Catalytic reduction of NO by CO over Cu/CexZr1- xO2 prepared by flame synthesis

Article abstract of DOI:10.1016/j.jcat.2010.04.001

Mixed oxides of 4% Cu/Ce_xZr_1-xO ₂ (x = 0, 0.25, 0.50, 0.75, 1) were synthesized by flame spray pyrolysis and characterized by N₂ adsorption, XRD, XPS, O ₂-TPD and H₂-TPR. The as-prepared catalysts were assessed for the equimolar reduction of NO by CO (3000 ppm each, space velocity 50,000 h^-1). Incorporation of Zr⁴⁺ in the form of solid solution with CeO₂ stabilizes the Cu⁺ species, which was in turn beneficial for the initial reductive chemisorption of NO to N₂O. A peculiar low-temperature activity giving 40% N₂ yield was found with the composition of Cu/Ce_0.75Zr_0.25O₂ at 150 °C. This is traced to the low-temperature activation of rapid CO oxidation (as probed by in situ DRIFTS), related to the abundance of surface reactive lattice oxygen sites and their high reducibility. At 250 °C and above, a N₂ yield of >85% (and ～100% at 350 °C) was obtained for all Zr-containing catalysts i.e. Cu/Ce_xZr_{1- x}O₂ (x < 1.0), with essentially no intermediate N ₂O detected in the exhaust gas stream. An organonitrogen mechanism is occurring in this case.

Full text of DOI:10.1016/j.jcat.2010.04.001

Journal of Catalysis 272 (2010) 210–219
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Catalytic reduction of NO by CO over Cu/CexZr1ꢀxO2 prepared by flame synthesis
Runduo Zhang a,c, ,1, Wey Yang Teoh b,1, Rose Amal b, Biaohua Chen a, Serge Kaliaguine c,
*
**
a State Key Laboratory of Chemical Resource Engineering, Faculty of Chemical Engineering, Beijing University of Chemical Technology, Beijing 100029, PR China
b ARC Centre of Excellence for Functional Nanomaterials, School of Chemical Sciences and Engineering, The University of New South Wales, Sydney, NSW 2052, Australia
c Department of Chemical Engineering, Laval University, Québec, Canada G1V 0A6
a r t i c l e i n f o
a b s t r a c t
Article history:
Mixed oxides of 4% Cu/CexZr1ꢀxO2 (x = 0, 0.25, 0.50, 0.75, 1) were synthesized by flame spray pyrolysis
and characterized by N2 adsorption, XRD, XPS, O2-TPD and H2-TPR. The as-prepared catalysts were
assessed for the equimolar reduction of NO by CO (3000 ppm each, space velocity 50,000 hꢀ1). Incorpo-
ration of Zr4+ in the form of solid solution with CeO2 stabilizes the Cu+ species, which was in turn bene-
ficial for the initial reductive chemisorption of NO to N2O. A peculiar low-temperature activity giving 40%
N2 yield was found with the composition of Cu/Ce0.75Zr0.25O2 at 150 °C. This is traced to the low-temper-
ature activation of rapid CO oxidation (as probed by in situ DRIFTS), related to the abundance of surface
reactive lattice oxygen sites and their high reducibility. At 250 °C and above, a N2 yield of >85% (and
ꢁ100% at 350 °C) was obtained for all Zr-containing catalysts i.e. Cu/CexZr1ꢀxO2 (x < 1.0), with essentially
no intermediate N2O detected in the exhaust gas stream. An organonitrogen mechanism is occurring in
this case.
Received 4 December 2009
Revised 29 March 2010
Accepted 1 April 2010
Available online 13 May 2010
Keywords:
NO reduction
deNOx
Flame spray pyrolysis
Copper
Ceria–zirconia
DRIFTS
Ó 2010 Published by Elsevier Inc.
1. Introduction
thermal and hydrothermal stability limit its practical application
in automobile catalytic muffler. The narrow pores of zeolites are
Removal of NOx from high-temperature combustion exhausts is
an important subject, with significant environmental impact. NOx
emission is known for causing acid rain and photochemical smog
in urban and industrial areas [1]. Recent advances in internal com-
bustion engines were focused on improving their efficiency, but
this is often a trade-off between drive supply and pollutants emis-
sion, e.g. NOx and CO, for a gasoline engine. Owing to the increas-
ingly stringent environmental regulations, the development of new
materials as three way catalysts (TWCs) with improved deNOx
function becomes necessary.
Among the options available, the use of copper-based catalysts
for selective catalytic reduction (SCR) of NOx is an interesting one,
from the viewpoint of cost (i.e. does not involve using precious
metals) as well as efficiency. In highly isolated form, copper species
provide efficient active sites for NO chemical adsorption and
reduction, while larger CuO-like aggregates are believed to be
active in oxidizing hydrocarbons [2]. In particular, Cu/ZSM-5
zeolite is a very effective catalyst for NOx reduction because of
the atomic dispersion of active metal [3]. Nevertheless, its poor
prone to blockage by carbonaceous and sulfurous deposits, with
resulting diffusion limitation. In this regard, oxide-based catalysts
would in principle be able to circumvent some of these limitations
[4].
Ceria is an attractive support that behaves more than just an
observer. In many catalytic reactions, its facile Ce4+/Ce3+ redox cy-
cle provides high oxygen storage capacity (OSC) [5]. When doped
with Cu+ ions, ceria acts as highly active deNOx catalysts with high
resistance to H2O and SO2 poisoning [6]. A major drawback of CeO2
catalyst lies however in its low sintering temperature, especially
under practical auto-exhaust atmosphere. This is however a well-
studied problem and over the years, an array of CeO2-based binary
oxides [4,7–11] have been designed to address the sintering prob-
lem. Of these, CeO2–ZrO2 is perhaps the most intensively studied
system. The incorporation of ZrO2 in solid solution with CeO2 not
only improves the latter thermal resistance, but more importantly
its redox capacity. The incorporation of Zr4+ lowers the energy for
Ce4+ reduction, thus enhancing CeO2 reducibility and its associated
oxygen mobility.
Conventional impregnation of Cu entity by wet techniques
often results in low dispersion of the active component and in
the appearance of large agglomerated CuO clusters, which is detri-
mental to deNOx activity. Hence, the present work explores the ra-
pid synthesis of novel Cu/CexZr1ꢀxO2 nanocomposites via flame
spray pyrolysis (FSP) for potential applications in SCR of NOx.
Besides being a single-step synthesis and yielding ready-to-use
catalysts, FSP has other obvious advantages, such as proven
* Corresponding author at: State Key Laboratory of Chemical Resource Engineer-
ing, Faculty of Chemical Engineering, Beijing University of Chemical Technology,
Beijing 100029, PR China.
** Corresponding author. Fax: +1 418 656 3810/5993.
E-mail addresses: zhangrd@mail.buct.edu.cn (R. Zhang), serge.kaliaguine@gch.
ulaval.ca (S. Kaliaguine).
1
R. Zhang and W.Y. Teoh contributed equally to this work.
0021-9517/$ - see front matter Ó 2010 Published by Elsevier Inc.
doi:10.1016/j.jcat.2010.04.001

R. Zhang et al. / Journal of Catalysis 272 (2010) 210–219
211
stoichiometric formation of CexZr1ꢀxO2 solid solutions [12] and
favorable strong interactions between the highly dispersed active
component and that of the functional support [13–15]. The pur-
pose of the present work is not only to assess the deNOx capability
of the composite catalysts, but more importantly, it seeks to clarify
the structural-performance relationship of the catalyst, where all
the components in Cu/CexZr1ꢀxO2 are of functional relevance.
