Thermal characteristics, Raman spectra and structural properties of new tellurite glasses within the Bi2O3-TiO2-TeO2 system

Article abstract of DOI:10.1016/j.jssc.2006.06.016

Within the Bi₂O₃-XO₂-TeO₂ (X=Ti, Zr) systems, a large glass-forming domain was found for X=Ti, but no glass formation was evidenced for X=Zr. Densities, glass transition (T_g), crystallization (T_c) temperatures and Raman spectra of the relevant glasses were studied as functions of the composition. The Raman spectra of the glasses were interpreted in terms of the structural transformations produced by the modifiers. It was established that the addition of Bi₂O₃ and TiO₂ content to TeO₂ glass influences the T_g temperature in a similar manner: this value progressively increases with the increase of the modifier concentration. However, the structural evolutions are different: (a) the addition of TiO₂ to TeO₂ glass keeps the polymerized framework structure in transforming a number of Te-O-Te bridges into the Te-O-Ti ones without producing any tellurite anions (i.e., the [TeO₃]^2- groups); (b) on the contrary, the addition of Bi₂O₃ destroys the glass framework by giving rise to the island-type [Te_nO_m]^2(m-2n)- complex tellurites anions, thus causing a depolymerization of the glass.

Full text of DOI:10.1016/j.jssc.2006.06.016

ARTICLE IN PRESS
Journal of Solid State Chemistry 179 (2006) 3252–3259
www.elsevier.com/locate/jssc
Rapid communication
Thermal characteristics, Raman spectra and structural properties of new
tellurite glasses within the Bi O –TiO –TeO system
2
3
2
2
a,b
a,
Ã
a
a
a
a
M. Udovic , P. Thomas , A. Mirgorodsky , O. Durand , M. Soulis , O. Masson ,
a
T. Merle-Mejean , J.C. Champarnaud-Mesjard
´
a
a
Science des Proc e´ d e´ s C e´ ramiques et de Traitements de Surface, UMR 6638 CNRS, Facult e´ des Sciences et Techniques, 123 Avenue Albert Thomas,
7060 Limoges Cedex, France
Jozef Stefan Institute, Jamova 39, 1000 Ljubljana, Slovenia
8
b
Received 20 March 2006; received in revised form 29 May 2006; accepted 18 June 2006
Available online 21 June 2006
Abstract
Within the Bi O –XO –TeO (X ¼ Ti, Zr) systems, a large glass-forming domain was found for X ¼ Ti, but no glass formation was
2
3
2
2
evidenced for X ¼ Zr. Densities, glass transition (T
studied as functions of the composition. The Raman spectra of the glasses were interpreted in terms of the structural transformations
produced by the modifiers. It was established that the addition of Bi and TiO content to TeO glass influences the T temperature in
a similar manner: this value progressively increases with the increase of the modifier concentration. However, the structural evolutions
are different: (a) the addition of TiO to TeO glass keeps the polymerized framework structure in transforming a number of Te–O–Te
bridges into the Te–O–Ti ones without producing any tellurite anions (i.e., the [TeO
g c
), crystallization (T ) temperatures and Raman spectra of the relevant glasses were
2
O
3
2
2
g
2
2
2
ꢀ
3
]
groups); (b) on the contrary, the addition of
2
(mꢀ2n)ꢀ
Bi O destroys the glass framework by giving rise to the island-type [Te O ]
2
complex tellurites anions, thus causing a
3
n
m
depolymerization of the glass.
r 2006 Elsevier Inc. All rights reserved.
Keywords: Tellurite glasses; Thermal analysis; Raman spectroscopy
1
. Introduction
pair related to the 5s orbital of tellurium atom. There is a
good reason to believe that the existence of such pairs is
Increasing interest is paid during the last decade to
TeO -based glasses for their remarkable optical properties,
good thermal stability, chemical durability, broad homo-
geneity range and low melting temperatures (see [1] and
references therein).
It should be emphasized that the non-linear optical
indices of tellurium oxide-based glasses are one of the
highest among oxide systems (more than one order higher
than in silica and silicate glasses), which makes those
materials very promising candidates for non-linear optical
applications and especially for electro-optic devices [1–6].
Customarily, the origin of the extraordinary non-linear
inherent in TeO polymorphs and in tellurites, but strictly
2
speaking, the extraordinary hyperpolarizability of those
pairs was never evidenced. Moreover, the recent ab initio
studies of (TeO ) clusters provide arguments to associate
2
2
n
the above-mentioned extraordinary optic properties with
the essential non-locality of the electron dielectric response,
which is found to be characteristic of the (TeO ) chain-like
2
n
polymers [7–9].
