Chiral N-heterocyclic carbene/pyridine ligands for the iridium-catalyzed asymmetric hydrogenation of olefins

Article abstract of DOI:10.1002/anie.201301251

Swapping N,P for C,N: Iridium complexes of bidentate pyridine-based C,N ligands with an N-heterocylic carbene (NHC) unit proved to be efficient and highly enantioselective hydrogenation catalysts. As a result of the lower acidity of iridium hydride intermediates produced from NHC-based complexes, these catalysts are much better suited than analogous N,P-ligand complexes for the hydrogenation of acid-sensitive substrates. Copyright

Full text of DOI:10.1002/anie.201301251

.
Angewandte
Communications
DOI: 10.1002/anie.201301251
Asymmetric Catalysis
Chiral N-Heterocyclic Carbene/Pyridine Ligands for the Iridium-
Catalyzed Asymmetric Hydrogenation of Olefins**
Andreas Schumacher, Maurizio Bernasconi, and Andreas Pfaltz*
The development of chiral iridium catalysts, which were
inspired by the Crabtree catalyst,[1] has considerably
enhanced the application range of the asymmetric hydro-
genation of olefins.[2] In contrast to rhodium and ruthenium
diphosphine complexes, the iridium-based catalysts do not
step.[4a,d] In addition to alkenes bearing only alkyl or aryl
=
substituents in the vicinity of the C C bond, a wide variety of
functionalized olefins with coordinating or noncoordinating
groups, as well as furans and indoles have been successfully
hydrogenated with high enantioselectivity using chiral iridium
N,P-complexes.[2,5]
=
require a coordinating group near the C C bond and,
therefore, enable the hydrogenation of a much wider range
of alkenes.
Burgess and co-workers have intro-
duced a new type of oxazoline-based
C,N ligand containing an N-heterocyclic
carbene (NHC) instead of a phosphine
unit.[6] The iridium C,N-complex
1 (Figure 1) with a 2-adamantyloxazo-
Most iridium catalysts developed so far are cationic
iridium(I) complexes derived from chiral bidentate N,P li-
gands containing an oxazoline or another N heterocycle as
a coordinating unit.[2,3] Typical examples are pyridine phos-
phinite complexes such as the catalyst shown in Scheme 1,
Figure 1. The iri-
line moiety and an N-(2,5-diisopropyl-
dium C,N-complex
phenyl) NHC unit proved to be an
1 from Burgess and
efficient and highly enantioselective
hydrogenation catalyst. Other deriva-
tives with different substituents at the
oxazoline or NHC ring, as well as NHC-
based ligands with other carbon scaf-
co-workers.[6] cod=
cyclo-1,5-octadiene.
folds,[7] gave only low to moderate enantioselectivities. The
complex 1 possesses distinct features which distinguish this
catalyst from iridium N,P-ligand complexes. As shown by
Burgess and co-workers, iridium hydride complexes which are
formed from 1 during hydrogenation are much less acidic than
analogous iridium hydrides derived from N,P-ligand com-
plexes.[8]
This difference in acidity, which can be explained by the
different electronic features of NHC and phosphine ligands,
has important consequences in the hydrogenation of acid-
sensitive substrates. Burgess et al. have compared the perfor-
mance of iridium N,P-ligand complexes and their catalyst 1 in
the hydrogenation of enol ethers. While the N,P complexes
produced substantial amounts of acid-induced side products,
the NHC-based catalyst gave only the desired hydrogenation
product.[8] We too encountered problems related to acid-
induced side reactions such as water elimination in the
hydrogenation of certain acid-sensitive allylic alcohols[9] or
desilylation of silyl-protected alcohols in hydrogenations with
N,P-ligand complexes.