N2. The DRIFTS cell (Harrick) was fitted with KBr windows and a
heating cartridge that allowed samples to be heated to 400 °C.
The pure samples were dried at 120 °C for 5 h before DRIFTS mea-
surements to eliminate the physisorbed water and finely ground in
a ceramic crucible. Prior to each spectrum recording, the sample
was flushed in a 20 mL minꢀ1 of He for 60 min for background col-
lection and then exposed to a 20 mL minꢀ1 of 2% NO, or 2% CO, or
2% NO + 2% CO, balanced in He at the desired temperature. All
spectra were measured at the respective temperatures and under
continuous flow of gas, with resolution of 4 cmꢀ1 and accumula-
tion of 100 scans.
2. Experimental
2.1. Catalysts preparation
2.3. Catalyst testing
A series of 4 wt.% Cu/CexZr1ꢀxO2 (x = 0, 0.25, 0.5, 0.75, 1) mixed
metal oxides was synthesized by
a one-step FSP technique
The catalytic activity evaluation of about 100 mg Cu/CexZr1ꢀxO2
catalysts for CO + NO reaction was performed in a fixed-bed reac-
tor under an atmosphere of 3000 ppm NO + 3000 ppm CO balanced
by He at a total flow rate of 60 mL minꢀ1, corresponding to a gas
hourly space velocity of 50,000 hꢀ1. The reactor was heated exter-
nally via a tubular furnace regulated by a temperature controller
(Omega CN 3240), with a thermocouple inserted in the catalyst
bed, over the range of 100–500 °C in steps of 50 °C. Reactant and
product contents in the reactor effluent were recorded only after
steady state was achieved at each temperature step. Nitrogen oxi-
des (NO and NO2) were analyzed using a chemiluminescence NO/
NO2/NOx analyzer (Model 200 AH, API Inc.). N2O and CO gases were
monitored using a FTIR gas analyzer (FTLA 2000, ABB Inc.). N2 was
analyzed with a gas chromatograph (GC) (Hewlett Packard 5890)
equipped with a TCD and columns of molecular 13 X (45–60 mesh,
2.4 m) combined with Silicone OV-101 (100–120 mesh, 0.6 m)
operating at an oven temperature of 30 °C.
[16,17]. Copper 2-ethylhexanoate (Sigma–Aldrich) was dissolved
in xylene (Riedel de-Haen, 98%) together with the corresponding
metal precursor for the respective metal oxide support (cerium
2-ethylhexanoate, Strem, and/or zirconium propoxide, Sigma–Al-
drich). The total concentration of Ce + Zr in the liquid precursor
was set at 0.5 M in all cases. During FSP, the liquid precursor was
fed to the nozzle at a flow rate of 5 mL minꢀ1, where it was dis-
persed by a surrounding 5 L minꢀ1 flow of O2 (1.5 bar). Combustion
of the dispersed droplets was ignited by a surrounding supporting
flame (1.5 L minꢀ1 CH4/3.2 L minꢀ1 O2). Fine aerosol nanoparticles
leaving the flame were collected on a glass fiber filter (Whatmann
GF/D) aided by a vacuum pump (Alcatel SD series). The aerosol
nanoparticles were scraped from the filter for direct use as catalyst
without further treatment.
2.2. Catalyst characterization
The specific surface area of the as-prepared nanoparticles was
determined from nitrogen adsorption equilibrium isotherms
(BET) at 77 K using an automated gas sorption system (Micromer-
itics Tristar 3000) operating in continuous mode. Powder X-ray dif-
fraction (XRD) patterns were recorded using a diffractometer (a
3. Results and discussion
3.1. Crystallite and surface characterization of Cu/CexZr1ꢀxO2
As-prepared mixed oxides of 4% Cu/CexZr1ꢀxO2 were synthe-
sized rapidly in a direct flame spray pyrolysis (FSP), showing com-
position-dependent specific surface areas of 68–88 m2 gꢀ1 (Fig. 1).
In the absence of Zr component, i.e. Cu/CeO2, X-ray diffraction
Philips X’Pert MPD) and Cu Ka radiation (k = 1.5406 Å) with a
0.1° step scan from 2h = 20–90°. X-ray photoelectron (XPS) spectra
were recorded using an ESCALAB220i-XL spectrometer (VG Scien-
tific) operating at 2 ꢂ 10ꢀ9 mbar, using a monochromated Al K
a
radiation (1486.60 eV). The binding energies were calibrated with
respect to the C 1s energy of incidental graphitic carbon at
284.9 eV. The Cu2p3/2 binding energy peak area of CuO standard
relative to that of the corresponding satellite peak (ICu2þ : Isat) was
determined to be 1.89. Based on this calibration, the relative Cu+
content of Cu/CexZr1ꢀxO2 samples was calculated by subtracting
the expected Cu2+ (based on the measured Cu2+ satellite peak area)
from the overlapping Cu2p3/2 peak areas at the 928.2–938.6 eV.
Temperature-programmed analysis was carried out using a
fixed-bed reactor. Prior to temperature-programmed desorption
of O2 (O2-TPD), the samples (ꢁ50 mg) were exposed to
50 mL minꢀ1 of 10% O2/He held at 500 °C for 1 h. The samples were
then cooled down to room temperature under the same gas flow.
Physisorbed O2 was eliminated by a 20-mL minꢀ1 He purge for
40 min. TPD was conducted by flowing 20 ml minꢀ1 of He with
temperature ramping at 10 °C minꢀ1. O2 signal was simultaneously
detected and recorded online using a mass spectrometer (MS, Bal-
zers) at m/z of 32. As for temperature-programmed reduction by
hydrogen (H2-TPR), the samples were reduced under 50 mL minꢀ1
of 5% H2/Ar along with temperature rising up to 900 °C at a
5 °C minꢀ1 ramping rate. The H2 consumption in effluent was mon-
itored by thermal conductive detector (TCD).
20
15
10
5
120
100
80
60
40
20
0
0
1.0
0.8
0.6
0.4
0.2
0.0
x in Cu/CexZr1-xO2
Fig. 1. Scherrer-determined crystallite size (solid circle) and specific surface area
(open triangle) of 4% Cu/CexZr1ꢀxO2 as a function of increasing Zr content. Also
shown are the crystallite size of similarly prepared pristine CeO2 and ZrO2 (open
circle).
In situ diffuse reflectance infrared Fourier transform spectra
(DRIFTS) were recorded on a FTIR spectrometer (Bruker TENSOR
27) equipped with a high-sensitive MCT detector cooled by liquid

212
R. Zhang et al. / Journal of Catalysis 272 (2010) 210–219
Cu2+
shake-up
Cu+
Cu2+
Cu-CeO2
Cu-Ce0.5Zr0.5O2
ZrO
Cu-Ce0.75Zr0.25O2
Cu-Ce0.25Zr0.75O2
Cu/ZrO
Cu/Ce Zr
O
O
Cu/Ce Zr
O
Cu/Ce Zr
Cu/CeO
Cu-ZrO2
CeO
10
20
30
40
50
60
70
80
90
CuO
x 2.5
Bragg angle (2 theta)
965
955
945
935
925
Fig. 2. XRD patterns of as-prepared CeO2, ZrO2 and 4% Cu/CexZr1-xO2 (x = 0–1.0).
Binding Energy (eV)
Fig. 3. XPS Cu2p3/2 lines of as-prepared 4%Cu/CexZr1ꢀxO2 and that of standard CuO
sample.
