Mechanistic investigation of the Ru-catalyzed hydroamidation of terminal alkynes

Article abstract of DOI:10.1021/ja111389r

The ruthenium-catalyzed hydroamidation of terminal alkynes has evolved to become a broadly applicable tool for the synthesis of enamides and enimides. Depending on the catalyst system employed, the reaction leads chemo-, regio-, and stereoselectively to a single diastereoisomer. Herein, we present a comprehensive mechanistic study of the ruthenium-catalyzed hydroamidation of terminal alkynes, which includes deuterium-labeling, in situ IR, in situ NMR, and in situ ESI-MS experiments complemented by computational studies. The results support the involvement of ruthenium-hydride and ruthenium-vinylidene species as the key intermediates. They are best explained by a reaction pathway that consists of an oxidative addition of the amide, followed by insertion of a π-coordinated alkyne into a ruthenium-hydride bond, rearrangement to a vinylidene species, nucleophilic attack of the amide, and finally reductive elimination of the product.

Full text of DOI:10.1021/ja111389r

ARTICLE
pubs.acs.org/JACS
Mechanistic Investigation of the Ru-Catalyzed Hydroamidation
of Terminal Alkynes
Matthias Arndt, Kifah S. M. Salih, Andreas Fromm, Lukas J. Goossen,* Fabian Menges, and
Gereon Niedner-Schatteburg*
Fachbereich Chemie and State Reaserch Center OPTIMAS, TU Kaiserslautern, Erwin-Schr€odinger-Strasse 52ꢀ54,
67663 Kaiserslautern, Germany
S Supporting Information
b
ABSTRACT: The ruthenium-catalyzed hydroamidation of terminal alkynes has
evolved to become a broadly applicable tool for the synthesis of enamides and
enimides. Depending on the catalyst system employed, the reaction leads
chemo-, regio-, and stereoselectively to a single diastereoisomer. Herein, we
present a comprehensive mechanistic study of the ruthenium-catalyzed hydro-
amidation of terminal alkynes, which includes deuterium-labeling, in situ IR,
in situ NMR, and in situ ESIꢀMS experiments complemented by computational
studies. The results support the involvement of rutheniumꢀhydride and rutheniumꢀvinylidene species as the key intermediates.
They are best explained by a reaction pathway that consists of an oxidative addition of the amide, followed by insertion of a
π-coordinated alkyne into a rutheniumꢀhydride bond, rearrangement to a vinylidene species, nucleophilic attack of the amide, and
finally reductive elimination of the product.
’ INTRODUCTION
efficient Ru catalysts and established the addition of amide-type
nucleophiles to terminal alkynes as a general method for the
synthesis of enamide derivatives. The same reaction principle is
the basis for a number of preparatively useful Ru-catalyzed ad-
dition reactions to alkynes, for example, their hydration with
formation of aldehydes,20 the addition of carboxylic acids to give
enol esters,21 their hydroamination with formation of imines or
enamines,22 their hydrothiolation to vinyl sulfides,23 and the ad-
dition of alcohols to form vinyl ethers.24
Over the last years, a range of customized protocols were dis-
closed for the anti-Markovnikov addition of various N-nucleo-
philic amides, thioamides, and imides across terminal CꢀC triple
bonds. They provide an expedient and chemo-, regio-, and stereo-
selective synthetic entry to enamides, thioenamides, and eni-
mides (Scheme 2).
With a catalyst system generated in situ from bis(2-methallyl)-
(cycloocta-1,5-diene)ruthenium(II) [(cod)Ru(met)2], tri-n-
butylphosphine (P(n-Bu)3), and 4-dimethylaminopyridine
(DMAP), tertiary (E)-enamides can be synthesized in high
yields and selectivities from terminal alkynes and secondary
amides.25 The stereoselectivity can be reversed in favor of the
corresponding (Z)-enamides when employing bis-(dicyclo-
hexylphosphino)methane (dcypm) and water instead of P(n-
Bu)3 and DMAP. The reaction proceeds smoothly even in the
presence of sensitive functional groups such as esters, ethers,
ketones, halides, or silanes. Various amides, anilides, ureas, bislac-
tams, carbamates, and even amide-type chiral auxiliaries can be
Enamides are valuable structural elements in natural products
with interesting biological activities and in pharmaceutical drug
lead compounds1 showing antibiotic,2 antitumor,3 anthelmintic,4
antifungal, and cytotoxic activities (Figure 1).5
In addition, enamides can serve as versatile synthetic intermedi-
ates, particularly in pericyclic and photochemical reactions for the
formation of heterocycles,6 [4 þ 2]-cycloadditions,7 cross-cou-
pling reactions,8 Heck olefinations,9 enantioselective additions,10
or asymmetric hydrogenations.11
Traditional syntheses include the condensation of aldehydes
and ketones with amides or dehydration of hemiaminals,12 the
Curtius rearrangement of R,β-unsaturated acyl azides,13 and the
elimination of β-hydroxy-R-silylamides (Peterson reaction).14
Several metal-catalyzed approaches have also been investigated,
such as the isomerization of N-allylamides15 and catalytic cross-
coupling reactions of amides and vinyl halides, pseudohalides or
enol ethers.16 Problems often encountered using these methods
are the harsh reaction conditions, the formation of (E)- and (Z)-
product mixtures, or the use of expensive or poorly available starting
materials.
Over the last years, a particularly convenient synthetic entry to
this important substrate class has emerged, namely, the addition
of amides to terminal alkynes (Scheme 1).
This reaction mode is the most atom-economic transforma-
tion of all the catalytic reactions based on carboxylic acid derivatives
that we have investigated over the last years.17 On the basis of
pioneering studies by Heider et al.18 and Watanabe et al.,19 who
were the first to observe that ruthenium complexes mediate the
addition of certain amides to terminal alkynes, we have developed
Received: January 3, 2011
Published: April 26, 2011
r
2011 American Chemical Society
7428
dx.doi.org/10.1021/ja111389r J. Am. Chem. Soc. 2011, 133, 7428–7449
|

Journal of the American Chemical Society
ARTICLE
be widely applied in organic synthesis. However, the reaction
mechanism has so far remained speculative.
The aims of the present study were to investigate the coordina-
tion type of the alkyne during the reaction, to clarify how the
regio- and stereochemistry is controlled, and to identify the rate-
determining step of the catalytic cycle.
’ MECHANISTIC CONSIDERATIONS
Numerous potential reaction mechanisms have to be evalu-
ated with appropriately designed mechanistic studies and control
experiments.
Several catalytic cycles have previously been proposed for the
Ru-catalyzed addition of nucleophiles such as amides, amines,
carboxylic acids, and water to CꢀC triple bonds. The first mecha-
nism for anti-Markovnikov-selective hydroamidations was pos-
tulated by Watanabe but was not supported by experimental data
(Scheme 3).19
This mechanism will further be referred to as Mechanism A. It
involves the oxidative addition of an amide, insertion of a π-co-
ordinated alkyne into the RuꢀN or RuꢀH bond, and reductive
elimination of the enamide product. Stabilizing interactions between
the oxygen atom of the carbonyl group and the Ru center in a four-
or six-membered intermediate (17 and 18) that forms following
alkyne insertion were used to explain the regioselectivity, in both
cases leading to the formation of the anti-Markovnikov product.
Uchimaru proposed a similar mechanism for the Markovnikov-
selective Ru-catalyzed hydroamination of terminal alkynes,22a in
which a π-coordinated alkyne inserts into the RuꢀN bond of a
Ru-amine species. The Markovnikov selectivity is explained by the
formation of a sterically less hindered Ru-enamine intermediate.
Indications for Mechanism A would be provided by a depen-
dence of the rate-determining step on the acidity of the amide and
the electronic and steric properties of the Ru-complex. Moreover,
Figure 1. Enamide substructure in bioactive or functional molecules.
Scheme 1. Addition of Amides to Terminal Alkynes
used as NꢀH nucleophiles. Recently, we have shown that for
many substrates, a similar level of activity can be achieved when
the catalytically active species is generated in situ from inexpen-
sive ruthenium trichloride hydrate (RuCl3 3H2O), P(n-Bu)3,
3
DMAP, K2CO3, and water.26
1
Ru-hydride species may be detectable via H NMR or electro-
Subsequently, modified catalyst systems were developed that
allowed extending the substrate scope to thioamides27 and imides.28
In this context, we made the discovery that for imides, which are
acidic NꢀH nucleophiles, the addition of a Lewis acid rather than
an auxiliary base is essential for achieving turnover of the catalyst.
The conversion of primary amides into secondary enamides is
challenging due to the higher nucleophilicity of secondary over
primary amides, leading to double vinylation products. However,
a catalyst system consisting of the Lewis acid ytterbium(III)
triflate in combination with (cod)Ru(met)2 and an electron-rich,
sterically demanding bidentate ligand (dcypb, 1,4-bis(dicyclohe-
xylphosphino)butane) allowed the selective conversion of pri-
mary amides to secondary enamides.29 This protocol gives access
to (Z)-enamides in high yields and selectivities, whereas the
(E)-enamides can be prepared by subsequent in situ double-
bond isomerization with triethylamine at higher reaction tem-
peratures in the same pot. Furthermore, after minor modifica-
tions, the same bimetallic system can be used for the Z-selective
addition of secondary amides and imides to terminal alkynes
in the absence of the primary amide functionality, yielding
the corresponding enamides and enimides with Z/E-selectivities
greater than 20:1.30
nspray ionization mass spectroscopy (ESIꢀMS) investigations if
concentrations are high enough, and the reaction of 1-deuter-
ioalkynes should give rise to a product deuterated exclusively in
the 1-position.
Dixneuf proposed a different mechanism to explain the selective
formation of anti-Markovnikov addition products in the addition
of carboxylic acids to alkynes. His pathway can directly be translated
to hydroamidation reactions (Scheme 4).32
This mechanism will further be referred to as Mechanism B. Its
key step is the formation of a Ruꢀvinylidene complex 20 via a
1,2-proton shift at the alkyne moiety, followed by an attack of a
nucleophile in the R-position to the ruthenium center. After
protonolysis of the ruthenium intermediate 25 and regeneration
of the active ruthenium species 19, an anti-Markovnikov enol
ester 26 or enamide 6ꢀ13 is formed. In contrast, the alternative
direct addition of a nucleophile to a coordinated alkyne should
result in the formation of the Markovnikov product. This mecha-
nism provides a sound explanation for the anti-Markovnikov se-
lectivity of the reaction and its limitation to terminal alkynes.
Experiments with isolated Ruꢀvinylidene complexes confirmed
that their reaction with nucleophiles will indeed lead to the addition
product. As an alternative to vinylidene formation via a 1,2-
proton shift, a sequence consisting of an oxidative addition of the
alkyne C(sp)ꢀH bond to ruthenium followed by a 1,3-proton
shift was also proposed as an entry to this mechanism.
The applicability of Ru-catalyzed hydroamidation reactions is
illustrated in Figure 2. The examples include the natural products
alatamide, lansiumamides A and B, lansamide I, and botrylla-
mides C and E.29,31 They demonstrate that the hydroamidation
of alkynes has meanwhile reached a high level of maturity and can
Mechanism B would again predict a dependence of the reaction
rate on the electronic and steric properties of the Ru-complex,
7429
dx.doi.org/10.1021/ja111389r |J. Am. Chem. Soc. 2011, 133, 7428–7449

Journal of the American Chemical Society
ARTICLE
Scheme 2. Ru-catalyzed Addition of N-Nucleophiles to Terminal Alkynes
Scheme 3. Mechanism A: Hydroamidation Pathway Postu-
lated by Watanabe
Figure 2. Representative example for the addition of terminal alkynes.
relatively slow reaction step. Therefore, in contrast to Mechanism
A, the acidity of the alkyne C(sp)ꢀH bond rather than that of the
amide NꢀH bond should have an additional influence on the
CꢀH bond cleaving vinylidene formation step. Therefore,
and the reaction would benefit from the stabilizing effect of
electron-donating ligands on the high oxidation state of the
ruthenium center. We assume the vinylidene formation to be a
7430
dx.doi.org/10.1021/ja111389r |J. Am. Chem. Soc. 2011, 133, 7428–7449

