Experimental and DFT studies on Sr-doped LaMnO3 catalysts for NOx storage and reduction

Article abstract of DOI:10.1039/c5cy00073d

The idea of rational design of perovskite catalysts for NO_x storage and reduction (NSR) starts from DFT studies on Sr-doped LaMnO₃ (001) plane models: a Mn-terminated plane (Mn-ter) has a higher activity than La- or Sr-ter planes; the number of A-site defects increases when Sr is doped on the surface; O-vacancy formation energies for both La- and Mn-ter gradually decrease with increased Sr loading, and the values for La-ter are always larger than for Mn-ter with the same Sr loading, indicating that the O-vacancy formation is facile on Mn-ter with Sr-doping. This model yields the highest reactivity for oxidation and the lowest energy barrier for O-vacancy formation. A surface tuning method for La_0.5Sr_0.5MnO₃ is introduced. With the exception of the promotion of surface area and pore volume, the ratio of Sr/La and the number of surface-active oxygen atoms and Mn⁴⁺ cations exposed on the outermost layers are improved with an increased contact time between the raw materials and dilute HNO₃. However, overtreatment leads to a less stable phase of MnO₂ with a high reducibility but significantly restrained NO_x adsorption. NSR performances under lean-burn/fuel-rich cycles are measured, and the results correspond well with the NO adsorption or oxidation behaviors and DFT calculations.

Full text of DOI:10.1039/c5cy00073d

Catalysis
Science &
Technology
View Article Online
View Journal
PAPER
Experimental and DFT studies on Sr-doped
LaMnO3 catalysts for NOx storage and reduction
Cite this: DOI: 10.1039/c5cy00073d
a
a
Yue Peng,ab Wenzhe Si, Junhua Li, John Crittenden b and Jiming Haoa
*
*
The idea of rational design of perovskite catalysts for NOx storage and reduction (NSR) starts from DFT
studies on Sr-doped LaMnO3 (001) plane models: a Mn-terminated plane (Mn-ter) has a higher activity than
La- or Sr-ter planes; the number of A-site defects increases when Sr is doped on the surface; O-vacancy
formation energies for both La- and Mn-ter gradually decrease with increased Sr loading, and the values
for La-ter are always larger than for Mn-ter with the same Sr loading, indicating that the O-vacancy forma-
tion is facile on Mn-ter with Sr-doping. This model yields the highest reactivity for oxidation and the lowest
energy barrier for O-vacancy formation. A surface tuning method for La0.5Sr0.5MnO3 is introduced. With
the exception of the promotion of surface area and pore volume, the ratio of Sr/La and the number of
surface-active oxygen atoms and Mn4+ cations exposed on the outermost layers are improved with an
increased contact time between the raw materials and dilute HNO3. However, overtreatment leads to a less
stable phase of MnO2 with a high reducibility but significantly restrained NOx adsorption. NSR perfor-
mances under lean-burn/fuel-rich cycles are measured, and the results correspond well with the NO
adsorption or oxidation behaviors and DFT calculations.
Received 17th January 2015,
Accepted 10th February 2015
DOI: 10.1039/c5cy00073d
www.rsc.org/catalysis
NOx and the active sites. One of the solutions is to use poly-
meric materials as soft templates for the preparation of
Introduction
Nitrogen oxides (NOx) are among the major air pollutants
emitted from power plants and diesel vehicles. Selective cata-
lytic reduction of NOx with NH3 has been demonstrated to be
effective for power plant exhaust gas, while NOx storage
reduction (NSR) technology is a promising method to elimi-
nate NOx from diesel vehicles.1–3 In a common NSR catalyst,
Pt/BaO/Al2O3, NO is oxidized to NO2 on Pt and subsequently
adsorbed on BaO as a nitrate species in the lean-burn
period.4,5 During the short fuel-rich period, hydrocarbons are
supplied to reduce NOx to N2. Recently, Sr-doped perovskite
oxides (ABO3) were discovered to be efficient platinum substi-
tutes in diesel oxidation and NSR reactions.6 A-sites of perov-
skites, such as La or Sr, are good sites for NOx adsorption and
B-sites (Mn or Co) are highly active for NO oxidation.7,8 The
oxidation activity depends on the reducibility, oxygen mobility
within the lattice, and the surface defects.9–11
porous perovskites.12–14 Though this method increases the
surface area up to 60 m2 g−1, the templates are too expensive
for large scale production in industry. Developing economical
methods for perovskite synthesis remains
a challenge.
