Unsupported versus alumina-supported Ni nanoparticles as catalysts for steam/ethanol conversion and CO2 methanation

Article abstract of DOI:10.1016/j.molcata.2013.11.006

Nickel nanoparticles (NPs, particle radius <8 nm) were prepared via reduction method in aqueous solution and were structurally and morphologically characterized by X-ray diffraction (XRD) and scanning electron microscopy (SEM). A DC-SQUID magnetometer was also used to determine the particle size. As a reference, a high loading (125% Ni wt_Ni/wt_support) Ni/Al₂O₃, prepared through wet impregnation technique, was also studied. The results of the catalytic tests in ethanol conversion in the presence of steam on unsupported Ni NPs at 523 K show already an ethanol conversion higher than 90% together with high yields in methane, CO and H ₂. These NPs actually behave as very active ethanol decomposition catalysts, in these conditions. Increasing the reaction temperature (673 K) they convert into good steam reforming catalysts. On the other hand, the conventional alumina-supported Ni catalyst is poorly active in dehydrogenation to acetaldehyde below 673 K, but behaves as a good steam reforming catalyst at this temperature. In CO₂ methanation at 773 K the best catalytic activity was obtained on 125% Ni/Al₂O₃, while Ni NPs result poorly active in catalysing this reaction. The characterization of the exhaust materials shows: (i) hexagonal Ni particles on both catalysts after ethanol conversion in the presence of steam, (ii) cubic nickel particles on both samples after methanation of CO₂. Mechanistic implications of these results are discussed.

Full text of DOI:10.1016/j.molcata.2013.11.006

Journal of Molecular Catalysis A: Chemical 383–384 (2014) 10–16
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Unsupported versus alumina-supported Ni nanoparticles as catalysts
for steam/ethanol conversion and CO methanation
2
a
b
a
c
Paola Riani , Gabriella Garbarino , Mattia Alberto Lucchini , Fabio Canepa ,
b,∗
Guido Busca
a
Università degli Studi di Genova, Dipartimento di Chimica e Chimica Industriale (DCCI), Via Dodecaneso, 31 16146 Genova, Italy
Università degli Studi di Genova, Dipartimento di Ingegneria Civile, Chimica e Ambientale (DICCA), Laboratorio di chimica delle superfici e catalisi,
b
P.zzale Kennedy, 1 16129 Genova, Italy
c
IMEM-CNR Unità di Genova and Università degli Studi di Genova, Dipartimento di Chimica e Chimica Industriale (DCCI), Via Dodecaneso,
31 16146 Genova, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 12 August 2013
Received in revised form 21 October 2013
Accepted 9 November 2013
Available online 20 November 2013
Nickel nanoparticles (NPs, particle radius <8 nm) were prepared via reduction method in aqueous solu-
tion and were structurally and morphologically characterized by X-ray diffraction (XRD) and scanning
electron microscopy (SEM). A DC-SQUID magnetometer was also used to determine the particle size.
As a reference, a high loading (125% Ni wtNi/wtsupport) Ni/Al2O3, prepared through wet impregnation
technique, was also studied. The results of the catalytic tests in ethanol conversion in the presence of
steam on unsupported Ni NPs at 523 K show already an ethanol conversion higher than 90% together
with high yields in methane, CO and H2. These NPs actually behave as very active ethanol decomposition
catalysts, in these conditions. Increasing the reaction temperature (673 K) they convert into good steam
reforming catalysts. On the other hand, the conventional alumina-supported Ni catalyst is poorly active
in dehydrogenation to acetaldehyde below 673 K, but behaves as a good steam reforming catalyst at this
temperature. In CO2 methanation at 773 K the best catalytic activity was obtained on 125% Ni/Al2O3,
while Ni NPs result poorly active in catalysing this reaction. The characterization of the exhaust materi-
als shows: (i) hexagonal Ni particles on both catalysts after ethanol conversion in the presence of steam,
Keywords:
Nickel nanoparticles
Ethanol conversion
Steam reforming
Methanation, Nickel catalysts
(ii) cubic nickel particles on both samples after methanation of CO2. Mechanistic implications of these
results are discussed.
©
2013 Elsevier B.V. All rights reserved.
1
. Introduction
The steam reforming of (bio)ethanol (ESR) represents a possible
ones [12–15], in relation to the lower production of carbon species
that finally cause catalyst deactivation [16], a very relevant problem
for ESR. In fact, unsupported cobalt nanoparticles (NPs) behave as
excellent catalysts for ESR at low temperature with poor methane
production [15], with performances even better than highly loaded
way to produce biosyngas and biohydrogen, as a step in the pro-
duction of renewable fuels [1–9]. At least three families of catalysts
are under study to perform efficiently this reaction. Nickel-based
catalysts are quite efficient in catalysing ESR giving rise to hydro-
gen yield only at quite high temperature. In these conditions, CO is
produced in relevant amounts, thus the syngas cannot be directly
fed to low temperature fuel cells. At low temperature, the thermo-
dynamics allows very small amount of CO, but ethanol conversion
may be slower and methane may be present in huge quantity in
the resulting syngas. This is what really occurs with noble metal
catalysts as well as with Ni ones in the low temperature ESR reac-
tion [10,11]. Cobalt catalysts appear more promising than Ni based
Co/Al O catalysts.
