Reactions of 5-(alkyl)thianthrenium and other sulfonium salts with nucleophiles

Article abstract of DOI:10.1002/1099-1395(200102)14:2<81::AID-POC338>3.0.CO;2-Z

A series of 5-(alkyl)thianthrenium triflates (3a-d, g-i) with alkyl (R) groups Me (a), Et (b), isoPr (c), 2-Bu (d), cyclopentyl (g), cyclohexyl (h) and cycloheptyl (i) were prepared by alkylation of thianthrene (Th) with alkyl formate and trifluoromethanesulfonic (triflic) acid. Benzylation (3f) was achieved with benzyl bromide and silver triflate. 5-(Neopentyl)thianthrenium perchlorate (3e) was prepared by reaction of thianthrene cation radical perchlorate with dineopentyl mercury. Methyl- (4a) and cyclohexyldiphenylsulfonium triflate (4b) were made by alkylation of diphenyl sulfide. Benzyldimethyl- (5a), dibenzylmethyl- (5b) and benzylmethylphenylsulfonium perchlorate (5c) were prepared in standard ways. Reactions of these sulfonium salts with iodide ion and thiophenoxide ion were studied for comparison with our earlier reported reactions of comparable 5-(alkoxy)thianthrenium and methoxydiphenylsulfonium salts. It is deduced that reactions of 3-5 with nucleophiles (Nu^-) I^- and PhS^- follow traditional S_N2 and E2C paths. Thus, the salts 3a-c, e and f gave virtually quantitative yields of RNu and Th, while small amounts of butene(s) were obtained from 3d. The cycloalkyl salts 3g-i gave amounts of cycloalkylNu and cycloalkene typical of competition of S_N2 and E2C routes in the classical reactions of cycloalkyl halides and tosylates with I^- and PhS^- ions. Whereas 4a gave only S_N2 products, 4b gave S_N2 and E2C products typical of S_N2/E2C competition. Among the salts 5a-c displacement of the benzyl group was dominant (5a) or exclusive (5b, c), thus exhibiting the preferential displacement of a benzyl group that has been fully documented in earlier studies of S_N2 reactions. Qualitative comparison showed that 3a (methyl) reacted much faster than 3e (neopentyl) with PhS^-. Unlike alkoxysulfonium salts, the salts 3-5 do not appear to undergo reactions at the sulfonium sulfur atom. Copyright

Full text of DOI:10.1002/1099-1395(200102)14:2<81::AID-POC338>3.0.CO;2-Z

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY
J. Phys. Org. Chem. 2001; 14: 81–89
Reactions of 5-(alkyl)thianthrenium and other sulfonium
salts with nucleophiles
Bo Liu and Henry J. Shine*
Department of Chemistry and Biochemistry, Texas Tech University, Lubbock, Texas 79409, USA
Received 19 May 2000; revised 26 September 2000; accepted 27 September 2000
ABSTRACT: A series of 5-(alkyl)thianthrenium triflates (3a–d, g–i) with alkyl (R) groups Me (a), Et (b), isoPr (c), 2-
Bu (d), cyclopentyl (g), cyclohexyl (h) and cycloheptyl (i) were prepared by alkylation of thianthrene (Th) with alkyl
formate and trifluoromethanesulfonic (triflic) acid. Benzylation (3f) was achieved with benzyl bromide and silver
triflate. 5-(Neopentyl)thianthrenium perchlorate (3e) was prepared by reaction of thianthrene cation radical
perchlorate with dineopentyl mercury. Methyl- (4a) and cyclohexyldiphenylsulfonium triflate (4b) were made by
alkylation of diphenyl sulfide. Benzyldimethyl- (5a), dibenzylmethyl- (5b) and benzylmethylphenylsulfonium
perchlorate (5c) were prepared in standard ways. Reactions of these sulfonium salts with iodide ion and thiophenoxide
ion were studied for comparison with our earlier reported reactions of comparable 5-(alkoxy)thianthrenium and
methoxydiphenylsulfonium salts. It is deduced that reactions of 3–5 with nucleophiles (Nu ) I and PhS follow
traditional SN2 and E2C paths. Thus, the salts 3a–c, e and f gave virtually quantitative yields of RNu and Th, while
small amounts of butene(s) were obtained from 3d. The cycloalkyl salts 3g–i gave amounts of cycloalkylNu and
cycloalkene typical of competition of SN2 and E2C routes in the classical reactions of cycloalkyl halides and tosylates
with I and PhS ions. Whereas 4a gave only SN2 products, 4b gave SN2 and E2C products typical of SN2/E2C
competition. Among the salts 5a–c displacement of the benzyl group was dominant (5a) or exclusive (5b, c), thus
exhibiting the preferential displacement of a benzyl group that has been fully documented in earlier studies of SN2
reactions. Qualitative comparison showed that 3a (methyl) reacted much faster than 3e (neopentyl) with PhS . Unlike
alkoxysulfonium salts, the salts 3–5 do not appear to undergo reactions at the sulfonium sulfur atom. Copyright
2000 John Wiley & Sons, Ltd.
KEYWORDS: thianthrenium; sulfonium; nucleophilic reactions
INTRODUCTION
derivatives to undergo E2C elimination with PhS .2
Instead, reaction at sulfonium sulfur was dominant, with
the formation of Th, ROH and diphenyl disulfide
(DPDS). We attributed this behavior to the remarkable
thiophilicity of PhS and to the ease of displacement of
an alkoxy group attached to sulfur in 1 and 2.2
Recently, we reported the reactions of 5-(alkoxy)thian-
threnium perchlorates (1) with iodide, bromide and
thiophenoxide ions.1,2 Compounds 1 were then relatively
new members of the class of alkoxysulfonium salts,
inviting study of their chemistry. We found that they
engaged in three types of reaction with halide ions,
namely, SN2 substitution, E2C elimination and reaction at
the sulfonium sulfur atom. The last type led to the
formation of thianthrene (Th), the alcohol (ROH)
corresponding with the 5-alkoxy group, and halogen
(iodine or bromine). Similar reactions occurred with the
more commonly known methoxydiphenylsulfonium tet-
rafluoroborate (2). Reactions of 1 and 2 with thiophen-
oxide (PhS ) were different, however. Little or no SN2
reaction occurred, in spite of thiophenoxide’s being a
better nucleophile than halide ions toward carbon, and no
elimination was obtained, even with the cyclohexyloxy
group, in spite of the propensity for cyclohexyl
Trialkyl-, triaryl- and alkylarylsufonium salts are well
known tricordinate organosulfur compounds, and num-
bers of their reactions with nucleophiles have been
reported. Among this class of compounds, 5-(alkyl)thian-
threnium salts (3) are not so well known, and to our
*Correspondence to: H. J. Shine, Department of Chemistry and
Biochemistry, Texas Tech University, Lubbock, Texas 79409, USA.
Contract/grant sponsor: Robert A. Welch Foundation; Contract/grant
number: D-0028.
