Reduction of NO by CO over nanoscale LaCo1-xCuxO3 and LaMn1-xCuxO3 perovskites

Article abstract of DOI:10.1016/j.molcata.2006.05.008

Nanoscale perovskites with nominal formula of La(Co, Mn)_1-xCu_xO₃ were generated by a novel method designated as reactive grinding and characterized by N₂ adsorption, X-ray diffraction (XRD), temperature programmed desorption (TPD) of O₂, NO, and CO, temperature programmed surface reduction (TPSR) of NO under CO/He flow. Activity tests of NO + CO reaction for those materials were also carried out in this study. A better catalytic performance (93% N₂ yield and 91% CO conversion at 500 °C) was found over LaCoO₃ compared to LaMnO₃ (76% N₂ yield and 76% CO conversion at 500 °C), with a reaction atmosphere of 3000 ppm NO and 3000 ppm CO in helium at a space velocity of 50,000 h^-1. The catalytic activity in NO + CO reaction for LaCoO₃ can be considerably improved via 20% Cu substitution, leading to a 97% N₂ yield and nearly complete CO conversion at 450 °C. This improvement was ascribed to the ease of generation of anion vacancies after Cu incorporation, which plays a crucial role in NO adsorption and dissociation. In addition, the enhancement in lattice oxygen mobility of Cu substituted perovskites promotes the CO oxidation and anion vacancy recovery giving another clue for this improvement. N₂O decomposition (68% N₂ yield at 500 °C) is much easier than NO decomposition (below 5% at T < 500 °C). Both NO and N₂O conversions are significantly improved by the reducing agent. A mechanism was proposed with dissociation of chemisorbed NO forming N₂ and/or N₂O, and oxidized perovskite surface, with continuous reduction by CO with the production of CO₂. O₂ has a strongly detrimental effect leading to the easy consumption of the reducing agent via CO oxidation.

Full text of DOI:10.1016/j.molcata.2006.05.008

Journal of Molecular Catalysis A: Chemical 258 (2006) 22–34
Reduction of NO by CO over nanoscale LaCo1−xCuxO3 and
LaMn1−xCuxO3 perovskites
Runduo Zhanga, Adrian Villanuevaa, Houshang Alamdarib, Serge Kaliaguinea,∗
a Department of Chemical Engineering, Laval University, Ste Foy (QC), Canada G1K 7P4
b Nanox Inc., 4975 rue Rideau, Local 100, Que., Canada G2E 5H5
Received 23 February 2006; received in revised form 24 April 2006; accepted 1 May 2006
Available online 16 June 2006
Abstract
Nanoscale perovskites with nominal formula of La(Co, Mn)1−xCuxO3 were generated by a novel method designated as reactive grinding and
characterized by N2 adsorption, X-ray diffraction (XRD), temperature programmed desorption (TPD) of O2, NO, and CO, temperature programmed
surface reduction (TPSR) of NO under CO/He flow. Activity tests of NO + CO reaction for those materials were also carried out in this study. A
better catalytic performance (93% N2 yield and 91% CO conversion at 500 ◦C) was found over LaCoO3 compared to LaMnO3 (76% N2 yield and
76% CO conversion at 500 ◦C), with a reaction atmosphere of 3000 ppm NO and 3000 ppm CO in helium at a space velocity of 50,000 h−1. The
catalytic activity in NO + CO reaction for LaCoO3 can be considerably improved via 20% Cu substitution, leading to a 97% N2 yield and nearly
complete CO conversion at 450 ◦C. This improvement was ascribed to the ease of generation of anion vacancies after Cu incorporation, which plays
a crucial role in NO adsorption and dissociation. In addition, the enhancement in lattice oxygen mobility of Cu substituted perovskites promotes the
CO oxidation and anion vacancy recovery giving another clue for this improvement. N2O decomposition (68% N2 yield at 500 ◦C) is much easier
than NO decomposition (below 5% at T < 500 ◦C). Both NO and N2O conversions are significantly improved by the reducing agent. A mechanism
was proposed with dissociation of chemisorbed NO forming N2 and/or N2O, and oxidized perovskite surface, with continuous reduction by CO
with the production of CO2. O2 has a strongly detrimental effect leading to the easy consumption of the reducing agent via CO oxidation.
© 2006 Elsevier B.V. All rights reserved.
Keywords: NO reduction; Carbon monoxide; Reactive grinding; High surface area; Perovskite
1. Introduction
Perovskites oxides having the general formula of ABO3,
where B usually designates a transition metal cation surrounded
by six oxygens in octahedral coordination, and A is a cation
of rare-earth metal coordinated by 12 oxygens, have been con-
sidered as promising alternatives to precious metals acting as
effective TWC due to their low cost, high thermal stability,
and diverse physicochemical properties. Some perovskites were
shown to be active in CO oxidation as early as in 1952 [3].
Since Libby showed their potential application for purification
of exhaust gases, perovskites have been recognized as valuable
catalysts for the catalytic reduction of NO by CO [4].
Among the perovskite-type oxides, lanthanum cobaltite and
lanthanum manganite have attracted more attention for this kind
of study due to their special oxygen nonstoichiometries [5].
Their properties can be further modified by the partial substi-
tution of A- or B-site cations. Much research has been done
concerning the A-site substitution and its effect in catalytic activ-
ities toward NO reduction by CO was considered to be related
with the defective structure and valence alternation after substi-
NO and CO, two main pollutants in exhaust gases from gaso-
line engines can cause serious environmental problems such
as photochemical smog, acid rain and respiratory diseases. It
would be an ideal way to simultaneously eliminate those pol-
lutants via catalytic reduction of NO using CO as a reducing
agent. Three-way-catalysts (TWC), mainly composed of plat-
inum and rhodium metal deposited on a CeO2-␥-Al2O3 coating,
werewidelystudiedforthisinterconversionbetweenNOandCO
[1]. Efforts for the replacement of supported noble metals are
however deemed necessary, because of the drawbacks of noble
metals in practical application, including high cost, volatiliza-
tion and sintering at high temperature [2].
∗
Corresponding author. Tel.: +1 418 656 2708; fax: +1 418 656 3810.
E-mail address: kaliagui@gch.ulaval.ca (S. Kaliaguine).
1381-1169/$ – see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2006.05.008

