Activated takovite catalysts for partial hydrogenation of ethyne, propyne, and propadiene

Article abstract of DOI:10.1016/j.jcat.2008.07.012

The gas-phase hydrogenation of ethyne, propyne, and propadiene was investigated over partially reduced Ni-Al mixed oxides derived from takovite, a hydrotalcite-type compound. The unique attributes of the hydrotalcite route leads to more active and selective catalysts compared to conventional Ni/Al₂O₃ prepared by impregnation. Tuning calcination and reduction conditions of the catalyst precursor is essential to optimize the hydrogenation performance. The best catalyst, calcined and reduced at 773 K, rendered stable propene yields up to ca. 65% and consisted of a Ni(Al)O_x solid solution with 55% of the total bulk nickel in reduced form and surface enrichment by aluminum. Sintering of NiO and crystallization of NiAl₂O₄ at high calcination temperature induce lower activity. The alkyne or diene conversion increases with the percentage of metallic Ni in the samples, while an optimal degree of nickel reduction maximizes the monoalkene selectivity. Below the optimum, oligomer formation is favored and above the optimum, alkane production increases. A similar pattern was found for the H₂/HC ratio. The alkene selectivity experiences a dramatic increase in early stages of the reaction, which correlated with the build-up of C-containing species on the catalyst (sub-)surface. These selectivity-enhancing species are formed at specific reaction temperatures, highlighting the relevance of the testing procedure on assessing hydrogenation catalysts. The catalytic performance is strongly influenced by the hydrocarbon substrate. In contrast to propyne and propadiene, ethyne hydrogenation led to a C₂H₄ yield of only 6%.

Full text of DOI:10.1016/j.jcat.2008.07.012

Journal of Catalysis 259 (2008) 85–95
Contents lists available at ScienceDirect
Journal of Catalysis
www.elsevier.com/locate/jcat
Activated takovite catalysts for partial hydrogenation of ethyne, propyne,
and propadiene
a
a
a
a,b,∗
Sònia Abelló , Danny Verboekend , Blaise Bridier , Javier Pérez-Ramírez
a
Institute of Chemical Research of Catalonia (ICIQ), Avinguda Països Catalans 16, 43007 Tarragona, Spain
Catalan Institution for Research and Advanced Studies (ICREA), Passeig Lluís Companys 23, 08010 Barcelona, Spain
b
a r t i c l e i n f o
a b s t r a c t
Article history:
The gas-phase hydrogenation of ethyne, propyne, and propadiene was investigated over partially reduced
Ni–Al mixed oxides derived from takovite, a hydrotalcite-type compound. The unique attributes of
the hydrotalcite route leads to more active and selective catalysts compared to conventional Ni/Al2O3
prepared by impregnation. Tuning calcination and reduction conditions of the catalyst precursor is
essential to optimize the hydrogenation performance. The best catalyst, calcined and reduced at 773 K,
rendered stable propene yields up to ca. 65% and consisted of a Ni(Al)Ox solid solution with 55%
of the total bulk nickel in reduced form and surface enrichment by aluminum. Sintering of NiO and
crystallization of NiAl2O4 at high calcination temperature induce lower activity. The alkyne or diene
conversion increases with the percentage of metallic Ni in the samples, while an optimal degree of nickel
reduction maximizes the monoalkene selectivity. Below the optimum, oligomer formation is favored and
above the optimum, alkane production increases. A similar pattern was found for the H2/HC ratio. The
alkene selectivity experiences a dramatic increase in early stages of the reaction, which correlated with
the build-up of C-containing species on the catalyst (sub-)surface. These selectivity-enhancing species are
formed at specific reaction temperatures, highlighting the relevance of the testing procedure on assessing
hydrogenation catalysts. The catalytic performance is strongly influenced by the hydrocarbon substrate.
In contrast to propyne and propadiene, ethyne hydrogenation led to a C2H4 yield of only 6%.
Received 6 May 2008
Revised 19 July 2008
Accepted 25 July 2008
Available online 28 August 2008
Keywords:
Selective hydrogenation
Alkyne
Diene
Alkene
Ni–Al oxides
Takovite
Hydrotalcite
Operando spectroscopy
Catalyst activation
Carbon deposits
©
2008 Elsevier Inc. All rights reserved.
1
. Introduction
supported nickel catalysts prepared by impregnation are more se-
lective to the alkane while displaying extensive oligomerization [7,
8,11,12]. In fact, nickel is frequently used for alkene hydrogena-
tion or total alkyne hydrogenation [1,3]. Besides, nickel requires
higher operating temperatures (up to 623 K) than palladium (323–
An important segment of commercial catalytic hydrogenations
in the petrochemical and fine (specialty) chemical industries is the
selective transformation of organic compounds with conjugated
or cumulative double bonds and/or triple bonds into the corre-
sponding monoalkene [1–5]. Supported palladium (<0.1 wt% Pd
on low-surface-area alumina) is the preferred catalyst for semi-
hydrogenation of alkyne (e.g. ethyne, propyne) and alkadiene (e.g.
propadiene, butadiene) impurities to olefins in C2–C5+ fractions
of steam crackers [2,4,6–8]. ‘Bare’ Pd catalysts experience over-
hydrogenation to the alkane and formation of higher hydrocar-
bons by oligomerization (green oil). Accordingly, modification of
monometallic Pd systems by promoters and/or additives is essen-
tial to increase the alkene selectivity and lifetime [3,5,8,9]. A repre-
sentative example of modified palladium catalyst for partial hydro-
genation of triple bonds in the fine chemical industry is the Lindlar
system, consisting of 5 wt% Pd supported on CaCO3 promoted by
lead diacetate and quinoline [10]. Compared to palladium catalysts,
353 K) [1]. Strategies to prevent total hydrogenation over nickel
catalysts include promotion by additional metals (e.g. Cu, Re, Pb,
Ag, Bi, Mo) [7,13,14], partial presulfiding, and continuous dosing of
small H2S amounts to the reactor [1].
Pd has monopolized recent open-literature publications on hy-
drogenation of alkynes, dienes, and alkenes [5,15–19], which was
traditionally focused on supported nickel (on alumina or pumice),
Raney nickel, and nickel boride catalysts [20–26]. Supported gold
nanoparticles also attract recent attention due to the superior
thermodynamic selectivity with respect to palladium [27]. In the
nineties, Monzón et al. [28] prepared Ni/NiAl2O4 by coprecipita-
tion of Ni –Al2 precursors (22 wt% Ni) followed by calcination at
1
1173 K and reduction at 773 K. Tests in ethyne hydrogenation at
4
23 K led to very low ethene yields (ca. 6%) and an extremely
−1
−1
high coking rate (up to 7.5 gcoke
g
h ). Incorporation of Zn, Cr,
cat
Corresponding author at: Institute of Chemical Research of Catalonia (ICIQ),
Avinguda Països Catalans 16, 43007 Tarragona, Spain. Fax: +34 977 920 224.
*
and/or Fe in the base Ni3–Al1 hydrotalcite increased the ethene
yield to 13% [29,30].
E-mail address: jperez@iciq.es (J. Pérez-Ramírez).
