Ni(II)-Mg(II)-Al(III) catalysts for hydrogen production from ethanol steam reforming: Influence of the Mg content

Article abstract of DOI:10.1016/j.apcata.2013.10.054

The Ni(II)Mg(II)Al(III) LDH were obtained by homogeneous precipitation urea method, after a proper thermal treatment resulted in an active catalyst for the ethanol steam reforming. After catalytic performance, catalysts were analyzed by temperature programmed oxidation (TPO), in order to quantify carbon deposits. Reduced Ni area was evaluated by H₂ static volumetric chemisorption measurements in a Micromeritics AutoChem II 2920 equipment. Thermogravimetric studies were carried out in a Shimadzu TGA-51H equipment, using a heating ramp of 10 K/min in air flow of 50 cm³/min. Experimental equipment used for catalytic evaluation consists of a quartz tubular reactor heated in electric oven at the reaction temperature, which is monitored by a thermocouple placed inside the reactor. The vaporized current is diluted with argon at the entrance of the reactor. It can be observed PXRD spectra of precursors with different Mg content, all samples show characteristic signals of LDH.

Full text of DOI:10.1016/j.apcata.2013.10.054

Applied Catalysis A: General 470 (2014) 398–404
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Review
Ni(II)–Mg(II)–Al(III) catalysts for hydrogen production from ethanol
steam reforming: Influence of the Mg content
Adriana Romeroa, Matías Jobbágyb, Miguel Labordea,
Graciela Baronettia, Norma Amadeoa,∗
a Laboratorio de Procesos Catalíticos, Departamento de Ingeniería Química, Facultad de Ingeniería Ciudad Universitaria, 1428 Buenos Aires, Argentina
b INQUIMAE, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Pabellón II, Ciudad Universitaria, 1428 Buenos Aires, Argentina
a r t i c l e i n f o
Article history:
Received 24 May 2013
Received in revised form 18 October 2013
Accepted 29 October 2013
Available online 14 November 2013
Keywords:
Bioethanol
Hydrogen production
LDH
Ni catalyst
Reforming
Contents
1.
2.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 398
Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 399
2.1.
2.2.
Synthesis and characterization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 399
Catalytic performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 399
3.
4.
Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 400
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 403
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 403
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 403
1. Introduction
product. However, hydrogen non-contaminant character depends
on the nature of the raw material used as source. In this sense,
steam reforming of ethanol, obtained from biomass, offers a true
green alternative for H2 production, due to its inherent renewable
character, low toxicity (unlike methanol) and the fact that it can be
obtained virtually sulphide-free [2,3]. Ethanol has relatively high
hydrogen content and in presence of water, is capable of producing
6 mol of H2 per mol of ethanol:
During the last centuries, human civilization has solved the
energy supply problem mainly on the bases of fossil sources of car-
bon and hydrocarbons and the associated technologies. In the last
three decades there have been growing concerns about the short-
age in energy supply in the world [1], which triggered the searching
for alternative sources and technologies. Most of the chemical
based alternatives consider hydrogen as an efficient energetic vec-
tor since it directly transforms chemical energy into electric current
by means of fuel cells technology, releasing steam as the only
C2H5OH + 3H2O → 2CO2 + 6H2
(I)
From a thermodynamic viewpoint, reforming reaction is highly
favorable, however it competes with multiple side reactions [2,4]
and in general, the outflow from the reactor contents a wide range
of liquids and gaseous products. Different catalytic systems have
been proved to be effective with variable hydrogen selectivities
∗
Corresponding author. Tel.: +54 1145763211; fax: +54 11 4576 3241.
E-mail addresses: norma@di.fcen.uba.ar,
normaamadeo@yahoo.com.ar (N. Amadeo).
0926-860X/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.apcata.2013.10.054

A. Romero et al. / Applied Catalysis A: General 470 (2014) 398–404
399
Table 1
[1–11]. Ni catalysts have been widely utilized in reforming of
Chemical composition and specific surface of precursors.
hydrocarbons because it promotes the rupture of the C–C bonds,
achieving high activity [11–14]. Noble-metal-catalysts (Rh, Pt, and
Pd) also present high activity and selectivity [9], however their high
cost limits their high-scale-use, in contrast to nickel based ones.
Steam reforming reaction takes place on the surface of a solid
catalyst. For this process usually nickel supported catalysts are
used. Water adsorbs preferentially on the catalysts support, formu-
lated to allow mobility for oxygen species, originated from water
dissociation, adsorbed on its surface, while hydrocarbon molecules
adsorb preferentially on the metal surface (metallic nickel), and the
steam reforming reaction takes place at the metal–support inter-
face [15–19].
In order to prevent undesired carbon formation, which is the
principal cause of poisoning in Ni based catalysts used in this pro-
cess, ternary Ni(II)–Mg(II)–Al(III) oxides were proposed, where the
presence of Mg(II) disfavor this undesired product [20].
