Hydrogen production by reforming of acetic acid using La-Ni type perovskites partially substituted with Sm and Pr

Article abstract of DOI:10.1016/j.cattod.2014.07.013

The condensation of gases from the pyrolysis of biomass leads to a liquid compound called bio-oil. This oil can be divided into two fractions: one aqueous and another non-aqueous. The aqueous fraction does not have a high market value and it is usually di

Full text of DOI:10.1016/j.cattod.2014.07.013

Catalysis Today 242 (2015) 71–79
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Hydrogen production by reforming of acetic acid using La–Ni type
perovskites partially substituted with Sm and Pr
K.A. Resendea, C.N. Ávila-Netoa, R.C. Rabelo-Netob, F.B. Noronhab, C.E. Horia,∗
a Federal University of Uberlândia, School of Chemical Engineering, Av. João Naves de Ávila, 2121, Bloco 1 K, 38408-144, Uberlândia, MG, Brazil
b National Institute of Technology, Division of Catalysis and Chemical Processes, Av. Venezuela, 82, Centro, 20081-312 Rio de Janeiro, RJ, Brazil
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online 22 August 2014
The condensation of gases from the pyrolysis of biomass leads to a liquid compound called bio-oil. This oil
can be divided into two fractions: one aqueous and another non-aqueous. The aqueous fraction does not
have a high market value and it is usually discarded. However, this mixture has considerable amounts of
organic compounds, which can be potential renewable sources of hydrogen. Steam reforming has been
the most studied reaction to produce hydrogen from the aqueous fraction of bio-oil. Due to the diversity
of organic compounds found in this aqueous fraction, model compounds have usually been used to study
this mixture. Among the compounds used to represent the aqueous fraction of bio-oil, acetic acid is the
most popular. Nickel-based catalysts are traditionally used for reforming reactions. An alternative to
combine high Ni content employed in catalysts, this kind of process with desirable values of dispersion
is to use perovskite-type precursors (LaNiO3). Therefore, the objective of this study was to investigate
the production of hydrogen from the aqueous fraction of bio-oil. More specifically, in this work, LaNiO3,
LaPrNiO3 and LaSmNiO3 were tested as precursors for catalysts in the reaction of steam and oxidative
reforming of acetic acid. During the reaction, the catalysts showed the formation of the same products: H2,
CO, CH4, CO2, C3H6O, but with different selectivities. All samples presented good selectivity for hydrogen
formation, and the presence of Pr and Sm just slightly affected the catalytic performance. Due to the large
accumulations of coke observed during the reaction of steam reforming, a small amount of O2 was added
to the mixture. The presence of this oxidant improved catalytic activity without reducing the amount
of hydrogen produced. Furthermore, it helped to reduce the deposits of coke, maintaining the catalytic
activity during 24 h of reaction.
Keywords:
Bio-oil aqueous fraction
Perovskite
Nickel
Acetic acid and hydrogen
© 2014 Elsevier B.V. All rights reserved.
1. Introduction
Pyrolysis consists on a thermal decomposition of biomass in
the absence of oxygen or water vapor [5]. During this process, a
Hydrogen is an important alternative for energy generation
and it presents several advantages, such as a high specific energy
(∼120 MJ kg−1), formation of water as the only product during the
combustion process, among others [1]. Nowadays, hydrogen pro-
duction occurs mainly through reforming of natural gas. However,
the goal of sustainable development requires the use of renewable
sources, such as biomass or water electrolysis [2]. Biomass is the
renewable energy source with greater availability worldwide and
it has a high-energy potential. According to Huber and Dumesic
[3], it is important to develop technologies that will make the use
of biomass economically viable. Gasification and pyrolysis are the
most used processes for the energetic use of biomass [4].
dark colored liquid called bio-oil is formed. It is a complex mixture
with different types of organic compounds, such as carboxylic acids,
ketones, aldehydes, sugars, alcohols, phenols, water and more com-
plex compounds [6,7]. When water is added to bio-oil, it separates
into two fractions. One of the fractions is non-aqueous and con-
tains lignin derivatives. This fraction has characteristics similar to
oils obtained from fossil fuels, which gives it a high economic value
[8]. The other fraction is aqueous and contains acids, ketones, alde-
hydes and others compounds in low concentrations. Therefore, this
aqueous phase does not have many applications and it is discarded
in most cases. The catalytic reforming of the aqueous fraction of
bio-oil to obtain hydrogen could be an interesting way to increase
the production and use of bio-oil [9]. According to Yan et al. [10],
this strategy has low costs and it is trouble-free. The aqueous frac-
tion of bio-oil usually has around 80 wt% of water, which could feed
the demand for water in the steam reforming process.
∗
Corresponding author. Tel.: +55 34 3239 4292; fax: +55 34 3239 4249.
E-mail addresses: cehori@ufu.br, cehori@gmail.com (C.E. Hori).
http://dx.doi.org/10.1016/j.cattod.2014.07.013
0920-5861/© 2014 Elsevier B.V. All rights reserved.

72
K.A. Resende et al. / Catalysis Today 242 (2015) 71–79
Steam reforming is the most reported reaction used to produce
2. Experimental
H2 from the aqueous fraction of bio-oil [11]. Many studies report
the production of H2 from steam reforming of the aqueous fraction
extracted from different types of bio-oils [10,12,13]. Acetic acid
is the most abundant organic compound in the aqueous fraction
of various types of bio-oil [14], and because of this, it is usually
reported as a representative molecule of this fraction. However,
other model molecules such as acetol [15], ethylene glycol [16],
acetone [17], among others [18], are also studied.
The major problem of steam reforming of the aqueous fraction
of the bio-oil is the formation of large amounts of coke, which
deposits on the catalyst surface. This accumulation contributes to
the deactivation of the catalyst and it also causes a decrease in H2
selectivity [19–23]. According to Hu and Lu [24], Ni has a signifi-
cant catalytic activity for this reaction and it is economically feasible
[16,25].
