Reversible disulfur monoxide (S2O)-forming retro-Diels-Alder reaction. Disproportionation of S2O to trithio-ozone (S3) and sulfur dioxide (SO2) and reactivities of S2O and S3

Article abstract of DOI:10.1021/ja047729i

5,6-Di-tert-buty(-2,3,7-trithiabicyclo[2.2.1]hept-5-ene 7-endo-oxide (4) was prepared by addition of S₂Cl₂ to 3,4-di-tert- butylthiophene 1-oxide (3) in high yield. The oxidation of 4 with dimethyldioxirane gave a 7:1 isomeric mixture of 5,6-di-tert-butyl-2,3,7- trithiabicyclo[2.2.1]hept-5-ene 2-endo-7-endo-dioxide (5a) and 2-exo-7-endo-dioxide (5b) quantitatively. The thermally labile 5 was shown to undergo a retro-Diels-Alder reaction that produces S₂O and 3 in a reversible way. The resulting S₂O was trapped by Diels-Alder reactions with dienes to give 3,6-dihydro-1,2-dithiin 1 -oxides in good yields. In the absence of the dienes, S₂O disproportionates to SO₂ and S₃, and the resulting S₃ underwent a 1,3-dipolar cycloaddition with 3 on its syn-π-face with respect to the S=O bond to give a trithiolane derivative, whereas in the presence of excess norbornene, it produced the 1,3-dipolar cycloadduct with norbornene in good yield. Thus, the retro-Diels-Alder reaction of 5 functions as an S₂O and S₃ source. DFT calculations at the B3LYP/6-311+G(3df,2p) level were carried out in order to explain why S₂O disproportionates to SO₂ and S₃ and why S₂O acts as a dienophile and not 1,3-dipole, whereas O₃ and S₃ serve as 1,3-dipoles.

Full text of DOI:10.1021/ja047729i

Published on Web 07/01/2004
Reversible Disulfur Monoxide (S2O)-Forming
Retro-Diels-Alder Reaction. Disproportionation of S2O to
Trithio-Ozone (S3) and Sulfur Dioxide (SO2) and Reactivities of
S2O and S3
Juzo Nakayama,* Satoshi Aoki, Jun Takayama, Akira Sakamoto,
Yoshiaki Sugihara, and Akihiko Ishii
Contribution from the Department of Chemistry, Faculty of Science, Saitama UniVersity,
Sakura-ku, Saitama, Saitama 338-8570, Japan
Received April 20, 2004; E-mail: nakaj@post.saitama-u.ac.jp
Abstract: 5,6-Di-tert-butyl-2,3,7-trithiabicyclo[2.2.1]hept-5-ene 7-endo-oxide (4) was prepared by addition
of S2Cl2 to 3,4-di-tert-butylthiophene 1-oxide (3) in high yield. The oxidation of 4 with dimethyldioxirane
gave a 7:1 isomeric mixture of 5,6-di-tert-butyl-2,3,7-trithiabicyclo[2.2.1]hept-5-ene 2-endo-7-endo-dioxide
(5a) and 2-exo-7-endo-dioxide (5b) quantitatively. The thermally labile 5 was shown to undergo a retro-
Diels-Alder reaction that produces S2O and 3 in a reversible way. The resulting S2O was trapped by
Diels-Alder reactions with dienes to give 3,6-dihydro-1,2-dithiin 1-oxides in good yields. In the absence of
the dienes, S2O disproportionates to SO2 and S3, and the resulting S3 underwent a 1,3-dipolar cycloaddition
with 3 on its syn-π-face with respect to the SdO bond to give a trithiolane derivative, whereas in the presence
of excess norbornene, it produced the 1,3-dipolar cycloadduct with norbornene in good yield. Thus, the
retro-Diels-Alder reaction of 5 functions as an S2O and S3 source. DFT calculations at the B3LYP/6-
311+G(3df,2p) level were carried out in order to explain why S2O disproportionates to SO2 and S3 and
why S2O acts as a dienophile and not a 1,3-dipole, whereas O3 and S3 serve as 1,3-dipoles.
Introduction
substrates and the application of S2O to organic synthesis in
solution under mild conditions. 4,5-Diphenyl-3,6-dihydro-1,2-
Disulfur monoxide (S2O; SdSdO), an analogue of sulfur
dioxide (SO2) which is one of the most common compounds
of sulfur, is formally derived from the latter by replacing the
oxygen atom by sulfur. S2O is suspected to be a component of
the surface and the atmosphere of Jupiter’s moon Io1 and has
been detected in the atmosphere of the planet Venus.2 Most
fundamentally, S2O is produced by incomplete combustion of
sulfur under reduced pressure.3,4 It is also generated by the
pyrolysis of ethylene episulfoxide,5 electronic discharge of a
mixture of SO2 and sulfur vapor,6 and reactions of heavy metal
oxides or sulfides with elemental sulfur or with gaseous
SOCl2.3,4,7 However, these methods, most of which are vapor
phase reactions, require forcing conditions, in addition to special
equipment and techniques. Therefore, these are not necessarily
convenient to investigate the reactions of S2O with organic
dithiin 1-oxide was found to liberate S2O, though not in a free
state, via a transition-metal-assisted retro-Diels-Alder reaction.8
Recently, we have reported the generation of S2O by thermal
decomposition of tetrathiolane 2,3-dioxide (1) (Ad ) 1-ada-
mantyl) in solution.9 The synthesis of 1 is, however, rather
laborious. More recently, we also found that readily obtainable
octasulfur monoxide (2)10 serves as an S2O equivalent.11 Here,
we report a new and clean method that generates free S2O by
a hetero retro-Diels-Alder reaction. We also report the com-
putational rationalization of the chemical properties of S2O.
(1) (a) Hapke, B. Icarus 1989, 79, 56-74. (b) Spencer, J. R.; McEwen, A. S.;
McGrath, M. A.; Sartoretti, P.; Nash, D. B.; Noll, K. S.; Gilmore, D. Icarus
1997, 127, 221-237.
(2) Na, C. Y.; Esposito, L. W. Icarus 1997, 125, 364-368.
(3) For reviews on the chemistry of S2O, see: (a) Schenk, P. W.; Steudel, R.
Angew. Chem., Int. Ed. Engl. 1965, 4, 402-409. (b) Murthy, A. R. V.;
Kutty, T. R. N.; Sharma, D. K. Int. J. Sulfur Chem. B 1971, 6, 161-175.
(4) For the latest review on sulfur-rich oxides SnO and SnO2 (n > 1), see
Steudel, R. Top. Curr. Chem. 2003, 231, 203-230.
(5) Dodson, R. M.; Srinivasan, V.; Sharma, K. S.; Sauers, R. F. J. Org. Chem.
1972, 37, 2367-2372.
(6) Schenk, P. W.; Holst, W. Z. Anorg. Allg. Chem. 1963, 319, 137-149.
(7) (a) Satyanarayana, S. R.; Murthy, A. R. V. Z. Anorg. Allg. Chem. 1964,
330, 245-250. (b) Schenk, P. W.; Steudel, R. Z. Anorg. Allg. Chem. 1966,
342, 253-262. (c) Schenk, P. W.; Steudel, R. Angew. Chem., Int. Ed. Engl.
1964, 3, 61.
Results and Discussion
Preparation of 5,6-Di-tert-butyl-2,3,7-trithiabicyclo[2.2.1]-
hept-5-ene 7-Oxide (4). In our continuing interest in the
reactions of 3,4-di-tert-butylthiophene 1-oxide (3),12,13 we have
(8) (a) Urove, G. A.; Welker, M. E. Organometallics 1988, 7, 1013-1014.
(b) Raseta, M. E.; Cawood, S. A.; Welker, M. E.; Rheingold, A. L. J. Am.
