Comparison of the catalytic activity of MO2 (M = Ti, Zr, Ce) for soot oxidation under NOx/O2

Article abstract of DOI:10.1016/j.jcat.2007.05.015

The catalytic activity of TiO₂, ZrO₂, and CeO₂ for soot oxidation under NO_x/O₂ was compared. Characterisation of the oxides by N₂ adsorption at -196 °C, Raman spectroscopy, and XRD revealed that the predominant crystalline phase of TiO₂ (rutile or anatase) and ZrO₂ (monoclinic or tetragonal) had no significant affect on the catalytic activity of these oxides for soot oxidation with NO_x/O₂. CeO₂, with the only allowed fluorite structure, is the most active oxide because accelerates the NO conversion to NO₂, whereas TiO₂ and ZrO₂ do not catalyse this reaction. Once NO₂ is produced by CeO₂, part of it reacts with soot, and a certain amount is stored on CeO₂. NO_x stored on CeO₂ under reaction conditions (around 450 °C) does not accelerate soot oxidation and can be evolved under N₂ flow but not under a NO_x/O₂ stream. NO_x storage on TiO₂ and ZrO₂ under reaction conditions was not detected. An additional benefit of CeO₂ with regard to TiO₂ and ZrO₂ is that the CeO₂-catalysed soot oxidation yields mainly CO₂, whereas TiO₂ and ZrO₂ yield higher percentages of CO. This can be attributed to the lowest CeO₂-catalysed soot oxidation temperature and to the ability of CeO₂ to catalyse CO oxidation to CO₂. The effect of thermal sintering at 800 °C (decreased BET surface area and increased crystal size) on CeO₂ is more important than on TiO₂ and ZrO₂, but 800 °C-calcined CeO₂ is still more active for soot combustion than the other oxides studied.

Full text of DOI:10.1016/j.jcat.2007.05.015

Journal of Catalysis 250 (2007) 75–84
www.elsevier.com/locate/jcat
Comparison of the catalytic activity of MO (M = Ti, Zr, Ce) for soot
2
oxidation under NO /O
x
2
∗
I. Atribak, I. Such-Basáñez, A. Bueno-López , A. García-García
Department of Inorganic Chemistry, University of Alicante, Ap. 99, E-03080 Alicante, Spain
Received 15 March 2007; revised 9 May 2007; accepted 20 May 2007
Available online 29 June 2007
Abstract
The catalytic activity of TiO , ZrO , and CeO for soot oxidation under NOx/O was compared. Characterisation of the oxides by N adsorption
2
2
2
2
2
◦
at −196 C, Raman spectroscopy, and XRD revealed that the predominant crystalline phase of TiO (rutile or anatase) and ZrO (monoclinic or
2
2
tetragonal) had no significant affect on the catalytic activity of these oxides for soot oxidation with NOx/O . CeO , with the only allowed fluorite
2
2
structure, is the most active oxide because accelerates the NO conversion to NO , whereas TiO and ZrO do not catalyse this reaction. Once
2
2
2
NO is produced by CeO , part of it reacts with soot, and a certain amount is stored on CeO . NOx stored on CeO under reaction conditions
2
2
2
2
◦
around 450 C) does not accelerate soot oxidation and can be evolved under N flow but not under a NOx/O stream. NOx storage on TiO
2 2 2
(
and ZrO under reaction conditions was not detected. An additional benefit of CeO with regard to TiO and ZrO is that the CeO -catalysed
2
2
2
2
2
soot oxidation yields mainly CO , whereas TiO and ZrO yield higher percentages of CO. This can be attributed to the lowest CeO -catalysed
2
2
2
2
◦
soot oxidation temperature and to the ability of CeO to catalyse CO oxidation to CO . The effect of thermal sintering at 800 C (decreased BET
surface area and increased crystal size) on CeO is more important than on TiO and ZrO , but 800 C-calcined CeO is still more active for soot
2
2
◦
2
2
2
2
combustion than the other oxides studied.
2007 Elsevier Inc. All rights reserved.
©
Keywords: Diesel exhausts; Soot; Metal oxide catalyst; CeO ; TiO ; ZrO ; NOx
2
2
2
1
. Introduction
Vehicles used in heavy-duty applications incorporate diesel
result in higher NOx emissions [3] and emanation of particulate
matter comprised primarily of carbon, on which heavy hydro-
carbons, sulfates, and water are adsorbed. NOx and particular
matter have been linked to serious environmental and health
problems.
Because the NOx and particulate emissions cannot be
avoided by engine modifications alone, postcombustion cat-
alytic processes for reducing the emission of both harmful sub-
stances must be developed. Selective catalytic reduction (SCR)
and NOx-storage catalysts have been proposed for NOx reduc-
tion. Catalytic regeneration of diesel particulate filters for soot
removal is also promising. A number of catalysts have been
investigated for pollution control in diesel exhausts, including
Pt [4], molten salts [5], perovskites [6], and different metal ox-
ides [7,8] among others.
engines. This includes buses, large trucks, and off-highway con-
struction mining equipment. Furthermore, diesel engines are
winning an increasing share of the worldwide light-duty vehi-
cle market [1]. Currently in Europe, about 100% of heavy duty,
6
0% of light-duty commercial vehicles, and 20% of passenger
cars are diesel-powered. The popularity of the diesel engine re-
volves around its fuel efficiency, reliability, and durability. The
high compression ratios and relatively high O concentrations
in the diesel combustion chambers are responsible for the better
fuel efficiency and lower CO and hydrocarbon emissions com-
pared with a gasoline engine [2]. However, these same factors
2
*
A soot oxidation catalyst (SOC) usually combines highly ac-
tive metals or metal oxides with different supports, including
Corresponding author. Fax: +34 965 90 34 54.
E-mail address: agus@ua.es (A. Bueno-López).