Copyright 2000 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 81–89

82
B. LIU AND H. J. SHINE
knowledge no studies have ever been reported of their
reactions with nucleophiles such as the halide and
thiolate ions. We are interested in these reactions
particularly as to how they compare with the reactions
of 1. We report here, therefore, the reactions of I and
PhS with the trifluoromethanesulfonate (triflate) and
perchlorate salts 3a–i, in which the alkyl group R is
successively Me (a), Et (b), isoPr (c), 2-butyl (d),
neopentyl (e), benzyl (f), cyclopentyl (g), cyclohexyl (h)
and cycloheptyl (i). Furthermore, although numbers of
reactions of sulfonium salts with nucleophiles are in the
literature, not many are to be found with I and PhS .
Consequently, for comparison with reactions of 3 we
have studied and report the reactions of I and PhS with
methyl- (4a) and cyclohexyldiphenylsulfonium triflate
(4b), and of Br and PhS with a number of sulfonium
perchlorates (5) containing the benzyl group, namely,
benzyldimethyl- (5a), dibenzylmethyl- (5b) and benzyl-
methylphenylsulfonium perchlorate (5c). Our interest in
these reactions, in comparison with reactions of 1 and 2,
was to find if halide and thiophenoxide ions would show
their usual carbon nucleophilicities or if, particularly with
PhS , reaction at sulfur would intervene.
yltins (R4Sn), with R being, for example, Me, Et, Bu,
allyl and vinyl (Ref. 3 for Me2Hg and Et2Hg, Ref. 4 for
R4Sn, R = Me, Et, Bu, vinyl).
Á
In these reactions, the radical R is formed initially in
‡
Á
oxidation of the organometal by Th , and if trapped by
‡
5
Á
Á
another Th gives the 5-(alkyl)thianthrenium ion. If R
‡
‡
Á
is easily oxidized to R by Th , the cation is trapped by
solvent acetonitrile and 3 is not formed.6 Hence this
method has limitations. Thianthrene, like other diaryl
sulfides, is not sufficiently nucleophilic to be alkylated by
methods7,8 that work well with dialkyl and alkyl aryl
sulfides, although Saeva was able to alkylate it by
reaction with p- cyanobenzyl bromide and silver triflate9
(Ref. 7 has numerous references to earlier methods).
Recently, Miyatake et al.7 made 3a in excellent yield by
reaction of Th with methyl formate and triflic acid, and
we used that method to prepare 3a–d, g–i in the present
work. This allowed us for the most part to avoid the use of
organomercurails and organotins for alkylating Th. We
were unable to prepare the 5-(neopentyl) triflate by this
method because of difficulty in preparing neopentyl
formate. Furthermore, we were unable to prepare that
triflate by alkylating Th with neopentyl iodide and silver
triflate. Therefore, we prepared 5-(neopentyl)thianthre-
nium perchlorate (3e) by our earlier method,6 the reaction
of Th ClO4 with dineopentylmercury. Alkylation of
Th with benzyl formate and triflic acid was unsuccessful.
We found that benzyl formate itself reacted violently
with triflic acid. Therefore, alkylation was carried out
with benzyl bromide and silver triflate.9 Attempts to
prepare 5-propylthianthrenium triflate (3j) were thwarted
by partial rearrangement that gave a mixture of 3j and 3c
(shown by NMR); this experience differs from that of
Miyatake et al., who prepared butyldiphenylsulfonium
triflate without rearrangement of the butyl group.7 The
same method was used for preparing 4a and b, while
customary alkylation techniques were used for preparing
5a–c.
‡
Á
Reactions of 3±5 with nucleophiles
The data in Table 1 show that quantitative or close to
quantitative yields of alkyl halide or alkyl phenyl sulfide
and Th were formed when 3a–i reacted with halide or
thiophenoxide ion in acetonitrile solution. Exceptions
were the reactions of 3h, in which the major product of
the cyclohexyl ring was cyclohexene, and 3g and i from
which smaller amounts of cycloalkene were obtained. A
similar pattern in results was obtained with 4a and 4b
(Table 2). Whereas the former gave quantitative yields of
methyl iodide and thioanisole, the latter gave mostly
cyclohexene. These results with 3h and 4b are typical of
the SN2/E2C behavior reported years ago by Winstein,
Eliel and co-workers,10 That is, that in reactions of
cyclohexyl halides and tosylates with halide and
thiophenoxide ions, both SN2 substitution and E2C
RESULTS
5-(Alkyl)thianthrenium salts (3)
Little is to be found on the preparation and chemistry of
these compounds in the literature. We reported from this
laboratory some years ago the preparation of a number of
‡
5-(alkyl)thianthrenium perchlorates, Th RClO4 , by
reaction of thianthrene cation radical perchlorate,
‡
Á
Th ClO4 , with dialkymercurials (R2Hg) and tetraalk-
Copyright 2000 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 81–89

REACTIONS OF SULFONIUM SALTS WITH NUCLEOPHILES
83
Table 1. Reactions of 5-(alkyl)thianthrenium salts (3)a with
halide and thiophenoxide ions (X ) in MeCN
Table 2. Reactions of alkyldiphenylsulfonium TFS (4)a salts
with iodide and thiophenoxide ions (X ) in MeCN
Products (%)b
Products (%)b
Compound Alkyl group
X
RX
Th Alkene
Compound Alkyl group
X
RX
Ph2S Alkene
3a
3a
3a
3b
3b
3c
3c
3d
3e
3e
3f
3f
3f
3g
3g
3h
3h
3i
Me
Me
Me
Et
Et
iPr
iPr
I
92
101
102
98
99
101
4a
4a
4b
4b
Me
Me
Cc
I
96
98
15
26e
98
PhS
PhSc
I
PhS
I
PhS
PhS
I
PhS
I
PhS
101
99
101
98
95
77d
71d
Cc6
98e
6
97
a
Trifluoromethanesulfonate.
Measured with GC.
Cyclohexyl.
100
103
94
99
b
c
d
102
100
97
99
96
99
99
101
101
103
99
100
99
2-Bu
neoPent
neoPent
PhCH2
PhCH2
PhCH2
Cf5
4.5d
Cyclohexene.
94
98
100
98
100
96
96
20
26
e Cyclohexyl phenyl sulfide and diphenyl sulfide appeared together as a
single GC peak which was assayed as Ph2S. The composition of the
mixture was calculated on the basis that equal (mmol) amounts of ene
and Ph2S and of RX and Ph2S has been formed.
PhS
Bre
PhS
PhSc
Ig
2.0
3.4
79
73
7.0
11
Cf5
PhSh
I
Ci6
a historical sense to emphasize the analogy of our weak-
base initiated eliminations from cycloalkyl groups with
those that were uncovered in earlier years.10,11 Valence
bond theory12d and temperature-dependent kinetic iso-
tope effects12g–i have been used in later years to support
an E2C transition state, but the concept itself remains
controversial.12c,e,f Small amounts of iodine were in-
dicated in some experiments by the color of the reaction
solution and were measured in two cases, with 3e (entry
9, Table 1, 7.3%) and 3i (entry 18, 8.6%). In all reactions
with NaSPh varying amounts of DPDS were also found.
