Proline-based P,O ligand/iridium complexes as highly selective catalysts: Asymmetric hydrogenation of trisubstituted alkenes

Article abstract of DOI:10.1002/anie.201104105

P,O joins the mix: P,O ligands (L1) form efficient iridium catalysts for the asymmetric hydrogenation of olefins. The proline-derived ligands lead to high enantioselectivities with several classes of alkenes, most notably with α,β-unsaturated carboxylic e

Full text of DOI:10.1002/anie.201104105

Communications
DOI: 10.1002/anie.201104105
Asymmetric Catalysis
Proline-Based P,O Ligand/Iridium Complexes as Highly Selective
Catalysts: Asymmetric Hydrogenation of Trisubstituted Alkenes**
Denise Rageot, David H. Woodmansee, Benoꢀt Pugin, and Andreas Pfaltz*
Heteroatom-based bidentate ligands are ubiquitous in the
field of asymmetric catalysis. However, only a few ligands that
coordinate to a transition metal through a carbonyl oxygen
atom have been reported. Tomioka and co-workers have
extensively studied the chiral amidophosphine LA1 (see
Scheme 1 for structure) in combination with rhodium or
copper in the asymmetric Michael addition of a variety of
organometallic reagents to a,b-unsaturated carbonyl com-
pounds with moderate to excellent enantioselectivities.[1–4] In
the coordination to rhodium(I), this P,O ligand was shown to
behave as a hemilabile ligand, in which the amide carbonyl
group is weakly bound to the transition metal.[1] More
recently Reek et al. have published on a rhodium-based
phosphine urea P,O-ligand system which gives moderate
enantioselectivity in asymmetric hydrogenation.[5,6]
in the iridium-catalyzed asymmetric hydrogenation
(Scheme 1).
As described by Tomioka and co-workers, a linear
approach was used to synthesize these P,O ligands with
introduction of the phosphine moiety prior to the substituents
at the N atom (Scheme 1).[4a–d] Starting from (S)-tert-butyl-2-
(bromomethyl)pyrrolidine-1-carboxylate (1) precursors
2
were obtained by nucleophilic substitution with various
metallated phosphines.[25] After removal of the N-Boc pro-
tecting group, the free amine was reacted with acetyl
chlorides, isocyanates, or carbamoyl chlorides, allowing
access to a variety of P,O ligands. The introduction of the
N substituents generally proceeded in high yield for both,
amides (43–98%) and ureas (61–93%). Oxygen-sensitive
compounds such as di-tert-butyl- or dicyclohexylphosphines
were protected as borane adducts to avoid oxidation.
Iridium complexes with chiral N,P ligands have emerged
as highly efficient catalysts for the asymmetric hydrogenation
of olefins.[7–24] In contrast to rhodium and ruthenium diphos-
phine catalysts, they do not require a coordinating group near
=
the C C bond and, therefore, can be applied to a wide range
of functionalized and unfunctionalized olefins. However,
there are still important substrate classes that give unsat-
isfactory results with known catalysts. Therefore, the search
for new ligands that could fill these methodological gaps
continues.
In a broad automated screening of various metal/ligand
combinations, an iridium complex formed in situ from the
ligand LA1 and [Ir(cod)]2BArF (cod = 1,5-cyclooctadiene;
BArF = tetrakis[bis-3,5-(trifluoromethyl)phenyl]borate) gave
promising enantioselectivities of up to 68% ee in the hydro-
genation of (E)-1,2-diphenylprop-1-ene. This result was
À
unexpected considering the assumed lability of the Ir O
bond. The ready accessibility and the modular nature of the
ligand LA1 prompted us to synthesize a diverse library of
proline-derived P,O ligands by systematic variation of the
substituents on the P and N atoms to evaluate their potential
[*] D. Rageot, Dr. D. H. Woodmansee, Prof. A. Pfaltz
University of Basel, Department of Chemistry
St. Johanns-Ring 19, 4056 Basel (Switzerland)
E-mail: andreas.pfaltz@unibas.ch
Scheme 1. Synthesis of the l-proline-based P,O ligands. 1a) KPPh2,
THF for R1 =Ph, n=0; 1b) o-Tol2PCl, Na, THF for R1 =o-Tol, n=0;
1c) R1 PH·BH3, nBuLi, THF for R1 =tBu, Cy, n=1. 2) HCl, 1,4-dioxane.