Another problem is determining how to modify the catalyst
surface to expose most of the active sites for efficient NOx
adsorption and reduction. The facile strategy is doping alkali
or rare-earth metals to the A-sites and transitional or precious
metals to the B-sites, which improves the reducibility of the
B-sites and provides surface-active oxygen.15,16 However, the
optimized redox sites cannot be easily observed due to the
nonstoichiometric nature of perovskites. This property limits
the control over the particle size and distribution, as well as
the surface anchorage.11 Therefore, it is necessary to design
reasonable models to predict these critical properties of
perovskites for a given reaction.
There are two problems that need to be resolved for the
application of these catalysts. Catalysts synthesized by tradi-
tional methods have low surface areas and large particle
sizes, thereby preventing adequate contact between gaseous
Density functional theory (DFT) is a powerful tool to eluci-
date the electronic structures of catalytic surfaces and their
reactivity.17,18 Researchers are also interested in the structure
of perovskite models. The (001) plane of LaMnO3 plays an
important role in oxygen-related processes. MnO2 (001) is
found to be the most stable termination for a solid oxide fuel
cell, and the interaction of O with a surface O-vacancy is the
rate-determining step for an oxidation reaction.19–23 Recently,
it has been reported that Sr could increase the oxidation
ability and diminish the inhibition of NO2 for La1−xSrxCoO3.
a State Key Joint Laboratory of Environment Simulation and Pollution Control,
School of Environment, Tsinghua University, Beijing, 100084, China.
E-mail: lijunhua@tsinghua.edu.cn; Fax: +1 404 894 5676; Tel: +86 10 62771093
b School of Civil and Environmental Engineering, Georgia Institute of Technology,
Atlanta, Georgia, 30332-0595, USA. E-mail: john.crittenden@ce.gatech.edu
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
In addition, the formation of vacancy sites is more facile on
CoO2-terminated surfaces.19 However, a systematic DFT study
on LaMnO3 perovskite, especially the relationship between
electronic structures and reducibility, remains incomplete.
In this work, we employed DFT calculations to study the
basic properties for a NSR reaction on Sr-doped LaMnO3
(001) planes and subsequently introduced a surface tuning
method to modulate the La0.5Sr0.5MnO3 catalysts. After char-
acterization by temperature programmed reduction (H2-TPR)
and X-ray photoelectron spectroscopy (XPS) spectra, the actual
NSR performance under alternating lean-burn/fuel-rich cycles
was also studied.
1
(1)
Ev  E(LSMv )  E(O2 )  E(LSM),
2
where EIJLSMv) is the energy of the reduced slab by removing
a surface oxygen atom, EIJO2) is the energy of oxygen, and
EIJLSM) is the energy of the slab. It is important to note that
a small value of Ev indicates the ease of forming O-vacancies.
Catalyst preparation
The perovskite was synthesized by a hydrothermal method.
LaIJNO3)3, SrIJNO3)2, KMnO4, and MnCl2 were mixed with a
20 g KOH solution with stirring. The solution was transferred
to a stainless steel lined Teflon vessel for hydrothermal treat-
ment at 260 °C for 24 h, and then the precipitate was washed
and dried at 60 °C. The precipitate was crushed and sieved
through meshes >60 times. A catalyst surface modification
method was then used to dissolve the sample with 3 M HNO3
for different lengths of time. The samples were first dried
and then calcined at 400 °C for 2 h. The raw catalyst was
denoted as LSM1, and the samples treated with dilute HNO3
for 10 min, 1 h, and 10 h were denoted as LSM2, LSM3, and
LSM4, respectively.
Experimental
Models selection and DFT details
For the theoretical calculations, we employed the ab initio
DFT plane wave based Materials Studio 5.5. The generalized
gradient approximation plus Hubbard model (GGA+U),
according to Perdew, Burke, and Ernzerhof, was used,24–26
where the value of U was set to be 5.0 for Mn.27 The typical
plane wave cut-off energy was Ecut = 450 eV and the spin
polarizations were also applied. The Monkhorst–Pack division
scheme was selected to generate a set of k-points within the
Brillouin zone.
Catalyst characterization
The BET surface area of the samples was investigated with a
Micromeritics ASAP 2020 apparatus. The crystal structure was
determined using XRD measurements (Rigaku, D/max-
2200/PC) between 20° and 80° at a step rate of 10° min−1,
operating at 40 kV and 30 mA, using Cu Kα radiation. A SEM
image was obtained with a Helios NanoLab 600i Dual Beam
System, FEI Company. XPS was performed with an
ESCALab220i-XL electron spectrometer from VG Scientific
using 300 W Al Kα radiation. The binding energy was
referenced to the C 1s line at 284.8 eV. The bulk element
atomic compositions were characterized by ICP-AES with an
IRIS Intrepid II XSP apparatus (Thermo Fisher Scientific
Inc.). H2-TPR was performed on a ChemiSorb 2720 TPx chem-
isorption analyzer under 10% H2/Ar gas flow (50 mL min−1),
at a rate of 10 °C min−1, up to 1000 °C. The H2 consumption
of each catalyst preparation was calculated by comparison
with that of a standard CuO sample.
