694
D. HEYES ET AL.
pentadienone.2 Substituents at the 4-phenyl have a
relatively small effect on product distribution with
electron-withdrawing groups favouring the 2H-pyran.
For the new compounds, with substituents at the 2- and 6-
phenyls, the effect on product distribution is similarly
small, and over the whole series ranges only from 11:89
for 8 (X = O) to 29:71 for 1 (X = O), a small shift in
favour of the 2H-pyran induced by electron-withdrawing
groups at either of the phenyls. For the asymmetrically
substituted pyrylium, 10 (X = O), the 4H-pyran: 2H-
pyran ratio is 16:84 with the 84% of 2H-pyran addition
being divided 38:36 between the possible isomers.
Figure 4. Possible conformations and valence isomerisation
of a pyridyl anion
zine18 which adopts a flattened boat conformation with
nitrogen atoms at prow and stern positions. One of the
nitrogen atoms has a planar arrangement of ligands, and
contributes an electron pair to the conjugated array; the
other is pyramidal with an axial substituent so that the
cyclic conjugation is interrupted. Figure 4 shows a
possible structure for the anion, with the methyl group in
an axial position, which would permit delocalization of
the negative charge to all phenylated positions. Semi-
empirical MO calculations (AM1) on the corresponding
N-benzyl anion19 support this picture. Unlike the pyranyl
series, there is no indication of valence isomerization.
We briefly compare the electrochemical behaviour of
this set of cations with those of substituted acetophe-
nones, for which Loufty and Loufty20 have measured
half-wave reduction potentials (also for acetonitrile
solutions). Their series included p-MeO and p-NO2
groups, and correlations with simple Hammett constants
were very poor but notably improved by exclusion of the
data for p-nitroacetophenone. Hammett plots, using
potentials scaled as above, then gave ꢁ = 11.75
(r2 = 0.945), reflecting a much greater demand on the
substituents.
For reactions of the pyryliums with borohydride, the
1,3,5-triphenylcyclopentadienone from ring opening of
the 2H-pyran is further reduced to the corresponding
dienol. This is a slow process compared with the initial
reduction of the cation and did not interfere with rate
measurements. It did, however, complicate measure-
ments of product ratios, especially for the more reactive
salts. For the least reactive pyrylium, 1 (X = O), the 4H-
pyran: 2H-pyran ratio is 3:97. As for the cyanoborohy-
dride reductions, phenyl substituents had only minor
effects on the product ratio, and in all cases the fraction of
4H-pyran is less than 7% (probably considerably less), so
that the more reactive borohydride shows a higher
selectivity for reaction via the 2-position of the cations.
With excess cyanoborohydride in acetonitrile, 2,4,6-
triphenylthiopyrylium, 4 (X = S), itself yields 2H- and
4H-thiopyrans in a 55:45 ratio. There is a similar muted
response of product ratio to substitution in the 4-phenyls,
but electron-withdrawing groups in the 2- and 6-phenyls
induce a more significant shift in favour of the 4H-
thiopyran with 11 (X = S) showing a 4H-pyran: 2H-pyran
ratio of 71:29. With the more reactive borohydride, the
product ratio for 4 (X = S) is unchanged, and the spread
across the whole series is slightly attenuated. Both 4H-
and 2H-thiopyrans are stable to the reaction conditions
with either of the two reagents but may be isolated and
equilibrated by heating in presence of the corresponding
thiopyrylium salt.21 Doddi and Ercolani22 reported
K = 7.1 favouring the 2H-thiopyran (25°C in CHCl3);
we repeated the equilibration and found K = 32 for
reaction in acetonitrile-d3 at 25°C.
The pyridinium salts, all inert to cyanoborohydride, all
reacted with excess borohydride to yield single dihy-
dropyridines, identified from their 1H NMR spectra as the
2H-isomers. For 2,4,6-triphenyl-N-methylpyridinium, 4
(X = NMe), for example, the product shows one-hydro-
gen signals at ꢂ 5.62 (dd, J = 6 and 2 Hz), ꢂ 5.40 (d,
J = 2 Hz) and ꢂ 5.14 (d, J = 6 Hz), readily assigned to
hydrogens at the 5-, 3- and 6-positions of the 1,2-
dihydropyridine 16;23 In the absence of a proton source,
further reduction by borohydride does not occur.24 An
authentic sample of the isomeric 1,4-dihydropyridine 17
was prepared quantitatively by reduction of the salt in
ethanol solution by sodium–mercury amalgam; its
symmetry and structure were confirmed by a two-
REACTION OF CATIONS WITH CYANOBORO-
HYDRIDE AND BOROHYDRIDE
Solutions of sodium cyanoborohydride and sodium
borohydride (ca 0.05 M) in acetonitrile could be prepared.
The borohydride solutions were notably sensitive to
traces of moisture but, provided that oxygen and moisture
were rigorously excluded, showed no sign of decom-
position or loss of reducing power on storage.
Product characterization and stability
The immediate products of reduction were determined
1
by H NMR spectroscopy of solutions of the cations
in CD3CN to which excess reducing agent had been
added.
As noted earlier, 2,4,6-triphenylpyrylium, 4 (X = O), is
reduced by cyanoborohydride in acetonitrile to the
corresponding 4H- and 2H-pyrans in a 25:74 ratio, with
the 2H-pyran suffering ring opening to 1,3,5-triphenyl-
Copyright 2002 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2002; 15: 689–700