Oxygen adsorption and catalytic performance in oxidative dehydrogenation of isobutane on chromium oxide-based catalysts

Article abstract of DOI:10.1039/a806136j

Oxygen adsorption on γ-Al₂O₃-, TiO₂ (anatase)-supported chromium oxide and on unsupported chromia in amorphic and crystalline form, has been studied with a gravimetric method in the temperature range 523-623 K. The equilibrium oxygen uptake at p → ∞ was the same for all the studied catalysts, the rate of the adsorption was, however, different, increasing in the order: CrO(x)/Al₂O₃ < CrO(x)/TiO₂ < Cr₂O₃ amorph. < Cr₂O₃ cryst. Measurements of oxidative dehyrogenation of isobutane on the catalysts given above have shown that the selectivity to isobutene decreases in the same sequence. The higher selectivity values and the change in the rate constant of the isobutane oxidative dehydrogenation with conversion, observed for the Al₂O₃- and TiO₂-supported chromium oxide catalysts, have been ascribed to the lower coverage of the surface of the catalyst with oxygen in the stationary state of the isobutane oxidative dehydrogenation, as compared with unsupported chromia.

Full text of DOI:10.1039/a806136j

View Article Online / Journal Homepage / Table of Contents for this issue
Oxygen adsorption and catalytic performance in oxidative
dehydrogenation of isobutane on chromium oxide-based catalysts
Jerzy S¡oczynski, Barbara Grzybowska, Ryszard Grabowski, Antonina Koz¡owska and
Krzysztof Wcis¡o
Institute of Catalysis and Surface Chemistry, Polish Academy of Sciences, ul. Niezapominajek,
30È239 Krakow, Poland
Recei¿ed 4th August 1998, Accepted 25th No¿ember 1998
Oxygen adsorption on c-Al O -, TiO (anatase)-supported chromium oxide and on unsupported chromia in
2
3
2
amorphic and crystalline form, has been studied with a gravimetric method in the temperature range 523È623
K. The equilibrium oxygen uptake at p ] O was the same for all the studied catalysts, the rate of the
adsorption was, however, di†erent, increasing in the order: CrO /Al O \ CrO /TiO \ Cr O amorph. \
x
2
3
x
2
2 3
Cr O cryst. Measurements of oxidative dehyrogenation of isobutane on the catalysts given above have shown
2
3
that the selectivity to isobutene decreases in the same sequence. The higher selectivity values and the change in
the rate constant of the isobutane oxidative dehydrogenation with conversion, observed for the Al O - and
TiO -supported chromium oxide catalysts, have been ascribed to the lower coverage of the surface of the
2
3
2
catalyst with oxygen in the stationary state of the isobutane oxidative dehydrogenation, as compared with
unsupported chromia.
groups, in which oxygen is bound with di†erent strengths. As
shown in the IR studies of Zecchina et al.,22 the CrxO
stretching frequency (an indicator of the CrwO bond energy)
for Cr O depends both on the coordination number of the
1
Introduction
Supported chromium oxide systems have been known, for
many decades, as active catalysts in a number of hydrocarbon
reactions such as polymerization,1 and dehydrogenation and
dehydrocyclization,2h6 and have been studied also in the
selective reduction of nitrogen oxides in the presence of
oxygen.7h9
Quite recently, we have found that supported chromium
oxides are active and selective in oxidative dehydrogenation
(OHD) of isobutane.10h12 The product of this reaction, iso-
butene, is a key reactant for the production of methyl tert-
butyl ether (an additive for lead-free petrol), and of acrylic
acid and methacrylates. The best results in the ODH of iso-
butane have been obtained with CrO /Al O and CrO /TiO
2
3
adsorbing chromium ions and on the environment of the
CrxO groups: several di†erent chromyl groups have been
identiÐed on the surface of chromia in these studies. In recent
years, the presence of di†erent oxygen species deriving from
di†erent types of chromiumÈoxygen complexes on the surface
of chromia and supported chromia systems, has been postu-
lated from studies with hydrogen temperature-programmed
reduction (TPR),11,23h27 X-ray photoelectron spectroscopy
(XPS),27h31 EPR32h34 and Raman and IR22,31,35h38 spectro-
scopies. Isolated and polymerized CrO polyhedra containing
x
Cr3`, Cr5`, and Cr6` ions have been identiÐed on the surface
of oxide supports, in addition to exposed Cr3` ions of di†er-
ent degree of coordinative unsaturation. The latter species are
mainly observed on the surface of pure Cr O . The presence
x
2
3
x
2
systems, which are active at relatively low temperatures 473È
523 K.10h12
2
3
Both the selective reduction of nitrogen oxides and the
ODH of isobutane involve adsorption of oxygen and reoxida-
tion of the catalyst surface as the reaction steps. Studies of the
interaction of oxygen with the catalyst are, hence, essential to
obtain an explanation of the mechanism of the catalyst action
and of the reaction mechanism on a molecular scale.
