Platinum/Apatite Water-Gas Shift Catalysts

Article abstract of DOI:10.1021/acscatal.5b01909

Water-gas shift (WGS) micro and membrane reactors are interesting components for compact H₂ production and purification devices, but they require catalysts with very high activity for optimum efficiency to minimize catalyst bed thickness and mass transfer limitations. On the other hand, activation of H₂O is known to be more challenging than CO in this reaction. Catalysts comprising ca. 2 nm large Pt particles on hydrophilic apatites are found to have very high WGS activity, with specific reaction rates exceeding those of a highly active Pt/CeO₂ catalyst by up to 50% at 573 K. These apatite-supported catalysts exhibit stable CO conversions at 673 K without showing any CH₄ formation tendencies up to 723 K. WGS activity increases with Ca/P ratio in the apatite, leveling off around Ca/P ≈ 1.75, and formate has been identified as the main reaction intermediate. The outstanding WGS performance is attributed to the superior activation of H₂O on these ionic oxides due to coordination of H₂O to Lewis acidic Ca²⁺ ions and H bonding to basic O atoms of PO₄^3- units. This renders H₂O molecules highly polarized and thus reactive on apatite surfaces with the ensuing formate-like intermediates being well stabilized through bonding to multiple Ca²⁺ ions, as well. Thus, apatites provide an intriguing alternative to increasingly expensive rare-earth oxides in high-performance noble-metal WGS catalysts not only for micro and membrane reactors.

Full text of DOI:10.1021/acscatal.5b01909

Research Article
pubs.acs.org/acscatalysis
Platinum/Apatite Water-Gas Shift Catalysts
Dengyun Miao, Andreas Goldbach,* and Hengyong Xu
Dalian National Laboratory for Clean Energy, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, 457 Zhongshan
Road, 116023 Dalian, People’s Republic of China
*
S Supporting Information
ABSTRACT: Water-gas shift (WGS) micro and membrane reactors are interesting
components for compact H production and purification devices, but they require
2
catalysts with very high activity for optimum efficiency to minimize catalyst bed thickness
and mass transfer limitations. On the other hand, activation of H O is known to be more
2
challenging than CO in this reaction. Catalysts comprising ca. 2 nm large Pt particles on
hydrophilic apatites are found to have very high WGS activity, with specific reaction rates
exceeding those of a highly active Pt/CeO catalyst by up to 50% at 573 K. These apatite-
2
supported catalysts exhibit stable CO conversions at 673 K without showing any CH4
formation tendencies up to 723 K. WGS activity increases with Ca/P ratio in the apatite,
leveling off around Ca/P ≈ 1.75, and formate has been identified as the main reaction
intermediate. The outstanding WGS performance is attributed to the superior activation
of H O on these ionic oxides due to coordination of H O to Lewis acidic Ca ions and H bonding to basic O atoms of PO43−
2
+
2
2
units. This renders H O molecules highly polarized and thus reactive on apatite surfaces with the ensuing formate-like
2
2+
intermediates being well stabilized through bonding to multiple Ca ions, as well. Thus, apatites provide an intriguing alternative
to increasingly expensive rare-earth oxides in high-performance noble-metal WGS catalysts not only for micro and membrane
reactors.
KEYWORDS: high-temperature water-gas shift, H O activation, formate intermediates, ionic oxides, reaction rate, activation energy
2
INTRODUCTION
expensive, so that current research aims at reducing the
13
■
amounts of both in catalyst formulations.
The water-gas shift (WGS) reaction (CO + H O ⇆ CO + H ,
2
2
2
−1
We are now investigating the HTS activity and stability of Pt
supported on hydroxyapatite (Ca (PO ) (OH) , HAP)
ΔH = −41.1 kJ mol ) is an important industrial process which
has attracted renewed attention in recent years in connection
with on the spot hydrogen production for fuel cells in
transportation and residential applications. Industrially it is
carried out in two stages: i.e., a high-temperature shift (HTS)
over Fe−Cr catalysts at 593−723 K followed by a low-
temperature shift (LTS) over Cu−Zn−Al catalysts at 473−523
1
0
4
6
2
because Au and Ru deposited on this irreducible ionic oxide
14
showed LTS activity before. The Au catalyst had been more
active than the Ru catalyst up to 593 K, but high-temperature
WGS activity and stability of the Au catalyst was not reported
while the Ru catalyst showed serious methanation activity
14
1
above 573 K. Apatites and related compounds are intriguing
support materials for catalysts because of their structural
stability, their hydrophilicity due to amphoteric acid and base
functionalities, and the possibility of cationic and anionic
K. However, these conventional catalysts are unsuitable for
compact hydrogen generation systems as desired for fuel cell
applications because of their weight and volume, long start-up
times, durability under steady-state and transient conditions,
15−27
2
,3
isomorphous substitutions.
For example, hydroxide and
and susceptibility to condensation and poisoning. Therefore,
precious metals supported on ceria and other reducible oxides
have been widely investigated, in particular as LTS catalysts for
phosphate ions were found to stabilize Au nanoparticles against
15
sintering at elevated temperatures. We focused on Pt catalysts
in our study because of their generally superior high-
2
−7
fuel cell applications.
Highly active WGS catalysts are also
28
temperature stability in comparison to e.g. Au WGS catalysts.
The catalytic activity of apatite catalysts can be substantially
required for achieving optimum efficiency with HTS micro-
8
9
reactors and Pd membrane reactors. For example, computa-
tional simulations show that an optimum membrane would
improve the efficiency of WGS Pd membrane reactors only by
16−19
modulated by changing the Ca/P ratio.
This results in
nonstoichiometric formulations Ca10−Z(HPO ) (PO )6−Z-
4
Z
4
10
(OH)2−Z with 0 ≤ Z ≤ 1, i.e. 1.50 ≤ Ca/P ≤ 1.67.
1
2%, whereas 76% could be gained by optimizing the catalyst.
Operation temperatures above 623 K are especially favorable in
the latter case, because hydrogen-selective Pd membranes are
significantly inhibited by CO below that threshold. However,
many precious-metal catalysts on reducible supports are not
sufficiently stable and progressively deactivate under typical
reformer outlet conditions. On the other hand, cost is a major
concern, since noble metals and rare-earth elements are
Hydroxyapatites with larger than stoichiometric Ca/P ratio
(
1.67) have been also reported where excess Ca2+ ions may be
20
11
charge balanced by carbonate or other foreign anions.
Received: August 28, 2015
Revised: December 20, 2015
12
©
XXXX American Chemical Society
775
DOI: 10.1021/acscatal.5b01909
ACS Catal. 2016, 6, 775−783

ACS Catalysis
Research Article
Therefore, we have explored the interdependence between
apatite stoichiometry and WGS activity of such Pt catalysts by
preparing HAP formulations with Ca/P ratios ranging between
determined from the desorption branch of the isotherms.
Platinum dispersions were derived from H2 and CO
chemisorption measurements at 323 K using an AutoChem
2910 system (Micromeritics) assuming that the molar ratio of
1
.58 and 2.00, and a fluoroapatite (FAP) form where hydroxide
ions were replaced by fluoride ions. In the following we will
adsorbed H to exposed Pt is 0.5 and 1 for adsorbed CO to
2
show that catalysts can be designed in this way that feature
exposed Pt. Each catalyst (50 mg) was first reduced in 10% H2/
Ar at 573 K for 1 h, heated to 698 K under Ar (He) and kept at
that temperature for 1 h, and then cooled to 323 K in Ar (He).
