A residue-free production of biaryls using supported gold nanoparticles

Article abstract of DOI:10.1016/j.jcat.2014.04.004

Small gold nanoparticles in the presence of O₂ are able to cleave CH bonds of non-activated arenes and facilitate catalytic formation of biaryls under mild conditions. This procedure avoids the common requirement of iodine, acids, bases, and/or other stoichiometrical additives to accomplish multiple turnovers and results in a zero-waste synthetic process. A number of unactivated arenes such as benzene, toluene, p-xylene, nitrobenzene, chlorobenzene, or phenol are selectively converted to the corresponding biaryls, with accumulated turnover numbers greater than 600. Kinetic measurements, in combination with IR, HRTEM, and STEM-HAADF data, suggest that the reaction takes place on gold atoms that are metallic in character, with the highest CC bond formation rate provided by gold nanoparticles of approximately 3 nm of diameter. A radical-free reaction mechanism is postulated on the basis of the kinetic measurements. The results represent a first stone for the waste-free preparation of biaryls using supported gold catalysts.

Full text of DOI:10.1016/j.jcat.2014.04.004

Journal of Catalysis 315 (2014) 41–47
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
A residue-free production of biaryls using supported gold nanoparticles
⇑
⇑
Pedro Serna , Avelino Corma
Instituto de Tecnología Química, Universidad Politécnica de Valencia–Consejo Superior de Investigaciones Científicas, Avda. de los Naranjos s/n, 46022 Valencia, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
Small gold nanoparticles in the presence of O2 are able to cleave CAH bonds of non-activated arenes and
facilitate catalytic formation of biaryls under mild conditions. This procedure avoids the common
requirement of iodine, acids, bases, and/or other stoichiometrical additives to accomplish multiple turn-
overs and results in a zero-waste synthetic process. A number of unactivated arenes such as benzene, tol-
uene, p-xylene, nitrobenzene, chlorobenzene, or phenol are selectively converted to the corresponding
biaryls, with accumulated turnover numbers greater than 600. Kinetic measurements, in combination
with IR, HRTEM, and STEM-HAADF data, suggest that the reaction takes place on gold atoms that are
metallic in character, with the highest CAC bond formation rate provided by gold nanoparticles of
approximately 3 nm of diameter. A radical-free reaction mechanism is postulated on the basis of the
kinetic measurements. The results represent a first stone for the waste-free preparation of biaryls using
supported gold catalysts.
Received 24 February 2014
Revised 2 April 2014
Accepted 3 April 2014
Keywords:
Gold catalysis
Supported gold nanoparticles
CAC bond formation
Waste-free formation of biaryls
Activation of non-activated arenes
Ó 2014 Elsevier Inc. All rights reserved.
1. Introduction
aromatic counterparts is functionalized [9,10], or upon substitu-
tion of iodine by lighter chlorine [11] or boronic acid group [12].
Biaryls are ubiquitous building blocks with application in agro-
chemistry, pharmacy, electronics, conductors, polymers, and liquid
crystals, among others [1–4]. The industrial demand of biaryls is
prominent, but the process to accomplish the formation of the
CAC bond is, in general, costly. The high costs are in part derived
from the difficulty of cleaving the ArAH bond directly, which
results in additional operations to pre-activate the arene prior to
the coupling step. These processes typically involve aryl halides
that are not ecofriendly to prepare (Friedel–Craft processes, [5]
using metal-halide salts, and the Sandmayer reaction, [6] starting
from aniline, hydrochloric acid and sodium nitrite are some exam-
ples). The aryl halide is then coupled with a second activated arene
in the presence of a metal catalyst (e.g., Suzuki [7] or Stille [8]
couplings). Stoichiometric amounts of additives such as bases in
solution [7] or metals such as Sn [8] are usually needed to facilitate
the removal of the halide from the metal and enable multiple
turnovers. The base/halide or metal/halide residues cause further
increment of the costs in operations for purification of the down-
stream effluents. The atom efficiency of the coupling step is, in this
scenario, often <30%, corresponding to approximately 2.3 kg of
waste per kilogram of desired product. Aware of these problems,
chemists have sought methods that limit the generation of
residues, for example through reactions where only one of the
A major breakthrough would be brought by catalysts able to
couple arenes that have not been activated in any manner before
reaction and that operate without the necessity of other additives
in the solution. In this sense, ‘‘unactivated arenes’’ react and yield
biaryls in the presence of some metal catalysts and oxygen donors,
although the number of examples is, up-to-date, scarce. For
example, gold salts in acidic solutions used in tandem with iodine
oxidants such as PhI(OAc)2 or PhI(OH)OTs activate the ArAH
moiety and catalyze some coupling reactions [13,14]—but the
use of iodine, and acids in solution is not yet avoided; Pd2+ facili-
tates the coupling of unactivates arenes with O2 as the oxidant
and water as the residue—but then the biaryl formation occurs
with reduction of Pd2+ to Pd0, which is an inactive species [15].
