Aminosulf(ox)ides as Ligands for Iridium(I)-Catalyzed Asymmetric Transfer Hydrogenation

Article abstract of DOI:10.1021/jo991700t

A new class of efficient catalysts was developed for the asymmetric transfer hydrogenation of unsymmetrical ketones. A series of chiral N,S-chelates (6-22) was synthesized to serve as ligands in the iridium(I)-catalyzed reduction of ketones. Both formic a

Full text of DOI:10.1021/jo991700t

3010
J . Org. Chem. 2000, 65, 3010-3017
Am in osu lf(ox)id es a s Liga n d s for Ir id iu m (I)-Ca ta lyzed Asym m etr ic
Tr a n sfer Hyd r ogen a tion
Danie¨lle G. I. Petra,† Paul C. J . Kamer,† Anthony L. Spek,‡ Hans E. Schoemaker,†,§ and
Piet W. N. M. van Leeuwen*,†
Institute of Molecular Chemistry, University of Amsterdam, Nieuwe Achtergracht 166,
NL-1018 WV Amsterdam, The Netherlands, Bijvoet Center for Biomolecular Research, Padualaan 8,
NL-3584 CH Utrecht, The Netherlands, and DSM Research, P.O. Box 18,
NL-6160 MD Geleen, The Netherlands
Received October 29, 1999
A new class of efficient catalysts was developed for the asymmetric transfer hydrogenation of
unsymmetrical ketones. A series of chiral N,S-chelates (6-22) was synthesized to serve as ligands
in the iridium(I)-catalyzed reduction of ketones. Both formic acid and 2-propanol proved to be
suitable as hydrogen donors. Sulfoxidation of an (R)-cysteine-based aminosulfide provided a
diastereomeric ligand family containing a chiral sulfur atom. The two chiral centers of these ligands
showed a clear effect of chiral cooperativity. In addition, aminosulfides containing two asymmetric
carbon atoms in the backbone were synthesized. Both the sulfoxide-containing â-amino alcohols
and the aminosulfides derived from 1,2-disubstituted amino alcohols gave rise to high reaction
rates and moderate to excellent enantioselectivities in the reduction of various ketones. The
enantioselective outcome of the reaction was favorably affected by selecting the most appropriate
hydrogen donor. Enantioselectivities of up to 97% were reached in the reduction of aryl-alkyl ketones.
In tr od u ction
irreversible reactions due to CO2 evolution, thus prevent-
ing racemization. As a result of this irreversibility,
reactions can be carried out at higher substrate concen-
trations (i.e., >1 M) and higher temperatures without the
disadvantage of decreasing enantiomeric excess. Formic
acid proved to be a very useful hydrogen donor in
ruthenium(II)-catalyzed transfer hydrogenation using
TsDPEN as the ligand, which was developed by Noyori
and co-workers [TsDPEN ) N-(p-tolylsulfonyl)-1,2-di-
phenylethylenediamine].2 Catalyst stability is an impor-
tant issue when using the azeotropic mixture of formic
acid/triethylamine as a hydrogen donor. Only a few
catalysts gave rise to high activities when formic acid was
used as a hydrogen donor in hydrogen transfer reactions.
Most catalysts used so far are only stable in 2-propanol.
One of the first successful classes of transfer hydroge-
nation catalysts that induced high enantioselectivities
was developed by Pfaltz and co-workers in 1991.3 These
catalysts contained C2-symmetric bioxazole ligands that
displayed good activity in iridium(I)-catalyzed hydrogen
transfer reactions. Aryl-alkyl ketones were readily re-
duced providing the corresponding alcohols in 47-91%
enantiomeric excess. Also Noyori’s diamine ligand (i.e.,
TsDPEN) was used in iridium(I)-catalyzed hydrogen
transfer reactions in 2-propanol, resulting in an enantio-
selectivity of 92% in the reduction of acetophenone.4
N-Acetyl-(S)-methionine-(R,S)-sulfoxide (AMSO) gave
rise to enantioselectivities of up to 71% in rhodium-
catalyzed enantioselective transfer hydrogenation in
2-propanol.5
The enantioselective synthesis of optically active sec-
ondary alcohols, which form an important class of
intermediates for fine chemicals and pharmaceuticals,
has been studied extensively. One of the most attractive
methods for synthesizing chiral alcohols is asymmetric
transfer hydrogenation of prochiral ketones. The product
alcohols are obtained in high yields and good enantio-
meric excesses, and the fractional yields of side products
are low. Furthermore, the reaction conditions are rela-
tively mild and do not require the use of molecular
hydrogen, because the organic solvent, often 2-propanol,
can serve as a hydrogen donor. Nitrogen-based ligands
are often the ligands of choice in systems developed so
far, inducing high enantiomeric excesses in combination
with high catalytic activities.
An inherent problem of this potentially useful asym-
metric catalytic reaction is the reversibility of the process,
because of the similarities between 2-propanol and the
product alcohol. Even if the reaction proceeds with
excellent kinetic enantiofacial differentiation, the equili-
bration frequently causes a decrease in the enantiomeric
purity of the alcoholic products. Also, the lack of a
thermodynamic driving force prevents complete conver-
sion. To minimize the unfavorable reaction in 2-propanol,
catalysis must be performed using substrate concentra-
tions as low as 0.1 M.1
An attractive alternative for 2-propanol as a hydrogen
donor is formic acid. The use of the latter results in
* Corresponding author. Fax: +31 20 5256456. E-mail: PWNM@
anorg.chem.uva.nl.
(2) Fujii, A.; Hashiguchi, S.; Uematsu, N.; Ikariya, T.; Noyori, R. J .
Am. Chem. Soc. 1996, 118, 2521-2522.
† University of Amsterdam.
(3) Mu¨ller, D.; Umbricht, G.; Weber, B.; Pfaltz, A. Helv. Chim. Acta
1991, 74, 232-240.