Adsorption of oxygen on pure chromia has been studied
since the 30s, initially at low temperatures.13h16 The results of
the early studies on this subject can be found in a monograph
by Trapnell.17 Winter18 included Cr O in his extensive
and distribution of di†erent [CrO ] species on the supports
x n
depend on the nature of the support, the chromium loading
and the preparation conditions. At higher Cr content, amorp-
hous and crystalline a-Cr O are also observed in the chro-
2
3
mium oxide/support systems.
Both low- and high-temperature adsorption of oxygen on
supported systems have been studied, mostly to determine the
area of the chromia phase on the support. Only in a few cases
have correlations been sought between the data on oxygen
chemisorption and the catalytic performance of chromia-
2
3
studies on isotopic exchanges of oxygen on oxides. The low-
temperature oxygen adsorption on supported chromia has
been proposed as a method for the determination of the area
of chromia on a support.16 High-temperature adsorption of
oxygen on pure and alumina-supported chromia, studied by
Voltz and Weller19 and Deren and Haber,20,21 has been
related to the oxidation of the surface Cr3` ions to Cr6`.
Oxygen chemisorption on chromia and supported chromia
systems has been found to be dissociative, even at room tem-
perature, as indicated by the high adsorption heat. Mono-
atomic oxygen species are present in di†erent surface CrxO
containing systems e.g. in the oxidation of CO to CO ,16 or
2
in polymerization.39
In the present work, oxygen chemisorption has been studied
on selected CrO /Al O and CrO /TiO catalysts which have
x
2
3
x
2
shown the highest activity and selectivity to isobutene in the
ODH of isobutane,11,12 and an attempt has been made to
correlate the data on oxygen adsorption with the catalytic
performance for this reaction. For comparison, the chemisorp-
tion of oxygen and isobutane ODH on unsupported chromia,
both in amorphous and crystalline forms have also been
examined.
Phys. Chem. Chem. Phys., 1999, 1, 333È339
333

View Article Online
Table 1 Basic characteristics of chromium catalysts
Since the ODH of isobutane is the Ðrst selective oxidation-
type reaction of hydrocarbons reported on chromium oxide-
based catalysts, it seemed of interest to check to what extent
the general principles of selective oxidation are fulÐlled in
catalysts of this type. Oxygen adsorption has been studied in
conditions close to those used in the catalytic reaction of the
ODH of isobutane: particular interest has been paid to mea-
surements of the kinetics of the oxygen sorption, which may
determine the oxidation state of the surface during the cata-
lytic oxidation reaction. Additionally, the state of the catalyst
surface has been checked by measurements of surface poten-
tial (work function) performed in situ, in the reaction mixture.
It has been shown previously40,41 that this technique can
provide some qualitative information about the reduction of
the catalyst during reactions on oxide catalysts.
Cr6`/atom nm~2
Fresh catalyst
SpeciÐc
surface
area
/m2g~1
After
reactionc
BRa
Cr
(wt.%)
tot
Catalyst
10 CrAl
BRa
TPRb
8.4
104
2.64
1.98
4.30
3.93
6.34
2.43
0.06
10 CrTi
0.6
68.4
68.4
8
28
210
È
3.56
È
È
0.71
È
a-Cr O
amorph. Cr O
2
3
2
3
a Determined by the BunsenÈRupp method. b Determined by the
TPR method (ref. 11). c Catalysts, ex situ, after ODH of IB at 523 K
for 10 h, then cooled down in the reaction mixture. d After evacuation
at 1 ] 10~4 kPa and 623 K for 24 h.
2
Experimental
BET method with argon as an adsorbate), and the content of
Cr ions of valency higher than ]3, both for fresh samples and
after ODH of isobutane.
Catalysts
The results indicate that, beside Cr3` ions, the surface of
the catalysts contains a considerable fraction of oxidized chro-
mium ions.
The supported catalysts CrO /Al O and CrO /TiO were
x
2
3
x
2
prepared as described previously.11,12 They were obtained by
impregnation with an aqueous solution of chromium nitrate
(p.a., POCH, Poland) of the Al O support (c-phase, basic
2
3
Chemisorption of oxygen
Merck, 111 m2 g~1, 60È200 lm), and the TiO (anatase,
2
Police, Poland, 8 m2 g~1) followed by drying for 5 h at 393 K
and calcination in a Ñow of air at 773 K.
This was studied by a gravimetric method using a S 3DV Sar-
torius microbalance coupled to
a conventional vacuum
The Cr content corresponded to ca. one theoretical mono-
system. 0.5 g of sample was degassed at 623È648 K at a pres-
sure below 1 ] 10~4 kPa for 24 h till a constant mass was
obtained and then cooled down to the measurement tem-
perature. Oxygen (99.0%), puriÐed by vacuum distillation, was
then introduced and the mass gain was recorded until con-
stant mass was attained. The measurements were carried out
in the temperature range 573È648 K and the oxygen pressure
range 0.5È15 kPa. In view of the large volume of the experi-
mental set-up the measurements were carried out in isobaric
conditions. The accuracy of the measurement was 1 lg. It was
checked, in preliminary experiments, that the oxygen uptake
was reversible in successive degassingÈadsorption cycles. The
control chemical analysis of the 10 CrAl sample after degass-
ing in vacuum and cooling to room temperature showed a
decrease in the total Cr6` content with respect to the fresh
sample of ca. 30%. The initial value of the Cr6` content was,
however, restored after the oxygen chemisorption measure-
ments. These facts suggest that the oxygen uptake consists of
reoxidation of the Cr3` ions, reduced by vacuum pretreat-
ment, to Cr6`.
layer of Cr O on the support surface. One monolayer of
2
3
Cr O , was estimated from the crystallographic data of the
2
3
oxide as equivalent to 10 atoms of Cr per 1 nm2 of the
support, in agreement with MacIver and Tobin.15 The
samples are denoted further in the text by symbols 10 CrAl
(Ti).