0021-9517/$ – see front matter © 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2007.05.015

76
I. Atribak et al. / Journal of Catalysis 250 (2007) 75–84
zeolites and metal oxides, among others. In many cases, the
role of the support in the catalytic activity of the metal is cru-
cial. For example, Oi-Uchisawa et al. [9] studied soot oxidation
under simulated diesel exhaust conditions with Pt supported on
different oxides and found that activity decreased in the order
Ta O > Nb O ≈ WO ≈ SnO ≈ SiO > TiO > Al O ≈
2. Experimental
2.1. Catalyst preparation
CeO2 and ZrO2 samples were prepared by precipitation of
the corresponding hydroxides (at pH 9) from aqueous solu-
tions of Ce(NO3)3·6H2O and ZrO(NO3)2·6H2O, respectively
2
5
2
5
3
2
2
2
2
3
ZrO2. Pt supported on zeolites (Na-Y, Ba-Y, Ferrierite, ZSM-
(
precursors provided by Aldrich, 99.9% purity), by dropping a
2
2, ETS-10, and AlPO-11), Al2O3, and TiO2 were also tested
diluted water solution of NH3. The hydroxides were dried in
for soot oxidation under simulated diesel exhaust conditions,
and a significant effect of the support also was reported in
this case [10]. Neeft et al. [11] studied catalysed soot oxida-
◦
◦
air at 110 C overnight and further calcinated at 600 or 800 C
for 2 h. TiO2 samples were prepared by thermal treatment of
TiO2 (P25, Degussa) in air for 2 h at the same temperatures.
tion by O with two catalysts containing a mixture of metals
2
a
The samples are designated MO2-T , where M = Ti, Zr, or Ce
[
Cu/K/Mo/(Cl)] supported on TiO2 and ZrO2, respectively, the
a
◦
and T = 600 or 800 C.
former being more active than the latter, and Braun et al. [12]
demonstrated that Mo/SiO2 is more active than Mo/TiO2 for
soot oxidation in air.
2
.2. Catalyst characterisation
On the other hand, the role of the support in catalysts is also
crucial to NO removal. Recently, a series of Pt/MO and Pt–
BET surface area of samples was determined by physical
x
2
◦
adsorption of N2 at −196 C in an automatic volumetric sys-
Ba/MO2 (M = Ce, Si, Zr) catalysts was investigated in terms
of NOx storage and subsequent reduction of the stored NOx
species by propene. Among the Ba-free catalysts, Pt/CeO2 ex-
hibited the highest NOx uptake, whereas Pt/SiO2 demonstrated
no NOx uptake [13]. A study of the chemisorptive properties
of Pt/Al2O3, Pt/CeO2, and Pt/ZrO2 catalysts and of the bare
supports toward NO and propylene concluded that the rela-
◦
tem (Autosorb-6B, Quantachrome) after degassing at 250 C
for 4 h. Raman spectra were recorded in a LabRam Jobin-Yvon
Horiba) spectrograph coupled to a microscope. A He/Ne laser
(
(
λ = 632.82 nm) source was used as the light source, and a
Perltier cooled CCd served as the detector. A power of approx-
imately 0.9 mW reached the sample through a ×50 VLWD
(
n/a = 0.50) objective for 3 s, and 15 consecutive spectra were
tive population and thermal stability of adsorbed NO species
x
added to obtain the final spectrum. Several expressions have
depended on the nature of the metal oxide used as the sup-
port [14]. Pt/ZrO2 promotes NO dissociation and nitrite/nitrate
decomposition at lower temperatures compared with Pt/Al2O3
and Pt/CeO2 catalysts. Investigation of the catalytic perfor-
mance of the three catalysts for the SCR of NO by propy-
lene revealed that the Pt/ZrO2 catalyst was much more active
than Pt/CeO2, with Pt/Al2O3 exhibiting an intermediate perfor-
mance [14].
But despite the decisive role of the supports demonstrated in
many cases, few studies have focused on comparisons of cat-
alytic activity of single oxides for soot oxidation under diesel
exhaust conditions. Van Doorn et al. [15] studied the catalytic
been proposed to estimate the ratio of tetragonal and mono-
clinic phases in ZrO samples, relying on the bands at 148,
2
−
1
−1
267 cm (tetragonal only), and the doublet at 180–190 cm
(due to monoclinic only) [17–19]. The most commonly used
expression, and also the one used in this study, is
ꢀ
ꢁ
−
1
−1
)
Ft = 0.97 It(148 cm ) + It(267 cm
ꢀ
−
1
−1
−1
)
× It(148 cm ) + It(267 cm ) + Im(180 cm
ꢁ
−1 −1
+
Im(190 cm
)
(1)
where Ii is taken to be the integral intensity of the band and no
polarization effects are considered [19–21].
X-ray diffractograms were measured on a Seifert powder
diffractometer using the CuKα radiation (λ = 0.15418 nm).
role of several metal oxides in soot combustion by O and
2
concluded that Al2O3 and SiO2 had no catalytic effect; TiO2
and ZrO2 had moderate activity, and CeO2, La2O2CO3, and
V2O5 exhibited substantial activity for soot combustion. Nu-
merous metal oxide catalysts for soot oxidation by O2 were also
screened by Neeft et al. [16], but those studies were carried out
◦
◦
Spectra were recorded between 10 and 60 (2θ) with a step
size of 0.05 and measuring for 3 s at each step. From XRD data,
different phases of the studied compounds were identified and
quantified. Phase composition on TiO2 samples was calculated
using
without NO in the gas mixture, and thus the conclusions can-
x
1
.26IR(110)
not be directly extrapolated to diesel exhaust conditions.
Considering these premises, the aim of the present study
was to compare the catalytic activity for soot oxidation under
FR =
(2)
IA(101) + 1.26IR(110)
where F is the rutile fraction, I (101) is the intensity of the
R
A
NO /O of three metal oxides (TiO , ZrO , and CeO ) com-
most intense anatase peak, and I (110) is the intensity of the
x
2
2
2
2
R
monly used as catalyst supports and even as catalysts in some
cases. These oxides were selected because they present differ-
most intense rutile peak [22,23].
Phase composition determination of ZrO samples was per-
2
ent activities for soot oxidation, with CeO the most active
formed in a similar manner, and the following equation presents
2
oxide and ZrO2 the least active oxide considering the studies
noted in this section [9,11]. Taking into account the highest ac-
tivity of CeO2, special attention was paid to the behaviour of
this oxide.
the expression to calculate the fraction of tetragonal phase from
XRD patterns:
It(111)
Ft =
(3)
It(111) + Im(−111) + Im(111)

I. Atribak et al. / Journal of Catalysis 250 (2007) 75–84
77
where Ft is the tetragonal fraction, It(111) is the intensity of
the tetragonal peak (111), and Im(−111) and Im(111) are the
intensities of the monoclinic peaks (−111) and (111), respec-
tively.
