Alkyne/alkene/allene-induced disproportionation of cationic gold(i) catalyst

Article abstract of DOI:10.1002/chem.201304271

The first detailed experimental study of the deactivation of cationic gold was conducted, and the influence of each component in the reaction system (substrate, counterion, solvent) on the decay process was examined. It was found that a substrate (alkyne/allene/alkene)-induced disproportionation of gold(I) may play a key role in the decay process. Our mechanism is supported by kinetic, XPS, voltammetry studies, and high-resolution ESI-MS data. The first detailed experimental study of the deactivation of cationic gold was conducted, and the influence of each component in the reaction system (substrate, counterion, solvent) on the decay process was examined. It was found that a substrate (alkyne/allene/alkene)-induced disproportionation of gold(I) may play a key role in the decay process (see scheme; TFa=atrifluoromethanesulfonyl).

Full text of DOI:10.1002/chem.201304271

DOI: 10.1002/chem.201304271
Full Paper
&
Homogeneous Catalysis
Alkyne/Alkene/Allene-Induced Disproportionation of Cationic
Gold(I) Catalyst
Manish Kumar,[a] Jacek Jasinski,[b] Gerald B. Hammond,*[a] and Bo Xu*[a]
Abstract: The first detailed experimental study of the deacti-
vation of cationic gold was conducted, and the influence of
each component in the reaction system (substrate, counter-
ion, solvent) on the decay process was examined. It was
found that a substrate (alkyne/allene/alkene)-induced dispro-
portionation of gold(I) may play a key role in the decay pro-
cess. Our mechanism is supported by kinetic, XPS, voltam-
metry studies, and high-resolution ESI-MS data.
Introduction
Gold catalysis, a landmark addition to the field of organic syn-
thesis,[1] is not without shortcomings, the main one being the
relatively low turnover numbers observed in the majority of
gold-catalyzed reactions. Indeed, high turnover numbers (or
low catalyst loadings) are observed only for few types of gold-
catalyzed reactions (e.g., hydration of alkyne).[2] In most gold-
catalyzed reactions, loadings in the 1–5% range are the
norm.[3] One reason for the low turnover numbers (TON) is the
relatively fast decay of cationic gold and its conversion into
nonreactive species (e.g., Au0—gold mirror and/or gold parti-
cles and [L2 AuI]+.[4] Despite this drawback, there has been no
comprehensive systematic investigations on the decay of ho-
mogeneous gold catalysis; a stark contrast with the extensive
work reported on catalyst deactivation in heterogeneous catal-
ysis.[5] Herein, we answer some fundamental and intriguing
questions on gold(I) decay, such as the effects of starting mate-
rial, nucleophile, solvent, counterion, and additives, and we
suggest a plausible pathway for gold(I) decay.
Scheme 1. Simplified representation of the gold catalytic cycle.
a gold s complex ([L-Au-S-Nu]) that eventually undergoes pro-
todeauration and the regeneration of cationic gold.
Complex [L-AuI]+XÀ (X=weakly coordinating counterion,
such as OTfÀ) is relatively stable by itself, but in many reactions
it decays at a fast rate (from few minutes to few hours). This
behavior suggests that a cationic gold catalyst may decay
rather rapidly in the presence of certain components in a reac-
tion system. Because [Au+(Ph3P)OTfÀ] (1) is one of the most
commonly used cationic gold catalysts,[3a,b,d] we chose it as our
model catalyst to study the deactivation process. Compound
1 in CDCl3 is relatively stable (31P NMR peak at d=27.4 ppm is
the only peak)[7] and also catalytically active for 24 h at room
temperature). First, we investigated the kinetic stability of [L-
Au+-S] (Scheme 1, S=alkyne/allene/alkene). After adding cy-
clohexene (10 equiv vs. 1) at room temperature, we noticed
significant amounts of gold mirror/black particles (Au0) in the
NMR tube, a telltale indication that 1 (0.01m in CDCl3) had un-
dergone significant decay. In the beginning, only a single
31P NMR peak at 29.5 ppm was recorded, consistent with
a gold-alkene p-complex or a fast equilibrium between free
[Au+(Ph3P)OTfÀ] and [Au(Ph3P)Àalkene]+OTfÀ (Figure 1). The
signal at d=29.5 ppm [gold(I)–alkene complex] decreased over
time, and a new signal at d=45.0 ppm increased over the
same time period. We assigned this new signal the structure
[Au+(Ph3P)2OTfÀ] (2). This assignment is in agreement with the
31P NMR shift value of 2 reported in the literature[8] and also
with the chemical shift recorded for our independently synthe-
Results and Discussion
A simplified gold catalytic cycle is shown in Scheme 1. The cat-
ionic gold complex (L-Au+) acts as carbophilic Lewis acid,[6] ac-
tivating an unsaturated CÀC bond to form a p complex [L-Au+
-S]. The interaction of [L-Au+-S] with a nucleophile (Nu) gives
[a] M. Kumar, Prof. Dr. G. B. Hammond, Prof. Dr. B. Xu
Department of Chemistry, University of Louisville
Louisville, Kentucky 40292 (USA)
E-mail: bo.xu@louisville.edu
gb.hammond@louisville.edu
[b] Dr. J. Jasinski
Materials Characterization
Conn Center for Renewable Energy Research
University of Louisville, Louisville, Kentucky 40292 (USA)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201304271.
