Supported catalysts for DeNOx reaction based on iron clays

Article abstract of DOI:10.1016/j.molcata.2004.05.025

Different clays have been employed as supports for iron oxide based catalysts prepared by the impregnation method. A basic characterization of the catalysts synthesized has been performed, which includes comparisons between the supports and iron-containing catalysts at bulk and surface levels. The characterization results are correlated with conversion levels achieved during NO_x reduction with propene over the iron clay catalysts. Comparison with the catalytic behaviour obtained over an alumina-supported iron oxide catalyst reveals the promising properties of iron clay catalysts for DeNO _x reactions.

Full text of DOI:10.1016/j.molcata.2004.05.025

Journal of Molecular Catalysis A: Chemical 219 (2004) 309–313
Supported catalysts for DeNO reaction based on iron clays
x
a
a,∗
b
b
C. Belver , M.A. Vicente , M. Fernández-Garc ´ı a , A. Mart ´ı nez-Arias
a
Dpto. Qu ´ı mica Inorgánica, Universidad de Salamanca, Plaza de la Merced, SN, E-37008 Salamanca, Spain
b
Instituto de Catálisis y Petroleoqu ´ı mica, CSIC, Campus Cantoblanco, E-28049 Madrid, Spain
Received 29 March 2004; received in revised form 13 May 2004; accepted 13 May 2004
Abstract
Different clays have been employed as supports for iron oxide based catalysts prepared by the impregnation method. A basic characterization
of the catalysts synthesized has been performed, which includes comparisons between the supports and iron-containing catalysts at bulk and
surface levels. The characterization results are correlated with conversion levels achieved during NO
catalysts. Comparison with the catalytic behaviour obtained over an alumina-supported iron oxide catalyst reveals the promising properties
of iron clay catalysts for DeNO reactions.
2004 Elsevier B.V. All rights reserved.
x
reduction with propene over the iron clay
x
©
Keywords: Clay catalysts; Iron active phase; Nitrogen oxides removal
1
. Introduction
incorporating the metal to the structure of the support, as
in iron zeolites, depositing it over the chosen solid (mainly
silicic or aluminic materials) or creating a new structure, as
in pillaring of smectitic clays with iron polycations. The re-
actions in which these solids have been employed are quite
different: benzene oxidation, ethylbenzene dehydrogena-
tion, methanol conversion or NOx reduction. The elimina-
tion of NOx has become one of the greatest challenges in
environmental protection. These compounds, which are es-
sentially originated from thermal combustion of fossil fuels
in stationary sources and engines of vehicles, are important
atmospheric pollutants, being involved in acid rain, ozone
destruction and photochemical pollution [5,6]. For these
reasons the NOx emissions must be controlled and severe
regulations have been implemented worldwide. Taking this
in mind, several technologies have been developed in or-
der to decrease the harmful NOx emissions; among them,
catalytic removal has shown to be among most efficient
ones, in particular the catalytic reduction employing either
ammonia or hydrocarbons as reductants [5,6].
Iron is an element of great interest in diverse fields such
as metallurgy, biology, geology and industrial processes in
which significant developments involving iron or iron ox-
ides materials like coloured pigments, catalysts, magnetic
materials or ceramics have been carried out. This techno-
logical relevance has motivated the study of iron properties,
focussing the attention over the nature of active states and its
chemistry. Among the latter, most important characteristics
are related to its redox properties, polymerisation capacity
and acidity [1]. An important objective in this respect is
related to the study of the behaviour of iron in solution,
because almost all iron compounds can be tailored from
solutions. Despite numerous studies of processes involved
in the aqueous chemistry of iron, the mechanism of the iron
hydrolysis is not fully understood. The versatility of iron
chemistry in solution arises from the existence of two stable
oxidation states, from the high reactivity of iron complexes
and the sensitivity of the hydrolysis reaction to relatively
minor changes in the experimental conditions [2–4].
The aim of this article is to analyse the preparation and
catalytic activity of supported catalysts based on Fe oxide
prepared by impregnation of different clays, employing
as supports an Al13 pillared clay and an acid activated
metakaolin while an alumina sample was also employed
as a reference support. All these solids were tested in the
catalytic reduction of NOx using propene as the reducing
Recent applications of iron involved the synthesis of het-
erogeneous catalysts. Their synthesis may be carried out
∗
Corresponding author. Tel.: +34 923 294489; fax: +34 923 294574.
E-mail address: mavicente@usal.es (M.A. Vicente).
1
381-1169/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2004.05.025

