On the role of external Co sites in NO oxidation and reduction by methane over Co-H-MFI catalysts

Article abstract of DOI:10.1016/j.jcat.2004.04.011

Co-H-MFI and Co-silica-alumina catalysts were comparatively characterized and their activities in CH₄-SCR and NO oxidation were compared. IR measurements evidence that Co sites distribute between internal cavities and external surface of Co-H-MFI; a large fraction of protonic sites are retained. Co-silica-alumina, unlike Co-H-MFI, is inactive in CH₄-SCR and catalyzes only NO oxidation and methane combustion. The CH₄-SCR activity is related to Co sites located in the zeolitic cavities. The role of NO₂ gas as intermediate in the CH₄-SCR reaction appears unlikely. The activity of zeolitic protons in NO oxidation seems negligible in the presence of cobalt.

Full text of DOI:10.1016/j.jcat.2004.04.011

Journal of Catalysis 225 (2004) 536–540
www.elsevier.com/locate/jcat
Research Note
On the role of external Co sites in NO oxidation and reduction by
methane over Co–H-MFI catalysts
G. Bagnasco,a M. Turco,a C. Resini,b T. Montanari,c M. Bevilacqua,c and G. Busca c,∗
a
Dipartimento di Ingegneria Chimica, Università “Federico II” di Napoli, p.le V. Tecchio, 80, I-80125, Napoli, Italy
b
Centro di ricerca Interuniversitario in Monitoraggio Ambientale, Università degli Studi di Genova, via Cadorna, 7, I-17100, Savona, Italy
c
Dipartimento di Ingegneria Chimica e di Processo, Università degli Studi di Genova p.le J. F. Kennedy, 1, I-16129, Genova, Italy
Received 29 March 2004; revised 8 April 2004; accepted 13 April 2004
Available online 2 June 2004
Abstract
Co–H-MFI and Co–silica–alumina catalysts were comparatively characterized and their activities in CH -SCR and NO oxidation were
4
compared. IR measurements evidence that Co sites distribute between internal cavities and external surface of Co–H-MFI; a large fraction
of protonic sites are retained. Co–silica–alumina, unlike Co–H-MFI, is inactive in CH -SCR and catalyzes only NO oxidation and methane
4
combustion. The CH -SCR activity is related to Co sites located in the zeolitic cavities. The role of NO gas as intermediate in the CH -SCR
4
2
4
reaction appears unlikely. The activity of zeolitic protons in NO oxidation seems negligible in the presence of cobalt.
2004 Elsevier Inc. All rights reserved.
Keywords: CH -SCR; NO oxidation; NO SCR; Co-MFI; Co–silica–alumina; Co sites; UV-vis spectroscopy; FT-IR spectroscopy; TPR
4
1. Introduction
Brønsted sites enhance the activity in oxidizing NO to NO2.
According to our recent work [8], the oxidation of NO to
NO2 is not beneficial as NO oxidation seems to be competi-
tive with respect to the overall CH4-SCR process; i.e., NO2
appears not to be an intermediate of the CH4-SCR process.
In the present work we will compare the activity of
Co-MFI zeolite and of Co–silica–alumina in the conversion
of NO in excess oxygen, both in the presence and in the ab-
sence of methane to gain information on the reactivities of
internal channel ions and external surface ions. The role of
NO2 in NO reduction is also discussed.
NO reduction from flue gases of power stations could be
performed using methane as reductant (CH4-SCR), accord-
ing to the following reaction:
CH4 + 2NO + O2 → N2 + CO2 + 2H2O.
Co-containing zeolites, such as Co-MFI and Co-FER, were
found by Armor [1] to be active under conditions that allow
application in high-dust configurations.
Several questions are still open about the active sites of
these catalysts for whose characterization the use of UV-vis
spectroscopy has been much emphasized [2–4]. The reaction
mechanism for these catalysts was also investigated [5,6],
but a substantial disagreement still exists about the role of
NO2 as a possible intermediate in the reaction. According
to Yan et al. [7] Co cations act in the oxidation of NO to
NO2 whereas Brønsted acid sites act in the actual reduction
by methane. On the contrary, Kauchy et al. [5] suggest that
(1)
2. Experimental
2.1. Preparation procedures
The preparation of Co–H-MFI (SiO2/Al2O3 = 50) has
been previously reported [8]. A similar procedure has been
applied to the production of Co–silica–alumina (Co–SA)
starting from silica alumina from STREM Chemicals (SiO2/
Al2O3 = 11.6). The resulting Co loadings are 1.7% for
Co–H-MFI and 11.4% (w/w) for Co–SA. The Co/Al atomic
ratios are 0.76 for Co–SA and 0.55 for Co–H-MFI.
