125
TaTe . In accordance, the Te NMR lines shift from δ 3400
2
(
VTe ) to 2600 (TaTe ). Again, the exact degrees of transfer
2
2
have not been calculated in NbTe2 or TaTe2 but they are
expected to be between those of ZrTe and VTe . Values of
2
2
Ϫ1.85 and Ϫ1.88 for the tellurium oxidation state in NbTe and
2
TaTe , respectively, are derived from Fig. 7.
2
Consider now the Group 9 transition-metal ditellurides. For
IrTe a tellurium oxidation state of Ϫ1.5 corresponds to a res-
2
onance at δ 4750, clearly deshielded relatively to the tellurium
signal of TiTe . The case of rhodium ditelluride is somewhat
2
different since the stoichiometric material does not exist. The
rhodium-poor phase is Rh1.15Te and the oxidation state is cal-
2
culated at Ϫ1.7. Such difference in oxidation state is consistent
with the 440 ppm paramagnetic shift observed between
Rh1.15Te and IrTe . In PtTe and PdTe the oxidation state is
125
2
2
2
2
Fig. 7 Plot of the Te NMR shift in transition-metal ditellurides vs.
tellurium oxidation state: , oxidation state calculated from extended-
Hückel and physical measurements; ᭺, interpolated. The dashed line
represents the estimated chemical shift contribution (see text)
expected to be Ϫ1 and the compounds display the most
deshielded resonances of the series. However their shift differ-
ence of almost 2000 ppm is totally unexpected.
1
25
Except for PtTe and PdTe , the relative shifts of Te NMR
2
2
lines are in good qualitative agreement with the concept of
Te→M electron transfer. Further discussion of the NMR
results would require the separation of CS and KS contribu-
tions. As equation (2) shows, the KS is proportional to the
electron density at the Fermi level. Band-structure calculations
performed on transition-metal tellurides have shown that the
orbitals participating at the Fermi level are essentially
on the nature and geometry of bonding electrons and on the
s character of the bonding orbitals. Owing to the latter, the
magnitude of the J interaction increases periodically with the
atomic number of the nuclei involved, in the same way the CS
125
195
1
range does. As an example, very large Te᎐ Pt J couplings
2
Ϫ
(
ca. 6 kHz) have been reported for [(PtCl ) TeMe ] in solu-
3 2 2
38 125 199
tion. In the solid state, Te᎐ Hg anisotropic J couplings of
ca. 7 kHz have been observed in the static Te NMR spectrum
of HgTe. Finally, conduction electrons may also be respon-
sible for another type of indirect coupling, for which the theory
5,6
anionic. Hence the KS should be correlated with the tellur-
ium oxidation state. On the other hand, an order of magnitude
for the possible influence of tellurium oxidation state variations
on the CS can be estimated from solution studies. For poly-
chalcogenide anions in solution Björgvinsson and Schrol-
125
39
40
predicts very long-range effects.
35
bingen have found that the resonance positions of species
125
Correlation between the Te total shift and the tellurium
oxidation state
ϪII
ϪII
ϪII
such as Te 2, Te
and Te
show a relative paramagnetic
3
4
shift of ca. 2000 ppm per electron. Assuming a linear depend-
ence between CS and tellurium oxidation state, the line drawn
in Fig. 7 corresponds to the estimated CS contribution for the
total NMR shift observed with transition-metal ditellurides.
The compounds WTe and HfTe which have the smallest elec-
125
We now discuss the relation between the Te NMR data and
the sp(Te) → d(M) electron transfer in transition-metal
binary ditellurides with a CdI -related structure. In practice,
2
this means attempting to correlate the NMR shift with the
tellurium oxidation state (from Ϫ to Ϫ in our compounds).
As far as the CS interaction is concerned, the oxidation of
tellurium should result in a progressive paramagnetic shift
through a deshielding effect. Also the KS is expected to vary as
a function of the characteristics of the Fermi surface of the
different compounds.
2
2
tron transfer and thus an oxidation state of Ϫ provide a refer-
ence for the CS. The distances from each point to the model
line give a crude estimation of the KS contribution. For
example, a similar order of magnitude (ca. 2500 ppm) is
obtained for the KS in TiTe , Rh Te , IrTe and PdTe while
2
1.15
2
2
2
the paramagnetic shift of PtTe must be explained by a signifi-
2
125
125
Fig. 7 shows the relation between Te NMR shifts and tel-
lurium oxidation states. The general trend is an increase of the
NMR shift when the tellurium oxidation state increases,
cantly larger KS (ca. 4500 ppm). Such high Te KS values are
11
consistent with those generally reported for heavy nuclei. In
our case, they should be related to the s-electron density on the
tellurium nuclei.
4ϩ
Ϫ2
2ϩ
Ϫ1
ranging from δ 550 for W Te to δ 7400 for Pt Te 2. On the
2
low-frequency side, the WTe and HfTe resonances are found
2
2
around δ 600 and 900. Their oxidation state is close to or
exactly Ϫ. The relative paramagnetic shift observed from
HfTe2 to ZrTe2 (at ca. 900 and 1800 ppm, respectively) is
consistent with an increased overlap between the p- and d-block
bands. Although the degree of the sp(Te) → d(Zr) electron
Conclusion
125
We have shown that solid-state Te NMR spectroscopy is a
sensitive tool to probe the local environment of tellurium in a
wide range of compounds. Our study of commercially available
tellurium compounds and transition-metal tellurides together
with the previously reported work allowed us to build up a
NMR shift database on this poorly studied nucleus. The Te
NMR shifts cover a large range of ca. 10 000 ppm versus tellu-
rium chemical environments, oxidation states and electronic
properties. The transition-metal tellurides resonate on the high-
frequency side of this scale. This observation has been attrib-
uted to the KS occurring for these conducting samples. In the
transition-metal ditelluride family the total Te NMR shift,
including CS and KS, correlates well with the expected tellur-
ium oxidation state: the paramagnetic shift increases with
increasing overlap of the anionic sp band and the cationic d
levels, providing a tool for the estimation of the Te→M elec-
tron transfer. Thus the present study, although only semiquanti-
tative, supports the previous electronic band structure calcul-
ations which have shown that weak anionic bondings occur.
transfer in ZrTe is not known, we can assume that its oxidation
2
state lies between that of HfTe and TiTe . From a cubic spline
2
2
125
fit (see Fig. 7), a tellurium oxidation state equal to about Ϫ1.95
can be estimated for ZrTe . On the other hand, TiTe gives a
2
2
markedly shifted signal at δ ca. 3800. The large paramagnetic
shift between ZrTe and TiTe is consistent with the important
2
2
electron transfer occurring in titanium ditelluride (oxidation
state calculated at Ϫ1.8).
125
For VTe calculations have indicated an electron transfer
2
5,6
comparable to that in TiTe2. Since in these compounds the
local environments of tellurium are very similar one expects
1
25
Te resonances at similar shifts and this is indeed observed
δ 3400 and 3800 for VTe and TiTe , respectively). Among
(
2
2
Group 5 transition metals the electronegativity decreases from
V (1.63) to Nb (1.60) and Ta (1.50) implying a decrease of the
sp(Te) → d(M) electron transfer from VTe to NbTe and
2
2
J. Chem. Soc., Dalton Trans., 1997, Pages 3741–3748
3747