Rhodium-phosphite SILP catalysis for the highly selective hydroformylation of mixed C4 feedstocks.
Rhodium–Phosphite SILP Catalysis for the Highly Selective
Hydroformylation of Mixed C4Feedstocks**
Michael Jakuttis, Andreas Sch?nweiz, Sebastian Werner, Robert Franke, Klaus-Diether Wiese,
Marco Haumann,* and Peter Wasserscheid*
The hydroformylation of alkenes catalyzed by dissolved
rhodium complexes is not only one of the largest applications
of homogeneous catalysis in industry,but also an established
benchmark reaction for testing immobilization concepts for
homogeneous catalysts.In recent years, ionic liquids (ILs) as
non-aqueous solvents for liquid–liquid biphasic hydroformy-
lation catalysis have been the subject of intensive study.
Important features of ILs compared to the industrial aque-
ous–organic biphasic catalysis (Ruhrchemie–Rh?ne–Poulenc
process), are their much better solubility for higher
alkenesand their compatibility with phosphite ligands,
which readily decompose by hydrolysis in water.
Despite these attractive features, we know of no large-
scale industrial application of ionic liquids in biphasic hydro-
formylation catalysis to date. Two important drawbacks of the
biphasic ionic liquid systems are the relatively high amounts
of expensive IL that are required and its intrinsically high
viscosity, which leads to slow mass transport between the two
To overcome these limitations, we, among others,have
in recent years developed the supported ionic liquid phase
(SILP) concept. SILP materials are prepared by dispersing a
solution of the catalyst complex in an ionic liquid as a thin,
physisorbed film on the large internal surface area of a porous
solid material.Since the film thickness of the ionic liquid is
within the nanometer range, diffusion problems are mini-
mized by the extremely small diffusion distances. Excellent
ionic liquid utilization is achieved; that is, the same catalytic
performance can be achieved with a much smaller total IL
amount compared to liquid–liquid biphasic systems. Because
ionic liquids typically have extremely low vapor pressures,
catalysis with SILP materials is particularly attractive in
continuous gas-phase contact. During catalysis the immobi-
lized catalytic ionic liquid film comes into contact solely with
gaseous reactants and products. For the continuous gas-phase
hydroformylation of pure 1-alkene feedstock, such as, pro-
pene and 1-butene, this concept has been demonstrated quite
successfully with good catalytic activity (turnover frequencies
(TOFs) up to 500 h?1in the case of propeneand 564 h?1in
the case of 1-butene) and excellent catalyst stability (up to
200 h time-on-stream in the case of propeneand 120 h in
the case of 1-butene) as was demonstrated using a Rh-SILP
catalyst modified with the sulfonated phosphine ligand
sulfoxantphos (1). The sulfoxantphos–rhodium catalyst is,
however, unable to react with internal alkenes such as 2-
butenes in either hydroformylation or isomerization. Thus, to
convert 1-butene and 2-butenes from a mixed technical C4
feedstock from steam-cracker into the desired linear penta-
nal, a different catalyst system is required. Rhodium–phos-
phite complexes are known to be capable of selective
internal alkenes in a classical monophase homogeneous
catalysis into linear aldehydes with good to excellent selec-
tivity.Most of these ligands, however, are highly air- and
moisture-sensitive, making it difficult to handle and use them
in large quantities and a real challenge to recycle rhodium–
Herein, we show how the new diphosphite ligand 2 in
form of a SILP catalyst system is applied in the continuous
gas-phase hydroformylation of an industrial mixed C4feed-
stock as illustrated in Scheme 1. Synthesizing 2 and using it in
Scheme 1. Left: sulfoxantphos (1) and the novel diphosphite (2)
ligands. Right: possible reaction pathways for hydroformylation of
mixed C4feed to yield predominantly n-pentanal.
[*] M. Jakuttis, A. Sch?nweiz, S. Werner, Prof. Dr. P. Wasserscheid
Lehrstuhl f?r Chemische Reaktionstechnik
Egerlandstrasse 3, 91058 Erlangen (Germany)
Dr. M. Haumann
Chemical Reaction Engineering
Friedrich-Alexander-University Campus Busan
1276 Jisa-Dong, Gangseo-Gu, Busan 618-230 (Republic of Korea)
Dr. R. Franke, Dr. K.-D. Wiese
Evonik Oxeno GmbH
Paul-Baumann-Strasse 1, 45772 Marl (Germany)
[**] A.S., S.W., M.H., and P.W. thank the German Science Foundation
for infrastructural support through the Erlangen Cluster of Excel-
lence “Engineering of Advanced Materials”. SILP=Supported ionic
Supporting information for this article is available on the WWW
? 2011 Wiley-VCH Verlag GmbH & Co. KGaA, WeinheimAngew. Chem. Int. Ed. 2011, 50, 4492–4495
homogeneous-catalyzed hydroformylation in organic solvents
has been recently described by B?rner?s group.
