Kinetics of propane oxydehydrogenation on metals oxides and metals phosphates catalysts: Evidence of compensation effects

Article abstract of DOI:10.1002/kin.20154

Kinetics of oxidative dehydrogenation of propane has been tested on three groups of catalysts; alumina-supported metal oxides (MO) (where metal is V, Cr, Ni, Z r, Mo, or Ba), alumina-supported rare earth metal oxides (RO) (where metal is Ce, Tb, Dy, Ho, T

Full text of DOI:10.1002/kin.20154

Kinetics of Propane
Oxydehydrogenation on
Metals Oxides and Metals
Phosphates Catalysts:
Evidence of Compensation
Effects
B. Y. JIBRIL
Petroleum and Chemical Engineering Department, Sultan Qaboos University, P.O. Box 33, AlKhod, PC 123, Muscat, Oman
Received 13 October 2004; accepted 11 October 2005
DOI 10.1002/kin.20154
Published online in Wiley InterScience (www.interscience.wiley.com).
ABSTRACT: Kinetics of oxidative dehydrogenation of propane has been tested on three groups
of catalysts; alumina-supported metal oxides (MO) (where metal is V, Cr, Ni, Zr, Mo, or Ba),
alumina-supported rare earth metal oxides (RO) (where metal is Ce, Tb, Dy, Ho, Tm, or Yb),
and metal phosphates (MP) (where metal is V, Cr, Mn, Ni, Zr, Mg, Ba, or Ce). They were found to
be active and exhibited different selectivities to propylene, ethylene, and CO (CO and CO ).
x
2
The Arrhenius parameters—apparent pre-exponential factor (ln Aapp) and activation energy
(Eapp)—were evaluated. Evidence of compensation effects was established through statistically
significant linear relationship between ln Aapp and Eapp. The rates of propane conversions and
CO productions on MO and MP showed common compensation line different from that of RO.
x
When the data for rates of production of propylene and ethylene were plotted, the line for the
ethylene rate on MO appeared above that of propylene rate, contrary to the observation on MP
and RO. An attempt was made to associate the differences in the behaviors of the catalysts with
differences in the ensembles of chemisorbed species that form the respective active centers.
ꢀ
C
2006 Wiley Periodicals, Inc. Int J Chem Kinet 38: 176–183, 2006
LPG. In addition, world propylene demands have been
increasing in recent years. Partly, these serve as motiva-
tions for attempts to find suitable catalysts for oxidative
dehydrogenation to replace or complement the nonox-
idative dehydrogenation route. The latter is energy
intensive, susceptible to catalysts deactivation, and lim-
ited by thermodynamic equilibrium. These problems
are absent in the case of oxidative dehydrogenation
(OXD) reactions. However, developing catalysts that
INTRODUCTION
Oxidative dehydrogenation of propane to propylene of-
fers prospects of using lower cost and more environ-
ment friendly route of utilizing the abundant supply of
Correspondence to: B. Y. Jibril; e-mail: baba@squ.edu.om.
2006 Wiley Periodicals, Inc.
c
ꢀ

KINETICS OF PROPANE OXYDEHYDROGENATION ON METALS OXIDES AND METALS PHOSPHATES CATALYSTS
177
exhibit high olefins yields at high selectivities proved
to be a challenging task due to the tendency of both the
alkane and olefins to produce CO and CO2, depending
on the nature of the catalyst [1,2].
results from within the activated complex [12]. This is
an ensemble of atoms/molecules, chemisorbed species,
activated reactant, and atoms or ions of the catalyst that
form the active center. This ensemble forms a “living
surface” that determines the rates of reactant conver-
sion and products’ distributions [11].