‡ Bijvoet Center for Biomolecular Research.
§ DSM Research.
(1) Noyori, R.; Hashiguchi, S. Acc. Chem. Res. 1997, 30, 97-102.
(4) Ter Halle, R.; Bre´he´ret, A.; Schulz, E.; Pinel, C.; Lemaire, M.
Tetrahedron: Asymmetry 1997, 8, 2101-2108.
10.1021/jo991700t CCC: $19.00 © 2000 American Chemical Society
Published on Web 04/14/2000

Iridium-Catalyzed Asymmetric Transfer Hydrogenation
J . Org. Chem., Vol. 65, No. 10, 2000 3011
Ta ble 1. Hyd r ogen Tr a n sfer Red u ction of Acetop h en on e
Sch em e 1. Asym m etr ic Tr a n sfer Hyd r ogen a tion
of Acetop h en on e
Usin g Liga n d s 1-5a
conv.
TOFinitial
c
entry
ligand
R
(1 h) (%)b
(mol mol-1 h-1
)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
1a
1b
1c
1d
2a
2b
3
4a
4b
4c
4d
5a
5b
1d
NH2
OH
SH
S-CH2-Ph
OH
S-Ph
-
NH2
OH
SH
S-CH3
OH
S-CH2-CH3
S-CH2-Ph
60
0.5
0.5
>99
0.4
0.3
0.1
20
2
6
10
0.5
<0.1
58d
50
5
6
250
4
7
To the best of our knowledge, the use of formic acid as
a hydrogen donor has never been investigated in iridium-
(I)-catalyzed asymmetric hydrogen transfer reactions. In
this paper, the development of new catalysts that are
stable in formic acid/triethylamine is described. The
resulting systems are among the most active catalytic
systems found so far for asymmetric hydrogen transfer
reactions based on iridium(I) complexes with simple
aminosulf(ox)ides as ligands.
2
70
20
10
40
-
-
700d
N,S-Chelates have been widely used in asymmetric
catalysis.6-15 They create additional possibilities com-
pared to N,N- and N,O-chelates, because the sulfur atom
becomes chiral when coordinated to the metal.16-21 A new
class of ligands is formed by oxidation of the sulfur atom.
Here, the sulfur atom itself is chiral and possesses
different electronic and polarity characteristics.
a
The reaction was performed at 60 °C using 4 mmol of substrate
in 3 mL of a formic acid/triethylamine (5:2) solution. Substrate:
b
[IrCl(COD)]2:ligand ) 400:1:5. Conversions were determined
by GLC analysis. c Initial turnover frequencies were measured
d
after 5 min of reaction time. Substrate:[IrCl(COD)]2:ligand )
1000:1:5.
Two series of N,S-chelates were synthesized and
optimized in terms of activity and selectivity toward the
reduction of acetophenone by variations in ligand sub-
stitution. The influences of the catalyst precursor, the
ligand stoichiometry, the reaction temperature, and the
substrate concentration were studied. Both formic acid
and 2-propanol proved to be useful as hydrogen donors.
The enantioselective outcome of the reaction was opti-
mized by selecting the most appropriate hydrogen donor.
Resu lts a n d Discu ssion
Two different series of achiral nitrogen-containing
ligands (1-5) were tested in the iridium(I)-catalyzed
transfer hydrogenation of acetophenone using formic acid
as a hydrogen donor (Scheme 1).
A standard set of conditions was used, consisting of 0.5
mol % [IrCl(COD)]2 as the catalyst precursor, 2.5 mol %
ligand in a 5:2 azeotropic mixture of formic acid/triethyl-
amine at 60 °C. Conversions were monitored during the
reaction. Table 1 shows that the use of aminosulfide 1d
(R ) S-CH2-Ph, entries 4 and 14, Table 1) resulted in
the formation of the most active catalyst. The use of
diamine-type ligands in iridium(I)-catalyzed transfer
hydrogenation (entries 1 and 8, Table 1) gave rise to
lower activities under these conditions. Introduction of
various substituents (R1 and R2) in the carbon backbone
of 1d provides optically active aminosulfides, fitting the
general structure depicted in Figure 1.
(5) Kvintovics, P.; J ames, B. R.; Heil, B. J . Chem. Soc., Chem.
Commun. 1986, 1810-1811.
(6) Anderson, J . C.; Harding, M. Chem. Commun. 1998, 393-394.
(7) Anderson, J . C.; J ames, D. S.; Mathias, J . P. Tetrahedron:
Asymmetry 1998, 9, 753-756.
(8) Anderson, J . C.; Cubbon, R.; Harding, M.; J ames, D. S. Tetra-
hedron: Asymmetry 1998, 9, 3461-3490.
(9) Allen, J . V.; Coote, S. J .; Dawson, G. J .; Frost, C. G.; Martin, C.
J .; Williams, J . M. J . J . Chem. Soc., Perkin Trans. 1 1994, 2065-2072.
(10) Chelucci, G.; Berta, D.; Fabbri, D.; Pinna, G. A.; Saba, A.;
Ulgheri, F. Tetrahedron: Asymmetry 1998, 9, 3461-3490.
(11) Christoffers, J .; Mann, A.; Pickardt, J . Tetrahedron 1999, 55,
5377-5388.
(12) Dawson, G. J .; Frost, C. G.; Martin, C. J .; Williams, J . M. J .
Tetrahedron Lett. 1993, 34, 7793-7796.
(13) Fulton, D. A.; Gibson, C. L. Tetrahedron Lett. 1997, 38, 2019-
2022.
(14) Koning, B.; Hulst, R.; Kelogg, R. M. Recl. Trav. Chim. Pays-
Bas 1996, 115, 49-55.
(15) De Vries, A. H. M.; Hof, R. P.; Staal, D.; Kellogg, R. M.; Feringa,
B. L. Tetrahedron: Asymmetry 1997, 8, 1539-1543.
(16) Berkessel, A.; Bats, J . W.; Bolte, M.; Neumann, T.; Seidel, L.