2
Dr. B. Pugin
Solvias AG, P.O. Box, 4002 Basel (Switzerland)
3a) R2COCl, NEt3, CH2Cl2 for R1 =Ph and R2 =tBu, CPh3; 3b) R2COCl,
K2CO3, CH2Cl2 for R1 =tBu, Cy, and R2 =1-Ad, CPh3; 3c) Ph3CNCO,
NEt3, CH2Cl2 for R1 =Ph and R2 =CPh3; 3d) R2NCO, K2CO3, CH2Cl2 for
R1 =tBu, Cy and R2 =1-AdNH, MesNH, Ph3CNH; 3e) iPr2NCOCl, NEt3
[**] Support of this work by the Swiss National Science Foundation and
the Federal Commission for Technology and Innovation (KTI) is
gratefully acknowledged. We thank Dr. Markus Neuburger for the
crystal-structure analysis.
for R1 =o-Tol and R2 =iPr2N; 3f) R2 NCOCl, K2CO3, CH2Cl2 for
2
R1 =tBu, Cy and R2 =iPr2N, Cy2N. 4) Et2NH. Ad=adamantyl, Boc=
tert-butoxycarbonyl, Cy=cyclohexyl, Mes=2,4,6-trimethylphenyl,
THF=tetrahydrofuran.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201104105.
9598
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 9598 –9601

Table 2: Screening of the urea-based ligands LU31–LU37 and LU41–LU45 in
the asymmetric hydrogenation of (E)-ethyl 3-phenylbut-2-enoate (S1).[a]
Crystals suitable for X-ray diffraction analysis of the
iridium/amidophosphine LA1 and iridium/ureaphosphine
L
U35 complexes were obtained to determine the coordination
mode of this new ligand class (Figure 1).[26] The solid-state
structures clearly indicate coordination of the P,O ligands in a
bidentate fashion to the metal center with both donors
bonding in a s fashion, which is consistent with the rhodium
amidophosphine complex reported previously.[1]
Entry
L*
R1
R2
Yield [%][b]
ee [%][c]
1
2
3
4
5
6
7
8
(S)-LU3
(S)-LU32
(S)-LU33
(S)-LU3
(S)-LU3
(S)-LU36
(S)-LU3
(S)-LU4
(S)-LU42
(S)-LU4
(S)-LU44
(S)-LU4
1
tBu
Cy
tBu
Cy
Ph
tBu
Cy
o-Tol
tBu
Cy
tBu
Cy
1-AdNH
1-AdNH
MesNH
MesNH
Ph3CNH
Ph3CNH
Ph3CNH
iPr2N
iPr2N
iPr2N
Cy2N
Cy2N
>99
>99
>99
>99
99
92 (R)
93 (R)
88 (R)
90 (R)
92 (R)
94 (R)
98 (R)
44 (R)
97 (R)
90 (R)
99 (R)
97 (R)
4
5
98
7
1
>99
>99
>99
>99
>99
>99
9
10
11
12
3
5
[a] Reaction conditions: See Table 1.
Figure 1. Crystal structures of complex [(R)-(LA1)Ir(cod)]BArF and [(S)-
(LU35)Ir(cod)]PF6; the counterions have been omitted for clarity. Red O,
blue N, magenta P, dark gray Ir, light gray C.
entries 1–6). We observed that very sterically demanding
substituents attached to the carbonyl group led to signifi-
cantly higher selectivities in combination with the stronger
electron-donating dialkyl phosphines. Carbamates 2a–d gave
only low to moderate enantioselectivities (Table 1, entries 7–
10).
The urea-based P,O ligands also gave high enantioselec-
tivities of up to 99% ee (Table 2, entry 11) under identical
screening conditions. Again, very bulky substituents attached
to the carbonyl function were essential for obtaining high
levels of enantioselectivity. Electron-rich alkyl-substituted
phosphines consistently demonstrated higher enantioselectiv-
ities than the corresponding diarylphosphines. In some cases
dicyclohexylphosphines were superior to di-tert-butylphos-
phine analogues, possibly as a result of steric overcrowding at
the metal center (Table 2, entries 1–4, 6,7). Ortho-tolyl-
substituted phosphines generally induced drastically lower
enantioselectivities (Table 2, entry 8).
