Iridium-catalyzed enantioselective hydrogenation of imines in supercritical carbon dioxide

Article abstract of DOI:10.1021/ja984309i

Supercritical carbon dioxide (scCO₂) was shown to be a reaction medium with unique properties for highly efficient iridium-catalyzed enantioselective hydrogenation of prochiral imines. Cationic iridium(I) complexes with chiral phosphinodihydrooxazoles, modified with perfluoroalkyl groups in the ligand or in the anion, were synthesized and tested in the hydrogenation of N-(1-phenylethylidene)aniline. Both the side chains and the lipophilic anions increased the solubility, but the choice of the anion also had a dramatic effect on the enantioselectivity with tetrakis-3,5- bis(trifluoromethyl)phenylborate (BARF) leading to the highest asymmetric induction. (R)-N-phenyl-1-phenylethylamine was formed quantitativley within 1 h in scCO₂ [d(CO₂) = 0.75 g mL^-1] at 40 °C and a H₂ pressure of 30 bar with enantiomeric excesses of up to 81% using 0.078 mol % catalyst. The use of scCO₂ instead of conventional solvents such as CH₂Cl₂ allowed the catalyst loading to be lowered significantly owing to a change in the rate profile of the reaction. The homogeneous nature of the catalytically active species under the reaction conditions was demonstrated and was found to depend strongly on the composition of the reaction mixture and especially on the presence of the substrate. Utilizing the selective extractive properties of scCO₂, the product could be readily separated from the catalyst, which could be recycled several times without significant loss of activity and enantioselectivity. High-pressure FT-IR and NMR investigations revealed that the reactivity of the products to form the corresponding carbamic acids plays an important role for the application of this new methodology.

Full text of DOI:10.1021/ja984309i

J. Am. Chem. Soc. 1999, 121, 6421-6429
6421
Iridium-Catalyzed Enantioselective Hydrogenation of Imines in
Supercritical Carbon Dioxide
Sabine Kainz, Axel Brinkmann, Walter Leitner,*,† and Andreas Pfaltz*
Contribution from the Max-Planck-Institut fu¨r Kohlenforschung, Kaiser-Wilhelm-Platz 1,
45470 Mu¨lheim/Ruhr, Germany
ReceiVed December 14, 1998
Abstract: Supercritical carbon dioxide (scCO2) was shown to be a reaction medium with unique properties
for highly efficient iridium-catalyzed enantioselective hydrogenation of prochiral imines. Cationic iridium(I)
complexes with chiral phosphinodihydrooxazoles, modified with perfluoroalkyl groups in the ligand or in the
anion, were synthesized and tested in the hydrogenation of N-(1-phenylethylidene)aniline. Both the side chains
and the lipophilic anions increased the solubility, but the choice of the anion also had a dramatic effect on the
enantioselectivity with tetrakis-3,5-bis(trifluoromethyl)phenylborate (BARF) leading to the highest asymmetric
induction. (R)-N-phenyl-1-phenylethylamine was formed quantitativley within 1 h in scCO2 [d(CO2) ) 0.75
g mL-1] at 40 °C and a H2 pressure of 30 bar with enantiomeric excesses of up to 81% using 0.078 mol %
catalyst. The use of scCO2 instead of conventional solvents such as CH2Cl2 allowed the catalyst loading to be
lowered significantly owing to a change in the rate profile of the reaction. The homogeneous nature of the
catalytically active species under the reaction conditions was demonstrated and was found to depend strongly
on the composition of the reaction mixture and especially on the presence of the substrate. Utilizing the selective
extractive properties of scCO2, the product could be readily separated from the catalyst, which could be recycled
several times without significant loss of activity and enantioselectivity. High-pressure FT-IR and NMR
investigations revealed that the reactivity of the products to form the corresponding carbamic acids plays an
important role for the application of this new methodology.
Introduction
better compatibility with the employed catalysts (temporary
protecting group).4 In favorable cases, the extractive properties
of scCO25 may allow remarkable efficient and simple separation
of catalysts and products.4,6
Supercritical carbon dioxide (scCO2; Tc ) 31 °C, pc ) 73.75
bar, dc ) 0.468 g mL-1) is receiving considerable and ever
increasing interest as an environmentally benign reaction
medium with unique properties for chemical synthesis1 and
especially for homogeneously catalyzed reactions.2 The misci-
bility of scCO2 with many gases and the absence of a liquid/
gas-phase boundary in the supercritical state results in the
maximum availability of gaseous reactants, avoiding potential
problems of mass-transfer limitations.3 Beneficial effects can
also arise from the high compressibility of scCO2, allowing for
selectivity changes by variation of the density with comparably
small changes in the reaction conditions.4 The chemical interac-
tion of CO2 with functional groups of the substrate can lead to
The use of scCO2 as a solvent for reactions involving
hydrogen as one of the substrates is particularly attractive.
Recent research has demonstrated enhanced reaction rates in
the ruthenium-catalyzed hydrogenation of CO2 to formic acid
and its derivatives3,7 or the rhodium-catalyzed hydroformylation
of olefins.6 Higher regioselectivity at comparable rates has also
been observed with certain catalytic systems for hydroformyl-
ation,8 and preliminary results on asymmetric hydroformylation
have been reported.9 The highly enantioselective hydrogenation
of CdC double bonds has been demonstrated with prochiral
R,â-unsaturated carboxylic acids as substrates.10,11 The phase
behavior of the reaction mixture and the solubility of the catalyst
were found to be of paramount importance for the successful
† Fax: +49-208-306-2993. E-mail: leitner@mpi-muelheim.mpg.de.
(1) Chemical Synthesis in Supercritical Fluids; Jessop, P. G., Leitner,
W., Eds.; Wiley-VCH: New York, 1999.
(2) Selected recent reviews: (a) Jessop, P. G.; Ikariya, T.; Noyori, R.
Science 1995, 269, 1065. (b) Morgenstern, D. A.; LeLacheur, R. M.; Morita,
D. K.; Borkowsky, S. L.; Feng, S.; Brown, G. H.; Luan, L.; Gross, M. F.;
Burk, M. J.; Tumas, W. In Green Chemistry; Anastas, P. T., Williamson,
T. C., Eds.; ACS Symposium Series 626; American Chemical Society:
Washington, DC, 1996; pp 132. (c) Dinjus, E.; Fornika, R.; Scholz, M. In
Chemistry under Extreme or Non-Classical Conditions; van Eldik, R.,
Hubbard, C. D., Eds.; Wiley: New York, 1996; pp 219. (d) Jessop, P. G.
In Fine Chemicals Catalysis II; Blackmond, D., Leitner, W., Eds.; Top.
Catal. 1998, 5, 95.
(5) (a) Zosel, K.; Angew. Chem. 1978, 90, 748; Angew. Chem., Int. Ed.
Engl. 1978, 17, 702. (b) McHugh, M. A.; Krukonis, V. J. Supercritical
Fluid Extraction, 2nd ed.; Butterworth: Boston, 1994.
(6) Koch, D.; Leitner, W. J. Am. Chem. Soc. 1998, 120, 13398.
(7) For general reviews on this area see: (a) Jessop, P. G.; Ikariya, T.;
Noyori, R. Chem. ReV. 1995, 95, 259. (b) Leitner, W. Angew. Chem. 1995,
107, 2391; Angew. Chem., Int. Ed. Engl. 1995, 34, 2207.
(8) (a) Rathke, J. W.; Klingler, R. J.; Krause, T. R. Organometallics
1991, 10, 1350. (b) Guo, Y.; Akgerman, A. Ind. Eng. Chem. Res. 1997,
36, 4581. (c) Bach, I.; Cole-Hamilton, D. J. Chem. Commun. 1998, 1463.
(9) Kainz, S.; Leitner, W. Catal. Lett. 1998, 55, 223.
(3) (a) Jessop, P. G.; Ikariya, T.; Noyori, R. Nature 1994, 368, 231. (b)
Jessop, P. G.; Hsiao, Y.; Ikariya, T.; Noyori, R. J. Am. Chem. Soc. 1996,
118, 344.
(10) Xiao, J.; Nefkens, S. C. A.; Jessop, P. G.; Ikariya, T.; Noyori, R.
Tetrahedron Lett. 1996, 37, 2813.
(4) Fu¨rstner, A.; Koch, D.; Langemann, K.; Leitner, W.; Six, C. Angew.
Chem. 1997, 109, 2562; Angew. Chem., Int. Ed. Engl. 1997, 36, 2466.
(11) Burk, J.; Feng, S.; Gross, M. F.; Tumas, W. J. Am. Chem. Soc.
1995, 117, 8277.
10.1021/ja984309i CCC: $18.00 © 1999 American Chemical Society
Published on Web 06/26/1999

6422 J. Am. Chem. Soc., Vol. 121, No. 27, 1999
Kainz et al.
Scheme 1
Scheme 2
application of scCO2 to asymmetric catalysis in these studies.
To increase the solubility of transition metal catalysts in scCO2,
it has been suggested to incorporate perfluorinated side chains
either directly in the metal-attached phosphine ligands12,13 or
in the anions in the case of cationic complexes.11
Recently, enantiopure phosphinodihydrooxazoles21-23 were
successfully used in the Ir-catalyzed enantioselective hydrogena-
tion of imines in CH2Cl2.24 We now report that it is possible to
replace the organic solvent with toxicologically and environ-
mentally benign scCO2 without loss of enantioselectivity, if the
phosphinodihydrooxazole-bearing Ir catalysts are suitably ad-
justed to the specific properties of the reaction medium.
Moreover, we provide the first evidence that the use of scCO2
can result in considerably higher catalyst efficiency owing to a
different rate profile rather than by simply increasing the overall
rate. Furthermore, scCO2 was utilized in an integrated process
as the medium for reaction and separation, which allowed easy
isolation of the pure product and efficient recovery of the
catalyst. Detailed investigations including high-pressure NMR
and in situ reflectance FT-IR spectroscopy provided insight into
the phase behavior, the solubility, and possible interactions of
the various reaction components with compressed CO2.
