Continuous-flow asymmetric hydrogenation of the β-keto ester methyl propionylacetate in ionic liquid-supercritical carbon dioxide biphasic systems

Article abstract of DOI:10.1002/adsc.201200724

A continuous-flow process for the asymmetric hydrogenation of methyl propionylacetate as a prototypical β-keto ester in a biphasic system of ionic liquid and supercritical carbon dioxide (scCO₂) is presented. An established ruthenium/2,2′-bis(diphenylphosphino)-1,1′-binaphthyl (BINAP) catalyst was immobilised in an imidazolium-based ionic liquid while scCO₂ was used as mobile phase transporting reactants in and products out of the reactor. The use of acidic additives led to significantly higher reaction rates and enhanced catalyst stability albeit at slightly reduced enantioselectivity. High single pass conversions (>90%) and good enantioselectivity (80-82% ee) were achieved in the first 80h. The initial catalyst activity was retained to 91% after 100h and to 69% after 150h time-on-stream, whereas the enantioselectivity remained practically constant during the entire process. A total turnover number of ～21,000 and an averaged space-time yield (STY_av) of 149g L^-1 h ^-1 were reached in a long-term experiment. No ruthenium and phosphorus contaminants could be detected via inductively coupled plasma optical emission spectrometry (ICP-OES) in the product stream and almost quantitative retention by the analysis of the stationary phase was confirmed. A comparison between batch-wise and continuous-flow operation on the basis of these data is provided. Copyright

Full text of DOI:10.1002/adsc.201200724

FULL PAPERS
DOI: 10.1002/adsc.201200724
Continuous-Flow Asymmetric Hydrogenation of the b-Keto Ester
Methyl Propionylacetate in Ionic Liquid–Supercritical Carbon
Dioxide Biphasic Systems
Jens Theuerkauf,a Giancarlo Franciꢀ,a,* and Walter Leitnera,b,*
a
Institut fꢁr Technische und Makromolekulare Chemie (ITMC), RWTH Aachen University, Worringerweg 1, D-52074
Aachen, Germany
Fax : (+49)-241-80-22177; e-mail: francio@itmc.rwth-aachen.de or leitner@itmc.rwth-aachen.de
Max-Planck-Institut fꢁr Kohlenforschung, Kaiser-Wilhelm-Platz 1, D-45470 Mꢁlheim an der Ruhr, Germany
b
Received: August 14, 2012; Revised: October 31, 2012; Published online: January 4, 2013
Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/adsc.201200724.
Abstract: A continuous-flow process for the asym- as the enantioselectivity remained practically con-
metric hydrogenation of methyl propionylacetate as stant during the entire process. A total turnover
a prototypical b-keto ester in a biphasic system of number of ~21,000 and an averaged space-time yield
ionic liquid and supercritical carbon dioxide (scCO2) (STYav) of 149 gLÀ1 hÀ1 were reached in a long-term
is presented. An established ruthenium/2,2’-bis(di- experiment. No ruthenium and phosphorus contami-
phenylphosphino)-1,1’-binaphthyl (BINAP) catalyst nants could be detected via inductively coupled
was immobilised in an imidazolium-based ionic plasma optical emission spectrometry (ICP-OES) in
liquid while scCO2 was used as mobile phase trans- the product stream and almost quantitative retention
porting reactants in and products out of the reactor. by the analysis of the stationary phase was con-
The use of acidic additives led to significantly higher firmed. A comparison between batch-wise and con-
reaction rates and enhanced catalyst stability albeit tinuous-flow operation on the basis of these data is
at slightly reduced enantioselectivity. High single provided.
pass conversions (>90%) and good enantioselectivi-
ty (80–82% ee) were achieved in the first 80 h. The Keywords: asymmetric ketone hydrogenation;
initial catalyst activity was retained to 91% after BINAP; carbon dioxide; continuous flow; ionic liq-
100 h and to 69% after 150 h time-on-stream, where- uids; ruthenium; supercritical fluids
Introduction
tary properties of these media result in a biphasic re-
action system that combines rapid mass transfer with
Homogeneous transition metal-catalysed asymmetric essentially quantitative phase separation. This allows
hydrogenation is an important method for the synthe- for an effective integration of catalytic reaction and
sis of high value chiral chemicals under mild reaction product separation in a single-step process.[6] Under
conditions.[1] The application of this technology, how- continuous flow conditions, the use of compressed
ever, is still facing typical obstacles associated with carbon dioxide as transport vector permits metal- and
homogeneous catalysed processes such as product solvent-free product isolation and even fractiona-
separation and catalyst recycling: Often, product iso- tion,[7,8] which can meet the strict purity criteria of the
lation requires resource-, time-, and labour-intensive pharmaceutical industry. The potential of this ap-
downstream processing to ensure complete removal proach has been successfully demonstrated so far in
of catalyst residues. Moreover, recovery and recycling hydroformylation,[9] C=C hydrogenation,[10] hydrovi-
of the often expensive ligands and precious metals are nylation,[11] and alkene metathesis.[12]
essential to achieve acceptable production costs.[2]
In this work, we have investigated the continuous
Different approaches have emerged in recent years flow asymmetric hydrogenation of b-keto esters to b-
to circumvent the aforementioned hurdles[3] and bi- hydroxy esters, which are useful building blocks for
phasic reaction systems based on ionic liquids (ILs) flavours and pharmaceuticals, for example, as key
and compressed carbon dioxide (scCO2) have been chiral intermediate in the synthesis of atorvastatin cal-
identified as a promising solution.[4,5] The complemen- cium (Lipitor).[13] The aim was to elucidate the funda-
Adv. Synth. Catal. 2013, 355, 209 – 219
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
209

Jens Theuerkauf et al.
FULL PAPERS
mental molecular and physico-chemical parameters reaction rates and low enantioselectivity were ob-
for the adjustment of the well-established Ru/BINAP served in most cases.[14e,15]
catalysts to this system as a basis for a fully integrated
reaction/separation process under stable long-term using the ionic liquid [dMEIm]AHCNUTGTRENNUNG
Thus, a series of batch experiments was carried out
[BTA] without co-sol-
operation.
vents to identify an appropriate catalyst precursor
(Table 1).