This point seems to be capable of explaining why the
addition of a modifying oxide to TeO glass (which is
2
necessary to ameliorate the glass stability) decreases, as a
rule, the linear and non-linear optical indices [1]. Actually,
2
ꢀ
optical properties of TeO -based glasses is attributed to a
2
to transform this glass into a tellurite structure, the O
high hyperpolarizability of a lone (i.e., localized) electron
atoms delivered by the modifier have to react with the
elementary neutral units of TeO2 glass (which can be
considered as TeO molecules or, more traditionally, as
Ã
Corresponding author. Fax: +33 5 55 45 72 70.
E-mail address: philippe.thomas@unilim.fr (P. Thomas).
2
TeO4/2 disphenoids) to produce a number of the charged
0022-4596/$ - see front matter r 2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.jssc.2006.06.016

ARTICLE IN PRESS
M. Udovic et al. / Journal of Solid State Chemistry 179 (2006) 3252–3259
3253
2
ꢀ
[
three-dimensional TeO framework would evolve into an
TeO3]
pyramid-like units. Consequently, the initial
behaviour of air and ice-quenched glassy samples were
determined by heat flux differential scanning calorimetry
DSC (Netzsch STA 409 apparatus). The powdered samples
(ꢁ3071 mg) were placed into covered gold crucibles under
pure flowing nitrogen atmosphere in order to prevent
potential oxidation of tellurium, and the DSC curves were
recorded between 20 and 800 1C using a heating rate of
2
2
(mꢀ2n)ꢀ
ensemble of complex [Te O ]
m
tellurite anions,
n
separated by cations [10]. As a result, the glass becomes
an island-type structure, and its microscopic polarization
mechanism becomes more localized, and the hyperpolar-
izabilty of the glass as a rule, becomes less pronounced.
However, a series of modifiers can even enhance the non-
linear refractive indices of tellurite glasses [1] by virtue of a
high polarizability inherent to the cations involved. Two
categories of such modifiers can be mentioned. The first
modifiers involve cations of heavy p-elements (e.g. those of
the sixth period , like Tl , Pb and Bi ) predisposed to
form the lone electron pairs in oxides. The second involve
cations of d-elements in which the highest orbitals are
weakly occupied, e.g. W , Nb , Ti , Zr , etc.
Among the last modifiers, TiO and ZrO are of special
interest. Actually, the X-ray diffraction patterns and the
10 1C/min. The glass transition temperature (T ), onset
g
crystallization temperature (T ) and temperature of the first
crystallization peak (T ) were determined as indicated in
o
x
Fig. 2. The densities of the glasses were measured on finely
ground powders by helium pycnometry (Accupyc 1330).
A structural aspect of the glass studies was realized using
Raman spectroscopy. The Raman spectra were recorded in
+
2+
3+
ꢀ
1
the 80–1000 cm range using a Jobin–Yvon spectrometer
6
+
5+
4+
4+
+
(64000 model) equipped with an Ar laser (514.5 nm
exciting line) and a CCD detector in a backscattering
geometry. Using such a detector, a good signal/noise ratio
needed two scans (during 100 s). The sample focalization
was done through a microscope ( ꢂ 100). The diameter of
the laser spot focused on the samples was about 1 mm.
Measurements were performed at low power (o100 mW)
of the excitation line, in order to avoid any damage of the
2
2
2
ꢀ
Raman spectra show that no [TeO ] pyramids, and thus
3
2
(mꢀ2n)ꢀ
no [Te O ]
anions appear in crystalline or glassy
structures within the TeO –TiO or TeO –ZrO systems
n
m
2
2
2
2
[
3,11–16] which would keep the framework nature of such
compounds, thus favouring their high polarizability and
hyperpolarizability.
Taking into account these points, it can be thought that
ternary TeO -based glassy systems including TiO (or
ZrO ) modifier, jointly with a modifier of the first category,
ꢀ
1
glasses. The spectral resolution was about 2.5 cm at the
exciting line.
2
2
3. Results
2
can provide the glasses which would offer the best
compromise on the high non-linear optical characteristics,
on the one hand, and the high mechanical and thermal
resistance, on the other hand. Keeping this in mind, we
have performed, for the first time, the studies of the
properties of the Bi O –TiO (ZrO )–TeO systems.
3.1. Glass formation
Neither by air nor by ice-quenching from 900 1C any
glassy phase was obtained in the Bi O –ZrO –TeO system,
or in the ZrO –TeO system. Crystallization of the a-TeO
2
3
2
2
2
3
2
2
2
2
2
2
Our aim was:
and ZrTe O lattices was always identified preventing the
3 8
formation of a stable glass, in line with early results [17].
