Mechanisms of syn -insertion of alkynes and allenes into gold-silicon bonds: A comprehensive experimental/theoretical study

Article abstract of DOI:10.1021/ja504024h

A detailed mechanistic study is reported for the syn-insertion of alkynes and allenes in the Au-Si bonds of complexes (R₃P)Au-SiR′Ph ₂ (R = Ph, Me and R′ = t-Bu, Ph). Kinetic experiments indicate that (i) the reaction is first-order in alkyne and gold silyl complex and (ii) it requires a rather low enthalpy of activation and a relatively large negative entropy of activation [ΔH^? = 13.7 (±1.6) kcal·mol^-1 and ΔS^? = -32.0 (±5.0) cal·mol^-1·K^-1 for the reaction of (Ph₃P)Au-Si(t-Bu)Ph₂ with methyl propiolate], in line with a bimolecular associative transformation. The different mechanistic pathways have been explored by DFT calculations. Accordingly, the reaction is found to proceed via a two-step inner-sphere mechanism: (i) first, the alkyne coordinates to the gold silyl complex to form a π-complex; (ii) the subsequent migratory insertion step is rate determining and occurs in a concerted manner. Provided dispersion effects are taken into account (B97D functional), the enthalpy of activation estimated theoretically [ΔH ^? = 11.5 kcal·mol^-1] is in good agreement with that measured experimentally. The influence of the π substrate (methyl propiolate, dimethyl acetylene dicarboxylate, phenyl acetylene, ethyl 2,3-butadienoate) has been analyzed theoretically, and the regioselectivity of the insertion has been rationalized. In particular, the unexpected selectivity observed experimentally with the allene is shown to result from the insertion of the terminal nonactivated C=C double bond into the Au-Si bond of (Ph ₃P)Au-SiPh₃, followed by an original isomerization (Au/Si exchange process). This study provides unambiguous evidence for coordination-insertion at gold and thereby strongly supports the possible occurrence of inner-sphere mechanisms during the functionalization of alkynes and allenes.

Full text of DOI:10.1021/ja504024h

Article
pubs.acs.org/JACS
Mechanisms of syn-Insertion of Alkynes and Allenes into Gold−
Silicon Bonds: A Comprehensive Experimental/Theoretical Study
Maximilian Joost,†,‡,⊥ Laura Estevez,§,∥,⊥ Sonia Mallet-Ladeira,⊥ Karinne Miqueu,*,§
Abderrahmane Amgoune,*,†,‡ and Didier Bourissou*,†,‡
†Universite
France
́ ́ ́ ́
de Toulouse, UPS, Laboratoire Heterochimie Fondamentale Appliquee, 118 route de Narbonne, F-31062 Toulouse,
‡CNRS, LHFA UMR 5069, F-31062 Toulouse, France
§Institut Pluridisciplinaire de Recherche sur l′Environnement et les Mater
Pau et des Pays de l′Adour, Helioparc, 2 Avenue du President Angot, 64053 Pau cedex 09, France
∥Departamento de Química Física, Universidade de Vigo, Facultade de Química Lagoas-Marcosende s/n, 36310 Vigo, Galicia, Spain
⊥Universite
de Toulouse, UPS, Structure Federative Toulousaine en Chimie Moleculaire, FR2599, 118 Route de Narbonne, F-31062
́ ́
iaux (UMR 5254), Equipe Chimie Physique, Universite de
́
́
́
́
́
́
Toulouse, France
S
* Supporting Information
ABSTRACT: A detailed mechanistic study is reported for the
syn-insertion of alkynes and allenes in the Au−Si bonds of
complexes (R3P)Au−SiR′Ph2 (R = Ph, Me and R′ = t-Bu, Ph).
Kinetic experiments indicate that (i) the reaction is first-order
in alkyne and gold silyl complex and (ii) it requires a rather
low enthalpy of activation and a relatively large negative
entropy of activation [ΔH‡ = 13.7 ( 1.6) kcal·mol−1 and ΔS‡
= −32.0 ( 5.0) cal·mol−1·K−1 for the reaction of (Ph3P)Au−
Si(t-Bu)Ph2 with methyl propiolate], in line with a bimolecular associative transformation. The different mechanistic pathways
have been explored by DFT calculations. Accordingly, the reaction is found to proceed via a two-step inner-sphere mechanism:
(i) first, the alkyne coordinates to the gold silyl complex to form a π-complex; (ii) the subsequent migratory insertion step is rate
determining and occurs in a concerted manner. Provided dispersion effects are taken into account (B97D functional), the
enthalpy of activation estimated theoretically [ΔH‡ = 11.5 kcal·mol−1] is in good agreement with that measured experimentally.
The influence of the π substrate (methyl propiolate, dimethyl acetylene dicarboxylate, phenyl acetylene, ethyl 2,3-butadienoate)
has been analyzed theoretically, and the regioselectivity of the insertion has been rationalized. In particular, the unexpected
selectivity observed experimentally with the allene is shown to result from the insertion of the terminal nonactivated CC
double bond into the Au−Si bond of (Ph3P)Au−SiPh3, followed by an original isomerization (Au/Si exchange process). This
study provides unambiguous evidence for coordination−insertion at gold and thereby strongly supports the possible occurrence
of inner-sphere mechanisms during the functionalization of alkynes and allenes.
to doubly activate terminal alkynes and form (σ,π) dinuclear
complexes has also been recently substantiated, which opens
new synthetic avenues.4 In a second step, a nucleophile adds to
the π-complex anti to gold, affording a vinyl gold complex with
trans arrangement of the incoming nucleophile and gold center
(Figure 1). The vinyl gold complex can eventually react further
or directly liberate the functionalized substrate, most commonly
via protodeauration.
Another pathway may also be envisioned for the reaction of
nucleophiles on π-systems upon gold activation. Instead of anti
nucleophilic attack on a π-complex, it involves coordination of
the nucleophile to gold and then syn-insertion of the π-
substrate into the Au−Nu bond. This inner-sphere mechanism
leads to cis vinyl gold complexes.
INTRODUCTION
■
Due to their excellent carbophilic Lewis acidity combined with
the unique functional group tolerance, gold complexes are very
powerful catalysts for the activation of CC multiple bonds
toward nucleophilic attack. Over the past 15 years, a broad
variety of valuable catalytic transformations have been
developed on this basis.1 In parallel, a thorough understanding
of the reactivity of gold complexes toward π-systems has been
gained thanks to experimental studies and high-level computa-
tional analyses.2
The addition of nucleophiles to alkynes, alkenes, and allenes
catalyzed by gold typically proceeds in a stepwise manner. The
first step is the activation of the π-substrate upon side-on
coordination to a cationic gold center. Such π-complexes of
gold have been extensively investigated, and the data collected
experimentally and computationally provide a detailed picture
of their structure and properties.3 Note that the ability of gold
Received: April 22, 2014
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja504024h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
RESULTS AND DISCUSSION
■
Kinetic Studies of the Reaction between the Gold Silyl
Complex 1 and Methyl Propiolate. Treatment of the gold
silyl complex 1 with methyl propiolate (MP) gives the vinyl
gold complex 2 with complete stereo- and regioselectivity.
