Continuous platinum-catalyzed enantioselective hydrogenation in
supercritical’ solvents
‘
Roland Wandeler, Niklaus Künzle, Michael S. Schneider, Tamas Mallat and Alfons Baiker*
Laboratory of Technical Chemistry, Swiss Federal Institute of Technology, ETH-Zentrum, CH-8092 Zürich,
Switzerland. E-mail: E-mail:baiker@tech.chem.ethz.ch; Fax: +41 1 6321163; Tel: +41 1 6323153
Received (in Cambridge, UK) 15th January 2001, Accepted 6th March 2001
First published as an Advance Article on the web 21st March 2001
Hydrogenation of ethyl pyruvate in ‘supercritical’ ethane in
a fixed bed reactor over cinchona-modified Pt/Al O affords
2 3
good ee at an exceptionally high rate, whereas in carbon
dioxide the catalytic performance under similar conditions is
inferior.
Application of supercritical (sc) fluids as solvents and reactants
1
–3
has great potential for optimizing chemical reactions.
The
or
continuous hydrogenation of organic compounds in sc CO
2
propane has recently been reported by Poliakoff and co-
4
workers. Here we show the first example of a continuous
asymmetric hydrogenation in a ‘sc’ fluid. The well studied
5
enantioselective hydrogenation of ethyl pyruvate (EP) over
2 3
cinchonidine (CD)-modified Pt/Al O has been chosen as a
model reaction (Scheme 1) to demonstrate the feasibility of the
process. A crucial point is that trace amounts of the chiral
modifier have to be fed continuously to the reactor in order to
Fig. 1 Conversion (2) and ee (5) of EP hydrogenation in dense CO
2
and
6
maintain a good ee with time-on-stream. We show that the
ethane at 30 °C and 100 bar. Molar ratio of solvent+EP+H = 500+1+10.
2
application of ‘sc’ethane as a solvent affords a remarkable
increase in reaction rate compared to the best conventional
solvent, toluene. Note that the widely used term supercritical is
deprived of any meaning in multi-component systems since
phase separation is still possible at conditions beyond the
mixture critical point or the critical point of the pure
21
EP-flow was 4.3 mmol min
.
reversed behaviour was observed when the sequence of solvents
was changed. The poor performance in CO is not completely
2
understood yet. A possible explanation is partial poisoning of
the platinum catalyst due to the formation of CO via the reverse
7
components. For convenience, ‘sc’ is used here in quotes for
8
water gas shift reaction (CO
2
+ H
2
" CO + H
2
O). As a
the solvent-rich phase at temperatures exceeding its mixture
critical point, irrespective of further liquid phases present.
consequence of these comparative studies further investigations
were performed in ethane.
Careful consideration of the phase behavior under reaction
conditions is critical for understanding the outcome of the
reaction. The phase behavior of the system under reaction
conditions was investigated in a computer controlled high-
pressure view cell of variable volume (23–63 ml), equipped
with on-line digital video imaging and recording. The magnet-
ically stirred cell consisted of a horizontal cylinder equipped
with a sapphire window covering the entire diameter and an
opposite, horizontally moving piston equipped with another
sapphire window for illumination of the system. The basic setup
of the computer-based approach and video imaging has been
described before.9
Scheme 1
Catalytic studies were carried out in a continuous stainless
steel tubular fixed-bed reactor with 12.5 mm inner diameter. A
mechanical mixture of 100 mg 5 wt% Pt/Al
759, metal dispersion: 0.27) and 900 mg Al O (110 m g
6
2 3
O (Engelhard
2
21
4
2 3
surface area) as diluent was employed, resulting in a catalyst
The reaction mixture ethane–EP–H
LLV equilibrium (EP–rich liquid, ethane-rich liquid and
ethane–H -rich gaseous phase) at 30 and 40 °C. The upper two
2
exhibited a three-phase
bed length of 15 mm. The catalyst was prereduced in situ at 400
2
1
°
C in H
2
. A flow of 1.0 ml min EP (Fluka, 97%) was mixed
(99.995%) and H (99.999%) in a
2
with ethane (99.5%) or CO
2
2
ethane-rich phases critically merged at around 40 °C and 70 bar.
Beyond this upper critical endpoint of the coexistence of the
ethane-rich phases, the system exhibited a two-phase equilib-
rium of a liquid EP-rich phase and a dense fluid (‘sc’) ethane-
rich phase, continuously blending to one phase with increasing
pressure. The pressure required for this transition depended
static mixer before entering the reactor. CD (Fluka, > 98%) was
fed together with EP (Fluka, 97%) at a molar ratio EP+CD of
500+1. The corresponding solution was prepared immediately
before the reaction and kept cool and in the dark to minimize
side-reactions. Conversion and ee were determined by GC
analysis without derivatization. Chemoselectivity to ethyl
lactate was always 100%. Enantiomeric excess is defined as
2
strongly on temperature (20–50 °C) and H
(1–9%) in the system. A similar phase behavior was found in the
system ethane–ethyl lactate–H as well as in the system ethane–
ethyl pyruvate–ethyl lactate–H . On this basis we assume that
the whole conversion range can be well modelled by the ethane–
EP–H system discussed above.
At low H concentration, the transition from a multiphase to
2
concentration
([R] 2 [S])/([R] + [S]).†
2
Preliminary studies using CO
2
and ethane as solvents
2
indicated that the latter is a better solvent for EP hydrogenation.
Fig. 1 shows the changes in conversion and ee induced during
continuous hydrogenation of EP when the solvent is changed
2
2
from dense CO
reaction rate and enantioselectivity with ethane. A similar but
2
to ethane. Note the prominent increase in
a homogeneous fluid phase (single phase) was accompanied by
a pronounced increase in EP conversion (Fig. 2, k = 2). In
DOI: 10.1039/b100511l
Chem. Commun., 2001, 673–674
673
This journal is © The Royal Society of Chemistry 2001