Chem. Ber./ Recl. 1997, 130, 891-897.
F igu r e 1. Ligands 6-22.
(17) Cargill Thompson, A. M. W.; Bardwell, D. A.; J effery, J . C.;
Rees, L. H.; Ward, M. D. J . Chem. Soc., Dalton Trans. 1997, 721-
726.
Optimization in terms of activity and selectivity toward
the transfer hydrogenation of acetophenone was carried
out using a variety of N,S-chelates containing different
R1-R4 substituents. Conversions and enantioselectivities
were monitored during the reaction. The enantioselec-
tivity proved to be constant in time for all catalytic
reactions described, unless stated otherwise.
(18) Kossenjans, M.; Martens, J . Tetrahedron: Asymmetry 1998, 9,
1409-1417.
(19) Mahapatra, A. K.; Datta, S.; Goswami, S.; Mukherjee, M.;
Mukherjee, A. K.; Chakravorty, A. Inorg. Chem. 1986, 25, 1715-1721.
(20) Sellmann, D.; Bail, P.; Knoch, F.; Moll, M. Chem. Ber. 1995,
128, 653-663.
(21) Sheldrick, W. S.; Exner, R. J . Organomet. Chem. 1990, 386,
375-387.

3012 J . Org. Chem., Vol. 65, No. 10, 2000
Petra et al.
Sch em e 3. Syn th esis of Liga n d s 9 a n d 10
Ta ble 2. Hyd r ogen Tr a n sfer Red u ction of Acetop h en on e
Usin g Liga n d s 6-12a
entry
ligand
time (h)
conv. (%)b
ee (%)c
confign
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
6a
6b
6c
6d
6e
7
8a
8b
9
10
11a
11b
12
12a
12b
20
1
66
1
20
20
1
1
1
20
40
40
1
96
26
94
98
75
24
97
58
97
97
20
<1
56
56
99
10
12
10
12
10
9
14
18
5
S
S
S
S
S
S
S
S
S
S
S
-
S
R
S
7
18
ndd
35
27
65
substituted aminosulfide 8a was obtained (Scheme 2).
The use of R1-substituted aminosulfides 8a and 8b in the
iridium(I)-catalyzed reduction of acetophenone did not
significantly affect the activity and selectivity of the
catalyst (entries 7 and 8, Table 2).
Sterically demanding R1 substituents in amino acid
derivatives were found to be very useful in asymmetric
borane reductions and enantiocontrolled catalytic addi-
tion of diethylzinc to benzaldehyde, resulting in high
enantioselectivities.18,22,23 Compound 9 was obtained from
(R)-cystine methyl ester hydrochloride by reaction with
phenylmagnesium bromide (Scheme 3).22 The substituted
aminosulfide 9, however, did not induce higher enantio-
selectivities compared to the unsubstituted 6d (entry 9,
Table 2).
R2-disubstitution was introduced to create more steric
bulk in the carbon backbone. Starting from (R)-penicil-
lamine, aminosulfide 10, containing two methyl substit-
uents on the R carbon atom, was synthesized (Scheme
3). However, the use of the substituted aminosulfide 10
did not induce higher enantioselectivities compared to
the unsubstituted 6d (entry 10, Table 2).
1
0.5
a
The reaction was performed at 60 °C using 4 mmol of substrate
in 3 mL of a formic acid/triethylamine (5/2) solution. Substrate:
b
[IrCl(COD)]2:ligand ) 400:1:5. Conversions were determined by
GLC analysis. c Determined by capillary GLC analysis using a
d
chiral cycloSil-B column. nd ) not determined.
N,S-Ch ela tes Ba sed on (R)-Cystein e. The introduc-
tion of a variety of R1 substituents could easily be
achieved in sulfur-containing amino acids. A number of
L-cysteine and L-methionine derivatives were synthesized
and applied in the iridium(I)-catalyzed reduction of
acetophenone. Optically pure S-methylcysteine (6a ) and
methionine (7) were used to study the effect of the ligand
bite angle. Table 2 shows that the two-carbon-bridged
ligand 6a displayed much higher activities than the
three-carbon-bridged methionine ligand 7, while the
selectivities were similar (entries 1 and 6, Table 2).
It has been reported that the catalyst activity or
selectivity is most strongly affected by substituents that
are in close proximity to the coordinating atoms.24 For
the sulfur-containing amino acid derivatives we describe
here, we assume a bidentate N,S-coordination fashion.
This assumption is supported by the results from Table
1, which show that ethanolamine, representative for N,O-
coordination, does not catalyze the reduction of acetophe-
none under the applied reaction conditions. Also, varying
the R1 and R2 substituents in 6d did not change the
enantioselective outcome of the reaction. Variations in
the R3 and R4 substituents on the nitrogen and sulfur
atoms are expected to give rise to larger effects on the
enantioselectivity, as they are positioned closer to the
catalytic center than the oxygen atom. Bulky R3 substit-
uents were introduced by substitution of the primary
amine functionality in 6d by a tosyl (11a ) and a tert-
butoxycarbonyl group (11b) (Scheme 4). The use of
N-substituted N,S-chelates in the iridium(I)-catalyzed
transfer hydrogenation of acetophenone resulted in a
considerable loss of catalytic activity, while the enantio-
selectivity was hardly affected (entries 11 and 12, Table
2). Because the introduction of a substituted amide not
only increases steric hindrance but also reduces the
donating capacity of the nitrogen atom, the cause of the
catalyst deactivation remains unresolved.
Sch em e 2. Liga n d s 8a a n d 8b
To investigate the difference in reactivity of amino
acids versus amino alcohols, 6b was compared with 6d .