The enantioselective reduction of the CdN double bond is
an important synthetic strategy for the preparation of optical
active amines (Scheme 1) and has received much attention over
the past few years, in both academic and industrial research.14
A variety of chiral Rh,15 Ir,16 Ru,17 and Ti18 complexes have
been studied as catalysts for the hydrogenation of imines,
whereby the late transition metal systems contained mainly
chiral diphosphine ligands.19 The first industrial application of
enantioselective imine hydrogenation was developed by a
research group at Ciba-Geigy.20 The catalyst, a chiral ferroce-
nyldiphosphine-iridium complex, shows extremely high activity
and unprecedented productivity (up to 106 turnovers) with
certain arylimines. The hydrogenation process is used to
synthesize a precursor of (S)-metolachlor, an important herbi-
cide, with 80% ee on a technical scale.
(12) Kainz, S.; Koch, D.; Baumann, W.; Leitner, W. Angew. Chem. 1997,
109, 1699; Angew. Chem., Int. Ed. Engl. 1997, 36, 1628.
(13) For recent applications of this concept in catalysis see: (a) Garrol,
M. A.; Holmes, A. B. Chem. Commun. 1998, 1395. (b) Morita, D. K.; Pesiri,
D. R.; Scott, A. D.; Glaze, W. H.; Tumas, W. Chem. Commun. 1998, 1397.
(c) Palo, D. R.; Erkey, C. Ind. Eng. Chem. Res. 1998, 37, 4203. (d)
Reference 6.
(14) (a) Noyori, R. Asymmetric Catalysis in Organic Synthesis; Wiley:
New York, 1994; pp 82. (b) Johansson, A. Contemp. Org. Synth. 1995, 2,
393. (c) James, B. R. In Catalysis of Organic Reactions, Chemical
Industries; Scaros, M. G., Prunier, M. L., Eds.; Dekker: New York, 1995;
Vol. 62, p 167.
Results and Discussion
Preparation of Ir Catalysts for Use in scCO2. Recent
research on catalysis in scCO2 and the known solubility data
25
for metal complexes in pure scCO2 suggested a very low
solubility in this medium for the cationic iridium phosphinodi-
hydrooxazole complexes used as catalysts for imine hydrogena-
tion in conventional organic solvents. Therefore, we synthesized
a series of complexes 5-7 with perfluorinated groups in the
ligand and/or the anion. In the ligands, perfluoroalkyl groups
were attached in the periphery of the chiral skeleton to minimize
interference with the chirality transfer at the active center. As
a “CO2-philic” counterion, the tetrakis-3,5-bis(trifluoromethyl)-
phenylborate anion (BARF) was chosen.
The unsubstituted 2-(phosphinoaryl)oxazoline ligand 3a was
obtained via orthometalation of the 2-aryloxazoline 1 and
subsequent treatment with chlorodiphenylphosphine 2a as
described earlier.26 Use of the perfluoroalkyl-substituted chlo-
rodiphenylphosphine 2b gave the new ligand 3b (Scheme 2).
The intermediate 2b was prepared in 67% yield as a colorless
solid by lithiation of 1-bromo-4-(3,3,4,4,5,5,6,6,7,7,8,8,8-tride-
cafluorooctyl)benzene27 with n-BuLi in diethyl ether, followed
by addition of the resulting solution to Cl2PNEt2 in THF, and
subsequent treatment with dry gaseous HCl.28,29
(15) (a) Vastag, S.; Bakos, J.; To¨ro¨s, S.; Takach, N. E.; King, R. B.;
Heil, B.; Marko´, L. J. Mol. Catal. 1984, 22, 283. (b) Bakos, J.; To´th, I.;
Heil, B.; Marko´, L. J. Organomet. Chem. 1985, 279, 23. (c) Kang, G.-J.;
Cullen, W. R.; Fryzuk, M. D.; James, B. R.; Kutney, J. P. J. Chem. Soc.,
Chem. Commun. 1988, 1466. (d) Becalski, A. B.; Cullen, W. R.; Fryzuk,
M. D.; James, B. R.; Kang, G.-J.; Rettig, S. J. Inorg. Chem. 1991, 30, 5002.
(e) Burk, M. J.; Martinez, J. P.; Feaster, J. E.; Cosford, N. Tetrahedron
1994, 50, 4399. For hydrogenation of benzylimines of aryl methyl ketones
in a two-phase system, see: (f) Bakos, J.; Orosz, A.; Heil, B.; Laghmari,
M.; Lhoste, P.; Sinou, D. J. Chem. Soc., Chem. Commun. 1991, 1684. (g)
Lensink, C.; deVries, J. G. Tetrahedron: Asymmetry 1992, 3, 235.
(16) (a) Spindler, F.; Pugin, B.; Blaser, H.-U. Angew. Chem., Int. Ed.
Engl. 1990, 29, 558. (b) Spindler, F.; Pugin, B. (Ciba-Geigy AG). EP Patent
0256982, 1988. (c) Ng Cheon Chan, Y. P.; Osborn, J. A. J. Am. Chem.
Soc. 1990, 112, 9400. (d) Ng Cheon Chan, Y. P.; Meyer, D.; Osborn, J. A.
J. Chem. Soc., Chem. Commun. 1990, 869. (e) Sablong, R.; Osborn, J. A.
Tetrahedron: Asymmetry 1996, 7, 3059. (f) Morimoto, T.; Achiwa, K.
Tetrahedron: Asymmetry 1995, 6, 2661. (g) Tani, K.; Onouchi, J.;
Yamagata, T.; Kataoka, Y. Chem. Lett. 1995. 955.
(17) (a) Oppolzer, W.; Wills, M.; Starkemann, C.; Bernardinelli, G.
Tetrahedron Lett. 1990, 31, 4117. (b) Fogg, D. E.; James, B. R. In Catalysis
of Organic Reactions, Chemical Industries; Scaros, M. G., Prunier, M. L.,
Eds.; Dekker: New York, 1995; Vol. 62, p 435. For the use of Ru-diamine
complexes in the enantioselective transfer-hydrogenation of imines, see:
(c) Uematsu, N.; Fujii, A.; Hashiguchi, S.; Ikariya, T.; Noyori, R. J. Am.
Chem. Soc. 1996, 118, 4916.
(18) (a) Willoughby, C. A.; Buchwald, S. L. J. Am. Chem. Soc. 1992,
114, 7562. Willoughby, C. A.; Buchwald, S. L. J. Org. Chem. 1993, 58,
7629. (b) Willoughby, C. A.; Buchwald, S. L. J. Am. Chem. Soc. 1994,
116, 8952. (c) Willoughby, C. A.; Buchwald, S. L. J. Am. Chem. Soc. 1994,
116, 11703.
(21) Sprinz, J.; Helmchen, G. Tetrahedron Lett. 1993, 34, 1769.
(22) Dawson, G. J.; Frost, C. G.; Williams, J. M. J., Coote, S. J.
Tetrahedron Lett. 1993, 34, 3149.
(23) (a) von Matt, P.; Pfaltz, A. Angew. Chem. 1993, 105, 614; Angew.
Chem., Int. Ed. Engl. 1993, 32, 566. (b) Review: Pfaltz, A. Acta Chem.
Scand. 1996, 50, 189.
(24) (a) Schnider, P.; Koch, G.; Pre´toˆt, R.; Wang, G.; Bohnen, F. M.;
Kru¨ger, C.; Pfaltz, A. Chem. Eur. J. 1997, 3, 887. (b) Schnider, P. Ph.D.
Thesis, University of Basel, 1996.
(25) Smart, N. G.; Carleson, T.; Kast, T.; Clifford, A. A.; Burford, M.
D.; Wai, C. M. Talanta 1997, 44, 137.
(26) Koch, G.; Lloyd-Jones, G. C.; Loiseleur, O.; Pfaltz, A.; Pre´toˆt, R.;
Schaffner, S.; Schnider, P.; von Matt, R. Recl. TraV. Chim. Pay-Bas 1995,
114, 206.
(27) Kainz, S.; Luo, Z.; Curran, D. P.; Leitner, W. Synthesis 1998, 1425.
(28) Casalnuovo, A. L.; RajanBabu, T. V.; Ayers, T. A.; Warren, T. H.
J. Am. Chem. Soc. 1994, 116, 9869.
(19) (a) Miao, C. K.; Sorcek, R.; Jones P.-J. Tetrahedron Lett. 1993,
34, 2259. (b) Sreckumar, R.; Pillai, C. N. Tetrahedron: Asymmetry 1993,
4, 2095.
(20) (a) Blaser, H.-U.; Spindler, F. In Fine Chemicals Catalysis I;
Blackmond, D., Leitner, W., Eds.; Top. Catal. 1997, 4, 275. (b) Togni, A.
Angew. Chem. 1996, 108, 1581; Angew. Chem., Int. Ed. Engl. 1996, 35,
1475.

Hydrogenation of Imines in scCO2
J. Am. Chem. Soc., Vol. 121, No. 27, 1999 6423
Scheme 3
Scheme 4
a
Table 1. Asymmetric Hydrogenation of Imines 8 in CH2Cl2
entry substrate (mmol) catalyst (mol %) conversion (%) ee (R) (%)
1
8a (4.15)
8a (5.05)
8a (5.00)
8a (4.17)
8a (4.18)
8b (4.00)
8b (3.83)
5a (0.15)
5b (0.11)
5b (0.06)
6b (0.13)
7a (0.15)
5a (0.3)
99.7
100.0
10.0
99.9
100.0
99.3
83
80
ndc
86
87
72
71
Treatment of the dimeric iridium complex [{Ir(cod)(µ-Cl)}2]
(4) (cod ) 1,5-cyclooctadiene) with ligand 3a or 3b in CH2Cl2
at 50 °C for several hours gave the complexes [Ir(cod)(3)]Cl.
The chloro complexes can be isolated, but in general, the
solutions were submitted directly to anion exchange by treatment
with aqueous solutions of [NH4][PF6], Na[BPh4], or NaBARF
at room temperature. Subsequent purification by flash chroma-
tography afforded the iridium complexes 5-7 in ca. 90% yield
(Scheme 3). Purification was also possible by crystallization,
albeit in substantially lower yield than by chromatography (see
the experimental procedure for 5b).
2
3b
4
5d
6
7
7a (0.15)
80.3
a p(H2) ) 100 bar, T ) 40 °C, reaction time 20 h. b In hexane. c Not
determined. d Reaction time 5 h.
a
Table 2. Asymmetric Hydrogenation of Imines 8 in scCO2
entry substrate (mmol) catalyst (mol %) conversion (%) ee (R) (%)
Complexes 5-7 were obtained as bright red solids containing
various amounts of CH2Cl2 (0.1-0.7 equiv) even after prolonged
drying under vacuum according to elemental and NMR analysis.
