Kinetics and selectivity of methyl-ethyl-ketone combustion in air over alumina-supported PdOx-MnOx catalysts

Article abstract of DOI:10.1016/j.jcat.2008.11.001

A study is presented of the kinetics and oxidation selectivity of methyl-ethyl-ketone (MEK) in air over bimetallic PdO_x(0-1 wt% Pd)-MnO_x(18 wt% Mn)/Al₂O₃ and monometallic PdO_x(1 wt% Pd)/Al₂

Full text of DOI:10.1016/j.jcat.2008.11.001

Journal of Catalysis 261 (2009) 50–59
Contents lists available at ScienceDirect
Journal of Catalysis
www.elsevier.com/locate/jcat
Kinetics and selectivity of methyl-ethyl-ketone combustion in air
over alumina-supported PdOx–MnOx catalysts
G. Arzamendi a, V.A. de la Peña O’Shea b,1, M.C. Álvarez-Galván b, J.L.G. Fierro b, P.L. Arias c, L.M. Gandía a,∗
a
Departamento de Química Aplicada, Edificio de los Acebos, Universidad Pública de Navarra, Campus de Arrosadía s/n, E-31006 Pamplona, Spain
Instituto de Catálisis y Petroleoquímica, CSIC, Cantoblanco, 28049 Madrid, Spain
Departamento de Ingeniería Química y del Medio Ambiente, Escuela Técnica Superior de Ingeniería de Bilbao, UPV/EHU, Alameda de Urquijo s/n, 48013 Bilbao, Spain
b
c
a r t i c l e i n f o
a b s t r a c t
study is presented of the kinetics and oxidation selectivity of methyl-ethyl-ketone (MEK) in air
Article history:
A
Received 13 August 2008
Revised 22 October 2008
Accepted 3 November 2008
Available online 22 November 2008
over bimetallic PdOx(0–1 wt% Pd)–MnOx(18 wt% Mn)/Al2O3 and monometallic PdOx(1 wt% Pd)/Al2O3
and MnOx(18 wt% Mn)/Al2O3 catalysts. Reaction rate data were obtained at temperatures in the 443–
523 K range and for MEK partial pressures in the reactor feed of between 6.5 and 126.6 Pa. Products
of both MEK combustion and partial oxidation reactions were found. Monometallic Pd/Al2O3 was the
most selective catalyst for complete oxidation whereas the partial oxidation of MEK in the presence of
manganese oxides was significant. The maximum yield for the partial oxidation products (acetaldehyde,
methyl-vinyl-ketone, and diacetyl) was always below 10%. Kinetic studies showed that the rates of CO2
formation over PdOx/Al2O3 were well-fitted by the surface redox Mars–van Krevelen (MvK) kinetic
Keywords:
Methyl-ethyl-ketone (MEK)
Palladium oxides
Manganese oxides
Bimetallic Pd–Mn catalysts
Catalytic combustion
Kinetics
expression and also by
a Langmuir–Hinshelwood (LH) model derived after considering the surface
reaction between adsorbed MEK and oxygen as the rate-determining step. In the case of the Mn-
containing catalysts the MvK model provides the best fit. Irrespective of the model, the kinetic parameters
for the bimetallic Pd–Mn catalysts were between the values obtained for the monometallic samples,
suggesting an additive rather than a cooperative effect between palladium and manganese species for
MEK combustion.
Selectivity
VOCs
© 2008 Elsevier Inc. All rights reserved.
1. Introduction
thermal stability and resistance to water or to deactivation by any
other compound present in the feed stream [3–5].
Catalytic combustion offers interesting advantages over conven-
tional flame combustion. It can be carried out at low temperature,
thereby reducing or even avoiding the use of auxiliary fuel, as
well as reducing the formation of thermal NOx. Moreover, catalytic
combustion is feasible at any fuel-to-oxidant ratio thus allowing
for the treatment of streams containing very low concentrations
of hydrocarbons [1]. There is growing interest in this technol-
ogy for the production of electricity with natural gas turbines, as
well as for the abatement of pollutant volatile organic compounds
(VOCs) [2]. Although there are a wide variety of active catalysts in
combustion reactions, two main groups can be distinguished: no-
ble metals (mainly Pt and Pd) and transition metal oxides (Co, Cr,
Cu, Mn and Ni) that can be present as single or mixed oxides (per-
ovskites, espinels and hexaaluminates). These catalysts may have
very different performances as concerns their activity or ability to
start the ignition of a given VOC at low temperature, selectivity,
The so-called ignition or light-off curves constitute the most
widely used manner for evaluating catalytic performance in hy-
drocarbon combustion studies. They are conversion-reaction tem-
perature graphs obtained by measuring the fuel or VOC conversion
as the reaction temperature is changed, which typically results in
a sigmoid or S-shaped curve [6]. It is thereby possible to deter-
mine the minimum space–time (or amount of catalyst) required to
achieve the complete combustion of the organic compounds con-
tained in a given gas stream at a sufficiently low temperature.
Nevertheless, the suitably design of a combustion reactor requires
knowing the kinetics of the reaction over the selected catalyst. At
present, there is a considerable amount of information in the lit-
erature based on light-off results, whereas kinetic and selectivity
studies are generally scarce.
Methyl-ethyl-ketone (MEK, 2-butanone) is a ketone widely used
as solvent in many industrial applications; however, it is hazardous
and its emission to the atmosphere has to be controlled according
to environmental regulations [7]. Catalytic combustion is a suitable
and relatively simple technology for controlling the emissions of
this VOC [8]. Nevertheless, to the best of our knowledge, there are
only four previous studies on the kinetics of the catalytic combus-
Corresponding author. Fax: +34 948 169606.
E-mail address: lgandia@unavarra.es (L.M. Gandía).
Present address: Instituto Madrileño de Estudios Avanzados en ENERGIA,
*
1
C/Tulipán s/n, 28933 Móstoles, Madrid, Spain.
0021-9517/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2008.11.001

G. Arzamendi et al. / Journal of Catalysis 261 (2009) 50–59
51
tion of MEK. The catalysts considered in these works were metallic
foams [8], Fe2O3 [9], Cr/ZrO2 [10] and Pt/Al2O3 [11].
Table 1
Chemical composition and BET specific surface area of alumina-supported mono-
and bimetallic Mn–Pd samples.
This paper studies the combustion of MEK over mono and
bimetallic alumina-supported Pd–Mn catalysts. The combinations
of certain noble metals (mainly Ag, Pd or Pt) and manganese have
been reported to improve catalytic performance, thermal stabil-
ity and resistance to deactivation by poisoning. Previous studies
with Pd–Mn/Al2O3 catalysts revealed a synergistic effect between
MnOx and PdOx species resulting in a remarkable improvement in
the activity for the complete oxidation of formaldehyde, methanol
[12–14] and methane [15,16]. The aim of this study is to investigate
the performance of Pd–Mn catalysts in the combustion of MEK,
with emphasis in the kinetics and selectivity; issues that have hith-
erto received little attention.
