Novel blended catalysts consisting of a TiO2 photocatalyst and an Al2O3 supported Pd-Au bimetallic catalyst for direct dehydrogenative cross-coupling between arenes and tetrahydrofuran

Article abstract of DOI:10.1039/c8ra02948b

Dehydrogenative cross-coupling (DCC) is a clean methodology to make C-C bonds by using abundant C-H bonds. The blended catalyst, developed in this study, consists of a TiO₂ photocatalyst and an Al₂O₃ supported Pd-Au bimetallic catalyst and shows superior activity to the conventional TiO₂ photocatalyst loaded with the corresponding metal co-catalyst for the direct DCC between various arenes and tetrahydrofuran, with concomitant evolution of hydrogen gas. The reactions were done under mild conditions without consuming any oxidising agent or other additional chemicals. This new approach of separating the photocatalyst and the metal catalyst parts allows their independent modification to improve the overall catalytic performance.

Full text of DOI:10.1039/c8ra02948b

RSC Advances
View Article Online
View Journal | View Issue
PAPER
Novel blended catalysts consisting of a TiO2
photocatalyst and an Al2O3 supported Pd–Au
bimetallic catalyst for direct dehydrogenative
cross-coupling between arenes and
tetrahydrofuran†
Cite this: RSC Adv., 2018, 8, 24021
a
ab
ab
*
Akanksha Tyagi,
Akira Yamamoto
and Hisao Yoshida
Dehydrogenative cross-coupling (DCC) is a clean methodology to make C–C bonds by using abundant
C–H bonds. The blended catalyst, developed in this study, consists of a TiO2 photocatalyst and an Al2O3
supported Pd–Au bimetallic catalyst and shows superior activity to the conventional TiO2 photocatalyst
loaded with the corresponding metal co-catalyst for the direct DCC between various arenes and
tetrahydrofuran, with concomitant evolution of hydrogen gas. The reactions were done under mild
conditions without consuming any oxidising agent or other additional chemicals. This new approach of
separating the photocatalyst and the metal catalyst parts allows their independent modification to
improve the overall catalytic performance.
Received 6th April 2018
Accepted 25th June 2018
DOI: 10.1039/c8ra02948b
rsc.li/rsc-advances
many cases, metal nanoparticles are loaded on TiO2 (M/TiO2),
that can accept the excited electrons from the TiO2 photo-
Introduction
catalyst, to decrease the electron–hole recombination and
increase the photocatalytic efficiency.4 In these cases, the metal
nanoparticles function as an electron receiver.
Dehydrogenative cross-coupling (DCC) is an efficient method-
ology to make carbon–carbon (C–C) bonds between two
different organic compounds.1 This direct route to make C–C
bonds yields hydrogen as a by-product, which can be used as
a fuel or for hydrogenation reactions. Different catalysts have
been developed to carry out the thermal dehydrogenation
reactions,2 but the scope for further improvement remains.
Recently, the use of an abundant, safe, and heterogeneous
TiO2 photocatalyst for DCC has become more attractive as it
works under mild reaction conditions.3 The photogenerated
holes on TiO2 can activate the C–H bonds in various organic
molecules to generate the corresponding radical species and
protons. The generated radicals can further react with other
radical species or molecules to give the coupling products by
making new C–C bonds, while the protons are reduced by the
photoexcited electrons to form hydrogen molecules. Thus, the
TiO2 photocatalyst efficiently brings out DCC with concomitant
hydrogen gas evolution, without any external oxidising agent. In
On the other hand, it is notable that the metal nano-
particles loaded on TiO2 sometimes can also participate
catalytically in some photocatalytic reactions, thus acting as
co-catalysts.3b–3e,5 For example, Pd nanoparticles catalyse
a reaction between an aromatic molecule and different carbon
centred radicals,3b–3d and Pt nanoparticles catalyse reactions
like photocatalytic hydrogenation,5a dehydrogenation,5c and
C–H bond activation.3e,5b To develop such a hybrid catalyst
system consisting of a photocatalyst and a metal co-catalyst,
one may plan some modications to improve the catalytic
activity, e.g., modifying the metal co-catalyst by increasing its
loading amount and dispersion, making an alloy with other
metal, varying its oxidation state and so on. However, it is
difficult to realise these alternations of the metal co-catalyst
part on the TiO2 surface without changing the photocatalytic
performance of the TiO2 part. It is because TiO2 itself is
susceptible to chemical or thermal treatments and its physical
properties like surface area, crystal phase, and crystal defects,
may be varied.6
aGraduate School of Human and Environmental Studies, Kyoto University, Yoshida
Nihonmatsu-cho, Sakyo-ku, Kyoto 606-8501, Japan. E-mail: yoshida.hisao.2a@
kyoto-u.ac.jp
bElements Strategy Initiative for Catalysts and Batteries (ESICB), Kyoto University,
Kyotodaigaku-Katsura, Nishikyo-ku, Kyoto 615-8520, Japan. E-mail: yoshida.hisao.
2a@kyoto-u.ac.jp
Here, a novel blended catalyst is proposed as one solution to
this problem, which is a physical mixture of a TiO2 photo-
catalyst and a supported metal catalyst. Since the two catalysts
can be prepared separately without affecting each other, they
can be modied independently to achieve high activity entirely.
Although several reports mentioned the use of a kind of the
† Electronic supplementary information (ESI) available: Synthesis of
4-(tetrahydrofuran-2-yl)benzonitrile and the results of photocatalytic direct DCC
between various arenes and THF are mentioned in the ESI. See DOI:
10.1039/c8ra02948b
This journal is © The Royal Society of Chemistry 2018
RSC Adv., 2018, 8, 24021–24028 | 24021

View Article Online
RSC Advances
Paper
blended catalyst including a photocatalyst,7 they were employed samples, Pd(3.0)/Al2O3 and Au(3.0)/Al2O3, were also prepared by
just for investigating the origin of the catalytic activity. a similar procedure.