For simplicity, all concentration values in this paper are based
on solvent-free complexes, because of the large molecular
weight of the compounds relative to CH2Cl2. The spectroscopic
data for the pairs of compounds a/b are very similar. The
introduction of the perfluoroalkyl chain leads to a slight decrease
of the 31P resonance from approximately 16.5 to 15.5 ppm for
the two sets of borate complexes 6a/b and 7a/b. For comparison,
the substitution of the [BAr4]- with the [PF6]- anion leads to a
variation of ∆δ ) 3 to lower field. The chemical shifts of the
olefinic proton and carbon nuclei of the cod ligand, which might
experience an electronic or steric change at the metal center
most directly, are identical within experimental error for the
whole series of complexes. This indirect evidence is fully in
line with our earlier 103Rh NMR-based finding that substitution
with ethylene-spaced perfluoroalkyl groups in the para position
of the PPh2 moiety keeps structural and electronic changes at
the metal center to a minimum.12
The lipophilicity of a compound is one of several factors
determining its solubility in scCO2, with higher lipophilicity
resulting generally in higher solubility. Although no quantitative
measurements have been performed, the Rf values obtained from
TLC using silica plates under identical conditions give some
indication of the lipophilicity of the cationic iridium complexes.
The Rf values increase in the order [Ir(cod)(3a)]Cl e [Ir(cod)-
(3b)]Cl , 6a < 6b , 7a < 7b, indicating that the anion has
a larger influence on the lipophilicity of these ionic complexes
than the ligand substitution pattern.
1
8a (0.15)
8a (4.15)
8a (4.13)
8a (4.17)
8a (4.16)
8a (4.22)
8a (4.19)
8a (4.53)
8b (3.85)
8b (3.87)
8b (2.4)
5a (0.15)
5b (0.14)
5b (0.13)
6a (0.22)
6b (0.14)
7a (0.14)
7a (0.15)
7b (0.09)
5a (0.3)
99.9
100
99.5
37
26
26
2
3b
4
11.0 (83.1d) ndc (60d)
5
99.8
100
100
100
68
81
79
80
ndc
6
7e
8
9
10
11
28.9
7a (0.15)
7a (0.25)
15.5 (28.4f) ndc
23
ndc
a d(CO2) ) 0.75 g mL-1, p(H2) ) 30 bar, T ) 40 °C, reaction time
20 h. b With addition of 0.5 mL of CH2Cl2. c Not determined. d Reaction
time 40 h. e p(H2) ) 10 bar. f Reaction time 70 h.
and with maximum stirring (magnetic stirring bar) for best
comparison. Reactions were conducted for a standard reaction
time of 20 h for screening purposes and monitored by GC for
quantitative comparison.
Complex 5a has been reported to catalyze the hydrogenation
of 8a to (R)-9a with ca. 84% ee in CH2Cl2 under 100 bar of H2
at room temperature using a catalyst loading of 0.1 mol %.24
We found that all new catalysts 5-7 lead to quantitative
hydrogenation of the imine 8a in the same solvent, giving rise
to formation of (R)-9a with 80-87% ee under similar conditions
(Table 1). As anticipated, the introduction of the perfluoralkyl
chain in the para position had little influence on the performance
of the catalyst, and almost identical enantioselectivities were
observed for complexes 5a and 5b. The anion had a marked
influence on the reaction rate (Table 3), but only minor effects
on the asymmetric induction in the organic medium, whereby
the borate complexes 6 and 7 gave marginally higher ee’s than
their [PF6]- congeners 5.30
The situation changed dramatically when the reaction was
carried out in scCO2 as a solvent (Table 2). Hydrogenation of
8a with complex 5a in scCO2 gave complete conversion to the
corresponding amine 9a after 20 h, but a disappointingly low
enantiomeric excess of 37% (R) was observed. Quantitative
conversion and similar poor enantioselectivity was obtained with
5b containing perfluoralkyl chains in the PPh2 group. The
unsubstituted borate complex 6a showed very low activity in
Iridium-Catalyzed Enantioselective Hydrogenation of
Imines in Supercritical CO2 and in CH2Cl2. The enantiose-
lective hydrogenation of N-(1-phenylethylidene)aniline (8a) to
give N-phenyl-1-phenylethylamine (9a) was used as a test
reaction (Scheme 4). Catalytic runs were performed in a
window-equipped stainless steel reactor (V ) 100 mL) fitted
with a dip tube (stainless steel capillary of 0.8 mm inner
diameter) and a HPLC valve to allow sampling for offline GC
analysis. Control experiments in CH2Cl2 were carried out with
identical amounts of substrate and catalysts per solvent volume
(30) The BARF anion was found to yield particularly active and selective
catalysts in the enantioselective hydrogenation of nonfunctionalized CdC
double bonds using the similar complexes in organic solvents: Lightfoot,
A.; Schnider, P.; Pfaltz, A. Angew. Chem. 1998, 110, 3047; Angew. Chem.,
Int. Ed. Engl. 1998, 37, 2897.
(29) A chiral C2-symmetric phosphinite with perfluoralkyl substituents
has been previously synthesized from 2b: Kainz, S.; Koch, D.; Leitner,
W. In SelectiVe Reactions of Metal-ActiVated Molecules; Werner, H.,
Schreier, W., Eds.; Vieweg: Wiesbaden, 1998; p 151.

6424 J. Am. Chem. Soc., Vol. 121, No. 27, 1999
Kainz et al.
scCO2 (11% conversion after 20 h), whereas 6b gave again
complete conversion after identical reaction time. Both com-
plexes 6a and 6b gave similar enantioselectivities between 60%
and 70% (R), i.e., considerably higher than the corresponding
complexes containing the [PF6]- counterion. Finally, the
complexes 7 containing the BARF anion led to quantitative
conversion within the standard reaction time, and (R)-9a was
now obtained with 81% and 80% ee using 7a (0.14 mol %)
and 7b (0.09 mol %), respectively. With 7a, the reaction
occurred still smoothly at a hydrogen pressure of only 10 bar,
giving 79% ee.
further addition of catalyst. Quantitative conversion to give (R)-
9a was again obtained with 78% ee, proving that the catalyst
had been dissolved from the compartment and homogeneously
dispersed throughout the reaction medium during the first run.
Similar experiments using glass inserts to separate the catalyst
from the other reaction components revealed that the solubility
of the metal species was strongly dependent on the composition
of the reaction mixture. Before introduction of H2, the bright
red complex 7a remained insoluble and unchanged in the
presence of imine 8a and CO2 even at temperatures and
pressures well above the point where complete dissolution of
the imine was observed. Introduction of H2 led to a rapid color
change from red to yellow and apparent complete dissolution
of the material, resulting in a pale yellow coloration of the
supercritical phase. After venting, 50-75% of the total amount
of iridium was found consistently in the reactor after removal
of the beaker. In the absence of substrate 8a, catalyst 7a also
turned yellow, but remained mostly insoluble in the compressed
CO2/H2 mixture (92% recovery of Ir in the beaker). Most
remarkably, the amine 9a also had no apparent effect on the
solubility of the metal species, and >90% iridium was recovered
in the glass beaker.31
Combination of Catalysis and Extraction (CESS Process)
and Catalyst Recycling. The solubility behavior of the catalyti-
cally active species suggested the possibility to separate 9a
selectively from the reaction mixture by supercritical fluid
extraction (SFE), leaving the catalyst in the reactor in active
form. Most recently, we have described the first examples4,6
for combinations of homogeneous catalysis and SFE using
scCO2, and we refer to such integrated processes as “catalysis
and extraction using supercritical solutions” (CESS). In the
present case, it was possible to extract the amine 9a selectively
without changing the conditions between the reaction and
extraction stages. Simply purging the reactor after complete
conversion with compressed CO2 at 40 °C and 110 bar allowed
9a to be collected in the form of colorless crystals in a cold
trap. Practically quantitative recovery of the amine was achieved
after 1 h of extraction using an amount of CO2 corresponding
to 100 L of gaseous CO2 at ambient conditions. The iridium
content of the product was determined by atomic absorption
spectroscopy (AAS) to be below 5 ppm.
Thus, it can be concluded that the enantioselective hydroge-
nation of N-(1-phenylethylidene)aniline (8a) can be carried out
in scCO2 with very high activity and with similar levels of
enantiocontrol as in CH2Cl2.24 However, scCO2 cannot be
regarded simply as a “nonpolar substitute solvent” for organic
liquids. For example, catalyst 5b proved insoluble and almost
completely inactive in hexane, a solvent with comparable density
and polarity as scCO2. Moreover, the substrate N-(1-phenyl-
ethylidene)benzylamine (8b), which is structurally closely
related to 8a, proved to be very difficult to hydrogenate in scCO2
(Table 2). Independently of the catalyst precursor, conversion
of 8b did not exceed 30% in scCO2 even at prolonged reaction
times or higher catalyst loadings. In contrast, the formation of
9b proceeded smoothly with 71-72% ee in CH2Cl2 (Table 1).
To gain more insight into the course of the reaction in scCO2
and to obtain a more detailed comparison to the conventional
solvent system, additional experiments were conducted using
catalyst 7a.
Phase Behavior and Solubility of the Metal Species. The
phase behavior of the reaction mixure and the nature of the
active species (homogeneous vs heterogeneous) are of major
concern for the design and the understanding of catalytic systems
operating in scCO2. In the present case, visual inspection showed
that up to 800 mg of the imine 8a were completely soluble in
100 mL of scCO2 of densities d(CO2) g 0.65 g mL-1 at or just
above the critical temperature of pure CO2. No separation of a
liquid phase was observed when 8a was converted quantitatively
to 9a during the whole course of reaction. Under the reaction
conditions, no undissolved solid could be observed, but very
small amounts would have been difficult to detect owing to the
reactor design. In fact, the difference in activity between 6a
and 6b may well be due to a solubility increase by introduction
of the perfluoroalkyl chains in the PPh2 group, and we cannot
fully exclude that parts of the initially charged precursor
complexes remain insoluble during catalytic runs also in other
cases. The homogeneous nature of the active intermediate
generated from catalyst precursor 7a, however, is indicated by
a pale yellow color of the reaction medium throughout the
reaction and could be verified in the following experiments.
