Catalytic dehydrogenation of isopentane with iridium catalysts

Article abstract of DOI:10.1002/anie.200704856

(Graph Presented) Activation by adding PPh₃: The catalytic dehydrogenation of isopentane to give isopentene and hydrogen with iridium catalysts ona silica gel support at 450°C proceeds with impressive conversion whenthe support is impregnated w

Full text of DOI:10.1002/anie.200704856

Angewandte
Chemie
DOI: 10.1002/anie.200704856
Heterogeneous Dehydrogenation
Catalytic Dehydrogenation of Isopentane with Iridium Catalysts**
Helmut G. Alt* and Ingrid K. Böhmer
Dedicated to Professor Wolfgang A. Herrmann on the occasion of his 60th birthday
The catalytic activation of alkanes under mild conditions to
give the corresponding alkenes and hydrogen[1] has tremen-
dous industrial implications because olefins are the most
important starting materials for large-scale industrial pro-
cesses such as olefin polymerization, dimerization, oligome-
rization, hydroformylation, and metathesis. In the search for a
process by which olefins can be produced from alkanes
without a steam cracker, iridium complexes have proved to be
the best catalysts so far.[2–5] An iridium catalyst with a pincer
ligand was reported to give the best performance,[6] but it had
the big disadvantage that an added external olefin was
required to eliminate the formed hydrogen from the equilib-
rium. The addition of such “sacrificial olefins” is not
economic and limits the commercial value of the process.
Figure 1. Correlation between the number of phosphine ligands and
the activity of the catalyst for the dehydrogenation of isopentane at
Catalysts that do not need a “sacrificial olefin” are rare,[7,8]
and typically they perform poorly in homogeneous solution.
We found that a series of iridium complexes, in combina-
tion with phosphorus-containing compounds on silica gel as
the support material, are able to activate alkanes in a fixed-
bed reactor to give the corresponding alkenes and hydrogen
with high selectivity and high activity. As a model compound
we chose isopentane, a refinery product with a high octane
number of 92, which cannot be added to gasoline because of
its low boiling point of 288C.
4008C.
We tested four different iridium complexes and found a
drastic dependence of the activity on the number of phos-
phine ligands (Figure 1). The activity of H2IrCl6 for isopen-
tane activation could be increased significantly when external
PPh3 was added to the reaction mixture. In addition, the
activity increased overproportionally at temperatures higher
than 3508C (see Figure 2). This behavior indicates the
formation of a new catalytic species at 3508C that is more
active than the original one.
Figure 2. Comparison of the activity of H2IrCl6 on silica gel without
PPh3 and with 4 equiv PPh3 dependingon the reaction temperature.
In another experiment we used the phosphine-free
catalyst bis(1,5-cyclooctadiene)iridium(I) tetrafluoroborate
on silica gel. In addition, catalysts with Ir/P ratios of 1:4 and
1:8 ratio were synthesized. Instead of an externally added
triphenylphosphine, the phosphine could also be integrated
into the catalyst through functionalized silica gel (see
Scheme 1). All of these catalysts were tested and compared
in C–H activation experiments (Figure 3). The conversion at
Scheme 1. Functionalization of the silica gel surface (also see the
Experimental Section).
[*] Prof. Dr. H. G. Alt, I. K. Böhmer
Laboratorium für Anorganische Chemie
Universität Bayreuth
4508C increased from 2.3% to 21.9% when a fourfold excess
of triphenylphosphine was added. The higher Ir/P ratio of 1:8
improved the conversion to 26.1%. A catalyst with phos-
phine-functionalized silica gel (Ir/P 1:4) also improved the
conversion of isopentane by threefold to 9.8% at 4508C.
The GC detection of reaction products containing phenyl
groups led to the assumption that the organometallic
Universitätsstrasse 30, 95440 Bayreuth (Germany)
Fax: (+49)921-55-2044
E-mail: helmut.alt@uni-bayreuth.de
[**] Financial support from ConocoPhillips, Bartlesville (USA), is
gratefully acknowledged.
Angew. Chem. Int. Ed. 2008, 47, 2619 –2621
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2619