2.1. Catalyst preparation
The perovskite precursors were synthesized by a precipita-
tion method adapted from de Lima et al. [37]. The first sample,
named LaNiO3, was prepared with equimolar amounts of Ni and
La (Ni/La = 1). In the two other samples, La atoms were partially
replaced by Pr or Sm. These samples are named LaPrNiO3 and
LaSmNiO3, where 0.1 is the fraction of replaced La atoms. To prepare
the LaNiO3 precursor, an aqueous solution of Na2CO3 (0.5 M)
was added rapidly and under vigorous stirring to a freshly pre-
pared aqueous solution of La(NO3)3·6H2O and Ni(NO3)2·6H2O until
reaching pH 8. In the case of the LaPrNiO3 and LaSmNiO3 precursors,
stoichiometric amounts of Pr(NO3)2·6H2O and Sm(NO3)3·6H2O
salts were also added to the above solution. Following, the pre-
cipitate was washed and filtered under vacuum for removing
remaining contaminant ions. The washed solid was then dried in
a furnace at 338 K for 20 h and calcined under flow of synthetic
air (50 mL min−1) in two stages: first at 723 K (heating rate of
Fatsikostas et al. [27] showed that Ni-based catalysts (17 wt%)
supported on La2O3 exhibited high activity and selectivity to hydro-
gen formation in steam reforming of ethanol. According to the
authors, the presence of La improved the catalyst performance.
When the catalyst promoted with La is subjected to reducing con-
ditions, this metal forms a thin LaOx layer on the surface of Ni
particles. This layer can react with CO2 to form lanthanum oxy-
carbonate (La2O2CO3), which combines with the deposited carbon
and cleans the Ni surface. The concentration of the active metal
also affects the rate of carbon accumulation. Li et al. [22] tested Ni-
based catalysts supported on ZrO2 with Ni content varying from 0
to 23% in steam reforming of acetic acid. They noticed an appre-
ciable increase in Ni particle size for higher concentrations of Ni,
which affected its dispersion. Furthermore, it was also shown that
the presence of large Ni crystallites promotes the formation of car-
bon [22,28]. Thus, the balance between the rates of formation and
oxidation of carbon deposits may be achieved by decreasing the
size of metal crystallites.
One way to combine both high loadings of Ni and high dispersion
is to obtain a catalyst from a method known as solid-phase crys-
tallization [29]. In this method, the active phase (Ni) is extracted
from a well-defined solid structure when it is submitted to a treat-
ment in a reducing atmosphere. Perovskites (ABO3) are examples
of solid structures used for this purpose [30–34]. In addition, they
have the ability to allow the substitution of cations in both pos-
itions A and B by other similar sizes elements, A’ and B’, which can
lead to better catalytic properties. Gallego et al. [35] studied the use
of perovskites La1−xAxNiO3–ı (A = Pr and Ce, x = 0–0.06) as precur-
sors of active catalysts for dry reforming of methane. The authors
observed that the addition of Ce and Pr to the perovskite-type struc-
ture, LaNiO3, decreased the Ni crystallite size. It also contributed
to improve the performance of the samples during the reaction. It
was shown that the rate of carbon accumulation on the catalyst
surface decreased as the amount of dopant (Ce and Pr) increased.
These observations were probably due to two factors: (i) the parti-
cle size of metallic Ni present in the doped catalysts decreased and
(ii) Ce and Pr have redox properties. Jahangiri et al. [36] also studied
a series of perovskite precursors doped with Sm (La1−xSmxNiO3,
x = 0–1). The results showed that the preparation method led to
the formation of a homogeneous and crystalline structure with a
small amount of Sm. The high activity of the catalyst doped with
high amounts of Sm was justified by its catalytic redox proper-
ties.
1.5 K min−1) for 4 h and finally at 1173 K (heating rate of 5 K min−1
for 6 h.
)
2.2. Characterization
X-ray powder diffraction (XRD) analyses were carried out (i)
in conventional mode at room temperature and (ii) during in situ
heating of the samples under a flow of H2 in He (5 vol%).
In conventional mode, XRD patterns of the calcined and
reduced/passivated samples were obtained with a Rigaku equip-
˚
ment, using Cu K␣ radiation (1.54056 A) with a Ni filter. The 2ꢀ
angle was swept from 10◦ to 80◦, with a step size of 0.02◦ and a
counting time of 2 s per step. The Scherrer equation was used to
estimate the crystallite mean diameter of NiO and Ni0 particles.
For the measurements of the crystallite mean diameter of metal-
lic Ni particles, the calcined samples were also reduced under pure
hydrogen (30 mL min−1) at 773 K for 1 h, purged under N2 at the
same temperature for 30 min and cooled to 298 K. Then, the reac-
tor was maintained at 203 K for 1 h, the catalyst was passivated
with a 5% O2/N2 mixture.
In situ XRD was carried out at the XPD-10B beamline of the
Brazilian Synchrotron Light Laboratory (LNLS). The samples were
first crushed, sieved to particle sizes smaller than 20 ␮m and homo-
geneously distributed over the support. XRD patterns were then
acquired during heating of the samples from room temperature
to 973 K, at 5 K min−1, with a holding time of 30 min, under a
200 mL min−1 flow of H2/He (5 vol%). Scans were carried out from
20◦ to 55◦, with a step size of 0.003◦ and a counting time of 1 s,
˚
using a wavelength of 1.65121 A and a resolution of 4.3 eV.
Temperature-programmed reduction (TPR) measurements
were conducted using a quadrupole mass spectrometer (OmniStar-
Balzers). The samples (50 mg) were reduced under a 2% H2/Ar
mixture (30 mL min−1) at a heating rate of10 K min−1 from 298 to
1273 K.
Temperature-programmed desorption (TPD) of acetic acid
experiments were performed in the same apparatus previously
described for TPR. In a typical TPD experiment, 100 mg of the cat-
alyst was reduced under pure H2 (30 mL min−1) at a heating rate
of 10 K min−1 up to 973 K for 2 h. After activation, the H2 flow was
shifted to Ar flow (30 mL min−1) and the samples were cooled to
room temperature. Adsorption of acetic acid was carried out at
390 K for 30 min by flowing argon (30 mL min−1) through a satu-
rator filled with acetic acid kept at 293 K. Then the sample was
purged with Ar flow for 30 min, and the temperature was lin-
early increased with a heating rate of 10 K min−1 up to 1023 K.
In order to improve the technology of obtaining H2 from
biomass, the objective of this work is to study the production of H2
from the aqueous fraction of bio-oil through reforming processes
using Ni-based catalysts. Activity and stability tests during steam
(SR) and oxidative reforming of acetic acid (OSR) were conducted
using catalysts derived from a La–Ni perovskite-type structure
doped with Pr and Sm.

