Catalytic removal of NOx and diesel soot over nanostructured spinel-type oxides

Article abstract of DOI:10.1016/j.jcat.2006.05.023

Nanostructured spinel-type oxide catalysts AB₂O₄ (where A = Co and Mn, and B = Cr and Fe), prepared by the solution combustion synthesis method and characterized by BET, XRD, FESEM, TEM, FTIR, and catalytic activity tests, proved to be effective in the simultaneous removal of soot and NO_x, the two prevalent pollutants in diesel exhaust gases in the temperature range of 350-450 °C. The activity order for soot combustion was found to be CoCr₂O₄ > MnCr₂O₄ > CoFe₂O₄, whereas the activity order for NO_x reduction was CoFe₂O₄ > CoCr₂O₄ > MnCr₂O₄. The best compromise between simultaneous abatement of soot and nitrogen oxide was therefore shown by CoCr₂O₄ catalyst; it could promote soot combustion and appreciable NO_x reduction below 400 °C, the maximum temperature reached in the exhaust line of a diesel engine. On the basis of oxygen temperature-programmed desorption, the prevalent catalytic combustion activity of the chromite catalysts could be explained by their higher concentration of suprafacial, weakly chemisorbed oxygen, which contributes actively to soot combustion by spillover in the temperature range of 300-500 °C. Tentative reaction pathways for the simultaneous reduction of NO_x are outlined as well.

Full text of DOI:10.1016/j.jcat.2006.05.023

Journal of Catalysis 242 (2006) 38–47
www.elsevier.com/locate/jcat
Catalytic removal of NO and diesel soot over nanostructured
x
spinel-type oxides
∗
D. Fino , N. Russo, G. Saracco, V. Specchia
Department of Materials Science and Chemical Engineering, Politecnico di Torino, C.so Duca degli Abruzzi 24, 10129 Torino, Italy
Received 22 March 2006; revised 23 May 2006; accepted 25 May 2006
Available online 30 June 2006
Abstract
Nanostructured spinel-type oxide catalysts AB O (where A = Co and Mn, and B = Cr and Fe), prepared by the solution combustion synthesis
2
4
method and characterized by BET, XRD, FESEM, TEM, FTIR, and catalytic activity tests, proved to be effective in the simultaneous removal of
◦
soot and NOx, the two prevalent pollutants in diesel exhaust gases in the temperature range of 350–450 C. The activity order for soot combustion
was found to be CoCr O > MnCr O > CoFe O , whereas the activity order for NOx reduction was CoFe O > CoCr O > MnCr O . The
2
4
2
4
2
4
2
4
2
4
2 4
best compromise between simultaneous abatement of soot and nitrogen oxide was therefore shown by CoCr O catalyst; it could promote soot
2
4
◦
combustion and appreciable NOx reduction below 400 C, the maximum temperature reached in the exhaust line of a diesel engine. On the basis
of oxygen temperature-programmed desorption, the prevalent catalytic combustion activity of the chromite catalysts could be explained by their
higher concentration of suprafacial, weakly chemisorbed oxygen, which contributes actively to soot combustion by spillover in the temperature
◦
range of 300–500 C. Tentative reaction pathways for the simultaneous reduction of NOx are outlined as well.
©
2006 Elsevier Inc. All rights reserved.
Keywords: Diesel particulate combustion; Nitrogen oxides reduction; CoCr O ; MnCr O ; CoFe O spinels; Reaction mechanism
2
4
2
4
2 4
1
. Introduction
blood, and even reach the brain [3]. Some investigations indi-
cate that such particles can induce inheritable mutations [4].
The automotive industry is currently facing serious chal-
lenges to reduce the environmental impact from vehicles since
The increased industrialization, and particularly increased
traffic, in Western and developing nations has been accom-
panied by a negative impact on air quality, the environment,
and human health. Diesel engines have both carbon monoxide
and unburned hydrocarbon outlet concentrations much lower
than those produced by spark-ignition engines. However, even
the most recent diesel engines (e.g., common rail) generate
nitrogen oxides and carcinogenic particulates, whose size (50–
European legislation introduced the Euro 4 step in January
−1
2
005, which for diesel cars corresponds to 0.25 g km of NOx
−1
and 0.025 g km of particulate matter (PM). Despite an 80%
reduction in diesel car emission standards over the levels of 10
years ago, the mean NOx emission level of diesel vehicles con-
tinues to be three times higher than their gasoline counterparts.
These limits cannot be accomplished by engine modifica-
tions [5], fuel pretreatments [6], or simply better tuning of
the combustion process [7]; thus, a convenient way of treating
diesel off-gases is needed. In this context, a possible solution
to reduce the PM emissions of vehicles is to retrofit them with
a particulate filter. Several diesel particulate filter systems are
currently on the market. Wall-flow monoliths are the most com-
mon type [8]. Adjacent channels in the wall-flow filters are
alternatively plugged at each end, thus forcing the gas to flow
through the porous wall, which acts as a filter medium. The
most important issue with diesel traps is the regeneration of the
2
00 nm) falls in the so-called lung-damaging range [1]. Many
toxicologic and epidemiologic studies have established adverse
health effects by particulate matter (PM10, PM2.5). There is in-
creasing evidence that several health effects are associated with
the ultrafine particles (diameter <100 nm) [2]. Recent research
shows that these can penetrate the cell membranes, enter the
*
Corresponding author. Fax: +39 011 5644699.
E-mail address: debora.fino@polito.it (D. Fino).
0021-9517/$ – see front matter © 2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2006.05.023

D. Fino et al. / Journal of Catalysis 242 (2006) 38–47
39
filters. Various strategies for this have been considered, includ-
ing catalyzed filter substrates [9–11], fuel-borne catalysts [12],
continuous regenerating traps [13], electric heaters [14], fuel
burners [15], and microwaves [16].
