Comparison of catalytic performance and coking resistant behaviors of cobalt- and nickel based catalyst with different Co/Ce and Ni/Ce molar ratio under SRE conditions

Article abstract of DOI:10.1016/j.apcata.2019.117334

A series of Co-xCe/CeO₂ and Ni-xCe/CeO₂ catalysts with different ceria amount introduced from Ce(NO₃)₃ × 6H₂O precursor, corresponding to Ce/Co(Ni) molar ratio from 0 to 1 were prepared by impregnation method. Among the tested catalysts, Co-0.1Ce/CeO₂ and Ni-0.1Ce/CeO₂ samples exhibited the smallest metal crystallites size, the strongest metal-support interaction, the best catalytic performance and the strongest resistance toward carbon deposition. Only an introduction of small amount of ceria from Ce(NO₃)₃ × 6H₂O precursor led to an increase in the number of active sites for breaking C[sbnd]C and C[sbnd]H bonds. Too high Ce/Co molar ratio caused rapid deactivation of cobalt-based catalysts because of faster coverage of active sites by encapsulating carbon. Whereas too high Ce/Ni molar ratio generated the thick and long carbon filamentous deposits causing the nickel-based catalysts deactivation including not only the coverage of nickel species by carbon but also limitation of the contact of reactants with nickel species.

Full text of DOI:10.1016/j.apcata.2019.117334

Journal Pre-proof
Comparison of catalytic performance and coking resistant behaviors of
cobalt- and nickel based catalyst with different Co/Ce and Ni/Ce molar
ratio under SRE conditions
Magdalena Greluk, Marek Rotko, Sylwia Turczyniak-Surdacka
PII:
S0926-860X(19)30489-2
DOI:
https://doi.org/10.1016/j.apcata.2019.117334
APCATA 117334
Reference:
To appear in:
Applied Catalysis A, General
Received Date:
Revised Date:
Accepted Date:
16 July 2019
28 September 2019
30 October 2019
Please cite this article as: Greluk M, Rotko M, Turczyniak-Surdacka S, Comparison of
catalytic performance and coking resistant behaviors of cobalt- and nickel based catalyst with
different Co/Ce and Ni/Ce molar ratio under SRE conditions, Applied Catalysis A, General
(2019), doi: https://doi.org/10.1016/j.apcata.2019.117334
This is a PDF file of an article that has undergone enhancements after acceptance, such as
the addition of a cover page and metadata, and formatting for readability, but it is not yet the
definitive version of record. This version will undergo additional copyediting, typesetting and
review before it is published in its final form, but we are providing this version to give early
visibility of the article. Please note that, during the production process, errors may be
discovered which could affect the content, and all legal disclaimers that apply to the journal
pertain.
©
2019 Published by Elsevier.

Comparison of catalytic performance and coking resistant behaviors of cobalt- and nickel
based catalyst with different Co/Ce and Ni/Ce molar ratio under SRE conditions
a,*
a
b
Magdalena Greluk , Marek Rotko , Sylwia Turczyniak-Surdacka
aDepartment of Chemical Technology, Faculty of Chemistry, Marie Curie-Sklodowska
University, Maria Curie-Sklodowska Sq. 3, 20-031 Lublin, Poland
bBiological and Chemical Research Centre, University of Warsaw, 101 Żwirki i Wigury
Street, 20-089 Warsaw, Poland
*Corresponding author. Tel.: +48 81 537 55 14; fax: +48 81 537 55 65.
E-mail adress: magdalena.greluk@poczta.umcs.lublin.pl (Magdalena Greluk).
Graphical abstract
Highlights

Co(Ni)-support interaction decreases with increase in Ce amount above
Co(Ni)/Ce=0.1
1


Optimum Ce amount for performance in SRE corresponding to molar ratio
metal/Ce=0.1.


Ce amount influences on morphology of coke deposits of Ni catalysts.
The Ni catalysts were more capable of cleaving the C–C bond in SRE than Co
samples.

The coke deposited on Co and Ni catalysts in SRE differs the type and morphology.
Abstract
A series of Co-xCe/CeO
2
and Ni-xCe/CeO catalysts with different ceria amount introduced
2
from Ce(NO
prepared by impregnation method. Among the tested catalysts, Co-0.1Ce/CeO2 and Ni-
.1Ce/CeO2 samples exhibited the smallest metal crystallites size, the strongest metal-support
interaction, the best catalytic performance and the strongest resistance toward carbon
deposition. Only an introduction of small amount of ceria from Ce(NO × 6H O precursor
3
)
3
× 6H
2
O precursor, corresponding to Ce/Co(Ni) molar ratio from 0 to 1 were
0
3
)
3
2
led to an increase in the number of active sites for breaking C-C and C-H bonds. Too high
Ce/Co molar ratio caused rapid deactivation of cobalt-based catalysts because of faster
coverage of active sites by encapsulating carbon. Whereas too high Ce/Ni molar ratio
generated the thin and long carbon filamentous deposits causing the nickel-based catalysts
deactivation including not only the coverage of nickel species by carbon but also limitation of
the contact of reactants with nickel species.
Keywords:
Hydrogen production; Ethanol steam reforming; Cobalt; Nickel; Ceria
1
. Introduction
Hydrogen may be produced by catalytic steam reforming of ethanol (SRE). An efficient
catalyst for hydrogen production from ethanol has to dissociate the C─C bond, maintain a low
carbon monoxide concentration by being a poison of platinum anodes in the low-temperature
fuel cells and be stable under catalytic operation. In the literature, supported cobalt and nickel
2

catalysts have shown promising results in SRE due to their low cost, relatively high activity
towards C─C bond cleavage and hydrogenation/dehydrogenation reactions. However, a
drawback in their performance in this process is the formation of carbonaceous species. The
conversion of ethanol and the product distribution is directly related to the catalyst
composition and the operational conditions in the reaction system. The support acts as a
dispersion medium, however, in most cases the nature of the support will influence the
product distribution and the metal support interactions. An ideal support to SRE would not
favour dehydration reactions (otherwise polymer formation may occur) and would have no (or
mild) capacity for other C─C bond formation reactions (e.g., making higher unsaturated
aldehydes by aldolization reactions) [1-5]. Moreover, there is some evidence that the use of
reducible oxides, like CeO
Al or SiO . Ceria improves catalyst stability due to their high oxygen storage capacity
OSC) and oxygen mobility. The high oxygen-exchange capacity of ceria is associated with
2
can result in additional benefits when compared to ir such as
2
O
3
2
(
4
+
3+
its ability to reversibly change oxidation states between Ce /Ce by reversibly
storing/releasing oxygen. The easily accessible oxygen can react with carbon species as soon
as it forms, and this process keeps the metal surface free of carbon, thus inhibiting
deactivation. Moreover, support such as ceria is known to be highly active toward WGS,
which promotes the reduction of carbon monoxide concentration in reformate. [2, 6, 7].
Generally, catalytic activity and coking resistant behaviors of catalysts are affected by
their surface and structural properties, so that the methodology involved in the preparation
step can lead to the obtainment of materials with important properties for applications in
catalytic processes [8]. Various synthesis techniques have been used preparing Co/CeO
Ni/CeO catalysts for SRE process, the most commonly utilized ones being wet impregnation
3, 8-15], incipient wetness impregnation [8, 16, 17] and conventional co-precipitation [8, 18-
2
and
2
[
2
0]. However, recent years have witnessed application of many novel synthesis techniques to
3

