In situ UV-VIS diffuse reflectance spectroscopy-on-line activity measurements: Significance of Crn+ species (n = 2, 3 and 6) in n-butane dehydrogenation catalyzed by supported chromium oxide catalysts

Article abstract of DOI:10.1039/a801710g

The dehydrogenation of n-butane over supported chromium oxide catalysts has been investigated at 500°C by in situ UV-VIS diffuse reflectance spectroscopy-on-line activity measurements as a function of the reaction time (0-60 min), the support composition (SiO₂:Al₂O₃ ratio) and the Cr loading (0-8 wt.%). The catalytic reaction is characterized by an initial induction period, during which the supported chromium oxides are reduced to Cr^3+/2+, and CO/CO₂ and cracking products (C₁, C₂ and C₃) are formed. After this initial induction period, but-1-ene and but-2-ene are produced with a selectivity up to 90%. The catalytic activity increases with increasing Cr content and Al₂O₃ content of the support, and is maximum after 5-15 min on stream. A relation between the dehydrogenation activity and the amount of Cr³⁺ formed upon the reduction of Cr⁶⁺ has been established.

Full text of DOI:10.1039/a801710g

View Article Online / Journal Homepage / Table of Contents for this issue
In situ UV–VIS di†use reÑectance spectroscopy–on-line activity
measurements
SigniÐcance of Crn‘ species (n = 2, 3 and 6) in n-butane dehydrogenation
catalyzed by supported chromium oxide catalysts
Bert M. Weckhuysen,* Abdelhamid Bensalem and Robert A. Schoonheydt
Centrum voor Oppervlaktechemie en Katalyse, Departement Interfasechemie, Katholieke
Universiteit L euven, Kardinaal Mercierlaan 92, 3001 Heverlee, Belgium
The dehydrogenation of n-butane over supported chromium oxide catalysts has been investigated at 500 ¡C by in situ UVÈVIS
di†use reÑectance spectroscopyÈon-line activity measurements as a function of the reaction time (0È60 min), the support composi-
tion (SiO : Al O ratio) and the Cr loading (0È8 wt.%). The catalytic reaction is characterized by an initial induction period,
2
2 3
during which the supported chromium oxides are reduced to Cr3`@2`, and CO/CO and cracking products (C , C and C ) are
2
1
2
3
formed. After this initial induction period, but-1-ene and but-2-ene are produced with a selectivity up to 90%. The catalytic
activity increases with increasing Cr content and Al O content of the support, and is maximum after 5È15 min on stream. A
2
3
relation between the dehydrogenation activity and the amount of Cr3` formed upon the reduction of Cr6` has been established.
reactions. The results will be discussed in relation to recent
publications on the nature of the active dehydrogenation site.
Introduction
Supported chromium oxide catalysts are crucial in chemical
industries for the production of important commodity chemi-
cals. Indeed, several large volume industrial processes rely on
Experimental
supported chromium oxides as a heterogeneous catalyst.
Chromium-on-silica is the famous Philips polymerization
Catalyst preparation
catalyst, responsible for about one third of the world pro-
duction of high density polyethylene (HDPE).1,2 This system
is still developing, and forms the basis of a new generation of
supported mono-alkene polymerization catalysts, as an alter-
native to the recently commercialized metallocenes. Similarly,
chromium-on-alumina is used as a light paraffin dehydroge-
nation catalyst employed in the Houdry CATOFIN and
CATADIENE processes.3,4 The Houdry CATOFIN process
is responsible for approximately 50% of the catalytically
derived world production of isobutane, for methyl tert-butyl
ether (MTBE), and 50% of the catalytically generated propyl-
ene, mostly used for the production of polypropylene.
The industrial importance of supported chromium oxide
catalysts has attracted a great deal of interest in industrial and
academic research towards the elucidation of the molecular
structure of chromium oxides and their relation with catalytic
performances.5h12 The state-of-the-art of these catalysts has
been recently summarized in a review paper.5 It has been
shown that almost all previous characterization studies were
conducted under controlled atmospheres, i.e., hydrated, dehy-
drated and reduced conditions, and the data obtained could
not be directly related to catalytic data. Therefore, in general,
spectroscopic techniques have to be developed which allow
measurements under conditions as close as possible to “realÏ
catalytic conditions. Thus, in situ spectroscopic studies, com-
bined with on-line activity/selectivity measurements, are
required to fully understand the catalytic action of supported
metal oxide catalysts.
