Influence of the support on the preferential oxidation of CO in hydrogen-rich steam reformates over the CuO-CeO2-ZrO2 system

Article abstract of DOI:10.1016/j.jcat.2003.09.006

Preferential oxidation of CO in hydrogen-rich steam reformates was studied using CuO-CeO₂, CuO-CeO₂-ZrO₂, and CuO-ZrO₂ catalysts. The composition of the support markedly influenced the PROX (preferential oxidati

Full text of DOI:10.1016/j.jcat.2003.09.006

Journal of Catalysis 221 (2004) 455–465
www.elsevier.com/locate/jcat
Influence of the support on the preferential oxidation of CO in
hydrogen-rich steam reformates over the CuO–CeO2–ZrO2 system
P. Ratnasamy,∗ D. Srinivas, C.V.V. Satyanarayana, P. Manikandan, R.S. Senthil Kumaran,
M. Sachin, and Vasudev N. Shetti
National Chemical Laboratory, Pune 411 008, India
Received 6 May 2003; revised 26 August 2003; accepted 2 September 2003
Abstract
Preferential oxidation (PROX) of CO in hydrogen-rich steam reformates was investigated using CuO–CeO , CuO–CeO –ZrO , and
2
2
2
2
CuO–ZrO catalysts. CuO (1–10 wt%) samples supported over high-surface-area (S
= 117–172 m /g), cubic CeO , CeO –ZrO , and
2
BET
2
2
2
ZrO were synthesized by coprecipitation. The composition of the support markedly influenced the PROX activity. Both CuO–CeO and
2
2
CuO–CeO –ZrO exhibited higher activity and selectivity in CO oxidation than CuO–ZrO . The adverse influence of H O was accen-
2
2
2
2
tuated in catalysts containing ZrO . Below 423 K and over CuO–CeO with less than 5 wt% CuO, the presence of H O in the feed
2
2
2
suppressed CO oxidation. H O had a negligible effect on H oxidation. The catalysts showed stable activity in long-term experiments
2
2
with the realistic feeds. The catalysts were characterized by XRD, surface area, TPR, diffuse reflectance UV–visible, EPR, and magnetic-
susceptibility techniques. While a small amount of copper might be incorporated in the CeO /ZrO fluorite lattice (forming a solid solution),
2
2
most of it was at the surface of the support as isolated monomeric (types I and II) and dimeric (type IV) copper oxo species, nano-sized
copper clusters containing magnetically interacting copper ions (type III) and a CuO-like, bulk phase. While the isolated copper oxo
species exhibited reversible reduction–oxidation behavior, the interacting copper and CuO-like phases exhibited irreversible reduction be-
havior. The amount and reducibility of CuO on different supports correlated with their CO oxidation activities and increased in the order
CuO–ZrO ꢀ CuO–CeO –ZrO < CuO–CeO .
2
2
2
2
2003 Elsevier Inc. All rights reserved.
Keywords: Preferential CO oxidation (PROX); PEM fuel cells; Supported CuO catalysts; CuO–CeO ; CuO–CeO –ZrO ; CuO–ZrO ; Influence of support
2
2
2
2
1. Introduction
moval in NH3 or H2 chemical plants, is impractical in the
fuel cell context, especially for “on-board” use in vehicles.
Preferential oxidation of CO to CO2 without simultaneously
oxidizing H2 to H2O (PROX) is one of the solutions. How-
ever, the latter oxidation is more exothermic and, hence, fa-
vored at high temperatures. Selective catalysts active at low
temperatures are needed. Supported Pt [4,5], Rh [6], Pd [7],
and Au [8] as well as base metal catalysts like CoO [9] and
CuO–CeO2 [10,11] have been investigated.
The PROX reaction has been extensively investigated
[12–20]. However, in many of the studies reported so far,
especially in the journal literature, the PROX reactor feed
did not include CO2, CH4, or H2O which occur in significant
concentrations in the effluents from real-life reforming-cum-
WGS processes and whose presence can lead to results not
anticipated from studies with only a mixture of CO, O2, and
H2. CH4, for example, can be oxidized to CO/CO2 under
PROX conditions, especially over noble metal catalysts, and
Selective CO removal from streams containing an excess
of H2 and significant amounts of CO2 and H2O is a current
challenge in heterogeneous catalysis research especially in
the preparation of H2 suitable for polymer electrolyte mem-
brane fuel cells (PEMFC) [1,2]. Hydrogen is usually gener-
ated from hydrocarbons or alcohols by steam/autothermal
reforming followed by the water–gas shift (WGS) reac-
tion [3]. Typical effluents from such a process contain about
0.3–1.0% of CO in a large excess of H2 (40–75%) and
about 20–25% CO2. CO levels, however, have to be re-
duced to below 100 and, preferably, below 10 ppm for use in
PEMFC [1]. Methanation, after removal of CO2 by pressure
swing adsorption, the conventional strategy used for CO re-
*
Corresponding author.
E-mail address: prs@ems.ncl.res.in (P. Ratnasamy).
0021-9517/$ – see front matter
doi:10.1016/j.jcat.2003.09.006
2003 Elsevier Inc. All rights reserved.

456
P. Ratnasamy et al. / Journal of Catalysis 221 (2004) 455–465
thereby lead to higher CO levels in the PROX reactor efflu-
ent. Again, in our laboratory, we had found that some of the
reported catalysts, especially those based on Au, which were
very active at very low temperatures (even below 373 K) ex-
hibited activity decay in long-term studies, probably due to
the strong adsorption of CO2 (an “acid gas”) and H2O on
some of the basic oxides used as supports for Au. On Pt–
Al2O3, Korotkikh and Farrauto [21] reported that CO2 does
not affect CO oxidation; H2O, however, decreased CO ox-
idation by about 10%. Such studies have not been reported
for base metal PROX catalysis.
2.1.1. Preparation of CuO–CeO2 catalysts
In the preparation of CuO (1 wt%)—CeO2 (99 wt%),
12.48 g of Ce(NO3)3 · 6H2O was dissolved in 288 ml of dis-
tilled water and 0.152 g of Cu(NO3)2 ·3H2O was dissolved
in 6.3 ml of distilled water. These solutions were mixed to-
gether and added dropwise to a continuously stirred solution
of KOH (0.1 M) taken in a 2 liter round bottom flask, placed
in a heating mantle, and fitted with a pH electrode, water-
cooled condenser and burette containing additional 0.1 M
KOH solution. During addition, the temperature of the flask
was maintained at 353 K and pH 10. The cations were pre-
cipitated in the form of their hydroxides. The mixture was
digested at 353 K for 3 h and then cooled to 298 K. The pre-
cipitate was separated and washed copiously with distilled
water until all the potassium ions were removed. It was then
dried in an air oven, at 393 K, for 8 h. Later the dried mater-
ial was crushed into powder and calcined in air, at 773 K, for
5 h, to get the final product. It may be noted that potassium
is an effective element for preferential oxidation of CO in a
H2-rich stream [30]. Absence of potassium ions in the final
product was determined by atomic absorption spectroscopy
(AAS, Varian Spectr SF-220).