The selectivities to olefins could be improved by
understanding the reaction kinetics as demonstrated re-
cently [3–7]. The activation of the alkane is believed
to be by the rate-determining abstraction of hydrogen
atom to form alkyl species [2]. Further reactions that
determine selectivities depend partly on the position
of the hydrogen atom on the alkyl group (i.e. primary,
secondary, or tertiary carbon) and partly on the nature
of the oxygen species (lattice, surface adsorbed, gas
phase) involved in the reaction. Thus, different mech-
anisms, Mars-van Krevelen (or redox) [1], Langmuir-
Hinshelwood [8], Rideal [4], and a mechanism that in-
volves desorbed propyl radicals with gas phase oxygen
[6], have been reported. The differences in the kinetics
and mechanisms of the reaction on different catalysts
are not well understood [2,9]. In fact, a recent report
considered such differences to be contradictory [10].
More insight into the reaction route could be gained
by considering the fact that the reactants, catalysts, and
surfacespeciesformensemblesthatdeterminethereac-
tion kinetics and mechanisms and thus conversion and
products distribution [11]. The ensembles differ from
one catalyst’s surface to another and from one alkane
to another. Such differences have been offered as a pos-
sible explanation for the observation of compensation
effects in heterogeneous catalysis [12–14].
Therefore, this paper presents results of study of ki-
netics of oxidative dehydrogenation of propane and ev-
idence of compensation effects when the reaction was
tested on three different classes of catalysts. The cata-
lysts are alumina-supported metal oxides (MO) (where
metal is V, Cr, Ni, Zr, Mo, or Ba), alumina-supported
rare-earth metal oxides (RO) (where metal is Ce, Tb,
Dy, Ho, Tm, or Yb), and metal phosphates (MP) (where
metal is V, Cr, Mn, Ni, Zr, Mg, Ba, or Ce). The data for
alumina-supported vanadium and molybdenum oxides
are new, while that of others have been reported ear-
lier in terms of conversions and selectivities [15,16].
The aim of the presentation was to shed more light
on the rate-determining abstraction of hydrogen from
propanetoformpropylspeciesandthefurtherreactions
of the propyl species that lead to propylene, ethylene,
and other products, based on the nature of the cata-
lysts involved. This was done by analyzing the rates
of propane conversions and respective formations of
propylene, ethylene, CO, and CO2.
EXPERIMENTAL
The compensation effect in heterogeneous catalysis
is realized when there is a linear correlation between
ln Aapp and Eapp. These parameters are obtained from
Arrhenius plots (ln (r) vs. 1/T ) (where Aapp is the ap-
parent pre-exponential factor that may include the as-
sociated reactants partial pressures, Eapp is the appar-
ent activation energy, and r is the rate of reaction at
temperature, T ). Thus, the temperature dependence of
a surface-catalyzed reaction rate could be represented
by r = A · e(−E/RT ). When a linear relation between ln
Aapp and Eapp is rigorously established and all Arrhe-
nius plots intersect at a point, the 1/T coordinate of
the point gives the “isokinetic temperature” (Ti). Then,
isokinetic relationship is said to be established among
the different primary variables (e.g. catalysts or reac-
tants) that are varied to obtain the plots. If the existence
of Ti cannot be rigorously established, but there is a sta-
tistically significant linear relation between ln Aapp and
Eapp, then the compensation effect is established [12].
The compensation effects have been observed in het-
erogeneous catalysis [12–14]. Although, the effects are
caused by variation of primary variables (such as cata-
lysts and reactants), it is suggested that the effect is not
specifically a function of such individual variables but
Catalyst Preparations
The catalysts like alumina-supported M-oxides (where
M is V, Cr, Ni, Zr, Mo, Ba, Ce, Tb, Dy, Ho, Tm,
or Yb) were prepared using impregnation method. In
each case, a predetermined amount of metal nitrate as
a precursor was added gradually with stirring to a crys-
tallizing dish containing a predetermined amount of
ꢀ-Al2O3 (50–200 m2/g, Riedel De Haen AG) as a sup-
port to make 5 wt% active component on the support.
