Dynamics and crystal chemistry of tellurites. II. Composition- and temperature-dependence of the Raman spectra of x(Tl2O) + (1 - x) Te2O glasses: Evidence for a phase separation?

Article abstract of DOI:10.1016/j.jpcs.2003.11.020

Immiscibility effects are suggested to be inherent in tellurite glasses with weak (low valence) cations. This hypothesis is used to explain the structural evolution of the x(Tl₂O)+(1-x)Te₂O glass system as a function of the x value and during temperature-induced crystallisation processes. The results of Raman measurements support the hypothesis.

Full text of DOI:10.1016/j.jpcs.2003.11.020

Journal of Physics and Chemistry of Solids 65 (2004) 981–993
www.elsevier.com/locate/jpcs
Dynamics and crystal chemistry of tellurites. II. Composition- and
temperature-dependence of the Raman spectra of x(Tl O) þ ð1 2 xÞ Te O
2
2
glasses: evidence for a phase separation?
O. Noguera, T. Merle-M e´ jean, A.P. Mirgorodsky , P. Thomas, J.-C. Champarnaud-Mesjard
*
Laboratoire de Science des Proc e´ d e´ s C e´ ramiques et de Traitements de Surface, UMR 6638 CNRS, Universit e´ de Limoges, Facult e´ des Sciences,
1
23, avenue Albert Thomas, Limoges, Cedex 87060, France
Received 10 June 2003; revised 10 June 2003; accepted 3 November 2003
Abstract
Immiscibility effects are suggested to be inherent in tellurite glasses with weak (low valence) cations. This hypothesis is used to explain the
structural evolution of the xðTl OÞ þ ð1 2 xÞTe O glass system as a function of the x value and during temperature-induced crystallisation
2
2
processes. The results of Raman measurements support the hypothesis.
q 2003 Elsevier Ltd. All rights reserved.
1
. Introduction
that its mechanism differs in nature from those proposed for
classic complex oxide glasses A B O :
k
i
j
In general, such glasses can be characterised as frozen
non-equilibrium systems in which (during their relaxation to
crystalline ground states) the initially disordered complex
anions B O and cations A evolve to regular lattices of the
Tellurite glasses are of special technological and
scientific interest because of their good visible and infrared
light transmittance and their high refractive indexes [1].
Moreover, they possess the highest non-linear susceptibility
among the known oxide glasses. The microscopic nature of
those properties is still far from final clarity, which
stimulates a special interest in the fundamental research of
i
j
relevant crystalline structures. This is the case for silicate
glasses in which the atomic arrangement in the complex
anions resembles that in the crystals, which shows the
primary role of the short-range factor (chemical bonding) in
the organisation of those solids.
the TeO -based compounds. This article is concerned with
2
the basic principles of their constitution. We present some
reasoning regarding the mechanism of the formation of
tellurite glasses, which brings up the question of immisci-
bility in those materials.
The experimental evidence arguing for that can be found
in the vibrational spectra of the silicate glasses which as a
rule, if not always, envelop those of the ‘parent’ crystals,
and manifest the same composition-dependence as the
spectra of the crystals. For example, the evolution of a
silicate glass system from the sodium meta-silicate to
sodium di-silicate formulation is accompanied by an
essential change of its vibrational spectrum, which reflects
a sharp distinction in the constitution of the relevant
complex anions consisting of the SiO chains, or Si O
Immiscibility behaviour is frequently observed for
classic’ glassy oxides like silicate glasses. It manifests
‘
the thermodynamic state of the glass, and influences a
variety of glass properties. Although several cases of such
behaviour have been observed for tellurite glasses [2], this
point is rarely discussed in the relevant literature (see e.g.
Ref. [1]). The present article is designed to bring to the
attention of the reader the ideas and facts which suggest that
3
2
5
layers, respectively [3].
However, it appears that such a systematic ‘glass–
crystal’ structural correspondence is not true for tellurites.
Actually, the complex anions in the tellurite crystals with
general formula nA O þ mTeO manifest a great diver-
immiscibility can be inherent in TeO -based glasses, and
2
*
Corresponding author. Tel.: þ33-5-55-45-74-33; fax: þ33-5-55-45-
2-70.
E-mail address: andrem@unilim.fr (A.P. Mirgorodsky).
7
p
r
2
sity of structures strongly depending on the n=m ratio.
0022-3697/$ - see front matter q 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jpcs.2003.11.020

982
O. Noguera et al. / Journal of Physics and Chemistry of Solids 65 (2004) 981–993
Consequently, their Raman spectra are highly variable and
sensitive to that ratio (see e.g. Figs. 9–11 in Ref. [4]). In
contrast to this, the spectra of the relevant xðA O Þ þ ð1 2
discussed in Section 5 in the light of the above-mentioned
hypothesis.
The preliminary studies of the xðTl OÞ þ ð1 2 xÞ Te O
n
r
2
2
xÞTeO glasses, as a rule, have only a weak dependence on
system (see Ref. [5] and references therein) have revealed a
large domain of a glassy state and four stoichiometric
crystalline tellurite structures: three stable ones, Tl2TeO3;
2
the x=1 2 x ratio: the band positions very frequently are kept
during the x value variation. Moreover, the shapes of the
spectra differ significantly from those of the related crystals
a-Tl
determined, and one metastable, b-Tl Te O , with unknown
Te O , Tl Te O for which the atomic positions were
2 5 2 3 7
2
(
see, e.g. Ref. [4]). So it appears that the atomic arrangement
2
2 5
structure. The Raman spectra of the crystals were well
documented and interpreted in detail within the lattice
dynamical model calculations [5] (with the exception of b-
in the complex anions of the tellurite glasses is not so
composition-dependent, and can essentially differ from that
in the parent tellurite crystals. In the present article we try to
understand the nature of such an ‘anomaly’ by using the
ideas tentatively expressed in Ref. [5] and developed in
more detail in Section 1 as a working hypothesis for
immiscibility (phase separation) of tellurite glasses.
Tl Te O ). In Fig. 1 those spectra are given jointly with the
2 2 5
spectra of the a- and g-crystalline polymorphs of TeO as
2
the basic data on the crystalline structures involved in the
present work. First, we consider the variation of the Raman
spectra of the above-mentioned system at ambient con-
ditions for x changing between 0 and 0.5. Second, for
selected values x, we consider the evolution of the spectra
related to crystallisation processes during heating.
In Section 6, the results thus obtained are summarised
and conclusions are offered in favour of the suggested
hypothesis.
The latter effect has a very important bearing on many
properties of glasses, and stimulates a large body of research
[
2]. Although its microscopic mechanism is not as yet clear,
the thermodynamic aspects of its origin seem to be
understood. There is a prerequisite for a separation of a
given glass into two phases X and Y (only this simple case is
considered in this article) if a close approach of the
structural units forming phase X is more favorable for
reducing potential energy E than the alternating (blending)
of those units with the units forming phase Y. However, this
would necessarily decrease entropy S; thus diminishing its
role in the free energy F ¼ E 2 TS: Consequently, with
increasing temperature, when the entropy factor becomes
sufficiently important, the phase separation can cease being
thermodynamically beneficial, and the system would
become homogeneous.
2
. Background
Tellurites belong to the category of ionic–covalent
complex oxides. The formation of those structures can be
formally described by the reaction
nApOr þ mBiOj ¼ AnpBmiOnrþmj
ð1Þ
in which atom B is implied to be more electronegative than
atom A. In such a case, the oxygen atoms brought by
modifier A O can form with atoms B chemical bonds
We would like to underline that immiscibility of the
tellurite glasses can be of significant scientific interest due to
its specific mechanism. Actually, this effect should be
essentially related to the particular constitution of the liquid
state of tellurites in which the charged entities (simple
p
r
stronger than those with atoms A. If so, the final structure
will consist of the B O complex anions and cations A.
mi nrþmj
Let us focus first on some structural and crystal chemistry
peculiarities of classic ionic–covalent complex oxides like
silicates, and rewrite Eq. (1) as
cations A and complex ortho-anions TeO ) would generally
3
occur along with the neutral entities, polar molecules TeO2:
The charge–charge (monopole–monopole) interactions
between the cations and anions are usually stronger than
their interactions (monopole–dipole) with the neutral
nA O þ mSiO ¼ A Si O
;
ð2Þ
p
r
2
np
m
nrþ2m
assuming that the silica is taken in its ground state—a quartz
lattice. Note that the oxygen spheres around the atoms of Si
are ‘saturated’ up to the maximal co-ordination number (four)
on both sides of Eq. (2). So the form of coordination
molecules TeO : This would favor the formation of the
2
aggregates of the charged units separately from those of
neutral ones. Consequently, the energy minimisation, i.e.
the thermodynamic factor governing the structural evolution
of the system, can be associated with the creation of the two
separated phases: that made of the charged entities, and that
made of the neutral ones.
polyhedrons around silicon atoms (SiO tetrahedrons) is kept
4
during that reaction and the total number of the Si–O bonds
does not vary. The same can be said about the co-ordination
polyhedrons of cation atoms A: very frequently those polyhe-
dra remain unchanged during the formation of the silicate
lattice (Eq. (2). Thus, it must be thought that this process is
mainly stimulated by the variation of the co-ordination
spheresaroundtheoxygenatoms, whichmakestheirpotential
wells deeper. This point is worth further consideration.
In the initial stage (left-hand side of Eq. (2)), all the
In this study, the experimental evidence of the phase
separation in tellurite glasses was mainly obtained by
Raman spectroscopy which we have used as a tool capable
of providing structural information about the composition-
and temperature-induced variations in the glassy and
crystalline compounds under consideration. The relevant
experimental data are presented in Section 4 and are
oxygen atoms in the SiO structure are the bridging ones,
2