The crystal sizes (D) of all the phases present in the catalysts
were estimated using Scherrer’s equation
equipped with a switched dual-column system and two serial
columns (a Porapak Q 80/100 for CO2 and N2O separation and
a Molecular Sieve 13X for O2, N2, and CO separation) and a
Chemiluminiscence NOx analyser for NO and NO2 analysis
(SIGNAL 400VM).
The model soot used in this study was carbon black from
Degussa S.A. (Printex-U), with 92.2% C, 0.6% H, 0.2% N, and
0.4% S. The ash and adsorbed hydrocarbon percentages were
Kλ
D =
(4)
β cosθ
<
0.1 and 5.2%, respectively, and the BET surface area was
2
where λ is the wavelength of the radiation used (CuKα), β is the
full width at half maximum of the diffraction peak considered,
K is a shape factor taken as 0.9 (1 being a perfect sphere), and
θ is the diffraction angle at which the peak appears.
95 m /g.
3
. Results
3
.1. Characterization of materials
2
.3. Catalytic activity
BET surface areas of the samples used in this study are
displayed in Table 1. TiO2-600 and ZrO2-600 had the same
BET area (47 m /g), with CeO2-600 exhibiting a lower value
Reactivity tests were carried out at atmospheric pressure in a
2
cylindrical fixed-bed reactor (1 cm of inner diameter) connected
to specific NDIR-UV gas analysers for NO, NO , CO, CO , and
2
◦
(
37 m /g). As expected, the samples calcined at 800 C had
2
2
◦
lower surface area values than those samples calcined at 600 C
due to sintering. Calcination temperature had a greater effect on
CeO2 compared with TiO2 and ZrO2; 84% of the area of CeO2
was lost when calcined at 800 C compared with that at 600 C,
with comparable values of 57% for ZrO2 and 66% for TiO2.
All samples were characterised by Raman spectroscopy; the
O2 (Fisher–Rosemount, models BINOS 1004, 100 and 1001).
Soot-catalysts mixtures (80 mg of catalyst + 20 mg of soot)
were prepared in loose contact [11] and diluted with 350 mg of
SiC. The addition of SiC decreases the flow resistance of the
sample bed and provides a heat sink, which decreases temper-
ature gradients in the sample. The reactive gas mixtures used
◦
◦
data are collected in Fig. 1. Raman spectra of TiO2 calcined at
−
1
(
5
500 ml/min, GHSV ∼30,000 h ) contained 0.05% NOx +
% O2 + balance N2 or 5% O2 + balance N2. Temperature-
programmed reactions (TPRs) were performed by raising the
◦
6
00 and 800 C are shown in Fig. 1a. TiO2 is known to appear
mainly in two different crystal structures (besides brookite): ru-
tile and anatase (designated R and A in Fig. 1a). The Raman
◦
◦
temperature from 25 to 700 C at a rate of 10 C/min. In addi-
◦
spectrum of TiO2 calcined at 800 C was consistent with the
◦
tion, isothermal reactions at 450 C were carried out, consisting
of heating the sample from room temperature up to 450 C
spectrum of pure rutile, even though the presence of anatase
cannot be ruled out. TiO2-800 showed bands at 142, 235, 446,
◦
under N2 flow, after which the inert gas was replaced by the
reactive mixture. Isothermal reactions were conducted until the
constant level of the different gases was reached.
Soot conversion profiles were determined from the CO and
CO2 evolved, and the selectivity of the different catalysts for
CO emission was determined using
−1
−1
and 608 cm . The bands at 142, 446, and 608 cm can be
assigned to the B1g, Eg, and A1g modes of rutile, respectively
−1
[
24,25]. The broad band at 234 cm has been identified as a
second-order phonon that disappears at low temperature [26].
The spectrum of TiO2-600 presented bands at 144, 394, 517,
−1
and 637 cm , which can be assigned to Eg, B1g, A1g, and
Eg of anatase, respectively [27]. TiO2-600 also presented fea-
CO/COx (%) = 100CO/(CO + CO2).
Blank experiments were also performed under the aforemen-
tioned experimental conditions but without soot, that is, using
only the catalyst.
(5)
tures of the Raman spectrum of rutile appearing as shoulders
−
1
at 446 and 607 cm , indicating a mixture of both rutile and
anatase phases on sample TiO -600, even though estimating
2
phase proportion by Raman spectroscopy for these materials is
In addition, TPRs were performed with selected catalysts
under the described conditions but following the gas compo-
sition with a gas chromatograph (HP model 6890 Plus Series)
not straightforward.
The Raman spectra of zirconia samples calcined at 600 and
◦
800 C is shown in Fig. 1b. When ZrO2 is synthesised by a pre-
Table 1
Phase composition, crystal sizes, and BET surface area of the catalysts
2
Sample
TiO -600
Phase (fraction)/particle size (nm)
SBET (m /g)
Rutile (0.36)/32; Anatase (0.64)/21
Rutile (1)/45
47
16
47
20
37
6
2
TiO -800
2
a
b
a
b
ZrO -600
Tetragonal (0.56 ; 0.50 )/13; Monoclinic (0.44 ; 0.50 )/11
2
a
b
a
b
ZrO -800
Tetragonal (0.24 ; 0.22 )/17; Monoclinic (0.76 ; 0.78 )/15
2
CeO -600
Fluorite (1)/18
Fluorite (1)/39
2
CeO -800
2
a
Phase fraction calculated by XRD.
Phase fraction calculated by Raman spectroscopy.
b

78
I. Atribak et al. / Journal of Catalysis 250 (2007) 75–84
Tetragonal zirconia presented bands at 148, 267, 322, 464,
−
1
and 648 cm . Some of these bands were unique, but most of
them overlapped significantly with bands due to the monoclinic
phase of zirconia, which presented bands a doublet at 180 and
1
7
90, 333, 380, 475, 502, a doublet at 539 and 559, 618, and
57 cm , some of which are not resolved in Fig. 1b due to the
−
1
analysis conditions used. The values obtained for the fraction
of tetragonal and monoclinic phase, along with those obtained
by XRD, are summarized in Table 1.