*
Corresponding author. Dipartimento di Ingegneria Chimica e di
Processo, Università degli Studi di Genova, p.le J. F. Kennedy, 1, pad. D,
I-16129 Genova, Italy. Fax: +39 010 353 6028.
E-mail address: Guido.Busca@unige.it (G. Busca).
0021-9517/$ – see front matter
doi:10.1016/j.jcat.2004.04.011
2004 Elsevier Inc. All rights reserved.

G. Bagnasco et al. / Journal of Catalysis 225 (2004) 536–540
537
2.2. Characterization techniques
near 15,000 cm−1. No other absorption occurs in the range
40,000–22,000 cm−1. The spectra we observe are similar to
those reported and deeply discussed by Dedechek et al. [3]
and by Drozdova et al. [4], for Co-MFI and other Co–H
zeolites. The absolute intensity ratio of these absorptions
roughly agrees with the Co amount ratio (Co–SA:Co–H-
MFI ≈ 7). These bands are not present in the spectra of Co-
free H-MFI and SA. The spectra of Co–H-MFI and Co–SA
are very similar qualitatively and can be both assigned
to low-coordination Co2+ species [4]. As discussed else-
where [9], these spectra are similar to those observed when
Co ions are dispersed on different oxide supports, like out-
gassed Co–alumina [10]. This suggests that the nature of
cobalt sites in both samples is similar and that this technique
does not allow differentiation of Co sites in zeolite cavities
from those located on the open surfaces of different oxides.
In Fig. 2 the IR spectra of Co–H-MFI after outgassing
at 773 K, and after subsequent contact at RT with pival-
onitrile vapors, are reported. It has been shown [11,12] that
this molecule, because of its steric hindrance, cannot enter
the cavities of MFI zeolites at RT and low pressure. On the
other hand, it is possible to detect its interaction with exter-
nal silanol groups and with two types of Al3+ Lewis sites of
H-MFI, also located at the external zeolitic surface. On Co–
H-MFI (Fig. 2) it is evident again that PN does not enter the
zeolite cavities. In fact, the band of bridging OH groups at
3620 cm−1 is essentially unperturbed. Interaction with ex-
ternal silanols is evident with their shift from 3747 cm−1 to
near 3440 cm−1. The corresponding CN-stretching band of
H-bonded PN is split at 2235 and 2250 cm−1. These fea-
tures disappear by outgassing. Additionally, a quite strong
band is formed at 2280 cm−1, with a sharp weaker peak at
2319 cm−1, that even increase after outgassing. The main
band at 2280 cm−1 is much more intense and defined than
for H-MFI and is assigned to PN interacting with cobalt
sites. This indicates that part of cobalt sites are located at the
external surface of the zeolite where they can interact with
PN. The spectra of hindered nitriles adsorbed on Co sites at
the external surface of Co–H-MFI are similar to those of the
same complexes on Co–SA.
Diffuse reflectance spectra (DR-UV-vis-NIR) have been
recorded by use of a Jasco V-570. The IR spectra were
recorded on a Nicolet Protégé 460 spectrophotometerFourier
transform instrument. The nitriles used for in situ adsorp-
tion were from Aldrich. Temperature-programmedreduction
(TPR) tests were carried out by a Micromeritics TPD/TPR
2900 instrument equipped with a TCD detector. The analyt-
ical procedures are reported elsewhere [9].
2.3. Catalytic activity measurements
Catalytic tests were performed in a flow laboratory plant
equipped with a fixed-bed reactor operating at atmospheric
pressure. The operation conditions were reacting mixture
composition: NO = 1500 ppm, CH4 = 1500 or zero ppm,
O2 = 25,000 ppm, He balance; T = 523–773 K; GHSV =
30,000–60,000 h−1. The analyses of NO, NO2, and N2O
were performed by a continuous spectrophotometer analyzer
ABB URAS 14 equipped with a NO2 converter. CH4, N2,
CO, and CO2 were detected by a HP 5890 gas chromato-
graph equipped with a TCD detector and a 5A-Porapak Q
molecular sieves double-packed column.
3. Results and discussion
3.1. Catalysts characterization
In Fig. 1 the UV-vis spectra of Co–H-MFI and of Co–
SA are reported, both obtained after evacuation at 773 K for
5 h. In both cases, absorptions are found that can be inter-
preted as a triplet near 19,000 cm−1, near 17,000 cm−1 and
The results of TPR measurements obtained with Co–H-
MFI and Co–SA are reported in Fig. 3 and in Table 1. The
data referring to Co–H-MFI have been reported in Ref. [8].
