Conformational preferences and enantiodiscrimination of phosphino-4-(1-hydroxyalkyl)oxazoline-metal-olefin complexes resulting from an OH-metal hydrogen bond

Article abstract of DOI:10.1021/jo051610q

Phosphinooxazolines carrying (1-hydroxy-1-phenyl)methyl and (1-methoxy-1-phenyl)methyl substituents in the 4 position of the oxazoline ring exhibit contrasting behavior in Pd- and Ir-catalyzed allylic alkylations. Whereas catalysts with the methoxy-containing ligand generally provide products with high ee's, use of catalysts prepared from the hydroxy-containing ligand results in products with low ee's or even racemates. DFT calculations suggest the presence of a hydrogen bond with Pd(0) as the proton acceptor in the hydroxy-containing olefin-Pd(0) complexes, which induces a conformational change in the ligand, leading to different stereoselectivity.

Full text of DOI:10.1021/jo051610q

Conformational Preferences and Enantiodiscrimination of
Phosphino-4-(1-hydroxyalkyl)oxazoline-Metal-Olefin Complexes
Resulting from an OH-Metal Hydrogen Bond
Anders Fro¨lander, Serghey Lutsenko, Timofei Privalov, and Christina Moberg*
Department of Chemistry, Organic Chemistry, Royal Institute of Technology,
SE 100 44 Stockholm, Sweden
kimo@kth.se
Received August 1, 2005
Phosphinooxazolines carrying (1-hydroxy-1-phenyl)methyl and (1-methoxy-1-phenyl)methyl sub-
stituents in the 4 position of the oxazoline ring exhibit contrasting behavior in Pd- and Ir-catalyzed
allylic alkylations. Whereas catalysts with the methoxy-containing ligand generally provide products
with high ee’s, use of catalysts prepared from the hydroxy-containing ligand results in products
with low ee’s or even racemates. DFT calculations suggest the presence of a hydrogen bond with
Pd(0) as the proton acceptor in the hydroxy-containing olefin-Pd(0) complexes, which induces a
conformational change in the ligand, leading to different stereoselectivity.
Introduction
polymeric support,7 or temperature;8 and the use of
mixtures of chiral and achiral ligands.9 A minor struc-
The steric properties of ligands employed in asym-
metric metal catalysis are crucial for efficient transfer
of chirality, and for control of the absolute configuration
of new stereocenters. Preparation of the two enantiomers
of a product usually requires access to two ligands of
opposite absolute configuration. Since the ultimate source
of a chiral ligand is often a naturally occurring compound
existing in only one enantiomeric form, access to com-
pounds derived from the opposite enantiomer may be
limited. It would therefore be highly desirable to be able
to modify the structure of the easily accessible enanti-
omer in such a way as to alter its enantioinducing
preference.1 Several examples of the reversal of enanti-
oselectivity caused by modifications of catalyst structure
or reaction conditions are known. These include N-
alkylations2 or other modifications of ligand structures;3
change of the metal ion,4 metal-to-ligand ratio,5 solvent,6
(3) (a) Evans, D. A.; Johnson, J. S.; Burgey, C. S.; Campos, K. R.
Tetrahedron Lett. 1999, 40, 2879-2882. (b) Sibi, M. P.; Chen, J.; Cook,
G. R. Tetrahedron Lett. 1999, 40, 3301-3304. (c) Chelucci, G.; Pinna,
G. A.; Saba, A.; Valenti, R. Tetrahedron: Asymmetry 2000, 11, 4027-
4036. (d) Li, X.-G.; Cheng, X.; Ma, J.-A.; Zhou, Q.-L. J. Organomet.
Chem. 2001, 640, 65-71.
(4) (a) Nishibayashi, Y.; Segawa, K.; Ohe, K.; Uemura, S. Organo-
metallics 1995, 14, 5486-5487. (b) Nishibayashi, Y.; Segawa, K.;
Takada, H.; Ohe, K.; Uemura, S. Chem. Commun. 1996, 847-848. (c)
Pe´rez Luna, A.; Bonin, M.; Micouin, L.; Husson, H.-P. J. Am. Chem.
Soc. 2002, 124, 12098-12099. (d) Du, D.-M.; Lu, S.-F.; Fang, T.; Xu,
J. J. Org. Chem. 2005, 70, 3712-3715. (e) Desimoni, G.; Faita, G.;
Guala, M.; Pratelli, C. J. Org. Chem. 2003, 68, 7862-7866. (f)
Desimoni, G.; Faita, G.; Guala, M.; Laurenti, A. Eur. J. Org. Chem.
2004, 3057-3062. (g) Mao, J.; Wan, B.; Wu, F.; Wang, R.; Lu, S. J.
Mol. Catal. A 2005, 232, 9-12. (h) Mao, J.; Wan, B.; Zhang, Z.; Wang,
R.; Wu, F.; Lu, S. J. Mol. Catal. A 2005, 225, 33-37.
(5) Shimizu, I.; Matsumoto, Y.; Shoji, K.; Ono, T.; Satake, A.;
Yamamoto, A. Tetrahedron Lett. 1996, 37, 7115-7118.
(6) Arseniyadis, S.; Valleix, A.; Wagner, A.; Mioskowski, C. Angew.
Chem., Int. Ed. 2004, 43, 3314-3317.
(7) Reed, N. N.; Dickerson, T. J.; Boldt, G. E.; Janda, K. D. J. Org.
Chem. 2005, 70, 1728-1731.
(1) Sibi, M. P.; Liu, M. Curr. Org. Chem. 2001, 5, 719-755.
(2) (a) Kimura, K.; Sugiyama, E.; Ishizuka, T.; Kunieda, T. Tetra-
hedron Lett. 1992, 33, 3147-3150. (b) Cobb, A. J. A.; Marson, C. M.
Tetrahedron: Asymmetry 2001, 12, 1547-1550. (c) Cobb, A. J. A.;
Marson, C. M. Tetrahedron 2005, 61, 1269-1279.
(8) (a) Sibi, M. P.; Gorikunti, U. M.; Liu, M. Tetrahedron 2002, 58,
8357-8363. (b) Casey, C. P.; Martins, S. C.; Fagan, M. A. J. Am. Chem.
Soc. 2004, 126, 5585-5592.
(9) Reetz, M. T.; Mehler, G. Tetrahedron Lett. 2003, 44, 4593-4596.
10.1021/jo051610q CCC: $30.25 © 2005 American Chemical Society
Published on Web 11/01/2005
9882
J. Org. Chem. 2005, 70, 9882-9891

Phosphino-4-(1-hydroxyalkyl)oxazoline-Metal-Olefin Complexes
center Pt-HN arrangement16 in zwitterionic Pt(II) com-
plexes with properly situated ammonium groups.17
We decided to explore whether the phenomenon ob-
served in 2-(1-hydroxyalkyl)pyridinooxazolines may ap-
pear in catalytic systems with other types of ligands
carrying suitably positioned hydroxyalkyl and alkoxy-
alkyl substituents. For this purpose we selected phos-
phinooxazolines (PHOX), first prepared independently by
the groups of Pfaltz,18 Helmchen,19 and Williams20 in
1993, which have proven to be highly versatile ligands
for a range of catalytic processes.21 A number of phos-
phinooxazolines carrying either hydroxy or methoxy
substituents were therefore prepared and studied.
FIGURE 1. Conformations of Pd(II)-allyl (A) and Pd(0)-
olefin (B) complexes with hydroxymethyl-substituted pyridi-
nooxazolines.
tural modification of a ligand leading to a catalytic system
serving as a pseudo-enantiomer was reported by Bala-
voine and co-workers, who found that palladium-cata-
lyzed allylic alkylations employing bisoxazoline ligands
with 4-hydroxybenzyl and 4-methoxybenzyl substituents,
with the same stereochemistry, provided products with
opposite absolute configuration.10 Their explanation for
this behavior was based on the assumption that the
hydroxy group in the ligand was interacting via a
hydrogen bond with the nucleophile.11
We found earlier that 2-(1-hydroxyalkyl)- and 2-(1-
alkoxyalkyl)pyridinooxazolines resulted in profoundly
different enantioselectivities in palladium-catalyzed al-
lylations of dimethyl malonate.12 This difference was
found to originate in the different conformational prefer-
ences of the two types of ligands,13 which we explained
by the presence of a hydrogen bond with palladium(0)
serving as the proton acceptor in the hydroxy-containing
complex.14 In the palladium(II)-allyl complexes with
2-(1-hydroxyalkyl) substituents, the hydroxy group was
aligned with the pyridine ring, anti to the metal (A,
Figure 1), whereas in the palladium(0)-olefin complex,
the hydroxy group pointed toward the palladium atom
with a N-C-C-O angle of -57° (B).13
For the conformational change accompanying the
reduction of palladium to affect the stereochemistry of a
catalytic reaction, it has to occur in the enantiodeter-
mining step, and be essentially completed in the transi-
tion state. For this reason, metal-catalyzed allylations
were considered to be suitable for our studies, because
the stereochemistry of the product is usually determined
during the nucleophilic attack, a step in which the
oxidation state of the metal is reduced by two units. This
situation was thus expected to lead to a conformational
change if the hydroxy proton becomes involved in hydro-
gen bonding to the low-valence metal.
Results and Discussion
Ligand Design and Synthesis. Two types of ligands,
one having a 4,5-disubstituted oxazoline ring (1-4) and
one having a large substituent in the 4 position of the
oxazoline ring (5-8), were considered to fulfill our
requirements.