0
021-9517/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2008.07.012

86
S. Abelló et al. / Journal of Catalysis 259 (2008) 85–95
In general, hydrogenation studies of acetylenics and dienes over
spectrometer. Before analysis, the sample was mixed with Li2B4O7
and LiI as tensoactive agent, and the mixture was fused at 1423 K
in a Perle’X-3 radiofrequency furnace.
nickel catalysts used a fixed condition for activation of the precur-
sor material (calcination and reduction) and catalytic tests have
been typically conducted with ethyne [12,20,26,31,32]. Compre-
hensive relationships between the catalyst synthesis, structure,
procedure for catalyst evaluation, and the resulting hydrorefining
performance have not been derived. Besides, studies assessing the
sensitivity to the hydrocarbon in terms of chain length (e.g. ethyne
and propyne) and type of insaturation (e.g. propyne and propadi-
ene) have not been reported. Monzón et al. [28] showed that the
degree of nickel reduced can be controlled by the reduction tem-
perature of the Ni–Al oxide, but the influence thereof on ethyne
hydrogenation was not examined. Phillipson et al. [25] observed
that the isomer selectivity in the hydrogenation of 1,3-butadiene is
affected by the reduction temperature of Ni/Al2O3. Apart from re-
duction conditions, well established is also that the calcination of
nickel oxide impacts on the reduction properties, as early shown
by Benton and Emmett [33]. Consequently, the possibility rises to
design rules for more efficient nickel-based hydrogenation catalysts
by proper selection of the synthesis and activation conditions.
Herein, we have investigated unpromoted reduced Ni–Al mixed
oxides derived from takovite (Ni–Al hydrotalcite) in the gas-phase
hydrogenation of ethyne, propyne, and propadiene. The hydrotal-
cite route makes it possible to obtain multimetallic mixed oxides
combining high metal loadings, high metal (inter)dispersion, and
high surface area [34–36]. Structure–performance relationships as-
sociated with the activation of the takovite precursor have been
derived by detailed characterizations and catalytic tests in a con-
tinuous fixed-bed reactor at different temperatures and hydrogen-
to-hydrocarbon ratios. Operando infrared spectroscopy monitored
the formation of carbon-containing species on the catalyst surface
under reaction conditions, which has been correlated with changes
in alkene selectivity.
Powder X-ray diffraction was studied in a Brucker-AXS D5005
θ–θ diffractometer equipped with a Bruker-AXS MRI high-tempera-
ture chamber and a diffracted beam graphite monochromator
using CuKα radiation. A thin layer of sample (ca. 30 mg) was
mounted on the Pt90–Rh10 heater strip by placing a few droplets
of a suspension of finely ground sample in ethanol followed by
drying under ambient conditions. The as-synthesized and calcined
−1
samples were measured at 298 K in N2 (100 ml STP min ). In situ
XRD patterns during reduction of the Ni–Al oxides were recorded
−1
in 5 vol% H2 in N2 (100 ml STP min ) at 423, 623, 773, and 1173 K
after 30 min equilibration at each temperature. The heating rate
−
1
between each temperature step was 5 K min . Data were col-
◦
◦
◦
lected at 2θ = 10 –70 with a step size of 0.05 and a counting
time of 1 s per step.
Thermal analysis was performed in a Mettler Toledo TGA/
SDTA851e microbalance. Analyses were performed in air (50 ml STP
−1
−1
min ), ramping the temperature from 298 to 1173 K at 5 K min .
N2 isotherms at 77 K were measured in a Quantachrome
Autosorb-1 MP analyzer. Before analysis, the samples were de-
gassed in vacuum at 393 K for 12 h.
Temperature-programmed reduction with hydrogen (H2-TPR)
was measured in a Thermo TPDRO 1100 unit equipped with
a thermal conductivity detector. Ca. 50 mg of Ni–Al oxide ob-
tained by calcination of takovite at 623, 773, and 1173 K were
loaded in the quartz microreactor (11 mm i.d.), pretreated in N2
−1
(20 ml STP min ) at 623 K for 1 h, and cooled to 323 K in the
same atmosphere. The analysis was carried out in a mixture of
−
1
5 vol% H2 in N2 (20 ml STP min ), ramping the temperature from
−1
323 to 1173 K at 10 K min . H2-TPR was also carried out over
prereduced samples. For this purpose, the calcined samples were
in situ reduced in 5 vol% H2 in N2 (42 ml STP min ) at different
temperatures (423, 623, and 773 K) and times (30 min and 5 h)
previous to the analysis.
−1
2
. Experimental
2
.1. Catalyst preparation and activation
X-ray photoelectron spectroscopy (XPS) measurements were ob-
tained using a VG-Microtech Multilab 3000 spectrometer equipped
with a MgKα (hν = 1253.6 eV) excitation source, a nine-channel-
tron detection system, and a hemispheric electron analyzer. The
sample was outgassed overnight at room temperature in a UHV
Ni–Al hydrotalcite with a nominal Ni/Al molar ratio of 3, known
as takovite, was synthesized by continuous coprecipitation us-
ing the ILDP method [37]. Briefly, aqueous solutions of 0.75 M
Ni(NO3)2·6H2O and 0.25 M Al(NO3)3·9H2O and the precipitating
agent (NaOH + Na2CO3, 1 M of each) were pumped into a 6 ml-
microreactor attached to a high-shear homogenizer rotating at
−8
chamber (<5 × 10
Torr). In some cases, before XPS measure-
ments, the specimen was transferred to a high-pressure reaction
−3
cell (10
Torr) attached to the analysis chamber of the spectrom-
1
3,500 rpm. The pH of the slurry was measured and controlled by
eter and pretreated in H2 at 773 K. Reduced samples were imme-
diately transferred to the analysis chamber. All binding energies
were referenced to the C 1s line at 284.6 eV, and the integrated
intensities were corrected by atomic sensitivity factors [38].
an in-line probe directly at the outlet of the precipitation chamber.
The synthesis was carried out at constant pH 10 with an average
residence time of 18 s. The product slurry was collected in a glass
vessel and aged at 313 K for 12 h under stirring, followed by fil-
tration, washing, and drying at 353 K for 12 h. The as-synthesized
sample, denoted as ‘AS,’ was calcined in static air at different tem-
2.3. Catalytic tests
−1
peratures (623, 773, or 1173 K) for 15 h using a ramp of 5 K min
.
The gas-phase hydrogenation of propyne, propadiene, and
ethyne was studied at ambient pressure in a MicroActivity Refer-
ence setup (PID Eng&Tech). The catalyst (0.15 g, sieve fraction 200–
400 μm) was loaded in the 12 mm i.d. quartz microreactor, heated
in He at 623 K for 30 min, and prereduced in 5 vol% H2 in He
Previous to catalytic tests (Section 2.3), the calcined samples were
treated in He at 623 K for 30 min and cooled or heated in the
same gas to the desired prereduction temperature (423, 623, or
7
3
73 K), which was done in a mixture of 5 vol% H2 in He during
0 min. Occasionally, a 5 h-reduction treatment was carried out.
−1
(42 ml STP min ) as described in Section 2.1. The catalysts were
tested isothermally at intervals of 50 K in the range 373–523 K
using a typical feed mixture of (alkyne or alkadiene)/H2/He =
−1
The ramp used in the temperature programs was 5 K min . Along
the manuscript, the samples are designated as ‘Cx–Ry,’ where ‘x’
and ‘y’ stand for the calcination and reduction temperatures, re-
spectively.
