Decomposition of acetic acid for hydrogen production over Pd/Al2O3 and Pd/TiO2: Influence of metal precursor

Article abstract of DOI:10.1016/j.molcata.2016.02.001

The effect of palladium precursor on the catalytic properties of Pd/Al₂O₃ and Pd/TiO₂ catalysts toward acetic acid decomposition was investigated. The catalysts were prepared by incipient wetness impregnation of titania and alumina with solutions of palladium acetylacetonate, palladium chloride and palladium nitrate. All the catalysts were characterized by N₂-physisorption, XRD, H₂-TPR, H₂-chemisorption, XPS and IR spectroscopy (DRIFTS in situ). Carbonaceous compounds were formed on the used catalysts and were investigated by TPO, TGA experiments and Raman spectroscopy. The results suggested that Pd/Al₂O₃ Nit (prepared using palladium nitrate) was the most active catalyst for conversion of acetic acid, followed by Pd/Al₂O₃ Aca (prepared using palladium acetylacetonate) which was the most selective for H₂. Alumina supported catalysts showed the best performance in acetic acid decomposition and hydrogen production.

Full text of DOI:10.1016/j.molcata.2016.02.001

G Model
MOLCAA-9766; No. of Pages14
ARTICLE IN PRESS
Journal of Molecular Catalysis A: Chemical xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Decomposition of acetic acid for hydrogen production over Pd/Al2O3
and Pd/TiO2: Influence of metal precursor
Laura M. Esteves, Maria H. Brijaldo, Fabio B. Passos∗
Departamento de Engenharia Química e de Petróleo, Universidade Federal Fluminense, Niterói, RJ, Brazil
a r t i c l e i n f o
a b s t r a c t
Article history:
The effect of palladium precursor on the catalytic properties of Pd/Al2O3 and Pd/TiO2 catalysts toward
acetic acid decomposition was investigated. The catalysts were prepared by incipient wetness impregna-
tion of titania and alumina with solutions of palladium acetylacetonate, palladium chloride and palladium
nitrate. All the catalysts were characterized by N2-physisorption, XRD, H2-TPR, H2-chemisorption, XPS
and IR spectroscopy (DRIFTS in situ). Carbonaceous compounds were formed on the used catalysts and
were investigated by TPO, TGA experiments and Raman spectroscopy. The results suggested that Pd/Al2O3
Nit (prepared using palladium nitrate) was the most active catalyst for conversion of acetic acid, followed
by Pd/Al2O3 Aca (prepared using palladium acetylacetonate) which was the most selective for H2. Alumina
supported catalysts showed the best performance in acetic acid decomposition and hydrogen production.
© 2016 Published by Elsevier B.V.
Received 1 October 2015
Received in revised form 30 January 2016
Accepted 1 February 2016
Available online xxx
Keywords:
Hydrogen
Palladium
Decomposition
Acetic acid
Precursor effect
Support effect
1. Introduction
fast pyrolysis of biomass is an alternative for its production and
consequent application in fuel cells [5].
The global growing energy consumption, the pollution caused
by conventional energy production technologies and the world’s
population growth have resulted in a quest for alternative and
renewable energy sources, including biomass [1]. The utilization
of lignocellulosic waste, such as forest residues and sawdust, is
especially sustainable and economically attractive because this
feedstock does not compete with food production [2]. For instance,
bio-oil derived from fast pyrolysis of biomass can be applied to
hydrogen production. Today, hydrogen is one of the most widely
investigated alternative energy carrier, due to its several environ-
mental advantages [3]. There are many ways to produce hydrogen,
but not all of them are a compromise between hydrogen economy
and environmental safety. Nowadays, hydrogen is mainly produced
at industrial scale by steam reforming of natural gas and naphtha,
partial oxidation of heavy residual oils or coal gasification. All these
feedstocks are fossil fuels derived, and thus they contribute for the
increased emissions of greenhouse gases [4]. In order to reduce
these emissions, hydrogen production from bio-oil derived from
Fast pyrolysis is a process in which biomass is rapidly heated
to high temperatures in the absence of oxygen. As result, biomass
decomposes to generate vapor, bio-oil and charcoal. The composi-
tion of bio-oil differs with the type of biomass and the conditions
of pyrolysis [5]. The condensation of these gases leads to a bio-oil
and it can be divided into two fractions: an aqueous phase and
a non-aqueous phase. The aqueous fraction consists of a range of
oxygenate compounds, which can be potential renewable sources
of hydrogen. Due to the amount of oxygenate compounds found in
the aqueous fraction, model compounds have usually been used to
investigate possible pathways and trends when converting bio-oil
at laboratorial scale. Acetic acid is one of the most popular com-
pounds used to represent the aqueous fraction of bio-oil [6] and one
of the major components in bio-oil (up to 19%) [6,7]. Nickel, other
transition metals and noble metals such as palladium, and different
supports (Al2O3, ZrO2, MgO, etc.) have already been tested in several
reactions like the steam reforming of acetic acid [7–10]. The main
disadvantage of this process is the high formation of carbonaceous
deposits on the catalyst surfaces causing the loss of activity.
The decomposition of acetic acid may produce a mixture of car-
bon monoxide and hydrogen according to the following chemical
reaction:
∗
Corresponding author at: Rua Passo da Pátria, 156. São Domingos, Niterói 24210-
240, Brazil. Fax: +55 2126295368.
E-mail addresses: lauraesteves@id.uff.br (L.M. Esteves), mh brijaldo@id.uff.br
(M.H. Brijaldo), fbpassos@vm.uff.br, fabiopassos@id.uff.br (F.B. Passos).
CH3COOH → 2CO + 2H2
(1)
http://dx.doi.org/10.1016/j.molcata.2016.02.001
1381-1169/© 2016 Published by Elsevier B.V.
Please cite this article in press as: L.M. Esteves, et al., J. Mol. Catal. A: Chem. (2016), http://dx.doi.org/10.1016/j.molcata.2016.02.001

