Journal of Natural Gas Chemistry 11(2002)70-78
Kinetics of the Oxidative Dehydrogenation of Isobutane
over Cr2 O3/La2 (CO3 )3
Yanping Sun,
Tracey A. Robson,
Trevor C. Brown∗
School of Biological, Biomedical & Molecular Sciences, University of New England, Armidale, NSW, 2351, Australia
[Manuscript received July 23, 2002; revised August 23, 2002]
Abstract: The oxidative dehydrogenation (ODH) of isobutane over Cr2 O3 /La2 (CO3 )3 has been investigated in a low-pressure Knudsen cell reactor, under conditions where the kinetics of the primary reaction
steps can be accurately determined. By heating the catalyst at a constant rate from 150–300 , temperature fluctuations due to non-equilibrium adsorption are minimized. The evolved gas profiles show that
ODH to isobutene and water is a primary reaction pathway, while carbon dioxide, which forms from the
catalyst during reaction, is the only other product. This CO2 evolution may enhance the activity of the
catalyst. Isobutene formation proceeds with the participation of lattice oxygen from the Cr 2 O3 /La2 (CO3 )3
catalyst. The intrinsic Arrhenius rate constant for the ODH of isobutane is
k(s−1 ) = 1011.5±2.2 exp{−((55 ± 5) − 4Hads kJmol−1 )/RT }
The small pre-exponential factor is expected for a concerted mechanism and for such a catalyst with a
small surface area and limited porosity.
Key words: catalytic kinetics, oxidative dehydrogenation, molecular flow, isobutane, isobutene,
Cr2 O3 /La2 (CO3 )3
(CH3 )3 CH(g) + O(catalyst)→
(CH3 )2 C=CH2 (g) + H2 O(g)
1. Introduction
There is a high and growing market demand for
isobutene, particularly because isobutene is a key reactant for the production of methacrylates and MTBE
(methyl-tert-butylether). Isobutene is typically produced in industry by the endothermic dehydrogenation of isobutane over a Cr2 O3 /Al2 O3 catalyst at ca.
900 K and a contact time of 1 s [1]. Disadvantages of
this process are that high selectivity is only achievable
at low conversions and the catalyst is readily deactivated by coke deposition.
Oxidative dehydrogenation (ODH) is, in principle, a more efficient and less energy intensive alternative to direct dehydrogenation. The overall reaction
for the ODH of isobutane is
∗
This is an exothermic process, which with an overall
positive entropy change leads to a negative Gibbs free
energy and hence, unlike dehydrogenation, is spontaneous at low temperatures. Coke deposition is generally not observed. Disadvantages of this process are
that CO2 and CO are also thermodynamically favored
products and low temperatures can result in low reaction rates. The ideal catalyst and conditions for ODH
would inhibit complete oxidation, but would allow the
reaction to proceed at a high rate.
Supported chromium oxides are known to selectively catalyze the oxidation of hydrocarbons [2-5].
Grabowski et al. [4] observed that, for the ODH
of isobutane, chromia supported on SiO2 , Al2 O3 ,
Corresponding author. Tel: +61 2 6773 2872; Fax: +61 2 6773 3268;
E-mail; [email protected].
2
Yanping Sun et al./ Journal of Natural Gas Chemistry Vol. 11 No. 1 2002
TiO2 , ZrO2 and MgO are active at relatively low
temperatures (200–400 ). The highest selectivities
(ca. 73% at 5% conversion) and highest yields (ca.
9%) are obtained for CrOx /TiO2 and a K-promoted
CrOx /Al2 O3 preparation. Moriceau et al. [5] reported
that Cr–Ce–O catalysts are active and selective (ca.
60% at 10% conversion) for the ODH of isobutane at
low temperature (270 ), and a maximum isobutene
yield of 8% could be achieved.
Hoang et al. [2] found that Cr2 O3 /La2 (CO3 )3 exhibits very high activity and selectivity (ca. 95% at
12.5% conversion) to the ODH of isobutane at low
temperatures (230–250 ), and CO2 is the only observed by-product. Therefore this catalyst is an ideal
candidate for more detailed kinetic studies.