R. Zhang et al. / Journal of Molecular Catalysis A: Chemical 258 (2006) 22–34
23
tution [6–10]. However, much less work was reported about the
effect of B-site substitution on the physicochemical properties
of perovskites as well as their catalytic performances for the
interconversion of NO and CO [11].
Copper containing catalysts are of special interest as they are
active in a wide range of reactions for transformation of nitro-
gen oxides, as reviewed by Centi and Perathoner [12]. Highly
dispersed Cu2+ was generally regarded as an active site for
NO reduction [13–17]. A good performance may therefore be
expected by incorporation of fully isolated copper cations into
the B-site of the perovskite structure.
The application of perovskites has to some extent been lim-
ited by their low specific surface areas resulting from the rather
high temperature at which they are usually generated (700–
800 ◦C) [18]. Some perovskite preparation methods including
microemulsion [19], citrate complexation [20], coprecipitation
[21], sol–gel process [22], spray–drying [23], freeze–drying
[24], and flame–hydrolysis [25], were attempted in order to
improve their surface areas. We have shown recently that a large
variety of perovskites can be synthesized at essentially room
temperature by reactive grinding which allowed to achieve a
specific surface area in excess of 100 m2/g [26–28].
The aim of the present work is to investigate the influ-
ence of Cu substitution on the physicochemical properties of
LaCo1−xCuxO3 and LaMn1−xCuxO3 (x = 0, 0.1, and 0.2) pre-
pared by reactive grinding, in order to clarify the relation
between perovskite properties and their catalytic behaviors in
the reaction between NO and CO. Attempt was also made to
propose an overall mechanism for the catalytic action of per-
ovskites in this reaction.
XRDpatternsofthesolidswererecordedusingadiffractome-
ter (D5000, Siemens) with monochroma◦tized Cu K␣ radiation
◦
◦
˚
(␭ = 1.5406 A) in the range 10 < 2θ < 80 at a scan rate of 0.05
every 2.4 s. The crystal phases were confirmed according to
JCPDS reference while the crystallite sizes were calculated on
the basis of Scherrer equation after Warren’s correction.
Temperature programmed desorption (TPD) of O2, NO, and
CO were performed in a fixed-bed reactor with the effluent gases
monitored using a quadrupole mass spectrometer (MS) (UTI
100). Prior to TPD of O2, NO, and CO experiments, 50 mg sam-
ples were treated with the various atmospheres of 10% O2 or
3000 ppm NO or 3000 ppm CO balanced by helium, respec-
tively, with a total flow rate of 20 cm3/min at 500 ◦C for 1 h,
then cooled down to room temperature under the same atmo-
sphere, andsubsequentlyflushedby20 cm3/minHefor40 minto
remove the physisorbed molecules. Finally, the reactor temper-
ature was raised up to 500 ◦C (800 ◦C for O2-TPD) at a constant
heating rate of 10 ◦C/min. O2, NO, CO and CO2 desorbed during
TPD experiments were simultaneously monitored by MS with
the mass numbers of 32, 30, 28 and 44, respectively.
Temperature programmed surface reduction (TPSR) of NO
underCO/Heflowwasperformed withthesamesamplepretreat-
ment as that during TPD of NO experiment. The thermodesorp-
tion was however performed under 1000 ppm CO/He instead of
He. The MS signals recorded in this case were those of mass
numbers 14, 32, 30, 28 and 44. The gaseous response obtained
by MS was calibrated using standard gas mixtures.
The catalyst activity tests of LaMn1−xCuxO3 and LaCo1−x
CuxO3 for the NO + CO reaction were carried out in a plug flow
reactor at atmospheric pressure. A mixture of 3000 ppm NO
and 3000 ppm CO at a total rate of 60 ml/min, corresponding to
a GHSV of 50,000 h−1 was passed through the catalyst bed con-
taining about 100 mg catalyst. In addition, the gas compositions
of 3000 ppm N2O + 3000 ppm CO or 3000 ppm NO or 3000 ppm
N2O or 3000 ppm CO or 3000 ppm CO + 1500 ppm O2 were also
adopted for the reactions of N2O + CO, NO or N2O decomposi-
tion, CO oxidation by perovskite or oxygen, respectively, over
LaCo0.8Cu0.2O3 under the same space velocity. The system was
heated externally in a tubular furnace, regulated by an Omega
controller in the temperature region of 100–500 ◦C via a thermo-
couple in contact with the catalyst bed. NO and NO2 in effluent
were analyzed by chemiluminescence NO/NO2/NOx analyzer
(Model 200AH, API Inc.). N2O and CO gases were monitored
using a FT-IR gas analyzer (FTLA 2000, ABB Inc.). N2 was
detected using a gas chromatograph (GC) (Hewlett Packard
5890) equipped with a thermal conductivity detector (TCD) and
columns of molecular 13X (45–60 mesh, 2.4 m) combined with
Silicone OV-101 (100–120 mesh, 0.6 m). Calibration was done
using standard gases containing known concentrations of the
components.
2. Experimental
Two series of LaCo1−xCuxO3 and LaMn1−xCuxO3 solid
solutions were prepared by reactive grinding. Powders of La2O3
(Alfa, 99.99%), CuO (Aldrich, 99.98%), and Co3O4 (Baker &
Adamson, 97.49%) or Mn3O4 (Baker & Adamson, 98.22%)
in desired proportions were milled under high energy, using a
SPEX grinder rotated at 1100 rpm, equipped with tempered steel
vial and three tempered steel balls. The La2O3 powder was pre-
liminarily calcined at 600 ◦C for 24 h in order to eliminate any
trace of lanthanum hydroxide. ZnO was added as a leachable
additive during grinding to further enhance the specific surface
areas of the products.
The specific surface area and pore size distribution of mixed
oxides were determinated by N2 adsorption at liquid nitro-
gen temperature from the linear part of the BET curve (P/
P0 = 0.01–0.10) and from the desorption branch of N2 isotherms
using the Barrett–Joyner–Halenda (BJH) formula, respectively.
Samples of about 200 mg were evacuated at 300 ◦C for 6 h prior
to the adsorption/desorption experiments.
The elemental analysis of the end products (Co, Mn, Cu) and
the residual impurities (Zn, Fe) was made by atomic absorption
spectroscopy (AAS) (1100B, Perkin-Elmer) after dissolution of
samples in a mixture of 10% HCl solution and concentrated HF.
The analysis of La was accomplished by inductively coupled
plasma (ICP) spectrometry (Optima 4300DV, Perkin-Elmer).
3. Results
3.1. Physicochemical properties
The chemical composition of the prepared oxides was deter-
mined by AAS and ICP showing values close to the nominal

24
R. Zhang et al. / Journal of Molecular Catalysis A: Chemical 258 (2006) 22–34
Table 1
Properties of LaCo1−xCuxO3 and LaMn1−xCuxO3 mixed oxides after calcination at 500 ◦ for 5 h
Sample
Chemical compositiona
Specific surface
area (m2/g)
Pore diameter
(nm)
Crystallite
size (nm)
Crystal phase
Perovskite structure
b
LaCoO3
La1.05Co1.0O3
La1.0Co0.86Cu0.14O3
La0.97Co0.76Cu0.24O3
La0.98Mn1.0O3
La0.97Mn0.91Cu0.09O3
La1.1Mn0.78Cu0.22O3
29.2
23.8
22.4
40.6
40.7
42.6
18.1
15.6
14.7
10.8
12.4
10.8
11.8
10.8
9.2
11.3
13.0
14.4
LaCoO3/La2O2CO3b/Co3O4
Rhombohedral
Rhombohedral
Rhombohedral
Rhombohedral
Rhombohedral
Rhombohedral
δ
b
LaCo0.9Cu0.1O3
LaCo0.8Cu0.2O3
LaMnO3
LaMn0.9Cu0.1O3
LaMn0.8Cu0.2O3
LaCoO3/Co3O4
LaCoO3/Co3O4
δ
b
δ
b
LaMnO3/La2O2CO3
δ
LaMnO3
LaMnO3/La2O2CO3b/CuOb
δ
δ
a
Fe and Zn can be detected as contaminants with a combined wt.% less than 2% of the total weight of prepared La(Co, Mn)1−xCuxO3 samples.
Minor phases.
b
formulation (Table 1). The BET surface areas and pore diam-
desorption curves using Lorentzian peak shape and reported in
Table 2.
eters of these materials synthesized by reactive grinding after
calcination at 500 ◦C for 5 h are also reported in Table 1, rep-
resenting specific surface areas of 20–30 m2/g for cobalt lan-
thanates and approximately 40 m2/g for manganese lanthanates
even after calcination. Higher surface areas and smaller pore
diameters were observed for LaMn1−xCuxO3 oxides compared
to LaCo1−xCuxO3 ones.
Two kinds of desorption peaks are present in the O2-TPD
spectrum of LaCoO3: the broad one appearing in the range
of 150–720 ◦C was ascribed to ␣-O2 desorption and related
to molecular O2 adsorbed on oxygen vacancies [27]. The
latter peak with a maximum at 797 ◦C was designated as ␤-O2
desorption and attributed to the liberation of O2 from the lattice
[27]. With the partial substitution of Co by Cu, the desorption of
␣-O2 is significantly enhanced in proportion to the substitution
degree. This result indicates that more oxygen vacancies can be
generated over Co-based perovskites after Cu incorporation into
B-site of lattice likely due to a positive charge compensation
in agreement with the previous report [30]. Additionally, the
␤-O2 desorption of Cu-substituted lanthanum cobaltites can
take place at relatively lower temperature (with the former
ones at about 765 ◦C and the latter ones at about 795 ◦C) with
respect to LaCoO3 suggesting the mobility of lattice O2 was
also improved via Cu substitution.
The crystallite phase identification of solid solutions was
accomplished by means of XRD with the results being given in
Table 1. The comparison of experimental spectra with JCPDS
charts indicates that all Co-containing samples are essentially
rhombohedral perovskite-type mixed oxides (JCPDS card 48-
0123). Other minor phases, such as Co3O4 (JCPDS card 42-
1467) in the case of all Co-based catalysts, and La2O2CO3
(JCPDS card 37-0804) in the case of LaCoO3, were detected
in addition to the major ABO3 perovskite phase. The XRD lines
of Mn-containing samples also mainly correspond to the rhom-
bohedral perovskite structure and with the formula LaMnO3.15
for the unsubstituted lanthanum manganite (JCPDS card 50-
0298). Minor phases, such as CuO (JCPDS card 80-1268) in the
case of LaMn0.8Cu0.2O3 and La2O2CO3 (JCPDS card 37-0804)
in the case of LaMnO3 and LaMn0.8Cu0.2O3, were detected
as crystalline impurities. The crystallite sizes of all prepared
perovskites were subsequently calculated by Scherrer equation
and listed in Table 1 with values (9–15 nm) close to those of
their pore diameters (10–18 nm) determined from N2 desorp-
tion isotherms. It is thus inferred that the present samples had a
somewhat porous structure formed via clustering of individual
primary particles in agreement with reference [29].
A minor ␣-O2 peak at T < 450 ◦C was found in the O2-TPD
profile for LaMnO3 associated with a slight enhancement in its
intensity upon increasing Cu content. A following intense des-
orption with a maximum at 608 ◦C was observed over lanthanum
manganite, showing a drop in its intensity after Cu substitu-
tion. The introduction of Cu ions in LaMnO3 perovskite lattice
3.2. Temperature programmed desorption of oxygen
(O2-TPD)
Due to the distinct oxygen nonstoichiometry of lanthanum
cobaltites (oxygen deficiency) and lanthanum manganites (oxy-
gen excess), the oxygen desorption behaviors over the above
two series of perovskite-type oxides were widely studied and
well documented to correlate nonstoichiometric oxygen of per-
ovskites with their catalytic characteristics in redox reactions
[30]. The O2 species adsorbed over Co- and Mn-based per-
ovskites generated by reactive grinding were thus investigated
via O2-TPD experiments as depicted in Fig. 1. The amounts
of O2 released were calculated after deconvolution of the O2
Fig. 1. TPD of O2 profiles of LaCo1−xCuxO3 and LaMn1−xCuxO3 perovskites
[15,16].