Chem. Eur. J. 2014, 20, 3113 – 3119
3113
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
vent, counterion, nucleophile,
and additive. In all the solvents
that we tested, 1 was stable for
24 h at room temperature. When
cyclohexene (10 equiv vs. gold)
was added, the decay process
began in earnest. A highly coor-
dinating solvent, such as CH3CN,
showed good aptitude in pre-
venting the deactivation of gold
catalyst (Figure 2c). As was men-
tioned above, [Au+(Ph3P)XÀ]
alone is relatively stable, but
when cyclohexene was added,
different degrees of decay were
observed in all cases (Figure 2d).
Compared to the commonly
Figure 1. Decay of 1 in the presence of cyclohexene, monitored by 31P NMR spectroscopy.
À
used OTfÀ counterion, NTf2 has
a greater ability to stabilize the
sized 2. This assignment was further confirmed by high-resolu-
tion ESI mass spectroscopy ([M-OTf]+ at m/z=721.1483; see
Figure S4 in the Supporting Information). We also recorded the
1H NMR of the reaction mixture, and found that cyclohexene
had remained unchanged during the decay process. However,
we noticed that higher concentrations of cyclohexene sped up
the decay (Figure 2a). Other unsaturated substrates commonly
used in gold catalysis, such as alkyne and allene (10 equiv vs.
1), also induced the decay of cationic gold 1 (Figure 2b).
cationic gold catalyst (Figure 2d).[9] Just as it was the case with
solvents and counterions, most of the additives we screened
did not induce the decay of cationic gold, but when we tested
how they influenced the decay kinetics in the presence of cy-
clohexene, we found that the decay of cationic gold slowed
down differently depending on the type of additives used (Fig-
ure 2e). Many N-heterocycles and hexamethylphosphoramide
(HMPA) stabilized the gold catalyst significantly; similarly, other
additives (LiNTf2, N-oxide, sodium sufinimide, KCTf3, N,N’-Di-
methyl-N,N’-trimethyleneurea (DMPU)) slowed down the decay
of gold catalyst 1 to different extents.
Next, we investigated the influence of other common com-
ponents on the kinetics of gold-catalyst decay, specifically, sol-
Figure 2. a) Decay of [Au+(Ph3P)OTfÀ] (1) in the presence of cyclohexene; b) decay of 1 in the presence of alkene/alkyne/allene; c) decay of 1 in various sol-
vents in the presence of cyclohexene; d) decay of [Au+(Ph3P)XÀ] in the presence of cyclohexene; e) decay of 1 in the presence of cyclohexene and additives;
and f) water effect on decay of 1.
Chem. Eur. J. 2014, 20, 3113 – 3119
www.chemeurj.org
3114
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
1
Because trace water is ubiquitous in the reaction system, we
investigated the effect of water on catalyst deactivation. We
found that water alone has little effect on the decay (Fig-
ure 2 f). Nevertheless, we observed a significantly faster deacti-
vation of gold catalyst 1, when 1 (0.01m) was mixed with both
cyclohexene (10 equiv) and water (1 equiv) at room tempera-
ture.
