Mechanistic studies of gold and palladium cooperative dual-catalytic cross-coupling systems

Article abstract of DOI:10.1021/cs400641k

Double-label crossover, modified substrate, and catalyst comparison experiments in the gold and palladium dual-catalytic rearrangement/cross- coupling of allenoates were performed in order to probe the mechanism of this reaction. The results are consisten

Full text of DOI:10.1021/cs400641k

Research Article
pubs.acs.org/acscatalysis
Mechanistic Studies of Gold and Palladium Cooperative Dual-
Catalytic Cross-Coupling Systems
Mohammad Al-Amin, Katrina E. Roth, and Suzanne A. Blum*
Department of Chemistry, University of California at Irvine, Irvine, California 92697-2025, United States
S
* Supporting Information
ABSTRACT: Double-label crossover, modified substrate, and catalyst
comparison experiments in the gold and palladium dual-catalytic rearrange-
ment/cross-coupling of allenoates were performed in order to probe the
mechanism of this reaction. The results are consistent with a cooperative
catalysis mechanism whereby (1) gold activates the substrate prior to
oxidative addition by palladium, (2) gold acts as a carbophilic rather than
oxophilic Lewis acid, (3) competing olefin isomerization is avoided, (4) gold
participates beyond the first turnover and therefore does not serve simply to
generate the active palladium catalyst, and (5) single-electron transfer is not
involved. These experiments further demonstrate that the cooperativity of
both gold and palladium in the reaction is essential because significantly
lower to zero conversion is achieved with either metal alone in comparison
studies that examined multiple potential gold, palladium, and silver catalysts and precatalysts. Notably, employment of the
optimized cocatalysts, PPh3AuOTf and Pd2dba3, separately (i.e., only Au or only Pd) results in zero conversion to product at all
monitored time points compared to quantitative conversion to product when both are present in cocatalytic reactions.
KEYWORDS: gold, palladium, cooperative catalysis, dual catalysis, cross-coupling, crossover, mechanism, intermediates
the broader understanding needed to design future dual-metal
systems.
INTRODUCTION
■
High-yielding reactions that maximize the number of bond-
forming processes per pot are attractive because they avoid the
isolation of intermediates,1 which is often the most time- and
cost-intensive part of a synthesis.2 One approach toward these
efficient multiple-bond-forming reactions is through dual
catalysis.3−16 For example, dual-catalytic cross-coupling reac-
tions with gold and palladium access novel reactivity
unavailable to single-metal systems, use substoichiometric
amounts of each catalyst, avoid the stochiometric trans-
metalation byproducts common to cross-coupling reactions
(except the Sonogashira reaction), and do not require the
isolation and purification of reaction intermediates.3−16 Less is
known, however, about the mechanisms available to dual-metal-
catalyzed systems as compared to their single-metal counter-
parts; this dearth of mechanistic understanding hinders the
design of additional dual-catalytic systems despite their
advantages.
We herein report mechanistic studies in the gold-and-
palladium cooperative catalytic cyclization/cross-coupling re-
action. These experiments include double-label crossover
studies and modified substrate studies that “isolate” individual
steps in the reaction. Experiments are consistent with a “gold
cyclization first” mechanism whereby gold catalyzes (i.e., lowers
the barrier for) a subsequent palladium-catalyzed oxidative
addition and further confirms5b the involvement of both of the
Au and Pd cocatalysts in the conversion to product. This
fundamental information provides insight into the reactions
available to gold and palladium together and thus contributes to
RESULTS AND DISCUSSION
■
We turned our attention to probing the mechanism of catalysis
in order to better understand and predict future chemo- and
regioselectivities and reactivity in Au/Pd and related dual-
catalytic systems. Our group had previously proposed a
consistent catalytic cycle for the Au/Pd dual-catalyzed synthesis
of butenolides. Optimized conditions are shown in eq 1. We
proposed a mechanism in which a carbophilic17 gold Lewis acid
first activated the C−C π system prior to involvement by
palladium (Scheme 1).5b Recent work by our group and others
suggests that oxophilic18 and azaphilic17a,19−21 activation by
gold and SET10b mechanisms may also be possible. In this
series of mechanistic studies, we investigated the precise order
of steps and the possibility of these alternative activation
pathways which will be called “palladium first”, “oxophilic”,
Received: August 4, 2013
Revised: December 27, 2013
Published: December 31, 2013
© 2013 American Chemical Society
622
dx.doi.org/10.1021/cs400641k | ACS Catal. 2014, 4, 622−629

ACS Catalysis
Research Article
Scheme 1. Original Proposed Gold Cyclization First
Mechanism for Au/Pd Dual-Catalytic Synthesis of
Butenolides
Scheme 2. Alternative Palladium First Possible Mechanism
for Au/Pd Dual-Catalytic Cyclizaiton/Cross-Coupling
organogold complex 4 and π-allyl Pd complex 5 would afford
the observed butenolide product 2.
“palladium π-allyl cation catalyzed”, and “SET” for clarity. Our
data are consistent with the originally proposed “gold
cyclization first” catalytic cycle wherein gold acts as a
carbophilic17 Lewis acid and generates an organogold cross-
coupling partner in situ.
We probed the potential of the Pd(0) oxidative addition with
our allenoate substrates through a series of double-label
crossover experiments.23 In the first experiment, allyl allenoates
1a and 1b (in a 1:1 mixture) were treated with catalytic and
high loading of palladium only, specifically, Pd2dba3 and PPh3
in CD2Cl2 (Scheme 3). These results were later compared to a
second crossover experiment under dual-catalytic conditions
with both gold and palladium (Scheme 4).5b
The reaction with palladium alone was monitored for the
presence of the crossover products from palladium-catalyzed
deallylation.24 If this deallylation occurs, the resulting
intermediates could recombine in an intermolecular fashion
to give crossover products. However, it must be noted that after
deallylation the resulting π-allyl palladium complex and
Summary of Previous Experiments. Kinetics. In our
initial communication,5b we reported the first-order kinetic
dependence of product formation on palladium concentration.
Product decomposition at higher loadings of gold (>5 mol %)
complicated measurement of the kinetic dependence on gold.
Over the examined loadings of gold and palladium, the reaction
exhibited substrate saturation kinetics, consistent with a resting
state with the substrate bound to the gold, palladium, or both.
