Challenging the scope of continuous, gas-phase reactions with supported ionic liquid phase (SILP) catalysts - Asymmetric hydrogenation of methyl acetoacetate

Article abstract of DOI:10.1016/j.apcata.2011.03.038

In the last years Supported Ionic Liquid Phase (SILP) catalysis has attracted growing interest for continuous gas phase reactions. The concept combines in a unique manner the advantages of traditional homogeneous and heterogeneous catalysis by making use of the extremely low vapor pressure of an ionic liquid film on support. In this way, continuous gas phase reactions with macroscopically solid but microscopically homogeneous catalyst materials have been successfully realized. However - so far - continuous gas-phase SILP catalysis has only been demonstrated for reactions with relatively simple selectivity problems (such as, regioselectivity, chemoselectivity). This study demonstrates the successful application of a SILP catalyst in continuous gas phase asymmetric hydrogenation. The hydrogenation of methyl acetoacetate (MAA) is presented using an immobilized dibromo[3-(2,5-(2R,5R)-dimethylphospholanyl-1) -4-di-o-tolylphosphino-2,5-dimethylthio-phene]-ruthenium (Ru1Br₂) catalyst.

Full text of DOI:10.1016/j.apcata.2011.03.038

Applied Catalysis A: General 399 (2011) 35–41
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Challenging the scope of continuous, gas-phase reactions with supported ionic
liquid phase (SILP) catalysts—Asymmetric hydrogenation of methyl acetoacetate
a
a
a
b
a,∗
Eva Öchsner , Martin J. Schneider , Carolin Meyer , Marco Haumann , Peter Wasserscheid
a
Lehrstuhl für Chemische Reaktionstechnik, Friedrich-Alexander-Universität Erlangen-Nürnberg, Egerlandstrasse 3, D-91058 Erlangen, Germany
Chemical Reaction Engineering, Friedrich-Alexander-University Campus Busan, 1276 Jisa-Dong, Gangseo-Gu, Busan 618-230, Republic of Korea
b
a r t i c l e i n f o
a b s t r a c t
Article history:
In the last years Supported Ionic Liquid Phase (SILP) catalysis has attracted growing interest for continuous
gas phase reactions. The concept combines in a unique manner the advantages of traditional homoge-
neous and heterogeneous catalysis by making use of the extremely low vapor pressure of an ionic liquid
film on support. In this way, continuous gas phase reactions with macroscopically solid but microscop-
ically homogeneous catalyst materials have been successfully realized. However—so far—continuous
gas-phase SILP catalysis has only been demonstrated for reactions with relatively simple selectivity
problems (such as, regioselectivity, chemoselectivity). This study demonstrates the successful applica-
tion of a SILP catalyst in continuous gas phase asymmetric hydrogenation. The hydrogenation of methyl
acetoacetate (MAA) is presented using an immobilized dibromo[3-(2,5-(2R,5R)-dimethylphospholanyl-
1)-4-di-o-tolylphosphino-2,5-dimethylthio-phene]-ruthenium (Ru1Br2) catalyst.
Received 5 January 2011
Received in revised form 8 March 2011
Accepted 18 March 2011
Available online 26 March 2011
Keywords:
Asymmetric catalysis
Hydrogenation
Ionic liquids
Ketones
Supported catalysts
SILP
© 2011 Elsevier B.V. All rights reserved.
1
. Introduction
In 2002, Mehnert et al. for the first time published the immobi-
liquid biphasic catalysis is the large specific exchange surface
between the reactant flow and the ionic catalyst solution. Together
with typical IL film thicknesses in the lower nanometer range
(depending on IL loading and support pore structure) catalytic
materials are obtained that provide very short diffusion distances
for the reactants through the relatively viscous ionic catalyst solu-
tion.
lization of a transition metal catalyst in a thin film of ionic liquid
immersed on a porous inorganic support [1], a concept for which
the term “Supported Ionic Liquid Phase” (SILP) catalysis was coined
shortly after [2]. Since then, the concept to apply solids coated with
an ionic liquid layer as catalyst has attracted steeply growing sci-
entific interest as evidenced by a number of recent reviews on the
topic [3].
SILP catalyst materials make use of the fact that many transi-
tion metal complexes dissolve well in ionic liquids [4]. To prepare
a SILP catalyst, the resulting ionic catalyst solution is dispersed
over the high internal surface area of a porous support. Due to
the very good wettability of ionic liquids on typical inorganic sup-
ports and as a result of strong capillary forces, the thus formed
material is a dry solid that still contains the ionic catalyst solu-
tion in the form of a thin liquid film in its pores. Consequently,
the dissolved catalyst still acts microscopically as a homogeneously
dissolved complex in its uniform ionic liquid environment. Macro-
scopically, however, a solid SILP material is obtained that can be
processed in reactor concepts traditionally applied in heteroge-
neous catalysis, e.g. fixed-bed reactors. A particular strength of
the SILP catalyst concept compared to traditional organic/ionic
In reviewing the recent development of SILP catalyst applica-
tions, it is obvious that almost all examples showing long-term
active and selective systems were obtained by applying the SILP
material in contact with a continuous gas flow of reactants. In
this way successful examples of hydroformylation [5], methanol
carbonylation [6], hydroamination [7], hydrogenation [8] and ultra-
low temperature water–gas shift reaction [9] have been achieved.
