, J.; Schwartz, J.; Urquhart, P. W. J. Organomet. Chem. 1974, 81,
C37.
(34) Reger, D. L.; Culbertson, E. C. J. Am. Chem. Soc. 1976, 98, 2789.
(35) Ozawa, F.; Ito, T.; Yamamoto, Y. J. Am. Chem. Soc. 1980, 102, 6457.
9
12070 J. AM. CHEM. SOC. VOL. 127, NO. 34, 2005

Transfer of Amido Groups from Isolated Rhodium(I)
A R T I C L E S
complex could be an intermediate in the conversion of 3a and
olefins to the rhodium product 5a and the imine.
However, we showed by independent synthesis that this
intermediate was the bis PEt3-ligated rhodium enamido complex
of the anion of the product imine, (PEt3)2Rh[(p-tol)NC(Ph)d
CH2]. This material was prepared independently by the reaction
of the potassium enamide of the N-tolylketimine with [(PEt3)2-
Rh(µ-Cl)]2 generated in situ and displayed the same 1H and 31
P
NMR signals of the intermediate that accumulated in small
amounts at the beginning of the reaction. The enamido complex
was fully characterized by NMR spectroscopic methods, and a
more comprehensive study of the synthesis, structure, and
reactivity of this complex will be reported separately.
We presume this complex was formed by proton transfer
between the ketimine product, or initial enamine product, and
the starting rhodium amide. Indeed, reaction of dimeric toluidide
2a and the ketimine generated some of the enamido complex.
However, this proton transfer occurred to only partial conver-
sion, and the equilibrium between the combination of p-toluidide
2a and free imine and the combination of p-toluidine and this
enamido complex lies toward the combination of p-toluidide
2a and free imine. Addition of p-toluidine to the isolated
enamido complex generated about 75% free 4a and 80% of 2a.
Because of the thermodynamics of this protonolysis, the
enamido complex is not present during reactions conducted with
added arylamine or after some of the amine has been generated
from hydrolysis by traces of water in the system. The monomeric
arylamido complexes with three PEt3 ligands are resistant to
formation of the enamido complex because the stability of the
enamido complex is largely derived from a polydentate coor-
dination mode. The trisphosphine amide 3a failed to react with
ketimine 4a even at 85 °C in C6D6.
E. Kinetic Orders of the Reagents and Additive. Data
revealing the order of the reaction in olefin and added phosphine
are shown in Figures 4 and 5. In the presence or absence of
added toluidine, the decay of 3a was first-order in olefin and
inverse first-order in the concentration of added PEt3. These
data suggest that reversible dissociation of PEt3 occurs prior to
the rate-determining step and that reaction with olefin occurs
prior to or as part of the rate-determining step. The rate constants
for the decay of the reactant or the appearance of the rhodium
product were unaffected by added amine. Thus, any involvement
of the amine would likely occur after the rate-determining step.
These data are inconsistent with reaction by Paths B and C.
The lack of an order in amine is inconsistent with reaction by
the Path B involving attack of free amine on the coordinated
olefin. The first-order rate behavior in 3a is inconsistent with
reaction by path C involving external attack of the rhodium
amide on a coordinated olefin. Instead, our kinetic data are
consistent with reaction by Paths A and D, which involve
insertion of olefin or the combination of reversible R-elimination
and rate-limiting [2 + 2] addition.
Figure 2. Representative plot of the decay of rhodium complex 3a (0.058
M) with styrene (5.2 M) in the presence of PEt3 (0.29 M) and p-toluidine
(0.006 M) in C7D8 at 105 °C. The curve for the consumption of 3a depicts
the results of an unweighted least-squares fit to y ) a exp(-bx) + c (a )
0.61 ( 0.01, b ) 0.00014 ( 0.00001, c ) 0.06 ( 0.01). The curve for the
accumulation of 4a depicts the results of an unweighted least-squares fit to
y ) -a exp(-bx) + c (a ) 0.354 ( 0.003, b ) 0.000120 ( 0.000003, c
) 0.360 ( 0.003).
