Modulating the catalytic behavior of non-noble metal nanoparticles by inter-particle interaction for chemoselective hydrogenation of nitroarenes into corresponding azoxy or azo compounds

Article abstract of DOI:10.1016/j.jcat.2018.11.011

Aromatic azoxy compounds have wide applications and they can be prepared by stoichiometric or catalytic reactions with H₂O₂ or N₂H₄ starting from anilines or nitroarenes. In this work, we will present the direct chemoselective hydrogenation of nitroarenes with H₂ to give aromatic azoxy compounds under base-free mild conditions, with a bifunctional catalytic system formed by Ni nanoparticles covered by a few layers of carbon (Ni@C NPs) and CeO₂ nanoparticles. The catalytic performance of Ni@C-CeO₂ catalyst surpasses the state-of-art Au/CeO₂ catalyst for the direct production of azoxybenzene from nitrobenzene. By means of kinetic and spectroscopic results, a bifunctional mechanism is proposed in which, the hydrogenation of nitrobenzene can be stopped at the formation of azoxybenzene with >95% conversion and >93% selectivity, or can be further driven to the formation of azobenzene with >85% selectivity. By making a bifunctional catalyst with a non-noble metal, one can achieve chemoselective hydrogenation of nitroarenes not only to anilines, but also to corresponding azoxy and azo compounds.

Full text of DOI:10.1016/j.jcat.2018.11.011

Journal of Catalysis 369 (2019) 312–323
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Modulating the catalytic behavior of non-noble metal nanoparticles by
inter-particle interaction for chemoselective hydrogenation of
nitroarenes into corresponding azoxy or azo compounds
⇑
Lichen Liu, Patricia Concepción, Avelino Corma
Instituto de Tecnología Química, Universitat Politècnica de València–Consejo Superior de Investigaciones Científicas (UPV-CSIC), Av. 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:
Aromatic azoxy compounds have wide applications and they can be prepared by stoichiometric or cat-
alytic reactions with H2O2 or N2H4 starting from anilines or nitroarenes. In this work, we will present
the direct chemoselective hydrogenation of nitroarenes with H2 to give aromatic azoxy compounds under
base-free mild conditions, with a bifunctional catalytic system formed by Ni nanoparticles covered by a
few layers of carbon (Ni@C NPs) and CeO2 nanoparticles. The catalytic performance of Ni@C-CeO2 catalyst
surpasses the state-of-art Au/CeO2 catalyst for the direct production of azoxybenzene from nitrobenzene.
By means of kinetic and spectroscopic results, a bifunctional mechanism is proposed in which, the hydro-
genation of nitrobenzene can be stopped at the formation of azoxybenzene with >95% conversion and
>93% selectivity, or can be further driven to the formation of azobenzene with >85% selectivity. By making
a bifunctional catalyst with a non-noble metal, one can achieve chemoselective hydrogenation of
nitroarenes not only to anilines, but also to corresponding azoxy and azo compounds.
Received 6 September 2018
Revised 31 October 2018
Accepted 8 November 2018
Keywords:
Ni@C
CeO2
Nitroarenes
Selective hydrogenation
Azoxy
Azo
Ó 2018 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
1. Introduction
some catalytic processes have been developed and aromatic azoxy
compounds can be obtained by the oxidative coupling of anilines
Selecting the desired reaction pathway among several possible
is a key objective for heterogeneous catalysis [1–3]. It has been
demonstrated in the literature that, by tuning the adsorption
properties of molecules on catalyst surface and the formation of
reaction intermediates, it is possible to achieve unique chemose-
lectivities by solid catalysts [4–7]. Moreover, by preparation of
multifunctional catalysts, transformations which require multiple
steps can be achieved in a one-pot process, leading to process
intensification that can reduce the energy consumption and
improve the efficiency of the global process [8–10].
Aromatic azoxy compounds are important intermediate com-
pounds with wide application as chemical stabilizers, polymeriza-
tion inhibitors, dyes, pigments, as well as precursors for materials
used in electronic displays and pharmaceuticals [11,12]. Moreover,
aromatic azoxy compounds can be further transformed into
hydroxyazo compounds through the Wallach rearrangement [13].
Conventionally, aromatic azoxy compounds are prepared by stoi-
chiometric reactions (such as the condensation between hydroxy-
lamine and nitrosobenzene), which is less atom-economic and may
produce hazardous by-products [14]. Therefore, in the last decade,
using metal oxides as catalyst and H2O2 as the oxidant (Fig. 1a)
[15,16]. Aromatic azoxy compounds can also be obtained through
the reduction of nitroarenes with hydrazine (see Fig. 1b) [17,18].
Compared with previous processes, these heterogeneous catalytic
systems are more efficient and sustainable, but the use of reac-
tants, such as H2O2 and hydrazine, is still required.
Considering the reaction pathways for hydrogenation of
nitroarenes into different products (see Fig. 2), it should be possible
to achieve the formation of azoxy and azo products if the reaction
goes through the condensation route (shown in Fig. 2). Indeed, in
the presence of a base, the intermediates formed during the hydro-
genation of nitroarenes can be transformed into azoxy and azo
products through the condensation route, as has been reported
for the hydrogenation of nitrobenzene to azoxybenzene in the
presence of n-butylamine as base and using Rh or Ir as metal cat-
alysts (see Fig. 1c) [19].
In recent years, the development of non-noble metal catalysts
as substitutes for noble metal catalysts is emerging in the field of
heterogeneous catalysis [20]. For chemoselective hydrogenation
reactions, it has been reported that Co-, Fe- and Ni-based catalysts
show promising catalytic performances for hydrogenation of
nitroarenes to anilines [21–23]. Recently, by studying the reaction
mechanism and identification of the active sites, non-noble
⇑
Corresponding author.
E-mail address: acorma@itq.upv.es (A. Corma).
https://doi.org/10.1016/j.jcat.2018.11.011
0021-9517/Ó 2018 The Authors. Published by Elsevier Inc.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

L. Liu et al. / Journal of Catalysis 369 (2019) 312–323
313
Fig. 1. Catalytic routes for the preparation of aromatic azoxy compounds by different approaches, including selective oxidation of anilines with H2O2 (a), reduction of
nitroarenes with N2H4 (b), hydrogenation of nitroarenes with H2 by noble metal catalysts in the presence of base (c) and direct hydrogenation of nitroarenes with H2 by non-
noble metal catalysts under base-free conditions (d).
and how they can interplay. It is of most interest that, based on the
nanoparticulate Ni@C catalyst, it is possible to chemoselectively
direct the hydrogenation of nitroarenes into corresponding azoxy,
azo or aniline products, and we are now able to chemoselectively
achieve any of the intermediates involved in the global process
of nitroarenes reduction.
2. Experiments
Fig. 2. Catalytic hydrogenation of nitroarenes into different products via different
routes.
TiO2 nanoparticles (from Sigma-Aldrich, product No. 718467-
100G), Al2O3 nanoparticles (from NanoScale NanoActiveÒ, with a
surface area of ꢀ300 m2/g), SiO2 nanoparticles (from Sigma-
Aldrich, product No. 236772-100G with
a surface area of
monometallic and bimetallic NPs have been prepared and show
better catalytic performance than the state-of-art heterogeneous
Au catalysts under the same reaction conditions for chemoselec-
tive hydrogenation of nitroarenes to the corresponding anilines
[24,25]. It has also been revealed that, the surface structure of
the non-noble metal catalysts has significant impact on the selec-
tivity for hydrogenation of nitroarenes [24,26]. Nitrobenzene tends
to be hydrogenated directly to aniline on a reduced Ni surface,
while nitrosobenzene can be produced as intermediate on a par-
tially reduced Ni surface.
In this work, we show that by combining Ni nanoparticles cov-
ered by a few layers of carbon (Ni@C NPs) and CeO2 NPs, a bifunc-
tional catalytic system can be prepared for the direct
chemoselective hydrogenation of nitroarenes into corresponding
azoxy compounds with high yields under base-free mild condi-
tions. Based on kinetic and spectroscopic studies, we discuss the
role of the two components in the Ni@C-CeO2 bifunctional catalyst
ꢀ480 m2/g), ZrO2 nanoparticles (from Sigma-Aldrich, product No.
544760-5G with a surface area of ꢀ25 m2/g) and CeO2 nanoparti-
cles (from Rhodia with a surface area of ꢀ120 m2/g) were used
directly as received without any treatments. Au/CeO2 catalyst with
1 wt% of Au was obtained from Johnson-Matthey and was directly
used for the catalytic test without any treatments.
Hydrotalcite (Mg/Al = 4) nanoparticles (with a surface area of
ꢀ80 m2/g) were prepared by a precipitation method. 8 mmol of
Mg(NO3)2Á6H2O and 20 mmol of Al(NO3)3Á9H2O were dissolved in
100 mL of H2O, resulting the formation of solution A. 40 mmol of
Na2CO3 and 200 mmol of NaOH was dissolved in 100 mL of H2O,
resulting in the formation of solution B. Solution A was dropped
into solution B slowly under vigorous stirring at room temperature.
After ꢀ1 h, the mixing of solution A and B was completed and the
suspension was aged at 65 °C for ꢀ18 h. The while precipitate was
obtained by filtration and washed with H2O. Finally, the solid pro-
duct was dry in an oven at 100 °C.