The research carried out worldwide in this field is aimed
mainly at two different solutions:
process can be formally split into the following two steps (using
the preparation of cobalt chromite as an example):
3
Co(NO3)2 + 6Cr(NO3)3 → 3CoCr2O4 + 12N2 + 30O2 (1)
and
0CO(NH2)2 + 30O2 → 20CO2 + 40H2O + 20N2.
2
(2)
1
. Filtration of particulate in a first catalytic converter, in
which the soot is trapped and burned out due to the pres-
ence of a catalyst thoroughly deposited onto the filter [17],
with NOx eliminated in a second catalytic converter ei-
ther by reaction with suitable reducing agents (e.g., light
hydrocarbons, fuel, ammonia [18]) or by direct decompo-
sition [19].
The first step is the endothermic reaction (1) for the real spinel
synthesis starting from the metal nitrate precursors, whereas the
second step is exothermic and accounts for the reaction (2) be-
tween oxygen derived from nitrates decomposition and urea. Of
course, some direct combustion of urea with atmospheric oxy-
gen cannot be excluded, because the preparation is carried out
◦
in air in an electric oven maintained at 500 C, hosting the pre-
2
. Simultaneous removal of soot and nitrogen oxides in a sin-
gle, suitably catalyzed trap [20–32].
cursor mixture placed in a porcelain pot. The overall set of reac-
tions is markedly exothermic, which leads to a thermal peak in
the reacting solid mixture greatly exceeding 1000 C for a few
◦
The second approach is clearly more ambitious than the first
seconds. To increase this sudden heat release, some NH4NO3
was used. Under these conditions, nucleation of spinel crys-
tals was induced, the growth of these crystals was limited, and
nanosized grains were obtained, as anticipated. Each prepara-
tion batch was tuned to produce 1 g of the desired catalyst.
After preparation, all catalysts were ground in a ball mill at
room temperature and submitted to physical and chemical char-
acterization.
one and has recently been explored in view of the consider-
able advantages that it may offer in terms of both investment
cost and pressure drop reduction. It is well known that 10–15%
of NOx is generally reduced during regeneration of soot-laden
traps. This reduction mostly involves NO2 molecules by direct
reaction with soot to form NO and, to a much lesser extent, N2
and N2O [33]. This reaction, exploited in Johnson Matthey’s
continuously regenerating technology (CRT) [34], is the basis
of the present study. Despite the fact that several technical prob-
lems remain to be solved (e.g., determining the best catalyst
deposition route on the trap and the best trap design so as to
allow intimate contact between the reactants and the catalyst
itself), a more urgent research topic is the development of a suit-
able catalyst capable of promoting both soot oxidation and NOx
reduction at comparatively low temperatures (possibly within
◦
A sample of the fresh CoCr2O4 was aged for 1 h at 900 C
in an oven to induce a decrease in the specific surface area and
to evaluate the effect of this parameter on the catalytic activity.
2.2. Catalyst characterization
Catalyst characterisation was carried out via the following
techniques:
◦
the range typical of diesel exhaust, 150–380 C). A recent pa-
per [35] has shown that mixed-type oxides are among the most
promising catalysts for this purpose.
– Chemical analysis of the catalyst was performed by atomic
absorption spectroscopy (AAS) in a Perkin-Elmer 1100B
spectrophotometer after dissolution in strongly acidic me-
dia, to verify the amount of each constituting element in the
prepared spinels.
– X-ray diffraction (XRD) was done in a PW1710 Philips
diffractometer equipped with a CuKα monochromator to
confirm that the desired spinel crystal structure was actu-
ally achieved.
– Field-emission scanning electron microscopy (FESEM)
with a Leo 50/50 VP with a GEMINI column and trans-
mission electron microscopy (TEM) with a Philips CM
30 T were carried out to analyze the microstructure of the
catalyst crystal agglomerates and the catalyst crystals them-
selves.
In the present work, nanostructured AB2O4 (with A = Co
and Mn, and B = Cr and Fe) spinel-type catalysts were pre-
pared via a tailored technique (i.e., solution combustion syn-
thesis [36]), characterized, and tested for their activity in the
simultaneous abatement of NOx and soot. The nanoscopic di-
mensions of the catalyst crystals thus obtained was pursued so
as to increase the number of contact points between the catalyst
and the particulate, a key feature of the reaction pathway, as
discussed later. On the basis of the experimental work reported
here, some conclusions are then drawn concerning the catalysts
tested.
2
. Experimental
–
Nitrogen adsorption for BET specific surface area (SSA)
measurements were performed with a Micromeritics ASAP
2
.1. Catalyst preparation
2
010.
Three spinel samples (CoCr2O4, MnCr2O4, and CoFe2O4)
– Temperature-programmed desorption (TPD) analyses were
also performed on all of the prepared catalysts in a Ther-
moquest TPD/R/O 1100 analyzer, equipped with a thermal
conductivity detector (TCD) and a quadrupole mass spec-
trometry (MS) detector (Baltzer Quadstar 422). A fixed
were prepared via a highly exothermic and self-sustaining
reaction, the so-called solution combustion synthesis (SCS)
method [36]. This technique is particularly suitable for pro-
ducing nanosized catalyst particles. The combustion synthesis

40
D. Fino et al. / Journal of Catalysis 242 (2006) 38–47
bed of catalyst was enclosed in a quartz tube and sand-
Similar conditions can be obtained by gently shaking the two
counterparts in a vessel. Most runs in the present investigation
were performed under tight contact conditions, because good
reproducibility is mandatory to provide a sound basis for mech-
anistic studies. Reference runs on a loose contact mixture in the
absence of catalyst (noncatalytic soot combustion) or the ab-
sence of soot (simple NO reduction by the catalyst) were also
performed for comparison purposes.