We now report an investigation of the preferential oxi-
dation of CO over the CuO–CeO2–ZrO2 catalyst system us-
ing a feed composition more representative of the feed to
a PROX reactor in a PEM fuel cell. CuO–CeO2 exhibits
significant activity in the total oxidation of CO [10]. The
well-known enhancement of the total oxidation activity of
CuO when supported on oxides like CeO2 was attributed
to a “synergistic” effect [11,22]. It is also known that addi-
tion of ZrO2 (to form CeO2–ZrO2 solid solution) improves
the oxygen-storage capacity of ceria and, in addition, re-
tards the growth of CeO2 crystallites [23]. Even though a
large amount of structural studies on the CuO-CeO2–ZrO2
system has been carried out [24–29], the influence of the
crystallite and particle size of cerium and zirconium ox-
ides on the dispersion, oxidation states, magnetic properties,
and catalytic activity, in PROX reactions, of copper cata-
lysts supported on the CeO2–ZrO2 support is not understood
completely. For example, what is the effect of varying the
dispersion of ceria and CuO on the selectivity for oxida-
tion of CO (vis-à-vis H2 and CH4) in the presence of CO2
and H2O? We have prepared samples of CuO–CeO2–ZrO2
with varying concentrations of the individual components
and characterized their physicochemical properties by tech-
niques such as XRD, temperature-programmed reduction,
UV–visible and electron spin resonance spectroscopies, and
magnetic susceptibility. The catalytic activity of these well-
characterized samples in the selective oxidation of CO in
a gas mixture with a composition similar to that present
in typical PROX reactor effluents is also reported. The re-
sults enable us to throw further light on the influence of
the structural characteristics of supported copper in these
systems on their catalytic activity and selectivity in PROX
reactions.
2.1.2. Preparation of CuO–ZrO2 catalysts
The CuO–ZrO2 catalysts were prepared in a similar
manner as described above. In a typical preparation of
CuO (1 wt%)–ZrO2 (99 wt%) (5 g batch), 9.28 g of
ZrO(NO3)2 ·xH2O dissolved in 400 ml distilled water and
0.152 g of Cu(NO3)2 · 3H2O dissolved in 6.3 ml distilled
water were used.
2.1.3. Preparation of CuO–CeO2–ZrO2 catalysts
The Ce:Zr ratio in these catalysts is 1:1 wt%. In the prepa-
ration of CuO (1 wt%)–CeO2 (49.5 wt%)–ZrO2 (49.5 wt%)
(5 g batch), 6.24 g of Ce(NO3)3 ·6H2O dissolved in 144 ml
of distilled water, 4.64 g of ZrO(NO3)2 ·xH2O dissolved in
200 ml of distilled water, and 0.152 g of Cu(NO3)2 ·3H2O
dissolved in 6.3 ml of distilled water were used. The cata-
lysts were prepared in a similar manner as described above.
For comparison, pure ceria, zirconia and ceria-zirconia
(1:1 wt%) samples were also prepared.
2.2. Catalyst characterization
A Rigaku Geigerflex X-ray diffractometer with Ni-
filtered Cu-Kα radiation (λ = 1.5406 Å, 40 kV, 30 mA) was
used to determine the crystallinity and phase purity of the
samples. The XRD patterns were recorded in the 2θ range of
20–90◦, with a scan speed of 2◦/min. The average crystallite
size was determined from the linewidths of the XRD peaks
corresponding to (111) reflection, using the Debye–Scherrer
equation. The specific surface area (SBET) was estimated
from the N2 adsorption/desorption isotherms, measured at
77 K, using a Coulter 100 instrument.
2. Experimental
2.1. Materials and catalyst preparation
Ce(NO3)3 ·6H2O
(SEMCO,
Mumbai;
99.9%),
ZrO(NO3)2 ·xH2O and Cu(NO3)2 · 3H2O (LOBA Chemie,
Mumbai; 99%) were used as sources of CeO2, ZrO2, and
CuO, respectively.
Temperature-programmed reduction (TPR) experiments
were performed using a Micromeritics AutoChem 2910 in-

P. Ratnasamy et al. / Journal of Catalysis 221 (2004) 455–465
457
strument. A weighed amount (20 mg) of the sample was
placed in a quartz reactor and treated with a O2 (22%)–He
(78%) gas mixture at 773 K for 2 h. A gas mixture of H2
(5%)–Ar (95%) was then passed (30.22 ml/min) through the
reactor. The temperature was raised to 1200 K at a heating
rate of 10 K/min. A standard CuO powder was used to cali-
brate the amount of H2 consumption.
Diffuse reflectance UV–visible (DRUV–visible) spectra
were recorded on a Shimadzu UV-2550 spectrophotome-
ter equipped with an integrating sphere attachment (ISR-
2200). Spectral grade BaSO4 was the reference material. In
situ EPR measurements were done on a Bruker EMX spec-
trometer operating at X-band frequency (ν ≈ 9.4 GHz) and
100-kHz field modulation. The calcined sample (40 mg) was
taken in a specially designed quartz EPR cell (o.d. = 4 mm)
fitted with greaseless stopcocks and provision for adsorp-
tion/desorption of gases. Prior to EPR measurements, the
samples were degassed at 773 K for 6 h (10−3 Torr) and then
treated with dry hydrogen (20 ml/h) for 1 h at elevated tem-
peratures. The spectra were recorded at both 298 and 77 K.
No marked changes in spectral resolution were observed at
lower temperatures.
Fig. 1. XRD profiles of CuO–CeO –ZrO samples.
2
2
Magnetic susceptibility measurements at 298 K (10–
15 mg catalyst) were carried out with a Lewis-coil force
magnetometer (Series 300, George Associates, USA) cou-
pled with a high-vacuum system (10−6 Torr) with provision
for in situ treatment in desired gases at elevated tempera-
tures. Corrections for sample holder and ferromagnetic im-
purities, if any, were also made.