Kinetic experiments were performed in NMR tubes (in C6D6,
using 10 equiv of MP per gold complex). 1H NMR monitoring
shows the progressive formation of complex 2 without a
detectable intermediate (Scheme 1). The respective concen-
trations of the reactants and product were followed over time
by integrating the resonance signals corresponding to the
OCH3 moiety of MP (δ = 3.12 ppm), the t-Bu group of
complex 1 (δ = 1.50 ppm), and the vinylic proton of complex 2
(δ = 8.40 ppm). According to kinetic plots (a) and (b), the
consumption of the gold silyl complex 1 follows a first-order
decay and the reaction is overall second-order (see Supporting
Information for details). Thus, the reaction is first-order with
respect to each reactant (rate law: v = kobs[1][MP]), in line
with a bimolecular process.19
Figure 1. Formation of vinyl gold complexes from alkynes: schematic
representation of the outer- and inner-sphere mechanisms.
The dichotomy between outer/inner-sphere mechanisms has
been frequently discussed in gold catalysis, in particular in the
frame of hydration,5 alkoxylation,6 and amination7 reactions.8
The outer-sphere mechanism has been supported by several
experimental observations and computational studies.9−13
Conversely, no direct unambiguous evidence for the inner-
sphere mechanism has been reported to date,14 and in fact, the
ability of gold complexes to lead to syn insertion products has
only been recently substantiated.15,16
Last year, we reported that gold(I) silyl complexes readily
react with unsaturated substrates, leading to isolable β-silyl vinyl
gold complexes.17,18 Terminal and internal alkynes, as well as
allenes were shown to insert into Au−Si bonds in a stereo- and
regioselective manner. This unprecedented reactivity for gold
prompted us to carry out a detailed experimental/computa-
tional mechanistic study that we report hereafter. Kinetic
experiments using multinuclear NMR monitoring have been
performed to determine the rate law and activation parameters.
The various possible mechanistic pathways have been explored
computationally (DFT calculations using different density
functionals), and the influence of the substrate (methyl
propiolate, dimethyl acetylene dicarboxylate, phenyl acetylene,
ethyl 2,3-butadienoate) has been analyzed. These studies
provide a detailed picture of the reaction mechanism and
substantiate a two-step inner-sphere pathway involving
coordination of the π-system followed by syn insertion into
the Au−Si bond. They also point out the involvement of an
original isomerization process in the case of ethyl 2,3-
butadienoate which explains the unexpected regioselectivity
observed experimentally.
The rate constant kobs was determined at different temper-
atures (25 to 65 °C, 10 °C steps), and the activation parameters
were extracted from the corresponding Eyring plot. Accord-
ingly, the insertion reaction was found to require a rather low
enthalpy of activation, ΔH‡ = 13.7 ( 1.6) kcal·mol−1. The
entropy of activation is negative and relatively large, ΔS‡ =
−32.0 ( 5.0) cal/(mol·K), in agreement with an associative
mechanism. Overall, the enthalpic and entropic terms
contribute about equally to the Gibbs free energy of activation,
ΔG‡ = 24.4 ( 2.3) kcal·mol−1 at 60 °C.
Reactions of alkynes with transition metal complexes
eventually display secondary kinetic isotope effects (KIE) as a
2
consequence of the change in hybridization from Csp to Csp in
the rate-determining step.20 Thus, methylpropiolate H/D
labeling studies have been carried out and 1H NMR
spectroscopy indicated a negligible kinetic isotope effect [kH/
kD = 1.04 ( 0.06)] in this case (see Supporting Information for
details).
Computational Study of the Reaction between the
Gold Silyl Complex 1 and Methyl Propiolate. Two
plausible pathways may be considered to explain the formation
of the syn insertion complex 2: (i) a one-step 1,2 addition of the
gold silyl complex 1 to the CC triple bond of MP or (ii) a two-
a
1
Scheme 1. Reaction of Complex 1 with Methylpropiolate Monitored by H NMR Spectroscopy
a
C6D6, [1]0 = 0.0483 mol/L, [MP]0 = 0.483 mol/L. (a) Plot of ln[1] vs time (s). (b) Plot of ln([MP]/[1]) vs time (s). (c) Plot of ln(kobs/T) vs 1/T
(K−1) (temperature range: 25−65 °C).
B
dx.doi.org/10.1021/ja504024h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
step pathway involving coordination of the alkyne to gold
followed by a migratory insertion process. Both mechanisms
were investigated by DFT calculations (B97D density func-
tional)21 on the actual reaction, i.e. without simplification of the
reactants.
1,2 Addition Pathway. The potential energy surface (PES)
was scrutinized to localize a transition state (TS) accounting for
a direct 1,2 addition of the Au−Si bond to the alkyne.
Optimizations from diverse starting geometries (Au−C1 and
Si−C2 bond distances ranging from 1.90 to 2.35 Å) did not
allow the localization of the expected in-plane (Au,Si,C1,C2)
four-center transition state structure. In fact, the only way to
find a transition state (TS) corresponding to the addition of the
Au−Si bond to MP was by decoordinating the PPh3 ligand
from gold (Scheme 2). Phosphine dissociation is not very
Scheme 2. 1,2 Addition of the Au−Si Bond to Methyl
Propiolate (with and without PPh3 Dissociation)
Figure 2. Reaction profile for the insertion of methyl propiolate into
the Au−Si bond of the gold silyl complex 1 (ΔG in kcal·mol−1),
computed at the B97D/SDD+f(Au)-6-31G** level of theory.
Enthalpy of activation ΔH‡ (in kcal·mol−1) in parentheses.
emphasize the importance of noncovalent interactions in such
systems which are relatively large (with 43 non-hydrogen
atoms) and feature main group elements (P,Si) as well as a gold
costly in energy (it is endergonic by 22.0 kcal·mol−1), but the
ensuing transition state for 1,2 addition is prohibitively high in
energy (ΔG‡ = 64.6 kcal·mol−1, ΔH‡ = 63.0 kcal·mol−1). It is
thus very unlikely that complex 2 is formed by 1,2 addition of
the Au−Si bond to MP (Figure S19).
center. Note that the value computed for ΔG‡
is
1→2
systematically overestimated, the best match being found with
the B97D functional (predicted: 30.1 kcal·mol−1, experimen-
tally determined: 23.2 kcal·mol−1 at 298 K). These deviations
arise from the difficulty in estimating accurately the entropic
contribution, which is especially challenging for reactions
involving a change in molecularity.23 Overall, the B97D
functional offers the best comprise in terms of accuracy vs
computing time, and it was thus preferred for the following
calculations.
Coordination−Insertion Pathway. While no in-plane TS
could be localized on the PES for the reaction of 1 with MP, an
out-of-plane four-center transition state TS1→2 associated with
the insertion of the alkyne into the Au−Si bond of 1 was
identified. Intrinsic Reaction Coordinate (IRC) calculations
indicate that TS1→2 is connected to the vinyl gold complex 2
and to a π-adduct π1→2 between MP and 1 (Figure 2). The
π1→2 intermediate species is slightly higher in energy than the
free reactants (ΔG = 5.2 kcal·mol−1), and its formation requires
a low activation barrier (ΔG‡ = 8.3 kcal·mol−1). Altogether, this
leads to a two-step coordination−insertion mechanism for the
formation of the vinyl gold complex 2. The reaction is
thermodynamically favored (overall Gibbs energy ΔG = −31.4
kcal·mol −1), and the rate-determining step is the insertion of
The structures of the key stationary points were then
examined thoroughly, starting from the intermediate π-complex
(Figure 3). The optimized geometry of π1→2 indicates that
methyl propiolate binds to gold in a side-on fashion and lies in
the same coordination plane as phosphorus and silicon. Upon
coordination, the CC triple bond elongates by 0.07 Å and the
P−Au−Si framework bends significantly (104.7°). Also
interesting to note is the nonsymmetric coordination of the
alkyne, the Au−C1 distance being substantially shorter than the
Au−C2 distance (2.108 Å vs 2.267 Å). To shed light on the
bonding situation, the molecular orbitals of π1→2 were analyzed.
