Sulfuranes Lacking Benzoannelation. Sulfuranes and Other Hypervalent Molecules Studied by 17O-NMR

Article abstract of DOI:10.1021/jo9909608

The rates of hydrolysis of benzoannelated vs nonbenzoannelated sulfuranes, viz. 5 vs 3 or 5 vs 4, were compared. Benzoannelation was found to provide very modest kinetic stabilization. Crystal structures of sulfuranes 3 and 4 were obtained and compared with each other, with a dibenzoannelated sulfurane, 17, and with a non-sulfurane analogue of 4. Bond length variations could be understood in the context of simple resonance arguments. ¹⁷O NMR studies of 3-5 showed that this technique was indeed sensitive to sulfurane structure. For example, the chemical shifts of the two carbonyl oxygens of 4 differed by over 20 ppm. Other hypervalent systems, mainly iodinanes, were studied by ¹⁷O NMR. A variety of theoretical methods were surveyed to test how well they could reproduce the geometry of 3. Density functional theory calculation outperformed ab initio geometry optimization at the MP2/3-21G(*) level. Finally, a cis-trans isomerization of the double bond of 11 and one of the two double bonds of 10 was studied.

Full text of DOI:10.1021/jo9909608

8
226
J . Org. Chem. 1999, 64, 8226-8235
Su lfu r a n es La ck in g Ben zoa n n ela tion . Su lfu r a n es a n d Oth er
Hyp er va len t Molecu les Stu d ied by 17O-NMR
Zwei-Chang Ho,1a Peter Livant,* William B. Lott,1b and Thomas R. Webb
Department of Chemistry, Auburn University, Auburn, Alabama 36849-5312
Received J une 15, 1999
The rates of hydrolysis of benzoannelated vs nonbenzoannelated sulfuranes, viz. 5 vs 3 or 5 vs 4,
were compared. Benzoannelation was found to provide very modest kinetic stabilization. Crystal
structures of sulfuranes 3 and 4 were obtained and compared with each other, with a dibenzo-
annelated sulfurane, 17, and with a non-sulfurane analogue of 4. Bond length variations could be
understood in the context of simple resonance arguments. 17O NMR studies of 3-5 showed that
this technique was indeed sensitive to sulfurane structure. For example, the chemical shifts of the
two carbonyl oxygens of 4 differed by over 20 ppm. Other hypervalent systems, mainly iodinanes,
were studied by 17O NMR. A variety of theoretical methods were surveyed to test how well they
could reproduce the geometry of 3. Density functional theory calculation outperformed ab initio
geometry optimization at the MP2/3-21G(*) level. Finally, a cis-trans isomerization of the double
bond of 11 and one of the two double bonds of 10 was studied.
In tr od u ction
Because of all the seminal work in sulfurane chemistry
naturally that the question of the necessity of benzo-
annelation in sulfurane design could be addressed by
measuring the relative kinetic stabilities of the sul-
furanes in the series 3-5.
by J . C. Martin and co-workers, the vast majority of
known sulfuranes embody principles of design which he
enunciated.2 These include the presence of a five-
membered ring spanning equatorial and apical positions
of the trigonal bipyramid, apical groups of high elec-
tronegativity, gem-dialkyl groups on the five-membered
ring, and benzoannelation of the five-membered ring.
This recipe implies generic structures 1 and 2, which
have been used with great success by Martin and his
group in numerous instances. Indeed, recent workers
refer to 2 as simply “the Martin ligand.”3
Exp er im en ta l Section
(
Z,Z)-3,3′-Th iod i(p r op en oic a cid ), 6. To a cooled (0 °C)
solution of 98% meso-dibromosuccinic acid (49.9 g, 177 mmol)
in 114 mL of absolute ethanol was added a solution of 55.00 g
(835 mmol) of 85% KOH pellets in 150 mL of ethanol. The
resulting mixture was refluxed 2 h. The mixture was filtered
while hot and washed twice with boiling ethanol. The filter
cake was dissolved in 120 mL water and acidified at 0 °C by
2 4
slow addition of 25% (v/v) aq H SO to pH < 2. The precipitate
was vacuum filtered, suspended in 150 mL of water, and
heated to reflux for 3 h. A small amount of solid remaining in
the reaction mixture was removed by filtration at room
temperature, and the filtrate was cooled to 10 °C. This was
treated with a solution of 10.0 g (152 mmol) of 85% KOH
pellets in 70 mL of water until pH > 10, followed by gradual
The necessity of benzoannelation came into question
1
7
in our laboratories as an outgrowth of our interest in
O
NMR studies of sulfuranes. Our early attempts to obtain
O NMR spectra of sulfuranes were unsuccessful. This
1
7
failure, we reasoned, may have been due to the sheer
bulk, and consequently the longer rotational correlation
times, of the sulfuranes under study, all of which
contained phenyl rings. Therefore, we considered simply
leaving off the phenyl ring(s) to allow the molecule to
tumble faster. Target sulfuranes 3 and 4 were devised
to help solve our 17O NMR problem. It then followed
addition of 22.3 g (92.9 mmol) Na
stirred under N 1 h at 10 °C and then 12 h at room
temperature. The solution was acidified at 0 °C by addition of
ice-cold 10% (v/v) aq H SO
2
S‚9H
2
O. This mixture was
2
2
4
to pH < 2. The resulting precipitate
was collected by vacuum filtration, washed twice with water,
and dried on the vacuum line, to give 7.94 g 6, 51.5% yield.
4
mp 233.5-235 °C. (lit. 245-247 °C (dec), 244-245 °C) IR
-
1
1
(
d
KBr): 1673, 1592, 1561, 897, 810 cm . H NMR (acetone-
1
3
(
1) (a) Present address: Ta-Ren J unior Pharmacy College, Ping-
6
): 7.591 (d, J ) 10.3 Hz, 1H), 6.041 (d, J ) 10.3 Hz, 1H)
): 167.10 (quat), 150.06 (CH), 116.09 (CH).
,1′-Spir obi(2,1-oxath iol-3-on e), 3. All manipulations were
performed in a N -filled glovebox. To a stirred solution of 1.00
C
Dong City, Taiwan, R.O.C. (b) Present address: Sir Albert Sakzewski
Virus Research Centre, Royal Children’s Hospital, Brisbane, Australia.
NMR (acetone-d
6
1
(2) Hayes, R. A.; Martin, J . C. In Studies in Organic Chemistry
2
(
4
Organic Sulfur Chemistry) Elsevier: New York, 1985; Vol. 19, pp 408-
83.
(
3) Kojima, S.; Tagaki, R.; Akiba, K. J . Am. Chem. Soc. 1997, 119,
(4) (a) Trofimov, B. A.; Vavilova, A. N. Sulfur Lett. 1985, 3, 189-
192. (b) Cain, E. N.; Warrener, R. N. Aust. J . Chem. 1970, 23, 73-82.
5
970-5971.