To compare our P,O ligand complexes with the state-of-
the-art iridium catalysts, we carried out a broad screening
including some demanding substrates for which efficient
catalysts are still lacking (Figure 2; for a comprehensive
compilation of the results, see the Supporting Information).
Alkenes bearing carbonyl or hydroxy groups seemed of
special interest, because Lewis basic substituents of this type
could severely affect the performance of the catalyst by
replacing the coordinating carbonyl group of the P,O ligand.
The proline-based P,O catalysts gave excellent selectiv-
ities with the much studied (E)-a-methylstillbene (S2), with
many of the better ligands giving results equal to the best
N,P systems. However, the more strongly coordinating allylic
alcohol S2 was reduced with much lower selectivity with these
catalysts in general, albeit with full conversion.
The library of P,O ligands was evaluated in the iridium-
catalyzed asymmetric hydrogenation of various prochiral
trisubstituted unfunctionalized and functionalized alkenes by
an in situ protocol for a rapid assessment of the selectivity/
reactivity profile. The precatalyst was generated through
complexation of [Ir(cod)2]BArF and the respective P,O ligand
in dichloromethane with subsequent addition of the stock
solution to the substrate. Preliminary investigations were
conducted with S1 as an initial screen (Table 1 and Table 2).
We were pleased to find that these P,O ligands showed high
activity and selectivity in the test reactions. Excellent
conversions and enantiomeric excesses of up to 98% were
obtained when amidophosphines were employed (Table 1,
Table 1: Screening of the amide- and carbamate-based ligands LA1–LA6
and 2a–d in the asymmetric hydrogenation of (E)-ethyl 3-phenylbut-2-
enoate (S1).[a]
Entry
L*
R1
R2
Yield [%][b]
ee [%][c]
1
2
3
4
5
6
7
8
9
(S)-LA1
(S)-LA2
(S)-LA3
(S)-LA4
(S)-LA5
(S)-LA6
(S)-2a
(S)-2b
(S)-2c
(S)-2d
Ph
tBu
Cy
Ph
tBu
Cy
tBu
65
>99
98
33 (R)
50 (R)
26 (R)
89 (R)
98 (R)
94 (R)
42 (R)
9 (R)
1-Ad
1-Ad
CPh3
CPh3
CPh3
OtBu
OtBu
OtBu
OtBu
88
>99
>99
74
99
80
Ph
o-Tol
tBu
Cy
78 (R)
75 (R)
Conjugate reduction of a,b-unsaturated ketones and
esters appears to be a particularly well-suited area of
application for these catalysts with both classes of alkenes
giving high ee values. Substrates S4a, S4b, and S6 were
recently reduced in record levels of selectivity in this
10
99
[a] Reaction conditions: [Ir(cod)2BArF] (2.50 mmol), ligand (2.50 mmol),
CH2Cl2 (0.5 mL), substrate (250 mmol). [b] Yields were determined by GC
analysis. [c] Enantioselectivities were determined by GC analysis using a
chiral stationary phase.
Angew. Chem. Int. Ed. 2011, 50, 9598 –9601
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 9599

Communications
Figure 3. Qualitative model rationalizing the enantioselectivity of Ir/
N,P ligand and Ir/P,O ligand catalysts.
major enantiomer is formed via an intermediate in which the
small H atom is pointing toward the large substituent at the
stereogenic center of the oxazoline ring, thus minimizing the
steric repulsion between the chiral ligand and the substrate.
Assuming an analogous coordination mode for P,O com-
plexes, a sterically similar situation results if we infer a ligand
conformation based on the crystal structures shown in
Figure 1. The large substituents of the amide or urea group
occupy the same region in space as the substituent in the
oxazoline ring of a phox ligand. This qualitative model
explains why rather bulky amide or urea groups are necessary
for high enantioselectivity and rationalizes the sense of
asymmetric induction.
