Ir-in-ceria : A highly selective catalyst for preferential CO oxidation

Article abstract of DOI:10.1016/j.jcat.2008.01.024

An Ir-in-ceria catalyst has been developed, in which most of the iridium particles are embedded in the ceria matrix through the redox reaction between Ce³⁺ and Ir⁴⁺, which occurred during co-precipitation (CP). This Ir-CP catalyst exhibited high activity for preferential CO oxidation under excess hydrogen conditions, and the selectivity to CO₂ remained nearly constant, at around 70%, with increasing reaction temperature. Temperature-programmed reduction and in situ diffuse reflectance infrared spectroscopy techniques were used to explore the structure of the Ir-CP catalyst and to correlate it with the catalytic performance. It was found that the CeO₂ support was activated by iridium and formed on the surface the active sites for CO oxidation. Due to the absence of extensively exposed Ir species on the Ir-CP catalyst, H₂ oxidation occurring on the Ir species and the ceria support at high temperatures was significantly suppressed, thus keeping the selectivity to CO₂ at a high level.

Full text of DOI:10.1016/j.jcat.2008.01.024

Journal of Catalysis 255 (2008) 144–152
www.elsevier.com/locate/jcat
“Ir-in-ceria”: A highly selective catalyst for preferential CO oxidation
a,b
a
a
a
c
a,∗
Yanqiang Huang , Aiqin Wang , Lin Li , Xiaodong Wang , Dangsheng Su , Tao Zhang
a
State Key Laboratory of Catalysis, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116023, People’s Republic of China
b
Graduate University of Chinese Academy of Sciences, Beijing 100049, People’s Republic of China
c
Fritz-Haber-Institut der Max-Planck-Gesellschaft, Faradayweg 4-6, 14195 Berlin, Germany
Received 14 January 2008; accepted 24 January 2008
Available online 6 March 2008
Abstract
An Ir-in-ceria catalyst has been developed, in which most of the iridium particles are embedded in the ceria matrix through the redox reaction
3+
4+
between Ce and Ir , which occurred during co-precipitation (CP). This Ir-CP catalyst exhibited high activity for preferential CO oxidation
under excess hydrogen conditions, and the selectivity to CO remained nearly constant, at around 70%, with increasing reaction temperature.
2
Temperature-programmed reduction and in situ diffuse reflectance infrared spectroscopy techniques were used to explore the structure of the
Ir-CP catalyst and to correlate it with the catalytic performance. It was found that the CeO support was activated by iridium and formed on the
2
surface the active sites for CO oxidation. Due to the absence of extensively exposed Ir species on the Ir-CP catalyst, H oxidation occurring on
2
the Ir species and the ceria support at high temperatures was significantly suppressed, thus keeping the selectivity to CO at a high level.
2
©
2008 Elsevier Inc. All rights reserved.
Keywords: Preferential CO oxidation; Iridium; Ceria; Selectivity; Ir/CeO
2
1
. Introduction
Polymer electrolyte membrane fuel cells (PEMFCs) oper-
ble metal (NM) catalysts exhibited better performance in the
range of 90–130 C compared with alumina-supported ones [5–
10]. A noncompetitive Langmuir–Hinshelwood mechanism has
been proposed to be responsible for the high activities of the
◦
ating with hydrogen have proven to be the most promising
candidates for transport applications [1]. The PEMFCs oper-
NM/CeO , in which the CO adsorbed on the NM reacts with
2
◦
ate at relatively low temperatures (∼80–120 C), and the anode
the oxygen activated on the ceria support. This mechanism in-
catalysts are easily poisoned by even traces of CO. Traditional
routes for hydrogen generation yield significant amounts of CO
volves the direct participation of the NM, which provides sites
for CO adsorption. Nevertheless, there is currently an ongoing
debate about the role of NM in these ceria-containing samples.
Hardacre et al. [11] observed enhanced rates of CO oxidation
on a Pt (111) single crystal encapsulated by ceria. Similarly,
Golunski et al. [12–14] found that the addition of the NM led to
a dramatic increase in the number of oxygen vacancies within
the ceria, which became active sites for CO oxidation. Recently,
Deng et al. [15] reported that Au/CeO2 catalysts can be effec-
tive for the PROX reaction even after removal of the neutral Au
clusters. In these examples, the activated ceria support, with-
out direct participation of the NM, can act as the active sites
for CO oxidation. A similar phenomenon also was found in the
WGS reaction; for example, the rate of the WGS reaction re-
mained the same when the metallic NM was removed [16] or
encapsulated in the ceria [17,18].
(ca. 5–15%) as the byproduct. A subsequent water–gas shift
(WGS) reaction can reduce the CO amount to about 0.5–1%
[2]. Thus, it is necessary to further reduce the amount of CO to
about 10 ppm while minimizing the consumption of hydrogen.
Several chemical and/or physical methods have been used to
obtain a high purification level of the hydrogen fuel. Among
these alternatives, preferential oxidation of CO under excess
hydrogen conditions (PROX) seems to be a promising process
[
3,4].
To date, various catalytic formulations have been investi-
gated in the PROX reaction. In particular, ceria-supported no-
*
Corresponding author. Fax: +86 411 84691570.
E-mail address: taozhang@dicp.ac.cn (T. Zhang).
0021-9517/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2008.01.024

Y. Huang et al. / Journal of Catalysis 255 (2008) 144–152
145
3
−1
Motivated by these experimental findings, we attempted to
design a highly selective PROX catalyst, in which the CO oxi-
dation reaction can proceed on the activated ceria surface alone
while the H2 oxidation occurring on the metal sites is signifi-
cantly suppressed. Toward this goal, we prepared an Ir-in-ceria
catalyst (with most of the Ir particles embedded in the ceria)
using co-precipitation (CP), as reported by Rajaram et al. [19].