Employing a blended catalyst to improve the activity is
The prepared samples were characterised by various tech-
uncommon, except for a few examples, such as a combination niques like TEM, XRD, XAFS, and UV-DRS. The transmission
of two kinds photocatalysts to build a Z-scheme for water electron microscopy (TEM) images of the various samples were
splitting,8a–c and a combination of a photocatalyst and an acid taken by a JEM-2200FS Field Electron Microscope. X-ray
catalyst for C–O coupling.8d Thus, this is the rst report for the diffraction (XRD) measurements of the Al2O3 support and
blended catalyst consisting of a photocatalyst and a metal various Al2O3 supported samples were carried out on a Lab X
catalyst for the enhancement of the catalytic activity. We have XRD-6000 Shimadzu. X-ray absorption ne structure (XAFS)
developed it for the sp2C–sp3C photocatalytic direct DCC was used to study the electronic state of the Pd and Au species
between various arenes and ethers like tetrahydrofuran (THF). in the Pd(x)Au(y)/Al2O3 samples. The Pd K-edge and Au LIII
-
In this study, we found that a blended catalyst consisting of edge XAFS measurements were carried out at the BL01B1 beam
a TiO2 photocatalyst and an Al2O3 supported Pd–Au bimetallic line of the synchrotron facility at Spring-8 in the RIKEN Har-
catalyst can convert various arenes and THF to their corre- ima institute (Hyogo, Japan). The measurements were done in
sponding DCC products and hydrogen, with higher activity than a transmission mode. The samples were mixed with a calcu-
the conventional metal loaded TiO2 photocatalyst, under mild lated amount of boron nitride, and the mixture was shaped
conditions without consuming any oxidising agent or other into a pellet which was used for measurements. The spectra
chemicals.
were analysed by Athena soware.10 The ultraviolet-visible
diffuse reectance spectroscopy (UV-DRS) measurements of
the monometallic Pd(3.0)/Al2O3 and Au(3.0)/Al2O3 samples,
and the bimetallic Pd(x)Au(y)/Al2O3 samples were carried out
on a JASCO V-570 instrument.
Experimental section
Catalyst preparation
The Pd loaded TiO2 sample (Pd(3.0)/TiO2) (the number in
parenthesis refers to the loading amount of the metal in
weight%), and the bimetallic Pd–Au loaded TiO2 sample
Procedure for the photocatalytic activity tests
(Pd(2.0)Au(1.0)/TiO2) was prepared by
a photodeposition
method as follows. For the Pd(3.0)/TiO2 sample, 4 g of TiO2 All chemicals were of analytical grade and used without further
powder (JRC-TIO-8 provided by Catalysis Society of Japan, purication; THF (Wako Pure Chemicals, 99%), benzene
anatase phase, surface area 335 m2 gꢀ1) was dispersed in ion- (Nacalai Tesque, 99%), benzaldehyde (Wako Pure Chemicals,
exchanged water (300 mL) and irradiated for 30 min from 99%), benzonitrile (Kishida Chemicals, 99%), toluene (Nacalai
a ceramic xenon lamp (PE300BUV). Then, 100 mL methanol and Tesque, 99%), and aniline (Kishida Chemicals, 99.5%). The
5.8 mL of an aqueous solution of PdCl2 (10.1 mg mLꢀ1 of Pd) cross-coupling product 4-(tetrahydrofuran-2-yl)benzonitrile was
was added to the suspension and the contents were stirred for synthesised according to the literature11 (details are mentioned
15 min without irradiation, followed by 1 h stirring under the in the ESI†).
light irradiation. It was then ltered off with suction, washed
The sp2C–sp3C photocatalytic direct DCC reactions between
with ion-exchanged water, and dried at 323 K for 12 h to get the different arenes and THF were carried out in a closed reactor of
Pd(3.0)/TiO2 sample. The bimetallic Pd(2.0)Au(1.0)/TiO2 sample a Pyrex test tube (volume ¼ 80 mL) and the xenon lamp
was prepared in a similar manner by using 2 g TiO2, 175 mL (PE300BUV) as the light source. The wavelength of the irradi-
water, 50 mL methanol, 6.5 mL of an aqueous solution of PdCl2 ated light was restricted to $350 nm, by using a long pass
(6.12 mg mLꢀ1 of Pd), and 4.1 mL of an aqueous solution of optical lter. The light intensity was maintained at 40 mW cmꢀ2
HAuCl4 (4.8 mg mLꢀ1 of Au).