O. Noguera et al. / Journal of Physics and Chemistry of Solids 65 (2004) 981–993
983
Fig. 1. Raman spectra of thallium–tellurite crystals, Tl
compound is labelled by a specific sign.
2
TeO , a-Tl
3
Te
2 2
O
5
, b-Tl
2
2
Te O
5
, Tl
2
3
Te O
7
, and of pure tellurium oxides, a-TeO
2
, and g-TeO
2
. Each
and the number of atoms O per single atom of Si (i.e. O/Si
ratio) is equal to two. During the formation of a silicate
structure as presented in Eq. (2), the nr oxygen atoms
bonds are ‘stronger’ than the Si–O bonds, reaction (2) cannot
be so energetically beneficial. Therefore, the notion of the
‘cation force’ is used in the crystal chemistry of silicates [3].
According to this notion, the weaker cation A is, the more
probable reaction (2) is.
‘delivered’ by the A Or modifier would enter the co-
p
ordination spheres of silicon atoms thus lowering the
potential energy of the structure. Consequently, 2nr=m
atoms O in each SiO tetrahedron (in the right-hand side of
Eq. (2)) become terminal ones thus increasing the O/Si ratio
Analogously to Eq. (2), the formation of a tellurite
structure can be described as
4
nA O þ mTeO ¼ A Te O
ð3Þ
assuming that the tellurium dioxide is taken in its ground
p
r
2
np
m
nrþ2m
(i.e. the number of atoms O bonded with atoms Si) without
changing the total number of Si–O bonds. Eventually, when
all the oxygen atoms in the SiO tetrahedra become terminal,
4
state structure, a-TeO paratellurite lattice, and that the Ap
2
i.e. the initial SiO framework (in Eq. (2)) is transformed into
2
O modifier involves ‘weak’ cations A (which will be
r
the ortho-structure (i.e. island-type) in which all the SiO4
tetrahedrons are isolated, the O/Si ratio would achieve its
maximum magnitude of 4. During such a transformation, the
always implied below).
It should be particularly emphasised that, in contrast to
Eq. (2), the form of the TeO polyhedrons and the number of
n
formal charge of the SiO tetrahedra varies from 0 to 24 e.
4
Te–O bonds on the right-hand side of Eq. (3) differ from
those on the left-hand side. Actually, tellurites are the salts
of tellurious acid H TeO in which the complex tellurite
(As a rule, all the tetrahedra in a given silicate structure are
identical, i.e. the total charge of the Si O anion is
m
nrþ2m
uniformly distributed between them). However, if the A–O
2
3
anion is of a pyramid-like TeO group formed by three
3