In contrast, the original concept of Mehnert and a number of
later publications [10] where the SILP material is contacted with
a liquid reaction mixture exhibited intrinsic problems regarding
catalyst stability. The latter concern (a) the lack of stability of the
liquid film on the support vs. mechanical removal by convection,
(b) the leaching of the very thin IL film into the liquid reaction phase
and (c) the leaching of the applied transition metal complex into
the liquid reaction phase.
While the first aspect might be solved by clever support selec-
tion, the other two require modifications of the support or addition
of ionic tags to the chiral ligand. One attempt to overcome these
limitations is the covalent linkage of ionic species to the support
[
11]. However, this implicates considerable additional synthetic
∗ Corresponding author. Tel.: +49 9131 85 27420; fax: +49 9131 85 27421.
E-mail address: wasserscheid@crt.cbi.uni-erlangen.de (P. Wasserscheid).
effort to make the SILP slurry system suitable for effective recycling.
0
926-860X/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2011.03.038

3
6
E. Öchsner et al. / Applied Catalysis A: General 399 (2011) 35–41
Scheme 1. Hydrogenation of methyl acetoacetate using dibromo[3-(2,5-(2R,5R)-dimethylphospholanyl-1)-4-di-o-tolylphosphino-2,5-dimethyl-thiophene]ruthenium as
catalyst.
This contribution describes our recent efforts to explore the
applicability and scope of continuous gas-phase SILP catalysis for
the asymmetric hydrogenation of prochiral ketones. This reaction
is important to produce optically pure alcohols that are applied
as building blocks in the production of pharmaceuticals and fine
chemicals [12,13]. Traditionally, the reaction is carried out in
organic solvents using chiral transition metal complexes [14,15].
Typically, complexes of precious metals, such as Ru, Rh or Ir, are
applied together with chiral ligands obtained from multi-step syn-
theses. Thus, efficient concepts and strategies are highly desirable
to recover these expensive catalysts from the reaction mixture
The applied SILP catalyst material consisted of a thin film of 1-
ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide on
silica gel 30 with dibromo[3-(2,5-(2R,5R)-dimethylphospholanyl-
1)-4-di-o-tolylphosphino-2,5-dimethyl-thiophene]ruthenium dis-
solved as catalyst therein. Such bidentate ligand modified
Ru-complexes are known from the literature to exhibit excellent
performance when dissolved in protic solvents such as methanol
[33]. Generally, the applied SILP catalysts contained 0.38 mass% Ru
dissolved in an ionic liquid volume that corresponded to 80% of the
silica pore volume.
[
16–18].
In the last decade, many different strategies have been
2. Experimental
developed to immobilize chiral catalyst complexes for asymmet-
ric hydrogenation to enable easy catalyst recycling and re-use
options. Important examples include the covalent anchoring of
chiral complexes on polymers [19], silica [20], zirconium phos-
phonate supports [21], and immobilization concepts with large
2
.1. Chemicals
All chemicals were handled using standard Schlenk
techniques. Commercially available chemicals were used
without further purification. For catalyst preparation, [bis(2-
methylallyl)(1,5-cyclooctadiene)ruthenium (II)] from Acros and
3
2
used. The hydrobromic acid (48%) was purchased from Fluka.
Commercial dry methanol and acetone from Acros (both 99.9%,
water content less than 50 ppm) were used. The silica gel 30
[
22] or polymerized ligands [23]. Furthermore, concepts for a
non-covalent immobilization of chiral metal complexes by ion
exchange [24] and embedding in membranes [25] are notewor-
thy. An alternative strategy for the recycling of catalysts for
asymmetric hydrogenation is immobilization in a liquid–liquid
biphasic system. Water [26], fluorinated phases [27], and ionic
liquids (ILs) [28] have been successfully used as immobilization
phases.
SILP catalysis has only recently been applied to asymmetric
hydrogenation in three reported examples dealing with the hydro-
genation of 2-cyclohexen-1-one, 1,3-cyclooctadiene and methyl
acetoacetate (MAA). All these examples have been studied in
slurry phase reactions using the solvents isopropanol, diethyl ether,
toluene, or hexane [29]. Moreover, the hydrogenation of acetophe-
none using either isopropanol [13] or hexane [30] as solvents has
been reported. Recently, SILP catalysts have also been used in com-
bination with supercritical carbon dioxide for the enantioselective
hydrogenation of dimethyl itaconate [31].