Chemisorption measurements were performed at that temper-
HTS activities exceeding those of Pt on CeO and other
2
reducible oxides and at the same time exhibit very good HTS
stability.
ature by injecting a series of 10% H /Ar (5% CO/He) pulses
2
EXPERIMENTAL SECTION
until saturation adsorption was reached. Transmission electron
microscopy (TEM) images were obtained on a JEM-2100
system (JEOL) with an acceleration voltage of 200 kV. The
samples were ultrasonically suspended in ethanol and placed
onto a carbon film supported over a Cu grid for that purpose.
Fourier transform infrared (FTIR) spectra were recorded in
■
Catalyst Preparation. All chemicals were of analytical
grade, were used without further purification, and were
procured from Kermel unless otherwise noted. Hydroxyapatite
1
5,23
powders were synthesized by a precipitation method
using
4
aqueous solutions of Ca(NO ) ·4H O and (NH ) HPO as
3
2
2
4 2
−1
−1
the 400−4000 cm range at a resolution of 4 cm on a Bruker
Tensor 27 spectrometer equipped with a deuterated triglycine
sulfate (DTGS) detector. Carefully ground catalyst samples
were diluted with KBr (mass ratio 1:100) and pressed into self-
supporting wafers (13 mm diameter) for ex situ measurements
in air. Undiluted catalysts were pressed into self-supporting
wafers, and the DTGS detector was replaced by a mercury−
cadmium−telluride (MCT) detector for in situ FTIR
spectroscopic analyses under WGS conditions using a quartz
cell equipped with CaF windows. Wafers were reduced in H
precursors and aqueous ammonia (25%) for adjustment of the
pH value of the solutions between 9 and 10.5. Typically 100
−1
mL of a 0.20 mol L (NH ) HPO aqueous solution (e.g., at
4
2
4
−
1
pH 9) was added dropwise to 100 mL of 0.33 mol L
Ca(NO ) aqueous solution (e.g., at pH 9) with vigorous
3
2
magnetic stirring at room temperature. The resulting
suspension was kept in a water bath for 2 h at 363 K and
then cooled to room temperature and aged for 6 h. Next the
precipitate was centrifuged, washed repeatedly with purified
water until pH ∼7, dried at 333 K for 12 h, and finally calcined
for 4 h in air at 753 K in a muffle furnace. The obtained HAP
powders were denoted HAP-9, HAP-9.5, HAP-10, and HAP-
2
2
at 573 K for 0.5 h and then purged in N at 723 K for 0.5 h and
2
finally cooled to 523 K in N before in situ measurements with
2
a 1.9% CO/3.2% H O/He mixture that had been produced by
1
0.5 according to the pH of the synthesis solutions. Further
2
bubbling a 2% CO/He mixture through water at room
temperature.
HAP-9 and HAP-10 powders with modified Ca content were
made by wet impregnation of the calcined materials with
Ca(NO ) solution. Those samples were denoted Ca-HAP-9
X-ray photoelectron spectroscopy (XPS) experiments were
carried out on a Thermo ESCALAB 250Xi instrument to
determine the molar ratios of elements at the catalyst surface
using Al Kα radiation (hν = 1486.6 eV). The C 1s peak was set
at 284.8 eV and taken as reference for binding energy
calibration. Temperature-programmed reduction (TPR) was
carried out in a homemade system furnished with a thermal
conductivity (TCD) detector. All TPR samples were pretreated
in Ar at 473 K for 2 h, cooled to room temperature, and then
3
2
and Ca-HAP-10, respectively. Fluoride-substituted apatite
denoted as FAP) was prepared by adding 0.247 g of NH4F
to the (NH ) HPO solution before combination with the
(
4
2
4
Ca(NO ) solution (both at pH 10). The remaining procedural
3
2
steps were the same as for preparation of the HAP powders.
Catalysts with nominally 1 wt % Pt were prepared by
impregnation of the calcined apatite powders. Appropriate
−1
amounts of 15.8 mg mL aqueous H PtCl (Sinopharm
2
6
−1
Chemical Reagent Co.) were added, and suspensions were kept
at room temperature for about 24 h before samples were dried
at 333 K for 12 h and calcined again for 4 h in air at 753 K in a
muffle furnace. For comparison, 1 wt % Pt/CaO, Pt/CaCO3,
Pt/γ-Al O and Pt/CeO catalysts were prepared by incipient
heated in 5% H
was calibrated by reducing 5 mg of CuO to Cu for
quantification of H consumption during TPR. Carbon
/Ar at a rate of 10 K min . The TCD detector
2
2
deposition after catalytic testing was probed by temperature-
programmed oxidation (TPO) in a home-built setup equipped
with a Pfeiffer OmniStar mass spectrometer (MS) to detect CO
2
3
2
wetness impregnation with aqueous H PtCl in addition.
2
6
Preparation of the Pt/CeO catalyst has been described
(28), CO (44), O (32), and H O (18). Values in parentheses
2 2 2
2
29
elsewhere in detail. CeO was obtained by calcination of
are monitored mass/charge ratios (m/z).
2
Ce(NO ) ·6H O (Sinopharm) at 723 K for 4 h in a muffle
Catalytic Testing. WGS activity tests were carried out in a
fixed-bed quartz reactor with a diameter of 8 mm at
atmospheric pressure and temperatures between 523 and 723
K. The calcined catalysts were ground to powder, pressed into a
pellet at 20 MPa, crushed, and sieved. The particle size used for
catalytic testing was 40−60 mesh (250−380 μm particles).
Typically 75 mg of catalyst was diluted with 100 mg of quartz
sand of the same particle size range and loaded into the quartz
reactor. The height of the catalyst bed was about 5 mm. The
reactor was heated under He to 523 K before the reactants were
introduced. Water was injected into a preheater with a
calibrated double-plunger pump (Elite P230), where it was
vaporized and mixed with the dry feed gas components. The
gas mixture for product-free activity tests consisted of 5% CO,
3
3
2
furnace. γ-Al O3 was prepared by calcination of pseudo
2
boehmite (Shandong Aluminum Co.) at 773 K for 4 h in a
muffle furnace.
Catalyst Characterization. X-ray diffraction (XRD)
patterns of the catalyst powders were recorded on a PANalytic
X’Pert Pro-1 instrument using Cu Kα radiation (λ = 0.1541
nm). The bulk Ca/P ratio and Pt content were analyzed with
an inductively coupled plasma optical emission spectrometer
(
ICP-OES, PerkinElmer Optima 7300DV) after the catalyst
samples had been digested with aqua regia at room temper-
ature. The specific surface area was analyzed by N adsorption
2
at 77 K using a QuadraSorb SI instrument (Quantachrome). All
samples were used as powders and degassed at 573 K for 6 h
under vacuum before N adsorption and pore sizes were
20% H O, and 75% He. Those measurements were carried out
2
2
7
76
DOI: 10.1021/acscatal.5b01909
ACS Catal. 2016, 6, 775−783

ACS Catalysis
Research Article
Table 1. Characteristic Properties of Catalysts
Ca/P molar ratio
a
SBET (m g−1)
b
2
c
d
d
sample
pH
Pt (wt %)
dispersion (%)
dPt (nm)
ICP
XPS
Pt/HAP-10.5
10.5
71
0.77
100
1.1
1.76
e
e
8
6
1.3
Pt/HAP-10
Pt/HAP-9.5
Pt/HAP-9
Pt/FAP
10
9.5
9
71
85
77
73
83
0.74
0.78
0.82
0.78
0.74
96
73
1.2
1.6
1.4
1.1
0.8
1.75
1.68
1.58
1.75
1.64
1.51
1.45
81
10
101
Pt/CeO2
150
e
e
5
8
1.9
Pt/Ca-HAP-9
Pt/Ca-HAP-10
9
0.73
0.79
1.74
2.00
10
a
b
c
d
e
Apatite synthesis pH. Support surface area (without Pt). ICP-OES analysis. H chemisorption. CO chemisorption.
