"Silver effect" in gold(I) catalysis: An overlooked important factor

Article abstract of DOI:10.1021/ja303862z

Clear experimental evidence from X-ray photoelectron spectroscopy and ³¹P NMR spectroscopy has been obtained for the first time to confirm that the combination of Ag⁺ cation with [L-Au]⁺ results in the formation of different complexes in solution. Re-evaluation of literature-reported gold-catalyzed reactions revealed a significant difference in the reactivities with and without silver. In extreme cases (more than "rare"), the conventional [L-Au]⁺ catalysts could not promote the reaction without the presence of silver. This investigation has therefore revealed a long-overlooked "silver effect" in gold catalysis and should lead to revision of the actual mechanism.

Full text of DOI:10.1021/ja303862z

Article
pubs.acs.org/JACS
“Silver Effect” in Gold(I) Catalysis: An Overlooked Important Factor
Dawei Wang,† Rong Cai,† Sripadh Sharma,† James Jirak,† Sravan K. Thummanapelli,†
Novruz G. Akhmedov,† Hui Zhang,‡ Xingbo Liu,‡ Jeffrey L. Petersen,† and Xiaodong Shi*,†
†C. Eugene Bennett Department of Chemistry, West Virginia University, Morgantown, West Virginia 26506, United States
‡Mechanical and Aerospace Engineering Department, West Virginia University, Morgantown, West Virginia 26506, United States
S
* Supporting Information
ABSTRACT: Clear experimental evidence from X-ray photo-
electron spectroscopy and 31P NMR spectroscopy has been
obtained for the first time to confirm that the combination of
Ag+ cation with [L-Au]+ results in the formation of different
complexes in solution. Re-evaluation of literature-reported
gold-catalyzed reactions revealed a significant difference in the
reactivities with and without silver. In extreme cases (more
than “rare”), the conventional [L-Au]+ catalysts could not
promote the reaction without the presence of silver. This
investigation has therefore revealed a long-overlooked “silver
effect” in gold catalysis and should lead to revision of the actual mechanism.
of reported and yet to be reported results in regard to whether
the reaction involves “actual” gold catalysis.
INTRODUCTION
■
Homogeneous gold catalysis is considered to be one of the
most important developments in organic synthesis during the
past decade.1 It has been generally accepted that [L-Au]+ is the
actual π-acid toward alkyne and alkene activation in Au(I)-
promoted reactions. Currently, the dominant approach for the
preparation of [L-Au]+ is through the treatment of [L-Au]Cl
with the corresponding silver salt to form AgCl. Since the AgCl
is not involved in the reaction, it has been assumed that
removing the AgCl or keeping it in the reaction mixture should
not make a difference, giving “identical” results (Scheme 1).
RESULTS AND DISCUSSION
■
Our interest in gold catalysis was initiated by studies of the use
of 1,2,3-triazole ligands to adjust transition-metal reactivity.3
These efforts led to the recent discovery of the air- and
moisture-stable triazole gold(I) compounds (TA-Au), which
gave interesting reactivity that could not be offered by simple
[L-Au]+. Some recent discoveries about the new reactivity of
the TA-Au catalysts include improved thermal stability,4 better
substrate stability,5 and excellent chemo- and stereoselectivity6
toward alkyne activation. On the basis of these successes, we
investigated the 1,2-migration of propargyl esters7 with the TA-
Au catalyst.
Scheme 1
In view of the importance of α-hydroxy methyl ketones
(Scheme 2A) in medicinal chemistry8 and the ready availability
of the enantiomerically pure propargyl alcohols, our initial
attempt was to investigate the possibility of direct alkyne
hydration to form the ketone with retention of stereochemistry
at the hydroxy position. However, after screening a series of the
Au(I) complexes, we found that no catalyst, not even TA-Au,
promoted this reaction effectively even at 80 °C. We then
looked into an alternative approach. It is known that propargyl
esters can undergo 1,2-migration with terminal alkynes.9
Therefore, it should be possible to trap the oxonium cation
with water and reach the desired α-hydroxy methyl ketone
through simple steps (Scheme 2B). However, according to the
literature, simple gold catalysts [L-Au]+ cause rapid epimeriza-
tion (within a few minutes). Therefore, we became interested
Although it has been reported in a few cases that the presence
of silver helps the reactivity of gold catalysts,2 it has still been
largely assumed that all of these reactions are catalyzed by the
gold cation (with silver playing a “supporting role”) as long as
the silver salt itself cannot catalyze the reaction. Herein we
report for the first time experimental evidence (based on
evaluations of 14 different types of previously reported gold-
catalyzed reactions) that gold catalysis is not simply activated
by [L-Au]+ as previously assumed. In some cases, absolutely no
reaction occurs with Ag-free [L-Au]+ catalysts. With the clear
experimental support described in this work, this long-
overlooked “silver effect” should lead to serious reconsideration
Received: April 22, 2012
Published: May 7, 2012
© 2012 American Chemical Society
9012
dx.doi.org/10.1021/ja303862z | J. Am. Chem. Soc. 2012, 134, 9012−9019

Journal of the American Chemical Society
Article
a b
,
Scheme 2. Proposed Asymmetric Synthesis of α-Hydroxy
Table 2. Reaction Substrate Scope
Ketones
in exploring whether the chemoselective TA-Au would
promote this reaction with good retention of chirality.