Thus, versions of pathways A-D with rate-limiting â-hydrogen
elimination were not considered further.
C. Conditions of the Kinetic Experiments. To begin to
distinguish between the possible mechanisms for the transfer
of the amido group, we determined the order of the reaction in
rhodium, olefin, added ligand, and added amine. Because the
kinetic expressions are more complex starting from the dimeric
species, because dimeric 2a could undergo first-order decay or
half-order decay, and because first and half-order rate behaviors
are difficult to distinguish, our kinetic studies focused on
reactions of monomeric 3a. Reactions of 3a were conducted
with a 0.029 M concentration of 3a, 0.4-2.6 M concentrations
of styrene, and 0.14-1.43 M concentrations of added PEt3.
Because p-toluidine is generated in 3-8% yield as a side
product, and this byproduct could affect the rate of the reaction,
we also conducted kinetic studies with 3.0 or 29 mM added
p-toluidine. Under all conditions, rhodium reactant 3a decayed
exponentially (Figure 2). These data show that the reactions
are first-order in 3a.
D. A Side Note Concerning the Curve of the Appearance
of Imine Product. The curve of the appearance of imine and
rhodium products from reaction of dimeric arylamides 2a-d
depended on the concentration of free amine and added
phosphine. In the absence of amine and phosphine, the appear-
ance of imine from 2a and styrene was sigmoidal (Figure 3).
1
Moreover, monitoring of the reaction by H and 31P NMR
spectroscopy revealed the accumulation of a rhodium complex
in small quantities that formed and decayed in concert with the
initial lag in the formation of the imine. This intermediate did
not accumulate during reactions of dimeric 2a-d conducted
with added phosphine or with any reaction of mononuclear
complexes 3a-d. Thus, this observation does not affect our
kinetic data with monomer 3a, but one could imagine that this
Prior literature and additional data argue against reaction by
Path D. The spontaneous generation of a low-valent, late-metal
imido complex from an amide group has little precedent,
although examples of the generation of late-metal carbene
intermediates from alkyl groups have been reported.38-42
(36) Bullock, R. M. In Transition Metal Hydrides; Dedieu, A., Ed.; VCH: New
York, 1992; p 263.
(37) Bullock, R. M.; Bender, B. R. In Encyclopedia of Catalysis; Horvath, I.
T., Ed.; Wiley: New York, 2003.
(38) Boutry, O.; Gutierrez, E.; Monge, A.; Nicasio, M. C.; Perez, P. J.; Carmona,
E. J. Am. Chem. Soc. 1992, 114, 7288.
(39) Luecke, H. F.; Arndtsen, B. A.; Burger, P.; Bergman, R. G. J. Am. Chem.
Soc. 1996, 118, 2517.
9
J. AM. CHEM. SOC. VOL. 127, NO. 34, 2005 12071

A R T I C L E S
Zhao et al.
Figure 3. (Left) Plot of the consumption of amide 2a and the formation of ketimine 4a over time. (Right) Plot of the consumption of amide complex 2a
and the formation of ketimine 4a over time in the presence of added p-toluidine (5 equiv). Reactions were carried out using 0.029 M 2a and 0.87 M styrene
1
in C6D6 at 85 °C and monitored by H NMR.
Figure 5. Plot of 1/kobsd vs [PEt3] for the reaction of rhodium complex 3a
(0.029 M) with styrene (2.57 M) in the presence of added PEt3 and
Figure 4. Plot of kobsd vs [styrene] for the reaction of rhodium complex
p-toluidine (0.003 M) in C7D8 at 105 °C. The curve depicts the results of
3a (0.029 M) with styrene in the presence of added PEt3 (0.29 M) and
an unweighted least-squares fit to y ) ax + b (a ) 4.5 ( 0.3 × 103, b )
p-toluidine (0.003 M) in C7D8 at 105 °C. The curve depicts the results of
9 ( 3 × 102).
an unweighted least-squares fit to y ) ax (a ) 1.50 ( 0.03 × 10-2).