Scheme 1. Diastereoselective synthesis of (R,R,R)-g-tocopherol ace-
tate.[4]
which for the first time have allowed the hydrogenation of
=
purely alkyl-substituted C C bonds with high efficiency and
enantioselectivity.[4] The potential of these catalysts is illus-
trated by the highly stereoselective, catalyst-controlled hydro-
genation of g-tocotrienyl acetate, thus introducing two
stereogenic centers in the desired R,R configuration in one
As 1 is the only efficient catalyst of this type available
today, it seemed desirable to explore other NHC-based C,N-
ligand systems which could serve as an alternative in reactions
that do not proceed well with catalyst 1. In view of the many
successful applications of pyridine phosphinite ligands
(Scheme 1), we decided to prepare a series of NHC pyridine
analogues (Scheme 2). The bicyclic pyridyl alcohols 2a–c
having a five-, six-, or seven-membered carbocyclic ring,
which were chosen as starting materials, were readily
prepared following published procedures[4b] and conveniently
obtained as enantiomerically pure R or S alcohols through
[*] Dr. A. Schumacher, M. Bernasconi, Prof. Dr. A. Pfaltz
Department of Chemistry, University of Basel
4056 Basel (Switzerland)
E-mail: andreas.pfaltz@unibas.ch
[**] Financial support from the Swiss National Science Foundation is
gratefully acknowledged. We thank Dr. Markus Neuburger (Labo-
ratory for Chemical Crystallography, University of Basel) for the
crystal structure analyses.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201301251.
7422
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 7422 –7425

Angewandte
Chemie
À
cantly longer than the Ir C bond to the NHC unit (2.05 ꢀ).
Space-filling models show that the NHC and the phosphinite
moieties occupy similar regions in space. In particular, the
chiral arrangement of the pyridine-bound phenyl group,
which plays a crucial role in the enantioselection process,[2]
is virtually identical in the two complexes. Therefore, similar
enantioselectivities were expected for NHC- and phosphinite-
based catalysts such as 10a and 11, or 10b and 12, respectively.
For the synthesis of the enantiomerically pure ligands 9a–
c the strategy of Fꢁrstner et al.[13] was chosen, and involved
imidazolium ring formation by condensation of the dihy-
drooxazolium salt 8 with a primary amine. The required
enantiomerically pure amines 4a and 4b were prepared in
essentially quantitative yields from the corresponding alco-
hols by mesylation and subsequent reaction with sodium
azide followed by reduction with Pd/C.[14] The seven-mem-
bered ring derivative 4c, however, did not form the azide and
racemic mesylate was isolated instead. MM2 calculations
suggest that the lack of reactivity of the mesylate is due to the
unfavorable conformation of the seven-membered ring,
which sterically impedes an SN2 reaction with azide. Thus,
the corresponding amine 4c was prepared by an oxidation/
oxime formation/reduction sequence (see the Supporting
Information).
Scheme 2. Synthesis of the iridium NHC pyridine complexes 10a–c:
a) MsCl, Et3N, DMAP (cat.), NaN3, THF, 08C, 20 min, then DMSO, RT,
2.5 h (quant.). b) Pd/C, H2, RT, 4 h, EtOH (quant.). c) nBuLi, RNH2,
THF (95%). d) HCO2H, Ac2O, THF (87%). e) Ac2O, HClO4 (aq.) (75–
80%). f) Toluene; Et2O, HClO4 (aq.) (41–65%). g) LiOtBu, THF, [{Ir-
(cod)Cl}2], RT, 2 h, then NaBArF, CH2Cl2, 30 min (44–72%). DMAP=
4-(N,N-dimethylamino)pyridine, DMSO=dimethylsulfoxide, Ms=me-
thanesulfonyl, THF=tetrahydrofuran.