Characterization of Potential Intermediate. Proposed
intermediate 5, with L = dba and L = ((PPh3)2), was present
in the catalytic reaction mixture prior to reaction completion, as
characterized by HRMS. This “build up” of a substrate-bound
organometallic species is consistent with the observed substrate
saturation kinetics.
Equilibria. A trapping experiment established that 1 and 3
were in rapid equilibrium in the presence of stoichiometric
gold. This result indicates a low barrier for the initial cyclization
step compared to subsequent steps. Thus, gold-mediated/
catalyzed cyclization is not rate-determining if it is indeed on
the reaction path.
Scheme 3. Crossover Experiment with Allenoate Substrates
1a and 1b with Only Palladium Demonstrates Crossover Is
Slower in the Absence of Gold
Motivation for Additional Data. As mentioned in the
Introduction, at least five plausible mechanistic pathways match
the previous data; these mechanistic pathways are considered
plausible in part due to studies by our group5f and others10b on
dual-metal systems after our original publication. Determi-
nation of the precise order and nature of steps informs on the
activation chemistry uniquely available to dual-metal systems, as
will now be described.
Experiments Inconsistent with a “Palladium First”
Mechanism. We first considered an alternative mechanism in
which the order of reactivity of Au and Pd is reversed from our
initial proposal: Pd(0) may first undergo oxidative addition into
the allyl group of 1, followed by a Au-catalyzed cyclization
(Scheme 2). The catalyst Pd2dba3 generally requires addition of
a phosphine ligand to promote allyl acetate oxidative addition.22
Thus, PPh3 dissociated from the PPh3AuOTf could bind to Pd
to promote the oxidative addition; this ligand exchange has
been characterized previously in our system by HRMS.5b For
these reasons, the palladium first mechanism is depicted in
Scheme 2 as containing “PPh3” in the oxidative addition step.
In the last step, a cross-coupling reaction between the
623
dx.doi.org/10.1021/cs400641k | ACS Catal. 2014, 4, 622−629

ACS Catalysis
Research Article
(Scheme 5), similar to our recently reported Au(I) azaphilic
activation of vinyl aziridines toward Pd(0) oxidative addition.5b
Scheme 4. Crossover Experiment with Allenoate Substrates
1a and 1b under Optimized Dual-Catalytic Conditions with
Both Gold and Palladium
Scheme 5. Alternative Possible Oxophilic Activation
Mechanism
carboxylate might not be solvent separated25 but rather
covalently attached in dichloromethane, a relatively nonpolar
solvent. Inaccessibility of the solvent-separated ion pairs may
retard crossover product formation, and this caveat should be
taken into account during data interpretation. The presence of
the crossover products with the optimized low loading of
catalytic palladium would indicate that the Pd(0) oxidative
addition step is feasible under the catalytic reaction conditions,
without the need for Au-catalyzed carbophilic17 activation of
the substrate first. The absence of crossover products, or a
slower rate of crossover product formation, with catalytic
palladium alone when compared to dual-catalytic conditions
would establish that gold is required for the observed rate of
crossover.
Accordingly, the crossover study was carried out for allyl
allenoates 1a and 1b in the presence of Pd2dba3 (5 mol %) and
PPh3 (5 mol %) in CD2Cl2 (Scheme 3, condition A). After 25
min, no other peaks except the starting allenoates were
observed by GCMS; thus 5.0 mol % of palladiumidentical to
reaction conditionsgenerates no conversion to either cross-
over product in the absence of gold.
In this proposed pathway, gold first coordinates to the carbonyl
oxygen, forming activated intermediate 7. The increased
cationic charge on the carbonyl would then lower the barrier
(i.e., catalyze) for oxidative addition of palladium to the allyl
C−O bond. This pathway provides an alternative mechanism
for formation of intermediate 6, also considered previously in
Scheme 2.
To test this hypothesis, a crossover experiment was
performed between two substrates that had the possibility for
Au(I) oxophilic activation but no opportunity for Au(I)-
catalyzed carbophilic cyclization (Scheme 6). These modified
Scheme 6. Crossover Experiment with Substrates 8a and 8b
To test if crossover could occur at presumably higher
concentrations of metal-bound intermediates, the same cross-
over experiment was performed with significantly higher
palladium loading, with 6 times the amount of Pd2dba3 (30
mol %) and PPh3 (30 mol %). Under these highly concentrated
substoichiometric conditions, the crossover products 1a′ and
1b′ were obtained along with competing internal olefin
isomerization, 1a″ and 1b″ (Scheme 3, condition B). The
time point 25 min was chosen for analysis because under
cocatalytic conditions with both 5 mol % of gold and 5 mol %
of palladium conversion to product is significant (69%). Thus
crossover does not occur with palladium alone under the
conditions or time scale of the optimized butenolide-forming
reaction but can be induced at significantly higher loadings.
Since oxidative addition into the allyl group (leading to
deallyation/solvent separation/recombination and crossover to
form products 1a′ and 1b′) does not occur with 5 mol % of
palladium (the optimized reaction conditions) in the absence of
gold, gold appears required to activate the substrate toward
oxidative addition. These data are most consistent with a “gold
first” mechanism operating under the optimized dual-catalytic
conditions because the presence of both gold and palladium
affects crossover (Scheme 4),5b whereas at identical concen-
tration of only palladium, no crossover occurs (Scheme 3).
Experiments Inconsistent with the “Oxophilic Activa-
tion” Mechanism. We also considered the possibility that
Au(I) may first act as an oxophilic18 Lewis acid and promote
the Pd(0) oxidative addition into the allyl group of allenoate 1
substrates were therefore employed to “isolate” steps of the
cocatalytic reaction for individual study. In Scheme 6, substrates
8a and 8b are not capable of cyclization due to the absence of
an allene moiety; however, oxidative addition is still possible.
Thus these substrates serve to isolate the oxidative addition
step of this reaction from other steps.
Substrates 8a and 8b were treated with PPh3AuCl/AgOTf (5
mol %) and Pd2dba3 (5 mol %) in CD2Cl2. After 1.5 h, no
reaction occurred. After 24 h, no crossover products were
624
dx.doi.org/10.1021/cs400641k | ACS Catal. 2014, 4, 622−629

ACS Catalysis
Research Article
1
formed; however, partial conversion (45%) of substrates 8a and
8b to their respective internal olefin isomerization products,
8a″ and 8b″, were observed by H NMR spectroscopy. The
Table 1. Comparison of H NMR Yields for Different
Precatalysts
1
lack of crossover products when no alkyne is present suggests
that gold-catalyzed cyclization lowers the barrier for oxidative
addition under the normal conditions where crossover does
occur. This result, however, does not completely rule out a
Pd(0) oxidative addition without gold; the π-allyl palladium
complex and carboxylate may not be free ion pairs25 in
dichloromethane, a relatively nonpolar solvent, which may
retard crossover product formation.