Catalyst
Mn (wt%)
Pd (wt%)
SBET (m2/g)
MnA
16.3
14.3
14.9
15.1
0
0
33.9
36.7
31.9
38.3
77.2
0.25PdMnA
0.50PdMnA
1PdMnA
PdA
0.25
0.49
0.92
0.92
were placed in a pretreatment chamber and degassed at 573 K. All
binding energies (BE) were referred to Al 2p line at 74.5 eV.
2.3. Catalytic performance
The MEK combustion reaction was carried out at atmospheric
pressure in a tubular (8 mm i.d.) fixed-bed Pyrex glass reactor.
Mass flow controllers (Bronkhorst) monitored and controlled the
flow of gases used to obtain the feed mixture and pretreat the cat-
alyst. Two types of catalytic tests were carried out to obtain the
light-off curves and the kinetic runs. In the case of the light-off
experiments, the catalyst (0.10 g) was diluted with 0.20 g of in-
ert solids (Pyrex glass beads, 0.1–0.2 mm) forming a bed of about
10–12 mm in length. A thermocouple placed inside the reactor,
in the center of the catalyst bed, monitored the reaction temper-
ature. Prior to each experiment, the catalyst was treated under
2. Experimental
2.1. Catalysts preparation
A commercial γ -alumina powder (PURALOX HP-14/150, CON-
DEA) was used as a carrier to prepare manganese and palladium
oxide catalysts. Before the incorporation of Mn and Pd, this sup-
port was dried at 423 K and calcined at 1073 K for 4 h. The
Mn-containing catalyst (MnA, where A refers to the alumina sup-
port), with a manganese loading of 18 wt%, was prepared by
impregnation with the appropriate volume (4 ml of solution per
gram of support) of an aqueous solution of manganese nitrate
(Mn(NO3)2·4H2O, Panreac, reagent grade). Excess water was re-
moved in a rotary evaporator at 353 K for 2 h. This catalyst pre-
cursor was then dried at 383 K for 4 h and calcined in static air at
773 K for 4 h. Aliquots of this catalyst were subsequently used to
prepare the bimetallic Mn–Pd catalysts. Suitable amounts of palla-
dium nitrate were employed for this purpose. The procedure was
the same as above described for the raw monometallic Mn catalyst.
Three binary palladium-manganese catalysts were thus prepared
by consecutive impregnation of the manganese sample with aque-
ous solutions of Pd(NO3)·H2O (Fluka, reagent grade). The volume of
solution was selected to achieve a final content of 0.25, 0.50 and
1.0 wt% Pd. The dried impregnates were calcined in air at 773 K for
4 h. These binary catalysts will be referred to hereafter as xPdMnA
(where x stands for the palladium loading as wt%). A monometallic
alumina-supported palladium catalyst containing 1.0 wt% Pd (PdA)
was also prepared according to the same procedure followed in the
bimetallic systems.
−1
100 cm3 min
(STP) of air (Praxair 99.999%) for 1 h at 673 K.
−1
The total feed flowrate was 550 cm3 min
(STP) with a MEK par-
tial pressure of 126.6 Pa (1250 ppmv), resulting in a weight hourly
space velocity (WHSV) referred to the amount of supported cata-
lyst in the reactor and total feed mass flowrate of 425 h−1. The
ignition curves were obtained after pretreatment in the decreasing
temperature mode starting from 673 K. The kinetic experiments
were conducted at constant temperature and the reactor operated
in the differential regime (MEK conversion <15%). After pretreat-
ment, the catalytic bed was allowed to cool down in order to reach
the temperature of the kinetic test (443–523 K). MEK partial pres-
sure in the reactor feed was varied between 6.5 and 126.6 Pa (64–
1250 ppmv). Online analysis of the product stream was performed
on a Hewlett Packard 6890 gas chromatograph, equipped with a
6 ft HayeSep Q column connected to a TCD for CO2 determination,
and an HP-INNOWax 30 m × 0.32 mm i.d. column connected to an
FID for MEK and partial oxidation products: acetaldehyde, acetic
acid, methyl-vinyl-ketone and diacetyl (2,3-butanedione) analyses.
3. Results and discussion
2.2. Catalysts characterization
3.1. Catalysts characterization
Elemental analysis was performed by Inductively Coupled
Plasma Atomic Emission Spectroscopy (ICP-AES) using a Perkin–
Elmer optima 3300 DV spectrophotometer. Specific surface areas
were calculated using the BET method from the nitrogen adsorp-
tion isotherms, recorded at the temperature of liquid nitrogen at
atmospheric pressure (77 K) on a Micromeritics apparatus, model
ASAP-2000, taking a value of 0.162 nm2 for the cross-sectional area
of the N2 molecule adsorbed at 77 K. Prior to the adsorption mea-
surements, samples were degassed at 413 K for 2 h.
3.1.1. Chemical analysis, specific surface area and XRD
The chemical analysis of both manganese and palladium was
performed by ICP-AES, with the analytical values being very close
to the nominal ones. Both chemical composition and specific BET
surface areas are compiled in Table 1. BET surface areas decrease
with respect to the bare alumina support (SBET = 79.9 m2/g) upon
manganese and palladium incorporation, with this decrease being
much higher after manganese introduction, since the high man-
ganese load should give way to big crystallites that may occlude
the mouth of the support pores, thus producing a substantial de-
crease in surface area. The BET surface areas obtained for bimetal-
lic catalysts show that the incorporation of such small amounts of
palladium to the Mn-loaded sample as well, as to the bare alu-
mina, did not substantially alter the specific surface area.
The phase composition of the calcined samples was analyzed
by X-ray diffraction (XRD) using a PANalytical X’Pert Pro Diffrac-
tometer operating with the following parameters: Cu K radiation
α
◦
(λ = 1.5405 Å), 45 mA, 40 kV, Ni filter, 2θ scanning range 5–80 ,
and scan step size of 0.03. Phase identification was made using the
reference database supplied with the equipment.
Photoelectron spectra (XPS) of the fresh and used catalysts were
acquired with a VG Escalab 200R spectrometer equipped with a
The crystalline phases of the mono and bimetallic alumina-
supported catalysts were studied by XRD (Fig. 1). The crystalline
phases found in the calcined alumina before metal impregnation
hemispherical electron analyzer and an Al K
(hν = 1486.6 eV,
α1
−19
1 eV = 1.6302 × 10
J) 120 W X-ray source. The powder samples

52
G. Arzamendi et al. / Journal of Catalysis 261 (2009) 50–59
Table 2
Core electron binding energies (eV) of alumina-supported mono- and bimetallic
Mn–Pd samples.
