Enantioselective hydrogenation of imines in ionic liquid/carbon dioxide media

Article abstract of DOI:10.1021/ja046129g

The enantioselective hydrogenation of N-(1-phenylethylidene)aniline using cationic iridium complexes with chiral phosphinooxazoline ligands was studied as a chemical probe to assess the potential of ionic liquid/carbon dioxide (IL/CO₂) media fo

Full text of DOI:10.1021/ja046129g

Published on Web 11/19/2004
Enantioselective Hydrogenation of Imines
in Ionic Liquid/Carbon Dioxide Media
Maurizio Solinas,† Andreas Pfaltz,§ Pier Giorgio Cozzi,| and Walter Leitner*,†,‡
Contribution from the Institut fu¨r Technische und Makromolekulare Chemie,
Worringer Weg 1, D-52074, Aachen, Germany, Max-Planck-Institut fu¨r Kohlenforschung,
Mu¨lheim an der Ruhr, Germany, Department Chemie, UniVersita¨t Basel, Basel, Switzerland, and
Dipartimento Chimico “G. Ciamician”, UniVersita` di Bologna, Bologna, Italy
Received June 30, 2004; E-mail: leitner@itmc.rwth-aachen.de
Abstract: The enantioselective hydrogenation of N-(1-phenylethylidene)aniline using cationic iridium
complexes with chiral phosphinooxazoline ligands was studied as a chemical probe to assess the potential
of ionic liquid/carbon dioxide (IL/CO2) media for, multiphase catalysis. The biphasic system leads to
activation, tuning, and immobilization of the catalyst that would be impossible in classical organic solvent
systems or in either of the two unconventional media separately. In particular it is demonstrated that (i) the
presence of CO2 can be beneficial or even mandatory for efficient hydrogenation in the IL; (ii) the precursor
is activated in the IL by anion exchange allowing one to use in situ catalysts; (iii) the anion of the IL greatly
influences the selectivity of the catalyst; (iv) the products are readily isolated from the catalyst solution by
CO2 extraction without cross contamination of IL or catalyst; and (v) the IL leads to enhanced stability of
the catalyst. These results are corroborated and rationalized on the basis of the physicochemical properties
of the biphasic medium and the chemical characteristics of the catalytic systems.
Scheme 1. Enantioselective Hydrogenation of N-(1-Phenylethyl-
idene)aniline (1) in IL/CO2 and Catalytic Systems Used in the
Present Studya
Introduction
Innovative methodologies for immobilization and efficient
recycling of organometallic catalysts are finding ever increasing
interest for the development of environmentally benign and
economically viable homogeneously catalyzed processes.1 Stimu-
lated by the seminal paper of Brennecke and Beckman2 on the
phase behavior of mixtures of ionic liquids (ILs)3 and super-
critical carbon dioxide (scCO2),4 several groups have shown
recently that the combination of these two media offers a highly
attractive approach to this problem.5,6 In the present paper, we
present our recent results on the activation, tuning, and
immobilization of chiral iridium catalysts in IL/CO2 for the
enantioselective hydrogenation reaction of imines (Scheme 1).7
The enantioselective hydrogenation of imines provides an
attractive route to enantiomerically enriched chiral secondary
a For the definition of ligands 5b-e, see Scheme 3.
† Institut fu¨r Technische und Makromolekulare Chemie.
‡ Max-Planck-Institut fu¨r Kohlenforschung.
amines as applied in the industrial production of (S)-meto-
lachlor.8 The chiral (phosphanodihydrooxazole)iridium catalyst
§ Universita¨t Basel.
| Universita` di Bologna.
(1) Chiral Catalyst Immobilization and Recycling; De Vos, D. E., Vankelecom,
I. F. J., Jacobs, P. A., Eds.; Wiley-VCH: Weinheim, 2000.
(2) Blanchard, L. A.; Hancu, D.; Beckman, E. J.; Brennecke, J. F. Nature 1999,
399, 28.
(3) (a) Welton, T. Chem. ReV. 1999, 99, 2071. (b) Wasserscheid, P.; Keim,
W. Angew. Chem., Int. Ed. 2000, 39, 3772. (c) Earle, M. J.; Seddon, K.
Pure Appl. Chem. 2000, 72, 1391. (d) Ionic Liquids in Synthesis;
Wasserscheid, P., Welton, T., Eds.; Wiley-VCH: Weinheim, 2003.
(4) (a) Chemical Synthesis Using Supercritical Fluids; Jessop, P. G., Leitner,
W., Eds.; Wiley-VCH: Weinheim, 1999. (b) Leitner, W. Acc. Chem. Res.
2002, 35, 746.
(5) (a) Bo¨smann, A.; Francio`, G.; Janssen, E.; Solinas, M.; Leitner, W.;
Wasserscheid, P. Angew. Chem., Int. Ed. 2001, 40, 2697. (b) Reetz, M. T.;
Wiesenho¨fer, W.; Francio`, G.; Leitner, W. Chem. Commun. 2002, 992. (c)
Reetz, M. T.; Wiesenho¨fer, W.; Francio`, G.; Leitner, W. AdV. Synth. Catal.
