Pd-supporting LaCoO3 catalyst with structured configuration for water gas shift reaction

Article abstract of DOI:10.1016/j.apcata.2014.03.005

The purpose of this study was to develop a structured perovskite-type oxide catalyst for the water gas shift (WGS) reaction, which could make lattice oxygen mobile at a low external heat energy by preparing it as a thin catalyst layer on a metal plate. The thickness of the formed LaCoO₃ layer was about 20 μm, which was confirmed by FE-SEM and EDX analysis. For enhancing its WGS reaction performance, a palladium (Pd) component was supported on the structured catalyst using various Pd precursors. When using a PdCl₂ aqueous solution which was prepared by filtration of the PdCl₂ slurry (the filtration liquid), the Pd-supporting LaCoO₃ structured catalyst showed a high and stable activity. The reason for such an enhanced activity was that the supported Pd was in a highly dispersed state, which was derived from the electrostatic interaction between [PdCl₃(H ₂O)]^- as the anionic charged precursor in the filtration liquid and the oppositely charged LaCoO₃ support surface confirmed by a UV-vis measurement. In order to remove the remaining chloride ligand from the Pd-supporting LaCoO₃ catalyst surface, a washing treatment was performed by immersing the as-made catalyst in water, NH₃ (aq) or NaOH (aq) at pH 11.5 for 30 min. Consequently, these washing processes were effective for eliminating the remaining chloride ion from the catalyst surface and improving the shift activity of the catalyst. The catalyst washed using NH₃ (aq) showed the highest activity among these catalysts.

Full text of DOI:10.1016/j.apcata.2014.03.005

Applied Catalysis A: General 477 (2014) 75–82
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Pd-supporting LaCoO3 catalyst with structured configuration for
water gas shift reaction
Ryo Watanabea, Yusuke Fujitaa, Tomohiro Tagawaa, Kazumasa Yamamotoa,
Takeshi Furusawab, Choji Fukuharaa,∗
a Department of Applied Chemistry and Biochemical Engineering, Graduate School of Engineering, Shizuoka University, 3-5-1, Johoku, Naka-ku,
Hamamatsu, Shizuoka 432-8561, Japan
b Department of Applied Chemistry, Faculty of Engineering, Utsunomiya University, Yoto 7-1-2, Utsunomiya 321-8585, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
The purpose of this study was to develop a structured perovskite-type oxide catalyst for the water gas shift
(WGS) reaction, which could make lattice oxygen mobile at a low external heat energy by preparing it as
a thin catalyst layer on a metal plate. The thickness of the formed LaCoO3 layer was about 20 ␮m, which
was confirmed by FE-SEM and EDX analysis. For enhancing its WGS reaction performance, a palladium
(Pd) component was supported on the structured catalyst using various Pd precursors. When using a
PdCl2 aqueous solution which was prepared by filtration of the PdCl2 slurry (the filtration liquid), the
Pd-supporting LaCoO3 structured catalyst showed a high and stable activity. The reason for such an
enhanced activity was that the supported Pd was in a highly dispersed state, which was derived from
the electrostatic interaction between [PdCl3(H2O)]− as the anionic charged precursor in the filtration
liquid and the oppositely charged LaCoO3 support surface confirmed by a UV–vis measurement. In order
to remove the remaining chloride ligand from the Pd-supporting LaCoO3 catalyst surface, a washing
treatment was performed by immersing the as-made catalyst in water, NH3 (aq) or NaOH (aq) at pH 11.5
for 30 min. Consequently, these washing processes were effective for eliminating the remaining chloride
ion from the catalyst surface and improving the shift activity of the catalyst. The catalyst washed using
NH3 (aq) showed the highest activity among these catalysts.
Received 7 November 2013
Received in revised form 14 February 2014
Accepted 1 March 2014
Available online 13 March 2014
Keywords:
Water gas shift reaction
Structured catalyst
Perovskite-type oxide
LaCoO3
PdCl2
© 2014 Elsevier B.V. All rights reserved.
1. Introduction
kinetically suppressed at low temperature. For the industrial pro-
cesses, the WGS reaction is performed using two stage catalytic
The fuel cell is one of the attractive technologies in terms of a
highly efficient and clean energy resource. Hydrogen for the poly-
mer electrolyte membrane fuel cell (PEMFC) is produced via the
steam reforming of hydrocarbons [1,2]. The reformed gas contains
not only H2, but also CO. The CO must be converted below 10 ppm
in the reformed gas because CO deactivates the Pt electrodes of fuel
cells [3]. The water gas shift (WGS, Eq. (1)) reaction is one of the
most effective reactions for CO reducing processes.
processes; i.e., a high temperature shift (HTS) and low tempera-
ture shift (LTS). The HTS is carried out with a Fe2O3–Cr2O3 catalyst
at a high temperature range (583–723 K) [4–6], and the LTS is
performed with a Cu/ZnO/Al2O3 catalyst for decreasing the CO
level to less than 1% in the low temperature range (483–523 K)
[7–15]. Recently, a variety of catalysts has been reported; a highly
active Pd/K/Fe2O3 catalyst at the mild temperature of 573 K [16],
an alkali-promoted Pt/SiO2 catalyst for the low temperature shift
[17] and sustainable Cu/Al2O3 catalyst for the daily start-up and
shutdown (DSS) like operation [18]. In addition, nano-sized gold
has emerged as a prominent catalyst for the WGS reaction after the
pioneering work of Haruta et al. [19], and a multitude of studies
for an effective Au catalyst has been performed, such as the novel
Au/La2O3 and Au/La2O2SO4, Au/Fe2O3 and Au/CeO2 as promising
low-temperature shift catalysts [20–22]. However, the price of pre-
cious metals has increased every year. In terms of the cost for
precious metals, a more efficient and low-cost catalyst needed to be
developed.