Models were cut from the (001) plane of LaMnO3 bulk
with a lattice constant of a = 3.88 Å and a space group of
Pm3m. We selected the seven-layer slab for the geometry
optimization, where the periodically repeated slabs were
separated by a vacuum gap of 15 Å. The slabs were termi-
nated by different planes: La–O and Mn–O were denoted
as La-ter and Mn-ter, respectively. A Sr-doped plane was
constructed by replacing the La atoms in the outermost
layer and sub-outermost layer with Sr atoms in the La-ter
and Mn-ter models. The models were denoted as Sr-ter
and Sr–Mn-ter (Fig. 1). To study the electronic structures of
the A-site defects, we deleted one surface La and Sr atom on
the La-ter and Sr-ter models. In addition, 1 to 4 surface La
cations were substituted for Sr atoms to study the
O-vacancies and their formation energy (Ev). The Ev value is
calculated as:
NSR tests
NO oxidation measurements were performed in a fixed-bed
quartz reactor using 100 mg of the catalyst. The feed gas mixture
contained 500 ppm NO, 3% O2, and the balance was N2.
Temperature programed desorption (NO-TPD) was performed in
a fixed-bed quartz reactor. Samples were purged under 1000
ppm NO at room temperature after being pretreated in N2 at 350
°C for 1 h. After isothermal removal of physically adsorbed NO,
the temperature was increased to 600 °C at a rate of 10 °C min−1.
The lean/rich cycles were evaluated at 250 °C in alternative lean-
burn (500 ppm NO, 5% O2) and fuel-rich (500 ppm NO, 1000
ppm C3H6) atmospheres. The total flow rate was 200 mL min−1.
Fig. 1 The optimized structures of (a) La-ter, (b) Mn-ter and the corre-
sponding Sr-doped (c) Sr-ter and (d) Sr–Mn-ter perovskite (001) planes.
Cyan balls are La, purple balls are Mn, green balls are Sr and red balls
are oxygen.
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Catalysis Science & Technology
Paper
The concentrations of the gases (NO, NO2, and C3H6) were
continually monitored by an FTIR spectrometer (MultiGas
TM 2030 FTIR). To better evaluate the catalytic activity,
kinetic parameters were calculated according to the follow-
ing equation, applied with the NO conversion:
integrals between ions and neighboring site electrons led to a
strong ionic interaction through the bridged oxygen between
Mn atoms.28 Compared to the top of the VBs of Mn 3d for
La-ter and Mn-ter, a slight movement to higher energy was
observed from −1.74 eV to −1.65 eV, indicating that the Mn 3d
states of Mn-ter exhibit stronger binding to the surface than
those of La-ter.
A new impurity level of Sr–Mn-ter occurred across the
Fermi energy, and the blue-shift of the VB became more
apparent than for Mn-ter. The results imply that Mn in the
perovskite lattice might have a higher reactivity when it is
exposed on the outermost layer and the reactivity could be
further improved by Sr. That is, to enhance the activity of the
perovskite during the oxidation process, more Mn–O needs
to be exposed, and dopants such as Sr are necessary to mod-
ify the proportions of the A-sites.
V
(2)
k    (1 x),
W
where k is the reaction rate constant (mL g−1 s−1), V is the
total gas flow rate (mL s−1), W is the mass of catalyst in the
reactor, and x is the NO conversion. The apparent activation
energy was calculated using the Arrhenius equation, which is
given by:
k = k0e−E /RT
,
(3)
a
Fig. 3 demonstrates the PDOS graphs of the La-ter and Sr-
ter planes and their corresponding A-site defect (Ad) models,
where a surface La or Sr cation was removed (dashed circle
on the right side of the figure). The Mn 3d states of La-ter
were nearly unchanged by eliminating La. Only the end of
the CB shifted 0.5 eV to a higher energy. Mn cations on the
sub-outermost layer of La-ter with Ad moved towards the
defect site to some extent (indicated by the arrow). In com-
parison to Mn 3d states of Sr-ter, the top of the VB and the
impurity level of the Ad model shifted to the Fermi energy,
and the bottom of the CB only presented slight changes. The
Ad of Sr-ter also showed differences from the Ad of La-ter
model. In addition to the movement of the Mn cations
towards the defect site, La on the third layer also moved up
to the defect site. The results indicate that Mn on the sub-
outermost layer can be activated by the formation of A-site
defects and can be further improved by the dopant Sr.
where Ea is the energy barrier, and it can be obtained from
the slope of the lnIJk) versus 1/T in formula 3.