R. Zhang et al. / Journal of Molecular Catalysis A: Chemical 258 (2006) 22–34
25
Table 2
Amounts of O2 desorbed from LaCo1−xCuxO3 and LaMn1−xCuxO3 perovskites during O2-TPD experiments [15,16]
Sample
Amount of oxygen desorbeda
Number of monolayers desorbedb
␣-O2 (␮mol g−1
)
Excess oxygen (␮mol g−1
)
␤-O2 (␮mol g−1
)
␣-O2
Excess oxygen
␤-O2
c
LaCoO3
133.2 (<720 ◦C)
252.2 (<720◦ C)
285.6 (<720 ◦C)
16.8 (324 ◦C)
22.6 (323 ◦C)
40.1 (301 ◦C)
–
–
–
112.4 (720–800 ◦C)
163.6 (720–800 ◦C)
171.6 (720–800 ◦C)
–
155.7 (760 ◦C)
196.1 (763 ◦C)
1.14
2.65
3.19
0.10
0.14
0.23
–
–
–
1.95
1.16
0.62
0.96
1.72
1.91
–
0.96
1.15
LaCo0.9Cu0.1O3
LaCo0.8Cu0.2O3
LaMnO3
LaMn0.9Cu0.1O3
LaMn0.8Cu0.2O3
316.7c (609 ◦C)
189.3c (646 ◦C)
106.8c (599 ◦C)
a
Calculated by deconvolution of the O2 desorption curves.
b
c
Calculated with 4 ␮mol m−2 of oxygen per monolayer (Ref. [27]).
The excess oxygen calculated according to the second O2 desorption peak under the region of 400–650 ◦C.
was reported to suppress its overstoichiometric (excess) oxy-
gen [5]. This peak is therefore associated with the desorption of
overstoichiometric oxygen of lanthanum manganite. This over-
stoichiometric oxygen desorbed in parent LaMnO3 was quanti-
fied and found to be consistent with the formula of LaMnO3.15
determined by XRD. This can be interpreted as another evi-
dence supporting the ascription of overstoichiometric oxygen
in O2-TPD profiles. The ␤-O2 desorption from Mn-based per-
ovskites are believed to correspond to those peaks appearing at
temperatures above 700 ◦C. It seems that the lattice O2 mobil-
ity of lanthanum manganite is improved again via substituting
B-site ion by Cu according to the results reported in Table 2.
The promotion in the mobility of lattice oxygen of lanthanum
cobaltite and manganite after Cu B-site substitution had also
been observed in our early H2-TPR studies [15,16]. The trace
amounts of ␣-O2 desorbing from LaMn0.8Cu0.2O3 suggest that
Cu substitution has indeed produced oxygen vacancies, which
can coexist with the overstoichiometric oxygen in the lattice.
However, this ␣-O2 desorption is severely suppressed compared
to LaCo1−xCuxO3 which is associated with the presence of over-
stoichiometric oxygen.
Fig. 2. MS signal of NO during TPD of NO of LaCo1−xCuxO3 and LaMn1−x
CuxO3 perovskites.
at 200 ◦C with one minor shoulder at 90 ◦C and another at
388 ◦C was observed in the TPD of NO profiles for LaCoO3.
Upon Cu substitution the low temperature peak is entirely sup-
pressed, the 200 ◦C one is substantially increased and shifted
to higher temperature so that it interferes with the minor high
temperature one. Similar NO desorption features were found for
LaMnO3 after NO adsorption, representing three superimposed
desorption peaks with maxima at 100, 202 and 311 ◦C, respec-
tively. With the increase of Cu percentage in LaMn1−xCuxO3,
the low temperature peak is totally suppressed whereas the high
temperature one grows and is slightly shifted towards higher
3.3. Temperature programmed desorption of NO (NO-TPD)
The MS signal of NO (m/e = 30) [along with O2 (m/e = 32)]
as function of desorption temperature was recorded during NO-
TPD under He flow for LaCo1−xCuxO3 and LaMn1−xCuxO3
perovskites as reported in Fig. 2. The amount of NO desorbed
from perovskites was further calculated based on deconvolu-
tion of these curves (Table 3). A broad NO desorption centered
Table 3
Amounts of NO desorbed from LaCo1−xCuxO3 and LaMn1−xCuxO3 perovskites during NO-TPD experiments
Physisorbed NOa (␮mol g−1
)
Nitrosyl speciesa (␮mol g−1
)
Nitrate speciesa (␮mol g−1
)
Chemisorbed NO
(␮mol g−1
)
(␮mol m−2
)
b
LaCoO3
2.4 (90 ◦C)
–
–
8.8 (100 ◦C)
5.7 (106 ◦C)
–
27.6 (201 ◦C)
39.9 (255 ◦C)
45.7 (283 ◦C)
15.9 (202 ◦C)
23.5 (210 ◦C)
31.5 (209 ◦C)
9.4 (388 ◦C)
37.0
39.9
45.7
21.7
29.5
41.2
1.27
1.68
2.04
0.53
0.72
0.96
LaCo0.9Cu0.1O3
LaCo0.8Cu0.2O3
LaMnO3
LaMn0.9Cu0.1O3
LaMn0.8Cu0.2O3
5.8 (311 ◦C)
6.0 (323 ◦C)
9.7 (361 ◦C)
a
Calculated by deconvolution of NO desorption curves.
Calculated with the specific surface areas in Table 1.
b