CO + H2O ꢀ CO2 + H2 ꢀH2098.15 = −41.2 kJ/mol
(1)
The WGS reaction is an exothermic reaction and the equilibrium
conversion is low at high temperature due to thermodynamic lim-
itations. Therefore, the reaction temperature should be lowered
to achieve a higher CO conversion, though the reaction rate is
∗
Corresponding author.
E-mail address: tcfukuh@ipc.shizuoka.ac.jp (C. Fukuhara).
http://dx.doi.org/10.1016/j.apcata.2014.03.005
0926-860X/© 2014 Elsevier B.V. All rights reserved.

76
R. Watanabe et al. / Applied Catalysis A: General 477 (2014) 75–82
La(NO3)3
Co(NO3)2
Citric acid
PdCl2
Dispersion in distillated water
Dissolution in distillated water
Stirring for 24 h
at room temperature
Evaporation to dryness
Dissolution in distillated water again
Dip-coating on metal plate
Filtration
PdCl2 filtration liquid
Scheme 2. Preparation of PdCl2 filtration liquid.
Calcination
follows: metal ion nitrates included in the perovskite type oxide
were dissolved in distilled water, then citric acid was added to the
solution at the 1/3 molar ratio of total metal ions/citric acid. It was
then evaporated to form a powder at 353 K for 10 h. Following that,
the powder was dissolved in H2O again and a viscid gel solution
was obtained. The adherence of the perovskite-type oxide onto the
star-shaped metal substrate (20 mm␸ × 30 mm length) was carried
out by immersion of the substrate in the viscid gel solution. The
metal substrate was stainless steel with a 100 ␮m thickness. After
the gel coating, calcination of the coated substrate was performed
at 673 K for 2 h. Finally, the substrate was calcined at 873 K for 10 h.
The prepared catalyst in this way was represented as the structured
LaCoO3 catalyst.
(673 K, 2h followed by 873 K, 10 h)
Structured LaCoO3 catalyst
Scheme 1. Preparation of the structured LaCoO3 catalyst.
A preliminary study showed that the reduction temperature of
the catalyst could influence the catalytic performance of the WGS
reaction [23,24]. A precious metal supported on a perovskite-type
oxide has the unique property; the reduction temperature of the
perovskite-type oxide is significantly reduced by supporting novel
metals such as platinum (Pt) and palladium (Pd). The reduction by
the loading of Pd or Pt resulted in an increased of WGS reaction
activity, in other words, the easier release of lattice oxygen, more
CO was oxidized to CO2 by the lattice oxygen in the LaCoO3, and also
accompanied by a change in the produced hydrogen. The mobility
of the lattice oxygen in the catalyst was found to be a key point for
achieving a high activity.
For the application for the perovskite-type catalyst to a fixed-
bed reaction system, the lattice oxygen in the catalyst at the center
of the reactor is not effectively utilized, compared to the catalyst
in the vicinity of the reaction wall because a severe temperature
gradient is generated from the reactor wall toward the center. For
effectively utilizing the lattice oxygen in the catalyst, we focused
on a plate type catalyst on a metal substrate [25–27]. The plate
type catalyst, which is commonly called a structured catalyst, is
composed of a regularly arranged catalyst on the metal plate. By
placing the catalyst on the metal plate, an effective heat transfer to
the reaction field could be achieved due to the conductional heat
transfer. Furthermore, such a catalyst could overcome the disad-
vantage of a poor heat conductivity of the perovskite-type oxides,
and high mobility of the lattice oxygen would be acquired at a low
temperature due to the effective utilization for the external heat
energy.
2.2. Loading method of Pd on the structured LaCoO3 catalyst
The Pd loading was performed by immersing the structured
LaCoO3 catalyst in an aqueous solution of the Pd precursor salt of
Pd(NH3)4Cl2 or Pd(ethylenediamine)2Cl2 (denoted as Pd(en)2Cl2)
for 2 min. The Pd-dipped catalyst was then dried at room temper-
ature for 30 min. These operations were repeated until a 1 wt% Pd
loading was achieved. In addition, Pd(acetylacetonate)2 (denoted as
Pd(acac)2), which was dissolved in an acetone solution, was used as
the precursor. PdCl2 was also used as the precursor, however, the
PdCl2 hardly dissolved in the H2O and organic solvents, unlike the
other reagents. Therefore, a solution with a trace amount of the Pd
component dissolved in H2O was utilized as the precursor. Namely,
the PdCl2 reagent was dispersed in H2O (denoted as the slurry) and
the slurry was stirred for 24 h at room temperature. Filtration of
the slurry was then performed, and the filtration liquid was used
as the Pd precursor solution. Scheme 2 shows steps in the prepa-
ration of this PdCl2 filtration liquid. The pH of this filtration liquid
was 2.5. After the Pd loading, the substrate was calcined at 823 K
for 2 h.
The washing procedures for the highly active catalyst were
applied in order to remove the chloride ligand from the Pd com-
plexes loaded on the support. The washing agents were water, an
aqueous solution of NH3 or NaOH at pH 11.5. The washing proce-
dure was employed by immersion of the Pd-supporting structured
catalyst without calcination in each solution for 30 min without
stirring. After immersion in each solution, the structured catalyst
was washed with distilled water for 2 min. The catalyst was then
calcined at 823 K for 2 h.
In this study, a plate-like LaCoO3 catalyst was developed on
the metal substrate by the following procedures; a viscid perov-
skite precursor solution was coated on a stainless steel substrate,
then the coated substrate was calcined several times. The WGS
performance of the plate-like LaCoO3 catalyst was then investi-
gated. Furthermore, for enhancing the WGS activity, the effects of
Pd supported on the structured LaCoO3 catalyst were examined by
changing the types of Pd precursors.
