Chinese Journal of Catalysis 34 (2013) 851–864

a v a i l a b l e a t w w w. s c i e n c e d i r e c t . c o m

j o u r n a l h o m e p a g e : w w w . e l s e v i e r. c o m / l o c a t e / c h n j c

Review (Special Issue in Memory of the 80th Birthday of Professor Jingfa Deng)
Interactions among supported copper‐based catalyst components
and their effects on performance: A review
DONG Lin a,b,*, YAO Xiaojiang a, CHEN Yi a
a School of Chemistry and Chemical Engineering, Key Laboratory of Mesoscopic Chemistry of MOE, Nanjing University, Nanjing 210093, Jiangsu, China
b Center of Modern Analysis, Jiangsu Key Laboratory of Vehicle Emissions Control, Nanjing University, Nanjing 210093, Jiangsu, China

A R T I C L E I N F O A B S T R A C T

Article history: Supported copper‐based catalysts are widely used in many industrial processes because of their
Received 22 February 2013 excellent catalytic performance and low cost. Understanding the interactions among the various
Accepted 2 April 2013 components of these catalysts is an important step in exploring the nature of the catalytic reactions
Published 20 May 2013 and provides a valuable scientific basis for the design of novel and efficient catalysts. This review
summarizes a number of our recent results in this field. In this work, copper oxide was loaded on
Keywords: CeO2, CexZr1xO2, and MoO3‐CeO2 supports to produce model catalysts. Investigations involving these
Supported copper‐based catalyst materials demonstrated that the dispersion of copper oxide on the surface of the oxide supports, as
Surface interaction well as some of the properties of these catalysts, can be explained on the basis of the incorporation
Incorporation model model, which we have previously proposed. Based on spectroscopic characterization of these cata‐
Carbon monoxide oxidation lysts and examination of their catalytic performance for the oxidation of CO or the selective catalytic
CO+NO reaction reduction of NO with CO/NH3, the relationships among the compositions, structures, and properties
NH3+NO+O2 reaction of these catalysts are discussed.
© 2013, Dalian Institute of Chemical Physics, Chinese Academy of Sciences.
Published by Elsevier B.V. All rights reserved.

1. Introduction such catalysts so as to broaden their range of applications and

to realize their full potential. For this reason, there is a re‐
Supported copper‐based catalysts offer exceptional catalytic quirement to research the surface compositions, structures,
performance at low cost and so are widely used in several in‐ and properties of supported copper‐based catalysts at the mo‐
dustrial processes. Examples include CuO/ZnO/Al2O3, the main lecular and atomic levels.
catalyst employed in steam reforming of methanol [1], It has long been recognized that the properties of supported
CuO/CeO2, an excellent catalyst for the preferential oxidation of species will always be affected to some extent by the charac‐
CO (the CO‐PROX reaction) [2], and CuO‐NiO/Al2O3, used in the teristics of the support materials [9]. There has been significant
synthesis of N‐methylpiperazine [3]. In addition to existing research into the nature of the interactions between supported
applications, supported copper‐based catalysts show promise species and their supports in the field of heterogeneous cataly‐
for applications such as the oxidation of volatile organic com‐ sis during recent years. Since the introduction of the strong
pounds (VOCs) [4], the water‐gas shift reaction [5], ester hy‐ metal‐support interaction [10] in the late 1970s and the pro‐
drogenolysis [6], NOx removal [7], and low temperature CO posal of the strong oxide‐oxide support interaction [11] in the
oxidation [8]. It is important to improve the performance of early 1980s, several research groups have studied the interac‐

* Corresponding author. Tel: +86‐25‐83592290; Fax: +86‐25‐83317761; E‐mail: [email protected]

This work was supported by the National Natural Science Foundation of China (20873060, 20973091), the National Basic Research Program of China
(2010CB732300), and Jiangsu Science and Technology Support Program (SBE201100389).
DOI: 10.1016/S1872‐2067(12)60592‐0 | http://www.sciencedirect.com/science/journal/18722067 | Chin. J. Catal., Vol. 34, No. 5, May 2013
 DONG Lin et al. / Chinese Journal of Catalysis 34 (2013) 851–864