Co/CeO
2
and Ni/CeO catalysts preparation for SRE reaction such as hydrothermal co-
2
precipitation [21], hydrothermal ultrasonic-assisted co-precipitation [21], co-precipitation in
reverse microemulsions [15, 17], solvothermal decomposition [17], colloidal crystal
templating [17] or one-step polymerization [22]. The conclusion drawn from these papers is
that SRE reaction on Co/CeO
2
and Ni/CeO catalysts is mainly dependant on the active
2
phase's particle and the interactions between the active metal and its support (SMSI) which
offers a route to control the structural properties of supported metals and, hence, their activity,
stability and resistance to carbon formation [20]. The form and degree of metal-support
interaction depend on many factors, including the content of metal and preparation methods.
According to Lovón et al. [23], the metal-support interactions strengthened with the increase
of cobalt content of Co/CeO catalyst from 5-20 wt.% resulted in better ethanol conversion, a
2
higher hydrogen selectivity and better stability of the catalyst with the highest amount of
cobalt. Also Carvalho et al. [24] observed that the enhancement of the metal-support
interactions with increase of cobalt content of Co/CeO catalyst from 5-20 wt.% led to the
2
better catalytic performance of the catalyst containing 20 wt.% of cobalt amount in SRE
reaction and a lower carbon deposition rate on it. Moreover, the increase of interaction
between metal and support of Co/CeO
the increase in cobalt content was also observed for the highest amount of metal ranging from
5-29 wt.% [25]. The enhancement of the metal-support interactions of Co/CeO and Ni/CeO
2
catalysts favoring the catalytic properties in SRE with
1
2
2
catalysts for SRE reaction was also possible by using the proper preparation method to obtain
materials. Song et al. [17] proposed novel synthesis techniques such as solverthermal
decomposition, colloidal crystal templating and reverse microemulsion methods to prepare
Co/CeO catalysts and the superior performance in SRE reaction of the catalyst prepared by
2
the last mentioned technique was attributed to the improved cobalt dispersion but also
enhanced metal-support interaction and increased metal-support interface. The influence of
4

the preparation method on the strength of the metal-support interaction resulting in the
catalyst properties in SRE reaction was also found for Ni/CeO catalysts by Fang et al [8]. Co-
precipitation method allowed to obtain the smallest sized NiO and CeO nanoparticles which
2
2
enhanced the interactions between the nickel and cerium species and not only led to very high
activities but also meant that the stability was maintained due to the formation of the graphitic
filamentous carbon, small and homogeneous in size compared to the filamentous carbon
formed over the catalysts issued from the other preparation methods. Moreover, Fang et al. [8]
demonstrated that the stronger metal-support interaction allowed to minimize the amount of
carbon deposition on the catalyst surface during SRE process. Also Liu et al. [26] noticed
dependence between the metal-support interaction and carbon formation on Ni-CeO catalyst
x
in SRE reaction stating that a strong-metal interaction between nickel and ceria facilitated
oxygen transfer and improved the oxidation of surface carbon. Moreover, the strong metal-
support interactions between nickel and ceria perturbed the electronic and chemical properties
of nickel atoms reducing their ability to break C─O bonds. Thus, the nickel-ceria interactions
substantially decreased the carbon monoxide methanation activity of nickel which is an
important factor to take into account when dealing with the steam reforming of ethanol on
Ni/CeO
Inspired by the findings in the literature mentioned above, the idea of this work was to
prepare Co/CeO and Ni/CeO catalysts exhibiting the strong metal-support interactions. It
2
catalyst as it was emphasized by Xu et al [27] and Zhou et al [28].
2
2
was assumed that the methodology used in their preparation, i.e. introduction of the metal
active phase in the presence of ceria to the support allows to strengthen the metal-support
interactions which stabilizes the loaded metal nanoparticles against thermal sintering during
SRE reaction. It was also supposed that subsequent addition of cobalt/nickel and CeO
2
to the
nano dispersed CeO support allows to modify the surface and structural properties of the
2
5

obtained catalysts, leading to the improvement of their catalytic activities and coking
resistance in SRE process.
With this aim, a series of Co/CeO
xCe/CeO ) catalysts with different Co/Ce and Ni/Ce molar ratio (x=0.1, 0.5 and 1) was
prepared by co-impregnation method in the presence of CA and characterized. For
comparison, the Co/CeO and Ni/CeO catalysts without ceria introduced from Ce(NO
O precursor were prepared by impregnation method in the presence of CA and
characterized. The amount of CeO used as the support remained constant for all catalysts. A
2
and Ni/CeO
2
(denoted as Co-xCe/CeO
2
and Ni-
2
2
2
3
3
) ×
6
H
2
2
correlation between physicochemical properties of these catalysts and their catalytic
performance in SRE reaction was investigated. Additionally, the deactivation behaviors of
these catalysts were investigated. The knowledge of this relation will help not only to
understand the effects of ceria promotion, but also to obtain the catalyst design concept. It is
also an interesting comparison of catalytic performance and deactivation behaviors of cobalt-
and nickel based catalyst under SRE conditions.
2
. Experimental
2
.1. Catalysts preparation
The Co/CeO
nano-dispersed ceria support (<25 nm, Aldrich) in the presence of CA as reported elsewhere
10-14]. The Co-xCe/CeO and Ni-xCe/CeO ) catalysts with different Ce/Co and Ce/Ni molar
ratio (x=0.1, 0.5 and 1) were prepared by co-impregnation the commercial nano-dispersed
ceria support (<25 nm, Aldrich) with a solution of Co(NO × 6H O, Ce(NO × 6H O and
CA or Ni(NO × 6H O, Ce(NO × 6H O and CA (the molar ratios of Co + Ce/CA or Ni +
Ce/CA was 1/1). After impregnation, the catalyst precursor was dried at 110°C for 12 h, then
2
and Ni/CeO
2
catalysts were prepared by impregnation of the commercial
[
2
2
3
)
2
2
3
)
3
2
3
)
2
2
3
)
3
2
-1
calcined at 500°C with the heating rate of 2°C min up to the calcination set point and
maintained for 1 h at that temperature. The amount of the Co(NO
3
)
2
× 6H
2
O and Ni(NO
3
)
2
×
6

6
H
2
O precursors was adjusted to reach a corresponding nominal metal loading of 10 wt% in
all catalysts. The cerium from Ce/Co or Ni/Ce molar ratio means that only cerium introduced
was from Ce(NO × 6H O precursor. The amount of the commercial nano-dispersed ceria
3
)
3
2
support used during the catalyst synthesis was fixed and equaled 10 g for all catalysts which
means that the contribution of the commercial ceria decreased whereas contribution of ceria
from Ce(NO
.2. Catalysts characterization
Catalysts characterization was performed using the conditions reported in previous
3
)
3
× 6H
2
O precursor increased with the increase Ce/Co or Ni/Ce molar ratio.
2
studies. The cobalt and nickel content in the catalysts was determined by X-ray fluorescence
method using a Canberra 1510 fluorescence spectrometer. The X-ray powder diffraction
patterns in Bragg-Brentano geometry were recorded on a Empyrean X-ray (PANalytical)
diffractometer using Cu Kα radiation (λ=0.154 nm). Textural properties of calcined catalysts
were determined by nitrogen adsorption measurements at -196°C using ASAP 2420
(
Micromeritics). Hydrogen chemisorption was measured in the ASAP 2020 apparatus
Micromeritics) These experiments were carried out on the samples after their reduction in
(
hydrogen flow successively for 1 h at 500ºC. Adsorption isotherms of hydrogen were
obtained at 130°C. The amount of surface metal atoms was calculated from the amount of
hydrogen chemisorbed, assuming that one hydrogen atom is adsorbed on the area occupied by
one surface cobalt/nickel atom (the stoichiometry of chemisorption is Co(Ni)/H=1/1) and that
2
the surface area occupied by one atom of hydrogen is equal to 0.065 nm [9]. The total H
2
uptake was determined by extrapolation of the linear part of the isotherm to zero pressure.
Reducibility of catalysts was investigated by the temperature-programmed reduction (TPR)
experiment carried out with the AutoChem II 2920 (Micromeritics). The reduction profile was
-1
obtained by passing 5% H /Ar flow at the rate of 30 mLmin through 0.05 g (0.3-0.6 mm) of
2
-1
the catalyst. The temperature was increased from RT to 750°C at the rate of 10 °C min . The
7

