Applied Catalysis B: Environmental 102 (2011) 362–371

Contents lists available at ScienceDirect

Applied Catalysis B: Environmental

journal homepage: www.elsevier.com/locate/apcatb

NOx removal efficiency and ammonia selectivity during the NOx

storage-reduction process over Pt/BaO(Fe, Mn, Ce)/Al2 O3 model catalysts. Part II:
Influence of Ce and Mn–Ce addition
N. Le Phuc, X. Courtois ∗ , F. Can, S. Royer, P. Marecot, D. Duprez
Laboratoire de Catalyse en Chimie Organique, Université de Poitiers, UMR 6503 CNRS, 40 Av. Recteur Pineau, Poitiers 86022, France

a r t i c l e i n f o a b s t r a c t

Article history: It was previously demonstrated in the first part of this work that NOx storage-reduction process over
Received 17 September 2010 Pt/BaO/Al2 O3 model catalyst is limited by the reduction step, with ammonia emission since H2 is not fully
Received in revised form 3 December 2010 consumed. The stored NOx reacts preferentially with the introduced H2 giving NH3 , than with NH3 in
Accepted 20 December 2010
order to produce N2 . Mn addition favors the NOx reduction with ammonia leading to better conversion
Available online 6 January 2011
and selectivity, but only at 400 ◦ C. In Part II, a special attention was focused on the role of Ce and Mn–Ce
addition in regard to the NOx conversion and the ammonia emission in the 200–400 ◦ C temperature
Keywords:
range. With ceria modified Pt/20Ba/Al catalyst, significant improvements are obtained from 300 ◦ C. In
NOx storage
NOx reduction
addition to the enhancement of the NOx + NH3 reaction, the ammonia selectivity is maintained at a lower
Ammonia level compared with Pt/Ba(Mn)/Al catalysts, even in the case of a large H2 excess. It is attributed to the
Barium ammonia oxidation into N2 via the available oxygen at the catalyst surface. A synergetic effect is observed
Lean/rich cycles between Mn and Ce when they are added simultaneously in Pt/Ba/Al catalyst.
Manganese © 2011 Elsevier B.V. All rights reserved.
Ceria

1. Introduction tion rate via the NOx reduction by H2 is higher than the ammonia
reaction rate with NOx to form N2 . It was also showed that H2 O
NOx storage reduction (NSR) catalysts are a possible way to inhibits the ammonia formation because it limits the formation
reduce NOx for diesel and lean burn engines [1]. They work mainly of CO via the reverse WGS reaction, CO being a precursor for the
in lean condition. During these periods, NOx are oxidized over pre- isocyanate formation, which leads to ammonia after hydrolysis.
cious metals and stored on basic compounds such as barium oxides, In absence of other introduced C-compounds, CO2 also favors the
mainly as nitrates. Periodically, the catalyst is submitted to short ammonia formation via the isocyanate route, with intermediate
periods for few seconds in rich conditions in order to reduce the formation of CO via the reverse WGS reaction.
trapped NOx into N2 [2,3]. In addition with deactivation by sul- Then, the influence of iron and manganese, both commonly pro-
fur poisoning [4,5] and thermal aging [6–8], another problem can posed in NSR formulation, was studied. Fe is reported to improve
be the NOx reduction selectivity. Indeed, in addition to N2 O, NH3 the catalyst sulfur resistance because it leads to the inhibition of
emission can be observed [9,10]. This work is mainly focused on bulk barium sulfates formation [12,13]. Fe is also reported to be
this ammonia emission. In the first part of this study [11], NOx active in NOx SCR by ammonia [14]. However, we showed that
removal efficiency of Pt/Ba/Al model catalyst was studied using Fe addition leads to a strong catalyst deactivation after successive
lean/reach cycling condition with H2 as reducer and CO2 and H2 O tests, probably due to interaction between iron and platinum. Mn
in the feed stream. It was established that the NOx reduction selec- can also participate in the NOx storage [15,16] and is active for
tivity strongly depends on the hydrogen conversion which was the NOx reduction by NH3 [14,17,18]. In fact, we found that Mn
introduced in the rich pulses: NH3 is emitted since hydrogen is addition induces different behaviors depending on the tempera-
not fully converted, whatever the NOx conversion rate. Moreover, ture test. At low temperature (200–300 ◦ C), Mn is a poison for the
the ammonia selectivity increases with the amount of unconverted reduction step. By contrast, at 400 ◦ C, Mn favors the NOx reduc-
hydrogen. This study has showed that Pt/Ba/Al catalyst is able to tion with ammonia, even if the introduced hydrogen is not fully
reduce NOx into N2 using NH3 as reducer, but the ammonia forma- converted, leading to a significant enhancement of the NOx con-
version and N2 selectivity. However, if a large hydrogen excess
is introduced, the ammonia selectivity becomes very close with
∗ Corresponding author. Tel.: +33 549 453 994; fax: +33 549 453 741. Pt/20Ba/Al and Pt/20BaMn/Al. Thus, the NOx conversion can be
E-mail address: [email protected] (X. Courtois). improved but the low temperature activity is still a problem. In

0926-3373/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcatb.2010.12.043
 N. Le Phuc et al. / Applied Catalysis B: Environmental 102 (2011) 362–371 363

this second part, influences of Ce and Ce–Mn addition were stud- 2.2. Catalyst characterizations
ied, especially toward ammonia emission. Cerium compounds are
well known in automotive catalysis for their oxygen storage/release 2.2.1. Specific surface measurement
behavior [19]. However, some interesting cerium properties were The BET surface areas and pore volumes were deduced from
put in evidence for NOx-trap systems. Ceria is claimed to improve N2 adsorption–desorption at −196 ◦ C carried out with a Tristar
the barium stability, with an inhibiting effect for the barium alu- 3000 Micromeritics apparatus. Prior to the measurement, the sam-
minate formation [20]. Barium–cerium interaction was evidenced ples were treated at 250 ◦ C under vacuum for 8 h to eliminate the
by BaCeO3 formation even if this specie is decomposed under adsorbed species.
NO2 –H2 O and destabilized under CO2 [8]. Migration of Ba ions
through CeZrOx compound was also observed by Liotta et al. [21] in 2.2.2. Platinum dispersion measurement
Pt–CeZrOx/Ba–Al2 O3 catalyst. This Ba–Ce interaction could allow The platinum dispersion was determined using the H2
a better control of the Ba dispersion as well as an improvement chemisorption method. The catalyst was first reduced under pure
of the resistance to SO2 poisoning [22]. These interesting proper- hydrogen at 500 ◦ C for 1 h and then flushed at the same tem-
ties toward sulfur poisoning regeneration are attributed to lower perature under argon for 3 h. The reactor was cooled down to
cerium sulfates stability compared with barium sulfates [23,24]. −85 ◦ C for Ce and Mn containing catalysts. Hydrogen was then
In addition, ceria compounds are able to store NOx [25,26,21]. For dosed on the sample until saturation. After flushing under argon
example, Ba/CeO2 material has exhibited a higher NOx storage effi- for 10 min, the sample was exposed to hydrogen again. The amount
ciency than Ba/Al2 O3 in the 200–400 ◦ C temperature range [20]. of chemisorbed hydrogen was taken as the difference between the
Similar results were obtained by Lin et al. [27] who investigated two hydrogen exposures.
the effect of La or Ce addition on the NOx storage properties of
Pt/Ba–Al2 O3 . The low temperature efficiency of ceria based stor- 2.2.3. XRD analysis
age material was also demonstrated with MnOx–CeO2 oxide [28]. X-ray powder diffraction was performed at room tempera-
Besides, cerium addition could improve the NOx removal effi- ture with a Bruker D5005 apparatus using a K␣ Cu radiation
ciency. Indeed, Pt/MgO–CeO2 catalyst was found to be active for ( = 1.54056 Å). The powder was deposited on a silicon monocrystal
low-temperature NO SCR by H2 of NO [29]. Concerning the ammo- sample holder. The crystalline phases were identified by compari-
nia selectivity, cerium could lead to lower emission. Indeed, in a son with the ICDD database files.
recent work about the NOx storage reduction (NSR) behavior of
Pt/Cex Zr1−x O2 catalysts, it was observed that the ammonia selec- 2.2.4. Temperature programmed reduction (TPR)
tivity decreases with the increase of the cerium loading [30]. Thus, Prior to the TPR test, the catalyst (50 mg) was first pretreated
the aim of the present work is to examine the influence of Ce addi- in situ under oxygen at 300 ◦ C for 30 min and cooled to room
tion on the NSR efficiency of Pt/Ba/Al model catalyst, with a special temperature. After flushing under argon for 45 min, the reduction
attention for the ammonia emission. Furthermore, association of was carried out from room temperature up to 800 ◦ C under a 1%
Mn and Ce addition is also studied. H2 /Ar mixture, using a 5 ◦ C min−1 heating rate. The sample was
maintained at 800 ◦ C for 30 min before cooling under argon. The
2. Experimental hydrogen consumption was followed by thermal conductivity.

