Characterization of copper oxide supported on ceria-modified anatase

Article abstract of DOI:10.1016/j.molcata.2004.04.032

The dispersion and reduction behavior of copper oxide in CuO/CeO₂/TiO₂ samples (copper oxide supported on ceria-modified anatase) was studied. Ceria could be highly dispersed on the surface of TiO₂ support, and corresponding to a dispersion capacity of ～ 6.98 Ce⁴⁺ ions/n sq m TiO₂. The dispersion of copper oxide on ceria-modified TiO₂ was below the loading amount of ceria. The reduction behavior of the copper oxide species correlated to the ceria loading. The activities of the catalysts, CuO/TiO₂, CuO/CeO₂/TiO₂, and CuO/CeO₂, in the NO + CO reaction at different temperature was studied to monitor the relationship between the catalytic activity vs. the composition of the catalysts. NO conversion increased in the order 4Cu-6Ce-Ti < CuO/TiO₂ < CuO/CeO₂ ≈ 4Cu-40Ce-Ti. N₂ selectivity and the NO conversion increased with the increasing temperature, except the decreasing NO conversion of CuO/TiO₂ sample at the reaction temperature from 300°-400°C.

Full text of DOI:10.1016/j.molcata.2004.04.032

Journal of Molecular Catalysis A: Chemical 219 (2004) 155–164
Characterization of copper oxide supported on ceria-modified anatase
a
a
a
b
a
Haiyang Zhu , Mingmin Shen , Yan Kong , Jianming Hong , Yuhai Hu ,
a
a,∗
a
c
c
Tiandong Liu , Lin Dong , Yi Chen , Can Jian , Zhong Liu
a
Key Laboratory for Mesoscopic Chemistry, Department of Chemistry, College of Chemistry and Chemical Engineering,
Nanjing University, Nanjing 210093, PR China
b
Center of Modern Analysis, Nanjing University, Nanjing 210093, PR China
c
Department of Chemistry and Environment, Tibet University, Lhasa 850000, China
Received 30 December 2003; received in revised form 30 April 2004; accepted 30 April 2004
Available online 7 June 2004
Abstract
XRD, LRS, XPS, TPR and ESR have been used to investigate the dispersion and reduction behavior of copper oxide in CuO/CeO
2 2
/TiO
samples (copper oxide supported on ceria-modified anatase). The results indicate that: (1) ceria can be highly dispersed on the surface of
+
2
TiO
ceria-modified TiO
of copper oxide decreases with the ceria loading. When the loading of ceria is greater than 6.98 Ce ions/nm TiO
oxide increases with ceria loading; (3) the reduction behavior of the copper oxide species are also shown to be correlated to the ceria loading.
For CuO/CeO /TiO samples with the low ceria loading of 3 wt.%, the dispersed copper oxide species could be simply regarded as two species
Cu–I and Cu–II) according to the differences in the reduction behavior. For the high ceria loading CuO/CeO /TiO samples, e.g. ceria loading
amount of 40 wt.%, the TPR profiles indicate that the reduction behavior of copper oxide is similar to that in CuO/CeO system, and the
sample should be approximately regarded as the composition of two systems, CuO/CeO and CeO /TiO . In addition, the activities of the
catalysts, CuO/TiO , CuO/CeO /TiO and CuO/CeO
, in the NO + CO reaction at different temperatures have been studied to monitor the
relationship between the catalytic activity versus the composition of the catalysts. For the samples with low copper oxide loading, the catalytic
activity varies as in this order: 4Cu–6Ce–Ti < CuO/TiO < CuO/CeO
≈ 4Cu–40Ce–Ti, which suggests that the dispersed ceria species
hinder the activity of surface copper species in 4Cu–6Ce–Ti sample. While for the samples with high ceria loading, the activities of copper
species are greatly enhanced due to the formation of crystalline CeO . Furthermore, for the CuO/CeO /TiO samples, especially for that with
2
support, and corresponding to a dispersion capacity (DC) of about 6.98 Ce4 ions/nm TiO
2
; (2) the dispersion of copper oxide on
4
+
2
2
is below the loading amount of ceria. For the samples with ceria loading prior to 6.98 Ce ions/nm TiO
2
, the dispersion
, the dispersion of copper
4
+
2
2
2
2
(
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
lower copper oxide and ceria loadings, the dispersion states of the copper and ceria are also discussed by the consideration of incorporation
model proposed previously [Catal. Lett. 12 (1992) 51].
©
2004 Elsevier B.V. All rights reserved.
Keywords: Anatase; Copper oxide species; Ceria; Dispersion capacity (DC); NO + CO
1
. Introduction
strongly depending on the experimental conditions such as
the loading amount of the oxide and the calcination tem-
perature. Accordingly, a large number of studies have been
devoted to exploring the interaction between the support
and the dispersed oxide species, and various explanations or
models concerning the nature of the interaction have been
proposed [1–7].
Titanium oxide is one of the major supports widely
applied in heterogeneous catalysis [8], and it has been
found that the catalytic activity of HDS catalyst can be
improved by choosing titanium oxide instead of ␥-Al2O3
as the support [9,10]. In recent years, TiO2, as a support,
Supported metal oxides are extensively used as heteroge-
neous catalysts in numerous chemical processes, e.g. ranging
from hydrodesulfurization, cracking, polymerization, and
partial oxidation of hydrocarbons to the selective reduc-
tion of nitrogen oxides. It is well established that the sup-
ported metal oxide species might have a variety of structures
∗
Corresponding author. Tel.: +86-25-83594945;
fax: +86-25-83317761.