2003, 345, 1221. (d) Ballivet-Tkatchenko, D.; Piquet, M.; Solinas, M.;
Francio`, G.; Wasserscheid, P.; Leitner, W. Green Chem. 2003, 5, 232.
(6) (a) Brown, R. A.; Pollett, P.; McKoon, E.; Eckert, C. A.; Liotta, C. L.;
Jessop, P. G. J. Am. Chem. Soc. 2001, 123, 1254. (b) Liu, F.; Abrams, M.
B.; Baker, R. T.; Tumas, W. Chem. Commun. 2001, 433. (c) Sellin, M. F.;
Webb, P. B.; Cole-Hamilton, D. J. Chem. Commun. 2001, 781. (d) Lozano,
P.; de Diego, T.; Carrie, D.; Vaultier, M.; Iborra, J. L. Chem. Commun.
2002, 692. (e) Hou, Z. S.; Han, B. X.; Gao, L.; Jiang, T.; Liu, Z. M.; Chang,
Y. H.; Zhang, X. G.; He, J. New J. Chem. 2002, 26, 1246. (f) Webb, P. B.;
Sellin, M. F.; Kunene, T. E.; Williamson, S.; Slawin, A. M. Z.; Cole-
Hamilton, D. J. J. Am. Chem. Soc. 2003, 125, 15577. (g) Jessop, P. G.;
Stanley, R. R.; Brown, R. A.; Eckert, C. A.; Liotta, C. L.; Ngo, T. T.;
Pollet, P. Green Chem. 2003, 5, 123.
(7) Preliminary results of this study have been reported in part as a conference
proceeding: Solinas, M.; Wasserscheid, P.; Leitner, W.; Pfaltz, A. Chem.-
Ing.-Tech. 2003, 75, 1153.
(8) (a) Blaser, H.-U.; Spindler, F. Top. Catal. 1998, 4, 275. (b) Blaser, H.-U.
AdV. Synth. Catal. 2002, 344, 17.
9
16142
J. AM. CHEM. SOC. 2004, 126, 16142-16147
10.1021/ja046129g CCC: $27.50 © 2004 American Chemical Society

Enantioselective Hydrogenation of Imines
A R T I C L E S
Table 1. Enantioselective Hydrogenation of N-(1-Phenylethyli-
Scheme 2. Set of Cations and Anions Used in the Present Work
dene)aniline (1) to (R)-Phenyl-(1-phenyl-ethyl)-amine (2) in IL/CO2
Systemsa
T
H2
[bar]
CO2
[g]
conversion
[%]
ee
[%]
entry
IL
cat.
[
°
C]
t [h]
1
2
3
4
5
6
7
8
9
[EMIM][BTA]
[EMIM][BTA]
[EMIM][BTA]
[4MBP][BTA]
[4MBP][BTA]
[PMIM][PF6]
[PMIM][PF6]
[PMIM][PF6]
[PMIM][PF6]
[BMIM][BF4]
[EMIM][BARF]e
[EMIM][BARF]e
3
40
40 100
30
22
22
22
22
22
22
22
22
22
22
22
22
3
97
3b
58
56
53
52
61
65
62
64
30
78
76
3
3
3
3
40
40
40
40
25
40
30 8.9
30
>99
>99
>99
>99
>99
>99
>99
92
30 9.8
35 7.5
30 8.0
30 8.6
30 9.2
30 7.6
30 8.9
30 9.0
3
4/5ac
4/5ad 40
10
11
12
3
3
3
40
40
0
>99
54
a Standard conditions unless noted otherwise: 1/Ir ) 500/1, V(reactor)
) 12 mL, V(IL) ) 2.0 mL. b 1/Ir ) 100/1. c 1/5a/Ir ) 250/1/1, preformation
in CH2Cl2. d 1/5a/Ir ) 250/1/1, in situ protocol. e [EMIM][BARF] ) 0.7-
0.8 g.
3 yields high levels of asymmetric induction in the hydrogena-
tion of prochiral N-alkyl and N-aryl imines9 as well as in the
hydrogenation of carbon-carbon double bonds.10 Some of us
have recently shown that this type of complexes can be used as
homogeneous catalysts for the hydrogenation of imines in scCO2
upon appropriate modification especially of the counterion.11
Although the product and the catalyst could be readily separated
using the CESS (catalysis and extraction using supercritical
solvents) procedure, an efficient immobilization of the catalyst
was not possible because of noticeable deactivation during the
recycling process.