Results and discussion
Influence of exposed surfaces, dopants, and A-site defects
As can be observed in Fig. 2, the projected density of state (PDOS)
of Mn 3d contains a valence band (VB) at −7.5 eV < E < 0
and a conduction band (CB) up to 3.0 eV without a strict
band gap. The PDOS graphs for the spin-up and spin-down
states are considerably different for Mn 3d on both sides of
the Fermi energy (E = 0), while O 2p shows a lower state at
these regions. One can conclude from the contribution of Mn
and O states to the PDOS that the ionic bond, rather than the
covalent component of the Mn–O bond, grows from the low-
est surface band to the top of the VB, even though the Mn–O
bond is not exposed on the outermost layer (La-ter). The large
polarization of the Mn 3d states from the small hopping
Fig. 2 Projected density of state (PDOS) graphs of Mn 3d and O 2p
orbitals in the La-ter, Mn-ter and Sr–Mn-ter models.
Fig. 3 The PDOS of the Mn 3d orbitals in the La-ter, Sr-ter and A-site
defect models, and the front views of the defect models.
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
the reactivity of Mn cations. Moreover, the surface with termi-
nated Mn atoms has a higher activity than those with termi-
nated La or Sr atoms, and the surface oxygen has a higher
mobility (low Ev) with more Sr-doped on the terminated Mn
surface. However, one must note that in the NSR reaction,
sufficient amounts of alkali metals are necessary for NO
adsorption directly from the gas phase or NO2 adsorption
from oxidation sites. The following work is based on the
results that improved the properties of the NSR reaction.33
Physicochemical properties of the perovskite
Fig. 5 presents the XRD patterns and SEM images of the
LSM1 to LSM4 catalysts. The SEM images showed that some
layer clusters grew up from the bulk materials (LSM2). They
became considerable on LSM3, and a large amount of clus-
ters were observed on LSM4. Combined with the XRD pat-
terns, these clusters might be MnO2 oxides. The main peaks
for LSM1 to LSM3 are attributed to the standard perovskite
phase (a = 3.88 Å with space group Pm3m), which exhibits
good agreement with the bulk model used in DFT. LSM4
displayed low crystallization properties compared to the other
three samples, and MnO2 diffraction peaks appeared. Except
for MnO2, both La2O3 and SrO are soluble in dilute HNO3.
With increased contact time, more A-site cations escape from
the perovskite structures and only MnO2 is left on the sur-
faces of the materials.
Table 1 lists the surface area, pore volume and Sr/La ele-
ment ratios for both the bulk and surfaces, which were all
improved with increased contact time between the raw mate-
rials and dilute HNO3. The surface areas was ranged from 0.36
m2 g−1 for LSM1 to 67.15 m2 g−1 for LSM4. The pore volumes of
the catalysts presented the same trend. We also prepared
another sample with a 72 h treatment, which had a surface
area of 178.95 m2 g−1, and a crystal phase which was mainly
β-MnO2 after calcined at 400 °C, indicating this modulate
method is effective for enlarging the surface area and pores
in a material.
Fig. 4 O-vacancy formation energy (Ev) for the La-ter and Mn-ter
models with increasing Sr-doping loadings.
O-vacancy formation energy
To directly obtain the relationship between the oxidation
activity and the influence factors discussed above, we calcu-
lated the O-vacancy formation energy (Ev) for both the La-ter
and Mn-ter models with different Sr coverages (Fig. 4). The Ev
values of the La-ter models were greater than those of the
Mn-ter models for all Sr-doping models, suggesting that
O-vacancies preferred to form on the Mn-ter surface, rather
than the La-ter surface, i.e. Mn-ter surfaces have a higher oxy-
gen storage/release capacity, thereby increasing the oxidation
activity during the reaction process. The results are in agree-
ment with the discussion of the electronic structures in
Fig. 2. With increased Sr loading, the Ev values of both the
La-ter and Mn-ter models decreased. The lowest value was
less than 2.0 eV for Sr–Mn-ter (the sub-outermost layers were
“totally” covered by Sr). It is important to note that we could
not prepare LaxSr1−xMnO3 materials with all of the Sr cations
exposed on the outermost layer because high levels of SrCO3
block the perovskite surface when excess SrIJNO3)2 precursors
are introduced during the preparation.29,30 Therefore, the
results above are only applicable for the DFT modeling. How-
ever, other researchers have also tried to synthesize
LaxSr1−xMnO3 with high Sr loading, and they found that a
value of Sr between 0.3 and 0.5 seemed better for the oxida-
tion reaction.7,31,32 Overall, the effect of Sr is to promote the
oxidation activity and O-vacancy formation in the LaMnO3
(001) planes. The absolute value of Ev for the La-ter and Mn-
ter models increased with increased Sr loading, revealing that
the improvement caused by Sr is greater in the Mn-ter model
than in the La-ter model. Consequently, the surface exposed
Mn–O bonds have a higher oxidation activity than the
exposed La–O or Sr–O bonds.