Propargyl ester 1a was prepared and reacted with different
gold catalysts under various conditions. As shown in Table 1,
a
General reaction conditions: 1 (0.25 mmol, 1.0 equiv) and TA-Au-1
(1.0 mol %) in acetone (2.5 mL) with H2O (3.0 equiv) at rt for 4−10
a
Table 1. Gold(I)-Catalyzed Propargyl Ester Hydration
b
h. The reactions were monitored by TLC. Isolated yields are shown.
Scheme 3. Alternative Reactivity Offered by the TA-Au
Catalyst
b
c
entry
cat
solvent
time (h)
yield (%)
1
2
[(PPh3)Au]+TfO−
dioxane
dioxane
dioxane
dioxane
dioxane
DCM
12
12
12
12
12
12
12
12
12
12
4
0
[(PPh3)Au]+SbF6
0
−
[(IPr)Au]+SbF6
0
−
3
4
TA-Au-1
TA-Au-2
TA-Au-1
TA-Au-1
TA-Au-1
TA-Au-1
TA-Au-1
TA-Au-1
TA-Au-1
87
85
32
<5
24
<5
11
89
5
The stereochemistry was evaluated by comparing the
enantiomeric excess (ee) values for the products and the
enantiomerically enriched propargyl ester starting materials. As
indicated in Scheme 4, excellent retention of chirality was
obtained with the TA-Au catalyst.
6
7
THF
8
MeOH
DMF
9
10
11
12
CH3NO2
CH3CN
acetone
Scheme 4. Excellent Retention Chirality with the TA-Au
Catalyst
d
4
99 (96)
a
General reaction conditions: 1a (0.25 mmol, 1.0 equiv) and 2a (1.0
mol %) in solvent (2.5 mL) with H2O (3.0 equiv) at room
temperature (rt) for 4−12 h. The reactions were monitored by thin-
b
c
d
layer chromatography (TLC). Type II [L-Au]+. NMR yields. The
isolated yield is given in parentheses.
both TA-Au-1 and TA-Au-2 effectively promoted the reaction.
Screening of the conditions revealed acetone to be the optimal
solvent, giving the desired α-acetoxy ketone in nearly quantitive
yield (1% loading). It should be noted that the [PPh3-Au]+ and
[IPr-Au]+ complexes could NOT promote this reaction at all
(entries 1−3), and most of the starting material 1a was
recovered. This result highlighted the strength of the triazole-
gold catalyst: a simple modification resulting in high efficiency
and new reactivity. The reaction substrate scope is shown in
Table 2.
As indicated in Table 2, TA-Au catalyst was highly effective
for this transformation, giving the desired products in excellent
yields. One good example that highlighted the unique reactivity
of the TA-Au catalyst compared with [L-Au]+ was the reaction
of 1,4-enyne 1j (Scheme 3). As reported in the literature,
treating 1j with PPh3AuCl/AgSbF6 gave the corresponding
cyclopentenone through a “carbene-like” intermediate.9 In
contrast, simply switching the catalyst to TA-Au-1 successfully
altered the reaction path to give 2j in excellent yield.
While we were excited to find that TA-Au offers new
reactivity relative to [L-Au]+ for this interesting reaction, Sahoo
and co-workers reported the same transformation using the
PPh3AuCl/AgSbF6 mixtures.10 This result was shocking, since
we had shown that [L-Au]+ could not promote this reaction
(Table 1, entries 1−3). These conflicting results were
significant and required further investigation. A direct
comparison of several gold catalysts, with or without the
presence of AgCl, was then carried out (Table 3).
As indicated in Table 3, entry 2, under the optimal
conditions reported by Sahoo (dioxane as solvent, 3 equiv of
H2O), none of the tested silver salts could catalyze the reaction,
and a majority of 1a was recovered. The catalysts [(PPh3)-
Au]+A−, which were prepared from Celite filtration of the
9013
dx.doi.org/10.1021/ja303862z | J. Am. Chem. Soc. 2012, 134, 9012−9019

Journal of the American Chemical Society
Article
a
To verify the difference between Celite filtration and paper
filtration, X-ray photoelectron spectroscopy (XPS) was applied.
