[2,3] Sigmatropic rearrangement of unstable sulfur ylides from allyl sulfonium salts. Comparative study of electrochemical reduction with the base method and mechanism elucidation by the MO method

Article abstract of DOI:10.1246/bcsj.77.1687

The cathodic reduction of sulfonium salts in acetonirile under the presence and absence of benzaldehyde were carried out and compared with the results of the base method. Under the presence of benzaldehyde, the electrochemical reduction gave epoxides as a

Full text of DOI:10.1246/bcsj.77.1687

ꢀ 2004 The Chemical Society of Japan
Bull. Chem. Soc. Jpn., 77, 1687–1695 (2004) 1687
[2,3] Sigmatropic Rearrangement of Unstable Sulfur Ylides from Allyl
Sulfonium Salts. Comparative Study of Electrochemical Reduction with
the Base Method and Mechanism Elucidation by the MO Method
ꢀ
Yuichi Okazaki, Fumio Ando, and Jugo Koketsu
Department of Applied Chemistry, College of Engineering, Chubu University, Kasugai 487-8501
Received January 13, 2004; E-mail: jkoketsu@isc.chubu.ac.jp
The cathodic reduction of sulfonium salts in acetonirile under the presence and absence of benzaldehyde were car-
ried out and compared with the results of the base method. Under the presence of benzaldehyde, the electrochemical
reduction gave epoxides as a result of a Corey–Chaykovsky reaction confirming ylide formation. The electrochemical
reduction of sulfonium salt without benzaldehyde yields the same [2,3] sigmatropic rearrangement product as those ob-
tained by the base method, in high yield. The reaction mechanism was studied by semi-emprical MO methods.
Organo-sulfur compounds have attracted attention as inter-
mediates in the syntheses of various organic compounds.1 Es-
pecially the sulfur ylides (abbreviated as S-Y hereafter) have
played an important role as sources of versatile carbanions.2
Corey–Chaykovsky reactions are well known as the important
reaction for getting a wide variety of epoxides derivatives.3,4
The reactions of S-Y with carbonyl compounds are different
from those of phosphonium ylide, because the carbanion of
S-Y attacks the carbonyl carbon and yields epoxides because
of the weak affinity of sulfur atom to oxygen.
Besides their reactions with many electrophiles as carban-
ions, the S-Ys appear to be involved in two types of rearrange-
ments as their characteristic reactions. Thomson and Stevens
reported sulfur versions of the Stevens rearrangement which
proceeded via a S-Y.5 This rearrangement has been regarded
as a radical process because the concerted mechanism is for-
bidden by the Woodward–Hoffmann rule. But by the MO cal-
culation, we have found the possibilities of a sigmatropic proc-
ess when the migrating groups are hydrogen, vinyl, and silyl
groups which could use their hypervalent bonding.6 On the
contrary, the [2,3] sigmatropic rearrangements are allowed
by the Woodward–Hoffmann rule, and many examples have
been reported so far. Parham and Groen reported the [2,3] sig-
matropic rearrangement of S-Y obtained from the reaction of
allyl sulfide with dichlorocarbene.7 In addition, Vedejs et al.
reported [2,3] sigmatropic rearrangement of S-Y which con-
tains allylic substituent to give macrocyclic molecules.8 An
ab initio theoretical study from a view point of stereochemistry
is reported and the energy differences in the transition state are
discussed.9 And the rearrangement of benzyl dimethylsulfoni-
um ion to o-methylbenzylmethylsulfide in the presence of
amide ion is a normal Sommelet–Hauser rearrangement
through the ylide as an intermediate.10 Okada et al.11 reported
the synthesis of thiepin derivatives through skeletal reconstruc-
tion of S-Y with two benzyl groups on the sulfur atom
ꢀ
(Scheme 1). Ab initio MO calculation at HF/6-31G basis
set was used to elucidate the reaction mechanism.12 We have
reported the Sommelet–Hauser rearrangement in electrochemi-
cally generated S-Y (Scheme 2).13
In this experiment, we used the electrochemical reduction
X
X
X
X
-H+
base
+
[1,3]sigmatropic
rearrangement
Sommelet-Hauser
rearrangement
S
+
S
SCH3
-
-
Y
SCH3
CH3
CH3
X=S, SO
Y=BF4
Scheme 1.
S
CH3
-
.
-H
+
+
e
.
Ph
CH2
S
CH2 Ph
S
Ph
CH2
S
-
Sommelet-Hauser
Rearrangement Product
Scheme 2.
Published on the web September 10, 2004; DOI 10.1246/bcsj.77.1687

1688 Bull. Chem. Soc. Jpn., 77, No. 9 (2004)
Rearrangement of S-ylide Generated by EC-Method
(m), 952 (m) cmꢃ1 1H NMR (400 MHz, CDCl3) ꢀ 5.81–5.73
;
method for allyl-substituted sulfonium salts to produce S-Ys
and compared the reaction products with those obtained from
the base method. We found that theꢁallylic S-Y generation oc-
curred under mild conditions at 22 C under a nitrogen atmo-
sphere. The electrochemical reduction method is found to be
simple and clean for giving S-Ys, and easy work-up is also
promised. We obtained the epoxides through the Corey–
Chaykovsky reactions of the allylic S-Y generated by the elec-
trochemical reduction method in the presence of benzalde-
hyde; the epoxides obtained are accompanied by a geometrical
isomerization. But when there are the methyl groups at both
terminals of the allyl group, the generated epoxides were
hydrogenated to the alcohol. Without benzaldehyde, the S-
Ys undergo the [2,3] sigmatropic rearrangement easily. The
reactions of electrogenerated S-Ys with benzaldehyde and
their [2,3] sigmatropic rearrangements were studied by using
allyl-(ATHT) and allyl-(ADES) to see the difference between
ring structure and open chain structure. In order to clarify the
mechanism of their reactions, we optimized all minimum or
saddle points of the reaction profile by a semi-emprical MO
method. The bond energies and spin density distributions were
also used to explain the reaction path in the electrochemical
reduction of the sulfonium salts.
(1H, m, H-4), 5.65 (2H, d, H-6, J ¼ 9:8 Hz) 5.43 (1H, d, H-7, J ¼
9:8 Hz), 4.32 (2H, d, H-3, J ¼ 7:3 Hz), 3.79–3.49 (4H, m, H-1),
2.32–2.26 (4H, m, H-2); 13C NMR (100 MHz, CDCl3) ꢀ 42.96
(C-1), 31.40 (C-2), 44.80 (C-3), 126.93 (C-4), 124.92 (C-5). Anal.
Calcd for C7H13BrS: C, 40.20; H, 6.26%. Found: C, 39.98; H,
6.51%.