(XRD) characteristic peaks belonging to well-crystallized cubic
phase CeO2 nanoparticles were obtained, with no peak associated
with Cu or any of its constituents detected despite its 4% loading
(Fig. 2). This indicates that the Cu phase was well-dispersed as will
be confirmed by the surface-sensitive X-ray photoelectron spec-
troscopy (XPS). It is interesting to note that despite the co-precip-
itation of both Cu and Ce components during its synthesis, the
presence of Cu did not alter the cell parameters of the cubic
CeO2. Similar behavior was also reported for Cu/CeO2 synthesized
by wet precipitation [18], reversed microemulsion [19] and urea–
nitrate combustion [20]. Likewise, the crystallite size of CeO2 (as
measured by X-ray line broadening) with and without the presence
of Cu was similar, 13 and 12 nm, respectively (Fig. 1). It is only at
much higher Cu loadings (>12%) that segregation of small CuO
clusters (1–2 nm) could be observed under high-resolution trans-
mission electron microscope[15].
With the addition of Zr component (x = 0.25–1.0), a gradual
shift in the Bragg peaks toward higher angle was evident. This cor-
responds nicely with the formation of cubic phase CexZr1ꢀxO2 solid
solutions with no indication of Ce or Zr rich phase segregation de-
tected, while Cu remains XRD invisible. The shift of CexZr1ꢀxO2
Bragg peaks to higher angles with respect to CeO2 is consistent
with a decrease in cell lattice parameters as Zr is incorporated in
the lattice [7]. As Ce4+ (1.01 Å) has a larger ionic radius than Zr4+
(0.80 Å), a longer distance between Ce4+ and O2ꢀ (dCeAO) would in-
deed be expected in the solid solution. As further shown in Fig. 1,
the Scherrer-determined crystallite size of CexZr1ꢀxO2 decreases
initially to 8 nm with the introduction of Zr (x = 0.75) but increases
monotonically as a higher fraction of Zr is introduced (up to 14 nm
in Cu/ZrO2). In the absence of cubic phase stabilization by Ce i.e.
Cu/ZrO2, a predominant tetragonal phase ZrO2 with a small frac-
tion of monoclinic was observed. The volume fraction of the latter
was estimated to be ꢁ0.15 [21]. In the case of pristine ZrO2, only
pure tetragonal phase was formed, as expected from prior FSP
studies [12,14].
Cu(I) surface
40
Cu(I+II) surface
Cu(I+II) bulk
30
20
10
0
0
0.2
0.4
0.6
0.8
1
x (Cu/CexZr1-xO2)
Fig. 4. XPS-determined surface composition of Cu+ and Cu2+ on 4% Cu/CexZr1ꢀxO2 as
a function of x. Solid triangles are the theoretical compositions assuming Cu, Ce and
Zr form perfect mixed oxide solid solutions.
ers the enrichment down to a value of 1.6 times for Cu/ZrO2, sug-
gesting a higher solid state solubility of Cu in ZrO2 and/or the
strong interactions of the two components, as will be further dis-
cussed below.
Among the surface Cu, there exist different fractions of Cu oxi-
dation states across all the different catalyst samples. For quantita-
tive analysis, the relatively close binding energy position of Cu2+
and that of Cu+ would render determination of their relative com-
positions by conventional XPS peak deconvolution inaccurate.
Hence, a more accurate approach is adopted in this instance, by
calibrating the relative peak areas of Cu2+ to that of the corre-
sponding satellite peak (for more details, please refer to the exper-
imental section). As shown in Fig. 4, although the total amount of
copper species (Cu+ + Cu2+) increases with Ce content, the ratio of
Cu+/Cu2+ is higher over ZrO2 support with approximately 89% of
the total surface Cu consisting of Cu+ in the Cu/ZrO2 sample, while
only 60% Cu+ was found on Cu/CeO2 surface. Stabilization of Cu+ in
the presence of Zr4+ is consistent with the strong interaction
XPS was adopted to study the speciation and dispersion of Cu
(which was invisible under XRD) on the composite catalysts. As
shown in Fig. 3, the Cu2+ species can be characterized by a
Cu2p3/2 binding energy of 934.8 eV with a shake-up or satellite
peak located at 938.6–948.6 eV, while the corresponding Cu+ has
a binding energy at 933.0 eV [22]. In all cases, the as-prepared cat-
alyst surfaces were Cu enriched, as depicted by the consistently
higher value of (Cu/M)surface (open square) relative to (Cu/M)bulk
(black triangle), where M = Ce + Zr (Fig. 4). More specifically, Cu
is most enriched on the CeO2 surface with 3.5 times higher surface
Cu/M relative to that of the bulk. The addition of Zr gradually low-

R. Zhang et al. / Journal of Catalysis 272 (2010) 210–219
213
between the two ions which led to lower surface dispersion. It was
previously proposed that the facile Ce4+/Ce3+ redox cycle could
promote the transformation of Cu+ to Cu2+ involving the following
step: Ce4+ + Cu+ ? Ce3+ + Cu2+, and hence resulting in a higher con-
tent of Cu2+ species in the Ce-containing samples [23]. However,
our XPS analysis on the Ce 3d peaks did not detect an appreciable
amount of Ce3+, at least within its limit of detection. Surface-dis-
persed Cu+ species is known to be the adsorption site for NO and
the formation of intermediate Cu+ANO adsorbed sites is crucial
for the activation of the NO molecule [24,25], whereas copper clus-
ters are less active for this deNOx process [26]. Hence, the high per-
centage of Cu+ species over Zr4+ may indeed favor catalytic NO
reduction as will be discussed below.
O2. Pristine ZrO2 alone did not exhibit any appreciable amount of
desorbed O2, suggesting that the presence of -O2 in Cu/ZrO2 to
be most likely originating from the anion vacancies created as a re-
a
sult of charge compensation during the substitutional doping of
Cu+/Cu2+ in ZrO2 lattice. Even so, the amount of
a-O2 was 5 times
lower than that previously observed for perovskite-type mixed
oxide of LaCoO3 despite the much lower specific surface area
(20–30 m2 gꢀ1) of the latter [28]. The lower amount of
a-O2 in
Cu/CexZr1ꢀxO2 probably concurs with the limited amount of substi-
tutional doping (i.e. low anion vacancies) as reflected by the Cu
surface enrichment.
The amount of b-O2 increases monotonically with the increase
in Ce content: from 75 l l
mol gꢀ1 for Cu/ZrO2 to 126 mol gꢀ1 for
Cu/CeO2 (Table 1). Two distinct b-O2 peaks could be observed at
ꢁ780 °C (b1-O2) and ꢁ860 °C (b2-O2), associated with the surface
and bulk lattice oxygen bonded to matrix metal ions, respectively.
The fact that no b-O2 was liberated in the case of pristine CeO2 and
ZrO2 suggests that the lattice distortion induced by the doping
with Cu ions lowers the energy for bulk oxygen release [29],
although the possibility also exists for provision of oxygen from
dispersed Cu. More interestingly, the b2-O2 peak, which was not
present on Cu/ZrO2, can be observed to rise as Ce content increases.
This is consistent with the general model of facile redox transfor-
mation of Ce4+ ? Ce3+ in homogeneous solid solution, which re-
sults in enhanced bulk lattice oxygen storage and exchange. In
the absence of Ce, that is Cu/ZrO2, only surface lattice oxygen could
be liberated, while the bulk lattice oxygen was comparatively less
mobile. Such effect of Ce is further exemplified by the H2-TPR stud-
ies presented in the next section.
3.2. Temperature-programmed desorption (TPD) of O2
Besides Cu speciation, the state of O species and their quantities
are equally important characteristics in the SCR of NOx. This can be
probed by the thermal desorption spectra of oxygen as illustrated
in Fig. 5, revealing the presence of a-O2 i.e. oxygen held on the cat-
alyst surface anion vacancies with a desorption temperature usu-
ally below ꢁ700 °C; as well as the b-O2 desorption which is
related to the release of lattice oxygen occurring usually at temper-
ature above 700 °C [27]. Quantification of the deconvoluted peaks
of the different desorbed oxygen is summarized in Table 1.
