Highly dispersed cobalt oxides nanoparticles on activated carbon fibres as efficient structured catalysts for the transfer hydrogenation of m-nitrostyrene

Article abstract of DOI:10.1016/j.cattod.2016.06.043

The facile preparation of structured catalysts featuring highly-dispersed non-precious metal (Fe, Ni, Co) oxide nanoparticles stabilized within a super-microporous network of activated carbon fibres (ACF) is described. The catalyst morphology was controll

Full text of DOI:10.1016/j.cattod.2016.06.043

G Model
CATTOD-10293; No. of Pages7
ARTICLE IN PRESS
Catalysis Today xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Highly dispersed cobalt oxides nanoparticles on activated carbon
fibres as efficient structured catalysts for the transfer hydrogenation
of m-nitrostyrene
a
a
a
b
a
Oliver Beswick , Alexander Parastaev , Igor Yuranov , Thomas LaGrange , Paul J. Dyson
a,c,∗
,
Lioubov Kiwi-Minsker
Insitute of Chemical Sciences and Engineering, Switzerland
Interdisciplinary Centre for Electron Microscopy (CIME), Ecole Polytechnique Fédérale de Lausanne (EPFL), Lausanne CH-1015, Switzerland
Regional Technological Centre, Tver State University, Tver 170100, Russian Federation
a
b
c
a r t i c l e i n f o
a b s t r a c t
Article history:
The facile preparation of structured catalysts featuring highly-dispersed non-precious metal (Fe, Ni, Co)
oxide nanoparticles stabilized within a super-microporous network of activated carbon fibres (ACF) is
described. The catalyst morphology was controlled over multiple levels from the nano-designed active
phase up to the macro-structure of the ACF support. A range of physicochemical techniques was used to
characterize the catalyst and rationalize the catalytic data during chemoselective transfer hydrogenation
of nitroarenes. The challenging reduction of m-nitrostyrene containing an easily reducible vinyl-group
was obtained under mild conditions with a near-quantitative yield of m-vinylaniline using hydrazine
hydrate as the reducing agent over CoOx/ACF as a structured catalyst.
Received 2 January 2016
Received in revised form 16 June 2016
Accepted 21 June 2016
Available online xxx
Keywords:
Transfer hydrogenation
Hydrazine hydrate
m-nitrostyrene hydrogenation
Cobalt-oxide nanoparticles
Activated carbon fibres
Structured catalysts
©
2016 Elsevier B.V. All rights reserved.
1
. Introduction
easy to handle circumventing safety-related issues. Hydrazine is
an ideal hydrogen source as the only by-products are N2 gas and
Functionalized anilines are key building blocks in the synthesis
H O.
2
of pharmaceuticals, dyes, pigments, polymers, and agrochemicals
1,2]. Usually, they are produced by the selective hydrogenation
Although noble metal based CTH catalysts such as Pd nanopar-
ticles (NPs) supported on carbon nanospheres [9] and Rh NPs
supported on polymer resin [10], etc. have been reported, there
are only few examples of non-precious metal based catalysts effi-
cient in the CTH of nitroarenes. Several CTH reactions were carried
[
of the corresponding nitroarenes over catalysts based on precious
metals (Pt, Pd, Ru, Rh, etc.) in the presence of gaseous hydrogen
and different modifiers or promoters [3]. However, these catalysts
are expensive and toxic and it would be advantageous to replace
them with earth abundant metals, (Fe, Co, Ni, Ce, etc.), which are
inexpensive and less toxic.
An attractive alternative to hydrogenation to convert
nitroarenes to anilines is catalytic transfer hydrogenation (CTH).
CTH refers to the addition of hydrogen to an organic molecule from
a hydrogen donor compounds such as hydrazine [4], formic acid
out over Co-based catalysts such as Co-Mo C/activated carbon [11],
2
and Co(II) substituted hexagonal mesoporous aluminophosphate
[12] with different hydrogen sources. The vinyl group was mainly
kept intact during the generation of m-vinylaniline (m-VA) from
m-nitrostyrene (m-NS) over nanosized Co oxide supported on N-
modified carbon but the highest yield (Ym-VA) reported was 93%
[5]. Previously this research group employed iron NPs modified by
1,10-phenanthroline that lead to highly selective transformation of
a range of nitroarenes using hydrazine, including the challenging
m-NS (Ym-VA = 96%). But so far, the highest yields in m-VA of 97%
was obtained during hydrogenation with gaseous hydrogen over
Pt-based catalysts (T = 348 K, PH2 = 10 bar) [13,14].
[
5], isopropanol [6], NaBH4 [7] and has become recently a focus of
research [8]. Indeed, CTH does not require hazardous pressurized
H2 gas with autoclaves and hydrogen donors are inexpensive and
∗
We have recently reported high performance of Fe-oxide NPs
supported on activated carbon fibres (ACF) showing close to 100%
Corresponding author at: Insitute of Chemical Sciences and Engineering,
Switzerland.