2
at GHSV = 150000 mL gcat−1 h−1. A reformate type gas mixture
reported in the literature. Hydroxyapatite with stoichiometric
composition (Ca/P = 1.67) formed at pH 9.5. The Ca/P ratio
did not exceed 1.75 in the precipitated apatite materials, but it
was increased up to 2.00 via subsequent impregnation of as-
synthesized materials with Ca(NO ) solutions. The surface
18
(
15% CO, 5% CO , 30% H O, 40% H , 10% N ) was used for
2 2 2 2
determination of specific reaction rates r and a long-term
stability test. The reactor tail gas was passed through a cooling
trap, where excess steam was condensed. The dry tail gas was
then fed to a gas chromatograph (Shimadzu GC-8A) equipped
with a TCD for CO, CO , and CH analysis.
3
2
Ca/P ratios resulting from XPS analyses (Table 1) were smaller
than the bulk values, in agreement with previous studies.
18,19,26
2
4
The 100−160 mesh fraction (96−150 μm particle size) of
the catalysts was diluted with quartz sand particles in the same
size range (100−160 mesh) and used for measurement of
specific reaction rates. The amount of catalyst was varied (20−
Textural properties of the apatite materials were determined
prior to impregnation with Pt. All adsorption−desorption
isotherms belong to type IV (Figure S2a in the Supporting
Information). The specific surface area S ranged between 71
BET
2
−1
200 mg) to limit CO conversion to 20% at maximum. The
and 85 m g (Table 1). Similar S
values have been
BET
15,22
height of the catalyst bed was about 8 mm. Samples were
subjected to 12 h of conditioning in the product-free feed
mixture at 673 K before these measurements. At each reaction
temperature the tail gas was analyzed at different reaction times
to calculate CO conversion. No methane was observed during
reported in the literature.
Substitution of hydroxide with
fluoride ions did not have an obvious impact on the surface
area. Interparticle pore size distributions (Figure S2b) were in
19−25
the typical range reported for apatite-supported catalysts
although noticeably reduced in FAP (5−20 nm) in comparison
to the HAP materials (10−30 nm). The noble-metal content of
all Pt/apatite catalysts was somewhat less than the targeted
amount, albeit very similar (0.73−0.82%, Table 1) according to
ICP-OES analyses. TEM showed that Pt was well dispersed on
the apatite catalysts (Figure S3 in the Supporting Information)
with most noble-metal clusters being smaller than 2 nm, in
good agreement with mean particle diameters and dispersions
measurement of the specific rates. Conversion of CO (XCO
)
was calculated according to
XCO = [
CO] − [CO]
in
out
[CO]in
(1)
where [CO] and [CO]out are CO concentrations in feed and
in
tail gas, respectively. Thermodynamic equilibrium WGS CO
conversions XCO,eq were calculated in an analogous manner by
replacing [CO]out in eq 1 with WGS equilibrium concentrations
derived from H and CO chemisorption (Table 1).
2
Unfortunately, analysis of Pt particle size was not as
straightforward for the Pt/CeO catalyst, although from the
2
[
CO]eq that were computed with the program Gaseq (http://
absence of Pt XRD peaks it is clear that Pt aggregates were
www.arcl02.dsl.pipex.com/).
smaller than ∼4 nm there as well. Platinum particles could not
30
be detected by TEM due to a lack of contrast between them
RESULTS
■
and the CeO
well suited for determining metal dispersions on reducible
oxides such as CeO because concomitant surface reduction of
the latter insinuates often exceedingly high dispersion values as
2
support. Hydrogen chemisorption is also not
Catalyst Characteristics. Table 1 summarizes physico-
chemical properties of the prepared apatite and Pt/apatite
materials. All diffraction patterns (Figure S1 in the Supporting
Information) correspond with the apatite structure (JCPDS 01-
2
3
0
for the present Pt/CeO2 catalyst (Table 1). Carbon
monoxide chemisorption yielded 58% Pt dispersion, but that
method is not very accurate either when dealing with reducible
oxides: Reactive CO adsorption at surface hydroxyl groups can
1
8,19
0
74-0566).
No diffraction peaks resulting from Pt or its
oxides were observed for any Pt/apatite sample, indicating that
the particle size of Pt species remained below the XRD
detection limit (ca. 4 nm). Apatite crystallinity improved with
increasing pH of the synthesis solution according to the width
and resolution of the Bragg reflections. Note that mixed-phase
structures including HAP precipitated at pH ≤ 8.5. Lattice
parameters a and c of the Pt/HAP-X and Pt/FAP samples
matched very well with those of standard apatite patterns
31
lead to overestimation of metal dispersion, while strong metal
support interaction in such materials can result in much too
33
32
large Pt particle size estimates as we previously experienced.
Thus, we consider the 1.9 nm resulting from CO chemisorption
as an upper limit for Pt particle size with the lower limit of 1.1
nm following from 100% Pt dispersion. Accordingly, the mean
(
Table S1 in the Supporting Information). Chemical analyses
confirmed that the Ca/P molar ratio of the apatite materials
Table 1) increased with increasing synthesis pH value, as
Pt particle size was not very different if at all on Pt/CeO from
that on the apatite-supported catalysts. Figure 1 and Figure S4
in the Supporting Information show FTIR spectra of the Pt/
2
(
7
77
DOI: 10.1021/acscatal.5b01909
ACS Catal. 2016, 6, 775−783

ACS Catalysis
Research Article
Figure 1. Infrared spectra of Pt/HAP-10, Pt/HAP-9, Pt/Ca-HAP-9,
and Pt/FAP catalysts.
Figure 2. Temperature-programmed reduction spectra of HAP, FAP,
and corresponding Pt catalysts.
HAP-X and Pt/FAP materials which are rather similar and are
in good agreement with reported apatite IR spectra. The bands
Pt. However, quantitative analysis of H uptake (Table S2 in
the Supporting Information) would suggest rather high H2
adsorption on the reduced Pt of the catalysts investigated here
2
−1
at 576, 603, 962, 1041, and 1095 cm are due to bending and
3
−
20,34
stretching vibrational modes of the PO4
tetrahedra.
(H /Pt ≈ 1). On the other hand, similar TPR patterns with two
2
These bands are considerably broadened in the samples
prepared at pH 9 and 9.5 (Figure S4), which hints at structural
disorder and is likely related to the lower crystallinity of these
materials (Figure S1 in the Supporting Information).
strong H consumption features have been frequently observed
2
for Pt supported on other irreducible support materials and
were typically attributed to Pt oxides that interact weakly (ca.