The spectra (summarized in Scheme 6) show that gold was
Table 3. Silver Effect in Alkyne Hydration
Scheme 6. XPS Spectra of (PPh3)AuCl/AgSbF6 Mixtures
[(PPh3)AuCl:AgSbF6 = 1:1.5] after Filtration through (A)
Regular Filter Paper and (B) Celite
yield
b
ee
entry
catalyst
1% TA-Au-1
conditions
(%)
(%)
1
2
3
4
5
6
5 h
87
0
90
−
−
0
c
10% AgA
2% [(PPh3)Au]+A−
12 h
12 h
12 h
8 h
c
0
2% (PPh3)AuCl/AgOTf
2% (PPh3)AuCl/AgSbF6
2% (PPh3)AuCl/AgOTf
90
97
92
0
5 h with N2
protection
90
2% [(PPh3)Au]+ SbF6 + 2% 12 h
AgSbF6
91
−
−
7
a
General reaction conditions: 1a (1.0 equiv) and catalyst in 2.5 mL of
dioxane/water (H2O, 3.0 equiv) at rt. Reactions were monitored by
b
c
−
−
TLC. NMR yields. A− = TfO−, SbF6 , BF4 .
observed in both filtrates. The binding energies (Au 4f7/2, 85.5
eV; Au 4f5/2, 89.2 eV) suggested the presence of Au(I) cations
after Celite filtration. The filtrate of (PPh3)AuCl/AgSbF6
through regular filter paper showed the silver signal (Scheme
6A: Ag 3d5/2, 368.7 eV; Ag 3d3/2, 374.7 eV). Interestingly, the
filtrate from Celite did not contain any silver (Scheme 6B),
even when an excess amount of silver salts was used in the
reaction with PPh3AuCl (Au:Ag = 1:1.5). These results
provided unambiguous evidence that the filtrate from Celite
contained only [(PPh3)Au]+ without the presence of any silver.
Similar results were observed with the (IPr)AuCl/AgSbF6
system. These XPS results support the observation that the
presence of silver influences the gold cation reactivity.
PPh3AuCl/Ag+A− mixtures, also could not promote this
reaction (entry 3). However, when the mixtures of PPh3AuCl
and Ag+A− were used directly (without filtration of AgCl), the
reaction worked as reported, though with complete stereo-
chemistry racemization (entries 4 and 5). Careful degassing and
N2 protection helped to avoid the racemization (entry 6).
Interestingly, a mixture of [(PPh3)Au]+TfO− (the inactive
complex obtained from filtration through Celite) and AgSbF6
promoted this reaction effectively (entry 7), which suggested
that the combination of gold and silver made the catalysis
possible.
Meanwhile, the gold(I) complexes are linearly coordinated,
with two ligands binding on the opposite sites of the gold.
Therefore, for complexes like (PR3)AuX, 31P NMR spectros-
copy was ideal for evaluating the electronic nature of the gold
cations in solution. To understand the solution behavior of
these two types of gold samples (with or without silver), 31P
NMR experiments were carried out, and the spectra are shown
Notably, filtration through Celite was a practical procedure
used in our group during the preparation of the TA-Au
complexes. In general, the silver salts were not stable and would
decompose to give silver nanoparticles or a precipitate of Ag2O
over time. This silver decomposition seemed to be accelerated
in the presence of solid AgX, forming a darker solution over
time. Filtration with regular filter paper could not completely
remove the color, even though the filtrate was transparent by
visual judgment. Filtration through Celite allowed the majority
of the [L-Au]+A− go through. This procedure worked for
various ligands. For example, Scheme 5 shows the crystal
structure of the recently obtained complex [(IPr)Au-
BTZ]+TfO− (BTZ = benzotriazole), which was prepared by
the reaction of BTZ and (IPr)AuCl/AgOTf Celite filtrate,
which gave the desired IPr-TA-Au complex in 92% yield.
in Scheme 7. In general, more cationic gold complexes gave 31
P
signals shifted further upfield (PPh3AuCl, 33.8 ppm; TA-Au-1,
30.8 ppm; Scheme 7A,B). [PPh3Au]+TfO− after Celite filtration
exhibited the 31P NMR signal at 27.1 ppm (Scheme 7D),
a
Scheme 7. 31P NMR Spectra of Different Au(I) Samples
Scheme 5. ORTEP Structure of [(IPr)Au-BTZ]+TfO−
Complex (with TfO− Omitted for Clarity)
a
The chemical shifts were calibrated using 85% H3PO4 as an internal
reference in a sealed capillary. The samples were prepared with 0.05 M
gold in CDCl3. The full spectra are shown in the Supporting
Information.
9014
dx.doi.org/10.1021/ja303862z | J. Am. Chem. Soc. 2012, 134, 9012−9019

Journal of the American Chemical Society
Article
suggesting the formation of highly cationic [PPh3Au]+ species.
However, (PPh3)AuCl/AgOTf with no filtration gave a
different chemical shift at 28.1 ppm (Scheme 7C). Clearly,
the presence of silver influenced the gold complexes in solution.
Comparisons of the 31P NMR chemical shifts among the
different samples before and after the Celite filtration are shown
in Table 4. Since the addition of AgOTf was intended to
occur ONLY in this reaction, we extended the investigation to
the alkyne hydration shown in Scheme 8.