In most cases, the amount of
a-O2 was marginal (<1.5
l
mol gꢀ1
)
with the exception of Cu/ZrO2, which contained 7.4
l
mol gꢀ1 of
a-
3.3. Temperature-programmed reduction (H2-TPR)
β
1-O2
The reducibility of Cu/CexZr1ꢀxO2 samples was investigated by
quantitative H2-TPR (Fig. 6, Table 2). In the case of Cu/CeO2, three
β
2-O2
reduction peaks at 138 (a-H2), 153 (b1-H2) and 178 °C (b2-H2)
α-O2
together with another one appearing at 810 °C were evident in
its TPR profile (Fig. 6). The former three peaks are related to the
speciation of the supported and highly dispersed Cu [18,20,
30,31]. These reduction temperatures are significantly lower than
that of the bulk and unsupported CuO, which reduces at ꢁ300 °C
[20]. The onset reduction temperature (Tonset, temperature at
which reduction of catalyst begins) reflects the presence of the
most easily reduced Cu species, which in the case of Cu/CeO2
was determined in our earlier studies to be the CuO dimer at
Tonset ꢁ 105 °C [15]. Note also that when compared to our earlier
publication [15], the current set of TPR has been optimized to re-
Cu/CeO2
Cu/Ce0.75 Zr0.25O2
Cu/Ce0.5 Zr0.5O2
Cu/Ce0.25Zr0.75O2
Cu/ZrO2
CeO2
ZrO2
solve the a- and b1-H2, which was previously embedded as a broad
hump. For all Cu/CexZr1ꢀxO2 samples containing Ce (x = 0.25–1.0),
the Tonset was consistently ꢁ105 °C, reflecting the presence of di-
meric CuO species, while the Tonset ꢁ 125 °C in Cu/ZrO2 reflects dif-
ferent and less reducible Cu species. In principle, the speciation of
Cu is strongly influenced by the interactions of Cu and the oxide
support, and although not shown here, speciation studies on the
samples by solid state electron paramagnetic resonance confirmed
the gradual transformation from Cu dimers in Ce-rich support to-
ward CuOx monomers and/or amorphous clusters [32]. While it ap-
pears that CuO dimer is present in all the Ce-containing samples,
its absolute content changes as a function of Zr content. This is in
200
400
600
800
1000
Temperature (°C)
Fig. 5. O2-TPD of 4% Cu/CexZr1ꢀxO2 as well as similar-prepared pristine CeO2 and
ZrO2.
Table 1
Amount of O2 desorbed from Cu/CexZr1ꢀxO2 samples during O2-TPD experiments.
fact observed with the shifting in
a reduction peak toward higher
Sample
Amount of O2 desorbed (
-O2 b1-O2
l )
mol gꢀ1
temperature with increasing Zr content.
a
b2-O2
Total
The b reduction peaks, which are associated with the size of
CuO (smaller size yielding lower reduction temperatures) [20],
did not change significantly among Ce-containing samples with
peak temperatures of b1-H2 and b2-H2 at ꢁ153 and 177 °C, respec-
tively. This corroborates the XPS analysis that the surface Cu spe-
cies were of fine and highly dispersed nature. On the contrary,
Cu/CeO2
1.5 (521 °C)
0.5 (527 °C)
0.8 (527 °C)
1.4 (548 °C)
7.4 (568 °C)
102.3 (776 °C)
108.1 (780 °C)
91.5 (784 °C)
81.1 (782 °C)
74.9 (779 °C)
23.4 (848 °C)
15.4 (872 °C)
2.4 (851 °C)
3.7 (857 °C)
127.2
124.0
94.7
Cu/Ce0.75Zr0.25O2
Cu/Ce0.5Zr0.5O2
Cu/Ce0.25Zr0.75O2
Cu/ZrO2
86.2
82.3

214
R. Zhang et al. / Journal of Catalysis 272 (2010) 210–219
of the Ce4+ to Ce3+. The highest reducibility was observed for sam-
18
12
6
a
ple Cu/Ce0.75Zr0.25O2 (sum of a- and b-H2 of 810 l
mol gꢀ1) where it
is well established that the substitution of Zr4+ in Ce4+ increases its
oxygen mobility [29], which in this case allows for the easy reduc-
tion of the latter.
The bulk capacity of lattice oxygen storage in Cu/CeO2 (through
Ce4+/Ce3+ redox) is further evident at elevated temperature around
810 °C (Fig. 6b). The content decreases gradually with the increase
in Zr content, as consistent with the trend of b2-O2 observed earlier
through O2-TPD. No appreciable amount of high-temperature
reduction band could be observed for Cu/ZrO2, again consistent
with the absence of b2-O2.
Cu/ZrO2
Cu/Ce0.25Zr0.75O2
Cu/Ce0.5Zr0.5O2
Cu/Ce0.75Zr0.25O2
Cu/CeO2
0
100
150
200
250
3.4. Catalytic reduction of NO by CO over Cu/CexZr1ꢀxO2
Temperature (oC)
The temperature dependence of catalytic reduction of NO by CO
over as-prepared catalysts is shown in Fig. 7a and b. In the absence
of Cu, the deNOx activity of pristine CeO2 was low with ꢁ20% NO
conversion (Fig. 7a) to N2 (Fig. 7b) at 300 °C, while similar values
were only observed with pristine ZrO2 at 500 °C. No intermediate
N2O was formed (Fig. 7c), while equimolar CO and NO conversions
were measured over these pristine supports (Fig. 7d), in agreement
with the overall stoichiometry: 2CO + 2NO ? N2 + 2CO.
An enhancement in NO conversion was observed over CeO2
with the addition of 4% Cu, resulting in an initiation at 200 °C
and reaching a value of 96% at 400 °C. Here, the initial NO adsorp-
tion and decomposition over highly dispersed Cu [24] was deemed
as an important step for NO + CO reaction. Improvement in NO
conversion is further evident with the addition of Zr i.e. Cu/
CexZr1ꢀxO2 (x < 1). Compared to Cu/CeO2, doping of the catalysts
with Zr presents some physicochemical advantages: strong inter-
actions of Cu and Zr leading to high Cu+/Cu2+ ratio, where NO
chemisorbs preferentially on Cu+ and enhanced surface oxygen
mobility. More specifically, they are beneficial for the intermediate
reduction of NO to N2O (Fig. 7c), where the latter was generated in
an expected decreasing order of Cu/ZrO2 > Cu/Ce0.25Zr0.75O2 > Cu/
Ce0.5Zr0.5O2 > Cu/Ce0.75Zr0.25O2.
1.2
0.8
0.4
0.0
b
Cu/CeO2
Cu/Ce0.75Zr0.25
O
2
Cu/Ce0.5Zr0.5
O
2
Cu/Ce0.25Zr0.75
O
2
Cu/ZrO2
600
700
800
900
Temperature (oC)
Fig. 6. H2-TPR of 4% Cu/CexZr1ꢀxO2 (x = 0–1.0) at (a) low temperature and (b) that at
high temperature.
the b reduction peaks of Cu/ZrO2 were shifted to higher tempera-
tures (with b2-H2 of 192 °C) as consistent with the relatively lower
Cu dispersion and more importantly demonstrates the strong
interactions between Cu and Zr. Another distinct difference be-
tween the TPR spectra of Cu/ZrO2- and Ce-containing samples is
the lower peak area of b2-H2 compared to that of b1-H2. This was
thought to be related to the poor reducibility to Cu0 in ZrO2 matrix
and could arise due to a few inter-related reasons: (1) strong inter-
actions between Cu+ and Zr4+; (2) poor surface Cu enrichment (i.e.
dispersion); and (3) the absence of Ce4+/Ce3+ redox-assisted reduc-
tion of Cun+ species. Except for Cu/ZrO2, the H2 consumption re-
While Cu+ sites are important for the chemisorption of CO and
NO, removal of atomic O (through surface lattice oxygen sites) pro-
duced from the decomposition of NO and N2O requires equal or
more rapid oxygen consumption through CO oxidation. Here, CO
has a dual role of removing the O while keeping the Cu sites in re-
duced state. The accumulation of N2O intermediate below 250 °C
suggests the limiting role of O removal, and the similar extent of
CO and NO conversions (and the absence of intermediate N2O)
for all catalysts above 250 °C corroborates the mentioned synergis-
tic effect between CO oxidation and NO reduction.