No crystalline Cr O was observed in the samples on XRD
2
3
examination, the Raman spectroscopy revealed, however, a
sharp band at ca. 560 cm~1, characteristic of Cr O , in addi-
2
3
tion to intense bands in the region 8000È1000 cm~1 ascribed
to the presence of polychromates for the 10 CrAl samples.11
The Raman spectra of the 10 CrTi sample did not show the
bands characteristic of the CrwO vibrations, owing to the low
amount of the deposited chromium phase. The bands corre-
sponding to the polychromates were observed, however, when
TiO of higher speciÐc surface area (and thus of higher
2
amount of chromium oxide phase at 1 monolayer content)
was used. Both samples contained chromium ions in an oxida-
tion state higher than 3], as revealed by chemical analysis
with the BunsenÈRupp (BR) method and the XPS tech-
nique.11,12 The BR method consists in treating a sample with
concentrated HCl at 373 K for 30 min and determining the
chlorine evolved by iodometric titration.42 In the case of the
10 CrAl sample, the total excess charge determined by this
method was attributed to Cr6` ions only,11 in the case of the
10 CrTi sample Cr5` ions were also detected by XPS and
EPR.12 For both samples, the results of the BR analysis were,
however, expressed in terms of a number of Cr6` ions per
nm2 of catalyst surface.
The rate of adsorption was obtained from the slope of a
curve of the mass gain vs. the adsorption time. For small time
intervals:
1
*m
r \
[(lmol O) s~1 m~2]
(1)
a
16m S *t
o
where *m is the mass gain of the sample (lg) in the time inter-
val *t(s), m the sample mass (g), S the speciÐc surface area of
o
the sample (m2 g~1).
In analysis of the adsorption kinetics the adsorption rate
related to the coverage of the surface with oxygen was used:
Amorphous Cr O was prepared following the method
2
3
described by Curry-Hyde and Baiker,24 by slow precipitation
from chromium nitrate with ammonia, drying and treatment
at 623 K for 3 h in a stream of hydrogen, followed by calcina-
tion at 573 K in air for 3 h. Crystalline Cr O was obtained
dH
R \
(s~1)
(2)
a
dt
2
3
by calcination of the amorphous chromia in air at 873 K for
6 h. XRD analysis of the chromium oxides thus obtained
showed a pattern characteristic of a-Cr O , in accordance
where H \ *m/(*m) , and (*m) is a mass gain correspond-
=
=
ing to equilibrium adsorption at a given temperature (a limit-
ing value of *m at time ] O and at p ] O).
2
3
with the ASTM data, no di†raction lines were observed for
the Ðrst (amorphous) sample.
In order to obtain the maximum coverage with adsorbed
oxygen, adsorption isotherms were determined at 623 K in the
pressure range 0.2È15 kPa. Measurements at pressures lower
than 0.2 kPa were not possible owing to the thermomolecular
Table 1 gives the basic characteristics of the studied
samples, including speciÐc surface area (determined by the
334
Phys. Chem. Chem. Phys., 1999, 1, 333È339

View Article Online
e†ects (thermal transpiration) in the vacuum microbalance
system, which caused apparent changes of the sample mass.
Surface potential
Surface potential U, was measured by the vibrating condenser
method; the method and the experimental set-up have been
described in ref. 43. The measurements were performed under
a Ñow of 20% O in Ar in the temperature range 323È723 K,
2
and in the Ñow of reaction mixture: isobutane (IB)ÈO ÈAr \
2
9.2/18.2/72.6 vol.% in the temperature range 523È603 K. The
total Ñow rate was 300 ml min~1. At each temperature the
samples were kept in a Ñow of O ÈAr, or IBÈO ÈAr mixture
2
2
for 3È5 h till a constant value of U was obtained. The mea-
surements were performed with increasing and decreasing
temperature: no change in the U values was observed within
the limits of the experimental error (^5 mV). The values of
the surface potential are given relative to the graphite elec-
trode, an increase in U indicating that the surface becomes
more negatively charged.
Fig. 1 Isotherms of oxygen adsorption on a-Cr O , 10 CrTi and 10
2
3
CrAl catalysts at 623 K.