Catalyst A, prepared in situ from [RuACTHNUTRGEN(UNG benzene)Cl2]2
Results and Discussion
and (S)-BINAP by Noyoriꢃs procedure,[16] resulted in
high enantioselectivity (95% ee). In line with the liter-
ature reports,[15] reduced activity (TOF=56 hÀ1) was
observed in IL as compared with the reaction in
Hydrogenation of MPA in Ionic Liquids
The hydrogenation of the b-keto ester methyl propio- protic solvents. Similar results were obtained with the
nylacetate (MPA) to the corresponding hydroxy ester preformed chloride-containing complex B (entry 2),
(MHP) was studied as the model system and readily while the iodide-containing complex C led to lower
available ruthenium-BINAP complexes were selected enantioselectivity as well as activity (entry 3), reflect-
as the catalysts (Scheme 1). Literature data indicate ing the importance of the nature of the halide li-
that the hydrogenation of b-keto esters in the pres- gands.[17] Halide-free complexes D[18] and E[19] were
ence of ionic liquid (Figure 1) is possible although it twice to ten times more active than complex B, re-
is often necessary to add protic co-solvents in order spectively. Unfortunately, both complexes resulted in
to promote the reaction and/or to achieve acceptable very low enantioselectivities and in the case of com-
enantioselectivities.[14] Only few attempts to use ionic plex E even reversed enantioselectivity was achieved,
liquid as the only solvent were undertaken and poor again underlining the pivotal role of the halide co-li-
gands for the catalyst performance.[17b]
Ionic liquids with different anions and cations were
then tested as solvents in the hydrogenation of MPA
using catalyst A (Table 2). Imidazolium-based ILs
gave superior results over ammonium-based solvents
(entries 1–4 vs. 5), but in general no drastic effects on
catalysis outcome were observed upon IL variation.
Scheme 1. Hydrogenation of the b-keto ester methyl propio-
nylacetate MPA.
Using [HO-EMIm]ACTHNUTRGNEUNG[BTA], a task-specific ionic liquid
bearing a hydroxy functionality and having a polarity
similar to that of methanol, did also not effect any im-
provement in the activity (entry 4). It was therefore
decided to explore the use of additives as a way to
improve the catalyst performance
Additives in the Hydrogenation of MPA in IL
Water is used as additive in the hydrogenation of keto
esters in alcoholic solvents to reduce acetal formati-
AHCTUNGTRENNUNG
on.[17a,20] The hydrogenation of ketones can even be
performed in pure water with high reaction rates and
Figure 1. Ionic liquids used in the hydrogenation of MPA as
solvent and/or additive.
very good enantioselectivities.[21] In ionic liquids, how-
Table 1. Hydrogenation of MPA in [dMEIm]ACTHNUTRGNEUNG
[BTA] with different Ru/(S)-BINAP precursors[a]
Entry
Catalyst
p(H2) [bar]
TOF [hÀ1]
Conv. [%]
ee [%]
_
_
1
2
[Ru
[Ru
[Ru
G
90
56
65
21
138
ꢀ660
45
52
17
99
95(S)
95(S)
89(S)
10(S)
13(R)
(p-cymene)ClPP]Cl, B
100
100
60
3
(p-cymene)IPP)]I, C
_
4[b]
5[c]
[Ru(h5-2,4-dimethylpentadienyl)PPCH3CN]
[BTA], D
[Ru
[BF4], E
60
ꢀ 99
_
[a]
PP=(S)-BINAP; IL=1.26 mL; MPA=1.88 mmol; MPA/Cat=500; T=608C, t=4 h.
[b]
[c]
t=3.6 h.
t=0.75 h.
210
asc.wiley-vch.de
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2013, 355, 209 – 219

Continuous-Flow Asymmetric Hydrogenation of the b-Keto Ester Methyl Propionylacetate
Table 2. Hydrogenation of MPA in different ILs.[a]
sulted in the formation of many side products
(entry 3), the use of p-TsOH led to an enhancement
of the reaction rate by a factor of 2.6 reaching a TOF
of 156 hÀ1 (entry 4). This increase in activity was ac-
companied by a slightly reduced enantioselectivity
and an ee of 81% was obtained. The addition of up to
40 equivalents of p-TsOH enhanced the reaction rate
by nearly 400% (entry 6), while higher amounts of
acid apparently did not affect further the reaction
outcome.
Entry IL
p(H2)
[bar]
TOF
[hÀ1]
Conv. ee[b]
[%]
[%]
1
2
3
4
5
N
N
ACHTUNGTRENNUNG
100
90
90
56
53
56
41
26
45
42
45
33
21
97
90
95
93
90
ACHTUNGTRENNUNG[BTA] 100
[a]
Catalyst=A; IL=1.26 mL; MPA=1.88 mmol; MPA/A=
500; T=608C, t=4 h.
S-enantiomer.
The same effect as with p-TsOH could be achieved
when the acid-functionalised IL 1-butyl-3-(4-sulfo-
[b]
AHCTUNGTREGbNNNU utyl)-imidazolium bis(trifluoromethanesulphonyl)-
ACHUTNGRENaNUG mide ([HSO3BBIm]ACHTUGNTREN[NUGN BTA]) was employed as acidic
ever, even small amounts of water can have a strong additive[25] resulting in increased reaction rates
negative effect on catalyst activity and even more on (TOF=102 hÀ1) and slightly lowered enantioselectivi-
enantioselectivity.[15b,22] Indeed, the addition of de- ty (82% ee, entry 7).
gassed water (4.0 equiv. with respect to the catalyst)
Chloride-containing additives were also tested in
AHCTUNGERTG[NNUN BTA] (water content: the hydrogenation of MPA. Higher TOF of 98 and
in virtually dry [dMEIm]
8 ppm) resulted in a drop in conversion and enantio- 85 hÀ1 rates were observed by adding 10–20 equiv. of
selectivity from 47% to 12% and 95% to 76% ee, re- [NBu4]Cl and [OMIm]Cl, respectively (entries 8 and
spectively (Table 3 entries 1 vs. 2). For this reason, 9). Slightly enhanced ee values of 97% were achieved
careful drying of all reaction components is essential confirming the beneficial effects of chloride
to obtain good and reproducible results in ionic liq- anions.[26,27]
uids in the present system.
Finally, a combination of p-TsOH and [NBu4]Cl
The beneficial effect of acidic additives in the hy- was used as additive to evaluate whether the activity
drogenation of b-keto ester was first reported by and the enantioselectivity can be simultaneously im-
King.[23] Addition of small amounts of hydrochloric proved. This mixture, however, did not lead to a syner-
acid significantly enhanced the reaction rate of the gistic effect and resulted in a TOF of 74 hÀ1 and ee
hydrogenation of methyl propionylacetate catalysed values of 82% (entry 10).
by [Et2NH2]
G
G
In summary, these results confirm the feasibility of
that other acidic additives like sulphuric acid and the use of catalyst A in a stationary IL phase, where-
para-toluenesulphonic acid (p-TsOH)[24] led also to by the role of the acidic additives requires special at-
significant improvements of reaction rates in conven- tention for activity and long-term stability.
tional solvents without any loss of enantioselectivity.