Because 31P, H NMR, or voltammetry data do not reflect the
change in the chemical state of gold itself, we used other
more direct techniques to investigate the change in the chemi-
cal state of gold. X-ray photoelectron spectroscopy (XPS) is
a quantitative spectroscopic technique that measures the ele-
mental composition, chemical state, and electronic state of the
elements that exist within a material.[13] We[14] and others[15]
have used XPS in the determination of the chemical state of
gold catalysts. The Au 4 f7/2 peaks can be resolved into a dou-
blet (spin–orbit splitting) and the binding energy of Au 4 f7/2
electrons in each gold oxidation state are usually sufficiently
wide apart to be differentiated using this technology.
We postulated two possible pathways to explain the decay
behavior (generation of Au0 from AuI). One pathway is dispro-
portionation, and the other is reduction of gold(I) by certain
reductants in the reaction system. Ceroni and co-workers[10] re-
ported that gold nanoparticles (Au0) can be produced by dis-
proportionation of Au+ ions. Some research groups have also
speculated that the deactivation of gold catalyst is due to dis-
proportionation.[4b,c] However, no direct experimental evidence
has been reported to support these statements. We decided to
investigate the actual gold valence change in the decay pro-
cess experimentally.
First, we tested our gold-standard samples ([AuICl(PPh3)] and
NaAuIIICl4) and found that the Au 4 f7/2 photoelectron peak was
located at a binding energy (BE) value of 85.7 and 87.5 eV, re-
spectively, which is quite consistent with literature reports (see
the Supporting Information).[16] Then, we tested our cationic
gold catalyst with various counterions. We prepared these cat-
ionic gold samples in chloroform; 24 h later, these solutions
were dried in high vacuum and were subjected to XPS analysis
(Figure 4a–c). As was expected, the gold valence in these
three samples is still AuI.
The redox potential of cationic gold should provide us with
quantitative information to determine the easiness of reduc-
tion of AuI in a reaction system. Laguna and co-workers have
studied the redox chemistry of AuI and AuIII complexes by
using cyclic voltammetry in CH2Cl2.[11] We also chose CH2Cl2 as
solvent to have a frame of reference to measure the reductive
potential of cationic AuI. The cyclic voltammogram of 1 re-
vealed a single irreversible redox event at a very negative
formal potential (À2.5 V vs. ferrocenium/ferrocene; Figure 3).
Compared with the formal potential of Ag+ in CH2Cl2
(+0.65 V),[12] the highly negative potential of 1 denotes that
a direct reduction of catalyst 1 could be very difficult. This ob-
Compound 1 (0.01m in CDCl3) was treated with cyclohexene
(10 equiv), and after 48 h the mixture was filtered, and the
liquid portion (sample 1, Figure 4d) and solid precipitate were
collected (sample 2, Figure 4e). We conducted XPS analyses of
Figure 3. Cyclic voltammogram (red line) of 1 (2 mm in CH2Cl2 with 0.1m tet-
rabutylammonium hexafluorophosphate (TBAHFP)). Blue line is background
CV of supporting electrolyte in CH2Cl2. Potential values are referenced versus
the ferrocenium/ferrocene couple.
servation was further confirmed by adding a strong reductant
to 1. When [CoCp*2] (reduction potential in CH2Cl2 is
À1.94 V)[12] was added to a solution of 1, no reduction took
place. [CoCp*2] is already a very strong reductant, most com-
pounds in a reaction system are far less reductive. Also, in our
decay-kinetic studies, we did not observe any change in the
1H NMR of cyclohexene in the solution. This combined data
suggests that the direct reduction of AuI is unlikely.
Figure 4. XPS spectra (AlKa =1486.6 eV) corresponding to the Au 4 f7/2 photo-
electron peaks.
Chem. Eur. J. 2014, 20, 3113 – 3119
www.chemeurj.org
3115
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
both samples. The majority of the solid sample (sample 2) was
Au0 (Figure 4e), and the liquid fraction was AuI (Figure 4d). Be-
cause we detected none or very little of AuIII component, it
was highly possible that a reductive agent, such as the phos-
phine ligand, reduced AuIII (because of the high oxidation
power of cationic AuIII itself). Next, we investigated the dispro-
portionation of gold(I) in the absence of phosphine ligand.
First, we tested a commercial AuICl sample; although the ma-
jority of it corresponded to AuI, there was also a small amount
of AuIII present (Figure 4 f). We wanted to investigate the be-
havior of cationic gold(I) (AuIOTf) under these conditions. AuICl
and AgOTf (1 equiv) were mixed in dry dichloromethane with
and without cyclohexene. Thus, after stirring for 30 min under
dark and anhydrous conditions, we observed the formation of
black particles. We collected the black particles (sample 3) and
conducted a XPS analysis of sample 3 (Figure 4 g), which de-
noted the presence of Au0, AuI, and AuIII. We conducted the
same experiment but in the company of cyclohexene and ob-
Figure 5. ESI-MS spectrum of sample 3.
