A modular approach to ligands for asymmetric π-allyl palladium catalysed additions

Article abstract of DOI:10.1039/a607200c

A route to phosphinodihydrooxazoles is reported, which is used to synthesize ten new ligands whose ability to asymmetrically catalyse the addition of dimethyl malonate to 1,3-diphenylprop-2-enyl acetate is investigated; the best gave a palladium complex w

Full text of DOI:10.1039/a607200c

View Article Online / Journal Homepage / Table of Contents for this issue
A modular approach to ligands for asymmetric p-allyl palladium catalysed
additions
Scott R. Gilbertson* and Cheng-Wei T. Chang
Department of Chemistry, Washington University, St. Louis, Missouri 63130, USA
A route to phosphinodihydrooxazoles is reported, which is
used to synthesize ten new ligands whose ability to asymmet-
rically catalyse the addition of dimethyl malonate to
1,3-diphenylprop-2-enyl acetate is investigated; the best
gave a palladium complex which catalyses the addition of
dimethyl malonate to 1,3-diphenylprop-2-enyl acetate in
99% yield and 97% ee.
a five-member chelate, gave poorer selectivity than the best
systems.10 Our approach makes ligands available that link the
phosphine and dihydrooxazole groups with a two-carbon alkyl
chain, thus forming a six-member chelate upon metal coordi-
nation and allowing a direct comparison to the six-member
chelates studied previously by others.8–12 The use of an alkyl
chain to connect the chelating functionality allows the introduc-
tion of a second chiral centre, which is next to the phosphine.
This results in diastereomeric ligands in which the two chiral
centres can, potentially, act in concert or dissonance. Results
with both diastereomeric possibilities are reported.
The synthesis of the library of ligands began with modular
building blocks 1, 2 and 3. We chose to use amino acids 3 and
phosphine containing acids (1 and 2) as our modules, allowing
the synthesis of new ligands by amide bond formation. For the
study reported here, we used two phosphine containing acids.
One was a diphenylphosphine derivative of acrylate 1, and the
other a derivative of cinnamic acid 2. The acrylate derivative 1
has been reported earlier as an intermediate in the synthesis of
diphenylphosphinoserine.1,2 The synthesis of the phenyl deriva-
tive 2 was performed in the same manner from cinnamic acid.
The synthesis of the ligands began with the coupling of
phosphine acids 1 or 2 to a given amino acid yielding the
phosphine sulfide amides 4–10 (Scheme 1). The desired ligands
were then obtained by a simple three step procedure. Reduction
of the methyl ester with LiBH4 followed by cyclization with
Burgess reagent14 [MeO2CNSO2NEt3] gave the dihydro-
oxazole phosphine sulfides 11–20. Reaction with Raney nickel
then removed the sulfur and gave the free phosphines. Reaction
of 2 with an amino acid gave two diastereomeric products.
These diastereomers were readily separated by chromatog-
raphy, after LiBH4 reduction, yielding both the (SS) and (RS)
pair of diastereomers. The ligands are fully characterized as the
phosphine sulfides. After reduction of the phosphine sulfides to
the free phosphines the ligands were treated with palladium and
used in catalysis without purification.†
The pace for the development of new asymmetric ligands is
often determined by the availability of new ligand structures.
Hence we have been interested in developing a building block
based approach to the systematic synthesis of new, structurally
varied ligands. Such an approach offers the potential for the
discovery of new synthetically useful metal–ligand complexes,
as well as the optimization of ligand systems that are known to
work to a limited extent. To date, this work has focused on the
synthesis and utility of amino acid based phosphine derivatives,
for the construction of peptide based ligands.1–4 We report here
that the building blocks used in that approach can also be
employed in the synthesis of a variety of phosphino-
dihydrooxazole ligands. This route allows access to new
phosphine ligands without the difficulties often associated with
C–P bond formation. We have used this chemistry to synthesize
ten new phosphinodihydrooxazole ligands. These ligands have
been screened for their ability to control the palladium catalysed
asymmetric addition of malonate to 1,3-diphenylprop-2-enyl
acetate.5–7 A number of successful ligands were discovered
with the best giving, at room temperature, the addition of
malonate with 97% ee. This study also resulted in the
observation of an interesting inverse temperature effect, in
which the best ligand 28 gives higher ees at room temperature
than at 230 °C.
As the starting point for our design we decided to use
phosphine dihydrooxazole ligands that were structurally similar
to those reported by Pfaltz, Helmchen and Williams.8–12 This
type of ligand has proven useful in asymmetric p-allyl
additions8–12 and asymmetric Heck reactions.13 To date, the
most successful ligands of this type have been six-member
chelates with the phosphine attached to the dihydrooxazole
through a phenyl ring. To the best of our knowledge only a
single system with the phosphine attached to the nitrogen ligand
by an alkyl bridge has been reported. That system, which forms
The ligands were investigated in two groups; ligands with
one chiral centre (Table 2) and ligands with two chiral centres
(Table 3). Of the ligands with one chiral centre, the ligands
derived from phenylalanine (R2 = Bn) 22 and valine (R2 = Pri)
23 gave the highest ees (87 and 86%) (Table 2, entries 2 and 3).
Table 1 Synthesis of phospine dihydrooxazole ligands
S
S
H
N
Yield
Compound (%)
Yield
Compound (%)
Ph2P
H2N
CO2Me
Ph2P
CO2Me
R1
R2
Ligand
CO2H
i
+
R1
R2
3
R1
O
R2
H
H
H
H
Me (S)
Bn (S)
Pri (S)
Ph (S)
Bn (S)
4
5
6
7
8
8
9
9
33
31
64
56
69a
69a
62a
62a
47a
47a
11
12
13
14
15
16
17
18
19
20
62
66
72
84
71
85
45
91
48
90
21
22
23
24
25
26
27
28
29
30
1 R1 = H
2 R1 = Ph
4–10
ii, iii
S
R2
Ph2P
R2
Ph (S)
N
O
N
Ph2P
iv
Ph (R) Bn (S)
Ph (S)
Pri (S)
Ph (R) Pri (S)
Ph (S)
Ph (R) Ph (S) 10
R1
R1
11–20
O
ligand 21–30
Ph (S) 10
Scheme 1 Reagents and conditions: i, EDC, HOBT, CH2Cl2, MeCN, room
temp.; ii, LiBH4, THF, 0 °C to room temp.; Burgess reagent, THF, reflux;
iv, Raney Ni, MeCN, 60–80%
a Yields reported for 8–10 are for the mixture of the two diastereomers.
Chem. Commun., 1997
975

View Article Online
The effect of solvent on the reaction was probed using ligand 23
and the Bu4N+/BSA cation base combination (entries 3, 8, 9 and
10).15 Of the solvents tested, CH2Cl2 was found to be the best
solvent for this ligand (entry 3). The counter ion and base
associated with the malonate anion have been shown to be
important in these additions and so their effect was also
investigated (entries 3, 5, 6 and 7). The optimal system for this
class of ligand was found to be (C6H13)4N+/BSA, as the counter
ion and base, and CH2Cl2 as the solvent (entry 6), giving an ee
of 90%.
Of the ligands with two chiral centres, the ligands derived
from valine were again found to perform the best. Both
combinations of the two chiral centres were looked at. It is
interesting to note that the chirality at the carbon next to the
phosphine had little effect on the selectivity of the catalysis, 90
vs. 93% ee for the (SS) vs. (RS) pair (Table 3, entries 3 and 4).
Ligand 28 with a (RS) configuration, was found to be the most
selective ligand (entry 4). The optimal reaction conditions for
this ligand were then determined. The best solvent system for
this ligand was found to be MeCN, and the best counter ion
system was (C6H13)4N+/BSA (entry 14).
In the case of the ligands reported by Pfaltz, where the
phosphine is attached to the dihydrooxazole through a phenyl
ring, a phenyl substituent next to the nitrogen gave the highest
selectivities, with isopropyl only slightly poorer.9 In the system
reported here, the best substituent, in the position next to
nitrogen, was isopropyl while phenyl was found to be the
poorest (ligand 28 vs. 30). The source of the difference between
the two systems is unknown, but probably has its origins in
different conformational preferences for the two complexes.
The positioning of an sp3 hybridized carbon, with a phenyl ring
attached, changes the canting of the phenyl rings on the
phosphine. This effect may be responsible for the difference
between these two ligands.
Through the use of modular building blocks we have
developed a new ligand for palladium catalysed p-allyl
additions. We are currently studying the use of these ligands in
other metal catalysed reactions. We are also using the phosphine
acid building blocks discussed here in the synthesis of other
collections of ligands for a variety of transition metal catalysed
reactions.
OAc
Ph
CH(CO2Me)2
Ph
CH2(CO2Me)2
Ph
Ph
cation/base
5 mol% [π-C3H5PdCl]2
10 mol% ligand
We gratefully acknowledge the Washington University
High-Resolution NMR Facility, partially supported by NIH
1S10R02004, and the Washington University Mass Spec-
trometry Resource Center, partially supported by
NIHRR00954, for their assistance.
Table 2 Results with ligand containing single chiral centre
Ligand
Footnotes
Entry (config.) Cation/base
Yield (%) ee (%)a Solvent
