New amidophosphine-phosphinites (tLANOPs) as chiral ligands for asymmetric hydrogenation reactions

Article abstract of DOI:10.1016/S0957-4166(98)00423-6

The new amidophosphine-phosphinite (AMPP) ligands 4a-g (called tLANOP ligands) derived from the chiral hydroxy amide (R)- or (S)-2-hydroxy-3,3,N- trimethylbutyramide have been prepared in 48-83% yield. The crystal structures of the square planar complexes

Full text of DOI:10.1016/S0957-4166(98)00423-6

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 9 (1998) 4043–4054
New amidophosphine-phosphinites (tLANOPs) as chiral ligands
for asymmetric hydrogenation reactions
Emil A. Broger, Wolfgang Burkart, Michael Hennig, Michelangelo Scalone ∗ and
Rudolf Schmid
Pharmaceuticals Division, Process Research and Structural Research, F. Hoffmann-La Roche AG, CH-4070 Basel,
Switzerland
Received 5 October 1998; accepted 20 October 1998
Abstract
The new amidophosphine-phosphinite (AMPP) ligands 4a–g (called tLANOP ligands) derived from the chiral
hydroxy amide (R)- or (S)-2-hydroxy-3,3,N-trimethylbutyramide have been prepared in 48–83% yield. The crystal
structures of the square planar complexes [(SP-4-3)-Pd((R)-dmphea)((S)-4a)]BF4 and [Rh((R)-4a)(COD)]BF4 have
allowed the absolute configurations of the ligands to be assigned. In both complexes the 7-membered chelating ring
of 4a has virtually the same twist-boat conformation. With this class of ligands the rhodium catalyzed asymmetric
hydrogenation of 4-oxoisophorone enol acetate gave (S)-phorenol acetate in up to 71% ee. The iridium catalyzed
asymmetric hydrogenation of the cyclic iminium salts 16a and 16b afforded after work-up the corresponding cyclic
secondary amine (S)-17 in up to 86% ee, when bulky groups were present on the phenyl substituents on the two
phosphorus atoms. © 1998 Elsevier Science Ltd. All rights reserved.
1. Introduction
Recently, amidophosphine-phosphinite (AMPP) ligands of formula 1–3 have been reported to give
efficient asymmetric catalysts for the hydrogenation of α-functionalized ketones1,2 or allylic alcohols.3
These ligands are particularly interesting from a practical point of view because they can be prepared
in very few steps from easily available natural products. Ligands of type 3 are derived from oxo-
prolinol and contain the carbonyl group in exocyclic position with respect to the 7-membered chelate
ring formed after coordination to a transition metal. They have so far been the most effective in
the asymmetric hydrogenation of, for example, dihydro-4,4-dimethyl-2,3-furandione, affording (S)-
pantolactone of 95–99% ee.4 The ligands of type 1 and 2, which have the carbonyl group endocyclic after
∗
Corresponding author. E-mail: michelangelo.scalone@roche.com
0957-4166/98/$ - see front matter © 1998 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(98)00423-6

4044
E. A. Broger et al. / Tetrahedron: Asymmetry 9 (1998) 4043–4054
coordination to the metal, have afforded (S)-pantolactone in 81–90% ee. The lower enantioselectivity was
ascribed to the lower rigidity of the backbone.1,4
We reasoned that the 7-membered chelate ring could be made more rigid by replacing the methyl
or the phenyl substituent of 1 or 2 by a t-butyl group. Indeed, molecular modelling calculations on
cationic rhodium complexes supported this hypothesis. Moreover, they indicated that, after coordination
to a metal, the chelate ring of 3a and 4a as well as 3b and 4b would have very similar preferred twist-
boat conformations, thus suggesting a similar rigidity of the backbone in these two ligand families.5
Consequently, we have synthesized the new tLANOP ligands 4a–g and tested them in two asymmetric
hydrogenation reactions.
2. Results and discussion
2.1. Synthesis and characterization of the tLANOP ligands
The starting hydroxy amides (R)- and (S)-6 were prepared by asymmetric hydrogenation of the
ketoamide 5 in the presence of chiral ruthenium catalysts (Scheme 1).6 After crystallization they were
isolated in 99% ee. Their absolute configuration was provisionally assigned by applying the general sense
of the asymmetric induction observed in the BINAP–Ru catalyzed hydrogenation of α-functionalized
ketones ((R)-BINAP gives (R)-alcohols)7,8 and was confirmed by determination of the crystal structure
of the two complexes (R,S)-10 and (R)-13a (vide infra). Then, the reaction of the dilithium salt of 6
with 2.1 molar equiv. of a chlorophosphine gave the tLANOP ligands 4a–g in 48–83% yield. These
phosphinylations proceeded selectively when BuLi was used in combination with a small amount (5%)
of a secondary amine (e.g. diisopropylamine). This avoided the intermediate precipitation of an insoluble
species, probably a monolithium salt of 6, and suppressed almost completely the concomitant formation
of by-products such as the diphosphine 7 and the n-butyl phosphine 8. If a stoichiometric amount
of diisopropylamine was employed, the phosphinylation was incomplete and substantial amounts of
(iPr2N)PPh2 were formed as a by-product. No formation of the desired ligands 4 was observed when
6 was treated in analogy to the procedures reported for 1–31,2,4 with a chlorophosphine in the presence
of a Et3N or DBU. Specifically, with Ph2PCl, rapid monophosphinylation at the oxygen atom was
observed at rt, but prolonged reaction times led only to formation of unidentified by-products. Under
the same conditions, no reaction at all was observed with Cy2PCl. Attempts to react the lithium salt of
the monophosphinylated species with Cy2PCl led only to complex mixtures of products.
The ligands 4a–g are white, air-sensitive solids which can be stored without decomposition for months
1
at rt under an argon atmosphere. They have been characterized by H, 31P NMR and IR spectroscopy,
microanalysis and mass spectrometry. Their NMR spectra are similar to those reported earlier for