Preparation of supports. SiO was prepared by mixing H O
2
2
at pH 2 (adding HCl) and tetraethyl orthosilicate (TEOS) (2/1,
v/v) during 5 h at room temperature. The mixture was titrated
under stirring to pH 6 with an NH OH solution at pH 9.5.
4
After 16 h of gelation, the gel was dried at 130 ¡C for 72 h and
calcined at 250 and 550 ¡C for 3 and 16 h, respectively. The
so-obtained cake was crushed. SiO Æ Al O with 20, 40 and 60
2
2 3
wt.% SiO was prepared following a modiÐed method of
2
Chen et al.13. Appropriate amounts of TEOS and aluminium
triisopropoxide were mixed in 128 ml of ethanol during 30
min at room temperature. After adding 35 ml of 1 M HCl, the
hydrolysis started and the solution was mixed for 1 h. The gel
was dried at 60 and 100 ¡C for 8 h, and calcined at 550 ¡C for
16 h. The samples of SiO Æ Al O were then crushed. Al O
2
2 3
2 3
was prepared by drying an ethanol solution of aluminium tri-
isopropoxide at 60 and 100 ¡C for 8 h. The dried solid was
calcined further at 550 ¡C for 16 h. The supports are denoted
as follows: SA-0 (Al O ), SA-20 (SiO Æ Al O with 20 wt.%
2
3
2
2 3
SiO ), SA-40 (SiO Æ Al O with 40 wt.% SiO ), SA-60
2 3
2
2
2
(SiO Æ Al O with 60 wt.% SiO ) and SA-100 (SiO ).
2
2 3
2
2
Preparation of catalysts. The supported chromium oxide
catalysts were prepared by the incipient-wetness method with
chromium(VI) oxide (CrO ). The catalysts were dried at 50 ¡C
3
for 8 h and granulated. The size fraction between 0.2 and 0.5
mm was used for in situ DRS measurements. The following
catalysts were prepared: 0.2 wt.% Cr/SA-0, 0.2 wt.% Cr/SA-
20, 0.2 wt.% Cr/SA-40, 0.2 wt.% Cr/SA-60, 0.2 wt.% Cr/SA-
100, 1.0 wt.% Cr/SA-0, 2.0 wt.% Cr/SA-0, 4.0 wt.% Cr/SA-0
and 8.0 wt.% Cr/SA-0.
In this work, we have investigated supported chromium
oxide catalysts in n-butane dehydrogenation reactions by
using in situ UVÈVIS di†use reÑectance spectroscopy. This
allows us to probe Cr6`, Cr3` and Cr2` in reaction condi-
tions, while the reaction products (CO/CO , CH , C H ,
In situ on-line GC analysis during di†use reÑectance
spectroscopy
2
4
2 6
C H , C H , C H , . . . ) can be monitored by on-line gas
In situ di†use reÑectance spectra were taken on a Varian Cary
5 UVÈVISÈNIR spectrophotometer equipped with a specially
designed Praying Mantis di†use reÑection attachment (DRA)
2
4
3 8
4 10
chromatography. The goal of this study is then to address the
role of Crn` (with n \ 6, 3 and 2) in alkane dehydrogenation
J. Chem. Soc., Faraday T rans., 1998, 94(14), 2011È2014
2011

View Article Online
of Harrick. A schematic drawing and the technical details of
The catalytic reaction is characterized by an initial induction
this cell are given in recent papers.14,15 The white reÑectance
period, during which CO/CO and some cracking products
are formed. The maximum selectivity towards these products
was reached after 3 min on-stream, and was ca. 50 and 35%
2
standard BaSO (Kodak) was used to take a baseline at 25 ¡C
4
in the Praying Mantis cell. The thickness of the catalyst bed
was 3 mm, and the amount of the catalyst used was ca. 40È65
mg. The catalyst was Ðrst treated in oxygen in the in situ cell
at 550 ¡C for 1 h, and then cooled in He to 500 ¡C. The cata-
lyst was then subjected to an 18% n-butane stream in He (50
ml min~1) at 1 atm. The measured in situ DRS spectra were
processed with Grams/386 (Galactic Industries Corp). On-line
gas chromatography analysis was done with an HP 5830 GC
instrument equipped with an FID detector, six-way valve inlet
system, and a packed column Ðlled with n-octane/porasil-C
for CO/CO and the cracking products, respectively. The
2
cracking products formed were CH , C H , C H , C H and
4
2 4 2 6 3 8
C H . The selectivity towards these products was highest
3
10
(ca. 50%) for 0.2 wt.% Cr/SA-40 and Cr/SA-60 catalysts. This
must be due to the acidity of the silicaÈalumina supports.