Several methods have been employed for the prepara-
tion of phosphinooxazolines.21 The two types of ligands
are known to be accessible starting from 2-halobenzoni-
triles and 2-halobenzoic acid chlorides, respectively,
(Scheme 1).22
Ligand 1 was prepared, following the first of these
procedures, from threoninol, which was reacted with
2-fluorobenzonitrile in the presence of cadmium acetate.
Protection of the alcohol function followed by nucleophilic
aromatic substitution using lithium diphenylphosphide
and final deprotection gave 1 (Scheme 2).
To avoid protection/deprotection of the alcohol function,
we employed a different route, starting by reaction of the
appropriate amino alcohol with 2-iodobenzonitrile, for the
syntheses of 2-4. The phosphine function was introduced
employing palladium-catalyzed coupling with diphe-
nylphosphine, either directly to afford 323 or after O-
methylation to give 2 and 424 (Scheme 3).
(15) (a) Epstein, L. M.; Shubina, E. S. Coord. Chem. Rev. 2002, 231,
165-181. (b) Calhorda, M. J. Chem. Commun. 2000, 801-809. (c)
Brammer, L.; Zhao, D.; Lapido, F. T.; Braddock-Wilking, J. Acta
Crystallogr., Sect. B 1995, 51, 632-640. (d) Braga, D.; Grepioni, F.;
Tedesco, E.; Biradha, K.; Desiraju, G. R. Organometallics 1997, 16,
1846-1856. (e) Desiraju, G. R. J. Chem. Soc., Dalton Trans. 2000,
3745-3751.
(16) Wehman-Ooyevaar, I. C. M.; Grove, D. M.; Kooijman, H.; van
der Sluis, P.; Spek, A. L.; van Koten, G. J. Am. Chem. Soc. 1992, 114,
9916-9924.
(17) (a) Wehman-Ooyevaar, I. C. M.; Grove, D. M.; van der Sluis,
P.; Spek, A. L.; van Koten, G. Chem. Commun. 1990, 1367-1369. (b)
Wehman-Ooyevaar, I. C. M.; Grove, D. M.; de Vaal, P.; Dedieu, A.;
van Koten, G. Inorg. Chem. 1992, 31, 5484-5493. (c) Pregosin, P. S.;
Ru¨egger, H.; Wombacher, F.; van Koten, G.; Grove, D. M.; Wehman-
Ooyevaar, I. C. M. Magn. Reson. Chem. 1992, 30, 548-551.
(18) von Matt, P.; Pfaltz, A. Angew. Chem., Int. Ed. 1993, 32, 566-
568.
Hydrogen bonds between protons bound to oxygen or
nitrogen and late transition metals in low-oxidation
states have indeed been observed in a number of cases.15
Experimental and computational evidence has, for ex-
ample, been provided for the involvement of Pt-HN
(19) Sprinz, J.; Helmchen, G. Tetrahedron Lett. 1993, 34, 1769-
1772.
(20) Dawson, G. J.; Frost, C. G.; Williams, J. M. J.; Coote, S. J.
Tetrahedron Lett. 1993, 34, 3149-3150.
hydrogen bonds exhibiting an essentially linear three-
(10) Hoarau, O.; A¨ıt-Haddou, H.; Daran, J.-C.; Cramaile`re, D.;
Balavoine, G. G. A. Organometallics 1999, 18, 4718-4723.
(11) A¨ıt-Haddou, H.; Hoarau, O.; Cramaile´re, D.; Pezet, F.; Daran,
J.-C.; Balavoine, G. G. A. Chem.sEur. J. 2004, 10, 699-707.
(12) Nordstro¨m, K.; Macedo, E.; Moberg, C. J. Org. Chem. 1997, 62,
1604-1609.
(21) Helmchen, G.; Pfaltz, A. Acc. Chem. Res. 2000, 33, 336-345.
(22) Peer, M.; de Jong, J. C.; Kiefer, M.; Langer, T.; Rieck, H.; Schell,
H.; Sennhenn, P.; Sprinz, J.; Steinhagen, H.; Wiese, B.; Helmchen, G.
Tetrahedron 1996, 52, 7547-7583.
(23) Ent-3 was prepared previously, but the procedure used was not
described: Lloyd-Jones, G. C.; Butts, C. P. Tetrahedron 1998, 54, 901-
914.
(13) Svensson, M.; Bremberg, U.; Hallman, K.; Cso¨regh, I.; Moberg,
C. Organometallics 1999, 18, 4900-4907.
(24) Ent-4 was prepared previously, but the procedure used was not
described: Aeby, A.; Consiglio, G. Inorg. Chim. Acta 1999, 296, 45-
51.
(14) Hallman, K.; Fro¨lander, A.; Wondimagegn, T.; Svensson, M.;
Moberg, C. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 5400-5404.
J. Org. Chem, Vol. 70, No. 24, 2005 9883

Fro¨lander et al.
SCHEME 1
SCHEME 2
SCHEME 4
aryl iodide with diphenylphosphine, afforded a consider-
ably higher overall yield of 7 in fewer steps.
SCHEME 3
DFT Computations of Palladium(0)-Olefin Com-
plexes. We decided to first study the product Pd(0)-
olefin complexes of ligand 1 using the B3LYP functional25
in Jaguar v4.0.26 The complexes were optimized using
the lacvp**/6-31G(d,p) basis set.27,28 It was assumed that
nucleophilic attack occurred at the allyl group from the
side opposite to the metal at the allylic position trans to
phosphorus,29 in accordance with experimental observa-
tions.30 Nucleophilic attack can occur either at the endo-
π-allyl-palladium complex (C1), leading to olefin com-
plex D1, or at the exo isomer (C2) leading to D2, with
the olefin ligand having opposite absolute configuration
(Figure 2). Syn-anti and anti-anti allyl complexes were
(25) (a) Becke, A. D.; J. Chem. Phys. 1993, 98, 5648-5652. (b) Lee,
C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785-789.
(26) Jaguar, version 4.0; Schro¨dinger, Inc.: Portland, OR, 1991-
2000.
(27) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299-310.
(28) (a) Hehre, W. J.; Ditchfield, R.; Pople, J. A. J. Chem. Phys. 1972,
56, 2257-2261. (b) Francl, M. M.; Pietro, W. J.; Hehre, W. J.; Binkley,
J. S.; Gordon, M. S.; DeFrees, D. J.; Pople, J. A. J. Chem. Phys. 1982,
77, 3654-3665. (c) Hariharan, P. C.; Pople, J. A. Theor. Chim. Acta
1973, 28, 213-222.
Ligands 5-8 were obtained from 2-iodobenzoyl chloride
as shown in Scheme 4. This procedure, replacing the
previously used nucleophilic substitution of the aromatic
fluoride by lithium diphenylphosphide used in the prepa-
ration of ent-722 by palladium-catalyzed coupling of an
(29) (a) Blo¨chl, P. E.; Togni, A. Organometallics 1996, 15, 4125-
4132. (b) Gilardoni, F.; Weber, J.; Chermette, H.; Ward, T. R. J. Phys.
Chem. A 1998, 102, 3607-3613.
(30) Steinhagen, H.; Reggelin, M.; Helmchen, G. Angew. Chem., Int.
Ed. 1997, 36, 2108-2110.
9884 J. Org. Chem., Vol. 70, No. 24, 2005

Phosphino-4-(1-hydroxyalkyl)oxazoline-Metal-Olefin Complexes
FIGURE 2. Nucleophilic attack at endo- and exo-π-allyl-
palladium complexes of ligand 1.
not considered, as such complexes have not been identi-
fied in solution or in the solid state.31
FIGURE 3. Two lowest-energy conformations of the product
olefin complex D2 arising from attack at the exo-π-allyl
complex. D2′ was found to be 2.6 kcal/mol lower in energy than
the most stable complex from an endo attack, and 2.1 kcal/
mol more stable than D2′′. All hydrogen atoms except the
hydroxy proton have been omitted for clarity.
To simplify the computations, the substituents on the
oxazoline ring, the phosphorus atom, and the olefin were
replaced by hydrogen atoms. Moreover, the nucleophile
was replaced by NH2-. The enantiodiscrimination is us-
ually assumed to originate from the formation of the most
favorable olefin complex.32 We found an olefin complex
(D2′) resulting from nucleophilic attack at the exo com-
plex to be 2.6 kcal mol-1 lower in energy than the most
stable complex attained after attack at the endo com-
plex.33 This is in accordance with the results by Helm-
chen, who has shown experimentally that the exo-allyl
complex is more stable than the endo isomer.34 The exo
Single-point calculations were performed using the
larger lacv3p**++/6-311+G(d,p) basis set. The results
confirmed complex D2′ to be lower in energy (1.3 kcal
mol-1) than D2′′. Single-point calculations were also
performed on a simulation of a THF solution. Again,
complex D2′ was found to be lower in energy (0.8 kcal
mol-1).
To model olefin complexes of ligand 2, the same
calculations were performed, with the hydroxy group
replaced by a methoxy group. Unlike the former ligand,
and as expected, the most stable conformation had the
methoxy group pointing away from palladium.