In reactions of the benzyl derivatives 5a–c (Table 3)
with bromide and thiophenoxide ions, the benzyl group
was displaced either exclusively (5b and c) or dominantly
(5a). That is, in reaction of 5a with PhS , some (6.4%)
thioanisole was formed in addition to benzyl methyl
sulfide, (94%). In reaction with Br , 5a gave not only 94–
95% of benzyl bromide and dimethyl sulfide, but also 7%
of benzyl methyl sulfide. The last product requires the
displacement of methyl from 5a as methyl bromide, but
we were unable to assay that compound with our GC
columns.
Ci6
PhS
I
PhS
Cj7
88
88
3i
Cj7
a
All salts were trifluoromethanesulfonates, except 3e, which was the
percholorate.
b Measured with GC, except I2.
c
An equal amount of PhSH was included.
d Unidentified, based on response factor for 2-butene.
e
Bu4NBr was used.
f Cyclopentyl.
g An average of three experiments.
h An average of two experiments.
i
Cyclohexyl.
Cycloheptyl.
j
elimination occurred, often in almost equal amounts.10,11
The concordance of the SN2 and elimination reactions, in
fact, led Winstein and co-workers to refer to the latter as
the ‘merged’ mechanism, that is, an elimination merged
with the pathway to substitution. This, then, appears to be
the situation with reactions of 3h and 4b and to a lesser
extent with 3g and i. We use the terminology E2C here in
Table 3. Reactions of benzyl(R1)(R2)sulfonium perchlorates (5) with bromidea and thiophenoxide ions (X ) in MeCN
Products (%)b
Compound
R1
R2
X
PhCH2X
PhCH2SR1
PhCH2SR2
R1SR2
R1X
R2X
c
c
5a
5a
5b
5b
5b
5c
5c
Me
Me
Me
Me
Me
Me
Me
Me
Br
PhS
Br
PhS
PhSd
Br
95
94
102
99
103
97
98
3.6
4.0
101
99
3.6
4.0
94
95
Me
3.2
3.2
PhCH2
PhCH2
PhCH2
Ph
100
94
100
Ph
PhS
a
The source of Br was KBr/18C6 except in the first entry, when Bu4NBr was used.
b Measured with GC.
c
MeBr was not measurable with our columns.
d 0.50 mmol each of NaSPh and PhSH.
Copyright 2000 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 81–89

84
B. LIU AND H. J. SHINE
The data in Tables 1 and 2 show that reactions of 3 and
4 with I and PhS do not have the variability of
reactions of 1 and 2. Regardless of the structure of R in 3
and 4 reactions followed either SN2 or E2C pathways.
The latter pathway was particularly marked with
cyclohexylsulfonium salts 3h and 4b, well in accordance
with earlier classical studies of cyclohexyl halides and
tosylate.10,11 Furthermore, our comparative results with
cyclopentyl- (3g) and cyclohexylthianthrenium (3h) salts
almost parallel those obtained in Winstein’s laboratory
with cyclopentyl and cyclohexyl tosylates.10b That is,
reactions of these tosylates (with Cl in acetone at 75°C)
gave 3.3 and 72.3% elimination, respectively, close to
our results with I and PhS (Table 1). The difference in
reactivities of the tosylates was attributed to the faster
SN2 reaction of cyclopentyl tosylate.10b Because reac-
tions of simple cyclopentyl and cycloheptyl derivatives
with thiolate ions could not be found in the literature, we
carried out those reactions ourselves by stirring the
cycloalkyl derivative overnight in MeCN solution
containing an excess of NaSPh. Cyclopentyl bromide
and tosylate each gave 97% of cyclopentyl phenyl sulfide
and small amounts of cyclopentene. Cycloheptyl bromide
gave 91% of cycloheptyl phenyl sulfide and 4.7% of
cycloheptene. Cycloheptyl tosylate gave 91 and 6% of
these products. Previously, McLennan11 reported the
reaction of cyclohexyl bromide with NaSPh, obtaining
55% of cyclohexene. Under our conditions, cyclohexyl
bromide gave 56% of cyclohexene and 44% of
cyclohexyl phenyl sulfide.
Our diagnosis of SN2/E2C reactions with 3 and 4 is
supported by the results of reactions of 5a–c, Table 3.
There it is seen that the benzyl rather than methyl group is
displaced either exclusively (5b, c) or dominantly (5a).
The preferred displacement of benzyl over methyl is
diagnostic of SN2 reactions and is in accordance with
observations in a variety of earlier studies.13 With respect
to sulfonium salts, reaction of azide ion with dimethyl(1-
phenylethyl)sulfonium chloride occurred exclusively at
the l-phenylethyl group14,15 and with inversion of
configuration.14 Thiourea reacted similarly with this
sulfonium ion, but methoxide reacted preferentially at
the methyl groups.15 Dorman and Love used preferential
attack at benzylic carbon in an adaptation of the
Merrifield peptide process. That is, a dialkyl sulfide
(e.g. Me2S, Et2S) leaving group was attached to the
benzylic CH2 group of the polymer, forming a polymer-
linked dialkylbenzylsulfonium ion. Reaction with car-
boxylate occurred mainly (90%) at the benzylic carbon
atom.16 King et al. found that dibenzylethylsulfonium
salts reacted with thiocyanate ion and thiourea exclu-
sively at a benzyl group, with no sign of ethyl-group
displacement.17
Scheme 1
DISCUSSION
In the reactions of alkoxysulfonium salts 1 and 2, akin to
3 and 4, reaction at sulfonium sulfur played a significant
role (halide ion) or dominant role (thiophenoxide ion).1,2
For example, reactions of iodide ion with 1 led to the
formation of RI and thianthrene 5-oxide [6, Eqn. (1)], to
alkene and 6 [Eqn. (2)] and to thianthrene (Th, 7), ROH
and iodine [Nu2, Eqn. (3)] in relative amounts depending
on the nature of R in the alkoxy group. These reactions
are shown in abbreviation in Scheme 1, with Nu = I ,
When the nucleophile was PhS , however, only 7, ROH
and diphenyl disulfide (Nu2 = PhSSPh, DPDS) were
formed, regardless of the nature of R. Similar behavior
was noted with 2. The reactions shown in Eqns (1) and (2)
(Nu = I ) were diagnosed as being SN2 and E2C
reactions, occurring in competition that became more
pronounced as R in RO was changed from primary to
secondary to cycloalkyl. As that change progressed,
furthermore, formation of 7, ROH and I2 became more
significant, and that was attributed to reaction of I at
sulfonium sulfur, increasing in scale as the nature of R
made SN2/E2C reactions more difficult. In contrast,
reactions of PhS with 1 and 2 were almost exclusively
according to Eqn. (3). The difference in behavior of PhS
and I , in spite of the well-known greater carbon
nucleophilicity of PhS , was attributed to the thiophili-
city of PhS , which not only directed its initial reaction to
sulfonium sulfur, but also its even more facile follow-up
reaction at the sulfenyl sulfur of the resulting phenylthio-
sulfonium ion (8), Scheme 2. The formation of 8 was
further attributed not only to the thiophilicity of PhS ,
but also to the easy displacement of alkoxide. Whether, in
the formation of 8, an intermediate sulfurane, should be
included, or whether direct displacement of alkoxide
occurred, could not be decided.