Interestingly, a high 40% N2 yield was observed at the low tem-
perature of 150 °C for sample Cu/Ce0.75Zr0.25O2 (see Fig. 7a and b).
The outstanding low-temperature activity may be attributed to the
combination of the abundance of surface lattice oxygen sites and
most importantly, the high reducibility of these sites at low
temperature (Table 2). As will be shown by in situ DRIFTS, the
lated to the reduction of Cu (a + b) were much higher than the
theoretical limit for all the catalysts for complete Cu reduction
(Table 2), in agreement with previously reported work on similar
systems [18,33]. In the case of Cu/ZrO2 where no Ce is present, only
slightly higher H2 consumption (373
cal value (350
l
mol gꢀ1) than the theoreti-
l
mol gꢀ1 based on XPS-determined Cu+/Cu2+
,
compositions) was measured. The result strongly implies the occur-
rence of H spillover effect over reduced surface Cu, reducing some
Table 2
Quantitative analysis of H2-TPR for Cu/CexZr1ꢀxO2 samples.
Sample
H2 consumption (l )
mol gꢀ1
a
b1
b2
Total
Theoreticala
Cu/CeO2
37 (138 °C)
63 (137 °C)
34 (141 °C)
31 (144 °C)
314 (153 °C)
250 (151 °C)
226 (155 °C)
337 (154 °C)
284 (156 °C)
309 (178 °C)
497 (176 °C)
538 (178 °C)
301 (177 °C)
89 (192 °C)
660
810
798
669
373
441
435
384
488
350
Cu/Ce0.75Zr0.25O2
Cu/Ce0.5Zr0.5O2
Cu/Ce0.25Zr0.75O2
Cu/ZrO2
a
Theoretical H2 consumption calculated based on 4% Cu and the percentage amount of Cu+ and Cu2+: CuO + H2 ? Cu + H2O; Cu2O + H2 ? 2Cu + H2O. Note: XPS-determined
surface Cu+/Cu2+ composition is assumed here as the bulk composition for each catalyst to allow calculation of these theoretical H2 consumptions.

R. Zhang et al. / Journal of Catalysis 272 (2010) 210–219
215
a
100
80
60
40
20
0
b
100
Cu/ZrO
Cu/Ce Zr
Cu/Ce Zr
Cu/Ce Zr
Cu/CeO
CeO
Cu/ZrO
O
Cu/Ce Zr
Cu/Ce Zr
O
O
O
O
80
Cu/Ce Zr
O
Cu/CeO
CeO
ZrO
ZrO
60
40
20
0
100
150
200
250
300
350
400
450
500
100
150
200
250
300
350
400
450
500
Temperature (oC)
Temperature (oC)
30
20
10
0
d
c
100
80
60
40
20
0
Cu/ZrO
Cu/Ce Zr
Cu/Ce Zr
Cu/Ce Zr
Cu/CeO
CeO
O
O
Cu/ZrO
O
Cu/Ce Zr
Cu/Ce Zr
Cu/Ce Zr
Cu/CeO
CeO
O
O
O
ZrO
ZrO
100
150
200
250
300
350
400
450
500
100
150
200
250
300
350
400
450
500
Temperature (oC)
Temperature (oC)
Fig. 7. (a) NO conversion, (b) N2 yield (c) N2O yield and (d) CO conversion in CO + NO reaction over as-prepared catalysts. Conditions: GHSV = 50,000 hꢀ1, 3000 ppm NO,
3000 ppm CO.
reducibility of these sites facilitated the rapid O removal (activa-
tion of rapid CO ? CO2) at 150 °C.
2.5
2.0
1.5
1.0
0.5
0.0
3.5. In situ DRIFT study
Based on the best performing 4% Cu/Ce0.75Zr0.25O2 (on the basis
of N2 yield at low temperature), the formation of surface species
during NO or CO adsorption (and sequential introduction of both)
over this catalyst was probed by means of in situ DRIFTS (Figs.
8–10). As evident in spectrum (a), the adsorption of 2% NO/He over
the catalyst at 300 °C gave rise to several peaks in the range of
1100–1700 cmꢀ1. Adsorption of NOx on surface Cu exists as config-
(d) NO-5min after (c)
(c) CO-20min
(b) CO-20min after (a)
(a) NO-5min
ꢂ
ꢀ
ꢁ
ꢃ
O
O
urations of chelating
Cunþ
NAO
as well as bridging
2400
2200
2000
1800
1600
1400
1200
ꢂ
ꢃ
CunþAO
CunþAO
Wavenumber cm-1
iNAO bidentate nitrate, [33–35] as reflected by the
most intense bands of asymmetric stretch (ANO2) at 1609 cmꢀ1
and shoulders at 1621 and 1582 cmꢀ1. Additional bridging nitrates
could be formed on the catalyst surface as reflected by the moder-
ate bands at 1213 cmꢀ1, while the N-containing compounds other
than nitrate species are marginal [36]. Upon switching to 2% CO/
He, the surface adNOx species were reduced, and hence a distinct
change in the DRIFT structure could be seen from spectrum (b)
Fig. 8. DRIFT spectra of adsorbates produced from the flow of (a) 2% NO/He for
5 min, (b) 2% CO/He for 20 min after (a), (c) 2% CO/He for 20 min, and (d) 2% NO/He
for 5 min after (c) over 4% Cu/Ce0.75Zr0.25O2 at 300 °C.
[34]. The appearance of an intense doublet at 2339 and
2360 cmꢀ1 is due to gaseous CO2. This illustrates the ability of
NO-pretreated sample to oxidize CO, which is in fact more active
than that over the same catalyst sample without NO pretreatment.
As direct comparison, Cu/Ce0.75Zr0.25O2 was exposed to 2% CO/
He at 300 °C without NO pretreatment (spectrum (c), Fig. 8). Here,
the carbonaceous-related bands were similar to those of Fig. 8b. On
closer inspection, the higher doublet peaks intensity of CO2 (2339
and 2360 cmꢀ1) and carbonates bands (1544 cmꢀ1) in spectrum (b)
of Fig. 8 strongly implies that NO pretreatment is beneficial for the
transformation of CO ? CO2 and carbonate. Further exposure of
in Fig.