Within the binary TiO –TeO system, glasses were
obtained in the range from 10 to 15 mol% of TiO using
(
(
i) to determine the glass formation domain;
ii) to follow the thermal behaviour of the system: i.e. to
2
2
2
establish their glass transition temperature (T ), onset
g
air-quenching. This domain was extended in the range
from 5 to 18 mol% of TiO under ice-quenching condi-
crystallization temperature (T ) and temperature of
o
2
the first crystallization peak (Tx);
iii) to study the structural evolution of the glasses as a
function of the composition.
tions. The results are in perfect accordance with previous
literature data [2,3] describing the glass formation in this
binary system.
(
Practically no glass forming domain was determined
within the Bi O –TeO system: only a very narrow glassy
2
. Experimental
2
3
2
compositional range (from 0 to 4 mol% of BiO1.5) could be
obtained when a small quantity of powder (about 100 mg)
was melted at 900 1C and then ice-quenched. Under such
conditions, a glass forming domain was evidenced up to
15 mol% of BiO1.5 within the Bi O –TiO –TeO system
To obtain glassy samples (weight: ꢁ500 mg), appropriate
amounts of commercial TiO (Aldrich, 99+) or ZrO
2
(
(
2
Aldrich, 99+), Bi O3 (Aldrich, 99.99+) and TeO2
2
prepared in the laboratory by decomposition at 550 1C
2
3
2
2
for 24 h of commercial ortho-telluric acid H TeO (Aldrich,
9
under air-quenched conditions, and was enlarged up to
25 mol% of BiO1.5 by ice-quenching (Fig. 1).
6
6
9.9%)) were first melted at 900 1C for 10 min in platinum
crucibles and then air/or ice-quenched. In order to extend
the glass forming range, the bottom of the platinum
crucible was quickly dipped in a freezing mixture consisting
of ice, ethanol and NaCl ( about ꢀ14 1C). Glass formation
domain was determined by using X-ray diffraction
3.2. Thermal properties and densities of glasses
The characteristics of some glasses are given in Table 1.
As an illustration, the DSC curves of the yBiO1.5
–
(
Guinier-De Wolff camera, CuKa radiation). The thermal
10TiO –(90ꢀy)TeO glasses with 0pyp25 are shown in
1
2
2

ARTICLE IN PRESS
M. Udovic et al. / Journal of Solid State Chemistry 179 (2006) 3252–3259
3254
2 3 2 2
Fig. 1. Glass forming domains at 900 1C under air or ice-quenching conditions within the Bi O –TiO –TeO system (melting time: 10 min; weight of the
samples: 500 mg).
Table 1
g o
Glass transition temperature (T ), onset crystallization temperature (T ),
temperature of the first crystallization peak (T ), thermal stability (T
x
o
g
ꢀT )
and density of glasses
Mol% Mol% Mol%
BiO1.5 TiO
T
g
T
o
T
x
T
o
ꢀT
g
Density
3
(g/cm )
2
TeO
2
(1C)
(1C)
(1C)
(1C)
0
0
0
0
5
5
5
0
0
5
5
0
0
5
5
10
15
18
5
95
90
85
82
90
85
80
85
80
80
75
75
70
65
313
335
356
372
321
338
362
322
341
324
347
332
351
354
361
397
427
446
365
425
441
361
410
352
401
357
390
387
368
407
441
457
375
438
450
368
429
357
416
361
397
405
48
62
71
74
44
87
79
39
69
28
54
25
39
33
5.57
5.47
5.39
5.33
5.73
5.66
5.58
5.94
5.86
6.13
6.04
6.30
6.23
6.44
10
15
5
1
1
1
1
2
2
2
10
5
10
5
10
10
Fig. 2. DSC curves of the yBiO1.5–10TiO
heating rate 10 1C/min).
2
–(90ꢀy)TeO glasses (0pyp25;
2
Fig. 2. (The coefficients in the formula representing various
glass compositions will be given in mol% throughout this
paper.) The thermal stability of the glasses was calculated
with the composition and reaches a maximum of 86.7 1C
for the 5BiO1.5–10TiO –85TeO glass. Increasing Bi O
amount augments the densities of the glasses, and
increasing amount of TiO lowers their densities.
2
2
2
3
as the difference in the first crystallization onset (T ) and
o
2
the glass transition temperature (T ). All the glasses were
g
homogeneous and yellow. It is worth noting that (i) these
data were obtained from glasses prepared under the same
ice-quenching conditions in order to give us comparable
results; (ii) strictly the similar characteristics were mea-
sured, when it was possible, from glassy samples elaborated
under air-quenching conditions.