K.A. Resende et al. / Catalysis Today 242 (2015) 71–79
73
The effluent composition was monitored by a quadrupole mass
spectrometer (OmniStarBalzers).
(A)
LaPrNiO3
Thermogravimetric analysis of the used catalysts was car-
ried out in a TA Instruments equipment (SDT Q600) in order to
determine the amount of carbon formed over the catalyst. Approx-
imately 12 mg of spent catalyst was heated under air flow from
room temperature to 1273 K at a heating rate of 20 K min−1 while
the weight change was measured. Other characterizations of the
used catalyst were not performed due to the small mass available.
LaSmNiO3
LaNiO3
NiO
2.3. Activity tests
Prior to the reactions, samples (10 mg of catalyst diluted
with 150 mg of SiC) were reduced under pure hydrogen flow
(30 mL min−1) at 973 K for 1 h. Steam reforming (SR) of acetic acid
was carried out in a fixed-bed tubular reactor (Hastelloy X) at 873 K
and atmospheric pressure using a Microactivity Reference equip-
ment (PID Eng & Tech.). The reactant mixture (water/acetic acid
molar ratio of 3) was injected with a piston pump (Gilson) and
vaporized in a He stream. The composition of the reactant mixture
was 13% acetic acid, 39% water, 48% helium, with the total flow
rate of 400 mL min−1. The reaction products were analyzed on line
by gas chromatography (Agilent 7890A) equipped with a thermal
conductivity detector (TCD) detector and a Porapark Q column.
Oxidative steam reforming (OSR) of acetic acid was performed
in a fixed-bed quartz reactor at 873 K and atmospheric pressure.
The water and acetic acid were fed using two parallel saturators
with Ar streams (100 mL min−1 each). The saturated streams were
mixed with O2 stream (5 mL min−1 each). The total feed consisted
of a 205 mL min−1 stream with 11% acetic acid, 13% steam, 2% O2
and 74% Ar. The effluent was analyzed by online gas chromatog-
raphy employing a Shimadzu gas chromatograph equipped with
a thermal conductivity detector (TCD), a flame ionization detector
(FID) and two columns: Carboxen 1010 Plot and Rtx®-5.
La4Ni3O10
LaNiO3- Standard
20 25 30 35 40 45 50 55 60 65 70 75 80
2
( ° )
(B)
LaSmNiO3
LaPrNiO3
LaNiO3
Ni0
La2O3
The molar fraction in dry basis of species i (yi) and the conver-
sion of acetic acid (XAc) for both steam reforming and oxidative
reforming were calculated by Eqs. (1) and (2), respectively.
20 25 30 35 40 45 50 55 60 65 70 75 80
(°)
2
Fi
yi (%) =
× 100
(1)
ꢀ
N
Fig. 1. XRD patterns of (A) calcined and (B) reduced and passivated LaNiO3, LaPrNiO3
and LaSmNiO3 samples. XRD patterns of NiO, La4Ni3O10, LaNiO3, Ni0 and La2O3
taken from the Inorganic Crystal Structure Database (collection codes 24014, 80279,
67714, 43397 and 24693, respectively).
j=1Fj
FAc,in − FAc
XAc (%) =
× 100
(2)
FAc,in
In Eqs. (1) and (2), Fi and Fj are the molar flows of species i
and j in the reactor outlet stream (H2, CH4, CO, CO2 and acetone),
respectively, and FAc,in is the molar flow of acetic acid in the inlet
stream.
indicates that these metals were incorporated in the perovskite-
type structure, at least partially. Similar results were reported in
the literature [35,36]. The three samples exhibit similar average
crystallite sizes, with values close to 20 nm. This value agrees with
previous results [39] and shows that the addition of Pr and Sm do
not alter significantly the average crystallite sizes of the precursors.
XRD patterns of the reduced and passivated samples are dis-
played in Fig. 1B. The diffractograms revealed the presence of lines
corresponding to La2O3 (26.1◦, 29.1◦, 29.9◦, 39.64◦, 46◦, 52◦, 55.6◦
and 56.1◦) and Ni0 (44.5◦ and 52◦) phases. These results indicate
that both perovskite-type structures, LaNiO3 and La4Ni3O10, were
destroyed during the reduction treatment, giving rise to Ni0 parti-
cles deposited over lanthanum oxide. Table 1 shows the size of Ni0
crystallites estimated by applying the Scherrer equation. The addi-
tion of Pr and Sm to the LaNiO3perovskite-type structure caused a
slight decrease in the size of Ni0 crystallites from 21 nm (LaNiO3)
to 18 nm (LaPrNiO3 and LaSmNiO3)[35,40–44].
3. Results
3.1. Catalyst characterization
Fig. 1A shows the XRD patterns of the calcined samples. All sam-
ples show diffraction lines around 23.35◦, 47.27◦ and 58.64◦. These
lines correspond to the lattice planes ofLaNiO3, where Ni atoms
have valence 3+, indicating that the calcination conditions were
sufficient to form the perovskite-type structure [33,38]. The sam-
ples also show a triple line around 32.3◦, 32.8◦ and 33◦, which may
be attributed to another perovskite-type phase, the La4Ni3O10. In
this type of structure, the Ni valence is close to 2.67+. Diffraction
lines around 37.3◦, 43.3◦ and 63.1◦ were also identified and may be
attributed to lattice planes (1 1 1), (2 0 0) and (2 2 0) of NiO. Since
the syntheses were performed with equimolar amounts of Ni and
La, the former results indicate that a small amount of Ni did not
enter the perovskite-type structure. Diffraction lines corresponding
to praseodymium or samarium oxides were not identified, which
Fig. 2 shows the temperature-resolved XRD patterns obtained
for the calcined LaNiO3 sample heated in a H2/He mixture (5 vol%).
At 473 K, the diffractograms reveal the presence of lines corre-
sponding to LaNiO3 (23.35◦ and 47.27◦), La4Ni3O10 (32.3◦) and NiO
(37.3◦ and 43.3◦), which were also observed in Fig. 1A. Between

74
K.A. Resende et al. / Catalysis Today 242 (2015) 71–79
Fig. 2. Temperature-resolved XRD patterns in (A) two dimensions (2D) with 2ꢀ = 20–55◦ and in (B) three dimensions (3D) with 2ꢀ = 28–35◦. Heating rate: 5 K min−1; carrier
flow: H2/He (5 vol%). XRD patterns of NiO, La4Ni3O10, LaNiO3, Ni0, La2O3 taken from the Inorganic Crystal Structure Database (collection codes 24014, 80279, 67714, 1917
and 24693, respectively) and La2Ni2O5 taken from Crespin et al. [45].