2.1. Catalysts preparation 2.2.5. Oxygen storage capacity (OSC)

The OSC was measured at 400 ◦ C under atmospheric pres-
The detailed preparation protocols are reported in Part I [11]. sure. The sample (5 mg) was continuously purged with helium
The reference catalyst contains 1 wt% Pt and 20 wt% BaO supported (30 mL min−1 ). Alternate pulses (0.265 mL) of pure O2 and pure CO
on alumina. Alumina powder (230 m2 g−1 ) was immersed in an were injected every 2 min [31]. The oxygen storage capacity (OSC)
ammonia solution and was firstly impregnated using a barium was calculated from the CO consumption after stabilization.
nitrate salt. After evaporation at 80 ◦ C and drying at 120 ◦ C, the
obtained powder was treated at 700 ◦ C under synthetic dry air. 2.3. Catalytic activity measurements
Platinum was then impregnated using a Pt(NH3 )2 (NO2 )2 aqueous
solution. After drying, the catalyst was pre-treated at 700 ◦ C for 4 h 2.3.1. NOx storage capacity (NSC) measurement
under N2 , and finally stabilized at 700 ◦ C for 4 h under a mixture The same protocol as described in Part I [11] was used in this
containing 10% O2 , 5% H2 O in N2 . study. The catalyst (60 mg) was first was pretreated in situ for
The modified samples were prepared using the same protocol 30 min at 550 ◦ C, under a 10% O2 , 10% H2 O, 10% CO2 and N2 mix-
except that the nitrate salts of MnIV and CeIII were simultaneously ture (total flow rate: 10 L h−1 ), and then cooled down to the storage
added with the barium salt. In this case, a part of alumina was temperature under the same mixture. The sample was then submit-
replaced to assure the desired “additive/Ba” molar ratio. For the ted to a lean mixture as reported in Table 1, at 200 ◦ C, 300 ◦ C and
cerium modified samples, Ce/Ba molar ratio was varied between 400 ◦ C. Both NO and NOx concentrations (NO + NO2 ) were followed
0.25 and 2. In addition, catalysts containing both Mn and Ce were by chemiluminescence. H2 O was removed prior to NOx analysis
also prepared. In this case, Mn/Ba molar ratio is always 1 (7.2 wt% with a membrane dryer. The NOx storage capacity was estimated
for Mn) and Ce is added with Ce/Ba molar ratio between 0.1 and 1. by the integration of the recorded profile for the first 60 s and the
The Ce and Mn–Ce modified catalysts are noted Pt/20BaCeX/Al and contribution of the reactor volume was subtracted. With the con-
Pt/20BaMnCeX/Al, respectively, X being the Ce/Ba molar ratio. ditions used in this test, 57.4 ␮mol NOx per gram of catalyst were

Table 1
Rich and lean gas compositions used for the NOx conversion test (60 s lean/3 s rich). Total flow rate: 10 L h−1 .

Gas NO H2 O2 CO2 H2 O N2

Rich – 1–6% – 10% 10% Balance

Lean 500 ppm – 10% 10% 10% Balance
364 N. Le Phuc et al. / Applied Catalysis B: Environmental 102 (2011) 362–371

Table 2
Catalysts composition, corresponding BET surface areas and pore volumes. (Mn/Ba)molar ratio = 1 in Pt/20BaMnCeX/Al catalysts.

Catalyst Ce/Ba molar ratio Ce loading (wt%) BET surface area (m2 g−1 ) Pore volume (cm3 g−1 )

Pt/20Ba/Al – – 127 0.36

Pt/20BaCe0.25/Al 0.25 4.6 119 0.31
Pt/20BaCe0.5/Al 0.50 9.1 109 0.30
Pt/20BaCe0.75/Al 0.75 13.7 92 0.29
Pt/20BaCe1/Al 1.00 18.3 80 0.22
Pt/20BaCe1.5/Al 1.50 27.4 72 0.18
Pt/20BaCe2/Al 2.00 36.5 65 0.15
Pt/20BaMn/Al – – 118 0.29
Pt/20BaMnCe0.1/Al 0.10 1.8 112 0.29
Pt/20BaMnCe0.2/Al 0.20 3.7 109 0.28
Pt/20BaMnCe0.5/Al 0.50 9.1 89 0.25
Pt/20BaMnCe1/Al 1.00 18.3 78 0.21

injected in 60 s, which corresponds to the lean period durations of

the NSR test in cycling conditions. Results are expressed as the NIOx
storage rate (%) for 60 s. In addition, the catalyst oxidation activity
was estimated as the NO2 /NOx ratio (%) at saturation (usually about
900 s).

2.3.2. NOx conversion in cycling conditions

Before measurement, the catalyst (60 mg) was treated in situ
at 450 ◦ C under 3% H2 , 10% H2 O, 10% CO2 and N2 for 15 min. The
sample was then cooled down to the desired temperature (200,
300 and 400 ◦ C) under the same mixture. The NOx conversion was
studied in cycling condition by alternatively switching between
lean (60 s) and rich (3 s) conditions using electro-valves. The gas
composition is described in Table 1. NO and NO2 were followed by
chemiluminescence, N2 O by specific FTIR, H2 by mass spectrome-
try. Before the analyzers, H2 O was trapped in a condenser at 0 ◦ C. As
described in Part I [11], the trapped water was analyzed by two dif- Fig. 1. X ray diffractogramms of Pt/20BaCeX/Al catalysts. (#) BaCO3 , (*) BaAl2 O4 ,
ferent HPLCs for NH4 + , NO2 − and NO3 − . NO2 − and NO3 − were added and (£) CeO2 .