26
R. Zhang et al. / Journal of Molecular Catalysis A: Chemical 258 (2006) 22–34
Table 4
Amounts of CO and CO2 desorbed from LaCo1−xCuxO3 and LaMn1−xCuxO3 perovskites during CO-TPD experiments
Sample
CO
CO2
Total carbonaceous species
T (◦C)
Amount (␮mol g−1
)
T (◦C)
Amount (␮mol g−1
)
Amount (␮mol g−1
)
LaCoO3
>200
>200
>200
>250
>250
>250
26.6
22.5
18.2
40.5
56.9
70.5
>200
>150
>150
>200
>200
>200
116.4
166.4
195.0
66.7
95.3
104.3
143.0
188.9
213.2
107.2
152.2
174.8
LaCo0.9Cu0.1O3
LaCo0.8Cu0.2O3
LaMnO3
LaMn0.9Cu0.1O3
LaMn0.8Cu0.2O3
temperature. Small O2 desorptions parallel with only the high-
temperature NO shoulder peaks were observed during NO-TPD
studies (not shown).
3.6. Activity tests
A blank test was performed under 60 ml/min of 3000 ppm CO
and 3000 ppm NO balanced by helium in an empty reactor (not
shown). No significant NO conversion (<2%) and CO conver-
sion (<5%) were observed up to 500 ◦C. Hence, the conversions
generated from homogenous reaction in the absence of catalyst
are essentially negligible under usual reaction conditions. Tem-
perature dependencies of N2 yields over Co- and Mn-based per-
ovskites are shown in Fig. 4a. Among the six tested samples, the
worst conversion to N2 is obtained in the case of LaMnO3 with
a progressive increase up to 76% at 500 ◦C. A higher N2 yield
with a maximum of 93% at 500 ◦C was observed over LaCoO3
with respect to LaMnO3. An obvious enhancement in N2 yields
is realized by Cu partial substitution for the two series of per-
ovskites resulting in an approximate order of LaCo0.8Cu0.2O3 >
LaCo0.9Cu0.1O3 > LaMn0.8Cu0.2O3 > LaMn0.9Cu0.1O3. N2O
is detectable in the effluent with maxima in the range of
270–330 ◦C depending on the solid composition as illustrated
in Fig. 4b. More significant N2O production and CO conversion
were found over LaCoO3 compared to those obtained over
LaMnO3 (see Fig. 4b and c). The N2O production and CO
conversion obtained over lanthanum cobaltite and lanthanum
manganite can be improved via Cu substitution. In general,
3.4. Temperature programmed desorption of CO (CO-TPD)
The desorption of CO (m/e = 28) and CO2 (m/e = 44) was
monitored by MS during the CO-TPD experiments for Co- and
Mn-based samples. A quantitative analysis of the various car-
bonaceous gases desorbed from perovskites is summarized in
Table 4. Both CO and CO2 desorptions were observed indicating
that the oxidation of CO into CO2 occurs over those lanthanum
perovskites during CO-TPD experiments, while Cu substitution
leads to more CO2 desorption over lanthanates suggesting a
promotion in CO oxidation by means of copper incorporation.
Simultaneously, it was observed that the amount of total car-
bonaceous species (CO + CO2) desorbed from perovskites also
increased after Cu substitution, suggesting that copper incorpo-
ration can also facilitate the generation of adsorption sites for CO
or CO2. Higher carbonaceous species was found over cobaltites
with respect to manganites under the same Cu substitution per-
centage.
3.5. Temperature programmed surface reduction of NO
(NO-TPSR) under CO/He
Fig. 3 represents the TPSR of adsorbed NO under 1000 ppm
CO/He flow over LaCo0.8Cu0.2O3. The amounts of desorbed NO
andgeneratedN2 werequantifiedaccordingtothedeconvolution
of their MS traces. The oxygen signal detected by MS (m/e = 32)
rapidly diminishes at 100–300 ◦C simultaneously with a con-
sumption of CO, while a recovery of this O2 signal takes place
at T > 300 ◦C. At temperatures lower than 180 ◦C, a NO des-
orption peak (15.6 ␮mol g−1) was observed in Fig. 3 whereas
it is absent in the NO desorption curve in Fig. 2. The NO des-
orption peak (45.7 ␮mol g−1) centered at 283 ◦C and its minor
shoulder for LaCo0.8Cu0.2O3 in Fig. 2 is essentially suppressed
into 3.9 ␮mol g−1 (in Fig. 3), indicating that the corresponding
adsorbed NO species has reacted with CO but that the product
of reaction is not desorbed over the temperature range of this
experiment. Some N2 desorption (3.5 ␮mol g−1) with a maxi-
mum at approximately 150 ◦C was indeed observed but not in
amounts comparable with the content of a desorbed NO reported
for the same sample in Fig. 2. The appearance of CO2 coincides
with the disappearance of CO.
Fig. 3. MS signals during TPSR of NO under CO/He flow of LaCo0.8Cu0.2O3,
conditions: flow rate = 20 ml/min, 3000 ppm NO, 1000 ppm CO, balanced by
He.

R. Zhang et al. / Journal of Molecular Catalysis A: Chemical 258 (2006) 22–34
27
Fig. 5. (a) N2 yield in various reactions over LaCo0.8Cu0.2O3. Conditions:
GHSV = 50,000 h−1, 3000 ppm NO, 3000 ppm N2O, 3000 ppm CO, balanced by
He. (b) CO conversion in various reactions over LaCo0.8Cu0.2O3. Conditions:
GHSV = 50,000 h−1, 3000 ppm NO, 3000 ppm N2O, 3000 ppm CO, 1500 ppm
O2, balanced by He.
the best performance is achieved over LaCo0.8Cu0.2O3 with
94% N2 yield and 99% CO conversion even at 400 ◦C. No
obvious NO2 formation was detected during activity tests of
the prepared perovskites under the atmosphere of 3000 ppm
NO and 3000 ppm CO.
For the best performing LaCo0.8Cu0.2O3, N2 yields in the
reactions of NO + CO, N2O + CO, NO decomposition and N2O
decomposition were investigated and the results are shown
in Fig. 5a with the objective of better understanding the NO
and CO transformation in the catalytic process. In the absence
of reducing agent, a little NO decomposition (<5%) occurs
at 350–500 ◦C. N2O decomposition is much easier than NO
decomposition starting at 300 ◦C and achieving a N2O conver-
sion of 68% at 500 ◦C, possibly related to the easy cleavage of
N–O bond in N2O. Nevertheless, it is noticed that N2O decom-
position does not take place at T < 300 ◦C. The transformation of
nitrogen oxides (NO and N2O) into N2 is significantly improved
Fig. 4. (a) N2 yield in CO + NO reaction over LaCo1−xCuxO3 and LaMn1−x
CuxO3 perovskites, conditions: GHSV = 50,000 h−1, 3000 ppm NO, 3000 ppm
CO, balanced by He. (b) N2O concentration in CO + NO reaction over LaCo1−x
CuxO3 and LaMn1−xCuxO3 perovskites, conditions: GHSV = 50,000 h−1
,
3000 ppm NO, 3000 ppm CO, balanced by He. (c) CO conversion in CO +
NO reaction over LaCo1−xCuxO3 and LaMn1−xCuxO3 perovskites, conditions:
GHSV = 50,000 h−1, 3000 ppm NO, 3000 ppm CO, balanced by He.