Neither H-MFI nor SA undergo significant reduction. H2
consumption is due only to the reduction of cobalt species.
The TPR signals of Co–H-MFI are much smaller than those
of Co–SA, according to the lower Co content. The three
components that appear in the curve of Co–H-MFI are at-
tributed to two different quite easily reducible species (pos-
sibly polynuclear and/or nanosized Co oxide particles) and
to hardly reducible zeolitic Co2+ ions [8,13]. The curve of
Co–SA shows two components only, with maxima at 648
and 1043 K. The amount of consumed H2 per mole of Co
is close to that corresponding to the reduction Co2+ → Co0
(Table 1). The high temperature signal represents the reduc-
tion of 77% of the overall Co ions and can be reasonably
Fig. 1. DR-UV-vis spectra of samples Co–H-MFI (a), Co–SA (b), both out-
gassed for 5 h at 773 K.

538
G. Bagnasco et al. / Journal of Catalysis 225 (2004) 536–540
Fig. 2. FT-IR spectra of Co–H-MFI after activation at 773 K (a), in contact with pivalonitrile 10 Torr (b), after successive outgassing at 573 K for 10 min (c).
signal at 648 K must be representative of some aggregated
Co oxide species, that undergo reduction more easily than
isolated Co2+ ions. The corresponding amount of H2 (Ta-
ble 1) shows that in Co–SA less than one-fourth of Co ions is
present in oxide-like form. Co3O4 is the most stable Co ox-
ide phase under the conditions applied in the catalyst prepa-
ration process but the TPR peak temperature is somewhat
lower than that measured for the reduction of a Co3O4 phase
under the same conditions [8]. Moreover, on the basis of the
low total H2 consumption per mole of Co (Table 1) and of
the absence of any spectroscopic evidence, it seems unlikely
that a segregated Co3O4 phase is present on the SA surface.
The TPR signal at 648 K is more likely due to the presence
of polynuclear Co2+ oxo-ions, that are reduced to Co0 nano-
Fig. 3. TPR curves of Co–H-MFI and Co–SA. Note that the pattern of
Co–H-MFI sample is reported increased 1 order of magnitude.
clusters by H2, similar to one of the two species observed for
Co-MFI samples [13].
Table 1
These data show that most cobalt species (those reduced
TPR results: peak temperatures, consumed H amounts and molar ratio
at the highest temperature) are isolated Co2+ with a similar
redox behavior in the two catalysts. In the case of Co–H-MFI
we showed previously [9] that the exchange of the internal
bridging OH groups is far from total; from the intensity of
the IR band of the residual internal OHs we can deduce that
no more than 50% protons have been actually exchanged.
From this datum and from Al and Co contents of the cata-
lyst (Co/Al = 0.55 a.r.), we can deduce that only 45% Co
species are in internal substitution position. Although this
evaluation is very rough, it seems evident that the TPR peak
at 1023 K is due not only to the reduction of internal Co
ions in substitutional position, but also to additional cobalt
in other positions either internal or external. These ions have
very similar redox behavior and UV-vis spectrum with re-
spect to those located on the open surface of Co–SA. More
easily reducible centers are also observed for both samples
2
H /Co
2
Sample
T
n
H /Co
H
peak
2
2
−1
−1
)
(K)
(mmol g
)
(mol mol
Co–H-MFI
559
723
1023
0.03
0.05
0.23
0.10
0.17
0.80
0.31
1.07
Co–SA
648
1043
0.39
1.31
0.21
0.69
1.70
0.90
attributed to the reduction of nearly isolated Co2+ ions. This
reduction occurs with great difficulty, as observed for Co–
H-MFI [8,13], possibly because it needs aggregation of Co
atoms to form Co–Co bonds as a preliminary step. The TPR

G. Bagnasco et al. / Journal of Catalysis 225 (2004) 536–540
539
and part of them is even more easily reducible on Co-MFI
than on Co–SA.
673 K. The selectivity to N2 increases with temperature from
20% at 623 K to 70% at 823 K. In the NO oxidation tests
NO conversion is very close to that observed in SCR tests
at T ꢀ 723 K, whereas it is lower at higher temperature. It
is close to the equilibrium conversion values at T ꢁ 673 K.
These data show that under CH4-SCR conditions,
3.2. Catalytic activity measurements
Catalytic activity of Co–H-MFI, of H-MFI, and of Co–
SA was studied for CH4-SCR and NO oxidation reactions.
