Chiral mixed secondary phosphine-oxide-phosphines: High-performing and easily accessible ligands for asymmetric hydrogenation

Article abstract of DOI:10.1002/anie.201002225

P&O: Combining secondary phosphine oxides (SPOs) with phosphines leads to highly effective chiral bidentate ligands for transition-metal-based catalysts. JoSPOphos and TerSPOphos are readily accessible from inexpensive starting materials. The steric and electronic properties of these modular ligands can be easily tuned. In the asymmetric hydrogenation of functionalized alkenes, their rhodium complexes reacted to give enantioselectivities of up to 99% ee and turnover frequencies of up to 20000 h^-1 Chemical equation Presented

Full text of DOI:10.1002/anie.201002225

Angewandte
Chemie
DOI: 10.1002/anie.201002225
Ligand Design
Chiral Mixed Secondary Phosphine-Oxide–Phosphines: High-
Performing and Easily Accessible Ligands for Asymmetric
Hydrogenation**
Heidi Landert, Felix Spindler, Adrian Wyss, Hans-Ulrich Blaser, Benoꢀt Pugin,*
Yann Ribourduoille, Bjꢁrn Gschwend, Balamurugan Ramalingam, and Andreas Pfaltz*
Dedicated to Professor Albert Eschenmoser on the occasion of his 85th birthday
Chiral diphosphines are the most frequently used ligands in
asymmetric catalysis.[1] In contrast, chiral secondary phos-
phine oxides (SPOs) are little explored as ligands. While their
chemical and physical properties are well known, their use in
asymmetric catalysis is still in its infancy.[2]
defined complexes. To avoid cumbersome resolution proce-
dures[2c,6,7] we used either a chiral backbone or a chiral
substituent, so that the chiral SPO unit could be built up in
diastereoselective reactions (Scheme 1).
SPOs are stable molecules which exist in equilibrium
between two tautomeric forms:[3] the preferred pentavalent
phosphine oxide and the trivalent phosphinous acid. When
two different substituents are attached to the phosphorus
atom, a configurationally stable, P-chiral group results which
can coordinate to metals either through the phosphorus atom
or through the oxygen atom.
To date, only a few examples of asymmetric catalytic
reactions with chiral SPOs have been described.[2] Ph-
(tBu)P(O)H, a monodentate P-chiral SPO gave approxi-
mately 80% ee in the palladium-catalyzed allylic alkylation,[4]
while over 90% ee was obtained with P-chiral diamino
phosphine oxides.[5] In asymmetric hydrogenation, Rh and Ir
complexes of monodentate chiral SPO ligands gave only
moderately active and selective catalysts (ee values up to
85%).[2c,6]
Scheme 1. Concept and generic structures of SPO–P ligands.
We thought that these somewhat disappointing results
might be due to an insufficient affinity of SPOs for Rh, Ir, or
Ru centers, the typical metals used in asymmetric catalytic
hydrogenations. Our idea was therefore to combine an SPO
with a PR2 substituent which should not only lead to stronger
coordination to the metal center but also should give better
Herein we present results for selected members of two
SPO–P ligand families based on a chiral ferrocenyl backbone
and a menthyl substituent, respectively (Scheme 1). The first
approach leads to ligands structurally similar to the well
known Josiphos[8] (therefore called JoSPOphos) while the
second gives menthyl derivatives (called TerSPOphos since
other terpene moieties are feasible). Both ligand families are
modular, allowing the ligand properties to be tuned by the
choice of the R and R’ groups. First tests showed that these
novel ligands give excellent enantioselectivities and high
turnover numbers for the hydrogenation of a variety of
functionalized alkenes.