Most heterogeneous catalysis reaction mechanisms have been investigated under high-pressure conditions where secondary reactions proliferate. As
a consequence observed products may not represent
the initial intrinsic reaction but result from a series of bimolecular steps. In this paper, the kinetics of the initial steps in the ODH of isobutane over
Cr2 O3 /La2 (CO3 )3 have been investigated in a Knudsen cell reactor [6-7]. Under low-pressure conditions,
the first steps in a heterogeneous reaction can generally be identified because gas-phase collisions are
effectively eliminated and surface bimolecular reactions are minimized. Furthermore, this technique allows for the determination of accurate rate parameters as the gas/surface collision frequency may be
precisely calculated from the molecular flow rate, the
geometric surface area of the catalyst and exit aperture area [6]. Also, by including temperature programming, temperature-fluctuations, resulting from
non-equilibrium adsorption on clean surfaces are minimized and a wide range of temperatures can be investigated over a short period of time, without an
internal standard.
2. Experiment
The low-pressure reaction system has been described in detail elsewhere [6-7]. Briefly, isobutane passes through a variable leak valve (GranvillePhillips), through 240 mm of 1mm ID capillary glass
tubing and into the Pyrex Knudsen cell containing
the catalyst. Reactant molecules then experience collisions with the reactor walls and catalyst surface,
prior to escaping through the exit aperture and into
a quadrupole mass spectrometer. Flow rate is monitored before each temperature-programmed experiment by recording the pressure decrease from the
pressure gauge situated, in a known volume, next to
the variable leak valve. The evolved-gas detection system is a Hewlett Packard 5995 quadrupole mass spectrometer. Details of the two Knudsen cells used in the
analysis are listed in Table 1. Collision numbers (Z)
are calculated from the ratio of the geometric surface
area of the catalyst (Scat ) to the area of the exit aperture (Sea ). The geometric surface area of the catalyst
[8,9] is used as an estimate for the exposed catalyst
surface area and may introduce errors of up to a factor of ten in the reported Arrhenius pre-exponential
factors. A value for Scat of 480 mm2 is used in all calculations. Measured exit aperture areas are corrected
by Clausing factors [10] to take into account the finite
width of the exit aperture.
Table 1. Key physical parameters for the two
Knudsen cell reactors used in this study
Cell parameters
Reactor 1
Reactor 2
Volume (mm3 )
8,700 ± 5
2,114 ± 5
Internal surface area (mm2 ) 2,570 ± 2
942 ± 2
Exit aperture area (mm2 )
0.16 ± 0.03
0.37 ± 0.03
Clausing factor
0.20 ± 0.03
0.28 ± 0.03
Collision number
17,100 ± 3,000 3,010 ± 200
Residence time (s)
1.5(M/T)1/2
0.16(M/T)1/2
Cr2 O3 /La2 (CO3 )3 was prepared as described by
Hoang et al.
[2].
A 1:8 mole ratio solution
of La2 (NO3 )3 ·6H2 O (CR, 99.999%, Aldrich Chemical Company, Inc.)
and Cr(NO3 )3 ·9H2 O (TR,
99%, Aldrich Chemical Company, Inc.) was dropwise added to a stirred aqueous solution of 24 g/L
NH4 HCO3 (AR, AJAX Chemical Ltd.). The resultant hydrogel was centrifuged, filtered and washed
with water to remove unreacted nitrates. Organic
residues were removed by washing with acetone with
two more centrifuging and filtering steps. The resulting solid catalyst was dried at room temperature under nitrogen for 2 h and then at 110
in air for
4 h. The final catalyst was obtained by calcination
at 300
in air for 2 h and then pressed binder-free
and crushed to particles of 0.15–0.23 mesh size. The
N2 BET surface area is 20 ± 10 m2 /g, while XRD
results indicate that the Cr2 O3 /La2 (CO3 )3 catalyst
3
Journal of Natural Gas Chemistry Vol. 11 No. 1 2002
is amorphous. Room temperature FT-IR spectra of
Cr2 O3 /La2 (CO3 )3 were recorded on a Perkin Elmer
1725X FT-IR spectrometer. An intense broad band
is apparent at 917 cm−1 and a sharp band at 1,079
cm−1 . This is in agreement with the reported FT-IR
spectrum on the same catalyst synthesized by Hoang
et al. [2].
3. Results
3.1. Temperature programmed reaction profiles
Temperature-programmed combined product and
reactant profiles were recorded for the ODH of isobutane over Cr2 O3 /La2 (CO3 )3 . Molecular flow rates
into the two reactors ranged from 1.74×1014 to
2.54×1014 molecule/s, and the heating rate was varied
from 2 to 10 /min.