According to these findings, the addition of acids
was evaluated in the hydrogenation of MPA in ionic
liquid solutions. Whereas the addition of tetrafluoro-
ACHTUNGTRENNUNGboric acid caused a reduced catalyst activity and re-
Table 3. Hydogenation of MPA in [dMEIm]ACTHNUTRGNEUNG
[BTA] with different additives.[a]
Entry
Additive (equiv.)
t [h]
TOF [hÀ1]
Conv. [%]
ee[b] [%]
1
–
4
4
4
1
1
1
4
4
4
1
59
15
43
156
237
270
102
98
47
12
34
62
48
55
82
77
65
15
95
76
69
81
80
80
82
97
97
82
2[c]
3[d]
4
H2O (4)
HBF4·Et2O (11)
p-TsOH (11)
p-TsOH (20)
p-TsOH (40)
U
ACHTUNGTRENNUNG
5
6
7
8
9
10
ACHTUNGTRENN[UNG BTA] (11)
C
85
74
A
[a]
IL=1.26 mL; Catalyst=A; MPA=1.88 mmol; MPA/A=500:1; T=608C; p(H2)=90 bar.
S-enantiomer.
p(H2)=100 bar; H2O/A=4:1.
Side products formed.
[b]
[c]
[d]
Adv. Synth. Catal. 2013, 355, 209 – 219
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
211

Jens Theuerkauf et al.
FULL PAPERS
Phase Behaviour and Partitioning in the System IL–
CO2
A sufficient CO2-solubility of the reagents as well as
a favourable partitioning of the products between
mobile and catalytic phase are prerequisites for the
use of CO2 as transport medium in a continuous flow
system. Organic compounds pressurised with CO2 can
exhibit complex phase behaviour, which have been
classified following practical criteria.[28] Upon pressur-
ising the substrate MPA with CO2 to p=80 bar at T=
408C, a large volume expansion of up to 500% was
observed indicating that MPA clearly belongs to class
II liquids. At higher CO2 pressures only one homoge-
neous phase could be observed.
Figure 2. Variation of CO2 partial pressure in the hydrogena-
tion of MPA in [dMEIm]ACTHUNGTRENNG[U BTA], V(IL)=1.03 mL, catalyst=
A, MPA=3.62 mmol, MPA/A=500:1, T=608C, t=4 h,
90 bar H2. Error bars determined from the standard devia-
tion of 3 experiments.
The partitioning of product and substrate between
ionic liquid and carbon dioxide expressed by the par-
tition coefficient KN was determined via high-pressure
NMR. A high-pressure sapphire NMR tube (Vtot =
0.93 mL) was loaded with 0.35 mL of an equimolar
mixture of MPA and MHP in the ionic liquid CO2. Enhanced hydrogen solubility[30] and reduced
[dMEIm]ACHTUNGTRENNUNG
[BTA] (ci =1.25 mol/L) mimicking a reaction viscosity of the ionic liquid phase[31] leading to im-
mixture at 50% conversion. The total substance con- proved mass transfer can explain this trend on the
centration in the IL was chosen to be of the same basis of increased hydrogen availability in the reactive
order of magnitude used later on in the continuous phase. Increasing the partial pressure further resulted
reactions. 1H NMR spectra were recorded at 408C in lower conversion (51%). This can be rationalised
prior and after pressurisation with CO2 (86 bar). The by the increased density of CO2, resulting in partial
residual concentration of both compounds in the ionic extraction of the substrate from the catalyst to the
liquid (bottom phase) after pressurisation with CO2 carbon dioxide phase in accordance with the partition
was measured using the signals of the ionic liquid as coefficient data (vide supra). The enantioselectivity
an internal standard assuming a negligible volume ex- was not affected by the presence of carbon dioxide
pansion of the IL-phase in the presence of CO2.[29] and ~95% ee was obtained over the whole pressure
The concentration in the CO2 phase was deduced range examined (Figure 2).
from the difference. Applying Eq. (1), the partition
The positive effect of CO2 at moderate pressures
was even more pronounced when the hydrogenation
was carried out in the presence of p-TsOH (25 equiv-
alents). After 1 h reaction time in presence of carbon
dioxide (50 bar) a substrate conversion of 73% was
reached corresponding to a four-fold increase of the
TOF to 363 hÀ1. Again a lower enantioselectivity was
coefficients KN,MPA =0.18 and KN,MHP =0.23 were de- obtained in the presence of the acid (ee 82%).
termined. These data indicate that both compounds
are dissolved preferentially in the ionic liquid phase.
However, the product displays a higher affinity (ca. Reactor Set-up for Continuous Flow Hydrogenation
30%) for the CO2-rich phase with respect to the sub-
strate. These solubility profiles meet the most favour- Based on previous experience,[32] a dedicated reactor
able situation for continuous flow operation, where set-up was built for the continuous flow hydrogena-
the product will be extracted preferentially by the tion experiments (Figure 3). The program LabViewꢄ
mobile phase from the catalytic phase.
(National Instruments) was used to record all data,
control flows and pressures, and to monitor safety pa-
rameters for automatic shutdown procedure. Sub-
strate delivery was accomplished by a double piston
HPLC pump capable of realising flows down to
Hydrogenation in Presence of CO2
The effect of carbon dioxide on the hydrogenation of 0.001 mLminÀ1 (E-5) and monitored by recording the
MPA in [dMEIm]ACTHUNRTGNEUNG[BTA] was examined in the next weight loss of the substrate container (I-6). Carbon
step (Figure 2). Conversion after 4 h reaction time in- dioxide flow was controlled via a heated, compressed
creased from 55% to 78% upon addition of 60 bar air regulated needle membrane valve connected to
212
asc.wiley-vch.de
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2013, 355, 209 – 219

Continuous-Flow Asymmetric Hydrogenation of the b-Keto Ester Methyl Propionylacetate
Figure 3. Flow scheme of the reactor for continuous flow reactions. E-7: CO2 dosing unit, MFC: hydrogen mass flow control-
ler, E-2: window reactor, E-9: back pressure regulator, E-12: gas-liquid separator, E-10: Fraction collector.
a calibrated liquid-flow device (E-7). Hydrogen flow CO2 and H2 flows, and total pressure, the delivery of
was regulated by a mass-flow controller (I-10).