28
R. Zhang et al. / Journal of Molecular Catalysis A: Chemical 258 (2006) 22–34
4. Discussion
4.1. O2-TPD study
A broad plateau-like ␣-O2 desorption was observed between
200 and 720 ◦C over the LaCoO3 perovskite. The intensity of
this ␣-O2 desorption is enhanced by Cu substitution because of
a positive charge deficiency after substitution of Co3+ by Cu2+
compensated by oxygen vacancies [27]. This ␣-O2 desorption
was ascribed to O2 adsorbed on anion vacancies, which can be
readily removed at relatively low temperature. Rather poor ␣-O2
coverage of LaMnO3 after O2 adsorption was found with respect
to LaCoO3 although its coverage can be slightly enhanced by
Cu substitution, indicating a low density of anion vacancies over
lanthanum manganites.
According to our previous work [27], the following process
was proposed to occur during calcination of Co- or Mn-based
perovskite:
Fig. 6. Effect of O2 feed concentration on catalytic performance in CO + NO
reaction over LaCo0.8Cu0.2O3, conditions: GHSV = 50,000 h−1, T = 350 ◦C,
3000 ppm NO, 3000 ppm CO, 0–2000 ppm O2, balanced by He.
(1) Transient generation of anion vacancy during calcination
of fresh sample, which seems more pronounced in the Cu
substituted lanthanum cobaltites or manganites.
by the CO reducing agent, resulting in N2 yields of 51% in
NO + CO reaction and 74% in N2O + CO reaction at 300 ◦C and
almost reaching 100% in both reactions at 500 ◦C. The better N2
yield in N2O reduction by CO is ascribed to the easier molecular
activation of N2O compared to NO.
(1)
CO conversions in reactions of NO + CO, N2O + CO, CO oxi-
dation by perovskite (no oxygen), CO + O2 over LaCo0.8Cu0.2
O3 are compared in Fig. 5b. CO oxidation can still carry on
with a maximum conversion of 42% at 400 ◦C in the absence
of O2, which is likely realized by consuming ␣-oxygen from
the perovskite surface. With the consumption of adsorbed O2
species by CO oxidation, a decline in CO oxidation occurs.
Simultaneously, the fast desorption of O2 species at relatively
higher temperature, confirmed by O2-TPD (Fig. 1) and TPSR
(Fig. 3) studies, also leads to this diminution in CO oxida-
tion. Nitrogen oxides (NO and N2O) promote the CO oxi-
dation up to 56 and 71% at 300 ◦C, respectively. N2O is
again more active than NO to donate its oxygen atom and
oxidize CO. Gaseous oxygen with the same atomic O con-
centration as 3000 ppm NO or N2O leads to the highest CO
oxidation achieving a value of 87% even at 300 ◦C. The
above results suggest an order of the efficiency in oxygen
utilization for CO oxidation as: oxygen molecule (O2) or ␣-
(where M=Co or Mn or Cu; n=2 or 3 or 4; ꢀ stands for
anion vacancy).
(2) Instantaneous formation of ␣-O2 by adsorbing O2 at anion
vacancy.
(2)
The two reactions (3) and (4) are proposed for the ␣-O2 des-
orption process:
(3)
(4)
oxygen > oxygen atom (O) in NO or N2O > lattice oxygen (O2−
)
A significant desorption of overstoichiometric oxygen (excess
O2) at 450–650 ◦C was observed over LaMnO3 during O2-TPD
experiments, whose intensity was suppressed by increasing Cu
substitution. This may be formulated as:
or ␤-oxygen.
The effect of O2 feed concentration on catalytic perfor-
mance of LaCo0.8Cu0.2O3 in the NO + CO reaction at 350 ◦C
was investigated with the results shown in Fig. 6. It was
observed that the N2 yield (80%) and N2O yield (19%) dur-
ing NO catalytic reduction in the absence of O2 decrease
rapidly accompanied with the increasing O2 feed concentration.
1200 ppm O2 in feed is sufficient to lead to a total inhibi-
tion in N2 yield resulting in a value below 5%. In contrast,
the CO conversion and NO2 yield are further improved by
gaseous O2 reaching the values of 100 and 12% at 2000 ppm O2,
respectively.
(5)
In Eq. (5) the excess oxygen charge is compensated by cation
vacancies (noted here as (ꢀ)) following the suggestions of van
Roosmalen et al. [31].
The ␤-oxygens in the surface lattice of two series of per-
ovskites can only be desorbed at high temperature (T > 700 ◦C)

R. Zhang et al. / Journal of Molecular Catalysis A: Chemical 258 (2006) 22–34
29
leading to the B-site ion reduction and anion vacancy generation.
upon Cu substitution in the case of lanthanum manganites. The
NO desorption peak related to nitrate species is however totally
immersed in the broad peak attributed to nitrosyl species due
to the shift of the latter to higher temperature at increasing Cu
content in lanthanum cobaltites. It is noticeable that only minor
nitrate species was formed over perovskites after NO adsorption,
showing rather as a shoulder than a peak compared to nitrosyl
species.
(6)
␤-O2 desorption is maintained via the diffusion of lattice oxygen
from the bulk to the surface:
The intensities of nitrosyl species for unsubstituted LaCoO3
and LaMnO3 is enhanced upon Cu substitution. Based on the
fact that more oxygen vacancies can be formed via substituting
trivalent cation (Co3+, Mn3+) by bivalent cation (Cu2+) (due to a
positive charge deficiency confirmed by O2-TPD experiments,
see Fig. 1), this intense NO desorption peak related to nitrosyl
species is likely formed by NO adsorption on anion vacancies as
verified by Shin et al via FT-IR study [36]. The NO chemisorp-
tion as a negatively charged form (NO−) was thought to be
the first important step for NO + CO reaction [37] because of
the back-donation which occurs through antibonding orbitals,
determines a weakening of the N–O bond when the net electron
transfer is from metal to NO. The formation of nitrosyl species
is formulated on the basis of the above discussion.
(7)
This ␤-oxygen desorption is promoted by Cu substitution as
indicated by the results of O2-TPD experiments (Fig. 1) in
accordance with the conclusion that Cu incorporation enhanced
the mobility of lattice O according to our earlier studies of the
reducibilities of LaCo1−xCuxO3 [15] and LaMn1−xCuxO3 [16]
by H2-TPR.
4.2. NO-TPD study
The formation of (mono-, di-) nitrosyl and (monoden-
tate, bidentate, bridging) nitrate species over LaCoO3 [8,32],
LaMnO3 [32], Cu/ZSM-5 [33] and Cu/MCM-41 [14] after
NO adsorption was mentioned in the literature. In this study,
the adsorbed N-containing species formed over prepared
LaCo1−xCuxO3 and LaMn1−xCuxO3 perovskites during NO
adsorption were investigated by means of monitoring the NO
desorption traces (m/e = 30) in the effluents by MS (Fig. 2).
The amounts of the various NO species desorbed were quan-
tified as listed in Table 3. It was observed that broad NO
desorption peaks occur at approximately 200–280 ◦C with
minor shoulders at their rising and falling branches, respec-
tively, during NO-TPD experiments for Co- and Mn-based
perovskites. Centi and Perathoner [12] concluded that the
thermal stability of adsorbed species increased with increas-
ing oxidation state of nitrogen in N-containing adspecies
according to transient catalytic experiments over Cu/ZSM-
5. In our previous TPD of NO + O2 studies over Co-, Mn-
based perovskites, NO peaks appearing at similar temperatures
as those in NO-TPD experiments (apart from a minor des-
orption at medium temperature and an intense one at high
temperature) were also found and ascribed to the desorp-
tions of physisorbed NO, nitrosyl species and nitrate species
depending on their thermal stability [15,16]. The NO peaks
obtained at low (90–110 ◦C), medium (200–210 ◦C), and high
(310–390 ◦C) temperatures in the present NO-TPD analyses
were thus correlated to physisorbed NO, nitrosyl and nitrate
species.
(8)
By taking into account the different specific surface areas of Co-
based and Mn-based perovskites, the total amount of chemically
adsorbed NO species was also quantified in terms of ␮mol m−2
.
This showed more than twice more desorption occurring over
Co-based samples compared to the corresponding Mn-based
samples at the same Cu contents (see Table 3). This tendency
coincides with the abundance of anion vacancies over lanthanum
cobaltites compared to lanthanum manganites.
4.3. CO-TPD study
´
According to the reports of Tascon et al., the inhibiting
effect of NO on subsequent CO adsorption is larger than the
inhibiting effect of CO on NO adsorption over both LaCoO3
and LaMnO3 [38]. It is thus speculated that CO appears to be
more weakly adsorbed than NO at the same site of perovskites.
The CO adsorption thus likely involves an adsorption at anion
vacancies nevertheless via coordinative bond without electron
transfer:
O2 desorption simultaneous with minor NO desorption peaks
occurring at 388 ◦C for LaCoO3 and 311 ◦C for LaMnO3 (not
shown) gives another clue for the ascription of these peaks to
nitrate species. This is in agreement with the report that the des-
orption of NO3− species would appear as a NO desorption peak
at high temperature (T > 300 ◦C) associated with O2 desorption
[15–17,34,35]. The intensity of the nitrate species peak increases
(9)
COisoxidizedintoCO2 overperovskitesduringCO-TPDexper-
iments (Table 4) and this CO oxidation is promoted by Cu
substitution leading to an enhancement in CO2 desorption. It was
reported that CO oxidation happens via a suprafacial catalytic
process involving a surface lattice atomic oxygen (O−) [39]. The