A careful analysis of the mass spectral profiles between 0 and 4% conversion indicates that the only
evolved gases are isobutene, carbon dioxide and water.
Other possible products, CO, acetone, propene and
methane or organic acids and aldehydes are not apparent in the mass spectra. The rate of water evolution
Table 2.
Mass spectrometer calibration factors and conversion factors (α)
for isobutane, isobutene and carbon dioxide
Isobutane(I58 )
Isobutene(I56 )
Carbon Dioxide(I44 )
I56 /I58
0.1294 ± 0.0028
—
—
I44 /I58
0.8958 ± 0.0028
—
—
α (abundance·s/mole)
(1.4 ± 0.2) × 1014
(1.8 ± 0.6) × 1015
(1.6 ± 0.8) × 1015
Typical temperature-programmed reaction profiles for isobutane decay, isobutene formation and carbon dioxide evolution are plotted in Figure 1, where
the heating rate is 5 /min and the initial flow rate
into Reactor 1 is 1.74 × 1014 molecule/s. Figure 1
shows that the abundance of isobutane slowly declines
from 220 to 285 , while the corrected abundances
of isobutene and carbon dioxide increase. Above
285 , the abundance of isobutane is effectively constant, while the abundances of isobutene and CO2 decrease. This adjustment in the high-temperature rate
of change in evolution rates with increasing temperature is most likely due to saturation of active sites,
limitations in the transfer of reactant to active sites
or poisoning by product intermediates. Fractional
could not be accurately determined as irreproducible
amounts of water desorbs from the catalyst during
temperature programming; these mass spectral profiles are influenced by factors such as pre-conditioning
of the reactor and the mass spectrometer. Hence, the
kinetics of water evolution was not determined. The
reactant and two products are characterized by m/e =
58 (isobutane), m/e = 56 (isobutene), m/e = 44 (carbon dioxide). Isobutane contributes to the characteristic product peaks, m/e = 56 and m/e = 44, while
isobutene does not affect the abundance of m/e = 44.
For isobutane, the fractional contributions relative to
m/e = 58 and m/e = 44, as well as mass-spectral
conversion factors (α (abundance·s/mole)) are measured at the commencement of every temperatureprogrammed experiment. For the products, separate
experiments using each product gas were carried out
over a range of flow rates to determine conversion factors. Table 2 lists fractional abundances and conversion factors for the respective peaks. The rate
of CO2 formation during temperature-programmed
ODH is determined by deducting contributions from
CO2 , which evolve from the catalyst under identical
conditions except that no isobutane is present.
isobutane decomposition and product yields were calculated from
fDecomposition =
fYield =
ICo 4 H10 − IC4 H10
ICo 4 H10
αC4 H10 IProduct
αProduct ICo 4 H10
(1)
(2)
Due to low yields fractional data, particularly for
isobutane decay, has a large signal-to noise ratio. As a
consequence Figure 2 shows only modeled percentage
isobutane decay, but both raw and modeled product
formation data are plotted. The model is simply a
quadratic fit to the data from equations (1) and (2).
The raw data, however, were used in the kinetic calculations. Figure 2 shows the low yield of isobutene,
4
Yanping Sun et al./ Journal of Natural Gas Chemistry Vol. 11 No. 1 2002
and indicates that CO2 is the major evolved species;
formation rate is greater than isobutane decomposition.
Figure 1. Corrected abundances for isobutane (m/e
= 58), isobutene (m/e = 56) and carbon
dioxide (m/e = 44) during the temperature programmed ODH of isobutane over
Cr2 O3 /La2 (CO3 )3 in Reactor 1. Isobutane flow rate was 1.74 × 1014 molecule/s
and heating rate 5
/min.
3.2. Mechanism for the formation of isobutene
The rate-determining step for the ODH of hydrocarbons is, for most systems, the breaking of a C–H
bond by surface oxygen to form an alkyl species [1113]. At low-pressure and in the absence of gaseous
oxygen, the rate-controlling mechanism for the ODH
of isobutane is
Here Sunactivated and Sactivated refer to non-catalytic
and catalytic sites on the Cr2 O3 /La2 (CO3 )3 surface.
