LETTER
Ytterbium Trifluoromethanesulfonate as an Efficient, Reusable Catalyst
837
Table Yb(OTf)3-Catalysed Reaction of 1,2-Epoxides with Alco-
ued at the reaction temperature for the required period
(monitored by TLC). After removal of solvents using a ro-
tavapor the residual mass is washed with water, the organ-
ic content is extracted with ethyl acetate (20 mL) and is
dried over anhydrous sodium sulfate. Partial evaporation
of the solvent followed by separation on a short silica gel
column (1.5 5 cm 60-120 mesh, petroleum ether/ethyl
acetate 94/6 as eluent) affords analytically pure -alkoxy
alcohol. The isolated yields are calculated on the basis of
the amount of epoxides used and 98-99% mass balance is
hols to form -Alkoxy Alcoholsa
1
obtained. The products are fully characterised by H
NMR, MS and for newly reported compounds, by C, H
analyses. The results obtained with some representative
epoxides are summarised in the Table. The Table shows
that in the presence of the catalyst, as low as 0.2 mol%, the
epoxides undergo facile reaction with various classes of
alcohols namely, aliphatic 1o, 2o, and 3o alcohols 2a-2g,
cyclic alcohol 2h and unsaturated alcohols 2i, 2j with ex-
cellent to good yield (entries 1-10). However, the reac-
tions with unsaturated alcohols are performed at the reflux
temperature of dichloromethane. When the reaction is
performed under identical conditions but in the absence of
Yb(OTf)3, 1a is recovered entirely unreacted (entry 15).
aReaction conditions: substrate epoxide (1.0 mmol), alcohol (7.0 ml)
otherwise mentioned. bProducts are characterised either on the basis
of spectroscopic data reported elsewhere9 or reported here. cIsolated
yield based on epoxide, values within parenthesis refer to GLC
1
yield.17 dLiquid; H NMR (CDCl3) : 7.2-7.4 (m, 5H), 4.3-4.4 (dd,
1H, J = 9.8 and 5.07 Hz), 3.5-3.65 (m, 2H), 3.25-3.4 (m, 2H), 2.5 (br,
OH) 1.5-1.7 (sextet, 2H, J = 7.78 Hz), 0.9 (t, 3H, J = 9.86 Hz); EIMS
e
1
e/z: 180. Liquid; H NMR (CDCl3) : 7.2 (s, 5H), 4.2-4.3 (dd, 1H,
J = 8.7 and 5.12 Hz), 3.4-3.6 (m, 2H), 3.2-3.35 (m, 2H), 2.69 (br, 1H),
1.4-1.55 (m, 2H), 1.2-1.4 (m, 2H), 0.8 (t, 3H, J = 8.7 Hz); EIMS e/z:
Products 3a-3j are identified as 2-alkoxy-2-phenyl etha-
nol by oxidation to the corresponding aldehyde, which
demonstrates that the epoxide 1a undergoes a completely
regioselective ring-opening under the influence of the cat-
alyst (Equation 1a). A similar facile reaction of 4 with al-
cohols affords trans-hydroxy ethers 5a and 5b in excellent
yield signifying that the reaction is highly anti-stereose-
lective (Equation 1b). However, the product obtained
from the reaction of 6 with MeOH resulted in a mixture of
exo- and endo-OMe isomers in the ratio 70:30, as deter-
mined by 1H NMR (Equation 1c). The results are obvious
consequence of the involvement of a carbocation interme-
diate.9,12 In the case of styrene oxide 1a the benzylic cat-
ion, being far more stabilised than the primary
carbocation,19 becomes the favourable site for the nucleo-
philic attack, thus giving rise to the formation of 2-alkoxy-
f
1
194. Liquid; H NMR (CDCl3) : 7.3 (m, 5H), 4.45-4.55 (dd,1H,
J = 9.04 and 4.76 Hz), 3.3-3.5 (m, 1H), 3.5-3.6 (m, 2H), 2.4 (br, 1H),
1.3-1.8 (m, 2H), 1.17 (d, 1H, J = 7.85 Hz), 1.05 (d, 1H, J = 8.0 Hz),
1.0-0.8 (m, 3H); EIMS e/z: 193. gLiquid; 1H NMR (CDCl3) : 7.3 (s,
5H), 4.5 (dd, 1H, J = 10.24 and 5.12 Hz), 3.55 (d, 2H, J = 5.6 Hz),
3.25 (m, 1H), 2.2 (br, 1H), 1.95 (m, 1H), 1.7 (m, 3H), 1.1-1.6 (m, 5H);
EIMS e/z: 220. hEpoxide (1.0 mmol), alcohol (2.0 mmol), methylene
chloride (7.0 ml). iLiquid; 1H NMR (CDCl3) : 7.25-7.4 (m, 5h), 5.8-
6.05 (m, 1H), 5.15-5.35 (m, 2H), 4.5 (dd, 1H, J = 8.72 and 4.18 Hz),
3.95-4.1 (m, 1H), 3.6-3.9 (m, 4H); EIMS e/z: 178. jLiquid; 1H NMR
(CDCl3) : 7.25-7.4 (m, 5H), 4.65 (dd, 1H, J = 9.21 and 5.63 Hz), 4.2
(dd, 1H, J = 15.88 and 2.56 Hz), 3.95 (dd, 1H, J = 15.88 and 2.56
Hz), 3.5-3.8 (m, 2H), 2.4 (t, 2H, J = 1.7 Hz); EIMS e/z: 176. kLiquid;
1H NMR (CDCl3) : 3.35 (s, 3H), 2.75-2.85 (m, 1H), 1.9-2.0 (m, 2H),
1.0-1.4 (m, 5H). EIMS e/z: 130. lExo-OMe:endo-OMe = 70:30 (by
1H NMR); EIMS e/z: 142. mCatalyst, recycled after single use, is
2-phenyl ethanol 3a-3j with essentially complete regiose- employed18. Solid; H NMR (CDCl3) : 7.3 (m, 10H), 4.9 (d, 1H,
lectivity. In the case of trans-stilbene oxide 1b the product
3k is obtained as mixture of two diastereoisomers in the
n
1
J = 3.5 Hz), 4.35 (d, 1H, J = 2.50 Hz), 3.25 (s, 3H), 2.3 (s, 1H) (threo
isomer); 7.3 (m, 10H), 4.7 (d, 1H, J = 9.34 Hz), 4.1 (d, 1H, J = 9.86
Hz), 3.35 (s, 3H), 2.5 (s, 1H) (erythro isomer); EIMS e/z: 228.
racemic form with a modest diastereoselectivity, as deter-
mined by 1H NMR20 and chiral HPLC21
(threo:erythro = 71:29, entry 16). The formation of two
diastereoisomers is again indicative of a dissociative
mechanism. The ratio threo:erythro is found to be virtual-
ly unaltered (within experimental error) when the reaction
is quenched at lower conversion level (40%, GLC). Fur-
thermore, a pure sample of the threo isomer (racemic mix-
ture) of 1,2-diphenyl-2-methoxy ethanol does not undergo
transformation to the erythro form under the experimental
conditions. Thus, the possibility of isomerisation of the
product by the influence of the catalyst is ruled out. When
1b is allowed to react with methanol in the presence of
L-(+)-diethyl tartrate (3 equivalents with respect to the
catalyst) under otherwise identical conditions, high dias-
tereoselectivity (threo:erythro = 97:3) is observed.21 Here
oThreo:erythro = 71:29 (determined by 1H NMR and by chiral
HPLC20).
again, both the isomers are formed as racemic mixtures.
The observed results can be plausibly explained by an in-
termediate complex formation between the epoxide and
chiral Lewis acid catalyst, in which the subsequent ap-
proach of the nucleophile is preferred form one face (at-
tack A in the Scheme). This results in an apparent
diastereo face differentiation. The reusability of the cata-
lyst has been tested by exposing the catalyst, isolated from
the reaction mixture (cf. Table, entry 2),18 to a fresh batch
of 1a under identical reaction conditions (entry 14) with
the yield of the product 3b remaining virtually unaltered
(within the experimental error).
Synlett 2001, No. 6, 836–838 ISSN 0936-5214 © Thieme Stuttgart · New York