Rhodium phosphine-π-Arene intermediates in the hydroamination of alkenes

Article abstract of DOI:10.1021/ja1057949

A detailed mechanistic study of the intramolecular hydroamination of alkenes with amines catalyzed by rhodium complexes of a biaryldialkylphosphine is reported. The active catalyst is shown to contain the phosphine ligand bound in a κ¹, η⁶ form in which the arene is π-bound to rhodium. Addition of deuterated amine to an internal olefin showed that the reaction occurs by trans addition of the N-H bond across the C=C bond, and this stereochemistry implies that the reaction occurs by nucleophilic attack of the amine on a coordinated alkene. Indeed, the cationic rhodium fragment binds the alkene over the secondary amine, and the olefin complex was shown to be the catalyst resting state. The reaction was zero-order in substrate, when the concentration of olefin was high, and a primary isotope effect was observed. The primary isotope effect, in combination with the observation of the alkene complex as the resting state, implies that nucleophilic attack of the amine on the alkene is reversible and is followed by turnover-limiting protonation. This mechanism constitutes an unusual pathway for rhodium-catalyzed additions to alkenes and is more closely related to the mechanism for palladium-catalyzed addition of amide N-H bonds to alkenes.

Full text of DOI:10.1021/ja1057949

ARTICLE
pubs.acs.org/JACS
Rhodium Phosphine-π-Arene Intermediates in the Hydroamination
of Alkenes
Zhijian Liu, Hideaki Yamamichi, Sherzod T. Madrahimov, and John F. Hartwig*
Department of Chemistry, University of Illinois, 600 South Mathews Avenue, Urbana, Illinois 61801-3602, United States
S Supporting Information
b
ABSTRACT: A detailed mechanistic study of the intramolecular
hydroamination of alkenes with amines catalyzed by rhodium
complexes of a biaryldialkylphosphine is reported. The active
catalyst is shown to contain the phosphine ligand bound in a κ1,
η6 form in which the arene is π-bound to rhodium. Addition of
deuterated amine to an internal olefin showed that the reaction
occurs by trans addition of the N-H bond across the CdC bond,
and this stereochemistry implies that the reaction occurs by
nucleophilic attack of the amine on a coordinated alkene. Indeed,
the cationic rhodium fragment binds the alkene over the secondary amine, and the olefin complex was shown to be the catalyst resting
state. The reaction was zero-order in substrate, when the concentration of olefin was high, and a primary isotope effect was observed. The
primary isotope effect, in combination with the observation of the alkene complex as the resting state, implies that nucleophilic attack of
the amine on the alkene is reversible and is followed by turnover-limiting protonation. This mechanism constitutes an unusual pathway
for rhodium-catalyzed additions to alkenes and is more closely related to the mechanism for palladium-catalyzed addition of amide N-H
bonds to alkenes.
’ INTRODUCTION
amines to alkenes.30 The second catalyst we reported contained an
aminophosphine analog of Xantphos and was particularly active for
cyclization of primary aminoalkenes.28
The hydroamination of alkenes has been studied for decades, but
only recently have complexes containing late transition metals been
identified that catalyze the additions of amines to unactivated
olefins.1-4 After catalysts for the addition of aniline to strained
alkenes5,6 and acrylamide7,8 were identified, palladium com-
plexes that catalyze the additions of amines to unstrained
vinylarenes,9,10 and complexes containing rhodium that add
amines to vinylarenes with the opposite regioselectivity11 in
high yields12 were discovered. Complexes of platinum were
then shown to catalyze the intramolecular additions of amines
to alkenes,13 and a simple iridium system has been shown to
cyclize aminoalkenes when they are conformationally biased for
intramolecular reactions.14,15 In parallel, a series of systems that
catalyze the intramolecular additions of amides and carbamates
to alkenes were uncovered.16-26
We recently reported the discovery of two rhodium catalysts that
are particularly active for the additions of alkylamines to alkenes27,28
and a recent study built upon the first of these systems to develop
enantioselective versions of this reaction.29 One of the catalysts we
reported was based on the combination of [Rh(COD)2]BF4 and a
biaryldialkylphosphine (L) first developed for palladium-catalyzed
cross coupling.27 This combination of metal and ligand was unique
forlate-metalsystemsandremainsa rare system because it catalyzes
the cyclization of aminoalkenes lacking substituents on the chain
linking the two functional groups that favor cyclization. This
combination also catalyzed the addition of amines to internal
alkenes and catalyzed additions of both primary and secondary
Although these studies revealed an unusually broad scope for
alkene hydroamination, the structure of the complex responsible for
the reactivity of the catalyst containing the biaryldialkylphosphine
and the basic steps of the reaction were not identified. Several types
of mechanisms have been shown to lead to hydroamination pro-
cesses catalyzed by late transition metals. One of these mechanisms
includes oxidative addition of an N-H bond, migratory insertion
of the alkene into the metal-nitrogen bond, and reductive
elimination to form a C-H bond.5 A second mechanism, more
commonly followed by lanthanide3,31,32 and group IV metals,32-36
involves migratory insertion of the alkene into the metal-amide,
followed by proton transfer.37 A third mechanism is thought to
occur by nucleophilic attack of an amine on a coordinated alkene,
followed by proton transfer to release the amine product.13,38
Finally, reactions of amines with alkynes catalyzed by some group
IV complexes have been shown to occur by [2 þ 2] additions of
alkenes to metal-imido intermediates,39,40 and reactions of some
aminoalkenes catalyzed by zirconium(IV) complexes have been
proposed to occur by an analogous mechanism.41,42
Rhodium catalysts are well-known to add H-X bonds to
alkenes by a pathway involving oxidative addition, migratory
insertion, and reductive elimination.43 Hydrogen, syngas,
Received: July 6, 2010
Published: February 10, 2011
r
2011 American Chemical Society
2772
dx.doi.org/10.1021/ja1057949 J. Am. Chem. Soc. 2011, 133, 2772–2782
|

Journal of the American Chemical Society
ARTICLE
silanes, boranes, and stannanes all react with alkenes by such a
pathway, and, therefore, these additions occur in a syn fashion
across the alkene.43-45 In the absence of mechanistic information
on the rhodium system that we identified, it was unclear if the
hydroamination process would occur by a mechanism akin to the
addition of these other reagents, or if the hydroamination would
occur by a pathway more akin to the addition of N-H bonds by
other catalysts.
Mechanistic studies have implied that palladium-catalyzed addi-
tionsof the N-H bonds of amides and carbamates to alkenes26 and
platinum-catalyzed additions of secondary amines to alkenes13
occur by nucleophilic attack on an alkene, followed by proton
transfer. Analogous steps have been shown to occur during the
addition of amines and carbamates to allenes catalyzed by com-
plexes of gold.46,47
Such a path involving nucleophilic attack on a coordinated
alkene would be unusual for a complex of rhodium for several
reasons. First, nucleophilic attack on an alkene bound to rhodium is
not a common reaction.48,49 Second, the only prior mechanistic
information on alkene amination catalyzed by rhodium, oxidative
aminations of vinylarenes with dialkylamines catalyzed by [Rh-
(COD)2]BF4 and PPh3, led to ambiguous conclusions about
whether the process occurred by nucleophilic attack on a coordi-
nated vinylarenes or oxidative addition of the amine, insertion of
the olefin, and β-hydrogen elimination.11 Although work from the
authors’ group on a different rhodium-catalyzed process implied
that nucleophilic attack on a vinylarene occurred as the first step,
nucleophilic attack on an alkene complex is different; the stable η3-
benzyl-type complex formed by attack on a coordinated
vinylarene12 is not formed from attack on a coordinated alkene.