3. Results and discussion
3.1. Structural and textural studies
Fig. 1 shows the XRD patterns of CeO2, CeO2 (50%)–
ZrO2 (50%), CuO (5%)–CeO2 (95%), and CuO (5%)–CeO2
(47.5%)–ZrO2 (47.5%). The reflections in the 2θ range of
28–80◦ indicate a distinct, cubic, fluorite structure [24–
29]. In CeO2–ZrO2-based catalysts the reflections shifted to
higher 2θ values compared to CeO2 (shown by dotted lines
in Fig. 1). The unit cell parameter (Table 1) of CeO2–ZrO2
is smaller (0.531 nm) than that of CeO2 (0.548 nm), suggest-
ing the formation of ceria–zirconia solid solutions [31]. The
ZrO2-based catalysts showed broad reflections and hence,
the unit cell and average crystallite size parameters could not
be determined with precision. No separate phase due to CuO
or Cu2O was detected, except in the samples with 10 wt%
CuO. A majority of the total CuO is present at the surface of
the support.
The presence of zirconium decreased the crystallite size
of ceria as well as CuO (Table 1). The crystallite size of
ceria decreased from 8.4 (for CuO (10%)–CeO2 (90%)) to
3.8 nm (for CuO (10%)–CeO2 (45%)–ZrO2 (45%)). The
corresponding decrease in the crystallite size of CuO was
from 4.1 to 3.2 nm. When CuO was also present the crystal-
lite size of ceria increased from 4.9 nm (for pure CeO2) to
8.4 nm (for CuO (10%)–CeO2 (90%)).
2.3. Catalytic activity studies
PROX reactions were carried out in a fixed bed, down-
flow, glass reactor (i.d. = 15 mm) using 1.5 g of the calcined
catalyst. The sample was made into a pellet (without any
binder), crushed, and sized (10–20 mesh). Then, it was acti-
vated at 573 K, for 3 h, in flowing hydrogen and then cooled
to 403–473 K. A synthetic feed gas mixture of volume com-
position H2 (74.17%)+CO (0.49%)+CO2 (23.26%)+CH4
(2.08%) was passed through a mass flow controller at a spec-
ified flow rate (2.5–10 liters/h; GHSV = 5000–20,000h−1
)
along with the required amount of oxygen in air (O2/CO =
0.5–1.5). Experiments were also conducted with a stream
containing added water (1.25 ml/h; p(H2O) = 200 Torr)
or an actual effluent from a steam reformer-cum-water–gas-
shift reactor processing LPG was passed directly (after ad-
dition of oxygen) into the PROX reactor. The products were
analyzed online using a CHEMITO 8610 gas chromatograph
equipped with a methanator. A Spherocarb column (1/8 inch
diameter × 8 feet length) was used for analysis. A calibrated
gas was used as a reference to estimate the concentrations of
gases.
SBET values increased in the series: CeO2 < CeO2–
ZrO2 < ZrO2 (Table 1). When CuO was also present, the
CuO–CeO2 samples showed a marked decrease in surface
area (from 150 to 117 m2/g). The decrease is not significant
in CuO–CeO2–ZrO2 (Table 1). Zirconia incorporation in the

458
P. Ratnasamy et al. / Journal of Catalysis 221 (2004) 455–465
Table 1
Structural parameters (from XRD) and BET surface areas of CuO-CeO –ZrO composites
2
2
a
Catalyst
Unit cell
parameter of the
support oxide (nm)
Average
crystallite size of the
support oxide (nm)
Unit cell
parameter
of CuO (nm)
Average
crystallite size
of CuO (nm)
S
BET
2
(m /g)
CeO
0.548
0.531
–
4.9
4.4
–
–
–
–
–
–
–
150
161
172
2
CeO (50%)–ZrO (50%)
2
2
ZrO
2
CuO (1%)–CeO (99%)
0.547
0.545
0.542
6.6
7.8
8.4
–
–
–
–
4.1
117
129
124
2
CuO (5%)–CeO (95%)
2
CuO (10%)–CeO (90%)
0.438
2
CuO (1%)–CeO (49.5%)–ZrO (49.5%)
0.532
0.533
0.530
4.4
4.2
3.8
–
–
–
–
3.2
168
148
136
2
2
CuO (5%)–CeO (47.5%)–ZrO (47.5%)
2
2
CuO (10%)–CeO (45%)–ZrO (45%)
0.437
2
2
b
b
CuO (1%)–ZrO (99%)
–
–
–
–
–
–
–
–
–
–
–
–
161
153
161
2
CuO (5%)–ZrO (95%)
2
b
CuO (10%)–ZrO (90%)
2
a
Values in parentheses indicate composition in wt%.
XRD peaks were broad.
b
support, hence, increases the surface area of the samples in
agreement with the XRD results where formation of smaller
crystallites is found in CeO2–ZrO2-based catalysts (Table 1;
compare rows 7–9 with 4–6).
3.2. Redox behavior, spectral and magnetic properties
3.2.1. Temperature-programmed reduction
Pure ZrO2 could not be reduced in H2 in the tempera-
ture range 298–1200 K. CeO2 showed two peaks at 658 and
794 K attributable to the reduction of surface and bulk ceria
compositions, respectively [23,28]. In the presence of zir-
conia the reducibility of ceria increased and the reduction
temperature decreased. In CeO2–ZrO2, the reduction of sur-
face and bulk ceria occurred at 620 and 804 K, respectively.
Reduction of CuO supported on CeO2, CeO2–ZrO2 and
ZrO2 started above 373 K [28,29]. In pure CuO this oc-
curred above 523 K. At least 2 to 3 reduction peaks were
observed in the supported catalysts, indicating the exis-
tence of different type of CuO species. CuO (5%)–ZrO2
(95%) showed overlapping reduction peaks at 465 K and
450 K (Fig. 2). These peaks in CuO (5%)–CeO2 (95%) ap-
peared at lower temperatures, 445 and 425 K, respectively
(Fig. 2). In CuO (5%)–CeO2 (47.5%)–ZrO2 (47.5%), they
occurred at still lower temperatures, 435 and 398 K, respec-
tively (Fig. 2). With different supports, the reducibility of
CuO increased in the order ZrO2 < CeO2 < CeO2–ZrO2.
TPR of CuO supported on CeO2 and ZrO2 was investigated
by several groups [32–34]. They observed CuO reduction
peaks in the range 398–448 K and above 473 K. The former
was attributed to arise from dispersed copper oxide clusters,
strongly interacting with the support while the latter (above
473 K) from CuO particles. The TPR peaks for our catalysts
(Fig. 2) reveal that the latter type of CuO particles is not
present in our catalyst samples and the CuO is mostly well
dispersed.