◦
The samples of CeO2 calcined at 600 and 800 C (Fig. 1c)
−1
presented a Raman characteristic band at about 465 cm
,
which can be unambiguously assigned to the F2g mode of
fluorite-type structures [30–32]. Band position did not change
when calcination temperature was modified; however, an in-
crease in the calcination temperature produced a significant
increase in the intensity of the F2g band and narrowing of the
band. Thus, increased crystallinity of the material with increas-
ing calcination temperature can be inferred.
Samples were also characterized by XRD; Fig. 2 collects
XRD patterns of all of the materials used in this study. XRD
◦
patterns of TiO2 calcined at 600 and 800 C are shown in
Fig. 2a. Sample TiO2-600 comprised a mixture of anatase and
rutile phases, whereas TiO2-800 did not show the main features
of anatase [A(101) and A(200)]. These results are congruent
with the observations from the Raman spectra of TiO2-600
and TiO2-800. The results of phase composition analysis, to-
gether with crystal size of each phase [calculated from Scher-
rer’s equation, using R(110) and A(101) peaks] are compiled
in Table 1. Narrower rutile peaks can be seen at higher calcina-
tion temperatures, and thus an increase in crystallite size can be
deduced.
XRD diffractograms of zirconia samples show that both
samples used in this study are mixtures of monoclinic and
tetragonal phases. The most important features of the diffrac-
◦
◦
togram are seen between 20 and 40 (2θ), with two intense
peaks [m(−111) and m(111)] due to the monoclinic phase and
one intense peak due to the tetragonal phase [t(111)]. From
Fig. 2b, it can be concluded that at the higher calcination tem-
perature (ZrO2-800), monoclinic phase was slightly more abun-
dant than tetragonal phase, whereas the opposite situation was
encountered at the lower temperature (ZrO2-600). As in the
case of TiO2, peaks corresponding to the monoclinic phase of
◦
ZrO2 calcined at 800 C were narrower, and, as shown in Ta-
ble 1, crystal sizes were larger.
XRD analysis of CeO2 is shown in Fig. 2c. No other phase
than ceria with a fluorite-type structure was detected; however,
crystal size changes can be seen in the diffractograms, with nar-
rower peaks in CeO2-800 compared with those in CeO2-600
are. Data on crystal size are compiled in Table 1. In general, in-
creased calcination temperature led to a significant drop in the
surface area of CeO2 and a consequent increase in crystal size
Fig. 1. Raman spectra of (a) TiO , (b) ZrO , and (c) CeO catalysts.
2
2
2
cipitation route, the predominant phase will be monoclinic or
tetragonal, depending on calcinations conditions (mainly tem-
perature, time, and atmosphere) [28]. It is quite common to
obtain materials in which a mixture of monoclinic and tetrago-
nal phases appears. The Raman spectra of ZrO2 shown in this
paper exhibit features of both monoclinic and tetragonal phases
(54%).
(
m and t, respectively, in Fig. 1b). The tetragonal phase is char-
3
.2. Catalytic activity
15
acterized by six Raman active modes (D , A1g +2B1g +3Eg),
4
h
5
whereas for the monoclinic phase (C ), 18 active Raman
Fig. 3 shows the soot conversion profiles obtained in TPRs
2h
modes are expected from symmetry analysis [29].
performed with TiO2, ZrO2, and CeO2 catalysts under NOx/O2,

I. Atribak et al. / Journal of Catalysis 250 (2007) 75–84
79
Fig. 3. Soot conversion in temperature programmed reactions under NOx/O
2
Fig. 2. XRD patterns of (a) TiO , (b) ZrO , and (c) CeO catalysts.
(solid symbols) with different catalysts. (a) TiO , (b) ZrO , and (c) CeO . (The
2
2
2
2 2 2
experiment represented by open symbols was carried out under O .)
2
along with the profile of the uncatalysed reaction. The cor-
responding temperatures for 50% soot conversion (T50%) are
included in Table 2. All of the metal oxides tested decreased
the soot oxidation temperature, with differences among them
observed. CeO2 catalysts were more active than their counter-
part ZrO2 and TiO2 catalysts, and in general, catalysts calcined
exhibiting less activity for soot oxidation than under NO /O
x
2
atmosphere, significant activity was maintained. The high ac-
tivity of CeO for soot combustion with O was attributed to the
2
2
production of active oxygen, that is, gas-phase O was incorpo-
2
rated to the CeO lattice, and atomic oxygen from the lattice
2
◦
◦
at 600 C were more active than those calcined at 800 C. The
was delivered to soot [7,8]. However, in the presence of NOx,
most active catalyst, CeO2-600, lowered the T50% parameter at
additional factors play important roles, as we discuss next.
◦
5
5 C with regard to the uncatalysed reaction. The catalytic ac-
The analysis of the CO and CO emission during the TPRs
2
tivity of CeO2-600 also was checked under O2 (Fig. 3c); despite
provides complementary information. From a practical stand-

80
I. Atribak et al. / Journal of Catalysis 250 (2007) 75–84
Table 2
Temperature for 50% soot conversion (T
) and selectivity to CO (determined
50%
with Eq. (5)) in temperature programmed reactions performed under NOx/O
2
◦
Sample
TiO -600
T
( C)
CO/COx (%)
5
0%
593
598
592
600
551
581
606
56.6
60.2
43.4
62.6
15.2
67.5
65.5
2
TiO -800
2
ZrO -600
2
ZrO -800
2
CeO -600
2
CeO -800
2
Uncatalysed
point, CO2 formation is preferable due to the high CO tox-
icity. The percentages of CO emitted during the TPRs per-
formed under NOx/O2 were calculated according to Eq. (5) and
are included in Table 2. The uncatalysed carbon gasification
yielded both CO and CO2, with CO the most abundant compo-
◦
nent (65.5%). All of the catalysts calcined at 600 C decreased
CO formation, and among these, CeO2-600 yielded the low-
◦
est amount of CO. In contrast, the catalysts calcined at 800 C
hardly improved CO2 formation with regard to CO.