473 and 623 K, the structures seem to reorganize and lose crys-
tallinity. The gradual increase in the width at half height of the lines
around 32.3◦, 32.8◦, 33◦ and 47.27◦ illustrates this fact. According
to Crespin et al. [44,45], the calcined sample suffers a deviation
from the oxygen stoichiometry when it is heated, which causes a
disorder in the perovskite-type structure. A third perovskite phase
(La2Ni2O5), with lines around 32.36◦ and 46.48◦, appears at 673 K
and seems to be stable until 723 K. For higher temperatures, the
La2Ni2O5 perovskite-type structure is destroyed and gives rise to
Ni0 particles deposited over lanthanum oxide, as already shown in
Fig. 1B. Indeed, La2O3 (26.1◦, 29.1◦, 29.9◦, 39.64◦, 46◦ and 52◦) and
Ni0 (44.5◦ and 52◦) are the unique phases observed for tempera-
tures higher than 873 K. This is an evidence that the perovskite-type
structure La4Ni3O10 follows the same reduction mechanism of
LaNiO3.
used in this study (2% H2/N2). According to the literature [37,40],
the reduction peak at low temperature is assigned to the reduction
of Ni3+ to Ni2+. Indeed, Fig. 2A shows that the reduction of Ni3+ to
Ni2+ starts around 600–700 K, which was also observed by Rivas
et al. 44]. The transition of LaNiO3 to La2Ni2O5 also occurred in this
temperature range (see Eq. (3) 44,45]). The peak at high tempera-
ture is attributed to the reduction of Ni2+ to Ni0 according to Eq. (4)
[29], which is also in agreement with the in situ XRD data.
2LaNiO3 + H2 → La2Ni2O5 + H2O
(3)
913 K
689 K
Fig. 3 presents the TPR profiles of LaNiO3, LaPrNiO3 and
LaSmNiO3 samples. The reduction profile of all perovskite-type
oxides exhibited two peaks at 689–702 K and 860–913 K. These
peaks are shifted to higher temperatures if compared to those
reported in the literature [31,40] and to those observed in the in situ
XRD analysis. This is likely due to the lower hydrogen concentration
900 K
LaSmNiO3
LaPrNiO3
690 K
702 K
860 K
Table 1
Average sizes of Ni0 crystallites on the reduced and passivated samples.
Sample
Sizea (nm)
LaNiO3
473
LaNiO3
LaPrNiO3
LaSmNiO3
21
18
18
373
573
673
773
873
973
1073
Temperature (K)
a
Determined by applying the Scherrer equation to the peak corresponding to
lattice plane (1 1 1) of Ni0.
Fig. 3. TPR profiles of LaNiO3, LaPrNiO3 and LaSmNiO3 samples.

K.A. Resende et al. / Catalysis Today 242 (2015) 71–79
75
Table 2
323
423
523
623
723
823
923
1023
H2 consumption during TPR analyses.
Sample
Theoretical
H2 consumed
Degree of
reduction
(%)
(2nd
PA)/(1st
PA)
H2 m/q=2 (x1)
−1
a
consumption
(␮mol gcat
)
−1
(␮mol gcat
)
LaNiO3
LaPrNiO3
LaSmNiO3
6108
6103
6080
5940
4172
3686
97
71
65
2.70
2.89
2.90
CH4 m/q=16 (x15)
CO m/q=28 (x2)
a
Sample weight used in TPR: 50 mg.
La2Ni2O5 + 2H2 → La2O3 + 2Ni0 + 2H2O
(4)
The addition of Pr and Sm to LaNiO3 perovskite-type oxide
changed the TPR profiles. The first peak was slightly shifted to
lower temperatures. This could be an indication of the incorpora-
tion of the dopants into the lattice structure of LaNiO3 [31,32]. On
the other hand, the second peaks were shifted to higher tempera-
tures and their intensity increased. These results suggest a higher
fraction of NiO segregated from the Perovskite-type structure on
the substituted perovskites. De Lima et al. [37] reported the pres-
ence of three peaks in the TPR profile of LaNiO3. The intermediate
peak was attributed to the reduction of Ni2+ to Ni0 of a NiO phase.
Taking into account the stoichiometry of reduction from Eqs. (3)
and (4), the hydrogen consumption in the second peak should be
twice of the one in the first peak, if the perovskite-type oxide is
the only phase present. However, the segregation of NiO from the
perovskite-type structure during calcination would lead to a higher
hydrogen uptake in the second peak. The presence of Pr and Sm
caused a displacement of the second peak to higher temperatures,
as well as an increase in its area. These results may indicate an
increase in the amount of nickel (Ni2+) outside the perovskite-type
structure. Table 2 lists the total hydrogen uptake and the intensity
ratio of the second and first reduction peaks. This ratio increased
with the addition of Pr and Sm, which indicates an increase in the
amount of nickel oxide segregated from the perovskite-type struc-
ture. The estimated values of the degree of reduction (D.R.) are also
presented in this table. LaNiO3 sample showed a D.R. of 97%. This
result indicates that most of the nickel was incorporated into the
perovskite-type structure. The D.R. was lower for doped samples,
following the order: LaNiO3 > LaPrNiO3 > LaSmNiO3. The decrease
in D.R. for the doped samples could be associated with the increase
in the amount of nickel outside the perovskite-type structure, when
Pr and Sm were added. This is likely due to the stoichiometry used
for the calculation of D.R. that considers the reduction of Ni3+ to
Ni0. The values of D.R. should be smaller when nickel Ni2+ from
segregated NiO phase is reduced to Ni0.
Acetone + Acetic acid
m/q=43 (x5)
CO2 m/q=44 (x2)
323
423
523
623
723
823
923
1023
Temperature (K)
Fig. 4. TPD profiles after reduction of the samples and subsequent saturation with
acetic acid. (Black solid line) LaNiO3, (red dash line) LaPrNiO3, (blue dash-dot line)
LaSmNiO3.