E. A. Broger et al. / Tetrahedron: Asymmetry 9 (1998) 4043–4054
4045
Scheme 1.
Table 1
Yields and 31P{1H} NMR data for the tLANOP ligands 4a–g
the AMPP ligands 1–3.4 The two 31P resonances were assigned accordingly: the P(N) and the P(O)
resonances appeared at 40–65 ppm and 110–139 ppm, respectively, with the exception of (2-furyl)-
tLANOP 4c, where the signals are shifted strongly upfield (Table 1).9
Since alcohols are employed frequently as solvents in asymmetric hydrogenations, we checked
1
the stability of two tLANOP ligands in deuteriomethanol with H and 31P spectroscopy. For 4a no
measurable decomposition was observed after 3 days at rt, whereas subsequent addition of water led to
ca. 10% of hydrolysis after one more week. Under the same conditions the ligand 4f, which contains the
bulky 3,5-tBu2-phenyl substituents, did not show any decomposition at all. As a comparison, oxoProNOP
3a was completely decomposed after 24 h in deuteromethanol.
In addition, the stereochemical integrity of the tLANOP ligands was checked in order to rule out
any partial racemization of the hydroxy amides (R)- and (S)-6 under the strongly basic phosphinylation
conditions. For this purpose, (S)-4a, (R)-4b and (R)-4f, were hydrolyzed with methanolic HCl and 6
was recovered. GC analysis showed that the ee of 6 was unchanged with respect to the starting hydroxy
amides used.
2.2. Synthesis of transition metal complexes
In order to establish unambiguously the absolute configuration of the tLANOP ligands, and in order to
probe their chelating capability and the conformational properties of the chelates, the palladium complex
(R,S)-10, the rhodium complex (R)-13a and the iridium complex (S)-13b were prepared.

4046
E. A. Broger et al. / Tetrahedron: Asymmetry 9 (1998) 4043–4054
The reaction of (S)-4a with [Pd((R)-dmphea)(NCMe)2]BF4 (R)-9 (dmphea=η2-Me2NCHMe-CC5H4)
in dichloromethane was carried out in analogy to published methods.10,11 If the reaction time was short
(10 min at rt), the kinetic product was isolated in high yield (Scheme 2). X-Ray analysis of crystals
grown at −30°C showed that this complex had the (SP-4-3) geometry of (R,S)-10. At longer reaction
times, rearrangement of (R,S)-10 to (SP-4-4)-(R,S)-11 was observed. The latter, however, decomposed
slowly in solution and could not be isolated.
Scheme 2.
The reaction of (R)-4a with 12a or of (S)-4a with 12b12 afforded the cationic rhodium cyclooctadiene
complex (R)-13a and the analogous iridium complex (S)-13b, respectively, in practically quantitative
yield. A detailed investigation describing mono- and dinuclear rhodium complexes of 4a will be
published soon.13
2.3. Crystal structure of (R,S)-10 and (R)-13a
ORTEP plots of the cationic components in (R,S)-10 and (R)-13a are shown in Fig. 1. Table 4
(see Experimental section) shows experimental parameters for the structure determination. Both crystal
structures confirmed the provisionally assigned absolute configuration of 4a and indirectly that of 6.
In both complexes the geometry around the metal atoms is distorted square-planar. In (R,S)-10 the
σ-coordinated carbon atom is located out of the palladium coordination plane by 0.493 Å. In (R)-13a, if
one takes the midpoints of the two coordinated olefinic bonds of cyclooctadiene as virtual monodentate
ligands, the plane defined by them and rhodium forms an angle of 15.8° with the plane containing the two
phosphorus atoms and rhodium.14 In spite of the different electronic and steric environments, however,
the topology of the seven-membered chelate ring and the spatial arrangement of the phenyl groups of
the PPh2 moieties in 4a in both (R,S)-10 and mirror-inverted (R)-13a are practically identical (cf. data
in Fig. 1). The twist-boat conformation of δ ((R,S)-10) or λ ((R)-13a) helicity is characterized by the
quasi-planarity of the amido function bound at P(2) and by the location of the t-butyl group in pseudo-
equatorial orientation. The phenyl substituents at the phosphorus atoms are in an alternating fashion
pseudo-equatorially and pseudo-axially oriented. In both complexes the metal–P(N) is longer than the
metal–P(O) bond, a phenomenon that seems to be general in transition metal complexes of amido-4 and
aminophosphine-phosphinite ligands.15