G Model
MOLCAA-9766; No. of Pages14
ARTICLE IN PRESS
L.M. Esteves et al. / Journal of Molecular Catalysis A: Chemical xxx (2016) xxx–xxx
2
The literature on decomposition of acetic acid to hydrogen pro-
duction is scarce [11–17] and its investigation may lead to a better
understanding of activity–structure correlations that could help
to optimize bio-oil conversion. For the conversion of acetic acid
and other carboxylic acids over metal supported catalysts acetate
species plays an important role as an intermediate. Acetic acid
decomposes to form acetate, liberating hydrogen and also produce
CO2 and C [13,14].
palladium–support interaction. The use of different palladium pre-
cursors has also been reported in literature as an important factor
to obtain good dispersion of palladium on the surface of the sup-
ports [26,27] to improve the catalytic activity.The main goal of this
work was to determine the effect of different palladium precursors
and supports on the reactivity of palladium supported catalysts for
decomposition of acetic acid. For this purpose, alumina and tita-
nia supported palladium catalysts were prepared using different
palladium precursors (chloride, acetylacetonate and nitrate). The
prepared catalysts were characterized by N2-physisorption, XRD,
H2-TPR, H2-chemisorption, XPS and IR spectroscopy (DRIFTS in situ)
in order to examine the effect of structural and morphological prop-
erties on the catalytic activity and stability. Aiming to verify and
classify the carbonaceous compounds on the surface of the cata-
lysts, TPO, TGA and Raman spectroscopy were conducted with the
used catalysts.
Bowker et al. [13] studied acetic acid adsorption and decom-
position on Pd (110) using temperature-programmed desorption
(TPD). They found the adsorption of acetic acid was efficient and
hydrogen and acetate formation occurred at room temperature.
The acetate decomposed between 47 ◦C and 167 ◦C to form CO2
and hydrogen and left carbon adsorbed on the surface. Acetate
was unstable at higher temperatures (>177 ◦C), but adsorption of
acetic acid continued at a steady-state rate and was not poisoned
by the buildup of carbon on the crystal. This was due to the fact
that most of the carbon deposited was lost from the surface to the
bulk in a facile manner above about 177 ◦C, therefore leaving a reac-
tive surface that was apparently continuously available for acetic
acid decomposition. In other studies, H2 was observed on Rh(110)
[14], while a combination of CO, CO2, H2, and H2O was reported
on Rh(111) [15]. Scharpf and Benziger [16] studied the decompo-
sition of acetic acid monomer an dimer on Ni(100). After acetic
acid monomer adsorbed molecularly at −103 ◦C, the acid hydro-
gen was irreversibly lost at −33 ◦C and bridge-bounded acetate
was formed which then underwent a reversible transformation to a
monodentate acetate above 47 ◦C and after that eventually decom-
posed to CO2, C (ad) and H2 at 162 ◦C. The authors found that the
key step in acetate decompositions is the C–C scission. In other
study, Madix et al. [17] also investigated the acetic acid decom-
position on Ni(110). Acetic acid adsorbed at 30 ◦C yielded gaseous
H2O and formed islands of adsorbed anhydride intermediates. In
their work the decomposition products were H2, CO2, CO, H2 and
surface carbon. Besides, the decomposition of the anhydride was
rate determining for the formation of CO2 and H2.
2. Experimental
2.1. Catalyst preparation
TiO2 (P-25, Sigma–Aldrich) and Al2O3 (Sba200, Sasol) supports
were calcined at 500 ◦C for 3 h. All the catalysts were prepared
by incipient wetness impregnation technique using acetylacetone
(Merck, 99%), palladium chloride (Aldrich, 99%) and palladium
nitrate (Merck, 99%) as precursors. The amount of Pd introduced
to the samples was 1 wt. %. After impregnation, the catalysts were
dried at 120 ◦C for 15 h and calcined at 500 ◦C (10 ◦C/min) for 3 h.
All catalysts, prior to characterization, were reduced at 500 ◦C
under H2 flow (30 mL/min) for 2 h and then heated to 800 ◦C under
He flow (30 mL/min) for 30 min. After this, the samples were cooled
to room temperature and then passivated at liquid nitrogen tem-
perature under a 1% O2/He flow. The Aca, Cl and Nit terminations
were used to denote Pd/Al2O3 and Pd/TiO2 catalysts prepared by
palladium acetylacetonate, chloride and nitrate, respectively.
The choice of support may influence the reaction pathways,
including intermediates, and product selectivity in carboxylic acid
formation [18]. Different types of metal–support interactions may
occur and these interactions can have a significant effect in deter-
mining catalytic activity and selectivity. In a recent work, Brijaldo
et al. [19] studied the acetic acid decomposition over Pd/SiO2,
Pd/Nb2O5, Pd/La2O3 and Pd/Fe2O3 for hydrogen production. The
products of acetic acid decomposition were H2, CO, CH4, CO2,
H2O and C2H4O. Hydrogen was only observed at temperatures
above 600 ◦C. Pd/Fe2O3 was the most active catalyst for conver-
sion of acetic acid and showed high hydrogen selectivity and good
catalytic stability. The Pd/Fe2O3 catalyst exhibited good catalytic
stability during acetic acid decomposition followed by Pd/Nb2O5,
Pd/La2O3 and Pd/SiO2. The excess of active sites turned Pd/Fe2O3
to the most active catalyst and the carbon type found after the
reaction corresponded to a highly ordered carbon. Despite the
mentioned investigations, there are not systematic studies about
the effect of the precursor and the support on acetic acid decom-
position over supported metal catalysts. Therefore, in this work,
the decomposition of acetic acid was carried out using palladium
supported catalysts [20]. Palladium active sites are influenced by
several factors, such as particle metal size, metal-support inter-
action and nature of precursor salts [21]. For palladium catalysts,
metal precursors are usually chlorides or other inorganic salts [22].
Different supports have been widely used to prepare palladium
catalysts, like Al2O3 and TiO2. These supports have different char-
acteristics: Al2O3 is an acid and non-reducible support, while TiO2
is a typical reducible support [23]. The influence of the support
has been studied in the literature [24,25] and it was concluded
that the nature of the support plays an important role in the
2.2. Catalyst characterization
2.2.1. X-ray diffraction
X-ray diffraction measurements of passivated catalysts were
carried out in a Rigaku Miniflex diffractrometer using a CuK␣
◦
˚
(1.540 A) as the incident X-ray source with a scan rate of 0.05 /min
in the range of 10 < 2ꢀ < 80◦.
2.2.2. BET surface area
BET surface area, pore volume and pore diameter dis-
tribution were measured with an ASAP 2020 Micromeritics
analyser, employing nitrogen physisorption at −196 ◦C. The sur-
face area of the reduced catalysts was estimated using the
Brunauer–Emmett–Teller (BET) technique and pore volume and
pore diameter distribution were calculated from the N2 desorption
curve using the Barrett–Joyner–Halenda (BJH) method. The weight
of the samples was around 1 g, for all catalysts and the samples
were previously dried under vacuum at 150 ◦C.
2.2.3. H2-temperature programmed reduction
H2-temperature programmed reduction of calcined catalysts
were performed in an U-shaped tubular quartz reactor coupled to
a mass spectrometer (Pfeiffer Prisma). The samples (500 mg) were
dried at 250 ◦C for 30 min in a He flow of 30 mL/min before TPR
analysis. After cooling to room temperature, TPR was conducted
using a mixture of 5% H2/Ar (30 mL/min) flowing through the sam-
ple and the temperature was raised at a heating rate of 10 ◦C/min
up to 1000 ◦C.
Please cite this article in press as: L.M. Esteves, et al., J. Mol. Catal. A: Chem. (2016), http://dx.doi.org/10.1016/j.molcata.2016.02.001

