Reactivity of electron-deficient alkynes on gold nanoparticles

Article abstract of DOI:10.1021/cs400362c

Propiolates cyclotrimerize in the presence of catalytic amounts of gold nanoparticles to give aryl benzoates in high yields and with turnover frequencies of thousands per hour. Types of alkynes other than propiolates do not react, and, if molecular oxygen

Full text of DOI:10.1021/cs400362c

Research Article
pubs.acs.org/acscatalysis
Reactivity of Electron-Deficient Alkynes on Gold Nanoparticles
Antonio Leyva-Perez,† Judit Oliver-Meseguer,† Jose R. Cabrero-Antonino,† Paula Rubio-Marques,†
́
́
Pedro Serna,† Saud I. Al-Resayes,‡ and Avelino Corma*,†,‡
†Instituto de Tecnología Química, Universidad Politec
los Naranjos s/n, 46022, Valencia, Spain
́
nica de Valencia-Consejo Superior de Investigaciones Científicas, Avenida de
‡Chemistry Department, College of Science, King Saud University, B.O. BOX 2455, Riyadh 11451, Saudi Arabia
S
* Supporting Information
ABSTRACT: Propiolates cyclotrimerize in the presence of
catalytic amounts of gold nanoparticles to give aryl benzoates
in high yields and with turnover frequencies of thousands per
hour. Types of alkynes other than propiolates do not react,
and, if molecular oxygen is present and dissociated by the gold
nanoparticles, electron-rich arenes engage with the propiolate
to form a new C−C bond. The activation of propiolates and
electron-rich arenes to form C−C bonds, beyond gold-
catalyzed Michael additions, constitutes a new example of
how and where gold nanoparticles modify the electronic
density of unsaturated C−C bonds and opens the door to
future transformations.
KEYWORDS: gold catalysis, supported nanoparticles, oxidative alkynylation of arenes, cyclotrimerization, C−H activation
with Au−ZnO (see Supporting Information, Table S1, entries
6−10) under these reaction conditions. A sample of Au−TiO2
was further reduced with phenylethanol under nitrogen
atmosphere and tested in situ for the cyclotrimerization of 1,
and no catalytic differences were found when compared with
the H2-reduced sample (Figure S1 in the Supporting
Information).13 With Au(I) chloride as a catalyst only minor
amounts of product were obtained after prolonged times, and
Au(III) salts were completely inactive for the cyclotrimerization
(entries 11 and 12 in Table S1 in the Supporting Information).
A literature search showed that there is only one precedent for
the gold-catalyzed cyclotrimerization of alkynes.14 There,
trisalkynegold(I) complexes decompose into cyclotrimerization
products after heating in a hexane/dichloromethane (DCM)
mixture, and a catalytic amount of AuSbF6 is also able to
accomplish the cyclotrimerization of the raw alkyne, but in both
cases the yields are <50%.
A solvent screening showed that only alkyl hydrocarbons or
chlorinated compounds are good solvents for the reaction while
coordinating solvents such as dioxane, toluene, or ethyl acetate
inhibit the reaction (Table S2 in the Supporting Information).
These observations connect with the feasibility of the reaction
in hexane/DCM mixtures after decomposition of gold(I)
complexes14 and reflect that gold nanoparticles present a high
sensitivity to external coordinating groups when activating
alkyne 1 for the cyclotrimerization reaction.
INTRODUCTION
■
Gold catalysis has become a trending topic in current research
due to the continuous discovery of new transformations and
unexpected reaction pathways.1−3 Different solid-supported
and homogeneous gold catalysts have been reported, and
although the latter allows a fine-tuning of the electronic/steric
properties of the metal through ligand interactions, the former
is preferred by economical, practical, and environmental
reasons.4−7 The electronic nature of gold nanoparticles as a
function of size, shape, and interaction with the support has
been studied for particular reactions, and some structure−
activity relationships (SARs) have been made8−10 though a
complete switch of reactivity for a single molecule by fine-
tuning the catalytic activity of the nanoparticle is rarely
presented. In this regard, the use of additives on the
nanoparticle to reach a better reactivity control can offer new
opportunities.11 As illustrated in Scheme 1, we will show that
(a) propiolates react selectively on gold nanoparticles to form
aryl rings, (b) electron-rich arenes seem to enhance the
cyclotrimerization, and (c) both molecules finally engage if the
gold nanoparticle is partially oxidized with molecular oxygen.
