Reduction of permanganate by thioanisole: Lewis acid catalysis

Article abstract of DOI:10.1021/jo9914162

The oxidation of sulfides and sulfoxides by permanganate in anhydrous acetone solutions is catalyzed by Lewis acids such as iron(III) chloride, zinc chloride, and mercury(II) chloride. The reaction kinetics unequivocally confirm that the function of these catalysts is to activate the oxidant by forming permanganate/Lewis acid complexes analogous to the protonation of MnO₄/^- by Bronsted acids. A Hammett analysis of the rate constants for the oxidation of a series of substituted thioanisoles gives a negative p value (- 1.11) indicative of an electron deficient transition state. No secondary kinetic isotope effect is observed when the hydrogens of the methyl group are replaced by deuterium. Despite previous observations that sulfoxides are preferentially oxidized in competitive experiments, sulfides are oxidized more rapidly when individual rates are measured. All of these observations are most consistent with a mechanism in which the reductant reacts with the oxidant via initial ligation.

Full text of DOI:10.1021/jo9914162

1008
J . Org. Chem. 2000, 65, 1008-1015
Red u ction of P er m a n ga n a te by Th ioa n isole: Lew is Acid Ca ta lysis
Ning Xie,† Robert A. Binstead,‡ Eric Block,§ W. David Chandler,† Donald G. Lee,*,†
Thomas J . Meyer,‡ and Mohan Thiruvazhi§
Departments of Chemistry, State University of New York at Albany, Albany, New York 12222,
University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599-3290, and
University of Regina, Regina, Saskatchewan, Canada S4S 0A2
Received September 8, 1999
The oxidation of sulfides and sulfoxides by permanganate in anhydrous acetone solutions is catalyzed
by Lewis acids such as iron(III) chloride, zinc chloride, and mercury(II) chloride. The reaction kinetics
unequivocally confirm that the function of these catalysts is to activate the oxidant by forming
permanganate/Lewis acid complexes analogous to the protonation of MnO-4 by Brønsted acids. A
Hammett analysis of the rate constants for the oxidation of a series of substituted thioanisoles
gives a negative F value (-1.11) indicative of an electron deficient transition state. No secondary
kinetic isotope effect is observed when the hydrogens of the methyl group are replaced by deuterium.
Despite previous observations that sulfoxides are preferentially oxidized in competitive experiments,
sulfides are oxidized more rapidly when individual rates are measured. All of these observations
are most consistent with a mechanism in which the reductant reacts with the oxidant via initial
ligation.
In tr od u ction
sulfides under certain conditions,1g,2 complexation pro-
vides a plausible explanation for the observed catalysis.
Alternatively, it is also possible that the catalysis could
be due to complexing of the oxidant by the Lewis acid. A
reaction between permanganate and a Lewis acid would
be analogous to the reaction between permanganate and
Brønsted acids which results in the formation of per-
manganic acid, HMnO4,3 a strong oxidant that is respon-
sible for the acid catalysis which has been observed
during several permanganate oxidations.4 It may also be
noted that formation of a Lewis acid/permanganate
complex has been proposed to account for the catalysis
observed when hydrocarbons are oxidized by permanga-
nate in the presence of boron trifluoride.5 When the
reductant is a hydrocarbon, it is not possible to attribute
the catalysis to formation of a Lewis acid/reductant
complex. A third reaction mechanism that must also be
considered is the possibility that the oxidant and reduc-
tant could form an intermediate which undergoes oxida-
tive decomposition with the assistance of a Lewis acid.
It has been possible to distinguish between these three
mechanisms from a study of the reaction kinetics with
the conclusion that the reaction most likely proceeds by
way of a permanganate/Lewis acid complex. Similar
results were obtained by using two different Lewis acids,
zinc chloride and mercury(II) chloride. The rate constants
Lewis acids are effective catalysts for several useful
oxidative processes including Oppenauer1a and Baeyer-
Villiger1b oxidations, conversion of thiols to disulfides by
reaction with sulfoxides1c or manganese dioxide,1d oxida-
tive rearrangements catalyzed by selenium dioxide,1e the
oxidation of unsaturated alcohols by manganate(VI),1f
and the oxidation of sulfides to sulfones by permangan-
ate.1g,h The latter reaction, oxidation of sulfides and
sulfoxides to sulfones by permanganate in anhydrous
acetone, is possible only if Lewis acids such as ferric
chloride or zinc acetate are present. The essential role
of Lewis acids in these reactions is exemplified by the
observation that thioanisole is converted into the corre-
sponding sulfone in 100% yield when treated with potas-
sium permanganate and ferric chloride in acetone for 30
min at -25 °C, while, in the absence of ferric chloride or
another Lewis acid, only a trace of sulfone is produced
under identical conditions. The purpose of this work is
to investigate the catalytic effect of Lewis acids on these
reactions and their relationship to the reaction mecha-
nism.
Three possibilities exist for Lewis acids to exert a
catalytic effect on the oxidation of sulfides by perman-
ganate. For example, the effect of a Lewis acid could be
-
to complex the sulfide (R2S + FeCl3 f R2S+-FeCl3
)
for
a
series of substituted thioanisoles and the
making it more sulfoxide-like (R2SdO T R2S+-O-). Since
it is known that sulfoxides are more rapidly oxidized than
(2) (a) Ogura, K.; Suzuku, M.; Tsuchihashi, G. Bull. Chem. Soc. J pn.