tions between supported species and supports in supported well as the stability of the coordination); and (5) because of
catalysts to distinguish these interactions from other phenom‐ differing surface vacancy densities on different supports, the
ena, using reaction mechanisms or models. Examples of re‐ shielding effects of the same anions will vary based on the
searchers involved in this work include Bell and Iglesia [12–16] supports, and when the anionic shielding effect increases to a
and Wachs in the United States [17–20], Knözinger in Germany certain degree, a surface close‐packed monolayer is formed.
[21], Coluccia in Italy [22], Reddy in India [23–25], and Xie and Experimental data show that this approximate model can
Tang in China [26–30]. The oxide supports used by the above quantitatively predict the dispersion capacity of different metal
research groups were generally polycrystalline powders simi‐ oxides (such as Li2O, NiO, CuO, and MoO3) on some metal oxide
lar to the supports used in actual catalysis, but in which various supports within the range of experimental error, such that
crystal planes coexist in unknown proportions, thus making these experimental results are consistent with the theoretical
quantitative analysis difficult. Taking advantage of progress in values. This model can also explain why the dispersion capaci‐
characterization techniques, Goodman et al. [31–34] used ties of different metal oxides on the same support are distinct
scanning tunneling microscopy to investigate the dispersion and why the same metal oxide will exhibit different dispersion
behavior of noble metals on the surface of a single crystal oxide capacities on different supports. In addition, it has explained
support. It is evident, however, that further work correlating why low valence metal oxides do not form surface close‐packed
dispersion behavior and structures of surface species with cat‐ monolayer dispersions on oxide supports.
alytic properties is required to fully understand the catalytic The incorporation of additives into a supported catalyst can
mechanism, as well as to improve existing catalysts and design change its catalytic performance significantly. For example, the
new catalysts. addition of WO3 or MoO3 to V2O5/TiO2 will improve its low
Compared with the interactions between metals and met‐ temperature activity in the selective catalytic reduction of NOx
al‐oxide supports, the interactions between metal‐oxides and by NH3 (NH3‐SCR reaction) remarkably. Although the interac‐
metal‐oxide supports are expected to exhibit greater variation tions among the components in ternary metal oxide systems
because of increased structural complexity based on the for‐ are more complex than in binary metal oxide systems, they are
mation of interacting surface species, solid solutions, and new closer to the interactions of a catalyst in actual application, and
compounds. Although it is necessary to elucidate the surface thus it is particularly important to further study this kind of
structure of the supports to study the interactions between catalyst.
supported species and the supports at the molecular and The investigation of copper‐based catalysts in catalytic CO
atomic levels, the single crystals of metal‐oxide supports re‐ elimination has been reviewed briefly in our previous publica‐
quired for study are difficult to obtain, meaning that direct ob‐ tion [37]. To allow future research work on this subject to pro‐
servation of the interactions between metal oxides and a single ceed in a systematic fashion, this paper offers a comprehensive
crystal support is also difficult. Based on this, to allow the in‐ review of our work to date concerning binary and ternary
vestigation of the interactions between different types of metal supported copper‐based catalysts with applications in CO oxi‐
oxides and commonly used supports, the preferentially ex‐ dation, as well as in CO+NO and NH3+NO+O2 reactions. Our
posed crystal planes of supports are primarily considered and present work is focused primarily on characterizing the physi‐
other co‐existing planes are ignored when considering the sur‐ cochemical properties and catalytic performance of cop‐
face structures of supports. Subsequently, based on various per‐based catalysts supported on CeO2, CexZr1‐xO2, and
characteristics of the supported species (such as valence state MoO3‐CeO2, exploring the interactions of the oxide components
and ionic radius), the so‐called incorporation model [35,36] is and the relationships among compositions, structures, and
employed to describe the interactions. There are five funda‐ catalytic properties with the purpose of providing a scientific
mental aspects of the catalyst that are taken into account in this basis to assist in the design of highly efficient supported cop‐
model: (1) the existence of appropriate vacancies on the sur‐ per‐based catalysts. Some representative examples of our re‐
face of the supports allows the supports to serve as the frame‐ search work are discussed in the following sections.
work for the spontaneous dispersion of metal ionic com‐
pounds; (2) the dispersion of supported ionic compounds on 2. CuO/CeO2
the surface of the supports is based on incorporation of cations
into surface vacancies, while keeping associated anions in close Since the early 1980s, the use of CeO2 in three‐way catalysts
proximity so as to maintain electric neutrality; (3) the disper‐ (TWCs) applied to automotive pollution control and in wa‐
sion capacities of different types of ionic compounds, such as ter‐gas shift reaction catalysts has become so widespread be‐
1‐1, 1‐2, 2‐1, 1‐3, and 2‐3, may vary because the shielding ef‐ cause of its excellent oxygen storage/release capacity (OSC)
fects produced by the accompanying anions differ; (4) different and redox properties, and now represents the most important
supports have different types of surface vacancies (tetrahedral, application of the rare earth oxides [38,39]. Until the late
octahedral, and cubic), and the same support may have more 1980s, noble metals (such as Rh and Pt) and CeO2 were used as
than one kind of vacancy (for example, both tetrahedral and the active components and support in TWCs with the purpose
octahedral vacancies exist on the surface of ‐Al2O3), and thus it of simultaneously converting CO, hydrocarbons (HCs), and NOx
is necessary to consider the steric or preferential effect when to CO2, H2O, and N2. In this process, it is necessary to maintain
discussing the incorporation of cations into vacancies (that is, the appropriate air‐to‐fuel ratio and to work at high tempera‐
the relationship between cation radius and vacancy size, as tures (> 327 °C). However, these conditions are difficult to meet
 DONG Lin et al. / Chinese Journal of Catalysis 34 (2013) 851–864

CeO2
500 Ce(III)Cu5
(200)
Ce(III)Cu15
CuO CuO 400 Ce(III)Cu30

TOF of CO (s1)
(111) (111) Ce(III)Cu50
(5) Ce(IV)Cu5
300
Intensity

Ce(IV)Cu15
(4) Ce(IV)Cu30
200 Ce(IV)Cu50
(3)
(2) 100

(1)
0
31 33 35 37 39 41 40 50 60 70 80
2/( o )
Temperature (oC)

