Reactions of nucleophiles with 5-(alkoxy)thianthrenium ions

Article abstract of DOI:10.1002/(SICI)1099-1395(199911)12:11<827::AID-POC198>3.0.CO;2-E

Reactions of 5-(alkoxy)thianthrenium perchlorates (1) with weakly basic nucleophiles Br^-, I^- and PhS^- (X^-) in MeCN and DMSO led to S_N2 substitution, E2C elimination, and reaction at sulfornium sulfur to extents depending on the structure of the alkoxy group (RO) in 1 and the nucleophile. Three types of reaction occurred with R = cyclopentyl (1a), cyclohexyl (1b), cis- (1c) and trans- 4-methylcyclohexyl (1d) and cycloheptyl (1e), and X^- = Br and I^-. That is, S_N2 reaction gave RX and thianithrene 5-oxide (ThO), E2C reaction gave cycloalkene and ThO and reaction at sulfonium sulfur gave X₂, thianthrene (Th) and cycloalkanol (ROH). Earlier work with R = Me (1f) and Et (1g) and X^- = I^-. Br^- had shown that only S_N2 reaction occurred. In contrast with reactions of halide ions, reactions of PhS^- with 1b-g occurred only at sulfonium sulfur, giving Th, ROH and PhSSPh (DPDS). For comparison with 1, reactions of Ph₂S⁺OMe (2) with I^- and PhS^- were carried out. Reaction with I^- gave only Ph₂S=O and Mel (S_N2), Reaction with PhS^- gave very little PhSMe (S_N2) but mainly Ph₂S, MeOH, and DPDS from reaction at sulfonium sulfur. The differences in nucleophilic pathways (PhS^- vs Br^- and I^-) in reactions with 1 and 2 are attributed to differences in thiophilicities of the nucleophiles. The thiophilicity of PhS^- dominates its reactions with 1 and 2. The direction toward products (Th, ROH and DPDS) in these reactions is compounded by the ease of displacement of alkoxide from 1 and 2 by PhS^-, and the ease with which, subsequently, thiophilic PhS^- attacks sulfenyl sulfur in the resulting phenylthiosulfonium ion. Copyright

Full text of DOI:10.1002/(SICI)1099-1395(199911)12:11<827::AID-POC198>3.0.CO;2-E

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY
J. Phys. Org. Chem. 12, 827–836 (1999)
Reactions of nucleophiles with 5-(alkoxy)thianthrenium
ions
Bo Liu, Henry J. Shine* and Wenyi Zhao
Department of Chemistry and Biochemistry, Texas Tech University, Lubbock, Texas 79409, USA
Received 3 May 1999; revised 7 July 1999; accepted 9 July 1999
ABSTRACT: Reactions of 5-(alkoxy)thianthrenium perchlorates (1) with weakly basic nucleophiles Br , I and
PhS (X ) in MeCN and DMSO led to SN2 substitution, E2C elimination, and reaction at sulfonium sulfur to extents
depending on the structure of the alkoxy group (RO) in 1 and the nucleophile. Three types of reaction occurred with
R = cyclopentyl (1a), cyclohexyl (1b), cis- (1c) and trans- 4-methylcyclohexyl (1d) and cycloheptyl (1e), and
X = Br and I . That is, SN2 reaction gave RX and thianthrene 5-oxide (ThO), E2C reaction gave cycloalkene and
ThO and reaction at sulfonium sulfur gave X2, thianthrene (Th) and cycloalkanol (ROH). Earlier work with R = Me
(1f) and Et (1g) and X = I , Br had shown that only SN2 reaction occurred. In contrast with reactions of halide ions,
reactions of PhS with 1b–g occurred only at sulfonium sulfur, giving Th, ROH and PhSSPh (DPDS). For comparison
‡
with 1, reactions of Ph2S OMe (2) with I and PhS were carried out. Reaction with I gave only Ph2S
=O and MeI
(SN2). Reaction with PhS gave very little PhSMe (SN2) but mainly Ph2S, MeOH, and DPDS from reaction at
sulfonium sulfur. The differences in nucleophilic pathways (PhS vs Br and I ) in reactions with 1 and 2 are
attributed to differences in thiophilicities of the nucleophiles. The thiophilicity of PhS dominates its reactions with 1
and 2. The direction toward products (Th, ROH and DPDS) in these reactions is compounded by the ease of
displacement of alkoxide from 1 and 2 by PhS , and the ease with which, subsequently, thiophilic PhS attacks
sulfenyl sulfur in the resulting phenylthiosulfonium ion. Copyright 1999 John Wiley & Sons, Ltd.
KEYWORDS: alkoxysulfonium ions; nucleophilic reactions; thiophilic reactions
INTRODUCTION
the rates of reaction (with primary and isopropyl R)
indicated that these reactions were SN2 displacements.
Small amounts of two side reactions were also observed.
One [Eqn. (2)] was an elimination that gave alkene and
ThO, and the other [Eqn. (3)] was reaction at sulfonium
sulfur that gave thianthrene (Th), halogen and the alcohol
(ROH).7
We have continued with studies of reactions of
nucleophiles with 1 and report here the reactions of 1a–
e with Br and I (Scheme 2) in which R is cyclopentyl
(1a), cyclohexyl (1b), cis- (1c) and trans-4-methylcy-
clohexyl (1d) and cycloheptyl (1e). As shown in Scheme
2, these reactions encompassed substitution [Eqn. (1)],
elimination [Eqn. (2)] and halogen formation [Eqn. (3)].
We also studied reactions of 1b–e with thiophenoxide ion
(PhS ). In contrast with halide ion reactions, those with
PhS occurred only at sulfonium sulfur, with the
formation of Th, cycloakanol (ROH) and diphenyl
disulfide (X2 = DPDS). For comparison with reactions
of 1 having cycloalkyl R, we studied also reactions of 5-
(methoxy)- (1f) and 5-(ethoxy)thianthrenium perchlorate
(1g) with PhS . For further comparison with reactions of
1, we studied reactions of methoxydiphenylsulfonium
tetrafluoroborate (2) with I and PhS .
Alkoxysulfonium salts have been known for about 60
years, since the first one was prepared by Meerwein.1–3
Meerwein’s method, the alkylation of a sulfoxide, is in
fact the one frequently used for preparing methoxy- and
ethoxysulfonium salts, utilizing commonly available
trimethyl- and triethyloxonium salts as alkylating
agents.4 Recently, we found that thianthrene cation
radical perchlorate (Th ClO4 ) reacts easily with a
variety of primary and secondary alcohols (Scheme 1),
permitting the isolation of crystalline 5-(alkoxy)thian-
threnium perchlorates (1).5–7 The availability of these
salts allowed us to begin a systematic study of their
reactions with nucleophiles. We found, for example, with
the use of 18O labeling, that water reacted with 5-
(cyclohexyloxy)thianthrenium perchlorate at sulfonium
sulfur, producing thianthrene 5-oxide (ThO) and cyclo-
hexanol.5 Reactions of 1 containing acyclic primary and
secondary alkyl groups (R) with bromide and iodide ions
were different, however. The major products were alkyl
halide and ThO [Scheme 2, Eqn. (1)]. The products and
‡
Á
*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.