2.3. WGS performance
2. Experimental
The WGS performance test of the prepared catalyst was per-
formed at 573 K in a conventional flow reactor system. After setting
the prepared catalyst in the reactor, the reactants of CO and H2O
were supplied at 12.2 and 24.4 ml min−1, respectively. The ratio of
2.1. Preparation of the structured LaCoO3 catalyst
Scheme 1 shows steps in the preparation of the structured
LaCoO3 catalyst. The preparation procedure of the catalyst was as

R. Watanabe et al. / Applied Catalysis A: General 477 (2014) 75–82
77
steam to carbon was 2. The catalyst weight was 0.5 g excluding
the substrate weight. The gaseous reactants and products, such as
CO, CO2 and CH4, after the trapping of steam in the effluent gas
through the cold trap were collected by a gas-tight syringe, then
injected into the off-line thermal conductivity detection gas chro-
matograph (GC-8A; Shimadzu Co., Ltd., Japan). After sampling at
10 min for the first plot, subsequent samplings were carried out
every 20 min. The CO conversion and CO2 selectivity were evalu-
ated using Eqs. (2) and (3), in which [CO], [CO2] and [CH4] denote
the concentration of CO, CO2 and CH4 in the outlet gas from the
reactor. The carbon balances of these investigations amounted to
≥98%.
[CO2] + [CH4]
CO conversion (%) =
CO2 selectivity (%) =
× 100
(2)
[CO] + [CO2] + [CH4]
[CO2]
× 100
(3)
[CO2] + [CH4]
2.4. Characterization of LaCoO3 catalysts and Pd precursors
The crystalline structure of the catalyst was obtained by an XRD
analysis (40 kV, 40 mA) using a X-ray diffraction meter (RINT-2200,
Rigaku Co., Ltd.) and CuK␣ radiation source filtered by Ni. The sur-
face structural properties of the catalyst were analyzed by Field
Emission-Scanning Electron Microscope (JSM-7001F, JEOL Co., Ltd.)
equipped with EDX and by a High Resolution-Transmission Elec-
tron Microscope (JEM-2010, JEOL Co., Ltd.).
Fig. 1. XRD pattern of the coated catalyst component peeled off the stainless steel
substrate.
3.2. WGS performance of structured LaCoO3 and Pd supported
catalyst
The H2-temperature-programmed-reduction (H2-TPR) mea-
surements were performed by a TAPS3000S (Bruker AXS Co., Ltd.).
The catalyst sample (20 mg) was placed in the center of the muffle.
The heating rate was 10 K min−1 from 323 to 873 K. The weight loss
of zero was defined as the starting weight of the catalyst at 323 K.
UV–vis spectroscopy measurements for clarifying the Pd species
in the filtration liquid of the PdCl2 slurry were carried out at room
temperature by a UV-vis instrument (UV-1700, Shimadzu Co., Ltd.)
in the range of 190–550 nm. Scans were performed with a data
In order to evaluate the catalytic performance of the struc-
tured LaCoO3 catalyst, the WGS reaction was carried out in the
reaction temperature range of 523–873 K. Fig. 3 represents the CO
conversion as a function of the reaction temperatures. The cat-
alytic activity over the structured LaCoO3 catalyst increased with a
change in the reaction temperature as follows; 0.7% at 523 K, 3.9%
at 573 K, 14.7% at 623 K, 39.9% at 673 K, 64.5% at 723 K and 78.3% at
773 K. Although CH4 was also produced as a by-product, the selec-
tivity of CH4 was less than 0.3% within the reaction temperature
range from 523 to 873 K. The structured LaCoO3 catalyst was a
highly active and selective catalyst for the WGS reaction at high
temperature, but the activity was not significantly high at the low
temperatures from 523 to 573 K.
The loading of a precious metal promoted the ability of releas-
ing the lattice oxygen, and such promotion resulted in an increased
WGS performance [23,24]. In order to enhance the WGS activity of
the structured catalyst, the LaCoO3 catalyst supported Pd compo-
nent was prepared and was its performance investigated. Table 1
shows the CO conversion and selectivity to CO2 and CH4 at the time
when a stable activity was obtained over the Pd-supported catalysts
prepared using various palladium precursors. The CO conversion
interval of 1 nm and a scan rate of 90 nm min−1
.
An X-ray photoelectron spectroscopy (XPS, ESCA-3400; Shi-
madzu Co., Ltd.) measurement was also performed using
non-monochromatic AlK␣ radiation. The pass energy of the ana-
lyzer was at 23.5 eV. Binding energy of C1s at 284.7 eV was used for
calibration.
3. Results and discussion
3.1. Morphology of the structured LaCoO3 catalyst
Fig. 1 represents the XRD pattern of the as-made catalyst which
was peeled off the stainless steel substrate. The peaks of the as-
made catalyst corresponded to the perovskite structure of LaCoO3
and no impurity phases were detected. For calculating the crystal-
lite sizes using the Scherrer equation and the peak at 2ꢁ = 32.9◦,
the diameter was estimated to be 15.2 nm. This value was close to
that in the literature [28,29]. Fig. 2(a) and (b) shows SEM images of
the surface and cross section of the catalyst, including the results
of the elemental analyses by EDX. As shown in Fig. 2(a), the cata-
lyst surface was composed of sintered spherical particles of 50 nm
diameter that produced a porous-like structure. From the cross sec-
tional view in Fig. 2(b), the thickness of the formed layer, which
was composed of La, Co and O elements, was about 20 ␮m. Based
on these results, a LaCoO3 catalyst layer was uniformly formed on
the stainless steel substrate. In addition, the pores, which were
observed on the LaCoO3 surface as shown in Fig. 2(a), were con-
sidered not to reach the substrate surface.