Fig. 1. XRD patterns of CuO/CeO2 catalysts with Cu2+ loadings of 0.90 Fig. 3. Turnover frequency (TOF) of CO per copper atom in CuO/CeO2
(1), 1.20 (2), 1.40 (3), 1.55 (4), and 1.70 mmol Cu2+/100 m2 CeO2 (5) catalysts as a function of temperature. In the legend, the Roman numer‐
[40]. al following Ce indicates the valence of Ce used as the precursor while
the numeral following Cu represents its molar percent in the catalyst
[8].
while a vehicle is in operation, especially because incomplete
combustion can emit a significant quantity of hydrocarbons on
sulting from the incorporation model, which considers that all
cold‐starting of the engine. Furthermore, because of anticipated
the cation vacancies on the (111) plane of CeO2 are occupied by
future worldwide restrictions concerning exhaust emissions,
Cu2+ and thus predicts 1.22 mmol CuO/100 m2 CeO2. CuO exists
investigations continue with the aim of further improving the
in a highly dispersed state when the copper content is below
performance of TWCs. A significant improvement applied by
the dispersion capacity, and no diffraction lines of CuO appear
some automobile companies in the mid‐1990s was the re‐
in the XRD patterns of the materials, whereas crystalline CuO
placement of CeO2 by CeO2‐ZrO2, which enhanced the catalyst’s
can be detected when the copper content exceeds the disper‐
thermal stability and OSC efficiency. This technique, when pub‐
sion capacity.
lished in 1997, prompted a great deal of interest in researching
CeO2 not only acts as the support but also performs the role
the modification of supports. In addition, there are ongoing
of promoter when CuO/CeO2 is used for CO oxidation, and this
attempts to find substitutes for precious metals in catalysts,
promotion effect is closely related to the ratio of Ce4+/Ce3+ and
and CuO has been identified by many research projects as a
their transition during the reaction. Ce salts (Ce(NO3)3·6H2O
good candidate for this purpose.
and (NH4)2Ce(NO3)6) with different valence states were used as
To investigate the existing status and redox properties of
precursors to prepare a series of CuO/CeO2 catalysts, and their
CuO supported on CeO2 as well as the interactions between CuO
catalytic performance for CO oxidation was measured [8]. The
and CeO2, we have prepared a series of CuO/CeO2 catalysts
corresponding results are presented in Fig. 3, from which sev‐
with different copper contents using a wet impregnation
eral conclusions can be made. First, comparing the two series of
method and characterized these materials by various spectro‐
samples prepared by different valence state precursors, we can
scopic methods [40]. The results of X‐ray diffraction (XRD, Fig.
see that the catalytic activities of Ce3+ samples are superior to
1) and quantitative analysis by X‐ray photoelectron spectros‐
those of Ce4+ samples at equivalent CuO loadings. This may be
copy (XPS, Fig. 2) demonstrate that the dispersion capacity of
because the content of Ce3+ in the former is higher, a situation
CuO on CeO2 is about 1.20 mmol CuO/100 m2 CeO2 (or 7.1
conducive to the formation of Cu+‐□ ‐Ce3+ (surface synergetic
Cu2+/nm2 CeO2). This value is consistent with the estimate re‐
oxygen vacancy (SSOV)), which further improves the catalytic
activities of these samples, as seen in Fig. 4. Second, comparing
the samples with different CuO loading amounts prepared with
0.4 the same precursor, the catalytic activity increases with in‐

CO CO2
ICu 2p /ICe 3d

SSOV
0.2 Cu2+-O-Ce4+ Cu+- -Ce3+
(Step 1)
CO/O2
CO2

(Step 3)

(Step 2)

O O
0.0 C O
0.00 2.26 4.52 6.78 9.04 11.30 13.56 15.82
Total content of copper/nm2 Cu - -Ce3+
+

Fig. 2. The relationship between CuO content and intensity ratio in Fig. 4. Proposed reaction mechanism for CO oxidation over CuO/CeO2
CuO/CeO2 samples as determined by XPS [40]. catalysts [8].
 DONG Lin et al. / Chinese Journal of Catalysis 34 (2013) 851–864

creasing CuO loading up until the appearance of bulk CuO. More stability [43,45–47].
bulk CuO is generated with further increases in CuO loading, The results of XRD analysis show that ceria‐rich CexZr1‐xO2
and the catalytic activity subsequently decreases dramatically. (x ≥ 0.5) solid solutions maintain the fluorite cubic structure,
This result is in agreement with the prediction of the incorpo‐ which is in accordance with the results of Ayastuy et al. [48].
ration model. Initially, the highly dispersed Cu2+ ions succes‐ Therefore, if the preferentially exposed surface of CexZr1‐xO2 (x
sively occupy the cation vacancies of the CeO2 surface, increas‐ ≥ 0.5) is also the (111) plane (Fig. 5), the same as in CeO2 [49],
ing the probability of Cu2+‐O‐Ce4+ formation. Further increases we can understand the dispersion of CuO and the coordination
in CuO loading after its dispersion capacity has been exceeded environment of Cu2+ supported on CexZr1‐xO2 (x ≥ 0.5).
will lead to the appearance of crystalline CuO, after which the The dispersion and reduction properties of catalysts based
catalytic activity per single Cu atom decreases because crystal‐ on CuO supported on Ce0.67Zr0.33O2 (denoted as CuO/CZ), as
line CuO makes no obvious contribution to the formation of well as their catalytic performance for the NO+CO reaction,
bridging Cu2+‐O‐Ce4+ units. Based on these results, we have have been investigated [50]. As shown in Fig. 6, the dispersion
proposed a catalytic reaction mechanism associated with SSOV capacity of CuO on a Ce0.67Zr0.33O2 support is between 0.66 and
[41]. In this mechanism, the surface oxygen in Cu2+‐O‐Ce4+ is 1.23, while the theoretical value is about 1.20. In addition, bulk
depleted by pre‐reduction to form surface synergetic oxygen CuO is detected in 1.23Cu‐CZ. Figure 7 shows that the peak
vacancies, which promote the CO oxidation reaction. areas for 0.66Cu‐CZ and 0.33Cu‐CZ at 224 °C are approximately
equal, indicating that this peak is related to the reduction of
3. Bulk modification of the support: CuO/CexZr1‐xO2 surface Ce4+ to Ce3+. The significant increase of the peak area in
1.23Cu‐CZ may be related to the formation of crystalline CuO,
Modification of the supports is a common means of improv‐ with the continued increase in 1.65Cu‐CZ due to the formation
ing catalytic performance. The properties of the supports can of more crystalline CuO, which is in agreement with the results
be modulated by the addition of appropriate oxides so as to of XRD. Figure 8 shows that the catalyst with copper content
obtain binary or multivariate solid solutions, spinel, perovskite
oxides, or mixed oxides. This is directly related to the interac‐
tions between these oxides and the supported active species. ◆ CuO
There is a great deal of literatures concerning the screening and
evaluation of causes of these interactions, and the results are ◆
◆
complex. CeO2 has a number of advantages, but its thermal 1.65Cu-CZ
Intensity