Palladium-Catalyzed Allylations. We next decided
to study the consequences of the theoretically observed
conformational preferences on the stereochemistry of the
products from palladium-catalyzed allylic substitutions
of rac-(E)-1,3-diphenyl-2-propenyl acetate (21) and rac-
3-cyclohexenyl acetate (22) (Scheme 5).
Substitutions of rac-(E)-1,3-diphenyl-2-propenyl ac-
etate with dimethyl malonate using [(η3-C3H5)PdCl]2 and
ligands 1-8 as precatalysts were first studied in THF
at 0 °C. For 1-6, only minor differences in enantioselec-
tivity were observed between reactions using hydroxy-
and alkoxy-containing ligands with the same substitution
pattern (Table 1), whereas 7 and 8 exhibited a somewhat
larger difference (88 and 99% ee, respectively).
complex was also shown to lead to the major product.30
For both types of olefin complexes (D1 and D2), the
most stable conformations were found to have the hy-
droxy proton pointing toward palladium. For the complex
arising from attack at the exo-allyl complex, this confor-
mation (D2′) was found to be 2.1 kcal mol-1 lower in
energy than a conformation having the hydroxy proton
pointing away from palladium (D2′′, Figure 3). Complex
D2′ was found to have an OH-Pd distance of 2.47 Å, a
Pd-O distance of 3.32 Å, an O-H distance of 0.98 Å, and
an O-H-Pd bond angle of 145°, all in accordance with
a weak hydrogen bond.15c Complex D2′′, in which the
hydroxy proton is not involved in hydrogen bonding, had
a slightly shorter O-H bond (0.97 Å). The closer proxim-
ity of palladium to the hydrogen atom than to the oxygen
atom and the slight elongation of the O-H bond in D2′
are arguments against donation of an oxygen lone pair
to palladium as the reason for the observed conformation.
Moreover, that type of interaction would rather be
favored in the Pd(II) complexes.
(31) Zehnder, M.; Schaffner, S.; Neuburger, M.; Plattner, D. A. Inorg.
Chim. Acta 2002, 337, 287-298.
Even if a hydrogen bond is present in the Pd(0)-olefin
complex, it might be not fully developed in the transition
state, and therefore not affect the stereochemistry of the
catalytic reaction to any greater extent. We reasoned that
a less polar solvent should result in a less reactive allyl
system, and therefore a later transition state for the
(32) Saitoh, A.; Achiwa, K.; Tanaka, K.; Morimoto, T. J. Org. Chem.
2000, 65, 4227-4240.
(33) Other local minima having the inner chelate cycle (Pd-N-C-
C-C-P) bent in other ways were also found, but they were all higher
in energy.
(34) Sprinz, J.; Kiefer, M.; Helmchen, G.; Reggelin, M.; Huttner, G.;
Walter, O.; Zsolnai, L. Tetrahedron Lett. 1994, 35, 1523-1526.
J. Org. Chem, Vol. 70, No. 24, 2005 9885

Fro¨lander et al.
TABLE 3. Reactions of rac-(E)-1,3-diphenyl-2-propenyl
Acetate with Acetylacetone at 25 °C Yielding 26
SCHEME 5
conversion (%)
(time (h))
ee (%)
(abs conf)
entry
ligand
solvent
1
2
3
4
5
6
1 (OH)
2 (OMe)
7 (OH)
7 (OH)
8 (OMe)
8 (OMe)
THF
100 (2)
100 (2)
94 (15)
55 (15)
95 (2)
80 (R)
84 (R)
THF
THF
30 (S)
toluene
THF
35 (S)
>99 (S)
>99 (S)
toluene
90 (2)
TABLE 4. Reactions of rac-3-cyclohexenyl with
Dimethyl Malonate in THF at 0 °C Yielding 27
conversion (%)
(time (h))
ee (%)
(abs conf)
entry
ligand
1
2
3
4
1 (OH)
2 (OMe)
7 (OH)
8 (OMe)
48 (15)
49 (15)
95 (20)
98 (10)
20 (S)
5 (S)
59 (R)
26 (R)
TABLE 1. Reactions of rac-(E)-1,3-diphenyl-2-propenyl
Acetate with Dimethyl Malonate in THF at 0 °C for 5 h
Yielding 25
entry
ligand
conversion (%)
ee (%) (abs conf)
1
2
3
4
5
6
7
8
1 (OH)
2 (OMe)
3 (OH)
4 (OMe)
5 (OH)
6 (OMe)
7 (OH)
8 (OMe)
99
100
100
100
100
100
100
100
95 (R)
95 (R)
92 (R)
88 (R)
97 (S)
98 (S)
88 (S)
99 (S)
TABLE 2. Reactions of rac-(E)-1,3-diphenyl-2-propenyl
Acetate with Dimethyl Malonate in Toluene at 25 °C
Yielding 25
conversion (%)
(time (h))
ee (%)
(abs conf)
entry
ligand
1
2
3
4
1 (OH)
2 (OMe)
7 (OH)
8 (OMe)
100 (1)
100 (1)
100 (1)
81 (18)
93 (R)
91 (R)
76 (S)
97 (S)
nucleophilic attack. To test this hypothesis, we performed
the reactions using ligands 1, 2, 7, and 8 in toluene at
25 °C.
FIGURE 4. Nucleophilic attack at endo- and exo-π-allyl-
palladium complexes of ligand 7.
A somewhat larger difference in enantioselectivity was
now observed between ligands 7 and 8 (76 and 97% ee,
respectively, Table 2). That this difference is not due to
the nature of the substituent, the presence of a hydroxy
group, seems clear from a comparison with the results
using ligands 1 and 2. The two ligands, carrying hydroxy
and methoxy substituents, respectively, do not exhibit
significantly different enantioselectivities. These ligands
lack bulky substituents in the 4 position of the oxazoline
ring, and their enantiodiscrimination is therefore not
expected to be affected to any greater extent by the
presence of a hydrogen bond, which leads to rotation
around the carbon-carbon bond in the 4 substituent.
Use of a nucleophile less reactive than malonate is also
expected to result in a later transition state for the
nucleophilic attack, and therefore possibly in a more
pronounced effect of a hydrogen bond leading to a
different conformation of the hydroxy-containing ligand.
Indeed, a considerable difference in enantioselectivity
between ligands 7 and 8 was observed in toluene (35 vs
>99% ee) as well as in THF (30 vs >99% ee) when
acetylacetone was the nucleophile (Table 3). At the same
time, ligands 1 and 2 gave products with similar enan-
tioselectivity (80 and 84% ee, respectively, in THF).
Ligands 1, 2, 7, and 8 were further studied in reactions
using rac-3-cyclohexenyl acetate as substrate. The enan-
tioselectivities were generally lower than those observed
using 1,3-diphenylpropenyl acetate, although significant
differences in ee were observed between ligands 7 and
8, as well as between 1 and 2 (Table 4). For this cyclic
substrate, use of the hydroxy-containing ligands resulted
in higher enantioselectivities, in contrast to reactions
with 1,3-diphenylpropenyl acetate.
To rationalize the contrasting experimental results
obtained using ligands 7 and 8 and the diphenyl sub-
strate 21, we performed calculations similar to those
described for the Pd(0)-olefin complexes of ligand 1 (D1
and D2) on the product olefin complexes of ligand 7 (F1
and F2, Figure 4). The only simplification made was the
replacement of the nucleophile by NH2-. The results gave
valuable qualitative information on the stereoselection.35
In analogy to the results obtained for ligand 1, we
found the Pd(0)-olefin complex F2 arising from attack
9886 J. Org. Chem., Vol. 70, No. 24, 2005

Phosphino-4-(1-hydroxyalkyl)oxazoline-Metal-Olefin Complexes
between reactions leading to F1′ and F2′ may be the
reason why high selectivity is not achieved with ligand
7 under certain conditions.
Similar calculations were performed on the Pd(0)-
olefin complexes of ligand 8. As expected, a complex
leading to the S product was found to be most stable.
Complexes leading to the R product were found to have
considerably higher energy (>3 kcal mol-1), thus explain-
ing the high enantioselectivities obtained using this
ligand.
Iridium-Catalyzed Allylations. Phosphinooxazolines
have also been found to induce high enantioselectivity
in Ir-catalyzed allylations.38 The detailed mechanism is
not yet known, but it has been established that the
reaction proceeds via a double inversion process.39 The
isomerization of the allyl-Ir intermediates is slower than
for allyl-Pd complexes, and memory effects can be
substantial.40 Allyl carbonates generally give faster reac-
tions than acetates, but have in some cases been shown
to provide lower enantioselectivities.41 Because of the
preferred octahedral geometry of allyl-Ir(III) complexes,
several complexes have to be considered. Some complexes
have been isolated and characterized, but it is unclear
which ones are leading to product and by which path they
do so.41
We compared 7 and 8 as ligands in Ir-catalyzed
allylations using (E)-cinnamyl acetate (28), (E)-cinnamyl
methyl carbonate (29), and (E)-3-(4-methoxyphenyl)-2-
propenyl methyl carbonate (30) (Table 5). The nucleophile
was either prepared before the reaction using NaH or
generated in situ using Cs2CO3. The latter method has
recently been used in Rh-catalyzed allylic alkylations
resulting in high enantioselectivities.42 When comparing
the results of the two ligands using the same reaction
conditions, we found that ligand 8 generally gave faster
reactions, higher branched:linear ratios, and considerably
higher ee’s. Ligand 8 afforded product 31 with an ee as
high as 90% (R; entry 6), whereas ligand 7 generated the
product with opposite absolute configuration with an ee
of 6% (entry 3). Substrate 30, having an electron-donating
substituent (entries 9 and 10), gave the product 32 with
higher branched:linear ratios than (and ee’s similar to)
31. Exchanging THF for toluene yielded comparable
results.