E-mail address: chem718@nju.edu.cn (L. Dong).
1
381-1169/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2004.04.032

1
56
H. Zhu et al. / Journal of Molecular Catalysis A: Chemical 219 (2004) 155–164
has been widely used in heterogeneous catalysts, especially
in photo-catalysis. It is well known that transition metal
oxides supported on TiO2 have been employed in various
2. Experimental
TiO2 support was prepared via hydrolysis of titanium
alkoxides, the product was washed, dried and then calcined
in flowing air at 500 C for 5 h. The anatase crystalline form
of the product was identified by XRD [27], and the BET
surface area is 83 m g
CeO2/TiO2 (ceria-modified TiO2) was prepared by im-
pregnating TiO2 with an aqueous solution of cerious nitrate
followed by drying at 100 C overnight and then calcined in
flowing air at 500 C for 7 h.
CuO/CeO2/TiO2 samples were prepared by impregnating
CeO2/TiO2 with an aqueous solution of cupric nitrate fol-
important reactions, for examples, V2O /TiO2 in the selec-
5
◦
tive catalytic reduction of NOx with NH3 and the oxidation
of sulfur dioxide to sulfur trioxide; MoO3/TiO2 in the se-
lective photo-oxidation of alcohol; and CoO–MoO3/TiO2
in hydrodesulfurization of hydrocarbon oils, etc. [8,11–14].
Ceria has proved to be a very important and potential
component in the three-way catalysts (TWCs), and been
widely investigated for the past few years. However, it
is an arduous task to define the role of the ceria in the
catalysis since multiple effects have been attributed to
this promoter [15], which include: promoting the noble
metal dispersion; increasing the thermal stability of the
2
−1
.
◦
◦
◦
lowed by drying at 100 C overnight and then calcined in
flowing air at 500 C for 7 h. For the sake of simplicity,
◦
␥
-Al2O3 support; promoting the water gas shift (WGS)
CuO/CeO2/TiO2 samples were noted as xCu–yCe–Ti, e.g.
4Cu–3Ce–Ti corresponds to the sample with copper oxide
and ceria loading amount of 4 and 3 wt.% respectively. The
results of BET surface area of ceria-modified TiO2 suggest
that the change of support surface area could be neglected
in this system, which is consistent to the results reported in
the literature [7].
X-ray diffraction (XRD) patterns were obtained with a
Shimadzu XD-3A diffractometer employing Ni-filtered Cu
K␣ radiation (0.15418 nm). The X-ray tube was operated at
35 kV and 15 mA. Quantitative XRD methods used in this
work are described elsewhere [28].
and steam reforming reactions; favoring catalytic activity at
the interfacial metal-support sites, promoting CO removal
through oxidation employing lattice oxygen; storing and
releasing oxygen under, respectively, lean and rich condi-
tions.
The catalysts containing transition metals, especially cop-
per, show a potential application for the treatment of exhaust
gas from automobiles [16–19], and special attention has also
been paid to this system as a substitute for noble metal con-
taining catalysts recently [20]. Generally, three forms of the
copper containing catalysts have been used for the investi-
gations, e.g. unsupported copper oxide catalysts [21], sup-
ported copper oxide catalysts and supported metal copper
catalysts [22,23]. As reported by Andersson and co-workers,
the dispersed copper oxide on TiO2 was active for the com-
plete oxidation of CO and toluene. However, the presence of
copper oxide species accelerates the sintering of TiO2 sup-
port, which prohibits the application of this catalyst [24].
Further study indicated that, doping a certain amount of ce-
ria in CuO/TiO2 greatly promoted the activity of CO oxi-
dation and enhanced the stabilization of this catalyst [25].
As documented in the literature, for the CuO/TiO2 sorbent,
both uptake rates and capacity of NOx significantly increase
as ceria species is doped [26].
Despite the studies on the Cu–Ce–Ti–O catalysts men-
tioned above, the role of the doped species and the nature
of the interactions among the doped species and the active
species, as well as the support, have not been solved.
In present work, the studies have been mainly focused
on attempting to [1] determine the dispersion capacity (DC)
and surface states of ceria on TiO2, [2] investigate the influ-
ence of the ceria loading amounts to the dispersion of cop-
per oxide in the CuO/CeO2/TiO2 catalysts, [3] explore the
relationship between the catalytic activity for NO + CO re-
action versus composition of the CuO/CeO2/TiO2 catalysts.
The results have shown that the dispersion of CuO and ceria,
the surface states of the supported binary oxide samples, as
well as the catalytic properties in “CO + NO” reaction are
strongly dependent on the loading amount of the individual
oxides.
Fourier transform laser Raman spectra (FT-LRS) were
recorded on a Bruker RFS-100 Fourier transform spectrom-
eter with an InGaAs detector cooled by liquid nitrogen. Ra-
man excitation at 1064 nm was provided by a Nd-YAG laser.
The laser power measured at the powder sample (∼30 mg)
was 100 mW, and spectra were accumulated for 50 scans at
−
1
4 cm resolutions in back-scattering geometry.
X-ray photoelectron spectroscopy (XPS) results were ob-
tained by using a V.G. Escalab MK II spectrometer equipped
with a hemispherical electron analyzer. The system was op-
erated at 13 kV and 20 mA using a magnesium anode (Mg
K␣, E = 1253.6 eV). A binding energy (BE) of 284.5 eV
for the C 1s level was used as an internal reference.
Electron spin resonance (ESR) spectra were recorded on
a Bruker electron spinning resonance spectrometer (EMX
EPR spectrometer) at room temperature. Before the deter-
mination, the catalysts were put into a glass tube, and then
◦
calcined in air at 300 C for 3 h.