−
1
2.5/7.5/90 and a total flow of 42 ml STP min (weight hourly space
velocity, WHSV = 16,800 ml h
−1
−1
g ). The tests were run starting
at the highest or lowest temperature, with a holding time of 5 h at
each temperature. The influence of the hydrogen-to-propyne ratio
2.2. Catalyst characterization
(1–5) was studied over selected catalysts at 523 K. For this pur-
The chemical composition of the as-synthesized Ni–Al hydro-
talcite was determined by X-ray fluorescence in a Philips PW 2400
pose, the inlet hydrocarbon concentration was kept at 2.5 vol% and
the hydrogen concentration was decreased from 12.5 to 2.5 vol%

S. Abelló et al. / Journal of Catalysis 259 (2008) 85–95
87
by balancing the mixture with He in order to keep the total flow
constant. The product gases were analyzed by a gas chromatograph
Agilent GC6890N) equipped with a GS-GasPro column and a ther-
mal conductivity detector. The turnover frequency was expressed
(
−1
−1
as [mole alkyne/alkadiene converted·s ·mole Ni reduced ]. The
selectivity to the alkene (alkane) was determined as the mole of
alkene (alkane) formed divided by the mole of propyne, propadi-
ene, or ethyne reacted. The alkene yield was determined as the
product of alkyne (alkadiene) conversion and alkene selectivity.
The fractional selectivity to oligomers (often referred to as green
oil) was determined as S(oligomers) = 1 − S(alkene) − S(alkane).
2
.4. Operando FTIR spectroscopy
Fourier transform infrared spectroscopy was carried out in a
Nicolet 5700 spectrometer (Thermo Scientific) using a Spectrat-
ech collector II diffuse reflectance (DRIFT) accessory equipped with
a high-temperature chamber, ZnSe windows, and a MCT detector.
The sample cell was filled with powdered catalyst and carefully
leveled off to reduce reflections off the sample surface and to en-
sure reproducible results. The experiments were carried out repro-
ducing the procedures detailed in Section 2.3. Gases were dosed
into the cell by digital mass flow controllers. Spectra were auto-
matically recorded every 80 s during 5 h using KBr (Aldrich, IR
spectroscopy grade) under reaction conditions as the background.
−
1
The range 650–4000 cm
was covered by co-addition of 32 scans
−1
at a nominal resolution of 4 cm
.
Fig. 1. XRD patterns of the as-synthesized takovite and the products of calcination
and in situ reduction at different conditions. Crystalline phases: (1) takovite, (2)
NiO, (Q) NiAl2O4 and (!) Ni. The reflections with asterisk belong to the Pt95–Rh5
strip where the samples were mounted.
3. Results and discussion
3.1. Hydrotalcite activation
3.1.1. As-synthesized sample
Fig. 1 shows the XRD patterns of the coprecipitated mate-
rial and the products of calcination and reduction at different
conditions. The as-synthesized sample (AS) presents characteris-
tic reflections of takovite (Ni6Al2(OH)16 CO3·4H2O, JCPDS 15-0087).
The molar Ni/Al ratio in the solid determined by XRF (3.3) was
close to the nominal ratio of 3. Takovite had the platelet-like
morphology characteristic of hydrotalcite-type compounds [34,39]
(
Fig. S1 in Supplementary information). The sample was purely
3
−1
mesoporous, with a total pore volume of 0.32 cm g
and a
2
−1
BET surface area of 227 m g . Thermogravimetric analysis in
air (Fig. 2) shows the typical two-step decomposition pattern of
these clays [39], with transition temperatures at 375 K (inter-
layer water removal) and 570 K (dehydroxylation and decarbon-
ation).
3
.1.2. Calcination
As shown in Fig. 1 and summarized in Table 1, the XRD patterns
of the materials calcined at 623 and 773 K showed NiO (bunsen-
ite, JCPDS 47-1049), while both NiO and NiAl2O4 spinel coexist in
the material calcined at 1173 K. No distinctive reflections associ-
ated with any crystalline form of alumina were detected in the
samples. The nickel and aluminum content in the oxides is 58 wt%
and 8 wt%, respectively. In agreement with the literature [39–41],
the NiO reflections in the calcined samples were shifted to higher
Fig. 2. Thermogravimetric profile of the as-synthesized takovite in air and derivative
of the weight loss (inset).
et al. [40] reported similar d spacings for the hydrotalcite-derived
Ni–Al oxides calcined at 573 and 723 K.
The cell parameter of the NiO phase in C1173 (a = 0.4171 nm)
resembles that of pure nickel oxide, suggesting that at this cal-
cination temperature aluminum is segregated from the solid so-
lution forming the additional NiAl2O4 spinel phase. It has been
reported that at ca. 1073 K, Ni(Al)Ox starts to transform into ad-
ditional phases, namely NiAl2O4 [42]. Both nickel oxide and nickel
2
θ values than those corresponding to pure NiO, due to Al3 sub-
+
stitution in the nickel oxide lattice and the formation of a solid
solution, denoted as Ni(Al)Ox. Puxley et al. [42] described this
solid solution as a highly disordered metastable mixed oxide, in-
termediate in nature between pure nickel oxide and nickel alumi-
nate. Hence, the cubic cell of the solid solution in C623 and C773
aluminate present a cubic close-packed lattice of O2 ions, only
−
2
+
differing in the way in which the cation vacancies are filled. Ni
(
a = 0.4151–0.4157 nm) is significantly smaller than that of NiO
ions occupy octahedral sites in NiO and tetrahedral sites in the
normal spinel NiAl2O4, respectively. The concentration of nickel in
(a = 0.4177 nm, JCPDS 47-1049). In line with our results, Alzamora

88
S. Abelló et al. / Journal of Catalysis 259 (2008) 85–95
Table 1
Characterization data of the takovite-derived samples
Sample codea
Phasesb
Ni reducedc (%)
SBET (m2
g )
−1
C623
C773
C1173
C623–R423
C623–R623
C773–R423
C773–R623
C773–R773
C773–R773, 5 hd
C1173–R773
NiO
NiO
–
–
–
275
206
62
250
238
178
170
155
166
55
NiO + NiAl2O4
NiO
0
NiO
NiO
NiO
17
0
6
NiO + Ni
55
91
37
NiO + Ni
NiO + NiAl2O4 + Ni
a
C and R stand for calcination and reduction treatments, respectively. The num-
bers indicate the temperature in K.
b
Identified by XRD.
Percentage of metallic nickel determined by H2-TPR.
Reduced for 5 h. Reduction time in the other samples was 30 min.
c
d
Td sites increases with temperature (>1273 K). At lower temper-
2
+
ature, due to the preference of Ni
ions for octahedral sites, an
inverse nickel aluminate spinel is favored [43]. Accordingly, both
NiO and NiAl2O4 with different degrees of inversion coexist in
C1173. Besides, as revealed by H2-TPR (vide infra), a certain amount
of Ni(Al)Ox still remains in this sample.
The samples calcined at 623 and 773 K present no appre-
ciable differences in average crystallite size (ca. 3 nm), as de-
termined by the Scherrer method using the (220) reflection at
◦
2
(
θ = 63 . Calcination at 1173 K leads to larger NiO crystallites
ca. 8 nm). The average crystallite size of the NiAl2O4 spinel in
C1173 amounted to 16 nm (estimated using the (440) reflection
◦
at 2θ = 65 ). The morphology of these oxides was studied by
TEM. Representative micrographs are shown in Supplementary in-
formation. Uniform nanoparticles of ca. 10 nm were observed in
C773. The sample calcined at 623 K combines platelets inherited
from the as-synthesized material and small nanoparticles (3–5 nm)
characteristic of the oxide. The transformation from lamellae in
Ni–Al hydrotalcite to nodular particles in the corresponding ox-
ide was reported elsewhere [39,44]. In agreement with XRD, C1173
presents larger particles, in the range of 10–25 nm, due to sin-
tering of NiO and crystallization of NiAl2O4. Nitrogen adsorption
reveals the progressive decrease of BET surface area upon increas-
ing the calcination temperature, particularly at 1173 K (Table 1).
