2
48
E. Kono et al. / Applied Catalysis A: General 489 (2015) 247–254
and K enhanced the redox properties of Fe O3 and increased the
X-ray photoelectron spectroscopy (XPS) measurements were
conducted using a PHI-5000 VersaProbe II (ULVAC-Phi Inc.), with
25 W Al K␣ emission as the X-ray source, to measure each elec-
tron state on the catalyst surface. After reaction, the catalysts were
treated in the reaction condition for 185 min and cooled to room
temperature in Ar purge. After reaction, the samples were carried
by a transfer-vessel filled with N2 to avoid air exposure. Binding
energies were calibrated with a C 1s peak at 284.8 eV.
2
WGS activity. The addition of Pd and K promoted the reduction
of the metal oxide and the re-oxidation of the lattice oxygen by
H O [18]. The Pd/K/Co O catalyst showed higher activity than
2
3
4
the Pd/K/Fe O catalysts [19]. Additionally, the activities did not
2
3
depend on the Pd loading weight. Catalysts that load potassium
exceeding 0.78 wt% have high activity for the WGS reaction. Fur-
thermore, the addition of potassium inhibited methane formation.
Pd/K/Co O shows a different trend from that of Pd/K/Fe O . We
To observe the fine structures of Pd, K, and Co, the X-ray
absorption fine structure (XAFS) measurement was conducted
at the BL14B2 station at SPring-8 (Hyogo, Japan). Each edge
was measured using a transmission method. The catalyst was
pressed into a pellet (7 mm). Pellets used for measuring Co K-
edge were diluted with BN, treated in the reaction condition
(H2O:CO:H2:N2:Ar = 30:6:42:13:9) for 65 min and cooled to room
temperature in Ar purge. Then, the pellet was packed into a gas-
barrier bag in N2 atmosphere. XAFS data were analyzed using
3
4
2
3
investigated the reasons why the Pd/K/Co O catalyst has high
3
4
activity for the WGS reaction by catalytic activity tests and charac-
terization including TPR, XPS, and XAFS. The reaction mechanism
was ascertained using Q-Mass and DRIFTS.
2
. Experimental
2.1. Catalyst preparation method
3
software (REX2000; Rigaku Corp.). Fourier transformation of k -
The Pd/K/Co O catalysts were prepared using an impregna-
weighted EXAFS spectra was obtained in the k-range of 0.3–1.2 nm.
3
4
tion method. A commercial Co O (Kanto Chemical Co. Inc.) was
3
4
used as a catalyst support. Potassium-supported catalysts were
prepared by impregnation with an aqueous solution of precur-
sor salt, K CO (Kanto Chemical Co. Inc.). The resulting slurry was
2.4. Transient response experiment
2
3
dried on a hot plate under continuous stirring with subsequent
calcination at 773 K for 1 h. Following that, Pd was impregnated
on potassium-supported catalysts with an acetone solution of Pd
Transient response experiments were conducted using a
quadrupole mass spectroscopy (Q-Mass; QGA; Hiden Analyti-
cal Ltd.). The gas compositions were the following: reaction
gas, H2O:CO:H2:He:Ar = 16.7:3.3:23.3:48.9:7.8, total flow rate of
(
OCOCH ) (Kanto Chemical Co. Inc.). The resulting slurry was dried
3 2
−
1
on a hot plate under continuous stirring with subsequent calcina-
tion at 773 K for 1 h. The loading amount of potassium was fixed to
300 mL min ; purge gas, He:Ar = 276.6:23.3, total flow rate of
300 mL min−1.
0
.78 wt%. The loading of Pd was 0.27, 0.53, 1.1, or 2.1 wt%. For the
Isotope exchange reaction experiments were conducted to
determine the contribution of lattice oxygen to the WGS reac-
standard impregnation, potassium was impregnated first and then
Pd was impregnated. For reverse impregnation, Pd was impreg-
nated first.