The Ir-in-ceria was chosen due to the fact that Ir has been less
widely explored than Pt or Au, and also because Ir/CeO2 can
be as active as Pt/CeO2 for the PROX reaction [20]. For com-
parison, an Ir-on-ceria catalyst (with the Ir particles deposited
mainly on the ceria surface) also was prepared by deposition–
precipitation (DP). It was found that the difference in the struc-
tures of these two catalysts affected their catalytic performance,
especially in terms of selectivity behavior. Thus, the present
work provides an example of a structure–reactivity relation-
ship.
of 67 cm min (STP), corresponding to a space velocity of
−
1
−1
40,000 mL h
g
. In some cases, a feed stream composed of
Cat
2 vol% CO and 1 vol% O2 with He balance was used to evalu-
ate the activity for CO oxidation. H2O was introduced into the
system by passing the stream through a saturator to create the
required H2O concentration. The effluent gas was analyzed us-
ing an online gas chromatograph system (Angilent GC-8800)
equipped with a thermal conductivity detector. CO2 and H2O
were detected as the only products. No methane was found un-
der our experimental conditions.
CO conversion (XCO) is calculated as
in
CO
out
CO
n
− n
XCO =
× 100(%).
(1)
in
n
CO
Selectivity (S) to CO2 is defined as the ratio of the oxygen
transformed into CO2 to the total oxygen consumed:
out
n
CO2
S =
× 100(%).
)
(2)
in
out
2
(n − nO2
2
. Experimental
O2
The relation between conversion and selectivity is defined as
2
.1. Catalyst preparation
SXO2
XCO =
λ(%),
(3)
1
00
To prepare the Ir-in-ceria catalyst, an aqueous solution of
where λ is the oxygen excess factor, by definition,
H2IrCl6 and Ce(NO3)3 was added to a heated NaOH solu-
tion (80 C) to form a black precipitate, as described previ-
ously [19]. After aging for 2 h, the suspension was filtered
and washed with deionized water several times. Then the solid
material was dried at 60 C overnight and finally calcined at
◦
in
2n
O2
λ =
.
(4)
in
n
CO
◦
2
.3. Characterization
◦
4
00 C for 2 h. The resultant samples, designated Ir-CP, had Ir
metal loadings ranging from 0.20 to 5.00 wt%. For comparison,
Au-CeO2 and Pt-CeO2 catalysts with 1.60 wt% of metal also
were prepared under the same conditions and designated Au-
CP-1.60 wt% and Pt-CP-1.60 wt%. To study the effect of the
synthesis parameters on the properties of the resultant catalysts,
we used various precipitants (urea, Na2CO3, and NaOH) and
Brunauer–Emmett–Teller (BET) surface areas of the cata-
◦
lysts were measured by nitrogen adsorption at −196 C using
a Micromeritics ASAP 2010 apparatus. The actual Ir loadings
of the catalysts were determined by inductively coupled plasma
(
ICP) analysis, and the chlorine residues in the catalysts were
determined by X-ray fluorescence (XRF). Powder X-ray dif-
fraction (XRD) patterns were obtained with a Rigaku (D/MAX-
βB) diffractometer equipped with an online computer.
◦
precipitation temperatures (ranging from 20 to 100 C). The re-
sultant samples are designated Ir-CP-X, where X represents the
precipitant (in which case, with the precipitation temperature
Temperature-programmed reduction (TPR) experiments
were carried out on a Micromeritics AutoChem II 2920 Au-
tomated Catalyst Characterization System. First, 100 mg of a
catalyst was loaded into a U-shape quartz reactor and pretreated
◦
fixed at 80 C) or the precipitation temperature (in which case,
NaOH was always used as the precipitant).
The Ir-on-ceria catalyst was prepared by the conventional
DP procedure, where the ceria support was suspended in an
aqueous solution of H2IrCl6 and the pH was adjusted to 8–
◦
with air at 400 C for 1 h to remove adsorbed carbonates and
hydrates. Then, after cooling to room temperature, the flow-
9
using NaOH as the neutralizer. After the same drying and
ing gas was switched to a 10 vol% H /Ar, and the catalyst was
2
◦
◦
−1
calcining treatments as for the Ir-CP catalyst, the catalyst, des-
ignated Ir-DP, was obtained. The ceria support was prepared by
precipitation from Ce(NO3)3 under preparation conditions sim-
ilar to those used for the Ir-CP catalyst.
heated to 900 C at a ramping rate of 10 C min
.
In situ diffuse reflectance infrared spectroscopy (DRIFTS)
spectra were acquired with a BRUKER Equinox 55 spectrome-
ter, equipped with a MCT detector and operated at a resolution
−
1
of 4 cm . Before each experiment, a 40-mg sample in a pow-
◦
2
.2. Activity test
der form was reduced in situ with H at 400 C for 2 h. After
cooling to 100 C, the gas for adsorption (CO alone or a PROX
2
◦
Catalytic performance was evaluated in a fixed-bed flow re-
reaction mixture) was introduced into the reaction cell at a to-
−
1
−1 −1
g ).
Cat
actor at atmospheric pressure using 100 mg of a catalyst diluted
with SiC. Before the test, the catalyst was reduced in situ with
tal flow rate of 100 mL min (GHSV = 120,000 mL h
The spectra were recorded as a function of the time on stream
for 60 min in all cases, against a background of the sample at
the reaction temperature under flowing He. All spectra were ob-
tained under steady-state conditions.
◦
hydrogen at 400 C for 2 h. After cooling to room temperature,
a reacting gas containing 2 vol% CO, 1 vol% O2, and 40 vol%
H2 in He was allowed to pass through the reactor at a flow rate

146
Y. Huang et al. / Journal of Catalysis 255 (2008) 144–152
Fig. 2. CO conversion as a function of reaction temperature over the Ir-CP-
1
.60 wt% in the presence of 5% H O (2), both 5% H O and 25% CO ("),
Fig. 1. CO conversion and selectivity to CO as a function of reaction temper-
2
ature in the PROX reaction over 100 mg of the Ir-CP-1.60 wt% (P), Ir-DP-
2 2 2
and none of them (P).
1
.60 wt% (Q) and Ir-DP-0.50 wt% (2) catalysts. Reacting gas composition:
−
1
−1
CO/O /He/H = 2/1/57/40 (vol%); GHSV = 40,000 mL h
g
.
2
2
Cat
◦
catalyst had good stability; CO conversion at 100 C remained
constant at around 68% over 10 h. On the other hand, addition
of 25% CO2 led to a significant decrease in CO conversion,
which may be associated with carbonate deposition. This de-
activation process became less prominent at high temperatures
due to the decomposition of carbonate species.