Alzamora et al. [40] also showed that the segregation of aluminum
from the Ni(Al)Ox solid solution and subsequent crystallization of
the NiAl2O4 spinel causes a drastic drop in surface area.
Fig. 3. H2-TPR profiles of the Ni–Al catalysts calcined and reduced at different con-
ditions.
aluminum in the NiO lattice. The broad reducibility range in C623
and C773 is likely due to the presence of nickel species in different
environments in terms of interaction with aluminum. This substan-
tiates the presence of a multiple range of structures in the model
by Puxley et al. [42].
It is well known that the presence of Al
creases the resistance of Ni
3+
in nickel oxide in-
2+
against reduction [39,42]. In analogy
with the work by Arnoldy and Moulijn [47] over the Co–Al system,
3
.1.3. Reduction
the number of Al3 ions in the surrounding of the Ni
+
2+
ion should
+
ions exert a lig-
H2-TPR was used to investigate the influence of the calcination
3
mainly determine the reduction temperature. Al
temperature on the reducibility of nickel in the samples and to
design prereduction treatments for deriving relationships between
the amount of nickel reduced and the hydrogenation performance
and effect, polarizing the more or less covalent Ni–O bonds, thus
increasing the effective charge of Ni2 ions, and as a consequence,
the lattice energy, resulting in a higher reduction temperature of
the intermediate Ni(Al)Ox phases as compared to NiO.
+
(Section 3.2). As shown in Fig. 3, the reduction of nickel occurs
in the broad range of 600–1050 K depending on the calcination
temperature of the takovite precursor. The unreduced C623 sample
shows a peak of hydrogen consumption centered at 800 K, with
a small peak at 645 K and a shoulder at 925 K. The unreduced
C773 sample shows a more symmetrical profile centered at 925 K
without the low-temperature contribution.
Pure NiO obtained by calcination of nickel nitrate at 773 K typ-
ically displays a single peak consumption at 673 K [45,46], more
or less matching the low-temperature peak in C623. The reduction
of nickel in Ni(Al)Ox solid solutions is hindered by aluminum [41],
causing the shift to higher temperatures. The reduction temper-
ature increases significantly from C623 (peak at 800 K) to C773
In the unreduced C1173 sample, the peak centered at 925 K
in C773 is unfolded in two peaks at 770 and 1050 K. The low-
temperature peak corresponds to NiO generated upon withdrawal
of aluminum from the Ni(Al)Ox solid solution, while the high-
temperature peak is attributed to the incipient NiAl2O4. The H2
consumption by NiO is much lower than that by NiAl2O4, which
does not account for the total segregation of aluminum from the
solid solution [41,48,49]. With a nominal Ni/Al ratio in the solid
of 3, the theoretical molar ratio NiO/NiAl2O4 is 5. Thus, a wide
range of oxidic structures coexist in this sample too, namely NiO,
Ni(Al)Ox, and NiAl2O4, covering the broad range of 800–1200 K.
H2-TPR was also carried out over in situ reduced samples using
different temperature and time. The prereduction temperature of
the samples never exceeded the calcination temperature. As shown
(peak at 925 K), strongly suggesting that the interaction of alu-
minum and nickel in the solid solution is enhanced in the latter
sample. This might be due to the favored solid-state diffusion of

S. Abelló et al. / Journal of Catalysis 259 (2008) 85–95
89
in Fig. 3, pretreatment of C623 in H2 at 423 K did not produce
detectable reduction of nickel species, while increasing the tem-
perature to 623 K led to the disappearance of the small peak at
6
45 K, attributed to free NiO. The area of the broad peak at 800 K
is somewhat decreased and slightly shifted to lower temperatures.
A similar behavior was observed for the C773 sample: prereduc-
tion at 423 K originated no change and the peak area gradually
decreased by H2 pretreatment at 623 and 773 K. The reduction
was particularly significant in the sample treated for 5 h. In the
same line, H2 pretreatment of C1173 at 773 K resulted in the dis-
appearance of the peak at 770 K due to complete NiO reduction,
and the decreased area of the peak at 1050 K is mainly due to
2
+
nickel reduction in Ni(Al)Ox phases, while Ni
in the spinel phase
is unreduced. The latter reasoning is grounded on the practically
unchanged XRD reflections of the NiAl2O4 phase upon reduction
(Fig. 1).
The fraction of metallic nickel was determined dividing the total
integral area of the prereduced samples by that of the unreduced
counterparts (Table 1). As expected, (i) the amount of reduced Ni
at a fixed calcination temperature increases with the prereduction
temperature (compare the C773–Ry series) and (ii) higher calci-
nation temperature requires higher prereduction temperature to
reach an equivalent percentage of metallic nickel (compare C623–
R623 and C773–R623 or C773–R773 and C1173–R773). In light of
these results, it can be concluded that the degree of bulk nickel
reduction in the oxide can be modulated by adjusting temperature
0
and time. For example, the amount of Ni in C773 is 0, 6, and 55%
upon prereduction at 423, 623, and 773 K for 30 min, respectively,
reaching 91% when the prereduction time at 773 K was 5 h.
H2-TPR results agree with in situ XRD studies during reduction
of the C773 sample (Fig. 1). The intensity of the NiO diffraction
lines in C773 and C773–R423 is virtually the same, and is slightly
reduced in C773–R623. Reflections of metallic Ni were not ob-
served in these samples. The intensity of NiO in C773–R773 signifi-
Fig. 4. Ni 2p3/2 core level spectrum of the C773 and C773–R773 samples.
Table 2
Comparative results of bulk and surface analyses over selected samples
0
Sample
Molar Ni/Al ratio (–)
Ni reduced (%)
cantly decreased and two broad reflections belonging to Ni (JCPDS
◦
◦
XRF
XPS
2
H -TPR
XPS
4
-850) emerged at 2θ = 44 and 52 . The degree of nickel reduc-
tion in C773–R773 was estimated as the ratio of the area of the
C773
3.3
3.3
1.4
1.0
0
55
0
15
◦
C773–R773
NiO (220) reflection at 2θ = 63 and the area of the same reflec-
tion in unreduced C773. The resulting value (57%) is in excellent
correspondence with that determined by H2-TPR (55%, Table 1).
Based on the (220) reflection, no NiO was identified in the C773–
R773 sample reduced for 5 h due to the complete nickel reduction.
The XRD pattern of C1173–R773 shows the virtual disappearance
of NiO, while NiAl2O4 reflections are still present, though exhibit-
ing a lower intensity compared to C1173. The average size of the
nickel crystallites in these reduced samples was estimated by the
C773–R773 (Fig. 4 and Table 2). The latter is the best hydrogena-
tion catalyst identified in our study (see Section 3.2). The Ni 2p3/2
core level spectrum of C773 shows a binding energy of 855.6 eV
and the corresponding shake-up satellite at ca. 6 eV above the
2+
principal line [50]. This peak, characteristic of Ni , is positioned
between the binding energy of NiO (854.1 eV [51]) and NiAl2O4
spinel (856.2 eV [52]). This further indicates the intimate interac-
◦
Scherrer method applied in the Ni (200) reflection at 2θ = 52 .
tion of Ni2 and Al
+
3+
in the catalyst, pointing once more to the
As expected, Ni0 crystallites in C773–R773 (3 nm) are smaller
than in C1173–R773 (6 nm) since the size of the metal is very
much inherited from the crystallite size of the nickel oxide pre-
cursor.
presence of metastable phases intermediate in nature between the
two structures. Noticeably, the atomic Ni/Al ratio of C773 deter-
mined by XPS amounts to 1.4, i.e. much lower than the bulk Ni/Al
ratio of 3.3 determined by XRF (Table 2). Consequently, the sur-
face of the oxide is enriched by aluminum. Various authors have
reported aluminum zoning on the surface of Ni–Al oxides derived
from thermal treatment of Ni–Al hydrotalcite [40,41,53]. For ex-
ample, based on alkaline leaching coupled with 27Al MAS-NMR
analyses, Rebours et al. [53] concluded that the oxide at 923 K
consists of NiO particles whose surface is decorated by aluminate-
type patches.