To complete multiple turnovers, Pd must be re-oxidized before
precipitation as palladium particles, usually in the presence of
acids and/or a second metal that allow Pd to shuttle in between
oxidation states [16–20]. Even then, the number of turnovers per
active site is low, typically <100 [16,17], and often <10 [20,21],
while significant amounts of hazardous wastes are generated.
To our knowledge, there is no precedent of a catalyst that
affords multiple turnovers in the coupling of unactivated arenes
without the generation of residues such as acids, bases, iodine
derivatives, metals salts, or other stoichiometrical additives. In
other words, we have not seen a catalyst that cleaves directly the
ArAH bond and subsequently forms a new CAC bond with O2 as
the only co-reagent and water as the only by-product.
⁄ Corresponding authors. Fax: +34 96 387 7800.
E-mail addresses: psername@itq.upv.es (P. Serna), acorma@itq.upv.es
(A. Corma).
http://dx.doi.org/10.1016/j.jcat.2014.04.004
0021-9517/Ó 2014 Elsevier Inc. All rights reserved.

42
P. Serna, A. Corma / Journal of Catalysis 315 (2014) 41–47
In the present investigation, we report such
a
catalyst,
STREM (AUROliteTM); d, the preceding catalyst after being used
in the coupling of benzene at 413 K and 12 bar for 100 h; e, sample
c after deposition of 0.5 wt% Pd by incipient wetness technique; f,
3 wt% Au/TiO2 sample prepared by deposition–precipitation of
HAuCl4 at pH = 9 and calcined under static atmosphere of air at
673 K. (B) Particle size distributions, as determined from the corre-
sponding HAADF-STEM and/or HRTEM images, of the preceding
catalysts and the following: a, sample c treated in 50 mL/min of
air at 1 bar and 673 K; b, sample d treated in 50 mL/min of air at
1 bar and 673 K; g, 3 wt% Au/TiO2 sample prepared by deposi-
tion–precipitation of HAuCl4 at pH = 6 and calcined under static
atmosphere of air at 673 K. (C) Dependence of the TOF per external
gold atom on the average particle size during the solvent-free cou-
pling of benzene at 12 bar of O2 and 413 K for the series of catalysts
in (B). The selectivity to biphenyl was always >98%. Experimental
details are provided in SI.
prompted by the realization that gold is active for a number of
CAC bond formation reactions not only when is present in the form
of a soluble metal salt, but also when arranged as small nanoparti-
cles [22–26], and that the latter work notably better in processes
that involve O2 dissociation [27–32]. Our results demonstrate
another remarkable feature of gold in heterogeneous catalysis
[33–37].
2. Experimental methods
2.1. Catalysts preparation
1 wt% Au/TiO2, 1 wt% Au/Al2O3, and 1 wt% Au/ZnO catalysts
were used as received from STREM (AUROliteTM catalysts). The
catalysts can be prepared following a deposition–precipitation
method from HAuCl4, as described elsewhere [32].
Pt/Al2O3, Rh/TiO2, Ni/TiO2, and Pd/TiO2 samples were prepared
by incipient wetness technique. H2PtCl6 (hexahydrate, Aldrich,
>37.5 as Pt), Ni(NO3)2 (hexahydrate, Fluka, >98.5%), RhCl3
(Aldrich, Rh content 40%) and PdCl2 (Aldrich, 99%) were used to
impregnate TiO2 (Degussa P-25) in water. As an example, 20 mL
of an aqueous solution containing 13.27 mg of H2PtCl6Á6H2O
was contacted with 1 g of TiO2 to prepare the 0.5 wt% Pt/TiO2 cat-
alyst. After a perfect mixing of the corresponding slurries, sam-
ples were dried at 373 K during 12 h. Some samples were
reduced in flow of H2 at 723 K for 3 h before reaction, as specified
in Table 1.
To prepare Au/TiO2 catalysts with variable particle sizes, HAuCl4
was deposited on TiO2 at a controlled pH. This method allows con-
trolling the degree of metal aggregation by choice of the metal
loading, the pH of the deposition, and the activation conditions,
as reported [38]. Table S1 (below) provides details of the synthesis
of Au/TiO2 catalysts characterized by average particle sizes of 7.2
and 14.4 nm, respectively.
2.2. Oxidative coupling of arenes
Catalytic testing was performed in a reinforced glass reactor
(2 mL volume) equipped with a temperature and pressure control
and stirred magnetically. Catalytic testing was performed in a rein-
forced glass reactor (2 mL volume) equipped with a temperature
and pressure control and stirred magnetically. The reactor vessel
was bought to Supelco (Reference 2-7037) and modified to allow
pressurization and/or extraction of liquids through a gas-tight nee-
dle. A 100
lL gas-tight syringe was used to get 15–20 lL aliquots at
the various reaction times.
Before each experiment, all the material was washed with
abundant acetone and dried at 383 K for >5 h. It is important to
avoid acetone and other polar molecules in the reaction mixture
to get optimal results. In a typical experiment, 891 mg of the
aromatic compound was placed in the reactor together with
30–70 mg of catalyst and 9 mg of dodecane as an internal standard.