Scheme 8. “Silver Effect” in the (IPr)Au-Catalyzed Simple
Alkyne Hydration
a
Table 4. 31P NMR Changes Caused by Celite Filtration
31P NMR chemical shift (ppm)
entry
equiv of AgOTf
before filtration
after filtration
1
2
3
4
5
0.50
0.75
0.90
1.5
31.5
30.2
29.2
28.1
28.1
30.5
29.3
28.7
27.1
27.1
2.0
a
The chemical shifts were calibrated using 85% H3PO4 as an internal
standard (0 ppm) in a sealed capillary. The samples were prepared
with 0.05 M gold in CDCl3.
remove the chloride anion from PPh3AuCl, it was anticipated
that more AgOTf would cause an upfield shift of the 31P signal
due to the formation of more cationic gold (entries 1−4). After
the stoichiometric amount of silver was used, no further
chemical shift changes were observed even with the addition of
excess AgOTf (entries 4 and 5). Notably, different chemical
shifts were observed after the sample was filtered through
Celite to remove the silver. Meanwhile, the addition of AgOTf
to the solution of [(PPh3)Au]+TfO− gave a slow 31P signal shift
from 27.1 ppm BACK to 28.1 ppm (see the detailed spectra in
the Supporting Information). The changes in the chemical
shifts were independent of the amount of AgOTf added: both
0.5 and 1.0 equiv of AgOTf shifted the 31P signals from 27.1 to
28.1 ppm with no further changes. NO chemical shift changes
were observed when AgCl was added to the silver-free
[(PPh3)Au]+TfO− solutions, suggesting that the influence is
from Ag+ instead of AgCl.
It is important to notice here that the TA-Au catalyst was
silver-free and could indeed effectively promote this reaction.
As revealed in the 31P NMR shown in Scheme 7, the TA-Au
complex exhibited the same downfield shift as the silver cation,
which suggests a similar effect from the triazole ligands, making
TA-Au a potentially good alternative for the combination of
gold and silver. Nevertheless, the XPS and NMR experiments
provided unambiguous evidence that the presence of silver
influences the reactivity of the [L-Au]+ cations. This discovery
is critical since it should lead to the revision of the current
assumption regarding the mechanism of homogeneous gold(I)
catalysis, as the actual catalyst might not be the gold alone! The
gold-catalyzed propargyl ester hydration reaction shown in this
study was one good example that highlighted the importance of
this silver effect: [(PPh3)Au]+TfO− could not promote the
reaction at all unless silver was added. One concern was
whether the presence of silver helped the hydration of the
oxonium cation intermediates, which therefore played the
important role in promoting the overall reaction as observed.
This was unlikely. The reaction with TA-Au, which did not
contain silver, led to an effective transformation with higher
reaction rates. This result suggested that the hydration of the
oxonium cation is fast and that silver is not necessary in this
step. Nevertheless, to ensure that this “silver effect” did not
It has been reported by Nolan and co-workers that
(IPr)AuCl/AgSbF6 is the optimal catalyst to promote
challenging internal alkyne hydrations.11 As shown in Scheme
8A, an excellent yield of ketone 4a was obtained when the
(IPr)AuCl/AgSbF6 mixture was used without filtration. When
3a was charged with [(IPr)Au]+ (the Celite filtrate of the same
catalyst mixture), no reaction occurred, even with increased
catalyst loading (5%) and prolonged reaction time (Scheme
8B). In contrast, addition of 2% AgSbF6 caused the catalyst to
become “active” again, giving ketone 4a in excellent yield
though with a longer reaction time (Scheme 8D). Notably,
silver salts alone could not promote this reaction (Scheme 8C).
Therefore, these results demonstrated again that the combina-
tion of gold and silver is the actual active catalyst for this
transformation, revising the previous assumption that [L-Au]+
is the actual catalyst. Reducing the amount of silver salt caused
the reaction rate to decrease. When 0.25 equiv of silver (relative
to gold) was used (Scheme 8F), the reaction could not reach
complete conversion (∼25% of the starting material was
recovered under identical conditions).
These experiments reconfirmed the shocking results shown
in previous case: the addition of silver is no longer an
“optimization” factor for gold catalysis. Instead, [L-Au]+, long
believed to be the effective catalyst, might not be, and instead,
mixtures of both silver and gold “accidentally” provided the
optimal catalysts combinations for effective transformation
(though we are not sure what types of complexes were formed
at this moment). This discovery is critical since it further
separates the overall “gold(I) catalysis” into three different
theoretical categories (Scheme 9): (A) “genuine” gold catalysis,
(B) Au/Ag bimetallic catalysis, and (C) silver-assisted gold
catalysis.
