2
40
REACTIONS OF LITHIUM WITH BENZALDEHYDE
is clearly seen to depend on the bulk concentration and for
this reason [1] was kept mostly within the range
stoichiometric amount of lithium wire, as previously
20
described for the lithium benzophenone ketyl. The concen-
trations of the organolithium derivatives and of the
independently prepared lithium ketyls were determined both
by the double titration technique using ethylene 1,2-dibro-
0
·20–0·25
M.
According to the observed results, adsorption of 1 would
seem to be a prerequisite for rapid electron transfer (as in
electrocayalsis) and is probably responsible for the ‘lag’ or
induction period observed in the formation of 2. Consistent
with this is the fact that the ‘sigmoid’ behaviour is more
pronounced at lower temperatures, with a reduction in the
13
mide and by reaction with diphenylacetic acid.
All glassware, syringes and needles were dried in a
vacuum oven and cooled in a desiccator; they were flushed
with dry nitrogen immediately before use.
29
‘
lag’ as the temperature increases. Nuzzo and Dubois have
studied the reaction of methyl bromide with magnesium
Reactions of 1 with lithium. Lithium wire was weighed
under ligroin, washed with hexane (then, for reactions in
THF, with THF) and cut into small pieces in the reaction
flask containing a small portion of solvent, under a stream
of dry argon. The reaction flask was capped with an ‘air-
tight’ stopper, and alternatively evacuated and flushed with
nitrogen several times. Reagents and solvents were trans-
ferred by syringes using techniques described for the
single crystals and found that CH Br is both physisorbed
3
and dissociatively chemisorbed at Ϫ150 °C; the results are
not altered by coadsorption with dimethyl ether. The
chemisorbed species desorbs at a temperature as low as
3
0
Ϫ130 °C.
Some of the important remaining questions concern the
nature of the transition state for the step(s) leading to
radicals and the nature of the steps leading from them to the
products. Some hypotheses are (1) that 1→5 occurs at the
lithium surface, (2) that 5 is not adsorbed at the surface but
instead diffuses freely in solution, where it can dimerize or
react with solvent or with other radicals or radical anions,
and (3) that the transformation of 5 into products (through
several steps) is initiated at the lithium surface and has a rate
that is proportional to the concentration of 5 at the surface.
Although controversial, the preferred mechanism at present
for Grignard reagent formation favors the interpretation in
35
manipulation of air-sensitive compounds. The products of
the reaction of 1 were isolated and characterized as reported
33
previously for similar reactions. The reactions of 1 with
lithium wire in the presence of CO were carried out by the
32
general procedure reported previously for phenyllithium.
The amounts of 1 on the lithium surface were determined by
difference between [1] and the [1] found in solution when
0
no reaction was observed; in the other cases a mass balance
was applied, taking into account the amount of 1 reacted.
Samples of the reaction mixture were quenched with acetic
3
1
which intermediate alkyl radical do not remain adsorbed.
anhydride or with a saturated solution of NH Cl, and
4
•
The present results seem also to indicate that PhCO does
analyzed by GC, using an OV-101 column (50–270 °C
temperature gradient).
not remain adsorbed.
EPR spectra. A Bruker ER 200D (X band) TE102 cavity
EXPERIMENTAL
was used for recording EPR spectra, using procedures
similar to those described for the EPR and NMR determina-
32
20, 21
Materials. THF was purified as previously described.
tions of the reaction of PhLi with CO.
The
Benzene was distilled over sodium wire and then refluxed
over lithium benzophenone ketyl and distilled immediately
prior to use. Hexane was purified by refluxing with
concentrated sulfuric acid for 2 h, then distilled and stored
over sodium hydroxide pellets; it was distilled over lithium
benzophenone ketyl immediately prior to use. Carbon
tetrachloride (Carlo Erba) was passed through an alumina
column and distilled. Solvents were stored under argon in
special vessels that allow delivery without air contamina-
tion. Benzaldehyde (UCB) was distilled at reduced pressure
under nitrogen and stored under dry argon. Carbon
monoxide was generated from the reaction of sulfuric acid
independently prepared radical anion solutions or the
reaction mixtures in preparative concentrations (0·2–1·5
M)
Ϫ5
were diluted to nearly 10
M
with THF. Optimization of the
final concentration was adjusted in each case. The tem-
perature was kept at 0 °C using a Dewar cavity with a
thermostated stream of air; measurements at higher tem-
peratures (30–60 °C) showed a fast decrease in the intensity
of the signals. Calibration was carried out at different
modulation amplitudes using a solution of galvinoxyl in
THF.
Kinetic measurements. The reaction was followed by taking
0·2 ml aliquots of the reaction mixture at time intervals,
quenching with 1 ml of acetic anhydride and analyzing by
GC after 48 h, using benzophenone as an internal standard.
The delay allowed complete esterification. Samples of
independently prepared standard references were treated
similarly and the GC response factors determined in each
case. Runs were made at least in triplicate and the
reproducibility was satisfactory. In the case of quenching
33
with 98% formic acid and treated as described previously.
Benzophenone (Fluka) was 99% pure and was used after
recrystallization from ethanol. Benzil and benzoin were
identified spectroscopically and by their melting points
against samples independently prepared by reported meth-
3
2
ods, and crystallized from ethanol, m.p. 94–95 °C (lit.
9
34
4–95 °C) and 135–137 °C) (lit. 135–137 °C), respec-
tively. The lithium benzyl ketyl was prepared by treatment
of the benzil in THF solution with a slight excess over the
with a saturated solution of NH Cl, 0·1 ml was used.
4
©
1997 by John Wiley & Sons, Ltd.
JOURNAL OF PHYSICAL ORGANIC CHEMISTRY, VOL. 10, 233–241 (1997)