984
O. Noguera et al. / Journal of Physics and Chemistry of Solids 65 (2004) 981–993
largely covalent Te-O bonds with length estimated at about
˚
.85 A. So, by definition, a quantity of those groups must
than atoms of oxygen, will necessarily possess two peculiar
features in the Raman spectrum [6]: (i) its high-frequency
part, occupied by the asymmetric stretching vibrations of the
X–O–X bridges, would have vanishing intensity; (ii)
the spectrum would be dominated by a strong band situated
the middle-frequency range and originating from the
symmetric stretching vibrations of those bridges. The classic
1
necessarily appear in the tellurite anion (right-hand side of
Eq. (3)), whereas they are absent in the a-TeO lattice (i.e.
2
on the left-hand side of Eq. (3)). Since the formation of the
TeO3 group corresponds to the saturation of the co-
ordination spheres around atoms of Te, the maximal value
of the ðnr þ 2mÞ=m ratio in Eq. (3) is equal to three. Such,
indeed, is the case if the modifier with mono-valence (i.e.
weak) cations A is taken in an equimolar proportion with
examples of such XO structures are the silica polymorphs.
2
They always have the strongest Raman scattering in the range
2
450–480 cm , whereas only very weak bands can be
1
2
observed higher than 800 cm [8].
1
TeO , so that Eq. (3) can be re-written as
2
Consequently, if the Raman spectrum of any oxide
structure containing the X–O covalent bonds has strong
bands in the highest-frequency range, whereas the middle-
frequency range is empty or contains only weak bands, it
can be thought that the X–O–X bridges are absent in the
structure, and those strong bands relate to the vibrations of
terminal X–O bonds. (The spectra of ortho-silicate lattices,
A O þ TeO ¼ A TeO ;
ð4Þ
This assures the formation of three Te–O bonds per atom
2
2
2
3
of Te. Consequently, the anion subsystem of the tellurite
lattice on the right-hand side of Eq. (4) would contain only
2
2
the ðTeO Þ ortho-groups.
3
The two following questions are important for our further
consideration: (i) what is the number of Te–O bonds per
one atom of Te on the left-hand side of Eq. (4)? (ii) what
fragment of the paratellurite lattice should be considered as
e.g. of ZrSiO
The spectrum of a-TeO
or of Mg SiO
, beautifully illustrate this).
(Fig. 1b) is dominated by the
4
2 4
2
band in the highest-frequency range, and only very weak
band is seen in the middle-frequency range. So, keeping in
mind the above-mentioned sentences, it can be concluded
that the Te–O bonds in the paratellurite lattice are mainly
terminal ones, and the Te –O –Te bridge polymerisation
its basic structural unit? Traditionally, the a-TeO structure
2
is regarded as a covalently bonded three-dimensional
framework (like the silica polymorphs) made of the TeO4
units called disphenoids. This implies that there are four
Te–O chemical bonds per one atom of Te (two equatorial
eq
ax
is practically absent. In line with this, the ab initio electron
structure calculations [9] of a cluster cut from the a-TeO
˚
bonds are short—1.87 A, and two axial ones are long—
2
˚
.12 A), and all those bonds form the Te–O–Te bridges.
2
lattice reveal that the order of the Te–O axial bonds is much
lower than that of the equatorial bonds (0.3 and 1.7,
respectively).
According to the basic principles of chemistry, the existence
of such bonds is an essential factor for the lattice energy
minimisation, and their number should not be lowered
during the reaction represented by Eq. (4). However, the
Note that the molecular description of a-TeO would
2
imply that the number of Te–O bonds per atom of Te
necessarily increases from two to three during the formation
of the tellurite lattice represented by Eq. (4). Taking into
account the above facts and speculations, we venture the
opinion that the latter description is more chemically
consistent.
framework description of the a-TeO lattice means that the
2
formation of the tellurite lattice, as presented by Eq. (4),
would necessarily lower the number of the Te–O bonds per
one atom of Te from four to three, and thus the adequacy of
such a description to ‘chemical reality’ seems, at the very
least, debatable [6].
Irrespective of the extent to which the a-TeO lattice can
2
There is another view which holds that the a-TeO lattice
2
be considered as a molecular crystal (its dualism is
mentioned in Ref. [6]), it is natural to think that these are
the A–O bonds of the modifier and the weak axial Te–O
can be regarded as a largely molecular structure built up
from the TeO molecules [6,7] which correspond to pairs of
2
the short equatorial bonds in the TeO4 disphenoids.
Consequently, the two long axial bonds of the disphenoids
are considered as (mainly electrostatic) contacts between
oppositely charged extremities of the neighbouring mol-
ecules. Thus, in contrast to the framework description, the
molecular description suggests that there are only two
covalent bonds per one atom of Te, and those bonds are
essentially terminal (non-bridging).
‘bonds’ of a-TeO which would be broken primarily when
2
the compounds indicated on the left-hand side of Eq. (3)
22
begin to melt. Theoretically, nr negative ions O and np
þ2r=p
positive ions A
dipole TeO
the melt during this process. To saturate the co-ordination
carried by the modifier, and m neutral
molecules carried by a-TeO would appear in
2
2
2
2
spheres around atoms of Te, the O anions would react
2
2
with the molecules, thus forming nr TeO pyramid-like
3
To judge which of those two descriptions is more
chemically consistent (and thus to answer the question: are
the Te–O bonds in the paratellurite lattice bridging or
terminal bonds?), we turn our attention to the Raman
dipole ortho-anions.
22
22
TeO þ O ¼ TeO3
ð5Þ
2
If the number of O 2 ions is sufficient to transform all the
2
spectrum of a-TeO (Fig. 1b). It is useful to recall, that any
2
homogeneously polymerized XO framework built up from
2
molecules, i.e. nr ¼ m; the melt would eventually consist of
2
r=pþ
and m ortho-anions TeO : Depending on
3
22
non-linear X–O–X bridges with atoms of X much harder
np cations A

