Explanation of "silver Effects" in Gold(I)-Catalyzed Hydroalkoxylation of Alkynes

Article abstract of DOI:10.1021/acscatal.5b01493

An extensive experimental NMR study of gold-catalyzed hydroalkoxylation was conducted to explain the influence of a silver salt additive on the gold-catalyzed process (silver effect). Addition of silver salt may have no effect or a negative or positive effect on gold-catalyzed hydroalkoxylation. However, silver was shown to be essentially innocent (plays no role) with regard to the mechanism of the catalytic process itself. The effect occurs only if silver induces variations in the fraction of in-cycle organogold intermediates and H⁺. This is associated with the formation of the argento vinyl gold species G, which was shown to be an off-cycle intermediate. This species is formed by trapping the vinyl gold species B with Ag⁺; in the same way, the diaurated species D (another possible off-cycle intermediate) is formed by trapping B with LAu⁺. The argento vinyl gold species G1 was extensively characterized in solution by various NMR techniques at different temperatures. Furthermore, bringing together our results and the topical research of others, we introduced classification of silver effects to eliminate the confusion around the interpretation of erratic effects. Reactions that are separately catalyzed either by silver or by gold were beyond the scope of this study (for those reactions a "true" silver effect would take place).

Full text of DOI:10.1021/acscatal.5b01493

Research Article
pubs.acs.org/acscatalysis
Explanation of “Silver Effects” in Gold(I)-Catalyzed Hydroalkoxylation
of Alkynes
Alexander Zhdanko* and Martin E. Maier*
Institut fur Organische Chemie, Universitat Tubingen, Auf der Morgenstelle 18, 72076 Tubingen, Germany
̈
̈
̈
̈
S
* Supporting Information
ABSTRACT: An extensive experimental NMR study of gold-
catalyzed hydroalkoxylation was conducted to explain the
influence of a silver salt additive on the gold-catalyzed process
(silver effect). Addition of silver salt may have no effect or a
negative or positive effect on gold-catalyzed hydroalkoxylation.
However, silver was shown to be essentially innocent (plays no
role) with regard to the mechanism of the catalytic process itself.
The effect occurs only if silver induces variations in the fraction of
in-cycle organogold intermediates and H+. This is associated with
the formation of the argento vinyl gold species G, which was
shown to be an off-cycle intermediate. This species is formed by
trapping the vinyl gold species B with Ag+; in the same way, the
diaurated species D (another possible off-cycle intermediate) is
formed by trapping B with LAu+. The argento vinyl gold species G1 was extensively characterized in solution by various NMR
techniques at different temperatures. Furthermore, bringing together our results and the topical research of others, we introduced
classification of silver effects to eliminate the confusion around the interpretation of erratic effects. Reactions that are separately
catalyzed either by silver or by gold were beyond the scope of this study (for those reactions a “true” silver effect would take
place).
KEYWORDS: gold catalysis, hydroalkoxylation, silver effect, reaction kinetics, reaction mechanism
Scheme 1. Reactivity of Vinyl Gold and Diaurated Species
1. INTRODUCTION
toward Ag+ Observed by Gagne2
́
For a long time silver salts AgX have been routinely used to
activate gold chloride complexes by halogen abstraction in
order to convert them to active gold catalysts or for the
synthesis of various gold complexes (eq 1).1 The precipitation
of AgCl is normally considered stoichiometric and complete.
Therefore, until recently the role of silver in gold catalysis has
not been discussed beyond this simple ligand exchange process.
However, recently there has been an increased amount of
experimental observations suggesting that silver is not totally
innocent in gold catalysis. The first mechanistic evidence was
the observation of mixed Au−Ag intermediates in reaction
mixtures, briefly reported by Gagne
́
and Weber in 2009.2 They
showed that silver is able to react with the key organogold
intermediates to generate new bimetallic intermediates
(Scheme 1). The diaurated species D0 and vinyl gold B0
react with AgNTf2 to give the argento vinyl gold species G0,
which was even isolated and characterized by NMR and MS.
Unfortunately, no X-ray analysis could be performed and the
exact structure remains speculative, on the basis of inspection of
previously known examples of various Au/Ag bimetallic
Received: July 16, 2015
Revised: August 27, 2015
© XXXX American Chemical Society
5994
DOI: 10.1021/acscatal.5b01493
ACS Catal. 2015, 5, 5994−6004

ACS Catalysis
Research Article
species.3 This suggests that the structure of G0 must have much
of the planar vinyl gold character with only side-on
coordination of silver, with maintenance of a Au−Ag
metallophilic interaction. In other words, the three-center−
two-electron interaction which was symmetrical in D0 now is
not equal in G0, with the C−Au bond dominating the
structure.2 In accordance with this notion we will apply the
term “argento vinyl gold” to emphasize the predominant vinyl
gold vs vinyl silver character of these species. Furthermore,
classification of various origins of the silver effect. This would
put in order an understanding of the erratic results associated
with this effect.
2. RESULTS AND DISCUSSION
In our experiments we used silver-free gold catalysts
−
−
Ph3PAuNCMe+SbF6 (1) and L2AuNCMe+SbF6 (2; L2 =
2-(di-tert-butylphosphino)biphenyl). For the observation of
silver effects the reactions were performed in the presence of
the silver salts AgOTf and AgSbF6. We found that argento vinyl
gold species were involved every time a silver effect took place.
Therefore, prior to discussion and explanation of silver effects it
is necessary to describe the properties and reactive pathways of
this species (sections 2.1−2.3.). The silver effect is then
explained for a number of catalytic hydroalkoxylation reactions
(section 2.4.). In section 2.5, we propose a generalized picture
of silver effects using the knowledge obtained in our research
and that of others.
́
Gagne and Weber observed G0 as a resting state of the catalytic
reaction, and this remains the only direct evidence of
involvement of Au/Ag bimetallic intermediates in gold catalysis.
In their case, the reaction was retarded by additions of AgNTf2
(which would be called a “negative silver effect”). In addition,
no investigation into the role of G in the reaction mechanism
was provided.