Temperature-programmed reduction (TPR) was carried
out in a quartz U-tube reactor, and 30 mg sample was used
for each measurement. Prior to the reduction, the sample
◦
was pretreated in an air stream at 100 C for 1 h and then
cooled to room temperature. After that, H2–Ar mixture (7%
H2 by volume) was switched on and the temperature was
◦
−1
increased linearly at a rate of 10 C min . A thermal con-
ductivity cell detected the consumption of H2 in the reac-
tant stream. Before each measurement, a TPR profile of
CuO (5 mg) was also obtained as a reference for the cal-
culations.

H. Zhu et al. / Journal of Molecular Catalysis A: Chemical 219 (2004) 155–164
157
The activities of the catalysts for NO + CO reaction were
measured under steady state, using a feed stream with a fixed
composition, NO 3.33%, CO 6.67% and He 90% by volume
as diluent. A quartz tube was employed as the reactor and
the requisite quantity of catalysts (100 mg for each test) was
◦
used. The catalysts were pretreated in N2 stream at 100 C
for 1 h and then heated to reaction temperature, after that, the
mixed gases were switched on. The reactions were carried
out at different temperatures with the same space velocity
−
1
−1
of 6000 ml g
h . Two columns and thermal conductivity
detector (TCD) were used for the purpose of analyzing the
production, column A was Porapak Q used for separating
N2O and CO2, and column B was 13X molecular sieve
(30–60 M) used for separating N2, NO and CO.
0
5
10
15
20
25
30
3
. Results and discussion
Ceria loading amount (wt%)
3
.1. Dispersion of ceria on TiO2
Fig. 2. Quantitative XRD result of CeO2/TiO2 samples.
Fig. 1 shows the XRD patterns of a series of CeO2/TiO2
support presents several Raman peaks centered at 144.3,
samples with different ceria loading amounts. For the sam-
ples with ceria loading amounts <16 wt.%, there are no typ-
ical diffraction peaks of crystalline CeO2 in the patterns,
except those of TiO2 supports. However, as the ceria load-
ing amounts are increased from 16 to 30 wt.%, the charac-
teristic peaks corresponding to crystalline CeO2 (centered
−1
395.7, 519.4 and 642.1 cm . For the samples with ce-
ria loading amounts of 8 and 10 wt.%, only Raman peaks
of TiO2 support are observed, while as the ceria loading
amounts reached 16 wt.%, the characteristic Raman peaks
−1
of crystalline CeO2 at 464 cm
Fig. 3(c)–(f).
appeared, as shown in
◦
at 2θ = 28.40 ) are observed, which means that crystalline
Fig. 4 shows the relationship between Raman inten-
CeO2 has formed in these samples besides the dispersed ce-
ria. An XRD quantitative analysis method was employed
to determine the dispersion capacity of ceria on TiO2 sup-
port [27,28], and the results indicated that the DC of ceria
−1
sity ratio and ceria loading, e.g. ICeO (462.5 cm )/ITiO2
(
2
−1
144.3 cm ). It can be seen that the DC of ceria on
TiO2 from Raman results is about 13.80 wt.%, i.e.
4+
2
6
.75 Ce ions/nm TiO2, which is close to that from XRD.
is about 14.64 wt.%, i.e. 7.23 Ce4 ions/nm TiO2, as shown
+
2
As discussed previously [27,29], the (0 0 1) plane of the
TiO2 (anatase) is considered the preferentially exposed
plane, and the dispersion capacity and the surface states of
in Fig. 2.
LRS results of a series of CeO2/TiO2 samples with
various ceria loading amounts are shown in Fig. 3. TiO2
¡ï
-CeO2
¡ï
-CeO2
¡ï
¡ï
h
¡ï
g
f
f
e
d
c
e
d
c
b
a
b
a
2
0
22 24 26 28 30 32 34 36 38 40
400
600
800
o
-1
2
/
Raman shifts (cm )
Fig. 1. XRD patterns: (a–h) for CeO2/TiO2 samples with ceria loading
amounts of 4, 6, 8, 10, 16, 20, 25, and 30 wt.%, respectively.
Fig. 3. Raman results: (a–f) for CeO2/TiO2 samples with ceria loading
amounts of 8, 10, 16, 20, 25 and 30 wt.%, respectively.

1
58
H. Zhu et al. / Journal of Molecular Catalysis A: Chemical 219 (2004) 155–164
¡
ï
-TiO2
-CuO
¡
ô
¡ï
¡ô
¡ï ¡ô
f
e
d
c
b
a
0
5
10
15
20
25
30
35
2
0
30
40
Ceria loading a mount(wt%)
o
2
/
Fig. 4. Quantitative Raman results of CeO2/TiO2 samples.
Fig. 6. XRD patterns: (a–f) for xCu–3Ce–Ti samples with copper oxide
loading amounts of 4, 6, 8, 12, 16 and 20 wt.%, respectively.
the active species could be expected by considering the sur-
face structure of TiO2. For anatase, the vacant site density
port copper oxide. Figs. 6 and 7 show the XRD patterns of
xCu–3Ce–Ti and xCu–40Ce–Ti, i.e. the lower and higher
ceria-containing samples were used as the examples, re-
spectively. In Fig. 6, no crystalline CeO2 formed in the
xCu–3Ce–Ti samples according to the absence of the typ-
ical diffraction peaks representing crystalline CeO2, and
the typical peaks of crystalline CuO appear for the sam-
ples with copper oxide loading amounts beyond 6 wt.%,
which indicates that crystalline CuO formed in these sam-
ples. In Fig. 7, a strong diffraction peak of CeO2 ap-
pears in all of the xCu–40Ce–Ti samples, and the peaks
of crystalline CuO can be clearly seen as copper oxide
loading amount is higher than 6 wt.% and the peak in-
tensities of CuO also increase with the copper loading.