water vapour formed during the reaction was removed in a cold trap placed in front of the
thermal conductivity detector. The susceptibility of active metal phase to its oxidation was
investigated by the temperature-programmed oxidation (TPO) experiment carried out with the
AutoChem II 2920 (Micromeritics) coupled with the quadruple mass spectrometer (HPR-20
with triple filter, Hiden Analitical). Prior to the main experiment, the catalyst sample (0.1 g,
0
.3-0.6 mm) was treated with the 6% H
and next it was cooled down to -70°C. The TPO measurement was performed using
%O /He and heating of the catalyst sample from -70°C to 600°C with the temperature ramp
of 10°C×min . In all cases, i.e. during the pre-treating and the main experiment, the total flow
2
/Ar mixture, at the temperature of 500°C for 1 hour
5
2
-1
-1
rate of the reaction mixture was 45 mL min . Studies of the catalyst coking under the steam
reforming of ethanol conditions were performed by the thermogravimetric method using the
TG121 microbalance system (CAHN), under dynamic conditions in a quartz reactor with a
continuous flow of ethanol-water vapours diluted with He at 420 °C. The molar ratio of
H
2
O/ethanol was equaled to 12/1 (ethanol concentration of 7.1 mol%). Prior to reaction,
catalyst sample (0.01 g, 0.3-0.6 mm) was reduced by passing 10 % H /He flow at the
temperature of 500°C for 1 hour. In all cases, i.e. during pre-treating and main experiment, the
2
-1
total flow rate of the reaction mixture was 70 mL min . Quantification of the carbon
formation rate (CFR) on catalysts was calculated according to the equation [29]:
mcarbon
CFR 
(1)
musedcatal ys t t
where:
mcarbon - is the mass of carbon deposited on the catalyst calculated from TG; musedcatal ys t - is the
mass of the catalyst, t – is the time of the reaction of steam reforming of ethanol.
SEM images were taken using a FIB-SEM Crossbeam 540 FEG (Zeiss) with an acceleration
voltage of 1.8 kV. For SEM observations, the samples were prepared by drying the solution of
8

catalyst diluted in 98% ethanol dropped on carbon tape at room temperature. The TEM and
STEM images of catalysts were obtained in scanning/transmission electron microscope
TALOS F200X (FEI Company), equipped with field emission gun (X-FEG), HAADF
detector and EDS spectrometer. Microscopic studies of the catalyst were conducted at an
accelerating voltage of the electron beam equal to 200 kV. The mapping was conducted in the
STEM mode by collecting point by point EDS spectrum of each of the corresponding pixels
in the map. The collected maps were presented in the form of a matrix of pixels with the color
mapped significant element and the intensity corresponding to percentage of the element.
2
.3. Catalytic performance test
The reaction of ethanol conversion was performed in a Microactivity Reference unit
(PID Eng & Tech.) under atmospheric pressure in a fixed-bed continuous-flow quartz reactor
over the catalyst (0.3-0.6 mm) reduced in situ with hydrogen with a flow rate of 100 ml min-1
at 500°C for 1 hour, prior to the reaction. The catalysts mass was selected as 0.05 or 0.1 g in
order to reach space velocity equal to 120000 and 60000 ml/g h, respectively. The catalyst
diluted at the weight ratio of 1/25 with the same size of quartz in order to avoid temperature
gradients and hot spots. The aqueous solution of ethanol with molar ratio H O/ethanol of 12/1
2
or 4/1 corresponded to the constant ethanol concentration around 7.7 mol% was supplied by a
mass controller (Bronkhorst) (5.20 g h-1 (H O/ethanol=12/1) and 2.21 g h-1
2
(
H O/ethanol=4/1)) to a michrochannel evaporator (Fraunhofer-ICT-IMM) heated by a
2
heating cartridge. The temperature was controlled with integrated K-type thermocouple
150°C). The reactant vapours were fed to the reactor at an actual flow rate of 100 mL·min-1
(
without diluting with any inert gas in order to work in similar conditions expected in a real
-1
fuel cell for H
2
O/ethanol molar ratio of 12/1 or at an actual flow rate of 38.5 mL min diluted
with argon with the flow rate of 61.5 mL/min for H
2
O/ethanol molar ratio of 4/1. The analysis
of the reaction mixture and the reaction products (all in the gas phase) were carried out on-line
9

by means of two gas chromatographs. One of them, Bruker 450-GC was equipped with two
columns, the first filled with a porous polymer Porapak Q (for all organics, CO and H
vapor) and the other one – capillary column CP-Molsieve 5Å (for CH and CO analysis).
2
2
O
4
Helium was used as a carrier gas and a TCD detector was employed. The hydrogen
concentration was analyzed by the second gas chromatograph, Bruker 430-GC, using a
Molsieve 5Å, argon as a carrier gas and a TCD detector. Argon was used as an internal
standard. In the case of the experiment with molar ratio H O/ethanol=12/1, argon was
2
introduced out of the reactor. Because argon flow rate in each steam reforming test was fixed,
based on the argon content in outlet, it was possible to calculate the selectivity to products in
each experiment.
The conversion of ethanol ( XEtOH ) and conversions of ethanol into individual carbon-
containing products ( XCP ) were calculated on the basis of its concentrations before and after
the reaction:
in
EtOH
out
EtOH
C
 C
X EtOH

100%
(2)
(3)
in
C
EtOH
out
niCi
XCP

100%
out

niCi
where:
in
EtOH
out
EtOH
C
- is the molar concentration of ethanol in the reaction mixture (mol%);
C
- is the
out
molar concentration of ethanol in the post-reaction mixture (mol%); Ci - is the molar
concentration of carbon-containing product in the post-reaction mixture (mol%); ni – is
number of carbon atoms in carbon-containing molecule of the reaction product.
The selectivity of hydrogen formation was determined as:
out
CH2
H selectivity 
100%
(4)
2
out
out
CH4
out
out
CH CHO
out
(CH ) CO
C  2C  2C
 2C
 3C
H
C H
2
2
4
3
3 2
1
0

where:
out
C
- is the molar concentration of the hydrogen-containing products in the post-reaction
mixture (mol%).
Blank runs (without catalyst) have been performed in order to confirm complete
vaporization of ethanol and water. These runs have been carried in an empty quartz reactor
and with maintaining the same conditions as in the runs with catalyst. The quantitate analysis
of the reaction mixture by means of Bruker 450-GC chromatograph confirmed that
H
2
O/ethanol molar ratio was equal to 12/1 or 4/1.
3
. Results and discussion
3
.1. Catalyst characterization
A summary of the physicochemical properties of the synthesized Co/CeO2, Co-
xCe/CeO
2
, Ni/CeO
2
and Ni-xCe/CeO
2
catalysts and CeO commercial support are presented
2
in Tables 1 and 2. It can be seen that the real content of cobalt or nickel was close to the
nominal value of 10 wt.%. All catalysts had a medium BET surface area which was
contributed to the support CeO
2
. There is no clear dependence between BET surface area and
× 6H O precursor. However, the highest
the amount of ceria introduced from Ce(NO
3
)
3
2
values of this parameter were obtained for catalysts with Ce/Co and Ce/Ni molar ratio equal
to 0.5. All catalysts possessed a mesoporous structure but it was found that a pore size and
pore volume slightly decreased with an increase in contribution of ceria introduced from Ce(
NO
3
)
3
× 6H O precursor. Two different techniques were used, hydrogen chemisorption (Table
2
1
) and XRD line broadening (Table 2), to measure metal particle size. Average crystallite
diameters estimated from these methods were found to be in poor agreement. Application of
hydrogen chemisorption method gives data about the amount of metal atoms exposed on the
sample surface [30]. Usually, the stoichiometry of the adsorption is assumed to be one
hydrogen atom for one surface metallic atom. However, it is well known that the presence of
1
1