3
. Results
3
.1. Activity test
Fig. 1 illustrates CO conversion and selectivity to CO2 on
For the CP process, the precipitant and precipitation tem-
perature were important parameters affecting the catalytic per-
formance of the resultant catalysts (Table 1). We investigated
the effect of precipitant using different bases (Na CO , urea,
both Ir-DP-1.60 wt% and Ir-CP-1.60 wt% samples. Note that
at low temperatures (<80 C), the Ir-DP-1.60 wt% catalyst was
more active than the Ir-CP-1.60 wt%; however, above 80 C,
◦
◦
2
3
the Ir-CP-1.60 wt% showed higher CO oxidation activity than
the Ir-DP-1.60 wt%. More importantly, the Ir-CP-1.60 wt%
and NaOH) as the precipitants, while keeping the precipitation
temperature at 80 C. The CO conversions on the Ir-CP-NaOH
◦
exhibited high selectivity to CO in the PROX reaction. The
were higher than those on the Ir-CP-urea, which in turn were
higher than those on the Ir-CP-Na CO . Thus, it seems that the
activities of the resultant Ir-CP catalysts for CO oxidation are
related to the base strength of the precipitants used. Clearly,
NaOH was the preferred base precipitant in our case.
2
selectivity to CO2 remained nearly constant, at around 70%,
2
3
◦
with increasing reaction temperature from 80 to 180 C. In con-
trast, over the Ir-DP-1.60 wt%, the selectivity decreased from
◦ ◦
5
7% at 80 C to only 33% at 180 C. Clearly, as desired, the
Ir-CP sample exhibited a wider temperature window to give a
high selectivity to CO2 and appeared to be more promising for
the PROX reaction. Because the Ir-CP was designed to have
an Ir-in-ceria structure, whereas the Ir-DP has an Ir-on-ceria
structure, it is reasonable to speculate that the difference in cat-
alytic performance between the Ir-CP and Ir-DP samples was
due to the difference in the number of the Ir sites exposed on
the ceria support surface. To confirm this, we reduced the metal
loading on the Ir-DP sample to 0.50 wt%, and found that this
Ir-DP-0.50 wt% sample exhibited higher selectivity than the
Ir-DP-1.60 wt% sample, although it had much lower activity
Choosing NaOH as the most suitable precipitant, we investi-
gated the effect of precipitation temperature (Table 1). CO con-
◦
version at 80 C was used as an index for the activity, whereas
◦
that at 180 C was regarded as an indicator for the selectiv-
◦
ity to CO . When CP was carried out at 20 C, the resulting
Ir-CP-20 C exhibited a very poor activity for CO oxidation. In-
creasing the precipitation temperature from 20 to 80 C resulted
2
◦
◦
in enhanced activity. For example, when precipitation was car-
◦
◦
ried out at 80 C, the CO conversion at 80 C was similar to that
◦
at 180 C, indicating that the catalyst was both highly active and
selective. However, on a further increase in precipitation tem-
◦
◦
◦
below 100 C.
perature from 80 to 100 C, CO conversion at 80 C remained
◦
The effects of H2O and/or CO2 on CO oxidation behav-
ior over the Ir-CP-1.60 wt% catalyst are shown in Fig. 2. The
presence of 5% H2O in the feed stream seemed to be slightly
favorable to the catalytic activity, especially at relatively high
temperatures. Similar results also were obtained on other ceria-
based catalysts [15,21]. A possible explanation for this finding
is that the presence of water accelerated the WGS reaction, and
high, but conversion at 180 C decreased. This indicates that the
selectivity to CO decreased when the catalyst was prepared at
2
◦
a precipitation temperature above 80 C. From the foregoing re-
sults, it can be concluded that the best precipitation temperature
◦
was 80 C.
Besides the precipitation parameters, the Ir loading also ex-
erted a significant effect on the catalytic performance for CO
oxidation (also Table 1). On increasing metal content from 0.20
thus more CO was converted into CO . A time-on-stream test
2
◦
under the presence of 5% H2O showed that the Ir-CP-1.60 wt%
to 1.60 wt%, the CO conversion at 80 C increased from 3

Y. Huang et al. / Journal of Catalysis 255 (2008) 144–152
147
Table 1
◦
Activity data for the PROX reaction on the Ir-CP-X catalysts at a reaction temperature of 80, 140, and 180 C
◦
◦
◦
−3 −1 c
Catalysts
Ir-CP-X
80 C
(%)
140 C
(%)
180 C
(%)
TOF (×10
s
)
a
b
X
X
(%)
X
X
(%)
X
X
(%)
O
2
CO
O
CO
O
CO
2
2
X: precipitant
Ir-CP-Na CO
Ir-CP-urea
Ir-CP-NaOH
1
29
68
4
40
92
3
44
77
25
100
100
14
35
68
100
100
100
0.8
4.1
10.1
2
3
X: precipitation temperature
◦
Ir-CP-20 C
7
26
54
68
68
8
33
71
92
100
13
67
71
77
54
18
100
100
100
100
43
68
68
68
54
97
100
100
100
100
1.6
2.9
7.2
10.1
13.0
◦
Ir-CP-40 C
◦
Ir-CP-60 C
◦
Ir-CP-80 C
◦
Ir-CP-100 C
X: Ir loading
Ir-CP-0.20 wt%
Ir-CP-0.50 wt%
Ir-CP-1.00 wt%
Ir-CP-1.60 wt%
Ir-CP-5.00 wt%
3
22
34
68
52
4
28
39
92
100
28
67
71
77
38
36
100
100
100
100
65
67
68
68
33
100
100
100
100
100
0.6
2.7
5.7
10.1
21.4
a
X
X
= CO conversion.
CO
b
c
= O conversion.
O
2
2
The TOFs of CO oxidation were calculated on the basis of the amount of reducible oxygen on the surface as measured by H -TPR; test condition: 2% CO–1%
2
◦
O –40% H –He; temperature: 100 C.
2
2
◦
to 68%, but the CO conversion at 180 C remained constant
(
around 68%). However, the Ir-CP-5.00 wt% sample exhibited
◦
low CO conversions at both 80 and 180 C, possibly due to
competitive H2 oxidation. Thus, there appears to be an opti-
mum metal loading for the Ir-CP samples, with the best results
found on the catalyst containing 1.60 wt% iridium. Conse-
quently, we chose the Ir-DP and Ir-CP samples with a nominal
metal loading of 1.60 wt% for further characterization stud-
ies.