Our conclusion about the results in Tables 1–3, then, is
that they are from SN2 and E2C reactions and do not
involve reaction at sulfonium sulfur. The reaction at
sulfonium sulfur that we have reported for 5-alkox-
Scheme 2
Copyright 2000 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 81–89

REACTIONS OF SULFONIUM SALTS WITH NUCLEOPHILES
85
ythianthrenium ions (1), particularly with PhS , does not
occur with the sulfonium salts 3–5. The difference
between reaction of 1, 2 and 3–5 is in the displacement of
the group attached to sulfonium sulfur. The alkoxy group
is easily displaced from 1 and 2 when a nucleophile
attacks sulfonium sulfur. Analogous reaction with 3–5
would require displacement of a carbanion and that does
not occur. The iodine measured in two reactions (3e, i), is
suggestive of reaction at sulfonium sulfur, but the
corresponding alkane was not found; the formation of
iodine may have been caused by inadvertent oxidation of
iodide ion. The DPDS that was found in reactions of
NaSPh is also attributed to concomitant oxidation of
PhS .
It is, of course, not possible to say that PhS , that we
have deemed to be so thiophilic, does not react at
sulfonium sulfur. Its thiophilicity suggests that this
reaction should occur with formation of a sulfurane [9,
Eqn. (4)], but if that does occur it must be in an unfruitful,
reversible way.
the methyl group of remaining 3a was slowly transferred
to solvent with the formation (deduced from the NMR) of
‡
d6-Me2SO Me. We saw no change in the NMR spectrum
of 3e to suggest the presence of a sulfurane 9. We
conclude that the result of the competitive reaction is also
diagnostic of an SN2 pathway.
More direct evidence for an SN2 pathway was sought
with the preparation of a 5-(alkyl)thianthrenium salt in
which the alkyl group would undergo inversion of
configuration in an SN2 reaction and retention of
configuration in ligand coupling. We chose first to
alkylate 7 with cis- and trans-4-methylcyclohexyl
formates. However, whereas we had no difficulty in
alkylating 7 with cyclohexyl formate, we were repeatedly
unsuccessful with the 4-methyl derivatives. Last, we
turned to alkylation with (S)-2-butyl formate,
[a]D = ‡7.9°,
made
from
(S)-(‡)-2-butanol,
[a]D = ‡13°. Alkylation of 7 to give the expected 5-
[(R)-2-butyl] thianthrenium triflate gave a product with a
very low rotation, namely [a]D = ‡0.9°, suggesting that
the 2-butyl group was mostly racemic. Reaction of the
product with PhS gave 2-BuSPh also with a very low
rotation, namely, [a]D = ‡0.8°. The positive rotation is
consistent with the formation of (S)-2-BuSPh
([a]D = ‡16°) and thus consistent with an SN2 pathway,
but the results are not considered to be strong enough for
a reliable diagnosis.
Is it possible that a sulfurane is formed, and through
ligand coupling18 leads to what we have diagnosed as an
SN2 product? We have no direct evidence for the
formation of 9 and indirect evidence that leads us to
argue on two counts against its leading to RSPh. First,
there is no reason to expect that ligand coupling in the
reactions of 5 would occur exclusively or preferentially
with the benzyl group. If ligand coupling was made
feasible by the formation of 9, both methyl and benzyl
groups would be able to participate in it. In the case of 5c,
even the phenyl group could be displaceable, forming
diphenyl sulfide. Second, with ligand coupling, the
cyclohexyl group in 3h and 4b would appear only as
cyclohexyl phenyl sulfide. There is no reason to believe
that the E2C reaction we observe could occur within 9.
The parallel of our reactions with 3g and h and cycloalkyl
tosylates10b and the similarities in our results with 3g–i
and the cycloalkyl bromides and tosylates are also
indicative of SN2/E2C rather than another route.
Nevertheless, we made several attempts to verify an
SN2 route to RSPh in the reactions of PhS with 3. In the
first, we set out to measure and compare the rates of
reaction of 3a (methyl) and 3e (neopentyl) with PhS by
NMR spectroscopy. We used DMSO-d6 rather than
MeCN-d3 as solvent because of the poor solubility of
NaSPh in the latter. We were able to measure the rate of
reaction of 3a but reaction of 3e was too slow to measure.
Consequently, we resorted to a qualitative, competitive
reaction between 3a and e for a deficiency of PhS . In
that case, all of 3a reacted rapidly forming 7 and MeSPh,
while the amount of 3e remained unchanged. Thereafter,
In conclusion, although we have not been able to
provide direct evidence against ligand coupling, all the
evidence we have is supportive of SN2 (and E2C)
reactions of 3–5.
EXPERIMENTAL
The following compounds used as GC controls and/or for
other preparations were obtained from commercial
sources: methyl, ethyl, isopropyl, neopentyl and cyclo-
hexyl iodide; cyclopentyl, cycloheptyl and benzyl
bromide; methyl and ethyl phenyl sulfide; and methyl
and ethyl formate. The preparations of cyclohexyl phenyl
sulfide and cyclopentyl and cycloheptyl iodide were
described earlier.2 Acetonitrile was dried by distillation
from P2O5 and again from CaCl2; methylene chloride
was dried by distillation from P2O5, and DMSO by
boiling over CaH2 followed by distillation under reduced
pressure. Column A used for all compounds except MeI,
Me2S and cyclopentene was 10% OV-101 on 80-100-
mesh Chromosorb-W HP, 4 ft  1/8 in i.d. stainless steel
(SS). The column was held at 50°C for 2 min and ramped
to 250°C at 12 ° min 1. The three compounds listed were
assayed on a column (B) of 20% BEEA on 60–80-mesh
Chromosorb P AW, 8 ft  1/8 in i.d. SS held at 50°C for
2 min and ramped to 100°C at 6°C min 1. In each case
the GC injector was set at 250°C and the detector at
300°C.