CH3COOH → CH2CO + H2O
2CH2CO → C2H4 + 2CO
(8)
(9)
There is also a peak corresponding to the formation of CO at
900 K on the TPD profile of all catalysts. This CO is likely a reac-
tion product of the oxidation of carbon deposited on the catalyst
surface during the decomposition of acetic acid. The oxygen used
for this reaction should come from the structure of La2O3 [37]. The
presence of La leads to the formation of mixed phases with cationic
vacancies which enhance the mobility of lattice oxygen anions. It
may favor the surface reaction between carbon-containing species
and reactive oxygen species [46].
Similar TPD profiles were reported by Basagiannis and Verykios
[26] using Ni/La2O3/Al2O3 catalysts. They suggested that the TPD
profiles stem from the decomposition reactions of acetic acid.
The addition of Pr and Sm did not change significantly the TPD
profiles, indicating that the reaction mechanism was not altered.
The only significant difference was the decrease in the intensity
of the H2 signal of LaPrNiO3 and LaSmNiO3 at 650 K. This result
suggests that the Pr and Sm addition minimized the formation of
H2 via decomposition of acetic acid (Eq. (5)).
TPD profiles obtained for all catalysts are shown in Fig. 4. All
catalysts showed similar desorption profiles and the following
compounds were formed: H2, CO, CO2, CH4, CH3COCH3, C2H4 and
CH3COOH. At low temperatures (300–400 K), the desorption of a
small amount of acetic acid is noticed. In the temperature range of
500–700 K, it is observed two peaks for the signals corresponding
to H2, CO, CO2, CO and CH4 that follows identical patterns. This is
likely due to the occurrence of acetic acid decomposition reaction
(Eqs. (5) and (6)) [26].
CH3COOH → 2CO + 2H2
CH3COOH → 2CH4 + 2CO2
(5)
(6)
3.2. Catalytic tests
It is also detected the formation of small amounts of CH3COCH3
and C2H4. These products were also observed by Basagiannis and
Verykios [26]. They proposed that acetone is produced by the
ketonization reaction (Eq. (7)). As far as ethylene production is
concerned, the formation of a ketene (CH2CO) intermediate was
proposed according to reactions (8) and (9).
3.2.1. SR of acetic acid
Acetic acid conversion and product distribution as a function of
time on stream (TOS) for SR of acetic acid at 873 K over LaNiO3,
LaPrNiO3 and LaSmNiO3 catalysts are shown in Fig. 5. The follow-
ing initial conversions were observed: LaNiO3-82%, LaSmNiO3-98%
and LaPrNiO3-90%. The main products formed were H2 and CO2,
2CH3COOH → (CH3)2CO + H2O + CO2
(7)

76
K.A. Resende et al. / Catalysis Today 242 (2015) 71–79
Table 3
SR (873K)
Rates of carbon accumulation under SR and OSR at 873 K.
(A)
Sample
rcarbon (mgcarbon gcat−1 h−1
)
50
40
30
20
10
0
100
80
60
40
20
0
SR (873 K)
OSR (873 K)
LaNiO3
LaPrNiO3
LaSmNiO3
16.6
14.4
26.9
2.16
3.65
3.78
indicating that SR of acetic acid is the main reaction taking place.
There is also the formation of small amounts of CO and acetone.
These results are in agreement with the transient experiments that
revealed the occurrence of acetic acid decomposition reaction (Eqs.
(5) and (6)) and ketonization reaction (Eq. (7)).
3.2.2. OSR of acetic acid
Fig. 6 shows the acetic acid conversion and product distribution
as a function of time on stream (TOS) for OSR of acetic acid at 873 K
over LaNiO3, LaPrNiO3 and LaSmNiO3 catalysts. The initial acetic
acid conversion was approximately the same for all catalysts. H2,
CO2, CO and small amounts of CH4 were formed during the OSR of
acetic acid.
0
4
8
12
16
20
24
Time on stream (h)
(B)
3.3. Characterization of post-reaction Ni-supported catalysts
50
100
80
60
40
20
0
Acetic acid conversions significantly decreased at the beginning
of the SR reaction tests for all catalysts. However, the addition of
oxygen to the feed significantly promoted the stability of the cata-
lysts. In order to determine the main cause of catalyst deactivation,
the used Ni-supported catalysts were characterized by TGA analy-
sis. TPO profiles of LaNiO3, LaPrNiO3 and LaSmNiO3 catalysts after
24 h of SR and OSR at 873 K are presented in Fig. 7A and B, respec-
tively. For SR, all catalysts showed one broad peak with maxima in
the range of 879–893 K, while for OSR, the peak maximum temper-
atures were slightly lower, around 823–828 K.
40
30
20
10
0
Table 3 lists the rate of carbon accumulation on the catalyst
surface after SR and OSR of acetic acid for 24 h TOS. For SR of
acetic acid, the amount of carbon formed followed the order:
LaPrNiO3 < LaNiO3 < LaSmNiO3.
0
4
8
12
16
20
24
Time on stream (h)
4. Discussion
(C)
In the present work, the catalytic activities and selectivities of
LaNiO3, LaPrNiO3 and LaSmNiO3 catalysts have been investigated in
the SR and OSR reactions. The product distribution was quite similar
for all catalysts, as it was previously observed in the transient exper-
iments. Wang et al. [18] studied the steam reforming of acetic acid
(SR) using theoretical (DFT, density functional theory) calculations
and proposed the reaction pathway described by Eqs. (10)–(13).
According to the authors, the initial decomposition of acetic acid on
the surface of Co-based catalysts may take two distinct routes. In
the first route, acetic acid is adsorbed with abstraction of an H atom
and converted into adsorbed acetate species (CH3COO*) (Eq. (10)).
Once formed the acetate species is easily decomposed into acetyl
species (CH3CO*) (Eq. (11)), which decomposes into CH3* and CO*
radicals (Eq. (12)). The radical CO* is oxidized or desorbed but the
radical CH3* may take two different pathways: (i) it is hydrogenated
to CH4 or (ii) it undergoes successive dehydrogenation and forms
carbon (C*) [19,20]. The deposited carbon is than oxidized with O*
atoms provided by water, cleaning the catalyst surface (Eq. (13)).