Catalytic activity measurements
The activity of the catalysts in the ODH of IB was measured
in a Ðxed-bed Ñow apparatus at 523 K. A stainless-steel
reactor (120 mm long, id 13 mm) was coupled directly to a
3
Results and discussion
Oxygen chemisorption
series of gas chromatographs. Isobutene, CO and CO were
found to be the main reaction products. The amounts of the
Fig. 1 shows typical isotherms of oxygen adsorption at 623 K
2
for 10 CrAl, 10 CrTi and a-Cr O catalysts. The experimental
2
3
degradation products (C , C ) and of oxygenates
data can be described by the Langmuir isotherm for the disso-
2
3
(methacrolein and methacrylic acids) were \1% of the total
ciative adsorption (the correlation coefficient \ 0.999):
amounts of products. The carbon balance calculated taking
into account isobutene, CO and CO was 97 ^ 2% for con-
*m
bp 0.5
2
O
H \
\
(3)
version of IB P 10%, at lower conversion the C balance was
(*m)
1 ] bp 0.5
=
O
poorer (90 ^ 5%). The reaction mixture contained 9.2 vol.%
IB (Phillips Petroleum Corp., 99%) in air. 0.5 ml catalyst
sample (grain size 0.63È1 mm) diluted (1 : 1) with glass beads
of the same diameter were used. It was checked in preliminary
measurements in which the mass and the grain size of a
sample were varied, that, under these conditions, transport
phenomena do not limit the reaction rate. For each sample
the measurements were performed at contact times varying
between 0.1 and 4 s which were obtained by changing the Ñow
rate of the reaction mixture. The total conversion of IB did
not exceed 20%.
where b is an adsorption coefficient and P the oxygen pres-
O
sure.
The lines given in Fig. 1 correspond to the values calculated
from eqn. (3). The values of maximum adsorption (a ) corre-
=
sponding to (*m) for di†erent catalysts are summarized in
=
Table 2. They do not di†er much and are equivalent to 7.5È
8.5% of an oxygen monolayer, corresponding to the closest
hexagonal oxygen packing of oxygen ions in oxides (14.7 at O
nm~2).
For all the catalysts studied the rate of oxygen uptake
(oxygen adsorption) was determined as a function of oxygen
pressure, temperature and degree of coverage of the surface.
As shown in Fig. 2 the initial rate of the oxygen adsorption (at
The catalytic data are presented in the form of plots of
selectivities to di†erent products S vs. the total conversion of
p
IB, conversion \ (C [ C )/C ] 100, S \ C /&C , where C
i
f
i
p
p
p
i
H \ 0) is proportional to p 0.5, which suggests the disso-
and C are the concentrations of IB at the entrance and the
exit of the reactor, respectively, and C is the concentration of
product p in exit gas. For a given sample the values of the
O
f
ciative character of the O adsorption. At constant oxygen
2
p
pressure, the adsorption rate decreases rapidly with the
increase in the oxygen coverage. Kinetics of this process can
be described by
conversion and S were reproducible within 2È5%. The total
rates of IB disappeance per unit of the catalyst surface, r
were also calculated for di†erent values of the total conver-
sion.
,
sp
R \ k p 0.5(1 [ H)e~aH
(4)
a
a O
A blank test without a catalyst or with glass beads in the
reactor showed no conversion of IB at the reaction tem-
perature. Pure alumina support was not active either, the IB
conversion at 523 K being \0.05%. No deactivation of the
catalysts was observed during 10 h of measurement.
where k is the adsorption rate constant and a is a constant.
a
Eqn. (4) is a generalized form of the Elovich equation.44
According to Elovich, Roginski and Zeldovich45 the exponen-
tial term in the equation indicates the energetic inhomogeneity
of the surface, the activation energy of the adsorption process
Table 2 Oxygen adsorption data for chromia catalysts
Temperature
/K
a a
E¡/
a/
=
a
Catalysts
(atom O) nm~2
ln k¡
kJ mol~1
kJ mol~1
a
10 CrAl
10 CrTi
a-Cr O
573È648
523È623
523È623
524È573
1.10
1.23
1.25
1.29
34.5
1.53
9.98
13.3
213 ^ 10
44 ^ 5
32.4
34.3
32.4
23.5
54 ^ 5
2
3
amorph. Cr O
98 ^ 10
2
3
a Maximum adsorption at 573 K.
Phys. Chem. Chem. Phys., 1999, 1, 333È339
335

View Article Online
Fig. 4 Surface potential values for 10 CrTi and 10 CrAl catalysts at
di†erent temperatures. Filled symbols: in O ÈAr atmosphere, open
2
symbols: in IBÈO ÈAr mixture.
2
Coefficient a does not vary markedly for the studied catalysts,
indicating that the extent of the surface inhomogeneity is
similar in all the samples.
Fig. 2 Initial rate of adsorption of oxygen as a function of the square
root of the oxygen pressure, p 0.5, for 10 CrAl and 10 CrTi catalysts.
O
Surface potential
increasing linearly with increase in H:
Fig. 4 shows the changes in the surface potential, U, with tem-
perature for the 10 CrAl and 10 CrTi catalysts in an oxygen
E \ E ] aH
(5)
0
atmosphere and in a Ñow of the IBÈO ÈAr reaction mixture.