Genuine gold catalysis refers to the type of reactions
promoted by [L-Au]+ with competitive or even better reactivity
than the [L-Au]+/Ag+ mixtures. The type-II bimetallic catalysis
includes the reactions promoted exclusively by the mixture of
gold and silver. One may argue that the group 11 metals have
similar properties with expected differences in reactivity toward
certain transformations. However, as shown in the two cases
discussed above, there are transformations that clearly could
not be promoted by either Au or Ag alone. Only the
9015
dx.doi.org/10.1021/ja303862z | J. Am. Chem. Soc. 2012, 134, 9012−9019

Journal of the American Chemical Society
Article
rely on the assistance of silver. Similar results were obtained in
the gold-catalyzed oxidative coupling reaction between vinyl-
gold and boronic acids (Figure 1B), which was first reported by
Zhang and co-workers.12b
Scheme 9. Refining Gold Catalysis on the Basis of Reactivity
Type II: Au/Ag Bimetallic Catalysis. As illustrated in
Scheme 8, Au/Ag bimetallic catalysis was a rather surprising
new type of reaction. Neither gold nor silver alone could
catalyze these reactions. Even more impressively, in our
investigation of the previously reported transformations, this
type of reactions was not rare at all. Besides the alkyne
hydration shown in Table 3 and Scheme 2, four more
transformations were identified with the exact same catalytic
manner, as shown in Figure 2.13
combination of these two metals could result in the formation
of an effective catalytic system. The type-III silver-assisted gold
catalysis, on the other hand, combines the transformations
promoted by either gold or silver or a combination of the two
metals. Different reactivities are usually associated with the
various types of catalysts employed. Generally, the combination
of silver and gold exhibits improved reactivity in comparison
with gold alone. With the awareness of these three theoretical
types of gold catalysis, we investigated several different
literature-reported “gold-catalyzed” reactions. All three types
of reactions were obtained during this investigation. Twelve
representative transformations are shown in Figures 1−3 to
highlight the generality of this long-overlooked “silver effect” in
gold catalysis.
Figure 1. Examples of genuine gold catalysis.
Figure 2. Examples of Au/Ag bimetallic catalysis.
Type I: “Genuine” Gold Catalysis. The relativistic effect
makes gold cations one unique π-acid in alkyne activation. In
fact, more and more efforts have been devoted to obtaining
“silver-free” gold catalysts with the intention of avoiding the
potential influence of silver cations. Therefore, one may assume
that among the numerous reported transformations, “pure”
Au(I)-promoted reactions (no need of assistance from silver)
should be plentiful. However, among the reactions we screened
(14 types of literature-reported reactions), only two examples
were confirmed that could be promoted effectively by silver-free
gold without loss of reactivity compared with gold/silver
mixture catalysts.
Again, these transformations were all previously assumed to
involve gold catalysis, since silver salts could not promote these
reactions at all. However, on the basis of the reinvestigation, the
pure gold catalyst could not promote these reactions. Addition
of the silver salts (AgCl could not activate the gold) resulted in
the reactivation of the gold catalyst, though longer reaction
times were needed, likely caused by the formation of the active
catalyst (similar to the 31P chemical shift change results).
Type III: Silver-Assisted Gold Catalysis. The examples
shown for type I and type II are two extreme cases. There
should be examples of type-III catalysis, in which both gold
cation and a mixture of gold and silver can promote the
reaction, likely with different reactivities. This type of gold
catalysis was also successfully identified, as shown in Figure 3.14
Although this type of reaction could be promoted by the
pure gold catalyst, the combination of gold and silver gave
much better results. In addition, these reactions could not be
promoted by silver alone, consistent with the literature results.
The first example is the Hashmi phenol synthesis,12a which
involves gold-promoted furan−yne rearrangement to phenol
through the formation of gold carbene intermediates (Figure
1A). The original conditions involved the treatment of XPhos-
Au-Cl with AgNTf2 for the preparation of [XPhos-Au]+. Our
investigation revealed similar reactivities for the silver-free gold
catalyst (after filtration through Celite) and the mixture of gold
and silver. Therefore, this interesting transformation does not
9016
dx.doi.org/10.1021/ja303862z | J. Am. Chem. Soc. 2012, 134, 9012−9019

Journal of the American Chemical Society
Article
filtration.16 The [(PPh3)Au-triazole]+·BF4 complex was
successfully prepared using the same method. However, the
attempts to grow this crystal were unsuccessful because of slow
decomposition. Nevertheless, comparison of the 19F NMR
spectra of AgBF4 and [(PPh3)Au]+·BF4− indicated the presence
of BF4− anion with no decomposition after Celite filtration (see
the Supporting Information). The fact that all of these anions
“survived” the Celite filtration strongly suggests that the silver
cations were the crucial “activation factor” for [L-Au]+ in the
−
́
above-mentioned studies. Notably, Gagne and co-workers
reported a similar Au−Ag dinuclear resting state in the catalytic
cycle in their mechanistic investigation of gold-catalyzed allene
hydroarylation, which is consistent with this “silver effect”.17
However, to be absolutely certain, the reactions using
(PPh3)AuMe/TfOH were conducted.