Here, we describe how the combination of ILs and scCO2
does not only allow one to overcome this problem, but permits
one to exploit fully the benefits of both systems through the
combination of the molecular interaction of the stationary IL
phase with the catalyst and the mass transfer properties of the
mobile CO2 phase. In particular, we observe that (i) the presence
of CO2 is beneficial or in certain cases even mandatory for
efficient hydrogenation in the IL at comparable partial pressures
of hydrogen; (ii) dissolving the catalyst in the IL leads to
activation by anion exchange and allows one to use in situ
systems that would not be applicable in conventional solvents;
(iii) the choice of the anion of the IL greatly influences the
selectivity of the catalyst; (iv) the products are readily removed
and isolated from the catalyst solution by CO2 extraction without
cross contamination of IL or catalyst; and (v) the IL leads to
greatly enhanced long-term stability of the iridium catalyst.
These results are corroborated and rationalized on the basis
of the physicochemical properties of the biphasic medium and
the chemical characteristics of the catalytic systems.
the study. In a typical experiment, the appropriate ionic liquid
(2.0 mL), the iridium complex 3 (3 × 10-3 mmol) and the
substrate 1 (1/3 ) 500/1) were loaded under argon in a window-
equipped stainless steel autoclave (V ) 12 mL). The reactor
was then pressurized with H2 and the desired amount of CO2
followed by heating under stirring to 40 °C for a standard
reaction time of 22 h. The products were collected for offline
GC and HPLC analysis from screening experiments by extrac-
tion of the IL phase with hexane after cooling and venting, or
alternatively isolated by CO2 extraction as described below.
Representative results are summarized in Table 1.
The ionic liquid ethyl-methyl-imidazolium bis(trifluoro-
methylsulfonyl)imide ([EMIM][BTA]) was used to assess the
influence of CO2 on the hydrogenation process using 3 as the
catalyst precursor. In the absence of CO2, conversion was only
marginal under standard conditions (p(H2) ) 30 bar, T ) 40
°C), and significant conversion required the use of hydrogen
pressures as high as 100 bar. However, quantitative formation
of the hydrogenation product was observed at a hydrogen partial
pressure of 30 bar in the presence of 8.9 g of CO2 under
otherwise identical conditions (Table 1, entries 1-3). Within
experimental error, there was no influence of CO2 on the
enantiomeric excess (Table 1, entries 2/3 and 4/5).
A similar beneficial effect of added CO2, albeit far less
dramatic, was observed with other ILs. Using the 4-methyl-N-
butyl pyridinium bis(trifluoromethylsulfonyl)imide ionic liquid
([4MBP][BTA]), quantitative conversion was observed with and
without CO2 under 30 bar of hydrogen within the standard
reaction time of 22 h (Table 1, entries 4 and 5). However, the
effect of the CO2 pressure was reflected again at shorter reaction
times. The hydrogenation of 1 in [4MBP][BTA] in the absence
of carbon dioxide with 30 bar of H2 pressure gave 25% of
conversion in 2 h reaction time (TOF ) 44 h-1), while at the
same condition, but with the presence of supercritical carbon
dioxide, 34% of conversion (TOF ) 60 h-1) was obtained
(Table 2, entries 3 and 4). At lower temperatures and in [BMIM]-
[PF6], the conversions again followed the same trend but were
too small to allow for quantitative conclusions. In this set of
screening experiments, the highest activities were achieved in
the presence of scCO2 at 40 °C in [PMIM][PF6], reaching
several hundreds turnover per hour at low catalyst loading.
In conventional solvents, the hydrogenation of 1 with catalyst
3 does not show a direct proportionality to the hydrogen
Results and Discussion
The hydrogenation of N-(1-phenylethylidene)aniline (1) to
give (R)-phenyl-(1-phenyl-ethyl)-amine (2) was used as a
benchmark reaction throughout the course of this study (Scheme
1). The set of cations and anions displayed in Scheme 2 was
used to assess the influence of the structural diversity of ILs
for the process. The cationic complex 3 containing [PF6]- as
the counterion was used as a precatalyst in the initial phase of
(9) Schnider, P.; Koch, G.; Pre´toˆt, R.; Wang, G.; Bohnen, F. M.; Kru¨ger, C.;
Pfaltz, A. Chem. Eur. J. 1997, 3, 887.
(10) (a) Lightfoot, A.; Schnider, P.; Pfaltz, A. Angew. Chem., Int. Ed. 1998,
37, 2897. (b) Blankenstein, J.; Pfaltz, A. Angew. Chem., Int. Ed. 2001, 40,
4445. (c) Blackmond, D. G.; Lightfoot, A.; Pfaltz, A.; Rosner, T.; Schnider,
P.; Zimmermann, N. Chirality 2000, 12, 442.
(11) Kainz, S.; Brinkmann, A.; Leitner, W.; Pfaltz, A. J. Am. Chem. Soc. 1999,
121, 6421.
9
J. AM. CHEM. SOC. VOL. 126, NO. 49, 2004 16143

A R T I C L E S
Solinas et al.
