Coprecipitated-like hydrotalcite-derived coatings on open-cell metallic foams by electrodeposition: Rh nanoparticles on oxide layers stable under harsh reaction conditions

Article abstract of DOI:10.1016/j.apcata.2018.04.014

Structured catalysts based on open-cell metallic foams coated by a catalytic film offer a great potential for intensification and optimization of catalytic processes. Here, we demonstrated the feasibility of the electrodeposition to synthesize in situ and

Full text of DOI:10.1016/j.apcata.2018.04.014

AppliedCatalysisA,General560(2018)12–20
Contents lists available at ScienceDirect
Applied Catalysis A, General
journal homepage: www.elsevier.com/locate/apcata
Coprecipitated-like hydrotalcite-derived coatings on open-cell metallic
foams by electrodeposition: Rh nanoparticles on oxide layers stable under
harsh reaction conditions
Phuoc Hoang Hoa,b, Wout de Nolfc, Francesca Ospitalia, Angela Gondolinid, Giuseppe Fornasaria,
Erika Scavettaa, Domenica Tonellia, Angelo Vaccaria, Patricia Benitoa,⁎
a
Dipartimento di Chimica Industriale “Toso Montanari”, Università di Bologna, Viale Risorgimento 4, 40136, Bologna, Italy
Institut für Technische und Makromolekulare Chemie, RWTH Aachen University, Worringerweg 2, 52074, Aachen, Germany
European Synchrotron Radiation Facility, 71 Avenue des martyrs, 38000, Grenoble, France
Institute of Science and Technology for Ceramics (ISTEC) of the National Research Council (CNR), Via Granarolo 64, I-48018, Faenza, RA, Italy
b
c
d
A R T I C L E I N F O
A B S T R A C T
Keywords:
Structured catalysts based on open-cell metallic foams coated by a catalytic film offer a great potential for
intensification and optimization of catalytic processes. Here, we demonstrated the feasibility of the electro-
deposition to synthesize in situ and quick Rh/Mg/Al hydrotalcite-type (HT) syngas catalyst precursors with
controlled composition, morphology and thickness around 5 to 20 μm on the surface of FeCrAlloy foams using a
two-compartment flow electrochemical cell. After calcination at 900 °C, catalytic coatings with properties si-
milar to those of conventional co-precipitated HT-derived catalysts were identified by synchrotron nano-XRF/
XRD tomography and HRTEM. The resulting structured catalysts, therefore, merged the properties of both HT-
derived catalysts and open-cell foams, namely, thermally stable nano MgO- and spinel-type phases where Rh was
dispersed and stabilized against sintering, and high mass and heat transfer. Moreover, the development of a
MgAl2O4 thin film in the support-coating interface, by chemical reaction between Mg2+ from the coating and
Al3+ from the support during calcination, increased the catalyst adhesion. Consequently, active and stable
performance was obtained under harsh reaction conditions in the catalytic partial oxidation of CH4 to syngas as a
model reaction. Even in the catalysts operating under severe reaction conditions for about 50 h, the coating was
stable and Rh metallic nanoparticles around 2 nm were still well dispersed.
Electrodeposition
Hydrotalcite
rhodium
FeCrAlloy foam
Structured catalyst
Syngas
1. Introduction
synthesis [3,5,6], and CO2 methanation [4,7]. For instance, in the
exothermic and fast catalytic partial oxidation (CPO) of CH4, the
Structured catalysts made of a 3D support coated by catalytic spe-
cies have gained increasing attention since their initial application in
automotive exhaust gas treatment [1,2]. The development of supports
of different shapes (open-cell foams, fibers), materials (ceramic and
metallic, including high corrosion resistant alloys) as well as experi-
mental and modelling studies demonstrated the ability of these type of
materials to enhance mass and heat transfer and decrease pressure
drop, making them the best option for catalytic processes operating at
high gas hourly space velocities, and/or strongly endo-/exothermic
[1,2]. In particular, open-cell metallic foams provide a disruptive and
tortuous flow path and hence a high flow mixing as well as supply or
release of the heat to/from the surface in comparison to honeycomb
monoliths, showing advantages in H2 and syngas production by re-
forming and partial oxidation [3,4], CH3OH and Fischer-Tropsch
pressure drop and formation of hot spots can be decreased with struc-
tured catalysts in comparison with their pelletized counterparts, thus
enhancing the activity and stability of the catalysts [8–11].
The first step in the development of structured catalysts is the de-
position of a highly active catalytic film, with suitable thickness,
homogeneity and adhesion to the support. The stability of the coating
determines the catalytic activity and in metal-based structured catalysts
also the resistance to oxidation of the metal supports [12–14]. Hydro-
talcite-type (HT) compounds, layered materials with general formula
3+
[M2+
M
(OH)2](An−
)
mH2O are widely used as catalyst pre-
1−x
x
x/n
cursors, and in particular for syngas production processes [15]. Hence,
they were applied as catalytic coatings on honeycombs, open-cell
foams, and fibers for the steam reforming [16], oxy-reforming [17], dry
reforming [18], partial oxidation [19,20], and ethanol reforming [21].
⁎
Corresponding author.
E-mail address: patricia.benito3@unibo.it (P. Benito).
https://doi.org/10.1016/j.apcata.2018.04.014
Received 2 February 2018; Received in revised form 1 April 2018; Accepted 15 April 2018
Availableonline17April2018
0926-860X/©2018ElsevierB.V.Allrightsreserved.

P.H. Ho et al.
AppliedCatalysisA,General560(2018)12–20
The well-known dip-coating of catalyst slurries or the in situ pre-
the mass transfer in the preparation of ZnO over Cr and Au electrodes
[46] and Cu on 3D graphite felts [47] by replenishing the solution in-
side the felt.
cipitation of HT compounds have been proposed to prepare structured
catalysts. The former method can deposit large catalyst loadings on
metallic monoliths (ca. 20 wt%) through long procedures adding ad-
ditives to the coating, which can modify the catalyst and decrease
catalytic performance [22]. The washcoating of foams is more chal-
lenging: to avoid pore clogging, low density slurries are used, de-
creasing the catalyst loading on SiC foams to 1–4 wt% [23]. On the
other hand, vertically aligned HT platelets can directly grow on the
structured support, i.e. a Ni foam, through homogeneous urea pre-
cipitation under hydrothermal conditions [17]. For anodized Al sup-
ports, i.e. Al2O3/Al, the substrate is the Al source and, besides urea
[24,25], NH3 [26,27] can be used as a precipitating agent. In these
cases, the adhesion of the HT compounds vertically grown on the sur-
face of the supports is improved. However, compared to coprecipitated
catalysts, the structured ones show worse reducibility and lower cata-
lytic activity in the total oxidation of ethanol, due to the formation of
spinel-like phases, high structural ordering of the products deposited on
the Al2O3/Al supports, and differences in composition [25]. Moreover,
in these methods the precipitation of HTs can take place not only on the
surface of the structured support but also in the bulk of the solution,
which is a drawback especially for noble metal-containing catalysts.