2
3
The presence of methane in the gas deriving from ESR reduces
the achieved hydrogen concentration and, in case, the fuel cell
performances. Two different ways may be responsible for the for-
mation of methane during ESR: one is the methanation reaction,
favoured thermodynamically more, lower is the reaction temper-
ature. The other is the decomposition of ethanol itself or of its
conversion products, like acetaldehyde and acetic acid. The reason
why some catalysts (mainly Co-based) seem to be more selective to
steam reforming products while others (Ni-based and noble metal
based) produce more methane it is still under discussion. Another
point still debated in the literature concerns the best metal oxida-
tion state for ESR, metallic or partly unreduced.
∗ Corresponding author at: University of Genoa, Department of Civil, Chemical and
Environmental Engineering, University of Genova, DICCA, Fiera del Mare, I-16129
Genoa, GE, Italy. Tel.: +390103536024; fax: +390103536028.
As a further step in our studies, we wanted to investigate the
behaviour of unsupported Ni NPs in the same reaction, to compare
their behaviour with that of unsupported Co NPs and of Ni/Al2O3
E-mail address: Guido.Busca@unige.it (G. Busca).
1
381-1169/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.molcata.2013.11.006

P. Riani et al. / Journal of Molecular Catalysis A: Chemical 383–384 (2014) 10–16
11
conventional catalysts. Additionally, we became interested to the
CO2 methanation reaction, an interesting possible way to convert
this greenhouse gas to useful products. This reaction can also give
information on the methane production step upon ESR. For these
reasons we tested our Ni based materials also for this reaction.
In this paper we will present our attempts to prepare
unsupported Ni metal NPs, their structural and morphological
characterization obtained coupling XRD and FE-SEM to magnetic
measurements, in order to achieve a precise particle size measure-
ment, and their behaviour in the cited catalytic reactions.
2
. Experimental
2.1. Materials preparation
•
NiCl2 6H O 99.5 wt.% and NaBH4 99.5 wt.% were purchased
2
from Sigma Aldrich and used without other purifications. The Ni
NPs were prepared by a reduction method in aqueous solution. Two
Fig. 1. FE-SEM image of as prepared Ni nanoparticles obtained from a dilute solution
(10 M, sample A). The NPs diameters range from 40 to 80 nm, but it is noteworthy
that on the upper right corner of this figure there is an enlarged nanoparticle that
seems to be formed by smaller NPs.
−
4
−4
different Ni ions concentrations were used, namely 10 M (A) and
− •
M (B) of NiCl2 6H O. To each Ni solution, in a three neck
2
2
2+
1
0
flask at room temperature under mechanical stirring in an argon
flux, a controlled NaBH4 excess was added as reducing agent. The
sudden appearance of a black fine suspension confirmed the reduc-
cooling down of the sample from RT down to 5 K; then, with the
magnetic field already switched on, the magnetization is measured
as a function of temperature.
2
+
tion of the Ni ions to metallic nickel. The solution is maintained
under vigorous stirring for 15 min, then Nickel NPs were collected
using a permanent magnet, washed in deionised water at least four
times and, finally, dried in open air.
2.3. Catalytic tests
In order to compare the catalytic activity of nickel NPs to that of
conventionally prepared catalysts, a heavily loaded Ni/Al O cata-
2
3
All the catalytic experiments were carried out in a fixed-bed
lyst has been prepared by conventional wet impregnation of Siralox
tubular silica glass flow reactor, operating isothermally, loaded
with 44 mg of catalyst mixed with 440 mg of silica glass particles
(60–70 mesh sieved).
For the experiments concerning either ethanol for ethanol/
steam conversion or 500 mg for methanation conversion in the
presence of steam the feed was constituted by a mixture of ethanol
and water with a molar ratio (water to ethanol) equal to 6 in He
2
5
/170 support (alumina with 5% (w/w) SiO from Sasol, 170 m /g)
2
•
using Ni(NO3) 6H O water solution as precursor. After impreg-
2
2
nation, drying at 363 K for 24 h and calcination at 973 K for 5 h
were performed. The composition of this catalyst, 125% Ni/Al O
2
3
(
wtNi/wtsupport), has been chosen in order to allow the formation of
a significant amount of Ni metal upon conditioning.
(
41.6% v/v) added as carrier gas. The gas hourly space velocity
−
1
was GHSV = 51,700 h . Moreover hydrogen yield was defined as
molH2out/(6 molethanol in).
2.2. Materials characterization
Instead, in the case of CO hydrogenation we fed a gaseous mix-
Microscopic analyses on all the samples were performed by the
2
ture of CO and H in the molar ratio 1 to 5 within any dilution in
SEM ZEISS SUPRA 40 VP, with a field emission gun. This instrument
is equipped with a high sensitivity “InLens” secondary electrons
detector and with an EDX microanalysis OXFORD “INCA Energie
2
2
−
1
a carrier gas. In this case the GHSV was equal to 52,300 h
.