Figure 2. Solubility of hydrogen in [EMIM][BTA] at a constant partial
pressure p(H2) ) 30 bar as a function of the added CO2 pressure as
determined by high-pressure 1H NMR spectroscopy (T ) 297 ( 2 K;
concentrations below the estimated detection limit of 0.01 mol L-1 were
set to zero).
pressurized with 30 bar of H2, the characteristic signal of
dissolved dihydrogen (ca. 4.0-4.5 ppm) could not be detected
even after prolonged shaking to ensure that the equilibrium
concentration was reached (Figure 1, trace a). On the basis of
the sensitivity of the NMR method, we estimate that the
concentration of hydrogen must be below c(H2) ≈ 0.01 mol
L-1 under these conditions. This is in line with the low solubility
of H2 in ILs determined recently through independent investiga-
tions by Dyson et al.17 After carbon dioxide was added to the
system, reaching a total pressure of 120 bar, the signal related
to the molecular hydrogen was clearly observed with a chemical
shift of 4.3 ppm (Figure 1, trace b). At the same time, the
viscosity of the ionic liquid decreased drastically, resulting in
an improved resolution of the NMR peaks. Once the CO2
pressure was released, a complete disappearance of the hydrogen
signal was observed (Figure 1, trace c). From integration of the
dihydrogen signal and the signals of the methylene protons in
the imidazolium cation, the concentration of hydrogen in this
system could be estimated to c(H2) ) 0.14 mol L-1. This value
compares well with the solubility of hydrogen in common
organic solvents at the same conditions and is approximately
an order of magnitude larger than in common ionic liquids in
the absence of CO2.17
The high-pressure NMR technique allows one to quantify the
enhancement of hydrogen solubility in ionic liquid by the effect
of carbon dioxide more systematically. As shown in Figure 2,
the H2 concentration in [EMIM][BTA] was found to increase
continuously at constant hydrogen pressure (30 bar) with the
additional pressure of carbon dioxide in the sapphire NMR tube.
In the range of CO2 pressure investigated here, the net
concentration of molecular hydrogen in the IL expressed in [mol
L-1] follows with good approximation (R2 ) 0.989) an
exponential curve of the type:
Figure 1. Part of the high pressure 1H NMR spectra of [EMIM][BTA]
under hydrogen pressure of p(H2) ) 30 bar (trace a), upon addition of CO2
to a total pressure of 120 bar (trace b), and after release of the system
pressure (trace c); the measurements were performed at T ) 297 ( 2 K.
Table 2. Effect of the Carbon Dioxide Pressure on the Rate of
Imine Hydrogenation with Catalyst 3 in Various ILs
T
H2
[bar]
CO2
[g]
conversion
[%]
TOF
1
entry
IL
sub./cat.
[
°
C]
t [h]
[h-
]
1
2
3
4
5
6
7
8
[EMIM][BTA]
[EMIM][BTA]
[4MBP][BTA]
[4MBP][BTA]
[BMIM][PF6]
[BMIM][PF6]
[BMIM][PF6]
[PMIM][PF6]
500/1 40
500/1 40
350/1 40
350/1 40
360/1 25
360/1 25
450/1 40
1160/1 40
30
30
30
30
30
30
30
30
22
8.9 22
2
3
>99
25
34
5
8
92
14
<1
>20
44
60
18
29
138
327
8.6
2
1
1
3
8.8
8.6
7.2
0.5
pressure. The exact reaction order is not known. However, it
has been noticed that the chiral iridium catalysts used in the
present study react quite sensitive on hydrogen availability
influenced by mass transfer properties of the system.9,10c,12 The
apparent “activation” of the catalytic system upon adding CO2
can thus be related to an improved H2 availability in the IL,
which may result from a strongly increased H2-solubility6g,13,14
and/or reduced viscosity15 in the presence of compressed CO2.
In the present case, we were able to assess these features
qualitatively and quantitatively by using high-pressure NMR
spectroscopy (Figure 1).
When a high-pressure sapphire single-crystal NMR tube16 was
charged with a pure sample of [EMIM][BTA] and then
(12) Pfaltz, A.; Blankenstein, J.; Hilgraf, R.; Hormann, E.; McIntyre, S.; Menges,
F.; Schonleber, M.; Smidt, S. P.; Wustenberg, B.; Zimmermann, N. AdV.
Synth. Catal. 2003, 345, 33 and references therein.
(13) (a) Blanchard, L. A.; Gu, Z.; Brennecke, J. F. J. Phys. Chem. B 2001, 105,
2437. (b) Anthony, J. L.; Maginn, E. J.; Brennecke, J. F. J. Phys. Chem.
B 2002, 106, 7315.
(14) For the effect of CO2 on hydrogenation rates in organic solvents, see:
Combes, G. B.; Dehgani, F.; Lucien, F. P.; Dillow, A. K.; Foster, N. R. In
Reaction Engineering for Pollution PreVention; Abraham, M. A., Hesketh,
R. P., Eds.; Elsevier: Amsterdam, 2000.
(15) (a) Baker, S. N.; Baker, G. A.; Kane, M. A.; Bright, F. V. J. Phys. Chem.