The constant a characterizes the degree of the energetic inho-
mogeneity of the surface, high values of a corresponding to a
high degree of inhomogeneity. The inhomogeneity of sites on
chromia supported on alumina has been previously reported
for nitric oxide adsorption on this system.46 Eqn. (4) can be
easily transformed to a linear form:
2
For both samples, the U values in oxygen are high and do not
vary much with temperature. The 10 CrTi catalyst shows
slightly higher values (by ca. 0.1 V) as compared with the 10
CrAl sample. The U values in the IBÈO ÈAr reaction mixture
2
are considerably lower (by 1È1.5 V) than those in oxygen, and
are signiÐcantly higher (by 0.3È0.5 V) for the 10 CrTi sample
as compared with the 10 CrAl one. The lower U values in the
reaction mixture, as compared with those in oxygen, indicate
the lower negative charge of the surface. The lowering of the U
values in the hydrocarbonÈoxygen reaction mixture has been
observed previously for vanadia titania catalysts41 and was
ascribed to the impoverishment of the surface in negatively
charged oxygen species: O2~, or O~ i.e. to partial reduction
of the catalyst surface. The present results suggest, then, that
the surface of both CrAl and CrTi catalysts is partially
reduced in the stationary state of the catalytic reaction, the
CrAl sample being more reduced than the CrTi catalyst. The
U values for the CrTi catalyst hardly change with the tem-
perature, whereas those for the 10 CrAl decrease slightly
between 523 and 573 K and remain costant at higher tem-
peratures. This suggests that, for the latter catalyst, the surface
is less oxidized at the higher reaction temperatures.
ln[RP ~0.5(1 [ H)~2] \ ln k [ aH
(6)
O
a
Fig. 3 presents kinetic data at 623 K for the studied
samples plotted in coordinates of eqn. (6). Analogous curves
have been found for lower adsorption temperatures. The
values of E¡ ln k¡ and a for the studied catalysts are reported
a
a
in Table 2. The values of the initial activation energy, E , and
0
the pre-exponential factor ln k¡ at H \ 0 for the process of
a
oxygen chemisorption di†er considerably. A compensation
e†ect can, however, be observed, the high values of E corres-
ponding to a high pre-exponential factor: the low values of E
are, on the other hand, accompanied by low values of ln k¡ .
0
0
a
Catalytic activity
Fig. 5 shows the dependence of the total conversion of iso-
butane on the contact time, and Fig. 6 and 7 give the selec-
tivities to isobutene, CO and CO as a function of IB
2
conversion, for all the catalysts under study. As seen, for all
the catalysts, the selectivity to isobutene decreases with the
increase in IB conversion (increasing contact time). This
implies a sequential reaction pathway in which isobutene
formed in the Ðrst step is further oxidized to CO . Extrapo-
2
lation to zero conversion indicates, moreover, that a fraction
of CO is formed by a parallel route. The selectivity to CO (a
2
minor product) increases with isobutene conversion for the 10
CrAl, 10 CrTi and amorphous Cr O , and does not change
2
3
with conversion for crystalline a-Cr O . Moreover, as indi-
2
3
cated by the data extrapolated to zero conversion, CO is
formed in a parallel reaction on chromium oxides, both in
amorphous and crystalline form, whereas for 10 CrAl and 10
CrTi it is produced only by the consecutive path.
Fig. 3 Rate of oxygen adsorption at 623 K as a function of oxygen
coverage, H, in coordinates of eqn. (6).
336
Phys. Chem. Chem. Phys., 1999, 1, 333È339

View Article Online
The mechanism of the isobutane ODH on the studied cata-
lysts can then be described by the following parallelÈ
consecutive scheme:
where IB and IBx stand for isobutane and isobutene, respec-
tively.
It can be mentioned that the parallelÈconsecutive mecha-
nism is typical for selective oxidation of hydrocarbons leading
to the formation of oxygenated products47 and has also been
found for the ODH of propane on V O /TiO catalysts.48
2
5
2
As seen from Fig. 6, the selectivities to isobutene at compa-
rable conversions are di†erent for di†erent catalysts, decreas-
ing in the order: 10 CrAl [ 10 CrTi [ amorph. Cr O [ a-
Fig. 5 Total conversion of IB as a function of contact time for
unsupported chromia and for 10 CrAl and 10 CrTi catalysts at 523 K.
2
3
Cr O . The selectivities to CO increase in the same
2
3
2
sequence. No distinct correlation is found between the nature
of the catalyst and the selectivity to CO: it can, however, be
observed that the selectivities to this product are rather low
and the error in the determination of the CO content may be
high, which does not permit detailed discussion on the CO
formation.
To compare the activity of the studied catalysts, the speciÐc
rate of IB disappearance referred to the unit of speciÐc surface
area, r at di†erent degrees of conversion, has been calculated
sp
by numerical di†erentiation of the conversionÈcontact time
curves. The measurements of r at di†erent partial pressures
sp
of IB, C and oxygen C have shown that the reaction order
IB
O
with respect to IB is 1, whereas that with respect to oxygen is
0.5, i.e.
r
\ k C C 0.5
IB
(7)
sp
r
O
Fig.