Fig. 2. TPR of CuO (5%)–ZrO (95%) (···), CuO (5%)–CeO (95%) (- - -),
2
2
and CuO (5%)–CeO (47.5%)–ZrO (47.5%) (—).
2
2
3.2.2. Electronic spectroscopy
Both pure CeO2 and CeO2–ZrO2 showed two UV bands
at 257 and 338 nm attributable to O2− → Ce4+ charge trans-
fer (CT) transitions. The former is a specular reflectance
band arising from the surface sites and the latter is a dif-
fuse reflectance band arising from bulk ceria [35]. Pure ZrO2
showed only one band at 215 nm. The extinction coeffi-
cients of these bands were different for different supports
and decreased in the order CeO2 > CeO2–ZrO2 > ZrO2.
The intensity of the 338-nm band in CeO2–ZrO2 (vis-à-
vis pure CeO2) increased due to the substitution of Zr4+ in
the CeO2 lattice and the consequent strain and lowering of
symmetry at the site of cerium. These CT bands shifted to
the lower energy side (higher wavelength) when CuO was

P. Ratnasamy et al. / Journal of Catalysis 221 (2004) 455–465
459
Fig. 3. DRUV–visible spectra showing the influence of support composition on the charge transfer (left) and d–d (right) transitions.
also present (Fig. 3, left panel). The shift is more noticeable
in CuO–CeO2 and CuO–ZrO2 than in CuO–CeO2–ZrO2.
In CuO (5%)–CeO2 (95%), these CT appeared at 267 and
345 nm (compared to 257 and 338 nm, respectively). In CuO
(5%)–CeO2 (47.5%)–ZrO2 (47.5%), they appeared at 260
and 345 nm. In CuO (5%)–ZrO2 (95%), the 215-nm band
shifted to 270 nm. These red shifts in the CT bands reveal
an increase in the crystallite size of CeO2 in the presence of
CuO. These observations reinforce similar conclusions from
the XRD data (Table 1).
The CuO-containing catalysts showed a broad, d–d band
in the visible region (Fig. 3, right panel). This band shifted
from 780 (Cu2+ in sixfold coordination) to 680 nm (Cu2+
in fourfold coordination) with a change in the support from
ZrO2 to CeO2–ZrO2 to CeO2, respectively [36–38]. A simi-
lar shift (from 800 to 763 nm in CuO–ZrO2, 768 to 645 nm
in CuO–CeO2–ZrO2, and 670 to 650 nm in CuO–CeO2) was
observed with an increase (from 1 to 10 wt%) in the CuO
content, indicating that the type of copper species in these
catalysts is influenced by both the support and the concen-
tration.
Species A and B could be generated, repeatedly, in sev-
eral cycles, indicating the facile redox behavior of CeO2 and
CeO2–ZrO2 supports. TPR had indicated that the reduction
of ceria (not shown in figure) is facilitated in the presence of
zirconia. The reduction of ceria occurred at 620 K in CeO2–
ZrO2 compared to 658 K in CeO2. Hence, higher amounts of
Ce3+ are formed in reduced CeO2–ZrO2 composites than in
pure CeO2. Both EPR and TPR studies reveal that zirconium
incorporation enhances the concentration of Ce3+
.
3.2.3.2. Redox behavior of supported copper At least four
types of paramagnetic copper species (I–IV) could be identi-
fied in the EPR spectrum of CuO (1%)–CeO2 (99%) (Fig. 5i,
curve a). Types I and II correspond to isolated, monomeric
Cu2+ species characterized by axial spectra with resolved
Cu hyperfine features in the parallel and perpendicular re-
gions [24–26,41,42]. These species are different in their g
ꢀ
and A (Cu) values (type 1: g = 2.300, A (Cu) = 126.4 G;
ꢀ
ꢀ
ꢀ
type 2: g = 2.339, A (Cu) = 124.8 G). Their g (2.049)
ꢀ
ꢀ
⊥
and A (Cu) (10.0 G) values are similar. The linewidth
⊥
of species I is smaller than that of species II (curve a).
The spin Hamiltonian parameters indicate that the geom-
etry around Cu2+ is distorted octahedral. These species (I
and II) are perhaps located in the lattice and surface sites
of ceria crystallites, respectively [41]. The type III species
characterized by a broad signal (linewidth = 210 G) cen-
tered at gav = 2.10 is attributed to dipolar, interacting Cu2+
ions forming a nano-sized 2-dimensional structure [41].
The relative concentration of the type III species increased
with copper content (Fig. 5ii). The visible band at 645–
680 nm in the diffuse reflectance spectrum corresponds
to this species. The type IV species showed two signals
at 3017 and 3716 G corresponding to a single-quantum,
perpendicular transitions of Cu2+–Cu2+ dimers at the sur-
face [42]. The exchange coupling constant (J) for these
dimers (type IV) was −52.5 cm−1 and the Cu–Cu separa-
3.2.3. EPR spectroscopy
3.2.3.1. Redox behavior of the support EPR spectra of
CeO2 reduced with dry hydrogen revealed the formation
of two types of paramagnetic species, A, due to Ce3+ [39]
−·
(g⊥ = 1.965 and g = 1.940) and B, due to O2 ions [40]
(g1 = 2.030, g2 =ꢀ2.016, and g3 = 2.011), the concentra-
tion of which depended on the reduction temperature (Tr)
(Fig. 4, left panel). Intensity of A signals reached a maxi-
mum at 573 K. Pure ZrO2 did not form these paramagnetic
species when reduced in hydrogen. CeO2–ZrO2 (Fig. 4, right
panel) behaved in a manner similar to pure CeO2. Upon re-
oxidation by exposure to air at 298 K, the signals due to
species A disappeared and of B decreased and broadened.

460
P. Ratnasamy et al. / Journal of Catalysis 221 (2004) 455–465
3+
−•
Fig. 4. EPR spectra (at 298 K) of Ce (A) and O
(B) ions generated on CeO (left) and CeO –ZrO (right) by reaction with H at elevated temperatures.
2 2 2 2
2
Fig. 5. (Left) EPR spectra of CuO (1%) supported on (a) CeO , (b) CeO –ZrO , and (c) ZrO . Signals due to different Cu species (I, II, III, and IV) are
2
2
2
2
indicated by arrows. (Right) Effect of CuO content on the EPR spectra of CuO (x%)–CeO ((100 − x)%) catalysts.
2
tion was 3.4 Å [42]. If the two copper ions of the dimer
were to occupy adjacent Ce4+ locations in the fluorite lat-
tice, their Cu–Cu separation would have been 5.41 Å. It
was, hence, proposed that these interacting, Cu2+–Cu2+
dimers in copper oxo species are at the surface of ce-
ria [41,42].