B 2001, 105, 9663. (b) Liu, Z.; Wu, W.; Han, B.; Dong, Z.; Zhao, G.;
Wang, J.; Jiang, T.; Yang, G. Chem. Eur. J. 2003, 9, 3897.
(16) Elsevier, C. J. J. Mol. Catal. 1994, 92, 285.
CO2/b
[H2] ) [H2]0 + AeP
where [H2]0 is the concentration in the absence of CO2 (defined
(17) Dyson, P. J.; Laurenczy, G.; Ohlin, C. A.; Vallance, J.; Welton, T. Chem.
Commun. 2003, 2418.
9
16144 J. AM. CHEM. SOC. VOL. 126, NO. 49, 2004

Enantioselective Hydrogenation of Imines
A R T I C L E S
Scheme 3. Screening of a Small Ligand Library 5a-e for
Enantioselective Hydrogenation of 1 in IL/CO2 Using the in Situ
Catalyst 4/5a-e
here as zero) and the constants have been calculated as A )
(3.0 ( 0.9) mmol L-1 and b ) (23.8 ( 2.0) bar. Obviously,
the concentration would be finally limited by the total amount
of hydrogen available, which is approximately 1 mol L-1 within
the total volume of the tube.
As revealed quantitatively in Figure 2, the system [EMIM]-
[BTA]/H2/CO2 exhibits a remarkable deviation from Henry’s
law, which predicts the hydrogen solubility in the liquid solvent
to be only a function of the H2 partial pressure. It has been
demonstrated very recently that the solubility of hydrogen in a
series of pure ionic liquids does indeed obey this rule and is
directly proportional to the H2 pressure.17 In our experiments,
the partial pressure of hydrogen is fixed to 30 bar, but its solu-
bility in [EMIM][BTA] ironically increases as it is diluted with
CO2. This result can be related to the exceptional solubility
properties of carbon dioxide which is largely soluble in ionic
liquids13 and in all proportions miscible with hydrogen.4a Beyond
arguments related to the hydrogen solubility, it is known that
the viscosity of ionic liquids is decreased by the effect of the
carbon dioxide pressure.15 The enhancement of hydrogen avail-
ability in the ionic liquid resulting from the combined effects
in the presence of CO2 leads to a higher reaction rate or, as in
the case of [EMIM][BTA], even in a surpassing of the threshold
concentration to make the catalytic cycle possible. This net
“activation” of the catalytic system illustrates nicely the potential
benefits arising from the combination of the two media.
In addition to the physicochemical properties, interactions on
the molecular level can also be used to activate and tune the
catalytic system through the choice of the anion in the IL/CO2
medium. Again, the cationic iridium catalysts used in the present
study serve as an illustrative example, as they are known to
respond very sensitively to the nature of the counterions in some
cases.10,11,18 To exploit this tuning effect in conventional
solvents, the synthesis of individual catalyst precursors of type
3 with varying anions is necessary. Using ILs as catalyst support
opens the possibility to use the same precursor and to introduce
the variation of the counterion with the choice of the IL. This
requires that the counterion of the catalyst precursor is fully
dissociated and effectively “diluted” in the pool of the ionic
solvent system. As shown in Table 1, this is indeed the case,
and remarkable effects are observed on the stereoselectivity of
the reaction upon variation of the anion in the IL using the
[PF6]--containing precursor of 3 in all cases. The ee values vary
from 30% with [BF4]- up to 78% with [BARF]-. As a general
trend, the performance of the catalyst improves with decreasing
coordinating ability of the counterion of the IL.19
Two different methods for the direct generation of the catalyst
were tested. In the first case, 4 and 5a were mixed in CH2Cl2
for 1 h at 50 °C, the solvent was removed, and the chloride
iridium complex was redissolved in the appropriate ionic liquid.
In an in situ protocol, the desired amount of metal precursor
and the ligand were dissolved directly in a mixture of 25%/
75% vol/vol of CH2Cl2 with the appropriate ionic liquid.22 The
organic solvent was removed under high vacuum before starting
the hydrogenation reaction. Almost identical results are observed
with either the direct preformation or the in situ protocol in
[PMIM][PF6] as compared to the isolated complex 3 in the same
IL (Table 1, entries 6, 8, and 9).
The simple and straightforward in situ protocol opens a rapid
and effective way to test libraries of structurally diverse ligands
in the enantioselective hydrogenation of imines. This was
verified for the series of ligands 5a-e23 as shown in Scheme
3. All catalytic solutions were prepared from 4/5a-e following
the in situ protocol in a 12 mL stainless steel reactor. After
charging the catalytic solution with substrate, H2 (30 bar), and
CO2 (8-9 g), the reaction was allowed to proceed at 40 °C.