30
R. Zhang et al. / Journal of Molecular Catalysis A: Chemical 258 (2006) 22–34
CO oxidation therefore was proposed as the following step:
significant surface coverage of carbonate species, giving another
explanation for this phenomenon.
N2 production was also detected during this experiment
implying that the reduction of nitrosyl species by CO is associ-
ated with the catalytic process.
(10)
Furthermore, the lower binding energy for this surface lattice
oxygen favors CO oxidation [40]. The improvement of CO
oxidation after Cu substitution can therefore be explained by an
enhanced mobility of perovskite lattice oxygen as found in the
present O2-TPD experiments and in previous H2-TPR studies
[15,16].
4.5. Activity test and reaction mechanism
The direct decomposition of NO would represent the most
attractive solution in NO emission control, because the reaction
does not require that any reactant be added to NO exhaust gas
and could potentially lead to the formation of only N2 and
O2. In principle, the simple NO decomposition is a thermody-
namically favored reaction at low temperature. However, this
reaction is very slow due to the high dissociation energy of NO
(153.3 kcal mol−1) and no sufficiently efficient catalysts have
been discovered so far [43]. Up to now Cu-zeolites are the best
catalysts with isolated Cu+ ions regarded as active centres in
NO decomposition [13]. Perovskites exhibit an activity better
(11)
In the presence of gaseous O2, this CO oxidation can also
be accelerated by ␣-O2 over perovskite [40] expressed as the
following reaction.
(12)
4.4. TPSR of NO under CO/He flow
than most of other metal oxides for this reaction, but at 500 ◦C
they are much less active than the best copper zeolites [44–46].
That NO hardly decomposes over LaCo0.8Cu0.2O3 perovskite at
<500 ◦C was confirmed again in this work (Fig. 5a).
It was observed in Fig. 3 that the NO desorption peaks related
to nitrosyl and nitrate species (observed in NO-TPD profile of
LaCo0.8Cu0.2O3 in Fig. 2) almost vanished. Only the desorp-
tion peak appearing at low temperature, which was ascribed to
physisorbed NO, is present in the NO-TPSR trace. This result
suggests that both nitrosyl and nitrate species are quite active
toward CO and fully consumed by this reducing agent in agree-
ment with the previous conclusion that the chemisorbed NO
plays an important role in NO + CO reaction catalyzed by per-
ovskites [41]. Nitrate species were also proven to be highly reac-
tive towards propene over perovskites [15–17], towards methane
over Ag/Al2O3 [34] and Ag-ZSM-5 [42] and resulted in a good
deNOx activity at high temperature. The reactivity of this nitrate
species towards CO is confirmed again in the present TPSR
study. Nevertheless, the nitrate species formed over perovskites
in the absence of gas phase oxygen is minor compared to nitrosyl
species according to NO-TPD study. Furthermore, NO reduc-
tion is already remarkable at T < 300 ◦C (see Figs. 4a and b)
whereas the nitrate species seems to be stable under this tem-
peraturerangeandonlycorrespondstohightemperatureactivity.
Nitrosyl species are therefore believed to make a most signifi-
cant contribution to NO reduction by CO and their role will be
emphasized in this study.
Nitrous oxide is not toxic but plays a role in ozone depletion.
N2O decomposition over LaCo0.8Cu0.2O3 was thereafter exam-
ined with the result in Fig. 5a showing a better activity in terms
of N2 yield (68% at 500 ◦C) than that achieved in NO decompo-
sition (4% at 500 ◦C). N2O is known to have a linear structure
N–N–O in which the N–N bond length is somewhat longer than
˚
that in triply bonded N2 (1.09 A) and in which the N–O distance
˚
is increased by about 0.03 A relative to free nitric oxide with a
bond order of 2.5, resulting in an easy cleavage of its N–O bond
and better dissociation behavior. However, enough energy is still
necessary for N–O bond breakage according to the result that
N2O was only decomposed as temperatures in excess of 300 ◦C
(Fig. 5a).
With the assistance of CO, N2 yields for both NO and
N2O transformations were significantly improved up to 80%
in NO + CO reaction and 91% in N2O + CO reaction even at
350 ◦C. Chien [41] et al. observed a higher NO adsorption
rate for actived (reduced) than for unactivated LaCoO3 and
La0.85Ba0.15CoO3. A reduced perovskite surface with a large
amount of anion vacancies was reported to be crucial for the
NO dissociation according to FTIR and EPR spectroscopic anal-
yses [8]. The role of CO seems therefore to maintain a reduced
surface, which is necessary for successive NO dissociation. CO
oxidation is also promoted by the O atom coming from NO or
N2O (Fig. 5b). From the above results it seems that these tested
perovskites act as “oxygen reservoir” transferring oxygen atoms
from nitrogen oxides to CO thus achieving NO reduction and CO
The MS signal of O2 decreased progressively in the range of
100–300 ◦C simultaneously with the CO consumption and CO2
formation due to a rapid CO oxidation (Fig. 3). Nevertheless, a
recovery of O2 signal was observed at T > 300 ◦C. This is likely
related to a fast desorption of O2 species at relatively high tem-
perature. The O2 consumption by CO to CO2 oxidation is also
slightly slowed down (see CO2 trace in Fig. 3) possibly due to a

R. Zhang et al. / Journal of Molecular Catalysis A: Chemical 258 (2006) 22–34
31
oxidation simultaneously due to their outstanding redox proper-
ties.
The high reactivity of nitrosyl species toward CO was estab-
lished via TPSR of NO under CO/He flow test. N2O formation
was observed with a maximum of 346 ppm at 300 ◦C and further
decreased at higher temperature (Fig. 4b). Based on our previ-
ousstudiesandpresentwork, thecatalyticmechanism(including
the adsorption and transformation of reactants over perovskite,
redox catalytic recycle, and the product components dependence
on reaction temperature) is proposed to involve the following
steps:
At low temperature, the dissociation of adsorbed NO species
occurring over reduced perovskite and yielding N2O and N2 was
recognized as the rate determining step for catalytic reduction
of NO by CO. The dimeric species of NO, such as N2O2, can
be an intermediate, the formation of which involves the N–N
bond formation and N–O bond cleavage [47,48]. Two paral-
lel reactions for chemisorbed NO dissociation occurring over
a reduced surface with N2O and N2 as the respective products
were assumed:
Fig. 7. Molar ratios of XNO/XCO and [CO2]/([N2] + [N2O]) in CO + NO reac-
tion over Co-based perovskites, conditions: GHSV = 50,000 h−1, 3000 ppm NO,
3000 ppm CO, balanced by He.
The ratio of NO conversion to CO conversion must be equal to
unity for Route (1) and two for Route (2), while the molar ratio
CO2/(N2 + N2O) in the products should be two for Route (1) and
one for Route (2) according to Eqs. (15) and (16). Therefore,
the plots of XNO/XCO as well as [CO2]/([N2] + [N2O]) versus
temperature provide an easy way to probe the reaction route
involved. The XNO/XCO values were found to be close to 1.5 at
T = 200 ◦C and 1 at T > 400 ◦C for lanthanum cobaltites. More-
over, the ratios of [CO2]/([N2] + [N2O]) for these catalysts are
close to 1.3 at T = 200 ◦C and 2 at T > 400 ◦C as depicted in
Fig. 7.
The reaction between CO and N2O (Route (4)) as well as
the decomposition of N2O (Route (3)) are accelerated by rising
temperature as depicted in Fig. 5a. As a result, all N2O generated
duringNOdissociationcanberapidlyconsumedandthereaction
occurring at high temperature can be simplified as Route (1) with
XNO/XCO = 1 and [CO2]/([N2] + [N2O]) = 2.
(13)
(14)
Reaction (14) involves the breakage of one N–O bond making
this reaction much easier than reaction (13) which results in two
N–O bond cleavages. This assumption is in agreement with the
activity results shown in Fig. 4b, indicating that reaction (14) is
dominating at low temperature. CO can be oxidized by Mn+O−
species as illustrated in Eq. (10) together with regeneration of
anion vacancies on the surface for a continuous NO dissociation.
The reaction routes occurring at low temperature can thus be
described as two parallel reactions:
The kinetic analysis may be interpreted as another evidence
supporting a mechanism dominated by Route (2) at low temper-
ature and then gradually progressing into Route (1) at elevated
temperature.
The higher conversion to N2 obtained over LaCoO3 com-
pared to LaMnO3 despite its lower surface area (Table 1)
was ascribed to its higher surface density of anion vacan-
cies (Table 2), which facilitates the adsorption of reactants
(Tables 3 and 4) and further dissociation of chemisorbed NO
(reactions (13) and (14)). Consequently, a significant improve-
ment in N2 yield was achieved after Cu incorporation into the
B-sites of lanthanum cobaltite. Besides the essential effect of
copper ions in the transformation of nitrogen oxides [12], Cu
substitution can promote the formation of anion vacancies. At
the same time, the regeneration of anion vacancies (see reac-
tion (11)) is also accelerated after Cu substitution due to the
enhancement in the mobility of lattice oxygen.
Route(1) : 2NO + 2CO → N2 + 2CO2
Route(2) : 2NO + CO → N2O + CO2(fast)
(15)
(16)
Although N2O was detected in the effluent at low tempera-
ture (see Fig. 4b), its further conversion as an intermediate is
not excluded [8,49,50]. Pomonis proposed a partially succes-
sive N2O decomposition (Route (3): 2N2O → 2N2 + O2) occur-
ring at low temperature simultaneously with reaction (16) over
La1−xSrxFeO3 [7,51,52] and LaMnO3 [52]. However, the N2O
decomposition over the present perovskite was found to be quite
difficultatT < 300 ◦C(seelineforN2OdecompositioninFig. 5a)
and is rather believed to occur only over a reduced perovskite
with the participation of CO to consume surface oxygen species.
Theoverstoichiometric(excess)oxygenpresentinlanthanum
manganite is also believed to participate in the oxidation of CO
but without the generation of any oxygen vacancies, so that the
Route(4) : N2O + CO → N2 + CO2
(17)