Journal of the American Chemical Society
ARTICLE
Scheme 4. Mechanism B: Hydroamidation in Analogy to Dixneuf’s Carboxylation of Terminal Alkynes
Scheme 5. Mechanism C: Hydroamidation in Analogy to
Wakatsuki’s Hydration of Terminal Alkynes
Scheme 6. Mechanism D: Hydroamidation via Oxidative
Addition of the Amide, Insertion of the Alkyne, and
RuꢀVinyl/ꢀVinylidene Rearrangement
kinetic isotope effects should be measurable in experiments with
1-deuterioalkynes, and the resulting product should exclusively
be deuterated in the 2-position, because proton shifts to the
internal alkyne C-atom are postulated. ESIꢀMS and in situ IR ex-
periments could help to verify reactive intermediates.
Wakatsuki et al. thus derived a different mechanism involving
Ruꢀvinylidene intermediates.20b Again, this may be translated to
hydroamidation reactions (Scheme 5).
Vinylidene intermediates have been confirmed for other ruthe-
nium-catalyzed addition reactions.32 One example is the addition
of alcohols to alkynes, where a 1,2-proton shift was observed in
isotopic labeling experiments.24b However, the intermediacy of
vinylidene intermediates does not necessarily call for proton shifts,
and alternative mechanisms for Ruꢀvinylidene formation have
been proposed for the hydration of terminal alkynes20b and for
stoichiometric reactions of terminal alkynes with rutheniumꢀ
hydride complexes.33 On the basis of computational studies,
Wakatsuki and Caulton concluded that vinylidene intermediates
are formed via rearrangement of Ru-vinyl species for these reactions,
whereas pathways via 1,2- or 1,3-proton shifts are energetically
unfavorable (see Schemes 5 and 6).
The key step in this mechanism, which will further be referred
to as Mechanism C, is the protonation of a π-coordinated alkyne
resulting in the formation of a cationic RuIVꢀvinyl intermediate
28. Its rearrangement to the RuꢀHꢀvinylidene species 29 was
proposed to be the rate-determining step. Addition of a nucleo-
phile and reductive elimination gives the aldehyde 33 or the enamide
6ꢀ13, respectively. In this mechanism, the alkyne C(sp)-proton
is transferred to the metal center and subsequently reattached to
its original carbon atom. For the hydration of alkynes, 1,2-proton
shifts were indeed not observed in deuteration studies. This mech-
anism offers an explanation for the anti-Markovnikov selectivity
and the limitation to terminal alkynes. However, it involves
cationic intermediates in the high oxidation states of þ4 or
even þ6, depending on whether the RudC bond in species 29 is
For the hydration of terminal alkynes, the absence of proton
shifts was corroborated by deuterium-labeling experiments.
7431
dx.doi.org/10.1021/ja111389r |J. Am. Chem. Soc. 2011, 133, 7428–7449

Journal of the American Chemical Society
ARTICLE
viewed as a covalent bond or a as the coordination of a neutral
carbene ligand to the metal center. This may be reasonable for
hydration reactions in aqueous solvents, but is less likely for
hydroamidations under almost neutral conditions in toluene.
If the hydroamidation proceeded via Mechanism C, the use of
protic and more polar solvents should result in a higher reaction
rate. Again, electron-donating ligands should enhance the catalyst
activity due to their stabilizing effect on higher oxidation states of the
ruthenium center, and sterically demanding ligands should facilitate
the reductive elimination of the product. During the slow vinylidene
formation step, the C(sp) proton is transferred to the ruthenium
center. This proton may be detectable via 1H NMR. When com-
paring the reactions of alkynes and 1-deuterioalkynes, a primary
kinetic isotope effect is expected, as a CꢀD rather than a CꢀH
bond has to be cleaved. The resulting product should be deuterated
exclusively in the 1-position.
Scheme 7. Redox-Neutral Mechanism E for the Hydroami-
dation of Terminal Alkynes
Caulton and co-workers investigated pathways leading to the
formation of Ruꢀvinylidene intermediates.33 They found that
π-coordinated alkynes readily insert into RuꢀH bonds to form
Ruꢀvinyl complexes. The Ruꢀvinyl species then rearrange to
RuꢀHꢀvinylidene complexes. They isolated all postulated Ru
intermediates and confirmed the pathways by DFT calculations.
These showed reasonable energy barriers for the proposed pathway,
and substantially higher barriers for an oxidative addition of the
alkyne followed by a 1,3-proton shift. It is possible to incorporate
this alternative route to vinylidene species into a new catalytic
cycle for hydroamidations, which will further be referred to as
Mechanism D (Scheme 6).
This catalytic cycle starts with the oxidative addition of an
amide to an active Ru0 species (34). After insertion of the alkyne
into the Ruꢀhydride bond and rearrangement to the vinylidene
species 39, the amide attacks the carbon atom R to the ruthenium
center, leading to the anti-Markovnikov enamide product
(6ꢀ13) via reductive elimination. The oxidation state of the
Ru center changes from 0 via þ2 to þ4 during the catalytic cycle,
or only from 0 to þ2 if the vinylidene ligand is interpreted as
neutral carbene. The vinylidene formation proceeds without
proton shifts, and at the end of the reaction, the alkyne proton
is linked to the original C-atom.
For Mechanism D, a dependence of the reaction rate on the
acidity of the NꢀH group of the amide as well as the C(sp)ꢀH
function of the alkyne should be detectable, resulting in measur-
able kinetic isotope effects both when N-deuterated amides or
1-[D]-alkynes are used as starting materials. Electron-rich ligands
will stabilize the high oxidation state of the ruthenium center, and
the use of sterically demanding ligands should facilitate the
reductive elimination step. Reaction intermediates might be
identifiable via ESIꢀMS and in situ IR experiments. It might
also be interesting to investigate by in situ NMR whether the Ru-
complexes are capable of activating NꢀH bonds.
experiments, we postulated a redox-neutral ligand exchange
mechanism, which will further be referred to as Mechanism E
(Scheme 7).
In the catalyst preformation step, all ligands initially bound to
the Ru-precursor are exchanged, with formation of rutheniumIIꢀ
amide complexes 44. In the first step of the catalytic cycle,
the alkyne coordinates to the ruthenium center (45). Depending
on the steric bulk of the phosphines, the amide attacks from either
the inner or the outer coordination resulting in the formation of
E- or Z-configured enamides (6ꢀ13). Bulky ligands are likely to
favor an external attack resulting in the formation of an (E)-
Ruꢀenamide complex 47, which releases the corresponding (Z)-
enamide (7, 9, 11, or 13) after protonolysis, along with a
regenerated active RuII-species 44. In contrast, an insertion of
the alkyne into the RuꢀN-bond of a coordinated amide is favorable
in the presence of smaller ligands, giving rise to (E)-enamides (6,
8, 10, or 12). Over the entire cycle, the Ru center remains in the
oxidation state þ2. A likely driving force behind the reaction is
the continuous exchange of basic for more acidic ligands at the
metal center.34 Mechanism E offers an explanation for the stereo-
chemistry of the hydroamidation, but cannot adequately address
the limitation to terminal alkynes and selectivity for the anti-
Markovnikov products.
If Mechanism E holds true, an inverse secondary kinetic
isotope effect should be detectable for 1-deuterioalkynes, be-
cause it involves a rehybridization from sp to sp2, which is more
favorable for the stronger CꢀD bond.35 Furthermore, no
rutheniumꢀhydride species should be detectable via 1H NMR.
The nucleophilicity of the amide should have a strong influence
on the reaction rate, and additives that increase the nucleophilic
character of the amide should enhance the reaction rate. The
ruthenium center remains in the oxidation state of þ2 and merely
activates the alkyne for nucleophilic attack by removing electron-
density from the π-system. Therefore, less electron-rich phosphines
should lead to a higher catalyst activity, whereas sterically deman-
ding phosphines should lead to a reduced activity because the
In our previous work, we had excluded reaction mechanisms
starting from Ru0 species because no coupling product of the
methylallyl ligands from the (cod)Ru(met)2 precursor could be
detected, and instead, free isobutene was observed via GCꢀMS
and 1H NMR spectroscopy, which appeared to support a redox-
neutral ligand-exchange reaction.29 A control experiment with
1-[D]-hex-1-yne showed no 1,2-proton shift during the reaction.28
In situ NMR and ESIꢀMS experiments of the catalyst preforma-
tion step confirmed that all ligands of (cod)Ru(met)2 are
exchanged, and that cationic Ru-amide-phosphine and Ru-
amide-phosphine-DMAP species are formed in the addition of
secondary amides to terminal alkynes.26 On the basis of these
7432
dx.doi.org/10.1021/ja111389r |J. Am. Chem. Soc. 2011, 133, 7428–7449

Journal of the American Chemical Society
ARTICLE
Table 1. Overview of Experimental Findings for a Set of Control Experiments
ligand-exchange reactions and the formation of the rutheniumꢀ
enamide intermediate are disfavored.
(5b), the deuterium was transferred to the 20-position of 6ab in
very high selectivity (entry 2). These results indicate that the
predominant mechanism does not involve proton shifts. The
trace formation of 20-deuterated product can be accounted for by
HꢀD exchange reactions at a relatively acidic site of the
hydroamidation product, possible competing mechanisms, or
unproductive vinylidene rearrangements.
The reaction of 2-pyrrolidone (1a) and 1-[D]-hex-1-yne (5a)
under Z-selective conditions (2 mol % (cod)Ru(met)2, 3 mol %
Cy2PCH2PCy2, 2 equiv H2O) also proceeded mostly without
proton shift (entry 3). The ratio between 20-deuterated (7ab)
versus 10-deuterated product (7aa) dropped to a moderate value
of 4:1 for the inversely deuterated starting materials (entry 4). In
both reactions, the deuteration grades in the products were only
moderate, which we attribute to the presence of water in the reaction
mixture responsible for background HꢀD exchange reactions.
Indeed, when we replaced the water, which is essential for an
effective hydroamidation protocol, by deuterium oxide, the deu-
teration rates in the products (7aa and 7ab) were high (entries 5
and 6). However, the regioselectivity of the deuteration was only
moderate, and in the reaction of 1-[D]-hex-1-yne (5a), a doubly
deuterated product was observed. All these findings suggest that
background HꢀD exchange overlays regioselective deuterium
incorporation in the presence of water.
The catalytic addition of succinimide (2a) to 1-[D]-hex-1-yne
(5a) under E-selective conditions (5 mol % (cod)Ru(met)2, 15
mol % P(i-Pr)3, 4 mol % Sc(OTf)3) proceeds without proton
shift and leads exclusively to the incorporation of the deuterium
in the 10-position of the corresponding enimide product (8aa,
entry 7). When using the inversely deuterated starting materials,
the deuterium is mainly incorporated in the 20-position of the
enimide product (8ab, entry 8). In addition to the expected
products, traces of the product deuterated in the R-position to
the imide carbonyl group were detected, which can be rationa-
lized by H/D exchange at this acidic position.36
An overview of all predictions we made in this paragraph for
Mechanisms AꢀE in a set of control experiments is presented in
Table 1. These predictions are made under the assumption that
the concentrations of all characteristic intermediates are high
enough for detection and that the reaction steps leading to possible
kinetic isotope effects (KIE) are slow.
All mechanisms presented above are in principle feasible, but
give contradictory predictions for the outcome of simple control
experiments and spectroscopic studies. The combined experi-
mental findings presented herein provide strong evidence that
the reaction proceeds via Mechanism D.
’ DEUTERATION STUDIES OF HYDROAMIDATIONS
We started our mechanistic investigation with hydroamidation
experiments using 1-deuterioalkynes. Mechanisms A, C, D, and E
predict that the deuterium should end up in the geminal position
to the amide nitrogen, whereas Mechanism B predicts that the
deuterium should be transferred to the vicinal carbon. In the
inverse experiment, with N-deuterioamides and nondeuterated
alkynes, the deuterium should be incorporated to the geminal
position to the amide nitrogen for Mechanism B and bind to the
vicinal atom for Mechanisms A, C, D, and E.
Deuteration studies were carried out for the additions of primary
and secondary amides as well as imides, using both the E- and
the Z-selective methods. The results are summarized in Table 2.
The addition of 2-pyrrolidone (1a) to 1-[D]-hex-1-yne (5a,
Deuteration grade, DG = 92%) using the E-selective protocol (2
mol % (cod)Ru(met)2, 6 mol % P(n-Bu)3, 4 mol % DMAP) led
to incorporation of the deuterium almost exclusively in the
10-position of the corresponding enamide (6aa), that is, in
geminal position to the amide (entry 1). The 20-deuterated
product (6ab) was detected in traces only. In the analogous
reaction of 1-[D]-2-pyrrolidone (1b, DG = 85%) and 1-hexyne
7433
dx.doi.org/10.1021/ja111389r |J. Am. Chem. Soc. 2011, 133, 7428–7449

Journal of the American Chemical Society
ARTICLE
Table 2. Hydroamidation with Deuterated Starting Materialsa
7434
dx.doi.org/10.1021/ja111389r |J. Am. Chem. Soc. 2011, 133, 7428–7449