The results suggest that crystalline ceria formed in all the
xCu–40Ce–Ti samples, and the ceria and ceria-modified
TiO2 will be the supports for the dispersion of copper oxide
simultaneously.
2
in (0 0 1) plane is about 6.98 sites/nm , shown in Fig. 5(a).
When ceria is dispersed on the surface of TiO2, Ce4 may
occupy the surface vacant site on TiO2, and the two accom-
panying oxygen anions will stay at the top of the occupied
site as capping oxygen, compensating the extra positive
charge. Consequently, the DC of ceria species is expected
+
about to be 6.98 Ce4 ions/nm TiO2, i.e. 14.27 wt.%, which
is basically in agreement with the XRD and Raman results.
The surface structure of the dispersed ceria species could
be tentatively shown as in Fig. 5(b).
+
2
3
.2. Dispersion of CuO on CeO2/TiO2 support
To investigate the variation of the dispersion of cop-
per oxide with the addition of different ceria loadings in
TiO2 support, the ceria-modified TiO2 supports, 3Ce–TiO2,
6
Ce–TiO2, 20Ce–TiO2 and 40Ce–TiO2, were used to sup-
4
+
Fig. 5. The schematic diagram for the incorporated Ce ions in the surface vacant sites on the (0 0 1) plane of TiO2 (anatase): (a) (0 0 1) plane of TiO2
(
anatase); (b) incorporation of the dispersed Ce4 species on the (0 0 1) plane of TiO2 (anatase) support.
+

H. Zhu et al. / Journal of Molecular Catalysis A: Chemical 219 (2004) 155–164
159
Table 1
¡ñ
-TiO2
Dispersion of copper oxide on ceria-modified TiO2 supports with various
ceria loading amounts
¡
ø
-CeO2
¡
ï
-CuO
¡ñ
¡ø
Samples
Dispersion of copper oxide on supports (wt.%)
¡ø
¡ï
¡ï
f
xCu–3Ce–Ti
xCu–6Ce–Ti
xCu–20Ce–Ti
xCu–40Ce–Ti
6.91
5.54
5.82
7.26
e
d
c
perse on the surface of formed crystalline CeO2. Along this
line, it seems to suggest that, the decrease of the disper-
sion of copper oxide on the surface of ceria-modified TiO2
support should be due to the occupation of the incorporated
b
a
4+
Ce ions into the surface vacancies of TiO2 support, and
the increase of the dispersion of copper oxide might be, at
least partly, attributed to the dispersion of copper oxide on
the surface of the formed CeO2 particles. As reported pre-
viously, copper oxide species can dispersed on the surface
of CeO2 with a DC of 6.98 ions/nm CeO2 [31]. The disper-
sion of copper oxide on ceria-modified TiO2 supports with
various ceria loading amounts is listed in Table 1.
2
0
30
40
o
2
/
Fig. 7. XRD patterns: (a–f) for xCu–40Ce–Ti samples with copper oxide
loading of 4, 6, 8, 12, 16 and 20 wt.%, respectively.
2
Quantitative XRD analysis of the dispersion of copper
oxide on ceria-modified TiO2 is shown in Fig. 8. The results
indicate that the dispersion of copper oxide is relative to the
ceria loading amounts. When the loading amounts of ceria
are ranging from 3 to 6 wt.%, the dispersion of copper oxide
decreases, but as the loading amounts of ceria increase from
To get more understanding about the surface prop-
erties of Cu–Ce–Ti catalysts, XPS was used to charac-
terize the samples: 4Cu–6Ce–Ti, CuO/CeO (in which,
2
2.41 Cu2 ions/nm CeO ) and CuO/TiO2 (in which,
+
2
2
4.94 Cu2 ions/nm TiO ). For the three samples, the copper
+
2
2
oxide should exist in dispersion state, as reported elsewhere
[27,31].
2
0 to 40 wt.%, it increases slightly.
With consideration of the incorporation model proposed
The Cu 2p3/2 binding energies in 4Cu–6Ce–Ti, CuO/TiO2
elsewhere [30], when the ceria loading amount is lower than
the DC of ceria on TiO2, the higher the ceria loading amount
is, the less the available surface vacant sites of TiO2 for cop-
per oxide are. However, when the ceria loadings are highly
beyond the DC, the surface vacant sites on TiO2 support
and CuO/CeO samples are 932.2, 931.9 and 932.2 eV, re-
spectively, shown as in Fig. 9. It has been well established
2
that for copper species the presence and absence of the
2
+
shake-up peak are the characteristic for Cu species and
+
Cu species [32]. For CuO/TiO and 4Cu–6Ce–Ti samples,
2
have been completely occupied by the Ce4 ions in the dis-
+
a shake-up peak can be observed evidently, which indicates
that Cu2 species exist in these two samples. In contrast,
+
persed ceria species, and copper oxide species would dis-
10
932.2
¡ï -- Cu2p3/2 shake-up
c
931.9
b
a
¡
ï
932.2
¡
ï
0
0
5
10 15 20 25 30 35 40 45
930
940
950
960
CeO loading amount/wt%
2
Binding Energy(eV)
Fig. 9. Cu 2p XPS spectra: (a) 4Cu–6Ce–Ti; (b) CuO/TiO2; (c) CuO/CeO2.