3.3. Raman spectra
The Raman spectra were considered as a source of an
information about structural properties of the glasses and
their composition dependences. First the study was
performed for the xTiO –(100ꢀx)TeO system in interval
2
2
Thus, it was found that the Tg value increases with
increasing concentration of both TiO2 and Bi O . The
between x ¼ 0 and 18. The spectra of such glasses and of the
related crystalline compounds with x ¼ 0 (a-TeO ), x ¼ 25
2
3
2
thermal stability of the glasses (difference T ꢀT ) changes
(TiTe O ) and x ¼ 100 (TiO rutile) are shown in Fig. 3.
o
g
3
8
2

ARTICLE IN PRESS
M. Udovic et al. / Journal of Solid State Chemistry 179 (2006) 3252–3259
3255
Fig. 3. Raman spectra of the xTiO
2
–(100ꢀx)TeO
2
glasses, and of the
Fig. 4. Raman spectra of the yBiO1.5–5TiO –(95ꢀy)TeO glasses,
2
2
a-TeO , TiTe and TiO crystalline lattices.
2
3
O
8
2
comparing to the spectra of pure TeO2 glass and of the Bi Te O and
2
4
11
2 3
Bi O crystalline lattices.
Then we studied the Raman spectra of the glasses within
the ternary yBiO1.5–xTiO –(100ꢀxꢀy)TeO system in
2
2
varying y, and keeping x ¼ 5, 10 and 15. The results are
given in Figs. 4 and 5, respectively, jointly with the spectra
of the crystalline structures Bi Te O and Bi O (The
2
4
11
2
3
spectra for x ¼ 15 have coincided practically with those for
x ¼ 10, and therefore are not presented here).
To characterize quantitatively the evolution of the bands
observed in the spectra of glasses in Figs. 3–5, their baseline
corrections and decompositions into the Gaussian-like
oscillators, A–D, were made and presented in Fig. 6(a)–(e)
in which, ‘‘by definition’’, (see caption to Fig. 6) oscillator
C keeps its intensity.
Figs. 3 and 6(a)–(c) show that:
(
i) the original (non-baseline-corrected) spectra of the
xTiO –(100ꢀx)TeO glasses in Fig. 3 demonstrate no
2
2
distinct changes with increasing x;
ii) oscillator A dominating the large band 400–500 cm
ꢀ
1
Fig. 5. Raman spectra of the yBiO1.5–10TiO
comparing to the spectra of pure TeO
Bi crystalline lattices.
2
–(90ꢀy)TeO
2
glasses,
(
2
2 4
glass and of the Bi Te O11 and
augments its frequency and relative intensity, and
oscillator B seems to become weaker;
2
O
3
(
iii) the highest frequency part of the spectra i.e. oscillators
C and D, remains practically intact.
4. Discussion
Figs. 4–6(b), (d), (e) show that with increasing content of
Bi O , the following three main points can be stated
Before discussing the relations between Raman spectra
2
3
and structural properties of the binary and ternary glassy
systems, we wish to touch on some peculiarities of the
spectra and chemical bonding in pure TeO glass and in the
regarding the spectra of the yBiO1.5–xTiO –(100ꢀxꢀy)
2
TeO glasses:
2
2
two crystalline lattices, TiTe O [14] and Bi Te O [18]
3
8
2
4
11
ꢀ
1
(
iv) oscillator A displaces from 438 cm
ꢀ
[Fig. 6(b)] to
4
v) oscillator B progressively losses its intensity and
which are related to the two modifiers used in our studies.
Thus, we would like to underline principal differences
between the constitutions of those lattices by analysing the
structures of those (Figs. 7 and 8) jointly with the relevant
Raman spectra (Figs. 3–6) with the aim to extend our
results into the field of the relevant binary and ternary
glasses.
1
15 cm [Fig. 6(e)], and becomes stronger;
(
vanishes at y ¼ 0:18;
(
vi) the frequency of oscillator D drops, whereas its
intensity progressively augments and becomes equal
to that of oscillator C for y ¼ 0:18.

ARTICLE IN PRESS
M. Udovic et al. / Journal of Solid State Chemistry 179 (2006) 3252–3259
3256
3 8
Fig. 7. Projection of the cubic cell of the TiTe O crystal lattice (space
group Iaꢀ3) on the (010) plan. Each black circle corresponds to two Ti
atoms having coinciding projections, which accounts for the doubling
coordination numbers regarding to oxygen atoms. The same can be said
about Te atoms in (x, 0.5, 0.25) and (x, 0.5, 0.75) positions. Only
˚
interatomic bonds shorter than 2 A are marked.
Fig. 6. Decomposition of baseline-corrected Raman spectra of some
glasses of the yBiO1.5–xTiO composition (using the
2
–(100ꢀxꢀy)TeO
2
Gaussian-like oscillators, A–D). The insets show the oscillators character-
istics: frequencies expressed in wavenumbers o, and areas S normalized
with respect to the area of oscillator C which is always kept equal to 1.