FIGURE 5. Conformations of Pd(0)-olefin complexes of
ligand 7. F1′, the most stable F1 conformation, was found to
be 1.3 kcal/mol higher in energy than F2′, the lowest-energy
F2 conformation.
at the exo-allyl complex to be lower in energy than F1
arising from attack at the endo isomer. As F2 represents
the S product and F1 the R product when dimethyl
malonate or acetylacetone is the nucleophile, this is in
agreement with the experimental results. The lowest
lying conformation of F2, F2′, has a nonplanar Pd-N-
C-C-C-P ring, in agreement with previous observations
by Helmchen on other phosphinooxazolines.36 The con-
formation of this ring determines the position of the
phenyl rings, which as in analogous complexes were
found to be essentially perpendicular to each other
(Figure 5). The axial phenyl group has its edge and the
equatorial its face toward the metal, a situation normally
leading to preferential reaction of the exo-allyl complex.
The hydroxy group in F2′ is pointing toward Pd with a
Pd-H bond length of 2.23 Å, a typical hydrogen bond
distance.15c
Several conformational minima were found for the Pd-
(0)-olefin complex F1 originating from the endo-allyl-
palladium complex E1, and thus having an olefin ligand
of opposite absolute configuration compared to F2. All
conformational minima having the ligand backbone bent,
and thus the phenyl groups arranged, in a way similar
Considering the ambiguities regarding the reaction
mechanism, we can draw no safe conclusion about the
origin of these interesting observations. Our results
demonstrate, however, that the two types of ligands also
in this reaction exhibit contrary behavior.
to F2′ were found to be considerably higher in energy
(>2.7 kcal mol-1) than F2′. A conformational minimum,
F1′, only 1.3 kcal mol-1 higher in energy than F2′, with
an OH-Pd interaction and the ligand backbone bent in
an opposite way and hence with its phenyl groups
differently oriented was found, however.37 Competition
Conclusions
In summary, we have shown that palladium-catalyzed
allylic alkylations of rac-1,3-diphenylpropenyl acetate
with malonate and acetylacetone using phosphinooxazo-
lines with (1-hydroxy-1-phenyl)methyl and (1-methoxy-
(35) Because of the large size of these systems, fully converged
structures were not obtained with the standard convergence threshold,
but the level of convergence of all the structures was identical and
within the standard accuracy of B3LYP/LACVP**.
(38) Janssen, J. P.; Helmchen, G. Tetrahedron Lett. 1997, 38, 8025-
8026.
(39) Bartels, B.; Garc´ıa-Yebra, C.; Rominger, F.; Helmchen, G. Eur.
J. Inorg. Chem. 2002, 2569-2586.
(36) Helmchen, G.; Kudis, S.; Sennhenn, P.; Steinhagen, H. Pure
Appl. Chem. 1997, 69, 513-518.
(40) Bartels, B.; Helmchen, G. Chem. Commun. 1999, 741-742.
(41) Garc´ıa-Yebra, C.; Janssen, J. P.; Rominger, F.; Helmchen, G.
Organometallics 2004, 23, 5459-5470.
(42) Hayashi, T.; Okada, A.; Suzuka, T.; Kawatsura, M. Org. Lett.
2003, 5, 1713-1715.
(37) The hydroxy group is in fact interacting with the palladium
from beneath because of extensive twisting of the ligand backbone.
See the Supporting Information for XYZ coordinates.
J. Org. Chem, Vol. 70, No. 24, 2005 9887

Fro¨lander et al.
TABLE 5. Reactions of 28, 29, or 30 with Dimethyl Malonate in THF at 65 °C
ee (%)
(abs conf)
entry
ligand
substrate
base
product
conversion (%)
b/l
1
2
7 (OH)
8 (OMe)
7 (OH)
8 (OMe)
7 (OH)
8 (OMe)
7 (OH)
8 (OMe)
7 (OH)
8 (OMe)
28
28
28
28
29
29
29
29
30
30
Cs2CO3
Cs2CO3
NaH
31
31
31
31
31
31
31
31
32
32
14 (5 days)
98 (3 days)
67 (5 days)
97 (2 days)
92 (3 days)
98 (3 days)
93 (5 days)
100 (2 days)
100 (20 h)
100 (20 h)
0.6
0.6
0.8
2.4
0.5
4.7
1.4
4.9
1.7
7.5
12 (R)
90 (R)
3 (S)
82 (R)
1 (S)
68 (R)
6 (S)
29 (R)
6 (R)
73 (R)
3
4
NaH
5
Cs2CO3
Cs2CO3
NaH
NaH
Cs2CO3
Cs2CO3
6
7
8
9
10
[(R,R)-2-(2-Diphenylphosphinophenyl)-5-methyl-4,5-
dihydrooxazol-4-yl]methanol (1). To a solution of 12 (1.04
g, 4.98 mmol) and dihydropyrane (2.0 g, 23.8 mmol) in CH2-
Cl2 (15 mL) was added p-toluenesulfonic acid (950 mg, 5.0
mmol) in portions to prevent boiling of the solvent. The
reaction was monitored by TLC. After 0.5 h, the resulting
mixture was washed with 2 M NaOH. The aqueous phase was
extracted with CH2Cl2, and the combined organic phases were
dried over Na2SO4, filtered, and concentrated. The product was
purified by flash chromatography on silica gel (eluent: EtOAc/
hexane 1:1) to afford (R,R)-2-(2-fluorophenyl)-5-methyl-4-(tet-
rahydropyran-2-yloxymethyl)-4,5-dihydrooxazole 13 (1.202 g,
82%, mixture of diastereomers) as a slightly yellow oil, which
was used directly in the next step.
SCHEME 6
A solution prepared from diphenylphosphine (1.12 g, 5.75
mmol) and BuLi (2.2 mL of a 2.5 M solution, 5.5 mmol) in THF
(7 mL) was added to a solution of 13 (1.202 g, 4.10 mmol) in
THF (10 mL) at -78 °C. The resulting mixture was stirred at
-78 °C for 0.5 h, and then allowed to warm to -5 °C before
addition of MeOH (1 mL). The solution was diluted with
diethyl ether, washed with water and then brine, and dried
over Na2SO4. After evaporation of the solvent, the residue was
purified by flash chromatography on silica gel (eluent: EtOAc/
hexane 1:2) to afford (R,R)-2-(2-diphenylphosphinophenyl)-5-
methyl-4-(tetrahydropyran-2-yloxymethyl)-4,5-dihydroox-
azole 14 (1.18 g, 63%, mixture of diastereomers) as a colorless
oil, which was used directly in the next step.
Compound 14 (1.18 g, 2.57 mmol) was dissolved in MeOH
(10 mL) containing p-toluenesulfonic acid (488 mg, 2.57 mmol).
After the reaction was completed (TLC), the solution was
diluted with CH2Cl2, and washed with 2 M NaOH. The
aqueous layer was extracted with CH2Cl2, and the combined
organic phases were dried over Na2SO4. After filtration and
concentration, the residue was purified by flash chromatog-
raphy on silica gel (eluent: EtOAc/hexane 2:1) to afford a
yellowish foam. This foam was dissolved in boiling hexane (15
mL), and the formed solution was allowed to cool slowly. At
50 °C, diethyl ether (5 mL) was added. The resulting mixture
was slowly cooled to 0 °C with the aid of an ice bath. The
precipitated solid was separated, washed with hexane, and
dried to afford 1 (835 mg, 87%) as a white solid: mp 115 °C;
[R]20D +142 (c 1.22, CH2Cl2); 1H NMR (CDCl3) δ 7.89 (1H, ddd,
J ) 1.3, 3.7, 7.6 Hz), 7.42-7.24 (12H, m), 6.92 (1H, ddd, J )
0.8, 4.1, 7.6 Hz), 4.51 (1H, dq, J ) 6.4, 6.4 Hz), 3.83 (1H, ddd,
J ) 3.5, 3.5, 6.8 Hz), 3.65 (1H, d, J ) 11.4 Hz), 3.29 (1H, m),
1.62 (1H, br s), 1.33 (3H, d, J ) 6.4 Hz); 13C NMR (CDCl3) δ
163.6 (d, J(C,P) ) 2.9 Hz), 138.7 (d, J(C,P) ) 21.8 Hz), 138.4
(d, J(C,P) ) 7.3 Hz), 137.4 (d, J(C,P) ) 9.7 Hz), 134.4 (d, J(C,P)
) 20.9 Hz), 134.1, 133.2 (d, J(C,P) ) 19.7 Hz), 131.7 (d, J(C,P)
) 19.8 Hz), 130.6, 129.6 (d, J(C,P) ) 2.6 Hz), 128.8, 128.63,
128.56 (d, J(C,P) ) 7.5 Hz), 128.5 (d, J(C,P) ) 7.4 Hz), 128.2,
77.5, 75.3, 63.9, 20.6; 31P NMR (CDCl3) δ -7.0. Anal. Calcd
for C23H22NO2P: C, 73.59; H, 5.91; P, 8.25. Found: C, 73.34;
H, 5.90; P, 8.05.