Fig. 8. Dispersion of copper oxide vs. ceria loading amounts.

1
60
H. Zhu et al. / Journal of Molecular Catalysis A: Chemical 219 (2004) 155–164
5
28.8
29.0
29.6
e
d
5
5
c
b
a
c
b
524
526
528
530
532
534
536
a
Binding Energy(eV)
200
400
Fig. 10. O 1s XPS spectra: (a) CuO/TiO2; (b) 4Cu–6Ce–Ti; (c) CuO/CeO2.
o
Temp./ C
Fig. 11. TPR profiles of xCu–3Ce–Ti samples: (a–c) with copper oxide
loading amounts of 4, 8 and 16 wt.%, respectively; (d) CuO/CeO2 sample
with a copper oxide loading amount of 4 wt.%; (e) CuO/TiO2 sample
with a copper oxide loading amount of 4 wt.%.
for CuO/CeO2 sample, almost no shake-up peak can be
observed in Fig. 9(c), which suggests that copper species
+
mainly exist as Cu species. It seems possible that the re-
duction of Cu2 to Cu occurs under the procedure of XPS
determination [25], and this should result from the proper-
ties of ceria with the oxygen storage and release capacity
+
+
as in Fig. 11(d) and (e). Generally speaking, the shapes of
TPR profiles varied with both the copper oxide and ceria
loadings, which also indicated the complexity of the re-
duction process in these samples. To further approach the
relationship between the amounts of various copper oxide
species and their reduction behavior, a quantitative analysis
has been tentatively employed in this section. For example,
a quantitative treatment for the sample with 4Cu–3Ce–Ti
has been shown in the dot line of Fig. 11(a), and the same
treatment has been reported elsewhere [33–35]. The cor-
responding quantitative data for the samples are listed in
Table 2.
[15]. In addition, the intensity ratio of the shake-up to Cu
2
p3/2 peak in 4Cu–6Ce–Ti sample is about 0.232, which is
much smaller than that in CuO/TiO2, i.e. 0.409, and this re-
sult seems to further support the suggestion that the reduc-
tion of copper oxide was resulted from the influence of ceria
addition. In addition, the peaks at ca. 952 eV in these sam-
ples attributed to Cu 2p1/2 peak varies in the similar trends
as Cu 2p3/2 peak.
Fig. 10 shows the O 1s XPS results of surface oxygen
species in CuO/TiO2, 4Cu–6Ce–Ti and CuO/CeO2 samples,
and the centers of O 1s binding energy are 529.6, 529.0 and
5
28.8 eV, respectively. For 4Cu–6Ce–Ti sample, doping of
ceria leads to the BE value shift of O 1s between those of
CuO/TiO2 and CuO/CeO2, which means that, in some ex-
tent, the properties of the surface oxygen species should be
the common contribution of O–Cu, O–Ce and O–Ti con-
tained. This result might suggest that, for the metal oxide
catalysts design, the individual properties of O–M from the
components included in the sample should be considered,
especially for the selective oxidations reaction via surface
crystallite oxygen.
c
b
3
.3. Reduction behavior of copper oxide in
CuO/CeO2/TiO2 samples
a
Figs. 11–13 show the TPR profiles of xCu–3Ce–Ti,
xCu–6Ce–Ti and xCu–40Ce–Ti samples respectively. In ad-
100
200
300
400
2
+
2
dition, the profiles of CuO/CeO2 (4.76 Cu ions/nm CeO2,
less than the dispersion capacity of copper oxide in
o
Temp/ C
CuO/CeO2 [31]) and CuO/TiO2 (4.94 Cu2 ions/nm TiO2)
+
2
Fig. 12. TPR profiles of xCu–6Ce–Ti samples: (a–c) with copper oxide
samples have also been employed for comparison, shown
loading amounts of 4, 8 and 16 wt.%, respectively.

H. Zhu et al. / Journal of Molecular Catalysis A: Chemical 219 (2004) 155–164
161
d
c
b
2+
Fig. 14. The schematic diagram for the incorporated Cu ions in the
surface vacant sites on the (0 0 1) plane of ceria-modified TiO2 (anatase).
a
2
00
400
as Cu–III, it should be resulted from the reduction of CuO
clusters in contact with TiO2 surface [27,36].
0
Temp/ C
Fig. 12 shows the TPR profiles of xCu–6Ce–Ti samples.
Two peaks around 160 and 290 C appear in all the three
Fig. 13. TPR profiles: (a) CuO/CeO2 sample with a copper oxide loading
amount of 4 wt.%; xCu–40Ce–Ti samples: (b–d) with copper oxide loading
amounts of 4, 8 and 16 wt.%, respectively.
◦
profiles, and which seems to suggest that, for these sam-
ples, two kinds of copper oxide species are the main ones.
◦
The reduction peaks around 290 C present two character-
As shown in Fig. 11(a), the TPR profile of 4Cu–3Ce–Ti
has a broad peak, which could be regarded as the contri-
istics, e.g. the shift of the peak to the higher temperature
and the peak intensity becoming stronger with the increas-
ing copper loading, which should be caused by the increas-
ing amount of the formed crystalline CuO, signed as Cu–IV.
bution of three individual peaks centered at 157, 177 and
◦
2
17 C. Considering the TPR profile of CuO/TiO2 sam-
◦
ple, peak 177 C should be ascribed to the reduction of
highly dispersed copper oxide species on the surface of TiO2,
signed as Cu–I, i.e. Cu–Ti. Comparing the peak at 157 C
◦
The peaks at about 160 C should be corresponding to the
Cu–II species, i.e. the highly dispersed copper oxide species
closely contacting with the dispersed ceria on TiO2, and the
amount of this species was calculated in Table 2. Noting
◦
of 4Cu–3Ce–Ti and TPR profile of CuO/CeO2 sample (Fig.