32
R. Zhang et al. / Journal of Molecular Catalysis A: Chemical 258 (2006) 22–34
Scheme1. OxygentransformationbehaviorintheprocessofNOcatalyticreduc-
tion by CO.
key process for a better activity in NO reduction. These anion
vacancies can capture the oxygen atom from nitrogen oxides
(NO or N2O) during NO dissociation to form surface lattice
O−. The formation of ␣-oxygen (O2−), which can further con-
vert into adsorbed O22− or O− after getting one electron [48,53],
likely takes place over perovskite when the reducing agent (CO)
is insufficient in feed. Both ␣-oxygen and surface lattice O−
may participate in the CO oxidation producing CO2 and oxygen
vacancies recovery for subsequent NO dissociation. Our previ-
ous studies of oxygen mobility in LaCoO3 perovskites [53,54]
showed that the adsorbed O2 could exchange with the surface,
and the exchanged atom could diffuse into the grain boundaries
and into the bulk of the solid. Oxygen transfer in NO catalytic
reduction by CO is illustrated as Scheme 1: the transfer of oxy-
gen from NO to CO involving the generation of surface lattice
oxygen atoms due to NO dissociation (Eqs. (8), (13) and (14)),
and reduction of surface lattice O− into oxygen vacancies by
CO (Eq. (10)). The high catalytic activity (especially at low tem-
perature) achieved over lanthanum cobaltite was ascribed to its
large density of surface oxygen vacancies providing more active
sites for reactant adsorption and reaction. Cu substitution further
promotes the catalytic performance through not only increasing
oxygen vacancies but also enhancing the lattice oxygen mobility.
Fig. 8. Effect of O2 feed concentration on the molar ratios of XNO/XCO and
[CO2]/([N2] + [N2O]) in CO + NO reaction over LaCo0.8Cu0.2O3, conditions:
GHSV = 50,000 h−1, T = 350 ◦C, 3000 ppm NO, 3000 ppm CO, 0–2000 ppm O2,
balanced by He.
oxygen vacancies abundant in cobalt lanthanate seem to make
a main contribution in NO + CO reaction compared with the
overstoichiometric oxygen of manganese lanthanate.
4.6. Effect of O2 feed concentration on the catalytic
behavior of LaCo0.8Cu0.2O3
A serious inhibiting effect of O2 was observed in Fig. 6 dur-
ing the reaction of 3000 ppm NO and 3000 ppm CO at 350 ◦C
over LaCo0.8Cu0.2O3. The corresponding ratios of XNO/XCO and
[CO2]/([N2] + [N2O]) as a function of O2 feed concentration are
plotted in Fig. 8, showing that XNO/XCO values diminish rapidly
upon increasing O2 concentration with values much lower than
those for route (1) and route (2) at O2 partial pressure above
800 ppm. The detrimental effect of oxygen became pronounced
at O2 concentration above 800 ppm, reflecting the fact that the
reaction was totally controlled by the unfavorable oxidation of
the reducing agent (Eq. (12)) leading to a far higher CO conver-
sion compared to NO conversion. NO2 is the main N-containing
product under these conditions.
5. Conclusion
Nanocrystalline perovskite-type oxides were prepared by
reactive grinding with specific surface areas of 20–30 m2/g for
Co-based perovskites and approximately 40 m2/g for Mn-based
perovskites, even after calcination at 500 ◦C for 5 h. The some-
what porous structure of these materials was achieved via clus-
tering of individual primary particles.
A broad ␣-O2 desorption was observed during O2-TPD for
lanthanum cobaltite, indicating the abundance of oxygen vacan-
cies present over the surface of this perovskite. More oxygen
vacancies were generated after partial substitution by Cu2+ due
to a positive charge deficiency. These anion vacancies seem to
play an important role in catalytic reduction of NO providing
the adsorption sites for NO and CO as well as promoting the
dissociation of adsorbed NO species. In contrast, a significant
desorption of overstoichiometric oxygen from lanthanum man-
ganite was found, which was suppressed by Cu incorporation
into lattice B-sites. The role of this excess oxygen seems to just
4.7. The oxygen transfer in NO catalytic reduction by CO
process
The interconversion of gaseous NO and CO due to homoge-
nous reaction is essentially negligible under usual reaction con-
ditions, whereas a good catalytic performance in NO + CO reac-
tion was achieved by catalysts, suggesting that oxygen transfer
between NO and CO seems to occur indirectly and involves
a catalytic process realized via perovskites. All of the atomic
oxygens in perovskites, NO, N2O, and O2 can be used for CO
oxidation(Fig. 5b). The oxygenspecies(␣-, and/oroverstoichio-
metric, and ␤-oxygens) of Co- and Mn-based perovskites were
also investigated by O2-TPD experiments (Fig. 1) and described
in the discussion part of this paper. Based on the catalytic mech-
anism proposed here, the generation of oxygen vacancies is the