1-phenyl)methyl substituents in the 4 position of the
oxazoline ring result in different enantioselectivities.
These differences are assumed to originate in the differ-
ent conformations of the two types of ligands. The
conformation of the methoxy-containing ligand results in
an exo-allyl complex being the most stable, leading to
products with S absolute configuration. For the hydroxy-
containing ligand, two Pd(0)-olefin complexes, originat-
ing from exo- and endo-allyl complexes, respectively, were
found to be of similar energy, which explains the low
enantioselectivity observed under certain conditions.
Iridium-catalyzed allylic alkylations employing the two
types of ligands also resulted in different enantioselec-
tivities. Also here, a hydrogen bond with Ir(I) serving as
the proton acceptor is assumed to be involved, but the
results are more difficult to rationalize because of insuf-
ficient knowledge of the mechanism of this reaction.
Experimental Section
[(R,R)-2-(2-Fluorophenyl)-5-methyl-4,5-dihydrooxazol-
4-yl]methanol (12). (R,R)-Threoninol (9, 2.00 g, 19.0 mmol),
2-fluorobenzonitrile (11, 2.32 g, 19.2 mmol), and cadmium
acetate dihydrate (250 mg, 0.94 mmol) were stirred in chlo-
robenzene (25 mL) at reflux for 16 h. The reaction mixture
was concentrated and purified by flash chromatography on
silica gel (EtOAc/CHCl3 continuous gradient: 0-50% CHCl3).
The solid obtained was crystallized from hexane/EtOAc to give
12 (1.1 g, 28%) as a white solid: mp 91 °C; [R]20D +76 (c 1.00,
CH2Cl2); 1H NMR (CDCl3) δ 7.81 (1H, m), 7.43 (1H, m), 7.07-
7.18 (2H, m), 4.67 (1H, m), 3.89-3.97 (2H, m), 3.66 (1H, m),
2.96 (1H, br s), 1.45 (3H, dd, J ) 2.4, 6.2 Hz); 13C NMR (CDCl3)
δ 161.3 (d, J(C,F) ) 5.5 Hz), 161.1 (d, J(C,F) ) 258.2 Hz), 133.0
(d, J(C,F) ) 8.9 Hz), 131.0, 123.9 (d, J(C,F) ) 3.8 Hz), 116.6
(d, J(C,F) ) 21.9 Hz), 115.8 (d, J(C,F) ) 5.5 Hz), 77.8, 75.1,
63.8, 20.6.
[(R,R)-2-(2-Iodophenyl)-5-methyl-4,5-dihydrooxazol-4-
yl]methanol (16a). (R,R)-Threoninol (9, 1.037 g, 9.88 mmol),
2-iodobenzonitrile (15, 2.6 g, 11.35 mmol), and cadmium
9888 J. Org. Chem., Vol. 70, No. 24, 2005

Phosphino-4-(1-hydroxyalkyl)oxazoline-Metal-Olefin Complexes
acetate dihydrate (133 mg, 0.5 mmol) were dissolved in
chlorobenzene (15 mL) and stirred at reflux for 16 h. The
concentrated reaction mixture was purified by flash chroma-
tography on silica gel (eluent: EtOAc), yielding 16a (1.236 g,
Hz), 138.7 (d, J(C,P) ) 25.5 Hz), 137.84 (d, J(C,P) ) 9.8 Hz),
137.83 (d, J(C,P) ) 12.7 Hz), 134.4 (d, J(C,P) ) 21.4 Hz), 133.8
(d, J(C,P) ) 20.7 Hz), 133.3 (d, J(C,P) ) 2.2 Hz), 131.4 (d,
J(C,P) ) 17.7 Hz), 130.4, 129.8 (d, J(C,P) ) 2.4 Hz), 128.6,
128.39, 128.37 (d, J(C,P) ) 7.2 Hz), 128.3 (d, J(C,P) ) 7.6 Hz),
127.8, 79.5, 74.1, 72.5, 58.9, 20.5; 31P NMR (CDCl3) δ -4.3.
Anal. Calcd for C24H24NO2P: C, 74.02; H, 6.21; P, 7.95.
Found: C, 73.92; H, 6.21; P, 7.79.
40%) as a white solid: mp 66 °C; [R]20 +63 (c 1.20, CH2Cl2);
D
1H NMR (CDCl3) δ 7.91 (1H, dd, J ) 0.8, 8.0 Hz), 7.58 (1H,
dd, J ) 1.6, 7.7 Hz), 7.37 (1H, ddd, J ) 1.1, 7.6, 7.6 Hz), 7.11
(1H, ddd, J ) 1.7, 7.8, 7.8 Hz), 4.68 (1H, dq, J ) 6.4, 6.4 Hz),
3.95 (1H, m), 3.88 (1H, dd, J ) 4.2, 11.4 Hz), 3.65 (1H, dd, J
) 4.7, 11.4 Hz), 2.67 (1H, br s), 1.47 (3H, d, J ) 6.3 Hz); 13C
NMR (CDCl3) δ 165.4, 140.3, 133.6, 131.7, 130.5, 127.9, 94.7,
78.8, 75.1, 63.9, 20.6.
[(R,R)-2-(2-Diphenylphosphinophenyl)-5-phenyl-4,5-
dihydrooxazol-4-yl]methanol (3).23 Compound 3 was pre-
pared in the same way as 2 starting from diphenylphosphine
(321 mg, 1.72 mmol), 16b (543 mg, 1.43 mmol), Et3N (1.0 mL,
7.17 mmol), CH3CN (10 mL), and Pd(OAc)2 (16.0 mg, 0.071
mmol). Flash chromatography on silica gel (diethyl ether/
[(R,R)-2-(2-Iodophenyl)-5-phenyl-4,5-dihydrooxazol-4-
yl]methanol (16b). Compound 16b was prepared in the same
way as 16a starting from (R,R)-2-amino-1-phenyl-1,3-pro-
panediol (10, 994 mg, 5.95 mmol), 2-iodobenzonitrile (15, 1.439
g, 6.28 mmol), cadmium acetate dihydrate (80 mg, 0.30 mmol),
and chlorobenzene (8 mL). Flash chromatography on silica gel
(eluent: EtOAc/hexane 1:1) yielded 16b (956 mg, 40%) as a
yellow oil: [R]20D -5.0 (c 1.48, CH2Cl2); 1H NMR (CDCl3) δ 7.98
(1H, m), 7.71 (1H, J ) 1.6, 7.7 Hz), 7.46-7.32 (6H, m), 7.15
(1H, ddd, J ) 1.7, 7.8, 8.0 Hz), 5.55 (1H, d, J ) 7.9 Hz), 4.35
(1H, ddd, J ) 3.9, 3.9, 7.8 Hz), 4.10 (1H, dd, J ) 3.7, 11.6 Hz),
3.80 (1H, dd, J ) 3.9, 11.7 Hz), 2.53 (1H, br s); 13C NMR
(CDCl3) δ 165.3, 140.5, 140.0, 133.2, 131.9, 130.6, 128.9, 128.5,
127.9, 125.9, 95.0, 83.3, 76.9, 63.7.
hexane 2:3) provided 3 (484 mg, 78%) as a white solid: mp 48
1
°C; [R]20 +77 (c 1.26, CH2Cl2); H NMR (CDCl3) δ 7.99 (1H,
D
ddd, J ) 1.5, 3.7, 7.6 Hz), 7.44-7.26 (17H, m), 6.97 (1H, ddd,
J ) 1.3, 4.0, 7.7 Hz), 5.34 (1H, d, J ) 7.4 Hz), 4.20 (1H, ddd,
J ) 3.1, 3.1, 7.4 Hz), 3.79 (1H, m), 3.44 (1H, m), 1.53 (1H, br
s); 13C NMR (CDCl3) δ 163.7 (d, J(C,P) ) 3.0 Hz), 140.5, 139.1
(d, J(C,P) ) 22.7 Hz), 138.5 (d, J(C,P) ) 7.5 Hz), 137.5 (d,
J(C,P) ) 9.7 Hz), 134.5 (d, J(C,P) ) 21.0 Hz), 134.3, 133.3 (d,
J(C,P) ) 19.7 Hz), 131.3 (d, J(C,P) ) 19.5 Hz), 130.8, 129.8
(d, J(C,P) ) 2.7 Hz), 128.9, 128.8, 128.7, 128.62 (d, J(C,P) )
7.5 Hz), 128.61 (d, J(C,P) ) 7.4 Hz), 128.33, 128.30, 125.8, 82.3,
77.2, 63.7; 31P NMR (CDCl3) δ -5.9.
(R,R)-2-(2-Iodophenyl)-4-methoxymethyl-5-methyl-4,5-
dihydrooxazole (17a). NaH (60% in oil, 187 mg, 4.67 mmol)
was suspended in THF (20 mL). 16a (1.187 g, 3.74 mmol) was
added in one portion. After hydrogen evaluation stopped,
dimethyl sulfate (533.2 mg, 4.23 mmol) was added dropwise.