ceria in a system will cause problems with the estimation of metal dispersion by
chemisorption-based methods. Particle size calculation is speculative because in the case of
metal supported on ceria, it has been observed that the support is able to chemisorb large
amounts of hydrogen via a spillover phenomenon [31]. The easier reduction of ceria in the
presence of transition metal, such as cobalt and nickel, was repeatedly confirmed [11, 32-34]
and explained on the basis of hydrogen dissociation at the metal surface and its subsequent
spillover onto the ceria support. However, the cobalt and nickel crystallites size were
determined by chemisorption method and presented in Table 1. Whereas table 2 lists the
crystallite sizes calculated using the Scherrer equation for unreduced and reduced cobalt- and
nickel-based catalysts. After reduction, the CeO
catalysts in comparison with crystallite size of unreduced catalysts. In addition, all catalysts
exhibited the similar CeO crystallites size regardless of the amount of introduced CeO from
Ce(NO × 6H O precursor. No signals from a cobalt-containing phase were detected after
reduction in the diffraction patterns of Co/CeO and Co-0.1Ce/CeO catalysts which indicates
that it was well-dispersed. But the addition of a larger amount of CeO from Ce(NO
O precursor led to an increase in the size of metallic cobalt crystallites to 14.4 and 17.4
nm for Co-0.5Ce/CeO and Co-Ce/CeO catalyst, respectively. In the case of nickel-based
catalysts, metallic nickel crystallites of Ni/CeO and Ni-0.5Ce/CeO were comparable (d=
10 nm) and metallic nickel crystallites of Ni-0.1Ce/CeO catalyst were the smallest (d=7.2
nm) among all studied nickel-based catalysts. The increase of Ce/Ni molar ratio to 1 caused,
however, a significant increase in the size of metallic nickel crystallites of Ni-Ce/CeO
2
crystallite size remained unchanged for all
2
2
3
)
3
2
2
2
2
3
3
) ×
6
H
2
2
2
2
2
~
2
2
catalyst to 21.4 nm. These results differ from these obtained for Ni-xCe/MMT [35] and Ni-
xCe/SBA-15 [36] catalysts which indicated that nickel particles are smaller with the increase
of the content of ceria promoter. However, Xiao et al. [37] also observed that only low
amount of CeO allowed to reduce nickel particle size and enhanced nickel dispersion of Co-
2
1
2

Ni/xCeO
2
-Al
2
3
O catalyst whereas too high content of this promoter caused agglomeration of
nickel particles. Moreover, the crystallites of NiO (before reduction) were mostly smaller than
the metallic nickel crystallites (after reduction) suggesting that the reduction process increased
the crystallite size by agglomeration or sintering of the active phase (Table 2) [4]. Whereas
the cobalt species with exception of Co-0.5Ce/CeO
under reduction conditions and the metallic cobalt particles become smaller in comparison
with the size of Co crystallites (Table 2) [38]. Probably the bonding between cobalt
2
catalyst undergone the re-dispersion
3
O
4
species and surface cerium atoms became stronger and thus more stable upon reduction,
which, in turn, led to the re-dispersion of metallic cobalt on ceria [38]. The XRD patterns of
Co-xCe/CeO
2
and Ni-xCe/CeO
2
catalysts before and after their reduction with hydrogen at
5
00 °C (see Supporting Information, Figs. S1a-d and S2a-d) showed reflections related to
fluorite face-centered cubic CeO
2
structure (JCPDS 34-0394). Moreover, the XRD patterns of
unreduced cobalt-based catalysts promoted with CeO
attributed to cubic spinel Co (JCPDS 43-1003). In the diffraction pattern of unreduced Ni-
.1Ce/CeO and Ni-0.5Ce/CeO catalysts, signals of nickel-containing phase were not
2
revealed reflection line at 2θ=36.8°
3
O
4
0
2
2
observed suggesting its small particle/crystalline sizes since the XRD technique presents
limitations especially for crystalline sizes below 2-4 nm [20, 39] and metal loading smaller
than 20 wt.% [39]. Only in the case of diffractograms of unreduced Ni/CeO2 and Ni-Ce/CeO
2
catalyst, very broad and small diffraction reflections at 2θ=37.2 and 43.1° corresponding to
cubic NiO phase (JCPDS 01-1239) were observed. The XRD patterns of all studied cobalt-
and nickel-based catalyst after their reduction with hydrogen at 500 °C for 1 hour confirm that
the active phase precursors (Co
or nickel (JCPDS 01-1258). Lines characteristic of Co
detected, suggesting that the Co or NiO reduction was complete. However, taking into
account the results obtained by Xiao et al [40], the XRD technique could give false
3
O
4
or NiO) were reduced to metallic cobalt (JCPDS 01-1277)
3
O
4
or NiO phases were no longer
3
O
4
1
3

information of particle size of the studied catalysts because cobalt/nickel nanoparticles could
be partially or fully covered by ceria introduced from CeO from Ce(NO × 6H O precursor.
Moreover, detailed analysis of diffractograms revealed that there was no shift in characteristic
reflection of CeO neither to the higher nor to the lower angles after ceria introduction from
Ce(NO × 6H O precursor which means that the interaction between Co and CeO and
NiO and CeO was not enhanced. It was also confirmed by the calculation of the lattice
parameter of CeO (Table 2) which remained almost unchangeable for all catalysts. These
2
3
)
3
2
2
3
)
3
2
3
O
4
2
2
2
results exclude solid solution formation in the studied cobalt-ceria and nickel-ceria systems
The morphological properties of promoted with ceria cobalt- and nickel-based catalysts
were studied by high-resolution STEM-EDS analysis (see Supporting Information, Figs. S3-
S8) which indicates the spatial distribution of the cobalt or nickel elements on the surface of
ceria. For the Co-0.1/CeO
2
catalyst, cobalt species were mostly distributed on the surface of
catalyst there were areas where cobalt and cerium
ceria (Fig. S3). For the Co-0.5Ce/CeO
2
species did not overlap suggesting that cobalt was not located on the surface of ceria and
remained unsupported. However, most of it was definitely supported (Fig. S4). Whereas in the
case of Co-Ce/CeO catalyst, the cobalt and cerium species were not in contact with each
2
other, i.e. empty, wide areas of the ceria and isolated unsupported cobalt species were mainly
found (Fig. S5). It means that the proposed method of the catalysts preparation indicates
limitation to the amount of cerium precursor which can be introduced with cobalt precursor
during co-impregnation the ceria support. This limitation was observed in a lesser degree for
the considered nickel-based catalyst samples. Nickel species are mostly distributed over ceria
support regardless of the amount of ceria introduced from Ce(NO
S6-S8). However, small areas that the nickel was unsupported were also observed for Ni-
.5Ce/CeO (Fig. S7) and Ni-Ce/CeO catalysts (Fig. S8). The results obtained for both
cobalt- and nickel-based catalysts suggest that preparation method shows a low repeatability
3
)
3
× 6H O precursor (Figs.
2
0
2
2
1
4

for catalysts with high amount of ceria introduced from Ce(NO
corresponding to Ce/Co(Ni) molar ratio higher than 0.5.
3
)
3
× 6H O precursor
2
The reduction behavior of the ceria support and cobalt- and nickel-based catalysts was
studied by H -TPR, as shown in Figs. 1a and b. The H -TPR profile of ceria showed peaks at
2
2
two different temperature ranges. First peak at low temperatures ranging from 240 to 500 °C
is assigned to the reduction of adsorbed and surface lattice oxygen of ceria. Whereas the
second peak above 580 °C is due to the reduction of bulk CeO
xCe/CeO catalysts (Fig. 1a), hydrogen consumption corresponding to first peaks below 200
C is attributed to the removal of oxygen species adsorbed on ceria oxygen vacancies [11, 28]
and the reduction peak centered at about 240 °C is due to the reduction of Co to CoO [44–
6]. Whereas the next peak appearing at temperature of about 290 °C – 300 °C is the result of
the reduction of CoO that weakly interacts with ceria to metallic cobalt [45, 46] or the
reduction of independent Co that weakly interacts with ceria directly to metallic cobalt
44]. This peak is small for both Co/CeO and Co-0.1Ce/CeO catalysts which suggests that
2
[41-43]. In the case of Co-
2
°
3
O
4
4
3
O
4
[
2
2
these catalysts contain small amount of cobalt species that interact weakly with the support.
But intensity of this peak increases with the increase of content of ceria introduced from
Ce(NO
3
)
3
× 6H O precursor which is in good accordance with STEM-EDS data which
2
indicated that amount of unsupported and isolated from ceria cobalt phase (weakly interacting
with the support) increased with the increase of Ce/Co molar ratio. The next peak, with the
maximum in the temperature range from 420 to 470 °C depending on the catalyst, is attributed
to the reduction of cobalt oxide species which strongly interacted with the ceria and it
probably overlapped with the surface reduction of ceria [44-46]. According to Wu et al. [47]
and Luo et al. [44], however, surface ceria reduction is attributed to peak detected at lower
temperature (240 °C) and overlapped with reduction of Co
3
O
4
to CoO. Undoubtedly, the last
catalysts is specifically
peak above 580 °C observed in H -TPR profiles of Co-xCe/CeO
2
2
1
5