E-mail address: lioubov.kiwi-minsker@epfl.ch (L. Kiwi-Minsker).
http://dx.doi.org/10.1016/j.cattod.2016.06.043
0
920-5861/© 2016 Elsevier B.V. All rights reserved.
Please cite this article in press as: O. Beswick, et al., Highly dispersed cobalt oxides nanoparticles on activated
carbon fibres as efficient structured catalysts for the transfer hydrogenation of m-nitrostyrene, Catal. Today (2016),
http://dx.doi.org/10.1016/j.cattod.2016.06.043

G Model
CATTOD-10293; No. of Pages7
ARTICLE IN PRESS
O. Beswick et al. / Catalysis Today xxx (2016) xxx–xxx
2
yield of the desired anilines during hydrazine assisted CTH of vari-
ety of a nitroarenes but not for a m-nitrostyrene which yield was
below 80% [15]. The ACF structured support consisting of arranged
microfilaments (d–20 ␮m) is characterized by high specific surface
the catalysts. Samples were calcined in air at 873 K for 2 h and
the remaining solid (metal oxide) was dissolved in aqua regia
(HNO :HCl (v/v) = 1:3) and diluted by demineralized water.
3
The specific surface areas (SSA) of the support and the cata-
lysts were measured by adsorption-desorption of N2 at 77 K using
a Sorptomatic 1990 instrument (Carlo Erba). Prior to analysis, sam-
2
−1
area (∼2000 m g ), high fluid permeability and easy handling. Slit
shaped and almost uniform ACF micropores (<2 nm) are opened
to the outer filament surface [16]. Such a network of micropores
allows extremely small metal oxide NPs to be stabilized. In addition,
internal mass transfer limitations within micro-sized filaments are
avoided. These supports were also successfully used in conven-
tional multiphase catalytic hydrogenations of nitrobenzene over
Pt/ACF [17–19] and bromates and nitrates over Pd/ACF [20,21].
Herein we describe the facile preparation of ACF-based struc-
tured catalysts with supported transition metal (Fe, Ni, Co) oxides
nanoparticles as an active phase. The catalyst morphology was
controlled at multiple levels starting from the nano-designed
metal oxides up to the macro-structure of the carbon support. A
CoOx/ACFHNO3 catalyst containing extremely small nanoparticles
−
2
ples were outgassed at 523 K for 2 h under vacuum (7 × 10 bar).
N
2
adsorption/desorption isotherms were recoded over the range
0.0005 ≤ P/P 0 ≤ 0.98. The SSA and the total pore volume were cal-
culated using the BET method [22].
X-ray diffraction (XRD) patterns were recorded on
a
Bruker/Siemens D500 incident X-ray diffractometer using CuK␣
◦
◦
radiation (␭ = 0.154 nm). The scan ranged from 30 ≤ 2␪ ≤ 85 at
◦
−1
0.02 step . The diffractograms were compared to JCPDS-ICDD
reference standards, i.e. Co (15-0806), CoO (43–1004) and Co O4
3
(74–1656).
X-ray photoelectron spectroscopy (XPS) analyses were conducted
on an Axis Ultra instrument (Kratos analytical). The source power
was maintained at 150 W and the emitted photoelectrons were
(
<2 nm) was the most effective in the CTH of m-nitrostyrene using
◦
hydrazine to afford the desired m-vinylaniline product in near-
quantitative yield. Considering the potential applications of this
catalytic system in batch as well as in flow-mode, we anticipate
that the results presented in this study will offer new avenues in the
sustainable synthesis of arylamines containing C C double bonds.
sampled from a 750 ␮m × 350 ␮m area at a take-off angle = 90 .
The analyser pass energy was 80 eV for survey spectra (0–1100 eV)
and 40 eV for high resolution spectra (Co 2p3/2 and Co 2p1/2). The
adventitious C (284.8 eV) and O (530.8 eV) 1 s peaks were used as
internal standards to compensate for any charging effects.
High angle annular dark-field scanning transmission electron
microscope (HAADF STEM) images were acquired using an FEI Tec-
nai Osiris instrument. Chemical mappings were obtained using the
“Super-X” energy-dispersive X-ray (EDX) spectroscopy detector.
The measurements were operated at the maximum accelerating
voltage of 200 keV. Specimens were prepared by infiltrating-
embedding the MOx/ACF catalysts into an EPON 812 epoxy
resin followed by its polymerization at 333 K for 24 h. Ultra-
microtomy (diamond grade) was used to cut 60 nm thick fibre
cross-sections that were eventually deposited on a holey carbon/Cu
grid (300 Mesh).
2
. Experimental
2.1. Materials
2
−1
Activated carbon fibres (ACF-K-20, SSA ∼2000 m g ) were
provided by Kynol Europa GmbH (Hamburg, Germany). Fe(III)
nitrate monohydrate (>98%, Fluka), Co(II) nitrate hexahydrate
(
>99%, Acros), Ni(II) nitrate hexahydrate (>98%, Fluka) were used
as precursors in the catalyst preparation. Co(II, III) oxide nanopar-
ticles (>50 nm, Sigma-Aldrich) were used as a powder catalyst
for comparison; p-nitroanisole (97%, Alfa Aeasar), p-nitrotoluene
Scanning electron microscopy (SEM) analyses were realized
with a Philips FEI XL30-FEG equipped with an Everhart-Thornley
secondary-electron (SE) detector using an accelerating voltage of
13 kV.