For instance, 38.5 g of Cr(NO3)3·9H2O was added to
100 g of γ-Al2O3 to obtain 5 wt% chromium oxide
on the support. This was continuously stirred, while
the excess deionized water was being evaporated. A
thick paste was obtained which was dried in an oven at
100◦C overnight. The catalyst was calcined under air at
600◦C for 3 h. Others were prepared in a similar man-
ner. The same support was used in all cases. The other
catalysts—M-Phosphates (where M is V, Cr, Mn, Ni,
Zr, Mg, Ba, or Ce)—were prepared using precipitation
method. The nitrates of the metals were also used as
precursors. In the subsequent discussion, the catalysts
are grouped into three groups (MO, RO, and MP). They
are alumina-supported metal oxides (where metal is V,

178
JIBRIL
Cr, Ni, Zr, Mo, or Ba), alumina-supported rare-earth
metal oxides (where metal is Ce, Tb, Dy, Ho, Tm, or
Yb), and metal phosphates (where metal is V, Cr, Mn,
Ni, Zr, Mg, Ba, or Ce). Details on catalyst preparation
are as reported elsewhere [15,16].
The catalysts’ samples were subjected to nitrogen
adsorption–desorption at 77 K, and the BET surface
areas, average pore diameters, and pore volumes of
the catalysts were measured using a commercial Mi-
crometrics 2400 apparatus. Prior to each adsorption–
desorption measurement, the samples (each of about
300 mg) were degassed at T = 473 K for 24 h. The
specific surface areas were determined using the linear
part of the BET equation.
of kinetic data were collected with negligible deactiva-
tion and the catalyst was presumably at same condition.
The performance of each catalyst is reported based
on the following: Conversion is defined as the mole
fraction of feed carbon present in the reaction prod-
ucts, while selectivity is the fraction of product carbon
in a particular product. Rate of consumption of propane
is X∗ Fin/w. The rate of production of component
i is Yi∗ Fout/w, where X is the conversion of propane,
Yi is the yield of component i, F is the flow rate
(mol/s), and w is the weight of the catalyst (g). These
rates were obtained at different temperatures. Since all
partial pressures were kept constants, the relation be-
tween rates and temperature was used to evaluate the
Arrhenius parameters.
Catalysts Testing
The catalysts were tested in a fixed bed, quartz reac-
tor, operated at atmospheric pressure, at the feed flow
rate of 75 cm3/min (26.7% propane, 6.6% oxygen, and
the balance helium), at the temperature range of 350–
450◦C (MO and RO) and 450–550◦C (MP). The resi-
dence time was maintained at about 0.60 s. The reactor
was a 70-mm long quartz tube of 7 mm internal diame-
ter (id) tapered to a 2-mm id. This removes the reaction
gases from the reaction zone as fast as possible, thereby
minimizing possible gas phase reactions. The catalysts
were placed in the reaction zone of the reactor and
supported on quartz wool directly above the junction
of this section with the bottom. The catalyst particle
size range was chosen based on a separate experiment
to minimize mass transfer limitations in the reactor.
In each case, 1 g of the catalyst was pretreated in a
stream of oxygen for 30–45 min at 450◦C. The line
was then evacuated with helium at the same temper-
ature for 30 min. Thereafter, the reactant gases were
flowed through the reactor at the desired temperatures.
A gas chromatograph (HP6890) was used for an
online analysis of both the feed and products streams.
The products flowed directly through a heat-traced line
to the GC-sampling valve. The hydrocarbons and aro-
matics were separated by HP-PLOT column and ana-
lyzed with flame ionization detector (FID), while O2,
CO, and CO2 were separated by MS and HayeSep-R
column and analyzed with thermal conductivity detec-
tor (TCD). All GC analyses were performed online by
software—HP Chemstation—provided with the equip-
ment. A blank test without a catalyst or with quartz
granules in the reactor showed negligible conversion
of propane at the reaction conditions. Carbon balances
were typically better than 95%. For some experiments,
triple runs were performed to ensure reproducibility.