All catalysts 4/5a-e showed high activity, and >90% conver-
sion was obtained within a standard reaction time of 22 h. The
enantioselectivities varied between 44% and 68%, and the (R)-
enantiomer was obtained as the major isomer. Interestingly, the
relative selectivities in the [PF6]--containing ionic liquid (5d
> 5a > 5b > 5c . 5e) are not identical to those observed with
the isolated [BARF]- complexes in conventional solvents (5a
= 5c > 5d > 5e . 5b). In the case of 5b, the in situ protocol
leads even to slightly higher ee even though the opposite trend
would be expected from the usual behavior of the [PF6]-- and
[BARF]--based catalysts.
Taking this one step further, the anion exchange with the IL
allows one to even use a simple catalyst formed directly from
the air-stable precursors [Ir(COD)(Cl)]2 (4) and the free ligand
5a in the appropriate molar ration (5a:Ir ) 1:1). It is important
to note that mixtures of the two components 4/5a do not lead
to active catalysts in conventional solvents as the coordination
strength of chloride is too high, thus poisoning the active site.20,21
Probably the most general interest in the system IL/CO2
relates to the recycling and immobilization of the organometallic
catalyst. Indeed, this new biphasic system provides by far the
most effective methodology to immobilize the cationic iridium
(18) Drago, D.; Pregosin, P. S.; Pfaltz, A. Chem. Commun. 2002, 286.
(19) (a) Chauvin, Y.; Oliver-Bourbigou, H. CHEMTECH 1995, 25, 26. (b)
Muldoon, M. J.; Gordon, C. M.; Dunkin, I. R. J. Chem. Soc., Perkin Trans.
2 2001, 4, 433.
(20) Porte, A. M.; Reibenspies, J.; Burgess, K. J. Am. Chem. Soc. 1998, 120,
9180.
(21) Chloride anions also poison the catalyst in IL, but only at much higher
levels than those introduced in the in situ protocol. It proved necessary,
however, to check the ILs for their chloride content and to remove any
chloride impurities as described in the Experimental Section.
(22) The low solubility of the iridium precursor 4 in pure ionic liquids prevented
the complete elimination of organic solvents during the preformation step.
More benign alternatives to CH2Cl2 such as methanol or acetone may also
be used.
(23) (a) Cozzi, P. G.; Zimmermann, N.; Hilgraf, R.; Schaffner, S.; Pfaltz, A.
AdV. Synth. Catal. 2001, 343, 450. (b) Cozzi, P. G.; Menges, F.; Kaiser, S.
Synlett 2003, 6, 833.
9
J. AM. CHEM. SOC. VOL. 126, NO. 49, 2004 16145

A R T I C L E S
Solinas et al.
deactivation occurred after the fourth cycle.11 This deactivation
was attributed to the sensitivity of the active intermediates to
adventitious oxygen during the batch-wise recycling. Indeed, it
was found that IL solutions of catalyst 3 are significantly less
air sensitive than solutions of 3 in conventional organic
solvents.24 A closer inspection of the data summarized in Figure
3 shows that only 81% conversion was achieved in run #6. This
was due to a leak in the high-pressure setup resulting in loss of
the pressurized gases including the reactant hydrogen. Conse-
quently, the catalyst solution was exposed to air for a short
period of time. Despite this, the catalyst solution showed the
same performance as in the previous runs again in run #7.
Following this serendipitous observation, catalyst solutions
[EMIM][BTA]/3 were found to be remarkably stable even upon
prolonged exposure to air (75% conversion and 56% ee after
20 h in air under standard reaction conditions), indicating that
the ionic liquid environment helps to stabilize the catalytically
active intermediates.
Figure 3. Batch-wise catalyst recycling for hydrogenation of imine 1 using
catalyst 3 in the system [BMIM][PF6]/scCO2.
In addition to the catalyst recycling, CO2 extraction provides
also a very convenient method for product isolation. Upon
depressurization, compound 2 was isolated directly from the
CO2 stream in a solvent-free and analytically pure form as a
colorless crystalline material. In the experiments summarized
in Figure 3, 98% of the employed material was recovered with
the total number of turnovers exceeding TON ) 3000. No
signals associated with the IL were detected during NMR and
HPLC analyses of the material. Samples from each run were
analyzed by ICP measurement, demonstrating that the iridium
content was below the detection limit of the instrument (<1
ppm) in all cases and that no significant leaching of metal
compounds into the product had occurred. Although the relative
amount of CO2 used for the isolation of a certain amount of 2
is still quite high in the batch procedure, this could be greatly
reduced under continuous-flow operation.5a,6c,f The engineering
aspects of this approach are the subject of ongoing studies and
will be reported separately.
Figure 4. Biphasic system IL/scCO2 during the extraction step of the batch-
wise recycling process. CO2 is bubbled through the intensively red
IL/catalyst solution via the capillary, extracting the product 2 into a colorless
solution leaving the reactor at the top for downstream processing.