Catalyst characteristics are determined by their physical-
chemistry, structural and textural properties, as they are: active
area, metal particle size, metal dispersion and reducibility. These
properties depend on metal–support interaction, and they could
be established on different stages of catalyst synthesis and thermal
treatment. For example, varying precursor material composi-
tion, preparation method and/or activation (calcination, reduction)
[1,21,22].
Sample
%wt Ni
LDH formula
Sg (m2/gcat
)
HT0,00
HT0,04
HT0,16
HT0,33
HT0,72
HT1,20
HT1,57
57.67
60.08
50.32
45.41
34.97
30.18
23.12
Ni0.69Al0.31(OH)2(CO3)0.15·nH2O
96.8
106.6
145.4
173.2
201.0
216.3
209.3
Ni0.72Mg0.03Al0.26(OH)2(CO3)0.13·nH2O
Ni0.60Mg0.10Al0.30(OH)2(CO3)0.15·nH2O
Ni0.51Mg0.17Al0.32(OH)2(CO3)0.16·nH2O
Ni0.28Mg0.39Al0.33(OH)2(CO3)0.17·nH2O
Ni0.30Mg0.37Al0.33(OH)2(CO3)0.17·nH2O
Ni0.26Mg0.41Al0.33(OH)2(CO3)0.17·nH2O
hydrogen flow of 100 ml/min. An “R” letter was added to the
nomenclature for reduced samples.
Samples were characterized by ICP chemical analysis, thermo-
gravimetric analysis (TGA), sorptometry (SBET), H2 chemisorption,
powder X-ray diffraction (PXRD), X-ray photoelectronic spec-
troscopy (XPS) and temperature programmed reduction (TPR).
After catalytic performance, catalysts were analyzed by temper-
ature programmed oxidation (TPO), in order to quantify carbon
deposits.
BET surface area was obtained in a Micromeritics ASAP 2020
equipment. Reduced Ni area was evaluated by H2 static volumetric
chemisorption measurements in a Micromeritics AutoChem II 2920
equipment. Ni metallic area was calculated assuming a Ni/H = 1
stoichiometry and that a Ni atom occupies 6.45 A . Solid compo-
2
˚
sition was analyzed in a Sequential Plasma Spectrometer ICP-AES
Shimadzu 1000 III. Thermogravimetric studies were carried out in
a Shimadzu TGA-51H equipment, using a heating ramp of 10 K/min
in air flow of 50 cm3/min. Solids were characterized by PXRD in a
Siemens D 5000 equipment (radiation Cu K␣).
Thermal decomposition of crystalline monophasic precursors
as Me(II)–Al(III)-based layered double hydroxides (LDHs) leads to
complex oxides, ideally composed of highly dispersed MeO and
MeAl2O4 mixtures [23–25]. However, the aforementioned stoi-
chiometric phases are only achieved after massive demixing at
high temperatures, an ill crystallized non-stoichiometric phases are
expectable at intermediate annealing [26]. Moreover, in a previous
work, we concluded that thermal treatments of Ni(II)–Al(III) LDHs
lead to synergetic effects between the elements in mixed oxide
structures, and after appropriate activation treatment, give rise to
well dispersed metal particles like a supported metal catalysts, with
the possibility of controlling metal–support interaction during the
synthesis and treatment stages [22].
XPS analysis was carried out in a multi-technical system (SPECS)
which has a hemispheric analyzer PHOIBOS 150 operating in fix
analyzer transmission mode (FAT). Spectra were acquired with a
step energy of 30 eV, using Mg K␣ radiation X-ray source operated
at 200 W and 12 kV. Work pressure in analysis chamber was lower
than 5 × 10−9 mbar. Spectra quantification was made by Casa XPS
Software [28] using the analyzer’s transmission function appropri-
ated values of the Scofield factors. Curves were deconvolucionated
by Gaussian and Lorentzian-type functions and Shirley-type back-
ground no-lineal subtraction. Binding energy values were corrected
using the aluminum oxide Al 2p line at 74.3 eV as reference [29].
TPR experiments were performed in a Micromeritics AutoChem
II 2920 equipment, using an amount of sample of 100 mg, N2/H2
flow of 100 ml/min (molar composition 98/2%) and a heating ramp
of 10 K/min in the range of 293–1173 K. Previously, the precursor
was calcined in oven at 623 K during 1 h in order to avoid anion
reduction to occur simultaneously with the Ni(II) reduction. Pre-
vious runs with different amounts of pure CuO were carried out
in order to quantify the detector signal. TPO experiments were
carried out in the same equipment than TPR ones, with the fol-
lowing conditions: an amount of 20 mg of sample was pretreated
in an Argon flow of 50 ml/min from 298 to 573 K, with a ramp of
20 K/min, in order to eliminate any adsorbate which may interfere
with the results. After that, the sample was cooled to 323 K. TPO
study was carried out from 323 to 1073 K, with a ramp of 10 K/min,
in an Argon/Air flow of 50 ml/min (90/10% molar). Both TPR and TPO
studies were carried out with Mass Spectrometry on line, in order
to detect H2, H2O, CO and CO2 species. It was utilized a ThermoStar
TM Pfeiffer Vacuum equipment.