314
L. Liu et al. / Journal of Catalysis 369 (2019) 312–323
2.1. Synthesis of the catalysts
heating the suspension at 120 °C. After the removal of the solvent,
the solid sample was calcined in flow air at 550 °C for 3 h with a
ramp rate of 5 °C/min from room temperature to 550 °C. After
the calcination in air, the sample was cooled to room temperature
and a subsequent reduction treatment by H2 was carried out at
450 °C for 3 h with a ramp rate of 5 °C/min from room temperature
to 450 °C. After the reduction treatment, the sample was cooled
down to room temperature in H2 flow and stored in a glass vial
in ambient environment.
2.1.1. Synthesis of Co@C and Ni@C nanoparticles
Co@C and Ni@C nanoparticles were synthesized by a carbon-
coating method, which was described in our recent work [25].
CeO2@C nanoparticles were prepared by a similar carbon-
coating approach. 1.0 g of CeO2 nanoparticles, 720 mg of glucose
and 20 mL distilled water was mixed to a suspension with the help
of ultrasonication. The suspension was then transferred to an auto-
clave and placed in an oven at 175 °C for 18 h. After the hydrother-
mal treatment, the solid product was collected by filtration and
washed with water and acetone. The dried solid was then annealed
in flow N2 at 600 °C for 2 h with a ramp rate of 10 °C/min from
room temperature to 600 °C. After the high-temperature anneal-
ing, the sample was cooled down to room temperature in N2 flow
and stored in a glass vial in ambient environment.
2.2. Characterizations
Powder X-ray diffraction (XRD) was performed with a HTPhilips
X’Pert MPD diffractometer equipped with a PW3050 goniometer
using Cu Ka radiation and a multisampling handler.
Raman spectra were recorded at ambient temperature with a
785 nm HPNIR excitation laser on a Renishaw Raman Spectrometer
(‘‘Reflex”) equipped with an Olympus microscope and a CCD detec-
tor. The laser power on the sample was 15mW and a total of 20
acquisitions were taken for each spectra.
Samples for electron microscopy studies were prepared by
dropping the suspension of solid sample in CH2Cl2 directly onto
holey-carbon coated copper grids. All the measurements were per-
formed in a JEOL 2100F microscope operating at 200 kV both in
transmission (TEM) and scanning-transmission modes (STEM).
STEM images were obtained using a High Angle Annular Dark Field
detector (HAADF), which allows Z-contrast imaging.
The amount of exposed surface Ni atoms in the Ni@C catalyst
was determined by H2 chemisorption at 100 °C in an ASAP 2010C
Micromeritics equipment by extrapolating the total gas uptakes
in the adsorption isotherms at zero pressure. Prior to the measure-
ments the sample was reduced under flowing pure H2 at 400 °C.
The percentage of exposed Ni atoms in the Ni@C NPs was esti-
mated from the total amount of chemisorbed H2 assuming an
adsorption stoichiometry H2/Ni of 1.
The FTIR spectra were collected with a Bruker Vertex 70 spec-
trometer equipped with a DTGS detector (4 cmÀ1 resolution, 32
scans). An IR cell allowing in situ treatments in controlled atmo-
sphere and temperature from 25 to 500 °C has been connected to
a vacuum system with gas dosing facility. Self-supporting pellets
(ca. 10 mg cmÀ2) were prepared from the sample powders and
treated at 200 °C in hydrogen flow (20 mL minÀ1) for 2 h followed
by evacuation at 10À4 mbar at 250 °C for 1 h. In the case of the
CeO2 sample only activation under vacuum at 120 °C have been
performed. After activation the samples were cooled down to
25 °C under dynamic vacuum conditions followed by nitrobenzene
dosing at 1.5 mbar. Once the physically absorbed and/or gas phase
nitrobenzene was removed in evacuation at 25 °C, H2 was co-
adsorbed at pressures between 0.4 mbar and 40 mbar. The IR spec-
tra were recorded after 35 min at a specific reaction temperature
(from 25 to 150 °C). Additional experiments were performed at dif-
ferent reaction times (from 20 min to 180 min) at each tempera-
ture. It should be noted that, the IR experimental procedure used
in this work is not operando. However, the above methodology
allows us to obtain information about the intrinsic reactivity of
surface sites in the solid catalyst in the presence of the reactant
molecule. With these information, it enables us to follow the reac-
tion pathways for the hydrogenation of nitrobenzene and
nitrosobenzene on CeO2 and Ni@C-CeO2 catalysts.
2.1.2. Synthesis of Cu@C and Fe@C nanoparticles
We have also prepared Cu@C and Fe@C nanoparticles by the
above-mentioned carbon-coating process. However, that method
is not applicable to Cu and Fe. Only bulk Cu and Fe (with particle
size larger than 1 lm) were obtained. Therefore, Cu@C and Fe@C
nanoparticles were synthesized by reducing metal-EDTA complex,
as described in our recent work [24]. More specifically, the Cu-
EDTA complex were prepared through a hydrothermal process.
First, 6.98 g Cu(NO3)2, 4.47 g Na2EDTA and 0.96 g NaOH are dis-
solved in 20 mL H2O. Then, 10 mL methanol was added to the
mixed aqueous solution temperature under stirring at room tem-
perature. After the formation of a homogeneous solution, 23 mL
of the purple solution was transferred into a 35 mL stainless steel
autoclave followed by static hydrothermal processing at 200 °C
for 24 h. After cooling to room temperature, the generated precip-
itates were filtered and washed with deionized water and acetone
several times followed by drying at 100 °C in air for 16 h. The
obtained complex was denoted as Cu-EDTA. Then, Cu@C NPs were
prepared by reduction of Cu-EDTA in H2 (50 mL/min) at 450 °C for
2 h with a ramp rate of 10 °C/min from room temperature to
450 °C. After the H2 reduction process at 450 °C, the sample was
cooled down to room temperature in H2 atmosphere. Then the
black solid product was stored in
environment.
a glass vial in ambient
Fe-EDTA complex was prepared by the same method, using Fe
(NO3)3Á9H2O as the precursor.
2.1.3. Synthesis of Ni-CeO2-C nanocomposites
The Ni-CeO2-C nanocomposites can be prepared by the carbon-
coating process in the presence of CeO2 nanoparticles. 1.0 g CeO2
nanoparticles, 720 mg of glucose, 0.5 g of Ni precipitate (obtained
from the precipitation of Ni(Ac)2 and sodium carbonate in glycol)
was mixed with 20 mL water to form a homogeneous suspension
with the help of ultrasonication. The suspension was then trans-
ferred to an autoclave and placed in an oven at 175 °C for 18 h.
After the hydrothermal treatment, the solid product was collected
by filtration and washed with water and acetone. The dried solid
was then annealed in flow N2 at 600 °C for 2 h with a ramp rate
of 10 °C/min from room temperature to 600 °C. After the high-
temperature annealing, the sample was cooled down to room tem-
perature in N2 flow and stored in
environment.
a glass vial in ambient
X-ray photoelectron spectra of the catalysts were recorded with
a SPECS spectrometer equipped with a Phoibos 150MCD-9 multi-
2.1.4. Synthesis of Ni/CeO2 by wetness impregnation
Ni/CeO2 with 5.0 wt% of Ni supported on CeO2 was prepared by
traditional wetness impregnation method. 123 mg of Ni(NO3)2-
Á6H2O was dissolved in 20 mL of water and then 0.5 g of CeO2
nanoparticles was added to the solution. The suspension was kept
stirring at room temperature for 2 h before removing the water by
channel analyzer using non-monochromatic AlK
diation. Spectra were recorded using analyser pass energy of 30 eV,
an X-ray power of 100 W and under an operating pressure of 10À9
mbar. The sample was pre-reduced by H2 in a pre-treatment cham-
ber at 120 or 200 °C and then transferred to analysis chamber for
a (1486.6 eV) irra-
-

