Inverted supercritical carbon dioxide/aqueous biphasic media for rhodium-catalyzed hydrogenation reactions

Article abstract of DOI:10.1002/chem.200601717

An inverted supercritical carbon dioxide (scCO₂)/aqueous biphasic system has been used as reaction media for Rh-catalysed hydrogenation of polar substrates. Chiral and achiral CO₂-philic catalysts were efficiently immobilised in scCO₂ as the stationary phase, while the polar substrates and products were contained in water as the mobile phase. Notably, product separation and catalyst recycling were conducted without depressurisation of the autoclave, The catalyst phase was reused several times with high conversion and product recovery of more than 85%. Loss of rhodium and phosphorus by leaching were found to be below the detection limit after the first two cycles in the majority of repetitive experiments. The reaction conditions were optimised with a minimum of experiments by using a simplex algorithm in a sequential optimisation. Total turnover numbers (TTNs) of up to 1600, turnover frequencies (TOFs) of up to 340 h^-1 and ee's up to 99% were obtained in repetitive batch operations. The scope of the devised catalytic system has been investigated and a semicontinuous reaction setup has been implemented. The chiral ligand (R,S)-3-H²F⁶-BINAPHOS allowed highly enantioselective hydrogenation of itaconic acid and methyl-2-acetamidoacrylate combined with a considerable catalyst stability in these reaction media.

Full text of DOI:10.1002/chem.200601717

DOI: 10.1002/chem.200601717
Inverted Supercritical Carbon Dioxide/Aqueous Biphasic Media for
Rhodium-Catalyzed Hydrogenation Reactions
[
a]
[a]
[a]
[a]
Katja Burgemeister, Giancarlo Franciò, Volker H. Gego, Lasse Greiner,
[
b]
[a]
Herbert Hugl, and Walter Leitner*
Abstract: An inverted supercritical
sion and product recovery of more
than 85%. Loss of rhodium and phos-
phorus by leaching were found to be
below the detection limit after the first
two cycles in the majority of repetitive
experiments. The reaction conditions
were optimised with a minimum of ex-
periments by using a simplex algorithm
in a sequential optimisation. Total turn-
over numbers (TTNs) of up to 1600,
turnover frequencies (TOFs) of up to
carbon dioxide (scCO )/aqueous biphas-
2
ꢀ
1
ic system has been used as reaction
media for Rh-catalysed hydrogenation
of polar substrates. Chiral and achiral
340 h and eeꢀs up to 99% were ob-
tained in repetitive batch operations.
The scope of the devised catalytic
system has been investigated and a
semicontinuous reaction setup has
been implemented. The chiral ligand
CO -philic catalysts were efficiently im-
2
mobilised in scCO2 as the stationary
phase, while the polar substrates and
products were contained in water as
the mobile phase. Notably, product
separation and catalyst recycling were
conducted without depressurisation of
the autoclave. The catalyst phase was
reused several times with high conver-
2
6
(R,S)-3-H F -BINAPHOS
allowed
highly enantioselective hydrogenation
of itaconic acid and methyl-2-acetami-
doacrylate combined with a considera-
ble catalyst stability in these reaction
media.
Keywords: enantioselectivity · hy-
drogenation · multiphase catalysis ·
rhodium · supercritical fluids
[
4,5]
Introduction
academia.
Also industrial applications were recently ac-
[6]
complished. In contrast to most conventional organic sol-
Multiphase catalysis represents a key technology for the im-
mobilisation of homogeneous catalysts and product separa-
vents, scCO is non-toxic, non-flammable and environmen-
2
tally benign. It enables fast mass transfer due to gas-like dif-
fusivity and at the same time offers a tuneable liquid-like
dissolving power. Mass-transfer limitations are strongly re-
duced for most gaseous reactants as they are fully miscible
[1]
tion. Among all possible solvent combinations for multi-
phase systems, aqueous-phase catalysis occupies a prominent
role as evidenced by the industrial success of the Ruhrche-
mie/Rhône-Poulenc oxo process (RCH/RP) for the hydro-
with CO . Therefore, hydrogenation reactions have been in-
2
[2,3]
formylation of propene.
In this process, a water-soluble
vestigated intensively using CO as the only reaction
2
[7]
rhodium catalyst is immobilised in the stationary aqueous
phase, whereas the products form a second phase that can
medium or in biphasic systems. When used in combination
with a second (immiscible) solvent, CO constitutes general-
2
[
8,9]
be easily separated. The gaseous reactants CO and H con-
ly the mobile phase, replacing the organic solvent.