Finally, transmission electron microscopy and nitrogen adsorp-
tion were carried out in order to check whether the reduction
treatments affect the morphological and textural properties of the
precursor oxides. The morphology of the calcined and reduced
samples did not apparently change (Supplementary information).
The reduction treatments did not significantly alter the BET sur-
face area of the samples either (Table 1).
The Ni 2p3/2 core level spectrum of C773–R773 displays an ad-
3
.1.4. Surface vs bulk analyses
The bulk analyses presented above, namely XRD and H2-TPR,
0
ditional peak at 852.9 eV, which is assigned to Ni [54], although
the peak due to Ni2 is still predominant (Fig. 4). By peak decon-
volution and taking into account the corresponding satellite peaks,
it was determined that only 15% of the nickel probed by XPS is
in metallic form. The amount of nickel reduced on the surface is
much lower than that of 55% determined in the bulk (Table 2).
+
indicated that the oxide resulting from calcination of takovite at
73 K consisted of a Ni–Al solid solution. However, these tech-
7
niques provide no insight on the chemical state of nickel on the
surface and the aluminum distribution within the catalyst parti-
cle. For this purpose, XPS analysis was carried out over C773 and

90
S. Abelló et al. / Journal of Catalysis 259 (2008) 85–95
(
a)
(b)
(
c)
Fig. 5. Propyne conversion (a), propene selectivity (b), and propane selectivity (c) vs time during propyne hydrogenation over the Ni–Al catalysts. Conditions: C3H4/H2/He
−1
−1
h
ratio = 2.5/7.5/90, T = 373–523 K, WHSV = 16,800 ml g
, and P = 1 bar.
This result can be tentatively explained considering the marked
aluminum zoning, hindering more significantly nickel reduction at
the surface of the catalyst. The atomic Ni/Al ratio in C773–R773
determined by XPS was 1.0 (Table 2), indicating that the reduc-
tion treatment enhances the relative enrichment of the surface by
aluminum with respect to the calcined material. It has been sug-
gested that the reduction of the samples causes the progressive
diffusion of aluminum ions to the surface of the crystallite [40].
373 K. Tests from the lowest to the highest temperature were
also conducted and have a strong impact on the catalytic perfor-
mance. This aspect deserves a separate discussion (Section 3.3).
The propene selectivity greatly increased with time-on-stream in
the first 2 h of reaction at 523 K (Fig. 5b), with a concomitant
decrease of propane selectivity (Fig. 5c) and propyne conversion
(Fig. 5a). After this period, the catalysts display stable performance.
As expected, the degree of propyne conversion decreases with de-
creasing temperature (Fig. 5a). The propene selectivity does not ex-
perience a marked dependence with temperature, although a slight
decrease when going from 523 to 373 K is generally observed. The
increase in propene selectivity at the beginning of the reaction was
more pronounced over C773–R773, going from 0 to 50% in the first
2 h and reaching ca. 80% after 5 h on stream. In this period, the
propane selectivity decreased from 50 to <5%. Monzón et al. [30,
55] attributed this behavior to the presence of two types of active
sites in the nickel catalyst: hydrogenolytic (leading to the alkane)
and hydrogenating (leading to the alkene). According to these au-
thors, the former sites experienced fast deactivation by coke in
ethyne hydrogenation, decreasing the alkane production. However,
we cannot exclude the presence of a single type of sites in the
catalyst that is transformed in early stages of the reaction, e.g. by
formation of C-containing deposits on the surface, thereby mini-
Finally, the Ni2 contribution in XPS is shifted to higher binding
energies upon reduction of the sample (856.3 eV in C773–R773
and 855.6 eV in C773), further accentuating the stronger interac-
+
tion of Ni2 and Al
+
3+
in the solid solution upon partial reduction.
This is due to both the higher surface concentration of aluminum
and the lower Ni2 /Al
provoking that the ligand effect of Al
tively lower number of Ni
+
3+
ratio in the oxide upon Ni formation,
0
3
+
is distributed over a rela-
2
+
.
3
3
.2. Hydrogenation performance
.2.1. Calcination and reduction conditions
Fig. 5 shows the propyne hydrogenation performance of the
takovite-derived catalysts calcined and prereduced at different con-
ditions. These tests were run at intervals of 50 K from 523 to

S. Abelló et al. / Journal of Catalysis 259 (2008) 85–95
91
mizing the pathway to the alkane. More details in this respect can
be found in Section 3.3.
The hydrogenation performance over the Ni–Al catalysts is de-
termined by (i) the nature of the oxidic nickel phase (determined
by the calcination conditions) and (ii) the amount of metallic
nickel (induced by the reduction conditions). The propene selectiv-
ity over most catalysts is in the range 60–80%. In general, the calci-
nation temperature has a stronger influence on the alkyne conver-
sion. Calcination at low temperature (623 K) is not appropriate due
to the fact that the degree of required metallic Ni to attain high
and stable propyne conversion is limited. On the other extreme,
high calcination temperatures (1173 K) cause low propyne con-
version due to extensive NiO sintering and formation of NiAl2O4.
The prereduction treatment strongly depends on selecting an ap-
propriate calcination temperature. The C773–R423 sample led to
low propyne conversion (ca. 30% at 523 K) and propene selectiv-
ity (ca. 30% at 473 K) since this sample was virtually unreduced
(Table 1). The lower activity of the latter sample with respect to
(a)
the C623–R623 catalyst is explained by the hindered nickel re-
ducibility attained by the diffusion of aluminum to the surface
when the calcination temperature increases. As expected, increas-
ing the prereduction temperature from 423 to 773 K guarantees
0
a sufficient amount of active Ni sites for hydrogenation. Extend-
ing the prereduction treatment from 30 min to 5 h results into
the highest propyne conversion, ca. 100% in the temperature range
studied, but this is accompanied by decreased propene selectiv-
ity. Overall, the best catalyst was C773–R773, displaying a C3H6
yield of 65%. To evaluate the importance of the synthesis route,
a Ni/Al2O3 catalyst was prepared by incipient wetness followed by
calcination and reduction at 773 K (C773–R773 IM). This sample,
whose characterization is provided in Supplementary information,
showed much lower propyne conversion and propene selectivity
(
C3H6 yield ∼6%).
The dependence of the average values of propyne conversion
and products selectivity on the degree of nickel reduction deter-
mined by H2-TPR is shown in Fig. 6a for the catalysts calcined at
773 K and reduced at different conditions (temperature and time).
(b)
The propyne conversion increases with the percentage of metallic
nickel in the samples, although the specific activity per mol of Ni0
Fig. 6. Conversion of C3H4 and selectivity to C3H6, C3H8, and oligomers (a) and
turnover frequency (b) as a function of the percentage of metallic nickel in differ-
ently reduced C773 catalysts (see Table 1). Data represent averaged values of the
(turnover frequency, TOF) experiences an opposite trend (Fig. 6b).