L. Liu et al. / Journal of Catalysis 369 (2019) 312–323
315
the XPS measurements. Peak intensities have been calculated after
nonlinear Shirley-type background subtraction and corrected by
the transmission function of the spectrometer. During data pro-
cessing of the XPS spectra, binding energy (BE) values were refer-
enced to C1s peak (284.5 eV) for the Ni@C samples and to the
Ce3d5/2 peak at 882.8 eV for the Ni@C-CeO2 samples. CasaXPS
software has been used for spectra treatment.
becomes the major product when conversion is >20%. From these
results, it can be speculated that the nitrosobenzene formed as a
primary product would be rapidly hydrogenated to aniline and
the formation of azoxybenzene would not be kinetically favorable
on this catalyst. However, when we reacted nitrosobenzene on
Ni@C NPs (see Fig. S4), azoxybenzene was obtained with high yield
and aniline only starts to form when the conversion of nitrosoben-
zene was >90%. Then, according to the reaction network in Fig. 2
and the catalytic results shown in Figs. S3 and S4, a catalyst in
which the rate of hydrogenation of nitrosobenzene to aniline is
decreased and nitrosobenzene is allowed to accumulate, may allow
to produce azoxybenzene (see summary in Fig. S5). In order to
achieve high selectivity for hydrogenation of nitrobenzene to
azoxybenzene, it would be critical to stabilize the nitrosobenzene
and phenylhydroxylamine intermediates, and to react them into
azoxybenzene.
Considering that the intermediates (nitrosobenzene and
phenylhydroxylamine) in the condensation reaction route (see
Fig. 2) can be stabilized and condensed in a basic environment,
we thought that the selectivity to the azoxybenzene product
should be improved in the presence of a solid catalyst with suitable
basicity. After screening a combination of Ni@C and metal oxides
with varying acid-base properties (see Table 1), it was found that
the combination of Ni@C and CeO2 NPs gives high selectivity to
azoxybenzene (94%) at full conversion of nitrobenzene, though
CeO2 alone in the absence of Ni@C NPs is not active for this reac-
tion. We have also noticed that, other metal oxides with different
acid-base properties such as Al2O3, TiO2, SiO2 or hydrotalcite
(Mg/Al = 4) give much lower selectivity to azoxybenzene.
2.3. Catalytic tests
The chemoselective hydrogenation of nitroarenes was per-
formed in batch reactors. The reactant, internal standard (hexado-
decane), solvent (toluene), powder catalyst as well as a magnetic
bar were added into the batch reactor. After the reactor was sealed,
air was purged out from the reactor by flushing two times with
10 bar of hydrogen. Then the autoclave was pressurized with H2
to the corresponding pressure. The stirring speed is kept at
1100 rpm and the size of the catalyst powder is below 0.05 mm
to avoid either external or internal diffusion limitation. Finally,
the batch reactor was heated to the target temperature. For the
kinetic studies, 50 lL of the mixture was taken out for GC analysis
at different reaction times. The products were identified by GC-MS.
In some cases, the catalyst was pre-reduced by H2 at an elevated
temperature (150 °C) for a certain time and then cooled down to
room temperature. Then nitrobenzene was injected into the batch
reactor and the temperature was raised to 120 °C again for the
hydrogenation reaction test.
3. Results and discussions
At this point, we assume that a bifunctional catalyst prepared
by supporting Ni on CeO2 NPs should work for the hydrogenation
of nitrobenzene to azoxybenzene. Following that assumption, a
supported Ni/CeO2 catalyst (with 5 wt% of Ni on CeO2) was pre-
pared by wet impregnation and it was then reduced by H2 at
450 °C and passivated by air at room temperature after the H2
reduction treatment. As it can be seen, the as-prepared Ni/CeO2
catalyst shows poor activity for hydrogenation of nitrobenzene,
which is probably caused by the re-oxidation of Ni when exposed
to air. Indeed, when the Ni/CeO2 catalyst was reduced in the reac-
tor before the catalytic test, the activity improved significantly
while selectivity was preserved, implying that metallic Ni is the
active component for hydrogenation of nitroarenes. Hou et al. have
studied the chemical states of Ni particles supported on CeO2 [18].
The XPS results show that, only part of the Ni can be reduced by H2
even after treatment at 500 °C. Therefore, at our current reaction
temperature (120 °C), only a small amount of Ni (<5%) in the sup-
ported Ni/CeO2 catalyst can be reduced (see the XPS results in
Fig. S6). The thin carbon layers in the Ni@C NPs can protect Ni from
deep oxidation by air and promote the partial reduction of Ni
under reaction conditions. Therefore, the Ni@C-CeO2 catalyst in
which Ni nanoparticles with a high reduction degree are mixed
with CeO2 is more active than the Ni/CeO2 in which only a small
part of Ni is reduced.
To figure out the reaction mechanism of Ni@C-CeO2 for direct
conversion of nitrobenzene to azoxybenzene, we have performed
the following experiments. Firstly, bare CeO2 is not active for
hydrogenation of nitrobenzene (see Fig. S7), inferring that the ini-
tial activation of nitrobenzene should be related with Ni@C NPs.
Afterwards, it is shown in Fig. S8 that nitrosobenzene can be selec-
tively hydrogenated into azoxybenzene by Ni@C-CeO2 with high
reaction rate. Furthermore, as shown in Figs. S9 and S10, only
hydrogenation of nitrosobenzene was observed during the com-
petitive hydrogenation tests with nitrobenzene and azoxybenzene.
It is found that nitrosobenzene can be directly and selectively
converted to azoxybenzene heating the nitrosobenzene in toluene
solvent (0.5 mol LÀ1) in the absence of solid catalyst with a
Non-noble transition metal nanoparticles covered by thin car-
bon layers (named as Co@C, Fe@C, Cu@C and Ni@C, respectively)
that could, in principle, catalyze the hydrogenation reaction, were
prepared according to our recent work (see the experimental
details). The morphologies of those as-prepared metal@C nanopar-
ticles have been characterized by TEM. As shown in Fig. 3, Ni NPs
show particle size ranging from 5 to 20 nm, with an average size
of ꢀ10 nm and those Ni NPs are covered by a few carbon layers
(less than 10 layers, mostly around 3–8 layers). Structural charac-
terizations (XRD and Raman) show that the as-prepared sample
contains metallic Ni NPs with disorder carbon surrounding them.
In the case of Co@C, the morphology is similar to Ni with a slightly
larger average particle size of ꢀ15 nm (see Fig. S1). For Cu@C and
Fe@C, the particle size is much larger than for Ni@C or Co@C (see
Fig. S2).
Firstly, we have studied the catalytic behavior of the several
types of non-noble metal NPs that in principle can work as cata-
lysts (Co@C, Ni@C, Fe@C and Cu@C NPs) for hydrogenation of
nitrobenzene under mild conditions (120 °C and 10 bar of H2). As
shown in Table 1, Fe@C and Cu@C NPs show poor activity for
hydrogenation of nitrobenzene while Co@C and Ni@C NPs are sen-
sibly more active. Indeed, the Ni@C sample achieves 42% conver-
sion after 150 min reaction time and gives aniline as the major
product (ꢀ90% selectivity) with azoxybenzene (ꢀ9% selectivity)
as byproduct. However, in the case of Co@C, very high selectivity
to aniline is observed, though conversion is lower.
Since a small amount of azoxybenzene (9.7% selectivity) and
azobenzene (3.2% selectivity) was obtained with Ni@C NPs, this
encourages us to study the kinetic profiles for hydrogenation of
nitrobenzene on this catalyst, to see the evolution of the different
products with reaction time. As shown in Fig. S3, nitrosobenzene
can be observed at the starting stage, which comes from a new
reaction pathway on Ni nanoparticles, being different to that
reported on noble metal catalysts (such as Au, Pt) [26]. The selec-
tivity to nitrosobenzene decreases with reaction time and aniline