RESULTS AND DISCUSSION
■
Nature of the gold catalyst. If ethyl propiolate 1 is heated
with a catalytic amount of Au−TiO2 in 1,2-dichlorobenzene
under air, the cyclotrimerization reaction occurs in good yields
and selectivity, as shown in Figure 1.12,13 The reaction rate is
not improved in nitrogen atmosphere, and no reaction occurs
in the presence of other gold-supported catalysts such as Au−
CeO2, Au−Fe2O3, Au−Al2O3, or Au−carbon, and very little
Received: May 15, 2013
Revised: July 3, 2013
Published: July 5, 2013
© 2013 American Chemical Society
1865
dx.doi.org/10.1021/cs400362c | ACS Catal. 2013, 3, 1865−1873

ACS Catalysis
Research Article
Scheme 1. Different Reaction Pathways for Ethyl Propiolate 1 on Supported Gold Nanoparticles
Figure 1. Top: Au−TiO2-catalyzed cyclotrimerization of 1. Reaction conditions: 1 (25.3 μL, 0.25 mmol), Au−TiO2 (1 wt %, 45 mg, 1 mol %), 1,2-
dichlorobenzene (0.5 mL, 0.5 M). Mass balances >95%, no other byproducts found. Middle: Initial rate for the cyclotrimerization of 1 as a function
of the crystal size for samples containing 1.0−1.5 wt % Au on TiO2 (left) and schematic drawing on the mechanism including the calculated equation
rate (right). Bottom: Initial rate for the cyclotrimerization of 1 as a function of the concentration of Au−TiO2 catalyst (left) and alkyne 1 (right).
Dodecane was used as an external standard.
If the higher catalytic activity found for gold on titania
compared to ceria comes from a more pronounced metallic
character of gold on the former support, as assessed by X-ray
photoelectron spectroscopy (XPS, Figure S2 in the Supporting
Information), gold nanoparticles with metallic character should
catalyze the cyclotrimerization reaction independently of the
support employed. This was confirmed after fully reducing a
sample of Au−CeO2 with phenylethanol in order to generate a
more metallic gold (Figure S2 in the Supporting Information),
and the catalytic performance of this reduced Au−CeO2 sample
increased for the cyclotrimerization of 1 with respect to that of
hydrogen-reduced Au−CeO2 (Table S1 in the Supporting
Information, entry 13). On the other hand, gold subnano-
particles (<1 nm) are quasi-molecular species with a tendency
to stabilize positive charges, thus accordingly they should be
inactive for the reaction. When a sample of Au−TiO2 was
treated with methyl iodide,15 the diffuse reflectance ultraviolet−
visible (DR-UV−vis) spectrum (Figure S3 in the Supporting
Information) showed that only small gold clusters remain on
the solid and that the plasmonic gold nanoparticles are absent.
When this sample of <0.75 nm gold particle size was submitted
to reaction conditions, no cyclotrimerization occurred,
suggesting the catalytic activity of metallic gold nanoparticles.
1866
dx.doi.org/10.1021/cs400362c | ACS Catal. 2013, 3, 1865−1873

ACS Catalysis
Research Article
Figure 2. Top: Possible mechanism for the gold-catalyzed cyclotrimerization reaction. Bottom left: Au−CeO2-catalyzed cyclotrimerization of 1 in
□
the presence ( × ) or not ( ) of 50 mol % of 2,6-di-tert-butylcatechol. Bottom right: Initial rate for the Au-supported-catalyzed cyclotrimerization of
1 in the presence or not of 20 mol % of 1,3,5-trimethoxybenzene 2. Reaction conditions: 1 (67.4 μL, 0.66 mmol), catalyst (0.5 mol %), 1,2-
dichlorobenzene (0.66 mL, 1 M), 120 °C. Dodecane was used as an external standard.
With these results in hand, we have to accept that the
catalytic activity for the cyclotrimerization of 1 is related to
electron-rich Au0 crystals, and it could very well occur that
there is an optimum size of the gold nanoparticle for catalyzing
the trimerization reaction. To check this, we prepared and
tested different Au−TiO2 solids with particle sizes ranging from
<1 to ∼10 nm,16,17 and Figure 1 shows that the initial catalytic
activity peaks at gold nanoparticles of ∼3 nm and levels off for
shorter and larger crystallite sizes. These results indicate that
the surface gold atoms on the nanoparticles facilitate the
reaction and that crystallites of 3 nm suppose a good
compromise between atoms exposed and metallic nature.
Reaction Mechanism. The accepted general mechanism
for the cyclotrimerization of alkynes involves the formation of a
five-membered organometallic ring that inserts a third alkyne
unit to finally collapse into the benzene ring and regenerate the
catalyst.18−26 However, in the case of gold nanoparticles, these
crowded organometallic rings are very improbable since gold
prefers mono- or bis-coordinated linear structures. Therefore,
we may think about an alternative mechanism. For doing that,
we decomposed the global reaction process into the elementary
steps of the reaction (Figure S4 in the Supporting
Information). It is clear that, depending on which is the
controlling step, a different kinetic equation could be written.
For instance, if the activation of the first alkyne molecule is the
rate-limiting step (see Figure 1), then the rate equation will
involve the concentration of the alkyne and gold, and a linear
correlation between the initial reaction rate and the initial
concentration of the alkyne and gold should be observed.
Results in Figure 1 show that this is indeed what occurs,
confirming that the first step is the rate-limiting. The calculated
kinetic rate constant indicates that ∼30 000 catalytic cycles per
hour are made in a 1 M solution of gold and alkyne. Dodecane
was used as an external standard during the catalytic
experiments.
during the catalytic cycle and, together with the kinetic
equation, gives a possible mechanism for the cyclotrimerization
reaction in which the rate-determining step would consist of
the σ-activation of 1 on gold, joining a second molecule of
alkyne to finally cyclize with the third unit, as shown in Figure
2. In accordance with this rate equation, none of the
intermediates were detected during reaction, indicating that
the attack to the second alkyne and the final trimerization are
much faster than the rate of activation of the first alkyne
molecule.