The variations among the runs were about 6%. In that
case, results reported are the average values. Thus, sets
RESULTS AND DISCUSSION
The oxidative dehydrogenation of propane was tested
on the catalysts. BET surface areas, pore volume, and
average pore diameter of a sample of the catalysts
are given in Table I. The rates of propane consump-
tions, propylene, ethylene, CO, and CO2 productions
at different reaction temperatures were evaluated. The
Arrheniusplot[ln(rate)vs. 1/T ]forpropaneconsump-
tion on the three groups of catalysts is displayed in
Figs. 1–3. Tables II and III show the linear regression
Table I BET Surface Areas, Pore Volume, and Average
Pore Diameter of a Sample of the Catalysts
BET
Pore
Surface
Area (m2/g)
Volume
(cm3/g)
Average Pore
˚
Sample
Diameter (A)
V-P-O
Cr-P-O
Mn-P-O
Ni-P-O
Zr-P-O
0.4
4.1
5.6
1.5
47.3
49.0
0.2
2.1
163.5
183.1
161.3
108.1
152.6
123.4
140.1
146.0
131.4
147.7
118.0
127.3
0.004
0.026
0.035
0.010
0.132
0.500
0.018
0.022
0.269
0.302
0.235
0.180
0.236
0.229
0.230
0.249
0.196
0.234
0.221
0.224
558.5
287.7
264.7
109.0
56.0
358.6
575.2
431.5
66.4
66.0
60.4
66.6
63.6
74.0
66.6
Ce-P-O
Mg-P-O
Ba-P-O
V-Al-O
Cr-Al-O
Ni-Al-O
Zr-Al-O
Mo-Al-O
Ba-Al-O
Ce-Al-O
Tb-Al-O
Dy-Al-O
Ho-Al-O
Tm-Al-O
Yb-Al-O
68.0
59.7
63.4
75.0
70.8

KINETICS OF PROPANE OXYDEHYDROGENATION ON METALS OXIDES AND METALS PHOSPHATES CATALYSTS
179
parameters for the analyses. The plots do not
exhibit unique intersection points. Thus, indicating ab-
sence of isokinetic temperature (Ti), within the range
explored. However, RO and MP show “factional” isoki-
netic points, common to some catalysts. The plots of
Ce, Tb, and Tm (Fig. 2) and Mg, Ba, Ce, and Ni (Fig. 3)
are intersected. When all the plots were combined on
one graph, MO and MP appeared together in the same
ln Aapp range, while RO appeared uniquely at a differ-
ent ln Aapp range. MO and MP together showed seven
catalysts that appeared to have a common isokinetic
temperature. However, the fact that there are no unique
differences among those that intersected and those that
did not make it difficult to assign fundamental signifi-
cance to such isokinetic temperature. For instance, the
less redox character of Mg, Ba, and Ce, which forms
the common Ti with Ni, may not be sufficient to lead
to a categorical conclusion.
Figure 1 Arrhenius plots for the rates of propane conver-
sion in oxidative dehydrogenation of propane on different
alumina-supported metal oxides catalysts.
Perhaps the isokinetic rate (or isokinetic temper-
ature) arises from ensembles that are energetically
similar and therefore statistically correlate with one
another as a primary variable (such as catalyst or re-
actant partial pressure) is changed. If there are strong
correlations among the ensembles, they may exhibit a
common isokinetic temperature. Otherwise, moderate
correlation may lead to linear ln Aapp vs. Eapp relation
only. Where there is segregation into groups of ensem-
bles with strong and moderate correlations, separate
compensation lines may be observed. This could be
a reason of multiple compensation lines in Fig. 4. As
could be observed from Figs. 1 and 3, MO and MP
appear on similar range of ln (rate), which is different
from that of RO. This perhaps reflects the similarity of
the ensembles involved in the conversion of propane
on the two groups of catalysts. Generally, MO and MP
involved reaction of propane with lattice oxygen [1,2].
In the case of RO, surface-adsorbed oxygen might be
dominant in abstracting hydrogen from propane [17].
Changing the catalysts in each group (RO as well as
MO and MP) may bring about different nature and de-
gree of change to the surface and lattice oxygen species.