Based on the combination of all the above described experi-
mental data, we are now in a good position to understand the
decay mechanism, which can be summed up as follows:
1
served a similar result. H NMR analysis indicated no change in
the structure of cyclohexene in the decay solution. This result
suggested that cyclohexene did not play a role as reductant.
All together, the experimental data supported the dispropor-
tionation of cationic gold(I) in the absence of phosphine
ligand.
1) Cationic AuI without a suitable ligand (e.g., AuIOTf) dispro-
portionates readily compared to its noncationic form (e.g.,
AuICl; Figure 4, sample 3).
Because XPS was conducted under high vacuum, and the
samples were tested in the solid state, there was some concern
that the gold catalyst might have changed under these condi-
tions. We wanted to further confirm the existence of gold(III)
by investigating the gold species directly in solution phase.
Electrospray ionization mass spectroscopy (ESI-MS) is a soft
ionization technique,[17] and is an especially suitable technique,
because cationic gold intermediates are charged species. In
a previous work, we succeeded in detecting the existence of
gold(III) species by using high-resolution (HR) ESI-MS.[18] First,
we checked the HR ESI-MS spectra of sample 1 (Figure 4). We
only detected the presence of various AuI species (e.g., [AuI-
(Ph3P)]+ (m/z 459.0571), [AuI(Ph3P)H2O]+ (m/z 477.0677), [AuI-
(Ph3P)2]+ (m/z 721.1483)). This observation is consistent with
our previous XPS analysis (Figure 4d). According to our previ-
ous studies of gold valence by using ESI-MS,[14] AuIII species are
difficult to detect under standard condition, but we have been
able to detect AuIII species by adding a bidentate ligand and
a halide (bipyridine and chloride).[14] Because AuIII complexes
have square-planar geometry, a bidentate ligand should great-
ly stabilize the AuIII cation during the ionization process, and
chloride will further stabilize cationic AuIII.[14] Therefore, we
added bipyridine and chloride to sample 1. Again, only AuI
species (e.g., [AuI(Ph3P)bipyridine]+ (m/z 615.1266), [AuI-
(Ph3P)2]+ (m/z 721.1483)) were detected. This result indicated
that there was no AuIII in the presence of phosphine ligand
(PPh3). We used the same method to investigate the ESI-MS of
sample 3. Various AuIII species were detected (Figure 5). Our HR
ESI-MS studies are consistent with our previous XPS studies,
that is, there was no AuIII species present (only AuI species) in
sample 1, but various AuIII species were detected in sample 3
(for more details, see the Supporting Information, section 5).
2) In the presence of a suitable ligand, cationic gold(I) decays
(or disproportionates) at a much slower pace. Complex
[Au+(Ph3P)OTfÀ] is stable for at least 24 h (Figure 2a), and
AuOTf disproportionates in less than 0.5 h (Figure 4). This
conclusion explains why ligands are so ubiquitous in gold
catalysis.
3) Cationic AuI p complexes with p donors (CÀC unsaturated
compounds, such as alkyne/allene/alkene) undergo relative
fast decay, whereas cationic AuI complexes with most
s donors (e.g., N-heterocycles and acetonitrile) are very
stable.
The detailed role of substrates (alkyne/allene/alkene) in sub-
strate-induced cationic gold decay is not very clear yet. Consid-
ering the great stabilization effect of the phosphine ligand (or
NHC ligand) against disproportionation, our hypothesis is that
an alkyne/allene/alkene may help to temporarily displace the
phosphine ligand from the gold(I) center by a trans effect. The
trans effect is the labilization (destabilization) of ligands that
are trans to certain other ligands in a metal complex. Although
most examples of trans effect are observed in square-planar
complexes (e.g., Pd, Pt), we propose that it may also operate
in linear metal complexes, such as gold(I) complexes.[19] And al-
kenes/alkynes are indeed the ligands that have shown the
strongest trans effect. Odell and co-workers demonstrated that
alkene and alkyne showed highest trans effect than any other
compounds in platinum complexes (trans effect: C2H4 ~iPrCH=
CHMe~Me2C(OH)CꢀC-C(OH)Me2 @Et3Sb>Ph3Sb>Me3P>
Et3P>Ph3P).[20]
Our proposed decay mechanism is shown in Scheme 2. First,
cyclohexene will complex with cationic gold to form a gold–
alkene p complex, then, because of the trans effect of the
Chem. Eur. J. 2014, 20, 3113 – 3119
www.chemeurj.org
3116
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
co-workers[26] have found that the gold resting
state is the gold–allene complex 5 in this reaction
(Scheme 3a). Intermolecular hydroamination of al-
kene[3a] is a similar example (Scheme 3b). In these
reactions, high loadings of gold catalyst (5% or
higher) are typically used. And in these reactions,
we did observe the formation of decay gold prod-
ucts 2 and Au0 in the reaction mixture, especially
Scheme 2. Proposed mechanism for substrate-induced cationic gold decay.