Some
2
stitute the third phase in this multi-ton-scale process.
examples of biphasic aqueous/CO systems have been re-
2
The use of supercritical carbon dioxide (scCO ) as reac-
tion medium for catalysis continues to attract interest from
ported in which the water layer containing a hydrophilic cat-
2
[10,11]
alyst forms the stationary phase.
These systems are truly
biphasic (sc/l) in nature. However, the scope of such systems
is limited to non-polar and/or volatile products, which can
be readily extracted by CO2 from the aqueous catalyst
phase.
The majority of products in the focus of the fine chemical
and pharmaceutical industry are densely functionalised mol-
ecules of low-to-medium polarity. For such compounds, it
seems attractive to devise an inverted aqueous biphasic
system, in which water is used as the product phase. Howev-
[
a] K. Burgemeister, Dr. G. Franciò, V. H. Gego, Dr. L. Greiner,
Prof. Dr. W. Leitner
Institut für Technische und Makromolekulare Chemie
RWTH Aachen, Worringerweg 1, D-52074 Aachen (Germany)
Fax : (+49)241-802-2177
E-mail: leitner@itmc.rwth-aachen.de
[
b] Prof. Dr. H. Hugl
LANXESS Deutschland GmbH
5
1369 Leverkusen (Germany)
2798
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2798 – 2804

FULL PAPER
er, very few examples are known in which the catalyst is im-
mobilised in an organic phase and the polar substrates and
Then, an aqueous solution of itaconic acid 1 was dosed into
the reactor with a high-pressure pump (Figure 1). The reac-
tion progress was monitored by taking samples of the water
layer through a needle valve at the bottom of the reactor.
[12]
products reside in the aqueous phase. The use of classical
organic solvents as the catalyst phase causes significant cross
contamination of the aqueous product phase and the result-
ing waste water stream. To enable the conversion of polar
compounds without losing the advantages of the combina-
tion of water and scCO , we have recently proposed the in-
2
[13]
verted system scCO /H O.
Rhodium-catalysed hydrofor-
2
2
mylation of water-soluble substrates was used to demon-
strate that this approach allows catalyst recycling and prod-
[14]
uct isolation from the aqueous layer.
Another example
has been reported where a CO -soluble palladium catalyst
2
allowed the synthesis of aqueous H O from H and O in
2
2
2
2
[15]
scCO /H O media. Furthermore, the same group showed
2
2
Figure 1. Schematic experimental setup of the high pressure window
equipped autoclave (I: inlet, G: gas inlet, P: dosing pump, O: outlet).
that H O generated in such a system can be used in situ as
the oxidant for the epoxidation of alkenes.
2
2
[16]
Herein, we present the application of an inverted super-
critical CO /aqueous biphasic system for the hydrogenation
of C=C double bonds in polar substrates. The system is
This catalytic system was reused in a repetitive batch se-
quence by using the following procedure: After quantitative
conversion or after a given time, the aqueous phase was re-
moved and a new batch of substrate was pumped into the
reactor without depressurisation of the autoclave. Variation
of the ligand-to-rhodium ratio (L:Rh) from 3:1 to 5:1 at
constant reaction time of one hour led to a significant in-
crease in conversion as well as catalyst retention. At an L/
Rh ratio of 3:1, the conversion decreased from 34% after
one hour in the initial reaction cycle to 8% in the fourth
run (Table 1, entries 1–4). In contrast, an L/Rh ratio of 5:1
led to a constant conversion of 95% per cycle. The catalyst
could be recycled three times without apparent loss of activ-
ity (Table 1, entries 5–8). Leaching of rhodium and phospho-
2
[17]
based on “CO -philic” rhodium catalysts that are immobi-
2
lised in scCO as the stationary phase, while substrates and
2
products are contained in water as the mobile phase. Achi-
ral and chiral catalysts are utilised efficiently and the poten-
tial for continuous-flow operation is demonstrated.
