Asymmetric reduction of perfluoroalkyl ketones with chiral lithium alkoxides

Article abstract of DOI:10.1016/j.tetlet.2006.02.078

Reduction of perfluoroalkyl ketones with chiral lithium alkoxides gave chiral α-perfluoroalkyl alcohols in high enantiomeric excesses. Interestingly, reaction of 2,2,2-trifluoroacetophenone (1) with lithium (S)-1-phenylethoxide (2) gave (S)-2,2,2-trifluoro-1-phenylethanol (3), while the same reaction of perfluorooctan-1-one (7) with 2 gave (R)-1H-1- phenylperfluorooctanol (8). Based on the speculation of mechanism, the order of steric effects on this reaction is estimated as C₇F₁₅ > substituted phenyl > CF₃.

Full text of DOI:10.1016/j.tetlet.2006.02.078

Tetrahedron Letters 47 (2006) 2821–2824
Asymmetric reduction of perfluoroalkyl ketones
with chiral lithium alkoxides
Yasser Samir Sokeirik, Kazuyuki Sato, Masaaki Omote, Akira Ando*
and Itsumaro Kumadaki*
Faculty of Pharmaceutical Sciences, Setsunan University, 45-1, Nagaotoge-cho, Hirakata 573-0101, Japan
Received 12 December 2005; revised 6 February 2006; accepted 6 February 2006
Available online 2 March 2006
Abstract—Reduction of perfluoroalkyl ketones with chiral lithium alkoxides gave chiral a-perfluoroalkyl alcohols in high enantio-
meric excesses. Interestingly, reaction of 2,2,2-trifluoroacetophenone (1) with lithium (S)-1-phenylethoxide (2) gave (S)-2,2,2-trifluo-
ro-1-phenylethanol (3), while the same reaction of perfluorooctan-1-one (7) with 2 gave (R)-1H-1-phenylperfluorooctanol (8). Based
on the speculation of mechanism, the order of steric effects on this reaction is estimated as C7F15 > substituted phenyl > CF3.
ꢀ 2006 Elsevier Ltd. All rights reserved.
Nowadays, asymmetric synthesis is one of the most
important topics in organic chemistry. For this purpose,
we have synthesized (Ra)-2,20-bis{(R)-1H-1-hydroxy-
perfluoroalkyl}biphenyls, axially dissymmetric ligands
with two chiral centers.1 These ligands showed high
asymmetric induction for the reaction of benzaldehyde
with diethylzinc in the presence of Ti(OiPr)4. This high
asymmetric induction owes to high stability and high
steric constraint of 1H-1-hydroxyperfluoroalkyl group.
Thus, we designed another ligand with this group and
ortho-nitrogen functionality. For this purpose, we
planned to synthesize tert-butyl N-[(perfluoroocta-
noyl)phenyl]carbamate by the reaction of the corre-
sponding aryllithium with ethyl perfluorooctanoate.2
Interestingly, we obtained not the objective ketone but
its reduction product, as shown in Scheme 1. We exam-
ined the mechanism of this reduction, and found that
lithium ethoxide played an important role in this reduc-
tion.2 This reaction proceeds with various aromatic per-
fluoroalkyl ketones in good to excellent yields, especially
when lithium isopropoxide is used.
Based on these results, we expected that asymmetric
reduction could be induced if a chiral lithium alkoxide
was used. In this letter, we would like to report the
asymmetric reduction of perfluoroalkyl ketones with
chiral lithium alkoxides carried out on the above
expectation.
Li
H
O
O
H
NHBoc
Li
BocN
BocN
O
CH3
H
C7F15
+ C7F15CO2C2H5
NHR O
C7F15
H
LiOC2H5
C7F15
H
BocN
OH
OC2H5
H
+
CH3CHO
C7F15
Scheme 1.
Keywords: Asymmetric reduction; Chiral transfer; Perfluoroalkyl ketone; Lithium alkoxide; a-(Perfluoroalkyl)carbinol; 1-Phenylethanol; Trifluoro-
acetophenone; Steric effect.
*
Corresponding authors. Tel.: +81 72 3241; fax: +81 72 850 7020; e-mail addresses: aando@pharm.setsunan.ac.jp; kumadaki@pharm.setsunan.ac.jp
0040-4039/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2006.02.078

2822
Y. S. Sokeirik et al. / Tetrahedron Letters 47 (2006) 2821–2824
First, we chose lithium menthoxide as chiral lithium alk-
oxide for the reduction of 2,2,2-trifluoroacetophenone
(1), since Nasipuri and his co-worker had reported that
dichloroaluminum menthoxide showed the highest
asymmetric induction of the alkoxide examined on the
reduction of 1.3 However, lithium menthoxide gave a
very low yield of the reduction product at room tempera-
ture, while raising the temperature to 70 ꢁC improved
the chemical yield but led to decrease of ee. We thought
that the presence of a-phenyl group would lower the
activation energy required for the reaction, allow the
reduction proceed at lower temperature, and hence
improve the enantioselectivity of the reaction. So, we
examined lithium (S)-1-phenylethoxide (2) as a reducing
agent. The reaction of 1 with 2 at room temperature
produced a reduced alcohol (3) in moderate yield
(53%) and ee (55%). Lowering the temperature from
room temperature to 0 ꢁC had a good effect in improve-
ment of the ee up to 80% but slight improvement on the
yield (63%), due to the formation of aldol adduct (4) as a
side product that was formed by the reaction of 1 and
acetophenone enolate. The absolute configuration of
the resulting alcohol (3) was determined to be (S) based
on the direction of optical rotation (Scheme 2).4
more bulky perfluoroheptyl ketones, 1-phenylperfluoro-
octan-1-one (7) and its ortho bromo (9) and ortho nitro
(11) derivatives, gave better chemical yields up to
80% of the reduction products (8, 10, and 12) of (R)-
configuration with high ee’s. Interestingly, absolute con-
figurations of latter three products are opposite to the
smaller analogs.
The determination of the absolute configuration of these
products seemed to be rather difficult, since Mosher
esters are not productive in our case of perfluoroalkyl
alcohols. We determined stereochemistry of these iso-
mers by chemical transformation.7 Thus, 6 was debro-
minated by lithiation followed by treatment with water
to the known (S)-1H-1-phenylperfluoro-1-butanol.6
The alcohol (10) from the ortho bromo compound (9)
was found to be (R) by comparing the retention time
on a chiral GC with that of authentic sample.1b The
nitro alcohol (12) from 11 was hydrogenated in the pres-
ence of Pd-C to the ortho amino compound (13), which
was converted to a diazonium salt. The diazonium salt
was heated with HBr–CuBr to give 10 or heated with
H3PO2 to yield the unsubstituted phenyl derivative (8).
So we could determine the absolute configurations of
8, 10, and 12 to be (R), as shown in Scheme 3.
This result encouraged us to investigate this chiral
reduction of other perfluoroalkyl ketones, since we
had reported that the byproduct from aldol reaction
was not obtained on the reduction of a larger perfluoro-
alkyl ketones with lithium alkoxide. The results with
other perfluoroalkyl ketones are shown in Table 1.5
The interesting point of the above results is that the
reduction of 1 and 5 with (S)-2 gave (S)-3 and 6, respec-
tively, while that of 7, 9, or 11 afforded (R)-8, 10, or 12.
Br
O
Br
C3F7
H
C3F7
H
Ph
(S)
H3C
BuLi
H2O
A little larger perfluoropropyl ketone, 1-(2-bromophen-
yl)perfluorobutan-1-one (5) gave the reduction product
of the same configuration as that of 3 in 78% yield with
87% ee without formation of an aldol product, while
C3F7
+
OH
OH
H
OLi
(S)
(S)
5
2
6
O
C7F15
OH
Ph
(S)
H3C
C7F15
+
H
(R)
H
OLi
H3PO2
OH
7
2
8
+
Br
O
Br
C7F15
N2
O
CF3
OH
H
Ph
2
H
CuBr
HBr
OLi
Ph
C7F15
S
H3C
( )
CF3
(R)
H
(R)
OH
+
Ph
(S)
+
H
OLi
9
10
3
1
2
NO2
O
NO2
NH2
C7F15
OH
C7F15
OH
O
HO CF3
O
H2
2
CF3
C7F15
H
H
(R)
(R)
Ph
Pd-C
4
11
12
13
Scheme 2.