CuO (1%)–ZrO2 (99%), on the other hand, contained
only two types of Cu2+ species, viz. isolated Cu2+ ions
(II) and interacting Cu2+ ions (III) (Fig. 5i, curve c). The
spin Hamiltonian parameters of these species are different
from CuO (1%)–CeO2 (99%) (Table 2). Unlike in CuO
(1%)–CeO2 (99%) (Fig. 5i, curve a), the perpendicular Cu-
hyperfine features could not be resolved in CuO (1%)–ZrO2
(99%) (Fig. 5i, curve c). The type III species of CuO–ZrO2
showed a signal at gav = 2.12.
CuO (1%)–CeO2 (49.5%)–ZrO2 (49.5%) showed spec-
tra (Fig. 5i, curve b) intermediate in shape to that of CuO
(1%)–CeO2 (99%) and CuO (1%)–ZrO2 (99%). Cu species,
types I, II, and III were detected. Similar to CuO (1%)–CeO2
(99%), the type I species of CuO (1%)–CeO2 (49.5%)–ZrO2
(49.5%) showed resolved Cu-hyperfine features at parallel
and perpendicular regions (Table 2). Signals of CuO (1%)–
CeO2 (49.5%)–ZrO2 (49.5%) were broader. Cu2+–Cu2+
dimers (type IV) could not be detected. Thus, different types
of CuO species were present in different amounts on CeO2,
ZrO2, and CeO2–ZrO2.

P. Ratnasamy et al. / Journal of Catalysis 221 (2004) 455–465
461
Table 2
EPR spin Hamiltonian parameters
Catalyst
Treatment
Assignment
Type
g
or g
g
or g , g
A (Cu) (G)
A
⊥
(Cu) (G)
ꢀ
1
⊥
2
3
ꢀ
3+
4+
CeO
H
reduction
Ce
A
B
1.940
2.030
1.965
–
–
–
–
2
2
−·
Ce –O
2.011, 2.016
2
2
3+
CeO (50%)–ZrO (50%)
H
reduction
Ce
A
B
1.940
2.028
1.964
–
–
–
–
2
2
2
4+
−·
Ce –O
2.009, 2.014
2+
CuO (1%)–CeO (99%)
–
–
Isolated Cu in lattice site
I
II
III
IV
2.300
2.339
2.10
2.049
2.049
2.10
126.4
124.8
Not resolved
10
2
2+
Isolated surface Cu
Interacting Cu
Isolated dimer
Not resolved
Not resolved
2+
a
2+
CuO (1%)–ZrO (99%)
Isolated surface Cu
Interacting Cu
II
III
2.336
2.12
2.063
2.12
130.0
Not resolved
Not resolved
Not resolved
2
2+
2+
CuO (1%)–CeO (49.5%)–
ZrO (49.5%)
2
–
–
Isolated Cu in lattice site
I
II
III
2.300
2.339
2.13
2.060
2.060
2.13
126.4
124.8
Not resolved
10
2
2+
Isolated surface Cu
Interacting Cu
Not resolved
Not resolved
2+
a
g
signals due to type IV were observed at 3017 and 3716 G; g signals were weak and masked by types I–III.
⊥
ꢀ
duced and reoxidized is more on CeO2 than on CeO2–ZrO2
support. The facile redox behavior and the abundance of the
reactive copper species influence the CO oxidation activity
of these catalysts (vide supra). In the absence of CuO, the
supports CeO2 and CeO2–ZrO2 generated Ce3+ ions and
−·
O2 species. Such species were not detected in supported
CuO catalysts.
3.2.4. Magnetic susceptibility
Prior to magnetic susceptibility measurements the sam-
ples were first evacuated and then purged with He (5–
10 ml/h) and again evacuated. Molar susceptibilities (χM)
of ZrO2, CeO2–ZrO2, and CeO2 after diamagnetic correc-
tion are 0.026×10−3, 0.029×10−3, and 0.060×10−3 c.g.s.
units, respectively. These values indicate weak paramag-
netism. The Mn impurity and paramagnetic ions like Ce3+
and defect centers generated by evacuation/dehydroxylation
are the probable causes for the weak paramagnetism of the
evacuated samples. The variation in magnetic susceptibility
(χM) of the samples with increasing CuO content is shown
in Fig. 7. The support influences the dispersion and mag-
netic state of copper significantly. From the linear increase in
magnetic susceptibility, it may be concluded that the concen-
tration of isolated, noninteracting Cu2+ ion increases with
copper content in CeO2–ZrO2 (it may be recalled that Cu2+
is mainly responsible for the observed magnetic suscepti-
bility). The presence of Cu1+ and interacting Cu2+ ions (at
high concentrations) in CeO2 and ZrO2 probably account for
their lower χM values at CuO content above 2 wt% (Fig. 7).
In other words, in samples containing more than 2 wt%
CuO, larger amounts of interacting Cu2+ ions/bulk CuO-like
phases are formed on ZrO2 and CeO2 supports compared to
CeO2–ZrO2. Apparently, CeO2–ZrO2 promotes a better and
more stable dispersion of copper oxide. These results rein-
force the conclusion from XRD (Table 1) that the average
crystallite size of CuO on CeO2–ZrO2 is indeed smaller than
that on CeO2.
2+
Fig. 6. Variation of Cu signal intensity of CuO (1%)–CeO (99%), CuO
2
(1%)–CeO (49.5%)–ZrO (49.5%), and CuO (1%)–ZrO (99%) as a func-
tion of reduction temperature.
2
2
2
The reducibility of CuO on various supports increased
in the order CuO–ZrO2 < CuO–CeO2 < CuO–CeO2–ZrO2
(Fig. 6). When Cu–CeO2 and Cu–CeO2–ZrO2 were reduced
at 573 K, almost all the copper was reduced. Reoxidation
by exposing the samples to air regenerated the EPR spec-
trum of types I, II, and IV species. Type III species could
not be regenerated. The reduction–reoxidation of types I,
II, and IV species (isolated/dispersed copper) could be re-
peated several times. A similar observation was noted also
on CeO2–ZrO2 supports. While 39.7% of total intensity
could be regenerated upon reduction followed by reoxida-
tion in CuO–CeO2, it was only 31.4% in CuO–CeO2–ZrO2.