to the unconverted NOx. The N2 selectivity is calculated assuming

no other N-compounds than NO, NO2 , N2 O, NH3 . Some tests were containing both Mn and Ce, the only detected phase containing
also performed using a Multigas FTIR detector (MKS 2030) without Mn is BaMnO3 (Fig. 2). CeO2 is detected only for the samples with
water trap system. Same results were then obtained. the higher Ce loadings: Pt20BaMnCe0.5 and Pt20BaMnCe1. As for
Pt/20BaCeX/Al samples, it also corresponds with a decrease of the
3. Results BaAl2 O4 diffraction peaks.
Concerning the platinum dispersion, Shinjoh et al. have clearly
3.1. BET, XRD and platinum dispersion demonstrated that platinum particles are better anchored on ceria-
based materials, compared with alumina support, due to strong
The BET specific surface areas of the studied samples are Pt–O–Ce bonds [32]. Then, a better platinum dispersion can be
reported in Table 2. Compared with the Pt/20Ba/Al reference cata- expected with the increase of the ceria loading. However, at the
lyst (127 m2 g−1 ), addition of cerium leads to a continuous decrease same time, a decrease of the BET surface area is also observed
of the specific surface areas, down to 65 m2 g−1 for Pt/20BaCe2/Al. It (Table 2). Finally, after the thermal treatments at 700 ◦ C, no evi-
can be attributed to the partial alumina substitution by ceria which dence of a significant change of the platinum dispersion is observed
should have a lower specific surface. A decrease of the pore vol-
ume is also observed, especially for Ce/Ba molar ratio higher than
0.75. The same trend is observed with Pt/20BaMnCeX/Al catalysts.
Mn addition to Pt/20Ba/Al leads to a small decrease of the BET
surface area from 127 to 118 m2 g−1 . Supplementary addition of
cerium induces a more significant area loss, especially for Ce/Ba
molar ratios equal to 0.5 and 1. It also corresponds to a significant
drop in the pore volume.
XRD patterns of the Pt/20BaCeX/Al catalysts are reported in
Fig. 1. The main detected crystalline phases are BaCO3 , BaAl2 O4 and
CeO2 . The intensities of the ceria diffraction peaks increase with the
ceria loading, but the particle sizes cannot be reasonably estimated
with the Scherrer equation since an enlargement of the main peak is
caused by the presence of BaAl2 O4 (2 = 32◦ ). The diffraction peaks
of BaAl2 O4 decrease with cerium loading, which is consistent with
the decrease of the alumina loading. BaCO3 is observed only for
Ce/Ba molar ratio 0.75 and higher. It suggests that the barium dis-
persion decreases with the cerium additions, which is consistent Fig. 2. X ray diffractogramms of Pt/20BaMnCeX/Al catalysts. (*) BaAl2 O4 , (+)
with the related BET specific surface areas losses. For the catalysts BaMnO3 , and (£) CeO2 .
 N. Le Phuc et al. / Applied Catalysis B: Environmental 102 (2011) 362–371 365

Table 3
NOx storage rate (%) for 60 s and NO2 /NOx ratio at saturation (between brackets). Catalyst weight: 60 mg; total NOx storage corresponds to 57.4 ␮molNOx /g in 60 s.

Catalyst NOx storage rate for 60 s (%) and (NO2 /NOx ratio at saturation (%))

200 ◦ Ca 300 ◦ Ca 400 ◦ Ca

Pt/20Ba/Al 57 (16) 73 (24) 83 (37)

Pt/20BaCe0.25/Al 52 (15) 71 (27) 82 (42)
Pt/20BaCe0.5/Al 60 (21) 68 (28) 76 (40)
Pt/20BaCe0.75/Al 63 (13) 73 (20) 79 (35)
Pt/20BaCe1/Al 58 (15) 70 (25) 79 (36)
Pt/20BaCe1.5/Al 53 (14) 62 (24) 76 (38)
Pt/20BaCe2/Al 47 (21) 57 (27) 70 (39)
Pt/20BaMn/Al 61 (14) 74 (21) 84 (38)
Pt/20BaMnCe0.1/Al 63 (16) 73 (22) 86 (40)
Pt/20BaMnCe0.2/Al 63 (16) 77 (21) 93 (44)
Pt/20BaMnCe0.5/Al 67 (13) 81 (21) 96 (41)
Pt/20BaMnCe1/Al 78 (13) 93 (24) 99 (43)
a
Temperature (◦ C).

depending on the Mn and/or Ce loading. It reaches around 10% for 3.3.1. NOx storage-reduction efficiency of Pt/20BaCeX/Al catalysts
all studied samples. 3.3.1.1. Effect of the cerium loading. The NOx removal efficiency
tests in cycling condition were performed at 200, 300 and 400 ◦ C
3.2. NOx storage capacity in order to study the influence of the Ce loading in Pt/20Ba/Al.
The results are reported in (Fig. 3a–c) depending on the temper-
Influences of Ce or Mn–Ce additives on Pt/20Ba/Al on the NOx ature test. First, note that the NOx conversion rate is always lower
storage rate for 60 s were investigated at 200, 300 and 400 ◦ C. For than the NOx storage rate whatever the temperature test. These
the Pt/20BaCeX/Al catalysts, results reported in Table 3 show that results indicate that the NOx conversion is not limited by the stor-
an optimal composition is observed with Pt20BaCe0.75. However, age step but by the reduction step. In addition, the introduced
the obtained storage rate is close to the one obtained with the hydrogen during the rich pulses in never totally converted. Then,
Pt/20Ba/Al reference catalyst, but the BET surface area of the Ce the amount if introduced reducer is not a limiting parameter. How-
modified sample is 30% lower. In fact, ceria is known to be able to ever, the H2 conversion rate is enhanced with the increase of the
store NOx, but ceria addition also induces a BET surface loss. Indeed, ceria loading. This was expected since ceria is easily reduced in rich
for the higher Ce loading, the BET surface loss is not balanced by atmosphere.
the ceria storage behavior. At 200 ◦ C, the influence of the ceria loading is not really
The influence of cerium addition in Pt/20BaMn/Al catalysts significant compared with Pt/20Ba/Al (Fig. 3a). The maximum
was also examined. Some storage improvements were obtained NOx conversion is obtained with Pt/20BaCe1/Al catalyst at 38%,
depending on the cerium loading and the temperature test. compared with 35% for the sample without ceria. The ammo-
At 200 ◦ C, a moderate storage improvement is observed for nia selectivity is always higher with the Ce-modified catalysts,
(Ce/Ba)molar ratio ≤ 0.5, whereas the NOx storage rate is significantly between 27 and 35% versus 22% for Pt/20Ba/Al. Finally, among
better with the Pt/20BaMnCe1/Al. It reaches 78% versus 57% for the the modified samples, the optimal Ce loading is obtained with
Pt/20Ba/Al reference catalyst. At 300 and 400 ◦ C, the influence of (Ce/Ba)molar ratio = 1. Pt/20BaCe1/Al exhibits both the higher NOx
the cerium loading becomes significant from (Ce/Ba)molar ratio = 0.2, conversion and the lower ammonia selectivity, even though
and increases with the cerium loading even if the BET sur- Pt/20BaCe1/Al catalytic behaviors are still close to those obtained
face areas decrease. The maximum storage rates, obtained with with the Pt/20Ba/Al reference catalyst. Besides, it should be remem-
Pt/20BaMnCe1/Al, are 93 and 99%, respectively. Finally, Mn or Ce bered that modification with Mn or Fe led to a catalyst deactivation
addition does not improve the storage rate for 60 s, but a benefi- at this temperature at 200 ◦ C [11].
cial effect is observed when Ce and Mn are added together. In this The influence of cerium addition is more significant at higher
case, the more the temperature is low, the higher cerium loading temperatures. At 300 and 400 ◦ C (Fig. 3b and c), Ce addition always
is needed. improves the NOx conversion compare with Pt/20Ba/Al. An opti-
Table 3 also reports the NO2 /NOx ratios measured at mal rate is obtained with Pt/20BaCe1/Al again. At 300 ◦ C, the NOx
saturation. The influence of the cerium loading is rather lim- conversion reaches 50%, versus 29% for Pt/20Ba/Al. At 400 ◦ C, the
ited compared with the influence of the temperature test. corresponding rates are 63 and 45%, respectively. For higher Ce
Generally, the cerium addition is rather negative at 200 ◦ C, loadings, the NOx conversion decreases. It can be attributed to
there is nearly no influence at 300 ◦ C, and a small benefi- the significant surface area losses, as previously discussed in Sec-
cial effect is observed at 400 ◦ C. However, after saturation, tion 3.2. Interestingly, at 300 and 400 ◦ C, the ammonia selectivity
the NOx storage capacity of the catalysts varies as follows: continuously decreases with the ceria loading, whatever the NOx
Pt/20BaMnCe1/Al  Pt/20BaCe1/Al > Pt/20BaMn/Al > Pt/20Ba/Al. conversion rates. At 300 ◦ C, the ammonia selectivity is a little
Then, the significant enhancement of the NOx storage rate for 60 s higher with (Ce/Ba)molar ratio = 0.25 compared with Pt/20Ba/Al (25%
when Ce and Mn are added together cannot be attributed to an versus 21%), but it decreases from 25 to 8% with the increase of
improvement of the NO oxidation rate, but to an improvement of the cerium loading until (Ce/Ba)molar ratio = 2. At 400 ◦ C, the ammo-
the proximity between the NO oxidation sites and the NOx storage nia selectivity is always lower with the Ce-containing catalysts. It
sites, which is known to be a key factor for a fast NOx storage [33]. varies between 22 and 4% for Pt/20BaCe0.25/Al and Pt/20BaCe2/Al,
respectively.
3.3. NOx storage-reduction efficiency These results are consistent with those obtained previously with
Pt/Cex Zr1−x O2 catalysts [30]. It was demonstrated that the ammo-
In this section, the influence of ceria additions was investigated nia selectivity obtained in similar condition test was dependant of
first in Pt/20Ba/Al reference sample, and secondly in the Mn mod- the ceria–zirconia composition. The more the support composition
ified catalyst (Pt/20BaMn/Al, (Mn/Ba)molar ratio = 1). was cerium loaded; the lower was the ammonia selectivity.
366 N. Le Phuc et al. / Applied Catalysis B: Environmental 102 (2011) 362–371