Recharging the reactor with new substrate 8a without any
addition of catalyst or ligand led to quantitative hydrogenation
of 8a within the standard reaction time with almost identical
levels of enantioselectivity in four subsequent experiments
(Figure 1). Longer reaction times were required for quantitative
conversion in subsequent runs, but the ee of the product
remained above 70%. The overall yield from the repeated
experiments shown in Figure 1 corresponds to a total turnover
number (TON, moles of 9a per mole of 7a) of 10 000, and the
isolated (R)-9a had an average ee of 76%.
In the first run, catalyst 7a (0.08 mol %) was placed in the
reactor inside a small glass beaker containing a small stirring
bar. The opening on the top of the beaker reached just slightly
below the lid of the reactor. Substrate 8a was introduced as a
solid into the reactor space around this compartment. The desired
amount of CO2 was then introduced very carefully at T ) 34
°C through a valve at the bottom of the reactor to allow the
presumably supercritical mixture of 8a and CO2 to come into
contact with 7a. To start the reaction, H2 was introduced and
the autoclave was further heated to the reaction temperature of
40 °C. After 20 h, the reactor was cooled and the product (98%
(R)-9a, 81% ee) isolated by CO2 extraction (vide infra). The
beaker was removed, and the reactor was charged again with
8a, CO2, and H2. Then, the second run was performed without
Comparison of the Enantioselective Hydrogenation of 8a
with Catalyst 7a in scCO2 and in CH2Cl2. Figure 2 and Table
3 summarize the course of formation of 9a during hydrogenation
of 8a with different loadings of catalyst 7a in scCO2 and CH2Cl2
under otherwise identical conditions. Following the course of
reaction by sampling and offline GC analysis under standard
conditions in scCO2 revealed that the reaction was too fast to
be monitored with reasonable accuracy at catalyst loadings >0.1
mol % (Table 3). Stepwise reduction of the catalyst loading
showed that quantitative conversion was achieved within less
(31) The observations made by visual control and the pronounced
influence of the composition of the mixture make other possible explanations
for the activity in the second run such as accidental mechanical extrusion
extremely unlikely.

Hydrogenation of Imines in scCO2
J. Am. Chem. Soc., Vol. 121, No. 27, 1999 6425
Scheme 5
40 bar (four experiments, ee ) 79.3-81.7%; see Table 2, entries
6 and 7, for representative results). It is important to note that
the ee’s were also independent of H2 pressure in CH2Cl2 between
30 and 100 bar (compare entry 5 in Table 3 with entry 5 in
Table 1). Typically, the ee’s were higher by 5-8% in CH2Cl2
than in scCO2 under comparable conditions. The initial TOF
was, however, approximately 2 times larger in the high-density
supercritical reaction medium. Most intriguingly, the overall
shape of the conversion versus time profile was remarkably
different in both solvent systems. Two distinct rate regimes were
observed in the organic solvent at 0.071 mol % 7a (Figure 2C),
an initial phase with a high rate followed by a significant rate
decrease after 30-40% conversion. The initial rate in CH2Cl2
was found to be proportional to the amount of catalyst in the
investigated concentration range (constant TOF; see Table 3),
showing that this rate profile is not governed by mass-transfer
limitations under these conditions. In sharp contrast, no rate
decrease was observed at the same iridium concentration in
scCO2 up to >90% conversion (Figure 2A). In scCO2, the rate
remained stable at even lower concentrations, corresponding to
Ir loadings as low as 0.014 mol %. Thus, >90% conversion of
the imine corresponding to >6800 turnovers could be achieved
within less than 6 h in scCO2 (Figure 2B), whereas more than
22 h was required to achieve similar conversion in CH2Cl2 using
a 5 times larger amount of 7a (TON ) 1400).
Figure 1. Catalyst recycling in the enantioselective hydrogenation of
8a to give (R)-9a using 7a as catalyst and scCO2 as medium for reaction
and separation (CESS process). The dark gray bars indicate the
enantiomeric excess, and the light gray bars show the time required
for quantitative conversion.
The iridium-catalyzed enantioselective hydrogenation of
imines proceeds through a multistep catalytic cycle which is
not known in detail. A highly simplified sequence including
only the most basic steps is depicted in Scheme 5. In this
scheme, the yellow species formed from the reaction of 7a with
H2 corresponds to the catalytically active complex 10. This
species reacts with 8a in a reversible reaction to give the
intermediate 11, whose reaction with H2 liberates the amine 9a
and regenerates 10. The rate profile in scCO2 indicates that the
reaction is zero order (or at least close to) in substrate and also
largely independent of the reaction time (see also the recycling
experiments!). This kinetic behavior is consistent with the
sequence shown in Scheme 5 if the equilibrium between 10
and 8a lies far to the side of 11 and is established much faster
than the subsequent product formation. This would also explain
why the catalytically active species 11, which is the only metal
species that is reasonably soluble in scCO2, remains in solution
until the imine 8a is consumed completely. In principle, the
same scheme may apply for the reaction in CH2Cl2, and indeed
the same color changes are observed in this medium. However,
the reaction profile in CH2Cl2 is highly nonlinear, demonstrating
a strong dependence of the rate on the reaction time and/or the
concentration of substrate. This may reflect a rate law with a
broken order in substrate or some deactivation process of the
catalyst.32 A precise mechanistic explanation for the greatly
enhanced catalytic efficiency observed in scCO2 has to await
further studies including detailed kinetic investigations in scCO2
and comparison to the yet unknown rate law in CH2Cl2. It is,
however, already clear from the current data that the enhance-
ment of the catalyst efficiency cannot be reduced to a mere
enhancement of “H2 availability”.
Figure 2. Conversion/time profile for the enantioselective hydrogena-
tion of imine 8a using catalyst 7a in scCO2 and CH2Cl2, respectively.
Table 3. Influence of Reaction Conditions on the Asymmetric
a
Hydrogenation of Imine 8a in scCO2 and CH2Cl2
catalyst
density
t
TOFd ee (R)
entry (mol %) solvent (g mL-1
)
(h)b TONc (h-1
)
(%)
1
2
3
4
5
6
7
7a (0.014) scCO2
7a (0.078) scCO2
7a (0.078) scCO2
7a (0.071) CH2Cl2
7a (0.16) CH2Cl2
5a (0.079) CH2Cl2
5a (0.15) CH2Cl2
0.75
0.75
0.65
8.8 6830 2140
1.2 1280 2820
1.5 1275 1280
74
78
73
86
86
84
83
24
1.5
30
1400 1200
625 1400
1260
645
600
600
3
a p(H2) ) 30 bar, 4.2 mmol of 8a, T ) 40 °C. b Time required for
>98% conversion. c Turnover number ) moles of 9a per mole of Ir at
time t. d Graphically determined initial turnover frequency for the
formation of 9a, given as moles of 9a per mole of Ir per hour.
than 9 h at 0.014 mol % 7a. Thus, in a single run, a TON )
6830 could be achieved with an initial turnover frequency (TOF,
moles of 9a formed per mole of 7a and hour) of 2140 h-1. The
initial TOF was found to vary somewhat with catalyst loading,
and a TOF of 2820 h-1 was obtained at 0.078 mol %. The ee
decreased slightly with decreasing amounts of 7a from 81%
(0.14 mol %) to 78% (0.078 mol %) and finally 74% (0.014
mol %). The initial reaction rate decreased by a factor of 2.2
when the density of CO2 was lowered from 0.75 g mL-1 to
0.65 g mL-1. At the same time, the ee decreased slightly from
78% to 73%. The latter observations could result from reduced
solubility of the active species at lower density, but a direct
effect of the density on rate and selectivity cannot be ruled out.
In scCO2, no significant variation of the ee was observed
upon varying the partial pressures of hydrogen between 10 and
(32) These results are also consistent with findings of an ongoing detailed
investigation of the reaction kinetics using 5a in CH2Cl2. Blackmond, D.;
Pfaltz, A. Unpublished results.

6426 J. Am. Chem. Soc., Vol. 121, No. 27, 1999
Kainz et al.
Figure 3. In situ FT-IR monitoring of the hydrogenation of imine 8a to give the secondary amine 9a with catalyst 7a in scCO2 using a high-
pressure reflectance probe. The marked and underlined bands can be identified at identical positions in control spectra of toluene solutions of 9a
and 8a, respectively.
Investigations Using High-Pressure Spectroscopy. Due to
the rapidly growing number of homogeneously catalyzed
reactions which can be carried out efficiently in scCO2, there
is an increasing demand for in situ techniques to monitor the
course of chemical syntheses in this reaction medium. In the
present case, the hydrogenation of 8a could be monitored readily
by in situ FT-IR spectroscopy utilizing a high-pressure reflec-
tance probe which was mounted directly to a standard window-
equipped stainless steel reactor (200 mL). This technique avoids
some of the problems frequently encountered with conventional
high-pressure IR cells such as difficulties in pumping the
reaction medium and inefficient mass transport owing to the
small path lengths. The high-pressure reflectance IR probe is
based on a silicon waver allowing for a spectral window from
4000 to 600 cm-1 and is fed through the lid of the reactor with
a home-built screw fitting. This probe design makes the whole
setup more flexible than reactors with integrated ATR sensors.
The probe is connected to a ASI/Mettler-Toledo ReactIR-1000
FT-IR system, allowing continuous sampling and subsequent
deconvolution and analysis of the acquired spectra.
those observed in conventional solvents (see the Supporting
Information). The arylamine 9a also gave a clear solution in
liquid or supercritical CO2. The spectra of 9a in CDCl3 and in
scCO2 were almost identical, the N-H resonance appearing as
a slightly broadened signal at 4.0 and 3.7 ppm, respectively. In
contrast, a white insoluble solid was formed when liquid CO2
was introduced into the NMR tube containing the secondary
amine 9b at room temperature. The solid remained mostly
insoluble upon raising the temperatures above Tc of pure CO2,
preventing the recording of conclusive NMR spectra. These
observations are in accord with the established higher tendency
of secondary alkylamines compared to arylamines to form
insoluble carbamates by interaction with liquified or supercritical
CO2.2b,34 Although no apparent phase separation or precipitate
was observed during hydrogenation of 8b under catalytic
conditions (<30% conversion), the carbamic acid or carbamate
from 9b can also be expected to form under these conditions.
The formation of a strongly coordinating carboxylate function
provides a plausible explanation for the inefficient hydrogenation
of 8b compared to 8a in scCO2,24 but other deactivation
processes may equally apply.