In the kinetic analysis, diffusion to and from active
sites (R2) is initially assumed to be rapid and not
rate-determining. As a consequence, the key reaction
steps are reversible adsorption (R1) and C–H bond
breaking (R3). The rate constants k−1 and k3 have
units 1/(s·site), while A1 PC4 H10 , where P is pressure,
is the isobutane/catalyst surface collision frequency.
The abundance of a mass spectral peak (IProduct),
which is uniquely characteristic of a primary product,
is proportional to the mass-spectral calibration factor
and the product flow rate [14].
dnProduct
IProduct
=
dt
αProduct
(3)
The product flow rate may be calculated from the collision number Z, the average reactant flow rate and
the ratio of first-order rate constants determined by
assuming a steady-state coverage of isobutane:
dnProduct
k3
dnC4 H10 −E1 /RT
=
Z
e
dt
k−1 + k3
dt
(4)
Here the product of Z, the average reactant flow rate
and exp(-E1 /RT) is equal to the effective collision frequency on the catalyst surface or k1 PC4 H10 , where
PC4 H10 is the isobutane pressure. As a first approximation equation (4) can be further simplified by noting that under most conditions k−1 k3 . Collision
numbers are listed in Table 1, while the reactant flow
rate is calculated from the mean of mass spectral
abundances entering and leaving the Knudsen cell.
Figure 2. Percentage product yields (isobutene and
carbon dioxide) and reactant decomposition (isobutane) during the temperature programmed ODH of isobutane over
Cr2 O3 /La2 (CO3 )3 in Reactor 1. Isobutane flow rate was 1.74 × 1014 molecule/s
/min.
and heating rate 5
k3 (ICo 4 H10 + IC4 H10 ) −E1 /RT
IProduct
=
Z
e
αProduct
k−1
2αC4 H10
(5)
ICo 4 H10 and IC4 H10 are abundances for isobutane before
and after reaction respectively, and αC4 H10 is the mass
spectral conversion factor (abundance·s/molecule).
5
Journal of Natural Gas Chemistry Vol. 11 No. 1 2002
The activation energy for adsorption (E1 ) of isobutane onto the Cr2 O3 /La2 (CO3 )3 is included in equations (4) and (5). Hence, the full rate expression with
Arrhenius parameters is
IProduct
=
(6)
αProduct
Io
+ IC4 H10
−E1 + E−1 − E3
A3
Z C4 H10
exp
A−1
2αC4 H10
RT
The unknowns in this expression are the preexponential factors and activation energies, and are
determined by taking natural logarithms of equation
(6) to give
o
(IC4 H10 + IC4 H10 ) αProduct
Z =
−ln
2αC4 H10
IProduct
A3
−E1 + E−1 − E3
(7)
ln
+
A−1
RT
pre-exponential factor for reaction (R3), can be calculated by estimating the isobutane desorption preexponential factor, A−1 . Pitt et al. [15] has calculated that for physisorbed systems, these factors
should be in the range 1011 –1013 s−1 , with mean
value A−1 = 1012.0±1.0 s−1 . The intrinsic activation energy for the ODH of isobutane can be calculated by adding the enthalpy of adsorption to the
apparent activation energy. However, no adsorption
enthalpies for isobutane on Cr2 O3 /La2 (CO3 )3 have
been reported in the literature. Table 3 lists apparent
and intrinsic Arrhenius pre-exponential factors and
apparent activation energies with corresponding standard deviations determined from three temperatureprogrammed experiments on the ODH of isobutane
over Cr2 O3 /La2 (CO3 )3 .
Hence, a plot of the left-hand side of equation
(7), which are all measurable quantities, against
1/RT produces a straight line with an intercept at
ln(A−1 /A3 ) and a slope of E3 +E1 -E−1 . The slope is
the intrinsic activation energy for reaction plus the
enthalpy of adsorption (E3 +4Hads ).
3.3. Model results
As the rise in isobutane decomposition and
isobutene formation decreased beyond 285
(see
Figure 1), due to poisoning, mass transfer limitations or saturation of active sites, this data was discarded for kinetic calculations. The Arrhenius plot
described by equation (7) is shown in Figure 3 for
the temperature-programmed experiment performed
in Reactor 1 with heating rate of 5 /min and initial
isobutane flow rate of 1.74 × 1014 molecule/s. The
Table 3.