1980, 53, 1414. (b) Poje, M.; Balenovic, K. Tetrahedron Lett. 1978, 1231.
(c) Block, E.; Corey, E. R.; Penn, R. E.; Renken, T. L.; Sherwin, P. E.;
Bock, H.; Hirabayshi, T.; Mohmand, S.; Solouki, B. J . Am. Chem. Soc.
1982, 104, 3119. (d) Rajagopal, S.; Sivasubramanian, G.; Suthakaran,
R.; Srinivasan, C. Proc. Indian Acad. Sci. 1991, 103, 637. (e) Fatiadi,
A. J . Synthesis 1989, 85.
(3) Frigerio, N. A. J . Am. Chem. Soc. 1969, 91, 6200.
(4) (a) Stewart, R. In Oxidation in Organic Chemistry, Part A.;
Wiberg, K. B., Ed., Academic Press: New York, 1965; p 35. (b) Verma,
R. S.; Reddy, M. J .; Shastry, V. R. J . Chem. Soc., Perkin Trans. 2 1976,
469. (c) Aitken, R. A.; Mesher, S. T. E.; Rosse, B. M. Synthesis 1997,
787.
* To whom correspondence should be addressed.
† University of Regina.
‡ University of North Carolina at Chapel Hill.
§ State University of New York at Albany.
(1) (a) Almeida, M. L. S.; Kocovsky, P.; Backvall, J .-E. J . Org. Chem.
1996, 61, 6587. (b) Kotsuki, H.; Arimura, K.; Araki, T.; Shinohara, T.
Synlett 1999, 462. (c) AbuOmar, M. M.; Khan, S. I. Inorg. Chem. 1998,
37, 4979. (d) Firouzabadi, H.; Abbassi, M.; Karimi, B. Synth. Commun.
1999, 29, 2527. (e) Shafer, C. M.; Morse, D. I.; Molinski, T. F.
Tetrahedron 1996, 52, 14475. (f) Firouzabadi, H.; Karimi, B.; Abbassi,
M. J . Chem. Res. S 1999, 236. (g) Block, E.; DeOrazio, R.; Thiruvazhi,
M. J . Org. Chem. 1994, 59, 2273. (h) Thiruvazhi, M. Ph.D. Dissertation,
SUNY-Albany, Albany, New York, 1993.
(5) Lau, T.-C.; Wu, Z.-B.; Bai, Z.-L.; Mak, C.-K. J . Chem Soc. Dalton
Trans. 1995, 695.
10.1021/jo9914162 CCC: $19.00 © 2000 American Chemical Society
Published on Web 01/28/2000

Reduction of Permanganate by Thioanisole
J . Org. Chem., Vol. 65, No. 4, 2000 1009
corresponding phenyl methyl sulfoxides are also com-
pared, and an attempt has been made to understand the
relationship of the relative rates of oxidation of sulfides
and sulfoxides with previously reported product studies.
Exp er im en ta l Section
Ma ter ia ls. Acetone, used as the solvent in all kinetic
experiments, was Fisher HPLC grade. Potassium permanga-
nate was BDH AnalaR grade. Anhydrous zinc chloride was
obtained from Aldrich and stored in a desiccator.
Thioanisole, p-methoxythioanisole, p-methylthioanisole, p-
fluorothioanisole, and p-chlorothioanisole, received (from Al-
drich) as liquids with 97% or greater purity, were carefully
distilled under reduced pressure and a middle portion of the
distillates collected. GC or HPLC analysis indicated purities
of 99.5 to 99.9% for the final products. The observed boiling
points were as follows: thioanisole, 114-115 °C/21 Torr;
p-methoxythioanisole, 139.0-139.5 °C/23 Torr; p-methylthio-
anisole, 114-115 °C/21 Torr; p-fluorothioanisole, 89-90 °C/
31 Torr; and p-chlorothioanisole, 131.0-131.5 °C/23 Torr.
p-Nitrothioanisole, obtained commercially, was recrystallized
from ethanol three times and purified by column chromatog-
raphy (silica gel; eluent: chloroform/ethyl acetate 1/1). GC
analysis indicated a purity of 99.2%, mp 71.8-72.7 °C (lit.6
70.5-71.5 °C).
F igu r e 1. Attempted pseudo-first-order plot.
reduced pressure leaving a yellow liquid (5.21 g, 74%). This
crude product was purified by careful distillation under
reduced pressure to give a colorless liquid (2.14 g, 30%), bp
113.5-114.4 °C /21 Torr, 1H NMR (CDCl3) δ: 2.40 (s, 3H), 7.10
(m, 2H), 7.20 (m, 2H); IR (cm-1) 3000 (m), 2050 (m), 1491 (s),
1589 (m).