3
10
C. Belver et al. / Journal of Molecular Catalysis A: Chemical 219 (2004) 309–313
agent. For comparing the efficiency of the different systems
for the reaction we mainly focus our attention on the char-
acterization of the solids; textural and bulk analyses having
been employed for this aim.
atomic absorption spectroscopy (AAS). The X-ray powder
diffraction patterns were carried out with a Siemens D-500
diffractometer, at 40 kV and 30 mA (1200 W), employing
Cu K␣ filtered radiation. Textural analyses were carried
out from the corresponding nitrogen adsorption–desorption
isotherms carried out at 77 K, employing a static volumet-
ric apparatus for this purpose (Micromeritics ASAP 2010
adsorption analyser). Prior to these measurements, the sam-
ples were degassed for 1 h at room temperature, and then at
2
. Experimental
The Fe impregnated solids were prepared by the incip-
◦
ient wetness impregnation method, a conventional method
employed for synthesizing supported catalysts [7]. The iron
component was incorporated over the supports by employ-
ing the ferric acetylacetonate complex, Fe(acac)3, as precur-
sor. The amount of precursor was calculated to obtain a ca.
110 C up to a pressure below 50 ␮mHg. Transmission elec-
tron microscopy (TEM) was performed using a Zeiss-902
microscope. Samples were ground and dispersed in ethanol
by using an ultrasonic apparatus; then, a drop of the sus-
pension was placed in a Cu grid and air dried before the
study.
8
wt.% of Fe2O3 in the final solids. The salt was dissolved
in the amount of acetone appropriate to achieve conditions
close to incipient wetness. The solids obtained were first
The catalysts were tested in the NOx reduction reaction
employing propene as reductant. The spatial velocity em-
◦
◦
−1
dried at 70 C overnight and then heated under air at 500 C
ployed was GHSV = 19,000 h and the feed composition
◦
during 4 h, employing a heating rate of 1 C/min.
of 0.1% C3H and 0.9% NO in Ar (v/v) was employed.
6
The supports employed were an Al13 pillared clay, an
acid activated metakaolin and a commercial alumina. The
Al13 pillared clay was prepared from a natural saponite,
whose characterization has been described in a previous
work [8]. The pillaring process was performed by a conven-
tional procedure [9]. The initial Al-polycation was prepared
by hydrolysis of AlCl3·6H2O precursor with diluted NaOH
The products of the reaction were analysed by infrared
spectroscopy using a Perkin-Elmer 1725X FTIR spectrom-
eter using a multiple reflection transmission cell (Infrared
Analysis Inc.). In all cases, the samples were taken to the
◦
measurement temperature with a ramp of 5 C/min and
stabilized for 45 min prior to analysis, in order to ensure sta-
tionary conditions. The conversion of NOx was calculated
as the ratio between the NOx consumption and the NOx
feed.
−
3+
with a ratio OH /Al = 2.2, followed by ageing at room
temperature for 24 h. This solution was added to previously
prepared clay suspensions (ratio mmol Al3 /g clay = 5)
obtaining a new suspension which, after 24 h ageing, was
washed by centrifugation and dialysis. The solid obtained
+
3. Results and discussion
◦
◦
was dried at 70 C and heated under air at 500 C during 4 h
◦
employing a heating rate of 1 C/min. The acid activated
The structural characterization of the natural clays em-
ployed as supports in this work has been extensively de-
scribed in previous reports [8,10]. In Table 1, the chemical
composition of supports and Fe impregnated solids are sum-
marized. The different nature of the supports employed, with
the metakaolin displaying a high silica content while Al13
is a magnesic silicate containing alumina pillars, is reflected
in the chemical composition. All the samples synthesized
metakaolin was prepared from a natural kaolin described in
a previous paper [10]. The metakaolin was prepared by cal-
◦
cination of the purified kaolin at 600 C for 10 h, employing
◦
a heating rate of 10 C/min. This metakaolin was submitted
to acid activation, a conventional method for improving the
surface and acid properties of clays [11]. The acid solution
employed was 6 M HCl and was added to the metakaolin
◦
with a ratio of 30 mL/g. The suspension was kept at 90 C un-
2 3
have an iron content quite similar, 7–8% Fe O ; considering
der reflux conditions for 6 h with constant stirring. The solid
obtained was separated by centrifugation, washed with dis-
tilled water until no chloride anions could be detected, dried
2 3
the structural amount of Fe O present in the original clays,
amounts very close to 7% are fixed in all impregnations. In
this sense, it must be noted that the impregnation method
permits to control the amount of active phase incorporated,
differences between the three solids being likely related to
◦
◦
at 50 C and heated at 500 C under the same conditions
employed for the former support. The last support employed
was a commercial ␥-alumina (Spheralite 505), chosen as
reference material. This material was supplied by Procatal-
yse and, as done for the other supports, it was submitted to
Table 1
Chemical composition of the supports and Fe impregnated solids
◦
calcination at 500 C before impregnation. The final samples
Samples
SiO2
Al2O3
Fe2O3
MgO
are hereafter referred to as Fe/support, where the supports
are denoted Al13 for the pillared clay, MK for the acid ac-
tivated metakaolin and Al2O3 for the commercial alumina.
Elemental chemical analyses of the solids were performed
by Activation Laboratories Ltd., Ancaster, Ont., Canada, us-
ing inductively coupled plasma spectroscopy (ICPS) and
Al13
MK
Al2O3
Fe/Al13
Fe/MK
Fe/Al2O3
53.54
89.08
–
50.19
82.49
–
14.33
7.86
98.00
13.56
7.87
1.07
0.35
–
8.27
7.11
7.26
28.98
0.34
–
26.91
1.02
–
93.04