The catalyst–carbon–SiO2 fixed bed was sandwiched be-
tween two quartz-wool layers, whereas the tip of a K-type ther-
mocouple was located well inside the bed itself. The reaction
temperature was controlled through a PID-regulation system
based on the measurements of an external K-type thermocou-
wiched between two quartz wool layers; before each TPD
run, the catalyst was heated under an O2 flow at a rate of
− ◦
1
4
0 Nml min up to 750 C. After 30 min at this temper-
ature as a common pretreatment, the reactor temperature
was lowered to room temperature with the same flow rate
of oxygen, thereby allowing complete oxygen adsorption
over the catalyst. Afterward, helium was fed to the reactor
−
1
at a rate of 10 Nml min , and the reactor was maintained
at room temperature for 1 h to purge any excess oxygen
◦
molecules. The catalyst was then heated to 1100 C at a
◦
−1
constant rate of 10 C min under a helium flow rate of
0 Nml min [37]. The quadrupole detector confirmed
−1
1
◦
that oxygen was the only desorbed species.
ple and varied during each TPR runs from 200 to 700 C at a
◦
−1
◦
–
Two Fourier transform infrared (FTIR) spectra were record-
ed to provide crucial information about the reaction mech-
anism over the most promising catalyst (CoCr2O4), pre-
rate of 5 C min . A 30-min stay at 200 C under He flow was
adopted as a common pretreatment to eliminate possible conta-
minants, such as adsorbed water. The analysis of the outlet gas
was performed via a CO/CO2/N2O NDIR analyzer, a NO/NO2
chemiluminescence analyzer, and an O2 paramagnetic analyzer
by ABB. Nitrogen was measured by gas chromatography using
a Hewlett Packard 5890 Series II, equipped with Porapak Q and
molecular sieve columns and a TCD.
treated as follows: (a) fresh catalyst, as prepared, and
◦
(
b) catalyst treated at 385 C under NO flow (2000 ppmv
in He) for 60 min and then cooled rapidly under pure He
gas flow, with the reactor then promptly removed from the
oven. A 200-mg pellet was formed by pressing a mixture
of each treated catalyst and KBr (1:100 weight ratio). FTIR
transmission spectra were then obtained in the wavenumber
Specific TPR analyses were performed on the most promis-
ing catalyst prepared (CoCr2O4) for the sake of gaining better
insight into the reaction mechanism:
−
1
range 1000–2000 cm using a Bruker Equinox 55 spec-
trophotometer equipped with an MCT cryodetector.
(
i) Runs at various inlet O2 partial pressures (0, 2, 10 vol%;
1000 ppmv NO; balance He).
2
.3. Catalytic activity assessment
(
ii) Runs at various inlet NO concentrations of (500, 1000,
2000 ppmv NO; 10 vol% O2).
The catalytic activity of the prepared catalysts was tested
in a temperature-programmed reaction (TPR) apparatus [38].
A standard gas mixture (1000 ppm NO, 10 vol% O2, balance
(iii) Runs at various catalyst/carbon ratios in the fixed bed (4:1,
9:1, 19:1, 1:0, being the amount of carbon equal to 5 mg,
whenever present; at 1000 ppmv NO, 10 vol% O2, balance
He).
−
6
3
−1
He) was fed at the constant flow rate of 1.66 × 10 Nm s
via a set of mass flow meters (Brooks) to a fixed-bed reactor
enclosed in a quartz tube placed in an electric oven. The tubu-
lar quartz reactor was loaded with 50 mg of a 9:1 by weight
mixture of powdered catalyst and carbon and 150 mg of SiO2
granules (0.3–0.7 mm); this inert material was added to reduce
the specific pressure drops across the reactor and to prevent
thermal runaways (maximum temperature difference across the
3. Results and discussion
All spinel catalyst samples were well crystallized by XRD
analysis (Fig. 1). All diffraction peaks expected according to
the reference JPCDS cards (CoCr2O4, PDF 221084; CoFe2O4,
PDF 221086; MnCr2O4, PDF 751614) were observed. No sec-
ondary phases could be detected by this technique. XRD has
a ± 4% precision, and thus the presence of amorphous phases
cannot be completely excluded. Moreover, chemical analysis
(dissolution and atomic absorption, O2 titration) confirmed,
within the limits related to its intrinsic precision (± 5%), that
the amounts of the various constituting elements (Co, Cr, Fe,
Mn, and O) were consistent with those used in the precursors
and was compatible with the phases detected by XRD.
◦
reactor, 20 C). The catalyst weight-to-flow rate ratio in the
−
3
standard runs was W/F = 27.1 kgsNm . Experiments were
performed by using instead of real diesel soot an amorphous
carbon (printex U by Degussa) with the following properties:
average particle size, 45 nm; 0.34 wt% ash after calcination at
◦
◦
8
00 C; and 12.2 wt% of moisture lost after drying at 110 C.
This material was chosen to avoid any interference due to the
presence of adsorbed HCs, sulfates, or fly ash present in real
diesel soot. Moreover, the carbon used is more difficult to
burn than diesel soot, making the results achieved conservative
In the present context, spinel crystals with size of the same
order of magnitude as that of the particulate nuclei (tens of
nanometers [41]) are expected to provide the highest specific
number of contact points between the soot directly captured
over the catalyst layer of diesel particulate traps [42] and the
crystals constituting such a layer. Fig. 2 shows a TEM picture
of the CoCr2O4 spinel-type catalyst produced via SCS, which
[
9,39,40]. A proper and reproducible mixing between catalyst
and carbon was routinely obtained by ball milling for 15 min.