1
0.1021/jo9909608 CCC: $18.00 © 1999 American Chemical Society
Published on Web 10/07/1999

Sulfuranes Lacking Benzoannelation
J . Org. Chem., Vol. 64, No. 22, 1999 8227
g (5.75 mmol) of 6, 6.10 g (77.2 mmol) of dried pyridine, and
treated gradually with a solution of 542 mg (5.00 mmol) of
tert-butyl hypochlorite in 10 mL of dry chloroform. The
resulting precipitate was vacuum filtered, washed sequentially
with dry pentane and dry chloroform, and dried on the vacuum
line. A 69.1% yield (684 mg) of 4 was obtained, mp 185.5-
2
9.0 g of dried chloroform was added dropwise 680 mg (6.30
mmol) tert-butyl hypochlorite (prepared by the method of
5
Walling and Mintz ). After stirring for 3 min, the mixture was
vacuum filtered. The collected solid was washed twice with
dry pentane and dried on the vacuum line, to give 654 mg of
-
1
1
186.5 °C. IR (KBr): 1715 cm . H NMR (acetone-d
6
): 8.122
3
1
1
, 66% yield. mp 160.5-161 °C. IR (KBr): 3108, 3067, 1707,
(d, J ) 7.2 Hz, 1H), 8.022-7.946 (m, 3H), 7.981 (d, J ) 5.3
-
1 1
13
188, 1080 cm . H NMR (acetone-d
6
): 7.778 (d, J ) 5.2 Hz,
Hz, 1H), 7.149 (d, J ) 5.3 Hz, 1H) C NMR (acetone-d
6
):
1
3
H), 7.194 (d, J ) 5.2 Hz, 1H) C NMR (acetone-d ): 167.17
6
167.43, 150.57, 140.54, 136.49, 134.95, 133.38, 132.19, 129.06,
1
7
17
(
quat), 146.79 (CH), 134.77 (CH). O NMR (59.4 MHz, CH
Cl ): 344 (CdO), 273 (-O-).
X-r a y Str u ctu r e of 3. Sulfurane 3 was recrystallized from
mixture at low temperature under strictly
2
-
2 2
128.88 O NMR (59.4 MHz, CH Cl ): 343 (CdO), 320 (CdO),
2
278 (-O-), 266 (-O-).
X-r a y Str u ctu r e of 4. Sulfurane 4 was recrystallized from
an acetone-CCl mixture at low temperature under strictly
4
anhydrous conditions. A crystal of approximate dimensions
an acetone-CCl
4
anhydrous conditions. A rhomboidal plate of approximate
dimensions 0.50 × 0.22 × 0.10 mm was selected; monoclinic,
0.10 × 0.15 × 0.20 mm was selected; monoclinic, a ) 5.585(4)
a ) 10.839(5) Å, b ) 4.337(2) Å, c ) 14.605(7) Å, â ) 105.41-
Å, b ) 17.216(7) Å, c ) 9.378(7) Å, â ) 91.16(6)°, V ) 901.5-
3
3
(
(
3)°, V ) 662.0(5) Å , Z ) 4. Space group C2/c. 698 reflections
581 independent, Rint ) 2.47%) were collected, of which 499
(10) Å , Z ) 4. Space group P2
1
/c. 1373 reflections (1173
independent, Rint ) 27.81%) were collected, of which 990 were
considered observed, F > 4.0σ(F). Solution and refinement
were carried out as in the case of 3, giving R ) 0.0734 and
wR ) 0.1076.
were considered observed, F > 4.0σ(F). The structure was
solved by direct methods and refined by full-matrix least
squares (on F), giving R ) 0.0416 and wR ) 0.0526. An
isotropic riding model was employed for hydrogens.
(Z)-3-(2-Ca r boxyp h en ylsu lfin yl)-2-p r op en oic Acid , 11.
Sulfurane 4 was hydrolyzed by stirring an acetone solution of
it for 2 d in a vessel open to the air. Removal of solvent and
drying on the vacuum line afforded a quantitative yield of
sulfoxide diacid 11, mp 161-162 °C. Anal. Calcd for
(
Z,Z)-3,3′-Su lfin yld i(p r op en oic a cid ), 10. An acetone
solution of sulfurane 3 was stirred for 1 d in a vessel open to
the air. Removal of solvent and drying on the vacuum line
afforded a quantitative yield of sulfoxide diacid 10, mp 132.5-
33 °C. IR (KBr): 1732, 1684 cm-1
1
10 8 5
C H O S: C, 50.00; H, 3.36; S, 13.35. Found: C, 49.73; H,
1
7
(
.
H NMR (acetone-d
6
):
-
1 1
1
3
3.35; S, 13.34. IR (KBr): 1740, 1671 cm . H NMR (acetone-
): 8.264 (d, J ) 8.0 Hz, 1H), 8.115 (d, J ) 8.0 Hz, 1H), 7.931
t, 1H), 7.700 (t, 1H), 6.715 (d, J ) 10.8 Hz, 1H), 6.338 (d, J )
0.4 Hz, 1H) 13C NMR (acetone-d
): 167.42 (quat), 165.89
quat), 150.21 (quat), 148.04 (CH), 134.43 (CH), 131.67 (CH),
31.20 (CH), 129.28 (quat), 127.51 (CH), 125.27 (CH).
-Mer ca p to-5-m eth ylben zoic Acid , 8. This compound
.056 (d, J ) 10.4 Hz, 1H), 6.500 (d, J ) 10.4 Hz, 1H) C NMR
d
6
acetone-d
6
): 165.20 (quat), 150.76 (CH), 128.12 (CH).
(
P ota ssiu m 2-Mer ca p toben zoa te. To a solution of 30.0 g
1
6
(191 mmol) of 98% 2-mercaptobenzoic acid in 90 mL of ethanol
(
1
at 0 °C was gradually added 12.0 g (182 mmol) of 85% KOH
pellets. After 30 min stirring at 0 °C, the precipitate was
filtered and dried in vacuo. This material was used directly
in the preparation of 7.
2
6
was prepared by the method of Allen and MacKay in 33%
yield after recrystallization from glacial acetic acid, mp 155-
(
Z)-3-(2-Ca r boxyp h en ylth io)-2-p r op en oic Acid , 7. A
1
56 °C (lit.7 155-157 °C). IR (KBr): 1679 cm
-1
.
1
H NMR
solution of 10.0 g (35.5 mmol) of 98% meso-2,3-dibromosuccinic
acid in 20 mL of absolute ethanol was cooled to 0 °C and
treated gradually with a solution of 10.0 g (152 mmol) of 85%
KOH pellets in 30 mL of ethanol. The mixture was refluxed 1
h and hot-filtered, washing the filter cake twice with boiling
ethanol. The filter cake was dissolved in 26 mL of water, cooled
(
(
(
(
acetone-d
6
): 7.892 (s, 1H), 7.367 (d, J ) 5.0 Hz, 1H), 7.251
1
3
d, J ) 5.0 Hz, 1H), 5.066 (s, 1H, SH), 2.328 (s, 3H). C NMR
acetone-d
CH), 132.99 (CH), 131.67 (CH), 126.69 (quat), 20.54 (CH
-(2-Ca r boxyp h en ylth io)-5-m eth ylben zoic Acid , 9. This
6
): 168.24 (quat), 135.93 (quat), 135.21 (quat), 134.36
3
).
2
compound was prepared by a modification of the method of
Protiva.8 A mixture of 8.41 g (50.0 mmol) of 8, 12.5 g (49.4
mmol) of 2-iodobenzoic acid, 15.0 g of 85% KOH (228 mmol),
and 0.505 g (7.95 mmol) of copper powder in 125 mL of water
was refluxed 8 h, cooled to room temperature, and filtered.