1
1(d)), it seems to suggest that this peak would be related to
◦
that, for the reduction temperature around 160 C, the pro-
the reduction of the highly dispersed copper oxide species
closely contacting with the dispersed ceria on TiO2, which
is signed as Cu–II, Ce–Cu–Ti. A possible scheme of the sur-
face copper species in 4Cu–3Ce–Ti sample should be shown
file shape, the peak temperature and the peak intensity did
not vary evidently with copper oxide loadings, which sug-
gest that Cu–II species, not Cu–I species, should be the pre-
dominant copper oxide in these samples. This result seems
easily to be understood because of the relatively more dis-
persed ceria species on the surface of TiO2 in these sam-
ples (compared with xCu–3Ce–Ti samples) and the increase
of the connection possibility of the dispersed copper ox-
ide species with the dispersed ceria species, as shown in
Fig. 14.
◦
as in Fig. 14. As for the reduction peak at 217 C, signed
Table 2
Amounts of copper oxide species in the xCu–yCe–Ti samples calculated
by quantitative TPR
Amount of copper oxide species
−
3
2+
(
10 mmol Cu )
Fig. 13 shows the TPR profiles of the xCu–40Ce–Ti sam-
ples, and the CuO/CeO2 sample (4.76 Cu2 ions/nm CeO2)
was used as the reference. The results indicate that the
profile-shape and the reduction peak temperatures are basi-
cally similar, and which suggested that there are some cop-
per species with the same states existing in all the samples.
Considering the fact that the full surface of TiO2 could be
covered by the dispersed ceria and the formed crystalline
ceria existed in all the xCu–40Ce–Ti samples, it seems rea-
sonable to suggest that the dispersed copper oxide species
on the surface ceria should be the common one in all the
+
2
Cu–I
Cu–II
Cu–III
Cu–IV
4Cu–3Ce–Ti
8Cu–3Ce–Ti
16Cu–3Ce–Ti
4Cu–6Ce–Ti
8Cu–6Ce–Ti
16Cu–6Ce–Ti
4Cu–40Ce–Ti
8Cu–40Ce–Ti
16Cu–40Ce–Ti
4.97
5.16
6.13
0
5.46
10.29
12.56
9.12
9.03
9.21
7.66a
5.79
8.68
17.56
6.71
0
6.34
25.14
0
7.68
29.29
0
0
0
9.65
21.31
11.82
21.80
19.74
1.82a
a
a
5.21
1.73a
8.93
8.73a
2.70
50.45
a
H2 consumption of copper oxide species during stepwise reduction,
2
+
+
samples, e.g. CuO/CeO . And this proposal should be fur-
and first row corresponds to reduction of Cu to Cu , and second row
corresponds to reduction of Cu to Cu , respectively.
2
+
0
ther supported by the comparison of the TPR results of

1
62
H. Zhu et al. / Journal of Molecular Catalysis A: Chemical 219 (2004) 155–164
the 4Cu–40Ce–Ti and CuO/CeO2 samples, i.e. the reduc-
tion temperature peaks in profiles Fig. 13(a) and (b) are
almost the same. In addition, the intensities of the reduc-
tion peaks gradually increase as the increasing copper load-
ing, and this should be resulted from the increasing amount
of the crystalline copper oxide formed in the samples, as
proved by XRD. In fact, for all the three xCu–40Ce–Ti sam-
ples, all vacant sites on the surface of TiO2 have been oc-
cupied by the dispersed ceria species and the extra ceria
would form crystalline CeO2, and consequently, the three
samples can be approximately regarded as the composi-
tion of two systems, CuO/CeO2 and CeO2/TiO2. Further-
more, it seems to suggest that, for CuO/CeO2/TiO2 cat-
alysts, the excess ceria would form the crystalline ceria,
which would result in the dispersion of copper oxide on
its surface and reduce the use of ceria as a promoter as
well as the TiO2 as the support in the ceria-modified cata-
lysts. Noteworthy, this result is different from that for the
CuO/WO3/CeO2 system reported previously [37], in which
copper oxide could continuously disperse on the surface of
tungsten covered ceria particles. As to the hazy shoulder
peaks appeared in the TPR profiles (c) and (d), as well as
the possible reduction peaks covered in the broad profiles,
a detailed description and discussion will be presented else-
where.
g=2.007
c
b
a
2000
2500
3000
3500
4000
4500
H(G)
Fig. 15. ESR results at room temperature: (a) CuO/TiO2; (b) CuO/CeO2;
c) 4Cu–6Ce–Ti.
(
oxidation reactions [38]. To obtain further understanding of
the influence of dispersed ceria on the surface of TiO , ESR
2
results of the surface oxygen species about 4Cu–6Ce–Ti,
2
+
2
CuO/CeO2 (in which, 2.41 Cu ions/nm CeO ) and
2
2
+
2
CuO/TiO (in which, 4.94 Cu ions/nm TiO ) have been
2
2
recorded and shown in Fig. 15. The ESR signal at g = 2.007
can be observed for all the three catalysts, which should
3
.4. “NO + CO” reaction activity of the Cu–Ce–Ti
−
be attributed to the adsorbed O2 species on the surface
catalysts in different compositions and temperatures
of catalysts [39,40]. However, the directions of the signal
are different, i.e. the signals of 4Cu–6Ce–Ti and CuO/TiO2
Table 3 shows the NO conversion and the N2 selectivity
◦
are opposite to that of CuO/CeO , which seems to suggest
of the Cu–Ce–Ti samples at 200 C for NO + CO reaction.
