Evidence That the α-Effects in Methyl Transfers from Aryldimethylsulfonium Salts Correlate with Single-Electron-Transfer Characteristics

Article abstract of DOI:10.1021/jo9606012

The ability of the anions of N-methylbenzohydroxamic acids to demonstrate the α-effect in methyl transfers from aryldimethylsulfonium salts correlates with the ability of these salts to demonstrate single-electron (SET) character in electrochemical experi

Full text of DOI:10.1021/jo9606012

J . Org. Chem. 1997, 62, 4795-4797
4795
Evid en ce Th a t th e r-Effects in Meth yl Tr a n sfer s fr om
Ar yld im eth ylsu lfon iu m Sa lts Cor r ela te w ith
Sin gle-Electr on -Tr a n sfer Ch a r a cter istics
K. R. Fountain* and Kamlesh D. Patel
Truman State University, Kirksville, Missouri 63501
Received April 1, 1996X
The ability of the anions of N-methylbenzohydroxamic acids to demonstrate the R-effect in methyl
transfers from aryldimethylsulfonium salts correlates with the ability of these salts to demonstrate
single-electron (SET) character in electrochemical experiments.
In tr od u ction
Recently, attention has been refocused on the role of
single-electron-transfer (SET) character in wave func-
tions of SN2 transition states, as predicted by Shaik et.
al.1 Examples of experimental approaches used to dem-
onstrate such SET character in an SN2 reaction include
an early report by Ashby3a where radicaloid intermedi-
ates in reactions putatively typical of SN2 classical
reactions exhibited large amounts of inversion of config-
F igu r e 1. Hoz model for the R-effect.
uration. A more recent demonstration of this effect was
reported in the identity reaction2 of bromide ion with
< phenyl (i.e., the system takes more character of the
top reaction in eq 1 in the phenyl case).
p-BrC6H4COCH2Br in cumene.3b This reaction gave
results consistent with disclosing the SET character by
trapping radicals in the identity reaction because sub-
stantial yields of p-BrC6H4COCH3 were obtained. The
scheme that rationalizes this behavior involves abstract-
ing a hydrogen atom from cumene, present in large excess
as the solvent, by the Br• radical component.
Factors that promote greater concertedness in this
sequence (synchronous capture of the single electron and
expulsion of the group) of alkylating agents include
weakening of the bond between S and the departing
group, implying that the phenyl S+-CH3 bond is the
weakest in the series. We will show below that this is
not decisively so. An alternative to this bond weakening
would be formation of a bond to the departing group by
another nucleophile, such as in an SN2 reaction. Such a
possibility would lower the energy necessary to expel a
group from the S atom. This set of salts allows us to
study nucleophilic attack and to form conclusions about
the degree of SET character in methyl transfers when
RdCH3.
An alternate experimental design is possible where the
ability of a substrate to accept an electron is correlated
with a nucleophile having, at least theoretically, the
possibility of donating an electron. Aryldimethylsulfo-
nium salts have recently been shown to be substrates
capable of receiving various degrees of SET character in
electrochemical experiments.4,14 The order of increasing
ability to undergo SET in these substrates, eq 1, is aryl
) phenyl < 1-naphthyl < 9-anthracenyl. Although all
It should be possible to investigate SET transfer from
nucleophiles to such substrates. The significance of this
strategy is that a theoretical model for the R-effect can
be tested. Hoz proposed that the R-effect, displayed by
nucleophiles possessing a pair of electrons neighboring
the reacting pair (as in hydrazine) reacting at the CdO
group, results from SET character in the wave function
of the transition state10 (TS), Figure 1. The wave function
of the TS includes this SET character because the
transfer of some SET character from an interaction of
the four electrons in the R-nucleophile would produce
stabilization of the three-electron character. The result-
ing three-electron splitting contribution is mixed into the
wavefunction of the TS in a stabilizing way that accounts
for a diminished Eact for this reaction.