Copyright 2000 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 81–89

86
B. LIU AND H. J. SHINE
Preparation of 5-(alkyl)thianthrenium tri¯ates. An
example is given with 3a. To a stirred mixture of
1.08 g of Th (5.0 mmol) and 600 mg (10.0 mmol) of
methyl formate, cooled in an ice-bath, was added 2.5 ml
of trifluoromethanesulfonic acid (triflic acid). The
mixture was removed from the ice-bath and stirred for
10 h at room temperature, after which it was poured into
100 ml of water. The resulting suspension was extracted
with 3 Â 100 ml of dichloromethane. The combined
dichloromethane solution was concentrated to 10 ml
and poured into 200 ml of diethylether to give 1.16 g
2H, J = 7.62 (ave), 1.18 (ave); 7.707, td, 2H, J = 7.55
(ave), 1.36 (ave); 4.707, m (mainly q), 1H, J = 6.0 (ave);
1.936–1.66, m, overlapping solvent peaks.
5-(Cycloheptyl)thianthrenium triflate (3i), 17%, m.p.
1
70–71°C after reprecipitation from dichloromethane. H
NMR, 200 MHz (CD3CN), ꢀ: 8.807, dd, 2H, J = 7.73,
1.38 (ave); 7.951, dd, 2H, J = 7.83, 1.46; 7.832, td, 2H,
J = 7.62 (ave), 1.52 (ave); 7.707, td, 2H, J = 7.54 (ave),
1.50 (ave): 4.501, m. 1H: 1.759–1.441, m. 12H.
5-(Benzyl)thianthrenium triflate (3f). Thianthrene
(864 mg, 4.0 mmol) and benzyl bromide (684 mg,
4.0 mmol) were dissolved in 20 ml of dichloromethane.
The solution was stirred while 514 mg (2.0 mmol) of
silver triflate was added. After 3 h of stirring the mixture
was filtered into 80 ml of dry diethylether. The
precipitated product was reprecipitated from MeCN with
diethylether, giving 560 mg (1.23 mmol, 61%) of 3f, m.p.
1
(3.05 mmol, 61 %) of 3a, m.p. 176–178°C. H NMR,
300 MHz (CD3CN) (J in Hz throughout) ꢀ: 8.116, dd(d),
2H, J = 7.90, 1.30 (ave), 0.267 (upfield peaks split only);
7.947, dd(d), 2H, J = 7.98, 1.07 (ave), 0.228 (upfield peak
split only); 7.806, td, 2H, J = 7.69 (ave), 1.41 (ave);
7.697, td, 2H, J = 7.69 (ave), 1.35 (ave); 3.160, s, 3H. 13C
NMR, ꢀ: 136.875, 135.501, 134.547, 131.328, 130.658,
119.587, 25.610.
Attempts to prepare 5-(methyl)thianthrenium perchlo-
rate by reaction of Th with methyl chloroformate and
perchloric acid failed.
1
87–88°C. H NMR 200 MHz (CD3CN), ꢀ: 4.935, s, 2H;
the NMR spectrum of 3f in the aromatic region was a
series of overlapping multiplets, ꢀ 7.987–7.058, from the
thianthrenium and phenyl rings. Also, 3f tended to
decompose slowly in MeCN so that the aromatic region
was further complicated by a small amount of thianthrene
and the departed benzyl group. Attempts to alkylate Th
with benzyl formate failed.
5-(Ethyl)thianthrenium triflate (3b) was prepared
similarly, using 2.16 g (10.0 mmol) of Th and 1.48 g
(20.0 mmol) of ethyl formate. The first precipitate from
pouring the dichloromethene solution into diethylether
was a yellow oil. This was dissolved in dichloromethane,
cooled in an ice–salt bath and a small amount of
diethylether was added, with scratching, giving 740 mg
5-(Neopentyl)thianthrenium perchlorate (3e). To a
‡
Á
stirred suspension of 630 mg (2.0 mmol) of Th ClO4
in 2 ml of MeCN was added dropwise a solution of
342 mg (1.0 mmol) of dineopentylmercury in 5 ml of
MeCN. The mixture was stirred until the color of Th.‡
had disappeared. The solvent was removed in a rotary
evaporator and the residue was dissolved in dichlor-
omethane. That solution was shaken with 1% aqueous
LiCl, separated, concentrated to about 5 ml and poured
into dry diethylether to precipitate 235 mg (0.61 mmol,
61%) of 3e, m.p. 125–127°C (decomp.). 1H NMR,
200 MHz (CD3CN), ꢀ:8.141 dd, 2H, J = 7.82, 1.42 (ave);
7.982, dd, 2H, J = 7.89, 1.229 (ave); 7.813, td, 2H,
J = 7.62 (ave), 1.52 (ave); 7.695, td, 2H, J = 7.60 (ave),
1.42 (ave); 3.710, s, 2H; 1.040, s, 9H.
1
(1.88 mmol, 19%) of pale yellow 3b, m.p. 65–67°C. H
NMR, 300 MHz (CD3CN), ꢀ: 8.098, dd, 2H, J = 7.95,
1.05 (ave); 7.945, dd, 2H, J = 7.95, 1.05 (ave); 7.829, td,
2H, J = 7.73 (ave), 1.40 (ave); 7.706, td, 2H, J = 7.65
(ave), 1.40 (ave); 3.674, q, 2H, J = 7.35 (ave); 1.179, t,
3H, J = 7.35 (ave). 13C NMR, ꢀ: 136.889, 135.714
135.353, 131.274, 130.593, 117.721, 37.115, 9.828.
5-(Isopropyl)thianthrenium triflate (3c), m.p. 75–
76°C, was obtained in 94% yield from reaction of Th
with isopropyl formate. 1H NMR, 200 MHz (CD3CN), ꢀ:
8.098, dd, 2H, J = 7.76, 1.50 (ave); 7.946, dd, 2H,
J = 7.93, 1.45 (ave); 7.845, td, 2H, J = 7.58 (ave), 1.43
(ave); 7.716, td, 2H, J = 7.58 (ave), 1.56 (ave); 4.469,
sept, 1H, J = 6.73 (ave); 1.273, d, 6H, J = 6.69 (ave).
5-(2-Butyl)thianthrenium triflate (3d), m.p. 61–62°C,
was obtained in 40% yield from alkylation of Th with 2-
butyl formate. 1H NMR, 300 MHz (CD3CN), ꢀ:8.090 and
8.055, overlapping dd, 2H, J = 7.88, 1.42 (ave) and 7.82,
1.32 (ave); 7.951, dd, 2H, J = 7.97, 1.29 (ave); 7.841, td,
2H, J = 7.58 (ave), 1.37 (ave); 7.712 and 7.164, over-
lapping td, 2H, J = 7.66 (ave), 1.26 (ave) and 7.73 (ave),
1.24 (ave). 200 MHz, ꢀ:4.402, sextet, 1H, J = 6.54 (ave);
1.600, quintet, 2H, J = 7.10 (ave); 1.214, d 3H, J = 6.89;
0.953, t, 3H, J = 7.36 (ave).