2
−
that the states of surface O2 species should be different in
these samples and the surface O2 species with the nega-
For comparison, reactivity data of CuO/CeO2 and CuO/TiO2
catalysts are also listed. The results show that the NO con-
version increases in the order: 4Cu–6Ce–Ti < CuO/TiO2 <
CuO/CeO2 ≈ 4Cu–40Ce–Ti. For the 4Cu–6Ce–Ti sam-
ple, the catalyst shows poor activity, which is lower than
that in CuO/TiO2 catalyst; while for the 4Cu–40Ce–Ti sam-
ple with high ceria loading, the catalyst shows high activ-
ity, which is close to that of CuO/CeO2. The results sug-
gested that, for the low ceria-modified TiO2, the existence
of dispersed ceria on the TiO2 support lowers the activity
for NO + CO reaction, and which is contrast with the re-
sults shown in Cu–Ce–Al catalyst [35], i.e. for the lower
ceria-modified ␥-Al2O3 catalyst, the activity is higher than
that of CuO/␥-Al2O3.
−
tive direction at g = 2.007 might be the active ones in the
NO + CO reaction.
Fig. 16 presents the relationship between the NO con-
version, the N2 selectivity of 4Cu–6Ce–Ti and CuO/TiO2
samples and reaction temperatures. The results show that
the N2 selectivity and the NO conversion increases with the
increasing temperature, except the decreasing NO conver-
sion of CuO/TiO2 sample at the reaction temperature from
◦
3
00 to 400 C. In addition, for the 4Cu–6Ce–Ti catalyst,
the color of the catalysts has changed from gray to dark, as
◦
well as slightly red after used at 400 C, which implies that
0
the reduction of surface copper species to Cu occurs dur-
ing the reaction. As discussed elsewhere [41], the nature of
the different behavior of the catalysts should be related to
the different reaction mechanism at different reaction con-
ditions.
As reported elsewhere, for metal-oxide catalysts, the state
of the surface oxygen species will play an important role in
To get more understanding about the behaviors of
xCu–yCe–Ti samples in NO + CO reaction, the varia-
tion of NO conversions and/or N2 selectivity at different
reaction time and temperatures have also been explored
Table 3
For NO + CO reaction, NO conversion and the N2 selectivity of the
◦
Cu–Ce–Ti, CuO/TiO2 and CuO/CeO2 samples at 200 C, reaction time 1 h
Sample number
4Cu–40Ce–Ti 4Cu–6Ce–Ti 8Cu–Ti 4Cu–Ce
(the results have not shown in this paper). However, it
NO conversion (%) 100
N2 selectivity (%) 94.1
0
–
40.6
38.9
100
100
seems difficult to explain the results shown in our exper-
iment for the time being, because of the complexity in

H. Zhu et al. / Journal of Molecular Catalysis A: Chemical 219 (2004) 155–164
163
1
00
d Cu-Ti
c Cu-Ce-Ti
b Cu-Ti
a Cu-Ce-Ti
5
0
0
2
00
300
400
500
o
Temp( C)
Fig. 16. Relationship between the NO conversions (ᮀ), the N2 selectivity (᭹) of 4Cu–6Ce–Ti and CuO/TiO2 samples and reaction temperature: (a) the
NO conversion of 4Cu–6Ce–Ti; (b) the NO conversion of Cu–Ti; (c) the N2 selectivity of 4Cu–6Ce–Ti; (d) the N2 selectivity of Cu–Ti.
this system, and a further study is being carried out in our
laboratory.
Acknowledgements
The financial supports of the Science and Technol-
ogy Cooperation Item of Jiang-su Province, China (No.
BG2001038) and the Special of Foundation for the
Doctor-subject of China (No. 20030284002) are gratefully
acknowledged.
4
. Conclusion
1
. From XRD and Raman results, the DC of ceria on the
surface of TiO2 is about 6.98 Ce4 ions/nm TiO2, which
is basically in agreement with the value predicted by the
incorporation model.
+
2
2
. The dispersion of copper oxide and the state of the
dispersed copper oxide species on ceria-modified TiO2
closely depend on the ceria loading amounts. For the
samples with low ceria loading amounts, the disper-
sion of copper oxide decreases with the ceria loading
amounts, while when the ceria loading amounts are be-
References
[1] H.P. Boehm, H. Knozinger, in: J.R. Anderson, M. Boudart (Eds.),
Analysis, Science and Technology, vol. 4, Springer-Verlag, West
Berlin, 1983, p. 39.
[2] Yu.I. Yermakov, B.N. Kuznetsov, V.A. Zakharov, Analysis by Sup-
ported Complexes, Elsevier, Amsterdam, 1981.
yond 6.98 Ce4 ions/nm TiO2, the dispersion of copper
oxide increases with the formation of ceria particles.
. TPR results show that, for the samples with ceria load-
ing amounts lower than its DC, two kinds of dispersed
copper oxide species are probably formed, e.g. Cu–I and
Cu–II, which is due to the connection or not connection
of dispersed copper oxide species with the dispersed ce-
ria species; as the ceria loading amounts are highly be-
yond the DC, CuO/CeO2/TiO2 systems can be regarded
as the composition of CuO/CeO2 and CeO2/TiO2.
+
2
[3] B. Delmon, M. Houalla, in: B. Delmon, P. Grange, P.A. Jacobs, G.
Poncelet (Eds.), Preparation of Catalysts II, Elsevier, Amsterdam,
1979, p. 447.