The data establish that under optimized cocatalytic
conditions the presence of both gold and palladium
cyclization/cross-coupling to generate butenolide product 5
outcompetes the olefin isomerization that was first detected in
the experiments described in Scheme 5. Given the longer time
scale for olefin isomerization (24 h) compared to butenolide
formation (1.5 h), it is unlikely that olefin isomerization is rapid
relative to cyclization to product. Therefore, olefin isomer-
ization does not appear to be fast enough to establish Curtin−
Hammett conditions whereby the starting olefin is rapidly
isomerizing between the vinyl ester and allyl ester, and only the
allyl ester provides a viable oxidative addition pathway that
leads to product formation.
Experiments Inconsistent with a “Palladium π-Allyl
Cation Catalyzed” Mechanism. We also considered the
possibility that π-allyl Pd cation 5, formed after one turnover in
Scheme 1, was the active catalyst for the remainder of the
reaction. Specifically, we considered that the role of gold might
be only during the first turnover, to generate 5 which then
catalyzed the reaction efficiently without gold in subsequent
turnovers. To determine whether π-allyl Pd cation could
catalyze the reaction alone, allenoate 1a was treated in the
absence of gold under three different conditions: with
Pd(allyl)(OTf) in the absence of ligand (Table 1, entry 8), in
the presence of dba (Table 1, entry 9), or in the presence of
PPh3 (Table 1, entry 10). The reactions were monitored by 1H
NMR spectroscopy at specific time points for comparison.
These control reactions displayed a slow, but nonzero,
formation of butenolide 2a (entries 8−10, at 1.5 h: 23% with
catalyst death, 13% with catalyst death, and 28% without
catalyst death, respectively). In entries 8−10, catalyst death was
indicated by a halt in further conversion after early time points.
For comparison, the gold-and-palladium dual-catalytic system
reached quantitative conversion after 1.5 h (entry 1). An active
role for the gold cocatalyst thus was demonstrated.
Experiments Inconsistent with “SET” Mechanism. We
examined if PPh3AuOTf was acting as a single-electron oxidant
for Pd(0) to form (PPh3)2Pd(OTf)2.26 Thus, the reaction may
proceed through a Pd(II)/Pd(0) mechanism in which the gold
does not bind to the substrate. In this mechanism, the gold
simply serves to oxidize the palladium precatalyst. To probe
this hypothesis, we treated allyl allenoate 1a with a palladium
precatalyst already in the Pd(II) oxidation state. To best mimic
catalytic conditions wherein palladium in both +2 and 0
oxidation states might be present, a palladium(0) precatalyst
was also added. Specifically, the potentially catalytic mixture of
2.5 mol % of (PPh3)2Pd(OTf)2 and 3.75 mol % of Pd2dba3 was
examined (Table 1, entry 11). The reaction was monitored by
1H NMR spectroscopy. The yield of butenolide product 2a was
low (7−9%) and remained approximately constant between 10
min and 1.5 h, consistent with catalyst death or single turnover
a
1H NMR yield (%)
entry
1
catalyst(s)/ligand (if added)
10 min 25 min 60 min 90 min
PPh3AuCl (5 mol %), AgOTf
(5 mol %), Pd2dba3 (5 mol %)
24
69
93
100
2
PPh3AuCl (5 mol %), AgOTf
(5 mol %)
0
0
0
0
3
4
Pd2dba3 (5 mol %)
0
0
0
0
Pd2dba3 (5 mol %), AgOTf
(5 mol %)
22
23
24
28
5
6
PPh3AuCl (5 mol %), NaBArF
(5 mol %), Pd2dba3 (5 mol %)
23
1
28
2
40
5
47
9
Pd2dba3 (5 mol %), NaBArF
(5 mol %)
7
8
NaBArF (5 mol %)
0
7
0
0
0
Pd2(allyl)2Cl2 (5 mol %), AgOTf
(10 mol %)
13
22
23
9
10
11
Pd2(allyl)2Cl2 (5 mol %), AgOTf
(10 mol %), dba (15 mol %)
0
10
7
6
14
8
12
17
8
13
28
9
Pd2(allyl)2Cl2 (5 mol %), AgOTf
(10 mol %), PPh3 (5 mol %)
Pd(PPh3)2Cl2 (2.5 mol %), AgOTf
(5 mol %), Pd2dba3 (3.75 mol %)
a
1H NMR yield was determined based on the ratio of product 2a
relative to mesitylene, which was used as an internal standard.
and no catalysis. In contrast, the Au/Pd dual-catalytic system
reached quantitative conversion at 1.5 h (Table 1, entry 1).
The cyclic voltammograms (CV) of PPh3AuCl, AgOTf,
PPh3AuCl plus AgOTf, and Pd2dba327 in dichloromethane also
suggest the absence of redox between the precatalysts in the
optimized reaction solvent (Figure 1 and also Supporting
Information Figures S1−S6). While Pd2dba3 shows redox
consistent with the studies of Pd-dba complexes by Amatore
and Jutand,27 the gold and gold/silver mixtures did not show
distinct redox features. Nevertheless, electron transfer may still
be possible between metal-based intermediates generated after
reaction initiation.10b
Thus, catalyst screening, conversion, and CV studies, when
considered together, are most consistent with an absence of
SET as a primary catalytic reaction pathway.
Experiments Consistent with Au/Pd Dual-Metal
Catalysis Mechanism. As mentioned previously, published
kinetics experiments from our group established a kinetic
dependence on both gold and palladium in this reaction.5b
Palladium displayed clean pseudo-first-order kinetics, and gold
displayed a clear kinetic dependence (however, the order in
gold could not be unambiguously determined due to catalyst
decomposition at higher loading).
