and DMSO-d6 (DMSO-d6 (0.1 M) shows no H-D ex-
change with liquid ammonia at room temperature after
1
1
Scheme 1
several weeks, monitored by H NMR). H NMR and
DEPT 135 spectra of dimedone in liquid ammonia show
only one proton attached to carbon 2 between the two
carbonyl groups, and the carbon shift of the carbonyl
carbons in liquid ammonia is about 10 ppm lower than
that at 200 ppm in CDCl3.11 Diethyl malonate, which has
an aqueous pKa of 12.912 and one of 15.9 in DMSO,13 is
not ionized in liquid ammonia but shows a broad single
peak of the central methylene protons, probably due to a
fast exchange of these protons with the solvent. Acetyla-
cetone and 2-acetylcyclohexanone with aqueous pKas of
9.0 and 10.1, respectively,12 (13.313and 15.514 respectively
in DMSO) react in liquid ammonia to give their corre-
sponding enamines, confirmed by GC-MS analysis.
Monocyano-activated carbon acids, such as benzyl cya-
nide with a pKa of 21.9 in DMSO,15 are not ionized in
liquid ammonia. Unlike other solvents, liquid ammonia
causes the relaxation time of the methylene protons of
benzyl cyanide to be fast compared with the aromatic
protons and, consequently, it takes 10 s relaxation time
to obtain the expected integration ratio between aromatic
protons and methylene protons in liquid ammonia.
indicating that there is less than 1% MDN with hydrogen
1
still attached. There are no H NMR signals associated
with the methylene carbon of MDN at ꢀ40 °C so if this is
due to a rapid exchangemechanism leading to a very broad
signal that is not observable, this lower temperature is
insufficient to slow the rate of exchange. Ionization of
MDN would transfer the protons to ammonia to form
ammonium ions but the addition of ammonium chloride
does not affect the 1H NMR spectrum. Of course, within
the ion-pair (eq 1) added ammonium ions would not alter
the equilibrium with the undissociated species.
The 13C NMR spectrum of MDN in liquid ammonia at
25 °C shows a high field carbon signal at ꢀ3.12 ppm,
consistent with the formation of a negatively charged
carbon, and contrasts with the methylene carbon signal
of 8.4 ppm for MDN in DMSO. These observations are
not due to a chemical reaction of MDN which is stable in
liquid ammoniaꢀafter two days at room temperature,
evaporation of the ammonia, acid neutralization and
extraction yields unreacted MDN. The carbon chemical
shifts of MDN with 2 equiv of NaH in liquid ammonia are
similar to those without the added base. Although it is not
clear whether MDN actually forms a carbon dianion in
liquid ammonia or simply forms a monoanion with un-
usual NMR properties, the signal at ꢀ3.12 ppm is signifi-
cantly further upfield compared with that of 11.5 ppm of
benzylmalonodinitrile monoanion in liquid ammonia.
The carbon shift of MDN in liquid ammonia is upfield,
but this could be due to a poorly solvated monoanion
leading to a relatively larger negative charge density on the
central carbon in liquid ammonia compared with other
solvents. Earlier work on NMR data for the ionization of
MDN is inconclusive, one17 reports the monoanion shift in
DMSO as ꢀ0.25 ppm, but another18 quotes the mono-
anion shift in HMPT as ꢀ2.1 and ꢀ2.2 ppm, but then gives
as shift of þ32 ppm for the dianion. Very highfield carbon
shifts are seen for carbon dianions or quasi-dianions in
other solvents, for example, carbon suboxide has a very
negative carbon shift of ꢀ14.6 ppm in CDCl3 at ꢀ40 °C,19
while iminopropadienones have shifts ofꢀ3.8 toꢀ6.8 ppm
in CDCl3 at room temperature.20 It has been proposed that
the resonance stabilization of R-cyano carbanions is not
Dicyano derivatives are, however, fully ionized in liquid
1
ammonia; the H NMR spectrum of benzylmalonodini-
trile does not show a proton attached to the methine
carbon and the 13C chemical shift is upfield (11.5 ppm)
for the central carbon compared with that in CDCl3, as
expected from an increased negative charge density in the
carbanion. Ionization in liquid ammonia is further sup-
ported by DEPT 135 spectra, which show no coupling
between the methine carbon and its attached proton pre-
sent in neutral benzylmalonodinitrile. The surprisingly
large downfield 13C shift of the cyano groups (145.1 ppm)
is similar to the 144.3 ppm reported for the cyano group of
the lithium salt of benzyl cyanide anion in THF.16
Malonodinitrile, MDN, with an aqueous pKa of 11.212
(11.1 in DMSO13) shows very unusual behavior in liquid
ammonia. In stark contrast to that observed in other
solvents, its 1H NMR spectrum at 25 °C shows no protons
attached to the central methylene carbon present in neutral
MDN. It would be surprising if both protons of MDN
have been removed to form a carbon dianion, but the lack
of a proton signal is not due to H-D exchange with the
deuteriated dimethyl sulfoxide used to “lock” the spectro-
meter as the same result is observed with toluene-d8 or
benzene-d6 as a lock. A similar spectrum is also seen when
one equivalent of acetonitrile is added using its three
methyl hydrogens as an internal standard, apparently
(12) Albert, A.; Serjeant, E. P. The Determination of Ionization
Constants: A Laboratory Manual, 3rd ed.; Chapman and Hall: London,
1984.
(13) Bordwell, F. G.; Harrelson, J. A., Jr.; Satish, A. V. J. Org. Chem.
1989, 54, 3101–3105.
(14) Kern, J. M.; Federlin, P. Tetrahedron 1978, 34, 661–670.
(15) Bordwell, F. G.; Bares, J. E; Bartmess, J. E.; McCollum, G. J.;
Van der Puy, M.; Vanier, N. R.; Matthews, W. S. J. Org. Chem. 1977, 42,
321–325.
(17) Abbotto, A.; Bradamante, S.; Pagani, G. A. J. Org. Chem. 1993,
58, 449–455.
(18) Vogt, H.-H.; Gompper, R. Chem. Ber. 1981, 114, 2884–2897.
(19) Williams, E. A.; Cargioli, J. D.; Ewo, A. Chem. Commun. 1975,
366–367.
(16) Crowley, P. J.; Leach, M. R.; Meth-Cohn, O.; Wakefield, B. J.
Tetrahedron Lett. 1986, 27, 2909–2912.
(20) Bibas, H.; Moloney, D. W. J.; Neumann, R.; Shtaiwi, M.;
Bernhardt, P. V.; Wentrup, C. J. Org. Chem. 2002, 67, 2619–2631.
Org. Lett., Vol. 13, No. 22, 2011
6119