E. A. Broger et al. / Tetrahedron: Asymmetry 9 (1998) 4043–4054
4047
Figure 1. Molecular structures (30% thermal ellipsoids) with atom numbering scheme, selected bond lengths (Å), an-
gles (°) and torsional angles (°). (R,S)-10: Pd(1)–P(1) 2.248(2), Pd(1)–P(2) 2.426(3), Pd(1)–N(35) 2.183(8), Pd(1)–C(38)
2.041(10), P(1)–O(1) 1.617(7), P(2)–N(4) 1.736(9); P(1)–Pd(1)–P(2) 94.53(9), C(38)–Pd(1)–N(35) 79.2(4), C(38)–Pd(1)–P(1)
89.5(3); C(2)–C(3)–N(4)–P(2) 5.3, O(5)–C(3)–N(4)–P(2) 179.9, Pd(1)–P(1)–O(1)–C(2) −32.5. (R)-13a: Rh(1)–P(1) 2.256(6),
Rh(1)–P(2) 2.304(6), P(1)–O(1) 1.635(15), P(2)–N(4) 1.711(19), Rh–C(olefin) 2.18 and 2.20; P(1)–Rh(1)–P(2) 93.4(2);
C(2)–C(3)–N(4)–P(2) 8.6, O(5)–C(3)–N(4)–P(2) 171.9, Rh(1)–P(1)–O(1)–C(2) 34.6
2.4. Asymmetric hydrogenations
The tLANOP ligands were tested in the asymmetric hydrogenation of: (a) 4-oxoisophorone enol
acetate 14 to (S)-phorenol acetate (S)-15, an intermediate in the synthesis of the natural pigment
zeaxanthin16; and (b) the cyclic iminium salts 16a,b to the octahydro-isoquinoline derivative (S)-17,
the so-called (S)-octabase,17 an intermediate in the synthesis of the antitussivum dextromethorphan.
In the rhodium catalyzed asymmetric hydrogenation of 14, the ee values achieved with the ligands of
type 4 are at best moderate (Table 2, entries 1 and 2). In contrast, with one ligand of type 3 (3b, entry
5) 95% ee is reached. The latter result comes close to the so far best ee of 98% ee achieved in this
hydrogenation with a DuPHOS-type diphosphine.18
In the iridium catalyzed asymmetric hydrogenation of the cyclic iminium salts 16a and 16b, increase
of the steric encumbrance of the aryl substituents at the phosphorus atoms in the ligands 4a and 4c–g
led to a remarkable improvement of the ee of 17, e.g. from 4% with 4a (Table 3, entry 7) to 78% with
4g (entry 13) for the fluoroborate salt 16a or even to 86% ee with 4f (entry 18) for the bisulphate salt
16b.19 Unfortunately, with these ligands the chemoselectivity was modest and substantial amounts of
stereoisomeric decahydro-isoquinoline derivatives were formed as by-products. The two ligands of the
oxoProNOP type tested (entries 14 and 15) gave higher chemoselectivity but low ees. With ligands of
both type 3 and 4 the presence of cyclohexyl substituents on the phosphorus atoms brought about the
highest chemoselectivity (96%) but, unfortunately, very low ees (entries 8, 15 and 17). Remarkably,
these ligands with e.g. (R)-configuration afforded (S)-17, whereas those with aryl substituents afforded
(R)-17.
In summary, we have shown that tLANOP ligands of type 4 can be readily prepared in high yields.
Preliminary results indicate that they can be very efficient in asymmetric hydrogenation reactions. Their
potential for other asymmetric catalytic reactions still remains to be explored.