Hence, the amount of reactive copper oxide that could be re-

462
P. Ratnasamy et al. / Journal of Catalysis 221 (2004) 455–465
3.3. Preferential oxidation of CO (PROX)
Catalytic activities of CuO supported on CeO2, CeO2–
ZrO2, and ZrO2 under different reaction conditions (e.g.,
temperature, O2/CO ratio, space velocity, and time on
stream) are presented in Tables 3–5 and Figs. 8–10. PROX
selectivity is defined as (oxygen consumed in CO oxida-
tion) ×100/(oxygen consumed in the oxidation of CO and
H2). Under our reaction conditions oxidation of CH4 was
not observed. Oxygen conversion was more than 99.8%. The
following conclusions are drawn:
1. CO oxidation increased with CuO content reaching a
maximum at 5 wt% (Table 3; rows 1–4). CuO supported
on CeO2 exhibited superior activity than CuO–CeO2–
ZrO2 and CuO–ZrO2 (Table 3; compare rows 3, 5 and
6).
2. With a water-containing feed (p(H2O) = 200 Torr), a
decrease in the CO oxidation activity was observed. This
decrease is more significant with ZrO2-containing cata-
lysts (Table 3, rows 5 and 6) and in the catalysts with a
lower CuO content (rows 1 and 2).
Fig. 7. Plot of molar magnetic susceptibility (χ ) as a function of CuO
M
content in CuO–CeO , CuO–CeO –ZrO , and CuO–ZrO .
2
2
2
2
3. The adverse influence of H2O on CO oxidation activity
diminished above 423 K (Table 4).
Table 3
Preferential oxidation of CO over CuO–CeO –ZrO catalysts
2
2
b
Catalyst
Feed without water
Feed with water
a
a
CO
H
Selectivity
(%)
CO
H
Selectivity
(%)
2
2
Composition Conversion Composition Conversion
Composition Conversion Composition Conversion
ppm
(%)
(v%)
(%)
ppm
(%)
(v%)
(%)
CuO (1%)–CeO (99%)
98
82
7
64
89
98.0
98.3
99.9
98.7
98.2
73.4
73.4
73.4
73.4
73.4
1.0
1.0
1.0
1.0
1.0
38.4
38.5
39.1
38.7
38.5
3070
1620
17
80
495
37.3
66.9
99.7
98.4
89.9
73.1
73.3
73.4
73.4
73.4
1.4
1.2
1.0
1.0
1.0
14.6
26.2
39.0
38.6
35.2
2
CuO (2%)–CeO (98%)
2
CuO (5%)–CeO (95%)
2
CuO (10%)–CeO (90%)
2
CuO (5%)–CeO (47.5%)–
2
ZrO (47.5%)
2
CuO (5%)–ZrO (95%)
2950
39.8
73.1
1.4
15.6
3500
28.6
73.1
1.5
11.2
2
−1
Reaction conditions: catalyst = 1.5 g, feed, H (74.17%) + CO (0.49%) + CO (23.26%) + CH (2.08%); O /CO = 1.25, GHSV = 5000 h ; temperature =
2
2
4
2
423 K; run time = 4 h.
a
Selectivity of oxygen toward CO oxidation = (oxygen consumed in CO oxidation) ×100/(total oxygen consumed in CO + H oxidation).
2
b
p(H O) = 200 Torr.
2
Table 4
Effect of temperature on the PROX activity over CuO (5%)–CeO (47.5%)–ZrO (47.5%)
2
2
b
Temperature
(K)
Feed without water
Feed with water
a
a
CO
H
Selectivity
(%)
CO
H
Selectivity
(%)
2
2
Composition Conversion Composition Conversion
Composition Conversion Composition Conversion
ppm
(%)
(v%)
(%)
ppm
(%)
(v%)
(%)
373
398
423
448
473
2760
775
89
443
930
43.7
84.1
98.2
90.9
81.0
73.6
73.5
73.4
73.4
73.3
0.8
0.9
1.0
1.0
1.1
27.0
38.1
38.5
35.7
31.8
3610
2330
495
140
852
26.3
52.4
89.9
97.1
82.6
73.6
73.5
73.4
73.4
73.3
0.8
0.9
1.0
1.0
1.1
18.5
27.7
35.2
38.1
32.4
−1
Reaction conditions: catalyst = 1.5 g, feed, H (74.17%) + CO (0.49%) + CO (23.26%) + CH (2.08%); O /CO = 1.25, GHSV = 5000 h
.
2
2
4
2
a
Selectivity of oxygen toward CO oxidation = (oxygen consumed in CO oxidation) ×100/(total oxygen consumed in CO + H oxidation).
2
b
p(H O) = 200 Torr.
2

P. Ratnasamy et al. / Journal of Catalysis 221 (2004) 455–465
463
Table 5
Preferential oxidation of CO in the effluent from LTS on CuO–CeO –ZrO
2
2
a
Catalyst
CuO (5%)–CeO (95%)
Conversion of CO (%)
Conversion of H (%)
2
Selectivity
99.95
99.5
53.0
2.8
0.8
0.8
20.3
47.0
32.1
2
CuO (5%)–CeO (47.5%)–ZrO (47.5%)
2
2
CuO (5%)–ZrO (95%)
2
−1
Reaction conditions: average feed composition, 5000 ppm CO + 20% CO + 70% H + balance H O + CH ; GHSV = 20,000 h , catalyst = 1.5 g,
2
2
2
4
O /CO = 1.25; temperature = 423 K.
2
a
Selectivity of oxygen toward CO oxidation = (oxygen consumed in CO oxidation) ×100/(total oxygen consumed in CO + H oxidation).
2
(Fig. 10, left panel). As expected the PROX selectivity
was high (94.9%) at a low value of O2/CO.
8. CO oxidation increased with an increase in O2/CO ra-
tio. Almost complete conversion of CO to CO2 was
observed beyond a O2/CO ratio of 1.25 (Fig. 10, right
panel).
9. Table 5 illustrates the PROX activity for a feed con-
sisting of the effluent from a fuel processor (steam
reformer-cum-high temperature (HTS)-cum-low tem-
perature water–gas-shift (LTS) reactors combination).
Indian LPG (liquefied petroleum gas), comprising most-
ly C3 and C4 hydrocarbons (both saturates and olefins),
was the feed to the steam reformer. 40% NiO–30%
CeO2–30% ZrO2 was the steam-reforming catalyst op-
erating at 923 K, GHSV = 20,000 h−1, and steam/
carbon = 3.0 [44]. Both the HTS (Fe–Cr oxides) and
LTS (CuO–ZnO–Al2O3) catalysts were commercial
samples from Sud-Chemie (India) operated under con-
ditions recommended by the catalyst manufacturer. With
this realistic feed, CuO–CeO2–ZrO2 exhibited higher
Fig. 8. Influence of support and temperature on the conversion of CO
PROX selectivity than CuO–CeO2 and CuO–ZrO2.
to CO . Reaction conditions: catalyst = 1.5 g; feed (H (74.17%) + CO
2
2
Long-term experiments with this feed (27 h) also con-
firmed the stability of this catalyst even in the presence
of realistic amounts of CO2 and H2O. The CO content
in the effluent from the PROX reactor was consistently
below 10 ppm for CuO–CeO2 and 50 ppm for CuO–
CeO2–ZrO2.