The dihydrooxazolium salt 8 was prepared in 75–80%
overall yield from bromoacetaldehyde diethylacetal accord-
ing to the procedures of Fꢁrstner et al.[13] Condensation with
the amines 4a–c led to the corresponding imidazolium salts
9a–c in yields of 41–65%. When the imidazolium salts were
treated with lithium tert-butoxide in THF in the presence of
[{Ir(cod)Cl}2] the desired Ir/NHC complexes were formed,
and after anion exchange were isolated as the BArF salts 10a–
c in 44–72% yield (Figure 3).
enzymatic kinetic resolution.[10] Initially, we planned to
introduce the NHC unit by converting the hydroxy group of
the pyridyl alcohol into a leaving group such as tosylate or
iodide with subsequent N alkylation of the corresponding
imidazole. However, this approach proved to be unsuitable
because of concomitant racemization. Nevertheless, the
racemic iridium complex 10b prepared in this way could be
crystallized and subjected to X-ray analysis.
Comparison of the X-ray structures of the iridium
complex 10b and its dicyclohexyl phosphinite analogue
show a very similar orientation of the 2-phenyl-tetrahydro-
quinoline component and a boat conformation of the chelate
ring in both cases (Figure 2).[11,12] The distance between the
pyridine nitrogen atom and the iridium center is 2.14 ꢀ in
both complexes, whereas the P-Ir bond (2.29 ꢀ) is signifi-
Figure 3. The iridium NHC pyridine catalysts and their phosphinite
analogues.
The new NHC-based iridium catalysts 10a–c were eval-
uated in the asymmetric hydrogenation of six typical test
substrates and compared with the analogous pyridine phos-
phinite catalysts 11 and 12 (Scheme 3). All three complexes
showed high catalytic activity and, with two exceptions (10c
with substrate 13; 10a with 15), led to full conversion within
2 hours under standard conditions at a 1 mol% catalyst
loading.[15] With the exception of the terminal alkene 16 all
Figure 2. Comparison of the X-ray structures of the iridium complex
10b Àand its dicyclohexyl phosphinite analogue. Hydrogen atoms and
BArF ions in both structures have been omitted for clarity.[11]
Angew. Chem. Int. Ed. 2013, 52, 7422 –7425
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7423

.
Angewandte
Communications
Scheme 4. Hydrogenation of acid-sensitive substrates. TMS=trime-
thylsilyl.
phosphinite complex 11 led to substantial amounts of the
deprotected saturated alcohol 21, while for the NHC-based
catalysts 10a and 10b silylether cleavage occurred to a much
lesser extent. The catalyst 10b afforded the saturated
silylether 20 in 95% yield and 91% ee with only 5% cleavage
product, which is almost identical to the results obtained with
the Burgess catalyst 1.
The acid-sensitive Boc-protected analogue of 19 also
underwent clean hydrogenation with the catalyst 10b to give
the Boc-protected saturated alcohol with 69% ee in essen-
tially quantitative yield, whereas the corresponding N,P li-
gand complex 12 gave a complex mixture of products, which
contained only 34% of the desired hydrogenation product
(see the Supporting Information). The catalyst 10b also
allowed hydrogenation of the substrate 22[5a] in quantitative
yield without notable cleavage of the tert-butylester group
(Scheme 4b). In contrast, hydrogenation with the analogous
pyridine phosphinite complex 12 afforded the acid 24 as the
main product.
In summary, we have shown that replacement of the
phosphinite group, in pyridine-based iridium complexes such
as 11 or 12, by a NHC unit leads to efficient and highly
enantioselective hydrogenation catalysts. In the hydrogena-
tion of various olefins the phosphinite- and NHC-based
catalysts showed very similar enantioselectivities. However,
as a result of the lower acidity of iridium hydride intermedi-
ates produced from NHC-based complexes, these catalysts
are much better suited for the hydrogenation of acid-sensitive
substrates. In addition, the new NHC-pyridine ligands may
also prove useful for other applications in asymmetric
catalysis.