O. Noguera et al. / Journal of Physics and Chemistry of Solids 65 (2004) 981–993
985
the properties of the cations, the latter entities can be
spatially separated or could associate (due to the dipole–
dipole attraction), thus forming complex anions framed
the pure TeO2 phase, and the monopole–monopole
þ2r=p
(charge–charge) forces which would unite the A
2
2
cations and the TeO anions into the A ðTeO Þ ortho-
3
p
3 r
from m TeO pyramids. In the event that there are not
3
tellurite phase. We note that neutral molecular entities are
absent during the formation of the classic complex oxide
glasses: silicates, phosphates.
2
enough O ions, i.e. nr , m; then nr ortho-anions TeO3
2
22
and m 2 nr neutral molecules TeO would be present in the
2
melt. During the crystallization process, the dipole–dipole
attractive forces would tend to associate those structural
In Section 5, we shall apply the above-mentioned
scheme to understand the mechanisms which govern the
composition- and temperature- induced structural evol-
utions of the tellurite glasses, observed in our experimental
studies.
units into the ðTe O
Þ complex anions (see Eq. (3))
m
nrþ2m
22
2nr2
ðm 2 nrÞTeO þ nrTeO3 ¼ ðTe Onrþ2mÞ
ð6Þ
2
m
22
In fact, negative extremities ðO Þ of the covalent Te–O
bonds belonging to a given unit would approach positive
3. Experimental
4þ
apexes ðTe Þ of neighbouring units, thus forming the Te–
O–Te linkage. Consequently, it can be thought that the
TeO molecules and the TeO3 ortho-ions are polymerised
there through the Te–O–Te bridges. However, strong bands
Glassy samples were prepared by first melting at 800 8C
during 1 h appropriate quantities of reagent grade TeO and
2
2
2
2
Tl TeO in a sealed gold tube, and then air-quenching the
2
2
1
3
in the range 300–500 cm , as a rule, are absent in the
Raman scattering of tellurite lattices with weak cations (see
e.g. Fig. 1(c)–(f)), which means that such a polymerisation
is practically absent, thus arguing for the ‘TeO2
melts. TeO was obtained by the decomposition at 550 8C of
2
commercial H TeO (Aldrich, 99.9%), and Tl TeO was
6
6
2
3
synthesised by heating at 350 8C for 18 h under nitrogen
atmosphere an intimate mixture of TeO2 and Tl CO
2
3
molecules þ TeO groups’ description of the tellurite
3
(Aldrich 99.9%). Glass formation domain was determined
by using X-ray diffraction (Guinier-De Wolff camera,
anions, corresponding to the left-hand side of Eq. (6).
To understand the principles of the short-range atomic
arrangement of the glassy tellurites, the questions of prime
Cu Ka radiation). Under these conditions, the glass-forming
domain ranged from 8 to 35 mol% Tl O. In order to extend
2
2
2
interest for us are: (i) Do the TeO and ðTeO Þ entities in
2
3
it up to pure TeO and Tl TeO compositions, the melts
2
2
3
tellurite glasses pooled in the structural fragments resemble
those in the relevant crystals? (ii) If not, how are those
entities distributed ?
were quickly dipped in a freezing mixture consisting of ice,
ethanol and NaCl kept at about 214 8C (cooling rate was of
4
about 10 8C/s).
The temperature-induced evolution of the Raman spectra
of the samples was studied between 20 and 600 8C with a
heating rate of 5 8C/min. The sizes of typical crystallites
observed during the crystallization varied from 2 to 10 mm.
The cooling conditions (rate of 10 8C/min, and a delay of
As was mentioned above, in general case (i.e. when
2
2
m . nr in Eq. (3)) the three types of structural units, TeO
þ2r=p
;
3
TeO2 and A ; can be found in the melted tellurite
represented by Eq. (3). To minimise the electrostatic energy
2
3
2
of the melt, the negatively charged ortho-groups TeO
þ2r=p
would tend to approach the positive ions A (rather than
the neutral molecules TeO ), thus expelling the TeO
30 min before the spectrum recording) allowed us to avoid
temperature gradients across the samples.
2
2
molecules which hinder this process. Theoretically, this
would result in an appearance of domains having chemical
composition A ðTeO Þ : Simultaneously, the domains
To ensure the temperature homogeneity of the samples,
a very small quantity of powder was placed in a thermal
cell (Linkam, model THS600) just above the thermocouple
position. To control the reproducibility of the results, the
spectra were repeatedly recorded for different samples after
an annealing time varying from 2 up to 30 min at each
temperature. (Only the selected measurements will be
presented in this article (Figs. 3–7).
p
3 r
enriched with pure dioxide TeO would appear. Thus, it is
2
highly likely that the tellurite melt would be essentially a
non-homogeneous system. This allows us to hypothesise
that (except in the case m ¼ nr) immiscibility is inherent for
tellurite glasses. Consequently, a phase separation into the
A (TeO ) and the TeO domains can be generally expected
p
3 r
2
The thermal cell was mounted in the path of a laser beam,
and the Raman spectra were recorded in back-scattering
mode using the XY model Jobin–Yvon spectrometer
in the A Te O
glasses so that their composition
formally given by Eq. (3)) can be more adequately
np
m
nrþ2m
(
described as a bi-phase one
þ
equipped with Ar laser ðl ¼ 514; 532 nm; power 50–
o
2
1
00 mW, spectral resolution 2.78 cm ) and a CCD
1
A Te O
np m
¼ nA ðTeO Þ þ ðm 2 nrÞTeO
ð7Þ
nrþ2m
p
3 r
2
detector. No specific polarisation effects have been studied
(neither polariser, neither analyser were included in the
optic scheme). The incident radiation passed through a
microscope (objective 100 £ or 50LW for thermal
According to the above consideration, immiscibility of a
tellurite glass is due to the two types of attractive forces
acting between the structural units generally existing in its
melt. Those forces are the intermolecular dipole–dipole
2
experiment). The laser spot size was about 1 mm . Due to
forces which would unite the neutral TeO molecules into
2
the confocal system of the spectrometer, the analysed

986
O. Noguera et al. / Journal of Physics and Chemistry of Solids 65 (2004) 981–993
3
volume of the sample was of about 1 mm , thus only one
crystallite could be interrogated at a time.
and the two above-mentioned bands, near 450 and
2
650 cm , relate to those fragments, respectively. No
1
2
1
intense band in the range 400–470 cm is observed in the
spectra of the thallium tellurite crystalline lattices (Fig. 1).
However, such a band always occurs in the spectra of the
xTl Oþð12xÞTeO glasses for 0:5 . x . 0; and its
4
. Results
2
2
4
.1. Composition dependence of the Raman spectra
intensity monotonically decreases with increasing x:
Another limiting case, x ¼ 0:5 corresponds to ortho-
tellurite Tl TeO . Its crystalline phase consists of isolated
Fig. 2 exhibits the Raman spectra of the xTl Oþð12xÞ
2
2
3
TeO glasses with x varying between 0 and 0.5. The
2
TeO pyramids separated by Tl cations, and has a Raman
3
spectrum of pure TeO ðx ¼ 0Þ is well known and discussed
2
spectrum (Fig. 1c) with the two characteristic bands: that
2
in the literature [10]. It has the two main bands situated
1
2
1
near 725 cm corresponds to the synchronous stretching
pulsation of Te–O bonds in the TeO pyramids, and the
near 450 and 650 cm . We wish to emphasize the
remarkably strong intensity of the former band, which
appears to be close to the intensity of the latter (dominant)
band. In the light of the general comments made in
Section 2 with respect to the Raman spectra of XO2
oxides, it can be hypothesized that polymerized Te–O–Te
3
2
1
band near 290 cm corresponds to a synchronous bending
of the O–Te–O angles [5]. The spectrum of the Tl TeO
2
3
glass (Fig. 3A) has broader features that envelop its
crystalline homologue so that the positions of those bands
are practically the same, which shows that the structure of
bridges and Te–O terminal bonds coexist in TeO glass,
2
Fig. 2. Composition dependence of the Raman spectra of the xTl
2
2
O þ ð1 2 xÞTeO glassy samples. All the spectra are presented with the same intensity in an
arbitrary scale.

O. Noguera et al. / Journal of Physics and Chemistry of Solids 65 (2004) 981–993
987
Fig. 3. (A) (a–c) Temperature-induced evolution of the Raman spectra of 0.5 Tl
spectra of 0.33 Tl glass (the labels are relative to the crystallised phases shown in Fig. 1); (C) (a–d) Temperature-induced evolution of the
O þ 0.66 TeO
Raman spectra of 0.25 Tl glass (the labels are relative to the crystallised phases shown in Fig. 1); (D) (a–f) Temperature-induced evolution of
O þ 0.75 TeO
the Raman spectra of pure TeO glass (the labels are relative to the crystallised phases shown in Fig. 1). Figures Ae, Bf, Ce, Dg represent the spectra recorded
2
2
O þ 0.5 TeO glass; (B) (a–e) Temperature-induced evolution of the Raman
2
2
2
2
2
after cooling to room temperature.
this glass is close to the structure of the Tl TeO crystalline
2
Actually, in the spectrum of 0.33Tl O þ 0.66TeO glass
3
2
2
2
1
lattice.
(Fig. 2i), the maximum intensity is observed at 725 cm
although in the relevant crystallised phase a-Tl Te , the
dominant band is located near 650 cm , and no band is
The main feature of the glass spectrum evolution for x
varying between 0 and 0.5 (Fig. 2) is well seen: with
increasing quantity of the Tl O modifier added to TeO , the
2
O
2 5
2
1
2
1
seen above 710 cm (Fig. 1e). For the case 0.25Tl
0.75TeO (Tl Te ), the situation is not so unambiguous:
the glass spectrum between 600 and 800 cm seems to be
compatible with the spectrum of the crystalline Tl Te
(although the maxima of their intensity do not coincide).
O–
2
2
2
initial spectrum of TeO continuously evolves into that of
2
2
2
O
3 7
2
1
Tl TeO : the two bands of the glassy TeO (near 450 and
2
3
2
2
1
6
50 cm ) disappear, making room for the lines of the
2
O
3 7
Tl TeO spectrum. During this evolution, one can see the
2
3
2
1
variation of the shape of the high-frequency band between
21
However, the strong band near 450 cm in the former
spectrum has no homologue in the latter; moreover, the
2
00 and 750 cm : the dominant peak shifts from 650 cm
1
6
2
pure TeO ) to 725 cm (pure Tl TeO ). A special case is
1
21
(
spectra between 600 and 800 cm for glass with x ¼ 0:33
2
2
3
the spectrum corresponding to x ¼ 0:18 for which the
dominant peak occupies the intermediate position (Fig. 2e).
This fact will be discussed below.
and for x ¼ 0:25 practically coincide, whereas their crystal-
line homologues differ dramatically.
The most pronounced difference between the spectra of
the glassy and crystallised states of the same composition is
observed for x ¼ 0:14; and this point will be discussed
below.
In principle, these observations are not new: similar
ones have already been reported [4]. Moreover, the same
results have been obtained for the tellurite glasses with
other modifiers having mono- and di-valent cations (see
Refs. [4,11], respectively). So we note that the spectra in
Fig. 2 represent a typical pattern of the tellurite glass
4.2. Thermal behaviour
Raman spectra evolution with varying TeO /modifier ratio.
2
The temperature-induced variations of the Raman
spectra of seven different compositions of glasses have
been studied: with x ¼ 0:5 (Fig. 3A), x ¼ 0:33 (Fig. 3B),
x ¼ 0:25 (Fig. 3C) and x ¼ 0 (Fig. 3D) which could exist in
forms of the mono-phase crystalline states, Tl TeO ,
The most intriguing fact observable in Fig. 2 is that the
intermediate compositions (with 0 , x , 0:5) manifest
spectra which differ critically from those of the related
crystal structures.
2
3