The industrial feedstock (raffinate 1 by Evonik Oxeno
GmbH) used containedisobutene
(25.6%), trans-2-butene (9.1%), cis-2-butene (7.0%), non-
reactive butanes (14.9%), and 1,3-butadiene (0.3%). To
selectively obtain the desired linear product n-pentanal,
hydroformylation of isobutene must be avoided and both 2-
butenes have to be isomerized prior to hydroformylation.
We modified a procedure in the literature to prepare the
Rh-2-SILP catalyst (see Experimental Section). 2 g of the
catalyst were placed in a fixed bed reactor. A gas mixing unit
supplied the gaseous reactant stream (the industrial C4feed/
syngas (H2/CO)). Online GC was used to analyze all reaction
products. The continuous reactor setup has been previously
describedand further details can be found in the Support-
In a first set of experiments, the Rh-2-SILP catalyst was
tested at different temperatures while the conversion and
selectivity were measured over time as shown in Figure 1.
At temperatures of 80, 90, and 1008 8C, the system showed
an exceptionally high regioselectivity toward n-pentanal of up
to 99.5%. Despite the high isobutene content of 43.1% in the
feedstock used, we observed no formation of 3-methyl butan-
1-al, and the 2-butenes were hydroformylated only after
isomerization, finally producing n-pentanal.
The Rh-2-SILP catalyst exhibited good activity (20%
conversion corresponds in these experiments to a TOF of
330 h?1.) but low catalyst stability. After an initial activation
phase, the catalyst deactivated rapidly with increasing tem-
perature. The increasing activity in the beginning of the
experiment is a typical sign of slow ligand decomposition over
time as a decreasing L/Rh ratio results in more active
hydroformylation catalysts.Once all the ligand has decom-
posed (note that the L/Rh ratio in the IL film was 10:1 in the
freshly prepared SILP materials), the homogeneously dis-
solved nature of Rh cannot be maintained and inactive Rh
clusters form under the reaction conditions.Our experi-
ments indicate that this process is temperature dependent and
results in an irreversible loss of activity.
Given the notorious susceptibility to hydrolysis of the
phosphite ligand it was clear that its decomposition is
attributable to traces of water present in the gaseous feed-
stock mix. For this reason, dried raffinate 1 from Evonik
Oxeno was used (water content less than 16 ppm) in
subsequent continuous hydroformylation experiments. As
ligand decomposition leads to the formation of phosphoric
acid, the latter being a catalyst for further ligand decom-
position, we decided to add an acid scavenger to the ionic
liquid film in addition to all the other measures to eliminate
traces of water in the system. Bis(2,2,6,6-tetramethyl-4-
piperidyl)sebacate (3; BTPS, from Evonik Oxeno GmbH)
was found to be a suitable acid scavenger that does not
interact or react with the active catalytic species. When BTPS
was added in a molar ratio 3/2=4:1 to the ionic liquid catalyst
solution, the stability of the system could be improved
significantly. Figure 2 shows an example of a long-term
catalytic experiment at 1008 8C applying these optimized
conditions for ligand preservation.
Remarkably, with these modifications the initial conver-
sion of 25% could be retained for more than 800 h on stream
without loss of the high selectivity. The average activity of the
Rh-2-SILP system was 410 h?1which resulted in an accumu-
lated turnover number (TON) of more than 350000. The
success of this stability test confirms that the phosphite ligand
2 can indeed be applied as a SILP catalyst through a
combination of drying procedures and acid scavenging with-
out adversely affecting the ligand or complex stability, thus
achieving a highly selective and active SILP catalyst system.
Figure 1. Hydroformylation of an industrial C4mixture (raffinate 1,
500 ppm H2O) in the presence of Rh-2-SILP catalyst. ptotal=10 bar,
praffinate1=2 bar, pH2=pCO=4 bar. Total volume flow=13.8 mLmin?1,
residence time=29 s, mSILP=3 g, wRh=0.2 wt%, 2/Rh=10:1, ionic
liquid loading=10 vol% [EMIM][NTf2] relative to total pore volume.