Catalyst
Al 2p
Mn 2p3/2
Pd 3d5/2
O 1s
MnA
74.5
74.5
74.5
74.5
74.5
641.9
641.9
642.0
641.9
–
–
531.3
531.4
531.3
531.3
531.4
0.25PdMnA
0.50PdMnA
1PdMnA
PdA
337.0
337.1
337.0
336.9
(A)
(A)
(B)
Fig. 2. (A) Mn 2p core-level spectra of MnA and xPdMnA catalysts; (B) Pd 3d core-
level spectra of PdA and xPdMnA catalysts.
(B)
Table 3
Fig. 1. X-ray diffraction patterns of: (A) Mn and Pd monometallic catalysts and
Surface atomic ratios for alumina-supported mono- and bimetallic Mn–Pd samples
derived from photoelectron spectroscopy.
(B) Mn monometallic and Mn–Pd bimetallic catalysts.
Catalyst
Mn/Al
Pd/Al
Pd/Mn
were mainly θ-Al2O3 monoclinic phase (JCPDS 79-1559) and cu-
bic γ -Al2O3 phase (JCPDS 01-1308), with a lower contribution. For
the Pd monometallic system, θ-Al2O3 monoclinic is also the major
observed phase, with a low contribution of cubic γ -Al2O3 phase
(JCPDS 80-0786). On the other hand, rhombohedral α-Al2O3 phase
(JCPDS 80-0786) and monoclinic θ-Al2O3 phase were found in the
mono- and bimetallic Mn-based catalysts. Manganese (IV) oxide
(MnO2, JCPDS 81-1947) is the major manganese-containing crys-
talline phase observed for MnA and the bimetallic catalysts. The
presence of other manganese oxide phases such as Mn2O3 and
Mn3O4 cannot be ruled out, but the characteristic diffraction lines
are very difficult to match, since these phases fit with the diffrac-
tion peaks of alumina. PdO oxide is detected by XRD analysis in
PdA and xPdMnA (x = 0.25 to 1) samples. An increase in the in-
tensity of the most important diffraction lines has been found at
MnA
0.25
0.25
0.42
0.49
–
–
–
0.25PdMnA
0.50PdMnA
1PdMnA
PdA
0.004
0.008
0.011
0.006
0.015
0.019
0.022
–
for the monometallic MnA and for the bimetallic xPdMnA fresh
samples were similar and close to 641.9 eV (Fig. 2). As the Mn
2p3/2 binding energies of Mn3+ do not differ substantially from
that of Mn4+, the conclusion derived from XPS is that Mn3+ and
Mn4+ species seem to coexist on the surface of the catalysts. On
the other hand, the XPS analysis for the Pd systems shows that
binding energies of Pd 3d5/2 peak (Table 2) are conclusive in the
presence of PdO species on the catalyst surface.
The Mn/Al, Pd/Al and Pd/Mn surface ratios are presented in Ta-
ble 3. An increase is observed in Pd surface exposition on both
alumina and manganese oxide phases with Pd content, indicat-
ing a deposition of palladium over both alumina and manganese
oxide phases. These data show that the Pd-free sample (MnA) ex-
hibits the lowest Mn/Al ratio which is higher for the 0.50PdMnA
and 1PdMnA catalysts. This finding could be due to a preferential
deposition of palladium particles over alumina and/or to some dis-
◦
◦
33.8 (101 plane) and 54.8 (112 plane) with increasing Pd content.
3.1.2. XPS
The chemical state of Mn and Pd (in the mono- and bimetallic
systems) and the relative surface proportion of the elements over
the alumina substrate were studied by XPS for all the catalyst sam-
ples. The binding energies of Mn 2p3/2 and Pd 3d5/2 core-levels are
compiled in Table 2. All the binding energies of the Mn 2p3/2 peak

G. Arzamendi et al. / Journal of Catalysis 261 (2009) 50–59
53
Fig. 4. Dependence of the relative energy between the Mn 3s main peak and its
satellite on the formal oxidation state of manganese in pure MnOx oxides and cat-
alysts. The inset Mn 3s spectrum and satellite show how the magnitude of the
splitting (δ BE) is measured.
Table 5
Surface atomic ratios for used alumina-supported mono- and bimetallic Mn–Pd
samples derived from photoelectron spectroscopy.
Catalyst
Mn/Al
Pd/Al
Pd/Mn
MnA-ua
0.44
0.50
0.43
0.41
–
–
–
0.25PdMnA-u
0.50PdMnA-u
1PdMnA-u
PdA-u
0.007
0.010
0.016
0.006
0.013
0.023
0.039
–
(A)
(B)
Fig. 3. (A) Mn 2p core-level spectra of MnA and xPdMnA used catalysts; (B) Pd 3d
core-level spectra of PdA and xPdMnA used catalysts.
a
The catalysts used in reaction are designated as xPdMnA-u (where x stands for
the palladium loading as wt%).
Table 4
Core electron binding energies (eV) of used alumina-supported mono- and bimetal-
lic Mn–Pd samples.
nitude of this splitting is proportional to (2S + 1), where S is the
local spin of the 3d electrons in the ground state [18].
We have firstly recorded the Mn 3s spectra of MnO, Mn3O4,
Mn2O3 and MnO2 pure oxides and measured their splitting val-
ues. As shown in Fig. 4, the Mn 3s spectrum of MnO displays the
largest splitting among all these oxides, and on going from MnO to
MnO2 the magnitude of the splitting decreases linearly. In the next
step we have measured the Mn 3s splitting for MnA and xPdMnA
catalysts and the corresponding values are indicated by arrows in
Fig. 4. From this plot it is clear that the manganese is in an oxi-
dation state between +3 and +4. The extreme cases are MnA and
0.25PdMnA catalysts whose δ BE values of Mn 3s level correspond
to manganese formal oxidation states of +3.76 and +3.37, respec-
tively. These new measurements, which agree with our previous
conclusions derived from the analysis of Mn 2p core-level spectra,
strongly support that Mn3+ and Mn4+ coexist on the surface of
the catalysts, in accordance also with the results obtained by XRD.
Surface atomic ratios were also calculated for the samples used
in the reaction (Table 5). As observed, Pd/Al ratios for bimetallic
catalysts increase in relation to those corresponding to fresh coun-
terparts. However, an increase in the Mn/Al ratio is observed for
used catalysts with a Pd content equal to, or lower than, 0.5% and
a decrease for catalysts with greater Pd content. This behavior pro-
duces a slight decrease in the Pd/Mn ratio for the used bimetallic
sample with 0.25 wt% Pd and an increase for systems with a higher
Pd content. The high increase in the Mn/Al ratio for used MnA and
0.25PdMnA samples in relation to fresh counterparts could be re-
lated to a preferential formation of strongly adsorbed species over
alumina during reaction that decrease the Al 2p signal, increas-
ing Pd/Al ratios in all bimetallic samples and the Mn/Al ratio in
samples with a high alumina surface exposition. In this regard,
it has been reported that aldol condensation reactions during the
combustion of ketones in the presence of supported Mn and Pd
catalysts may lead to the formation of strongly bound surface com-
pounds [19–21]. This seems to be a general feature of carbonyl
compounds carrying hydrogen atoms at their α position when ad-
Catalyst
MnA-ua
Al 2p
Mn 2p3/2
Pd 3d5/2
O 1s
74.5
74.5
641.7
641.8
–
531.2
531.4
0.25PdMnA-u
335.0 (29)
337.1 (71)
335.1 (25)
337.1 (75)
335.0 (21)
337.1 (79)
335.1 (34)
337.2 (66)
0.50PdMnA-u
1PdMnA-u
PdA-u
74.5
74.5
74.5
641.9
641.9
–
531.4
531.4
531.3
a
The catalysts used in reaction are designated as xPdMnA-u (where x stands for
the palladium loading as wt%).
aggregation of manganese particles during palladium impregnation
given the acidity of the nitrate palladium solution (pH ≈ 2), since
the pH of the solution decreases as Pd content increases.