988
O. Noguera et al. / Journal of Physics and Chemistry of Solids 65 (2004) 981–993
Fig. 4. Temperature-induced evolution of the Raman spectra of 0.38 Tl
2
2
O þ 0.62 TeO glass (the labels are relative to the crystallized phases shown in Fig. 1).
Tl Te O , Tl Te O , TeO , respectively, and three compo-
2
The two Tl Te O crystalline forms, a- and b-forms, can be
2 2 5
2
2
5
2
3
7
sitions with x ¼ 0:38 (Fig. 4), x ¼ 0:29 (Fig. 5) and x ¼
expected to evolve by heating glass of this composition [5].
2
1
0
:14 (Fig. 6) for which the crystalline states were not found.
For the case x ¼ 0:5 (Fig. 3A), the Tl TeO lattice
First, a sharp peak arises at about 750 cm (Fig. 3B-b)
signalling the appearance of a small amount of the b-
Tl Te O structure, whereas most of the sample remains
2
3
crystallises from the glass phase at a temperature of about
1
2
2 5
50 8C.
For the case x ¼ 0:33 (Tl Te O ), a vigorous evolution
glassy. Several minutes later, the lines near 272 and
1
2
653 cm begin to develop indicating formation of the a-
Tl Te O lattice which is completed at about 160 8C
2
2 5
of the glass spectrum (Fig. 3B) is observed at about 140 8C.
2
2 5
Fig. 5. Temperature-induced evolution of the Raman spectra of 0.29 Tl
2
2
O þ 0.71TeO glass (the labels are relative to the crystallized phases shown in Fig. 1).

O. Noguera et al. / Journal of Physics and Chemistry of Solids 65 (2004) 981–993
989
Fig. 6. Temperature-induced evolution of the Raman spectra of 0.14 Tl
2
2
O þ 0.86 TeO glass (the labels are relative to the crystallized phases shown in Fig. 1).
(
Fig. 3B). The last spectrum (Fig. 3B-f), recorded at room
of the initial tellurite glass and that of the final structures
crystallising within its volume.
temperature, has no significant difference comparing to that
in Fig. 3B-e.
While heating the composition with x ¼ 0:25 (Fig. 3C)
the lines related to a- and b-Tl Te O structures first appear
2
5. Discussion and commentary
2 5
at about 140 8C against the background of the glassy bulk
spectrum. This picture is kept up to 260 8C, and then, all the
volume becomes a well-crystallised Tl Te O structure.
5.1. Composition dependence of the spectra
2
3 7
As seen in Fig. 6, the crystallisation of a pure TeO glass
2
Some general comments can be made relating the high-
2
1
begins at a temperature of about 240 8C with the formation
of the g-TeO structure which then transformed progress-
frequency bands (between 600 and 800 cm ) seen in all the
spectra in Figs. 1 and 2. The most intense bands correspond
to the vibrations of the Te–O terminal bonds [4–6]. Since
such bonds are present in any complex Te O anion, the
2
ively into an a-TeO lattice which becomes the single phase
2
at 300 8C.
m
t
The evolution of the spectra for compositions with x ¼
strongest lines in the Raman spectra of tellurites are always
situated in that range. However, their positions are highly
variable and therefore hardly can be characteristic for any
particular type of Te O anions. For example, even in the
0
:38; 0.29, and 0.14 (Figs. 4–6, respectively), will be
discussed in more detail below; we note that the latter
composition represents a beautiful example showing the
dramatic difference between the general view of the spectrum
m
t
2
2
Raman spectra of the isolated [TeO3] ortho-anions, which