G Model
MOLCAA-9766; No. of Pages14
ARTICLE IN PRESS
L.M. Esteves et al. / Journal of Molecular Catalysis A: Chemical xxx (2016) xxx–xxx
3
2.2.4. H2-chemisorption
carried out between 300 up to 800 ◦C under 100 mL/min of a 5%
of CH3COOH (Sigma–Aldrich—maintained at 42 ◦C) and 95% of He
mixture with each temperature being kept for 50 min on stream.
Acetic acid was fed to the reactor by bubbling He through a satura-
tor containing this acid at 42 ◦C. The effluent gases were analyzed
by a gas chromatograph (Varian CP 3800) equipped with a thermal
conductivity detector and two capillary columns (Carboxen 1010
plot and CP-PoraBond Q). Prior to the reaction, the catalysts were
dried and reduced for 2 h at 500 ◦C with pure H2 (30 mL/min). Then,
the samples were submitted to He flow (30 mL/min) and heated to
800 ◦C at a rate of 10 ◦C/min and remained at this temperature for
15 min. The reaction temperature was varied from 300 to 800 ◦C,
with each temperature being kept for 50 min on stream. The flow
rate of the gases was controlled using Brooks Instruments’ mass
flow controllers. The total inlet flow rate was 100 mL/min, consist-
ing of 5% acetic acid and 95% helium. The decomposition reaction
was also investigated under isothermal condition at 700 ◦C for 12 h
in order to observe catalyst deactivation.
H2-chemisorption uptakes were measured in a Micromeritics
ASAP 2010C automated gas sorption analyser. The pre-treatment of
the samples (500 mg) consisted of drying in He at 150 ◦C for 30 min.
Then, the samples were evacuated at 150 ◦C during 60 min followed
by reduction with H2 at 500 ◦C for 2 h. After that, the samples were
evacuated for 60 min at the reduction temperature to remove any
residual H2, before cooling in vacuum to 100 ◦C for analysis [23].
The chemisorption analysis were carried out at 100 ◦C to avoid the
formation of palladium hydride that is generally formed at room
temperature and decomposes around 60–80 ◦C. Pd dispersion was
calculated using the total chemisorption uptakes.
2.2.5. X-ray photoelectron spectroscopy
X-ray photoelectron spectroscopy of passivated catalysts was
carried out using an ESCALAB 250Xi Thermo Scientific equipment
with monochromatic Al K␣ X-ray (1486.6 eV) with a pressure of
approximately 1 × 10−9 mbar. The samples were mounted on a
sample holder by means of a double-sided adhesive copper tape.
The analyser was operated using an energy step size of 0.05 eV and a
pass energy of 25.0 eV. The Flood Gun was used to neutralize charge
build up on the surface of the samples. All spectra were calibrated
using the C1s peak at 284.6 eV and Ar+- etching was used to attempt
to break of oxygen film created during the passivation.
To verify the absence of mass transfer effects, the Weisz–Prater
criterion was calculated for all the catalysts used in the acetic acid
decomposition [28]. The values obtained for Weisz–Prater crite-
rion were below 0.3. Therefore, the results reveal the absence of all
transport limitations and guarantees that internal diffusion resis-
tance did not affect the reaction. Mole fraction (yj), conversion of
acetic acid (XC
) and selectivity of hydrogen (SH2 ) were calcu-
H
O
2
4
2.2.6. Acetic acid and CO DRIFTS
lated using the foll2owing expressions:
Diffuse reflectance infrared with Fourier transform spec-
troscopy was used to examine the nature of the adsorbed species on
the catalysts. The system consisted in a FT-IR spectrometer (Bruker,
Vertex 70) equipped with a diffuse reflectance cell with Praying
Mantis geometry and a reaction chamber with ZnSe windows (Har-
rick, HVC-DRP-4). All spectra were obtained after average of 350
nj
ꢀ
yj =
ntotal
yCO + yCH + yCO + 2yC
H
2
O
4
2
2yC
+ yCO + yCH + yCO + 42yC
XC
=
× 100
O
H
O
2
4
2
H
O
H
4
2
4
2
4
2
2
scans and a resolution of 4 cm−1
.
yH
The adsorption of acetic acid was carried out at room tempera-
ture using an acetic acid/He mixture which was obtained by passing
He flow (30 mL/min) in a saturator containing acetic acid main-
tained at 42 ◦C. Before the spectrum was acquired, the passivated
catalysts were dried at 250 ◦C for 30 min with He flow (30 mL/min),
reduced under H2 flow (30 mL/min) at 400 ◦C (10 ◦C/min) for 2 h
to eliminate the oxide layer formed during the passivation step,
cooled under He flow until 30 ◦C and at this temperature, a back-
ground interferogram was collected. After adsorption of acetic acid,
the catalyst was purged under flowing helium (30 mL/min) and
then a new interferogram was taken at 30 ◦C and related to the
background reference to obtain the spectrum of adsorbed acetic
acid.
The acetic acid decomposition surface reaction was also fol-
lowed by DRIFTS from 30 ◦C to 400 ◦C. The pretreatment of the
samples was identical to that describe above. During the cool-
ing step under He flow, the backgrounds were taken at 400, 300,
200,100 and 30 ◦C. Then, under acetic acid/He mixture (30 mL/min)
the interferograms were taken at 30, 100, 200, 300 and 400 ◦C
and related to the background reference (at same temperature) to
obtain the spectrum of adsorbed acetic acid. The cell was kept at
each temperature for 15 min.
2
SH
=
× 100
O
2
yH + 2yCH + yH O + 2yC
H
4
2
4
2
2
2.2.8. Temperature programmed oxidation (TPO),
thermogravimetry and mass spectroscopy (TG–MS)
Temperature programmed oxidation (TPO) experiments of used
sample were performed using a thermo-gravimetric analyzer (Shi-
madzu DTG-60H) coupled to a quadrupole mass spectrometer
(HPR-20, Hiden). The samples were weighed and dried at 200 ◦C
(10 ◦C/min) with a He flow (30 mL/min) during 30 min, followed
by cooling to room temperature. Then, the samples were weighed
again and the temperature was increased until 900 ◦C (10 ◦C/min)
with 5% O2/He continuous flow (50 mL/min). The CO2 produced by
the oxidation of deposited carbon species was monitored by mass
spectroscopy.
2.2.9. Raman spectroscopy
Raman spectra were taken in a Confocal Raman Microscope
alpha 300, with Witec software, using a 50× objective lens and
green laser with 532 nm wavelength. The integration time was 1.0 s
and the number of scans was 400.
Infrared spectroscopy of adsorbed carbon monoxide was also
measured and the pretreatment was similar to that used in acetic
acid DRIFTS analysis. After drying, reduction and purging of the
samples, a background was taken and then 5% CO/He (30 mL/min)
flow was admitted into the cell during 30 min. The samples were
purged with He (30 mL/min) and under He flow the interferograms
were obtained at room temperature.
3. Results and discussion
3.1. Catalyst characterization
Fig. 1 shows the XRD results for passivated catalysts. The diffrac-
tograms of titania supported catalysts shows characteristic peaks of
TiO2 with anatase phase with a low part of rutile phase [29]. Typical
diffraction peaks of ␥-Al2O3 were observed for alumina supported
catalysts. There was not any metallic palladium or any diffraction
line of their oxides probably be due to the low palladium content.
2.2.7. Catalytic activity tests
Decomposition of acetic acid was performed in a quartz reactor
at atmospheric pressure using 100 mg of catalyst. The reaction was
Please cite this article in press as: L.M. Esteves, et al., J. Mol. Catal. A: Chem. (2016), http://dx.doi.org/10.1016/j.molcata.2016.02.001

G Model
MOLCAA-9766; No. of Pages14
ARTICLE IN PRESS
4
L.M. Esteves et al. / Journal of Molecular Catalysis A: Chemical xxx (2016) xxx–xxx
Fig. 1. XRD patterns of passivated catalysts.
Table 1
BET surface area (SBET), pore size, pore volume and H/Pd ratio uptakes obtained for several catalysts.
Catalyst
Pd/Al2O3 Aca
Pd/Al2O3 Cl
Pd/Al2O3 Nit
Pd/TiO2 Aca
Pd/TiO2 Cl
Pd/TiO2 Nit
SBET (m2/g)
172
87
0.46
16
180
93
0.47
25
193
79
0.47
16
4.5
18.1
222
0.10
2
23
˚
Pore size (A)
170
0.014
3
191
0.11
3
Pore volume (cm3/g)
H/Pd (%)
BET surface area, pore volume, pore size and H/Pd ratio obtained
for the several catalysts are listed in Table 1. Alumina supported
catalysts exhibited higher surface area than titania supported cata-
lysts. Pd/Al2O3 Nit presented the highest surface area of all catalysts
(193 m2/g), with a pore volume of 0.47 cm3/g. BET surface areas of
the catalysts with same support were similar, except for Pd/TiO2
Aca whose surface area was 4.5 m2/g because during the treatment
at 800 ◦C some pores collapsed and as consequence, the surface area
decreased. Also, at high temperature the anatase phase is converted
into the more stable rutile (Fig. 1) and unfortunately, this phase
transition to rutile is associated with a significant loss in surface
area [30].
Fig. 2 shows the adsorption isotherms of Pd/Al2O3 Aca, Pd/Al2O3
Cl and Pd/Al2O3 Nit catalysts were type IV isotherms. The isotherms
showed a clear H2 type hysteresis, which indicates ‘ink bottle’ type
pores. Pd/TiO2 Aca, Pd/TiO2 Cl and Pd/TiO2 Nit catalysts presented
type III isotherms with H3 type hysteresis which represents ‘slit-
like’ pores [31].
Titania supported catalysts, showed very low H/Pd values, when
compared with alumina supported catalysts. Titania is a partially
reducible support and high temperature reduction (500 ◦C), leads
to partially reduced species (TiO2-x) which can easily migrate over
metal particles, causing low capacities of hydrogen adsorption. This
phenomenon is called SMSI effect (strong metal-support interac-
tion) [32,33]. On the other hand, for alumina supported catalysts,
pre-treatment at 800 ◦C that could cause sintering of active phase.
Sintering is usually significant when material is heated above its
Tamman temperature ∼650 ◦C for Pd [34]. However, among all
alumina supported catalysts, the Pd/Al2O3 Cl showed the highest
dispersion (25%). Some authors [21,35] affirm that chlorides inter-
act with the hydroxyl groups of supports and favor the dispersion
of the metallic phase.
H2-TPR profiles of the calcined catalysts are shown in Fig. 3. All
the catalysts exhibited a peak at room temperature, ascribed to the
reduction of palladium oxide (PdO) to metallic palladium (Pd0).
Besides the reduction of PdO, hydrogen absorption in the
reduced palladium and adsorption on the metallic surface are also
responsible for the hydrogen consumption at room temperature
[36]. This fact was taken in account when the hydrogen con-
sumption at room temperature was calculated. For all supported
catalysts, the results showed negative TPR peaks in the range of
60–80 ◦C which are attributed to Pd hydride (␤-PdH2) decomposi-
tion, formed at room temperature from the absorption of hydrogen
within the structure of metallic Pd [32,36,37]. Generally, larger
particles (>2 nm) have higher storage capacity than the smaller
ones to form hydrides. The decomposition of palladium hydrides
takes place at higher temperatures in the case of small palladium
particles [38]. Alumina supported catalysts exhibited a peak at
∼355 ◦C possible due to the reduction of small particles of PdO,
which interact most strongly with the support. Similar results have
been reported with different types of interaction between palla-
dium and the support [38–41]. For supported palladium catalysts
prepared with palladium chloride, the peak may also include the
hydrogen consumption due to PdOxCly reduction. These species are
Please cite this article in press as: L.M. Esteves, et al., J. Mol. Catal. A: Chem. (2016), http://dx.doi.org/10.1016/j.molcata.2016.02.001