Figure 3. Arrhenius-type plot of ln (k −1 /k 3 )
against 1/RT for the ODH of isobutane
Kinetic parameters obtained for ODH of isobutane
over Cr2 O3 /La2 (CO3 )3
Experiment No.
1
2
3
E3 +∆Hads (kJ/mol)
58 ± 4
54.2 ± 1.5
53.7 ± 2.2
55 ± 5
logA−1 /A3
A3 (s−1 )
0.1 ± 0.4
1011.9±1.4
0.52 ± 0.15
1011.5±1.2
0.79 ± 0.22
1011.2±1.2
0.5 ± 0.6
1011.5±2.2
4. Discussion
4.1. Infrared Spectroscopy of Cr2 O3 /La2 (CO3 )3
There have been a number of infrared spectroscopy studies of the formation of surface chromate
[16,17]. For supported chromium oxide on alumina,
Average
zirconia, titania, and silica IR spectra are characterized by a high-frequency group consisting of bands
at 1,000–1,050 cm−1 assigned to isolated or nonisolated Cr(V)=O species and a lower frequency group
consisting of bands in the 850–950 cm−1 region attributable to Cr(VI) poly-chromate or dichromate
species. Hoang et al. [3] has shown that no Cr(V)=O
6
Yanping Sun et al./ Journal of Natural Gas Chemistry Vol. 11 No. 1 2002
species are apparent in the Cr2 O3 /La2 (CO3 )3 synthesized for this study, while an IR band at 917
cm−1 corresponds to the formation of surface chromate species, and the band at 1,079 cm−1 is from
La2 (CO3 )3 . This catalyst has also been investigated by XPS and temperature-programmed decomposition (TPD), showing that the active phase for
ODH of isobutane is surface oxide chromates in which
chromium is chemically bound to the oxygen of lanthanum carbonate, and not to bulk chromium oxide
or LaCrO3 [3].
4.2. Oxygen species on the ODH of isobutane over
Cr2 O3 /La2 (CO3 )3
Lattice oxygen, and not gas-phase or physisorbed
oxygen, must be involved in the dehydrogenation of
isobutane and formation of water. This is because
all temperature-programmed reaction profiles were
recorded in the absence of gas-phase oxygen, and following repeated calcinations to 400 . As a consequence, the ODH of isobutane follows a Mars-van
Krevelen mechanism, which is common for many selective oxidation reactions, whereby adsorbed hydrocarbon reacts with lattice oxygen [18-20].
of the Cr2 O3 /La2 (CO3 )3 catalyst has previously been
reported by Hoang et al. [2] There are two explanations for the low apparent activation energy for the
ODH of isobutane on Cr2 O3 /La2 (CO3 )3 .
Table 4. Apparent activation energies for
ODH of isobutane and n-butane
Reactant
Catalyst
E3 +∆Hads Reference
(kJ/mol)
isobutane Cr2 O3 /La2 (CO3 )3 55.3 ± 5.3 This work
isobutane
CrOx /SiO2
83
[4]
isobutane
KCrOx /SiO2
64
[4]
isobutane
CrOx /Al2 O3
19
[4]
isobutane
KCrOx /Al2 O3
88
[4]
isobutane
CrOx /TiOI 2
106
[4]
isobutane
/TiOI
83
[4]
2
72
[4]
/TiOI
78
[4]
isobutane
isobutane
KCrOx
CrOx
/TiOII
K-CrOx
2
2
4.3. Kinetic Parameters for ODH of isobutane
Haber [18] has indicated that ODH and complete
oxidation to CO2 proceed via two distinct pathways.