The electrodeposition, specifically the electro-base generation
method, is “ideally” the best way to coat metallic supports in situ with
HT-type materials [28,29]. Unlike the above commented precipitation
methods, the pH increases quickly and near the surface of the support
due to the reduction of nitrate, oxygen and water by application of a
cathodic potentiostatic pulse [30]; therefore, the solid is selectively
precipitated on the support in a short time. Thus, we proposed this
technique to synthesize HT compounds in situ on FeCrAlloy open-cell
foams to prepare structured catalysts for the catalytic partial oxidation
(CPO) of CH4 to syngas [19]. However, the control of the composition is
challenging, as nominal and real compositions deviate. For instance,
during the preparation of thick Rh/Mg/Al HT coatings (10–15 μm), we
observe the sequential precipitation of layers with different composi-
tions as the synthesis time proceeds, which modifies composition,
crystalline phases, Rh reducibility and metal particle size and, conse-
quently, film stability and catalytic performance [31,32].
The challenging preparation of these Rh/Mg/Al HT films can be
related to the formation of thick and insulating layers, and the in-
efficient replenishment of the electrolytic solution on the surface of the
electrode at long times [30]. Although differences in the precipitation
pH, diffusion coefficients of the cations and the shape of the support
could also play a role in this phenomenon. Indeed, deviations in the
expected composition of Ni/Co- [33–35] and Fe-containing compounds
[36–38] deposited on Ni foams are reported notwithstanding thin films
are prepared at short times. To the best of the authors’ knowledge, only
one article reports the coating by Fe-Co hydroxides of Ni foams with
controlled composition by continuous electrodeposition at a constant
current density of 10 mA cm−2 for 1800 s [39].
In porous supports, such as open-cell foams, limited mass transfer
inside the pores, and generation of polarization potentials or changes in
the local current density do not guarantee a uniform composition and
thickness of deposits over the complete geometry [40,41]. Electro-
depositions performed under forced convection, by pumping or agita-
tion, and through pulse depositions, with stirring between the pulses,
can improve the properties of the coating. The mechanical stirring re-
plenishes the electrolytic solution and decreases the thickness of the
diffusion layer during the electrodeposition [42,43]; however, in the
coating of complex shaped supports an inhomogeneous flow rate and
hence strongly scattering current density can be generated, creating
inhomogeneity in the coating [44]. The use of short electrochemical
pulses, allowing sufficient time (aided by stirring) between depositions,
replenishes the local ion concentration creating a more homogeneous
Fe distribution in Ni-Fe/Oxihydroxide on a Ni foam as well as mod-
ifying the texture of the film [45]. Finally, the use of flow cells improves
The aim of this work was to control the electrodeposition of Rh-
containing HT compounds on the surface of small pore open-cell me-
tallic foams. The final goal was to obtain, after calcination, a stable
coating with controlled properties (i.e. Rh loading, thickness, Rh me-
tallic particle size and dispersion) not possible until now, to further
improve the activity and stability of structured catalysts for syngas
production. To achieve that, the electrodepositions were performed
under flow in a homemade flow-cell double-compartment electro-
chemical set-up. To study the feasibility of the procedure, two Rh/Mg/
Al samples with different Rh loading (2 and 5% as atomic ratio) were
prepared, like in our previous work [32], the high Rh loadings were
used to balance the low amount of catalytic coating. The catalysts were
tested in the CPO of CH4, and characterized at the different stages of
their life-time by XRD, electron (FE-SEM, HRTEM) and synchrotron
combined nano-XRF/XRD tomography, which allowed to study the
distribution of the elements and crystalline phases in the coating and
the foam support in the nano-scale rather than in the micro-scale pre-
viously reported by us [12,13].
2. Experimental section
2.1. Catalyst preparation
All chemicals were purchased from Sigma Aldrich: Mg(NO3)2·6H2O
(> 99%), Al(NO3)3·9H2O (> 98%), Rh(NO3)3 solution (10 wt% Rh in
HNO3). FeCrAlloy open-cell foam cylinders (80 ppi, pores per inch, and
10 mm × 11.9 mm), cut from a panel, were used as structured supports.
Electrodepositions were performed in a homemade double-com-
partment flow electrochemical cell by FAVS Gnudi, using a potentiostat
(Autolab, PGSTAT128N, Eco Chemie) with GPES software. A Pt coil
(0.4 mm diameter and 40 cm in length) and a saturated calomel elec-
trode (SCE) were used as counter and reference electrode (C.E. and
R.E.), respectively. The working electrode was the FeCrAlloy foam in
the cylindrical shape and it was assembled by a three-pronged Pt
electrical contact as described elsewhere [31]. The working and counter
electrode compartments were separated by a glass frit tube to avoid the
mixture of the electrochemical reaction products. The reference elec-
trode was in electrolytic contact with the main compartment via a
Luggin capillary placed close 1 mm to the surface of the foam cylinder.
All potentials were reported with respect to SCE. Prior to coating, the
foam was subsequently washed in acetone, water and then dried at
60 °C for 24 h.
The electrodepositions were performed by circulating the electro-
lyte during the coating process in the working electrode chamber with a
flow rate of 2 mL min−1 by a peristaltic pump. Preliminary studies in-
dicated that a lower (or no) or higher flow gave poorer results in terms
of coating properties and catalytic activity. The synthesis conditions
(potential applied, time and composition of the electrolytic solution)
were like those used in depositions performed in the conventional
single-compartment cell, i.e. a −1.2 V vs SCE pulse was applied for
2000 s to the foam immersed in an aqueous metal nitrate solution
(0.06 M) containing Rh/Mg/Al atomic ratio (a.r.) 5/70/25 or 2/70/28
[32], named Rh5 or Rh2 respectively. It should be noted that in these
electrolytes, the set-up reached a low uncompensated resistance (Ru) ca.
1–2 Ω in all of experiments. After electrodeposition, the coated foams
were thoroughly washed with distilled water and dried at 40 °C for
24 h. The samples were then calcined at 900 °C for 12 h with a
10 °C min−1 heating rate.