We next considered whether the single-metal precatalysts
could catalyze the reaction when examined over a time range
analogous to the optimized cocatalytic conditions. For this
comparison, we treated allenoate 1a under optimized
conditions with 5 mol % of PPh3AuCl and 5 mol % of
AgOTf in the presence of 5 mol % of Pd2dba3 (Table 1, entry
1). The reaction was monitored by 1H NMR spectroscopy after
10, 25, 60, and 90 min intervals and showed 24, 69, 93, and
100% conversions, respectively. Next, the allenoate 1a was
treated with only Au and only Pd. In these experiments, no
625
dx.doi.org/10.1021/cs400641k | ACS Catal. 2014, 4, 622−629

ACS Catalysis
Research Article
In this mechanism, substrate cyclization by gold creates a
transmetalation cross-coupling partner in situ for further
reactivity with a palladium π-allyl intermediate. Furthermore,
the mechanistic data support the cooperative role of both
catalysts in that a role of gold is to lower the barrier (i.e.,
catalyze) for oxidative addition by palladium. In this way, the
gold and palladium cocatalysts behave differently from a purely
tandem catalyst system where each catalyst acts on the
substrate separately; instead the gold and palladium act
cooperatively. This understanding of the pathways available in
gold and palladium systems is expected to aid in the predictive
design of future dual-catalytic reactions.
EXPERIMENTAL SECTION
■
General. All chemicals were used as received from
commercial suppliers unless otherwise noted. Precatalysts
PPh3AuCl and Pd2dba3 were purchased from Strem Chemical
Co. Dichlorobis(triphenylphosphine)palladium(II) was pur-
chased from Alfa Aesar. Methylene chloride-d2 was dried over
CaH2, degassed using three freeze, pump, thaw cycles, and
vacuum-transferred prior to use. All gold and palladium
substrate rearrangement and control reactions were performed
in an inert atmosphere (N2) box unless otherwise noted.
Analytical and preparatory TLC was performed on Merck F250
TLC plates. Flash chromatography was performed on Dynamic
Absorbents 43-60 μm silica gel. 1H NMR spectra were acquired
on either the GN-500, AV-600, or Bruker CRYO-500
spectrometers. 13C NMR spectra were acquired on the Bruker
CRYO-500 spectrometer. Chemical shifts (δ) are reported in
parts per million (ppm) relative to tetramethylsilane and
referenced to the residual protiated solvent peak (1H NMR).
High-resolution mass spectrometry (HRMS) was performed at
the facility operated by the University of California, Irvine.
Synthesis of 1a and 1 b. The two allenoate substrates 1a
and 1b were prepared according to the previously reported
Figure 1. Cyclic voltammograms of (a) PPh3AuCl, (b) PPh3AuCl +
AgOTf, and (c) Pd2dba3, measured in dry CH2Cl2 (0.1 M TBAP)
under air-free conditions with a scan rate of 0.1 V s−1.
butenolide product 2a was observed by 1H NMR spectroscopy
(Table 1, entries 2 and 3). When we treated allenoate 1a with 5
mol % of Pd2dba3 in the presence of 5 mol % of AgOTf (Table
1, entry 4), the yield of butenolide product 2a was lowered as
compared to the Au/Pd dual-catalytic system (Table 1, entry
1). Thus, a kinetic role for the Au/Pd cocatalyst was
demonstrated definitely.
A “silver effect” in gold catalysis, where residual silver from
the salt metathesis reagent AgOTf is noninnocent, was reported
early by our group5a and then by others.28 To probe if silver
was noninnocent under the conditions of the dual-catalytic
rearrangement, a reaction with a silver-free metathesis reagent
was explored. When employing NaBArF (instead of AgOTf),
the rate of butenolide product 2a formation was lowered
somewhat (entry 5) compared to the gold/palladium catalytic
system (entry 1). This lower yield could be due to the absence
of Ag(I) or to the effect of the changing the counterion to BArF
instead of OTf.
literature5 and H NMR spectra were consistent with that
1
literature.5
Synthesis of Substrate 8a. In the glovebox, p-methyl
benzoic acid (90 mg, 0.66 mmol) was dissolved in dry DMF
(2.0 mL) and NaHCO3 (92.0 mg, 1.10 mmol) was added. The
solution was stirred for 2 min, then 1-bromo-2-methyl-1-
propene (0.11 mL, 1.1 mmol) was added dropwise and the
resulting solution was allowed to stir overnight. Next, the
solution was concentrated in vacuo and purified by column
chromatography (35:1 hexanes/EtOAc). The resulting oil was
put under vacuum (∼12 mTorr) overnight, which afforded 8a
1
(65.0 mg, 51%) as a clear, colorless oil. H NMR (CDCl3, 500
MHz): δ 1.85 (s, 3H), 2.42 (s, 3H), 4.74 (s, 2H), 4.99 (m, 1H),
5.08 (m, 1H), 7.25 (d, J = 8.0 Hz, 2H), 7.97 (dd, J = 1.7, 6.5
Hz, 2H). 13C NMR (CDCl3, 125 MHz): δ 19.6, 21.7, 68.0,
112.8, 127.5, 129.2, 129.7, 140.2, 143.7, 166.4. HRMS (ESI):
[M + Na]+ calcd for C12H15O2, 191.1072; found, 191.1070.
Synthesis of Substrate 8b. In the glovebox, p-methoxy
benzoic acid (2.00 g, 1.32 mmol) was dissolved in dry DMF
(4.1 mL) and NaHCO3 (332 mg, 3.95 mmol) was added. The
solution was stirred for 2 min, then allyl bromide (0.33 mL, 3.9
mmol) was added dropwise and the resulting solution was
allowed to stir overnight. Next, the solution was concentrated
in vacuo and purified by column chromatography (20:1
hexanes/EtOAc). The resulting oil was put under vacuum
(∼12 mTorr) overnight, which afforded 8b (210.0 mg, 84%) as
a clear, yellow oil. The 1H NMR spectrum was consistent with
that previously reported in the literature.29
CONCLUSION
■
The summary of mechanistic crossover studies and control
reactions are consistent with a “gold cyclization first”
mechanistic hypothesis. Specifically, (1) gold activates the
substrate prior to oxidative addition by palladium, (2) gold acts
as a carbophilic rather than oxophilic Lewis acid, (3) competing
olefin isomerization is avoided, (4) gold participates beyond the
first turnover and therefore does not serve simply to generate
the active palladium catalyst, and (5) single-electron transfer is
not involved.
626
dx.doi.org/10.1021/cs400641k | ACS Catal. 2014, 4, 622−629

ACS Catalysis
Research Article
Crossover Experiment between 1a and 1b (Condition
A: Catalytic). In a glovebox, Pd2dba3 (2.3 mg, 0.0025 mmol),
PPh3 (0.7 mg, 0.003 mmol), 1a (9.7 mg, 0.049 mmol), and 1b
(8.30 mg, 0.049 mmol) were weighed in separate dram vials.