The aim of this work is to study in which extent the partial
replacement of Ni(II) by Mg(II) in a family of LDHs affect the nature
of their mixed oxides in terms of their reducibility, activity, product
distribution and carbon deposition during ethanol steam reforming
process.
2. Experimental
2.1. Synthesis and characterization
Mother solutions of Al(III), Ni(II) and Mg(II) nitrates (0.5 mol/l,
each) were prepared dissolving Al(NO3)3·9H2O, Ni(NO3)2·6H2O
and Mg(NO3)2·6H2O in distilled water, respectively. Catalyst pre-
cursors were prepared by homogeneous precipitation method
based on urea hydrolysis [22], for which solutions containing
urea–Ni(II)–Mg(II)–Al(III) in proper ratios for each catalyst; were
aged at 363 K during 24 h in PP bottles. The reaction was quenched
submitting the bottles into ice-bath. The precipitated precursors
were centrifuged, washed with cold distilled water and dried at
343 K. Total concentrations for the starting solutions of urea and
cations [Ni(II) + Mg(II) + Al(III)] were 0.5 mol/l and 5.0 × 10−2 mol/l,
respectively. Initial molar ratio of [Ni(II) + Mg(II)]/Al(III) were 2 [27].
The catalysts were named “HTx”, being x = molar ratio Mg(II)/Ni(II),
as indicated in Table 1.
2.2. Catalytic performance
Experimental equipment used for catalytic evaluation consists
of a quartz tubular reactor (Ø = 9.2 mm) heated in electric oven at
the reaction temperature, which is monitored by a thermocouple
placed inside the reactor. The ethanol and water feed is pumped
After that, the precursors were reduced with a heating ramp of
10 K/min up to 993 K, holding such temperature for 2 h, in pure

400
A. Romero et al. / Applied Catalysis A: General 470 (2014) 398–404
Fig. 1. PXRD spectra of precursors.
Fig. 2. Thermal decomposition.
with a syringe pump and evaporated before the entrance of the
reactor, using nitrogen as carrier. The vaporized current is diluted
with argon at the entrance of the reactor. Inlet and outlet compo-
sition analysis were carried out by GC in an Agilent Technologies
6890N equipment.
Ethanol conversion and product yields are defined as follows:
yis − yio
yoEtOH − ysEtOH
yoEtOH
X =
Yi =
yoEtOH
where yo: molar fraction in the inlet; ys: molar fraction in the outlet.
Two series of experiences were carried out. In all experiences,
the operation conditions were chosen to prevent mass transfer lim-
itations and guarantee isothermal conditions.
First, in order to study the effect of the Mg content over the
activity and yields of the catalysts, the operative conditions were:
catalyst mass: 8.0 mg; reaction temperature: 823 K, water/ethanol
molar ratio: 5.5, inert flow: 350 ml/min. The catalyst was diluted
1:10 with an inert solid. These conditions assured ethanol conver-
sion to be lower than 100%.
Fig. 3. PXRD spectra of reduced samples. (ꢀ) NiO; (ꢁ) Ni0; (♦) spinel-like phase.
Then, we have studied the effect of the Mg content over the
carbon deposit on the catalyst, employing the following operative
conditions: catalyst mass: 30.0 mg; reaction temperature: 923 K,
water/ethanol molar ratio: 5.5, inert flow: 400 ml/min. There was
no dilution of the catalyst. After that, the samples were analyzed
by TPO.
reported in a previous work [22], the evolution of thermal decom-
position presents the same pattern, regardless of the study being
carried out in oxidizing or reducing atmosphere [34].
In Fig. 3 PXRD patterns for all reduced samples are shown. In
order to study the evolution of the structure during thermal treat-
ment, the pattern of the precursor HT0.33 calcined at 973 K (700 ◦C)
was included (first pattern from below, HT0.33-C700).
3. Results and discussion
After thermal decomposition, the initial structure collapses
because of the loss of carbonates (decarboxylation) and hydroxyl
groups (dehydration), transforming the solid into an oxide mixture.
The sample HT0.33-C700 pattern shows characteristic reflec-
tions of NiO (Bunsenite, JCPDS 22-1189); because MgO (JCPDS
4-0829) has a similar NaCl-type structure, observed reflections at
high angles could be consistent with Ni1 – qMgqO solid solution.
Since characteristic reflections of NiO, MgO and Ni1 − qMgqO solid
solution are very close, they are impossible to distinguish at low
values of 2ꢀ [35–39]. It has to be noticed that the inclusion of Al(III)
in the lattice also affects these reflections, hindering the distinction
of separated phases.