50
40
30
20
10
0
100
80
60
40
20
0
0
4
8
12
16
20
24
Time on stream (h)
CH3COOH∗ → CH3COO ∗ + H∗,AE = 45.6 kJ mol−1
CH3COO∗ → CH3CO ∗ + O∗,AE = 84.8 kJ mol−1
CH3CO∗ → CH3∗ + CO∗,AE = 77.9 kJ mol−1
C ∗ + O∗ → CO∗,AE = 77.1 kJ mol−1
(10)
(11)
(12)
(13)
Fig. 5. Stability results at 873 K for SR on (A) LaNiO3, (B) LaPrNiO3 and (C) LaSmNiO3
catalysts. The feed is a 400 mL min−1 stream with 13 vol% acetic acid, 40 vol% steam
and 47 vol% He (S/F = 3). (+) Acetic acid conversion, H2
) and acetone ( ).
(
), CH4
(
), CO ( ), CO2
(

K.A. Resende et al. / Catalysis Today 242 (2015) 71–79
77
ORA (873 K)
100
(A)
(A)
90
50
40
30
20
10
0
100
80
80
LaNiO3
LaPrNiO3
LaSmNiO3
70
60
60
50
40
20
0
40
573
673
773
873
973
1073
Temperature (K)
(B)
100
90
80
70
60
50
40
0
4
8
12
16
20
24
Time on stream (h)
(b)
50
40
30
20
10
0
100
80
60
40
20
0
LaNiO3
LaPrNiO3
LaSmNiO3
573
673
773
873
973
1073
Temperature (K)
Fig. 7. TGA profiles of LaNiO3, LaPrNiO3 and LaSmNiO3 samples.
In the SR, the selectivity to H2 and CO decreases, whereas the
0
4
8
12
16
20
24
formation of acetone and CO2 increases in the first 2 h with TOS.
Then, the selectivity to H2 and CO continuously increased, while
the production of CO and acetone decreased. After 18 h TOS, ace-
tone is no longer detected for LaPrNiO3 and LaSmNiO3 catalysts.
Ketonization is an alternative route in the conversion of acetic acid
over polycrystalline oxide samples [17]. It is a bimolecular reaction
in which two molecules of acetic acid are coupled to produce the
symmetric ketone (in this case, acetone), also generating water and
CO2 as byproducts (Eq. (14)).
Time on stream (h)
(C)
50
100
40
30
20
10
0
80
60
40
20
0
2CH3COOH → CH3COCH3 + CO2 + H2O
(14)
Basagiannis and Verykios [26] carried out the TPD of acetic acid
over Al2O3, La2O3 and Ni/La2O3/Al2O3. Ketonization reaction with
the formation of acetone and CO2 occurred over both supports.
However, this reaction did not take place over the catalyst. Accord-
ing to the authors, metallic nickel promotes the decomposition
reaction of acetic acid. In the present work, the increasing forma-
tion of acetone and CO2 is likely due to the oxidation of the metallic
Ni particles by the water in the feed. Then, the NiO particles formed
were reduced again to metallic Ni particles by the syngas produced
during the SR reforming reaction. The formation of metallic Ni par-
ticles that are the active sites for the SR of acetic acid led to an
increase in the selectivity to H2 and CO and the disappearance of
acetone. The selectivity to CO2 did not decrease significantly likely
due to the occurrence of water gas-shift reaction, which produces
CO2 at the expense of CO selectivity decrease. The reduction of the
NiO particles formed during reaction seems to be more effective
for the doped catalysts. The oxidation of the metallic particles by
0
4
8
12
16
20
24
Time on stream (h)
Fig. 6. Stability results at 873 K for OSR on (A) LaNiO3, (B) LaPrNiO3 and (C)
LaSmNiO3 catalysts. OSR conditions: 205 mL min−1 stream with 11 vol% acetic acid,
13 vol% steam, 2 vol% O2 and 74 vol% Ar (S:F:O2 = 1.08:1:0.18). (+) Acetic acid con-
version, H2
(
), CH4
(
), CO ( ), CO2
(
) and acetone ( ).

78
K.A. Resende et al. / Catalysis Today 242 (2015) 71–79
the feed has also been reported for the SR of ethanol over Ni- or Co-
supported catalysts [20]. The absence of methane is associated with
the activity of Ni particles for SR under this reaction conditions.
On the contrary to SR, in the OSR the acetone was formed only in
trace amounts, indicating that ketonization reaction did not occur
for OSR. Kugai et al. 47] reported that the addition of water and oxy-
gen promotes C C bond cleaving, which can explain the decrease
in acetone formation during OSR in comparison with SR [37]. The
addition of O2 to the feed also improved catalyst stability. In the
SR, the conversions decrease during the reaction time for all sam-
ples, while the SRO conversions were stable. De Lima et al. [37]
also observed that the use of a small oxygen amount increased
the stability of LaNiO3 in the oxidative steam reforming of ethanol.
These authors showed that in the presence of oxygen, carbon for-
mation was significantly decreased, which resulted in enhanced
stability. Furthermore, the addition of oxygen to the feed increased
the selectivity to CO and decreased the formation of CO2. This result
is also related to the inhibition of ketonization rate reaction, which
contributes to decrease the formation of CO2. Medrano et al. 48]
also noted the decrease in the CO2 yields during the SRO of acetic
acid, with the correspondent decrease in CO formation. Accord-
ing to them, the partial oxidation of acetic acid (Eq. (15)) is taking
place in the presence of oxygen. They also reported a decrease in
the H2 yield when the percentage of oxygen addition increases. In
this work, the decrease in the H2 yield was not observed, probably
because of the small O2 concentration used.
CH4 → C + 2H2
(20)
The extent of each reaction depends on both reaction conditions
and the catalyst used [22,50]. While low reaction temperatures
favor the formation of carbon through reactions (18) and (19), car-
bon formation via reactions (19) and (20) are the main routes at
higher temperatures. The removal of carbon deposits from the sur-
face can be represented by Eq. (13).
According to Gallego et al. [35], praseodymium oxide (Pr2O3)
has redox properties, which may explain the lower rate of carbon
accumulation in the sample LaPrNiO3. Praseodymium oxide can
promote the cleaning of the catalyst surface according to Eq. (21).
4PrO2 + C∗ → 2Pr2O3 + CO2
(21)
The acid/base properties of the support, which exhibits activity
for this reaction, also determine the reaction pathways and, as a
consequence, also influence catalyst stability. The support can play
a major role in the acetic acid conversion reactions by assisting in
the removal of carbon or suppressing its formation.