Results and Discussion
Recycling experiments and kinetic investigations: As a first
benchmark system, phosphine-modified rhodium catalysts
were prepared from [Rh
diene) and tri-[3-
3,3,4,4,5,5,6,6,7,7,8,8,8-trideca-
fluoroctyl)phenyl]phosphane
A
H
R
U
G
2
4
(
[
a]
[
b]
[c]
[c]
Conversion
%]
TOFmax
Rh leaching
[ppm]
P leaching
[ppm]
2
6
[18]
(
3-H F -TPP)
genation of itaconic acid (1) as
prototypical substrate
Scheme 1). The catalysts were
for the hydro-
ꢀ1
[
A
H
R
U
G
]
A
H
R
U
G
[%]
A
H
R
U
G
[%]
34
20
14
8
95
95
95
95
70
24
20
14
252
194
192
192
19.5
4.4
2.1
2.0
14.2
3.6
18.9
4.3
2.0
1.9
14.0
3.5
0.6
0.6
0.9
0.8
6.2
0.8
1.0
0.8
0.6
0.6
1.0
0.8
4.0
0.5
0.6
0.5
a
(
prepared in situ by mixing the
ligand and the metal complex
in CH Cl . After evaporation of
2
2
1.0
0.6
0.9
0.6
the solvent, the residual solid
was re-dissolved in scCO2 and
subsequently the solution was
2
6
A
H
R
U
G
2
]BF
4 2
0.02 mmol, L=3-H F -TPP, substrate 1 (4.0 mmol, 0.2m in H O,
2
)=30 bar, T=408C, t=1 h@1000 rpm. [b] Compare Figure 2. [c] Rh
pressurised with
hydrogen. and P content measured by ICP methods.
rus into the aqueous layer was determined by inductively
coupled plasma (ICP) atomic emission spectrometry
(
Table 1). Significant amounts of rhodium and phosphorus
leached to the aqueous layer when an L/Rh=3:1 ratio was
used. By utilising a larger excess of ligand (L/Rh=5:1) the
leaching decreased below 1 ppm in the second and third
cycle for phosphorus and rhodium, respectively, and re-
mained low in the following cycles.
Scheme 1. Rhodium-catalysed hydrogenation of itaconic acid 1.
Chem. Eur. J. 2007, 13, 2798 – 2804
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2799

W. Leitner et al.
A series of experiments with short sampling intervals
were conducted to investigate the reaction course during the
repetitive runs. It was found that a short induction period
occurred during the first reaction cycle. The reaction rates
in following cycles are comparable within the experimental
error and no loss of activity is apparent. A total turnover
number (TTN) of 1000 was achieved within the four cycles
was not iterated further as the next calculated parameter set
would have exceeded the maximum autoclave pressure. Sev-
eral conclusions about the influence of individual parame-
ters can be drawn from this set of experiments. In general,
higher conversions were obtained by increasing the amount
of CO at constant temperature. This can be attributed to
2
higher substrate availability due to increasing solvent power
ꢀ1
with a maximum turnover frequency of 252 h (Figure 2).
of the dense CO and the decrease of hydrogen partial pres-
2
sure which act as an anti-solvent in the given system. If the
temperature is increased at a given CO amount, the accel-
2
eration of the catalytic cycle surpasses potential negative ef-
fects of a reduced CO density. The L/Rh ratio and the hy-
2
drogen concentration seem to be close to optimum values
and are not changed significantly.
Hydrogenation of other polar substrates: The successful ap-
plication of this catalyst system to further substrates de-
pends both on the intrinsic reactivity and on the partition
coefficient of each particular substrate within the biphasic
system. Table 3 presents the hydrogenation results of some
[
a]
Table 3. Hydrogenation of other substrates.
Substrate
Conversion [%]
Figure 2. Conversion/time profiles of repetitive batch hydrogenations of
itaconic acid 1 (experimental: points (symbol/entry in Table 1: square:5,
caret:6, triangles:7, circles:8, maximum reaction rates: lines).
Z-2-butenoic acid
tiglic acid
maleic acid
50
>99
15
>99
20
[b]
methyl-2-acetamidoacrylate (3)
[
c]
2
-acetamidoacrylic acid
[
c]
fumaric acid
>99
>99
Optimisation of reaction conditions: A simplex algorithm
a-acetamidocinnamic acid
for function minimisation according to Nelder and Mead
2
6
[19]
[a] Reaction conditions: [Rh
A
H
R
U
G
2 4
(cod) ][BF ] 0.02 mmol, 3-H F TPP
A
H
R
U
G
was used for the optimisation of the reaction conditions.
Maximised conversion was set as the target parameter of
1
.12 mmol, substrate 4 mmol in 20 mL H
2
O, Vtot =100 mL, m
A
C
H
T
R
E
U
N
G
2
(CO )=
6
2 g, p(H )=30 bar, T=568C, Reaction time=1 h@1000 rpm. [b] Cat-
alyst was reused four times (see Table 4). [c] Catalyst was reused four
times (data not shown).
2
the optimisation. The amounts of CO and hydrogen, the
2
temperature, and the L/Rh ratio were selected as the opti-
misation parameters to be varied within reasonable limits.
Substrate, rhodium amount, stirring speed, and reaction
time were kept constant. The conditions of the five initial
selected polar substrates in the inverted scCO /H O biphasic
2
2
[20]
experiments were randomly calculated.