Table 1
CO conversion and selectivity over the Pd/LaCoO3 catalysts prepared with different
types of palladium precursors.
Pd precursor
CO conversion/%
Selectivity/%
CO2
CH4
Pd(acac)2
20.3
12.1
13.2
58.6
0.3
99.4
93.3
99.8
98.3
100.0
100.0
0.6
6.7
0.2
1.7
0.0
0.0
Pd(NH3)4Cl2
Pd(en)2Cl2
PdCl2 filtrate
Bare LaCoO3
a
Pd/Al2O3
1.0
Reaction condition: CO and H2O were supplied at 12.2 and 24.4 ml min−1
,
respectively; Catalyst weight was 0.5 g excluding the substrate weight; Reaction
temperature was 573 K.
a
PdCl2 filtrate was used as Pd precursor.

78
R. Watanabe et al. / Applied Catalysis A: General 477 (2014) 75–82
Fig. 2. SEM images and elemental analyses by EDX for (a) the surface and (b) the cross section of the structured LaCoO3 catalyst.
was as follows: 20.3% over the catalyst prepared using Pd(acac)2,
12.1% over Pd(NH3)4Cl2 and 13.2% over Pd(en)2Cl2. However, when
using the filtration liquid of the PdCl2 slurry, a higher conversion
of 58.6% was obtained. It was found that the types of Pd precursors
seriously affected the WGS performance and the use of the filtration
liquid was the most suitable. Since a Pd/Al2O3 catalyst prepared
using the filtration liquid and a bare LaCoO3 catalyst showed a sig-
nificantly low activity of 1.0% and 0.3% CO conversion at 573 K, as
shown in Table 1, the activity progression by loading Pd would be
due to a synergistic effect between the LaCoO3 and the Pd compo-
nent. In other words, the mobility of the lattice oxygen would be
enhanced by the supported Pd on the LaCoO3 and such progression
might improve the WGS performance of the structured catalyst.
Additionally, the CH4 selectivity was only 1.7% over the Pd/LaCoO3
catalyst prepared using the PdCl2 filtration liquid. It was found that
the Pd/LaCoO3 catalyst was highly selective for the WGS reaction.
catalysts. The left axis denotes the weight loss and the right axis
represents the differential value of the weight loss (ꢀ weight loss)
which was obtained by dividing the weight loss by the tempera-
ture. From Fig. 4, a two-step reduction over the bare LaCoO3 was
observed at 654 and 837 K, which was known to occur; LaCoO3
was reduced to LaCoO2.5 (1/2La2O3 + CoO) corresponding to the
Co3+ reduction to Co2+ at about 654 K, followed by LaCoO2.5 being
reduced to LaCoO1.5 (1/2La2O3 + Co) corresponding to the Co2+
reduction to Co0 at about 837 K. By supporting the Pd compo-
nent, the reducibility of the Co cation was enhanced as shown in
Fig. 4. The as-made Pd/LaCoO3 catalyst has three peaks at 386, 461
and 809 K as confirmed by the ꢀ weight loss, although the bare
LaCoO3 catalyst has two peaks at a higher temperature. Chiarello
et al. reported that loading Pd promoted the reduction of the Co
cation in LaCoO3, and Co3+ to Co2+ was reduced in two steps [30].
The two-step reduction was explained by Eqs. (4) and (5) on the
catalyst:
3.3. Role of Pd for WGS reaction
LaCoO3 + (1/3)H2 → (1/3)La3Co3O8 + (1/3)H2O
(4)
(5)
(1/3)La3Co3O8 + (1/6)H2 → (1/2)La2Co2O5 + (1/6)H2O
For clarifying the promotion by the supported Pd, H2-TPR mea-
surements were performed over the bare LaCoO3, the as-made
Pd/LaCoO3 and the Pd/LaCoO3 after the WGS reaction (denoted
as spent catalyst). Fig. 4 shows the reduction behavior of these
They also found that these reactions progressed by incorpo-
ration of Pd as the B-site in LaCoO3. Due to the incorporation of
0
-5
0.25
0.2
100
Co3+→Co2+
Co2+→Co0
Equilibrium
80
Bare LaCoO3
As-made Pd/LaCoO3
Spent Pd/LaCoO3
0.15
60
40
20
0
-10
-15
-20
809
837
0.10
386
654
0.05
0
461
578
300
500
700
900
500
600
700
800
900
Temperature / K
Reaction temperature / K
Fig. 4. Reduction behaviors in presence of H2 over the bare LaCoO3, the as-made
Fig. 3. CO conversion over the structured LaCoO3 catalyst.
Pd/LaCoO3 and the spent Pd/LaCoO3 catalyst.

R. Watanabe et al. / Applied Catalysis A: General 477 (2014) 75–82
79
Fig. 5. Overview of the filtration liquids with different pH values.
Pd, the reduction of LaCoO3 occurred through the intermediate
phases with the general formula LanConO3n−1. The initial and sec-
ond reduction steps corresponded to n = 3 (Eq. (4)) and 2 (Eq. (5)),
respectively. From Fig. 4, the two-step reduction of Co3+ to Co2+
was confirmed over the as-made Pd/LaCoO3 catalyst as stated by
Chiarello. Therefore, the supported Pd might be initially incorpo-
rated into the B-site of the LaCoO3.