After this initial induction period, but-1-ene and but-2-ene
were produced selectively. The selectivity towards alkenes
reached a maximum of ca. 90% after 10 min, and decreased
slightly with increasing time on-stream.
(
100È120 mesh, Durapack, Applied Science Laboratories). The
thermodynamically expected conversion at 500 ¡C will be ca.
0%. However, as will be shown below, the conversions
Fig. 2 compares the catalytic activity of 0.2 wt.% Cr cata-
lysts as a function of the support composition (SiO : Al O
2
2 3
2
ratio). The dehydrogenation activities were measured after 10
min on-stream. It is clear that the catalytic activity decreased
from 0.44% for a Cr/SA-0 catalyst, through 0.35% for a Cr/
SA-40 to 0.18% for a Cr/SA-100 catalyst. Thus, the dehydro-
obtained in the in situ cell are much lower. Indeed, the best
catalyst employed in this study has a conversion below 10%.
Thus, the in situ DRS cell cannot be considered as an ideal
Ðxed bed reactor, and gas by-pass and eventually an inhomo-
geneous temperature proÐle have to be invoked to explain the
lower conversions.
genation activity decreased with increasing SiO content of
2
the support. The dependency of the dehydrogenation activity
on Cr loading is illustrated in Fig. 3 for Cr/Al O catalysts. It
2
3
is shown that the conversion towards alkenes increased
almost linear with increasing Cr loading.
Results
Catalytic performance
SigniÐcance of Crn‘ (with n = 6, 3 and 2) in n-butane
dehydrogenation
The dehydrogenation of n-butane over supported chromium
oxide catalysts has been investigated at 500 ¡C in the in situ
cell as a function of the support composition, the Cr loading,
and the reaction time. As an example, we will discuss the cata-
lytic performance of the 0.2 wt.% Cr/SA-0 catalyst in detail
In situ UVÈVIS di†use reÑectance spectra were monitored
between 300 and 800 nm as a function of time on-stream for
(
Fig. 1). Other Cr catalysts had a similar activity and selec-
tivity pattern, although their exact conversions/selectivities
were di†erent. It is also important to notice that the catalytic
activity of the empty in situ cell at 500 ¡C was below 0.10%,
while the selectivities towards but-1-ene and but-2-ene were
ca. 50%.
Fig. 1A shows that the catalytic activity of the 0.2 wt.%
Cr/SA-0 catalyst reached a maximum after 6 min on-stream,
and decreased with increasing reaction time. In addition, the
selectivity pattern of the reaction was quite complex (Fig. 1B).
Fig. 2 A, Conversion (%) and, B, Cr2` : Cr3` ratio as a function of
the support composition for n-butane dehydrogenation at 500 ¡C over
0
.2 wt.% Cr catalysts. The Cr2` : Cr3` ratio was derived from pre-
vious CO reduction experiments.5,17 The DRS spectra obtained after
CO reduction of the supported Cr catalysts at 600 ¡C for 30 min were
deconvoluted in a set of Gaussian bands typically for Cr3` and Cr2`.
The ratio of the absorption maxima was used to determine the
Cr2` : Cr3` ratio.