XPS analysis was also performed on Mn and Mn–Pd used cata-
lysts (Fig. 3). The photoelectron binding energies of Mn 2p3/2 and
Pd 3d5/2 core-levels are compiled in Table 4. All the binding en-
ergies of the Mn 2p3/2 peak for the monometallic MnA and for
the bimetallic xPdMnA (x = 0.25 to 1) used catalysts have similar
values to those obtained for the fresh samples, showing that the
reaction does not change the net oxidation state of Mn species.
Furthermore, besides PdO, a certain proportion of metal palladium
species (Pd0) was found on the catalyst surface after reaction. As
observed in previous works [15,16], the presence of manganese
oxide stabilizes PdO against reduction to Pd0 under reaction con-
ditions.
As the Mn 2p3/2 binding energies of Mn3+ do not differ
substantially from that of Mn4+, we followed the approach of
Galakhov et al. [17] to probe the valence state of manganese ions
in the catalysts by studying the Mn 3s energy region. The split-
ting of the Mn 3s core-level spectra originates from the exchange
coupling between the Mn 3s hole and Mn 3d electrons. The mag-

54
G. Arzamendi et al. / Journal of Catalysis 261 (2009) 50–59
−1
0.5 wt% Pt on alumina catalyst at WHSV of 145–582 h
[11], and
100% conversion of 600 ppmv of MEK at 580–640 K over 2.3 wt%
−1
Pt catalysts supported on pillared clays at WHSV of 432 h
[20].
In short, as concerns the light-off curves, the performance of the
Pd–Mn/Al2O3 catalysts is similar to that of supported Pt catalysts
under comparable MEK concentration and space velocity condi-
tions.
3.3. Catalytic performance: formation of partially oxidized products
The question of the possible formation of partial oxidation
products is a very important issue in catalytic combustion, since
this technology is intended for the complete oxidation of VOCs or
methane to CO2 and H2O, whereas the presence of partially oxi-
dized products in flue gases could give rise to environmental and
health concerns. Regarding the combustion of MEK over supported
Pt catalysts, 100% selectivity for the complete oxidation prod-
ucts was found in previous works [11,20]. However, when using
Pd–Mn/Al2O3 catalysts, acetaldehyde, methyl-vinyl-ketone, diacetyl
and trace amounts of acetic acid are also formed. Fig. 6 shows the
yields (moles of MEK converted into a given product/moles of MEK
fed into the reactor) of the main reaction products obtained with
the MnA, 0.50PdMnA (a representative bimetallic catalyst) and PdA
catalysts. The lowest formation of partially oxidized products takes
place over the monometallic Pd catalyst, with maximum yields of
acetaldehyde and methyl-vinyl-ketone below 1% and even lower
yields of diacetyl. However, in presence of manganese oxides, the
formation of these products is favored at low and intermediate
MEK conversion levels. With the MnA and 0.50PdMnA catalysts,
maximum yields of acetaldehyde, diacetyl and methyl-vinyl-ketone
were 5–6%, 2–2.5% and 0.5–1%, respectively. The presence of a
higher or lower amount of Pd, as in the 1PdMnA and 0.25PdMnA
catalysts, does not substantially change these values, which points
to manganese species as the one mainly responsible for MEK par-
tial oxidation. The yields of acetaldehyde, methyl-vinyl-ketone and
diacetyl increase with reaction temperature, passing through a
maximum value near the ignition point, and then decreasing to
become zero at higher temperatures, when selectivity to CO2 is
100%. This can be clearly seen in Fig. 7, where the evolution of the
yields is plotted against MEK conversion. Irrespective of the cat-
alyst, maximum acetaldehyde and diacetyl yields are attained at
about 50 and 30% MEK conversion, respectively. The situation is
more complex in the case of methyl-vinyl-ketone. The maximum
yield of this product is attained at low conversion (20%) over PdA,
whereas it is shifted to 50% MEK conversion in the presence of
manganese species.
Diacetyl is an interesting naturally-occurring vicinal diketone
that is used in the food and beverage industries because of its
characteristic buttery flavor. A number of papers have been ded-
icated to the direct synthesis of diacetyl through the partial oxida-
tion of MEK [25–30]. This reaction is also of interest from a core
perspective, since it represents the functionalization of a molecule
at a carbon atom in the α position of an existing carbonyl group.
McCullagh et al. [28,30] proposed a general reaction network for
the oxidation of MEK that, adapted to the conditions relevant in
catalytic combustion where CO2 and H2O are ultimately produced,
is schematized in Fig. 8. The scheme is complex and includes 4
main reaction pathways in parallel. The first one is the forma-
tion of diacetyl and methyl-vinyl-ketone from a common reaction
intermediate that was identified as acetoin (3-hydroxybutan-2-
one) [30]. The second pathway is the direct complete oxidation of
MEK. The third and fourth reaction pathways give rise to the for-
mation of acetaldehyde and acetic acid. Intermediate 2 (Fig. 8) has
been reported to be a diol (butane-2,3-diol), which is oxidatively
cleaved to form 2 molecules of acetaldehyde [28]. The inclusion
of this reaction is important because it explains the larger pro-
Fig. 5. Ignition curves for MEK combustion in air (PMEK,0 = 126.6 Pa; 1250 ppmv;
WHSV = 425 h−1) over the Pd–Mn series of catalysts: MnA ("), 0.25PdMnA (!),
0.50PdMnA (2), 1PdMnA (1) and PdA (P).
sorbed on the surface of metal oxide-based catalysts with suitable
acid–base properties, as pointed out by Busca et al. [22].