In both cases products analysis was performed with a gas-
chromatograph Agilent 4890 equipped with a Varian capillary
column “Molsieve 5A/Porabond Q Tandem” and TCD and FID detec-
tors in series. Between them a nickel catalyst tube was employed
ꢀ
ꢀ
4
50 × 3 . Samples for SEM analysis were suspended in ethanol and
exposed to ultrasonic vibrations to decrease the aggregation. A drop
of the resultant mixture was finally deposed on a Lacey Carbon
copper grid.
to reduce CO and CO to CH . Products analysis in the first case was
2
4
also performed with a GC/MS (ThermoFisher), in order to have a
precise identification of the involved compounds.
X-ray diffraction patterns of both samples were obtained using
a vertical powder diffractometer X’Pert with Cu K␣ radiation
◦
One of the considered parameters was temperature, that was
varied step by step in-between 523 and 773 K waiting for each point
the steady state; moreover, in order to understand something about
catalyst conditioning we have done the experiments first raising
the temperature and after decreasing it.
(
ꢀ = 0.15406 nm). The patterns were collected in the 25–100 2ꢁ
◦
range with a step of 0.03 and a counting time for each step of 12 s.
Powder patterns were indexed by comparing experimental results
to the data reported in the Pearson’s Crystal data database [17].
DC magnetization was performed in a dc-Superconducting
Quantum
Interference
Device
(SQUID)
magnetometer
(
MPMS—Magnetic Properties Measurement System, Quantum
3. Results and discussion
−
6
Design) with resolution better than 10 emu. The room tempera-
ture (RT) magnetic hysteresis cycle was obtained in the 0–5 Tesla
3.1. Sample A Ni NPs
(
T) ꢂ H magnetic field range. The thermal dependence of the
0
magnetisation (5–300 K range) was obtained following the zero
field cooling (ZFC)—field cooling (FC) procedure. In a ZFC process
the sample is cooled to the lowest temperature in absence of a
magnetic field. At that temperature the field is switched on and
the magnetization is recorded as a function of temperature. In a FC
procedure the same magnetic field of the ZFC is applied before the
Fig. 1 reports a SEM image of as prepared Ni NPs obtained
−
4
from the dilute solution (10 M). The apparent diameters of the
observed particles range from 40 to 80 nm, but, as it is possible
to see from the image of a single particle reported in the upper
right corner of the same figure, these are actually aggregates of
even smaller NPs. The X-ray analysis carried out on the as cast

1
2
P. Riani et al. / Journal of Molecular Catalysis A: Chemical 383–384 (2014) 10–16
Fig. 3. X-ray diffractograms of (a) Ni NPs annealed sample A (marked peaks are of
cubic Ni phase) and (b) fresh Ni125%/Al2O3 (marked peaks are of NiO).
Fig. 2. ZFC–FC curves (5–300 K temperature range, ꢂ0H = 0.01 T) of the as prepared
A sample. In the inset the room temperature magnetic hysteresis cycle is shown.
sample showed a diffraction pattern, characteristic of an “amor-
phous” state.
In Fig. 2 the magnetic data relative to these NPs obtained from
SQUID measurements are presented. It displays the ZFC–FC curves
obtained in the 5–300 K temperature range with an applied mag-
netic field ꢂ H of 0.01 T, while in the inset of the same figure the
0
room temperature (RT) magnetic hysteresis cycle is reported.
The ZFC–FC curve displays a very sharp maximum centred
at 16 K, the typical feature of the blocking temperature (TB) of
superparamagnetic NPs, followed by an abrupt decrease of the
magnetisation to very low values above 50 K. The good agreement
between TB and the bifurcation of the ZFC–FC curves (18.5 K) means
that these NPs are in a very narrow dimensional range. A prelimi-
nary estimate of the real dimensions of our Ni NPs can be obtained
from TB since this temperature is related to the size of the nanopar-
ticle by the equation:
Fig. 4. Room temperature magnetic hysteresis of the sample A after annealing at
773 K for 6 h.
Kan.V
The X-ray diffraction pattern of the annealed NPs, presented in Fig. 3
(a line), displays the cubic crystal structure of nickel.
TB =
ꢀ
ꢁ
(1)
ꢃm
ꢃ
kB ln
0
The RT magnetic hysteresis, shown in Fig. 4, confirms the diffrac-
tion data, giving evidence of a typical Langevin behaviour with a
where Kan. is the anisotropy constant, V is the average volume of the
NP, kB is the Boltzmann constant, ꢃm is the measurement time and
ꢃ0 is a constant called attempt time (10 –10 s). A typical value
2
saturation value (∼52 A m /kg) comparable with that of bulk Ni
−9
−10
2
[21] (M
sat.