(
>98%, Fluka), m-nitrostyrene (97%, stabilized, Acros), nitroben-
zene (99%, Acros), p-chloronitrobenzene (99%, Sigma Aldrich),
m-xylene ( >98%, Fluka) and hydrazine monohydrate (∼100%,
Merck) were used as received without further purification. All gases
Temperature-programmed reduction in hydrogen (TPR) was car-
ried out using a Micromeritics Autochem II 2920 system. Samples
(
>99.99%) were purchased from Carbagas Switzerland.
3
−1
−1
were heated in a 20 cm min 5% (v/v) H /Ar flow at 2 K min to
2
2
.2. Catalyst preparation
873 K. The outlet gas stream after passing through a liquid N2 trap
was monitored by a TCD to measure H2 concentration variations.
Data were acquired/manipulated with TPR Win software.
TM
The MOx/ACF (M = Fe, Co or Ni) catalysts were prepared as fol-
lows. The ACF supports were impregnated with ethanolic solutions
of metal oxide precursors (nitrates) ensuring complete filling of
the pores. Prior to impregnation, some ACF were pre-treated in a
Temperature-programmed decomposition (TPD) measurements
3
−1
were conducted by heating samples in a He flow (50 cm min
)
−
1
at 6 K min from RT up to 1173 K. The outlet gas flow composition
was monitored by a mass spectrometer (Thermostar TM GSD 300
T2, Pfeiffer Vacuum).
1
5 wt.% HNO aq. solution at 373 K for 15 min (denoted as ACF
).
3
HNO3
The concentration of a precursor solution was adjusted to reach a
desired metal loading in the range of 4–7 wt.% for the samples used
for catalytic testing and up to 10 wt.% for the catalyst characterisa-
tions. The loading is indicated in the text and in the figure captions.
Impregnated ACF samples were then dried in air at room temper-
ature (RT) over-night and thermally treated in a tubular reactor
2.4. Catalytic transfer hydrogenation: set-up and procedure
Liquid-phase CTH reactions were carried out in a commer-
3
cial semi-batch stainless steel stirred reactor (150 cm , Büchi AG,
3
−1
(
50 cm × 3 cm i. d.) under an Ar flow (280 cm min ) to decom-
Uster, Switzerland) equipped with 4 wall baffles. The reaction
temperature was controlled using an oil heating bath Ministat
125 (Huber Kältemaschinenbau GmbH, Germany) connected to a
reactor jacket. A 6-blade disk turbine impeller ensured intensive
mixing. The catalyst was fixed on four wire-mesh blades attached
directly to the stirrer shaft. The stirrer was driven by a magnetic
drive and equipped with a speed controller (cyclone 075/cc 075,
Büchi AG, Uster, Switzerland). The reactor temperature, pressure
and stirring speed were monitored via a control unit (bpc 6002/bds
mc, Büchi AG, Uster, Switzerland). Prior to each experiment the
reactor was charged with 0.3 g of the catalyst and 100 cm3 of an
pose the nitrates precursor. The reactor was heated from RT to 623 K
−
1
(
temperature ramp − 6 K min ) and maintained at this tempera-
ture for 1 h. Finally, the catalysts were passivated in a 2.8% v/v air/Ar
3
−1
flow (145 cm min ) at RT during 1 h in order to avoid pyrophoric
oxidation on air.
2.3. Catalyst characterization
Atomic absorption spectroscopy (AAS) was carried out using a Shi-
madzu AA-6650 spectrometer to determine the metal content of
Please cite this article in press as: O. Beswick, et al., Highly dispersed cobalt oxides nanoparticles on activated
carbon fibres as efficient structured catalysts for the transfer hydrogenation of m-nitrostyrene, Catal. Today (2016),
http://dx.doi.org/10.1016/j.cattod.2016.06.043

G Model
CATTOD-10293; No. of Pages7
ARTICLE IN PRESS
O. Beswick et al. / Catalysis Today xxx (2016) xxx–xxx
3
ethanol solution of nitroarenes (0.12 M) and 1 g of m-xylene (inter-
nal standard for GC analyses). After flushing (4 times) with N under
2
agitation (∼2000 rpm) to remove air, the reactor was heated up
3
33 K and equilibrated for 5 min. Injection of 1.8 mL of the reduc-
ing agent (3 equivalents), i.e. hydrazine monohydrate, defined the
reaction start. A typical reaction was performed at T = 333 K, under
atmospheric pressure with an agitation speed of 2000 rpm. The
initial substrate to metal molar ratio of 50 was used by adjusting
the mass of catalyst introduced in the reactor. A non-invasive liq-
uid sampling system allowed the controlled withdrawal of small
3
aliquots (∼0.2 cm ) from the reactor during the reaction. In a series
of blank experiments, no conversion of the substrates was detected
in the absence of the catalyst or in the presence of the ACF support
only.