Our results show that proline-derived phosphines bearing
bulky amide or urea groups at the pyrrolidine N atom form
efficient Ir catalysts for the asymmetric hydrogenation of
olefins. These new P,O ligands give high enantioselectivities
with several classes of alkenes, most notably with a,b-
unsaturated carboxylic esters and ketones, where they
match or even surpass the ee values reported for the best
N,P and C,N ligands. The ligands are readily prepared in
enantiomerically pure form starting from inexpensive chiral
precursors. Their modular nature easily allows structural
tuning by variation of the coordinating P and O units. We are
currently exploring further applications of these ligands
including other metal-catalyzed reactions.
Figure 2. Selected hydrogenation results; see the Supporting Informa-
tion for reaction conditions.
laboratory with pyridyl phosphinite based catalysts.[13] The
easily prepared proline-based P,O ligands give comparable
results to these systems but with the added advantage of using
inexpensive starting materials from the chiral pool for the
ligand syntheses. The very challenging a-aryl ester S7 was
reduced in moderate yield and good selectivity. Perhaps more
impressive is the level of enantioselectivity obtained with
methyl ketone S5a, which was reduced by using several
catalysts, with selectivities much higher than previously
reported for the N,P based systems.[20c] Conversion in the
hydrogenation of S5a was always complete, but sometimes
accompanied, as reported, by the reduction to the saturated
alcohol in variable and uncontrolled amounts.[20c] To over-
come this issue, the reaction was carried out with a catalyst
that was isolated instead of using the in situ complexation
method (see the Supporting Information). By using this
procedure essentially no overreduction was observed. The
corresponding phenyl ketone S5b gave similar results to that
of the methyl ketone S5a.
Received: June 15, 2011
Published online: August 31, 2011
Keywords: asymmetric catalysis · asymmetric hydrogenation ·
.
iridium · ligand design · P,O ligand
[1] M. Kuriyama, K. Nagai, K.-I. Yamada, Y. Miwa, T. Taga, K.
Tomioka, J. Am. Chem. Soc. 2002, 124, 8932 – 8939.
[2] For reviews on hemilabile ligands, see: a) P. Braunstein, F. Naud,
Angew. Chem. 2001, 113, 702 – 722; Angew. Chem. Int. Ed. 2001,
40, 680 – 699; b) P. Braunstein, J. Organomet. Chem. 2004, 689,
3953 – 3967; c) A. Bader, E. Lindner, Coord. Chem. Rev. 1991,
108, 27 – 110.
[3] Q. Chen, M. Kuriyama, X. Hao, T. Soeta, Y. Yamamoto, K.-I.
Yamada, K. Tomioka, Chem. Pharm. Bull. 2009, 57, 1024 – 1027.
[4] a) Y. Nakagawa, K. Matsumoto, K. Tomioka, Tetrahedron 2000,
56, 2857 – 2863; b) M. Kanai, Y. Nakagawa, K. Tomioka,
Tetrahedron 1999, 55, 3843 – 3854; c) Y. Nakagawa, M. Kanai,
Y. Nagaoka, K. Tomioka, Tetrahedron 1998, 54, 10295 – 10307;
Although experimental structural information about the
catalytic cycle is still lacking, computational studies of Ir/
N,P ligand and Ir/C,N ligand catalysts suggest that the enan-
tiodiscrimination occurs during the migratory insertion
step.[7,19a–c,21–24] In an Ir/phox complex the olefin is proposed
to coordinate trans to the P atom as shown in Figure 3. The
9600 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 9598 –9601

d) Y. Nakagawa, M. Kanai, Y. Nagaoka, K. Tomioka, Tetrahe-
dron Lett. 1996, 37, 7805 – 7808; e) M. Kanai, K. Koga, K.
Tomioka, Tetrahedron Lett. 1992, 33, 7193 – 7196.
[5] J. Meeuwissen, R. Detz, A. J. Sandee, B. de Bruin, M. A. Siegel,
A. L. Spek, J. N. H. Reek, Eur. J. Inorg. Chem. 2010, 2992 – 2997.
[6] J. Meeuwissen, R. J. Detz, A. J. Sandee, B. de Bruin, J. N. H.
Reek, Dalton Trans. 2010, 39, 1929 – 1931.
Hall, J. Am. Chem. Soc. 2004, 126, 16688 – 16689; c) J. Zhou,
J. W. Ogle, Y. Fan, V. Banphavichit(Bee), Y. Zhu, K. Burgess,
Chem. Eur. J. 2007, 13, 7162 – 7170; d) J. Zhao, K. Burgess, J. Am.
