7
0
Can. J. Chem. Vol. 84, 2006
Scheme 1.
PPh /NXS or NXSac(1.2/1.2) X = Br, Cl, I
PhO
3
PhOCH CH(OH)CH X
O
2
2
r.t.,1% aq. CH CN
3
PPh /NXS or NXSac (2.5/2.5)
3
PhOCH CHXCH X
2
2
reflux, CH CN
3
X = Br, Cl
PhO
'
X = Br, Cl, I
O
PPh /NXS or NXSac(1.2/1.2) and
3
PhOCH CHX'CH X
2
2
PPh /NX'S or NX'Sac (2.5/2.5), reflux, CH CN
3
3
tive conversion of oxiranylmethyl phenyl ether to 1-bromo-
-phenoxy-2-propanol. Aqueous acetonitrile is required to
hydrolyze the intermediate to the corresponding vicinal
halohydrine. Similar reaction conditions were then applied
for the conversion of structurally different epoxides to their
corresponding vic-halo alcohols. The results are shown in
Table 1.
and to introduce a mild and novel method for this
transformation, we applied our method to epoxides. For this
transformation, the successive use of two N-halo imides hav-
ing different halide ions produces the desired unsymmetrical
vic-dihalide in high yield. To optimize the reaction condi-
tions, oxiranylmethyl phenyl ether was added to a mixture of
PPh –NCS or PPh –NCSac with the molar ratio of 1.2:1.2 in
3
3
3
Under these reaction conditions, the ethereal bonds, ester
groups, carbon–carbon double bond, and activated aromatic
ring do not interfere with the formation of halo alcohols and
remain intact. Electronic and steric factors control the orien-
tation of the attack of halide ion. In the case of styrene ox-
ide, owing to electronic factors, the attack of the halide ion
occurs mainly at the more substituted position. However, in
other epoxides that bear alkyl or electron-withdrawing
groups, owing to the combination of steric and electronic
factors, the attack occurs at the less hindered site of the ring
and 1-halo-2 alcohols are produced as the major product
acetonitrile at room temperature. When all the epoxide was
consumed, the mixture was poured into another mixture con-
taining PPh3–NBS or PPh3–NBSac (2.5:2.5) in acetonitrile
and refluxed for 8 h. After completion of the reaction, 2-
bromo-1-chloro-3-phenoxypropane was obtained in 85%
yield (Table 3, entry 2). Since this attempt was successful,
we applied it to the preparation of unsymmetrical vic-
dihalides from other epoxides. By using the proper halide
anion in each step, we could easily control the formation of
vic-dihalides with high regioselectivity. The results obtained
for this study are tabulated in Table 3. In these reactions, as
we expected, in the first step, the first halide ion attacks the
epoxide ring from the same side of the formation of vic-halo
alcohols. In the second step, the resulting halo alcohol reacts
(
Table 1).
We also investigated the possibility of converting epoxides
to symmetrical vicinal dihalides using PPh –NXS or PPh –
3
3
with the mixture of PPh –NXS or PPh –NXSac and the sec-
NXSac (X = Cl, Br). For this purpose, oxiranylmethyl
phenyl ether (Table 2, entries 1 and 2) was treated as a
model compound with the PPh –NBS or PPh –NBSac re-
3
3
ond halide anion displaces the hydroxyl group to produce
the desired unsymmetrical dihalide. It was observed that
when one of the nucleophiles was iodide, because of the
possibility of the elimination of IX (X = Cl, Br), alkene for-
mation was a competing pathway and lowered the yield of
vic-dihalides considerably (Table 3, entries 3 and 6).
3
3
agent system using the optimized molar ratio of 2.5:2.5 in
refluxing acetonitrile, and the corresponding dibromide was
obtained in high yield. The use of NCS or NCSac in this
method under the same reaction conditions as used for the
preparation of dibromide offers the possibility of producing
vic-dichlorides. We extended this procedure to other
epoxides to obtain symmetrical vic-dihalides. The results of
this investigation are shown in Table 2.
Although the exact mechanism of these reactions is not
clear, it is suggested that, at first, a positively charged adduct
(I) (31) is formed from the reaction of PPh with NBS
3
(Scheme 2). This positively charged adduct interacts with
the oxygen atom of the epoxide and produces (II). When R-
(–)-styrene oxide, the formation of the corresponding β-
bromohydrin, which did not show any optical rotation, ac-
counts for the formation of the racemic product through the
intermediacy of a benzylic carbocation. Nucleophile attack
of the halide ion on (II) produces the intermediate (III).
This intermediate in the presence of water could produce
triphenylphosphine oxide, succinimide, and the correspond-
ing vic-halo alcohol. Similarly, the attack of another halide
ion on this intermediate produces the corresponding vicinal
symmetrical or unsymmetrical dihalide.
However, the preparation of vic-diiodides by this method
was not successful. This result is in accordance with the
literature (30), which indicates that alkene formation is fa-
voured owing to the ease of elimination of molecular iodine.
In all the reactions of epoxides with PPh –NIS or PPh –
3
3
NISac (2.5:2.5), alkene formation was the only pathway. For
example, the addition of oxiranylmethyl phenyl ether to the
mixture of PPh –NBS or PPh –NBSac (2.5:2.5) in refluxing
3
3
acetonitrile produced allyl phenyl ether as the only product
after 0.5 h. Therefore, the use of a PPh –NIS or PPh –NISac
3
3
mixture offers a new mixed reagent system for the
deoxygenation of epoxides into their corresponding alkenes.
As shown in Table 3, entries 1 and 4, when the first
To overcome the problems of using highly acidic condi-
tions for the conversion of epoxides to vic-dihalides (27),
nucleophile used was bromide,
regioisomers was obtained. It is assumed that, in these cases,
a
mixture of two
©
2006 NRC Canada