An extended work reported in 2012 by Shi et al.
demonstrated that the rate (and yield) of many gold-catalyzed
reactions can be largely influenced by the presence of Ag+ (in
addition to already present gold catalyst).4 These observations
not only questioned the innocence of silver in gold catalysis but
also questioned fundamental mechanistic principles of gold
catalysis at that time. Therefore, an in-depth study seemed
necessary to clarify the situation. In the following years there
have been increased efforts to investigate and explain this “silver
effect” in more detail. Notable examples include recent works
from the groups of Echavarren,5 Hammond, and Hu (to be
discussed below).6 Due to the ambiguous role of silver, several
alternative methods for activation of a gold catalyst have been
developed that avoid the use of silver salts.7,8
2.1. Synthesis of Argento Vinyl Gold in Situ and
Reactions with Nucleophiles. In 2013 we reported the
formation of diaurated species D using an alkynol, a gold
catalyst, and the non-nucleophilic base 1,8-bis(dimethylamino)-
naphthalene (Proton Sponge, PrSp).10 Using AgOTf as an
additional reactant, the same method was initially attempted for
the synthesis of the corresponding argento vinyl gold species.
However, AgOTf was found to be incompatible with PrSp,
causing immediate formation of a black precipitate (reduction
of silver). Therefore, tBu2Py was used as a redox-stable base.
Thus, simple treatment of a concentrated THF-d8 solution
containing a gold catalyst (1 or 2) and AgOTf with a CDCl3
solution containing tBu2Py and alkynol (S1 or S2) led to
immediate, quantitative formation of the corresponding
organometallic species, accompanied by minor competitive
enol ether formation. Argento vinyl gold species G1 and G4
were formed exclusively (Table 1, entries 1 and 4), while G2
and G3 were accompanied by a minor formation of the
In our laboratory we have been involved in mechanistic
investigations of gold catalysis. Recently we have reported
extensive experimental studies on gold(I)-catalyzed hydro-
alkoxylation of alkynes (Scheme 2).9 Herein, using this model
Scheme 2. Mechanism of Gold-Catalyzed Hydroalkoxylation
Table 1. Stoichiometric Formation of Argento Vinyl Gold in
Situ
reactant (amount in equiv)
a
entry
alkynol
gold
tBu2Py
AgOTf
G:D
1:0
1
2
3
4
S1 (1.1)
S2 (2.0)
S1 (1.1)
S2 (1.5)
2 (1.0)
1 (1.0)
1 (1.0)
2 (1.0)
1.5
1.3
1.6
2
1.1
1.1
1.1
1.4
1:0.1
1:0.03
1:0
a
Legend: 1, L = PPh3; 2, L = L2 = 2-(di-tert-butylphosphino)biphenyl.
reaction with a well-established mechanism, we provide a
systematic mechanistic experimental study to disclose the
origins of the silver effect. In this work any silver effect is
regarded as the influence of silver on the rate of disappearance
of the starting alkyne S, because formation of enol ether C from
S is a gold-catalyzed step, while formation of acetal L from C is
only a Brønsted acid catalyzed process. On the basis of our
observations and the works of others we propose a generalized
5995
DOI: 10.1021/acscatal.5b01493
ACS Catal. 2015, 5, 5994−6004

ACS Catalysis
Research Article
corresponding diaurated species D2 and D3 (entries 2 and 3).
An important requirement for successful formation of argento
vinyl gold is that the silver salt should be dissolved prior to the
reaction; otherwise, formation of D and/or enol ether will
predominantly occur. However, AgOTf is negligibly soluble in
CDCl3. This is why it was dissolved in a minimum amount of
THF-d8. The synthesis of G2−G4 in CDCl3 occurs under
kinetic control, because D2−D4 are thermodynamically
favored (to be discussed below). Attempts to isolate any of
the argento vinyl gold species in the crystalline state were
unsuccessful. Therefore, all characterizations and reactions were
performed in solution, using enriched samples prepared in situ.
1H NMR spectra of G1−G4 (prepared in situ) exhibit simple
patterns, despite the molecules being unsymmetrical. Cooling
these solutions to −60 °C led to split spectra in case of G1, G3,
and G4. Since G2 was unstable in solution and equilibrated
toward D2 within several minutes, no clear spectrum was
obtained for this species at low temperature. A likely reason for
the degeneracy of the CH2 and C(CH3)2 protons at room
temperature is a fast silver exchange, occurring in the presence
of a slight excess of AgOTf. Indeed, when the amount of
residual AgOTf was sufficiently low (or excluded), we could
observe split spectra of G1 and G4, revealing the unsym-
metrical structure even at room temperature. Most reliably this
was achieved by fast and quantitative reduction of the residual
AgOTf by PrSp. Alternatively, the samples could be slowly
reduced by CaH2 (a routine procedure to dehydrate a solution;
G stays unreactive). Upon addition of fresh AgOTf (as a
solution in minimum THF), the split spectrum is simplified,
indicating resumption of the fast silver exchange. The 1H NMR
spectrum of G1 in the mode of fast and slow silver exchange is
shown in Figure 1 as an example. We suppose that the silver
very intriguing, since this signal is often not found. In our case,
it could be observed as a doublet at 125.1 ppm (2JP = 93.7 Hz).
The 2J phosphorus coupling through the two bonds of the C−
1
Au−P link appears to be greater than the J coupling through
the single CAr−P bond (36.5 Hz). In addition, it is comparable
to the value typical for vinyl gold species (ca. 105 Hz) and is
more distant from the value typical for diaurated species (ca. 60
Hz).11 This is in accordance with a planar vinyl gold moiety
being only slightly disturbed by the side-on coordination of
silver. The enol ether α-carbon atom appeared at 173.3 ppm as
a doublet (3JP = 9.1 Hz). The coupling to the magnetically
active Ag was not observed in any spectrum at room
temperature, presumably because of the fast silver exchange
in solution.
In order to observe the coupling to Ag, we recorded NMR
spectra of G1 at −20 °C under conditions of slow silver
1
exchange. No evidence was found for coupling in the H
spectrum. This is not surprising, since there are no protons
located in close proximity to the Ag atom. In contrast, the 13C
spectrum was much more indicative. The CAu atom and the
enol ether α-carbon atom were additionally split, appearing
now as doublets of doublets. The assignment of Ag coupling
was made from a comparison of multiplicity in 13C spectra
recorded at room temperature and at −20 °C, with the
knowledge of the P coupling from the room-temperature
1
spectrum. This way JAg = 46 Hz was determined for the CAu
atom and 2JAg = 4 Hz for the enol ether α-carbon atom. These
data are consistent with the assigned structure of G1. To the
best of our knowledge, this is the first time C−Ag coupling has
been determined for any species containing a quaternary
CAu(Ag) carbon atom.
Next, we attempted to elucidate the bonding and
coordination chemistry at silver and found it very intriguing.
To get some insight, we conducted NMR titrations of G1 at
room temperature with various nucleophiles in CDCl3 (Me2S,
Lut = 2,6-lutidine, TMTU = tetramethylthiourea, PPh3, OAc−,
1
Cl−), monitoring H, 31P, and 19F NMR spectra after each
addition of the nucleophile. No clear picture about the
coordination mode of silver could be obtained in most cases.