8 except for the now-attenuated bands at 1100–
1700 cmꢀ1 for nitrate species, which nearly vanished. Instead, the
spectrum now consists predominantly of carbonaceous species,
with a remarkable band at 1544 cmꢀ1 correlated to Cun+
A
m(C@O)
of bidentate carbonate species [37,38], while the peaks maxima
at 1521 cmꢀ1 were ascribed to Cun+ s(COOꢀ) of carboxylates
Am
[39,40]. More specifically, the band at 1435 cmꢀ1 is likely caused
by nitrite species resulting from the partial reduction of nitrates

216
R. Zhang et al. / Journal of Catalysis 272 (2010) 210–219
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
a
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0.00
10 min
9 min
8 min
7 min
6 min
5 min
4 min
3 min
2 min
1 min
220 0 210 0 200 0 19 00 18 00
10min
9min
8min
7min
6min
5min
4min
3min
2min
1min
2200
2000
1800
1600
1400
1200
Wavenumber cm-1
80
60
40
20
0
b
chemisorbed CO (2116 cm-1)
chemisorbed NO (1876 cm-1)
nitrates (1586 + 1609 + 1621 cm-1)
1
2
3
4
5
6
7
8
9
10
Time (min)
Fig. 9. (a) Dynamic changes of DRIFT spectra of adsorbates produced from the flow of 2% NO/He + 2% CO/He at 50 °C over 4% Cu/Ce0.75Zr0.25O2 as a function of time and (b)
time dependence of the integrated areas of the IR peak in the range of 1100–2200 cmꢀ1
.
the CO-saturated sample with 2% NO/He at the same temperature
(spectrum (d), Fig. 8) resulted in the appearance of bands related to
adNOx species, similar to that of spectrum (a).
monocarbonyl species passed through a maximum at 4 min before
its total disappearance at t > 7 min (Fig. 9a and b). The diminution
of monocarbonyl is accompanied by the onset of new IR peaks
1876 and 1890 cmꢀ1 at t > 4 min, attributed to nitrosyl (chemi-
sorbed NO) and physisorbed molecular NO, respectively [41–43].
Hence, it is possible that a competitive adsorption between chem-
isorbed NOx and chemisorbed CO for common active sites is occur-
ring, with the latter (especially nitrates) being dominant due to
higher thermal stability. This is evident by the strong surface cov-
erage of bidentate nitrates (1236 and 1550–1650 cmꢀ1) and mono-
To further understand the NO and CO co-interactions, the DRIFT
spectra of the catalyst surface were monitored in the presence of
the (NO + CO) mixture as a function of time at 50 °C (Fig. 9a), as
well as function of temperature from 50 to 400 °C (Fig. 10a). The
dynamic changes of the surface coverage of various corresponding
species were determined according to a deconvolution of the re-
lated spectra with the results being depicted in Figs. 9 and 10b.
At low temperature of 50 °C, NO and CO interacted with the ac-
tive sites resulting in their simultaneous chemisorptions as evi-
denced by the DRIFT spectra in Fig. 9a. In line with literature
assignment [37], the IR band at 2116 cmꢀ1 for chemisorbed CO
background with the typical roto-vibrational fringes at 2114 and
2175 cmꢀ1 for gaseous CO was initially found. When the adsorp-
tion time (t) is extended, the intensity of band at 2116 cmꢀ1 for
dentate nitrates (1517 cmꢀ1
) at prolonged exposure period
t > 7 min. Similar findings where CO appears to be more weakly ad-
sorbed than NO was reported over anionic vacancy sites of La(Fe,
Co, Mn)1ꢀxCuxO3 perovskite-type oxides [44].
Fig. 10a illustrates the DRIFT study of species formed over 4%
Cu/Ce0.75Zr0.25O2 at various temperatures. Here, it is evident that
gaseous CO2 (2339 and 2356 cmꢀ1 doublet) becomes apparent at

R. Zhang et al. / Journal of Catalysis 272 (2010) 210–219
217
7
6
5
4
3
2
1
0
a
0.08
0.07
0.06
0.05
350 C
0.04
300 C
0.03
250 C
200 C
150 C
100 C
50 C
0.02
0.01
0.00
2300 2200 2100 2000 1900 1800
350 C
300 C
250 C
200 C
150 C
100 C
50 C
2400
2200
2000
1800
1600
1400
1200
Wavenumber cm-1
120
100
80
60
40
20
0
b
nitrates (1586 + 1609 + 1621 cm-1)
chemisorbed NO (1876 cm-1, x 50)
physisorbed N2O (2237 cm-1, x 400)
NCO (2190-2280 cm-1, x 100)
carbonates (1563 cm-1)
gaseous CO2 (2339 + 2356 cm-1, x 10)
50
100
150
200
250
300
350
Temperature oC
Fig. 10. (a) Dynamic changes of DRIFT spectra of adsorbates produced from the flow of 2% NO/He + 2% CO/He for 5 min over 4% Cu/Ce0.75Zr0.25O2 as a function of temperature
and (b) time dependence of the integrated areas of the IR peak in the range of 1100–2500 cmꢀ1
.
150 °C, coinciding with the peculiarly high N2 yield seen earlier in
Fig. 7b. The observation provides evidence for the low-temperature
activation of the CO oxidation (limiting step). The oxidation of CO
is in principle dependent on the removal of O, which is in turn a
function of the catalyst reducibility [15]. The onset activation of
CO oxidation step results in the corresponding decomposition of
surface nitrates (Fig. 10a and b). Careful scrutiny of DRIFT spectra
did not detect any trace amount of monocarbonyl species at all
temperatures, which was expected due to displacement by nitrosyl
as discussed earlier. Hence, the CO2 generation is possibly realized
via an Eley–Rideal mode involving the interaction of chemisorbed
nitrate species with the gas-phase carbon monoxide. On the other
hand, some minor formation of N2O intermediate can be detected
carboxylates at 1532 cmꢀ1 grew at the expense of bridging
(1586, 1609 and 1621 cmꢀ1) and monodentate (1517 cmꢀ1) ni-
trates. The lack of m(NAN) is again in agreement with the absence
of N2O in the outlet gas stream at high temperatures. The growth of
a new nitrites band (1435 cmꢀ1) with increasing temperature indi-
cates the occurrence of partial reduction of nitrates to lower the N-
valence. Similarly, the bridging nitrate band red-shifted monoto-
nously from 1236 to 1194 cmꢀ1 with the increasing temperature
from 50 to 300 °C can be attributed to the formation of bridging
nitrito species [36,41].
As shown in Fig. 10b, a maximum amount for chemisorbed NO
was achieved at 200 °C. Thereafter, the nitrosyl species vanished as
the temperature reaches 300 °C. Interestingly, a relatively broad
band at 2190–2280 cmꢀ1 was discernable at temperature above
300 °C and is likely due to isocyanate (NCO) species [46,47]. The
band at 2260–2270 cmꢀ1 was previously assigned to isocyanate
complex bound to surface metal atoms [48], while the vibration
of Cu+ANCOꢀ that lies ca. 2200 cmꢀ1 belongs to that of Cu+ANCO
from the m m(NAO) stretching was im-
(NAN) at 2237 cmꢀ1 and
mersed at 1276 cmꢀ1 [41,45] in agreement with the outlet gas
analysis during catalytic testing (Fig. 7c).
At 250 °C, a sudden rapid increase in the bidentate carbonate
bands 1563 and 1276 cmꢀ1 [36], together with that of s(COOꢀ)
m

218
R. Zhang et al. / Journal of Catalysis 272 (2010) 210–219
Oxidation of carbonaceous species:
vibration [45]. These isocyanate species are formed following NO
dissociation on the metallic surface and subsequent spillover onto
the support oxides where they accumulate, whereas a blue shift in
the IR band to higher frequency could be seen accompanying with
this NCO spillover process [49,50]. Here, it is likely that accumula-
tion of isocyanate arises from the partial reduction of nitrate/ni-
trite/nitrosyl species at the ceria–zirconia sites, which may be
less capable of catalyzing full reduction to N2 in the absence of vic-
inal Cu. It is envisaged that the high-temperature activation of CO
oxidation at >300 °C provided more reduced ceria–zirconia sites
for the partial reduction of nitrate, resulting in the intensification
of NCO signal at 2190–2280 cmꢀ1 (Fig. 10a and b).