990
O. Noguera et al. / Journal of Physics and Chemistry of Solids 65 (2004) 981–993
theoretically must contain a sole intense stretching vibration
a synchronous pulsation motion of the TeO pyramid),
the positions of the relevant bands depend strongly on the
glass can be regarded as that of theTl TeO glassto which the
2 3
(
two bands are added specifying the spectrum of a glassy
3
2
TeO : one near 450 cm , and one near 650 cm . To
1
21
2
2
1
cations and vary over a large range of 600–800 cm (Fig. 9
in Ref. [3]).
In such a situation, special attention must be paid to the
interpret this fact, it is sufficient to use the idea stated in
Section 2, according to which the Tl Te O glass would
2
2 5
consist of the two phases: the one is rich in Tl TeO and the
2 3
2
1
2
bands can indicate the presence of the following structures
50–500 cm range in which the three well-pronounced
other inTeO2.
It appears, that using such an idea, we reveal the
universal factor which governs the evolution of the spectra
in Fig. 2 with the x value decreasing from 0.5 to 0. Actually,
the relevant increasing intensities of the bands specifying
or structural units found within the xTl O þ (1 2 x)TeO
2
2
system, namely:
21
1) a broad intense band located near 400–500 cm is
(
(
glassy TeO
Tl TeO can be naturally associated with a continuous
substitution of the phase rich in the Tl TeO glass by that
rich in the TeO glass.
However, whereasa competitionbetweentheTl
TeO phases is clearly observed in the spectra in the bottom
2
, and decreasing intensities of those specifying
indicative of pure TeO glass (see Fig. 3D);
2
2
3
2
1
2) a band near 290–300 cm (which is clearly observed
in the spectra of both crystalline and glassy forms of
Tl TeO in Fig. 3A) is indicative of the isolated TeO
2
3
2
TeO and
2 3
2
3
3
groups;
3) a line near 272 cm , well separated in the spectrum in
2
2
1
(
and in the top of Fig. 2, the situation is not so unambiguous for
the spectra presented in the intermediate part of Fig. 2, and
this point is worth further consideration. The most intriguing
is the view of the Raman spectrum for x ¼ 0:18 (Fig. 2e).
Actually, a dominant peak in that spectrum has a position to
2
Fig. 3B, is indicative of the (Te O ) pyro-groups.
2
2
5
Actually, this line corresponds to symmetric vibrations
of the Te– O –Te bridges. Among the structures
ax ax
under consideration, such bridges exist only in the
2
2
(
crystal [5]. So, in principle, the picture in the range
Te O )
pyro-groups found in the a-Tl Te O
5
which no vibrations from the spectra of the Tl TeO or TeO
2 3 2
2
5
2
2
glassy phases can be directly attributed. The question is: does
that peak correspond to some peculiar structure, or does it just
result from a superposition of the spectra of those two phases?
First, it was found that the peculiarities of the spectrum for
x ¼ 0:18 can be reproduced (Fig. 7a) by summing the two
experimental spectra, those for x ¼ 0:21 and for x ¼ 0:14;
which are the nearest neighbours of that spectrum in Fig. 2.
Then, a Gaussian effective oscillator model capable of
reproducing the high-frequency parts of the spectra of the
two limiting compositions, with x ¼ 0:5 (Tl TeO ) and with
2
50–500 cm can quite unambiguously indicate the
1
2
structural peculiarities of the system during the x value
variation. We note first some points related to the
spectra of TeO , Tl TeO , and Tl Te O in glassy and
2
2
3
2
2 5
crystalline phases.
As it was mentioned in Eq. (1) (Section 2), the two
principal bands in the spectrum of TeO glass (Fig. 3D-a)
2
point to the fact that this compound contains quite
polymerized Te–O–Te bridges jointly with terminal Te–O
bonds. Such a structural peculiarity is absent in the
paratellurite lattice, but it is present in the lattice of g-TeO2.
The latter fact is documented by the X-ray diffraction study
2
3
x ¼ 0 (TeO ), was developed. For both those spectra, it was
2
sufficient to use the models including only the two
2
1
oscillators: near 656 and 722 cm for x ¼ 0:5; and near
2
1
654 and 758 cm for x ¼ 0:
[
by the coexistence of the two intense bands, near 426 and
12], and well visualized in the Raman spectrum in Fig. 1(a)
The parameters of those four oscillators were used as an
initial set to describe the spectrum for x ¼ 0:18 (Fig. 7b). The
fitting procedure accurately reproduced the shape of the latter
spectrum with the following final oscillator frequencies: 661,
21
6
80 cm (Fig. 3D-c) related to the Te–O–Te bridges and to
the Te–O terminal bonds, respectively, found in that lattice
12]. Thus, it can be thought that the structure of TeO glass is
2
1
[
722, 665, and 763 cm , which differ from the oscillator
positions in the spectra of the Tl TeO and TeO pure phases
2
more closely related to the lattice of g-TeO than to that of
2
2
3
2
paratellurite[13]. (Wenote thattheg-TeO latticeisnotstable
2
by less than 1%. A further analysis has shown that the domain
of the high-frequency vibrations in all the spectra in Fig. 2
can be reproduced in keeping the above-mentioned oscillator
frequencies, but changing the oscillator intensities in line
with the variation of the proportion between the Tl TeO and
and usually evolves into the paratellurite lattice at heating).
In contrast to this, the situation with the Tl TeO3
2
composition seems to be quite unequivocal: the spectrum
of the glass with x ¼ 0:5 envelopes the spectrum of the
Tl TeO crystal (Fig. 3A) thus indicating the occurrence of
2
3
TeO phases. So it appears that all the glasses in our system,
2
2
3
the same structural units in both phases.
However, this is not the case for the Tl Te O compo-
according to their Raman spectra can be regarded as
consisting of those two phases in accordance to Eq. (7).
2
2 5
sition (Fig. 3B). In that case, the spectrum of the glass differs
1
2
significantly from that of the crystal. No band near 270 cm
related to Te– O –Te bridges is seen in the glass spectrum,
5.2. Crystallisation of the glasses
ax ax
which gives reason to think that Te O groups are absent in
2
Let us discuss how the data on the crystallisation of the
tellurite glasses presented in Section 4 can be interpreted
5
the glassy state. Moreover, the spectrum of the Tl Te O
2 5
2

O. Noguera et al. / Journal of Physics and Chemistry of Solids 65 (2004) 981–993
991
If x . ð1 2 2xÞ; i.e. x . 0:33; the limiting reagent in Eq.
(
8) is TeO , and the system would evolve as follows
2
xðTl TeO Þ þ ð1 2 2xÞTeO
2
3
2
¼
If x , 0:33; i.e. the limiting reagent in the right-hand side of
ð1 2 2xÞTl Te O þ ð3x 2 1ÞTl TeO
ð9Þ
2
2
5
2
3
Eq. (8) is Tl TeO , the reaction would pass as
2
3
xðTl TeO Þ þ ð1 2 2xÞTeO
2
3
2
¼
xðTl Te O Þ þ ð1 2 3xÞTeO
ð10Þ
Let us consider in more detail some particular cases
2
2
5
2
corresponding to x varying between 0.5 and 0.
For 0:5 . x . 0:33; the temperature-induced evolution
of the glass would result in the two crystalline structures
having formulas as written in the right-hand side of Eq. (9).
Such a case (with x ¼ 0:38) is presented in Fig. 4a in which
the occurrence of the two initial glassy phases, Tl TeO and
2
3
TeO , is manifested by the characteristic bands near 300 and
2
2
50 cm , respectively. At heating, the latter phase
1
4
2
1
vanishes: the band near 450 cm is present in Fig. 4a but
is absent in Fig. 4(b) and (c) (see inset). Consequently, the
a-Tl Te O crystal lattice appears (bands near 270 and
2
2 5
2
1
6
59 cm in Fig. 4b) with a small amount of the b-lattice
2
1
(
weak peak near 750 cm in Fig. 4b). At the same time, the
remainder of Tl TeO glass crystallizes also, which is
2
3
2
manifested by a sharp line near 727 cm and a weak one
1
2
near 290 cm in Fig. 4c (see inset). Since Tl Te O and
1
2
2 5
Fig. 7. Experimental Raman spectrum for x ¼ 0:18 (solid line) compared
with that obtained as a combination of the spectra for x ¼ 0:21 and x ¼ 0:14
Tl TeO cannot react, the evolution is ended.
2 3
For x ¼ 0:33; the heated glass would directly evolve to a
(
(
dotted line). (b) Decomposition of the Raman spectrum for x ¼ 0:18
single Tl Te O crystalline phase (see Eqs. (9) and (10)).
2
2 5
correction of the baseline) into individual bands (see text).
This case corresponds to Fig. 3B which shows that a small
quantity of the b-Tl Te O phase first appears (at about
2
2 5
within the above-mentioned idea. According to that, the
initial glassy system with composition xTl O þ
140 8C), and then, when the temperature slightly increases,
all the volume quickly develops into the a-Tl Te O
2 2 5
structure.
2
ð1 2 xÞTeO ¼ Tl Te1 2 xO22x would consist of the two
2
2x
phases, and could be presented more adequately (taking into
account Eq. (7)) by expression
For 0:33 . x . 0:25; the Tl Te O phase would be
2 2 5
formed at heating and coexist with the remainder of TeO
Eq. ((10)). At a further temperature-increase, the two phases
on the right-hand side of Eq. (10) would react, so that the
2
xðTl TeO Þ þ ð1 2 2xÞTeO
ð8Þ
2
3
2
which we shall consider as a starting point of our discussion.
The two phases under consideration, namely, the
limiting cases, x ¼ 0:5 and x ¼ 0; are trivial; they
correspond to the mono-phase glasses which at heating
would crystallise into the Tl TeO or TeO2 lattices
final product would contain the Tl
in an amount corresponding to that of TeO
2
Te
3
O
7
tri-tellurite phase
(which is the
2
limiting reagent in this case), and some quantity of
Tl Te O would remain unreacted
2
2 5
2
3
xðTl Te O Þ þ ð1 2 3xÞTeO
2
2
5
2
(
Fig. 3(A) and (D), respectively). In all other cases, i.e.
for 0:5 . x . 0; the two phases indicated in Eq. (8)
would be initially present in the glass separately. At
heating, when the entropy factor becomes sufficient to mix
them, they would enter into reaction due to which a new
Tl Te O phase would appear
¼
This is precisely what can be seen from the spectra in Fig. 5
ð1 2 3xÞTl Te O þ ð4x 2 1ÞTl Te O
ð11Þ
2
3
7
2
2
5
which show first the appearance of the b-Tl Te O lattice
2
2 5
transforming then into the a-Tl Te O , and, finally, the
2 5
2
2
5
2
formation of the Tl Te O lattice coexisting with the
2 3 7
Tl TeO þ TeO ¼ Tl Te O
2
3
2
2
2
5
remainder of the Tl Te O phase.
2
2 5
The pathway of this reaction depends on the relationship
among the two initial components in Eq. (8).
For the case x ¼ 0:25; the system would pass an
intermediate step presented by Eq. (10), and finally would