[*] H. Landert, Dr. F. Spindler, A. Wyss, Dr. H.-U. Blaser, Dr. B. Pugin
Solvias AG
P.O. Box, 4002 Basel (Switzerland)
Fax: (+41)616-866-311
E-mail: benoit.pugin@solvias.com
Homepage: www.solvias.com
Two routes were developed for the preparation of the
JoSPOphos ligands (Scheme 2). In route 1 the phosphine
group was introduced before the SPO group, starting from
(R)-N,N-dimethyl-1-[(S)-2-bromoferrocenyl]ethylamine (3),
obtained by lithiation/bromination of the (R)-Ugi amine.[9]
The dimethylamino group was exchanged for the desired PR2
group to give ferrocenyl phosphine bromides 4 with retention
of configuration. JoSPOphos ligands 1a–d were obtained by
treating 4a or 4b with BuLi at low temperature, subsequent
addition of the chosen dichlorophosphine, and finally hydrol-
ysis with water. Since surprisingly the SPO moiety withstood
Dr. Y. Ribourduoille, B. Gschwend, Dr. B. Ramalingam, Prof. A. Pfaltz
Department of Chemistry, University of Basel, Organic Chemistry
St. Johanns-Ring 19, 4056 Basel (Switzerland)
Fax: (+41)612-671-103
E-mail: andreas.pfaltz@unibas.ch
[**] We thank Raphael Aardoom for his experimental contributions.
Financial support from the Federal Commission for Technology and
Innovation (KTI) is gratefully acknowledged.
Supporting information for this article (detailed procedures for the
ligand syntheses and for the hydrogenation experiments ) is
available on the WWW under http://dx.doi.org/10.1002/anie.
201002225.
Angew. Chem. Int. Ed. 2010, 49, 6873 –6876
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6873

Communications
comparison of ligands 1a and 1a’ which differ only in the
absolute configuration of the SPO moiety. The Rh complex of
1a (prepared with [Rh(nbd)2]BF4) gave a 31P NMR spectrum
with two doublets of doublets at d = 113.1 ppm and d =
58.4 ppm (both JRhP = 165 Hz, JPP = 35 Hz), showing that the
Rh center is coordinated to both phosphorus atoms. With
ligand 1a’ a P,P complex giving rise to two doublets of
doublets at d = 132.3 ppm and d = 57.9 ppm (both JRhP
=
168 Hz, JPP = 39 Hz) as well as a P,O complex (PR2: doublet
of doublet at d = 40.5 ppm; JRhP = 174 Hz, JPP = 2 Hz; SPO:
doublet at d = 66.2 ppm; JPP = 2 Hz) was detected. We assume
that the different behavior of 1a and 1a’ is due to the steric
interactions of the tert-butyl group of the SPO moiety and the
ferrocenyl backbone (see Scheme 3).
Scheme 3. Coordination modes of ligands 1a and 1a’ with Rh.
Scheme 2. Synthesis and absolute configuration of the JoSPOphos
ligands 1a–d. Reagents and conditions: a) 1. sBuLi, Et2O, 2. (BrF2C)2
or (BrCl2C)2; b) HPR2, AcOH; c) 1. nBuLi, TBME; 2. Cl2PR’; 3. hydroly-
sis; d) 1. sBuLi, Et2O; 2. Cl2PR’; 3. hydrolysis; e) HPR2, AcOH. TBME=
tert-butyl methyl ether.
The TerSPOphos ligands 2a–c were prepared starting
from 2-bromoiodobenzene (6) which was metallated and then
treated with a chlorophosphine to give 7 (Scheme 4).
Lithiation of 7 and reaction with dichloro[(À)-menthyl]phos-
phine[12] yielded the chlorophosphine intermediates which
were hydrolyzed with 0.1m NaOH to give the SPO–P ligands
in good yields and with diastereomeric ratios of around 10:1.
The pure ligands 2a–c were obtained by recrystallization or
column chromatography. A single-crystal X-ray analysis of a
ZnBr2 complex of ligand 2b allowed the absolute config-
uration of the major epimer of 2a–c to be assigned as
(SSPO).[11] Also in this case, the zinc ion coordinates to the
oxygen atom.
heating in acetic acid, the “reverse” procedure, that is, first the
introduction of the SPO group to give 5, and subsequent
exchange of the NMe2 moiety, was another option (route 2).
In this way the lithiation/bromination step and the isolation of
3 could be avoided.
