Вы находитесь на странице: 1из 11

Journal of Non-Crystalline Solids 242 (1998) 154164

Thermal stability and heat capacity changes at the glass


transition in K2OWO3TeO2 glasses
T. Kosuge a, Y. Benino a, V. Dimitrov a, R. Sato b, T. Komatsu
b

a,*

a
Department of Chemistry, Nagaoka University of Technology, Kamitomioka-cho, Nagaoka 940-2188, Japan
Department of Material Engineering, Tsuruoka National College of Technology, Oaza-Ioka, Tsuruoka 997, Japan

Received 9 April 1998; received in revised form 10 July 1998

Abstract
The thermal stability and heat capacity changes in the glass transition region of K2 OWO3 TeO2 glasses (glass
formation range 2090 mol% TeO2 ) have been studied to examine the structural relaxation behavior. The glasses with
6070 mol% TeO2 and with both K2 O and WO3 are thermally stable against crystallization. It is proposed from Raman
spectral analyses that TeO4 trigonal bipyramids change to TeO3 trigonal pyramids with the addition of K2 O and that
TeOW bonds are formed in the substitution of WO3 for TeO2 . Heat capacity changes of DCp 4858 J mol1 K1
(DCp Cpl Cpg , Cpg and Cpl are the heat capacities of the glasses and supercooled liquids, respectively), and ratios
Cpl =Cpg 1.61.8 are obtained for xK2 O  xWO3  (100 ) 2x)TeO2 glasses. The DCp and Cpl =Cpg increase with decreasing TeO2 content, indicating an increase in thermodynamic fragility with decreasing TeO2 content. But, the kinetic
fragility estimated from the activation energy for viscous ow is almost constant irrespective of TeO2 content. These
behaviors have been analyzed using the congurational entropy model proposed by Adam and Gibbs. The results
indicate that in K2 OWO3 TeO2 glasses, TeOTe bonds are weak and bond breakings occur easily in the glass
transition region, leading to large congurational entropy changes and thus large DCp . 1998 Elsevier Science B.V.
All rights reserved.

1. Introduction
Tellurium oxide (TeO2 )-based glasses are of
scientic and technical interest on account of their
various unique properties, and have been considered as promising materials for use in optical
ampliers because of their low phonon energy or
nonlinear optical devices because of their large
third-order non-linear susceptibility [1,2]. Recently, optically transparent TeO2 -based glass-ceramics showing second harmonic generation have been

*
Corresponding author. Tel.: +81-258 47 9313; fax: +81-258
47 9300; e-mail: komatsu@chem.nagaokaut.ac.jp

discovered [3,4]. The structure of TeO2 -based


glasses is also of interest, because there are two
types of the basic structural units, i.e. TeO4 trigonal bipyramid (tbp) and TeO3 trigonal pyramid
(tp). In order to develop TeO2 -based glasses as
new optical functional materials, an understanding
of the thermal stability against crystallization and
the structural relaxation behavior in the glass
transition region is necessary, but such information is scarce.
It is known that a pure TeO2 does not become a
glass under usual quenching rates and the addition
of other elements is needed to form TeO2 -based
bulk glasses. Imaoka and Yamazaki [5] examined
the glass-forming regions in various TeO2 -based

0022-3093/98/$ see front matter 1998 Elsevier Science B.V. All rights reserved.
PII: S 0 0 2 2 - 3 0 9 3 ( 9 8 ) 0 0 8 0 0 - X

T. Kosuge et al. / Journal of Non-Crystalline Solids 242 (1998) 154164

systems. Among the glass-forming regions reported in their paper, it is noted that the systems of
R2 OWO3 TeO2 (R: Li, Na, K) have extremely
wide glass-forming regions. These results suggest
that R2 OWO3 TeO2 glasses, which are available
with a wide range in TeO2 content, are suitable for
the study of the structure, properties, thermal
stability, structural relaxation and crystallization
behaviors of TeO2 -containing glasses.
In this paper, the thermal stability against
crystallization and the heat capacity changes in the
glass transition region of various K2 OWO3 TeO2
glasses have been examined. The structural changes due to the substitution of K2 O and WO3 for
TeO2 have been studied using Raman scattering
spectroscopy. To the best of our knowledge, the