All reaction mixtures exhibited fast ligand exchange at silver on
the NMR time scale, as evidenced by a single set of signals of
the enol ether core and the ligands (at room temperature).
However, some of these experiments provided useful
information. Thus, stoichiometric substitution of the OTf−
ion by 1 equiv of OAc− at G1 could be confirmed by the
absence of further changes in the 1H, 31P, and 19F NMR spectra
of G1 at >1 equiv of OAc− (Scheme 3). The strong
Scheme 3. Stoichiometric Reactions of G1 with Various
Nucleophiles (OAc−, Cl−, and PPh3)
Figure 1. Dynamic behavior of G1 in the presence of AgOTf.
exchange might occur either by a nucleophile-assisted
dissociative pathway or through a doubly argented intermedi-
ate. We prefer the latter scenario, which was also proposed by
Gagne
́
to explain the spectrum of G0.2
1
Complex G1 was characterized in CDCl3 solution using H,
31P, 19F, 13C and various two-dimensional NMR techniques
(these spectra were recorded under conditions of fast silver
exchange at room temperature). Using CF3CO2Me as an
internal standard, the overall 1:1:1 organic core:Au:Ag ratio
1
could be confirmed by H and 19F NMR spectra. Observation
of the CAu carbon signal in the 13C NMR spectrum of G1 is
5996
DOI: 10.1021/acscatal.5b01493
ACS Catal. 2015, 5, 5994−6004

ACS Catalysis
Research Article
nucleophiles PPh3 and Cl− caused stoichiometric cleavage of
G1 to form vinyl gold species B1 and Ag(PPh3)n+ (n = 2−4) or
AgCl precipitate. No evidence for the alternative cleavage to
vinyl silver species RAgL and LAuNu+ was obtained; thus, this
pathway appears improbable. Whereas the reaction with Cl−
required 1 equiv of the nucleophile for complete cleavage, the
reaction with PPh3 required 2 equiv. This last fact suggests that
G1 can accept 1 equiv of PPh3 at silver without detachment of
B1, whereas the second equivalent of PPh3 causes the cleavage.
G1 was resistant to cleavage in the presence of excess Me2S,
Lut, or OAc−. Partial cleavage was observed in the presence of 4
equiv of TMTU. In cases where incomplete cleavage of G1
took place either due to a reversible reaction (TMTU) or
reaction before the equivalence point (PPh3 and Cl−), the enol
ether core exhibited a single set of signals, indicating a fast
equilibrium between G1 and B1. This all characterizes argento
vinyl gold as a very reactive species, which would undergo fast
silver exchange (in the presence of extra Ag+), fast ligand
exchange at silver (in the presence of extra Nu), and fast
exchange of the vinyl gold fragment at silver (in the presence of
extra vinyl gold).
Scheme 4. Possible Mechanisms of Protodemetalation of
Argento Vinyl Gold
accurate determination of chemical kinetics that would prove
this mechanism more rigorously. Therefore, we turned to a
catalytic study instead.
Theoretical analysis of a gold-catalyzed hydroalkoxylation in
the case of complete and immediate formation of argento vinyl
gold as a single resting state comes to different results
depending on whether it undergoes protodemetalation by a
nucleophile-assisted dissociative mechanism or an SE2 mech-
anism. An analysis of the kinetics with the anticipated
nucleophile-assisted mechanism is shown in Scheme 5. Using
In order to get a more detailed insight into these fast
exchange reactions, we performed some of them at low
temperature. NMR titration of G1 with PPh3 at −40 °C
revealed the formation of an intermediate that was assigned to
the 1:1 adduct G1·PPh3. A 31P NMR spectrum of G1·PPh3 at
low temperature exhibited one signal for L2 and a pair of
doublets due to a single PPh3 group coupled to the two silver
Scheme 5. Derivation of the Rate Law for a Catalytic
Reaction Having G as a Single Resting State
1
isotopes (1J(31P−109Ag) = 641 Hz, J(31P−107Ag) = 555 Hz).
1
Integration of 31P as well as H NMR spectra confirmed the
molecular composition of G1·PPh3. Upon further PPh3
addition G1·PPh3 was cleaved to yield B1, which was also
observed separately at this temperature (Scheme 3). Similar
NMR titration of G1 with TMTU at −40 °C did not reveal the
formation of a separate species, because even at this
temperature the exchange processes were fast.
These observations suggest that silver in G1 has a
coordination number of 2. Therefore, OTf− must be covalently
bound to silver in a CDCl3 solution of G1 in the absence of
sufficiently strong nucleophiles. Rather, in CD3OD solution G1
appears to be a cationic complex. This follows from the
observation of a 19F resonance at a constant shift of −80.1 ppm,
the same as that with AgOTf, PrSpH+OTf−, and TfOH,
unambiguously indicating complete dissociation of OTf− from
silver in CD3OD.
2.2. Mechanism of Protodeauration of Argento Vinyl
Gold. In order to understand the role of argento vinyl gold in
the catalytic cycle, it is necessary to ascertain the mechanism by
which it can be protodemetalated, releasing gold and silver back
into solution. Similarly to what was previously described for
diaurated species, also here there is a choice between the
nucleophile-assisted dissociative mechanism and bimolecular
SE2 mechanism (Scheme 4).
Preliminary experiments were conducted with the argento
vinyl gold species G1 generated in situ. Thus, TfOH,
CF3CO2H, and tBu2PyH+OTf− all caused instant (<5 min)
protodemetalation of this compound in MeOD at room
temperature (G1 was stable toward tBu2PyH+OTf− in CDCl3,
however). Reaction with AcOH occurred more slowly and was
found to be inhibited in the presence of additional AgOTf but
accelerated in the presence of Me2S. These results already
suggest the nucleophile-assisted dissociative mechanism.
However, it was difficult to find suitable conditions to allow
the steady-state approximation for vinyl gold species B (always
present in undetectable amounts, eq 5) and the steady-state
approximation for G (always present in full amount, keeping all
the gold of the system, eq 6), one easily comes to the following
expression for B (eq 7). Putting this into the reaction rate (eq
8) gives the following final expression for the rate law (eq 9).
According to this analysis, the reaction should be zero order in
substrate, first order in gold, first order in acid, and minus first
order in silver. An analysis of the kinetics that would take place
in the case of an SE2 mechanism is given in the Supporting
Information. It leads to zero order in substrate, first order in
gold, first order in acid, and zero order in silver. Therefore, the
order in silver in a catalytic process would be indicative of the
actual mechanism of protodemetalation of argento vinyl gold.