Conversion (&,~, ^) and selectivity (&, *, ^) plotted over time at
80 (&, &), 90 (~, *), and 1008 8C (^, ^). EMIM=1-ethyl-3-methylimida-
Figure 2. Hydroformylation of an industrial C4mixture (raffinate 1, less
than 16 ppm H2O) in the presence of Rh-2-SILP catalyst. T=1008 8C,
ptotal=10 bar, praffinate1=2 bar, pH2=pCO=4 bar. Total volume flow=29.2
mLmin?1, residence time=15 s, mSILP=3 g, wRh=0.2 wt%, 2/
Rh=10:1, 3/2=4:1, ionic liquid loading=10 vol% [EMIM][NTf2] rela-
tive to total pore volume. Conversion (&) and selectivity (&) plotted
Angew. Chem. Int. Ed. 2011, 50, 4492–4495? 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Kinetic investigations revealed a first order dependence
on substrate partial pressure in the temperature range
between 70 and 1008 8C. Varying the partial pressures of
hydrogen and carbon monoxide showed that H2has a slightly
positive and CO a slightly negative influence on the reaction
rate. These observations are in accordance with the well
accepted Wilkinson mechanism for ligand-modified rhodium
catalysis.In the temperature range under investigation, the
Rh-2-SILP catalyst had an activation energy of EA=
The performance of the new Rh-2-SILP catalyst system
could be further optimized by increasing the temperature to
1208 8C and the total pressure to 25 bar. Under these con-
ditions, the TOF reached 3600 h?1which corresponds to a
space–time yield (STY) of 850 kg n-pentanal per cubic meter
per hour. This is—to our knowledge—the highest STY
reported for a SILP catalyst and a value that clearly outper-
forms most industrially realized, homogeneously catalyzed
processes. Under these harsher conditions, the selectivity
toward n-pentanal always remained above 99%. Note that
this ligand-induced, homogeneous catalyst performance was
achieved with a material that macroscopically has the
appearance of a dry solid and by using a highly diluted
(43% inert isobutene) industrial feedstock. Inductively
coupled plasma atomic emission spectrometry (ICP–AES)
analysis of aldehyde product condensate did not reveal any
Rh in the product (detection limit ca. 100 ppb). Increasing the
temperature to boost catalytic productivity even further was
unfeasible because the thermal instability of the phosphite
Our catalytic studies were accompanied by
spectroscopic investigations to learn more about the nature
and stability of the Rh-2-SILP catalyst. A fresh solution of the
[Rh(acac)(CO)2] precursor (acac=acetylacetonato) together
with a tenfold excess of 2 in the ionic liquid [EMIM][NTf2]
was measured (Figure 3a). A strong signal at d=145.3 ppm
corresponding to the free ligand can be seen together with a
doublet centered at d=147.5 ppm (JRh-P=296 Hz) resulting
from the rhodium–phosphite species.In a comparative
experiment, a freshly prepared Rh-2-SILP material was
washed with dried dichloromethane to remove the ionic
liquid catalyst solution from the silica support and this
washing solution was analyzed by NMR spectroscopy. As
expected the same signals at d=145.3 and 147.5 ppm were
found (Figure 3b) indicating that contact with the support
material during the SILP synthesis does not change the
catalyst complex or the free ligand. Even a SILP catalyst that
had been stored in the glovebox for 10 months showed still
the same31P NMR spectrum after the material was washed
with CH2Cl2(Figure 3c). Weak signals in the range between
d=5 and 8 ppm, however, correspond to formation of a small
amount of oxidized ligand species.This result indicates that
the rhodium–phosphite SILP catalysts can be stored in the
complete absence of air and water for prolonged times.
We also performed spectroscopic investigations on SILP
catalyst material that had been in catalytic use for more than
800 h (see Figure 2). The sample was taken from the reactor,
washed with CH2Cl2and the washings analyzed by NMR
spectroscopy(Figure 3d). Inthis case,a slightlydifferentpeak
pattern is observed with additional signals appearing at d
?172 ppm and 152 ppm and in the range between d=?2
and 8 ppm. Signals for the intact ligand and the rhodium
complex ligand are also present. For technical reasons, it was
unavoidable that the SILP material should briefly come into
contact with air and moisture during removal from the
reactor. However, a reference sample of the Rh-2 complex
that was placed in H2O2solution for 10 min showed no
significant sign of decomposition or oxidation products. It can
therefore be assumed that the decomposition products were
formed during catalysis without affecting the selectivity of the
system over 800 h. The nature of these decomposition
products and their role in the Rh-2 catalyzed hydroformyla-
tion under the given conditions is not yet clear. Note that
additional1H NMR experiments with the same SILP washing
solution revealed very intense signals arising from the ionic
liquid, demonstrating the presence of the supporting film still
after 800 h time-on-stream.