The mixture was stirred at room temperature for 30 min,
whereafter diethyl ether (20 mL) and 25% aqueous ammonia
solution (10 mL) were added. The organic phase was separated,
washed with 2 M NaOH, water, and brine, and dried over Na2-
SO4. After evaporation of the solvent, the residue was purified
by flash chromatography on silica gel (eluent: EtOAc/hexane
1:1) to give 17a (1.075 g, 87%) as a colorless oil: [R]20D +44 (c
1.02, CH2Cl2); 1H NMR (CDCl3) δ 7.91 (1H, m), 7.59 (1H, ddd,
J ) 1.8, 7.7, 7.7 Hz), 7.35 (1H, m), 7.09 (1H, m), 4.67 (1H,
ddq, J ) 2.0, 6.3, 6.3 Hz), 4.01 (1H, dddd, J ) 2.2, 4.6, 7.1, 7.1
Hz), 3.69 (1H, ddd, J ) 2.2 Hz, 4.5, 9.4 Hz), 3.39-3.47 (4H,
m), 1.46 (3H, dd, J ) 2.2, 6.3 Hz); 13C NMR (CDCl3) δ 164.6,
140.3, 133.9, 131.5, 130.6, 127.7, 94.5, 80.1, 74.4, 73.1, 59.3,
20.9.
(R,R)-2-(2-Diphenylphosphinophenyl)-4-methoxyme-
thyl-5-phenyl-4,5-dihydrooxazole (4).24 Compound 4 was
prepared in the same way as 2 starting from diphenylphos-
phine (375 mg, 2.01 mmol), 17b (682 mg, 1.735 mmol), Et3N
(1.5 mL, 10.8 mmol), CH3CN (10 mL), and Pd(OAc)2 (19.5 mg,
0.087 mmol). Flash chromatography on silica gel (EtOAc/
hexane 1:3) afforded 4 (639 mg, 82%) as a viscous colorless
1
oil: [R]20 -40 (c 1.18, CH2Cl2); H NMR (CDCl3) δ 7.92 (1H,
D
ddd, J ) 1.0, 3.4, 7.5 Hz), 7.33-7.15 (15H, m), 7.08 (2H, dd, J
) 1.5, 8.0 Hz), 6.80 (1H, ddd, J ) 1.0, 4.2, 7.5 Hz), 5.10 (1H,
d, J ) 7.0 Hz), 4.10 (1H, dt, J ) 4.6, 7.4 Hz), 3.34 (1H, dd, J
) 4.5, 9.7 Hz), 3.18 (3H, s), 2.86 (1H, dd, J ) 7.9, 9.6 Hz); 13
C
NMR (CDCl3) δ 163.6 (d, J(C,P) ) 3.7 Hz), 140.6, 139.4 (d,
J(C,P) ) 26.7 Hz), 138.1 (d, J(C,P) ) 12.4 Hz), 138.1 (d, J(C,P)
) 10.9 Hz), 134.5 (d, J(C,P) ) 21.4 Hz), 133.8 (d, J(C,P) )
20.5 Hz), 133.7 (d, J(C,P) ) 1.5 Hz), 131.1 (d, J(C,P) ) 17.8
Hz), 130.7, 130.2 (d, J(C,P) ) 2.3 Hz), 128.7, 128.52 (d, J(C,P)
) 7.0 Hz), 128.51, 128.46, 128.4 (d, J(C,P) ) 7.4 Hz), 128.0,
127.9, 125.7, 83.7, 74.6, 74.1, 59.0; 31P NMR (CDCl3) δ -4.4.
(R,R)-2-(2-Iodophenyl)-4-methoxymethyl-5-phenyl-4,5-
dihydrooxazole (17b). Compound 17b was prepared in the
same way as 17a starting from NaH (60% in oil, 150 mg, 3.75
mmol), THF (15 mL), 16b (853 mg, 2.25 mmol), and dimethyl
sulfate (333.3 mg, 2.64 mmol). Flash chromatography on silica
gel (eluent: EtOAc/hexane 2:1) provided 17b (795 mg, 90%)
as a colorless oil: [R]20D -17 (c 1.16, CH2Cl2); 1H NMR (CDCl3)
δ 7.97 (1H, dd, J ) 0.9, 8.0 Hz), 7.73 (1H, dd, J ) 1.6, 7.7 Hz),
7.30-7.46 (6H, m), 7.13 (1H, ddd, J ) 1.7, 7.8, 7.9 Hz), 5.53
(1H, d, J ) 7.1 Hz), 4.41 (1H, ddd, J ) 4.5, 6.9, 7.0 Hz), 3.80
(1H, dd, J ) 4.4, 9.7 Hz), 3.66 (1H, dd, J ) 6.7, 9.7 Hz), 3.46
(3H, s); 13C NMR (CDCl3) δ 164.4, 140.5, 140.4, 133.3, 131.7,
130.8, 128.7, 128.2, 127.8, 125.8, 94.7, 84.2, 75.0, 74.2, 59.3.
(R,R)-1-[2-(2-Iodophenyl)-4,5-dihydrooxazol-4-yl]etha-
nol (19a). A solution of 2-iodobenzoyl chloride (18, 1.704 mg,
6.4 mmol) in CH2Cl2 (6 mL) was added dropwise at 0 °C to a
solution of (R,R)-threoninol (9, 631 mg, 6.0 mmol), and Et3N
(3.0 mL, 21.5 mmol) in CH2Cl2 (14 mL), and the mixture was
stirred at room temperature for 0.5 h. TsCl (1.37 mg, 7.2 mmol)
was added, and the solution was brought to reflux for 16 h.
The solvent was removed under reduced pressure, and the
residue was purified by flash chromatography on silica gel
(eluent: EtOAc) to afford 19a (841 mg, 44%) as a colorless
1
oil: [R]20 -90 (c 0.77, CH2Cl2); H NMR (CDCl3) δ 7.96 (1H,
D
dd, J ) 1.0, 8.0 Hz), 7.64 (1H, dd, J ) 1.7, 7.7 Hz), 7.40 (1H,
dt, J ) 1.1, 7.6 Hz), 7.13 (1H, dt, J ) 1.7, 7.7 Hz), 4.51 (1H,
m), 4.32-4.22 (2H, m), 3.80 (1H, m), 2.37 (1H, d, J ) 5.6 Hz),
1.35 (3H, d, J ) 6.3 Hz); 13C NMR (CDCl3) δ 165.3, 140.5,
133.1, 131.8, 130.5, 127.8, 94.9, 73.1, 70.0, 69.7, 19.6.
(R,R)-2-(2-Diphenylphosphinophenyl)-4-methoxyme-
thyl-5-methyl-4,5-dihydrooxazole (2). Diphenylphosphine
(642 mg, 3.45 mmol), 17a (1.025 g, 3.1 mmol), and Et3N (2.5
mL, 17.9 mmol) were dissolved in CH3CN (20 mL). Pd(OAc)2
(34.0 mg, 0.15 mmol, 5 mol %) was added, and the solution
was brought to reflux for 4 h. The mixture was absorbed on
silica, and purified by flash chromatography on silica gel
(eluent: diethyl ether/hexane 1:3) to yield 2 (875 mg, 73%) as
(R,R)-[2-(2-Iodophenyl)-4,5-dihydrooxazol-4-yl]phenyl-
methanol (19b). A solution of 2-iodobenzoyl chloride (18, 847
mg, 3.2 mmol) in CH2Cl2 (2 mL) was added dropwise at 0 °C
to a solution of (R,R)-2-amino-1-phenyl-1,3-propanediol (10,
501 mg, 3.0 mmol) and Et3N (1.5 mL, 10.8 mmol) in CH2Cl2
(7 mL), and the mixture was stirred at room temperature for
0.5 h. DMAP (12.2 mg, 0.1 mmol) was added followed by TsCl
(686 mg, 3.6 mmol), and the solution was brought to reflux
for 16 h. The reaction mixture was diluted with EtOAc, washed
with 0.02 M HCl, water, 2 M NaOH, and brine, and dried over
Na2SO4. After filtration and evaporation of the solvent, the
a slightly yellow oil: [R]20 +7.1 (c 1.46, CH2Cl2); 1H NMR
D
(CDCl3) δ 7.89 (1H, ddd, J ) 1.3, 3.5, 7.5 Hz), 7.37-7.25 (12H,
m), 6.84 (1H, ddd, J ) 1.0, 4.3, 7.6 Hz), 4.32 (1H, dq, J ) 6.3,
6.3 Hz), 3.77 (1H, ddd, J ) 4.6, 6.7, 8.2 Hz), 3.36 (1H, dd, J )
4.6, 9.5 Hz), 3.25 (3H, s), 2.80 (1H, dd, J ) 8.6, 9.2 Hz), 1.20
(3H, d, J ) 6.3 Hz); 13C NMR (CDCl3) δ 163.8 (d, J(C,P) ) 3.1
J. Org. Chem, Vol. 70, No. 24, 2005 9889

Fro¨lander et al.
residue was purified by flash chromatography on silica gel
(eluent: EtOAc/hexane 1:1) followed by recrystallization from
EtOAc/hexane to afford 19b (668 mg, 59%) as a white solid:
mp 104 °C; [R]20D -78 (c 2.16, CH2Cl2); 1H NMR (CDCl3) δ 7.97
(1H, d, J ) 8.0 Hz), 7.63 (1H, dd, J ) 1.6, 7.7 Hz), 7.46 (2H,
m), 7.42-7.29 (4H, m), 7.13 (1H, dt, J ) 1.6, 7.8 Hz), 4.61 (2H,
m), 4.31 (1H, m), 4.21 (1H, m), 3.43 (1H, br s); 13C NMR
(CDCl3) δ 165.6, 140.6, 139.9, 132.8, 131.9, 130.6, 128.6, 128.3,
127.9, 127.0, 94.9, 77.2, 73.4, 69.6.