The Raman spectrum of pure TeO glass is known for
2
years [1]. Its interpretation justified by model calculations
has been recently given in [19]. The main structural
hypothesis which can be deduced from this is that TeO2
Fig. 8. Perspective view of the Bi Te O unit cell (space group P12 /n1).
2
4
11
1
˚
Only interatomic bonds shorter than 2 A are marked.
glass is a framework made of chain-like fragments in which

ARTICLE IN PRESS
M. Udovic et al. / Journal of Solid State Chemistry 179 (2006) 3252–3259
3257
2
three ortho-ions [TeO3] (see coordination polyhedrons
ꢀ
the TeO units are well polymerized, thus forming strong
2
(mainly covalent) and rather symmetric Te–O–Te intra-
chain bridges in which both the Te–O bonds have lengths
around atoms Te1, Te2 and Te4 in Fig. 8) and one
molecule TeO in which Te atom has two complementary
2
˚
of about 2.0 A. Such bridges are typical neither for the
complex tellurite anions, nor for the pure TeO lattices (see
2
neighbouring oxygen atoms, thus forming a so-called
disphenoid (see the coordination polyhedron around atom
Te3 which forms two strong ‘‘molecular’’ bonds with
atoms O11 and O10, and two much weaker contacts with
atoms O6 and O5). Thus, the strong Te–O–Te bridges are
absent in this structure.
a- and b-TeO structures [19,20]), with the exception of
2
g-TeO2 [21]. Symmetric stretchings vibrations of those
bridges give rise to a large and intense band situated near
ꢀ
1
4
30 cm
(see the spectrum of TeO2 glass in Fig. 3),
described as oscillator A in Fig. 6(a), whereas the
asymmetric stretchings have weak intensities and are
localized above 700 cm (oscillator D). The inter-chain
linkages are realised via ‘‘standard’’ Te–O–Te bridges [i.e.,
Consequently, in its Raman spectrum (Figs. 4 and 5), the
ꢀ
1
area between 400 and 500 cm
whereas the two strong lines are present above 600 cm
is practically empty,
ꢀ1
ꢀ
1
.
ꢀ
1
The highest of them (near 720 cm ) has a position and an
intensity characteristic for the synchronous pulsation
units
[20,22]. The second band (645 cm ) can be attributed to
those which are formed by the so-called axial bonds
˚
2
ꢀ
(
(
having 2.10–2.15 A in length) and equatorial bonds
vibrations of the terminal bonds in the [TeO ]
ꢀ
3
1
˚
1.85–1.90 A) similar to the Te–O–Te linkages in the a-
and b-TeO lattices as well as in many tellurite structures],
the symmetric stretching vibrations of the TeO molecules
2
2
whose symmetric and asymmetric vibrations are situated
in condensed mediums (see e.g. [10,19]). As for the two
ꢀ1
ꢀ
1
ꢀ1
near 480 cm [oscillator B in Fig. 6(a)] and near 660 cm
oscillator C in Fig. 6(a)], respectively.
The Raman spectrum of crystalline TiTe O (Fig. 3) is
bands around 400 cm , our lattice dynamical calculations
show that they are associated with Te–O–Bi linkages.
[
It is still worth underlining that the absence of intense
ꢀ1
3
8
ꢀ
1
dominated by a relatively strong band near 440 cm and a
lines between 400 and 500 cm
in the spectrum just
ꢀ
1
weaker band near 650 cm (a high frequency part of the
infrared spectrum of this lattice can be found in Ref. [16]).
It must be particularly underlined that the absence of
strong bands in the high-frequency (stretching) region of
the spectrum shows that this structure is a framework [15]
in which neither Ti–O bonds, nor Te–O bonds can be
considered as terminal ones, but forming the covalently
bonded Te–O–Ti bridges. Therefore, the interpretation of
those bands in terms of localized Ti–O or Te–O bond
vibrations is unambiguous: actually, they correspond to the
motions of oxygen atoms in the potential wells formed by
the Te–O–Ti linkages (Fig. 7).
indicates that the neutral TeO molecules and the charged
2
2
[TeO3] pyramids form the [Te O ] complex tellurite
ꢀ
6ꢀ
4
11
anions without polymerizing via the Te–O–Te linkage.
According to the Raman spectroscopy data, such a
situation seems to be customary for complex tellurite
structures (e.g. Refs. [10,22]).