For ODH two hydrogen atoms are abstracted, water
desorbs and a new C–C bond forms, while for complete oxidation hydrogen abstraction is followed by
oxygen atom addition. As discussed below, it is most
likely that for the current system, CO2 is a product
from the catalyst surface, rather than from complete
oxidation of isobutane. Hence, in the following kinetic
analysis of the ODH of isobutane, CO2 formation is
not included in the calculations. Experimental activation energies reported in the literature for the ODH
of isobutane and n-butane over metal oxides are listed
in Table 4. These are apparent values and represent a
combination of the adsorption enthalpy (∆Hads ) and
the intrinsic reaction activation energies (E3 ). These
apparent activation energies for the ODH of isobutane
and n-butane are strongly dependent on the type of
catalyst and the nature of the support. For the ODH
of isobutane over Cr2 O3 /La2 (CO3 )3 the activation energy (55 ± 5 kJ/mol) is at the lower end of these energies, but greater than that reported for isobutane over
CrOx /Al2 O3 (19 kJ/mol). The high ODH reactivity
isobutane
CrOx /MgO
124
[4]
isobutane
K-CrOx /MgO
83
[4]
isobutane
CrOx /ZrO2
70
[4]
isobutane
K-CrOx /ZrO2
93
[4]
n-butane
α-NiMoO4
39.4
[30]
n-butane
3%Cs-NiMoO4
75.9
[30]
1. The Cr–O bond energy in Cr2 O3 is low
and Cr3+ reduction potential is high. Previous
temperature-programmed reduction, oxygen TPD
and allyl iodide probe reaction measurements [21]
have shown that the surface oxygen on chromium
oxide-based catalyst is active at relatively low temperatures. Furthermore it has been reported that on the
surface of chromium oxides Cr3+ ions readily oxidize
to Cr6+ with uptake of oxygen, to form surface Cr6+ –
O; labile oxygen species [19,22]. The apparent activation energy is the sum of ∆Hads and the activation
energy of the rate-determining step, which for ODH
is assumed to be C–H fission. Generally, molecular
adsorption of isobutane on chromates is very weak
[22], so ∆Hads values are small. As a result, apparent activation energies often reflect the bond fission
step. Activation of the C–H bond involves a concerted reaction with lattice oxygen atoms, leading to
alkoxide and hydroxyl groups as well as reduction of
metal cations. Therefore, the intrinsic activation energy depends on the reduction potential of the cations
and on the strength of the metal–O–H bonds. More
Journal of Natural Gas Chemistry Vol. 11 No. 1 2002
reducible cations and low metal–O–H bond energies
lead to smaller activation energies, and to higher ODH
rates. The cleavage of C–H bonds in alkanes, however,
also depends on the electron density (basicity) of the
lattice oxygen anions that abstract the H atoms [23].
2. The support has an indirect effect on the apparent activation energy. The catalytic properties of
supported chromium oxide catalysts have been attributed to surface chromium species which form as
a result of metal-support interaction [24], while bulk
oxides are inactive phases [25]. Hoang et al. [24]
showed that lanthanum carbonate is a particularly
effective support to chromium oxide for the ODH of
isobutane. The introduction of La2 (CO3 )3 changes the
acidity and basicity of chromium oxide, which weakens oxygen bonds or creates more reducible metal
cations.
Intrinsic pre-exponential factors (1011.5±2.2 s−1 ),
listed in Table 3, are similar to typical factors observed for concerted unimolecular gas-phase reactions where entropy changes between reactant and
transition-state are negligible (ekT /h ) [26]. This confirms that the rate-determining step involves a concerted rearrangement. Implicit in calculating the
ODH pre-experimental factors is that the coverage of
external sites is equal to the coverage of active sites.
As the magnitude of these factors is not large, only
the external surface of Cr2 O3 /La2 (CO3 )3 must be involved in the ODH of isobutane, or diffusion into micropores of the catalyst is rapid. For this catalyst,
the surface area is low (20 m2 /g), indicating limited
micropore structure.
An assumption made in deriving equation (5) is
that k−1 k3 . However, the apparent activation energy and pre-exponential factor are both sufficiently
small that k3 maybe similar in magnitude to k−1 . If
k3 = k−1 then equation (7) becomes
o
(IC4 H10 + IC4 H10 ) αProduct
Z =
−ln
2αC4 H10
IProduct
1
ln + (−E1 /RT ) (8)
2
In this case the Arrhenius plot, shown in Figure 3, will
have an intercept at ln (1/2) and the apparent activation energy equals the activation energy for isobutane adsorption on the catalyst surface. Of the kinetic parameters listed in Table 3 the apparent preexponential factor of log (0.5 ± 0.6) is in close agreement with ln (1/2), but the activation energy (55 ±
5 kJ/ mol) is well in excess of that expected for adsorption.
7
4.4. Formation of Carbon Dioxide
Figure 1 contains a representative plot of corrected m/e = 44 abundance characteristic of evolved
carbon dioxide plotted against temperature. The
abundance has been corrected for background CO2
as well as contributions from isobutane. This figure shows that a large amount of CO2 is produced
during reaction. Evolution rate is approximately 4–
10 times larger than that evolved during background
measurements. The yield of CO2 , shown in Figure 2,
is ca. 8 times greater then the yield of isobutene.