= 58.6 A m /kg). Under the hypothesis of a distribution
for the logarithm is 25. Since the anisotropy constant for bulk Ni is
strongly temperature dependent and the blocking temperature is
very low (TB = 16 K), in agreement with [18,19] we decided to use
the anisotropy constant value at 4.2 K in order to calculate the aver-
of dimensions and, consequently, a distribution of NP magnetic
moments present in the real nanosystems, we can fit the exper-
imental magnetisation data with the following Langevin equation
(2):
5
3
age volume of our NPs from TB. So, taking Kan. = 12 × 10 erg/cm
b
ꢂ
[
20], we obtain an average radius around 2 nm. The RT hystere-
ꢃ
ꢃ
ꢄ
VNP (r) × 0.62 × ꢂB × H
sis cycle, obtained in the −5 to 5 T magnetic field range, confirms
the very low dimensions of the Ni NPs: the behaviour, completely
anhysteretic without any evidence of coercivity, is completely lin-
ear in the whole magnetic field range investigated and does not
show any trend to reach the saturation. The absence of any curva-
ture in the low field magnetisation curve is a fingerprint of the
lack of ferromagnetic NPs. Numerical simulations, based on the
Langevin model for superparamagnetism fix as an upper limit for
the average radius of these NPs the value of 1 nm, in agreement
with the amorphous structure detected by XRD.
The amorphous nature of these nanostructures suggested the
study of the structural and magnetic properties of these NPs after
an annealing procedure, in order to know if they crystallize and also
to have a rough indication on their sintering behaviour. Thus the
NPs were annealed at 773 K for 6 h. After this procedure, they were
analysed again by X-ray diffraction and magnetic measurements.
M (H, r) =
coth
V
× kB × T
Ni
a
ꢄ
VNi × kB × T
−
× f (r) × dr
(2)
VNP (r) × 0.62 × ꢂB × H
Where r is the radius of the NP, VNP(r) is the volume of the nanopar-
ticle, 0.62ꢂB is the magnetic moment (in Bohr magnetons) per Ni
atom, VNi is the volume of the Ni atom, kB is again the Boltzmann
constant, H is the magnetic field, T the temperature, a and b are the
integral limits (in our fit a = 0.1 nm and b = 30 nm), f(r) is the log nor-
mal distribution of radii, currently adopted in this kind of analysis,
given by (3):
ꢅ
ꢆ
ꢀ
ꢁ
2
r
ln
1
rm
f (r) =
√
× exp
(3)
2
r × ꢄ × 2ꢅ
2 × ꢄ

P. Riani et al. / Journal of Molecular Catalysis A: Chemical 383–384 (2014) 10–16
13
the manifest indication of a broad distribution of NPs dimensions
and, consequently, of different blocking temperatures (TB’s), ran-
ging from 30 to 45 K. On the basis of the above maximum limits
taking the mean Ni anisotropy constant in this temperature span
5
3
(
Kan. = 10 × 10 erg/cm ), we obtain a radii distribution between 2.9
and 3.5 nm.
Furthemore, also the bifurcation at high temperature could be
representative of a set of NPs with larger dimensions and, conse-
quently, higher TB’s. Under this hypothesis, a reasonable upper limit
of the radius of this set of Ni NPs could be 7 nm.
3.3. Characterization of alumina supported catalysts
In Fig. 3 (line b) the diffraction pattern of the 125% Ni/Al O3
2
catalyst is shown. The 125% Ni/Al O3 catalyst pattern shows the
2
characteristic peaks of the NiO phase (rock salt or periclase struc-
ture, cF8 NaCl type), together with those, very weak and not well
detectable from the background, given by the support, that is a
transition alumina with a tetragonal structure such as ␦-Al2O3 or
Fig. 5. FE-SEM image of as prepared Ni nanoparticles obtained from a concentrated
solution (10 M, sample B).
−
2
ꢀ
␥ -Al O [22]. The analysis of the crystal size of the nickel oxide
2 3
phase, carried out using the Scherrer equation gives rise to values
of about 36 nm for the most intense peak (2 0 0) of the NiO phase.
3.4. Catalytic activity in the conversion of ethanol in the presence
of steam
The experiment were performed in a tubular fixed bed reactor
so, in order to prevent the risk of NPs losses we decided to use the B
type sample that presents bigger agglomerates than A type sample.
In Table 1 the catalytic behaviour of the sample B of Ni NPs in the
conversion of ethanol in the presence of steam is summarized. The
catalytic data indicate that this material is very active (conversion
>
90%) already at very low temperature (523 and 573 K) in convert-
ing ethanol in the presence of steam. The obtained C-products are
methane and CO (selectivity in the range 40–45% for both, with a
selectivity towards this reaction up to 85%) together with acetalde-
hyde. Thus, it catalyzes, at least formally, mainly the decomposition
of ethanol:
CH CH OH → CH CO + H
3
2
4
2
Fig. 6. 5–300 K ZFC–FC curves (ꢂ0H = 0.01 T) of the B sample. In the inset the room
temperature magnetic hysteresis cycle is sketched.
with very small conversion of water, small production of CO2,
and consequently, a low hydrogen yield.
This behaviour could result either from a high steam reforming
activity towards ethanol:
Here rm is the average (mean) radius and ꢄ the distribution width.
The fit, reported as continuous line in the same figure, gives as
results rm = 4 nm and ꢄ = 0.5. The annealing process had the effect
to increase the NPs radius (from ∼1 nm to a mean value of 4 nm)
changing thus the microstructure from amorphous to crystalline.
CH CH OH + H O → 2CO + 4H
3
2
2
2
coupled with a very efficient methanation:
CO + 3H2 → CH + H O
4
2
3.2. Sample B Ni NPs
or from a direct ethanol decomposition step.