Journal of the American Chemical Society
ARTICLE
Table 2. Continued
a Reaction conditions: Amide or imide (1.00 mmol), alkyne (2.00 mmol), (cod)Ru(met)2 (2 mol %), P(n-Bu)3 (6 mol %), DMAP (4 mol %), toluene
(3 mL), 100 °C, 15 h, selectivity determined by 1H NMR; DG = deuteration grade. b Cy2PCH2PCy2 (3 mol %) instead of P(n-Bu)3, H2O (2.00 equiv)
instead of DMAP. c Cy2PCH2PCy2 (3 mol %) instead of P(n-Bu)3, D2O (2.00 equiv) instead of DMAP. d The doubly deuterated product was mainly
e
formed with a DG of 91%. (cod)Ru(met)2 (5 mol %), P(i-Pr)3 (15 mol %), Sc(OTf)3 (4 mol %), DMF-d7 (3 mL), 60 °C, 15 h, isolated yields,
selectivities determined by 1H NMR. f DMF (3 mL) instead of DMF-d7. g A product deuterated in the 3-position of the imide was detected in traces.
h (cod)Ru(met)2 (2 mol %), P(n-Bu)3 (6 mol %), Sc(OTf)3 (4 mol %), DMF (3 mL), 60 °C, 15 h, isolated yields, selectivities determined by 1H NMR.
i (cod)Ru(met)2 (5 mol %), dcypb (6 mol %), Yb(OTf)3 (4 mol %), DMF (3 mL), 60 °C, 6 h, isolated yields, selectivity determined by 1H NMR.
Similar results were observed also for the Z-selective protocol
(2 mol % (cod)Ru(met)2, 6 mol % P(n-Bu)3, 4 mol % Sc-
(OTf)3). Thus, the reaction of succinimide (2a) with 1-[D]-hex-
1-yne (5a) yielded the enimide mainly deuterated in the
10-position (9aa, entry 9), whereas the addition of N-[D]-
succinimide (2b, DG = 81%) to 1-hexyne (5b) led to the
formation of the product with near-quantitative deuterium
incorporation in the 20-position (9ab, entry 10). The hydro-
amidation therefore also proceeded without proton shift, as did
the reaction of the secondary amides 1a and 1b.
The reaction of benzamide (3a) with 1-[D]-hex-1-yne (5a)
under Z-selective conditions (5 mol % (cod)Ru(met)2, 6 mol %
dcypb, 4 mol % Yb(OTf)3) furnished the corresponding enam-
ide 10aa deuterated exclusively in the 10-position (entry 11).
Analogously, the deuterium was selectively incorporated in
the reaction of N,N-[D2]-benzamide (3b, DG = 75%) and
1-hexyne (5b) under otherwise identical conditions (10ab,
entry 12). It is worth mentioning that, in contrast to the
previously reported hydroamidation protocol for the addition
of primary amides,29 we had to perform these two deuteration
experiments without water, because in the presence of water
HꢀD exchange reactions overlaid regioselective deuterium
incorporation.
’ KINETIC INVESTIGATIONS OF HYDROAMIDATIONS
We next investigated the kinetics of hydroamidations by
means of in situ IR spectroscopy. Using a ReactIR 45 m FT-IR
spectrometer (3.2 scans/s, 8 cmꢀ1 resolution), we monitored the
hydroamidation of 2-pyrrolidinone (1a) and 1-hexyne (5b) by
the disappearance of the CꢀC triple bond valence oscillation of
1-hexyne (5b) at 2122 cmꢀ1 and the appearance of a CdO
valence oscillation of the enamide product (6a) at 1725 cmꢀ1
.
Figure 3 illustrates the starting material consumption as well as
the product formation and the temperature inside the reaction
vessel over a period of 30 min. The reaction starts almost im-
mediately after the reaction vessel is placed in an aluminum block
preheated to 100 °C and is complete within several minutes.
Evaluation of the spectroscopic data with the iC IR software
using the ConcIRT algorithm also allowed to detect the appear-
ance of short-lived reaction intermediates by their CdC vibra-
tion stretches at 1607 cmꢀ1 (see Supporting Information). This
frequency is in an area typical for rutheniumꢀvinylidene species.37
However, in-depth studies under carefully optimized conditions
would be required to unambiguously confirm these species.
Attempts to slow down the reaction by lowering the tempera-
ture were unsuccessful. Below 100 °C, the hydroamidation was
sluggish, and undesired alkyne dimerization became the predo-
minant process. Even at the optimum reaction temperature, the
catalyst preformation did not always proceed at the same speed,
so that the variability of overall reaction rates was rather high.
It was therefore unfeasible to perform comparative kinetic
studies with deuterated and nondeuterated substrates in separate
vessels to determine kinetic isotope effects. Instead, we performed
On the basis of these results, the formation of vinylidene in-
termediates via a 1,2-proton shift or via oxidative addition of the
alkyne followed by a 1,3-proton shift as postulated in Mechanism
B can be ruled out for both E- and Z-selective hydroamidation
reactions. Mechanisms A, C, D, and E are all in agreement with the
results of the present deuteration studies.
7435
dx.doi.org/10.1021/ja111389r |J. Am. Chem. Soc. 2011, 133, 7428–7449

Journal of the American Chemical Society
ARTICLE
Figure 3. In situ IR-experiment: Concentration trends for the hydroamidation of 1-hexyne (5b) with 2-pyrrolidinone (1a).
competition experiments in which one of the starting materials
was added in large excess as a 1:1 mixture of its deuterated and
nondeuterated form (Table 3). From the ratio of hydrogen versus
deuterium incorporation in the enamide product, we calculated
the relative reaction rates.
Table 3. Determination of KIE Values by Competition
Experimentsa
In the reaction of 2-pyrrolidinone (1a) with 1-hexyne (4
equiv, DG adjusted to 50%), the enamide (6a/6aa) was obtained
with a DG of 39% (entry 1). This translates to a kinetic isotope
effect (KIE) of 1.6. Similarly, the reaction of succinimide (2a)
with 1-hexyne (4 equiv, DG adjusted to 50%) gave a product
(9a/9aa) with 30% DG, translating to a KIE of 2.3 (entry 2). In
the analogous reaction of benzamide (3a), the product DG was
40% (10a/10aa), corresponding to a KIE of 1.5 (entry 3).
In all cases, normal kinetic isotope effects(KIE=kH/kD =nH/nD)
greater than 1 were observed. During the reaction, the hybridiza-
tion of the C(1) carbon changes from sp to sp2 and any intermediate
should thus have a higher or the same p-character as the starting
material. However, a reaction step during which the p-character
of a CꢀH bond increases is known to lead to an inverse secondary
KIE. Thus, the KIE should always have a value smaller than 1,
unless the sp CꢀH bond is cleaved in a slow reaction step, and a
sp2 CꢀH bond is reinstated at the same carbon atom in a later,
non rate-determining step.35 This implies that, despite their
relatively small magnitudes, the values must result from a primary
KIE. Mechanism A, which does not involve a bond cleavage but
rather an insertion of the alkyne into a RuꢀH or RuꢀN bond
with sp to sp2 rehybridization, can thus be excluded as the main
reaction pathway. Mechanism E is also incompatible with the
observed KIE > 1, as in this process, the CꢀH bond is also not
cleaved.
In Mechanisms C and D, a rearrangement step from a ruthe-
niumꢀvinyl to a rutheniumꢀhydrideꢀvinylidene species takes
place, during which the C(sp)ꢀH/D bond is cleaved. The KIE
values greater than 1 are in good agreement with these pathways.
Although primary KIEs for deprotonation processes usually have
values of 4ꢀ7, the relatively low values of 1.5ꢀ2.3 observed here
are easily rationalized when considering that the overall catalytic
process consists of several steps. For example, if one reaction step
is slower by a factor of 5 because of an isotope effect, but this
individual reaction step requires only one-forth of the catalyst
turnover time, the overall KIE would be 2 rather than 5.
a Reaction conditions: Amide (1.00 mmol), 1-hexyne (2.00 mmol), 1-[D]-
hex-1-yne (2.00 mmol), (cod)Ru(met)2 (2 mol %), P(n-Bu)3 (6 mol %),
DMAP (4 mol %), toluene (3 mL), 100 °C, 15 h, product ratio
1
determined by H NMR. b Determined by the ratio of nondeuterated
c
to deuterated product yields. Sc(OTf)3 (4 mol %) instead of DMAP,
DMF (3 mL) instead of toluene, 60 °C, 15 h. d (cod)Ru(met)2 (5 mol %),
dcypb (6 mol %), Yb(OTf)3 (4 mol %), DMF (3 mL), 60 °C, 6 h.
Indeed, in the reaction of 2-pyrolidinone (4 equiv, DG adjusted to
50%) and hex-1-yne (5b) (Scheme 8), the enamide (6a/6ab) was
obtained with a DG of 30%, translating to a KIE of 2.3. The fact that a
noticeable KIE is again observed revealsthatthisstepisalsorelatively
slow. The KIE value for each individual step is thus probably larger
than 4, resulting in a smaller observed KIE for the overall reaction
because the catalytic cycle involves at least two slow steps.
7436
dx.doi.org/10.1021/ja111389r |J. Am. Chem. Soc. 2011, 133, 7428–7449

Journal of the American Chemical Society
ARTICLE
Scheme 8. Competition Hydroamidation Experiment with 1-[D]-2-Pyrrolidinone (1b)
Figure 4. In situ 2H NMR experiments with 1-[D]-hex-1-yne (5b) and 1-[D]-2-pyrrolidinone (1b). (a) Free 1-[D]-hex-1-yne; (b) triplet signal; (c)
oligomerization products; (d) free 1-[D]-2-pyrrolidinone; (e) 10-deuterium of the enamide; (f) 20-deuterium of the enamide; (g) benzene-d6.
Another possible explanation for the relatively small values is
that the transition state is a nonlinear one.38
Overall, only Mechanisms C and D, both involving Ruꢀ
vinylidene intermediates formed without proton shifts, are in
agreement with the outcome of both the deuteration studies and
the kinetic investigations.
alkyne, which can be attributed to alkyne oligomerization products.
Some of them might also originate from Ruꢀvinyl species, but an
unambiguous assignment of signals to such intermediates was
not possible. When both 1-[D]-hex-1-yne (5b) and 2-pyrrolidi-
none (1a) were present, the same triplet (b) was detected along
with a strong signal for the C(1)-deuterated enamide product
(e). Alkyne oligomerization products (c) and C(2)-deuterated
enamide (f) were detected only in traces.
’ NMR STUDIES OF HYDROAMIDATIONS
This triplet (b) was not in good agreement with literature
NMR data for ruthenium alkyne and vinyl complexes (Figure 5).
Terminal protons of coordinated alkynes should appear around 5
ppm and have a large HꢀP coupling.39 RuꢀH and Ruꢀvinyl
species should also appear at chemical shifts different than 1.77
We next performed in situ NMR studies with the goal of
identifying potential Ru-bound organic fragments. However, the
reaction proceeds very rapidly when using increased amounts of
Ru catalyst, and the spectra contained numerous signals pertain-
ing to ligands, additives, and byproducts that often obscured relevant
species. To specifically monitor species derived from the alkyne
and amide starting materials, we used deuterated derivatives and
2
ppm.33 Considering the small value of DꢀH couplings in H
NMR spectra40 and the broad shape of triplet b, this signal might
be assigned to the vinylic proton of a Ruꢀvinylideneꢀphosphine
species. For a known trans-[(dppm)2(Cl)RudCdCHnBu]PF6
complex (dppe = bisdiphenylphosphino methane) with the iden-
tical 1-hexyne-derived vinylidene ligand, Dixneuf et al. reported a
triplet of quintets at 2.5 ppm in the1H NMR spectrum (Figure 5).41
Coupling constants of 7.9 and 2.8 Hz for the vinylic proton at the
vinylidene entity could be determined, resulting from a coupling
to two neighboring protons and a long-range coupling to four
phosphine atoms. The chemical shift of triplet b is in good range
for such a vinylic proton, considering that the Ru center should
be more electron-rich because of the stronger donor capacity of
alkyl phosphine compared with the aryl phosphine ligands. That
should lead to an upfield shift of the proton signal. The observed
coupling constant of 2.2 Hz is in agreement with a long-range
2
followed the reaction by H NMR. This greatly simplified the
NMR spectra obtained (Figure 4). Unfortunately, the sensitivity
of this spectroscopic method was determined to be rather low.
2H NMR spectra of a mixture containing the catalyst (20 mol %
(cod)Ru(met)2, 60 mol % P(n-Bu)3, 40 mol % DMAP) and
1-[D]-hex-1-yne (5a) only were recorded in toluene, using ben-
zene-d6 as an internal standard (Figure 4 and Supporting Infor-
mation).
At room temperature, only the signal pertaining to the starting
material was visible (a), confirming that the catalyst activation
requires elevated temperatures. Indeed, at 100 °C, a new broad
triplet (b) was detected at δ = 1.77 ppm (J = 2.2 Hz). Moreover, a
series of signals in the range of 4.89ꢀ6.20 ppm (c) rapidly appeared
and increased in intensity with concomitant consumption of the
7437
dx.doi.org/10.1021/ja111389r |J. Am. Chem. Soc. 2011, 133, 7428–7449

