SimplePhos as efficient ligand for the copper-catalyzed kinetic resolution of cyclic vinyloxiranes with Grignard reagents

Article abstract of DOI:10.1055/s-2008-1077901

Cyclic vinyloxiranes react with Grignard reagents under copper catalysis. Highly efficient kinetic resolution is achieved with the new SimplePhos family of chiral phosphorus ligands. Stereodivergence could be exploited to get either the S_N2 or

Full text of DOI:10.1055/s-2008-1077901

LETTER
1797
SimplePhos as Efficient Ligand for the Copper-Catalyzed Kinetic Resolution
of Cyclic Vinyloxiranes with Grignard Reagents
SimplePhos
Ligae
n
E
fficient
K
in
a
etic
R
esolu
u
tion of
C
ycl
d
ic Vinylepoxide
s
Millet, Alexandre Alexakis*
Department of Organic Chemistry, University of Geneva, 30 Quai Ernest Ansermet, 1211 Geneva, Switzerland
Fax +41(22)3793215; E-mail: alexandre.alexakis@chiorg.unige.ch
Received 15 March 2008
In this communication we report our results on the cop-
Abstract: Cyclic vinyloxiranes react with Grignard reagents under
copper catalysis. Highly efficient kinetic resolution is achieved with
the new SimplePhos family of chiral phosphorus ligands. Stereodi-
vergence could be exploited to get either the SN2 or SN2¢ product
with up to 96% ee.
per-catalyzed kinetic resolution of cyclic vinyloxiranes
with Grignard reagents with SimplePhos ligands. From a
mechanistic point of view this reaction is a copper-cata-
lyzed allylic substitution.7a,b
Key words: epoxides, asymmetric synthesis, Grignard reagents,
catalysis, copper
L1 R1 = Ph, R2 = R2' = Ph, R3 = Me
L2 R1 = Ph, R2 = R2' = 2-naphthyl, R3 = Me
L3 R1 = Ph, R2 = R2' = 2-MeC6H4, R3 = Me
L4 R1 = 2-naphthyl, R2 = R2' = Ph, R3 = Me
L5 R1 = 3,5-Me2C6H3, R2 = R2' = Ph, R3 = Me
L6 R1 = 3-CF3C6H4, R2 = R2' = Ph, R3 = Me
L7 R1 = 4-MeC6H4, R2 = R2' = Ph, R3 = Me
L8 R1 = Ph, R2 = R2' = 2-MeOC6H4, R3 = Me
L9 R1 = R2 = Ph, R2' = 2-naphthyl, R3 = Me
L10 R1 = Ph, R2 = R2' = Ph, R3 = Et
R2
N
R1
R3
R3
Asymmetric carbon–carbon bond formation is an impor-
tant issue in modern organic chemistry. Copper catalysis
is a well-documented method to achieve this goal.1
Among all the different methodologies to produce enan-
tiomerically pure chiral centers, kinetic resolution is the
oldest but still a very useful and successful one.2 Chiral al-
lylic alcohol can be easily prepared by kinetic resolution
of racemic vinyloxiranes (Scheme 1). Copper-catalyzed
kinetic resolution of such racemic compound with hard
nucleophile is not very well documented. Dialkylzinc re-
agents have been successfully used in combination with a
chiral copper complex based on phosphoramidite.3 Simi-
lar complexes can also promote the kinetic resolution of
vinyloxiranes with trimethylaluminium.4 We also recent-
ly showed that more easily and commercialy available
Grignard reagents are able to perform this kinetic resolu-
tion in the presence of chiral ferrocenyl-based ligand and
copper.5 However, all these methodologies have several
drawbacks, zinc and aluminum reagents lead to good to
excellent enantioselectivities (up to 96% ee) but only the
transfer of commercially available methyl and ethyl group