8 presents the plot of the r /C C 0.5 ratio,
(corresponding to the rate constant k ) vs. the total conversion
of IB for the studied catalysts. The horizontal straight lines
sp IB
O
r
observed for amorphous and crystalline Cr O indicate that
Fig. 6 Selectivities to isobutene and CO as a function of the total
2
3
2
for these samples eqn. (7) is valid, i.e. k is constant over a wide
IB conversion for unsupported chromia and for 10 CrAl and 10 CrTi
r
conversion range. For the 10 CrAl and 10 CrTi catalysts, the
catalysts. Reaction temperature, 523 K.
rate constant decreases with conversion, the e†ect being par-
ticularly pronounced for the latter catalyst. Such behaviour
may be ascribed to a decrease in the number of active centres
Fig. 7 Selectivity to CO as a function of the total IB conversion for
unsupported chromia and for 10 CrAl and 10 CrTi catalysts. Reaction
temperature, 523 K.
Fig. 8 Rate constant for the disappearance of IB in the ODH reac-
tion, as a function of the total IB conversion for unsupported chromia
and 10 CrAl as 10 CrTi catalysts. Reaction temperature, 523 K.
Phys. Chem. Chem. Phys., 1999, 1, 333È339
337

View Article Online
(the pre-exponential factor in the Arrhenius equation) when
the total conversion increases, and/or to the change in the
reaction mechanism (e.g., the change in the rate-determining
step), which could give rise to di†erent activation energies for
the reaction.
ported Cr O . Moreover, the dissociation of the oxygen mol-
2
3
ecule requires the presence of the two reduced centres in close
vicinity:52 this condition should be more easily fulÐlled in the
case of pure chromia, which is reduced to a greater extent,
than for that dispersed on the support CrO groups, reduced
x
to a lesser extent. Another factor, which may a†ect the reoxi-
dation rate, is the ease of electron transfer accompanying the
passage from an oxygen molecule to an O2 species, which
requires four electrons.
4
Discussion
A tentative explanation of the properties of the studied cata-
lysts in the ODH of IB can be proposed by comparing the
catalytic data with those obtained in the oxygen adsorption
studies, supplemented by chemical analysis and surface poten-
tial measurements. Let us observe, in the Ðrst place, that pre-
liminary studies with the pulse method of the ODH of IB on
Cr O and 10 CrAl catalysts have shown practically the same
initial activity and selectivity in the presence and absence of
gaseous oxygen.49 This suggests that the redox, Mars and Van
Krevelen mechanism,50 common for most of the selective oxi-
dation reactions,47 operates also in the case of IB ODH on
chromia-based catalysts. According to this mechanism, the
oxidation reaction proceeds in two steps: (a) reduction of an
oxidized catalyst by a hydrocarbon molecule, with incorpor-
ation of the catalyst oxygen into the oxidation products (water
in the case of the ODH reactions), and (b) reoxidation of the
reduced catalyst involving dissociative adsorption of the
gaseous oxygen present in the reaction mixture and the elec-
tron transfer. It is logical to assume that, in the chromia-based
One may speculate that electron transfer should be easy
on
the
chromia
surface,
occurring
along
the
Cr3`wOwCr6`wOwCr3`wO chains. When the CrO units
x
are isolated on the support surface electrons can be trans-
ferred to the oxygen molecule from only a limited number of
Cr ions. Such an e†ect could account for the high activation
energy of the oxygen chemisorption (reoxidation) observed for
the CrAl catalysts, which leads to the low rate of reoxidation.
In the case of the CrTi catalysts, for which the value of the
activation energy has been found to be much lower, electron
transfer could be easier, since the reducible ions, Ti4`, of the
support may participate in the exchange of electrons between
the distant Cr3` ions. The low rate of oxygen chemisorption
for this catalyst is due to the low value of the pre-exponential
3
3
factor k for which no explanation can be given at present.
0
The di†erence in the oxygen adsorption rates may also
account for the di†erent course of the rate constant vs. conver-
sion curves observed for unsupported Cr O and for 10 CrAl
2
3
catalysts, the centres for step (a) involve Cr6`O groups,
present in the studied catalysts as indicated by chemical
and 10 CrTi catalysts. It can be suggested that, for unsup-
ported chromia, the high rate of catalyst reoxidation ensures
constant, high coverage of the surface with oxidized species,
even at high conversions. The reaction rate may then be deter-
mined by the rate of reduction of the catalyst with IB [step
(a)]. For the 10 CrAl and 10 CrTi catalysts, with increase in
conversion the surface becomes more and more reduced, as
the relatively slow reoxidation step cannot keep pace with the
enhanced reducibility of the surface [both the primary IB
x
analysis, XPS and Raman spectroscopy,11,12 whereas exposed
Cr3` ions provide centres for the catalyst reoxidation (the
oxygen uptake). The surface potential measurements and the
chemical analysis indicate that the catalysts are partially
reduced during the catalytic reaction i.e. both Cr6` and Cr3`
species do exist on the surface. It follows, from the Mars and
Van Krevelen mechanism, that the coverage of the catalyst
surface with oxygen in the stationary state of the oxidation
reaction depends on the relative rates of the catalyst reduction
and reoxidation. The higher rates of oxygen chemisorption
observed for unsupported Cr O , both amorphous and crys-
talline, as compared with those for 10 CrAl and 10 CrTi cata-
lysts, suggest higher coverage with oxygen during the IB
ODH reaction in the case of Cr O . This is indeed conÐrmed
by the results of chemical analysis, which show a higher
content of hexavalent Cr species equivalent to surface oxygen
after the IB reaction for unsupported chromia, as compared
with the 10 CrAl catalyst. Higher coverage with oxygen
implies that the number of oxygen atoms in the vicinity of the
reacting hydrocarbon molecule is higher, which could account
reaction and consecutive isobutene transformations to CO
participate at higher conversions in the reduction step (a)].