After reductive activation, all the precursors evolved into a
metallic Ni phase (JCPDS 4-0850). Additional low intensity reflec-
tions at 2ꢀ = 36.3; 59.5 and 65.3◦ were detected, suggesting
incipient formation of a spinel-like phase (JCPDS 10-0339 NiAl2O4,
21-1152 MgAl2O4); which is expected, given the abundant pres-
ence of Al(III) in the solid. Previous reports based on neutron
diffraction studies indicate that such mixed phases crystallize in an
incipient way at 923 K [40], while massive segregation (demixing)
of stable phases takes place at 1273 K [41].
Table 1 shows the chemical composition and specific surface for
the samples. It can be seen that specific surface increases when Mg
content does [30].
In Fig. 1 it can be observed PXRD spectra of precursors with dif-
ferent Mg content, all samples show characteristic signals of LDH.
Peak sharpness (directly related to crystallinity) increases when
Mg(II) content does [31].The main interbasal reflections (0 0 3),
(0 0 6), in addition to the characteristic (1 1 0) and (1 1 3) reflections
close to 2ꢀ = 60◦; indicate the presence of crystalline carbonate-
intercalated hydrotalcite-like structure [32]. There is no evidence
of Ni(OH)2, MgCO3 nor AlO(OH) phase segregation [27,33].
Thermal decomposition occurs in two endothermic steps, typi-
cal for LDHs [23]. The first step, the transition is a constant weight
loss in the range of 300–500 K, which corresponds to reversible loss
of interlaminar water, without structure collapse. The second step,
between 500 and 700 K, corresponds to irreversible dehydroxyla-
tion of brucitic layers and loss of carbonate anions.
TGA results for the sample without Mg (HT0.00) were reported
in a previous work [13]. Fig. 2 shows the obtained results in thermal
decomposition. Total weight loss in % was between 30 and 37% for
all samples, confirming the LDH nature of precursors, and these
results were taking into account for further calculations. As we
TPR profiles for precursors with different ratios Mg/Ni are com-
pared in Fig. 4. It can be seen two broad peaks. According to Mass
Spectrometer equipment monitoring on line, the first peak, at low

A. Romero et al. / Applied Catalysis A: General 470 (2014) 398–404
401
Fig. 4. TPR profiles for the samples with different ratios Mg/Ni.
Table 2
Ni reducibility for all precursors.
Sample
% Red Ni
T˛ (K)
Tˇ (K)
˛/(˛ + ˇ) area ratio
HT0.00
HT0.04
HT0.16
HT0.33
HT0.72
HT1.20
HT1.57
94
97
95
93
94
86
90
838
818
848
840
858
880
900
968
978
0.49
0.50
0.37
0.26
0.24
0.19
0.18
1003
1013
1023
1058
1063
Fig. 5. 2nd TPR event fitting.
Table 3
Chemisorption data and intrinsic conversion.
Sample
ANi (m2/gcat
)
ANi (m2/gNi
)
dp (nm)
YH2 /m2 metallic Ni
HT0.00
HT0.04
HT0.16
HT0.33
HT0.72
HT1.20
HT1.57
17.67
18.91
18.41
14.70
14.99
10.45
9.54
30.65
31.47
36.58
32.38
42.85
34.63
41.25
18.3
17.8
15.4
17.4
13.1
16.2
13.6
2.91
1.43
2.51
5.60
5.06
6.86
7.23
temperature (430–630 K), corresponds to the reduction of carbon-
ate anions, which are not totally eliminated during pretreatment
(heat in air at 623 K for 60 min) [42]. It is remarkable that this event
shifts to lower temperatures when Ni content increases. It is pos-
sible that at these low temperatures, the reduction of very small
superficial particles of Ni has begun, having a catalytic effect on the
reduction of nearby carbonates [33]. The second event, a very broad
peak (650–1150 K) corresponds to the reduction of Ni(II) species
associated to mixed oxides [41,43].
Fig. 6. Ethanol conversion and H2, CO and CO2 yields.
Quantitative analysis of the TPR profiles showed that Ni has been
reduced between 86 and 97%, as presented in Table 2.
In order to identify the nature of these Ni(II) species in the sec-
ond event, a peak fitting of the profiles was made. Fig. 5 shows these
results, which had an accuracy of R2 = 0.997.
It can be seen that, for both fitting peaks (˛ and ˇ), maximum H2
consumption slightly shifts to higher temperatures as Mg(II) con-
tent increases. This fact is related to a stronger interaction of Ni(II)
atoms with neighboring cations (see Table 2). It is known that a
higher Mg(II) proportion makes Ni(II) less reducible, thus maxi-
mum H2 consumption moves to higher temperatures [24,32,38].
According to the literature, peak ˛ can be assigned to the reduc-
tion of Ni(II) in Ni1 − qMgqO matrix [35–38], and peak ˇ corresponds
to Ni(II) in a non-stoichiometric spinel-like phase [44,45], con-
firmed by PXRD. Particularly, this spinel-like phase competes for
divalent cations, taking up preferably Mg(II), and then Ni(II) when
necessary [22].
Fig. 7. Methane, ethylene and acetaldehyde yields.