There are three possible sources of CO2 from the
Cr2 O3 /La2 (CO3 )3 catalysis of isobutane.
1. Complete catalytic oxidation of hydrocarbons
to COx in the presence of gas-phase oxygen is generally described by a Langmuir-Hinshelwood mechanism in which adsorbed hydrocarbon reacts with dissociatively adsorbed oxygen [18,20,27]. In the Knudsen cell, gaseous O2 is excluded, so this pathway is
not apparent. Furthermore, CO is not observed during reaction, as also shown by Hoang et al. [2], and so
gas-phase oxidation of isobutane is not occurring. For
oxidation catalysis of hydrocarbons via lattice oxygen
the initial step is hydrogen atom abstraction to form
a carbocation. The relative rates of ODH to complete oxidation depend on the relative rates of C–C
and C–H bond breaking within adsorbed alkyl species,
and on the propensity for C–O bond formation [11].
The breaking of C–C bonds require concurrent interaction of two oxygen atoms with two neighboring carbon atoms. Such interaction is favored in the case
of linear alkanes, which can adopt a planar configuration, parallel to the catalyst surface. Owing to
the quasi-tetrahedral configuration of isobutane, the
tert-carbon atom would be located at a greater distance from the surface than the primary carbons, and
so C–C fission is difficult. Moreover, the most stable location of the carbocation’s positive charge is on
this more distant tert-carbon, which may hinder C–O
bond formation. Complete oxidation cannot be dismissed as a mechanism to CO2 , but the absence of CO
and the decreased possibility of C–C bond breaking
and C–O bond formation, should mean a low selectivity for this pathway.
2. A cracking mechanism is possible whereby a
methyl group is eliminated, decomposes to carbon and
is subsequently oxidized to CO2 . For this pathway,
other products such as propene or ethene should be
observed, however none were found in the mass spectral profiles.
8
Yanping Sun et al./ Journal of Natural Gas Chemistry Vol. 11 No. 1 2002
3. Removal of oxygen atoms from the catalyst surface during the ODH of isobutane brings about a simultaneous removal of CO2 from the catalyst/support
system. This view is supported by the results of
Hoang et al. [3] They found that the CO3 2− /La
ratios in Cr2 O3 /La2 (CO3 )3 obtained by XPS analysis after reaction were much lower than the prereaction ratio of 1.5. The similarity in the trends
of increasing isobutene and carbon dioxide evolution
rates with increasing temperature makes this the most
likely source of CO2 during reaction. This mechanism can also explain the relative activities and selectivities of supported chromium oxide catalysts studied by Hoang et al. [24,28] That is, the reactivity
of Cr2 O3 supported on La2 (CO3 )3 is greater than
that on other lanthanum compounds even through
La2 (CO3 )3 possesses a lower surface area (20 m2 /g).
Mamedov et al. [29] found that the ODH of isobutane to isobutene over Cr–Mn–O/Al2 O3 catalyst was
three times greater in the presence of CO2 than in
the absence of CO2 . Mamedov et al. showed that
the CO2 reoxidized the catalyst and removed the hydrogen from the product mixture by the reverse water
gas shift reaction therefore reducing the occurrence of
the reverse reaction.
5. Conclusion
A low-pressure Knudsen cell was used to
investigate the ODH of isobutane over the
Cr2 O3 /La2 (CO3 )3 catalyst. Despite the low yields
of isobutane, the results were found to be sufficiently
sensitive, to calculate accurate rate parameters.
The evolved products of the ODH are isobutene,
water and carbon dioxide. The formation of isobutene
proceeds with participation of the lattice oxygen on
the surface of the Cr2 O3 /La2 (CO3 )3 catalyst. The
large amount of CO2 emission is most likely to be
from the catalyst surface immediately following ODH
of isobutane, and so is not a primary product of the
reaction. This evolved CO2 may enhance the activity
of the Cr2 O3 /La2 (CO3 )3 catalyst.
The geometric surface area of the catalyst and the
effective area of the exit aperture were used to determine the collision frequency of the isobutane molecule.