2.5. Analytical method
Fig. 1. TPD profile of Co(NO3)2/ACF ( CO;
CO2;
NO). Co loading 4.3 wt.%.
The composition of the reaction mixture was determined
using
USA). The instrument was equipped with
split/splitless injector, flame ionization detector and Elite-35MS
5% phenyl (methylpolysiloxane) capillary column (i. d. = 0.32 mm,
a
gas chromatograph (AutoSystem XL, Perkin-Elmer,
a
programmed
3
length = 30 m, film thickness of 0.25 ␮m) with m-xylene as an inter-
nal standard.
The conversion of the substrate A is defined as:
CA0 − CAt
XA (%) =
× 100
(1)
C
A
0
where CA0 and CAt represent the molar concentrations of the sub-
strate at the start of the reaction and at time t, respectively.
The yield of product B with respect to substrate A is defined as:
CB − CBt ꢀA
0
SBA (%) = C
·
At
× 100
ꢀB
(2)
Fig. 2. XRD patterns of (A) CoOx/ACF, Co loading 7.3 wt.%; commercial Co3O4 NPs
<50 nm): (B) as-received and (C) after reduction; JCPDS-ICDD reference diffrac-
tograms for (D) Co (15–0806), (E) CoO (43–1004) and (F) Co3O4 (74–1656).
− C
A
0
(
where CB0 and CBt represent the molar concentrations of the prod-
uct B at the start of the reaction and at time t, respectively; ␯A and
␯
B are the stoichiometric coefficients.
The yield towards the product B with respect to substrate A is
thermal treatments. The CO and CO2 peaks at higher temperature
(∼830–K) may be ascribed to partial reduction of metal oxide by
carbon. It is noteworthy that the position of these peaks depends
on the metal used and shifts to lower temperatures when passing
from Fe to Ni (Fig. S1A, B).
The freshly prepared catalysts were found to be pyrophoric in
contact with air which is an indication of extremely high dispersion
of reduced metal oxide nanoparticles formed on the ACF surfaces.
Therefore, a passivation step involving a flow of diluted air was
conducted as a final step.
defined as:
CB0 − CBt
YBA = XA · SBA
=
× 100
(3)
C
A
0
The catalyst activity response was evaluated measuring the ini-
tial transformation rate, −R − the number of moles of substrate
0
converted in the reactor per unit of time referred to the number of
moles of metal. The initial activity was calculated for the reaction
time t required to reach a conversion of 20% as:
3
.2. Catalyst characterization
CA0 − CAt
1
−
R0 (molA · mol− · s−1) =
· V
(4)
M
t · n
The MOx/ACF catalysts were found to be XRD inactive. Fig. 2A
where V is the volume of the reaction mixture, t is the time in
seconds and n is the number of moles of metal.
contains (as a representative example) the XRD pattern of the
CoO /ACF catalyst. The fact that no peak is detected suggests
x
that CoOx particles supported on ACF are extremely small and/or
3
. Results and discussion
are amorphous [23]. The presence of a highly dispersed phase
confirmed by STEM (see below) supports these observations. For
3.1. Catalyst preparation
comparison, commercial Co O4 nanopowder was found to consist
3
of a Co O4 phase (Fig. 2B, F), and reduction with hydrogen affords
3
The catalysts were prepared by decomposing metal oxide pre-
metallic Co (Fig. 2C, D).
cursors (nitrates) supported on ACF under an inert atmosphere.
In order to optimize the preparation conditions, TPD of the Co,
Fe and Ni nitrates supported on ACF was studied. Fig. 1 shows
that complete Co oxide precursor decomposition takes place in the
range 430–600 K. The NO peak corresponds to the decomposition
of NO3 groups. Detected CO and CO2 (1st broad peak) are pro-
duced through oxidation of the ACF support by generated nitrogen
oxides. Therefore, a temperature of 623 K was chosen for standard
A comparison of the H -TPR patterns recorded for the CoOx/ACF
2
and commercial Co O4 nanopowder (Fig. 3) revealed that reduc-
3
tion of both catalysts takes place in the same temperature region
◦
(580–640 K). The second broad peak (∼800 C) of the CoOx/ACF is
assigned to methanation of carbon. This result indicates that the
Co-active phase after catalyst passivation consists of an oxide.