3
4
[
4] H. Knozinger, E. Taglauer, in: J.J. Spivey, S.K. Agarwal (Eds.),
Catalysis, vol. 10, A Specialist Periodical Report, The Royal Society
of Chemistry, London, 1993, p. 1.
[5] F.E. Massoth, Adv. Catal. 27 (1978) 265.
[6] L. Dong, Y. Hu, F. Xu, D. Lu, B. Xu, Z. Hu, Y. Chen, J. Phys.
Chem. 104 (2000) 78.
[7] L. Dong, Y. Hu, M. Shen, T. Jin, J. Wang, W. Ding, Y. Chen, Chem.
Mater. 13 (2001) 4227.
[
[
8] S. Matsuda, A. Kato, Appl. Catal. 8 (1983) 149.
9] H.B. Jones, R. Smith, US Patent 4 206 038, 1980.
. In NO + CO reactions, the NO conversion increases
as: 4Cu–6Ce–Ti < CuO/TiO2 < CuO/CeO2
≈
[10] Y. Okamoto, A. Maezawa, T. Imamaka, J. Catal. 120 (1989) 29.
[11] M.A. Reiche, E. Ortelli, A. Baiker, Appl. Catal. B 23 (1999) 187.
4
Cu–40Ce–Ti. For 4Cu–6Ce–Ti sample, the existence
[
[
[
12] J.P. Dunn, P.R. Koppula, H.G. Stenger, I.E. Wachs, Appl. Catal. B
9 (1998) 103.
13] Y.C. Liu, G.L. Griffin, S.S. Chan, I.E. Wachs, J. Catal. 94 (1985)
08.
14] K.Y.S. Ng, E. Gulari, J. Catal. 92 (1983) 340.
of dispersed ceria on TiO2 would restrain the catalytic
activity of copper oxide species; and for 4Cu–40Ce–Ti
sample, the catalyst would be approximately regarded
as the composition of two systems, CuO/CeO2 and
CeO2/TiO2, and result in a similar catalytic activity to
CuO/CeO2.
1
1
[15] J. Kašpar, P. Fornasiero, M. Graziani, Catal. Today 50 (1999) 285.
16] R.T. Rewick, H. Wise, J. Catal. 40 (1975) 301.
[

1
64
H. Zhu et al. / Journal of Molecular Catalysis A: Chemical 219 (2004) 155–164
[
17] F. Boccuzzi, E. Guglielminotti, G. Martra, G. Cerrato, J. Catal. 146
[30] Y. Chen, L.F. Zhang, Catal. Lett. 12 (1992) 51.
[31] L. Dong, Y.S. Jing, Y. Chen, Sci. China, Ser. B 40 (1997) 24.
[32] D. Brigs, M.P. Seah (Eds.), Practical Surface Analysis by Auger and
X-ray Photoelectron Spectroscopy, Wiley, New York, 1983.
[33] P. Zimmer, A. Tschope, R. Birringer, J. Catal. 205 (2002) 339.
[34] F. Pinna, T. Fantinel, G. Strukul, A. Benedetti, N. Pernicone, Appl.
Catal. A 149 (1997) 341.
(1994) 449.
[
[
[
18] M.C. Wu, D.W. Goodman, J. Phys. Chem. 98 (1994) 9874.
19] H. Knozinger, Adv. Catal. 25 (1976) 184.
20] P.Y. Lin, M. Skoglundh, L. Lowendahl, J.E. Otterstedt, L. Dahl, K.
Jansson, M. Nygren, Appl. Catal. B 6 (1995) 237.
[21] J.W. London, A.T. Bell, J. Catal. 31 (1973) 96.
[
22] R. Hierl, H.P. Urbach, H. Knozinger, J. Chem. Soc., Faraday Trans.
[35] K. Hadjiivanov, M. Mihaylov, D. Klissurski, P. Stefanov, N.
Abadjieva, E. Vassileva, L. Mintchev, J. Catal. 185 (1999)
314.
[36] P. Larsson, A. Andersson, Appl. Catal. B 24 (2000) 175.
[37] L. Dong, Y.H. Hu, F. Xu, D. Lu, B. Xu, Z. Hu, Y. Chen, J. Phys.
Chem. 104 (2000) 78.
88 (1992) 355.
[23] A.R. Balkenede, H. den Daas, M. Huisman, O.L.J. Gijzeman, J.W.
Geus, Appl. Surf. Sci. 47 (1991) 341.
[24] P. Larsson, A. Andersson, L.R. Wallenberg, B. Svensson, J. Catal.
163 (1996) 279.
[
[
[
25] P. Larsson, A. Andersson, J. Catal. 179 (1998) 72.
26] W. Li, R. Yang, Energy Fuel 11 (1997) 428.
27] B. Xu, L. Dong, Y. Chen, J. Chem. Soc., Faraday Trans. 94 (1998)
[38] P.J. Gellings, H.J.M. Bouwmeester, Catal. Today 58 (2000) 1.
[39] D. Dvoranova, V. Brezova, M. Mazur, M.A. Malati, Appl. Catal. B
37 (2002) 91.
1
905.
[40] J. Soria, A. Martinez-Arias, J.C. Conesa, J. Chem. Soc., Faraday
Trans. 91 (1995) 1669.
[
28] Y.C. Xie, Y.Q. Tang, Adv. Catal. 37 (1990) 1.
[29] B. Xu, L. Dong, Y.N. Fan, Y. Chen, J. Catal. 193 (2000) 88.
[41] G. Centi, S. Perathoner, Appl. Catal. A 132 (1995) 179.