The mildly reduced C773–R623 sample (6% of nickel reduced) is
ca. 5 times more active than the C773–R773 (55% of nickel re-
duced), although the TOF does not change at higher degrees of
Ni reduction. The propane selectivity increases with the percent-
age of metallic nickel in the samples and the opposite occurs with
the selectivity to oligomers. The volcano-type dependence of the
propene selectivity indicates that an optimal degree of bulk re-
duction (ca. 55% in C773–R773) maximizes the S(C3H6) at 80%.
The maximum in propene selectivity is likely related to the ex-
istence of an optimum between the rate of hydrogenation and the
rate of oligomerization [5]. The relative significance of these two
paths controls the selectivity to the desired product. A low degree
of Ni reduction induces a lower rate of hydrogenation due to the
limited number of sites to dissociate H2. Consequently, the recom-
bination of neighboring intermediate hydrocarbon fragments into
oligomers is favored. Contrarily, at high degrees of nickel reduction,
5
h reaction period at 473 K in Fig. 7. The TOF is expressed as mole C3H4 converted
per second and mole of Ni0.
(sub-)surface Ni-hydride species can be formed, increasing the
alkyne conversion and the selectivity to the over-hydrogenated hy-
drocarbon.
3
.2.2. Hydrogen-to-hydrocarbon ratio
The average values of propyne conversion and products selectiv-
ity over the optimal C773–R773 catalyst using feed H2-to-propyne
ratios ranging from 5 to 1 and 523 K are shown in Fig. 7. The
pattern in this figure qualitatively resembles those obtained with
the different degree of nickel reduction in Fig. 6. The propyne con-
Fig. 7. Conversion of C3H4 and selectivity to C3H6, C3H8, and oligomers vs the
feed hydrogen-to-propyne ratio over C773–R773. Conditions: T = 523 K, WHSV =
−1
−1
h
16,800 ml g
, and P = 1 bar.

92
S. Abelló et al. / Journal of Catalysis 259 (2008) 85–95
version increases on increasing the partial pressure of H2 at the
reactor inlet, since the reaction rate is largely governed by the
high concentration of adsorbed hydrogen atoms and nickel hy-
dride species [8]. The hydrogenation of alkenes and alkynes has
been classically described by the Horiuti–Polanyi mechanism [56],
consisting of the successive addition of atomic hydrogen to the
substrate. The selectivity to propane increases with the H2/C3H4
ratio and the selectivity to oligomers displays the opposite trend.
A high coverage of hydrogen favors total hydrogenation, while low
hydrogen coverage favors oligomerization of the adsorbed hydro-
carbon [57]. At high partial H2 pressure, the eventual formation
of a nickel hydride phase could account for the high C3H8 pro-
duction. Analogously, high H2/HC ratios over palladium catalysts
form the hydrogen-saturated β-hydride phase, which is highly re-
active but unselective [5,19]. In consequence, there is an optimal
hydrogen pressure at which the selectivity to the partially hydro-
genated product is maximized. The propene selectivity over C773–
R773 was ca. 70% at H2/C3H4 = 3 and the selectivity to oligomers
was ca. 30% (i.e. no alkane was produced). Going slightly over or
below this optimal ratio causes a significant decrease of the olefin
selectivity.
(a)
3
.2.3. Sensitivity to the hydrocarbon
We have assessed whether the promising catalytic results using
propyne can be extrapolated to related unsaturated hydrocarbons
such as ethyne and propadiene. The performance of selected cat-
alysts in the hydrogenation of propyne, propadiene, and ethyne is
compared in Fig. 8. Average values of conversion, selectivity, and
yield at 473 K were plotted, since the catalysts exhibited stable
performance at this temperature. The results indicate the supe-
rior activity of the Ni–Al catalysts in the hydrogenation of both
propyne and propadiene as compared to ethyne (Fig. 8a). The high-
est conversions were attained with propadiene, which is due to
the high reactivity of the diene [8]. The conversion of propadi-
ene is in the relatively narrow range of 65–80% despite marked
differences in calcination and reduction of the samples. The se-
lectivity to propene was significantly lower with propadiene than
with propyne (Fig. 8b), due to favored oligomerization and certain
propane formation over the former substrate. It has been previ-
ously suggested that during reaction over nickel catalysts, propa-
diene is converted into C6 products due to the isomerization of
the half-hydrogenated product (2-propenyl), giving a free radical
that can attack a neighboring adsorbed propadiene molecule [8,
(b)
22]. Considering the stronger adsorption of an alkadiene with re-
spect to a monoene, the lower C3H6 selectivity with propadiene
can also be explained taking into account the higher probability of
C–C bond formation if both terminal carbon atoms are susceptible
of being fully hydrogenated. In general, and particularly for C773–
R773, the C3H6 yield was higher using propyne (up to 62%) than
propadiene (up to 47%).
In great contrast to the tests with C3s, ethyne hydrogenation
led to extremely low alkyne conversion (<12%). The ethene se-
lectivity was well measurable only over the C773–R773 catalyst,
amounting to ca. 55%. This leads to a C2H4 yield of 6%, in excellent
agreement with values over precipitated nickel catalysts reported
by Peña et al. [28]. The gap between ethyne and propyne has
been attributed to the extensive oligomerization with the C2 [5,
(
c)
Fig. 8. Hydrocarbon conversion (a), alkene selectivity (b), and alkene yield (c) over
selected catalysts in propyne, propadiene, and ethyne hydrogenation. Conditions:
1
−1
h ,
(
C3H4 or C2H2)/H2/He ratio = 2.5/7.5/90, T = 473 K, WHSV = 16,800 ml g−
and P = 1 bar.
(20%) compared to propyne and propadiene (15%). Interestingly,
only one step of weight loss occurs in all the samples (transi-
tion temperature around 620 K), suggesting a similar nature of
the coke deposits. N2 adsorption revealed a lower surface area
in the used catalysts due to coke residues. The BET surface area
2
−1
8
,12,21,24], causing very fast deactivation by coking and thus low
of the fresh C773–R773 catalyst (155 m g ) markedly decreased
2
−1
ethene yields. The presence of the extra carbon of propadiene
and propyne as compared to ethyne minimizes oligomer forma-
tion, as the steric hindrance prevents adjacent adsorption of the
approaching molecule during the C–C bond forming step [5,8,
to 77 m g
115 m g
after ethyne hydrogenation, while values of 128 and
−1
were obtained after propyne and propadiene hydro-
2
genation, respectively. Accordingly, ethyne should be used with
extreme precaution to generally establish the potential of cata-
lysts for hydrogenation of higher alkynes or dienes. Despite the
marked performance differences among the substrates, the C773–
R773 sample was the best catalyst for the three hydrocarbons in-
vestigated.
21]. The more pronounced coking with ethyne was confirmed by
thermogravimetric analysis of the used catalysts in air (Fig. 9).
The total weight loss due to combustion of the carbonaceous de-
posits was higher in the catalyst subjected to ethyne hydrogenation

S. Abelló et al. / Journal of Catalysis 259 (2008) 85–95
93
3
.3. Activation during reaction
Fig. 5 showed the pronounced increase of propene selectiv-
ity (from 0 to 70%) coupled to the decreased propane selectivity
during the first stages of the reaction at 523 K. Afterwards, the
S(C3H6) is practically independent on the reaction temperature
down to 373 K. The outcome of the test is spectacularly differ-
ent if conducted from the lowest to the highest temperature. As
shown in Fig. 10, the fresh C773–R773 catalyst shows a C3H8 se-
lectivity of ca. 80% at 373 and 423 K, with no C3H6 production. At
473 K and more significantly at 523 K, the selectivity to propene
increases at the expenses of propane, resembling the activation be-
havior in Fig. 5. As expected, once the catalyst is activated under
reaction conditions, the propene selectivity does not change when
going to lower temperature. This result highlights the importance
of the testing procedure for the evaluation of hydrogenation cata-
lysts and indicates that selectivity-enhancing species are formed in
the specific temperature range of 473–523 K.