Scheme 3.
Table 1. Reduction of perfluoroalkyl ketones with lithium (S)-1-phenylethoxide
Ketone
Temp (ꢁC)
Yielda (%)
eeb (%)
Abs config.
2,2,2-Trifluoroacetophenone (1)
2,2,2-Trifluoroacetophenone (1)
1-(2-Bromophenyl)perfluorobutan-1-one (5)
1-Phenylperfluorooctan-1-one (7)
1-(2-Bromophenyl)perfluorooctan-1-one (9)
1-(2-Nitrophenyl)perfluorooctan-1-one (11)
rt
0
0
0
0
0
53
61
78
80
77
68
55
80
87
82
84
91
(S)-3c
(S)-3
(S)-6d
(R)-8
(R)-10e
(R)-12
a Isolated yield.
b Determined by chiral GC.
c Determined by the sign of optical rotation.4
d (S)-6 was debrominated by treatment with BuLi followed by treatment with 2 M HCl, and the sign of optical rotation was compared with the value
reported.6
e Determined by comparing the chiral GC retention time to that of a predetermined authentic sample prepared by our group.1b

Y. S. Sokeirik et al. / Tetrahedron Letters 47 (2006) 2821–2824
2823
CH3
(S)
H3C
O
Rs
Rs
Rs: CF3, C3F7, RL: Ar (S)
Rs: Ar, RL: C7F15 (R)
Li
O
H
+
RL
OH
H
O
RL
Scheme 4.
3. Nasipuri, D.; Bhattacharya, P. K. J. Chem. Soc., Perkin
Trans. 1 1977, 576–578.
4. Compound 3 is commercially available..
We have reported that the CBS reduction of perflu-
oroalkyl ketones gave a similar result and rationalized
this inversion of optical rotation as the steric effects
5. General procedure for the reduction. To a solution of (S)-1-
phenylethanol (0.46 mmol, 56 mg, 0.055 mL) in dry toluene
(3.0 mL) was added n-BuLi (1.58 M in hexane, 0.29 mL,
0.46 mmol) at 0 ꢁC. To the resulting mixture was added
slowly the corresponding perfluoroalkyl ketone (0.38
mmol). The mixture was stirred for additional 16 h, and
then quenched with 2 N HCl. After separation of the two
phases, the aqueous phase was extracted with Et2O (3 ·
5 mL). The combined organic layer was dried over MgSO4,
and evaporated under vacuum. The residue was purified as
shown for each case.
of perfluoroalkyl groups and phenyl group; C7F15
ꢀ
Ph P C3F7 ꢀ CF3.1 Surprisingly, the CBS reduction
of 11 gave only 10% ee. This suggests that introduction
of a nitro group to ortho position of a phenyl group
decreases the steric difference between C7F15 and Ar
group and that our reduction using lithium alkoxide still
differentiates this small difference. Further, the CBS
reduction of 7 gave 8 in 60% ee,1b while this reduction
of 7 gave 8 of much higher ee. This also shows that
the selectivity of the reduction of this report is much
higher than that of CBS reduction.
(S)-2,2,2-Trifluoro-1-phenylethanol (3). The residue was
separated by column chromatography (SiO2, 5% Et2O
in hexane) to give 3 (40.1 mg, 61%), which was identified
by comparison of the spectral data of the commercially
available authentic sample.4 1H NMR (CDCl3) d: 7.38–7.71
(5H, m), 4.98–5.06 (1H, m), 2.62 (1H, s, disappeared with
D2O). 19F NMR (CDCl3) d: (from C6H5CF3): ꢁ15.25 (3F,
The absolute configurations of 3, 6, 8, 10, and 12 are
reasonably explained by a chair-like six-membered tran-
sition state, where the larger substituents occupy the
equatorial position and the smaller one the axial posi-
tion in the transition state. We assume that the order
of steric effects is C7F15 > Ar P C3F7 > CF3, where Ar
represents (un)substituted phenyl groups. If this
assumption is correct, 1 and 5 would give (S)-products
preferentially, and 7, 9, or 11 would give the correspond-
ing (R)-products. These are consistent with the observed
results, which supports our assumption (Scheme 4).
22
d, J = 6.2 Hz). IR (neat) cmꢁ1: 3450. ½aꢂD +33.02 (c 0.92,
CHCl3). 4,4,4-Trifluoro-3-hydroxy-1,3-diphenyl-butan-1-
one (4) was eluated with 10% Et2O in hexane as a colorless
oil (28.0 mg, 25%). 1H NMR (CDCl3) d: 7.94–7.91 (2H, m),
7.59–7.55 (3H, m), 7.50–7.47 (2H, m), 7.37–7.31 (3H, m),
4.04 (1H, d, J = 17 Hz), 3.64 (1H, d, J = 17 Hz). 19F NMR
(CDCl3) d: ꢁ17.59 (3F, s). MS m/z 294 (M+). HRMS calcd
for C16H13F3O2: 294.087 (M+), found: 294.087. IR (neat)
cmꢁ1: 3472, 1672.
In conclusion, treatment of perfluoroalkyl aromatic
ketones can be reduced to chiral a-perfluoroalkyl alco-
hols with chiral lithium 1-phenylethoxide in high enan-
tiomeric excesses. This reduction is especially useful
for reduction of aromatic perfluoroalkyl ketones with
a large perfluoroalkyl group, and (S)-1-phenylethoxide
gives (R)-perfluoroalkyl alcohols in good yields and high
ee, while trifluoromethyl ketone gives (S)-alcohol. This
reduction is useful for reduction of perfluoroalkyl
ketones, of which the CBS reduction does not give good
results. This reaction is a modification of Meerwein–
Pondorf–Verley’s reduction. This reaction is an equilib-
rium reaction and cannot be applied for chiral reduction
of hydrocarbon analogs, but a perfluoroalkyl group sta-
bilizes adjacent sp3 carbon. Thus, this method is quite
useful for synthesis of chiral perfluoroalkyl carbinols.
(S)-1H-1-(2-Bromophenyl)perfluorobutan-1-ol (6). The resi-
due was treated with NaBH4 (70 mg, 1.84 mmol) in MeOH
(3 mL) for 18 h at room temperature to reduce the side
product, acetophenone, and to facilitate its removal. The
mixture was quenched with water and extracted with Et2O
(3 · 5 mL). The organic phase was dried over MgSO4 and
evaporated under vacuum. The residue was purified by
column chromatography (SiO2, 5% Et2O in hexane) to give
22:2
6 (127 mg, 78%) as a colorless oil, ½aꢂD +22.39 (c
1.68, CHCl3). 1H NMR (CDCl3) d: 7.70–7.66 (1H, m),
7.61–7.58 (1H, m), 7.43–7.37 (1H, m), 7.29–7.23 (1H, m),
5.86 (1H, m), 3.16 (1H, br s, disappeared with D2O). 19F
NMR (CDCl3) d: ꢁ18.31 to ꢁ18.99 (3F, m), ꢁ53.25 (1F,
m), ꢁ63.22 to ꢁ64.04 (3F, m). MS m/z 354 (M+). HRMS
calcd for C10H679BrF7O: 353.949 (M+), found: 353.949,
HRMS calcd for C10H681BrF7O: 355.947 (M+), found:
355.946. IR (neat) cmꢁ1: 3520.
(R)-1H-1-Phenylperfluorooctan-1-ol (8). The residue was
purified by column chromatography (SiO2, 5% Et2O in
hexane) to give 8 (144 mg, 80%) as white crystals. Mp 63–
References and notes
1
64 ꢁC. H NMR (CDCl3) d: 7.49–7.36 (5H, m), 5.23–5.18
(1H, m), 2.53 (1H, s, disappeared with D2O). 19F NMR
(CDCl3) d: ꢁ18.36 (3F, m), ꢁ53.50 (1F, m), ꢁ57.10 to
ꢁ60.35 (8F, m), ꢁ62.50 (2F, m), ꢁ63.25 (1F, m). MS m/z
476 (M+). HRMS calcd for C14H7F15O: 476.025 (M+),
1. (a) Omote, M.; Kominato, A.; Sugawara, M.; Sato, K.;
Ando, A.; Kumadaki, I. Tetrahedron Lett. 1999, 40, 5583–
5585; (b) Omote, M.; Nishimura, Y.; Sato, K.; Ando, A.;
Kumadaki, I. Tetrahedron Lett. 2005, 46, 319–322; (c)
Omote, M.; Nishimura, Y.; Sato, K.; Ando, A.; Kumadaki,
I. J. Fluorine Chem. 2005, 126, 407–409.
21:5
found: 476.025. IR (KBr) cmꢁ1: 3480. ½aꢂD +15.67 (c 0.95,
CHCl3).