(measured at 360 nm ꢁ 60 nm). The reaction test involved the
The various Al2O3 supported samples (monometallic Pd(3.0)/ pre-treatment of the catalyst (photocatalyst or the blended
Al2O3 and Au(3.0)/Al2O3 and bimetallic Pd(x)Au(y)/Al2O3) were catalyst) by light irradiation from the xenon lamp for 1 h. Next,
prepared in a similar manner to the reported method9 with the test tube was sealed with a silicon septum and aer a 10 min
some modications, as mentioned below. 2 g of the Al2O3 argon purge, the reactants (arene and THF) were added and the
powder (JRC-ALO-7 provided by the Catalysis Society of Japan, g- resultant suspension was magnetically stirred under the light
phase, 180 m2 gꢀ1) and the desired volume of the aqueous irradiation for the desired time. Aer the reaction, a part of the
solutions of PdCl2 and HAuCl4 were dispersed in 60 mL of ion- gaseous phase was collected in an air-tight syringe and analysed
exchanged water. The suspension was vigorously stirred at by GC-TCD (Shimadzu, GC-8A). The liquid phase was rst
room temperature for 15 min. The pH of the suspension was diluted by ethanol (2 mL), ltered, and analysed by GC-MS
adjusted to 10 by using an aqueous solution of NaOH (1 M). The (Shimadzu, GCMS-QP5050A) by using decane as an internal
resultant slurry was stirred for 24 h at room temperature. Then, standard. The amount of all DCC products was approximately
the contents were ltered and dried overnight in an electric determined from the calibration curve of 4-(tetrahydrofuran-2-
oven at 323 K. Finally, the powder was reduced at 423 K under yl)benzonitrile. The GC-MS yield of the products (% Y) was
pure H2 for 30 min to get the Pd(x)Au(y)/Al2O3 samples. The calculated as % Y ¼ 100 ꢂ [total amount of the DCC products
total loading amount of the metals on the support was xed to 3 (mmol)/initial amount of the arene (mmol)]. The selectivity to the
weight%. For comparison, the monometallic Al2O3-supported DCC products, based on the arene, (% S) was calculated as % S
24022 | RSC Adv., 2018, 8, 24021–24028
This journal is © The Royal Society of Chemistry 2018

View Article Online
Paper
RSC Advances
Fig. 1 TEM images and particle distribution histograms of (a) monometallic Pd(3.0)/Al2O3, (b) Au(3.0)/Al2O3, and (c) bimetallic Pd(2.0)Au(1.0)/
Al2O3 samples.
¼ 100 ꢂ [total amount of DCC product (mmol)/total amount of these diffractions were observed in the XRD patterns of the
the products from arene (mmol)].
prepared monometallic Pd(3.0)/Al2O3 and Au(3.0)/Al2O3, and
bimetallic Pd(x)Au(y)/Al2O3 samples. This could be due to the
small size and high dispersion of the metal nanoparticles, as
revealed by TEM. Another reason could be the broad diffrac-
tions from the Al2O3 support around 2q ¼ 38ꢃ, 46ꢃ, and 67ꢃ,12
which may have interfered with the very small and broad
diffractions from the metal nanoparticles in the prepared
samples.
Fig. 3 shows the normalised Pd K-edge X-ray absorption near
edge structures (XANES) of the reference Pd foil (Fig. 3a), the
monometallic Pd(3.0)/Al2O3 sample (Fig. 3b), and the Pd(x)
Au(y)/Al2O3 samples (Fig. 3c–f). The spectral features in the
prepared samples were clearly different from those of the foil,
which would mainly originate from the property of the nano-
particles. The samples containing a large amount of Au
exhibited slightly different shape from others (Fig. 3f), sug-
gesting the Pd atoms are affected by the Au atoms and have local
structure or electronic state different from those in the Pd
metal. Also, a clear shi in the absorption edge towards the
Results and discussion
Characterisation of Pd(x)Au(y)/Al2O3 samples
Fig. 1 shows the TEM images and particle distribution histo-
grams of three Al2O3 supported samples, monometallic Pd(3.0)/
Al2O3 (Fig. 1a), monometallic Au(3.0)/Al2O3 (Fig. 1b), and
bimetallic Pd(2.0)Au(1.0)/Al2O3 (Fig. 1c) that exhibited the
highest catalytic activity as mentioned later (Pd/Au molar ratio:
3.5). Clearly, all samples contained small and well-dispersed
nanoparticles of 3–4 nm in size.
Fig. 2 shows the XRD patterns of the Al2O3 support (Fig. 2a),
the monometallic Pd(3.0)/Al2O3 (Fig. 2b), monometallic Au(3.0)/
Al2O3 sample (Fig. 2f), and the bimetallic Pd(x)Au(y)/Al2O3
samples (Fig. 2c–e). According to literature, the Pd (111), (200),
and (220) diffractions should appear at 2q ¼ 40.1ꢃ, 46.5ꢃ, and
68.1ꢃ, respectively (ICSD coll. code 41517) while the Au (111),
(220), and (311) diffractions should be at 2q ¼ 38.1ꢃ, 64.5ꢃ, and
77.1ꢃ, respectively (ICSD coll. Code 52 249). However, none of
Fig. 3 Normalised Pd K-edge XANES of metallic Pd as a reference, and
the monometallic and bimetallic catalyst samples: (a) a Pd foil, (b)
Pd(3.0)/Al2O3, (c) Pd(2.5)Au(0.5)/Al2O3, (d) Pd(2.0)Au(1.0)/Al2O3, (e)
Pd(1.5)Au(1.5)/Al2O3, and (f) Pd(0.5)Au(2.5)/Al2O3.
Fig. 2 XRD patterns of the support, the monometallic and the bime-
tallic samples: (a) Al2O3, (b) Pd(3.0)/Al2O3, (c) Pd(2.5)Au(0.5)/Al2O3, (d)
Pd(1.5)Au(1.5)/Al2O3, (e) Pd(0.5)Au(2.5)/Al2O3, and (f) Au(3.0)/Al2O3.
This journal is © The Royal Society of Chemistry 2018
RSC Adv., 2018, 8, 24021–24028 | 24023

View Article Online
RSC Advances
Paper
Fig. 6 UV-DRS of the Al2O3-supported samples; (a) Pd(3.0)/Al2O3, (b)
Pd(2.5)Au(0.5)/Al2O3, (c) Pd(2.0)Au(1.0)/Al2O3, (d) Pd(1.5)Au(1.5)/Al2O3,
(e) Pd(0.5)Au(2.5)/Al2O3, and (f) Au(3.0)/Al2O3.