In Fig. 5, a FE-SEM image of a sample of Ni NPs obtained from
At 523 and 573 K in the increasing and in the decreasing tem-
perature experiments, a CH4 to CO product ratio very near to 1
is observed: this trend seems to be more consistent with a direct
ethanol decomposition step.
At 673 K methane selectivity is still high but lower, selectiv-
ity to CO falls in favour of that to CO2 (56%), water conversion
a more concentrated synthesis, is presented. Very large NPs aggre-
gates are formed; the apparent diameters of the observed particles
range from 10 to 30 nm, but, also in this case, these are actually
aggregates of even smaller NPs. The X-ray diffraction pattern of the
Ni NPs prepared with a more concentrated Ni salt solution (1B) dis-
plays the nickel cubic crystal structure, isostructural to Cu, as the
previous sample A annealed at 773 K for 6 h. This diffraction pattern
has been omitted in Fig. 3 for simplicity.
Fig. 6 reports the RT ZFC–FC magnetisation curves with an
applied magnetic field of 0.01 T obtained from the same sam-
ple. In the inset, the RT hysteresis cycle is also displayed. The
ZFC curve presents a maximum, corresponding to the blocking
temperature, around 40 K. The two curves diverge at higher tem-
perature (Tbif. ∼ 160 K). The relatively broad maximum observed is
also becoming more significant. The CH /CO + CO2 ratio is well
4
below 1. This means that either steam reforming or water gas shift
activities appeared. A further increase of the temperature up to
773 K and the successive reduction to 673 K does not change very
much the behaviour observed, with predominant steam reforming
activity: this trend leads to hydrogen yields in the 50–60% range,
mainly limited by the still significant methane selectivity (around
30%). Thermodynamic calculations show that the composition of
the product gas at 773 K is near that forecast if steam reforming

1
4
P. Riani et al. / Journal of Molecular Catalysis A: Chemical 383–384 (2014) 10–16
Table 1
Ethanol and water conversions, C-product selectivities and hydrogen yields on Nickel NPs.
T Furnace
K)
Ni NP
(
X, C2H6O
X, H2O
S, CH4
S, CO
S, CO2
S, CH2CH2
S, CH3CHO
S, CH3COCH3
Y, H2
523
573
673
773
673
573
523
0.92
0.97
1.00
0.94
1.00
0.17
0.19
0.01
0.04
0.22
0.28
0.26
0.02
0.02
0.40
0.44
0.41
0.26
0.33
0.42
0.40
0.42
0.41
0.03
0.07
0.04
0.40
0.33
0.01
0.10
0.56
0.67
0.63
0.18
0.26
0.01
0.01
0.00
0.00
0.01
0.00
0.00
0.13
0.03
0.00
0.00
0.00
0.00
0.00
0.03
0.00
0.00
0.00
0.00
0.00
0.00
0.17
0.22
0.44
0.59
0.54
0.05
0.07
Table 2
Ethanol and water conversions, C-product selectivities and hydrogen yields on 125% Ni/Al2O3.
T Furnace
K)
Ni 125
(
X, C2H6O
X, H2O
S, CH4
S, CO
S, CO2
S, CH2CH2
S, CH3CHO
S, C4H10O
S, C4H8O2
Y, H2
523
573
673
773
673
573
523
0.01
0.05
0.46
1.00
1.00
0.31
0.10
0.00
0.00
−0.04
0.34
0.19
0.00
0.00
0.00
0.00
0.00
0.21
0.46
0.08
0.04
0.00
0.00
0.00
0.08
0.02
0.09
0.05
0.00
0.02
0.00
0.72
0.52
0.04
0.02
0.00
0.06
0.36
0.00
0.00
0.02
0.00
1.00
0.47
0.27
0.00
0.00
0.64
0.83
0.00
0.40
0.36
0.00
0.00
0.11
0.06
0.00
0.04
0.00
0.00
0.00
0.01
0.00
0.00
0.01
0.02
0.70
0.38
0.06
0.02
Table 3
CO2 conversion and C-product selectivities on 125% Ni/Al2O3 and on Ni NPs.
of ethanol, methanation/steam methane reforming and water gas
shift approach equilibrium.
Interestingly, in the decreasing temperature experiments at 573
and 523 K Ni NPs lost their high activity, thus becoming poorly
active. This behaviour is remarkably different from that of the 125%
Ni/Al O catalyst, summarized in Table 2. This catalyst has low
T Furnace
K)
Ni 125%
X, CO2
Ni NPs
X, CO2
(
S, CO
S, CH4
S, CO
S, CH4
523
573
673
0.00
0.07
0.42
0.71
0.40
0.04
0.01
0.00
0.18
0.19
0.14
0.24
0.25
0.00
0.00
0.82
0.81
0.86
0.76
0.75
1.00
0.00
0.00
0.03
0.06
0.004
0.00
0.00
0.00
0.00
0.50
0.50
0.51
0.00
0.00
0.00
0.00
0.50
0.50
0.49
0.00
0.00
2
3
activity in the conversion of ethanol in the same conditions at 523 K
and 573 K, producing acetaldehyde as the major product. At 673 K
it is active in converting ethanol also to ethylene and diethylether,
the products of ethanol dehydration, together with acetaldehyde.