5-(Cyclohexyl)thianthrenium triflate (3h) was pre-
pared from 5.0 mmol of Th, 10.0 mmol of cyclohexyl
formate and 2.5 ml of triflic acid. The product was
reprecipitated from dichloromethane with ether three
times, giving 1.86 g (4.15 mmol, 83%) of 3h, m.p. 80–
1
81°C. H NMR, 200 MHz (CD3CN), ꢀ:8.057, dd, 2H,
J = 7.76, 1.33; 7.947, dd, 2H, J = 7.89, 1.29 (ave); 7.834,
td, 2H, J = 7.63 (ave), 1.37 (ave); 7.701, td, 2H, J = 7.56
(ave), 1.44 (ave); 4.310, quintet, 1H, J = 7.38 (ave);
1.817–1.753, m, 2H; 1.664–1.544, m, 5H; 1.413–1.268,
m 3H.
Diphenylmethylsulfonium triflate (4a). Reaction of
930 mg (5.0 mmol) of diphenyl sulfide, 600 mg
(10.0 mmol) of methyl formate and 2.5 ml of triflic acid
5-(Cyclopentyl)thianthrenium triflate (3g), 85% yield,
m.p. 90–92°C after reprecipitation from dichloro-
1
gave 1.56 g (4.46 mmol, 89%) of 4a, m.p. 95–97°C. H
1
methane. H NMR, 200 MHz (CD3CN), ꢀ:8.114, dd,
NMR, 200 MHz (CD3CN), ꢀ: 7.968–7.903, m, 4H;
7.695–7.580, m, 6H; 3.688, s, 3H. Lit. m.p. 94–97.5°C,
2H, J = 7.77, 1.40 (ave); 7.945, d, 2H, J = 7.81; 7.835, td,
Copyright 2000 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 81–89

REACTIONS OF SULFONIUM SALTS WITH NUCLEOPHILES
87
from alkylation of diphenyl sulfide with methyl iodide
and silver triflate.19
(CD3CN), ꢀ:7.423–7.362, m, 2H; 7.331–7.173, 2m, 3H;
3.374, heptet, 1H, J = 6.66; 1.291, d, 6H, J = 6.66.
Neopentyl phenyl sulfide, 62%, obtained as an oil from
reaction of neopentyi iodide with sodium thiophenoxide
was purified on a column of silica gel, and had a
Cyclohexyldiphenylsulfonium triflate (4b), prepared
similarly to 4a, pale yellow solid after two precipitations
from dichloromethane, m.p. 96–98°C. 1H NMR,
200 MHz (CD3CN), ꢀ:7.982–7.939, m, 4H; 7.845–
7.684, m, 6H; 4.535, br m, 1H; 1.995–1.840, br m,
overlapping solvent; 1.759–1.407, br m, 7H. Compound
4b was obtained as a semi-solid by Kang and Ku in the
reaction of cyclohexyl phenyl sulfide with diphenylio-
donium triflate.20
1
satisfactory H NMR spectrum.27
2-Butyl phenyl sulfide, 36%, obtained from reaction of
2-butyl iodide with NaSPh in MeCN had one GC peak
1
after purification on a column of silica gel. H NMR,
200 MHz (CDCl3), ꢀ: 7.419–7.358, m, 2H; 7.325–7.159,
m, 3H; 3.163, sextet, 1H, J = 6.63 (ave); 1.598, m, 2H;
1.272, d, 3H, J = 6.69; 1.007, t, 3H, J = 7.36 (ave).
(S)-2-Butyl phenyl sulfide. (R)-2-Butyl iodide (as-
sumed configuration28) was prepared by reaction of (S)-
2-butanol with NaI and Me3SiCl in MeCN following the
procedure of Olah et al.29 The reaction solution was
diluted with diethylether and the ether solution was
washed successively with water, sodium thiosulfate
solution and brine, and dried over Na2SO4. The ether
solution was concentrated to a small volume and used for
reaction with NaSPh in MeCN. Work-up and chroma-
tography on silica gel gave 2.3% of (S)-(‡)-2-butyl
phenyl sulfide, with a single GC peak and [a]D (room
temperature) in MeCN ‡15.3°. A second preparation,
16% yield, had [a]D = ‡17.3°.
Preparation of benzylsulfonium perchlorates (5). The
method of Aggarwal et al. was used.8 To a solution of
2.0 mmol of sodium perchlorate in a minimum amount of
acetone were added 2.0 mmol of a dialkyl sulfide and
2.0 mmol of alkyl halide. The solution was stirred for 2
days at room temperature. The solid that formed was
filtered off, the filtrate was evaporated under reduced
pressure and the residue was dissolved in a small amount
of dichloromethane. The solid that remained was filtered
off and diethylether was added to the filtrate to precipitate
the product.
Benzyldimethylsulfonium perchlorate (5a), from di-
methyl sulfide and benzyl bromide, 75%, m.p. 104–
1
105°C. H NMR, 200 MHz (CD3CN), ꢀ: 7.461, s, 5H;
Cyclopentyl phenyl sulfide, 71%, was obtained
similarly from cyclopentyl bromide as an oil25,30 with a
single GC peak. 1H NMR, 200 MHz (DMSO-d6), ꢀ:
7.38–7.13, m, 5H; 3.597, m, 1H; 2.09–2.00, m, 2H; 1.82–
1.56, m, 6H.
Cycloheptyl phenyl sulfide, 80%, was obtained from
cycloheptyl bromide as an oil with a single GC peak. 1H
NMR, 200 MHz (CDCl3), ꢀ: 7.392–7.186, m, 5H; 3.340,
heptet, 1H, J = 4.40 (ave); 2.079, m, 2H; 1.459–1.777, m,
10H.
4.722, s, 2H; 2.907, s, 6H.
Dibenzylmethylsulfonium perchlorate (5b), from ben-
zyl methyl sulfide and benzyl bromide, 68%, m.p. 120–
121°C. 1H NMR, 200 MHz (CD3CN), ꢀ:7.530–7.482, m,
4H; 7.394–7.314, m, 6H; 4.849, d, 2H, J = 12.6; 4.693, d,
2H, J = 12.6; 2.692, s, 3H. Lit. m.p. 122°C, and
diastereotopic benzylic CH2 1H NMR, 60 MHz
(DMSO-d6), ꢀ: 5.32, d, 2H, J = 13.0; 5.16, d, 2H,
J = 13.0.21
Benzylmethylphenylsulfonium perchlorate (5c), from
thioanisole and benzyl bromide, 4 days of stirring, 6.4%,
m.p. 132–133°C. 1H NMR, 200 MHz (CD3CN), ꢀ:
7.791–7.605, m, 5H; 7.434–7.321, m, 3H; 7.193, d, 2H,
J = 7.49; diastereoptopic CH2, 4.878, d, 1H, J = 12.8;
4.691, d, 1H, J = 12.6; 3.169, s, 3H. Lit. m.p. 115–117°C,
1H NMR, 60 MHz (acetone-d6), ꢀ: 8.00–7.57, m, 5H;
7.25, br s, 5H; 5.10, d, 2H, J = 4.4; 3.41, s, 3H.22 The
triflate salt had a 1H NMR spectrum (CD3CN) similar to
that of 5c.23
Formates. Formates were prepared as follows. Isopropyl
formate was prepared in 60% yield by reaction of
isopropyl alcohol with formic acid catalyzed by con-
centrated H2SO4, b.p. 62–65°C. 1H NMR, 200 MHz
(CDCl3), ꢀ:8.017, s, 1H; 5.139, sept of d, 1H, J = 6.26,
1.00; [these splittings are similar to these for ethyl
formate in the AIST (Agency of Industrial Science and
Technology), Japan, database. (http://www.go.jp)],
1.296, d, 6H, J = 6.29. Lit. b.p. 68.8-68.9°C.31
(S)-(‡)-2-Butyl formate.