4048
E. A. Broger et al. / Tetrahedron: Asymmetry 9 (1998) 4043–4054
Table 2
Asymmetric hydrogenation of 4-oxoisophorone enol acetate 14a)
3. Experimental section
All manipulations of oxygen- and moisture-sensitive materials were conducted under an argon at-
mosphere. All solvents were distilled under argon before use. Hydrogen and argon were of 99.9999%
purity. The complexes, [Ru(OAc)2((R)- or (S)-BIPHEMP)] BIPHEMP=(6,60-dimethylbiphenyl-2,20 -
diyl)bis(diphenylphosphine),20 were prepared according to reported procedures.6 The chlorophosphines
were purchased or prepared according to known methods;21 their high purity was essential for obtaining
the tLANOP ligands in high purity. The AMPP ligands 3a–c were prepared according to a known general
procedure.1 Before use, basic alumina (Camag) was dried at 200°C under vacuum and stored under
argon. Nuclear magnetic resonance spectra (NMR) were measured on a Bruker 250E (1H 250 MHz and
31P 101.26 MHz) spectrometer in CDCl3 using TMS (1H) as an internal standard and 85% H3PO4 (31P)
as an external standard. Optical rotations were measured on a Perkin–Elmer 241 spectrometer. Melting
points were determined on a Büchi 510 and are uncorrected.
3.1. (R)-2-Hydroxy-3,3,N-trimethyl-butyramide (R)-6
In a glove box (O2 content<1 ppm) the catalyst solution was prepared by dissolving [Ru(OAc)2((R)-
BIPHEMP)] (385 mg, 0.50 mmol) in a 0.025 molar methanolic HCl solution (40 ml) and dichloro-
methane (40 ml) and stirring at rt for 1.5 h. This solution was transferred into a catalyst addition device.
A 2 l Hastelloy autoclave was charged with 522 (71.6 g, 0.50 mol) and methanol (310 ml), then sealed,
purged with argon and finally with hydrogen. The catalyst solution was allowed to flow into the autoclave
with a slight overpressure and the hydrogenation was carried out at 60°C with a constant pressure of 60
bar. After 20 h the reaction mixture was drained off and evaporated. Crystallization of the crude residue
from diisopropyl ether (250 ml) and cyclohexane (275 ml) at 5°C afforded (R)-6 (43.9 g, 60.5%) as a
white crystalline solid, mp 69–71°C. 1H NMR (250 MHz, CDCl3, 20°C): δ 0.99 (s, 9H, tBu), 2.84 (d,
20
3H, NCH3, J=5.0), 3.08 (d, 1H, OH, J=5.6), 3.71 (d, 1H, CHOH, J=5.6), 6.32 (b, 1H, NH). [α]D
+61.0 (c 1, CHCl3), ee=98.7% (determined by GC on an OV-61/p-DiMe-β-cyclodextrin chiral column).

E. A. Broger et al. / Tetrahedron: Asymmetry 9 (1998) 4043–4054
4049
Table 3
Asymmetric hydrogenation of 16·HXa)
MS (EI) m/z 146 (M+, 2%), 112 (M−H2O−CH3, 1%), 89 (M−C4H8, 100%). IR (KBr, ν, cm−1): 1651
(b, C_O). Anal. calcd for C7H15NO2: C, 57.90; H, 10.41; N, 9.65. Found: C, 57.97; H, 10.60; N, 9.61.
3.2. (S)-2-Hydroxy-3,3,N-trimethyl-butyramide (S)-6
The asymmetric hydrogenation of 5 (55 g, 0.384 mol) in the presence of [Ru(OAc)2((S)-
BIPHEMP)]/HCl (296 mg, 0.384 mmol) afforded, after crystallization, (S)-6 (34.4 g, 62%) with
20
mp 60–62°C, [α]D −59.9 (c 1, CHCl3), ee=99.2%. Anal. found for C7H15NO2: C, 58.02; H, 10.49; N,
9.59.
3.3. Preparation of tLANOP ligands 4a–g
3.3.1. (R)-Ph-tLANOP (R)-4a
In a typical experiment, (R)-6 (3.88 g, 26.7 mmol) was dissolved in a 500 ml Schlenk tube containing
THF (200 ml) and diisopropylamine (0.5 ml) and reacted at −78°C with n-butyllithium in hexane (28.5
ml, 53.5 mmol). After 1 h a solution of chlorodiphenylphosphine (11.8 g, 53.5 mmol) in THF (30 ml)