Fig. 4 Expanded normalised Pd K-edge XANES of (a) a Pd foil refer-
ence foil, (b) the monometallic Pd(3.0)/Al2O3 sample, and the bime-
tallic (c) Pd(2.5)Au(0.5)/Al2O3 and (d) Pd(2.0)Au(1.0)/Al2O3 samples.
suggesting that the Au atoms became electron rich on the
support. A clear absorption shoulder peak was observed around
11 923 eV in the Au foil and the monometallic Au(3.0)/Al2O3
sample, which corresponds to the transition from the lled core
2p3/2 level to the vacant d orbitals.14 The intensity of this
shoulder peak gradually decreased with increasing the Pd
content in the bimetallic Pd(x)Au(y)/Al2O3 samples, indicating
the lling of the d orbitals in Au, which would arise due to the
higher energy was observed in the order: the reference Pd foil
(Fig. 4a), the monometallic Pd(3.0)/Al2O3 sample (Fig. 4b), the
bimetallic Pd(2.5)Au(0.5)/Al2O3 (Fig. 4c) and Pd(2.0)Au(1.0)/
Al2O3 samples (Fig. 4d). The shi from the Pd foil to the
monometallic Pd(3.0)/Al2O3 sample is oen observed in the
supported metal nanoparticles, and can be attributed to the
acidic property of the Al2O3 support and nanosize of the parti-
cles. The further shis from the monometallic Pd to the
bimetallic Pd–Au samples suggest that the Pd species became
electron decient upon the introduction of Au.13 This would
arise if the Pd and Au atoms in nanoparticles are in intimate
contact, and a transfer of electron density occurs from Pd atoms
to Au atoms, i.e., Pdd+–Audꢀ. Fig. 5 shows the normalised XANES
of Au LIII-edge. The XANES features of the prepared Al2O3 sup-
ported samples, especially at the higher energy region, were
similar to that of Au foil indicating that the Au species in these
samples was almost metallic. However, the Au absorption edge
was clearly shied in the prepared supported samples at
11 919.7 eV (Fig. 5b–e) from the Au foil at 11 920.2 eV (Fig. 5a),
transfer of electron density from Pd to Au atoms, i.e., Pdd+–Audꢀ
.
Thus, the XANES results revealed that the Pd and Au atoms had
an intimate contact in the bimetallic Pd(x)Au(y)/Al2O3 samples,
which affected their local structure and facilitated the electron
transfer. As a result, the Pd atom are proposed to be electron
decient in the bimetallic Pd(x)Au(y)/Al2O3 samples.
Fig. 6 shows the UV-DR spectra of different Al2O3 supported
samples. The monometallic Pd(3.0)/Al2O3 sample showed
a very broad band over almost the entire range without any
clear obvious peak (Fig. 6a). With the introduction of Au to the
samples, the spectra gradually changed (Fig. 6b–e) and the
plasmonic peak of Au nanoparticles became clearly visible
around 520 nm in wavelength for the Pd(0.5)Au(2.5)/Al2O3
sample and the monometallic Au(3.0)/Al2O3 sample (Fig. 6e
and f). However, the spectra of the bimetallic samples could
not be simulated by that of the two monometallic Pd(3.0)/
Al2O3 and Au(3.0)/Al2O3 samples, which suggests that the Pd
and Au species are not independently present on the surface of
the support but in close contact, and thus they affect the
electronic properties of each other, which would provide
unique catalytic property.
Based on these results we conclude that the Pd and Au
atoms in the bimetallic Pd(x)Au(y)/Al2O3 samples have a strong
interaction, making the Pd atoms to be electron decient.
Photocatalytic direct DCC between benzene and THF with
different catalysts
Fig. 5 Normalised Au LIII-edge XANES of a metallic Au as a reference,
and the monometallic and bimetallic catalyst samples: (a) an Au foil, (b)
the monometallic Au(3.0)/Al2O3 sample, and the bimetallic (c) Pd(0.5)
Au(2.5)/Al2O3, (d) Pd(1.5)Au(1.5)/Al2O3, and (e) Pd(2.0)Au(1.0)/Al2O3
samples.
Various photocatalysts like TiO2, Pd(3.0)/TiO2 and Pd(2.0)
Au(1.0)/TiO2, and Al2O3 supported metal catalysts like Pd(3.0)/
Al2O3, Au(3.0)/Al2O3 and Pd(x)Au(y)/Al2O3 were used for the
photocatalytic direct DCC between arenes and THF.
24024 | RSC Adv., 2018, 8, 24021–24028
This journal is © The Royal Society of Chemistry 2018

View Article Online
Paper
RSC Advances
Table 1 Photocatalytic direct DCC between benzene and THF with Table 2 Photocatalytic direct DCC between benzene and THF with
various catalystsa
the blended catalyst consisting of TiO2 photocatalyst and the sup-
ported Pd–Au bimetallic catalystsa
Entry
Photocatalyst
Al2O3 supported catalyst
1a (mmol)
% S
1
2
3
4
5
6c
TiO2
TiO2
TiO2
TiO2
TiO2
None
Pd(2.5)Au(0.5)
Pd(2.0)Au(1.0)
Pd(1.5)Au(1.5)
Pd(0.5)Au(2.5)
Au(3.0)
13.1
15.2
5.8
2.2
0.0
>99
>99
>99
>99
n.a.b
n.a.
Entry
Catalyst
1a (mmol)
% S
Pd(2.0)Au(1.0)
0.0
1
2
3
Pd(3.0)/TiO2
TiO2
Pd(2.0)Au(1.0)/TiO2
TiO2 + Pd(3.0)/Al2O3
Pd(3.0)/Al2O3
8.9
0.0
1.4
11.9
0.0
>99
n.a.b
>99
>99
n.a.
a
All reaction conditions and abbreviations were the same as those
b
c
described in Table 1. Not applicable. The reaction was done
without the TiO2 sample.