Scheme 3. Asymmetric hydrogenation of the substrates 13–18 with the
iridium NHC-based catalysts 10a–c and the analogous pyridine
phosphinite complexes 11 and 12. [a] 1.0 bar H2.
other substrates gave excellent enantioselectivities, compara-
ble to those obtained with analogous pyridine phosphinite
complexes 11 or 12. Overall, the complexes 10a and 10b
outperformed 10c, which gave inferior results in the hydro-
genation of 13, 16, and 18. In addition to catalyst 10b an
analogous complex bearing an N-mesityl substituent at the
NHC ring was prepared and tested. Overall, this catalyst
proved to be less active, thus giving lower conversion and in
most cases lower ee values.
The enantioselectivities of the catalysts 10a and 10b in the
hydrogenation of alkenes 13, 15, and 18 were similar to those
reported for the Burgess catalyst 1.[6] For substrates 16 and 17
the enantioselectivities of the pyridine- and oxazoline-based
catalysts were distinctly different, and is not surprising
considering the differences in geometry and steric demand
between these ligand systems. For the terminal alkene 16 the
Burgess catalyst 1 was superior (89% ee versus 63% ee with
10a), whereas the Z-alkene 17 gave better results with
catalysts 10a–c (92–94% ee versus 80% ee with 1). These
results indicate that the pyridine-derived ligands are a useful
addition to the oxazoline-derived catalyst 1, and enhance the
application range of NHC-based iridium catalysts.
To test the compatibility of the new NHC-based catalysts
with acid-sensitive substrates, we studied the hydrogenation
of the silyl-protected allylic alcohol 19 (Scheme 4a). For
comparison the reaction was also carried out with the Burgess
catalyst 1 and the pyridine phosphinite complex 11. All
catalysts gave full conversion within 2 hours at 50 bar hydro-
gen pressure at room temperature and a catalyst loading of
1 mol%. As expected, hydrogenation with the pyridine
Experimental Section
Representative procedures for the synthesis of imidazolium salts 9
and the corresponding Ir-NHC complexes 10.
7424 www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 7422 –7425

Angewandte
Chemie
(S)-3-(2,6-Diisopropylphenyl)-1-(2-phenyl-5,6,7,8-tetrahydroqui-
nolin-8-yl)-1H-imidazol-3-ium perchlorate ((S)-9b): The amine (S)-
4b (700 mg, 3.12 mmol, 1.00 equiv) was added to a suspension of 8
(1.22 g, 3.12 mmol, 1.00 equiv) in toluene (40 mL) and the reaction
mixture was stirred at room temperature for 6 h. During the reaction
separation of a second phase was observed. Stirring was discontinued,
the toluene phase was removed with a pipette, and the residue was
triturated with Et2O (3 ꢂ 10 mL). Toluene (25 mL) was added to the
remainder and followed by 70% aq. HClO4 (0.27 mL, 1.00 equiv).
The mixture was stirred at 808C overnight, then the solvent was
evaporated and the residue taken up in CH2Cl2 (40 mL; if a precipitate
appeared at this point, it was filtered off prior to further processing).
NH3 (1.10 mL, 7m in MeOH, 2.50 equiv) was then added to the clear
solution and the precipitate of NH4ClO4 was filtered off through
a small plug of celite. Evaporation of the solvent and crystallization of
the final product with CH2Cl2/Et2O (1:5) provided the corresponding
imidazolium salt (S)-9b as an analytically pure, white powder (1.08 g,
2.02 mmol, 65%).
ligands of this type have been independently developed:Q. B.
Liu, C. B. Yu, Y. G. Zhou, Tetrahedron Lett. 2006, 47, 4733 –
4736.
[5] a) D. H. Woodmansee, M.-A. Mꢁller, L. Trçndlin, E. Hçrmann,
A. Pfaltz, Chem. Eur. J. 2012, 18, 13780 – 13786; b) A. Ganic, A.
Pfaltz, Chem. Eur. J. 2012, 18, 6724 – 6728; c) A. Baeza, A. Pfaltz,
Chem. Eur. J. 2010, 16, 2036 – 2039; d) C. Valla, A. Baeza, F.