Accordingly, the main interaction between 1 and MP (Figure
S20) involves the HOMO of the gold silyl complex (a
combination of σAuSi and dyz(Au) orbitals) and the LUMO of
methyl propiolate (π*CC). NPA charges calculated from
Natural Bond Order analysis (NBO) also revealed a net charge
transfer of 0.47 e− from the gold silyl complex 1 to methyl
propiolate (Table S6). The bonding situation found in π1→2
differs from that observed upon coordination of alkynes to
(L)Au+ fragments.3 For example, upon coordination of MP to
the alkyne into the Au−Si bond (ΔG‡
= 30.1 kcal·mol−1).
1→2
At this stage, several density functionals (B97D, M06,
B3PW91) were evaluated in order to determine the most
appropriate method (Table S4). With the B3PW91 functional,
the value predicted for ΔH‡1→2 is overestimated by 10 kcal·mol
−1
compared to that determined experimentally (13.7 kcal·mol
−1). Significantly better results were obtained with B97D and
M06, two density functionals taking into account dispersion
22
interactions. In this case, the values computed for ΔH‡
1→2
(11.5 and 14.2 kcal·mol−1, respectively) deviate by less than 2.5
1
kcal·mol−1 from that estimated by H NMR. These results
C
dx.doi.org/10.1021/ja504024h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 3. (Top) Intermediate π1→2 and transition state TS1→2 computed for the insertion of methyl propiolate into the Au−Si bond of 1. Selected
bond distances (Å) and angles (deg): π1→2: AuC1, 2.108; AuC2, 2.267; C1C2, 1.280; AuSi, 2.405; AuP, 2.395; PAuSi, 104.7; TS1→2: AuC1, 2.125;
AuC2, 2.318; C1C2, 1.323; AuSi, 2.587; AuP, 2.273; C1Si, 2.357; PAuSi, 130.8. (Bottom) Plots of the HOMO of intermediate π1→2 and
corresponding fragment orbitals ( 0.04 au isosurfaces). B97D/SDD+f(Au)-6-31G** level of theory.
(Ph3P)Au+, there is a net charge transfer of 0.12 e− from the
alkyne to the metal fragment, and the main orbital interaction
involves the HOMO of MP (πCC) and the LUMO of
(Ph3P)Au+ (σ*PAu) (Figures S20 and S21).
computational investigations were performed with complex 3,
featuring PMe3 instead of PPh3 at gold. The insertion reaction
of MP proceeds faster with 3 than with 1 (1 h vs 8 h at 60 °C)
and is again fully stereo- and regioselective, affording the β-silyl
vinyl gold complex 4 as the sole product (Scheme 3).
The structure of the transition state TS1→2 corresponding to
the insertion step was then studied. Starting from π1→2, the
reaction involves migration of the silyl group from Au to C1
with a concomitant shift of gold toward C2, resulting overall in
the syn insertion of MP into the Au−Si bond. The most
important geometrical modifications from π1→2 to TS1→2 are
(i) the displacement of the silyl group out of the (Au,C1,C2)
plane (C1C2AuSi = −34.5° in TS1→2 vs −7.0° in π1→2) and (ii)
the widening of the P−Au−Si bond angle (130.8°), which
altogether bring the Si atom close to C1 (SiC1 = 2.357 Å). It is
also interesting to note that, from an electronic point of view,
the charge transfer from 1 to MP increases further by 0.33 au in
Scheme 3. Syn Insertion of Methylpropiolate into Au−Si
Bond of Complex 3
1
The kinetic data obtained by H NMR monitoring of the
reaction between 3 and MP (5 equiv) showed here also a first-
order dependence in each of the reactants (see Supporting
Information). An Eyring plot was obtained from the temper-
ature dependence of the rate constant (15−55 °C), and the
following activation parameters were extracted: ΔH‡ = 14.5
( 0.7) kcal·mol −1 (vs 13.7 kcal·mol−1 for complex 1) and ΔS‡
= −21.4 ( 2.3) cal/(mol·K) (vs −32.0 cal/(mol·K) for
complex 1), corresponding to a Gibbs free energy of activation
ΔG‡ of 21.6 kcal·mol−1 at 60 °C (that is ∼3 kcal·mol−1 lower
than that found for 1, ΔG‡ = 24.4 kcal·mol−1). The main
difference comes from the entropic term which is significantly
lower with complex 3. According to DFT calculations, the
reaction proceeds by coordination−insertion and the reaction
profile computed for 3 is similar to that found for complex 1
(Figure S24). The enthalpy of activation predicted by DFT
(ΔH‡3→4 = 13.3 kcal·mol−1 with the B97D functional) fits with
that determined experimentally (Table S5). Moreover, the
intermediate π-complex π3→4 and transition state TS3→4 for the
insertion step display similar structures to those for π1→2 and
TS1→2. The prevailing orbital interaction arises from the
HOMO of the gold silyl complex (σAuSi/dyz(Au) combination)
and the LUMO of methyl propiolate (π*CC), and it is
TS1→2
.
Note also that the C1 atom already develops sp2 character in
π1→2 (the AuC1 bond is short at 2.125 Å and the C1C2H bond
angle is significantly bent at 146°) and thus does not experience
a significant change in hybridization in the rate-determining
step. This is consistent with the negligible KIE observed
experimentally.
The formation of the other conceivable regioisomer of the
vinyl gold complex 2′ has also been investigated computation-
ally for comparison (syn insertion of MP in the Au−Si bond,
with the silyl group introduced in the α instead of β position of
the CO2Me substituent). The most favorable pathway to form
2′ is again a two-step coordination−insertion process (Figure
S22). The transition state TS1→2′ for the rate-determining step
stands 6−7 kcal·mol−1 higher in energy than TS1→2 (both in H
and G), and the regioisomer 2′ was found to be less stable than
2 by 4 kcal·mol−1 (ΔG = −27.6 kcal·mol−1), in agreement with
the selective formation of 2 as observed experimentally.
Reaction between the Gold Silyl Complex 3 and
Methyl Propiolate. In order to confirm the conclusions
drawn from the reaction of 1 with MP, similar kinetic and
D
dx.doi.org/10.1021/ja504024h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Scheme 4. Competitive Insertion of Dimethyl Acetylene Dicarboxylate in the Au−Si and Au−P Bonds of the Phosphine Gold
Silyl Complex 1
associated with a substantial charge transfer from 3 to MP (by
0.54e− au in π3→4 according to NBO charges).
5a). According to 31P NMR, the two species are formed in
approximately equal amounts. After workup, the compounds
were isolated in pure form and their structures were
unambiguously ascertained by spectroscopic and crystallo-
graphic means (Figure 4). The 31P and 29Si NMR data of
Thus, the reaction of complex 3 with MP corroborates the
results obtained with 1 and the two-step coordination−
insertion process seems decidedly favored over other
conceivable mechanistic pathways. The gold silyl complexes 1
and 3 first react with methyl propiolate to give π-complexes.