Kin etic Mea su r em en ts. It was found experimentally that
ferric chloride, the catalyst used for reactions at -25 °C,1
promoted a rapid reduction of permanganate by the solvent
at ambient temperatures, thereby preventing its use in kinetic
studies. However, zinc chloride and mercury(II) chloride were
found to catalyze the reduction of permanganate by thioanisole
at convenient rates without promoting a corresponding rapid
reaction between the oxidant and the solvent. The kinetic
study of the reduction of permanganate by sulfides and
sulfoxides was therefore conducted using anhydrous acetone
solutions containing different concentrations of zinc chloride
or mercury(II) chloride. In a typical experiment, a solution of
zinc chloride in anhydrous acetone was sealed in a 50 mL
Erlenmeyer flask and immersed in a constant-temperature
bath for 1 h. While the zinc chloride solution was being
thermostated, a permanganate solution was prepared by
placing a few milligrams of KMnO4 in a 50 mL Erlenmeyer
flask and adding 40 mL of anhydrous acetone. After swirling
for ca. 30 s, the supernant was transferred, by use of a
disposable pipet, to another flask suspended in the constant-
temperature bath. The flask was stoppered and sealed to
prevent contact with moisture in the air. An aliquot of the zinc
chloride solution (2.0 mL) was then transferred to a 10 mm
cuvette, and a stock solution of thioanisole in anhydrous
acetone (0.10 mL) was added with a microliter syringe. The
cuvette was placed in the thermostated cell compartment of
an HP8452 diode array spectrophotometer, and after a few
minutes permanganate solution (0.50 mL) was added by use
of a microliter syringe. The cuvette was quickly inverted
several times to ensure good mixing and spectra were collected
every five seconds until the reaction was complete. The initial
concentration of permanganate was approximately 2 × 10-4
M for all experiments, while the concentration of zinc chloride
was varied from 5 × 10-4 to 7 × 10-3 M, and that of thioanisole
from 1 × 10-2 to 0.12 M. The reaction rates were determined
by monitoring the decrease in absorbance at 528 nm and
application of standard procedures for kinetic studies.
The corresponding sulfoxides were prepared by oxidation
of the sulfides with hydrogen peroxide in acetone. In a typical
procedure, a flask containing the sulfide (0.12 mol) dissolved
in 125 mL of acetone was cooled in an ice-water bath, and
30% H2O2 (20 g) was added. The mixture was stirred until TLC
showed no sulfide remained. The product was then extracted
with chloroform (4 × 50 mL), and the extracts were dried over
anhydrous MgSO4. Solvent was removed by flash evaporation,
and the remaining mixture of sulfone and sulfoxide was
separated by column chromatography (silica gel; eluent:
chloroform/ethyl acetate 1/1). Final purification was by re-
crystallization or distillation under reduced pressure. The
products were characterized by use of nuclear magnetic
resonance and infrared spectroscopy as indicated below: Phen-
yl methyl sulfoxide (obtained in 24% yield), bp 161.2-162.0
°C/21 Torr; 1H NMR (CDCl3) δ: 2.70 (s, 3H), 7.50 (m, 3H), 7.65
(d, 2H); IR (cm-1) 1045 (s), 1481 (s), 1582 (s), 2900 (s), 3010
(s). p-Methoxyphenyl methyl sulfoxide (obtained in 30% yield)
mp 42.1-43.0 °C (lit.7 43 °C), 1H NMR (CDCl3) δ: 2.67 (s, 3H),
3.95 (s, 3H), 7.00 (d, 2H), 7.80 (d, 2H); IR (cm-1) 1048 (s), 1255
(s), 1497 (s), 1595 (s), 2837-3200 (s). Methyl p-methylphenyl
sulfoxide (obtained commercially), mp 41.5-42.5 °C (lit.7 42-
1
43 °C), H NMR (CDCl3) δ: 2.38 (s, 3H), 2.65 (s, 3H), 7.28 (d,
2H), 7.68 (d, 2H): IR (cm-1) 1056 (s), 1448 (m), 3000 (s).
p-Fluorophenyl methyl sulfoxide (obtained in 68% yield), bp
135-136 °C /21 Torr, 1H NMR (CDCl3) δ: 2.65 (s, 3H), 7.15
(d, 2H), 7.60 (d, 2H): IR (cm-1) 1049 (s), 1087 (s), 1493 (s),
1590 (m), 2900-3100 (w).
Methyl-d3 p-methylphenyl sulfide was prepared from the
reaction of p-thiocresol with dimethyl sulfate-d6. A 100 mL
flask fitted with a reflux condenser, a dropping funnel, and a
stirrer was immersed in an ice water bath. p-Thiocresol (6.47
g, 00.051 mol), tetrabutylammonium bisulfate (0.340 g, 0.001
mol), dichloromethane (20 mL), and 50% NaOH (10.4 g) were
added. The solution was stirred vigorously for 30 min, and
(CD3)2SO4 (7.93 g, 0.06 mol) was added dropwise over 1 h. After
stirring at room temperature for an additional 18 h, TLC
analysis indicated that no p-thiocresol remained. Concentrated
NH4OH (2 mL) was added, and after stirring for another 30
min, water (30 mL) was added to dissolve precipitated white
solids. The aqueous layer was separated and extracted with
dichloromethane (5 × 20 mL). The organic layers were washed
with 33% NaOH solution (2 × 20 mL) and then with water
until neutral to pH paper. After being dried over anhydrous
magnesium sulfate for 12 h, the solvent was evaporated under
Th e Ra te La w
Under pseudo-first-order conditions, the approximate
linearity of plots of ln(A - Af) vs time suggests that the
reaction is first order with respect to permanganate. The
slight upward curvature of many of these plots (see, for
example, Figure 1) indicates that the reaction is subject
to autocatalysis, a well-known phenomenon when MnO2
is one of the products.8 Possible errors caused by this
curvature were avoided by the use of initial reaction rates
to determine the rate constants. To prevent the introduc-
tion of bias, the initial rates were determined by fitting
(6) Lindberg, B. J .; Schroder, B. Acta Chem. Scand. 1970, 24, 3089.
(7) Cerniani, A.; Modena, G. Gazz. Chim. Ital. 1953, 89, 843.