related to bulk reduction of ceria [12, 25]. The same peak is also observed in H
2
-TPR profiles
of all considered Ni-xCe/CeO catalysts and it is assigned the same process (Fig. 1b). Because
2
the adsorbed oxygen molecules are very reactive oxygen species and ready to be reduced by
hydrogen at low temperature, the peak located below 220 °C can be ascribed to the reduction
of the adsorbed oxygen species in ceria oxygen vacancies [48]. The next broad peak in H
2
-
TPR profiles of Ni-xCe/CeO catalysts in the temperature range from about 240 to 420 °C
2
included the reduction of a dispersed NiO phase interacting weakly with ceria and step–by-
step reduction of a well-dispersed phase interacting strongly with ceria [49]. However, it
could not be excluded that surface ceria reduction also occurred in the discussed temperature
range. As can be seen in both Fig. 1a and Fig 1b, the peak of reduction of a well-dispersed
phase interacting strongly with ceria shifts gradually to lower value after increasing
Ce/Co(Ni) molar ratio from 0.1 to 1, indicating weakened interaction between cobalt or nickel
species and ceria at higher promoter content [49]. A summary of H
Table 3. The highest values of H/M and the hydrogen consumption were obtained for
Co/CeO and Ni/CeO catalysts. These values were higher than the calculated ones for the
reduction of Co or NiO to metallic cobalt or nickel, respectively (based on the
cobalt/nickel amount calculated from XRF) which is typical of CeO supported catalysts and
2
-TPR results is shown in
2
2
3
O
4
2
is a consequence of the high hydrogen spillover rate and confirmation of surface reduction of
ceria below 500 °C along with cobalt/nickel species reduction [50]. An examination of the
data reported in Table 3 reveals a gradual suppression of hydrogen consumption with the
increase of the ceria introduced from Ce(NO
values is Co/CeO (2.91) > Co-0.1Ce/CeO (2.78) > Co-0.5Ce/CeO
(2.30) > Ni-0.1Ce/CeO (2.23) > Ni-0.5Ce/CeO
2.03). Moreover, in the case of Co-0.5Ce/CeO , Co-Ce/CeO
, quite a good agreement between calculations of H/M values and their expectations
3
)
3
× 6H
2
O precursor. The order in the H/M
2
2
2
(2.74) > Co-Ce/CeO
2
2
(
2.71) and Ni/CeO
2
2
2
(2.07) > Ni-Ce/CeO
(
2
2
, Ni-0.5Ce/CeO and Ni-
2
Ce/CeO
2
1
6

2
.67 and 2.0 for Co-xCe/CeO
2
and Ni-xCe/CeO catalysts, respectively can be noticed.
2
Because the spillover phenomenon of hydrogen occurs when there is a direct interaction
between the ceria surface and the transition metal, a gradual suppression of hydrogen
consumption with the increase of the ceria introduced from Ce(NO
3
)
3
× 6H O precursor is not
2
surprising. The EDS maps (Figs. S4-S5 and S7-S8) show that there are areas of the lack of
intimate contact between cobalt/nickel species and ceria in the case of catalysts with higher
Ce/Co(Ni) molar ratio. Besides the intimate contact between metal dissociation source and
ceria, the high dispersion and small size of metal particles are crucial to facilitate the
hydrogen spillover. The well dispersed metal particles close to the ceria activate the hydrogen
dissociation and by hydrogen spillover favors the reduction of ceria surface [51, 52]. Sermon
and Bond [53] emphasized the importance of the degree of dispersion and contact of the
catalyst phases in studies on hydrogen spillover and the confusion which can arise in
comparing systems in which these parameters are unknown indicating that the extent of
hydrogen spillover can be lesser in the catalysts where contact between the two phases is
insufficient. Moreover, it should be remembered that spillover phenomena depends on many
other factors such as the size of the ceria nanoparticle the stability of the Ce 4f levels [54],
morphology of ceria [55], the hydroxylation of ceria surface [56], the dispersion state of ceria
[
57].
Literature data suggests that metallic nickel is the active phase for C-C bond cleavage
26, 27, 58]. There remains a disagreement on the role of metallic cobalt and CoO during SRE
[
reaction. Numerous papers assumed that the metallic cobalt is the most active form of cobalt
in SRE [16, 59-61] but the literature does not provide a definitive consensus on this issue.
0
2+
Recently, the ideas that Co /Co ratio forms during SRE might favor different reaction paths
has been gaining ground. According to Martono and Vohs [62, 63], metallic cobalt is most
active for decomposition and decarboxylation reactions while CoO favors dehydrogenation of
1
7

ethanol to acetaldehyde. The similar conclusion were drawn by Passos et al [4] who also
stated that the presence of CoO species plays a role in the selectivity towards ethanol
dehydrogenation to acetaldehyde and formation of methane because of acetaldehyde
decomposition. Karim et al. [52] found that metallic cobalt is active for water gas shift
reaction whereas methanation increases if CoO phase becomes more predominant. Whereas
Turczyniak et al. [12] showed that in the presence of metallic cobalt carbon monoxide is
favored , while CoO species promotes carbon dioxide and acetaldehyde yields. The profiles of
O
2
-TPO for CeO
are presented in Figs. 2a and b. In case of both cobalt- and nickel-based catalysts a few
oxidation peaks are observed. Because similarly to Luo et al. [44] studies , O -TPO profile of
2
, Co-xCe/CeO
2
and Ni-xCe/CeO catalysts, depending on their composition,
2
2
ceria indicates that there was no detectable oxygen consumption peak for pure ceria it could
be supposed that all peaks observed in Figs. 2a and b result from oxidation of cobalt or nickel
species. Moreover, Luo et al. [44] showed that for unsupported cobalt metal, a broad oxygen
consumption peak occurs at 393 °C due to the oxidation of metallic cobalt to Co
obtained by Sewell et al. [64], however, indicated that O -TPO profiles of unsupported cobalt
metal are characterized by two oxygen consumption maxima at 298 and 583 °C
corresponding to the sequential oxidation of metallic cobalt first to CoO and then to Co
The maxima of the last peaks present in the O -TPO profiles of all ceria supported cobalt
3
4
O . Results
2
3
4
O .
2
catalysts (Fig. 2a) are below 300 °C which indicates that interaction between cobalt species
and ceria can contribute to easier oxidation of cobalt species as it was suggested by Luo et al.
[
44]. Furthermore, oxidation of cobalt species of ceria supported cobalt catalysts is rather
complex as can be seen in Fig. 2a. As it was suggested by Greluk et al. [11], the first peaks
can be attributed to oxidation of the surface layer of cobalt crystallites whereas a broad
oxygen consumption peaks at higher temperature of Co-xCe/CeO
2
catalysts probably resulted
from oxidation of metallic cobalt. The shift of high-temperature oxidation peaks towards
1
8