2.2. Characterization techniques
Scanning electron microscopy (SEM) coupled to energy dispersive
spectrometry (EDS) was performed by using an EP EVO 50 Series
13

P.H. Ho et al.
AppliedCatalysisA,General560(2018)12–20
Instrument (EVO ZEISS) equipped with an INCA X-act Penta FET®
Precision EDS microanalysis and INCA Microanalysis Suite Software
(Oxford Instruments Analytical) to provide images of the spatial var-
iation of elements in a sample. The accelerating voltage was 20 kV and
the spectra were collected in duration 60 s. As-prepared deposited,
calcined and spent coated foams were analyzed in 10–15 regions of
interest. To better investigate the thickness and composition of the
coating, foams were cut vertically and embedded in an epoxydic resin
(Kit EpoFix, Struers) under vacuum, then mechanically polished till
one-micron finish and coated with Au.
Field emission scanning electron microscopy (FE-SEM) analyses
were conducted on a Zeiss LEO Gemini 1530 equipped with an
Everhart-Thornley (E–T) secondary electron detector and a Scintillator
BSE detector – KE Developments CENTAURUS. The accelerating vol-
tages were 5 or 10 kV.
crystal orientations around the rotation axes are measured when rota-
tion is continuous as opposed to several discrete orientations when
rotation is step-wise), an imperfect rotation and normalization for the
changing X-ray flux caused artefacts. Nevertheless, the tomographic
reconstructions gave a good indication on the chemical composition of
metal support and coating with 200 nm nominal resolution.
2.3. Catalytic partial oxidation of CH4 experiments
CPO tests were conducted at atmospheric pressure using a quartz
reactor (i.d. 10.0 mm) with one foam cylinder. The foam fitted well
with the wall of the reactor to minimize any by-passing. Before carrying
out the tests, the catalyst was reduced in situ with 7.0 L h−1 H2/N2 (1/
1) at 750 °C for 2 h.
The activity of the catalysts was evaluated by modifying the reac-
tion conditions (Table S1), i.e. setting up the oven at 750 and 500 °C,
modifying the gas hourly space velocity (GHSV) from 126,000 to
23,000 h−1 and feeding different concentrations of reagents (CH4/O2/
N2 = 2/1/20 and 2/1/4 v/v). GHSV were calculated considering the
total foam volume, much higher values will be obtained considering
only the catalytic coating, e.g. 9.81*106−2.325*106 h−1. Reaction
conditions either feeding diluted gas mixtures (CH4/O2/N2 = 2/1/
20 v/v) or at a low oven temperature reduced the amount of heat de-
veloped by the exothermic reactions, making it possible to highlight the
differences among catalysts. On the other hand, tests with a con-
centrated gas mixture (CH4/O2/N2 = 2/1/4 v/v) and high oven tem-
perature (750 °C) were used to study the stability of the catalysts
against deactivation by sintering, coking and oxidation. Control tests
were carried out after each reaction condition by feeding the initial one
(CH4/O2/N2 = 2/1/20 v/v, 126,000 h−1) to check deactivation/acti-
vation processes.
The X-ray diffraction (XRD) analysis of the film grown on the foam
was carried out using a PANalytical X’Pert diffractometer equipped
with a copper anode (λmean = 0.15418 nm) and a fast X’Celerator de-
tector. Wide-angle diffractogram was collected over 2θ range from 5 to
70° with a step size of 0.07° and counting time 60 s.
High resolution transmission electron microscopy (HRTEM) char-
acterization was carried out by a TEM/STEM FEI TECNAI F20 micro-
scope, equipped with an EDS analyzer. Powder catalysts were collected
by scratching the powder from the foam surface and then suspending it
in ethanol under ultrasounds for 20 min. The suspension was subse-
quently deposited on a Cu grid with lacey quanti-foil carbon film and
dried at 100 °C before doing the measurement. Fast Fourier
Transformation (FFT) was applied to determine the interplanar spacing
of the crystals. Particle size distribution was processed considering
around 150 particles in three different zones for each sample.
Micro-Raman spectra were recorded with a Renishaw Raman Invia
spectrometer configured with a Leica DMLM microscope using Ar+
laser source (λ = 514.5 nm, Pout = 30 mW considering the decrease in
power due to the plasma filter). In each measurement, the laser power
was set by 10% of the source and the signal was accumulated by 4
individual spectra with an acquisition time of 10 s.
Combined nano X-ray Fluorescence (n-XRF) and nano X-ray Powder
Diffraction (n-XRD) measurements in tomographic mode were per-
formed at the ID16b beamline of the European Synchrotron Radiation
Facility (ESRF, Grenoble, France). The polychromatic beam from the
undulator source was monochromatized to 29.6 keV using a Si (111)
double crystal monochromator and subsequently focused on the sample
by means of a Kirkpatrick-Baez mirror system, achieving a spot size of
80 × 70 nm (h × v). Two 2 × 3-element silicon drift detectors (KETEK
GmbH D) collected the fluorescent radiation produced by the irradiated
The stability of the best catalysts was investigated through three
reaction cycles including shut-down to room temperature and start-up
to the reaction temperature, between them maintaining an inert flow of
N2. Each cycle consisted of two reaction conditions: i) Toven = 500 °C,
CH4/O2/N2 = 2/1/20 v/v, GHSV = 30,500 h−1
; ii) Toven = 750 °C,
CH4/O2/N2 = 2/1/4 v/v, GHSV = 30,500 h−1
.
The effluent products were analysed on-line after water condensa-
tion by a PerkinElmer Autosystem XL gas chromatograph, equipped
with two TCDs and two Carbosphere columns using He as a carrier gas
for CH4, O2, CO, and CO2 measurements, and N2 for H2 analysis.
Oxygen conversion was complete in all tests. CH4 and O2 conversion
and the selectivity in H2 and CO were calculated as reported elsewhere
[50].
specimen under 90° from the incoming X-rays.
A
Frelon 2 K
3. Results and discussion
(51 × 51 μm2) detector positioned behind the sample was used to
register the Debye rings produced by interference of the X-ray photons
elastically scattered by the sample. For the tomographic experiments,
samples were prepared by isolating individual struts (about 50 μm)
from the foams and gluing them to the tip of glass capillaries, allowing
an easy rotation in the beam. The sample moved horizontally with a
0.2 μm step size and was rotated over 180° with a 1.6–2.4° interval,
depending on the sample. The data treatment was performed as re-
ported elsewhere using the Spectrocrunch software for the analysis of
the XRF signals [48] and the XRDUA software for the analysis of the
XRD data [13,49].