4
5c
a
Reaction conditions: 2 mL (22.4 mmol) of benzene, 2 mL (24 mmol) of
THF, 50 mg of photocatalyst or 100 mg of the blended catalyst (50 mg of
each the photocatalyst and the metal catalyst) was used, the reaction
time was 1 h. The amount of 1a was approximately determined from
the calibration curve of 4-(tetrahydrofuran-2-yl)benzonitrile. Other
conditions and selectivity calculations are mentioned in the
out the possibility of plasmonic photocatalysis by the Pd metal
nanoparticles as reported in some studies,15 and conrming
that the coexistent TiO2 photocatalyst is essential. These results
suggest that the formation of 1a is a hybrid of TiO2 photo-
catalysis and Pd metal catalysis. This revelation means that we
can modify the property of the two catalytic components inde-
pendently to improve the product yield. This time we have
focused on the modication of the Pd catalyst part, and exam-
ined the bimetallic Pd(x)Au(y)/Al2O3 catalysts.
b
c
Experimental section. Not applicable. The reaction was done
without the TiO2 sample.
To begin with, the DCC of benzene (1) and THF was exam-
ined with different catalysts (Table 1). The reaction with the
Pd(3.0)/TiO2 sample gave 2-phenyltetrahydrofuran (1a) as the
only DCC product (Table 1 entry 1), along with small amounts of
side products like octahydro-2,20-bifuran, a homocoupling
product of THF, and some unknown products (amounts not
shown). The gas phase contained hydrogen (Fig. S1†), but, due
to these side reactions, the amount of hydrogen and the DCC
products was not balanced. Hence, the amount of hydrogen
couldn't be determined precisely. The reaction done with
a pristine TiO2 sample did not produce 1a (Table 1, entry 2),
indicating that Pd loading was necessary for the DCC reaction.
Attempts were made to modify the Pd metal part by simul-
taneous loading Au on the TiO2 photocatalyst so as to make
bimetallic Pd–Au nanoparticles (Pd(2.0)Au(1.0)/TiO2). However,
this sample gave much lower yield of 1a (Table 1, entry 3) than
the monometallic Pd(3.0)/TiO2 sample (Table 1, entry 1). Since
this modication was not helpful for the enhancement of
catalytic activity, we decided to separate the metal part from the
TiO2 part and designed a blended catalyst consisting of a TiO2
photocatalyst and an Al2O3 supported metal catalyst.
At the beginning, the reaction was done with a blended
catalyst consisting of a pristine TiO2 sample and a mono-
metallic Pd(3.0)/Al2O3 sample (Table 1, entry 4). The reaction
gave a much larger amount of 1a and a slightly smaller amount
of hydrogen (not shown) than that with the Pd(3.0)/TiO2 sample
(Table 1, entries 1 and 4). This means that the Pd nanoparticles
can contribute to the formation of 1a even deposited on a photo-
inactive support like Al2O3 instead of TiO2, or rather work more
efficiently on Al2O3. This result showed that, in this reaction
system, the Pd nanoparticles function not just as an electron
receiver on the TiO2 photocatalyst but also as a catalyst. More-
over, the reaction done with the Pd(3.0)/Al2O3 sample alone,
under light irradiation did not yield 1a (Table 1, entry 5), ruling
Photocatalytic direct DCC between benzene and THF with
a TiO2 photocatalyst blended with Pd(x)Au(y)/Al2O3 catalysts
Table 2 shows the results of the DCC between benzene and THF
with blended catalysts consisting of the TiO2 photocatalyst and
various bimetallic Pd(x)Au(y)/Al2O3 catalysts. When compared
to the result with the monometallic Pd(3.0)/Al2O3 catalyst (Table
1, entry 4), an introduction of a small amount of Au to Pd
increased the yield of 1a (Table 2, entries 1 and 2). However,
a further increase in the Au content had negative effect on the
yield (Table 2, entries 3 and 4) and the monometallic Au(3.0)/
Al2O3 catalyst was inactive for the formation of 1a (Table 2, entry
5). These results indicate that Pd was the catalytically active
component of these bimetallic samples. Also, an optimum ratio
of Pd and Au is needed for the high activity of the bimetallic
samples. This suggests that the increase of the Au content
would withdraw the electron density of the active Pd atoms to
enhance the reaction, and then the decreasing number of the
active Pd sites would decrease the activity. Furthermore, the
bimetallic Pd(2.0)Au(1.0)/Al2O3 catalyst did not yield 1a or H2
gas in the absence of the TiO2 photocatalyst (Table 2, entry 6),
conrming that they could not catalyse the DCC reaction alone
and required TiO2 photocatalysis. This result also ruled out the
possibility of plasmonic catalysis by the bimetallic samples for
the production of 1a and H2.
Scope of arene molecule for the photocatalytic direct DCC
with THF
The scope of the substituted arenes was examined for the
photocatalytic direct DCC reaction with THF by using different
catalysts. The results are shown in Tables S1–S4 in the ESI.†
Fig. 7 shows the representative results of photocatalytic direct
This journal is © The Royal Society of Chemistry 2018
RSC Adv., 2018, 8, 24021–24028 | 24025

View Article Online
RSC Advances
Paper
The reaction between benzonitrile (3) and THF also gave
a mixture of three DCC products (collectively shown as 3a),
along with benzamide (3b) (Table S2†). The blended catalyst
with the bimetallic catalyst again exhibited the highest activity
for the formation of the DCC products (Table S2,† entry 5),
although the selectivity was lower than the photocatalyst alone
due to the increased production of 3b (Table S2,† entry 3). It is
notable that, however, the improved reaction condition resulted
in almost complete DCC selectivity (% S > 99, Fig. 7, 3a).