the experiments are actually EC type, where electron
capture (E) precedes expulsion of a radical (C), the timing
of the E steps and the C steps becomes closer (concert-
edness of the capture of the electron and the expulsion
of the radical) in the order 9-anthracenyl < 1-naphthyl
The R-effect results in increased nucleophilic reactivity
compared to a nucleophile having no R-electrons but of
the same pKa value as the R-nucleophile.5 Buncel as well
as our group have both determined that R-effects occur
in methyl transfers from sulfates6 and sulfonates7 with
* To whom correspondence should be addressed. Phone: (816) 785-
4633. FAX: (816) 785-4045. E-mail: SC18@truman.edu.
X Abstract published in Advance ACS Abstracts, J une 15, 1997.
(1) Shaik, H.; Schlegel, B.; Wolfe, S. Theoretical Aspects of Physical
Organic Chemistry: The SN2 Mechanism; J ohn Wiley & Sons, Inc.:
New York, 1992; Chapter 4.
(2) See ref 1, Chapter 5, for a discussion of identity reactions.
(3) (a) Ashby, E. C.; Pham, T. N. Tetrahedron Lett., 1987, 28, 3183.
(b) Haberfield, P. J . Am. Chem. Soc. 1995, 117, 3314.
(4) Andrieux, C. P.; Robert, M.; Saeva, F. D.; Save´ant, J .-M. J . Am.
Chem. Soc. 1994, 116, 7864.
(5) Hoz, S.; Buncel, E. Isr. J . Chem. 1985, 26, 313.
(6) (a) Buncel, E.; Chuaqui, C. J . Org. Chem. 1980, 45, 3621. (b)
Buncel, E.; Wilson, H.; Chuaqui, C. J . Am. Chem. Soc. 1982, 104, 4896.
S0022-3263(96)00601-9 CCC: $14.00 © 1997 American Chemical Society

4796 J . Org. Chem., Vol. 62, No. 14, 1997
Fountain and Patel
Ta ble 1. r-Effects in th e Ar S+Me2 System in Meth a n ol
by Com p etition Exp er im en ts a t 29.5 °C
Ta ble 2. MCA (k ca l/m ol) a n d MRA Va lu es for th e Ar yl
Meth yl Th ioeth er s
aryl
phenyl
1-naphthyl
9-anthracenyl
R-effect
PhSMe 1-naphthylSMe 9-anthracenylSMe
13.3 ( 2.5
32.5 (12.0
68.4 ( 14.0
MCA PM3
-99.11
-99.5
-111.1
-20.77
-103.7
-78.44
-31.84
3-21G* -53.3
-18.12
MRA
dilute gas-phase reaction (eq 3).13 The results are sum-
marized in Table 2.
hydrogen peroxide anions, N-methylbenzohydroxamate
anions (NMBH), and N-phenylhydroxylamines (NPHA).
Typical ranges for these R-effects are 2.5-11, expressed
BMe+ h B + Me+
(3)
as a ratio of rate constants, kR-nuc/knuc
.
We now report that the R-effect demonstrated in
methyl transfers from aryldimethylsulfonium salts in eq
1 with NMBH anions (reaction 2) correlates with the
ability of the substrates to accept an electron.
Discu ssion
The sizes of the R-effects in Table 1 follow the order of
the ability of the aryldimethylsulfonium salts to accept
an electron, as demonstrated earlier by Save´ant4 et. al.
The R-effect with the phenyldimethylsulfonium salt is
very similar to the R-effect reported by us, determined
by direct kinetic studies.9 In all these cases, in Save´ant’s
study, the capture of a single electron from an electrode
was part of an EC process. In an EC process the rate-
limiting step is a first-order chemical step (“C”) following
a fast electron-transfer step (“E”). From theoretical and
experimental electrochemical criteria it was possible for
Save´ant to evaluate the degree of concertedness of these
two processes. The rates of the two process in eq 1
approach each other as the aryl system becomes smaller.