NO oxidation tests were carried out by feeding the same
NO and O2 concentrations as in CH4-SCR tests, and no
CH4. Preliminary tests ascertained that no reaction occurred
without catalysts, under our experimental conditions. H-MFI
shows very weak activity in both reactions. In CH4-SCR
tests over Co–H-MFI the reaction products are N2, NO2,
CO2, and H2O whereas N2O and CO were not produced.
In NO oxidation tests only NO2 was formed. The results
are shown in Fig. 4 where total NO conversion and NO2
yield for the samples Co–H-MFI and Co–SA, respectively,
are reported. N2 is the only other N-containing product. The
N2 yield is not plotted in the figures because it is equal to
the difference between total NO conversion and NO2 yield,
in the precision of N balance (see Section 2). The conver-
sions are compared with the equilibrium values of NO oxi-
dation to NO2. The behavior of Co–H-MFI in the CH4-SCR
tests (Fig. 4, top) was previously studied [8]. NO conversion
shows a maximum value of about 30% at 673 K and slightly
decreases with temperature. NO2 is the main product at tem-
peratures lower than 773 K and reaches a maximum value at
1
NO + O2 = NO2,
(2)
2
besides the reduction Reaction (1) also NO oxidation (2)
occurs. Reaction (1) prevails at temperatures higher than
723 K, when NO oxidation (2) becomes thermodynamically
limited. On the other hand, the analysis of the conversion of
methane, that increases with temperature from 15% at 623 K
to 65% at 823 K, provides evidence for the additional occur-
rence of the oxidation of methane to CO2:
CH4 + 2O2 → CO2 + 2H2O.
(3)
At 620 K Reaction (1) does not occur at all, nitrogen
not being detected among the products. This means that, at
this stage, both Reactions (2) and (3) are faster than Reac-
tion (1). By increasing the temperature, Reaction (1) starts
to occur while the selectivity to NO2 decreases and methane
conversion steadily increases. It is evident that at T = 670–
770 K without methane the NO2 yield corresponds to that
forecast by thermodynamics, indicating that the equilibrium
is reached. In the presence of methane the conversion of
NO is higher by an amount corresponding to the fraction
of NO that is reduced by methane to N2. Below 670 K the
equilibrium of Reaction (2) is not reached and conversion in-
creases with temperature because of the increase of reaction
rate. This shows that Co–H-MFI acts as an efficient cata-
lyst for Reaction (2) and, at the same time, catalyses Reac-
tions (1) and (3). At high temperature, when Reaction (2) is
no more favored, Reaction (3) becomes progressively faster
and faster. This means that Reactions (2) and (3) compete
with Reaction (1).
The behavior of Co–SA (Fig. 4, bottom) in the CH4-SCR
reaction is quite different. NO conversion and NO2 yield are
practically the same, evidencing that Co–SA catalyzes the
oxidation of NO to NO2, whereas it does not catalyze at all
the CH4-SCR reaction, no nitrogen being observed among
the reaction products. On the other hand, CH4 conversion
and formation of CO2 are noticeable, indicating the occur-
rence of Reaction (3). The behavior of this catalyst in NO
oxidation is similar to that of Co–H-MFI, showing a maxi-
mum in NO conversion at about 670 K and values close to
the equilibrium at temperatures higher than 673 K. The NO
conversion obtained in these conditions is markedly higher
than that obtained in SCR tests, indicating that the NO oxi-
dation activity of this catalyst is partly inhibited by the pres-
ence of CH4. This effect can be due to H2O (produced from
CH4 oxidation) that can compete with NO for adsorption on
the active sites, thus partially inhibiting NO oxidation. The
similar activity of the two catalysts in NO oxidation at low
temperature, despite the much higher cobalt content in Co–
SA, can be related to the easier reducibility of part of cobalt
Fig. 4. NO and CH conversions and NO yield as a function of tempera-
4
2
ture in tests of CH -SCR and NO oxidation on Co–H-MFI (top) and Co–SA
4
(bottom). The dotted lines represent NO conversion in equilibrium condi-
tions.