3.2. Catalytic performance: ignition curves
The ignition curves for MEK combustion over the Pd–Mn/Al2O3
series of catalysts are shown in Fig. 5. Curves are depicted as CO2
yield (moles of MEK converted into CO2/moles of MEK fed into the
reactor) versus reaction temperature. If the results obtained at rela-
tively low MEK conversion are considered, it is clear that the activ-
ity increases with Pd content. In fact, taking T10 as the temperature
at which CO2 yield is 10%, this temperature decreases as follows:
555 K (MnA) > 546 K (0.25PdMnA) > 543 K (0.50PdMnA) > 538 K
(1PdMnA) > 516 K (PdA). Thus, in the range of relatively low tem-
peratures, the monometallic Pd/Al2O3 catalyst is the most active
in the series. However, this situation changes at higher tempera-
tures (and MEK conversions), as can be seen from the values of T90
(temperature at which the CO2 yield is 90%): 613 K (PdA) > 602 K
(MnA) > 590 K (0.25PdMnA ≈ 0.50PdMnA) > 578 K (1PdMnA).
This means that in the range of relatively high reaction temper-
atures, the bimetallic 1PdMnA sample becomes the most active
catalyst. Of course, this is the catalyst with the highest load of
metal oxides, but the differences between the performance at low
and high temperatures would suggest some type of cooperative ef-
fect between MnOx and PdOx species in these catalytic systems
[12,16]. As seen in Fig. 5, the shape of the ignition curve obtained
with the PdA catalyst differs notably from those of the other cat-
alyst. Indeed, as previously mentioned, whereas its shape shows
a higher activity at relatively low temperatures (and MEK conver-
sions), the activity is less favored by the increase in temperature
and/or much more detrimentally affected by the decrease in MEK
concentration than for the other catalysts. This behavior indicates
(vide infra) that the apparent activation energy and reaction order
for MEK are lower and higher, respectively, over PdOx than MnOx.
−1
At the weight hourly space velocity of 425 h
and MEK con-
centration in air of 1250 ppmv considered in this work, the Pd–
Mn/Al2O3 catalysts allow for the complete oxidation of MEK into
CO2 at 620 K, whereas a higher temperature of 643 K is required in
the case of the monometallic Pd/Al2O3. For the purpose of compar-
ison with previous studies, Pina et al. [23] reported the complete
combustion of 1750 ppmv of MEK at 473 K over a catalytic mem-
brane containing 0.17 wt% Pt. Irusta et al. [24] obtained 100% con-
−1
version of 1600 ppmv of MEK in air and WHSV of 178 h
at 548
and 573 K over lanthanum-based Mn and Co perovskites, respec-
tively. In previous papers, we reported the complete conversion of
467 ppmv of MEK in air in the 600–650 K range over a commercial

G. Arzamendi et al. / Journal of Catalysis 261 (2009) 50–59
55
Fig. 7. Yields of the MEK oxidation products (PMEK,0 = 126.6 Pa, 1250 ppmv;
−1
) as a function of MEK conversion over the MnA ("),
0.50PdMnA (1) and PdA (P) catalysts.
WHSV = 425
h
Fig. 6. Yields of the MEK oxidation products (PMEK,0 = 126.6 Pa, 1250 ppmv;
WHSV = 425 h−1) as a function of reaction temperature over the MnA, 0.50PdMnA
and PdA catalysts: CO2 ("), acetaldehyde (!, AC), methyl-vinyl-ketone (e, MVK)
and diacetyl (1, DAC).
portion of acetaldehyde compared to acetic acid that is typically
obtained in spite of the fact that the acid is more difficult to oxi-
dize than the aldehyde. Finally, Intermediate 3 (Fig. 8) is the enol
form of MEK (CH3COH=CHCH3), which is oxidized to acetaldehyde
and acetic acid. According to McCullagh et al. [28], the participa-
tion of this intermediate has to be included in any mechanism of
MEK oxidation, especially at high oxygen partial pressures. Takita
et al. [26,27] found that basic oxides (Co3O4, NiO) promoted the
formation of diacetyl, whereas the acidic ones (V2O5, MoO3, WO3,
Cr2O3) mainly catalyzed the scission reactions to C2 moieties.
Fig. 8. Reaction network for methyl-ethyl-ketone (MEK) oxidation: diacetyl (DAC),
methyl-vinyl-ketone (MVK), acetaldehyde (ACH) and acetic acid (AcOH). Scheme
adapted from McCullagh et al. [28,30].

56
G. Arzamendi et al. / Journal of Catalysis 261 (2009) 50–59
The results of this study are basically consistent with the above-
described reaction network. As shown in Fig. 6, at low reaction
temperatures below 450 K there is practically no production of
CO2, and the only reaction products are acetaldehyde and diacetyl
over the Mn-containing catalysts, and acetaldehyde and methyl-
vinyl-ketone over the monometallic Pd catalyst. As the temperature
increases, the yield of all the reaction products increases, until a
maximum yield for MEK partial oxidation products is achieved at
MEK conversions between 20 and 50% depending on the catalyst
and the reaction product being considered. On the other hand, the
rate of the reactions leading to CO2 production from MEK and the
partially oxidized compounds is more favored by the increase in
temperature. As a result, the yield of CO2 increases continuously
until this carbon oxide becomes the only reaction product at tem-
peratures in the 620–643 K range. The main difference between
the results of this study and previous reports on MEK partial oxida-
tion is perhaps the extremely low amounts of acetic acid produced
by the Pd–Mn/Al2O3 catalysts. On the other hand, none of the in-
termediate products has been detected with our analysis method.
This seems to be influenced by the very low MEK concentration
and large oxygen excess prevailing in our experimental conditions
that favor the scission and combustion reactions instead of the par-
tial oxidation preserving the carbon skeleton; as well as, of course,
by the catalyst nature. As a matter of fact, MEK combustion on
Pt catalysts under similar reaction conditions proceeds with the
sole production of CO2 [11,20]. Regarding acetic acid, it should be
noted that this product comes from the cleavage of the enol form
of MEK (Fig. 8). It seems likely that the complete oxidation of this
intermediate is easy under our reaction conditions, thereby lead-
ing to an undetectable acetic acid production. This would indicate
that the cleavage of the diol intermediate (butane-2,3-diol) is the
main source of acetaldehyde. Ketones record high chemical reactiv-
ity on metal oxides, and species such as enolates and carboxylates
have been identified upon their adsorption on transition metal ox-
ides [22,31,32], alumina [33,34] and silica [35,36]. These species,
as well as alkoxides, are considered to be involved in the reac-
tion pathways leading to the combustion of hydrocarbons [22,32,
37,38]. In our case, the role of the support cannot be discarded. In
fact, acetone interacts with Lewis acid sites of γ -Al2O3 through the
oxygen atom to form an enol that is considered to be the interme-
diate in the carbon–carbon bond scission producing CO2 and H2O
at high temperatures [33,34]. Reed et al. [39] have proposed that
the migration of adsorbed acetone intermediates from the silica
support to Mn sites is involved in the mechanism of the complete
oxidation of this ketone with ozone on MnOx/SiO2 catalysts.
3.4. MEK combustion kinetics
This kinetic study is based on the data of CO2 formation rate
(RCO ) obtained at steady state, operating the reactor under a dif-
2
ferential regime of MEK conversion (XMEK < 0.15) and calculated
according to
Fig. 9. Power-law rate model plot for the production of CO2 from MEK combustion
over the MnA, 0.50PdMnA and PdA catalysts at the indicated temperatures.