MPA was started with a rate of 0.6 g/h. Under these
A stainless steel autoclave (10 mL) equipped with conditions, a residence time of 17 min can be estimat-
two thick-glass windows and a magnetic stirring bar ed for the 10-mL reactor volume.[34]
was used as central reactor unit (E-2). Capillaries for
Figure 4 shows the results obtained without any ad-
incoming and outcoming flow were connected to the ditive at two different catalyst loadings (run 1:
upper part of the autoclave. The total pressure in the 15 mmol; run 2: 30 mmol). Conversion reached a maxi-
system was controlled via a back pressure regulator mum after 5 h time-on-stream (18% in run 1, 35% in
consisting of a compressed air regulated and heated run 2) and slowly decreased afterwards. At the begin-
needle membrane valve (E-9).
ning, high enantioselectivities of up to 96% ee were
The product was separated from the effluent CO2 obtained, which also decreased in the course of the
flow by customised gas-liquid separator (E-12). The experiments (Figure 4). The average TOFav calculated
use of the separator allowed for continuous time-re-
solved sampling of the product mixture in a fraction
collector (E-10). The outgoing gas stream from the
separator was then passed through a cooling trap con-
taining toluene at À808C (E-1). Mass balance could
be closed up to 95% at the end of each continuous
experiment.
Continuous Hydrogenation Experiments
In a preliminary experiment, the density of the
mobile phase (0.49Æ0.01 gmLÀ1) and the partial pres-
sure of H2 (pH2 of 57Æ3.5 bar) were determined
under the typical conditions used for the continuous
flow catalytic experiments (H2 flow=3.3 Ls/h, CO2
Figure 4. Continuous hydrogenation of MPA in [dMEIm]
[BTA]/scCO2 in the absence of additives. Conditions: IL=
flow=9.3 Ls/h, T=608C; ptot =250 bar).[33] Catalyst A
was dissolved in [dMEIm]ACTHUNTGRNEUNG[BTA] (3 mL), MPA
3 mL; V/t(H2)=3.3 Ls/h, V/t
0.6 g/h, T=608C, ptot =250 bar; open symbols: A=15 mmol;
introduced in the reactor. After setting temperature, closed symbols: A=30 mmol.
ACHTUGTNRENGUN(CO2)=9.3 Ls/h, m/tACHTUNTGREN(NUGN MPA)=
(1.6 mL, S/C=500) added, and the resulting solution
Adv. Synth. Catal. 2013, 355, 209 – 219
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
213

Jens Theuerkauf et al.
FULL PAPERS
from the slope of accumulated turnovers in the first
8 h on stream was identical in both runs (49 hÀ1).
Again, no phosphorus and ruthenium could be de-
tected in the collected samples via ICP-OES measure-
1
Product samples obtained in the first phase of the ment. In the H NMR spectrum of the ionic liquid so-
reaction had a slightly yellow colour, but no IL con- lution recorded after reaction, the signals of the aro-
1
tamination could be detected by H NMR spectrosco- matic protons of p-TsOH at d=7.38 and 7.75 ppm
py. Similarly, ruthenium content in the product frac- were no longer present. This indicates that the addi-
tion never exceeded the detection limit of 5 ppm of tive was depleted from the catalyst phase by CO2 ex-
the ICP-OES analysis. High retention of Ru-BINAP traction over the extended reaction period.
species was confirmed by 31P NMR analysis of the re-
To prevent extraction of the additive, the acid-func-
actor content showing a set of signals in the character- tionalised IL [HSO3BBIm][BTA] was used (Figure 6).
AHCTUNGTRENNUNG
istic region between 35–45 ppm. No evidence for the Gratifyingly, an even higher single pass conversion of
formation of phosphine oxides[35] or for trimeric Ru- 95% was observed with 80% ee after 10 h time-on-
BINAP complexes[17b] could be detected, ruling out stream, corresponding to a maximum TOF of 138 hÀ1
such transformations as major source for deactivation. and STY of 168 gLÀ1 hÀ1. Switching off the stirrer in
Similar to the situation under batch conditions, the the reactor unit led to a rapid but fully reversible de-
addition of p-TsOH as acidic additive strongly en- crease of turnover rate, reflecting the importance of
hanced catalyst activity and stability. In the presence mass transport in the biphasic reaction system.
of 5 equiv. of p-TsOH, conversions >50% were
The permanent immobilisation of the acidic addi-
1
achieved over 165 h reaching a maximum value of tive was confirmed by H NMR analysis of the sta-
73% after 55 h (Figure 5, open symbols). The TOFav tionary phase after 140 h. Nevertheless, the deactiva-
(101 hÀ1) was increased two-fold as compared to ex- tion behaviour was very similar to that observed with
periments without additives corresponding to a space- p-TsOH, as conversion had steadily decreased to 40%
time yield (STY) of 114 gLÀ1 hÀ1.[36] Notably, nearly (TON=18,900) at this stage with a constant ee of 80–
constant levels of enantioselectivity between 85% and 82%. These results indicate that the acid plays
79% ee were obtained over the entire experiment.
a major role in the initial phase of the reaction for
By doubling the amount of p-TsOH to 10 equiv.cat the formation of the active species, but has a less
a further increase in activity (TOFav =122 hÀ1) and direct influence on its performance and stability
catalyst lifetime could be obtained. After a short in- under turnover conditions.
duction period, stable conversion of 81–83% was ob-
Similarly to all previous runs, also in this continu-
served for over 70 h time-on-stream. Afterwards, the ous flow experiment no traces of ionic liquid or cata-
1
conversion decreased similarly to the previous experi- lyst components were detected via H and 31P NMR
ment but was still 40% after 185 h time-on-stream. in the collected samples. The analysis of the catalyst
Again, stable enantioselectivities around 80% ee were solution before and after the continuous flow experi-
registered over the entire experiment. A total turn- ment showed no difference in metal and phosphorus
over number (TON) of ~19,000 with a STY of content within the accuracy of the measurement and
132 gLÀ1 hÀ1 was achieved in this experiment (TON= Ru and P mass-balance could be closed for all experi-
21,000 and STY=149 gLÀ1 hÀ1 including residual ments within the error margins.
product in the reactor and cooling traps).