R. Zhang et al. / Journal of Molecular Catalysis A: Chemical 258 (2006) 22–34
33
accelerate the oxidation of the reducing agent, with less contri-
bution in NO reduction since no anion vacancies are generated
during CO oxidation by this overstoichiometric oxygen.
Cu partial substitution can also enhance not only the abun-
dance of anion vacancies but also the mobility of lattice oxygen
in lanthanates to accelerate the oxidation of CO by surface lattice
oxygen with regeneration of the anion vacancies. Accordingly,
a significant improvement in catalytic performance of NO + CO
reaction was achieved by introduction of Cu ions into the B-sites
of perovskites.
[5] L. Lisi, G. Bagnasco, P. Ciambelli, S.D. Rossi, P. Porta, G. Russo, M. Turco,
J. Solid State Chem. 146 (1999) 176–183.
[6] V.C. Belessi, C.N. Costa, T.V. Bakas, T. Anastasiadou, P.J. Pomonis, A.M.
Efstathiou, Catal. Today 59 (2000) 347–363.
[7] A.A. Leontiou, A.K. Ladavos, P.J. Pomonis, Appl. Catal. A: Gen. 241
(2003) 133–141.
[8] L. Forni, C. Oliva, T. Barzetti, E. Selli, A.M. Ezerets, A.V. Vishniakov,
Appl. Catal. B: Environ. 13 (1997) 35–43.
[9] V.C. Belessi, T.V. Bakas, C.N. Costa, A.M. Efstathiou, P.J. Pomonis, Appl.
Catal. B: Environ. 28 (2000) 13–28.
[10] A.A. Leontiou, A.K. Ladavos, G.S. Armatas, P.N. Trikalitis, P.J. Pomonis,
Appl. Catal. A: Gen. 263 (2004) 227–239.
At low temperature, the chemisorbed NO (nitrosyl species)
was dissociated over perovskites with the formation of both N2
and N2O as well as an oxidized surface. The oxidized surface
can be reduced by CO in these conditions with regeneration
of anion vacancies for the continuous reaction. N2O was further
reducedbyCOandthistransformationbecomesmoresignificant
as temperature increases.
NO decomposition is difficult over LaCo0.8Cu0.2O3 as tem-
perature below 500 ◦C. In contrast, N2O can be decomposed
at T > 300 ◦C, achieving a N2 yield of 68% at 500 ◦C and
50,000 h−1 GHSV. Nevertheless, both NO and N2O transfor-
mations are obviously promoted in the presence of CO.
O2 strongly inhibits N2 yield by suppressing the reducing
agent through complete CO oxidation. As a result the total
NO + CO reaction in the presence of O2 was found to be severely
limited by CO oxidation.
Inpreviousworks[15,16]wehavestudiedthecatalyticreduc-
tion of NO by propylene over the same catalysts. In these studies
we have found that the main NO reduction process involved
surface nitrate species NO3−. This is in contrast with the con-
clusionsofthepresentwork, whichconsidersthenitrosylspecies
NO− as the main adsorbed reactant (Eqs. (13) and (14)). The
difference between the two systems is due to the higher effi-
ciency of CO (compared to propene) in reducing the perovskite
surface (Eqs. (10)–(12)).
[11] S.D. Peter, E. Garbowski, V. Perrichon, B. Pommier, M. Primet, Appl.
Catal. A: Gen. 205 (2001) 147–158.
[12] G. Centi, S. Perathoner, Appl. Catal. A: Gen. 132 (1995) 179–259.
[13] H. Yahiro, M. Iwamoto, Appl. Catal. A: Gen. 222 (2001) 163–181.
[14] Y. Wan, J.X. Ma, Z. Wang, W. Zhou, S. Kaliaguine, J. Catal. 227 (2004)
242–252.
[15] R.D. Zhang, A. Villanueva, H. Alamdari, S. Kaliaguine, Appl. Catal. B:
Environ. 64 (2006) 220–233.
[16] R.D. Zhang, A. Villanueva, H. Alamdari, S. Kaliaguine, Appl. Catal. A:
Gen. 307 (2006) 85.
[17] R.D. Zhang, A. Villanueva, H. Alamdari, S. Kaliaguine, J. Catal. 237 (2006)
368–380.
[18] M. Crespin, W.K. Hall, J. Catal. 69 (1981) 359–370.
[19] A.E. Giannakas, T.C. Vaimakis, A.K. Ladavos, P.N. Trikalitis, P.J. Pomonis,
J. Colloid Interface Sci. 259 (2003) 244–253.
[20] X. Li, H.B. Zhang, X.X. Liu, S.J. Li, M.Y. Zhao, Mater. Chem. Phys. 38
(1994) 355–362.
[21] C. Xiulan, L. Yuan, Chem. Eng. J. 78 (2001) 205–209.
[22] G. Sinquin, C. Petit, J.P. Hindermann, A. Kiennemann, Catal. Today 70
(2001) 183–196.
[23] R.J.H. Voorhoeve, D.W. Johnson Jr., J.P. Remeika, P.K. Gallagher, Science
195 (1977) 827–833.
[24] J. Kirchnerova, D. Klvana, Int. J. Hydrogen Energy 19 (1994) 501–506.
[25] R. Leanza, I. Rossetti, L. Fabbrini, C. Oliva, L. Forni, Appl. Catal. B:
Environ. 28 (2000) 55–64.
[26] S. Kaliaguine, A. van Neste, (Process For Synthesizing Perovskites Using
High Energy Milling, USA), US Patent US 6017504 (2000) 10 pp.
[27] S. Kaliaguine, A. van Neste, V. Szabo, J.E. Gallot, M. Bassir, R. Muzychuk,
Appl. Catal. A: Gen. 209 (2001) 345–358.
[28] V. Szabo, M. Bassir, A. Van Neste, S. Kaliaguine, Appl. Catal. B: Environ.
43 (2003) 81–92.
[29] S. Royer, A. van Neste, R. Davidson, S. McIntyre, S. Kaliaguine, Ind. Eng.
Chem. Res. 43 (2004) 5670–5680.
[30] P. Porta, S.D. Rossi, M. Faticanti, G. Minelli, I. Pettiti, L. Lisi, M. Turco,
J. Solid State Chem. 146 (1999) 291–304.
[31] J.A.M. van Roosmalen, E.H.P. Cordfunke, R.B. Helmholdt, H.W. Zand-
bergen, J. Solid State Chem. 110 (1994) 100–105.
[32] L.A. Isupova, A.A. Budneva, E.A. Paukshtis, V.A. Sadykov, J. Mol. Catal.
A: Chem. 158 (2000) 275–280.
In the presence of propene and oxygen in the gas phase, the
surface concentration of O2− species is higher (compared to CO
−
alone) and therefore the NO3 surface density is also higher. It
is this species, which is then reduced by propene. The nitrosyl
species are in lower concentration (again compared to CO in the
gas phase) and do not react with propene.
Thus the NO catalytic reduction is entirely different with
these two reducing agents.
[33] G. Centi, Mater. Eng. 5 (1994) 79–85.
[34] X. She, M. Flytzani-Stephanopoulos, J. Catal. 237 (2006) 79–93.
[35] B. Coq, D. Tachon, F. Figue´ras, G. Mabilon, M. Prigent, Appl. Catal. B:
Environ. 6 (1995) 271–289.
Acknowledgements
[36] S. Shin, H. Arakawa, Y. Hatakeyama, K. Ogawa, K. Shimomura, Mater.
Res. Bull. 14 (1979) 633–639.
[37] V.A. Matyshak, O.V. Krylov, Catal. Today 25 (1995) 1–87.
[38] J.M.D. Tasco´n, L.G. Tejuca, C.H. Rochester, J. Catal. 95 (1985) 558–
566.
The financial support of NSERC through its industrial chair
program is gratefully acknowledged. We thank Nanox Inc. for
the preparation of the perovskite samples.
[39] R.J.H. Voorhoeve, Advanced Materials in Catalysis, New York, 1977. p.
129.
[40] B. Viswanathan, Catal. Rev. -Sci. Eng. 34 (4) (1992) 337–354.
[41] M.W. Chien, I.M. Pearson, K. Nobe, Ind. Eng. Chem., Prod. Res. Dev. 14
(1975) 131–134.
[42] C. Shi, M. Cheng, Z. Qu, X. Yang, X. Bao, Appl. Catal. B: Environ. 36
(2002) 173–182.
[43] V.I. Paˆrvulescu, P. Grange, B. Delmon, Catal. Today 46 (1998) 233–316.
References
[1] W.P.A. Jansen, J.M.A. Harmsen, A.W.D.v.d. Gon, J.H.B.J. Hoebink, J.C.
Schouten, H.H. Brongersma, J. Catal. 204 (2001) 420–427.
[2] H.S. Gandhi, G.W. Graham, R.W. McCabe, J. Catal. 216 (2003) 433–442.
[3] G. Parravano, J. Chem. Phys. 20 (1952) 342–343.
[4] W.F. Libby, Science 171 (1971) 499–500.

34
R. Zhang et al. / Journal of Molecular Catalysis A: Chemical 258 (2006) 22–34
[44] Y. Teraoka, H. Fukuda, S. Kagawa, Chem. Lett. (1990) 1–4.
[45] H. Yasuda, N. Mizuno, M. Misono, J. Chem. Soc., Chem. Commun. (1990)
1094–1096.
[46] C. Tofan, D. Klvana, J. Kirchnerova, Appl. Catal. A: Gen. 223 (2002)
275–286.
[47] J. Valyon, W.K. Hall, J. Phys. Chem. 97 (1993) 1204–1212.
[48] H.X. Dai, H. He, P.H. Li, Zh.L. Gao, C.-T. Au, Catal. Today 90 (2004)
231–244.
[49] L.G. Tejuca, J.L.G. Fierro, J.M.D. Tasco´n, Adv. Catal. 36 (1989) 237–327.
[50] P.J. Gellings, H.J.M. Bouwmeester, Catal. Today 12 (1992) 1–101.
[51] V.C. Belessi, P.N. Trikalitis, A.K. Ladavos, T.V. Bakas, P.J. Pomonis, Appl.
Catal. A: Gen. 177 (1999) 53–68.
[52] A.E. Giannakas, A.K. Ladavos, P.J. Pomonis, Appl. Catal. B: Environ. 49
(2004) 147–158.
[53] S. Royer, D. Duprez, S. Kaliaguine, Catal. Today 112 (2006) 99–102.
[54] S. Royer, D. Duprez, S. Kaliaguine, J. Catal. 234 (2005) 364–375.