As previously reported over different metals [5,8,30,55,58], par-
tial blocking of the surface by strongly held oligomers and coke
deposits positively influenced the alkene selectivity during alkyne
hydrogenation. Recently, it has been reported that dissolution of
carbon (derived from fragmentation of alkyne molecules) in in-
terstitial sites of Pd generates a metastable subsurface palladium
carbide that turns the hydrogenation selective [19].
Fig. 9. TGA profiles in air of C773–R773 after hydrogenation of ethyne, propyne, and
propadiene according to the test in Fig. 8.
Further insights into the activation process, displayed over all
substrates investigated here, were offered by operando infrared
spectroscopy over the C773–R773 sample. Fig. 11 shows the evo-
lution of the CH stretching region in infrared during propadiene
hydrogenation at 523 K, which is indicative of the build up of C-
containing deposits in the course of the reaction. Similar spectra
were observed using propyne and acetylene (not shown for con-
−1
ciseness). The bands at 2960 and 2925 cm
are attributed to the
asymmetric CH stretching of CH3 and CH2 groups, respectively [59,
−1
−1
and the shoulder at 2856 cm
60]. The bands at 2871 cm
belong to the symmetric CH stretching of CH3 and CH2 groups,
respectively. These bands were previously assigned to the forma-
tion of a carbonaceous layer on 2 wt% Pd/γ -Al2O3 on exposure to
ethyne [60].
Fig. 12 shows the build up of C-containing deposits (obtained
by integration and normalization of the infrared bands) during
propadiene and ethyne hydrogenation and the corresponding C3H6
Fig. 10. Selectivity to propene and propane vs time during propyne hydro-
genation over C773–R773. Conditions: C3H4/H2/He ratio = 2.5/7.5/90, WHSV =
−1
−1
and C H4 selectivity vs time-on-stream. The results with propyne
16,800 ml g
h
, and P = 1 bar. The temperature was increased stepwise from
2
373 to 523 K and then decreased back to 373 K.
were similar to those with propadiene. In agreement with the
Fig. 11. Operando DRIFT spectra in the CH stretching region during propadiene hydrogenation over C773–R773. Conditions: C3H4/H2/He ratio = 2.5/7.5/90, T = 523 K, WHSV
−1
−1
h
=
16,800 ml g
, and P = 1 bar.

9
4
S. Abelló et al. / Journal of Catalysis 259 (2008) 85–95
NiAl2O4 at high calcination temperature induce lower activity. The
hydrocarbon conversion increases with the percentage of metallic
Ni in the samples, while maximizing the monoalkene selectivity
requires an optimal degree of nickel reduction. Over-hydrogenation
occurs above the optimum, while oligomerization is favored below
the optimum. The same pattern applies with the feed hydrogen-to-
hydrocarbon ratio. Operando infrared spectroscopy gives evidence
of the intimate relation between the increased alkene selectivity
in early stages of the reaction and the formation of C-containing
species on the catalyst (sub-)surface. These selectivity-enhancing
species are formed at specific reaction temperatures, stressing the
relevance of the testing procedure on assessing hydrogenation cat-
alysts. In contrast with the C3s, ethyne hydrogenation led to much
lower conversion and ethene selectivity as well as fast deactiva-
tion due to coking. Consequently, the latter should not be used as
a model molecule to generally establish the potential of catalysts
for hydrogenation of higher hydrocarbons.
Fig. 12. Correlation between the normalized integral of the absorption bands
−1
in the 2800–3100 cm
range during operando DRIFT (lines) and the alkene
selectivity (symbols) during propadiene and ethyne hydrogenation over C773–
Acknowledgments
R773. Conditions: (C3H4 or C2H2)/H2/He ratio = 2.5/7.5/90, T = 523 K, WHSV =
−1
−1
h
16,800 ml g
, and P = 1 bar.
This research was sponsored by the Spanish MEC (CTQ2006-
0
2
1562/PPQ, PTQ05-01-00980, Consolider-Ingenio 2010: grant CSD-
006-00003) and the ICIQ Foundation. D.V. thanks the Erasmus
higher tendency of ethyne to form oligomers (see Section 3.2.3),
the development of the coke-related bands over the catalyst was
remarkably faster compared to propadiene and propyne.
The increase in propene selectivity during propadiene hydro-
genation correlates with the formation of carbonaceous deposits.
On the other hand, the ethene selectivity increases slowly dur-
ing the first two hours of reaction to ca. 40%, exhibiting a less
evident relation with the coke build up determined by infrared
spectroscopy. The formation of coke during ethyne hydrogenation
does not facilitate an increase in alkene selectivity as strong as
with propyne and propadiene. This suggests that the extensive cok-
ing process deactivates a significant fraction of active sites. This
result is supported by the overall higher oligomer selectivity in
this reaction (60% in the reaction with ethyne compared to 40%
for propadiene).
programme for a fellowship. Dr. F. Coloma (Universidad de Ali-
cante) is acknowledged for discussions on the XPS analyses.
Supplementary information
The online version of this article contains additional supple-
mentary information.
Please visit DOI: 10.1016/j.jcat.2008.07.012.
References
[
1] R.J. Farrauto, C.H. Bartholomew, Fundamentals of Industrial Catalytic Processes,
Blackie Academic and Professional, London, 1997, p. 430.
2] F. Cavani, F. Trifirò, Catal. Today 34 (1997) 269.
[
Based on these results, it is plausible that the hydrogenation
behavior of nickel is similar to that of palladium, in which the
population of (sub-)surface sites governs the hydrogenation steps
on the surface [19]. Consequently, nickel hydride-type species favor
total hydrogenation. The latter species are not operative when sur-
face C-containing deposits or even nickel carbide-type species are
formed under certain reaction conditions, boosting the selectivity
to the alkene. Further studies are required to ascertain the struc-
ture and nature of these species remain unclear, requiring further
studies.
[3] H.F. Rase, Handbook of Commercial Catalysts, CRC Press, Boca Raton, 2000,
p. 105.
[4] S. Nishimura, Handbook of Heterogeneous Catalytic Hydrogenation for Organic
Synthesis, John Wiley and Sons, New York, 2001, p. 63.
[
5] Á. Molnár, A. Sárkány, M. Varga, J. Mol. Catal. A 173 (2001) 185.
[6] M.L. Derrien, in: L. Cˇ ervený (Ed.), Catalytic Hydrogenation, in: Stud. Surf. Sci.
Catal., Elsevier, Amsterdam, 1986, p. 613.
[7] Y.J. Ryu, J.R. Adams, W.A. Groten, CA 02488051 (2004).
[8] G.C. Bond, Metal-Catalyzed Reactions of Hydrocarbons, Springer, New York,
2005, p. 395.
[9] N.A. Khan, S. Shaikhutdinov, H.-J. Freund, Catal. Lett. 108 (2006) 159.
[10] H. Lindlar, Helv. Chim. Acta 35 (1952) 446.
11] J. Sheridan, J. Chem. Soc. (1945) 133.
[
[
[
12] G.C. Bond, J. Chem. Soc. (1958) 2705.
13] C. Trinh Dinh, J. Bournonville, US 4504593 (1984).