As ligand oxidation, cluster formation, catalyst
leaching and additive depletion can all be ruled out as
Figure 5. Continuous hydrogenation of MPA with p-TsOH
as additive. Conditions: see Figure 4, A=30 mmol. Open
symbols: p-TsOH=5 equiv.; closed symbols: p-TsOH=
10 equiv.
Figure 6. Continuous
[HSO3BBIm][BTA] as additive. Conditions: see Figure 4,
A=30 mmol; [HSO3BBIm][BTA]=10.5 equiv.
hydrogenation
of
MPA with
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
214
asc.wiley-vch.de
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2013, 355, 209 – 219

Continuous-Flow Asymmetric Hydrogenation of the b-Keto Ester Methyl Propionylacetate
Figure 7. Continuous hydrogenation of MPA in [EMIm]
[BTA]=10 equiv. Variation of reaction conditions every 8–10 h. Standard conditions (I, III; V, VII, IX): V/t(H2)=3.3 Ls/h,
V/t(CO2)=10.5 Ls/h, m/t(MPA)=0.6 g/h, T=608C, ptot =250 bar. Variations: (II): T=708C, (IV): V/t(H2)=1.7 Ls/h; (VI): V/
t(H2)=1.7 Ls/h, V/t(CO2)=5.4 Ls/h; (VIII): m/t(MPA)=0.9 g/h, T=708C.
ACHTUNGTRNE[UNNG BTA]/scCO2. Conditions: A=14 mmol; IL=3 mL [HSO3BBIm]
A
ACHTUNGTRENNUNG
A
ACHTUNGTRENNUNG
major source for deactivation, the introduction of achieved. From the accumulated TON an average
water into the reactive phase remains most likely the TOF of 203 hÀ1 was calculated. Raising the tempera-
limiting factor. Water is constantly delivered as a con- ture by 108C in period II led to an increase of conver-
taminant of the substrate (60–110 ppm) and of the sion to 95% (TOF 290 hÀ1) accompanied by a drop in
CO2 stream (ꢁ5 ppm) contacting or even accumulat- ee to 70%. Restoring initial conditions after 10 h
ing in the catalyst phase. A 25-fold excess relative to (period III) resulted in similar conversion (71%) and
the catalyst amount is estimated over a typical run. slightly reduced enantioselectivity level (ee 77%) as
As shown in the batch experiments, water is deleteri- before in period I.
ous for the catalyst performance and even a four-fold
In the next step (period IV), hydrogen flow was set
excess resulted already in a clear depletion of the cat- to 1.7 Ls/h corresponding to a reduction of the H2 par-
alyst activity. Thus, water is most probably the main tial pressure by 50% in the mobile phase. Conversion
source of catalyst deactivation in the long term.
dropped to ~33% over 10 h, consistent with a first
To explore the limits for optimisation, the response order rate dependence in hydrogen pressure. After
of the system to the variation of different reaction pa- switching back to standard conditions (period V),
rameters was investigated. Temperature, hydrogen a maximum conversion of only 52% was observed (cf.
flow, and residence time were varied in 8–10 h peri- period V and III). This indicates that catalyst stability
ods. Conversion, enantioselectivity, and the phase be- is somewhat reduced under lower H2 pressure.
haviour in the reactor at the end/beginning of
Next, carbon dioxide and hydrogen flows were re-
a period were monitored and are depicted in Figure 7. duced by 50% leaving the substrate flow unchanged
The catalyst concentration was set to 14 mmol to (period VI). Visual inspection revealed a constant rise
enable positive and negative variation of single pass of the filling level in the reactor resulting from insuffi-
conversion.
cient extraction of the substrate/product mixture
In the first 10 h time-on-stream (period I) a maxi- under these conditions. Thus, the actual residence
mum conversion of 69% and an ee value of 83% were time in the reactor is no longer determined by the
Adv. Synth. Catal. 2013, 355, 209 – 219
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
215

Jens Theuerkauf et al.
FULL PAPERS
Table 4. Calculated figures for the production of 100 kg of
flow rates, and therefore no steady operating mode
could be adjusted. The conversion measured at the
outlet increased steadily, but this apparent “activa-
tion” reflects merely the increasing contact time be-
tween substrate and catalyst.
By switching back to standard conditions (period
VII) the product mixture could be extracted and the
volume of the catalyst phase reduced to the initial
level. Similar conversion and enantioselectivity as at
the end of period V were observed, indicating no sig-
nificant catalyst deactivation during the last 20 h
time-on-stream.
MHP through batch[16b] and continuous-flow hydrogenation.
Batch
Continuous flow
Number of batch
Reactor unit
4
50 L
100 L
350 g
1
5.4 L
2.7 L
25 g
Reaction solvent[a]
Catalyst A[a]
Reaction time
Processing time]
Pressure
4ꢅ6 h
1–2 weeks
4 bar (H2)
1008C
96 h
<1 week
250 bar (H2/CO2)
608C
Temperature
[a]
Cumulative values for four batches.
In period VIII, the substrate flow was increased to
0.9 ghÀ1 and the temperature to 708C, under other-
wise identical conditions. As expected, enantioselec-
tivity dropped from 71% to 63%, whereas single pass almost quantitative single pass conversion and 80–
conversion increased to 71%. This corresponds to an 82% ee. No contamination of IL or catalyst compo-
apparent TOF of 307 hÀ1 that exceeds the initial maxi- nents was detected in the product. The initial catalyst
mum despite the significant deactivation resulting activity was retained at 91% after 100 h and 69%
from more than 70 h time-on-stream. Again a slight after 150 h time-on-stream, with traces of water as the
volume increase of the catalyst phase was observed at most likely source for deactivation.
the end of period VIII indicating that the extraction
capacity of the mobile phase was not able to balance 100 kg of product can be produced within less than
the incoming flow completely. a week using 25 g of Ru/BINAP catalyst in 2.7 L of
Based on these data, it can be estimated that
Restoring the initial conditions (period IX) led to IL in a simple compact processing unit comprising
a decrease in conversion and slightly higher enantio- a 5.4-L reactor (Table 4). To yield the same amount
selectivity. The comparison of the conversion values of product under conventional condition, four batches
at the end of zone I (69%) and zone IX (45%) indi- of 6 h reaction time in a 50-L reaction vessel would
cates a rate of catalyst deactivation by 35% over the be necessary, each requiring 87.5 g of catalyst and
80 h time-on-stream, which is fully consistent with the 25 L of MeOH as solvent. Including all processing
results obtained under constant conditions.