The subsequent
system.
experiments were obtained by applying the algorithm as re-
flection of the experimental conditions with the least favour-
able result at the hyperplane given by the other parameter
sets (Table 2).
The highest efficacy was achieved with methyl-2-acetami-
doacrylate (3). Both, favourable partition and quantitative
conversion is achieved in hydrogenation of this substrate.
The catalyst could be recycled four times without noticeable
deterioration of performance, reaching a TTN of 750. The
leaching of rhodium and phosphorus decreased with the
number of repetitive batches performed (Table 4).
Within six experiments, an increase of conversion from
8
% to 86% could be achieved. The optimisation procedure
Table 2. Parameters for the hydrogenation of 1 obtained by simplex opti-
misation.
Semi-continuous reaction: Encouraged by the promising re-
sults of the batchwise recycling, a semi-continuous reaction
was carried out to assess the possibility of continuous proc-
essing. The hydrogenation of itaconic acid (1) was per-
formed in a 35 mL reactor with a catalyst/substrate ratio of
[
a]
Entry
m
A
C
H
T
R
E
U
N
G
(CO
2
)
T
L/Rh
p(H
2
)
Conv.
[%]
[g]
[8C]
ACHTREUNG
[bar]
1
2
3
4
5
6
50
60
55
55
55
62
40
40
58
46
46
56
5.0
5.0
5.0
5.8
5.2
5.6
30
30
30
30
38
34
8
9
65
17
22
86
1
0
:60, an L/Rh ratio of 6.2:1 and a substrate concentration of
.1 molL . A fixed amount of aqueous non-reactive phase
ꢀ1
was replaced at regular time intervals with the same volume
of a substrate solution without depressurisation. The shorter
the intervals compared to the averaged residence time are,
the closer the system approaches the characteristic of a con-
2
6
[
a] Reaction conditions: [Rh
A
H
R
U
G
2
]
A
H
R
U
G
4
[BF ] 0.02 mmol, L=3-H F TPP, sub-
strate 4 mmol in 20 mL H O, 30 min, 1000 rpm.
2
2800
www.chemeurj.org
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2798 – 2804

Multiphase Catalysis
FULL PAPER
Table 4. Hydrogenation of methyl-2-acetamidoacrylate 3 with catalyst re-
cycling.
obtained. Recycling of the catalyst phase resulted in a de-
crease of both conversion and enantioselectivity (Figure 4).
The instability of this system does not seem to result from
[
a]
Cycle
Conv. [%]
2 h
Leaching [ppm]
P
1
h
Rh
1
2
3
4
5
78
61
>99
>99
>99
81
2.3
1.7
0.5
0.5
0.3
4.9
2.1
1.5
0.4
0.5
92
>99
>99
>99
[
a] Reaction conditions: see footnote [a] in Table 3.
tinuously stirred tank reactor (CSTR). In these experiments,
a third of the total aqueous solution volume was replaced
every 30 min. After an induction period of approximately
Figure 4. Conversion and enantiomeric excess (ee) of the hydrogenation
of itaconic acid 1 with (R)-Cl-MeO-BIPHEP-modified rhodium catalyst.
1
.5 h, a stable conversion of 87% (ꢁ3%) was achieved for
the residual duration of the experiment (five hours in total;
Figure 3).
catalyst leaching, as no contamination of rhodium and phos-
phorus in the aqueous phase was detectable.
Stable performance was achieved with the same catalyst
for the hydrogenation of 3 leading to quantitative conver-
sion throughout five cycles. The enantioselectivity was only
modest (about 40% in all cycles; Figure 5). Also in these ex-
Figure 3. Semi-continuous hydrogenation of itaconic acid 1 (experimental
points: circles; average conversion after 1.5 h: straight line).
Figure 5. Conversion and enantiomeric excess (ee) of the hydrogenation
Asymmetric hydrogenation reactions: To extend this ap-
proach to enantioselective hydrogenation, the chiral bis-
of methyl-2-acetamidoacrylate
rhodium catalyst.
3 with (R)-Cl-MeO-BIPHEP-modified
A
H
R
U
G
[21]
BIPHEP
enantioselective hydrogenation
in the inverted biphasic system
scCO /H O.
periments, no leaching of rhodium and phosphorus into the
product phase was measurable.