4 · YCO
2
(RCO ) =
,
(1)
2
ln(RCO ) = lnkPL + α · ln PMEK
,
(3)
2
(W /FMEK,0
)
where kPL is the kinetic constant of the power-law model which
includes dependence on the oxygen partial pressure that can be
assumed constant, since oxygen is in considerable excess with re-
spect to the stoichiometrically necessary amount. PMEK is the MEK
partial pressure and α the apparent reaction order for MEK. A plot
where YCO is the measured yield of CO2 (moles of MEK converted
2
into CO2/moles of MEK fed into the reactor) and W /FMEK,0 the
space–time referred to the amount of catalyst loaded into the reac-
tor (W ) and the MEK molar flow rate in the reactor feed (FMEK,0).
In these experiments the selectivity for CO2 was always above 60%.
of the natural logarithms of RCO and PMEK, including the results
2
obtained with the catalysts MnA, 0.50PdMnA and PdA is shown
in Fig. 9. As can be seen, the power-law model fits well the data
obtained with the monometallic Pd catalyst. In the case of the cat-
alysts containing Mn, the fit is also reasonable, but a dependence
of the apparent reaction order for MEK on the reaction temper-
3.4.1. Power-law rate model
Our first approach was to consider an empirical power-law (PL)
rate model
α
(RCO ) = kPL · PMEK
,
(2)
2

G. Arzamendi et al. / Journal of Catalysis 261 (2009) 50–59
57
In spite of the fact that the variation of the kinetic parame-
ters with the composition and temperature renders the power-law
model invalid, the results obtained can be considered as indicative
of catalyst features. Actually, the results provided by this model
explain the differences among the ignition curves described in
Section 3.2 (Fig. 5). As mentioned earlier, the lowest apparent ac-
tivation energy corresponds to PdA, whereas the apparent reaction
order for MEK is much lower over the Mn-containing catalysts than
over the monometallic Pd sample. The combination of these effects
explains the relatively good performance shown by the bimetal-
lic Pd–Mn catalysts at high values of MEK conversion resulting
in lower temperatures to achieve complete combustion, and also
the superior performance of PdA at relatively low temperatures.
The fact that PdA is the sample best described by the power-law
model can be related to the high selectivity of this catalyst towards
complete oxidation products. In the case of the Mn-containing cat-
alysts, CO2 and water are probably produced from a variety of
substrates. As a result, it is not striking that a simple expression
like Eq. (2) provides worse results with the least selective cata-
lysts.
Fig. 10. Apparent reaction orders for MEK (α) according to the power-law kinetic
model for the combustion of MEK over the Pd–Mn series of catalysts: MnA ("),
0.25PdMnA (!), 0.50PdMnA (2), 1PdMnA (1) and PdA (P).
3.4.2. Mars–van Krevelen and Langmuir–Hinshelwood models
In a previous work, we found that the kinetics of the combus-
tion of MEK in air at low concentrations over a Pt/Al2O3 catalyst
is well described by both the first-order Mars–van Krevelen (MvK)
rate equation and a Langmuir–Hinshelwood (LH) kinetic expres-
sion, derived from a mechanism assuming surface reaction be-
tween MEK and atomic oxygen adsorbed on the same type of sites
as the rate-determining step [11]. These rate equations can be writ-
ten as follows:
kMvK · PMEK
kMvK · PMEK
(RCO
)
=
=
,
(4)
MvK
2
ꢀ
ν·kMvK
1 + kMvK · PMEK
1 +
· PMEK
kox·P O2
ꢀ
kLH · KMEK
·
K
· P ·PMEK
O2
O2
(RCO
)
LH
=
2
2
(1 + KMEK · PMEK
)
ꢀ
kLH · PMEK
Fig. 11. Apparent activation energy according to the power-law kinetic model for the
combustion of MEK over the Pd–Mn series of catalysts: MnA ("), 0.25PdMnA (!),
0.50PdMnA (2), 1PdMnA (1) and PdA (P).
=
,
(5)
2
(1 + KMEK · PMEK
)
where kMvK and kox are the kinetic constants of MEK combustion
and catalyst reoxidation steps according to the Mars–van Kreve-
len mechanism, and ν is the stoichiometric coefficient of O2 for
MEK combustion (5.5). On the other hand, kLH is the kinetic con-
stant of the surface reaction, which is the rate-determining step
ature can be appreciated from the change in the slope of the
straight lines. This effect is more clearly shown in Fig. 10, where
the value of α is displayed as a function of the reaction tempera-
ture for the three representative catalysts. Only in the case of the
monometallic Pd catalyst does the apparent reaction order remain
relatively unaffected, with a value of about 0.6. In contrast, α de-
creases with reaction temperature from around 0.4 at 443 K to
values close to 0 over the bimetallic xPdMnA catalysts and even
to negative values (−0.2) over the monometallic Mn catalyst at
523 K. On the other hand, the corresponding Arrhenius plots show
that the apparent activation energy for MEK combustion decreases
when MEK partial pressure increases in the range of low concen-
tration of the ketone, as shown in Fig. 11. In the case of the PdA
catalyst, this parameter remains almost constant, but the variation
for the Mn-containing samples is remarkable. Taking, for exam-
ple, PMEK = 126.6 Pa (1250 ppmv) as reference, the resulting val-
ues are 70 kJ/mol for PdA, 82 kJ/mol for 1PdMnA, 86 kJ/mol for
MnA and 0.50PdMnA and 95 kJ/mol for 0.25PdMnA. Clearly, the
monometallic Pd catalyst provides the lowest apparent activation
energy, with a value very similar to the one obtained for a 0.5 wt%
Pt/Al2O3 catalyst, also with a power-law kinetic model (73 kJ/mol,
PMEK = 191 Pa) [11]. Higher values are obtained for the catalysts
containing MnOx, which is also in line with results previously re-
ported by other researchers for MEK combustion over transition
metal oxides as Fe2O3 (117 kJ/mol) [9].
in the Langmuir–Hinshelwood mechanism adopted; KMEK and KO
2
are the equilibrium constants for the adsorption of MEK and oxy-
gen, respectively. These kinetic models have also been considered
in this study, and fittings of the rate data to Eqs. (4) and (5) are
shown in Fig. 12 as dotted and solid lines, respectively. As in the
case of Pt, kinetic data from the monometallic Pd catalyst are sat-
isfactorily described by both MvK and LH kinetic models. However,
for the Mn-containing catalysts, fitting to the surface redox Mars–
van Krevelen model is the best at very low MEK partial pressures.