Figure 1. ORTEP diagram of [Rh(L)(C2H4)]BF4 (1). Ellipsoids are
shown at 50% probability. All hydrogen atoms are omitted for clarity.
the phosphine in CH2Cl2. This reaction is depicted in eq 2 and
formed a single complex 1, which contained one coordinated
alkene and one L. NMR spectroscopy and X-ray diffraction
showed that the ligand L was bound in the tethered η6, κ1 mode
shown in this equation. A series of complexes of Cp and arene
ligands tethered to phosphorus donors have been prepared, but
few contain just a two-carbon linker. Such a binding mode was
observed previously by Faller et al. between ruthenium and this
biaryldialkylphosphine when studying complexes for reactions
catalyzed by Lewis acids51 and by Werner et al. with a rhodium
complex of a ligand containing an aliphatic bridge between the
arene and the phosphine.52
To begin to reveal the factors controlling the activity of this
catalyst and the selectivity for products from hydroamination vs
olefin isomerization and oxidative amination,12 we have conducted
mechanistic studies of the process in eq 1. We report the structure
of the active form of the rhodium catalyst, the relative binding
affinity of this species toward alkenes and secondary amines, the
resting state of the catalyst, the turnover-limiting step of the
catalytic cycle, and the stereochemistry of the addition process.
These data support a mechanism involving trans addition and allow
us to propose an experimentally supported hypothesis to explain
the selectivity for formation of products from hydroamination over
oxidative amination.
An ORTEP diagram of 1 is shown in Figure 1. This structure
clearly contains the biaryldialkylphosphine bound in an η6, κ1
mode. The dimethylamino group is nearly coplanar with the arene
P
ring, and the nitrogen is almost perfectly planar ( CNC = 359.8°),
suggesting that the dimethylamino group serves as a strong electron
donor to the coordinated arene ring. The arene ring is not strictly
planar. The C6 carbon atom lies out of the arene plane by 0.11 Å,
and the angle consisting of the centroid of the arene and the two
carbon atoms of the C-C bond between the coordinated and
bridging arene rings is 170.6°. Thus, the arene-phosphine chelate
contains observable strain. The Rh-P distance (2.2383(8) Å) is
0.06 Å shorter than the average Rh-P distance involving phos-
phine ligands for the compounds in the CCD and similar to the
2.251(2) Å Rh-P distance in a closely related rhodium complex
containing an η6, κ1-arene-phosphine ligand linked by a three-
carbon aliphatic tether.52
’ RESULTS AND DISCUSSION
1. Identification of the Metal-Ligand Core of the Active
Catalyst. The catalyst we initially reported was generated from
the biaryldialkylphosphine (L) in eq 250 and [Rh(COD)2]BF4,
but the structure of the species in the catalytic system was
unknown. Mixing these two materials (in the dioxane solvent
of the initial catalytic system) led to only a small amount of
product from binding of L. To favor coordination of the ligand,
we studied the reaction of the phosphine with the species formed
from [Rh(ethylene)2Cl]2 and AgBF4. To ensure solubility, we
conducted the reaction of this species, as generated in THF, with
Finally, the metric parameters of the bound ethylene are only
slightly distorted from those of free ethylene and close to those of
classic Pd(II) or Pt(II) species. The C-C distance of 1.380(6) Å in
1 is similar to the 1.37 Å distance in Zeise’ salt,53 shorter than the
2773
dx.doi.org/10.1021/ja1057949 |J. Am. Chem. Soc. 2011, 133, 2772–2782

Journal of the American Chemical Society
ARTICLE
Scheme 1
1.43 Å distance in [Pt(PPh3)2(C2H4)],54 similar to the 1.38
and 1.39 Å distances in [Pd(COD)Cl2],55 similar to the 1.35 Å
distance in a cationic Pd-ethylene complex containing a P,O
ligand,56 slightly shorter than the 1.40(3) Å distance in the related
Rh-ethylene complex,52 and slightly longer than the 1.31-1.34 Å
distance in a typical free alkene.57 The Rh-C distances (2.134(3)
Å for both) are similar to those of known rhodium-ethylene
complexes (2.10(4), 2.15(4), 2.162(17), and 2.172(17) Å for a
related half-sandwich-Rh-ethylene complex).52 These data indicate
that the structure of the complex is more akin to olefin complexes,
such as Zeise’ salt, that react with nucleophiles than to electron-rich
complexes, such as Pt(PPh3)2(C2H4), that react with electrophiles.
Solution NMR spectra indicate that the η6, κ1 mode is also
Figure 2. ORTEP diagram of the cationic portion of [Rh(L)-
(NCMe)2]BF4 (2). Ellipsoids are shown at 50% probability. All hydro-
gen atoms are omitted for clarity.
2.096(3) Å) for the nitrile trans to the phosphine is longer than the
Rh-N distance (2.028(3) and 2.022(3) Å) for the nitrile located
trans to the η2-arene ligand, implying that the phosphine has a
larger trans influence than the η2-coordinated π system of the
arene. The 1H NMR signal for the hydrogen on the bound ortho
carbon was observed at 5.42 ppm, and the 13C NMR signals for the
two bound carbons were observed at 88.5 and 89.1 ppm with larger
JRh-C values (9 and 12 Hz, respectively) than those in complex 1
(2-3 Hz).
Treatment of this acetonitrile complex with the secondary
aminoalkene 3a led to coordination of the olefinic portion of 3a,
not the amino group, to form complex 4a as a 1.6:1 ratio of two
diastereomers. This binding mode of the reactant was clearly
revealed by upfield signals of the bound alkene at 2.54-3.05
ppm in the 1H NMR spectrum and upfield signals at 43-46 and
63-66 ppm with nearly equivalent JRh-C values (13 and 11 Hz,
respectively) in the 13C NMR spectrum of 4a. This binding of an
alkene over an amine is unusual for a cationic complex.49 The η6
binding mode of the arene is evidenced by the signals for the
coordinated ring in the 1H NMR spectrum upfield of the aromatic
region and signals for the bound C-H in the 13C{1H} NMR
spectrum between 90 and 110 ppm.
Because the combination of [Rh(COD)2]BF4 and L also
catalyzes the cyclization of primary aminoalkenes and this class of
substrate generally has been less reactive toward late-metal catalysts
for hydroamination, we tested the binding of primary amine 3b to
ethylene complex 1 by addition of the aminoalkene to the ethylene
complex, as shown in Scheme 2. In this case, addition of the primary
aminoalkene 3a occurred without the need to generate the nitrile
complex as an intermediate. Direct reaction of 1-SbF6 with 3b
formed 4b in high yield.
The structure of the primary aminoalkene complex 4b was much
different from that of the secondary aminoalkene complex 4a. Both
the amine andthe alkene portions of the primary aminoalkene bind
to rhodium in 4b, and two diastereomers of the adduct are
observed. The ratio of diastereomers was ca. 3:1 in methylene
chloride, but over 10:1 in dioxane. These two diastereomers result
from different relative configurations created by binding to the two
faces of the arene and the alkene. The ratios of these diastereomers
1
adopted by the complex in solution. In particular, the H NMR
signals for the amino-substituted arene are observed between
5.42 and 6.68 ppm. The 13C NMR signals for the same arene are
observed between 89.6 and 141.4 ppm and contain coupling to
rhodium.