4
1−43
5
23 K) and strongly (ca. 713−753 K) with the support.
3−
Interestingly, the broadening of the PO4 bands disappeared
after impregnation of Pt/HAP-9 with additional Ca2+ ions. In
fact, the FTIR spectrum of Ca-impregnated Pt/Ca-HAP-9 was
practically identical with that of Pt/HAP-10 (Figure 1). Weak
Therefore, we attribute the two strong H consumption peaks
2
of our Pt/apatite materials also to reduction of PtO species
x
interacting weakly and strongly with the support. Reduction of
Pt oxides supported on the apatite materials is apparently
−1
4
1
bands at 1450 and 1413 cm can be attributed to carbonate
somewhat facilitated with respect to e.g. γ-Al O and HZSM-
2
3
2
0,35
42,43
species
which might have originated from CO adsorption
2
5.
Furthermore, the low-temperature H consumption with
a maximum below 373 K is most certainly due to Pt species,
2
4
+
during preparation or subsequent handling in air. The band at
1
36
1
385 cm− can be assigned to nitrite species resulting from
as TPR of PtO yields a reduction peak centered at 368−378
2
44
incomplete decomposition of residual nitrate from the synthesis
solution during calcination. Broad bands at 3446 and 1637
K. The quantitative H consumption analysis also suggests the
2
presence of highly oxidized Pt species, as total H /Pt ratios
2
−
1
37
−
cm are due to adsorbed water. The characteristic OH
ranged between 1.5 and 2 for most catalysts (Table S2).
Catalytic Reaction. Figure 3 shows CO conversion over
Pt/HAP-X catalysts as a function of temperature in the
product-free reaction mixture. Carbon dioxide was the only
carbonaceous product over these catalysts: i.e., no CH4
formation was observed. Catalyst activity increased with Ca/P
ratio, with XCO reaching equilibrium conversion over Pt/HAP-
2
0,38
−1
librational and stretching bands
at 632 and 3572 cm− are
missing in the Pt/FAP spectrum. This indicates that OH ions
inside the channels of the HAP structure can be completely
−
substituted by F ions, in agreement with literature
3
4,39,40
reports.
Figure 2 shows H TPR spectra of Pt/HAP-10 and Pt/FAP
2
and the corresponding noble-metal-free materials. There was
no appreciable H consumption up to 973 K in case of the
2
fluoroapatite and also HAP-9 and HAP-9.5 (Figure S5 in the
Supporting Information), while the other apatite materials
yielded weak reduction peaks around 723 K (773 K for HAP-
1
0.5; Figure S5) and stronger peaks around 923 K. In general,
Pt reduction started around 323 K, yielding a first maximum or
shoulder around 358−366 K and then two strong H2
consumption peaks around 483 K and 673−713 K, although
the weak initial maximum was missing and the second strong
reduction feature shifted to lower temperatures (608−678 K)
in the case of HAP-9 and HAP-9.5. These patterns are similar
to those reported by Takarroumt et al. for a series of Pt/HAP
catalysts with Ca/P ≈ 1.60 and 0.42−2.40% Pt, although the
second strong Pt reduction peak above 573 K was observed
24
only at higher Pt loadings. They attributed H consumption
below that temperature to reduction of Pt oxides and Pt ions
2
2+
on the surface and within the apatite lattice and H uptake at
higher temperatures to adsorption of hydrogen on the reduced
2
Figure 3. CO conversion over Pt/apatite and Pt/CeO catalysts.
2
7
78
DOI: 10.1021/acscatal.5b01909
ACS Catal. 2016, 6, 775−783

ACS Catalysis
0 and Pt/HAP-10.5 at 673 K. Moreover, CO conversion over
Research Article
7
−1
1
Thinon et al. reported a somewhat lower value of 65 kJ mol
these two catalysts exceeded significantly that over a highly
for a 2% Pt/CeO catalyst that exhibited also a rather moderate
2
WGS-active Pt/CeO2 catalyst except at the lowest test
specific WGS reaction rate (Table 2). This shows that
precursor materials and preparation methods are very
important for the activity of WGS catalysts supported on
reducible oxides where metal−support interactions play a
significant role.
Figure 4 and the tabulated reaction rates confirm that all Pt/
apatite catalysts are more WGS active than our Pt/CeO2
catalyst except for Pt/HAP-9 and Pt/HAP-9.5. In fact, the
specific reaction rates are very similar for the catalysts
supported on apatites with Ca/P ≥ 1.75 and exceed that of
the ceria-supported catalyst by ca. 50% in the case of Pt/FAP
and Pt/Ca-HAP-10. This assessment also stands when TOFs
are considered. It is unclear whether turnover frequencies over
the apatite catalysts are indeed lower than those over our Pt/
temperature. This previously reported Pt/CeO catalyst had
2
29
the same nominal noble-metal loading. In contrast, Pt/HAP-9
and Pt/HAP-9.5 were much less WGS active and XCO did not
reach equilibrium even at 723 K. Interestingly, after the Ca/P
ratio of Pt/HAP-9 was raised to the same level as that of Pt/
32
HAP-10 via impregnation with Ca(NO3) solution, CO
2
conversion over Pt/Ca-HAP-9 approached that over Pt/HAP-
1
0 over the entire temperature range (Figure S6 in the
Supporting Information). This corroborates that the Ca/P ratio
is important for WGS activity of Pt supported on
hydroxyapatites even though addition of further Ca to HAP-
10 did not trigger a further substantial XCO enhancement
(
Figure S6, Pt/Ca-HAP-10). Figure 3 shows in addition that
CO conversion over Pt/FAP was similar to that over the Pt/
CeO catalyst because of the uncertainty associated with Pt
2
CeO reference catalyst. Note that HAP-10 did not exhibit any
2
dispersion on CeO (vide supra). We employed the lowest
2
WGS activity in the range from 523 to 723 K under these test
conditions. Furthermore, Pt supported on CaO and CaCO3
yielded rather low CO conversion in this temperature range
plausible dispersion for calculation so that the tabulated TOF is
the upper limit for this Pt/CeO catalyst. Even if correct, that
2
value only would signal that larger Pt particles on CeO are
2
(
Figure S7 in the Supporting Information). Finally, we tested
more WGS active than smaller particles on apatite supports.
However, a larger Pt fraction will be inaccessible in larger
particles, reducing the utilization efficiency of the precious
metal. Therefore, from a practical (i.e., cost) point of view it is
more interesting to consider reaction rates normalized to Pt
amount, which are included in Table 2 as well. These also
facilitate comparisons with literature catalysts, as they account
for differing noble-metal contents. From these normalized
reaction rates it is obvious that the best apatite-supported
catalysts rank among the top Pt WGS catalysts reported so far.
the WGS activity of nominally 1% Pt on another irreducible
support, i.e. γ-Al O , which also yielded much smaller X
2
3
CO
values than all of the Pt/apatite catalysts except Pt/HAP-9 and
Pt/HAP-9.5 (Figure S7).
In order to assess the WGS activity of the apatite-supported
catalysts quantitatively, we measured specific reaction rates in
the reformate-type mixture between 573 and 673 K. Figure 4
Only a strongly sodium promoted Pt/TiO catalyst gave a
2
45
higher specific WGS rate at 573 K. The latter catalyst was
tested in a rather diluted mixture with no CO2 present,
however.