Reactants were obtained from Sigma–Aldrich with purities above
99% and used as received. The reactor atmosphere was purged with
O2 at room temperature, pressurized with 12 bar of O2 and placed
into a silicon oil bath pre-heated at the desired reaction tempera-
ture. We systematically assigned time = 0 (start of the kinetic
experiment) 30 s after the reactor had been immersed into the
silicon oil bath. During the experiment, the pressure was kept at
12 bar, and the stirring rate was fixed at 700 r.p.m. Aliquots were
taken at different times until the end of the experiment. The com-
position of these aliquots was determined with a gas chromato-
graph equipped with a FID detector and a 30 m HP-5 capillary
column. Conversions and selectivities were calculated from the
GC areas of the products corrected with the response factors
determined experimentally. The products were identified by mass
spectrometry using a GC/MS device (Agilent MDS-5973) equipped
with a quadrupole electron-impact ionization detector (spectra
provided in SI). Pure, commercially available compounds were
used to compare the biaryls mass spectra and their GC retention
times. We also performed HPLC analyses to evaluate the potential
presence of heavier compounds. A Varian ProStar 240 device
The bimetallic Au@Pd/TiO2 catalyst was synthesized by impreg-
nation of the Au/TiO2 catalyst with a solution of PdCl2 to have a
final Au/Pd mol ratio of approximately 1. This catalyst was thor-
oughly washed with deionized water and dried at 373 K for 12 h
before reaction.
The following information guides the identification of each of
the samples in Fig. 2: (A) c, 1 wt% Au/TiO2 sample supplied by
Table 1
Catalytic performance of several metal-based samples for the O2-assisted coupling of
benzene at 413 K and 12 bar.
Entry Catalyst
% Med (mol) TOFe (hÀ1
)
TONf (mol/mol)
a
1
2
3
4
5
6
7
8
9
1%Au/TiO2
0.022
0.041
0.041
0.014
0.075
0.298
–
0.043
0.022
0.022
0.01
382
0.21
0
0
0
0
0
0.02
244
254
0
230
<1
0
0
0
0
0
<1
164
158
0
b
c
1% Pd/TiO2
1% Pd/TiO2
c
0.5% Pt/TiO2
c
1% Ni/TiO2
b
5% Rh/TiO2
b
TiO2
b
1% Au@0.5% Pd/TiO2
a
equipped with a column Mediterranea C18 (5
l
m, 25 Â 0.46 mm)
1% Au/Al2O3
10
11
12
1% Au/ZnOa
HAuCl4
was used; the mobile phase was acetonitrile/ethanol in a 30:70
ratio, and the flow was 0.5 mL/min; and detection was done using
a PDA UV–Vis detector at a wavelength of 254 nm.
(CH3)Au(PPh3)
0.01
0
0
Some reactions were scaled-up (15 g of substrate) and the prod-
ucts purified by removal of the starting reactants in a rotatory
evaporator at temperatures in the range 313–333 K. Isolated yields
were calculated and are reported in the SI. 1H NMR and 13C NMR
spectra were recorded in CDCl3 or CD2Cl2 with tetramethylsilane
as the internal standard at 298 K on a Bruker Avance 300 (spectra
provided in SI).
Turnovers calculated per metal atom on the external surface of the nanoparticles
[33]. Experimental details provided in SI.
a
Supplied by STREM (AUROlite™ catalysts).
Dried at 373 K before reaction.
b
c
Preactivated in H2 at 723 K before reaction.
Mol of metal  mol of benzeneÀ1  100.
d
Turnover frequency (mol of benzene converted  mol of metalÀ1  hÀ1).
e
f
Turnover number (mol of benzene converted  mol of metalÀ1).

P. Serna, A. Corma / Journal of Catalysis 315 (2014) 41–47
43
Table 2
Catalytic performance of Au/TiO2 for the O2-assisted coupling of various arenes at 12 bar.
e
Sustrate
Ta (K)
% Aub (mol)
TOFc (hÀ1
)
TONd (mol/mol)
SBiaryl (%)
Benzene
Toluene
p-Xylene
1,2,4-Trimethylbenzene
Chlorobenzene
Nitrobenzene
Phenol
413
373
373
373
413
413
373
0.022
0.026
0.030
0.034
0.032
0.035
0.027
382
176
18
230
206
71
8
88
98.5
97.6
98.1
98.4
98.7
98.6
80.3
6
368
176
85
175
336
Turnovers calculated per metal atom on the external surface of the nanoparticles. Experimental details provided in SI.
a
Reaction temperature.
b
Mol of metal  mol of benzeneÀ1  100.
c
Turnover frequency (mol of benzene converted  mol of metalÀ1  hÀ1).
d
Turnover number (mol of benzene converted  mol of metalÀ1).
e
Selectivity to the biaryl.
to work in the high- and low (77 K)-temperature range. Prior to CO
adsorption experiments, the sample has been evacuated at 298 K
in a vacuum (10–6 mbar) for 1 h. In one case, the catalyst has been
treated in a 4% CO/He flow at 298 and 373 K for 1 h prior to the CO
adsorption experiment. CO adsorption experiments have been per-
formed at 77 K in the 0.2–20 mbar range. Spectra were recorded
once complete coverage of CO at the specified CO partial pressure
has been achieved.