2. Coordination Chemistry of the Active Catalyst. 2.1.
Synthesis and Structures of Rhodium Aminoalkene Com-
plexes. Rhodium complexes of the secondary and primary ami-
noalkenes were generated indirectly from ethylene complex 1. The
ethylene in complex 1 was not displaced by the secondary
aminoalkene in a closed system, but the complex of this substrate
was generated by the method shown in Scheme 1. Dissolution of
the ethylene complex in acetonitrile released ethylene from the
system and formed the acetonitrile complex 2. Although we could
not determine the number of coordinated acetonitriles in complex
2 because it is stable only in acetonitrile, we were able to gather data
on the solid-state structure by X-ray diffraction. An ORTEP
diagram for this compound is shown in Figure 2.
This compound crystallized with two molecules in the asym-
metric unit. The biaryldialkylphosphine is bound in an η2, κ1 mode,
rather than the η6, κ1 mode of the ethylene complex 1. The
Rh-Cipso distance (2.219(3) and 2.216(3) Å) in the η2-arene is
roughly 0.05 Å longer than that in the η6-arene of 1 (2.166(3) Å),
and the Rh-Cortho distance (2.214(3) and 2.215(3) Å) in 2 is 0.06
Å shorter than that in 1 (2.271(3) Å). The dimethylamino group
P
on the arene is more pyramidized (
= 352.6 and 356.7°)
CNC
P
than that in the ethylene complex 1 ( CNC = 359.8°). Despite the
change in hapticity of the arene, the Rh-P distance (2.2058(9) and
2.2147(9) Å) is only 0.02 Å shorter than that in complex 1
(2.2383(8) Å). Finally, the Rh-N distance (2.104(3) and
2774
dx.doi.org/10.1021/ja1057949 |J. Am. Chem. Soc. 2011, 133, 2772–2782

Journal of the American Chemical Society
ARTICLE
Scheme 2
were invariant in the individual solvents, but sharp signals were
observed in the NMR spectra of these species, implying that the
two stereoisomers equilibrate rapidly on the laboratory time scale,
but slowly on the NMR time scale. Binding of the alkene was
evidenced by upfield shifts of the olefinic resonances located at
4.3-4.8, 3.37-3.6, and 2.8-2.9 ppm in the complex. 103Rh-13C
coupling was observed in all four of the olefinic 13C NMR signals
corresponding to the two diastereomeric forms of 4b (J = 15 and
14 Hz for the resonances at 52 and 55 ppm corresponding to the
vinylidene carbons and 15 and 16 Hz for the resonances at 68 and
69 ppm corresponding to the internal alkene carbons of the two
diastereomers). The similarity of the chemical shifts and coupling
constants for the analogous resonances of the two diastereomers
implies that the two isomers result from binding of the metal to the
two faces of the alkene, rather than from a cis and trans disposition
of the amino or alkene group to the phosphorus atom.
Figure 3. ORTEP diagram of the cationic portion of
[Rh(L)(CH2dCHCH2CPh2CH2NH2)]SbF6 (4b-SbF6). Ellipsoids
are shown at 50% probability. The counterion is omitted for clarity.
Binding of the amine portion of the primary aminoalkene was
deduced from the 2H NMR and 31P NMR chemical shifts and the
2
value of JRh-P. We used H NMR spectroscopy to measure the
chemical shift of the amino protons so that they would be clearly
distinguished from the many cyclohexyl resonances. The 2H NMR
chemical shift of 4b-d2 (1.53 ppm) was downfield of that of the free
primary aminoalkene 3b-d2 (0.71 ppm), indicating that the amine
was significantly perturbed by binding of the substrate. In addition,
the 31P NMR chemical shifts of the diastereomers of complex 4b
(47.8 and 46.6 ppm) and the JRh-P values (143 and 142 Hz) were
closer to those of the amine complex (56.9 ppm, 159 Hz, vide infra)
than to those of complex 1 containing a bound alkene and an η6
arene ligand (71.7 ppm, 191 Hz). Finally, the reduced hapticity of
the arene ligand was indicated by the downfield shift of the arene
resonances from the region of η6-arene complexes observed for
complex 1 to the more typical region of aromatic resonances.
Single crystals of 4b-SbF6 were obtained from a mixture of
CH2Cl2 and pentane, and the structure in the solid state is
consistent with that deduced from the solution NMR data. An
ORTEP diagram of 4b-SbF6 is shown in Figure 3. This
complex clearly adopts a structure with the ligand bound in a
κ1-η2 form through the CdC bond containing the carbon ipso
to the second aryl ring and the unsubstituted ortho carbon.
Such a binding mode is distinct from that in which a metal is
bound in a κ1 mode to the arene portion of the ligand at the
carbon ipso to the second aryl group.58 The degree of
pyramidalization of the dimethylamino group on the arene in
and the Rh-C distances of 2.111(7) and 2.144(8) Å straddle
those of 1.
To gather further information on the binding of primary
amines to the LRhþ fragment, we prepared and obtained struc-
tural data on a complex containing two primary amine donors.
The reaction of ethylene complex 1 with ethylenediamine in
THF at room temperature generated a complex 5 in which both
amino groups are bound to the metal (eq 3). This binding of
two dative donors was, again, accompanied by a reduction in the
hapticity of the arene portion of the ligand. This reduced
hapticity in solution was evidenced by the upfield 13C NMR
resonances for the ipso and the unsubstituted ortho carbons
(81.3 and 85.3 ppm) and typical aromatic resonances for the
other carbons in the bound ring.
The structure deduced by NMR was corroborated by X-ray
diffraction (Figure 4). In this structure, the biaryldialkylphosphine
ligand clearly bound in an η2, κ1 mode with a Rh-P distance
(2.2120(9) Å) that is 0.03 Å shorter than that in complex 1, which
contains the η6-arene, and nearly identical to the distance in acetonitrile-
complex 2, which contains the η2-arene. The Rh-Cipso distance
(2.180(3) Å) is similar to that in 1 (2.166(3) Å) and 0.03 Å longer
than that in 2, but the Rh-Cortho distance (2.191(3) Å) is 0.08 Å
shorter than that in 1 (2.271(3) Å) and 0.2 Å longer than that in
P
P
4b-SbF6 ( CNC = 350.2°) is more akin to that in 2 (
=
CNC
2 (2.214(3) and 2.215(3) Å). Again, the dimethylamino group on
352.6 and 356.7°) than to that in the ethylene complex 1
P
P
the arene is more pyramidized ( CNC = 350.8°) than that in the
(
CNC = 359.8°). The amino group is bound to rhodium with
P
ethylene complex 1 (
= 359.8°) and similar to that in 2
a Rh-N distance (2.124(7) Å) that is longer than the Rh-N
distances to the acetonitrile ligands in 2. The alkene CdC
distance remains that of an electrophilic alkene; the CdC
distance of the coordinated alkene (1.375(12) Å) is within
experimental error of the CdC distance in ethylene complex 1,
CNC
(352.6 and 356.7°). Finally, like the Rh-N distances to the
acetonitrile ligands in 2, the Rh1-N1 distance (2.145(3) Å) to
the nitrogen trans to the phosphine is longer than the Rh1-N2
distance (2.101(3) Å) located trans to the η2-arene ligand, and the
2775
dx.doi.org/10.1021/ja1057949 |J. Am. Chem. Soc. 2011, 133, 2772–2782

Journal of the American Chemical Society
ARTICLE
Scheme 3
Figure 4. ORTEP diagram of the cationic portion of [Rh(L)(ethyl-
enediamine)]BF4 (5). Ellipsoids are shown at 50% probability. Counter-
ion, solvent, and all hydrogen atoms other than that of amino group are
omitted for clarity.