3
.2. Routine determinations
The BET surface areas of the Ir-CP-1.60 wt% and Ir-DP-
2
−1
1
.60 wt% were 120 and 106 m g , respectively. The Ir load-
ings on the two catalyst samples, as determined by ICP, were
similar (1.50 wt% for the Ir-CP-1.60 wt% and 1.47 wt% for the
Ir-DP-1.60 wt%); however, the surface Ir/Ce atomic ratio on
the Ir-CP-1.60 wt% sample, determined by XPS analysis, was
significantly lower than that on the Ir-DP-1.60 wt% (0.0035 vs
Fig. 3. H -TPR profiles of ceria, Ir-CP-1.60 wt% and Ir-DP-1.60 wt% samples.
2
catalytic performance without the need to consider these para-
meters [6,20].
0
.0091). This implies that most of the Ir particles on the Ir-
3
.3. H2 TPR
CP-1.60 wt% are embedded in the ceria matrix, in accordance
with the Ir-in-ceria structure. No Cl residues were detected by
XRF analysis in either sample. XRD patterns of the Ir-CP-
The H2-TPR technique is often used to characterize ceria-
1
.60 wt% and Ir-DP-1.60 wt% were similar (results not shown).
supported catalysts, because it can provide useful information
on the generation of oxygen vacancies as well as on metal–
support interactions. The H2-TPR profiles of CeO2, Ir-CP-
1.60 wt% and Ir-DP-1.60 wt% are shown in Fig. 3. The TPR
profile of CeO2 presented two reduction peaks. The broad
Only those diffraction peaks corresponding to crystalline CeO2
were observed, and neither an IrO2 nor an Ir phase was de-
tected, irrespective of preparation method. This is likely due to
the very small sizes of the Ir particles in both samples; CeO2
particle sizes were estimated to be ∼6 nm according to the
Debye–Scherrer equation, as confirmed by HRTEM. Thus, the
Ir-CP-1.60 wt% and Ir-DP-1.60 wt% had similar metal load-
ings, similar support particle sizes, and complete removal of
chlorine, which allowed us to analyze the difference in their
◦
peak at around 480 C was assigned to the reduction of sur-
◦
face oxygen, whereas the peak at around 800 C was due to
bulk ceria reduction [22,23]. In agreement with previous studies
[20,24,25], the loading of iridium had no affect on the reduction
of bulk ceria, but effectively promoted the reduction of surface

148
Y. Huang et al. / Journal of Catalysis 255 (2008) 144–152
Table 2
Reducibility of Ir/CeO catalysts and ceria support as measured by H -TPR
2
2
−1
)
Cat
Sample
H consumed (µmol g
2
Peak 1 (T )
Peak 2 (T )
◦
CeO
Ir-DP-1.60 wt%
Ir-CP-1.60 wt%
1160 (470 C)
2
◦
◦
937 (89 C)
223 (295 C)
◦
◦
1071 (140 C)
178 (350 C)
oxygen on ceria, associated with a new signal appearing at low
temperatures. The peak temperatures and corresponding hydro-
gen consumptions are listed in Table 2.
For the Ir-DP-1.60 wt% sample, the major reduction peak
◦
at around 90 C should be an overlap profile of the reduction
of the Ir4 species and the most easily reducible CeO2, be-
+
−
1
Fig. 4. Typical DRIFTS spectra of Ir-CP-1.60 wt% and Ir-DP-1.60 wt% samples
cause the amount of H2 consumed (∼937 µmol g ) greatly
Cat
◦
obtained under steady-state CO adsorption conditions at 100 C.
exceeded that required for the reduction of iridium oxide. (The
4+
calculated amount of hydrogen consumed for reducing Ir to
metallic Ir is only 150 µmol g .) In addition, the reduction
peak at about 300 C would be due to the reduction of surface
−
1
duction peaks (Table 3). Varying the precipitant from Na CO ,
2
3
Cat
◦
urea to NaOH, the amount of hydrogen consumed for the Ir-CP
−
.
Cat
1
samples increased from 750, to 892, and then to 1071 µmol g
CeO2, which has no strong interaction with iridium [19,24].
The Ir-CP-1.60 wt% sample showed an initial reduction peak
In addition, both the Ir loading and the precipitation temper-
ature had a significant affect on the reducibility of the ceria
support. A higher precipitation temperature or a higher Ir load-
◦
◦
at about 140 C, 50 C higher than that of the Ir-DP-1.60 wt%
sample. This lower reduction temperature should be related to
the Ir-DP-1.60 wt% sample’s slightly higher activity at low tem-
peratures (Fig. 1). Moreover, the hydrogen consumption cor-
responding to the first peak was greater for Ir-CP-1.60 wt%
than for Ir-DP-1.60 wt%, leading to more oxygen vacancies on
the former catalyst. As described previously [19], the NM in a
ing resulted in higher H consumption and a lower reduction
2
temperature. Comparing the H2 consumption (Table 3) with
the activity level (Table 1) demonstrates a direct correlation
between the reducibility of the surface ceria and its catalytic
activity for CO oxidation. Higher hydrogen consumption often
was observed on the highly active samples.
NM/CeO -CP catalyst exists mainly in a metallic state, regard-
2
less of the thermal treatment. Moreover, the NM species are
mostly blocked within the internal surface of the solid. Thus, we
can speculate that the first reduction peak of the Ir-CP-1.60 wt%
should be mainly associated with the reduction of surface oxy-
3
.4. In situ DRIFTS
To further verify whether the Ir particles in the Ir-CP sample
gen of CeO . Our observations seem to be consistent with the
were embedded in the ceria matrix, as desired, we investigated
the CO adsorption behavior over the Ir-CP-1.60 wt% at 100 C
using in situ DRIFTS. For comparison, we also performed the
same experiments on the Ir-DP-1.60 wt% catalyst. As shown
in Fig. 4, CO adsorption on the Ir-DP-1.60 wt% sample pro-
duced a strong band at 2050 cm and a shoulder at 2063 cm ,
which can be attributed to CO species linearly adsorbed on the
different Ir sites [30–32]. In contrast, CO adsorption on the Ir-
2
◦
idea of a “junction effect” [26] arising from the contact between
ceria and a metal with a high work function.