(R)-1H-1-(2-Bromophenyl)perfluorooctan-1-ol (10). The
residue was purified by column chromatography (SiO2,
2. Sokeirik, Y. S.; Sato, K.; Ando, A.; Kumadaki, I. J.
Fluorine Chem. 2006, 127, 150–152.

2824
Y. S. Sokeirik et al. / Tetrahedron Letters 47 (2006) 2821–2824
2% Et2O in hexane) to produce 10 (162 mg, 77%) as a white
mass, which is purified by column chromatography (SiO2,
2% Et2O in hexane). The resulting solid was recrystallized
from hexane to give 6.50 g (51%) of pale yellow crystals.
Mp 53.0–54.0 ꢁC. H NMR (CDCl3) d: 8.34–8.31 (1H, m),
7.91–7.87 (1H, m), 7.83–7.79 (1H, m), 7.50–7.48 (1H, m).
19F NMR (CDCl3) d: ꢁ17.56 (3F, m), ꢁ53.50 to ꢁ55.30
(2F, m), ꢁ57.50 to ꢁ61.50 (8F, m), 62.50–64.00 (2F, m).
MS m/z 520 (M+1)+. HRMS calcd for C14H5F15NO3:
520.003 (M+1)+, found: 520.003. IR (KBr) cmꢁ1: 1748.
Diazotization of 13: 1H-1-(2-bromophenyl)perfluorooctan-1-
ol (10).
solid. This was identified with the authentic sample by the
following spectral data.1b Mp 42 ꢁC (lit. 43.5 ꢁC). 1H NMR
(CDCl3) d: 7.70–7.67 (1H, m), 7.61–7.59 (1H, m), 7.43–7.39
(1H, m), 7.29–7.26 (1H, m), 5.90 (1H, m), 2.78 (1H, br s
disappeared with D2O). 19F NMR (CDCl3) d: ꢁ17.59 to
ꢁ18.19 (3F, m), ꢁ51.50 (1F, m), ꢁ58.17 to ꢁ59.84 (8F, m),
1
22:2
ꢁ62.78 to ꢁ63.35 (2F, m), ꢁ65.32 (1F, m). ½aꢂD +13.81
(c 0.115, CHCl3).
(R)-1H-1-(2-Nitrophenyl)perfluorooctan-1-ol (12). The resi-
due was first dried at 60 ꢁC at 1 mmHg to remove the
volatile material and purified by column chromatography
(SiO2, 5% Et2O in hexane) to give 10 (134 mg, 68%) as a
1H-1-(2-Aminophenyl)perfluorooctan-1-ol (13, 40 mg) was
treated with 48% HBr (1 mL) and cooled in ice bath to
0 ꢁC. To this suspension was added a solution of sodium
nitrite (20 mg) in H2O (0.5 mL) so slowly as not to allow the
temperature exceed 0 ꢁC. After complete addition, the
mixture was left standing for 15 min and then added at once
to a hot solution of CuBr (40 mg) in 48% HBr (1 mL). The
mixture was boiled for additional 1 h. The mixture was
extracted with Et2O. The organic layer was dried over
MgSO4 and evaporated under vacuum. The residue was
passed through a flash column chromatography (SiO2, 5%
Et2O in hexane) to remove coloring material and analyzed
without further purification with GC with the authentic
sample 10 (a capillary column: TC-5 (GL Sciences Inc,
0.25 mm · 15 m), temp 80–200 ꢁC. Retention time: Authen-
tic sample (CBS reduction): 6.85 min. Diazonium produced
sample: 6.87 min. Chiral column: GAMMA DEXTM (Spe-
lco, 0.25 mm · 30 m). Temp 150 ꢁC, Authentic sample
(CBS reduction): major peak; 12.98 min, minor peak;
13.41. Sample produced in this experiment: major peak;
12.97 min, minor peak; 13.40).
1H-1-Phenylperfluorooctan-1-ol (8). 1H-1-(2-Aminophenyl)-
perfluorooctan-1-ol (40 mg, 13) was dissolved in 50% hypo-
phosphorous acid (2 mL) and cooled in an ice bath to 0 ꢁC,
To this suspension was added NaNO2 (solid, 20 mg) slowly,
and then the mixture was left standing for 15 min at the
same temperature, then the mixture was boiled for 1 h.
After cooled to room temperature, the mixture was
extracted with Et2O. The organic phase was dried over
MgSO4, and evaporated under vacuum. The residue was
passed through a flash column chromatography (SiO2, 10%
Et2O in hexane) to remove coloring material and analyzed
with GC without further purification to compare the
retention time with compound 8 (nonchiral column, temp
80–200 ꢁC/min). Retention time: sample from chiral reduc-
tion with Li alkoxide; 4.296 min. Sample from diazotiza-
tion: 4.303. Chiral column temp 150 ꢁC. The former
sample: minor peak; 6.963 min, major peak; 7.240 min.
The latter sample: minor peak; 6.940, major peak; 7.246
minor peak; 7.246 min.
1
yellow oil. H NMR (CDCl3) d: 8.03–7.99 (2H, m), 7.74–
7.71 (1H, m), 7.58–7.54 (1H, m), 6.50 (1H, m), 3.92 (1H, br
s, disappeared with D2O). 19F NMR (CDCl3) d: ꢁ17.50 to
ꢁ18.33 (3F, m), ꢁ52.50 (1F, m), ꢁ58.17 toꢁ60.82 (8F, m),
ꢁ62.78 to ꢁ63.35 (2F, m), ꢁ65.32 (1F, m). MS m/z 521
(M+) HRMS calcd for C14H6F15NO3: 521.011 (M+),
21:7
found: 521.011. IR (neat) cmꢁ1: 3528. ½aꢂD +167.16
(c 0.107, CHCl3).
6. The reported [a]D for (S)-isomer is +23.2. Ramachandran,
V.; Teodorvic, A.; Brown, H. C. Tetrahedron 1993, 49,
1725–1738.
7. Experiments for synthesis of starting material and deter-
mination of the absolute configurations are as follows.
1H-1-Phenylperfluorobutan-1-ol. A solution of 6 (50 mg) in
THF (3 mL) was treated with n-BuLi (0.17 mL of 2.55 M
solution in hexane, 0.42 mmol) at ꢁ78 ꢁC, The reaction
mixture was stirred for 2 h at the same temperature, then
quenched with 2 N HCl and extracted with Et2O. The
collected organic phase was dried over MgSO4 and evapo-
rated under vacuum. The residue was purified by flash
column chromatography (SiO2, 10% Et2O in hexane)
to give 1H-1-phenylperfluorobutan-1-ol (27 mg, 70%) as a
22
colorless oil ½aꢂD +22.7876 (c 1.06, EtOH). Ee = 78%
(checked by chiral GC). 1H NMR (CDCl3) d: 7.46–7.42
(5H, m), 5.17 (1H, m), 2.61 (1H, br s, disappeared with
D2O). MS m/z 354 (M+).
1-(2-Nitrophenyl)perfluorooctan-1-one (11). To a solution
of 2-bromonitrobenzene (5.00 g, 24.8 mmol) in THF
(120 mL) at ꢁ100 ꢁC was added n-BuLi in hexane
(1.58 M, 17.2 mL, 27.2 mmol) over a period of 45 min.
The resulting mixture was stirred for an additional 1 h at
the same temperature. To this solution was added methyl
perfluorooctanoate (6.55 mL, 27.2 mmol) over a period of
20 min, and the resulting mixture was kept at this tempera-
ture under stirring for additional 6 h. The mixture is
quenched with 2 M HCl (50 mL), and the whole mixture
was extracted with Et2O. The Et2O layer was dried over
MgSO4, and evaporated under vacuum to give an oily black