In conclusion, we have demonstrated the applicability of
the sophisticated diphosphite ligand 2 in the continuous gas-
phase hydroformylation of a diluted C4olefin mixture using
the SILP approach. Given the complex composition of the
industrial feedstock, the obtained selectivity of more than
99% n-pentanal under all the applied conditions is remark-
able. It indicates that the Rh-2-SILP materialis ahighly active
isomerizationcatalyst thatselectively addsCO/H2to 1-butene
and leaves 2-butenes and the large amount of isobutene
(43.1% of the feedstock mixture) completely untouched with
respect to direct hydroformylation. Furthermore, it is note-
worthy that the neutral ligand 2 can be applied efficiently in a
continuous gas-phase SILP process. In an ionic liquid–organic
biphasic operation, the same ligand would leach quickly into
the product phase and structural modification (e.g., attach-
ment of ionic groups) would be necessary to immobilize it in
the ionic liquid. Compared to former SILP hydroformylation
a) Rh-2 in [EMIM][NTf2], dissolved in CH2Cl2; b) fresh Rh-2-SILP,
washed with CH2Cl2, c) Rh-2-SILP, stored for 10 months under Ar,
washed with CH2Cl2; d) used Rh-2-SILP, 800 h on stream, washed with
31P NMR solution spectra of Rh-2 hydroformylation catalysts.
? 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2011, 50, 4492–4495
work where ionic ligands with ionic immobilizing groups were
always used (such as e.g. in sulfoxantphos, 1) this important
new finding will certainly pave the way for the application of a
much broader range of non-ionic ligand systems in future
SILP catalysis studies.
As expected, 2 was found to be sensitive to hydrolysis. If,
however, a dried feed gas is used and an acid scavenger added
to the immobilized ionic liquid on support, the catalyst
stability can be extended to more than 30 days time-on-
stream. During this operation time a total turnover number of
more than 350000 was achieved. Under slightly harsher
reaction conditions (1208 8C, 25 bar total pressure) the perfor-
mance of Rh-2-SILP can be increased with the same highly
diluted feed to a TOF of 3600 h?1and a space–time yield of
850 kgn-pentanalm?3h?1. This productivity level can be main-
tained for at least 10 h time-on-stream without visible
deactivation or loss in selectivity.
Our work demonstrates that SILP catalysis is suitable for
performing homogeneous catalysis with complex, sensitive,
and expensive ligand systems in a highly productive manner
using simple fixed-bed technology. In this particular case
where rhodium–diphosphite SILP systems were investigated,
significant space–time yield and excellent selectivity could be
achieved in a highly substrate-specific isomerization/hydro-
formylation reaction sequence. This result makes these
systems relevant for a future industrial use.
All syntheses have been carried out under inert atmosphere in a
glovebox. Details of the experimental setup have been reported
elsewhere and can be found in the Supporting Information. The pre
andpostanalysisoftheSILPcatalystwascarriedoutusinga 400 MHz
Jeol NMR spectrometer. Further details can be obtained in the
Both the ligand 2 and stabilizer 3 were obtained from Evonik
Oxeno GmbH. [Rh(CO)2(acac)] (Aldrich) was dissolved in water-
free CH2Cl2and stirred for 5 min. Ligand 2, dissolved in CH2Cl2, was
added in a tenfold excess (2/Rh=10:1), and the resulting yellow
solution was stirred for another 5 min. For stabilizer addition, a
fourfold excess (3/2=4:1) in CH2Cl2was added. Afterward, the ionic
liquid [EMIM][NTf2] (Merck KGaA) was added to the solution.
After 5 min stirring, the appropriate amount (see Supporting
Information) of calcinated silica (Merck KGaA) was added and the
suspension was stirred for 15 min. The CH2Cl2was removed in vacuo,
and a pale yellow powder was obtained (for details see Supporting
Information). The SILP catalyst was stored under argon until further
Received: November 15, 2010
Published online: April 8, 2011
phosphates · supported ionic liquid phase
Keywords: butene · hydroformylation · isomerization ·
 “Hydroformylation”: C. D. Frohning, C. W. Kohlpaintner in
Applied Homogeneous Catalysis with Organometallic Com-
pounds (Eds.: B. Cornils, W. A. Herrmann), Wiley-VCH,
 “Catalyst Separation, Recovery and Recycling”: Catalysis by
Metal Complexes, Vol. 30 (Eds.: D. Cole-Hamilton, R. Tooze),
Springer, Dordrecht, 2006.
 a) “Transition Metal Catalysis in Ionic Liquids”: P. Wassersc-
heid, P. S. Schulz in Ionic Liquids in Synthesis, Wiley-VCH,
Weinheim, 2007; b) V. I. P?rvulescu, C. Hardacre, Chem. Rev.
