Applied Catalysis B: Environmental 75 (2007) 189–200

www.elsevier.com/locate/apcatb

Potential rare earth modified CeO2 catalysts for soot oxidation

I. Characterisation and catalytic activity with O2
K. Krishna, A. Bueno-López, M. Makkee *, J.A. Moulijn
Catalysis Engineering, DelftChemTech, Delft University of Technology, Julianalaan 136, NL 2628 BL Delft, The Netherlands
Received 2 February 2007; received in revised form 6 April 2007; accepted 13 April 2007
Available online 19 April 2007

Abstract
Ceria (CeO2) and rare-earth modified ceria (CeReOx with Re = La, Pr, Sm, Y) catalysts are prepared by nitrate precursor calcination and are
characterised by BET surface area, XRD, H2-TPR, and Raman spectroscopy. Potential of the catalysts in the soot oxidation is evaluated in TGA
with a feed gas containing O2. Seven hundred degree Celsius calcination leads to a decrease in the surface area of the rare-earth modified CeO2
compared with CeO2. However, an increase in the meso/macro pore volume, an important parameter for the soot oxidation with O2, is observed.
Rare-earth ion doping led to the stabilisation of the CeO2 surface area when calcined at 1000 8C. XRD, H2-TPR, and Raman characterisation show
a solid solution formation in most of the mixed oxide catalysts. Surface segregation of dopant and even separate phases, in CeSmOx and CeYOx
catalysts, are, however, observed. CePrOx and CeLaOx catalysts show superior soot oxidation activity (100% soot oxidation below 550 8C)
compared with CeSmOx, CeYOx, and CeO2. The improved soot oxidation activity of rare-earth doped CeO2 catalysts with O2 can be correlated
with the increased meso/micro pore volume and stabilisation of external surface area. The segregation of the phases and the enrichment of the
catalyst surface with unreducible dopant decrease the intrinsic soot oxidation activity of the potential CeO2 catalytic sites. Doping CeO2 with a
reducible ion such as Pr4+/3+ shows an increase in the soot oxidation. However, the ease of catalyst reduction and the bulk oxygen-storage capacity
is not a critical parameter in the determination of the soot oxidation activity. During the soot oxidation with O2, the function of the catalyst is to
increase the ‘active oxygen’ transfer to the soot surface, but it does not change the rate-determining step, as evident from the unchanged apparent
activation energy (around 150 kJ mol1), for the catalysed and un-catalysed soot oxidation. Spill over of oxygen on the soot surface and its
subsequent adsorption at the active carbon sites is an important intermediate step in the soot oxidation mechanism.
# 2007 Elsevier B.V. All rights reserved.

Keywords: Diesel soot; Oxidation; Rare earth oxide; Mixed oxides

1. Introduction engine sector will lead a larger amount of soot particulate

The emission of the various pollutants from the exhaust
gases of the different energy sources leads to a serious
atmospheric pollution and climate change [1,2]. Increasingly
stringent environmental legislations, on the reduction of NOx
and soot from the various sources of exhaust gases require the
development of a new devices/technology that can abate these
pollutants. Compared with the exhaust gases from the gasoline-
powered engines diesel engine exhaust gases are relatively
clean with respect to the unburnt hydrocarbons and carbon
monoxide. The large volume of the diesel engine market and its
ever increasing demand of the diesel engines in the heavy-duty
* Corresponding author. Tel.: +31 152 78 13 91; fax: +31 152 78 50 06.
E-mail address: m.makkee@tudelft.nl (M. Makkee).
0926-3373/$ – see front matter # 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcatb.2007.04.010
matter emission and these emissions are of the concern and,
therefore, particulate traps or after treatment devices are
necessary [3–5]. The direct oxidation of these soot particulates
needs high temperatures (around 600 8C), and when the engine
is fitted with non-catalytic traps frequent high temperature
regenerations are required. These regenerations are inefficient
and inconvenient [3–5]. The high temperature regeneration of
the un-catalysed filter generally carried out by injecting a diesel
fuel into the exhaust stream. This strategy leads to an
uncontrolled exotherm, which can destroy a filter. A catalysed
soot oxidation is the way to decrease the soot burn-off
temperature in a controlled fashion, and, thereby, increasing the
overall fuel efficiency of the diesel engine.
One of the major problems in decreasing the catalysed soot
oxidation temperature is the contact between the catalyst
surface and the soot particles [6,7]. To overcome the above
190 K. Krishna et al. / Applied Catalysis B: Environmental 75 (2007) 189–200
major problem catalysts, that work based on different
principles, have been developed [3–8]. The use of the fuel-
borne catalysts incorporates the catalyst within the soot particle
and increases the number of contact points and, therefore,
decreases the soot oxidation temperature significantly, from
600 to 400 8C. Similarly, the molten salt catalyst can wet the
soot surface efficiently, increase the number of contact points
between the catalyst and the soot, and, as a result, decrease the
soot oxidation temperature with O2. NO2 is a more powerful
oxidant than O2, and catalysts that convert NO to NO2,
efficiently decrease the soot oxidation temperature. Commer-
cial technologies based on the fuel-borne catalysts or NO2 as an
oxidant are presently used in the automobile engines.
CeO2 is one of the extensively used catalytic components in
many of the above described after-treatment technologies due
to its high activity in the redox reactions. CeO2 is used in a well-
known three-way catalyst for CO, HC, and NOx abatement, as a
fuel-borne catalyst, and in the catalysed soot filters in
elimination of the soot particulates [5,9]. The fuel-borne ceria
catalyst leads to the formation of the CeO2 nano-particles
trapped within the soot particle. The redox properties of CeO2
will generate the active species responsibly for the oxidation
uniformly throughout the soot precursor and lead to an efficient
oxidation. Thereby, a decrease of the combustion temperature
will results in an easier and controllable filter regeneration.
CeO2 is also a major component in many catalysed soot filters
or pre-filters. However, not many fundamental studies are
reported in the open literature using CeO2 or CeO2 containing
materials in the soot oxidation [10].
CeO2 alone as a catalyst or as a support in passive
regeneration of particulate filter is probably of little interest
because of its low textural stability for high-temperature
reactions, usually encountered in the exhaust gases. When
exposed to a high temperature the surface area of CeO2
decreases drastically and at the same time it could lose its redox
properties and oxygen storage ability. Modification of CeO2
with various ions is known to improve the stability towards the
sintering and also the oxidation activity of the resulting
catalysts. This modification was attributed to the changes in the
redox properties and the creation of the oxygen vacancies; both
improve the oxygen exchange with the catalyst and the oxygen
storage capacity [9]. The modification by doping with a
transition metal and rare earth oxide will stabilise the surface
area and will improve the redox/oxygen storage properties of
Table 1
Materials and catalysts
Precursors used Decomposition temperature (8C)a
Ce(NO3)36H2O 230
Ce(NO3)36H2O + La(NO3)36H2O >375
Ce(NO3)36H2O + Pr(NO3)36H2O >375
Ce(NO3)36H2O + Sm(NO3)36H2O
Ce(NO3)36H2O + Y(NO3)36H2O >375
Ce(NO3)36H2O
Ce(NO3)36H2O + La(NO3)36H2O
Ce(NO3)36H2O + Pr(NO3)36H2O
a
Decomposition measured in TGA in He.
CeO2. However, many of such studies were mainly focussed in
three-way catalysis. In only a few instances it is shown that the
dopants in fact could decrease the activity of particular
reactions [11]. Recently, it is shown that the La3+ modified
CeO2 catalysts can improve the soot oxidation activity through
the participation of the lattice oxygen [12,13]. Though the
surface and bulk redox properties of CeO2 are responsible for
the soot oxidation, they are not thoroughly exploited and
conclusively related to the soot oxidation reaction. If the
properties of the reducible oxide CeO2, and the active sites
responsibly for the soot oxidation, are well understood then the
tuning of such materials by the modifications to make them a
more active catalyst might be possible and, thereby, expand the
window of operation in the periodical soot abatement
technology. In the same way, it has been reported that
Ce0.5Zr0.5O2 presents better catalytic activity for soot oxidation
than pure CeO2 [14]. On the contrary, cerium–zirconium
solid solution doped with rare earth (La, Pr, Sm, Tb) have been
studied for soot oxidation by O2, concluding that the
samples containing La, Pr, Sm or Tb do not achieve lower
temperature of combustion than the parent cerium–zirconium
solid solution [15].
The present study aims at the determination of the properties
of the CeO2 catalyst responsibly for the soot oxidation with O2.
In this study CeO2 is modified by its doping with the different
rare-earth ions (La3+, Pr4+/3+, Sm3+, and Y3+) and its effects on
the soot oxidations are studied with O2 and correlated with the
surface and the structural-properties of the catalysts. Further-
more, the tentative mechanistic steps in the catalysed soot
oxidation are proposed.
2. Experimental
2.1. Catalyst preparation
The catalysts are prepared from of the Ce3+, La3+, Pr3+,
Sm3+, and Y3+ nitrate precursors (Aldrich, 99.9%) (Table 1).
The required amount of the nitrate precursor of either single or
mixed components (5 g) is dissolved in H2O (3.3 g). This
aqueous solution is dried at 100 8C for 5 h, followed by a
heating to either 700 or 1000 8C in static air, where it is
maintained for 2 h. g-Al2O3, calcined at 1000 8C for 5 h is used
as an inert material for analysing the un-catalysed soot
oxidation activity. The catalyst is denoted as CeReOx_y, where
Calcination (8C/2 h in air) Catalyst
1000 CeO2_1000
1000 CeLaOx_1000
1000 CePrOx_1000
1000 CeSmOx_1000
1000 CeYOx_1000
700 CeO2_700
700 CeLaOx_700
700 CePrOx_700
 K. Krishna et al. / Applied Catalysis B: Environmental 75 (2007) 189–200 191