-(2,5-(2R,5R)-dimethylphospholanyl-1)-4-di-o-tolylphosphino-
,5-dimethylthiophene from Evonik Degussa GmbH were
1
3
−1
(particle size 0.2–1 mm, pore volume 0.37 cm g , Merck KGaA)
and silanized silica gel 60 (particle size 0.063–0.2 mm, pore
3
−1
volume 0.54 cm g , Merck KGaA) were used as support mate-
rials (see Supporting information Table S-1 for further support
characterization). The ionic liquid [EMIM][NTf ] (H O < 500 ppm,
Cl < 100 ppm, Merck KGaA) was used for SILP preparation. For the
miniplant runs the carrier gas helium (purity: 4.6) and hydrogen
2
2
−
(purity: 5.0) from Linde AG and methyl acetoacetate (purity: 99%)
from Aldrich was used. During some experiments—marked with
“Oxisorb guard bed for He and H ”—the gases were contacted
2
®
with an OXISORB (Messer Industriegase GmbH, guaranteed
final impurities of oxygen <5 ppm) guard bed prior to addi-
tion to the reactor to avoid contamination of the catalyst with
oxygen.
Our paper focuses on the continuous, gas phase, SILP catalyzed,
asymmetric hydrogenation of methyl acetoacetate to methyl
hydroxybutyrate (see Scheme 1). It builds on our recently reported
work on the liquid phase hydrogenation of MAA in mixtures of
methanol and ionic liquids [32]. Particular challenges of trans-
ferring asymmetric hydrogenation reactions into continuous gas
phase systems concern the very unusual reaction conditions (in
2.2. Catalyst preparation
The catalyst was prepared in situ according to literature
[34]. A mixture of one equivalent of [bis(2-methylallyl)(1,5-
cyclooctadiene)ruthenium (II)] and two equivalents of the ligand
1 were placed in a Schlenk flask under argon atmosphere. Dry ace-
tone (5 mL) and a methanolic solution (0.29 M) of hydrobromic
acid (0.33 mL) were added and the solution was stirred 30 min
at room temperature. This solution was heated to 30 C in an oil
bath and the solvent was removed under high vacuum over 1 h.
The catalyst dibromo[3-(2,5-(2R,5R)-dimethylphospholanyl-1)-4-
◦
our case T > 100 C and residence times ꢀ < 3 s) that are linked to
this synthetic concept. While a continuous gas phase operation
provides various advantages over batch-wise processing, it is, how-
ever, linked to certain restrictions. All feedstocks and products have
to be volatile at reaction conditions. The partial pressure of the
reactants has to be low enough in order to avoid condensation
in the catalyst bed and therefore high throughput for high-boiling
reactants is an issue.
◦

Fig. 1. Flow scheme of our continuous asymmetric hydrogenation reactor setup equipped with two different reactors: Berty reactor (G) and tubular reactor (H).

3
8
E. Öchsner et al. / Applied Catalysis A: General 399 (2011) 35–41
di-o-tolylphosphino-2,5-dimethyl-thiophene]ruthenium (Ru1Br2)
was obtained as a brown solid powder.
To synthesize the SILP catalyst, Ru1Br2 was dissolved in 15 mL
dry and degassed methanol. Afterwards the porous support and
the ionic liquid were added to the catalyst solution and the mix-
ture was stirred under argon for 90 min. In the next step the solvent
methanol was removed by high vacuum treatment and an impreg-
nated support was obtained. The SILP system was stored under
argon atmosphere until further use in continuous experiments.
2.3. Continuous gas phase experiments
The continuous gas phase experiments were carried out in a
fixed bed reactor plant equipped with a Berty-type reactor [35]
and a tubular reactor (see Fig. 1). In the Berty reactor (95 mL vol-
ume), a gas phase circulation (loop flow) is generated by a turbine
on top of the reactor vessel and intensively contacted with the solid
SILP catalyst in the reactor. The reactor has been originally devel-
oped for kinetic measurements of classical heterogeneous catalysts
because of its CSTR-like gradient-free concentration profile. The
applied tubular reactor was 30 cm long and had a reaction volume
of 35 mL.
Fig. 2. Conversion vs. time-on-stream profiles of the SILP catalyzed MAA hydro-
genation in a continuous gas-phase reaction—variation of IL loading. (Reaction
conditions: Berty reactor; L/Ru = 2:1, [EMIM][NTf2], silica 30 (0.2–1 mm), 0.38 wt.%
◦
Ru, 125 C, 10 bar, 5000 rpm, 200 NmL/min He, 75 NmL/min H2, 0.25 g/h MAA,
ꢀ
= 2.4 s.)