Then we experimentally determined the kinetics of
cyclization of S1 in the presence of various amounts of catalyst
5997
DOI: 10.1021/acscatal.5b01493
ACS Catal. 2015, 5, 5994−6004

ACS Catalysis
Research Article
2, AgOTf, and TsOH in CD3OD (Scheme 6). In each run gold
rested as the single species G1, directly observable by 1H NMR.
Scheme 7. Possible Mechanisms of Transmetalation of
Diaurated Species
Scheme 6. Kinetics of Gold Catalyzed Reaction in the
Presence of Ag+ and H+
We hoped that the slow reaction of D4 with Ag+ would give
us insight into the actual reaction mechanism. This reaction was
performed at a constant initial concentration of D4 using
various amounts of Ag+ (10−80 equiv) and catalyst 2 (0−2
equiv) in MeOD (Scheme 8). Unfortunately, the kinetics of
Scheme 8. Slow and Reversible Transmetalation of D4
Protodemetalation of G1 becomes visible only upon complete
consumption of S1 (the signal of L2Au(Sol)+ appears). [S1]
decreased almost linearly with time; therefore, zero order in
substrate was accepted for further analysis (Supporting
Information). The analysis indicated first order in gold, first
order in acid, and, importantly, minus first order in silver. This
finding is fully consistent with a nucleophile-assisted
dissociative mechanism and not with the an associative SE2
mechanism. Accordingly, argento vinyl gold must be essentially
considered as an off-cycle intermediate, with clear resemblance
to the known properties of diaurated species.12
2.3. Reactivity of Silver Salt toward Diaurated
Species. Following the work of Gagne,2
we studied the
́
transmetalation of diaurated species by silver salts. To our
surprise, diaurated species D2 did not exhibit any appreciable
reactivity, giving no argento vinyl gold G2 species (eq 10).
this reaction could not be analyzed to differentiate between the
two possible mechanisms, because the reaction appeared to be
reversible and did not reach 100% conversion even in the
presence of 80 equiv of Ag+. However, as a small reward, we
were able to determine the equilibrium constant (Keq = 0.025
0.005). Since D2 is >1000 times more stable than D4, it is not
surprising that D2 does not show any signs of transmetalation
even in the presence of a large excess of a silver salt.13
We found that the equilibrium between diaurated species D
and argento vinyl gold species G can be strongly influenced by
the addition of a nucleophile (Scheme 9). Thus, using Me2S,
Lut, TMTU, and PPh3 it is possible to entirely shift the
equilibrium toward G even if it was completely on the side of D
before the addition of a nucleophile. The main driving force is
provided by binding gold to form LAuNu+. Clearly, there is the
opposite driving force provided by binding silver to form
Rather, the opposite reaction was found to occur immediately
(eq 11). In contrast, the less stable diaurated species D4 slowly
reacted with Ag+ in methanol, affording G4. This trans-
formation can be described by two mechanisms: a nucleophile-
assisted dissociative mechanism and a bimolecular SE2
mechanism (Scheme 7). In contrast to protodeauration of
diaurated species that occurs only by the nucleophile-assisted
mechanism, here also the SE2 mechanism appears likely (the
transition state would be stabilized by multiple metallophilic
interactions). It is easy to see that the transmetalation must be
considered a reversible process (if the strong preference for one
of the sides is not known beforehand).
+
Ag(Nu)n (n = 2−4), but it does not play a role, being
dominated by gold, as indicated by a model study of the
binding preferences.14 The stability of LAuNu+ increases in the
order Nu = Me2S < Lut < TMTU < PPh3;15 therefore, the
ability of a nucleophile to shift the equilibrium between D and
G increases in the same order.
These findings allow the conclusion that not every diaurated
species D can be transmetalated by silver salts. This appears to
5998
DOI: 10.1021/acscatal.5b01493
ACS Catal. 2015, 5, 5994−6004

ACS Catalysis
Research Article
Scheme 9. Influence of a Nucleophile on Transmetallation of
Diaurated Species
Scheme 10. No Silver Effect on Hydroalkoxylation of S3
Catalyzed by 2
cases was a L2Au(S3)+/L2Au(Sol)+ mixture. This indicates that
silver has no affect on changing the rate-limiting step (that is,
vinyl gold B formation) and on changing the L2Au(S3)+/
L2Au(Sol)+ ratio. This unambiguously proves that silver does
not cooperate with gold and that all catalytic activity is
exclusively due to the gold catalyst alone. This conclusion was
additionally confirmed in the intramolecular cyclization of 6-
phenylhexyn-5-ol-1 (PhCC(CH2)4OH), which also did not
exhibit any significant silver effect (see the Supporting
Information).
This finding suggests that a possible way for silver to
influence a gold-catalyzed hydroalkoxylation is through
formation of the off-cycle argento vinyl gold species G. This
conclusion is corroborated in a series of further experiments. In
addition to organogold resting states, here we need to consider
the concentration of concomitant acid because all the reactions
described below are acid dependent (in the absence of silver
they are characterized by the formation of B or D as a resting
state).
be a reversible reaction that obviously depends on the
thermodynamic stability of D and G. In most cases D would
be reluctant to undergo transmetalation; rather, the opposite
reaction would be favored. In the case where D is highly
destabilized by the steric bulk of the LAu, formation of G would
be more favored (the stability of G is obviously less sensitive to
the steric bulk of the LAu moiety and the enol ether core in
comparison to that of D). In addition, binding of LAu+ by an
additional nucleophile provides a driving force that may shift
this equilibrium toward G (this may take place under catalytic
conditions; see below).
2.4. Influence of Silver Additives on Catalysis by Gold
Catalysts. Having briefly described the properties of argento
vinyl gold, we turn to an explanation of silver effects in catalysis.
Here MeOD was used as a solvent to ensure complete
solubility of the catalytic system and also to eliminate any
counterion effects. We performed a series of catalytic runs with
and without silver additives and monitored the reaction
progress directly by NMR. Particular attention was given to
the observation of catalytic intermediates in situ during the
whole process. Using these observations as well as a knowledge
of the properties of argento vinyl gold (sections 2.1−2.3.) and
the mechanism of gold-catalyzed hydroalkoxylation (shown in
Scheme 2) it was possible to unambiguously define the role of
silver and explain the silver effect in every case.