316
L. Liu et al. / Journal of Catalysis 369 (2019) 312–323
Fig. 3. Structural characterization of Ni@C nanoparticles. (a and b) TEM and high-resolution TEM image. (c) XRD pattern of Ni@C nanoparticles, being the typical
diffraction pattern of metallic Ni (PDF code:96-210-0650), without peaks corresponding to NiO or other species. (d) Raman spectrum of Ni@C nanoparticles, showing the
presence of Raman bands corresponding to D and G bands. The ratio between these two bands suggests that, the carbon layers covered on Ni nanoparticles are disordered.
Table 1
Catalytic performance of various non-noble metal catalysts for hydrogenation of nitrobenzene.a
Entry
Sample
Time/min
Conversion/%
Selectivity to Azoxy (1b)
Selectivity to Azo (1c)
Selectivity to Aniline (1d)
1
2
3
4
5
6
7
8
Ni@C
Co@C
Cu@C
Fe@C
Ni@C+CeO2
CeO2
Ni/CeO2
Ni/CeO2-in situ reduced
Ni@C+Al2O3
Ni@C + TiO2
Ni@C+SiO2
Ni@C+Hydrotalcite
Co@C+CeO2
Cu@C+CeO2
Fe@C+CeO2
150
210
240
240
50
300
840
150
110
145
150
75
43.4
9.1
<3
<3
100
<2
9.7
–
–
–
94.1
–
75.6
94.5
10.7
–
3.2
–
–
–
4.1
–
6.2
0.9
5.6
15.9
1.6
5.6
–
85.5
>99.0
–
–
1.8
–
18.3
4.6
83.8
84.1
98.4
74.3
22.2
–
5.9
35.1
83.7
45.7
44.7
84.8
13.3
<5
9
10
11
12
13
14
15
–
6.5
77.8
–
360
360
360
–
–
<2
–
–
a
Reaction conditions: 1 mmol of nitrobenzene, 2 mL toluene as solvent, 120 °C and 10 bar of H2. 1.5 mg of metal nanoparticles and 10 mg of metal oxide as co-catalyst. The
yield of different products are calculated based on hexadecane as internal standard. For Ni/CeO2 (containing 5 wt% of Ni), 20 mg of the solid catalyst was used as catalyst.

L. Liu et al. / Journal of Catalysis 369 (2019) 312–323
317
selectivity of ꢀ90% at > 90% conversion. It should be noted, when
using Ni@C+CeO2 as the catalyst for hydrogenation of nitrosoben-
zene, the initial reaction rate is 4 times faster than the correspond-
ing homogeneous reaction, indicating that Ni@C+CeO2 is
catalytically active for the hydrogenation of nitrosobenzene to
azoxybenzene.
However, this homogeneous reaction is strongly related to the
concentration of nitrosobenzene in the solvent. As shown in
Fig. S11, the initial reaction rate decreases significantly when
reducing the concentration of nitrosobenzene from 0.5 mol LÀ1 to
0.05 mol LÀ1 and 0.025 mol LÀ1. As we already know, the concen-
tration of nitrosobenzene intermediate produced by Ni@C
nanoparticles under our reaction conditions is very low, even
lower than 0.025 mol LÀ1 (as we can see in Fig. S12). Therefore,
under our reaction conditions for the direct hydrogenation of
nitrobenzene to azoxybenzene, the influence of homogeneous
reaction is negligible when compared to the catalytic reaction rate
by Ni@C-CeO2 catalyst (see Fig. S13).
(>95%). To show the influence of stirring rate on the selectivity
for azoxybenzene, the amount of CeO2 was decreased to 2 mg,
which will cause less contact between Ni@C and CeO2 NPs. The cat-
alytic results presented in Fig. S16 show that the selectivity to
azoxybenzene also decreases with the stirring rate, which should
be caused by the insufficient diffusion of nitrosobenzene from
the Ni@C to CeO2 nanoparticles.
In another experiment, the surface of CeO2 is partially blocked
by the carbon coating treatment and then used as the co-catalyst
with Ni@C for hydrogenation of nitrobenzene. As can be seen in
Fig. S17, the initial reaction rate decreases slightly compared to
the naked CeO2. More importantly, the selectivity to azoxybenzene
also decreases, which further confirms the participation of CeO2 for
achieving high selectivity to azoxybenzene.
To further study the interaction of the reactants and intermedi-
ates with the surface of the catalyst, we have also performed the
in situ IR studies on the CeO2 NPs and Ni@C-CeO2 catalyst. As
shown in Fig. 4a, when CeO2 was used alone as catalyst after
Since the concentration of nitrosobenzene as intermediate in
the reaction mixture can be very low during the hydrogenation
of nitrobenzene, we have also carried out a competitive hydro-
genation experiment with a mixture of 90% nitrobenzene and
10% of nitrosobenzene at 60 °C. As shown in Fig. S14a, when only
Ni@C was used as the catalyst, nitrosobenzene was firstly hydro-
genated to both azoxybenzene and aniline while nitrobenzene
was not converted until nearly most of the nitrosobenzene was
consumed. Indeed, when the conversion of nitrosobenzene reaches
ꢀ70%, the hydrogenation of nitrobenzene started to occur, imply-
ing
a competitive adsorption between nitrosobenzene and
nitrobenzene on the surface of Ni@C nanoparticles. When Ni@C-
CeO2 was used as the catalyst (see Fig. S14b), nitrosobenzene
was also firstly selectively hydrogenated to azoxybenzene with a
higher reaction rate and >90% selectivity to azoxybenzene. More
importantly, hydrogenation of nitrobenzene only occurs when
almost all the nitrosobenzene in the reaction mixture was con-
verted. These results infer that, hydrogenation of nitrobenzene
on Ni@C is more preferable when the concentration of nitrosoben-
zene is low (<0.02 mol LÀ1), and that the presence of CeO2 can
accelerate the selective conversion of nitrosobenzene to azoxyben-
zene, which explains the higher reactivity and selectivity of Ni@C-
CeO2. Since nitrosobenzene is not observed as the intermediate
when using Ni@C-CeO2 as the catalyst, it implies that nitrosoben-
zene may be preferentially absorbed by the Ni@C-CeO2 composite
catalyst and converted further to azoxybenzene.
On the basis on the above control experiments, it can be
deduced that the introduction of CeO2 to Ni@C nanoparticles can
selectively adsorb and convert the low-concentration nitrosoben-
zene rapidly to avoid the over-hydrogenation reaction to aniline
on Ni@C nanoparticles. However, for doing that, nitrosobenzene
formed on the Ni@C NPs should diffuse and adsorb on the CeO2
NPs or the interfacial sites between Ni@C-CeO2 nanoparticles. For
effective hydrogenation of the nitrosobenzene, the H2 activated
on the Ni@C NPs should also reach the absorption sites of
nitrosobenzene. Then, it will be on the surface of the CeO2 NPs or
the Ni@C-CeO2 interfacial sites, where the azoxybenzene is pro-
duced. Nevertheless, the presence of Ni@C NPs is key to produce
the nitrosobenzene and to activate H2.
Since the model described above requires the diffusion of inter-
mediate products from one catalyst particle to another, mass trans-
portation should be maximized to achieve maximum efficiency.
Therefore, we study the influence of stirring speed in the reactor
on the rate of reaction. The results presented in Fig. S15 show that
the initial reaction rate increases when increasing the stirring rate
up to 900 rpm. After that, further increasing the stirring speed has
no effects on the initial reaction rate. Nevertheless, in the presence
of 10 mg of CeO2, selectivity to azoxybenzene is always high
Fig. 4. (a) In situ IR spectra of hydrogenation of nitrosobenzene (NSB) on CeO2 with
10 mbar H2. Spectra before H2 were collected immediately after NSB dosing (black)
and also after 10 min (gray) at 25°. After H2 dosing, spectra were collected at 25 °C
(red), 70 °C (blue) and 120 °C (green). (b). In situ IR spectra of hydrogenation of
nitrobenzene on CeO2 with 10 mbar H2. Before H2 dosing, the spectrum was
collected after the introduction of nitrobenzene at 25 °C (black). After H2 dosing IR
spectra were collected at 25 °C (red), 70 °C (blue) and 120 °C (green). IR spectra
have been acquired after 35 min at each temperature. DM, cis-dimer of nitrosoben-
zene; AN, aniline; AOB, azoxybenzene; AB, azobenzene; PHA, phenylhydroxy-
lamine. (For interpretation of the references to colour in this figure legend, the
reader is referred to the web version of this article.)