The aryl ) phenyl was most concerted case, followed by
the 1-naphthyl case and then the 9-anthracenyl case. All
of the cases deal with EC mechanisms, but with varying
degrees of mixture of the E and C processes. The single
electron is known to enter an S-C σ* orbital and not the
π* system of the aryl ring.4,14
The correlation of the sizes of the R-effects in the table
with the order of ease of the E process and, inversely,
with the degree of concerted EC reactivity in the series
indicates the following: 1. The R-effect in methyl transfer
involves SET character in the transition state. 2. The
degree of concertedness of an EC process controls the size
of the R-effect in SN2 reactions.
We have previously discussed the R-effect at carbon
atoms as involving an intrinsic effect, shown at the most
simple carbon atom in CH3 groups.7a This effect is
modified by substitutions of other atoms or groups on the
C atom undergoing nucleophilic attack, such as phenyl
(giving a benzyl), or dO, giving the CdO group. These
groups give larger R-effects than at methyl due to their
ability to disperse the extra charge.7a This argument is
similar to the one Save´ant advanced in explaining the
increased ability of the 9-anthracenyl salts to stabilize
SET character in the sulfonium salts. The extra charge
is dispersed into the more extensive π system in those
systems. The present results with the N-methylbenzo-
hydroxamate transition states are entirely consistent
with that idea.
Exp er im en ta l Section
The experiments were performed as a series of competition
experiments using a 10 M excess of both p-ClC6H4CON(Me)-
O- (p-ClNMBH; pKa ) 12.04 in MeOH7a) and 3-NO2C6
H4O-Na+ (pKa ) 12.44 in MeOH8) in a thermostated water
bath at 29.5 °C in reaction with a 1.0 M amount of the
sulfonium salts. These salts were prepared as reported.3 The
reactions were allowed to stand for 1 h and were then analyzed
using a Hewlett-Packard 5890 GCMS employing a 12 ft cross-
linked methyl silicone gum column. Authentic materials
identified the retention times and gave authentic mass spectra
for comparison of products. The areas were digitally inte-
grated. Each determination was carried out in, at least,
duplicate. The GCMS has been shown previously to be nearly
linear in relating areas to grams.7a The ratios of the areas
were then multiplied by the ratios of the molecular weights
to give the R-effects as kR-nuc/knuc
.
The R-effects are summarized in Table 1.
Computational chemistry for the methyl cation affinities
was performed on a Gateway P166 using either Hyperchem
4.5 or 5.0. Geometries were obtained at the PM3 and 3-21G*
levels of theory. Computation of all the vibrational spectra
showed all positive frequencies and, hence, energy minima for
all structures. Computations at the ab initio 3-21G* level were
performed with Gaussian 94 on an SGI Indy or G94W on the
Gateway P166. The methyl radical affinities were defined
analogously to the methyl cation affinities as in reaction 3 and
are at the PM3 (UHF) level of theory. The issue of local vs
global minima was settled by rotating the SMe2 group around
the aryl-S bond in various possible conformations until the
lowest energy was obtained; in all cases, this conformation was
one in which at least one S-CH3 bond was nearly perpen-
dicular to the plane of the aryl ring.
We have previously reported that the R-effect in the
substituted phenyldimethylsulfonium series is most prob-
ably electronic in origin.9 The application of the tool of
increasing electron demand in a series of ring substituted
N-methylbenzohydroxamate anions gave a Bro¨nsted-type
plot that intersected the plot of the substituted phenolate
anions (non-R-nucleophiles). The present correlation of
the size of the R-effect with increasing ease of SET
reception combined with the behavior these sulfonium
Resu lts
The methyl cation affinities (MCA)13 and methyl radi-
cal affinities (MRA) were computed from the ∆Hf pre-
dicted by the quantum calculations using the ∆H for the
(7) (a) Fountain, K. R.; Fountain, D. P.; Michaels, B.; Meyers, D.