The highest conversion is obtained with the catalyst HT0.00, as
shown in Fig. 6. With increasing Mg(II) content, at the beginning
the conversion diminishes, reaching a minimum, then increases a
little to remain constant.
Hydrogen, CO and CO2 yields have a similar behavior, after a
minimum for Mg/Ni = 0.04, they remain constant at 70, 20 and 55%
respectively.
Methane yield kept constant and lower than 3% (Fig. 7) for all
samples. Acetaldehyde yield stabilizes fewer than 10%. Ethylene
yield shows higher values (almost 30%) for samples with low Mg
content, but it decreases abruptly to values under 1% when Mg
content increases.
On the other hand, the performance of the catalysts was tested
in the ethanol steam reforming reaction. Figs. 6 and 7 and Table 3
show the catalytic results.

402
A. Romero et al. / Applied Catalysis A: General 470 (2014) 398–404
Fig. 9. Scheme of the evolution of the precursors during thermal treatment, as a
function of Mg(II) content.
particular case of the ternary Ni(II)–Mg(II)–Al(III) LDHs, the start-
ing oxide can be described as an ill crystallized spinel-like phase,
in which all the cations share a single cubic oxygen framework.
While Mg(II) and Al(III) can occupy either tetrahedral or octahedral
positions within it, Ni(II) cations are exclusively placed at octahe-
dral ones, irrespective of the Ni(II) to Mg(II) ratio [39]. With further
thermal treatment, the parent spinel-like metastable phase splits
into two stable phases (demixing). In Ni-rich samples (Mg/Ni = 0
to 0.33) it is expected in the total demixing scenario, the segre-
gation of pure NiO and ternary Ni1 − pMgpAl2O4 spinel, assuming
that Mg(II) cations preferentially accumulate in the latter phase,
in accordance with a previous report [53], made up with all the
available Al(III) in the hydroxycarbonate precursors. Eqs. (1) and
(2) show this evolution.
Fig. 8. Intrinsic activity (H2 yield per metallic area unit) and TPR % ˛ area vs. %w Ni.
Several authors consider ethylene as one of the principal respon-
sible for coke deposition in steam reforming [46–49] and attribute
to Mg(II) the capacity of giving higher mobility to anions OH− and,
thus, favoring ethanol dehydrogenation to acetaldehyde instead of
dehydration to ethylene [1,30].
The sample HT0.33 has the best results set: high conversion
(about 70%), the highest H2 yield (82%) and the lowest ethylene
and acetaldehyde yields.
In order to study the nature of the active sites for all catalysts,
it was analyzed the H2 yield per m2 of metallic surface (parame-
ter related with intrinsic activity). For this purpose, metallic area
and active particle size were measured by H2 chemisorption. The
obtained results are indicated in Table 3. In this regard, as Mg
content increases, Ni area (m2/gcat) decreases [30]. However, Ni
metallic area per gNi remains constant, which indicates that Mg(II)
content does not affect the metallic Ni dispersion.
Eq. (1): Decomposition
NixMgyAlz(OH)2(CO3)z/2 · nH2O → sl − NixMgyAlzO1+(z/2)
z
+
CO2 + (n + 1)H2O x + y + z = 1
(1)
2
Eq. (2): Demixing (for Ni-rich samples)
z
sl − NixMgyAlzO1+(z/2)
→
Ni(1−(2y/z))Mg(2y/z)Al2O4
2
ꢀ
ꢁ
H2 yield per m2 of metallic surface is shown in the last column of
Table 3 and in Fig. 8 as a function of %w Ni(II). It can be seen that the
catalysts present several active sites with different intrinsic activ-
ity. However, Mg(II) content does not affect uniformly the nature
of the active sites: the samples with low Mg(II) content present
low intrinsic activity (Mg/Ni = 0 up to 0.33); on the other hand, the
samples with higher Mg(II) content present higher intrinsic activ-
ity (Mg/Ni = 0.33–1.57).It seems to be a break point in the tendency
for the sample HT0.33.
Besides, when TPR % ˛ area (related to % Ni1 − qMgqO solid solu-
tion present in the sample) is analyzed in Fig. 8, both tendencies
observed for intrinsic activity are present. In a previous work [22]
we had proposed a model to explain in a schematic approach the
structural and textural changes in the surface for Ni-rich samples
(Mg/Ni = 0 up to 0.33) when they were submitted to different ther-
mal treatments.
According to this model, during the thermal decomposition
a complex phase segregation process takes place, in which the
involved cations redistribute in order to reach their correspondent
stable, commonly stoichiometric, oxides [50–52], before the birth
of metallic Ni particles, when reducing atmospheres are employed.