(a) Pt/20Ba Pt/20Ba Pt/20Ba Pt/20Ba Pt/20Ba Pt/20Ba Pt/20Ba 1% H2 2% H2 3% H2 4% H2 5% H2 6% H2 7% H2 8% H2

200oC /Al Ce0.25 Ce0.5 Ce0.75 Ce1 Ce1.5 Ce2
/Al /Al /Al /Al /Al /Al 100%
100%
90%
90%
80%
80%
70%
70%
60% 60%

50% 50%
40% 40%
30% 30%
20%
20%
10%
10%
0%
0%
Temp. 200oC 200oC 200oC 200oC 200oC 200oC 200oC Temp. 400°C 400°C 400°C 400°C 400°C 400°C 400°C 400°C
H2 conv. 73% 69% 71% 76% 80% 81% 84% H2 conv. 100% 100% 97% 90% 78% 77% 68% 63%
N2 selec. 78% 65% 70% 70% 73% 69% 69% N2 selec. 99% 94% 91% 79% 79% 78% 78% 77%
NH3 selec. 22% 35% 30% 30% 27% 31% 31% NH3 selec. 1% 6% 9% 21% 21% 22% 22% 23%
NOx conv. 35% 35% 35% 35% 38% 32% 32% NOx conv. 37% 47% 63% 65% 66% 67% 67% 67%

(b) Pt/20Ba Pt/20Ba Pt/20Ba Pt/20Ba Pt/20Ba Pt/20Ba Pt/20Ba Fig. 4. Pt/20BaCe/Al catalyst. NOx storage/reduction efficiency test at 400 ◦ C. NOx
300oC /Al Ce0.25 Ce0.5 Ce0.75 Ce1 Ce1.5 Ce2 conversion (%) into N2 ( ) and into NH3 () and related data. Influence of H2
/Al /Al /Al /Al /Al /Al
concentration in the rich pulses (1–8%).
100%
90%
80% the hydrogen concentration up to 2% leads to a significant improve-
70%
ment of the NOx conversion, at 47%. H2 is still totally converted
60%
and the ammonia selectivity is still low, at 6%. 3% H2 is a transi-
50%
tional concentration. The NOx conversion increases again, at 63%,
40%
few hydrogen remains (97% are converted) and ammonia selectiv-
30%
ity becomes significant, at 9%. From 4 to 8% H2 , the NOx conversion
20%
is rather constant at 65–67%. In opposition with results obtained
10%
0%
with Pt/20Ba/Al and Pt/20BaMn/Al [11], the ammonia selectivity
Temp. 300°C 300°C 300°C 300°C 300°C 300°C 300°C
becomes almost constant at 21–23%, despite the increase of the
H2 conv. 70% 79% 84% 86% 90% 90% 92% hydrogen concentration. Then, it confirms the beneficial effect of
N2 selec. 79% 75% 81% 86% 87% 92% 92% Ce addition on the ammonia selectivity, even in the case of a large
NH3 selec. 21% 25% 19% 16% 13% 8% 8% hydrogen excess.
NOx conv. 29% 36% 41% 45% 50% 42% 40%

3.3.1.3. Conclusion about the cerium influence. Finally, Ce addition

(c) Pt/20Ba Pt/20Ba Pt/20Ba Pt/20Ba Pt/20Ba Pt/20Ba Pt/20Ba leads to a noteworthy NOx conversion improvement at 300 and
400oC /Al Ce0.25 Ce0.5 Ce0.75 Ce1 Ce1.5 Ce2
/Al /Al /Al /Al /Al /Al 400 ◦ C which can be attributed to an improvement of the NOx
100% reduction by ammonia formed in situ. A slight improvement can
90% be obtained at 200 ◦ C, depending on the ceria loading. For compar-
80% ison, Mn addition leads to similar NOx conversion improvement
70% at 400 ◦ C, but also to a deactivation of the catalyst at 200 ◦ C and
60%
300 ◦ C. However, the remarkable behavior of the Pt/20BaCeX/Al
50%
catalysts is to limit the ammonia selectivity, with a positive effect
40%
of the cerium loading. Nevertheless, for the higher cerium con-
30%
tent, the BET surface area drop induces a lower NOx conversion
20%
rate. Then the optimal (Ce/Ba)molar ratio was found to be 1. With
10%
this Pt/20BaCe1/Al catalyst, the ammonia selectivity reaches a lim-
0%
Temp. 400°C 400°C 400°C 400°C 400°C 400°C 400°C
ited rate of approximately 22–23% at 400 ◦ C if a large hydrogen
H2 conv. 85% 88% 91% 93% 97% 98% 99% excess is introduced during the rich pulses, whereas Pt/20Ba/Al and
N2 selec. 67% 78% 79% 87% 91% 93% 96% Pt/20BaMn/Al catalysts exhibited ammonia selectivity until 40%
NH3 selec. 33% 22% 21% 13% 9% 7% 4% with the same experimental conditions.
NOx conv. 45% 52% 54% 54% 63% 57% 54% Further investigations were done with the aim to com-
Fig. 3. Influence of the cerium loading in Pt/20BaCeX/Al catalyst: NOx stor- bine the beneficial effects of cerium and manganese additions.
age/reduction efficiency test at 200 ◦ C (a), 300 ◦ C (b) and 400 ◦ C (c) with 3% H2 in the Pt/20BaMn/Al formulation was selected and it was modified by
rich pulses. NOx conversion (%) into N2 ( ) and into NH3 () and related data. ceria addition. A part of alumina was then replaced in order to
obtain four (Ce/Ba) molar ratios between 0.1 and 1.
3.3.1.2. Effect of H2 concentration in the rich pulses. Then, cerium
addition dramatically decreases ammonia emission, even if the 3.3.2. NOx storage-reduction efficiency of Pt/20BaMnCeX/Al
introduced hydrogen is not fully converted. In order to check this catalysts
point, additional measurements were carried out at 400 ◦ C with The influence of cerium addition on the NOx storage-reduction
Pt/20BaCe1/Al. The H2 concentration in the rich pulses was varied efficiency of Pt/20Mn/Al was studied at 200, 300 and 400 ◦ C using
between 1 and 8%. Results are reported in Fig. 4. With 1% H2 , the first 3% H2 in the rich pulses. For comparison, results obtained with
NOx conversion is limited at 37%, the introduced reducer is totally the Pt/20Ba/Al reference catalyst are also reported in Figs. 5 and 6,
converted and there is nearly no ammonia emission. The increase of depending on the temperature test.
 N. Le Phuc et al. / Applied Catalysis B: Environmental 102 (2011) 362–371 367