Addition of acids to L-Au-Me or L-Au-Ph is another general
strategy that has been reported in the literature for the
preparation of [L-Au]+ (with formation of methane or
benzene). This process is extremely sensitive to solvent
(making L-Au-Cl/AgA a better choice in practice). For
example, addition of TfOH to (PPh3)AuMe in either
CH2Cl2, CH3NO2 or CH3CN caused immediate catalyst
decomposition, forming gold nanoparticles or a gold mirror
within minutes. Therefore, it is very difficult to compare
directly the reactivities of [L-Au]+ formed from (PPh3)AuMe/
TfOH and (PPh3)AuCl/AgOTf with Celite filtration. For-
tunately, solvent screening revealed that dioxane is an optimal
solvent with much slower catalyst decomposition using the
(PPh3)AuMe/TfOH approach. This allowed the head-to-head
comparison of the hydrations of 1a to investigate the claimed
“silver effect”. As shown in Figure 5, [PPh3Au]+TfO− obtained
Figure 3. Examples of silver-assisted gold catalysis.
Notably, two other reactions were also identified that could be
promoted by both gold and silver. However, a gold/silver
mixture or even gold alone exhibited better performance
(Figure 4).15
Figure 5. [PPh3Au]+TfO− from PPh3AuMe/TfOH.
Figure 4. Examples of silver-promoted reactions.
from PPh3AuCH3/TfOH gave the exactly same results as
observed for [PPh3Au]+TfO− prepared from Celite filtration:
without the addition of AgOTf, [PPh3Au]+TfO− exhibited no
reactivity, providing unambiguous evidence of the “silver effect”
in this type of gold catalysis.
The Activation Factor in Gold Catalysis. While the
influence of silver on gold catalysis was shocking and crucial,
the TA-Au catalysis suggested that silver salts are not the only
viable “activation factor”. Interestingly, the two genuine gold
catalysis examples shown in Figure 1 both used the
coordinating anion Tf2N−. In general, The L-Au-NTf2
complexes were different than the gold catalysts formed with
noncoordinated counteranions. As indicated in Scheme 10, the
(PPh3)AuNTf2 gave a 31P NMR chemical shift at 31.1 ppm,4a
The Effect of Celite Filtration. The results shown above
clearly suggest the influence of silver salts in gold catalysis. One
concern was whether the Celite filtration caused some changes
that deactivated the catalysts. Sequential addition of silver salts
to the Celite filtrates resulted in the reactivation of the catalyst
to give the observed results. The three most commonly applied
noncoordinated anions in literature-reported gold catalysis are
TfO−, SbF6 , and BF4 . The crystal structure shown in Scheme
−
−
5 confirmed that TfO− was filtered through the Celite with no
−
decomposition. Similarly, as we reported previously, the SbF6
anion was used as the counteranion to form the corresponding
gold complexes with no decomposition through Celite
9017
dx.doi.org/10.1021/ja303862z | J. Am. Chem. Soc. 2012, 134, 9012−9019

Journal of the American Chemical Society
Article
Scheme 10. Comparison of 31P NMR Chemical Shifts for
Different Au(I) Samples
much more complicated than the previously assumed simple
[L-Au]+ model.
CONCLUSION
■
In view of the huge amount of effort that has been put into
homogeneous gold catalysis, it is clear that a better under-
standing of the nature of the reaction is crucial for further
advance of this research field. The exact complexes formed by
mixing silver and gold cations are not certain at this moment.
With the noncoordinating anions (TfO−, SbF6 , and BF4 ), the
reactions indicated little counteranion influence, suggesting that
anion-bridged gold−silver cocomplexes do not form. The
addition of AgCl did not activate gold catalysis in all of the
tested cases, ruling out the formation of [Ag−Cl−Au]+ as the
actual catalyst. The relatively large number of examples tested
in this work confirm the generality of this phenomenon that the
long-overlooked influence of silver in gold catalysis is crucial
and real.
−
−
which is similar to the chemical shift of TA-Au (30.8),
suggesting the coordination of Tf2N− anion with [(PPh3)Au]+
in solution.
The silver-free TA-Au effectively promoted the hydration of
1, suggesting that other coordination ligands, such as 1,2,3-
triazoles, could also be used as the “activation factor” to initiate
the gold catalysis instead of silver. To investigate the difference
in the gold catalysts formed from coordinating anions and
noncoordinating anions, silver-free (PPh3)AuNTf2 and IPr-Au-
NTf2 were prepared with Celite filtration. XPS confirmed that
neither sample had silver present. Several reactions were re-
evaluated with these silver-free catalysts, and the results are
summarized in Figure 6.
ASSOCIATED CONTENT
* Supporting Information
Detailed experimental procedures, spectral data for all new
compounds, and a CIF file for [(IPr)Au-BTZ]OTf. This
material is available free of charge via the Internet at http://
pubs.acs.org.
■
S
AUTHOR INFORMATION
Corresponding Author
Xiaodong.Shi@mail.wvu.edu
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
We thank the NSF (CAREER-CHE-0844602) for financial
support.