Figure 3 shows the changes of the IR spectra during
hydrogenation of 8a using catalyst 7a under typical reaction
conditions. There were several well-resolved absorption bands
growing in due to the formation of the product, and other bands
disappeared concomitantly. No bands indicating the formation
of an intermediate or side product(s) were detected. The
positions of the nonoverlapping dissappearing signals at 1216
and 1643 cm-1 and the growing band at 1262 cm-1 are almost
identical to those of toluene solutions or of the neat compounds
8a and 9a, respectively (see the Supporting Information). Most
notably, the spectra provided no indication for the formation
of the carbamic acid or the ammonium carbamate from 9a and
CO2.
Conclusion
Taken together, the results of this study demonstrate that
scCO2 is an attractive alternative reaction medium for the
enantioselective hydrogenation of prochiral imines. From a
practical point of view, scCO2 provides a number of advantages
over the organic solvents generally used in imine hydrogenation.
In addition to the low toxicity and the environmentally benign
character of the supercritical reaction medium, its utilization
allows lower catalyst loading as well as efficient product
isolation and catalyst recycling (CESS process).
However, it is also evident that scCO2 cannot be regarded a
simple “substitute solvent”, and successful “transfer“ of reactions
Further information on the solubility and the chemical
behavior of the substrates and products in scCO2 was provided
by high-pressure NMR spectroscopy using sapphire tubes with
a titanium pressure head.33 Both substrates 8a and 8b were
readily soluble in scCO2 (40 °C, d(CO2) ) 0.75 g mL-1) at
concentrations of 10 mg mL-1. The spectra were identical to
(33) (a) Roe, D. C. J. Magn. Reson. 1985, 63, 388. (b) Horva´th, I.; Ponce,
E. C. ReV. Sci. Instrum. 1991, 62, 1104. (c) The tube used here was
manufactured in the laboratories of C. J. Elsevier, J. van’t Hoff Research
Institute, University of Amsterdam.
(34) (a) Francis, A. W. J. Phys. Chem. 1954, 58, 1099. (b) Ashraf-
Khorassani, M.; Taylor, L. T.; Zimmermann, P. Anal. Chem. 1990, 62, 1177.

Hydrogenation of Imines in scCO2
J. Am. Chem. Soc., Vol. 121, No. 27, 1999 6427
stirred at 0 °C overnight. Dry gaseous HCl was bubbled through the
clear solution over a period of 30 min at room temperature. After
degassing, filtration, and concentration, the product precipitated at -20
°C as colorless crystals. The product was isolated by filtration and
washed twice with Et2O to yield 2b (3.38 g, 3.70 mmol, 67%): 1H
NMR (200.1 MHz, CD2Cl2) δ 2.40 (m, 4H; CH2CH2CF2), 2.90 (m,
4H; CH2CF2), 7.25 (m, 4H; CH(3)), 7.52 (m, 4H; CH(2)); 13C{1H}
NMR (50.3 MHz, CD2Cl2) δ 26.7 (t, JFC ) 4 Hz; CH2CH2CF2), 33.0
(t, JFC ) 22 Hz; CH2CF2), 129.1 (d, JPC ) 7 Hz; CH(3)), 132.3 (d, JPC
) 25 Hz; CH(2)), 137.5 (d, JPC ) 33 Hz; C(1)), 141.7 (s; C(4)), 105-
125 (m; CFx); 19F NMR (282.4 MHz, CDCl3) δ -81.2 (t, 6F, J ) 11.4
Hz), -114.8 (m, 4F), -122.1 (m, 4F), -123.1 (m, 4F), -123.7 (m,
4F), -126.4 (m, 4F); 31P{1H} NMR (80.2 MHz, CDCl3) δ 81.3 (s);
MS (70 eV, DEI) m/z (%) 912 (100) [M•+].
from conventional solvents to this medium requires detailed
knowledge about the physicochemical properties of the reaction
mixture and about possible chemical interactions of reaction
components with the supercritical fluid. In the present case, it
was found that the substrate is of key importance for solubilizing
the catalytically active species. Furthermore, high levels of
asymmetric induction were observed only with catalysts con-
taining the BARF anion, although other complexes also showed
sufficient solubility for catalysis. Additional research is clearly
needed to elucidate the peculiar properties of this anion in
transition metal catalysis in more detail.
In situ high-pressure NMR and reflectance FT-IR spec-
trosocpy provided useful information about possible interactions
of substrates and products with scCO2 and allowed a rational-
ization of the different reactivities of structurally similar
substrates. A very remarkable observation of the present study
is the fact that catalyst 7a shows a considerably higher efficiency
in scCO2 compared to conventional organic solvent resulting
from a different rate profile rather than a simple increase in the
overall rate of the process. We can only speculate on the reasons
for this beneficial effect at present. Among various possibilities,
different chemical interactions and coordination abilities of the
different solvent systems with catalytic active species35 might
be of particular importance. The techniques and methodologies
described here should prove useful for further elucidation of
this system and also for other applications of scCO2 as a reaction
medium in homogeneous catalysis.
Synthesis of (-)-(4S)-2-(2-{Bis[4-(3,3,4,4,5,5,6,6,7,7,8,8,8-trideca-
fluorooctyl)phenylphosphanyl}phenyl]-4-isopropyl-4,5-dihydroox-
azole (3b). A solution of sec-BuLi (1.3 M in cyclohexane, 4.25 mL,
5.53 mmol) was added dropwise to a cooled (-65 °C) mixture of 3
(1.01 g, 5.03 mmol) and TMEDA (0.83 mL, 5.53 mmol) in hexane
(90 mL). After stirring for 1.5 h at -65 °C the yellow solution was
allowed to warm to 0 °C, whereby the color changed to orange. A
solution of 2b (4.68 g, 5.03 mmol) in THF (30 mL) was added at 0
°C, and the resulting mixture was stirred overnight. After hydrolysis
with saturated aqueous NH4Cl (10 mL) at room temperature, the product
was extracted with Et2O (2 × 10 mL), and the combined organic layers
were dried over Na2SO4, filtered, and evaporated to dryness. Recrys-
tallization from MeOH/Cl2FCCF2Cl (2:1) afforded the desired product
3b as a colorless solid (1.21 g, 1.14 mmol, 21%): 1H NMR (300.1
MHz, CDCl3) δ 0.73 (d, 3H, J ) 6.7 Hz; CH(CH3)2), 0.83 (d, 3H, J )
6.7 Hz, CH(CH3)2), 1.51 (m, 1H; CH(CH3)2), 2.40 (m, 4H; CH2CF2),
2.94 (m, 4H; CH2CH2CF2), 3.85-3.94 (m, 2H; CH2(5), CH(4)), 4.20
(m, 1H, CH2(5)), 6.90 (m, 1H; CH(4’)), 7.16-7.42 (m, 10H; arom CH),
7.94 (m, 1H; CH(6’)); 13C NMR (75.5 MHz, CDCl3) δ 18.3 (CH(CH3)2),
18.8 (CH(CH3)2, 26.2 (br, CH2CH2CF2), 32.7 (t, JFC ) 22 Hz; (CH2CH2-
CF2)), 32.8 (t, JFC ) 22 Hz CH2CH2CF2), 32.8 (CH(CH3)2, 70.1 (CH2-
(5)), 73.2 (CH(4)), 124.4, 128.1, 128.2, 128.3, 129.9 (d, JPC ) 3 Hz),
105-125 (m, CFx), 130.3, 133.7, 134.1 (d, JPC ) 21 Hz), 134.7 (d, JPC
) 21 Hz) (arom CH), 132.0 (d, JPC ) 19 Hz), 136.6 (d, JPC ) 10 Hz),
138.8 (d, JPC ) 24 Hz), 139.5 (d, JPC ) 14 Hz) (arom C), 162.7 (d, JPC
) 3 Hz, C(2)). 19F NMR (282.4 MHz, CDCl3) δ -81.5 (t, 6F, J )
11.4 Hz), -115.1 (m, 2F), -115.2 (m, 2F), -122.5 (m, 4F), -123.5
(m, 4F), -124.2 (m, 4F), -126.8 (m, 4F); 31P{1H} NMR (121.5 MHz,
CDCl3) δ -6.9 (s); MS (70 eV, DEI) m/z (%) 1065 (15) [M•+]. Anal.
Calcd for C40H30F26N1O1P1 (1065.7): C, 45.08; H, 2.84; F, 46.35; N,
1.31; P, 2.91. Found: C, 45.19; H, 2.88; F, 46.60; N, 1.39; P, 2.96.
Experimental Section
Safety warning: Conducting catalytic reactions and spectroscopic
inVestigations in scCO2 requires handling of highly compressed gases
and must be carried out only using suitable equipment and under
appropriate safety conditions.
All manipulations of air-sensitive materials and all catalytic experi-
ments were carried out under argon atmosphere. Solvents were dried,
purified, and degassed according to standard methods36 and stored under
argon. The gases (H2 (99.9%), argon (99.999%), and CO2 (99.995%))
(Messer Griesheim) were used without further purification. [{Ir(cod)-
(µ-Cl)}2] (4) (99%, Strem), n-BuLi (1.7 M in n-hexane, Aldrich), and
sec-BuLi (1.3 M in cyclohexane, Aldrich) were commercial products
and used as received. N,N,N’,N’-Tetramethylethylendiamine (TMEDA;
37
Fluka) was dried and distilled over LiAlH4 prior to use. Cl2PNEt2
and NaBARF38 were synthesized according to known procedures. The
syntheses of 1, 8a, 8b, 3a,26 5a,24 and 1-bromo-4-(3,3,4,4,5,5,6,6,7,7,8,8,8-
tridecafluorooctyl)benzene27 are described elsewhere.