higher temperatures with the increase of ceria content introduced from Ce(NO
precursor is related to the decrease of metal dispersion degree [44]. Metallic cobalt particles
of Co/CeO and Co-0.1Ce/CeO catalysts oxidized more easily at low temperature than those
of Co-0.5Ce/CeO and especially of Co-Ce/CeO catalysts because cobalt particles for the
catalysts with lower amount of ceria introduced from Ce(NO × 6H O precursor were
3
)
3
× 6H
2
O
2
2
2
2
3
)
3
2
smaller. It means that the higher dispersion of cobalt species corresponds to lower oxidation
temperature of metallic cobalt [44]. According to Luo et al. [44], high cobalt dispersion can
ensure a more profound contact and interaction between cobalt species and ceria, hence
facilitating oxygen supply from ceria to cobalt species and promoting the oxidation of cobalt
phase. The similar conclusions can be drawn for Ni-xCe/CeO
2
catalysts (Fig. 2b). In O -TPO
2
profiles of these catalysts, there are also low-temperature peaks which can be attributed to
oxidation of the surface layer of nickel crystallites and a broad oxygen consumption peaks at
higher temperature resulted probably from oxidation of metallic nickel. The shapes of O
TPO profiles and the locations of oxygen consumption peaks for Ni/CeO , Ni-0.1Ce/CeO
and Ni-0.5Ce/CeO
of these catalysts is comparable as it was indicated by hydrogen chemisorption or XRD
results (Tables 1 and 2). But the nickel crystallites of Ni-Ce/CeO catalyst are much larger
and its O -TPO profile differs from the remaining nickel-based catalysts. It is noticeable that
it is more difficult to oxidize the nickel species of this catalyst.
.2. Catalytic performance test of Co-xCe/CeO and Ni-Ce/CeO
reaction
The activity test of Co-xCe/CeO
2
-
2
2
2
catalysts are very similar (Fig. 2a) because the size of nickel crystallites
2
2
3
2
2
catalysts in SRE
2
and Ni-Ce/CeO catalysts in SRE reaction was
2
evaluated by the reactants conversion and the selectivity to the products. To compare the
catalytic properties of the catalyst and influence of ceria content on it, the condition of the
process which was not to ensure complete ethanol conversion was chosen. Because space
1
9

velocity of 60000 mL/g h allowed 100 % ethanol conversion over nickel-based catalysts (see
Supporting Information, Fig. S9), it was increased to 120000 mL/g h in case of catalytic tests
for these catalysts by decreasing the catalyst mass. The comparison of catalysts at full
conversion is meaningless because obtained results can be the consequence of an excess of
contact time (i.e. of catalyst mass) for the test conditions selected. Therefore, it cannot be sait
if the whole catalyst sample in the catalytic bed is really working or part of it cannot work
because ethanol is exhausted before reaching it. It is also possible that an activity decay is not
observed because, if catalyst amount is in excess, the decay is compensated by a higher
amount of catalyst taking part in the reaction (without still reaching the total of the bed). The
results obtained from catalytic tests at 420 °C for H
hours of SRE reaction over Co-xCe/CeO and Ni-Ce/CeO
b (see also Supporting Information, Figs. S10 and S11).
2
O/EtOH molar ratio of 12/1 after 21
2
2
catalysts are shown in Figs. 3a and
As it was suggested by Song et al. [65], the first step of a typical reaction pathway for
ethanol decomposition over cobalt sites involves the formation an ethoxy species through the
dissociative adsorption of ethanol. The second step consists of eliminating hydrogen from the
ethoxy intermediate which produces an aldehyde intermediate. Acetaldehyde can desorb to
the gas phase, decompose to methane and carbon monoxide, or undergo further abstraction
which leads to the formation of acetyl species [66]. This reaction pathways observed for
ethanol adsorption over nickel sites are very similar. As demonstrated previously by Gates et
al. [67], the decomposition of ethanol on the nickel surface forming these products occurs by
a sequence of bond scission: (i) the scission of the H–O– bond of the ethanol molecule to
form adsorbed ethoxy; (ii) the scission of the –C–H bond of the CH
an adsorbed intermediate of acetaldehyde; (iii) the scission of C–C bond of intermediate of
acetaldehyde adsorbed to form adsorbed –CH and –CO [68]. It is well known that the
oxidation products, resulting from ethanol transformation (e.g. acetate) are present to greater
2
group in ethoxy to form
3
2
0

extent over metal oxide supported metals because of the availability of oxygen on surfaces of
these materials [66]. Both acetaldehyde and acetyl species can be oxidized to acetate species
by hydroxyl groups or by oxygen from the support. The acetate species can be further
oxidized to carbonate species that decompose to give carbon dioxide. Finally, acetaldehyde
can contribute to the formation of acetone through condensation of acetate or acetyl species.
These reactions take place mainly over the support and depend on the nature of the metal
oxide [66].
The discussion on results concerns only those obtained after 21 hours of the process and
presented in Figs. 3a and b. With this aim, bar graphs were chosen because they are better for
comparing the small differences in the data among the groups of Co-xCe/CeO
Ni-xCe/CeO (Fig. 3b) catalysts. Using bar graphs (Figs. 3a and b) data can be shown in a
clear manner and any differences in activity and selectivity to products of the catalysts during
SRE process in dependence of ceria amount introduced from Ce(NO × 6H O precursor can
2
(Fig. 3a) and
2
3
)
3
2
be noticed better. Moreover, error bars as graphical representations of the uncertainty were
used on Figs. 3a and b to indicate the standard deviation in reported measurement. Usually the
decreasing of catalyst stability is determined by measuring activity or/and selectivity as a
function of time. The changes in trend over the time displayed as line plots (see Supporting
Information, Figs. S10 and S11) are not discussed. However, the loss over time of catalytic
activity and selectivity is observed in the case of all catalysts (see Supporting Information,
Figs. S10 and S11) because of carbon formation (described in detail in the section 3.3). The
time of 21 hours of SRE reaction was sufficient to observe the differences in catalytic
performance between catalysts, although it was too short to obtain the steady state of the
catalysts under SRE conditions. It means that the non-steady-state behavior caused by the
deactivation of the catalyst is illustrated on Fig. 3a and b.
2
1

The increase of ceria amount introduced from Ce(NO
3
)
3
× 6H O precursor does not
2
have significant influence on the ethanol conversion over cobalt-based catalysts but slight
differences in the selectivity to some products in dependence of the Ce/Co molar ratio after 21
hours of SRE process are observed (Fig. 3a). All cobalt-based catalyst, with the exception of
Co-Ce/CeO
reaction, i.e. hydrogen (~75% for Co/CeO
.5Ce/CeO ) and carbon dioxide (~39% for Co/CeO
Co-0.5Ce/CeO ) after 21 hours of SRE process. The formation of these products is slightly
lower only in the presence Co-Ce/CeO catalyst (H ~72%, CO ~ 32%) (Fig. 3a). As it was
2
sample, exhibit a comparable selectivity to two most desirable products of the
, ~76% for Co-0.1Ce/CeO , ~75% for Co-
, ~41% for Co-0.1Ce/CeO , ~38% for
2
2
0
2
2
2
2
2
2
2
reported by Zhang [69], in the case of ceria, water is chemisorbed on vacancies of its surface
forming hydroxyls, which migrate through oxygen vacancies on the ceria surface. These
oxygen species (including hydroxyls) originating from water can oxidize CO and CH
x
to form
H
2
and CO . Therefore, if there are sufficient oxygen species on the catalyst surface, the
x
intermediate products can be fully converted to carbon dioxide and hydrogen by WGS
reaction (reaction 1) [33]:
CO + H
All cobalt-based catalysts exhibit highest carbon dioxide formation (~39% for Co/CeO
41% for Co-0.1Ce/CeO , ~38% for Co-0.5Ce/CeO2, ~32% for Co-Ce/CeO ) in comparison
with production of carbon monoxide (~6% for Co/CeO , ~5% for Co-0.1Ce/CeO , ~5% for
Co-0.5Ce/CeO2, ~7% for Co-Ce/CeO ) after 21 hours of SRE reaction which suggests that
WGS reaction is favored by cobalt-based catalysts supported on ceria and is consistent with
conclusion of Marcos et al. obtained for Co/CeO catalyst [3]. Also similarly to results
obtained in this study, Marcos et al. [3] showed that high ethanol conversion of Co/CeO
catalyst was directed to the formation of acetaldehyde (~13% for Co/CeO , ~12% for Co-
.1Ce/CeO , ~16% for Co-0.5Ce/CeO2, ~19% for Co-Ce/CeO ) and acetone (~39% for
2
O ↔ CO
2
+H
2
(1)
2
,
~
2
2
2
2
2
2
2
2
0
2
2
2
2