On the other hands, the spent catalyst did not show a two-
step reduction, but one moderate reduction at about 578 K was
observed. The reason for the one moderate reduction might be
derived from no-interaction of Pd with the bulk Co3+ in the perov-
skite structure because the Pd originally incorporated into the
B-site appeared on the surface during the WGS reaction. Such sur-
face Pd clusters triggered the dissociative adsorption of H2, and
H adatoms were subsequently formed. The H adatoms lead to the
consumption of the lattice oxygen at a lower reduction tempera-
ture as compared to the bare LaCoO3. Such a reduction promotion,
which meant improvement of the releasing ability of the lattice
oxygen, might improve the WGS performance. A preliminary study
showed that the reaction mechanism of the WGS reaction was Mars
van Krevelen using lattice oxygen by Eqs. (6) and (7) [24]:
catalysts at a steady-state was as follows; 19.5% (99.5%) for pH = 1,
65.9% (98.3%) for pH = 2.5, 61.7% (98.2%) for pH = 3, 16.2% (100%) for
pH = 5, 3.2% (100%) for pH = 7, 2.9% (100%) for pH = 9, 2.7% (100%)
for pH = 11 and 22.0% (99.8%) for pH = 13. The value in parenthesis
was the selectivity for CO2, which was not shown in Fig. 6. The
selectivity for CO2 was almost the same over these catalysts. In
terms of the activity, the optimum pH value of the filtration liquid
was 2.5 for producing a high WGS performance. The reason for such
an optimization will be discussed in the next section.
3.4. Relationship between WGS performance and Pd state of the
filtration liquid
For elucidating the reason for the high performance obtained
over the catalyst prepared using the filtration liquid of PdCl2, TEM
measurements were carried out. Fig. 7(a) and (b) shows TEM images
of the as-made and the spent Pd/LaCoO3 catalyst, respectively.
Here, the spent catalyst denotes the catalyst after a 190-min WGS
reaction. In Fig. 7(a), the Pd species were not clearly determined.
The size of Pd on the as-made catalyst might be less than 1 nm, or
Pd might be incorporated into the B-site of the LaCoO3. As for the
spentcatalystin Fig. 7(b), therewas a portionindicatedby the arrow
whose color was darker than the LaCoO3 layer. The supported Pd
might be the darker portion because heavier elements such as Pd
appeared darker due to scattering of the electrons in the catalyst in
the TEM measurements. The supported Pd was sintered during the
reaction and its particle size grew to about 2–5 nm. Although the
CO + O2− → CO2 + Vox + 2e−
(6)
(7)
lat
H2O + Vox + 2e− → H2 + O2−
lat
In these equations, O2− and Vox denote the lattice oxygen and lat-
lat
tice vacancy in the catalyst, respectively. The key point of the WGS
performance enhancement was improvement of the mobility of the
lattice oxygen. Hence, the supported Pd produced a high mobility
of the lattice oxygen and enhanced the WGS performance.
100
80
60
40
20
0
In addition, as shown in Fig. 4, a smaller weight loss was
observed over the spent catalyst as compared to the as-made
Pd/LaCoO3. The imbalance of the release rate (Eq. (6)) and the
regeneration rate of the lattice oxygen (Eq. (7)) was considered to
produce a lattice vacancy in LaCoO3. In other words, the release rate
of the lattice oxygen was higher than its regeneration rate. There-
fore, the lattice oxygen diffused from the bulk to the surface of the
LaCoO3 catalyst, and a lattice vacancy was subsequently produced.
The synergistic effect as already mentioned could be controlled
by changing the pH of the filtration liquid. Excluding pH = 2.5, the
adjustment of the filtration liquid pH was carried out by adding
NaOH aq. or HNO3 aq. to the original filtration liquid. Fig. 5 shows a
photograph of the filtration liquids with different pHs. By increasing
the pH of the filtration liquid from 2.5 to 9, precipitation of the
Pd component was found to proceed in the solution. On the other
hand, when the pH further increased from 9 to 13, the precipitated
Pd component was dissolved again. The Pd-supporting structured
catalyst was prepared with these solutions, and the activity test
was performed. Fig. 6 shows the CO conversion over the catalysts
using the filtrate of pH = 1, 2.5 (the original filtration liquid), 3, 5, 7,
9, 11 and 13. The result showed that the CO conversion over these
0
50
100
150
200
Time on stream / min
Fig. 6. Effect of pH of the filtration liquid on CO conversion at 573 K: (♦) pH = 1, (᭹)
pH = 2.5 (original), (ꢁ) pH = 3, (ꢂ) pH = 5, (ꢀ) pH = 7, (ꢃ) pH = 9, (ꢁ) pH = 11 and (ꢄ)
pH = 13.

80
R. Watanabe et al. / Applied Catalysis A: General 477 (2014) 75–82
Fig. 7. TEM images of (a) the as-made Pd/LaCoO3 catalyst and (b) the spent Pd/LaCoO3 catalyst.
particle growth proceeded, the particle size was found to be quite
small after the reaction. For investigating the advantage of the pro-
posed method, we compared the different catalysts prepared by
the impregnation method. Supporting data 1 shows TEM images
of the as-made catalyst prepared by the impregnation method. In
supporting data 1, particles indicated by the arrow were considered
to be Pd. Its diameter was larger than 6 nm. Comparing the particle
size of the catalyst prepared with the filtration liquid to that of the
catalyst prepared by the impregnation method, the smaller Pd was
obtained over the catalyst prepared with the filtration liquid. This
proposed method using the filtration method had the advantage of
Pd particle size.
The Pd species in the filtration liquid was identified by a UV–vis
measurement. Fig. 8 shows the UV–vis spectra of the original filtra-
tion liquid and the solutions after several dip-coatings. From Fig. 8,
two peaks at 210 and 238 nm were observed, and these peaks were
found to decrease with the change in number of dip-coatings. These
two peaks were assigned to the complex of the [PdCl3(H2O)]− [31].
The decrease in the absorption peaks in the spectrum is due to
the decrease in the [PdCl3(H2O)]− concentration of the solution.