A
solution of 7.4 g
Sul®des. Sulfides were prepared as follows. Benzyl
phenyl sulfide was prepared by stirring a mixture of
benzyl bromide (2.8 g, 16 mmol) and thiophenol (1.5 g,
14 mmol) with Cs2CO3 (4.6 g, 14 mmol) under argon for
48 h. The mixture was poured into 50 ml of 2 M NaOH.
Extraction with diethylether gave 2.5 g (12.5 mmol,
89%) of product, m.p. 37–38°C. Lit. m.p. 40–41°C.24,25
Phenyl isopropyl sulfide, 82%, from reaction of
isopropyl iodide with PhSNa in MeCN solution was
purified on a column of silica gel. The product was an
(0.10 mol) of (S)-(‡)-2-butanol, 10.5 g (0.20 mol) of
88% formic acid and 10.8 g (0.10 mol) of Me3SiCl in
20 ml of MeCN was stirred at room temperatue for 24 h
and poured into water. To this was added solid NaHCO3
until gas evolution stopped. Three extractions with
diethylether, work-up and distillation gave 4.7 g (46%)
of product, b.p. 90–92°C, [a]D in MeCN ‡7.9 °. This
product was used for the attempted preparation of (R)-3d
as described for 3d itself.
Cyclohexyl formate was prepared by reaction of
cyclohexanol (0.15 mmol) with formic acid (0.20 mmol)
1
oil26 that gave a single GC peak. H NMR, 200 MHz
Copyright 2000 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 81–89

88
B. LIU AND H. J. SHINE
and chlorotrimethylsilane (0.15 mmol) in 20 ml of
MeCN.32 The mixture was stirred overnight and poured
into water. The aqueous solution was neutralized with
solid NaHCO3 and extracted three times with diethyl-
ether. After work-up and fractional distillation, the
portion of b.p. 30–32°C (3.5 mmHg) was freed from an
phenyl sulfide. The yields of these compounds were
calculated as stated in Table 2.
Reactions of benzylsulfonium salts (5) with bromide
and thiophenoxide ions. The source of bromide ion was
mainly KBr in the presence of 18-crown-6 (KBr/18C6).
Tetrabutylammonium bromide was used with 5a but was
avoided thereafter because it gave rise to spurious GC
peaks. Naphthalene and 2-butanone were used together as
GC standards. When 5a was used the products designated
as PhCH2SR1 and PhCH2SR2 were the same
(PhCH2SMe) so that half of the total yield is entered in
each column (entries 1 and 2). The same applies to R1X
and R2X (MeSPh, entry 2). We were unable to assay
MeBr with our columns (entry 1). The experimental
procedures were the same as with 3 and 4.
1
impurity on a column of silica gel and had H NMR,
200 MHz (CDCl3), ꢀ:8.046, d, 1H, J = 0.80, [these
splittings are similar to those for ethyl formate in the
AIST (Agency of Industrial Science and Technology),
Japan, database (http://www.go.jp)]; 4.891, tt, 1H,
J = 8.51 (ave), 4.31 (ave); 1.906–1.700, m, 4H; 1.600–
1.232, m, 6H. Lit. b.p. 94.5–95.0°C (97–98 mmHg).33
Cyclopentyl formate was prepared similarly, b.p. 30–
32°C (6.5 mmHg), 1H NMR, 200 MHz (CDCl3),
ꢀ:7.871, s, 1H; 5.139, irreg. m, 1H; 1.754–1.469, irreg.
m, 8H.
Cycloheptyl formate was prepared similarly, b.p. 55–
58°C (5 mmHg), 1H NMR, 200 MHz (CDCl3), ꢀ: 8.023,
s, 1H; 5.065, hept of d, 1H, J = 4.29 (ave), 0.92 (ave).
1.974–1.870, m, 2H; 1.781–1.429, m, 10H.
Competition between 3a and 3e in reaction with
thiophenoxide ion.
A
solution of 19.6 mg
(5,17 Â 10 5 mol) of 3a and 19.9 mg (5.16 Â 10 5 mol)
of 3e was prepared in 1 g of DMSO-d6. A solution of
3.11 mg (2.36 Â 10 5 mol) of NaSPh was prepared in 1 g
of the solvent and the two solutions were mixed. The
NMR spectrum recorded after 11 min showed the singlet
of MeSPh at 2.47 ppm. Integration of that singlet and of
the singlet for the CH2 group of 3e at 3.99 ppm showed
that 88% of 3a had been converted into MeSPh. After
18 h no change in the spectrum of 3e could be found, but
the remaining 3a had transferred its methyl group to the
solvent, visible at 3.99 ppm. The singlet of the CH2 group
of neopentyl phenyl sulfide was not seen.
Reaction of sulfonium salts (3 and 4) with iodide ion.
An example is given with 3a. In a 10 ml volumetric flask
were placed 60.4 mg (0.159 mmol) of 3a, 67.5 mg
(0.407 mmol) of KI, 21.9 mg (0.171 mmol) of naphtha-
lene as GC standard and 10 ml of MeCN. The septum-
capped mixture was stirred overnight and was assayed
three times for Th (column A) and MeI (column B),
giving 0.158 Æ 0.002 mmol (99.4%) of Th and
0.146 Æ 0.020 mmol (91.8%) of MeI. Reactions of 3b–i
and 4a and b were carried out similarly. In some cases,
the reaction solution was noticeably yellow or brown at
the end of stirring and in some of these cases the solution
was titrated for iodine with aqueous sodium thiosulfate
after GC analysis, e.g. with 3e (entry 9, Table 1) and 3i
(entry 18). In the reaction of 3g, in which a large amount
of ene (cyclohexene, 79%) was formed, three assays were
carried out after overnight stirring and three more after a
second night’s stirring (results in parentheses), giving
78.6 (78.6)% of cyclohexene. 20.1 (20.7)% of cyclohexyl
iodide and 103 (103)% of Th.