It was previously reported that in XPS spectra of FeOx/ACF the
position of Fe 2p3/2 and Fe 2p1/2 peaks (∼711 eV and 725 eV, respec-
−
Please cite this article in press as: O. Beswick, et al., Highly dispersed cobalt oxides nanoparticles on activated
carbon fibres as efficient structured catalysts for the transfer hydrogenation of m-nitrostyrene, Catal. Today (2016),
http://dx.doi.org/10.1016/j.cattod.2016.06.043

G Model
CATTOD-10293; No. of Pages7
ARTICLE IN PRESS
O. Beswick et al. / Catalysis Today xxx (2016) xxx–xxx
4
Fig. 3. H2-TPR of (A) CoOx/ACF, Co loading 6.3 wt.% (after nitrates decomposition)
and (B) commercial Co3O4 NPs.
tively) and the presence of a characteristic satellite peak at ∼718 eV
indicates that Fe-based NPs in the passivated FeOx/ACF catalyst are
composed of Fe O [15]. In Co 2p XPS spectra of the synthesized
2
3
CoOx/ACF (Fig. 4) a non-symmetric Co 2p3/2 peak (780.9 eV) with
a satellite (786.0 eV) are observed which does not correspond to
metallic Co (Co 2p3/2 main peak is at 778.1 eV without any satellite)
or Co O (Co 2p3/2–779.6 eV) [24]. As-received commercial Co O4
3
4
3
NPs are characterized by a Co 2p3/2 peak at 780.1 eV. In this case, the
position and relative intensity of Co 2p3/2 satellite peak differ con-
siderably to those of the CoOx/ACF. The XPS pattern of the CoOx/ACF
is very similar to a spectrum of CoO (and/or Co(OH) ) [24].
2
The XPS spectrum of reduced Co O4 NPs appears to be a super-
3
position of spectra of different phases: Co O and CoO (and/or
3
4
Co(OH) ), presumably due to the slow oxidation of the NPs in air.
2
The low temperature N2 adsorption isotherms measured for
the ACF support and MOx/ACF catalysts have a profile of the type
I (according to the IUPAC classification) typical for microporous
materials (pore diameter <2 nm). The SSA and total pore volume
2
−1
and
(
Vp) calculated for the ACF support are ∼2170 ± 50 m g
3
−1
∼
1.01 ± 0.02 cm g , respectively. After deposition of metal oxide
(
6–10 wt.%) a small (4–6%) decrease of SSA and pore volume was
detected. The formation of extremely small metal oxide NPs can
explain an absence of micropore blocking with a negligible decrease
of SSA and Vp.
STEM-HAADF analysis of the catalysts was carried out to visu-
alize the active phase of the MOx/ACF catalysts. In a typical image
(
Fig. 5) NPs (bright contrast) with an average size of ∼2 nm homo-
geneously distributed throughout ACF (dark contrast) are observed.
Fig. 5C II acquired by TEM-EDX mapping shows that the zones of
high intensities correspond to Co-based NPs. Oxygen was found to
be distributed over the carbon support with the highest concentra-
tions at the Co locations confirming that NPs are indeed Co-oxide
compounds. The morphologies of NPs deposited on as-supplied and
Fig. 4. XPS profiles of (A) CoOx/ACF (after nitrates decomposition) (Co loading
.3 wt.%, commercial Co3O4 NPs (<50 nm): (B) as-received and (C) after reduction.
7
gen content reduces the nanoparticle density, thereby altering its
contrast in HAADF images and visible above the ACF support.
For a given nanoparticle diameter, a higher content oxygen and
lower density can reduce the signal-to-background ratio (SBR) to
the extent that they cannot be discerned over background signal,
especially when there are variations in the support thickness and
density (cf. supplementary information, S2–S7).
Such a Co oxide composition, with a higher number of oxy-
gen per cobalt atom, is presumably due to the increased amount
of O-containing groups on the carbon surface after the HNO3 pre-
treatment. These groups participate in the precursor ion exchange
during the impregnation step and act as seeds for CoOx formation
during the thermal treatment resulting in different composition
of final NPs and influencing the catalytic properties. Indeed, the
ACF pre-treatment with nitric acid was already shown to increase
tremendously the oxygen content by more than an order of mag-
HNO pre-treated ACF are different. In the latter case, the morphol-
3
ogy also varies depending on metal type. The average size of the
Ni-based NPs (∼1.4 nm) was negligibly affected by the support pre-
treatment (∼1.3 nm) whereas no Co oxide NPs could be observed
by STEM in the CoOx/ACF
(Fig. 5DI) on the FEI Tecnai Osiris,
HNO
3
despite EDX mapping shows cobalt well distributed throughout
the carbon surface (Fig. 5DII). This indicates either the formation
of NPs below the detection limit (<0.5 nm), or that there is a higher
oxygen-lattice concentration. The first hypothesis was ruled out
on the basis of the supplementary imaging of the CoOx/ACFHNO3
catalyst performed with an aberration-corrected microscope (FEI
Titan Themis). Nanoparticles ranging from ∼1–3 nm were revealed
(Fig. 5DIII).