steps of make-up and downstream purification com-
prising separation of solvent and catalyst, a conserva-
tive estimate leads to a total working time of approxi-
mately one to two weeks to isolate the product. Thus,
the continuous-flow process is considerably more re-
Conclusions
The present study demonstrates that enantioselective source efficient and requires much less man power for
hydrogenation of b-keto esters can be conducted operation. It is conducted at considerably higher pres-
under continuous-flow conditions in ionic liquid/ sure but this is offset largely by a ten times reduced
scCO2 biphasic systems using established Ru-BINAP equipment size.
catalysts. Using methyl propionylacetate as example
Whereas the mobile phase is vented in the labora-
for this class of compounds, a representative method- tory set-up described here, it could be easily recycled
ology for a systematic approach to successful imple- at the scale of Table 4. Commercially available equip-
mentation has been developed. The key elements of ment designed for small-scale natural product extrac-
this strategy are the integrated evaluation and optimi- tion could be readily adapted to this kind of process-
sation of the catalytic system on the molecular level ing scale. A major factor of operating costs and
(catalyst and transformation), the meso-scale (solubil- energy requirement are the compression cycles for
ity and phase behaviour,) and the macro-scale (reac- CO2 between isolation and reaction.[37] Using temper-
tor unit and process scheme).[6a] This approach can be ature as separation variable[7] or adsorption and ab-
readily adopted to other targets of this substrate class, sorption processes[4b] would allow nearly isobaric op-
providing an efficient general outline for the valida- eration conditions, however.
tion of continuous-flow methods in specific synthetic
applications.
The major drawback of the present case study is
the reduced enantioselectivity of the Ru/BINAP cata-
The highest activity and stability was observed with lyst in the acidic IL matrix (80–82% vs. 97–98% in
a catalyst prepared according to the Noyori in situ methanol). Although acceptable for certain agro-
method[16] using [dMEIm]
and [HSO3BBIm]
space-time yields of 180 gLÀ1 hÀ1 were obtained at purity for pharmaceutical end products.
[BTA] as stationary phase chemicals[38] and APIꢃs intermediates,[39] this may ne-
ACHTUNGTRENNUNG
AHCTUNGTRENGN[UN BTA] as acidic additive. Maximum cessitate more efforts to obtain the required optical
216
asc.wiley-vch.de
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2013, 355, 209 – 219

Continuous-Flow Asymmetric Hydrogenation of the b-Keto Ester Methyl Propionylacetate
valve, pressure read-out, and two thick-glass borosilicate
Although a detailed economic analysis will depend
strongly on an individual company framework, these
basic data suggest that continuous-flow asymmetric
hydrogenation in IL/scCO2 can offer a viable process
alternative to classical batch operation. Critical fac-
tors to be addressed include the product extraction ef-
ficiencies and the method for product isolation from
the CO2 stream on a process level and the control of
catalyst performance by immobilisation matrix on
a molecular level.
windows allowing for visual observation. After heating the
autoclave at 608C, it was pressurised with CO2 and then
with H2, and the stirrer was switched on (750 rpm). After
the reaction time was over, the autoclave was cooled to
room temperature, slowly vented, and the reaction mixture
analysed.
General Procedure for Continuous Flow Experiments
The system was purged with a flow of carbon dioxide for at
least 2 h at 120 bar to exclude any oxygen from the appara-
tus (Figure 3). Afterwards, pressure was reduced to 0.2 bar
and a preformed mixture of precatalyst, additive and sub-
strate (2 mL) dissolved in ionic liquid (3 mL) was filled into
the reactor. The system was pressurised with carbon dioxide
to 250 bar. Substrate, hydrogen, and carbon dioxide flows
were set to the desired value, and stirrer (500 rpm) and
product sampling at fraction collector was started. See cap-
tions of Figure 4 to Figure 7 for details.
Experimental Section
Caution: High-pressure experiments with compressed gases
are potentially dangerous and must be carried out with suita-
ble equipment under appropriate safety regulation only.
All manipulations of air- and moisture-sensitive material
were carried out under an argon atmosphere using Schlenk
techniques. Solvents were dried and purified according to
standard methodologies prior to use. The gases H2
(99.999%) and CO2 (99.995%) were used without further
purification. Methyl propionylacetate was dried over 3 ꢆ
molecular sieves, decanted, distilled and degassed via three
freeze-pump-thaw cycles prior to use. Dry methanol was
used as received and stored under argon on 3 ꢆ molecular
Determination of Conversion and Enantiomeric
Purity
In batch experiments, conversion was determined by
1H NMR spectroscopy of the crude reaction mixture. In con-
tinuous flow experiments, conversion was determined by
1
GC analysis and validated by H NMR spectroscopy.
sieves. [dMEIm]
according to the literature.[40] [EMIm]
and [Me3BuN][BTA] were purchased from Solvent Innova-
tion (Merck). [HO-EMIm]ACTHNUGTENR[NUG BTA] was prepared according to
N
The enantiomeric purity of MHP was determined from
the corresponding TMS-derivative using following proce-
dure: an aliquot of the product mixture (ca. 4 mg) was dis-
solved in CH2Cl2 (1 mL), treated with N-methyl-N-(trime-
thylsilyl)trifluoroacetamide (50 mL) and analysed via GC
(Chiraldex CB).
E
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
5 ppm except for the case of [HOEMIm]ACHTNUTRGNEUNG[BTA] (340 ppm).
Water content was determined by a Metrohm KF 756 as
average from two independent measurements. NMR spectra
were measured on a Bruker AV-400 spectrometer. Chemical
shifts are given relative to TMS by using the solvent signal
Acknowledgements
We thank the German Ministry of Science and Education
(BMBF: WING program ChirAmAl) and the European
Union (FP-7 integrated project SYNFLOW, http://synflo-
w.eu) for financial support, and Heraeus for the donation of
precious metal precursors. Furthermore, we thank Jennifer
Julis and Andreas Kapelski for experimental support and the
group of Prof. A. Behr (TU Dortmund) for ICP-OES meas-
urements.
1
as internal reference for H NMR as well as for 13C NMR
and H3PO4 (85%) as external reference for 31P NMR spec-
troscopy. Ruthenium and phosphorus content were deter-
mined via ICP-OES (Thermo Scientific, Iris Intrepid) at the
University of Dortmund.
Catalyst Preparation
Catalyst A was prepared in situ following a reported proce-
dure.[16] Volatiles were removed under reduced pressure and
the resulting solid material was used without further purifi-
cation. Complexes B,[17b] C,[17b] D[18] and E[19] were prepared
according to literature procedures.