3. Results and discussion
3.1. Aerobic-coupling of arenes by gold and other supported metal
catalysts
Fig. 1. Activity (TON, blue bars; and TOF, red bars) of various catalysts in the O2-
assisted coupling of benzene at 413 K and 12 bar. The insert shows the performance
(TON, blue bars; TOF, red bars; and selectivity to the corresponding substituted
biaryls, green bars) of Au/TiO2 (3.2 nm average diameter) in the coupling of various
arenes. Details of experimental conditions provided in SI. (For interpretation of the
references to color in this figure legend, the reader is referred to the web version of
this article.)
When benzene is contacted with gold nanoparticles on TiO2
(Au/TiO2) in the presence of O2 at 413 K and 12 bar under
solvent-free conditions, a turnover frequency (TOF) and turnover
number (TON) of 382 hÀ1 and 230, respectively, are obtained
(reaction time = 118 h). TOF and TON are calculated here per sur-
face atom in the supported gold nanoparticles, as estimated from
the HAADF-STEM and HRTEM micrographs given in Figure S1
[39–41]. Gold nanoparticles on other supports such as ZnO and
Al2O3 are also active for the O2-assisted coupling of benzene
(Table 1), but gold species such as HAuCl4 or (CH3)(PPh)3Au in
solution are not under the same reaction conditions. The selectivity
to biphenyl with the supported gold nanoparticles is close to 99%,
and blank experiments with TiO2 show that the support is inactive.
To further check that gold species in solution, leached from the cat-
alyst surface during the reaction, are not responsible for the CAC
bond formation, we removed the Au/TiO2 catalyst from the reac-
tion solution after several turnovers, which resulted in cessation
of the catalytic activity (Figure S2). Catalyst removal was carried
out by filtration at the reaction temperature (413 K).
The catalytic behavior of gold nanoparticles for coupling of ben-
zene appears to be quite unique, since other metals such as Rh, Pt,
Pd, or Ni supported on TiO2 are not active for the formation of
biphenyl. Only the TiO2-supported Pd catalyst gives trace amounts
of the coupling product provided that the sample has not been
reduced in H2 before reaction. Nevertheless, the TON of supported
Pd was lower than 1 after a 24 h reaction (Table 1). We infer that
Pd2+ species in the unreduced Pd/TiO2 sample are responsible for
a stoichiometric transformation of benzene into biphenyl, as has
been reported previously [15]. In fact, when adding Pd to Au/
TiO2, the activity of the gold catalyst drops from 382 hÀ1 to
0.02 hÀ1 (TOF measured at 413 K and 12 bar of O2). This is due to
the incorporation of Pd onto the gold surface in a core–shell
conformation (Fig. 2e and Figure S3), with Pd, we infer, acting as
a catalyst poison.
2.3. Electron microscopy
The experiments have been performed in a JEOL-JEM-2010F
microscope with a LaB6 electron gun operated at 200 kV both in
scanning-transmission mode (STEM). This microscope has a struc-
tural resolution of 0.19 nm and allows forming, in STEM mode,
electron probes with diameters down to 0.5 nm suitable for high-
spatial resolution analytical investigation. STEM images were
obtained using a High Angle Annular Dark Field detector (HAADF),
which allows Z-contrast imaging. Analytical electron microscopy
was performed with an ATW type EDS detector and the INCA
Energy TEM platform. EDS spectra were collected at specific points
of the samples and along electron-beam paths going from vacuum
through surface positions of the metal particles and the support.
Particle size distributions were determined upon measurement
of, at least, 200 particles per sample.
Samples for electron microscopy studies were prepared by
depositing small amounts of the powders directly onto holey-car-
bon coated Cu grids. Excess powder was removed from the grids by
gentle blowing of air with a nozzle.
2.4. FTIR experiments
Fourier transform infrared (FTIR) spectra have been obtained on
a Biorad FTS-40A spectrometer equipped with a DTGS detector.
The experiments have been carried out in a homemade IR cell able

44
P. Serna, A. Corma / Journal of Catalysis 315 (2014) 41–47
3.2. Influence of gold particle size on catalytic performance
expected for a fresh catalyst of a similar average particle size. Thus,
while some sintering is observed, this can only be partially respon-
sible for the observed catalytic activity decay.
We have found that the particle size of supported gold strongly
influences the activity of Au/TiO2, being the TOF greater as the gold
crystallite size decreases (Fig. 2). Notice, however, that a maximum
of activity is observed for a 3 nm diameter, approximately, with a
decrease in activity for further size decrease. We infer, thus, that
most dispersed forms of gold (small clusters or isolated atoms)
are less effective than gold crystallites of ꢀ3 nm diameter for this
reaction, as it also occurs with the homocoupling of iodobenzene
[42]. To test this hypothesis, Au/TiO2 with a 3.2-nm average parti-
cle size was treated with iodomethane. This treatment has been
reported to generate gold clusters and isolated gold atoms [43].