(0.49%)+CO (23.26%)+CH (2.08 v%)) + water (p(H O) = 200 Torr);
2
4
2
−1
O /CO = 1.25, GHSV = 5000 h ; run time = 4 h: (a) CuO (5%)–CeO
2
2
2
(47.5%)–ZrO (47.5%); (b) CuO (5%)–CeO (95%); (c) CuO (5%)–ZrO
2
2
(95%); (d) CeO ; (e) CeO (50%)–ZrO (50%).
2
2
2
4. The CO oxidation activity/selectivity increased with
temperature reaching a maximum at 448–478 K (Ta-
ble 4) due either to increased H2 oxidation at higher tem-
peratures and/or to the growth of CuO particles above
around 470 K [43].
5. The support has a marked effect on the catalytic ac-
tivity (Fig. 8). With water-containing feed (p(H2O) =
200 Torr), the CO oxidation activity decreased in the
order CuO–CeO2 > CuO–CeO2–ZrO2 > CuO–ZrO2.
Copper-free supports also catalyzed the oxidation of CO
and H2. But their activities were considerably lower
than the copper-containing catalysts. H2 oxidation takes
place on both the metal and the support.
Both the PROX activity and selectivity of the CuO are
significantly affected by interactions with the underlying
support. Bera et al. [45] attributed the promoting effect of
CeO2 in the CuO–CeO2–ZrO2 system for CO oxidation to
the lowering of the redox potentials of the Cu2+/Cu1+ and
Cu1+/Cu0 couples in a CeO2 matrix in relation to CuO and
CuO–ZrO2. In agreement with our results (TPR and EPR)
they also observed that the Cu species in a CuO–CeO2 ma-
trix require less energy to be reduced and reoxidized than
in the case of pure CuO and CuO–ZrO2. The oxidation of
CO on metals is known to obey a single site Langmuir–
Hinshelwood kinetic model [46] where CO and O2 compete
for the same sites. In the case of a metal–reducible oxide sys-
tem (like our reduced Cu–CeO2–ZrO2), we may expect that
while CO may adsorb on copper sites, O2 (and H2 and H2O)
may adsorb on both the metal (Cu) and support (CeO2–
ZrO2) (Fig. 8). The availability of additional surface oxygen
6. CuO–CeO2 and CuO–CeO2–ZrO2 samples exhibited a
stable activity in long-term (18 h) experiments (Fig. 9).
7. At high space velocities, in the presence of water in
the feed, while the H2 conversion was not affected, the
CO conversion and PROX selectivity were decreased

464
P. Ratnasamy et al. / Journal of Catalysis 221 (2004) 455–465
Fig. 9. Influence of time on stream and water in feed on CO and H conversions and CO selectivity. Reaction conditions: catalyst = 1.5 g; feed (H
2
2
−1
(74.17%) + CO (0.49%) + CO (23.26%) + CH (2.08 v%)) + water (p(H O) = 200 Torr); O /CO = 1.25; GHSV = 5000 h : catalyst: CuO (5%)–CeO
2
4
2
2
2
(95%) (left panel); CuO (5%)–CeO (47.5%)–ZrO (47.5%) (right panel).
2
2
Fig. 10. Effect of space velocity (GHSV; left panel) and O /CO (right panel) on the PROX activity/selectivity. Reaction conditions: catalyst (CuO (5%)–CeO
2
2
(95%) = 1.5 g; feed (H (74.17%) + CO (0.49%) + CO (23.26%) + CH (2.08 v%)) + water (p(H O) = 200 Torr); run time = 4 h.
2
2
4
2
and the consequent higher activity of CuO–CeO2 and CuO–
CeO2–ZrO2 in CO oxidation are, thus, understandable.
Both the XRD and magnetic susceptibility data indi-
cate that the dispersion of supported CuO decreases in the
series CeO2–ZrO2 > CeO2 > ZrO2. EPR and magnetic
susceptibility data indicate that while a small amount of
copper is incorporated in the CeO2/ZrO2, fluorite lattice
(forming a solid solution), most of it is present on the sur-
face as isolated, monomeric copper as well as dimeric cop-
per oxo species containing magnetically interacting copper
ions, nano-sized copper oxide clusters, and a bulk, CuO-like
phase. While the isolated and dispersed CuO species exhibit
reversible reduction-oxidation behavior, the bulk CuO-like
phase exhibits irreversible reduction behavior. The amount
of reducible and reoxidizable CuO content is more on CeO2
(39.7%) than on CeO2–ZrO2 catalysts (31.4%) (EPR). The
concentration of the bulk-like CuO phase decreases in the
series ZrO2 > CeO2 > CeO2–ZrO2. The PROX selectivity
of these samples correlated with the concentration of the dis-
persed nano-sized CuO species.
4. Summary and conclusions
In this study, the influence of composition and prepara-
tion procedure in the CeO2–ZrO2 system on the structure,
reduction behavior, and catalytic activity of supported CuO
in the preferential oxidation of CO in the presence of excess
H2 was investigated. The CO oxidation activity/selectivity
of CuO on different supports increases in the order CuO–

P. Ratnasamy et al. / Journal of Catalysis 221 (2004) 455–465
465
ZrO2 < CuO–CeO2–ZrO2 ꢀ CuO–CeO2. A catalyst of
composition containing 5% CuO exhibited stable activity in
long-term experiments, using as feed the effluent from an
actual steam reformer-cum-water–gas-shift reactor combi-
nation generating hydrogen from commercial liquefied pe-
troleum gas. The adverse influence of H2O was accentuated
in catalysts containing ZrO2. Below 423 K and over CuO–
CeO2 with less than 5 wt% CuO, the presence of H2O in
the feed suppressed CO oxidation. The high dispersion and
facile reducibility of CuO on CeO2 and CeO2–ZrO2 sup-
ports are responsible for their superior activity/selectivity
over CuO–ZrO2 catalysts.
[22] J.B. Wang, W.H. Shih, T.-J. Huang, Appl. Catal. A 203 (2000) 191.
[23] H. Vidal, J. Kašpar, M. Pijolat, G. Colon, S. Bernal, A. Cordón, V.
Perrichon, F. Fally, Appl. Catal. B 27 (2000) 49.