alkene, the phosphine ligand will be temporarily replaced to
form complex 2 and free AuOTf. The disproportionation of
AuOTf generates Au0 and AuIII(OTf)3. This disproportionation
may be enhanced by the presence of trace water in the
system. This statement is supported by a literature report that
stated that a AuI complex underwent disproportionation in
aqueous solution, whereas AuI was stable in organic sol-
vents.[21] The disproportionation of AuOTf or its alkene/alkyne
complex can be a relatively fast process (Figure 4, sample 3).
This statement also has literature support: Dias and co-work-
ers[22] reported that AuI alkene or alkyne complexes without li-
gands are only stable at very low temperature, turning into
black particles at room temperature. It is known that AuIII will
oxidize the phosphine ligand,[23] and itself will be reduced to
AuI again. We propose that Ph3P is oxidized to Ph3P(OTf)2,[24]
which could then be hydrolyzed to give Ph3P=O in the pres-
ence of water. The presence of Ph3P=O has been confirmed by
ESI-MS and TLC (by comparison with an authentic sample). Fur-
thermore, when Au(OTf)3 (generated from AuCl3/AgOTf) was
mixed with PPh3, [Au+(Ph3P)2OTfÀ] (2) and Ph3P=O were ob-
tained. This behavior could explain why we were not able to
detect AuIII in gold solutions using XPS and ESI-MS (sample 1).
We now can apply our decay mechanistic study to explain
some unanswered questions:
Scheme 3. Catalyst decay in different gold catalyzed reactions.
when triflate was used as counterion. A typical example of
type II reactions is the cyclization of propargyl amide (7)
(Scheme 3c); its catalyst resting state is a [L-Au-S-Nu] type
intermediate 9.[25,27] Thus, the decay of active catalyst is ex-
pected to be slow. Indeed, low catalyst loadings (0.1%)
have been achieved for this reaction.[28] Similarly, the alkyne
hydration (Scheme 3d) reported by Nolan and co-work-
ers[2c] and the intramolecular addition of a diol to an alkyne
reported by Hashmi and co-workers[2d,29] have been ach-
ieved with very high TONs. In these reactions, the turnover-
limiting determining stage is most likely the protodeaura-
tion of [L-Au-S-Nu]-type complex (Scheme 1).[25]
1) What causes the decay of cationic gold? The CÀC unsatu-
rated compounds (alkyne/allene/alkene)-induced cationic
gold decay is the major reason for cationic gold decay. It
should be noted that the starting material is not the only
source of CÀC unsaturated compounds; the products often
contain CÀC unsaturated compounds as well.