‘ReOx’ denotes as the 10 wt% doped rare-earth oxide and ‘y’ is the TGA experiments the set-up is optimised to eliminate the
the calcination temperature used to prepare the catalysts. possible heat and mass transfer effects [17].

2.2. Characterisation 3. Results

The BET surface areas and the N2 adsorption isotherms of
catalysts are measured by the N2 physisorption at 196 8C in
an automatic volumetric system (Autosorb-6, Quantachrome).
The X-ray diffractogrammes are recorded in a Philips X-ray
diffractometer, PW 1840, with Ni-filtered Cu Ka radiation
(l = 0.15418 nm). Data is collected between 2u = 6–708 with a
step size of 0.028 (total collection time 4 h). Crystallite size is
calculated using the Scherrer equation:
Kl
D¼
b cos u
3.1. Characterisation
3.1.1. Catalyst preparation
Table 1 shows the catalyst precursors used and their
decomposition temperature. Seven hundred or 1000 8C
calcinations are applied to make sure all the catalyst precursors
decompose to the oxides. The decomposition temperature
values of the respective precursors suggest that CeO2 is formed
first and on top of the CeO2 the dopant nitrates decompose to
form oxides and, subsequently, the high temperature treatment
(1)
leads to the solid solutions.

where l is the X-ray wavelength, K the particle shape factor, 3.1.2. N2 adsorption
and the b is the full width at half maximum of peak at an angle The BET surface areas of the various materials prepared are
u. given in Table 2. The surface area of CeO2 decreased from 57 to
The temperature programmed reduction by H2 (H2-TPR) is 2 m2 g1, when calcined at 700 and 1000 8C, respectively. The
carried out in a tubular quartz reactor (5 mm internal diameter), rare-earth ions selected in this study are known to form a solid
coupled to a TCD detector for analysing the H2 consumption. solution, and, as a consequence also stabilise the CeO2
100 mg of the sample, diluted with SiC packed between two crystallites towards sintering resulting in a less severe surface
quartz wool plugs, is reduced from room temperature to area decline [18]. However, the surface areas decreased for the
1000 8C at a heating rate of 10 8C/min, in 30 ml/min flow of CeReOx_700 catalysts compared with CeO2_700. The BJH
7.7 vol.% H2 in Ar. CuO is used as a reference material to analysis of the N2 adsorption isotherms (Fig. 1) of the
quantify the H2 consumption. CeO2_700 catalyst shows a narrow pore size distribution
The Raman spectra of fresh catalysts are recorded in a around 10 nm. Doping with rare-earth ions led to a bimodal
Renishaw Raman Imaging Microscope, system 2000 with a pore size distribution. The pore size also shifted to from 10 to
20 mW Ar laser (514 nm). The Raman spectrometer is around 15 nm (meso-pore range) and above 50 nm (meso/
calibrated using a silicon wafer. macro pore range). The effect of the rare-earth ion dopant on
the increase of surface area is only evident for the catalysts
2.3. Soot oxidation in TGA prepared by a 1000 8C calcination. For the CeReOx_1000
catalysts the surface area is mainly arising from the meso–
The thermo gravimetric analysis (TGA) of the soot macro pore range and considered as the external surface area.
oxidation of the various samples with O2 is carried out in a
Mettler Toledo, TGA/SDTA851e instrument. The catalytic 3.1.3. XRD
activity is measured in a loose (mixed with a spatula) and a tight Fig. 2 shows the powder X-ray diffractogrammes of the
contact (grounded in a mortar) mode with catalyst–soot selected catalysts prepared by a 1000 8C calcination.
mixtures in 4:1 (w/w) ratio, respectively [6,12]. Printex-U CeO2_1000 shows the very sharps peaks characteristic for
from Degussa S.A., is used as a model soot whose crystalline cubic phase. Doping with the different rare-earth cat
characterisation is reported elsewhere [6]. Weighed amount
of the sample (having less than 1 mg of soot) is placed in a TGA Table 2
crucible and heated with the different heating rates in 100 ml/ BET surface area of different materials and aT10 and T50 temperature
min air or He to 800 8C.
Catalyst BET surface Temperature (8C)
The apparent activation energy of the soot oxidation is area (m2 g1)
determined by the Ozawa [16] method using the following T10 T50
expression: Al2O3_1000 – 539 618
CeO2_1000 2 499 576
d logðFÞ Ea CeLaOx_1000 16 414 459
¼ 0:4567 (2) CePrOx_1000 11 425 484
dð1=T x Þ R CeSmOx_1000 5 450 520
CeYOx_1000 2 500 572
where F is the heating rate used, Tx the temperature corre- CeO2_700 57 388 450
sponding to x% soot conversion, Ea is the apparent activation CeLaOx_700 48 388 425
energy in kJ mol1. Ea can be estimated from the slope of the CePROx_700 44 378 422
a
least squares straight line fit of log(F) versus 1/Tx plot. Prior to Temperature at 10 and 50% soot conversion, experimental conditions in Fig. 8.
192 K. Krishna et al. / Applied Catalysis B: Environmental 75 (2007) 189–200