1-(2-Butenyl)tetrahydrothiophꢁenium Bromide (b): (E)-Iso-
mer; White crystal; mp 75.4–76.8 C; IR (KBr Tablet) 2951 (m),
2066 (m), 1850–1680 (w), 1662 (m), 1458 (m), 1419 (m), 1258
(m), 975 (m) cmꢃ1 1H NMR (400 MHz, CDCl3) ꢀ 6.22–6.17
;
(1H, m, H-4), 5.52 (1H, s, H-6), 4.35–4.34 (2H, d, H-3, J ¼ 9:7
Hz), 3.79–3.55 (4H, m, H-1), 2.40–2.37 (4H, m, H-2), 1.86 (3H,
s, H-7); 13C NMR (100 MHz, CDCl3) ꢀ 43.00 (C-1), 28.69 (C-
2), 44.16 (C-3), 139.11 (C-4), 111.63 (C-5), 17.93 (C-7). Anal.
Calcd for C8H15BrS: C, 43.05; H, 6.77%. Found: C, 42.89; H,
6.58%.
1-(3-Methyl-2-butenyl)tetrahydrothiophenium Bromide (c):
ꢁ
White crystal; mp 79.0–81.2 C; IR (KBr Tablet) 2937 (m), 2362
(w), 1850–1680 (w), 1662 (m), 1446 (m), 1415 (m), 1245 (m), 846
(m) cmꢃ1; 1H NMR (400 MHz, CDCl3) ꢀ 5.47–5.44 (1H, m, H-4),
4.36 (2H, d, H-3, J ¼ 9:8 Hz), 3.80–3.48 (4H, m, H-1), 2.37–2.35
(4H, m, H-2), 1.86 (3H, s, H-6), 1.81 (3H, s, H-7); 13C NMR (100
MHz, CDCl3) ꢀ 41.11 (C-1), 28.88 (C-2), 42.93 (C-3), 146.82 (C-
4), 111.60 (C-5), 19.22 (C-6), 25.61 (C-7). Anal. Calcd for
C9H17BrS: C, 45.57; H, 7.22%. Found: C, 45.68; H, 7.16%.
ADES. 1-Allyldiethylsulfonium Bromide (d): White crys-
tal; mp 68.3–70.6 ꢁC; IR (KBr Tablet) 2939 (m), 2061 (m), 1850–
1650 (w), 1635 (m), 1456 (m), 1423 (m), 1263 (m), 995 (m), 952
(m) cmꢃ1; 1H NMR (400 MHz, CDCl3) ꢀ 5.64–5.62 (1H, m, H-4),
5.54 (2H, d, H-6, J ¼ 16:3 Hz) 5.32 (1H, d, H-7, J ¼ 9:8 Hz),
4.29 (2H, d, H-3, J ¼ 7:32 Hz), 3.44–3.41 (4H, m, H-1), 1.24–
1.20 (6H, m, H-2); 13C NMR (100 MHz, CDCl3) ꢀ 33.03 (C-1),
9.26 (C-2), 41.06 (C-3), 126.96 (C-4), 123.53 (C-5). Anal. Calcd
for C7H15BrS: C, 39.82; H, 7.16%. Found: C, 39.96; H, 6.99%.
Experimental
Materials. Tetrahydrothiophene, diethyl sulfide, 3-bromopro-
pene, 1-bromo-2-butene, 4-bromo-2-methyl-2-butene, benzalde-
hyde, and acetonitrile are commercial products and used were af-
ter being purified by distillation. Sodium amide and tetraethyl-
ammonium perchlorate were used without further purification.
1H NMR and 13C NMR were recorded on JEOL JNM-A series in-
struments in CDCl3 solvent. Mass spectra were obtained on a SHI-
MADZU GCMS-QP2010. For preparative column chromatogra-
phy, DB-5MS (30 m ꢂ 0.320 mm ꢂ 0.50 mm) were employed.
Each melting point was exhibited on a SEIKO INSTRUMENTS
TG-DTA SSC/5200H. IR spectra were recorded on a SHIMA-
DZU FTIR-8700. Current was recorded on NIKKO KEISOKU
DIGITAL COULOMB METER NDCM-1. Column chromatogra-
phy for purification was performed using Mallinckrodt silica gel
(100 mesh) with CHCl3 as eluent.
Synthesis of the Sulfonium Salts. In diethyl ether, the sulfo-
nium salts are synthesized by the reaction of tetrahydrothiophene
or diethyl sulfide with equimolar amounts of the corresponding
allyl bromides; 3-bromopropene, 1-bromo-2-butene or 4-bromo-
2-methyl-2-butene under argon at r.t. The melting point, IR spec-
tra, 1H NMR, 13C NMR spectra data, and elemental analysis of the
ATHT and ADES are shown below. The numbering systems of
the compounds are specified as described in Chart 1.
1-(2-Butenyl)diethylsulfonium Bromide (e):
(E)-Isomer;
White crystal; mp 70.4–71.5 ꢁC; IR (KBr Tablet) 2950 (m),
2064 (m), 1850–1680 (w), 1662 (m), 1456 (m), 1419 (m), 1259
(m), 975 (m) cmꢃ1 1H NMR (400 MHz, CDCl3) ꢀ 6.35–6.30
;
(1H, m, H-4), 5.58 (1H, s, H-6), 4.57 (2H, d, H-3, J ¼ 9:7 Hz),
3.76–3.66 (4H, m, H-1), 1.83 (3H, s, H-7) 1.27–1.24 (6H, m, H-
2); 13C NMR (100 MHz, CDCl3) ꢀ 32.32 (C-1), 8.90 (C-2),
40.51 (C-3), 138.76 (C-4), 115.45 (C-5), 17.25 (C-7). Anal. Calcd
for C8H17BrS: C, 42.67; H, 7.61%. Found: C, 42.60; H, 7.87%.
1-(3-Methyl-2-butenyl)diethylsulfonium
Bromide
(f):
White crystal; mp 73.1–74.3 ꢁC; IR (KBr Tablet) 2930 (m),
1850–1680 (w), 1662 (m), 1450 (m), 1415 (m), 1261 (m), 979
(w), 839 (w) cmꢃ1 1H NMR (400 MHz, CDCl3) ꢀ 5.56–5.50
;
(1H, m, H-4), 4.56 (2H, d, H-3, J ¼ 9:8 Hz), 3.78–3.66 (4H, m,
H-1), 1.86 (3H, s, H-6), 1.80 (3H, s, H-7) 1.27–1.24 (6H, m, H-
2); 13C NMR (100 MHz, CDCl3) ꢀ 33.45 (C-1), 10.10 (C-2),
38.27 (C-3), 147.35 (C-4), 109.45 (C-5), 19.36 (C-6), 25.62 (C-
7). Anal. Calcd for C9H19BrS: C, 45.19; H, 8.01%. Found: C,
45.39; H, 8.07%.