Chem. Soc. 2009, 131, 13236 – 13237; e) Y. Zhu, K. Burgess, Adv.
Synth. Catal. 2008, 350, 979 – 983; f) S. Nanchen, A. Pfaltz, Chem.
Eur. J. 2006, 12, 4550 – 4558.
=
[20] For the asymmetric hydrogenation of the C C bond of a,b-
[7] X. Cui, K. Burgess, Chem. Rev. 2005, 105, 3272 – 3296.
[8] T. L. Church, P. G. Andersson, Coord. Chem. Rev. 2008, 252,
513 – 531.
[9] D. H. Woodmansee, A. Pfaltz, Top. Organomet. Chem. 2011, 34,
31 – 76.
[10] S. J. Roseblade, A. Pfaltz, Acc. Chem. Res. 2007, 40, 1402 – 1411.
[11] M. Schçnleber, R. Hilgraf, A. Pfaltz, Adv. Synth. Catal. 2008,
350, 2033 – 2038.
[12] S. Kaiser, S. P. Smidt, A. Pfaltz, Angew. Chem. 2006, 118, 5318 –
5321; Angew. Chem. Int. Ed. 2006, 45, 5194 – 5197.
[13] D. H. Woodmansee, M.-A. Mꢀller, M. Neuburger, A. Pfaltz,
Chem. Sci. 2010, 1, 72 – 78.
unsaturated ketones, see: a) W.-J. Lu, Y.-W. Chen, X.-L. Hou,
Angew. Chem. 2008, 120, 10287 – 10290; Angew. Chem. Int. Ed.
2008, 47, 10133 – 10136; b) S.-M. Lu, C. Bolm, Angew. Chem.
2008, 120, 9052 – 9055; Angew. Chem. Int. Ed. 2008, 47, 8920 –
8923; c) S.-M. Lu, C. Bolm, Chem. Eur. J. 2008, 14, 7513 – 7516.
[21] C. Hedberg, K. Kꢁllstrçm, P. Brandt, L. K. Hansen, P. G.
Andersson, J. Am. Chem. Soc. 2006, 128, 2995 – 3001.
[22] T. L. Church, T. Rasmussen, P. G. Andersson, Organometallics
2010, 29, 6769 – 6781.
[23] P. Brandt, C. Hedberg, P. G. Andersson, Chem. Eur. J. 2003, 9,
339 – 347.
[24] K. H. Hopmann, A. Bayer, Organometallics 2011, 30, 2483 –
2497.
[25] For the synthesis of (S)-tert-butyl-2-(bromomethyl)pyrrolidine-
1-carboxylate (1), see: B. Hinzen, H. Broetz, R. Endermann, K.
Henninger, H. Paulsen, S. Raddatz, S. Anlauf, Ger. Offen. DE
10308107A1, 2004.
[26] Crystal data for (R)-[(LA1)Ir(cod)] and (S)-[(LU35)Ir(cod)]: see
the Supporting Information. CCDC 829325 and 829326 contain
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.
[14] S. Li, S.-F. Zhu, C.-M. Zhang, S. Song, Q.-L. Zhou, J. Am. Chem.
Soc. 2008, 130, 8584 – 8585.
[15] A. Baeza, A. Pfaltz, Chem. Eur. J. 2009, 15, 2266 – 2269.
[16] A. Baeza, A. Pfaltz, Chem. Eur. J. 2010, 16, 2036 – 2039.
[17] T. Yoshinari, K. Ohmori, M. G. Schrems, A. Pfaltz, K. Suzuki,
Angew. Chem. 2010, 122, 893 – 897; Angew. Chem. Int. Ed. 2010,
49, 881 – 885.
[18] C. Valla, A. Baeza, F. Menges, A. Pfaltz, Synlett 2008, 3167 –
3171.
[19] For C,N ligands containing a N-heterocyclic carbene and an
oxazoline unit, see: a) Y. Zhu, Y. Fan, K. Burgess, J. Am. Chem.
Soc. 2010, 132, 6249 – 6253; b) Y. Fan, X. Cui, K. Burgess, M. B.
Angew. Chem. Int. Ed. 2011, 50, 9598 –9601
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 9601