4050
E. A. Broger et al. / Tetrahedron: Asymmetry 9 (1998) 4043–4054
was added during 30 min and the mixture was stirred for 1 h. Then the cold bath was removed and the
yellow clear solution evaporated. The residue was dried under high vacuum, taken up in diethyl ether
and the suspension filtered under argon through basic alumina (60 g). The filter cake washed with diethyl
ether (100 ml). The combined filtrates were evaporated and the residue dried under high vacuum. The
resulting yellow oil was cooled to −180°C with liquid nitrogen, cold pentane (50 ml) was added and the
mixture treated in an ultrasonic bath for 5 min, during which time white crystals formed. After 24 h at
5°C the supernatant was removed, the crystals were rinsed with pentane (2×20 ml, 0°C) and dried under
20
high vacuum to afford (R)-4a (8.41 g, 61%) as a white crystalline solid, mp 95–96°C, [α]D +37.5 (c
1, CHCl3). 1H NMR (250 MHz, CDCl3, 20°C): δ 1.06 (s, 9H, tBu), 2.73 (s, 3H, NCH3), 5.86 (dd, 1H,
3JPH=4JPH=9.0, CHO), 7.15–7.60 (m, 20H). MS (EI) m/z 514.3 (M+H+, 90%). IR (KBr, ν, cm−1): 1663
(s, C_O). Anal. calcd for C31H33NO2P2: C, 72.50; H, 6.48; N, 2.73; P, 12.06. Found: C, 72.45; H, 6.50;
N, 2.50; P, 12.02.
The other tLANOP ligands with (R) or (S) configuration were prepared by procedures similar to that
described for (R)-4, starting from the corresponding hydroxy amide and chlorophosphine.
3.3.2. (S)-Ph-tLANOP (S)-4a
20
10.6 g, 67% yield, colorless crystals, mp 95–96°C, [α]D −37.7 (c 1, CHCl3). Anal. found for
C31H33NO2P2: C, 72.42; H, 6.56; N, 2.46; P, 12.00.
3.3.3. (R)-Cy-tLANOP (R)-4b
Filtered through alumina in pentane, the residue after evaporation was not recrystallized. 8.20 g, 69%
yield, colorless solid, moisture sensitive, [α]D20 +18.8 (c 1, CHCl3). 1H NMR (250 MHz, CDCl3, 20°C):
3
δ 0.99 (s, 9H, tBu), 1.0–2.2 (m, 44H, Cy), 2.89 (s, 3H, NCH3), 5.26 (dd, 1H, JPH=4JPH=7.5, CHO).
MS (EI) m/z 537 (M+, 5%), 454 (M+−C6H11, 50%). IR (KBr, ν, cm−1): 1665 (s, C_O). Anal. calcd for
C31H57NO2P2: C, 69.24; H, 10.68; N, 2.60; P, 11.52. Found: C, 69.12; H, 10.66; N, 2.54; P, 11.22.
3.3.4. (S)-Cy-tLANOP (S)-4b
Filtered through alumina in pentane, the residue after evaporation was not recrystallized. 4.20 g, 66%
20
yield, colorless solid, moisture sensitive, [α]D −18.9 (c 1, CHCl3). Anal. found for C31H57NO2P2: C,
68.92; H, 10.87; N, 2.41; P, 11.48.
3.3.5. (R)-(2-Furyl)-tLANOP (R)-4c
1
2.77 g, 48% yield, colorless solid. H NMR (250 MHz, CDCl3, 20°C): δ 1.20 (s, 9H, tBu), 3.12 (s,
3H, NCH3), 5.9–6.1 (m, 6H), 6.5–7.3 (m, 7H). MS (EI) m/z 474.3 (M+H+, 30%). IR (KBr, ν, cm−1):
1681 (s, C_O). Anal. calcd for C23H25NO6P2: C, 58.35; H, 5.32; N, 2.96; P, 13.09. Found: C, 58.12; H,
5.29; N, 2.91; P, 12.93.
3.3.6. (S)-(Diphol)-tLANOP (S)-4d
1
4.74 g, 62% yield, colorless solid. H NMR (250 MHz, CDCl3, 20°C): δ 1.19 (s, 9H, tBu), 2.35 (s,
3
3H, NCH3), 6.07 (dd, 1H, JPH=4JPH=8.7, CHO). IR (KBr, ν, cm−1): 1669 (s, C_O). Anal. calcd for
C31H39NO2P2: C, 73.08; H, 5.74; N, 2.75; P, 12.16. Found: C, 72.79; H, 5.76; N, 2.88; P, 12.20.
3.3.7. (R)-(3,5-Xyl)-tLANOP (R)-4e
2.23 g, 73% yield, colorless solid. 1H NMR (250 MHz, CDCl3, 20°C): δ 1.05 (s, 9H, tBu), 2.15–2.50
(m, 24H), 2.78 (s, 3H, NCH3), 5.82 (dd, 1H, 3JPH=4JPH=8.0, CHO), 6.8–7.7 (m, 12H).