Experimental Section
catalysts of type 3 known to date. For the representative set of
experiments shown in Figure 3, a solution of [BMIM][PF6]
containing the catalyst 3 was used at T ) 40 °C and p(H2) )
40 bar in the presence of scCO2 (d ) 0.68 g mL-1). The reaction
time was reduced to only 3 h from the screening procedures,
still leading to complete conversion in the first cycle. This cor-
Safety Warning. The handling of pressurized gases and supercritical
fluids requires the use of suitable high-pressure equipment and must
be carried out under rigorous safety conditions only.
General Procedures and Materials. All manipulations of air- and
moisture-sensitive material were carried out under a positive pressure
of argon. 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. The syntheses of N-(1-
Phenylethylidene)aniline 19,25 and the phosphanodihydrooxazole-iridium
catalyst 39 were carried out according to literature procedures. Ionic
liquids were synthesized according to known procedures26 or obtained
from Solvent Innovation.27 In addition to standard purification, all ILs
were tested for chloride impurities which were removed by washing
CH2Cl2 solutions of the IL with water until the washing tested negative
for Cl- with AgNO3. 1H NMR, 13C NMR, and 31P NMR spectra were
recorded in deuterated solvents on a Bruker DPX 300 spectrometer.
GC-MS analyses were performed on a HP-MASS 5973 chromatograph.
HPLC analyses were performed with a JASCO PU-2080 chromatograph
responds to an average turnover frequency of TOF > 150 h-1
.
Supercritical CO2 was shown to be effective at moderate
temperatures and pressures for the quantitative extraction of
product 2 after the hydrogenation reaction from the catalyst
solution. Figure 4 shows a view of the biphasic mixture in the
high pressure cell during the extraction process with the
intensively red IL-catalyst phase on the bottom and the colorless
CO2-product phase on top. The capillary used for introduction
of CO2 and the bubbles purging through the IL phase are also
visible. Following the extraction, the reactor was charged again
with another batch of imine 1 for the successive hydrogenation
reaction. The catalyst solution showed stable activity and enan-
tioselectivity under these conditions for at least seven cycles.
With one exception, conversion remained above 90% in all runs
and the enantiomeric excess stayed constant at 62% ( 1%.
The long-term stability of the catalytic solution is in positive
contrast to the previous CO2-based immobilization method for
catalysts of type 3 using the CESS procedure, where a significant
(24) A similar observation was made for a Rh-MeDuPHOS-based catalyst:
Guernik, S.; Wolfson, A.; Herskowitz, M.; Greenspoon, N.; Geresh, S.
Chem. Commun. 2001, 2314.
(25) (a) Taguchi, K.; Westheimer, F. H. J. Org. Chem. 1971, 36, 1570. (b)
Samec, J. S. M.; Ba¨ckvall, J. E. Chem.-Eur. J. 2002, 8, 2955.
(26) Bonhoˆte, P.; Dias, A. P.; Papageorgiou, N.; Kalyanasundaram, K.; Gra¨tzel,
M. Inorg. Chem. 1996, 35, 1168.
(27) http://www.solvent-innovation.com.
9
16146 J. AM. CHEM. SOC. VOL. 126, NO. 49, 2004

Enantioselective Hydrogenation of Imines
A R T I C L E S
equipped with a MD-2010 detector using a CHIRACEL OD-H column.
Values for conversion and ee’s from representative single experiments
are quoted in the tables. Reproducibility of the catalytic runs was
checked for selected examples and was found to be well within 5%
relative experimental error.
Summary and Conclusion
In the present work, the iridium-catalyzed hydrogenation of
imines was used as a chemical probe to evaluate systematically
the potential of ionic liquid/supercritical CO2 reaction media
for multiphase catalysis. It is demonstrated that the application
of the biphasic system leads to activation, tuning, and im-
mobilization of the catalyst system that would be impossible in
classical organic solvent systems or in either of the two
unconventional media separately.
Hydrogenation of N-(1-Phenylethylidene)aniline 1 in IL/CO2
Using Catalyst 3. In a typical experiment, a 12 mL window equipped
stainless steel reactor was charged under argon atmosphere with the
appropriate IL (2 mL), substrate 1 (293 mg, 1.5 mmol), and catalyst 3
(2.5 mg, 3.0 × 10-3 mmol). The reactor was pressurized with 30 bar
of hydrogen followed by addition of 8.5 g of CO2 using a compressor
(corresponding to a total pressure of 150 bar at room temperature).
The reactor was warmed to 40 °C, and the reaction was allowed to
proceed for 22 h. After cooling to room temperature, all volatiles were
vented through a cold trap (-70 °C). The IL was recovered from the
autoclave and extracted several times with n-hexane until complete
recovery of the reaction product. The contents of the cold trap and the
extracted material were combined and analyzed by GC-MS and HPLC
to determine conversion and enantiomeric excess, respectively.