G Model
MOLCAA-9766; No. of Pages14
ARTICLE IN PRESS
L.M. Esteves et al. / Journal of Molecular Catalysis A: Chemical xxx (2016) xxx–xxx
5
Fig. 2. N2 adsorption–desorption isotherms for palladium-supported catalysts.
Fig. 3. TPR profiles of the investigated catalysts.
formed during the calcination step on catalysts prepared with PdCl2
as palladium precursor [42]. In the case of titania supported cata-
lysts the high temperature peaks are attributed to the reduction of
titania to form partially reduced species, TiOx [43]. This fact is con-
sistent with H2-chemisorption results, in which titania supported
catalysts presented low H/Pd ratios attributed to a decoration of
metallic particles by partially reduced species. Hydrogen consump-
tion during palladium oxide reduction is shown in Table 2.
Pd/Al2O3 Nit presented the higher H2-consumption at room
temperature among all alumina supported catalysts and as men-
tioned earlier, for alumina supported catalysts, the hydrogen
consumption at high temperatures was assigned to PdO species
Please cite this article in press as: L.M. Esteves, et al., J. Mol. Catal. A: Chem. (2016), http://dx.doi.org/10.1016/j.molcata.2016.02.001

G Model
MOLCAA-9766; No. of Pages14
ARTICLE IN PRESS
L.M. Esteves et al. / Journal of Molecular Catalysis A: Chemical xxx (2016) xxx–xxx
6
Table 2
H2 consumption during temperature programmed reduction at room temperature and at high temperatures and degree of reduction.
Catalysts
Room temperature
High temperatures
Degree of reduction (%)
H2 consumption (␮mol H2/gcat
)
H2 consumption (␮mol H2/gcat)
Pd/Al2O3 Aca
Pd/Al2O3 Cl
Pd/Al2O3Nit
Pd/TiO2 Aca
Pd/TiO2 Cl
28.8
44.4
56.0
21.3
102.1
41.1
77.9
68.8
53.3
–
–
–
98.7
100
100
43.4
94.5
55.4
Pd/TiO2 Nit
(2045 cm−1) and bridged-bonded CO (1905 cm−1) for Pd/TiO2 Aca
catalyst. Titania is a reducible support and it is able to decorate pal-
ladium atoms and hinder CO adsorption. This finding is explained
by SMSI effect, which was detected in H2-TPR, H2-chemisorption
and XPS. For alumina supported catalysts, there were two bands
around 1905 cm−1 and 2045 cm−1, typical of CO adsorbed on metal-
lic palladium. These bands can be assigned to bridge-bonded and
linear bonded adsorbed CO, respectively. The Pd/Al2O3 Aca catalyst
showed only one band at 1905 cm−1 for bridged-bonded CO, while
Pd/Al2O3 Cl and Pd/Al2O3 Nit catalysts exhibited two bands due
to bridged-bonded CO (Pd/Al2O3 Cl (1905 cm−1 and 1970 cm−1))
and Pd/Al2O3 Nit (1910 cm−1 and 1970 cm−1). The intensity of
bridge-bonded bands were higher than linear bonded, which was
a consequence of the level of dispersion [47] observed for alumina
supported catalysts. Binding of carbon monoxide to metal surfaces
occurs via a 5␴ donation of the highest occupied molecular orbital
(HOMO) to the metal surface and 2␲* back-donation from the metal
surface to the lowest unoccupied molecular orbital (LUMO) [48].
Bridged-bonded CO bands differ in position depending on several
factors such as coordination of metal atoms, geometry of adsorp-
tion sites and surface coverage degree. On supported palladium
catalysts, bridged-bonded CO located at higher frequencies may
be ascribed to CO adsorption on Pd(100) surfaces and the bands
located at lower frequencies are associated to CO adsorption on
Pd(111) surfaces [49]. Therefore, Pd/Al2O3 Cl and Pd/Al2O3 Nit cat-
alysts contain both Pd(100) and Pd(111) orientation surface sites,
while CO absorption at lower frequencies for Pd/Al2O3 Aca suggest
the presence of surface sites with Pd(111) orientation. This behav-
ior is in agreement with Monteiro et al. [21] who studied the effect
of palladium precursor on the oxidation of carbon monoxide over
Pd/Al2O3 and Pd/CeO2/Al2O3 catalysts.
Fig. 6 shows infrared spectra after adsorption of acetic acid
at room temperature on alumina and titania supported cata-
lysts. All the catalysts exhibited molecular adsorption of acetic
acid. The band attributed to weakly chemisorbed molecular acetic
acid interacting with surface hydroxyl groups can be assigned
to carbonyl stretching v(C O). Undissociated acetic acid forms
monomers and dimers in both phases, liquid and gas, at approx-
imately 1782–1790 cm−1 and 1175 cm−1, respectively [50]. At
room temperature, the bands around 1724 cm−1 and 1294 cm−1
were assigned to some vibrations of dimers of acetic acid. For
Pd/Al2O3 Aca this molecular adsorption is characterized by a band
at 1722 cm−1 for v(C O), 1372 cm−1 for v(C O), 1294 cm−1 for v(C-
C), 1072 cm−1 for ␳(CH3) and 960 cm−1 for ␥(O H) [46]. Similar
bands were detected for the other catalysts as shown in Fig. 6 and
Table 4.
more strongly bonded with the support. Hydrogen uptakes found
for alumina supported catalysts were near the stoichiometric
amount necessary for the reduction of PdO which was reflected in
degree of reduction indicating that at temperatures above 450 ◦C
all palladium was reduced. On the other hand, except for Pd/TiO2
Cl, all the titania supported catalyst showed lower hydrogen con-
sumption and consequently low degree of reduction at room
temperature when compared with alumina supported catalysts.
TPR results indicated the reduction at 500 ◦C under hydrogen flow
is sufficient to totally reduce Pd in the examined catalysts.
The passivated catalysts have also been characterized by XPS
and the results for Pd 3d region are shown in Fig. 4 and Table 3.
The data indicated both Pd2+ and Pd0 species exist in all the cat-
alysts. The binding energies at 335.2 eV and 340.4 eV corresponds
to Pd0 3d5/2 and Pd0 3d3/2, respectively while the binding energies
at 336.2 eV and 341.4 eV corresponds to Pd2+ 3d5/2 and Pd2+ 3d3/2
respectively [23,44].
,
The peak areas of metallic phases are higher than oxide phases
indicating that reduction and passivation steps were efficient. The
existence of Pd2+ species is due to the palladium oxide shell formed
during the passivation step. The XPS spectra of Pd/Al2O3 Cl dis-
played an additional small peak at 337.3 eV which is attributed
to the PdCl2 species from palladium precursor [44]. Pd/support
ratios for the alumina supported catalysts were higher than the
ratios obtained for titania supported catalysts which indicated that
alumina supported catalysts had more palladium on their surfaces.
SMSI was also identified for the titania supported catalysts
because significant binding energy (BE) shift was observed for
O1s and Ti2p (not shown). The negative shift of Ti 2p3/2 indicated
that titania is partially reduced into Ti3+ and oxygen vacancies
were formed [30,45]. When oxygen vacancies are formed at metal-
support interface of titania supported catalysts, oxygen from the
gas phase could be dissociatively adsorbed and the other being
coordinated to the five-coordinated Ti4+ sites. The O1s BE region
is rather complex because it contains contribution not only from
several oxygen species (such TiO2, PdO and hydroxyl groups) but
also from Pd. The O1s BE of passivated titania is slightly lower and
a significant shoulder peaks at approximately 532 eV, correspond-
ing to the chemisorbed oxygen were also observed [30,45]. These
results clearly indicated the existence of an SMSI effect and this
behavior is in agreement with H2-chemisorption results and con-
firms the presence and the influence of the decoration effect in the
case of titania supported catalysts, causing a low exposure of the
active phase.
Infrared spectroscopy of adsorbed CO for all the catalysts is
shown in Fig. 5. In the case of palladium catalysts, two major peaks
are commonly attributed to linearly-bonded CO (above 2000 cm−1
)
For Pd/Al2O3 Aca and Pd/Al2O3 Cl catalysts, the surface acetate
species were characterized by a band at 1540 cm−1 for antisymmet-
ric carboxylate (vas(COO)) and a band at 1426 cm−1 for symmetric
carboxylate (vs(COO)). The splitting of carboxylate stretching
frequencies determines the configuration of acetate species as
monodentate, chelating bidentate or bridging bidentate configura-
tion [46]. In monodentate configuration, the splitting between the
symmetric ␯s(COO) and anti-symmetric ␯a(COO) bands is greater
than 160 cm−1 and the acetate is bound to the surface through only
and bridged-bonded CO (below 2000 cm−1) [21]. The difference in
the position of CO absorption frequencies demonstrates that the
nature of Pd precursors plays a role on the Pd sites properties. Also,
the precursor–support interaction influences the Pd particle size
distribution, which determines the nature of sites available on the
surface.
CO adsorption bands on titania catalysts were not well
defined, however they can be assigned to linearly-bonded CO [46]
Please cite this article in press as: L.M. Esteves, et al., J. Mol. Catal. A: Chem. (2016), http://dx.doi.org/10.1016/j.molcata.2016.02.001