From the foregoing results, we can postulate that the acti-
vated ceria can be easily reduced at low temperatures, and thus
be responsible for the creation of oxygen vacancies. Oxygen
vacancies are known to be chemically active and play an im-
portant role in low-temperature CO oxidation [27–29]. Accord-
ingly, we characterized the Ir-CP catalysts with different pre-
cipitants and precipitation temperatures, as well as with various
Ir loadings, giving special attention to their low-temperature re-
−
1
−1
0
CP-1.60 wt% sample yielded only a very weak band of at least
10-fold lower intensity than that on the Ir-DP-1.60 wt% sam-
ple. This result strongly suggests that most of the Ir particles
Table 3
Reducibility of the Ir-CP-X catalysts as measured by H -TPR
2
Precipitants
Sample
Metal loadings
Precipitation temperatures
−
1
−1
−1
H consumed (µmol g
2
)
Sample
H consumed (µmol g
2
)
Sample
H consumed (µmol g
2
)
Cat
Cat
Cat
◦
a
◦
◦
◦
Ir-CP-Na CO
2
750 (147 C)
Ir-CP-0.20 wt%
Ir-CP-0.50 wt%
830 (197 C)
Ir-CP-20 C
580 (183 C)
3
◦
◦
◦
915 (182 C)
Ir-CP-40 C
758 (145 C)
◦
◦
◦
◦
Ir-CP-urea
892 (140 C)
Ir-CP-1.00 wt%
Ir-CP-1.60 wt%
973 (150 C)
Ir-CP-60 C
803 (141 C)
◦
◦
◦
1071 (140 C)
Ir-CP-80 C
1071 (140 C)
◦
◦
◦
◦
Ir-CP-NaOH
a
1071 (140 C)
Ir-CP-5.00 wt%
1343 (117 C)
Ir-CP-100 C
1116 (136 C)
The number in brackets indicates the first reduction peak temperature.

Y. Huang et al. / Journal of Catalysis 255 (2008) 144–152
149
Fig. 6. Typical DRIFTS spectra of Ir-CP-1.60 wt% sample obtained under dif-
◦
ferent steady-state conditions at 100 C.
Fig. 5. Typical DRIFTS spectra of Ir-CP-1.60 wt% and Ir-DP-1.60 wt% samples
obtained under steady-state PROX conditions at 100 C.
◦
in the Ir-CP-1.60 wt% sample were embedded in the ceria sup-
port, whereas those on the Ir-DP-1.60 wt% sample were mainly
exposed on the ceria surface.
Fig. 5 compares the DRIFTS spectra on the Ir-CP-1.60 wt%
and Ir-DP-1.60 wt% samples during the PROX reaction at
◦
−1
1
00 C. The bands in the 1700–1200 cm region can be as-
cribed to different carbonates [7,8,18,33,34]. The broad bands
−
1
centered at 3421 cm are assigned to hydrogen-bonded ad-
sorbed water on the ceria surface [7,8], whereas the bands at
−1
around 2360 cm are associated with gas-phase CO . Com-
2
paring the spectra on the two catalysts raises some interesting
points. First, the main difference between them concerns the
strength of the CO adsorption band. The low intensity of the
CO adsorption band on the Ir-CP-1.60 wt% should relate to the
Fig. 7. Typical DRIFTS spectra of Ir-CP-NaOH and Ir-CP-Na2CO3 samples
obtained under steady-state PROX conditions at 100 C.
◦
Ir-in-ceria structure. Second, the band for gas-phase CO was
2
stronger on Ir-CP-1.60 wt% than on Ir-DP-1.60 wt%, indicating
greater CO oxidation activity on the former. Third, the bands
bonyl species were very weak on both samples, implying that
the CP procedure could yield an Ir-in-ceria structure regardless
of whether NaOH or Na2CO3 was used as the precipitant. In
contrast to the appearance of a stronger band for gas-phase CO2
on the Ir-CP-NaOH, neither CO2 nor H2O was detected on the
Ir-CP-Na2CO3 sample. These observations are consistent with
the results obtained in the catalytic reactor tests. Moreover, the
stretching vibration bands at 1700–1200 cm were relatively
weak on the Ir-CP-Na2CO3. Considering that such species are
generally formed on activated ceria supports [34], we thus can
propose that the ceria on the Ir-CP-Na2CO3 was not sufficiently
activated to form extensive oxygen vacancies.
−
1
attributed to hydrogen-bonded water (centered at 3421 cm
)
were stronger on Ir-DP-1.60 wt% than on Ir-CP-1.60 wt%, sug-
gesting greater H oxidation activity on the former, in agree-
2
ment with the H2-TPR findings. In addition, we found that CO
formed several carbonates by reacting with the surface oxygen
on both samples, and the intensity of these species was greater
on Ir-DP-1.60 wt% than on Ir-CP-1.60 wt%.
−
1
The DRIFTS spectra on the Ir-CP-1.60 wt% recorded un-
der different reaction environments (PROX mixture, CO alone,
and CO oxidation) are shown in Fig. 6. Initially, the spectra
were collected under the PROX condition after a steady-state
condition was established; later, the H and O were removed
4. Discussion
2
2
from the stream. In this case, the gas-phase CO2 bands dis-
appeared quickly, along with the hydrogen-bonded adsorbed
water. Subsequently, oxygen was introduced to the stream. The
An effective PROX catalyst should have a wide operation
window to avoid the need for precise temperature control. How-
ever, in most cases, the selectivity to CO2 tends to decrease
with increasing reaction temperature due to the competitive ox-
idation of H2. To address this problem, we prepared a highly
selective ceria-supported iridium catalyst with a special Ir-in-
ceria structure (as revealed by the DRIFTS spectra in Figs. 4
gas-phase CO band reappeared, but was smaller than that un-
2
der the PROX condition, implying a lower activity level in the
absence of H2.