Reduction of the amino acid 6b to the amino alcohol 6d
provided a second major increase of the reaction rate
without a change in enantioselectivities (entries 2 and
4, Table 2). To optimize enantioselectivities for this
system, the substituents R1-R4 were systematically
varied. Starting from the commercially available S-
benzyl-(R)-cysteinol (6d ), the hydroxyl group in the R1
substituent was functionalized with a benzyl group, and
after deprotection of the amine functionality, the oxygen-
(22) Li, X.; Xie, R. Tetrahedron: Asymmetry 1997, 8, 2283-2285.
(23) Trentmann, W.; Mehler, T.; Martens, J . Tetrahedron: Asym-
metry 1997, 8, 2033-2043.
(24) Zassinovich, G.; Mestroni, G. Chem. Rev. 1992, 92, 1051 and
references therein.

Iridium-Catalyzed Asymmetric Transfer Hydrogenation
J . Org. Chem., Vol. 65, No. 10, 2000 3013
Sch em e 4. Syn th esis of Liga n d s 11a a n d 11b
from either ethanol or 2-propanol. The configuration of
diastereomer 12a [i.e., S-benzyl-(R)-cysteinol (S)-sulfox-
ide] was determined by X-ray analysis.
When the 1:1 diastereomeric mixture of 12a and 12b
was used in catalysis, the enantiomeric excess increased
to 35% from 12% for 6d . At the same time, the catalyst
activity decreased by a factor of 2 (entry 13, Table 2).
Interestingly, a marked difference in stereoselection was
found when the two diastereomers were used separately.
The diastereomer of S-benzyl-(R)-cysteinol (S)-sulfoxide
(12a ) in combination with [IrCl(COD)]2 created the
opposite product configuration (i.e., R) compared to the
diastereomeric 1:1 mixture (giving the S product) and had
a similar reaction rate (entry 14, Table 2). The use of
the diastereomer of S-benzyl-(R)-cysteinol (R)-sulfoxide
12b induced an increase in enantioselectivity of up to
65%, giving the S product. At the same time, the reaction
rate was similar to that of 6d , resulting in >99%
conversion after 1 h (entry 15, Table 2). Thus, a clear
effect of chiral cooperativity was observed between the
sulfoxide functionality and the R position of the amino
alcohol. In addition, the use of the diastereomeric mixture
of ligands nicely shows that the asymmetric transfer
hydrogenation reaction is kinetically controlled, resulting
in the S product.
Sch em e 5. Syn th esis of Liga n d 6e
Sch em e 6. Syn th esis of Liga n d 12a
N,S-Ch ela tes Der ived fr om (Nor )ep h ed r in e a n d
2-Am in od ip h en yleth a n ol. Introduction of two chiral
centers in the carbon backbone using different R1 and
R2 substituents was achieved starting from (1R,2S)-(nor)-
ephedrine and (1R,2S)-2-aminodiphenylethanol. Ephe-
drine-based â-aminothiols and -disulfides catalyze the 1,2
addition of diethylzinc to aromatic aldehydes, producing
the corresponding alcohols in high enantiomeric excess
and yield.31-33 Dieter and co-workers34 showed that
optically pure aminosulfides containing a tertiary amine
functionality could easily be obtained from N-substituted
ephedrine derivatives. For the synthesis of aminosulfides
containing a secondary amine, a different method was
used by Kellogg and co-workers.35 Scheme 7 shows that
N,S-chelates containing a primary amine functionality
could be obtained starting from norephedrine and 2-ami-
nodiphenylethanol using the latter synthetic strategy.
Starting from (1R,2S)-norephedrine, aziridine 13 was
synthesized using Mitsunobu conditions. Stereoselective
ring opening of the aziridine with several thiol nucleo-
philes provided the aminosulfides 16-18, in which R1 )
methyl. Similarly, aminosulfides 19 and 20, with R1 )
phenyl, were synthesized from aziridine 14, which was
obtained from (1R,2S)-2-aminodiphenylethanol. The R4
substituent at the sulfur atom was varied using different
sulfur nucleophiles. The results of the use of these ligands
in iridium(I)-catalyzed transfer hydrogenation of ac-
etophenone using formic acid as a hydrogen donor are
summarized in Table 3.
a
Diastereomers 12a and 12b were separated by repeated
crystallization from ethanol.
An N,S-chelate containing a bulky R4 substituent on
the sulfur atom was synthesized by reduction of S-trityl-
(R)-cysteine (6c) to S-trityl-(R)-cysteinol (6e) (Scheme 5).
Standard LiAlH4 reduction failed in the case of these
sulfur-containing amino acids. A milder method, devel-
oped by Abiko et al. for large-scale reductions of amino
acids using sodium borohydride and sulfuric acid, pro-
vided the desired product (6e).25 Replacement of the
benzyl group by a trityl group resulted in a dramatic
decrease in activity, while the selectivity was identical
in both cases (entries 3 and 5, Table 2).
An additional possibility for creating a more sterically
hindered environment around the sulfur moiety is oxida-
tion of the sulfur atom, which provides a different class
of ligands. The presence of a sulfoxide moiety also
changes the electronic properties of the ligand. Moreover,
sulfoxidation results not only in an extra substituent on
the sulfur atom but also in an additional chiral center in
the ligand. A 1:1 mixture of two diastereomers of S-
benzyl-(R)-cysteinol sulfoxide (12a and 12b) was formed
using hydrogen peroxide as the oxidant (Scheme 6). All
attempts to achieve asymmetric sulfoxidation, using
either chiral titanium catalysts26-29 or chloroauric acid,30
failed. The polar sulfoxide products could not be sepa-
rated from the reaction mixture by extraction or chro-
matographic methods. We succeeded, however, in sepa-
rating the two diastereomers by repeated crystallization
Table 3 shows that the variation of both the R1
substituent and the R4 substituent largely affects the
outcome of the reaction. Both the activity and the
(31) Hof, R. P. Ph.D. Thesis, Groningen University, Groningen, The
Netherlands, 1995; Chapter 3.
(32) Poelert, M. A.; Hof, R. P.; Peper, N. C. M. W.; Kellogg, R. M.