The rate of segregation and the observed intermediate phases
are strongly dependent on the nature of the involved cations; in the
2y
z
+
+ x − 1 NiO
(2)
However, according to the presented model, the second group
of samples (Mg/Ni = 0.33–1.57) has a different behavior. In the total
demixing scenario for these Mg-rich samples, it is expected the seg-
regation of pure MgAl2O4 spinel and Ni1 − qMgqO solid solution. In
the oxide phase, only part of the Ni(II) present would be reducible,
depending of the strength of the interaction Ni(II)–Mg(II), since
both cations form a solid solution oxide. In order to study this
possibility, an XPS study was made for the sample HT1.57 in its
reduced form. It was found that surface molar fraction of Al(III),
Mg(II) and Ni(II) was 0.54, 0.41 and 0.05 respectively. In compar-
ison to the bulk results (shown in Table 1), it is observed that the
surface becomes poor in Ni(II), indicating a spinel surface coverage
of about 85% of the corresponding theoretical composition lines.
Eqs. (3) and (4) explain this behavior. Fig. 9 shows the evolution of
the solids according to Mg(II) content.
Eq. (3): Decomposition
NixMgyAlz(OH)2(CO3)z/2 · nH2O → Sl − NixMgyAlzO1+(z/2)
z
+
CO2 + (n + 1)H2O x + y + z = 1
(3)
2

A. Romero et al. / Applied Catalysis A: General 470 (2014) 398–404
403
Table 4
TPO results.
Sample
CO2 maximum
temperature
CO2 area
(×100)
%wt Ni (relative to
HT0.00)
HT0.00
HT0.33
HT0.72
HT1.57
875
867
860
832
8.857
3.612
1.827
0.519
100%
52%
34%
15%
et al. [1,30] who claims that higher Mg(II) content leads to a higher
capacity for water adsorption-dissociation and oxygen species
mobility. All these characteristics produce materials with better
resistance to carbon deposition. They conclude that resistance to
carbon deposition increases when metal dispersion increases and
metallic particle size decreases. Higher Mg(II) content leads to
smaller metallic area and metallic particles. Carbon growing on Ni0
particle is enhanced when particle size is greater. Our results sup-
port this fact, since the sample with lowest particle size presents
the lowest coke deposition.
Fig. 10. TPO profiles of different samples.
Eq. (4): Demixing (for Mg-rich samples)
ꢀ
ꢁ
z
z
sl − NixMgyAlzO1+(z/2)
→
MgAl2O4
+
x + y −
Ni1−qMgqO
2
2
All profiles show an only CO2 peak. This fact leads to think that an
only type of coke is formed in the bed. Therefore, basing in Kitiyanan
et al. results [55], we can conclude that this only type of coke is a
mixture of monotubular and multitubular nanofibers.
y − (z/2)
q =
(4)
x + y + (z/2)
The two observed tendencies indicate that in the first group (Ni-
rich samples), reducible Ni(II) (the active phase in steam reforming)
comes from the NiO phase, meanwhile in the second group (Mg-
rich samples), comes from the Ni1 − qMgqO solid solution phase.
The model predicts a break point in the behavior, for the case of
the solid having just enough Mg(II) to form a pure MgAl2O4 spinel
and NiO phases. This is in good agreement with our characteriza-
tion results. The presence of different environment for the active
sites in the samples is reflected by the results of activity in steam
reforming. The samples with low Mg(II) content present a mixed
spinel-like phase, MgxNi1 − xAl2O4, which would be more acidic
than pure MgAl2O4 [9], presented in the samples with higher Mg(II)
content. This fact favors the dehydration of ethanol to ethylene in
samples with low Mg(II) content.
This type of coke does not lead to an activity loss, but a bed
porosity decrease and thus an increase in the reactor pressure.
The obtaining of an only CO2 peak, and no CO peak, indicates
a lineal relation between gasified CO2 amount and formed coke
amount on the catalyst surface during reaction.
Taking this into account, it is possible to conclude that an
increase in the Mg(II) content in the precursor leads to a decrease
in the coke deposition during reaction.
4. Conclusions
The Ni(II)–Mg(II)–Al(III) LDH obtained by homogeneous precip-
itation urea method, after a proper thermal treatment resulted in
an active catalyst for the ethanol steam reforming. The results show
that addition of a basic cation such as Mg(II) redounds favorably in
the resistance to carbon deposition. However, in terms of activity
and product distribution in ESR reaction, there is an optimal Mg/Ni
ratio = 0.33, which shows the best results. This is related to the fact
that when Mg/Ni ratio varies, the interaction of Ni0 in the oxide
matrix changes, and so it does the nature of the active sites.
It can be concluded from this report that the appropriate ratio
between the cations Ni(II)–Mg(II)–Al(III), and after a proper reduc-
tion treatment, would allow to control the interaction of the Ni0
active centers with the environment and the solids structure,
redounding in an improved activity, H2 yield and resistance to
carbon deposition.
Besides, it is noticeable that TPR ˛ and ˇ events shift at lower
temperatures when Mg(II) content increases, without showing a
break point (see Table 2). This must be due to the fact that dur-
ing thermal treatment, reduction occurs simultaneously with the
demixing (phase segregation).