318
L. Liu et al. / Journal of Catalysis 369 (2019) 312–323
pre-activation in vacuum at 120 °C before the in situ IR experi-
ments, nitrosobenzene (1617, 1546, 1481 and 1472 cmÀ1) was
adsorbed on CeO2 and the hydrogenation of nitrosobenzene was
observed together with the formation of the dimer of nitrosoben-
zene (1401 cmÀ1) [27] Several reaction products were observed
in the IR spectra, including phenylhydroxylamine (1443 cmÀ1),
azoxybenzene (1568, 1374 and 1357 cmÀ1), azobenzene
(1550 cmÀ1) and aniline (1589, 1285 cmÀ1). Interestingly, the con-
version of nitrosobenzene during the in situ IR experiments is high,
suggesting the activity of CeO2 for hydrogenation reaction [28].
When the reactant was changed to nitrobenzene (1515 and
1345 cmÀ1), the formation of small amounts of azoxybenzene
(1374 and 1357 cmÀ1), azobenzene (1550 cmÀ1), and aniline
(1582, 1493 and 1285 cmÀ1) was still observed (see Fig. 4b)
[29,30]. However, the conversion of nitrobenzene is much lower
than that of nitrosobenzene, implying that CeO2 alone is not active
enough for hydrogenation of nitrobenzene, probably due to its low
activity for H2 activation. Furthermore, the hydrogenation of
nitrobenzene and nitrosobenzene on Ni@C-CeO2 has also been fol-
lowed by in situ IR. As shown in Fig. 5, high conversion of
nitrosobenzene (1617, 1546, 1481 and 1472 cmÀ1) (Fig. 5a) and
nitrobenzene (1515 and 1345 cmÀ1) (Fig. 5b) can be observed
together with the formation of the dimer of nitrosobenzene
(1401 cmÀ1), azoxybenzene (1572, 1376 and 1357 cmÀ1), azoben-
zene (1550 cmÀ1), nitrosobenzene (1486 cmÀ1) as well as aniline
(1602, 1498 and 1280 cmÀ1) are confirmed according to the IR
bands. For comparison, we have also followed the hydrogenation
of nitrobenzene on Ni@C + TiO2 catalyst presented above. As can
be seen in Fig. S18, only IR bands corresponding to aniline are
observed in the IR spectra, which is in line with the catalytic results
(see Table 1). The IR results indicate that, nitrosobenzene can be
further hydrogenated to phenylhydroxylamine and then form
azoxybenzene on CeO2. The role of Ni@C NPs involves the activa-
tion of H2 and production of nitrosobenzene as the primary inter-
mediate. As a result of the synergistic effects of Ni@C and CeO2,
nitrobenzene can be firstly hydrogenated to nitrosobenzene and
then to azoxybenzene.
The pristine Ni@C NPs and the sample after H2 pre-reduction
treatment have been measured by XPS. As shown in Fig. S19a,
40.7% of metallic Ni can be observed in the pristine Ni@C NPs. After
pre-reduction by H2 at 120 °C, the percentage of metallic Ni
increases to 54.2%. Even after pre-reduction treatment at 200 °C,
the amount of metallic Ni is 65.6%, suggesting the difficulty to
achieve fully reduced Ni nanoparticles under our current reaction
conditions for hydrogenation of nitrobenzene in batch reactor.
Interestingly, after being physically mixed with CeO2, the amount
of metallic Ni in the Ni@C NPs decreases and the NiO species
become more difficult to be reduced by H2 (see Fig. S19b). These
results refer interaction between Ni@C and CeO2 NPs, which mod-
ulates the redox properties of Ni@C NPs. During our catalytic tests,
we have observed that, the mixture of Ni@C and CeO2 in the sol-
vent (toluene) can form a magnetic composite. As shown in
Fig. S20, after stirring and heating at 60 °C for a short time, the
physical mixture of Ni@C and CeO2 was absorbed by the magnetic
bar, implying that some interaction may occur between the Ni@C
and CeO2 nanoparticles. As a consequence, the non-magnetic
CeO2 nanoparticles will also stick to the magnetic bar due to the
formation of a nanoparticulate composite with the magnetic
Ni@C. There are a number of –OH, –C = O and –COOH groups in
the thin carbon layers in the Ni@C NPs, which can interact with
the surface of CeO2 nanoparticles. The formation of interfacial con-
tact between Ni@C and CeO2 nanoparticles have been confirmed by
TEM-EDS analysis (see Figs. S21–S24).
Fig. 5. (a) In situ IR spectra of hydrogenation of nitrosobenzene on Ni@C + CeO2
with 10 mbar H2. Before the introduction of H2, immediately after NSB dosing at 25°
(black) and grey 10 min later). After H2 dosing, the spectra were collected at 25 °C
(red), 70 °C (blue) and 120 °C (green). (b) In situ IR spectra of hydrogenation of
nitrobenzene on Ni@C + CeO2 with 10 mbar H2. Before H2 dosing, spectrum was
collected after the introduction of nitrobenzene at 25 °C (black). After H2 dosing, IR
spectra were collected at 25 °C (red), 70 °C (blue) and 120 °C (green), respectively.
IR spectra have been acquired after 35 min at each temperature. DM, cis-dimer of
nitrosobenzene; AN, aniline; AOB, azoxybenzene; AB, azobenzene; NB, nitroben-
zene; NSB, nitrosobenzene. (For interpretation of the references to colour in this
figure legend, the reader is referred to the web version of this article.)
catalyst. As shown in Fig. S25, the pre-reduced catalyst shows sig-
nificantly higher initial activity for hydrogenation of nitrobenzene.
More importantly, the high selectivity (>90%) to azoxybenzene is
preserved after the pre-reduction treatment. These results indicate
that metallic Ni is the active phase in the Ni@C NPs and the NiO
present in the Ni@C NPs are not active, but has no influence on
the chemoselectivity. Since the Ni@C NPs are not fully reduced
under our current conditions (as we discussed before on the XPS
results), the nitrosobenzene intermediate formed on Ni@C can be
transferred to CeO2. According to our previous work [26], on the
fully reduced Ni surface, the nitrosobenzene can be further hydro-
genated to aniline directly before the desorption from the surface
of Ni@C NPs in the presence of abundant hydrogen.
We have investigated the influence of reaction conditions by
in situ IR spectroscopy. As displayed in Fig. S26a, when nitroben-
zene is co-absorbed with low-pressure H2 (0.4 mbar) on pre-
reduced Ni@C-CeO2 catalyst, aniline is not observed. When the
To further clarify the active sites in Ni@C NPs, we have also pre-
reduced the catalyst in the reactor before the hydrogenation reac-
tion and then tested the catalytic behavior of the pre-reduced