The ratios of the surface-active oxygen and the Mn4+ cat-
ions were calculated from the XPS spectra (Fig. 6) and are
summarized in Table 1. The O 1s peaks can be separated into
two peaks, referred to as the lattice oxygen at 529.3–530.3 eV
In summary, in addition to controlling the catalyst surface
areas or pore volumes, Sr improves the number of adsorption
sites, modifies the surfaces with A-site defects, and enhances
Fig. 5 SEM images and XRD patterns of the catalysts.
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Catalysis Science & Technology
Paper
Table 1 Surface area (SA), pore volume (PV), element atomic ratio of Sr and La, proportion of surface adsorbed oxygen, Mn4+ determined by XPS for
the four perovskite materials
Sample
SA (m2 g−1
)
PV (cm3 g−1
)
Sr/Laa
Sr/Lab
Oad/Olattb (%)
Mn4+/Mn3+b (%)
LSM1
LSM2
LSM3
LSM4
0.36
5.90
22.22
67.15
0.02
0.03
0.09
0.20
0.90
0.92
0.93
1.01
0.80
0.94
0.99
1.10
13.7
27.9
32.1
33.9
42.0
62.8
63.7
65.8
a
Calculated from the ICP results of the catalyst powder. b Calculated from the XPS spectra.
Fig. 6 (a) O 2p and (b) Mn 3d XPS spectra of the catalysts.
Fig. 7 H2-TPR profiles of the perovskite materials in the temperature
range of 30–1000 °C.
(Olatt) and the chemisorbed surface oxygen at 530.9–531.9 eV
(Oad). Oad is highly active in oxidation reactions due to its
higher mobility compared to the lattice oxygen (Olatt).33 The
spectra of Mn 2p can also be separated into two peaks, attrib-
uted to Mn3+ (641.1 and 652.8 eV) and Mn4+ (643.0 and 654.3
eV).29 They were both enhanced by elevating the contact time:
LSM1 only had 13.7% Oad/Olatt and 42.0% Mn4+/Mn3+, while
LSM4 yielded 33.9% Oad/Olatt and 65.8% Mn4+/Mn3+. Com-
bined with the DFT results, this tuning method increases
both the number of surface-active oxygen atoms and the con-
centration of Mn4+ cations, leading to more O-vacancy
formation.34
attributed to the reduction of Mn4+ to Mn3+ on MnO2 sur-
faces, and the peaks in the range of 397 °C to 505 °C for the
samples were due to the reduction of Mn4+ on the perovskite
surfaces. The H2 consumption peaks for the catalysts are
listed in Table 2 and the peaks were enhanced with increased
contact time. That is, the method improved both the quantity
(peak area) and reducibility (peak position) of the Mn4+ cat-
ions on the perovskite structures.
Fig. 7 illustrates the H2-TPR profiles of the catalysts. Two
broad bands centered at 505 °C and 755 °C for LSM1 can be
attributed to the reduction of Mn4+ to Mn3+ for charge com-
pensation from Sr incorporation, and the reduction of Mn3+
to Mn2+, respectively.34–36 With an increased contact time,
the peak at 505 °C for LSM1 shifted to a lower temperature,
at 397 °C for LSM4. Another new peak centered at 293 °C
appeared for LSM3 and grew considerably for LSM4. Com-
bined with the structural characterization information, this
reduction peak can be attributed to being mainly from the
contribution of MnO2 (Mn4+ to Mn3+). Previous work on a
comparative study on α-Mn2O3 and β-MnO2 obtained similar
TPR peaks at approximately 300 °C on pure β-MnO2 mate-
rial.37 By calculating the ratio of H2 consumption, this peak
was attributed to the reduction of MnO2 to Mn3O4, i.e. partial
reduction of β-MnO2. Therefore, this peak can be specifically
NSR performance
Fig. 8(a) displays the Arrhenius plots of NO oxidation over
the samples. The apparent energy barrier for LSM1 was rela-
tively high (72.7 kJ mol−1), however the energy barriers for
the other samples were within 44–50 kJ mol−1, and only a slight
decrease appeared between LSM3 and LSM4. The results dem-
onstrated that with increased Mn4+ cations and surface-active
oxygen atoms, NO oxidation activity is significantly improved.