Beyond the textural model depicted herein, low yields of hydro-
gen for the Ni rich samples results from undesired C2 products as
ethylene and acetaldehyde, which subsequently drive to coke for-
mation. This fact could obey to the occurrence of Al(III) acid sites;
their presence was documented for related solids obtained by the
straight decomposition of Mg–Al LDH and Ni–Al LDH [54]. The
latter, in particular reflected a delayed crystallization of NiAl2O4,
respect to MgAl2O4, leaving significant amounts of surface amor-
phous alumina, silent to PXRD inspection, in which acid Al(III) sites
are likely. The aforementioned behavior seems to remain, at least
partially in our samples, particularly on those in which Ni ions are
more abundant than Mg ones within the parent LDH lattice.
On the other hand, it is well known that the addition of Mg(II)
considerably improves the resistance of the catalyst to the coke
deposition. In order to analyze this effect, kinetic experiments have
been carried out using a mass of catalyst of 30 mg, increasing the
residence time, and reaching total conversion. Thus, by increasing
the permanence time of carbon precursors in the reactor, it favors
coke deposition.
Acknowledgements
Authors would like to thank to ANPCyT, CONICET and UBA, for
the economic support. Also, thanks are given to ANPCyT for Grant
PME 8-2003 to finance the purchase of the UHV Multi Analysis
System. MJ also acknowledge CONICET for the PIP 5215 project
financial support. The careful and devoted revision provided by the
expert reviewer is deeply appreciated by the authors.
References
TPO technique was applied on these used samples in order to
carry out a qualitative analysis about the type and amount of coke
formed on the bed. Fig. 10 and Table 4 show these results.
A pronounced diminish of the CO2 amount is observed when
proportion Mg/Ni increases. These results are according with Melo
[1] F. Melo, N. Morlanés, Cat. Today 133–135 (2008) 383–393.
[2] J. Comas, F. Marin˜o, M. Laborde, N. Amadeo, Chem. Eng. J. 98 (1–2) (2004) 61–68.
[3] F. Marin˜o, E. Cerrella, S. Duhalde, M. Jobbágy, M. Laborde, Int. J. Hydrogen Energy
23 (1998) 1095–1101.
[4] S. Cavallaro, Energy Fuels 14 (2000) 1195–1199.
[5] J. Llorca, N. Homes, J. Sales, P. Ramírez de la Piscina, J. Catal. 209 (2002) 306–317.

404
A. Romero et al. / Applied Catalysis A: General 470 (2014) 398–404
[6] F. Marin˜o, G. Baronetti, M. Jobbágy, M. Laborde, Appl. Catal. A: Gen. 238 (2003)
[32] V. Rives, M.A. Ulibarri, Coordin. Chem. Rev. 181 (1999) 61–120.
[33] G.J. de, A.A. Soler-Illia, M. Jobbágy, A.E. Regazzoni, M.A. Blesa, Chem. Mater. 11
(1999) 3140–3146.
[34] O. Lebedeba, D. Tichit, B. Coq, Appl. Catal. A: Gen. 183 (1999) 61–71.
[35] O.W. Pérez-López, A. Senger, N.R. Marcilio, M.A. Lansarin, Appl. Catal. A: Gen.
303 (2006) 234–244.
41–54.
[7] J. Llorca, P. Ramírez de la Piscina, J.A. Delmon, J. Sales, N. Homes, Appl. Catal. B:
Environ. 43 (2003) 355–369.
[8] S. Cavallaro, V. Chiodo, S. Freni, N. Mondillo, F. Frusteri, Appl. Catal. A: Gen. 249
(2003) 119–128.
[9] F. Aupretre, C. Descorme, D. Duprez, D. Casanave, D. Uzio, J. Catal. 233 (2005)
464–477.
[10] A. Akande, R. Idem, A. Delai, Appl. Catal. A: Gen. 287 (2005) 159–175.
[11] A. Haryanto, S. Fernando, N. Murali, S. Adhikari, Energy Fuels 19 (2005)
2098–2106.
[36] J. Feng, Y. Ding, Y. Guo, X. Li, W. Li, Fuel 109 (2013) 110–115.
[37] M. Kong, Q. Yang, J. Fei, X. Zheng, Int. J. Hydrogen Energy 37 (2012)
13355–13364.
[38] M. Kong, Q. Yang, W. Lu, Z. Fan, J. Fei, X. Zheng, T.D. Wheelock, Chin. J. Catal. 33
(2012) 1508–1516.
[12] S. Velu, K. Suzuki, M. Vijayaraj, S. Barman, C.S. Gopinath, Appl. Catal. B: Environ.
55 (2005) 287–299.
[13] V. Mas, M.L. Dieuzeide, M. Jobbágy, G. Baronetti, N. Amadeo, M. Laborde, Catal.
Today 133–135 (2008) 319–323.
[14] F. Frusteri, S. Freni, V. Chiodo, L. Spadaro, O. Di Biasi, G. Bonura, S. Cavallaro,
Appl. Catal. A: Gen. 270 (2004) 1–7.