In line with this, the evolution of TeO glass into a
2
tellurite structure is associated, as a rule, with depolymer-
ization effects because of which the large and intense band
ꢀ
1
400–500 cm inherent in the spectrum of pure TeO glass
2
eventually disappears. During this evolution, the formation
ꢀ
2
of the [TeO3] ortho-anions gives rise to a strong Raman-
ꢀ
1
active band near 720–730 cm , whereas the TeO mole-
Moreover, we venture the opinion, that from the formal
chemical point of view, the TiTe O lattice cannot be
considered as a tellurite structure: actually, neither simple
2
cules which rest ‘‘intact’’ manifest themselves by the band
3
8
ꢀ
1
near 650 cm . Therefore, it can be said that the high-
frequency part of the Raman spectrum of the Bi Te O
11
2
ꢀ
4ꢀ
[
there (see Fig. 7) in contrast to structures like Zn Te O , in
TeO3] ortho-anions, nor complex [Te O ] anions exist
3
8
2
4
lattice has the features of ‘‘classic’’ spectra of the
2
2
3
8
(mꢀ2n)ꢀ
which such complex anions are the fundamental structural
units [16]. So it can be specified as a crystalline phase of the
[Te O ]
m
complex tellurite anions which are dis-
n
cussed in [10,15].
2
addition, we wish to remark that the resemblance of the
band positions in the spectra of pure TeO glass and of
2
5TiO –75TeO ‘‘solid solution’’ of TiO2 in TeO . In
2 2 2
4.1. xTiO –(100ꢀx)TeO glasses
2
2
TiTe O structures (Fig. 3) is dictated by the resemblance
3
The lowering glass density (over all the glass forming
domain from 5 to 18 mol% of TiO ) with increasing x can
be naturally associated with the contribution of relatively
lightweight atoms of titanium. The densities of obtained
glasses are quite comparable to those previously reported
in literature [23].
8
of the chemical nature of the binding between the oxygen
atoms and their nearest neighbours, and, correspondingly,
of the lengths of the bonds forming the Te–O–Te and
Ti–O–Te bridges in those structures, respectively.
2
The spectrum of the Bi Te O lattice [Figs. 4 and 5]
4
2
11
reflects a significant difference between its chemical
constitution and that of the TiTe O lattice. Actually,
The main features of the structural evolution of the
xTiO –(100ꢀx)TeO glasses at increasing x should be
3
8
2
2
2
(mꢀ2n)ꢀ
while no [Te O ]
n
tellurite anions are present in the
6
deduced from the interpretation of the changes in the
spectra in Figs. 3 and 6(a)–(c), i.e., from the explication of
the items (i)–(iii) indicated in Section 3.3.
m
ꢀ
latter structure (Fig. 7), the [Te O ]
4
anions can be
11
readily recognized in the former (Fig. 8). According to a
schema proposed in [10], those complex anions can be
regarded as framed of the four dipole structural blocks:
First, we concern with the item (i). The absence of
the bands related to pure titanium dioxide framework

ARTICLE IN PRESS
M. Udovic et al. / Journal of Solid State Chemistry 179 (2006) 3252–3259
3258
(
characterized by the three-fold coordination of atoms of
Within this ternary system, for a fixed concentration of
TiO , the increasing content of Bi O provides a practically
oxygen) in the spectra of the xTiO –(100ꢀx) glassy
2
2
2
3
samples (see Fig. 3) allows us to think that those are
monophase systems in which TiO6 polyhedrons are
interconnected via TeO4/2 disphenoids like in the TiTe O
linear increase of glass densities which is readily explained
by the very high molecular weight of the latter oxide.
We focus now on the evolution of the spectra in
Figs. 4–6(b), (d), (e) with the aim to comment the items
(iv)–(vi) in Section 3.3 in terms of the structural changes in
the Bi O –TiO –TeO glasses. A sharp strengthening of the
3
8
lattice. This implies the existence of the two types of
bridges in those glasses: Te–O–Ti and Te–O–Te bridges.
The former bridges would manifest their presence in the
spectra by bands similar to those observed for TiTe O .
However, positions and relatively weak intensities of those
bands (see Fig. 3) cannot influence dramatically the initial
shape of the spectrum (i.e. that of pure TeO glass), which
2
3
2
2
oscillator D which lowers its position with increasing y [see
Fig. 6(b), (d), (e)] indicates beyond any reasonable doubt
3
8
2
ꢀ
that an increasing number of [TeO3] ortho-groups are
formed in the glasses, and the latter evaluate to a ‘‘classic’’
tellurite structures with complex anions resembling those in
2
can account for point (i).