Estimated kinetic and statistical parameters for all the catalysts are
included in Table 6. The statistical parameters evaluated have been
the model selection criterion (MSC) and the coefficient of deter-
mination (DC) [9]. As it can be seen, the values of MSC and DC
are consistently higher for the Mars–van Krevelen than for the
Langmuir–Hinshelwood model, especially for the bimetallic Pd–Mn
catalysts. However, these results of rate data fitting are not conclu-
sive regarding the reaction mechanism. In fact, a recent paper by
Vannice [40] shows that the original derivation of the Mars–van
Krevelen rate equation is inconsistent and that it has no physical
relevance. Moreover, rate expressions derived from more realistic
Langmuir–Hinshelwood models can fit the kinetic data as well as
or even better than the Mars–van Krevelen rate equation. Indeed,

58
G. Arzamendi et al. / Journal of Catalysis 261 (2009) 50–59
Table 6
Estimated kinetic and statisticala parameters of the Mars–van Krevelen (MvK) and
Langmuir–Hinshelwood (LH) rate equations for the combustion of MEK over the
catalysts indicated.b
Parameter
MnA
2.2
0.62
152
0.3
0.021
114
14
18.0
15.0
0.962
0.949
0.25PdMnA 0.50PdMnA 1PdMnA PdA
kMvK × 108 [mol/(gcat Pa s)]
1.5
1.0
1.4
1.2
ꢀ
−1
kMvK [Pa
]
0.21
0.10
0.10
0.02
c
EMvK [kJ/mol]
133
115
117
76
ꢀ
kLH × 108 [mol/(gcat Pa s)]
0.4
0.4
0.7
1.1
−1
KMEK [Pa
]
0.015
0.010
0.014
0.006
d
ELH [kJ/mol]
108
99
94
75
e
−ꢀHMEK,LH [kJ/mol]
10
137.6
34.7
11
35.1
19.7
2.0
44.7
43.5
3.0
70.8
47.4
MSCMvK
MSCLH
DCMvK
DCLH
0.995
0.973
0.990
0.963
0.992
0.987
0.986
0.979
a
The statistical parameters considered are the model selection criterion (MSC)
and the coefficient of determination (DC).
b
Values of the kinetic and equilibrium constants are given at 473 K.
Activation energy for MEK combustion according to the Mars–van Krevelen
c
model.
d
Activation energy for MEK combustion according to the Langmuir–Hinshelwood
model.
e
Enthalpy for MEK adsorption according to the Langmuir–Hinshelwood model.
the power-law model (70 kJ/mol), which also fitted the rate data
well. On the other hand, EMvK was as high as 152 kJ/mol for the
MnA catalyst. In the case of the bimetallic samples, the activation
energy was 115–117 kJ/mol for 0.50PdMnA and 1PdMnA, and in-
creased to 133 kJ/mol for the bimetallic sample with the lowest
Pd content. The kinetic constants at 473 K of the two models, kMvK
ꢀ
and kLH, are similar to each other and from sample to sample. In
contrast, the enthalpy and equilibrium constant for MEK adsorp-
tion are much higher over Mn than Pd species, which indicates a
greater affinity between MEK and manganese oxides.
The presence of a cooperative effect between palladium and
manganese species in PdOx/MnOx catalysts for CO oxidation has
been reported [41,42]. This effect was described considering that
MnOx supplied oxygen onto Pd where the oxidation reaction takes
place. An improved catalytic performance was also found upon in-
creasing the oxidation state of Mn [42]. Requies et al. [16] have
also found a positive effect in the presence of Mn species that
increased the stability of PdO during methane combustion over
Pd–Mn monolith catalysts. However, this reaction is conducted at
much higher temperatures that the ones required for MEK com-
bustion. In our case, the XPS results revealed that the net oxidation
state of Mn does not change during reaction. On the other hand, in
the case of palladium, Pd0 is found in the used samples, whereas
PdO is the dominant species in the fresh catalysts. Moreover, the
analysis of the Mn/Al, Pd/Al and Pd/Mn surface ratios suggests a
preferential deposition of Pd onto alumina and a strong interaction
of Mn with the support. This situation, together with the relatively
low reaction temperatures, would render the manifestation of co-
operative effects of the type described in previous reports more
difficult [16,42]. Therefore, the performance in the MEK combus-
tion reaction of the bimetallic Pd–Mn/Al2O3 catalysts in this study
seems to be dominated by the additive rather than the cooperative
effect of Pd0, PdO and MnOx species.
Fig. 12. Rates of CO2 formation for MEK as
a function of MEK partial pressure
at the indicated temperatures. Points are experimentally obtained values with the
MnA ("), 0.50PdMnA (2) and PdA (P) catalysts. Dotted and solid lines correspond
to the fitting of the kinetic data to the Mars–van Krevelen (Eq. (4)) and Langmuir–
Hinshelwood (Eq. (5)) rate equations, respectively.
4. Summary and conclusions
a LH model including the reaction between oxygen from the gas
phase and adsorbed MEK can lead to a rate equation that is math-
ematically equivalent to the MvK one that fits the kinetic data
equally well [11]. In our case, irrespective of the model, MvK or LH,
the values of the kinetic parameters for the bimetallic Pd–Mn cat-
alysts are intermediate between the values for the monometallic
MnA and PdA samples. This can be clearly illustrated by analyz-
ing the results for the activation energy (EMvK). This parameter
adopted 76 kJ/mol for the PdA catalyst, close to the value given by
The catalytic combustion of MEK in air at low concentrations
takes place over Pd–Mn/Al2O3 catalysts with the formation of ac-
etaldehyde, methyl-vinyl-ketone, and diacetyl as well as CO2 and
water. The partial oxidation products are formed in parallel, being
favored over the Mn-containing catalysts at low and intermediate
levels of MEK conversion, below the ignition point. Monometal-
lic Pd/Al2O3 was the most selective catalyst for MEK complete
oxidation. The only reaction products at 100% MEK conversion

G. Arzamendi et al. / Journal of Catalysis 261 (2009) 50–59
59
were CO2 and H2O over all the catalysts considered in this study.
MnO2 and PdO were the most abundant species in the fresh cat-
alysts. XPS analysis performed on the samples used after reaction
revealed that whereas the oxidation state of Mn did not signif-
icantly change, both Pd0 and PdO coexisted in the palladium-
containing catalysts. In what concerns the combustion kinetics,
the classical first-order redox Mars–van Krevelen (MvK) mech-
anism and a Langmuir–Hinshelwood (LH) model assuming the
surface reaction between adsorbed MEK and oxygen as the rate-
determining step, have been considered. Both models satisfactorily
fitted the rates of CO2 formation obtained with the monometal-
lic PdOx/Al2O3 catalyst. In the case of the bimetallic Pd–Mn cat-
alysts, the MvK model provided the best fit. Irrespective of the
model, the kinetic parameters for the bimetallic catalysts were in-
between the values for the monometallic samples. The activation
energy and MEK enthalpy of adsorption were considerably lower
over PdOx than MnOx. The selectivity and kinetics of MEK com-
bustion over the bimetallic catalysts is dominated by an additive
rather than cooperative effect between palladium and manganese
species.