Chem. Soc. 1989, 111, 8268-8270. (c) Raseta, M. E.; Mishra, R. K.;
Cawood, S. A.; Welker, M. E.; Rheingold, A. L. Organometallics 1991,
10, 2936-2945. (d) Hayes, B. L.; Welker, M. E. Organometallics 1998,
17, 5534-5539.
9
10.1021/ja047729i CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 9085-9093
9085

A R T I C L E S
Nakayama et al.
Scheme 1
Scheme 2
Figure 1. Temperature-dependent 1H NMR spectra of the methine protons
of 7 in CD2Cl2
mer, the compound of our desire, in addition to thiophenes 10
and 11, with the product ratio being dependent upon the
conditions. Further, either treatment of the crude 7 with water
or subjection of 7 to silica gel column chromatography gave 4
in reasonable yields (50-60%) together with 10 and 11 in small
amounts. 4 was obtained in an optimized yield of 98% when 7
was treated with aqueous NaHCO3 solution. An additional
curious but reproducible observation is the nearly quantitative
formation of 10 on treatment of 7 with MeOH, where MeOH
acted as a reducing agent. Treatment of 7 with tert-BuOH also
produced 10 in 78% yield in addition to 4 in 22% yield.
We then treated 7 with tetramethylethylene and morpholine
with expectation of obtaining a more stable derivative.15
Although the reactions produced the expected products 12 and
13, respectively and quantitatively, they are still thermally labile
and decomposed completely, when allowed to stand at room-
temperature overnight, to give 4 and 10 as the principal products.
Interestingly, the methyl protons of one of the two tert-butyl
groups and one of the two methine protons of 7 appeared as
broad singlets in the room temperature 1H NMR spectrum. On
lowering the temperature, the proton signals for both of the
methine and both of the tert-butyl groups became quite broad.
At -15 °C, each singlet split into two very broad singlets, and
at -45 °C, into two clear singlets with the signal intensity ratios
of 50:50 at -15 °C and 56:44 at -45 °C in CDCl3 and of 56:
44 at -15 °C, 65:35 at -45 °C, and 71:29 at -75 °C in CD2-
Cl2 (Figure 1). The peak-broadening at room temperature and
the peak-splitting at low temperatures were also observed for
the 13C NMR. These phenomena can best be explained by
assuming a rapid equilibrium between two of the four possible
stereoisomers 7a-7d, where the equilibrium position, i.e., the
isomer ratio is dependent upon the temperature and solvent, and
planned the preparation of compounds 5 and 6 via 4 by use of
312 as the starting material (Scheme 1). Compounds 5 and 6
are expected to generate S2O and diatomic sulfur S2, respec-
1
tively, by retro-Diels-Alder reactions.
Thus, addition of S2Cl2 to 3 was examined in CH2Cl2 with
the expectation of obtaining adduct 7, 8, or 9 as the precursor
of 4 (Scheme 2). The addition took place, providing a plethora
of curious but interesting observations (see also Experimental
Section). When the solvent was removed immediately after a
solution of 3 and S2Cl2 in CH2Cl2 had been stirred for only 1
min, 1,4-adduct 714 was obtained quantitatively as a yellow oil.
Neither 8 nor 9 was formed. 7 is highly labile, and thus leaving
the crude 7 neat or as a solution at room-temperature brought
about the decomposition that produced 4 as a single stereoiso-
(9) (a) Ishii, A.; Nakabayashi, M.; Nakayama, J. J. Am. Chem. Soc. 1999, 121,
7959-7960. (b) Ishii, A.; Nakabayashi, M.; Jin, Y.-N.; Nakayama, J. J.
Organomet. Chem. 2000, 611, 127-135.
1
the equilibrium takes place slightly faster than the H NMR
(10) (a) Steudel, R.; Latte, J. Angew. Chem., Int. Ed. Engl. 1974, 13, 603-604.
(b) Steudel, R.; Sandow, T. Inorg. Synth. 1982, 21, 172-174.
time scale at room temperature, whereas it becomes slow enough
at lower temperatures, allowing us to observe the two isomers
as different molecules. The equilibrium between 7a and 7b or
(11) Ishii, A.; Kawai, T.; Tekura, K.; Oshida, H.; Nakayama, J. Angew. Chem.,
Int. Ed. 2001, 40, 1924-1926.
(12) Nakayama, J.; Yu, T.; Sugihara, Y.; Ishii, A. Chem. Lett. 1997, 499-500.
(13) (a) Nakayama, J.; Sano, Y.; Sugihara, Y.; Ishii, A. Tetrahedron Lett. 1999,
40, 3785-3788. (b) Otani, T.; Sugihara, Y.; Ishii, A.; Nakayama, J.
Tetrahedron Lett. 1999, 40, 5549-5552. (c) Otani, T.; Sugihara, Y.; Ishii,
A.; Nakayama, J. Chem. Lett. 2000, 744-745. (d) Otani, T.; Sugihara, Y.;
Nakayama, J. Tetrahedron Lett. 2000, 41, 8461-8465. (e) Nakayama, J.;
Otani, T.; Sughihara, Y.; Sano, Y.; Ishii, A.; Sakamoto, A. Heteroatom
Chem. 2001, 12, 333-348. (f) Nakayama, J.; Takayama, J.; Sugihara, Y.;
Ishii, A. Chem. Lett. 2001, 758-759. (g) Nakayama, J.; Furuya, T.; Ishii,
A.; Sakamoto, A.; Otani, T.; Sugihara, Y. Bull. Chem. Soc. Jpn. 2003, 76,
619-625. (h) Otani, T.; Takayama, J.; Sugihara, Y.; Ishii, A.; Nakayama,
J. J. Am. Chem. Soc. 2003, 125, 8255-8263. (i) Takayama, J.; Fukuda, S.;
Sugihara, Y.; Ishii, A.; Nakayama, J. Tetrahedron Lett. 2003, 44, 5159-
5162. (j) Nakayama, J. J. Synth. Org. Chem., Jpn. 2003, 61, 1106-1115.
(14) Bromine adds to 3 and the related compounds to produce 1,4-cis-mode
adducts exclusively.13g The two 1,4-cis-mode bromine adducts (syn- and
anti-adducts to the SdO bond) to 3 are stable enough to be isolated by
usual methods (silica gel column chromatography and crystallization).
Therefore, the SSCl group of 7 plays a crucial role to the mutual conversion
between the two species, and the carbenium ion 14b would not be a probable
intermediate.
(15) For a recent review on the chemistry of the SnCl (n > 1) group, see Abu-
Yousef, I. A.; Harpp, D. N. Sulfur Rep. 2003, 24, 255-282. A recent paper
on reactions of the SSCl group: Mloston, G.; Majchrzak, A.; Senning, A.;
Søtofte, I. J. Org. Chem. 2002, 67, 5690-5695.
9
9086 J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004

(S O)-Forming Retro-Diels−Alder Reaction
2
A R T I C L E S
Scheme 3
Scheme 5
elimination of Cl2 from 15 produces 4; the smell witnesses the
generation of Cl2. The hydroxide ion that originates from
NaHCO3 is more nucleophilic than is the chloride, and might
also abstract chlorine in 15 to promote the formation of 4. A
route involving elimination of HCl from 7 to give 16 is not
possible. The direct determination of the stereochemistry of the
sulfinyl group in 4 was not possible because the single crystals
for X-ray diffraction analysis could not be obtained despite
numerous efforts.
Scheme 4
Incidentally, attempted syntheses of 17a-c by reactions of
3 with SCl2, S3Cl2, and Se2Cl2 were all unsuccessful.