2) Why is it that a high turnover can be achieved in some re-
actions but not in others? The decay process is highly de-
pendent on the resting state of the gold catalyst. Based on
our most recent study,[25] we have classified gold-catalyzed
reactions according to their turnover limiting stage: in
type I, the formation of gold–s complex ([L-Au-S-Nu] in
Scheme 1) is the turnover limiting stage; in type II, the re-
generation of cationic gold catalyst (e.g., protodeauration)
from [L-Au-S-Nu] is the turnover-limiting stage. For type I
reactions, the resting state is often [L-Au+-S] (Scheme 1);
thus, a relatively fast decay is to be expected, and so, turn-
over numbers will be low in general. On the other hand, in
type II reactions, the resting state is usually not [L-Au+-S],
so the decay is relatively slow and high TON are to be ex-
pected. According to our recent study,[25] the hydroamina-
tion of allene is a type I reaction (Scheme 3a). Toste and
3) What can we do to reduce the decay of an active gold cat-
alyst? Our studies indicate that the use of s donors, such
as N-heterocycles and CH3CN, will enhance the stability of
the cationic gold,[30] but s donors also have a significant
detrimental effect on reactivity, so their use is normally not
a good option. But it is still possible to disrupt the gold–
alkyne/alkene interaction and therefore reduce the decay
by using suitable ligands and counterions. Recently, we and
others have reported that the use of a ligand containing
a o-biphenyl steric handle can reduce the decay of cationic
gold.[25,31] As shown in Scheme 3e, the o-biphenyl motif will
clearly interrupt the bonding of the gold center with an
alkene/alkyne, which may be the reason it slowed down its
decay. A steric handle, such as o-biphenyl will not reduce
the reactivity though; actually, for many types of gold-cata-
Chem. Eur. J. 2014, 20, 3113 – 3119
www.chemeurj.org
3117
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
J. H. Wright, F. S. Tham, S. Quinlivan, Angew. Chem. Int. Ed. 2013, 52,
3172–3176.
lyzed reactions, ligands with biphenyl motif often give
faster reactions.[1c,32] In Figure 2d, we demonstrated that
some counterions are capable of stabilizing the cationic
[3] a) J. Zhang, C.-G. Yang, C. He, J. Am. Chem. Soc. 2006, 128, 1798–1799;
b) X. Yao, C.-J. Li, J. Am. Chem. Soc. 2004, 126, 6884–6885; c) C. F.
Bender, R. A. Widenhoefer, Org. Lett. 2006, 8, 5303–5305; d) H. H. Jung,
P. E. Floreancig, Org. Lett. 2006, 8, 1949–1951; e) T. Dombray, A. L.
Blanc, J.-M. Weibel, P. Pale, Org. Lett. 2010, 12, 5362–5365.
[4] a) H. Duan, S. Sengupta, J. L. Petersen, N. G. Akhmedov, X. Shi, J. Am.
Chem. Soc. 2009, 131, 12100–12102; b) T. J. Brown, M. G. Dickens, R. A.
Widenhoefer, J. Am. Chem. Soc. 2009, 131, 6350–6351; c) N. D. Shapiro,
F. D. Toste, Proc. Natl. Acad. Sci. USA 2008, 105, 2779–2782; d) A. Guꢂri-
not, W. Fang, M. Sircoglou, C. Bour, S. Bezzenine-Lafollꢂe, V. Gandon,
Angew. Chem. Int. Ed. 2013, 52, 5848–5852; e) G. Lemiꢃre, V. Gandon, N.
Agenet, J.-P. Goddard, A. de Kozak, C. Aubert, L. Fensterbank, M. Malac-
ria, Angew. Chem. 2006, 118, 7758–7761; Angew. Chem. Int. Ed. 2006,
45, 7596–7599.
À
À
gold. NTf2 is a notable example. The use of NTf2 will not
À
reduce reactivity; in many reactions, NTf2 actually enables
faster reactions compared with a commonly used counter-
ion, such as OTfÀ.[9]
It should be noted that the alkyne/alkene/allene-induced dis-
proportionation of cationic gold(I) species may generate small
gold clusters or gold nanoparticles under proper conditions; in
turn, these clusters exhibit high reactivity in some gold-cata-
lyzed reactions.[2e,33] Therefore, disproportionation may not be
the end of the story for gold catalysts. We believe that if rela-
tively large gold particles or a gold mirror are generated
through disproportionation, the reactivity of the gold catalyst
will be lost, but if we can control this disproportionation pro-
cess so as to generate small gold clusters or gold nanoparti-
cles, we may then be able to engender new reactivities.
[5] C. H. Bartholomew, Appl. Catal. A 2001, 212, 17–60.
[6] A. S. K. Hashmi, Gold Bull. 2003, 36, 3–9.
[7] A. S. K. Hashmi, A. M. Schuster, S. Litters, F. Rominger, M. Pernpointner,
Chem. Eur. J. 2011, 17, 5661–5667.
[8] a) C.-G. Yang, C. He, J. Am. Chem. Soc. 2005, 127, 6966–6967; b) W. E.
Brenzovich, D. Benitez, A. D. Lackner, H. P. Shunatona, E. Tkatchouk,
W. A. Goddard, F. D. Toste, Angew. Chem. 2010, 122, 5651–5654; Angew.
Chem. Int. Ed. 2010, 49, 5519–5522; c) M. Bardajꢁ, P. Uznanski, C.