[Rh(COD)2]BF4 in the presence of 20 equiv of the primary
aminoalkene and 1 equiv of the ligand in dioxane led to a solution
containing free phosphine ligand and no detectable coordinated
phosphine. Instead, the complex [Rh(COD)(1°-aminoalke-
ne)2]BF4 (6) was observed. Consistent with the competitive
binding of the aminoalkene and the phosphine, reaction of the
primary aminoalkene complex 4b with an excess (9 equiv) of free
aminoalkene 3b in the same solvent led to a roughly 1: 1 ratio of 4b
to free phosphine.60 In contrast to reaction of the primary
aminoalkene complex 4b with excess of the primary aminoalkene
3b, reaction of the secondary aminoalkene complex 4a with an
excess of the secondary aminoalkene 3a in dioxane did not lead to
the generation of free phosphine ligand.
distance trans to phosphorus is longer than that of the amino group
in 4b-SbF6 (2.124(7) Å), which is trans to phosphorus.
2.2. Relative Binding Abilities of the Aminoalkenes and
the Phosphine. To reveal factors controlling the reactivity of
this catalyst system toward primary and secondary amines and to
probe pathways for catalyst decomposition, we studied the relative
binding affinities of the rhodium for the primary aminoalkene, the
secondary aminoalkene, and the phosphine ancillary ligand. These
studies showed that binding of the primary aminoalkene was
favored over binding of the secondary amine. Treatment of
secondary aminoalkene complex 4a with 1.5 equiv of the primary
aminoalkene 3b led to quantitative formation of the primary
aminoalkene complex 4b (eq 4). In this case, the pseudo square-
planar geometry of the primary aminoalkene complex and the
bidentate binding mode of the primary aminoalkene within this
structure likely account for the exclusive binding of this substrate to
the LRh fragment.59
The cationic bis-amine complex 6 was isolated and fully
characterized and is related to the bis-amine complex [Rh-
(COD)(morpholine)2]þ reported by Beller.11 The similarity
between the chemical shifts of the olefinic resonances of 6(4.9-5.3
ppm) and those of the free aminoalkene (5.0-5.46 ppm), along
with the lack of Rh-C coupling to the alkene, indicate that the
amine nitrogen, not the alkene CdC bond, is bound to rhodium.
The connectivity of this species was confirmed by X-ray diffraction,
although the quality of the crystal did not allow detailed reporting of
metrical parameters.
3. Evaluation of the Competence of the Observed Species
to be Intermediates. Ethylene complex 1, acetonitrile complex
2, and secondary and primary aminoalkene complexes 4a,b are
competent to be intermediates in the cyclization of primary and
secondary aminoalkenes 3a-c or to lead to an intermediate by a
low-energy process. Secondary aminoalkene complex 4a reacted in
THF at20 °C within 5 h to form the cyclic hydroamination product
7a in quantitative yield (eq 5). The rhodium from this process
precipitated from solution, and we were not able to characterize it.
The reaction of N-p-chlorobenzylaminoalkene 3c catalyzed by
2.5 mol % of 4a with 5 mol % added COD (to help stabilize the
catalyst, vide infra) in dioxane-d8 at 70 °C gave 7c in 98% yield after
26 h (eq 6).
Primary aminoalkene complex 4b also underwent cyclization.
This complex reacted in dioxane at 100 °C within 5 min to form the
cyclic amine 7b in 53% yield (eq 7). This reaction conducted with
addition of a different primary aminoalkene occurred in a higher
These studies also showed that the phosphine and the
aminoalkene bound competitively to the Rh(COD)þ and Rh-
(aminoalkene)þ fragments (Scheme 3). The reaction of
2776
dx.doi.org/10.1021/ja1057949 |J. Am. Chem. Soc. 2011, 133, 2772–2782

Journal of the American Chemical Society
ARTICLE
80% yield.61 Reaction of primary aminoalkene 3b catalyzed by
5 mol % of4b with 10 mol % added COD in dioxane at 100 °C gave
7b in 82% yield after 3 h (eq 8). The yield of this reaction was
somewhat higher in the absence of added COD and similar in the
presence of added L.
complex 4b was observed before heating, and the signal of 4b
remained the major species observed, along with a small amount of
free L, after heating for 5 min at 100 °C. As the catalytic reaction
progressed, the signal of 4b decayed, and the signal due to free L
increased. After 10 min at 100 °C, about 30% of the ligand had
dissociated. At the end of the reaction, new four doublet signals and
the signal of free L was observed. In the reactions catalyzed by the
combination of Rh(COD)2BF4 and L, the red color of complex 4b
was not observed at room temperature. Instead, the yellow bis-
amine complex 6 was likely generated, as suggested by the yellow
color of the solution. Heating of this mixture at 100 °C generated a
small amount of 4b, but free L was the major species observed by
31P NMR spectroscopy, and the ratio of 4b to L was approximately
1:20. Thus, binding of the aminoalkene over the phosphine limits
the reactivity of this catalyst for cyclization of aminoalkenes.
5. Reactions of Deuterated Amines . 5.1. Probes for Direct
versus Indirect Addition Pathways. The cyclic amine could
form by direct addition of the N-H bond to the alkene, or it could
occur by oxidative amination, followed by addition of the released
hydrogen to the resulting enamine. To distinguish between these
mechanisms, we conducted the reaction of the deuterated N-p-
chlorobenzylamine 3c-N-d1 shown in eq 9.
Oxidative amination, followed by addition of the released
equivalent of H-D, would form the two isotopomers 7cA and
7cB in eq 9. However, direct addition would form the single
isotopomer 7cA. Consistent with direct addition, 3c-N-d1 reacted
in the presence of 5 mol % of [Rh(COD)2]BF4 and 6% of L to
form 7cA in 93% yield, with 93% deuterium content and without
formation of 7cB.
5.2. Stereochemistry of the Hydroamination Process.
Several catalytic cycles can be envisioned for the observed hydro-
amination process starting from substrate complexes 4a and 4b
(Scheme 4). Complex 4a containing the secondary amine could
react by the combination of external attack of the amine on the
coordinated alkene and protonolysis of the resulting Rh-C bond
(path A). Complex 4b containing the primary amine could react by
an analogous path after dissociation of the amino group. Complex
4a could react by the combination of olefin dissociation, N-H
bond activation, and olefin insertion into the amide or hydride
ligand, and complex 4b could react by N-H bond activation and
analogous olefin insertion (path B). Finally, these complexes could
react by the generation of an amido complex, perhaps by depro-
tonation of the substrate bound to the metal through the amino
group, followed by insertion of the alkene to form an aminoalkyl
complex that undergoes protonolysis to release the amine product
(path C).