1
g of the SILP catalyst was filled into the Berty reactor (G) or
in the tubular reactor (H) under helium atmosphere and the com-
plete rig was flushed with helium. The rig was pressurized with
It is highly noteworthy that the use of the ionic liquid film on
the support was found to be absolutely essential to realize any cat-
alytic hydrogenation in the system. By just impregnating the silica
with the same Ru-precursor/ligand without ionic liquid, no cat-
alytic activity was observed. Additionally, it was even found that a
relatively high degree of pore filling ˛ (˛ = ionic liquid volume/pore
volume) with the ionic liquid is beneficial, probably to protect
the catalyst from unfavorable chemical interaction with the sup-
port surface. High ˛ values result in higher maximum conversions
1
0 bar helium and left under pressure for 15 min while monitor-
ing the pressure. If no pressure drop was observed, the reactor
was heated to reaction temperature under a continuous flow of
helium. Helium and hydrogen were fed into the rig via mass flow
controllers (MFC (A) and (B), Bronkhorst F-201CV-200-RAD-11-
V/F-201CV-200-RAD-33-V). The carrier gas helium was used to
evaporate the substrate MAA in a CEM (controlled evaporation
and mixing)-unit (F) (Bronkhorst) and to flush the top section
of the Berty reactor. MAA and MeOH were taken from a reser-
voir in the liquid state and fed into the CEM-unit via LIQUI-flows
(
e.g. ˛ = 1, X = 32.4%) and a slightly shortened activation phase. This
result also indicates that rising ionic loading ˛ does not seem to
cause visible mass transport limitation effects, a fact that is cor-
related to the relatively slow catalytic reaction. In contrast to the
conversion profiles, the enantiomeric excess profile in the stable
phase is not affected by a different IL-loading (ee in the range of
(
D, E) (D: Bronkhorst L13-RAD-11-K-10S, E: Bronkhorst L13V02-
RAD-11-K-10S). In order to ensure proper mixing, the preheated
hydrogen and helium–MAA flows were combined in a mixing unit,
which consisted of a 7 mm Swagelok filter filled with glass wool.
The gas mixture could then either enter the Berty reactor (G),
or enter the tubular reactor (H), or exit the system via a bypass.
After the reactor, the gas mixture passed a 7 ␮m filter in order to
avoid contamination of the pressure valve with solid particles. A
back-pressure regulator valve (I) was used to maintain the desired
reaction pressure and outlet gas flow. After the regulator valve the
reactants were condensed (J) and samples were collected at reg-
ular intervals in an auto-sampler (K). The collected samples were
diluted with diethyl ether and afterwards directly analyzed via gas
chromatography on a Varian 3900 equipped with a Lipodex E col-
umn (length 25 m, inner diameter 0,25 mm, temperature program:
8
2–85%). After 30 h time-on-stream (TOS) slight changes in enan-
tioselectivity were observed, comparing the different IL loadings.
With these first experiments it became evident that the SILP
catalyst concept is in principle able to promote asymmetric hydro-
◦
◦
◦
◦
◦
8
0 C–20 min isotherm, 80 C → 200 C, 10 C/min, 200 C–10 min
◦
isotherm, carrier gas: 0.4 mL/min helium, detector: FID, 250 C).
3
. Results and discussion
The results of our initial hydrogenation runs to test the influ-
ence of ionic liquid loading in the Berty reactor are presented
in Figs. 2 and 3. In all these experiments, the conversion and ee
vs. time-on-stream showed comparable profiles. During the ini-
tial activation process conversion and enantiomeric excess increase
steadily. The activation process is followed by a period of stable
catalytic operation, concerning both selectivity and activity. Sub-
sequently, a deactivation period, in which the enantiomeric excess
and conversion decrease and the chiral catalyst degrades over time,
is observed.
Fig. 3. ee vs. Time-on-stream profiles of the SILP catalyzed MAA hydrogenation
in
a continuous gas-phase reaction—variation of IL loading. (Reaction condi-
◦
tions: L/Ru = 2:1, [EMIM][NTf2], silica 30 (0.2–1 mm), 0.38 wt.% Ru, 125 C, 10 bar,
5000 rpm, 200 NmL/min He, 75 NmL/min H2, 0.25 g/h MAA, ꢀ = 2.4 s.)

E. Öchsner et al. / Applied Catalysis A: General 399 (2011) 35–41
39
Table 1
NH3-TPD measurement of the applied supports.
Support
Surface aciditya (␮mol/g)
Silica 60 (silanized)
Silica 30 (calcinated)
Silica 30 (dried)
43.23
159.16
276.35
a
Determined by NH3 temperature programmed desorption.
genation in a continuous gas-phase reaction with reasonable
◦
enantiomeric excess at 125 C. The moderate conversions of around
3
0% have to be rated in the light of the very short catalyst contact
time of only 2.4 s (!) in these experiments.
Encouraged by these initial results, we investigated in the next
step the influence of different silica based support materials char-
acterized by different surface acidities. Table 1 shows the analysis
of surface acidity of the applied support materials as determined
by NH -TPD measurements.
3
The lowest acidity was found for silanized silica in which
surface Si–OH groups have been modified by treatment with
dichloro-dimethyl-silane. The calcinated silica showed a signifi-
cantly reduced amount of Si–OH groups compared to dried silica.
In the calcinated sample, a large proportion of the Si–OH groups
had been removed during the calcination process. The SILP sys-
tem with the silanized silica support showed low catalytic activity
Fig. 4. Conversion vs. time-on-stream profiles of the SILP catalyzed MAA hydro-
genation in a continuous gas-phase reaction—influence of the additive methanol.