C. Belver et al. / Journal of Molecular Catalysis A: Chemical 219 (2004) 309–313
311
Fig. 2. TEM image obtained for the iron pillared clay.
2
values, 150–180 m /g, but they differ in their porosity prop-
erties. The MK and the pillared clay present microporosity,
which corresponds to 50 and 75%, respectively, of the to-
tal area, while the commercial alumina is not a porous ma-
terial. When these supports are impregnated with the iron
precursor, their surface properties can become modified as
a function of the support material employed. Thus, the iron
2
Fig. 1. Diffraction patterns of the supports and iron impregnated samples.
pillared clay presents a high surface area, 217 m /g, higher
than the original support, which may be explained by the in-
corporation of the Fe-complexes to the internal galleries of
the pillared clay, expanding these galleries during the elim-
ination of the gases generated by combustion of the organic
moieties during the calcination process. The microporosity
surface is also higher, ca. of 60% of the total area, which
indicates that the impregnation method induces relatively
minor changes in the surface properties of this support. In
contrast, the metakaolin suffers an important modification
their different water content during the various steps of the
synthesis procedure.
The structural characteristics of the supports are not quite
modified upon iron incorporation by impregnation. Thus,
the X-ray diffraction patterns of the iron-containing solids
show the same reflections as the support samples (Fig. 1).
This behaviour means that the supports are quite stable; even
the pillared clay maintains its layered ordered structure, its
basal plane distance remaining constant at 18.7 Å. Notewor-
thy, although the iron percent in the samples is quite high,
the diffraction patterns show no evidence with respect to the
formation of any crystalline phase of iron oxide or iron oxi-
hydroxide, as could be expected to be present after calcining
after iron incorporation, its surface area is reduced from
2
1
50 to 30 m /g while the microporosity almost disappears.
These data reveal the low stability of the metakaolin as cat-
alytic support. It is an unstable solid which is converted into
low surface area silicates during the deposition of an active
phase by impregnation, probably due to the sintering of the
particles during the synthesis process. This behaviour can
explain the fact that metakaolin is not usually employed as
catalytic support. Thus, some works can be found on the em-
ployment of acid activated kaolins directly as catalysts, but
not as catalyst supports, and mainly used for acid-catalysed
reactions, due to the improvement of the acidic and textural
properties achieved during the acid treatment [12–14]. The
iron impregnated alumina sample presents a surface area
quite similar to the original support. This material is quite
stable, its textural and structural properties do not apprecia-
bly change as a consequence of the impregnation process,
which justifies its common use as catalytic support [15,16].
The catalytic properties of the iron clay samples were
studied for NOx reduction employing a hydrocarbon, namely
propene, as reductant agent. The reaction monitored was
◦
the solids at 500 C. This effect may be due to the formation
of small iron oxide particles very disperse over the supports,
as can be seen in Fig. 2 where the TEM image shows small
dark zones attributable to iron-containing particles over the
pillared clay particles.
The surface properties of the solids are summarized in
Table 2. All supports have analogous specific surface area
Table 2
Surface properties of the supports and Fe impregnated solids
2
2
2
Samples
SBET (m /g)
SMP (m /g)
SEXT (m /g)
Al13
MK
Al2O3
Fe/Al13
Fe/MK
Fe/Al2O3
162
150
185
217
30
124
75
–
116
5
38
75
185
101
25
NOx +C3H under stoichiometric conditions, where the ex-
6
191
–
191
pected products could likely be CO2, H2O and N2. Accord-