Studies of reactions of alkoxysulfonium ions with
nucleophiles are by no means new. The largest systematic
Copyright 1999 John Wiley & Sons, Ltd.
CCC 0894–3230/99/110827–10 $17.50

828
B. LIU, H. J. SHINE AND W. ZHAO
Scheme 1
studies have been in reactions of hydroxide ion, alkoxide
ions and alcohols with methoxy- and ethoxysulfonium
ions (since those studies were used in characterizing the
exchange of alkoxy ligands at the sulfur atom), the
racemization of sulfoxides and the oxidation of alcohols
to aldehydes and ketones.2,4,8–12 As far as we are aware,
systematic or deliberate studies of reactions of alkoxy-
sulfonium ions with halide ions have not been reported.
Reactions of chloride and bromide ion with particular
alkoxysulfonium ions, coincidental, for example, with,
and arising from studies of, oxidations of alcohols, have
been reported,13–16 and we shall refer further to these
reactions. We shall refer further, also, to the reduction of
sulfoxides by HI17 and the role of halide ions in the
racemization of sulfoxides,18 and in the exchange of
ethoxide groups in an ethoxysulfonium ion.19 Reports of
reactions of alkoxysulfonium ions with thiolate ions are
also not numerous. Oae and Kim20 have reported the
reaction of ethoxydiphenylsulfonium tetrafluoroborate
with p-tolyl thiolate and Kobayashi and co-workers21,22
have reported reactions of arylmethylmethoxysulfonium
salts with PhS . We shall refer to these reactions also.
We have been unable to find any other reports of
reactions of thiolate ions with alkoxysulfonium salts.
RESULTS AND DISCUSSION
Reactions with I and Br
Reactions were carried out in dry acetonitrile (MeCN).
The source of I was KI. Two sources of Br were used.
Because KBr is not sufficiently soluble in MeCN, it was
used, as one of the sources, in the presence of 18-crown-6
(18C6). A second source was tetrabutylammonium
bromide (TBAB). Reaction products, except I2 and Br2,
were assayed with quantitative gas chromatographic
(GC) analysis. Because salts such as 1a–g decompose
when injected into a hot gas chromatograph inlet,5 it was
necessary to ensure that reaction was complete before the
final assay was made. To do that, assays of the reaction
mixture solution were made at timed intervals until, when
reaction appeared to be complete, completion was
verified by injecting a small amount of aqueous K2CO3
into the reaction solution. Salts such as 1a–g are
converted into ThO and ROH by aqueous K2CO3.
Therefore, completion of reaction with halide ion was
indicated by invariant assay before and after injection of
aqueous K2CO3. In Tables 1 (for I ) and 2 (for Br ) only
the final assay of the dry reaction solution is reported.
Iodine was assayed by titration of an aliquot of solution
after addition of aqueous K2CO3. Assay of Br2 could not
be made directly, and therefore it was assayed after
conversion into I2 by addition of KI.
Reactions with I . Six products are listed in Table 1:
cycloalkene (ene), cycloalkyl iodide (RI), ThO, Th,
cycloalkanol (ROH) and I2. The amount of ene was small
(6.6%) and the amount of RI large (89%) from 1e. In
contrast, the amount of ene was large (about 50%) and RI
small (2–6%) from 1b and c. From 1a, 21% of ene and
55% of RI were obtained, whereas from 1d, the amounts
were approximately 35 and 38%. It is apparent that 1e
Scheme 2
Copyright 1999 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 12, 827–836 (1999)

REACTIONS OF NUCLEOPHILES WITH 5-(ALKOXY)THIANTHRENIUM IONS
Table 1. Products of reactions of 1a±e with KI
829
Product (%)
Th
a
Compound
ene
RI
55
6.0
2.5e
38
89
ThO
ROH
c
I2
Ratiob
1a
21
53
52
35
78
55
52
76
95
19
44
48
23
5.4
17
43
47
23
0.97
1.07
1.05
0.96
1.01
1b
1cd
1df
1e
43
46
21
tr
6.6
5.3
a
Yield (%) based on starting amount of 1.
b Sum of (ene ‡ RI)/ThO.
c
Peak overlapped with solvent.
d Average of two experiments. Errors in tabulated averages were 2.0 (ene), 0.6 (RI) and 1.0 (all other products.
e
Assay based on response factor for cis-(4-methyl)iodocyclohexane because pure trans isomer could not be prepared.
f Average of two experiments. Errors in tabulated averages were 1.0 (ene and I2), 2.0 (RI and Th), 3.0 (ThO) and 4.0 (ROH).
behaved much like a primary salt7 in undergoing mainly
SN2 displacement, whereas 1a and d to some extent, and
1b and c to major extents, succumbed to elimination.
The duality of pathways, substitution and elimination,
reported in Table 1 (and Table 2), and occuring to a much
lesser extent in our earlier work, parallels competing
substitution and elimination that was uncovered in
reactions of alkyl halides and tosylates with weakly
basic nucleophiles about 40 years ago. The elimination
reaction is a lesser known but well recognized part of the
spectrum of E2 eliminations. When first reported by
Winstein et al.23 with cyclohexyl substrates, it was
called, for want of a better understanding, the ‘merged’
mechanism, connoting an elimination that occurred in a
pathway destined for SN2 substitution. Thus, this
elimination accompanied SN2 reactions of weakly basic,
good nucleophiles, and gave rise to the classification,
albeit somewhat controversially, E2C elimination.24–26
This is the elimination that iodide ion causes in reactions
with 1a–e, much more pronounced than was obtained in
our earlier work with primary and acyclic secondary
alkyloxy groups.7 The data in Table 1 show that E2C
reaction becomes more competitive with SN2 on going
from cycloheptyl to cyclopentyl and, lastly, cyclohexyl
alkyl groups. Reactions with the cyclohexyl compounds
1b–d are particularly striking in that they parallel results
reported years ago for cyclohexyl halides and tosylates.
That is, Winstein et al.23 noted that cis-4-tert-butylcy-
clohexyl tosylate gave more ene than the trans isomer in
reactions with I , Br and PhS . Eliel and Haber27 found
in reactions with PhS that the rate of elimination from
cis-4-tert-butylcyclohexyl bromide was 66 times faster
than from the trans isomer. In those cis substrates, the
leaving group is axially oriented. That is also the
orientation of the thianthreniumoxy group in the ring of
1c,6 and Table 1 shows that more ene (52%) is obtained
from the 1c than from its trans isomer, 1d (35%). Thus,
1c and d have the characteristics of E2C reactions
observed earlier with much simpler substrates. It is
noticeable that the amounts of ene and RI formed also
from the cyclohexyloxy salt (1b) are similar to those
formed from 1c rather than 1d, which has an equatorial
thianthreniumoxy group. At first sight, this was puzzling,
since the dominant orientation of the thianthreniumoxy
group in cyclohexyl ring of 1b is equatorial.6 However,
1b, but not 1c and d, undergoes ring inversion in solution,
and at room temperature approximately 30% has the
axially oriented thianthreniumoxy group.6 We deduce,
therefore, that the axial conformer reacts in the E2C
mode more rapidly than the equatorial conformer, as
found by earlier workers with simpler systems, thus
leading 1b into a pattern of reaction comparable to that
for 1c. It is apparent that, in SN2/E2C terminology, ThO
is a good leaving group, and that the cyclopentyl and
cycloheptyl rings of 1a and e allow more easily for SN2
displacement than the cyclohexyl systems 1b–d. We have
not been able to find studies of or references to E2C-type
reactions with cyclopentyl and cycloheptyl derivatives
comparable to those of cyclohexyl halides and tosylates.