Preparation and Thermal Properties of 5,6-Di-tert-butyl-
2,3,7-trithiabicyclo[2.2.1]hept-5-ene 2,7-Dioxide (5). The oxi-
dation of 4 by a slight excess of dimethyldioxirane (DMD) at
-18 °C, followed by removal of the solvent at -18 °C,
furnished 2,7-dioxide 5 quantitatively as a 7:1 mixture of two
diastereomers (Scheme 5). These are thermally highly unstable
and decompose even at room-temperature both in solution and
as the solid. The major diastereomer was isolated by crystal-
lization at -18 °C and the structure determined to be 5a by
X-ray crystallographic analysis (Figure 2). Thus the minor
diastereomer was assigned as 5b. The 1H NMR chemical shift
data of the methine protons of 5b are also in harmony with this
structure; for example, Ha of 5a, located in the deshielding zone
of the newly introduced SdO group, is more deshielded (δ 6.28)
than both methine protons of 5b (δ 5.81, 5.93).19 The same
conclusion is also reached by inspection of the chemical shift
values of the tert-butyl protons. This assignment in turn means
that the SdO bond in 4 is syn to the S-S bond.
Brief heating of a crude 7:1 mixture of 5a and 5b in boiling
CHCl3 gave trithiolane 18 in 43% yield (86% yield based on
the sulfur atom) in addition to 3 in 45% yield (Scheme 6). The
structure of 18 was determined by X-ray crystallographic
analysis (Figure 3). The formation of 18 is easily explainable.
The retro-Diels-Alder reaction of 5 gives 3 and S2O. The
resulting S2O disproportionates to S3 and SO2.9,11,20 Finally, 1,3-
dipolar cycloaddition of S3 with 3 furnishes 18, by reaction
exclusively on the syn-π-face of 3 with respect to the SdO
bond consistent with Diels-Alder reactions of 3 which occur
on the syn-π-face.13h-j
between 7c and 7d is the likely candidate for such an
equilibrium, where the equilibrium might be attained through
intermediary formation of the episulfonium ion 14a,14 although
we cannot say which combination is more probable from the
NMR data. Other combinations, i.e., the equilibria between 7a
and 7c, 7a and 7d, 7b and 7c, and 7b and 7d, seem least
probable because such equilibria require that inversion of the
configuration at >CHSSCl takes place rapidly, i.e., addition of
S2Cl2 to 3 is a rapid reversible reaction.16 However, the
experimental observation does not fulfill the requirement
1
because the H NMR spectrum shows the independent signals
of 3 and the product 7, when 3 and S2Cl2 were mixed up in the
ratio 1:1 in CDCl3. Isomer 8 is not appreciably involved in the
equilibrium because the 13C NMR of 7 shows two sp2-carbon
peaks at δ 143 and 147, reasonable values for sp2 carbons that
carry a tert-butyl group.13g-i,17
Interestingly, both methine and tert-butyl protons of both 5a
and 5b appeared as broad singlets in the 1H NMR spectrum at
25 °C. This led us to determine variable temperature 1H NMR
spectra in the range from -40 to 50 °C. In addition, progress
of the decomposition of 5 was followed (Figure 4). Although
To account for the conversion of 7 to 4, we suggest the
following tentative mechanism. Initially 7 isomerizes to the 1,2-
adduct 8 through 14a. An intramolecular addition in 8 produces
the carbenium ion 15, which is stabilized by neighboring group
participation of the two sulfur atoms.18 Water, through its highly
polar nature, probably promotes the formation of 15. Finally,
(19) (a) Green, C. H.; Hellier, D. G. J. Chem. Soc., Perkin Trans. 2 1972, 458-
463. (b) Green, C. H.; Hellier, D. G. J. Chem. Soc., Perkin Trans. 2 1973,
243-252. (c) Pritchard, J. G.; Lauterbur, P. C. J. Am. Chem. Soc. 1961,
83, 2105-2110. (d) Green, C. H.; Hellier, D. G. J. Chem. Soc., Perkin
Trans. 2 1975, 190-193. (e) Buchanan, G. W.; Hellier, D. G. Can. J. Chem.
1976, 54, 1428-1432.
(20) On annealing matrix-isolated S2O decomposes primarily to SO2 and S3;
Tang, S.-Y.; Brown, C. W. Inorg. Chem. 1975, 14, 2856-2858. In the gas
phase, S2O survives for several days below 100 Pa and at temperatures
below 20 °C.4
(16) Pyramidal inversion at the sulfinyl center would not be the case because
such inversion does not take place for the two 1,4-cis-mode bromine adducts
to 3.13g
(17) Nakayama, J.; Hasemi, R.; Yoshimura, K.; Sugihara, Y.; Yamaoka, S. J.
Org. Chem. 1998, 63, 4912-4924.
(18) For neighboring group participation of sulfur, see Block, E. Reactions of
Organosulfur Compounds, Academic Press: New York, 1978; pp 141-
145, and references therein.
9
J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004 9087

A R T I C L E S
Nakayama et al.
Figure 3. Molecular structure of 18.
Figure 2. Molecular structure of 5a.
rotation of the tert-butyl groups17 or conformational change of
the trithiolane ring.22
Scheme 6
It is well documented that Diels-Alder reactions of 3 with
a great number of dienophiles take place exclusively on the syn-
π-face with respect to the SdO bond.13h,i If we take this account
into consideration, then the ratio 7:1 of 5a and 5b in CDCl3
will not change regardless of the number of times the reversible
addition of S2O to 3 occurs (Scheme 7). Indeed, letting a 12:1
mixture of 5a and 5b in CDCl3 stand at room temperature gave
9:1 and 7:1 mixtures after 3 and 6 h, respectively.
discussion here is concentrated only on the signals of the
methine protons of 5 and those of 3 and 18, similar phenomena
were observed for the tert-butyl protons. Thus, at lower
temperatures, for example, at -35 °C, the methine protons of
both 5a and 5b appeared as clear doublets because of long-
range coupling, and the R-protons of 3 that exists as an impurity
were observed as a clear singlet (Figure 4a). At 15 °C, the
methine proton signals of 5 became broad singlets, while the
R-proton signal of 3 became obscure by signal-broadening
(Figure 4b). In addition, the methine and vinyl protons signals
of 18, which formed by decomposition of 5, began to appear.
At 25 °C, the methine proton signals of 5 became quite broad
(Figure 4c). At the same time, the signal intensity of 18
increased, whereas the R-protons signal of 3 is virtually unseen
by signal-broadening notwithstanding that both 18 and 3 are
produced by decomposition of 5. Letting the sample stand at
25 °C for 12 h resulted in the further decomposition of 5 (Figure
4d). This was re-cooled and, again at -35 °C, the methine
protons of 5a and 5b became clear doublets and the R-protons
of 3 a singlet, whereas the signals of 18 appeared as rather broad
singlets (Figure 4e). Warming of the sample to 50 °C brought
about the decomposition of most 5, where the R-protons of 3
appeared as very broad singlet, whereas methine protons signals
of 5 are unseen because of broadening; the signals of 18 were
observed as sharp singlets (Figure 4f). After the decomposition
of 5 had been completed, both signals of 3 and 18 appeared as
clear singlets at 25 °C (Figure 4g).
2,7-Dioxide 5 as the S2O Source. The above results indicate
that the retro-Diels-Alder reaction of 5 serves as a clean source
of S2O. Indeed, when the crude 5 was heated with excess dienes
in refluxing CH2Cl2, the resulting S2O reacted with the dienes
to give Diels-Alder adducts 19a,b8,9,11 in high yields (Scheme
8). On the other hand, the reaction of 5 with an equimolar
amount of (Ph3P)2Pt(C2H4) took place quickly even at room
temperature to give a platinum complex 2023 in 60% yield. Thus,
20 is likely formed directly by reaction of 5 with the metallic
reagent; contribution of free S2O would be small if any.