Side-by-side reactions of secondary aminoalkene 3a catalyzed by
the combination of [Rh(COD)2]BF4 and L, by isolated acetonitrile
complex 2, and by isolated aminoalkene complex 4a showed that
the reactions catalyzed by the preformed η6, κ1 species are faster
thanthosecatalyzed bytheoriginalsystem. Thehalf-lifeofthelatter
two reactions was about 8 h, whereas only 10% conversion to
product was observed after this time from the reaction of 3a
catalyzed by [Rh(COD)2]BF4 and L.
4. Identification of the Catalyst Resting State. To deter-
mine the resting state of the catalyst, we monitored by 31P NMR
spectroscopy reactions of the secondary and primary aminoalkene
complexes catalyzed by several catalyst precursors and aminoalkene
complexes. Monitoring the reactions of the secondary aminoalkene
3a catalyzed by 2.5 mol % [Rh(COD)2]BF4 and 3 mol % L or by
2.5 mol % acetonitrile complex 2 in dioxane showed that the
secondary aminoalkene complex 4a was the only rhodium-phos-
phine complex observed throughout either reaction. Some catalyst
decomposition over the course of this reaction was observed; the
rhodium catalyst was lost from the system over time as a black
precipitate and some free L accumulated. The addition of COD
helped preserve the active catalyst, and, therefore, some of the
kinetic studies (vide infra) were conducted with added COD.
Reactions of primary aminoalkene 3b catalyzed by 5 mol % of
acetontrile complex 2 or 5 mol % of Rh(COD)2BF4 and 6 mol % L
in dioxane were also monitored by 31P NMR spectroscopy. In
reactions catalyzed by 2, only the signal of primary aminoalkene-
To distinguish mechanisms involving nucleophilic attack on an
alkene from those involving migratory insertion of the alkene, we
determined the stereoisomer formed by hydroamination of a 1,
2-disubstituted alkene. Reaction by nucleophilic attack, followed by
protonolysis, would yield the product of anti addition across the
alkene. Reaction by a pathway involving insertion, followed by
reductive elimination or protonolysis, would yield the product of
syn addition across the alkene.
2777
dx.doi.org/10.1021/ja1057949 |J. Am. Chem. Soc. 2011, 133, 2772–2782

Journal of the American Chemical Society
ARTICLE
Scheme 4
Scheme 5
Scheme 6
If the stereochemistry of this process is to be evaluated, the
catalyst must lead to addition across an internal alkene. Togenerate
a substrate for such an addition, we prepared aminoalkenes 8a-N-d1
and 8b-N-d2 in eq 10. Reaction of aminoalkene 8a-N-d1 in the
presence of 5 mol % [Rh(COD)2]BF4 and 6 mol % L led to the
cyclization product 9a-d, and reaction of aminoalkene 8b-N-d2 in
the presence of this catalyst led to the cyclization product 9b-d.
Although the reaction of 8a-N-d1 occurred in modest yield,62
a
rhodium-catalyzed hydroamination results from the unusual
structure of the metal-ligand complex or the unusual reaction
of an N-H bond by a rhodium system, we determined the
stereochemical outcome of the addition of D2 and DSiEt3 to
norbornene catalyzed by 2. When examined, rhodium-catalyzed
hydrogenations and hydrosilylations occur by syn addition to
alkenes.
The reaction of norbornene with 50 psi D2 at 100 °C or DSiEt3
at 110 °C catalyzed by acetonitrile complex 2 gave the stereo-
iosomers of norbornane-d2 and exo-2-deuterio-exo-3-triethylsilyl-
norbornane that reflect a cis addition (Scheme 5). Thus, the trans
addition pathway results from the amine reagent, not from the
particular coordination sphere of the rhodium catalyst.
7. Kinetics of the Addition Process. To identify the turnover-
limiting step of the catalytic process and any reversible processes
that would precede this step, we conducted kinetic studies on the
cyclization process. Complexes 4a and 4b containing acyclic
secondary aminoalkene bound through the alkene (4a) and the
acyclic primary aminoalkene bound through the olefin and amino
group (4b) were the observed resting states when reactions were
initiated with acetonitrile complex 2 or the aminoalkene complexes
as catalyst, at least at early reaction times before substantial com-
petitive decomposition of the catalyst occurred (vide supra). Two
mechanisms leading to the trans addition products could operate
with complexes 4a and 4b as the resting state (Scheme 6). First, the
reaction could occur by turnover-limiting attack of the amino group
on the coordinated alkene directly within the olefin-bound ami-
noalkene complex 4a or after dissociation of the amino group in
complex 4b in which the amino and olefin groups are both bound
to rhodium. Alternatively, an initial, thermodynamically unfavor-
able, reversible attack of the amine on the alkene, followed by
turnover-limiting proton transfer to release the aminoalkene, could
lead to the product and would occur with the observed resting
states. These two scenarios can be distinguished by the kinetic
isotope effect. A primary isotope effect would not be observed if
sufficient amount of pure product was isolated to obtain spectral
data to analyze the stereochemical outcome of the addition. The
reaction of 8b-d occurred in higher yield. The N-D of 9b-d
exchanged upon workup, but deuterium incorporation at the new
position β to the amino group in both products was about 75%.
The relative stereochemistry in the product was assessed by a
combination of 1H-COSY and 1D-NOE spectroscopy. Details of
this analysis are provided as Supporting Information. In brief, an
NOE was observed between Ha and Hd in the products of eq 10,
indicating a 1,3-cis diaxial relationship between these two hydro-
gens, and an NOE was observed between Ha and Hb, indicating a
cis 1,2-relationship. The resonance corresponding to Hd located
trans to the amino nitrogen is the one at the chemical shift in the 2H
NMR spectrum of the product 9-d formed from reaction of 8-N-d1.
Thus, the amino group in the product is trans to the deuterium, and
the addition of the N-H bond across the alkene occurs in an anti
fashion.
6. Evaluation of the Stereochemistry of Additions of H2
and Silanes Catalyzed by Acetonitrile Complex 2. As noted
in the Introduction, most rhodium-catalyzed reactions of X-H
bonds across alkenes lead to products of syn addition, while
complexes of many other metals catalyze reactions of N-H
bonds across alkenes to form products of anti addition. To
determine if the relative configuration in the product of our
2778
dx.doi.org/10.1021/ja1057949 |J. Am. Chem. Soc. 2011, 133, 2772–2782

Journal of the American Chemical Society
ARTICLE
Figure 5. Profiles for the decay of secondary aminoalkenes 3c and 3c-
N-d1 in the presence of 1.25 mol % [Rh(L)(NCMe)2] (2) and 10 mol %
COD at 100 °C in dioxane-d8. The linear fit to the decay of 3c
corresponds to a fit to the first 67 data points (2600 s); the linear fit
to the decay of 3c-N-d1 corresponds to a fit to the first 150 data points
(6000 s).
Figure 6. Profiles for the decay of secondary aminoalkene 3c in the
presence of 1.25% [Rh(L)(NCMe)2]BF4 (2) and 1 equiv (vs 3c) of
COD at 100 °C in dioxane-d8. The curve corresponds to an exponential
decay of 2 with a rate constant of 3.9 ꢀ 10-4 s-1
.
nucleophilic attack were turnover limiting, but a primary isotope
effect would be observed if nucleophilic attack were reversible and
followed by turnover-limiting proton transfer.