1010 J . Org. Chem., Vol. 65, No. 4, 2000
Xie et al.
F igu r e 3. Plot indicating the order with respect to zinc
chloride. Slope ) 1.12 ( 0.02; r ) 0.999.
F igu r e 2. Plot indicating the order with respect to perman-
ganate. Slope ) (1.59 ( 0.02) × 10-3 s-1; r ) 0.999.
the absorbance/time data to a standard curve, defined
by eq 1, as described in previous publications.9
A ) a0 + a1t + a2t2 + a3t3 + a4t4 + a5t5
(1)
Differentiation with respect to time of the equation
obtained from computer modeling of the data, and setting
t ) 0, gives the initial rate, (-dA/dt)i, which has been
designated as Ri. Because permanganate is slowly re-
duced by acetone in the absence of other reductants, it
is necessary to correct the observed initial rate for the
“blank rate” which has been designated as Rb. This
correction did not reduce the initial rate by more than
10% under any conditions. A plot of Ri - Rb (the initial
rate) vs Ai (the initial absorbance) is linear (Figure 2),
confirming that the reaction is first order with respect
to permanganate.
F igu r e 4. Variation of initial rate with reductant concentra-
tion.
The use of pseudo-first-order conditions and initial
reaction rates minimizes the possibility of contributions
to the observed rate by consecutive oxidations in which
the initial product, a sulfoxide, is converted into the
ultimate product, a sulfone.1,10
When the concentration of thioanisole was held in a
constant large excess while the concentration of zinc
chloride was varied, a plot of ln(Ri - Rb)/Ai vs ln[ZnCl2]
was linear with a slope of 1.1 ( 0.1 (Figure 3). The rate
law is therefore also first order with respect to the Lewis
acid.
When the initial rate was plotted against the concen-
tration of thioanisole, the order with respect to thioani-
sole was observed to change from approximately unity
at low concentrations to approximately zero at high
concentrations as in Figure 4. A similar observation was
made when mercury(II) chloride was used as the catalyst.
Rea ction P r od u cts. Sequential scanning (Figure 5)
indicated that the disappearance of permanganate coin-
cides with the appearance of a product with a spectrum
that conforms to Rayleigh’s law for light scattering by
F igu r e 5. Sequential scans (top to bottom at 528 nm) for the
reduction of permanganate (1.62 × 10-4 M) by thioanisole (0.12
M) in anhydrous acetone containing zinc chloride (5 × 10-4
M).
colloidal particles, ln A ) -4λ + C.11 The plot in Figure
6 indicates a good correspondence between the predic-
tions of this law and the final spectrum obtained.
Furthermore, if the solutions are allowed to stand
undisturbed for 24 h, a brown solid, characteristic of
manganese dioxide, slowly precipitates.12 It appears,
therefore, that the initial observable product is colloidal
MnO2.
(8) (a) Mata-Perez, F.; Perez-Benito, J . Z. Phys. Chem. Neue Folge
1984, 141, 213. (b) Lee, D. G.; Perez-Benito, J . F. J . Org. Chem. 1988,
53, 5725. (c) Perez-Benito, J . F.; Arias, C. Int. J . Chem. Kinet. 1991,
23, 717. (d) Insausti, M. J .; Alvarez-Macho, M. P.; Mata-Perez, F.
Collect. Czech. Chem. Commun. 1992, 57, 2331. (e) Insausti, M. J .;
Mata-Perez, F.; Alverez-Macho, M. P. Int. J . Chem. Kinet. 1995, 27,
507.
(9) (a) Lee, D. G.; Lee, E. J .; Chandler, W. D. J . Org. Chem. 1985,
50, 4306. (b) Chandler, W. D.; Lee, E. J .; Lee, D. G. J . Chem. Ed. 1987,
64, 878.
(11) (a) Laidler, K. J .; Meiser, J . H. Physical Chemistry, 2nd ed.;
Houghton Mifflin: Toronto, 1995; p 858. (b) Perez-Benito, J . F.; Arias,
C. J . Colloid Interface Sci. 1992, 152, 70. (c) Perez-Benito, J . F.; Lee,
D. G. Can J . Chem. 1985, 63, 3545.
(10) (a) May, B. L.; Yee, H.; Lee, D. G. Can. J . Chem. 1994, 72, 2249.
(b) Ogura, K.; Suzuki, M.; Tsuchihashi, G. Bull. Chem. Soc. J pn. 1980,
53, 1414.
(12) (a) Perez-Benito, J . F.; Brillas, E.; Pouplana, R. Inorg. Chem.
1989, 28, 390. (b) Mata-Perez, F.; Perez-Benito, J . F. Can. J . Chem.
1985, 63, 988.

Reduction of Permanganate by Thioanisole
J . Org. Chem., Vol. 65, No. 4, 2000 1011
The rate constant, k1, was found to be independent of
the structure of the reductant (Table 1), as would be
expected for a reaction between permanganate and zinc
chloride as in Scheme 2. The effect of substitutents on
the values of kobs in Table 1 must be associated with k2
because k1 and k-1 would not be dependent on the
structure of the reductant.
Discu ssion
The primary purpose of this investigation is to under-
stand the nature of the observed catalysis that occurs
when sulfides are oxidized by permanganate in the
presence of Lewis acids. Possible reaction mechanisms
involving the complexation of zinc chloride with the
reductant, the oxidant, or an intermediate have been
summarized in Schemes 1, 2, and 3, respectively. The
experimental results, which give a rate law that is first
order in oxidant, first order in Lewis acid, and of an order
between unity and zero in reductant, clearly indicate that
the function of the catalyst is to complex permanganate
as in Scheme 2.
