1506
J.C. Champarnaud-Mesjard et al. / Journal of Physics and Chemistry of Solids 61 (2000) 1499–1507
two lattices, it was found that: (i) The high-frequency vibra-
tions ꢀv Ͼ 500 cmϪ1 represent the intramolecular Te–Oeq
stretching motions; the number of those vibrations always
corresponds to the number of Te–Oeq bonds; (ii) Among the
Raman-active stretching vibrations, there is always one
with very high intensity which dominates all the spectra.
This is absolutely opposite to the view of the Raman
spectra of typical “homogeneous” framework-like XO2
structures built up from symmetric X–O–X bridges (see,
e.g. Raman spectra of silica [15]). In such structures, the
high-frequency region of the spectra consists of the stretch-
ing asymmetrical vibrations nasym of the X–O–X bridges.
Therefore, they would necessarily have low Raman-intensity,
which is dictated by asymmetrical bond length changes
(within the bridges) inherent for these vibrations; (iii) The
Te–Oax potentials (jointly with O–O ones) dominate the
frequency interval below 400 cmϪ1, thus describing the inter-
molecular forces.
and of nasym of the a-bridges, respectively, whereas the
vibration near 820 cm-1 (attributed to the B1 species) is
more delocalized.
The high intensity of the line near 683 cmϪ1 unambigu-
ously indicates that, from the point of view of crystal chem-
istry, its description as nasym of the a-bridges is not adequate
to “chemical reality”. Actually, this is due to the totally
symmetric pulsation of all the covalent Te–Oeq bonds
˚
(1.86 A), that the line in question has such a high intensity
(such a picture is characteristic for the strongest high-
frequency vibrations in the Raman spectra of a- and
b-TeO2). Formally, those are the asymmetric vibrations
nasym of the Te–eqOax–Te bridges, but in reality the Te–
Oeq bond contribution essentially predominates [5]. More-
over, the calculations show that no vibration can be speci-
fied as nsym of the a-bridge; so the term “a-bridge” has rather
geometrical than chemical sense. In fact, it must be con-
sidered as the strong (largely covalent) Te–Oeq bond
˚
The discussed properties manifest well a quasi-molecular
character of the a and b-TeO2 lattices [5], i.e. the essential
anisotropy of the electron distribution in their TeO4 group-
ings, or, in other words, the strong asymmetry of the chemi-
cal bonding in the Te–O–Te bridges.
(1.86 A) adjusted with the weak electrostatic Te–Oax
˚
contact (2.20 A) interconnecting the above mentioned
helical chains.
The two vibrations of the s-bridge demonstrate other
situation: the line near 426 cmϪ1 (nsym) is rather intense,
whereas that near 611 cmϪ1 (nasym) is much weaker. Thus
this resembles the properties of X–O–X bridge vibrations in
the Raman spectra of such typical (“homogeneous”) XO2
frameworks as various polymorphs of silica [18]. For exam-
ple, the A1-spectrum of a-quartz contains a very intense line
near 464 cmϪ1 and a weak line near 1080 cmϪ1. Those lines
are unambiguously attributable to nsym and to nasym of the
Si–O–Si bridge, respectively [19]. According to our present
treatment of the g-TeO2 spectrum, the two lines near 426
and 611 cmϪ1 are the homologues of the two above
mentioned lines of quartz, respectively.
Let us concentrate now on the data obtained for g-TeO2.
In the light of the issue made above, the central attention
must be paid to the point which differentiates this structure
from the a- and b-TeO2 ones. That is the existence of two
types of Te–O–Te bridges in g-TeO2, respectively, less
symmetric (we call it a-bridge) and much more symmetric
(s-bridge) than the bridges in a- and b-TeO2. In the a-bridge,
˚
the shortest Te–Oeq bond of 1.86 A (close to that in the TeO2
molecule), is alternated with a much longer Te···O axial
˚
contact (2.20 A). In contrast to this, the characteristics of
the s-bridge indicate much more isotropy in the Te–O bond-
ing. This bridge is shorter than any other Te–O–Te bridge
existing in the crystalline lattices of TeO2 and tellurates
Thus the spectrochemical properties indicate that a
Te–axOeq–Te polymerization occurs in the s-bridges,
and is practically absent in the a-bridges. At the same
time, in extrapolating the above mentioned quantum-
chemical estimates [16] of the orders of the different
tellurium–oxygen bonds, it can be thought that, in a-
˚
[17]. The bond lengths (1.95 and 2.02 A) are very close to
their average magnitude, i.e. the bridge is almost
“balanced”.
In contrast to the Raman spectrum of a-TeO2, which
manifests the quasi-molecular character of this lattice [5],
it is likely that the vibrational properties of g-TeO2 indicate
the occurrence of the intermolecular bonding (polymeriza-
tion) within the Te(1)–O(2)–Te(2)–O(2) chain (Fig. 7a).
Let us consider the Raman spectrum of g-TeO2 (Fig. 2)
and focus our attention on the six lines observable above
400 cmϪ1: 426, 611, 645, 683, 752 and 812 cmϪ1. The
assignment of these lines to the calculated vibrations is
given in Table 5. The analysis of their shapes (eigenvectors)
leads us to the following issue: the lines near 426 and
611 cmϪ1 correspond to the totally symmetric combinations
(A1 species) of the nsym and nasym stretching vibrations in the
s-bridges, respectively; the line near 683 cmϪ1 to the totally
symmetric combination (A1 species) of the nasym vibrations
of the a-bridge. The weak lines near 645 and 752 cmϪ1
correspond to A2-combinations of nasym of the s-bridges
˚
bridge, the order of the Te–O (1.86 A) bond is nearly 2,
˚
and that of the Te···O (2.20 A) contact is about ten
times lesser. At the same time, the orders of the
bonds in the s-bridge are much closer to their average
value of about 1. Consequently, the g-TeO2 lattice can
be considered as a chain structure with well-pronounced
covalent bonding along the z-axis, and with mainly
ionic inter-chain interactions. In this case, the lattice
can be compared with the chain structure of crystalline
SeO2 [20] (this latter compound is isovalent to TeO2, and,
in a gaseous phase, also exists as an isolated three atomic
molecule).
The calculated bulk modulus of g-TeO2 (Table 5), mainly
˚
determined by interactions of atoms separated by 2.5–3.3 A,
is found to be about two times lower as that of b-TeO2 and
three times lower than that of a-TeO2.