(-)-(4S)-2-(2-{Bis[4-(3,3,4,4,5,5,6,6,7,7,8,8,8-tridecafluorooctyl)-
phenyl]phosphanyl}phenyl)-4-isopropyl-4,5-dihydrooxazole-(η4-1,5-
cyclooctadienyl)iridium(I) Hexafluorophosphate (5b). A solution of
3b (456 mg, 0.428 mmol) and 4 (143 mg, 0.213 mmol) in CH2Cl2 (10
mL) was stirred at 50 °C for 3 h in a sealed tube under Ar. After cooling
to room temperature, the red solution was treated with an aqueous
solution of NH4PF6 (0.4 M, 2 × 10 mL), washed with water (10 mL),
and dried over Na2SO4. Crystallization from CH2Cl2/CH3OH and drying
at 10-4 mbar afforded 5b as a bright red powder (336.4 mg, 0.22 mmol,
52%): 1H NMR (300.1 MHz, CD2Cl2) δ -0.04 (d, 3H, J ) 6.7 Hz;
CH3), 0.91 (d, 3H, J ) 7.1 Hz; CH3), 1.40-1.61 (m, 1H; CH2 of cod),
1.63-1.86 (m, 1H; CH2 of cod), 1.96-2.24 (m, 1H; CH2 of cod; 1H;
CH(CH3)2), 2.24-2.10 (m, 1H; CH2 of cod), 2.30-2.73 (m, 4H; CH2
of cod; 4H; CH2CF2), 2.90-3.13 (m, 1H; CH of cod; 4H; CH2CH2-
CF2), 3.24-3.36 (m, 1H; CH of cod), 4.29-4.16 (m, 1H; CH(4)), 4.52
(d, 2H; J ) 6.5 Hz; CH2(5)), 4.94-5.07 (m, 1H; CH of cod), 5.16-
5.24 (m, 1H; CH of cod), 7.03-7.13 (m, 2H; arom CH), 7.30-7.81
(m, 9H; arom CH), 8.22-8.27 (m, 1H, arom CH); 13C NMR (75.5
MHz, CD2Cl2): δ ) 12.9 (CH(CH3)2), 18.9 (CH(CH3)2, 26.6 (CH2-
CH2CF2), 26.8 (CH2CH2CF2), 26.9 (CH2 of cod), 28.9 (CH2 of cod),
32.6 (t, JFC ) 22 Hz; CH2CF2), 32.7 (CH2 of cod), 33.4 (CH(CH3)2),
36.5 (d, JPC ) 5 Hz; CH2 of cod), 63.3 (CH of cod), 64.0 (CH of cod),
69.1 (CH2(5)), 70.9 (CH(4)), 94.4 (d, JPC ) 13 Hz; CH of cod), 98.4
(d, JPC ) 11 Hz; CH of cod), 105-125 (m; CFx), 129.2 (d, JPC ) 10
Hz), 130.2 (d, JPC ) 11 Hz), 132.8 (d, JPC ) 2 Hz), 134.1 (d, JPC ) 11
Hz), 134.3 (d, JPC ) 8 Hz), 134.4 (d, JPC ) 7 Hz), 134.5 (d, JPC ) 2
Analytical Methods. Specific rotations were recorded on a JASCO
DIP-360 polarimeter (5.0 cm, 20 °C, concentration in grams per 100
mL of solution, estimated error (5%). NMR spectra were recorded on
a Bruker AC 200 or a Bruker AMX 300 spectrometer. Chemical shifts
δ are recorded in parts per million relative to external CFCl3 for 19F,
1
H3PO4 for 31P, and TMS for H and 13C, using the solvent resonance
as secondary standard if possible. Elemental analysis and AAS
measurements were carried out in the microanalytical laboratory Kolbe,
Mu¨lheim/Ruhr.
Synthesis of Bis[4-(3,3,4,4,5,5,6,6,7,7,8,8,8-tridecafluorooctyl)phen-
yl]chlorophosphine (2b). A solution of n-BuLi (1.7 M in n-hexane,
10 mL, 17 mmol) was added slowly to 1-bromo-4-(3,3,4,4,5,5,6,6,7,7,
8,8,8-tridecafluorooctyl)benzene27 (8.77 g, 17.4 mmol) in 50 mL of
Et2O at -50 °C. The yellow cloudy solution was added slowly to a
solution of Cl2PNEt2 (1.51 g, 8.6 mmol) in 20 mL of THF at 0 °C and
(35) For the coordination chemistry of CO2 and its relevance for catalysis,
see: Leitner, W. Coord. Chem. ReV. 1996, 153, 257.
(36) Perrin, D. D.; Armagedo, W. L. F. Purification of Laboatory
Chemicals, 3rd ed.; Pergamon Press: Oxford, 1988.
(37) Butters, T.; Winter, W. Chem. Ber. 1984, 117, 990.
(38) (a) Brookhart, M.; Grant, B.; Volpe, A. F., Jr. Organometallics 1992,
11, 3920. (b) Nishida, H.; Takada, N.; Yoshimura, M.; Sonoda, T.;
Kobayashi, H. Bull. Chem. Soc. Jpn. 1984, 57, 2600.

6428 J. Am. Chem. Soc., Vol. 121, No. 27, 1999
Kainz et al.
Hz), 135.8 (d, JPC ) 12 Hz) (arom CH), 121.2 (d, J ) 59 Hz), 128.3,
128.7, 129.0, 129.4, 144.9 (d, JPC ) 2 Hz), 144.0 (d, JPC ) 2 Hz)
(arom C), 164.3 (d, JPC ) 7 Hz, C(2)); 19F NMR (282.4 MHz, CD2-
Cl2) δ -73.6 (d, 6F, JPF ) 713 Hz), -81.3 (t, 6F, J ) 8.5 Hz), -114.6
(m, 2F), -114.8 (m, 2F), -122.2 (m, 4F), -123.2 (m, 4F), -123.8
(m, 4F), -126.5 (m, 4F); 31P{1H} NMR (81.0 MHz, CD2Cl2) δ 18.0
(s), -141.3 (sep, JPF ) 710.6 Hz); 11B{1H} NMR (64.2 MHz, CD2-
Cl2) δ -6.5 (s); [R]589 ) -134 (CHCl3, c ) 0.3, 20 °C); MS (ESI/pos
(MeOH)) m/z (%) 1366 (100) [M+, 193Ir], isotope cluster 1364-1368;
calcd (obsd) 54 (51), 30 (33), 100 (100), 52 (54), 14 (12). Anal. Calcd
for C48H42F32Ir1N1O1P2 (1511.1): C, 38.15; H, 2.80; F, 40.23; Ir, 12.72;
N, 0.93; P, 4.10. Found: C, 38.18; H, 2.86; F, 40.33; Ir, 12.58; N,
0.96; P, 3.98.
2.13 (m, 1H; CH2 of cod), 2.20-2.57 (m, 4H; CH2 of cod, 4H, CH2-
CF2), 2.88-3.00 (m, 1H; CH of cod; 4H, CH2CH2CF2), 3.17-3.26
(m, 1H; CH of cod), 4.00 (dt, 1H, J ) 9.0 Hz, 2.7 Hz; CH(4)), 4.13 (t,
1H, J ) 9.7 Hz; CH2(5)), 4.29 (dd, 1H, J ) 9.6 Hz, 3.3 Hz; CH2(5)),
4.82-4.92 (m, 1H; CH of cod), 4.93-5.02 (m, 1H; CH of cod), 6.77
(m, 4H; arom CH), 6.90-7.00 (m, 2H; arom CH; 8H; arom CH), 7.19-
7.32 (m, 8H; arom CH; 5H; arom CH), 7.45-7.63 (m, 4H; arom CH),
8.05-8.10 (m, 1H; arom CH); 13C NMR (75.5 MHz, CD2Cl2) δ 12.9
(CH(CH3)2), 18.9 (CH(CH3)2, 26.6 (m; CH2CH2CF2), 26.8 (m; CH2-
CH2CF2), 27.0 (CH2 of cod), 28.9 (CH2 of cod), 32.6 (t, JFC ) 22 Hz;
CH2CF2),), 32.7 (CH2 of cod), 33.4 (CH(CH3)2), 36.5 (d, JPC ) 5 Hz;
CH2 of cod), 63.5 (CH of cod), 64.2 (CH of cod), 69.0 (CH2(5)), 70.9
(CH(4)), 94.1 (d, JPC ) 13 Hz; CH of cod), 98.0 (d, JPC ) 10 Hz; CH
(-)-(4S)-2-(2-Diphenylphosphanylphenyl)-4-isopropyl-4,5-dihy-
drooxazole-(η4-1,5-cyclooctadienyl)iridium(I) Tetrakisphenylborate
(6a). A solution of 3a (60.7 mg, 0.16 mmol) and 4 (54.6 mg, 0.08
mmol) in CH2Cl2 (10 mL) was stirred at 50 °C for 4.5 h in a sealed
tube under Ar. After cooling to room temperature, a suspension of Na-
[BPh4] (111.3 mg, 0.33 mmol) in H2O (25 mL) was added and the
mixture stirred for several hours. Anion exchange was monitored by
TLC (silica gel, CH2Cl2; [Ir(cod)(3a)]Cl, Rf ) 0; [Ir(cod)(3a)]BPh4, Rf
) 0.40). The aqueous layer was separated, extracted with CH2Cl2 (4
× 5 mL), and dried over Na2SO4. Flash colomn chromatography on
silica gel (eluent CH2Cl2) and evaporation of the solvent afforded 6a
as a bright red powder (143.8 mg, 0.14 mmol, 89%): 1H NMR (300.1
MHz, CD2Cl2): δ ) -0.06 (d, 3H, J ) 6.7 Hz; CH3), 0.87 (d, 3H, J
) 7.1 Hz, CH3), 1.42-1.57 (m, 1H; CH(CH3)2), 1.65-1.79 (m, 1H;
CH2 of cod), 1.95-2.08 (m, 2H; CH2 of cod), 2.08-2.20 (m, 1H; CH2
of cod), 2.45-2.69 (m, 4H; CH2 of cod), 3.03-3.16 (m, 1H; CH of
cod), 3.31-3.36 (m, 1H; CH of cod), 4.10 (dt, 1H, J ) 9.5 Hz, 2.9
Hz; CH(4)), 4.22 (t, 1H, J ) 9.5 Hz; CH2(5)), 4.39 (dd, 1H, J ) 9.5
Hz, 3.4 Hz; CH2(5)), 4.91-5.01 (m, 1H; CH of cod), 5.01-5.11 (m,
1H; CH of cod), 6.87 (m, 4H; arom CH), 7.03 (m, 8H; arom CH),
7.09-7.16 (m, 2H; arom CH), 7.29-7.37 (m, 8H; arom CH), 7.37-
7.73 (m, 11H; arom CH), 8.15-8.20 (m, 1H; arom CH); 13C NMR
(75.5 MHz, CD2Cl2) δ 12.7 (CH(CH3)2); 18.9 (CH(CH3)2; 26.9 (CH2
of cod), 28.9 (CH2 of cod), 32.7 (CH2 of cod), 33.3 (CH(CH3)2); 36.5
(d, JPC ) 5 Hz; CH2 of cod), 63.5 (CH of cod), 64.1 (CH of cod), 68.9
(CH2(5)), 70.9 (CH(4)), 93.9 (d, JPC ) 13 Hz; CH of cod), 97.7 (d, JPC
) 11 Hz; CH of cod), 129.2 (d, JPC ) 12 Hz), 130.1 (d, JPC ) 11 Hz),
132.3 (d, JPC ) 3 Hz); 132.8 (d, JPC ) 2 Hz), 133.0 (d, JPC ) 3 Hz),
133.7 (d, JPC ) 10 Hz), 134.3 (d, JPC ) 10 Hz), 134.4 (d, JPC ) 8 Hz),
134.6 (d, JPC ) 2 Hz), 135.3 (d, JPC ) 12 Hz) (aromm CH), 122.9 (d,
J ) 58 Hz), 128.7, 129.0, 129.2, 129.9, 130.6 (arom C), 164.1 (d, JPC
) 2 Hz, C(2)), 122.1 (s, 4C; arom CH), 125.9-126.0 (m, 8C; arom
CH), 136.3 (m, 8C; arom CH), 164.5 (q, JCB ) 49.5 Hz); 31P{1H) NMR
(121.5 MHz, CD2Cl2) δ 16.5 (s); 11B{1H} NMR (64.2 MHz, CD2Cl2)
δ -7.0 (s); MS (ESI/pos (CH2Cl2)) m/z (%) 674 (100) [M+, 193Ir],
isotope cluster 672-676; calcd (obsd) 57 (65), 21 (23), 100 (100), 36
(39), 7 (7); (ESI/neg (CH2Cl2)) m/z (%) 319 (100) [M-, 11B], isotope
cluster 318-321; calcd (obsd) 23 (36), 100 (100), 26 (37), 3 (5)
(deviation according to unfavorable signal/noice ratio). Anal. Calcd
for C56H56B1Ir1N1O1P1 (×0.6 mol of CH2Cl2) (1044.0): C, 65.12; H,
5.52; N, 1.34; P, 2.97. Found: C, 64.96; H, 5.97; N, 1.16; P, 3.08.