Ce4þAOꢀ þ CO ! Ce4þACOOꢀ ! CO2 þ Ce3þ
Ce4þAOꢀ þ CO2 $ Ce3þ þ carbonate
ð9Þ
ð10Þ
Further conversion of N2O to N2 requires the activation of CO
conversion (to CO2) in a synergistic manner (Eqs. (5) and (10)).
At >150 °C, the latter reaction becomes apparent which is in turn
dependent on the abundance of reactive lattice oxygen surface
sites and their reducibility (Eq. (10)). Between 150 and 250 °C,
the rate of N2O formation decreases gradually as a result of more
rapid CO oxidation.
At the higher temperature region (>250 °C), however, the reac-
tion involves NCO intermediate orignating from the interaction of
chemisorbed NO and gaseous CO (Eq. (6)), followed by the succes-
sive reaction with NO and/or oxygen to yield the N2 and CO2 prod-
ucts (Eqs. (7) and (8)). The high reactivity of NCO radical with NO
or O2 was once established by means of kinetic studies [57].
Adsorption of CO2 and the formation of surface carbonates (Eq.
(10)), which was relatively minor at low temperature, becomes
the dominant at above 250 °C. The possibility also exists for the
formation of Oꢀ2 ACe4+ superoxide species [58], which could inter-
transform via a sequence of Oꢀ2 $ O22ꢀ ! 2Oꢀ at elevated temper-
ature [59] and further contributing to the high-temperature redox
on the catalyst surface.
3.6. Mechanistic of CO + NO reaction over Cu/CexZr1ꢀxO2
In general, the proposed deNOx mechanisms reported in the
literature can be categorized as redox mechanism as well as
the organonitrogen mechanism. The former consists of an initial
dissociation–chemisorption NO yielding N2O intermediate and
an oxidized surface, while the role of CO is to further reduce
N2O into the final product of N2 as well as to restore the initial
catalytic surface [44,51]. In contrast, the latter organonitrogen
mechanism involves the generation of isocyanate (NCO) com-
pounds as intermediates [52,53]. During an actual catalytic
reduction process of NO, more than one reaction mechanism
can occur simultaneously, and the relative contribution of each
mechanism essentially depends on the catalyst nature, reductant
type and reaction conditions (temperature, concentration, etc.)
[54–56].
4. Conclusions
Based on the present deNOx activity and DRIFT results, it can be
reasonably deduced that there are a few key aspects of the func-
tional characteristics of Cu/CexZr1ꢀxO2, which are in turn instru-
mental to assess both redox and/or organonitrogen mechanisms
for NO reduction by CO. At low temperatures (50–250 °C), the
presence of highly dispersed Cu+ promotes the direct chemisorp-
tion of NO as nitrosyl species (Eq. (1)). At the same time, some
oxidation of chemisorbed NO could also take place with the assis-
tance of Ce4+AOꢀ to form nitrites (Eq. (2)) and nitrates (Eq. (3)).
Further reduction of adNOx to N2O involves the removal of O
which can be conveniently abstracted by adjacent reduced lattice
i.e. Ce3+ and the associated oxygen vacancies (Eq. (4)). These
reductant sites could either be formed from the reduction of
Ce4+ by CO (Eqs. (9) and (10)) and/or by formation of nitrites
and nitrates (Eqs. (2) and (3)). In fact, N2O could also form by
reducing the chemisorbed NO on reduced ceria–zirconia surface,
but as discussed above, its further reduction is limited in the ab-
sence of vicinal Cu. Here, the intermolecular reaction between
two neighboring adsorbed nitrosyl species is beneficial for the
NAN bond formation to yield the related diazo compounds (Eq.
(4)).
In this work, the catalytic deNOx activity of a series of flame syn-
thesized 4% Cu/CexZr1ꢀxO2 (x = 0, 0.25, 0.50, 0.75, 1.0) nanocom-
posites was studied. The as-prepared catalysts consisted of highly
dispersed Cu, with the addition of Zr4+ being beneficial for the sta-
bilization of Cu+ species, presumably through the strong interac-
tions between the two components. The presence of copper and
doping with Zr4+ (in the form of solid solution) strongly enhances
the reactive lattice oxygen content as well as site reducibility, with
spillover effects observed during H2-TPR for all Ce-containing cat-
alysts. The highest surface reducibility was measured for sample
Cu/Ce0.75Zr0.25O2.
Based on the results of DRIFT and activity tests, a dissociation
mechanism was proposed for the CO + NO reaction at T < 250 °C
over Cu/CexZr1ꢀxO2 firstly involving the reductive chemisorption
of NO on Cu+ sites. Here, the coupling nitrosyl species was further
dissociated to intermediate N2O, by reductive Ce3+ sites and its
associated oxygen vacancies. This is the dominant reaction step
at <150 °C. Note that the proposed mechanism for this unusually
low-temperature process involves Cu dimmers (Eq. (4)) recently
discussed in reference [32]. For the most surface reducible sample
of Cu/Ce0.75Zr0.25O2, the activation of CO oxidation over the catalyst
at T > 150 °C promoted the rapid formation of N2, with almost no
intermediate N2O detected beyond 250 °C. Essentially, the deNOx
activity was comparable for all Zr-containing catalysts i.e. Cu/
CexZr1ꢀxO2 (x < 1.0) likely following an organonitrogen mechanism
at temperatures above 250 °C.
Dissociation and reduction of chemisorbed NOx:
At low-temperature region (50–250 °C)
Cuþ þ NO ! Cu2þNOꢀ
ð1Þ
ð2Þ
ð3Þ
Cu2þNOꢀ þ Ce4þAOꢀ $ Cu2þANOꢀ2 ðnitriteÞ þ Ce3þ
CuþANO2 þ Ce4þAOꢀ $ Cu2þANOꢀ3 ðnitrateÞ þ Ce3þ
2Cu2þNOꢀ þ Ce3þ ! Cu2þ ꢃ ꢃ ꢃ NOꢀ ꢃ ꢃ ꢃ NOꢀ ꢃ ꢃ ꢃ Cu2þ þ Ce3þ
Cu2þN2O2ꢀCu2þ þ Ce4þAOꢀ ! 2Cuþ þ N2O þ Ce4þAOꢀ
Cu2þN2O2ꢀCu2þ þ Ce3þ ! 2Cuþ þ N2 þ Ce4þAOꢀ
At high-temperature region (>250 °C)
!
ð4Þ
ð5Þ
Acknowledgments
The financial supports of NSERC through its industrial chair pro-
gram (R.Z. and S.K.), NSFC (R.Z. and B.C.) with Grant Nos. 20821004,
20877007 and 20977004 and (R.Z.) Program for New Century
Excellent Talents in University are greatly acknowledged. The work
was partly supported by the ARC through its Centre of Excellence
program (W.Y.T. and R.A.).
CuþANO þ CO þ Ce3þ ! CuþANCO þ Ce4þOꢀ
CuþANCO þ CuþANO ! 2Cuþ þ N2 þ CO2
CuþANCO þ Ce4þOꢀ ! Cuþ þ Ce3þ þ 1=2N2 þ CO2
ð6Þ
ð7Þ
ð8Þ

R. Zhang et al. / Journal of Catalysis 272 (2010) 210–219
[26] Y. Wan, J. Ma, Z. Wang, W. Zhou, S. Kaliaguine, J. Catal. 227 (2004) 242.
219
References
[27] S. Kaliaguine, A. van Neste, V. Szabo, T.E. Gallot, M. Bassir, R. Muzychuk, Appl.
Catal. A 209 (2001) 345.