for the catalysts prepared by a 700 8C calcination; however, the

Fig. 1. BJH analysis of N2 adsorption isotherms over selected catalysts as
indicated.
Fig. 2. XRD of selected catalysts (higher angles is not shown).
ions decreased the intensity of the peaks and shifted them to the
different 2u values. Only a single phase corresponding to the
cubic fluorite type structure is observed in all CeReOx_1000
catalysts. In the case of the La3+ and Sm3+ modified catalysts,
the 2u shifted to the lower values, suggesting the incorporation
of the relatively larger ions into the CeO2 cubic structure
forming solid solution [18]. The similar observations are made
intensity of CeO2 and the decrease in intensity on the rare-earth
ion doping are much smaller compared with the CeO2_1000
and CeReOx_1000 catalysts.
In Table 3 the various parameters calculated from the XRD
are given. The lattice constant of CeO2_1000 is estimated as
0.5406 nm, which increased to 0.5441 nm for the La3+ doping.
This is consistent with the incorporation of the larger size La3+
ions (0.116 nm) that will expand the CeO2 lattice. Comparison
of the CeLaOx lattice constant with the materials of a similar
composition, reported in the literature (which follows Vegard’s
law [19]), suggests that most of the La3+ introduced is in fact
forming a solid solution. Praseodymium ions form a solid
solution very efficiently with CeO2, however, only a small
change in the peak positions are observed in XRD (not shown).
The lattice parameter of CePrOx_1000 (0.5402) was smaller
than that expected for the incorporation of Pr3+ and also smaller
than that observed over the CeO2_1000 (0.5406 nm). Praseo-
dymium doping leads to the solid solution containing both the
Pr4+ and Pr3+ ions into the resulting material. If extensive
amount of Pr3+ (radii 0.113 nm) is incorporated in the lattice of
CeO2 it should lead to a significant increase in the lattice
constant. However, the small decrease observed in the lattice
constant is consistent with the incorporation of the majority of
the praseodymium as Pr4+, having a slightly smaller ionic
radius (0.096 nm) compared with Ce4+ (0.097 nm). The second
possibility which cannot be ruled out is the dispersion of PrOx
on the surface of CeO2 with out forming solid solution. A small
increase in the lattice constant of CeO2 is observed on the Sm3+
doping, though one expected it to be somewhat larger if all the
Sm3+ (with ionic radii of 0.106 nm) is incorporated into the
CeO2 lattice. Therefore, the possible presence of SmOx as a
separate phase cannot be ruled out though it is not observed
from the XRD. A similar comparison of the CeYOx_1000 with
the CeO2_1000 lattice constants also indicates the existence of
the separate phases in the former catalyst.
The crystallite size calculated from the Scherrer equation
and the surface area calculated from the crystallite size
assuming spherical shape are reported in Table 3. CeO2_1000
and CeYOx_1000 have crystallites larger than 140 nm. The size
decrease to around 50 nm for the different rare earth cat-ion
doping, which is also in agreement with the trend of the

