2700
O. A. Hamed et al. / Tetrahedron Letters 53 (2012) 2699–2701
ketone and 2-hydroxy ketone as by-products. In a related study we
showed that oxidation of ketones with Pd(II) bimetallic complex A
as catalyst in the absence of LiCl and at lower concentration of
a modest ee of 68% for the (R)-enantiomer, while the same sub-
strate in the absence of acid (Run 3) was very slow in the conver-
sion (28% of SM consumed after 2 days) but the asymmetric
induction was higher (82%). Thus acid is beneficial for the forward
progress of the reaction, since acid catalyzes ketone enolization,
although higher acidity may ultimately compromise the ee of the
product. The conversions for all other reactions were much higher,
and workup was performed when >90% of the starting material
was consumed by GC. Run 4 with propiophenone included 2.0 M
of LiCl, in addition to cupric chloride, and resulted in a lower ee rel-
CuCl2 produces predominantly the racemic
The results together suggest the possibility of a new and direct
asymmetric synthesis of -hydroxyketones by incorporation of a
a
-hydroxy ketones.14
a
chiral ligand in the place of the achiral ligand cyclohexane-1,2-dia-
mine in the coordination sphere of bimetallic complex A. Herein
we report the success of this direct approach and describe a new
procedure for the direct asymmetric catalytic
ketones.
a-hydroxylation of
ative to Run 3. Butyrophone gave the corresponding (R)-
yketone in 71% ee, quite comparable to the 68% ee for the shorter
homolog of Run 2. Asymmetric -hydroxylation of the sterically
a-hydrox-
Chiral bimetallic catalyst B (Table 1) was chosen for this study
because of its ease of preparation and success in asymmetric
halohydrin formation. The complex was prepared by reacting tet-
rakis(acetonitrile) palladium(II) tetrafluoroborate complex with
1-phenylhexane-1,3,5-trione followed by (R)-BINAP.13 The result-
ing asymmetric Pd(II) complex B was then utilized in asymmetric
oxidation of various symmetrical and unsymmetrical ketones in
an aqueous solution of THF in the presence of cupric chloride, lith-
ium chloride, and a catalytic amount of trifluoroacetic acid.
The oxidation results are summarized in Table 1, and the abso-
a
larger 1,2-diphenylethanone afforded the (R)-enantiomer (Run 6),
and replacing (R)-BINAP with (S)-BINAP produced the opposite
enantiomer as expected, affording (S)-2-hydroxyketone (Run 7).
Employment of the more bulky 1-(3-chlorophenyl)-propan-1-one
gave (R)-a-hydroxyketone in 92% ee (Run 8), the highest observed
in this series. The similar substrate 1-(3,5-difluorophenyl)-2-
hydroxypropan-1-one gave the product with the (R)-configuration
also in a high ee of 90%. Employing 3,3,3-trifluoro-1-phenylpro-
pan-1-one afforded (R)-furoin, with the opposite configuration rel-
ative to the generally-observed configuration (fluorine reverses the
Cahn–Ingold–Prelog priorities), and the opposite configuration of
what is drawn for Table 1. Run 11 employing furyl furfuryl ketone
lute configurations of the resulting a-hydroxy ketones were deter-
mined based on the comparison of the sign of optical rotation with
literature values.15 We first examined the oxidation of the sym-
metrical cyclic ketone cyclohexanone (Run 1) which produced a
modest ee of 67% for the (R)-enantiomer, then proceeded with
unsymmetrical aralkyl ketones. Propiophenone (Run 2) also gave
gave
Mechanistically, we propose that the
initial enolization, which is accelerated by acid, giving rise to a
-bound enol palladium species. An equilibrating mixture of E
a-hydroxyketone in quite high ee.
a-hydroxylation involves
p
Table 1
Asymmetric synthesis of
a-hydroxyl ketones using chiral bimetallic palladium
and Z-enols should exist in solution, although the two stereoiso-
mers may have quite different binding constants to palladium. A
2,1-insertion involving a palladium-bound solvent water molecule
complex Ba
O
should install the
a-hydroxyl, and reductive elimination then af-
O
[Pd2(triketone)(L*-L*)]
R1
fords the product plus the reduced catalyst, which is reoxidized
by CuCl2, and the resulting CuCl is oxidized by oxygen to complete
the catalytic cycle. Running the reaction under an inert atmosphere
led to only slightly lower isolated yields of products and with no
impact on enantioselectivities, consistent with the presence of ex-
cess CuCl2.
*
R1
R2
CuCl2, O2
R2
OH
CF3CO2H, H2O, THF
1-11
L*-L* = (R)-BINAP
ee = 61-92%
Run
R1
R2
Yieldb (%)
eec (%)
R/S
The result of Run 4 that was conducted at high [Clꢀ] affording a
lower ee suggests that the mode of H2O addition may be different
from that of other runs. In this case of high [Clꢀ], the chloride ion
would replace the solvent in the coordination sphere of Pd(II) dis-
couraging syn hydroxypalladation (2,1-insertion) and giving rise to
1
2
CH2CH2CH2CH2
64
64
36
52
61
52
52
56
53
48
72
67
68
82
51
71
85
87
92
90
89
91
Rg
Rh
Rh
Rh
Ri
CH3
CH3
CH3
CH3CH2
Ph
Ph
Ph
Ph
Ph
Ph
Ph
3d
4e
5
6
Rj
7f
8
Ph
Sj
the possibility of nucleophilic attack of water on the
p-complex,
CH3
CH3
CF3
(3-Cl)Ph
3,5-di-F-Ph
Ph
Rk
Rl
leading to the opposite stereochemistry. Several earlier studies
showed that anti-hydroxypalladation predominates at a high con-
centration of LiCl.16
9
10
11
Rm
Rn
2-furyl
2-furyl
In summary, enantioselectivities of the catalytic alpha-hydrox-
ylation were at least 70% and reached a maximum of 92%. The E/Z
enol ratio would be expected to impact the enantioselectivity, but
as noted the two stereoisomers would have different binding
constants to the palladium catalyst, and different rates of attack
once bound. Further, the presence of excess chloride decreases
the enantioselectivities. With modest increases in enantioselectiv-
ities, which may be realized by varying chiral auxiliaries and reac-
tion conditions, this catalytic oxidation procedure should compete
with much more elaborate and involved procedures for preparing
a
All runs contain 0.08–0.2 mmol of catalyst in 20 mL 4:1 THF/H2O, 0.5 M in
CuCl2, with a catalytic amount of CF3CO2H. t = 25 °C.
b
Yields for isolated products after column chromatography. Absolute configu-
ration as drawn except for run 10.
c
ee’s were determined by 1H NMR utilizing Eu(hfc)3 chiral shift reagent.
No acid catalyst was used in this run.
d
e
Contains 2.0 M of LiCl in addition to CuCl2.
(S)-BINAP was used rather than (R)-BINAP.
f
For (R) ½a 2D0
ꢁ
= +26.20 (c = 1.35, CHCl3)15a; Run 1: ½a D20
ꢁ
= +14.0 (c = 2.0, CHCl3).
= +82.2 (c = 2.0, CHCl3)15c
;
g
For (R) ½a 2D0
ꢁ
= +81.0 (c = 1.5, CHCl3), ee = 96%15b; ½a D20
ꢁ
h
Run 2: ½a 2D0
ꢁ
= +60.3 (c = 2.0, CHCl3); Run 3: ½a D20
ꢁ
= +69.8 (c = 2.0, CHCl3); Run 4:
½
a 2D0
ꢁ = +46.2 (c = 2.0, CHCl3).
i
For (S) ½ ꢁ
a 2D0 = ꢀ30.8 (c = 2.24, CHCl3), ee = 95% in Davis,15d but for (S)
optically active
a-hydroxy ketones. Ongoing work is focused on
understanding the mechanism, improving the enantioselectivities,
and exploring other nucleophiles.
½
a 2D0
ꢁ
= +40.5 (c = 0.3, CHCl3) in Krawczyk15c; Run 5: ½a D20
ꢁ
= +23.7 (c = 2.0, CHCl3).
For (R) ½a 2D0
ꢁ
= ꢀ170.1 (c = 1.0, benzene)15b; ½a D20
ꢁ
= ꢀ230.5 (c = 1.0, benzene)15f
;
j
for (S) ½a 2D0
ꢁ
= +138.4 (c = 0.25, CHCl3)15c, ½a D20
ꢁ
= +114.9 (c = 1.5, acetone)15e; Run 6
= +141.8 (c = 0.2, CHCl3).
½
a 2D0
ꢁ
= ꢀ139.6 (c = 0.2, CHCl3); Run 7 ½a D20
ꢁ
Acknowledgments
For (R) ½a 2D0
ꢁ
= +64.2 (c = 1.2, CHCl3)15g; Run 8: ½a D20
ꢁ
= +66.5 (c = 2.0, CHCl3).
= +46.8 (c = 2.0, CHCl3).
= ꢀ7.8 (c = 0.2, CHCl3).
= +59.2 (c = 1.0, CHCl3).
k
l
For (R) ½a 2D0
ꢁ
= +50.0 (c = 1.0, CHCl3)15h; Run 9 ½a D20
ꢁ
For (S) ½a 2D0
ꢁ
= +8.6 (c = 0.2, CHCl3)15h; Run 10 ½a D20
ꢁ
m
n
American Chemical Society Petroleum Research Fund for
support of this research through ACS PFR #48511-AC1. NSF Grant
For (R) ½a 2D0
ꢁ
= +62.7 (c = 0.9, CHCl3)15h; Run 11 ½a D20
ꢁ