This may lead to clear separate compensation line for
the propane conversion on RO.
Figure 2 Arrhenius plots for the rates of propane conver-
sion in oxidative dehydrogenation of propane on different
alumina-supported rare-earth metals oxides catalysts.
Oxidative dehydrogenations of lower alkanes have
been severally reported to take place via the Mars Van
Krevelen (MVK) or redox mechanism on oxide cata-
lysts [1,2]. It is assumed that the rate-determining ab-
straction of hydrogen from alkane to alkyl species in-
volves reaction with lattice oxygen, thereby reducing
the catalyst. Further reaction of the alkyl species to pro-
duce alkene also involves lattice oxygen. The oxygen
from the gas phase reoxidizes the catalyst. The prod-
ucts’distributions, especiallytheamountofCOx and/or
otheroxygen-containingcompoundsproduced, depend
Figure3 Arrheniusplotsfortheratesofpropaneconversion
in oxidative dehydrogenation of propane on different metals
pyrophosphates catalysts.

180
JIBRIL
Table II Regression Coefficients of ln (rate) vs. 1/T for Different Rates on Metal Phosphates (MP) at 450–550◦C,
Alumina-Supported Metals (MO) and Alumina-Supported Rare-Earth Metals (RO) at 350–450◦C
Regression Coefficient for Different Rates
Data
Catalysts
Points
C3H8
C3H6
C2H4
CO2
CO
MO
Cr
Zr
Ni
Ba
Mo
V
5
5
2
5
5
5
0.840
0.980
–
0.970
0.890
0.910
0.842
0.992
–
0.998
0.917
0.95
0.988
0.991
–
0.995
0.797
0.963
0.795
0.991
–
0.9499
0.913
0.875
–
0.985
–
0.914
0.748
0.911
RO
Tm
Dy
Yb
Ce
Tb
Ho
5
5
5
5
5
5
0.991
0.969
0.986
0.916
0.944
0.973
0.991
0.869
0.995
0.926
0.936
0.978
0.994
0.978
0.993
0.978
0.989
0.991
0.977
0.803
0.959
0.894
0.86
0.973
0.768
0.963
0.864
0.788
0.688
0.732
MP
V
Zr
4
5
5
5
3
3
3
5
0.989
0.956
0.976
0.999
0.999
0.993
0.928
0.997
0.969
0.958
0.978
0.995
0.995
0.99
0.994
0.997
0.995
0.999
0.999
0.993
0.993
0.997
0.992
0.806
0.583
0.583
0.987
1.000
0.952
0.996
0.993
0.997
0.983
0.983
0.957
1.000
0.74
Cr
Mg
Mn
Ni
Ba
Ce
0.981
0.998
0.988
Table III Regression Parameters of ln (Aapp) vs. Eapp for Alumina-Supported Metals (MO), Alumina-Supported
Rare-Earth Metals (RO), and Metal Phosphates (MP)
Catalysts
C3H8
C3H6
C2H4
CO2
CO
MO
RO
MP
R
0.999
0.327
0.997
0.595
0.995
0.445
0.999
0.311
0.995
0.641
Std. Error
Intercept
Slope
−13.2 0.18
−13.8 0.39
−17.2 0.91
−14.7 0.16
−14.2 0.31
0.686 0.01
0.669 0.03
0.745 0.04
0.695 0.02
0.708 0.04
R
0.998
0.164
0.999
0.087
0.987
0.853
0.999
0.134
14.8 0.12
0.741 0.02
0.998
0.147
Std. Error
Intercept
Slope
−8.6 0.20
−13.9 0.08
−15.6 1.59
−13.94 0.12
0.687 0.02
0.647 0.06
0.709 0.06
0.699 0.02
R
0.999
0.243
0.999
0.252
0.999
0.379
0.998
1.012
0.999
0.703
Std Error
Intercept
Slope
−12.8 0.21
−13.77 0.22
−13.84 0.61
−15.91 0.45
−15.38 0.36
0.624 0.01
0.628 0.01
0.620 0.01
0.611 0.02
0.622 0.01
on the redox character of the catalysts. The OXD of
lower alkanes over the most widely reported V-based
catalysts such as V Mg O produce mainly dehydro-
genation and combustion products, conforming to the
assumption of redox mechanism [2]. Molybdenum-
based systems in addition to alkenes produce acrolein.