Now we concentrate on the items (ii) and (iii). Note first
that the peculiarities of the evolution of spectra indicated
there resemble those observed in the Raman spectra of
the Bi Te O
4
lattice. In other words, an increasing
11
2
modifier (Bi O ) content provokes a ‘‘normal’’ progressive
2
3
transformation of the TeO2 glass network into a typic
tellurite structure, i.e., into an ensemble of alternate
complex tellurite anions and monoatomic cations.
pure TeO during its glass–liquid transformation [19] at
2
heating. A vanishing of the weak shoulder of the band
4
wing above 700 cm (see Fig. 4 in [19]), was interpreted as
ꢀ
1
00–500 cm simultaneously with the strengthening of the
The evolution of the band occupying interval between
ꢀ1
ꢀ
1
400 and 500 cm (items iv and v), is in line with the above
1
ꢀ
indication of the breaking of the inter-chain bridges in
TeO at melting leading to the appearance of the Te–O
terminal bonds in the liquid.
conclusion. Actually, its shift towards 400 cm
just
indicates a decay of the initial framework (y ¼ 0) and its
depolymerization with increasing y, i.e. the replacement of
the strong Te–O–Te and Te–O–Ti bridges by weaker
Te–O–Bi linkages which were discussed above, when
considering the spectrum of crystalline Bi Te O . It can
2
The similar idea comes in mind when analysing the
spectra in Figs. 3 and 6(a)–(c). Reasoning from the above-
mentioned peculiarities of the crystal chemistry of the
TiTe O lattice, we are led to conclude that the glass
2
4
11
be added, that, theoretically, the formation of any
3
8
structures within the xTiO –(100ꢀx)TeO system can be
yBiO1.5–xTiO –(100ꢀxꢀy)TeO glass structure implies a
2
2
2
2
specified in terms of solid solutions of TiO in TeO , as it
2
possibility for appearance of another types of bridges like
Ti–O–Ti, Bi–O–Bi and Ti–O–Bi, but none of them can be
‘‘revealed’’ from the spectra.
2
was proposed above for the TiTe O crystalline lattice.
3
8
Actually, the formation of those glasses with increasing x
would change the coordination polyhedrons neither
around Ti atom nor around Te atom, and would result
in the following alterations: (a) the disappearance of the
We note that the transformation of glassy TeO network
2
induced by an addition of modifiers with cations having a
valence of one or two, as a rule, reduces the T values [1].
g
‘
decay of oscillator B), which are weak since they involve
˚
the weak axial Te–O bonds (about 2.15 A in length); (b) the
formation of stronger Te–O–Ti bridges (made of two
˚
relatively strong bonds having 1.95 A in length) replacing
just mentioned Te–O–Te linkages. These account for the
items (ii) and (iii), respectively.
‘standard’’ inter-chain Te–O–Te linkages (leading to a
Therefore, it is important to underline that the T values of
g
our glasses always increases with the progressive additions
of TiO2 or Bi O . Thus, it can be suggested that the
2
3
substitution of the Te–O–Te bridges by new Te–O–M
bridges in which the M-elements have valences of three,
four and higher, would augment thermal and mechanical
resistance of tellurite glasses, whether they are of frame-
work-type or not.
The fact of such a transformation is indirectly confirmed
by our another observations: glass transition temperature
T increases with increasing TiO content, thus indicating
5. Summary and conclusions
g
2
the substitution of the Te–O–Te bridges inherent of pure
TeO2 by the stronger chemical linkages Te–O–Ti. In
general, our interpretation of the experimental data is in
line with results early obtained by Sabadel et al. [3,11,12].
A glass forming domain was evidenced in the yBiO1.5
–
xTiO –(100ꢀxꢀy)TeO system, but no glass formation
2
2
was observed in the yBiO –xZrO –(100ꢀxꢀy)TeO
1
.5
2
2
system. For y ¼ 0, the evolution of the Raman spectra of
4
.2. Bi O –TiO –TeO glasses
2
the TiO –TeO glasses at increasing TiO content allows us
2
2
3
2
2
2
to conclude that no tellurite anions are built up in that
structure, and the Te–O–Te linkages are progressively
substituted by Te–O–Ti, due to which the framework
arrangement of the glass is kept. In contrast to this, the
addition of Bi O results in breaking off the initial glass
Under ice-quenching conditions, the glass forming
domain for the Bi O –TeO binary system extends in a
2
3
2
very narrow range of composition (only up to ꢁ4 mol% of
BiO ). However, the addition to this system of the third
1
.5
2
3
component, TiO , results in a broad glass domain
2
extending in some cases up to 25 mol% of BiO1.5 (Fig. 1).
framework, thus provoking its depolymerization, and in
forming local complex tellurite anions interconnected via

ARTICLE IN PRESS
M. Udovic et al. / Journal of Solid State Chemistry 179 (2006) 3252–3259
3259
atoms of Bi. The increase of the T values observed in both
g
[8] O. Noguera, M. Smirnov, A.P. Mirgorodsky, T. Merle-Me
´
jean, P.