155

study of the structural relaxation behavior of


TeO2 -based glasses by measurements of heat capacity is limited to a few systems of Li2 OTeO2 ,
AgIT1O0:5 TeO2 , Na2 OTeO2 and Li2 ONa2 O
TeO2 [610].
2. Experimental procedure
The nominal compositions examined in the
present study are shown in Table 1. Commercial
powders of reagent K2 CO3 (99.5%), WO3 (99%)
and TeO2 (99%) were mixed and melted in a platinum crucible at around 800C for 30 min in an
electric furnace. The batch weight was 20 g. The
quenched glasses were prepared by pouring the

Table 1
Values of glass transition, Tg , crystallization onset, Tx , and melting, Tm , temperatures, density, q, and refractive index, n, for
K2 OWO3 TeO2 glasses
Glasses

Tg (C)

Tx (C)

Tm (C)

q (g/cm3 )

K2 O

WO3

TeO2

(2)

(2)

(2)

(0.01)

(0.02)

0
5
10
0
10
15
20
0
10
15
20
5
10
15
20
25
22.5
23.75
15
20
25
27.5
15
20
25
20
25
30
25
30

10
5
0
20
10
5
0
30
20
15
10
35
30
25
20
15
22.5
23.75
35
30
25
27.5
45
40
35
50
45
40
55
50

90
90
90
80
80
80
80
70
70
70
70
60
60
60
60
60
55
52.5
50
50
50
45
40
40
40
30
30
30
20
20

330
306
281
353
308
275
235
368
344
307
269
388
361
331
300
252
296
294
370
331
281
280
396
363
314
380
345
290
369
325

427
416
416
506
456
370
320
536
509

628
687
459
630
619
487
630
604

416

454

5.79
5.60
5.08
5.97
5.31
4.98
4.57
6.11
5.50
5.13
4.79
5.94
5.61
5.35
5.02
4.62
4.96

2.17
2.11
2.05
2.17
2.04
1.99
1.92
2.16
2.05
1.99
1.92
2.11
2.05
2.00
1.92
1.86
1.90

538

636

419
375
511
486
452
458
439
377
439
422

482
477
665
630
504
684
544
511
550
550

5.51
5.13
4.83
4.82

1.99
1.92
1.87
1.84

5.32
5.08

1.92
1.86

5.21

1.85

5.02

1.62

156

T. Kosuge et al. / Journal of Non-Crystalline Solids 242 (1998) 154164

melts onto an iron plate and by pressing the sample


rapidly to a thickness of 12 mm using another iron
plate. The glassy state in the quenched samples was
conrmed by X-ray diraction (XRD) analyses at
room temperature using CuK a radiation.
Glass transition, Tg and crystallization onset, Tx ,
temperatures were determined using dierential
thermal analyses (DTA) at a heating rate of 10 K
min1 . Densities of the glasses were determined by
the Archimedes method using distilled water as the
immersion liquid. Refractive indices at a wavelength of 632.8 nm (HeNe laser) were measured
at room temperature using an ellipsometer (Mizojiri Optical, DVA-36 model). The relative permittivity, er , at room temperature in the frequency
range from 0.1 to 100 kHz was measured by an
LCR meter. Ion-sputtered gold lms were used as
electrodes for the measurements. Viscosities in the
glass transition region (107 1011 Pa) were measured by a penetration method within an accuracy
of 5%. Raman scattering spectra at room temperature for the glasses were measured in the wave
number range from 200 to 1000 cm1 using an Ar
laser (wavelength 514.5 nm, laser power 100 mW)
and with a JASCO NR-1100 type Laser Raman
Spectrometer.
The annealed glasses, which were obtained by
annealing the quenched glasses at a temperature Tg
+10C for 15 min and by cooling to a temperature
below Tg at a rate of 1 K min1 , were used for the
measurements of heat capacity changes. Heat capacities, Cp , in the glass transition region in a
heating process were measured using dierential
scanning calorimetry, DSC, (Rigaku Thermoex
TAS 200, DSC8230D). The heating rate was 10 K
min1 . The sample weight was around 30 mg.
Standard specimens of alumina (sapphire) were
used as a heat capacity standard. The molecular
weight of glasses was calculated based on batch
composition.
3. Results
3.1. Some properties and thermal stability
The glass-forming region in the K2 OWO3
TeO2 system obtained in the present study is