■
REFERENCES
■
(1) For selected recent reviews of homogeneous Au catalysis, see:
(a) Rudolph, M.; Hashmi, A. S. K. Chem. Commun. 2011, 47, 6536−
6544. (b) Nolan, S. P. Acc. Chem. Res. 2011, 44, 91−100. (c) Hashmi,
A. S. K. Angew. Chem., Int. Ed. 2010, 49, 5232−5241. (d) Díez-
́
Gonzalez, S.; Marion, N.; Nolan, S. P. Chem. Rev. 2009, 109, 3612−
3676. (e) Gorin, D. J.; Sherry, B. D.; Toste, F. D. Chem. Rev. 2008,
108, 3351−3378. (f) Widenhoefer, R. A. Chem.Eur. J. 2008, 14,
5382−5391. (g) Arcadi, A. Chem. Rev. 2008, 108, 3266−3325.
(h) Furstner, A.; Davies, P. W. Angew. Chem., Int. Ed. 2007, 46, 3410−
3449. (I) Gorin, D. J.; Toste, F. D. Nature 2007, 446, 395−403.
(2) Patrick, S. R.; Boogaerts, I. I. F.; Gaillard, S.; Slawin, A. M. Z.;
Nolan, S. P. Beilstein J. Org. Chem. 2011, 7, 892−896.
(3) (a) Sengupta, S.; Duan, H.; Lu, W.; Petersen, J. L.; Shi, X. Org.
Lett. 2008, 10, 1493−1496. (b) Duan, H.; Sengupta, S.; Petersen, J. L.;
Shi, X. Organometallics 2009, 28, 2352−2355. (c) Liao, W.; Chen, Y.;
Duan, H.; Liu, Y.; Petersen, J. L.; Shi, X. Chem. Commun. 2009, 6436−
6438. (d) Yan, W.; Wang, Q.; Chen, Y.; Petersen, J. L.; Shi, X. Org.
Lett. 2010, 12, 3308−3311.
(4) (a) Duan, H.; Sengupta, S.; Petersen, J. L.; Akhmedov, N. G.; Shi,
X. J. Am. Chem. Soc. 2009, 131, 12100−12102. (b) Duan, H.; Yan, W.;
Sengupta, S.; Shi, X. Bioorg. Med. Chem. Lett. 2009, 19, 3899−3902.
(5) Chen, Y.; Yan, W.; Akhmedov, N.; Shi, X. Org. Lett. 2010, 12,
344−347.
(6) (a) Wang, D.; Ye, X.; Shi, X. Org. Lett. 2010, 12, 2088−2091.
(b) Wang, D.; Gautam, L. N. S.; Bollinger, C.; Harris, A.; Li, M.; Shi,
X. Org. Lett. 2011, 13, 2618−2621. (c) Wang, D.; Zhang, Y.; Cai, R.;
Shi, X. Beilstein J. Org. Chem. 2011, 7, 1014−1020.
Figure 6. Examples of Tf2N− as the activation factor. NMR yields are
shown.
As indicated, the silver-free L-Au-NTf2 showed good
reactivity in these transformations, which could not be
promoted by [L-Au]+TfO−. Therefore, in addition to silver
salts and 1,2,3-triazoles, the coordinating anion Tf2N− was
identified as another activation factor in gold catalysis. Notably,
for the intermolecular cyclopropanation reaction (Figure 6E),
the combination of silver and Tf2N− exhibited improved
reactivity in comparison with (PPh3)AuNTf2 alone, which
suggests the different influence of various activation factors.
Overall, the re-evaluation of literature-reported gold-
catalyzed reactions confirmed the rather shocking observation
that gold catalysis is greatly influenced by the presence of silver
salts, which were previously considered as inactive reagents. It is
clearly suggested that the gold catalysis mechanism could be
9018
dx.doi.org/10.1021/ja303862z | J. Am. Chem. Soc. 2012, 134, 9012−9019

Journal of the American Chemical Society
Article
(7) For selected examples of gold-catalyzed 1,2-shifts, see:
(a) Jimenez-Nunez, E.; Echavarren, A. M. Chem. Commun. 2007,
́
́
̃
333−346. (b) Barluenga, J.; Riesgo, L.; Vicente, R.; Lopez, L.; Tomas,
M. J. Am. Chem. Soc. 2007, 129, 7772−7773. (c) Hashmi, A. S. K.;
Hutchings, G. J. Angew. Chem., Int. Ed. 2006, 45, 7896−7936.
(d) Wang, S.; Zhang, L. J. Am. Chem. Soc. 2006, 128, 14274−14275.
(e) Zhang, L.; Wang, S. J. Am. Chem. Soc. 2006, 128, 1442−1443.
(f) Gorin, D. J.; Dube, P.; Toste, F. D. J. Am. Chem. Soc. 2006, 128,
14480−14481. (g) Johansson, M. J.; Gorin, D. J.; Staben, S. T.; Toste,
F. D. J. Am. Chem. Soc. 2005, 127, 18002−18003. (h) Mamane, V.;
Gress, T.; Krause, H.; Furstner, A. J. Am. Chem. Soc. 2004, 126, 8654−
̈
8655. For selected examples of 1,3-shifts, see: (i) Zhang, L. J. Am.
Chem. Soc. 2005, 127, 16804−16805. (j) Zhang, L.; Wang, S. J. Am.