Dry CD2Cl2 (0.1 mL) was added via syringe to the vial
containing 1a. The solution was transferred to the vial
containing 1b. Dry CD2Cl2 (0.1 mL) was used as a rinse.
Next, dry CD2Cl2 (0.1 mL) was added via syringe to the dram
vial containing Pd2dba3, and this solution was then added to the
dram vial containing PPh3. Next, dry CD2Cl2 (0.05 mL) was
used as a rinse and added to the Pd2dba3/PPh3 vial. The
Pd2dba3/PPh3 solution was added to the dram vial containing
the substrates 1a and 1b. Again, CD2Cl2 (0.05 mL) was used as
a rinse. The final solution was transferred to a J. Young tube.
Dry CD2Cl2 (0.1 mL) was used as a rinse and added to the J.
Young tube. The reaction was monitored by GCMS, and after
25 min, peaks of two allenoates 1a and 1b were observed by
GCMS at RT of 13.83 and 11.31, respectively, and no peaks
from crossover products 1a′ and 1b′ were detected by GCMS.
Crossover Experiment between 1a and 1b (Condition
B: High Loading). In a glovebox, Pd2dba3 (8 mg, 0.009
mmol), PPh3 (2 mg, 0.009 mmol), 1a (5.9 mg, 0.030 mmol),
and 1b (5 mg, 0.03 mmol) were weighed in separate dram vials.
Dry CD2Cl2 (0.1 mL) was added via syringe to the vial
containing 1a. The solution was transferred to the vial
containing 1b. Dry CD2Cl2 (0.1 mL) was used as a rinse.
Next, dry CD2Cl2 (0.1 mL) was added via syringe to the dram
vial containing Pd2dba3, and this solution was then added to the
dram vial containing PPh3. Next, dry CD2Cl2 (0.05 mL) was
used as a rinse and added to the Pd2dba3/PPh3 vial. The
Pd2dba3/PPh3 solution was added to the dram vial containing
the substrates 1a and 1b. Again, CD2Cl2 (0.05 mL) was used as
a rinse. The final solution was transferred to a J. Young tube.
Dry CD2Cl2 (0.1 mL) was used as a rinse and added to the J.
Young tube. The reaction was monitored by GCMS and after
25 min, the peaks of two allenoates 1a and 1b, the peaks of
their crossover products 1a′ and 1b′, and peaks of two
isomerized products 1a″ (isomer of 1a′) and 1b″ (isomer of
1b) were observed by GCMS at RT of 13.66, 11.20, 14.62,
1 0 . 0 5 , 1 7 . 3 2 , a n d 1 4 . 2 5 i n t h e r a t i o o f
0.10:0.34:0.30:0.19:0.03:0.04, respectively.
isomerization products 8a″ and 8b″ (0.42:0.14:0.11:0.33,
respectively). Products were determined based on comparison
1
of the H NMR spectrum to known literature compounds.
Control Experiments with 1a (Table 1). Entry 1. With
PPh3AuCl/AgOTf, Pd2dba3. In a glovebox, AgOTf (0.6 mg,
0.002 mmol), PPh3AuCl (1.2 mg, 0.0025 mmol), Pd2dba3 (2.3
mg, 0.0025 mmol), and 1a (9.7 mg, 0.049 mmol) were weighed
in separate dram vials. Dry CD2Cl2 (0.1 mL) was added via
syringe to the vial containing PPh3AuCl. The solution was
transferred to the vial containing AgOTf. Dry CD2Cl2 (0.05
mL) was used as a rinse. Next, the solution was transferred to
the vial containing 1a. Dry CD2Cl2 (0.1 mL) was used as a
rinse. The solution was transferred to the vial containing
Pd2dba3. Dry CD2Cl2 (0.1 mL) was used as a rinse. The final
solution was transferred to a J. Young tube. Dry CD2Cl2 (0.15
mL) was used as a rinse and added to the J. Young tube, and
then mesitylene (6 mg, 7 μL, 0.05 mmol) was added as an
1
internal standard. The reaction was monitored by H NMR
spectroscopy after 10, 25, 60, and 90 min intervals and showed
a 24, 69, 93, and 100% conversion, respectively (determined
based on the ratio of butenolide product 2a to mesitylene).
Entry 2. With PPh3AuCl/AgOTf. In a glovebox, AgOTf (0.6
mg, 0.002 mmol), PPh3AuCl (1.2 mg, 0.0025 mmol), and 1a
(9.7 mg, 0.049 mmol) were weighed in separate dram vials. Dry
CD2Cl2 (0.1 mL) was added via syringe to the vial containing
AgOTf. The solution was transferred to the vial containing
PPh3AuCl. Dry CD2Cl2 (0.1 mL) was used as a rinse. Next, the
solution was transferred to the vial containing 1a. Dry CD2Cl2
(0.1 mL) was used as a rinse. The final solution was transferred
to a J. Young tube. Dry CD2Cl2 (0.2 mL) was used as a rinse
and added to the J. Young tube, and then mesitylene (6 mg, 7
μL, 0.05 mmol) was added as an internal standard. The
reaction was monitored by 1H NMR spectroscopy after 10, 25,
60, and 90 min intervals and showed a zero conversion in each
case (determined based on the ratio of butenolide product 2a
to mesitylene).