2,7-Dioxide 5 as the S3 Source. The retro-Diels-Alder
reaction of 5 is also expected to serve as the S3 source through
disproportionation of S2O. Indeed, decomposition of crude 5
in the presence of excess norbornene provided 21,9,11,24 the 1,3-
dipolar cycloadduct of S3 with norbornene, in 82% yield
(Scheme 9). The decomposition of 5 in the presence of excess
cyclopentadiene produced, the 1,3-dipolar cycloadduct, 2225 in
80% yield, and not the Diels-Alder adduct, 24. This probably
means that 24, even if it formed, undergoes the cycloreversion
to S2O and cyclopentadiene. The reaction also produced 23 in
58% yield, which formed by reaction of cyclopentadiene with
3, the cycloreversion counterpart of S2O. The decomposition
in the presence of norbornadiene did not give the expected
(21) For some examples of temperature-dependent 1H NMR spectra of the
equilibrating systems, see cycloheptatriene-norcaradiene system: (a) Vogel,
E.; Wendisch, D.; Roth, W. R. Angew. Chem. 1964, 76, 432-433. (b)
Schro¨der, G.; Oth, J. F. M.; Mere´nyi, R. Angew. Chem., Int. Ed. Engl.
1965, 4, 752-761. (c) Ciganek, E. J. Am. Chem. Soc. 1967, 89, 1458-
1468. Benzene oxide-oxepin system: (d) Vogel, E.; Gu¨nther, H. Angew.
Chem., Int. Ed. Engl. 1967, 6, 385-401. Tri-tert-butylcyclopropenyl
azide: (e) Curci, R.; Lucchini, V.; Kocienski, P. J.; Evans, G. T.; Ciabattoni,
J. Tetrahedron Lett. 1972, 3293-3296.
The signal-broadening of 5 and 3 can be explained by
assuming a reversible reaction21 between 5 and (3 + S2O) that
takes place, probably in a solvent cage, at 15 °C or higher
temperatures and is frozen at the lower temperatures (Scheme
7). The broadening, brought about by a rise in temperature,
cannot be explained by restricted rotation of the tert-butyl
groups. On the other hand, the signal-broadening of 18 at low
temperatures can be explained as a result of either restricted
(22) Sugihara, Y.; Abe, K.; Nakayama, J. Heteroatom Chem. 1999, 10, 638-
643.
(23) Ishii, A.; Murata, M.; Oshida, H.; Matsumoto, K.; Nakayama, J. Eur. J.
Inorg. Chem. 2003, 20, 3716-3721.
(24) (a) Bartlett, P. D.; Ghosh, T. J. Org. Chem. 1987, 52, 4937-4943. (b)
Leste-Lasserre, P.; Harpp, D. N. Tetrahedron Lett. 1999, 40, 7961-7964.
(25) Steliou, K.; Gareau, Y.; Milot, G.; Salama, P. J. Am. Chem. Soc. 1990,
112, 7819-7820.
9
9088 J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004

(S O)-Forming Retro-Diels−Alder Reaction
2
A R T I C L E S
Scheme 7
Scheme 8
Scheme 9
adduct of 3 with norbornadiene, is ruled out because a separate
experiment showed that the reaction of 3 with norbornadiene
did not afford 28 under the above conditions. Incidentally, during
the preparation of this paper, the rotational spectrum and
geometrical structure of S3 were communicated in this journal.27
1
Approach to S2 Source, 5,6-Di-tert-butyl-2,3,7-trithiabi-
cyclo[2.2.1]hept-5-ene (6). The 7-oxide 4 is thermally stable,
in marked contrast to 5, and shows no tendency, at least under
mild conditions, to undergo the retro-Diels-Alder reaction that
produces1S2 and 3. Meanwhile, 6 might undergo retro-Diels-
Alder reaction readily to generate 1S2 because the formation of
aromatic 10 serves as the driving force. Consequently, we then
attempted conversion of 4 to 6.
Heating 5 with Lawesson’s reagent (LR), a reagent that
reduces sulfoxides to the corresponding sulfides,28 in toluene
at 100 °C provided 10 and 1,2,3,4-tetrathiocin 3029 in 23% and
31% yields, respectively (Scheme 10). The formation of 30 is
Figure 4. Temperature-dependent 1H NMR spectra of 5, 3, and 18 in the
region δ 5.5-7.1 and the progress of the decomposition of 5: (a) at -35
°C; (b) at 15 °C; (c) at 25 °C; (d) at 25 °C after letting stand at 25 °C for
12 h; (e) at -35 °C; (f) at 50 °C after determination of (e); (g) at 25 °C
after completion of decomposition of 5.
[2+2+2] (homo-Diels-Alder) cycloadduct 27,26 but instead
gave adducts 25 (5%) and 26 (43%). Thus, S2O is inert to
norbornadiene and disproportionates to S3 and SO2. The
resulting S3 undergoes a 1,3-dipolar cycloaddition with nor-
bornadiene to provide 25 initially, which then reacts with 3 to
give 26. The formation of 26 by reaction of S3 with 28, the
(27) McCarthy, M. C.; Thorwirth, S.; Gottlieb, C. A.; Thaddeus, P. J. Am. Chem.
Soc. 2004, 126, 4096-4097.
(28) Rasmussen, J. B.; Jørgensen, K. A.; Lawesson, S.-O. Bull. Soc. Chim. Belg.
1978, 87, 307-308.
(29) Monocyclic 1,2,3,4-tetrathiocine is a rare class of compounds. For tricyclic
1,2,3,4-tetrathiocines, see (a) Bigoli, F.; Pellinghelli, M. A.; Atzei, D.;
Deplano, P.; Trogu, E. F. Phosphorus, Sulfur 1988, 37, 189-194. (b)
Schroth, W.; Felicetti, M.; Hintzsche, E.; Spitzner, R.; Pink, M. Tetrahedron
Lett. 1994, 35, 1977-1980.
(26) (a) Blomquist, A. T.; Meinwald, Y. C. J. Am. Chem. Soc. 1959, 81, 667-
672. (b) Williams, J. K.; Benson, R. E. J. Am. Chem. Soc. 1962, 84, 1257-
1258.
9
J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004 9089

A R T I C L E S
Nakayama et al.
Scheme 10
Figure 5. Bond angles (°) and bond lengths (Å) of SO2, S2O, and S3
calculated at B3LYP/6-311+G(3df) level.
Figure 6. Experimental bond angle (°) and bond length (Å) data of SO2,
25
S2O,36 and S3 in the gas phase.
Figure 7. Thermochemistry of the disproportionation of S2O to SO2 and
S3.
explained by skeletal rearrangement of the initially formed
311+G(3df) level.36,37 For comparison, the experimental bond
thiosulfoxide intermediate 29.30 Meanwhile, 10 could be
27
lengths and bond angles of SO2, S2O,38 and S3 are given in
1
produced by extrusion of S2 from 6 that formed from 29 by
Figure 6. The calculated structures are in good agreement with
the experimental ones. The calculations on enthalpy of the
reaction predict that the disproportionation (2S2O f SO2 + S3)
is 10.3 kcal mol-l exothermic (see Figure 7). The calculations
also revealed that OdSdS and OdSdO are more stable than
their isomers SdOdS and OdOdS by 63.8 and 116.5 kcal
mol-1, respectively.