2
6
The CO -philic ligand (R,S)-3-H F -BINAPHOS was then
2
2
2
[24]
Although the ligand itself is
tested in the same reaction. The chiral phosphane–phos-
phite framework of BINAPHOS was originally designed for
poorly soluble in scCO , a suffi-
2
cient solubility of the resulting
rhodium complex could be ach-
ieved by using the CO -philic
2
BARF (BARF=tetrakis-[3,5-bis(trifluormethyl)phenyl]bo-
[
22]
rate) as the counterion
3
1
and through the addition of
[23]
vol% benzotrifluoride as co-solvent. Hydrogenation of
was performed under the previously optimised conditions
except that an L/Rh ratio of 3:2 was employed, considering
the fact that Cl-MeO-BIPHEP is a bidentate ligand. In the
first cycle, a conversion of 89% and an ee of 53% (S) were
Chem. Eur. J. 2007, 13, 2798 – 2804
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2801

W. Leitner et al.
[25]
hydroformylation reactions, but recent research has dem-
onstrated its potential also for enantioselective hydrogena-
[26,27]
tion, especially in scCO media.
2
Again, the complex [Rh(cod) ]BARF was used as the pre-
A
H
R
U
G
2
cursor and the catalyst was preformed as before. No undis-
solved material was detected by visual inspection and the
water layer remained colourless, whereas the scCO phase
2
showed the yellow colour typical of rhodium–phosphane
complexes. Enantioselective hydrogenation of 1 led to quan-
titative conversion under standard conditions with an ee of
94% (S) (Figure 6). Recycling of the catalyst was straight-
Figure 7. Conversion and enantiomeric excess (ee) of the hydrogenation
of methyl-2-acetamidoacrylate 3 with (R,S)-3-H F -BINAPHOS-modified
2
6
rhodium catalyst.
Conclusion
The inverted biphasic system scCO /H O provides a promis-
2
2
ing approach to combine efficient transformation and sepa-
ration for organometallic catalytic processes leading to
highly polar and water-soluble products. Immobilisation of
the catalyst in the scCO phase can be achieved utilising es-
2
tablished concepts to design achiral or chiral “CO -philic”
2
[29]
Figure 6. Conversion and enantiomeric excess (ee) of the hydrogenation
of itaconic acid 1 with (R,S)-3-H F -BINAPHOS-modified rhodium cata-
catalysts.
The system is completely VOC-free and the
2
6
product-containing water phase is readily isolated with very
low levels of contamination of catalyst components (below
ICP detection limits in favourable cases). Continuous pro-
cessing is possible for which no energy consuming pressure
lyst.
forward, and only after seven cycles and a long hold up time
did the activity start to decrease. A TTN >1600 was reached
in the repetitive batch series. These data are encouraging as
they demonstrate a noticeable stability of the potential sen-
sitive catalytic system in the unconventional and fairly
cycle of a scCO -phase is required and liquid water consti-
2
tutes the mobile phase in this system. The resulting relative-
ly low operating costs, the high reaction rates (TOF
2
ꢀ1
ꢂ10 h ), excellent asymmetric induction (up to 99% ee)
and a remarkable long-term stability are salient features of
this novel approach to multiphase catalysis.
[28]
acidic medium (pH 3 under experimental conditions).
2
6
With the (R,S)-3-H F -BINAPHOS-based system, leach-
ing of rhodium and phosphorus into the aqueous layer was
found to be 1.6 and 4.9 ppm, respectively. This indicates a
Experimental Section
better retention in the scCO phase of the latter catalyst
2
compared to the previous achiral catalyst system. Two ef-
fects are likely to contribute to the better retention of the
General: Reactions involving air-sensitive materials were performed
under an atmosphere of dry argon using standard Schlenk techniques.
Water was distilled and degassed before use. Hydrogen (99.999%), argon
(99.995%) and carbon dioxide (99.9995%) were purchased from Messer
2
6
(
R,S)-3-H F -BINAPHOS-derived catalyst. One is the use
of the BARF anion, which is expected to increase the solu-
2
6
[30]
[21]
2
6
Griesheim. 3-H F TPP,
Cl-MeO-BIPHEP,
and (R,S)-3-H F -BINA-
bility of the metal complex in scCO relative to the tetra-
2
[
26]
PHOS were synthesised according to literature procedures. NMR spec-
tra were recorded on a Bruker DPX 300 spectrometer. Rhodium and
phosphorus contents were determined by inductively coupled plasma
(ICP) atomic emission spectrometry (average value over five replicates;
error ꢁ0.1 ppm, detection limit typically 0.7 to 1.2 ppm depending on the
amount of sample available).
fluoroborate analogue and which is not susceptible to hy-
drolysis. The second is the enhanced catalyst stability due to
the chelation of the bidentate ligand.
For the hydrogenation of 3, quantitative conversion and
an ee of 97.3% (R) were achieved in the first run (Figure 7).