−
1
tion. After heating of the catalyst to 573 K (5 K min ) in He purge,
the reaction gas was supplied at 573 K for 1 h. Then, 30% H218
O
To investigate the structure of K and Pd as active sites, other cat-
alysts were prepared using a physical mixture of 2.2 wt%Pd/Co O
(purity 98%; Taiyo Nippon Sanso Corp.) was supplied for 10 min
to exchange the lattice oxygen in/on the catalyst. The reaction gas
was supplied again to detect the product gas. To observe the lattice
3
4
and 1.56 wt%K/Co O . Each was prepared using the impregnation
3
4
18
method, and mixed together at a 1:1 ratio using a planetary ball
mill in a dry condition, or in 2-propanol solvent.
oxygen behavior during reaction, the reaction gas containing H2
O
16
18
(H2 O:H2 O:CO:H2:He:Ar = 11.2:5.5:3.3:23.3:48.9:7.8, total flow
−
1
rate of 300 mL min ) was supplied at 573 K for 1 h. The product
gas was detected using a Q-Mass. Nine parent peaks were scanned
using a mass spectrometer: m/e = 2 (H ), 15 (CH ), 18 (H O), 20
2.2. Catalytic activity tests
2
4
2
18
16
18
16
16 18
Catalytic activity tests were performed in a fixed bed flow reac-
(H2 O), 28 (C O), 30 (C O), 44 (C O2), 46 (C
(C18O2).
O
O), and 48
tor at atmospheric pressure. A Pyrex glass tube with 8 mm outer
diameter and 6 mm inner diameter was used as a reactor. The cat-
alyst (250–500 m particle size) with an amount of 40 or 80 mg,
was charged in the reactor. Reaction conditions were the following:
2.5. DRIFT measurements
−
1
H O:CO:H :N :Ar = 30:6:42:13:9; total flow rate = 178 mL min
;
2
2
2
−
1
W/F = 1.50 or 2.99 g-cat h mol ; reaction temperature at 573 K. The
product gas was analyzed using a GC-FID/TCD (GC-8A; Shimadzu
Corp.).
Adsorption species and reaction intermediates were inves-
tigated using diffuse reflectance infrared Fourier transform
spectroscopy (DRIFT; FT/IR-4200; Jasco Corp.). The experiment
−
1
−1
conditions were 600–4000 cm measurement range, 1 cm mea-
−
1
2.3. Characterization of catalysts
surement interval, 4 cm
limit of resolution, and 64 times
cumulated numbers. A peak that appeared from 3500–4000 cm
−
1
The oxidation–reduction property of catalysts during reaction
was attributed to H O. This peak was corrected with a background
2
was investigated using thermogravimetry (TG; TGA-50; Shimadzu
Corp.). The catalyst weight loss was measured under heating to
peak that was measured before CO intake. The total gas flow was
100 mL min . All catalysts were pre-reduced. The catalysts were
−
1
−
1
5
73 K, which is the reaction temperature (20 K min , in Ar purge).
When the temperature reached 573 K, hydrogen, CO or a sim-
ulated reaction gas (H O:CO:H :Ar = 7.2:2.4:16.8:73.6; total flow
heated to 673 K for 1 h in Ar purge, and were reduced by 10% H for
2
30 min at 473 K.
The spectrum for CO adsorption was measured to observe on
the surface adsorbed species of the catalyst. After the catalyst was
reduced, 3% CO gas was supplied at 298 K. Then the spectrum was
measured. The spectrum was measured until no spectrum change
was observed.
2
2
−1
rate = 100 mL min ) was supplied. The catalyst weight was 20 mg
for each case.
Metal dispersions of Pd were measured using CO chemisorption
(
BEL-CAT-55; BEL Japan Inc.). The stoichiometric factor of CO/Pd
was assumed as 1.0 for the calculation of metal dispersion. All cat-
alysts were reduced in hydrogen at 473 K for 30 min and cooled to
To identify the peaks in each spectrum, H–D exchange measure-
ments were conducted. After the catalyst was reduced and cooled
to 298 K, the spectrum was measured under 3% CO gas flow for
3
23 K in He flow. Then, 10% CO/He was dosed 20 times in He flow.