These mixing conditions are normally known as “tight contact”
conditions [17] and were used in the present study because
they allow for better reproducibility. The soot–catalyst contact
conditions in a real catalytic trap are less intensive, however,
and are usually referred to as “loose contact” conditions [17].
2
−1
displays the highest SSA value (59.0 m g ). The spinel crys-

D. Fino et al. / Journal of Catalysis 242 (2006) 38–47
41
Fig. 1. XRD spectra of the three spinel-type catalysts prepared.
tals of the considered catalyst are much smaller than 20 nm,
perfectly in line with the dimensions of particulate nuclei and
in substantial agreement with the SSA measured. As expected,
this SSA was somehow reduced after calcination at high tem-
◦
perature (900 C), even if it remains higher than that of the two
other catalytic materials prepared.
The FESEM view of the microstructure of the crystal aggre-
gates of the CoCr2O4 catalyst (Fig. 3) enlightens a very foamy
structure, a typical feature of catalysts synthesized by SCS. This
should once again strongly favor the contact between soot and
catalyst even under loose contact conditions and, once coated
on the wall flow trap channel walls, it should allow just a slight
increase in the pressure drop [42].
Beyond the SSA data, Table 1 summarizes the carbon com-
bustion peak temperatures (the temperature corresponding to
the maximum carbon conversion rate to CO2, Tcomb), the maxi-
Fig. 2. TEM micrograph of the CoCr O catalyst crystals.
2
4
Fig. 3. FESEM view of CoCr O catalyst crystal agglomerates obtained by combustion synthesis.
2
4

42
D. Fino et al. / Journal of Catalysis 242 (2006) 38–47
Table 1
the abatement of NOx. The complete set of products is not re-
BET analysis and catalytic activity results
ported here for the sake of clarity; it is provided later in Fig. 6. It
can be stressed that CO2 is not the only carbon oxidation prod-
uct: CO is produced in larger amounts as long as the combustion
temperature range of the different catalysts is shifted to higher
Catalyst
BET
T
T
NO conversion
comb
red
2
−1
◦
◦
(m g
)
( C)
( C)
at T (%)
red
CoCr O
2
Aged CoCr O
59
39
37
10
396
399
427
446
385
389
331
345
36
29
25
43
4
2
4
temperatures. As testified by the lowering of the CO peak ar-
2
MnCr O
2
4
eas in Fig. 4, CO2 selectivity diminishes in the following order:
CoCr2O4 > MnCr2O4 > CoFe2O4 > noncatalytic combustion.
Furthermore, explorative TPR runs, performed by feeding
fixed beds of the CoCr O in the absence of carbon (see Fig. 7
CoFe O
2
4
mum NO conversion achievable, and the related NO abatement
peak temperature (Tred) for all of the catalysts investigated.
The lower the Tcomb value, the higher the activity toward soot
combustion. Because the main goal of the present investigation
was the development of a catalyst for the combined abatement
of soot and NOx, the CoCr2O4 catalyst was selected for fur-
ther investigation because it is characterized by the best carbon
combustion activity, coupled with significant NO conversion.
Despite the fact that the other two catalysts promote NO con-
version at lower temperatures, MnCr2O4 provides the lowest
NO abatement yield, and CoFe2O4 provides the poorest soot
combustion activity.
2
4
later) or with the carbon in the absence of catalyst (see Fig. 5
later), lead to just a small reduction of nitrogen oxides. This
confirms that the simultaneous presence of both carbon and
catalyst is necessary to enable the desired NO reduction. As
described later, NO reduction to N actually occurs only when
2
the simultaneous combustion of soot takes place (see the case
of CoCr O in Fig. 6).
2
4
Transient thermal analysis studies (see the oxygen TPD plots
in Fig. 4b) were quite helpful in understanding the behavior
of the catalysts toward soot oxidation. In line with other stud-
ies on perovskite-type catalysts [36,43,44], the superior activ-
ity of chromites toward carbon oxidation indeed should result
from their significantly higher capacity to deliver active oxy-
gen species compared with that of ferrites. Another key issue is
Furthermore, as anticipated from the data given in Table 1,
the plots in Fig. 4a show a peculiar feature of the CoCr2O4 cat-
alyst: the abatement of soot seems to occur simultaneously with
Fig. 4. Results of the TPR (a) (feed concentrations: O = 10 vol%, NO = 1000 ppmv, He = balance; catalyst–soot mass ratio = 9:1; catalyst–soot contact condi-
2
−
3
tions = tight; W/F = 27.1 kgsNm ) and oxygen TPD (b) runs performed on the prepared catalysts.

D. Fino et al. / Journal of Catalysis 242 (2006) 38–47
43
Fig. 5. NO and CO outlet concentration plots obtained during TPR runs performed with the CoCr O catalyst at different operating conditions (feed composition:
2
2
4
−
3
1
000 ppmv NO, 10 vol% O , He = balance; catalyst–soot mass ratio = 9:1; W/F = 27.1 kgsNm ).
2
Fig. 6. Concentration plots of the outlet gaseous species in a TPR run performed with the CoCr O catalyst under tight contact conditions (feed concentrations:
2
4
−
3
O = 10 vol%, NO = 1000 ppmv, He = balance; catalyst–soot mass ratio = 9:1; W/F = 27.1 kgsNm ).
2
the temperature at which such oxygen species are desorbed: the
lower the temperature, the higher the soot combustion activity.