The photocatalytic direct DCC reaction of toluene (4) and
aniline (5) with THF also proceeded (Tables S3 and S4†). Both
reactions gave a mixture of the corresponding DCC product
isomers, 4a and 5a, along with some side products. Photo-
catalytic oxidation of toluene (4) by adsorbed water to benzal-
dehyde (4b) and benzyl alcohol (4c) was a competitive reaction
to the DCC to give 4a. The blended catalyst with the bimetallic
catalyst drastically improved the selectivity to the DCC products
(Table S3,† entry 5). Under the optimised conditions shown in
Fig. 7, 90% selectivity to the DCC products 4a was obtained for
toluene. The reaction between aniline (5) and THF, on the
contrary, yielded DCC products with poor selectivity. Due to the
high reactivity of the NH2 group of aniline (5), the homocou-
pling of aniline to give di(phenyl)diazine (5b) was the main
reaction rather than the DCC between aniline and THF (Table
S4†).17 NH2 group would be readily oxidised by the photo-
generated holes on TiO2, which is suggested from the fact that
ammonia can be easily oxidised to become amide radicals.18
This resulted in the very low selectivity to the DCC products, 5a,
even under the optimised conditions (Fig. 7).
Table 3 shows the comparison of the total yield (%) of the
DCC products (mmol) and selectivity (%, in the parenthesis) for
the reaction using above-mentioned substrates (1–5). For the
arene without any substituent (1) and the ones with electron
withdrawing substituents like CHO (2) and CN (3), the blended
catalyst consisting of the TiO2 photocatalyst and the bimetallic
Pd(2.0)Au(1.0)/Al2O3 catalyst exhibited a higher activity to form
the corresponding DCC products than the metal loaded TiO2
photocatalysts (Pd(3.0)/TiO2 and Pd(2.0)Au(1.0)/TiO2) and the
blended catalyst with the monometallic catalyst (TiO2 + Pd(3.0)/
Al2O3) (Table 3, entries 1–3). So, the new blended catalyst, with
the bimetallic catalyst, developed in this study is the most
selective and active catalyst for the photocatalytic direct DCC
reactions between these arenes and THF. Also, the lower reac-
tivity of the arenes with electron releasing substituents (4 and 5)
(Table 3, entries 4 and 5) indicates that an electron decient
arene is more reactive towards the cross-coupling reaction than
an electron rich arene. These results suggest that the photo-
generated THF radical species is nucleophilic, so that it would
prefer to react with an electron decient aromatic ring. This
property of THF radical is opposite to a photocatalytically
generated electrophilic OH radical observed in our previous
study.17a These results are consistent with the fact that the
electron decient Pd species in the Pd–Au bimetallic catalyst is
more active than pure Pd metal catalyst for the DCC reactions.
This electron decient Pd species can withdraw the electron
density from an adsorbed arene molecule to accelerate its
reaction with the nucleophilic THF radical.
Fig. 7 Photocatalytic direct DCC between various arenes and THF
with a blended catalyst consisting of a TiO2 photocatalyst and
a
bimetallic Pd(2.0)Au(1.0)/Al2O3 catalyst. The general reaction
conditions and % S calculation were the same as Table 1. The definition
and calculation of the yield (% Y) is mentioned in the Experimental
section. Briefly, 50 mL of an arene, 2 mL (24 mmol) of THF, and 25 mg
of each catalyst (totally 50 mg of the blended catalyst) was used for
12 h reaction. 50 mL of each arene (1–5) corresponds to 541, 490, 485,
470, and 548 mmol, respectively.
DCC between different arenes (1–5) and THF with the blended
catalyst consisting of a TiO2 photocatalyst and an Al2O3 sup-
ported Pd–Au bimetallic catalyst. Substituted benzenes with
electron withdrawing (2 and 3) and electron releasing (4 and 5)
substituents successfully underwent the DCC reaction with
THF. Benzaldehyde (2) and a benzonitrile (3) noticeably gave
appreciable yields of the DCC products with high selectivity.
The details such as the regioselectivity of the DCC products have
not been fully claried yet since the syntheses of the reference
products are not easy by the conventional methods. In other
words, these results show the importance of the further devel-
opment of these new catalyses for the photocatalytic DCC
reactions.
Benzaldehyde (2) and THF reacted to give a mixture of ortho,
meta, and para substituted products (collectively shown as 2a)
along with the oxidation and reduction products of 2, i.e., benzoic
acid (2b) and benzyl alcohol (2c), respectively, as undesired by-
products (Table S1†). As mentioned in the ESI,† benzoic acid
(2b) was the major product of this reaction, which resulted in
a low selectivity to 2a, the DCC products. The autoxidation of
benzaldehyde (2) to benzoic acid (2a) is well known and the
presence of adsorbed water on the TiO2 photocatalyst might
promote this reaction.16 Although both the pristine TiO2 and
Pd(3.0)/TiO2 samples showed low selectivity to 2a (Table S1,†
entries 1 and 2), the blended catalyst consisting of the TiO2
photocatalyst and the bimetallic Pd(2.0)Au(1.0)/Al2O3 catalyst
slightly increased the formation of the DCC products and sup-
pressed the formation of the side products, which improved the
selectivity to 2a (Table S1,† entry 5). Change in the reaction
conditions resulted in a high selectivity of 75% (Fig. 7, 2a).