Table 3
Various parameters calculated from XRD
Catalyst Ionic sizea (nm) Phases a0 (nm) CS (nm) Surface areab
(m2 g1)
Re4+ Re3+
CeO2_1000 0.097 0.114 Cubic 0.5406 >140 6
CeLaOx_1000 – 0.116 Cubic 0.5441 44 19
CePrOx_1000 0.096 0.113 Cubic 0.5402 79 11
CeSmOx_1000 – 0.106 Cubic 0.5413 89 9
CeYOx_1000 – 0.102 Cubic 0.5404 >175 5
CeO2_700 – Cubic – 12 70
CeLaOx_700 – Cubic 0.5440 12 70
CePrOx_700 – Cubic 0.5400 14 60
a
Not calculated from XRD, a0 = lattice constant, CS = crystallite size.
b
From XRD crystallite size, assuming spherical shape.
 K. Krishna et al. / Applied Catalysis B: Environmental 75 (2007) 189–200
measured BET surface area (Table 2). The extent of the
incorporation of the different ions into a solid solution seems to
influence the crystallite size; for example, more La3+ is
incorporated into the lattice of CeO2 (from lattice parameter
estimation). This will result in a larger stabilisation of the CeO2
particles and, thereby, maintaining a small crystallite size.
Similarly CePrOx_1000 and CeSmOx_1000 resulted in
relatively smaller crystallites of around 70 nm. The surface
area calculated from the crystallite size measured by the XRD
assuming the spherical particles is also in a fair agreement with
the measured BET surface area (Tables 2 and 3) over the
different catalysts indicating most of it is arising from the
external surface area. For the catalysts prepared by the 700 8C
calcination, the crystallite size remained around 12 nm, also in
a fair agreement with the BET surface area of these catalysts.
3.1.4. H2-TPR
Fig. 3 shows the H2-TPR profiles of the selected catalysts
used in this study. CeO2_1000 shows a prominent H2-
consumption peak centred on 870 8C corresponding to the
bulk reduction. In the CeReOx_1000 catalysts, the H2 reduction
peak broadened and shifted towards low temperatures. The shift
in the H2-TPR curves towards a lower temperature is due to the
increased number of the surface/subsurface reducible Ce4+ ions
leading to an increased H2 consumption [20]. The TPR profile
of CePrOx_1000 shows a large H2 consumption peak around
490 8C along with the broad peak below 700 8C, corresponding
to the surface Ce4+ reduction followed by the bulk CeO2
reduction. The H2-consumption peak around 490 8C is due to
the reduction of Pr4+ to Pr3+, confirming the existence of the
large amounts of Pr4+ in the CePrOx_1000 solid solution. This
is in an agreement with the XRD. The PrOx containing catalysts
can be easily reduced compared with CeO2 as shown before
[21]. However, from the H2-TPR, it is not obvious whether the
Pr4+ is present on the surface or in the bulk of the CePrOx_1000
crystallites. Even in an oxidising atmosphere, exposure to
1000 8C, leads to the reduction of the majority of Pr4+ to Pr3+.
The presence of a large number of Pr3+ in the lattice is shown to
exert a lattice strain in the solid solution and, thereby, pushing
Fig. 3. H2-TPR of selected catalysts.
193
the Pr3+ onto the surface [22]. Therefore, it is fair to assume that
the Pr4+ ions are near to the surface in the CePrOx_1000
crystallites.
3.1.5. Raman spectroscopy
The Raman spectra of the different catalysts are shown in
Fig. 4. CeO2_1000 shows a Raman shift at 464 cm1,
corresponding to the oxygen breathing frequency around the
Ce4+ ions [18,23]. The low intensity band observed around
570 cm1 is correlated to the oxygen vacancies and around
1200 cm1 to the surface O22 ions. When CeO2 is doped with
ReOx, the decrease in the intensity and the small negative shifts
in the Raman bands are observed. It is proposed that with a
decreasing particle size the Raman peak energy decreases and
the line width increases due to the increased lattice constant
[18,23]. The influence of the particle size on the broadening of
the Raman bands is not expected to be significant in this study
as the crystallite size of the catalysts studied are quite large. The
broadening of the peaks can be, therefore, entirely related to the
oxygen vacancies and dilution of the CeO2 lattice. For the
coloured CePrOx_1000 catalyst the optical absorption is largely
responsible for a strong decrease in the intensity of Raman
bands. CePrOx_1000 showed relatively a large number of
oxygen vacancies compared to the other catalysts (Fig. 4, inset).
The XRD and H2-TPR results did not suggest the formation of
such a large number of vacancies due to the Pr3+ in the
CePrOx_1000 solid solution (lattice constant should increase
and less H2 consumption around 470 8C). The following
explanation can be given for the relatively large number of the
observed vacancies. The Pr4+/3+ forms a good solid solution
with the CeO2. There could also be the segregation of the Pr4+
ions onto the surface (as suggested by the H2-TPR). Pr4+ loses
the oxygen easily on heating to around 300 8C and the Ar-laser
power used in the Raman spectrometer could cause such a local
heating of the sample, creating the observed vacancies. As the
Raman technique is a surface characterisation technique, the
large number of the oxygen vacancies in the cubic structure of
the CePrOx_1000 catalyst shows the surface segregation of the
Pr3+/Pr4+ ions. The inset in Fig. 4 shows the Raman shift over
Fig. 4. Raman spectra of catalysts as indicated.
194 K. Krishna et al. / Applied Catalysis B: Environmental 75 (2007) 189–200
the different catalysts above 1000 cm1. The bands observed
for CeO2_1000 and CeLaOx_1000 around 1200 cm1 can be
assigned to the O22 groups on the lattice. CeSmOx_1000 and
CeYOx_1000 show, however, a number of bands around these
frequency regions. Though we did not attempt to characterise
the nature of the species responsible for these bands,
considering the XRD analysis we can fairly correlate them
to the separate phases of SmOx and YOx.
3.2. Activity tests
3.2.1. Soot oxidation with O2
Fig. 6 shows the soot conversion during the oxidation in O2
for the catalyst–soot mixtures in the loose contact mode. The
TGA data is normalised by removing the weight loss (below
300 8C) due to H2O and the adsorbed hydrocarbons/very
reactive surface oxygen groups of the soot. The onset
temperature T10, the temperature at 10% soot conversion, is
above 530 8C for all the catalysts studied. Using Al2O3, the soot
oxidation (referred to as the un-catalysed soot oxidation) is
complete between 500 and 670 8C, with T50, the temperature at
50% soot conversion, at 605 8C. The presence of the ceria-
based catalysts decreased T50, by about 5–15 8C. However, the
onset temperature of the soot oxidation is not influenced. The
influence of the catalysts in the loose contact mode is very
limited, in agreement with the previous studies [3,6,7,12].
Among the materials examined in the loose contact mode,
CePrOx_1000 is found to have a better activity. The soot
oxidation activity followed the trend CePrOx_1000 >
CeLaOx_1000  CeO2_1000 > Al2O3_1000.
Fig. 5 shows the soot oxidation activity on single component
oxides in the tight contact with the soot. Except Al2O3_1000 all
the other oxides showed decrease in the soot oxidation
temperature. The soot oxidation temperature followed the trend
of SmOx_1000  LaOx_100 > PrOx_1000  CeO2_1000 
YOx_1000 > Al2O3_1000. Despite the absence of redox
sites, Sm3+, La3+ and Y3+ oxides show a decrease in the soot
oxidation temperature. These oxides are shown to have a
Fig. 5. Soot oxidation with O2 over selected catalysts as indicated. Reaction
conditions: TGA, catalyst:soot—4:1 (w/w), contact = loose, heating rate =
10 8C/min, air flow = 100 ml/min.
Fig. 6. Soot oxidation with O2 over single component oxide catalysts. Reaction
conditions: TGA, catalyst:soot—4:1 (w/w), contact = tight, heating rate =
15 8C/min, air flow = 100 ml/min.
surface peroxide like species that could participate in the soot
oxidation [24]. In the present investigation, the surface area of
these materials is not measured and the extent of their
participation compared with the lattice oxygen of CeO2_1000 is
not evident.
Using O2 as the oxidising agent generally it is difficult to
rank the catalysts under the loose contact conditions, as all of
the catalysts show a similar temperature region for the soot
oxidation. Therefore, the influence of the rare-earth ion doping
on the intrinsic activity of CeO2 is studied in the tight contact
mode [25,26]. Fig. 7 shows the TGA results of the catalysts,
prepared by 1000 8C calcination in the tight contact with the
soot. With the Al2O3_1000, no appreciable change of an onset
temperature T10 (451 8C) or T50 (618 8C) is observed in the soot
oxidation compared with the loose contact mode obviously
because Al2O3 does not have the redox or oxygen storage
catalytic sites. In the tight contact mode the CeO2_1000 and
CeReOx_1000 catalysts showed a considerable decrease in the
soot oxidation temperatures at all conversion levels (T10 and T50
Fig. 7. Soot oxidation with O2. Reaction conditions: TGA, catalyst:soot—4:1
(w/w), contact = tight, heating rate = 15 8C/min, air flow = 100 ml/min.
 K. Krishna et al. / Applied Catalysis B: Environmental 75 (2007) 189–200
shifted to lower temperatures by about 40–150 8C). Among all
the catalysts CeLaOx_1000 showed a superior performance
followed by CePrOx_1000, which decreased the T10 and T50
temperatures by more than 100 8C. The soot conversion with
a catalyst in the tight contact mode followed a trend of
CeLaOx_1000 > CePrOx_1000  CeSmOx_1000 > CeYO-
x_1000 > CeO2_1000 > Al2O3_1000. The catalysts prepared by
the 700 8C calcination (Fig. 8) are more active than the catalysts
prepared by the 1000 8C calcination (Fig. 7). The soot conversion
followed the order CePrOx_700  CeLaOx_700 > CeO2_700.
The BET surface areas of the various catalysts studied and
T10 and T50 temperatures of soot oxidation are given in Table 2.
No correlation is observed between the BET surface area and
the soot oxidation in the loose contact mode or in the tight
contact mode for the catalysts prepared by the 700 8C
calcination. In contrast for the catalysts prepared by the
1000 8C calcination in the tight contact with the soot, a clear
correlation of the decrease in the soot oxidation temperature
with the increase in the surface area is seen. It is not possible to
correlate the total surface area with the soot conversion in the
tight contact mode if the catalyst has the micro-pores, as the
soot particles do not fit in the micro-pores, even under the tight
contact conditions [25]. The N2 adsorption isotherm suggests
that the surface area of all the catalysts, prepared by the 1000 8C
calcination, is arising from the external surface area, which is
accessible for the spread of the soot, especially in the tight
contact mode. The CeReOx_700 catalysts have the higher
external surface area compared with the CeO2_700 and all the
other catalysts. The higher external surface area explains the
superior performance of the CeReOx_700 in spite of having the
relatively lower BET surface area compared to the CeO2_700.
To evaluate the effect of the rare-earth doped CeO2 and the
resulting changes in the redox properties, evident from the
characterisation of the catalysts on the soot oxidation activity, it
is necessary to compare the catalysts having a similar external
surface area under the tight contact conditions. Fig. 9 shows the
soot conversion over the surface area normalised CeO2_1000
and CePrOx_1000. The soot conversion over CeLaOx_1000 is
Fig. 8. Soot oxidation with O2 over catalysts as indicated. Reaction conditions:
TGA, catalyst:soot—4:1 (w/w), contact = tight, heating rate = 15 8C/min, air
flow = 100 ml/min.
195
Fig. 9. Comparison of soot oxidation over surface area normalised CeO2 with
other catalysts. CeO2* represents surface area of CeO2 normalised to that of
CePrOx. Reaction conditions: TGA, catalyst:soot—4:1 (w/w), contact = tight,
heating rate = 15 8C/min, air flow = 100 ml/min.
also shown for a comparison. The soot conversion temperature
decreased over the surface area of the normalised CeO2_1000
catalyst compared with that of the CePrOx_1000 or even the
CeLaOx_1000, having a higher surface area. This observation
shows that the intrinsic redox behaviour of the unmodified
CeO2 is more efficient in the oxidation of the soot. Also it can
be expected that over the surface area normalised CeO2 more
number of the redox sites exist compared with the
CeLaOx_1000 catalyst, where the surface segregation of
La3+ is expected, which is not reducible and may not directly
participate in the soot oxidation. However, La3+ could improve
the rate of the oxidation due to the creation of the oxygen
vacancies, which were not observed in the present study. In the
catalysts studied the modification of the surface redox
properties (shown by the TPR) may not be as important as
that of the external surface area.
In Fig. 10, the ratio of the external surface area of the catalyst
to that of the initial soot surface area is plotted against the
temperature (Tx) at the different conversion levels (x). The soot
Fig. 10. Effect of surface area ratio of catalyst to initial soot, on Tx, temperature
at ‘x%’ soot conversion. Reaction conditions: TGA, catalyst:soot—4:1 (w/w),
contact = tight, heating rate = 15 8C/min, air flow = 100 ml/min.
196 K. Krishna et al. / Applied Catalysis B: Environmental 75 (2007) 189–200