The 125% Ni/Al O catalyst only at 773 K allows the full conver-
773
673
5
5
73
23
2
3
sion of ethanol with steam mainly to CO and hydrogen, i.e. gives
rise to steam reforming even if selectivity to methane is still quite
high, >20%. However, in the decreasing temperature experiment it
still is very active at 673 K producing CH , CO, CO and H , with-
CO + H2 → CO + H O
2
2
4
2
2
out any ethylene nor acetaldehyde. This shows a “conditioning”
effect, after which steam reforming activity remains also at 673 K.
We can remark that this catalyst is loaded in an oxide form, without
any reduction pretreatment. Thus, such a “conditioning” effect may
be associated to the reduction to Ni metal. Interestingly the ratio
while Ni NPs are still inactive. At 673 K CO2 conversion increases
to 42% on Ni-alumina while is very low on Ni NPs sample, with
a different product distribution; on Ni/Al O3 the main product
2
remains CH with 84% selectivity, while on Ni NPs the ratio between
4
selectivity to CO and CH4 is almost 1. At 773 K CO2 conversion on
CH /CO + CO observed at 673 K in the lowering reaction temper-
Ni/Al O reaches the value of 71% with 86% selectivity to methane
4
2
2 3
ature experiment is only slightly lower than 1. This product ratio
suggests that ethanol decomposition may also occur as the main
reaction, possibly followed by water gas shift reaction. However,
even after conditioning, the catalyst loses activity at lower tem-
perature give rise again mainly to dehydration to acetaldehyde at
and 14% to CO. The same behaviour is observed at each tempera-
ture in the increasing and the decreasing temperature experiments.
Instead, Ni NPs remain almost inactive (6% conversion) also at 773 K
again producing 50% CO and 50% CH4.
5
73 and 523 K. Thus in the decreasing reaction temperature exper-
3
.6. Characterization of spent catalysts
iments at 573 and 523 K, when it is supposed to be “conditioned”,
the catalyst recover a behaviour similar to that of the fresh catalyst
The data concerning the characterization of the spent catalysts
(
supposed to be still oxidized).
after ethanol conversion in the presence of steam are reported in
Figs. 7–10. XRD patterns of both spent catalytic materials (Fig. 7 bot-
tom, lines a and b) present the peaks of the metastable hexagonal
Nickel polymorph (hP2-Mg type), usually assumed to be stabilized
by carbon, while the XRD pattern of the spent 125% Ni/Al2O3 cat-
alyst (indicated as a in the Fig. 7) shows additionally the peaks of
the alumina support and residual peaks of the NiO precursor.
The FE-SEM micrograph of the spent catalytic Ni NPs shows
(see Fig. 8) aggregation phenomena of the single Ni NPs, par-
tially coalesced to each other. Moreover, it is possible to observe
the appearance of carbon nanotubes. As it is well known, these
3.5. Catalytic activity in CO2 methanation
In Table 3 the data on CO2 hydrogenation experiments are
reported. At the lowest reaction temperature (523 K) no conversion
of CO is observed on both catalysts. At 573 K we have a little con-
version of CO on Ni-alumina catalyst producing mainly methane,
and CO as a byproduct
2
2
CO + 4H → CH + 2H O
2
2
4
2

P. Riani et al. / Journal of Molecular Catalysis A: Chemical 383–384 (2014) 10–16
15
Fig. 7. X-ray diffractograms of spent catalytic materials: (a) Ni125%/Al2O3 after
ethanol + steam conversion cycle (marked peaks * are residual NiO precursor;
h = hexagonal Ni), (b) Ni NPs after ethanol + steam conversion cycle (marked peaks
h = hexagonal Ni), (c) Ni125%/Al2O3 after CO2 methanation (marked peaks c = cubic
Ni) and (d) Ni NPs after CO2 methanation (marked peaks c = cubic Ni).
Fig. 9. FE-SEM image of NiNPs after CO2 methanation.
Fig. 10. FE-SEM image of Ni125/alumina after CO2 methanation.
Fig. 8. FE-SEM image of spent NiNPs after ethanol conversion in the presence of
steam. It is possible to observe the appearance of carbon nanotubes.
still with high methane coproduction. By decreasing back reaction
temperature, they lose their high low-temperature activity. They
actually change their crystal structure from cubic to hexagonal
nickel. It can be supposed that cubic Ni is more active in ethanol
decomposition while hexagonal Ni is more active in steam reform-
ing (at 673 K). A different behaviour is found for Ni/Al O : when
fresh (and unreduced) they are moderately active in ethanol dehy-
drogenation and dehydration at low temperature, likely with the
contribution of the uncovered support surface, as shown in previ-
ous experiments [22], becoming active in steam reforming at 773,
carbonaceous materials contribute to the increase of pressure
losses upon time on stream and may also cause a decrease the cat-
alytic activity. The formation of carbon nanotubes is also evident
on the FE-SEM pictures of 125% Ni/Al O (not reported here).
2
3
2
3
XRD patterns of both spent catalytic materials after methanation
experiments are reported in Fig. 7. Both the spent catalytic materi-
als Ni125%/alumina (indicated as c in the figure) and spent NiNPs
(
indicated as d) show the peaks of the cubic nickel crystal structure,
and 673 K when “conditioned”. However, the high CH /CO + CO2
4
isostructural to Cu. The FE-SEM image of the spent NiNPs, reported
in Fig. 9) shows an important agglomeration of the particles and
also the absence of carbon nanotubes after reaction.