The subsequent insertion step is rate determining and occurs in
a concerted manner (migration of the silyl group from Au to
C1 with a concomitant shift of Au toward C2, resulting in syn
insertion of MP in the Au−Si bond). The regioselectivity of the
reaction is well-reproduced theoretically and follows the
polarity of the alkyne, with a preference for the silyl group to
attack MP at the terminal electrophilic carbon atom (Michael-
type addition).
The reaction between (Ph3P)AuMe and methylpropiolate
(MP) was investigated for comparison. Computationally, the
reaction was predicted to also proceed via a two-step inner-
sphere mechanism (Figure S25). It is thermodynamically
favored (ΔG = −36.8 kcal·mol−1), but the rate determining
insertion step requires a significantly higher activation barrier
(ΔG‡ = 40.3 kcal·mol−1, ∼10 kcal·mol−1 higher than the one
found for the gold-silyl complex 1). Analysis of the molecular
orbitals revealed that the main interaction between the
(Ph3P)AuMe complex and MP in the π-complex and in the
TS also involves the HOMO of the gold complex (σAuC orbital)
and the LUMO of the alkyne (π*CC orbital) (Figure S26).
The HOMO of the Au−Me complex is lower in energy than
that of the gold silyl complex 1 (by 0.4 ev), and thus the
HOMO−LUMO gap increases. Accordingly, the charge
transfer from the Au−Me complex to MP at the TS is
significantly lower than that computed for the Au−Si complex
(0.80 e− for Au−Si vs. 0.53 e− for Au−Me). In line with
computational studies, no reaction was observed experimentally
between the (Ph3P)AuMe complex and MP under the same
conditions compared to the silyl complexes (60 °C, several
hours). All attempts to carry out the reaction at higher
temperatures or during prolonged reaction times resulted only
in the decomposition of the (Ph3P)AuMe complex.
Influence of the Alkyne on the Reaction of Gold Silyl
Complexes. In order to probe the influence of the alkyne on
reactivity and selectivity, we then investigated the reaction of
complex 1 with a more activated alkyne than MP, namely
dimethyl acetylenedicarboxylate (DMAD), and a less activated
alkyne than MP, namely phenyl acetylene (PA).
Reaction of the Phosphine Gold Silyl Complex 1 with
Dimethyl Acetylene Carboxylate. Complex 1 readily reacts
with DMAD, and complete conversion is achieved within 1−2
h at 60 °C using only 1 equiv of the alkyne (Scheme 4). As
expected, the reaction yields the vinyl gold complex 5a as the
result of syn insertion of DMAD in the Au−Si bond of 1.
However, 31P NMR monitoring indicates the concomitant
formation of a second complex 5b exhibiting a resonance signal
at significantly higher field (δ = 16.1 ppm for 5b vs 40 ppm for
Figure 4. Molecular views of complexes 5a (left) and 5b (right) in the
solid state (thermal ellipsoids set at 50% probability; hydrogen atoms
and solvent molecules are omitted for clarity).
complex 5b are very different from those of the vinyl gold
complex 5a. Interestingly, complex 5b displays a 29Si NMR
resonance signal at δ = 31.4 ppm (vs −7.7 ppm for complex
5b), very close to that of the starting gold silyl complex 1 (δ =
35.4 ppm), suggesting that the Au−Si bond is retained in 5b. X-
ray diffraction analysis revealed a zwitterionic structure for 5b
that results from the syn insertion of DMAD in the Au−P bond
(instead of Au−Si bond) of 1. Interestingly, 5b slowly converts
to 5a upon prolonged heating (60 °C, overnight), indicating
that 5a is thermodynamically more stable than 5b.
According to calculations, direct 1,2-addition and coordina-
tion−insertion are unlikely to account for the insertion of
DMAD into the Au−P bond of 1.24 Instead, the reaction profile
leading to 5b was found to involve the dissociation of PPh3
(ΔGdiss = 22.0 kcal·mol−1) from complex 1 (facilitated by the
strong trans influence of the silyl ligand),25 followed by Michael
addition of PPh3 to DMAD to form a phosphonium enolate
species.26 Finally, nucleophilic addition of the vinyl anion to the
gold silyl fragment affords complex 5b. Both, the addition of
Ph3P to DMAD and the nucleophilic addition of the vinyl
anion to gold silyl are barrierless processes (see Supporting
Information for details, Figure S27).
The reaction profile leading to complex 5a was explored in
detail, and as for methyl propiolate, the syn insertion of
dimethyl acetylene dicarboxylate in the Au−Si bond of 1
proceeds via a two-step coordination−insertion pathway
(Figure S28).27 The formation of the intermediate π-complex
π1→5a is slightly endergonic (ΔG = 3.3 kcal·mol−1), similarly to
that found for methyl propiolate (ΔG = 5.2 kcal·mol−1). But
the transition state TS1→5a corresponding to the rate-
determining insertion step is readily accessible in energy, and
the activation barrier for the syn insertion of DMAD is
estimated to ΔH‡
= 3.2 kcal·mol−1 (ΔG‡1→5a= 22.4 kcal·
1→5a
mol−1). This is about 8 kcal·mol−1 lower than with MP and
agrees well with the experimental observations (a faster reaction
E
dx.doi.org/10.1021/ja504024h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
of 1 with DMAD than with MP). Overall, the process leading
to 5a is thermodynamically favored (ΔG = −28.0 kcal·mol−1)
compared to that leading to 5b (ΔG = −9.8 kcal/mol), in line
with experimental observations.
leading to 6b (ΔH‡
= 26.0 kcal·mol−1, ΔG‡
= 45.6
1→6b
1→6b
kcal·mol−1) is substantially larger than that computed for the
formation of 6a, in agreement with the complete regioselec-
tivity observed experimentally.
Reaction of Ethyl 2,3-Butadienoate with Gold Silyl
Complexes. We reported previously that the replacement of
the Si(t-Bu)Ph2 group by SiPh3 induces a substantial increase in
reactivity of the Au−Si bond. Accordingly, complex 7 was
found to react readily with ethyl 2,3-butadienoate (EB),
affording the gold vinyl complex 8a as the sole product
(Scheme 6). Surprisingly, the insertion reaction involves the
terminal nonactivated double bond of the allene, and the silyl
group is introduced in the γ-position to the CO2Et substituent.
To gain insight into this rather unexpected selectivity, the
mechanism of the reaction between 7 and EB was investigated
computationally.
The molecular structure of π1→5a deserves some comments.
Here also the PAuSi fragment is strongly bent (111.4°) and the
alkyne coordinates in the PAuSi plane. Despite the symmetric
nature of DMAD, the π-complex is dissymmetric (dAuC1(trans to P)
= 2.132 Å, dAuC2(trans to Si) = 2.270 Å) as the result of the
stronger trans influence of the silyl over the phosphine ligand.28
The lower activation barrier found for DMAD compared to
MP is consistent with the smaller HOMO(1)−LUMO(alkyne)
energy gap (the LUMO of DMAD is lower in energy than that
of MP by 0.2 eV) and the higher 1 → alkyne charge transfer
(0.85 e− au for DMAD vs 0.80 e− for MP, according to NBO
charges). The enhanced electro-deficient character of DMAD
also favors Michael addition of PPh3 and probably explains the
competitive insertions in the Au−Si and Au−P bond of 1, while
MP selectively inserts in the Au−Si bond.
At this stage, it is important to note that allenes have been
frequently used in the context of gold-catalyzed π-activation,
and computational analyses of the reaction mechanisms have
suggested the possible existence of several isomeric structures
for the key gold allene complexes.29 In addition to the classical
η2-allene form, the involvement of η1-structures (bent allene,
planar allylic, and zwitterionic carbene) has been discussed.
Insertion of EB into Au−Si bonds may thus be quite different
and more complicated than that of alkynes, as studied formerly.
First, we studied the reaction of complex 7 with the terminal
C1C2 bond of the allene, considering the formation of the
two regioisomers 8a and 8b (Figure 6). In both cases, the
insertion reaction is exergonic and occurs via a two-step
coordination−insertion pathway.30 The formation of the vinyl
gold complex 8a (ΔG7→8a = −23.1 kcal·mol−1) is thermody-
namically favored by 3 kcal·mol−1 over the formation of the
alkyl gold complex 8b (ΔG7→8b = −20.2 kcal·mol−1), in
agreement with experimental observations. The reaction starts
by the coordination of the terminal C1C2 bond of BE to
gold, which is endergonic by about 10 kcal·mol−1. The resulting
complexes π7→8a and π7→8b adopt η2-allene structures, and their
relative stability is opposite to that of the insertion products 8a/
8b (π7→8a, leading to the more stable insertion product, is less
stable than π7→8b by about 3 kcal·mol−1). The bonding
situation is similar in both π-complexes, the main orbital
interaction involving the HOMO of the gold silyl complex and
the LUMO+1 (π*C1C2) of the allene (Figure S31). However,
their geometric features are slightly different: the allene binds
symmetrically in π7→8a (AuC1, 2.200 Å; AuC2, 2.189 Å)
whereas the gold center is tilted toward the central carbon atom
of EB in π7→8b (AuC1 = 2.381 Å; AuC2 = 2.113 Å).
Reaction of the Phosphine Gold Silyl Complex 1 with
Phenyl Acetylene. We have previously shown that complex 1
reacts with PA to give the corresponding vinyl gold complex 6a
(Scheme 5). As for MP, the reaction proceeds with complete
stereo- and regioselectivity but it requires a longer reaction time
(3 weeks vs 8 h).
Scheme 5. Syn Insertion of Phenyl Acetylene in the Au−Si
Bond of the Gold Silyl Complex 1
Calculations support a two-step coordination−insertion
pathway for the formation of 6a from 1 and PA (Figure 5).
The associated activation barrier amounts to ΔH‡
= 21.4
1→6a
kcal·mol−1 (ΔG‡1→6a = 39.6 kcal·mol−1). This value is about 10
kcal·mol−1 higher than that predicted for MP, in line with
experimental observations. The intermediate π-complex π1→6a
is also found somewhat uphill in energy (ΔGπ1→6a = 9.8 kcal·
mol−1). The overall structures of the intermediate π1→6a and
transition state TS1→6a are very similar to those observed with
MP. The alkyne is dissymmetrically coordinated to gold, the
stronger interaction involving the terminal carbon atom. In
π1→6a, the AuC(H) and AuC(Ph) distances equal 2.148 and
2.387 Å, respectively. The only noticeable difference between
PA and MP relates to the magnitude of their interaction with
the gold silyl complex 1. PA forms a slightly looser π-complex
than MP (as apparent from the lengthened AuC distances).
Consistently, analysis of the frontier molecular orbitals indicates
a less favorable HOMO(1)−LUMO(alkyne) interaction for PA
than for MP (the LUMO of PA is 0.5 eV higher in energy than
that of MP). A smaller charge transfer is also found from
complex 1 to PA (0.71 e−) than to MP (0.80 e−).
Careful examination of the transition states associated with
the insertion step (Figure 7) revealed important features and
striking differences between the two pathways. TS7→8a leads to
the insertion product obtained experimentally 8a (vinyl gold
complex with the silyl group in γ-position to the CO2Et
substituent). It corresponds to a direct 1,2-insertion of the
C1C2 double bond of BE in the Au−Si bond of 7. In this case,
the TS adopts a planar four-membered ring structure. The silyl
group migrates from Au to C1 while the gold center shifts
toward C2. The process does not involve the adjacent C
CH(CO2Et) fragment: the C2C3 bond remains short (1.345 Å,
vs 1.460 Å for the C1C2 bond), C3 remains trigonal planar
(whereas C1 is pyramidalized, sum of the bond angles Σα =
348°), and the terminal CH2/CH(CO2Et) moieties remain
quasi perpendicular (θ = 82.8°). The corresponding activation
To complete the computational study, the formation of the
other regioisomer 6b (not observed experimentally) was
considered. The corresponding intermediate π1→6b adopts a
symmetric structure (with essentially equal AuC distances of
2.198 and 2.207 Å) and lies 3.2 kcal·mol−1 higher in energy
than π1→6a. The activation barrier for the insertion of PA
barrier is very high (ΔH‡
= 20.0 kcal·mol−1, ΔG‡
=
7→8a
7→8a
F
dx.doi.org/10.1021/ja504024h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 5. Reaction profiles for the formation of complexes 6a and 6b as the result of insertion of phenyl acetylene in the Au−Si bond of complex 1
(ΔG in kcal·mol−1), computed at the B97D/SDD+f(Au)-6-31G** level of theory. Enthalpies of activation ΔH‡ (in kcal·mol−1) in parentheses.
Scheme 6. Reaction of Ethyl 2,3-Butadienoate with the Gold Silyl Complex 7
38.4 kcal·mol−1), and it is very unlikely that complex 8a is
actually formed by this route.
constitutive fragments, i.e. the gold silyl complex and the
allene. The energy required to distort each fragment from its
ground-state structure to its TS geometry was estimated via
The other transition state TS7→8b adopts an out-of-plane
structure. The silicon atom is tilted out of the AuC1C2 plane
and approaches C2 (SiC2, 2.422 Å). In this case, the allene
moiety is significantly distorted and tends to adopt planar allyl-
type geometry. The terminal CH2/CH(CO2Et) fragments are
no longer perpendicular (θ = 41°), and the C1C2/C2C3
distances tend to equalize (1.411 and 1.421 Å, respectively).
The gold center interacts strongly with C2 (AuC2, 2.146 Å;
PAuC2, 170.9°), and the overall structure can be described as
an η1-allyl gold complex with a silyl group bridging over Au and
single-point energy calculations (ΔE‡
and ΔE‡dist‑allene).
dist‑AuSi
The interaction energy between the two fragments ΔE‡ was
int
directly given by the difference between the activation energy
(ΔE‡) and the overall distortion energy ΔE‡strain = ΔE‡
+
dist‑AuSi
ΔE‡dist‑allene. The results obtained for TS7→8a and TS7→8b are
disclosed in Figure 8. The so determined interaction energy
ΔE‡int is significantly larger for TS7→8b (−63.1 kcal·mol−1) than
for TS7→8a (−48.7 kcal·mol−1). This is consistent with the fact
that the interaction between the two fragments is dominated by
the transfer of electron density from the gold complex to the
allene. And as mentioned above, the adjacent CCH(CO2Et)
group remains a spectator in TS7→8a but actively participates in
TS7→8b (due to the planarization of the allene framework),
which contributes to the enhancement of the interaction of the
gold silyl complex with the otherwise unactivated C1C2
double bond of EB. As expected, the energy required to distort
the allene is larger for TS7→8b than for TS7→8a, but the
difference is relatively small (3.8 kcal·mol−1) and more than
compensated by the lower distortion energy of the gold silyl
complex for TS7→8b than for TS7→8a (11.9 kcal·mol−1). Thus,
both terms, the strain energy and the interaction energy, pull
together and favor the formation of 8b over 8a.
C2. This results in a much lower activation barrier (ΔH‡
=
7→8b
−2.7 kcal·mol−1, ΔG‡
= 17.0 kcal·mol−1), suggesting that
7→8b
the formation of 8b is kinetically much more favorable than
that of 8a.
Thus, in this case, the structures and energies of the two
regioisomeric transition states are very different from each
other. Moreover, the energy profiles computed for the
formation of 8a and 8b are not consistent with the ready and
selective formation of 8a, as observed experimentally. To gain
insights into the difference between these two insertion
reactions, transition states TS7→8a and TS7→8b were analyzed
using the activation strain model,31,32 which enables decom-
position of the activation energy ΔE‡ into two components, the
strain energy ΔE‡
and the interaction energy ΔE‡int. Each
A possibility to reconcile the reaction profiles predicted
computationally and the selective formation of regioisomer 8a
strain
transition state was considered as the interaction of its
G
dx.doi.org/10.1021/ja504024h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 6. Two-step mechanism for the insertion of the terminal double bond of the allene into the Au−Si bond of the gold silyl complex 7 (ΔG in
kcal·mol−1), computed at the B97D/SDD+f(Au)-6-31G** level of theory. Enthalpies of activation ΔH‡ (in kcal·mol−1) in parentheses.
Figure 7. Main geometric features of the transition states associated
with the insertion of the terminal double bond of ethyl 2,3-
butadienoate in the Au−Si bond of complex 7 (B97D/SDD+f(Au)-
6-31G** level of theory).
Figure 8. Distortion/interaction analysis of the transition states with
the insertion of the terminal double bond of ethyl 2,3-butadienoate in
observed experimentally would likely involve the kinetically
favored insertion product 8b subsequently rearranging into the
thermodynamically favored vinyl gold complex 8a. Consis-
tently, a transition state (TS8b→8a) linking both products could
be located on the PES. It is located only 12.3 kcal·mol−1 higher
in energy compared to the initial reactants (and thus lower in
the Au−Si bond of complex 7 (energies in kcal·mol−1).
energy than the insertion transition states TS7→8a and TS7→8b
)
and can certainly be reached under the experimental conditions
(heating at 60 °C overnight). Interconversion of the alkyl and
vinyl gold complexes 8b and 8a is a rather unusual process.
TS8b→8a corresponds to the exchange of the relative positions of
gold and silicon on C1/C2, and its structure deserves
comments (Figure 9). The geometry resembles that of
TS7→8b, the allene moiety being further distorted toward planar
allyl-type geometry: the terminal CH2/CH(CO2Et) fragments
are rotated by only θ = 18°, and the C1C2/C2C3 distances are
almost identical at 1.410 and 1.433 Å, respectively. The gold
Figure 9. Main geometric features of the transition state associated
with the isomerization between the gold complexes 8b and 8a (B97D/
SDD+f(Au)-6-31G** level of theory).
H
dx.doi.org/10.1021/ja504024h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
center deviates only slightly from the plane of the C1C2C3 allyl
fragment and interacts strongly with C2 (AuC2, 2.163 Å;
PAuC2, 177.9°). The silicon atom also interacts mainly with C2
(2.038 Å) but sits above the η1-allyl gold fragment, slightly
tilted toward C1 (SiC1, 2.493 Å vs SiC3, 2.689 Å).
uk/data_request/cif. Detailed experimental conditions and
procedures, theoretical details, and analytical data. This material
is available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
For the sake of comprehensiveness, insertion of the C2C3
bond of EB into the Au−Si bond of complex 7 was also
considered. The electron-withdrawing CO2Et group signifi-
cantly activates this double bond, the corresponding π* orbital
(which is actually the LUMO of EB) being 1.1 eV lower in
energy than the π*C1C2 (LUMO+1) orbital. Despite this
electronic balance, the two conceivable π-adducts involving the
C2C3 bond of EB (Figure S32) are about isoenergetic with
π7→8a and π7→8b involving C1C2. This is probably due to
steric shielding between the CO2Et group at C3 and the bulky
(Ph3P)AuSiPh3 fragment, which counterbalances electronic
effects. Yet, insertion of the C2C3 into the Au−Si bond was
found to be less favored thermodynamically (ΔG7→9a = −15.5
kcal·mol−1; ΔG7→9b = −18.5 kcal·mol−11) and to require higher
activation barriers (ΔH‡7→9a = 20.5 kcal·mol−1, ΔG‡7→9b = 41.3
kcal·mol−1 and ΔH‡7→9b = 5.7 kcal·mol−1, ΔG‡7→9b = 25.0 kcal·
mol−1). Thus, insertion of the activated C2C3 double bond
of EB is both thermodynamically and kinetically disfavored over
the insertion of the terminal C1C2 bond.
Corresponding Authors
karinne.miqueu@univ-pau.fr
amgoune@chimie.ups-tlse.fr
dbouriss@chimie.ups−tlse.fr
Author Contributions
⊥M.J. and L.E. contributed equally to this work.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
Financial support from the Centre National de la Recherche
Scientifique, the Universite
́
de Toulouse, and the Agence
Nationale de la Recherche (ANR-10-BLAN-070901) is grate-
fully acknowledged. The theoretical work was granted access to
HPC resources of Idris under Allocation 2013 (i2013080045)
made by Grand Equipement National de Calcul Intensif
(GENCI). L.E. thanks the Xunta de Galicia for a postdoctoral
contract under the I2C program.
Altogether, the computational study carried out on the
reaction between complex 7 and ethyl 2,3-butadienoate
provides mechanistic insights and explains the rather
unexpected selectivity observed experimentally. Insertion of
the terminal nonactivated CC double bond of EB into the
Au−Si bond is favored and first gives the alkyl gold complex 8b,
which subsequently rearranges into the vinyl gold complex 8a
via an original Au/Si exchange process.
REFERENCES
■
(1) (a) Hashmi, A. S. K.; Hutchings, G. J. Angew. Chem., Int. Ed.
2006, 45, 7896. (b) Hashmi, A. S. K. Chem. Rev. 2007, 107, 3180.
(c) Li, Z.; Brouwer, C.; He, C. Chem. Rev. 2008, 108, 3239. (d) Arcadi,
A. Chem. Rev. 2008, 108, 3266. (e) Gorin, D. J.; Sherry, B. D.; Toste,
F. D. Chem. Rev. 2008, 108, 3351. (f) Jimenez-Nunez, E.; Echavarren,
A. M. Chem. Rev. 2008, 108, 3326. (g) Furstner, A. Chem. Soc. Rev.
̈
2009, 38, 3208. (h) Corma, A.; Leyva-Perez, A.; Sabater, M. J. Chem.
Rev. 2011, 111, 1657. (i) Boorman, T. C.; Larrosa, I. Chem. Soc. Rev.
2011, 40, 1910. (j) Bandini, M. Chem. Soc. Rev. 2011, 40, 1358.
(k) Hashmi, A. S. K.; Toste, F. D. Modern Gold Catalyzed Synthesis, 1st
ed.; Wiley-VCH Verlag GmBH & Co. KGaA: Weinheim, 2012. (l) Acc.
Chem. Res. 2014, issue 3; special issue edited by Friend, C. M. and
Hashmi, A. S. K. on gold catalysis.
CONCLUSION
■
This comprehensive experimental/theoretical study affords a
detailed mechanistic picture for the reaction of alkynes and
allenes with gold silyl complexes leading to syn insertion
products. The combination of kinetic studies, isolation and
characterization of vinyl gold complexes, and DFT calculations
provide evidence that the reaction is bimolecular and proceeds
through a two-step coordination−insertion pathway. The π-
substrates first coordinate to gold and then insert into the Au−
Si bond in a syn fashion through a concerted inner-sphere
mechanism. DFT calculations also give insight into the
regioselectivity of the insertion reactions, in particular in the
case of ethyl 2,3-butadienoate for which an original isomer-
ization step (Au/Si exchange) explains the unexpected
regioselectivity observed experimentally.
These studies substantiate unequivocally the ability of gold to
promote the functionalization of π-substrates through an inner-
sphere mechanism (alternatively to the commonly encountered
outer-sphere nucleophilic addition)15,33 and open the way to
new useful chemical transformations. We are currently
investigating the generality of coordination−insertion reactions
at gold with the aim to increase further the versatility of this
peculiar element.
(2) For recent reviews, see: (a) Furstner, A.; Davies, P. W. Angew.
̈
Chem., Int. Ed. 2007, 46, 3410. (b) Hashmi, A. S. K. Angew. Chem., Int.
Ed. 2010, 49, 5232. (c) Liu, L. P.; Hammond, G. B. Chem. Soc. Rev.
2012, 41, 3129.
(3) For recent reviews, see: (a) Schmidbaur, H.; Schier, A.
Organometallics 2010, 29, 2. (b) Brooner, R. E. M.; Widenhoefer, R.
A. Angew. Chem., Int. Ed. 2013, 52, 11714.
(4) (a) Gomez-Suarez, A.; Nolan, S. P. Angew. Chem., Int. Ed. 2012,
51, 8156. (b) Braun, I.; Asiri, A. M.; Hashmi, A. S. K. ACS Catal. 2013,
3, 1902. (c) Hashmi, A. S. K. Acc. Chem. Res. 2014, 47, 864.
(5) Lein, M.; Rudolph, M.; Hashmi, A. S. K.; Schwerdtfeger, P.
Organometallics 2010, 29, 2206.
(6) (a) Teles, J. H.; Brode, S.; Chabanas, M. Angew. Chem., Int. Ed.
1998, 37, 1415. (b) Krauter, C. M.; Hashmi, A. S. K.; Pernpointner, M.
ChemCatChem 2010, 2, 1226.
(7) (a) Mizushima, E.; Hayashi, T.; Tanaka, M. Org. Lett. 2003, 5,
3349. (b) Nishina, N.; Yamamoto, Y. Angew. Chem., Int. Ed. 2006, 45,
3314. (c) Lavallo, V.; Frey, G.; Donnadieu, B.; Soleilhavoup, M.;
Bertrand, G. Angew. Chem., Int. Ed. 2008, 47, 5224. (d) Zeng, X.;
Soleilhavoup, M.; Bertrand, G. Org. Lett. 2009, 11, 3166. (e) Kinjo, R.;
Donnadieu, B.; Bertrand, G. Angew. Chem., Int. Ed. 2011, 50, 5560.
(8) For addition of water and methanol to alkynes, see: Casado, R.;
Contel, M.; Laguna, M.; Romero, P.; Sanz, S. J. Am. Chem. Soc. 2003,
125, 11925.
ASSOCIATED CONTENT
* Supporting Information
CCDC numbers for compounds 5a and 5b, CCDC 995108−
995109, contain the supplementary crystallographic data for
this paper. These data can be obtained free of charge from The
Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.
■
S
(9) For mechanistic studies of gold catalyzed hydroamination of
allenes, see: Wang, Z. J.; Benitez, D.; Tkatchouk, E.; Goddard, W. A.,
III; Toste, F. D. J. Am. Chem. Soc. 2010, 132, 13064.
I
dx.doi.org/10.1021/ja504024h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(10) The anti addition of amines to alkenes and allenes has been
supported experimentally through labeling experiments, see: LaLonde,
R. L.; Brenzovich, J.; Benitez, D.; Tkatchouk, E.; Kelley, K.; Goddard,
W. A., III; Toste, F. D. Chem. Sci. 2010, 1, 226.
Menges, F.; Niedner-Schatteburg, G. J. Am. Chem. Soc. 2011, 133,
7428.
(21) Grimme, S. J. Comput. Chem. 2006, 27, 1787.
(22) Grimme, S.; Diedrich, C.; Korth, M. Angew. Chem., Int. Ed.
2006, 45, 625.
(11) For theoretical studies of the hydroamination of alkynes, see:
(23) For discussions on the difficult estimation of entropic efffects
with multimolecular processes, see: Braga, A. A. C.; Ujaque, G.;
Maseras, F. Organometallics 2006, 25, 3647.
(24) The only TS connecting 1 + DMAD and 5b located on the PES
is associated with direct 1,2-addition (in-plane four-membered ring
TS). But the corresponding activation barrier (ΔG‡ = 52.6 kcal·mol−1)
is prohibitively high and the formation of 5b is very unlikely to result
from this process (Figure S27).
́ ́
(a) Kovacs, G.; Lledos, A.; Ujaque, G. Angew. Chem., Int. Ed. 2011, 50,
11147. (b) Liu, X.-Y.; Guo, Z.; Dong, S. S.; Li, X.-H.; Che, C.-M.
Chem.Eur. J. 2011, 17, 12932.
(12) Theoretical investigations of gold catalyzed hydrogenation of
alkenes also support an outer-sphere mechansim; see: Comas-Vives,
A.; Ujaque, G. J. Am. Chem. Soc. 2013, 135, 1295.
(13) In addition, all the gold vinyl complexes isolated from the attack
of nucleophiles (carboxamide, hydride, fluoride) to alkynes display a
(25) For discussions on phosphine dissociation from gold, see:
Kumar, M.; Jasinski, J.; Hammond, G. B.; Xu, B. Chem.Eur. J. 2014,
20, 3113.
trans configuration; see: (a) Akana, J. A.; Bhattacharyya, K. X.; Muller,
̈
P.; Sadighi, J. P. J. Am. Chem. Soc. 2007, 129, 7736. (b) Tsui, E.;
Muller, P.; Sadighi, J. P. Angew. Chem., Int. Ed. 2008, 47, 8937.
̈
(26) Phosphines are known to readily react with electron-poor
alkynes to give zwitterionic phosphonium enolates. See for example:
Zhu, X.-F.; Henry, C. E.; Kwon, O. J. Am. Chem. Soc. 2007, 129, 6722.
(27) The direct 1,2-addition pathway is associated with a much
higher activation barrier (ΔG‡ = 38.4 kcal·mol−1), and the two-step
mechanism involving phosphine dissociation/1,2-addition/phosphine
association requires an even higher activation barrier (ΔG‡ = 55.1 kcal·
mol−1) (Figure S29).