The resulting material gives almost no activity for the coupling of
benzene. We stress, nevertheless, that some iodine remains
adsorbed on the catalyst surface,[43] which may also poison the
active sites [42,44].
Fig. 3A shows the number of turnovers as a function of the
reaction time in the coupling of benzene with the 3-nm average
size Au/TiO2 catalyst. The kinetic curve evidences that the rate of
formation of biphenyl is highest at the shortest contact times,
decreasing substantially afterward. Fig. 3, thus, evidences the
occurrence of catalyst deactivation. However, the reaction rate sta-
bilizes after approximately 5 h and is fairly sustained for the next
120 h, indicating the approach to a steady state situation.
At this point, the reacted catalyst was characterized by IR spec-
troscopy after being dried to remove the excess of physisorbed
benzene. The results in Figure S4 evidence the accumulation of aro-
matic compounds on the catalyst surface, as inferred from the
growth of IR bands at 3034 and 3066 cmÀ1, ascribed to the @CAH
stretch in arenes, and the appearance of other bands in the range
1400–1600 cmÀ1, typical of the CAC stretching vibrations in aro-
matics. When the reacted catalyst was subsequently probed with
CO at 77 K, no IR bands attributable to metallic Au0 species were
observed, in contrast to the spectra of the fresh Au/TiO2 catalyst,
which shows intense bands in the range 2096–2125 cmÀ1 region
characteristic of Au0 species in this class of materials [45–47].
These results suggest the presence of aromatic compounds on
top of the gold nanoparticles, and we point to these surface species
blocking the Au0 sites as potentially responsible for the observed
catalyst deactivation.
The aromatic compounds can be removed from the catalyst sur-
face without further aggregation of the gold nanoparticles by calci-
nation in air at 673 K for 3 h. TEM of the regenerated sample
shows, in fact, that gold redisperses, to some extent, over the
TiO2 support when calcined at 673 K (Figure S1, G vs. H). Interest-
ingly, the regenerated catalyst behaves nearly identically with
fresh benzene from conversion and selectivity standpoints, as
shown in Fig. 2 and Figure S5 along three consecutive runs. The
number of turnovers achieved with the AuTiO2 catalyst after three
recycles is >600. We note that due to the high arene/gold ratio in
our batch experiments (ꢀ4500), and because of the occurrence of
catalyst deactivation, the conversion generally achieved with the
supported gold catalysts is low after a single run (ꢀ2%). The
conversion can be increased to ꢀ4%, and the reaction time short-
ened, by removal of the deactivated catalyst and addition of regen-
erated material every few hours (Fig. 4). This makes the Au/TiO2
catalyst to complete a similar number of transformations (ꢀ0.4
vs. ꢀ0.7 mmol) than with the HAuCl4/PhI(OAc)2/HOAc/benzene
system (0.02/1/17.5/20, mol ratio),[13] but with a remarkably
3.3. Catalyst regenerability
To elucidate the origin of the catalyst deactivation, the catalyst
was recovered after reaction, and characterized by HAADF-STEM
microscopy. The images evidence that the average particle size
increases from approximately 3.2 nm to approximately 4.5 nm
after reaction, with the formation of a significant fraction of nano-
particles in the range 7–12 nm (Fig. 3D) that are, in contrast, neg-
ligible in the fresh Au/TiO2 catalyst (Fig. 3B). Nevertheless, taking
into account the correlation between the catalytic activity (per
external gold atom) and the average particle size (Fig. 2), it is evi-
dent that the activity of the used catalyst is much lower than that
Fig. 2. (A) HRTEM micrographs of some representative gold catalysts used for the coupling of non-activated arenes in O2 (image e corresponds to a bimetallic Pd@Au catalyst).
(B) Particle size distributions of several supported gold catalysts used in the present work. (C) Dependence of the TOF per external gold atom on the average particle size
during the solvent-free coupling of benzene at 12 bar of O2 and 413 K for the series of catalysts in (B). The selectivity to biphenyl was always >98%. Experimental details are
provided in SI and in the Section 2.

P. Serna, A. Corma / Journal of Catalysis 315 (2014) 41–47
45
Fig. 3. (A) Evolution of the TON (per external gold atom of the fresh catalyst) with time during the solvent-free coupling of benzene in O2 at 413 K and 12 bar, using an Au/
TiO2 sample (AUROlite™). The selectivity to biphenyl was >98%. (B and D) Particle size distributions of the preceding catalyst before and after a 100 h reaction, respectively. (C
and E) IR spectra of the fresh and used catalyst, respectively, using CO as a probe molecule.
greater TON (ꢀ240 vs. ꢀ35, per total gold atoms used) and no
agrees well with our observation that palladium hinders the
activity of gold for establishment of a new CAC bond, as depicted
hereinabove (Fig. 2 and Table 1).
iodine, acids or bases needed.