Journal of the American Chemical Society
ARTICLE
Figure 5. 1H- and 31P NMR chemical shifts for selected rutheniumꢀphosphine complexes.
Figure 6. In situ 1H NMR experiments with 2-pyrrolidinone (1a) at ambient temperature after heating to 100 °C for 5 min. (a) 1H NMR experiment;
(b) 1H{31P}-NMR experiment.
coupling to two phosphine atoms. The broad shape might indicate
that the coupling to the neighboring protons, which is expected to
be arround 1 Hz for such a DꢀH coupling,40 could not be resolved
in this in situ 2H NMR experiment. These results suggest that, in
the absence of the amide, a Ruꢀvinylidene species is formed,
possibly via 1,2-proton shift. This is in agreement with the observa-
tion by Dixneuf that alkyne oligomerization reactions proceed via
vinylidene species.32,42
In an attempt to find evidence for RuꢀH species resulting
from the oxidative addition of amides, we recorded the 2H NMR
spectrum for a mixture of 1-[D]-2-pyrrolidinone (1b) with the
catalyst system (40 mol % (cod)Ru(met)2, 120 mol % P(n-Bu)3,
80 mol % DMAP) in toluene using benzene-d6 as an internal
standard (Figure 4). However, it contained only the signal for the
free amide (d). Subsequent addition of 1-[D]-hex-1-yne (5a) led
to a rapid disappearance of this signal with concomitant appear-
ance of the signals for vinylic deuterium atoms of the enamide
product (e and f). The reaction simply proceeded too rapidly to
allow a detection of any intermediates by 2H NMR.
We next investigated the reaction mixtures with the help of
2 1
1H NMR, as this is more sensitive than H NMR. In the H
NMR spectrum of a mixture containing the catalyst (20 mol %
(cod)Ru(met)2, 60 mol % P(n-Bu)3, 40 mol % DMAP) and
1-hexyne (5b) but no amide (see Supporting Information), multiple
signals in the range of 4.66ꢀ6.12 ppm were observed, but none
were detected below 0 ppm in the region characteristic for RuꢀH
species. Such signals would have been expected for Mechanisms B
2
and C. A corresponding signal for triplet b observed in the H
1
NMR experiment at 1.77 ppm could not be found in the H
NMR spectrum since the area from 0 to 3 ppm was overlaid with
multiple broad proton signals of alkyl groups, for example, from
the phosphine ligands.
In contrast, after briefly heating a mixture of the catalyst and
2-pyrrolidinone (1a) to 100 °C, the 1H NMR spectrum showed
several new signals below 0 ppm (Figure 6): A group of peaks
consisting of a duplet of triplets, two triplets and two quartets
appeared between ꢀ8 and ꢀ23 ppm with coupling constants
between 22 and 33 Hz for the triplets and quartets and 108 Hz for
7438
dx.doi.org/10.1021/ja111389r |J. Am. Chem. Soc. 2011, 133, 7428–7449

Journal of the American Chemical Society
ARTICLE
Figure 7. In situ 31P NMR experiments with 2-pyrrolidinone (1a) measured at 22 °C after heating to 100 °C for 5 min. (a) 31P NMR experiment;
(b) 31P{1H}-NMR experiment.
Figure 9. In situ P,P-COSY experiments with 2-pyrrolidinone (1a) at
Figure 8. In situ H,P-HMQC experiments with 2-pyrrolidinone (1a) at
ambient temperature after heating to 100 °C for 5 min.
ambient temperature after heating to 100 °C for five minutes.
signal at ꢀ37.26 ppm could unambiguously be assigned to free
the duplet. Because of the extreme upfield shift of these signals,
they are likely to originate from RuꢀH species.33 Coupling con-
stants of 20ꢀ30 Hz are typical for cis HꢀRuꢀP couplings in
ruthenium hydride species stabilized by phosphines. The observed
coupling constants of 108 Hz are in the expected range for trans
HꢀRuꢀP coupling in such complexes.43 That the observed
couplings indeed resulted from HꢀRuꢀP interactions could be
tri-n-butylphosphine.
H,P-HMQC (Heteronuclear Multiple Quantum Coherence)
and P,P-COSY (Correlation spectroscopy) experiments were
used to elucidate which of these phosphine signals correspond to
the RuꢀH species (Figures 8 and 9).
The intensity of the proton signals at ꢀ8.55 and ꢀ18.66 ppm
species was sufficient to detect them in the 2D NMR experi-
ments. The H,P-HMQC (Figure 8) showed cross-peaks between
the RuꢀH-signal at ꢀ8.55 and phosphorus signals at ꢀ8.20
ppm and 13.77 ppm, which are in a reasonable range for Ru-
coordinated phosphines. The signal at ꢀ8.20 ppm integrates as
1
verified by phosphorus-decoupled H NMR experiments with
the same sample. Here, all signals changed to singlets.
The 31P-spectrum revealed that many different phosphine
species are present in the reaction mixture (Figure 7). Only the
7439
dx.doi.org/10.1021/ja111389r |J. Am. Chem. Soc. 2011, 133, 7428–7449

Journal of the American Chemical Society
ARTICLE
Scheme 9. In Situ Formation of RuꢀH-Phosphine Com-
plexes Starting from (Cod)Ru(met)2 and the Resulting
1H-NMR Coupling Patterns
carbonꢀRuꢀphosphorus couplings. However, Ruꢀvinylidene
species should appear at chemical shifts higher than 300 ppm,
Ruꢀvinyl species between 120 and 170 ppm, and Ruꢀalkyne
π-complexes around 75 ppm.44
1
The corresponding H NMR spectrum showed a duplet of
triplets at around ꢀ7.4 ppm and a duplet of triplets of duplets at
around ꢀ8.7 ppm. Before the addition of phenylacetylene-2-13C
(5c), the signal at ꢀ8.7 ppm appeared as a duplet of triplets with
HꢀP couplings of 108 and 27 Hz, whereas after the addition of
phenylacetylene-2-13C (5c), the HꢀP couplings changed in part
to 87 and 27 Hz and an additional coupling of 7 Hz appeared
which must result from a proton-carbon coupling. The HꢀC
coupling might originate from a π-coordination of phenylacety-
lene-2-13C (5c) to species 52 (Scheme 9), the formation of a
Ruꢀvinylideneꢀhydride such as 29 and 39, or a Ruꢀhydrideꢀ
enamide species such as 31 and 40.
one, the signal at 13.77 ppm as two P-atoms in the 31P- spectrum.
In the P,P-COSY spectrum (Figure 9), a cross-peak between the
two phosphorus signals is observed, confirming that all signals
belong to a single Ru-species likely to contain one hydride and
three phosphine ligands. The best interpretation for the HꢀP
coupling constants for the proton signal at ꢀ8.55 ppm (JHꢀP = 27
and 108 Hz) is that two of the phosphines are in cis- and one is in
trans-position to the RuꢀH bond (Scheme 9, 52).
After heating the reaction mixture briefly to 100 °C, the spectra
revealed that the reaction was already complete, as expected for
this highly reactive alkyne.
We continued our in situ NMR studies with the investigation
of the Z-selective hydroamidation of secondary amides (see
Supporting Information). The first generation catalyst system
for the Z-selective hydroamidation of secondary amides proved
to be unsuitable for NMR investigations because of the high
amount of water (8 equiv) required for this catalyst system,25
causing massive solubility problems and preventing an effective
shimming during the measurements. Therefore, we used a modified
protocol of the second generation catalyst system.30 The experi-
ments were performed in DMF-d7 instead of chlorobenzene and
the paramagnetic Yb(OTf)3 was substituted by diamagnetic
Sc(OTf)3. Because of the high catalyst loading (40 mol %
(cod)Ru(met)2, 45 mol % dcypb, 80 mol % Sc(OTf)3, 2-pyrro-
lidinone (1a), DMF-d7) only broad signals were detected in the
1H NMR spectrum. After centrifugation of the suspension in the
NMR tube and briefly heating to 100 °C, the 1H NMR spectrum
showed several signals between ꢀ15 and ꢀ25 ppm. These
signals below 0 ppm confirm the formation of Ruꢀhydride
species under Z-selective reaction conditions, indicating a similar
reaction mode for this protocol. Unfortunately, no characteristic
HꢀP couplings could be determined. The corresponding 31P
NMR spectrum showed several low intensity phosphorus signals
between ꢀ20 and 70 ppm.
The H,P-HMQC (Figure 8) also showed a cross-peak for the
proton signal at ꢀ18.66 ppm and the phosphorus signal at
1
52.69 ppm. A quartet was observed for this proton in the H
1
NMR which merged to a singlet in the P-decoupled H NMR
spectrum. In addition, the P,P-COSY (Figure 9) showed a clear
cross-peak between the phosphorus signals at 52.69 and 19.26 ppm.
The signal at 52.69 ppm integrates as one P-atom, and the signal
at 19.26 ppm as two P-atoms in the 31P- spectrum. These com-
bined findings indicate an octahedral Ru-species with three phos-
phine ligands and one DMAP ligand in plane (Scheme 9, 54).
Two phosphorus signals are expected for species 54, one for the
two P-atoms next to, and one for the P-atom in trans position to
1
the DMAP ligand. Therefore, the observed quartet in the H
NMR spectrum must actually be a duplet of triplets with nearly
identical cis HꢀP couplings, which leads to an overlay of two
triplets to one quartet.
1
The remaining upfield signals in the H NMR showed very
similar splitting patterns that disappear in the P-decoupled proton
spectrum and are also likely to originate from RuꢀH phosphine
complexes. Beside the duplet of triplets at ꢀ8.55 ppm and
the pseudo quartet at ꢀ18.66 ppm, two triplets at ꢀ11.50
and ꢀ12.51 ppm and one quartet at ꢀ22.77 ppm were detected
in the 1H NMR. Their coupling patterns and coupling constants
in the range of 24ꢀ36 Hz also suggest RuꢀHꢀphosphine
complexes with, respectively, two and three phosphines coordi-
nated in cis-position to the hydride. Scheme 9 provides for an
overview of possible RuꢀH species that would be in agreement
with the spectroscopic data obtained.
The results from these NMR studies strongly suggest that
RuꢀHꢀphosphine complexes are present in the reaction mix-
tures of hydroamidation reactions. The experiments also show
how readily RuꢀH species are formed when an amide is added to
a Ruꢀphosphine catalyst, presumably via oxidative insertion into
the NꢀH bond.19,22,45 These observations, in combination with
the findings of Caulton et al., who reported that RuꢀH-
complexes rapidly react with alkynes under formation first of
Ruꢀvinyl and then of Ruꢀvinylidene complexes,33 all point in
the direction of Mechanism D. After all, Mechanisms A and D
are the only ones that involve RuꢀH species formed via an
oxidative addition of NꢀH nucleophiles. In Mechanisms E,
RuꢀH species are not involved at all and can be ruled out. In
Mechanisms B and C, RuꢀH species are formed in the reaction
of a Ru-precursor with an alkyne without participation of the
amide. However, in the NMR experiments, the fact that RuꢀH
species were observed in the presence of amide but not in the
presence of alkyne contributes to the evidence against Me-
chanisms B and C. Considering that the results of the labeling
experiments and the kinetic studies are incompatible with
We next added phenylacetylene-2-13C (5c) to the NMR sample
containing 2-pyrrolidinone (1a), 40 mol % (cod)Ru(met)2, 120
mol % P(n-Bu)3, 80 mol % DMAP and toluene-d8 which had
shown the signals for RuꢀH species, and performed further 13C-
and 1H NMR experiments (see Supporting Information).
The proton-decoupled 13C NMR spectrum showed multiple
signals below 200 ppm. A triplet detected at 214 ppm and a
duplet of duplets of duplets at 216 ppm seemed to originate from
Ruꢀalkyne intermediates, since no signals of an organic frag-
ment should appear at chemical shifts higher than 200 ppm.
It is not straightforward to assign these signals. The cou-
pling constants between 8 and 28 Hz seem to result from
7440
dx.doi.org/10.1021/ja111389r |J. Am. Chem. Soc. 2011, 133, 7428–7449