It is possible that the Co based nanocatalysts have increased
oxygen content due to the HNO pre-treatment. The increased oxy-
3
Please cite this article in press as: O. Beswick, et al., Highly dispersed cobalt oxides nanoparticles on activated
carbon fibres as efficient structured catalysts for the transfer hydrogenation of m-nitrostyrene, Catal. Today (2016),
http://dx.doi.org/10.1016/j.cattod.2016.06.043

G Model
CATTOD-10293; No. of Pages7
ARTICLE IN PRESS
O. Beswick et al. / Catalysis Today xxx (2016) xxx–xxx
5
Fig. 5. Representative STEM-HAADF images of FeOx,/ACF (A), NiOx/ACF (B),
CoOx/ACF (C1) and Co/ACFHNO3 (D1). STEM-HAADF image of Co/ACFHNO3 obtained
on the FEI Titan Themis (D3). MOx NPs appear bright on the dark background (ACF
support). Co (C2) and O (C3) EDX images for CoOx/ACF. Co EDX for Co/ACFHNO3 (D2).
Metal loading (wt.%) for the samples A, B, C, D = 7.0, 9.8, 7.3, 5.3, respectively.
Fig. 7. Concentration vs reaction time for the CTH of p-CNB to p-CAN over FeOx/ACF
(
A), CoOx/ACF (B) and NiOx/ACF (C) ( p-CNB; p-CAN; p-chloroazobenzene;
p-chloroazoxybenzene; p-chloronitrosobenzene; aniline). Reaction conditions:
−
1
−1
T = 333 K; PH2 = 10 bar; Cp-CNB,0 = 1.2 × 10 mol l , molar ratio p-CNB: M = 50. Metal
Fig. 6. CTH of p-CNB with hydrazine: catalyst activity (solid bars) and maximum
p-CAN yield (hatched bars) over MOx/ACF catalysts. Reaction conditions: T = 333
loading (wt.%): (Fe, Co, Ni) = 6.1, 6.0, 6.2, respectively.
−
1
−1
K; PH2 = 10 bar; Cp-CNB,0 = 1.2×10 mol l , molar ratio p-CNB:M = 50. Metal loading
(
wt.%) of Fe, Co, Ni = 6.1, 6.0, 6.2, respectively.
In the case of the FeOx/ACF catalyst, the desired p-CAN prod-
uct is formed directly (Fig. 7A) and no intermediates were detected
in the course of the reaction. In contrast, intermediates (nitroso-
, azo- and azoxyarenes) were found using the CoO /ACF catalyst.
In the latter part of the reaction, these intermediates are slowly
nitude [25]. Such a high surface concentration of functionalities
leads to the formation of small nanoparticles with cobalt at higher
oxidation state.
x
converted to p-CAN (Fig. 7B). These observations are in line with
the widely accepted reaction network for nitrobenzene hydrogena-
tion [3] implying two possible routes, i.e. direct and condensation
(Scheme 1).
3
.3. Catalytic transfer hydrogenation
The MOx/ACF catalysts were tested in the CTH of p-
chloronitrobenzene (p-CNB) with hydrazine used as a model
reaction. The initial specific activities and maximum yields
obtained using Fe-, Co- and Ni-based catalysts are presented in
Fig. 6. The corresponding liquid bulk concentrations as a function
of reaction time are presented in Fig. 7.
Under the working conditions, the FeOx/ACF and CoOx/ACF cat-
alysts are characterized by ∼2 orders of magnitude higher activity
than the NiOx/ACF catalyst. Notably, the selectivity and maximum
yield of the desired p-CAN was close to 100% in all cases.
Although the full mechanism of heterogeneous CTH of
nitroarenes over transition metal oxides remains unclear, one can
affirm that H-transfer from hydrazine to the substrate in a catalytic
cycle is associated with metal transition from one oxidation state
to another (Fe(III) ↔ Fe(II), Co(III) ↔ Co(II), etc.) [26,27]. Since the
most likely redox couple for Ni is Ni(II) ↔ Ni(I), this difference may
explain the lower activity of the NiOx/ACF. We also found that com-
III
II
mercial highly crystalline Co O (Co O Co O) NPs are essentially
3
4
2
3
catalytically inactive in the CTH of nitroarenes with hydrazine. A p-
Please cite this article in press as: O. Beswick, et al., Highly dispersed cobalt oxides nanoparticles on activated
carbon fibres as efficient structured catalysts for the transfer hydrogenation of m-nitrostyrene, Catal. Today (2016),
http://dx.doi.org/10.1016/j.cattod.2016.06.043

G Model
CATTOD-10293; No. of Pages7
ARTICLE IN PRESS
6
O. Beswick et al. / Catalysis Today xxx (2016) xxx–xxx
Scheme 1. Reaction network for the reduction of nitroaromatic compounds.
CNB conversion of ∼1% was detected after 3 h reaction over the
as-received Co O NPs. Over reduced/passivated Co O NPs after
3
4
3
4
the same reaction time, however, a conversion of >75% of p-CNB was
obtained despite the increased NPs size during reduction (see Figs.
S8 and S9). This suggests that a highly dispersed mixture of Co O
2
3
and CoO phases is probably responsible for the observed activity.