Tetrahedron: Asymmetry 1994, 5, 31-34.
(33) Kang, J .; Lee, J . W.; Kim, J . I. J . Chem. Soc., Chem. Commun.
1994, 2009.
(34) Dieter, R. K.; Deo, N.; Lagu, B.; Dieter, J . W. J . Org. Chem.
1992, 57, 1663-1671.
(35) Poelert, M. A.; Hof, R. P.; Peper, N. C. M. W.; Kellogg, R. M.
Heterocycles 1994, 37, 461-475.
(25) Abiko, A.; Masamume, S. Tetrahedron Lett. 1992, 33, 5517.
(26) Baldenius, K.-U.; Kagan, H. B. Tetrahedron: Asymmetry 1990,
1, 597-610.
(27) Brunel, J .-M.; Diter, P.; Duetsch, M.; Kagan, H. B. J . Org.
Chem. 1995, 60, 8086-8088.
(28) Brunel, J .-M.; Kagan, H. B. Synlett 1996, 404-406.
(29) Pitchen, P.; Dun˜ach, E.; Deshmukh, M. N.; Kagan, H. B. J . Am.
Chem. Soc. 1984, 106, 8188-8193.
(30) Natile, G.; Bordignon, E. Inorg. Chem. 1976, 15, 246-248.

3014 J . Org. Chem., Vol. 65, No. 10, 2000
Sch em e 7. Syn th esis of Liga n d s 16-21
Petra et al.
Sch em e 8. Syn th esis of Liga n d 22a
a
Diastereomers 22a and 22b were separated by column chro-
matography.
Ta ble 4. Hyd r ogen Tr a n sfer Red u ction of Acetop h en on e
Usin g Va r iou s Ir id iu m P r ecu r sor sa
conv.
Ta ble 3. Hyd r ogen Tr a n sfer Red u ction of Acetop h en on e
entry ligand
Ir precursor
[IrCl(COD)]2
[Ir(COD)]BF4
[IrCl(COE)2]2
[IrCl(1,5-diMe-COD)]2
[IrCl(CO)3]n
(1 h) (%)b ee (%)c confign
Usin g Liga n d s 16-22ba
1
2
3
4
5
6
6d
6d
6d
6d
6d
6d
98
98
8
56
18
<1
12
12
16
15
15
ndd
S
S
S
S
S
-
entry
ligand
time (h)
conv. (%)b
ee (%)c
confign
1
2
3
4
5
6
7
8
16
17
18
19
20
21
22a
22b
3
3
3
3
18
5
52
>99
>99
20
2
91
23
41
65
42
58
42
32
2
S
S
S
R
S
S
S
S
[IrCl2(Cp)*]2
a
The reaction was performed at 60 °C using 4 mmol of substrate
in 3 mL of a formic acid/triethylamine (5/2) solution. Substrate:
6
6
32
9
b
[IrCl(COD)]2:ligand ) 400:1:5. Conversions were determined by
GLC analysis. c Determined by capillary GLC analysis using a
a
d
The reaction was performed at 60 °C using 4 mmol of substrate
in 3 mL of a formic acid/triethylamine (5:2) solution. Substrate:
chiral cycloSil-B column. nd ) not determined.
b
[IrCl(COD)]2:ligand ) 400:1:5. Conversions were determined by
sulf(ox)ide catalysts in terms of activity and selectivity
by varying the catalyst precursor, the ligand stoichiom-
etry, the reaction temperature, and the hydrogen source.
In flu en ce of th e Ca ta lyst P r ecu r sor . The influence
of the nature of the catalyst precursor on the reaction
rate and enantioselectivity of this reaction was studied.
Different anions, different dienes, and an iridium car-
bonyl precursor were used. Table 4 shows that changing
GLC analysis. c Determined by capillary GLC analysis using a
chiral cycloSil-B column.
enantioselectivity increased on changing the R4 substit-
uents from phenyl (16) to isopropyl (17) and to benzyl
(18). Introduction of bulky phenyl groups at both posi-
tions R1 and R4 (19) resulted in a change of product
configuration from S to R. The presence of three phenyl
groups in close proximity to the metal center probably
results in a change in conformation of the complex. Also,
interaction between an aromatic substituent and the
phenyl ring of the substrate could play a role. When the
steric strain is released slightly by changing R4 to benzyl
(20), the product configuration changes back to S. To
improve the enantioselectivities even further in this
series of ligands, the nitrogen atom was functionalized.
Starting from (1R,2S)-ephedrine, ligand 21, which con-
tains a methyl substituent on the nitrogen atom, was
obtained (see Scheme 7).31 As shown in Table 3, introduc-
tion of the methyl substituent at the R3 position in 21
induced a decrease in both reaction rate and asymmetric
induction (entry 6). In analogy to 12a ,b, the sulfide
functionality in 18 was oxidized to form 22a ,b (Scheme
8). The diastereomers were separated by column chro-
matography.
the anion from Cl- to BF4 does not affect the catalytic
-
reaction. The presence of a diene in the active catalyst
in rhodium-catalyzed transfer hydrogenation using 2-pro-
panol as a hydrogen source was proved by Lemaire and
co-workers.36 Table 4 shows that the presence of a diene
is also important for obtaining high reaction rates in
iridium(I)-catalyzed hydrogen transfer reactions using
formic acid as the H-donor (entries 1 and 3). The use of
various catalyst precursors only marginally affected the
stereoselective outcome of the reaction. The catalyst
activity did not improve using different iridium precur-
sors compared to [IrCl(COD)]2.
In flu en ce of th e Liga n d Stoich iom etr y. In rho-
dium(I)- and iridium(I)-catalyzed transfer hydrogenation,
2 equiv of ligand per metal atom is typically used. In a
mechanistic investigation concerning the ligand-to-
rhodium ratio, however, Lemaire and co-workers postu-
lated that the active catalyst is most likely a rhodium
complex having only one diamine ligand coordinated to
the metal atom.36 The ligand-to-metal ratio was varied
in a range from 1 to 5 equiv, using ligands 1d , 6d , 12b,
and 18 in the iridium(I)-catalyzed reduction of acetophe-
As observed for 12a ,b, the two diastereomers 22a ,b
showed a large difference in reaction rate and enantio-
selectivity. Oxidation of the sulfur atom as in 22a ,b,
however, did not provide any further improvement in the
iridium(I)-catalyzed transfer hydrogenation of acetophe-
none (entries 7 and 8, Table 3).
Ca ta lyst Op tim iza tion . In the following, we describe
our attempts to further optimize the iridium(I) amino-
(36) Bernard, M.; Guiral, V.; Delbecq, F.; Fache, F.; Sautet, P.;
Lemaire, M. J . Am. Chem. Soc. 1998, 120, 1441-1446.

Iridium-Catalyzed Asymmetric Transfer Hydrogenation
J . Org. Chem., Vol. 65, No. 10, 2000 3015
Ta ble 5. Hyd r ogen Tr a n sfer Red u ction of Acetop h en on e
Usin g Va r iou s Liga n d :Ir id iu m Ra tiosa
Ta ble 7. Hyd r ogen Tr a n sfer Red u ction of Acetop h en on e
Usin g 2-P r op a n ol a s a H Don or a
conv.
entry
ligand
time (h)
conv. (%)b
ee (%)c
confign
entry
ligand
ligand:Ir
(0.5 h) (%)b
ee (%)c
confign
1
2
6d
12b
6d
16
17
18
19
20
21
1/20
1/20
1/20
1
1
1
1
1
1
1
7/45
6/24
2/15
83
88
96
40
82
22
85
33/33
16/13
19/18
24
73
65
2
80
82
39
R/S
R/S
S/S
R
S
S
R
R
S
R
1
2
3
4
5
6
7
8
9
1d
1d
1d
6d
6d
1:1
2.5:1
5:1
1:1
2.5:1
5:1
1:1
2.5:1
5:1
1:1
2.5:1
5:1
73
90
91
64
70
82
61
99
99
44
45
47
-
-
-
-
-
S
S
S
S
S
S
S
3d
4
-
9
5
6
7
8
9
10
12
12
65
65
65
60
65
65
6d
12b
12b
12b
18
18
18
22a
10
11
12
a
The reaction was performed at 20 °C using 4 mmol of substrate
in a 0.1 M 2-propanol solution. Substrate:[IrCl(COD)]2:ligand:
S
S
b
tBuOK ) 400:1:5:12.5. Conversions were determined by GLC
a
The reaction was performed at 60 °C using 4 mmol of substrate
in 3 mL of a formic acid/triethylamine (5:2) solution. Substrate:
analysis. c Determined by capillary GLC analysis using a chiral
d
cycloSil-B column. In situ catalyst incubation time ) 30 min.
b
[IrCl(COD)]2:ligand ) 400:1:2-10. Conversions were determined
by GLC analysis. c Determined by capillary GLC analysis using a
chiral cycloSil-B column.
shown in Table 7. The change in product configuration
during the reaction is most likely due to a change in the
ligand coordination fashion. Under the basic reaction
conditions, the alcohol functionality is deprotonated,
resulting in ligand coordination involving nitrogen and
oxygen donor atoms instead of nitrogen and sulfur. The
experiment described in entry 1 was repeated using a
longer incubation time of the in situ catalyst. After
addition of the basic cocatalyst (i.e., tBuOK), the reaction
mixture was stirred for 30 min before adding the sub-
strate (entry 3, Table 7). This resulted in an S product
configuration that was constant in time. Both the catalyst
activity and selectivity dropped when longer incubation
times were used.
Ta ble 6. Hyd r ogen Tr a n sfer Red u ction of Acetop h en on e
a t Va r iou s Tem p er a tu r esa
conv. (%)
entry
ligand
T (°C)
(time (h))b
ee (%)c
confign
1
2
3
4
5
6
7
8
9
6d
6d
6d
12b
12b
12b
18
60
40
20
60
40
20
60
40
20
98 (1)
24 (1)
15 (5)
99 (1)
38 (1)
57 (5)
50 (1)
12 (1)
11 (5)
12
19
19
65
67
80
58
69
69
S
S
S
S
S
S
S
S
S
18
18
Using the aminosulfides derived from (nor)ephedrine
and 2-aminodiphenylethanol (16-22a ), the reaction rates
and the enantioselectivities substantially increased with
2-propanol as the H-donor instead of formic acid. Ligand
20 induced 80% ee using 2-propanol as the hydrogen
source. As was also found with formic acid as the
hydrogen donor (see also Table 3), the product configu-
ration changed to R when more bulky phenyl substitu-
ents were introduced. However, the system using 2-pro-
panol seems to be more sensitive to this change in ligand
structure, as suggested by the fact that the product
configuration changed to R also in the cases of ligands
16, 20, and 22a , in contrast to the results obtained with
the system using formic acid as the H-donor. The change
in product configuration can only be explained by differ-
ent catalytic intermediates in the enantioselectivity-
determining step.
Mech a n istic Con sid er a tion s. From a mechanistic
point of view, two general pathways can be envisaged
for transfer hydrogenation (Figure 2): a stepwise process
through a hydride complex (the “hydridic route”, path A)
and a concerted process in which the hydrogen is directly
transferred from the hydrogen donor to the substrate
(“direct hydrogen transfer”, path B).24
a
The reaction was performed at the indicated temperature
using 4 mmol of substrate in 3 mL of a formic acid/triethylamine
b
(5:2) solution. Substrate:[IrCl(COD)]2:ligand ) 400:1:5. Conver-
sions were determined by GLC analysis. c Determined by capillary
GLC analysis using a chiral cycloSil-B column.
none. When the ligand-to-metal ratio was changed from
1 to 2.5, both the activity and enantioselectivity increased
only slightly (Table 5). This is most likely due to
enhanced stabilization of the iridium complex, prohibiting
the formation of unsaturated iridium species that might
lead to racemic phenyl ethanol.36 Addition of more than
2.5 equiv of ligand did not affect the outcome of the
reaction any further.
In flu en ce of th e Rea ction Tem p er a tu r e. The effect
of the temperature on the enantioselectivity of the
reaction was tested using ligands 6d , 12b, and 18.
Decreasing the reaction temperature in a range from 60
to 20 °C resulted in an increase in the enantioselectivity
in all cases (Table 6). Using the diastereomer S-benzyl-
(R)-cysteinol (R)-sulfoxide (12b) as a ligand, the largest
difference in enantioselectivity was observed, ranging
from 65% ee at 60 °C to 80% ee at 20 °C (entries 4-6,
Table 6). In the case of ligand 18, the beneficial temper-
ature effect also resulted in a marked difference in
enantioselection (entries 7-9, Table 6).
In flu en ce of th e Hyd r ogen Sou r ce. An alternative
hydrogen source for the azeotropic mixture of formic acid
and triethylamine is 2-propanol. The results of the
experiments using 2-propanol as a hydrogen donor in
iridium(I)-catalyzed reduction of acetophenone with vari-
ous N,S-chelates are presented in Table 7. Using the
amino-acid-based aminosulf(ox)ides (6d and 12b), the
initial product configuration proved to be R. At higher
conversions, the product configuration changed to S, as
For transition-metal-catalyzed transfer hydrogenation,
some key features have been discovered that support the
hydridic route (path A). Noyori and co-workers37 showed
that the structure of the supposed active species in
ruthenium(II)-catalyzed transfer hydrogenation is a ru-
thenium(II) 16-electron complex bearing TsDPEN as the
ligand and an η6-arene moiety. In the presence of 2-pro-
panol, an 18-electron ruthenium hydride species is formed
(37) Haack, K.-J .; Hashiguchi, S.; Fujii, A.; Ikariya, T.; Noyori, R.
Angew. Chem., Int. Ed. Engl. 1997, 36, 285-288.

3016 J . Org. Chem., Vol. 65, No. 10, 2000
Petra et al.
Sch em e 9. Asym m etr ic Tr a n sfer Hyd r ogen a tion
of Keton es 23a -25
F igu r e 2. Two mechanistic pathways for hydrogen transfer
reactions.
that catalyzes the reduction of various ketones. Isolation
of these complexes confirms that the ruthenium-catalyzed
transfer hydrogenation takes place by way of a metal
hydride rather than a metal alkoxide, as is active in the
Meerwein-Ponndorf-Verley reduction. Lemaire and co-
workers36 showed that the active complex in rhodium-
(I)-catalyzed transfer hydrogenation is most likely a
rhodium hydride complex with one diamine and one
diene coordinated to the metal. When the active catalyst
is presumed to be a metal hydride, the enantioselective
outcome of the reaction is expected to be independent of
the hydrogen donor that is involved. However, it should
be noted that solvent effects can also give rise to differ-
ences in the enantioselective outcome of the reaction.
The results presented in Tables 3 and 7 show that the
enantioselectivity of the reaction is largely affected by
the use of different hydrogen donors. Therefore, a mecha-
nistic route in which the hydrogen is transferred to the
substrate in a concerted process cannot be excluded.
Enantiomer-discriminating H-transfer reactions via the
direct hydrogen transfer mechanism can give rise to
different enantioselectivities using various hydrogen
donors. In the examples shown here (Table 3 versus Table
7), a large difference in enantioselectivity is observed
when changing the H-donor, even resulting in the op-
posite product configurations in some cases. In addition,
Noyori and co-workers1 showed that the presence of a
NH moiety in the ligand is of crucial importance in
obtaining high activities and enantioselectivities. It was
therefore proposed that the ruthenium(II)-catalyzed trans-
fer hydrogenation occurs by hydrogen bond formation
between the NH linkage of the amine ligand and the
carbonyl functionality of the substrate.1 In the transition
state, a six-membered ring is formed, after which both
the metal hydride and the amine proton of the ligand
are donated to the substrate (so-called bifunctional
catalysis). DFT calculations on possible catalytic inter-
mediates confirmed this hypothesis.38 However, for iri-
dium(I)- and rhodium(I)-catalyzed hydrogen transfer
reactions, this theory proved to be inconsistent. Using
the latter two metal catalysts in combination with
diimine-type ligands (e.g., bipyridines, phenanthrolines,
and bioxazoles) lacking an amine proton, the reduction
of ketones is successful as well.3,24
In summary, the enantioselective outcome of the
reaction in the iridium(I)-catalyzed transfer hydrogena-
tion is strongly influenced by the type of hydrogen donor
(i.e., formic acid or 2-propanol). This change in enantio-
selectivity might be caused by the interference of a direct
hydrogen transfer path (path B, Figure 2), resulting in a
different catalytic intermediate in the enantioselectivity-
determining step of the reaction.
Va r ia tion s of th e Su bstr a te. To investigate the scope
of the transfer hydrogenation systems presented above,
ketones other than acetophenone (23a ) were subjected
to reduction using ligands 12b, 18, and 20 (see Scheme
9). The results are presented in Table 8. The enantiose-
lective outcome of the reaction was significantly influ-
enced by the electronic and steric properties of the
substrate. The enantioselectivity increased when the
substrate contained more steric bulk, as in 23b, 23e, and
23f. An electron-withdrawing substituent in the sub-
strate (23c) resulted in a higher enantioselectivity,
whereas an electron-donating substituent (23d ) gave rise
to a lower ee. An exception to the trend is the rigid and
bulky ketone R-tetralone (25). Reduction of this substrate
resulted in moderate enantioselectivities of the corre-
sponding alcohol (26), which might be due to the fixed
ring conformation of this substrate.
The use of 2-propanol as a hydrogen donor in the
asymmetric transfer hydrogenation of acetophenone is
(38) (a) Alonso, D. A.; Brandt, P.; Nordin, S. J . M.; Andersson, P. G.
J . Am. Chem. Soc. 1999, 121, 9580-9588. (b) Petra, D. G. I.; Reek, J .
N. H.; Handgraaf, J .-W.; Meijer, E. J .; Dierkes, P.; Kamer, P. C. J .;
Brussee, J .; Schoemaker, H. E.; Van Leeuwen, P. W. N. M. Chem.-
Eur. J . Manuscript accepted.

Iridium-Catalyzed Asymmetric Transfer Hydrogenation
J . Org. Chem., Vol. 65, No. 10, 2000 3017
Ta ble 8. Hyd r ogen Tr a n sfer Red u ction of Keton es
Ta ble 9. Hyd r ogen Tr a n sfer Red u ction Usin g Liga n d 20
in 2-P r op a n ol a t Differ en t Su bstr a te Con cen tr a tion sa
23a -f a n d 25
conv.
conv. (h) (%)b
(time (h))
entry ketone ligand H donor (1 h) (%)c ee (%)d confign
entry ketone conc. (M)
ee (%)c confign
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
23a
23a
23a
23b
23b
23b
23c
23c
23c
23d
23d
23d
23e
23e
23e
23f
23f
23f
25
12b
18
20
12b
18
20
12b
18
20
12b
18
20
12b
18
20
12b
18
20
12b
18
20
HCOOHa
HCOOHa
iPrOHb
99
>99
82
82
92
95
>99
97
99
95
80
52
>99
>99
99
>99
>99
97
65
65
80
73
77
92
52e
52e
77e
65
65
88
79
79
97
70
59
87
55
55
37
S
S
R
S
S
R
S
S
R
S
S
R
S
S
R
S
S
R
S
S
R
1
2
3
4
5
23a
23a
23a
23e
23e
0.1
0.2
0.3
0.1
0.2
82 (1)
80d
82d
81d
97
R
R
R
R
R
92 (0.5)
94 (0.5)
99 (1)
HCOOHa
HCOOHa
iPrOHb
99 (1)
96
a
The reaction was performed at 20 °C using 4 mmol of substrate
HCOOHa
HCOOHa
iPrOHb
in a 2-propanol solution. Substrate:[IrCl(COD)]2:ligand:tBuOK )
b
400:1:5:12.5. Conversions were determined by GLC and/or 300-
MHz 1H NMR analysis. c Determined using chiral HPLC analysis
HCOOHa
HCOOHa
iPrOHb
d
with a Chiracel OD column, unless otherwise specified. Deter-
mined by capillary GLC analysis using a chiral cycloSil-B column.
HCOOHa
HCOOHa
iPrOHb
(I)-catalyzed transfer hydrogenation of prochiral ketones.
Both the sulfoxide-containing â-amino alcohol (12b) and
the aminosulfides derived from 1,2-disubstituted amino
alcohols (16-21) gave rise to high reaction rates and
moderate to excellent enantioselectivities in the reduction
of various ketones. The enantioselective outcome of the
reaction was directed by selecting the appropriate ligand
and hydrogen donor. Depending on the ligand structure
and stoichiometry, the substrate, the reaction tempera-
ture, and the hydrogen donor, enantioselectivities of up
to 97% could be reached. The best results were obtained
with the aminosulfide ligand (1S,2R)-2-amino-1,2-diphen-
yl-1-benzylthioethane (20), which turned out to be a good
ligand for the control of asymmetric iridium-catalyzed
transfer hydrogenation of prochiral ketones.
HCOOHa
HCOOHa
iPrOHb
HCOOHa
HCOOHa
iPrOHb
98
92
39
25
25
a
The reaction was performed at 60 °C using 4 mmol of substrate
in 3 mL of a formic acid/triethylamine (5:2) solution. Substrate:
[IrCl(COD)]2:ligand ) 400:1:5. The reaction was performed at 20
°C using 4 mmol of substrate in a 0.1 M 2-propanol solution.
Substrate:[IrCl(COD)]2:ligand:tBuOK ) 400:1:5:12.5. c Conver-
sions were determined by GLC and/or 300-MHz 1H NMR analysis.
b
d
Determined using chiral HPLC analysis with a Chiracel OD
column, unless otherwise specified. e Determined by capillary GLC
analysis using a chiral cycloSil-B column.
restricted by the fact that the reaction must be performed
using a substrate concentration as low as 0.1 M, because
of the structural similarities of the product alcohol and
the H-donor.1 To become economically feasible, the reac-
tion should be performed using higher substrate concen-
trations and consequently less 2-propanol. The effect of
different substrate concentrations was investigated using
ligand 20 in the iridium(I)-catalyzed reduction of ac-
etophenone (23a ) and 1′-acetonaphthone (23e). Table 9
shows that a higher substrate concentration of up to 0.3
M does not significantly affect the enantioselective
outcome of the reaction.
Ack n ow led gm en t. Financial support from the In-
novation Oriented Research Program (IOP-Katalyse) is
gratefully acknowledged. Wim G. J . de Lange is ac-
knowledged for technical assistance.
Su p p or tin g In for m a tion Ava ila ble: Experimental de-
tails; characterization data for compounds 6e-22b; IR, 1H
NMR, 13C NMR, HRMS, and elemental analysis data; GC and
HPLC conditions for the separation of the enantiomers of 24a -
26, as well as crystal structure refinement data for compound
12a , including atomic coordinates, isotropic and anisotropic
displacement parameters, and a listing of bond angles and
bond lengths. This material is available free of charge via the
Internet at http://pubs.acs.org.
Con clu d in g Rem a r k s
Two series of N,S-chelates were synthesized that were
very effective as a new class of ligands for the iridium-
J O991700T