stability is not good and the high‐temperature sintering effect

can severely affect its reactivity [39,42,43]. It is well known 1.23Cu-CZ
that a CuO/ZrO2 catalyst employing ZrO2 as a support exhibits
excellent catalytic performance for CO oxidation. Moreover, the 0.66Cu-CZ
thermal stability of CuO/ZrO2 is better than that of CuO/CeO2 in
the reforming reaction of methanol steam [44]. The phase dia‐ 0.33Cu-CZ
gram of CeO2‐ZrO2 indicates that the solid solution can be
formed over a relatively large concentration range [39]. The 10 20 30 40 50 60 70 80
radius of Zr4+ (0.084 nm) is smaller than that of Ce4+ (0.097 2/( o )
nm), so the incorporation of Zr4+ into the fluorite cubic lattice Fig. 6. XRD patterns of xCu‐CZ catalysts with different CuO loadings
can lead to a decrease in cell volume and lattice parameter in [50].
the ceria‐rich CeO2‐ZrO2 solid solution. Furthermore, the ther‐
mal stability of the support, mobility of bulk lattice oxygen, and 188 231
oxygen storage/release capacity of the catalyst are also im‐ crystalline CuO
145
proved. In fact, the CeO2‐ZrO2 solid solution has already taken
the place of CeO2 as a support because of its excellent thermal 1.65Cu-CZ
H2 consumption

1.23Cu-CZ

0.66Cu-CZ
187
154
224
0.33Cu-CZ

100 150 200 250 300 350

Temperature (oC)

Fig. 5. Crystal structure (a) and preferentially exposed (111) plane (b) Fig. 7. Temperature‐programmed reduction by H2 (H2‐TPR) profiles of
of CeO2 or CexZr1xO2 [49]. xCu‐CZ catalysts with different CuO loadings [50].
 DONG Lin et al. / Chinese Journal of Catalysis 34 (2013) 851–864

100 100

98 98 319
309 (8)
NO conversion (%)

N2 selectivity (%)
96 96
203 303
94 94 (7)
163
205 299

Intensity
165 (6)
92 92 210 299 (5)
165
90 90 210
168 231 263 (4)
88 88 170 (3)
234
170 (2)
0.33Cu-CZ 0.66Cu-CZ 1.23Cu-CZ 1.65Cu-CZ
235
172 (1)
Fig. 8. NO conversion and N2 selectivity over xCu‐CZ catalysts with
different CuO loadings at a reaction temperature of 300 °C [50]. 100 200 300 400 500
Temperature (oC)
around the dispersion capacity exhibits the best catalytic per‐
Fig. 10. H2‐TPR profiles of CuO/CZ samples containing 0.3 (1), 0.6 (2),
formance. 0.9 (3), 1.2 (4), 1.8 (5), 2.4 (6), and 3.0 mmol CuO/100 m2 CZ (7), and
These results also indicate that it is very important to give crystalline CuO (8) [49].
sufficient consideration to the dispersion capacity value of ac‐
tive species, as well as the preparation of high dispersion sam‐ the preceding section, indicating that the highly dispersed CuO
ples by appropriate methods to avoid the waste of active spe‐ is the main active species, but also that an excessive loading
cies in the preparation of supported catalysts. decreases the catalytic activity estimated according to the
Samples with different CuO loading amounts (denoted as turnover frequency (TOF).
CuO/CZ) were prepared by an impregnation method employ‐
ing the Ce0.5Zr0.5O2 support [49]. The XRD patterns of these 4. Surface modification of the support: CuO/MoO3‐CeO2
materials (Fig. 9) show that the dividing point at which crystal‐
line CuO appears is at 0.9 mmol CuO/100 m2 CZ. H2‐TPR results Modification of the supports by means of oxide‐support in‐
(Fig. 10) also confirm that there are only two peaks at 170 and teractions is widely used in industry. We have previously in‐
235 °C attributed to the reduction of highly dispersed CuO for vestigated the catalytic performance of copper oxide supported
those samples with copper content below 0.9 mmol CuO/100 on MoO3‐modified CeO2 (MoO3‐CeO2). MoO3 was chosen be‐
m2 CZ, while the peak at 309 °C due to the reduction of crystal‐ cause it is widely used in many important catalytic reactions
line CuO can only be found in samples with higher loadings. such as olefin metathesis, oxidation, and hydrogenation pro‐
The catalytic activities of these samples for low temperature CO cessing. Furthermore, its low Taman temperature facilitates the
oxidation are displayed in Fig. 11, and the sample with 1.2 dispersion of MoO3 on oxide supports. Partially MoO3‐covered
mmol CuO/100 m2 CZ exhibits the best catalytic performance. and fully MoO3‐covered (via a close‐packed monolayer) sup‐
These results are consistent with the discussion of CuO/CeO2 in ports can be easily prepared by adjusting the loading of MoO3,
which is beneficial in terms of investigating the physicochemi‐
cal properties of these systems. Previous research work has
o CuO shown that MoO3‐CeO2 catalysts may be applied to many cata‐
lytic reactions, such as CO low temperature oxidation and wa‐
o o

(7) 70
Intensity

(6) 60
125 oC
CO conversion (%)

(5) 50
(4)
40
(3)
(2) 30
(1) 20
100 oC
20 25 30 35 40 45 50 55 10
2 /( o ) 0.0 0.6 1.2 1.8 2.4 3.0 3.6
Copper oxide loading (mmol CuO/100 m2 CZ)
Fig. 9. XRD patterns of CuO/CZ samples containing 0.3 (1), 0.6 (2), 0.9
(3), 1.2 (4), 1.8 (5), 2.4 (6), and 3.0 mmol CuO/100 m2 CZ (7) [49]. Fig. 11. Catalytic activities of CuO/CZ catalysts [49].
 DONG Lin et al. / Chinese Journal of Catalysis 34 (2013) 851–864

ter electrolysis [51,52]. With regard to the latter, Hang et al.