Co/CeO
1 hours of SRE process. According to Carvalho et al. [24], the formation of acetaldehyde in
the presence of Co/CeO catalysts suggests that the basicity of ceria (owing to their basic
2
, ~36% for Co-0.1Ce/CeO
2
, ~38% for Co-0.5Ce/CeO2, ~37% for Co-Ce/CeO ) after
2
2
2
2
-
lattice O sites on the surface) stabilized the adsorbed acetaldehyde. As it was explained by
the authors [24], the oxygen radicals on the surface of ceria can favor the dissociative
adsorption of ethanol (reaction 2) with consequent dehydrogenation to acetaldehyde (reaction
3
) as follows:
2
-
-
-
C
C
2
2
H
5
5
OH + O → C
2
H
5
O + OH
(2)
(3)
-
-
+O2-
H
O + OH → CH
3
CHO + H
2
Because ethanol dehydrogenation to acetaldehyde is enhanced on the basic supports [70], also
Sohn et al [71] suggested that this reaction mainly takes place on the reduced surface having
3
+
Ce sites, which are more basic. The results presented in Fig. 3a show that Co/CeO
.1Ce/CeO catalysts exhibit the lowest selectivity to acetaldehyde in comparison with two
remaining cobalt-based catalysts with higher Ce/Co molar ratio. The selectivity to
acetaldehyde is comparable for Co/CeO and Co-0.1Ce/CeO catalysts and slightly increases
with the increase of ceria amount introduced from Ce(NO × 6H O precursor. Barroso et al.
72] suggested that the highest formation of hydrogen and the lowest acetaldehyde production
2
and Co-
0
2
2
2
3
)
3
2
[
of ceria containing cobalt-based catalysts come out of the highest dispersion of cobalt in the
presence of ceria. Also Marcos et al. [3] ascribed the higher hydrogen yield of the mixed
Co/La
2
O
3
-CeO
2
-SiO
2
with varied La
2
O
3
and CeO content to smaller cobalt species crystallite
2
size, greater metal area and dispersion which meant more active sites available to react in
comparison with the catalyst without mentioned promoters. Because in the case of considered
Co-xCe/CeO
Table 1 and Table 2) leading to larger cobalt active phase for breaking C–C and C–H bonds,
both Co/CeO and Co-0.1Ce/CeO catalysts indicated less selectivity to acetaldehyde after 21
2
catalysts, only low Ce/Co molar ratio allows to improve cobalt dispersion
(
2
2
2
3

hours of SRE process (Fig 3a). Because besides the presence of aldehyde among the products
of SRE reaction over all studied Co-xCe/CeO catalysts also acetone was detected in the
2
significant amount, it suggests that cobalt-based catalysts are not very active in breaking C–C
bonds at 420 °C. The presence of acetone which is formed either by coupling of two acetate
molecules or the disproportionation of two acetyl species [66] is very undesirable because
acetone can polymerize to produce carbon which would cause catalyst deactivation [73].
Since small amount of ethylene in the presence of all Co-xCe/CeO
was also detected, it is likely that dehydration of ethanol occurred (reaction 4):
OH → C +H
The obtained results indicates that the lowest amount of C2 and C3 products was observed
over Co-0.1Ce/CeO catalysts which suggests that the presence of cobalt species with a small
2
(~1.5% for all catalysts)
C
2
H
5
2
H
4
2
O
(4)
2
particle size and enhancement of metal-support interactions guarantee higher activity in C─C
bond cleavages and is in good accordance with results obtained by Song et al [17] and
Carvalho et al. [24]. The low selectivity to methane over all catalysts (~3% for all catalysts)
after 21 hours of SRE process suggests that the Co-xCe/CeO
reforming of methane (reaction 5) [74]:
2
catalysts promoted the steam
CH
4
+H
2
O → CO + 3H
2
(5)
Both polimeryzation of ethylene (reaction 6) and methane dehydrogenation (reaction 7) are
the possible routes for carbon deposition on cobalt-based catalysts. However, the carbon
formation on the studied Co-xCe/CeO
cannot be excluded:
2
catalysts because Boudouard reaction (reaction 8)
C
2
H
4
→ coke
→ C + H
CO → C + CO
(6)
(7)
(8)
CH
4
2
2
2
4

According to Carvalho et al. [24]. both methane decomposition and/or the Boudouard reaction
include hydrogen and carbon dioxide as by-products and their contribution can influence the
high selectivity to mentioned products under SRE conditions over Co/CeO
2
catalysts. From
the distribution of the main products (H , CO, CO , CH and C , acetaldehyde, acetone), it
2
2
4
2
H
4
can be inferred that the predominant reactions are: the dehydrogenation of ethanol to
acetaldehyde (reaction 9), followed by the complete acetaldehyde decomposition to carbon
monoxide and methane (reaction 10) and/or decomposition of ethanol (reaction 11), followed
by WGS reaction (reaction 1) and dehydration of ethanol (reaction 4). Moreover, aldol
condensation reaction of acetaldehyde to acetone, carbon monoxide and hydrogen (reaction
1
2) occurred over the Co-xCe/CeO under studied conditions rather than the steam reforming
2
of acetaldehyde (reactions 13) as it is suggested by large amount of acetone detected among
by-products of SRE reaction.
C
2
H
5
OH → CH
CHO → CH
OH → CH
CHO → CH
CHO + H
Regarding the products distribution, all Ni-xCe/CeO
products as Co-xCe/CeO
ethanol transformations over Ni-xCe/CeO
in the case of Co-xCe/CeO catalysts. However, nickel displayed a higher activity for the
3
CHO + H
+ CO
+ CO + H
COCH
2
(9)
CH
3
4
(10)
C
2
H
5
4
2
(11)
2
CH
3
3
3
+ CO + H
2
(12)
(13)
CH
3
2
O → 2CO + 3H
2
2
catalysts displayed the same
2
catalysts but with different distributions (Fig. 3b). It suggests that
catalysts proceeded through the same reactions as
2
2
methanation leading to increased methane selectivity which is highly undesirable since it
consumes hydrogen [70] (reactions 14).
CO + 2H
2
→ CH
4
+ H
2
O
(14)
2
5