Communications
If one compares the phosphorus content of the catalyst
systems before and after the reactions, a strong dependence
between iridium and phosphorus is apparent. This indicates
that the phenyl groups are lost but the phosphorus content
before and after the run is about the same. After the reaction
with an Ir/P ratio of 1:8, a slight loss of phosphorus is
apparent, but this can be expected owing to an insufficient
interaction of phosphorus with the metal.
To gain further insight, we conducted the dehydrogen-
ation of isopentane for 5 h in a fixed-bed reactor at 150, 250,
350, 450, and 5508C. After the runs, the carbon and
phosphorus contents of the samples were analyzed. A small
amount of PPh3 was lost at room temperature. It is
remarkable that the P/Ir ratio drops to a value of 3:1, which
remains constant above 2508C. At this temperature a drastic
loss of the phenyl groups takes place, which is evident from
the decreasing C/P ratio. This value drops below 18, which
would be expected for intact PPh3 groups. It is not possible to
use the carbon contents as absolute values owing to the
amount of cyclooctadiene in the catalyst, leads to a value
higher than 18 in the first samples. Above temperatures of
5008C coking becomes evident. The thermodynamic equilib-
rium of the dehydrogenation of isopentane can be calculated
with computer software (Aspen or HSC). Figure 5 shows two
Figure 3. Comparison of conversion rates of catalysts with and without
triphenylphosphine.
complexes decompose partly at elevated temperatures by
losing the phenyl groups of triphenylphosphine; similar
results were found previously for other iridium catalysts.[7]
To test this for the systems with externally added phosphines,
the carbon and phosphorus contents of three catalysts before
and after the reactions were analyzed (see Figure 4).
Figure 5. Calculation of the thermodynamic equilibrium of the dehy-
drogenation reaction of isopentane at different temperatures (light
gray: Aspen, dark gray: HSC Chemistry, black: experimental results).
Figure 4. Carbon content of dehydrogenation catalysts before (black)
and after the reaction (shaded).
The data are interesting in a number of aspects. First of all,
they show that the main carbon content, presumably due to
triphenylphosphine, disappears during the reaction. This
confirms the assumption of the loss of phenyl groups.
Apparently this loss does not affect the activity in a negative
way: it was detected by thermogravimetric analysis (TGA) at
temperatures below 4008C, whereas the maximum conver-
sion occurs at higher temperatures. This could mean that the
loss of the phenyl groups is required for the formation of a
more active species. Completely phosphine-free catalyst
systems had lower conversion rates. All investigated catalysts
maintained their high activities over days, indicating that the
eliminated phenyl groups from the triphenylphosphine
ligands or additives cannot play the role of hydrogen-
consuming (“sacrificial”) components. We also found that
the amount of coking on the active catalysts is negligibly low.
Therefore this should not be the reason for deactivation.
curves that are not identical. However, the experimental
results (29.9% yield at 4508C) fit the predictions very well.
The actual catalyst is probably an iridium phosphide or an
iridium–phosphorus cage or nanocluster. These conclusions
are based on comparative measurements between active and
less active catalysts before and after the dehydrogenation
reaction. Less active catalysts showed bigger conglomerates
of iridium metal that formed during the reaction and are
apparently not able to activate CH bonds. Since the dehy-
drogenation reaction is an endothermic equilibrium reaction,
it is advantageous to perform this as a continuous reaction in a
fixed-bed reactor at high temperatures and to separate the
products. The same reactions in a batch reactor (closed
system) gave much lower yields.
2620 www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 2619 –2621