3
12
C. Belver et al. / Journal of Molecular Catalysis A: Chemical 219 (2004) 309–313
Fig. 3. NOx and propene conversion levels observed for the iron impregnated clay samples in comparison with a Fe/Al2O3 reference catalyst.
ingly, the effluent analysis detected those products although
a small amount of NH3 was also detected as byproduct in
concentration below 500 ppm. The NOx conversions for the
samples studied are shown in Fig. 3, along with the respec-
tive hydrocarbon conversions. The Fe/Al13 and Fe/Al2O3
catalysts exhibit high NOx conversion levels, values close
4. Conclusions
These results show the most optimum behaviour of the
alumina as catalytic support, particularly at lower reac-
tion temperatures. Nevertheless, the high conversion levels
reached by the iron pillared clay opens a promising way to
obtain functional catalysts for this reaction from pillared
clays. In contrast, the kaolinitic material does not appear
suitable for this reaction. The loss of its original surface
properties after the impregnation process may be the reason
of its low catalytic activity.
◦
to 80% being achieved at 500 C. However, differences
◦
between them are somewhat greater at 450 C, at which
the Fe/Al2O3 achieves a conversion level of 70% whereas
the Fe/Al13 only reaches 50%. This difference does not ap-
pear to correlate closely with the hydrocarbon conversion
levels comparatively attained by the samples, levels higher
than 90% being achieved in both cases at the mentioned
temperatures. Such slight differences in the activities of
Fe/Al13 and Fe/Al2O3 may be due to differential influences
of reaction intermediates in each case, in turn attributable
to small changes in the physicochemical properties of the
iron oxide particles present in each case as a consequence
of differences in the dispersion levels obtained for the in-
corporated metal. This hypothesis, rather than attribution of
the observed differences to changes in the support nature, is
mainly based on the fact that this reaction has been shown
to be mainly influenced by the chemical and electronic
properties of the active iron oxide phase of the catalysts
Acknowledgements
CB and MAV are grateful for the financial support from
Ministerio de Ciencia y Tecnolog ´ı a (MAT2002-03526) and
Junta de Castilla y León (SA012/04). CB acknowledges the
predoctoral grant from Ministerio de Ciencia y Tecnolog ´ı a,
PFI Program (ref. PN-05424782). MFG and AMA thank the
CSIC for financial support.
References
[
17].
The behaviour of Fe/MK is quite different to that exhibited
by the other systems. Its NOx conversion is quite lower, only
0% being achieved, while the propene conversion does not
[1] C.M. Flynn, Chem. Rev. 84 (1984) 31.
[
2] M. Schultz, W. Burckhardt, S.T. Barth, J. Mater. Sci. 34 (1999) 2217.
3] T.G. Spiro, S.E. Allerton, J. Renner, A. Terzis, R. Bils, P. Saltman,
J. Am. Chem. Soc. 88 (1996) 2721.
[
4
reach 65%. The structural collapse evidenced in the textural
analysis of this system certainly plays a significant role on
this catalytic behaviour. In contrast, the high acidity of this
support, mainly Brönsted acidity created by the acid activa-
tion treatment, does not appear to have a particular role in
improving the catalytic behaviour of the final impregnated
catalysts.
[
4] J.P. Jolivet, C. Chanéac, E. Tronc, Chem. Commun. (2004) 481.
[5] A. Fritz, V. Pitchon, Appl. Catal. B Environ. 13 (1997) 1.
[6] V.I. P aˆ rvulexcu, P. Grange, B. Delmon, Catal. Today 46 (1998) 233.
[7] M. Che, O. Clause, Ch. Marcilly, Preparation of Solid Catalysts,
Wiley-VCH, Weinheim, 1999, p. 185.
[8] C. Belver, M.A. Bañares-Muñoz, M.A. Vicente, Appl. Catal. B En-
viron., 50 (2004) 101.
[9] N. Lahav, U. Shani, J. Shabtai, Clays Clay Miner. 26 (1978) 107.

C. Belver et al. / Journal of Molecular Catalysis A: Chemical 219 (2004) 309–313
10] C. Belver, M.A. Bañares-Muñoz, M.A. Vicente-Rodr ´ı guez, Chem.
[14] R.J. Lussier, J. Catal. 129 (1991) 225.
313
[
[
Mater. 14 (2002) 2033.
[15] A.J. van Dillen, R.J.A.M. Terörde, D.J. Lensveld, J.W. Geus, K.P.
de Jong, J. Catal. 216 (2003) 257.
11] M.A. Vicente, C. Belver, R. Trujillano, M. Suárez, M.A. Bañares,
V. Rives, Applied Study of Cultural Heritage and Clays, Biblioteca
de Ciencias, CSIC, Madrid, 2003, p. 519.
[16] B.N. Shelimov, J.-F. Lambert, M. Che, B. Didillon, J. Mol. Catal.
A Chem. 158 (2000) 91.
[12] J.C.D. Macedo, C.J.A. Mota, S.M.C. de Menezes, V. Camorim, Appl.
Clay Sci. 8 (1994) 321.
[17] K.A. Bethke, M.C. Kung, B. Yang, M. Shah, D. Alt, C. Li, K.K.
Hung, Catal. Today 26 (1995) 169.
[13] M. Perissinotto, M. Lenarda, L. Storaro, R. Ganzerla, J. Mol. Catal.
A Chem. 121 (1997) 103.