[28] R. Zhang, A. Villanueva, H. Alamdari, S. Kaliaguine, Appl. Catal. B 64 (2006) 220.
[29] G. Balducci, J. Kašpar, P. Fornasiero, M. Graziani, M. Saiful Islam, J. Phys. Chem.
B 102 (1998) 557.
[30] W. Liu, M. Flytzani-Stephanopoulos, Chem. Eng. J. 64 (1996) 283.
[31] L. Kundakovic, M. Flytzani-Stephanopoulos, Appl. Catal. A 171 (1998) 13.
[32] R. Kydd, Ph.D. Thesis, The University of New South Wales, Sydney, 2009.
[33] Y. Chi, S.S.C. Chuang, J. Catal. 190 (2000) 75.
[34] B. Wichterlova, Z. Sobalik, A. Vondrova, Catal. Today 29 (1996) 149.
[35] Y.H. Yeom, M. Li, W.M.H. Sachtler, E. Weitz, J. Catal. 238 (2006) 100.
[36] T. Venkov, K. Hadjiivanov, A. Milushev, D. Klissurski, Langmuir 19 (2003) 3323.
[37] L. Liu, Y. Chen, L. Dong, J. Zhu, H. Wan, B. Liu, B. Zhao, H. Zhu, K. Sun, L. Dong, Y.
Chen, Appl. Catal. B 90 (2009) 105.
[38] A.B. Hungria, N.D. Browning, R.P. Erni, M. Fernandez-Garcia, J.C. Conesa, J.A.
Perez-Omil, A. Martinez-Arias, J. Catal. 235 (2005) 251.
[39] A. Davydov, IR spectroscopy applied to surface chemistry of oxides, Nauka,
Novosibirsk, 1984.
[40] Y. Wan, J. Ma, Z. Wang, W. Zhou, S. Kaliaguine, Appl. Catal. B 59 (2005) 235.
[41] K. Hadjiivanov, Catal. Rev. Sci. Eng. 42 (2000) 71.
[42] J. Szanyi, M. Paffett, J. Catal. 164 (1996) 232.
[43] A. Aylor, S. Larsen, J. Reimer, A.T. Bell, J. Catal. 157 (1995) 592.
[44] R. Zhang, H. Alamdari, S. Kaliaguine, J. Catal. 242 (2006) 241.
[45] J.W. Labdon, A.T. Bell, J. Catal. 31 (1973) 96.
[46] N. Bion, J. Saussey, M. Haneda, M. Daturi, J. Catal. 217 (2003) 47.
[47] O. Gorce, F. Baudin, C. Thomas, P. D. Costa, G. Djéga-Mariadassou, 54 (2004) 69.
[48] M.L. Unland, J. Phys. Chem. 77 (1973) 1952.
[49] K. Almusaiteer, S.C. Chung, J. Catal. 184 (1999) 189.
[50] N. Macleod, R.M. Lambert, Chem. Commun. (2003) 1300.
[51] R. Burch, P.J. Millington, A.P. Walker, Appl. Catal. B 4 (1994) 65.
[52] F. Solymosi, T. Bánsági, J. Catal. 202 (2001) 205.
[1] G. Centi, S. Perathoner, Appl. Catal. A 132 (1995) 179.
[2] P. Carniti, A. Gervasini, V.H. Modica, N. Ravasio, Appl. Catal. B 28 (2000) 175.
[3] S. Bennici, A. Gervasini, N. Ravasio, F. Zaccheria, J. Phys. Chem. B 107 (2003)
5168.
[4] M. Fernandez-Garcia, A. Martinez-Arias, J.C. Hanson, J.A. Rodriguez, Chem. Rev.
104 (2004) 4063.
[5] M. Fernandez-Garcia, E.G. Rebollo, A.G. Ruiz, J.C. Conesa, J. Soria, J. Catal. 172
(1997) 146.
[6] W. Shan, W. Shen, C. Li, Chem. Mater. 15 (2003) 4761.
[7] R. Di Monte, J. Kašpar, J. Mater. Chem. 15 (2005) 633.
[8] R.K. Usman, G.W. Graham, W.L.H. Watkins, R.W. McCabe, Catal. Lett. 30 (1995)
53.
[9] A. Bensalem, F.B. Verduraz, M. Delamar, G. Bugli, Appl. Catal. A 121 (1995) 81.
[10] M. Lou, J. Chen, J. Lu, Z. Feng, C. Li, Chem. Mater. 13 (2001) 1491.
[11] A.I. Kozlov, D.H. Kim, A. Yezerets, P. Andersen, H.H. Kung, M.C. Kung, J. Catal.
209 (2002) 417.
[12] W.J. Stark, M. Maciejewski, L. Mädler, S.E. Pratsinis, A. Baiker, J. Catal. 220
(2003) 35.
[13] W.Y. Teoh, R. Amal, L. Mädler, S.E. Pratsinis, Catal. Today 120 (2007) 203.
[14] W.Y. Teoh, R. Setiawan, L. Mädler, J.-D. Grunwaldt, R. Amal, S.E. Pratsinis,
Chem. Mater. 20 (2008) 4069.
[15] R. Kydd, W.Y. Teoh, K. Wong, Y. Wang, Q. Zeng, J. Scott, A. Yu, J. Zou, R. Amal,
Adv. Funct. Mater. 19 (2009) 369.
[16] L. Mädler, W.J. Stark, S.E. Pratsinis, J. Mater. Res. 17 (2002) 1356.
[17] W.Y. Teoh, L. Mädler, D. Beydoun, S.E. Pratsinis, R. Amal, Chem. Eng. Sci. 60
(2005) 5852.
[18] K.A. Pokrovski, A.T. Bell, J. Catal. 241 (2006) 276.
[19] X. Wang, J.A. Rodriguez, J.C. Hanson, D. Gamarra, A. Martinez-Arias, M.
Fernandez-Garcia, J. Phys. Chem. B 110 (2006) 428.
[20] G. Avrouropoulos, T. Ioannides, Appl. Catal. A 244 (2003) 155.
[21] H. Toraya, M. Yashmura, S. Somiya, J. Am. Ceram. Soc. 67 (1984) C119.
[22] J.F. Moulder, W.F. Stickle, P.E. Sobol, K.D. Bomben, Handbook of X-ray
Photoelectron Spectroscopy, in: J. Chastain (Ed.), Perkin-Elmer Corporation,
Eden Prairie, MN, 1992.
[53] W.C. Hecker, A.T. Bell, J. Catal. 85 (1984) 389.
[54] R.H.H. Smits, Y. Iwasawa, Appl. Catal. B 6 (1995) L201.
[55] V.I. Pârvulescu, P. Grange, B. Delmon, Catal. Today 46 (1998) 233.
[56] M.C. Kung, H.H. Kung, Top. Catal. 28 (2004) 105.
[57] B. Atakan, J. Wolfrum, Chem. Phys. Lett. 178 (1991) 157.
[58] A. Martinez-Arias, M. Fernandez-Garcia, C. Belver, J.C. Conesa, J. Soria, Catal.
Lett. 65 (2000) 197.
[23] A. Aboukais, A. Bennani, C. Lamonier-Dulongpont, E. Abi-Aad, G. Wrobel,
Colloids Surf. A 115 (1996) 171.
[24] G.P. Ansell, A.F. Diwell, S.E. Golunski, J.W. Hayes, R.R. Rajaram, T.J. Truex, A.P.
Walker, Appl. Catal. B 2 (1993) 81.
[59] R. Zhang, A. Villanueva, H. Alamdari, S. Kaliaguine, J. Mol. Catal. A 258 (2006)
22.
[25] T. Curtin, P. Grange, B. Delmon, Catal. Today 35 (1997) 121.