From this standpoint, Fig. 4b provides insight into why the ac-
tivity order toward soot combustion is CoCr2O4 > MnCr2O4 >
CoFe2O4. Comparing the plots of the CoCr2O4 and MnCr2O4
catalysts suggest that the temperature at which the oxygen
species can be released is even more important than the amount
of such oxygen species. A certain role should also be played
by the SSA (Table 1), which reaches the highest values for the
fresh CoCr2O4 catalyst. However, despite the fact that the aged
CoCr2O4 catalyst exhibits a much lower SSA than the fresh
one, it remains more active toward soot combustion than its Mn-
or Fe-based counterparts.
those related to tight contact. Because most of the runs per-
formed in the present investigation were based on tight contact
conditions between catalyst and carbon, the results cannot be
considered representative of actual operating conditions in a
diesel particulate catalytic trap. In any case, even under loose
contact conditions, where the number of contact points between
the two catalyst and carbon counterparts is expected to be sig-
nificantly lower, the catalytic C/NO/O2 reactions occur signif-
icantly faster than for the reference noncatalytic ones (see the
dotted lines in Fig. 5).
Fig. 6 shows the complete set of gas outlet concentration
plots of the various components involved in a standard TPR run
on the CoCr2O4 catalyst. The scenario is rather complex. CO2
is by far the main carbon oxidation product, with a selectivity
greatly exceeding 95%. CO may play some role as a reducing
agent; however, this role should be minor, because NO reduc-
By comparing the results of the runs performed under loose
or tight contact conditions, Fig. 5 shows how Tcomb and Tred
◦
values related to loose contact are some 60–80 C higher than

44
D. Fino et al. / Journal of Catalysis 242 (2006) 38–47
gaseous molecules. The catalyst acts to some extent as a NOx
storage material. As soon as the temperature exceeded 600–
◦
6
50 C and no carbon remained to exert a reducing effect,
NO/NO2 desorption occurred, as demonstrated by the increased
NO and NO2 outlet concentration to values even higher than
the inlet values. NO chemisorption thus must be considered re-
versible. Finally, we note that this phenomenon seems to be
strictly related to some functional modification of the catalyst
surface (e.g., reduction) occurring during the catalyzed com-
bustion of carbon. In fact, the results plotted in Fig. 7 for the
TPR run carried out in the absence of carbon show only very
limited intrinsic activity of the catalyst toward NO abatement.
The data plotted in Fig. 7 provide the following indications
as well:
–
The role of the catalyst-to-soot ratio on the catalytic com-
bustion of carbon is positive, as expected. The Tcomb value
becomes lower and lower as the ratio increases (Fig. 7b).
This can be attributed mainly to the increased number of
contact points between catalyst and carbon at the beginning
of each TPR run.
–
The absolute value of the negative peak of the NO plot in-
creases with the catalyst-to-soot ratio (Fig. 7a). This is a
sign that the C/O2/NO reaction is simultaneous and simi-
larly influenced by the contact conditions between catalyst
and carbon.
If the role of O2 feed concentration on the studied reaction
system is considered, then the following arguments can be put
forward based on the data plotted in Fig. 8:
Fig. 7. NO (a) and CO (b) outlet concentration plots obtained during TPR runs
2
performed with the CoCr O catalyst at different catalyst-to-carbon weight
2
4
ratios (feed composition: 1000 ppmv NO, 10 vol% O , He = balance; cata-
2
lyst–soot contact conditions = tight).
–
As the O2 concentration increases, the catalytic combustion
of carbon occurs at lower and lower temperatures; that is,
tion by CO on metal oxide catalysts is known to be strongly
inhibited by the presence of oxygen [45], which accounted for
the peak of CO outlet concentration versus temperature
becomes narrower and steeper (Fig. 8b). This finding is not
2
1
0 vol% of the feed gas mixture in all TPR experiments per-
formed.
The main product of NO reduction is N . However, N O for-
surprising, because oxygen is expected to have an evident
promoting effect on carbon combustion.
2
2
– The occurrence of catalytic combustion even in the absence
of oxygen in the feed is conversely quite peculiar. This is
should be ascribed to the intrinsic redox activity provided
to the catalyst by some of its elements (e.g., Cr, Co). In this
case, carbon oxidation is directly related to the reduction
of the catalyst itself and is accompanied by a significantly
◦
mation is rather intensive at low temperatures (around 300 C).
Conversely, nitrogen dioxide outlet concentrations, formed to
some extent from the fed NO and O2 according to the NO +
(
1/2)O2 ↔ NO2 equilibrium, generally follows the NO trend.
NO2 should actually contribute to direct carbon combustion, but
with a minor effect on N2 generation, because the main reduc-
tion product in this case is NO itself, as mentioned earlier [33].
A particular feature of the NOx concentration plots becomes
apparent as soon as most of the carbon conversion has occurred:
NO and NO2 concentrations are somehow depleted, and a sec-
reduced selectivity to CO (down to 35%), with CO as the
2
main product.
– NO reduction occurs simultaneously with carbon combus-
tion independently of the temperature range at which this
reaction occurs (Fig. 8a).