(
Reaction conditions: Berty reactor; L/Ru = 2:1, [EMIM][NTf2], silica 30 (0.2–1 mm),
◦
ionic liquid loading ˛ = 0.8, 0.38 wt.% Ru, 125 C, 10 bar, 5000 rpm, 200 NmL/min He,
75 NmL/min H2, 0.125 g/h MAA, 0.152 g/h MeOH, ꢀ = 2.4 s, oxisorb guard bed for He
and H2.)
(
XMax = 5.8%; see Supporting information, Fig. S-1). In contrast,
SILP catalysts immobilized on calcinated or dried silica resulted
troscopy (ICP-AES). No leaching of Ru (catalyst), P (ligand) and S
(ligand and ionic liquid) was detected in the liquid product (detec-
tion limit: 1 ppm). Additionally, leaching of the Br-ligand (resulting
in a stability-loss of the transition metal complex) was investigated.
During one experiment the product stream was passed through an
acidic aqueous solution of silver nitrate, but no precipitation of AgBr
was observed.
As another potential reason for the observed deactivation
behavior, the accumulation of organic compounds or side-products
in the SILP catalyst was considered. Analysis of the product stream
by means of gas chromatography and mass spectroscopy (GC–MS)
indicated the formation of high-boiling side-products such as
transesters and aldol adducts in traces. However, largely deacti-
vated SILP catalysts (after 60 h TOS) showed no re-activation after
vacuum treatment for 8 h (see Supporting information, Fig. S-4).
Accordingly, side-product formation is not the only reason for the
observed deactivation.
in better conversions (calcinated SiO : XMax = 17.7%, dried SiO :
2
2
XMax = 20.4%). Additionally, the catalyst activation time was sig-
nificantly shortened by using non-calcinated silica instead of the
calcinated support (19 h TOS for full activation vs. 27 h TOS). Wolf-
son et al. [36] described the positive influence of acids on the
catalytic activity of Ru-based asymmetric hydrogenation of MAA
and postulated a modification of the general catalytic cycle origi-
nally described by Noyori et al. [34,37]. The modified catalytic cycle
involves protonation of the carbonyl group of the substrate by an
acid prior to hydride insertion. Furthermore, it was stated that the
presence of acidic solids in the reaction mixture could enhance
the reaction in aprotic solvents. We assume that in our case the
acidic Si–OH groups enhance the catalytic activity of the Ruthenium
catalyst in the same manner.
The enantioselectivity was also influenced by the acidity of
the support material in a certain range. Use of the silanized sil-
ica support reduced the ee in the time range of stable activity
to 70% compared to ee = 85% for calcinated SiO2 (see Supporting
information, Fig. S-2). The SILP catalyst with dried silica showed an
ee of 80% in the stable activity phase.
Furthermore, the influence of the reaction temperature on the
stability of the SILP catalyst was tested, because the applied tran-
sition metal catalyst Ru1Br has been originally designed for lower
2
◦
reaction temperatures in liquid phase reactions (50–80 C) [33].
Further experiments focused on the catalyst activation time
and the influence of hydrogen pre-treatment. In this context, a
SILP catalyst was flushed with hydrogen for 16 h prior to addi-
tion of the MAA substrate. Remarkably, the catalyst activation could
indeed be shortened from 27 h to 20 h in this way (see Supporting
information, Fig. S-3). This finding hints for an acceleration of the
catalyst activation process in a hydrogen saturated system. This
accelerated activation process can be explained by the postulated
reaction mechanisms of Noyori and Ohkuma [37] and Wolfson et al.
Noteworthy, time to full activation of the SILP catalyst was
strongly depending on the reaction temperature, with 31 h
◦
◦
time-on-stream for 125 C, but 63 h for 105 C (see Supporting
information, Fig. S-5). The period of stable catalytic operation was
significantly increased at lower reaction temperatures (35 h TOS
◦
◦
at 105 C compared to 10 h TOS at 125 C); on the other hand
a lower maximum conversion was observed. Remarkably, enan-
◦
tiomeric excess was up to 88% at 105 C and decreased very slowly
over TOS (still 80% ee after 140 h TOS; see Supporting information,
Fig. S-6).
[
36]. The latter include the conversion of the Ru-dihalide catalyst
precursor to the Ru-hydride by reaction with hydrogen followed
by elimination of HX. This activation step should be accelerated in
a hydrogen saturated SILP-system
The following experiments aimed for a better understanding of
the reasons for the observed catalyst deactivation in our continuous
experiments. To exclude the rather unlikely option that the active
Ru-complex leaves the SILP material via the gaseous product phase
by sublimation, condensed samples from the continuous reactor
were analyzed by inductive coupled plasma-atom emission spec-
Ru1Br2 is known to exhibit particularly high activities in alco-
holic and protic solvents, especially in methanol [34c]. Additionally,
methanol increases the hydrogen solubility in the thin ionic liquid
film. Therefore we expected an increase in activity and enan-
tioselectivity by continuously co-feeding of gaseous methanol
into our reactor system (for details see Experimental Part). The
corresponding conversion/ee vs. time-on-stream profiles of the
◦
continuous experiments in the Berty reactor at 125 C are presented
in Figs. 4 and 5.