In both variants, the JoSPOphos ligands 1 were obtained
in good yields with a diastereomeric ratio of typically around
10:1 and purified either by crystallization or by chromatog-
raphy on silica gel. While the stereogenic carbon atom (R-
configuration) and the ferrocene ring (Sp-configuration) had
the same absolute configuration in both routes (controlled by
the absolute configuration of the Ugi amine), the configu-
ration of the SPO group depended on the nature of R’. With
R’ = Ph (1b and 1d), both routes yielded preferentially RSPO
.
In contrast, for R’ = tBu (1a and 1c), route 1 gave RSPO
whereas route 2 gave mainly SSPO isomers allowing the
controlled preparation of either epimer. Variation of the
hydrolysis conditions[10] in route 2 also gave access to a small
sample of the SSPO ligand 1b’.
The absolute configuration of all the JoSPOphos ligands,
as well as the coordination mode (P versus O coordination)
were determined by single-crystal X-ray analysis of a rhodium
norbornadiene (nbd) tetrafluoroborate complex of ligand
(SSPO)-1a’ and of a ZnBr2 complex of ligand (RSPO)-1b.[11] As
expected, the oxophilic zinc ion is coordinated to the oxygen
atom in the phosphine oxide but the coordination behavior of
rhodium is more subtle. Of particular interest was the
Scheme 4. Synthesis of the TerSPOphos ligands 2a–c. Reagents and
conditions: a) 1. iPrMgCl, THF; 2. ClPR2; b) 1. nBuLi, THF; 2. (l-men-
thyl)PCl2; 3. hydrolysis.
The ligands were tested in hydrogenation experiments,
using standard substrates (Scheme 5) to show the scope and
limitations for their synthetic applications. Most tests were
carried out with a Symyx HTS robot which uses plates with
6874 www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 6873 –6876

Angewandte
Chemie
Similar results were obtained for the TerSPOphos ligands.
Also in these cases most substrates are hydrogenated with
ee values in the range of 94% to over 99%. The fact that
ligands 2a and 2b with PAr2 groups give similar enantiose-
lectivities to 2c (R = Cy) indicates that the electronic nature
of the phosphine group hardly affects the ee value.
A few reactions with MAA and DMI were carried out in
50 mL reactors with s/c = 200–1000 at a hydrogen pressure of
1 bar. For all the ligands, the reactions were usually complete
within 5 min (implying turnover frequencies (TOF) in the
range of 2000–20000 hÀ1), showing that both types of ligands
yield very active catalysts for disubstituted alkenes.
Scheme 5. Test substrates for hydrogenation.
Ligand families 1 and 2 were also tested for the
96 vials (for reaction conditions see Table 1). Selected hydro-
genations at higher substrate to catalyst ratios (s/c) were
carried out in 10–50 mL reactors.
ruthenium- and rhodium-catalyzed hydrogenation of a
series of a-and b-ketoesters. The results indicate that the
hydrogenation of such substrates with SPO–P ligands is not
straightforward and that the struc-
ture/selectivity match is quite
Table 1: Enantioselectivities obtained with JoSPOphos and TerSPOphos in rhodium-catalyzed hydro-
narrow. The best results were
obtained with JoSPOphos ligand
1a (R’ = tBu), for the ruthenium-
catalyzed hydrogenation of EOP
(92% ee) and the rhodium-cata-
lyzed hydrogenation of KPL
(89% ee). On the positive side, a
catalyst formed in situ from 1a and
[{RuCl2(p-cymene)}2] was highly
active and productive, giving com-
plete conversion within less than
17 h for the hydrogenation of EOP
at a s/c of 5000.
In conclusion, the combination
of an SPO and a phosphine group
leads to ligands which form highly
effective hydrogenation catalysts.