Journal of the American Chemical Society
ARTICLE
Figure 10. Assignment of [(n-Bu)3PꢀC4H7]þ (m/z = 257.2) as the product of an allylic substitution reaction of methylallyl ligands with tri-n-
butylphosphine.
Mechanisms A, B, and E, Mechanism D at this stage appears to
be most likely.
cyclooctatriene (C8H10) or anionic cyclooctadienyl fragments
(C8H102-) provides a simple explanation for the observed signal
patterns. During the electrospray process, several toluene
molecules seemed to condense to the detected cationic Ru-
fragments. Such species are highly unstable and unlikely to exist
in the reaction solution in significant abundances.
’ IN SITU ESIꢀMS EXPERIMENTS
The electronspray ionization mass spectroscopy (ESIꢀMS)
has evolved as a key technology in the investigation of reac-
tion mechanism. In pioneering work, Chen et al., for example,
used ESIꢀMS to identify key intermediates for the Ru-catalyzed
olefin metathesis,46 and Pfaltz et al. and di Lena et al. used ESIꢀ
MS technology to develop more efficient and more selective
catalyst systems.47 Intrigued by the predictive power of these
investigations, we performed a series of in situ ESIꢀMS experi-
ments with the goals of investigating the catalyst preformation
step and identifying reaction intermediates that further support
any of the proposed catalytic cycles.
For the toluene solution of P(n-Bu)3 and (cod)Ru(met)2, three
signals of Ru species at m/z = 505.3, 761.5, and 779.5 were detected
matching the calculated patterns of [Ru(cod)(P(n-Bu)3)(tol)-
(H)]þ (P(n-Bu)3 =PC12H27), [Ru(met)(cod)(P(n-Bu)3)2(tol)]þ
and [Ru(cod)(P(n-Bu)3)(tol)4)] þ Hþ.
For the toluene solution of (cod)Ru(met)2, P(n-Bu)3 and
DMAP, two strong signals at m/z = 783.5 and 803.5 were detected,
that matched the calculated patterns for [Ru(P(n-Bu)3)2(tol)3-
(H)]þ and [Ru(P(n-Bu)3)3(tol)(H)2] þ Hþ.
Whenever P(n-Bu)3 was present the reaction solution, a
series of three signals at m/z = 217.2, 257.2, and 417.5 was
detected matching the calculated patterns of [OPC12H26]þ,
[(n-Bu)3PꢀC4H7]þ and [O(PC12H26)2]þ fragments (Figure 10).
The signals for [OPC12H26]þ and [O(PC12H26)2]þ fragments
might have been caused by partial oxidation of P(n-Bu)3, but
[(n-Bu)3PꢀC4H7]þ fragment must be seen as evidence for a re-
duction process at the Ru center, in which a phosphine reacts
with a methylallyl ligand with formation of a phosphonium salt
that can be further deprotonated to the corresponding phos-
phorus ylide species ((n-Bu)3PdCH(C3H5)). This coupling is
known to occur in the reaction of bis(2-methylallyl)palladium
chloride dimer with phosphines giving rise to Pd0ꢀphosphine
complexes such as Pd(P(n-Bu)3)4.48
There is ample literature evidence that the cod ligand is im-
mediately replaced by more strongly coordinating phosphine
and/or DMAP ligands, giving rise to a LnRu(met)2 species (L =
phosphine, DMAP).26,29 To differentiate between the proposed
mechanisms, the decisive question to be addressed is whether the
methylallyl ligands remain bound to RuII (consistent with Mechan-
isms B and C), become protonated and replaced by another ligand
(consistent with Mechanisms B, C and E), or are cleaved via a
reductive elimination step leading to Ru0 species (consistent with
Mechanisms A and D).
In this context, we investigated several combinations of
(cod)Ru(met)2, P(n-Bu)3 and DMAP by ESIꢀMS, both at room
temperature and after briefly heating to 100 °C. The spectra
obtained at room temperature from all possible solutions showed
low overall intensities and no clear assignment to Ru-species was
possible (see Supporting Information). This is not surprising, as
the expected neutral Ru(II) complexes with strongly coordinat-
ing counterions would be hard to detect. However, after briefly
heating the solutions to 100 °C, the overall intensities increased
and clear signals for RuII species could be detected in all cases,
except for the toluene solution of (cod)Ru(met)2 alone.
The solution of (cod)Ru(met)2 and DMAP showed three
signals at m/z = 371.1, 625.2, and 716.3 which could only be
matched with the calculated pattern for [Ru(met)(DMAP)(tol)]þ
(met = C4H7ꢀ, DMAP = C7H10N2, tol = toluene, C7H8),
[Ru(met)2(cod-H2)(DMAP)(tol)2] þ Hþ (cod-H2 = C8H10)
and [Ru(met)2(cod-H2)(DMAP)(tol)3]þ, illustrating that
DMAP indeed coordinates to the Ru center and is able to replace
the cod ligand. The dehydrogenation of the cod ligand leading to
This suggests that during catalyst preformation, the nucleo-
philic P(n-Bu)3 attacks one of the methylallyl ligands with formation
of a phosphonium salt. In this process, RuII is reduced to Ru0
while the second methylallyl ligand is protonated, for example, by
the acidic proton of the phosphonium salt, with formation of
isobutene. The resulting phosphorus ylide species are known to
decompose to the corresponding phosphine oxide and the alkene.
Indeed, upon heating the reaction mixture to 100 °C, the
intensity of the signal at m/z = 257.2 decreased and a signal at
m/z = 217.2 was detected, which is characteristic for trin-
butylphosphine oxide ((n-Bu)3PdO). The formation of isobu-
tene as well as (n-Bu)3PdO in hydroamidation reactions was
previously confirmed via GC- and NMR-spectroscopy.29a
We had previously dismissed mechanisms starting from Ru0
intermediates (Mechanisms A and D), because we had never
detected byproducts resulting from the reductive elimination of
7441
dx.doi.org/10.1021/ja111389r |J. Am. Chem. Soc. 2011, 133, 7428–7449

Journal of the American Chemical Society
ARTICLE
Figure 11. ESIꢀMS spectra for the E-selective hydroamidation of 2-pyrrolidinone (1a) and 1-hexyne (5b). (a) Heating to 100 °C for 5 min without
1-hexyne; (b) 40 min reaction time; (c) 150 min reaction time.
allyl fragments and had thus favored Mechanism E. However, this
new experimental data gives at least indirect evidence for the
formation of LnRu0 species, which unfortunately are particularly
hard to detect by ESIꢀMS, because such species cannot be ionized
by ligand dissociation.
We attempted to overcome this hurdle by using charged
phosphine ligands. Unfortunately, the spectra obtained using
[2-(dicyclohexylphosphino)ethyl] trimethylammonium chlor-
ide [(C6H11)2P(C2H4)N(CH3)3Cl] instead of P(n-Bu)3 were
inconclusive. Only few signals of ruthenium intermediates were
detected, and they resulted mostly from ruthenium chloride
species. In fact, we observed a strong signal at m/z = 225.2, which
could unambiguously be assigned to a [(cy)2P(C2H4)]þ frag-
ment (cy = cyclohexyl = C6H11), indicating that the trimethyla-
mine entity is cleaved under the reaction conditions. The
exchange of chloride by other counterions, for example, hexa-
fluorophosphate, did not have any beneficial effect on the com-
plexity of the resulting spectra. We also tried to induce a ligand
7442
dx.doi.org/10.1021/ja111389r |J. Am. Chem. Soc. 2011, 133, 7428–7449

Journal of the American Chemical Society
ARTICLE
Figure 12. ESIꢀMS spectrum of a mixture of the catalyst and 2-pyrrolidinone (1a) under Z-selective hydroamidation conditions and after 5 min at
100 °C.
exchange of P(n-Bu)3 by [(C6H11)2P(C2H4)N(CH3)3]þ in a
preformed catalyst solution (2 mol % (cod)Ru(met)2, 6 mol %
P(n-Bu)3, 4 mol % DMAP) by adding the charged phosphine
ligand afterward. Unfortunately, we could not observe any new
Ru species in the corresponding ESIꢀMS spectra. The charged
amonium phosphine ligand was simply not stable enough for
in situ ESIꢀMS experiments. The syntheses of other non-amo-
nium-basedcharged phosphine ligands, which are also suitable for
hydroamidation reactions, are underway.
Next, we added 2-pyrrolidinone (1a) to the catalyst system
consisting of 2 mol % of (cod)Ru(met)2, 6 mol % of P(n-Bu)3 and 4
mol % of DMAP at 100 °C. At this stage, signals at m/z = 712.3 and
792.4 appeared, which, based on their location and exact isotope
pattern, could be assigned to [Ru(P(n-Bu)3)2(DMAP)(pyr)]þ
(pyr =2-pyrrolidinyl anion, C4H6NOꢀ) and [Ru(P(n-Bu)3)3-
(pyr)]þ (Figure 11, Spectrum a).26
These species can result from the dissociation of a 2-pyrroli-
dinyl anion ([M-(pyr)]þ) of [Ru(P(n-Bu)3)2(DMAP)(pyr)2]
and [Ru(P(n-Bu)3)3(pyr)2] (Mechanisms B, C and E) but also
from [Ru(P(n-Bu)3)2(DMAP)(pyr)(H)] and [Ru(P(n-Bu)3)3-
(pyr)(H)] by protonation and H2 release ([M þ H]þ-H2)49
during the ionization process (Mechanisms A and D).
1-Hexyne (5b) was then injected to the mixture and further
samples were taken after 5, 15, 30, 40, 150, and 470 min at 100 °C
(Figure 11, Spectra b and c, and Supporting Information).
Already after 5 min, the signals at m/z = 712.3 and 792.4
disappeared. Strong signals appeared at m/z = 798.4 and 833.5
and weaker signals at m/z = 593.3, 675.4, 716.4, 757.5, 880.5,
921.7, 1003.8, and 1085.9.
ESIꢀMSꢀCIDꢀMS (CID = collision-induced dissociation)
was in agreement with the assignment and showed mainly the
cleavage or the gradual decay of DMAP and P(n-Bu)3 ligands. This
intermediate is in agreement with all postulated mechanisms.
The signal at m/z = 716.4 matched the calculated pattern for
[Ru(P(n-Bu)3)(DMAP)2(pyr)(hex)(H)] þ Hþ (hex = C6H10)
and the weak signal at m/z = 631.3 that of [Ru(P(n-Bu)3)-
(DMAP)2(hex)(H)]þ. Both species are likely to result from the
ionization of a [Ru(P(n-Bu)3)(DMAP)2(pyr)(hex)(H)] com-
plex. This might result from a Ruꢀvinyl complex such as 28 or 38
with a suitable ligand sphere (m/z = 833.5), via vinyl/vinylidene
rearrangement with concomitant dissociation of a P(n-Bu)3
ligand (consistent with Mechanisms C and D). The weak signals
at m/z = 751.5, 836.5, and 918.6 might be explained by [Ru(P(n-
Bu)3)2(DMAP)2(H)]þ, [Ru(P(n-Bu)3)2(DMAP)2(pyr)(H)] þ
Hþ and [Ru(P(n-Bu)3)2(DMAP)2(enamide)(H)] þ Hþ
fragments (enamide = C10H16NOꢀ) resulting from the ioniza-
tion of [Ru(P(n-Bu)3)2(DMAP)2(pyr)(H)] (Mechanisms A and
D) and [Ru(P(n-Bu)3)2(DMAP)2(enamide)(H)] (Mechanisms
A to E).
The species at m/z = 798.4 matched the pattern calculated for
[Ru(P(n-Bu)3)(DMAP)2(enamide)(vinyl)] þ Hþ and might
result from ionization of [Ru(P(n-Bu)3)(DMAP)2(enamide)-
(vinyl)]. The CID-fragmentation of this species showed the
dissociation of a P(n-Bu)3 ligand and of a fragment with the mass
of the enamide product (6a). As it contains two 1-hexyne (5b)
molecules, this species is likely to be an intermediate in the
formation of double alkyne insertion enamide products, which
are observed as minor side products in hydroamidations. Its
intensity decreases sharply after the first minutes of the reaction,
while all other signals discussed above increased and the signal at
m/z = 833.5 remained at a high level.
The signals at m/z = 833.5 and 918.6 matched the calculated
patterns for Ruꢀvinyl species [Ru(P(n-Bu)3)2(DMAP)2(vinyl)]þ
(vinyl = C6H11ꢀ) and [Ru(P(n-Bu)3)2(DMAP)2(pyr)(vinyl)]
þ Hþ (Figure 11, Spectrum b). They are both likely to result
from the ionization of the [Ru(P(n-Bu)3)2(DMAP)2(pyr)-
(vinyl)] complex. Further fragmentation of this species via
After the hydroamidation reaction was complete (150 min)
the intensity of the signals pertaining to hydroamidation inter-
mediates decreased, and a set of signals that had been detected at
7443
dx.doi.org/10.1021/ja111389r |J. Am. Chem. Soc. 2011, 133, 7428–7449