Scheme 7
The profiles for cyclization of 3c and 3c-N-d1 catalyzed by 2 in
the presence of a small amount of COD to stabilize the catalyst
(vide supra) are shown in Figure 5. These decays indicate that the
reaction is zero-order in 3c when the concentration of 3c is much
larger than that of the added COD. When the substrate concentra-
tion is high, the added COD does not compete with 3c for the open
coordination site. Under these conditions, kH/kD is 2.5 ( 0.6. This
clear primary isotope effect contrasts with that of the cyclization of
primary aminoalkenes by a rhodium catalyst containing the
aminophosphine derivative of Xantphos.28 Consistent with com-
petitive binding of COD and the aminoalkene, the reaction was
first-order in the aminoalkene 3c when the reaction was run with 1
equiv of COD (Figure 6).
The zero-order rate behavior with a small amount of added
COD, primary isotope effect, and the presence of 4a as the resting
state imply that the turnover-limiting step involves proton transfer
in a transition state comprised solely of the components contained
within 4a. These data are consistent with the mechanism involving
reversible nucleophilic attack of the amine on the alkene, followed
by turnover-limiting intramolecular protonation of the Rh-C
bond, as shown in the bottom scenario in Scheme 6.
The kinetic behavior of the reactions of primary amines was more
complex due to equilibria involving competitive binding of the
amine and the phosphine and the shorter lifetime of the catalyst.
Thus, reactions of the primary amine catalyzed by acetonitrile
complex 2 occurred with neither simple first nor simple zero-order
decay profiles. 31P NMR spectra indicated that a significant amount
of free phosphine was contained in this system. Thus, some
rhodium of the rhodium in the system lacks coordinated phosphine.
This observation suggested that the rates of the hydroamination
would be positive order in added phosphine. If the active catalyst
contains the phosphine, and if the rhodium lacking phosphine and
rhodium bound by phosphine are in equilibrium, then the reaction
would occur with a positive order in this added ligand, as summar-
ized in Scheme 7. Moreover, if the active catalyst contains the
phosphine and the equilibrium between the bound and free species
lies further toward the bound species at higher concentrations of
added ligand, then the order in ligand should decrease as the
population of rhodium is shifted toward the active ligated species. If
the rhodium were fully bound by phosphine and remained in this
state throughout the catalytic cycle (or if the barrier for exchange
between free and bound rhodium is high enough that these species
are not in equilibrium), then the reaction would be zero-order in
phosphine. If reversible ligand dissociation occurs prior to the
turnover-limiting step or the active catalyst lacks the phosphine
ligand (Scheme 7), then the reaction would occur with an inverse
order in phosphine.
The profiles for cyclization of primary aminoalkenes 3b and 3b-
N,N-d2 catalyzedby2in the absence of COD are shownin Figure 7.
From initial rates, the value of kH/kD was found to be ca. 2.4 ( 0.2.
At the end of the reaction, rhodium black was not observed,
and four unidentified doublet signals and free L were observed by
31P NMR spectroscopy.
A second set of experiments was run in the presence of 10 mol %
added COD, which is present in the catalyst precursor and was
present in our preparative reactions of secondary aminoalkenes
initiated with isolated acetonitrile complex 2.27 The added COD
appeared to stabilize the catalyst, but to retard the cyclization. We
measured the reaction rates as a function of the concentration of
ligand and obtained several pieces of data that argue for reaction
through the observed amine complex ligated by the phosphine in
equilibrium with an inactive phosphine-free complex. First, as
2779
dx.doi.org/10.1021/ja1057949 |J. Am. Chem. Soc. 2011, 133, 2772–2782

Journal of the American Chemical Society
ARTICLE
Figure 7. Profiles for the decay of primary aminoalkenes 3b and 3b-N,
N-d2 in the presence of [Rh(L)(NCMe)2] (2) at 100 °C in dioxane-d8
with no added COD. From a least-squares fit for initial rates using the
first eight data points, kH/kD = 2.4 ( 0.2.
Figure 8. Profiles for the decay of primary aminoalkene 3b in the
presence of 5% [Rh(L)(NCMe)2]BF4 (2), 18% L, and 10% COD at
100 °C in dioxane-d8. From a least-squares fit to the data over the first
50% conversion, kH/kD = 2.3 ( 0.2.
shown in Figure 8, reactions catalyzed by 2 with 18 mol % added
ligand occurred with a profile that was consistent with a zero-order
reaction. This reaction profile was analogous to that in Figure 5 for
the reaction of the secondary aminoalkene 3c. Under these
conditions, kH/kD was found to be 2.3 ( 0.2. This value is
indistinguishable from the value obtained in the absence of added
COD. Second, a plot of kobs versus [L] (Figure 9) had a clear
positive slope, indicating that the reaction occurs with a positive
dependence on free ligand. We hesitate to fit these data to a precise
equation because the catalyst lifetime is short at the lower [L], but
this plot is qualitatively consistent with an equilibrium in the
catalytic system between a phosphine-free rhodium that is cataly-
tically inactive and phosphine-bound rhodium that is catalytically
active. Consistent with this assertion, the reaction of primary
aminoalkene 3b catalyzed by 5 mol % phosphine-free bis-amine
complex [(COD)Rh(aminoalkene)2]þ (aminoalkene = 3b) 6 in
dioxane-d8 was slow, and after 24 h gave about 10% of an
aminoalkene containing an internal olefin from isomerization,
15% imine, and only 5% cyclized product (eq 11).
Figure 9. Plot of kobs versus [L] for the reaction of primary aminoalkene
3b in the presence of 5% [Rh(L)(NCMe)2]BF4 (2), 0-100% L, and
10% COD at 100 °C in dioxane-d8.
salt with an alkyl complex containing the η6, κ1-bound ligand L.
We did not find a means to conduct this reaction on a β-
aminoalkyl species without elimination of the amino group.
Thus, we prepared instead an alkylrhodium complex lacking β-
hydrogens and containing ligand L.
Neopentylrhodium complex 11 was prepared by the addition
of neopentylmagnesium chloride to the rhodium chloride com-
plex 10 at room temperature, as shown in eq 12. This species was
unstable at room temperature and was, therefore, characterized
in solution (see the Supporting Information for data). The
presence of a rhodium-bound neopentyl group was revealed by
the coupling of the neopentyl methylene carbon and hydrogens
to the 103Rh nucleus, and the π-bound arene ring was identified
by the presence of resonances upfield of the typical aromatic
region.
The reaction of tri-n-butylammonium chloride with 11
formed the chloride complex 10 (82%) and neopentane
(quantitative yield) at room temperature in only 10 min. The
rate of this process is much faster than the catalytic reaction
that requires hours at elevated temperatures. This result
implies that the proposed protonolysis of the alkylrhodium
8. Independent Synthesis of an LRh-Alkyl Complex and
Evaluation of the Rate of Protonolysis. If the hydroamina-
tion process occurs by nucleophilic attack of the amine on the
alkene, followed by protonolysis, then an alkylrhodium com-
plex containing the η6, κ1-bound ligand L must undergo
protonolysis with a rate that is equal to or greater than that
of the catalytic process. If nucleophilic attack is a thermody-
namically unfavorable and reversible step that occurs prior to
protonolysis, then this protonolysis process must be faster than
the catalytic reaction. Many alkyl complexes of the late transi-
tion metals are kinetically stable to mild acids. Thus, it was
not clear a priori whether an alkylrhodium complex of ligand L
would undergo fast proton transfer.
