4452 J. Am. Chem. Soc., Vol. 121, No. 18, 1999
Costentin et al.
1
b
solution. Solvated electrons in liquid ammonia, sodium
Scheme 2
1j
1b,d,3
1c
amalgam in the same solvent, light,
electrodes, electro-
generated electron donors,1 ferrous salts, have been used for
g,k
4
this purpose.
However, there are several examples where the reaction works
without external stimulation. They mostly concern Kornblum
reactions1 in which the substrates are benzyl or cumyl deriva-
tives activated by one or more electron-withdrawing substituents
on the phenyl ring (generally nitro groups). A few examples of
a
5
similar reactions have been reported with aromatic iodides.
What is the mechanism of these reactions, the so-called
“
thermal” SRN1 reactions? The only species that might provide
1
a
stimulating electrons to the system is the nucleophile itself.
However, in most cases, the nucleophile is such a poor electron
donor that electron transfer from the nucleophile to the substrate
is expected to be extremely slow, apparently too slow to serve
as a viable initiation step. Previous attempts to discuss this
question have come to the same conclusion even if the values
used for the pertinent standard potentials were rather uncertain
for the reaction with homogeneous electron donors, giving rise
to a slightly uphill reaction. The observation that a concerted
reaction is taking place with poor electron donors such as the
2
-nitropropanate ion thus implies that a change in mechanism
has occurred as a result of the attending change in driving force.
This novel piece of evidence that the driving force controls the
transition between stepwise and concerted mechanisms in
dissociative electron transfers will be discussed in conjunction
with previous observations pointing to the occurrence of the
same phenomenon.
and if the kinetic factors controlling initiation, propagation, and
termination were not precisely taken into account.5a,6 So far,
the reason thermal SRN1 reactions work has thus remained
mysterious.
The work reported below aimed at solving this enigma. Our
strategy was to select one model pair of reactants, identify the
reaction products, measure the reaction kinetics, determine all
of the thermodynamic and kinetic factors of initiation, propaga-
tion, and termination that allow one to predict what should be
the reaction kinetics if the nucleophile were the electron-donor
initiator, and compare the result with experiment. The example
we selected for this purpose was the reaction of 4-nitrocumyl
chloride with the 2-nitropropanate ion. It will be shown that
the results rule out the possibility that initiation could be a simple
outersphere electron transfer from the nucleophile to the
substrate, leading to an anion radical that would cleave in a
successive step. The alternative is a dissociative electron transfer
where the electron transfer from the nucleophile and the breaking
of the carbon-halogen bond in the substrate are concerted. After
estimation of the pertinent thermodynamic and kinetic param-
eters, it will be shown that the predicted kinetics satisfactorily
reproduce the experimental results. Albeit with fewer details
and less precision, we will see that the dissociative electron-
transfer mechanism also explains the results gathered in previous
studies of thermal SRN1 reactions of 4-nitrocumyl chloride with
other nucleophiles. Although, for phenyl iodide, the lack of
pertinent thermodynamic and kinetic data precludes a detailed
analysis of the reaction, it will be shown that the initiation step
cannot be of the outersphere type in this case, too.
Results and Discussion
Reaction of 4-Nitrocumyl Chloride with 2-Nitropropanate
Ion. Reaction Kinetics. The reaction was carried out under
pseudo-first-order conditions (excess of 2-nitropropanate ions)
in acetonitrile at 25 °C, under an argon atmosphere in a light-
protecting vessel. The 2-nitropropanate ion was introduced as
the tetramethylammonium salt. Two products were formed
(
Scheme 2). One is the expected C-substitution product. The
other is an unstable compound which decomposes into the
-nitrocumyl alcohol during workup and may thus be ascribed
4
8
to O-substitution. With 4-nitrobenzyl derivatives, the C-
substitution product is considered to result from the SRN1
reaction and the O-substitution product from a SN2 substitution.
1a
Since steric hindrance at the reacting carbon prevents the SN2
reaction to occur with 4-nitrocumyl chloride, we are led to
conclude that both the C- and the O-substitution products result
9
from a SRN1 reaction.
The concentrations of 4-nitrocumyl chloride and of the two
substitution products vary with time as represented in Figure
1
a. The half-reaction time is 41 min. Repeated runs showed a
good reproducibility of the time variations.
The SRN1 character of the reaction was ascertained by the
effect of light irradiation and of the addition of a radical trap.
Under light irradiation (Figure 1b), the half- reaction time is
considerably shortened (3 instead of 41 min). Addition of di-
tert-butyl nitroxide completely quenched the reaction: neither
the C-substitution product nor the O-substitution product was
observed after 4 h. This last experiment confirms the SRN1
character of the reaction. Since the radical trap may only react
As shown below, the electrochemical reduction of 4-nitro-
cumyl chloride, a slightly uphill process, follows a stepwise
7
mechanism, as does 4-nitrobenzyl chloride. The same is true
(3) (a) Several possibilities have been envisaged for the mechanism of
light initiation, in particular, the excitation of a nucleophile-charge-transfer
3b,c
complex. In the reaction of thiolates with 1-iodoadamantane, the initiating
3
d
electron is photoejected from the thiolate ion. (b) Hoz, S.; Bunnett, J. F.
J. Am. Chem. Soc. 1977, 99, 4690. (c) Fox, M. A.; Yonnathan, J.; Fryxell
J. Org. Chem. 1983, 48, 3109. (d) Ahbala, M.; Hapiot, P.; Houmam, A.;
Jouini, M.; Pinson, J.; Sav e´ ant, J.-M. J. Am. Chem. Soc. 1995, 117, 11488.
•
with the R radicals that have escaped the solvent cage where
•
•
-
R , Nu , and X have been formed, this experiment also indicates
(
(
4) Galli, C.; Gentili. J. Chem. Soc., Perkin Trans. 2 1993, 1135.
5) (a) Kim, J. K.; Bunnett, J. F. J. Am. Chem. Soc. 1970, 92, 7463. (b)
(8) (a) The O-substitution product is also obtained, albeit in lesser yield,
upon reaction with the lithium salt of 2-nitropropanate.8b Ion pairing of the
negative end of the 2-nitropropanate ion by the countercation is expected
Scamehorn, R. G.; Bunnett, J. F. J. Org. Chem. 1977, 42, 1449. (c) Swarz,
J. E.; Bunnett, J. F. J. Org. Chem. 1979, 44, 340. (d) Scamehorn, R. G.;
Hardacre, J. M.; Lukanich, J. M.; Sharpe, L. R. J. Org. Chem. 1984, 49,
+
+
to be stronger with Li than with (CH3)4N , thus explaining why the
percentage of O-alkylation is less in the first case than in the second. (b)
Kornblum, N.; Davies, T. M.; Earl, G. W.; Holy, N. L.; Manthey, J. W.;
Musser, M. T.; Swiger, R. T. J. Am. Chem. Soc. 1968, 90, 6219.
(9) Ambident reactivity of nucleophiles is not unprecedented. For
examples in aromatic SRN1 reactions see ref 1k and references therein.
4
881.
6) (a) Eberson, L. J. Mol. Catal. 1983, 20, 27. (b) Eberson, L. Acta
Chem. Scand. B 1984, 38, 439.
7) Andrieux, C. P.; Le Gorande, A.; Sav e´ ant, J.-M. J. Am. Chem. Soc.
992, 114, 6892.
(
(
1