NOx elimination profiles obtained with the series of catalysts
studied are given in Fig. 4. Most of the catalysts exhibited a sin-
gle band, and the temperatures of maximum NOx elimination
and of the onset of NOx elimination corresponding to catal-
ysed reactions decreased with regard to those for uncatalysed
reactions. The sample CeO2-600 showed unique behaviour; the
NOx elimination profile obtained with this catalyst was com-
◦
◦
plex and ranged from approximately 225 C to above 700 C.
This broad profile suggests that several processes are occurring,
as we discuss later.
Considering the temperatures of maximum NO elimination
x
◦
with the series of oxides calcined at 600 or 800 C, the activ-
ity for NOx elimination increases in the order ZrO2 ꢀ TiO2 <
CeO2, in agreement with the trend of activity for soot oxida-
tion deduced from Fig. 3 and from the T50% values included in
Table 2.
O2 conversion profiles corresponding to selected TPRs per-
formed under NOx/O2 are plotted in Fig. 4 together with those
of NOx elimination. Fig. 4a shows quite parallel NOx and
O2 elimination profiles during the TiO2-600-catalysed reaction,
with the NOx elimination onset temperature slightly lower than
the O2 elimination onset temperature. Quite similar behaviour
can be observed for the catalyst ZrO2-600 shown in Fig. 4b;
that is, the NOx and O2 elimination profiles are concurrent.
This type of behaviour, with parallel O2 and NOx elimination
curves, was also observed in the reactions performed with all
Fig. 4. NOx elimination in temperature programmed reactions under NO /O
x
2
(solid symbols) and O2 elimination in some of these experiments (open sym-
bols). Catalysts: (a) TiO , (b) ZrO , and (c) CeO .
2
2
2
◦
of the catalysts calcined at 800 C (O2 profiles not included for
brevity). This suggests that in all of these experiments, NOx and
O2 were consumed due to reduction with soot once the reaction
temperature for its gasification was achieved. The NOx elim-
ination profiles obtained with TiO2-600 and ZrO2-600 were
slightly different, whereas the soot conversion profiles (Fig. 3)
corresponding to these experiments were very similar. In the
experiment performed with TiO -600, NO elimination at ca.
reduction with regard to O2 combustion than ZrO2-600 in this
temperature range.
The CeO2-600 catalyst exhibited a unique behaviour, with
NOx and O2 elimination profiles not following parallel trends.
As mentioned earlier, NOx elimination occurred within a wide
◦
◦
range of temperatures, from ca. 225 C to >700 C. In con-
trast, the corresponding O2 elimination curve exhibited a single
2
x
◦
◦
4
50–500 C was higher than that using ZrO2-600 as a catalyst,
peak between 325 and 700 C. These differences between O2
suggesting that TiO2-600 is slightly more selective toward NOx
and NOx elimination profiles indicate that, in addition to the

I. Atribak et al. / Journal of Catalysis 250 (2007) 75–84
81
Fig. 5. Blank temperature programmed reactions (without soot) performed with
◦
Fig. 6. Isothermal experiments at 450 C under NOx/O2 performed with
◦
catalysts calcined at 600 C. (a) NO conversion to NO and (b) NOx retention
2
CeO -600: (a) catalysed soot oxidation and (b) blank experiment.
2
on catalysts. (The profile obtained with the catalyst CeO -600 from the soot
2
oxidation experiment is included for comparison.)
same than that measured in the absence of catalyst (Fig. 5a).
In the same way, the total concentration of NOx was constant
throughout the experiments performed with these two oxides,
and their NOx elimination profiles (included in Fig. 5b) ex-
hibited the background level. In contrast, CeO2-600 was quite
◦
aforementioned NOx and O2 reduction by soot at ca. 550 C,
an additional NOx elimination pathway was present at lower
temperatures. Finally, note that a small amount of NOx was ad-
sorbed on CeO2-600 at room temperature, and then released
◦
◦
effective for NO oxidation to NO2 from 300 C (Fig. 5a). The
between 150 and 200 C, as shown on Fig. 4c.
NO2 profile obtained with this catalyst increased with tem-
perature until the thermodynamic equilibrium of Eq. (6) was
fulfilled, and then decreased at higher temperatures following
thermodynamics. The highest NO2 concentration was reached
In an attempt to provide some insight into the NOx elimina-
tion processes occurring in the catalysed soot oxidation experi-
ments, we carried out blank TPRs (without soot) with the three
◦
catalysts calcined at 600 C. The NO percentage in the out-
2
◦
let gas mixture, calculated on the basis of the inlet NOx level,
is plotted as a function of temperature in Fig. 5a, and the NOx
elimination profiles are included in Fig. 5b. It is important to
note that the NO2 level in the gas stream is about zero at the en-
trance of the reactor, with NO being the main component of the
NOx binary mixture. NO can be oxidised to NO2 by O2 accord-
ing to
at 450 C. This ability of CeO2 in NO2 production was reported
previously [33]. In addition, NOx is stored on CeO2-600 above
◦
ca. 250 C, as shown in Fig. 5b. Physical adsorption of NOx
can be ruled out in this temperature range, and chemical NOx–
CeO2 interaction is expected. NOx desorption did not occur
◦
even though the catalyst was heated up to 700 C. Fig. 5b also
compiles the NOx elimination profile corresponding to the soot
oxidation in TPRs performed with the CeO2-600 (previously
included in Fig. 4c). Surprisingly, NOx elimination decreased
in the presence of soot with regard to the blank experiment.
NO + 1/2O2 ꢁ NO2.
(6)
The predicted NO2 level considering the thermodynamic
equilibrium of this reaction is represented in Fig. 5a as a dotted
line.
◦
Isothermal soot oxidation at 450 C was performed under
NOx/O2 with the CeO2-600 catalyst, and a blank experiment
(using the same catalyst but without soot) was performed under
similar conditions. The soot conversion and NOx and O2 elim-
TiO2-600 and ZrO2-600 exhibited no activity for NO2 pro-
duction, and the NO2 level reached with these oxides is the

82
I. Atribak et al. / Journal of Catalysis 250 (2007) 75–84
Fig. 8. CeO -600-catalysed soot oxidation by O in temperature programmed
reactions performed with fresh catalyst and with the pre-treated catalyst.
