∆σ = 212 ppm, whereas the anisotropy for the unhydrated oxide
has 207 ppm, i.e. to the values for the uncomplexed TMPPO
(Fig. 5). The difference between ∆σ (unhydrated minus
hydrated) in the pure oxide is 5 ppm. Complexes with two dis-
tinct phosphorus signals have shown an increasing difference
between the anisotropy values as the pKa of the phenol
decreases (6 ppm for phenol–TMPPO to 21 ppm for 2,5-
dinitrophenol–TMPPO, see Table 3). Comparing the
anisotropy differences for the solid complexes which have two
phosphorus signals with the differences between δiso 31P values,
the higher sensitivity displayed by the former parameters as
compared with their isotropic counterparts becomes more
apparent.
Conclusions
Fig. 5 Phosphorus ∆σ values for the solid-state complexes as a
function of the pKa of the phenols
The hydrogen transfer process has been studied in a number of
complexes between substituted phenols and TMPPO. Proton
and phosphorus-31 NMR spectroscopies have been the chosen
methodologies to monitor the behaviour of the adducts in the
solution state, while high-resolution CPMAS 13C and 31P NMR
have been used to study them in the solid phase. The complexes
show the same variation pattern in the two phases. The acid–
base reaction seems to have a larger extent as the pKa of the
substituted phenol decreases. The evidence supporting this
hypothesis is that the phenolic 1H NMR signal moves to higher
frequencies. The shifting continues until the proton is sub-
stantially transferred from the phenol to the oxide residue
(picric acid complex). On the other hand, δ31P for the complexes
in solution showed the same trend, its value increasing as the
acidity of the phenolic hydrogen increases. The adducts studied
in the solid state showed a reduction of the 31P shielding
anisotropy with the acid strength of the associated phenol,
while 13C NMR results displayed higher shift differences
between phenolic C1 and C4 signals with decreasing pKa values.
difference in δiso is the result of the same effect on every com-
plex, this being the influence exerted by the hydration water
linked to the TMPPO residue in the adduct. To help assess the
validity of this hypothesis, we tried to find supporting evidence
by studying other parameters obtained by measuring the com-
ponents of the phosphorus-31 shielding tensor.
As discussed in earlier work, the 31P shielding tensor com-
ponents (σii), anisotropy (∆σ) and the asymmetry (η)14,23–27 are
highly sensitive parameters that reflect changes in the surround-
ings of the phosphorus nucleus. It should be noticed that
changes in phosphorus-31 δiso values occupy a smaller range
than ∆σ for the same complexes, showing the enhanced sensitiv-
ity of the latter parameter (see Table 3). In all these compounds,
the asymmetry is zero, indicating that all oxide residues are
axially symmetric within experimental error not only in pure
TMPPO but also in the adducts. Spinning sideband manifolds
are, however, insensitive20,28,29 to values of η less than ca. 0.2.
The anisotropies of each type of phosphorus nucleus for pure
TMPPO are 207 and 212 ppm, the former corresponding to the
hydrated moiety and the latter for the residue that is not linked
to water.22 Fig. 5 illustrates ∆σ values of the complexes as a
function of the pKa of the phenol. For complexes giving rise to
two phosphorus signals, both anisotropy values increase as the
pKa of the substituted phenol increases, following the tendency
discussed previously for the isotropic chemical shift. In line
with previous reports,7,30–32 this is the predicted outcome of a
relatively more asymmetric environment of the phosphorus
nucleus. The change in the anisotropy values is controlled by
relative differences among individual π P᎐O bond orders.32
Hence, the described increase of ∆σ accompanies the increase
of π electron density along the involved bond produced by
a smaller contribution of the (TMPPO–H)ϩ structure with
weaker proton donors. Thus, this effect should be the result
of a smaller degree of hydrogen transfer when the substituted
phenol increases its pKa value. Therefore, the principal com-
ponents of the 31P shielding tensor assume values correspond-
ing to a more tetrahedral phosphorus environment as the pKa
of the phenol increases.
Acknowledgements
Robin M. Harrison is gratefully acknowledged for useful com-
ments on crystallographic data. C. M. L. and A. C. O. acknow-
ledge Universidad Nacional de Rosario, the British Council
and Fundación Antorchas for financial support.
References
1 P. Schuster, G. Zundel and C. Sandorfi, The Hydrogen Bond, Recent
Developments in Theory and Experiments, North Holland,
Amsterdam, 1976.
2 T. L. Brown, L. G. Butler, D. Y. Curtin, Y. Hiyama, I. C. Paul and
R. B. Wilson, J. Am. Chem. Soc., 1982, 104, 1172.
3 J. Emsley, Struct. Bonding (Berlin), 1984, 56, 147.
4 F. Herbstein, M. Kapon, G. M. Reisner, M. S. Lehman, R. B. Kress,
R. B. Wilson, W. I. Shiau, E. N. Duesler, I. C. Paul and D. Y. Curtin,
Proc. R. Soc. London, Ser. A, 1985, 399, 295.
5 D. R. Clark, J. Emsley and F. Hibbert, J. Chem. Soc., Chem.
Commun., 1988, 1252.
6 A. J. Vila, C. M. Lagier and A. C. Olivieri, J. Chem. Soc., Perkin
Trans. 2, 1990, 1615.
7 C. M. Lagier and A. C. Olivieri, Solid State NMR, 1994, 3, 163.
8 C. M. Lagier, M. Zuriaga, G. Monti and A. C. Olivieri, J. Phys.
Chem. Solids, 1996, 57, 1183.
9 M. Ilczyszyn, Z. Latajka and H. Ratajczak, Org. Magn. Reson.,
1980, 13, 132.
10 B. Brycki, Z. Szafran and M. Szafran, Pol. J. Chem., 1980, 54, 221.
11 B. Brycki and M. Szafran, J. Chem. Soc., Perkin Trans. 2, 1982,
1333.
Taking into account the crystal structure of TMPPO, in
which the hydrated molecule of the oxide gives a smaller value
of ∆σ, it is reasonable to infer that, in complexes presenting
two different values of anisotropy, the smaller one would
correspond to a hydrated oxide moiety. Of course, ∆σ values
for both hydrated and unhydrated complexes should vary as a
function of pKa of the phenol. This change has to be as dis-
cussed previously, i.e. ∆σ values for the adducts should increase
as the extent of the hydrogen transfer reaction diminishes.
However, anisotropies of hydrated TMPPO residues in the
complexes should tend to a higher value than those of un-
hydrated TMPPO moieties since pure water-linked oxide has
12 B. Brycki and M. Szafran, J. Chem. Soc., Perkin Trans. 2, 1984, 223.
13 T. Keil, B. Brzezinski and G. Zundel, J. Phys. Chem., 1992, 96, 4421.
14 C. M. Lagier, U. Scheler, G. McGeorge, M. G. Sierra, A. C. Olivieri
and R. K. Harris, J. Chem. Soc., Perkin Trans. 2, 1996, 1325.
15 F. Guillan, J. P. Senguin, L. Nadjo, R. Uzan, F. Membrey and J. P.
Doucet, J. Chem. Soc., Perkin Trans. 2, 1984, 1139.
16 M. Ilczyszyn, H. Ratajczak, Z. Latajka and K. Skowronek, Magn.
Reson. Chem., 1988, 26, 445.
J. Chem. Soc., Perkin Trans. 2, 1998
1795