The filtrate was acidified with HCl, and the resulting precipi-
tate was filtered and washed twice with water. Recrystalliza-
tion from glacial acetic acid afforded 7.55 g of diacid 9 (53.0%
2 4
to 0 °C, and treated slowly with ice-cold 25% (v/v) aq H SO
,9
until the pH was less than 2. The resulting precipitate was
vacuum-filtered, suspended in 28 mL water, and refluxed 3
h. After cooling to 10 °C, it was treated with 4.37 g (22.8 mmol)
of potassium 2-mercaptobenzoate and stirred 1 h while cold
and then overnight at room temperature. It was filtered, and
the filtrate, cooled to 0 °C, was acidified with ice-cold 10% (v/
2 4
v) aq H SO to pH < 2. The precipitate which formed was
suction-filtered, washed with water, and dried on the vacuum
-
1 1
yield), mp 203-204 °C. IR (KBr): 1699, 1676, 1559 cm . H
NMR (acetone-d ): 7.949 (d, J ) 7.8 Hz, 1H), 7.731 (s, 1H),
6
line, affording 3.55 g (44.6% yield) of diacid 7, mp 184-186
-
1
1
7.432-7.271 (m. 3H), 7.205 (d, J ) 8.0 Hz, 1H), 7.045 (d, J )
°
C. IR (KBr): 1680, 1593, 1582, 1565 cm . H NMR (acetone-
7
1
.8 Hz, 1H), 2.383 (s, 3H) 13C NMR (acetone-d
67.77 (quat), 140.88 (quat), 138.71 (quat), 135.12 (CH), 133.67
6
): 168.04 (quat),
d
6
): 7.943 (dd, J ) 7.8, 1.5 Hz, 1H), 7.69-7.61 (m, 2H), 7.543
(
d, J ) 10.2 Hz, 1H), 7.453 (dt, J ) 7.4, 1.4 Hz, 1H), 6.042 (d,
1
3
(CH), 133.41 (quat), 132.91 (CH), 131.85 (CH), 131.69 (CH),
31.57 (quat), 131.45 (CH), 126.67 (CH), 20.82 (CH ).
,1-Ben zoxa th iol-3-on e-1-sp ir o-1′-(5′-m eth yl-2′,1′-ben -
zoxa th iol-3-on e), 5. This compound was prepared following
the method of Lam and Martin.10 In a N
-filled glovebox, a
J ) 10.2 Hz, 1H) C NMR (acetone-d
6
): 167.61 (quat), 167.16
1
3
(quat), 149.65 (CH), 139.34 (quat), 133.33 (CH), 133.00 (quat),
2
1
32.12 (CH), 131.29 (CH), 128.08 (CH), 115.21 (CH).
X-r a y Str u ctu r e of 7. Crystals were obtained by slow
2
evaporation of a methanol solution of 7, and a diamond-shaped
crystal of approximate dimensions 0.20 × 0.22 × 0.22 mm was
solution of 285 mg (0.990 mmol) of 9 in 2.00 g (25.3 mmol) of
dry pyridine was treated gradually with 108 mg (0.995 mmol)
of tert-butyl hypochlorite. After 5 min, the precipitate was
vacuum filtered, washed with 20 mL of dry pentane, and
vacuum-dried. The solid was stirred 20 min in boiling dried
chosen; monoclinic, a ) 5.273(2) Å, b ) 13.846(5) Å, c ) 13.532-
3
(
4) Å, â ) 97.53(3)°, V ) 979.5(6) Å , Z ) 4. Space group P2
1
/
n. A total of 1530 reflections (1292 independent, Rint ) 9.54%)
were collected, of which 1291 were taken to be observed (F >
6
.0σ(F)). Solution and refinement were carried out as for 3,
(
6) Allen, C. F. H.; MacKay, D. D. Organic Syntheses; Wiley: New
York, 1943; Coll. Vol. 2, pp 580-583.
7) Krollpfeiffer, F.; Schultze, H.; Schlumbohm, E.; Sommermeyer,
giving R and wR of 0.0537 and 0.0623 respectively
,1-Ben zoxa th iol-3-on e-1-sp ir o-1′-(2′,1′-oxa th iol-3′-on e),
. All manipulations were performed in a N -filled glovebox.
2
(
4
2
E. Chem. Ber. 1925, 58, 1654.
A solution of 1.00 g (4.46 mmol) of 7 and 360 mg (4.56 mmol)
of dry pyridine in 45 mL of dry chloroform was stirred and
(8) J ilek, J . O.; Seidlov a´ , V.; Sv a´ tek, E.; Protiva, M. Monatsch. Chem.
1965, 96, 182-207.
(9) Adzima, L. J .; Chiang, C. C.; Paul, I. C.; Martin, J . C. J . Am.
Chem. Soc. 1978, 100, 953-962.
(
5) Mintz, M. J .; Walling, C. Organic Syntheses; Wiley: New York,
(10) Lam, W. Y.; Martin, J . C. J . Am. Chem. Soc. 1981, 103, 120-
1
973; Coll. Vol. 5, pp 184-187.
127.

8
228 J . Org. Chem., Vol. 64, No. 22, 1999
Ho et al.
toluene and hot-filtered, and the filtrate was cooled in the
refrigerator. The resulting solid was collected by vacuum
filtration, 78 mg, 48% yield. mp 264-266 °C. Anal. Calcd for
Sch em e 1
15 10 4
C H O S: C, 62.93; H, 3.52; S, 11.20. Found: C, 63.06; H,
-
1 1
3
d
8
.39; S, 11.31. IR (KBr): 1736, 1722 cm . H NMR (acetone-
): 8.158-8.087 (m, 2H), 7.996-7.923 (m, 4H), 7.766 (d, J )
.4 Hz), 1H), 2.517 (s, 3H) 13C NMR (acetone-d
): 167.41
6
6
(quat), 146,52 (quat), 142.87 (quat), 139.54 (quat), 137.93 (CH),
1
36.49 (CH), 134.96 (CH), 130.72 (quat), 130.63 (quat), 129.82
1
7
3
(CH), 129.58 (CH), 129.05 (CH), 128.84 (CH), 21.14 (CH ) O
2 2
NMR (59.4 MHz, CH Cl ): 325 (CdO), 270 (-O-).
2
-(2-Ca r boxyp h en ylsu lfin yl)-5-m eth ylben zoic Acid , 12.
Sulfurane 5 (100 mg) was dissolved in 50 mL of boiling acetone.
After stirring 4 d in air, the solvent was removed on the rotary
evaporator, and the residue was dried on the vacuum line.
Attempts to determine the mp were frustrated by the dehy-
dration of 12 to sulfurane 5 in the mp capillary. IR (KBr):
-
1 1
1
1
7
1
6
717, 1690 cm . H NMR (acetone-d ): 7.990 (d, J ) 7.6 Hz,
H), 7.875 (d, J ) 8.0 Hz, 1H), 7.787 (s, 1H), 7.743 (t, 1H),
.682 (d, J ) 8.0 Hz, 1H), 7.582 (t, 1H), 7.525 (d, J ) 8.0 Hz,
H), 2.400 (s, 3H) 13C NMR (acetone-d
): 166.86 (quat), 166.73
6
(
1
(
quat), 150.27 (quat), 146.90 (quat), 134.24 (CH), 133.75 (CH),
31.63 (CH), 130.96 (CH), 130.53 (CH), 130.37 (quat), 127.57
CH), 127.28 (CH), 20.98 (CH ).
Ser en d ip itou s Syn th esis of (E,Z)-3,3′-Su lfin yld i(p r o-
3
p en oic a cid ), 13. A vial containing a methanol solution of
Z,Z sulfoxide diacid 10 was covered completely with aluminum
foil, a tiny pinhole was made in the foil to allow very slow
escape of the solvent, and the vial was allowed to stand. After
roughly 50 days, the vial contained a small amount of
methanol and some crystals, which were subsequently shown
1
1
by X-ray crystallography to be 13. IR (KBr): 3057, 1709,
-
1 1
1
1
6
6
664, 1614 cm . H NMR (acetone-d ): 7.830 (d, J ) 15.0 Hz,
H), 6.782 (d, J ) 10.0 Hz, 1H), 6.514 (d, J ) 15.0 Hz, 1H),
.500 (d, J ) 9.8 Hz, 1H).
(
E)-3-(2-Ca r boxyp h en ylsu lfin yl)-2-p r op en oic Acid , 15.