Moreover, apart from the oxidation ability, NOx adsorption is
another important property for NSR technology.
Fig. 8(b) presents the NO-TPD profiles of the four samples
in the temperature range of 50–600 °C, and the total NO
desorption values normalized by surface area are listed in
Table 3. The total NOx desorption significantly decreased
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
Table 2 H2 consumption calculated from the H2-TPR profiles
high temperature NO peak can be attributed to the nitrite or
nitrate species with strong bonding, and the NO2 desorption
peak can be attributed to the nitrate species.38,39 In compari-
son to the peaks for LSM4 and LSM3, the intensity of the low
temperature NO peak increased significantly. However, the
intensity of the high temperature NO peak decreased, even
when it was normalized by the surface area. Based on the
crystal structures of the two samples, the low temperature
peak is tentatively assigned to NO adsorption on MnO2,
which is easily desorbed at 200 °C, and it seems ineffective
for the NSR process (stable adsorption above 250 °C). The
concentration of NO2 is enhanced for LSM4, which shows a
good agreement with the NO oxidation.
H2 consumption (a.u.)
Sample
<300 °C
<600 °C
<800 °C
LSM1
LSM2
LSM3
LSM4
0
0
5.0
20.6
10.0
11.1
11.4
13.8
4.5
5.7
4.0
1.0
The NO adsorption/desorption behaviors of the samples
were determined in the periodically alternating lean-burn/
fuel-rich atmospheres, as shown in Fig. 9. After the samples
reached steady states, LSM1 only provided limited NO cap-
ture: seventy percent of NO escapes out of the reactor
directly, and NO2 reaches equilibrium within 1 min. For
LSM2 and LSM3, the escaped NO concentrations are nearly
100 ppm. The equilibrium time of NO2 in a lean-burn period
is reduced; approximately 2 min for LSM2 and 1.5 min for
LSM3. The results indicate that increased contact time with
the catalysts could improve the NO storage process. Only a
small amount of NO is obtained for LSM4 and NO2 does not
achieve a steady concentration within 3 min, implying that
the LSM4 catalyst exhibits the best NO oxidation activity
among the samples. The improvement in this NSR perfor-
mance originates from the synergistic effect of positive fac-
tors: an increase in surface area, surface-active oxygen atoms
and Mn4+ cations, as well as negative factors, such as a
decrease in the number of adsorption sites. Under the fuel-
rich conditions, the trapped NOx is released and reduced by
C3H6 and no N2O is detected during the entire period. The
NOx maximums of the samples were improved with increased
Fig. 8 (a) Arrhenius plot of NO oxidation and (b) NO-TPD profiles of
the perovskite materials.
with an increased contact time, suggesting the adsorption
sites for NOx are not improved by surface tuning. Consider-
ing the decrease of La and Sr element proportions, it can be
concluded that the major adsorption sites for NOx are La or
Sr cations, rather than Mn cations or MnO2. Moreover, the Sr
loading from ICP and XPS exhibited the same values for
LSM3 and LSM4. This might result from the fact that only
some of the Sr atoms are successfully doped in the perovskite
lattice and the remnant form SrCO3 during the preparation
process. With increased contact time, most of the SrCO3 is
dissolved in HNO3 and the remaining Sr cations are enriched
on the surface of the materials. Two NO peaks and a NO2
peak can be observed on the NO-TPD profiles for the cata-
lysts. The low temperature NO peak (below 300 °C) can be
attributed to the weakly adsorbed NO or nitrite species, the
Table 3 Element atomic compositions of Sr, La and Mn, and total NOx
desorption calculated by NO-TPD over the four perovskite materials
Sr loading
(%)
La loading
(%)
Mn loading
(%)
a
NOx
(mmol
m−2
Sample
ICP
XPS
ICP
XPS
ICP
XPS
)
LSM1
LSM2
LSM3
LSM4
13.0
12.9
10.6
4.0
27.1
25.8
10.0
4.1
14.4
14.0
11.4
4.0
33.9
27.4
10.1
3.7
23.0
23.4
24.5
29.2
39.0
46.8
67.5
83.2
2.19
0.19
0.14
0.06
Fig. 9 Lean/rich cycles of the perovskite materials operated at 300
°C. Lean-burn period: 3 min, gas composition: 500 ppm NO/5% O2;
fuel-rich period: 2 min, gas composition: 500 ppm NO/1000 ppm
C3H6.
a
Calculated from NO-TPD normalized by SA.