Fig. 1 A, Conversion (%) vs. time on-stream (min) for n-butane
dehydrogenation at 500 ¡C over a 0.2 wt.% Cr/SA-0 catalyst in the in
situ UVÈVIS DRS cell. B, Selectivity (%) vs. time on-stream (min) for
n-butane dehydrogenation at 500 ¡C over a 0.2 wt.% Cr/SA-0 catalyst
in the in situ UVÈVIS DRS cell: (a) CO/CO ; (b) cracking products
Fig. 3 Conversion (%) vs. Cr loading (wt.%) for n-butane dehydrogen-
ation at 500 ¡C over Cr/SA-0 catalysts
2
(
C , C and C ) and (c) but-1-ene and but-2-ene.
1
2
3
2012
J. Chem. Soc., Faraday T rans., 1998, V ol. 94

View Article Online
each supported chromium oxide catalyst treated in n-butane
at 500 ¡C. The scan time was 1 min, and 60 di†erent spectra
were recorded during 60 min. An example of a set of in situ
di†use reÑectance spectra is given in Fig. 4 for the 0.2 wt.%
Cr/SA-0 catalyst. Fig. 4 shows a gradual decrease of the
absorption maximum at ca. 360 nm with increasing reaction
time at the expense of a new weak band with an absorption
maximum at ca. 590 nm. Fig. 4 also shows the presence of an
isobestic point at ca. 490 nm, suggesting the presence of two
di†erent Cr species. The Ðrst absorption band is typical for the
presence of Cr6`, whereas the broad absorption band at 590
nm is indicative for the presence of Cr3`.5,16,17 In the case of
Cr/SA-100 catalysts, the absorption maximum of reduced Cr
is at ca. 625 nm, which is typical for the formation of Cr2`,
although some Cr3` may contribute to the spectra as well.
Indeed, the absorption maximum at 590 nm shifts to higher
energy with decreasing Al O content of the support, which is
Fig. 5 Intensity proÐles of the absorption maxima at 360 and 590
nm for a 0.2 wt.% Cr/SA-0 catalyst treated at 500 ¡C in 18% n-butane
in N as a function of time on-stream
2
2
3
linearly with increasing Cr loading up to 2 wt.% Cr (Fig. 6).
The deviation from linearity at higher Cr loadings must be
due to the inherent limitations of the KubelkaÈMunk theory
indicative of an increase of the Cr2` : Cr3` ratio with decreas-
ing Al O content of the support. Thus, the di†erences in Cr
2
3
speciation are related to the support composition, i.e., mainly
[
eqn. (1)].15
Cr3` on Al O , a mixture of Cr3` and Cr2` on SiO Æ Al O
2
3
2
2 3
and, mainly Cr2` on SiO .5,17
2
By monitoring the intensity of the absorption bands at ca.
90È625 nm and 360 nm as a function of time on-stream, one
Discussion
5
The goal of this work was to address the role of Cr6`, Cr3`
and Cr2` in alkane dehydrogenation reactions by using in situ
on-line GC analysis during UVÈVIS DRS spectroscopy. First
of all, it is important to stress that, although the conversions
obtained in the in situ cell are always low, substantial di†er-
ences in catalytic performance were obtained between the dif-
ferent Cr catalysts. This suggests that this method will be
useful for characterizing other supported metal oxide catalysts
in catalytic action. Furthermore, our experiments clearly show
that Cr6` was instantly reduced to Cr3`@2` after exposing the
catalyst to a stream of n-butane. Thus, the initially low selec-
tivity towards alkenes must be explained by the cracking and
oxidation properties of high-valence chromium species. After
5È10 min on-stream, all Cr6` was reduced to Cr3`@2`, and
the catalytic activity reached a maximum. The selectivity
towards but-1-ene and but-2-ene, on the other hand, was ca.
90%, and similar values are obtained for Cr/Al O
can measure the amount of Cr6` and Cr3`@2` during cata-
lytic action, according to eqn. (1):15
1 [ R )2
K
S
I
(Crn`) \ (
=
\
\ kC(Crn`)
(1)
KM
2R
=
with I (Crn`), the KubelkaÈMunk intensity; R , the di†use
KM
=
reÑectance of the Cr catalyst; K and S, the absorption and
scattering coefficients; k, a proportionality coefficient and
C(Crn`), the amount of Crn`. However, it is important to
notice here that previous work5,17 has shown that eqn. (1) is
only valid for low loaded Cr catalysts and, as a consequence,
the I (Crn`) values will deviate from linearity at higher Cr
KM
loadings. As an example, the intensity proÐles of the 360 and
590 nm band for the 0.2 wt.% Cr/SA-0 catalyst treated in n-
butane are given in Fig. 5, and similar plots were obtained for
other Cr catalysts. This Ðgure shows a gradual decrease of the
intensity of the 360 nm band with increasing time on-stream.