ATHT. 1-Allyltetrahydrothiophenium Bromide (a): White
crystal; mp 71.3–73.2 ꢁC; IR (KBr Tablet) 2947 (m), 2077 (w),
1850–1650 (w), 1633 (m), 1456 (m), 1421 (m), 1259 (m), 1001
1
2
(a) R=H, R'=H, X=Br
(b) R=Me, R'=H, X=Br
(c) R=Me, R'=Me, X=Br
R
Electrochemical Reduction. In a 0.1 dL electrolytic cell
which has a porous cup made of ceramics as a divided membrane,
0.2 mol of the sulfonium salt was dissolved in 0.05 dL in acetoni-
trile at the cathodic side of the cell. In the anodic side, 0.01 dL of
acetonitrile that contains 0.01 mol of tetraethylammonium per-
chlorate was used as electrolyte. As the electrodes, two platinum
plates (2 ꢂ 2 cm) were installed in the cell. The cathodic reduction
was carried out under condition of ꢃ2:5 V vs SCE at 22 ꢁC in ni-
H
C
4
-
+
6
X
CH2
C
s
3
5
R'
7
1
2
(d) R=H, R'=H, X=Br
(e) R=Me, R'=H, X=Br
(f) R=Me, R'=Me, X=Br
R
H
C
4
-
+
6
X
CH2
C
s
3
5
R'
7
Chart 1.

Y. Okazaki et al.
Bull. Chem. Soc. Jpn., 77, No. 9 (2004) 1689
trogen gas flow. After 1 F/mol electricity based on the sulfonium
salts was passed, the reaction products were dissolved in diethyl
ether and extracted with the same amount of water to remove in-
organic salts. And the organic solution was dried over Na2SO4.
After the solvent was removed by evaporation, the products were
purified by column chromatography on silica gel adsorbent using
CHCl3 as eluent. In case of the Corey–Chaykovsky reactions, an
equimolar amount of benzaldehyde was dissolved in the cathodic
solution and the same work-ups were carried out.
Base Method. In acetonitrile solvent, 0.2 mol of the sulfoni-
um salts were dissolved and an equimolar amount of sodium
amide was added. Then the mixture was refluxed for 2 h. The
work-up method was the same as in the case of the electrochem-
ical reduction method.
Computational Method. Semi-empirical calculations were
carried out using PM3 Hamiltonian implemented in Prime Power
200 at Chubu university computer center. Each transition state
possesses only one imaginary frequency for the vibrational analy-
sis, corresponding to the appropriate coordinate of the reaction.
An IRC calculation was performed for the transition state to con-
firm that it is a true transition state.
Reaction Products. The reaction products are the same in
both electrochemical and base methods. And also the yields in
the methods are comparable. When the sulfonium salt is reacted
by each method, the rearranged products are obtained in high yield
and the yields become higher as the numbers of terminal methyl
groups increased. The products [2] and [8] were the mixtures of
(R)- and (S)-isomers which can be differentiated by 1H NMR
and 13C NMR. However, they could not be separated by column
chromatography, so the combined yields are shown in Table 1.
During the electrochemical reduction of sulfonium salts in the
presence of benzaldehyde, the epoxide was formed with geomet-
rical isomer. And when the methyl groups exist at the two termi-
nals of the allyl group, the generated epoxide is hydrogenated to
the alcohol. The reaction products from ATHT and ADES are
summarized in Table 1; they were numbered as [1]–[9] and their
yields are recorded in the same table. The ratio of geometrical iso-
mer in epoxide is determined by the peak area in GC–MS spectra
because it can not be isolated. Their structures are determined
based on the physical properties, spectral data, and elemental anal-
yses listed below.
Products from ATHT.
2-Allyltetrahydrothiophene [1]:
Yellow liquid; bp 145–147 ꢁC; IR (KBr Coat) 2958 (m), 2927
(m), 2860 (m), 1850–1650 (w), 1685 (m), 1458 (m), 1377 (m),
1
1260 (m), 952 (m), 912 (m) cmꢃ1; H NMR (400 MHz, CDCl3)
ꢀ 7.27 (1H, m, H-6), 5.76–5.69 (2H, m, H-7), 5.01–4.93 (1H,
m, H-4), 3.31 (2H, t, H-5, J ¼ 7:8 Hz), 2.79 (2H, t, H-1, J ¼
7:6 Hz), 2.28–1.97 (2H, m, H-2), 1.81–1.51 (2H, m, H-3);
13C NMR (100 MHz, CDCl3) ꢀ 41.38 (C-1), 29.89 (C-2), 32.01
(C-3), 47.89 (C-4), 36.39 (C-5), 136.39 (C-6), 115.84 (C-7); MS
m=z 129 [Mþ], 113 (16), 100 (29), 87 (34). Anal. Calcd for
C7H12S: C, 65.56; H, 9.43%. Found: C, 65.16; H, 9.81%.
2-(1-Methylallyl)tetrahydrothiophene [2]: Mixture of (R)-
and (S)-isomers. Yellow liquid; bp 146–148 ꢁC; IR (KBr Coat)
3076 (m), 2947 (s), 2862 (m), 1850–1650 (w), 1637 (m), 1440
(m), 1371 (m), 995 (m), 914 (m) cmꢃ1; MS m=z 141 [Mþ], 129
(12), 113 (28), 100 (41), 87 (54). Anal. Calcd for C8H14S: C,
67.54; H, 9.92%. Found: C, 67.50; H, 10.06%. Main fraction.