Namely, the [PdCl3(H2O)]− was loaded on the structured LaCoO3
catalyst. The PdCl2 is known to be hardly dissolved in H2O, however,
a trace amount of Pd was dissolved as [PdCl4]2−, [PdCl3(H2O)]− and
PdCl2(H2O)2. Also, the state of Pd was dependent on the concentra-
tion of the chloride ion and proton in the solution [32]. As shown
in Fig. 8, only the [PdCl3(H2O)]− species was observed in the filtra-
tion liquid in this study. The following reaction (Eq. (8)) proceeded
in the filtrate after a trace amount of PdCl2 was dissolved in the
solution as [PdCl4]2− [33].
[PdCl4]2− + H2O ꢀ [PdCl3(H2O)]− + Cl−
(8)
By the way, the pH of the filtrate of the PdCl2 slurry was 2.5. Braisted
et al. reported that the LaCoO3 support surface carried a net pos-
itive charge in the solution at a low pH [34]. Therefore, the Pd
loading on the support surface was considered to be driven by
an electrostatic interaction between the anionic charged precur-
sor of [PdCl3(H2O)]− and the oppositely charged support surface.
The electrostatic interaction might produce highly dispersed Pd on
the structured catalyst. Table 2 shows the concentrations of the
[PdCl3(H2O)]− in the filtrate of the PdCl2 slurry and after the dip-
coating (2 min, 5 times) on the structured LaCoO3 catalyst which
were measured by ICP. The concentration of the as-made filtrate
was 793 ppm, on the other hand, the concentration of the filtrate
after the dip-coating was 589 ppm. The decrease in the concentra-
tion of [PdCl3(H2O)]− proceeded by only immersing the structured
catalyst in the filtration liquid, which was due to the electrostatic
interaction between the Pd precursor and the support. The reason
for the optimum pH of the Pd precursor solution was considered
to be derived from the Pd state in the precursor solution. It was
reported that a Pd(OH)2 species precipitated by increasing the pH
of the filtration liquid as noted by the following reaction (Eq. (9))
[33].
[PdCl3(H2O)]− + H2O ꢀ Pd(OH)2 + 3Cl− + 2H+
(9)
There was no electrostatic interaction between Pd(OH)2 and the
cationically charged support surface, which might lead to a low
Pd dispersion state. It might cause a significantly lower activ-
ity over the catalyst prepared using the filtrate at a pH of more
than 3. In addition, the low CO conversion when pH = 1 might
be due to the chloride ion in the Pd precursor solution. The
solution (pH = 1) was prepared from the original Pd precursor
solution (pH = 2.5) by adjusting with HCl. Namely, the chloride
ion was more contained in the solution (pH = 1). To confirm the
effect of the chloride ion on the catalytic performance, the Pd
solution, which nearly 4-fold chloride ion was contained in, was
prepared. The prepared catalyst was denoted as the Pd/LaCoO3
(Cl excess). Supporting data 2 shows the CO conversion over
the Pd/LaCoO3 catalyst prepared with the filtration liquid, which
1.2
210 nm
1.0
1
a
c
a: PdCl2 filtrate
b: Dip 1time
c: Dip 2times
d: Dip 3times
e: Dip 4times
f: Dip 5times
b
0.8
0.6
0.4
0.2
0
d
e
f
238 nm
Table 2
Concentration of the [PdCl3(H2O)]− in the filtrate of the PdCl2 slurry and after the
dip-coating.
200 220 240 260 280 300
Wavelength / nm
Condition
Concentration/ppm
As-made filtrate
Filtrate after 5 dippings for 2 min
793
589
Fig. 8. UV–vis spectra of the PdCl2 filtrate and solutions after the dip-coating.

R. Watanabe et al. / Applied Catalysis A: General 477 (2014) 75–82
81
(a)
(b)
C
200
200
ICl/IPd= 0.15
ICl/IPd= 0.11
Washing by H2O
NaOH treatment
NH3 treatment
After
reaction
ICl/IPd= 0.10
ICl/IPd= 0.51
As-made
Before
calcination
As-made
204
202
200
198
196
204
202
200
198
196
Binding energy / eV
Binding energy / eV
Fig. 9. XPS spectra of Cl2p for the Pd/LaCoO3 catalyst (a) in each step and (b) with washing treatment of basic solution and/or water.
was abbreviated as the Pd/LaCoO3 (original), and the Pd/LaCoO3
(Cl excess) catalyst. The CO conversion at a steady-state was 65.9%
for the Pd/LaCoO3 (original) catalyst and 6.6% for the Pd/LaCoO3
(Cl excess) catalyst. Such a result indicated that the chloride ion
inhibited the catalytic activity for WGS reaction. The activity sup-
pression by chloride ion was reported by other researchers. Oh et al.
reported that the dual role of Cl− in suppressing the CO preferential
oxidation (CO + 1/2O2 → CO2) activity over the Au/Al2O3 catalyst
was accelerating the agglomeration of Au particles and poisoning
the active sites [35]. The work of Lin et al. suggested that the low
catalytic activities for CO oxidation in the Au/TiO2 catalyst with high
chloride contents may also be due to an inhibitive effect [36]. They
prepared catalysts by introducing AuCl3 on TiO2 using the incipi-
ent wetness method and observed that the activity for the Au/TiO2
reduced with H2 at 500 ◦C was 40-fold higher than those reduced
at 200 ◦C. The improvement in catalytic property was accompanied
by a threefold decrease in the residual Cl− content of the catalyst
but no significant change in the particle size of Au. In our results,
the residual Cl− had the impact on the catalytic properties, which
might be due to the poisoning the active sites. The low CO con-
version when pH = 1 would be due to the chloride ion in the Pd
precursor solution.
and treated catalysts. From Fig. 9(b), the chloride ion was detected
and the remaining chloride ion of the as-made catalyst was partially
eliminated on the support surface, based on the decline of the Cl2p
peak intensity. From the XPS spectra of Pd3d over the untreated
and treated catalysts, Pd was confirmed to be the divalent state as
shown in supporting data 3. The Pd3d5/2 and Pd3d3/2 spin-orbital
splitting photoelectrons for these catalysts were located at bind-
ing energies of 336.9 and 342.4 eV, which were assigned to Pd2+
.