4
0
E. Öchsner et al. / Applied Catalysis A: General 399 (2011) 35–41
Fig. 7. ee vs. Time-on-stream profiles of the SILP catalyzed MAA hydrogenation in
a continuous gas-phase reaction—influence of reactor type and temperature. (Reac-
Fig. 5. ee vs. Time-on-stream profiles of the SILP catalyzed MAA hydrogenation
in a continuous gas-phase reaction—influence of the additive methanol. (Reaction
conditions: Berty reactor; L/Ru = 2:1, [EMIM][NTf2], silica 30 (0.2–1 mm), ionic liquid
loading ˛ = 0.8, 0.38 wt.% Ru, 125 C, 10 bar, 5000 rpm, 200 NmL/min He, 75 NmL/min
H2, 0.125 g/h MAA, 0.152 g/h MeOH, ꢀ = 2.4 s, oxisorb guard bed for He and H2.)
tion conditions: L/Ru = 2:1, [EMIM][NTf2], silica 30 (0.2–1 mm), ionic liquid loading
◦
˛
= 0.8, 0.38 wt.% Ru, 105 C, 10 bar, 5000 rpm (Berty), 200 NmL/min He, 75 NmL/min
◦
H2, 0.125 g/h MAA, 0.152 g/h MeOH, ꢀ = 2.4 s, oxisorb guard bed for He and H2.)
latter may fill and block the open pore volume of our SILP catalyst
material. Figs. 6 and 7 show the direct comparison of the contin-
uous gas phase hydrogenation in the tubular reactor vs. the same
In the experiment with added methanol, catalyst activation was
much faster and a higher activity in the period of stable catalytic
operation was obtained (XMax = 87.3%). This confirms that the pres-
ence of MeOH shows the same activating effects than previously
reported for the reaction in organic solvents [34c]. However, the
period of stable catalytic operation (in both cases 10 h) was not
extended by added methanol. Surprisingly, the decrease in activ-
ity with the alcoholic additive is smoother and catalytic activity is
observed up to a TOS of 160 h (X = 5.7%). Unexpectedly, the enan-
tiomeric excess is not improved by adding methanol.
◦
experiments at 105 C in the Berty reactor.
Indeed, a very positive effect of the plug flow reactor on the cat-
alyst stability was observed. After 35 h TOS full catalyst activation
was achieved and stable catalytic operation over at least 70 h was
realized. Thus, for more than 100.000 residence times 70% con-
◦
version (XMax = 74%) with 75–80% ee could be realized at 105 C
reaction temperature. During the total reaction time of the experi-
ment in the tubular reactor a total turnover number of the catalyst
of 2500 was achieved.
◦
Additional experiments were carried out at 105 C to realize
longer catalyst stability and higher ee’s. At this temperature, exper-
iments in the Berty reactor (G in Fig. 1) and in the tubular reactor
4. Conclusion
(
H in Fig. 1) were compared. The reason for this reactor comparison
was our suspicion that the broad residence time distribution (RTD)
of the Berty reactor, which is identical to the RTD of an ideal tank
reactor, could lead to preferred formation of higher boiling consec-
utive reaction products (such as transesterification products). The
In conclusion, we have successfully extended the scope of
the SILP catalyst concept to continuous gas phase reaction with
a complex, ligand influenced enantioselectivity. In detail, we
have demonstrated for the first time a continuous, gas phase,
asymmetric hydrogenation reaction. In the hydrogenation of
methyl acetoacetate using Ru1Br in 1-ethyl-3-methylimidazolium
2
bis(trifluoromethyl-sulfonyl)imide on silica 30, an enantiomeric
excess in the range of 65–82% ee could be obtained for more than
◦
1
00 h time-on-stream continuous operation (experiment at 105
in the plug flow reactor). Moreover, it was demonstrated, that the
ionic liquid film on the support is absolutely essential for catalytic
activity in the system and a relatively high degree of pore filling
with the ionic liquid is beneficial. The catalyst activation times can
be shortened by using non-calcinated silica instead of the calci-
nated support and by hydrogen pre-treatment. The combination
of thermal degradation of the SILP catalyst and the formation of
high-boiling side-products were identified as the main threats for
catalyst stability. By changing the reactor from a back-mixed Berty-
type to a tubular reactor, the accumulation of side-products in the
SILP catalyst could be significantly reduced and much higher cata-
lyst stability was observed.
Remarkably, our continuous gas phase hydrogenation has been
achieved for a substrate of relatively low volatility by using helium
as carrier gas in the reactor. Based on these results, we expect
that the use of SILP catalyst technology in continuous gas phase
reactions can be in principal transferred to any substrate that has
enough vapor pressure to be analyzed by gas chromatography. SILP
catalyst materials in continuous gas-phase contact offer a very sim-
Fig. 6. Conversion vs. time-on-stream profiles of the SILP catalyzed MAA hydrogena-
tion in a continuous gas-phase reaction—influence of reactor type and the additive
methanol. (Reaction conditions: L/Ru = 2:1, [EMIM][NTf2], silica 30 (0.2–1 mm), ionic
liquid loading ˛ = 0.8, 0.38 wt.% Ru, 105 C, 10 bar, 5000 rpm (Berty), 200 NmL/min
He, 75 NmL/min H2, 0.125 g/h MAA, 0.152 g/h MeOH, ꢀ = 2.4 s, oxisorb guard bed for
◦
He and H2.)