It is also clear that almost all Cr6` was reduced to Cr3` after
3 catalysts
2
in industrial practice.3,4
Another comment concerns the oxidation state of the active
Cr species for alkane dehydrogenation, which has been the
subject of debate and controversy for many years in the liter-
ature.3 According to several authors the active species is Cr3`,
while other authors propose both Cr2` and Cr3`, or solely
Cr2` as the active dehydrogenation site.6h12 The almost
linear relationship between the catalytic activity and the
amount of Cr3` formed upon Cr6` reduction, as measured by
in situ DRS spectroscopy, indicates that Cr3` is the active
species for n-butane dehydrogenation (Fig. 6). Further evi-
dence comes from Fig. 2, which shows the relationship
between the Cr2` : Cr3` ratio on the catalyst surface (as
determined by DRS after CO reduction at 600 ¡C5,17) and the
10 min on-stream. Indeed, the amount of Cr3` formed upon
the reduction of Cr6` gradually increased with reaction time,
and reached a maximum after 10È20 min on-stream. It is also
important to notice that there was still a small increase in the
absorption intensities with reaction time after the reduction of
surface Cr6` took place. This is most probably due to the
formation of coke on the catalyst surface, which is responsible
for the deactivation of the catalyst.
By comparing the amount of Cr3` formed (upon the
reduction of Cr6`) in Cr/SA-0 catalysts di†ering in their Cr
content, we noticed that the amount of Cr3` increased almost
Fig. 4 In situ UVÈVIS di†use reÑectance spectra of a 0.2 wt.%
Cr/SA-0 catalyst treated at 500 ¡C in 18% n-butane in N
tion of time on-stream
2 as a func-
Fig. 6 Amount of Cr3` formed upon Cr6` reduction vs. conversion
(%) for Cr/SA-0 catalysts di†ering in their Cr content
J. Chem. Soc., Faraday T rans., 1998, V ol. 94
2013

View Article Online
catalytic activity. It is clear that the catalytic activity increases
with decreasing Cr2` : Cr3` ratio. However, it is also clear
that our experiments do not allow us to rule out the possible
activity of Cr2` sites. Indeed, an alternative explanation could
be that Cr3` species are more active than Cr2` species. In
addition, the dÈd absorption band is always very broad, and
di†erent Cr2`@3` species may be involved in catalysis.
dehydrogenation activity of Cr2` species cannot be ruled out,
it seems that Cr3` was more active than Cr2`.
B. M. W. is a postdoctoral fellow of the Fonds voor Wetens-
chappelijk Onderzoek [ Vlaanderen (FWO). A. B. acknow-
ledges a junior fellowship from the Onderzoeksraad of the K.
U. Leuven. This work was Ðnancially supported by the
Geconcerteerde Onderzoeksactie (GOA) of the Flemish Gov-
ernment and by the FWO.
Further evidence for the signiÐcance of Cr3` species
(
formed upon Cr6` reduction) in alkane dehydrogenation
reactions can be found in the literature.6,10,12 Masson et al.12
and Hakuli et al.6 found that high-valence chromium species
were the precursors of the active Cr3` sites of the catalyst
because (1) the amount of Cr6` decreased in parallel with the
initial activity, and (2) Cr3` was the only species detected by
XPS measurements in a reduced Cr/Al O catalyst. In addi-
References
1
B. M. Weckhuysen and R. A. Schoonheydt, Catal. T oday, in
press, and references therein.
2
3
M. P. McDaniel, Adv. Catal., 1985, 33, 47, and references therein.
B. M. Weckhuysen and R. A. Schoonheydt, Catal. T oday, 1998 in
press, and references therein.