540
G. Bagnasco et al. / Journal of Catalysis 225 (2004) 536–540
sites in Co–H-MFI than in Co–SA. The above reported data
show that Co–H-MFI catalyst is active both in CH4-SCR re-
action and in NO oxidation to NO2 and that these reactions
are competing. Actually, CH4-SCR activity appears mostly
above 673 K, i.e., when the oxidation of NO to NO2 becomes
less thermodynamically favored. The Co–SA catalyst, on the
contrary, seems to be inactive in CH4-SCR, but active in NO
oxidation to NO2, as well as in the oxidation of methane to
CO2. Over both catalysts the equilibrium of NO oxidation
to NO2 is reached above 673 K. In the region where equi-
librium is not reached, the activity of Co–H-MFI is a little
higher than that of Co–SA. The characterization data show
that part of the Co centers are located at the exterior of Co–
H-MFI. These external sites look similar to those observed
on the open surface of Co–SA. This suggests that Co ions
located at open external surfaces of zeolites are likely not
(or very poorly) active in CH4-SCR, although being likely
active in NO oxidation to NO2. The active sites for SCR of
NO by methane and other hydrocarbons are likely those lo-
cated in the interior of the zeolite cavities, in agreement with
the data reported by Satsuma et al. [14], who showed that
hydrocarbons not accessing the zeolite cavities do not re-
duce NO over Co zeolites, or give very slow reaction. The
cavity sites are apparently also more active in NO oxidation
than those on the open surface. The data reported here show
that the activity in NO oxidation to NO2 does not neces-
sarily imply CH4-SCR activity. The less NO oxidation is fa-
vored, the more CH4-SCR on Co–H-MFI occurs and the lack
of any correlation between the activity in NO oxidation to
NO2 and CH4-SCR suggests that the two reactions are more
likely competitive than successive. Oxidation of NO can ac-
tually occur at the catalyst surface, although NO2 is not
produced in the gas. Formation of oxidized surface species
like nitrates by NO oxidative adsorption, followed by de-
composition back to NO, has been frequently observed [15].
According to previous work [16], it is likely that the reac-
tion mechanism involves the preliminary adsorption of NO
vate the CH4 molecule by subtracting a hydrogen atom and
then forming a nitro- or nitroso-methane intermediate. The
desorption of NO2 from this intermediate can only occur
when NO2 is thermodynamically stable, but it can, in any
case, be considered as a side reaction occurring through a
parallel pathway. The active intermediate for CH4-SCR is
probably formed in the zeolite cavities. In fact, if formed on
the open surface of Co–SA it is not reduced by methane. The
far higher catalytic activity of Co–H-MFI with respect to
H-MFI in NO oxidation to NO2 allows us to rule out a rele-
vant effect of protonic sites of zeolites in favoring CH4-SCR
due to NO oxidation. Residual protonic sites in the zeolite
cavities could more probably be involved in the formation of
the active intermediate or in the adsorption of the reducing
hydrocarbon.
References
[1] J.N. Armor, Catal. Today 26 (1995) 147.
[2] J. Dedechek, B. Wichterlova, J. Phys. Chem. B 103 (1999) 1462.
[3] J. Dedechek, D. Kauchy, B. Wichterlova, Micropor. Mesopor.
Mater. 35–36 (2000) 483.
[4] L. Drozdova, R. Prins, J. Dedechek, Z. Sobalik, B. Wichterlova,
J. Phys. Chem. B 106 (2002) 2240.
[5] D. Kauchy, A. Vondrovà, J. Dedechek, B. Wichterlova, J. Catal. 194
(2000) 318.
[6] S.A. Beloshapkin, E.A. Paukshtis, V.A. Sadykov, J. Mol. Catal. A 158
(2000) 355.
[7] J.-Y. Yan, H.H. Kung, W.M.H. Sachtler, M.C. Kung, J. Catal. 175
(1998) 294.
[8] C. Resini, T. Montanari, L. Nappi, G. Bagnasco, M. Turco, G. Busca,
F. Bregani, M. Notaro, G. Rocchini, J. Catal. 214 (2003) 179.
[9] T. Montanari, M. Bevilacqua, C. Resini, G. Busca, J. Phys. Chem.
B 108 (2004) 2120.
[10] T. Montanari, G. Busca, in preparation.
[11] M. Trombetta, G. Busca, J. Catal. 187 (1999) 521.
[12] T. Armaroli, M. Trombetta, A. Gutièrrez Alejandre, J. Ramirez Solis,
G. Busca, Phys. Chem. Chem. Phys. 2 (2000) 3341.
[13] X. Wang, H. Chen, W.M.H. Sachtler, Appl. Catal. B 29 (2001) 47.
[14] A. Satsuma, A. Shichi, T. Hattori, CATTECH 7 (2003) 42.
[15] G. Busca, L. Lietti, G. Ramis, F. Berti, Appl. Catal. B 18 (1998) 1.
[16] E. Ivanova, K. Hadjiivanov, D. Klissurski, M. Bevilacqua, T. Armaroli,
G. Busca, Micropor. Mesopor. Mater. 46 (2001) 299.
x−
that is oxidized by O2 forming an adsorbed NOy species
(y = 2, 3) bound to a Co2+ center. This species could acti-