992
O. Noguera et al. / Journal of Physics and Chemistry of Solids 65 (2004) 981–993
crystallise in the pure tri-tellurite phase Tl Te O (see Eq.
2
the a-TeO structure. This point of the evolution corre-
2
3
7
(
11) with x ¼ 0:25). This is in full accordance with the
sponds to the right- hand side of Eq. (10) with x ¼ 0:14 (i.e.
to composition 0.14 Tl Te O þ 0.58 TeO ).
experimental evidence (Fig. 3C): first, the formation of the
Tl Te O crystalline structure is seen (Fig. 3(b) and (d))
which exists in parallel with a glassy TeO phase. At further
2
2
5
2
At this point, (Fig. 6(b) and (c)), the crystalline
structures, a-Tl Te O and TeO , occur simultaneously
2
2 5
2
2
2
5
2
temperature increase, those two phases disappear simul-
taneously forming the single Tl Te O structure (Fig. 3C-e).
For x , 0:25; after the first stage of the reaction (Eq.
the latter being represented by the a- and g-forms, whereas
no indication of a glassy phase is seen. According to Eq.
(12), this system would finally evolve into the Tl Te O and
2
3 7
2
3 7
(
and the crystallisation process would be ended as
10), the Tl Te O phase would become a limiting reagent,
TeO crystal structures. This corresponds to the spectrum in
2
2
2
5
Fig. 6d in which the latter structure is represented by the a-
TeO phase.
2
xðTl Te O Þ þ ð1 2 3xÞTeO
2
2
5
2
¼
so that the two crystalline phases, Tl Te O and TeO , can
xTl Te O þ ð1 2 4xÞTeO
ð12Þ
2
3
7
2
6
. Summary and conclusion
2
3
7
2
be finally observed. Such a case is presented in Fig. 6 for
x ¼ 0:14; which corresponds to composition 0.14
Tl TeO 0.72 TeO (see expression 8). The initial spectrum
The working ideas of this article can be summarised as
follows.
2
3þ
2
(
Fig. 6a) consists of an intense hump between 400 and
(a) As a starting point, liquid tellurium dioxide TeO is
2
2
00 cm , and a broad band in the 600–800 cm range
1
21
5
dominated by the line near 650 cm . Such a picture is
considered as consisting of TeO molecules, taking
2
2
1
into account a quasi-molecular nature of the para-
tellurite crystalline lattice, a-TeO (the ground state of
TeO2).
typical for tellurite glasses rich in TeO phase.
2
2
At about 250 8C, the situation changes drastically (see
2
Fig. 6b). The sharp peaks near 426, 680 and 820 cm arise
1
(b) Consequently, the initial state of melted tellurites
described by Eq. (3) (with relatively weak cations) can
be considered as containing m moles of the TeO
indicating the appearance of the g-TeO2 lattice. Conse-
21
quently, the line near 650 cm (see Fig. 6a) related to the
2
2
2
TeO glassy phase becomes very weak, and the broad band
2
molecules, nr moles of the O anions and np moles of
þ2r=p
2
1
occupying the 600–800 cm
typical for the Tl TeO glass: its maximum shifts from 650
range assumes the shape
the A
cations. Due to collisions between the
2
2
molecules and the O anions, the nr moles of TeO
2
2
3
2
1
22
to 725 cm
(Fig. 6b) corresponding to synchronous
would transform into nr moles of the TeO3 complex
anions (Eq. (4)), whereas the m 2 nr moles of TeO2
would rest intact. Because of the Coulomb attraction,
stretching vibrations of the TeO3 pyramids. Such an
evolution of the spectrum unequivocally shows that the
glassy TeO phase crystallizses into the g-TeO lattice,
2
2
þ2r=p
the TeO3 complex anions and the A cations
would form the domains of the A (TeO ) composition.
As a result, the neutral TeO molecules, (being drawn
2
2
whereas the glassy Tl TeO phase remains intact.
3
2
p
3 r
This evidence is readily interpretable within the
hypothesis that the initial (totally glassy) sample contains
the two phases, TeO glass and Tl TeO glass, existing
2
from those domains) would unite (owing to the dipole–
dipole attraction forces) into the TeO phase. Thus,
2
2
3
2
separately. Actually, since the molar percentage of the TeO2
phase is of about five times as large as that of the Tl TeO
finally, the melted tellurite would be a two-component
liquid represented by Eq. (7).
2
3
phase (they are in ratio 72:14, respectively), the contribution
of the former phase to the spectrum in Fig. 6a far exceeds
that of the latter phase and makes it poorly observable. At
We wish to underline that the physical factor
providing the main thermodynamic condition for the
phase separation in our system, namely, its free energy
minimisation, is that the electrostatic (monopole–
monopole) interactions between the above-mentioned
charged units, i.e. between the cations and anions, must
be stronger than their interactions with the neutral
units, molecules of TeO2.
heating up to 250 8C, the TeO glass disappears transform-
2
ing into the g-TeO2 lattice. Consequently, the ‘glass
contribution’ to the spectrum is now dominated by that of
the Tl TeO glass which becomes clearly observable (see
2
3
Fig. 6b). Thus, it appears that the evidence in Fig. 6(a) and
b) provides one of the strongest experimental arguments in
favour of the idea stated in the end of Section 2.
(
(c) So, the glass obtained by the rapid cooling of the just-
mentioned liquid would consist of two phases: that of
the TeO2 formulation, and that of the A (TeO )
2
1
Note that the hump between 400 and 500 cm is weak
but not vanishing in Fig. 6b, which indicates that some
quantity of TeO remains in the glassy state. At further
p
3 r
formulation. A variation of the m=n ratio would govern
the proportion between those phases.
2
heating (Fig. 6c) glassy phases of TeO2 and Tl TeO
2
(d) Consequently, the anion subsystem of the nA O þ
p r
22
3
disappear, thus giving rise to the a-Tl Te O lattice (a
2
mTeO tellurite glass would be localized within the
2
5
2
2
characteristic band near 270 cm is clearly observed in
1
A (TeO ) phase in view of the nr (TeO ) ortho-
p
3 r
3
2
)
nr2
Fig. 6c). In parallel, some quantity of g-TeO inverts into
2
groups, and therefore the (Te O
m
complex
nrþ2m