Nevertheless, we sought experimental data that could distin-
guish between Paths A and D. One could imagine that deuterium
labeling could distinguish between these pathways because the
two mechanisms would place the starting N-H hydrogen in a
different position in the initially formed enamine product.47
However, tautomerization of the enamine in this case scrambles
the position of the label and prevents gaining information from
this experiment.
Instead, measurements of isotope effects and reactivity of
secondary amido complexes could distinguish between these
mechanisms. The order in olefin requires that the R-elimination
be reversible. If this step occurred, then one would expect to
observe an equilibrium isotope effect when the reaction is
conducted with toluidide complex 3a-d1 with deuterium in the
N-H position. If reductive elimination of the resulting metal-
lacycle were rate-limiting, then one would also expect to observe
an isotope effect, although this isotope effect could be normal
or inverse.48-50 In contrast to these predictions, the rate constants
Further, few late-metal imido complexes react with alkenes, and
products from cycloadditions of imido complexes have been
limited to those formed from olefins containing activated C-C
multiple bonds in R,â-unsaturated carbonyl compounds.32 Most
of the reactions of olefins with imido complexes generate
aziridine,43,44 and it is not clear if the imido group is transferred
by an outer-sphere process of if a concerted [2 + 2] reaction to
form an azametallacyclobutane occurs. Thus, the first and the
second steps of pathway D, which are akin to the Green
mechanism for formal insertion of olefins into metal alkyl
groups,45,46 have little precedent for the reaction of a low-valent,
late-metal amido complex.
(40) Ho, V. M.; Watson, L. A.; Huffman, J. C.; Caulton, K. G. New J. Chem.
2003, 27, 1446.
(41) Paneque, M.; Poveda, M. L.; Santos, L. L.; Carmona, E.; Lledos, A.; Ujaque,
G.; Mereiter, K. Angew. Chem., Int. Ed. 2004, 43, 3708.
(42) Clot, E. C., J.; Lee, D.; Sung, S.-Y.; Appelhans, L.; Faller, J. W.; Crabtree,
R. H.; Eisenstein, O. J. Am. Chem. Soc. 2004, 126, 8795.
(43) Waterman, R.; Hillhouse, G. L. J. Am. Chem. Soc. 2003, 125, 13350.
(44) Liang, J. L.; Huang, J. S.; Yu, X. Q.; Zhu, N. Y.; Che, C. M. Chem.-Eur.
J. 2002, 8, 1563.
(45) Green, M. L. H. Pure Appl. Chem. 1978, 50, 27.
(46) Ivin, K. J.; Rooney, J. J.; Stewart, C. D.; Green, M. L. H.; Mahtab, R. J.
Chem. Soc., Chem. Commun. 1978, 604.
(47) This type of labeling experiment ruled out the Green mechanism for olefin
insertion into alkyls in the context of olefin polymerization: Grubbs, R.
H.; Coates, G. W. Acc. Chem. Res. 1996, 29, 85.
9
12072 J. AM. CHEM. SOC. VOL. 127, NO. 34, 2005

Transfer of Amido Groups from Isolated Rhodium(I)
A R T I C L E S
for reaction of the N-D complex with added D2N-p-Tol were
indistinguishable from the rate constants for the reaction of the
N-H complex with added toluidine (3.4(0.2 for 3a-d1 vs
3.0(0.2 × 10-4 s-1 for 3a).
cyclization of lanthanide amido complexes generated in situ
from aminoalkenes. The paucity of clear examples of insertions
of olefins into amido complexes stands in contrast to the many
examples of the transfer of alkyl and aryl groups to olefins,
usually by insertion mechanisms.
Formation of the analogous products from reactions of
secondary amido complexes, which lack an N-H proton, more
clearly distinguishes between Paths A and D. Path D cannot
occur with a secondary amido complex. Although the N-
methylanilido complex 3g, which possesses the least steric and
electronic perturbation from the arylamido complexes, undergoes
â-hydrogen elimination without any reaction with the olefin,
the dianisylamido complex 3f did transfer the amido group to
olefins. The transfer of this diarylamido group from rhodium
to stryene was much slower and occurred in lower yield than
the transfer of the arylamido groups, presumably because the
diarylamido complexes are much more hindered and are less
electron-donating than the monoarylamido complexes. Never-
theless, we did observe reaction of the dianisylamido complex
3f with an excess of styrene to form measurable amounts of
enamine. The formation of enamine from this reaction provides
strong experimental evidence against Path D.