The high primary beam energy of 29.6 keV, necessary to measure
diffraction in transmission, combined with the measurement performed
in air, caused the detection limits for low Z elements like Mg and Al to
be poor due to the low absorption cross-section and air absorption re-
spectively. Artefacts in the tomographic reconstruction (visible as rings
or arcs) were primarily caused by the decision to use continuous rota-
tion scanning and step-wise translation (as opposed to continuous
translation and step-wise rotation as used in previous work). Although
this scanning mode improved the particle statistics in diffraction (all
3.1. Electrodeposited hydrotalcite-type compounds
Fig. 1a–d depicted SEM images of the surface and cross-sections of
foams coated with Rh5 electrolytic solution; similar results were ob-
tained for Rh2 samples (Fig. S1), while Fig. 1e showed optical images of
the foam after coating. A 4.6 wt% catalyst precursor deposited on the
surface of the foams (both outer surface and strut cavities) regardless of
the nominal Rh content. The layer thickness, estimated from the ana-
lysis of cross-section images, ranged between 5–20 μm on the outer
surface of the struts and nodes connecting the struts; the thinner coat-
ings were found inside the foam cylinder structure. Very thin layers
(less than 1.5 μm) coated the cavities within the struts. The morphology
and size of the particles was not modified with the flow, unlike as re-
ported for the deposition of ZnO in an impinging cell, probably due to
the low flow rate [46]. Platelet-like and spherical agglomerates of na-
noparticles (100–400 nm) connected in a gel-like manner (Fig. 1a, b),
characteristic of electrodeposited materials [32], were observed.
Average deposit compositions (Rh/Mg/Al atomic ratio) were ob-
tained from EDS analyses of the surface of at least 5 foams, the results
14

P.H. Ho et al.
AppliedCatalysisA,General560(2018)12–20
Fig. 1. SEM images and composition of Rh5 catalyst precursor: a) medium magnification; b) high magnification; c and d) cross-section of an embedded foam; e) an
overview of one coated foam; f) XRD pattern of the fresh coating, and the inset Table showed the compositions (identified by EDS) of selected area and average
values. The numbers in the table correspond to the regions identified in figures a, c, d.
are summarized in table in Fig. 1. Rh, Mg and Al evenly distributed in
the coating, a feature also confirmed by EDS analysis performed in
cross-sections (Fig. 1c, d), and the estimated composition was close to
the nominal one. However, in a few points, Mg-rich and Al-rich layers
were identified, the former in films with platelet-like particles and the
latter in layers thicker than 20 μm.
nominal Rh loading, close to the ones used in conventional powder
catalysts [51]. The morphology and size of the catalyst particles were
not largely altered by the thermal treatment and the film showed a
small number of cracks and detachments. The stability against sintering
of the particles may be related to the properties of HT-derived catalysts;
it is well-known that the particle shape and size of HT compounds was
maintained after calcination [52] and relatively high surface areas can
be kept even after calcination at 900 °C [15]. To explain the enhanced
adhesion of the film, the absence of the Rh- and Al- rich layer, which
easily detached such as reported in our previous work [32], and the
interaction between support and coating elements should be considered
[13]. The latter involved the chemical reaction between Mg2+ from the
coating and Al3+ from the foam giving rise to a MgAl2O4 phase. During
calcination, the HT decomposed, and the foam was oxidized; at around
500 °C reactive MgO-type and γ-Al2O3 phases developed in the coating
and foam surface respectively, which could react at higher tempera-
tures developing the spinel phase. SEM/EDS analysis of cross-sections
confirmed an enrichment in Al of the coating near to the metal support,
more remarkable in thin films (Fig. 2c).
To better study the distribution of crystalline phases in the struc-
tured catalysts, combined nano-XRF/XRD tomography analyses using
synchrotron radiation were performed on selected struts of Rh5 and
Rh2 catalysts. Struts coated with thick Mg/Al and Mg-rich layers as well
as thin layers were analyzed (Fig. 3). The chemical composition of
metal support and coating was obtained from tomographic re-
constructions with a 200-nm nominal resolution, while the detection
limits for low Z elements, like Mg and Al, were poor due to the low
absorption cross-section and air absorption respectively.
Using averaged diffraction patterns, the support was shown to
consist of Fe-Cr and chromium-iron carbide phases. Due to the large
domain sizes with respect to the nanometric X-ray beam, the support
can only be visualized by use the Fe and Cr fluorescence which showed
a homogeneous Fe distribution while Cr was enriched in certain areas.
In the coating, the crystalline phases periclase, MgO, and spinel,
MgAl2O4, were identified, characteristic of HT compounds calcined at
900 °C [15]. Their domain sizes were small enough to form Debye rings
in the diffraction pattern and hence their spatial distribution could be
visualized. The distribution of the phases depended on the coating
composition. In Mg/Al thick layers, both MgO and MgAl2O4 phases
distributed in the coating (Fig. 3a), but the concentration of the spinel
phase increased close to the foam surface. In Mg-rich layer (Fig. 3c), the
spinel phase formed a thin layer of 0.2–2 μm close to the foam surface,
XRD patterns of the as prepared coated foams showed the char-
2−
acteristic reflections of HT compounds intercalated with CO3
−
(Fig. 1f), while no evidence of NO3 intercalation and side phases
formation were found [30].
The above results suggest that it was possible to coat cylinders of
FeCrAlloy open-cell foams (80 ppi) by the electro-base generation
method with HT compounds of rather controlled composition and
crystalline phases whenever the layer thickness is below 20 μm. This
behavior was likely related to the separation of the anodic and cathodic
compartments and the forced flow through the foam cylinder. The latter
favors the replenishment of the electrolyte near the foam, although
surfaces placed on the inner part of the cylinders were still less coated
than those in the outer ones, probably because a part of the electrolyte
bypassed the foam in the cell, and the slowest flow in the inner part of
the cylinder. The imperfect control of the flow may explain the devia-
tions in the composition at a few points of the surface, although the
effect of the formation of an insulating layer and some potential gra-
dients could not be discarded. The current profiles generated during the
electrodeposition were identical for both Rh5 and Rh2 electrolytic so-
lutions (Fig. S2). The shape of the profiles was very similar to the one
performed with nitrate bath containing only Mg2+ and Al3+ (total
metal cations 0.06 M, Mg/Al = 3/1 as a.r.) in the single compartment
electrochemical cell [30]. This indicated that the electrochemical pro-
cesses may proceed in the same way regardless of Rh content in the
electrolyte and as a result the in situ precipitation occurred under si-
milar conditions. This explained why the quality of the coating was
similar in both Rh5 and Rh2 catalysts.
