Palladium-catalyzed deracemization of allylic carbonates in water with formation of allylic alcohols: Hydrogen carbonate ion as nucleophile in the palladium-catalyzed allylic substitution and kinetic resolution

Article abstract of DOI:10.1021/ja034109t

The palladium-catalyzed deracemization of racemic cyclic and acyclic allylic methyl carbonates in water in the presence of N,N′-(1R,2R)-1,2-cyclohexanediylbis[2-(diphenylphophino)benzamide] proceeds with high enantioselectivities to give the corresponding allylic alcohols in high yields. This deracemization involves a palladium-catalyzed allylic substitution with the in-situ-formed hydrogen carbonate ion and an irreversible decomposition of the intermediate allylic hydrogen carbonates, with formation of the corresponding allylic alcohols. The palladium-catalyzed reaction of racemic cyclic allylic acetates with potassium hydrogen carbonate in water in the presence of the chiral bisphosphane proceeds with a highly selective kinetic resolution to give the corresponding allylic alcohols and allylic acetates. Copyright

Full text of DOI:10.1021/ja034109t

Published on Web 04/29/2003
Palladium-Catalyzed Deracemization of Allylic Carbonates in Water with
Formation of Allylic Alcohols: Hydrogen Carbonate Ion as Nucleophile in the
Palladium-Catalyzed Allylic Substitution and Kinetic Resolution
Bernhard J. Lu¨ssem and Hans-Joachim Gais*
Institut fu¨r Organische Chemie der RWTH Aachen, Prof.-Pirlet Strasse 1, D-52056 Aachen, Germany
Received January 10, 2003; E-mail: Gais@RWTH-Aachen.de
Scheme 1. Proposed Mechanism of the Palladium-Catalyzed
The complete conversion of a racemate to one enantiomer
without intermediate separation of materials (deracemization) is one
of the current challenges in asymmetric synthesis.1 In particular,
deracemization processes using abiotic chiral catalysts are much
sought after. Allylic alcohols are targets of high synthetic impor-
tance because of their many applications.2 Their deracemization3,4
has been achieved by two methods. The first involves a dynamic
kinetic resolution mediated by lipases and achiral transition metal
catalysts.5 While this method is very effective for acyclic allylic
alcohols, it is less suited for cyclic substrates, and it requires the
application of two catalysts. The second features a palladium(0)-
catalyzed substitution of symmetrically substituted allylic esters with
carboxylate ions in the presence of a chiral ligand for the palladium
atom.6 While this method gives high enantioselectivities for cyclic
allylic esters, it has not yet been applied to symmetrically substituted
acyclic esters. Furthermore, fine-tuning of the reaction conditions
is required in order to avoid racemization of the product ester since
it can also serve as a substrate for the catalyst.
Deracemization of Allylic Carbonates in Watera
a The same scheme is proposed for racemic symmetrically substituted
acyclic carbonates. L* ) chiral ligand.
Scheme 2. Palladium-Catalyzed Deracemization of Cyclic Allylic
Carbonates in Water/CH2Cl2
It occurred to us that, by choosing a cyclic or an acyclic racemic
allylic carbonate as the substrate and utilizing water as part of the
reaction medium, the palladium-catalyzed allylic substitution could
perhaps be turned into a facile deracemization of cyclic and acyclic
allylic carbonates, leading to allylic alcohols. Racemic methyl
carbonates of type rac-1 are known to react with a chiral palladium
catalyst with formation of the π-allyl palladium complex 2,
containing a methyl carbonate ion (Scheme 1).7 In the presence of
water, 2 should form the π-allyl palladium complex 3, containing
the hydrogen carbonate ion.8a,b We speculated that the hydrogen
carbonate ion in turn could substitute the π-allyl palladium ion of
3, resulting in the allylic hydrogen carbonate 4, which should
decompose in water with formation of the allylic alcohol 5 and
CO2.8c,d Moreover, the allylic carbonate-to-allylic alcohol transfor-
mation would be rendered irreversible. A prerequisite for the
attainment of 5 with high ee values is that the decomposition of 4
is faster than the re-formation of 3. The advantages of this method
are that it would require only water and no external nucleophile
and would give directly the allylic alcohol.
Table 1. Palladium-Catalyzed Deracemization of Allylic
Carbonates
a
b
entry
substrate
time, h
alcohol
yield, %
ee, %
1
2
3
4
5
rac-1a
rac-1b
rac-1c
rac-6a
rac-6b
16
24
52
22
24
5a
5b
5c
7a
7b
91
94
94
80
94
43
97
g99
89
g99
a Isolated yield. b Determined by GC on chiral stationary phase.
Scheme 3. Palladium-Catalyzed Deracemization of Acyclic Allylic
Carbonates in Water/CH2Cl2
We were pleased to see that, in a test experiment, treatment of
the racemic cyclohexenyl carbonate rac-1b with 2 mol % of Pd2-
(dba)3‚CHCl3 (dba ) dibenzylideneacetone) and 16 mol % of PPh3
in a mixture of water and CH2Cl2 furnished, after 24 h, the racemic
allylic alcohol rac-5b in 96% yield (Scheme 2). Having demon-
strated the feasibility of this concept, we studied reactions using
BPA9 as ligand for the palladium atom. To our delight, a similar
treatment of rac-1b with Pd2(dba)3‚CHCl3 and BPA in water/CH2-
Cl2 gave alcohol 5b with 97% ee in high yield (Table 1, entry 2).
Extension of the deracemization to the cycloheptenyl carbonate rac-
1c under similar conditions afforded the allylic alcohol 5c with
g99% ee in high yield (entry 3). The deracemization of the
cyclopentenyl carbonate was high yielding but less selective.
Treatment of rac-1a with Pd2(dba)3‚CHCl3 and BPA in water/CH2-
Cl2 gave the allylic alcohol 5a with only 43% ee in high yield (entry
1). However, a more efficient deracemization of the corresponding
acetate could be achieved by using the hydrogen carbonate ion as
an external nucleophile (vide infra).
Fortunately, this deracemization method, which features a
scission of the C-OCO bond of the allylic ester, could also be
extended to acyclic carbonates (Scheme 3). Treatment of the
racemic pentenyl carbonate rac-6a with Pd2(dba)3‚CHCl3 and BPA
in water/CH2Cl2 furnished the allylic alcohol 7a with 89% ee in
80% yield (entry 4). Deracemization of the heptenyl carbonate rac-
6b under similar conditions afforded the allylic alcohol 7b with
g99% ee in high yield (entry 5).
9
6066
J. AM. CHEM. SOC. 2003, 125, 6066-6067
10.1021/ja034109t CCC: $25.00 © 2003 American Chemical Society

C O MMU N I C A T I O N S
Scheme 4. Palladium-Catalyzed Kinetic Resolution and
Deracemization of Cyclic Allylic Acetates with KHCO3
with 1.4 equiv of KHCO3 in water/CH2Cl2 in the presence of Pd2-
(dba)3‚CHCl3 and BPA slowed significantly at approximately 50%
conversion of the starting acetates. Chromatography afforded the
enantioenriched acetates 8b and 8c and the highly enantioenriched
alcohols 5b and 5c in high yields (entries 2 and 3). The stereo-
chemical course of the kinetic resolution and the substitution with
the hydrogen carbonate ion is the same as that with other
nucleophiles in the presence of BPA.12 We note that the palladium-
catalyzed resolution of acetates rac-8b and rac-8c with the hydrogen
carbonate ion in water gives the same products as a hydrolase-
catalyzed resolution, which involves, however, an O-CO bond
cleavage and proceeds in these cases with all enzymes studied with
low enantioselectivities.3b
Table 2. Palladium-Catalyzed Kinetic Resolution and
Deracemization of Cyclic Allylic Acetates with KHCO3
acetate
alcohol
time,
h
convn,
%
yield,
%
ee,
%
yield,
%
ee,
%
a
b
c
b
c
entry
substrate
Acknowledgment. We thank the Deutsche Forschungsgemein-
schaft (SFB 380) for financial support.