The use of a chiral backbone or a
chiral substituent at the SPO center
genations of six functionalized alkenes (see Scheme 5).[a]
Entry Ligand R (P) R’ (SPO) Config. SPO MAC
AC
MAA
DMI
Z-EAC E-EAC
1
2
3
4
5
6
7
8
9
1a
1a’
1b
1b’
1c
1c’
1d
2a
2b
2c
Ph
Ph
tBu
tBu
Ph
R
S
R
S
R
S
R
S
S
S
+38[b] +98[b] +71[c] +95[c,f] +25[c,f] À96[b,f])
À97[b] À99[b] À97[b] À98[b] +61[c] +94[c]
tBu
tBu
tBu
tBu
Ph
l-Men Ph
l-Men 4-Tol
l-Men Cy
+90[b] +98[b] +99[d] +94[c] À98[c]
À99[c]
+70[c]
À58[b]
Ph
–
–
–
À84[c]
À1[c]
tBu
tBu
Ph
+75[b] À11[b] +94[c] +19[c] À51[c]
À99[b] À99[b] À98[b] À93[c] +76[b] +76[b]
+85[b] +98[b] +93[c] +99[b] À72[b]
À90[b]
–
–
À95[e]
–
–
–
À96[b] À99[b] À98[b] À98[b] +68[b] +94[c]
10
À94[b] À98[b] À96[b]
–
–
–
[a] ee values ꢀ90% are in bold. The reactions were performed at room temperature, 1 bar H2 pressure,
with a s/c of 100 giving complete conversions in less than 2 h. The catalysts were prepared in situ by
mixing 1.1 equivalent ligand with 1 equivalent of a rhodiumprecursor. [b] Rh precursor=[Rh(nbd)2]BF4;
solvent=EtOH. [c] Rh precursor=[Rh(nbd)2]BF4; solvent=THF. [d] Rh precursor=[Rh(cod)Cl]2;
solvent=1,2-dichloroethane. [e] As [b] but s/c 200. [f] Reaction time 14 h. nbd=norbornadiene,
cod=1,5-cyclooctadiene.
Most experiments were performed with six functionalized
alkenes and selected results are shown in Table 1 for rhodium
JoSPOphos (entries 1–7) and rhodium TerSPOphos com-
plexes (entries 8–10). Both ligand families show excellent
catalytic performance and many catalysts gave high enantio-
selectivities with several substrates. Notably, ligand 1b gave
ee values in the range of 90% to over 99% with all substrates,
which is quite exceptional. Of special interest is the fact that
E- and Z-EAC afford products with the same absolute
configuration, allowing the use of E/Z-mixtures.[13] Ligand 1b
with a phenyl group on the SPO moiety and tBu groups on the
phosphine outperforms ligand 1a where the phenyl and tBu
groups are transposed. Ligand 1b also outperforms, ligands
1c and 1d which have only tBu or Ph groups, respectively. The
absolute configuration of the phosphorous center seems to
dominate the sense of induction: in almost all cases tested to
date, the product absolute configuration changes when going
from RSPO to SSPO ligands. The influence of the other
stereogenic units is less predictable, but it appears that for
R’ = tBu the (R,Sp,SSPO) isomer (e.g. 1c’) is superior to the
RSPO isomer (e.g. 1c) whereas for R’ = Ph, the reverse
behavior is observed
allows easy access to this modular class of ligands. We have
found that SPO–P ligands can coordinate to metal centers
either through both phosphorus atoms or through one
phosphorus and an oxygen atom. Although at present we do
not have any experimental evidence, we assume that the P,P
complex rather than the P,O complex is the active catalyst.
Our results show that the corresponding Rh and Ru
complexes exhibit excellent activities and enantioselectivities
in the hydrogenation of functionalized alkenes and moderate
enantioselectivity for ketoesters. Thus the combination of a
SPO and a phosphine unit in a chelating ligand appears to be a
promising approach to generate high-performing ligands.
Preliminary work has shown that this concept can be
extended to analogues of 1 with other chiral backbones,
such as biaryls, or analogues of 2 with different aryl systems or
other terpenes as the chiral moiety.
Received: April 15, 2010
Published online: August 16, 2010
Keywords: alkenes · asymmetric catalysis · hydrogenation ·
.
phosphane oxides · rhodium
Angew. Chem. Int. Ed. 2010, 49, 6873 –6876
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
6875

Communications
Sieron, W. Wieczorek, T. Jeraphagnon, I. W. C. E. Arends, U.
Hanafeld, A. J. Minnaard, Synthesis 2009, 2061 – 2065.
[8] a) A. Togni, C. Breutel, A. Schnyder, F. Spindler, H. Landert, A.