E. A. Broger et al. / Tetrahedron: Asymmetry 9 (1998) 4043–4054
4051
3.3.8. (R)-(3,5-tBu)-tLANOP (R)-4f
3.78 g, 83% yield, colorless solid. 1H NMR (250 MHz, CDCl3, 20°C): δ 1.10 (s, 9H, tBu), 1.25–1.29
(4s, 72H, tBu), 2.79 (s, 3H, NCH3), 5.89 (dd, 1H, 3JPH=4JPH=8.7, CHO), 7.2–7.6 (m, 12H). MS (ESI)
m/z 963.3 (M+H+, 25%). IR (KBr, ν, cm−1): 1675 (s, C_O). Anal. calcd for C63H97NO2P2: C, 78.62;
H, 10.16; N, 1.46; P, 6.44. Found: C, 77.72; H, 10.12; N, 1.36; P, 6.27.
3.3.9. (S)-(3,5-tBu)-tLANOP (S)-4f
3.41 g, 55% yield, colorless solid. Anal. found for C63H97NO2P2: C, 78.32; H, 10.02; N, 1.38; P, 6.14.
3.3.10. (S)-(3,5-tBu,4-MeO)-tLANOP (S)-4g
Filtered though alumina in pentane, the residue after evaporation was not recrystallized. 2.63 g,
1
60% yield, colorless solid, contains ca. 25% LiCl. H NMR (250 MHz, CDCl3, 20°C): δ 1.03 (s,
9H, tBu), 1.25–1.45 (m, 72H, tBu), 2.77 (s, 3H, NCH3), 3.64–3.70 (4×s, 12H, OCH3), 5.76 (dd, 1H,
3JPH=4JPH=9.0, CHO), 7.0–7.7 (m, 8H). MS (TSP) m/z 970.6 (M−tBu−OMe). IR (KBr, ν, cm−1): 1673
(s, C_O). Anal. calcd for C67H105NO6P2·0.25LiCl: C, 73.62; H, 9.68; N, 1.28; P, 5.67; Cl, 0.81. Found:
C, 73.34; H, 9.82; N, 1.00; P, 5.63; Cl, 0.77.
3.4. [(SP-4-3)-Pd((R)-dmphea)((S)-4a)]BF4 (R,S)-10
In a 50 ml Schlenk tube, a solution of [Pd((R)-dmphea)(NCMe)2]BF4 (R)-911 (105 mg, 0.25 mmol) and
(S)-4a (129 mg, 0.25 mmol) in dichloromethane (10 ml) was stirred for 10 min at rt. After evaporation,
the residue was washed with diethyl ether (10 ml) and dried for 2 h under vacuum, affording (R,S)-10
(211 mg, 98%) as a colorless solid. 1H NMR (250 MHz, CDCl3, 20°C): δ 0.67 (s, 9H, tBu), 1.45 (bs, 3H,
CHCH3), 2.18, 2.27, 2.40 (3×m, 9H), 3.62 (m, 1H, CHCH3), 5.04 (dd, 1H, 3JPH=6.3, 4JPH=6.4, CHO),
6.3–8.3 (m, 24 arom H). 31P NMR (101.25 MHz): δ 73.3 (d, JPP=41) (P–N); 136.6 (d, JPP=41) (P–O).
IR (KBr, ν, cm−1): 1689 (m, C_O), 1084 (s, BF4). Anal. calcd for C41H33BF4NO2P2Pd: C, 57.60; H,
5.54; N, 3.28. Found: C, 56.42; H, 5.54; N, 3.08. Within 24 h at rt (R,S)-10 was completely converted to
a mixture of (R,S)-11 (31P NMR signals at 78.8 (d) and 122.2 (d), JPP=35) and several by-products (main
31P NMR signals at 37 (m) and 101–103 (m)). Attempts to isolate (R,S)-11 by crystallization resulted
only in decomposition.
3.5. [Rh((R)-4a)(COD)]BF4 (R)-13a
In a 100 ml Schlenk tube, to a suspension of [Rh(COD)2]BF4 (144 mg, 0.355 mmol) in THF (10 ml)
was added dropwise a solution of (R)-4a (189 mg, 0.368 mmol) in THF (10 ml). The orange-yellow
solution was stirred for 5 min, evaporated and dried under high vacuum. The oily residue was stirred in
diethyl ether and the supernatant removed. The solid residue was dried for 12 h in high vacuum, affording
1
(R)-13a (278 mg, 97%) as a yellow crystalline material. H NMR (250 MHz, CDCl3, 20°C): δ 0.90 (s,
9H, tBu), 2.27 (d, 3H, NCH3, J=5.7), 2.2–2.6 (m, 8H), 4.38, 4.60, 4.76, 5.16 (4×bm, 4H, COD), 5.21
(dd, 1H, CHO, 3JPH=10.0, 4JPH=5.0), 7.2–8.2 (m, 20H). 31P NMR (101.25 MHz): δ 83.4 (dd, 1JRhP=170,
JPP=25) (P–N); 131.3 (dd, 1JRhP=167, JPP=25) (P–O). IR (KBr, ν, cm−1): 1685 (m, C_O), 1083 (s, BF4).
Anal. calcd for C39H45BF4NO2P2Rh: C, 57.73; H, 5.59; N, 1.73; P, 7.63. Found: C, 57.51; H, 5.59; N,
1.46; P, 7.41.