Fig. 11. Ozawa plots over CeO2 at different soot conversion levels. F-Different
heating rates applied during soot oxidation, and Tx (temperature at ‘x%’
conversion). Reaction conditions: TGA, catalyst:soot—4:1 (w/w), contact =
tight, F = 15, 20, 25 and 30 8C/min, air flow = 100 ml/min.
conversion below 475 8C can be considered entirely arising
from the catalysed oxidation (as the onset of un-catalysed soot
oxidation is above 530 8C). Above 475 8C, the resulting soot
conversion is due to the contribution from both the catalysed
and un-catalysed soot oxidation. From the figure it can be
derived that, if the majority of the soot has to be converted by
the catalysed oxidation, the catalyst external surface area to the
initial soot surface area ratio should be above 0.4 (La3+ and
Pr3+/4+ modified CeO2). From Fig. 10 we can also conclude that
the La3+ and Pr3+/4+ modified CeO2 catalysts are more efficient
in catalysing the soot oxidation (around 75% catalysed soot
conversion at 475 8C).
3.2.2. Activation energy
By applying the different heating rates during the soot
oxidation in TGA, the Ozawa plots at the various soot
conversion levels are constructed (an example for CeO2 is
shown in Fig. 11). The plots of the logarithmic heating rates
used versus the inverse temperatures at the various soot
conversion levels showed a good linear fit. Such experiments
are carried out for all the catalysts. In Table 4 the soot oxidation
activation energies (Ea) estimated from the slopes of the linear
fits are given. The activation energy for the un-catalysed
reaction is around 150 kJ mol1. The presence of a catalyst,
Fig. 12. Soot oxidation over the catalysts as indicated in He. Reaction con-
ditions: TGA, catalyst:soot–4:1 (w/w), contact = tight, heating rate = 10 8C/
min, He flow = 100 ml/min.
even under the tight contact mode, did not influence this
activation energy significantly. The small changes observed in
the activation energy are due to the nature of the TGA technique
used (oxygen diffusion and heat transfer influence). The
unchanged activation energy indicates that only the rate of the
reaction is increased in the presence of the catalyst, but the soot
oxidation reaction mechanism does not change.
3.2.3. Bulk oxygen for soot oxidation
Fig. 12 shows the weight loss due to the soot oxidation in He
with the catalyst–soot mixtures in the tight contact mode. The
weight loss is arising from the soot oxidation due to the
participation of the surface/bulk lattice oxygen. The soot
oxidation activity of the PrOx_1000 catalyst is included in
Fig. 12, as it can contribute a significant amount of the bulk
oxygen for the soot oxidation compared with the other
catalysts. If all the reducible bulk oxygen, as observed from the
H2-TPR, is involved in the soot oxidation it should lead to
around 6% weight loss. However, no appreciable weight loss is
observed over the CeO2_1000 or CeLaOx_1000 catalysts,
which indicates that the bulk reduction properties of the
catalysts, which only show above the 700 8C H2-TPR (Fig. 3),
is not important in the determination of the soot oxidation
properties. The PrOx_1000 catalyst shows more than two orders
of magnitude soot oxidation in He due to the participation of the
bulk O2. PrOx_1000 was, however, found to be a poor soot