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Catalysis Science & Technology
Paper
contact time, suggesting the NO oxidation ability was
enhanced.
and 21221004), the National High-Tech Research and the
Development (863) Program of China (2013AA065401 and
2013AA065304) and the International Postdoctoral Exchange
Fellowship Program of China (20130032). The authors appre-
ciate the support of the Brook Byers Institute for Sustainable
Systems, Hightower Chair, and Georgia Research Alliance at
Georgia Institute of Technology.
The outlet C3H6 peaks significantly decreased: nearly 800
ppm for LSM1 and 150 ppm for LSM4, indicating that more
C3H6 can react with the trapped NOx to fulfil the NSR pro-
cess. The NO concentrations of the samples increased under
the fuel-rich period, which shows the same trend as the NO-
TPD results. Moreover, the NO peak widened for LSM4 com-
pared with LSM2 and LSM3, indicating that perovskite with
excess MnO2 decreases the NSR performance. This can be
attributed to the lack of effective adsorption sites such as
alkali metals on MnO2, and the fact that the adsorption pro-
cess of NO is weak and NO is easily desorbed from the sur-
face of MnO2 under NSR working temperatures (250–300 °C).
The NO-TPD profiles of LSM4 also showed similar results.
Notes and references
1 F. Prinetto, G. Ghiotti, I. Nova, L. Lietti, E. Tronconi and P.
Forzatti, J. Phys. Chem. B, 2001, 105, 12732–12745.
2 N. Phuc, X. Courtois, F. Can, S. Berland, S. Royer, P. Marecot
and D. Duprez, Catal. Today, 2011, 176, 424–428.
3 F. Frola, M. Manzoli, F. Prinetto, G. Ghiotti, L. Castoldi and
L. Lietti, J. Phys. Chem. C, 2008, 112, 12869–12878.
4 C. DiGiulio, V. Komvokis and M. Amiridis, Catal. Today,
2012, 184, 8–19.
Conclusions
Combining DFT calculations with experimental results, three
aspects of LaxSr1−xMnO3 catalysts have been modulated and
evaluated for NSR technology:
(1) The surface area and pore volume of the catalysts are
increased by the modification process because of the corro-
sion effect of dilute HNO3 to La2O3 and SrO rather than
MnO2, which provides more effective surfaces for NOx
adsorption or oxidation;
(2) La or Sr cations in the outermost layer of the perov-
skite are removed by the modification process, promoting
the number of O-vacancies and the reactivity of the Mn4+ cat-
ions in the perovskite lattice. From the calculated models, a
Mn-terminated surface with Sr-doped in the A-sites of the
LaMnO3 (001) plane yields the highest activity for oxidation
and the lowest energy barrier for O-vacancy formation. This
exhibits good agreement with the apparent energies barriers
for NO oxidation;
(3) The crystallized MnO2 oxides might form on the outer-
most layer of the perovskite if the contact time with HNO3
becomes too long. Although the reducibility of the material is
further enhanced, the number of NO adsorption sites dra-
matically decreases even when normalized by the surface
area. The adsorbed NOx is unstable and can easily desorb
from the catalyst surface, leading to a NO adsorption block-
age in the NSR process.
5 N. Phuc, X. Courtois, F. Can, S. Royer, P. Marecot and D.
Duprez, Appl. Catal., B, 2011, 102, 353–361.
6 C. Kim, G. Qi, K. Dahlberg and W. Li, Science, 2010, 327,
1624–1627.
7 X. Li, Y. Dong, H. Xian, W. Hernandez, M. Meng, H. Zou, A.
Ma, T. Zhang, Z. Jiang, N. Tsubaki and P. Vernoux, Energy
Environ. Sci., 2011, 4, 3351–3354.
8 L. Guo, H. Xian, Q. Li, D. Chen, Y. Tan, J. Zhang, L. Zheng
and X. Li, J. Hazard. Mater., 2013, 260, 543–551.
9 M. Pena and J. Fierro, Chem. Rev., 2001, 101, 1981–2018.
10 J. Zhu, H. Li, L. Zhong, P. Xiao, X. Xu, X. Yang, Z. Zhao and
J. Li, ACS Catal., 2014, 4, 2917–2940.
11 D. Neagu, G. Tsekouras, D. Miller, H. Ménard and J. Irvine,
Nat. Chem., 2013, 5, 916–923.
12 J. Xu, J. Liu, Z. Zhao, J. Zheng, G. Zhang, A. Duan and G.
Jiang, Catal. Today, 2010, 153, 136–142.
13 H. Arandiyan, H. Dai, J. Deng, Y. Wang, H. Sun, S. Xie, B.
Bai, Y. Liu, K. Ji and L. Li, J. Phys. Chem. C, 2014, 118,
14913–14928.