O. Noguera et al. / Journal of Physics and Chemistry of Solids 65 (2004) 981–993
993
anions (existing in the relevant crystalline tellurites)
would be absent in the glass.
Actually, the origin of the immiscibility tendency of
many oxide systems is usually correlated with their ionic
nature, i.e. with the charge–charge interactions of the
constituent ions. To explain the formation of different
phases, the theory appeals to the difference of such
From the above discussions, the following points
were expected for the evolution of the vibrational
spectra of the xTl O þ ð1 2 xÞTeO glass system
2
2
2
caused by the variation of its composition for 0:5 .
x . 0; and by the temperature-induced crystallization:
e) the vibrational spectra of that system at ambient
conditions would be interpretable as a superposition
parameters as ionic field strength z=R ; ionic potential z=R
(z and R are effective charge and radius, respectively),
effective ionic volume, ionic polarizability, etc. In the case
of tellurite glasses, a more simplified concept can be used,
namely, that taking into account the coexistence of the two
types of structural units, charged and neutral ones, and
regarding consequently the two different types of inter-unit
potentials, coulombic ones (charge–charge) and those of
van der Waals (dipole–dipole).
(
(
of the spectra of the TeO and Tl TeO glassy phases.
2
2
3
f) the crystallization of such glasses would be a multi-
step process beginning with an equimolar mixing of
those phases, so that the Tl Te O structure would
2
2 5
always be formed at the first step. In the presence of the
remainder of TeO (if x , 0:33), the system would
2
Generally, the results of the present study favour the idea
that the tellurite glasses (with relatively weak cations) have
a strong tendency to an immiscibility behaviour which
evolve at further heating to the Tl Te O structure
2
3 7
which can form a single phase (if x ¼ 0:25), or can
coexist with crystalline Tl Te O (Eq. (11)) or with
2 2 5
results in the separation into the pure TeO phase and that of
2
crystalline TeO (Eq. (12)).
2
the A (TeO ) formulation having ortho-tellurite structure.
p 3 r
It can be believed that this article can stimulate further more
direct experimental and theoretical investigations aimed to
gain the insight into the structural organisation and crystal-
lization mechanism of those glasses.
The experimental results of our study show that:
–
At room temperature, the composition-dependence of
the Raman spectra of all the thallium tellurite glasses
with 0:5 . x . 0 corresponds to above point (e), i.e.
indicates the presence of the two phases: TeO and
2
Tl TeO .
3
2
References
–
–
In line with point (d), the glass spectra do not indicate
2
the existence of the (Te O
m
nr2
)
complex anions
nrþ2m
[
[
1] R.A.F. El-Mallawany, Tellurite Glasses Handbook: Properties and
Data, CRC Press, Boka Raton, FL, 2002.
which appear at further (temperature-induced) crystal-
lization of the glasses;
2] W. Vogel, Glass Chemistry, Springer, Berlin, 1994.
The temperature-induced crystallization processes pass
as predicted by point (f), and their particularities are
in line with the details preliminarily discussed
in Section 5.2.
[3] A.N. Lazarev, Vibrational spectra and structures of silicates,
Consultants bureau, NY 1972.
[
[
[
4] T. Sekiya, N. Mochida, A. Ohtsuka, M. Tonokawa, J. Non-Cryst.
Solids 144 (1992) 128.
5] A.P. Mirgorodsky, T. Merle-M e´ jean, P. Thomas, J.-C. Champarnaud,
B. Frit, J. Phys. Chem. Solids 63 (2002) 545.
The three above-mentioned facts can be considered as
6] A.P. Mirgorodsky, T. Merle-M e´ jean, J.-C. Champarnaud, P. Thomas,
B. Frit, J. Phys. Chem. Solids 61 (2000) 501.
the basic experimental arguments for the phase separation in
tellurite glasses which was predicted in Section 2 as a
consequence of an immiscibility behaviour inherent for
them. We suppose that the specific factor governing such a
behaviour of those glasses is the presence of the neutral
[7] P.A. Thomas, J. Phys. C: Solid State Phys. 21 (1988) 4611.
[8] K.J. Kingma, R.J. Hemley, Am. Mineralog. 79 (1994) 269.
[
9] D.S. Yakovlev, A.P. Mirgorodsky, A.V. Tulub, B.F. Sschegolev, Opt.
Spectrosc. 92 (2002) 449.
[
10] T. Sekiya, N. Mochida, A. Ohtsuka, M. Tonokawa, Nippon
Seramikkusu Kyokai Gakujutsu Ronbunshi, J. Ceram. Soc. Jpn 97
units (TeO molecules) in the glass-forming melt. This point
2
which seems to be typical for tellurites is not typical for the
classic glass-formers manifesting a phase separation: for
example, the existence of the SiO molecules (homologues
(
1989) 1435.
[11] T. Sekiya, N. Mochida, A. Ohtsuka, J. Non-Cryst. Solids 168 (1992)
28.
1
2
[
12] J.C. Champarnaud-Mesjard, S. Blanchandin, P. Thomas, A.P.
Mirgorodsky, T. Merle-M e´ jean, B. Frit, J. Phys. Chem. Solids 61
of the linear CO molecules) seems to be highly unlikely in
2
liquid silica, much less in silicates. Consequently, the
mechanism of the phase separation in the tellurite glasses
seems to be more physically transparent than that occurring
in the classic ones.
(
2000) 1499.
[
13] O. Noguera, T. Merle-M e´ jean, A.P. Mirgorodsky, M.B. Smirnov, P.
Thomas, J.C. Champarnaud-Mesjard, J. Non-Cryst. Solids 330 (2003)
50–60.