Table 4
Apparent activation energy (Ea) calculated by Ozawa plot at different soot conversions
Catalyst Apparent activation energy [kJ mol1] at ‘x%’ conversion
5 10 15 20 30 50 75
Al2O3 – 159  5 153  1 183  35 177  34 143  2 142  3
CeO2 168  12 156  10 146  4 139  5 135  7 129  1 133  2
CeLaOx 132  10 139  4 133  4 135  3 133  3 123  2 127  4
CePrOx 155  7 144  7 145  5 140  8 138  10 130  9 133  10
CeSmOx 148  12 153  9 144  11 146  10 145  6 153  12 148  43
CeYOx 134  28 143  20 141  19 139  17 139  16 143  10 140  6
For reaction conditions see Fig. 10.
 K. Krishna et al. / Applied Catalysis B: Environmental 75 (2007) 189–200
Table 5
Apparent activation energy of soot oxidation over PrOx with oxygen and in He
Catalyst Apparent activation energy
[kJ mol1] at T5%
PrOx–O2 145  38
PrOx–He 234  9
For reaction conditions see Fig. 9 for soot oxidation with O2, and Fig. 12 for soot
oxidation in He.
oxidation catalyst (not shown) compared with the rare-earth ion
doped catalysts. The above results rule out the involvement of
the catalyst bulk reduction properties in the soot oxidation.
Table 5 gives the activation energies of the soot oxidation in
O2 and in He over the PrOx_1000 in the tight contact with the
soot. The activation energy of around 150 kJ mol1 (Table 4)
during the oxygen assisted soot oxidation, which is in
agreement with the generally reported values in literature
[3]. When the soot oxidation is carried out in He, where one can
assume the lattice oxygen, mainly bulk oxygen for PrOx_1000,
participates in oxidation, the apparent activation energy is
found to be around 240 kJ mol1. This high activation energy
indicates that the participation of the bulk oxygen in the soot
conversion is rather limited.
4. Discussion
The effect of the rare-earth ion doping on CeO2, which is
thought to influence the catalytic properties, in many instances
in a positive way, is investigated here in more detail. A very
simple catalyst preparation method was selected with emphasis
on the high calcination temperature, though it may not yield to
the high surface area materials. The surface areas obtained in
this paper are more realistic (Table 2) considering the fact that
the high temperatures usually encountered during the soot
oxidation in the exhaust gases, where there is the frequently
uncontrolled burning of the soot, leading to the low surface area
materials. The soot oxidation over the various catalysts has
been studied using O2 (Figs. 5–10) in the loose and tight contact
conditions. The loose contact did not show a significant
decrease in the soot oxidation temperature irrespective of the
nature or the surface area of the catalyst, consistent with
reported literature [6,7,11,14]. When the soot particle is not in
close contact with the catalyst surface, the catalyst lattice
oxygen cannot be transferred to the soot surface [12].
Therefore, it is adopted to test the catalysts using the tight
contact condition, which is often used in literature to prioritise
the catalysts especially using the O2 as an oxidant [12,25,26].
The soot conversion temperature strongly decreased over the
rare-earth modified CeO2 catalysts (Figs. 7 and 10). Rare-earth
ion modification leads to a number of changes in the catalyst
surface properties, which can explain the observed improve-
ment in activity. The possible reasons for the increase in activity
of the rare-earth doped CeO2 catalysts in the soot oxidation
with O2, is thoroughly probed by the different catalyst
characterisation techniques.
One of the parameters that can explain the differences in the
catalysed soot oxidation temperature with O2 is the catalyst
197
surface area. The comparison of the BET surface area of all the
catalysts to that calculated from the crystallite size measured
from the XRD (Tables 2 and 3), in the respective catalysts,
suggest that most of it is arising from the external surface of
crystallites. The CeO2_700 had higher BET surface area
(58 m2 g1) compared with the CeReOx_700 catalysts
(<48 m2 g1). There is no significant change in the XRD
crystallite size of the different catalysts prepared by the 700 8C
calcination (around 12 nm). The BET and XRD results indicate
that the sintering of CeO2 crystallites is not significant during
catalyst preparation by heating to 700 8C. A sharp pore size
distribution observed around 10 nm in CeO2_700 could be due
to the intra particle voids. Though the pore volume is relatively
more, it is not accessible for the spread of the soot (around
30 nm particles) even in the tight contact mode. Doping with
La3+ or Pr3+/4+ led to an increase in this intra particle void
volume (to around 15 nm) and especially led to an important
increase in the macro-pore volume (>50 nm, Fig. 1), indicating
the more open structure of CeReOx_700 particles. This open
structure leads to an increased accessibility of the CeReOx
redox sites to the soot and, therefore, increases the soot
oxidation rate (Fig. 7).
The sintering CeO2 (CS > 140 nm) seems to be more
important at high catalyst preparation temperature (1000 8C).
XRD analysis (Fig. 2) shows that the dopants in general
decrease the crystallite size, which can be correlated to a
decreased sintering of CeReOx_1000. This is also reflected in
the increased BET surface area of these catalysts (Table 2).
Most of the surface area in the catalysts, prepared by the
1000 8C calcination, is arising from the macro-pores or external
surface area, which is accessible for the spreading of the soot.
Under the catalyst preparation and the soot oxidation conditions
used a decrease in the crystallite size not only increases the
external surface area but also leads to a good mixing of the
catalyst and soot, thereby, decreasing the oxidation tempera-
ture. However, these experiments do not reveal the intrinsic
activity of modified catalytic sites.
Fig. 13 depicts how the soot particle could be present under
the tight contact conditions over the catalysts having the
different external surface areas. Simple calculations reveal that
a monolayer of soot particles, under the experimental
conditions used, can be easily spread, over the catalyst having
8 m2 g1 surface area. Because of the high surface area of
CeLaOx_1000 catalyst it can be assumed that a monolayer of
primary soot particles can be spread over the catalyst, though
this may not be easy even under the tight contact conditions
used. Considering the low surface area of CeO2_1000, soot
Fig. 13. Depiction of primary soot particle in a tight contact with (a) CeO2 and
(b) CeLaOx catalysts under ideal spreading.
198 K. Krishna et al. / Applied Catalysis B: Environmental 75 (2007) 189–200
particles can only be spread in multi layers under the conditions
used for the soot oxidation. Though the surface of CeO2_1000
is intrinsically more active due to the restricted accessibility of
the catalyst active sites to the gas phase oxygen and to the soot
layers lead to the lower soot conversions as observed. The
intrinsic activity of the redox catalysts, therefore, cannot be
obtained by the simple tight contact experiments using the fixed
weight of catalyst for analysing the effect of the rare-earth ion
doping on the soot oxidation, which changes the CeO2 surface/
bulk redox properties. The external catalyst surface area has to
be normalised (Fig. 9) for studying such a soot oxidation. When
the surface area was normalised the intrinsic activity of
CeO2_1000 was found to be higher than the rare-earth ion
doped CeO2 catalysts.
The soot oxidation activity may also depend on the nature
and the extent of the incorporation of the dopant into the CeO2
structure, which influence the surface properties of the catalyst
to the different extents. Among all the catalysts CeLaOx showed