[52] reported that MoO3 did not incorporate into the CeO2 lat‐
tice in MoO3/CeO2 samples, but rather catalytic activity result‐
ed from the formation of Mo‐O‐Ce on the surface of the materi‐
al, such that the best catalytic activity was obtained when MoO3
loading was at the extent of its dispersion capacity, at which
point MoO3 species formed a close‐packed monolayer. This
result agrees with our own findings concerning CuO/CeO2. In
addition, Rathod et al. [53] investigated the application of
mixed oxides in organic reactions in an attempt to substitute
Fig. 12. Schematic representation of dispersed CuO on the (111) plane
these for conventional inorganic acid catalysts because the
of a CeO2 support that has been monolayer MoO3‐modified (a) and
current catalysts are toxic, environmentally unfriendly and partial layer MoO3‐modified (b) [55].
difficult to recycle. MoO3/CeO2‐ZrO2 exhibited its highest activ‐
ity for the catalytic synthesis of benzimidazole derivatives
when the loading of MoO3 was 20 wt%, and this catalytic sys‐ some of the supported CuO on the surface of the MoO3 is in
tem was also shown to be easy to recycle and recover. direct contact with CeO2. These two cases are shown in Fig.
We have previously examined catalysts based on CuO sup‐ 12(a) and (b).
ported on MoO3‐modified CeO2 and have investigated their The XRD results obtained for CuO/0.8Mo‐Ce and
catalytic performance when applied to low temperature CuO/0.4Mo‐Ce are shown in Fig. 13. It can be seen from Fig.
NH3‐SCR [54,55]. These samples are denoted as CuO/0.8Mo‐Ce 13(b) and (d) that the dispersion capacity of CuO is 2.06
and CuO/0.4Mo‐Ce in the following discussion. The 0.8Mo‐Ce mmol/100 m2 on the monolayer MoO3‐modified support
material approaches the dispersion capacity obtained when (0.8Mo‐Ce), which is significantly higher than that of unmodi‐
MoO3 is supported on CeO2. According to the incorporation fied CeO2 (1.20 mmol CuO/100 m2 CeO2). This result shows that
model, the surface of CeO2 in this material is covered by a close‐packed monolayer MoO3 modification is beneficial to the
close‐packed monolayer formed by oxygen ions accompanied dispersion of CuO. A possible explanation for this phenomenon
by Mo6+. As a result, the supported CuO does not directly con‐ is that CuO disperses in the tetrahedral vacancies formed by
tact the CeO2. In the case of the 0.4Mo‐Ce samples, the CeO2 the close‐packed monolayer of oxygen associated with the
surface is modified only by a partial layer, and consequently MoO3. The estimated theoretical dispersion capacity of this

(a) (b)
 CuO  CuO
Intensity

Intensity

 
(6) (8)
  (7)
(5) (6)
(4) (5)
(3) (4)
(2) (3)
(2)
(1) (1)

20 30 40 50 20 30 40 50
2/( o ) 2/( o )

(c) (d)
ICuO(111)/ICeO2(111)
ICuO(111)/ICeO2(111)

0.73 2.06

0.0 0.5 1.0 1.5 2.0 0 1 2 3 4 5 6

CuO loading (mmol/100 m2 CeO2) CuO loading (mmol/100 m2 CeO2)

Fig. 13. XRD patterns of xCu/yMo‐Ce samples with different MoO3 and CuO loadings. (a) xCu/0.4Mo‐Ce with CuO loadings of 0 (1), 0.3 (2), 0.6 (3), 0.9
(4), 1.2 (5), and 1.8 mmol/100 m2 (6); (b) xCu/0.8Mo‐Ce with CuO loadings of 0 (1), 0.3 (2), 0.6 (3), 0.9 (4), 1.8 (5), 2.4 (6), 3.6 (7), and 4.8 mmol/100
m2 (8); Quantitative XRD results of xCu/0.4Mo‐Ce (c) and xCu/0.8Mo‐Ce (d) samples [55].
 DONG Lin et al. / Chinese Journal of Catalysis 34 (2013) 851–864

(6)
800 900
965
(5)
235
(4)
Intensity

Intensity
(4)
(3)
196 (3)
(2)
170
(2)
(1)
(1)

150 200 250 300 350 400 600 700 800 900 1000
Temperature (oC) Raman shift (cm1)