3.2. Structured catalysts
The control in the deposition of the HT precursors influenced the
properties of the coating both after calcination at 900 °C for 12 h
(Fig. 2a–d) and catalytic tests (vide infra). A ca. 40% weight loss of HT
precursors occurred during calcination, thus the total Rh loadings in the
structured catalyst, considering also the foam weight, were estimated to
be ca. 0.26 and 0.12 wt%, for Rh5 and Rh2 respectively considering the
15

P.H. Ho et al.
AppliedCatalysisA,General560(2018)12–20
Fig. 2. SEM images and composition of Rh5
calcined catalyst: a and b) images at medium
and high magnification; c and d) images of
cross-section of one embedded foam. The inset
of figure c showed the EDS spectrum of se-
lected point (green dot) in figure c (For inter-
pretation of the references to colour in this
figure legend, the reader is referred to the web
version of this article).
while the periclase phase formed a thicker layer up to 10–15 μm on top.
In thin films, such as those in the inner cavities of the struts (Fig. 3a, c)
and in some struts inside the foam cylinders (Fig. 3b), only the spinel
phase developed. These results confirmed the solid-state reaction be-
tween Mg2+ from the coating and Al3+ from the foam during calci-
nation.
or periclase lattice parameters or identification of other Rh containing
phases). This was most likely due to the low Rh loading [13]. Spatially
Rh was colocalized with periclase (Fig. 3a–c) or with spinel in its ab-
sence (Fig. 3b). This was characteristic of coprecipitated HT-derived
catalysts Rh-MgO or Rh-MgAl2O4.
The Rh loading did not have a characteristic effect on coating
thickness or crystalline composition. No Al2O3 phases, expected to de-
velop due to the oxidation of the foam, could be detected, probably due
to the coating support-interaction. The Rh distribution was visible due
to its fluorescence but no direct indication of its inclusion in a crys-
talline structure could be derived from diffraction (e.g. change in spinel
3.3. Catalytic partial oxidation of CH4
The performances of Rh5 and Rh2 structured catalysts in the CPO of
CH4 to syngas were evaluated with tests at 750 and 500 °C oven tem-
perature, concentration of the reaction mixture (CH4/O2/N2 = 2/1/20
and 2/1/4 v/v) and GHSV (126,000–23,000 h−1). Average CH4
Fig. 3. XRF and XRD distributions on the catalysts. a) Rh2; b) Rh5 (thin coating); c) Rh5 (Mg-rich thick coating).
16

P.H. Ho et al.
AppliedCatalysisA,General560(2018)12–20
Fig. 4. CH4 conversion over Rh5 and Rh2 cat-
alysts at several reaction conditions. The data
correspond to the average of four and three
catalytic tests over different samples for Rh5
and Rh2, respectively. A longer time was re-
quired to reach stable performances in the in-
itial reaction condition for Rh2, the 0 in the x-
scale only corresponded to Rh5 catalyst.
conversion, H2 and CO selectivities and H2/CO ratio values, obtained
over four and three different Rh5 and Rh2 coated cylinders, respec-
tively, were summarized in Fig. 4 and Table S1 to assess the reprodu-
cibility of the performance.
Moving from 30,500 to 126,000 h−1 a higher amount of heat developed
inside the catalytic bed (higher inlet temperature), and the effect of the
increase in the GHSV, short contact time, was not remarkable on the
activity, producing a H2/CO = 2.0 syngas. A H2-rich gas was instead
obtained at small GHSV (H2/CO = 2.4) due to the contribution of the
water gas shift (WGS) reaction.
The catalysts obtained from Rh5 HT-tailored coatings outperformed
our previous structured catalysts in terms of activity and stability [32] -
also considering that here tests were performed at doubled GHSV (same
gas flow but using one instead of two foam cylinders) - and performance
was close to those of coprecipitated powder catalysts [15]. At 750 °C,
stable, high and reproducible conversion and selectivity were reached
after a short activation period in the initial reaction condition (CH4/O2/
N2 = 2/1/20 v/v, 126,000 h−1), probably due to the reduction of
hardly reducible Rh3+ species identified by μ-XRF/XANES in our pre-
vious work [32]. In diluted reaction conditions, conversion steadily
increased up to 92% by decreasing GHSV due to a lower amount of CH4
to be converted. A syngas with H2/CO molar ratio around 1.90 was
produced, Table S1, due to a slightly lower selectivity in H2 than in CO.
Increasing the concentration of the feed, stable CH4 conversions close to
90% were still achieved, despite the higher partial pressure of the re-
actants and the higher temperatures developed in the bed that could
deactivate the catalyst. The H2/CO ratio slightly decreased since the
reverse water gas shift (rWGS) reaction may be favored at concentrated
conditions due to higher temperatures [53]. The stability of the cata-
lysts was confirmed by control tests, in which an 83% constant con-
version was obtained.
The performance of the Rh5 catalyst was compared with data from
the literature considering specific activity (moles of methane converted
in an hour, at a given temperature, scaled on the mass of active metal,
molCH4/gRh h [54]) in Table S2. However, it should be considered that
reaction conditions were not similar, which modify catalytic activities.
The activity of Rh5 electrodeposited catalyst was close to that reported
for Rh/Ce16Zr84Ox on a honeycomb monolith [54] and Rh directly
deposited on a foam, although the later sintered (vide infra) [55]. In
comparison with Ni-based catalyst coated on Ni foam [20], not shown
in the Table S2, the performance of the electrodeposited structured
catalysts was comparable in terms of CH4 conversion but avoiding
carbon formation (vide infra).
The decrease in the metal loading in Rh2 catalysts largely decreased
the performance and the reproducibility of the data at 750 °C oven
temperature in all reaction conditions; CH4 conversion values were
about 20% less than Rh5 conversions with a larger standard deviation
(change from 1.5 to 7.7% depending the reaction conditions) as seen in
Fig. 4. Moreover, the catalyst activated more slowly in the initial tests
at 126,000 h−1. The largest differences between the different tested
samples may be related to the heat generated inside the catalytic bed; at
low catalytic performance, the contribution of the heat developed to
sustain the reaction may be more remarkable and it may change from
sample to sample. The selectivities in CO and H2 were 12–18% lower
than those of Rh5 sample, with H2/CO ratios close to 2. However, the
performance was rather stable with time-on-stream. The differences in
activity were more remarkable by setting the oven temperature at
500 °C; indeed, decreasing the external heat the Rh2 catalyst was not
active.