L. Liu et al. / Journal of Catalysis 369 (2019) 312–323
319
H2 pressure is increased to 10 or 40 mbar (Figs. S26b and c), aniline
appears in the IR spectra. By comparing the product distributions
in Figs. S16a and S28b, it can be seen that, after pre-reduction
treatment, the selectivity to azoxybenzene decreases from ꢀ85%
to ꢀ60% when using 1.5 mg of Ni@C and 2 mg of CeO2 as the cata-
lyst for hydrogenation of nitrobenzene. Therefore, the surface
properties of Ni@C NPs and their ability for the activation of H2
can also be a factor that influence the product distributions in
the hydrogenation of nitrobenzene with bifunctional Ni@C-CeO2
catalyst, although it is not so clear under our current reaction con-
ditions in the batch reactor.
To further prove the role of Ni@C, the hydrogenation of
nitrobenzene has been performed with the same amount of CeO2
(20 mg) while varying the amount of Ni@C (0.5–2.5 mg). As pre-
sented in Fig. S27, high selectivity to azoxybenzene has been
obtained in all the catalytic tests. The initial reaction rate increases
when increasing the amount of Ni@C catalyst and the selectivity to
azoxybenzene remains high.
The influence of the amount of CeO2 on the catalytic behavior of
Ni@C has also been studied. As shown in Fig. S28, addition of just
2 mg of CeO2 as co-catalyst can dramatically improve the activity
for hydrogenation of nitrobenzene. The selectivity to azoxyben-
zene is also improved from <10% to ꢀ60%. Further increasing the
amount of CeO2 to 5 mg will lead to >80% selectivity. When
20 mg of CeO2 is used as the co-catalyst, >95% selectivity to azoxy-
benzene can be achieved.
The above results indicate that at low mass ratios of CeO2/Ni@C
(5 mg/1.5 mg), the process for the formation of the azoxybenzene
is controlled by the condensation step between the intermediates
(nitrosobenzene and phenylhydroxylamine) and reduction step
on the CeO2. However, with a higher mass ratio of CeO2/Ni@C
(20 mg/2.5 mg), the reaction is controlled by the transformation
of nitrobenzene to nitrosobenzene, and reactivity is promoted by
increasing the amount of Ni@C. In other words, and as it occurs
within any bifunctional catalysts, both components should be opti-
mized to maximize the overall activity and selectivity [31–33].
In conventional supported metal catalysts (such as Au/TiO2,
Pt/TiO2, etc.), spectroscopic results and theoretical calculations
have proved that nitrobenzene is adsorbed at the metal-oxide
interface [34,35]. H2 is activated and dissociated on metal nanopar-
ticles and then transfer to the metal-oxide interface for hydrogena-
tion reaction (see Fig. 6a). In this work, Ni@C interact with CeO2
nanoparticles through inter-particle interaction, which is different
to conventional supported metal catalysts.
As we have mentioned before, the evolution of the intermedi-
ates during the hydrogenation of nitrobenzene with Ni@C or
Ni@C-CeO2 catalyst (as shown in Figs. S3 and S14) indicates that
the presence of CeO2 can immediately transform nitrosobenzene
into azoxybenzene once nitrosobenzene is formed. Therefore, it is
highly likely that, nitrosobenzene can be preferentially absorbed
on CeO2. Considering the above catalytic and IR results, we propose
that reaction intermediate species and activated H2 are transported
not only on a single catalyst particle, but between catalyst parti-
cles. As shown in Fig. 6b, H2 is activated on Ni@C NPs and the acti-
vated H species hydrogenate nitrobenzene into nitrosobenzene,
which is a rate-limiting step in the reaction. Meanwhile, activated
H can transfer to the interfacial sites between Ni@C and CeO2
nanoparticles through inter-particle spillover, and convert
nitrosobenzene into phenylhydroxylamine and making the con-
densation between nitrosobenzene and phenylhydroxylamine for
azoxybenzene [36–38]. As discussed before, the interfacial interac-
tion between Ni@C and CeO2 NPs lead to the formation of a Ni@C-
CeO2 interface, which facilitate the inter-particle H spillover from
Ni@C to CeO2 NPs. Therefore, by employing Ni@C as the active sites
for H2 activation and nitrobenzene hydrogenation, and using CeO2
as a co-catalyst for stabilization and further transformation of
Fig. 6. (a) Schematic illustration of direct hydrogenation of nitrobenzene to
azoxybenzene and then to azobenzene on Au/CeO2 catalyst. H2 molecules are
activated on Au nanoparticles and then transfer to the Au-CeO2 interface to catalyze
the hydrogenation of nitrobenzene. Nitrosobenzene and phenylhydroxylamine are
the intermediates during the transformation. (b) Schematic illustration of direct
chemoselective hydrogenation of nitrobenzene to azoxybenzene with Ni@C-CeO2
catalyst. Interfacial contact is formed between Ni@C and CeO2 nanoparticles under
reaction conditions. H2 molecules are activated on Ni@C nanoparticles and then
hydrogenate nitrobenzene to nitrosobenzene. The nitrosobenzene intermediate will
be preferentially absorbed by the interfacial sites between Ni@C and CeO2
nanoparticles, and further hydrogenated into azoxybenzene.
nitrosobenzene, the direct hydrogenation of nitrobenzene to
azoxybenzene can be achieved under base-free mild conditions
with high activity and selectivity. The structural of Ni@C-CeO2
bifunctional catalyst based on inter-particulate interaction pro-
posed in this work is different to conventional metal-support cat-
alysts, [39–42] providing a new strategy to design supported
metal catalysts.
It should be mentioned that, in the above proposed mechanism,
the hydrogenation of nitrobenzene and hydrogenation of
nitrosobenzene may occur on two different sites. However, it is
also possible that, nitrobenzene is absorbed on CeO2 and then con-
verted into azoxybenzene in the presence of activated H from the
Ni@C nanoparticles. According to our above experimental results,
we cannot rule out the above possibility. However, considering
the results obtained in the mass transfer experiments (see
Fig. S16), the intimate contact between Ni@C and CeO2 NPs to form
Ni@C-CeO2 interfacial sites is key to achieve high selectivity to
azoxybenzene. Thus, we favor the mechanism that involves two
separated reactive sites for hydrogenation of nitrobenzene to
nitrosobenzene and for the formation of azoxybenzene,
respectively.
We have compared the catalytic performance of Ni@C-CeO2
with the state-of-art Au/CeO2 for hydrogenation of nitrobenzene
under the same conditions. As shown in Fig. 7, with a higher
amount of solid catalyst (50 mg of Au/CeO2 vs. 1.5 mg of Ni@C
+ 10 mg of CeO2), Au/CeO2 showed a much lower reaction rate.
More importantly, Au/CeO2 gives 90% selectivity to azoxybenzene

320
L. Liu et al. / Journal of Catalysis 369 (2019) 312–323
Fig. 7. Comparison between Au/CeO2 (1 wt% of Au) and Ni@C-CeO2 for hydrogenation of nitrobenzene under the same conditions. (a) Catalytic result of Ni@C-CeO2 for
hydrogenation of nitrobenzene to azoxybenzene. Small amount of azobenzene and aniline were obtained as by-products. Reaction conditions: 1 mmol of nitrobenzene, 2 mL
toluene as solvent, 1.5 mg Ni@C and 10 mg of CeO2 as catalyst, 120 °C and 10 bar of H2. (b) Catalytic result of Au/CeO2 for hydrogenation of nitrobenzene. Both azoxybenzene
and azobenzene were observed in the products. Reaction conditions: 1 mmol of nitrobenzene, 2 mL toluene as solvent, 50 mg of Au/CeO2 containing 0.5 mg of Au as catalyst,
120 °C and 10 bar of H2. (c) Initial reaction rates (mmolH2∙hÀ1∙mgÀ1) normalized to the mass of metal of Au/CeO2 and Ni@C-CeO2 catalysts for hydrogenation of nitrobenzene.
(d) Initial TOF values of Au/CeO2 and Ni@C-CeO2 for hydrogenation of nitrobenzene, calculated based on the surface atoms of Au and Ni in both systems. The amount of
surface Au atoms in the Au/CeO2 sample is calculated according to the size distribution of Au nanoparticles. In the case of Ni@C-CeO2 catalyst, the amount of exposed surface
Ni atoms is calculated based on the chemisorption measurement using H2 as probe molecules.
at 94% conversion while Ni@C-CeO2 gives 94% selectivity of azoxy-
benzene at 100% conversion. More importantly, Ni@C-CeO2 cata-
lyst can achieve full conversion of nitrobenzene with a much
shorter reaction time than Au/CeO2, especially if considering the
lower quantity of solid catalyst used with Ni@C-CeO2. The initial
reaction rates normalized to the mass of metal catalysts and the
initial turnover frequencies (TOFs) have also been calculated based
on the exposed surface atoms in the Au/CeO2 and Ni@C-CeO2 cat-
alysts. Ni@C-CeO2 shows better performance than Au/CeO2, indi-
cating that the combination of Ni@C and CeO2 is a good catalyst
for the above-mentioned process.
components can be merged into a single solid catalyst. To achieve
that aim, we have prepared a Ni-CeO2-C catalyst by one-pot
hydrothermal synthesis and post-annealing treatment in N2 at
600 °C (see experimental section for details). As shown in Fig. 8,
Ni nanoparticles and CeO2 nanoparticles are mixed together and
both of them are covered by a few layers of carbon. The Ni-CeO2-
C sample also contains a mixture of metallic Ni and NiO according
to the XPS results (see Fig. S29). The catalytic results shown in
Table 2 indicate that, Ni-CeO2-C sample is also active and selective
for the hydrogenation of nitrobenzene to azoxybenzene. The activ-
ity of Ni-CeO2-C is higher than the Ni@C-CeO2 catalyst, which can
be attributed the intimate contact between Ni and CeO2 in the Ni-
CeO2-C nanocomposites. However, if CeO2 is replaced by other
From a practical point of view, it will be more convenient for the
application of this Ni@C-CeO2 bifunctional system if the separated

L. Liu et al. / Journal of Catalysis 369 (2019) 312–323
321
Fig. 8. (a and b) High-resolution TEM images of the Ni-CeO2-C catalyst, showing the presence of Ni nanoparticles and CeO2 nanoparticles with thin carbon layers surrounding
them. (c–f) STEM image and the corresponding elemental mapping of that area.
Table 2
Catalytic performance of Ni-MOx-C catalyst prepared by one-pot hydrothermal synthesis for hydrogenation of nitrobenzene.a
Sample
Time/min
Conversion/%
Selectivity to Azoxy
Selectivity to aniline
Selectivity to azo
Ni-CeO2-C
Ni-TiO2-C
Ni-Al2O3-C
Ni-ZrO2-C
35
97.1
38.2
58.9
19.8
92.4
–
1.3
21.9
4.9
2.7
–
7.4
–
190
195
190
>99
91.3
78.1
a
Reaction conditions: 1 mmol of nitrobenzene, 2 mL toluene as solvent, 120 °C and 10 bar of H2. 10 mg of solid catalyst. The amount of Ni in the Ni-MOx-C catalysts is ca.
15 wt%. The yield of different products are calculated based on hexadecane as internal standard.