Next we investigated the cyclization of 3-pentyn-1-ol (S2) in
CD3OD in the presence of catalyst 1 (Scheme 11). No silver
effect was observed. In the absence of any additives this
reaction is characterized by complete formation of D2 (leaving
only undetectable steady-state concentrations of other gold
species).9 NMR examination confirmed that the same situation
Scheme 11. Absence of Acid and Silver Effects on
Cyclization of S2 in the Presence of Catalyst 1
At first, it is most convenient to describe and explain a silver
effect for a reaction where transition from π-complex A to vinyl
gold species B is rate limiting, because for this easy case there is
no need to consider any off-cycle intermediates (D or G),
protodemetalation, and concentration of acid in the system. At
the same time this reaction type allows probing the role of
silver in the initial steps of the gold catalytic cycle. A suitable
reaction is intermolecular hydroalkoxylation of hexyne-3 (S3),
which was previously investigated.9 Repeating this reaction in
the presence of AgOTf did not reveal any detectable silver
effects (Scheme 10). The resting state of the catalyst in both
5999
DOI: 10.1021/acscatal.5b01493
ACS Catal. 2015, 5, 5994−6004

ACS Catalysis
Research Article
is observed in the presence of silver additive. Given the
aforementioned stability of D2 in comparison to G2 (section
2.3.) it is not surprising that silver was unable to form G2 and
influence the distribution of gold intermediates in any
appreciable way. In other words, if the diaurated species is
too stable, the silver effect does not take place.
Scheme 13. Changes in the Catalytic System Caused by
Addition of Ag+
A completely different situation was observed in a similar
study of the cyclization of 3-pentyn-1-ol (S2) in CD3OD in the
presence of catalyst 2 (Scheme 12). In the absence of silver
Scheme 12. Positive Acid and Silver Effects on Cyclization of
S2 in the Presence of Catalyst 2
In this case the reaction is not accompanied by formation of
any detrimental off-cycle intermediates at all (which means that
the gold catalysis runs at its full possible power) and yet the
concentration of H+ is directly increased. This observation
supports our explanation about the positive silver effect being
associated with favorable changes in material balance of the
catalytic system.
Furthermore, the kinetics of the silver-assisted reaction and
the influence of the amount of AgOTf additive on the strength
of the silver effect were analyzed. It is clear to see that the
strength of the positive silver effect initially increases (going
from 0 until 0.25% Ag, curves 1−3) and then decreases as more
silver is present (from 0.25 until 2.5% Ag, curves 3−6). This is
associated with the interplay of negative and positive factors.
The initial gradual increase of the positive effect must be
associated with the favorable changes in material balance of the
catalytic system (replacement of D4 by G4 as a resting state,
hence, the increase of in-cycle organogold and H+). However,
once the formation of G4 is stoichiometrically complete (which
in our case takes place already at 0.25% Ag, curve 3), addition
of more silver no longer benefits the reaction this way. Rather,
the excess of silver shifts the equilibrium KG further toward G4,
now reducing the total amount of in-cycle organogold, albeit
not affecting H+ (Scheme 13). This explanation is supported by
establishing the minus first order in silver, taking place from
0.25 until 2.5% Ag, as shown in curves 3−6 (see the Supporting
Information for calculations), which is the same result as was
obtained in Scheme 6. In addition, the zero order in S2 is
obeyed with a moderate accuracy for this range of silver
additive.
Similar explanations can be applied to explain the positive
acid and silver effects that were observed in the cyclizations of
4-phenylbut-3-yn-1-ol (S4) and 4-(4-cyanophenyl)-butyn-3-ol-
1 (S5) in the presence of catalyst 2 (Schemes 14 and 15) and
the cyclization of S2 catalyzed by diaurated species D4
(Supporting Information). In the case of S4 the resting state
was entirely changed from D5 to G5 upon addition of silver,
resembling the situation explained for alcohol S2 above. Only
partial changes in resting states were noted in the case of
alcohol S5 (see the Supporting Information for more details).
As a further case we examined the reaction of 2,2-dimethyl-4-
phenylbut-3-yn-1-ol (S1), which is unable to form diaurated
species with catalyst 2 because of steric reasons associated with
additive this reaction is also accompanied by formation of the
corresponding diaurated species (D4).9 However, because of
the low catalyst loading (0.16% of 2) and the lower stability of
D4 in comparison with that of D2, the diaurated species does
not form instantly but accumulates gradually during the first
few minutes of the reaction, resulting in appearance of the rush
period on the kinetic curve (curve 1). When it was performed
in the presence of various amounts of AgOTf additive, this
reaction exhibited a positive silver effect (curves 2−6). NMR
examination indicated that D4 was (almost) completely
replaced by the corresponding argento vinyl gold G4, which
was observed as a single resting state. Since the formation of G4
is immediate, the rush period disappears. For an explanation of
this effect it is necessary to consider the material balance of the
catalytic system. Thus, in the presence of silver we have 1 equiv
of G4 and 1 equiv of H+ instead of 1/2 equiv of D2 and 1/2
equiv of H+ before (per 1 equiv of LAu+ as the reference,
Scheme 13). Second, KD > KG because D4 is more stable than
G4 (see section 2.3).16 This means that G4 must be a less
detrimental off-cycle resting state for the reaction than D4. As a
consequence, the gold/silver-catalyzed reaction benefits from
having a higher fraction of in-cycle organogold intermediates
and a higher concentration of H+. This adds together to
account for the overall positive silver effect. For comparison we
also determined the effect of a Brønsted acid additive (curve 7),
which appeared to be considerably stronger than a silver effect.
6000
DOI: 10.1021/acscatal.5b01493
ACS Catal. 2015, 5, 5994−6004

ACS Catalysis
Research Article
Scheme 14. Acid and Silver Effects on Cyclization of S4
Scheme 16. Negative Silver Effect on the Cyclization of S1 in
the Presence of Catalyst 2
Scheme 15. Acid and Silver Effects on Cyclization of S5
other gold complexes instead. In 2012 Straub et al. performed a
reaction of a very bulky IPr**AuCl with AgSbF6 under almost
non-nucleophilic conditions and isolated [IPr**AuClAg]+,
where silver is adjacent to the Au−Cl bond (eq 12).17 This
finding shows that an ideal halogen exchange reaction (eq 13)
would be reversible under virtually non-nucleophilic con-
ditions.14 This complex is considered as the very early
intermediate of eq 1.