To assess the rate of protonation of the proposed alkylrho-
dium intermediate, we conducted the reaction of an ammonium
2780
dx.doi.org/10.1021/ja1057949 |J. Am. Chem. Soc. 2011, 133, 2772–2782

Journal of the American Chemical Society
ARTICLE
Scheme 8
hydroamination when other rhodium catalysts have led to
predominant11,67 or substantial12 amounts of products from oxida-
tive amination. With a η6, κ1-bound ligand on Rh(I), attack of the
amine on the coordinated alkene generates a coordinatively
saturated, 18-electron aminoalkyl complex. Such a coordination
sphere disfavors β-hydrogen elimination due to the lack of an open
coordination site.68-70 This lack of an open coordination site
appears to make this aminoalkyl complex sufficiently stable that the
rate of proton transfer to form amine exceeds that of β-hydrogen
elimination.
The use of these data to design more active and stable catalysts
based on the π,σ-binding mode of the ligand in this catalyst is in
progress, and two weak points must be addressed to improve upon
the existing system. First, the alkene complex containing the η6, κ1-
bound L contains a significant degree of strain. The bond angles in
the structure and the alternative η2, κ1-binding mode of several
species reveal this strain, and the precipitation of a rhodium species
during the reaction suggests that this strain limits the stability of the
metal-ligand complex at elevated temperatures in the presence of
high concentrations of amine and alkenedonors. Second, the rate is
limited by the combination of nucleophilic attack by the amine and
proton transfer. One would expect that these two steps are favored
by opposite changes to the electronic properties of the metal. Thus,
an analysis of the sensitivity of these steps to changes in the ancillary
ligand is needed to design a system in which the combination of
these two steps is faster than it is with the current system. Studies to
address these various issues will be the subject of future studies.
intermediate containing a pendant ammonium acid is suffi-
ciently fast to be part of the catalytic cycle.
9. Overall Mechanism of the Hydroamination Process.
The identification of η6, κ1 biaryldialkylphosphine complexes as the
active catalyst, olefin complex 4a and amine complex 4b as the
resting states, trans stereochemistry for addition to the alkene, and
primary kinetic isotope effect altogether point to the mechanism
shown in Scheme 8. In this scheme, the aminoalkene displaces
solvent or residual COD ligand to form an alkene complex when
the substrate contains a secondary amine or to form an amine
complex when the substrate contains a primary aminoalkene. This
binding is then followed by reversible attack of the amino group on
the bound alkene. Such nucleophilic attack on a bound alkene is not
common for rhodium systems,49,63 but is a common reaction for
cationic olefin complexes of other late transition-metal systems.64,65
Cyclization of primary amines has been more challenging to develop
with late transition-metal catalysts than cyclization of secondary
amines,13,14,28 and this study illustrates explicitly that the greater
binding affinity of the primary amine is the likely origin of this trend.
The aminoalkyl intermediate resulting from this nucleophilic
attack then undergoes protonolysis. This protonolysis appearsto be
an intramolecular process66,38 because the reaction occurs with
zero-order kinetics from the resting state containing the bound
substrate. This observation precludes the participating of a second
metal or second substrate prior to or during the turnover-limiting
step. This protonolysis process can occur by direct transfer of the
proton to the metal, followed by reductive elimination to form a
C-H bond or by proton transfer directly to the Rh-C bond.
Recent computational studies of the protonolysis of an iridium
ammoniumalkyl complex implied that protonolysis of the metal
was favored,15 but direct protonolysis of the metal-C bond could
be favored for the rhodium complex because of the lower basicity of
second row metals versus third-row metals. In either case, the
overall pathway in Scheme 8, particularly the anti addition, is more
closely related to that of palladium- and platinum-catalyzed hydro-
amination reactions than that of rhodium-catalyzed additions of H2,
H-Si, H-B, and H-C bonds across alkenes.
’ ASSOCIATED CONTENT
S
Supporting Information. Experimental procedures, com-
b
pound characterization, and kinetic plots. This material is available
free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
jhartwig@illinois.edu
’ ACKNOWLEDGMENT
We thank the NIH for support (R37-GM055382) and John-
son-Matthey for a gift of RhCl3.
’ REFERENCES
(1) Hartwig, J. F. Pure Appl. Chem. 1999, 71, 1417.
(2) Muller, T. E.; Hultzsch, K. C.; Yus, M.; Foubelo, F.; Tada, M.
Chem. Rev. 2008, 108, 3795.
(3) Hong, S.; Marks, T. J. Acc. Chem. Res. 2004, 37, 673.
(4) M€uller, T. E. In Encyclopedia of Catalysis; Horvꢀath, I. T., Ed.;
Wiley-Interscience: Hoboken, 2003; Vol. 3, p 518.
(5) Casalnuovo, A. L.; Calabrese, J. C.; Milstein, D. J. Am. Chem. Soc.
1988, 110, 6738.
(6) Dorta, R.; Egli, P.; Zurcher, F.; Togni, A. J. Am. Chem. Soc. 1997,
119, 10857.
(7) Seligson, A. L.; Trogler, W. C. Organometallics 1993, 12, 744.
(8) Kawatsura, M.; Hartwig, J. F. Organometallics 2001, 20, 1960.
(9) Kawatsura, M.; Hartwig, J. F. J. Am. Chem. Soc. 2000, 122, 9546.
(10) Utsunomiya, M.; Hartwig, J. F. J. Am. Chem. Soc. 2003, 125, 14286.
(11) Beller, M.; Trauthwein, H.; Eichberger, M.; Breindl, C.; Herwig,
J.; Muller, T. E.; Thiel, O. R. Chem.-Eur. J. 1999, 5, 1306.
(12) Utsunomiya, M.; Kuwano, R.; Kawatsura, M.; Hartwig, J. F. J.
Am. Chem. Soc. 2003, 125, 5608.
10. Proposed Origins of the Selectivity for Hydroamina-
tion over Oxidative Amination. Our identification of the
elementary steps of the catalytic cycle and the precise species
undergoing these steps allows us to make an experimentally
grounded proposal about why this system leads to products from
2781
dx.doi.org/10.1021/ja1057949 |J. Am. Chem. Soc. 2011, 133, 2772–2782

Journal of the American Chemical Society
ARTICLE
(13) Bender, C. F.; Widenhoefer, R. A. J. Am. Chem. Soc. 2005, 127,
1070.
(51) Faller, J. W.; D’Alliessi, D. G. Organometallics 2003, 22, 2749.
(52) Werner, H.; Canepa, G.; Ilg, K.; Wolf, J. Organometallics 2000,
(14) Hesp, K. D.; Stradiotto, M. Org. Lett. 2009, 11, 1449.
(15) Hesp, K. D.; Tobisch, S.; Stradiotto, M. J. Am. Chem. Soc. 2010,
132, 413.
19, 4756.
(53) Love, R. A.; Koetzle, T. F.; Williams, G. J. B.; Andrews, L. C.;
Bau, R. Inorg. Chem. 1975, 14, 2653.
(16) LaLonde, R. L.; Sherry, B. D.; Kang, E. J.; Toste, F. D. J. Am.
Chem. Soc. 2007, 129, 2452.
(17) Kimura, M.; Tanaka, S.; Tamaru, Y. Bull. Chem. Soc. Jpn. 1995,
68, 1689.
(18) Liu, G. S.; Lu, X. Y. Org. Lett. 2001, 3, 3879.