Thus, our kinetic data rules out Paths B and C, and several
factors argue against reaction by Path D. Both the observation
of some C-N bond formation with the secondary amide and
the lack of an isotope effect of the N-H proton from reactions
of the primary arylamides argue against reaction by Path D.
Thus, Path A, which occurs by olefin insertion into the amide,
remains the only path in Scheme 4 that is consistent with all of
our experimental data.
Our current study demonstrates that amido groups can be
transferred from simple late-transition-metal amido complexes
to vinylarenes and alkenes to form imines. Although we have
not directly observed the immediate product from an insertion
process, detailed mechanistic data are consistent with a pathway
involving insertion of the olefin into the rhodium-amido linkage,
followed by â-hydrogen elimination and tautomerization of the
resulting enamine to the ketimine. The mechanistic data are
inconsistent with several plausible alternative reaction pathways.
At the current time, we do not have a firm explanation for
the origin of the regioselectivity of the C-N bond formation.
The regioselectivity of the reaction of the rhodium amides with
propene and 1-hexene is the same as the typical regioselectivity
of insertion of alkenes into metal-alkyl bonds,52,53 but the
regioselectivity of the reaction of the rhodium amides with
vinylarenes contrasts with the typical regioselectivity of insertion
of vinylarenes into late-transition metal-alkyl54 and -aryl
bonds. The reactions of late metal aryl complexes typically form
stilbenes, rather than 1,1-disubstituted olefins, after â-hydrogen
elimination.55,56 Thus, studies to understand the origins of this
selectivity by Path A, or another path that could be revealed by
further data, and efforts to observe directly the initial products
from reaction of the amido complex with olefins will be the
focus of future studies.
Acknowledgment. Financial support for this work was
provided by the Department of Energy, Office of Basic Energy
Sciences. We thank Boehringer-Ingelheim for an unrestricted
gift.
4. Conclusions
The catalytic amination of olefins has been a long-term goal,
and several mechanisms for hydroamination reactions have been
deduced recently. Although oxidative addition of amine51 and
transfer of the resulting amido group to an olefin can be
envisioned to be part of a catalytic cycle for hydroamination,
the intermolecular transfer of the amido group from an isolated
amido complex to an unactivated olefin had not been observed
previously. The closest examples involve insertion of the
strained norbornene into an iridium amido complex and the
Supporting Information Available: Experimental details,
kinetic plots, and full structural characterization of 1a and 2e
(CIF and PDF). This material is available free of charge via
the Internet at http://pubs.acs.org.
JA052473H
(52) Svejda, S.; Johnson, L.; Brookhart, M. J. Am. Chem. Soc. 1999, 121, 10634.
(53) Tempel, D. J.; Johnson, L. K.; Huff, R. L.; White, P. S.; Brookhart, M. J.
Am. Chem. Soc. 2000, 122, 6686.
(54) Rix, F. C.; Brookhart, M.; White, P. S. J. Am. Chem. Soc. 1996, 118, 2436.
(55) Beletskaya, I., P.; Cheprakov, A. V. Chem. ReV. 2000, 100, 3009.
(56) Demeijere, A.; Meyer, F. E. Angew. Chem., Int. Ed. Engl. 1995, 33, 2379.
(48) Janak, K. E.; Parkin, G. J. Am. Chem. Soc. 2003, 125, 13219.
(49) Janak, K. E.; Parkin, G. J. Am. Chem. Soc. 2003, 125, 6889.
(50) Bullock, R. M.; Bender, B. R. In Encyclopedia of Catalysis; Istva´n H.,
Ed.; Wiley and Sons: New York, 2002.
(51) Zhao, J.; Goldman, A. S.; Hartwig, J. F. Science 2005, 307, 1080.
9
J. AM. CHEM. SOC. VOL. 127, NO. 34, 2005 12073