Since the metallic gold atoms present an electron unpairing
6s1 configuration that gives a possibility to radical catalysis,18
the radical inhibitor 2,6-di-tert-butylcatechol was introduced in
the reaction medium in order to check if the cyclotrimerization
on gold is a radical reaction, and the results in Figure 2 show
that the reaction was indeed inhibited, which confirms that
radical intermediates are present during the catalytic cycle.27b
It is not surprising that nanosized gold with metallic
character catalyzes the cyclotrimerization of alkynes since
other metals reported as catalysts for this reaction lie in low-
oxidation states.19−26 In many cases, arene-type ligands have
been employed as electron-donors to stabilize the metal
charge.19b,22−24 We added electron-rich arenes in the reaction
medium to test a possible enhancement of the cyclo-
trimerization reaction, and the results in Figure 2 show that
Au supported solids become in general more active for the
cyclotrimerization of 1 in the presence of methoxy-substituted
arenes. It was also found that this catalytic enhancement is
maximized for catalytic amounts of the arene (Figure S5 in the
Supporting Information) and decreases as the arene additive
becomes less electron-rich, and finally a complete inhibition of
the reaction occurs for neutral arenes such as mesitylene.
Carbon−Carbon Bond Forming Reaction. In view that
ethyl propiolate 1 and 1,3,5-trimethoxybenzene 2 are probably
coadsorbed onto gold nanoparticles, we considered the
possibility that both could be engaged in C−C bond forming
reactions. On one hand, He and Shi have reported that
aromatic rings and propiolates give Michael-type additions after
C−H activation of the ring in gold-catalyzed homogeneous
conditions.28 On the other hand, Nevado and de Haro have
To shed more light on the mechanism, deuterated ethyl
propiolate 1-d1 was prepared27a and used as substrate. The
results (also in Figure S4 in the Supporting Information)
showed a kinetic isotopic effect (KIE) of 4.2. The observed KIE
denotes that the terminal C−H bond is somehow broken
1867
dx.doi.org/10.1021/cs400362c | ACS Catal. 2013, 3, 1865−1873

ACS Catalysis
Research Article
Scheme 2. Results for the Reaction between Trimethoxybenzene 2 and Ethyl Propiolate 1 in the Presence of Au−C (5 mol %)
a
and Molecular Oxygen
a
Reaction conditions: 2 (21.0 mg, 0.125 mmol), 1 (38.0 μL, 0.375 mmol), 1,2-dichlorobenzene (0.5 mL). Percentages refer to GC yields. Complete
mass balance for the arene and ∼80−99% for the alkyne. Reaction stops after 1 is transformed into 3 (not shown).
Scheme 3. Isotopic Experiments for the Oxidative Alkynylation between Trimethoxybenzene 2 and Ethyl Propiolate 1 in the
Presence of Gold-Supported Nanoparticles and Molecular Oxygen
Scheme 4. Tentative Mechanism for the Oxidative Alkynylation between Trimethoxybenzene 2 and Ethyl Propiolate 1 in the
Presence of Gold-Supported Nanoparticles and Molecular Oxygen
recently reported that gold phosphine complexes catalyze the
oxidative alkynylation of electron-rich rings in solution through
Au(I)/(III) redox cycles after treatment with strong iodine
oxidants and base.29 Since it has been reported that gold
nanoparticles dissociate molecular oxygen on unsaturated
atoms to form oxygen reactive species (ORS),30 it was
envisaged that the oxidative alkynylation of the arene may
occur on the gold nanoparticle provided that gold could act as a
bifunctional catalyst that activates the organic molecules and
molecular oxygen at the same time. The ORS released after
oxygen dissociation would act as the base needed to quench the
protons generated during the reaction. If this was so, it would
represent, to our knowledge, the first example of a
homogeneous reaction catalyzed by Au(I)/Au(III) redox cycles
that is also catalyzed by a gold heterogeneous system.31 The
result in Scheme 2 shows that the oxidative propiolation of 2 to
give 4 indeed occurs in reasonable yields with Au−carbon as a
catalyst, after optimization (see also Table S3 in the Supporting
Information), and that the Michael-type addition reaction does
not compete. Minor amounts of product 6, coming from either
the oxidative cleavage of the triple bond or a hydration/
retroaldolic sequence in 4, were observed in some cases.
1868
dx.doi.org/10.1021/cs400362c | ACS Catal. 2013, 3, 1865−1873

ACS Catalysis
Research Article
Figure 3. Scope for the cyclotrimerization and oxidative alkynylation of different propiolates 1a−e and trimethoxybenzene 2 in the presence of gold-
supported nanoparticles and oxygen. aGC yields, between brackets isolated yields. bRatio isomers 4:1; with 0.015 mol % Au, 77% yield; TON, 5000.
d
e
f
g
c3 equiv of propiolate. 2 mol % Au. 5 equiv of 1a. Ratio of isomers 1:1. Reaction temperature 120 °C.