(19) Zhang, Z.; Liu, C.; Kinder, R. E.; Han, X.; Qian, H.;
Widenhoefer, R. A. J. Am. Chem. Soc. 2006, 128, 9066.
(20) Zhang, Z.; Bender, C. F.; Widenhoefer, R. A. J. Am. Chem. Soc.
2007, 129, 14148.
(54) Cheng, P. T.; Nyburg, S. C. Can. J. Chem. 1972, 50, 912.
(55) Rettig, M. F.; Wing, R. M.; Wiger J. Am. Chem. Soc. 1981, 103,
2980.
(56) Malinoski, J. M.; White, P. S.; Brookhart, M. Organometallics
2003, 22, 621.
(57) Tables of Interatomic Distances and Configuration in Molecules
and Ions; Sutton, L. E., Ed.; The Chemical Society: London, 1958.
(58) Christmann, U.; Pantazis, D. A.; Benet-Buchholz, J.; McGrady,
J. E.; Maseras, F.; Vilar, R. n. J. Am. Chem. Soc. 2006, 128, 6376.
(59) For a brief analysis of the strictly κ1 binding of primary and
secondary aminoalkenes to the rhodium catalyst containing an amino-
phosphine derivative of Xantphos, see ref 28. In this case, the primary
amine binds 50-100 times stronger than the secondary aminoalkene,
which adopts both N-bound and alkene-bound forms.
(60) A species of unknown structure was detected in the 31P NMR
spectrum during this displacement of ligand, perhaps corresponding to a
complex with the phosphine bound without an interaction of the
rhodium with the arene.
(21) Kinder, R. E.; Zhang, Z.; Widenhoefer, R. A. Org. Lett. 2008, 10,
3157.
(22) Giner, X.; Najera, C. Org. Lett. 2008, 10, 2919.
(23) Bender, C. F.; Widenhoefer, R. A. Chem. Commun. 2006, 4143.
(24) Zhang, J. L.; Yang, C. G.; He, C. J. Am. Chem. Soc. 2006, 128,
1798.
(25) Liu, X. Y.; Li, C. H.; Che, C. M. Org. Lett. 2006, 8, 2707.
(26) Michael, F. E.; Cochran, B. M. J. Am. Chem. Soc. 2006, 128,
4246.
(27) Liu, Z.; Hartwig, J. F. J. Am. Chem. Soc. 2008, 130, 1570.
(28) Julian, L. D.; Hartwig, J. F. J. Am. Chem. Soc. 2010, 132, 13813.
(29) Shen, X.; Buchwald, S. L. Angew. Chem., Int. Ed. 2010, 49, 564.
(30) For a recent example of a system that adds both primary and
secondary amines, see ref 15.
(61) It is possible that the second amine increases the rate of
dissociation of a portion of the κ2-bound aminoalkene or that it binds
the rhodium product and prevents decomposition of the reactant by the
initial unsaturated rhodium product, but we did not conduct detailed
studies on this effect.
(62) We did not seek to optimize the yield of this process.
(63) Hauger, B. E.; Huffman, J. C.; Caulton, K. G. Organometallics
1996, 15, 1856.
(64) Hartwig, J. F. Organotransition Metal Chemistry; University
Science Books: Sausalito, CA, 2010; Chapter 11: Nucleophilic Attack
on Coordinated Ligands.
(31) Gagnꢀe, M. R.; Stern, C. L.; Marks, T. J. J. Am. Chem. Soc. 1992,
114, 275.
(32) Stubbert, B. D.; Marks, T. J. J. Am. Chem. Soc. 2007, 129, 6149.
(33) Gribkov, D. V.; Hultzsch, K. C. Angew. Chem., Int. Ed. 2004, 43,
5542.
(34) Knight, P. D.; Munslow, I.; O’Shaughnessy, P. N.; Scott, P.
(65) Crabtree, R. H. The Organometallic Chemistry of the Transition
Metals, 4th ed.; Wiley: New York, 2005; Chapter 8, p 560.
(66) For an example of protonolysis of a cationic alkylpalladium
species containing nitrogen in the β position, see ref 38.
(67) Beller, M.; Eichberger, M.; Trauthwein, H. Angew. Chem., Int.
Ed. Engl. 1997, 36, 2225.
(68) McCarthy, T. J.; Nuzzo, R. G.; Whitesides, G. M. J. Am. Chem.
Soc. 1981, 103, 3396.
(69) Cross, R. J. In The Chemistry of the Metal-Carbon Bond;
Hartley, F. R., Patai, S., Eds.; John Wiley: New York, 1985; Vol. 2, p 559.
(70) Hartwig, J. F. Organotransition Metal Chemistry; University
Science Books: Sausalito, CA, 2010; Chapter 10: Elimination Reactions.
Chem. Commun. 2004, 894.
(35) M€uller, C.; Saak, W.; Doye, S. Eur. J. Org. Chem. 2008, 2731.
(36) Leitch, D. C.; Payne, P. R.; Dunbar, C. R.; Schafer, L. L. J. Am.
Chem. Soc. 2009, 131, 18246.
(37) For a recent study providing evidence for alkene insertion
during a hydrohydrazination proess, see: Hoover, J. M.; DiPasquale, A.;
Mayer, J. M.; Michael, F. E. J. Am. Chem. Soc. 2010, 132, 5043.
(38) Cochran, B. M.; Michael, F. E. J. Am. Chem. Soc. 2008, 130,
2786.
(39) Walsh, P. J.; Baranger, A. M.; Bergman, R. G. J. Am. Chem. Soc.
1992, 114, 1708.
(40) Baranger, A. M.; Walsh, P. J.; Bergman, R. G. J. Am. Chem. Soc.
1993, 115, 2753.
(41) Wood, M. C.; Leitch, D. C.; Yeung, C. S.; Kozak, J. A.; Schafer,
L. L. Angew. Chem., Int. Ed. 2007, 46, 354.
(42) Gott, A. L.; Clarke, A. J.; Clarkson, G. J.; Scott, P. Chem.
Commun. 2008, 1422.
(43) Hartwig, J. F. Organotransition Metal Chemistry; University
Science Books: Sausalito, CA, 2009; Chapter 16: Hydrofunctionaliza-
tion and Oxidative Functionalization of Olefins.
(44) Modern Rhodium-Catalyzed Organic Reactions; Evans, P. A., Ed.;
Wiley-VCH: Weinheim, 2005.
(45) Crabtree, R. H. The Organometallic Chemistry of the Transition
Metals, 4th ed.; Wiley: New York, 2005; Chapter 9, p 560.
(46) Han, X.; Widenhoefer, R. A. Angew. Chem., Int. Ed. 2006, 45,
1747.
(47) Kennedy-Smith, J. J.; Staben, S. T.; Toste, F. D. J. Am. Chem.
Soc. 2004, 126, 4526.
(48) For an example in which the reaction of a rhodium-olefin
complex led to displacement of the alkene, rather than addition, see
ref 49.
(49) Hahn, C.; Sieler, J.; Taube, R. Chem. Ber. 1997, 130, 939.
(50) Old, D. W.; Wolfe, J. P.; Buchwald, S. L. J. Am. Chem. Soc. 1998,
120, 9722.
2782
dx.doi.org/10.1021/ja1057949 |J. Am. Chem. Soc. 2011, 133, 2772–2782