Once the activity of the Rh5 catalysts was ascertained, their stability
was studied. A 28 h time-on-stream test was performed after the above
activity tests over two different samples. Three reaction cycles were run
at 30,500 h−1, including tests at 500 and 750 °C with diluted and
concentrated gas mixtures, respectively (Fig. 5). After every cycle, the
reactor was shut-down by cooling to room temperature under N2 at-
mosphere. The conversion was 48% and 88% at 500 and 750 °C, re-
spectively, during the first cycle. Only a slight deactivation of 2 and 1%
at low and high oven temperatures, respectively, was observed at the
end of the tests, suggesting a good stability with time-on-stream of the
catalysts under the reaction conditions here studied.
Catalytic results confirmed that the deposition of well-controlled HT
compounds on the foams provided high and reproducible performances
whenever the Rh loading was 5% a.r. To correlate the catalyst prop-
erties and performances, the spent catalysts were characterized.
The high activity of Rh5 structured catalysts was further demon-
strated decreasing the effect of the heat supplied during tests with the
2/1/20 v/v gas mixture at 500 °C. All the samples converted CH4, ex-
ceeding 55% at high GHSV values and only a slight deactivation
(ca. < 1%). In these tests the heat developed and its transfer in the
catalytic bed played an important role in the activity and selectivity.
3.4. Spent catalysts
The temperature and gas composition varied along the catalytic bed
in CPO tests, namely in the first region of the bed exothermic reactions
occur, while in the second part the endothermic steam reforming and
17

P.H. Ho et al.
AppliedCatalysisA,General560(2018)12–20
The above commented techniques did not give any evidence about
the formation of carbon deposits on the surface of the catalysts;
therefore, to better investigate this feature micro-Raman analyses were
performed along the length of the spent foam cylinders, Fig. S3. No
carbon was present in the coating at the inlet of the catalytic bed, where
oxidation reactions occurred, and only in some few regions of the foam
surface at the middle and outlet of the bed the characteristic D and G
bands of carbon were registered at ca. 1340 and 1590 cm−1, respec-
tively. They were more intense at the middle of the bed, i.e. where CH4
reforming took place; however, their detection only in a few regions
and their low intensity suggested a small carbon formation over these
catalysts after 50 h time-on-stream. This confirmed the better stability
of Rh-based structured catalyst against carbon formation, also fostered
by HT-derived support, than Ni-based structured catalyst, where gra-
phite phase was detected in spent catalyst by XRD [20].
Fig. 5. CH4 conversion during stability tests on Rh5 samples at
GHSV = 30,500 h−1. The data correspond to the average of two catalytic tests
over different samples.
4. Conclusions
The feasibility of the electrodeposition to in situ synthesize Rh/Mg/
Al hydrotalcite-type (HT) compounds with controlled properties on the
surface of FeCrAlloy foams of small pore size in an effective and facile
way was demonstrated for the first time. The use of a two-compartment
flow-cell probably due to separation of cathodic and anodic compart-
ments and the replenishment of the electrolyte solution in the elec-
trode-electrolyte interface, allowed to evenly deposit 5–20 μm Rh/Mg/
Al coatings with a HT structure and composition like the nominal one,
avoiding the sequential precipitation of the solids previously reported.
Only some deviations were found in few regions.
Consequently, after calcination at 900 °C and reduction treatment,
coatings were mainly made of catalysts with properties close to co-
precipitated ones, i.e. MgO and MgAl2O4 with well dispersed Rh na-
noparticles (average 2–3 nm), and with a good thermal stability de-
creasing the detachment and crack formation. Thus, in electrodeposited
catalytic coatings with controlled composition, a high Rh amount could
be loaded to balance the low catalyst amount, avoiding any sintering
phenomena. Moreover, the simultaneous thermal decomposition of the
HT structure and the oxidation of the alloy led to the development of a
MgAl2O4 film at the interface between the foam and the coating due to
the chemical interaction between Mg2+ from the coating and Al3+ from
the support, which increased the adherence to the support. These
properties, together with the low carbon formation tendency of the
catalysts, provided active and stable performances in the CPO of CH4,
whenever the Rh loading in the coating was 5% a.r. A thicker coating
should be necessary to achieve similar performances with the 2% a.r.
catalyst to balance the number of active species. The Rh5 structured
catalyst prepared by the simple electrodeposition method and con-
taining ca. 0.26 wt% total Rh, exhibited comparable specific activity
performance than Rh-based structured catalysts reported in the litera-
ture.
the water gas shift or reverse water gas shift take place [53,56].
Therefore, different deactivation processes (sintering, detaching, and
coking) could take place in catalysts depending on the position in the
bed. To study these phenomena, Rh5 and Rh2 spent catalysts were
characterized by FE-SEM, n-XRF/XRD tomography, HRTEM, and micro-
Raman. The results commented for Rh5 samples correspond to catalysts
after long-term tests, i.e. after ca. 50 h time-on-stream.
In general, the morphology of the coating was preserved during
catalytic tests, i.e. no particle sintering and extended detachment oc-
curred, as evidenced in Fig. 6a. Only was found some partial detach-
ment at the inlet of the catalytic bed, where the temperatures were the
highest. Combined n-XRF/XRD tomography further confirmed the sta-
bility of the structured catalysts under the reaction conditions here
reported. The distributions of periclase and spinel phases and Rh XRF
signals were similar than those in fresh catalysts (Fig. 6f). Most im-
portantly, the alloy was stable under CPO reaction conditions and no
further oxides segregated. From these results, it could be stated that
despite the absence of the Al2O3 protective layer, catalyst coating
preserved the foam from oxidation under partial oxidation tests.
Only in a few regions of Rh5 sample, aggregates of Rh particles
(15–30 nm) could be identified by FE-SEM (Fig. 6b), although the
presence of Rh all over the surface of Rh5 and Rh2 catalysts was con-
firmed by EDS. To get insight into the properties of Rh particles, some
powder was scraped from the upper part of the spent foam cylinders,
namely at the inlet of the catalytic bed, and investigated by HRTEM/
STEM-HAADF, Fig. 6c–h.