1H NMR (400 MHz, CDCl3) ꢀ 7.25 (1H, m, H-7), 5.77–5.70
(2H, m, H-8), 4.99–4.89 (1H, m, H-4), 3.22–3.01 (1H, m, H-5),
Table 1. Reduction Products and Their Yields by ATHT and ADES
Synthetic methods and yield
Electro reduction
Salt Electro reduction
Yield/%
75
Yield/%
Base method
Yield/%
88
with PhCHO
1
1
2
2
[1]
[4]
[1]
[2]
[3]
[7]
[8]
[9]
6
7
S
S
2
1
3
4
E 45
Z 37
5
3
3
8
CH CH CH
H2C
(a)
4
4
6
7
6
7
C
H
C
O
CH2
CH2
H2C
5
H2C
5
H
1
1
2
3
2
3
[2]
[3]
[7]
[8]
[9]
[5]
2
HC
7
8
S
S
3
4
2E,4E 31
2E,4Z 22
5
6
9
CH CH CH
4
CH
4
CH
(b)
(c)
(d)
(e)
(f)
76
84
52
74
74
8
79
92
62
75
90
8
7
7
C
5
5
CH2
CH2
C
1 CH3
O
H
H
H3C 6
H3C 6
1
1
2
3
2
3
[6]
S
4
S
4
8
6
9
6
3
C
4
CH
5
2
7
5
62
9
8
10
8
9
CH CH2
H3C
H3C
H3C
6
6
CH2
CH2
C
C
C
C
1CH3 11 OH
5
5
H
H
H3C
H3C
7
7
1
1
2
3
[4]
2
3
7
S
S
1
2
3
4
E 43
Z 32
8
CH CH CH
H2C
4
4
6
7
6
7
C
O
C
CH2
CH2
H2C
H2C
5
H
H
5
1
1
2
3
2
3
[5]
2
HC
7
8
S
S
3
4
2E,4E 31
2E,4Z 21
5
6
9
4
CH
CH CH CH
4
CH
7
C
8
7
8
5
5
CH2
CH2
C
1 CH3
O
H
H
H3C 6
H3C 6
1
1
2
3
2
3
[6]
S
4
S
4
8
9
6
4
CH
5
3
C
2
7
8
9
8
9
64
10
H3C
H3C
CH CH2
H3C
6
6
CH2
C
CH2
C
C
C
5
5
H
1CH3 11 OH
H
H3C
H3C
7
7

1690 Bull. Chem. Soc. Jpn., 77, No. 9 (2004)
Rearrangement of S-ylide Generated by EC-Method
2.78–2.71 (2H, m, H-1) 2.32–2.01 (2H, m, H-2), 1.95–1.82 (2H,
m, H-3), 1.03–1.00 (3H, d, H-6, J ¼ 6:6 Hz); 13C NMR (100
MHz, CDCl3) ꢀ 44.02 (C-1), 30.98 (C-2), 32.10 (C-3), 54.87
(C-4), 34.90 (C-5), 19.26 (C-6), 142.42 (C-7), 114.10 (C-8). Mi-
nor fraction. 1H NMR (400 MHz, CDCl3) ꢀ 7.25 (1H, m, H-7),
5.77–5.70 (2H, m, H-8), 4.99–4.89 (1H, m, H-4), 3.22–3.01
(1H, m, H-5), 2.78–2.71 (2H, m, H-1) 2.32–2.01 (2H, m, H-2),
1.95–1.82 (2H, m, H-3), 1.03–1.00 (3H, d, H-6, J ¼ 6:6 Hz);
13C NMR (100 MHz, CDCl3) ꢀ 44.92 (C-1), 30.62 (C-2), 32.22
(C-3), 54.92 (C-4), 35.34 (C-5), 19.84 (C-6), 142.32 (C-7),
114.01 (C-8).
(C-7), 127.72 (C-8), 128.06 (C-9); MS m=z 160 [Mþ], 145 [Mþ
ꢃ CH3] (15), 131 (29), 117 (43), 105 (55), 91 (69).
2-Methyl-5-phenyl-2-pentene-4-ol [6]:
Yellow liquid; bp
124–126 ꢁC; IR (KBr Coat) 3398 (s), 2972 (m), 2914 (m),
1850–1750 (w), 1701 (m), 1492 (m), 1450 (m), 1363 (m), 1026
;
(m), 758 (m), 700 (s) cmꢃ1 1H NMR (400 MHz, CDCl3) ꢀ
7.34–7.19 (3H, m, H-8,9,10), 5.13 (1H, d, H-3, J ¼ 6:8 Hz),
4.59 (1H, t, H-4, J ¼ 6:6 Hz), 2.44–2.38 (2H, m, H-6), 1.80
(1H, s, OH) 1.69 (3H, s, H-2, trans) 1.56 (3H, s, H-1, cis);
13C NMR (100 MHz, CDCl3) ꢀ 25.73 (C-1), 17.78 (C-2), 144.17
(C-3), 119.17 (C-4), 73.89 (C-5), 38.04 (C-6), 135.18 (C-7),
125.72 (C-8), 127.17 (C-9), 128.13 (C-10); MS m=z 176 [Mþ],
161 [Mþ ꢃ CH3] (15), 143 (33), 128 (48), 117 (59), 107 (69),
103 (73), 91 (85). Anal. Calcd for C12H16O: C, 81.77; H,
9.15%. Found: C, 82.14; H, 8.73%.
2-(1,1-Dimethylallyl)tetrahydrothiophene [3]: Yellow liq-
uid; bp 148–150 ꢁC; IR (KBr Coat) 3080 (m), 2929 (s), 2860
(m), 1850–1650 (w), 1637 (m), 1487 (m), 1414 (m), 1377 (m),
;
1002 (m), 912 (m) cmꢃ1 1H NMR (400 MHz, CDCl3) ꢀ 7.31
(1H, m, H-8), 5.98–5.91 (2H, m, H-9), 4.99–4.89 (1H, m, H-4),
2.80–2.72 (2H, m, H-1) 2.16–2.11 (2H, m, H-2), 1.97–1.82 (2H,
m, H-3), 1.06 (6H, s, H-6,7); 13C NMR (100 MHz, CDCl3) ꢀ
39.53 (C-1), 31.29 (C-2), 31.72 (C-3), 59.70 (C-4), 31.98 (C-5),
25.02 (C-6), 24.93 (C-7), 145.56 (C-8), 111.41 (C-9); MS m=z
156 [Mþ], 141 [Mþ ꢃ CH3] (15), 129 (27), 113 (43), 100 (56),
87 (69). Anal. Calcd for C9H16S: C, 69.16; H, 10.32%. Found:
C, 69.31; H, 10.33%.
Products from ADES. Ethyl 1-Methyl-3-butenyl Sulfide
[7]: Yellow liquid; bp 140–142 ꢁC; IR (KBr Coat) 3082 (m),
2960 (m), 2825 (m), 1850–1650 (w), 1643 (m), 1438 (m), 1375
(m), 1261 (m), 914 (m) cmꢃ1 1H NMR (400 MHz, CDCl3) ꢀ
;
7.20 (1H, m, H-6), 5.76–5.69 (2H, m, H-7), 5.03–4.98 (1H, m,
H-4), 2.52–2.47 (2H, m, H-5), 2.51–2.44 (2H, m, H-1), 1.20–
1.17 (6H, m, H-2,3); 13C NMR (100 MHz, CDCl3) ꢀ 38.97 (C-
1), 14.75 (C-2), 20.69 (C-3), 41.24 (C-4), 24.27 (C-5), 135.67
(C-6), 116.78 (C-7); MS m=z 131 [Mþ], 117 (14), 105 (26), 91
(40). Anal. Calcd for C7H12S: C, 64.55; H, 10.83%. Found: C,
64.45; H, 11.03%.