322
L. Liu et al. / Journal of Catalysis 369 (2019) 312–323
Table 4
metal oxides (TiO2, Al2O3 and ZrO2), the selectivity to azoxyben-
zene drops dramatically, which is similar to the results obtained
before with separated Ni@C and metal oxides.
Scope study on direct hydrogenation of nitroarenes into corresponding azo
compounds.a
Reactant
Product
Yield/%
87%
We have also studied the scope for direct hydrogenation of
nitroarenes into corresponding azoxy compounds. As shown in
Table 3, high selectivity can be achieved for a variety of nitroarenes
with different functional groups. Notably, 3-iodonitrobenzene can
be selectivity hydrogenated into corresponding azoxy compounds,
which cannot be achieved by supported Au catalysts. Therefore,
the combination of Ni@C NPs and CeO2 can catalyze the formation
of azoxy compounds with high chemoselectivity by direct hydro-
genation of nitroarenes under mild conditions.
68%
58%
According to the reaction scheme shown in Fig. 2, azoxyben-
zene could be further hydrogenated into azobenzene by simply
controlling the reaction time, which offers the possibility to obtain
either the corresponding azoxy or azo compounds [29,43]. Then,
after achieving high yield of azoxybenzene from nitrobenzene with
Ni@C-CeO2 catalyst, the hydrogenation reaction was continued and
to our delight, azobenzene can be obtained with yield of 87%.
Moreover, we have also studied the scope of Ni@C-CeO2 catalyst
for direct hydrogenation of various nitroarenes into corresponding
azo compounds, through the azoxy compounds as intermediates.
As shown in Table 4, several substituted azo compounds can be
obtained from moderate to good yields under bass-free mild con-
ditions [44]. Especially, it is possible to prepare aromatic azo com-
pounds with iodo-substituted group with Ni@C-CeO2 catalyst,
62%
91%
71%
70%
Table 3
Scope study on direct hydrogenation of nitroarenes into corresponding aromatic
a
azoxy products.a
Reaction conditions: 1 mmol of nitrobenzene, 2 mL toluene as solvent, 90 °C
and 10 bar of H2. 2.0 mg of Ni@C nanoparticles and 20 mg of CeO2 as co-catalyst.
The yield of different products are calculated based on hexadecane as internal
standard.
Reactant
Product
Yield/%
94%
88%
NH2
NO2
H2
Co@C, Fe@C, Ni@C...
95%
92%
91%
R
R
H2
Ni@C+CeO2
H2
R
R
O
N
H2
N
N
N
Ni@C+CeO2
R
R
58%
Fig. 9. Chemoselective hydrogenation of nitroarenes to different products with
non-noble metal catalysts. By choosing different catalysts and co-catalysts, it is
possible to obtain various products (azoxy, azo and aniline) by tuning the reaction
conditions and reaction time.
98%
92%
84%
which is not possible with Au-based catalysts. At this point, based
on the above results and taking into consideration of the high reac-
tivity and low cost of Ni@C-CeO2 catalyst, this catalyst can be a
promising substitute for Au/CeO2. At present, combining the
results shown in this work and our previous works, we are able
to selectively convert nitroarenes into either aniline, azoxy or azo
compounds by means of nanoparticulate non-noble metal catalysts
(see Fig. 9) [21–25].
88%
4. Conclusions
a
Reaction conditions: 1 mmol of nitrobenzene, 2 mL toluene as solvent, 120 °C
and 10 bar of H2. 2.0 mg of Ni@C nanoparticles and 20 mg of CeO2 as co-catalyst.
The yield of different products are calculated based on hexadecane as internal
standard.
Bifunctional non-noble Ni@C-CeO2 catalysts are found to be
highly active and chemoselective catalysts for direct hydrogenation