The DRIFTS spectra collected on Ir-CP-Na2CO3 (a nonac-
tive catalyst) and Ir-CP-NaOH (a highly active catalyst) under
◦
the PROX condition at 100 C are compared in Fig. 7. The car-
and 5), on which H oxidation on the metallic Ir sites can be
2

150
Y. Huang et al. / Journal of Catalysis 255 (2008) 144–152
significantly suppressed. In contrast, the Ir-DP-1.60 wt% sam-
ple unambiguously exhibited an Ir-on-ceria structure. The dif-
ferent structures of the Ir-CP-1.60 wt% and Ir-DP-1.60 wt%
determined their different behaviors in the PROX reaction, es-
pecially in terms of selectivity to CO2. The Ir-CP-1.60 wt%
sample exhibited high CO oxidation activity in the PROX reac-
tion and almost no loss in selectivity in the temperature range
◦
8
0–180 C, whereas the Ir-DP-1.60 wt% catalyst, due to its high
H2 oxidation activity, demonstrated a continuous drop in selec-
tivity.
To gain insight into this structure–reactivity relationship, we
changed precipitants from NaOH to Na2CO3 and compared the
catalytic performance of the resultant catalysts. Fig. 7 and Ta-
ble 1 clearly show that the Ir-CP-Na2CO3 sample exhibited
very poor activity and selectivity even though it also had an
Ir-in-ceria structure. This unexpected result reminds us that the
metal-in-ceria structure is not the only factor determining cat-
alytic performance; some other intrinsic factors must be affect-
ing the activity and selectivity of the Ir-CP catalyst. Considering
the several preparation parameters that influence catalytic per-
formance (Table 1), we believed that analyzing the interaction
between Ir and ceria during the preparation process would help
us uncover those unknown factors in the Ir-in-ceria structure
that are actually responsible for the high selectivity to CO2.
◦
Fig. 8. CO conversions at 80 C plot for the hydrogen consumed for the sur-
face CeO reduction over the Ir-in-ceria catalysts. The scatters are experimental
data. The lines are obtained by linear fitting.
2
reducible oxygen on the surface as measured by H2-TPR [15].
Although this calculation may underestimate the TOF values by
considering that not all of the surface oxygen vacancies partici-
pate in the reaction, it can help us better understand the reaction
on the different Ir-CP catalysts. As shown in Table 1, the TOFs
increased with increasing base strength, precipitation tempera-
ture, and Ir loading. When the precipitation proceeded under a
It is well known that Ce3 can be oxidized to form Ce
+
4+
hydroxyl species in alkaline environments and in the presence
of oxygen [19,35,36]. In a basic solution, the redox potential
less strongly basic environment (such as with urea or Na CO )
or at a lower temperature (below 80 C), the redox reaction be-
2
3
◦
can be quantified by the equation E
= [0.787 −
Ce(4+)/Ce(3+)
tween Ce3 and Ir occurred to a lesser extent or did not occur
at all. In such cases, the interaction between Ir and ceria was not
strong enough to activate the surface ceria, resulting in lower
activity. In contrast, when the precipitation was conducted at a
higher temperature or more Ir species were incorporated into
+
4+
∗
−20
0
.0592 pH] V [37], corresponding to Ksp values of 1.6×10
−47
and 2.5 × 10
for Ce(OH)3 and Ce(OH)4, respectively. A de-
crease in emf will occur with increasing the pH, thus facili-
tating the oxidation of Ce3 to Ce ; however, the reaction is
+
4+
thermodynamically unfavorable due to the high emf. On the
other hand, if CeO2 were formed instead of Ce(OH)4, then
oxidation would be strongly favored due to its very low Ksp
◦
the ceria support, as in the case of Ir-CP-100 C and Ir-CP-
.00 wt%, a stronger interaction between Ir and ceria was seen.
In these cases, the ceria was overactivated, and the resulting Ir-
CP catalysts were more active for the undesired H oxidation
5
−
63
(
1.25 × 10 ). In that case, the redox potential can be quanti-
∗
fied by the equation E
37]. In this CP process, the oxidation of Ce to Ce is cou-
= [−0.112−0.0592 pH] V
2
Ce(4+)/Ce(3+)
3+
4+
reaction as well, leading to decreased CO conversion and se-
lectivity to CO2 [39]. This finding demonstrates that the ceria
support must be properly activated to obtain the desired selec-
tivity to CO2.
For comparison, we also calculated the TOF on Ir-DP-
1.60 wt% on the basis of the reducible oxygen atoms, and found
it to be slightly higher than that on the Ir-CP-1.60 wt% catalyst
(Table 4). But if the extensively exposed Ir sites on the Ir-DP-
1.60 wt% were considered the active sites for CO oxidation and
the dispersion was assumed to be 100%, then its TOF value
[
4+
0
pled with the reduction of Ir to Ir . Our XRD examination of
the as-synthesized (uncalcined) Ir-CP catalysts revealed the for-
mation of a phase-pure CeO2. Moreover, the black color of the
as-synthesized Ir-CP (compared with the green-colored Ir-DP)
also is indicative of the metallic state of iridium. Therefore, the
key to this coupled redox process is the formation of CeO2 in
solution, which is favored under strongly basic conditions and
at high temperatures [38].
This redox-precipitation procedure, besides the formation of
an Ir-in-ceria structure, induces strong, extensive interactions
between Ir and ceria. The metal–support interactions can be
roughly quantified by the reducibility of the surface ceria pro-
−
1
(0.184 s ) was one order of magnitude greater than that on the
−
1
Ir-CP-1.60 wt% catalyst (0.010 s ). Such a high activity level
of the Ir-DP-1.60 wt% would cause undesirable H2 oxidation
and thereby decrease the selectivity to CO2.
Because Au/CeO2 and Pt/CeO2 have been extensively stud-
ied for the PROX reaction [5–10,15], we prepared Au-CP-
1.60 wt% and Pt-CP-1.60 wt% catalysts under the same pre-
cipitation conditions and evaluated their catalytic performance
for the PROX reaction in comparison with Ir-CP-1.60 wt%. The
◦
moted by iridium. Fig. 8 plots the CO conversions at 80 C
Table 1) versus the amount of H2 consumed for surface ceria
(
reduction (Table 3). The clear strongly positive correlation be-
tween them implies that oxygen vacancies on the ceria surface,
formed with the aid of iridium, may act as the active sites in
this reaction [12–14]. Consequently, we calculated turnover fre-
quencies (TOFs) of CO oxidation on the basis of the amount of
◦
selectivities of these catalysts at 100 and 180 C, as well as their

Y. Huang et al. / Journal of Catalysis 255 (2008) 144–152
151
Table 4
Comparison of reducibility, TOFs of CO oxidation, and selectivity to CO
2
Catalyst
H consumed
Selectivity to CO
TOF
2
2
−
1
◦
◦
−3 −1 b
s )
(µmol g
)
100 C
180 C
(×10
Cat
◦
a
Ir-DP-1.60 wt%
Ir-CP-1.60 wt%
Pt-CP-1.60 wt%
Au-CP-1.60 wt%
937 (89 C)
54
71
66
75
33
68
30
73
14.0
10.1
11.6
7.2
◦
1071 (140 C)
◦
1442 (34 C)
◦
536 (154 C)
a
The number in brackets indicates the first reduction peak temperature.