Amiens, B. Chaudret, A. Laguna, Chem. Commun. 2002, 598–599; d) S.
Attar, W. H. Bearden, N. W. Alcock, E. C. Alyea, J. H. Nelson, Inorg. Chem.
1990, 29, 425–433.
[9] N. Mꢂzailles, L. Ricard, F. Gagosz, Org. Lett. 2005, 7, 4133–4136.
[10] G. Bergamini, P. Ceroni, V. Balzani, M. Gingras, J.-M. Raimundo, V. Moran-
di, P. G. Merli, Chem. Commun. 2007, 0, 4167–4169.
[11] U. Koelle, A. Laguna, Inorg. Chim. Acta 1999, 290, 44–50.
[12] N. G. Connelly, W. E. Geiger, Chem. Rev. 1996, 96, 877–910.
[13] J.-H. Choy, Y.-I. Kim, J. Phys. Chem. B 2003, 107, 3348–3350.
[14] W. Wang, J. Jasinski, G. B. Hammond, B. Xu, Angew. Chem. 2010, 122,
7405–7410; Angew. Chem. Int. Ed. 2010, 49, 7247–7252.
[15] a) M. P. Casaletto, A. Longo, A. Martorana, A. Prestianni, A. M. Venezia,
Surf. Interface Anal. 2006, 38, 215–218; b) M. P. de Almeida, L. M. D. R. S.
Martins, S. A. C. Carabineiro, T. Lauterbach, F. Rominger, A. S. K. Hashmi,
A. J. L. Pombeiro, J. L. Figueiredo, Catal. Sci. Technol. 2013, 3, 3056–
3069.
Conclusion
The substrate (alkyne/allene/alkene)-induced cationic gold
decay (disproportionation) may be the main reason behind
cationic gold deactivation, but it should be noted that there
could be other decay pathways. Further investigations on how
to reduce gold-catalyst decay in particular and homogeneous
metal-catalyst decay in general are ongoing in our laboratory.
Acknowledgements
[16] J.F. Moulder, J. Chastain, Handbook of X-Ray Photoelectron Spectroscopy,
Physical Electronics Division, Perkin-Elmer, Eden Prairie, 1992.
[17] a) R. Colton, A. D’Agostino, J. C. Traeger, Mass Spectrom. Rev. 1995, 14,
79–106; b) W. Henderson, B. K. Nickleson, L. J. McCaffrey, Polyhedron
1998, 17, 4291–4313.
[18] B. Xu, W. Wang, G. B. Hammond, J. Fluorine Chem. 2011, 132, 804–810.
[19] R. Dçpp, C. Lothschꢄtz, T. Wurm, M. Pernpointner, S. Keller, F. Rominger,
A. S. K. Hashmi, Organometallics 2011, 30, 5894–5903.
We are grateful to the NSF for financial support (CHE-111316).
We acknowledge Professor Craig A. Grapperhaus (University of
Louisville) for allowing us the use of his electrochemistry facili-
ty and for helpful discussions. We also acknowledge the useful
electrochemistry discussions held with Professor Richard M.
Baldwin (University of Louisville) and the use of the CREAM
Mass Spectrometry Facility (University of Louisville), funded by
NSF/EPSCoR (EPS-0447479).
[20] T. P. Cheeseman, A. L. Odell, H. A. Raethel, Chem. Commun. 1968, 1496–
1498.
[21] C. H. Gammons, Y. Yu, A. E. Williams-Jones, Geochim. Cosmochim. Acta
1997, 61, 1971–1983.
Keywords: catalyst deactivation
disproportionation · homogeneous catalysis · gold
·
catalyst decay
·
[22] a) H. V. R. Dias, M. Fianchini, T. R. Cundari, C. F. Campana, Angew. Chem.
2008, 120, 566–569; Angew. Chem. Int. Ed. 2008, 47, 556–559; b) A.
Das, C. Dash, M. Yousufuddin, M. A. Celik, G. Frenking, H. V. R. Dias,
Angew. Chem. Int. Ed. 2012, 51, 3940–3943; c) A. Das, C. Dash, M. A.
Celik, M. Yousufuddin, G. Frenking, H. V. R. Dias, Organometallics 2013,
32, 3135–3144; d) M. Fianchini, H. Dai, H. V. R. Dias, Chem. Commun.
2009, 0, 6373–6375.