Reaction Intermediates. Figure 5 shows in situ FTIR
spectra of the Pt/HAP-10.5 catalysts recorded at 523 K under
N and a WGS mixture consisting of 1.9% CO and 3.2% H O
2
2
in He. The spectrum under an N atmosphere is practically
2
identical with that reported by Canepa et al. for hydroxyapatite
5
1
−1
nanoparticles. Accordingly, the bands in the 1300−1600 cm
range are due to carbonate species, the band centered at 1630
cm originates from H O molecules (bending mode) on the
surface, and the group of bands between 1900 and 2250 cm
−1
2
−1
are overtones and combinations of the fundamental phosphate
modes below 1250 cm− (cf. Figure 1). In addition there is a
1
Figure 4. WGS rates in reformate-type feed gas in the temperature
range 573−673 K.
broad and asymmetric adsorption spreading out from 2250 to
3
800 cm− which arises from superposition of stretching modes
1
of surface hydroxy groups and H O molecules involved in H
shows Arrhenius plots of the determined reaction rates, from
which activation energies Eact were derived. These are given in
Table 2 together with the specific reaction rates at 573 K and
corresponding turnover frequencies (TOFs) calculated from Pt
dispersion (Table 1). Reaction rates varied only marginally with
space velocity (Figure S8 in the Supporting Information),
demonstrating that diffusion within the catalyst bed had
negligible impact under our kinetic test conditions. The table
also includes such data of Pt WGS catalysts from the
2
51
bonding. Several additional bands emerged after switching to
−1
the WGS atmosphere. The intense band at 2051 cm can be
readily assigned to CO bonded to Pt on the basis of literature
−
1
52
(
2059 cm ). On the other hand, bands at 1349, 1378, 1588,
−1
and 2858 cm agree very well with the symmetric (ν (CO))
s
and asymmetric C−O stretching (ν (CO)) and the C−H
as
bending (δ(CH)) and stretching (ν(CH)) modes reported for
−
1
formate anions in aqueous solution (ν
(CO) 1351 cm
,
s
7
,13,45−49
−1
−1
53
literature
that were tested in product-containing
δ(CH) 1383 cm , νas(CO) 1585 cm , ν(CH) 2803) and icy
matter (ν (CO) 1350 cm , δ(CH) 1380 cm , νas(CO) 1580
cm ). Furthermore, the band at 1756 cm matches well
with the stretching mode reported for gaseous formic acid at
ν(CO) 1767 cm . Thus, formate appears to be the key
−1
−1
(
CO , H ) reaction mixtures. The WGS activation energies of
s
2
2
−1
54
−1
the apatite-supported catalysts are at the upper end of the range
reported for those literature catalysts but are close to the 82 kJ
mol found for the Pt/CeO catalyst prepared here. Note that
−1
−1 55
2
7
79
DOI: 10.1021/acscatal.5b01909
ACS Catal. 2016, 6, 775−783

ACS Catalysis
Research Article
Table 2. Specific WGS Reaction Rates r of Pt-Based Catalysts at 573 K and Atmospheric Pressure
1
r (molCO molPt−1 s−1
TOF (s−1
Eact (kJ mol−1
catalyst
Pt/HAP-10.5
Pt (wt %)
r (μmolCO gcat− s−1
)
)
)
)
ref
0.77
0.74
0.78
0.82
0.78
0.79
0.73
0.74
2
21.6
21.2
2.9
0.548
0.558
0.072
0.55
0.58
0.10
0.05
0.60
83
82
90
89
81
87
83
82
65
70
58
23
44
80
this work
Pt/HAP-10
Pt/HAP-9.5
Pt/HAP-9
this work
this work
a
a
1.7
0.040
this work
Pt/FAP
24.0
24.9
16.5
16.0
0.596
0.617
0.444
0.424
0.146
0.585
0.260
0.400
0.078
1.143
0.055
0.064
0.048
0.151
0.248
0.490
0.470
0.211
this work
Pt/Ca-HAP-10
Pt/Ca-HAP-9
Pt/CeO2
this work
this work
b
0.73
this work
c
Pt/CeO2
15
7
c
Pt/CeO /Al O
3
0.9
27
7
2
2
c
Pt/ZrO2
1.5
20
7
c
Pt/TiO2
1.9
39
7
c
Pt/Fe O3
1.5
6
7
2
d
Pt-4Na-TiO2
Pt/MgO-CeO2
Pt/6VCeO2
Pt/ZrO2
1
58.6
3.82
45
46
47
48
48
13
13
49
49
d
1
2.8
e
1.15
0.95
0.97
0.5
3.8
86
f
2.33
0.41
0.43
0.33
0.65
f
Pt/Ti/ZrO2
Pt/CeO2
7.51
g
6.3
g
Pt/Ca0.5Ce0.5O1.5
0.5
12.5
h
i
Pt/Ce(La)O NaCN leached
3.06
73.7
75
x
h
i
Pt/Ce(La)Ox
4.15
45
75
a
b
c
d
Rate at 623 K. Calculated for 58% Pt dispersion. Reactant composition: 10% CO, 20% H O, 10% CO , 30% H , 30% Ar. Reactant composition:
2
2
2
e
f
2
.83% CO, 5.66% H O, 37.74% H , 53.77% He. Reactant composition: 6.0% CO, 60.0% H O, 1.6% CO , 16.0% H , 0.4% CH , 16.0% N . Reactant
2 2 2 2 2 4 2
g
composition: 7.0% CO, 22.0% H O, 8.5% CO , 37% H , 25.5% N . Reactant composition: 3.0% CO, 26.2% H , 29.9% H , 3.7% N , 37.2% He. Pt
dispersion was estimated at 75% assuming 1.5 nm hemispherical particles according to an established method.
atomic Pt/Ce/La ratio assuming CeO and La O . Reactant composition: 11% CO, 26% H O, 7% CO , 26% H , 26% He.
2
2
2
2
2
2
2
2
9,50 h
Pt content calculated from bulk
i
2
2
3
2
2
2
Figure 5. In situ FTIR spectra of Pt/HAP-10.5 catalyst under an N2
and WGS atmosphere at 523 K.
Figure 6. Stability of Pt/HAP-10.5 and Pt/CeO2 catalysts in
reformate-type feed gas at 673 K.
2
−1
intermediate in the WGS reaction over apatite-supported Pt
catalysts.
Catalyst Stability. Stability of the Pt/HAP-10.5 and Pt/
Information), while S
had decreased from 78 m g to 53
BET
2
−1
m g with a concomitant increase of the average pore size
from 10 to 16 nm (Figure S10 in the Supporting Information).
This may indicate blocking of the smaller support pores. TPO
CeO catalysts was tested in the reformate-type mixture at 673
2
K (Figure 6). CO conversion dropped somewhat over both
catalysts during the first 24 h but then remained constant at
experiments yielded a weak CO signal (m/z 44) between 523
and 823 K, but this was again practically identical with that
from a fresh catalyst sample (Figure S11 in the Supporting
2
∼52% for Pt/HAP-10.5 and ∼47% for Pt/CeO2 for the
remainder of the 100 h experiment. This demonstrates that the
apatite-supported catalyst retained its superior WGS perform-
ance during continuous operation. It was characterized by
XRD, BET, TEM, and TPO measurements after this test. The
XRD pattern of the stability-tested sample was identical with
that of the fresh sample (Figure S9 in the Supporting
Information). In addition, the H O signals (m/z 18) were
2
rather similar before and after WGS testing (Figure S11). Thus,
CO2 evolution was likely due to contraction of airborne
carbonaceous contaminants during handling of the samples
under ambient atmosphere, and coke formation remained
below the detection level of our facilities after 100 h operation
7
80
DOI: 10.1021/acscatal.5b01909
ACS Catal. 2016, 6, 775−783

ACS Catalysis
Research Article
of Pt/HAP-10.5 in reformate. Finally, Figure 7 shows TEM
images and Pt particle size distributions of the fresh and used
surface for H O remains nearly as high as that of the (010)
surface.