3,4-Epoxy-4-phenyl-1-butene [4]: Mixture of (E)- and (Z)-
isomers; Yellow liquid; bp 180–208 ꢁC; IR (KBr Coat) 3085
(m), 3031 (m), 2991 (m), 1850–1650 (w), 1637 (m), 1496 (m),
1456 (m), 1440 (m), 1388 (m), 985 (m), 925 (m), 871 (s), 752
(s), 698 (s) cmꢃ1. Anal. Calcd for C10H10O: C, 82.16; H,
6.89%. Found: C, 81.99; H, 7.10%. (E)-isomer. 1H NMR (400
MHz, CDCl3) ꢀ 7.18–7.10 (3H, m, H-6,7,8), 5.60–5.52 (1H, m,
H-2), 5.36–5.25 (1H, m, H-1, cis), 5.17–5.07 (1H, m, H-1, trans),
3.59 (1H, d, H-4, J ¼ 2:0 Hz), 3.19 (1H, t, H-3, J ¼ 2:2 Hz);
13C NMR (400 MHz, CDCl3) ꢀ 119.20 (C-1), 134.94 (C-2),
59.87 (C-3), 62.59 (C-4), 131.92 (C-5), 125.24 (C-6), 127.95
(C-7), 128.23 (C-8); MS m=z 146 [Mþ], 131 [Mþ ꢃ CH3] (15),
117 (29), 105 (41), 89 (56). (Z)-isomer. 1H NMR (400 MHz,
CDCl3) ꢀ 7.18–7.10 (3H, m, H-6,7,8), 5.60–5.52 (1H, m, H-2),
5.36–5.25 (1H, m, H-1, cis), 5.17–5.07 (1H, m, H-1, trans), 4.05
(1H, d, H-4, J ¼ 4:4 Hz), 3.47 (1H, t, H-3, J ¼ 4:2 Hz); 13C NMR
(100 MHz, CDCl3) ꢀ 121.57 (C-1), 131.92 (C-2), 58.47 (C-3),
59.44 (C-4), 136.83 (C-5), 126.17 (C-6), 127.44 (C-7), 127.87
(C-8); MS m=z 146 [Mþ], 131 [Mþ ꢃ CH3] (15), 117 (29), 105
(41), 89 (56).
1,2-Dimthyl-3-butenyl Ethyl Sulfide [8]: Mixture of (R)- and
ꢁ
(S)-isomers. Yellow liquid; bp 142–144 C; IR (KBr Coat) 3078
(m), 2945 (s), 2862 (m), 1850–1650 (w), 1647 (m), 1440 (m),
1374 (m), 999 (m), 914 (m) cmꢃ1; MS m=z 145 [Mþ], 131 (14),
117 (14), 105 (26), 91 (40). Anal. Calcd for C8H14S: C, 66.60;
1
H, 11.18%. Found: C, 66.50; H, 11.06%. Main fraction. H NMR
(400 MHz, CDCl3) ꢀ 7.23 (1H, m, H-7), 5.80–5.73 (2H, m, H-8),
4.99–4.94 (1H, m, H-4), 2.51–2.49 (2H, m, H-1), 2.49 (2H, t, H-5,
J ¼ 7:6 Hz), 1.19–1.16 (6H, m, H-2,3), 1.03–1.00 (3H, d, H-6,
J ¼ 6:8 Hz); 13C NMR (100 MHz, CDCl3) ꢀ 42.53 (C-1), 14.69
(C-2), 17.25 (C-3), 44.96 (C-4), 24.72 (C-5), 18.21 (C-6),
1
140.91 (C-7), 114.39 (C-8). Minor fraction. H NMR (400 MHz,
CDCl3) ꢀ 7.23 (1H, m, H-7), 5.80–5.73 (2H, m, H-8), 4.99–4.94
(1H, m, H-4), 2.51–2.49 (2H, m, H-1), 2.49 (2H, t, H-5, J ¼ 7:6
Hz), 1.19–1.16 (6H, m, H-2,3), 1.03–1.00 (3H, d, H-6, J ¼ 6:8
Hz); 13C NMR (100 MHz, CDCl3) ꢀ 42.29 (C-1), 15.46 (C-2),
17.64 (C-3), 44.84 (C-4), 24.97 (C-5), 18.21 (C-6), 141.39 (C-
7), 114.33 (C-8).
4,5-Epoxy-5-phenyl-2-pentene [5]: Mixture of (2E,4E)- and
(2E,4Z)-isomers. Yellow liquid; bp 161–200 ꢁC; IR (KBr Coat)
3075 (m), 3029 (m), 2991 (m), 1850–1650 (w), 1635 (m), 1492
(m), 1465 (m), 1440 (m), 1388 (m), 984 (m), 921 (m), 875 (s),
752 (s), 696 (s) cmꢃ1. Anal. Calcd for C11H12O: C, 82.46; H,
7.55%. Found: C, 82.62; H, 7.25%. (4E)-isomer. 1H NMR (400
MHz, CDCl3) ꢀ 7.16–7.08 (3H, m, H-7,8,9), 5.25–5.22 (1H, t,
H-3, J ¼ 4:3 Hz), 4.95–4.86 (1H, m, H-2), 3.58–3.55 (1H, d, H-
5, J ¼ 4:2 Hz), 3.13–3.11 (1H, t, H-4, J ¼ 10:0 Hz) 1.59–1.55
(3H, d, H-1, J ¼ 11:3 Hz); 13C NMR (100 MHz, CDCl3) ꢀ
17.49 (C-1), 124.50 (C-2), 135.32 (C-3), 59.60 (C-4), 62.48 (C-
5), 125.11 (C-6), 129.22 (C-7), 127.22 (C-8), 127.95 (C-9); MS
m=z 160 [Mþ], 145 [Mþ ꢃ CH3] (15), 131 (29), 117 (43), 105
1,2,2-Trimethyl-3-butenyl Sulfide [9]: Yellow liquid; bp
ꢁ
143–146 C; IR (KBr Coat) 3082 (m), 2968 (s), 2927 (m), 2871
(m), 1850–1650 (w), 1637 (m), 1456 (m), 1413 (m), 1373 (m),
1
1066 (m), 1002 (m), 912 (m) cmꢃ1; H NMR (400 MHz, CDCl3)
ꢀ 7.29 (1H, m, H-8), 5.91–5.83 (2H, m, H-9), 5.01–4.96 (1H, m,
H-4), 2.60–2.52 (2H, m, H-1), 1.26–1.23 (6H, m, H-2,3), 1.11
(6H, s, H-6,7); 13C NMR (100 MHz, CDCl3) ꢀ 40.77 (C-1),
14.74 (C-2), 17.76 (C-3), 50.71 (C-4), 26.19 (C-5), 25.19 (C-6),
23.71 (C-7), 146.23 (C-8), 111.58 (C-9); MS m=z 159 [Mþ],
144 [Mþ ꢃ CH3] (15), 131 (28), 117 (42), 105 (54), 91 (68). Anal.