Experiments with AuCl and HAuCl4 (Table S3 in the
Supporting Information, entries 1−2) showed that, as
expected,28 the Lewis-catalyzed Michael addition to form 5
occurs in moderate yields with neat cationic gold, and cationic
gold stabilized on CeO2 and Fe2O3 also gave the addition
product 5 preferentially to 4 (entries 5−7). These results
confirm that the unsaturated gold atoms on the nanoparticles,
having a cationic character, catalyze the Michael-type addition,
as salts do. To check this, a new experiment was performed in
the absence of oxygen (entry 12) and no product 4 was found
but only 5 and 3, which confirms that the oxidative alkynylation
only occurs under the action of oxygen and that, otherwise, the
Michael-type addition proceeds on the cationic gold present. In
order to further check that the oxidative coupling product 4
does not come from the Michael product, the cinnamylic ester
5 was prepared independently28 and introduced under oxidative
conditions. No reaction was observed in that case. With these
results in hand, in principle, cationic supported gold remains
aside of the oxidative process. Oxygen dissociation on small
gold nanoparticles has been proved elsewhere,31 and electro-
voltammetry of the Au−carbon solid confirmed the feasibility
of the redox process.32,33
Deuterated ethyl propiolate 1-d1 was prepared and Scheme 3
shows that the oxidative alkynylation is inhibited when this was
used as a reactant, only Michael addition occurring. Notice that
1-d1 also retards the cyclotrimerization reaction (KIE
observed), and these results suggest that the deprotonation of
the propiolate occurs before the oxidative coupling. If this is so,
1869
dx.doi.org/10.1021/cs400362c | ACS Catal. 2013, 3, 1865−1873

ACS Catalysis
Research Article
Scheme 5. Failed Cyclotrimerization Reaction of Triyne 7 (Top) and Selective Intermolecular Cyclotrimerization of 8 and 1e
a
(bottom) under Au−TiO2 Catalyzed Conditions
a
GC yields.
◇
□
Figure 4. Top left: Reusability of Au−TiO2 in the cyclotrimerization of 1, in terms of final yield ( ) and initial rate ( ). Top right: Hot filtration
□
test for the cyclotrimerization of 1 by filtering at 20% ( ) and 50% (Δ) conversion of 1. Bottom left: Percentage of contribution in the reaction rate
for the cyclotrimerization of 1 of the leached species after filtration of Au−CeO2. Au−TiO2 gave similar results. Bottom right: Hot filtration test by
□
filtering at 10% ( ) for the oxidative alkynylation of 1 and 2. Reaction conditions for the cyclotrimerization: 1 (67.4 μL, 0.66 mmol), catalyst (1 mol
% left, 0.5 mol % right), n-decane (0.66 mL, 1 M), 120 °C. Reaction conditions for the oxidative alkynylation: 1 (38.0 μL, 0.375 mmol), 2 (21 mg,
0.125 mmol), 1,2-dichlorobenzene (0.5 mL), catalyst (5 mol %), 120 °C, molecular oxygen (6 bar, ∼1 mmol, ∼8 equiv). Dodecane was used as an
external standard.
the cyclotrimerization should compete with the alkynylation,
and this is indeed observed (see Table S3 in the Supporting
Information). The results would explain the moderate
conversion of arene 2 since propiolate consumption occurs
even in high excess.
released by the alkyne and the arene after gold activation, and a
final reductive elimination process engages both molecules into
a new C−C bond and regenerates the metallic gold sites. This
reaction pathway is in accordance with one proposed by
Nevado and de Haro in solution.29
Trimethoxybenzene 2-d3 was also prepared and tested in
reaction, and Scheme 3 shows that hydrogen exchange was also
observed, even in the absence of the gold catalyst. Deuterium
scrambling between 2-d3 and 1 confirms somehow the
interaction of the arene with gold. Nevertheless, none of
these two processes are the rate-limiting step of the reaction
since the H-exchange rate is much faster than the coupling rate
(seven times faster for the arene), so it must be assumed that
the final coupling constitutes the rate-limiting step.
A tentative mechanism for the alkynylation of trimethox-
ybenzene 2 on Au−carbon is depicted in Scheme 4. The
different steps proposed are supported by kinetic, isotopic, and
spectroscopic experiments. Overall, small gold nanoparticles
dissociate molecular oxygen, probably only in the presence of
the reactants, to give basic ROS that neutralize the protons
Scope of the Reactions. Now it seems possible to control
the reactivity of propiolates toward the cyclotrimerization
reaction or the oxidative alkynylation of arenes by using the
appropriate gold-supported nanoparticles under oxygen or not.
A brief scope of the reactions was evaluated after preparing
different propiolates,34 and the results in Figure 3 reveal that
the propiolates engage intermolecularly with different diynes
for the cyclotrimerization reaction and also with the arene 2 for
the oxidative coupling in good yields. Remarkably, a turnover
number (TON) of 5000 can be obtained for the cyclo-
trimerization of 1 which constitutes, as far as we know, the
highest recorded for an intermolecular cyclotrimerization of
alkynes.