Previously, the [2+2] self-dimer 32 was proposed as the
intermediate of the disproportionation of S2O.20 However, the
[2+3] dimer 31, which would be formed by a symmetry allowed
1,3-dipolar cycloaddition, seems to be a more likely intermediate
both kinetically and thermodynamically. We then calculated the
enthalpy of formation of 31 and 32 at B3LYP/6-311+G(3df)
level. The calculations predicted that the formation of 32 is
largely endothermic by 15.7 kcal mol-1, whereas the formation
of 31 is endothermic only by 1.6 kcal mol-1 (Figure 7).39 Thus,
31 would be more favorable as an intermediate. In addition,
the calculations predicted that the [2+3] dimerization is
symmetry allowed and the energy gap between HOMOπ (-8.83
eV) and LUMOπ (-4.27 eV) of S2O is as small as 4.56 eV
(Figure 8).40 This will render the 1,3-dipolar dimerization
loss of the sulfur atom. Competitively, 10 might be also formed
from 30 together with S3. Indeed, heating 30 with norbornene
in refluxing xylene afforded 21, together with 10, in good yield
based on the consumed 30.31 The above results are suggestive
but not conclusive of the intermediary formation of 6 as the
1S2 source.31a,32
Although the reduction of 4 by Me2SiCl2/Zn33 produced 10
in 63% yield, its probable counterpart 1S2 could not be trapped
by 2,3-dimethyl-1,3-butadiene. This might mean that 1S2 (or 6)
reacted more quickly with metallic zinc than with the diene, if
it had formed. Attempted reduction by Me3SiCl (or Me2SiCl2)/
PhSH and by Ac2O/iso-PrOH34 resulted in the quantitative
recovery of 4.
Computational Rationalization of Properties of S2O and
S3. We have observed that 1) S2O disproportionates to SO2 and
S3, and 2) S2O acts as a dienophile and not a 1,3-dipole, whereas
O3 and S3 serve as 1,3-dipoles. To rationalize these observations,
we have carried out DFT calculations.
Figure 5 shows the bond angles and bond lengths of SO2,
S2O,35 and S3 obtained by DFT calculations at B3LYP/6-
(30) (a) Shimada, K.; Kodaki, K.; Aoyagi, S.; Takikawa, Y.; Kabuto, C. Chem.
Lett. 1999, 695-696. (b) Ishii, A.; Yamashita, R.; Saito, M.; Nakayama,
J. J. Org. Chem. 2003, 68, 1555-1558.
(36) The calculations were carried out at both B3LYP/6-311+G(d) and B3LYP/
6-311+G(3df) levels, although the results of the latter level calculations
are described here; for the former level calculations and the vibrational
models and frequencies of these molecules, see Supporting Information.
(37) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M.
A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.; Stratmann,
R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin,
K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi,
R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.;
Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.;
Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz,
J. V.; Baboul, A. G.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.;
Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham,
M. A.; Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.;
Gill, P. M. W.; Johnson, B. G.; Chen, W.; Wong, M. W.; Andres, J. L.;
Head-Gordon, M.; Replogle, E. S.; Pople, J. A. Gaussian 98 Revision A.7;
Gaussian, Inc.: Pittsburgh, PA, 1998.
(31) The formation of 21 does not necessarily require the involvement of S3
since 21 is formed by reaction of norbornene with other sulfur species.23
See also (a) Steliou, K.; Salama, P.; Brodeur, D.; Gareau, Y. J. Am. Chem.
Soc. 1987, 109, 926-927. (b) Okuma, K.; Kuge, S.; Koga, Y.; Shioji, K.;
Wakita, H.; Machiguchi, T. Heterocycles 1998, 48, 1519-1522.
(32) (a) Schmidt, M.; Go¨rl, U. Angew. Chem., Int. Ed. Engl. 1987, 26, 887-
888. (b) Harpp, D. N.; MacDonald, J. G. J. Org. Chem. 1988, 53, 3812-
3814. (c) Steliou, K.; Gareau, Y.; Milot, G.; Salama, P. J. Am. Chem. Soc.
1990, 112, 7819-7820. (d) Clennan, E. L.; Stensaas, K. L. Org. Prep.
Proced. Int. 1998, 30, 551-600.
(33) Nagasawa, K.; Yoneta, A.; Umezawa, T.; Ito, K. Heterocycles 1987, 26,
2607-2609.
(34) Numata, T.; Togo, H.; Oae, S. Chem. Lett. 1979, 329-332.
(35) For previous ab intio calculation studies on S2O, see (a) Ivanic, J.; Atchity,
G. J.; Ruedenberg, K. J. Chem. Phys. 1997, 107, 4307-4317. (b) Fueno,
T.; Buenker, R. J. Theor. Chim. Acta 1988, 73, 123-134. A DFT calculation
study predicted a cyclic ground-state structure of S2O.; (c) Jones, R. O.
Chem. Phys. Lett. 1986, 125, 221-224.
(38) Marsden, C. J.; Smith, B. J. Chem. Phys. 1990, 141, 325-334.
(39) B3LYP/6-311+G(d) level calculations predicted that the formation of 31
is 1.3 kcal mol-1 exothermic.
9
9090 J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004

(S O)-Forming Retro-Diels−Alder Reaction
2
A R T I C L E S
Scheme 11
Scheme 12
Figure 8. π-HOMO and π-LUMO of S2O.
favorable by lowering the transition state energy with effective
HOMOπ-LUMOπ interactions.
Both experimental (1.885 Å) and calculated (1.894 Å) S-S
bond lengths in S2O are shorter than both experimental (1.917
Å) and calculated (1.922 Å) S-S bond lengths in S3. But the
S-O bond length in S2O is slightly longer than that in SO2.
The S-S bond length of S2O is also slightly shorter than those
of [(RO)2SdS], which are of the order of 1.90-91 Å.41 These
indicate that the S-S bond order in S2O is greater than that in
S3 because of greater contribution of the canonical structure
33′ than 33′′. The same conclusion is also reached by inspection
of the HOMOπ of S2O in which the sulfur atoms are connected
by a π-bonding orbital (Figure 8). This would explain why S2O
acts as a dienophile toward dienes, whereas S3 does not. For
another point of view, the energy gap between the LUMOπ of
S2O and the HOMO (-6.53 eV) of 2,3-dimethyl-1,3-butadiene
is only 2.26 eV because of the low-lying LUMO (-4.27 eV)
of S2O, and thus effective interactions between these orbitals
take place to lower the transition state energy of the Diels-
Alder reaction.
obtainable as isolable, stable compounds, if we could devise a
suitable method for their synthesis.
Finally, we examined the thermochemistry of 1,3-dipolar
cycloadditions of O3, S2O, and S3 with norbornene by B3LYP/
6-311+G(3df,2p) level calculations. Results of the calculations
are summarized in Scheme 12. It is well-known that a number
of addition reactions of norbornene take place exclusively at
its exo-face.42 The calculations at the enthalpy of adducts predict
that exo-1,3-dipolar adducts (34, 35, and 21) are more stable
than the corresponding endo-1,3-dipolar adducts (34′, 35′, and
21′) by 3.6, 3.0, and 3.0 kcal mol-1, revealing that the exo-
selection is in harmony with thermodynamic stability of the
adducts. The 1,3-dipolar cycloadditions of O3 and S3 at the exo-
face of norbornene are exothermic by as large as 63.9 and 28.5
kcal mol-1, respectively. Unexpectedly, even 1,3-dipolar cy-
cloaddition of S2O is 13.8 kcal mol-1 exothermic, though less
so than that of O3 and S3. In the latter case, however, it must
be difficult to reach the fully symmetrical transition state by
concurrent two-bond formation between S2O and norbornene
because the HOMO lobes of sulfur and oxygen in S2O are not
Calculations at B3LYP/6-311+G(3df,2p) level predicted that
the Diels-Alder reaction of S2O with 2,3-dimethyl-1,3-buta-
diene that forms 19a is exothermic by 8.5 kcal mol-1, while
the reaction that gives 19a′ is less exothermic by 4.1 kcal mol-1
(Scheme 11). Thus, the former reaction is thermodynamically
more favorable than the latter by 4.4 kcal mol-1 in accordance
with the exclusive formation of 19a. However, the 19a-forming
reaction is exothermic only by 8.5 kcal mol-1. This in turn
means that 19a, when heated, can undergo the retro-Diels-
Alder reaction that regenerates S2O. The difference of enthalpy
of formation of 19a and 19a′ is much smaller than we expected,
in other words, 19a′ is much stabler than we expected. Thus,
thionosulfinate esters R-S(dS)-O-R such as 19a′ may be
(40) (a) Huisgen, R. In 1,3-Dipolar Cycloaddition Chemistry; Padwa, A., Ed.;
John Wiley: New York, 1983; Vol. 1, Chapter 1. (b) Houk, K. N.;
Yamaguchi, K. In 1,3-Dipolar Cycloaddition Chemistry; Padwa, A. Ed.;
John Wiley: New York, 1983; Vol. 2, Chapter 13.