The TOFs of CO oxidation were calculated on the basis of the amount of
b
reducible oxygen on the surface as measured by H -TPR; test condition: 2%
2
◦
CO–1% O –40% H –He; temperature: 100 C.
2
2
◦
TOFs at 100 C, are given in Table 4. Of the three NM-CP cata-
lysts (NM = Ir, Pt, Au), the Pt-CP-1.60 wt% exhibited the high-
est TOF value, demonstrating its highly active nature; however,
such high activity also caused competitive H2 oxidation, result-
Fig. 9. CO conversion as a function of reaction temperature over the Ir-CP-
1.60 wt% under different reaction conditions.
◦
ing in a dramatic decrease in selectivity to CO2 at 180 C. In
contrast to the Pt-CP-1.60 wt% catalyst, the Au-CP-1.60 wt%
catalyst yielded the best results in terms of selectivity to CO2,
while exhibiting the lowest activity for CO oxidation. These
catalysts’ differing performance likely is related to their differ-
ent natures, with different work functions (Pt > Ir > Au) [26].
Clearly, the Ir-CP-1.60 wt% catalyst is the best choice in terms
of both activity and selectivity.
tional effect on CO oxidation. In fact, the presence of H2 and/or
H2O has been reported to enhance the rates of CO oxidation
on ceria-based catalysts [9,15,21]. In this case, we tentatively
attribute this promotion to the formation of bicarbonates (by
reaction of carbonates with OH and/or H2O), which decom-
pose at a higher rate than carbonates [42,43]. Under the PROX
−1
conditions, the significant broadening above 1611 cm may
Based on the foregoing discussion regarding the active sites
responsible for the catalytic activity of the Ir-CP catalysts, we
believe that CO oxidation could proceed via the Mars–van
Krevelen mechanism [40,41]. The iridium particles embedded
in the ceria matrix and interacting strongly with the ceria sup-
port could weaken the surface Ce–O bonds and facilitate the
formation of more reactive oxygen [26]. In CO oxidation, the
CO takes an oxygen atom from the ceria surface and creates
an oxygen vacancy. The gas-phase oxygen will adsorb on the
vacancies and be activated by the electron-rich environment
created by the vacancies. The activated O2 reacts with CO to
form a carbonate, which decomposes to release CO2 and heal
the oxygen vacancy; thus, a catalytic cycle from CO to CO2 is
completed without direct participation of the Ir for providing
CO adsorption sites.
imply the formation of bicarbonate [7], as shown clearly in
Fig. 6.
Finally, it should be kept in mind that our experimental
results do not provide absolute and incontrovertible evidence
of total encapsulation of the Ir species. The exposed Ir, even
in a very low concentration, would contribute to the reaction
through the Langmuir–Hinshelwood mechanism.
5. Conclusion
δ+
3+
The redox-precipitation reaction between Ir and Ce oc-
curring during the CP procedure resulted in Ir particles becom-
ing embedded in the ceria matrix. The resulting Ir-in-ceria cata-
lyst exhibited a wide temperature window for producing a high
Along with the reaction mechanism, another important point
is the promoting role of H2. In accordance with the DRIFTS
study illustrated in Fig. 6, we found that the CO conversions
without the presence of H2 were significantly lower than those
under the PROX conditions (Fig. 9). One of the promoting
mechanisms of the oxidation of H2 to CO may be due to the
WGS reaction originating from the oxidation of H2 to H2O on
the basis of previous reports [6,7]. Consequently, we investi-
gated the WGS reaction over the Ir-CP-1.60 wt% catalyst using
a gas composition of 2% CO and 2% H2O balanced with He. As
shown in Fig. 9, the WGS reaction began at temperatures above
selectivity to CO and appeared to be a highly selective catalyst
2
for the PROX reaction. The desired selectivity was probably
due to the absence of extensively exposed Ir species on the ceria
support, allowing suppression of hydrogen oxidation at higher
temperatures. Our findings provide promising possibilities for
preparing improved ceria-based catalysts for the PROX reac-
tion.
Acknowledgments
◦
◦
1
40 C, and the CO conversion was only 10% at 220 C. Thus,
we can conclude that the contribution from the WGS reaction
to CO oxidation was minor in the PROX reaction, which agrees
well with the effect of 5% H2O on the PROX activity (Fig. 2).
We also fed 2% H2O to the reacting gas for CO oxidation and
found that even a small amount of H2O had a significant promo-
Financial support was provided by the National Science
Foundation of China (NSFC) for Distinguished Young Schol-
ars (20325620) and NSFC grants (20673116, 20773124). The
authors thank R. Prins and S.T. Oyama for help with the English
and fruitful discussions.

152
Y. Huang et al. / Journal of Catalysis 255 (2008) 144–152
References
[21] J.L. Ayastuy, A. Gil-Rodríguez, M.P. González-Macrcos, M.A. Gutiérrez-
Ortiz, Int. J. Hydrogen Energy 31 (2006) 2231.
[22] A. Trovarelli, Catal. Rev. Sci. Eng. 38 (1996) 439.
[23] A. Trovarelli (Ed.), Catalysis by Ceria and Related Materials, Imperial
College Press, London, 2002, Chapter 4, p. 85.
[24] L. Tournayan, N.R. Marcilio, R. Frety, Appl. Catal. 78 (1991) 31.
[25] B. Zhang, X. Tang, Y. Li, W. Cai, Y. Xu, W. Shen, Catal. Commun. 7
(2006) 367.