(R,R)-2-(2-Iodophenyl)-4-(1-methoxyethyl)-4,5-dihy-
drooxazole (20a). NaH (60% in mineral oil, 37.84 mg, 0.946
mmol, washed three times with hexane) was suspended in
THF (5 mL) at 0 °C. 19a (200 mg, 0.63 mmol) dissolved in
THF (7 mL) was added, and the mixture was stirred at 0 °C
for 1 h. MeI (58.9 µL, 0.946 mmol) was added dropwise, and
the solution was stirred in a thawing ice bath for 16 h. The
solution was diluted with diethyl ether (20 mL) and washed
with H2O (20 mL) and sat. NH4Cl (20 mL). The combined
water phases were washed with diethyl ether (3 × 20 mL),
and the combined organic phases were concentrated and dried
over Na2SO4 to afford 20a (192.9 mg, 92%) as a yellowish oil:
[R]20D -40 (c 0.71, CH2Cl2); 1H NMR (CDCl3) δ 7.93 (1H, dd, J
) 0.9, 7.9 Hz), 7.63 (1H, dd, J ) 1.7, 7.7 Hz), 7.37 (1H, dt, J
) 1.1, 7.6 Hz), 7.11 (1H, dt, J ) 1.7, 7.7 Hz), 4.60 (1H, m),
4.40 (2H, m), 3.71 (1H, dq, J ) 4.8, 6.3 Hz), 3.42 (3H, s), 1.21
(3H, d, J ) 6.3 Hz); 13C NMR (CDCl3) δ 165.0, 140.5, 133.6,
131.6, 130.8, 127.8, 94.5, 77.3, 69.8, 68.5, 56.8, 14.0.
(R,R)-2-(2-Iodophenyl)-4-(methoxyphenylmethyl)-4,5-
dihydrooxazole (20b). Compound 20b was prepared in the
same way as 17a starting from NaH (60% in oil, 130 mg, 3.25
mmol), THF (10 mL), 19b (984 mg, 2.6 mmol), and dimethyl
sulfate (986.4 mg, 7.82 mmol). Flash chromatography on silica
gel (eluent: EtOAc/hexane 2:3) provided 20b (914 mg, 90%)
as a yellowish oil: [R]20D -75 (c 0.82, CH2Cl2); 1H NMR (CDCl3)
δ 7.88 (1H, m), 7.48-7.26 (7H, m), 7.07 (1H, dt, J ) 1.8, 7.7
Hz), 4.75 (1H, m), 4.42 (1H, dd, J ) 2.0, 6.1 Hz), 4.25 (2H, m),
3.34 (3H, m); 13C NMR (CDCl3) δ 165.4, 140.1, 137.4, 133.7,
131.5, 130.7, 128.3, 128.1, 127.8, 127.6, 94.4, 84.9, 71.3, 69.0,
57.1.
(d, J(C,P) ) 7.6 Hz), 128.0, 77.2, 69.5, 67.6, 56.6, 13.3; 31P NMR
(CDCl3) δ -4.8.
[(R,R)-2-(2-Diphenylphosphinophenyl)-4,5-dihydroox-
azol-4-yl]phenylmethanol (7).22 Compound 7 was prepared
in the same way as 2 starting from diphenylphosphine (267.5
mg, 1.44 mmol), 19b (453 mg, 1.2 mmol), Et3N (1.0 mL, 7.2
mmol), CH3CN (10 mL), and Pd(OAc)2 (13.6 mg, 0.060 mmol).
Flash chromatography on silica gel (eluent: EtOAc/hexane 1:2)
afforded 7 (392 mg, 75%) as a white solid: mp 109 °C; [R]20
D
1
-94 (c 0.90, CH2Cl2); H NMR (CDCl3) δ 7.93 (1H, ddd, J )
1.3, 3.7, 7.6 Hz), 7.41 (1H, ddd, J ) 1.1, 7.6, 7.6 Hz), 7.39-
7.24 (16H, m), 6.95 (1H, ddd, J ) 0.9, 3.9, 7.7 Hz), 4.47 (1H,
ddd, J ) 7.5, 7.5, 9.6 Hz), 4.17 (1H, m), 4.11 (1H, dd, J ) 2.2,
7.5 Hz), 4.00 (1H, m), 2.88 (1H, br s); 13C NMR (CDCl3) δ 164.3
(d, J(C,P) ) 1.8 Hz), 140.1, 138.9 (d, J(C,P) ) 23.7 Hz), 138.2
(d, J(C,P) ) 8.1 Hz), 137.7 (d, J(C,P) ) 10.3 Hz), 134.4, 134.2
(d, J(C,P) ) 20.4 Hz), 133.4 (d, J(C,P) ) 19.8 Hz), 131.4 (d,
J(C,P) ) 19.8 Hz), 130.9, 129.7 (d, J(C,P) ) 2.8 Hz), 128.7,
128.7, 128.6, 128.5 (d, J(C,P) ) 8.1 Hz), 128.5 (d, J(C,P) ) 8.1
Hz), 128.3, 128.1, 127.1, 77.4, 73.5, 69.4; 31P NMR (CDCl3) δ
-5.8.
(R,R)-2-(2-Diphenylphosphinophenyl)-4-(methoxyphe-
nylmethyl)-4,5-dihydrooxazole (8). Compound 8 was pre-
pared in the same way as 2 starting from diphenylphosphine
(535.0 mg, 2.87 mmol), 20b (970 mg, 2.47 mmol), Et3N (2.0
mL, 14.4 mmol), CH3CN (15 mL), and Pd(OAc)2 (27 mg, 0.12
mmol). Flash chromatography on silica gel (eluent: diethyl
ether/hexane 1:3) afforded 8 (910 mg, 82%) as a white solid:
mp 48 °C; [R]20D +16 (c 0.61, CH2Cl2); 1H NMR (CDCl3) δ 7.82
(1H, ddd, J ) 1.3, 3.2, 7.1 Hz), 7.43-7.22 (15H, m), 7.18 (2H,
m), 6.91 (1H, m), 4.46 (1H, m), 3.90 (1H, d, J ) 6.7 Hz), 3.85
(1H, t, J ) 9.3 Hz), 3.75 (1H, t, J ) 8.6 Hz), 3.23 (3H, s); 13C
NMR (CDCl3) δ 165.1 (d, J(C,P) ) 1.7 Hz), 138.5 (d, J(C,P) )
24.8 Hz), 137.9 (d, J(C,P) ) 11.9 Hz), 137.6, 137.5 (d, J(C,P)
) 9.9 Hz), 134.11 (d, J(C,P) ) 21.1 Hz), 134.06 (d, J(C,P) )
21.0 Hz), 133.4 (d, J(C,P) ) 1.7 Hz), 131.7 (d, J(C,P) ) 18.9
Hz), 130.3, 130.1 (d, J(C,P) ) 3.0 Hz), 128.59, 128.57, 128.4
(d, J(C,P) ) 7.1 Hz), 128.3 (d, J(C,P) ) 7.0 Hz), 128.1, 127.9,
127.8, 127.6, 85.0, 71.1, 68.4, 56.8; 31P NMR (CDCl3) δ -4.7.
Anal. Calcd for C29H26NO2P: C, 77.15; H, 5.80; P, 6.86.
Found: C, 76.95; H, 5.92; P, 6.65.
General Procedure for the Palladium-Catalyzed Al-
lylic Alkylations. Ligand (0.012 mmol, 2.4 mol %), [(η3-C3H5)-
PdCl]2 (1.83 mg, 0.0050 mmol, 2.0 mol % Pd), and substrate
(0.50 mmol) were dissolved in solvent (1 mL), and stirred for
1 h at room temperature. A solution/suspension of the nucleo-
phile was prepared by adding acetylacetone/dimethyl malonate
(0.70 mmol) dropwise to a suspension of NaH (60% in mineral
oil, 24.0 mg, 0.60 mmol) in 5 mL of solvent. The two solutions
were mixed at -78 °C, and stirred at the reported reaction
temperature for the appropriate time. The samples taken were
filtered through silica using EtOAc. The reaction mixture was
purified by flash chromatography on silica gel (eluent: EtOAc/
hexane 1:6).