Journal of the American Chemical Society
ARTICLE
Figure 13. ESIꢀMS spectrum of a mixture of the catalyst, 2-pyrrolidinone (1a) and 1-hexyne (5b) under Z-selective hydroamidation conditions and
after 5 min at 100 °C.
very low intensities earlier on, now became predominant.
The signals at m/z = 593.4, 675.4, 757.4, 839.5, 921.6,
1003.7, and 1085.5 (Figure 11, Spectrum c) matched the
calculated pattern for a series of RuIII complexes with the
chemical composition of [Ru(P(n-Bu)3)2(pyr)(hex)x(H)2] þ
Hþ bearing up to six 1-hexyne molecules (5b, x = 0ꢀ6). These
species seem to be intermediates of alkyne oligomerization,
which becomes the main reaction once most of the amide is
consumed.
We next injected 1-hexyne (5b) to the mixture and further
samples were taken after 5, 25, 130, and 360 min at 100 °C
(Figure 13 and Supporting Information).
After 5 min of heating in the presence of 1-hexyne (5b), the
signals at m/z = 919.5 and 1002.5 disappeared and strong signals
at m/z = 593.4, 757.5, 839.6, 919.6, and 1001.7 as well as weaker
signals at m/z = 649.4, 675.4, 731.5, 859.5, 1083.7, and 1165.8
appeared. The masses often deviated by 2, 4, 6, or 8 mass units
from expected molecular formulas. This can be explained with ex-
periments by Leitner et al. who found that dcypb and bis-(di-
cyclohexylphosphino)propane (dcypp) ligands at Ru-centers
easily dehydrogenate.50 This thermal activation of sp3 CꢀH bonds
of bisphosphine Ruꢀbisallyl complexes leads to η3-cyclooctenyl
bridged Ru-complexes and extrusion of up to three protons from
one cyclohexyl ring. We believe that similar dehydrogenation
reactions took place under our hydroamidation reaction condi-
tions. This would explain why signals that match [M]þ ꢀ 2, 4, 6,
or 8 were detected.
The signal at m/z = 916.6 matched the calculated pattern of a
[Ru(dcypm-H2)(hex)5(H)]þ fragment (dcypm-H2 = Cy2P-
(CH2)P(Cy)(C6H9), formed via extrusion of H2) and was con-
firmed to be nonidentical with the species at m/z = 915.5, observed
in the absence of 1-hexyne (5b, Figure 12). The ESIꢀMSꢀ
CIDꢀMS (see Supporting Information) of both signals were
completely different. Whereas for the peak at m/z = 915.6, inter
alia, the stepwise dissociation of four 1-hexyne (5b) molecules was
observed, and for the peak at m/z = 915.5, stepwise decay of two
phosphine ligands was observed. We believe that the species
corresponding to the dominating signal at m/z = 915.6 is formed
via dissociation of a 2-pyrrolidinyl anion ([M-(pyr)]þ) of
[Ru(dcypm-H2)(hex)5(pyr)(H)] or via protonation and H2 re-
lease ([M þ H]þ-H2)49 of [Ru(dcypm)(hex)5(H)2], both pre-
sumably intermediates in alkyne oligomerization side reactions.
The smaller signal at m/z = 1001.7 could be assigned to a
[Ru(dcypm)2(vinyl)]þ fragment likely to originate from a
Next, we investigated the Z-selective hydroamidation of 2-pyrro-
lidinone (1a) and 1-hexyne (5b) in an analogous set of ESIꢀMS
experiments.
A reaction mixture of the catalyst (2 mol % (cod)Ru(met)2,
3 mol % dcypm, 8 equiv water) and 2-pyrrolidinone (1a) in
toluene was heated to 100 °C for 5 min. The spectrum obtained
at this stage showed two major signals at m/z = 919.5 and 1002.5,
which, based on their location and exact isotope pattern, could be
assigned to [Ru(dcypm)2(H)]þ (dcypm = bis-(dicyclohexyl-
phosphino)methane, P2C25H46) and [Ru(dcypm)2(pyr)]þ
fragments (Figure 12).
Both cationic species seem to result from the same Ru-
intermediate [Ru(dcypm)2(pyr)(H)]. The formation of a Ru-
intermediate such as 16 and 35 with a suitable ligand sphere can
be explained best via oxidative addition of an amide to a neutral
Ru0-phosphine species (consistent with Mechanisms A and D). It
is most likely that the cationic species are formed via dissociation
of a 2-pyrrolidinyl anion ([M-(pyr)]þ) or protonation and H2
release ([M þ H]þ-H2)49 of [Ru(dcypm)2(pyr)(H)] during the
ionization process. Further fragmentation of both species via
ESIꢀMSꢀCIDꢀMS (see Supporting Information) was in
good agreement with this assignment and showed the same
fragmentation species formed via gradual decay of dcypm and
additionally via the cleavage of one 2-pyrrolidinyl anion from
[Ru(dcypm)2(pyr)]þ in the case of the species at m/z =
1002.5.
7444
dx.doi.org/10.1021/ja111389r |J. Am. Chem. Soc. 2011, 133, 7428–7449

Journal of the American Chemical Society
ARTICLE
Scheme 10. Catalytic Cycle for the E-Selective Hydroamidation of 1-Hexyne (5b) and 2-Pyrrolidinone (1a), L = ligand (P(n-Bu)3
or DMAP)
[Ru(dcypm)2(vinyl)(pyr)] species with the general formula 17,
18 or 38, formed by insertion of one 1-hexyne molecule (5b) into
a RuꢀH bond of 16 or 35 (consistent with Mechanisms A and D).
The corresponding Ruꢀhydrideꢀvinylidene species 39 would
show the identical mass and isotope pattern. This is so, since both
bidentate phosphine ligands would remain attached to the Ru
center even if one of the RuꢀP bonds is cleaved during the vinyl/
vinylidene rearrangement step (consistent with Mechanism D).
Four RuIII species could be assigned to the signals at m/z =
593.4, 649.4, 675.4, and 731.5 matching the calculated patterns
of [Ru(dcypm-H2)(pyr)(H)]þ, [Ru(dcypm)(pyr)(H)(H2O)3]þ,
[Ru(dcypm)(pyr)(vinyl)]þ and [Ru(dcypm)(pyr)(vinyl)-
(H2O)3]þ fragments. These species occur with only low inten-
sities. They are most likely formed by oxidation side reaction
during the sample extraction and injection into the ESIꢀMS
instrument. Nevertheless, these RuIII complexes—bearing 2-pyr-
rolidinyl-, hydrido-, and/or 1-hexenyl ligands—suggest that an
oxidative addition step of the amide and an insertion step of the
alkyne into a RuꢀH bond is involved in the catalytic cycle of
the Z-selective hydroamidation (consistent with Mechanisms A
and D).
The five signals at m/z = 757.5, 839.6, 859.5, 1083.7, and 1165.8
match the calculated patterns of [Ru(dcypm)x(hex)y(H)]þ species
bearing one or two dcypm ligands (x = 1 or 2) and up to four
1-hexyne molecules (5b, y = 1ꢀ4). These species are also likely
formed via dissociation of a 2-pyrrolidinyl anion ([M-(pyr)]þ) of
[Ru(dcypm)x(hex)y(pyr)(H)] or via protonation and H2
release ([M þ H]þ-H2)49 of [Ru(dcypm)x(hex)y(H)2]. They
constitute intermediates of the alkyne oligomerization side
reaction.
After 25 min of heating, the intensities of all Ru-species remained
at a high level and two new Ru-species at m/z = 865.5 and 947.6
could be detected (see Supporting Information). These new signals
matched the calculated pattern of [Ru(dcypm)(hex)3(OH)(tol)]þ
and [Ru(dcypm)(hex)4(OH)(tol)]þ fragments and are most likely
formed via protonation and H2 release ([M þ H]þ-H2)49 of [Ru-
(dcypm)(hex)3(H)(OH)(tol)].
The intensities of these two signals increased strongly in the
spectra obtained after 130 and 360 min of heating, while the in-
tensities of other signals decreased, indicating that the former
correspond to oligomerization intermediates which predominate
once most of the amide has been consumed.
7445
dx.doi.org/10.1021/ja111389r |J. Am. Chem. Soc. 2011, 133, 7428–7449

Journal of the American Chemical Society
ARTICLE
’ CONCLUSIONS FROM THE EXPERIMENTAL STUDIES
secondary kinetic isotope effect caused by a change in hybridiza-
tion of the alkyne-C(1) carbon atom bond from sp to sp2.
The same experimental findings also rule out the redox-neutral
Mechanism E, which does not involve Ruꢀvinylidene intermedi-
ates, but instead proceeds via an attack of the amide nucleophile
to a π-coordinated alkyne. This pathway is in good agreement
with the isotope labeling and the ESIꢀMS studies, but offers no
explanation for the observation of RuꢀH-species after the ad-
dition of the amide to the catalyst system.
Mechanism B involves the formation of Ruꢀvinylidene species
via 1,2-proton shift followed by an attack of the amide nucleo-
phile. It must be excluded based on the results of the deuterium
labeling studies, which unambiguously showed that in contrast to
Ru-catalyzed additions of other nucleophiles, hydroamidations
do not involve a shift of the terminal alkyne proton to the internal
sp-carbon.
Overall, most of the mechanisms under consideration were in
disagreement with one or more experimental findings. Mecha-
nism A, which involves an oxidative addition of the amide followed
by an insertion of the alkyne into either the RuꢀH or the RuꢀN
bond, correctly predicted the findings of the deuterium labeling
experiments. It is also consistent with the detection of RuꢀH
species following addition of the amide to the Ru-catalyst and
with most species detected by ESIꢀMS. However, it must be
dismissed on the basis that a normal kinetic isotope effect was
observed, while Mechanism A would have predicted an inverse
Scheme 11. Selectivity-Determining Step of the Hydroami-
dation Reaction
Mechanism C, which involves the formation of Ruꢀvinyl
species and their rearrangement to Ruꢀhydrideꢀvinylidene inter-
mediates, is in agreement with most of the experimental findings.
It correctly predicts the results of the deuterium labeling experi-
ments and the observed normal kinetic isotope effect, and is in
good agreement with most of the species observed in the ESIꢀMS
studies. However, some of the cationic RuIV intermediates (28,
29 or 31) should have been easily detectable by ESIꢀMS. One
might also have expected the detection of RuꢀH-species in the
1H NMR in the presence of 1-hexyne (5b), rather than after the
addition of 2-pyrrolidinone (1a) to the catalyst system. More-
over, it is unlikely that a protonation step with formation of
cationic Ruꢀvinyl species is a favorable pathway in a nonpolar
solvent under almost neutral conditions. The catalytic cycle
involves only Ru-species in high oxidation states and offers no
Figure 14. Relative energies and optimized structures of potential hydroamidation intermediates.
7446
dx.doi.org/10.1021/ja111389r |J. Am. Chem. Soc. 2011, 133, 7428–7449