Fig. 14. H2‐TPR profiles of xCu/yMo‐Ce samples. (1) 0.3Cu/0.4Mo‐Ce; Fig. 16. LRS of xCu/0.8Mo‐Ce samples, in which x is 0 (1), 0.3 (2), 0.6
(2) 0.6Cu/0.4Mo‐Ce; (3) 0.3Cu/0.8Mo‐Ce; (4) 0.6Cu/0.8Mo‐Ce; (5) (3), and 0.9 mmol/100 m2 (4) [55].
0.3Cu/Ce; (6) 0.8Mo‐Ce [55].
0.6Cu/0.8Mo‐Ce), there is no low temperature peak associated
material, 2.44 mmol/100 m2, is close to the measured value, with the reduction of Cu‐O‐Ce species, which demonstrates that
although somewhat higher, perhaps because the introduction the strong interaction between the close‐packed monolayer of
of CuO results in changes to the close‐packed monolayer struc‐ MoO3 and CeO2 obstructs contact between supported CuO and
ture of the surface MoO3. CeO2. Consequently, the interaction between CuO and CeO2 is
Based onthe above analysis, the dispersed Cu2+ on 0.8Mo‐Ce very weak in these samples.
is in a four‐coordination environment and is more difficult to Laser Raman spectroscopy (LRS) results indicate that the
reduce in comparison with Cu2+ on the unmodified CeO2, which dispersion of MoO3 on the surface of CeO2 shifts the peak asso‐
is in a five‐coordination environment. From the XRD patterns in ciated with the highly dispersed Mo=O end group to higher
Fig. 13(a) and (c), the CuO dispersion capacity is only wavenumbers with increases in the MoO3 loading, from 921
0.73mmol CuO/100 m2 CeO2 on the partial layer cm–1 for 0.2Mo/Ce to 964 cm–1 for 0.8Mo/Ce. The microstruc‐
MoO3‐modified support (0.4Mo‐Ce), which is lower than that of ture of the highly dispersed surface MoO3 species changes from
unmodified CeO2 (1.20 mmol CuO/100 m2 CeO2), indicating a common single Mo‐O tetrahedral species at low loadings (Fig.
that the random distribution of MoO3 on CeO2 does not pro‐ 15(a), left) to a polyoctahedral molybdenum oxide species at
mote the dispersion of CuO. One possible reason is that the high loadings, as per the diagram presented in Fig. 15(b), left.
formation of tetrahedral vacancies on the close‐packed mono‐ As a result of coordination with CuO on these modified
layer is hindered because O2– accompanied with Mo6+ can supports, the Mo=O peak obtained from LRS is reduced. In
shield some vacancies on the CeO2. Copper species can also samples supporting CuO on 0.4Mo‐Ce, Mo species exhibit a
interact with CeO2 and MoO3 on the surface of 04Mo‐Ce. The tetrahedral structure while in those supporting CuO on
above discussion is consistent with the H2‐TPR results dis‐ 0.8Mo‐Ce, the original regular polyoctahedral structure is dis‐
cussed below. rupted because of weakening or even breaking of some
In the H2‐TPR profiles of 0.3Cu/Ce, 0.3Cu/0.4Mo‐Ce, and Mo–O–Mo bonds. This process is illustrated in Fig. 15(a) and
0.6Cu/0.4Mo‐Ce samples in Fig. 14, the peaks at 170 and 196 °C (b), right.
reflect the reduction of interacting species of CuO‐CeO2, while The M=O bond of octahedral (~965 cm–1) and tetrahedral
the shoulder peak at 235 °C is due to the reduction of Cu‐O‐Mo (~900 cm–1) species in these Cu/0.8Mo‐Ce samples changes
species. It is worth noting that among the samples modified with increasing CuO loadings, as confirmed by LRS results (Fig.
with a close‐packed oxygen monolayer (0.3Cu/0.8Mo‐Ce and 16). In 0.3Cu/0.8Mo‐Ce, 0.6Cu/0.8Mo‐Ce, and 0.9Cu/0.8Mo‐Ce
(a) Cu
O O O O
Mo Mo
O O O O

Cu Cu Cu

(b) O O
O O O O O O O O O O O
O
Mo Mo Mo Mo Mo Mo
O O O O O O O O O O O O

Fig. 15. Structures of molybdenum surface species in xCu/0.4Mo‐Ce (a) and xCu/0.8Mo‐Ce (b) [55].
 DONG Lin et al. / Chinese Journal of Catalysis 34 (2013) 851–864

(a) 10 1355 (b) 1355 (c) 1345

10 10
400 oC
400 oC 400 oC
350 oC
350 oC 350 oC
300 oC
300 oC

Transmittance (%)
Transmittance (%)

300 oC

Transmittance (%)
250 oC 250 oC
250 oC
200 oC 200 oC
200 oC
150 oC 150 oC
150 oC
100 oC 100 oC
100 oC
50 oC 50 oC
L1 50 oC
1255 25 oC
25 oC
25 oC
1630 B L1 B L2
1447 L2
1255 1221 1453 L1 1235
1624 1253
1442 B 1624
1700 1600 1500 1400 1300 1200 1100 1700 1600 1500 1400 1300 1200 1700 1600 1500 1400 1300 1200
Wavenumber (cm1) Wavenumber (cm1) Wavenumber (cm1)
Fig. 17. NH3 adsorption as monitored via in situ FT‐IR spectra of xCu/0.8Mo‐Ce samples. (a) 0.8Mo‐Ce; (b) 0.3Cu/0.8Mo‐Ce; (c) 0.6Cu/0.8Mo‐Ce [55].

samples, as the CuO loading increases, the proportions of octa‐ the role of B acid sites may become more important. These
hedral structures decrease to approximately 0.20, 0.12, and results demonstrate that the NO+NH3+O2 catalytic reaction
0.05, respectively. These results indicate that the introduction may progress by different reaction mechanisms in different
of CuO gradually transitions the octahedral Mo species to the temperature zones.
tetrahedral structure and decreases the degree of polymeriza‐ In addition, we have investigated interactions among the
tion of Mo species because of the strong interaction between oxide components of multi‐oxide systems formed by CuO sup‐
CuO and Mo species. ported on CexZr1‐xO2 and TixSn1‐xO2 supports modified by oxides
In situ FTIR of NH3 adsorption at different temperatures (such as CuO/MnOx‐CexZr1‐xO2 [56], CuO/CoOx‐CexZr1‐xO2 [57],
was carried out to investigate the relationship between differ‐ CuO/WO3‐CexZr1‐xO2 [58,59], CuO/TixSn1‐xO2 [60], CuO/VOx‐
ent Mo species and surface acidity in these Cu/Mo‐Ce samples TixSn1‐xO2 [61] and others). The dispersion of active compo‐
(Fig. 17). The results show that Mo species in the polymeric nents and the relationships among compositions, structures,
octahedral structure form surface Brönsted (B) acid sites, while and properties in these catalysts have also been investigated,
those in isolated tetrahedral structures form weak Lewis acid especially with regard to the oxidation of CO and the selective
sites (L1). The introduction of CuO evidently reduces the quan‐ catalytic reduction of NO.
tity of B acid sites and produces new and stronger Lewis acid Research concerning interactions among catalytic oxide
sites (L2). Table 1 lists the quantitative results; the behavior of components is a very wide and complex field of study, and it is
0.4Mo‐Ce and 0.8Mo‐Ce samples is similar, but the acid important that this work continues, so as to allow progress in
strength and quantity of acid sites in the latter are higher. both the theoretical understanding and the practical applica‐
The catalytic performance of some CuO/MoO3‐CeO2 samples tions of these materials, especially in heterogeneous catalysis
during the NH3‐SCR reaction as a function of temperature is and material preparation processes. Previous investigations of
shown in Fig. 18. Generally speaking, the series of 0.8Mo‐Ce surface models and the nature of the interactions, as well as the
samples exhibit higher activity than the 0.4Mo‐Ce catalysts. As
shown in Table 1, the former has more acid sites than the latter,
which indicates that surface acid sites play an important role in 0.4Mo-Ce 0.3Cu/0.4Mo-Ce
70 0.6Cu/0.4Mo-Ce 0.8Mo-Ce
NH3‐SCR. The addition of CuO improves the catalytic activity
0.3Cu/0.8Mo-Ce
between 200 and 250 °C, which is related to the increase in L 60
NO conversion (%)