B.; Salmon, J . K.; Van Galen, D. A.; Yu, P. Can. J . Chem. 1991, 69,
798. (b) Fountain, K. R.; Hutchinson, L. K.; Mulhearn, D. C.; Xu, Y.
B. J . Org. Chem. 1993, 58, 7883.
(9) Fountain, K. R.; Dunkin, T. W.; Patel, K. D. J . Org. Chem. 1997,
62, 2738.
(8) Rochester, C. H.; Rossall, B. Trans. Faraday Soc. 1969, 65, 1005.

Methyl Transfers from Aryldimethylsulfonium Salts
J . Org. Chem., Vol. 62, No. 14, 1997 4797
salts on increasing electron demand in the nucleophiles
indicates that the R-effect, in this system at least, is an
intrinsic electronic one, and it involves some SET behav-
ior as well.
The expulsion of a methyl group, either as a cation or
a radical, should reflect the ability of the S atoms in the
series of compounds to hold on to these two kinds of
methyl groups. This ability should be reflected in the
methyl cation affinities of the series of methyl aryl
thioethers or in the (analogously defined) methyl radical
affinities. The methyl aryl thioether best able to receive
a methyl cation or a methyl radical should be the worst
at giving it up to a nucleophile if the thermodynamic
argument holds. We computed the methyl cation affini-
ties at two different levels of theory (PM3 and 3-21G*)
for the series phenyl, naphthyl, and 9-anthracenyl methyl
thio ethers. The results are summarized in Table 2, and
they demonstrate that, at least at these levels of theories,
the trends for these parameters do not correlate the
experimental data. The size of the R-effect increases
regularly in the order the phenyl < naphthyl < 9-an-
thracenyl. Admittedly, the 3-21G* MCA values predict
the 9-anthracenyl should give a larger effect than the
1-naphthyl, but it should give a much smaller R-effect
than the phenyl case if the thermodynamic argument
holds. The MRA affinities are in the completely reverse
order, predicting that the 9-anthracenyl case would be
worst at transferring a methyl radical. These data
indicate that our conclusions that substantial SET char-
acter is involved in the R-effect are at least sound at these
levels of theory.
These conclusions support the Shaik1 idea of inclusion
of some SET character into the transition-state wave
function of even an SN2 methyl group transfer. The
finding of correlation of the increasing size of the R-effect
with increasing ease of acceptance of SET character
supports the Hoz model of inclusion of some three-
electron splitting character into the TS wave function,
even in methyl group transfers. Previously the Hoz
model was meant only to refer to the R-effects at groups
such as >CdO.10
Possible alternative explanations for our observations
deserve mention here. A reviewer has suggested that
when the reaction is placed on a More O’Farrall-J encks
diagram a reasonable deduction is that the R-effect has
nothing to do with the electronic nature of the nucleophile
but reflects an increase in the bond formation between
the C atom and the nucleophile at the transition state.
The transition state changes in the direction of increased
bond function, as suggested by Dixon and Bruice.11
J encks has also recently pointed to a possible thermo-
dynamic origin for the R-effect in nucleophilic attack at
phosphorus.12 Attacking these alternatives is not an easy
prospect. We offer the following approach.
(10) Hoz, S. J . Org. Chem. 1982, 47, 3545.
(11) Bruice, T. S.; Dixon, J . E. J . Am. Chem. Soc. 1971, 93, 3248,
6592.
(12) Herschlag, D.; J encks, W. P. J . Am. Chem. Soc. 1990, 112, 1951.
(13) McMahon, T. B.; Heinis, T.; Nichol, G.; Hovey, J . K.; Kebarle,
P. J . Am. Chem. Soc. 1988, 110, 7591.
Ack n ow led gm en t . We thank the NSF for a RUI
grant (No. CHE9626663) for support of this work. We
are grateful to the administration of Truman State
University for an internal grant in support of this work.
(14) Saeva, F. D.; Brestin, D. T.; Martic, P. A. J . Am. Chem. Soc.
1989, 111, 1328.
J O9606012