Sulfoxide diacid 11 (32.0 mg, 0.133 mmol) was dissolved in a
mixture of 5 mL of water and 10 mL of methanol and kept 15
d in a closed container. Removal of solvent afforded 31.8 mg
Sch em e 2
-1
of 15, 99.4% yield. mp 219-220 °C. IR (KBr): 1709, 1681 cm
.
1
H NMR (acetone-d
6
): 8.183 (d, J ) 7.75 Hz, 1H), 8.096 (d, J
)
8.0 Hz, 1H), 8.037 (d, J ) 15.0 Hz, 1H), 7.935 (t, 1H), 7.714
1
3
(
t, 1H), 6.555 (d, J ) 15.0 Hz, 1H) C NMR (acetone-d
67.48 (quat), 165.05 (quat), 153.39 (CH), 147.89 (quat), 135.36
CH), 132.00 (CH), 131.55 (CH), 124.75 (CH), 124.31 (CH),
6
):
1
(
1
24.24 (quat).
Kin etics of Hyd r olysis of Su lfu r a n es. Two sulfuranes
were dissolved in an NMR tube in 10:1 (v/v) acetone-d
6
:DMSO-
d
6
. No procedures for drying the solvents were used. Since the
1
HDO peak in the H NMR spectrum was much larger than
any sulfurane signal (typically by a factor of 7 or more), and
2
since that implies that an even larger amount of D O was
present, no extra water was required to obtain an excess
concentration of water. The NMR probe was maintained at
4
0 ( 1 °C by the NMR hardware. Precise timing of data
calculate a rate constant, which was identical within experi-
mental error to that calculated from the growth of product 12.
Two kinetic runs were made in the 3 vs 5 case. Three runs
were made in the 4 vs 5 case.
collection was provided by a short program written locally. An
internal standard (anisole for 4 vs 5; 1,1,2,2-tetrachloroethane
for 3 vs 5) was included. Hydrolyses were followed for roughly
2
.5 half-lives, typically, and first-order treatment of the data
yielded satisfactory linear plots. For 5, the loss of the methyl
signal vs an anisole aromatic signal at 6.90 ppm was followed.
For 4, the loss of the olefinic signal at 7.15 ppm vs the same
anisole signal was followed. In the case of 3 vs 5, both olefinic
signals of 3 were unusable because of overlap with aromatic
signals of other species. Therefore, from the t ) ∞ signal
intensity of product 10 was subtracted the signal (6.41 ppm)
intensity of 10 at each time to give the intensity of 3 at each
time. Similarly the methyl signal intensity of 12 at each time
was subtracted from the methyl signal intensity of 12 at t )
Resu lts
Sulfuranes 3-5 were synthesized as shown in Schemes
1 and 2. Yields of sulfide diacids 6 and 7, 52% and 45%,
respectively, each reflect the overall yield of three steps
12
done in one pot, since propiolic acid was not isolated.
It was found that good temperature control during the
final acidification procedure leading to either 6 or 7 was
essential to avoid cis-trans isomerization; 5-10% of the
undesired trans isomer was produced if acid were added
too quickly.
∞
to give the intensity of 5 at each time. As a check, the loss
of signal intensity of 5 as a function of time was used to
(
11) Webb, T. R.; Ho, Z. C.; Livant, P. Acta Crystallogr. 1995, C51,
3
04-306.
(12) Wilson, C. J .; Wenzke, H. H. J . Am. Chem. Soc. 1935, 57, 1265.

Sulfuranes Lacking Benzoannelation
J . Org. Chem., Vol. 64, No. 22, 1999 8229
through sulfur normal to the O-S-O bond and bisecting
the C-S-C bond angle, so that a ) g, b ) h, and so on.
In preparing a sample of sulfoxide diacid 10 for X-ray
crystallography, it was found that slow evaporation of a
methanol solution of 10 gave crystals of the isomeric 13.
The E,E isomer, 14, was not detected. The isomerization
1
of 10 to 13 was studied by H NMR and a rate constant
of 5.8 ( 0.1 × 10 s-1 was determined. It was found that
sulfoxide acid 11 in methanol underwent an analogous
isomerization at room temperature, with k ) 5.2 ( 0.1
×
-7
F igu r e 1. Structure of 3 as determined by X-ray crystal-
lography. A C axis passes through S, bisecting both the
2
O-S-O bond angle and the C-S-C bond angle.
-
1
10-7 s .
1
The H NMR cis-coupling constant across an endocyclic
carbon-carbon double bond is known to change as a
function of ring size; in particular the cis-coupling in
cyclopentene is 5.1 Hz13 or 5.4 Hz. It was found that
the cis-couplings in 3 and 4 were 5.2 and 5.3 Hz,
respectively, as compared to over 10 Hz for their precur-
sors 6 and 7, and over 10 Hz for their hydrolysis products
14
1
0 and 11. Therefore, in the oxidations leading to 3 or 4,
Samples of 10 protected from light isomerized at the
same rate as did samples of 10 exposed to ambient
daylight. Under conditions sufficient to isomerize 10 or
1, 6 and 7 were unchanged. A sample of 6 in methanol
containing DMSO did not undergo isomerization. A
cyclization to the desired sulfuranes could be quickly
verified by H NMR.
1
Products of hydrolysis of 3, 4, and 5, namely 10, 11,
and 12, were obtained quantitatively by stirring an
acetone solution of the appropriate sulfurane several days
in air. In attempting to determine the mp of 12, dehydra-
tion, presumably to sulfurane 5, was noted. This was
verified on a larger scale by heating 12 to 195-210 °C
for 12 h in a closed flask evacuated to 0.2 Torr. A
quantitative yield of sulfurane 5 was obtained.
1
sample of 10 in methanol containing a trace of HCl gave
1
a product mixture having an extremely complicated H
NMR spectrum; over 40 peaks were noted. A sample of
1
1 containing about 10 mol % KOH in methanol-d
4
-
6
-1
isomerized with a rate constant of 1.0 ( 0.1 × 10
s .
The rate constants for the isomerization of 11 in several
Hydrolysis at 40 °C of solutions containing pairs of the
1
solvents were measured: rate constants (relative to that
sulfuranes (3 and 5, or 4 and 5) was followed by H NMR.
in methanol) were D
2 6 3 6
O:DMSO-d :CD OD:acetone-d 3.9:
Ratios of pseudo-first-order rate constants were: k
.20 ( 0.24; k /k ) 1.79 ( 0.13.
3 5
/k )
1
.5:1.0:0.2. (The rate in acetone solvent was inconve-
8
4
5
niently slow, so the rate constant is an estimate based
on the observation of 25% isomerization after 40 days).
1
7
O NMR spectra of sulfuranes 3, 4, and 5 are reported
in Chart 1, along with 17O data for other hypervalent
molecules and nonhypervalent reference compounds.
Discu ssion
Benzoannelation imparts kinetic stability to sul-
furanes: 3, lacking phenyls, hydrolyzes 8.2 times faster
than 5, having two phenyls. It is salutary to note that
the factor of 8.2 is quite small. By comparison, the “five-
membered ring effect”, another Martin design feature,
produces kinetic stabilization of at least several orders
of magnitude.15
To adequately interpret this subtle rate enhancement,
one should be able to write the rate-determining-step
X-ray crystallographic structure determinations were
carried out on sulfide diacid 7, sulfurane 3, and sulfurane
. Figure 1 shows an ORTEP plot of the structure of 3.
Important geometrical parameters from these determi-
nations are presented in Table 1. In 3, a 2-fold axis passes
(rds), one should have structural data pertinent to all
species involved in the rds, and one should have data
which would permit one to disentangle any solvent effect.
4
(Even though the solvent used was identical for 3 and 5,
it is possible that differences in solvation of 3 and 5 and
of intermediates derived from them might account for a
(
13) Smith, G. V.; Kriloff, H. J . Am. Chem. Soc. 1963, 85, 2016-
2
2
017.