of cod), 105-125 (m; CFx), 129.3 (d, JPC ) 11 Hz), 130.2 (d, JPC
)
11 Hz), 132.9 (d, JPC ) 3 Hz), 134.1 (d, JPC ) 10 Hz), 134.3, 134.4,
134.5, 135.7 (d, JPC ) 12 Hz) (arom CH), 121.1 (d, J ) 59 Hz), 128.2,
128.7, 128.9, 129.0, 129.1, 144.1 (d, JPC ) 3 Hz), 144.9 (d, JPC ) 3
Hz) (arom C), 164.2 (d, JPC ) 7 Hz, C(2)), 122.1 (s, 4C; arom CH),
126.0 (m, 8C, arom CH), 136.3 (m, 8C; arom CH,), 164.5 (q, JCB
)
49.8 Hz); 19F NMR (282.4 MHz, CD2Cl2) δ -81.3 (t, 6F, J ) 11.4
Hz), -114.5 (t, 2F, J ) 11.4 Hz), -114.7 (t, 2F, J ) 11.4 Hz), -122.2
(br; 4F), -123.2 (br; 4F), -123.8 (br; 4F), -126.4 (br; 4F); 31P{1H}
NMR (121.5 MHz, CD2Cl2) δ 15.4 (s); 11B{1H} NMR (64.2 MHz,
CD2Cl2) δ -6.5 (s); [R]589 ) -184 (CHCl3, c ) 0.25, 20 °C); MS
(ESI/pos (CH2Cl2)) m/z (%) 1366 (100) [M+, 193Ir], isotope cluster
1364-1368; calcd (obsd) 55 (57), 30 (30), 100 (100), 52 (48), 14 (13);
(ESI/neg (CH2Cl2)) m/z (%) 319 (100) [M-, 11B], isotope cluster 318-
321; calcd (obsd) 23 (21), 100 (100), 26 (25), 3 (3). Anal. Calcd for
C72H62B1F26Ir1N1O1P1 (1685.3): C, 51.31; H, 3.71; F, 29.31; N, 0.83;
P, 1.84. Found: C, 51.17; H, 3.89; F, 29.88; N, 0.88; P, 1.41.
(-)-(4S)-2-(2-{Diphenylphosphanylphenyl)-4-isopropyl-4,5-dihy-
drooxazole-(η4-1,5-cyclooctadienyl)iridium(I) Tetrakis(3,5-bis(tri-
fluoromethyl)phenyl)borate (7a). A solution of 3a (219.9 mg, 0.59
mmol) and 4 (199.6 mg, 0.29 mmol) in CH2Cl2 (10 mL) was stirred at
50 °C for 1.5 h in a sealed tube under Ar. After cooling to room
temperature, a suspension of NaBARF (1.08 g, 1.22 mmol) in H2O
(40 mL) was added, and the resulting mixture was stirred at room
temperature. Anion exchange was monitored by TLC (silica gel, CH2-
Cl2; [Ir(cod)(3a)]Cl, Rf ) 0; NaBARF, Rf ) 0; [Ir(cod)(3a)]BARF, Rf
) 0.75). The layers were separated, the aqueous layer was extracted
with CH2Cl2 (4 × 5 mL), and the combined organic phases were dried
over MgSO4. Flash column chromatography on silica gel (eluent CH2-
Cl2) and evaporation of the solvent afforded 7a as a bright red powder
(822.4 mg, 0.54 mmol, 91%): 1H NMR (300.1 MHz, CD2Cl2) δ -0.06
(d, 3H, J ) 6.7 Hz; CH3), 0.88 (d, 3H, J ) 7.1 Hz; CH3), 1.44-1.57
(m, 1H; CH(CH3)2), 1.65-1.78 (m, 1H; CH2 of cod), 1.98-2.07 (m,
2H; CH2 of cod), 2.08-2.19 (m, 1H; CH2 of cod), 2.24-2.69 (m, 4H;
CH2 of cod), 3.04-3.12 (m, 1H; CH of cod), 3.32-3.38 (m, 1H; CH
of cod), 4.18 (dt, 1H, J ) 9.2 Hz, 2.5 Hz; CH(4)), 4.40 (t, 1H, J ) 9.5
Hz; CH2(5)), 4.50 (dd, 1H, J ) 9.5 Hz, 3.7 Hz; CH2(5)), 4.92-5.02
(m, 1H; CH of cod), 5.03-5.11 (m, 1H; CH of cod), 7.09-7.16 (m,
2H; arom CH), 7.41-7.74 (m, 23H; arom CH), 8.17-8.22 (m, 1H;
arom CH);13C NMR (75.5 MHz, CD2Cl2) δ 12.6 (CH(CH3)2), 18.9 (CH-
(CH3)2, 26.9 (CH2 of cod), 28.9 (CH2 of cod), 32.7 (CH2 of cod), 33.4
(CH(CH3)2), 37.0 (d, JPC ) 5 Hz; CH2 of cod), 63.8 (CH of cod), 64.3
(CH of cod), 68.9 (CH2(5)), 71.0 (CH(4)), 93.9 (d, JPC ) 13 Hz; CH
of cod), 97.5 (d, JPC ) 11 Hz; CH of cod), 129.2 (d, JPC ) 10 Hz),
130.2 (d, JPC ) 10 Hz), 132.4 (d, JPC ) 3 Hz), 132.7 (d, JPC ) 2 Hz),
133.1 (d, JPC ) 3 Hz), 133.7 (d, JPC ) 10 Hz), 134.2 (d, JPC ) 8 Hz),
134.4 (d, JPC ) 7 Hz), 134.7 (d, JPC ) 2 Hz), 135.4 (arom CH), 125.0
(q, JFC ) 272 Hz), 122.5, 128.9, 129.3 (q, JFC ) 31 Hz), 129.5, 129.8,
130.6 (arom C), 164.3 (d, JPC ) 7 Hz, C(2)), 117.9 (m, 4C; arom CH),
135.3 (br, 8C; arom CH), 162.2 (q, JCB ) 49.9 Hz); 19F NMR (282.4
MHz, CD2Cl2) δ -63.3 (s); 31P{1H} NMR (121.5 MHz, CD2Cl2) δ
16.5 (s); 11B{1H} NMR (64.2 MHz, CD2Cl2) δ -6.7 (s); [R]589 ) -186
(CHCl3, c ) 0.25, 20 °C); MS (ESI/pos (CH2Cl2)) m/z (%) 674 (100)
[M+, 193Ir], isotope cluster 672-676 calcd (obsd) 57 (75), 21 (25), 100
(100), 36 (41), 7 (6); (ESI/neg (CH2Cl2)) m/z (%) 863 (100) [M-, 11B],
isotope cluster 862-865; calcd (obsd) 23 (24), 100 (100), 35 (36), 6
(6). Anal. Calcd for C64H48B1F24N1O1Ir1P1 (1536.8): C, 50.02; H, 3.15;
(-)-(4S)-2-(2-{Bis[4-(3,3,4,4,5,5,6,6,7,7,8,8,8-tridecafluorooctyl)-
phenyl]phosphanyl}phenyl)-4-isopropyl-4,5-dihydrooxazole-(η4-1,5-
cyclooctadienyl)iridium(I) Tetrakisphenylborate (6b). A solution of
3b (141.8 mg, 0.15 mmol) and 4 (50.4 mg, 0.08 mmol) in CH2Cl2 (10
mL) was stirred at room temperature overnight and then at 50 °C for
2 h in a sealed tube under Ar. After cooling to room temperature, a
suspension of Na[BPh4](100 mg, 0.29 mmol) in H2O (25 mL) was
added and stirred for several hours. Anion exchange was monitored
by TLC (silica gel, CH2Cl2; [Ir(cod)(3b)]Cl, Rf ) 0; [Ir(cod)(3b)]BPh4,
Rf ) 0.50). The organic layer was separated, and the aqueous layer
extracted with CH2Cl2 (4 × 5 mL). The combined organic fractions
were washed with H2O and dried over Na2SO4. Flash colomn
chromatography on silica gel (eluent CH2Cl2) and evaporation of the
solvent afforded 6b as a bright red powder (244.9 mg, 0.15 mmol,
97%): 1H NMR (300.1 MHz, CD2Cl2) δ -0.05 (d, 3H, J ) 6.7 Hz;
CH3), 0.78 (d, 3H, J ) 7.1 Hz; CH3), 1.18-1.47 (m, 1H; CH(CH3)2),
1.56-1.71 (m, 1H; CH2 of cod), 1.86-1.98 (m, 2H; CH2 of cod), 1.99-

Hydrogenation of Imines in scCO2
J. Am. Chem. Soc., Vol. 121, No. 27, 1999 6429
B, 0.70; F, 29.67; Ir, 12.51; N, 0.91; P, 2.02. Found: C, 50.32; H,
3.20; B, 0.61; F, 29.84; Ir, 12.29; N, 0.87; P, 2.03.