It is known that hydrazine is a weak reducing agent to transform
Co-oxide to metallic Co. But to rule out a possibility of in-situ Co0
NPs formation as an active phase, the control experiments were
performed. First, CoOx/ACF catalyst was reduced in a 17% v/v H /Ar
2
mixture during one hour at 673 K and before the reaction was in-
situ activated for 1 h (T = 423 K; PH2 = 10 bar). No transformation of
p-CNB to p-CAN was observed over this pre-reduced catalyst dur-
ing the first 30 min. After exposing this catalyst to air for 16 h, the
activity rose to a level close to the usual transformation rate.
Fig. 8. CTH of m-nitrostyrene (m-NS) by hydrazine over FeOx/ACF and CoOx/ACF,
with and without HNO3 pre-treatment of the ACF support: catalyst activity (solid
bars) and maximum m-vinylaniline yield (hatched bars). Reaction conditions: 333 K;
3.4. Effect of solvent
−
1
m-NS concentration 0.032 mol l ; molar ratio m-NS:M = 20; molar ratio N2H4/m-
NS = 3. Metal loading (wt.%): (Fe/ACF, Fe/ACFHNO , Co/ACF, Co/ACFHNO ) = 6.5, 6.7,
3
3
It is known that CTH reactions are solvent sensitive [28,29].
4.3, 5.3, respectively.
In our case the choice was limited by the solubility of hydrazine
and consequently only alcohols were tested as solvents. Some-
times water was added to increase the solvent polarity. It is known
that alcohols can also serve as a hydrogen source [30]. However,
no conversion was detected in a control experiment carried out in
ethanol under the standard reaction conditions over CoOx/ACF in
the absence of hydrazine. Data for the CTH of p-CNB over CoOx/ACF
in different solvents are presented in Fig. S10. No clear dependency
of the catalyst activity on solvent properties was observed. The
highest catalytic activity was obtained in ethanol. It is worth not-
ing that the yield of the desired product was close to 100% in all the
solvents used but after different reaction times.
alysts in the CTH of m-nitrostyrene, which contains an easily
reducible vinyl-group. This compound is often taken as a model
for hydrogenation of complex nitroaromatics with multiple sub-
stituted groups in which the preservation of the C C double bond
during reduction is the main concern. The results presented in Fig.
8 demonstrate that FeOx/ACF is more active than CoOx/ACF in the
CTH of m-nitrostyrene using hydrazine, but is less selective toward
desired m-VA product.
Considerable improvements in the yield of m-VA were obtained
using ACF pre-treated by HNO3 as supports for the NPs catalysts
(Fig. 8). The pre-treatment was found to give more active Fe and
The excellent results obtained with the FeOx/ACF and CoOx/ACF
catalysts for p-CNB hydrogenation prompted us to test these cat-
Please cite this article in press as: O. Beswick, et al., Highly dispersed cobalt oxides nanoparticles on activated
carbon fibres as efficient structured catalysts for the transfer hydrogenation of m-nitrostyrene, Catal. Today (2016),
http://dx.doi.org/10.1016/j.cattod.2016.06.043

G Model
CATTOD-10293; No. of Pages7
ARTICLE IN PRESS
O. Beswick et al. / Catalysis Today xxx (2016) xxx–xxx
7
Co-based ACF catalysts. The highest m-VA selectivity and yield of
References
9
4
1
9.4% was obtained over the CoOx/ACFHNO3 .
[
[
[
1] T. Kahl, K.-W. Schröder, F.R. Lawrence, W.J. Marshall, H. Höke, R. Jäckh,
Aniline, in Ullmann’s Encyclopedia of Industrial Chemistry, Wiley-VCH Verlag
GmbH & Co. KGaA, 2000.
2] J. Hartwig, S. Shekar, Q. Shen, F. Barrios-Landeros, Z. Rappoport, The
Chemistry of Anilines, John Wiley & Sons, Inc, West Sussex, England, 2007
. Conclusions
Structured catalysts featuring highly-dispersed non-precious
metal (Fe, Co, Ni) oxide NPs stabilized within super-
[
CrossRef] There is no corresponding record for this reference.
a
3] H.-U. Blaser, H. Steiner, M. Studer, Chemcatchem 1 (2009) 210–221.
microporous network of activated carbon fibres (ACF) have been
developed and used for chemoselective reduction of substituted
nitroarenes using hydrazine as hydrogen transfer agent. The cat-
alyst morphology was controlled over multiple levels from the
nano-designed active phase up to the macro-structure of the ACF
support.
The main advantages of the ACF structured supports include
the facile formation of extremely small (<2 nm) NPs and their
stabilization within a micro-porous network which prevents
their agglomeration. Moreover, the micro-sized diameters of the
carbon filaments in combination with their slit-type short micro-
pores eliminate any mass-transfer limitations allowing a high
selectivity.
[4] P.L. Gkizis, M. Stratakis, I.N. Lykakis, Catal. Commun. 36 (2013) 48–51.
[
5] R.J. Jagadeesh, D. Banerjee, P. Arockiam, H. Junge, K. Junge, M.-M. Pohl, J.
Radnik, A. Brueckner, M. Beller, Green Chem. 17 (2014) 898–902.
6] M.B. Gawande, H. Guo, A.K. Rathi, P.S. Branco, Y. Chen, R.S. Varma, D.-L. Peng,
RSC Adv. 3 (2013) 1050–1054.