E. Öchsner et al. / Applied Catalysis A: General 399 (2011) 35–41
41
ple way of catalyst/product separation and allow application of
well-known catalytic systems without additional synthetic effort
for ligand modification. The finding of this study that even com-
plex ligand induced selectivities can be tackled by this continuous
gas-phase technology should open further fields of application.
[10] (a) H. Hagiwara, Y. Sugawara, T. Hoshi, T. Suzuki, Chem. Commun. (2005) 2942;
(
(
b) H. Hagiwara, K.H. Ko, T. Hoshi, T. Suzuki, Chem. Commun. (2007) 2838;
c) C. Brochwitz, A. Feldhoff, U. Kunz, M. Vaultier, A. Kirschning, Lett. Org. Chem.
3
(2006) 442.
[11] (a) C.P. Mehnert, Chem. Eur. J. 11 (2005) 50–56;
b) B. Gadenne, P. Hesemann, J.J.E. Moreau, Chem. Commun. (2004) 1768–1769.
12] T. Ohkuma, H. Ooka, S. Hashiguchi, T. Ikariya, R. Noyori, J. Am. Chem. Soc. 117
9) (1995) 2675–2676.
[13] L.-L. Lou, X. Peng, K. Yu, S. Liu, Catal. Commun. 9 (2008) 1891–1893.
(
[
(
Acknowledgements
[
[
[
14] W. Tang, X. Zhang, Chem. Rev. 103 (2003) 3029–3069.
15] W.S. Knowles, Angew. Chem. Int. Ed. 41 (2002) 1998–2007 (nobel lecture).
16] M. Heitbaum, F. Glorius, I. Escher, Angew. Chem. Int. Ed. 45 (2006)
4732–4762.
We gratefully thank Dr. Renat Kadyrov (Evonik Degussa GmbH)
for fruitful discussions on ligand selection and for providing
ligand samples. The authors acknowledge support by the Fed-
eral Ministery of Education and Research (BMBF), the Deutsche
Forschungsgemeinschaft within the Erlangen Excellence Cluster
[17] A.F. Trindade, P.M.P. Gois, C.A.M. Afonso, Chem. Rev. 109 (2009)
18–514.
18] J.M. Fraile, J.I. Graci, J.A. Mayoral, Chem. Rev. 109 (2009) 360–417.
4
[
[19] (a) D.J. Bayston, J.L. Fraser, M.R. Ashton, A.D. Baxter, M.E.C. Polywka, E. Moses,
J. Org. Chem. 63 (1998) 3137–3140;
“
Engineering of Advanced Materials and SYNFLOW within the EU’s
(
3
b) T. Ohkuma, H. Takeno, Y. Honda, R. Noyori, Adv. Synth. Catal. 343 (2001)
69–375.
Seventh Framework Programme“.
[
20] (a) P.N. Liu, J.G. Deng, Y.Q. Tu, S.H. Wang, Chem. Commun. (2004) 2070–2071;
(b) B. Kesanli, W. Lin, Chem. Commun. (2004) 2284–2285.
Appendix A. Supplementary data
[
[
[
21] A. Hu, H.L. Ngo, W. Lin, J. Am. Chem. Soc. 125 (2003) 11490–11491.
22] C. Köllner, B. Pugin, A. Togni, J. Am. Chem. Soc. 120 (1998) 10274–10275.
23] R.T. Halle, E. Schulz, M. Spagnol, M. Lemaire, Tetrahedron Lett. 41 (2000)
3323–3326.
24] W.P. Hems, P. McMorn, S. Riddel, S. Watson, F.E. Hancock, G.J. Hutchings, Org.
Biomol. Chem. 3 (2005) 1547–1550.
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.apcata.2011.03.038.
[
[
References
25] D. Tas, C. Thoelen, I.F.J. Vankelecom, P.A. Jacobs, Chem. Commun. (1997)
2323–2324.
[
[
1] C.P. Mehnert, R.A. Cook, N.C. Dispenziere, M. Afeworki, J. Am. Chem. Soc. 124
2002) 12932–12933.
2] (a) A. Riisager, P. Wasserscheid, R. van Hal, R. Fehrmann, Catal. Lett. 90 (2003)
49–153;
b) A. Riisager, P. Wasserscheid, R. van Hal, R. Fehrmann, J. Catal. 219 (2) (2003)
52–455.
3] (a) P. Tundo, A. Perosa, Chem. Soc. Rev. 36 (2007) 532–550;
[26] S. Zeror, J. Collin, J.C. Fiaud, L.A. Zouioueche, J. Mol. Catal. A: Chem. 256 (2006)
85–89.