Rh nanoparticles were found both on the external surface or em-
bedded in ill- and well-defined nano-hexagonal shaped particles Fig. 6c,
d, due to the bulk characteristics of catalysts [57]. STEM-EDS analyses
confirmed that they contained both Mg and Al. High magnification
images confirmed that Rh was dispersed on both MgO and MgAl2O4
phases, Fig. 6e, in agreement with n-XRF/XRD results. A narrow Rh
metal particle size distribution between 1–5 nm was obtained for both
Rh2 and Rh5 catalysts (Fig. 6g, h), which slightly shifted toward
smaller dimensions for the former; however, it was not possible to es-
tablish a relationship between particle size and crystalline phases as
previously reported for HT-derived catalysts [15]. The small Rh° par-
ticle size, despite the high Rh loading, i.e. 5% a.r., was characteristic of
HT-derived catalysts, overcoming the Rh sintering observed for Rh di-
rectly deposited on the foam [55].
The versatility of HT-type compounds as precursors of catalysts for
H2 and syngas production processes using both methane and biomass-
derived feedstocks, and for many others catalytic processes, and the
advantages of metallic foams vs honeycomb monoliths will open new
applications of the electrodeposited materials once their properties are
under control.
Acknowledgements
The financial support from the University of Bologna (FARB) is
acknowledged. P.H. Ho thanks to SINCHEM grant for funding during
his PhD course. SINCHEM is a joint-doctorate programme selected
under Erasmus Mundus Action 1 programme (FBA 2013-0037). For the
assistance in nano-XRF/XRD tomography measurements at ESRF
beamline ID16b we thank J. Villanova. Thanks also go to Dr. Y. Zhang
for his help during the measurements at ESRF and to F. Corticelli for FE-
SEM analysis at Bologna-IMM-CNR.
Hence, the stable and homogeneous coating on the open-cell foam
and the presence of well dispersed Rh nanoparticles provided active
and stable performance for oxidation and reforming reactions taking
place in the CPO of CH4. The low activity of Rh2 catalyst, despite si-
milar characterization results, suggested that a thicker loading was
necessary to balance the lower number of active sites.
18

P.H. Ho et al.
AppliedCatalysisA,General560(2018)12–20
Fig. 6. Characterization of spent catalysts: (a, b) FE-SEM images, (c, d, e) HRTEM and STEM-HAADF images; (e) FFT of the indicated regions of the TEM image; (f) n-
XRF/XRD distributions; (g, h) Rh particle size distributions of Rh2 and Rh5, respectively.
Appendix A. Supplementary data
[3] A. Delparish, A.K. Avci, Fuel Process. Technol. 151 (2016) 72–100.
[4] M. Aghayan, D.I. Potemkin, F. Rubio-Marcos, S.I. Uskov, P.V. Snytnikov,
I. Hussainova, ACS Appl. Mater. Interfaces 9 (2017) 43553–43562.
[5] A. Montebelli, C.G. Visconti, G. Groppi, E. Tronconi, S. Kohler, H.J. Venvik,
R. Myrstad, Appl. Catal. A 481 (2014) 96–103.
Supplementary material related to this article can be found, in the
online version, at doi:https://doi.org/10.1016/j.apcata.2018.04.014.
[6] J.C. Park, N.S. Roh, D.H. Chun, H. Jung, J.I. Yang, Fuel Process. Technol. 119
(2014) 60–66.
References
[7] M. Frey, T. Romero, A.C. Roger, D. Edouard, Catal. Today 273 (2016) 83–90.
[8] L. Bobrova, N. Vernikovskaya, V. Sadykov, Catal. Today 144 (2009) 185–200.
[9] G. Landi, P.S. Barbato, S. Cimino, L. Lisi, G. Russo, Catal. Today 155 (2010) 27–34.
[10] T. Maffei, S. Rebughini, G. Gentile, S. Lipp, A. Cuoci, M. Maestri, Chem. Ing. Tech.
86 (2014) 1099–1106.
[1] E. Tronconi, G. Groppi, C.G. Visconti, Curr. Opin. Chem. Eng. 5 (2014) 55–67.
[2] J. Gascon, J.R. van Ommen, J.A. Moulijn, F. Kapteijn, Catal. Sci. Technol. 5 (2015)
807–817.
19

P.H. Ho et al.
AppliedCatalysisA,General560(2018)12–20
[11] M. Hettel, E. Daymo, O. Deutschmann, Comput. Chem. Eng. 109 (2018) 166–178.
[12] F. Basile, P. Benito, S. Bugani, W. De Nolf, G. Fornasari, K. Janssens, L. Morselli,
E. Scavetta, D. Tonelli, A. Vaccari, Adv. Funct. Mater. 20 (2010) 4117–4126.
[13] P. Benito, W. De Nolf, G. Nuyts, M. Monti, G. Fornasari, F. Basile, K. Janssens,
F. Ospitali, E. Scavetta, D. Tonelli, A. Vaccari, ACS Catal. 4 (2014) 3779–3790.
[14] O.H. Laguna, M.I. Domínguez, M.A. Centeno, J.A. Odriozola, Chem. Eng. J. 302
(2016) 650–662.
[15] F. Basile, P. Benito, G. Fornasari, A. Vaccari, Appl. Clay Sci. 48 (2010) 250–259.
[16] F. Basile, P. Benito, G. Fornasari, V. Rosetti, E. Scavetta, D. Tonelli, A. Vaccari, Appl.
Catal. B 91 (2009) 563–572.
[37] X. Lu, C. Zhao, Nat. Commun. 6 (2015) 6616.
[38] X. Zhu, C. Tang, H.F. Wang, B.Q. Li, Q. Zhang, C. Li, C. Yang, F. We, J. Mater. Chem.
A 4 (2016) 7245–7250.
[39] L. Jiang, Y. Sui, J. Qi, Y. Chang, Y. He, Q. Meng, F. Wei, Z. Sun, Y. Jin, Part. Part.
Syst. Charact. 34 (2017) 1600239.
[40] V.A. Ettel, Apparatus and foam electroplating process, WO patent 121549 (2007),
to Inco Limited, https://patentscope.wipo.int/search/en/detail.jsf?docId=
WO2007121549.
[41] A. Jung, M. Weinmann, H. Natter, M. Aliofkhazraei, A.S.H. Makhlouf (Eds.),
Handbook of Nanoelectrochemistry: Electrochemical Synthesis Methods, Properties
and Characterization Techniques, Springer, 2016, pp. 941–970.
[42] J. Zhang, E.J. Juárez-Pérez, I. Mora-Seró, B. Viana, T. Pauporté, J. Mater. Chem. A 3
(2015) 4909–4915.
[17] R. Chai, Y. Li, Q. Zhang, S. Fan, Z. Zhang, P. Chen, G. Zhao, Y. Liu, Y. Lu,
ChemCatChem 9 (2017) 268–272.
[18] R. Chai, S. Fan, Z. Zhang, P. Chen, G. Zhao, Y. Liu, Y. Lu, ACS Sustain. Chem. Eng. 5
(2017) 4517–4522.
[43] C. Devivier, V. Tagliaferri, F. Trovalusci, N. Ucciardello, Mater. Des. 86 (2015)
[19] F. Basile, P. Benito, G. Fornasari, M. Monti, E. Scavetta, D. Tonelli, A. Vaccari, Catal.
Today 157 (2010) 183–190.
[20] R. Chai, Z. Zhang, P. Chen, G. Zhao, Y. Liu, Y. Lu, Micropor. Mesopor. Mater. 253
(2017) 123–128.
[21] A. Casanovas, N.J. Divins, A. Rejas, R. Bosch, J. Llorca, Int. J. Hydrogen Energy 42
(2017) 13681–13690.
[22] I. Reyero, I. Velasco, O. Sanz, M. Montes, G. Arzamendi, L.M. Gandía, Catal. Today
272–278.
[44] T. Loewenstein, A. Hastall, M. Mingebach, Y. Zimmermann, A. Neudeck,
D. Schlettwein, Phys. Chem. Chem. Phys. 10 (2008) 1844–1847.
[45] A.S. Batchellor, S.W. Boettcher, ACS Catal. 5 (2015) 6680–6689.
[46] M. Nobial, O. Devos, B. Tribollet, J. Cryst. Growth 327 (2011) 173–181.
[47] R. Abdallah, R. Geneste, T. Labasque, H. Djelal, F. Fourcade, A. Amrane, S. Taha,
D. Floner, J. Electroanal. Chem. 727 (2014) 148–153.
216 (2013) 211–219.
[23] N. Gokon, Y. Yamawaki, D. Nakazawa, T. Kodama, Int. J. Hydrogen Energy 35
(2010) 7441–7453.
[24] Z. Lü, F. Zhang, X. Lei, L. Yang, S. Xu, X. Duan, Chem. Eng. Sci. 63 (2008)
4055–4062.
[25] F. Kovanda, K. Jirátová, Appl. Clay Sci. 53 (2011) 305–316.
[26] F.J. Echave, O. Sanz, L.C. Almeida, J.A. Odriozola, M. Montes, Stud. Surf. Sci. Catal.
175 (2010) 639–642.
[27] C. Zhao, J. Wu, L. Yang, G. Fan, F. Li, Ind. Eng. Chem. Res. 56 (2017) 4237–4244.
[28] L. Indira, P.V. Kamath, J. Mater. Chem. 4 (1994) 1487–1490.
[29] C. Mousty, A. Walcarius, J. Solid State Electrochem. 19 (2015) 1905–1931.
[30] P.H. Ho, M. Monti, E. Scavetta, D. Tonelli, E. Bernardi, L. Nobili, G. Fornasari,
A. Vaccari, P. Benito, Electrochim. Acta 222 (2016) 1335–1344.
[31] P. Benito, M. Monti, W. De Nolf, G. Nuyts, G. Janssen, G. Fornasari, E. Scavetta,
F. Basile, K. Janssens, F. Ospitali, D. Tonelli, A. Vaccari, Catal. Today 246 (2015)
154–164.
[48] M. Cotte, E. Pouyet, M. Salomé, C. Rivard, W. De Nolf, H. Castillo-Michel, T. Fabris,
L. Monico, K. Janssens, T. Wang, P. Sciau, L. Verger, L. Cormier, O. Dargaud,
E. Brun, D. Bugnazet, B. Fayard, B. Hesse, A.E. Pradas del Real, G. Veronesi,
J. Langlois, N. Balcar, Y. Vandenberghe, V.A. Solé, J. Kieffer, R. Barrett, C. Cohen,
C. Cornu, R. Baker, E. Gagliardini, E. Papillon, J. Susini, J. Anal. At. Spectrom. 32
(2017) 477–493.
[49] W. De Nolf, F. Vanmeert, K. Janssens, J. Appl. Crystallogr. 47 (2014) 1107–1117.
[50] A. Ballarini, P. Benito, G. Fornasari, O. Scelza, A. Vaccari, Int. J. Hydrogen Energy
38 (2013) 15128–15139.
[51] C. Berger-Karin, M. Sebek, M.-M. Pohl, U. Bentrup, V.A. Kondratenko, N. Steinfeldt,
E.V. Kondratenko, ChemCatChem 4 (2012) 1368–1375.
[52] J.C.A.A. Roelofs, J.A. van Bokhoven, A.J. van Dillen, J.W. Geus, K.P. de Jong,
Chem. Eur. J. 8 (2002) 5571–5579.
[53] M. Maestri, D. Livio, A. Beretta, G. Groppi, Ind. Eng. Chem. Res. 53 (2014)
10914–10928.
[54] A. Scarabello, D.D. Nogare, P. Canu, R. Lanza, Appl. Catal. B. 174–175 (2015)
308–322.
[32] P. Benito, G. Nuyts, M. Monti, W. De Nolf, G. Fornasari, K. Janssens, E. Scavetta,
A. Vaccari, Appl. Catal. B 179 (2015) 321–332.
[55] E. Verlato, S. Barison, S. Cimino, F. Dergal, L. Lisi, G. Mancino, M. Musiani,
L. Vazquez-Gomez, Int. J. Hydrogen Energy 39 (2014) 11473–11485.
[56] R. Horn, K.A. Williams, N.J. Degenstein, A. Bitsch-Larsen, D. Dalle Nogare,
S.A. Tupy, L.D. Schmidt, J. Catal. 249 (2007) 380–393.
[33] Y. Liu, X. Teng, Y. Mia, Z. Chen, J. Mater. Chem. A 5 (2017) 24407–24415.
[34] S. Yoon, J.Y. Yun, J.H. Lim, B. Yoo, J. Alloys Compd. 693 (2017) 964–969.
[35] Y. Liu, N. Fu, G. Zhang, M. Xu, W. Lu, L. Zhou, H. Huang, Adv. Funct. Mater. 27
(2017) 1605307.
[57] P. Benito, V. Dal Santo, V. De Grandi, M. Marelli, G. Fornasari, R. Psaro, A. Vaccari,
[36] Z. Li, M. Shao, H. An, Z. Wang, S. Xu, M. Wei, D.G. Evans, X. Duan, Chem. Sci. 6
(2015) 6624–6631.
Appl. Catal. B 179 (2015) 150–159.
20