Fig. 7. Thermal treatment of CeO -600 under N flow after the isothermal ex-
2
2
2
2
◦
periment at 450 C under NOx/O (previously illustrated in Fig. 6b).
2
Fig. 8 shows the soot conversion profile thus obtained, along
with the curve corresponding to a similar experiment carried out
with fresh catalyst. As can be deduced from Fig. 8, NOx stored
on CeO2 at 450 C did not improve soot conversion, which oc-
curred at even slightly higher temperature with the pretreated
catalyst compared with the fresh catalyst.
ination curves are illustrated in Fig. 6a, and those obtained in
the blank experiment are shown in Fig. 6b. The soot conversion
percentages (Fig. 6a) increased with time until total consump-
◦
tion, and both NO and O were eliminated from the gas stream
x
2
during soot gasification. In this case, CO2 was the only car-
bon product; CO emission was not detected. Special attention
should be given to the NO elimination curve, which showed
two peaks, one appearing at the beginning of the soot oxidation
experiment and the second appearing once the soot consump-
x
4
. Discussion
4
.1. Comparison of MO2 catalysts
tion was almost complete. In contrast, O elimination reached
2
a maximum level early and then decreased progressively un-
til zero. Note that little O elimination occurred after total soot
It is generally known that surface area is of great importance
2
conversion.
in gas–solid catalysed reactions and that it becomes even more
importantce in reactions in which catalyst (oxide) and one of
the reactants (soot) must be in close contact. External surface
area is the factor mainly responsible for the BET areas of the
materials used in this study; thus, a relationship between sur-
face area and particle size is common. For CeO2, the results of
the present study suggest that the relationship between the BET
The NO and O elimination profiles were different in the
x
2
blank experiment. In this case, a maximum elimination of both
NOx and O was reached after a few minutes, after which
2
the levels progressively decreased to zero. The maximum NOx
elimination level reached in the blank experiment exceeded the
maximum level reached during the catalysed soot oxidation re-
action, whereas the opposite situation was observed for the O2
profiles.
2
surface areas of CeO2-600 and CeO2-800 (37 and 6 m /g, re-
◦
spectively) and their catalytic activity (T50% = 551 and 581 C,
Once the blank experiment was finished, the NOx/O2 mix-
ture was replaced by N2 and the temperature was raised up to
respectively) is not directly related to the soot-catalyst contact.
As can be deduced from the catalytic tests, the high activity of
CeO2-600 for soot oxidation can be attributed to the conversion
of NO to NO2; for this reason, high surface area is preferable.
Note that NO2 had a much greater oxidizing effect than either
◦
9
00 C. As observed in Fig. 7, NO and NO2 released, indicat-
ing that the previous NOx chemisorption is reversible and NOx
stored on CeO2-600 is further recovered under inert atmosphere
◦
(
but not under NOx/O2, as mentioned).
NO or O2. Calcination of CeO2 at high temperature (800 C)
◦
Finally, after a blank experiment at 450 C under NOx/O2
had a negative effect on its activity with regard to calcination
◦
(
similar to that included in Fig. 6b), the gas flow was replaced
at 600 C due to sintering, thereby decreasing the BET surface
by N2 and the furnace was cooled down quickly (after 20 min
the temperature was below 250 C). The gas composition was
monitored during this time and evidences of emission of NO
area and also affecting some other properties of the oxide (e.g.,
oxygen storage capacity, oxygen mobility) that could be related
to NO2 production [33,34]. Calcination of CeO2 at 800 C also
◦
◦
or NO were not observed. Once the furnace reached room
temperature, the CeO2-600 was removed from the reactor and
mixed with soot. A TPR was then performed with this pre-
decreased its selectivity for CO2 formation, instead of CO, as a
soot oxidation product.
2
Despite the fact that a catalyst’s BET surface area affects its
catalytic activity, the nature of the catalyst is a more decisive
factor than BET when MO2 (M = Ti, Zr, Ce) oxides are com-
treated catalyst using the gas mixture O /N . This experiment
2
2
was done to determine the eventual contribution of the NOx
◦
2
species stored on the catalyst at 450 C to the soot oxidation
pared. For instance, CeO2-600, with 37 m /g, is much more
process.
active for soot oxidation than TiO2-600 or ZrO2-600, both with

I. Atribak et al. / Journal of Catalysis 250 (2007) 75–84
83
2
4
7 m /g. This is because CeO2 is able to catalyse the NO oxi-
It is known that CeO2 is a very effective catalyst for this re-
action through oxidation using lattice oxygen [36]. The results
obtained here suggest that TiO2 is more active than ZrO2 in
catalysing reaction (7).
dation to NO2, as mentioned, whereas TiO2 and ZrO2 are not.
The effect of calcination temperature on TiO2 and ZrO2
catalytic activity for soot oxidation is more complicated than
that described for CeO2. The phase composition of TiO2 and
ZrO2 depends on the calcination temperature. Rutile phase is
favoured for TiO2 calcined at high temperature; monoclinic
phase is favoured for ZrO2 calcined at high temperature. Taking
into account that the differences in catalytic activity of TiO2-
4.2. On the mechanism of CeO2-catalysed NOx-soot reaction
Depending on the reaction temperature, CeO2 can partici-
pate in the CeO2-catalysed NOx-soot reaction mechanism in
different ways. At low temperature, CeO2 adsorbs NOx, as can
be deduced from the inset of Fig. 4c. It has been reported [32]
6
00 vs TiO2-800 and ZrO2-600 vs ZrO2-800 are minor, it is
possible to conclude that the phase composition does not play a
crucial role on the activity of TiO and ZrO for soot oxidation.
◦
that NO2 stored at low temperature (200 C [33]) can be decom-
2
2
In the case of TiO2, this observation was already reported [35].
Taking this idea into consideration, the minor differences in ac-
tivity between TiO -600 and TiO -800 and between ZrO -600
posed at higher temperatures, and that the NO2 thus evolved
contributes to the accelerated oxidation of soot. But in a con-
tinuously regenerated soot trap, the temperature must be main-
tained above 200 C; otherwise, soot is not converted to CO2.