(
[27] E. de Wolf, B.-J. Deelman, in: J.G. de Vries, C.J. Elsevier (Eds.), Handbook of
Homogeneous Hydrogenation, Wiley-VCH, Weinheim, 2007, pp. 1377–1388.
[28] P. Wasserscheid, P. Schulz, in: J.G. de Vries, C.J. Elsevier (Eds.), Handbook of
Homogeneous Hydrogenation, Wiley-VCH, Weinheim, 2006, pp. 1389–1421.
[29] A. Wolfson, I.F.J. Vankelecom, P.A. Jacobs, Tetrahedron Lett. 44 (2003)
1195–1198.
[30] K.L. Fow, S. Jaenicke, T.E. Müller, C. Sievers, J. Mol. Catal. A: Chem. 279 (2008)
239–247.
[31] U. Hintermair, T. Höfener, T. Pullmann, G. Franciò, W. Leitner, ChemCatChem 2
(2010) 150–154.
1
(
4
[
(
(
(
(
b) Y. Gu, G. Li, Adv. Synth. Catal. 351 (2009) 817–847;
c) S. Werner, M. Haumann, P. Wasserscheid, Annu. Rev. Chem. Biomol. Eng. 1
2010) 203–230;
d) C. Van Doorslaer, J. Wahlen, P. Mertens, K. Binnemans, D. De Vos, Dalton
Trans. 39 (2010) 8377–8390.
[
[
4] P. Wasserscheid, P.S. Schulz, T. Welton (Eds.), Ionic Liquids in Synthesis, 2nd
ed., Wiley-VCH, Weinheim, 2007, pp. 369–463.
5] (a) A. Riisager, R. Fehrmann, S. Flicker, R. van Hal, M. Haumann, P. Wasserscheid,
Angew. Chem. Int. Ed. 44 (2005) 815–819;
[32] (a) E. Öchsner, K. Schneiders, K. Junge, M. Beller, P. Wasserscheid, Appl. Catal.
A 364 (1–2) (2009) 8–14;
(b) T. Floris, P. Kluson, M.J. Muldoon, H. Pelantova, Catal. Lett. 134 (2010)
279–287.
(
(
(
b) A. Riisager, R. Fehrmann, M. Haumann, P. Wasserscheid, Eur. J. Inorg. Chem.
2006) 695–706;
c) A. Riisager, R. Fehrmann, M. Haumann, B.S.K. Gorle, P. Wasserscheid, Ind.
[33] (a) R. Kadyrov, Chim. Oggi 26 (5) (2008) 41–42;
(b) R. Kadyrov, J. Tarabocchia, M.L. Prunier (Eds.), Catalysis of Organic Reactions,
CRC Press, Boca Raton, 2009, pp. 211–216;
Eng. Chem. Res. 44 (2005) 9853–9859;
(c) R. Kadyrov, R.M. Koenigs, C. Brinkmann, D. Voigtlaender, M. Rueping, Angew.
Chem. Int. Ed. 48 (2009) 7556–7559.
(
(
(
d) A. Riisager, R. Fehrmann, M. Haumann, P. Wasserscheid, Top. Catal. 40 (1–4)
2006) 91–102;
e) M. Haumann, K. Dentler, J. Joni, A. Riisager, P. Wasserscheid, Adv. Synth.
[34] (a) R. Noyori, T. Ohkuma, M. Kitamura, H. Takaya, N. Sayo, H. Kumobayashi, S.
Akutagawa, J. Am. Chem. Soc. 109 (1987) 5856–5858;
(b) M. Kitamura, T. Ohkuma, S. Inoue, N. Sayo, H. Kumobayashi, S. Akutagawa,
T. Ohta, H. Takaya, R. Noyori, J. Am. Chem. Soc. 110 (1988) 629–631.
[35] J.M. Berty, Chem. Eng. Prog. 70 (5) (1974) 78–84.
[36] A. Wolfson, I.F.J. Vankelecom, S. Geresh, P.A. Jacob, J. Mol. Catal. A: Chem. (2004)
21–26.
[37] (a) R. Noyori, T. Ohkuma, Angew. Chem. Int. Ed. 40 (2001) 40–73;
(b) R. Noyori, M. Kitamura, T. Ohkuma, Proc. Natl. Acad. Sci. U.S.A. 101 (2004)
5356–5362.
Catal. 349 (3) (2007) 425–431.
6] A. Riisager, B. Jørgensen, P. Wasserscheid, R. Fehrmann, Chem. Commun. (2006)
[
[
[
[
9
94–996.
7] O. Jimenez, T.E. Müller, C. Sievers, A. Spirkl, J.A. Lercher, Chem. Commun. (2006)
974–2976.
8] M. Ruta, G. Laurenczy, P.J. Dyson, L. Kiwi-Minsker, J. Phys. Chem. C 112 (2008)
7814–17819.
2
1
9] S. Werner, N. Szesni, A. Bittermann, M.J. Schneider, P. Härter, M. Haumann, P.
Wasserscheid, Appl. Catal. A 377 (2010) 70–75.