Different functionalities including ether, chloride, and nitro
groups are tolerated under the reaction conditions. Dimethy-
1870
dx.doi.org/10.1021/cs400362c | ACS Catal. 2013, 3, 1865−1873

ACS Catalysis
Research Article
lacetylene dicarboxylate (DMAD) 1e reacts with other alkynes
(3h, 3j−k) and with itself (3l) when forcing the reaction
conditions. The reactivity of internal alkynes strongly supports
a radical mechanism. In any case, the cyclotrimerization does
not proceed unless one of the partners is a propiolate. For
instance, a benchmark substrate for the cyclotrimerization
reaction12 such as the entropic and sterically favored triyne 7
was prepared35a and tested under Au−TiO2 catalyzed
conditions, and the result in Scheme 5 shows that no reaction
occurs. However, if the nonreactive internal alkyne is converted
in a propiolate such as triyne 8,35b it does react with DMAD 1e
to give the cyclotrimerization product 9 having a free,
unreacted, terminal alkyne. Other common cyclotrimerization
partners such as nitriles12 do not react either. These results
illustrate the high specificity of Au−TiO2 toward propiolates.
Catalyst Stability and Reusability. At this point, an
extensive study on the reusability and stability of the catalyst
under reaction conditions was carried out. As shown in Figure
4, Au−TiO2 could be reused up to six times for the
cyclotrimerization of 1 in n-decane after filtering the solid
just at complete conversion (22 min), but a slight decrease in
both the initial rate and final conversion was observed.
Hot filtration tests revealed that no active species are present
in solution, but inductively coupled plasma atomic emission
spectroscopy analysis of the filtrates and X-ray fluorescence
measurements of the reused solid showed that ∼5% of the total
gold in the solid was leached to the solution during reaction,
which explains the gradual loss of activity. After this result, the
possible leaching and activity of gold species in 1,2-
dichlorobenzene was scrutinized by filtering the solid at
different times and comparing the initial rate before and after
filtration. The results in Figure 4 show a progressive
contribution of leached species to the catalytic activity, but
this contribution is only significant at conversions of 1 >20%,
even at long reaction times, which denotes that the initial rate
measurements reflect only catalytically active supported gold
species. The filtration hot test was also carried out for the
oxidative alkynylation of 1 and 2 and revealed that no active
species are present in solution with 1,2-dichlorobenzene as a
solvent, which confirms that the catalytic process occurs onto
the solid. The presence of inactive gold in solution during the
reaction fits well with the lack of catalytic activity of
uncoordinated atoms on the solid since the latter are easily
leached into the solution. A similar effect has been observed for
gold-catalyzed Sonogashira reactions.4a,b
showed that no chlorides remain on the solid after treatment
under hydrogen at high temperature. Alternatively, neat
phenylethanol (5 mL per gram of solid) was used as reducing
agent in a preheated oil bath at 160 °C for 1 h, washing with
acetonitrile and diethyl ether and drying under vacuum. The
commercial samples used were supplied by the following:
Strem, Aurelite (Au−TiO2 1 wt %, Au−ZnO, 1 wt %, Au−
Al2O3 1 wt %); and Johnson-Matthey, Au−CeO2 (1 wt %) and
Au−carbon (2.3 wt %).
Preparation of Propiolates. Ethyl (1a, referred to as 1
throughout the text for simplification), tert-butyl (1d), and
dimethyl (1e) were supplied by Aldrich. The chloro- and nitro-
substituted ones were prepared according to a modified
standard procedure34 as follows. Synthesis of functionalized
terminal propiolates 1b and 1c: A mixture of DCC (2.32 g) and
DMAP (0.090 g) in DCM/Et2O (5/25 mL) was added
dropwise over 20 min to a solution of the corresponding
alcohol (10.0 mmol) and propiolic acid (0.678 mL, 11.0 mmol)
in Et2O (5 mL) at −30 °C. The reaction mixture was stirred
overnight at room temperature and quenched with saturated
NH4Cl (15 mL), and the aqueous layer was extracted with
Et2O (3 × 15 mL). The combined organic extracts were dried
over magnesium sulfate, concentrated under reduced pressure,
and purified by column chromatography to give the
corresponding terminal propiolate.
The triynes 7 and 8 were prepared according to reported
procedures.35
Typical Procedure for the Cyclotrimerization of 1. The
solid catalyst (1 mol % of metal respect to 1) was placed in a 2
mL vial equipped with a magnetic stirrer. 1,2-Dichlorobenzene,
n-decane, or n-octane (0.66 mL) and ethyl propiolate 1 (67.6
μL, 0.66 mmol) were added, and the vial was sealed. The
resulting mixture was placed in a preheated oil bath at 120 °C
and magnetically stirred for the corresponding time. Aliquots
were periodically taken, poured into diethyl ether or CH2Cl2
(1.0 mL), filtered through a 0.2 μm PTFE filter syringe if
needed, and submitted to GC and GC−MS analysis after
dodecane (5.6 μL, 0.05 mmol) was added as external standard.
In some cases dodecane (14.6 μL, 0.066 mmol) was added to
the reaction as internal standard. At the end of the reaction,
diethyl ether or CH2Cl2 (1 mL) was added after cooling the vial
and the mixture was stirred for a few minutes at room
temperature, solids were filtered off, and the resulting filtrates
were analyzed as the aliquots. For reusing, n-decane (1 mL)
was used as a solvent and the solid was washed twice with
decane (1 mL) before the next run. The resulting solid after six
uses was washed with CH2Cl2 (3 mL, 3 times) and dried under
vacuum before analyses. Isolation of the products was
performed by column chromatography or preparative TLC
(15% AcOEt/hexane) to give 3 as a mixture of triethyl
benzene-1,2,4-tricarboxylate (3a, major) and triethyl benzene-
1,3,5-tricarboxylate (3b, minor) in a 4:1 ratio. The reaction
could be scaled up to 5 mmol without loss of efficiency.
Triethyl benzene-1,2,4-tricarboxylate 3a: colorless oil (40 mg,
60%). Rf (20% AcOEt/hexane): 0.42. GC−MS (m/z, M+•
294), major peaks found: 294 (11%), 249 (100%), 221 (100%),
193 (90%). IR (cm−1): 2982, 1727, 1307, 1283, 1245, 1110. 1H
NMR (δ, ppm; J, Hz): 8.31 (1H, dd, J = 1.6, 0.5), 8.10 (1H, dd,
J = 8.1, 1.6), 7.67 (1H, dd, J = 8.0, 0.5), 4.32 (6H, q, J = 7.1),
1.32 (9H, t, J = 7.1). 13C NMR (δ, ppm): 167.0 (C), 166.4 (C),
164.8 (C), 136.1 (C), 132.5 (C), 132.0 (C), 131.9 (CH), 129.9
(CH), 128.7 (CH), 61.8 (CH2), 61.7 (CH2), 61.5 (CH2), 14.1
(CH3), 14.0 (CH3), 13.9 (CH3). HRMS (ESI) [(M + H)+;
CONCLUSIONS
■
Propiolates give different reactions on supported gold nano-
particles depending on the nature of the particle and the
amount of oxygen present: cyclotrimerization products on
metallic nanoparticles, Michael additions with 1,3,5-trimethox-
ybenzene on cationic gold species, and oxidative alkynylation of
1,3,5-trimethoxybenzene for terminal propiolates if enough
molecular oxygen dissociates on the gold nanoparticle.
EXPERIMENTAL SECTION
■
16,17
Preparation of the Solid Catalysts. Au−TiO2
and
4a,6
Au−CeO2
solids were prepared according to a reported
procedure. In some cases gold was incorporated onto the solids
by the incipient wetness methodology with a solution of AuCl
in acetonitrile, drying in an oven at 100 °C after impregnation
and reducing under H2 at the specified temperature. XPS
1871
dx.doi.org/10.1021/cs400362c | ACS Catal. 2013, 3, 1865−1873

ACS Catalysis
Research Article
(4) (a) Corma, A.; Juar
́
ez, R.; Boronat, M.; San
́
chez, F.; Iglesias, M.;
calculated for C11H15O3: 195.1021] found m/z 195.1014.
Triethyl benzene-1,3,5-tricarboxylate 3b: white solid (10 mg,
15%). Rf (20% AcOEt/hexane): 0.52. GC−MS (m/z, M+ 294),
major peaks found: 294 (27%), 266 (36%), 249 (100%), 238,
(24%), 221 (66%), 210 (26%), 193 (31%). IR (cm−1): 2996,
1726, 1718, 1239. 1H NMR (δ, ppm; J, Hz): 8.77 (3H, s), 4.37
(6H, q, J = 7.1), 1.36 (9H, t, J = 7.1). 13C NMR (δ, ppm):
165.0 (C × 3), 134.4 (CH × 3), 131.4 (C × 3), 61.7 (CH2 ×
3), 14.3 (CH3 × 3). HRMS (ESI) [(M + H)+; calculated for
C11H15O3: 195.1021] found m/z 195.1014.
Typical Procedure for the Oxidative Alkynylation. The
solid catalyst (5 mol % of metal respect to 2) and
trimethoxybenzene 2 (21 mg, 0.125 mmol) were placed in a
double-walled glass 2 mL reactor equipped with a magnetic
stirrer and a manometer. 1,2-Dichlorobenzene (0.5 mL) and
ethyl propiolate 1 (38.0 μL, 0.375 mmol) were added, and the
reactor was closed. Molecular oxygen (6 bar, ∼ 0.9 mmol) was
introduced at room temperature, and the resulting mixture was
magnetically stirred in a preheated oil bath at 120 °C for the
required time. Aliquots were periodically taken, poured into
CH2Cl2 (1 mL), filtered through a 0.2 μm PTFE filter syringe,
and submitted to GC and GC−MS analysis after dodecane (5.6
μL, 0.025 mmol) was added as an external standard. At the end
of the reaction, CH2Cl2 (1 mL) was added after cooling the
reactor and the mixture was stirred for a few minutes at room
temperature, solids were filtered off, and the resulting filtrates
were analyzed as the aliquots.
García, H. Chem. Commun. 2011, 47, 1446. (b) Boronat, M.; Combita,
D.; Concepcion, P.; Corma, A.; García, H.; Juaarez, R.; Laursen, S.;
Lopez-Castro, J. D. J. Phys. Chem. C 2012, 116, 24855. (c) Daniel, M.-
C.; Astruc, D. Chem. Rev. 2004, 104, 293.
(5) Grirrane, A.; Corma, A.; García, H. Science 2008, 322, 1661.