2
60
Combining IR spectroscopy and microgravimetry, it has been
inferred that H O molecules are coordinated approximately in a
2
2+
58
1
:1 ratio to surface Ca ions in the first hydration layer. The
computational modeling provided more detailed insight into
the interaction between H O and HAP at the molecular level in
2
2+
addition. Unsaturated surface Ca cations act as Lewis acids to
which water molecules coordinate through their O atom, while
a rather strong H bond is simultaneously formed with a basic O
atom of a neighboring PO43 group.
−
60,61
Hence, the large H O
2
binding energies on HAP surfaces are due to a synergetic
bonding situation that renders the adsorbed H O molecules
2
highly polarized if not dissociated. In these states H O will be
2
very reactive toward electrophilic molecules such as CO. The
dual H O bonding also allows for a rationale for the observed
2
2+
WGS activity dependence on Ca/P ratio: Ca ions are essential
for anchoring H O, and thus their concentration determines
2
the density of these reactant molecules on the HAP surface in
the first place. The secondary H bonding requires a sufficient
number of accessible O atoms from PO4 units in the vicinity
of the cations, on the other hand. This sets an upper limit for
Figure 7. TEM images and Pt particle size distributions of fresh (a, c)
and used (b, d) Pt/HAP-10.5 catalyst samples.
3−
the number of fully activated H O molecules on the HAP
2
surface, which could explain why WGS activity levels off around
Ca/P ≈ 1.75. Note that the surface Ca/P ratio (Table 1)
approached the value of stoichiometric HAP (i.e., 1.67) at this
Pt/HAP-10.5 samples. The Pt particle size had increased
slightly from ca. 1 nm to ca. 2 nm after 100 h, but it remained
quite uniform. This minor Pt particle growth and decrease of
surface area might be responsible for the initial XCO decline
during the stability test, but the rapid stabilization of CO
conversion shows that these apatite catalysts are structurally
stable under typical WGS conditions at 673 K.
overall composition. Presumably, H O activation benefits from
2
an optimum balance between Ca2 and PO4 ions on the
+
3−
apatite surface. Activation of additional H O molecules is
2
2
+
probably not feasible via simultaneous coordination to Ca and
H bonding to PO43 ions above that boundary. Obviously this
−
implies that H O activation is a key step in the WGS reaction
on these Pt/apatite catalysts. Preliminary kinetic analyses reveal
2
DISCUSSION
■
The high-temperature WGS activity of Pt/apatite catalysts
clearly correlates with the Ca/P ratio of the support material.
Addition of Ca is known to promote the WGS conversion rates
that the H
10.5 (n = 0.81) than for Pt/CeO
indicating a greater influence of H
WGS catalysts as well.
The similarity of activation energies further suggests that the
rate-determining steps involve similar activated species on Pt
WGS catalysts supported on apatite and reducible oxides such
2
O reaction order n is noticeably larger for Pt/HAP-
(n = 0.55) at 623 K,
O on the apatite-supported
2
2
1
3,56
57
on Pt/CeO
and Pt/TiO2 significantly. Incorporation of
2
2+
Ca into ceria and titania lattices enhances the reducibility and
mobility of oxygen.
improves the mobility of O-bound reaction intermediates such
as formates, carbonates, and carboxylates and thus facilitates
their decomposition into CO and H . Kondarides and co-
workers attributed the higher reaction rates to the creation of
oxygen vacancies at the Pt support interface that are capable of
adsorption and activation of both CO and H O.
these scenarios can apply in the present case, since apatites are
not easily reduced and TPR peaks below 773 K can be readily
attributed to PtO species in our catalysts.
13,57
Linganiso et al. suggested that this
as CeO . The exact nature of the relevant intermediates and the
2
13
mechanistic route(s) for the WGS reaction are still a matter of
2
2
2
debate. Two general types of reaction pathways are under
discussion in the case of Pt/CeO : i.e., associative mechanisms
2
64
56,57
62,63
Neither of
involving e.g. formate
and/or carboxylate species and
2
regenerative redox routes involving oxygen transfer from the
65
support. Such a redox mechanism does not appear to be
viable for WGS on Pt/apatites, considering the irreducible
character of the support. Formate species have been frequently
x
The noble metal is generally thought to be responsible for
CO activation in supported WGS catalysts, while the oxide
activates H O. The intense Pt−CO band in the in situ FTIR
spectrum suggests that CO is activated on Pt in HAP catalysts
62,63
proposed
as crucial intermediates in associative mecha-
is
2
nisms, on the other hand, and their decomposition into CO
deemed to be rate-determining. The in situ FTIR analysis of
2
2
2
as well. It is also plausible that apatites assume the H O
the Pt/HAP-10.5 catalyst under the CO/H O/He atmosphere
2
2
activating role in the catalysts investigated here, given their very
large hydrophilicity. Ab initio calculations for the two most
(Figure 5) strongly suggests that formate is also the key
reaction intermediate in the WGS reaction over apatite-
supported catalysts.
58
important HAP surfaces confirmed that their affinity is high for
5
9,60
H O.
2
While H O adsorbs molecularly on the (001) surface,
Ab initio calculations have also been employed to investigate
2
51
it dissociates spontaneously on the (010) surface in a
the interaction of formic acid with HAP surfaces. Formic acid
was found to adsorb molecularly on the (001) surface with
barrierless, highly exothermal process (−250 to −320 kJ
−1
−1
mol ), giving rise to new surface terminations (Ca−OH and
binding energies of around 190 kJ mol but to dissociate
5
1,60
PO−H).
Calculated H O binding energies (−50 to −100
spontaneously on the water-reconstructed (010) surface with a
2
−
1
−1 51
kJ mol ) agree well with calorimetric adsorption enthalpies,
formate ion binding energy of ca. 250 kJ mol . Formate ions
revealing also that the affinity of the reconstructed (010)
can also form on the (001) surface, albeit with a somewhat
7
81
DOI: 10.1021/acscatal.5b01909
ACS Catal. 2016, 6, 775−783

ACS Catalysis
Research Article
−
1
51
lower binding energy (ca. 170 kJ mol ). This demonstrates
that apatite surfaces can stabilize formate ions rather well once
they are formed there. Formate acts as a bidentate ligand with
test at 673 K after a minor initial activity loss. The latter is likely
due to growth of mean Pt particle size from ca. 1 nm to ca. 2
nm, since TPO of a fresh and the stability tested sample yielded
2
+
its two O atoms coordinating to two separate Ca ions on the
001) HAP surface. The bonding situation is more complex
on the reconstructed (010) surface, where one formate O atom
bonds to two Ca ions. The other formate O atom bonds to a
third Ca2 ion and also forms a hydrogen bond with the nearby
H atom of hydroxide ions resulting from H O dissociation on
the (010) surface. Similar adducts between Pt-activated CO
minor and practically identical CO emission patterns. This
2
51
(
very favorable set of characteristics makes Pt/apatite catalysts
promising candidates for application in single-stage, high-
temperature WGS devices that require catalyst beds with large
spatial and minimal mass transfer limitations such as micro and
membrane reactors.
2+
+
2
51
and apatite-activated H O molecules can be readily envisioned
ASSOCIATED CONTENT
Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acscatal.5b01909.