Reactions that give RI and ene must also give ThO
[Eqns (1) and (2)]. Therefore, the sum of the yields of RI
and ene should equal the yield of ThO. That this is
reasonably correct in our results is shown by the ratio
(RI ‡ ene)/ThO in Table 1.
Earlier, we noted that although we have not been able
to find systematic studies of reactions of halide ions with
alkoxysulfonium salts in the literature, reactions of
chloride and bromide ions with particular alkoxysulfo-
niumionshadbeenreported byCoreyandco-workers.13–16
The reactions originated in studies of oxidation of
primary and secondary alcohols through the agency of
an alkoxydimethylsulfonium ion (5) formed in situ from
a chlorosulfonium chloride (3) and/or succinimidosulfo-
nium chloride and bromide (4, Scheme 3). Corey and co-
workers noted that in some cases oxidation was
accompanied by or superseded by halogenation, with
the formation of R1R2CHX and Me2SO. For example,
when the succinimido chloride 4 was used, benzhydrol
and 2-cyclohexenol gave the corresponding alkyl chlor-
ides rather than the ketones.13 Reaction at 25°C with
Copyright 1999 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 12, 827–836 (1999)

830
B. LIU, H. J. SHINE AND W. ZHAO
Scheme 3
benzhydrol, 2-cyclohexenol and benzyl alcohol in the
absence of triethylamine gave, in fact, >95% yields of
alkyl chlorides.14 Analogously, geraniol was converted
into geranyl bromide with the use of N-bromosuccini-
mide. Corey and co-workers attributed alkyl halide
formation to cases in which the sulfoxonium ion
intermediate (5, Scheme 3) is relatively unstable14 and
contains alcohols which correspond with stabilized
carbocations.13,15
An analogous chlorination has been reported by
Johnson and Jones29 in the formation of alkoxysulfonium
ions by the oxidation of sulfides with tert-butyl- and
isopropyl hypochlorites at 78°C. The initial tert-
butoxy- and isopropyloxysulfonium chlorides were found
to be too unstable to be isolated, but could be stabilized
into isolable hexachloroantimonates by the addition of
antimony chloride. Without this stabilization the initial
salts decomposed into sulfoxide and tert-butyl- and
isopropyl chloride; quantitative data for chloroalkane
formation were not provided. Obviously, the formation of
tert-butyl chloride could not be from an SN2 displacement
and, by analogy, isopropyl chloride was not formed in
that way either. The mechanism of chloroalkane forma-
tion in these cases remains unidentified.
It is evident that our reactions of 1 with iodide (and
bromide) ion differ from the halogenations accompany-
ing these oxidation reactions. That is, our first report of
alkyl halide formation was with examples of 1 that were
not unstable in solution and that contained primary and
acyclic secondary alkyl groups unlikely to represent
stabilized carbocations. In those reactions the rates of
formation of primary alkyl halides clearly fitted SN2
characteristics. Furthermore, 5-(neopentyloxy)thianthre-
nium perchlorate failed to give neopentyl- or any other
alkyl iodide, a finding that was consistent with SN2
characteristics of alkyl iodide formation. In our present
work, the behavior of the cyclohexyloxy salts 1b–d fit
clearly into the SN2/E2C duality of earlier reports. In only
one respect are Corey and co-workers’ results analogous
to ours, and that is that saturated alcohols, ordinarily inert
to conversion into halides in competition with oxidation,
could be converted during lengthy reaction times in the
absence of triethylamine. Thus, cyclohexylmethanol
gave 43% of the chloride after 100 h at 5°C, while
cyclohexylmethanol and 2-phenylethanol gave 70% of
the bromides after 36 h at 20°C in reaction with
derivative 4.14 These reactions would appear to be SN2
displacements analogous to ours. Corey and co-workers
commented that the use of reagents derived from other
sulfides may extend the scope of their halogenation
reaction. Our use of the isolable alkoxythianthrenium
salts fulfils that prognosis, at least insofar as primary and
acyclic secondary alkyl groups are concerned. Suited also
to halide formation is the methoxydiphenylsulfonium ion
(2), which as the tetrafluoroborate gave methyl iodide and
diphenyl sulfoxide quantitatively in reaction with iodide
ion (see Experimental section).
Table 1 lists three more products, Th, ROH and I2.
These come from a third type of reaction of compounds 1
with halide ion, namely reaction at the sulfur atom
(Scheme 4). This reaction was observed to a small extent
in our earlier work,7 but is seen here to be significant in
reactions of the cycloalkoxy derivatives. It occurs to the
greatest extent with 1b and c, in which, seemingly, SN2
displacement is the most difficult. Thus, reaction at
sulfur, as with elimination, competes with the SN2
Scheme 4
Copyright 1999 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 12, 827–836 (1999)

REACTIONS OF NUCLEOPHILES WITH 5-(ALKOXY)THIANTHRENIUM IONS
Table 2. Products of reactions of 1a, b and e with bromide ion
831
Product (%)
ThO
Compound Sourcea of Br
ene
RBr
RBr2
Th
5.1
3.5
45
8.9
5.1
ROH
Br2
Ratiob
c
d
1a
TBAB
KBr–18C
KBr–18C
TBAB
tr
89
89
4.2
89
97
—
—
94
100
55
93
97
—
—
6.2
2.0
8.2
9.6
0.95
0.91
0.93
0.97
1.00
c
d
1a
1.5
3.9
1.0
0
1be
1ef
1eg
43
45
1.3
3.7
c
—
c
c
KBr–18C
—
—
a
TBAB is tetrabutylammonium bromide; KBr–18C is potassium bromide ‡ 18 crown-6.
b Sum of (ene ‡ RBr ‡ RBr2)/ThO.
c
Not sought or measured.
d Peak overlapped with solvent.
e
Average of two experiments. Errors in tabulated averages were 2.4 (ene), 0.6 (RBr), 1.0 (RBr2 and ThO), 2.0 (Th), 3.0 (ROH) and 4.8 (Br2).
f Average of two experiments. Errors in tabulated averages were 1.0 (ene), 5.0 (RBr), 3.0 (ThO), 4.1 (Th), 1.3 (ROH) and 3.4 (Br2).
g Average of two experiments. Errors in tabulated averages were 1.0 (RBr), 2.0 (ThO), 0.3 (Th) and 1.8 (ROH).
reaction. Reaction at sulfur should give equal amounts of
Th, ROH and I2, and this is borne out by the data in Table
1.
Reactions of halide ions at the sulfur atom of
oxysulfonium ions, with concomitant formation of a
sulfurane intermediate, have been proposed earlier.