At 373 K, the CAC formation rate on our Au/TiO2 catalyst is lar-
gely determined by the number of methyl substituents, which
defines the degree of sterical hindrance. Benzene and toluene yield
the corresponding biaryls at similar rates, whereas a notable
decrease in activity is observed for the di- and, especially, the tri-
substituted methylbenzenes (Fig. 1, insert). In contrast, we do
not observe a straightforward dependence of the reactivity of
substituted arenes with the electron-withdrawing character of
the substituent. For example, despite the presence of the deactivat-
ing halide function in chlorobenzene, this molecule reacts at a rate
comparable to that of toluene (368 hÀ1 vs. 356 hÀ1, respectively),
whereby the ACH3 group acts as a moderately activating group.
With nitrobenzene, which incorporates a strongly deactivating
ANO2 group, the reaction rate is only 2-fold lower than that of tol-
uene (176 hÀ1 vs. 356 hÀ1, respectively), and no reaction at all is
observed with the relatively less deactivated iodobenzene. The
presence of the activating AOH group in phenol does not result
in an increase in the coupling activity with regard to toluene
(85 hÀ1 vs. 176 hÀ1) (see Fig. 1, insert). We remark that Au/TiO2
catalyzes the O2-assisted conversion of the strongly deactivated
nitrobenzene into dinitrobiphenyls at 413 K with high selectivity
(>98%), a TOF of 176 hÀ1 and, at least, 175 turnovers completed,
improving recent results with Pd salts in the presence of trifluoro-
acetic acid (approximately 5 turnovers, and a TOF < 0.2 hÀ1 at
378 K) [17].
3.4. Effect of arene substituents
The 3.2 nm Au/TiO2 catalyst provides high selectivity to the cor-
responding substituted biaryls from arenes that incorporate one or
several methyl groups, chloride, nitro, or hydroxyl substituents,
provided that the reaction temperature is properly selected.
Methyl groups in mono-, di-, and tri-substituted benzenes (tolu-
ene, p-xylene, and 1,2,4-trimethylbenzene) underwent oxidation
to aldehydes and carboxylic acids when the temperature was
413 K, with selectivities to dimethylbiphenyl, tetramethylbiphe-
nyl, and hexamethylbiphenyl, respectively, lower than 80%. The
selectivity to coupling products, however, increases remarkably
(>98%) if the reaction is performed at a milder temperature
(373 K, Table 2). This result is consistent with earlier reports show-
ing that gold nanoparticles catalyze the oxidation of alkylbenzenes,
even toward full combustion, when the reaction temperature is
sufficiently high [48,49]. At lower temperatures, and in solvent-
free conditions, in contrast, biaryls that retain the substituents of
the starting arenes are obtained. Interestingly, recent reports show
that toluene in the absence of a solvent oxidizes to benzaldehyde,
benzoic acid, and benzyl benzoate on bimetallic Au/Pd catalysts at
353–433 K, without formation of biaryl products [50]. This result
On the other hand, we observe a mixture of the six possible
regioisomers during the transformation of toluene, chlorobenzene,
and nitrobenzene that cannot be rationalized according to the clas-
sical electrophilic substitution pattern, in contrast to the results
with soluble Au salts in the presence of iodine oxidants [13].
3.5. Reaction mechanism: first insights from kinetic data
The coupling reaction should require, on the one hand, the
dissociation of O2 and, on the other hand, the CAH bond scission.
Since we have not observed any regioselectivity during the
coupling of, for example, toluene, a working hypothesis is that
the CAC bond formation involves carbon radicals. The homolytic
sequestration of H has been identified as a reaction step in a
number of gold-catalyzed oxidation reactions, such as the alcohol
oxidation [51,52], or in the oxidative coupling of anilines to azo-
benzenes [53], although the investigation of the ArAH cleavage
Fig. 4. Conversion vs. time curves for the solvent-free oxidative coupling of toluene
catalyzed by Au/TiO2. Open symbols refer to a single catalytic experiment over
approximately 210 h. Solid circles refer to conversion accumulated along several
shorter consecutive runs (approximately 24 h duration) with the catalyst regener-
ated in air at 673 K in between each run. Reaction conditions stated in Table 2.

46
P. Serna, A. Corma / Journal of Catalysis 315 (2014) 41–47
Fig. 5. Influence of the pressure of O2 (A) and the concentration of arene (B) in the initial rate of benzene coupling at 373 K and 413 K, respectively. The selectivity to biphenyl
was always >98%. Experiment A was performed under solvent-free conditions, and experiment B at a pressure of O2 of 12 bar.
by gold nanoparticles is scarce. However, the electron-induced
dissociation of benzene to phenyl radicals and subsequent obser-
vation of biaryl products have been demonstrated [54]. Consistent
with the radical mechanism, when we introduce radical scavengers
such as hydroquinone into the reaction, the initial benzene to
biphenyl coupling rate falls from 382 hÀ1 to <0.02 hÀ1 on the 3.2-
nm average size Au/TiO2 catalyst.
Acknowledgments
The research was supported by Project CONSOLIDER INGENIO
(MULTICAT), The SEVERO OCHOA Program for centers of excel-
lence, the European Union Seventh Framework Programme (FP7/
2007-2013) under Grant Agreement PIOF-GA-2009-253129 (P.S.).