Reactions of cycloalkyl bromides and tosylates with
NaSPh. As an example,
a solution of 38.6 mg
(0.161 mmol) of cyclopentyl tosylate and 50.3 mg
(0.381 mmol) of NaSPh in 10 ml of MeCN containing
naphthalene and 2-butanone as GC standards was stirred
overnight. GC assay gave 97% of cyclopentyl phenyl
sulfide and 2% of cyclopentene. Similar reactions were
carried out with other cycloalkyl derivatives.
Acknowledgement
Reaction of sulfonium salts (3 and 4) with thiophen-
oxide ion. An example is given with 3a, the procedure
being similar to that with KI. Compound 3a (58.2 mg,
0.153 mmol), NaSPh (54.3 mg, 0.411 mmol) and
naphthalene (22.2 mg, 0.173 mmol) were used. Assays
on column A gave 0.154 mmol (100.7%) of thioanisole
and 0.154 mmol (100.7%) of Th. Diphenyl disulfide
(0.0193 mmol) was also formed and is attributed, because
of the quantitative yields of products, to oxidation of
some of the unused thiophenoxide ion, either in solution
or in the GC inlet. Reactions of 3b–i and 4a and b were
carried out similarly. In all of these reactions not all of the
NaSPh was dissolved. In the reaction of 4b, column A
could not separate diphenyl sulfide (Ph2S) and cyclohexyl
We thank the Robert A. Welch Foundation for support,
Grant D-0028.
REFERENCES
1. Zhao W, Shine HJ. Can. J. Chem. 1988; 76: 695–702.
2. Liu B, Shine HJ, Zhao W. J. Phys. Org. Chem. 1999; 12: 827–836.
3. Bandlish BK, Porter WC Jr, Shine HJ. J. Phys. Chem. 1978; 82:
1168–1171.
4. Lochynski S, Boduszek B, Shine HJ. J. Org. Chem. 1991; 56: 914–
920.
5. Sugiyama K, Shine HJ. J. Org. Chem. 1983; 48: 143–146.
6. Lochynski S, Shine HJ, Soroka M, Venkatachalam TK. J. Org.
Chem. 1990; 55: 2702–2713.
Copyright 2000 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 81–89

REACTIONS OF SULFONIUM SALTS WITH NUCLEOPHILES
89
7. Miyatake K, Yamamoto K, Endo K, Tsuchida E. J. Org. Chem.
1998; 63: 7522–7524.
8. Aggarwal VK, Thompson A, Jones RVH. Tetrahedron Lett. 1994;
35: 8659–8660.
H, Yoshihara M. Russ. J. Org. Chem. 1996; 32: 265–269; (c) Oae
S, Kawai T, Furukawa N, Iwasaki F. J. Chem. Soc., Perkin Trans. 2
1987; 405–414; (d) Oae S. Phosphorus Sulfur 1986; 27: 13–29; (e)
Oae S. Croat. Chem. Acta 1986; 59: 129–151; (f) LaRochelle RW,
Trost BM. J. Am. Chem. Soc. 1971; 93: 6077–6086; (g) Trost BM,
LaRochelle R, Atkins RC. J. Am. Chem. Soc. 1967; 91: 2175–
2177.
9. Saeva FD. Tetrahedron 1986; 42: 6123–6129.
10. (a) Winstein S, Darwish D, Holness NJ. J. Am. Chem. Soc. 1956;
78: 2915; (b) Beltrame P, Biale G, Lloyd DJ, Parker AJ, Ruane M,
Winstein S. J. Am. Chem. Soc. 1972; 94: 2240–2255; (c) Eliel EL,
Ro RS. J. Am. Chem. Soc. 1957; 79: 5995–6000; (d) Eliel EL,
Haber RG. J. Am. Chem. Soc. 1959; 81: 1249–1254.
11. McLennan DJ. J. Chem. Soc. 1966; 705–708.
12. (a) McLennan DJ. Tetrahedron 1975; 31: 2999–3010; (b) Ford
WT. Acc. Chem. Res. 1973; 6: 410–415; (c) Bordwell FG,
Mrozack SR. J. Org. Chem. 1982; 47: 4813–4815; (d) Pross A,
Shaik SS. J. Am. Chem. Soc. 1982; 104: 187–195; (e) Muir DM,
Parker AJ. Aust. J. Chem. 1983; 36: 1667–1673; (f) McLennan DJ,
Lim G-C. Aust. J. Chem. 1983; 36: 1821–1830; (g) Kwart H, Wilk
K, Chatellier D. J. Org. Chem. 1983; 48: 756–758; (h) Kwart H,
Gaffney A. J. Org. Chem. 1983; 48: 4502–4508; (i) Kwart H,
Gaffney A, Wilk KA. J. Org. Chem. 1983; 48: 4509–4513.
13. Streitwieser A. Chem. Rev. 1956; 56: 571–752.
19. Kevill DN, Anderson SW. J. Org. Chem. 1986; 51: 5029–5032.
20. Kang J, Ku BC. Bull. Korean Chem. Soc. 1985; 6: 375–376.
21. Kondo K, Mislow K. Tetrahedron Lett. 1967; 1325–1328.
22. Umemura K, Matsuyama H, Watanabe N, Kobayashi M,
Kamagata N. J. Org. Chem. 1989; 54: 2374–2383.
23. Kevill DN, Ismail NHJ. J. Chem. Soc., Perkin Trans. 2 1998;
1865–1868.
24. Shriner RL, Struck HC, Jorison WJ. J. Am. Chem. Soc. 1930; 52:
2060–2066.
25. Reid EE. Organic Chemistry of Bivalent Sulfur, vol. II. Chemical
Publishing Co.: New York, 1960.
26. Fehnel EA, Carmack M. J. Am. Chem. Soc. 1949; 71: 84–93.
27. Landini D, Rolla F. Org. Synth. 1978; 58: 143–147.
28. Jung ME, Ornstein PL. Tetrahedron Lett. 1977; 2659–2662.
29. Olah GA, Narang SC, Gupta BGB, Malhotra R. J. Org. Chem.
1979; 44: 1247–1251.
14. Harvey SH, Hoye PAT, Hughes ED, Ingold CK. J. Chem. Soc.
1960; 800–806.
15. Hoffmann HMR. J. Chem. Soc. 1965; 823.
30. Danen WC, Newkirk DD. J. Am. Chem. Soc. 1976; 98: 516–520.
31. Anderson RB, Rowley HH. J. Phys. Chem. 1943; 47: 454–463.
32. Brook MA, Chan TH. Synthesis 1983; 201–203.
33. Knight HB, Koos RE, Swern D. J. Am. Chem. Soc. 1953; 75: 6212–
6215.
16. Dorman LC, Love J. J. Org. Chem. 1969; 34: 158–165.
17. King JF, Tsang GTY, Abdel-Malik MM, Payne NC. J. Am. Chem.
Soc. 1985; 107: 3224–3232.
18. (a) Oae S. Pol. J. Chem. 1996; 70: 1491–1505; (b) Oae S, Ishihara
Copyright 2000 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2001; 14: 81–89