Again, the product was easily isolated in quantitative yield
Caution: Working with highly compressed gases must be carried out only
using suitable equipment and with appropriate safety precautions.
directly from the water layer and the scCO phase could be
2
General procedure for batch and repetitive batch hydrogenation experi-
ments: Ligand and rhodium precursor (see text and Tables for details)
were dissolved in dichloromethane (3 mL) and stirred at room tempera-
ture for 30 min. This solution was transferred into a window-equipped
ꢀ
1
re-used several times. TTN=1000 and TOF=200 h were
achieved and the enantiomeric excess increased from 97 to
99% (R) over the course of five cycles. The contamination
1
00 mL autoclave (Tmax =1208C; pmax =200 bar) and the solvent was re-
of the aqueous layer with rhodium and phosphorus was low
for the initial experiment (1.4 ppm and 5.2 ppm, respective-
ly) and below the detection limit for the successive experi-
ments.
moved under reduced pressure. Alternatively, the ligand and rhodium
precursor were introduced in the autoclave as solids. No significant differ-
ence was observed between these two procedures. The reactor was pres-
2 2
surised with H (typically 30 bar) and then filled with CO (typically
2802
www.chemeurj.org
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2798 – 2804

Multiphase Catalysis
FULL PAPER
5
0 g). The reaction mixture was heated to 408C and the aqueous solution
ꢀ
1
of the substrate (20 mL, 0.2 molL ) was added to the autoclave. The re-
action was started by stirring at 1000 rpm, at which a highly disperse
scCO /water mixture is formed rapidly. Samples of the reaction mixture
2
2
6
Semi-continuous hydrogenation: 3-H F TPP (217.2 mg, 0,167 mmol) and
Rh(cod) ]BF (10.9 mg, 0,027 mmol) were dissolved in dichloromethane
3 mL). This solution was transferred into a 35 mL window-equipped au-
toclave (Tmax =1208C; pmax =400 bar) and the solvent was removed under
reduced pressure. The reactor was pressurised with H (40 bar), and
[
(
A
T
E
N
2
4
(
ꢂ0.5 mL) were withdrawn periodically from a needle valve placed at
the bottom of the reactor. The stirrer was switched off during sampling
to allow for complete phase separation. The products were isolated
simply by removing the water under reduced pressure and the conversion
2
heated to 558C. An aqueous solution of itaconic acid (15 mL,
.1 molL ) was added. The reactor was then filled with CO
pressure of 260 bar was reached. The reaction was started by stirring
500 rpm). Every thirty minutes the stirrer was turned off and 5 mL of
the water phase was withdrawn and replaced with the same volume of
fresh substrate solution. The pressure was readjusted to the initial value
ꢀ1
0
2
until a total
1
was determined by H NMR spectroscopy. The recovery of the substrate/
product mixture from the autoclave ranged from 85 to 90%. The repeti-
tive batch experiments were carried out by removing the water phase
completely and exchanging it with a substrate solution using a HPLC
pump.
(
2
by addition of CO . The hydrogen consumption was considered as negli-
Simplex optimisation: the conversion of 1 was defined as the target for
gible. The reaction was restarted by stirring. Quantification was carried
out as described above for batch experiments.
the optimisation sequence. The amount of CO
the H
ble parameters [Eq. (1)].
2
, the reaction temperature,
2
pressure and the ligand-to-rhodium ratio were chosen as the varia-
ꢀ
ꢁ
ꢂꢃ
nðLÞ
c ¼ f mðCO
2
Þ,T,pðH
2
Þ,
ð1Þ
nðRhÞ
Acknowledgements
We thank the team of Prof. Dr. A. Behr (University of Dortmund) for
the ICP measurements. Financial support from the BMBF (ConNeCat
light house project “Regulated Systems for Multiphase Catalysis,
ꢀ
1
The amount of itaconic acid (4 mmol, 0.2 molL ) and of rhodium pre-
cursor (5 mol%), stirring speed (1000 rpm) and reaction time (30 min)
were kept constant throughout the optimisation study. To calculate the
starting simplex, the initial conditions and the interval for each variable
have been set (Table 5). By using Equation (2) and the kF values accord-
0
5
3C0338D), the Deutsche Forschungsgemeinschaft (SFB 380 and SFB
40), and the Fonds der Chemischen Industrie are gratefully acknowl-
edged.
Table 5. Starting conditions of the simplex optimisation.
[
[
1] Multiphase Homogeneous Catalysis (Eds.: B. Cornils, W. A. Herr-
mann, D. Vogt, I. Horvath, H. Olivier-Bourbigon, W. Leitner, S.
Mecking), Wiley-VCH, Weinheim, 2005.