(E)-2-(1,3-Diphenylprop-2-enyl)malonic Acid Dimethyl
Ester (25). Conversions and enantiomeric excesses were
determined by a chiral OD-H (0.46 cm Ø × 25 cm) HPLC
column (eluent: degassed 2-propanol/hexane 1:99), with a flow
rate of 0.5 mL/min, UV detection at 254 nm, tR(R) ) 21.3 min,
tR(S) ) 22.9 min. Absolute configurations were assigned by
comparing the sign of the optical rotation with literature
data.19
(R,R)-1-[2-(2-Diphenylphosphinophenyl)-4,5-dihydroox-
azol-4-yl]ethanol (5). Compound 5 was prepared in the same
way as 2 starting from diphenylphosphine (156.4 mg, 0.70
mmol), 19a (222.0 mg, 0.84 mmol), Et3N (0.59 mL, 4.2 mmol),
CH3CN (5 mL), and Pd(OAc)2 (7.86 mg, 0.035 mmol). Flash
chromatography on silica gel (eluent: EtOAc/hexane 1:2)
provided 5 (392 mg, 75%) as a colorless oil: [R]20D -110 (c 1.41,
1
CH2Cl2); H NMR (CDCl3) δ 7.91 (1H, ddd, J ) 1.1, 3.6, 7.6
Hz), 7.40 (1H, dt, J ) 1.0, 7.6 Hz), 7.37-7.22 (11H, m), 6.93
(1H, m), 4.34 (1H, dd, J ) 8.0, 9.5 Hz), 4.13 (1H, m), 4.05 (1H,
t, J ) 7.9 Hz), 3.39 (1H, m), 2.08 (1H, m), 1.11 (3H, d, J ) 6.3
Hz); 13C NMR (CDCl3) δ 163.9, 138.6 (d, J(C,P) ) 22.8 Hz),
138.3 (d, J(C,P) ) 8.0 Hz), 137.8 (d, J(C,P) ) 9.3 Hz), 134.5,
134.2 (d, J(C,P) ) 20.5 Hz), 133.3 (d, J(C,P) ) 19.7 Hz), 131.6
(d, J(C,P) ) 20.0 Hz), 130.7, 129.6, 128.64, 128.57, 128.51 (d,
J(C,P) ) 6.9 Hz), 128.50 (d, J(C,P) ) 6.9 Hz), 128.3, 73.3, 70.1,
69.6, 19.9; 31P NMR (CDCl3) δ -5.8.
(R,R)-2-(2-Diphenylphosphinophenyl)-4-(1-methoxy-
ethyl)-4,5-dihydrooxazole (6). Compound 6 was prepared
in the same way as 2 starting from diphenylphosphine (45.0
mg, 0.242 mmol), 20a (222.0 mg, 0.201 mmol), Et3N (0.17 mL,
1.21 mmol), CH3CN (2 mL), and Pd(OAc)2 (2.26 mg, 0.010
mmol). Flash chromatography on silica gel (eluent: EtOAc/
(E)-3-(1,3-Diphenylprop-2-enyl)pentane-2,4-dione (26).
Conversions and enantiomeric excesses were determined by
a chiral OJ (0.46 cm Ø × 25 cm) HPLC column (eluent:
degassed 2-propanol/hexane 5:95), with a flow rate of 0.5 mL/
min, UV detection at 254 nm, tR(R) ) 37.0 min, tR(S) ) 43.9
min. Absolute configurations were assigned by comparing the
sign of the optical rotation with literature data.43
hexane 1:9) provided 6 (48.4 mg, 62%) as a slightly yellowish
1
oil: [R]20 -11 (c 1.10, CH2Cl2); H NMR (CDCl3) δ 7.91 (1H,
D
ddd, J ) 1.5, 3.6, 7.6 Hz), 7.38-7.26 (12H, m), 6.88 (1H, ddd,
J ) 1.2, 4.1, 7.7 Hz), 4.36 (1H, ddd, J ) 4.7, 7.7, 10.0 Hz),
4.08 (2H, m), 3.41 (1H, m), 3.29 (3H, s), 0.75 (3H, d, J ) 6.3
Hz); 13C NMR (CDCl3) δ 164.3 (d, J(C,P) ) 2.8 Hz), 139.0 (d,
J(C,P) ) 25.6 Hz), 138.1 (d, J(C,P) ) 12.6 Hz), 137.9 (d, J(C,P)
) 9.7 Hz), 134.3 (d, J(C,P) ) 21.1 Hz), 133.83 (d, J(C,P) )
20.7 Hz), 133.79, 131.5 (d, J(C,P) ) 19.0 Hz), 130.5, 130.1 (d,
J(C,P) ) 2.7 Hz), 128.6, 128.5, 128.4 (d, J(C,P) ) 6.9 Hz), 128.3
(43) Hayashi, T.; Yamamoto, A.; Hagihara, T.; Ito, Y. Tetrahedron
Lett. 1986, 27, 191-194.
9890 J. Org. Chem., Vol. 70, No. 24, 2005

Phosphino-4-(1-hydroxyalkyl)oxazoline-Metal-Olefin Complexes
2-(3-Cyclohexenyl)malonic Acid Dimethyl Ester 27.
Conversions were determined by GC-MS. Enantiomeric ex-
cesses were determined by a chiral OD-H (0.46 cm Ø × 25
cm) HPLC column (eluent: degassed 2-propanol/hexane 0.25:
99.75), with a flow rate of 0.5 mL/min, UV detection at 220
nm, tR(R) ) 30.5 min, tR(S) ) 32.0 min. Absolute configurations
were assigned by comparing the sign of the optical rotation
with literature data.44
determined by GC-MS. Enantiomeric excesses were deter-
mined by a chiral OD-H (0.46 cm Ø × 25 cm) HPLC column
(eluent: degassed 2-propanol/hexane 0.25:99.75), with a flow
rate of 0.5 mL/min, UV detection at 220 nm, tR(R) ) 44.4 min,
tR(S) ) 51.7 min. Absolute configurations were assigned by
comparing the sign of the optical rotation with literature
data.45
1-(4-Methoxyphenyl)-2-propenyl Malonic Acid Dimeth-
yl Ester 32a and (E)-3-(4-Methoxyphenyl)-2-propenyl
Malonic Acid Dimethyl Ester 32b. Conversions and
branched:linear ratios were determined by GC-MS. Enantio-
meric excesses were determined by a chiral OD-H (0.46 cm Ø
× 25 cm) HPLC column (eluent: degassed 2-propanol/hexane
5:95), with a flow rate of 0.5 mL/min, UV detection at 220 nm,
tR(R) ) 13.5 min, tR(S) ) 15.0 min. Absolute configurations
were assigned by comparing the sign of the optical rotation
with literature data.46
Computations. The following computational methodology
was applied. First, geometry optimizations of structures were
carried out using B3LYP functional25 with the lacvp**/6-31G-
(d,p) basis set.27,28 In the second step, B3LYP energies were
evaluated for the optimized geometry using the much larger
triple-ú basis, lacv3p**++/6-311+G(d,p), with additional dif-
fuse and polarization functions. All computations were per-
formed with the Jaguar v4.0 suite of ab initio quantum
chemistry programs.26 Solvent (THF) was represented with the
following parameters: ꢀ ) 7.58, probe radius rp ) 2.524 Å.
Gas-phase optimized structures were used to estimate the
effect of solvent on the reaction energy within the self-
consistent reaction field model as implemented in the Jaguar
computational package.
General Procedure for the Iridium-Catalyzed Allylic
Alkylations Using NaH as Base. Ligand (0.011 mmol, 4.4
mol %), [IrcodCl]2 (3.36 mg, 0.0050 mmol, 4.0 mol % Ir), and
substrate (0.25 mmol) were dissolved in THF (1 mL), and the
solution was stirred for 15 min at room temperature. A solution
of the nucleophile was prepared by adding dimethyl malonate
(62.9 µL, 0.55 mmol) dropwise to a suspension of NaH (60%
in mineral oil, 20.0 mg, 0.50 mmol) in THF (3 mL). The
nucleophile solution was added dropwise to the catalyst
solution at room temperature, and stirred at 65 °C for the
appropriate time. The samples taken were filtered through
silica using EtOAc. The reaction was quenched by the addition
of H2O (10 mL). The product was extracted with diethyl ether
(10 mL), washed with brine (10 mL), dried over MgSO4, and
purified by flash chromatography on silica gel (eluent: EtOAc/
hexane 1:9) to yield a mixture of branched and linear products.
General Procedure for the Iridium-Catalyzed Allylic
Alkylations Using Cs2CO3 as Base. Ligand (0.011 mmol,
4.4 mol %), [IrcodCl]2 (3.36 mg, 0.0050 mmol, 4.0 mol % Ir),
substrate (0.25 mmol), and Cs2CO3 (162.9 mg, 0.50 mmol) were
suspended in THF (2 mL), and stirred for 15 min at room
temperature. Dimethyl malonate (62.9 µL, 0.55 mmol) was
added dropwise at room temperature, and the mixture was
stirred at 65 °C for the appropriate time. The samples taken
were filtered through silica using EtOAc. The reaction was
quenched by the addition of H2O (10 mL). The product was
extracted with diethyl ether (10 mL), washed with brine (10
mL), dried over MgSO4, and purified by flash chromatography
on silica gel (eluent: EtOAc/hexane 1:9) to yield a mixture of
branched and linear products.
Acknowledgment. This work was supported by the
Swedish Research Council.
Supporting Information Available: General experimen-
tal methods.1H NMR spectra of compounds 1-8, 12, 16a, 16b,
17a, 17b, 19a, 19b, 20a and 20b. XYZ coordinates of D2′, D2′′,
F1′, and F2′. This material is available free of charge via the
Internet at http://pubs.acs.org.
1-Phenyl-2-propenyl Malonic Acid Dimethyl Ester 31a
and (E)-3-Phenyl-2-propenyl Malonic Acid Dimethyl
Ester 31b. Conversions and branched:linear ratios were
JO051610Q
(44) Evans, D. A.; Campos, K. R.; Tedrow, J. S.; Michael, F. E.;
Gagne´, M. R. J. Am. Chem. Soc. 2000, 122, 7905-7920 (see the
Supporting Information).
(45) Trost, B. M.; Hachiya, I. J. Am. Chem. Soc. 1998, 120, 1104-
1105.
(46) Alexakis, A.; Polet, D. Org. Lett. 2004, 6, 3529-3532.
J. Org. Chem, Vol. 70, No. 24, 2005 9891