The selectivity to methane was slightly lower for the Ni-0.5Ce/CeO
2
(~17%) and Ni-Ce/CeO
~14%) catalysts among the studied nickel-based catalysts (~19% and for Ni/CeO and ~23%
) after 21 hours of SRE process (Fig. 3b) which indicates that methanation was
× 6H
precursor. Simultaneous increase of hydrogen and carbon monoxide production in the
presence of the Ni-0.5Ce/CeO (H ~84%, CO ~ 8%) and Ni-Ce/CeO (H ~84%, CO ~ 15%)
catalysts can suggests that catalysts with molar ratio of Ce/Ni above 0.5 promoted the steam
reforming of methane (reaction 5) (H ~75%, CO ~ 7% for Ni/CeO , H ~81%, CO ~ 7% for
2
(
2
Ni-0.1Ce/CeO
2
suppressed in the presence of a high content of ceria introduced from Ce(NO
3
)
3
2
O
2
2
2
2
2
2
2
Ni-0.1Ce/CeO2,) It is in accordance with hypothesis that ceria promotes steam reforming
reactions due to an improvement in the adsorption and dissociation of water molecules on the
surface of catalyst [75]. But promotional effect of ceria on the reactions responsible for
carbon formation, methane cracking (reaction 7) may not be excluded [29]. On the other hand,
Xu et al. [27] proposed that the strong metal-support interactions between nickel and ceria
perturb the electronic and chemical properties of nickel adatoms reducing their ability to
break C─O bonds and substantially decrease the carbon monoxide methanation activity of
nickel which suggests that methane production over Ni-0.1Ce/CeO catalysts should be the
2
lowest because it exhibited the highest nickel active phase dispersion (Table 1 and Table 2)
and the strongest metal-support interactions (Fig. 1). However, the highest selectivity to
methane (~23%) observed in the presence of this catalyst are in contradiction with this thesis.
On the other hand, this catalyst allowed to significantly reduce acetaldehyde formation to
~
8% in the comparison to other studied nickel-based catalysts after 21 hours of SRE process
~12% for Ni/CeO , ~14% for Ni-0.5Ce/CeO2, ~17% for Ni-Ce/CeO ) (Fig. 3b). Therefore,
both high selectivity to hydrogen (~81%) and low selectivity to acetaldehyde (~8%) over Ni-
.1Ce/CeO catalyst could be ascribed to smaller nickel crystallite size with respect to other
studied nickel-based catalysts (Table 2) due to the dispersing effect obtained by introduction
(
2
2
0
2
2
6

of low amount of ceria from Ce(NO
3
)
3
× 6H O precursor [34, 75]. Similarly as in the case of
2
cobalt-based catalysts, a smaller crystallite size of nickel active phase (Table 2) results in
enhanced metal-support interactions which leads to increase metal active sites for breaking C–
C and C–H bonds and facilitates the ethanol conversion and other C
Ebiad et al. [76] ascribed superior activity of Ni/CeO -ZrO catalyst containing 2% of the
active phase to the small nickel particle size. The increase of nickel active phase dispersion
and metal-support interaction obtained by introduction of CeO to Al -La support was
also an explanation for the complete conversion of ethanol in SRE process in the presence of
obtained Ni/Al -La -CeO catalyst given by Osorio-Vargas et al. [29]. Similar
2
intermediates. Also
2
2
2
2
O
3
2
O
3
2
O
3
2
O
3
2
conclusions were drawn for promoting effect of ceria in the case of SRE reaction over Ni-
Ce/MMT [35] and CeNi/SBA-15 [36] catalysts. In the case of studied catalysts ethanol
conversion after 21 hours of SRE process is the highest over Ni-0.1Ce/CeO
follows the order of Ni-0.1Ce/CeO (~52%) > Ni/CeO (~48%) ~ Ni-0.5Ce/CeO
Ni-Ce/CeO (~43%) (Fig. 3b) which corresponds to the sequence in which nickel crystallites
size of these catalysts increases (Table 2). In comparison to ethanol conversion obtained for
Co-xCe/CeO catalysts (Fig 3a and Fig. S9) under the same SRE conditions (for space
velocity of 60000 mL/g h), the activity of Ni-xCe/CeO catalysts (Fig. S10) is much higher.
These results are in good accordance with those obtained by Zhang et al [77] which also
indicated that Ni/CeO catalyst had stronger capacity for breaking C─C bond in ethanol than
Co/CeO catalyst. Fig. 3a indicates that the conversion of Co-xCe/CeO catalysts was directed
to the formation of acetone which is also with good agreement with the results obtained by
Zhang et al [77]. Whereas the production of this product in the case of Ni-xCe/CeO catalysts
Fig. 3b) ,despite the two times higher space velocity (1200000 mL/g h), was rather small
~6% for Ni/CeO , ~3% for other nickel-based catalysts) indicating that these catalysts are
2
catalyst and it
2
2
2
(~47%) >
2
2
2
2
2
2
2
(
(
2
more capable of cleaving the C–C bond than the studied cobalt-based ones which was also
2
7

confirmed by the lower selectivity to ethylene (<0.7% for all nickel-based catalysts)
demonstrated by Ni-xCe/CeO catalysts after 21 hours of SRE process. Because the Ni-
2
xCe/CeO
xCe/CeO
2
2
catalysts exhibited lower selectivity to C2 and C3 intermediates than the Co-
catalysts, more C1 products in their presence were produced. It is mainly carbon
dioxide which was produced in much larger amounts than carbon monoxide indicating nickel-
based catalysts’ capability to facilitate WGS reaction (reaction 1). Since the highest CO /CO
ratio of 8.7 is demonstrated over Ni-0.1Ce/CeO catalyst after 21 hours of SRE process (Fig.
b), it can suggest that WGS reaction was especially favored by the small addition of ceria
from Ce(NO × 6H O precursor to nickel active phase which can be attributed to the strong
2
2
3
3
)
3
2
interaction between nickel crystallites and ceria occurring in case of this catalysts as it was
confirmed by TPR method (Fig. 1b).
It must be emphasized that all considerations about catalytic properties of co-xCe/CeO
2
and Ni-xCe/CeO catalyst concern the state which was obtained after 21 hours of SRE
2
process. Because both cobalt-and nickel-catalysts underwent gradual deactivation due to
carbon formation (described in detail in the section 3.3) under SRE conditions, it is possible
that observed dependence could slightly change with the further increase in time.
3
.3. Analysis of the carbon deposits on the used Co-xCe/CeO
To examine the possible carbon deposition on the catalysts during SRE process, post
reaction characterization experiments were performed. The amount of carbon deposition over
the Co-xCe/CeO2 and Ni-xCe/CeO2 catalysts under SRE conditions at 420 °C for H O/EtOH
2
and Ni-xCe/CeO
2
catalysts
2
molar ratio of 12/1 was determined by using TG analysis (Fig 4a and b and Table 4) whereas
to investigate the structure and morphology of the carbonaceous materials deposited on the
used catalysts exposed to SRE reaction during catalytic tests under mentioned conditions, the
SEM (Fig. S12) and TEM analysis was examined (Figs. 5 and 6).
2
8

The redox properties of ceria and the high lability of lattice oxygen are among the most
important factors contributing to carbon minimization by ceria as a promoter under SRE
conditions. The highly mobile oxygen in ceria can instantaneously react with carbon species
formed during the reaction and maintain a clean metal surface whereas the oxygen vacancies
are readily replenished by steam from the feed [35, 75]. As it is suggested by Osorio-Vargas
et al. [29] the oxygen lattice provided by ceria may oxidize the solid carbon according to the
following reaction:
C
(s) + O(l) → O(l-1) + CO
(15)
On the other hand, studies by Senanayake et al [78] indicate that the electronic
perturbations between nickel and ceria can limit the ability of nickel to break C–O bonds
which could probably suppress the activity of carbon monoxide disproportion to produce
carbon deposits [66]. Whereas Yang [79] suggested that electrons released from the oxygen
vacancies in ceria could migrate through the Ni-CeO
interface to the unfilled d-orbital of Ni0
2
0
and due to the electron-rich property of Ni , decomposition of methane could be effectively
inhibited. The conclusion of these studies concerning influence of ceria on the minimization
of carbon during steam reforming reactions suggest that both these promoters reduce the rate
of carbon deposition and accelerates the rate of carbon gasification [75].
The results obtained from the thermogravimetric studies (Fig. 4a and b and Table 4)
seem to indicate that introduction of ceria from Ce(NO
3
)
3
× 6H O precursor to cobalt- and
2
nickel-based catalysts allow to minimize the amount of carbon formed on their surface under
SRE conditions. However, the detailed analysis of the products distribution after 21 hours of
this process (Figs 3a and b) and the SEM and TEM images (Figs. S12 and 5 and 6) rather
suggests that only small amount of ceria corresponding to Co(Ni)/Ce molar ratio of 0.1
prevents the cobalt- and nickel-based catalysts from carbon accumulation.
2
9