In our work, the TPR profiles are in agreement with the in situ
XRD data (Fig. 2). These analyses showed that the perovskite phases
LaNiO3 and La4Ni3O10 were destroyed after reduced producing
metallic Ni particles deposited over La2O3. Fatsikostas et al. [27]
proposed that the La2O3 could react with CO2 to produce La2O2CO3
(Eq. (22)). This lanthanum oxycarbonates can react with carbon
deposited to form CO and regenerate La2O3 by promoting the
removal of carbon (Eq. (23)).
CH3COOH + O2 → 2CO + 2H2O
(15)
La2O3 + CO2 → La2O2CO3
(22)
(23)
The different peaks observed in the TPO profiles of the cat-
alysts used in the SR can be associated with different types of
carbon formed during the reaction such as amorphous carbon, fil-
amentous carbon and graphitic carbon. The oxidation peak located
below 673 K is ascribed to amorphous carbon, whereas the peak
above 773 K corresponds to carbon nanotubes. The oxidation of
graphite occurs at higher temperatures (973 K) [49]. Basagiannis
and Verykios [26] reported a TPO profile of Ni/La2O3/Al2O3 catalyst
after SR of acetic acid at 773 or 973 K, which revealed the pres-
ence of two peaks. The low-temperature oxidation peak (623 K)
was attributed to the carbon deposited on the surface of metallic
particles. The peak at high temperature (923 K) was due to the oxi-
dation of carbon filaments. Therefore, in this work, the TPO profiles
indicate the formation of carbon filaments during SR and OSR of
acetic acid for all catalysts. The addition of oxygen to the feed in
the OSR significantly decreased the amount of carbon deposited
over all catalysts. Oxygen from the feed enhances the gasification
rate of the carbon deposits formed, which improves catalyst stabil-
ity. In this case, the amount of carbon formed over all catalysts was
approximately the same.
La2O2CO3 + C → La2O3 + 2CO
The partial substitution of La by Pr decreased the amount of
carbon formed. According to Gallego et al. [35], the sample derived
from LaPrNiO3 presents redox properties, which are related to the
presence of praseodymium oxide (Pr2O3). This could explain the
lower coke deposition observed for this sample. Pr2O3 can promote
the cleaning of the surface of the catalyst following equation.
4PrO2 + C → 2Pr2O3 + CO2
(24)
5. Conclusions
The data show that the precipitation method was effective
to form the perovskite-type structure. All samples showed good
reducibility and the addition of dopants (Pr and Sm) changed
slightly the reduction profiles. This structure is destroyed during
the reduction process, giving rise to Ni0 particles dispersed over
La2O3. The presence of La4Ni3O10 had no effect on the properties
of the catalyst formed after reduction. This can be confirmed by
the XRD of the passivated samples and by the in situ XRD anal-
ysis. The presence of dopants did not alter significantly the XRD
patterns of the precursors nor the average crystallite sizes. Data
from steam reforming of acetic acid showed that the perovskite-
type are selective catalytic precursors for hydrogen production.
This high selectivity may be explained by the fact that the sur-
face of these samples adsorb and decompose the acetic acid as
observed in TPD analysis. The values of selectivity for the three
precursors studied during the reaction of steam reforming were
very similar. Thus, it can be affirmed that the changes obtained by
doping the LaNiO3 structure with Pr or Sm did not alter signifi-
cantly the catalytic performance of these materials. TGA analysis
showed that the catalyst doped with Pr had the lowest accumu-
lation of coke, probably due to the redox properties of this metal.
The catalytic oxidative reforming tests showed that the presence of
a small amount of O2 contributes to reduce the formation of coke
which helps the catalyst does not deactivate during the reaction.
Among the conditions studied, the best selectivity for hydrogen was
The main reactions that may contribute to coke formation
during acetic acid conversion reactions are as follows [50]: (i) eth-
ylene formation, followed by polymerization to coke (Eq. (16)); (ii)
ketonization reaction followed by dehydration to mesityl oxide
(MO) (Eq. (17)); (iii) the Boudouard reaction (Eq. (18)); (iv) the
reverse of carbon gasification (Eq. (19)); and (v) the decomposition
of hydrocarbons such as methane (Eq. (20)) and ethylene.
C2H4 → coke
(16)
2CH3COCH3 → (CH3)2C(OH)CH2COCH3
→ (CH3)2C CHCOCH3 + H2O
(17)
2CO → CO2 + C
(18)
(19)
CO + H2 → H2O + C

K.A. Resende et al. / Catalysis Today 242 (2015) 71–79
79
obtained at 873 K with a little amount of O2. It was concluded that
among the precursors studied, LaNiO3 showed the best cost benefit.
The addition of Pr and Sm in the crystal structure of perovskite-type
increased the cost of the catalyst, without significantly affecting the
catalytic.
[21] S.C. Dantas, K.A. Resende, R.L. Rossi, A.J. Assis, C.E. Hori, Chem. Eng. J. 197 (2012)
407–413.
[22] Z. Li, X. Hu, L. Zhang, S. Liu, G. Lu, Appl. Catal. A: Gen. 417–418 (2012) 281–289.
[23] J.A. Medrano, M. Oliva, J. Ruiz, L. García, J. Arauzo, J. Anal. Appl. Pyrol. 85 (2009)
214–225.
[24] X. Hu, G. Lu, Appl. Catal. B: Environ. 99 (2010) 289–297.
[25] S. Thaicharoensutcharittham, V. Meeyoo, B. Kitiyanan, P. Rangsunvigit, T. Rirk-
somboon, Catal. Today 164 (2011) 257–261.
[26] A.C. Basagiannis, X.E. Verykios, Appl. Catal. A: Gen. 308 (2006) 182–193.
[27] A.N. Fatsikostas, D.I. Kondarides, X.E. Verykios, Catal. Today 75 (2002) 145–155.
[28] D.L. Trimm, Catal. Today 49 (1999) 3–10.
Acknowledgments
The authors wish to acknowledge the financial support provided
CAPES, FAPEMIG, CNPq, and Vale S.A. The authors are also grateful
to the Brazilian Synchrotron Light Laboratory (LNLS) for providing
facilities and technical support at the D10B-XPD beamline.
[29] R. Peren˜íguez, V.M. González-DelaCruz, J.P. Holgado, A. Caballero, Appl. Catal.
B: Environ. 93 (2010) 346–353.