Entry 3. With Pd2dba3. In a glovebox, Pd2dba3 (2.3 mg,
0.0025 mmol) and 1a (9.7 mg, 0.049 mmol) were weighed in
separate dram vials. Dry CD2Cl2 (0.1 mL) was added via
syringe to the vial containing Pd2dba3. Next, the solution was
transferred to the vial containing 1a. Dry CD2Cl2 (0.1 mL) was
used as a rinse. The final solution was transferred to a J. Young
tube. Dry CD2Cl2 (0.3 mL) was used as a rinse and added to
Crossover Experiment between 8a and 8b. In a
glovebox, PPh3AuCl (1.2 mg, 0.0025 mmol), AgOTf (0.6 mg,
0.003 mmol), Pd2dba3 (2.3 mg, 0.0025 mmol), 8a (4.8 mg,
0.025 mmol), and 8b (4.8 mg, 0.025 mmol) were weighed in
separate dram vials. Dry CD2Cl2 (0.1 mL) was added via
syringe to the vial containing 8a. The solution was transferred
to the vial containing 8b. CD2Cl2 (0.05 mL) was used as a
rinse. Next, CD2Cl2 (0.1 mL) was added via syringe to the
dram vial containing PPh3AuCl, and this solution was then
added to the dram vial containing AgOTf. Next, CD2Cl2 (0.05
mL) was used as a rinse and added to the PPh3AuCl/AgOTf
vial. The PPh3AuCl/AgOTf solution was added to the dram
vial containing the substrates 8a and 8b. Again, CD2Cl2 (0.05
mL) was used as a rinse. To the dram vial containing Pd2dba3
was added the solution containing substrate and PPh3AuCl/
AgOTf. The vial previously containing Pd was rinsed with
CD2Cl2 (0.1 mL) and added to the final solution. The final
solution was transferred to a J. Young tube. CD2Cl2 (0.05 mL)
was used as a rinse and added to the J. Young tube. The
1
the J. Young tube. The reaction was monitored by H NMR
spectroscopy after 10, 25, 60, and 90 min intervals and showed
a zero conversion in each case (determined based on the ratio
of butenolide product 2a to allenoate 1a).
Entry 4. With Pd2dba3/AgOTf. In a glovebox, AgOTf (0.6
mg, 0.002 mmol), Pd2dba3 (2.3 mg, 0.0025 mmol), and 1a (9.7
mg, 0.049 mmol) were weighed in separate dram vials. Dry
CD2Cl2 (0.1 mL) was added via syringe to the vial containing
AgOTf. The solution was transferred to the vial containing
Pd2dba3. Dry CD2Cl2 (0.1 mL) was used as a rinse. Next, the
solution was transferred to the vial containing 1a. Dry CD2Cl2
(0.1 mL) was used as a rinse. The final solution was transferred
to a J. Young tube. Dry CD2Cl2 (0.2 mL) was used as a rinse
and added to the J. Young tube, and then mesitylene (6 mg, 7
μL, 0.05 mmol) was added as an internal standard. The
reaction was monitored by 1H NMR spectroscopy after 10, 25,
60, and 90 min intervals and showed a 22, 23, 24, and 28%
conversion, respectively (determined based on the ratio of
butenolide product 2a to mesitylene).
1
reaction was monitored by H NMR spectroscopy after 1.5 h
and after 24 h. After 1.5 h, no reaction had occurred. After 24 h,
starting materials 8a and 8b were observed, along with olefin
627
dx.doi.org/10.1021/cs400641k | ACS Catal. 2014, 4, 622−629

ACS Catalysis
Research Article
Entry 5. With PPh3AuCl/NaBArF, Pd2dba3. In a glovebox,
PPh3AuCl (0.7 mg, 0.002 mmol), NaBArF (1.8 mg, 0.0015
mmol), Pd2dba3 (1.8 mg, 0.0015 mmol), and 1a (6 mg, 0.03
mmol) were weighed in separate dram vials. Dry CD2Cl2 (0.1
mL) was added via syringe to the vial containing PPh3AuCl.
The solution was transferred to the vial containing NaBArF.
Dry CD2Cl2 (0.05 mL) was used as a rinse. Next, the solution
was transferred to the vial containing 1a. Dry CD2Cl2 (0.1 mL)
was used as a rinse. The solution was transferred to the vial
containing Pd2dba3. Dry CD2Cl2 (0.1 mL) was used as a rinse.
The final solution was transferred to a J. Young tube. Dry
CD2Cl2 (0.15 mL) was used as a rinse and added to the J.
Young tube, and then mesitylene (4 mg, 4 μL, 0.03 mmol) was
added as an internal standard. The reaction was monitored by
1H NMR spectroscopy after 10, 25, 60, and 90 min intervals
and showed a 23, 28, 40, and 47% conversion, respectively
(determined based on the ratio of butenolide product 2a to
mesitylene).
Entry 6. With Pd2dba3/NaBArF. In a glovebox, NaBArF (1.8
mg, 0.0015 mmol), Pd2dba3 (1.8 mg, 0.0015 mmol), and 1a (6
mg, 0.03 mmol) were weighed in separate dram vials. Dry
CD2Cl2 (0.1 mL) was added via syringe to the vial containing
NaBArF. The solution was transferred to the vial containing
Pd2dba3. Dry CD2Cl2 (0.1 mL) was used as a rinse. Next, the
solution was transferred to the vial containing 1a. Dry CD2Cl2
(0.1 mL) was used as a rinse. The final solution was transferred
to a J. Young tube. Dry CD2Cl2 (0.2 mL) was used as a rinse
and added to the J. Young tube, and then mesitylene (4 mg, 4
μL, 0.03 mmol) was added as an internal standard. The
reaction was monitored by 1H NMR spectroscopy after 10, 25,
60, and 90 min intervals and showed a 1, 2, 5, and 9%
conversion, respectively (determined based on the ratio of
butenolide product 2a to mesitylene).
mL) was added via syringe to the vial containing Pd2(allyl)2Cl2.
The solution was transferred to the vial containing AgOTf. Dry
CD2Cl2 (0.05 mL) was used as a rinse. The solution was
transferred to the vial containing dba. Dry CD2Cl2 (0.1 mL)
was used as a rinse. Next, the solution was transferred to the
vial containing 1a. Dry CD2Cl2 (0.1 mL) was used as a rinse.
The final solution was transferred to a J. Young tube. Dry
CD2Cl2 (0.15 mL) was used as a rinse and added to the J.
Young tube, and then mesitylene (6 mg, 7 μL, 0.05 mmol) was
added as an internal standard. The reaction was monitored by
1H NMR spectroscopy after 10, 25, 60, and 90 min intervals
and showed a 0, 6, 12, and 13% conversion, respectively
(determined based on the ratio of butenolide product 2a to
mesitylene).
Entry 10. With Pd2(allyl)2Cl2/AgOTf, PPh3. In a glovebox,
AgOTf (1.3 mg, 0.0051 mmol), Pd2(allyl)2Cl2 (0.9 mg, 0.002
mmol), PPh3 (0.7 mg, 0.002 mmol), and 1a (9.7 mg, 0.049
mmol) were weighed in separate dram vials. Dry CD2Cl2 (0.1
mL) was added via syringe to the vial containing Pd2(allyl)2Cl2.
The solution was transferred to the vial containing AgOTf. Dry
CD2Cl2 (0.05 mL) was used as a rinse. The solution was
transferred to the vial containing PPh3. Dry CD2Cl2 (0.1 mL)
was used as a rinse. Next, the solution was transferred to the
vial containing 1a. Dry CD2Cl2 (0.1 mL) was used as a rinse.