4052
E. A. Broger et al. / Tetrahedron: Asymmetry 9 (1998) 4043–4054
3.6. [Ir((S)-4a)(COD)]BF4 (S)-13b
The analogous reaction of [Ir(COD)2]BF4 with (S)-4a in THF afforded (S)-13b as a red crystalline
1
material. H NMR (250 MHz, CDCl3, 20°C): δ 0.84 (s, 9H, tBu), 2.30 (d, 3H, NCH3, J=6.3), 2.0–2.6
(m, 8H), 4.05, 4.15, 4.40, 4.90 (4×bm, 4H, COD), 5.10 (dd, 1H, CHO, 3JPH=9.6, 4JPH=4.0), 7.2–8.1 (m,
20H). 31P NMR (101.25 MHz): δ 71.0 (s) (P–N); 113.0 (s) (P–O). MS (ISP) m/z 814.5 (M+). IR (KBr, ν,
cm−1): 1686 (m, C_O), 1083 (s, BF4). Anal. calcd for C39H45BF4NO2P2Rh: C, 52.00; H, 5.04; N, 1.56;
P, 6.88. Found: C, 51.93; H, 4.95; N, 1.51; P, 6.43.
3.7. Hydrolysis of tLANOP ligands
A solution of (S)-4a (180 mg, 0.353 mmol) in dichloromethane (5 ml) was treated with water (1 ml),
25% HCl (0.1 ml) and methanol (1 ml). The mixture was stirred at rt for 1.5 h before sodium carbonate
(3.2 g) was added and the suspension stirred for 30 min. After filtration the filtrate was evaporated and
dried under vacuum to afford an oily residue consisting of a mixture of (S)-6 with 99.7% ee (GC analysis)
and diphenylphosphine oxide. The analogous hydrolysis of (R)-4b afforded (R)-6 of 98.2% ee, that of
(R)-4f afforded (R)-6 of 99.1% ee.
3.8. X-Ray structural analysis of (R,S)-10 and (R)-13a
Crystallographic data are summarized in Table 4. (R,S)-10: colorless crystals (0.3×0.3×0.2 mm)
suitable for X-ray diffraction were grown by slow diffusion of a diethyl ether layer into a CH2Cl2
solution of (R,S)-10 at −30°C. 5669 independent reflections were measured at rt on a Siemens R3m/V
diffractometer equipped with a graphite monochromator using Mo Kα radiation. The θ range for
data collection was 1.42–28.00°. (R)-13a: brick red crystals (0.4×0.3×0.25 mm) were grown by slow
evaporation of a solution in CH2Cl2/Et2O at rt. 3466 independent reflections were measured at 173
K on a Siemens P4 diffractometer equipped with a Siemens rotating anode M18XHF and a graphite
monochromator using Cu Kα radiation. The θ range for data collection was 4.28–63.94°.23 For the
refinement, 444 restraints were used. The Flack parameter was refined to 0.09(4), supporting the
established absolute configuration. Both structures were solved by direct methods and all non-hydrogen
atoms were refined anisotropically (with the exception of (R)-13a, where the solvent molecule and the
counterion were refined isotropically) using full-matrix least-square technique and Fourier synthesis
based on F2 (SHELX-package).24 H atoms were constrained in the calculated positions.
3.9. Asymmetric hydrogenation of 14
In a glove box a solution of [Rh(COD)2]BF4 (11.4 mg, 0.051 mmol) and (R)-4a (13.1 mg, 0.051 mmol)
in ethyl acetate (6.5 ml) was stirred for 60 min, after which time 14 (0.50 g, 2.57 mmol) was added. The
autoclave was removed from the box and the hydrogenation carried out under 10 bar of hydrogen and at
rt for 18 h. After evaporation of the solvent, a sample of the crude ester 15 was converted to the alcohol
(NaOMe/MeOH, 30 min, 50°C) and, after addition of few drops of acetic acid, analyzed by GC (column:
OV-61/p-DiMe-β-CD). The ee of (S)-15 was 71%.

E. A. Broger et al. / Tetrahedron: Asymmetry 9 (1998) 4043–4054
4053
Table 4
Crystal data
3.10. Asymmetric hydrogenation of 16a
In a glove box a solution of [IrCl(COD)]2 (13.4 mg, 0.020 mmol), (R)-4f (57.7 mg, 0.060 mmol) and
tetrabutylammonium iodide (59.1 mg, 0.16 mmol) in toluene (4 ml) was stirred for 30 min, after which
time 16a (0.343 g, 1.0 mmol) and methanol (4 ml) were added. The autoclave was removed from the
box and the hydrogenation carried out under 100 bar of hydrogen and at rt for 44 h. Evaporation of the
solvent and basic work-up afforded crude (R)-17, which was converted to the amide of (−)-camphanic
acid (DCC, DMAP, CH2Cl2, 60°C, 30 min). GC analysis (column: OV-240 OH, Ohio Valley Chemicals,
Marietta, OH) showed the conversion to be complete. The ee of (R)-17 was 72% and the chemoselectivity
48%.
Acknowledgements
We wish to thank our colleagues from the Analytical Services for NMR Spectra (Dr. W. Arnold),
elemental analyses (Dr. S. Müller), GC analyses (Dr. W. Walther), P. Schönholzer for the X-ray structure
determinations and Dr. D. Bur for molecular modeling calculations. We also thank M. Glatz, A. Meili,
D. Spiess and J. Stadelmann for skillful technical assistance and Dr. Y. Crameri for helpful discussions.