Menges, A. Pfaltz, Synlett 2008, 3167 – 3171; e) Y. Zhu, K.
´
Burgess, Adv. Synth. Catal. 2008, 350, 979 – 983; f) I. Gazic Smi-
´
lovic, E. Casas-Arcꢄ, S. J. Roseblade, U. Nettekoven, A. Zanotti-
ˇ
ˇ
ˇ
Gerosa, M. Kovacevic, Z. Casar, Angew. Chem. 2012, 124, 1038 –
1042; Angew. Chem. Int. Ed. 2012, 51, 1014 – 1018; g) J. Mazuela,
P.-O. Norrby, P. G. Andersson, O. Pꢅmies, M. Diꢄguez, J. Am.
Chem. Soc. 2011, 133, 13634 – 13645; h) P. Cheruku, J. Diesen,
P. G. Andersson, J. Am. Chem. Soc. 2008, 130, 5595 – 5599; i) S.
Li, S.-F. Zhu, C.-M. Zhang, S. Song, Q.-L. Zhou, J. Am. Chem.
Soc. 2008, 130, 8584 – 8585.
[6] a) M. T. Powell, D. R. Hou, M. C. Perry, X. Cui, K. Burgess, J.
Am. Chem. Soc. 2001, 123, 8878 – 8879; b) M. C. Perry, X. Cui,
M. T. Powell, D. R. Hou, J. H. Reibenspiess, K. Burgess, J. Am.
Chem. Soc. 2003, 125, 113 – 123; c) M. C. Perry, K. Burgess,
Tetrahedron: Asymmetry 2003, 14, 951 – 961.
[7] a) S. Nanchen, A. Pfaltz, Chem. Eur. J. 2006, 12, 4550 – 4558;
b) S. Nanchen, A. Pfaltz, Helv. Chim. Acta 2006, 89, 1559 – 1573;
c) C. Bolm, T. Focken, G. Raabe, Tetrahedron: Asymmetry 2003,
14, 1733 – 1746.
[8] Y. Zhu, Y. Fan, K. Burgess, J. Am. Chem. Soc. 2010, 132, 6249 –
6253.
[9] A. Schumacher, M. G. Schrems, A. Pfaltz, Chem. Eur. J. 2011, 17,
13502 – 13509.
[10] M. Maywald, A. Pfaltz, Synthesis 2009, 3654 – 3660.
[11] a) CCDC 923981 (rac-10b), and 923982 (dicyclohexyl phosphin-
ite analogue), contains the supplementary crystallographic data
for this paper. These data can be obtained free of charge from
The Cambridge Crystallographic Data Centre via www.ccdc.
cam.ac.uk/data_request/cif; b) The data for dicyclohexyl-phos-
phinite analog was acquired by S. Kaiser, Dissertation, Univer-
sity of Basel, 2005. (http://edoc.unibas.ch/288/).
(h4-1,5-Cyclooctadiene)-{(S)-8-(3-(2,6-diisopropylphenyl)-2,3-
dihydro-carbene-1-yl)-2-phenyl-5,6,7,8-tetrahydroquinoline}-iri-
dium(I) tetrakis(3,5-bis(trifluoromethyl) phenyl)borate ((S)-10b):
LiOtBu (0.48 mL, 1m in hexane, 1.50 equiv) was added to a suspension
of (S)-9b (170 mg, 0.32 mmol, 1.00 equiv) and [{Ir(cod)Cl}2] (107 mg,
0.16 mmol, 0.50 equiv) in THF (10 mL) at room temperature. The
orange suspension turned dark brown and was stirred for 20 h at 708C
under argon. After cooling to room temperature the volatiles were
removed in vacuo. NaBArF (365 mg, 0.41 mmol, 1.30 equiv) dissolved
in CH2Cl2 (5 mL) was added and the reaction was stirred for 30 min at
room temperature. Then the solvent was evaporated and the residue
was purified by column chromatography (SiO2, 2.5 ꢂ 15 cm, MTBE,
then CH2Cl2) to give complex (S)-10b as an orange solid (364 mg,
0.23 mmol, 72%).