24026 | RSC Adv., 2018, 8, 24021–24028
This journal is © The Royal Society of Chemistry 2018

View Article Online
Paper
RSC Advances
Table 3 Comparison of the results of the photocatalytic direct DCC between various arenes and THF with the photocatalysts and the blended
catalystsa
Yield of DCC products, Y (%)b
Entry
Arene
DCC products
Pd(3.0)/TiO2
Pd(2.0)Au(1.0)/TiO2
TiO2 + Pd(3.0)/Al2O3
TiO2 + Pd(2.0)Au(1.0)/Al2O3
1c
2d
3e
4f
1
2
3
4
5
1a
2a
3a
4a
5a
0.04(>99)
1.24(36)
0.10(80)
0.05(64)
0.02(19)
0.006(>99)
0.42(32)
0.08(82)
0.02(74)
0.003(6)
0.05(>99)
1.06(43)
0.06(68)
0.01(50)
0.02(14)
0.06(>99)
1.73(46)
0.27(57)
0.03(90)
0.02(23)
5g
a
2 mL (19–24 mmol) of arene, 2 mL (24 mmol) of THF, and 50 mg of the photocatalyst or 100 mg of the blended catalyst were used, other conditions
were the same as those described in Table 1. b The total yield, Y (%), of DCC products. The value in parenthesis shows % S. The denition of % Y and
% S is mentioned in the Experimental section. c 2 mL (22.4 mmol) of benzene (1). d 2 mL (19.9 mmol) of benzaldehyde (2). e 2 mL (19.4 mmol) of
benzonitrile (3). 2 mL (19.01 mmol) of toluene (4). g 2 mL (21.9 mmol) of aniline (5).
f
As mentioned before, electron decient arenes exhibited
the highest activity for this reaction. So, the nature of Pd
catalysis in this photocatalytic reaction is proposed to be the
activation of the aromatic ring by withdrawing its electron
density, which facilitates its reaction with a photogenerated
THF radical. The presence of an electron decient Pd species,
formed by the introduction of Au, would promote this
reaction.
According to this mechanism for the DCC, it is also proposed
that the THF radical species, photocatalytically formed on the
TiO2 surface, can migrate to the separated metal catalyst. This
suggests that the life of the THF radical species is not so short in
these condition. Additionally, the coupling of two THF radical
species can give the homocoupling products like octahydro-2,20-
bifuran.3e But, as results show, this was a very minor reaction,
meaning that the THF radical would selectively react with the
activated arene molecule to give the DCC products.
Proposed reaction mechanism for the photocatalytic direct
DCC between arenes and THF
Based on the mechanism proposed in the previous study3c and
the above results, the following reaction mechanism can be
proposed for the photocatalytic direct DCC reaction between
arenes and THF with the blended catalyst, which is a hybrid of
TiO2 photocatalysis and metal catalysis by Pd–Au bimetallic
nanoparticles (Fig. 8). The photo-excitation of TiO2 generates
electrons and holes in the conduction band and the valence
band, respectively (i). The photogenerated hole oxidizes the
THF molecule to form a radical species and proton (ii). The
proton is reduced to hydrogen radical by the photoexcited
electron on TiO2 (iii). The THF radical can migrate to the
bimetallic catalyst and attack to the carbon in the aromatic ring
of the activated arene molecule, adsorbed on the Pd–Au bime-
tallic catalyst (iv), which gives the DCC product with the elimi-
nation of a hydrogen radical (v). The two hydrogen radicals can
combine to give a hydrogen molecule either on the TiO2 surface
or on the bimetallic catalyst (vi).
Conclusions
In the present work, we found that the photocatalytic direct
DCC between various arenes and THF was a hybrid of TiO2
photocatalysis and Pd metal catalysis. The Pd metal catalyst can
work even if they were on a photo-inactive support like Al2O3.
Based on these ndings, we developed a novel blended catalyst
consisting of a TiO2 photocatalyst and an Al2O3 supported Pd–
Au bimetallic catalyst that exhibited higher activity and selec-
tivity for the photocatalytic direct DCC between arenes and THF
than the conventional metal loaded TiO2 photocatalyst alone.
Although the reported yields are still low, this work has
provided some new insights about the mechanisms of the
photocatalytic organic synthesis reactions and opened new
prospects for the catalyst design. This blended catalyst has wide
exibility for modication, like, the photocatalyst part by using
Fig.
8 Proposed mechanism for the photocatalytic direct DCC
between arenes and THF by the blended catalyst consisting of a TiO2
photocatalyst and a supported Pd–Au bimetallic catalyst.
This journal is © The Royal Society of Chemistry 2018
RSC Adv., 2018, 8, 24021–24028 | 24027

View Article Online
RSC Advances
Paper
a co-catalyst or by changing the photocatalyst itself, the metal
catalyst part by changing the loading amount, dispersion, or the
metal component, and also the ratio of the two catalytic
components. Thus, the blended catalyst can be further
improved and become available for other photocatalytic direct
DCC reactions and also other reaction systems.
5 (a) B. Ohtani, K. Iwai, S. Nishimoto and S. Sato, J. Phys. Chem.
B, 1997, 17, 3349; (b) J. M. Herrmann, Top. Catal., 2006, 39, 3;
(c) Y. Shiraishi, Y. Sugano, S. Tanaka and T. Hirai, Angew.
Chem., Int. Ed., 2012, 49, 1656.
6 (a) J. G. Yu, H. G. Yu, B. Cheng, X. J. Zhao, J. C. Yu and
W. K. Ho, J. Phys. Chem. B, 2003, 107, 13871; (b) J. Zhang,
Q. Xu, Z. Feng, M. Li and C. Li, Angew. Chem., Int. Ed.,
2008, 47, 1766; (c) C. H. Lin, J. H. Chao, C. H. Liu,
J. C. Chang and F. C. Wang, Langmuir, 2008, 24, 9970.
7 (a) W. Zhang, Y. Wang, Z. Wang, Z. Zhong and R. Xu, Chem.