0.6Cu/0.8Mo-Ce
acid sites in these samples. At 300–350 °C, the introduction of 50
CuO leads to a decrease in catalytic activity, perhaps because
40
Table 1
Peak areas of Brönsted acid and Lewis acid sites on xCu/yMo‐Ce sam‐ 30
ples as measured by chemisorption of NH3 at 25 °C [55].
20
Peak area
Sample 10
Brönsted (B) Lewis (L1 + L2)
0.4Mo/Ce 640 44
0
0.3Cu/0.4Mo‐Ce 535 535 200 250 300 350
0.6Cu/0.4Mo‐Ce 50 975 Temperature (oC)
0.8Mo/Ce 2530 59
0.3Cu/0.8Mo‐Ce 660 707 Fig. 18. NO conversion over different catalysts during the NH3+NO+O2
0.6Cu/0.8Mo‐Ce 540 1585 reaction [55].
 DONG Lin et al. / Chinese Journal of Catalysis 34 (2013) 851–864

development of modern spectroscopic techniques and theoret‐ [11] Lund C R F, Dumesic J A. J Phys Chem, 1981, 85: 3175
ical methods, currently provide favorable circumstances that [12] Olthof B, Khodakov A, Bell A T, Iglesia E. J Phys Chem B, 2000, 104:
should allow a more thorough understanding of these catalysts. 1516
Our own work to date can be seen as only a drop in the ocean, [13] Xie S B, Iglesia E, Bell A T. J Phys Chem B, 2001, 105: 5144
[14] Xie S B, Iglesia E, Bell A T. Langmuir, 2000, 16: 7162
and additional systematic investigations of the modification of
[15] Cheung P, Liu H C, Iglesia E. J Phys Chem B, 2004, 108: 18650
both active species and supports are required to find practical
[16] Dai H X, Bell A T, Iglesia E. J Catal, 2004, 221: 491
applications for these catalytic systems and to provide a scien‐ [17] Wachs I E. Catal Today, 1996, 27: 437
tific basis for the creation and development of improved cata‐ [18] Hardcastle F D, Wachs I E, Horsley J A, Via G H. J Mol Catal, 1988,
lysts. 46: 15
[19] Wang X, Wachs I E. Catal Today, 2004, 96: 211
5. Summary and outlook [20] Tian H J, Wachs I E, Briand L E. J Phys Chem B, 2005, 109: 23491
[21] Knözinger H, Taglauer E. In: Spivey J J, Agarwal S K eds. Catalysis.
The dispersion of CuO (the main active species) supported Vol. 10. Cambridge: The Royal Society of Chemistry, 1993. 1
on the surface of CeO2, CexZr1‐xO2, and MoO3‐CeO2 has been [22] Coluccia S, Marchese L. Catal Today, 1998, 41: 229
investigated. Based on the incorporation model, differences in [23] Reddy B M, Sreekanth P M, Yamada Y, Xu Q, Kobayashi T. Appl
Catal A, 2002, 228: 269
the dispersion behavior of CuO on different supports, as well as
[24] Reddy B M, Chowdhury B, Smirniotis P G. Appl Catal A, 2001, 211:
variations in the corresponding physicochemical properties of
19
the materials, have been explained. We have attempted to es‐ [25] Reddy B M, Chowdhury B. J Catal, 1998, 179: 413
tablish a structure‐activity relationship by using a variety of [26] Xie Y C, Tang Y Q. Adv Catal, 1990, 37: 1
characterization methods (including in situ and quasi situ) [27] Yan J H, Liu Y J, Gui L L, Tang Y Q. Acta Phys‐Chim sin, 1993, 9: 13
combined with measurements of the performance of these [28] Luo S C, Gui L L, Tang Y Q. Acta Phys‐Chim Sin, 1996, 12: 341
catalysts for CO oxidation as well as CO+NO and NH3+NO+O2 [29] Fu X Z, Luo S C, Gui L L, Luo X H. Acta Chim Sin, 1992, 50: 342
model reactions. There is still some uncertainty, however, re‐ [30] Fu X Z, Luo S C, Gui L L, Tang Y Q. Prog Natural Sci, 1991, 1: 543
garding the reaction kinetics of such catalysts in terms of elec‐ [31] Goodman D W. J Catal, 2003, 216: 213
tron transfer and reaction mechanisms related to valence tran‐ [32] Hess C, Ozensoy E, Goodman D W. J Phys Chem B, 2003, 107:
sitions of the copper species. A more comprehensive evaluation 2759
[33] Xu C, Oh W S, Goodman D W. J Phys Chem B, 2000, 104: 10310
of these catalysts is required, including studies of their re‐
[34] Chen M S, Goodman D W. Science, 2004, 306: 252
sistance to water and sulfur poisoning and aging characteris‐
[35] Chen Y, Dong L, Jin Y S, Xu B, Ji W J. Stud Surf Sci Catal, 1996, 101:
tics, and we intend to pursue such research in the future. 1293
[36] Chen Y, Zhang L F. Catal Lett, 1992, 12: 51
CHEN Yi notes: We previously published a paper entitled [37] Yu Q, Gao F, Dong L. Chin J Catal, 2012, 33: 1245
“The interaction between the double metal oxides and oxide [38] Trovarelli A. Catal Rev‐Sci Eng, 1996, 38: 439
support components” in the Journal of Fudan University (Natu‐ [39] Kašpar J, Fornasiero P, Graziani M. Catal Today, 1999, 50: 285
ral Science) to commemorate the anniversary of the passing of [40] Dong L, Jin Y S, Chen Y. Sci China (Ser B), 1996, 26: 561
Professor DENG Jingfa in March, 2002. Professor Kangnian [41] Li D, Yu Q, Li S S, Wan H Q, Liu L J, Qi L, Liu B, Gao F, Dong L, Chen
later suggested that we also write a paper to commemorate the Y. Chem‐Eur J, 2011, 17: 5668
80th anniversary of Professor Jingfa’s birth. In the intervening [42] Kašpar J, Fornasiero P, Hickey N. Catal Today, 2003, 77: 419
[43] Wang S P, Zheng X C, Wang X Y, Wang S R, Zhang S M, Yu L H,
ten years, additional work in this field has been performed by
Huang W P, Wu S H. Catal Lett, 2005, 105: 163
Professor DONG Lin’s research group at Nanjing University.
[44] Oguchi H, Kanai H, Utani K, Matsumura Y, Imamura S. Appl Catal A,
Professor Dong therefore submits this paper to the special is‐ 2005, 293: 64
sue of this journal in fond memory of Professor Deng. [45] Liotta L F, Macaluso A, Longo A, Pantaleo G, Martorana A, De‐
ganello G. Appl Catal A, 2003, 240: 295
References [46] Kakuta N, Ikawa S, Eguchi T, Murakami K, Ohkita H, Mizushima T. J
Alloys Comp, 2006, 408‐412: 1078
[1] Jones S D, Hagelin‐Weaver H E. Appl Catal B, 2009, 90: 195 [47] Larese C, Granados M L, Mariscal R, Fierro J L G, Lambrou P S,
[2] Jobbágy M, Marino F, Schönbrod B, Baronetti G, Laborde M. Chem Efstathiou A M. Appl Catal B, 2005, 59: 13
Mater, 2006, 18: 1945 [48] Ayastuy J L, Gurbani A, González‐Marcos M P, Gutiérrez‐Ortiz M A.
[3] Guo S H, Chen X R. J Hebei University Technol, 2009, 38(3): 24 Int J Hydrogen Energy, 2012, 37: 1993
[4] Hu C Q, Zhu Q S, Jiang Z, Zhang Y Y, Wang Y. Microporous Mesopo‐ [49] Chen H L, Zhu H Y, Wu Y, Gao F, Dong L, Zhu J J. J Mol Catal A, 2006,
rous Mater, 2008, 113: 427 255: 254
[5] Li Y, Fu Q, Flytzani‐Stephanopoulos M. Appl Catal B, 2000, 27: 179 [50] Liu L J, Liu B, Dong L H, Zhu J, Wan H Q, Sun K Q, Zhao B, Zhu H Y,
[6] Brands D S, Poels E K, Bliek A. Appl Catal A, 1999, 184: 279 Dong L, Chen Y. Appl Catal B, 2009, 90: 578
[7] Bera P, Priolkar K R, Sarode P R, Hegde M S, Emura S, Kumashiro [51] Mohamed M M, Katib S M A. Appl Catal A, 2005, 287: 236
R, Lalla N P. Chem Mater, 2002, 14: 3591 [52] Huang H, Guo Z C. J Inorg Mater, 2012, 27: 939
[8] Qi L, Yu Q, Dai Y, Tang C J, Liu L J, Zhang H L, Gao F, Dong L, Chen Y. [53] Rathod S B, Lande M K, Arbad B R. Bull Korean Chem Soc, 2010, 31:
Appl Catal B, 2012, 119‐120: 308 2835
[9] Selwood P W. Adv Catal, 1951, 3: 27 [54] Zhu J, Gao F, Dong L H, Yu W J, Qi L, Wang Z, Dong L, Chen Y. Appl
[10] Tauster S J, Fung S C, Garten R L. J Am Chem Soc, 1978, 100: 170 Catal B, 2010, 95: 144
 DONG Lin et al. / Chinese Journal of Catalysis 34 (2013) 851–864