Examples include the racemization of sulfoxides by
HCl18a and the exchange of a sulfoxide’s oxygen atom
with water, catalyzed by HCl.18b Landini and co-work-
ers,17a,b interestingly, found that iodide ion in perchloric
acid solution caused the reduction of sulfoxides, whereas
bromide and chloride ions under the same condition
caused only racemization. In all of these cases, reaction
was initiated by the attack of halide ion at the sulfur atom
of the protonated sulfoxide, that attack being, in fact,
identified as the rate-determining step.17a,b Our results
differ a little from those of Landini and co-workers in that
both iodide and bromide ion caused to some extent what
is in effect the reduction of compounds 1, whereas only
iodide caused the acid-catalyzed reduction of sulf-
oxides.17a The same workers reported an order in thio-
philicities of the halide ions, Cl , Br , I as 1:3:87.
Similar ordering in the thiophilicities of halide ions
have been listed by Kice31 for catalyses of reactions of
nucleophiles at sulfenyl and sulfynyl sulfur atoms. In
analogy with such cases we might expect in our reactions
that iodide would give much more Th and halogen than
bromide ion in reactions with 1. That expectation is seen
only in reactions of 1a, however. It may be that the
expected order in thiophilicities is compromised by the
complexity of our competing reaction systems.
Reactions with Br . These reactions were with 1a, b and
e, and the source of bromide ion was either potassium
bromide in the presence of 18-crown-6 or tetrabutylam-
monium bromide. In analogy with iodide ion reactions,
three pathways were observed (Scheme 2). The dominant
pathway with the cyclopentyloxy (1a) and cyclohepty-
loxy (1e) derivatives was bromoalkane formation. This
was the minor pathway with the cyclohexyloxy deriva-
tive (1b). Instead, elimination and reaction at sulfonium
sulfur were the major fates of 1b. A complication in the
reaction of 1b was the addition of Br2, formed as in Eqn.
(3), to cyclohexene formed as in Eqn. (2) (Scheme 2),
giving trans-1,2-dibromocyclohexane, designated RBr2
in Table 2. It is possible that small amounts of
dibromocycloalkane were formed also from 1a and e
but we did not search for those products. In terms of
stoichiometry, therefore, the sum of the yields of ene,
RBr and RBr2 should equal the yield of ThO. We have
expressed that equality as a ratio in Table 2. Also, the
yields of Th and ROH should each equal the sum of the
yields of Br2 plus RBr2. This is also shown to be
approximately the case in Table 2.
The summary, then, of our reactions of 1 with I and
Br and of methoxydiphenylsulfonium ion with I is that
SN2 displacement of an alkyl group prevails until
inhibited by structural features of the alkyl group,
whereupon E2C reaction and reaction at sulfur become
significant. Reaction at sulfur is shown in Scheme 4 to
lead to a sulfurane (6). The sulfurane itself may react with
halide ion, or as proposed by others in other reactions,
17a,b may lead to a halosulfonium ion (7), from which
products are formed. The loss of an alkoxy group from an
alkoxysulfonium ion in reaction with a nucleophile has
been described as facile in even the early work of
Meerwein.30 We are unable to distinguish between the
two pathways.
Reactions with PhS
Reactions were carried out in MeCN and DMSO over a
period of 2 h. Completion of reaction was verified in
some cases by following separately the disappearance of
the sulfonium ion with NMR spectroscopy. A small
excess of sodium thiophenoxide (PhSNa) over the
amount of sulfonium salt (1 and 2) was used, correspond-
ing to the stoichiometry of Schemes 5 and 6. Thiophenol
(PhSH), approximately equivalent to the amount of
Copyright 1999 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 12, 827–836 (1999)

832
B. LIU, H. J. SHINE AND W. ZHAO
Table 4. Products of reaction of 5-methoxy- (1f) and 5-
ethoxythianthrenium perchlorate (1g) with PhSNa
Product (%)a
Run Compound Solvent ROHb
Th DPDSd ThO
1
2
3
4
1f
1f
1g
1g
DMSO
MeCN
DMSO
MeCN
98
99
99
99
95
104
105
102
103
1.7
2.5
1.0
4.2
105
96
c
Scheme 5
a Average of three assays in each run. The average errors for assays of
ROH, Th and ThO were: (run 1) 0.05, 1.7, 0.18; (run 2) 0.93, 0.30,
0.15; (run 3) 0.03, 1.3, 0.22; (run 4) –, 0.12, 0.18. In all experiments the
molar ratio of PhSNa to 1 was ca 2.5, and PhSH equivalent to the
amount of PhSNa was also used.
b
On column E.
EtOH overlapped the solvent in GC traces.
c
d After separation by column chromatography; GC on column A. Yield
is based on the stoichiometry of Scheme 5.
Scheme 6
PhSNa, was included in all runs. Unless PhSH was used,
large amounts of sulfoxide (ThO from 1, Ph2SO from 2)
were formed. (Insofar as using added PhSH is concerned,
we found that this was the practice in one of the earlier
studies,27 but not in others. Where E2C occurred, of
course, thiol was formed in situ, as pointed out by Cook et
al.32 with p-nitrothiophenoxide and by Bordwell and
Mrozack33 with p-methoxythiophenoxide. Routine use of
an excess of PhSH was noted by Arnone et al.34 in studies
of SRN1 reactions.) In the presence of PhSH the products
of reaction were substantially only alkanol (ROH), the
parent sulfide (Th from 1, Ph2S from 2) and diphenyl
disulfide (DPDS). None of the SN2 product (PhSR) was
obtained from reactions of 1 and only small amounts
from reactions of 2. PhSH itself did not react with the
alkoxysulfonium salts. The results are listed in Tables 3–
5. In these Tables, the yields of products, except DPDS,
are from GC analyses of reaction solutions. Because an
excess of PhSNa and PhSH was used, the yields of DPDS
were assayed by GC but only after PhSH and unused
PhSNa had been removed as described in the Experi-
mental section. We were able to assay without difficulty
the formation of cycloalkanols from reactions of 1b–e in
MeCN, and of MeOH and EtOH from reactions of 1f and
g and 2 in DMSO. Direct assay of MeOH and EtOH from
reactions of 1f and g and 2 in MeCN were not possible,
apparently because of the insolubility of NaOMe and
NaOEt in MeCN. These alkoxides were not converted
into the alcohols in these circumstances. Conversion
resisted the addition of a large excess of PhSH, and was
achieved only with the addition of a small amount of
water (0.2 ml) to the reaction solution (10 ml) after
reaction was complete. Assay of MeOH by GC was then
successful and is reported in Tables 4 and 5. Assay of
EtOH was not possible even with this treatment because
the GC peak overlapped that of MeCN. Some control
experiments on the fate of sodium alkoxides in MeCN are
included in the Experimental section. We conclude that
the formation of cycloalkanol from 1b–e in MeCN was
driven to completion by reaction of PhSH with soluble
Table 5. Products of reaction of methoxydiphenylsulfonium
tetra¯uoroborate (2) with PhSNa
Table 3. Products of reaction of 1b±e with PhSNa
Product (%)a
Product (%)
Run Compound ROHb
Thb
DPDSb,c
ThOd
Run Solvent MeOHa Ph2Sb DPDSc Ph2SOb PhSMeb
1
2
3
4
1b
1c
1d
1e
99
101
96
94
94
96
97
105
100
103
99
5.1
7.7
3.7
6.2
1
2
MeCN
DMSO
98
94
98
89
110
102
tr
9.7
1.6
3.1
a Column E was used. An average of three assays is given with errors
2.1% (run 1) and 1.0% (run 2). MeOH was found in run 1 only after
adding 0.2 ml of water to the suspension before assay. MeOH was
found and assayed in run 2 without addition of water.