Commun., 2010, 46, 7631; (b) H. Cheng, X. J. Lv, S. Cao,
Z. Y. Zhao, Y. Chen and W. F. Fu, Sci. Rep., 2016, 6, 19846;
(c) X. Xu, G. Liu, C. Randorn and J. T. S. Irvine, Int. J.
Hydrogen Energy, 2011, 36, 13501.
8 (a) K. Sayama, K. Mukasa, R. Abe, Y. Abe and H. Arakawa,
Chem. Commun., 2001, 2416; (b) Y. Sasaki, H. Nemoto,
K. Saito and A. Kudo, J. Phys. Chem. C, 2009, 113, 17536; (c)
K. Maeda, ACS Catal., 2013, 3, 1486; (d) E. Wada, A. Tyagi,
A. Yamamoto and H. Yoshida, Photochem. Photobiol. Sci.,
2017, 16, 1744.
Conflicts of interest
There are no conicts to declare.
Acknowledgements
We are grateful to Prof. Ken-ichi Fujita and his laboratory
members at Kyoto University for their help to synthesize the
specic compound. We thank Dr Tsutomu Kiyomura, Kyoto
University, for TEM measurements which was supported by
Kyoto University Nano Technology Hub in "Nanotechnology
Platform Project" sponsored by the Ministry of Education,
Culture, Sports, Science and Technology (MEXT), Japan. We
acknowledge the kind help of the laboratory members for XAFS
measurements (Spring-8, Hyogo, Japan, proposal no.:
2017A1477, 2016A1170). A. Tyagi would like to thank JICA for
providing the nancial assistance under IITH–JICA ‘Friendship’
project.
9 K. Taniguchi, X. Jin, K. Yamaguchi and N. Mizuno, Chem.
Commun., 2015, 51, 14969.
10 B. Ravel and M. Newville, J. Synchrotron Radiat., 2005, 12,
537.
11 D. R. Heitz, J. C. Tellis and G. A. Molander, J. Am. Chem. Soc.,
2016, 138, 12715.
´
12 L. Samain, A. Jaworski, M. Eden, D. M. Ladd, D. K. Seo,
Notes and references
¨
F. J. G. Gracia and U. Haussermann, J. Solid State Chem.,
1 (a) C. Li, Acc. Chem. Res., 2009, 42, 335; (b) C. S. Yeung and
V. M. Dong, Chem. Rev., 2011, 111, 1215.
2014, 217, 1.
13 A. F. Lee, C. J. Baddeley, C. Hardcare, R. M. Ormerod and
R. M. Lambert, J. Phys. Chem., 1995, 99, 6096.
14 H. Miura, K. Endo, R. Ogawa and T. Shishido, ACS Catal.,
2017, 7, 1543.
2 (a) J. J. H. B. Sattler, J. R. Martinez, E. S. Jimenez and
B. M. Weckhuysen, Chem. Rev., 2014, 114, 10613; (b)
M. R. Nabid, Y. Bide and B. Etemadi, New J. Chem., 2017,
41, 10773; (c) K. A. Goulas, Y. Song, G. R. Johnson,
J. P. Chen, A. A. Gokhale, L. C. Grabow and F. D. Toste,
Catal. Sci. Technol., 2018, 8, 314.
3 (a) C. Vila and M. Rueping, Green Chem., 2013, 15, 2056; (b)
H. Yoshida, Y. Fujimura, H. Yuzawa, J. Kumagai and
T. Yoshida, Chem. Commun., 2013, 49, 3793; (c) A. Tyagi,
T. Matsumoto, T. Kato and H. Yoshida, Catal. Sci. Technol.,
2016, 6, 4577; (d) E. Wada, T. Takeuchi, Y. Fujimura,
A. Tyagi and H. Yoshida, Catal. Sci. Technol., 2017, 7, 2457;
(e) A. Tyagi, A. Yamamoto, T. Kato and H. Yoshida, Catal.
Sci. Technol., 2017, 7, 2616.
15 (a) T. T. Trinh, R. Sato, M. Sakamoto, Y. Fujiyoshi,
M. Haruta, H. Kurata and T. Teranishi, Nanoscale, 2015, 7,
12435; (b) W. Yang, Y. Xiong, L. Zou, Z. Zou, D. Li, Q. Mi,
Y. Wang and H. Yang, Nanoscale Res. Lett., 2016, 11, 283.
16 (a) M. Conte, H. Miyamura, S. Kobayashi and V. Chechik,
Chem. Commun., 2010, 46, 145; (b) M. Sankar, E. Nowicka,
E. Carter, D. M. Murphy, D. W. Knight, D. Bethell and
G. J. Hutchings, Nat. Commun., 2014, 5, 3332.
17 (a) H. Yuzawa, M. Aoki, K. Otake, T. Hattori, H. Itoh and
H. Yoshida, J. Phys. Chem. C, 2012, 116, 25376; (b)
H. Yuzawa, J. Kumagai and H. Yoshida, J. Phys. Chem. C,
2013, 117, 11047.
4 (a) A. J. Bard, Science, 1980, 207, 139; (b) C. M. Wang,
A. Heller and H. Gerischer, J. Am. Chem. Soc., 1992, 114,
5230; (c) S. Vaidyanathan, W. E. Eduardo and P. V. Kamat,
J. Am. Chem. Soc., 2004, 126, 4943; (d) J. Yang, D. Wang,
H. Han and C. Li, Acc. Chem. Res., 2013, 46, 1900.
18 H. Yuzawa, T. Mori, H. Itoh and H. Yoshida, J. Phys. Chem. C,
2012, 116, 4126.
24028 | RSC Adv., 2018, 8, 24021–24028
This journal is © The Royal Society of Chemistry 2018