2] B. Cornils, J. Falbe, Proceedings of the 4th International Symposium
on Homogeneous Catalysis, Leningrad 1984, 487; G. Kessen, B. Cor-
nils, J. Hibbel, H. Bach, W. Gick, EP-0216151, 1986 (Hoechst AG).
C. W. Kohlpaintner, R. W. Fischer, B. Cornils, Appl. Catal .A 2001,
Parameter
(CO ) [g]
T [8C]
L/Rh
X
0
DX
m
A
C
H
T
R
E
U
N
G
2
50
40
5
10
20
1
p(H
2
) [bar]
30
10
2
21, 219.
[
[
[
3] Aqueous-Phase Organometallic Catalysis (Eds.: B. Cornils, W. A.
Herrmann), Wiley-VCH, Weinheim, 2004.
4] Chemical Synthesis Using Supercritical Fluids, (Eds.: P. G. Jessop, W.
Leitner), Wiley-VCH, Weinheim, 1999.
5] P. G. Jessop, T. Ikariya, R. Noyori, Chem. Rev. 1999, 99, 475; W.
Leitner, Acc. Chem. Res. 2002, 35, 746.
Table 6. Values of kF of the simplex optimisation.
Entry
kF value for
m
A
C
H
T
R
E
U
N
G
(CO
2
)
T
L/Rh
2
p(H )
1
2
3
4
5
0
1.00
0.50
0.50
0.05
0
[6] P. Licence, J. Ke, M. Sokolova, S. K. Ross, M. Poliakoff, Green
Chem. 2003, 5, 99; see also http://www.thomas-swan.co.uk.
[7] W. Leitner in Handbook of Homogeneous Hydrogenation, Vol. 3
(Eds.: J. G. de Vries, C. J. Elsevier), Wiley-VCH, Weinheim, 2007,
p. 1361.
0.87
0.29
0.29
0
0.82
0.20
0
0.79
[
8] Hydrogenation: R. A. Brown, P. Pollet, E. McKoon, C. A. Eckert,
C. L. Liotta, P. G. Jessop, J. Am. Chem. Soc. 2001, 123, 1254; M. Sol-
inas, A. Pfaltz, P. G. Cozzi, W. Leitner J. Am. Chem. Soc. 2004, 126,
ing to the algorithm shown in Table 6, the first five working points were
calculated.
1
6142; D. J. Heldebrant, P. G. Jessop, J. Am. Chem. Soc. 2003, 125,
5
600.
X
i
¼ X
o
þ ðkF DXÞ
ð2Þ
[
9] Other reactions: M. F. Sellin, P. B. Webb, D. J. Cole-Hamilton,
Chem. Commun. 2001, 781; P. B. Webb, M. F. Sellin, T. E. Kunene, S.
Williamson, A. M. Z. Slawin, D. J. Cole-Hamilton, J. Am. Chem.
Soc. 2003, 125, 15577; A. Bçsmann, G. Franciò, E. Janssen, M. Soli-
nas, W. Leitner, P. Wasserscheid, Angew. Chem. 2001, 113, 2769;
Angew. Chem. Int. Ed. 2001, 40, 2697; Z. S. Hou, B. X. Han, L. Gao,
T. Jiang, Z. M. Liu, Y. H. Chang, X. G. Zhang, J. He, New J. Chem.
2002, 26, 1246; D. Ballivet-Tkatchenko, M. Picquet, M. Solinas, G.
Franciò, P. Wasserscheid, W. Leitner Green Chem. 2003, 5, 232; Z.
Hou, N. Theyssen, A. Brinkmann, W. Leitner, Angew. Chem. 2005,
The obtained results were then listed in descending order relative to the
conversion and the barycentre of the best four experiments is calculated
according to Equation (3).
!
1
4
!
!
!
!
S
a,b,c,d
¼
ð a þ b þ c þ d Þ
ð3Þ
1
17, 1370; Angew. Chem. Int. Ed. 2005, 44, 1346; A. Serbanovic,
The following experiment was calculated [Eq. (4)] by reflection of the
point of lowest conversion onto the barycentre calculated before
L. C. Branco, M. N. da Ponte, C. A. M. Afonso, J. Organomet.
Chem. 2005, 690, 3600; L. C. Branco, A. Serbanovic, M. N. da Ponte,
C. A. M. Afonso, Chem. Commun. 2005, 107; M. T. Reetz, W. Wie-
senhçfer, G. Franciò, W. Leitner, Adv. Synth. Catal. 2003, 345, 1221.