For evaluating whether or not the chloride ion was eliminated, the
ratio of the intensity for Cl2p- to Pd3d-level photoelectron peaks
(ICl/IPd), which reflected the atomic ratio of Cl to Pd elements on
the surface layers, was calculated. These values are also shown in
Fig. 9(b). The values of ICl/IPd were 0.51 for the catalyst without
treatment, 0.10 for the NH3-treated catalyst, 0.11 for the NaOH-
treated catalyst and 0.15 for the water-washed catalyst. The water
or base treatment was found to be effective for elimination of the
chloride ion on the catalyst.
Fig. 10 shows the CO conversion over the untreated catalyst and
the treated catalysts. The value in parentheses is the CO2 selectiv-
ity at a steady-state. The product except for CO2 was CH4 over the
untreated and the treated catalysts. By washing the Pd-supported
catalyst, the CO conversion was improved as follows: 83.6% at
10 min for washing with water; 82.0% for the NH3 and 85.4% for
3.5. Effect of post-treatment of the Pd-supported catalyst on
performance
100
80
The [PdCl3(H2O)]− of the Pd precursor contains chloride ions in
its Pd complex. Generally, the presence of chloride ions causes a
metal particle growth and deactivates the active site for the CO
adsorption [35,37]. The XPS spectra of Cl2p over the Pd/LaCoO3
catalyst before calcination, as-made and after the reaction were col-
lected for elucidating whether or not chloride ions were present.
Fig. 9(a) shows the XPS spectra of these catalysts. Surface chloride
ion was detected over these catalysts. The chloride ion remained
and seemed not to be affected by the thermal treatment on the
support surface, although the calcination was performed at the
high temperature of 823 K for 2 h to eliminate the chloride ion on
the support surface. Xu et al. reported that the remaining chlo-
rine on the catalyst surface could be removed by washing [38]. A
washing procedure was then applied to remove the chloride lig-
ands from the [PdCl3(H2O)]−. For the chloride removal, a washing
procedure by a basic agent was performed because the presence
of a basic agent would permit the substitution of the chloride ions
by hydroxyl groups. The washing agents were aqueous solutions of
NH3 and NaOH. Additionally, the effect of washing with water was
also examined. Fig. 9(b) shows the Cl2p XPS spectra of the untreated
60
Untreated catalyst
(98.3)
Washing by H2O (97.6)
40
20
0
NH3 treatment
NaOH treatment
(96.9)
(97.6)
0
50
100
150
200
Time on stream / min
Fig. 10. CO conversion over the Pd/LaCoO3 structured catalysts: (᭹) untreated, (♦)
with H2O washing, (ꢅ) with NaOH aq. washing (pH = 11.5), (ꢃ) with NH3 aq. washing
(pH = 11.5).

82
R. Watanabe et al. / Applied Catalysis A: General 477 (2014) 75–82
the NaOH. Especially, the CO conversion over the NH3-treated cat-
alyst increased with time on stream and a 91.3% conversion was
obtained at 190 min. Post-treatment by NH3 was the most effec-
tive for improvement of the WGS activity. The CO2 selectivity over
the treated catalysts was as follows; 96.9% over the NH3-treated
catalyst, 97.6% over the NaOH-treated catalyst and 97.6% over the
water-washed catalyst. Compared with the CO2 selectivity of 98.3%
over the untreated structured catalyst, the treated catalyst showed
a slightly lower selectivity. Such a slightly lower selectivity was due
to the production of methane which was produced from the unre-
acted CO and the produced hydrogen. Although a trace amount of
chloride ion remained on the support surface, the washing process
had a positive effect for further improving the WGS performance
over the Pd-supported structured catalyst.
In summary, the structured LaCoO3 catalyst prepared with
utilizing a singular precursor of [PdCl3(H2O)]− showed high perfor-
mance for the WGS reaction. Additionally, the Pd loading method
was a very simple, which was only dipping in the precursor solu-
tion with a few times, and the resulting Pd particle size was a quite
small. Further enhancement of the catalytic performance would
be expected by the optimization of washing treatment with NH3
for eliminating the chloride ligand. Its simple dip-coating method
and washing procedure might contribute to the effective loading of
active metals on the structured catalyst.
References
[1] S. Ahmed, M. Krumpelt, Int. J. Hydrogen Energy 26 (2001) 291–301.
[2] G.G. Scherer, Solid State Ionics 94 (1997) 249–257.
[3] T.V. Choudhary, D.W. Goodman, Catal. Today 77 (2002) 65–78.
[4] S. Oki, R. Mezaki, J. Phys. Chem. 77 (1973) 1601–1605.
[5] S. Oki, R. Mezaki, Ind. Eng. Chem. Res. 27 (1988) 15–21.
[6] M. Tinkle, J.A. Dumesic, J. Catal. 103 (1987) 65–78.
[7] C. Rhodes, B.P. Williams, F. King, G.J. Hutchings, Catal. Commun. 3 (2002)
381–384.
[8] Y. Lei, N.W. Cant, D.L. Trimm, J. Catal. 239 (2006) 227–236.
[9] A.F. Ghenciu, Curr. Opin. Solid State Mater. Sci. 6 (2002) 389–399.
[10] M.J.L. Ginés, N. Amadeo, M. Laborde, C.R. Apesteguía, Appl. Catal. A: Gen. 131
(1995) 283–296.
[11] O. Ilinich, W. Ruettinger, X. Liu, R. Farrauto, J. Catal. 247 (2007) 112–118.