G Model
MOLCAA-9766; No. of Pages14
ARTICLE IN PRESS
L.M. Esteves et al. / Journal of Molecular Catalysis A: Chemical xxx (2016) xxx–xxx
7
Fig. 4. XPS spectra of the passivated catalysts in the Pd 3d region.
Table 3
Binding energies of Pd 3d5/2 and Pd 3d3/2 peaks of the different samples.
Catalyst
Binding energy (eV)
Pd0 3d5/2
Pd/support
Pd0 3d3/2
Pd2+ 3d5/2
Pd2+ 3d3/2
Pd/Al2O3Aca
Pd/Al2O3Cl
Pd/Al2O3Nit
Pd/TiO2Aca
Pd/TiO2Cl
335.0
335.0
335.2
335.2
335.2
335.2
340.3
340.1
340.4
340.4
340.4
340.5
336.3
335.9
336.2
336.1
336.5
336.2
341.7
341.2
341.4
340.4
341.8
341.5
0.027
0.018
0.059
0.018
0.023
0.010
Pd/TiO2Nit
Fig. 5. Infrared spectra of CO adsorption for all Pd/Al2O3 and Pd/TiO2 catalysts.
of its oxygen atoms. In the case of the chelating bidentate mode, the
splitting is considerably lower than ∼160 cm−1, with the acetate
bounded to one surface metal cation through both of its oxygen
atoms. The bridging configuration is characterized by a splitting
larger than of the chelating bidentate, but it is similar to the ionic
value and acetate is bound to the surface with two oxygen atoms
bound to two metal sites [46]. For Pd/Al2O3 Aca and Pd/Al2O3 Cl cat-
alysts, the splitting was 114 cm−1 that suggest a chelating bidentate
configuration. The Pd/Al2O3 Nit catalyst exhibited a band charac-
teristic of molecular acetic acid at 1722 cm−1 and there was only
a band characteristic of acetate species at 1426 cm−1 (vs(COO)).
The band at 1600 cm−1 was assigned to the O
C O stretch and
Please cite this article in press as: L.M. Esteves, et al., J. Mol. Catal. A: Chem. (2016), http://dx.doi.org/10.1016/j.molcata.2016.02.001

G Model
MOLCAA-9766; No. of Pages14
ARTICLE IN PRESS
L.M. Esteves et al. / Journal of Molecular Catalysis A: Chemical xxx (2016) xxx–xxx
8
Fig. 6. DRIFT spectra of surface species formed following adsorption of acetic acid on alumina (left) and titania (right) palladium supported catalysts.
Table 4
Vibrational mode assignments for surface species formed by acetic acid adsorption at room temperature on palladium catalysts.
Band
Infrared band wavenumber (cm−1
)
Pd/Al2O3 Aca
Pd/Al2O3 Cl
Pd/Al2O3 Nit
Pd/TiO2 Aca
Pd/TiO2 Cl
Pd/TiO2 Nit
Al2O3
TiO2
v(C O)HOAc
v(C O)HOAc
v( C)HOAc
␥(O H) HOAc
␳(CH3)
1722
1372
1294
960
1072
1540
1426
1722
1371
1294
960
1070
1540
1426
1722
–
1294
960
1062
–
1726
1363
1294
963
1050
1548
1466
1726
–
1294
–
–
1566
1427
1726
–
1294
962
1050
1573
1426
1722
1372
1294
960
–
1727
–
1294
960
–
vas(COO)acetate
vs(COO)acetate
–
–
–
k
1426
Acetate species at 1426 cm−1 (symmetric carboxylate (vs(COO))),
disappears almost completely upon heating from 300 ◦C up to
400 ◦C, while the band at 1545 cm−1(antisymmetric carboxylate
(vas(COO))) decreased its intensity with temperature increased.
At 300–400 ◦C, bands at 2306 cm−1 (CO2 v(CO)) and 1642 cm−1
(␦(OH) water) were also detected, which suggests that CO2 and H2O
can result from acetate species decomposition (bidentate acetate,
bridging acetate and acyl species) [53]. Similar bands were detected
for Pd/Al2O3 Cl catalyst with some differences in intensity and
position of the characteristic bands. For Pd/Al2O3 Nit the intensity
of molecular acetic acid band also decreased when the tempera-
ture was elevated. At 200 ◦C, there was no evidence of molecular
acetic acid as in previous results. With temperature increase, bands
at 1790 cm−1 (monomeric acetic acid), 1680 cm−1 (acyl species)
and 2372 cm−1 (CO2) were detected. Disappearance of the bands
assigned to molecular acetic acid, mainly at temperatures between
300 and 400 ◦C, suggests that dissociated species of acetic acid are
the precursors of CO2 and H2O.
corresponded to a monodentate acetate species [51]. In the case of
the spectra for alumina supported catalysts, the bands for acetate
species were of low intensity and mainly demonstrated the adsorp-
tion of molecular acetic acid.
Molecular adsorption of acetic acid and acetate species were
formed on the titania supported catalysts, as identical bands were
observed during acetic acid adsorption on alumina supported cat-
alysts, but with minor intensity as shown in Fig. 6 and Table 4.
This behavior indicates a lower adsorption of acetic acid and lower
affinity of acetic acid with titania possibly due to the high particle
coverage of palladium by the partially reduced species of the sup-
port. The same bands that were observed for the titania catalysts
were also observed for pure TiO2. This may indicate that palladium
is less accessible on the surface of catalysts for the adsorption of
acetic acid and its dissociated species.
Alumina supported catalysts exhibited both acetic acid and CO
adsorption possible due to the availability palladium on catalysts
surfaces. However, titania supported catalysts showed lower acetic
acid adsorption and scarce CO adsorption. The above-mentioned
results indicate that on all catalysts, the acetic acid was adsorbed
mainly through the acetate species.
For titania supported catalysts, at temperatures between 100
and 400 ◦C, molecular (∼1722 cm−1) and monomeric (∼1790 cm−1
)
acetic acid was detected. At 400 ◦C, for Pt/TiO2 Aca and Pd/TiO2 Nit
catalysts, characteristic bands of CO2 were observed at ∼2037 cm−1
with the disappearance of monomeric acetic acid and intense
decrease in intensity of acetate species.
Fig. 7 shows infrared spectra after adsorption of acetic acid on
several catalysts at room temperature, 100, 200, 300 and 400 ◦C. At
room temperature, both molecular acetic acid and acetate species
were formed, as identical bands were observed earlier.
For Pd/Al2O3 Aca catalysts, at 100 ◦C, a band at 1489 cm−1
was assigned to bridging bidentate acetate species, the most sta-
ble configuration and it decomposes with further heating [52].
The molecular adsorption of acetic acid decreases with the tem-
perature increase indicating that molecular acetic acid is weakly
bounded. The bands at 1789 cm−1 and 1682 cm−1 are indicative of
acetic acid monomer and acyl species (v(C O)), respectively [50].
3.2. Decomposition of acetic acid
The results of conversion of acetic acid during the decom-
position reaction on the several catalysts are shown in Fig. 8.
Alumina supported catalysts were the most active catalysts and
presented almost 100% of acetic acid conversion above 600 ◦C. In
contrast, titania supported catalysts were not significantly active
Please cite this article in press as: L.M. Esteves, et al., J. Mol. Catal. A: Chem. (2016), http://dx.doi.org/10.1016/j.molcata.2016.02.001