In Fig. 10 the FE-SEM micrograph of spent Ni-alumina catalyst
is reported: also in this case the absence of carbon nanotubes is
evident.
ratio obtained at 673 K may indicate that they may also produce
ethanol decomposition. This catalyst recovers the initial low activ-
ity for dehydrogenation (but not in dehydration) in the lowering
reaction temperature experiments, at the end of our cycle. Also this
catalyst contains hexagonal Ni when spent, but also some NiO.
The ability of Ni catalysts to catalyse ethanol decomposition has
already been reported in the literature. As for example, Homs et al.
[23] cited ethanol decomposition propensity of Ni/ZnO catalysts
while, several years before, Gates et al. [24] provided evidence of
ethanol decomposition on Ni 111 monocrystal surface. The data
we report here suggest that ethanol decomposition likely goes
through a first dehydrogenation step to formaldehyde, thus being
actually due to decomposition of acetaldehyde. In fact, on both Ni
NPs and on “conditioned” Ni/Al O (in the decreasing temperature
4
. Discussion
The experiments reported here provided some data which can
give some contribution to a better understanding the mechanism
of ethanol steam reforming on Ni catalysts.
Unsupported Ni NPs are very active in catalysing low temper-
ature ethanol decomposition, even in the presence of steam at
very low temperature (523, 573 K). They convert into good steam
reforming catalysts at high temperature (673, 773 K) although
2
3
experiment), we find production of acetaldehyde at lower tempera-
ture, and high selectivity to methane and CO at higher temperature.

1
6
P. Riani et al. / Journal of Molecular Catalysis A: Chemical 383–384 (2014) 10–16
We previously found on different Ni-MgO-Al O3 and Ni-ZnO-
studies provided evidence of the role of Nickel metal particles in
2
Al O catalysts methane selectivity higher than that forecasted
decomposing ethanol through a previous acetaldehyde production
step as a first step in ethanol steam reforming. This behaviour dif-
ferentiates Ni catalysts from Co catalysts. Additionally, a role of the
support in the ESR over supported catalysts is envisaged. A support
effect is also evident in the methanation of CO2 on Ni/Al O , sup-
2
3
by thermodynamics by ESR on Ni catalysts [25], suggesting that
methane indeed comes from some decomposition step more that
from methanation. However, we attributed methane formation to
surface acetate decomposition: this was based on experiments of
acetic acid steam reforming, that produced high methane selectiv-
ity on Ni catalysts [26] and on IR spectroscopic experiments [25,26],
as well on the evidence that the higher the water amount in the
feed, the lower the acetaldehyde selectivity. We supposed that at
higher water concentrations the catalyst is more oxidized and oxi-
dizing, thus favouring acetaldehyde conversion to acetic acid or
surface acetate species, that can more easily decarboxylate [27].
The data might be rationalized supposing that two mechanism
can concur in producing methane: the ethanol decomposition via
acetaldehyde on the metal particles, as reported by several authors
2
3
posed to justify the high activity of this catalyst and the very low
activity of Ni NP in this reaction. This CO methanation mechanism
2
occurring in part on the support may also justify the permanence
of Ni in the cubic crystal structure in the case of CO methanation
2
on 125% Ni/Al O , in contrast to what happens upon ethanol steam
2
3
reforming where both unsupported and supported spent catalysts
contain hexagonal nickel.
References
[
[
1] A. Haryanto, S. Fernando, N. Murali, S. Adhikari, Energ. Fuel 19 (2005)
098–2106.
2] P.D. Vaidya, A.E. Rodrigues, Chem. Eng. J. 117 (2006) 39–49.
[
2,28], and the acetate decomposition on the support or over an
2
incompletely reduced phase.
These mechanisms occur to a much lower extent on Cobalt
catalysts that show very low methane selectivities also at low tem-
perature, which is supposed to be due to the higher activity of cobalt
to oxidize acetate surface intermediates [15]. At 773 K the activity
toward steam reforming reaction giving CO2 as main product is
higher on Cobalt NPs than Ni ones. A characterization of the spent
materials underlined that all of them produce carbonaceous species
that could cause deactivation in more extended experiments.
It seems quite interesting too that 125% Ni/Al O is also a very
[3] M. Ni, D.Y.C. Leung, M.K.H. Leung, Int. J. Hydrogen Energy 32 (2007) 3238–3247.
[
[
[
4] F. Frusteri, S. Freni, J. Power Sources 173 (2007) 200–209.
5] P. Ramirez de la Piscina, N. Homs, Chem. Soc. Rev. 37 (2008) 2459–2467.
6] A. Bshish, Z. Yaako, N. Narayanan, R. Ramakrishnan, A. Ebshish, Chem. Pap. 65
(2011) 251–266.
[7] G. Nahar, V. Dupont, Biofuels 3 (2012) 167–191.
[
[
8] N. Bion, D. Duprez, F. Epron, ChemSusChem 5 (2012) 76–84.