◦
ond negative peak can be seen at temperatures around 450 C,
– For O inlet concentrations >2 vol%, the shape of the NO
2
beyond that ascribable to the simultaneous C/O2/NO reaction
occurring below 400 C. The intensity and the area of this peak
outlet concentration peak is roughly comparable to that
◦
shown in Figs. 5–7, whereas in the absence of O , the NO
2
seem to be related to the amount of catalyst present, if the
curves of Fig. 7 are considered. Furthermore, this second nega-
tive peak is not accompanied by the formation of nitrogen or
chemisorption reduction to N (selectivity >95%) becomes
extremely intensive. It is very likely that under these con-
ditions, NO chemisorbed species play a nonnegligible role
in the carbon oxidation process. Specific runs (not reported
here for space considerations) carried out at different NO
concentrations and in the absence of oxygen confirm this
2
N O, as is the first peak. All of these findings indicate that
2
NO and probably NO2 chemisorption likely occurred, which
hampers the possibility to close nitrogen balances among the

D. Fino et al. / Journal of Catalysis 242 (2006) 38–47
45
Fig. 8. NO (a) and CO (b) outlet concentration plots obtained during TPR runs
Fig. 9. NO conversion (a) and CO outlet concentration (b) plots obtained
2
2
performed with the CoCr O catalyst at different O inlet concentrations (feed
during TPR runs performed with the CoCr O catalyst at different NO inlet
2
4
2
2 4
composition: 1000 ppmv NO, He = balance; catalyst-to-carbon ratio = 9:1;
concentrations (10 vol% O ; He = balance; catalyst-to-carbon ratio = 9:1; cat-
2
−
3
−3
catalyst–soot contact conditions = tight; W/F = 27.1 kgsNm ).
alyst–soot contact conditions = tight; W/F = 27.1 kgsNm ).
hypothesis. It must be stressed, however, that these condi-
tions are unconventional for a lean engine like the diesel
one, in which the O2 concentration in the exhaust ranges
from 2% (heavy load) up to 20% (idle).
Table 2
Main tentative pathways for the simultaneous reduction of NO and combustion
of soot over the CoCr O catalyst (Legend: O , NO = oxygen, nitrogen ox-
2
4
ad
ad
ide chemisorbed over the catalyst; C = active carbon site; C[i, j] = i, j species
f
adsorbed over carbon; g = gas phase)
Finally, the NO conversion plots, shown in Fig. 9 as a
function of the inlet NO concentration in TPR runs over the
CoCr2O4 catalyst, are very useful for drawing inference about
the NO reaction order in the C/NO/O2 reaction system. Be-
cause the NO percent conversion increases with the inlet NO
concentration, such a reaction order should probably be >1.
Tred remains nearly similar, because it is governed by the si-
multaneous occurrence of the combustion process, which is not
markedly influenced by variations in NO inlet concentration
C/O reaction (adapted from [20])
2
O (g) ↔ 2O
(a)
(b)
(c)
(d)
(e)
2
ad
C + O → C[O]
f
ad
C[O] + O → CO (g)(+C )
ad
2
f
C[O] + (1/2)O2 → CO2(g)(+Cf)
C[O] → CO(g)(+Cf)
C/NO/O reaction (adapted from [23–26])
2
NO(g) ↔ NO
(f)
(g)
(h)
(i)
ad
Cf + NOad → C[N, O]
(
carbon conversion plots are omitted in Fig. 9 for this reason).
C[N, O] + NOad → CO2(g) + N2(g)(+Cf)
C[N, O] + NO(g) → CO (g) + N (g)
This parameter does not significantly affect the temperature
range and intensity of the second NO peak, which was previ-
ously attributed to NO chemisorption. This should be a sign
that this last process should in any case attain saturation of the
chemisorption sites created during carbon combustion over the
catalyst surface.
On the basis of the experimental evidence described herein,
as well as reports in the literature [46–51], the reaction steps
listed in Table 2 can be hypothesized. In line with earlier re-
2
2
C[N, O] + NO (+C ) → CO (g) + C[N, N]
(j)
ad
f
2
C[N, N] → C + N (g)
(k)
(l)
2
NO(g) + (1/2)O ↔ NO
2
2
NO (g) ↔ NO + O
(m)
(n)
(o)
(p)
(q)
2
ad
ad
NO (g) + C → NO(g) + CO
2
C[N, N] + O → N O(g) + C
ad
2
N O → O + N
2
ad
2
N O + O → O + N
2
ad
2
2

46
D. Fino et al. / Journal of Catalysis 242 (2006) 38–47
ports [46,47], the key players in carbon oxidation should be the
following ones:
–
–
Oxygen species adsorbed (Oad) on the catalyst surface and
obtained by dissociative chemisorption of gas oxygen (re-
action (a)).
Free carbon sites (C ), particularly active, obtained in the
f
early stages of the combustion process (not considered
here) and renewed over the catalyst surface as long as
combustion is occurring (see below). These sites can be
considered those locations on the carbon surface where an
oxidation process has just occurred. A deeper insight into
this has been provided previously [48].
Fig. 10. Transmission FT-IR spectra obtained for the following materials:
The Oad species attack the reactive free carbon (C ) to give
f
◦
(
a) fresh CoCr O catalyst, as prepared; (b) catalyst treated at 385 C under
2
4
the oxygen-containing active intermediate C[O] over carbon
surface (reaction (b)), which in turn produces CO2 by react-
ing with either Oad (reaction (c)) or molecular oxygen (reac-
tion (d)), thereby leaving another reactive free carbon species
NO flow rate (2000 ppmv) for 60 min and then flushed with He and cooled
rapidly by removing it from the oven.
Finally, oxygen could be involved in the formation of N2O
(
Cf) available for the reaction processes. Evidence of the pres-
(
reaction (o)). In this context, very recent studies [55] have
ence of such intermediate C[O] species was provided by Kureti
et al. [32] using DRIFT spectroscopy. Finally, CO production
from the reaction intermediate (reaction (e)) is hypothesized to
shown that N2O can be decomposed over spinel catalysts above
◦
3
50 C via pathways (p) and (q). This may well be respon-
sible for limiting the overall formation of N2O to rather low
temperatures, as detected experimentally. Deeper insight into
the reaction mechanism perhaps may be obtained using a high-
temperature spectroscopic technique (e.g., DRIFTS [32,56]) to
directly assess the presence of the hypothesised reaction inter-
mediates over the catalyst and carbon surfaces.
account for the fact that CO selectivity is high, but not com-
2
plete.