96
a
Average of two experiments with 1b–d and three with 1e. All
experiments were in MeCN at room temperature.
b Errors in tabulated averages were 1–5.
b Column A was used. An average of three assays is given with errors
in yields (%) for Ph2S, Ph2SO, PhSMe, respectively, of 0.80,–, 0.03
(run 1) and 0.40, 0.17, –(run 2).
c After separation by column chromatography; GC on column A. Yield
is based on the stoichiometry of Scheme 5.
d Errors in tabulated averages were 3.3 (1b), 4.1 (1c), 1.6 (1d) and 1.0
(1e).
c After separation by column chromatograph; GC on column A. Yield
is based on the stoichiometry of Scheme 6.
Copyright 1999 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 12, 827–836 (1999)

REACTIONS OF NUCLEOPHILES WITH 5-(ALKOXY)THIANTHRENIUM IONS
833
quantitative amounts of Th, ROH and DPDS in reaction
with PhS . None of the displacement products, PhSMe
and PhSEt, were found. We chose to work with 2 so as to
characterize the mode of reaction of simpler alkoxysul-
fonium salts with PhS and I . Reaction with I (see
Experimental section) gave 98–99% of MeI and Ph2SO,
indicative of SN2 displacement of methyl. Only small
amounts of diphenyl sulfide (Ph2S) were formed, that
would be consistent with reaction at sulfur. Reaction with
PhS , however (Scheme 6), gave predominantly Ph2S,
MeOH and DPDS. About 9% of ThO was formed in
reaction in DMSO, suggestive of SN2 displacement, but
only 3.1% of PhSMe was found (Table 5).
Scheme 7
alkoxide. The formation of MeOH and EtOH from 1f and
g and 2 in DMSO was similarly driven or may have been
driven to completion by small amounts of water in the
incompletely dried solvent.
Our results show that in contrast with halide ions PhS
has a strong preference for reaction at sulfur in reactions
with 1 and 2. Oae and Kim20 found analogously that only
Ph2S (90%) and p-tolyl disulfide (80%) were formed
when p-tolyl thiolate reacted with ethoxydiphenylsulfo-
nium tetrafluoroborate in anhydrous ethanol. The relative
thiophilicity of PhS compared with halide ions is all the
more noteworthy when related to the carbon nucleophi-
licity of these nucleophiles. For example, the ratio of
reactivities PhS /I in methanol solution toward methyl
tosylate (MeOTs) is 321 and toward methyl iodide (MeI)
it is 313.35,36 In sulfolane solution, this ratio of
reactivities toward MeI is 37000.37 The ratio for
displacement of methyl from cobalt-bound methyl in
methyl cobalt(III)phthalocyanine in dimethylacetamide
solution is 41000.37 Intrinsically, therefore, especially in
analogy with reaction with MeOTs, one would expect
PhS to react more easily than I (and Br ) at the alkyl
group and displace ThO from 1 and Ph2SO from 2.
Reaction at both sulfonium sulfur (43%) and carbon
(12%) did ocur in reaction of PhS with methoxymethyl-
p-tolylsulfonium tetrafluoroborate (8) in dichloro-
methane (Scheme 7).21 In that case, methoxide ion,
displaced from the methoxysulfonium ion by PhS ,
reacted further with 8 to give dimethyl ether. No such
reaction was found in our work.
Reactions with 1b±e
Table 3 shows the substantially quantitative formation of
cycloalkanol, Th and DPDS. Small amounts of ThO were
obtained. In principle, ThO should be formed from either
SN2 displacement of the cycloalkyl group or from E2C
elimination, but no indications of PhSR and ene were
found in the GC traces. These reactions of 1b–e are thus
in great contrast with those of the halide ions, in which
SN2 and E2C were dominant. The reactions are also
unlike those of cyclohexyl halides and tosylates with aryl
thiolates. For example, Eliel and Haber27 found almost
equal amounts of substitution and elimination in the
reactions of PhS with bromocyclohexane and cis- and
trans-4-(tert-butyl)bromocyclohexane, the ratio of rates
for the two reaction types being 1.13, 1.04 and 1.17,
respectively. Reaction of cis-4-tert-butylcyclohexyl to-
sylate gave 40% of 4-tert-butylcyclohexene.27 The last
reaction in the hands of Winstein and co-workers23 gave
48% of elimination. McLennan28 obtained 55% of
cyclohexene from reaction of PhS with bromocyclo-
hexane. We repeated that reaction and confirmed that in
our hands, too, large amounts of cyclohexene were
formed (see Experimental section). Bordwell and Mro-
zack33 reported that, whereas 9-methylfluorene anion led
to >90% substitution, and 6-bromo-2-naphthoxide ion
led to >90% elimination, in reaction with bromocyclo-
hexane, 4-methoxythiophenoxide ion caused equal
amounts of substitution and elimination. Thus, although
our compounds 1b–d can be regarded as cyclohexanes
carrying a good leaving group (ThO), they have no
propensity to engage in the SN2 and E2C reactions that
suit a good leaving group. The nucleophilic nature of
PhS is such that, instead, given the opportunity, it
prefers to react with 1b–d at sulfonium sulfur.
The collective data show that reaction of PhS with the
sulfur atom of an alkoxysulfonium ion is very facile. In
terms of Scheme 4, the sulfurane (6) and phenylthio-
sulfonium ion (7) are formed preferably. Once 7 is
formed, furthermore, it must react rapidly with thiolate
ion irreversibly. Measurements of thiophilicity of thiolate
ions comparable to those of halide ions17a,31,38 do not
seem to have been made, although Kice and Rogers39
found that SCN reacted more rapidly than I at the
‡
sulfenyl sulfur atom of PhSS (OH)Ph. Edwards and
Pearson40 noted that, among nucleophiles, RS and PhS
rank high in their ability to break a sulfur-sulfur bond,
‡
Reactions with 1f and g and 2. Preference for reaction
at sulfonium sulfur is also seen in the reactions of PhS
with 1f,g (Table 4) and with methoxydiphenylsulfonium
tetrafluoroborate (2, Table 5). We chose 1f and g because
we had found earlier7 that they underwent almost total
SN2 reaction with I and Br . In contrast, each gave
while Kice31 deduced that ions such as R2S SPh undergo
nucleophilic substitution at sulfenyl sulfur with tremen-
dous alacrity. More recent reports have brought out the
very facile, SN2-like nature of the attack of thiolate ions
on the disulfide bond,41 from which, by comparison, the
even more facile SN2-like displacement of sulfonium
Copyright 1999 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 12, 827–836 (1999)

834
B. LIU, H. J. SHINE AND W. ZHAO
sulfur (e.g. in 7) by thioloate would be expected. The
summation of our work and the earlier reports20–22 is that
thiolate ions react selectively at the sulfonium sulfur
atom of alkoxysulfonium ions, facilitated, perhaps, by the
facile expulsion of the alkoxy group.20 Once that occurs,
it is followed by another very facile product-forming
reaction of thiolate at the sulfenyl sulfur atom of (in our
cases) a phenylthiosulfonium ion.
partly soluble in fresh MeCN. We did not proceed
further. The insolubility of PhSNa in MeCN has been
noted earlier.45 Thermogravimetric analysis showed that
the PhSNa did not contain volatile material such as water.
Preparation of 1a±g. The preparations of 1b–d, f and g
from reactions of thianthrene cation radical perchlorate
(caution: a warning about explosiveness has been
given46) have already been decribed.6,7 Compounds 1a
(66%) and 1e (69%) were prepared similarly and had
m.p. (decomp.) 63–68 and 70–71°C, respectively. Each
had the characteristic aromatic 1H NMR spectrum of a 5-
(alkoxy)thianthrenium perchlorate.6,7 The multiplet for
H-1' of 1a at ꢀ 5.12 and of 1e at ꢀ 4.84 were, as expected,
approximately 0.8 ppm downfield from the signals of the
corresponding alcohols.
EXPERIMENTAL
Solvents were dried as described earlier.6 Cycloheptanol,
iodocyclohexane, bromocyclopentane, bromocyclohex-
ane, bromocycloheptane and trans-1,2-dibromocyclo-
hexane were obtained from Aldrich Chemical and
cyclopentanol from Arapahoe Chemical. Iodocyclopen-
tane was prepared42 by reaction of cyclopentanol (1.0 g,
12 mmol) with iodotrimethylsilane (5 g, 25 mmol) in
40 ml of CH2Cl2 and was purified by column chromato-
graphy on 60–100-mesh silica gel with pentane as eluent,
giving 0.62 g (3.16 mmol, 26%) of product with a
Assays of products. Products of reaction (except
halogen) of 1a–e with halide ions were assayed by GC
on a Varian Model 3700 gas chromatograph and a
Spectra-Physics Model 4290 integrator–recorder. Three
stainless-steel packed columns were used, each of 1/8 in
i.d.: A, 10% OV-101 on 80–100-mesh Chrom-WHP, 4 ft;
B, 10% Carbowax 20 M on Chrom-WHP, 6 ft; C, 10%
OV-17 on 80–100-mesh Chrom Q11, 6 ft. Columns A
and C were held at 50°C for 2 min and ramped to 250°C
at 12°C min 1; column B was heated similarly, but to
only 100°C. Assays of products of reactions with PhSNa
were made with column A and with two other columns:
D, capillary SE-54 (Supelco No. 2-4001) and E, capillary
SPB-20 (Supelco No. 2-4086). Each was 30 m Â
0.25 mm i.d., film thickness 0.25 mm. D was heated at
1
satisfactory H NMR spectrum. Iodocycloheptane was
prepared similarly and was distilled after chromatogra-
phy failed to give a satisfactory product, giving 1.4 g
(6.25 mmol, 71%), b.p. 56–57°C (2 mmHg); lit.43 b.p.
92°C (14 mmHg). cis-(4-Methyl)iodocyclohexane was
prepared from trans-4-methylcyclohexanol (830 mg,
7.28 mmol) and iodotrimethylsilane (2.90 g, 14.5 mmol)
in 40 ml of CHCl3. Only a small amount of product was
obtained after 2 days of stirring, and stirring was
continued at room temperature for 20 days. Iodine was
removed by washing with sodium thiosulfate and water.
Evaporation of the dried solution and distillation gave
460 mg (2.05 mmol, 28%) of product, b.p. 45–46°C
(2 mmHg). The product had a satisfactory 1H NMR
spectrum corresponding to the cis structure, concordant
with the mechanism of formation.42 Attempts to prepare
trans-(4-methyl)iodocyclohexane from cis-4-methylcy-
clohexanol gave a mixture of cis- and trans-iodides.
Physical data for the isomeric iodides could not be found
in the literature. Cyclohexyl phenyl sulfide was prepared
by reaction of iodocyclohexane with PhSNa, and was
separated from DPDS by chromatography on a column of
silica gel eluted with light petroleum ether, b.p. 37–56°C.
1
50°C for 2 min and ramped at 12°C min to 250°C; E
was heated similarly and ramped to 100°C. In all cases
the injector was at 250°C and detector at 300°C.
Reactions of 1 with KI. An example is given with 1a
(entry 1, Table 1). A solution of 1a (44.0 mg,
0.110 mmol) and KI (46.9 mg, 0.394 mmol) in 5 ml of
MeCN containing both naphthalene and 2-butanone as
GC standards was stirred in a septum-capped volumetric
flask. Aliquots were taken for GC analysis after 35, 140
and 280 min of stirring, after which 0.3 ml of 2 M K2CO3
was injected. Stirring was continued for 2 h and GC
analysis was repeated. The products (mmol  102) of the
third and fourth (in parentheses) assays were cyclopen-
tene 2.30 (2.20), iodocyclopentane 6.00 (6.00), Th 2.10
(2.13) and ThO 8.60 (8.80). Cyclopentanol was not
assayed because its GC peak was too broad and shallow
and overlapped with the solvent peak. The third assays
are listed in Table 1. An aliquot was taken after the third
assay for titration of iodine. For this, 0.3 ml of K2CO3
was added to the aliquot and was followed by titration
with Na2S2O3. Column E was used for all products except
cyclopentene, for which column B was used.
1
The product44 had a satisfactory H NMR spectrum.
Methoxydiphenylsulfonium tetrafluorborate (2) was pre-
pared as described earlier.4 Sodium thiophenoxide
(PhSNa) was prepared by reaction of thiophenol (PhSH)
with dispersed sodium in boiling diethyl ether under
argon. The product (95%) was filtered and washed with
dry diethyl ether. It was completely soluble in MeOH and
DMSO but only partly soluble in MeCN. For example,
21 mg remained undissolved when 66 mg in 10 ml of
MeCN were treated with a sonic vibrator. Similarly, 36%
remained undissolved when 18-crown-6 was included in
a heated solution. The recovered solid was itself only
Reactions of 1b–e were carried out similarly with
small variations in concentrations and sampling times.
Copyright 1999 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 12, 827–836 (1999)

REACTIONS OF NUCLEOPHILES WITH 5-(ALKOXY)THIANTHRENIUM IONS
835
Reactions of 1 with TBAB. An example is given with 1a
(entry 1, Table 2). A solution of 1a (41.6 mg,
0.104 mmol) and TBAB (115 mg, 0.357 mmol) in 5 ml
of MeCN containing both naphthalene and 2-butanone as
GC standards was stirred in a septum-capped volumetric
flask. Aliquots were taken for GC analysis after 10, 50,
120, 285 and 345 min of stirring. Products (mmol  102)
after the third and fifth (in parentheses) assays were
cyclopentene trace(trace), bromocyclopentane 9.33
(9.29), Th 0.523 (0.530) and ThO, 9.92 (9.79). The fifth
assays are listed in Table 2. After 6 h of stirring, 500 mg
of KI were added to the solution and the liberated iodine
was titrated with Na2S2O3. The result is expressed as Br2
in Table 2. Column B was used for cyclopentene and
column C for all other products.
Reaction of methoxydiphenylsulfonium tetra¯uorobo-
rate (2) with sodium thiophenoxide in the presence of
thiophenol. An example is given with entry 2, Table 5. A
solution of 71.9 mg (0.237 mmol) of 2 and 60 ml (64 mg,
0.582 mmol) of PhSH in 10 ml of DMSO containing both
biphenyl and 2-butanone as GC standards was stirred in a
septum-cappped volumetric flask with 76.8 mg
(0.582 mmol) of PhSNa for 2 h. The solution was assayed
three times on column A, giving in mmol (average)
0.212 Æ 0.020 (89%) of Ph2S, 0.023 Æ 0.002 (9.7%) of
Ph2SO, 0.357 Æ 0.002 (151%, based on the amount of 2)
of DPDS, 0.00745 Æ 0.0009 (3.1%) of PhSMe and
0.222 Æ 0.0036 (94%) of MeOH. Following the GC
assays, the solution was poured into 50 ml of 2 M KOH
and worked up as described earlier to give 105 mg of a
mixture of biphenyl, Ph2S and DPDS. GC assay gave
0.200 mmol (84%) of Ph2S and 0.242 mmol (102%,
based on the amount of 2) of DPDS. This assay of DPDS
is reported in Table 5.
Reactions of 1 with KBr±18-crown-6. An example is
given with 1b (part of entry 3, Table 2). A solution of
427 mg (0.103 mmol) of 1b in 5 ml of a solution that was
0.0717 M in KBr and 0.132 M in 18-crown-6 and
containing both naphthalene and 2-butanone as GC
standards was stirred in a septum-capped volumetric
flask. Aliquots were taken for GC assay after 10, 170, 235
and 370 min. The fourth assay is listed in Table 2. After
that assay, 500 mg of KI were added to the solution and
iodine was titrated as usual. Column B was used for
assaying cyclohexene and column C for all other
products.
Reactions of 1b and f and 2 with sodium thiophen-
oxide in the absence of thiophenol. The PhSNa was
crystallized from ethanol, washed with hexane and dried
under vacuum. A solution of 34.0 mg (0.0821 mmol) of
1b in 5 ml of MeCN containing naphthalene as GC
standard was stirred with 35.6 mg (0.0270 mmol) of
PhSNa for 2 h. GC assay gave 0.0524 mmol (64%) of Th,
0.0311 mmol (38%) of ThO, 0.0771 mmol (94%) of
cyclohexanol and 0.0521 mmol (63%) of DPDS. An
analogous reaction with 1f gave 55% of Th, 43% of ThO
and 62% 0f DPDS. The GC peak for MeOH overlapped
that for the solvent (column A). Similarly, 2 gave 46% of
Ph2S, 54% of Ph2SO, 71% of DPDS and 53% of MeOH
(column E). These reactions may be compared with those
in the presence of PhSH (Tables 3–5).
Reactions of 1 with sodium thiophenoxide in the
presence of thiophenol. An example is given with 1b
(part of entry 1, Table 3). A solution of 71.0 mg
(0.171 mmol) of 1b and 45 ml (48.3 mg, 0.439 mmol) of
PhSH in 10 ml of MeCN containing biphenyl as GC
standard was stirred with 57.8 mg (0.438 mmol) of
PhSNa in a septum-capped volumetric flask for 2 h.
The solution was assayed twice on column A giving
in mmol (average) 0.169 Æ 0.0015 (99%) of cyclohex-
anol, 0.157 Æ 0.0001 (92%) of Th, 0.0029 Æ 0.0005
(1.7%) of ThO and 0.215 Æ 0.002 (126%, based on the
amount of 1b) of DPDS. Following the GC assays, the
solution was poured into 50 ml of 2 M NaOH and the
aqueous solution was extracted with 3 Â 25 ml of diethyl
ether. The dried (MgSO4) ether solution was evaporated
to dryness and the solid residue was placed on a column
of silica gel from which 94.3 mg of a mixture of biphenyl,
Th and DPDS was eluted with light petroleum. This
mixture was dissolved in 5 ml of MeCN and assayed by
GC on column A to give 0.148 mmol (87%) of Th and
0.175 mmol (102%, based on 1b) of DPDS. This assay of
DPDS is reported in Table 3.
Reaction of 2 with KI. A solution of 61.6 mg
(0.203 mmol) of 2 and 90.8 mg (0.547 mmol) of KI in
10 ml of dry DMSO containing both biphenyl and 2-
butanone as GC standards was stirred for 2 h in a septum-
capped volumetric flask. CG assay on column A gave
0.198 mmol (98%) of Ph2SO and 0.200 mmol (99%) of
MeI. Traces of MeOH and Ph2S were also found.
Reaction of bromocyclohexane with sodium thiophen-
oxide. A solution of 43.3 mg (0.266 mmol) of bromocy-
clohexane and 91.6 mg (0.694 mmol) of PhSNa in 10 ml
of DMSO containing biphenyl as GC standard was stirred
for 1 h. GC assay gave 0.172 mmol (65%) of cyclohex-
ene, 0.111 mmol (42%) of cyclohexyl phenyl sulfide and
0.0372 mmol (11%) of DPDS. PhSH was formed but
could not be assayed because its GC peak overlapped that
of the solvent. Column A was used.
Similar reactions with 1c–e were carried out. In some
cases, both biphenyl and 2-butanone (for ROH assay)
were used as internal GC standards. In these runs,
columns A and D were used. Reactions with 1f and g
were carried out similarly but in DMSO and MeCN
(Table 4).
Experiments with sodium alkoxides. (a) NaOMe was
prepared as a dry powder from reaction of Na with dry
MeOH. A suspension of 13.8 mg (0.256 mmol) of NaOMe
Copyright 1999 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 12, 827–836 (1999)

836
B. LIU, H. J. SHINE AND W. ZHAO
2. M. A. Khuddus and D. Swern, J. Am. Chem. Soc. 95, 8393 (1973).
3. J. March, Advanced Organic Chemistry, 4th ed., p. 1194. Wiley–
Interscience, New York (1992).
4. C. R. Johnson and W. G. Phillips, J. Org. Chem. 32, 1926 (1967).
5. W. Zhao and H. J. Shine, Tetrahedron Lett. 37, 1749 (1996).
6. W. Zhao, H. J. Shine and B. R. Whittlesey, J. Org. Chem. 62, 8693
(1997).
7. W. Zhao and H. J. Shine, Can. J. Chem. 76, 695 (1998).
8. S. Oae, Organic Sulfur Chemistry: Structure and Mechanism, pp.
86–87, 161–167. CRC Press, Boca Raton, FL (1991).
9. N. J. Leonard and C. R. Johnson, J. Am. Chem. Soc. 84, 3701
(1962).
10. C. R. Johnson, J. Am. Chem. Soc. 85, 1020 (1963).
11. C. R. Johnson and D. McCants Jr, J. Am. Chem. Soc. 87, 23 (1965).
12. C. R. Johnson and W. G. Phillips, J. Am. Chem. Soc. 91, 682
(1969).
13. E. J. Corey and C. U. Kim, J. Am. Chem. Soc. 94, 7586 (1972).
14. E. J. Corey, C. U. Kim and M. Takeda, Tetrahedron Lett. 4339
(1972).
15. E. J. Corey and C. U. Kim, Tetrahedron Lett. 919 (1973).
16. E. J. Corey and C. U. Kim, J. Org. Chem. 38, 1233 (1973).
17. (a) D. Landini, G. Modena, F. Montanari and G. Scorrano, J. Am.
Chem. Soc. 92, 7168 (1970); (b) D. Landini, G. Modena, U.
Quintily and G. Scorrano, J. Chem. Soc. 2041 (1971); (c) S.
Tamagaki, M. Mizuno, H. Yoshida, H. Hirota and S. Oae, Bull.
Chem. Soc. Jpn. 44, 2456 (1971).
18. (a) K. Mislow, T. Simmons, J. T. Melillo and A. L. Ternay Jr, J.
Am. Chem. Soc. 86, 1452 (1964); (b) H. Yoshida, T. Numata and S.
Oae, Bull. Chem. Soc. Jpn. 44, 2875 (1971).
19. C. R. Johnson and J. J. Rigau, J. Am. Chem. Soc. 91, 5398 (1969).
20. S. Oae and Y. H. Kim, Bull. Chem. Soc. Jpn. 42, 3528 (1969).
21. K. Tsumori, H. Minato and M. Kobayashi, Bull. Chem. Soc. Jpn.
46, 3503 (1973).
in 10 ml of MeCN containing toluene as GC standard was
stirred for 30 min. GC on column E failed to detect MeOH.
PhSH (71 mg, 0.645 mmol) was injected into the suspen-
sion and stirring was continued for 30 min. GC again
failed to detect MeOH. Water (0.2 ml) was injected and
stirring was continued for 30 min. A small amount of solid
remained. GC gave 0.269 mmol (105%) of MeOH. A
similar experiment with 10.5 mg (0.194 mmol) of NaOMe
and 1.22 g (11.1 mmol) of PhSH again gave MeOH
(0.185 mmol, 95%) only after the addition of 0.2 ml of
water. The same type of experiments were carried out with
EtSH. With 10.5 mg (0.194 mmol) of NaOMe and 36 ml
(0.484 mmol) of EtSH, MeOH (0.192 mmol, 99%) was
obtained only after the addition of water. When a larger
amount of EtSH (11.2 mmol) was used, MeOH was
detectable again only after the addition of water, but assay
was thwarted by overlap of the MeOH peak with the very
large peak from EtSH.
(b) The sodium salt of cyclohexanol was prepared by
heating Na with cyclohexanol under reflux. The white
precipitate was filtered, washed with dry diethylether and
dried under vacuum. To 20.7 mg (0.170 mmol) of the salt
was added 10 ml of dry MeCN containing naphthalene as
GC standard. Most of the salt dissoved. After stirring for
30 min, GC assay on column A gave 0.162 mmol (95%)
of cyclohexanol. The experiment was repeated with
redried (P2O5) MeCN and again cyclohexanol (98%) was
found without the need to add PhSH or water.
22. H. Minato, T. Miura and M. Kobayashi, Chem. Lett. 701 (1975).
23. S. Winstein, D. Darwish and N. J. Holness, J. Am. Chem. Soc. 78,
2915 (1956).
24. W. T. Ford, Acc. Chem. Res. 6, 410 (1975).
25. D. J. McLennan, Tetrahedron 31, 2999 (1975).
26. J. March, Advanced Organic Chemistry, 4th ed., pp. 997, 998.
Wiley–Interscience, New York (1992).
27. E. L. Eliel and R. G. Haber, J. Am. Chem. Soc. 81, 1249 (1959).
28. D. J. McLennan, J. Chem. Soc. B 705 (1966).
29. C. R. Johnson and M. P. Jones, J. Org. Chem. 32, 2014 (1967).
30. H. Meerwein, K.-F. Zenner and R. Gipp, Justus Liebigs Ann.
Chem. 688, 67 (1965).
31. J. L. Kice, Acc. Chem. Res. 1, 58 (1968).
32. D. Cook, R. E. J. Hutchinson, J. K. MacCleod and A. J. Parker, J.
Org. Chem. 39, 534 (1974).
33. F. G. Bordwell and S. R. Mrozack, J. Org. Chem. 47, 4813 (1982).
34. C. Arnone, G. Consiglio, V. Frenna and D. Spinelli, J. Org. Chem.
62, 3093 (1997).
Reaction of 1b with PhSH. A solution of 16.2 mg
(0.039 mmol) of 1b and 28.3 mg (0.257 mmol) of PhSH
was made in 3 ml of CD3CN containing napthalene as
GC standard. Reaction (decrease of the peaks of 1b at 8.4
and 4.54 ppm) had not occurred after 18 h. After the
solution had been heated for 1 h at 100°C, GC showed
cyclohexene, ThO, PhSH and DPDS. A solution of 1b in
MeCN gave ThO (105%) and cyclohexene (100%) when
injected into the hot GC inlet.
35. R. G. Pearson and J. Songstad, J. Org. Chem. 32, 2899 (1967).
36. R. G. Pearson, H. Sobel and J. Songstad, J. Am. Chem. Soc. 90, 319
(1968).
Acknowledgements
37. W. Galezowski and E. S. Lewis, J. Phys. Org. Chem. 7, 90 (1994).
38. D. Landini and F. Montanari, J. Chem. Soc. Chem. Commun. 86
(1968).
39. J. L. Kice and T. E. Rogers, J. Org. Chem. 41, 255 (1976).
40. J. O. Edwards and R. G. Pearson, J. Am. Chem. Soc. 84, 16 (1962).
41. S. M. Bachrach and D. C. Mulhearn, J. Phys. Chem. 100, 3535
(1996), and references cited therein.
We thank the Robert A. Welch Foundation for support
(Grant D-0028) and Professors J. L. Kice, E. S. Lewis, C.
J. Stirling, N. Furukawa and J. Drabowicz for helpful
discussions.
42. M. E. Jung and P. L. Ornstein, Tetrahedron Lett. 30, 947 (1973).
43. L. Ruzicka, P. A. Plattner and H. Wild, Helv. Chim. Acta 28, 395
(1945).
REFERENCES
44. E. L. Eliel and R. S. Ro, J. Am. Chem. Soc. 79, 5995 (1967).
45. K. Iwai and H. J. Shine, J. Org. Chem. 46, 271 (1981).
46. H. J. Shine and W. Yueh, J. Org. Chem. 59, 3553 (1994).
1. H. Meerwein, E. Battenberg, H. Gold, E. Pfeil and G. Wilfang, J.
Prakt Chem. 154, 83 (1939).
Copyright 1999 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 12, 827–836 (1999)