The latter has been produced via oxidation of propylene
[18].
However, based on the earlier studies [4,8], the as-
sumption of redox cycle may not hold in some cases

KINETICS OF PROPANE OXYDEHYDROGENATION ON METALS OXIDES AND METALS PHOSPHATES CATALYSTS
181
of the interaction) may not easily furnish the required
electrons, and the oxygen incorporation may be slower.
This rather slow incorporation increases the population
of the oxygen species on the surface of nonredox metal
oxides such as RO. This behavior arises from chemical
nature of the rare-earth atoms. Their electronic config-
uration in the outmost 6s and 5d orbitals are similar.
Added electrons begin to enter the fairly deep 4f or-
bitals. Due to this, the 4f are screened from outside by
outer electrons in the 6s and 5d. So, the addition of
valence electron brings about no significant change in
their properties and limit interaction with outside. This
may partly be the source of the differences in the com-
pensation line shown by RO compared with MO and
MP.
Figure 5 shows the compensation plot for the rates
of production of propylene and ethylene on MO, RO,
and MP. Not all points lie on a straight line. It was ob-
served that the catalysts based on nonredox (Ba, Mg) or
less redox (Ce) metals tended to exhibit the highest ap-
parent activation energies compared with other metals
among the MO and MP groups. Linear regression anal-
yses of the data points show that the points for the rate
of production of ethylene on MO somewhat appeared
on a different line from the principal compensation line
(inset). Perhaps this is in accordance with lower pro-
duction of ethylene on MO. After the rate-determining
abstraction of hydrogen atom, the resulting ensemble
involving MO has less susceptibility toward break-
ing the propyl species or propylene to produce ethy-
lene. Since the surface adsorbed electrophilic oxygen
species is associated with electrophilic attack on propy-
lene/propyl species, the observation indicates low pop-
ulation of such oxygen on MO [19]. Alternatively,
Figure 4 Compensation effect plots for the conversion of
propane in oxidative dehydrogenation of propane on different
metals oxides and metals pyrophosphate catalysts.
as pointed out in the Introduction. The rate of lattice
oxygen replenishment may be slower than the rate of
removal. This may lead to a deviation from simple re-
dox mechanism as observed when propane OXD was
testedonrare-earth-oxide-basedcatalysts. Althoughno
oxygenates other than COx are observed, there is sig-
nificant formation of ethylene, similar to an earlier re-
port [6]. Surface adsorbed electrophilic oxygen may be
involved in the reaction on RO as against nucleophilic
latticeoxygeninthecaseofMOandMP[17]. Whenthe
lattice oxygen is involved in the reaction, there could
be increase in the surface reaction in the lattice. Since
the reaction is exothermic, the local heat release may
not be dissipated uniformly. The nonuniformities in the
surface temperature in turn result into local changes in
the surface-adsorbed species. This causes change in the
surface diffusion [11]. The diffusion is limited where
lattice oxygen is not involved in the reaction, as in the
caseofRO. Suchlimiteddiffusionleadstoalargepopu-
lation of surface adsorbed electrophilic oxygen species
that attack the electro-rich center on propylene to pro-
duce larger amount of ethylene.
Furthermore, one energetically possible path for
incorporation of lattice oxygen in the redox MO is
considered to be by homolytic dissociation of surface
adsorbed O2 to form an O•. The O• may then be incor-
porated simultaneously via two-electron transfer from
two redox metals. This is demonstrated by Scheme 1
[11]. The less redox metals (depending on the nature
•
ꢀ
O
2 ꢁ
2O
Figure 5 Compensation effect plots for the production
of propylene and ethylene in oxidative dehydrogenation of
propane on different metals oxides and metals pyrophosphate
catalysts.