The active Cf carbon sites are likely to play a crucial role
in the simultaneous reduction of NO as well [49,50]. This
ties NO reduction to the combustion of carbon, which gen-
erates such C sites. C[N,O] species over the carbon surface
f
(
reaction (g)) by combination of a carbon active site Cf and
4
. Conclusions
a NO molecule adsorbed on the catalyst surface, NOad (reac-
tion (f)). The desired nitrogen formation occurs by reaction
of C[N,O] with a further NOad molecule (reaction (h)) or di-
rectly with a gaseous NO molecule (reaction (i)). Evidence of
chemisorbed nitrogen oxides on oxide catalysts for diesel soot
combustion is not infrequent at the temperatures of interest in
the aforementioned reaction [32,35,51]. The FTIR spectra in
Some spinel-type catalysts were prepared, characterized,
and tested for potential applications in the treatment of diesel
exhaust gases as promoters of simultaneous NOx and soot re-
moval. A typical feature of the most promising catalyst tested
(
CoCr2O4) was the simultaneity of NO and soot abatements,
occurring at temperatures lower than those enabled by earlier
developed perovskite-type catalysts [35]. This success is related
in part to the particular synthesis method adopted—solution
combustion synthesis, which enables the formation of fine cat-
alyst crystals (<20 nm) and a very foamy structure of their
agglomerates. This catalyst microstructure helps in maximiz-
ing the number of contact points between catalyst and carbon,
a key step in promoting carbon combustion and, consequently,
NO reduction.
An experimental test campaign is currently in progress to
verify the potential of CoCr2O4 at a catalytic trap level on actual
exhaust gases. Along with this, deeper mechanistic studies are
in progress to possibly provide further evidence beyond that
given here of the proposed mechanisms for this rather complex
solid–solid–gas reaction system.
Fig. 10 for the CoCr O catalyst show in particular that NO
2
4
can be easily adsorbed on the catalyst surface, because the ab-
−
1
sorbance peaks at 1370 and 1395 cm can be ascribed to
formation of surface nitrites on both Cr and Co sites [52,53]
◦
and remain present on the surface of the catalyst at 385 C
(Fig. 10b).
The reaction order >1 versus NO partial pressure can be ex-
plained by the fact that this critical step requires the presence
of two NO-related species (i.e., NO , C[N,O], NO(g)). In line
ad
with this, it may also be hypothesized that N2 may be generated
via a further reaction intermediate C[N,N] (reaction (j)), which
is then decomposed via reaction (k). Based on the experimental
evidence (see Fig. 5), NO2 formation from NO catalytic oxida-
tion (reaction (l)) and, consequently, its decomposition (reac-
tion (m)) should proceed in a rather limited manner. However,
NO should play a certain role in soot combustion according to
2
References
reaction (n) [33], because its production should be maximized
◦
◦
at 300–400 C as a consequence of kinetic (<300 C) and ther-
[1] L. Shirnamé-Moré, Diesel Exhaust: A Critical Analysis of Emissions, Ex-
◦
modynamic limitations (>400 C) [54].
posure, and Health Effects, A Special Report of the Institute’s Diesel

D. Fino et al. / Journal of Catalysis 242 (2006) 38–47
47
Working Group, Health Effects Institute, Cambridge, MA (1995) pp. 221–
42.
[26] Y. Teraoka, K. Nakano, W.F. Shangguan, S. Kagawa, Catal. Today 27
(1996) 107.
2
[
[
[
[
2] D.M. Brown, M.R. Wilson, W. MacNee, V. Stone, K. Donaldson, Toxicol.
Appl. Pharmacol. 175 (2001) 191.
3] G. Oberdörster, Z. Sharp, V. Atudorei, A. Elder, R. Gelein, W. Kreyling,
C. Cox, Inhal. Toxicol. 16 (2004) 437.
4] C.M. Somers, B.E. McCarry, F. Malek, J.S. Quinn, Science 304 (2004)
[27] W.F. Shangguan, Y. Teraoka, S. Kagawa, Appl. Catal. B 12 (1997) 237.
[28] W.F. Shangguan, Y. Teraoka, S. Kagawa, Appl. Catal. B 16 (1998) 149.
[29] Y. Teraoka, K. Kanada, S. Kagawa, Appl. Catal. B 34 (2001) 73.
[30] S. Xiao, K. Ma, X. Tang, H. Shaw, R. Pfeffer, J.G. Stevens, Appl. Catal.
B 32 (2001) 107.
[31] W. Weisweiler, K. Hizbullah, S. Kureti, Chem. Eng. Technol. 25 (2002)
140.
[32] S. Kureti, W. Weisweiler, K. Hizbullah, Appl. Catal. B 43 (2003) 281.
[33] B.J. Cooper, J.E. Thoss, SAE Paper No. 890404 (1989).
[34] R. Allanson, B.J. Cooper, J.E. Thoss, A. Uusimäki, A.P. Walker, J.P. War-
ren, SAE Paper No. 2000-01-0480 (2000).
[35] D. Fino, P. Fino, G. Saracco, V. Specchia, Appl. Catal. B 43 (2003) 243.
[36] A. Civera, M. Pavese, G. Saracco, V. Specchia, Catal. Today 83 (2003)
199.
[37] N. Russo, D. Fino, G. Saracco, V. Specchia, J. Catal. 229 (2005) 459.
[38] D. Fino, N. Russo, G. Saracco, V. Specchia, Chem. Eng. Sci. 59 (2004)
5329.
1008.