Calcd for C9H16S: C, 68.28; H, 11.46%. Found: C, 68.20; H,
11.82%.
1
(55), 91 (69). (4Z)-isomer. H NMR (400 MHz, CDCl3) ꢀ 7.16–
7.08 (3H, m, H-7,8,9), 5.82–5.70 (1H, t, H-3, J ¼ 4:3 Hz),
4.95–4.86 (1H, m, H-2), 3.99–3.98 (1H, d, H-5, J ¼ 4:1 Hz),
3.44–3.41 (1H, t, H-4, J ¼ 12:8 Hz), 1.42–1.40 (3H, d, H-1, J ¼
6:8 Hz); 13C NMR (100 MHz, CDCl3) ꢀ 17.66 (C-1), 126.15 (C-
2), 131.21 (C-3), 58.40 (C-4), 59.32 (C-5), 126.07 (C-6), 128.06
Discussion
In our previous report,13 the generation of S-Y from benzyl
sulfonium salts by the electrochemical reduction has been con-
firmed in two ways: by the reaction with benzaldehyde giving

Y. Okazaki et al.
Bull. Chem. Soc. Jpn., 77, No. 9 (2004) 1691
R
+
-
.
S
CH CH
C
C
H
+
A
B
R'
R
R
-
+
e
S
CH2
CH C
.
S
CH2
CH
C
R'
R'
R
+
S
CH2
CH
.
+ H
R'
-
Scheme 3.
epoxides as a result of Corey–Chaykovsky reaction, and also
by the Sommelet–Hauser rearrangement that is a [2,3] sigma-
tropic rearrangement of a substituent on the S atom onto ben-
zene ring. Their reaction occurred under gentle conditions. In
this study, we found generation of S-Y from allyl sulfonium
salts by the electrochemical reduction.
Without catalyst, we prepared allyl sulfonium salts by the
electrophilic reaction of allyl halides with aliphatic sulfides;
sulfonium salts [b] and [e] were obtained as (E)-isomer only.
This S-salt is more difficult to synthesize than benzyl sulfoni-
um salts because of the unstability of the carbocations supplied
by the allyl halides.
OH
OH
H
H
LiNEt2
+
CH
CH2
H3C C
C
CH2
O
3
2
1
no
E-1
88% 2, 12% 3
100% 2, 0% 3
OH
H3C
Z-1
Scheme 5.
CH3
CH3
H
C
H
C
+
-
S
CH CH
C
S
+
+
PhCHO
C
HC
Ph
H3C
O
The generation of S-Ys by a base such as sodium amide has
been confirmed by reactions with many carbonyl compounds.
And without the presence of carbonyl compounds, we have re-
ported that the benzyl S-Y give symmetrical epoxides derived
from the benzyl groups on the salt as a result of an oxidation
reaction caused by coexisting oxygen.13 However, the allyl S-
Y gave only the rearrangement compounds and no epoxide de-
rived from the oxidation of the allyl group on the S-Y, because
the allyl group has weak affinity to oxygen and is not oxidized
to aldehyde.
Electrochemical generation of S-Y progresses as one-elec-
tron reduction, since sulfonium salts disappeared completely
when 1 F/mol of electric current has been passed. And the
electrochemical reduction might proceed through 2-way mech-
anisms to generate S-Ys as described in Scheme 3. The two
routes were designated as A and B route based on the position-
al difference of hydrogen abstraction. In the presence of benz-
aldehyde, the electrochemical reduction of sulfonium salts, (a),
(b), (d), and (e), gave epoxides in high yields as epoxidation of
S-Y with benzaldehyde followed the A route.
In the case of allyl S-Ys, the epoxides obtained from the A
route isomerized, giving (E)- and (Z)-isomers, although the
benzyl S-Ys gave (E)-epoxide stereoselectively. The detailed
mechanisms of Corey–Chaykovsky reaction of S-Ys have been
calculated by Aggarwal et al.14 using the B3LYP density func-
tion. The reaction of S-Ys with aldehyde proceeds in three-step
reactions through the most stable route, which is shown in
Scheme 4, and the stereoselectivity is determined on the b-
step, where the charged groups rotate to anti from gauche
form. In the case of benzyl S-Ys, the rotation barrier is high
and on b-step, erythro-betaine is stabler than threo-betaine
and only the (E)-epoxide is obtained stereoselectively. And
in the reaction of allyl S-Y, the reason of the contamination
H
H3C
H
C
H3C
C
H
H
H3C
H
H
C
C
HC
CH Ph
HC
C C Ph
C
HC
CH2 Ph
H3C
H3C
O
H3C
OH
OH
Scheme 6.
of (E)- and (Z)-epoxides is because the rotation barrier from
erythro-betaine to threo-betaine is smaller than in the case of
benzyl S-Y, because of the allyl group being a less bulky group
than the benzyl group.
But on sulfonium salts (c) and (f), which have the methyl
groups at both ends of the allyl group, this reduction gives al-
cohols only which were considered as the addition compounds
of H atom to the epoxide. Production of alcohol by the base-
catalyzed ꢁ- or ꢂ-proton elimination pathway from epoxide
have been reported by R. P. Thummel et al.15,16 as shown in
Scheme 5. Electrochemical reduction of epoxides has not been
reported so far. The active H radical comes from the sulfonium
salts or the other activated minor products at the cathode might
cause the hydrogenation of epoxides, because the usual hydro-
gen molecule generation was not observed at the cathodic elec-
trode in this reaction. And the electron donative property of
methyl groups helps the electrophilic radical addition reaction
to the strained epoxide. The reduction process is shown in
Scheme 6.
Without benzaldehyde, the products of [1]–[3] and [7]–[9]
were found to be the rearrangement products formed through
an unstable ylide B, as shown in Scheme 7. The yield of prod-
uct [7] is rather low because of its poor stability. A possible
rearrangement mechanism through stable ylide A is shown
in route A1 that may give eight-membered products. Both
routes A1 and B1 will give transition states obeying the
Woodward–Hoffmann rule, i.e., 5 orbitals and 6 electrons.
But to proceed via unstable Ylide B giving only intriguing
product B indicates that unstable Ylide B may exist in equilib-
rium with stable Ylide A. The B route is supposed to proceed
through the most stable TS to rearrange.
-
O
O
+
-
O
+
S
S
+
O
+
S
+
S
Without any substituent on the allyl group, it would be dif-
ficult to differentiate between the [2,3] Sommelet–Hauser
route B1 and the [1,2] Stevens rearrangement route B2, al-
-
a-step
b-step
c-step
Scheme 4.