(a) Pt/20Ba Pt/20Ba Pt/20Ba Pt/20Ba Pt/20Ba Pt/20Ba (a) Pt/20Ba Pt/20Ba Pt/20Ba Pt/20Ba Pt/20Ba Pt/20Ba
200oC /Al Mn MnCe0.1 MnCe0.2 MnCe0.5 MnCe1 /Al Mn MnCe0.1 MnCe0.2 MnCe0.5 MnCe1
3%H2
/Al /Al /Al /Al /Al /Al /Al /Al /Al /Al
100% 100%
90% 90%
80% 80%
70% 70%
60% 60%
50% 50%
40% 40%
30% 30%
20% 20%
10% 10%
0% 0%
Temp. 200°C 200°C 200°C 200°C 200°C 200°C Temp. 400°C 400°C 400°C 400°C 400°C 400°C
H2 conv. 73% 59% 72% 75% 89% 91% H2 conv. 85% 100% 100% 100% 100% 100%
N2 selec. 78% 61% 60% 51% 60% 67% N2 selec. 67% 93% 94% 96% 100% 100%
NH3 selec. 22% 39% 40% 46% 40% 33% NH3 selec. 33% 7% 6% 4% 0% 0%
NOx conv. 35% 22% 30% 36% 42% 46% NOx conv. 45% 62% 63% 71% 76% 70%
(b) Pt/20Ba Pt/20Ba Pt/20Ba Pt/20Ba Pt/20Ba Pt/20Ba (b) Pt/20Ba Pt/20Ba Pt/20Ba Pt/20Ba Pt/20Ba Pt/20Ba
300oC /Al Mn MnCe0.1 MnCe0.2 MnCe0.5 MnCe1 /Al Mn MnCe0.1 MnCe0.2 MnCe0.5 MnCe1
/Al /Al /Al /Al /Al 6%H2
/Al /Al /Al /Al /Al
100% 100%
90% 90%
80% 80%
70% 70%
60% 60%
50% 50%
40%
40%
30%
30%
20%
20%
10%
10%
0%
0%
Temp. 300°C 300°C 300°C 300°C 300°C 300°C Temp. 400°C 400°C 400°C 400°C 400°C 400°C
H2 conv. 70% 66% 80% 86% 99% 100% H2 conv. 63% 77% 83% 87% 94% 100%
N2 selec. 79% 51% 66% 68% 79% 93% N2 selec. 57% 64% 73% 79% 88% 100%
NH3 selec. 21% 49% 34% 32% 21% 7% NH3 selec. 43% 36% 27% 21% 12% 0%
NOx conv. 29% 21% 29% 40% 50% 53% NOx conv. 48% 69% 74% 89% 95% 94%
Fig. 5. Influence of the cerium loading in Pt/20BaMnCeX/Al catalyst: NOx stor-
Fig. 6. Influence of the cerium loading in Pt/20BaMnCeX/Al catalyst: NOx stor-
age/reduction efficiency test at 200 ◦ C (a) and 300 ◦ C (b) with 3% H2 in the rich
age/reduction efficiency test at 400 ◦ C with 3% H2 (a) or 6% H2 (b) in the rich pulses.
pulses. NOx conversion (%) into N2 ( ) and into NH3 () and related data.
NOx conversion (%) into N2 ( ) and into NH3 () and related data.

At 200 and 300 ◦ C (Figs. 5a and b), the NOx conversion is always
lower than the NOx storage rate (Table 3), indicating that the whole amount. However, it was previously found that manganese addi-
process is limited by the reduction step again at these temperatures. tion improves the NOx conversion and the selectivity of Pt/20Ba/Al
At 200 ◦ C, it was already reported that Mn addition led to catalyst at this temperature. Further cerium additions improve
a catalyst deactivation. Addition of cerium allows a continuous again the deNOx activity, and a maximum NOx conversion rate of
increase of the NOx conversion which becomes higher than the 76% is observed with Pt/20BaMnCe0.5/Al. In the same time, the
conversion obtained with the Pt/20Ba/Al reference catalyst for ammonia selectivity, which was already low with Pt/20BaMn/Al
(Ce/Ba)molar ratio ≥ 0.2. Ammonia selectivity remains high at low (7%) becomes nil from (Ce/Ba)molar ratio = 0.5 in Pt/20BaMnCeX/Al
cerium content (around 40%). It decreases at high cerium loading catalysts. Thus, with this condition with full H2 conversion, the
(33% for Pt/20BaMnCe1/Al), even if the selectivity remains higher in situ formed ammonia is able to react with the remaining stored
than the value obtained with Pt/20Ba/Al, at 22%. NOx to produce N2 , with a promoting effect of the cerium loading.
Compared with results obtained at 200 ◦ C, the NOx conversion With the higher cerium loading (Pt/20BaMnCe1/Al), the NOx
at 300 ◦ C increases more significantly with the cerium load- conversion reaches only 70%, even if this sample exhibits the higher
ing, from 22% for Pt/20BaMn/Al to 53% for Pt/20BaMnCe1/Al. NOx storage rate, at 99% for 60 s. This result can be attributed to the
(Ce/Ba)molar ratio = 0.1 is enough to observe similar NOx conver- important hydrogen consumption due to the catalyst reduction and
sion than Pt/20Ba/Al. The influence of the cerium loading on the the NOx conversion could be limited by the reducer amount. Then,
ammonia selectivity is more significant at 300 ◦ C compared with supplementary tests were performed with higher H2 concentration
results at 200 ◦ C. The ammonia selectivity continuously decreases, in the rich pulses (6%), with the aim to have no limitation by the
down to 7% for Pt/20BaMnCe1/Al. However, with this catalyst, the reducer amount.
introduced hydrogen is fully converted, which is favorable for low With 6% H2 in the rich pulses (Fig. 6b), the hydrogen con-
ammonia emission. The influence of the hydrogen concentration sumption is total only with Pt/20BaMnCe1/Al and the influence
on the NOx conversion and selectivity is examined in a next part of the cerium loading is more pronounced. A significant NOx con-
below. version improvement is observed with the cerium content, from
At 400 ◦ C with 3% H2 in the rich pulses, the hydrogen conver- 69% with Pt/20BaMn/Al to 94–95% with the two catalysts with the
sion reaches 100% with all the modified catalysts (Fig. 6a). As a higher cerium loading. These values can be considered as opti-
consequence, the NOx conversion could be limited by the reducer mal values since the NOx storage rates of these two samples reach
368 N. Le Phuc et al. / Applied Catalysis B: Environmental 102 (2011) 362–371

again, as seen in Section 3.3.1 with Pt/20BaCe1/Al catalyst. More-

over, the NOx conversion is improved when Mn and Ce are added
together.