a decreased of around 150 8C compared to the un-catalysed
soot oxidation (Fig. 7). The XRD measurements show that the
La3+ forms a solid solution efficiently compared with all the
other ions used in this study and, therefore, the catalyst
CeLaOx_1000 is also relatively more active. With Pr3+/4+,
though it is not possible to fully ascertain the extent of solid
solution formation with CeO2, looking at the improvement in
the surface area (11 m2 g1), we can assume that it forms a
solid solution efficiently. The CePrOx_1000 activity is less than
that observed for the CeLaOx_1000. However, in spite of its
lower surface area of the CePrOx_700 (44 m2 g1) showed a
slightly better soot conversion compared with CeLaOx_700
(48 m2 g1). The formation of the solid solution does not seem
to be an important criterion for the extent of the soot oxidation.
Most of the praseodymium ions are present in 4+-oxidation
state in the CePrOx catalyst as evident from the XRD, which
leads to a surface enrichment with the oxygen, responsibly for
the soot oxidation. However, the presence of the non-reducible
ions such as La3+ leads to a surface enrichment with the oxygen
vacancies, especially when it is segregated to the surface, which
could decrease the overall oxidation function of the catalyst.
For praseodymium this can lose the oxygen relatively easily
thus could oxidise the soot at a relatively lower temperature.
From the XRD it was also observed, that Sm3+ and Y3+ did not
form efficiently the solid solutions and the presence of separate
phases in the CeSmOx_1000 and CeYOx_1000 catalysts can be
expected. This also explains the lower extent of the surface area
improvement and its relatively lower soot oxidation activity
compared with the CeLaOx_1000 and CePrOx_1000 catalysts.
The improvements in the activity of the various reactions
observed in the rare-earth ion doped CeO2 based catalysts are
mainly correlated to the changes in the catalyst redox
properties. The catalyst surface and bulk properties calculated
based on the XRD, BET surface area, and H2-TPR are given in
Table 6. Except CeYOx_1000, all the catalysts consumed
around 0.25 moles of H2 per mole of CeO2. The soot oxidation
has no correlation with the total H2 consumption. H2-TPR also
reveals that the catalyst bulk reduction takes place above
700 8C and is not involved in the soot oxidation. This is further
Table 6
Parameters calculated based on H2-TPR, XRD and BET surface area over
catalysts as indicated
Catalyst H2/Ce4+ No. of surface No. of reduced
(mole/mole) Re4+ ionsa (g1 cat) layersb
CeO2_1000 0.25 4.35E+18 8
CeLaOx_1000 0.24 3.44E+19 3
CePrOx_1000 0.27 2.40E+19 14
CeSmOx_1000 0.27 1.09E+19 3
CeYOx_1000 0.20 4.36E+18 7
a
Assuming all the surface ions are Re4+.
b
Calculated from H2 consumption below 475 8C.
supported by the soot oxidation carried out in He (Table 5;
Fig. 12). Higher amounts of soot is oxidised over PrOx_1000
under He (Fig. 12), which is due to the easy bulk reduction.
The bulk reduction of CeLaOx_1000 and CeO2_1000 occurs at
a higher temperature and, therefore, resulted in an insignif-
icant soot oxidation under He. However, in the presence of the
oxygen the CePrOx_1000 and CeLaOx_1000 catalysts are
more active compared with the PrOx_1000. This suggests the
involvement of only the surface reducible layers in the soot
oxidation. Further, the high activation energy for the soot
oxidation involving bulk oxygen (around 250 kJ mol1)
suggests that the bulk oxygen diffusion to the surface is
much slower and only the surface lattice oxygen is important
in determining the soot oxidation under the oxidation
conditions.
Though the nature of the reductant in the H2-TPR and the
soot oxidation (carbon) are different, the trends in the catalyst
surface reduction temperature in the H2-TPR and the soot
oxidation onset temperature are very well coincided, which
suggests the involvement of the surface reducible layers in the
soot oxidation. From the H2 consumption below 475 8C (the
temperature of interest for catalysed soot oxidation over these
catalysts), the XRD and BET surface area of the catalysts, it is
estimated that 3–14 surface layers can be reduced in these
catalysts and they are possibly involved in the soot oxidation
(Table 6). The smaller the number reducible layers over
CeLaOx_1000 (3 layers) compared with the CeOx_1000 (8
layers) indicates the possibility of the surface/subsurface
segregation of the non-reducible La3+ in CeLaOx_1000
catalyst. The studies dealing with the migration of the dopant
to the surface are reported in the literature especially when high
temperatures are used in the catalyst preparation [19]. The
segregation of the non-reducible ion to the surface leads to an
oxygen deficient surface. Thus the intrinsic activity per unit
surface area of the catalyst decreases. This indicates that the
rate of the soot oxidation strongly depends on the surface area,
because the CeOx_1000 having a low surface area will lead to a
smaller contribution of the number of the Ce4+ ions though
more surface layers can be reduced. The larger surface area
materials will, however, contribute to an increased number of
the Ce4+ ions at the surface and, thereby, to the soot oxidation
rates. In the case of the CePrOx_1000, 14 layers is probably an
overestimation as it is not possible to distinguish between the
surface and the bulk reduction of Pr4+ ion. However, the results
 K. Krishna et al. / Applied Catalysis B: Environmental 75 (2007) 189–200
suggest that the surface reducible layers are important and
participate in soot oxidation.
The soot takes the lattice oxygen from the surface of CeO2
creating vacancies, which are quickly filled by the gas phase or
subsurface oxygen. It is shown that the rate of the gas phase
oxygen reacting with the vacancies on the catalyst surface is
much faster than the rate of the direct soot oxidation [12]. The
ease of the surface reduction may also influence the soot
oxidation rate, which is not correlated in our previous studies.
The H2-TPR shows that the CePrOx_1000 surface is more
easily reducible due to the surface enrichment with Pr4+ ions.
However, this catalyst is relatively less active compared with
the CeLaOx_1000, which suggests that the ease of the surface
reduction is not that important for soot oxidation with O2. The
presence of the oxygen vacancies is often thought to favour
the oxygen mobility over the catalyst surface and improve the
rates of oxidation reactions. From the Raman spectra the
CePrOx_1000 catalyst (Fig. 4) indicates the possibility of more
number of such species under the soot oxidation conditions.
However, this catalyst was less active than CeLaOx_1000,
which has relatively less number of such vacancies. The
mobility of ‘‘O’’ is also, therefore, not an important parameter
in determining the soot oxidation rate. The Raman spectra
further revealed the presence of the separate phases in
CeSmOx_1000 and CeYOx_1000 catalysts, which is consistent
with the low BET surface area and also the low soot conversions
found over these catalysts.
From the surface area of the printex-U soot and assuming the
graphite density, the primary soot particle size to be around
30 nm is estimated, which is in a good agreement with that
observed from the TEM [27]. Thus the majority of the soot
surface area is arising from the external surface. From the
surface areas (Table 2) and the soot conversion under the tight
contact conditions (Fig. 10), it is estimated that roughly
1900 nm2 catalyst surface area is needed for the catalysed
oxidation of 30 nm soot particle. If the catalyst is assumed to be
flat surface, 1900 nm2 is a few orders of magnitude higher than
the real contact area between the soot and the catalyst. The
following proposal can be drawn on which the active oxygen