[23] P. Braunstein, H. Lehner, D. Matt, K. Burgess, M. J. Ohlmeyer, Inorganic
Synthesis, Vol. 27, Wiley, New York, 2007, pp. 218–221.
[24] Ph3P(OTf)2 and even Ph3PO were not detected in 31P NMR spectra in the
decay solutions, possible reason is that their 31P NMR peaks are broad
in reaction system.
[25] W. Wang, G. B. Hammond, B. Xu, J. Am. Chem. Soc. 2012, 134, 5697–
5705.
[26] Z. J. Wang, D. Benitez, E. Tkatchouk, W. A. Goddard III, F. D. Toste, J. Am.
Chem. Soc. 2010, 132, 13064–13071.
[27] A. S. K. Hashmi, J. P. Weyrauch, W. Frey, J. W. Bats, Org. Lett. 2004, 6,
4391–4394.
[1] a) A. S. K. Hashmi, Chem. Rev. 2007, 107, 3180–3211; b) M. Rudolph,
A. S. K. Hashmi, Chem. Soc. Rev. 2012, 41, 2448–2462; c) D. J. Gorin,
B. D. Sherry, F. D. Toste, Chem. Rev. 2008, 108, 3351–3378; d) Z. Li, C.
Brouwer, C. He, Chem. Rev. 2008, 108, 3239–3265; e) P. Garcia, M. Malac-
ria, C. Aubert, V. Gandon, L. Fensterbank, ChemCatChem 2010, 2, 493–
497; f) A. Arcadi, Chem. Rev. 2008, 108, 3266–3325.
[2] a) J. H. Teles, S. Brode, M. Chabanas, Angew. Chem. 1998, 110, 1475–
1478; Angew. Chem. Int. Ed. 1998, 37, 1415–1418; b) E. Mizushima, K.
Sato, T. Hayashi, M. Tanaka, Angew. Chem. 2002, 114, 4745–4747;
Angew. Chem. Int. Ed. 2002, 41, 4563–4565; c) N. Marion, R. S. Ramon,
S. P. Nolan, J. Am. Chem. Soc. 2009, 131, 448–449; d) M. C. Blanco
Jaimes, C. R. N. Bçhling, J. M. Serrano-Becerra, A. S. K. Hashmi, Angew.
Chem. Int. Ed. 2013, 52, 7963–7966; e) J. Oliver-Meseguer, J. R. Cabrero-
Antonino, I. Domꢁnguez, A. Leyva-Pꢂrez, A. Corma, Science 2012, 338,
1452–1455; f) A. S. K. Hashmi, Science 2012, 338, 1434; g) V. Lavallo,
Chem. Eur. J. 2014, 20, 3113 – 3119
www.chemeurj.org
3118
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[28] S. Doherty, C. H. Smyth, J. G. Knight, S. A. K. Hashmi, Nat. Protocols
2012, 7, 1870–1883.
[31] M. Touil, B. Bechem, A. S. K. Hashmi, B. Engels, M. A. Omary, H. Rabaꢅ, J.
Mol. Struct. 2010, 957, 21–25.
[29] a) A. S. K. Hashmi, A. Loos, S. Doherty, J. G. Knight, K. J. Robson, F. Ro-
minger, Adv. Synth. Catal. 2011, 353, 749–759; b) A. S. K. Hashmi, A.
Loos, A. Littmann, I. Braun, J. Knight, S. Doherty, F. Rominger, Adv. Synth.
Catal. 2009, 351, 576–582.
[32] A. Leyva, A. Corma, J. Org. Chem. 2009, 74, 2067–2074.
[33] D. Canseco-Gonzalez, A. Petronilho, H. Mueller-Bunz, K. Ohmatsu, T. Ooi,
M. Albrecht, J. Am. Chem. Soc. 2013, 135, 13193–13203.
[30] a) A. S. K. Hashmi, L. Schwarz, J.-H. Choi, T. M. Frost, Angew. Chem. 2000,
112, 2382–2385; Angew. Chem. Int. Ed. 2000, 39, 2285–2288; b) A. S. K.
Hashmi, T. M. Frost, J. W. Bats, J. Am. Chem. Soc. 2000, 122, 11553–
11554.
Received: November 1, 2013
Revised: December 31, 2013
Published online on February 12, 2014
Chem. Eur. J. 2014, 20, 3113 – 3119
www.chemeurj.org
3119
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim