LETTER
595
Studies towards the Taming of the ‘Carbocation’ in the Regioselective Ring
Opening of Epoxides to Allylic Alcohols
R
egioselective Rin
e
gOpening
l
of
E
po
e
xides to
A
l
n
A
lcohols A. Chapman, Karim Herbal, William B. Motherwell*
Christopher Ingold Laboratories, Department of Chemistry, University College London, 20 Gordon Street, London, WC1H 0AJ, UK
Fax +44(20)76797524; E-mail: w.b.motherwell@ucl.ac.uk
Received 23 December 2009
Dedicated with respect and much gratitude for a long friendship to Professor Gerald Pattenden FRS, an acknowledged master of molecular
acrobatics in the terpenoid field.
O
Abstract: Regioselective isomerisation of epoxides to allylic alco-
hols can be achieved using p-toluenesulfonic acid in the presence of
O
H
OSO2CF3
OSiMe3
1,3-dimethylimidazolidin-2-one.
Key words: epoxide ring opening, allylic alcohols, isomerisation
LiBr
i) Me3SiOTf
LiClO4
ii) DBU
O
The regioselective isomerisation of an unsymmetrically
substituted epoxide into an allylic alcohol is a very valu-
OSiMe3
able transformation for organic synthesis.1 The Sharpless
protocol2 of ring opening with phenylselenide anion fol-
Scheme 1
lowed by hydrogen peroxide induced syn elimination of
the resultant selenoxide is now a recognised method for
OH
+
O
OH
TsOH (1 equiv)
DMF, r.t., 1.5 h
introduction of the double bond at the less hindered termi-
nus. For alkene formation at the more hindered site, the
use of a stoichiometric amount of a strong chelating base
possessing an oxophilic metal cation, such as lithium di-
ethylamide or diethylaluminium 2,2,6,6-tetramethylpipe-
ridide, has often proven to be the method of choice.3
+
O
OTs
(6%)
(6%)
OH
(6%)
+
+
OCHO
O
(6%)
(31%)
By way of contrast, efforts to control the evolution of an
incipient carbocation intermediate generated by coordina-
tion of a Brønsted or Lewis acid to the oxygen lone pair
can be fraught with danger, inasmuch as several compet-
ing pathways including hydride shift, alkyl shift, proton
loss, and nucleophilic capture can all compete.4 This situ-
ation is summarised in Scheme 1 for the simple case of 1-
methylcyclohexene oxide and clearly indicates that care-
ful selection of both cation and counterion is required to
direct a desired pathway.
Scheme 2
In light of this observation, we reasoned, as shown in
Scheme 3, that the selection of a tetrasubstituted urea as a
more potent nucleophile for capture of the incipient car-
bocation would also lead to an intermediate with some ad-
ditional degree of basic character to trigger an in situ
elimination.
In terms of allylic alcohol formation, our own studies
were therefore inspired by a method developed by
Noyori5 involving sequential ring opening with trimethyl-
silyl triflate in the presence of 2,6-lutidine followed by
anti elimination from the unstable diequatorial triflate us-
ing DBU (Scheme 1). The results of a simple preliminary
experiment in which efforts were made to replace the sil-
icon electrophile by a proton from p-toluenesulfonic acid
and the nucleophile by DMF in order to create a positively
charged leaving group were completely unsuccessful and
merely served to illustrate the aforementioned complexity
of the situation (Scheme 2).
H
••
NR2
O
R
O
OH
N
OH
OTs
R
Scheme 3
Irrespective of the merits or otherwise of the above hy-
pothesis (vide infra), we now report our preliminary ob-
servations in this area. Following on from an initial
screening of several Brønsted acids, ureas, and solvent
systems, p-toluenesulfonic acid (0.5 mol equiv) and the
stealth nucleophile,6 1,3-dimethylimidazolidin-2-one (2.0
mol equiv), were selected for further study in order to
evaluate the scope and utility of this inexpensive system.7
The results of this study are shown in Table 1 and reveal
several features of interest.
SYNLETT 2010, No. 4, pp 0595–0598
0
1.
0
3.
2
0
1
0
Advanced online publication: 08.02.2010
DOI: 10.1055/s-0029-1219373; Art ID: D37609ST
© Georg Thieme Verlag Stuttgart · New York