The final solution was transferred to a J. Young tube. Dry
CD2Cl2 (0.15 mL) was used as a rinse and added to the J.
Young tube, and then mesitylene (6 mg, 7 μL, 0.05 mmol) was
added as an internal standard. The reaction was monitored by
1H NMR spectroscopy after 10, 25, 60, and 90 min intervals
and showed a 10, 14, 17, and 28% conversion, respectively
(determined based on the ratio of butenolide product 2a to
mesitylene).
Entry 11. With Pd(PPh3)2Cl2/AgOTf, Pd2dba3. In a glovebox,
AgOTf (0.6 mg, 0.002 mmol), Pd(PPh3)2Cl2 (0.5 mg, 0.001
mmol), Pd2dba3 (1.7 mg, 0.0019 mmol), and 1a (9.7 mg, 0.049
mmol) were weighed in separate dram vials. Dry CD2Cl2 (0.1
mL) was added via syringe to the vial containing Pd(PPh3)2Cl2.
The solution was transferred to the vial containing AgOTf. Dry
CD2Cl2 (0.05 mL) was used as a rinse. Next, the solution was
transferred to the vial containing 1a. Dry CD2Cl2 (0.1 mL) was
used as a rinse. The solution was transferred to the vial
containing Pd2dba3. Dry CD2Cl2 (0.1 mL) was used as a rinse.
The final solution was transferred to a J. Young tube. Dry
CD2Cl2 (0.15 mL) was used as a rinse and added to the J.
Young tube, and then mesitylene (6 mg, 7 μL, 0.05 mmol) was
added as an internal standard. The reaction was monitored by
1H NMR spectroscopy after 10, 25, 60, and 90 min intervals
and showed a 7, 8, 8, and 9% conversion, respectively
(determined based on the ratio of butenolide product 2a to
mesitylene).
Entry 7. With NaBArF. In a glovebox, NaBArF (1.8 mg,
0.0015 mmol) and 1a (6 mg, 0.03 mmol) were weighed in
separate dram vials. Dry CD2Cl2 (0.1 mL) was added via
syringe to the vial containing NaBArF. Next, the solution was
transferred to the vial containing 1a. Dry CD2Cl2 (0.1 mL) was
used as a rinse. The final solution was transferred to a J. Young
tube. Dry CD2Cl2 (0.3 mL) was used as a rinse and added to
1
the J. Young tube. The reaction was monitored by H NMR
spectroscopy after 10, 25, 60, and 90 min intervals and showed
a zero conversion in each case (determined based on the ratio
of butenolide product 2a to mesitylene).
Entry 8. With Pd2(allyl)2Cl2/AgOTf. In a glovebox, AgOTf
(1.3 mg, 0.0051 mmol), Pd2(allyl)2Cl2 (0.9 mg, 0.002 mmol),
and 1a (9.7 mg, 0.049 mmol) were weighed in separate dram
vials. Dry CD2Cl2 (0.1 mL) was added via syringe to the vial
containing Pd2(allyl)2Cl2. The solution was transferred to the
vial containing AgOTf. Dry CD2Cl2 (0.05 mL) was used as a
rinse. Next, the solution was transferred to the vial containing
1a. Dry CD2Cl2 (0.1 mL) was used as a rinse. The final solution
was transferred to a J. Young tube. Dry CD2Cl2 (0.25 mL) was
used as a rinse and added to the J. Young tube, and then
mesitylene (6 mg, 7 μL, 0.05 mmol) was added as an internal
ASSOCIATED CONTENT
* Supporting Information
■
S
Details of experimental procedures and NMR spectra. This
material is available free of charge via the Internet at http://
pubs.acs.org.
1
standard. The reaction was monitored by H NMR spectros-
copy after 10, 25, 60, and 90 min intervals and showed a 7, 13,
22, and 23% conversion, respectively (determined based on the
ratio of butenolide product 2a to mesitylene).
Entry 9. With Pd2(allyl)2Cl2/AgOTf, dba. In a glovebox,
AgOTf (1.3 mg, 0.0051 mmol), Pd2(allyl)2Cl2 (0.9 mg, 0.002
mmol), dba (1.8 mg, 0.0075 mmol), and 1a (9.7 mg, 0.049
mmol) were weighed in separate dram vials. Dry CD2Cl2 (0.1
AUTHOR INFORMATION
Corresponding Author
*E-mail: blums@uci.edu.
■
Notes
The authors declare no competing financial interest.
628
dx.doi.org/10.1021/cs400641k | ACS Catal. 2014, 4, 622−629

ACS Catalysis
Research Article
(19) Lorenz, I.-P.; Krinninger, C.; Wilberger, R.; Bobka, R.;
ACKNOWLEDGMENTS
■
Piotrowski, H.; Warchhold, M.; Noth, H. J. Organomet. Chem. 2005,
̈
We thank for funding the NSF for a CAREER award (CHE-
0748312) to S.A.B., and the NIH (1R01GM098512-01).
690, 1986.
(20) Yamamoto, Y. J. Org. Chem. 2007, 72, 7817.
(21) For a proposed activation by bidentate binding of Au(I)
simultaneously to an alkyne and an aziridine, see: Kern, N.; Blanc, A.;
Miaskiewicz, S.; Robinette, M.; Weibel, J.-M.; Pale, P. J. Org. Chem.
2012, 77, 4323.
REFERENCES
■
(1) Shenvi, R. A.; O’Malley, D. P.; Baran, P. S. Acc. Chem. Res. 2008,
42, 530.
(2) Vaxelaire, C.; Winter, P.; Christmann, M. Angew. Chem., Int. Ed.
(22) Amatore, C.; Jutand, A.; Meyer, G.; Mottier, L. Chem.Eur. J.
1999, 5, 466.
2011, 50, 3605.
(23) Anslyn, E. V.; Dougherty, D. A. Modern Physical Organic
Chemistry; University Science Books: Sausalito, CA, 2006; p 76.
(24) Trost, B. M.; Van Vranken, D. L. Chem. Rev. 1996, 96, 395.
(25) Amatore and co-workers demonstrated that an analogous
complex, (π-allyl)Pd(PPh3)2OAc, does not exist as a free ion pair in
THF. See ref 22.
(26) Weber, D.; Gagne, M. R. Chem. Commun. 2011, 47, 5172.