In 2013 Echavarren et al. described the formation of bridged
chloronium complexes (LAu)2Cl+ as products of incomplete
halogen abstraction (eq 14).5 Since (LAu)2Cl+ is a precatalyst
that has to react through a ligand exchange equilibrium (eq 15),
it is clear that the best theoretical performance of 1 equiv of
(LAu)2Cl+ (with regard to gold) cannot be higher than the
activity of 0.5 equiv of formal LAu+. We estimated the
equilibrium constant of ligand exchange of (L2Au)2Cl+ with
MeCN (eq 16), which indicates that (L2Au)2Cl+ would be
slightly more stable toward reaction with a substrate than
L2AuNCMe+.
such a bulky substrate/ligand combination. In the absence of
any additives this reaction is characterized by complete
formation of vinyl gold B1, having protodeauration as the
rate-limiting step.9 Application of silver should bind this vinyl
gold into argento vinyl gold. In complete accordance with our
expectations, a negative silver effect was indeed observed
(Scheme 16). As expected, vinyl gold completely disappeared
and argento vinyl gold G1 was observed by NMR as the only
species. The kinetics of this reaction was analyzed in section
2.2.
The findings allow the conclusion that silver effects described
in this section are caused only by changes of the concentrations
of in-cycle organogold intermediates and H+, induced by the
formation of the new off-cycle intermediate G, and that silver
has no direct role in the catalytic cycle itself.
CH2Cl2
2L2AuCl + Ag+X− ⎯⎯⎯⎯⎯⎯→⎯ (L2Au)2Cl+ + AgCl↓
(14)
(L2Au)2Cl+ + Substr ⇌ L2AuSubstr+ + L2AuCl
(15)
Keq≈0.4
(L2Au)2Cl+ + MeCN HoooooooI L2AuNCMe+ + L2AuCl
CDCl3
(16)
2.5. Classification of Silver Effects. Our research and the
topical reports found in the literature indicate that the silver
effect may be encountered in many situations. However, it is
not always “silver” in nature (that is, silver is not involved in the
catalytic process itself). In order to eliminate confusion around
the interpretation of various silver effects, we discuss briefly the
silver effects found by other researchers in this area and
demonstrate that all effects have, in their essence, a common
origin.
In 2013 yet another complex was identified by Jones et al. as
a component of LAuCl/Ag+ mixtures: the mixed silver−gold
chloride complex 3 eq 17.18 It also exists under almost non-
2.5.1. Silver Effect Associated with Incomplete Halogen
Exchange (Eq 1). Under certain conditions the halogen
abstraction reaction may not reach completion, leading to
6001
DOI: 10.1021/acscatal.5b01493
ACS Catal. 2015, 5, 5994−6004

ACS Catalysis
Research Article
nucleophilic conditions and immediately precipitates AgCl if a
nucleophile (e.g., a substrate) is added to a solution of 3 (eq
18). A combination of eqs 17 and 18 completes the
stoichiometry of eq 1.
Scheme 17. Negative Effect of MeCN on Reaction Rate
Should Not be Confused with a Silver Effect
From these findings it is easy to conclude that precipitation
of AgCl and the presence of a nucleophile (even a weak one,
such as H2O or MeCN) or a counterion able to bind gold are
the two driving forces which are simultaneously required for the
halogen exchange (eq 1) to be complete.
Furthermore, yet another factor should be taken into
account: the simple solubility of the starting AgX salt. This
idea came to mind after examination of the paper by
Echavarren, where they observed incomplete halogen abstrac-
tion with AgNTf2 and AgOTf in CH2Cl2.5 Presumably, the low
solubility of these salts in CH2Cl2 or CHCl3 alone might be the
likely reason for the incomplete reaction: a certain part of AgX
may remain in the precipitate. In our own practice, we always
dissolved silver salts in a minimum amount of acetone or
MeCN prior to a halogen abstraction reaction. It never failed
under these conditions. In particular, L2AuOTf was synthesized
in quantitative yield from L2AuCl and AgOTf taken in a 1:1
ratio.
Without giving it any further investigation, we propose the
following condition for reliable and 100% complete halogen
exchange to use in laboratory practice: use the starting silver
salt not as a solid but as a solution in a minimum amount of a
weakly coordinating solvent (acetone, MeOH, THF, or
MeCN). This ensures immediate and complete reaction strictly
according to eq 1, avoiding accumulation of any intermediate
complexes such as [LAuClAg]+, (LAu)2Cl+, and 3.
S + LAuCl + AgX → LAuS+X− + AgCl↓
(19)
LAuNCMe+ (eq 20). The gold hydrate can be quantitatively
prepared by rotary evaporation of a commercial LAuNCMe+
S + LAuNCMe+ ⇌ LAuS+ + MeCN
(20)
catalyst from EtOH (two times) and then from CDCl3 (one
time), as exemplified in eq 21. This procedure can be
considered as a simple way to boost a silver-free gold catalyst.
2.5.2. Silver Effect Associated with Reactivation of a Gold
Catalyst That Was Poisoned by Impurity or Celite Filtration.
It was demonstrated that the reactivity of gold catalysts
generated in situ from a LAuCl/AgX system can be
dramatically decreased if they are filtered through Celite prior
to the reaction.4,19 Hammond and Hu nicely explained that this
has nothing to do with silver itself but with poisoning of a gold
catalyst by traces of high gold/proton affinity impurities in
Celite.6 They described reactivation of a gold catalyst by using
H+, Ag+, and other Lewis acids as sacrificial electrophiles to
bind possible catalyst poisons and set the gold catalyst free.
2.5.3. Silver Effect Wrongly Associated with the Negative
Role of MeCN. In connection with the use of a LAuCl/AgX
system instead of a commercially available family of silver-free
catalysts LAuNCMe+X− we especially stress that for some
substrates the LAuCl/AgX system would work more efficiently
than LAuNCMe+ but that this is not necessarily a silver effect!
This might be the effect of the absence of MeCN in the system.
In fact, many alkynes and alkenes have very weak binding
affinity to the catalyst, and for them even MeCN ligand would
be like a catalyst poison. This is nicely exemplified by cis/trans
isomerization of enol ether C1 (Scheme 17), which is much
more efficiently catalyzed by a gold hydrate, a catalyst lacking
MeCN.20 Another example is a reaction of enyne S7 taken from
the literature (the authors did not interpret this effect).5 We
believe the confusing role of MeCN is encountered (but
overlooked) in many gold-catalyzed reactions of substrates with
very low binding affinity to gold. Those reactions would have
LAuNCMe+ as a resting state, and often they would not involve
any neutral organogold intermediates in their mechanism. It is
clear to see that this phenomenon has nothing to do with silver
itself. Simply, the LAuCl/AgX system provides the activated
LAuS+ (S = substrate) complex more easily (eq 19) than does
2.6. Outlook. The knowledge obtained in our research
(sections 2.1−2.4) and the literature (section 2.5) adds
together to explain the origin of silver effects within the
mechanism of gold-catalyzed hydroalkoxylation (Scheme 18).