Therefore, NOx storage at low temperature would not be ex-
2
2
2
◦
and ZrO2-800 could be attributed to the extent of soot-catalyst
contact, which is related to their BET surface area. Because
TiO and ZrO do not catalyse the NO oxidation to NO , the
pected to play a role in an actual trap.
2
2
2
◦
effect of surface area on this reaction can be ruled out.
With regard to NO and O elimination, both gases were
At higher temperatures (above 250 C), CeO2 catalyses the
oxidation of NO to NO2 by Eq. (6), as shown in Fig. 5a. In the
presence of soot, part of the NO2 produced reacts with soot and
part is chemisorbed on the oxide. The high efficiency of CeO2
for NO2 production is the main reason for the higher activity
of this oxide for soot oxidation compared with TiO2 or ZrO2.
The competition between NO2 adsorption on CeO2 and NO2
reaction with soot can be clearly seen in Fig. 6. The amount of
NOx eliminated at the beginning of the soot oxidation isother-
x
2
consumed during the uncatalysed soot oxidation reaction.
This process is not selective toward NOx reduction; above
◦
ca. 500 C, O2 oxidises soot to CO and CO2, and NOx is re-
duced. TPRs followed by gas chromatography were performed
without catalyst and with the CeO -600 catalyst; the results
2
show that N2 was the product of NOx reduction; formation
of N2O was not observed. TiO2 and ZrO2 catalysts exhibited
O2 and NOx elimination profiles qualitatively similar to those
observed in the uncatalysed process. NOx was reduced, along
◦
mal reaction at 450 C (first peak of the NOx profile in Fig. 6a)
was less than the amount of NOx eliminated in the absence of
soot (Fig. 6b). In the presence of soot, part of the NO2 produced
by CeO2 became NO due to soot oxidation and was not avail-
able for retaintion on CeO2. However, once the soot was almost
consumed, a second retention peak occurred (Fig. 6a), because
although NO2 consumption by soot no longer occurred to any
great extent, the maximum NOx storage capacity of CeO2 was
not yet fulfilled. Another difference in behaviour among the
three MO2 oxides studied under NOx/O2 mixtures was the lack
of NOx accumulation on ZrO2 and TiO2.
◦
with a massive O2 consumption above ca. 500 C, indicating
that these catalysts are not selective for NOx reduction by soot.
The NOx elimination profiles obtained with the CeO2 catalysts
(
mainly CeO2-600 and, to a much lower extent, CeO2-800)
are quite different from those of TiO2 and ZrO2. Along with
◦
the aforementioned NOx reduction above 500 C, another NOx
elimination pathway occurred at lower temperature. We discuss
this in detail in the next section.
Finally, an additional difference among the MO catalysts
2
studied here is the different selectivity for the formation of CO2
instead of CO as the soot oxidation product. This can be seen
in the data presented in Table 2. A general feature of carbon
The NOx stored on CeO2 during catalysed soot oxidation
under a NOx or O2 stream at 450 C did not contribute to soot
◦
oxidation. As can be deduced from Fig. 8, the fresh CeO2-600
for soot oxidation by O2 and the same catalyst pretreated un-
combustion reactions is that the CO/CO ratio increases with
2
◦
temperature, because CO2 is preferred at low temperature and
CO is preferred at high temperature. Considering this idea, the
highest selectivity of CeO -600 for CO formation could be due
der NOx/O2 at 450 C had about the same catalytic activity. In
addition, the results obtained in this study do not demonstrate
the amount of NOx stored on CeO2 released under NOx/O2
atmosphere. This is supported by the fact that the NOx elim-
ination profiles shown in Fig. 5b corresponding to CeO2-600
did not reach negative values within the range of temperatures
studied. The amount of NOx stored on CeO2 only was released
under a NOx-free gas stream, as shown in Fig. 7, which includes
NOx evolution profiles obtained under N2.
2
2
on one hand to its highest activity for soot oxidation. CeO2-
00-catalysed soot combustion occurs in the lowest tempera-
6
ture range, and thus, CO formation is the lowest one. However,
this general argument is not valid for explaining some of the
data given in Table 2. For instance, TiO2-600 and ZrO2-600
◦
had about the same T50% value (593 and 592 C, respectively),
whereas their selectivity for CO2 formation differed dramati-
cally (CO = 56.6 and 43.4%, respectively). This is because of
the differing activities of the catalysts in accelerating the reac-
tion
5. Conclusions
Based on our comparison of the catalytic activity of TiO2,
ZrO2, and CeO2 for soot oxidation under NOx/O2, we can draw
the following conclusions:
CO + 1/2O → CO .
(7)
2
2

84
I. Atribak et al. / Journal of Catalysis 250 (2007) 75–84
•
CeO2-600 is more active than TiO2-600 and ZrO2-600 for
soot oxidation because it accelerates the conversion of NO
to NO2, whereas TiO2 and ZrO2 do not catalyse this re-
action. The catalytic activity of all these three oxides de-
creases when calcined at 800 C, with CeO2 maintaining
the highest activity among the three oxides studied despite
the significant decrease in BET surface area.
[5] B.A.A.L. van Setten, J. Bremmer, S.J. Jelles, M. Makke, J.A. Moulijn,
Catal. Today 53 (1999) 613.
[6] S. Biamino, P. Fino, D. Fino, N. Russo, C. Badini, Appl. Catal. B 61 (2005)
2
97.
7] A. Bueno-López, K. Krishna, M. Makkee, J.A. Moulijn, Catal. Lett. 99
2005) 203.
[8] A. Bueno-López, K. Krishna, M. Makkee, J.A. Moulijn, J. Catal. 230
2005) 237.
[
◦
(
(
[9] J. Oi-Uchisawa, A. Obuchi, R. Enomoto, S. Liu, T. Nanba, S. Kushiyama,
•
•
The phase composition of TiO2 (rutile or anatase) and ZrO2
Appl. Catal. B 26 (2000) 17.
10] K. Villani, W. Vermandel, K. Smets, D. Liang, G. Van Tendeloo, J.A.
Martens, Environ. Sci. Technol. 40 (2006) 2727.
[11] J.P.A. Neeft, O.P. van Pruissen, M. Makkee, J.A. Moulijn, Appl. Catal.