[11] G. Arzamendi, R. Ferrero, A.R. Pierna, L.M. Gandía, Ind. Eng. Chem. Res. 46
(2007) 9037.
[12] M.C. Álvarez-Galván, V.A. de la Peña O’Shea, J.L.G. Fierro, P.L. Arias, Catal. Com-
mun. 4 (2003) 223.
[13] M.C. Álvarez-Galván, B. Pawelec, V.A. de la Peña O’Shea, J.L.G. Fierro, P.L. Arias,
Appl. Catal. B: Environ. 51 (2004) 83.
[14] V.A. de la Peña O’Shea, M.C. Álvarez-Galván, J.L.G. Fierro, P.L. Arias, Appl. Catal.
B: Environ. 57 (2005) 191.
[15] V.A. de la Peña O’Shea, M.C. Álvarez-Galván, J. Requies, V.L. Barrio, P.L. Arias,
J.F. Cambra, M.B. Güemez, J.L.G. Fierro, Catal. Commun. 8 (2007) 1287.
[16] J. Requies, M.C. Álvarez-Galván, V.L. Barrio, P.L. Arias, J.F. Cambra, M.B. Güe-
mez, A. Manrique Carrera, V.A. de la Peña O’Shea, J.L.G. Fierro, Appl. Catal. B:
Environ. 79 (2008) 122.
[17] V.R. Galakhov, M. Demeter, S. Bartkowski, M. Neumann, N.A. Ovechkina, E.Z.
Kurmaev, N.I. Labachevskaya, Ya.M. Mukoskii, J. Mitchell, D.L. Ederer, Phys. Rev.
B 65 (2002) 113102.
[18] I. Barrio, I. Legórburu, M. Montes, M.I. Domínguez, M.A. Centeno, J.A. Odriozola,
Catal. Lett. 101 (2005) 151.
[19] M. Paulis, L.M. Gandía, A. Gil, J. Sambeth, J.A. Odriozola, M. Montes, Appl. Catal.
B: Environ. 26 (2000) 37.
[20] A. Gil, M.A. Vicente, J.-F. Lambert, L.M. Gandía, Catal. Today 68 (2001) 41.
[21] L.M. Gandía, M.A. Vicente, A. Gil, Appl. Catal. B: Environ. 38 (2002) 295.
[22] G. Busca, E. Finocchio, G. Ramis, G. Ricchiardi, Catal. Today 32 (1996) 133.
[23] M.P. Pina, S. Irusta, M. Menéndez, J. Santamaría, R. Hughes, N. Boag, Ind. Eng.
Chem. Res. 36 (1997) 4557.
[24] S. Irusta, M.P. Pina, M. Menéndez, J. Santamaría, J. Catal. 179 (1998) 400.
[25] M. Ai, J. Catal. 89 (1984) 413.
Acknowledgments
[26] Y. Takita, K. Inokuchi, O. Kobayashi, F. Hori, N. Yamazoe, T. Seiyama, J. Catal. 90
(1984) 232.
[27] Y. Takita, F. Hori, N. Yamazoe, T. Seiyama, Bull. Chem. Soc. Jpn. 60 (1987) 2757.
[28] E. McCullagh, J.B. McMonagle, B.K. Hodnett, Appl. Catal. A: Gen. 93 (1993) 203.
[29] E. McCullagh, N.C. Rigas, J.T. Gleaves, B.K. Hodnett, Appl. Catal. A: Gen. 95
(1993) 183.
G.A. and L.M.G. gratefully acknowledge the financial support for
this work provided by the Spanish Ministry of Science and In-
novation (MAT2006-12386-C05). M.C.A.-G. acknowledges financial
support from the MCYT in the Ramón y Cajal research program.
[30] E. McCullagh, N.C. Rigas, J.T. Gleaves, B.K. Hodnett, Stud. Surf. Sci. Catal. 82
(1994) 853.
References
[31] E. Finocchio, R.J. Willey, G. Busca, V. Lorenzelli, J. Chem. Soc., Faraday Trans. 93
(1997) 175.
[32] G. Busca, E. Finocchio, V. Lorenzelli, G. Ramis, M. Baldi, Catal. Today 49 (1999)
453.
[33] B.E. Hanson, L.F. Wieserman, G.W. Wagner, R.A. Kaufman, Langmuir 3 (1987)
549.
[34] V. Ermini, E. Finocchio, S. Sechi, G. Busca, S. Rossini, Appl. Catal. A: Gen. 190
(2000) 157.
[1] R.E. Hayes, S.T. Kolaczkowski, Introduction to Catalytic Combustion, Gordon and
Breach Science Publishers, Amsterdam, 1997, p. 1.
[2] B.K. Hodnett, Heterogeneous Catalytic Oxidation, John Wiley & Sons, Chichester,
2000, p. 189.
[3] R. Prasad, L.A. Kennedy, E. Ruckenstein, Catal. Rev.-Sci. Eng. 26 (1984) 1.
[4] J.J. Spivey, in: G.C. Bond, G. Webb (Eds.), Catalysis, vol. 8, The Royal Society of
Chemistry, Cambridge, 1989, p. 157.
[35] T. Yamaguchi, Y. Nakano, K. Tanabe, Bull. Chem. Soc. Jpn. 51 (1978) 2482.
[36] J.C. McManus, Y. Harano, M.J.D. Low, Can. J. Chem. 47 (1969) 2545.
[37] E. Finocchio, R.J. Willey, G. Ramis, G. Busca, V. Lorenzelli, Stud. Surf. Sci. Catal.
101 (1996) 483.
[5] M.F.M. Zwinkels, S.G. Järås, P.G. Menon, T.A. Griffin, Catal. Rev.-Sci. Eng. 35
(1993) 319.
[6] F. Duprat, Chem. Eng. Sci. 57 (2002) 901.
[7] G. Ware (Ed.), Reviews of Environmental Contamination and Toxicology,
vol. 103, Springer, Berlin, 1988, p. 165.
[38] J. Haber, Stud. Surf. Sci. Catal. 110 (1997) 1.
[8] J.C. Lou, C.L. Chen, Hazard. Waste Hazard. Mater. 12 (1995) 37.
[9] G. Picasso Escobar, A. Quintilla Beroy, M.P. Pina Iritia, J. Herguido Huerta, Chem.
Eng. J. 102 (2004) 107.
[39] C. Reed, Y. Xi, S.T. Oyama, J. Catal. 235 (2005) 378.
[40] M.A. Vannice, Catal. Today 123 (2007) 18.
[41] S. Imamura, Y. Tsuji, Y. Miyake, T. Ito, J. Catal. 151 (1995) 279.
[42] J.S. Park, D.S. Doh, K.-Y. Lee, Top. Catal. 10 (2000) 127.
[10] V.R. Choudary, G.M. Deshmukh, Chem. Eng. Sci. 60 (2005) 1575.