G Model
MOLCAA-9766; No. of Pages14
ARTICLE IN PRESS
L.M. Esteves et al. / Journal of Molecular Catalysis A: Chemical xxx (2016) xxx–xxx
9
Fig. 7. DRIFT spectra of surface species formed during adsorption of acetic acid, at different temperatures on alumina and titania palladium supported catalysts.
at temperatures below 700 ◦C. Conversion curves indicate that
higher temperature favored acetic acid conversion and the fol-
lowing sequence of catalytic activity was pointed out: Pd/Al2O3
Nit > Pd/Al2O3 Aca > Pd/Al2O3Cl> Pd/TiO2 Aca > Pd/TiO2 Cl > Pd/TiO2
Nit.
Table 5 summarizes the initial rate for the acetic acid decom-
position (−rA), the turnover frequency (TOF) for the Pd supported
Please cite this article in press as: L.M. Esteves, et al., J. Mol. Catal. A: Chem. (2016), http://dx.doi.org/10.1016/j.molcata.2016.02.001

G Model
MOLCAA-9766; No. of Pages14
ARTICLE IN PRESS
L.M. Esteves et al. / Journal of Molecular Catalysis A: Chemical xxx (2016) xxx–xxx
10
Table 5
Initial rate, TOF for acetic acid decomposition over Pd supported catalysts at 400 ◦C and carbon amount of carbon deposits after reaction.
Catalysts
−rA(mmol/gcat h)
TOF (s−1
)
Carbon amount (g/gcat)
Pd/Al2O3Aca
Pd/Al2O3 Cl
Pd/Al2O3Nit
Pd/TiO2 Aca
Pd/TiO2 Cl
9.8
4.3
13.0
7.1
0.5
1.1
0.18
0.05
0.24
0.82
0.065
0.10
0.7
0.7
2.4
0.5
0.3
0.6
Pd/TiO2 Nit
it was only detected at 800 ◦C. H2 selectivity followed a similar
behavior to the conversion of acetic acid. Pd/Al2O3 Nit and Pd/Al2O3
Aca were more active at higher temperatures and exhibited higher
hydrogen formation. Catalysts prepared with palladium acetylace-
tonate as precursor were more selective to H2 production. H2 yield
was higher at 800 ◦C (52%) for the Pd/Al2O3 Aca catalyst.
As observed by Brijaldo et al. [19], the products of acetic acid
decomposition were H2, CO, CH4, CO2, H2O and acetaldehyde
(C2H4O) and Fig. 10 shows the reactant and product distribution
(molar fraction) for all the catalysts as a function of reaction tem-
perature. For alumina supported catalysts, at 300 ◦C only CO2 and
H2O were observed. Then, at 400 ◦C acetaldehyde (C2H4O) was also
observed. Acetic acid decomposition could yield acetaldehyde and
oxygen, but only acetaldehyde was detected because the oxygen
produced quickly reacted with acetic acid to form CO2 and H2O.
Acetaldehyde also comes from the hydrogenation of acetic acid
mainly at temperatures between 500 and 600 ◦C. When the tem-
perature was increased, the same behavior was observed at 500 ◦C
and 600 ◦C and three additional reactions were proposed: acetic
acid decomposition to CO and H2, methanation derived from the
reaction between the produced H2 and CO2 and a water–gas shift
reaction.
Fig. 8. Conversion of acetic acid during decomposition over palladium supported
catalysts. Reaction conditions: Total flow: 100 mL/min (5% C2H4O2; 95% He). Catalyst
weight: 100 mg.
At 700 ◦C, hydrogen detection began for Pd/Al2O3 Aca and
Pd/Al2O3 Nit and there was the formation of a high amount of CO,
which suggests that the decomposition of acetic acid into H2 and CO
was favored from this temperature. Other important reaction is the
acetic acid reforming which was observed for all alumina supported
catalysts. As can be seen in Fig. 10, when H2 was observed at tem-
peratures between 700 and 800 ◦C, the amount of water decreased
while CO2 increased. The Boudouard reaction has to be taken in
account for all catalysts. The several products mentioned above can
be explained by the following chemical reactions:
1
2
C2H4O2 → C2H4O + O2
(2)
C2H4O2 + 2O2 → 2CO2 + 2H2O
C2H4O2 → 2CO + 2H2
(3)
(4)
(5)
(6)
(7)
(8)
(9)
C2H4O2 + H2 → C2H4O + H2O
CO2 + 4H2 → CH4 + 2H2O
CO + H2O ↔ CO2 + H2
Fig. 9. Selectivity toward hydrogen production over all the catalysts during acetic
acid decomposition Reaction conditions: Total flow: 100 mL/min (5% C2H4O2; 95%
He). Catalyst weight: 100 mg.
C2H4O2 + 2H2O → 2CO2 + 4H2
2CO → C + CO2
catalysts and the carbon amount present in the catalysts after
decomposition of acetic acid. Turnover frequencies at 400 ◦C were
calculated based on H2 chemisorption as estimate of the number
of active sites. The catalysts presented TOF values in the fol-
lowing order: Pd/TiO2 Aca > Pd/Al2O3 Nit > Pd/Al2O3 Aca > Pd/TiO2
Nit > Pd/TiO2 Cl > Pd/Al2O3 Cl. These results suggest that catalysts
prepared with acetylacetonate and nitrate were more active for
acetic acid decomposition under clean conditions. Catalysts pre-
pared with palladium chloride showed the lower TOF. Additionally,
the results indicate the formation of metal-interface sites for titania
supported catalysts. Fig. 9 shows the selectivity toward hydro-
gen production over all the catalysts. Hydrogen was only detected
above 600 ◦C and for catalysts prepared with palladium chloride
Titania supported catalysts presented a low activity for the con-
version of acetic acid as observed above (Fig. 8). The same products
were observed when compared with alumina supported catalysts,
but with low molar fractions. At temperatures between 300 and
600 ◦C, water, CO2 and some acetaldehyde were observed.
The results indicated that palladium precursor and also the sup-
port have strong influence in the catalytic behavior of the prepared
catalysts. Also, the results indicate that palladium acetylacetonate
is a suitable precursor of palladium catalysts for decomposition
of acetic acid because it was the most selective to hydrogen. On
Please cite this article in press as: L.M. Esteves, et al., J. Mol. Catal. A: Chem. (2016), http://dx.doi.org/10.1016/j.molcata.2016.02.001