Concerning the O2 activation, gold nanoparticles on various
metal oxides have been shown to catalyze a number of reactions
that involve O2 dissociation. [28,30,32,53–56] We emphasize, how-
ever, that other noble metals, exemplified by palladium, are also
good oxidation catalysts, but they are not catalytically active for
the O2-assisted coupling reaction of non-activated aromatic, as
demonstrated here, even at temperatures and pressures markedly
higher than those strictly needed to form biaryls with the Au nano-
particles (Fig. 1).
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2014.04.004.
References
[1] G. Bringmann, R. Walter, R. Weirich, Angew. Chem., Int. Ed. Engl. 29 (1990)
977.
[2] M. Sainsbury, Tetrahedron 36 (1980) 3327.
To find out which is the rate-determining reaction step on the
supported gold catalysts, initial rates were measured at variable
partial pressures of O2 while keeping constant the concentration
of benzene, and vice versa. In the latter case, benzene was diluted
in decane, a solvent that does not react under our reaction condi-
tions and is expected to interact weakly with the gold nanoparti-
cles. Solvents such as acetone, ethanol, dichloroethane, or acetic
acid were found to be, in contrast, unsuitable (no biaryl formation
observed).
These experiments show that the initial coupling rate of ben-
zene is little influenced by the O2 pressure in the range 1–12 bar
(Fig. 5B), although no activity is observed in the total absence of
O2. At 373 K and 1 bar of O2, for example, more than 175 turnovers
toward biphenyl can be accomplished with the 3.2-nm average
size Au/TiO2 catalyst, a result that greatly differs from the high
O2 pressures (35–50 bar) often required with Pd salts in the addi-
tional presence of acids and other metal functions to provide just
a few catalytic turnovers [18–20].
[3] J. Hassan, M. Sévignon, C. Gozzi, E. Shulz, M. Lemaire, Chem. Rev. 102 (2002)
1359.
[4] N. Miyaura, in: L.S. Libeskind (Ed.), Advances in Metal-organic Chemistry, vol.
6, Jai, London, 1998, pp. 187–243.
[5] R.W. Boyd, R. Morrison, Organic Chemistry, Prentice Hall, Englewood Cliffs, NJ,
1992. p. 947.
[6] H.J. Lucas, E.R. Kennedy, Org. Synth. Coll. 2 (1939) 351.
[7] (a) N. Miyamura, T. Yanagi, A. Suzuki, Synth. Commun. 11 (1981) 513;
(b) A. Suzuki, Pure Appl. Chem. 63 (1991) 419.
[8] D. Milstein, J.K. Stille, Am. Chem. Soc. 100 (1978) 3636.
[9] C.-L. Sun, H. Li, D.-G. Yu, M. Yu, X. Zhou, X.-Y. Lu, K. Huang, S.-F. Zheng, B.-J. Li,
Z.-J. Shi, Nat. Chem. 2 (2010) 1044.
[10] N.R. Deprez, D. Kalyani, A. Krause, M.S. Sanford, J. Am. Chem. Soc. 128 (2006)
4972.
[11] R.N. Dhital, C. Kamonsatikul, E. Somsook, K. Bobuatong, M. Ehara, S. Karanjit, H.
Sakurai, J. Am. Chem. Soc. 134 (2012) 20250.
[12] J. Guzman, A. Corma, Angew. Chem., Int. Ed. 44 (2005) 2242.
[13] A. Kar, N. Mangu, H.M. Kaiser, M.K. Tse, J. Organomet. Chem. 694 (2009) 524.
[14] L.T. Ball, G.C. Lloyd-Jones, C.A. Russell, Science 337 (2012) 1644.
[15] Y. Liu, K. Murata, M. Inaba, J. Mol. Catal. A 256 (2006) 247.
[16] S. Mukhopadhyay, G. Rothenberg, G. Lando, K. Agbaria, M. Kazanci, Y. Sasson,
Adv. Synth. Catal. 343 (2001) 455.
[17] L. Zhou, W. Lu, Organometallics 31 (2012) 2124.
[18] M.O. Unger, R.A. Fouty, J. Org. Chem. 34 (1969) 18.
[19] A.K. Yatsimirsky, S.A. Deiko, A.D. Ryabov, Tetrahedron 39 (1983) 2381.
[20] I.V. Kozhevnikov, React. Kinet. Catal. Lett. 6 (1977) 401.
[21] S. Mukhopadhyay, G. Rothenberg, G. Lando, K. Agbaria, M. Kazanci, Y. Sasson,
Adv. Synth. Catal. 5 (2001) 455.
On the other hand, the rate of biphenyl formation increases lin-
early with the concentration of benzene in the solution (Fig. 5A),
and we infer, thus, that the CAC formation rate is not controlled
by the activation of O2, but controlled by the activation of the
arene, with AuAO species potentially playing a role during the
ArAH cleavage.
[22] H.A. Wegner, M. Auzias, Angew. Chem., Int. Ed. 50 (2011) 8236.
[23] S.K. Beaumont, G. Kyriakou, R.M. Lambert, J. Am. Chem. Soc. 132 (2010) 12246.
[24] A. Corma, R. Juárez, M. Boronat, F. Sánchez, M. Iglesias, H. García, Chem.