(41) (a) Harpp, D, N.; Steliou, K.; Cheer, C. J. J. Chem. Soc., Chem. Commun.
1980, 825-826. (b) Snyder, J. P.; Nevins, N.; Tardif, S. L.; Harpp, D. N.
J. Am. Chem. Soc. 1997, 119, 12 685-12 686. (c) Zysman-Colman, E.;
Abrams, C. B.; Harpp, D. N. J. Org. Chem. 2003, 68, 7059-7062. (d)
Tanaka, S.; Sugihara, Y.; Sakamoto, A.; Ishii, A.; Nakayama, J. J. Am.
Chem. Soc. 2003, 125, 9024-9025.
(42) For the exo-selection and enhanced reactivities of norbornene, see (a)
Schleyer, P. von R. J. Am. Chem. Soc. 1967, 89, 701-703. (b) Brown, H.
C.; Hammar, W. J.; Kawakami, J. H.; Rothberg, I.; Vander Jagt, D. L. J.
Am. Chem. Soc. 1967, 89, 6381-6382. (c) Inagaki, S.; Fujimoto, H.; Fukui,
K. J. Am. Chem. Soc. 1976, 98, 4054-4061. (d) Huisgen, R.; Ooms, P. H.
J.; Mingin, M.; Allinger, N. L. J. Am. Chem. Soc. 1980, 102, 3951-3953.
9
J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004 9091

A R T I C L E S
Nakayama et al.
was stirred only for 1 min at room temperature. The solvent was
immediately removed under reduced pressure to give the yellow oily
residue, which was stirred for another one minute. The resulting crude
7 was stirred with a aqueous saturated solution of NaHCO3 (15 mL)
for a while. The mixture was extracted with CH2Cl2. The CH2Cl2 layer
was washed with water, dried over MgSO4, and evaporated. The
crystalline residue was washed with a small amount of pentane to give
127 mg (98%) of practically pure 4. 4: mp 195-196 °C (dec); pale
yellow crystals; 1H NMR (CDCl3, 200 MHz) δ 1.35 (s, 18H), 5.38 (s,
2H); 13C NMR (CDCl3, 50 MHz) δ 32.8, 35.1, 80.0, 141.1; IR (KBr)
1096 cm-1 (SdO). Anal. Calcd for C12H20OS3: C, 52.13; H, 7.29.
Found: C, 52.37; H, 7.33.
Treatment of the crude 7 (prepared from 0.47 mmol of 3) with water
(10 mL), and not NaHCO3, also gave 4, but in decreased yield (82 mg,
59%), in addition to a mixture of 10 and 11 (26%). Subjection of the
crude 7 (prepared from 0.47 mmol of 3) to silica gel column
chromatography also produced 4, but in decreased yield (75 mg, 54%),
in addition to a mixture of 10 and 11 (35%). Treatment of the crude 7
(prepared from 0.47 mmol of 3) with MeOH (1 mL) gave 10 (89 mg,
96%).
of the same size, much greater at the sulfur atom than at the
oxygen (Figure 8). This would explain why S3 undergoes 1,3-
dipolar cycloadditions, whereas S2O prefers to act as a dieno-
phile, even though the difference of the exothermicity described
above would not be a negligible factor.
In conclusion, the retro-Diels-Alder reaction of 5 that
produces S2O and 3 is reversible and serves as an excellent
source of both S2O and S3, and will set the next stage for
developing the organic chemistry of these rather little explored
species.
Experimental Section
Preparation of 5,6-Di-tert-butyl-2,3,7-trithiabicyclo[2.2.1]hept-
5-ene 7-Oxide (4) by Reaction of 3,4 -Di-tert-butylthiophene 1-oxide
(3) with S2Cl2. (a) Isolation of the Intermediate 7. 3,4-Di-tert-
butylthiophene 1-oxide (3) (100 mg, 0.47 mol) and S2Cl2 (64 mg, 0.47
mmol) were dissolved in CH2Cl2 (5 mL) and stirred for only one minute
at room temperature. The solvent was immediately removed under
reduced pressure to give a yellow oily residue, which was stirred for
another 1 min to quantitatively provide the intermediate 7. For
Preparation of 5,6-Di-tert-butyl-2,3,7-trithiabicyclo[2.2.1]hept-
5-ene 2,7-Dioxide (5). A 82 mM solution of DMD in acetone (5.0
mL, 0.41 mmol) was added to a solution of 4 (103 mg, 0.37 mmol) in
CH2Cl2 (3 mL) at -18 °C. After the mixture had been stirred for 3.5
h at -18 °C, the solvent was removed below -18 °C to furnish a 7:1
mixture of 5a and 5b quantitatively as crystalline solid. Crystallization
of the crude product at -18 °C from CH2Cl2/hexane gave 69 mg (63%)
1
temperature-dependent H NMR spectra in CD2Cl2, see Figure 1. 7:
yellow oil; 1H NMR (400 MHz); CDCl3 as the solvent at 298 K δ 1.44
(s, But, 9H), 1.49 (broad s, But, 9H), 5.61 (s, 1H), 5.85 (broad s, 1H),
at 258 K δ 5.57 (s, 0.5H), 5.72 (s, 0.5H), 5.78 (s, 0.5H), 6.03 (s, 0.5H),
at 228 K δ 5.60 (s, 0.44H), 5.78 (s, 0.56H), 5.83 (s, 0.44H), 6.06 (s,
0.56H) (tert-butyl signals are omitted); CD2Cl2 as the solvent at 298 K
δ 1.43 (s, But, 9H), 1.49 (broad s, But, 9H), 5.62 (s, 1H), 5.85 (broad
s, 1H); C6D6 as the solvent at 298 K δ 1.01 (s, But, 9H), 1.36 (broad
s, But, 9H), 4.97 (s, 1H), 5.80 (broad s, 1H); 13C NMR (CDCl3, 100.6
MHz) 298 K δ 32.1, 32.2 (broad), 36.4, 36.5, 79.0, 86.0 (broad), 143.4,
146.8 (broad); 13C NMR (CD2Cl2, 100.6 MHz) 198 K δ 32.3, 32.7,
36.87, 36.94, 78.2, 90.4, 144.3, 145.1 (major isomer); 32.2, 33.3, 36.6,
37.5, 76.0, 86.3, 147.8, 150.9 (minor isomer); UV/Vis (CH2Cl2) λmax
(ꢀ) 266 nm (3800, shoulder).
1
of practically pure 5a as colorless crystals: mp < 90 °C (dec); H
NMR (CDCl3, 400 MHz, 238 K) δ 1.33 (s, 9H), 1.49 (s, 9H), 5.99 (d,
J ) 1.9 Hz, Hb), 6.28 (d, J ) 1.9 Hz, Ha); 13C NMR (CDCl3, 100.6
MHz, 238 K) δ 31.6, 32.1, 34.4, 36.7, 82.5, 92.6, 138.5, 149.8; IR
(KBr) 1093 (SdO), 1078 (SdO) cm-1. Anal. Calcd for C12H20O2S3:
C, 49.28; H, 6.89. Found: C, 49.32; H, 6.54. 5b: 1H NMR (CDCl3,
400 MHz, 238 K) δ 1.31 (s, 9H), 1.33 (s, 9H), 5.81 (d, J ) 2.0 Hz, Ha
or Hb), 5.93 (d, J ) 2.0 Hz, Ha or Hb). For definition of Ha and Hb, see
Scheme 5.
The following were observed when the reaction was carried out in
1
Thermal Decomposition of a 7:1 Mixture of 5a and 5b: Forma-
tion of 3 and 18. A solution of a crude 7:1 mixture of 5a and 5b,
prepared from 44 mg (0.15 mol) of 4, in CHCl3 (3 mL) was heated at
50 °C for 1 h. After the solvent had been removed, the residue was
purified by silica gel column chromatography and then by GPC (gel
permeation chromatography) to give 15 mg (45%) of 3 and 21 mg
(43% or 86% based on the sulfur atom) of 18. 18: mp 132-133 °C
CDCl3 and monitored by H NMR. On stirring even for 1 h, 3 still
remained unreacted, while a slight decomposition of the resulting 7
was observed. 3 was consumed completely after 3 h with considerable
decomposition of the resulting 7 that gave rise to 3,4-di-tert-butyl-
thiophene (10) and 3,4-di-tert-butyl-2-chlorothiophene (11) in addition
to 5,6-di-tert-butyl-2,3,7-trithiabicylo[2.2.1]hept-5-ene 7-oxide (4).
Thus, in a dilute solution the addition of S2Cl2 to 3 is rather slow, but
seemingly accelerated when the solution was concentrated or the solvent
was removed.
1
(dec); faint yellow crystals (from CH2Cl2/hexane); H NMR (CDCl3,
400 MHz) δ 1.27 (s, 9H), 1.42 (s, 9H), 5.59 (s, 1H), 6.66 (s, 1H); 13
C
NMR (CDCl3, 100.6 MHz) δ 30.5, 34.2, 37.7, 40.5, 89.4, 101.1, 135.8,
161.8; IR (KBr) 1032 (SdO) cm-1. Anal.Calcd for C12H20OS4 C, 46.71;
H, 6.53. Found: C, 46.82; H, 6.54.
(b) Derivation of 12 and 13 from 7. A solution of tetramethyleth-
ylene (40 mg, 0.47 mmol) in CH2Cl2 (3 mL) was added to 7 (prepared
from 0.47 mmol of 3). Immediately after the addition, the solvent was
removed under reduced pressure to give the thermally unstable adduct
12 quantitatively as viscous colorless oil. The reaction of 7 with
morpholine (two molar amounts) was done in a similar way to give
the thermally unstable adduct 13 quantitatively as viscous colorless
oil; the resulting hydrochloride salt of morpholine was removed by
Trapping of S2O by Dienes: Formation of 19. A solution of a
crude mixture of 5a and 5b, prepared from 21 mg (0.077 mol) of 4,
and 2,3-dimethylbutadiene (126 mg, 1.54 mmol) in CH2Cl2 (3 mL)
was heated at 30 °C for 15 h. After the solvent had been removed, the
residue was analyzed by 1H NMR with dibenzyl as the internal standard.
The analysis revealed the formation of 19a9a,11 and 3 in 82% and 86%
yields, respectively. Similarly, 19b5,11 was formed in 86% yield along
with 3 in 80% yield. 19a: 1H NMR (CDCl3, 200 MHz) δ 1.95 (s, 3H),
2.02 (s, 3H), 3.18 (d, J ) 13.7 Hz, 1H), 3.23 (d, J ) 13.1 Hz, 1H),
3.76 (d, J ) 13.7 Hz, 1H), 3.93 (d, J ) 13.1 Hz, 1H). 19b: 1H NMR
(CDCl3, 200 MHz) δ 3.65 (d, J ) 13.2 Hz, 1H), 3.65 (d, J ) 13.2 Hz,
1H), 4.23 (d, J ) 13.2 Hz, 1H), 4.47 (d, J ) 13.2 Hz, 1H), 7.10-7.26
(m, 10H).
1
filtration. 12: colorless oil; H NMR (CDCl3, 400 MHz) δ 1.40 (s,
9H), 1.45 (s, 9H), 1.60 (s, 3H), 1.63 (s, 3H), 1.72 (s, 3H), 1.73 (s, 3H),
5.17 (s, 1H), 5.74 (s, 1H); (C6D6, 400 MHz) δ 1.12 (s, 9H), 1.17 (s,
9H), 1.51 (s, 3H), 1.52 (s, 3H), 1.54 (s, 3H), 1.55 (s, 3H), 5.02 (s, 1H),
5.71 (s, 1H); 13C NMR (CDCl3, 100.6 MHz) δ 29.2, 29.3, 31.9, 32.0,
32.7, 33.4, 36.3, 36.4, 53.4, 59.7, 81.2, 83.0, 144.9, 145.4; MS (FAB)
m/z 431 (MH+). 13: colorless oil; 1H NMR (CDCl3, 400 MHz) δ 1.40
(s, 9H), 1.46 (s, 9H), 2.96-3.01 (m, 2H), 3.05-3.10 (m, 2H), 3.68-
3.78 (m, 4H), 5.40 (s, 1H), 5.70 (s, 1H); 13C NMR (CDCl3, 100.6 MHz)
δ 32.0, 32.1, 36.3 (overlapping of two peaks), 55.7, 66.8, 80.9, 83.0,
144.7, 145.5.
Formation of Platinum Complex 20. A solution of a crude mixture
of 5a and 5b, prepared from 11 mg (0.04 mol) of 4, and (Ph3P)2Pt-
(C2H4) (30 mg, 0.04 mmol) in toluene (7 mL) was stirred for 75 min
at room temperature. The solvent was removed and the residue was
purified by silica gel column chromatography and then by GPC to give
(c) Optimized Procedure for the Preparation of 4. A solution of
S2Cl2 (64 mg, 0.47 mmol) and 3 (100 mg, 0.47 mol) in CH2Cl2 (5 mL)
9
9092 J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004

(S O)-Forming Retro-Diels−Alder Reaction
2
A R T I C L E S
7 mg (79%) of 3 and 19 mg (60%) of 20.23 20: mp 201-202 °C (dec);
faint yellow crystals;1H NMR (CDCl3, 400 MHz) δ 7.15-7.23 (m,
12H), 7.26-7.35 (m, 18H); 31P NMR (CDCl3, 162 MHz) δ 17.9 [d,
30: mp 111-112 °C; faint yellow crystals (CH2Cl2/MeOH); 1H NMR
(CDCl3, 200 MHz) δ 1.20 (s, 18H), 6.63 (s, 2H); 13C NMR (CDCl3,
50 MHz) δ 31.5, 36.8, 120.7, 164.2; UV/Vis (hexane) λmax (ꢀ) 212
(11200, shoulder), 232 (7650, sh), 302 nm (135, sh); MS (EI) m/z 292
(M+). Anal. Calcd for C12H20S4 C, 49.27; H, 6.89. Found: C, 49.21;
H, 6.91.
Thermolysis of 30 in the Presence of Norbornene. A solution of
30 (20 mg, 0.069 mmol) and norbornene (132 mg, 1.4 mmol) in xylene
(5 mL) was heated at reflux for 42 h. The solvent was removed under
reduced pressure. The residue was purified by silica gel column
chromatography and then by HPLC to give 9.7 mg (48%) of 30, 5.4
mg (40%) of 10, and 5.5 mg (42%) of 21.
2
1
1J(195Pt-31P) ) 4366 Hz, J(31P-31P) ) 7 Hz], 18.1 [d, J(195Pt-31P) )
2
3411 Hz, J(31P-31P) ) 7 Hz]; IR (KBr) 1095 (SdO) cm-1
.
Trapping of S3 with Norbornene: Formation of 21. A solution
of a crude mixture of 5a and 5b, prepared from 34 mg (0.12 mol) of
4, and norbornene (175 mg, 1.86 mmol) in CHCl3 (7 mL) was heated
at reflux for 1 h. After the solvent had been removed, the residue was
purified by silica gel column chromatography and then by GPC to give
1
10 mg (82%) of 219,11,24a and 22 mg (83%) of 3. 21: yellow oil; H
NMR (CDCl3, 300 MHz) δ 1.07 (dt, J ) 10.3 Hz, J ) 1.9 Hz, 1H),
1.18-1.27 (m, 2H), 1.71-1.76 (m, 2H), 1.94 (dt, J ) 10.3 Hz, J )
1.9 Hz, 1H), 2.48 (m, 2H), 3.65 (d, J ) 1.9 Hz, 2H).
X-ray Crystallographic Analysis of 5a and 18. The crystal data
were recorded on a Mac Science DIP3000 diffractometer equipped with
a graphite monochromator. Oscillation and nonscreen Weissenberg
photographs were recorded on the imaging plates of the diffractometer
by using Mo-KR radiation (λ ) 0.71073 Å), and the data reduction
was made by MAC DENZO program system. The cell parameters were
determined and refined by using the MAC DENZO for all observed
reflections. The structure was solved by direct methods using SIR-
9744 and refined with full-matrix least-squares (SHELXL-9745) using
all independent reflections. Absorption corrections were done by a
multiscan method (SORTAV46). The non-hydrogen atoms were refined
anisotropically. Crystal data for 5a: C12H20O2S3, Mw 292.48, colorless
cubes, 0.26 × 0.13 × 0.12 mm3, orthorhombic, space group Pbca, a
) 8.6590(14), b ) 11.644(2), c ) 27.952(5) Å, V ) 2789.0(8) Å3, Z
) 8, Dcalcd ) 1.393 g cm-3, µ (Mo-ΚR) ) 0.52 mm-1, 2819 independent
reflections, 155 parameters; R1 ) 0.074 (I > 2σ(I), 1522 reflections),
wR2 ) 0.195 (for all), S ) 1.002; temperature 153 K. Crystal data for
18: C12H20OS4, Mw 308.55, colorless cubes, 0.22 × 0.13 × 0.12 mm3,
triclinic, space group P-1, a ) 8.0020(12), b ) 9.0860(13), c )
11.798(3) Å, R ) 69.516(6), â ) 83.248(6), γ ) 67.72 (2) °, V )
Trapping of S3 with Cyclopentadiene: Formation of 22 and 23.
A solution of a crude mixture of 5a and 5b, prepared from 66 mg
(0.24 mol) of 4, and cyclopentadiene (158 mg, 2.4 mmol) in CHCl3 (5
mL) was heated at reflux for 3 h. The mixture was evaporated and the
residue was purified by silica gel column chromatography and then by
GPC to give 16 mg (80%) of 22,25 6 mg (11%) of 3, and 38 mg (58%)
1
of 23. 22: yellow oil; H NMR (CDCl3, 300 MHz) δ 2.51-2.61 (m,
1H), 3.04-3.15 (m, 1H), 4.76 (m, 1H), 5.21 (m, 1H), 5.63 (m, 1H),
5.85 (m, 1H). 23: mp 150-151 °C; colorless crystals (from CH2Cl2/
1
hexane); H NMR (CDCl3, 200 MHz) δ 1.19 (s, 9H), 1.24 (s, 9H),
1.99-2.12 (m, 1H), 2.45-2.61 (m, 1H), 3.42-3.45 (m, 1H), 3.89-
3.96 (m, 1H), 3.99 (dd, J ) 3.9 Hz, J ) 1.5 Hz, 1H), 4.05 (dd, J ) 4.2
Hz, J ) 1.5 Hz, 1H), 5.57-5.68 (m, 2H); 13C NMR (CDCl3, 50 MHz)
δ 31.6, 32.4, 32.9, 33.7, 34.5, 38.8, 50.8, 68.6, 68.9, 129.4, 132.8, 140.2,
142.8; IR (KBr) 1075 (SdO) cm-1. Anal. Calcd for C17H26OS: C,
73.33; H, 9.41. Found: C, 73.26; H, 9.60.
Trapping of S3 with Norbornadiene: Formation of 25 and 26.
A solution of a crude mixture of 5a and 5b, prepared from 70 mg
(0.25 mol) of 4, and norbornadiene (233 mg, 2.53 mmol) in CHCl3 (5
mL) was heated at reflux for 8 h. The mixture was evaporated and the
residue was purified by silica gel column chromatography and then by
GPC to give 2.7 mg (5%) of 25,24b,43 2.5 mg (5%) of 3, and 44 mg
743.5(2) Å3, Z ) 2, Dcalcd ) 1.128 g cm-3, µ (Mo-ΚR) ) 0.47 mm-1
,
2565 independent reflections, 154 parameters; R1 ) 0.073 (I > 2σ(I),
1301 reflections), wR2 ) 0.198 (for all), S ) 1.001; temperature 298
K.
1
(43%) of 26. 25: yellow oil; H NMR (CDCl3, 300 MHz) δ 1.72 (dt,
Acknowledgment. This work was supported by a Grant-in-
Aid for Scientific Research (No. 16350019) from the Ministry
of Education, Culture, Sports, Science and Technology of Japan.
J ) 9.4 Hz, J ) 1.9 Hz, 1H), 2.48 (d, J ) 9.4 Hz, 1H), 2.91 (m, 2H),
4.06 (d, J ) 1.9 Hz, 2H), 6.38 (t, J ) 1.9 Hz, 2H). 26: mp 211-212
°C; faint yellow crystals (from CH2Cl2/hexane); 1H NMR (CDCl3, 400
MHz) δ 1.25 (s, 18H), 1.67-1.71 (m, 1H, Hf), 1.80-1.83 (m, 1H,
He), 2.51-2.52 (m, 2H, Hc), 2.77-2.78 (m, 2H, Hb), 3.81 (d, J ) 2.0
Hz, 2H, Hd), 4.02 (dd, J ) 2.1 Hz, J ) 2.1 Hz, 2H, Ha) (for hydrogen
labeling, see Scheme 9); 13C NMR (CDCl3, 100.6 MHz) δ 29.7, 32.2,
34.4, 42.8, 45.4, 68.4, 71.3, 140.9; IR (KBr) 1079 (SdO) cm-1; MS
(EI) m/z 400 (M+); HRMS (EI) Calcd for C19H28OS4 (M+): 400.1023,
Found: 400.1023.
6,7-Di-tert-butyl-1,2,3,4-tetrathiocin (30). A mixture of 4 (50 mg,
0.18 mmol) and Lawesson’s reagent (46 mg, 0.11 mol) in toluene (50
mL) was heated at 100 °C for 24 h. The solvent was removed under
reduced pressure and the residue was stirred with hexane (10 mL). The
insoluble materials were removed and the filtrate was evaporated. The
residue was purified successively by silica gel column chromatography,
GPC, and HPLC to give 8 mg (23%) of 10 and 16 mg (31%) of 30.
Supporting Information Available: Temperature-dependent
1H NMR spectra of methine protons of 7 in CDCl3, X-ray
crystallographic data of 5a and 18, and DFT calculation data at
B3LYP/6-311+G(d,p) and B3LYP/6-311+G(3df,2p) levels,
including the vibrational models and frequencies (PDF). This
material is available free of charge via the Internet at
http://pubs.acs.org.
JA047729I
(43) We thank Prof. D. Harpp for providing us the 1H NMR data of 25.
(44) Altomare, A.; Cascarano, G.; Giacovazzo, C.; Guagliardi, A. J. Appl.
Crystallogr. 1993, 26, 343-350.
(45) Sheldrick, G. M. SHELXL-97: Program for Crystal Structure Refinement;
Go¨ttingen University: Germany, 1997.
(46) Blessing, R. H. Acta Crystallogr. 1995, A51, 33-38.
9
J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004 9093