Graphical Abstract
Chin. J. Catal., 2013, 34: 851–864 doi: 10.1016/S1872‐2067(12)60592‐0
Interactions among supported copper‐based catalyst
components and their effects on performance: A review
DONG Lin *, YAO Xiaojiang, CHEN Yi
Nanjing University
This review summarizes the relationships among the compo‐
sitions, structures and properties of the copper‐based cata‐
lysts CuO/CeO2, CuO/CexZr1xO2, and CuO/MoO3‐CeO2 when
applied to model reactions involving the oxidation of CO or
the selective catalytic reduction of NO by CO/NH3.

[55] Yu W J, Zhu J, Qi L, Sun C Z, Gao F, Dong L, Chen Y. J Colloid Inter‐ Q, Dong L, Chen Y. Catal Commun, 2009, 10: 741
face Sci, 2011, 364: 435 [59] Li X W, Shen M M, Hong X, Zhu H Y, Gao F, Kong Y, Dong L, Chen Y. J
[56] Liu L J, Yu Q, Zhu J, Wan H Q, Sun K Q, Liu B, Zhu H Y, Gao F, Dong Phys Chem B, 2005, 109: 3949
L, Chen L. J Colloid Interface Sci, 2010, 349: 246 [60] Dong L H, Liu L J, Lv Y Y, Zhu J, Wan H Q, Liu B, Gao F, Wang X S,
[57] Liu L J, Chen Y, Dong L H, Zhu J, Wan H Q, Liu B, Zhao B, Zhu H Y, Dong L, Chen Y. J Mol Catal A, 2012, 365: 87
Sun K Q, Dong L, Chen Y. Appl Catal B, 2009, 90: 105 [61] Dong L H, Zhang L L, Sun C Z, Yu W J, Zhu J, Liu L J, Liu B, Hu Y H,
[58] Li X W, Wan H Q, Liu B, Yang L J, Zhu H Y, Chen H L, Zhao X, Sun K Gao F, Dong L, Chen Y. ACS Catal, 2011, 1: 468