[
[
[
[
7] D.R. Petkar, B.S. Kadu, R.C. Chikate, RSC Adv. 4 (2014) 8004–8010.
8] D. Wang, D. Astruc, Chem. Rev. 115 (2015) 6621–6686.
9] Y.-M. Lu, H.-Z. Zhu, W.-G. Li, B. Hu, S.-H. Yu, J. Mater. Chem. A 1 (2013)
3783–3788.
2
3
[10] N.R. Guha, D. Bhattacherjee, P. Das, Tetrahedron Lett. 55 (2014) 2912–2916.
11] Z. Zhao, H. Yang, Y. Li, X. Guo, Green Chem. 16 (2014) 1274–1281.
12] S.K. Mohapatra, S.U. Sonavane, R.V. Jayaram, P. Selvam, Tetrahedron Lett. 43
[
[
(
2002) 8527–8529.
[13] A. Yarulin, C. Berguerand, I. Yuranov, F. Cardenas-Lizana, I. Prokopyeva, L.
Kiwi-Minsker, J. Catal. 321 (2015) 7–12.
14] C. Berguerand, A. Yarulin, F. Cardenas-Lizana, J. Warna, E. Sulman, D.Y.
Murzin, L. Kiwi-Minsker, Ind. Eng. Chem. Res. 54 (2015) 8659–8669.
[
The challenging reduction of m-nitrostyrene containing an eas-
ily reducible vinyl-group was achieved under mild conditions
[15] O. Beswick, I. Yuranov, D.T.L. Alexander, L. Kiwi-Minsker, Catal. Today 249
2015) 45–51.
(
[
[
16] Y. Matatov-Meytal, M. Sheintuch, Appl. Catal. a-Gen. 231 (2002) 1–16.
17] S.R. de Miguel, J.I. Villella, E.L. Jablonski, O.A. Scelza, C.S.M. de Lecea, A.
Linares-Solano, Appl. Catal. a-Gen. 232 (2002) 237–246.
(
333 K, 1 atm) with a near-quantitative yield (99.4%) of m-
vinylaniline using hydrazine hydrate and the CoOx/ACFHNO3
catalyst.
[
[
18] I.M. Vilella, S.R. de Miguel, O.A. Scelza, Chem. Eng. J. 114 (2005) 33–38.
19] M.C.M. Perez, C.S.M. deLecea, A.L. Solano, Appl. Catal. A: Gen. 151 (1997)
4
61–475.
The potential applications of this catalytic system in batch as
[20] F. Cardenas-Lizana, B. Bridier, C.C.K. Shin, J. Perez-Ramirez, L. Kiwi-Minsker,
ChemCatChem 4 (2012) 668–673.
21] Y. Marco, E. Garcia-Bordeje, C. Franch, A.E. Palomares, T. Yuranova, L.
Kiwi-Minsker, Chem. Eng. J. 230 (2013) 605–611.
well as in flow-mode are considerable. Therefore, we anticipate
that the catalysts presented in this study will offer new approaches
for the synthesis of aryl amines containing C C double bonds.
[
[22] S. Brunauer, P.H. Emmett, E. Teller, J. Am. Chem. Soc. 60 (1938) 309–319.
[
[
23] J. Zhu, K. Kailasam, A. Fischer, A. Thomas, ACS Catal. 1 (2011) 342–347.
24] M.C. Biesinger, B.P. Payne, A.P. Grosvenor, L.W.M. Lau, A.R. Gerson, R.S. Smart,
Appl. Surf. Sci. 257 (2011) 2717–2730.
Acknowledgment
[
[
[
25] G.B. Baur, O. Beswick, J. Spring, I. Yuranov, L. Kiwi-Minsker, Adsorption 21
(2015) 255–264.
26] P.S. Kumbhar, J. Sanchez-Valente, J.M.M. Millet, F. Figueras, J. Catal. 191 (2000)
We thank the EPFL and the Russian Science Foundation (Grant
5-19-20023) for financial support.
1
467–473.
27] M. Benz, A.M. van der Kraan, R. Prins, Appl. Catal. a-Gen. 172 (1998) 149–157.
Appendix A. Supplementary data
[28] R.A.W. Johnstone, A.H. Wilby, I.D. Entwistle, Chem. Rev. 85 (1985) 129–170.
[
29] D.B. Zhao, Z.F. Fei, T.J. Geldbach, R. Scopelliti, P.J. Dyson, J. Am. Chem. Soc. 126
2004) 15876–15882.
30] J. Toubiana, Y. Sasson, Catal. Sci. Technol. 2 (2012) 1644–1653.
(
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.cattod.2016.06.
[
043.
Please cite this article in press as: O. Beswick, et al., Highly dispersed cobalt oxides nanoparticles on activated
carbon fibres as efficient structured catalysts for the transfer hydrogenation of m-nitrostyrene, Catal. Today (2016),
http://dx.doi.org/10.1016/j.cattod.2016.06.043