* E-mail: srg@wuchem.wustl.edu
1
2
3
4
5
6
7
8
9
21 (S)
22 (S)
23 (S)
24 (S)
23 (S)
23 (S)
23 (S)
23 (S)
23 (S)
23 (S)
Bu4N+/BSA
Bu4N+/BSA
Bu4N+/BSA
Bu4N+/BSA
K+/BSA
67
78
94
82
87
62
82
87
86
37
66
90
81
11
64
78
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
THF
CN
THF
C6H6
† General procedure for p-allyl addition: The phosphinooxazoline ligand
3
was mixed with [Pd(h -C3H5)Cl]2 in degassed MeCN. After 30 min,
1,3-diphenylprop-2-enyl acetate (10 equiv.) in MeCN was added. To this
solution at the desired reaction temperature, a solution of dimethyl malonate
(30 equiv.), Bu4NF (30 equiv.) and BSA (30 equiv.) in MeCN was added
over 1 h. After complete reaction, as judged by TLC, the reaction mixture
was worked up extractively.
(C6H13)4N+/BSA
(C6H13)4NBr/Me/NaH 22
Bu4N+/BSA
Bu4N+/BSA
Bu4N+/BSA
93
14
25
References
10
1 S. R. Gilbertson, G. Chen and M. McLoughlin, J. Am. Chem. Soc., 1994,
116, 4481.
2 S. R. Gilbertson and X. Wang, J. Org. Chem., 1996, 61, 434.
3 S. R. Gilbertson and R. V. Pawlick, Angew. Chem., Int. Ed. Engl., 1996,
35, 902.
4 S. R. Gilbertson and G. W. Starkey, J. Org. Chem., 1996, 61, 2922.
5 C. G. Frost, J. Howarth and J. M. J. Williams, Tetrahedron: Asymmetry,
1992, 3, 1089.
6 O. Reiser, Angew. Chem., Int. Ed. Engl., 1993, 32, 547.
7 B. M. Trost and D. L. Van Vranken, Chem. Rev., 1996, 96, 395.
8 G. C. Lloyd-Jones and A. Pfaltz, Angew. Chem., Int. Ed. Engl., 1995, 34,
462.
a
The enantiomeric excesses were determined by chiral shift reagent
[Eu(hfbc)3] (ref. 11). BSA = bis(trimethylsilyl)acetamide.
Table 3 Results with ligands containing two chiral centresa
Ligand
(config.)
Entry
Cation/base
Yield (%) ee (%) Solvent
1
2
3
4
5
6
7
8
9
10
11
12
13
14
25 (S,S)
26 (R,S)
27 (S,S)
28 (R,S)
29 (S,S)
30 (R,S)
27 (S,S)
27 (S,S)
28 (R,S)
28 (R,S)
28 (R,S)
28 (R,S)
28 (R,S)
28 (R,S)
Bu4N+/BSA
Bu4N+/BSA
Bu4N+/BSA
Bu4N+/BSA
Bu4N+/BSA
Bu4N+/BSA
Bu4N+/BSA
Bu4N+/BSA
Bu4N+/BSA
Bu4N+/BSA
Bu4N+/BSA
Bu4N+/BSA
TBAOAc/BSA
100
53
91
80
56
37
91
43
80
33
48
49
93
84
93
90
93
22
54
90
65
93
85
88
79
82
97
MeCN
MeCN
MeCN
MeCN
MeCN
MeCN
MeCN
CH2Cl2
MeCN
CH2Cl2
THF
9 P. von Matt and A. Pfaltz, Angew. Chem., Int. Ed. Engl., 1993, 32,
566.
10 J. Sprinz and G. Helmchen, Tetrahedron Lett., 1993, 34, 1769.
11 J. V. Allen, S. J. Coote, G. J. Dawson, C. G. Frost, C. J. Martin and
J. M. J. Williams, J. Chem. Soc., Perkin Trans. 1, 1994, 2065.
12 G. J. Dawson, C. G. Frost and J. M. J. Williams, Tetrahedron Lett.,
1993, 34, 3149.
13 O. Loiseleur, P. Meier and A. Pfaltz, Angew. Chem., Int. Ed. Engl.,
1996, 35, 200.
14 P. Wipf and S. Venkatraman, Tetrahedron Lett., 1996, 37, 4659.
15 B. M. Trost and D. J. Murphy, Organometallics, 1985, 4, 1143.
C6H6
MeCN
MeCN
(C6H13)4N+/BSA 99
Received in Corvallis, OR, USA, 21st October 1996; Com.
6/07200C
a General procedure given in the footnote†.
976
Chem. Commun., 1997