(
14) Laszlo, P.; Schleyer, P. v. R. J . Am. Chem. Soc. 1963, 85, 2017-
(15) (a) Martin, J . C.; Perozzi, E. F. J . Am. Chem. Soc. 1974, 96,
3155-3168. (b) Perozzi, E. F.; Martin, J . C. Ibid. 1972, 94, 5519-5520.
018.

8
230 J . Org. Chem., Vol. 64, No. 22, 1999
Ta ble 1. Com p a r ison of Su lfu r a n e Str u ctu r a l P a r a m eter s fr om X-Ra y Cr ysta llogr a p h y
Ho et al.
Ch a r t 1. 17O NMR Ch em ica l Sh ifts
nontrivial portion of such a small kinetic effect). We do
not profess to have met all these requirements. Never-
theless, we offer the following thoughts.
heterolysis of the axial S-O bond to be the rds, as shown
below for 3. We focus on the so-called equatorial angle,
which is labeled “fl” in Table 1. Structural studies of
sulfonium salts would suggest the C-S-C angle in 16
is smaller than its value in 3.17 Similarly, the C-S-C
angle in the sulfonium carboxylate resulting from het-
On the basis of numerous mechanistic studies of
hydrolyses of sulfuranes structurally similar to those
under discussion here,2 it is reasonable to take the
,16

Sulfuranes Lacking Benzoannelation
J . Org. Chem., Vol. 64, No. 22, 1999 8231
erolysis of 5 would be smaller than the equatorial angle
of 5. That is, the rate-determining step probably involves
a contraction of the equatorial angle. If this is indeed the
case, a “least-motion” argument can explain our kinetic
results for 3 and 5.
It is more reasonable to meet these disparate require-
ments with two different axial ligands rather than to
expect two of the same ligand to do both jobs. If, for
example, X were the ligand which was better suited to
excess electron density and Y were the ligand suited to
stabilization of an adjacent sulfonium ion center, which
is to say resonance structure b were quite important, and
c not so, then, reasoning from b and b′,23 one predicts
that the S-X bond in unsymmetrical sulfurane R
2
SXY
The equatorial angle of 3 is 105.3(2)°. The equatorial
angle of 5 is not known but should be nearly identical to
that of sulfurane 17, which is simply 5 lacking the methyl
group. The X-ray data for 17 are of poor quality,
unfortunately, (R ) 0.146 (observed), R ) 0.172 (all))
owing to difficulty in obtaining a well-shaped crystal.18
The reported equatorial angle for 17 is 107.8°. This is
consistent with well-measured equatorial angles reported
for other spirobicyclic dioxasulfuranes having two equa-
should be longer than the S-X bond in symmetrical
sulfurane R
sulfurane R
in symmetrical sulfurane R
2
SX
SXY should be shorter than the S-Y bond
SY . In the present case, it
2
, and the S-Y bond in unsymmetrical
2
2
2
is clear from examining the S-O bond lengths that in 4
the carboxylate group of the benzoannelated five-
1
9
19
20
torial phenyls (108.1(4)°, 107.6(3)°, 107.9(1)°, 112.4-
2
1
(
1)° ). Even with the worst esd reported for 17, 2.2°, the
equatorial angle of 17 (and, by extension, 5) is larger than
that of 3. The least motion argument posits that 3, having
a smaller equatorial angle contraction to undergo than
does 5, would undergo the rds faster than 5.
The kinetic behavior of the intermediate case, 4, having
one phenyl, does not fall at the midpoint, which would
be hydrolysis 4.1 times faster than 5. Rather, 4 is more
kinetically stable; namely, only 1.8 times the hydrolysis
rate of 5. This is congruent with earlier ab initio calcula-
membered ring (hereinafter the B (for “benzoate”) ring)
is the ligand which assumes the role of accommodating
negative charge and the carboxylate of the nonbenzoan-
nelated five-membered ring (hereinafter the A (for “acry-
late”) ring) is the ligand which is better able to stabilize
the partially positive sulfur. This conclusion may be
restated more simply in terms of resonance structures:
4b is more important than 4c.
2
tions which found sulfuranes R SXY with nonidentical
axial ligands (X, Y) are thermodynamically more stable
than the average of the thermodynamic stabilities of the
2
2
2 2 2 2
two relevant symmetrical sulfuranes, R SX and R SY .
We proceed with caution since calculations of thermody-
namic stability of a reactant may not suffice to explain
its kinetic behavior. The explanation advanced to ratio-
nalize the calculations was that to take energetic advan-
tage of a charge-separated resonance structure, b or c,
one axial ligand should be particularly capable of accom-
modating negative charge, and the other should be
particularly capable of stabilizing an adjacent sulfonium
center, conceivably via resonance structure b′.
An alternative argument is as follows: since a carbon-
carbon double is shorter than a benzene carbon-carbon
bond (e.g., in 4, bond e, 1.321(9) Å, is shorter than bond
k, 1.371(9) Å), the shorter S-O bond in the A ring and
the longer S-O bond in the B ring are merely structural
consequences of carbon-carbon bond lengths in their
respective rings. However, this line of reasoning would
predict a “short” S-O bond in 3, contrary to fact (in 3, e
is 1.309(3) Å but S-O is 1.839(2) Å).
(
16) (a) Kapovits, I.; R a´ bai, J .; Ruff, F.; Kucsman, AÄ .; Tan a´ cs, B.
Tetrahedron 1979, 35, 1875-1881. (b) Vass, E.; Ruff, F.; Kapovits, I.;
R a´ bai, J .; Szab o´ , D. J . Chem. Soc., Perkin Trans. 2 1993, 855-859. (c)
Vass, E.; Ruff, F.; Kapovits, I.; Szab o´ , D.; Kucsman, AÄ . Ibid. 1997,
2
061-2068. (d) AÄ dam, T.; Ruff, F.; Kapovits, I.; Szab o´ , D.; Kucsman,
AÄ . Ibid. 1998, 1269-1275.
17) Perozzi, E. F.; Paul, I. C. In The Chemistry of the Sulphonium
Group; Stirling, C. J . M., Patai, S., Eds.; Wiley: London, 1981; pp 15-
7.
(
7
(
18) K a´ lm a´ n, A.; Sasv a´ ri, K.; Kapovits, I. Acta Crystallogr. 1973,
B29, 355-357.
19) Perozzi, E. F.; Martin, J . C.; Paul, I. C. J . Am. Chem. Soc. 1974,
6, 6735-6744.
20) Adzima, L. J .; Duesler, E. N.; Martin, J . C. J . Org. Chem. 1977,
2, 4001-4005.
21) Lam, W. Y.; Duesler, E. N.; Martin, J . C. J . Am. Chem. Soc.
981, 103, 127-135.
22) Eggers, M. D.; Livant, P. D.; McKee, M. L. J . Mol. Struct.
(
If 4b is more important than 4c, one might expect a
lower carbonyl stretching frequency in the B ring than
9
4
1
(
24
in the A. Unfortunately, only a broad signal at 1715
(
-
1
cm
was observed in the IR spectrum of 4, so this
(
prediction cannot be checked.
(
THEOCHEM) 1989, 186, 69-84.
(
(
The least-motion argument involving contraction of the
equatorial angle, which was suggested as an explanation
of the relative reactivity of 3 and 5, is not invalidated by
23) Livant, P. J . Org. Chem. 1986, 51, 5384-5389.
24) Livant, P.; Martin, J . C. J . Am. Chem. Soc. 1977, 99, 5761-
5
767.

8
232 J . Org. Chem., Vol. 64, No. 22, 1999
Ho et al.
including in the comparison the equatorial angle of 4,
06.8(3)°. This value is between the 105.3(2)° for 3 and
1
the 107.8° for 5’s model, 17, and the reactivity of 4 is
between that of 3 and 5.
17O NMR Sp ectr a . Although our synthesis of nonben-
zoannelated sulfuranes 3 and 4 enabled us to quantify
the importance of benzoannelation in sulfurane design,
the original goal of increasing tumbling rate to produce
observable 17O NMR lines was rendered moot by our
1
7
observation that O NMR spectra which were unobserv-
able in hot acetonitrile or toluene solvent were observable
using room-temperature dry methylene chloride. We were
at a loss to explain the improvement since the viscosities
of acetonitrile and methylene chloride are not so different
shift. Unfortunately, the carbonyl 17O chemical shift of
2
2
6 is unknown, and would be difficult to measure because
6 is thermally unstable. A proper “substitute” for 26 in
this comparison would be a carbonyl compound having
a carbonyl stretching frequency close to that of 26 and
whose carbonyl 1 O chemical shift is known. Benzoyl
2 2 3
CH Cl : 0.393 cP at 30 °C; CH
CN: 0.345 cP at 25 °C),25
but welcomed the good fortune and undertook a study of
various hypervalent systems of interest to us (see Chart
).
17O NMR chemical shifts have a diamagnetic and a
paramagnetic component, with the paramagnetic term
providing the bulk of the variation of chemical shifts
within a given class of compounds. The diamagnetic
term is important when comparing compounds in which
different nuclei are bonded to oxygen, for example water
vs dimethyl ether.
The paramagnetic term has three important compo-
nents, according to the well-known Karplus-Pople equa-
tion: (i) (∆E) term, in which ∆E is an average electronic
excitation energy, sometimes approximated using the
first band in the electronic spectrum, (ii) <r
p-orbital “size” term, and (iii) QAA and QAB (A ) oxygen,
B ) neighboring atom), which, roughly speaking, include
the “charge” on oxygen and the 2p-2p π bond order
between oxygen and its neighbor(s). This final component
has been illustrated27 by comparing the O chemical
shifts of acetone (569 ppm) and protonated acetone (310
ppm). The former has an “undiluted” π bond and is thus
shifted far downfield. The latter may be viewed as a
resonance hybrid of a double-bonded form (chemical shift
like that of acetone) and a single-bonded form (chemical
shift like that of 2-propanol, 38 ppm).28 The shift of 310
ppm is then an approximate average of 569 and 38 ppm.
(
7
-
1 30
chloride (1763 cm
) or benzoyl peroxide (1792 and 1772
-
1
31
17
cm
)
suggest themselves. The O NMR spectrum of
1
benzoyl peroxide (BPO) has not been reported; therefore
we measured it. It consists of a single broad signal (W1/2
≈
2 2
700 Hz) at 340 ppm (CH Cl solvent). The absence of
two signals suggests the operation of some type of
exchange process, conceivably involving impurities. The
spectrum was identical whether BPO was taken directly
from the jar or whether it was freshly recrystallized, so
impurities would seem not to be the culprit. Oxygen
2
6
32
scrambling processes in diacyl peroxides are known, but
occur on a much longer time scale, or not at all. Another
33
-1
possible explanation is that the rotational correlation
times of the two types of oxygens are sufficiently different
that one type is observable, but the other is so broadened
that it is unobservable. For the sake of logical complete-
ness, it should be noted that it is possible (but unlikely)
that BPO’s carbonyl oxygen and peroxide oxygen have
the same chemical shift. In any event, benzoyl peroxide’s
one signal at 340 ppm is a mystery. Therefore, benzoyl
chloride by default may be considered as a substitute for
-
3
>2p, a
17
2
6. Its one signal at 483 ppm is unambiguously due to
17
the carbonyl oxygen. This fits in with the carbonyl
O
signal of 5 (325 ppm) lying between that of a model
compound possessing low π-bond order (18, 271 ppm) and
that of one possessing high π-bond order (PhCOCl, 483
ppm).
Might a similar consideration of π bond orders suffice
to explain the shifts observed for 3, 4, and 5 (and 17)?
Earlier work on carbonyl stretching frequencies in sul-
furanes24 showed quite a significant contribution of
carboxylate anion resonance forms in carboxy sulfuranes
and thus might serve as a guide for π-bond order. For
The carbonyl oxygens of 4 absorb at 343 and 320 ppm.
Our assignment of the 343 ppm signal to the A-ring
carbonyl of 4 and the 320 ppm signal to the B-ring
carbonyl is based on the similarity of these chemical
shifts to those of 3 (344 ppm) and 5 (325 ppm) (or 17,
322 ppm), respectively. Assuming these assignments are
correct, we may state that the A-ring carbonyl has higher
π-bond order than the B-ring carbonyl, which in turn
means that the carboxyl group in the A-ring has less
carboxylate anion character than the carboxyl group in
the B-ring, or, as was stated before, resonance structure
4b is more important than resonance structure 4c.
Therefore, the conclusion based on 17O NMR chemical
shifts is consistent with the conclusion based on axial
bond lengths. It should be noted that a single crystal is
not required for NMR spectroscopy, and the unmistak-
able 20 ppm difference in chemical shifts produced by
-
1
example, the 1724 cm frequency observed for 17 lies
-
1
between the 1609 cm for 18 (full carboxylate anion
-
1
character; low π-bond order) and the 1790 cm for 26
2
9
(
“no” carboxylate anion character; high π-bond order).
1
7
Similarly, the O chemical shift of the carbonyl oxygen
of 5 should lie between that of 18 and 26, if the π-bond
order effect were the dominant effect on the chemical
(
25) Handbook of Chemistry and Physics, 48th ed.; Weast, R. C.,
Ed.; Chemical Rubber Co.: Cleveland, 1967-68; pp F35-F40.
26) Kintzinger, J .-P. In NMR Basic Principles and Progress, Diehl,
(
P., Fluck, E., Kosfeld, R., Eds.; Springer-Verlag: Berlin, 1981; Vol. 17,
pp 1-64.
(
27) Klemperer, W. G. In The Multinuclear Approach To NMR
(30) Overend, J .; Scherer, J . R. Spectrochim. Acta 1960, 16, 773-
782.
Spectroscopy; Lambert, J . B., Riddell, F. G., Eds., D. Reidel: Dordrecht,
Holland, 1983 (ISBN 90-277-1582-3); Chapter 11.
(31) Davidson, W. H. T. J . Chem. Soc. 1951, 2456-2461.
(32) (a) Martin, J . C.; Hargis, J . H. J . Am. Chem. Soc. 1969, 91,
5399-5400. (b) Goldstein, M. J .; Haiby, W. A. J . Am. Chem. Soc. 1974,
96, 7358-7359.
(
28) Christ, H. A.; Diehl, P.; Schneider, H. R.; Dahn, H. Helv. Chim.
Acta 1961, 44, 865-880.
29) Field, L.; Giles, P. M., J r.; Tuleen, D. L. J . Org. Chem. 1971,
6, 623-626.
(
3
(33) Greene, F. D. J . Am. Chem. Soc. 1959, 81, 1503-1506.

Sulfuranes Lacking Benzoannelation
J . Org. Chem., Vol. 64, No. 22, 1999 8233
two carbonyls even as similar as the A-ring carbonyl of
Ab In itio Ca lcu la tion s of Su lfu r a n e Str u ctu r e.
The theoretical description of hypervalent molecules in
general and sulfuranes in particular is a problem of long-
standing interest to theoreticians.38 Accurate calculations
of sulfurane structure have been done most often for
1
7
4
and the B-ring carbonyl of 4 recommends O NMR
spectroscopy as a sensitive tool for investigating other
oxygen-containing hypervalent systems.
Some exploratory results are presented in Chart 1.
Compounds 20-22 are iodinanes (10-I-3 species) which
have at least one carboxyl ligand. Chloroiodinane 20
exhibits a carbonyl oxygen absorption at 328 ppm, which
is quite similar to the carbonyl absorptions of sulfuranes
2
2,39
SF
4
,
a small molecule with an experimental geometry
available for comparison. However, we would argue that
fluorine-containing systems, hypervalent or not, are
inherently atypical due to the extreme electronegativity
of fluorine. Further, “Martin-type” sulfuranes, which
account for a large fraction of structurally well-character-
5
2
and 17. This may be fortuitous. The “ether” oxygen of
0 is shielded relative to those of 5 and 17 by 41-47 ppm.
Part of this may be a bond angle dependence. Because
of the larger covalent radius of iodine, the ether oxygen
of 20 must have a larger bond angle than the ether
oxygen of 5 or 17. This would produce an upfield shift
4
ized sulfuranes, do not resemble SF . Therefore, the
ability of a theoretical method to provide accurate
geometries of hypervalent sulfur species of practical
interest to experimentalists might be more validly as-
(
cf. ether oxygen of γ-butyrolactone, 178.5 ppm vs that
sessed using a test molecule other than SF
X-ray structure of 3 in hand, and noting that the number
of atoms in 3 is fairly small, and that the C symmetry
4
. With the
3
4
of δ-valerolactone, 166.5 ppm). The remainder of the
difference is probably a consequence of the diamagnetic
2
-
-
35
term (e.g. ClO
3
297 ppm vs IO
3
206 ppm). Iodinanes
axis reduces the size of the problem still further, we
propose 3 as a more suitable test molecule with which to
evaluate theoretical methods.
2
1 and 22 each exhibit only one oxygen absorption,
unexpectedly. The peak for 21 is quite broad, but the one
for 22 is sharp, by 17O standards. We cannot at present
offer a good explanation of these spectra; they are
included in the hope that others with greater expertise
will be motivated to propose an explanation. The crystal
structures of iodoarene dicarboxylates usually show
significant “nonbonded” contacts between carbonyl oxy-
gen and iodine, either intra- or intermolecularly.36 It is
therefore not unreasonable to consider the possibility of
degenerate [1,3] sigmatropic shifts of iodine from -O-
to dO as a route to oxygen scrambling. Another possible
route to oxygen scrambling is dissociation to an iodonium
cation carboxylate anion pair followed by I-O bond
formation. However, we have no data to support or refute
Acting on this suggestion, we carried out geometry
optimizations of 3 using a wide variety of methods,
including molecular mechanics, semiempirical, and ab
initio calculations. Our purpose was to survey methods
which could be of practical value in doing calculations
on the larger molecules that experimentalists routinely
synthesize and study. Accordingly, calculations employ-
ing very large basis sets and/or sophisticated correlation
treatments were purposely excluded. All calculations
were carried out using the SPARTAN 4.0 system of
programs.40 J ohnson has pointed out that DFT results
are to an unfortunate extent a function of the details of
implementation employed by the particular program.
SPARTAN 4.0 uses numerical basis sets, with DN
purported to be roughly equivalent to 6-31G and DN* to
41
either of these suggestions.
The 17O chemical shift of chlorosulfurane 24, 115 ppm,
is in a sparsely populated region of the chemical shift
range. It is difficult to find a compound with which to
compare 24. Dimethyl sulfite (-O-) happens to absorb
at 115 ppm too,2 but this may be misleading since the
oxygen in 24 is attached to a tertiary carbon. Compare
the chemical shifts of methanol (-37 ppm) and 2-(2-
phenylsulfinyl)phenyl)-2-propanol (69 ppm), a difference
of over 100 ppm. A better comparison might be tetra-
(
38) The following is not an exhaustive list of leading references:
(
a) Kar, T.; Marcos, E. S. Chem. Phys. Lett. 1992, 192, 14-20. (b) Reed,
A. E.; Weinhold, F. J . Am. Chem. Soc. 1986, 108, 3586-3593. (c) Ponec,
R.; Krack, M.; J ug, K. Theor. Chim. Acta 1996, 93, 165-176. (d)
Musher, J . I. In Conformational Analysis - Scope and Present
Limitations; Chiurdoglu, G., Ed.; Academic: New York, 1971; pp 177-
189. (e) Bartell, L. S.; Gavezzotti, A. J . Mol. Struct. (THEOCHEM)
1983, 91, 331-336. (f) Harcourt, R. D. J . Mol. Struct. (THEOCHEM)
8
1
989, 186, 131-165. (g) Mayer, I. Ibid. 1989, 186, 43-52. (h) AÄ ngy a´ n,
J . G. Ibid. 1989, 186, 61-67. (i) Volatron, F. Ibid. 1989, 186, 167-
184. (j) Cooper, D. L.; Cunningham, T. P.; Gerratt, J .; Karadakov, P.
B.; Raimondi, M. J . Am. Chem. Soc. 1994, 116, 4414-4426. (k) Csonka,
G. I.; Loos, M.; Kucsman, A.; Csizmadia, I. G. J . Mol. Struct.
3
7
methylethylene sulfate (-O- 168.1) or tetramethyl-
ethylene sulfite (-O- 194.7 ppm).37 It is then difficult
to rationalize the “shift” from 168.1 or 194.7 ppm to 115
ppm. A possible explanation may involve a bond angle
change. Comparing ethylene sulfite (-O- 155.5 ppm) to
trimethylene sulfite (-O- 126.1 ppm) to tetramethylene
sulfite (-O- 145.0 ppm) shows that geometry changes
(
THEOCHEM) 1994, 315, 29-35. (l) Budzelaar, P. H. M. J . Org. Chem.
1998, 63, 1131-1137. (m) Harcourt, R. D. Int. J . Quantum Chem. 1996,
6
3
(
0, 553-566. (n) Cioslowski, J .; Mixon, S. T. Inorg. Chem. 1993, 32,
209-3216. (o) H a¨ ser, M. J . Am. Chem. Soc. 1996, 118, 7311-7325.
p) Chen, M. M. L.; Hoffmann, R. J . Am. Chem. Soc. 1976, 98, 1647-
1653. (q) Vetter, R.; Z u¨ licke, L. Chem. Phys. 1986, 101, 201-209. (r)
Kutzelnigg, W. Angew. Chem., Int. Ed. Engl. 1984, 23, 272-295. (s)
Musher, J . I. Ibid. 1969, 8, 54-68. (t) Kar, T.; Marcos, E. S. Chem.
Phys. Lett. 1992, 192, 14-20. (u) Reed, A. E.; Weinhold, F. J . Am.
Chem. Soc. 1986, 108, 3586-3593.
3
7
can sometimes have a large effect. The C-O-S bond
angle in 24 is 115.9(2)°.23 It is also possible to argue the
sulfur to which the oxygen in 24 is bonded is so dissimilar
to either a sulfite sulfur or a sulfate sulfur that any
comparison of sulfurane 24 to either a sulfite or a sulfate
is invalid.
(
39) (a) J ug, K.; Iffert, R. J . Comput. Chem. 1987, 8, 1004-1015.
(
b) Alkorta, I. Theor. Chim. Acta 1994, 89, 1-12. (c) Marsden, C. J .;
Smart, B. A. Aust. J . Chem. 1994, 47, 1431-1440. (d) Cheung, Y.-S.;
Chen, Y.-J .; Ng, C. Y.; Chiu, S.-W.; Li, W.-K. J . Am. Chem. Soc. 1995,
1
17, 9725-9733. (e) Innes, E. A.; Csizmadia, I. G.; Kanada, Y. J . Mol.
Struct. (THEOCHEM) 1989, 186, 1-8. (f) Hay, P. J . J . Am. Chem. Soc.
1977, 99, 1003-1012. (g) Baird, N. C.; Taylor, K. F. Can. J . Chem.
1981, 59, 814-816. (h) Radom, L.; Schaefer, H. F., III. Aust. J . Chem.
1975, 28, 2069-2072. (i) Pietro, W. J .; Francl, M. M.; Hehre, W. J .;
DeFrees, D. J .; Pople, J . A.; Binkley, J . S. J . Am. Chem. Soc. 1982,
104, 5039-5048.
(
34) Boykin, D. W.; Sullins, D. W.; Eisenbrun, E. J . Heterocycles
989, 29, 301-306.
35) (a) Alei, M. J . Chem. Phys. 1965, 43, 2904-2906. (b) Dwek, R.
A.; Luz, Z.; Peller, S.; Shporer, M. J . Am. Chem. Soc. 1971, 93, 77-79.
36) (a) Stergioudis, G. A.; Kokkou, S. C.; Bozopoulos, A. P.;
1
(
(
Rentzeperis, P. J . Acta Crystallogr. 1984, C40, 877-879. (b) Kokkou,
S. C.; Cheer, C. J . Ibid. 1986, C42, 1748-1750. (c) Kokkou, S. C.; Cheer,
C. J . Ibid. 1986, C42, 1159-1161.
(40) Wavefunction, Inc., 18401 Von Karman Ave. #370, Irvine, CA
92715.
(41) J ohnson, B. G. In Modern Density Functional Theory: A Tool
For Chemistry; Seminario, J . M., Politzer, P., Eds. Elsevier: New York,
1995; pp 169-219.
(
37) Hellier, D. G.; Liddy, H. G. Magn. Reson. Chem. 1988, 26, 671-
6
74.

8
234 J . Org. Chem., Vol. 64, No. 22, 1999
Ch a r t 2. Bon d Len gth s of 3 Ca lcu la ted by Va r iou s Meth od s
Ho et al.
Ch a r t 3. Bon d An gles of 3 Ca lcu la ted by Va r iou s Meth od s
6
-31G*. The LSD approximation is used with the Vosko,
the experimental geometry of 3. The results are pre-
sented graphically in Charts 2 and 3. The bond lengths
and angles are named as in Table 1. In both charts, the
left-to-right order of plotting within each group of bars
is identical to the top-to-bottom order in the legend, and
Wilk, and Nusair (VWN) exchange-correlation func-
tional.42
The only criterion upon which the calculations were
judged was how well the calculated geometry agreed with
“deviation” means simply the calculated value minus the
(42) Vosko, S. H.; Wilk, L.; Nusair, M. Can. J . Phys. 1980, 58, 1200.
observed value.

Sulfuranes Lacking Benzoannelation
J . Org. Chem., Vol. 64, No. 22, 1999 8235
The three-center four-electron axial bond is at the
heart of the structure of a trigonal bipyramidal hyper-
valent like 3. This parameter, S-O bond length “a”, is
the left-most bar in each group in Chart 2. Two other
important parameters are the O-S-O angle (the arc of
which passes between the two equatorial substituents),
Sch em e 3
“ag”, and the C-S-C equatorial angle, “fl”, which are,
respectively, the two left-most bars in each group in
Chart 3.
Not surprisingly, molecular mechanics with the Sybyl
force field was quite poor. Bond lengths were not as bad
as with DFT/VWN/DN, but calculated bond angles were
worthless. The calculated axial angle (“ag”) of 180.0° was
unacceptable (cf. 174.9(1)° observed) and the equatorial
angle (“fl”) of 119.88° (cf. 105.3(2)° observed) was worse
than that. Also, it is fair to call the DFT treatment using
the DN basis a crashing failure.
It was worrisome that MNDO/d, although acquitting
itself respectably with regard to bond lengths, erred on
the axial angle (“ag”) so much as to exceed 180° and
render the sulfur slightly saddle-shaped.
not undergo the isomerization. The sulfinyl group must
be intramolecular, as 6 in methanol solvent containing
dimethyl sulfoxide did not undergo the isomerization. The
reaction changes when 10 in methanol is treated with a
trace of acid. A large number of products are formed,
many of which are methyl esters. No attempt was made
to further characterize these products. The effect of base
is to accelerate the isomerization rate constant by a
modest factor of 1.7. A tentative mechanism consistent
with these observations is presented in Scheme 3.
In the case of 11, there is only one double bond which
may isomerize. In 10, there are two, yet only one
isomerizes. This raises the question of why the other
double bond does not isomerize too. A plausible sugges-
tion is that 10 has an internal hydrogen bond between
sulfinyl oxygen and one carboxylic acid, as shown in
Scheme 3. On conversion to 13, this hydrogen bond may
be retained and the lower energy of a trans double bond
results in lower energy overall for 13 relative to 10.
However, for 13 to isomerize to 14, the system must give
up the intramolecular hydrogen bond to gain a trans
double bond. Since the energy lowering of the cis-trans
The simple HF calculations using polarized basis sets,
3
-21G(*) and 6-31G*, were very roughly similar. The
axial length (“a”) was strongly underestimated using
-21G(*), though. When a correlation treatment was
3
included (MP2), it was anticipated that better geometries
would be obtained. While the bond angles were clearly
better, it was surprising to find the bond lengths had
gotten worse. The rms bond length deviations for 3-21G-
(*), 6-31G*, and MP2/3-21G(*) were, respectively, 0.025,
0
.020, and 0.039 Å. However MP2/3-21G(*) did a much
better job on the axial length (“a”), a deviation of 0.0098
Å compared to -0.0498 Å and -0.0293 Å for 3-21G(*)
and 6-31G*, respectively.
Clearly the method of choice, judging from these
results, is DFT using a polarized basis (DFT/VWN/DN*).
The rms bond length deviation was 0.008 Å and the rms
bond angle deviation was 0.90°. Such a result is some-
what unexpected, since the VWN functional has been
shown to perform much more poorly than several other
functionals.43
Dou ble Bon d Isom er iza tion . Sulfoxide diacid 10
isomerizes to 13 on standing in methanol. The double
bond isomerization is very slow, having a half-life of
about 14 days at room temperature. Sulfoxide acid 11
undergoes the same slow isomerization in methanol, with
a half-life of about 15 days. Apparently, the sulfinyl group
is required, as the analogous sulfide acids 6 and 7 did
4
4
isomerization, typically less than 2 kcal/mol, is much
less than the bond energy of a hydrogen bond, roughly 5
kcal/mol,45 the isomerization of 13 to 14 is unfavorable.
Ack n ow led gm en t. We thank Dr. K. P. Sarathy of
the Auburn University NMR Center for a generous
allocation of instrument time. Also, we thank Dr. David
Young for helpful comments concerning the calculations.
(
43) Scheiner, A. C.; Baker, J .; Andzelm, J . W. J . Comput. Chem.
997, 18, 775-795.
44) Akimoto, H.; Sprung, J . L.; Pitts, J . N., J r. J . Am. Chem. Soc.
972, 94, 4850-4855.
45) (a) Hadzi, D.; Rajnvajn, J . J . Chem. Soc., Faraday Trans. 1
1
1
1
Su p p or tin g In for m a tion Ava ila ble: Additional details
of crystallographic studies of 3, 4, and 7, and Cartesian
coordinates and final energies for all calculations. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(
(
973, 69, 151-155 report H-bonds between carboxylic acids and
various sulfoxides having ∆G298 in the range -4.4 to -6.4 kcal/mol (b)
J oesten, M. D.; Schaad, L. J . Hydrogen Bonding; Dekker: New York,
1
974.
J O9909608