2
The reaction rate would then fall as a result of the smaller
number of active Cr6`O species. On the other hand, at
2
3
x
higher conversions, catalyst reoxidation [step (b)] may
become the rate-determining step, which could also lead to a
decrease in the total activity.
Detailed kinetic studies of IB ODH, now in progress, are
necessary to distinguish between these two ways of accounting
for the fall in activity with IB conversion, observed for 10
CrAl and CrTi catalysts.
2
3
Acknowledgements
for the higher selectivity to CO observed for unsupported
chromia. It has been recognized that one of the prerequisites
2
The assistance of Mrs. K. Samson, M. Eng. and T. Bobinska,
M. Eng. in the preparation and chemical analysis of the cata-
lysts, Mrs. Z. Czu¡a in the BET determinations, and Mrs. I.
Gressel in the technical preparation of the typescript is grate-
fully acknowledged. The surface potential measurements were
performed by K. Wcis¡o, M. Eng. at the Laboratoire de
Catalyse URA 402 CNRS, University of Lille I, France; we
want to thank Prof. Yolande Barbaux for enabling us to make
these measurements and for the discussion. This research was
supported by the State Committee for ScientiÐc Research,
KBN under the project No 3 TO9A 128 09.
for a selective oxidation catalyst is a limited number of active
oxygen species surrounding an adsorbed hydrocarbon mol-
ecule: if this number is high, the molecule, subjected to simul-
taneous attack of several oxygens at di†erent carbon atoms,
undergoes total oxidation to carbon oxides.51
The lower reoxidation rate of the supported chromium
oxide catalysts as compared with pure chromia may be due to
the lower surface concentration of the Cr3` centres on which
adsorption of the oxygen molecule takes place. TPR, TPD
and allyl iodide probe reaction measurements11,12 have
shown that the Cr6`wO bond energy is higher for the sup-
ported samples than for unsupported chromia, most probably
due to the presence of Cr6`wOwAl and Cr6`wOwTi bonds
in the supported catalysts, stronger than those in
Cr6`wOwCr6` groups on the surface of Cr O .
References
1
2
M. P. McDaniel, Adv. Catal., 1985, 33, 48, and references therein.
S. De Rossi, G. Ferraris, S. Fremiotti, V. Indovina and
A. Cimino, Appl. Catal. A, 1993, 106, 125.
2
3
The vacuum pretreatment prior to the chemisorption mea-
surements will, then, lead to more extensive reduction of the
Cr6`wO groups, with formation of higher amounts of the
oxygen adsorption active centres, Cr3` in the case of unsup-
3
4
S. De Rossi, G. Ferraris, S. Fremiotti, E. Garrone, G. Ghiotti,
M. C. Campa and V. Indovina, J. Catal., 1994, 148, 36.
D. SanÐlippo, F. Buonomo, G. Fusco, M. Lupieri and I. Miracca,
Chem. Eng. Sci., 1992, 47, 2313.
338
Phys. Chem. Chem. Phys., 1999, 1, 333È339

View Article Online
5
6
S. Udomsak and R. G. Anthony, Ind. Eng. Chem. Res., 1996, 35,
47.
30 A. Rahman, M. M. Mohamad, M. Ahmed and A. M. Aitari, Appl.
Catal. A, 1995, 121, 234.
31 B. M. Weckhuysen, R. A. Schoonheydt, F. e. Mabbs and D. Col-
lison, J. Chem. Soc., Faraday T rans., 1996, 92, 2431.
32 C. Groeneveld, P. P. M. M. Wittgen, A. M. Van Kersbergen,
P. L. M. Mestrom, C. E. Nuijten and G. C. A. Schuit, J. Catal.,
1979, 59, 153.
33 K. Kohler, C. W. Schlapfer, A. von Zelewsky, J. Nickl, J. Eng-
weiler and A. Baiker, J. Catal., 1993, 143, 201.
34 V. Sharf, H. Schneider, A. Baiker and A. Wokaun, J. Catal., 1994,
145, 464.
35 P. J. M. Carrott and N. Sheppard, J. Chem. Soc., Faraday T rans.
1, 1983, 79, 2425.
F. Cavani, M. Koutyrev, F. TriÐro, A. Bartolini, D. Ghisletti,
R. Iezzi, A. Santucci and G. Del Piero, J. Catal., 1996, 158, 236.
T. P. Kobylinski and B. W. Taylor, J. Catal., 1973, 31, 450.
H. Bosch and F. Janssen, Catal. T oday, 1988, 2, 369.
H. E. Curry-Hyde and A. Baiker, Ind. Eng. Chem. Res., 1990, 29,
1985.