This scheme illustrates two main conclusions. (1) Ag+ is found
to be essentially innocent with regard to the mechanism of the
catalytic process. (2) The silver effect (positive or negative)
occurs only if silver induces variations of concentrations of in-
cycle organogold intermediates and H+, which can be
associated with formation of the new off-cycle intermediate
argento vinyl gold G, incomplete halogen abstraction by Ag+, or
reactivation of a poisoned catalyst by Ag+. The steps that cause
these variations are shown in blue.
3. CONCLUSION
The research initiated by Gagne (Scheme 1) was expanded to
́
complete the explanation of the silver effect (Scheme 18). In
situ NMR monitoring of gold-catalyzed hydroalkoxylation in
the presence of a silver salt revealed the formation of argento
vinyl gold G in all cases when a silver effect was operative.
Stoichiometric studies indicated argento vinyl gold G to be a
reactive species, undergoing in the presence of Ag+, Nu, or vinyl
gold B, correspondingly, fast silver exchange, fast ligand
exchange at silver, and fast vinyl gold exchange at silver.
Sufficiently strong nucleophiles will cause cleavage of G to yield
B.
Protodemetalation of G was found to occur by the
nucleophile-assisted dissociative mechanism (through B).
Therefore, G is considered as an off-cycle intermediate of the
catalytic process (the same as diaurated species D).
Considering the material balance of the catalytic system, it
6002
DOI: 10.1021/acscatal.5b01493
ACS Catal. 2015, 5, 5994−6004

ACS Catalysis
Research Article
Complete experimental procedures and detailed NMR
spectra (PDF)
Scheme 18. Mechanism of Gold(I)-Catalyzed
Hydroalkoxylation (in Black) Showing the Steps
Responsible for the Silver Effect (in Blue)
AUTHOR INFORMATION
Corresponding Authors
*E-mail for A.Z.: vinceeero@gmail.com, .
■
*E-mail for M.E.M.: martin.e.maier@uni-tuebingen.de.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
We thank Dr. K. Eichele and the Institut fur Anorganische
Chemie for giving us access to their NMR spectrometer.
■
̈
Financial support by the state of Baden-Wurttemberg is greatly
̈
acknowledged.
REFERENCES
■
(1) For reviews on gold catalysis, see: (a) Rudolph, M.; Hashmi, A. S.
K. Chem. Soc. Rev. 2012, 41, 2448−2462. (b) Corma, V.; Leyva-Perez
́
,
A.; Sabater, M. J. Chem. Rev. 2011, 111, 1657−1712. (c) Bandini, M.
Chem. Soc. Rev. 2011, 40, 1358−1367. (d) Boorman, T. C.; Larrosa, I.
Chem. Soc. Rev. 2011, 40, 1910−1925. (e) Shapiro, N. D.; Toste, F. D.
Synlett 2010, 2010, 675−691. (f) Sengupta, S.; Shi, X. ChemCatChem
2010, 2, 609−619. (g) Bongers, N.; Krause, N. Angew. Chem., Int. Ed.
2008, 47, 2178−2181. (h) Gorin, D. J.; Sherry, B. D.; Toste, F. D.
́
́
Chem. Rev. 2008, 108, 3351−3378. (i) Jimenez-Nunez, E.; Echavarren,
̃
A. M. Chem. Rev. 2008, 108, 3326−3350. (j) Li, Z.; Brouwer, C.; He,
C. Chem. Rev. 2008, 108, 3239−3265. (k) Arcadi, A. Chem. Rev. 2008,
108, 3266−3325. (l) Muzart, J. Tetrahedron 2008, 64, 5815−5849.
(m) Shen, H. C. Tetrahedron 2008, 64, 7847−7870. (n) Widenhoefer,
R. A. Chem. - Eur. J. 2008, 14, 5382−5391. (o) Gorin, D. J.; Toste, F.
was demonstrated that formation of G will induce changes in
the concentrations of in-cycle organogold intermediates and
H+, and only these changes are responsible for the observed
effect (not silver itself). The effect can be positive or negative
(see explanations in the main text). No silver effect takes place
if there is no accumulation of G in the reaction mixture. We
foresee that the formation of mixed silver−gold acetylide
complexes [RCC(AuL)(Ag)]+ might contribute to a silver
effect in reactions of some terminal alkynes.21 This subclass
remains to be explored.
Despite the ability to influence the rate of gold-catalyzed
hydroalkoxylation, Ag+ is found to be essentially innocent with
regard to the mechanism of the catalytic process itself. This is
thought to be valid for many reaction types, including those for
which the simultaneous presence of silver and gold components
was previously (erroneously) suggested to be crucial.4 Some
reactions that are separately catalyzed either by silver or by gold
are beyond the scope of this conclusion (for those reactions a
“true” silver effect associated with direct participation of silver
in the catalytic process would indeed take place).22
In addition to the formation of G, a silver effect may be
associated with an incomplete eq 1 or reactivation of a gold
catalyst by silver from halide poisoning. Whatever reason
applies, the chemical sense of a silver effect is the same: a
change (decrease or increase) in the concentration of nonsilver
catalytic cycle participants. Therefore, a gold catalysis
practitioner should not think of silver as a necessary component
of the catalytic cycle but rather must take care of speciation of
the catalytic system to ensure the maximum concentration of
catalytic cycle participants.
D. Nature 2007, 446, 395−403. (p) Furstner, A.; Davies, P. W. Angew.
̈
́
́
Chem., Int. Ed. 2007, 46, 3410−3449. (q) Jimenez-Nunez, E.;
̃
Echavarren, A. M. Chem. Commun. 2007, 333−346. (r) Hashmi, A.
S. K. Chem. Rev. 2007, 107, 3180−3211. (s) Hashmi, A. S. K.;
Hutchings, G. J. Angew. Chem., Int. Ed. 2006, 45, 7896−7936.
́
(2) Weber, D.; Gagne, M. R. Org. Lett. 2009, 11, 4962−4965.
(3) Examples of Au−Ag complexes with a bridging carbon:
(a) Contel, M.; Garrido, J.; Gimeno, M. C.; Jones, P. G.; Laguna,
A.; Laguna, M. Organometallics 1996, 15, 4939−4943. (b) Contel, M.;
Jimen
Trans. 1994, 2515−2518. (c) Fernan
A.; Lagunas, M. C.; Lopez-de-Luzuriaga, J. M.; Monge, M.; Montiel,
M.; Olmos, M. E.; Puelles, R. C.; Sanchez-Forcada, E. Chem. - Eur. J.