[24] A. Martínez-Arias, M. Fernández-García, J. Soria, J.C. Conesa,
J. Catal. 182 (1999) 367.
[25] A. Martínez-Arias, M. Fernández-García, O. Gálvez, J.M. Coronado,
J.A. Anderson, J.C. Conesa, J. Soria, G. Munuera, J. Catal. 195 (2000)
207.
[26] A. Martínez-Arias, M. Fernández-García, A.B. Hungría, A. Iglesias-
Juez, O. Gálvez, J.A. Anderson, J.C. Conesa, J. Soria, G. Munuera,
J. Catal. 214 (2003) 261.
[27] J. Xiaoyuan, L. Guanglie, Z. Renxian, M. Jianxin, C. Yu, Z. Xiaoming,
Appl. Surf. Sci. 173 (2001) 208.
[28] M.-F. Luo, Y.-J. Zhong, X.-X. Yuan, X.-M. Zheng, Appl. Catal. A 162
(1997) 121.
[29] G. Avgouropoulos, T. Ioannides, Appl. Catal. A 244 (2003) 155.
[30] H. Tanaka, S.-I. Ito, S. Kameoka, K. Tomishige, K. Kunimori, Catal.
Commun. 4 (2003) 1.
References
[31] A. Martínez-Arias, M. Fernández-García, V. Ballesteros, L.N. Sala-
manca, J.C. Conesa, C. Otero, J. Soria, Langmuir 15 (1999) 4796.
[32] W. Liu, M. Flytzani-Stephanopoulos, Chem. Eng. J. 64 (1996) 283.
[33] L. Kundakovic, M. Flytzani-Stephanopoulos, Appl. Catal. A 171
(1998) 13.
[1] D.L. Trimm, Z. Ilsen Önsan, Catal. Rev. Sci. Tech. 43 (2001) 31.
[2] T.V. Choudhary, D.W. Goodman, Catal. Today 77 (2002) 65.
[3] J.R. Rostrup-Nielson, T. Rostrup-Nielson, Cattech 6 (2002) 150.
[4] M.J. Kahlich, H.A. Gasteiger, R.J. Behm, J. Catal. 171 (1997) 93.
[5] H. Igarashi, H. Uchida, M. Suzuki, Y. Sasaki, M. Watanabe, Appl.
Catal. A 159 (1997) 159.
[34] G. Wrobel, C. Lamonier, A. Bennani, A. D’Huysser, A. Aboukais,
J. Chem. Soc., Faraday Trans. 92 (1996) 2001.
[35] A. Bensalem, F. Bozon-Verduraz, M. Delamar, G. Bugli, Appl. Catal.
A 121 (1995) 81.
[6] S.-I. Ito, T. Fujimori, K. Nagashima, K. Yuzaki, K. Kunimori, Catal.
Today 57 (2000) 247.
[36] H. Praliaud, S. Mikhailenko, Z. Chajar, M. Primet, Appl. Catal. B 16
(1998) 359.
[37] S. Velu, K. Suzuki, M. Okazaki, M.P. Kapoor, T. Osaki, F. Ohashi,
J. Catal. 194 (2000) 373.
[38] P.I. Paulose, G. Jose, V. Thomas, G. Jose, N.V. Unnikrishnan, M.K.R.
Warrier, Bull. Mater. Sci. 25 (2002) 69.
[39] J.L.G. Fierro, J. Soria, J. Sanz, J.M. Rojo, J. Solid State Chem. 66
(1987) 154.
[40] M. Che, J.F.J. Kibblewhite, A.J. Tench, M. Dufaux, C. Naccache,
J. Chem. Soc., Faraday Trans. I 69 (1973) 857.
[41] P.G. Harrison, I.K. Ball, W. Azelee, W. Daniell, D. Goldfarb, Chem.
Mater. 12 (2000) 3715.
[42] A. Aboukais, A. Bennani, C.F. Aïssi, G. Wrobel, M. Guelton, J.C.
Vedrine, J. Chem. Soc., Faraday Trans. 88 (1992) 615.
[43] M.M. Gunter, T. Ressler, R.E. Jentoft, B. Bems, J. Catal. 203 (2001)
133.
[7] W.H. Cheng, React. Kinet. Catal. Lett. 58 (1996) 329.
[8] R.M. Torres, A. Sanchez, K. Ueda, K. Tanaka, M. Haruta, J. Catal. 168
(1997) 125.
[9] Y. Teng, H. Sakurai, A. Ueda, T. Kobayashi, Int. J. Hydrogen En-
ergy 24 (1999) 355.
[10] G. Avgouropoulos, T. Ioannides, H.K. Matralis, J. Batista, S. Hocevar,
Catal. Lett. 73 (2001) 33.
[11] W. Liu, M. Flytzani-Stephanopoulos, J. Catal. 153 (1995) 304.
[12] M. Haruta, Cattech 6 (2002) 102.
[13] M.J. Kahlich, H.A. Gasteiger, R.J. Behm, J. Catal. 182 (1999) 430.
[14] G. Avgouropoulos, T. Ioannides, Ch. Papadopoulou, J. Batista, S. Ho-
cevar, H.K. Matralis, Catal. Today 75 (2002) 157.
[15] G.K. Bethke, H.H. Kung, Appl. Catal. A 194–195 (2000) 43.
[16] R.J.H. Grisel, B.E. Nieuwenhuys, J. Catal. 199 (2001) 48.
[17] M.M. Schubert, V. Plzak, J. Garche, R.J. Behm, Catal. Lett. 76 (2001)
143.
[44] P. Ratnasamy, D. Srinivas, H.S. Soni, A.J. Chandwadkar, H.S. Potdar,
C.S. Gopinathand, B.S. Rao, Stud. Surf. Sci. Catal. 135 (2001) 1270.
[45] P. Bera, S. Mitra, S. Sampath, M.S. Hegde, Chem. Commun. (2001)
927.
[18] Y.-F. Han, M.J. Kahlich, M. Kinne, R.J. Behm, Phys. Chem. Chem.
Phys. 4 (2002) 389.
[19] S.-M. Zhang, W.-P. Huang, X.-H. Qiu, B.-Q. Li, X.-C. Zheng, S.-H.
Wu, Catal. Lett. 80 (2002) 41.
[46] R.J. Farrauto, C. Bartholomew, Introduction to Industrial Catalytic
Processes, Chapman & Hall, London, 1997, Chaps. 1, 38, 39.
[20] D.H. Kim, M.S. Lim, Appl. Catal. A 224 (2002) 27.
[21] O. Korotkikh, R. Farrauto, Catal. Today 62 (2000) 249.