from the catalyst surface is apparently able to reach a height of
up to 30 nm in the oxidation of the soot (Fig. 14a). This can only
happen when one considers the transfer of the active oxygen to
the carbon active sites on the soot by the oxygen spillover
mechanism, but not by gas-phase diffusion. If the active oxygen
could diffuse through the gas phase, the high surface area of the
Fig. 14. (a and b) Oxygen spill over mechanism in catalysed soot oxidation [28].
199
CeO2_700 should be more active compared with the
CeReOx_700 catalysts. Furthermore, it is impossible to spread
the primary soot particles on the catalyst surface and the
significant amount of the soot particles can be expected to be
present as clusters (secondary particles). Therefore, the
spillover oxygen is not only occurring at the primary particle
level (30 nm), but also to much greater distances (Fig. 14b).
The activation energy estimation in the tight contact mode
using the catalysts will be reasonably better than that calculated
using the loose contact mode. Because of the good contact
between the catalyst and the soot, the heat dissipation will be
more efficient and its effect will not influence the activation
energy calculations. The apparent activation energy (Ea) is
found to be around 150 kJ mol1 (Table 5), over the various
catalysts studied under the different contact conditions. The Ea
values are also found to be similar at the different soot
conversion levels, though the nature of soot could have changed
with the conversion. The unchanged Ea values under the various
conditions in the present study indicate that the catalyst only
increases the rate of reaction, through the well-known redox
cycle, Ce4+ $ Ce3+. The soot oxidation mechanism involves
the following important steps [3]; (1) the adsorption of the
oxygen, (2) the spillover of the oxygen to the carbon active
sites, (3) the chemisorption of the active oxygen to form the
surface oxygen complexes (SOC) and (4) the desorption of the
SOC. The usage of a catalyst will increase the rate of ‘active
oxygen’ adsorption onto soot surface [12]. Catalysts having the
various reducible properties are expected to show the
differences in the rate of the ‘active oxygen’ adsorption and
the transfer to the soot surface. However, this reaction step did
not influence Ea. Therefore, the ‘active oxygen’ adsorption and
the transfer from the catalyst to the soot surface can be ruled out
as the rate-determining step. The spillover of the oxygen on the
carbon active sites mainly occurs by the surface diffusion and
for such a process Ea can be expected to be much lower than the
150 kJ mole1. Similarly for the desorption of SOC the Ea is
found in the range of 50–70 kJ mole1, which is much lower
than the Ea observed in the soot oxidation. The observed Ea
seems to be due to the chemisorption of the spillover oxygen on
the carbon active sites to form the SOC. Because the role of the
catalyst is not involved in the chemisorption of the active
oxygen to form the SOC, the un-catalysed and catalysed soot
oxidation have the similar Ea values. This also explains why
there was no significant change in the Ea, using catalysts having
the different surface oxidation properties.
200 K. Krishna et al. / Applied Catalysis B: Environmental 75 (2007) 189–200
5. Conclusions
Ceria (CeO2) and the rare-earth modified ceria (CeReOx)
catalysts are characterised thoroughly and various changes in
the surface and bulk properties of the catalysts are correlated to
the observed soot oxidation activity with O2. Rare-earth ion
doped CeO2 generally lead to an improved soot oxidation
activity using O2. CeLaOx presented superior activity when in
the tight contact with soot, decreasing the soot oxidation
temperature by about 150 8C. The soot oxidation with O2 under
the tight contact mode can be strongly correlated with the
external surface area of the catalyst, which is increased by the
rare-earth ion doping. The intrinsic soot oxidation activity of
the CeO2 is higher than the rare-earth ion doped CeO2 catalyst,
which indicates that the doping with the rare-earth cat ions in
fact decreases the activity of CeO2. Rare earth ion doped
catalysts led to a solid-solution formation resulting in a
decreased crystallite size of doped catalysts. The enrichment of
the catalyst surface with an unreducible dopant (still present in
solid solution as in the case of La3+) and the segregation of the
phases (in the case of Sm3+ and Y3+ ions) blocked/decreased the
activity of the potential catalytic sites of CeO2 responsibly for
the soot oxidation. The ease of the reduction or the bulk oxygen
storage capacity is not very important parameter in determining
the soot oxidation over the studied catalysts. Only the first few
surface reducible layers of catalyst participate in the soot
oxidation with O2 are of importance. The similar activation
energies (150 kJ mol1) observed over the various catalysts for
the soot oxidation with O2 indicated that the catalyst is only
facilitating the oxygen transfer to the soot surface and is not
changing the reaction mechanism. The spillover of the oxygen
followed by the chemisorption of the active oxygen is the two
very important steps in the reaction mechanism (Fig. 14).
Acknowledgements
The authors want to thank BASF (former Engelhard
Corporation) for their financial support, and the Spanish
MEC for the contract to ABL.
References
[1] B.J. Finlayson, J.N. Pitts, Science 276 (1997) 1045.
[2] C.M. Somers, B.E. McCarry, F. Malek, J.S. Quinn, Science 304 (2004)
1008.
[3] B.R. Stanmore, J.F. Brilhac, P. Gilot, Carbon 39 (2001) 2247.
[4] R. Allansson, C. Goersmann, M. Lavenius, P.R. Phillips, A.J. Uusimak,
A.P. Walker, SAE Paper 2004-01-0072, Detroit, 2004.
[5] T. Campenon, P. Wouters, G. Blanchard, P. Macaudiere, T. Seguelong,
SAE Paper 2004-01-0071, Detroit, 2004.
[6] J.P.A. Neeft, M. Makkee, J.A. Moulijn, Fuel Proc. Technol. 47 (1996) 1.
[7] B.A.A.L. van Setten, M. Makkee, J.A. Moulijn, Catal. Rev. Sci. Eng. 43
(2001) 489.
[8] J. Suzuki, S. Matsumoto, Top. Catal. 28 (2004) 171.
[9] S.I. Matsumoto, Catal. Today 90 (2004) 183.
[10] E.E. Miro, F. Ravelli, M.A. Ulla, L.M. Cornaglia, C.A. Querini, Catal.
Today 53 (1999) 631.
[11] S. Zhao, R.J. Gorte, Appl. Catal. A 248 (2003) 9.
[12] A. Bueno-López, K. Krishna, M. Makkee, J.A. Moulijn, J. Catal. 230
(2005) 237.
[13] A. Bueno-López, K. Krishna, M. Makkee, J. Moulijn, Catal. Lett. 99
(2005) 203.
[14] X. Wu, D. Liu, K. Li, D. Weng, Catal. Commun. 8 (2007) 1274.
[15] E. Aneggi, C. de Leitenburg, G. Dolcetti, A. Trovarelli, Catal. Today 114
(2006) 40.
[16] T. Hatakeyama, Z. Liu, 1998, p. 47.
[17] J.P.A. Neeft, F. Hoornaert, M. Makkee, J.A. Moulijn, Thermochim. Acta
287 (1996) 261.
[18] J.R. Mcbride, K.C. Hass, B.D. Poindexter, W.H. Weber, J. Appl. Phys. 76
(1994) 2435.
[19] K.M. Ryan, J.P. McGrath, R.A. Farrell, W.M. O’Neill, C.J. Barnes, M.A.
Morris, J. Phys.: Condens. Matter 15 (2003) L49.
[20] J. Kaspar, P. Fornasiero, M. Graziani, Catal. Today 50 (1999) 285.
[21] H. He, H. Dai, C.T. Au, Catal. Today 90 (2004) 245.
[22] A. Hartridge, M.G. Krishna, A.K. Bhattacharya, Mater. Sci. Eng. B: Solid
State Mater. Adv. Technol. 57 (1999) 173.
[23] J.E. Spanier, R.D. Richard, D. Robinson, F. Zhang, S.-W. Chan, I.P.
Herman, Phys. Rev. B 64 (2001) 245407.
[24] W.Z. Weng, H.L. Wan, J.M. Li, Z.X. Cao, Angew. Chem. Int. Ed. 43
(2004) 975.
[25] D. Fino, P. Fino, G. Saracco, V. Specchia, Appl. Catal. B 43 (2003) 243.
[26] V.G. Milt, C.A. Querini, E.E. Miro, M.A. Ulla, J. Catal. 220 (2003) 424.
[27] L. Pesant, J. Matta, F. Garin, M.J. Ledoux, P. Bernhardt, C. Pham, C.
Pham-Huu, Appl. Catal. A 266 (2004) 21.
[28] G. Mul, F. Kapteijn, C. Doornkamp, J.A. Moulijn, J. Catal. 179 (1998)
258.