Journal of the American Chemical Society
ARTICLE
explanation for the detection of phosphonium salts known to
arise from reductive deallylation processes.
mixtures of Z-selective hydroamidations, we believe that analo-
gous species as those presented in Scheme 10 (58ꢀ63) are also
present. The ligands are dcypm instead of P(n-Bu)3 and DMAP
before. However, in each step in which a vacant coordination site
is required in the catalytic cycle (Scheme 10), one RuꢀP bond of
the bidentate ligand is cleaved, and therefore, the ligand remains
coordinated to the Ru center leading to sterically more demand-
ing Ru complexes. Hence, we conclude that the steric bulk of the
ligand sphere is the decisive factor in directing the orientation of
the substituent at the vinylidene moiety. As stretched in
Scheme 11, the attack of the amide would then lead to enamides
with different stereoselectivities depending on the preferred
orientation of the vinylidene moiety relative to the amide. More
in-depth studies are required to fully understand how the choice
of ligands affects the stereoselectivity.
Mechanism D is similar to Mechanism C in that it also involves
Ruꢀvinyl intermediates that rearrange to Ruꢀhydrideꢀvinylidene
complexes. It is thus also in full agreement with the deuteration
studies and correctly predicts the observed normal kinetic isotope
effect. In contrast to Mechanism C, it starts from a Ru0 species and
thus offers a good explanation for the detection of phosphonium
salts and of many Ru-species in the ESIꢀMS experiments. More-
over, the findings of the NMR studies are best explained by this
mechanism, which starts with the oxidative addition of amides to
the Ru-center with formation of RuꢀH species. This would
explain why after the addition of amides to the catalyst system, 1H
NMR signals below 0 ppm were detected, whereas these were not
observed when only the alkyne was added to the catalyst.
For these combined reasons, the possible catalytic pathways
can be narrowed down to a mechanism closely related to the pro-
posed Mechansim D. On the basis of the experimental evidence, we
believe that our model reaction, the E-selective hydroamidation
of 1-hexyne (5b) with 2-pyrrolidinone (1a), proceeds via the
mechanism depicted in Scheme 10.
The catalyst preformation proceeds via a reductive allylation
process with release of a phosphorus ylide (57) and isobutene
(43) and formation of a coordinatively unsaturated Ru0 species
bearing several neutral ligands. As the starting point for the depicted
catalytic cycle we choose complex 58 with two phosphine and
two DMAP ligands on the basis that intermediates with this com-
bination of neutral ligands showed particularly strong signals in
the in situ ESIꢀMS experiments. However, similar catalytic cycles
with any combination of DMAP, phosphines, and solvent molecules
would also be viable (MS signals at m/z = 751.5 [M þ H]þ).
Oxidative addition of the amide (1a) gives rise to an octahedral
Ruꢀhydride complex 59. This must be a slow step to be in
accordance with the KIE of 2.3 found during the kinetic studies.
The detection of signals at m/z = 751.5 and 836.5, which would
match the [M-(pyr)]þ and [M þ H]þ fragments of complex 59, is
in agreement with this pathway. In the next step, an alkyne (5b)
coordinates to this species with dissociation of one neutral ligand
(60). The alkyne then inserts into the RuꢀH bond and the open
coordination site created in the process is filled with a neutral
ligand, leading to RuII-vinyl intermediate 61. Signals at m/z =
833.5 and 918.6 can be explained by the presence of [M-(pyr)]þ
and [M þ H]þ fragments of species 61. The high intensity of the
signal at m/z = 833.5 along with the observed KIE when using
deuterated alkynes, indicates that this step is comparatively slow.
1,2-Hydride shift in RuIIꢀvinyl complex 61 then gives RuIVꢀ
Hꢀvinylidene species 62. Only weak ESIꢀMS signals could be
assigned to 62 (m/z = 631.3 [M-(pyr)]þ and 716.4 [M þ H]þ).
Because of the δþ polarization at C(1) of the vinylidene moiety,
62 is susceptible to attack of the amide ligand to afford species
63. A neutral ligand is likely to coordinate and to refill the empty
coordination site. The signals detected at m/z = 751.5 ([M-(en-
amide)]þ) and 918.6 ([M þ H]þ) could be assigned to inter-
mediate 63. Finally, reductive elimination releases the enamide 6a,
regenerating the original catalytic species 58 and closing the catalytic
cycle of the hydroamidation.
’ COMPUTATIONAL STUDIES
The Ru intermediates of the proposed catalytic cycle (Scheme 10,
Mechanism D) have been identified based on ESIꢀMS and to
some extent on NMR data. Since the information on the likely
configuration of these intermediates is still limited, we used DFT
calculations to look at the stability of each individual structure
and possible spatial arrangements of the ligands. All calculations
were performed using the Gaussian 03 or Gaussian 09 software
package,51 with B3LYP52/6-311þG(2d,p)53//B3LYP/6-31G-
(d)54 for H, C, N, O, P and Stuttgart RSC 1997 ECP55 for Ru.
A scaling factor for anharmonic corrections of vibrational fre-
quencies of f = 0.9804 was used.56
A stable minimum was found for every postulated intermedi-
ate within the catalytic cycle. The calculated structures along with
total energies and Gibbs free energies are depicted in Figure 14.
The hydrogen atoms of the phosphine and DMAP ligands are
omitted for clarity. Larger pictures of the optimized structures are
included in the Supporting Information. The oxidative addition
of 2-pyrrolidinone (1a) to the Ru species 58 with formation of the
Ruꢀhydride complex 59 was calculated to be exothermic by
ΔrEtot = ꢀ32.2 kcal molꢀ1 and exergonic by ΔrG298 = ꢀ16.7 kcal
molꢀ1. The exchange reaction of DMAP by 1-hexyne (5b)
to give a π-coordinated complex is almost thermoneutral
(ΔrEtot = ꢀ0.3 kcal molꢀ1, ΔrG298 = ꢀ0.6 kcal molꢀ1). The
insertion of the alkyne in the RuꢀH bond and refilling of the
empty coordination site with one additional DMAP is exothermic
and exergonic (ΔrEtot = ꢀ20.4 kcal molꢀ1, ΔrG298 = ꢀ2.7 kcal
molꢀ1). The following formation of vinylidene species 62 with
the concomitant release of one phosphine ligand is endothermic
(ΔrEtot = 10.6 kcal molꢀ1) but exergonic (ΔrG298 = ꢀ12.6 kcal
molꢀ1). The subsequent addition of the amide to the vinylidene
moiety and the concomitant coordination of a neutral ligand is
slightly endothermic (ΔrEtot = 3.3 kcal molꢀ1) and highly ender-
gonic (ΔrG298 = 24.2 kcal molꢀ1). The last reaction step, the
reductive elimination of the enamide product requires only a
small amount of total energy but releases a large amount of Gibbs
free energy (ΔrEtot = 5.1 kcal molꢀ1, ΔrG298 = ꢀ10.4 kcal molꢀ1).
Overall, the computational studies support the conclusions drawn
from the mechanistic studies. They confirm that the proposed
catalytic cycle involves stable intermediates with comparable
energies. Extensive computational studies using strongly simpli-
fied model systems are underway with the goal of calculating the
transition states and obtaining reliably predicted kinetic isotope
effects.
On the basis of the findings of the deuteration studies and the
kinetic investigations, we assume that this pathway is also valid
for related hydroamidations with other NꢀH nucleophiles. The
studies presented above did not reveal any fundamental mecha-
nistic differences between the E- and Z-selective protocols. On
the basis of the NMR and ESIꢀMS studies with reaction
7447
dx.doi.org/10.1021/ja111389r |J. Am. Chem. Soc. 2011, 133, 7428–7449

Journal of the American Chemical Society
ARTICLE
In summary, the results of our in-depth mechanistic studies of
the hydroamidation support a catalytic cycle with ruthenium
hydride and vinylidene species as the key intermediates. We thus
propose that the reaction proceeds via an oxidative addition of
the amide, followed by insertion of a π-coordinated alkyne into a
rutheniumꢀhydride bond, rearrangement to a vinylidene spe-
cies, nucleophilic attack of the amide, and finally reductive elimina-
tion of the product. This catalytic cycle is in agreement with all
experimental results and is supported by DFT calculations that
confirm the stability of all reaction intermediates.
(11) (a) van den Berg, M.; Minnaard, A. J.; Haak, R. M.; Leeman, M.;
Schudde, E. P.; Meetsma, A.; Feringa, B. L.; de Vries, H. M.; Maljaars,
E. P.; Wilians, C. E.; Hyett, D.; Boogers, A. F.; Hendricks, J. W.; de Vries,
J. G. Adv. Synth. Catal. 2003, 345, 308–323. (b) Blaser, H.-U.; Malan, C.;
Pugin, B.; Spindler, F.; Steiner, H.; Studer, M. Adv. Synth. Catal. 2003,
345, 103–151.
(12) (a) Dupau, P.; Le Gendre, P.; Bruneau, C.; Dixneuf, P. H.
Synlett 1999, 1832–1834. (b) Wang, X.; Porco, J. A., Jr. J. Org. Chem.
2001, 66, 8215–8221. (c) Bayer, A.; Maier, M. E. Tetrahedron 2004,
60, 6665–6677. (d) Adam, W.; Bosio, S. G.; Turro, N. J. J. Org. Chem.
2004, 69, 1704–1715. (e) Burk, M. J.; Casy, G.; Johnson, N. B. J. Org.
Chem. 1998, 63, 6084–6098.
(13) (a) Brettle, R.; Mosedale, A. J. J. Chem. Soc., Perkin Trans. 1
1988, 2185–2195. (b) Kuramochi, K.; Watanabe, H.; Kitahara, T. Synlett
2000, 397–399. (c) Sato, M. J. Org. Chem. 1961, 26, 770–779.
(14) (a) Ager, D. J. Synthesis 1984, 384–398. (b) F€urstner, A.;
Brehm, C.; Cancho-Grande, Y. Org. Lett. 2001, 3, 3955–3957.
(15) Krompiec, S.; Pigulla, M.; Kuꢀznik, N.; Krompiec, M.; Marciniec,
B.; Chadyniak, D.; Kasperczyk, J. J. Mol. Catal. A: Chem. 2005, 225, 91–101.
(16) (a) Wallace, D. J.; Klauber, D. J.; Chen, C.-Y.; Volante, R. P. Org.
Lett. 2003, 5, 4749–4752; (b) Jiang, L.; Job, G. E.; Klapars, A.; Buchwald,
S. L. Org. Lett. 2003, 5, 3667–3669; (c) Pan, X.; Cai, Q.; Ma, D. Org. Lett.
2004, 6, 1809–1812; (d) Brice, J. L.; Meerdink, J. E.; Stahl, S. S. Org. Lett.
2004, 6, 1845–1848; (e) Han, C.; Shen, R.; Su, S.; Porco, J. A., Jr. Org. Lett.
2004, 6, 27–30;(f)Tracey, M. R.;Hsung, R. P.;Antoline, J.;Kurtz, K. C. M.;
Shen, L.; Slafer, B. W.; Zhang, Y. Sci. Synth. 2005, 21, 387–475; (g) Klapars,
A.; Campos, K. R.; Chen, C.; Volante, R. P. Org. Lett. 2005, 7, 1185–1188;
(h) Bolshan, Y.; Batey, R. A. Angew. Chem. 2008, 120, 2139–2142; Angew.
Chem. Int. Ed. 2008, 47, 2109–2112.
(17) (a) Goossen, L. J.; D€ohring, A. Adv. Synth. Catal. 2003,
345, 943–947. (b) Goossen, L. J.; Paetzold, J.; Winkel, L. Synlett 2002,
10, 1721–1723. (c) Goossen, L. J.; Ghosh, K. Chem. Commun 2001,
20, 2084–2085. (d) Goossen, L. J.; Goossen, K.; Rodríguez, N.;
Blanchot, M.; Linder, C.; Zimmermann, B. Pure Appl. Chem. 2008,
80, 1725–1731.
(18) Heider, M.; Henkelmann, J.; Ruehl, T. EP 646571 1995 [Chem.
Abstr. 1995, 123, 229254].
(19) Kondo, T.; Tanaka, A.; Kotachi, S.; Watanabe, Y. J. Chem. Soc.
Chem. Commun 1995, 413–414.
’ ASSOCIATED CONTENT
S
Supporting Information. Experimental procedures and
b
full spectroscopic data of the deuterium-labeling, the in situ IR,
the in situ NMR, the in situ ESIꢀMS and the competition ex-
periments. This material is available free of charge via the Internet at
http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
goossen@chemie.uni-kl.de; gns@chemie.uni-kl.de
’ ACKNOWLEDGMENT
We thank the DFG and NanoKat and OPTIMAS for financial
support, Umicore for donating chemicals, Mettler Toledo for
giving us access to a ReactIR spectrometer, and the DAAD (K.S.
M.S) and Landesgraduiertenf€orderung Rheinland-Pfalz (M.A.
and A.F.) and Hans-B€ockler-Stiftung (F.M.) for scholarships,
and Dr. M. Blanchot for technical assistance. Part of this work
was performed in preparation of the new transregional colla-
borative research center SFB/TRR 88 3MET.
’ REFERENCES
(1) Yet, L. Chem. Rev. 2003, 103, 4283–4306.
(2) Sugie, Y.; Dekker, K. A.; Hirai, H.; Ichiba, T.; Ishiguro, M.;
Shiomi, Y.; Sugiura, A.; Brennan, L.; Duignan, J.; Huang, L. H.; Sutcliffe,
J.; Kojima, Y. J. Antibiot. 2001, 54, 1060–1065.
(3) (a) McDonald, L. A.; Swersey, J. C.; Ireland, C. M.; Carroll, A. R.;
Coll, J. C.; Bowden, B. F.; Fairchild, C. R.; Cornell, L. Tetrahedron 1995,
51, 5237–5244. (b) Boyd, M. R.; Farina, C.; Belfiore, P.; Gagliardi, S.;
Kim, J. W.; Hahakawa, Y.; Beutler, J. A.; Mckee, T. C.; Bowman, B. J.;
Bowman, E. J. J. Pharmacol. Exp. Ther. 2001, 297, 114–120.
(4) Davyt, D.; Entz, W.; Fernandez, R.; Mariezcurrena, R.; Mombrꢀu,
A. W.; Salda~na, J.; Domínguez, L.; Coll, J.; Manta, E. J. Nat. Prod. 1998,
61, 1560–1563.
(5) (a) Vidal, J.-P.; Escale, R.; Girard, J.-P.; Rossi, J.-C. J. Org. Chem.
1992, 57, 5857–5860. (b) Erickson, K. L.; Beutler, J. A.; Cardellina, J. H.;
Boyd, M. R. J. Org. Chem. 1997, 62, 8188–8192. (c) Jansen, R.;
Washausen, P.; Kunze, B.; Reichenbach, H.; H€ofle, G. Eur. J. Org. Chem.
1999, 1085–1089.
(20) (a) Tokunaga, M.; Wakatsuki, Y. Angew. Chem. 1998, 110,
3024–3027; Angew. Chem. Int. Ed. 1998, 37, 2867–2869; (b) Tokunaga,
M.; Suzuki, T.; Koga, N.; Fukushima, T.; Horiuchi, A.; Wakatsuki, Y.
J. Am. Chem. Soc. 2001, 123, 11917–11924; (c) Grotjahn, D. B.; Incarvito,
C. D.; Rheingold, A. L. Angew. Chem. 2001, 113, 4002–4005; Angew.
Chem. Int. Ed. 2001, 40, 3884–3887; (d) Chevallier, F.; Breit, B. Angew.
Chem. 2006, 118, 1629–1632; Angew. Chem. Int. Ed. 2006, 45,
1599–1602. (e) Labonne, A.; Kribber, T.; Hintermann, L. Org. Lett.
2006, 8, 5853–5856. (f) Hintermann, L.; Kribber, T.; Labonne, A.;
Paciok, E. Synlett 2009, 2412–2416.
(21) (a) Rotem, M.; Shvo, Y. Organometallics 1983, 2, 1689–1691.
(b) Mitsudo, T.; Hori, Y.; Yamakawa, Y.; Watanabe, Y. J. Org. Chem.
1987, 52, 2230–2239. (c) Ruppin, C.; Dixneuf, P. H. Tetrahedron Lett.
1986, 27, 6323–6324. (d) Philippot, K.; Devanne, D.; Dixneuf, P. H.
J. Chem. Soc. Chem. Commun. 1990, 1199–1200. (e) Neveux, M.; Seiller,
B.; Hagedorn, F.; Bruneau, C.; Dixneuf, P. H. J. Organomet. Chem. 1993,
451, 133–138. (f) Goossen, L. J.; Paetzold, J.; Koley, D. Chem. Commun.
2003, 706–707.
(6) Carbery, D. R. Org. Biomol. Chem. 2008, 6, 3455–3460.
(7) (a) Stevenson, P. J.; Graham, I. ARKIVOC 2003, 7, 139–144. (b)
Gaulon, C.; Dhal, R.; Chapin, T.; Maisonneuve, V.; Dujardin, G. J. Org.
Chem. 2004, 69, 4192–4202.
(8) Roff, G. J.; Lloyd, R. C.; Turner, N. J. J. Am. Chem. Soc. 2008,
126, 4098–4099.
(9) Willans, C. E.; Mulders, J. M. C. A.; de Vries, J. G.; de Vries,
A. H. M. J. Organomet. Chem. 2003, 687, 494–497.
(10) Matsubara, R.; Nakamura, Y.; Kobayashi, S. Angew. Chem.
2004, 116, 1711–1713; Angew. Chem. Int. Ed. 2004, 43, 1679–1681.
(22) (a) Uchimaru, Y. Chem. Commun 1999, 1133–1134; (b)
Tokunaga, M.; Eckert, M.; Wakatsuki, Y. Angew. Chem. 1999, 111,
3416–3419; Angew. Chem. Int. Ed. 1999, 38, 3222–3225. (c) Kondo, T.;
Okada, T.; Suzuki, T.; Mitsudo, T.-a. J. Organomet. Chem. 2001,
622, 149–154. (d) Fukumoto, Y.; Dohi, T.; Masaoka, H.; Chatani, N.;
Murai, S. Organometallics 2002, 21, 3845–3847. (e) Shimada, T.;
Yamamoto, Y. J. Am. Chem. Soc. 2003, 125, 6646–6647. (f) Yi, C. S.;
Yun, S. Y.; Guzei, I. A. J. Am. Chem. Soc. 2005, 127, 5782–5783. (g) Li, Y.;
Marks, T. J. Organometallics 1996, 15, 3770–3772. (h) Li, Y.; Marks, T. J.
J. Am. Chem. Soc. 1998, 120, 1757–1771.
7448
dx.doi.org/10.1021/ja111389r |J. Am. Chem. Soc. 2011, 133, 7428–7449