[
1] L. Shore, R.J. Farrauto, in: W. Vielstich, A. Lamm, H.A. Gasteiger (Eds.),
Handbook of Fuel Cells: Fundamentals Technology and Applications,
Part 2, vol. 3, John Wiley & Sons, West Sussex, 2003, p. 211.
2] D.L. Trimm, Z.L. Onsan, Catal. Rev. Sci. Eng. 43 (2001) 31.
3] D.L. Trimm, Appl. Catal. A 296 (2005) 1.
[
[
[
[
[
[
4] S.H. Oh, R.M. Sinkevitch, J. Catal. 142 (1993) 254.
5] F. Mariño, C. Descorme, D. Duprez, Appl. Catal. B 54 (2004) 59.
6] A. Wootsch, C. Descorme, D. Duprez, J. Catal. 225 (2004) 259.
7] O. Pozdnyakova, D. Teschner, A. Wootsch, J. Kröhnert, B. Steinhauer,
H. Sauer, L. Toth, F.C. Jentoft, A. Knop-Gericke, Z. Paál, R. Schlögl,
J. Catal. 237 (2006) 1.
[26] J.C. Frost, Nature 334 (1988) 577.
[27] Q. Fu, A. Weber, M. Flytzani-Stephanopoulos, Catal. Lett. 72 (2001) 87.
[
28] A.M. Venezia, G. Pantaleo, A. Longo, G.D. Carlo, M.P. Casaletto, F.L.
Liotta, G. Deganello, J. Phys. Chem. B 109 (2005) 2821.
[
29] J. Guzman, S. Carrettin, J.C. Fierro-Gonzalez, Y. Hao, B.C. Gates, P. Con-
cepción, A. Corma, Angew. Chem. Int. Ed. 44 (2005) 4778.
30] F. Solymosi, É. Novák, A. Molnár, J. Phys. Chem. 94 (1990) 7250.
31] A. Erd o˝ helyi, K. Fodor, G. Suru, Appl. Catal. A 139 (1996) 131.
32] A. Goguet, F.C. Meunier, D. Tibiletti, J.P. Breen, R. Burch, J. Phys. Chem.
B 108 (2004) 20240.
[
8] O. Pozdnyakova, D. Teschner, A. Wootsch, J. Kröhnert, B. Steinhauer,
H. Sauer, L. Toth, F.C. Jentoft, A. Knop-Gericke, Z. Paál, R. Schlögl,
J. Catal. 237 (2006) 17.
[
[
[
[9] O. Pozdnyakova-Tellinger, D. Teschner, J. Kröhnert, F.C. Jentoft, A.
Knop-Gericke, R. Schlögl, A. Wootsch, J. Phys. Chem. C 111 (2007)
5426.
[33] C. Li, Y. Sakata, T. Arai, K. Domen, K. Maruya, T. Onishi, J. Chem. Soc.
[
10] D. Teschner, A. Wootsch, O. Pozdnyakova-Tellinger, J. Kröhnert, E.M.
Vass, M. Hävecker, S. Zafeiratos, P. Schnörch, F.C. Jentoft, A. Knop-
Gericke, R. Schlögl, J. Catal. 249 (2007) 316.
Faraday Trans. 85 (1989) 1451.
[
[
34] A. Holmgren, B. Andersson, D. Duprez, Appl. Catal. B 22 (1999) 215.
35] C. Milone, M. Fazio, A. Pistone, S. Galvagno, Appl. Catal. B 68 (2006)
[
[
[
11] C. Hardacre, R.M. Ormerod, R.M. Lambert, J. Phys. Chem. 98 (1994)
2
8.
36] F. Arena, G. Trunfio, J. Negro, B. Fazio, L. Spadaro, Chem. Mater. 19
2007) 2269.
37] A.J. Bard, R. Parsons, J. Jordan, in: IUPAC (Ed.), Standard Potential in
Aqueous Solution, Marcel Dekker, New York, 1985, p. 617.
1
0901.
12] S. Golunski, H. Hatcher, R. Rajaram, T. Truex, Appl. Catal. B 5 (1995)
67.
[
[
[
[
(
3
13] S. Golunski, R. Rajaram, N. Hodge, G. Hutchings, C.J. Kiely, Catal. To-
day 71 (2002) 107.
38] D. Terribile, A. Trovarelli, J. Llorca, C. Leitenburg, G. Dolcetti, J. Catal.
[
[
14] S. Golunski, R. Rajaram, CATTECH 6 (2002) 30.
15] W. Deng, J.D. Jesus, H. Saltsburg, M. Flytzani-Stephanopoulos, Appl.
Catal. A 291 (2005) 126.
178 (1998) 299.
39] D. Gamarra, G. Munuera, A.B. Hungría, M. Fernández-García, J.C.
Conesa, P.A. Midgley, X.Q. Wang, J.C. Hanson, J.A. Rodríguez, A. Mar-
tínez-Arias, J. Phys. Chem. C 111 (2007) 11026.
[
[
16] Q. Fu, H. Saltsburg, M. Flytzani-Stephanopoulos, Science 301 (2003) 935.
17] C.M.Y. Yeung, K.M.K. Yu, Q.J. Fu, D. Thompsett, M.I. Petch, S.C.J.
Tsang, J. Am. Chem. Soc. 127 (2005) 18010.
[40] P. Mars, D. Van Krevelen, Chem. Eng. Sci. (special supplement) 3 (1954)
41.
[
[
[
18] C.M.Y. Yeung, F.C. Meunier, R. Burch, D. Thompsett, S.C.J. Tsang,
[
[
41] V. Shapovalov, H. Metiu, J. Catal. 245 (2007) 205.
42] M.M. Schubert, A. Venugopal, M.J. Kahlich, V. Plzak, R.J. Behm,
J. Catal. 222 (2004) 32.
J. Phys. Chem. B 110 (2006) 8540.
19] R. Rajaram, J.W. Hayes, G.P. Ansell, H.A. Hatcher, U.S. patent 5,993,762,
filed on 30 Nov. 1999.
20] Y.Q. Huang, A.Q. Wang, X.D. Wang, T. Zhang, Int. J. Hydrogen En-
ergy 32 (2007) 3880.
[43] Y. Denkwitz, Z. Zhao, U. Hörmann, U. Kaiser, V. Plzak, R.J. Behm,
J. Catal. 251 (2007) 363.