G Model
MOLCAA-9766; No. of Pages14
ARTICLE IN PRESS
L.M. Esteves et al. / Journal of Molecular Catalysis A: Chemical xxx (2016) xxx–xxx
11
Fig. 10. Product distribution over palladium catalysts during acetic acid decomposition Reaction conditions: Total flow: 100 mL/min (5% C2H4O2; 95% He). Catalyst weight:
100 mg.
the other hand, palladium chloride showed to be less adequate
to prepare catalysts for decomposition of acetic acid, since chlo-
ride species inhibited the catalytic reaction. DRIFTS results showed
that the mode of acetic acid adsorption was influenced by the
support. Acetate formation was favored on alumina supported cat-
alysts, but a molecular mode of adsorption was preferred on titania
Please cite this article in press as: L.M. Esteves, et al., J. Mol. Catal. A: Chem. (2016), http://dx.doi.org/10.1016/j.molcata.2016.02.001

G Model
MOLCAA-9766; No. of Pages14
ARTICLE IN PRESS
L.M. Esteves et al. / Journal of Molecular Catalysis A: Chemical xxx (2016) xxx–xxx
12
supported catalysts. Besides, catalytic activity varied appreciably
with apparent metal dispersion. Alumina supported catalysts pre-
sented higher adsorption capacity than titania supported catalysts
due to SMSI effect. Taking into account our previous work [19],
we suggest that hydrogen selectivity is influenced by several fac-
tors: metal site distribution, metal-support interface effect and
nature of the support. In addition, the presence of dissociated acid
acetic species also favors the hydrogen production from acetic acid
decomposition. Although SiO2 and Al2O3 are non-reducible sup-
ports, Pd/SiO2 and Pd/Al2O3 presented different catalytic behaviors.
Pd/SiO2 catalyst was not able to produce H2, while Pd/Al2O3 cat-
alysts presented high hydrogen formation. This fact was due to
the existence of dissociated species of acetic acid as monoden-
tate, bidentate chelating and bridging acetates and acyl species on
Pd/Al2O3. On the other hand, Pd/SiO2 mainly exhibited adsorption
of molecular acetic acid. Other factor that contributed to high per-
formance of Pd/Al2O3 was the dispersion and orientation of active
phase. Pd/Al2O3 catalysts showed 16–20% dispersions (Table 1) and
the CO-DRIFTS results indicated a high density of corners and edges
and such imperfections (Fig. 4) while Pd/SiO2 only presented 2% dis-
persion and low distribution of palladium as was detected in the
DRX analysis as consequence of agglomeration of active phase. The
metal-support interface effect also contributes to hydrogen pro-
duction. The catalysts supported on reducible oxides, like Pd/TiO2
(this work) and Pd/Fe2O3, Pd/Nb2O5 and Pd/La2O3 [19] presented
formation of hydrogen. They exhibited Pd0–MOx (MOx: partially
reduced support species), which favored the adsorption of dissoci-
ated species of acetic acid.
Fig. 11. Results of thermal stability tests in the acetic acid decomposition at 700 ◦C.
Reaction conditions: Total flow: 100 mL/min (5% C2H4O2; 95% He). Catalyst weight:
100 mg.
supported catalyst, the conversion decreased during the first 3 h
and then remained low and constant. The deactivation of these cat-
alysts is due to carbon accumulation onto the surface of palladium.
Raman spectroscopy and TPO experiments were useful to char-
acterize carbon deposits on the surface of the catalysts after
decomposition of acetic acid (Figs. 12 and 13). The amount of carbon
deposit was defined as the mass of carbon deposited per mass of cat-
alyst (Table 5). Fig. 12 shows the Raman spectra of the used catalysts
recorded at room temperature. All catalysts exhibited a D band,
centered at ∼1344 cm−1 and a G band, centered at ∼1586 cm−1
[54,55]. These results indicate that there may be two distinct carbon
species on the catalyst surfaces after the acetic acid decomposi-
tion. The D band is associated to the Raman mode of C C bond
stretching due to disordered defective carbon structures and it is
generally associated to an amorphous carbon or amorphous car-
bon hydrogenate [55] and the G band corresponds to the ordered
and well graphitized carbon forms. These results indicate that
the used catalysts presented a mixture of disordered and ordered
3.3. Stability tests
The results of stability tests under isothermal conditions
(700 ◦C) are shown in Fig. 11. As observed, all the catalysts exhib-
ited the same trend of stability, so the conversions decreased with
the time. In the first 4–5 h alumina supported catalysts showed
better stability than titania supported catalysts. Pd/Al2O3 Nit exhib-
ited the best stability among all catalysts. Pd/Al2O3 Cl evidenced a
slight increase of conversion of acetic acid in the first 3 h, but after
that time the conversion decreases and remained with low conver-
sion as consequence of catalysts deactivation. In the case of titania
Fig. 12. Raman spectra of used palladium supported catalysts after acetic acid decomposition.
Please cite this article in press as: L.M. Esteves, et al., J. Mol. Catal. A: Chem. (2016), http://dx.doi.org/10.1016/j.molcata.2016.02.001

G Model
MOLCAA-9766; No. of Pages14
ARTICLE IN PRESS
L.M. Esteves et al. / Journal of Molecular Catalysis A: Chemical xxx (2016) xxx–xxx
13
Fig. 13. TPO profiles of the carbon species present on supported palladium catalysts after acetic acid decomposition.
carbon structures, but D band was more intense, which indicate
there was more disordered carbon than ordered carbon. For titania
supported catalysts there were additional bands between 100 cm−1
and 700 cm−1 (not shown), which are attributed to the support,
predominantly the anatase phase, which is in agreement with the
XRD results of passivated titania supported catalysts [43]. Fig. 13
illustrates the TPO profiles due to the oxidation of carbon species
produced on the supported palladium catalysts after decompo-
sition of acetic acid. The Raman observations after the catalytic
tests revealed two types of carbon (ordered and disordered) and
in fact all the catalysts exhibited broad and not well-defined TPO
peaks. Pd/TiO2 Aca and Pd/TiO2 Cl catalysts showed more clearly
the presence of three different species of carbon because there
were three shoulders at 480 ◦C, 585 ◦C and 670 ◦C in the large
peak. The amount of carbon in the used samples is displayed in
Table 5. Pd/Al2O3 Nit displayed the higher amount of deposited
carbon (2.4 g/gcat) followed by Pd/Al2O3 Aca > Pd/Al2O3 Cl > Pd/TiO2
Nit > Pd/TiO2 Aca > Pd/TiO2Cl. The catalyst deactivation observed in
the stability tests was attributed to the carbon deposits that cover
the metallic particle and the support. Although Pd/Al2O3 Nit was
the most active catalyst during the stability tests, loss of activity
was observed due to formation of carbon deposits.
and palladium acetylacetonate is a suitable precursor of palladium
catalysts for decomposition of acetic acid. This behavior was influ-
enced by the different surface site composition depending on the
used precursors. Pd/Al2O3 Nit is a suitable precursor for acetic
acid conversion due to the presence of different coordination sites
(Pd(100) and Pd(111)) which were responsible for the good activ-
ity, due to the higher presence of site defects that favored acetic
acid dissociation as evidenced by the DRIFTS of adsorbed acetic
acid experiments. On the other hand, for Pd/Al2O3 Aca catalyst
only Pd(111) was observed which favored the good selectivity for
hydrogen. TiO2 supported catalysts showed SMSI behavior, causing
an increase in the turnover frequency. DRIFTS results showed that
acetate species were important on transformation of acetic acid and
on hydrogen production, as evidenced on alumina supported cata-
lysts. In the stability tests, all catalysts showed loss of activity with
time due to the formation of carbon on their surfaces. The deposited
carbon, for all catalysts, was a mixture of ordered and disordered
carbon.
Acknowledgements
The authors are grateful to Capes, CNPq and Faperj for the finan-
cial support.
4. Conclusions
References
The use of different palladium precursors and mainly different
supports played an important role on activities of the catalysts
for acetic acid decomposition. Alumina supported catalysts were
more active and selective on hydrogen than titania supported cat-
alysts at higher temperatures. The present work shows that the
decomposition of acetic acid over palladium supported catalysts
can be performed with high activity at temperatures above 700 ◦C
[1] G. Mao, H. Zou, G. Chen, H. Du, J. Zuo, Renew. Sustain. Energy Rev. 52 (2015)
1823–1833.
[2] S. Nanda, R. Azargohar, A.K. Dalai, J.A. Kozinski, Renew. Sustain. Energy Rev.
50 (2015) 925–941.
[3] F. Barbir, Energy 34 (2009) 308–312.
[4] J.D. Holladay, J. Hu, D.L. King, Y. Wang, Catal. Today 139 (2009) 244–260.
[5] S. Czernik, R. French, Int. J. Hydrogen Energy 39 (2014) 744–750.
Please cite this article in press as: L.M. Esteves, et al., J. Mol. Catal. A: Chem. (2016), http://dx.doi.org/10.1016/j.molcata.2016.02.001