Hydrogenation of N-(1-Phenylethylidene)aniline 1 in IL/CO2
Using the in Situ Catalysts 4/5a-e. In a typical experiment a 12 mL
window equipped stainless steel reactor was charged under argon
atmosphere with [Ir(COD)Cl]2 (4, 7.8 mg, 12 × 10-3 mmol) and the
appropriate ligand 5a-e (24.2 × 10-3 mmol, 5a-e/Ir ) 1:1). A mixture
of the appropriate IL (2 mL) and CH2Cl2 (0.7 mL) was added, and the
resulting solution was stirred for 1-3 h at 50 °C. The organic solvent
was removed by evacuating the reactor for 2 h under stirring. Next,
substrate 1 (1.17 g, 6.0 mmol) was introduced under argon, and the
reactor was pressurized with 30 bar of hydrogen followed by addition
of ca. 8.5 g of CO2 using a compressor. The reactor was allowed to
proceed at 40 °C for 22 h followed by workup and analysis as described
above.
Recycling Experiments. A 24 mL window equipped stainless steel
autoclave was charged, under argon atmosphere, with [BMIM][PF6]
(3.2 mL), substrate 1 (0.87 g, 4.46 mmol), and catalyst 3 (8.2 mg, 0.01
mmol). The reactor was subsequently pressurized with hydrogen (40
bar) and filled with 13.7 g of CO2 using a compressor (total pressure
) 145 bar at 22 °C). The reaction was allowed to proceed at 40 °C for
3 h. The product 2 of the reaction was extracted by flushing CO2 through
the reactor at 160 bar and 45 °C for 3 h with an exit flow set to
approximately 700-800 mL min-1. The product was collected in two
sequential cold traps kept at 0 and -60 °C, leading to greater than
95% recovery in all cases. In the cases of quantitative conversion, the
product was recovered as a colorless microcrystalline material. Samples
were analyzed for conversion and ee as described above with additional
ICP analysis to assess the iridium content. After the extraction, the
reactor was charged again with the same amount of substrate, hydrogen,
and CO2 for the subsequent catalytic cycle.
Using imine hydrogenation as a benchmark transformation,
it was possible to rationalize the observed effects on the basis
of the physicochemical and molecular characteristics of the
reaction system. Whereas the unique properties of the super-
critical phase are particularly important for aspects such as mass
transfer, gas availability, and product isolation, the ionic
environment of the catalyst phase provides beneficial molecular
interactions with the cationic organometallic intermediates. Most
intriguingly, certain disadvantages of one medium can be
ameliorated through the complementary characteristics of the
other one. Among the most striking examples for this paradigm
is the significant increase of hydrogen solubility in the IL in
the presence of compressed CO2 that was quantified in this work
by high-pressure NMR. Another example is the feeble solvent
power of scCO2 for cationic catalysts that is turned into an
advantage as the catalyst is immobilized through a virtually
quantitative partitioning into the IL phase.
In summary, the results presented in this study emphasize
the remarkable potential of the IL/CO2 biphasic system for the
future development of multiphase catalysis. They provide also
guidelines for how to exploit the characteristics of this system
by focusing on certain benefits offered by the unique set of
complementary properties. This type of information is required
to assess the scope and limitations of individual techniques as
the number of novel approaches for biphasic catalysis based
on scCO2 as the mobile phase is increasing (“cartridge cataly-
sis”).28
Acknowledgment. This work was supported by the German
Ministery of Research and Education (bmbf) under the ConNe-
Cat program as part of the lighthouse project “Regulated
Systems for Multiphase Catalysis - smart solvents/smart
ligands”. Additional financial support by the Fonds der Che-
mischen Industrie is gratefully acknowledged. We thank Prof.
Dr. Peter Wasserscheid for fruitful discussion and Solvent
Innovation for the donation of ionic liquids. This paper is
dedicated to Prof. Dr. Wilhelm Keim on the occasion of his
70th birthday in recognition of his pioneering achievements in
multiphase catalysis.
High-Pressure NMR Investigation. A high-pressure sapphire NMR
tube (outer diameter ) 5 mm, V ) 0.9 mL) was charged, under argon
atmosphere, with [EMIM][BTA] (0.5 mL) and one drop of C6D6 for
the lock signal. The NMR tube was placed inside a cylindrical safety
shield made from polycarbonate, pressurized at room temperature with
30 bar of H2, and agitated by gentle shaking behind a second safety
shield. The tube was lowered into the magnet directly from the
JA046129G
1
polycarbonate cylinder using a home-built device. After the H NMR
spectrum was recorded at ambient temperature (297 ( 2 K), the tube
was removed from the spectrometer into the safety cylinder, connected
to a CO2 supply, and pressurized stepwise up to a total pressure of 130
bar with CO2 at room temperature.
(28) (a) Cole-Hamilton, D. J. Science 2003, 299, 1702. (b) Leitner, W.; Scurto,
A. M. In Aqueous-Phase Organometallic Catalysis, 2nd ed.; Cornils, B.,
Herrmann, W. A., Eds.; Wiley-VCH: Germany, 2004; Chapter 7.4. (c)
Leitner, W. Pure Appl. Chem. 2004, 76, 635.
9
J. AM. CHEM. SOC. VOL. 126, NO. 49, 2004 16147