References
[1] J. G. de Vries, C. J. Elsevier, (Eds.), Handbook of Ho-
mogeneous Hydrogenation, Wiley-VCH, Weinheim,
2006.
[2] H. U. Blaser, E. Schmid, Asymmetric catalysis on indus-
trial scale, Vol. 1, Wiley-VCH, Weinheim, 2004.
[3] a) D. J. Cole-Hamilton, R. P. Tooze, (Eds.), Catalyst
Separation, Recovery and Recycling – Chemistry & Pro-
cess Design, Springer, Heidleberg, Berlin, 2006; b) D. J.
Cole-Hamilton, Science 2003, 299, 1702–1706.
General Procedure for Batch Catalysis
Catalyst A (3.5 mg 3.75 mmol), MPA (0.24 mL, 1.88 mmol),
ionic liquid (1.26 mL) and, additive were placed in a Schlenk
tube and stirred for at least 1 h until a homogenous orange
solution was obtained. The solution was transferred via a sy-
ringe to a 10-mL stainless steel autoclave, which was previ-
ously evacuated and purged with argon (3 cycles). The auto-
clave was equipped with a magnetic stirring bar, needle
[4] a) Multiphase Homogeneous Catalysis, 1st edn., (Eds.:
B. Cornils, W. A. Herrmann, I. T. Horvꢇth, W. Leitner,
S. Mecking, H. Olivier-Bourbigou, D. Vogt), Wiley-
Adv. Synth. Catal. 2013, 355, 209 – 219
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
217

Jens Theuerkauf et al.
FULL PAPERS
VCH, Weinheim, 2005; b) Handbook of Green Chemis-
try, Vol. 4, Supercritical Solvents, (Eds.: W. Leitner,
P. G. Jessop), Wiley-VCH, Weinheim, 2010.
3064–3076; b) K. Mashima, T. Hino, H. Takaya, J.
Chem. Soc. Dalton Trans. 1992, 2099–2107; c) K. Mashi-
ma, K. Kusano, T. Ohta, R. Noyori, H. Takaya, J.
Chem. Soc. Chem. Commun. 1989, 1208–1210.
[18] J. A. Wiles, C. J. A. Daley, R. J. Hamilton, C. G. Leong,
S. H. Bergens, Organometallics 2004, 23, 4564–4568.
[19] A. Doppiu, A. Salzer, Eur. J. Inorg. Chem. 2004, 2244–
2252.
[20] a) A. Wolfson, I. F. J. Vankelecom, P. A. Jacobs, J. Or-
ganomet. Chem. 2005, 690, 3558–3566; b) L. Bartek, M.
Drobek, M. Kuzma, P. Kluson, L. Cerveny, Catal.
Commun. 2005, 6, 61–65.
[5] W. Leitner, Acc. Chem. Res. 2002, 35, 746–756.
[6] a) U. Hintermair, G. Franciꢀ, W. Leitner, Chem.
Commun. 2011, 47, 3691–3701; b) M. I. Burguete, E.
Garciꢇ-Verdugo, S. V. Luis, Beilstein J. Org. Chem.
2011, 7, 1347–1359.
[7] T. Harwardt, G. Franciꢀ, W. Leitner, Chem. Commun.
2010, 6669–6671.
[8] M. T. Reetz, W. Wiesenhçfer, G. Franciꢀ, W. Leitner,
Adv. Synth. Catal. 2003, 345, 1221–1228.
[9] a) M. F. Sellin, P. B. Webb, D. J. Cole-Hamilton, Chem.
Commun. 2001, 781–782; b) 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–
15588; c) A. C. Frisch, P. B. Webb, G. Zhao, M. J. Mul-
doon, P. J. Pogorzelec, D. J. Cole-Hamilton, Dalton
Trans. 2007, 47, 5531–5538; d) U. Hintermair, G. Zhao,
C. C. Santini, M. J. Muldoon, D. J. Cole-Hamilton,
Chem. Commun. 2007, 1462–1464; e) U. Hintermair, Z.
Gong, A. Serbanovic, M. J. Muldoon, C. C. Santini,
D. J. Cole-Hamilton, Dalton Trans. 2010, 39, 8501–8510.
[10] U. Hintermair, T. Hçfener, T. Pullmann, G. Franciꢀ, W.
Leitner, ChemCatChem 2010, 2, 150–154.
[11] A. Bçsmann, G. Franciꢀ, E. Janssen, M. Solinas, W.
Leitner, P. Wasserscheid, Angew. Chem. 2001, 113,
2769–2771; Angew. Chem. Int. Ed. 2001, 40, 2697–2699.
[12] R. Duque, E. ꢈchsner, H. Clavier, F. Caijo, S. P. Nolan,
M. Mauduit, D. J. Cole-Hamilton, Green Chem. 2011,
13, 1187–1195.
[13] a) H. Shimizu, I. Nagasaki, K. Matsumura, N. Sayo, T.
Saito, Acc. Chem. Res. 2007, 40, 1385–1393; b) M. A.
Schwindt, M. P. Fleming, Y.-K. Han, L. M. Hodges,
D. A. Johnston, R. P. Micheli, C. R. Roberts, R. Snyder,
R. J. Topping, K. Pꢁntener, M. Scalone, Org. Process
Res. Dev. 2007, 11, 524–533; c) A. Korostylev, V. An-
drushko, N. Andrushko, V. I. Tararov, G. Kçnig, A.
Bçrner Eur. J. Org. Chem. 2008, 840–846; d) N. An-
drushko, V. Andrushko, V. Tararov, A. Korostylev, G.
Kçnig, A. Bçrner, Chirality, 2010, 22, 534–541.
[14] a) H. L. Ngo, A. Hu, W. Lin, Chem. Commun. 2003,
1912–1913; b) A. Hu, H. L. Ngo, W. Lin, Angew. Chem.
2004, 116, 2555–2558; Angew. Chem. Int. Ed. 2004, 43,
2501–2504; c) K. H. Lam, L. Xu, L. Feng, J. Ruan, Q.
Fan, A. S. C. Chan, Can. J. Chem. 2005, 83, 903–908;
d) E. ꢈchsner, K. Schneiders, K. Junge, M. Beller, P.
Wasserscheid, Appl. Catal. A: Gen. 2009, 364, 8–14;
e) E. V. Starodubtseva, M. G. Vinogradov, V. A. Pavlov,
L. S. Gorshkova, V. A. Ferapontov, Russ. Chem. Bull.
Int. Ed. 2004, 53, 2172–2174.