Angewandte
Chemie
0.1 bar. The reaction temperature range was from 300 to 4508C,
Experimental Section
measured by a thermocouple positioned in the catalyst bed. The
products were collected in a sample collector cooled with ice. Samples
were removed hourly and the products identified by GC. Depending
on the temperature, an isomeric mixture of 2-methyl-2-butene, 2-
methyl-1-butene, and 3-methyl-1-butene was generated. To calculate
the activity, the sum of the three isomers was used. The activity for the
dehydrogenation reaction is denoted as conversion of isopentane to
isopentene in percent.
The catalysts were synthesized by supporting different metal com-
plexes on silica gel by the method of “incipient wetness”. This
impregnation method maximizes the dispersion on the surface. First,
the amount of liquid absorbable by the support was determined. For
example, toluene was added dropwise to one gram of silica (Davicat
SI 1102) until the liquid was visible around the particles. The required
amount of solvent could be calculated by the weight difference. A
metal complex(H IrCl6, [(cod)2Ir]+ BF4À, [(cod)(PCy3)Ir(py)]+ PF6
,
À
2
[(PPh3)2Ir(Cl)(CO)], or [(PPh3)3Ir(H)(CO)]) was dissolved in exactly
the amount of a suitable solvent (THF, toluene, CH2Cl2, water) that
was required to fill all pores of the support. The solution was added
dropwise to the support material and the supported catalyst was dried
under a nitrogen stream, by heating, and/or under high vacuum.
The metal content on the support was in a range of 0.3–1.0 wt%.
To examine the influence of triphenylphosphine on the activity of the
metal complexes, 4 or 8 equiv of triphenylphosphine was dissolved in
n-pentane and added to the supported catalyst by the technique of
“incipient wetness”. In experiments with [(cod)2Ir]+ BF4À the silica gel
support was functionalized with phosphine groups before use. The
silica gel surface can be modified by the reaction of its surface
hydroxy groups with bifunctional linker agents[9–12] (see Scheme 1).
For this reaction, a solution of 2-(diphenylphosphanyl)ethyltrieth-
oxysilane (188 mg) in pentane (7.2 g) was added dropwise to silica gel
(6.0 g). The impregnated silica gel was heated to 1208C for 3 h under
nitrogen. After the support had cooled to room temperature, a
solution of [(cod)2Ir]+ BF4À was added by the technique of “incipient
wetness”.
All synthesized catalysts were tested in a fixed-bed reactor for the
dehydrogenation of isopentane without additional activation steps
and without a “sacrificial olefin”. The catalysts were analyzed before
and after the C–H activation experiments by combustion analysis
(ConocoPhillips Inc., Bartlesville, USA) for the content of carbon,
hydrogen, and phosphorus.
For C–H activation experiments, a sample of the catalyst (5 g)
was placed in a stainless-steel reactor tube and held in place by layers
of glass wool and glass beads. The packed reactor was plumbed into a
pipe system inside of a heating unit. After the system had been purged
with nitrogen and heated to 3008C, the feed isopentane was pumped
up-flow by a syringe pump with a weight hourly space velocity
(WHSV) of 1.9 through the catalyst bed. The pressure was about
Received: October 19, 2007
Published online: February 27, 2008
À
Keywords: C H activation · dehydrogenation ·
heterogeneous catalysis · iridium · isopentane
.
[1] T. Graening, Nachr. Chem. 2007, 55, 836.
[2] “Activation of Unreactive Bonds and Organic Synthesis”: S.
Murai, Topics in Organometallic Chemistry, Vol. 3, Springer,
Berlin, 1999, and references therein.
[3] R. H. Crabtree, J. M. Mihelcic, J.-M. Quirk, J. Am. Chem. Soc.
1979, 101, 7738.
[4] A. H. Janowicz, R. G. Bergman, J. Am. Chem. Soc. 1983, 105,
3929.
[5] W.-W. Xu, G. P. Rosini, M. Gupta, C. M. Jensen, W. C. Kaska, K.
Krogh-Jespersen, A. S. Goldman, Chem. Commun. 1997, 2273.
[6] I. Göttker-Schnetmann, P. White, M. Brookhart, J. Am. Chem.
Soc. 2004, 126, 1804.
[7] T. Aoki, R. H. Crabtree, Organometallics 1993, 12, 294.
[8] F. Liu, A. Goldman, Chem. Commun. 1999, 655.
[9] N. J. Meehan, A. J. Sandee, J. N. H. Reek, P. C. J. Kamer,
P. W. N. M. van Leeuwen, M. Poliakoff, Chem. Commun. 2000,
1497.
[10] H. Yang, H. Gao, R. J. Angelici, Organometallics 2000, 19, 622.
[11] Y. Wang, T. J. Su, R. Green, Y. Tang, D. Styrkas, T. N. Danks, R.
Bolton, J. R. Lu, Chem. Commun. 2000, 587.
[12] C. Merckle, S. Haubrich, J. Blꢀmel, J. Organomet. Chem. 2001,
627, 44.
Angew. Chem. Int. Ed. 2008, 47, 2619 –2621
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
2621