4054
E. A. Broger et al. / Tetrahedron: Asymmetry 9 (1998) 4043–4054
References
1. Roucoux, A.; Agbossou, F.; Mortreux, A.; Petit, F. Tetrahedron: Asymmetry 1993, 4, 2279.
2. Carpentier, J.-F.; Mortreux, A. Tetrahedron: Asymmetry 1997, 8, 1083 and references therein.
3. Ait Ali, M.; Alloud, S.; Karim, A.; Roucoux, A.; Mortreux, A. Tetrahedron: Asymmetry 1995, 6, 369.
4. Roucoux, A.; Thieffry, L.; Carpentier, J.-F.; Devocelle, M.; Méliet, C.; Agbossou, F.; Mortreux, A.; Welch, A. J.
Organometallics 1996, 15, 2440.
5. The preferred calculated conformation of 4a in a hypothetical cationic rhodium complex was practically identical to that
determined later in 10 and 13a by X-ray (D. Bur, unpublished results).
6. Heiser, B.; Broger, E. A.; Crameri, Y. Tetrahedron: Asymmetry 1991, 2, 51.
7. Kitamura, M.; Ohkuma, T.; Inoue, S.; Sayo, N.; Kumobayashi, H.; Akutagawa, S.; Ohta, T.; Takaya, H.; Noyori, R. J. Am.
Chem. Soc. 1988, 110, 629.
8. Noyori, R. Asymmetric Catalysis in Organic Synthesis, John Wiley & Sons, Inc., 1994.
9. Allen, W. D.; Taylor, B. F. J. Chem. Soc., Dalton Trans. 1982, 51.
10. (a) Roberts, N. K.; Wild, S. B. J. Am. Chem. Soc. 1979, 101, 6254; (b) J. Chem. Soc., Dalton Trans. 1979, 2015; (c) Otsuka,
S.; Nakamura, A.; Kano, T.; Tani, K. J. Am. Chem. Soc. 1971, 93, 4301; (d) Tani, K.; Brown, L. D.; Ahmed, J.; Ibers, J.
A.; Yokota, M.; Nakamura, A.; Otsuka, S. ibid. 1977, 99, 7876.
11. Fornies, J.; Navarro, R.; Sicilia, V. Polyhedron 1988, 28, 219. During the preparation of the manuscript, the following
cognate papers appeared: Falvello, L. R.; Fernandez, S.; Navarro, R.; Pascual, I.; Urriolabeitia, E. P. J. Chem. Soc., Dalton
Trans. 1997, 763; and Barco, I. C.; Falvello, L. R.; Fernandez, S.; Navarro, R.; Urriolabeitia, E. P. J. Chem. Soc., Dalton
Trans. 1998, 1699.
12. Schrock, R. R.; Osborn, J. A. J. Am. Chem. Soc. 1971, 93, 2397.
13. Feiken, N.; Pregosin, P.; Trabesinger, G. to be published in Organometallics, 1998.
14. Armstrong, S. K.; Brown, J. M.; Burk, M. J. Tetrahedron Lett. 1993, 34, 879.
15. (a) Naili, S.; Carpentier, J.-F.; Agbossou, F.; Mortreux, A.; Nowogrocki, G.; Wignacourt, J.-P. Organometallics 1995, 14,
401. (b) Cesarotti, E.; Chiesa, A.; Ciani, G.; Sironi, A. J. Organomet. Chem. 1983, 251, 79.
16. Mayer, H. Pure & Appl. Chem. 1979, 51, 535.
17. (a) Den Hollander, C. W.; Leimgruber, W.; Mohacsi, E. Ger. Offen. DE 2003486 to F. Hoffmann-La Roche Ltd, August
13, 1970, CAN 73:77081. (b) Imwinkelried, R. Chimia 1997, 51, 300.
18. (a) Broger, E. A.; Crameri, Y.; Schmid, R.; Siegfried, T. EP 691325 A1 960110 to F. Hoffmann-La Roche AG, CAN
124:260450. (b) Scalone, M. ChiraTech 97, Philadelphia, PA, 11–13 November 1997.
19. Broger, E. A.; Scalone, M.; Wehrli, C. EP 850931 A2 980701 to F. Hoffmann-La Roche AG, CAN 129:109000.
20. Schmid, R.; Cereghetti, M.; Heiser, B.; Schönholzer, P.; Hansen, H.-J. Helv. Chim. Acta 1988, 71, 897.
21. RajanBabu, T. V.; Casalnuovo, A. L.; Ayers, T. J. Am. Chem. Soc. 1994, 116, 9869.
22. Seifert, D.; Hui, R. C. Tetrahedron Lett. 1984, 25, 5251.
23. The high correlation coefficients in the crystal structure of (R)-13a reflect the rather poor quality of the crystal and the
errors due to absorption effects (use of the Cu-radiation). Moreover, the solvent molecule (CH2Cl2) and the counterion
(BF4−) show some degree of disorder. Therefore, the somewhat weaker quality of the structural analysis should not affect
the correctness of the structure of (R)-13a.
24. Sheldrick, G. M. Acta Crystallography Section A 1990, 46, 467.