14 Y. Liu, H. Dai, Y. Du, J. Deng, L. Zhang and Z. Zhao, Appl.
Catal., B, 2012, 119–120, 20–31.
15 T. Ishihara, H. Matsuda and Y. Takita, J. Am. Chem. Soc.,
1994, 116, 3801–3803.
16 E. Edri, S. Kirmayer, M. Kulbak, G. Hodes and D. Cahen,
J. Phys. Chem. Lett., 2014, 5, 429–433.
Therefore, we tentatively propose the goal of rationally
designing perovskite catalysts for NSR technology: to improve
the surface area or pore volume, to expose more transitional
metals in the outermost layer of the catalyst which can pro-
vide adequate adsorption sites, and to increase the doping
amount to the catalyst. This improvement could result in
SrCO3 crystallization on the surface to cover the oxidation
sites.
17 P. Maitarad, D. Zhang, R. Gao, L. Shi, H. Li, L. Huang, T.
Rungrotmongkol and J. Zhang, J. Phys. Chem. C, 2013, 117,
9999–10006.
18 P. Maitarad, J. Han, D. Zhang, L. Shi, S. Namuangruk and T.
Rungrotmongkol, J. Phys. Chem. C, 2014, 118, 9612–9620.
19 S. Choi, M. Penninger, C. Kim, W. Schneider and L.
Thompson, ACS Catal., 2013, 3, 2719–2728.
20 E. Kotomin, R. Evarestov, Y. Mastrikov and J. Maier, Phys.
Chem. Chem. Phys., 2005, 7, 2346–2350.
21 E. Kotomin, Y. Mastrikov, E. Heifets and J. Maier, Phys.
Chem. Chem. Phys., 2008, 10, 4644–4649.
Acknowledgements
The authors gratefully acknowledge the financial support of
the National Natural Science Foundation of China (21407088
22 Y. Mastrikov, E. Heifets, E. Kotomin and J. Maier, Surf. Sci.,
2009, 603, 326–335.
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
23 M. Kuklja, E. Kotomin, R. Merkle, Y. Mastrikov and J. Maier,
Phys. Chem. Chem. Phys., 2013, 15, 5443–5471.
24 A. Becke, Phys. Rev. A: At., Mol., Opt. Phys., 1988, 38,
3098–3100.
25 B. Delley, J. Chem. Phys., 1990, 92, 508–517.
26 J. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett.,
1996, 77, 3865–3868.
32 M. de Jong, I. Bergenti, W. Osikowicz, R. Friedlein, V.
Dediu, C. Taliani and W. Salaneck, Phys. Rev. B, 2006, 73,
052403.
33 K. Huang, X. Chu, L. Yuan, W. Feng, X. Wu, X. Wang and S.
Feng, Chem. Commun., 2014, 50, 9200–9203.
34 K. Huang, X. Chu, W. Feng, C. Zhou, W. Si, X. Wu, L. Yuan
and S. Feng, Chem. Eng. J., 2014, 244, 27–32.
27 C. Sun, Y. Wang, J. Zou and S. Smith, Phys. Chem. Chem.
Phys., 2011, 13, 11325–11328.
28 V. Avdeev and V. Tapilin, J. Phys. Chem. C, 2010, 114, 3609–3613.
29 F. Teng, W. Han, S. Liang, B. Gaugeu, R. Zong and Y. Zhu,
J. Catal., 2007, 250, 1–11.
30 S. Liang, F. Teng, G. Bulgan and Y. Zhu, J. Phys. Chem. C,
2007, 111, 16742–16749.
31 M. Zhi, G. Zhou, Z. Hong, J. Wang, R. Gemmen, K. Gerdes,
A. Manivannan, D. Ma and N. Wu, Energy Environ. Sci.,
2011, 4, 139–144.
35 E. Crumlin, E. Mutoro, Z. Liu, M. Grass, M. Biegalski, Y. Lee,
D. Morgan, H. Christen, H. Bluhm and Y. Shao-Horn, Energy
Environ. Sci., 2012, 5, 6081–6088.
36 J. Fierro, J. Tascón and L. Tejuca, J. Catal., 1985, 93, 83–91.
37 X. Tang, J. Li, L. Sun and J. Hao, Appl. Catal., B, 2010, 99,
156–162.
38 Y. Peng, J. Li, X. Huang, X. Li, W. Su, X. Sun, D. Wang and J.
Hao, Environ. Sci. Technol., 2014, 48, 4515–4520.
39 Y. Peng, C. Liu, X. Zhang and J. Li, Appl. Catal., B,
2013, 140–141, 276–282.
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2015