B 12 (1997) 21.
12] S. Braun, L.G. Appel, M. Schmal, Appl. Surf. Sci. 201 (2002) 227.
13] M. Piacentini, M. Maciejewski, A. Baiker, Appl. Catal. B 72 (2007) 105.
14] A. Kotsifa, D.I. Kondarides, X.E. Verykios, Appl. Catal. B 72 (2007) 136.
15] J. van Doorn, J. Varloud, P. Meriaudeau, V. Perrichon, M. Chevrier,
C. Gauthier, Appl. Catal. B 1 (1992) 117.
(
monoclinic or tetragonal) has no significant affect on the
[
catalytic activity of these oxides. The main feature of these
oxides that affects their activity is BET surface area, which
may be related to better soot-catalyst contact.
[
[
[
[
During CeO2-catalysed soot oxidation, NOx is stored on
this oxide. This is not the case for TiO2- or ZrO2-catalysed
soot oxidation. However, NOx stored on CeO2 does not ac-
celerate soot oxidation. NOx stored on CeO2 evolves sub-
sequently under N2 flow but not under NOx/O2 flow.
During CeO2-catalysed soot oxidation under NOx/O2, once
NO2 is produced by CeO2, there is a competition between
the NO2-soot reaction and NOx storage on the catalyst.
CeO2-600 catalysed soot oxidation yields mainly CO2,
whereas TiO2-600- and ZrO2-600-catalysed soot oxidation
yield higher percentages of CO. This additional benefit of
CeO2 compared with the other oxides can be explained
considering two facts: (i) CeO2-catalysed soot oxidation
occurs at the lowest temperature, and (ii) CeO2 catalyses
[16] J.P.A. Neeft, M. Makkee, J.A. Moulijn, Chem. Eng. J. 64 (1996) 295.
[
[
[
17] E. Djurado, P. Bouvier, G. Lucazeau, J. Solid State Chem. 149 (2000) 399.
18] B.-K. Kim, J.-W. Hahn, K.R. Han, J. Mater. Sci. Lett. 16 (1997) 669.
19] L.D. Huy, P. Laffez, P. Daniel, A. Jouanneaux, N.T. Khoi, D. Simeone,
Mater. Sci. Eng. B 104 (2003) 163.
•
•
[
20] D.R. Clarke, J. Am. Ceram. Soc. 65 (1982) 284.
[21] M.S. Kaliszewski, G. Behrens, A.H. Heuer, M.C. Shaw, D.B. Marshall,
G.W. Dransmanri, R.W. Steinbrech, A. Pajares, F. Guiberteau, F.L. Cum-
brera, A. Dominguez-Rodriguez, J. Am. Ceram. Soc. 77 (1994) 1185.
22] C. Su, B.Y. Hong, C.M. Tseng, Catal. Today 96 (2004) 119.
23] R.A. Spurr, H. Myers, Anal. Chem. 29 (1957) 760.
24] S.R. Dhage, V.D. Choube, V. Samuel, V. Ravi, Mater. Lett. 58 (2004)
2310.
25] H. Cheng, J. Ma, Z. Zhao, L. Qi, Chem. Mater. 7 (1995) 663.
26] S.P.S. Porto, P.A. Fleury, T.C. Daman, Phys. Rev. 154 (1967) 522.
27] S.J. Rigby, A.H.R. Al-Obaidi, S.-K. Lee, D. McStay, P.K.J. Robertson,
Appl. Surf. Sci. 252 (2006) 7948.
28] L.A. Cortez-Lajas, J.M. Hernández-Enríquez, A. Castillo-Mares, J.L.
Rivera-ARmetna, G. Sandoval-Robles, L.A. García-Serrano, R. García-
Alamilla, Rev. Mexicana Ing. Quím. 5 (2006) 321.
[
[
[
◦
the oxidation of CO to CO2. CeO2 calcination at 800 C
◦
[
[
[
instead of 600 C has a negative affect on the selectivity of
this oxide toward CO2 production.
Acknowledgments
[
Financial support was provided by the Spanish Ministry of
Education and Science (project CTQ2005-01358) and a con-
tract of ABL (Ramon y Cajal Program), which is co-funded by
the Generalitat Valenciana and the University of Alicante.
[
[
[
29] D. Michel, M. Perez, Y. Jorba, R. Collonges, J. Raman Spectrosc. 5 (1976)
163.
30] L.N. Ikryannikova, A.A. Aksenov, G.L. Markaryan, G.P. Muraveva, B.G.
Kostyuk, A.N. Kharlanov, E.V. Lunina, Appl. Catal. A 210 (2001) 225.
31] A. Mineshige, T. Taji, Y. Muroi, M. Kobune, S. Fujii, N. Nishi, M. Inaba,
Z. Ogumi, Solid State Ionics 135 (2000) 481.
References
[
32] Y. Li, Q. Zhu, X. Zhang, B. Xu, J. Catal. 221 (2004) 584.
[
33] A. Setiabudi, J. Chen, G. Mul, M. Makke, J.A. Moulijn, Appl. Catal. B 51
[
1] R.H. Hammerle, D.A. Ketcher, R.W. Horrocks, G. Lepperhoff, G. Hüth-
wohl, B. Lüers, SAE Technical Paper Series, Paper Number 942043, 1994.
2] J.C. Summers, S. Van Houtte, D. Psaras, Appl. Catal. B 10 (1996) 139.
3] J.H. Johnson, S.T. Bagley, L.D. Gratz, D.G. Leddy, SAE Technical Paper
Series, Paper Number 940233, 1992.
(2004) 9.
[
[
[34] E. Rocchini, A. Trovarelli, J. Llorca, G.W. Graham, W.H. Weber, M. Ma-
ciejewski, A. Baiker, J. Catal. 194 (2000) 461.
[35] I. Atribak, I. Such-Basáñez, A. Bueno-López, A. García-García, Catal.
Commun. 8 (2007) 478.
[4] J. Oi-Uchisawa, S. Wang, T. Nanbaa, A. Ohi, A. Obuchi, Appl. Catal. B 44
(2003) 207.
[36] J. K a˘ spar, P. Fornasiero, M. Graziani, Catal. Today 50 (1999) 285.