́
(27) Amatore, C.; Broeker, G.; Jutand, A.; Khalil, F. J. Am. Chem. Soc.
1997, 119, 5176.
(3) Sammis, G. M.; Danjo, H.; Jacobsen, E. N. J. Am. Chem. Soc.
2004, 126, 9928.
(4) Allen, A. E.; MacMillan, D. W. C. Chem. Sci. 2012, 3, 633.
(5) For dual-catalytic gold and palladium reactions, see: (a) Shi, Y.;
Peterson, S. M.; Haberaecker, W. W., III; Blum, S. A. J. Am. Chem. Soc.
2008, 130, 2168. (b) Shi, Y.; Roth, K. E.; Ramgren, S. D.; Blum, S. A. J.
Am. Chem. Soc. 2009, 131, 18022. (c) Lauterbach, T.; Livendahl, M.;
́
Rosellοn, A.; Espinet, P.; Echavarren, A. M. Org. Lett. 2010, 12, 3006.
(d) Jones, L. A.; Sanz, S.; Laguna, M. Catal. Today 2007, 122, 403.
(e) Panda, B.; Sarkar, T. K. Chem. Commun. 2010, 46, 3131.
(f) Hirner, J. J.; Roth, K. E.; Shi, Y.; Blum, S. A. Organometallics 2012,
31, 6843.
(6) For palladium-catalyzed reactions to functionalize stoichiometric
organogold reagents (i.e., monocatalytic reactions), see: (a) Shi, Y.;
Ramgren, S. D.; Blum, S. A. Organometallics 2009, 28, 1275.
(28) Wang, D.; Cai, R.; Sharma, S.; Jirak, J.; Thummanapelli, S. K.;
Akhmedov, N. G.; Zhang, H.; Liu, X.; Petersen, J. L.; Shi, X. J. Am.
Chem. Soc. 2012, 134, 9012.
(29) Corey, E. J.; Guzman-Perez, A.; Noe, M. C. J. Am. Chem. Soc.
1995, 117, 10805.
(b) Hashmi, A. S. K.; Lothschutz, C.; Dopp, R.; Rudolph, M.;
̈
̈
Ramamurthi, T. D.; Rominger, F. Angew. Chem., Int. Ed. 2009, 48,
8243. (c) Hashmi, A. S. K.; Dopp, R.; Lothschutz, C.; Rudolph, M.;
̈
̈
Riedel, D.; Rominger, F. Adv. Synth. Catal. 2010, 352, 1307.
(d) Weber, D.; Gagne,
́
M. R. Chem. Commun. 2011, 47, 5172.
-Varela, M.; Sarandeses, L. A.; Sestelo, J. P.
́
(e) Pena-Lοpez, M.; Ayan
Chem.Eur. J. 2010, 16, 9905.
́
̃
(7) Jacobsen, E. N. Acc. Chem. Res. 2000, 33, 421.
(8) Liu, Z.; Henderson, J. A.; Sasaki, T.; Kishi, Y. J. Am. Chem. Soc.
2009, 131, 16678.
(9) (a) Trost, B. M.; Luan, X. J. Am. Chem. Soc. 2011, 133, 1706.
(b) Trost, B. M.; Luan, X.; Miller, Y. J. Am. Chem. Soc. 2011, 133,
12824.
(10) (a) Wu, H.; He, Y.-P.; Gong, L.-Z. Adv. Synth. Catal. 2012, 354,
́
975. (b) Weber, D.; Gagne, M. R. Chem. Commun. 2011, 47, 5172.
(11) (a) Nakao, Y.; Yada, A.; Ebata, S.; Hiyama, T. J. Am. Chem. Soc.
2007, 129, 2428. (b) Koester, D. C.; Kobayashi, M.; Werz, D. B.;
Nakao, Y. J. Am. Chem. Soc. 2012, 134, 6547.
(12) Graham, T. J. A.; Shields, J. D.; Doyle, A. G. Chem. Sci. 2011, 2,
980.
(13) Nicewicz, D. A.; MacMillan, D. W. C. Science 2008, 322, 77.
(14) Skucas, E.; MacMillan, D. W. C. J. Am. Chem. Soc. 2012, 134,
9090.
(15) DiRocco, D. A.; Rovis, T. J. Am. Chem. Soc. 2012, 134, 8094.
(16) Burns, N. Z.; Written, M. R.; Jacobsen, E. N. J. Am. Chem. Soc.
2011, 133, 14578.
(17) (a) Gorin, D. J.; Sherry, B. D.; Toste, F. D. Chem. Rev. 2008,
108, 3351. (b) Liu, L.-P.; Xu, B.; Mashuta, M. S.; Hammond, G. B. J.
Am. Chem. Soc. 2008, 130, 17642. (c) Dopp, R.; Lothschutz, C.;
̈
̈
Wurm, T.; Pernpointner, M.; Keller, S.; Rominger, F.; Hashmi, A. S. K.
Organometallics 2011, 30, 5894. (d) Hashmi, A. S. K.; Lothschutz, C.;
̈
Dopp, R.; Rudolph, M.; Ramamurthi, T. D.; Rominger, F. Angew.
̈
Chem., Int. Ed. 2009, 48, 8243. (e) Hashmi, A. S. K.; Lothschutz, C.;
Dopp, R.; Ackermann, M.; De Buck Becker, J.; Rudolph, M.; Scholz,
C.; Rominger, F. Adv. Synth. Catal. 2012, 354, 133. (f) Furstner, A.
Chem. Soc. Rev. 2009, 38, 3208. (g) Hashmi, A. S. K. Angew. Chem., Int.
Ed. 2010, 49, 5232.
̈
̈
̈
(18) For examples of Au(I) as an oxophilic Lewis acid, see:
(a) Vidhani, D. V.; Cran, J. W.; Krafft, M. E.; Manoharan, M.;
Alabugin, I. V. J. Org. Chem. 2013, 78, 2059. (b) Jagdale, A. R.; Youn,
S. W. Eur. J. Org. Chem. 2011, 3904. (c) Dombray, T.; Blanc, A.;
Weibel, J.-M.; Pale, P. Org. Lett. 2010, 12, 5362. (d) Lin, C.-C.; Teng,
T.-M.; Odedra, A.; Liu, R.-S. J. Am. Chem. Soc. 2007, 129, 3798.
629
dx.doi.org/10.1021/cs400641k | ACS Catal. 2014, 4, 622−629