2009, 15, 6222−6233. (d) Fernandez, E. J.; Laguna, A.; Lopez-de-
Luzuriaga, J. M.; Montiel, M.; Olmos, M. E.; Perez, J.; Puelles, R. C.
Organometallics 2006, 25, 4307−4315. (e) Ruiz, J.; Riera, V.; Vivanco,
M.; García-Granda, S.; García-Fernandez, A. Organometallics 1992, 11,
4077−4082. (f) Vicente, J.; Chicote, M. T.; Alvarez-Falcon, M. M.;
́
ez, J.; Jones, P. G.; Laguna, A.; Laguna, M. J. Chem. Soc., Dalton
́
dez, E. J.; Hardacre, C.; Laguna,
́
́
́
́
́
́
́
Jones, P. G. Organometallics 2005, 24, 4666−4675. (g) Xie, Z.-L.; Wei,
Q.-H.; Zhang, L.-Y.; Chen, Z.-N. Inorg. Chem. Commun. 2007, 10,
1206−1209. (h) Mazhar-Ul-Haque; Horne, W.; Abu-Salah, O. M. J.
Crystallogr. Spectrosc. Res. 1992, 22, 421−425.
(4) 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, 9012−9019.
(5) Homs, A.; Escofet, I.; Echavarren, A. M. Org. Lett. 2013, 15,
5782−5785.
(6) Kumar, M.; Hammond, G. B.; Xu, B. Org. Lett. 2014, 16, 3452−
3455.
(7) For activation of gold hydroxo complexes with a Brønsted acid,
́
see: Gaillard, S.; Bosson, J.; Ramon, R. S.; Nun, P.; Slawin, A. M. Z.;
Nolan, S. P. Chem. - Eur. J. 2010, 16, 13729−13740.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acscatal.5b01493.
■
(8) For anion exchange of LAuCl with an alkali-metal salt, see:
S
(a) Furstner, A.; Alcarazo, M.; Goddard, R.; Lehmann, C. W. Angew.
̈
Chem., Int. Ed. 2008, 47, 3210−3214. (b) Kleinbeck, F.; Toste, F. D. J.
Am. Chem. Soc. 2009, 131, 9178−9179. (c) Lau, V. M.; Gorin, C. F.;
6003
DOI: 10.1021/acscatal.5b01493
ACS Catal. 2015, 5, 5994−6004

ACS Catalysis
Research Article
Kanan, M. W. Chem. Sci. 2014, 5, 4975−4979. For activation of
LAuCl by other Lewis acids, see: (d) Lavallo, V.; Frey, G. D.; Kousar,
S.; Donnadieu, B.; Bertrand, G. Proc. Natl. Acad. Sci. U. S. A. 2007, 104,
13569−13573. (e) Guer
Bezzenine-Lafollee, S.; Gandon, V. Angew. Chem., Int. Ed. 2013, 52,
5848−5852. (f) Fang, W.; Presset, M.; Guerinot, A.; Bour, C.;
Bezzenine-Lafollee, S.; Gandon, V. Chem. - Eur. J. 2014, 18, 5439−
́
inot, A.; Fang, W.; Sircoglou, M.; Bour, C.;
́
́
́
5446. For activation of LAu(phthalimide) by Brønsted and Lewis
acids, see: (g) Han, J.; Shimizu, N.; Lu, Z.; Amii, H.; Hammond, G. B.;
Xu, B. Org. Lett. 2014, 16, 3500−3503.
(9) Zhdanko, A.; Maier, M. E. Chem. - Eur. J. 2014, 20, 1918−1930.
(10) Zhdanko, A.; Maier, M. E. Chem. - Eur. J. 2013, 19, 3932−3942.
(11) Seidel, G.; Lehmann, C. W.; Furstner, A. Angew. Chem., Int. Ed.
̈
2010, 49, 8466−8470.
́
(12) Brown, T. J.; Weber, D.; Gagne, M. R.; Widenhoefer, R. A. J.
Am. Chem. Soc. 2012, 134, 9134−9137.
(13) The stability of various diaurated species was quantitatively
evaluated: Zhdanko, A.; Maier, M. E. Organometallics 2013, 32, 2000−
2006.
(14) In connection with this intermetallic ligand exchange of Nu
between formal LAu+ and Ag+, we established that the LAu+ unit has
higher potency than a “naked” Ag+ to bind a Nu (Supporting
Information).
(15) Zhdanko, A.; Strobele, M.; Maier, M. E. Chem. - Eur. J. 2012, 18,
̈
14732−14744.
(16) Formation of G4 in place of the more stable D4 should not be
misleading, because under the catalytic reaction conditions G4 may
become kinetically or even thermodynamically favored. For example,
note that G4 forms immediately and is reluctant to equilibrate to D4.
The equilibration toward the more stable D4 may be strongly retarded
because there are always negligible concentrations of free vinyl gold B4
and free LAu(Sol)+. In addition, this does not necessarily contradict
the thermodynamics, because an additional driving force is certainly
provided by binding LAu+ by the reactive substrate, eventually turning
both gold and silver into G4, disfavoring formation of D4 to some
extent (Le Chatelier’s principle). This driving force, however, was
unable to cause any changes in resting state in the aforementioned
reaction with catalyst 1 (D2 forms quickly and is simply much more
stable than G2).
(17) Weber, S. G.; Rominger, F.; Straub, B. F. Eur. J. Inorg. Chem.
2012, 2012, 2863−2867.
(18) Zhu, Y.; Day, C. S.; Zhang, L.; Hauser, K. J.; Jones, A. C. Chem. -
Eur. J. 2013, 19, 12264−12271.
(19) Sota, Y.; Yamamoto, M.; Murai, M.; Uenishi, J.; Uemura, M.
Chem. - Eur. J. 2015, 21, 4398−4404.
(20) Isomerization of E1 is found to be purely a gold-catalyzed and
not a proton-catalyzed process under these conditions (unpublished).
̌ ́ ́
(21) Jasíkova, L.; Roithova, J. Organometallics 2013, 32, 7025−7033.
(22) For examples of such reactions, see: (a) Pennell, M. N.; Turner,
P. G.; Sheppard, T. D. Chem. - Eur. J. 2012, 18, 4748−4758. (b) Su, Y.;
Zhang, Y.; Akhmedov, N. G.; Petersen, J. L.; Shi, X. Org. Lett. 2014,
16, 2478−2481.
6004
DOI: 10.1021/acscatal.5b01493
ACS Catal. 2015, 5, 5994−6004