(6) Gonzalez-Arellano, C.; Abad, A.; Corma, A.; García, H.; Iglesias,
M.; Sanchez, F. Angew. Chem., Int. Ed. 2007, 46, 1536.
́
́
́
́
́
(7) Navalon, S.; Martin, R.; Alvaro, M.; Garcia, M. Angew. Chem., Int.
Ed. 2010, 49, 8403.
(8) (a) Kanuru, V. K.; Kyriakou, G.; Beaumont, S. K.; Papageorgiou,
A. C.; Watson, D. J.; Lambert, R. M. J. Am. Chem. Soc. 2010, 132,
8081. (b) Beaumont, S. K.; Kyriakou, G.; Lambert, R. M. J. Am. Chem.
Soc. 2010, 132, 12246. (c) Kyriakou, G.; Beaumont, S. K.; Humphrey,
S. M.; Antonetti, C.; Lambert, R. M. ChemCatChem 2010, 2, 1444.
(9) Herzing, A. A.; Kiely, C. J.; Carley, A. F.; Landon, P.; Hutchings,
G. J. Science 2008, 321, 1331.
(10) Hughes, M. D.; Xu, Y. J.; Jenkins, P.; McMorn, P.; Landon, P.;
Enache, D. I.; Carley, A. F.; Attard, G. A.; Hutchings, G. J.; King, F.;
Stitt, E. H.; Johnston, P.; Griffin, K.; Kiely, C. J. Nature 2005, 437,
1132.
(11) Toma, H. E.; Zamarion, V. M.; Toma, S. H.; Araki, K. J. Braz.
Chem. Soc. 2010, 21, 1158.
(12) (a) For reviews on the cyclotrimerization reaction see Shaaban,
M. R.; El-Sayed, R.; Elwahy, A. H. M. Tetrahedron 2011, 67, 6095.
(b) Kotha, S.; Brahmachary, E.; Lahiri, K. Eur. J. Org. Chem. 2005,
4741.
(13) (a) The Stratakis group has uncovered that Au−TiO2 is an
excellent catalyst for the disilylation of alkynes and isomerization of
epoxides; see: Lykakis, I. N.; Psyllaki, A.; Stratakis, M. J. Am. Chem.
Soc. 2011, 133, 10426. (b) Raptis, C.; Garcia, H.; Stratakis, M. Angew.
Chem., Int. Ed. 2009, 48, 3133.
(14) Das, A.; Dash, C.; Yousufuddin, M.; Celik, M. A.; Frenking, G.;
Dias, H. V. R. Angew. Chem., Int. Ed. 2012, 51, 3940.
(15) Sa, J.; Goguet, A.; Taylor, S. F. R.; Tiruvalam, R.; Kiely, C. J.;
́
Nachtegaal, M.; Hutchings, G. J.; Hardacre, C. Angew. Chem., Int. Ed.
2011, 50, 8912.
Hot Filtration Tests. Two parallel reaction mixtures were
followed by GC, taking aliquots periodically, under typical
reaction conditions, and one of them was filtered hot at a
determined conversion through a 0.2 μm PTFE filter. The
resulting filtrates were placed under the same conditions
(stirring and temperature) as the original reaction and also
followed by GC.
(16) (a) Corma, A.; Serna, P.; Concepcion, P.; Calvino, J. J. J. Am.
Chem. Soc. 2008, 130, 8748. (b) Boronat, M.; Concepcion
́
, P.; Corma,
ASSOCIATED CONTENT
* Supporting Information
General procedures, syntheses and characterization of catalysts,
substrates, and products, reaction procedures, additional tables
and figures, and NMR spectra of compounds. This material is
available free of charge via the Internet at http://pubs.acs.org.
■
A.; Gonzalez, S.; Illas, F.; Serna, P. J. Am. Chem. Soc. 2007, 129, 16230.
́
S
(17) (a) Chirea, M.; Freitas, A.; Vasile, B. S.; Ghitulica, C.; Pereira, C.
M.; Silva, F. Langmuir 2011, 27, 3906. (b) Oliver-Meseguer, J.;
Cabrero-Antonino, J. R.; Domínguez, I.; Leyva-Per
Science 2012, 338, 1452.
́
ez, A.; Corma, A.
(18) Martin, R.; Menchon, C.; Apostolova, N.; Victor, V. M.; Alvaro,
M.; Herance, J. R.; Garcia, H. ACS Nano 2010, 4, 6957.
(19) (a) Reported metal catalysts for the cyclotrimerization reaction
include cobalt, rhodium, ruthenium, titanium, molybdenum, and
palladium, among others. See this and the following reference for
selected examples. These metals can perform the organometallic ring
pathway. For cobalt see: Agenet, N.; Gandon, V.; Vollhardt, K. P. C.;
Malacria, M.; Aubert, C. J. Am. Chem. Soc. 2007, 129, 8860. (b) Rhyoo,
H.-Y.; Lee, B. Y.; Yu, H. K. B.; Chung, Y. K. J. Mol. Catal. 1994, 92, 41.
(20) For rhodium see: Tanaka, K.; Toyoda, K.; Wada, A.; Shirasaka,
K.; Hirano, M. Chem.Eur. J. 2005, 11, 1145.