(c) Hashmi, A. S. K.; Schuster, A. M.; Rominger, F. Angew. Chem., Int.
Ed. 2009, 48, 8247. (d) Hashmi, A. S. K. Gold Bull. 2009, 42, 275.
́
(e) Weber, D.; Tarselli, M. A.; Gagne, M. R. Angew. Chem., Int. Ed.
2009, 48, 5733. (f) Zeng, X.; Kinjo, R.; Donnadieu, B.; Bertrand, G.
Angew. Chem., Int. Ed. 2010, 49, 942. (g) Egorova, O. A.; Seo, H.; Kim,
Y.; Moon, D.; Rhee, Y. M.; Ahn, K. H. Angew. Chem., Int. Ed. 2011, 50,
11446. (h) Rosca, D.-A.; Smith, D. A.; Hughes, D. L.; Bochmann, M.
Angew. Chem., Int. Ed. 2012, 51, 10643.
(28) A similar situation was reported for the trans-[PtCl(SiMePh2)
(PMe2Ph)2] complex. Due to the trans influence of the silyl group, the
Pt−Cl distance is elongated by about 0.15 Å compared to that of
(14) Gold(I) amide and gold(I) phenoxide complexes have been
recently isolated, and their reactivity towards alkynes has been
investigated. No reaction was observed even at high temperatures,
indicating that these compounds are unlikely intermediates in the
insertion of π-substrates into Au−N bonds or Au−O bonds; see:
(a) Johnson, M. W.; Shevick, S. L.; Toste, F. D.; Bergman, R. G. Chem.
Sci. 2013, 4, 1023. (b) Oonishi, Y.; Gomez-Suarez, A.; Martin, A. R.;
Nolan, S. P. Angew. Chem., Int. Ed. 2013, 52, 9767.
(15) Stradiotto et al. have reported gold catalyzed hydroamination of
internal alkynes with dialkylamines. The stereo- and regioselectivity of
the reactions combined with stoichiometric reactions led the authors
to propose an inner-sphere insertion mechanism; see: Hesp, K. D.;
Stradiotto, M. J. Am. Chem. Soc. 2010, 132, 18026. Recent mechanistic
studies by Zhdanko and Maier provide further insight and point out
the involvement of conformationally flexible auro-iminium salts.
Accordingly, the hydroamination proceeds by outer-sphere insertion
and rotation around the C−CAu bond of these auro-iminium salts is
responsible for the syn arrangement of the final insertion products.
See: (b) Zhdanko, V.; Maier, M. Angew. Chem., Int. Ed. 2014,
DOI: 10.1002/anie.201402557.
(PtCl )2−; see: Kapoor, P.; Lovqvist, K.; Oskarsson, Å. Acta Crystallogr.
̈
4
1995, C51, 611.
(29) For leading contributions on the gold chemistry of allenes, see:
(a) Gandon, V.; Lemiere, G.; Hours, A.; Fensterbank, L.; Malacria, M.
̀
Angew. Chem., Int. Ed. 2008, 47, 7534. (b) Malacria, M.; Fensterbank,
L.; Gandon, V. Top. Curr. Chem. 2011, 302, 157. (c) Montserrat, S.;
Ujaque, G.; Lopez, F.; Mascarenas, J. L.; Lledos, A. Top. Curr. Chem.
2011, 302, 225. (d) Krause, N.; Winter, C. Chem. Rev. 2011, 111,
1994. (e) Montserrat, S.; Faustino, H.; Lledos, A.; Mascarenas, J. L.;
Lopez, F.; Ujaque, G. Chem.Eur. J. 2013, 19, 15248. (f) Yang, W.;
Hashmi, A. S. K. Chem. Soc. Rev. 2014, 43, 2941.
(30) A similar two-step mechanism has been computed for
palladium-catalyzed sila-stannation of allenes; see: Wang, M.; Cheng,
L.; Hong, B.; Wu, Z. J. Comput. Chem. 2009, 30, 1521.
(31) For selected mechanistic computational studies including
distortion/interaction analyses, see: (a) Ess, D. H.; Houk, K. N. J.
Am. Chem. Soc. 2007, 129, 10646. (b) Ess, D. H.; Houk, K. N. J. Am.
Chem. Soc. 2008, 130, 10187. (c) Hong, X.; Liang, Y.; Houk, K. N. J.
Am. Chem. Soc. 2014, 136, 2017. (d) Green, A. G.; Liu, P.; Merlic, C.
A.; Houk, K. N. J. Am. Chem. Soc. 2014, 136, 4575.
(32) van Zeist, W.-J.; Bickelhaupt, F. M. Org. Biomol. Chem. 2010, 8,
3118.
(33) For cis-disilylation and silaboration of alkynes catalyzed by
supported gold nanoparticles, see: (a) Gryparis, C.; Kidonakis, M.;
Stratakis, M. Org. Lett. 2013, 15, 6038. (b) Gryparis, C.; Stratakis, M.
Org. Lett. 2014, 16, 1430.
(16) (a) For syn carboauration catalyzed by Pd complexes, see: Shi,
Y.; Ramgren, S. D.; Blum, S. A. Organometallics 2009, 28, 1275. (b)
For two-step boraauration of terminal alkynes giving cis gold vinyl
complexes, see: Ye, H.; Lu, Z.; You, D.; Chen, Z.; Li, Z. H.; Wang, H.
Angew. Chem., Int. Ed. 2012, 51, 12047.
(17) Joost, M.; Gualco, P.; Mallet-Ladeira, S.; Amgoune, A.;
Bourissou, D. Angew. Chem., Int. Ed. 2013, 52, 7160.
(18) For recent contributions involving the insertion of lakynes into
Cu−Si bonds, see: (a) Fujihara, T.; Tani, Y.; Semba, K.; Terao, J.;
Tsuji, Y. Angew. Chem., Int. Ed. 2012, 51, 11487. (b) Iannazzo, L.;
Molander, G. A. Eur. J. Org. Chem. 2012, 4923. (c) Calderone, J. A.;
Santos, W. L. Angew. Chem., Int. Ed. 2014, 53, 4154. (d) Linstadt, R. T.
H.; Peterson, C. A.; Lippincott, D. J.; Jette, C. I.; Lipshutz, B. H.
Angew. Chem., Int. Ed. 2014, 53, 4159.
(19) Experiments in the presence of an excess of PPh3 (0.5, 1, and 2
equiv) were also carried out to investigate the influence of added
phosphine on the reaction rate. However, the addition of free
phosphine resulted in the immediate polymerization of MP whatever
the ratio of PPh3. See: Simionescu, C. I.; Bulachovschi, V.; Grovu-
Ivanoiu, M.; Stanciu, A. J. Macromol. Sci., Chem. 1987, A24, 611.
(20) (a) Gomez-Gallego, M.; Sierra, A. M. Chem. Rev. 2011, 111,
4857. (b) Weiss, C. J.; Marks, T. J. J. Am. Chem. Soc. 2010, 132, 10633.
(c) Weiss, C. J.; Wobser, S. D.; Marks, T. J. Organometallics 2010, 29,
6308. (d) Arndt, M.; Salih, K. S. M.; Fromm, A.; Goossen, L. J.;
J
dx.doi.org/10.1021/ja504024h | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX