Optically active iridium imidazol-2-ylidene-oxazoline complexes: Preparation and use in asymmetric hydrogenation of arylalkenes

Article abstract of DOI:10.1021/ja028142b

This work explores the potential of iridium complexes of the N-heterocyclic carbene oxazoline ligands 1 in asymmetric hydrogenations of arylalkenes. The accessible carbene precursors, imidazolium salts 2, and robust iridium complexes 5 facilitated a disco

Full text of DOI:10.1021/ja028142b

Published on Web 12/05/2002
Optically Active Iridium Imidazol-2-ylidene-oxazoline
Complexes: Preparation and Use in Asymmetric
Hydrogenation of Arylalkenes
Marc C. Perry, Xiuhua Cui, Mark T. Powell, Duen-Ren Hou,
Joseph H. Reibenspies, and Kevin Burgess*
Contribution from the Chemistry Department, Texas A & M UniVersity, P.O. Box 30012,
College Station, Texas 77842
Received August 14, 2002. Revised Manuscript Received September 27, 2002
Abstract: This work explores the potential of iridium complexes of the N-heterocyclic carbene oxazoline
ligands 1 in asymmetric hydrogenations of arylalkenes. The accessible carbene precursors, imidazolium
salts 2, and robust iridium complexes 5 facilitated a discovery/optimization approach that featured preparation
of a small library of iridium complexes, parallel hydrogenation reactions, and automated analysis. Three of
the complexes (5ab, 5ad, and 5dp) and a similar rhodium complex (6ap) were studied by single-crystal
X-ray diffraction techniques. This revealed molecular features of 6ap, and presumably the corresponding
iridium complex 5ap, that the others do not have. In enantioselective hydrogenations of arylalkenes complex
5ap was the best for many, but not all, substrates. The enantioselectivities and conversions observed
were sensitive to minor changes to the catalyst and substrate structure. Ligands with aliphatic N-heterocyclic
carbene substituents gave complexes that are inactive, and do not lose the 1,5-cyclooctadiene ligands
under the hydrogenation conditions. Experiments to investigate this unexpected observation imply that it
is of a steric, rather than an electronic, origin. Temperature and pressure effects on the conversions and
enantioselectivities of these reactions had minimal effects for some alkenes, but profound effects for others.
In one case, the enantioselectivities obtained at high-pressure/low-temperature conditions were opposite
to those obtained under high-temperature/low-pressure conditions (-64% enantiomeric excess versus +89%
enantiomeric excess); a transformation from one prevalent mechanism to another is inferred from this.
The studies of pressure dependence revealed that many reactions proceeded with high conversions, and
optimal enantioselectivities in approximately 2 h when only 1 bar of hydrogen was used. Deuterium-labeling
experiments provide evidence for other types of competing mechanisms that lead to D-incorporation at
positions that do not correspond to direct addition to the double bond.
Electron-rich carbenes have the potential to become the most
interesting transition-metal ligands to emerge from the 1990s.
Several factors have converged to bring this about. First,
synthetic methods to prepare complexes of these ligands have
improved. Early synthetic routes based on insertion of metals
into the carbon-carbon double bonds of highly electron-rich
alkenes1-4 were inconvenient because the electron-rich alkenes
are sensitive and not particularly accessible compounds. The
discovery of carbenes that are stable at ambient temperatures5-9
rapidly led to investigations of isolated carbenes with transition-
metal compounds,10 and interest in that area has expanded
since.9,11-13 In some cases, salts of small heterocycles can be
converted directly to transition-metal carbene complexes provid-
ing an even more facile route to these materials.14-16
A second reason for the growing interest in transition-metal
complexes of electron-rich carbenes is that they have several
characteristics that make them valuable in catalysis. They tend
to be air-stable materials in which the carbene ligands bind to
the metal more strongly than electron-rich phosphines.17 Their
powerful σ-donating and weak π-accepting properties give metal
* Corresponding author. E-mail: burgess@tamu.edu.
(1) Hitchcock, P. B.; Lappert, M. F.; Terrenos, P.; Wainwright, K. P. J. Chem.
Soc., Chem. Commun. 1980, 1180-1181.
(9) Herrmann, W. A.; Ko¨cher, C. Angew. Chem., Int. Ed. Engl. 1997, 36, 2162-
2187.
(2) Hitchcock, P. B.; Lappert, M. F.; Terrenos, P. J. Organomet. Chem. 1982,
239, C26-C30.
(10) Arduengo, A. J., III; Gamper, S. F.; Calabrese, J. C.; Davidson, F. J. Am.
Chem. Soc. 1994, 116, 4391-4394.
(3) Coleman, A. W.; Hitchcock, P. B.; Lappert, M. F.; Maskell, R. K.; Mu¨ller,
J. H. J. Organomet. Chem. 1983, 250, C9-C14.
(4) Doyle, M. J.; Lappert, M. F.; Pye, P. L.; Terrenos, P. J. Chem. Soc., Dalton
Trans. 1984, 2355-2364.
(11) Herrmann, W. A.; Elison, M.; Fischer, J.; Ko¨cher, C.; Artus, G. R. J. Chem.
Eur. J. 1996, 2, 772-780.
(12) Weskamp, T.; Bo¨hm, V. P. W.; Herrmann, W. A. J. Organomet. Chem.
2000, 600, 12-22.
(5) Arduengo, A. J., III; Harlow, R. L.; Kline, M. J. Am. Chem. Soc. 1991,
113, 361-363.
(13) Herrmann, W. A. Angew. Chem., Int. Ed. 2002, 41, 1290-1309.
(14) O¨ fele, J. J. Organomet. Chem. 1968, 12, 42-43.
(15) Wanzlick, V. H.-W.; Scho¨nherr, H.-J. Angew. Chem. 1968, 80, 154.
(16) Herrmann, W. A.; Elison, M.; Fischer, J.; Kocher, C.; Artus, G. R. Angew.
Chem., Int. Ed. Engl. 1995, 34, 2371-2374.
(6) Arduengo, A. J., III; Dias, H. V. R.; Harlow, R. L.; Kline, M. J. Am. Chem.
Soc. 1992, 114, 5530-5534.
(7) Arduengo, A. J.; Krafczyk, R.; Schmutzler, R. Tetrahedron 1999, 55,
14523-14534.
(17) Huang, J.; Schanz, H.-J.; Stevens, E. D.; Nolan, S. P. Organometallics
1999, 18, 2370-2375.
(8) Regitz, M. Angew. Chem., Int. Ed. Engl. 1996, 35, 725-728.
9
10.1021/ja028142b CCC: $25.00 © 2003 American Chemical Society
J. AM. CHEM. SOC. 2003, 125, 113-123
113

A R T I C L E S
Perry et al.
centers that are electron-rich relative to the corresponding
phosphine complexes.18 Consequently, the carbene complexes
tend to be highly active in oxidative addition reactions, but that
activity can be tempered by sterically shielding the metal with
umbrella-like aromatic substituents. For these reasons, com-
plexes of electron-rich carbenes are proving to be robust and
active catalysts for several different reactions19-28 especially
in alkene metathesis.29
Chiral modification of electron-rich carbene ligands is a
logical progression in the field. Chiral, optically active, imid-
azolinylidines (alternatively called dihydroimidazolylidines)
were first reported several decades ago3 and several others have
been reported more recently.30-32 Chiral, unsaturated analogues
of the imidazolinylidine ligands, that is, chiral imidazolylidines,
have also been reported. Most of these are monodentate systems
with chiral N-substituents,33-35 or metalated derivatives36 that
might be expected to behave as a monodentate carbene complex
in many catalytic reactions. A binaphthyl system37 and the
oxazoline-containing chelate38 are the only robust, chiral,
bidentate, unsaturated carbene complex types reported to date.
Figure 1. The concept of “reactions of libraries with libraries” applied to
the carbene-oxazoline ligands 1.
Our primary interest in catalysis is to demonstrate high-
throughput methods can be used to accelerate catalyst discovery
and optimization.41-45 We believe that the slow step in such
approaches is preparation of the catalyst library, rather than
screening, so any advance that enables faster production of
ligands with good molecular architectures for asymmetric
catalysis is significant. Consequently, our thoughts settled on
the new ligand design 1 (Figure 1). These are constrained,
bidentate systems that seemed suitable for asymmetric catalysis.
Moreover, they could be accessible via the imidazolium salts 2
obtained from the oxazoline electrophiles 3 and the imidazoles
4; a relatively large library of ligand precursors 2 therefore could
be generated from two very small libraries of synthetic
constituents. This is a case in which reactions of libraries with
libraries46-49 could allow the production of a much bigger
collection of compounds. For instance, if 20 imidazoles 4 were
reacted with 5 oxazolines 3, then 100 different ligand precursors
could be prepared. Practically, that approach would be facilitated
by the characteristics of the products 2; these salts might be
purified simply by precipitation from apolar solvents, and, unlike
phosphines, the imidazolium salts 2 are easily handled, robust,
air-stable materials.
Asymmetric hydrogenations of alkenes that possess little or
no coordinating functionality were identified as a worthy test
application of the target ligand set. Substrate types suitable for
asymmetric hydrogenations include trisubstituted and tetra-
substituted alkenes (Figure 2). If these substrates have large
trans-orientated substituents, then desirable types of asymmetric
catalyst topographies are easier to envisage (ligands that fill
diagonally situated quadrants of space, e.g., C2-symmetric
ligands). However, there simply are not many homogeneous
Despite these efforts to prepare optically active, electron-
rich carbene complexes, there have been very few successful
applications of these complexes in asymmetric catalysis. Before
the work described here was submitted for communication,39
the only asymmetric catalysis featuring N-heterocyclic carbene
ligands involved an unspecified Heck reaction that gave
enantiomeric excesses (ee’s) of no more than 8%,33 and
hydrosilylation of an arylmethyl ketone with enantiomer ex-
cesses of 32% or less.34 Since then, such ligands have been
applied in an intramolecular cyclization reaction giving a product
of 76% ee,31 and in symmetry-breaking metathesis reactions
that give excellent ee’s in a limited number of reactions.32,40
(18) Huang, J.; Stevens, E. D.; Nolan, S. P.; Petersen, J. L. J. Am. Chem. Soc.
1999, 121, 2674-2678.
(19) Huang, J.; Nolan, S. P. J. Am. Chem. Soc. 1999, 121, 9889-9890.
(20) Huang, J.; Grasa, G.; Nolan, S. P. Org. Lett. 1999, 1, 1307-1309.
(21) Baratta, W.; Herrmann, W. A.; Rigo, P.; Schwarz, J. J. Organomet. Chem.
2000, 593-4, 489-493.
(22) Bo¨hm, V. P. W.; Weskamp, T.; Gsto¨ttmayr, C. W. K.; Hermann, W. A.
Angew. Chem., Int. Ed. 2000, 39, 1602-1604.
(23) Xu, L.; Chen, W.; Xiao, J. Organometallics 2000, 19, 1123-1127.
(24) Baratta, W.; Herrmann, W. A.; Kratzer, R. M.; Rigo, P. Organometallics
2000, 19, 3664-3669.
(25) Bo¨hm, V. P. W.; Gsto¨ttmayr, C. W. K.; Weskamp, T.; Herrmann, W. A.
J. Organomet. Chem. 2000, 595, 186-190.
(26) Weskamp, T.; Bo¨hm, V. P. W.; Herrmann, W. A. J. Organomet. Chem.
1999, 585, 348-352.
(27) Herrmann, W. A.; Reisinger, C.-P.; Spiegler, M. J. Organomet. Chem. 1998,
557, 93-96.
(28) Chen, C. A.; Ren, L.; Decken, A.; Crudden, C. M. Organometallics 2000,
19, 3459-3461.
(29) Trnka, T. M.; Grubbs, R. H. Acc. Chem. Res. 2001, 34, 18-29.
(30) Scholl, M.; Ding, S.; Lee, C. W.; Grubbs, R. H. Org. Lett. 1999, 1, 953-
956.
(31) Lee, S.; Hartwig, J. F. J. Org. Chem. 2001, 66, 3402-3415.
(32) Van Veldhuizen, J. J.; Garber, S. B.; Kingsbury, J. S.; Hoveyda, A. H. J.
Am. Chem Soc. 2002, 124, 4954-4955.
(33) Enders, D.; Gielen, H.; Raabe, G.; Runsink, J.; Teles, J. H. Chem. Ber.
1996, 129, 1483-1488.
(41) Burgess, K.; Lim, H.-J.; Porte, A. M.; Sulikowski, G. A. Angew. Chem.,
Int. Ed. Engl. 1996, 35, 220-222.
(34) Herrmann, W. A.; Goossen, L. J.; Kocher, C.; Artus, G. R. J. Angew. Chem.,
Int. Ed. Engl. 1996, 35, 2805-2807.
(42) Burgess, K.; Moye-Sherman, D.; Porte, A. M. In Molecular DiVersity and
Combinatorial Chemistry; Chaiken, I. M., Janda, K. D., Eds.; American
Chemical Society: Washington, DC, 1996; pp 128-136.
(43) Burgess, K.; Porte, A. M. In AdVances in Catalytic Processes; Doyle, M.
P., Ed.; JAI Press: Greenwich, CT, 1997; Vol. 2, pp 69-82.
(44) Porte, A. M.; Reibenspies, J.; Burgess, K. J. Am. Chem. Soc. 1998, 120,
9180-9187.
(35) Herrmann, W. A.; Goossen, L. J.; Artus, G. R. J.; Kocher, C. Organo-
metallics 1997, 16, 2472-2477.
(36) Enders, D.; Gielen, H.; Raabe, G.; Runsink, J.; Teles, J. H. Chem. Ber.
1997, 130, 1253-1260.
(37) Clyne, D. S.; Jin, J.; Genest, E.; Galluci, J. C.; RajanBabu, T. V. Org.
Lett. 2000, 2, 1125-1128.
(45) Burgess, K.; Porte, A. M. Tetrahedron: Asymmetry 1998, 9, 2465-2469.
(46) Nielsen, J.; Jensen, F. R. Tetrahedron Lett. 1997, 38, 2011-2014.
(47) Baldino, C. M.; Casebier, D. S.; Caserta, J.; Slobodkin, G.; Tu, C.; Coffen,
D. L. Synlett 1997, 488-490.
(38) Herrmann, W. A.; Goossen, L. J.; Spiegler, M. Organometallics 1998, 17,
2162-2168.
(39) Powell, M. T.; Hou, D.-R.; Perry, M. C.; Cui, X.; Burgess, K. J. Am. Chem.
Soc. 2001, 123, 8878-8879.
(48) Neustadt, B. R.; Smith, E. M.; Nechuta, T.; Zhang, Y. Tetrahedron Lett.
1998, 39, 5317-5320.
(40) Seiders, T. J.; Ward, D. W.; Grubbs, R. H. Org. Lett. 2001, 3, 3225-
3228.
(49) Boger, D. L.; Chai, W.; Jin, Q. J. Am. Chem. Soc. 1998, 120, 7220-7225.
9
114 J. AM. CHEM. SOC. VOL. 125, NO. 1, 2003

Carbene Complexes for Asymmetric Hydrogenation
A R T I C L E S
Scheme 1. Typical Synthesis of the Chirons 3
Figure 2. Types of prochiral alkenes.
Chart 1. Asymmetric Variants of Crabtree’s Catalyst
organometallic compounds known that will hydrogenate these
more hindered substrates.
One class of homogeneous organometallic compounds that
will hydrogenate tri- and tetrasubstituted alkenes are the cationic
iridium complexes of the type [(1,5-cyclooctadiene; COD)Ir-
(py)(PR3)]+ including Crabtree’s catalyst (Chart 1; R ) cyclo-
hexyl).50 Pfaltz and co-workers recognized chiral phosphine-
oxazoline ligands are similar to cis-disposed, monodentate,
pyridine/phosphine ligand combinations.51-59 Subsequently, they
prepared several different types of P/N-ligands, complexed them
with the “(COD)Ir” fragment, tested them as hydrogenation
catalysts, and obtained some excellent enantioselectivities. The
groups of Buchwald60,61 and of Marks62 have used titanocene,
zirconocene, and cyclopentadienyl lanthanide complexes as
hydrogenation catalysts for the same types of substrates.
Excellent enantioselectivities have been obtained, but the
catalysts are more difficult to prepare (highly air-sensitive) and
harsher, less experimentally convenient conditions are required.
Overall, some very important developments have occurred in
the area of catalytic hydrogenations of aryl-substituted and
unfunctionalized alkenes, but many challenges still remain.
These include increased enantioselectivities for a broader range
of substrates, via reactions at lower H2 pressures, and involving
more accessible catalysts.
This work investigates similarities between Crabtree’s cata-
lyst, Pfaltz’s phosphine-oxazoline complexes, some phosphine-
oxazoline complexes prepared in these laboratories (JM Phos
ligands),63,64 and the iridium carbene-oxazoline complexes 5.
For this study, generation of a small library of complexes was
ideal for two reasons. First, the ligands could be made easily,
as outlined above. Second, iridium complexes of this type tend
to be air-stable, and easily chromatographed if necessary, so a
library of imidazolium-oxazolines 2 could potentially be
transformed into the corresponding iridium complexes 5 without
excessive effort.
Results and Discussion
1. Preparation of the Oxazoline Electrophiles 3. Three
syntheses of these chirons were developed in studies focused
on the preparation of JM Phos ligands.63,64 They each have
different advantages and drawbacks regarding efficiency, scale-
up, and scope for divergence. One of the most efficient is shown
in Scheme 1, for oxazoline 3a. This route and the other syntheses
were used to prepare the iodides 3a-3f.
2. Preparation of the Imidazoles 4. Scheme 2 summarizes
the reported methods that were used to obtain the imidazoles
4a-4p. The first synthesis (Scheme 2a) is a direct 1-pot
condensation that works well for aliphatic amines.65 Only a
limited number of aromatic amines react well in these reactions;
(50) Crabtree, R. Acc. Chem. Res. 1979, 12, 331-337.
(51) Schnider, P.; Koch, G.; Pretot, R.; Wang, G.; Bohnen, F. M.; Kruger, C.;
Pfaltz, A. Chem. Eur. J. 1997, 3, 887-892.
(52) Lightfoot, A.; Schnider, P.; Pfaltz, A. Angew. Chem., Int. Ed. 1998, 37,
2897-2899.
(53) Hilgraf, R.; Pfaltz, A. Synlett 1999, 1814-1816.
(54) Blackmond, D. G.; Lightfoot, A.; Pfaltz, A.; Rosner, T.; Schnider, P.;
Zimmermann, N. Chirality 2000, 12, 442-449.
(55) Escher, I. H.; Pfaltz, A. Tetrahedron 2000, 56, 2879-2888.
(56) Helmchen, G.; Pfaltz, A. Acc. Chem. Res. 2000, 33, 336-345.
(57) Blankenstein, J.; Pfaltz, A. Angew. Chem., Int. Ed. 2001, 40, 4445-4447.
(58) Cozzi, P. G.; Zimmermann, N.; Hilgraf, R.; Schaffner, S.; Pfaltz, A. AdV.
Synth. Catal. 2001, 343, 450-454.
(59) Menges, F.; Pfaltz, A. AdV. Synth. Catal. 2002, 344, 40-44.
(60) Broene, R. D.; Buchwald, S. L. J. Am. Chem. Soc. 1993, 115, 12569-
12570.
(61) Troutman, M. V.; Appella, D. H.; Buchwald, S. L. J. Am. Chem. Soc. 1999,
121, 4916-4917.
(62) Conticello, V. P.; Brard, L.; Giardello, M. A.; Tsuji, Y.; Sabat, M.; Stern,
C. L.; Marks, T. J. J. Am. Chem. Soc. 1992, 114, 2761-2762.
(63) Hou, D.-R.; Burgess, K. Org. Lett. 1999, 1, 1745-1747.
(64) Hou, D.-R.; Reibenspies, J.; Burgess, K. J. Org. Chem. 2001, 66, 206-
215.
(65) Gridnev, A. A.; Mihaltseva, I. M. Synth. Commun. 1994, 24, 1547-1555.
9
J. AM. CHEM. SOC. VOL. 125, NO. 1, 2003 115

A R T I C L E S
Perry et al.
Scheme 2. Synthesis of the Imidazoles: (a) Direct Condensation
to Give 4a-j, (b) N-Arylation Affording 4k-m, and (c) a Ring
Closure-Oxidation-Reduction Sequence Furnishing 4n-q
Table 1. Carbene-Oxazoline Complexes Prepared
1
2
1
2
5
R
R
5
R
R
ab 1-Ad tBu
ac 1-Ad CHPh2
ad 1-Ad Cy
aq 1-Ad
ba tBu
bb tBu
bc tBu
bp tBu
2,5-tBu2C6H3
1-Ad
tBu
af
1-Ad 2,4,6-Me3C6H2
CHPh2
an 1-Ad 3,5- tBu2-4-MeOC6H2
ao 1-Ad 2,5-Et2C6H3
2,6-iPr2C6H3
cp CHPh2 2,6-iPr2C6H3
ap 1-Ad 2,6-iPr2C6H3
dp Ph
2,6-iPr2C6H3
always required. The last route that was used (Scheme 2c)
involves multiple steps and is only applicable to aromatic
amines.67 This approach works well for scale-up purposes
because no chromatographic separations are required.
3. Preparation of the Imidazolium Salts 2. The imidazolium
salts 2 were prepared via SN2 displacements of iodide from the
electrophiles 3 by the imidazoles 4. At the end of each reaction,
the solvent was removed, and the residue was washed with
diethyl ether, then dried under vacuum; the products were pure
enough to use in the complex formation steps.
4. Preparation of the Iridium Complexes 5. Table 1 shows
how the complexes 5 were prepared by using 1.5 equiv of tert-
butoxide to deprotonate the imidazolium salt to give the
corresponding carbene. In general, the complexes were air-stable
and easily purified via flash chromatography. In some cases,
especially where the substituents involved were bulky, the
reaction was not clean, and the desired complex was not isolated
(e.g., 5ep where R1 ) 3,5-tBu2C6H3; R2 ) 2,6-iPr2C6H3).
However, in most cases where a particular R1 and R2 combina-
tion does not appear in Table 1, we simply did not attempt to
make it. This is because the screening process described below
began as soon as the first batches of complexes were prepared,
and continued as more were made, so we were able to focus
our library of complexes in response to the data emerging from
the screens.
5. Crystallographic Analyses of Illustrative Complexes.
Three iridium complexes, 5ab, 5ad, 5dp, and the rhodium
complex 6ap were analyzed via single-crystal X-ray diffraction
(Supporting Information). The latter was made because we were
unable to obtain X-ray crystals of the complex that later emerged
as the most enantioselective in many reactions, 5ap. The
coordination spheres of the Ir atoms in 5ab and 5ad and of the
two such cases are shown (compounds 4f and 4h), but several
that were tried did not work (e.g., R2 ) 3,5-Me2C6H3, and
1-Nap). Scheme 2b shows another one-step route to the
compounds. That approach works well for unhindered aromatic
amines,66 but scale-up is inconvenient because 1 equiv of 1,10-
phenanthroline is used, and this basic heterocycle can be difficult
to separate from the imidazole product; in fact, a column is
(66) Kiyomori, A.; Marcoux, J.-F.; Buchwald, S. L. Tetrahedron Lett. 1999,
40, 2657-2660.
(67) Johnson, A. L.; Kauer, J. C.; Hsarma, D. C.; Dorfman, R. I. J. Med. Chem.
1969, 12, 1024-1028.
9
116 J. AM. CHEM. SOC. VOL. 125, NO. 1, 2003

Carbene Complexes for Asymmetric Hydrogenation
A R T I C L E S
Figure 3. Comparison of structures 5ab, 5ad, and 6ap, side-on with COD
ligand omitted (BARF- counterion not shown).
Rh atom in 6ap can be described as pseudo-square planar. The
chiral ligands in the complexes exist as U-shaped entities in
which the imidazolylidine and oxazoline substituents are
projected above the metal square plane (Figure 3). For com-
plexes 5ab and 5ad the imidazolylidine substituent, i.e., the tert-
butyl group of 5ab and the cyclohexyl of 5ad, only fill space
above the metal. The corresponding group for 6ap, 2,6-iPr2C6H3,
is larger, it interacts more with the oxazoline substituent (1-
Ad), and projects into an area in front of, below, and to the left
of the Rh-square plane (when viewed in the orientation indicated
in Figure 3). This fundamental difference between the complex
with the 2,6-iPr2C6H3-substituent and the rest, may, by inference,
provide an insight into the unique reactivity of corresponding
iridium complex 5ap (vide infra).
6. Enantioselective Hydrogenations with Complexes 5.
(a) Effects Of Catalyst Structure. Figure 4a summarizes yield
and enantioselectivity data that were obtained in the early
screens with some representative complexes 5 in which the
N-heterocyclic carbene ligand substituent (R2) was always 2,6-
di-iso-propylphenyl, but the oxazoline substituent R1 was varied.
Both the yield and enantioselectivity of the catalyst were highly
dependent on R1. Unsatisfactory chemical and optical yields
Figure 4. Yield and enantioselectivity data for asymmetric hydrogenations
(a) using catalysts with four different oxazoline substituents; and (b) using
t
catalyst 5bp and 5ap to compare R1 ) Bu and 1-Ad for three different
alkenes.
9
J. AM. CHEM. SOC. VOL. 125, NO. 1, 2003 117

A R T I C L E S
Perry et al.
were obtained for R1 ) Ph, tBu, and CHPh2, but both the results
were excellent for R1 ) 1-Ad (i.e., 1-adamantyl, complex 5ap).
The contrast between the catalysts where R1 is tert-butyl (5bp)
and 1-adamantyl (5ap) in these initial screens was remarkable,
so further experiments were undertaken to probe if the trends
observed also apply to asymmetric hydrogenations of other
substrates (Figure 4b). In direct comparisons for three different
alkenes, the 1-admantyl complex 5ap gave superior chemical
and optical yields to the tert-butyl complex 5bp. These
observations show slight topographical differences in the ligand
that can dramatically affect the performance of the catalyst.
The imidazolylidine substituent (R2) can also have dramatic
effects on the results obtained for a given substrate. Figure 5
shows data that illustrate this. Catalysts 5ao and 5ap are
structurally similar, differing only slightly in the N-heterocyclic
carbene substituent, but 5ao is marginally more effective for
one of the substrates shown and 5ap gives better data for the
other.
A surprising observation was made when the screen was
expanded to include other variations of the heterocyclic carbene
portion (R2) while retaining the 1-adamantyl R1 substituent.
Hydrogenation of the 4-methoxyphenyl-substituted alkene shown
in reaction 1 with catalyst 5ap (R2 ) 2,6-iPr2C6H4) gave good
conversion and high ee as expected. However, a similar catalyst
5ad that has R2 ) cyclohexyl (Cy) gave almost no detectable
conversion. Moreover, in the latter case, the catalyst was
recovered unchanged at the end of the reaction; the COD ligand
had not been removed. This was confirmed in an experiment
in which the two complexes were maintained under 1 atm of
H2 in CDCl3 for 3 h; catalyst 5ap was transformed into a mixture
of products whereas 5ad was unchanged. Lack of reactivity of
1,4-cyclooctadiene complexes is not surprising because this type
of phenomenon has been reported several times before in the
literature,68,69 but the contrast between catalysts 5ap and 5ad
is remarkable. Other experiments (data not shown) indicated
complexes 5ab and 5ad were unreactive as catalysts in similar
reactions; both have aliphatic R2 substituents on the N-
heterocyclic carbene.
Figure 5. A comparison to illustrate the effects of minor changes to the
imidazolylidine substituent R2, that is, for 5ao (R2 ) 2,6-Et2C6H3) and 5ap
(R2 ) 2,6-iPr2C6H3).
the N-heterocyclic carbene substituent. There was, however, no
significant difference in the metal carbonyl IR stretches for the
two complexes. Apparently, the reactivity differences observed
in reaction 1 most probably have steric origins.
(b) Some Effects of Changing the Alkene Substrate. As a
result of the data described above and other studies (Supporting
Information), the subsequent work focused on complex 5ap (R1
) 1-Ad, R2 ) 2,6-iPr2C6H3). Figure 6 shows chemical and
optical yields for hydrogenations of six alkenes having methyl,
ethyl, and iso-propyl substituents in cis- and trans-orientations
Two dicarbonyl derivatives 7 and 8 were made to investigate
possible electronic differences at the iridium center caused by
(68) Drexler, H.-J.; Baumann, W.; Spannenberg, A.; Fischer, C.; Heller, D. J.
Organomet. Chem. 2001, 621, 89-102.
(69) Borner, A.; Heller, D. Tetrahedron Lett. 2001, 42, 223-225.
9
118 J. AM. CHEM. SOC. VOL. 125, NO. 1, 2003

Carbene Complexes for Asymmetric Hydrogenation
A R T I C L E S
Figure 6. Yield and enantioselectivity data for hydrogenation of 2-(4-
methoxyphenyl)-alkenes (Ar ) 4-MeOC6H4) using catalyst 5ap.
relative to the aryl substituent. In general, E-alkenes produced
better results than their cis-isomers.
(c) Temperature and Pressure Effects. The next step in
this study was to investigate temperature and pressure effects.
Figure 7a shows the effects of temperature in the hydrogenations
of three isomeric cis-, trans-, and 1,1-disubstituted alkenes. The
cis- and trans-alkenes show almost no temperature dependence,
but there is a remarkable enantioselectivity variation for the 1,1-
disubstituted alkene. These experiments show that the variation
of enantioselectivities with temperature is substrate dependent,
and that it is impossible to generalize on the effects of
temperature for a range of substrate classes.
Figure 7b shows the pressure effects on the hydrogenation
of six different alkenes. Three of these substrates show no
significant yield or enantioselectivity variations with pressure.
For both isomers of 2-(4′-methoxyphenyl)-4-methylpent-2-ene,
the enantioselectivities of the hydrogenation reactions varied
with pressure, and ee/pressure correlations were observed for
the 1,1-disubstituted alkene (the same substrate that showed
interesting ee/temperature dependence in Figure 7a). Because
enantioselective hydrogenations of 1,1-disubstituted alkenes are
generally difficult, a set of experiments were undertaken to
optimize the enantioselectivity of hydrogenation of this substrate
in faVor of either enantiomer. Two aspects of the data obtained
in those experiments (Figure 8) are notable. First, the enantio-
selectivity, and presumably the prevailing mechanism, switches
abruptly on varying the conditions between -15 °C, 85 bar and
25 °C, 1 bar. High concentrations of hydrogen in solution (low
temperature and high H2 pressure) favor formation of the (S)-
enantiomer, whereas the (R)-antipode is preferred when the
concentration of hydrogen in solution is low. Second, the
Figure 7. (a) Temperature effects in hydrogenations of three isomeric
alkenes and (b) pressure effects in hydrogenations of five alkenes
(throughout Ar ) MeOC6H4).
9
J. AM. CHEM. SOC. VOL. 125, NO. 1, 2003 119

A R T I C L E S
Perry et al.
intermediate that precedes the enantioface-determining step in
the process, particularly under low [H2] conditions. Crabtree
has discussed the possibility that two alkenes are coordinated
in hydrogenations involving his catalysts.72
Most of the prior work on enantioselective, iridium-mediated
hydrogenations has featured reactions performed at 50 bar, and
relatively little emphasis was placed on the effects of using lesser
pressures. The data shown in Figure 7 show that the reactions
Figure 8. Optimization of enantioselectivity of hydrogenation of one
substrate based on the temperature and pressure effects outlined in Figure
7, parts a and b. Throughout Ar ) 4-MeOC6H4.
variation of enantioselectivity observed is much greater than
Pfaltz and co-workers reported for the same substrate using their
catalysts (e.g., 62% ee at 50 bar and 85% ee at 1 bar).59
Pronounced variation of enantioselectivities with hydrogen
concentration in solution is indicative of two (or even more)
different mechanisms that happen to give opposite enantiomers;
one pathway predominates at low [H2] (high temperature and
low pressure) and the other dominates at high [H2]. The variation
of enantioselectivities between the extremes is surprising,
especially because several other substrates do not exhibit this
behavior.
The enantioselectivity changes noted above imply the prevail-
ing mechanism varies with the amount of hydrogen in solution,
that is, [H2] impacts the first irreversible step that sets the
enantioface selection. Oxidative addition of hydrogen across
iridium tends to occur at lower temperatures than similar
transformations for rhodium complexes.50,70,71 However, it is
conceivable that at higher temperatures and low H2 pressures
oxidative addition of H2 is slow enough to directly impact the
enantiofacial selectivity of the reaction, whereas at high [H2]
another step in the mechanism is critical. Another explanation
for the temperature and pressure effects is that the types of
intermediates that might be involved are P and Q below, that
is, two alkenes might be coordinated to the metal in the
9
120 J. AM. CHEM. SOC. VOL. 125, NO. 1, 2003

Carbene Complexes for Asymmetric Hydrogenation
A R T I C L E S
Scheme 3. A Possible Mechanism To Account for “Unexpected”
Deuterium Incorporation Illustrated for Isomeric Z- and
1,1-Disubstituted Alkenes That Do Give Anomalous Incorporation,
and an E-alkene That Does Not
Scheme 3 shows a possible mechanism for the incorporation
of deuterium at unexpected positions as indicated. It is based
on similar mechanisms proposed by Brown and co-workers for
similar reactions of Crabtree’s catalyst.73 This is not the only
possibility, but formation of allyl intermediates neatly explains
why deuteration of the isomeric alkenes shown in Scheme 3a,b
gives deuterium incorporation at anomalous positions. We
propose that the third isomeric alkene shown in Scheme 3c does
not give anomalous incorporation because a more congested
allyl intermediate would be involved.
Comparison of the data from reactions 3-6 shows that the
carbene complex 5ap gives less of these competing processes
than the phosphine oxazoline 9, at least in the reactions explored
to date. This is significant because the enantioselectivies are
easier to control for reactions that have fewer competing
mechanistic pathways.
Reactions 7-10 explore a different issue related to the two
substrates that were previously found to undergo profound
variations of enantioselectivities with hydrogen pressure in the
hydrogenation process, probably because of a change in the
predominant mechanism of the hydrogenation process on
varying the conditions between high H2 pressure/low temper-
ature and low H2 pressure/high temperature. If the competing
pathways in Figures 7 and 8 were the dominant “second
mechanism” then near-complete double-bond migration before
hydrogenation would be expected under some conditions. The
data show that for both substrates this is not so: there is
marginally less deuterium scrambling under high-H2-pressure/
low-temperature conditions (compare reactions 7 and 8, and 9
and 10); hence, there is no evidence for complete double-bond
tend to proceed to completion with optimal ee’s when only one
atmosphere of hydrogen is used. This is important with respect
to experimental convenience and scale-up issues.
(d) Deuterium Labeling Studies. In labeling experiments,
incorporation of deuterium at positions that do not correspond
to direct addition to the double bond are indicative of competing
mechanisms. Reaction 2 illustrates a case where only direct
addition is observed (values quoted relative to the maximum,
set at 1.00, with the results of indirect addition shown in red),
but evidence for indirect incorporation pathways was obtained
for several other substrates. Mass spectra data indicate that when
deuterium scrambling is observed, more than two deuteriums
can be incorporated into the product.
(70) Kimmich, B. F. M.; Somsook, E.; Landis, C. R. J. Am. Chem. Soc. 1998,
120, 10115-10125.
(71) Crabtree, R. H.; Uriarte, R. J. Inorg. Chem. 1983, 22, 4152-4154.
9
J. AM. CHEM. SOC. VOL. 125, NO. 1, 2003 121

A R T I C L E S
Perry et al.
shift before the reduction step, that is, double-bond migration
does not correlate with the mechanistic change inferred from
the temperature/pressure effects shown in Figure 6.
Conclusions
N-Heterocyclic carbene ligands are useful in asymmetric
hydrogenation processes. Their complexes are reasonably stable
and easy to prepare from robust imidazolium salts, which in
turn are conveniently accessible by combining libraries of
oxazoline electrophiles and imidazole nucleophiles. Many of
the reactions performed with the carbene complexes can be run
at room temperature and 1 bar of hydrogen; hence, the reactions
are experimentally convenient, whereas the same reactions using
P-oxazoline-based catalysts generally have been performed
using 50 bar of hydrogen.
When this study was initiated, there was no guiding concept
for how the ligands should be designed. None of the proposed
ligand structures were C2-symmetric, nor do they have PAr2
aryl groups to provide edge-face arrangements about the metal.74
To be effective, ligands 1 had to provide asymmetric environ-
ments for catalysis in other ways. It would have been very
difficult to predict ideal combinations of the ligand R1 and R2
groups required for this, especially because the mechanistic
uncertainties preclude reliable predictions via modeling. Efficient
routes to a small library of catalysts and parallel screening
methods were therefore valuable for this project.
The structural data that have now been obtained (Figure 3)
hint at the desirable molecular characteristics of ligand 1ap
relative to two others (1ab and 1ad). Figure 9a shows this ligand
in complex 6ap. In this graphic, the area of space around the
metal center in complex 6ap is divided into eight regions on
the basis of the C-Rh-N plane, so the Rh-C and Rh-N bonds
make angles of approximately 45° with the x and z axes. The
quadrants of space may then be labeled NW, NE, SW, SE, front
and back. In 6ap the 2,6-iPr2C6H3 imidazolylidine substituent
penetrates into the front NW quadrant, and a little into the front
SW quadrant. Ligand 1ab in the corresponding iridium complex,
where R2 ) tBu, has less steric influence on events in the front
quadrants around the metal (Figure 9b).
This analysis does not imply that all carbene oxazoline ligands
should be modeled on ligand 1ap because, as illustrated in
Figure 4c, optimal results will be obtained by modifying the
ligand to match specific substrates. However, it is likely that
the steric presence of the ligand in the octants of space around
the metal where substrate binds will tend to be critical.
The mechanisms of hydrogenation of alkenes with the
catalysts presented here, those for Pfaltz’ complexes, and even
the achiral systems developed by Crabtree more than 25 year
ago, are unknown. The work described here illustrates factors
that complicate mechanistic work in this area. For instance,
temperature and pressure effects are such that, for some
substrates, but not others, different mechanisms prevail accord-
ing to the conditions. Scrambling of deuterium labels demon-
strate that minor competing pathways are also operative.
Nevertheless, high enantioselectivities were obtained here for
several substrates.
Figure 9. Octants of space around the metal in (a) 6ap and (b) 5ab.
Carbene oxazoline ligands are potentially useful for other
types of asymmetric catalysis, including other hydrogenations,
some oxidative processes, allylations, Heck and related cou-
plings, hydroformylations, and hydrosilylations. Efforts are
underway in these laboratories to explore those reactions.
Approaches based on syntheses of small libraries of catalysts
coupled with parallel screening and automated analyses are
likely to lead to some interesting discoveries. In general, this
strategy aligned with structural and mechanistic work to focus
the library syntheses, is likely to prove the most efficient way
to discover and optimize catalysts for many diverse reaction
types.
(72) Crabtree, R. H.; Demou, P. C.; Eden, D.; Mihelcic, J. M.; Parnell, C. A.;
Quirk, J. M.; Morris, G. E. J. Am. Chem. Soc. 1982, 104, 6994-7001.
(73) Brown, J. M.; Derome, A. E.; Hughes, G. D.; Monaghan, P. K. Aust. J.
Chem. 1992, 45, 143-153.
(74) Steinhagen, H.; Reggelin, M.; Helmchen, G. Angew. Chem., Int. Ed. Engl.
1997, 36, 2108-2110.
9
122 J. AM. CHEM. SOC. VOL. 125, NO. 1, 2003

Carbene Complexes for Asymmetric Hydrogenation
A R T I C L E S
Acknowledgment. Financial support for this work was
provided by The Robert Welch Foundation and Johnson Matthey
plc. We thank Dr. Thomas A. Colacot (JM) and Dr. Shane
Stichy (TAMU/LBMS-Applications Laboratory) for help and
advice, and Jing Liu for preparation of imidazole 4q.
procedures for the preparations of the ligands, complexes, and
hydrogenation/deuteration procedures, discussion of assignments
of absolute configuration and tabulated details of these (PDF).
This material is available free of charge via the Internet at
http://pubs.acs.org.
Supporting Information Available: Comparison of enantio-
selectivities with some relevant literature values, experimental
JA028142B
9
J. AM. CHEM. SOC. VOL. 125, NO. 1, 2003 123