4
. Conclusions
[
14] H.M. Putman, J.R. Adams, WO 2006/08336 A2 (2006).
Reduced Ni–Al oxides obtained from takovite are highly active,
[15] E.W. Hansen, M. Neurock, J. Catal. 196 (2000) 241.
selective, and stable catalysts for partial hydrogenation of propyne
and propadiene, without addition of promoters. The intrinsic fea-
tures of the hydrotalcite route in terms of attaining mixed oxides
with high metal loadings, high metal (inter)dispersion, and high
surface area are beneficial for the performance, which largely ex-
ceeds that of Ni/Al2O3 prepared by conventional impregnation.
Proper selection of activation conditions of the takovite precursor
[16] E.W. Hansen, M. Neurock, J. Phys. Chem. B 105 (2001) 9218.
[
17] D. Teschner, E. Vass, M. Hävecker, S. Zafeiratos, P. Schnörch, H. Sauer, A. Knop-
Gericke, R. Schlögl, M. Chamam, A. Wootsch, A.S. Canning, J.J. Gamman, S.D.
Jackson, J. McGregor, L.F. Gladden, J. Catal. 242 (2006) 26.
[18] W. Huang, W. Pyrz, R.F. Lobo, J.G. Chen, Appl. Catal. A 333 (2007) 254.
[19] D. Teschner, J. Borsodi, A. Wootsch, Z. Révay, M. Hävecker, A. Knop-Gericke,
S.D. Jackson, R. Schlögl, Science 320 (2008) 86.
[20] J. Sheridan, J. Chem. Soc. (1944) 373.
[21] G.C. Bond, J. Sheridan, Trans. Faraday Soc. 48 (1952) 651.
(calcination and reduction of the derived mixed oxide) is decisive
[22] G.C. Bond, J. Sheridan, Trans. Faraday Soc. 48 (1952) 658.
for the catalytic behavior, as they determine the nature of the ox-
ide (crystalline phase structure, metal distribution, dispersion, and
morphology) and the degree of nickel reduction. The best cata-
lyst, calcined and reduced at 773 K, rendered stable propene yields
up to ca. 65% and consisted of a Ni(Al)Ox solid solution with 55%
of the total bulk nickel in reduced form and surface enrichment
by aluminum. Extensive sintering of NiO and crystallization of
[
23] G.C. Bond, G. Webb, P.B. Wells, J.M. Winterbottom, J. Catal. 1 (1962) 74.
[24] R.S. Mann, K.C. Khulbe, J. Catal. 13 (1969) 25.
[25] J.J. Phillipson, P.B. Wells, G.R. Wilson, J. Chem. Soc. A (1969) 1351.
[
[
[
26] P.B. Wells, G.R. Wilson, J. Chem. Soc. (1970) 2442.
27] Y. Segura, N. López, J. Pérez-Ramírez, J. Catal. 247 (2007) 383.
28] J.A. Peña, J. Herguido, C. Guimon, A. Monzón, J. Santamaría, J. Catal. 159 (1996)
13.
[29] J.C. Rodríguez, A.J. Marchi, A. Borgna, A. Monzón, J. Catal. 171 (1997) 268.
3

S. Abelló et al. / Journal of Catalysis 259 (2008) 85–95
95
[
30] A. Monzón, E. Romeo, C. Royo, R. Trujillano, F.M. Labajos, V. Rives, Appl. Catal.
[45] F.-W. Chang, M.-S. Kuo, M.-T.T.-C. Hsieh, J. Chem. Tech. Biotechnol. 79 (2004)
691.
A 185 (1999) 53.
[
31] R.G. Oliver, P.B. Wells, J. Catal. 47 (1977) 364.
[46] K.V.R. Chary, P.V.R. Rao, V. Vishwanathan, Catal. Commun. 7 (2006) 974.
[47] P. Arnoldy, J.A. Moulijn, J. Catal. 93 (1985) 38.
[48] D. Tichit, F. Medina, B. Coq, R. Dutartre, Appl. Catal. A 159 (1997) 241.
[49] F. Basile, L. Basini, M.D. Amore, G. Fornasari, A. Guarinoni, D. Matteuzzi, G.D.
Piero, F. Trifirò, A. Vaccari, J. Catal. 173 (1998) 247.
[
[
[
[
[
32] Y. Okamoto, K. Fukino, T. Imanaka, S. Teranishi, J. Catal. 74 (1982) 173.
33] A.F. Benton, P.H. Emmett, J. Am. Chem. Soc. 46 (1924) 2728.
34] F. Cavani, F. Trifirò, A. Vaccari, Catal. Today 11 (1991) 173.
35] M. Bellotto, B. Rebours, O. Clause, J. Lynch, J. Phys. Chem. B 100 (1996) 8527.
36] J. Pérez-Ramírez, G. Mul, F. Kapteijn, J.A. Moulijn, J. Mater. Chem. 11 (2001) 821.
[50] L. Zhang, J. Zhu, X. Jiang, D.G. Evans, F. Li, J. Phys. Chem. Solids 67 (2006) 1678.
[51] A.N. Mansour, Surf. Sci. Spectra 3 (1996) 239.
[37] S. Abelló, J. Pérez-Ramírez, Adv. Mater. 18 (2006) 2436.
[
38] D. Briggs, M.P. Seah (Eds.), Practical Surface Analysis, vol. 1, Auger and X-Ray
Photoelectron Spectroscopy, John Wiley and Sons, Chichester, 1994.
39] O. Clause, B. Rebours, E. Merlen, F. Trifiró, A. Vaccari, J. Catal. 133 (1992) 231.
40] L.E. Alzamora, J.H. Ross, E.C. Kruissink, L.L. van Reijen, J. Chem. Soc. Faraday
Trans. 77 (1981) 665.
[52] L. Salvati, L.E. Makovsky, J.M. Stencel, F.R. Brown, D.M. Hercules, J. Phys.
Chem. 85 (1981) 3700.
[53] B. Rebours, J.-B. d’Espinose de la Caillerie, O. Clause, J. Am. Chem. Soc. 116
(1994) 1707.
[
[
[54] F. van Looij, J.W. Geus, J. Catal. 168 (1997) 154.
[
41] O. Clause, M.G. Coelho, M. Gazzano, D. Matteuzzi, F. Trifirò, A. Vaccari, Appl.
Clay Sci. 8 (1993) 169.
[55] C. Guimon, A. Auroux, E. Romeo, A. Monzón, Appl. Catal. A 251 (2003) 199.
[56] J. Horiuti, M. Polanyi, Trans. Faraday Soc. 30 (1934) 1164.
[57] J.A. Peña, A. Monzón, J. Santamaría, J. Catal. 142 (1993) 59.
[58] S.D. Jackson, N.J. Casey, J. Chem. Soc. Faraday Trans. 91 (1995) 3269.
[59] N. Sheppard, J.W. Ward, J. Catal. 15 (1969) 50.
[
42] D.C. Puxley, I.J. Kitchener, C. Komodromos, N.D. Parkyns, in: G. Poncelet, P.
Grange, P.A. Jacobs (Eds.), Preparation of Catalysts III, in: Stud. Surf. Sci. Catal.,
Elsevier Science, 1983, p. 237.
[
43] R.K. Datta, R. Roy, Am. Mineral. 53 (1968) 1456.
[60] J. Wood, M.J. Alldrick, J.M. Winterbottom, E.H. Stitt, S. Bailey, Catal. Today 128
(2007) 52.
[44] J. Pérez-Ramírez, S. Abelló, N.M. van der Pers, J. Phys. Chem. C 111 (2007) 3642.