Received: February 12, 2013
Published online: June 13, 2013
Keywords: alkenes · asymmetric catalysis · hydrogenation ·
.
iridium · N heterocyclic carbene
[12] One of the reviewers pointed out that because of the NHC unit
bearing two different N substituents, the complex 10b contains
À
a chirality axis formed by the Ir C bond as an additional
[1] R. Crabtree, Acc. Chem. Res. 1979, 12, 331 – 337.
chirality element (see, e.g.: D. Enders, H. Gielen, J. Runsink, K.
Breuer, S. Brode, K. Boehn, Eur. J. Inorg. Chem. 1998, 913 – 919;
A. Ros, M. Alcarazo, J. Iglesias-Siguenza, E. Diez, E. Alvarez, R.
Fernandez, J. M. Lassaletta, Organometallics 2008, 27, 4555 –
4564). However, no experimental evidence for the formation
of diastereomeric complexes was found. The conformation
[2] a) X. Cui, K. Burgess, Chem. Rev. 2005, 105, 3272 – 3296; b) K.
Kꢃllstrçm, I. Munslow, P. G. Andersson, Chem. Eur. J. 2006, 12,
3194 – 3200; c) S. J. Roseblade, A. Pfaltz, Acc. Chem. Res. 2007,
40, 1402 – 1411; d) T. L. Church, P. G. Andersson, Coord. Chem.
Rev. 2008, 252, 513 – 531; e) D. H. Woodmansee, A. Pfaltz, Top.
Organomet. Chem. 2011, 34, 31 – 76; f) D. H. Woodmansee, A.
Pfaltz, Chem. Commun. 2011, 47, 7912 – 7916; g) A. Cadu, P. G.
Andersson, J. Organomet. Chem. 2012, 714, 3 – 11.
[3] For iridium catalysts based on chiral P,O ligands, see: D. Rageot,
D. H. Woodmansee, B. Pugin, A. Pfaltz, Angew. Chem. 2011,
123, 9772 – 9775; Angew. Chem. Int. Ed. 2011, 50, 9598 – 9601.
[4] a) S. Bell, B. Wꢁstenberg, S. Kaiser, F. Menges, T. Netscher, A.
Pfaltz, Science 2006, 311, 642 – 644; b) S. Kaiser, S. P. Smidt, A.
Pfaltz, Angew. Chem. 2006, 118, 5318 – 5321; Angew. Chem. Int.
Ed. 2006, 45, 5194 – 5197; c) A. Wang, B. Wꢁstenberg, A. Pfaltz,
Angew. Chem. 2008, 120, 2330 – 2332; Angew. Chem. Int. Ed.
2008, 47, 2298 – 2300; d) A. Wang, R. P. Fraga, E. Hçrmann, A.
Pfaltz, Chem. Asian J. 2011, 6, 599 – 606; e) Pyridine phosphinite
À
around the Ir C bond is likely dictated by the geometric
constraints of the chelate ring, which favor the geometry found
in the crystal structure and destabilize other conformers.
[13] A. Fꢁrstner, M. Alcarazo, V. Cꢄsar, C. W. Lehmann, Chem.
Commun. 2006, 2176 – 2178.
[14] J. Uenishi, M. Hamada, Synthesis 2002, 625 – 630.
[15] Reduction of the catalyst loading to 0.25 mol% in the hydro-
genation of the substrate 13 with complex 10a did not affect the
enantioselectivity. However, conversion dropped from > 99% to
28% when the catalyst loading was decreased from 0.5 to
0.25 mol%.
Angew. Chem. Int. Ed. 2013, 52, 7422 –7425
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7425