2
■
as intermediates of the WGS reaction on apatite-supported
catalysts. The requirement of multiple Ca2+ surface ions for
formation of formate-like activated complexes may also play a
role in the large sensitivity of WGS reaction rates toward the
surface Ca/P ratio of apatites. Moreover, formate reaction
intermediates may explain why apatite-supported catalysts were
*
S
Additional catalyst characterization data (PDF)
less WGS active than the Pt/CeO catalyst below 523 K
AUTHOR INFORMATION
Corresponding Author
A.G.: e-mail, goldbach@dicp.ac.cn; tel, + 86 411 8437 9229.
2
■
(
Figure 3). Decomposition of intermediary formate adducts
could be hampered on Pt/HAP at lower temperatures because
*
of their better stabilization on apatite than on CeO surfaces.
2
Notes
Thus, the observed WGS activity trends of Pt/apatite catalysts
can be well understood in the framework of an associative
mechanism involving formate-like reaction intermediates.
Irrespective of mechanistic details, it is obvious that the
outstanding high-temperature WGS activity of optimized Pt/
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
Financial support by the Chinese Academy of Sciences through
the External Cooperation Program (Helmholtz-CAS Joint
Research Group on Integrated Catalytic Technologies for
Efficient Hydrogen Production, grant no. GJHZ1304) is
gratefully acknowledged.
apatite catalysts is due to the very efficient H O activation on
2
the ionic apatite surface. In fact, H O is more difficult to
2
activate than CO between the two WGS reactants, due to its
2
thermodynamic stability. The present results suggest that H O
2
activation is superior on apatites in comparison with CeO2,
where the WGS reaction is aided by facile redox cycling of the
latter material. The specific HTS reaction rate on the Pt/CeO2
catalyst tested here is among the highest for such Pt catalysts
supported on reducible oxides reported in the literature (Table
REFERENCES
■
(1) Sun, Y.; Hla, S. S.; Duffy, G. J.; Cousins, A. J.; French, D.;
Morpeth, L. D.; Edwards, J. H.; Roberts, D. G. Appl. Catal., A 2010,
390, 201−209.
2
). Therefore, the oxygen storage capacity of reducible oxides
(2) Azzam, K. G.; Babich, I. V.; Seshan, K.; Lefferts, L. J. Catal. 2007,
251, 153−162.
appears to be less important for improving the (high-
(
2
(
3) Kalamaras, C. M.; Dionysiou, D. D.; Efstathiou, A. M. ACS Catal.
temperature) WGS activity than efficient H O activation.
2
012, 2, 2729−2742.
4) Jacobs, G.; Chenu, E.; Patterson, P. M.; Williams, L.; Sparks, D.;
Thomas, G.; Davis, B. H. Appl. Catal., A 2004, 258, 203−214.
5) Panagiotopoulou, P.; Kondarides, D. I. J. Catal. 2004, 225, 327−
36.
6) Liu, X. S.; Ruettinger, W.; Xu, X. M.; Farrauto, R. Appl. Catal., B
005, 56, 69−75.
(7) Thinon, O.; Diehl, F.; Avenier, P.; Schuurman, Y. Catal. Today
Evidently the latter can be more readily achieved on suitable
ionic materials such as apatites. However, note that structural
−
OH ions of HAP are not involved in H O adsorption and
2
(
3
(
2
60
activation as H O moves away from those basic ions.
Therefore, it is not surprising that their substitution with F
2
−
ions has no effect on the WGS activity and reaction rates on
optimized Pt/FAP and Pt/HAP are practically the same.
2
008, 137, 29−35.
CONCLUSION
(8) Piermartini, P.; Schuhmann, T.; Pfeifer, P.; Schaub, G. Top. Catal.
■
2
011, 54, 967−976.
9) Bi, Y.; Xu, H.; Li, W.; Goldbach, A. Int. J. Hydrogen Energy 2009,
4, 2965−2971.
10) Ma, D. H.; Lund, C. R. F. Ind. Eng. Chem. Res. 2003, 42, 711−
17.
(11) Li, H.; Goldbach, A.; Li, W.; Xu, H. J. Membr. Sci. 2007, 299,
The water-gas shift activity of Pt/apatite catalysts is strongly
correlated with increasing Ca/P ratio in the support material,
until it levels off around Ca/P ≈ 1.75. The WGS activity of
such optimized catalysts exceeds that of a highly active Pt/
(
3
(
7
CeO catalyst prepared here with the same noble metal loading
2
above 523 K, with specific reaction rates being up to 50%
higher at 573 K. This extraordinary WGS performance is
130−137.
̈
(12) Zalc, J. M.; Sokolovskii, V.; Loffler, D. G. J. Catal. 2002, 206,
1
(
69−171.
attributed to superior activation of H O and stabilization of
2
13) Linganiso, L. Z.; Jacobs, G.; Azzam, K. G.; Graham, U. M.;
ensuing formate-like reaction intermediates on these hydro-
59,60
Davis, B. H.; Cronauer, D. C.; Kropf, A. J.; Marshall, C. L. Appl. Catal.,
A 2011, 394, 105−116.
philic, ionic oxides. Literature ab initio modeling studies
indicate that simultaneous coordination of H O to Lewis acidic
2
(
47.
14) Venugopal, A.; Scurrell, M. S. Appl. Catal., A 2003, 245, 137−
2+
3−
Ca ions and H bonding to basic O atoms of PO4 units
1
render H O molecules highly polarized and thus reactive on
2
(15) Zhao, K.; Qiao, B.; Wang, J.; Zhang, Y.; Zhang, T. Chem.
apatite surfaces. Moreover, the apatite-supported catalysts do
not exhibit any methanation activity up to 723 K but good
stability, as CO conversion remained constant during a 100 h
Commun. 2011, 47, 1779−1781.
(16) Ghantani, V. C.; Lomate, S. T.; Dongare, M. K.; Umbarkar, S. B.
Green Chem. 2013, 15, 1211−1217.
7
82
DOI: 10.1021/acscatal.5b01909
ACS Catal. 2016, 6, 775−783

ACS Catalysis
Research Article
(
17) Ogo, S.; Onda, A.; Iwasa, Y.; Hara, K.; Fukuoka, A.; Yanagisawa,
K. J. Catal. 2012, 296, 24−30.
18) Tsuchida, T.; Kubo, J.; Yoshioka, T.; Sakuma, S.; Takeguchi, T.;
Ueda, W. J. Catal. 2008, 259, 183−189.
19) Yan, B.; Tao, L. Z.; Liang, Y.; Xu, B. Q. ACS Catal. 2014, 4,
931−1943.
20) Domínguez, M. I.; Romero-Sarria, F.; Centeno, M. A.;
Odriozola, J. A. Appl. Catal., B 2009, 87, 245−251.
21) Phonthammachai, N.; Ziyi, Z.; Jun, G.; Fan, H. Y.; White, T. J.
Gold Bull. 2008, 41, 42−50.
22) Carvalho, D. C.; Pinheiro, L. G.; Campos, A.; Millet, E. R. C.; de
(51) Canepa, P.; Chiatti, F.; Corno, M.; Sakhno, Y.; Martra, G.;
Ugliengo, P. Phys. Chem. Chem. Phys. 2011, 13, 1099−1111.
(52) Kuriyama, M.; Tanaka, H.; Ito, S.; Kubota, T.; Miyao, T.; Naito,
S.; Tomishige, K.; Kunimori, K. J. Catal. 2007, 252, 39−48.
(53) Ito, K.; Bernstein, H. J. Can. J. Chem. 1956, 34, 170−178.