removed in vacuo, and the product was isolated by Kugelrohr distillation
(typically >99% recovery). The enantiomeric excess was determined
(-)-(4S)-2-(2-{Bis[4-(3,3,4,4,5,5,6,6,7,7,8,8,8-tridecafluorooctyl)-
phenyl]phosphanyl}phenyl)-4-isopropyl-4,5-dihydrooxazole-(η4-1,5-
cyclooctadienyl)iridium(I) Tetrakis(3,5-bis(trifluoromethyl)phenyl)-
borate (7b). Treating 5b (30 mg, 0.02 mmol) in CH2Cl2 (10 mL) with
a suspension of NaBARF (1.0 g, 1.13 mmol) in H2O (20 mL) at room
temperature for 0.5 h followed by workup as described for 7a gave
[Ir(cod)(3b)]BARF (7b; Rf ) 0.85) as a bright red powder (38.1 mg,
0.54 mmol, 86%): 1H NMR (300.1 MHz, CD2Cl2) δ -0.04 (d, 3H, J
) 6.7 Hz; CH3), 0.89 (d, 3H, J ) 7.1 Hz; CH3), 1.45-1.62 (m, 1H;
CH(CH3)2), 1.65-1.83 (m, 1H; CH2 of cod), 1.98-2.08 (m, 2H; CH2
of cod), 2.07-2.21 (m, 1H; CH2 of cod), 2.32-2.67 (m, 4H; CH2 of
cod; 4H; CH2CF2), 2.98-3.08 (m, 1H; CH of cod; 4H; CH2CH2CF2),
3.29-3.35 (m, 1H; CH of cod), 4.18 (dt, 1H, J ) 8.8 Hz, 3.1 Hz;
CH(4)), 4.42 (t, 1H, J ) 9.5 Hz; CH2(5)), 4.51 (dd, 1H, J ) 9.5 Hz,
3.8 Hz; CH2(5)), 4.91-5.03 (m, 1H; CH of cod), 5.04-5.13 (m, 1H;
CH of cod), 7.03-7.09 (m, 2H; arom CH), 7.31-7.44 (m, 7H; arom
CH), 7.54-7.75 (m, 16H; arom CH);13C NMR (75.5 MHz, CD2Cl2) δ
12.8 (CH(CH3)2), 18.8 (CH(CH3)2, 26.6 (m; CH2CH2CF2), 26.8 (m;
by HPLC (Daicel Chiralcel OD-H column, 254 nm, 0.5 mL min-1
,
n-heptane/2-propanol, 90:10; tR ) 12.2 (S), 15.9 (R) min). In all
experiments the (R)-enantiomer was formed predominantly.
Experiments in CH2Cl2 were carried out in a similar reactor (V )
50 mL) with a solvent volume of 20 mL, either under the conditions
described in the literature24 or following the above standard procedure
for comparison. The amount of catalyst and substrate was corrected
for the change in reaction volume (100 mL vs 20 mL) for the
supercritical and the liquid reaction medium, respectively. Stirring was
effected using a Teflon-coated magnetic stirring bar at the maximum
stirring rate.
In Situ High-Pressure FT-IR Spectroscopy Experiment. A
window-equipped stainless steel reactor (V ) 200 mL) was charged
with the catalyst 7a (10.5 mg, 6.8 × 10-3 mmol) under argon
atmosphere. The reactor was closed with the lid being connected to
the high-pressure reflectance probe of a ASI/Mettler-Toledo ReactIR-
1000 FT-IR system. CO2 was introduced using a compressor to adjust
to the desired density of the reaction medium of d(CO2) ) 0.75 g mL-1
.
CH2CH2CF2), 26.9 (CH2 of cod), 28.9 (CH2 of cod), 32.6 (t, JFC
)
Hydrogen was introduced to adjust to a partial pressure of 30 bar, the
reactor was heated to 40 °C, and the background spectrum was recorded
after temperature equilibration. To start the reaction, imine 8a (1.54 g,
7.81 mmol) was injected with positive CO2 pressure from a sample
compartment connected to the reactor via a ball valve. The reaction
was monitored by collecting spectra in 2 min intervals over a period
of 4 h.
High-Pressure NMR Investigations. A 5 mm high-pressure sap-
phire NMR tube33 (V ) 0.9 mL) was charged with a solution of
compound 8 or 9 (approximately 0.05 mmol in 0.6 mmol of C6D6)
under argon atmosphere. The NMR tube was pressurized with CO2
(0.65 g) using a compressor to adjust to the desired density of CO2
(0.75 g mL-1). Measurements were conducted at a temperature setting
of 40 °C on the NMR spectrometer, allowing at least 30 min for
temperature equilibration.
22.4 Hz; CH2CF2), 32.7 (CH2 of cod), 33.4 (CH(CH3)2); 36.5 (d, JPC
) 5 Hz; CH2 of cod), 63.7 (CH of cod), 64.3 (CH of cod), 69.0 (CH2-
(5)), 71.0 (CH(4)), 94.1 (d, JPC ) 13 Hz; CH of cod), 97.8 (d, JPC
)
10 Hz; CH of cod), 105-125 (m; CFx), 128.9 (d, JPC ) 2 Hz), 129.0
(d, JPC ) 2 Hz), 129.3 (d, JPC ) 11 Hz), 130.3 (d, JPC ) 11 Hz), 132.8
(d, JPC ) 2 Hz), 134.1 (d, JPC ) 10 Hz), 134.3, 134.4 (d, JPC ) 7 Hz),
134.6 (d, JPC ) 2 Hz), 135.7 (d, JPC ) 13 Hz) (arom CH), 121.1 (d, J
) 59 Hz), 125.0 (q, JPC ) 272 Hz), 126.3, 128.1, 129.2, 129.3 (qq, JFC
) 31 Hz, 3 Hz), 129.5, 144.2 (d, JPC ) 2 Hz), 145.0 (d, JPC ) 2 Hz)
(arom C), 117.9 (m, 4C; arom CH), 135.2 (br, 8C, arom CH), 162.2
(q, JCB ) 49.9 Hz), 164.3 (d, JPC ) 6 Hz, C(2)); 19F NMR (282.4
MHz, CD2Cl2) δ -62.9 (s, 18F), -81.1 (t, 6F, J ) 9.0 Hz;), -114.4
(t, 2F, J ) 11.4 Hz;), -114.6 (t, 2F, J ) 11.4 Hz;), -122.0 (br; 4F),
-123.0 (br; 4F), -123.7 (br; 4F), -126.3 (br; 4F); 31P{1H} NMR
(121.5.0 MHz, CD2Cl2) δ 15.5 (s); 11B{1H} NMR (64.2 MHz, CD2-
Cl2) δ -6.6 (s); MS (ESI/pos (CH2Cl2)) m/z (%) 1366 (100) [M+, 193Ir],
isotope cluster 1364-1368; calcd (obsd) 55 (69), 30 (32), 100 (100),
52 (58), 14 (15); (ESI/neg (CH2Cl2)) m/z (%) 863 (100) [M-, 11B],
isotope cluster 862-865; calcd (obsd) 23 (31), 100 (100), 35 (42), 6
(6). Anal. Calcd for C80H54B1F50Ir1N1O1P1 (2229.4): C, 43.10; H, 2.44;
F, 42.61; N, 0.63; P, 1.39. Found: C, 42.75; H, 2.49; F, 43.68; N,
0.89; P, 1.86.
Acknowledgment. This work was supported by the Max-
Planck-Gesellschaft, the Deutsche Forschungsgemeinschaft
(Gerhard-Hess-Programm), the Fonds der Chemischen Industrie,
and the European Community (COST-D10). Generous gifts of
chemicals by Clariant GmbH, Werk Gendorf are gratefully
acknowledged. We thank Dr. R. Mynott, W. Wisniewski, and
K. Wittmann at MPI Mu¨lheim (NMR) and Dr. R. Ferwerda
and Dr. H. Behl at Mettler Toledo/Applied Systems (FT-IR)
for support during the high-pressure spectroscopy studies and
J. Rosentreter (Mu¨lheim) for GC analyses. Special thanks are
due to Professors D. G. Blackmond, Mu¨lheim, C. J. Elsevier,
Amsterdam, and P. G. Jessop, Davis, for stimulating discussions.
This paper is dedicated to Prof. R. Selke on the occasion of his
65th birthday.
Hydrogenation of N-(1-Phenylethylidene)aniline (8a). In a typical
experiment, a window-equipped stainless steel reactor (V ) 100 mL)
with a Teflon-coated magnetic stirring bar was charged with the
substrate (820 mg, 4.2 mmol) and the catalyst (9.5 mg, 6.0 × 10-3
mmol, s/c ) 700, 0.14 mol % catalyst) under argon atmosphere. A
weighted amount of CO2 was introduced using a compressor to adjust
to the desired density of d(CO2) ) 0.75 g mL-1. The reactor was heated
to 33 °C, and hydrogen was introduced to adjust to a partial pressure
of 30 bar at this temperature as determined from a calibration curve.
The final reaction temperature of 40 °C was reached by further heating
within less than 5 min, and the mixture was kept at that temperature
for a standard reaction time of 20 h. Modifications of this procedure
are indicated in the figure captions and footnotes of the tables.
For standard workup, the reactor was cooled to room temperature
and vented carefully through a cold trap. The contents of the trap and
the reactor were collected by washing with CH2Cl2. The solvent was
Supporting Information Available: 1H NMR spectra of
compounds 8a,b and 9a in CDCl3 and scCO2 and reflectance
FT-IR spectra of 8a and 9a as neat compounds, in toluene, and
in scCO2 (PDF). This material is available free of charge via
the Internet at http://pubs.acs.org.
JA984309I