has been reported. With Grignard reagents both primary
and secondary alkyl groups can be transferred, but with
lower enantioselectivities (up to 90% ee with primary one
and up to 74% ee with secondary one). More recently, our
group described the synthesis and the applications of a
new class of phosphorus ligand SimplePhos and their ap-
plication in copper-catalyzed allylic substitution.6
P
R1
R2'
Figure 1 Ligands used for this study
Different SimplePhos ligands were tested (Scheme 1) and
the simplest L1 was used to perform a screening of exper-
imental conditions (Scheme 2, Table 1). The nature of the
copper salt and the solvent are known to be crucial for
both regioselectivity and enantioselectivity.8,9 First of all,
the reaction was performed in a mixture of CH2Cl2 and the
Et2O content of the Grignard reagent, with CuTC (copper
2-thiophene carboxylate) as the best conditions described
for the allylic substitution with Grignard reagents and this
family of ligand.6,7c,d
Table 1 Screening of Experimental Conditions
Entry Copper salt Solvent Conversiona SN2¢/SN2c ee SN2¢
(yield,b %)
50 (47)
17
(%)d
e
e
e
e
e
1
2
3
4
5
6
7
CuTC
CuCN
CH2Cl2
CH2Cl2
95:5
92:8
96:4
90:10
96:4
97:3
96:4
88
14
CuOAc·H2O CH2Cl2
20 (17)
18 (15)
15
62
Cu(OTf)2
CuBr
CH2Cl2
CH2Cl2
Et2O
48
88
R
O
R'M' (0.5 equiv)
CuTC
45
40
R
OH
Cu/L*
n
n
OH
n
CuTC
PhMe
50
46
1 n = 1
2 n = 2
a Determined by GC-MS with undecane as internal standard.
b Isolated yield.
Scheme 1
c Determined by 1H NMR.
d Determined by chiral GC.
e Mixture of CH2Cl2 and the Et2O content of the Grignard reagent.
SYNLETT 2008, No. 12, pp 1797–1800
x
x
.x
x
.2
0
0
8
Advanced online publication: 02.07.2008
DOI: 10.1055/s-2008-1077901; Art ID: G09308ST
© Georg Thieme Verlag Stuttgart · New York

1798
R. Millet, A. Alexakis
LETTER
Bu
Table 2 Ligand Studya
OH
OH
Entry
Ligand
Conversion SN2¢/SN2c
ee SN2¢
BuMgCl (0.5 equiv)
O
+
(%)b
(%)d
Cu (3 mol%), L1 (3 mol%)
solvent, –78 °C, 3 h
Bu
1
2
L1
L2
L3
L4
L5
L6
L7
L8
L9
L10
47
50
45
20
50
30
45
43
42
43
95:5
94:6
90:10
93:7
97:3
90:10
97:3
92:8
98:2
83:17
88
70
22
12
88
52
66
38
84
34
2
3a
3b
Scheme 2 Test reaction
3
Full conversion was obtained (50% maximum conver-
sion) with a very good SN2¢/SN2 ratio and a good enantio-
meric excess (Table 1, entry 1). Then, different copper
salts were tested and all showed good regioselectivities,
but a dramatic effect on the conversion was observed:
none of them led to more than 20% conversion (Table 1,
entries 2–5). And only CuBr led to a good enantioselectiv-
ity (88% ee, Table 1, entry 5).
4
5
6
7
8
9
The choice of the solvent is very important in terms of
enantiomeric excess. When more polar solvent, such as
Et2O, was used the enantiomeric excess decreases down
to 40% (Table 1, entry 6) and aromatic solvent had the
same effect on the reaction (Table 1, entry 7). Our inves-
tigations continued by modifying the structure of the
ligand (Figure 1).
10
a Reaction conditions: 2 (1 mmol), CuTC (3 mol%), ligand (3 mol%),
BuMgCl in Et2O (0.5 equiv.), CH2Cl2, –78 °C.
b Determined by GC-MS with undecane as internal standard.
c Determined by 1H NMR.
d Determined by chiral GC.
We first studied the effect of the bulkiness of the ligand by
changing the aromatic ring of the amine part from a phe-
nyl to a 2-naphthyl (Table 2, entry 2) or o-tolyl (Table 2,
entry 3). In both cases a decrease of the enantioselectivity
was observed (70% for the 2-naphthyl and only 22% for
the o-tolyl). An increase of the bulkiness on the phospho-
rus part L4 led to an impressive lost of conversion and
enantioselectivity (Table 2, entry 4).
mental to the reaction with a decrease of the enantiomeric
excess down to 66% (Table 2, entry 7).
Finally, other modifications were studied such as the in-
troduction of a chelating group on the aromatic part of the
amine L8, C1-symmetrical ligand L9 and the replacement
of a methyl group by an ethyl one on the amine part
(Table 2, entries 8–10). But none of these modifications
led to any improvement (38%, 84%, and 34% ee, respec-
tively).
Then, we tried to increase the bulkiness of the two aryl
groups on the phosphorus by changing the two phenyl
group for two xylyl L5, but no improvement was observed
on the reaction (Table 2, entry 5).
With this set of conditions in hand we tested the scope of
the substrate and Grignard reagents (Scheme 3).
Then we decided to study the electronic effect of the aro-
matic part on the phosphorus atom. Ligands bearing elec-
tron-deficient aryl group (3-F3CC6H4) L6 gave a bad
result with moderate conversion, regioselectivity, and low
enantiomeric excess (Table 2, entry 6). A small increase
of the electronic density on the aromatic was also detri-
The very reactive five-membered ring 1 led to very good
selectivities with simple primary Grignard reagents
(Table 3, entries 1 and 2) with respectively 90% ee and
92% ee, with ethyl and butyl Grignard, and more than
90:10 regioselectivity. However, isolated yield remains
low around 20%. Nevertheless, these results remain better
than the one obtained previously on this substrate.3 Mov-
ing to a secondary Grignard reagent had a positive effect
RMgCl (0.51 equiv)
CuTC/L1 (3 mol%)
R
O
+
R
OH
CH2Cl2, –78 °C
OH
n
n
n
1 n = 1
2 n = 2
3a n = 2 R = Bu
3b n = 2 R = Bu
4b n = 1 R = Et
5b n = 1 R = Bu
6b n = 1 R = i-Pr
7b n = 1 R = CH2TMS
8b n = 2 R = Et
4a n = 1 R = Et
5a n = 1 R = Bu
6a n = 1 R = i-Pr
7a n = 1 R = CH2TMS
8a n = 2 R = Et
9a n = 2 R = CH2CH2Ph 9b n = 2 R = CH2CH2Ph
10a n = 2 R = i-Bu
11a n = 2 R = i-Pr
10b n = 2 R = i-Bu
11b n = 2 R = i-Pr
12a n = 2 R = c-Hex
13a n = 2 R = c-C5H9
14a n = 2 R = CH2TMS
12b n = 2 R = c-Hex
13b n = 2 R = c-C5H9
14b n = 2 R = CH2TMS
Scheme 3
Synlett 2008, No. 12, 1797–1800 © Thieme Stuttgart · New York

LETTER
SimplePhos Ligands for the Efficient Kinetic Resolution of Cyclic Vinyloxiranes
1799
Table 3 Scope of Substrate and Grignard Reagentsa
on the yield, but the enantiomeric excess decreases some-
what while regioselectivity remained at the same level
(Table 3, entry 3). Finally, the highly hindered and func-
tionalized TMSCH2MgCl gave a poor SN2¢/SN2 ratio of
58:42 but an interesting 72% ee with a good yield. This is
the first successful result with this Grignard reagent.
Entry
n
R
Product Conversionb
(yield,c %)
SN2¢/ eeSN2¢
SN2d
(%)e
1
2
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
Et
Bu
4a
5a
18
20
91:9
90
90:10
92:8
92
The six-membered ring was then investigated. Simple pri-
mary Grignard reagents showed good reactivities with al-
most perfect regioselectivities and good enantioselectivities
(80% ee for ethyl and 88% ee for butyl) with good conver-
sion and yield (Table 3, entries 5 and 6). Optical rotation
of the acetate derivative of 3a allowed us to determine the
absolute configuration 1S,4S {[a]D20 –155 (c 0.76 CHCl3),
ee = 88%].5 Increasing the bulkiness of the Grignard re-
agent in b-position induces a slight decrease in the
enantioselectivity of the process 73% ee with PhCH2CH2
and 79% ee for i-Bu (Table 3, entries 7 and 8). The ee can
be increased above the 80’s by decreasing the conversion
to 25% (Table 3, entry 9).