Tijani, J. Am. Chem. Soc. 1994, 116, 4062 – 4066; b) H. U. Blaser,
W. Brieden, B. Pugin, F. Spindler, M. Studer, A. Togni, Top.
Catal. 2002, 19, 3 – 16.
[9] a) J. W. Han, N. Tokunaga, T. Hayashi, Helv. Chim. Acta 2002,
85, 3848 – 3854; b) P. Barbaro, C. Bianchini, G. Giambastiani, A.
Togni, Tetrahedron Lett. 2003, 44, 8279 – 8283.
[10] Hydrolysis was initially performed by adding water to the
monochlorophosphine intermediates. We then realized that
whereas hydrolysis conditions have little influence in route 1,
they can strongly affect the diastereomeric ratio of the SPOs in
route 2. Best results were obtained when the reaction mixture
was poured onto a stirred mixture of NEt3/water 1:10.
[11] X-Ray structures will be described in more details in a forth-
coming Full Paper.
[12] Dichloro[(À)- or (+)-menthyl]phosphine was prepared as de-
scribed in the literature: a) M. Minato, T. Kaneko, S. Masauji, T.
Ito, J. Organomet. Chem. 2006, 691, 2483 – 2488; b) A. Hinke, W.
Kuchen, Phosphorus Sulfur 1983, 15, 93 – 98. In our hands, the
product was always contaminated with up to 10% of a side
product, which we assumed to be the corresponding neomenthyl
analogue.
[1] M. Thommen, H.-U. Blaser, Phosphorus Ligands in Asymmetric
Catalysis (Ed.: A. Bꢀrner) Wiley-VCH, Weinheim, 2008,
pp. 1457 – 1471.
[2] see literature cited in a) N. V. Dubrovina, A. Bꢀrner, Angew.
Chem. 2004, 116, 6007 – 6010; Angew. Chem. Int. Ed. 2004, 43,
5883 – 5886; b) L. Ackermann, Synthesis 2006, 1557 – 1571; c) X.-
b. Jiang, A. J. Minnaard, B. Hessen, B. L. Feringa, A. L. L.
Duchateau, J. G. O. Andrien, J. A. F. Boogers, J. G. de Vries,
Org. Lett. 2003, 5, 1503 – 1506.
[3] M. J. Gallagher, The Chemistry of the Organophosphorus
Compounds, Vol 2 (Ed.: F. R. Hartley), Wiley, New York, 1992.
[4] W.-M. Dai, K. K. Y. Yeung, W. H. Leung, R. K. Haynes,
Tetrahedron: Asymmetry 2003, 14, 2821 – 2826.
[5] a) T. Nemoto, T. Matsumoto, T. Masuda, T. Hitomi, K. Hatano,
Y. Hamada, J. Am. Chem. Soc. 2004, 126, 3690 – 3691; b) T.
Nemoto, T. Masuda, T. Matsumoto, Y. Hamada, J. Org. Chem.
2005, 70, 7172 – 7178.
[6] a) X.-b. Jiang„ M. van den Berg, A. J. Minnaard, B. L. Feringa,
J. G. de Vries, Tetrahedron: Asymmetry 2004, 15, 2223 – 2229;
b) A. Galland, C. Dobrota, M. Toffano, J.-C. Fiaud, Tetrahedron:
Asymmetry 2006, 17, 2354 – 2357.
[7] see for example, a) O. I. Kolodiazhnyi, E. V. Gryshkun, N. V.
Andrushko, M. Freytag, P. G. Jones, R. Schmutzler, Tetrahedron:
Asymmetry 2003, 14, 181 – 183; b) A. Leyris, J. Bigeault, D. Nuel,
L. Giordano, G. Buono, Tetrahedron Lett. 2007, 48, 5247 – 5250;
c) J. Holt, A. M. Maj, E. P. Schudde, K. M. Pietrusiewicz, L.
[13] H.-J. Drexler, J. You, S. Zhang, C. Fischer, W. Baumann, A.
Spannenberg, D. Heller, Org. Process Res. Dev. 2003, 7, 355 –
361.
6876 www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 6873 –6876