(54) Schutte, W. A.; Boogert, A. C. A.; Tielens, A. G. G. M.; Whittet,
D. C. B.; Gerakines, P. A.; Chiar, J. E.; Ehrenfreund, P.; Greenberg, J.
M.; van Dishoeck, E. F.; de Grauuw, T. Astron. Astrophys. 1999, 343,
(
(
1
(
9
66−976.
55) Maco
Fausto, R.; Ras
56) Panagiotopoulou, P.; Papavasiliou, J.; Avgouropoulos, G.;
Ioannides, T.; Kondarides, D. I. Chem. Eng. J. 2007, 134, 16−22.
57) Panagiotopoulou, P.; Kondarides, D. I. Appl. Catal., B 2011, 101,
38−746.
58) Bertinetti, L.; Tampieri, A.; Landi, E.; Ducati, C.; Midgley, P. A.;
Coluccia, S.; Martra, G. J. Phys. Chem. C 2007, 111, 4027−4035.
59) Astala, R.; Stott, M. J. Phys. Rev. B: Condens. Matter Mater. Phys.
008, 78, 075427.
60) Corno, M.; Busco, C.; Bolis, V.; Tosoni, S.; Ugliengo, P.
Langmuir 2009, 25, 2188−2198.
61) Bolis, V.; Busco, C.; Martra, G.; Bertinetti, L.; Sakhno, Y.;
Ugliengo, P.; Chiatti, F.; Corno, M.; Roveri, N. Philos. Trans. R. Soc., A
(
(
̧
̂
as, E. M. S.; Lundell, J.; Pettersson, M.; Khriachtchev, L.;
anen, M. J. Mol. Spectrosc. 2003, 219, 70−80.
̈
̈
(
(
Sousa, F. F.; Filho, J. M.; Saraiva, G. D.; da Silva Filho, E. C.; Fonseca,
M. G.; Oliveira, A. C. Appl. Catal., A 2014, 471, 39−49.
(
(
7
(
23) Huang, J.; Wang, L.-C.; Liu, Y.-M.; Cao, Y.; He, H.-Y.; Fan, K.-
N. Appl. Catal., B 2011, 101, 560−569.
24) Takarroumt, N.; Kacimi, M.; Bozon-Verduraz, F.; Liotta, L. F.;
Ziyad, M. J. Mol. Catal. A: Chem. 2013, 377, 42−50.
25) Gangarajula, Y.; Kedharnath, R.; Gopal, B. Appl. Catal., A 2015,
06, 100−108.
26) Sugiyama, S.; Miyamoto, T.; Hayashi, H.; Moffat, J. B. J. Mol.
(
(
2
(
(
5
(
Catal. A: Chem. 1998, 135, 199−208.
(
(
(
27) Kozlowski, J. T.; Davis, R. J. ACS Catal. 2013, 3, 1588−1600.
28) Luengnaruemitchai, A.; Osuwan, S.; Gulari, E. Catal. Commun.
2
(
(
012, 370, 1313−1336.
2
003, 4, 215−221.
29) Cavusoglu, G.; Miao, D.; Lichtenberg, H.; Carvalho, H. W. P.;
Xu, H.; Goldbach, A.; Grunwaldt, J.-D. Appl. Catal., A 2015, 504, 381−
90.
30) Perrichon, V.; Retailleau, L.; Bazin, P.; Daturi, M.; Lavalley, J. C.
Appl. Catal., A 2004, 260, 1−8.
31) Jacobs, G.; Ricote, S.; Graham, U. M.; Patterson, P. M.; Davis, B.
H. Catal. Today 2005, 106, 259−264.
32) Bernal, S.; Calvino, J. J.; Cauqui, M. A.; Gatica, J. M.; Lop
Cartes, C.; Perez Omil, J. A.; Pintado, J. M. Catal. Today 2003, 77,
85−406.
62) Shido, T.; Iwasawa, Y. J. Catal. 1993, 141, 71−81.
(
63) Jacobs, G.; Davis, B. H. Int. J. Hydrogen Energy 2010, 35, 3522−
3
536.
3
(
(
64) Grabow, L. C.; Gokhale, A. A.; Evans, S. T.; Dumesic, J. A.;
Mavrikakis, M. J. Phys. Chem. C 2008, 112, 4608−4617.
65) Gorte, R. J.; Zhao, S. Catal. Today 2005, 104, 18−24.
(
(
(
́
ez
́
3
(
(
33) Liu, B.; Goldbach, A.; Xu, H. Catal. Today 2011, 171, 304−311.
34) Zhang, Y.; Wang, J.; Yin, J.; Zhao, K.; Jin, C.; Huang, Y.; Jiang,
Z.; Zhang, T. J. Phys. Chem. C 2010, 114, 16443−16450.
(
35) Chen, W.; Huang, Z.; Liu, Y.; He, Q. Catal. Commun. 2008, 9,
5
(
(
16−521.
36) Lynn, A. K.; Bonfield, W. Acc. Chem. Res. 2005, 38, 202−207.
37) Lafon, J. P.; Champion, E.; Bernache-Assollant, D. J. Eur. Ceram.
Soc. 2008, 28, 139−147.
38) Yasukawa, A.; Kandori, K.; Ishikawa, T. Calcif. Tissue Int. 2003,
2, 243−250.
39) Boukha, Z.; Kacimi, M.; Pereira, M. F. R.; Faria, J. L.;
Figueiredo, J. L.; Ziyad, M. Appl. Catal., A 2007, 317, 299−309.
40) Boukha, Z.; Kacimi, M.; Ziyad, M.; Ensuque, A.; Bozon-
(
7
(
(
Verduraz, F. J. Mol. Catal. A: Chem. 2007, 270, 205−213.
(
(
41) Yu, C.; Ge, Q.; Xu, H.; Li, W. Catal. Lett. 2006, 112, 197−201.
42) Zhang, Y.; Zhou, Y.; Qiu, A.; Wang, Y.; Xu, Y.; Wu, P. Catal.
Commun. 2006, 7, 860−866.
43) Zhang, Y.; Zhou, Y.; Huang, L.; Zhou, S.; Sheng, X.; Wang, Q.;
Zhang, C. Chem. Eng. J. 2015, 270, 352−361.
44) Borgna, A.; Le Normand, F.; Garetto, T.; Apesteguia, C. R.;
Moraweck, B. Catal. Lett. 1992, 13, 175−188.
45) Zhu, X.; Shen, M.; Lobban, L. L.; Mallinson, R. G. J. Catal.
011, 278, 123−132.
46) Duarte de Farias, A. M.; Barandas, A. P. M. G.; Perez, R. F.;
Fraga, M. A. J. Power Sources 2007, 165, 854−860.
47) Duarte de Farias, A. M.; Bargiela, P.; Rocha, M. d. G. C.; Fraga,
M. A. J. Catal. 2008, 260, 93−102.
48) Hwang, K.-R.; Ihm, S.-K.; Park, S.-C.; Park, J.-S. Int. J. Hydrogen
Energy 2013, 38, 6044−6051.
49) Fu, Q.; Saltsburg, H.; Flytzani-Stephanopoulos, M. Science 2003,
01, 935−938.
50) Bergeret, G.; Gallezot, P. In Handbook of Heterogeneous
Catalysis; Ertl, G., Knozinger, H., Schuth, F., Weitkamp, J., Eds.;
Wiley-VCH: Weinheim, Germany, 2008; Vol. 3, pp 738−765.
(
(
(
2
(
(
(
(
3
(
̈
̈
7
83
DOI: 10.1021/acscatal.5b01909
ACS Catal. 2016, 6, 775−783