[12] P.J. Guo, L.F. Chen, G.B. Yu, Y. Zhu, M.H. Qiao, H.L. Xu, K.N. Fan, Catal. Commun.
10 (2009) 1252–1256.
[13] T. Shishido, M. Yamamoto, D. Li, Y. Tian, H. Morioka, M. Honda, T. Sano, K.
Takehira, Appl. Catal. A: Gen. 303 (2006) 62–71.
[14] T. Shishido, M. Yamamoto, I. Atake, D. Li, Y. Tian, H. Morioka, M. Honda, T. Sano,
K. Takehira, J. Mol. Catal. A: Chem. 253 (2006) 270–278.
[15] H. Yahiro, K. Murawaki, K. Saiki, T. Yamamoto, H. Yamaura, Catal. Today 126
(2007) 436–440.
[16] Y. Sekine, T. Chihara, R. Watanabe, Y. Sakamoto, M. Matsukata, E. Kikuchi, Catal.
Lett. 140 (2010) 184–188.
[17] Y. Zhai, D. Pierre, R. Si, W. Deng, P. Ferrin, A.U. Nilekar, G. Peng, J.A. Herron,
D.C. Bell, H. Saltsburg, M. Mavrikakis, M. Flytzani-Stephanopoulos, Science 329
(2010) 1633–1636.
[18] S. Nishimura, T. Shishido, K. Ebitani, K. Teramura, T. Tanaka, Appl. Catal. A 387
(2010) 185–194.
[19] M. Haruta, S. Tsubota, T. Kobayashi, H. Kageyama, M.J. Genet, B. Delmon, J. Catal.
144 (1993) 175–192.
[20] J.D. Lessard, I. Valsamakis, M. Flytzani-Stephanopoulos, Chem. Commun. 48
(2012) 4857–4859.
4. Conclusion
[21] D. Andreeva, V. Idakiev, T. Tabakova, A. Andreev, J. Catal. 158 (1996)
354–355.
[22] Q. Fu, A. Weber, M. Flytzani-Stephanopoulos, Catal. Lett. 77 (2001) 87–95.
[23] Y. Sekine, H. Takamatsu, S. Aramaki, K. Ichishima, M. Takada, M. Matsukata, E.
Kikuchi, Appl. Catal. A 352 (2009) 214–222.
[24] R. Watanabe, Y. Sekine, H. Takamatsu, Y. Sakamoto, S. Aramaki, M. Matsukata,
E. Kikuchi, Top. Catal. 53 (2010) 621–628.
[25] C. Fukuhara, H. Ohkura, K. Gonohe, A. Igarashi, Appl. Catal. Gen. 279 (2005)
195–203.
[26] C. Fukuhara, Y. Kamata, A. Igarashi, Appl. Catal. Gen. 330 (2007) 108–116.
[27] V. Meille, Appl. Catal. Gen. 315 (2006) 1–17.
[28] Y. Zhu, R. Tan, T. Yi, S. Ji, X. Ye, Li. Cao, J. Mater. Sci. 35 (2000) 5415–5420.
[29] N. Yi, Y. Cao, Y. Su, W.-L. Dai, H.-Y. He, K.-N. Fan, J. Catal. 230 (2005)
249–253.
[30] G.L. Chiarello, J.-D. Grunwaldt, D. Ferri, F. Krumeich, C. Oliva, L. Forni, A. Baiker,
J. Catal. 252 (2007) 127–136.
[31] C.T. Tait, D.R. Janecky, P.S.Z. Rogers, Geochim. Cosmochim. Acta 55 (1991)
1253–1264.
The novel structured LaCoO3 perovskite-type catalyst was
developed for effectively converting CO via the WGS reaction.
Among the various Pd-supported LaCoO3 catalysts, the catalyst
prepared by the filtration liquid of the PdCl2 slurry showed the
highest activity. High activity was due to the electrostatic inter-
action between [PdCl3(H2O)]− as the anionic charged precursor
and the oppositely charged LaCoO3 support surface, which pro-
duced highly dispersed state of Pd on the support. For removing
the chloride ligand from the [PdCl3(H2O)]−, washing procedures
of the Pd supported catalyst were applied using water, NH3 and
NaOH. Consequently, the catalytic performance was improved by
all washing processes. Especially, the post-treatment by NH3 was
the most effective for improvement of WGS performance.
[32] J.-F. Boily, T.M. Seward, Geochim. Cosmochim. Acta 69 (2005) 3773–3789.
[33] O.B. Bel’skaya, T.I. Gulyaeva, A.B. Arbuzov, V.K. Duplyakin, V.A. Likholobov,
Kinet. Catal. 51 (2010) 105–112.
Acknowledgement
[34] M.P. Braisted, L.M. Cirjak, D.L. Feke, I. Manas-Zloczower, Part. Part. Syst. Charact.
23 (2006) 480–483.
[35] H.-S. Oh, J.H. Yang, C.K. Costello, Y.M. Wang, S.R. Bare, H.H. Kung, M.C. Kung, J.
Catal. 210 (2002) 375–386.
[36] S.D. Lin, M. Bollinger, M.A. Vannice, Catal. Lett. 17 (1993) 245–262.
[37] C.K. Costello, M.C. Kung, H.-S. Oh, Y. Wang, H.H. Kung, Appl. Catal. Gen. 232
(2002) 159–168.
This work was supported in part by JST, A-Step
(AS242Z02697M). We greatly thank Prof. Shigeki Tsukui (Osaka
Prefecture University) for ICP measurement.
Appendix A. Supplementary data
[38] Q. Xu, K.C.C. Kharas, A.K. Datye, Catal. Lett. 85 (2003) 229–235.
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/j.apcata.
2014.03.005.