3.3.3. Conclusion on the influence of the Mn, Ce and Mn–Ce

additive on the NSR behavior
Finally, Ce addition in Pt/20Ba/Al leads to slight improvement
of the NOx conversion at 200 ◦ C. The enhancement of the catalytic
properties is more significant at higher temperature, and especially
at 400 ◦ C. It can be attributed to an improvement of the reaction
between the ammonia formed in situ and the remaining stored
NOx.
At 200 and 300 ◦ C, cerium addition compensates the inhibit-
ing effect of Mn addition previously observed with Pt/20BaMn/Al
catalyst, especially with the higher cerium loading. At 400 ◦ C,
high NOx conversions can be obtained, near 100%, with very low
ammonia emission. Comparison of the NSR behaviors at 400 ◦ C of
Pt/20Ba/Al, Pt/20BaMn/Al, Pt/20BaCe1/Al and Pt/20BaMnCe0.5/Al
catalysts is reported in Fig. 8. It shows that Pt/20BaMnCeX/Al cat-
alysts exhibit the beneficial effects of both Mn and Ce for the NOx
conversion. Moreover, it also puts in evidence that addition of
cerium in Pt/20Ba/Al and in Pt/20BaMn/Al leads to a limitation of
the maximum ammonia selectivity at a low level compared with
Pt/20Ba(Mn)/Al catalysts. The ammonia selectivity tends to 40%
with Pt/20Ba/Al and Pt/20BMn/Al, and it is two times lower with
the Ce-containing catalysts.
Ji et al. [35] have showed somewhat similar results. They have
tested catalysts with various ceria contents and they have observed
that the ammonia selectivity decreased with the increase of the
ceria loading. They have proposed that ammonia could react with
Fig. 7. Pt/20BaMnCe0.5/Al catalyst (60 mg): NOx storage/reduction efficiency test
the available oxygen from the ceria surface. Then, the reducibility
at 300 ◦ C (a) and 400 ◦ C (b). NOx conversion (%) into N2 ( ) and into NH3 () and of the samples was studied, using H2 temperature programmed
related data. Influence of H2 concentration in the rich pulses. reduction (TPR) and oxygen storage capacity (OSC) measurements.

96–99%, and because a NOx desorption without reduction is usually 3.4. Role of the catalyst reducibility
observed during the process [34]. In addition with the NOx conver-
sion improvement, the ammonia selectivity clearly decreases with 3.4.1. TPR measurements
the cerium loading. It suggests that cerium promotes the reaction First, it was previously showed that Pt/Ba/Al catalyst did not
between ammonia and stored NOx, even if the introduced hydrogen exhibit significant H2 consumption using the same TPR operating
is not totally converted. condition [23]. In addition, whatever the studied catalyst, no signif-
For a better understanding of the Mn–Ce modified catalysts icant H2 consumption was observed below 100 ◦ C.The TPR profiles
behaviors, tests with various H2 concentrations were performed of the Pt/20BaMnCeX/Al catalysts are reported in Fig. 9. The reduc-
at 300 and 400 ◦ C. Pt/20BaMnCe0.5/Al was selected because tion of 20BaMn/Al material without platinum was also studied.
Pt/20BaMnCe1/Al exhibits very high H2 consumption. Its reduction profile (not shown) exhibits H2 consumption in the
At 300 ◦ C (Fig. 7a), the NOx conversion increases from 24 to 54% 100–800 ◦ C temperature range, with a main peak near 550–600 ◦ C
with the increase of the H2 concentration in the rich pulses from and two broad peaks near 300 and 750 ◦ C. These temperatures
1 to 4%. Higher hydrogen concentration does not improve the NOx are higher than those reported by Christel et al. [36] for the
conversion rate anymore. In opposition, the ammonia selectivity reduction of unsupported MnO2 (2MnO2 + H2 → Mn2 O3 + H2 O
is limited to 21–23% when H2 is fully converted, and it increases near 250 ◦ C, 3Mn2 O3 + H2 → 2Mn3 O4 + H2 O at 280 ◦ C, and
until 34% with 6% H2 in the rich pulses. Thus, at 300 ◦ C, ammonia Mn3 O4 + H2 → 3MnO + H2 O near 450 ◦ C). This stabilization of
is always detected even if the introduced H2 is fully converted. It the manganese oxide can be attributed to interactions with other
means that the reaction rate between the in situ formed ammonia compounds such as barium, leading for instance to BaMnO3 as
and the stored NOx is lower compared with the ammonia formation detected by XRD (see Section 3.1). The Pt/20BaMn/Al sample
rate, i.e. NOx react preferentially with H2 to form ammonia than (symbol () in Fig. 9) exhibits one main peak in the 250–450 ◦ C
with ammonia itself to form N2 at this temperature. temperature range and a small H2 consumption in the 600–800 ◦ C
At 400 ◦ C (Fig. 7b), the NOx conversion rate increases with the temperature range. Then, platinum dramatically favors the man-
hydrogen concentration until it reaches 95% with 6% H2 , which ganese reduction, probably due to the hydrogen activation. In
can be considered as a maximum conversion value. Since the addition, this result indicates that platinum and manganese are
introduced hydrogen is fully converted, no ammonia emission is close together. Assuming that manganese is reduced into MnII at
observed, in opposition with results obtained at 300 ◦ C, indicating the end of the TPR test (after 30 min at 800 ◦ C), H2 consumption
that the ammonia reactivity is significantly improved at 400 ◦ C. H2 quantification allows the evaluation of the initial manganese mean
remains for introduced concentration of 5% and higher, and ammo- redox state according MnOx + (x − 1) H2 → MnO + (x − 1) H2 O. The
nia is then detected. However, the ammonia selectivity is limited to obtained manganese redox state is 3.2, indicating that MnIII species
16–17% from 7% H2 in the rich pulses. Then, the beneficial effect of such as Mn2 O3 are predominant, with also the presence of MnIV
cerium addition on the maximum ammonia selectivity is observed (BaMnO3 , MnO2 ).
 N. Le Phuc et al. / Applied Catalysis B: Environmental 102 (2011) 362–371 369

100
100

80
80

NOx conversion (%)

NH3 selectivity (%)

60
60

40 40

20 20

0 0
1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8

[H2 ] (%) [H2] (%)

Fig. 8. NOx conversion rate (full symbols) and NH3 selectivity (open symbols) measured at 400 ◦ C depending on hydrogen concentration in the rich pulses for Pt/20Ba/Al (,
♦), Pt/20BaMn/Al (, ), Pt/20BaCe1/Al (, ) and Pt/20BaMnCe0.5/Al (䊉, ).

Compared with Mn containing samples, 20BaCe/Al and The Pt/20Ba/Al reference catalyst exhibits an OSC value of
Pt/20BaCe/Al samples tested with the same protocol exhibit very 28 ␮molO g−1 . Addition of manganese leads to a two times higher
low hydrogen consumption (results not shown). However, the typ- OSC whereas TPR experiments show a high reducibility. In oppo-
ical two steps reduction of CeO2 was obtained. The reduction of sition with the TPR results, catalyst modified by ceria addition
the CeIV surface species occurs first (at about 300 ◦ C without plat- presents redox properties in cycling condition significantly higher
inum, and near 200 ◦ C when the promoting effect of platinum is than Pt/20BaMn/Al, with 121 ␮molO g−1 versus 61 ␮molO g−1 ,
observed). The reduction of bulk ceria is obtained for higher tem- respectively. TPR experiments also showed that the reducibility of
perature, it is not achieved at 800 ◦ C [37,38]. Pt/20BaMnCeX/Al samples was not significantly affected by ceria
The influence of Ce loading on the TPR profiles of addition. On the contrary, the OSC of Pt/20BaMn/Al is more than
Pt/20BaMnCeX/Al samples is also reported in Fig. 9. In fact, three times enhanced after Ce addition (Pt/20BaMnCe1/Al catalyst,
cerium addition induces only small changes. The total H2 con- Table 4). Then, compared with manganese addition, ceria addition
sumption is not really affected by the ceria addition; only a small largely improves the available oxygen on the catalyst surface in
shift of the main manganese reduction peak is noticeable, from transient condition, with a synergetic effect of the simultaneous
320 to 340 ◦ C. presence of manganese and cerium.
Finally, the TPR measurements do not allow us to obtain a cor-
relation between the samples reducibility and the catalytic activity 3.4.3. Discussion
in cycling condition. Ceria loading induces significant improvement In the first part of this study [11], it was demonstrated that
in NSR measurements but the reducibility evaluated by H2 -TPR is ammonia is one of the intermediate in the NOx reduction pro-
mainly dependant of the presence of manganese. However, cat- cess. With Pt/20Ba/Al catalysts, ammonia emission is due to the
alytic tests were performed in cycling condition, in opposition fact that the ammonia formation rate, via the NOx + H2 reaction, is
with the TPR measurements. Then, further investigations were lower than the reaction rate between this in situ formed ammo-
performed to evaluate the reducibility of the catalysts in cycling nia and the stored NOx in order to obtain N2 . At 400 ◦ C, there is no
conditions. ammonia emission if the introduced hydrogen is fully converted.
This general trend is always observed whatever the composition of
3.4.2. OSC measurements the studied catalysts, but only at 400 ◦ C (Fig. 7). In accordance with
OSC measurements were performed at 400 ◦ C using the CO–O2 the OSC measurements, the higher is the OSC, the higher is the
pulsed method. Results obtained for Pt/20Ba/Al, Pt/20BaMn/Al, hydrogen consumption and ammonia is emitted for higher hydro-
Pt/20BaCe/Al and Pt/20BaMnCe1/Al and catalysts are reported in gen concentration. For temperature below 400 ◦ C, ammonia can be
Table 4. observed even if hydrogen is fully converted (Fig. 7). It indicates that
NOx + NH3 reaction rate remains limited, even with the more active
catalysts for this reaction. When H2 is not fully converted, exami-
nation of the obtained results indicates some correlation between
H2 consumption (a.u.)

the amount of remaining hydrogen and the ammonia selectivity,

but there is a large dispersion of the points depending on the
catalyst composition and the catalytic test condition (figure not
shown).
Mn addition to Pt/20Ba/Al can induce a significant improve-
ment of the NOx conversion, but only at 400 ◦ C. It was attributed
to an enhancement of the NOx reduction with ammonia. Ceria
also improves the NOx reduction with ammonia (Figs. 3 and 6),
from 300 ◦ C. Contrary to manganese, ceria clearly leads to lower
ammonia emission, even if the introduced hydrogen is not fully
100 200 300 400 500 600 700 800
converted (large H2 excess, Fig. 8). Then, the available oxygen
from ceria is supposed to participate to the ammonia oxidation.
Temperature (ºC)
Two products can be considered, N2 and/or NOx, both leading to
Fig. 9. H2 -TPR profiles of Pt/20BaMn/Al (), Pt/20BaMnCe0.2/Al (), lower ammonia selectivity. Fig. 8 shows that, for high hydrogen
Pt/20BaMnCe0.5/Al (♦) and Pt/20BaMnCe1/Al (×) catalysts. concentration, Pt/20BaCe/Al exhibits similar NOx conversion com-
370 N. Le Phuc et al. / Applied Catalysis B: Environmental 102 (2011) 362–371

Table 4
Oxygen storage capacities (OSC, ␮molO g−1 ) measured at 400 ◦ C for Pt/20Ba/Al, Pt/20BaMn/Al, Pt/20BaCe/Al et Pt/20BaMnCe1/Al catalysts.