L. Liu et al. / Journal of Catalysis 369 (2019) 312–323
323
[18] T. Hou, Y. Wang, J. Zhang, M. Li, J. Lu, M. Heggen, C. Sievers, F. Wang, Peculiar
hydrogenation reactivity of Ni–Nid+ clusters stabilized by ceria in reducing
nitrobenzene to azoxybenzene, J. Catal. 353 (2017) 107–115.
[19] Z.P. Zhang, X.Y. Wang, K. Yuan, W. Zhu, T. Zhang, Y.H. Wang, J. Ke, X.Y. Zheng,
C.H. Yan, Y.W. Zhang, Free-standing iridium and rhodium-based
hierarchically-coiled ultrathin nanosheets for highly selective reduction of
nitrobenzene to azoxybenzene under ambient conditions, Nanoscale 8 (2016)
15744–15752.
[20] L. Liu, A. Corma, Metal catalysts for heterogeneous catalysis: from single atoms
to nanoclusters and nanoparticles, Chem. Rev. 118 (2018) 4981–5079.
[21] F.A. Westerhaus, R.V. Jagadeesh, G. Wienhofer, M.M. Pohl, J. Radnik, A.E.
Surkus, J. Rabeah, K. Junge, H. Junge, M. Nielsen, A. Bruckner, M. Beller,
Heterogenized cobalt oxide catalysts for nitroarene reduction by pyrolysis of
molecularly defined complexes, Nat. Chem. 5 (2013) 537–543.
[22] R.V. Jagadeesh, A.E. Surkus, H. Junge, M.M. Pohl, J. Radnik, J. Rabeah, H. Huan, V.
Schunemann, A. Bruckner, M. Beller, Nanoscale Fe2O3-based catalysts for
selective hydrogenation of nitroarenes to anilines, Science 342 (2013) 1073–
1076.
of nitroarenes into corresponding azoxy or azo products under mild
conditions. Notably, the Ni@C-CeO2 catalyst surpasses the perfor-
mance of state-of-art Au/CeO2 catalyst for direct transformation of
nitroarenes to azoxy compounds. Nitroarenes can be reduced to cor-
responding nitroso compound on Ni@C nanoparticles then diffuse
to interfacial sites between Ni@C and CeO2 nanoparticles. H2 is also
activated on Ni@C NPs and transfer to the interfacial sites where the
chemoselective hydrogenation of nitrosobenzene to azoxybenzene
occurs. The Ni@C-CeO2 catalyst enables us to select the product to
be obtained (azoxy, azo or aniline compound). Considering the
hydrogenation mechanism proposed in this work, it could be possi-
ble to tune the catalytic properties of metal nanoparticles by differ-
ent combinations of metal-oxide(or other co-catalysts) for various
chemoselective hydrogenation reactions.
[23] G. Hahn, J.-K. Ewert, C. Denner, D. Tilgner, R. Kempe, A reusable mesoporous
nickel nanocomposite catalyst for the selective hydrogenation of nitroarenes
in the presence of sensitive functional groups, ChemCatChem 8 (2016) 2461–
2465.
Acknowledgments
[24] L. Liu, P. Concepción, A. Corma, Non-noble metal catalysts for hydrogenation: a
facile method for preparing Co nanoparticles covered with thin layered carbon,
J. Catal. 340 (2016) 1–9.
[25] L. Liu, F. Gao, P. Concepción, A. Corma, A new strategy to transform mono and
bimetallic non-noble metal nanoparticles into highly active and
chemoselective hydrogenation catalysts, J. Catal. 350 (2017) 218–225.
[26] R. Millán, L. Liu, M. Boronat, A. Corma, A new molecular pathway allows the
chemoselective reduction of nitroaromatics on non-noble metal catalysts, J.
Catal. 364 (2018) 19–30.
This work has been supported by the European Union through
the SynCatMatch project (ERC-AdG-2014-671093). Financial sup-
ports by the Spanish Government-MINECO through the program
‘‘Severo Ochoa” (SEV-2016-0683) are gratefully acknowledged.
The authors also thank the Microscopy Service of UPV for kind help
with TEM and STEM measurements.
[27] S. Meijers, V. Ponec, J. Catal. 160 (1996) 1–9.
[28] G. Vile, B. Bridier, J. Wichert, J. Perez-Ramirez, Ceria in hydrogenation
catalysis: high selectivity in the conversion of alkynes to olefins, Angew.
Chem. Int. Ed. 51 (2012) 8620–8623.
Appendix A. Supplementary material
[29] D. Combita, P. Concepción, A. Corma, Gold catalysts for the synthesis of
aromatic azocompounds from nitroaromatics in one step, J. Catal. 311 (2014)
339–349.
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.jcat.2018.11.011.
[30] G. Richner, J.A. van Bokhoven, Y.M. Neuhold, M. Makosch, K. Hungerbuhler, In
situ infrared monitoring of the solid/liquid catalyst interface during the three-
phase hydrogenation of nitrobenzene over nanosized Au on TiO2, Phys. Chem.
Chem. Phys. 13 (2011) 12463–12471.
[31] F. Jiao, J. Li, X. Pan, J. Xiao, H. Li, H. Ma, M. Wei, Y. Pan, Z. Zhou, M. Li, S. Miao, J.
Li, Y. Zhu, D. Xiao, T. He, J. Yang, F. Qi, Q. Fu, X. Bao, Selective conversion of
syngas to light olefins, Science 351 (2016) 1065–1068.
[32] J. Zecevic, G. Vanbutsele, K.P. de Jong, J.A. Martens, Nanoscale intimacy in
bifunctional catalysts for selective conversion of hydrocarbons, Nature 528
(2015) 245–248.
References
[1] A. Corma, Heterogeneous catalysis: understanding for designing, and
designing for applications, Angew. Chem. Int. Ed. 55 (2016) 6112–6113.
[2] R. Ye, T.J. Hurlburt, K. Sabyrov, S. Alayoglu, G.A. Somorjai, Molecular catalysis
science: perspective on unifying the fields of catalysis, Proc. Natl. Acad. Sci.
USA 113 (2016) 5159–5166.
[3] R. Schlogl, Heterogeneous catalysis, Angew. Chem. Int. Ed. 54 (2015) 3465–
3520.
[4] P. Serna, A. Corma, Transforming nano metal nonselective particulates into
chemoselective catalysts for hydrogenation of substituted nitrobenzenes, ACS
Catal. 5 (2015) 7114–7121.
[33] X. Kong, Y. Zhu, H. Zheng, X. Li, Y. Zhu, Y.-W. Li, Ni Nanoparticles inlaid Nickel
phyllosilicate as
a metal-acid bifunctional catalyst for low-temperature
hydrogenolysis reactions, ACS Catal. 5 (2015) 5914–5920.
[34] M. Boronat, P. Concepcion, A. Corma, S. Gonzalez, F. Illas, P. Serna, A Molecular
mechanism for the chemoselective hydrogenation of substituted
nitroaromatics with nanoparticles of gold on TiO2 catalysts: a cooperative
effect between gold and the support, J. Am. Chem. Soc. 129 (2007) 16230–
16237.
[35] A. Corma, P. Serna, P. Concepcion, J.J. Calvino, Transforming nonselective into
chemoselective metal catalysts for the hydrogenation of substituted
nitroaromatics, J. Am. Chem. Soc. 130 (2008) 8748–8753.
[36] R. Prins, Hydrogen spillover. Facts and fiction, Chem. Rev. 112 (2012) 2714–
2738.
[37] S. Lee, K. Lee, J. Im, H. Kim, M. Choi, Revisiting hydrogen spillover in Pt/LTA:
Effects of physical diluents having different acid site distributions, J. Catal. 325
(2015) 26–34.
[38] W. Karim, C. Spreafico, A. Kleibert, J. Gobrecht, J. VandeVondele, Y. Ekinci, J.A.
van Bokhoven, Catalyst support effects on hydrogen spillover, Nature 541
(2017) 68–71.
[39] L. Liu, X. Gu, Y. Cao, X. Yao, L. Zhang, C. Tang, F. Gao, L. Dong, Crystal-plane
effects on the catalytic properties of Au/TiO2, ACS Catal. 3 (2013) 2768–2775.
[40] L. Liu, C. Ge, W. Zou, X. Gu, F. Gao, L. Dong, Phys. Chem. Chem. Phys. 17 (2015)
5133–5140.
[41] D.I. Enache, J.K. Edwards, P. Landon, B. Solsona-Espriu, A.F. Carley, A.A. Herzing,
M. Watanabe, C.J. Kiely, D.W. Knight, G.J. Hutchings, Solvent-free oxidation of
primary alcohols to aldehydes using Au-Pd/TiO2 catalysts, Science 311 (2006)
362–365.
[42] M. Cargnello, V.V. Doan-Nguyen, T.R. Gordon, R.E. Diaz, E.A. Stach, R.J. Gorte, P.
Fornasiero, C.B. Murray, Control of metal nanocrystal size reveals metal-
support interface role for ceria catalysts, Science 341 (2013) 771–773.
[43] X. Liu, H.Q. Li, S. Ye, Y.M. Liu, H.Y. He, Y. Cao, Gold-catalyzed direct
hydrogenative coupling of nitroarenes to synthesize aromatic azo
compounds, Angew. Chem. Int. Ed. 53 (2014) 7624–7628.
[5] M. Filice, J.M. Palomo, Cascade reactions catalyzed by bionanostructures, ACS
Catal. 4 (2014) 1588–1598.
[6] M.J. Climent, A. Corma, S. Iborra, M.J. Sabater, Heterogeneous catalysis for
tandem reactions, ACS Catal. 4 (2014) 870–891.
}
}
[7] G. Szollosi, Asymmetric one-pot reactions using heterogeneous chemical
catalysis: recent steps towards sustainable processes, Catal. Sci. Technol. 8
(2018) 389–422.
[8] B.H. Lipshutz, N.A. Isley, J.C. Fennewald, E.D. Slack, On the way towards greener
transition-metal-catalyzed processes as quantified by E factors, Angew. Chem.
Int. Ed. 52 (2013) 10952–10958.
[9] R.A. Sheldon, The E factor 25 years on: the rise of green chemistry and
sustainability, Green Chem. 19 (2017) 18–43.
[10] A. Leyva-Perez, P. Garcia-Garcia, A. Corma, Multisite organic-inorganic hybrid
catalysts for the direct sustainable synthesis of GABAergic drugs, Angew.
Chem. Int. Ed. 53 (2014) 8687–8690.
[11] H.E. Bigelow, Azoxy compounds, Chem. Rev. 9 (1931) 117–167.
[12] M.M. Russew, S. Hecht, Photoswitches: from molecules to materials, Adv.
Mater. 22 (2010) 3348–3360.
[13] E. Buncel, Catalysis in strongly acidic media and the wallach rearrangement,
Acc. Chem. Res. 8 (1975) 132–139.
[14] S. Sakaue, T. Tsubakino, Y. Nishiyama, Y. Ishii, Oxidation of aromatic amines
with
hydrogen
peroxide
catalyzed
by
cetylpyridinium
heteropolyoxometalates, J. Org. Chem. 58 (1993) 3633–3638.
[15] S.S. Acharyya, S. Ghosh, R. Bal, Catalytic oxidation of aniline to azoxybenzene
over CuCr2O4 spinel nanoparticle catalyst, ACS Sustain. Chem. Eng. 2 (2014)
584–589.
[16] H. Tumma, N. Nagaraju, K.V. Reddy, Titanium (IV) oxide, an efficient and
structure-sensitive heterogeneous catalyst for the preparation of
azoxybenzenes in the presence of hydrogen peroxide, Appl. Catal. A: Gen.
353 (2009) 54–60.
[17] M.N. Pahalagedara, L.R. Pahalagedara, J. He, R. Miao, B. Gottlieb, D. Rathnayake,
S.L. Suib, Room temperature selective reduction of nitrobenzene to
azoxybenzene over magnetically separable urchin-like Ni/Graphene
nanocomposites, J. Catal. 336 (2016) 41–48.
[44] W. Liu, L. Zhang, W. Yan, X. Liu, X. Yang, S. Miao, W. Wang, A. Wang, T. Zhang,
Single-atom dispersed Co–N–C catalyst: structure identification and
performance for hydrogenative coupling of nitroarenes, Chem. Sci. 7 (2016)
5758–5764.