AUTHOR INFORMATION
Corresponding Author
*E-mail: acorma@itq.upv.es. Phone: +34963877800. Fax:
+349638 77809.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
(21) For ruthenium see: Cadierno, V.; García-Garrido, S. E.; Gimeno,
J. J. Am. Chem. Soc. 2006, 128, 15094 and ref 22.
A.L.-P. thanks CSIC for a contract. J.O.-M. thanks ITQ for a
postgraduate scholarship. J.R.C.-A. and P.R.-M. thank MECD
for the concession of a FPU contract. Financial support by the
Severo Ochoa program and Consolider-Ingenio 2010 (proyecto
MULTICAT) from MICIINN is acknowledged, and also the
King Saud University. P.S. thanks European Union Seventh
Framework programme (PIOF−GA−2009−253129).
(22) Yamamoto, Y.; Arakawa, T.; Ogawa, R.; Itoh, K. J. Am. Chem.
Soc. 2003, 125, 12143.
(23) For titanium see: Ozerov, O. V.; Patrick, B. O.; Ladipo, F. T. J.
Am. Chem. Soc. 2000, 122, 6423.
(24) For molybdenum see: Ardizzoia, G. A.; Brenna, S.; LaMonica,
G.; Maspero, A.; Masciocchi, N. J. Organomet. Chem. 2002, 649, 173.
(25) For palladium see: Lin, Y.-Y.; Tsai, S.-C.; Yu, S. J. J. Org. Chem.
2008, 73, 4920 and ref 26.
(26) In the context of heterogeneous catalysis, palladium has shown
good catalytic activity; see for instance: Ormerod, R. M.; Lambert, R.
M. J. Chem. Soc., Chem. Commun. 1990, 1421. We tested commercial
Pd−carbon for the cyclotrimerization of 1 under our standard
REFERENCES
■
(1) Corma, A.; Leyva-Per
1657.
́
ez, A.; Sabater, M. J. Chem. Rev. 2011, 111,
(2) Leyva-Per
́
ez, A.; Corma, A. Angew. Chem., Int. Ed. 2012, 51, 64.
(3) Hashmi, A. S. K. Angew. Chem., Int. Ed. 2010, 49, 5232.
1872
dx.doi.org/10.1021/cs400362c | ACS Catal. 2013, 3, 1865−1873

ACS Catalysis
Research Article
conditions, and an equimolecular product distribution 3a:3b was
found after complete conversion of 1. This catalyst was also tested for
the oxidative alkynylation of 1 and 2, and no product was found.
(27) (a) Schwier, T.; Gevorgyan, V. Org. Lett. 2005, 7, 5191.
(b) García-Mota, M.; Cabello, N.; Maseras, F.; Echavarren, A. M.;
Perez-Ramírez, J.; Lopez, N. ChemPhysChem 2008, 9, 1624.
́
(28) (a) Shi, Z.; He, C. J. Org. Chem. 2004, 69, 3669. (b) Reetz, M.
T.; Sommer, K. Eur. J. Org. Chem. 2003, 3485.
(29) (a) De Haro, T.; Nevado, C. J. Am. Chem. Soc. 2010, 132, 1512.
(b) For a recent review by the same authors on C-H activation by
gold see: De Haro, T.; Nevado, C. Synthesis 2011, 16, 2530.
(30) (a) For a recent review see: Boronat, M.; Corma, A. J. Catal.
2011, 284, 138. (b) Rodríguez, J. A.; Feria, L.; Jirsak, T.; Takahashi, Y.;
Nakamura, K.; Illas, F. J. Am. Chem. Soc. 2010, 132, 3177. (c) Boronat,
M.; Corma, A. Dalton Trans. 2010, 39, 8538.
(31) (a) Alves, L.; Ballesteros, B.; Boronat, M.; Cabrero-Antonino, J.
́
R.; Concepcion, P.; Corma, A.; Correa-Duarte, M. A.; Mendoza, E. J.
Am. Chem. Soc. 2011, 133, 10251. (b) Engle, K. M.; Mei, T.-S.; Wang,
X.; Yu, J.-Q. Angew. Chem., Int. Ed. 2011, 50, 1478.
(32) Private communication from Prof. Domen
related article see: Domenech, A.; Leyva-Perez, A.; Al-Resayes, S. I.;
Corma, A. Electrochem. Commun. 2012, 19, 145.
(33) Leyva-Perez, A.; Domenech, A.; Al-Resayes, S. I.; Corma, A.
ACS Catal. 2012, 2, 121.
́
ech. For a recent
́
́
́
́
(34) Hodgson, D. M.; Talbot, E. P. A.; Clark, B. P. Org. Lett. 2011,
13, 5751.
(35) (a) Weding, N.; Jackstell, R.; Jiao, H.; Spannenberg, A.; Hapke,
M. Adv. Synth. Catal. 2011, 353, 3423. (b) Zhang, W.; Sprafke, J. K.;
Ma, M.; Tsui, E. Y.; Sydlik, S. A.; Rutledge, G. C.; Swager, T. M. J. Am.
Chem. Soc. 2009, 131, 8446.
1873
dx.doi.org/10.1021/cs400362c | ACS Catal. 2013, 3, 1865−1873