[21] a) M. Berthod, C. Saluzzo, G. Mignani, M. Lemaire,
Tetrahedron: Asymmetry 2004, 15, 639–645; b) T. La-
mouille, C. Saluzzo, R. ter Halle, F. Le Guyader, M.
Lemaire, Tetrahedron Lett. 2001, 42, 663–664.
[22] Often the negative effect of water was attributed to the
generation of HF by hydrolysis of fluorinated ionic liq-
uids anions like BF4 and PF6. In our experiments, how-
ever, FAP and BTA anions were used, which are re-
garded as inert against hydrolysis.
[23] S. A. King, A. S. Thompson, A. O. King, T. R. Verho-
even, J. Org. Chem. 1992, 57, 6689–6691.
[24] a) M. Kitamura, M. Yoshimura, N. Kanda, R. Noyori,
Tetrahedron 1999, 55, 8769–8785; b) A. E. Sollewijn
Gelpke, H. Kooijman, A. L. Spek, H. Hiemstra, Chem.
Eur. J. 1999, 5, 2472–2482; c) A. Wolfson, I. F. J. Vanke-
lecom, S. Geresh, P. A. Jacobs, J. Mol. Catal. A: Chem.
2004, 217, 21–26; d) V. A. Pavlov, E. V. Starodubtseva,
M. G. Vinogradov, V. A. Ferapontov, O. R. Malyshev,
G. L. Heise, Russ. Chem. Bull. Int. Ed. 2000, 49, 728–
731.
[25] F. M. A. Geilen, B. Engendahl, A. Harwardt, W. Mar-
quardt, J. Klankermayer, W. Leitner, Angew. Chem.
2010, 122, 5642–5646; Angew. Chem. Int. Ed. 2010, 49,
5510–5514.
[26] R. Noyori, T. Ohkuma, M. Kitamura, H. Takaya, N.
Sayo, H. Kumobayashi, S. Akutagawa, J. Am. Chem.
Soc. 1987, 109, 5856–5858.
[27] D. B. Williams, M. E. Stoll, B. L. Scott, D. A. Costa,
W. J. Oldham Jr, Chem. Commun. 2005, 1438–1440.
[28] U. Hintermair, W. Leitner, P. G. Jessop, in: Handbook
of Green Chemistry, Vol. 4, Supercritical Solvents,
(Eds.: W. Leitner, P. G. Jessop), Wiley-VCH, Wein-
heim, 2010, pp 103–120.
[29] S. N. V. K. Aki, B. R. Mellein, E. M. Saurer, J. F. Bren-
necke, J. Phys. Chem. B 2004, 108, 20355–20365.
[30] M. Solinas, A. Pfaltz, P. G. Cozzi, W. Leitner, J. Am.
Chem. Soc. 2004, 126, 16142.
[31] A. Ahosseini, E. Ortega, B. Sensenich, A. M. Scurto,
Fluid Phase Equilib. 2009, 286, 72–78.
[15] a) T. Floris, P. Kluson, L. Bartek, H. Pelantova, Appl.
Catal. A: Gen. 2009, 366, 160–165; b) M. Berthod, J.-M.
Joerger, G. Mignani, M. Vaultier, M. Lemaire, Tetrahe-
dron: Asymmetry 2004, 15, 2219–2221.
[16] a) M. Kitamura, M. Tokunaga, T. Ohkuma, R. Noyori,
Tetrahedron Lett. 1991, 32, 4163–4166; b) M. Kitamura,
M. Tokunaga, T. Ohkuma, R. Noyori, Org. Synth. 1993,
71, 1.
[32] a) C. Roosen, J.-L. Muller, M. Kaever, H. Kronenberg,
R. Thelen, S. Aey, T. Harwardt, W. Leitner, L. Greiner,
J. Supercrit. Fluids 2009, 48, 33–35; b) U. Hintermair,
C. Roosen, M. Kaever, H. Kronenberg, R. Thelen, S.
Aey, W. Leitner, L. Greiner, Org. Process Res. Dev.
2011, 15, 1275–1280.
[33] The density was directly determined by weighing the
reactor filled the mobile phase. The NIST Super-
TrappTM program was used to calculate the CO2/H2
ratio from the mixture density.
[17] a) K. Mashima, K.-h. Kusano, N. Sato, Y. Matsumura,
K. Nozaki, H. Kumobayashi, N. Sayo, Y. Hori, T. Ishi-
zaki, S. Akutagawa, H. Takaya, J. Org. Chem. 1994, 59,
218
asc.wiley-vch.de
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2013, 355, 209 – 219

Continuous-Flow Asymmetric Hydrogenation of the b-Keto Ester Methyl Propionylacetate
[34] Considering the high solubility of CO2 in the IL, the
entire reactor volume including the IL-phase was con-
sidered for the calculation of the residence time.
[37] C. M. Gordon, W. Leitner, in: Catalyst Separation, Re-
covery and Recycling – Chemistry & Process Design,
(Eds.: D. J. Cole-Hamilton, R. P. Tooze), Springer, Hei-
dleberg, Berlin, 2006, p 215.
[35] 2,2’-Bis(diphenylphosphino)-1,1’-binaphthyl monoxide:
31P NMR (CDCl3): d=À14.1 (s), 28.7 (s); 2,2’-bis(di-
[38] H.-U Blaser, W. Brieden, B. Pugin, F. Spindler, M.
Studer, A. Togni, Topics in Catalysis 2002, 19, 3–16.
[39] J. Klosin, C. R. Landis, Acc. Chem. Res. 2007, 40, 1251–
1259.
[40] J. D. Holbrey, W. M. Reichert, R. P. Swatloski, G. A.
Broker, W. R. Pitner, K. R. Seddon, R. D. Rogers,
Green Chem. 2002, 4, 407–413.
phenylphosphino)-1,1’-binaphthyl
(CDCl3): d=29.4 ppm.
oxide:
31P NMR
[36] In analogy to heterogeneous catalytic systems, we
define the space-time yield (STY) for biphasic systems
as the mass flow of the product per volume of the cata-
lyst phase (ghÀ1 LrVÀ1). For simplicity, the initial volume
of the ionic liquid phase is considered.
[41] K. Shimojo, N. Kamiya, F. Tani, H. Naganawa, Y.
Naruta, M. Goto, Anal. Chem. 2006, 78, 7735–7742.
Adv. Synth. Catal. 2013, 355, 209 – 219
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
219