2
3
tion, De Rossi et al.10 found that the dehydrogenation activity
of Cr/Al O and Cr/ZrO catalysts was due to coordinatively
4
F. Buonomo, D. SanÐlippo and F. TriÐro, in Handbook of Heter-
ogeneous Catalysis, ed. G. Ertl, H. KnoŽ zinger, J. Weitkamp,
2
3
2
unsaturated Cr3` sites. In view of this work, it is important to
underline that our analysis (1) is not able to give a more
detailed picture of the coordination geometry and poly-
merization degree of the active Cr3` site, and (2) does not
exclude the possibility that the active Cr3` species could also
arise from reduction of Cr5`. Indeed, De Rossi et al.10 have
concluded from EPR measurements that Cr5`-species are the
precursors of the active Cr3` species, and in situ EPR mea-
surements are under way in our laboratory to substantiate
this point.
WileyÈVCH, Weinheim, 1997, p. 2140, and references therein.
B. M. Weckhuysen, I. E. Wachs and R. A. Schoonheydt, Chem.
Rev., 1996, 96, 3327, and references therein.
5
6
7
A. Hakuli, A. Kytokivi, A. O. Krause and T. Suntola, J. Catal.,
1
996, 161, 393.
F. Cavani, M. Koutyrev, F. TriÐro, A. Bartolini, D. Ghisletti, R.
Iezzi, A. Santucci and G. Del Piero, J. Catal., 1996, 158, 236.
H. J. Lugo and J. H. Lunsford, J. Catal., 1985, 91, 155.
A. Kytokovi, J. P. Jacobs, A. Hakuli, J. Merilainen and H. H.
Brongersma, J. Catal., 1996, 162, 190.
8
9
1
0
S. De Rossi, G. Ferraris, S. Fremiotti, E. Garrone, G. Ghiotti, M.
C. Campa and V. Indovina, J. Catal., 1994, 148, 36; S. De Rossi,
G. Ferraris, S. Fremiotti, V. Indovina and A. Cimino, Appl.
Catal., 1993, 106, 125.
Conclusions
The following conclusions can be made.
11 B. M. Weckhuysen and I. E. Wachs, J. Phys. Chem., 1996, 100,
1
4 437; B. M. Weckhuysen and I. E. Wachs, J. Phys. Chem. B,
997, 101, 2793.
(
1) In situ on-line GC analysis during UVÈVIS DRS spec-
1
troscopy is a useful technique for studying supported metal
oxide catalysts in catalytic action.
1
2
J. Masson, J. M. Bonnier, P. H. Duvigneaud and B. Delmon, J.
Chem. Soc., Faraday T rans. 1, 1977, 73, 1471.
(
2) The dehydrogenation activity of supported chromium
13 K. C. Chen, T. Tsuchiya and J. D. Mackenzie, J. Non-Cryst.
Solids, 1986, 81, 227.
oxide catalysts increased with increasing Cr content, and
Al O content of the support, and a maximal conversion was
14 A. Bensalem, B. M. Weckhuysen and R. A. Schoonheydt, J. Phys.
Chem. B, 1997, 101, 2824.
2
3
reached after 5È15 min on-stream. The selectivity towards
but-1-ene and but-2-ene was ca. 90%, and deactivation of the
catalysts was due to coke formation.
1
5
B. M. Weckhuysen and R. A. Schoonheydt, Catal. T oday, sub-
mitted.
1
6
B. M. Weckhuysen, A. A. Verberckmoes, A. R. De Baets and R.
A. Schoonheydt, J. Catal., 1997, 166, 160.
(
3) The initial reduction of Cr6` after exposing the catalyst
to n-butane was responsible for the formation of CO/CO and
cracking products (C , C and C ).
17 B. M. Weckhuysen, L. M. De Ridder and R. A. Schoonheydt, J.
Phys. Chem., 1993, 97, 4756.
2
1
2
3
(
4) The dehydrogenation activity was proportional to the
amount of Cr3` formed upon Cr6` reduction. Although the
Paper 8/01710G; Received 2nd March, 1998
2014
J. Chem. Soc., Faraday T rans., 1998, V ol. 94