9] L.V. Mattos, G. Jacobs, B.H. Davis, F.B. Noronha, Chem. Rev. 112 (2012)
4
094–4123.
[10] P. Ciambelli, V. Palma, A. Ruggiero, Appl. Catal. B: Environ. 96 (2010) 18–27.
11] P. Panagiotopoulou, X.E. Verykios, Int. J. Hydrogen Energy 37 (2012)
6333–16345.
[
2
3
1
active catalyst for CO2 methanation, while Ni NPs are not, both
actually containing cubic Ni. A possible explanation is that the sup-
port (alumina or ionic Ni dispersed on alumina) may actually act in
this reaction, possibly participating to the carbon dioxide adsorp-
tion step. The lack of formation carbon nanotubes as well as of
C-stabilized hexagonal nickel during these experiments may sup-
[12] V.A. de la Pe n˜ a O’Shea, R. Nafria, P. Ramírez de la Piscina, N. Homs, Int. J.
Hydrogen Energy 33 (2008) 3601–3606.
[
[
[
[
[
[
13] B. Bayram, I. Ilgaz Soykal, D. von Deak, J.T. Miller, U.S. Ozkan, J. Catal. 284 (2011)
7–89.
14] G. Busca, U. Costantino, T. Montanari, G. Ramis, C. Resini, M. Sisan, Int. J. Hydro-
gen Energy 35 (2010) 5356–5366.
15] G. Garbarino, P. Riani, M.A. Lucchini, F. Canepa, S. Kawale, G. Busca, Int. J. Hydro-
gen Energy 38 (2013) 82–91.
16] R. Espinal, E. Taboada, F. Molins, R.J. Chimentao, F. Medina, J. Llorca, Appl. Catal.
B: Environ. 127 (2012) 59–67.
17] Pearson Crystal Data: Crystal structure database for inorganic compounds,
Release 2009/2012, ASM International, The Material Information Society.
18] P.C. Fonseca, G.F. Goya, R.F. Jardim, R. Muccillo, N.L.V. Carreno, E. Longo, E.R.
Leite, Phys. Rev. B 66 (2002) 104406.
7
port the idea that the CO methanation reaction may in large part
2
occur on the support, the metal being mostly associated to the acti-
vation of hydrogen. A similar hypothesis was proposed previously
for conversion of CO into CO and methanol on Cu/ZnO-Al O cata-
2
2
3
lysts [29]. Additionally, it has previously been found on Ni/Al O the
2
3
possibility of a CO methanation path occurring through oxygenate
intermediates such as formate and methoxy groups, that convert
into gas-phase methane through hydrogenation [30]. A mecha-
nism through oxygenates (carbonates, formates, and methoxydes)
adsorbed on the support with activation of hydrogen on the metal
would avoid a contact between carbon atoms and the metal, thus
avoiding growth of nanotubes and conversion into the hexagonal
form, notoriously stabilized by carbon. Maybe, in actual condi-
tions (higher pressures) this mechanism may occur together with
a carbide-like mechanism occurring on the metal.
[19] Y. Hou, S. Gao, J. Mater. Chem. 13 (2003) 1510–1512.
[
[
[
[
20] R. O’Handley, Modern Magnetic Materials: Principles and Applications, John
Wiley & Sons, New York, 2000.
21] E.P. Wolhfarth (Ed.), Ferromagnetic Materials, 1, North Holland Publishing
Company, Amsterdam, 1980, pp. 1–70.
22] G. Garbarino, A. Lagazzo, P. Riani, G. Busca, Appl. Catal. B: Environ. 129 (2013)
460–472.
23] N. Homs, J. Llorca, P. Ramírez de la Piscina, Catal. Today 116 (2006) 361–366.
[24] S.M. Gates, J.N. Russell Jr., J.T. Yates Jr., Surf. Sci. 171 (1986) 111–134.
[
25] C. Resini, T. Montanari, L. Barattini, G. Ramis, G. Busca, S. Presto, P. Riani, R.
Marazza, M. Sisani, F. Marmottini, U. Costantino, Appl. Catal. A: Gen. 355 (2009)
8
3–93.
[26] L. Barattini, G. Ramis, C. Resini, G. Busca, M. Sisani, U. Costantino, Chem. Eng. J.
53 (2009) 43–49.
1
[
27] G. Busca, U. Costantino, T. Montanari, G. Ramis, C. Resini, M. Sisani, Intern. J.
Hydrogen Energy 35 (2010) 5356–5366.
5
. Conclusions
[
[
28] V. Mas, G. Baronetti, N. Amadeo, M. Laborde, Chem. Eng. J. 138 (2008) 602–607.
29] M.A. Larrubia Vargas, G. Busca, U. Costantino, F. Marmottini, T. Montanari, P.
Patrono, F. Pinzari, G. Ramis, J. Mol. Catal. A: Chem. 266 (2007) 188–197.
The catalytic behaviour of Ni NPs in the conversion of ethanol
in the presence of steam and in the methanation of CO2 has been
studied and compared with that of heavily charged Ni/Al O . These
[30] V. Sanchez-Escribano, M.A. Larrubia Vargas, E. Finocchio, G. Busca, Appl. Catal.
A: Gen. 316 (2007) 68–74.
2
3