Commun. 47 (2010) 1446.
[25] H. Tsunoyama, H. Sakurai, N. Ichikuni, Y. Negishi, T. Tsukuda, Langmuir 20
(2004) 11293.
[26] J. Han, Y. Liu, R. Guo, J. Am. Chem. Soc. 131 (2009) 2060.
[27] G.J. Hutchings, Catal. Today 100 (2005) 55.
4. Conclusions
In summary, our results provide valuable insights on a new pro-
cess for the formation of biaryls in O2 using re-usable solid gold
catalysts. Further work is needed to understand the exact mode
of action of supported gold nanoparticle in the coupling of unacti-
vated arenes, so that more efficient catalysts that, for example,
circumvent deactivation issues, can be designed. Nonetheless, we
consider this advance an important achievement on route toward
a waste-free production of biaryls.
[28] A.S.K. Hashmi, G.J. Hutchings, Angew. Chem., Int. Ed. 45 (2006) 7896.
[29] G.C. Bond, C. Louis, D.T. Thompson, in: Catalysis by Gold, Imperial College
Press, London, 2006.
[30] M. Boronat, A. Corma, Dalton Trans. 39 (2010) 8538.
[31] T. Mallat, A. Baiker, Annu. Rev. Chem. Biomol. Eng. 3 (2012) 11.
[32] M. Haruta, Catal. Today 36 (1997) 153.
[33] M.D. Hughes, Y.-J. Xu, P. Jenkins, P. McMorn, P. Landon, D.I. Enache, A.F. Carley,
G.A. Attard, G.J. Hutchings, F. King, E.H. Stitt, P. Johnston, K. Griffin, C.J. Kiely,
Nature 43 (2005) 1132.
[34] P. Claus, Appl. Catal. A 291 (2005) 222.

P. Serna, A. Corma / Journal of Catalysis 315 (2014) 41–47
47
[35] A. Corma, P. Serna, Science 313 (2006) 332.
[36] M. Haruta, N. Yamada, T. Kobayashi, S. Iijima, J. Catal. 115 (1989) 301.
[37] M. Haruta, Nature 437 (2005) 1098.
[47] M. Boronat, P. Concepción, A. Corma, J. Phys. Chem. C 113 (2009) 16772.
[48] H. Wu, L. Wang, Z. Shen, J. Zhao, J. Mol. Catal. A: Chem. 351 (2011) 188.
[49] H. Wu, L. Wang, J. Zhang, Z. Shen, J. Zhao, Catal. Commun. 12 (2011) 859.
[38] P. Serna, M. Boronat, A. Corma, Top Catal. 54 (2011) 439.
[39] R. Van Hardeveld, F. Hartog, Surf. Sci. 15 (1969) 189.
[40] R.E. Benfield, J. Chem. Soc. Faraday Trans. 88 (1992) 1107.
[41] A. Abad, A. Corma, H. García, Chem. Eur. J. 14 (2008) 212.
[42] M. Boronat, D. Combita, P. Concepción, A. Corma, H. García, R. Juárez, S.
Laursen, J.D. López-Castro, J. Phys. Chem. C 116 (2012) 24855.
[43] J. Sá, S.F.R. Taylor, H. Daly, A. Goguet, R. Tiruvalam, Q. He, C.J. Kiely, G.J.
Hutchings, C. Hardacre, ACS Catal. 2 (2012) 552.
[50] L. Kesavan, R. Tiruvalam, M.H. Ab Rahim, M.I. bin Saiman, D.I. Enache, R.L.
Jenkins, N. Dimitratos, J.A. Lopez-Sanchez, S.H. Taylor, D.W. Knight, C.J. Kiely,
G.J. Hutchings, Science 331 (2011) 195.
[51] P. Ionita, M. Conte, B.C. Gilbert, V. Chechik, Org. Biomol. Chem. 5 (2007) 3504.
[52] M. Conte, H. Miyamura, S. Kobayashi, V. Chechik, J. Am. Chem. Soc. 131 (2009)
7189.
[53] A. Grirrane, A. Corma, H. García, Science 322 (2008) 1661.
[54] D. Syomin, J. Kim, B.E. Koel, J. Phys. Chem. B 105 (2001) 8387.
[55] J.T. Miller, A.J. Kropf, Y. Zha, J.R. Regalbuto, L. Delannoy, C. Louis, E. Bus, J.A. van
Bokhoven, J. Catal. 240 (2006) 222.
[44] R.N. Dhital, C. Kamonsatikul, E. Somsook, Y. Sato, H. Sakurai, Chem. Commun.
49 (2013) 2542.
[45] H. Liu, A.I. Kozlov, A.P. Kozlova, T. Shido, K. Asakura, Y. Iwasawa, J. Catal. 185
(1999) 252.
[56] J.A. van Bokhoven, C. Louis, J.T. Miller, M. Tromp, O.V. Safonova, P. Glatzel,
Angew. Chem., Int. Ed. 45 (2006) 4651.
[46] F. Boccuzzi, A. Chiorino, S. Tsubota, M. Haruta, J. Phys. Chem. 100 (1996) 3625.