The collision frequency, the ratio of conversion factors of reactant and product as well as their measured abundances during the reaction allowed for an
accurate calculation of the kinetic parameters for the
ODH of isobutane over Cr2 O3 /La2 (CO3 )3 in the ab-
sence of gaseous oxygen. These kinetic parameters are
expressed as the intrinsic Arrhenius rate constant:
k3 =1011.5±2.2 exp{-((55±5)-∆Hads)/RT}
The magnitude of the pre-exponential factor indicates
that the ODH of isobutane proceeds on the external
catalyst surface and involves a concerted rearrangement of the adsorbed species. The apparent activation energy is significantly less than most energy
barriers previously reported in the literature for the
ODH of isobutane. Hence Cr2 O3 /La2 (CO3 )3 is an
active catalyst for the oxidative dehydrogenation of
isobutane.
Acknowledgement
Financial support from the Australian Research
Council is grateful acknowledged. Yanping Sun is also
appreciative of UNERS and OPRS awards.
References
[1] Huff M, Schmidt L D. J Catal, 1995, 155: 82
[2] Hoang M, Mathews J F, Pratt K C. J Catal, 1997,
171: 320
[3] Hoang M, Hughes A, Mathews J F, Pratt K C E.
J Catal, 1997, 171: 313
[4] Grabowski R, Grzybowska B, Sloczynski J,
Wcislo K. Appl Catal A, 1996, 144: 335
[5] Moriceau P, Grzybowska B, Barbaux Y, Wrobel
G, Hecquet G. Appl Catal A, 1998, 168: 269
[6] Sun Y, Brown T C. J Catal, 2000: 194: 301
[7] Sun Y, Brown T C. Int J Chem Kinet, 2002, 34:
467
[8] Keyser L F, Moore S B, Leu M T. J Phys Chem,
1991, 95: 5496
[9] Underwood G M, Li P, Usher C R, Grassian V H.
J Phys Chem A, 2000, 104: 819
[10] Iczkowski R P, Margrave J L, Robinson S M. J
Phys Chem, 1963, 67: 229
[11] Khodakov A, Yang J, Su T, Iglesia E, Bell A T.
J Catal, 1998, 177: 343
[12] Chen K, Iglesia E, Bell A T. J Catal, 1999, 186:
325
[13] Chen K, Iglesia E, Bell A T. J Catal, 2000, 192:
197
[14] Golden D M, Spokes G N, Benson S W. Angewandate Chemie, 1973, 12: 534
[15] Pitt I G, Gilbert R G, Ryan K R. J Chem Phys,
1995, 102: 3461
[16] Cimino A, Cordischi D, Febbraro S, Gazzoli D,
Indovina V, Occhiuzzi M, Valigi M. J Mol Catal,
Journal of Natural Gas Chemistry Vol. 11 No. 1 2002
1989, 55: 23
[17] Indovina V, Cordischi D, De Rossi S, Ferraris G.
J Mol Catal, 1991, 68: 53
[18] Haber J. Handbook of Heterogeneous Catalysis.
Vol 5. Weinheim: VCH, 1997. 2253
[19] Sloczyniski J, Grzybowska B, Grabowski R, Kozlowska A, Wcislo K. Phys Chem Chem Phys,
1999, 1: 333
[20] Sloczynski J. Appl Catal A, 1996, 146: 401
[21] Grzybowska B, Sloczynski J, Grabowski R,
Wcislo K, kozowska A, Stoch J, Zielinski J. J
Catal, 1998, 178: 687
[22] Grzybowska B, Sloczynski J, Grabowski R,
Keromnes L, Wcislo K, Bobinska T. Appl Catal
A, 2001, 209: 279
9
[23] Banare M A. Catal Today, 1999, 51: 319
[24] Hoang M, Mathews J F, Pratt K C. React Kinet
Cata Lett, 1997, 61: 21
[25] McDaniel M P. J Catal, 1982, 76: 17
[26] Benson S W. Thermochemical Kinetics. New
York: Wiley, 1976
[27] Stern D L, Grasselli R K. J Catal, 1997, 167: 560
[28] Hoang M, Mathews J F, Pratt K C. Chemtech,
1999, 29: 45
[29] Mirzabekova S R, Mamedov A K. Kinet Catal,
1994, 35 (6): 834
[30] Madeira L M, Portela M F, Mazzocchia C, Kaddouri A, Anouchinsky R. Catal Today, 1998, 40:
229