G Model
MOLCAA-9766; No. of Pages14
ARTICLE IN PRESS
L.M. Esteves et al. / Journal of Molecular Catalysis A: Chemical xxx (2016) xxx–xxx
14
[6] L. Garcia, R. French, S. Czernik, E. Chornet, Appl. Catal. A: Gen. 201 (2000)
225–239.
[32] G. Neri, G. Rizzo, L. De Luca, A. Donato, M.G. Musolino, R. Pietropaolo, Appl.
Catal. A: Gen. 356 (2009) 113–120.
[7] L.M. Esteves, M.H. Brijaldo, V.M. Ribeiro, F.B. Passos, Rev. Virtual Quim. 6
(2014) 1062–1075.
[33] H. Rojas, G. Borda, P. Reyes, M. Brijaldo, J. Valencia, J. Chil. Chem. Soc. 56
(2011) 793–798.
[8] K.K. Pant, P. Mohanty, S. Agarwal, A.K. Dalai, Catal. Today 207 (2013) 36–43.
[9] K. Takanabe, K.I. Aika, K. Inazu, T. Baba, K. Seshan, L. Lefferts, J. Catal. 243
(2006) 263–269.
[10] X. Hu, G. Lu, J. Mol. Catal. A: Chem. 261 (2007) 43–48.
[11] C. Houtman, M.A. Barteau, Surf. Sci. Lett. 248 (1991) 57–76.
[12] J.L. Davis, M.A. Barteau, Surf. Sci. Lett. 256 (1991) 50–66.
[13] M. Bowker, C. Morgan, J. Couves, Surf. Sci. 555 (2004) 145–156.
[14] Y. Li, M. Bowker, J. Catal. 142 (1993) 630–640.
[34] R.J. Liu, P.A. Crozier, C.M. Smith, D.A. Hucul, J. Blackson, G. Salaita, Appl. Catal.
A: Gen. 282 (2005) 111–121.
[35] D.O. Simone, T. Kennelly, N.L. Brungard, R.J. Farrauto, Appl. Catal. 70 (1991)
87–100.
[36] F.B. Noronha, D.G.A. Aranda, A.P. Ordine, M. Schmal, Catal. Today 57 (2000)
275–282.
[37] M.G. Musolino, G. Apa, A. Donato, R. Pietropaolo, F. Frusteri, Appl. Catal. A:
Gen. 325 (2007) 112–120.
[15] C.J. Houtman, N.F. Brown, M.A. Barteau, J. Catal. 145 (1994) 37–53.
[16] E.W. Scharpf, J.B. Benziger, J. Phys. Chem. 91 (1987) 5531–5534.
[17] R.J. Madix, J.L. Falconer, a. M. Suszko, Surf. Sci. 54 (1976) 6–20.
[18] Z.F. Pei, V. Ponec, Appl. Surf. Sci. 103 (1996) 171–182.
[19] M.H. Brijaldo, H.A. Rojas, J.J. Martínez, F.B. Passos, J. Catal. 331 (2015) 63–75.
[20] M.J. MacCarrone, C.R. Lederhos, G. Torres, C. Betti, F. Coloma-Pascual, M.E.
Quiroga, J.C. Yori, Appl. Catal. A: Gen. 441–442 (2012) 90–98.
[21] R. Monteiro, L.C. Dieguez, M. Schmal, Catal. Today 65 (2001) 77–89.
[22] C.L. Pieck, C.R. Vera, E.M. Peirotti, J.C. Yori, Appl. Catal. A: Gen. 226 (2002)
281–291.
[38] S. Bhogeswararao, D. Srinivas, J. Catal. 327 (2015) 65–77.
[39] N. Seshu Babu, N. Lingaiah, P.S. Sai Prasad, Appl. Catal. B: Environ. 111–112
(2012) 309–316.
[40] F. Cárdenas-Lizana, S. Gómez-Quero, M.A. Keane, Appl. Catal. A: Gen. 334
(2008) 199–206.
[41] F. Cárdenas-Lizana, S. Gómez-Quero, C. Amorim, M.A. Keane, Appl. Catal. A:
Gen. 473 (2014) 41–50.
[42] F.B. Noronha, M.A.S. Baldanza, R.S. Monteiro, D.A.G. Aranda, A. Ordine, M.
Schmal, Appl. Catal. 210 (2001) 275–286.
[43] L.M. Sikhwivhilu, N.J. Coville, D. Naresh, K.V.R. Chary, V. Vishwanathan, Appl.
Catal. A: Gen. 324 (2007) 52–61.
[44] H. Karhu, A. Kalantar, I.J. Väyrynen, T. Salmi, D.Y. Murzin, Appl. Catal. A: Gen.
247 (2003) 283–294.
[23] X. Wang, G. Wu, N. Guan, L. Li, Appl. Catal. B: Environ. 115–116 (2012) 7–15.
[24] W.J. Shen, M. Okumura, Y. Matsumura, M. Haruta, Appl. Catal. A: Gen. 213
(2001) 225–232.
[25] S. Scirè, S. Minicò, C. Crisafulli, Appl. Catal. A: Gen. 235 (2002) 21–31.
[26] L.M.T. Simplício, S.T. Brandão, E.A. Sales, L. Lietti, F. Bozon-Verduraz, Appl.
Catal. B: Environ. 63 (2006) 9–14.
[45] H. Huang, D.Y.C. Leung, J. Catal. 280 (2011) 60–67.
[46] W. Rachmady, M.A. Vannice, J. Catal. 207 (2002) 317–330.
[47] S. Rieck, A.T. Bell, J. Catal. 54 (1987) 46–54.
[27] N.M. Kinnunen, M. Suvanto, M.A. Moreno, A. Savimäki, K. Kallinen, T.-J.J.
Kinnunen, T.A. Pakkanen, Appl. Catal. A: Gen. 370 (2009) 78–87.
[28] H.S. Fogler, Elements of Chemical Reaction Engineering, Prentice Hall, New
Jersey, 1999.
[29] M. Hosseini, S. Siffert, H.L. Tidahy, R. Cousin, J. Lamonier, Cat. Today 122
(2007) 391–396.
[30] J.J. Dodson, H.E. Hagelin-weaver, J. Mol. Catal. A: Chem. 410 (2015) 271–279.
[31] K.S.W. Sing, D.H. Everett, R.A.W. Haul, L. Moscou, R.A. Pierotti, J. Rouquérol, T.
Siemieniewska, Pure Appl. Chem. 57 (1985) 603–619.
[48] G. Blyholder, J. Phys. Chem. A 68 (1964) 2772–2777.
[49] A. Palazov, G. Kadinov, Ch. Bonev, D. Shopov, J. Catal. 74 (1982) 44–54.
[50] R. Haley, M. Tikhov, Catal. Lett. 76 (2001) 125–130.
[51] Z. Li, F. Calaza, F. Gao, W.T. Tysoe, Surf. Sci. 601 (2007) 1351–1357.
[52] A.R. Garcia, J.L. da Silva, L.M. Ilharco, Surf. Sci. 415 (1998) 183–193.
[53] M.A. Hasan, M.I. Zaki, L. Pasupulety, Appl. Catal. A: Gen. 243 (2003) 81–92.
[54] J. Xu, W. Zhou, Z. Li, J. Wang, J. Ma, Int. J. Hydrogen Energy 34 (2009)
6646–6654.
[55] J.E. Herrera, D.E. Resasco, J. Catal. 221 (2004) 354–364.
Please cite this article in press as: L.M. Esteves, et al., J. Mol. Catal. A: Chem. (2016), http://dx.doi.org/10.1016/j.molcata.2016.02.001