5] T. Kawatani, K. Mari, I. Fukano, K. Sugakawa, T. Koyama, Technology
for Meeting the 1994 USA, Exhaust Emission Regulations on Heavy-Duty
Diesel Engine, SAE Paper No. 932654 (1993).
[
6] K. Slodowske, An engine manufacturer’s prospective on diesel fuels and
the environment, in: Proceedings of the SAE Panel Presentation at SAE
Fuels and Lubrificant Meeting, Detroit, 20 October 1993.
[7] H. Hiroyasu, M. Arai, H. Nakanishi, SAE Paper No. 800252 (1980).
[8] A.G. Konstandopoulos, H.J. Johnson, SAE Paper No. 890405 (1989).
[9] D. Fino, N. Russo, C. Badini, G. Saracco, V. Specchia, AIChE J. 49 (2003)
2173.
[
[
[
[
10] P. Marecot, A. Fackhe, L. Pirault, C. Geron, G. Mabilon, M. Prigent,
J. Barbies, Appl. Catal. B 5 (1994) 57.
[39] A. Yezerets, N.W. Currier, H.A. Eadler, A. Suresh, P.F. Madden, M.A.
Branigin, Catal. Today 88 (2003) 17.
[40] A. Yezerets, N.W. Currier, D.H. Kim, H.A. Eadler, W.S. Epling, C.H.F.
Peden, Appl. Catal. B Environ. 61 (2005) 120.
[41] M. Ambrogio, G. Saracco, V. Specchia, C. van Gulijk, M. Makkee, J.A.
Moulijn, Sep. Purif. Technol. 27 (2002) 195.
[42] D. Fino, P. Fino, G. Saracco, V. Specchia, Chem. Eng. Sci. 58 (2003) 951.
[43] D. Fino, N. Russo, G. Saracco, V. Specchia, J. Catal. 217 (2003) 367.
[44] L. Forni, I. Rossetti, Appl. Catal. B 38 (2002) 29.
[45] B. Viswanathan, Catal. Rev.-Sci. Eng. 34 (1992) 337.
[46] A.F. Ahlström, C.U.I. Odenbrand, Appl. Catal. 60 (1990) 143.
[47] A.F. Ahlström, C.U.I. Odenbrand, Appl. Catal. 60 (1990) 157.
[48] B.R. Stanmore, J.F. Brilhac, P. Gilot, Carbon 39 (2001) 2247.
[49] Y. Teraoka, W.F. Shangguan, K. Jansson, M. Nygren, S. Kagawa, Bull.
Chem. Soc. Jpn. 72 (1999) 133.
11] C.L. Liang, K.J. Baumgard, R.A. Gorse, J.E. Orban, J.M.E. Storey, J.C.
Tan, J.E. Thoss, W. Clark, SAE Paper No. 2000-01-1876 (2000).
12] J.M. Valentine, J.D. Peter-Hoblyn, G.K. Acres, SAE Paper No. 2000-01-
1934 (2000).
13] P. Hawker, N. Myers, V.H.Th. Hüthwohl, B. Bates, L. Magnusson,
P. Bronnenberg, SAE Paper No. 970182 (1997).
[14] P. Kojetin, F. Janezich, L. Sura, D. Tuma, SAE Paper No. 930129 (1993).
[15] H. Houben, R. Miebach, J.E. Sauerteich, SAE Paper No. 942264 (1994).
[16] M. Gautam, S. Popuri, B. Rankin, M. Seehra, SAE Paper No. 1999-01-
3
565 (1999).
17] B.A.A.L. Van Setten, M. Makkee, J.A. Moulijn, Catal. Rev.-Sci. Eng. 43
2001) 489.
[
[
(
18] M. Makkee, H.C. Krijnsen, S.S. Bertin, H.P.A. Calis, C.M. van den Bleek,
J.A. Moulijn, Catal. Today 75 (2002) 459.
[
[
19] C.Y. Lee, K.Y. Chai, B.H. Ha, Appl. Catal. Part B Environ. 5 (1994) 7.
20] K. Yoshida, S. Makino, S. Sumiya, G. Muramatsu, R. Helferich, SAE Pa-
per No. 892046 (1989).
[50] W.F. Shangguan, Y. Teraoka, S. Kagawa, Appl. Catal. B Environ. 12
(1997) 237.
[51] A.L. Goodman, E.T. Bernard, V.H. Grassian, J. Phys. Chem. 88 (2001)
6443.
[21] F. Kapteijn, A.J.C. Mierop, G. Abbel, J.A. Moulijn, J. Chem. Soc. Chem.
Commun. 16 (1984) 1085.
[52] J. Laane, J.R. Ohlsen, Prog. Inorg. Chem. 27 (1980) 465.
[53] FDM FTIR Spectra of Minerals and Inorganic Compounds, Copyright©
1997, Fiveash Data Management, Inc. http://www.fdmspectra.com/.
[54] J. Despres, M. Koebel, M. Elsener, A. Wokaun, PSI Scientific Rep. 2001 5
(2002) 64.
[
[
22] N. Kakuta, S. Sumiya, K. Yoshida, Catal. Lett. 11 (1991) 71.
23] K. Matsuoka, H. Orikasa, Y. Itoh, P. Chambrion, A. Tomita, Appl. Catal.
B 26 (2000) 89.
[
24] Y. Teraoka, K. Nakano, S. Kagawa, W.F. Shangguan, Appl. Catal. B 5
(
1995) L181.
[55] N. Russo, D. Fino, G. Saracco, V. Specchia, Catal. Today, in press.
[56] G. Mul, F. Kapteijn, J.A. Moulijn, Carbon 37 (1999) 401.
[
25] W.F. Shangguan, Y. Teraoka, S. Kagawa, Appl. Catal. B 8 (1996) 217.