1
2
3
rac-8a
rac-8b
rac-8c
18
48
56
100
49
50
89
46
45
88
94
g99
48
48
72
67
Supporting Information Available: Sample experimental proce-
dures for the deracemization of rac-1b and the resolution of rac-8b,
and details of ee value determination, optical rotations, and assignment
of absolute configuration for all compounds described (PDF). This
material is available free of charge via the Internet at http://pubs.acs.org.
a Determined by GC. b Isolated yield. c Determined by GC on chiral
stationary phase.
The results described thus far did not allow us to exclude the
possibility that water or the hydroxide anion and not the hydrogen
carbonate ion is the nucleophile reacting with the π-allyl palladium
ion. Therefore, a number of test experiments were carried out with
the cyclopentenyl acetate rac-8a (Scheme 4). Treatment of rac-8a
with Pd2(dba)3‚CHCl3 and BPA in water/CH2Cl2 did not lead to
the formation of the allylic alcohol 5a; the acetate was recovered
in practically quantitative yield. This was not surprising, since water
is known to be a very poor nucleophile in palladium-catalyzed
allylic substitution.10 Having observed no reaction of rac-8a in
water, we repeated the experiment, but this time in the presence of
1.4 equiv of KHCO3. This measure led to the isolation of 5a with
88% ee in 89% yield (Table 2, entry 1). Since an aqueous solution
of KHCO3 (1.75 M) has a pH of 8.2, the hydroxide ion could have
also acted as the nucleophile. This hypothesis was tested by
submitting acetate rac-8a to catalysis in water of a similar pH but
omitting KHCO3. Mixtures of rac-8a, Pd2(dba)3‚CHCl3, BPA, and
CH2Cl2 were treated with either H2O/NaH2PO4/Na2HPO4 of pH
8.2 or H2O/NaOAc of pH 8.4. In both cases, formation of 5a could
not be observed, and the acetate was recovered. These results show
that the hydrogen carbonate ion acts as the nucleophile and provide
strong evidence for the deracemization of allylic carbonates with
formation of the corresponding allylic alcohols, as proposed in
Scheme 1.
The formation of 5a with 88% ee in the palladium-catalyzed
reaction of acetate rac-8a with aqueous KHCO3 is remarkable.
Treatment of carbonate rac-1a with water alone gave the allylic
alcohol with only 43% ee. To see whether the added nucleophile
is responsible for this difference in selectivity, the reaction of rac-
1a in water was repeated, but this time with the addition of 1.4
equiv of KHCO3. Again, alcohol 5a was obtained with only 43%
ee. The difference in enantioselectivities of the reactions of
carbonate rac-1a and acetate rac-8a with the hydrogen carbonate
ion may perhaps be ascribed to the operation of a nucleofuge-
dependent “memory effect”.11
References
(1) (a) Noyori, R.; Tokunaga, M.; Kitamura, M. Bull. Chem. Soc. Jpn. 1995,
68, 36. (b) Ba¨ckvall, J.-E.; Huerta, F. F.; Alexander, B. E. Chem. Soc.
ReV. 2001, 30, 321.
(2) (a) Lipshutz, B. H.; Sengupta, S. Org. React. 1992, 41, 135. (b) Katsuki,
T.; Martin V. S. Org. React. 1996, 48, 1. (c) Charette, A. B.; Beauchemin,
A. Org. React. 2001, 58, 1. (d) Hanson, R. M. Org. React. 2002, 60, 1.
(3) For the kinetic resolution of allylic alcohols, see: (a) Johnson, R. A.;
Sharpless, K. B. In Catalytic Asymmetric Synthesis; Ojima, I., Ed.; Wiley-
VCH: New York, 2000; p 231. (b) Gais, H.-J.; Theil, F. In Enzyme
Catalysis in Organic Synthesis; Drauz, K., Waldmann, H., Eds.; Wiley-
VCH: Weinheim, 2002; Vol. 2, p 335.
(4) For the asymmetric synthesis of allylic alcohols, see: (a) Corey, E. J.;
Helal, C. J. Angew. Chem., Int. Ed. 1998, 37, 1986. (b) Itsuno, M. Org.
React. 1998, 52, 395. (c) Magnus, A.; Bertilsson, S. K.; Andersson, P. G.
Chem. Soc. ReV. 2002, 31, 223.
(5) (a) Williams, M. J.; Allen, J. V. Tetrahedron Lett. 1996, 37, 1859. (b)
Kim, M.-J.; Jung, J. Y.; Lim, I. T.; Lee, D.; Suh, J. H.; Choi, Y. K. J.
Org. Chem. 1999, 64, 8423. (c) Lee, D.; Huh, E. A.; Kim, M.-J.; Jung,
H. M.; Koh, J. H.; Park, J. Org. Lett. 2000, 2, 2377.
(6) (a) Trost, B. M.; Organ, M. G. J. Am. Chem. Soc. 1994, 116, 10320. (b)
For the use of phenol nucleophiles, see: Trost, B. M.; Tsui, H.-C.; Toste,
F. D. J. Am. Chem. Soc. 2000, 122, 3534.
(7) Moreno-Mana˜s, M.; Pleixats, R. In Handbook of Organopalladium
Chemistry for Organic Synthesis; Negishi, E., de Meijere, A., Eds.;
Wiley: New York, 2002; Vol. 2, p 1707.
(8) (a) Behrendt. W.; Gattow, G.; Dra¨ger, M. Z. Anorg. Allg. Chem. 1973,
397, 237. (b) Ozawa, F.; Son, T.; Ebina, S.; Osakada, K.; Yamamoto, A.
Organometallics 1992, 11, 171. (c) Behrendt. W.; Gattow, G. Z. Anorg.
Allg. Chem. 1973, 398, 198. (d) For the use of NaHCO3/water in the direct
palladium-catalyzed enantioselective conversion of racemic isoprene
monoepoxide to the corresponding vinylglycidol, see: Trost, B. M.;
McEachern, E. J. J. Am. Chem. Soc. 1999, 121, 8649.
(9) (a) Trost, B. M.; Lee, C. In Catalytic Asymmetric Synthesis; Ojima, I.,
Ed.; Wiley-VCH: Weinheim, 2000; p 593. (b) Fairlamb, I. J. S.; Lloyd-
Jones, G. C. Chem. Commun. 2000, 2447.
(10) Beletskaya, I. P.; Cheprakov, A. V. In Handbook of Organopalladium
Chemistry for Organic Synthesis; Negishi, E., de Meijere, A., Eds.;
Wiley: New York, 2002; Vol. 2, p 2957.
(11) Fairlamb, I. J. S.; Lloyd-Jones, G. C.; Vyskocˇil, Sˇ.; Kocˇovsky´, P. Chem.
Eur. J. 2002, 8, 4443 and earlier pertinent work cited therein.
(12) (a) Gais, H.-J.; Eichelmann, H.; Spalthoff, N.; Gerhards, F.; Frank, M.;
Raabe, G. Tetrahedron: Asymmetry 1998, 9, 235. (b) Frank, M.; Gais,
H.-J. Tetrahedron: Asymmetry 1998, 9, 3353. (c) Lloyd-Jones, G. C.;
Stephen, S. C. Chem. Commun. 1998, 2321. (d) Trost, B. M.; Hembre, E.
J. Tetrahedron Lett. 1999, 40, 219. (e) Gais, H. J.; Spalthoff, N.; Jagusch,
T.; Frank, M.; Raabe, G. Tetrahedron Lett. 2000, 41, 3809. (f) Dominguez,
B.; Hodnett, N. S.; Lloyd-Jones, G. C. Angew. Chem., Int. Ed. 2001, 40,
4289.
The use of the hydrogen carbonate ion in the palladium-catalyzed
substitution allows not only deracemization but also an interesting
kinetic resolution of the racemic allylic substrate. The reactions of
the cyclohexenyl acetate rac-8b and cycloheptenyl acetate rac-8c
JA034109T
9
J. AM. CHEM. SOC. VOL. 125, NO. 20, 2003 6067