[30] G. Pecchi, P. Reyes, R. Zamora, C. Campos, L.E. Cadús, B.P. Barbero, Catal. Today
133 (2008) 420–427.
[31] S.O. Choi, S.H. Moon, Catal. Today 146 (2009) 148–153.
[32] A.K. Norman, M.A. Morris, J. Mater. Process. Technol. 92 (1999) 91–96.
[33] Z. Zhang, M. Greenblatt, J. Solid State Chem. 117 (1995) 236–246.
[34] J.L.G. Fierro, J.M.D. Tascón, L.G. Tejuca, J. Catal. 93 (1985) 83–91.
[35] G.S. Gallego, J.G. Marín, C. Batiot-Dupeyrat, J. Barrault, F. Mondragón, Appl.
Catal. A: Gen. 369 (2009) 97–103.
References
[1] S.A. Chattanathan, S. Adhikari, N. Abdoulmoumine, Renew. Sustain. Energ. Rev.
16 (2012) 2366–2372.
[2] A. Tanksale, J.N. Beltramini, G.M. Lu, Renew. Sustain. Energ. Rev. 14 (2010)
166–182.
[36] A. Jahangiri, H. Pahlavanzadeh, H. Aghabozorg, Int. J. Hydrogen Energy 37
(2012) 9977–9984.
[37] S.M. de Lima, A.M. da Silva, L.O.O. da Costa, J.M. Assaf, G. Jacobs, B.H. Davis, L.V.
Mattos, F.B. Noronha, Appl. Catal. A: Gen. 377 (2010) 181–190.
[38] J.L. García-Mun˜oz, J. Rodríguez-Carvajal, P. Lacorre, J.B. Torrance, Phys. Rev. B
46 (1992) 4414–4425.
[39] P. Gordes, N. Christiansen, E.J. Jensen, J. Villadsen, J. Mater. Sci. 30 (1995)
1053–1058.
[3] G.W. Huber, J.A. Dumesic, Catal. Today 111 (2006) 119–132.
[4] E.C. Vagia, A.A. Lemonidou, Appl. Catal. A: Gen. 351 (2008) 111.
[5] N. Iwasa, T. Yamane, M. Takei, J. Ozaki, M. Arai, Int. J. Hydrogen Energy 35 (2010)
110–117.
[6] M. Garcia-Perez, A. Chaala, H. Pakdel, D. Kretschmer, C. Roy, Biomass Bioenergy
31 (2007) 222–242.
[40] S.M. de Lima, A.M. da Silva, L.O.O. da Costa, J.M. Assaf, L.V. Mattos, R. Sarkari, A.
Venugopal, F.B. Noronha, Appl. Catal. B: Environ. 121 (2012) 1–9.
[41] H. Li, Q. Liang, L.Z. Gao, S.H. Tang, Z.Y. Cheng, B.L. Zhang, Z.L. Yu, C.F. Ng, C.T. Au,
Chin. Chem. Lett. 12 (2001) 839–842.
[7] D. Mohan, PittmanF C.U. Jr., P.H. Steele, Energy Fuels 20 (2006) 848–889.
[8] Z. Luo, S. Wang, Y. Liao, J. Zhou, Y. Gu, K. Cen, Biomass Bioenergy 26 (2004)
455–462.
[9] J.A. Medrano, M. Oliva, J. Ruiz, L. García, J. Araizo, Energy 36 (2011) 2215–2224.
[10] C. Yan, F. Cheng, R. Hu, Int. J. Hydrogen Energy 35 (2010) 11693–11699.
[11] R. Trane, S. Dahl, M.S. Skjøth-Rasmussen, A.D. Jensen, Int. J. Hydrogen Energy
37 (2012) 6447–6472.
[12] D. Wang, S. Czernik, E. Chorne, Energy Fuels 12 (1998) 19–24.
[13] L. Garcia, R. French, S. Czernik, E. Chornet, Appl. Catal. A: Gen. 201 (2000)
225–239.
[14] X. Hu, G. Lu, J. Mol. Catal. A: Chem. 261 (2007) 43–48.
[15] M.C. Ramos, A.I. Navascués, L. García, R. Bilbao, Ind. Eng. Chem. Res. 46 (2007)
2399–2406.
[42] R. Peren˜iguez, V.M. Gonzalez-delaCruz, A. Caballero, J.P. Holgado, Appl. Catal.
B: Environ. 123–124 (2012) 324–332.
[43] D.G. Mustard, C.H. Bartholomew, J. Catal. 67 (1981) 186–206.
[44] M.E. Rivas, C.E. Hori, J.L.G. Fierro, M.R. Goldwasser, A. Griboval-Constant, 184,
Journal of Power Sources (2008 265-275).
[45] M. Crespin, P. Levitz, L. Gatinea, J. Chem. Soc. Faraday Trans
1181–1194.
2 (1983)
[46] U. Olsbye, O. Moen, A. Slagtern, I.M. Dahl, Appl. Catal. A: Gen. 228 (2002)
289–303.
[47] J. Kugai, S. Velu, C. Song, Catal. Lett. 101 (2005) 255–264.
[48] J.A. Medrano, M. Oliva, J. Ruiz, L. Garcia, J. Arauzo, Int. J. Hydrogen Energy 33
(2008) 4387–4396.
[49] B. Kitiyanan, W.E. Alvarez, J.H. Harwell, D.E. Resasco, Chem. Phys. Lett. 317
(2000) 497–503.
[16] E.C. Vagia, A.A. Lemonidou, Int. J. Hydrogen Energy 32 (2007) 212–223.
[17] R. Martinez, M.C. Huff, M.A. Barteau, J. Catal. 222 (2004) 404–409.
[18] S. Wang, X. Li, L. Guo, Z. Luo, Int. J. Hydrogen Energy 37 (2012) 11122–11131.
[19] C.N. Ávila-Neto, J.W.C. Liberatori, A.M. da Silva, D. Zanchet, C.E. Hori, F.B.
Noronha, J.M.C. Bueno, J. Catal. 287 (2012) 124–137.
[50] L.V. Mattos, G. Jacobs, B.H. Davis, F.B. Noronha, Chem. Rev. 112 (2012)
4094–4123.
[20] C.N. Ávila-Neto, D. Zanchet, C.E. Hori, R.U. Ribeiro, J.M.C. Bueno, J. Catal. 307
(2013) 222–237.