1692 Bull. Chem. Soc. Jpn., 77, No. 9 (2004)
Rearrangement of S-ylide Generated by EC-Method
+
though from Ylide A the different products A1 and A2 are
expected. We have reported the mechanisms about the [1,2]
Stevens rearrangement of P-Y and As-Y and found that sigma-
tropic route might be more preferable than radical route with
the migrating groups6 such as hydrogen, vinyl, and silyl
groups. In this study, we should take into consideration the
[1,2] Stevens rearrangement A2, B2 and radical routes A3,
B3, which will proceed through the intermediate IA and IB,
as the alternative routes. However in the reaction, the sigma-
tropic [2,3] Sommelet–Hauser route is supposed to be most
preferable because the activation energies to transition states
TS A1 or TS B1 on sigmatropic route are expected to be re-
duced by allylic substituent. In order to elucidate the rear-
rangement mechanisms, we calculated the energy surfaces
and the geometries of all the reactants, intermediates and tran-
sition states on the surface of the A and B routes by using the
PM3 Hamiltonian; an energy diagram of the reaction of ATHT
that has no substituent on the allyl group is shown in Fig. 1. In
this figure, the heat of formation of proton was subtracted from
that of sulfonium ion; also, from the heats of formation of each
epoxide, that of benzaldehyde was taken away.
The difference in the heat of formations between stable
Ylide A and unstable Ylide B is 9.9 kcal/mol. For the reaction
of S-Y with benzaldehyde, the experimental product is Epox-
ide A, though the heats of formation of the two products were
nearly equal. Because this reaction is similar to that of the ben-
zyl S-Y;13 the stable Ylide A must react with benzaldehyde
immediately after its formation. On the other hand, the rear-
rangement reaction proceeds via unstable Ylide B, giving only
intriguing Product B. At the PM3 level, the Sommelet–Hauser
rearrangement at B route is easier by about 28–34 kcal/mol
than any other route, and it is clear that the dramatic stability
of TS B1 is the main factor of route determination.
δ
S
CH
S
CH
CH
-
δ
CH
C
H
C
H
H
A1
A2
A3
H
TS A1
Product A1
+
δ
S
S
-
+
-
δ
S
CH
CH
CH
CH
CH
CH
C
C
H
H
C
H
H
H
H
TS A2
Ylide A
Product A2
-
+
δ
δ
S
CH
CH
CH
S
H
CH
C
C
H
H
H
TS A3
IA
+
δ
CH2
CH
S
-
δ
H
C
B1
B2
B3
H
TS B1
+
CH2
CH
S
+
S
CH2
CH
δ
S
-
CH2
CH
-
C
δ
H
C
H
C
H
H
H
H
Product B
Ylide B
TS B2
+
δ
CH2
CH
H2C
S
S
CH
-
δ
H2C
H
C
H
TS B3
IB
Scheme 7.
H
C
-H+
+
S
CH2
CH2
Sulfonium ion
114.0
93.5
92.5
TS B3
TS B2
TS A2
TS A3
91.0
87.0
A
B
87.0 TS A1
TS B1
59.2
IA
47.5
46.6
Ylide B
IB
41.0
37.6
Ylide A
Z 36.4 Epoxide A
Epoxide B Z 36.8
35.4
35.4
E
E
H
C
H
H2C
H
S
C
CH2
CH2
C
C
C
S
+
H
O
H
Ph
O
H
-PhCHO
C
-PhCHO
Product A2
Product A1
Ph
10.1
6.0
6.6
Product B
Fig. 1. Energy surfaces in the A and B routes.

Y. Okazaki et al.
Bull. Chem. Soc. Jpn., 77, No. 9 (2004) 1693
The B route, from unstable Ylide B via TS B1 to Product B,
must be the real path of this rearrangement; it involves a cleav-
age of the S–7C ꢃ bond. In case of having a benzyl group on
the S-Y, the rich ꢄ electrons of phenyl group may be expected
to interact with the partial positive charge generated on 7C at
TS B1 resulting in much stable transition state. We calculated
the differences of electron density changes for allyl S-Y be-
tween Ylide A and TS A1, and also between Ylide B and
TS B1 (represented as ꢀAl) and compared with those of ben-
zyl S-Y (represented as ꢀBn) as shown in Fig. 2. As a result, it
emerged that the partial positive charge on S atom of TS B1
was delocalized to allyl group almost equally with the case
of benzyl group; in addition, the delocalization of partial neg-
ative charge on 4C to 5C appeared to be larger in the allyl
group than in the benzyl group. This resonance stabilization
by delocalization of the electron at anionic and cationic charge
on TS B1 decreases the activation energy to TS B1 from un-
stable Ylide B. On the contrary, the partial positive charge
on S atom of TS A1 was distributed locally to adjacent atoms
and was not enough to stabilize TS A1. This is the reason why
no rearrangement via A route occurs.
As another possible mechanism in B route, the [1,2] Stevens
rearrangement gave the TS B2 with higher activation energy
than the TS B1, but this route is more favored than TS B3, rad-
ical route, by 1.0 kcal/mol although it is forbidden by the
Woodard–Hoffmann rule, i.e., 3 orbitals and 4 electrons. To
elucidate the factor of stabilization of TS B2, we compared
0.09 Al
1C
2C
2C
2C
2C
3C
-0.01 Al
1C
-0.46 Al
1C
1C
3C
-0.01 Al
0.10 Al
-0.15 Al
3C
3C
0.26 Al
0.22 Bn
S
S
4C
4C
-0.64
-4.04
4C
S
-0.28 Al
-0.34 Bn
4C
-0.70
-0.52 Al
-0.54 Bn
P
0.69 Al
0.31 Bn
0.21
0.76
-5.66
-5.50
7C
-3.15
7C
0.38 Al
0.35 Bn
5C
-0.11 Al
0.04 Bn
7C
7C
5C
0.18
6C
-0.06 Al
-0.11 Bn
0.36 Al
0.05
0.01 Al
0.22 Bn
0.24 Bn
6C
6C
-0.20 Al
-0.24 Bn
6C
5C
5C
TS A1
TS B1
E=43.7 kcal/mol
E=58.7 kcal/mol
Fig. 2. The differences of electron densities between Ylide
A and TS A1, and between Ylide B and TS B1. Al is allyl
S-Y and Bn is benzyl S-Y.
P-TS B2
S-TS B2
Fig. 3. Electron densities of the S-TS B2 and P-TS B2.