[15] D.L. Trimm, A.A. Adesina, Praharso, N.W. Cant, Catal. Today 93–95 (2004) 17–22.
[16] X. Wang, R.J. Gorte, Appl. Catal. A: Gen. 224 (2002) 209–218.
[17] Q. Ming, T. Healey, L. Allen, P. Irving, Catal. Today 77 (2002) 51–64.
[18] J.R. Rostrup-Nielsen, Catalysis science and technology, in: J.R. Anderson, M.
Bouduart (Eds.), Catalytic Steam Reforming, vol. 5, Springer-Verlag editions,
1984 (Chapter 1).
[19] J.R. Rostrup Nielsen, J. Sehested, Adv. Catal. 47 (2002) 65–138.
[20] J.R. Rostrup-Nielsen, J. Catal. 33 (1974) 184–201.
[21] D. Tichit, B. Coq, CATTECH 7 (2003) 206–217.
[39] Y. Wang, H. Liu, B. Xu, J. Mol. Catal. A: Chem. 299 (2009) 44–52.
[40] M. Gazzano, W. Kagunya, D. Matteuzzi, A. Vaccari, J. Phys. Chem. B 101 (23)
(1997) 4514–4519.
[41] O. Clause, M. Gazzano, F. Trifirò, A. Vaccari, L. Zatorski, Appl. Catal. 73 (2) (1991)
217–236.
[42] T. Tichit, F. Medina, B. Coq, R. Dutartre, Appl. Catal. A: Gen. 159 (1997) 241–258.
[43] J. Rodríguez, J. Hanson, A. Frenkel, J. Kim, M. Pérez, J. Am. Chem. Soc. 124 (2002)
346–354.
[44] K. Park, K.Y. Kim, L. Lu, T.H. Lim, S.A. Hong, H.I. Lee, Fuel Cells 07 (3) (2007)
211–217.
[45] A. Penkova, L. Bobadilla, S. Ivanova, M.I. Domínguez, F. Romero-Sarria, A.C.
Roger, M.A. Centeno, J.A. Odriozola, Appl. Catal. A: Gen. 392 (2011) 184–191.
[46] A.J. Vizcaíno, P. Arena, G. Baronetti, A. Carrero, J.A. Calles, M.A. Laborde, N.
Amadeo, Int. J. Hydrogen Energy 33 (2008) 3489–3492.
[47] A. Carrero, J.A. Calles, A.J. Vizcaíno, Chem. Eng. J. 163 (3) (2010) 395–402.
[48] F. Aupretre, C. Descorme, D. Duprez, D. Casanave, D. Uzio, J. Catal. (2005)
464–477.
[22] A. Romero, M. Jobbágy, M. Laborde, G. Baronetti, N. Amadeo, Catal. Today 149
(2010) 407–412.
[23] F. Cavani, F. Trifirò, A. Vaccari, Catal. Today 11 (1991) 173–301.
[24] A. Vaccari, Appl. Clay Sci. 14 (1999) 161–198.
[49] H. Wang, Y. Liu, L. Wang, Y.N. Qin, Chem. Eng. J. 145 (2008) 25–31.
[50] M. Bellotto, B. Rebours, O. Clause, J. Phys. Chem. US 100 (20) (1996) 8527–8534.
[51] M. Bellotto, B. Rebours, O. Clause, J. Phys. Chem. US 100 (20) (1996) 8535–8542.
[52] E. Sileo, M. Jobbágy, C. Paiva-Santos, A. Regazzoni, J. Phys. Chem. B 109 (2005)
10137–10141.
[53] F. Basile, L. Basini, M. D’Amore, G. Fornasari, A. Guarinoni, D. Matteuzzi, G. Del
Piero, F. Trifirò, A. Vaccari, J. Catal. 173 (1998) 247–256.
[54] B. Rebours, J.B.D. De La Caillerie, O. Clause, J. Am. Chem. Soc. 116 (5) (1994)
1707–1717.
[25] T. Borowieki, Appl. Catal. A: Gen. 4 (1987) 207–220.
[26] Johnsen, E.Norby Rune, J. Phys. Chem. C 113 (44) (2009) 19061–19066.
[27] Y. Yang, X. Zhao, F. Zhang, Chem. Mater. 24 (2012) 81–87.
[28] N. Fairley, Casa XPS Software, 2.3.13 Version, 2007-12-03.
[29] C. Wagner, A. Naumkin, A. Kraut-Vass, J. Allison, C. Powell, J. Rumble, NIST X-
Ray Database 20, Version 3.4 (Web version). NIST Standard Reference Database
20, National Institute of Standards and Technology, Gaithersburg, USA.
[30] F. Melo, N. Morlanés, Catal. Today 107/108 (2005) 458–466.
[31] M.J. Holgado, V. Rives, M.S. San Román, Appl. Catal. A: Gen. 214 (2001) 219–228.
[55] B. Kitiyanan, W.E. Alvarez, J.H. Harwell, D.E. Resasco, Chem. Phys. Lett. 317
(2000) 497–503.