Catalyst Pt/20Ba/Al Pt/20BaMn/Al Pt/20BaCe/Al Pt/20BaMnCe1/Al

−1
OSC (␮molO g ) 28 61 121 403

pared with Pt/20BaMn/Al, both catalysts being more efficient than and selectivity), but only at 400 ◦ C, whereas the activity was inhib-
Pt/20Ba/Al. However, the ammonia selectivity is significantly lower ited at 200 and 300 ◦ C. With ceria modified Pt/20Ba/Al catalyst, no
and limited with the ceria containing catalysts (Ce and Mn–Ce mod- deactivation is observed at 200 ◦ C and significant improvements
ified catalysts). As a consequence, the main product of the ammonia were obtained from 300 ◦ C. In addition to the enhancement of the
oxidation by the available oxygen from the support should be NOx + NH3 reaction, ceria addition led to a limitation of the ammo-
N2 since oxidation into NOx should decrease the NOx conver- nia selectivity at a lower level compared with Pt/20Ba(Mn)/Al
sion. Finally, ceria promotes both the NOx + NH3 reaction, from catalysts. It was attributed to the ammonia oxidation into N2 via
300 ◦ C, and the ammonia oxidation into N2 via the available oxy- the available oxygen at the catalyst surface. A synergetic effect
gen at the catalyst surface. A synergetic effect is observed when was observed between Mn and Ce in Pt/20BaMnCeX/Al catalysts.
the catalyst contains both Ce and Mn: OSC is largely enhanced A near total NOx conversion can be observed at 400 ◦ C, and ammo-
and, at the same time, NOx conversion and ammonia emission are nia is limited to a lower level. These results are correlated with
improved. Nearly total NOx conversion into N2 can be obtained a large improvement of the oxygen mobility for the Mn–Ce con-
(Fig. 6) at 400 ◦ C, which indicates high NOx reduction rates (NOx taining catalysts, even if this parameter should be not the only
reduction into NH3 and then NH3 reaction with stored NOx or oxy- one.
gen in order to obtain N2 ). However, the NOx reduction efficiency Al lower temperature (300 ◦ C), the reduction rates are still the
at 300 ◦ C is still limited, notably due to insufficient NOx + NH3 limiting steps of the NSR process, notably because of the insufficient
reaction rate (ammonia emission even in the case of total H2 con- NOx + NH3 reaction rate.
version).
In addition to the improvement of the reaction between NH3 References
and the stored NOx with the increase of the ceria loading, two other
phenomenons can be considered. [1] W.S. Epling, L.E. Campbell, A. Yezeerets, N.W. Currier, J.E. Parks II, Catal. Rev. 46
Firstly, the increase of the H2 consumption with the increase of (2004) 163–245.
[2] T. Kobayashi, T. Yamada, K. Kayano, SAE Technical Papers 970745, 1997, p. 63.
the ceria loading indicates that a part of the support is also reduced [3] S. Matsumoto, Cattech 4 (2000) 102–109.
during the rich pulses, in accordance with the OSC measurements [4] C. Sedlmair, K. Seshan, A. Jentys, J.A. Lercher, Catal. Today 75 (2002)
carried out in cycling condition. Then, the NO decomposition over 413–419.
[5] E.C. Corbos, X. Courtois, N. Bion, P. Marecot, D. Duprez, Appl. Catal. B 80 (2008)
the reduced ceria can be considered [39]. Nevertheless, such a reac-
62–71.
tion should not induce a decrease of the ammonia yield, as observed [6] J. Li, J. Theis, W. Chun, C. Goralski, R. Kudla, J. Ura, W. Watkins, M. Chattha, R.
for example in Fig. 3c. Hurley, SAE Technical Paper No 2001-01-2503, 2001.
However, the increase of the H2 consumption with the ceria [7] D. Uy, A.E. O’Neill, J. Li, W.L.H. Watkins, Top. Catal. 95 (2004) 191–201.
[8] M. Casapu, J.D. Grunwaldt, M. Maciejewski, M. Wittrock, U. Göbel, A. Baiker,
loading also leads to a lower H2 pressure at the end of the catalytic Appl. Catal. B 63 (2006) 232–242.
bed. Then, a lower NH3 selectivity could be expected in the tail of [9] R.D. Clayton, M.P. Harold, V. Balakotaiah, Appl. Catal. B 84 (2008) 616–630.
the catalytic bed, which would be compatible with the observed [10] L. Lietti, I. Nova, P. Forzatti, J. Catal. 257 (2008) 270–282.
[11] N. Le Phuc, X. Courtois, F. Can, S. Berland, S. Royer, P. Marecot, D. Duprez, Appl.
results. Two points allow us to decline this hypothesis. First, what- Catal. B 102 (2011) 362–371.
ever the studied catalyst, no ammonia emission is observed only [12] K. Yamazaki, T. Suzuki, N. Takahasi, K. Yojota, M. Sugiura, Appl. Catal. B 30
in the case of total H2 conversion, and only at 400 ◦ C. Secondly, (2001) 459–468.
[13] P.T. Fanson, M.R. Horton, W.N. Delgass, J. Lauterbach, Appl. Catal. B 46 (2003)
at 300 ◦ C, the ammonia selectivity reaches around 20% even if the 393–413.
introduced hydrogen is fully converted (Fig. 7a, Pt/20BaMnCe0.5/Al [14] J. Huang, Z. Tong, Y. Huang, J. Zhang, Appl. Catal. B 78 (2008) 309–314.
catalyst), but the beneficial effect of the cerium loading on both NOx [15] U. Bentrup, A. Bruckner, M. Richter, R. Fricke, Appl. Catal. B 32 (2001)
229–241.
conversion and ammonia selectivity is observed (Figs. 5b and 3b). [16] J. Xiao, X. Li, S. Deng, F. Wang, L. Wang, Catal. Commun. 9 (2008) 563–567.
Then, the decrease of the ammonia yield in presence of remain- [17] X. Liang, J. Li, Q. Lin, K. Sun, Catal. Commun. 8 (2007) 1901–1904.
ing H2 conversely with the increase of the N2 yield suggests that [18] Z. Wu, B. Jiang, Y. Liu, Appl. Catal. B 79 (2008) 347–355.
[19] F. Dong, A. Suda, T. Tanabe, Y. Nagai, H. Sobukawa, H. Shinjoh, M. Sugiura, C.
the reaction between NH3 and the stored NOx is the main cause
Descorme, D. Duprez, Catal. Today 93–95 (2004) 827–832.
of the NOx reduction improvement. The NO decomposition over [20] M. Eberhardt, R. Riedel, U. Göbel, J. Theis, E.S. Lox, Top. Catal. 30/31 (2004)
reduced ceria cannot be excluded, but it seems to be negligi- 135–142.
ble. [21] L.F. Liotta, A. Macaluso, G.E. Arena, M. Livi, G. Centi, G. Deganello, Catal. Today
75 (2002) 439–449.
Nevertheless, the activity improvement obtained by cerium [22] H. Kwak, D.H. Kim, J. Szanyi, C.H.F. Peden, Appl. Catal. B 84 (2008) 545–551.
addition should not be attributed only to its OSC properties. [23] E.C. Corbos, S. Elbouazzaoui, X. Courtois, N. Bion, P. Marecot, D. Duprez, Top.
Indeed, in a recent study, Pt/Cex Zr1−x O2 catalysts were studied Catal. 42–43 (2007) 9–13.
[24] H. Mahzoul, L. Limousy, J.F. Brilhac, P. Gilot, J. Anal. Appl. Pyrol. 56 (2000)
[30]. The increase of the cerium content led to an increase of 179–193.
the NOx conversion and, simultaneously, to a decrease of the [25] S Philipp, A. Drochner, J. Kunert, H. Vogel, J. Theis, E.S. Lox, Top. Catal. 30/31
ammonia selectivity (i.e. improvement of the NOx + NH3 reaction). (2004) 235–238.
[26] P. Svedberg, E. Jobson, S. Erkfeldt, B. Andersson, M. Larsson, M. Skoglundh, Top.
Among the studied composition, the best activity was obtained Catal. 30 (2004) 199–206.
with Pt/Ce0.8 Zr0.2 O2 , which exhibited the higher cerium content, [27] H.Y. Lin, C.J. Wu, Y.W. Chen, C.H. Lee, Ind. Eng. Chem. Res. 45 (2006)
whereas Pt/Ce0.58 Zr0.42 O2 exhibited the higher OSC. 134–141.
[28] M. Machida, D. Kurogi, T. Kijima, Chem. Mater. 12 (2000) 3165–3170.
[29] C.N. Costa, A.M. Efstathiou, Appl. Catal. B 72 (2007) 240–252.
4. Conclusion [30] P.N. Lê, E.C. Corbos, X. Courtois, F. Can, P. Marecot, D. Duprez, Appl. Catal. B 93
(2009) 12–21.
[31] S. Kacimi, J. Barbier.Jr., R. Taha, D. Duprez, Catal. Lett. 22 (1993) 343–350.
In the first part of this study, it was shown that Mn addition to [32] H. Shinjoh, M. Hatanaka, Y. Nagai, T. Tanabe, N. Takahashi, T. Yoshida, Y. Miyake,
Pt/20Ba/Al led to an improvement of the NOx reduction (conversion Top. Catal. 52 (2009) 1967–1971.
 N. Le Phuc et al. / Applied Catalysis B: Environmental 102 (2011) 362–371 371

[33] B. Pereda-Ayo, R. Lopez-Fonseca, J.R. Gonzalez-Velasco, Appl. Catal. A 363 [36] L. Christel, A. Pierre, D.A.M. Rousset Abel, Thermochim. Acta 306 (1997) 51–59.
(2009) 73–80. [37] H.C. Yao, Y.F. Yu Yao, J. Catal. 86 (1984) 254–265.
[34] E. Fridell, M. Skoglundh, B. Westerberg, S. Johansson, G. Smedler, J. Catal. 183 [38] E. Rogemond, R. Fréty, V. Perrichon, M. Primet, S. Salasc, M. Chevrier, C. Gauthier,
(1999) 196–209. F. Mathis, J. Catal. 169 (1997) 120–131.
[35] Y. Ji, J.-S. Choi, T.J. Toops, M. Crocker, M. Naseri, Catal. Today 136 (2008) [39] G. Ranga Rao, P. Fornasiero, R. Di Monte, J. Kaspar, G. Vlaic, G. Balducci, S.
146–155. Meriani, G. Gubitosa, A. Cremona, M. Graziani, J. Catal. 162 (1996) 1–9.