Chem. Soc. 2006, 128, 1442−1443. (k) Buzas, A.; Istrate, F.; Gagosz, F.
Org. Lett. 2006, 8, 1957−1959. (l) Wang, S.; Zhang, L. Org. Lett. 2006,
8, 4585−4587. (m) Wang, S.; Zhang, L. J. Am. Chem. Soc. 2006, 128,
8414−8415. (n) Buzas, A.; Gagosz, F. J. Am. Chem. Soc. 2006, 128,
12614−12615. (o) Wang, S.; Zhang, L. J. Am. Chem. Soc. 2006, 128,
14274−14275. (p) Marion, N.; Nolan, S. P. Angew. Chem., Int. Ed.
2007, 46, 2750. (q) Marco-Contelles, J.; Soriano, E. Chem.Eur. J.
2007, 13, 1350. (r) Barluenga, J.; Riesgo, L.; Vicente, R.; Lopez, L.;
Tomas, M. J. Am. Chem. Soc. 2007, 129, 7772. (s) Shi, F.-Q.; Li, X.;
Xia, Y.; Zhang, L.; Yu, Z.-X. J. Am. Chem. Soc. 2007, 129, 15503−
15512.
(8) (a) Liu, R.; Lin, Z.; Zhu, T.; Fang, Y.; Gu, Q.; Zhu, W. J. J. Nat.
Prod. 2008, 71, 1127−1132. (b) Wang, S.-K.; Huang, M.-J.; Duh., C.-
Y. J. Nat. Prod. 2006, 69, 1411−1416. (c) Chen, F.-C.; Peng, C.-F.;
Tasi, I.-L.; Chen, I.-S. J. Nat. Prod. 2005, 68, 1318−1323. (d) Liu, Y.-
B.; Su, E.-N.; Li, J.-L.; Yu, S.-S.; Liu, J.; Li, Y. J. Nat. Prod. 2009, 72,
229−237.
(9) Shi, X.; Gorin, D. J.; Toste, F. D. J. Am. Chem. Soc. 2005, 127,
5802−5803.
(10) Ghosh, N.; Nayak, S.; Sahoo, A. J. Org. Chem. 2011, 76, 500−
511.
(11) Marion, N.; Ramon, R. S.; Nolan, S. P. J. Am. Chem. Soc. 2009,
131, 448−449.
(12) (a) Figure 1A: Hashmi, A. S. K.; Weyrauch, J. P.; Rudolph, M.;
Kurpejovic, E. Angew. Chem., Int. Ed. 2004, 43, 6545−6547. (b) Figure
1B: Zhang, G.; Peng, Y.; Cui, L.; Zhang, L. Angew. Chem., Int. Ed.
2009, 48, 3112−3115.
(13) (a) Figure 2A: Zhang, L.; Wang, S. J. Am. Chem. Soc. 2006,
128, 1442−1443. (b) Figure 2B: Staben, S. T.; Kennedy-Smith, J. J.;
Toste, F. D. Angew. Chem., Int. Ed. 2004, 43, 5350−5352. (c) Figure
2C: see ref 7g. (d) Figure 2D: Luzung, M. R.; Markham, J. P.; Toste,
F. D. J. Am. Chem. Soc. 2004, 126, 10858−10859.
(14) (a) Figure 3A: Ramon, R. S.; Marion, N.; Nolan, S. P.
Tetrahedron 2009, 65, 1767−1773. (b) Figure 3B: Marion, N.;
́
Carlqvist, P.; Gealageas, R.; de Fremont, P.; Maseras, F.; Nolan, S. P.
Chem.Eur. J. 2007, 13, 6437−6451. (c) Figure 3C: see ref 9. (d)
Figure 3D: Reich, N. W.; Yang, C. G.; Shi, Z.; He, C. Synlett 2006,
1278−1280.
́ ́
(15) (a) Figure 4A: Marion, N.; Díez-Gonzalez, S.; de Fremont, P.;
Noble, A. R.; Nolan, S. P. Angew. Chem. 2006, 118, 3729−3732. (b)
Figure 4B: Yu, M.; Zhang, G.; Zhang, L. Org. Lett. 2009, 11, 3646−
3649.
(16) Chen, Y.; Wang, D.; Petersen, J. L.; Akhmedov, N.; Shi, X.
Chem. Commun. 2010, 46, 6147−6149.
(17) (a) Weber, D.; Gagne,
(b) Weber, D.; Tarselli, M. A.; Gagne,
2009, 48, 5733−5736.
́
M. R. Org. Lett. 2009, 11, 4962−4965.
́
M. R. Angew. Chem., Int. Ed.
9019
dx.doi.org/10.1021/ja303862z | J. Am. Chem. Soc. 2012, 134, 9012−9019