3
i-Pr
6a
32
76
4
CH2TMS
Et
7a
41f
58:42
95:5
72g
82
5
8a
48 (42)
47
6
Bu
3a
95:5 88 (–)
7
CH2CH2Ph
i-Bu
9a
42
>95:5
97:3
97:3
96:4
95:5
83:17
95:5
95:5
87:13
73
79
83
84
93
88
96
88
45k
8
10a
10a
11a
11a
12a
12a
13a
14a
46 (42)
25 (20)
49 (45)
18 (15)
55 (44)
37 (32)
38 (35)
36 (30f)
9h
10
11i
12
13j
14
15
i-Bu
i-Pr
i-Pr
Using secondary Grignard reagents allowed a good enan-
tioselectivity for the kinetic-resolution process and espe-
cially when the reaction was stopped below 50%
conversion with ee above 90% and good regioselectivities
(Table 3, entries 11 and 13). When the reaction was al-
lowed to reach 50% conversion the ee decreased slightly
below 90% (Table 3, entries 10, 12, and 14).
c-Hex11
c-Hex
c-C5H9
CH2TMS
a Reaction conditions see Scheme 3.
Finally, the highly hindered and functionalized
TMSCH2MgCl gave a moderate regioselectivity and a
good yield. Interestingly, even if the ee of the SN2¢ is not
high (45%) the SN2 product is obtained in good enantiose-
lectivity (70%, Table 3, entry 15).
b Determined by GC-MS with undecane as internal standard.
c Isolated yield of allylic alcohol.
d Determined by 1H NMR.
e Determined by chiral GC.
f Isolated yield of the mixture.
g The ee of SN2 product was not determined.
h Grignard: 0.3 equiv.
It has already been shown that in some cases a regiodiver-
gent kinetic resolution of such vinyloxiranes can be per-
formed.10 We tried to perform such resolution with our
system by adding 1.5 equivalents of i-PrMgCl
(Scheme 4). We were delighted to see that we managed to
obtain the SN2¢ product 11a with still a moderate ee (54%
ee) but the SN2 product in enantiomerically pure form
(>99% ee). No traces of the other enantiomer were detect-
ed.
i Grignard: 0.25 equiv.
j Grignard: 0.4 equiv.
k Enantiomeric excess of SN2 product: 70%.
ty. Furthermore, SimplePhos ligands gave access to a
wider range of Grignard reagents (functionalized Grig-
nard reagents) with the same or higher enantioselectivities
than the ferrocenes-based ligands.5 Reactions were also
cleaner with this new methodology. Additional studies are
in progress with other substrates and especially on the re-
giodivergent kinetic resolution process.
In conclusion we have shown that SimplePhos ligands are
able to perform the kinetic resolution of racemic cyclic vi-
nyloxirane with moderate to good regio- and enantiose-
lectivities and from low-to-good yields. We were also
able to perform a regiodivergent kinetic resolution with
secondary Grignard reagents yielding the homoallylic al-
cohol in moderate yield but in very high enantioselectivi-
Acknowledgment
The authors thank the Swiss National Research Foundation (no. 20-
068095.02), and BASF for a generous gift of chiral amines.
Scheme 4 Regiodivergent kinetic resolution of 2 with i-PrMgCl
Synlett 2008, No. 12, 1797–1800 © Thieme Stuttgart · New York

1800
R. Millet, A. Alexakis
LETTER
(9) Marino, J. P.; Floyd, D. M. Tetrahedron Lett. 1979, 25, 675.
(10) Bertozzi, F.; Crotti, P.; Macchia, F.; Pineschi, M.; Feringa,
B. L. Angew. Chem. Int. Ed. 2001, 40, 930.
References and Notes
(1) Modern Organocopper Chemistry; Krause, N., Ed.; Wiley-
VCH: Weinheim, 2002.
(2) Vedjes, E.; Jure, M. Angew. Chem. Int. Ed. 2005, 44, 3974;
and references cited therein.
(3) (a) Badalassi, F.; Crotti, P.; Macchia, F.; Pineschi, M.;
Arnold, A.; Feringa, B. L. Tetrahedron Lett. 1998, 39, 7795.
(b) For a review, see: Pineschi, M. New J. Chem. 2004, 28,
657; and references cited therein.
(4) Equey, O.; Alexakis, A. Tetrahedron: Asymmetry 2004, 15,
1531.
(5) Millet, R.; Alexakis, A. Synlett 2007, 435.
(6) Palais, L.; Mikhel, I. S.; Bournaud, C.; Micouin, L.; Falciola,
C. A.; Vuagnoux-d’Augustin, M.; Rosset, S.; Bernardinelli,
G.; Alexakis, A. Angew. Chem. Int. Ed. 2007, 46, 7462.
(7) (a) Yorimitsu, H.; Oshima, K. Angew. Chem. Int. Ed. 2005,
44, 4435. (b) Alexakis, A.; Malan, C.; Lea, L.; Tissot-
Croset, K.; Polet, D.; Falciola, C. Chimia 2006, 60, 124.
(c) Tissot-Croset, K.; Polet, D.; Alexakis, A. Angew. Chem.
Int. Ed. 2004, 43, 2426. (d) Alexakis, A.; Croset, K. Org.
Lett. 2002, 4, 4147. (e) For a very recent review, see:
Falciola, C. A.; Alexakis, A. Eur. J. Org. Chem. 2008, DOI:
10.1002/ejoc.200800025.
(11) Typical Procedure
In a dried Schlenk tube under N2 atmosphere were placed
CuTC (5.7 mg, 0.03 mmol) and (S,S)-L1 (13.5 mg, 0.03
mmol). Then, CH2Cl2 was added (6 mL) and the mixture was
stirred at r.t. for 20 min. The tube was cooled down to –78 °C
and 1,3-cyclohexadiene monoepoxide 1 (100 mL, 1 mmol)
was added. c-HexMgCl (250 mL, 2.0 M in Et2O, 0.5 mmol)
was added dropwise over a period of 3 min. The reaction was
stirred for 3 h and quenched with MeOH and (NH4)2SO4
before being allowed to reach r.t. The aqueous phase was
extracted with CH2Cl2 and the organic layer was dried over
MgSO4 and the solvent removed on rotary evaporator (55%
conversion, 88% ee). The crude product was purified by
flash chromatography on SiO2 (pentane then pentane–Et2O,
20
75:25) yielding 79 mg (44% yield) of a white solid. [a]D
–128.2 (c 0.87, CHCl3, ee = 88%). 1H NMR (400 MHz,
CDCl3): d = 5.68 (m, 2 H), 4.20 (br, 1 H), 2.09 (m, 1 H), 1.97
(m, 1 H), 1.74–1.60 (m, 6 H), 1.45–1.35 (m, 3 H), 1.22–1.12
(m, 4 H), 1.07–0.97 (m, 2 H). 13C NMR (100 MHz, CDCl3):
d = 133.3, 130.1, 67.5, 42.4, 41.1, 32.7, 30.2, 29.7, 26.7,
23.7.
(8) Alexakis, A.; Benhaim, C.; Rosset, S.; Humam, M. J. Am.
Chem. Soc. 2002, 124, 5262.
Synlett 2008, No. 12, 1797–1800 © Thieme Stuttgart · New York

Copyright of Synlett is the property of Georg Thieme Verlag Stuttgart and its content may not
be copied or emailed to multiple sites or posted to a listserv without the copyright holder's
express written permission. However, users may print, download, or email articles for
individual use.