Thomas, J.C. Champarnaud-Mesjard, Phys. Rev. B 68 (2003)
94203/1–094203/10.
9] A.P. Mirgorodsky, M. Smirnov, M. Soulis, P. Thomas, T. Merle-
Mejean, Phys. Rev. B 73 (2006) 134206/1–134206/13.
10] O. Noguera, T. Merle-Mejean, A.P. Mirgorodsky, P. Thomas, J.C.
the cases indicates a relative strength of the Ti–O–Te and
Bi–O–Te bridges replacing the Te–O–Te ones.
0
[
´
[
[
[
[
´
Acknowledgments
Champarnaud-Mesjard, J. Phys. Chem. Solids 65 (2004) 981–993.
11] J.C. Sabadel, P. Armand, P.E. Lippens, D. Cachau-Herreilat, E.
Philippot, J. Non-Cryst. Solids 244 (1999) 143–150.
Three of us, M. Udovic, O. Durand and M. Soulis are
grateful to the Conseil Regional du Limousin for financial
´
support.
12] J.C. Sabadel, P. Armand, F. Terki, J. Pelous, D. Cachau-Herreilat, E.
Philippot, J. Phys. Chem. Solids 61 (2000) 1745–1750.
13] J.C. Sabadel, P. Armand, E. Philippot, Phys. Chem. Glasses 41 (1)
(
2000) 17–23.
References
[
[
14] G. Meunier, J. Galy, Acta Crystallogr. B 27 (1971) 602–608.
15] O. Noguera, Thesis of the University, No. 48-1999, Limoges,
November 2004.
[
1] R.A.F. El-Mallawany, Tellurite Glasses Handbook: Properties and
Data, CRC Press, Boca Raton, FL, 2002.
[16] M. Arnaudov, V. Dimitrov, Y. Dimitriev, L. Markova, Mater. Res.
Bull. 17 (1982) 1121–1129.
[
[
2] K. Sae-Hoon, T. Yoko, J. Am. Ceram. Soc. 78 (4) (1995) 1061–1065.
3] J.C. Sabadel, P. Armand, D. Cachau-Hereillat, P. Baldeck, O. Foclot,
A. Ibanez, E. Philippot, J. Solid State Chem. 132 (1997) 411–419.
4] S. Le Boiteux, P. Segonds, L. Canioni, L. Sarger, E. Fargin, T.
Cardinal, C. Duchesne, G. Le Flem, J. Appl. Phys. 81 (1997)
[17] T. Sekiya, M. Mochida, A. Soejima, J. Non-Cryst. Solids 191 (1995)
115–123.
[
[
[18] H.J. Rossel, M. Leblanc, G. Fe
M.R. Taylor, Aust. J. Chem. 45 (1992) 1415–1425.
[19] O. Noguera, T. Merle-Mejean, A.P. Mirgorodsky, M. Smirnov, P.
´
rey, D.J.M. Bevan, D.J. Simpson,
1
481–1487.
5] B. Jeansannetas, S. Blanchandin, P. Thomas, P. Marchet, J.C.
Champarnaud-Mesjard, T. Merle-Mejean, B. Frit, V. Nazabal, E.
´
Thomas, J.C. Champarnaud-Mesjard, J. Non-Cryst. Solids 330
(2003) 50–60.
[20] A.P. Mirgorodsky, T. Merle-Mejean, J.C. Champarnaud-Mesjard,
´
´
Fargin, G. Le Flem, M.O. Martin, B. Bousquet, L. Canioni, S. Le
Boiteux, P. Segonds, L. Sarger, J. Solid State Chem. 146 (1999)
P. Thomas, B. Frit, J. Phys. Chem. Solids 61 (2000) 501–509.
[21] J.C. Champarnaud-Mesjard, S. Blanchandin, P. Thomas, A.P.
3
29–335.
6] J. Lin, W. Huang, Z. Sun, C.S. Ray, D.E. Day, J. Non-Cryst. Solids
36 (2004) 189–194.
7] O. Noguera, M. Smirnov, A.P. Mirgorodsky, T. Merle-Me
Thomas, J.C. Champarnaud-Mesjard, J. Non-Cryst. Solids 345&346
2004) 734–737.
[
[
´
Mirgorodsky, T. Merle-Mejean, B. Frit, J. Phys. Chem. Solids 61
(2000) 1499–1507.
3
´
jean, P.
[22] T. Sekiya, M. Mochida, A. Ohtsuka, M. Tonokawa, J. Non-Cryst.
Solids 144 (1992) 128–144.
[23] V. Dimitrov, T. Komatsu, J. Non-Cryst. Solids 249 (1999) 160–179.
(