X
new ¼ 2Sa,b,c,dꢀe
ð4Þ
Chem. Eur. J. 2007, 13, 2798 – 2804
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2803

W. Leitner et al.
[
10] Hydrogenation: B. M. Bhanage, Y. Ikushima, M. Shirai, M. Arai,
Chem. Commun. 1999, 1277; S. Fujita, S. Akihara, M. Arai, J. Mol.
Catal. A 2006, 249, 223; R. J. Bonilla, B. R. James, P. G. Jessop,
Chem. Commun. 2000, 941; B. Jacobson, C. T. Lee, Jr., K. P. John-
son, W. Tumas, J. Am. Chem. Soc. 1999, 121, 11902.
[18] S. Kainz, D. Koch, W. Baumann, W. Leitner, Angew. Chem. 1997,
109, 1699; Angew. Chem. Int. Ed. Engl. 1997, 36, 1628; D. Koch, W.
Leitner, J. Am. Chem. Soc. 1998, 120, 13398.
[19] J. A. Nelder, R. Mead, Comput. J. 1965, 7, 308.
[20] M. Otto, Chemometrie, VCH, 1997, 127.
[
11] Other reactions: B. M. Bhanage, Y. Ikushima, M. Shirai, M. Arai,
Tetrahedron Lett. 1999, 40, 6427; J. P. Hanrahan, K. J. Ziegler, J. P.
Galvin, J. D. Holmes, Langmuir 2004, 20, 4386.
12] G. Verspui, G. Elbertse, F. A. Sheldon, M. A. P. J. Hacking, R. A.
Sheldon, Chem. Commun. 2000, 1363; G. Verspui, G. Elbertse, G.
Papadogianakis, R. A. Sheldon, J. Organomet. Chem. 2001, 621, 337.
R. M. Deshpande, S. S. Divekar, B. M. Bhanage, R. V. Chaudhari, J.
Mol. Catal. 1992, 75, L19.
13] As suggested previously, we follow the notation of catalyst or reac-
tive phase and product or non-reactive phase, respectively see M. F.
Eckstein, M. Peters, J. Lembrecht, A. C. Spiess, L. Greiner, Adv.
Synth. Catal. 2006, 348, 1591; M. F. Eckstein, J. Lembrecht, J. Schu-
macher, W. Eberhard, A. C. Spiess, M. Peters, C. Roosen, L. Grein-
er, U. Kragl, Adv. Synth. Catal. 2006, 348, 1597; see also refer-
ence [7].
[21] B. Driessen-Hçlscher, J. Kralik, F. Agel, C. Steffens, Adv. Synth.
Catal. 2004, 346, 979.
[
22] M. J. Burk, S. Feng, M. F. Gross, W. Tumas, J. Am. Chem. Soc. 1995,
17, 8277; A. Wegner, W. Leitner, Chem. Commun. 1999, 1583; S.
Kainz, A. Brinkmann, W. Leitner, A. Pfaltz, J. Am. Chem. Soc.
999, 121, 6421.
1
[
1
[
23] S. Lange, A. Brinkmann, P. Trautner, K. Woelk, J. Bargon, W. Leit-
ner, Chirality 2000, 12, 450.
[
[
24] G. Franciò, W. Leitner, Chem. Commun. 1999, 1663.
25] K. Nozaki, N. Sakai, S. Mano, T. Higashijima, T. Horiuchi, H.
Takaya, J. Am. Chem. Soc. 1997, 119, 4413.
[
[
[
[
26] G. Franciò, K. Wittmann, W. Leitner, J. Organomet. Chem. 2001,
6
21, 130.
27] Y. Yan, Y. Chi and X. Zhang, Tetrahedron: Asymmetry 2004, 15,
173.
2
28] C. Roosen, M. Ansorge-Schumacher, T. Mang, W. Leitner, L. Grein-
er, Green Chem. 2007 DOI: 10.1039/b613345b.
[
[
[
14] M. McCarthy, H. Stemmer, W. Leitner, Green Chem. 2002, 4, 501.
15] D. Hꢂncu, E. J. Beckman, Green Chem. 2001, 3, 80.
16] D. Hꢂncu, J. Green, E. J. Beckman, Ind. Eng. Chem. Res. 2002, 41,
[
[
29] W. Leitner, Nature 2000, 405, 129.
30] S. Kainz, Z. Luo, D. P. Curran, W. Leitner, Synthesis 1998, 1425.
4
466; E. J. Beckman, Green Chem. 2003, 5, 332.
[
17] For a preliminary communication see: K. Burgemeister, G. Franciò,
Received: November 30, 2006
Published online: February 13, 2007
H. Hugl, W. Leitner, Chem. Commun. 2005, 6026.
2804
www.chemeurj.org
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 2798 – 2804