F igu r e 6. Rayleigh plot for the reaction product (MnO2). Slope
) 3.86 ( 0.04; r ) 0.999.
Sch em e 1. Rea ction Sequ en ce a n d Ra te La w
Wh en th e In ter m ed ia te is a
Su lfid e/Lew is Acid Com p lex
R2S + ZnCl2 yk1z R2SZnCl2
k-1
F igu r e 7. Reciprocal plot in verification of eq 3. Slope ) 0.536
( 0.008 M s; Intercept ) 29.1 ( 0.6 s; r ) 0.999.
k2
R2SZnCl2 + MnO-4
8 R2SO + MnO-3 + ZnCl2
slow
Manganate(V) ion, MnO-3 , formed when an oxygen
atom is transferred from permanganate to a sulfide would
be a very reactive entity under these conditions; it is
known to be stable only in highly alkaline solutions (6-
10 M NaOH).13 If formed in a neutral solution in the
presence of an oxidizable solvent such as acetone, reduc-
tion to manganese dioxide, the observed product, would
be instantaneous.14
R2SO + MnO-4 fast8 R2SO2 + MnO3-
MnO-3 solvent8 MnO2
fast
k1k2[R2S][MnO-4 ][ZnCl2]
rate )
k-1 + k2[MnO-4 ]
Sch em e 2. Rea ction Sequ en ce a n d Ra te La w
Wh en th e In ter m ed ia te is a
Resu lts
P er m a n ga n a te/Lew is Acid Com p lex
The experimentally determined rate law (first order
in oxidant, first order in Lewis acid, and of variable order
between unity and zero in reductant) is of the form
expected if zinc chloride acts as a catalyst by combining
with permanganate prior to the redox step, as in Scheme
2.
MnO-4 + ZnCl2 yk1z MnO4ZnCl-2
k-1
k2
MnO4ZnCl-2 + R2S
8 R2SO + MnO3ZnCl-2
slow
The derived rate law may be restated as in eqs 2 and
3 where kobs ) k1k2/k-1. From eq 3 it is apparent that
values for k1 and kobs can be obtained from the intercepts
and slopes, respectively, of plots of Ai/(Ri - Rb) vs 1/[RSR′]
as in Figure 7.
R2SO + MnO4ZnCl-2 fast8 R2SO2 + MnO3ZnCl-2
MnO3ZnCl2- solvent8 MnO2ZnCl2
fast
k1k2[R2S][MnO-4 ][ZnCl2]
rate )
(Ri - Rb)/Ai ) kobs[RSR′][ZnCl2]/(1 + k2/k-1[RSR′])
k-1 + k2[R2S]
(2)
To understand the catalytic nature of zinc chloride in
these reactions, it is necessary to begin with a consider-
ation of the chemical and electronic structure of the
oxidant. A tetrahedral ion with four equivalent, highly
polarized oxos,15 permanganate ion may be represented
by the use of resonance structures as in Scheme 4.
Ai/(Ri - Rb) ) 1/(kobs[ZnCl2] [RSR′]) + 1/(k1[ZnCl2])
(3)
(13) Arndt, D. Manganese Compounds as Oxidizing Agents in
Organic Chemistry; Open Court: LaSalle, Illinois, 1981; p 172.
(14) (a) Zahonyi-Budo, E.; Simandi, L. I. Inorg. Chim. Acta 1996,
248, 81. (b) Zahonyi-Budo, E.; Simandi, L. I. Inorg. Chim. Acta 1992,
191, 1. (c) J aky, M. Acta Chim. Hung. 1992, 129, 487. (d) Simandi, L.
I.; J aky, M.; Schelly, Z. A. J . Am. Chem. Soc. 1984 106, 6866.
(15) (a) Sharpless, K. B.; Teranishi, A. Y.; Blackvall, J . E. J . Am.
Chem. Soc. 1977, 99, 3120. (b) Howard, W. A.; Parkin, G. J . Am. Chem.
Soc. 1994, 116, 606.

1012 J . Org. Chem., Vol. 65, No. 4, 2000
Xie et al.
Ta ble 1. Ra te Con sta n ts for th e Oxid a tion of Th ioa n isoles a n d Meth yl P h en yl Su lfoxid es by P er m a n ga n a te in
An h yd r ou s Aceton e Solu tion s
catalyst
[ZnCl2] ) 1.16 × 10-3
M
[HgCl2] ) 0.120 M
reductant
k1 (M-1 s-1
)
kobs (M-2 s-1
)
k1 (M-1 s-1
)
kobs (M-2 s-1
)
p-methoxythioanisole
p-methylthioanisole
thioanisole
p-fluorothioanisole
p-chlorothioanisole
25.9 ( 0.1
26.1 ( 0.1
26.9 ( 0.4
27.8 ( 0.4
26.6 ( 0.1
23.5 ( 0.2
27.9 ( 0.4
25.9 ( 1,2
26.9 ( 1.0
26.2 ( 3.1
3630 ( 50
2480 ( 111
1580 ( 20
1430 ( 20
951 ( 6
4.68 ( 0.07
4.85 ( 0.05
4.83 ( 0.06
4.72 ( 0.14
317 ( 12
252 ( 7
203 ( 5
183 ( 12
p-nitrothioanisole
228 ( 5
methyl p-methoxyphenyl sulfoxide
methyl p-methylphenyl sulfoxide
methyl phenyl sulfoxide
methyl p-fluorophenyl sulfoxide
334 ( 5
147 ( 4
97 ( 1
51 ( 1
Sch em e 3. Rea ction Sequ en ce a n d Ra te La w
Wh en th e In ter m ed ia te is a
P er m a n ga n a te/Su lfid e Com p lex
R2S + MnO-4 yk1z R2SMnO-4
k-1
k2
R2SMnO-4 + ZnCl2
8 R2SO + MnO3ZnCl-2
slow
R2SO + MnO-4 fast8 R2SO2 + MnO3-
MnO3ZnCl2- solvent8 MnO2ZnCl2
F igu r e 8. Hammett plot for the oxidation of substituted
thioanisoles by permanganate in anhydrous acetone solutions
containing zinc chloride as a Lewis acid catalyst. [ZnCl2] )
1.16 × 10-3 M. T ) 23.0 ( 0.1 °C. Slope ) -1.11. r ) 0.999.
fast
MnO-3 solvent8 MnO2
fast
the positive charge on manganese, making it more
susceptible to nucleophilic attack by the sulfide. A similar
effect may account for the high reactivities of zinc and
copper permanganates which have previously been de-
scribed in the literature.19
The validity of the reaction sequence in Scheme 2 was
confirmed by the observation that similar kinetics were
obtained when mercury(II) chloride was used as the
Lewis acid catalyst. Because of the proclivity of mercury-
(II) to complex sulfides, its use would provide the best
opportunity to observe a reaction sequence such as the
one depicted in Scheme 1 where the intermediate is a
Lewis acid/sulfide complex. The magnitude of the rate
constants, k1 and kobs, were decreased (Table 1) when
mercury(II) chloride was used as the catalyst, but an
identical rate law was observed.
k1k2[R2S][MnO-4 ][ZnCl2]
rate )
k-1 + k2[ZnCl2]
Molecular orbital calculations using the Hartree-
Fock-Slater Discrete Variational Method (HFS-DVM),
reported by Ziegler, Rauk, and Baerends, indicate that
the manganese atom in permanganate has a d-orbital
population close to Mn2+ in which two electrons have
been removed from the 4s orbital.16 Symmetry adapted
cluster (SAC) and SAC-configuration interaction (SAC-
CI) calculations have indicated a somewhat smaller
charge (+1.04) on manganese with, of course, a charge
of -2.04 shared by the oxygens.17 These calculations
confirm the assumption that the metal oxos are polarized
as indicated in Scheme 4.15
A quantitative examination of the way in which sub-
stituents affect reaction rates can provide an improved
understanding of the mechanism of a reaction. For the
oxidation of substituted thioanisoles, the Hammett plot
is observed to have a negative slope indicating a transi-
tion state with decreased electron density. The observa-
tion that a better correlation is observed when σ substit-
uent values are used (Figure 8) as compared to σ+
substituent values (Figure 9) indicates that the transition
state is not stabilized by direct resonance with substit-
uents in the para position.
Reaction of a Lewis acid, such as zinc chloride, with
permanganate must involve formation of a coordinate
covalent bond using electrons from the permanganate
HOMO, which has primarily oxygen p-orbital character.
Assuming, according to the most recent calculations, that
the structure of permanganate is best represented by
resonance structure 2, the reaction with zinc chloride
could be depicted as in eq 4.18
(18) An alternative representation of the permanganate/zinc chloride
complex would be
The overall effect of this process would be to increase
(16) Ziegler, T.; Rauk, A.; Baerends, E. J . Chem. Phys. 1976, 16,
209.
(17) J itsuhiro, S.; Nakai, H.; Hada, M.; Nakatsuji, H. J . Chem. Phys.
1994, 101, 1029.
(19) (a) Wolfe, S.; Ingold, C. F. J . Am. Chem. Soc. 1983, 105, 7755.
(b) Lee, D. G.; Noureldin, N. A. J . Am. Chem. Soc. 1983, 105, 3188.

Reduction of Permanganate by Thioanisole
J . Org. Chem., Vol. 65, No. 4, 2000 1013
a single electron transfer, it must therefore occur as a
fast step prior to the rate-determining step for the
reaction.28
An alternative approach to an understanding of the
mechanism for sulfide oxidation by permanganate is to
consider the rate-limiting step to be ligand formation; i.e.,
the attachment of the sulfide to manganese by a coordi-
nate covalent bond as in eq 8.29
F igu r e 9. Hammett plot for the oxidation of substituted
thoanisoles by permanganate in anhydrous acetone solutions
containing zinc chloride as a Lewis acid catalyst. [ZnCl2] )
1.16 × 10-3 M. T ) 23.0 ( 0.1 °C. Slope ) -0.79. r ) 0.973.
The product of this reaction, 5, would be the interme-
diate that most closely resembles the transition state.
The development of a positive charge on sulfur as the
reaction progresses is consistent with the observed nega-
tive slope of the Hammett plot (Figure 8), and the
inability of this intermediate to delocalize the charge via
resonance with substituents on the aromatic ring agrees
with the observation that a better correlation is observed
with σ-values as compared to σ+-values (Figures 8 and
9). This intermediate, 5, would also be analogous to the
agostic complex that has recently been proposed as an
intermediate in the reaction of hydrocarbons with per-
manganate.30 The absence of a secondary isotope effect
when the methyl hydrogens are replaced by deuterium
is also consistent with this intermediate because hyper-
conjugation of the type depicted in Scheme 5 would not
be possible. This intermediate could also be easily
converted into products as indicated in eqs 9-11.
The lack of correlation with σ+-values indicates that a
single electron transfer (SET) is not likely the rate-
limiting step of the reaction. The product of such a
reaction (eq 5) would be a radical cation which could be
stabilized by resonance with substituents such as p-
methoxy and p-fluoro (eqs 6 and 7).20 Other chemical,21
electrochemical,22 and enzyme-catalyzed reactions23 that
involve rate-limiting single electron transfers from sulfur
to yield radical cation intermediates, are known to give
better Hammett correlations when σ+ substituent con-
stants are used. If the transition state resembles a radical
cation, as predicted by the Hammond postulate24 for a
SET mechanism, a better correlation should have been
observed with σ+-values.
The lack of an observable secondary kinetic isotope
effect when the hydrogens on the carbon adjacent to the
sulfur are replaced by deuterium is also an indication
that the rate-limiting step is not a single electron
transfer. The reactions of thioanisole with other oxidants
that have been proposed to proceed via SET mechanisms
show substantial secondary isotope effects when the
methyl hydrogens are replaced by deuterium.25 In those
cases it is believed that the secondary isotope effects are
related to the ability of the methyl group to stabilize the
radical cation by hyperconjugation.26,27 This effect, which
arises from the possibility for stronger orbital overlap
between a sulfur p-orbital containing a single electron
and an adjacent C-H σ-bond as compared to an adjacent
C-D σ-bond, can be illustrated by use of resonance
structures as in Scheme 5. If the reaction is initiated by
It has been assumed, as proposed in the literature,29a
that the first stable product of this reaction sequence is
a sulfoxide which would subsequently be oxidized to a
sulfone, the observed product of the reaction. Perman-
(23) Oae, S.; Watanbe, Y.; Fujimori, K. Tetrahedron Lett. 1982, 23,
1189.
(24) March, J . Advanced Organic Chemistry, 4th ed.; Wiley: New
York, 1992; p 215.
(25) Acquayne, J . H.; Muller, J . G.; Takeuchi, K. J . Inorg. Chem.
1993, 32, 160.
(26) Carroll, F. A. Perspectives on Structure and Mechanism in
Organic Chemistry; Brooks/Cole: Toronto, 1998; p 262.
(27) Hess, R. A.; Hengge, A. C.; Cleland, W. W. J . Am. Chem. Soc.
1998, 120, 2703.
(20) Aplin, J . T.; Bauld, N. L. J . Chem. Soc., Perkin Trans. 2 1997,
853.
(21) Ganesan, T. K.; Rajagopal, S.; Bharathy, J . B.; Sheriff, A. I. M.
J . Org. Chem. 1998, 63. 21.
(22) (a) Bauld, N. L.; Aplin, J . T.; Yueh, W.; Loinaz, A. J . Am. Chem.
Soc. 1997, 119, 11381. (b) Bauld, N. L.; Aplin, J . T.; Yueh, W.; Endo,
S.; Loving, A. J . Phys. Org. Chem. 1998, 11, 157.
(28) (a) Kochi, J . K. Angew. Chem., Int. Ed. Engl. 1988, 27, 1227.
(b) Bockmann, T. M.; Hubig, S. M.; Kochi, J . K. J . Am Chem. Soc. 1998,
120, 2826.
(29) (a) Bohra, A.; Sharma, P. K.; Banerji, K. K. J . Org. Chem. 1997,
62, 3562. (b) Lee, D. G.; Chen, T. J . Org. Chem. 1991, 56, 5346.
(30) Collman, J . P.; Chien, A. S.; Eberspacher, T. A.; Brauman, J .
I. J . Am. Chem. Soc. 1998, 120, 425.

1014 J . Org. Chem., Vol. 65, No. 4, 2000
Sch em e 4. Reson a n ce Str u ctu r es Dep ictin g P er m a n ga n a te Ion
Xie et al.
Sch em e 5. Reson a n ce Str u ctu r es Th a t Illu str a te th e Sta biliza tion by Hyp er con ju ga tion of th e Ra d ica l
Ca tion F or m ed by a Sin gle Electr on Tr a n sfer fr om Th ioa n isole
Sch em e 6. Com p etitive Oxid a tion of Su lfid es a n d
Su lfoxid es by P er m a n ga n a te in An h yd r ou s
Aceton e Solu tion s Con ta in in g Zin c Ch lor id e a s a
Ca ta lyst
that the electron density at sulfur has decreased in the
transition state and with the experimentally determined
rate law. Both mechanisms also accommodate product
formation directly. On the basis of the available kinetic
data it is, therefore, not possible to distinguish between
them.
Holm33 has suggested a mechanism similar to the one
depicted in eq 8 for the transfer of oxygen from a
molybdenum(VI) oxo complex to triethyl- and triphen-
ylphosphine. The reaction is described as a nucleophilic
attack on the ModO bond by interaction with its vacant
π* orbital.34 Since π* orbitals have the same shape and
signs as the permanagnate 2e molecular orbital, the
reactions appear to be mechanistically similar.
The asymmetric oxidation of thioanisole,35 and other
sulfides, by (oxo)(salen)manganese(V) chloride is also
consistent with the idea that attack by a nucleophile on
manganese oxos involves some interaction with the metal
center. It is not likely that direct interaction of a sulfur
p-orbital and an orbital localized on oxygen would lead
to an enantioselective reaction.
Since a similar rate law and substitutent effects are
observed it seems reasonable to expect that a similar
mechanism must pertain for the oxidation of sulfoxides.
Because the sulfur-oxygen double bond is quite polar
(R2SdO T R2S+-O-), a mechanism analogous to eq 8
would likely occur with ligand formation through oxygen
rather than sulfur as in eq 14.
ganate would concurrently be reduced to trioxomanga-
nate(V), MnO3-, as in eqs 9 and 10. Such manganate(V)
compounds are known to be extremely vigorous oxidants
which would be rapidly reduced by the solvent, acetone,
to give the observed product, manganese dioxide as in
eq 11.14 Since the product of the initial reaction, a
sulfoxide, would be readily oxidized to the corresponding
sulfone,10 it is likely that some of the permanganate
would be reduced by secondary oxidation, as in eq 10.
However, the use of initial reaction rates ensures that
the measured rate constants are for the reaction depicted
in eq 8; no sulfoxide is present initially.
It has also been suggested that the initial interaction
of the oxidant and reductant could actually be between
sulfur and one of the permanganate oxygens.31 Nucleo-
philic attack at an electron rich oxygen may seem
improbable; however, electrostatic repulsions between
two electron-rich atoms could be reduced if electrons were
simultaneously transferred from the oxo into vacant
manganese d-orbitals as depicted in eq 12. Although this
process has been referred to in the literature as an “SN2
type mechanism”, it is more properly viewed as a remote
nucleophilic attack on the metal. The resulting interme-
diate, 6, would be easily converted into products as
indicated in eq 13. A similar mechanism has been
proposed for the oxidation of sulfides by ruthenium(IV)
oxo complexes.32
(14)
(12)
(13)
A mechanism analogous to the one proposed in eq 12
would, however, likely involve an initial interaction
(31) Banerji, K. K. Tetrahedron 1988, 44, 2969.
(32) Roecher, L.; Dobson, J . C.; Vining, W. J .; Meyer, T. J . Inorg.
Chem. 1987, 26, 779.
(33) (a) Schultz, B. E.; Gheller, S. F.; Muetterties, M. C.; Soctt, M.
J .; Holm, R. H. J . Am. Chem. Soc. 1993, 115, 2714. (b) Holm, R. H.
Chem. Rev. 1987, 87, 1401.
(34) Holm, R. H.; Berg, J . M. Acc. Chem. Res. 1986, 19, 363.
(35) (a) Palucki, M.; Hanson, P.; J acobsen, E. N. Tetrahedron Lett.
1992, 33, 7111. (b) Chellamani, A.; Alhaji, N. I.; Rajagopal, S.; Sevvel,
R.; Srinivasan, C. Tetrahedron 1995, 51, 12677.
Both of the mechanisms depicted in eqs 8 and 12 are
consistent with the Hammett analysis which indicates

Reduction of Permanganate by Thioanisole
J . Org. Chem., Vol. 65, No. 4, 2000 1015
between sulfur and a manganese oxo as in eq 15.
fact that sulfoxides react slower than sulfides when
individual rates are measured, suggests a mechanism in
which the sulfoxide undergoes ligation more completely,
but converts to products more slowly. With reference to
Scheme 6, Kso must be much larger than Ks, while ks is
larger than kso. This description of the mechanism is
consistent with all of the experimental evidence including
both kinetic and product studies. It is also consistent with
the knowledge that sulfoxides are, in general, much
better ligands for simple metal cations than sulfides. It
therefore appears that the mechanisms presented in eqs
8 and 14 are most consistent with the experimental
evidence.
Con clu sion s
1. A kinetic study has shown conclusively that the
catalytic effect of Lewis acids on the oxidation of sulfides
and sulfoxides by permanganate is due to complexation
of the oxidant with the catalyst prior to reaction with
the reductant.
2. Substituent and isotope effect studies indicate that
a single electron transfer mechanism is not likely for the
oxidation of sulfides by permanganate.
3. Permanganate reacts faster with sulfides than
sulfoxides when individual rates are measured, but in
competitive reactions, sulfoxides react preferentially.
4. The mechanisms summarized in eqs 8 and 14 and
in Scheme 6 are consistent with all of the available
experimental evidence.
The negative Hammet F value observed for the oxida-
tion of sulfoxides is consistent with both of the mecha-
nisms presented in eqs 14 and 15, as is the observation
that a slightly better correlation is obtained with σ rather
than σ+ values; i.e., F ) -2.18, r ) 0.97 and F+ ) -0.85,
r ) 0.96. A distinction between these two mechanisms
on the basis of kinetics alone is therefore not possible.
However, evidence from product studies can be of as-
sistance since it is known from competitive studies that
permanganate oxidizes sulfoxides to sulfones in the
presence of sulfides1a,10 as indicated by the examples in
eqs 16 and 17. Such preferential oxidations, despite the
Ack n ow led gm en t. N.X. and D.G. L. are pleased to
acknowledge financial assistance from the Natural
Sciences and Engineering Research Council of Canada
(NSERC). E.B. acknowledges support from the National
Science Foundation. D.G.L. also gratefully acknowl-
edges the hospitality of the Department of Chemistry
at the University of North Carolina, Chapel Hill, where
this work was initiated.
J O9914162