O• + 2M(n − 1)+ → Mn+O2−Mn+
Scheme 1

182
JIBRIL
different ensembles might be involved in the respec-
tive routes to ethylene and propylene.
On the other hand, RO has been associated with high
population of the electrophilic oxygen species that ren-
ders the catalysts less selective to propylene or at least
high rate of production of ethylene in OXD of propane
[17]. This could be due to their lower ability to per-
form redox reaction and less tendency to replenish the
lattice oxygen as discussed. This brings about lower
oxygen mobility and lower diffusion in the catalysts. It
has been observed that the reaction kinetics and mecha-
nism on oxide catalysts could be strongly influenced by
the chemisorption of the reactant themselves and by the
population and nature of the adsorbed species. For RO
(at 350–450◦C), a fast surface hopping movement of
different electrophilic oxygen adsorbed species is pos-
sible [11]. Thus, increasing the tendency toward elec-
trophilic attack on electron-rich propylene and propyl
species thereby exhibits high selectivity to ethylene
and COx .
Figure 6 Compensation effect plots for the production of
carbon dioxide and carbon monoxide in oxidative dehydro-
genation of propane on different metals oxides and metals
pyrophosphate catalysts.
duction of CO by MO and RO than MP. The low CO on
MP is in accordance with earlier report that has shown
MP to desorb no surface oxygen up to 550◦C [20].
These differences are further demonstrated in Table IV
that shows the conversions and selectivities to products.
The MO exhibited higher selectivities to olefins than
RO, but lower than MP at comparable conversions. The
selectivities on MO range from 33.9–61.0% while on
The compensation plots for the rate of production
of CO and CO2 are shown in Fig. 6. The rates data for
MP lie on the principal compensation line. There is an
indication of higher tendency to CO on RO (inset). The
points for rate of production of CO from MO and RO
appear to cluster on a common line but separate from
thatofMP. Thisindicatesahighertendencytowardpro-
Table IV Conversions and Products Distributions for Different Catalysts on Metal Phosphates (MP) at 450◦C,
Alumina-Supported Metals (MO), and Alumina-Supported Rare-Earth Metals (RO) at 350◦C
Conversion (%)
C3H8
Selectivity (%)
C2H4
Catalysts
C3H6
MO
CO2
CO
Cr
Zr
Mo
V
8.5
1.0
6.9
12.5
3.5
41.4
35.9
61.0
57.7
3.7
1.0
0.0
0.0
0.4
2.0
0.4
29.0
18.5
15.0
52.0
41.8
35.1
20.5
26.6
42.3
Ba
RO
Tm
Dy
Yb
Ce
Tb
Ho
3.0
6.7
6.0
2.5
4.7
8.0
7.6
35.0
5.2
21.9
21.1
31.4
2.3
1.4
3.6
0.0
1.9
2.8
47.0
25.6
44.0
47.7
34.9
31.1
43.2
38.0
46.3
30.4
41.3
33.7
MP
V
Zr
2.8
9.7
5.5
0.5
4.3
0.2
3.0
72.4
56.8
30.6
70.6
41.0
82.3
26.6
2.5
2.5
7.0
25.7
34.2
0.0
13.2
0.0
18.2
14.4
20.1
0.0
28.1
0.0
Cr
Mg
Mn
Ni
Ba
13.1
21.4
15.2
17.7
6.8
48.3
17.1

KINETICS OF PROPANE OXYDEHYDROGENATION ON METALS OXIDES AND METALS PHOSPHATES CATALYSTS
183
RO the values are 8.8–36.4%. MP showed the highest
selectivities range of 33.4–100%. For the three groups
of catalysts, the average values of selectivities to COx
are 77.2% (RO), 50.8% (MO), and 32.3% (MP). In ad-
dition, the MO and MP show less tendencies to produce
ethylene per unit propylene than RO. As discussed ear-
lier, this is associated with the surface adsorbed elec-
trophilic oxygen species which is lowest on MP and
hence the lowest selectivity to COx . More work is be-
ing conducted at our laboratory to give further insight
into the relation between compensation effect and the
catalysts’ performance.