2007, 107, 2615–2665; c) M. Haumann, A. Riisager, Chem. Rev.
2008, 108, 1474–1497; d) H. Olivier-Bourbigou, L. Magna, D.
Morvan, Appl. Catal. A 2010, 373, 1–56.
 C. W. Kohlpaintner, R. W. Fischer, B. Cornils, Appl. Catal. A
2001, 221, 219–225.
 D. Camper, C. Becker, C. Koval, R. Noble, Ind. Eng. Chem. Res.
2005, 44, 1928–1933.
 F. Favre, H. Olivier-Bourbigou, D. Commereuc, L. Saussine,
Chem. Commun. 2001, 1360–1361.
 a) C. P. Mehnert, R. A. Cook, N. C. Dispenziere, M. Afeworki, J.
Am. Chem. Soc. 2002, 124, 12932–12933; b) A. Riisager, P.
Wasserscheid, R. van Hal,R. Fehrmann,J. Catal.2003, 219, 452–
455; c) A. Riisager, R. Fehrmann, S. Flicker, R. van Hal, M.
Haumann, P. Wasserscheid, Angew. Chem. 2005, 117, 826–830;
Angew. Chem. Int. Ed. 2005, 44, 815–819; d) Y. Gu, G. Li, Adv.
Synth. Catal. 2009, 351, 817–847.
 a) C. P. Mehnert, E. J. Mozeleski, R. A. Cook, Chem. Commun.
2002, 3010–3011; b) C. P. Mehnert, Chem. Eur. J. 2005, 11, 50–
56; c) A. Riisagera, R. Fehrmann, M. Haumann, P. Wassersc-
heid, Eur. J. Inorg. Chem. 2006, 695–706; d) A. Riisager, R.
Fehrmann, M. Haumann P. Wasserscheid, Top. Catal. 2006, 40,
 a) M. J. Earle, J. M. S. S. Esperanca, M. A. Gilea, J. N. Canon-
gia Lopes, N. J. L. P. N. Rebelo, J. W. Magee, K. R. Seddon, J. A.
Widegren, Nature 2006, 439, 831–834; b) P. Wasserscheid,
Nature 2006, 439, 797.
 A. Riisager, R. Fehrmann, M. Haumann, B. S. K. Gorle, P.
Wasserscheid, Ind. Eng. Chem. Res. 2005, 44, 9853–9859.
 M. Haumann, K. Dentler, J. Joni, A. Riisager, P. Wasserscheid,
Adv. Synth. Catal. 2007, 349, 425–431.
 M. Haumann, M. Jakuttis, S. Werner, P. Wasserscheid, J. Catal.
2009, 263, 321–327.
 A. Behr, D. Obst, C. Schulte, T. Schosser, J. Mol. Catal. A 2003,
 “Catalyst Preparation and Decomposition”: P. W. N. M. van
Leeuwen in Catalysis by Metal Complexes: Rhodium Catalyzed
Hydroformylation (Eds.: P. W. N. M. van Leeuwen, C. Claver),
Kluwer, Dordrecht, 2000.
 A. B?rner et al., unpublished results.
 “Phosphines as Ligands”: P. W. N. M. van Leeuwen, C. P. Casey,
G. T. Whiteker in Catalysis by Metal Complexes: Rhodium
Catalyzed Hydroformylation (Eds.: P. W. N. M. van Leeuwen, C.
Claver), Kluwer, Dordrecht, 2000.
 a) B. L. Booth, M. J. Else, R. Fields, R. N. Haszeldine, J.
Organomet. Chem. 1971, 27, 119–131; b) E. Billig, J. D. Jamer-
son, R. L. Pruett, J. Organomet. Chem. 1980, 192, C49.
 a) J. F. Young, J. A. Osborn, F. A. Jardine, G. J. Wilkinson, J.
Chem. Soc. Chem. Commun. 1965, 131; b) D. Evans, J. A.
Osborn, G. J. Wilkinson, Chem. Soc. 1968, 3133; c) D. Evans, G.
Yagupsky, G. J. Wilkinson, J. Chem. Soc. 1968, 2660.
 A. van Rooy, P. C. J. Kamer, P. W. N. M. van Leeuwen, K.
Goubitz, J. Fraanje, N. Veldman, A. L. Spek, Organometallics
1996, 15, 835–847.
have been calculated using ACD Labs NMR predictor software.
31P NMR shifts for possible degradation and oxidation products
Angew. Chem. Int. Ed. 2011, 50, 4492–4495? 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim