An investigation by means of correlation analysis into the mechanisms of oxidation of aryl methyl sulfides and sulfoxides by dimethyldioxirane in various solvents

Article abstract of DOI:10.1039/b714707d

Relative rate constants have been measured for the oxidation of aryl methyl sulfides and sulfoxides by dimethyldioxirane in acetone, in mixtures of acetone with aprotic co-solvents of both higher and lower relative permittivity, and in aqueous acetone mixtures. Correlation analyses of the effects of substituents in the different solvents show that, with one exception, reactions take place via a single step mechanism in which the formation of the new SO bond and the elimination of acetone occur concertedly. The exception was oxidation of the sulfides in aqueous acetone containing the highest proportion of water of those studied (20% v/v). Here, the behaviour of the reaction is consistent with a two-step mechanism in which the oxidant reversibly attacks the sulfide to form an open-chain sulfonium betaine that subsequently fragments to sulfoxide and acetone. There is no evidence for the participation of an intermediate dioxathietane as has been found in the case of sulfide oxidations by (trifluoromethyl)methyldioxirane in CH₂Cl₂ and similar aprotic solvents. It is not justified to generalise a mechanism involving a betaine, with or without a derived dioxathietane, to the reactions of dimethyldioxirane in acetone. This journal is The Royal Society of Chemistry.

Full text of DOI:10.1039/b714707d

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
An investigation by means of correlation analysis into the mechanisms of
oxidation of aryl methyl sulfides and sulfoxides by dimethyldioxirane in
various solvents†
Peter Hanson,* Ramon A. A. J. Hendrickx‡ and John R. Lindsay Smith*
Received 24th September 2007, Accepted 5th December 2007
First published as an Advance Article on the web 11th January 2008
DOI: 10.1039/b714707d
Relative rate constants have been measured for the oxidation of aryl methyl sulfides and sulfoxides by
dimethyldioxirane in acetone, in mixtures of acetone with aprotic co-solvents of both higher and lower
relative permittivity, and in aqueous acetone mixtures. Correlation analyses of the effects of
substituents in the different solvents show that, with one exception, reactions take place via a single step
mechanism in which the formation of the new SO bond and the elimination of acetone occur
concertedly. The exception was oxidation of the sulfides in aqueous acetone containing the highest
proportion of water of those studied (20% v/v). Here, the behaviour of the reaction is consistent with a
two-step mechanism in which the oxidant reversibly attacks the sulfide to form an open-chain
sulfonium betaine that subsequently fragments to sulfoxide and acetone. There is no evidence for the
participation of an intermediate dioxathietane as has been found in the case of sulfide oxidations by
(trifluoromethyl)methyldioxirane in CH2Cl2 and similar aprotic solvents. It is not justified to generalise
a mechanism involving a betaine, with or without a derived dioxathietane, to the reactions of
dimethyldioxirane in acetone.
from one to two.5 These observations were explained in terms of
Introduction
a cyclic transition state with the participation of a protic solvent
Numerous oxidations of organosulfur compounds by hydroper-
molecule allowing proton transfer without the separation of full
oxidic reagents have been studied over the past five decades.1–13
ionic charges (Scheme 1b); in the absence of a protic solvent,
a second molecule of the hydroperoxide may fulfil this role.3
Entries 1–8 of Table 1 illustrate typical activation parameters for
oxidation of sulfides to sulfoxides, i.e. low DH‡ and relatively large
negative DS‡ values. The oxidation of sulfoxides to sulfones is
generally more difficult, and entries 9–11 of Table 1 indicate that
changes in both activation parameters contribute to a higher free
energy of activation for these substrates.
Table 2 gives Hammett reaction constants, q, for a selection of
oxidations of sulfides and sulfoxides by hydroperoxidic oxidants.
The small negative q values found for both types of substrate
are consistent with a modest demand for electron density from
sulfur in the rate-determining transition states. In general, the q
values found for sulfoxides are smaller than those found for sulfides
(note especially entries 5 and 6 where the oxidant and reaction
medium are the same), which indicates that these oxidations fail
to follow the reactivity–selectivity principle: oxidations of the more
reactive sulfides are more selective than those of the less reactive
sulfoxides. Thus although both oxidations result in the formation
of a new S–O bond and their general mechanisms are believed to
be comparable, there must be subtler differences between the two
types of transition state than a simple application of the Hammond
postulate would suggest.
The early expectation was that the oxidation of sulfides would
involve the formation of a sulfonium intermediate by SN2 dis-
placement of nucleophilic S upon the peroxo bond (Scheme 1a);1
however, inconsistencies with this assumption became apparent.
For example, reaction rates are strongly dependent neither on the
relative permittivity of the solvent2 nor on the ionic strength3 as
would be expected for a reaction involving a large separation of
charge; on the other hand, reaction rates are strongly dependent
on the protic nature of the solvent4 and, if an aprotic solvent
is used, the order of reaction in the hydroperoxide may increase
Scheme 1
Dimethyldioxirane, 1, is a three-membered cyclic peroxide
which has been used for the oxidation of various substrates
including sulfides and sulfoxides.14,15 The mechanistic evidence
indicates that, in general, the oxidant is electrophilic. For example,
in the epoxidation of alkenes, 1-(phenyl)propene16 is 103 times
more reactive than ethyl trans-3-(phenyl)propenoate17 and, for the
Department of Chemistry, University of York, Heslington, York, YO10 5DD,
UK. E-mail: ph8@york.ac.uk, jrls1@york.ac.uk
† Electronic supplementary information (ESI) available: Substituent con-
stants; relative permittivities; evaluation of f X and f H for various Dq. See
DOI: 10.1039/b714707d
‡ Present address: Department DMPK & Bio-analysis, AstraZeneca, AB
221 87 Lund, Sweden. E-mail: ramon.hendrickx@astrazeneca.com
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 745–761 | 745
©

View Article Online
Table 1 Activation parameters for the oxidation of selected sulfides, R1SR2, and sulfoxides, R1S(O)R2, by hydroperoxidic reagents, XOOH
Entry
R1
R2
X
Solvent
DH‡/kJ mol−1
DS‡/J mol−1 K−1
Ref.
Sulfides
1
2
3
4
5
6
7
8
(CH2)2–O–(CH2)2
(CH2)2–O–(CH2)2
(CH2)2–O–(CH2)2
(CH2)2–O–(CH2)2
H
H
t-Bu
t-Bu
PhC(O)
MeC(O)
P(O)(OH)2
S(O)2OH
H2O
MeOH
H2O
MeOH
CHCl3
MeCO2H
MeCN, 40% v/v H2O
H2O, 0.5% v/v MeCN
54.6
60.9
56.7
57.5
35.7
43.5
47.3
17.4
−113.4
−121.8
−113.4
−142.8
−121.8
−96.2
3
3
3
3
2
6
7
8
4-O2N–Ph
4-O2N–Ph
HO2CCH2
Me
Ph
Ph
Ph
−123.0
−124.0
Me
Sulfoxides
9
10
11
Ph
Ph
Ph
Me
HO2CCH2
Me
PhC(O)
MeC(O)
P(O)(OH)2
H2O
MeCO2H
MeCO2H, 50% v/v H2O
58.2
59.0
56.3
−78.7
−105.0
−89.3
9
6
10
Table 2 Hammett reaction constants, q, for reactions of organosulfur compounds with hydroperoxidic oxidants
Entry
Substrate
Oxidant
Solvent
q
Ref.
1
2
3
4
5
6
7
8
ArSMe
ArSPh
ArSMe
ArSMe
ArSCH2CO2H
ArS(O)CH2CO2H
ArS(O)Me
ArS(O)Ph
H2O2/HClO4
H2O2/HClO4
3-Cl-PhCO3H
H3PO5
MeCO3H
MeCO3H
PhCO3H
EtOH (6% v/v H2O)
EtOH (6% v/v H2O)
MeOH
MeCN (40% v/v H2O)
MeCO2H
MeCO2H
Dioxan (50% v/v H2O)
Dioxan (50% v/v H2O)
−1.13
−0.98
−0.90
−1.09
−1.07
−0.49
−0.64
−0.54
11
11
12
7
6
6
13
13
PhCO3H
16
17
=
=
substituted series ArCH CH2 and trans-ArCH CHCO2Et
reacting in Me2CO, negative reaction constants, q, of −0.90 and
−1.53, respectively, were obtained. Although sulfides have been
reported by Ballistreri and co-workers18 to be ca. 30 times more
reactive than sulfoxides, similar q values were reported for the
oxidation of ArSMe and ArS(O)Me by Murray and co-workers19
(−0.77 and −0.76, respectively, in Me2CO); thus, unlike the
alkenes and, as was the case with the hydroperoxidic oxidants, the
organosulfur compounds do not follow the reactivity–selectivity
principle on reaction with 1. The relative reactivity of sulfides
and sulfoxides in reaction with 1, the size and sign of their
reaction constants and the similarity of these with those of
hydroperoxidic oxidants (Table 2) imply the oxidative sequence
sulfide → sulfoxide → sulfone has mechanisms in which O-transfer
is concerted with elimination of Me2CO in both steps (Scheme 2a).
Recently, Asensio and co-workers20 found that (trifluo-
romethyl)methyldioxirane, 2, a more electrophilic oxidant than
1 on account of the fluorination, oxidises a range of sulfides
directly to sulfones even in the presence of an excess of sulfide.
A mechanism was proposed (Scheme 2b) in which the sulfide
reacts with 2 to form a sulfonium betaine, 3, that rapidly cyclises
to a dioxathietane (cyclic sulfurane) 4, faster than it eliminates
MeCOCF3 (i.e. k2 > k3); 4 then reacts with 2 faster than does
the initial sulfide (i.e. k4 > k1) to produce sulfone plus two
molecules of MeCOCF3. The conclusions regarding the nature
and behaviour of 4 were based on an investigation of the effects of
variation in substitution, solvent composition and temperature on
the sulfone/sulfoxide ratio, supported by convincing isotopic (2H
and 18O) tracer studies. It was concluded that, although sulfoxide-
to-sulfone oxidation is a known reaction of 2 (reaction 5, k5 in
Scheme 2b), for which Ballistreri et al.18 had obtained q = −0.34
Scheme 2
in CHCl3, it is unimportant in comparison to the oxidation of 4
in the conversion of sulfides into sulfones, though no detail of the
latter was given.
It was further suggested that oxidation of sulfides by 1 also
follows a path analogous to Scheme 2b but with reactions 3 and
5 now superseding reactions 2 and 4. This suggestion implies a
change in the nature of the rate- determining step (reaction 1, k1
in Scheme 2b) from that illustrated in Scheme 2a: it implies that
the oxidation of sulfides by 1 involves the formation of an entity
with a full sulfonium charge (the betaine) despite the low reaction
constant found by Murray19 and the similarity with oxidation by
746 | Org. Biomol. Chem., 2008, 6, 745–761
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
Table 3 Relative rate constants for the oxidation of aryl methyl sulfides
and sulfoxides by dimethyldioxirane in acetone at 293 K
hydroperoxides where low q values are taken to indicate transition
states in which only partial charge separation occurs.
The purpose of this paper is to report a fuller investigation than
hitherto into the effects of substituents on the oxidation of aryl
methyl sulfides and sulfoxides by 1, and into the dependence of
such effects on the nature of the reaction medium, with a view
to clarifying the reaction mechanism and to understanding the
differences between the behaviour of sulfides and sulfoxides.
a
a
Aryl substituent, X
(kX/kH)sulfide
(kX/kH)sulfoxide
4-OMe
4-Me
4-Cl
4-C(O)Me
4-CN
4-NO2
H
3-OMe
3-Cl
1.581 0.091
1.262 0.069
0.659 0.068
0.385 0.036
0.322 0.018
0.243 0.026b
1.000 0.000
0.811 0.087
0.574 0.071
0.496 0.035
0.371 0.018
0.282 0.032
1.475 0.144
1.317 0.067
0.701 0.072
0.384 0.036
0.297 0.034
0.215 0.019b
1.000 0.000
0.679 0.082
0.502 0.058
0.383 0.050
0.333 0.022
0.300 0.038
Results and discussion
3-CF3
3-CN
3-NO2
(i) Oxidations in acetone solution
The previous investigation by Murray and co-workers,19 into
the effects of ring-substituents on the oxidations of aryl methyl
sulfides and sulfoxides by 1 in acetone, had involved only a few
4-substituted substrates. We therefore decided first to extend these
series and to include 3-substituted members. Oxidations were
carried out at 293 K in which (with two exceptions) the substituted
aryl methyl sulfides and sulfoxides, in turn, were in competition
with the corresponding unsubstituted parent compound for a
limited amount of 1. (The exceptions were, in each series, the
slowly reacting 4-nitrophenyl compound, which was made to
compete against the corresponding 4-cyano analogue). On the
assumption that the kinetics have a first-order dependence on
each reactant, in each experiment a 15–20-fold excess of both
competing substrates over the oxidant was used, hence ensuring
pseudo-first-order conditions for the oxidant with respect to each.
For each series the yields of reaction products at completion of
reaction were quantified by GC (See Experimental). Here our
method differed from that of Murray and co-workers,19 who had
measured substrate depletions of up to 25%. Given the much
smaller depletion of substrates in our case (5–7%), we felt safe
in the assumption that the molar ratio of the products at the end
of reaction approximates the ratio of the rates at which they were
produced and hence, when the initial molar ratio of the competing
substrates, r0, was taken into account, the relative rate constant,
kX/kH [eqn (1)]:
a Uncertainties are the 95% confidence intervals. b Measured as
(k4-NO /k4-CN) × (k4-CN/kH).
2
the rR scale, giving the best fit of the data as judged by the various
statistics tabulated.
Entry 1 of Table 4 presents details of the Hammett correlation
for the oxidation of the full data set of 3-and 4-substituted sulfides.
The squared correlation coefficient, R2, indicates that 99.4% of the
variation in log(kX/kH)sulfide is explained by the simple linear model;
the correlation is highly significant as indicated by the high value
of F and the low value of Fsignif, and precise as indicated by the low
standard error of the estimate, s; the fit of the model to the data as
expressed by Exner’s statistic,25 w, is ‘good’. Most importantly,
the reaction constant, q, is very close to that of Murray and
co-workers,19,26 confirming their result. Entries 2 and 3 are the
Hammett correlations of the separate para (4-substituted) and
meta (3-substituted) subsets of experimental data. As with the full
set, the intercepts are insignificant. Comparison of the respective
values of s and w for the subsets shows that the data of the para
subset are less scattered than those of the meta subset although
their reaction constants, q, do not differ significantly.
Entries 4 and 5 present dual parameter correlations of
log(kX/kH)sulfide for the separate subsets. The use of two explanatory
variables does not lead to a statistically improved account of
the data in either case but, interestingly, neither shows its best
correlation to involve rI with rBRA, which is expected since these
constants, in principle, represent the separate inductive and
mesomeric ‘components’ of the Hammett substituent constants.24
For the para subset (entry 4), the best dual parameter correlation
is given by rI with r0R with both variables statistically significant
at the 0.01 level. If this observation has chemical significance,
it suggests that, in the observed reactivity, the sulfur centre is
insulated from the direct mesomeric effects of substituents.27
However, the observation could be due to experimental error, as
OMe is the only substituent of those in the subset for which the
value differs between the rBRA and rR0 scales.24 If k4-OMe/kH were
to be 1.65 rather than 1.58, i.e. still within the 95% confidence
interval (see Table 3), the best correlation would involve rI with
rBRA. For the meta subset (entry 5), the best correlation is given by
rI with r−R, with significance levels of 0.01 and 0.03, respectively.
The deterioration in statistical quality between entries 3 and 5, as
judged by the values of and s and w, is not very marked and, of
the five substituents in the subset, only Cl has the same value in
both the r−R and rBRA scales; it therefore seems unlikely that this
ꢀ
[Product(X)]
[Product(H)]
−d[Substrate(X)] −d[Substrate(H)]
=
=
dt
dt
kX[Substrate(X)]0[1]
kH[Substrate(H)]0[1]
kX
= r
(1)
0 kH
The combined yields of products corresponded closely (95–
98%) to the amount of oxidant consumed, showing that side-
reactions such as self-degradation of the oxidant21 were not
significant. Product analysis also showed that, in the case of sulfide
oxidation, no sulfone was produced. Table 3 presents the relative
rate constants determined and Table 4 a correlation analysis of
the dependence of the logarithms of these constants upon various
substituent constants. Simple correlations of log(kX/kH) by single
substituent constants (rm, rp, r+p )22 (r0m)23 are given together with
dual parameter correlations which use the constants rI and rR of
Taft and co-workers.24 There are 4 scales of rR falling in the order
r−R → r0R → rBRA → rR+ for decreasing −M character and increasing
+M character (see ESI, Section S1†). We report results only for
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 745–761 | 747
©

View Article Online
observation is accidental. As the 3-substituents cannot conjugate
directly with the sulfur centre, the observation implies their −M
effects augment the polarisation of the aromatic p-system induced
by their inductive effects (see 5a–d), giving an accumulation of
positive charge adjacent to the S-bearing ring C-atom, which will
inhibit the development of similar charge on sulfur in the transition
state for oxidation.
The relative importance of the different electronic effects in the
overall substituent effect may be assessed from the ratio, k, of
the individual regression coefficients: k = qR/qI (see Table 4). For
the para subset the mesomeric effect apparently predominates,
whereas for the meta subset the inductive effect is the more
important, as expected. However, since the rI and various rR scales
have different ranges, the value of k may be somewhat misleading
as a measure of relative electronic effects. This difficulty can be
reduced by the standardisation of all variables by unit normal
scaling, which eliminates the intercept c and gives the scale of
each variable a mean value of zero and a standard deviation and
variance of 1.28 Thus Est = fIbI + fRbR where Est is the standardised
experimental observable, log(kX/kH), fI is the standardised rI scale,
fR is the appropriate standardised rR scale and bI and bR are
the corresponding regression coefficients. In Table 4 we compare
the standardised inductive and mesomeric contributions as kst
(i.e. bR/bI). It is seen that by this assessment, for the para subset,
the I and M effects contribute about equally to the combined
substituent effect whereas for the meta subset, the I effect is twice
as important as the M effect.
Entry 6 of Table 4 presents details of the Hammett correlation
for the oxidation of the full data set of 3-and 4-substituted
sulfoxides; as for the sulfides, the value of q obtained fully
confirms that reported by Murray and co-workers.19 Entries 7 and
8 (Hammett correlations of the separate para and meta subsets)
show that, as with the sulfides, the meta subset of sulfoxides gives
results which are more scattered than those of the para subset. The
Hammett correlation for the meta subset (entry 8) is ‘poor’ by the
criterion of Exner’s statistic, w;25 it is marginally improved by use
of r0m (entry 9) in place of rm but, nevertheless, remains ‘poor’. The
reaction constants for the separate subsets show greater variation
than those of the sulfides, probably as a result of the greater scatter
in the meta subset. The greater scatter in both meta subsets in
comparison to their respective para counterparts is most likely to
be due to experimental error. The range of (kX/kH)sulfide in the meta
subset is half that in the para subset, and the range of (kX/kH)sulfoxide
in the meta subset is less than one third of that in the para subset
but the uncertainty intervals on all the values average ca. 0.05
(Table 3). The uncertainty, being a greater proportion of the range
for the meta subsets, produces their greater scatter.
The best dual parameter correlation of log(kX/kH)sulfoxide for the
para subset is given by rI with r−R, with both variables significant
at the 0.01 level (entry 10); in terms of s and w, this correlation is
only marginally poorer than that with rp (entry 7). The mesomeric
effects of the 4-substituents play a greater role relative to their
inductive effects for the sulfoxides (kst = 1.32) than for the sulfides
748 | Org. Biomol. Chem., 2008, 6, 745–761
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
(kst = 1.09). The reason for this probably arises from the partial
positive charge on the S-atom in the sulfoxides. Such charge will
counter the p-polarisation of the aromatic sextet induced by the
substituents (Ip effect) thus relatively enhancing their mesomeric
effects. The application of a dual parameter correlation to the meta
subset of sulfoxides fails: no rR scale is significant at a level ≤0.1
and correlation by rI alone is poor.
Although subdivision of the data sets for both series of
oxidations and of the explanatory variables used to express them
allows possible differences between the subsets to be articulated,
the fact remains that the best overall account of the reactivity of
both series is given by the Hammett equation applied to the whole
data set in each case. Hence, if the data for one series are plotted
against those of the other there is a linear correlation (Fig. 1). The
line in Fig. 1 is given by eqn (2):
of sulfoxides by 1 are identical to those affecting the oxidation of
sulfides, implying the mechanisms are essentially the same.
(ii) Oxidations in acetone with aprotic co-solvents
Table 5 presents relative rate constants measured at 293 K for
oxidations performed in mixtures of acetone with aprotic co-
solvents, namely 60% v/v hexane and 80% v/v DMF. Acetone
was necessarily present as 1 was synthesised and isolated in acetone
solution. The constraint on the extent of dilution by hexane was
the need for a sufficient solubility of more polar substrates such
as methyl 4-nitrophenyl sulfoxide; DMF itself reacts slowly with
1, which limited the use of this co-solvent to oxidations of sulfides
only.
Table 6 details the correlation analysis of the logarithms of the
relative rate constants of the reactions studied. The substrates were
restricted to 4-substituted compounds which had given data with
less scatter than 3-substituted analogues when oxidised in acetone;
however, for the reactions in 60% v/v hexane in acetone, these
included 4-methoxy-3-methylphenyl methyl sulfide and sulfoxide.
In correlation analyses including these compounds, additivity of
substituent constants was assumed for the disubstituted phenyl
group. For the sulfides, a better Hammett correlation, as judged
by the values of s and w, was obtained by inclusion of the point so
parameterised than by exclusion of data for this substrate. Entry 1
reports the inclusive Hammett correlation for which the goodness
of fit, as indicated by w, is ‘good’. The chemically important
result is that the reaction constant, q, obtained in this solvent
is twice as large as that obtained in acetone alone (see later
for discussion of solvent effects on reaction constants). Entry 2
reports the corresponding best dual parameter correlation; this
gives no statistical improvement but the best rR scale is rBRA (with
both variables significant at the 0.01 level) whereas for reaction in
acetone alone the best scale was r0R (cf. Table 4 entry 4). This is
not the consequence of inclusion of the disubstituted substrate, as
the same rR scale remains the best if it is excluded. The balance
of the inductive and mesomeric effects in the mixed solvent (kst =
1.17) is somewhat greater than that in acetone alone (kst = 1.09),
consistent with the change of rR scale to one giving greater weight
to the +M effects of substituents.
log(kX/kH)sulfoxide = (1.067 0.084) log(kX/kH)sulfide
(2)
with n = 12, R2 = 0.970, F = 361.6, Fsignif = 3.51 × 10−9, s =
0.047 and w = 0.188. The standard error of the estimate, s, is of
very similar magnitude to the uncertainty in the plotted ordinates
(+0.042 to −0.046) calculated from the average of the 95%
confidence intervals given in Table 3. Since the log(kX/kH) terms
are proportional to the differences in free energy of activation
between the substituted and unsubstituted substrate in both series,
the 1 : 1 correspondence shown by eqn (2) indicates that, in
acetone, the factors whereby the substituents affect the oxidation
Log(kX/kH)sulfoxide in 60% v/v hexane in acetone (entry 3) is
poorly correlated by Hammett’s rp. The use of r+p in place of
rp improves the linear model somewhat (entry 4) and also allows
recognition of the point for the 4-Me substituent as a statistical
outlier,29 deletion of which gives a much improved correlation
Fig. 1 Plot of the logarithms of relative rate constants for the oxidation
of aryl methyl sulfoxides by 1 in acetone versus those for the oxidation of
the corresponding sulfides.
Table 5 Relative rate constants for the oxidation of aryl methyl sulfides and sulfoxides by dimethyldioxirane in acetone with aprotic co-solvents at 293 K
Co-solvent: 60% v/v hexane
Co-solvent: 80% v/v DMF
a
a
a
Aryl substituent, X
(kX/kH)sulfide
(kX/kH)sulfoxide
(kX/kH)sulfide
4-OMe-3-Me
4-OMe
4-Me
H
4-Cl
4-CN
4-NO2
3.149 0.391
2.477 0.289
1.312 0.165
1.000 0.000
0.494 0.048
0.094 0.011
0.059 0.012b
2.373 0.088
2.068 0.109
1.765 0.133
1.000 0.000
0.721 0.072
0.445 0.023
0.404 0.024b
1.444 0.085
1.063 0.072
1.000 0.000
0.668 0.027
0.352 0.016
a Uncertainties are the 95% confidence intervals. b Measured as (k4-NO /k4-CN) × (k4-CN/kH).
2
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 745–761 | 749
©

View Article Online
(entry 5). Entry 6 reports the corresponding dual parameter
correlation. The data are best expressed by rI and r+R at significance
levels of 0.02 and 0.01, respectively. There is thus a striking contrast
between the effect of substituents on the oxidation of sulfoxides
by 1 in the two solvent systems so far considered (cf. Table 4
entry 10). Although in both solvent systems the mesomeric effects
predominate, kst = 1.32 in acetone alone and 1.94 in the 60% v/v
hexane in acetone, it is the −M effects that are the more important
in acetone whereas in the mixed solvent it is the +M effects (see
later for discussion).
Entries 7 and 8 relate to the oxidation of sulfides by 1 in 80% v/v
DMF in acetone. As with the oxidations of sulfides in other
solvents, the data are correlated by rp (entry 7) but the smaller
values of s and w shown by the dual parameter correlation (rI and
rBRA) (entry 8), indicate that the latter gives a more precise account
of the data. Fig. 2 shows plots of the logarithms of the relative rate
constants for oxidation of sulfides in 60% v/v hexane in acetone
and 80% v/v DMF in acetone versus those for oxidation in acetone
alone. Both are linear but with different gradients. The lines are
given by eqn (3) and eqn (4).
–
acetone
log(kX/kH)hsuelxfiadnee
= (2.002 0.151) log(kX/kH)sauclefitodnee
(3)
with n = 6, R2 = 0.994, F = 878.28, Fsignif = 7.72 × 10−6, s = 0.049,
and w = 0.092 and
–acetone
log(kX/kH)DsuMlfidFe
= (0.891 0.168) log(kX/kH)sauclefitodnee
(4)
with n = 5, R2 = 0.979, F = 183.59, Fsignif = 8.69 × 10−4, s = 0.035,
and w = 0.189.
Fig. 2 Plots of the logarithms of relative rate constants for the oxidation
of aryl methyl sulfides by 1 in acetone with aprotic co-solvents versus
those for their oxidation in neat acetone: 1, 60% v/v hexane in acetone; 2,
80% v/v DMF in acetone.
It is seen from eqn (3) that the effect of dilution of acetone
(relative permittivity, er = 21.0130a) with the less polar hexane
(er = 1.8930b), giving a mixture with er = 7.9231a (see ESI, Section
S2†), is to double the overall substituent effect on the oxidation
of sulfides by 1 relative to that in neat acetone, whereas eqn (4)
shows that dilution by the more polar DMF (er = 38.2530a), giving
a mixture with er = 33.6031b (see ESI, Section S2†), has little effect
(a t-test shows the difference in the reaction constants of entry 2
in Table 4 and entry 7 in Table 6 is not significant at the 0.05
level of probability). It thus seems that, in the less polar solvent
750 | Org. Biomol. Chem., 2008, 6, 745–761
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
mixture, solvation cannot stabilise the development of charge on
S as much as in acetone alone, which results in a greater demand
for electron density upon the substituents and consequently in an
enhanced reaction constant. In the more polar solvent mixture,
the fact that the reaction constant remains essentially unchanged
suggests that, for each substituent, the extent of charge separation
has reached an upper limit in acetone and increasing the solvent’s
relative permittivity produces no further increase.
Fig. 3 shows the contrasting Hammett plots obtained for the
oxidation of sulfides and sulfoxides in 20% aqueous acetone. The
plot for sulfoxides is normal (cf. entry 7 in Table 8a) but that for
sulfides is curved; indeed, it is well represented by a quadratic
equation in rp (Table 8b). The coefficients found for this equation
are somewhat sensitive to the small differences between values of
log(kX/kH)sulfide close to zero. Entry 9 of Table 8b gives the statistical
analysis for the full set of sulfides. If the sole di-substituted
substrate is excluded on grounds of its structural difference, the
decrease in s and w in entry 10 indicates improved precision in the
fit of the residual data; however, the 95% confidence intervals for
the regression coefficients are larger than those in entry 9. If the
datum for the di-substituted substrate is restored and that for the
4-OMe substrate excluded on the grounds that log(kX/kH)sulfide is
out of sequence with those adjacent in the rp scale, a significant
improvement is seen: Fsignif, s and w are all decreased relative
to entry 2, and the 95% confidence intervals on the regression
coefficients are decreased relative to entry 1. Curve 2 in Fig. 3 is
given by eqn 5 (cf. entry 11 of Table 8b).
(iii) Oxidations in aqueous acetone
Relative rate constants for the oxidation of aryl methyl sulfides
and sulfoxides by 1 in various aqueous acetone mixtures at 293 K
are given in Table 7 and the corresponding correlation analyses
of their logarithms in Table 8a and b. Entry 1 of Table 8a reports
the Hammett correlation for the oxidation of sulfides in acetone
containing 1% by volume of water and entry 2 the best dual
parameter correlation (rI and rBRA) of the same data. Comparison
of the values of s and w between these two entries shows that
the latter entry gives the better account of the data. Analogous
results for oxidations of the same sulfides in 5% v/v aqueous
acetone are given in entries 3 and 4 and, again, a dual parameter
correlation (now rI and r0R) gives the more precise account of the
data. Comparison of the regression coefficients of entries 1 and 3
with that of entry 2 in Table 4 shows that the Hammett reaction
constant increases considerably with the proportion of water in
the solvent (by 57% for 1% water and by 75% for 5% water).
Similar scrutiny of the regression coefficients of entries 2 and 4
together with those of entry 4 in Table 4 shows that the changes
in the Hammett reaction constant are matched by changes in
both qI and qR. The finding that r0R is preferred to rRBA in two
of the three cases might arise as experimental error as suggested
above [see (i)].
2
log(kX/kH)sulfide = −(0.566 0.123)rp − (1.279 0.193)rp (5)
The Hammett correlations for the oxidation of sulfoxides in 1%
and 20% aqueous acetone are reported in Table 8a, entries 5 and
7, respectively. Comparison of their reaction constants with that
for oxidation in acetone alone (Table 4 entry 7) shows that the
values are very similar (qmean = −0.751 0.044), though a small
decrease with increasing proportion of water may be discerned (by
6% for 1% water and by 11% for 20% water). The dual parameter
correlations of the same data sets (entries 6 and 8 in Table 8a)
do not give improved precision, but comparison with entry 10 in
Table 4 shows a consistent preference for the r−R scale for reactions
in neat acetone and aqueous acetone solvents.
Fig. 3 Hammett plots for the oxidation of aryl methyl sulfides and
sulfoxides in acetone containing 20% v/v water: 1, sulfoxides; 2, sulfides.
Curvature in a Hammett plot, where the gradient changes
from zero through increasingly negative values as rp increases,
is diagnostic of the occurrence of a change of rate-determining
step within the range of substituents considered. It is apparent
from Fig. 3 that only substrates having rp > 0 show a significant
Table 7 Relative rate constants for the oxidation of aryl methyl sulfides and sulfoxides by dimethyldioxirane in aqueous acetone at 293 K
1% v/v Water
5% v/v Water
20% v/v Water
a
a
a
a
a
Aryl substituent, X
(kX/kH)sulfide
(kX/kH)sulfoxide
(kX/kH)sulfide
(kX/kH)sulfide
(kX/kH)sulfoxide
4-OMe-3-Me
4-OMe
4-Me
H
4-Cl
4-C(O)Me
4-CN
4-NO2
—
—
—
1.152 0.116
0.905 0.106
1.089 0.125
1.000 0.000
0.599 0.075
0.264 0.034
0.126 0.020
0.056 0.009b
1.900 0.062
1.462 0.135
1.559 0.087
1.000 0.000
0.803 0.067
0.448 0.026
0.372 0.014
0.295 0.017b
1.712 0.133
1.213 0.126
1.000 0.000
0.456 0.041
0.265 0.025
0.114 0.016
0.099 0.009b
1.441 0.091
1.309 0.052
1.000 0.000
0.702 0.063
0.402 0.028
0.331 0.022
0.230 0.014b
1.844 0.146
1.322 0.171
1.000 0.000
0.476 0.041
0.210 0.022
0.112 0.015
0.071 0.005b
a Uncertainties are the 95% confidence intervals. b Measured as (k4-NO /k4-CN) × (k4-CN/kH).
2
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 745–761 | 751
©

View Article Online
752 | Org. Biomol. Chem., 2008, 6, 745–761
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
substituent effect: a step which is initially not rate-determining be-
comes so for substrates having electron-withdrawing substituents,
consistent with the step concerned being an electrophilic attack
on the substrate.
where Ke is the equilibrium constant for the reversible reaction 1
(pre-equilibrium).
Applying this notation to both substituted and unsubstituted
substrates (initially equimolar and in large excess over 1), it follows
that the experimental relative rate constant, (kX/kH)sulfide, is given
by eqn (8).
This finding can be understood in terms of Scheme 3. Here, in
reaction 1, an aryl methyl sulfide, S, undergoes an electrophilic
attack by 1 in which the O–O bond stretches and breaks
heterolytically to give a betaine, B; the separation of charges can be
assisted by the high polarity of the medium and by solvation of the
ionic centres by water. The betaine reacts in three ways determined
by the internal site at which the anionic oxygen attacks. If the
site is the S-bonded oxygen, the result is reversal of reaction 1
(i.e. reaction −1). The principle of microscopic reversibility relates
this reaction to the scissoring of the OCO angle, which shortens
the O-to-O distance in the betaine. If the site is the adjacent
carbon atom, B fragments to give the sulfoxide product, P, and
acetone (reaction 2). This reaction can originate from the same
ground-state conformation as reaction −1 but corresponds to the
‘asymmetric stretching’ of the two C–O bonds in B, i.e. shortening
of C–O− concurrent with stretching of C–OS. The third possibility
is that the anionic oxygen attacks the sulfonium centre to produce
a dioxathietane, D, as found by Asensio and co-workers20 for 2 as
oxidant (reaction 3, cf. Scheme 2b). This reaction would originate
from a rotamer of B different from that assumed for reactions 2 and
−1. However, in the present case, the fact that no sulfone results
(providing that S is in excess over 1) shows that D, if formed, does
not undergo further oxidation. Since the yield of P is equivalent
to the amount of 1 used, any formation of D must be reversible;
D is thus a cul-de-sac which merely stores B until it reacts either
by reaction −1 or by reaction 2.
ꢁ
ꢂ
kX
kH
KeXkX2 (1 + fH)
=
(8)
H
H
Ke k2 (1 + fX)
sulfide
and taking logarithms,
ꢁ
ꢂ
ꢁ
ꢂ
ꢁ
ꢂ
ꢁ
ꢂ
kX
kH
KeX
Ke
k2X
k2H
1 + fX
1 + fH
log
= log
+ log
− log
H
sulfide
Application of the Hammett equation to the first two terms on
the right hand side gives eqn (9),
ꢁ
ꢂ
ꢁ
ꢂ
kX
kH
1 + fX
1 + fH
log
= rp(qe + q2) − log
(9)
sulfide
As reaction 1 is an electrophilic attack on S, the pre-equilibrium
will lie further to the right the more strongly electron-donating
is the substituent X. The reaction constant qe will therefore be
negative. In reaction 2, the full ionic charge on S in B is reduced to
+
−
=
the dipolar charge of the sulfoxide, P (S –O ↔ S O), a process
disfavoured by electron-donating X; q2 is therefore expected to be
positive. If qe and q2 are of similar absolute magnitudes it follows
that the first term on the right hand side of eqn (9) will be small.
If, as suggested above, the excitation to the transition state for
reaction −1 originates in an OCO in-plane bending (scissoring)
vibration in B whereas that for reaction 2 originates in an OCO
bond-stretching vibration of higher energy, in the absence of any
substituent effect and other factors such as desolvation being
equal, it is expected that kH2 < k−H1, i.e. 0 < f H < 1.
The behaviour of f X can be inferred from a comparison of
hypothetical Hammett plots for reactions 2 and −1 (Fig. 4). As
both reactions are disfavoured by electron-donating substitution,
these are drawn with positive gradients intersecting the ordinate at
rp = 0 such that logkH2 < logk−H1 as required by the above argument.
To agree with the observation of a change in rate-determining step
for electron-withdrawing substituents, the intersection of the plots
must occur at a positive value of rp. This requires that q2 > q−1.
Scheme 3
A kinetic analysis with application of the steady state approxi-
mation to B and D shows that, in general,
d[P]
dt
d[S] k1k2[S][1]
= −
=
(6)
dt
k−1 + k2
The usual assumption for simplifying eqn (6) is that either
k−1 ꢀ k2 or vice versa, which allows expression of a curved
Hammett plot in terms of two intersecting linear arms. However,
the degree of curvature implicit in the parabolic character of the
present plot suggests that such an assumption is not justified and
that, within the substituent range, k−1 and k2 are presumably of
comparable magnitudes.
Suppose f = k2/k−1. Then
d[P] k1k2[S][1] Kek2[S][1]
Fig. 4 Hypothetical Hammett plots for reactions −1 and 2 in Scheme 3
=
=
(7)
(see text).
dt
k−1(1 + f )
1 + f
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 745–761 | 753
©

View Article Online
As f X is defined as kX2 /k−X1,
substituents, which accounts for the observed very small variation
in log(kX/kH)sulfide
.
logf X = logkX2 − logk−X1
Applying the Hammett equation,
ꢁ
ꢂ
kX
kH
= −0.566 rp + log(1 + 0.07)
= −0.566 rp + 0.029
(13)
sulfide
logf X = (logkH2 + rpq2) − (logk−H1 + rpq−1)
We have (qe + q2) i.e. (q1 − q−1 + q2) = −0.566; now we also have
Dq i.e. (q2 − q−1) = 2.45. Subtraction of the latter equation from
the former gives q1 = −3.02. A reaction constant of this magnitude
is typical of a simple reaction step that creates a full positive ionic
charge on the first atom attached to the substituted aromatic ring
(cf. quaternisation of ArNMe2),32 as envisaged in the formation of
B (Scheme 3).
A reaction constant is usually obtained as the gradient of a linear
Hammett plot, but this is not possible when the plot is curved.
However, the gradient of the tangent to the curve at any value of rp
represents a ‘local reaction constant’ at that value. Differentiation
of eqn (5) with respect to rp allows the determination of the
gradient at particular points:
= logf H + rpDq,
where Dq = (q2 − q−1) > 0. Hence,
f H = f X/antilog(rpDq)
Substituting for f H in eqn (9) gives eqn (11),
(10)
(11)
ꢁ
ꢂ
ꢁ
ꢂ
kX
kH
1 + fX
log
= rp(qe + q2) − log
1 + fX/antilog(rpDq)
sulfide
Comparison of eqn (11) with eqn (5) indicates (qe + q2) =
−(0.566
0.123), i.e. a small figure as required for qe and q2
being of opposite sign and similar magnitude and that,
ꢁ
ꢂ
1 + fX
log
= (1.279 0.193) r2p
d [log(kX/kH)sulfide
drp
]
1 + fX/antilog(rpDq)
= −0.566 − (2 × 1.279 rp)
(14)
whence,
Thus, at rp = 0.78 (for X = 4-NO2) the ‘local reaction constant’
is −2.56, which is smaller in absolute magnitude than the value
of −3.02 obtained above. As shown in Table 9 for X = 4-NO2,
kX2 is just over five times larger than k−X1 and therefore is not
sufficiently large for kX−1 to be neglected in comparison with it. In
the expression kX = k1Xk2X/(k−X1 + kX2 ) [cf. eqn (6)] the denominator
is therefore significantly larger than kX2 in the numerator, and kX1
is underestimated, and hence so is the reaction constant, if the
two are assumed to cancel. Also, it should be noted that the
slowest step in the reaction sequence changes between H and NO2,
and therefore (kX/kH)sulfide does not strictly compare like with like.
Nevertheless, it is clear that as substituents become more electron-
withdrawing, the ‘local reaction constant’ approaches a value that
is typical for the formation of a full positive ionic charge on the
atom attached to the substituted ring.
The observed behaviour of the oxidation of aryl methyl sulfides
by 1 in aqueous acetone containing 20% v/v water is thus
consistent with the betaine-involving mechanism proposed in
Scheme 3. This is different from the behaviour of the same
oxidation in acetone alone. It is thereby demonstrated that sulfide
oxidation by 1 in acetone alone does not proceed via a betaine
intermediate as suggested by Asensio and co-workers,20 and that
the earlier inference of a concerted mechanism as in Scheme 2a is
most probably correct. Furthermore, the absence of any sulfone
product when reaction does proceed via a betaine intermediate
shows that betaine formation does not necessarily lead to the
formation of a dioxathietane intermediate that is more easily
oxidised than the initial sulfide. Conceivably, as suggested in
Scheme 3, a dioxathietane (D) might form reversibly in aqueous
acetone and prove to be inert to oxidation by 1 but, more
probably we suggest, no dioxathietane is produced on account of
solvation of the ionic centres of the betaine by water (see below).
The solvent-dependence of the behaviour of 2 as an oxidant of
sulfides supports this suggestion.20b The sulfone/sulfoxide product
ratio characteristic of reaction in polar aprotic solvents (CH2Cl2,
MeCN, Me2CO) is increased by admixture of non-polar aprotic
additives (e.g. CCl4) but markedly decreased by admixture of
antilog(1.279 r2p) − 1
fX =
(12)
1 − antilog(1.279 r2p − rpDq)
For assumed values of Dq, values of f X can hence be found for a
substituent from its Hammett constant. Values of f X were obtained
via eqn (12) for 0 < Dq < 4, which is a reasonable range in the
context of Hammett correlation of side-chain reactivity. All values
of Dq in this range gave negative f X for substituents having rp < 0,
which is physically meaningless. {Evidently for these substituents,
the second term on the right of eqn (9) reduces to −log[1/(1 + f H)],
i.e. log(1 + f H), which is constant}. Negative values of f X were also
found for 0 < Dq < 1 and rp > 0, which eliminates Dq in this range
from further consideration. Positive values of f X were obtained for
Dq ≥ 1 and rp > 0. In order further to narrow the range of Dq,
values of f H were found via eqn (10) from the calculated f X values,
the requirement being that f H should be constant irrespective of
the value of f X and should lie in the range 0–1. This requirement is
well met for Dq = 2.45, and the value of f H found is 0.07. It is also
found that, as anticipated, for electron-withdrawing X, kX2 and k−X1
are of similar magnitudes as all f X = kX2 /k−X1 < 6 (Table 9 and ESI,
Section S3†).
The finding that f H is 0.07 and that f X is without meaning for
substituents with rp < 0 transforms eqn (9) into eqn (13) for such
Table 9 Evaluation of f X and f H for Dq = 2.45
a
b
c
Substituent, X
rp
f X
f H
4-OMe-3-Med
4-OMe
4-Me
H
4-Cl
4-C(O)Me
4-CN
4-NO2
−0.34
−0.27
−0.17
0.00
−0.05
−0.05
−0.05
Indeterminate
0.25
−0.32
−0.23
−0.13
Indeterminate
0.07
0.23
0.50
0.66
1.24
2.86
0.07
0.07
0.07
0.78
5.40
a Ref. 22. b See eqn (12). c See eqn (10). d Additivity of substituent constants
assumed: ‘rp’ = rOp Me + rmMe
.
754 | Org. Biomol. Chem., 2008, 6, 745–761
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
protic additives (CF3CH2OH, t-BuOH, CF3CO2H, MeCO2H).
Since sulfones derive from a dioxathietane whereas sulfoxides
derive from a betaine (Scheme 2b), it is evident that the effect
produced by an additive depends on its ability to influence the
relative stability of the betaine and the dioxathietane.§ When
both 1 and 2 oxidise sulfides in conditions involving intermediate
betaines, the course of reaction depends crucially on the solvation
of the betaine: protic nucleophilic solvent components solvate the
betaine, thus favouring the formation of sulfoxide over sulfone.
additions increase er to 21.24 and 24.08, respectively31c–e (see ESI,
Section S2†), which are less than that caused by addition of
80% v/v DMF; hence, in view of the latter having scant effect on
the reaction constant, the observed behaviour must derive from
another property of the aqueous solvent. The most obvious is
the ability of water to interact with the transition state both by
lone-pair donation at the electron-deficient sulfur centre and by
hydrogen-bonding to the distal oxygen atom as shown in 6.
(iv) The solvent dependence of the concerted mechanisms
It is now clear that the oxidations of the sulfides investigated
here proceeded by a concerted mechanism with the exception
of the reaction in 20% v/v H2O in acetone. The oxidations of
sulfoxides showed no evidence of a change in mechanism as a
result of variation in the medium and we conclude that, for them,
reaction was concerted in all the media studied. As the regression
coefficients for both kinds of substrate show variation with the
medium, differences in transition state character might be inferable
from them.
Interactions such as these serve to stabilise individually the
partially charged centres and bring about an increased weakening
of the O–O bond in the dioxirane moiety relative to that in neat
acetone. The bond-weakening results in increased magnitudes
of partial charge with a consequent increase in the demand
for electron density from the substituents. Their differential
abilities to supply it results in the larger reaction constants
observed. Progressive increase in the proportion of water leads to a
combination of these effects with an increase in er, resulting in such
O–O bond weakening that formation of a betaine intermediate
becomes viable and a change of mechanism ensues as was observed
for 20% v/v H2O (er = 34.30).31c–e
Sulfoxides. The Hammett reaction constants for the oxidation
of sulfoxides by 1 are similar for reactions in acetone alone and
in 1% v/v and 20% v/v water in acetone, qmean = −0.751 0.044
(cf. Table 4 entry 7, Table 8a entries 5 and 7). Evidently, neither
the smaller nor the larger increase in water content of the solvent
has much effect on the relative rate constants. The changes in er
and hydrogen-bonding capacity of the solvent, which necessarily
accompany the addition of water, must either negate one another
or affect the absolute rate constants by similar combined factors
across the substituent range, and so cancel in the relative rate
constants. (The following paper examines the effect of solvent
variation on absolute rate constants). For oxidation of sulfoxides
in 60% v/v hexane in acetone, the Hammett correlation was poor
(Table 6 entry 3); a correlation was found by using r+p (Table 6
entry 5) that was only fair by Exner’s criterion despite a reduced
data set. Comparison of the reaction constants (Table 4 entry 7
and Table 6 entry 5) reveals a 59% reduction on transfer of the
oxidation from acetone to 60% v/v hexane in acetone, indicating
that the reaction becomes less selective in the solvent of lower
polarity. Although standardisation of the variables in multiple
regression facilitates comparison of the coefficients derived from
those having different scales, the comparisons for sulfoxides are
uninformative as, for all the solvent systems examined, |bR| > |bI|
and kst does not show systematic variation. The notable finding is
that, whereas f+R (i.e. standardised r+R) is the best mesomeric scale
for the standardised dual parameter correlation of reactions in
60% v/v hexane in acetone (consistent with r+p in single parameter
correlation), in acetone alone and in aqueous acetone mixtures,
the best mesomeric scale is f−R (cf. Table 6 entry 6, Table 4 entry 10
and Table 8a entries 6 and 8).
Sulfides. Activation to the transition state for the single-step
oxidation of sulfides by 1 involves a separation of partial charges
(Scheme 2a). This process is, in principle, dependent on the relative
permittivity, er, of the medium. Above [see (ii)], it was found
that the Hammett reaction constant for the oxidation of sulfides
increases two-fold from its value in acetone (er = 21.0130a), when er
is reduced to 7.92 by admixture of 60% v/v hexane (er = 1.8930b,31a
)
(see ESI, Section S2†). We suggested the reason to be that, in the
mixed solvent, where the effectiveness of the medium in stabilising
charge-separation is reduced, there is a greater demand upon
the substituents for electron density than in acetone alone. By
contrast, it was found that an increase in er to 33.60 caused by
admixture of 80% v/v DMF (er = 38.2530a,31b) (see ESI, Section
S2) had little effect: the measured reaction constant changed in
the opposite sense to that caused by reduction in er, but the
decrease was small and not statistically significant. Evidently,
the absolute rate constants for the oxidation of all the sulfides
examined change by an approximately constant factor in response
to increase in er and the reaction constant, determined from the
relative rate constants, kX/kH, thus remains virtually unchanged.
An alternative explanation considered is that, in the 80% v/v
DMF mixture, the transition states are preferentially solvated by
acetone and are thereby shielded from the effects of change in
the composition of the outer solvent shell. This, however, begs
the question as to why a similar situation does not occur in
the hexane–acetone mixture where the strength of dipole–dipole
interaction between the transition state and acetone molecules
would be expected to be greater than that in the DMF–acetone
mixture; preferential solvation therefore seems unlikely.
It was found that significant increases in the Hammett reaction
constant occur on addition of small amounts of water (1% and
5% v/v (er = 80.1030c) to acetone [see (iii) above]. These small
§ Note added in proof: In a study of the oxidation of PhC≡CH by 1
in CCl4,62 PhSMe, used in standardising solutions of 1 in CCl4, was
oxidised to sulfoxide–sulfone mixtures whereas, in acetone, only sulfoxide
was produced. This observation is explicable by the onset of the two-step
mechanism in the less polar solvent as occurs with 2. We are grateful to
the Editor for drawing our attention to this paper.
Fig. 5 compares the relative rate constants for oxidations of
sulfoxides by 1 in acetone and 60% v/v hexane in acetone
(cf. Tables 3 and 5). The striking feature is that, in the less
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 745–761 | 755
©

View Article Online
electron-accepting. The ground-states of such species are expected
to be less well stabilised in solvents of low polarity than in
more polar solvents. If ground-state stabilities were to be the
dominant factor in determining the pattern of activation energies
for oxidations of sulfoxides, the relative disposition of the curves
of Fig. 5 would be explained.
Fig. 5 Illustration of the disposition of the logarithms of relative rate
constants of aryl methyl sulfoxides when oxidised by 1 in different solvents:
᭹, neat acetone; ᭺, 60% v/v hexane in acetone.
polar solvent, both electron-withdrawing groups and electron-
donating groups increase the relative rate constants in comparison
to their values in acetone (the curves necessarily have the point
for the unsubstituted substrate in common). This fact implies
the operation of a solvent-dependent effect that varies across the
range of substituents. Sulfoxides have a polarised Sd+–Od− bond
(see below) which may be represented by the canonical structures
7a ↔ 7b. The dipolar positive charge on S results in the S(O)Me
group exerting a strong electron-attracting effect that polarises
an attached phenyl ring. This is demonstrated by its positive
substituent constants: rm = 0.52, rp = 0.49, rI = 0.50.22,24 The
near equality of rm and rp shows that, despite possession of a
lone-pair of electrons on S, the S(O)Me group exerts no significant
+M effect, and the similar value for rI shows its electronic effect
is essentially inductive. Structure 8 can therefore be envisaged to
contribute to the overall hybrid of methyl phenyl sulfoxide. Placing
a heteroatom-containing substituent para to the sulfoxide group
will introduce an opposing Ip polarisation of the ring. However,
of the substituents used, only CN and NO2 (rI 0.56 and 0.65,
respectively24) exert −I effects stronger than S(O)Me; thus, for all
the other substituents, the direction of net polarisation of the ring
will be the same as in 8. Furthermore, a +M substituent such
as 4-OMe should enhance the net polarisation (see 9a–b). This
expectation is based on the fact that S(O)Me has a substituent
constant r−p = 0.72,22 which implies that it can accept p-electron
density mesomerically from a conjugated strong donor. Given
the pyramidal disposition of the attachments to S, it is unlikely
that this ‘mesomeric acceptance’ involves an increased C(1)–S
bond order but rather that the relative weighting of canonical
structures 7a and 7b is adjusted in favour of 7a as a result of
increased negative charge at C(1). Strongly electron-withdrawing
substituents such as CN and NO2 will have the opposite effect.
Their combined −I and −M effects are sufficient to reverse the
polarisation of the aromatic ring caused by the S(O)Me group and
to deplete C(1) of p-electron density, leading to adjustment of the
relative weighting of canonical structures 7a and 7b in favour of
the latter (see 10a–c). It is thus clear that aryl methyl sulfoxides are
polarised molecules in which the net charge-separation depends
on the substitution and is expected to be most marked for the
most strongly acting substituents, whether electron-donating or
The SO bonds in sulfones are less polar than those in sulfoxides.
For example, the molecular dipole moment of Me2S (l = 1.47 D;
4.900 × 10−30 C m)33 is markedly increased on oxidation to Me2SO
(l = 3.90 D; 1.300 × 10−29 C m),33 whereas the second oxidation to
Me2SO2 produces a much smaller increment (l = 4.49 D; 1.498 ×
10−29 C m).33 Assuming the differences in molecular dipole moment
between that of Me2S and those of its two oxidation products are
essentially due to the SO bond moment(s) and, allowing for the
angular separation of the SO bonds in the sulfone (121◦),34 it is
clear that the SO bonds of the sulfone are much less polarised
than that of the sulfoxide. An analogous situation is expected
for any triad ArSMe, ArS(O)Me and ArS(O)2Me. That being the
case, it is expected that the transition state for the oxidation of
any sulfoxide to the sulfone, in which the properties of the latter
are developing, will be less polarised than the initial sulfoxide,
particularly so the later the transition state occurs in the reaction
coordinate. Theoretical calculations bear out this expectation.35
Transfer of the oxidation reaction from a polar medium to one
less polar should therefore lead to a relative stabilisation of the
transition state irrespective of the nature of Ar.
Three factors therefore appear to influence the relative reactivi-
ties of ring-substituted methyl phenyl sulfoxides: (1) the electronic
character of the substituent, which governs the availability of the
sulfur lone-pair to the electrophilic oxidant and hence determines
the order of reactivities in any solvent; (2) the effect of solvent on
the energy of the ground state of the sulfoxides, which is strongly
substituent-dependent; and (3) the effect of solvent on the energy
of the transition states, which largely depends on the bonding
changes occurring at S. These factors combine to produce the
relative reactivities illustrated in Fig. 5 and the reduced selectivity
observed on transfer to a less polar solvent. They also produce
the predominance of −M effects in polar solvents (acetone and
aqueous acetone) and, by contrast, of +M effects in less polar
60% v/v hexane in acetone.
In the polar solvents, the effect of substituents is dominated
by the ability of −M substituents to impede the electrophilic
attack of 1. The operation of a −M effect serves to increase
dipolar positive charge at C(1) whereas the electrophilic attack
increases electron-demand at the adjacent S atom. The former
effect must impede the latter change and, if solvation is more
756 | Org. Biomol. Chem., 2008, 6, 745–761
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
effective in stabilising the substituent-induced polarisation in the
ground state of the sulfoxide than in stabilisation of the transition
state for oxidation, the dominant effect of −M substituents can be
understood. It appears that hydrogen bonding is not important in
the solvation process, as the −M effect predominates in acetone
alone, which at best would form weak H-bonds with the sulfoxidic
oxygen and the relative reactivities show only slight decrease in the
presence of added water. (In the following paper it is shown that the
dominant property in determining the effect of solvent on the rates
of oxidation of methyl 4-nitrophenyl sulfoxide in aqueous acetone
is its polarity/polarisability). By contrast, in the less polar solvent,
the substituent-dependent polarisations caused by M effects of
either kind are not as well stabilised by solvation, and ground-
state energies are consequently raised. The less polar solvent
also facilitates excitation to the transition states for oxidation on
account of the decrease in SO bond dipolarity relative to that of the
substrates. There is, nevertheless, partial positive charge at S in the
transition states that is stabilised by electron-donation from +M
substituents accumulating negative charge at C(1), so accounting
for their predominance in this medium.
good agreement with that previously reported by Murray and
co-workers19 for a smaller data set. The reaction constant was
increased two-fold, i.e. the selectivity of reaction increased, when
the oxidations were carried out in 60% v/v hexane in acetone.
Since partial charges separate in the transition state for oxidation
of a sulfide by 1 and the sulfoxide reaction product is highly
polarised, both the transition state and, particularly, the product
will be relatively less well solvated than the reactants in a solvent
of lower polarity. A shift of transition state to a position later
in the reaction coordinate, with resultant increased selectivity, is
therefore consistent with the Hammond postulate.
2. The reaction constant measured when sulfide oxidations were
carried out in 80% v/v DMF in acetone was scarcely different
from that measured in acetone alone, indicating that increasing the
relative permittivity of the solvent does not shift transition states
to positions in the reaction coordinate significantly earlier than
those in acetone. This is consistent with the extent of transition
state dipolarity being maximised for each sulfide in acetone so
that, although increasing solvent polarity will stabilise each indi-
vidual maximally-polarised transition state, there is no differential
substituent-dependent stabilisation and the relative rate constants,
and hence the reaction constant, remain unchanged.
Comparison between sulfides and sulfoxides. In the Introduc-
tion it was noted that, on reaction with both hydroperoxidic
oxidants and dimethyldioxirane, sulfides and sulfoxides
do not follow the reactivity-selectivity principle: although,
experimentally, sulfoxides are generally less reactive than sulfides,
the oxidants do not discriminate more between substituted
sulfoxides than between substituted sulfides. Theoretical studies
indicate sulfoxides to be intrinsically more reactive than sulfides
towards electrophilic oxidants in the gas-phase but that the
relative reactivity is inverted by solvation.35 The discussion above
attests to the importance of solvation in determining the reactivity
of sulfoxides and the following paper demonstrates the relative
reactivity of sulfides and sulfoxides to be solvent-dependent. Why
sulfoxides should have a high intrinsic reactivity in the gas phase
which is maintained between substituted sulfoxides in solution,
leading to the failure of the reactivity-selectivity principle, can be
understood qualitatively as follows. The separation of the partial
charges of the Sd+–Od− bond requires a notional input of energy.
Suppose this is compensated by a concomitant stabilisation of
the lone-pair orbital on S as a consequence of the increase in the
electronegativity of S due to its increased positive charge. The
energy of the lone-pair orbital also depends on the substituent.
For each sulfoxide there is thus a substituent-dependent ‘balance’
between the extent of charge separation and the energy level of
the lone-pair orbital on S. When the lone-pair orbital interacts
with the LUMO of an electrophile on forming a transition state
this balance is perturbed. Collapse of the charge separation raises
the energy of the lone-pair orbital, hence increasing the strength
of its interaction with the LUMO and facilitating the formation
of the new bond to the electrophile. The sulfoxide SO bond thus
activates the electrophilic attack by providing electron density,
and the extent to which it does so is governed by the particular
substituent.
3. In aqueous acetone mixtures (1% v/v H2O and 5% v/v
H2O) Hammett reaction constants for oxidation of sulfides were
increased (by 57% and 75%, respectively) relative to that in
acetone alone. Since the relative permittivities of these solvent
mixtures are less than that of 80% v/v DMF in acetone (where
the effect was negligible) it is clear that factors specific to the
water co-solvent are operative. It is suggested these are solvation
of the individual centres of charge: lone-pair donation to Sd+ and
hydrogen bonding to Od− (cf. 6), which allows a greater extent of
charge separation than can occur under the influence of increased
relative permittivity alone.
4. Increasing the water content further to 20% by volume
results in a change of kinetic behaviour, although the relative
permittivity is very similar to that of 80% v/v DMF in acetone.
The solvation by water of the separate centres of charge and the
raised relative permittivity together enable the electrophilic attack
of dimethyldioxirane upon a sulfide to produce an intermediate
sulfonium betaine, a process that is rate-determining for sulfides
with electron-withdrawing substituents and for which a character-
istic Hammett reaction constant of −3.0 can be deduced. There is
no evidence of the betaine cyclising to a dioxathietane as reported
for sulfide oxidations by (trifluoromethyl)methyldioxirane.20
5. The oxidation of aryl methyl sulfoxides to the corresponding
sulfones by dimethyldioxirane occurs via a single-step concerted
mechanism irrespective of the solvent composition. The behaviour
of sulfoxides can be understood in terms that stem from their
high and variable polarity, arising from their polarised SO bond
and its interaction with substituents and solvent. The reaction
constant is only slightly diminished (by 11%) between acetone
and 20% v/v aqueous acetone as reaction media, which indicates
that the substituent effect on the oxidation of sulfoxides to sulfones
is rather insensitive to the specific solvating properties of water.
Although solvation reduces the reactivity of sulfoxides relative to
sulfides, sulfoxides are activated to electrophilic attack by their
polarised SO bonds, a fact which accounts for the failure of
the reactivity-selectivity principle when they are compared with
sulfides reacting with the same electrophile.
Conclusions
1. The Hammett reaction constant found for the oxidation of
aryl methyl sulfides by dimethyldioxirane, 1, in acetone is in
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 745–761 | 757
©

View Article Online
3-(Trifluoromethyl)phenyl methyl sulfide
Experimental
Colourless oil (2.20 g, 57%); m/z 192 (100%, M+), 177 (13), and
159 (18); mmax/cm−1(CH2Cl2) 2926 (CH), 1169s and 1105s (CF);
dH (270 MHz, CDCl3) 2.52 (s, 3H) and 7.35–7.54 (m, 4H); dC
(67.9 MHz, CDCl3) 16.4, 122.4, 122.8, 124.4 (q, J 249.9, 1C),
130.0, 130.3, 133.2 (q, J 33.3, 1C) and 141.0.39
General
Infra-red spectra, obtained for solutions in CH2Cl2, were recorded
on an Ati Mattson Genesis Series FTIR instrument. NMR
spectra, obtained for solutions in CDCl3, were recorded using
a JEOL JNM-EX270 (270 MHz) spectrometer and are referenced
to residual CHCl3. GC analysis was performed on a Varian 3000
CX chromatograph served by a Varian Star Workstation 4.5; the
column used was an Econocap FFAP column from Alltech (30 m,
0.32 mm, 0.25 lm). For column chromatography, ICN Silica 32–
3-Cyanophenyl methyl sulfide
Colourless solid (1.64 g, 55%), mp 39.8–41.0 ◦C (lit.40 40 ◦C); m/z
149 (100%, M+), 134 (22), 116 (30); mmax/cm−1(CH2Cl2) 2926 (CH)
and 2231s C≡N; dH (270 MHz, CDCl3) 2.50 (s, 3H), and 7.31–7.48
(m, 4H); dC (67.9 MHz, CDCl3) 15.3, 113.0, 118.4, 128.2, 128.7,
129.2, 130.3 and 140.9.
˚
63, 60 A, was used and thin layer chromatography was performed
using DC-Alufolien Kieselgel 60 F254 tlc plates supplied by Merck.
Data processing for correlation analysis was carried out using
SPSS 10 for Windows and the Excel add-in Essential Regression
2.219.28
Methyl 3-nitrophenyl sulfide
Pale yellow solid (2.03 g, 60%), mp 89.1–90.3 ◦C (lit.41 90 ◦C); m/z
169 (100%, M+), 123 (30), 108 (31) and 45 (44); mmax/cm−1(CH2Cl2)
2952 (CH), 1519s and 1349s (NO2); dH (270 MHz, CDCl3) 2.54 (s,
3H), 7.42 (t, J 7.7, 1H), 7.52 (ddd, J 7.9, 1.7 and 1.2, 1H), 7.96
(ddd, J 7.9, 2.2, 1.0, 1H) and 8.03 (t, J 1.9, 1H); dC (67.9 MHz,
CDCl3) 15.3, 119.6, 120.0, 129.3, 131.9, 141.5 and 148.5.42
Materials
Solvents (Fisher) were of either analytical or HPLC grade and
were used as supplied. Methyl phenyl sulfide and its 4-substituted
derivatives were sourced commercially (Aldrich, Fluka or Lan-
caster) in the highest available purity. Methyl phenyl sulfoxide and
sulfone (Aldrich) were further purified by column chromatography
(silica, 10% v/v acetone in CH2Cl2).
Oxidations
Sulfoxides and sulfones were prepared by oxidation of the
corresponding sulfides in procedures which differed only in
the stoichiometric proportion of oxidant used. The sulfide (ca.
12 mmol) was oxidised in chilled (ice-water bath) aqueous
Methyl 3-substituted-phenyl sulfides
Methyl 3-substituted-phenyl sulfides were synthesised from the
corresponding anilines by adaptation of the procedure of Oae,
Shinhama and Kim36 to use tert-butyl nitrite in place of tert-
butyl thionitrate. The appropriate aniline (20 mmol) was added to
dimethyl disulfide (20 cm3) and the mixture stirred magnetically.
After heating to ca. 100 ◦C, tert-BuONO (3.10 g, 30 mmol)
was added in five portions over 10 min (Vigorous evolution of
N2!). After a further 20 min stirring, the mixture was cooled to
room temperature and the solvent removed in vacuo (Stench!).
The reaction mixture was purified by column chromatography
(silica, 5% v/v diethyl ether in hexane); after rotary evaporation
of the bulk of eluting solvent, vestigial amounts were removed by
pumping under vacuum at ambient temperature.
R
acetone (50% v/v, 40 cm3) by addition, with stirring, of Oxoneꢀ
(2KHSO5·KHSO4·K2SO4) (0.55 or 1.5 mol equiv. depending on
the product required). Reactions were normally stirred overnight.
On completion, the reaction mixtures were diluted with water
(30 cm3) and extracted with CH2Cl2 (3 × 30 cm3). The extracts
were washed with water (50 cm3) and dried (Na2SO4). Evaporation
of the solvent gave the required products, which were purified
by column chromatography. In all cases gradient elution was
used; initially CH2Cl2 (100%), and finally acetone (10% v/v)
in CH2Cl2.
3-Methoxyphenyl methyl sulfoxide
Colourless oil (72%); m/z 170 (61%, M+), 155 (100), 92 (25) and
76 (26); mmax/cm−1(CH2Cl2) 3051 and 2966 (CH), 1250 (CO) and
1068s (SO); dH (270 MHz, CDCl3) 2.76 (s, 3H), 3.82 (s, 3H), 6.96
(ddd, J 8.2, 2.4 and 0.7, 1H), 7.08 (dd, J 7.7, and 2.4, 1H), 7.10–
7.25 (m, 1H) and 7.36 (t, J 8.0, 1H); dC (67.9 MHz, CDCl3) 43.9,
55.4, 107.8, 115.4, 117.2, 130.2, 147.0 and 160.3.37
3-Methoxyphenyl methyl sulfide
Colourless oil (0.31 g, 10%); m/z 154 (100%, M+), 121 (47);
mmax/cm−1(CH2Cl2) 2926 (CH) and 1248s (CO); dH (270 MHz,
CDCl3) 2.47 (s, 3H), 3.79 (s, 3H), 6.68 (ddd, J 8.0, 2.4 and 0.7, 1H),
6.75–6.86 (m, 2H) and 7.20 (t, J 8.0, 1H); dC (67.9 MHz, CDCl3)
15.7, 55.2, 110.5, 112.1, 118.7, 129.6, 139.8 and 159.8.37,38
3-Methoxyphenyl methyl sulfone
3-Chlorophenyl methyl sulfide
◦
Colourless solid (83%), mp 46.1–47.3 C (lit.43 47 ◦C); m/z 186
Colourless oil (1.17 g, 37%); m/z 158 (100%, M+), 143 (18), 125
(33) and 108 (32); mmax/cm−1(CH2Cl2) 2926 (CH) and 768 (CCl);
dH (270 MHz, CDCl3) 2.49, (s, 3H) and 7.05–7.27 (m, 4H); dC
(67.9 MHz, CDCl3) 15.5, 124.4, 124.9, 125.7, 129.7, 134.7 and
146.6.38
(39%, M+), 171 (46) and 107 (100); mmax/cm−1(CH2Cl2) 3066 and
2965 (CH), 1322s and 1152s (SO2); dH (270 MHz, CDCl3) 2.96 (s,
3H), 3.85 (s, 3H), 7.05 (t, J 8.1, 1H), 7.16 (dt, J 7.9 and 2.1, 1H)
7.28 (dt, 7.9 and 1.9, 1H), and 7.46 (t, J 2.1, 1H); dC (67.9 MHz,
CDCl3) 44.3, 57.4, 109.8, 115.4, 118.6, 131.2, 148.0 and 161.2.37
758 | Org. Biomol. Chem., 2008, 6, 745–761
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
3-Chlorophenyl methyl sulfoxide
J 7.7, 1H), 7.99 (ddd, J 7.7, 1.5 and 1, 1H), 8.34 (ddd, J 8.2, 2.4
and 1.2, 1H) and 8.49 (t, J 1.9, 1H); dC (67.9 MHz, CDCl3) 43.9,
118.9, 125.7, 129.2, 130.6, 148.5 and 148.6.
Colourless solid (86%), mp 42.2–43.4 C (lit.44 43–44 ◦C); m/z
174 (76%, M+), 159 (100), 131 (41), 111 (37), 84 (82) and 74 (42);
mmax/cm−1(CH2Cl2) 3052 and 2983 (CH), 1089s (SO) and 782 m
(CCl); dH (270 MHz, CDCl3) 2.69 (s, 3H), 7.40–7.48 (m, 3H) and
7.62 (m, 1H); dC (67.9 MHz, CDCl3) 44.6, 122.2, 124.2, 131.2,
131.8, 136.3 and 148.4.45
◦
Methyl 3-nitrophenyl sulfone
◦
Pale yellow solid (82%), mp 147.5–148.2 C (lit.46 147–148 ◦C);
m/z 201 (33%, M+), 186 (39), 139 (100), and 122 (60);
mmax/cm−1(CH2Cl2) 3080 and 2987 (CH), 1537s and 1353s (NO2),
1327s and 1161s (SO2); dH (270 MHz, CDCl3) 3.14 (s, 1H), 7.83
(t, J 7.7, 1H), 8.30 (dt, J 7.7 and 1.2, 1H), 8.53 (ddd, J 8.2, 2.2
and 1.2, 1H) and 8,81 (t, J 1.7, 1H); dC (67.9 MHz, CDCl3) 44.3,
122.9, 128.3, 130.9, 133.0, 142.3 and 148.7.
3-Chlorophenyl methyl sulfone
Colourless solid (54%), mp 105.2–106.4 C (lit.46 106 ◦C); m/z
◦
190 (63%, M+), 175 (56), 128 (57), 111 (100) and 74 (56);
mmax/cm−1(CH2Cl2) 3066 and 2929 (CH), 1321s and 1157s (SO2);
dH (270 MHz, CDCl3) 3.05 (3H), 7.51 (t, J 7.7, 1H), 7.62 (ddd, J
8.0, 2.2 and 1.2, 1H), 7.82 (ddd, J 7.8, 2.9 and 1.2, 1H) and 7.91 (t,
J 1.9, 1H); dC (67.9 MHz, CDCl3) 45.0, 126.1, 128.2, 131.4, 134.5,
136.2 and 142.8.
4-Methoxy-3-methylphenyl methyl sulfoxide
Colourless oil (85%), m/z 184 (16%, M+), 169 (100); mmax/cm−1
(CH2Cl2) 3049 and 2966 (CH), 1255s (CO) and 1047s (SO); dH
(270 MHz, CDCl3) 2.20 (s, 3H), 2.65 (s, 3H), 3.81 (s, 3H), 6.86 (d,
J 8.1, 1H), and 7.30–7.45 (m, 2H): dC (67.9 MHz, CDCl3) 16.3,
44.0, 55.5, 110.2, 123.1, 125.6, 128.4, 135.8 and 160.0.51
3-(Trifluoromethyl)phenyl methyl sulfoxide
Colourless oil (58%); m/z 208 (87%, M+), 193 (100), 165 (28), 145
(50) and 75 (26); mmax/cm−1(CH2Cl2) 3062 and 2983 (CH), 1168s
and 1105s (CF) and 1091s (SO); dH (270 MHz, CDCl3) 2.72 (s,
3H), 7.60–7.79 (m, 3H) and 7.89 (t, J 0.7, 1H); dC (67.9 MHz,
CDCl3) 43.8, 120.5, 123.9 (q, J 244.4, 1C), 126.7, 127.6, 128.8,
131.6 (q, J 27.4, 1C) and 147.2.39
4-Methoxy-3-methylphenyl methyl sulfone
◦
Colourless solid (68%), mp 64.2–64.9 C (lit.52 64–65 ◦C); m/z
200 (95%, M+), 185 (100), 137 (51), 121 (84), 91 (95) and 77 (34);
mmax/cm−1(CH2Cl2) 2969 (CH), 1307s and 1130s (SO2) and 1260
(CO); dH (270 MHz, CDCl3) 2.20 (s, 3H), 2.96 (s, 3H), 3.82 (s, 3H),
6.86 (d, J 8.9, 1H), 7.62 (dd, J 2.2 and 1.6, 1H) and 7.70 (d, J 8.6,
1H); dC (67.9 MHz, CDCl3) 16.2, 44.9, 55.7, 109.7, 127.2, 128.1,
129.4, 131.5 and 161.9.
3-(Trifluoromethyl)phenyl methyl sulfone
Colourless solid (67%), mp 58.3–59.5 ◦C (lit.47 58–60 ◦C); m/z 224
(17%, M+), 209 (19), 162 (42) and 145 (100); mmax/cm−1(CH2Cl2)
3019 and 2967 (CH), 1327s and 1153s (SO2); dH (270 MHz, CDCl3)
3.10 (s, 3H), 7.75 (t, J 7.8, 1H), 7.93 (d, J 7.7, 1H) and 8.13–8.22
(m, 2H); dC (67.9 MHz, CDCl3) 44.4, 123.7 (q, J 240.3, 1C), 124.5,
130.2, 130.5, 130.7, 133.2 (m, 1C), 141.5.47
4-Methoxyphenyl methyl sulfoxide
◦
Colourless solid (65%), mp 42.0–43.6 C (lit.44 44–45 ◦C); m/z
170 (15%, M+) and 155 (100); mmax/cm−1(CH2Cl2) 3050 and 2966
(CH), 1250s (CO) and 1043 (SO); dH (270 MHz, CDCl3) 2.70 (s,
3H), 3.85 (s, 3H), 7.03 (d, J 8.7, 2H) and 7.70 (d, J 8.7, 2H): dC
(67.9 MHz, CDCl3) 43.8, 55.4, 114.7 (2C), 125.3 (2C), 136.4 and
161.8.53,54
3-Cyanophenyl methyl sulfoxide
Colourless solid (72%), mp 69.0–69.8 C (lit.48 69–70 ◦C); m/z
◦
165 (85%, M+), 150 (100); mmax/cm−1(CH2Cl2) 2983 (CH), 2235s
(C≡N) and 1080s (SO); dH (270 MHz, CDCl3) 2.76 (s, 3H), 7.67
(t, J 7.7, 1H)), 7.79 (dt, J 7.7 and 1.2, 1H), 7.86 (dt, J 7.7 and 1.5,
1H) and 7.95 (t, J 1.7, 1H); dC (67.9 MHz, CDCl3) 43.9, 113.8,
117.5, 127.2, 127.6, 130.2, 134.3 and 147.8.
4-Methoxyphenyl methyl sulfone
Colourless solid (76%), mp 118.9–119.8 ◦C (lit.47 118–120 ◦C); m/z
186 (84%, M+), 171 (100), 123 (45), 107 (66), 92 (44) and 77 (65);
mmax/cm−1(CH2Cl2) 3056 and 2842 (CH), 1319 and 1145 (SO2) and
1298 (CO); dH (270 MHz, CDCl3) 2.99 (s, 3H), 3.85 (s, 3H), 6.98
(d, J 8.7, 2H) and 7.83 (d, J 8.7, 2H); dC (67.9 MHz, CDCl3) 44.7,
55.6, 114.4 (2C), 129.4 (2C), 132.1 and 163.6.47,55
3-Cyanophenyl methyl sulfone
Colourless solid (87%), mp 100.9–101.6 C (lit.49 101–103 ◦C);
◦
m/z 181 (28%, M+), 166 (37), 119 (53), and 102 (100);
mmax/cm−1(CH2Cl2) 3066 and 2929 (CH), 2237 (C≡N), 1327s and
1149s (SO2); dH (270 MHz, CDCl3) 3.09 (s, 3H), 7.74 (t, J 7.8, 1H),
7.94 (dt, J 7.8 and 1.5, 1H), 8.17 (dt, J 7.8 and 1.0, 1H) and 8.23 (t,
J 0.7, 1H); dC (67.9 MHz, CDCl3) 44.4, 114.0, 116.8, 130.5, 131.1,
131.4, 136.8 and 142.0.
Methyl 4-methylphenyl sulfoxide
◦
Colourless solid (68%), mp 41.9–42.9 C (lit.45 40 ◦C); m/z 154
(54%, M+), 139 (100) and 91 (30); mmax/cm−1(CH2Cl2) 3041 and
2981 (CH) and 1045 (SO); dH (270 MHz, CDCl3) 2.41 (s, 3H), 2.70
(s, 3H), 7.33 (d, J 8.0, 2H) and 7.54 (d, J 8.0, 2H); dC (67.9 MHz,
CDCl3) 21.2, 43.8, 123.2 (2C), 129.8 (2C), 141.3 and 142.3.45,53,54,56
Methyl 3-nitrophenyl sulfoxide
◦
Pale yellow solid (72%), mp 114.3–115.7 C (lit.50 115–116 ◦C);
Methyl 4-methylphenyl sulfone
m/z 185 (96%, M+), 170 (100), 139 (12), 124 (40), 75 (27) and
48 (29); mmax/cm−1(CH2Cl2) 3068 and 2873 (CH), 1535s and 1351s
(NO2) and 1076s (SO); dH (270 MHz, CDCl3) 2.79 (s, 3H), 7.75 (t,
Colourless solid (86%), mp 86.1–86.7 ◦C (lit.47 85–87 ◦C); m/z 170
(35%, M+), 155 (36), 107 (30), and 91 (100); mmax/cm−1(CH2Cl2)
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 745–761 | 759
©

View Article Online
3056 and 2929 (CH), 1315s and 1149s (SO2); dH (270 MHz, CDCl3)
2.41 (s, 3H), 3.00 (s, 3H), 7.33 (d, J, 8.3, 2H), and 7.79 (d, J, 8.3,
2H); dC (67.9 MHz, CDCl3) 21.5, 44.5, 127.2 (2C), 129.8 (2C),
137.5 and 144.5.47
3099 and 2863 (CH), 1529s and 1346s (NO2) and 1056s (SO);
dH (270 MHz, CDCl3) 2.78 (s, 3H), 7.82 (d, J 8.5, 2H) and 8.37 (d,
J 8.5, 2H); dC (67.9 MHz, CDCl3) 43.7, 124.4 (2H), 126.6 (2C),
149.5 and 153.2.45,53,60
4-Chlorophenyl methyl sulfoxide
Methyl 4-nitrophenyl sulfone
◦
Colourless solid (64%), mp 45.0–46.3 C (lit.44 46–47 ◦C); m/z
◦
Pale yellow solid (74%), mp 141.5–141.8 C (lit.60 142–144 ◦C);
174 (57%, M+), 159 (100), 131 (27), 111 (22) and 75 (33);
mmax/cm−1(CH2Cl2) 2981 (CH) and 1053s (SO); dH (270 MHz,
CDCl3) 2.73 (s, 3H), 7.51 (d, J 8.4, 2H) and 7.61 (d, J 8.4, 2H);
dC (67.9 MHz, CDCl3) 43.9, 124.8 (2C), 129.5 (2C), 137.0 and
144.1.45,54,56
m/z 201 (31%, M+), 186 (35), 139 (100) and 122 (60);
mmax/cm−1(CH2Cl2) 3066 and 2967 (CH), 1535s and 1349s (NO2),
1326 m and 1157s (SO2); dH (270 MHz, CDCl3) 3.12 (s, 3H), 8.16
(d, J 8.9, 2H) and 8.42 (d, J 8.9, 2H); dC (67.9 MHz, CDCl3) 44.6,
124.9 (2C), 1.29.3 (2C), 164.2 and 151.2.55,60
4-Chlorophenyl methyl sulfone
Dimethyldioxirane, 1
◦
Colourless solid (67%), mp 96.5–96.9 C (lit.57 97 ◦C); m/z
The method adopted was that of Adam and co-workers.61 Water
(60 cm3), acetone (50 cm3) and NaHCO3 (14.5 g, 0.17 mol) were
added to a 1000 cm3 two-necked, round-bottomed reaction flask
equipped with a magnetic stirrer, and cooled (to 5–10 ◦C) by
immersion in an ice-water bath. The reaction flask was connected
by means of a U-tube to a Dewar condenser and thence to a
two-necked receiving flask. The condenser and receiving flask
were cooled to −78 ◦C by means of dry ice/acetone. With
190 (35%, M+), 175 (45), 12 (37), 111 (100) and 75 (45);
mmax/cm−1(CH2Cl2) 3056 and 2934 (CH), 1319s and 1155s (SO2)
and 775 (CCl); dH (270 MHz, CDCl3) 3.04 (s, 3H), 7.53 (d, J 8.5,
2H), 7.62 (d, J 8.5, 2H); dC (67.9 MHz, CDCl3) 44.4, 128.8 (2C),
129.6 (2C), 138.9 and 140.4.58
4-Acetylphenyl methyl sulfoxide
R
cooling and vigorous stirring, Oxoneꢀ (2KHSO5·KHSO4·K2SO4)
Colourless solid (82%), mp 107.1–108.0 ◦C (lit.56 110 ◦C); m/z 182
(100%, M+), 167 (98), 152 (75) and 139 (20); mmax/cm−1(CH2Cl2)
3006 and 2979 (CH), 1689s (C O) and 1052s (SO); dH (270 MHz,
CDCl3) 2.57 (s, 3H), 2.69 (s, 3H), 7.67 (d, J 8.6, 2H) and 8.02 (d, J
8.6, 2H); dC (67.9 MHz, CDCl3) 26.6, 43.6, 123.5 (2C), 128.9 (2C),
138.8, 150.7 and 196.9.56
(30 g, 48.7 mmol) was added in 5 portions at 3 min intervals.
Three min after the last addition, the ice-water bath was removed,
a moderate vacuum (80–100 Torr) was applied and, with vigorous
stirring, the reaction mixture was distilled. Typically, 20–30 cm3 of
pale yellow dimethyldioxirane in acetone distillate was collected
(Caution: strained peroxide—but dilute and no hazards reported).
The concentration of 1 (0.04–0.10 mol dm−3, 2–6%) was deter-
mined by oxidation of MeSPh and GC analysis of the MeS(O)Ph
produced using 1,4-dibromobenzene as internal standard. The
dried (Na2SO4) dimethyldioxirane/acetone solution was stored
at −20 ◦C.
=
4-Acetylphenyl methyl sulfone
◦
Colourless solid (73%), mp 128.9–129.3 C (lit.46 127–128 ◦C);
m/z 198 (22%, M+), 183 (100) and 121 (30); mmax/cm−1(CH2Cl2)
=
3064 (CH), 1693s (C O), 1321s and 1153s (SO2); dH (270 MHz,
CDCl3) 2.63 (s, 3H), 3.05 (s, 3H), 8.00 (d, J 8.6, 2H) and 8.09 (d, J
8.6, 2H); dC (67.9 MHz, CDCl3) 26.8, 44.1, 127.7 (2C), 129.0 (2C),
140.7, 144.0 and 196.6.55
Competitive oxidations
Stock solutions in acetone of each of the reactant sulfides and sul-
foxides and their respective product sulfoxides and sulfones were
prepared freshly as required. Gas-chromatographic response fac-
tors were determined for each product using 1,4-dibromobenzene
as internal standard. The chromatographic conditions for both
the calibration and the analysis of reaction products employed
the earlier mentioned FFAP capillary column with temperature
programming: 1 min at 60 ◦C then increasing to 230 ◦C (the
4-Cyanophenyl methyl sulfoxide
Colourless solid (58%), mp 86.3–87.9 ◦C (lit.59 86–88 ◦C); m/z 165
(73%, M+), 150 (100), 122 (43) and 102 (48); mmax/cm−1(CH2Cl2)
2986 (CH), 2233s (C≡N) and 1075s (SO); dH (270 MHz, CDCl3)
2.79 (s, 3H) and 7.75–7.95 (m, 4H); dC (67.9 MHz, CDCl3) 43.6,
114.6, 117.6, 124.2 (2H), 132.8 (2H) and 151.3.45
◦
maximum for the column) at 16 C min−1. For all analyses the
4-Cyanophenyl methyl sulfone
split injection mode was used with, normally, a 1 ll injection
volume; the exceptions were mixtures containing the 3- and 4-
nitrophenyl sulfoxides and sulfones where 3 ll was used to improve
the response. For the competitive oxidations, measured volumes—
around 1 cm3 of stock solutions of the competing substrates in
acetone (8 × 10−3 to 2 × 10−2 mol dm−3)—were added to a small re-
action flask such that there was a 1–1.5-fold molar excess of the less
reactive compound; in mixed solvent studies the co-solvent was
added at this juncture. The mixtures were equilibrated at 20.0
Colourless solid (68%), mp 140.0–142.1 C (lit.47 142–143 ◦C);
◦
m/z 181 (26%, M+), 166 (28), 119 (75) and 102 (100);
mmax/cm−1(CH2Cl2) 3062 and 2929 (CH), 2235 (C≡N), 1322s and
1151s (SO2); dH (270 MHz, CDCl3) 3.08 (s, 1H), 7.88 (d, J 8.3, 2H)
and 8.07 (d, J 8.3, 2H); dC (67.9 MHz, CDCl3) 44.5, 117.4, 117.9,
128.5 (2C), 133.6 (2C) and 144.7.47
Methyl 4-nitrophenyl sulfoxide
◦
0.1 C for 15 min. Standardised dimethyldioxirane solution (see
Pale yellow solid (82%), mp 150.6–151.9 C (lit.60 151–152 ◦C);
m/z 185 (100, M+), 170 (52) and 50 (41); mmax/cm−1(CH2Cl2)
above) was diluted to ca. 5 × 10−3 mol dm−3 with the appropriate
◦
solvent and also thermally equilibrated. An aliquot (0.1 cm3) was
760 | Org. Biomol. Chem., 2008, 6, 745–761
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
added to the substrate mixture, which was shaken vigorously for
ca. 10 s and then allowed to stand for 3 min before analysis. Each
competitive oxidation was repeated three to five (n) times and the
mean value of kX/kH and its standard deviation smean found from
the molar product ratios and the known initial molar ratios (r0) of
competing substrates [see eqn (1)]. The standard error of the mean
(smean/ n) was multiplied by the appropriate Student’s t-factor for
(n − 1) degrees of freedom to give the 95% confidence intervals.
and the substituent, such as phenylethanoic acids and 3-substituted
benzene derivatives (see ref. 23).
28 On standardisation, each variable is centred by subtraction of its
mean from the individual values and scaled by division of the
differences by the standard deviation of the variable. D. D. Step-
pan, J. Werner and R. P. Yeater, Essential Regression (add-in for
Microsoft Excel), 1998 and 2006, erbbook, p. 13; freely available at
http://www.geocities.com/jowerner98/index.html.
√
29 See ref. 23b, p. 30.
30 CRC Handbook of Chemistry and Physics, 84th edn, ed. D. R. Lide,
CRC Press, Boca Raton, 2003: (a) Table 6-159, (b) Table 6-166,
(c) Table 6-156.
Acknowledgements
31 (a) Values interpolated from data for the binary solvent systems (see
ESI, Section 2†) given by: F. Mato and F. Fernandez-Polanco, An.
Quim., 1976, 72, 280; (b) D. S. Gill and H. Schneider, Indian J. Chem.,
Sect. A, 1980, 19, 313; (c) P. S. Albright, J. Am. Chem. Soc., 1937, 59,
2098; (d) A. M. Nilsson and P. Beronius, Z. Phys. Chem., 1972, 79, 83;
(e) K. Singh and P. P. Sinha, J. Indian Chem. Soc., 1977, 54, 975, all
from the DETHERM database accessed via Chemical Data Service,
http://cds.dl.ac.uk.
We thank AstraZeneca for a studentship (RAAJH) and Drs
A. M. Davis and N. Gensmantel for their interest and encour-
agement. We are grateful to Dr A. C. Whitwood for assistance
with IT.
32 (a) K. J. Laidler, J. Chem. Soc., 1938, 1786; (b) E. Hertel and J. Dressel,
References and notes
Z. Phys. Chem., 1933, B23, 281.
1 E. J. Behrman and J. O. Edwards, Prog. Phys. Org. Chem., 1967, 4, 93.
2 R. Curci, R. A. DiPrete, J. O. Edwards and G. Modena, J. Org. Chem.,
1970, 35, 740.
3 M. A. P. Dankleff, R. Curci, J. O. Edwards and H.-Y. Pyun, J. Am.
Chem. Soc., 1968, 90, 3209.
4 C. G. Overberger and R. W. Cummins, J. Am. Chem. Soc., 1953, 75,
4783.
5 L. Bateman and K. R. Hargrave, Proc. R. Soc. London, Ser. A, 1954,
224, 389; L. Bateman and K. R. Hargrave, Proc. R. Soc. London, Ser.
A, 1954, 224, 399.
33 SI values from G. Liessmann, W. Schmidt and S. Reiffarth, in Data
Compilation of the Saechsische Olefinwerke, Boehlen, Germany, 1995,
available from DETHERM database accessed via Chemical Data
Service, http://cds.dl.ac.uk.
34 See ref. 30, Table 9–33.
35 (a) J. J. W. McDouall, J. Org. Chem., 1992, 57, 2861; (b) A. G. Baboul,
H. B. Schlegel, M. N. Glukhovtsev and R. D. Bach, J. Comput. Chem.,
1998, 19, 1353; (c) R. A. A. J. Hendrickx, unpublished work.
36 S. Oae, K. Shinhama and Y. H. Kim, Bull. Chem. Soc. Jpn., 1980, 53,
2023.
6 G. Kresze and W. Schramm, Chem. Ber., 1961, 94, 2060.
7 D. Thenraja, P. Subramanian and C. Srinivasan, Tetrahedron, 2002, 58,
4283.
8 C. A. Bunton, H. J. Foroudian and A. Kumar, J. Chem. Soc., Perkin
Trans. 2, 1995, 33.
9 R. Curci and G. Modena, Tetrahedron Lett., 1966, 22, 1235.
10 R. Suthakaran, S. Rajagopal and C. Srinivasan, Tetrahedron, 2001, 57,
1369.
11 G. Modena and L. Maioli, Gazz. Chim. Ital., 1957, 87, 1306.
12 M. Bonchio, S. Campestrini, V. Conti, F. Di Furia and S. Moro,
Tetrahedron, 1996, 51, 12363.
37 P. J. Kropp, G. W. Breton, J. D. Fields, J. C. Tung and B. R. Loomis,
J. Am. Chem. Soc., 2000, 122, 4280 (supplementary information).
38 D. Chianelli, L. Testaferri, M. Tiecco and M. Tingoli, Synthesis, 1982,
475.
39 P. Charlesworth, W. Lee and W. S. Jenks, J. Phys. Chem., 1996, 100,
15152, (ESI†).
40 O. Exner and M. Budesinsky, Magn. Reson. Chem., 1989, 27, 27.
41 V. G. Avramenko, G. N. Pershin, V. D. Nazina, T. N. Zykova, P. Sarma,
Zh. F. Sergeeva, L. B. Rusanova and N. N. Suvurov, Pharm. Chem. J.,
1974, 8, 205.
42 M. P. Hay, S. A. Gamage, M. S. Kovacs, F. B. Pruijn, R. F. Anderson,
A. V. Patterson, W. R. Wilson, J. M. Brown and W. A. Denny, J. Med.
Chem., 2003, 46, 169.
13 G. Modena and A. Cerniani, Gazz. Chim. Ital., 1959, 89, 843.
14 R. W. Murray, Chem. Rev., 1989, 89, 1187.
15 W. Adam, R. Curci and J. O. Edwards, Acc. Chem. Res., 1989, 22, 205.
16 A. L. Baumstark and P. C. Vasquez, J. Org. Chem., 1988, 53, 3437.
17 R. W. Murray and D. L. Shiang, J. Chem. Soc., Perkin Trans. 2, 1990,
349.
18 F. P. Ballistreri, G. A. Tomaselli, R. M. Toscano, M. Bonchio, V. Conte
and F. Di Furia, Tetrahedron Lett., 1994, 35, 8041.
19 R. W. Murray, R. Jeyaraman and M. K. Pillay, J. Org. Chem., 1987, 52,
746.
20 (a) G. Asensio, R. Mello and M. E. Gonza´lez-Nu´n˜ez, Tetrahedron Lett.,
1996, 37, 2299; (b) M. E. Gonza´lez-Nu´n˜ez, R. Mello, J. Royo, J. V. R´ıos
and G. Asensio, J. Am. Chem. Soc., 2002, 124, 9154.
21 J. O. Edwards, R. H. Pater, R. Curci and F. Di Furia, J. Photochem.
Photobiol., 1979, 30, 63.
22 C. Hansch, A. Leo and R. W. Taft, Chem. Rev., 1991, 91, 165.
23 (a) S. Wold and M. Sjo¨stro¨m, Chem. Scr., 1976, 9, 200; (b) S. Wold and
M. Sjo¨stro¨m, in Correlation Analysis in Chemistry, ed. N. B. Chapman
and J. Shorter, Plenum Press, New York and London, 1978, p. 25.
24 S. Ehrenson, R. T. C. Brownlee and R. W. Taft, Prog. Phys. Org. Chem.,
1973, 10, 1.
43 M. E. Heppenstall and S. Smiles, J. Chem. Soc., 1938, 899.
44 D. Landini, G. Modena, F. Montanari and G. Scorrano, J. Am. Chem.
Soc., 1970, 92, 7168.
45 Sung S. Kim, K. Nehru, Sang S. Kim, D. W. Kim and H. C. Jung,
Synthesis, 2002, 2484.
46 L. Cutress, T. B. Grindley, A. R. Katritzky, M. Shome and R. D.
Topsom, J. Chem. Soc., Perkin Trans. 2, 1974, 268.
47 W. Zhu and D. Ma, J. Org. Chem., 2005, 70, 2696.
48 U. Folli and D. Iarossi, Gazz. Chim. Ital., 1969, 99, 1297.
49 X. Creary, A. F. Sky, G. Philips and D. E. Alonso, J. Am. Chem. Soc.,
1993, 115, 7584.
50 F. Ruff and A. Kucsman, J. Chem. Soc., Perkin Trans. 2, 1975, 509.
51 K. Ganapathy and M. Ramanujam, J. Indian Chem. Soc., 1983, 60, 958.
52 J. M
e´tivier (Socie´te´ des Usines Chimiques de Rhone-Poulenc), US Pat.
2803580, 1957.
53 M. H. Ali and S. Stricklin, Synth. Commun., 2006, 1779.
54 A. Lattanzani, S. Piccirillo and A. Scetti, Eur. J. Org. Chem., 2006, 713.
55 J. M. Baskin and Z. Wang, Org. Lett., 2002, 4, 4423, (ESI†).
56 V. G. Shukla, P. D. Salgaonkar and K. G. Akamanchi, J. Org. Chem.,
2003, 68, 5422, (ESI†).
25 (a) O. Exner, Collect. Czech. Chem. Commun., 1966, 31, 3222; (b) J.
Shorter, in Correlation Analysis of Organic Reactivity, Research Studies
Press (Wiley), Chichester, 1982, ch. 7.
57 D. Gilbert, J. Org. Chem., 1963, 28, 1945.
58 M. Peyronneau, M.-T. Boisdon, N. Roques, S. Mazi
e`res and C. Le
26 There are inconsistencies in the data of Murray and co-workers:19 the
quoted values of log(kX/kH)sulfide disagree with the values of (kX/kH)sulfide
given for X = 4-Me and 4-NO2. The reaction constant based on the
quoted logarithms is −0.77; that based on the actual logarithms of the
values of (kX/kH)sulfide is −0.82.
Roux, Eur. J. Org. Chem., 2004, 4636.
59 F. Fringuelli, R. Pellegrino, A. Piermattei and F. Pizzo, Synth.
Commun., 1994, 24, 2665.
60 M. Nose and H. Suzuki, Synthesis, 2002, 1065.
61 W. Adam, J. Bialas and L. Hadjiarapoglou, Chem. Ber., 1991, 124, 2377.
62 K.-P. Zeller, M. Kowallik and P. Haiss, Org. Biomol. Chem., 2005, 3,
2310.
27 r0R measures the resonance component of rp0 and rm0 , which were
determined on systems lacking conjugation between the reaction centre
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 745–761 | 761
©