BIBLIOGRAPHY
1. Mamedov, E. A.;CortesCorberan, V. ApplCatalA1995,
127, 1.
2. Kung, H. H.; Kung, M. C. Appl Catal A 1997, 157, 105.
3. Le Bars, J.; Vedrine, J. C.; Auroux, A.; Trautmann, S.;
Baerns, M. Appl Catal A 1992, 88, 179.
4. Andersson, S. L. T. Appl Catal A 1994, 112, 209.
5. Albonetti, S.; Blanchard, G.; Burattin, P.; Cavani, F.;
Masetti, S.; Trifiro, F. Catal Today 1998, 42, 283.
6. Buyevskaya, O. V.; Baerns, M. Catal Today 1998, 42,
315.
7. Resini, C.; Panizza, M.; Arrighi, L.; Sechi, S.; Busca,
G.; Miglio, R.; Rossini, S. Chem Eng J 2002, 89, 75.
8. Huff, M.; Schmidt, L. D. J Catal 1994, 149, 127.
9. Michaels, J. N.; Stern, D. L.; Grasselli, R. K. Catal Lett
1996, 42, 139.
CONCLUSION
10. Baldi, M.; Finocchio, E.; Pistarino, C.; Busca, G. Appl
Catal A 1998, 173, 61.
11. Centi, G.; Cavani, F.; Trifiro, F. Selective Oxidation
by Heterogeneous Catalysis; Kluwer Academic/Plenum
Publishers: London, 2001; p. 392.
12. Bond, G. C.; Keane, M. A.; Kral, H.; Lercher, J. A.; Catal
Rev – Sci Eng 2000, 42, 323.
13. Bautista, F. M.; Campelo, J. M.; Garcia, A.; Luna, D.;
Marinas, J. M. J Catal 1989, 116, 338.
14. Cheung, T. K.; Jentoft, F. C.; d’Itri, J. L.; Gates, B. C.
Chem Eng Sci 1997, 52, 4607.
15. Jibril, B. Y.; Al-Zahrani, S. M.; Abasaeed, A. E. Catal
Lett 2001, 74, 145.
16. Jibril, B. Y.; Al-Zahrani, S. M.; Abasaeed, A. E. Catal
Lett 2003, 85, 57.
17. Zang, W.; Zhou, X.; Tang, D.; Wan, H.; Tsai, K. Catal
Lett 1994, 23, 103.
The propane oxidative dehydrogenation has been
tested on three groups of catalysts—alumina-supported
metal-oxides, alumina-supported rare-earth oxides,
and metal phosphates. The catalysts exhibited differ-
ent degrees of conversions and products distributions.
The rates of conversion of propane, rates of produc-
tion of propylene, ethylene, CO and CO2 were sta-
tistically analyzed; and the values of Arrhenius pa-
rameters, apparent pre-exponential factors (Aapp) and
apparent activation energy (Eapp), were obtained. Plot
of ln Aapp vs. Eapp formed statistically significant lin-
ear relationships, indicating evidence of compensation
effects. An analysis of the data for the rates of propane
conversions and production of COx indicated that sim-
ilar group of species might be involved in the case of
MO and MP. While production of propylene and ethy-
lene showed similarity on RO and MP. Thus, the com-
pensation effect could be used as a simple criterion
for gaining insight into the kinetics of the reaction.
However, more work is needed for a more categorical
conclusion.
18. Stern, D. L.; Grasselli, R. K. J Catal 1997, 167, 560.
19. Pantazidis, C.; Bucholz, A.; Zanthoff, S. A.; Schuurman,
H. W.; Mirodatos, Y. A. Catal Today 1998, 40, 207.
20. Takita, Y.; Sano, K.; Kurosaki, K.; Kawata, N.;
Nishiguchi, H.; Ito, M.; Ishihara, T. Appl Catal A 1998,
167, 49.