Journal of the American Chemical Society
ARTICLE
(23) Koelle, U.; Rietmann, C.; Tjoe, J.; Wagner, T.; Englert, U.
Organometallics 1995, 14, 703–704.
2000, 122, 8204–8214. (c) Frech, C. M.; Blacque, O.; Schmalle, H. W.;
Berke, H.; Adlhart, C.; Chen, P. Chem.—Eur. J. 2006, 12, 3325–3338.
(47) (a) Markert, C.; Pfaltz, A. Angew. Chem. 2004, 116, 2551–2554;
Angew. Chem. Int. Ed. 2004, 43, 2498–2500; (b) Markert, C.; Neuburger,
M.; Kulicke, K.; Meuwly, M.; Pfaltz, A. Angew. Chem. 2007, 119,
5996–5999; Angew. Chem. Int. Ed. 2007, 46, 5892–5895; (c) Markert,
C.; Rosel, P.; Pfaltz, A. J. Am. Chem. Soc. 2008, 130, 3234–3235; (d)
Teichert, A.; Pfaltz, A. Angew. Chem. 2008, 120, 3408–3410; Angew.
Chem. Int. Ed. 2008, 47, 3360–3362. (e) di Lena, F.; Matyjaszewski, K.
Chem. Commun. 2008, 6306–6308. (f) di Lena, F.; Matyjaszewski, K.
Dalton Trans. 2009, 8884–8890.
(48) Kuran, W.; Musco, A. Inorg. Chim. Act. 1975, 12, 187–193.
(49) Reinhardt, B. M.; Niedner-Schatteburg, G. J. Phys. Chem. A
2002, 106, 7988–7992.
(50) Six, C.; Gabor, B.; G€orls, H.; Mynott, R.; Philipps, P.; Leitner,
W. Organometallics 1999, 18, 3316–3326.
(51) (a) Gaussian 03, Revision E.01, Gaussian, Inc.: Wallingford
CT, 2004. (b) Gaussian 09, Revision A.02, Gaussian, Inc.: Wallingford
CT, 2009; for full citations see the Supporting Information.
(52) (a) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988,
37, 785–789. (b) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652.
(c) Stephens, P. J.; Devlin, J. F.; Chabalowski, C. F.; Frisch, M. J. J. Phys.
Chem. 1994, 98, 11623–11627.
(24) (a) Gemel, C.; Trimmel, G.; Slugovc, C.; Kremel, S.; Mereiter,
K.; Schmid, R.; Kirchner, K. Organometallics 1996, 15, 3998–4004. (b)
Varela-Fernꢀandez, A.; Gonzꢀalez-Rodríguez, C.; Varela, J. A.; Castedo, L.;
Saꢀa, C. Org. Lett. 2009, 11, 5350–5353. (c) Liu, P. N.; Su, F. H.; Wen,
T. B.; Sung, H. H.-Y. Chem.—Eur. J. 2010, 16, 7889–7897.
(25) Goossen, L. J.; Rauhaus, J. E.; Deng, G. Angew. Chem. 2005,
117, 4110–4113; Angew. Chem. Int. Ed. 2005, 44, 4042–4045.
(26) Goossen, L. J.; Arndt, M.; Blanchot, M.; Rudolphi, F.; Menges,
F.; Niedner-Schatteburg, G. Adv. Synth. Catal. 2008, 350, 2701–2707.
(27) Goossen, L. J.; Blanchot, M.; Salih, K. S. M.; Karch, R.; Rivas-Nass,
A. Org. Lett. 2008, 10, 4497–4499.
(28) Goossen, L. J.; Blanchot, M.; Brinkmann, C.; Goossen, K.;
Karch, R.; Rivas-Nass, A. J. Org. Chem. 2006, 71, 9506–9509.
(29) (a) Goossen, L. J.; Salih, K. S. M.; Blanchot, M. Angew. Chem.
2008, 120, 8620–8623; Angew. Chem. Int. Ed. 2008, 47, 8492–8495. (b)
Goossen, L. J.; Blanchot, M.; Salih, K. S. M.; Goossen, K. Synthesis
2009, 2283–2282.
(30) Buba, A. E.; Arndt, M.; Goossen, L. J. J. Organomet. Chem. 2010,
696, 170–178.
(31) Goossen, L. J.; Blanchot, M.; Arndt, M.; Salih, K. S. M. Synlett
2010, 1685–1687.
(32) (a) Bruneau, C.; Dixneuf, P. H. Angew. Chem. 2006, 118,
2232–2260; Angew. Chem. Int. Ed. 2006, 45, 2176–2203. (b) Rigaut, S.;
Touchard, D.; Dixneuf, P. H. Coord. Chem. Rev. 2004, 248, 1585–1601.
(33) Olivꢀan, M.; Clot, E.; Eisenstein, O.; Caulton, K. G. Organome-
tallics 1998, 17, 3091–3100.
(34) DFT calculations confirmed that the reaction of 2-pyrrolidi-
none (1a) and deprotonated N-((E)-hex-1-enyl)pyrrolidin-2-one, lead-
ing to the 2-pyrrolidinyl anion and the corresponding enamide (6a), is
(53) Krishnan, R.; Binkley, J. S.; Seeger, R.; Pople, J. A. J. Chem. Phys.
1980, 72, 650–654.
(54) Hariharan, P. C.; Pople, J. A. Theor. Chim. Acta 1973, 28,
213–222.
(55) Andrae, D.; H€aussermann, U.; Dolg, M.; Stoll, H.; Preuss, H.
Theor. Chim. Acta 1990, 77, 123–141.
(56) Wong, M. W. Chem. Phys. Lett. 1996, 256, 391–399.
both exotherm (ΔrEtot = ꢀ30.88 kcal molꢀ1) and exergonic (ΔrG298
=
ꢀ30.02 kcal molꢀ1).
(35) Although there is no experimental proof in literature that the
rehybridisation from sp to sp2 causes an inverse isotope effect, it must be
expected based on the analogy to sp2/sp3 rehybridizations. See for
example: Carey, F. A.; Sundberg, R. J. Advanced Organic Chemistry, Third
edition; Plenum Press: New York, London, 1990; pp 216ꢀ218. The
same assumption was made, e.g., in: Ipaktschi, J.; Mohsseni-Ala, J.; Uhlig,
S. Eur. J. Inorg. Chem. 2003, 4313ꢀ4320.
(36) Maegawa, T.; Fujiwara, Y.; Inagaki, Y.; Monguchi, Y.; Sajiki, H.
Adv. Synth. Catal. 2008, 350, 2215–2218.
(37) Rigaut, S.; Perruchon, J.; Guesmi, S.; Fave, C.; Touchard, D.;
Dixneuf, P. H. Eur. J. Inorg. Chem. 2005, 447–460.
(38) Ryabov, A. D. Chem. Rev. 1990, 90, 403–424.
(39) Ciardi, C.; Reginato, G.; Gonsalvi, L.; de los Rios, I.; Romerosa,
A.; Peruzzini, M. Organometallics 2004, 23, 2020–2026.
(40) Choe, J.-I.; Choi, H.-S.; Kuczkowski, R. L. Magn. Reson. Chem.
1986, 24, 1044–1047.
(41) Touchard, D.; Haquette, P.; Pirio, N.; Toupet, L.; Dixneuf,
P. H. Organometallics 1993, 12, 3132–3139.
(42) (a) Bruneau, C.; Dixneuf, P. H. Acc. Chem. Res. 1999,
32, 311–323. (b) Vijayaraj, T. A.; Sundararajan, G. J. Mol. Catal. A
1995, 99, 47–54.
(43) Caballero, A.; Jalꢀon, F. A.; Manzano, B. R. Chem. Commun
1998, 1879–1880.
(44) (a) Guerchais, V.; Lapinte, C.; Thepot, J. Y.; Toupet, L.
Organometallics 1988, 7, 604–612. (b) Maurer, J.; Linseis, M.; Sarkar,
B.; Schwederski, B.; Niemeyer, M.; Kaim, W.; Zli, S.; Anson, C.; Zabel,
M.; Winter, R. F. J. Am. Chem. Soc. 2008, 130, 259–268. (c) Jung, S.; Ilg,
K.; Brandt, C. D.; Wolf, J.; Werner, H. Eur. J. Inorg. Chem.
2004, 469–480. (d) Bassetti, M.; Cadierno, V.; Gimeno, J.; Pasquini,
C. Organometallics 2008, 27, 5009–5016.
(45) Cabeza, J. A.; Riera, V. J. Organomet. Chem. 1989, 376, C23–C25.
(46) (a) Hinderling, C.; Adlhart, C.; Chen, P. Angew. Chem. 1998,
110, 2831–2835; Angew. Chem. Int. Ed. 1998, 37, 2685–2689. (b)
Adlhart, C.; Hinderling, C.; Baumann, H.; Chen, P. J. Am. Chem. Soc.
7449
dx.doi.org/10.1021/ja111389r |J. Am. Chem. Soc. 2011, 133, 7428–7449