2C
3C
1.52
1C
1.51
2C
1.51
3C
1.52
113.8
1C
111.2
111.7
111.7
118.6
112.7
113.8
1.83
1.52
4C
1.50
4C
1.84
S
113.4
S
3.27
108.3
112.6
2.27
3C
2.22
1.52
5C
1.53
1.75
2C
1.67
7C
118.1
133.6
102.9
1.52
108.3
7C
131.9
1.52
4C
A1
120.0
1.37
5C
108.6
1C
1.85
119.3
1.49
1.33
1.39
109.6
6C
1.85
6C
97.8
124.9
Product A1
S
TS A1
109.3
38.7
3.47
2.81
1.62
7C
2C
1.52
1.52
77.0
1C
3C
132.7
1.43
5C
3C
127.5
111.7
1.52
112.6
114.5
2C
1.52
109.3
1.34
1.52
4C
A2
110.7
1C
1.82
1.51
6C
109.5
115.8
2.85
104.2
1.83
125.6
S
2.10
4C
S
1.53
7C
108.0
1.77
115.4
109.9
1.84
Ylide A
2.12
64.6
125.4
1.33
129.6
128.3
1.49
6C
5C
7C
5C
1.34
1.42
6C
TS A2
Product A2
Fig. 4. Optimized structures of the A route.

1694 Bull. Chem. Soc. Jpn., 77, No. 9 (2004)
Rearrangement of S-ylide Generated by EC-Method
Fig. 5. Optimized structures of the B route.
Table 2. Dihedral Angles of Optimized Structures in A and B Routes
A
B
Ylide A
TS A1
TS A2
Product A1
Product A2
Ylide B
TS B1
TS B2
Product B
S–C7–C6–C5
C4–S–C7–C6
C4–S–C6–C5
S–C4–C5–C6
C4–C5–C6–C7
3.98
49.26
49.40
40.41
ꢃ26:22
12.11
22.27
34.06
40.86
ꢃ40:37
ꢃ1:02
116.76
122.49
99.76
1.26
0.16
1.17
1.81
ꢃ90:23
129.81
86.54
54.67
ꢃ21:43
ꢃ70:80
32.27
ꢃ25:36
ꢃ21:90
48.18
ꢃ59:42
26.81
ꢃ21:30
ꢃ24:89
52.91
ꢃ138:25
75.24
ꢃ16:99
ꢃ10:39
57.29
ꢃ63:17
22.14
ꢃ38:13
ꢃ74:31
134.18
ꢃ37:79
ꢃ2:74
the electron density of TS B2 of S-Y (S-TS B2) with that of P-
Y (P-TS B2) which might be difficult to migrate using sp3 hy-
brid orbital in the allyl group; the results are shown in Fig. 3.
In the P-TS B2, the neutralized 7C has low affinity with the
negatively charged 4C, on the contrary, the positively charged
7C on S-TS B2 has high affinity with the 4C atom. For the cat-
ionic sigmatropic route, the activation energy of S-TS B2 is
lower than P-TS B2 by 15.0 kcal/mol. However, the activation
energies of TS-B2 and TS B3 are much higher than that of TS
B1 of the Sommelet–Hauser rearrangement.
The most favorable B route from TS B1 to Product B is a
[2,3] sigmatropic rearrangement and is suprafacial, as the dihe-
dral angle of C4–S–C7–C6 changes from 32.3ꢁ to 26.8ꢁ, keep-
ing the tetrahydrothiophene ring upon the migration. The geo-
metrical figures of the A and B routes are shown in Figs. 4
and 5 and the relevant dihedral angles are listed in Table 2.
In the case of allyl S-Ys which have the methyl substituents
at the end of allyl group, the mechanism is evidently the B1
route, that is, the [2,3] sigmatropic Sommelet–Hauser route,
judging from the products obtained. As a result, it is obvious
that the [2,3] sigmatropic rearrangement of allyl group is a pre-
dominant route in the electrochemical reduction of the S-salt.
Summary
The electrochemical reductions of allyl sulfonium salts gave
allyl S-Ys with one-electron transfer reactions. The presence
of S-Ys was verified by Corey–Chaykovsky reactions with
benzaldehyde, giving (E)- and (Z)-mixture of the correspond-
ing epoxides. The reactions of sulfonium salts having the
methyl groups at both ends of the allyl group gives alcohols
which were considered to be the addition compounds of H
atom generated at cathodes to the epoxides. Without the pres-
ence of benzaldehyde, the S-Ys gave Sommelet–Hauser rear-
rangement products in good yields. And this rearrangement
was ascertained in the reaction of sulfonium salts with strong
base, sodium amide, and the yields of the two methods are not
so different. By the base method, no oxidation of S-Ys was ob-
served. This fact is different from the case of the S-Ys having a
benzyl group on the S atom. The MO calculations by using
semi-empirical PM3 Hamiltonian clarified that the most
plausible route among the possible rearrangements was the
Sommelet–Hauser rearrangement.

Y. Okazaki et al.
Bull. Chem. Soc. Jpn., 77, No. 9 (2004) 1695
Tetrahedron Lett., 19, 519 (1978).
Y. D. Wu and K. N. Houk, J. Org. Chem., 56, 5657 (1991).
References
9
1 E. J. Corey, K. Lin, and M. Jautelat, J. Am. Chem. Soc., 90,
2724 (1968).
10 W. Ando, T. Yagihara, and T. Migita, Tetrahedron Lett.,
10, 1425 (1969).
2 A. G. Hortmann and D. A. Robertson, J. Am. Chem. Soc.,
89, 5974 (1967).
11 K. Okada and M. Tanaka, J. Chem. Soc., Perkin Trans. 1,
2002, 2704.
3 E. J. Corey and M. Chaykovsky, J. Am. Chem. Soc., 87,
1353 (1965).
12 K. Okada, M. Tanaka, and R. Takagi, J. Phys. Org. Chem.,
16, 271 (2003).
4 O. Tsuge and I. Shinkai, Bull. Chem. Soc. Jpn., 43, 3514
(1970).
13 Y. Okazaki, F. Ando, and J. Koketsu, Bull. Chem. Soc.
Jpn., 76, 2155 (2003).
5
6
T. Tompson and T. S. Stevens, J. Chem. Soc., 1932, 69.
K. Makita, J. Koketsu, F. Ando, Y. Ninomiya, and N.
14 V. K. Aggarwal, J. N. Harvey, and J. Richardson, J. Am.
Chem. Soc., 124, 5747 (2002).
Koga, J. Am. Chem. Soc., 120, 5764 (1998).
W. E. Parham and S. H. Groen, J. Org. Chem., 29, 2214
(1964); 30, 728 (1965).
E. Vedejs, M. J. Mullins, J. H. Renga, and S. P. Singer,
15 R. P. Thummel and B. Rickborn, J. Org. Chem., 37, 3919
(1972).
16 R. P. Thummel and B. Rickborn, J. Org. Chem., 37, 4250
(1972).
7
8