7
8
9
10 R. Grabowski, B. Grzybowska, J. S¡oczynski and K. Wcis¡o,
Appl. Catal., 1996, 144, 335.
11 B. Grzybowska, J. S¡oczynski, R. Grabowski, K. Wcis¡o,
A. Koz¡owska, J. Stoch and J. Zielinski, J. Catal., 1998, 178, 687.
12 B. Grzybowska, J. S¡oczynski, R. Grabowski, K. Wcis¡o,
A. Koz¡owska and J. Stoch, Polish J. Chem., 1998, 72, 2159.
13 D. A. Dowden and W. E. Garner, J. Chem. Soc., 1939, 893.
14 R. A. Beebe and D. A. Dowden, J. Am. Chem. Soc., 1938, 60,
2912.
36 F. D. Hardcastle and I. E. Wachs, J. Mol. Catal., 1988, 46, 173.
37 M. Schraml-Marth, A. Wokaun, M. E. Curry-Hyde and
A. Baiker, J. Catal., 1992, 133, 415.
38 B. M. Weckhuysen and I. E. Wachs, J. Chem. Soc., Faraday
T rans., 1996, 92, 1969.
15 D. S. MacIver and H. H. Tobin, J. Phys. Chem., 1960, 64, 451.
16 J. M. Bridges, D. S. MacIver and H. H. Tobin, in Proceedings of
the 2nd International Congress on Catalysis, 1961, vol. 2, p. 2161.
17 B. M. Trapnell, Chemisorption, Butterworths, London, 1955.
18 E. R. S. Winter, Adv. Catal., 1958, 10, 196.
19 S. W. Weller and S. E. Voltz, J. Am. Chem. Soc., 1954, 76, 4695,
4701.
20 J. Deren, J. Haber, A. Podgorecka and J. Burzyk, J. Catal., 1963,
39 H. Charcosset, A. Revillon and A. Guyot, J. Catal., 1967, 8, 326,
334.
40 Y. Barbaux, J. P. Bonnelle and Y. P. BeauÐls, J. Chem. Res.,
1979(S), 48, (M) 0650.
41 B. Grzybowska, Y. Barbaux and J. P. Bonnelle, J. Chem. Res.,
1981(S), 48, (M) 0555.
42 J. Deren, J. Haber and J. S¡oczynski, Chemia Analityczna, 1961, 6,
659.
2, 161.
21 J. Deren and J. Haber, in Studies on the Physico-chemical and
Surface Properties of Chromium Oxides, Polish Academy of
Sciences-Ceramics Commission, Polish ScientiÐc Publishers,
PWN Krakow, 1969.
43 D. Courcot, L. Gengembre, M. Guelton, Y. Barbaux and
B. Grzybowska, J. Chem. Soc., Faraday T rans., 1994, 90, 895.
44 V. Ponec, Z. Knor and S. Cerny, Adsorption on Solids, Butter-
worths, London, 1974.
45 S. L. Kiperman, Introduction to Kinetics of Heterogeneous Cata-
lytic Reactions (in Russian), Nauka, Moscow, 1964.
46 K. Otto and M. Shelef, J. Catal., 1969, 14, 226.
47 A. Bielanski and J. Haber, Oxygen in Catalysis, Marcel Dekker,
New York, 1991, and references therein.
48 J. S¡oczynski, R. Grabowski, K. Wcis¡o and B. Grzybowska-
Swierkosz, Polish J. Chem., 1997, 71, 1585.
49 B. Grzybowska and I. Gressel, in preparation.
50 P. Mars and D. W. van Krevelen, Chem. Eng. Sci., Sp. Suppl.,
1954, 3, 41.
22 A. Zecchina, S. Coluccia, L. Cerruti and E. Borello, J. Phys.
Chem., 1971, 75, 2783.
23 K. Erhardt, M. Richter, U. Roost and G. Ohlman, Appl. Catal.,
1985, 17, 23.
24 H. E. Curry-Hyde, H. Mush and A. Baiker, Appl. Catal., 1990, 65,
211.
25 W. Grunert, W. Sa†ert, R. Feldhaus and K. Anders, J. Catal.,
1986, 99, 149.
26 J. Engweiler, J. Nickl, A. Baiker, K. Kohler, C. W. Schlapfer and
A. Von Zelewsky, J. Catal., 1994, 145, 141, and references therein.
27 Ch. Fountzoula, H. K. Matralis, Ch. Papadopoulou, G. A. Voy-
atzis and Ch. Kordulis, J. Catal., 1997, 172, 391.
28 A. Cimino, B. A. De Angelis, A. Luchetti and G. Minelli, J.
Catal., 1976, 45, 316.
51 L. J. Callahan and R. K. Grasselli, AIChE, 1963, 9, 755.
52 M. Najbar, W. Wal and J. Chrz¸szcz, Stud. Surf. Sci. Catal., 1990,
55, 779.
29 W. Grunert, E. S. Shpiro, R. Feldhaus, K. Anders, G. V. Antoshin
and Kh. M. Minachev, J. Catal., 1986, 100, 138.
Paper 8/06136J
Phys. Chem. Chem. Phys., 1999, 1, 333È339
339