Fig. 1. Glass-forming region in the K2 OWO3 TeO2 system.


(s) glass, (g) partially crystallized, (d) crystallized.

shown in Fig. 1. It should be pointed out that glass


formation is observed in compositions with 2090
mol% TeO2 . The region shown in Fig. 1 is slightly
narrower at low TeO2 content than that reported
by Imaoka and Yamazaki [5]. The DTA curves of
xK2 O  xWO3  (100 ) 2x)TeO2 glasses with
x 10, 15 and 25 are shown in Fig. 2 as an example. In the glass with x 10, the glass transition
of Tg 308 C and a crystallization onset of
Tx 456 C are observed. The dierence between
Tg and Tx , i.e. DT Tx Tg , in this glass is 148 C,
indicating that the thermal stability against crystallization is considerably high. In the glass with
x 15, a glass transition of Tg 307 C, is observed, but no crystallization peak is detected. A
similar DTA pattern showing no clear crystallization peak was obtained for the glass with x 20.
In the glass with x 25, a crystallization peak is
clearly observed at around 419 C. The values of
Tg , Tx and the melting temperature, Tm , for the
other glasses are given in Table 1. The thermally
stable glasses, in which crystallization peaks were
not observed in the DTA curves taken at a heating
rate of 10 K min1 , are obtained at compositions
with 6070 mol% TeO2 and with both K2 O and
WO3 . As seen in Table 1, the substitution of K2 O
for TeO2 or WO3 gives a rapid decrease in Tg . On

T. Kosuge et al. / Journal of Non-Crystalline Solids 242 (1998) 154164

157

larizability of the glasses is comparable to that of


TeO2 . It is found from Table 1 that the refractive
index increases almost linearly with increasing
density. The frequency dependence of dielectric
constant, er , at room temperature for xK2 O 
xWO3  (100 ) 2x)TeO2 glasses is shown in Fig. 3.
The values of er decrease gradually with decreasing
TeO2 content. But, the glass with x 25 still has a
large value of er 19, implying again that the
polarizability of WO3 is large.
The activation energies for viscous ow, Eg, in
the glass transition region for xK2 O  xWO3 
(100 ) 2x)TeO2 glasses were evaluated by using
the Arrhenius equation, g A exp Eg=RT , where
g is a viscosity, A is a constant, R is the gas constant and T is a temperature. The evaluated values
of Eg for these glasses are given in Table 2. The
experimental error in Eg is around 3%. It is seen
that Eg is in the range of 500552 kJ mol1 and
decreases slightly with decreasing TeO2 content.
These values are comparable to those (418595 kJ
mol1 ) of (20 ) x)Li2 O  xNa2 O  80TeO2 glasses
[11], but are smaller than those (730870 kJ mol1 )
of xLi2 O  (100 ) x)TeO2 glasses [6].

Fig. 2. DTA patterns for some xK2 O  xWO3  (100 ) 2x) TeO2
glasses. Tg and Tx are glass trasition and crystallization onset
temperatures, respectively. Heating rate was 10 K min1 .

the other hand, Tg increases due to the substitution


of WO3 for TeO2 . These results suggest that K2 O
acts as network modier and breaks the network
structure and WO3 would take part in the network
structure as network former.
The values of density, q, and refractive index, n,
for the K2 OWO3 TeO2 glasses are given in Table 1. For xK2 O  xWO3  (100 ) 2x)TeO2 glasses,
they decrease almost linearly with decreasing TeO2
content. For the (30 ) x)K2 O  xWO3  70TeO2
glasses, they increase with increasing WO3 content.
As seen in Table 1, for example, the value of
n 2:05 for the 10K2 O  90TeO2 glass is the same
as that for the 10K2 O  20WO3  70TeO2 glass,
implying that the contribution of WO3 to the po-

Fig. 3. Relative permittivity, er , at room temperature in the


frequency range of 0.1100 kHz for xK2 O  xWO3 
(100 ) 2x)TeO2 glasses. (X) x 5, (s) x 10, (m) x 15, (d)
x 20, (n) x 25.

158

T. Kosuge et al. / Journal of Non-Crystalline Solids 242 (1998) 154164

Table 2
Values of the glass transition temperature, Tg , the activation energy for viscous ow near the glass transition temperature, Eg, the Eg/
Tg ratio, the degree of fragility estimated from Eq. (1), m, the heat capacities of glasses, Cpg , and supercooled liquids, Cpl , the heat
capacity change in the glass transition region, D Cp Cpl Cpg , and the product of Tg and D Cp for xK2 O xWO3 (100 ) 2x)TeO2
glasses
Sample

Tg (K)

Eg (kJ mol1 )

Eg/Tg (kJ mol1 K1 )

Cpg

Cpl (J mol1 K1 )

DCp

Tg D Cp (J mol1 )

x5
x 10
x 15
x 20
x 25
x 27.5

579
581
580
573
554
553

552
534
511
500

0.95
0.92
0.89
0.90

50
48
47
47

72
73
74
75
73
72

120
121
122
126
130
130

48
48
48
51
57
58

)28
)28
)28
)29
)31
)32

Fig. 4. Raman scattering spectra at room temperature for


xK2 O  xWO3  (100 ) 2x)TeO2 glasses.

3.2. Raman scattering spectra


The Raman scattering spectra for xK2 O  xWO3 
(100 ) 2x)TeO2 glasses are shown in Fig. 4.
Similar Raman spectra are obtained for 20K2 O 
xWO3  (100 ) x)TeO2 glasses. Each spectrum was
deconvoluted by using seven symmetrical Gaussian functions, considering peak assignments in
Raman scattering spectra for K2 OTeO2 and
WO3 TeO2 glasses reported so far [1214]. As an
example, the tting result for 25K2 O  25WO3 
50TeO2 glass is shown in Fig. 5. The peaks at
around 470, 610, 670, 730, 790, 860 and 920 cm1
are designated here as peak A, B, C, D, E, F and
G, respectively. The relative peak intensity of
seven peaks obtained by peak deconvolutions
in xK2 O  xWO3  (100 ) 2x)TeO2 and 20K2 O 

000
000
000
000
000
000

Fig. 5. Raman scattering spectrum at room temperature


25K2 O  25WO3  50TeO2 glass. The spectrum was deconvoluted by using symmetrical Gaussian functions.

xWO3  (80 ) x)TeO2 glasses are shown in Fig. 6.


The present peak assignments are performed in
accordance with Raman spectra data cited above
[1214]. Peak A at around 460 cm1 is assigned to
symmetrical stretching vibrations of TeOTe
linkages, Peak B at around 615 cm1 is assigned to
a vibration of the continuous network composed
of TeO4 tbp, Peak C at around 670 cm1 is assigned to antisymmetric vibrations of TeOTe
linkages constructed by two unequivalent TeO
bonds, Peak D at around 720 cm1 is assigned to
stretching vibrations between Te and non-bridging
oxygen (NBO) of TeO31 polyhedra and TeO3 tp,
Peak E at around 780 cm1 is assigned to TeO
stretching vibrations of TeO31 polyhedra and
TeO3 tp. The Peaks F and G at around 840 cm1
and 930 cm1 , respectively, are connected mainly

T. Kosuge et al. / Journal of Non-Crystalline Solids 242 (1998) 154164

159

obtained. Since the intensity of the peaks A and C


decreases rapidly with increasing K2 O or WO3
content, the number of TeOTe linkage decreases, indicating a decrease in network connectivity.
The intensity of Peaks D and E in xK2 O  xWO3 
(100 ) 2x)TeO2 glasses increases with increasing K2 O and WO3 content, but that in 20K2 O 
xWO3  (80 ) x)TeO2 glasses is almost constant
with WO3 content. This means that TeO4 tbp
structures change to TeO31 or TeO3 tp with increasing K2 O content, but such a change is not
generated from the addition of only WO3 . Since
the intensity of Peak A, attributable to symmetrical stretching vibrations of TeOTe linkages in
20K2 O  xWO3  (80 ) x)TeO2 glasses, decreases
with increasing WO3 content, but the intensity of
Peaks D and E, attributable to stretching vibrations related to TeO31 polyhedra and TeO3 tp,
does not change with increasing WO3 content, the
formation of TeOW bonds is strongly suggested.
Indeed, the formation of WOTe linkages in
WO3 TeO2 glasses has been proposed [12,1417].
The position of Peak G, which is distinct from
the other peaks, is shown in Fig. 7. It is seen that
the vibration frequency of this peak xK2 O  xWO3 

Fig. 6. Relative intensities of seven peaks in Raman scattering


spectra for (a) xK2 O  xWO3  (100 ) 2x)TeO2 glasses and for
(b) 20K2 O  xWO3  80 xTeO2 glasses. (n) peak A at around
470 cm1 , (d) peak B at around 610 cm1 (m) peak C at around
670 cm1 , () Peak D at around 730 cm1 , (h) Peak E at
around 790 cm1 , (s), peak F at around 860 cm1 and (n) peak
G at around 920 cm1 . The lines are drawn as a guide for the
eye.

with tungsten-oxygen stretching vibration of both


WO4 and WO6 groups [1316]. It is dicult to
separate them because the neighbouring frequency
ranges of the dierent vibrating species. In this
relation, Peak G, which is not overlapped with
other peaks, can be assigned to the stretching vibrations of WO and W@O terminal bonds associated with WO4 and WO6 polyhedra,
respectively [1316].
From Fig. 6, the following information on the
structural changes in K2 OWO3 TeO2 glasses is

Fig. 7. Positions of the peak at around 920 cm1 in Raman


spectra for (s) xK2 O  xWO3  (100 ) 2x)TeO2 glasses and for
(d) 20K2 O  xWO3  80 xTeO2 glasses. The lines are drawn
as a guide for the eye.

160

T. Kosuge et al. / Journal of Non-Crystalline Solids 242 (1998) 154164

(100 ) 2x)TeO2 glasses is almost constant, while in


the spectra of 20K2 O  xWO3  (80 ) x)TeO2
glasses it increases with increasing WO3 content. A
similar shift has been reported in infrared spectra
of WO3 TeO2 glasses by Dimitrov et al. [15]. They
suggested that when the WO3 concentration is
small, the tungsten enters the glass structure in
WO4 groups to form TeOW bonds, and, with
the increase in WO3 content, WO6 groups are also
formed. Briey, the analysis of the Raman spectra
presented here for K2 OWO3 TeO2 glasses shows
that a breaking of TeOTe network bonds and
the formation of TeO3 groups occur due to addition of K2 O and that TeOW bonds appear in the
structure due to the substitution of WO3 for TeO2 .
3.3. Heat capacity changes at Tg
The heat capacity, Cp , in the glass transition
region for the 10K2 O  10WO3  80TeO2 sample is
shown in Fig. 8, as an example. A jump in Cp is
clearly observed at the transformation from the
glassy state to the supercooled liquid state. The
relaxational overshoot in the heat capacity is also
clearly observed. Similar patterns in Cp vs. temperature curves are observed in other xK2 O 
xWO3  (100 ) 2x)TeO2 samples. From Cp vs.

Fig. 8. Heat Capacity, Cp , in the glass transition region for the


10K2 O  10WO3  80TeO2 sample. Cpg and Cpl are the heat capacities of the glasses and supercooled liquids, respectively. DCp
is equal to dierence between Cpl and Cpg at the glass transition
temperature Tg . Heating rate was 10 K min1 .

temperature curves, we determined the heat capacities of glasses, Cpg , and of supercooled liquids,
Cpl , and the heat capacity changes at the glass
transition, DCp Cpl Cpg . The values of Cpg
(200C), Cpl (350C) and DCp for 10K2 O 
10WO3  80TeO2 sample are 73, 121 and 48 J
mol1 K1 respectively. These are the average
values of three measurements. The values of Cpg
and Cpl for xK2 O  xWO3  (100 ) 2x)TeO2 samples are 7275 and 120130 J mol1 K1 , respectively. The values of DCp and Cpl =Cpg for xK2 O 
xWO3  (100 ) 2x)TeO2 samples are shown in
Fig. 9. It is seen that these values in the samples
with x 5, 10 and 20 are almost the same, but they
increase gradually in samples with x 2030.
It should be pointed out that the value of
Cpl =Cpg 1.651.81 obtained in the xK2 O  xWO3 
(100 ) 2x)TeO2 samples is extremely large compared with Cpl =Cpg 1:1 in SiO2 or GeO2 sample.
The values of Cpl =Cpg in various glass-forming
samples are given in Fig. 10 [18]. It is clear that the
value of Cpl =Cpg in K2 OWO3 TeO2 glasses is
comparable to those in KNO3 Ca(NO3 2 having

Fig. 9. Heat capacity changes, DCp , at the glass transition temperature and Cpl =Cpg ratios for xK2 O  xWO3  (100 ) 2x)TeO2
samples. The lines are drawn as a guide for the eye.

T. Kosuge et al. / Journal of Non-Crystalline Solids 242 (1998) 154164

Fig. 10. Values of Cpl =Cpg ratio for various glass-forming liquids. The data except for the K2 O WO3 TeO2 samples are
taken from Ref. [18].

ionic bonding character and in gycerol, a molecular-type liquid.


4. Discussion
4.1. Thermal stability and structure
In various binary MOTeO2 and ternary R2 O
Nb2 O5 TeO2 systems (R Li, K), glasses thermally stable against crystallization are obtained at
a composition around 80 mol% TeO2 , where MO
is a modier [11,1922]. It has been proposed that
the rate of homogeneous nucleation becomes low
in glasses with around 80 mol% TeO2 [22]. In the
K2 OWO3 TeO2 system, however, the most thermally stable glasses are obtained in the composition of 6070 mol% TeO2 , as demonstrated in the
present study. This means that WO3 contributes
signicantly to the thermal stability against crystallization.
Based on the Raman scattering spectra obtained in the present study, it is concluded that in
the ternary K2 OWO3 TeO2 glasses the network

161

structure is basically composed of TeO4 , TeO3 and


WO4 (and WO6 ) units, and the breaking of
TeOTe bonds occurs due to the addition of K2 O
and further TeOW bonds are formed due to the
substitution of W6 for Te4 . The melting temperatures of TeO2 and WO3 crystals are 452C and
1473C, respectively, implying that WO bonds in
K2 OWO3 TeO2 glasses are much stronger than
TeO bonds [23]. These structural and bonding
features would cause the compositional dependence of Tg in K2 OWO3 TeO2 glasses, i.e. the
substitution of K2 O for TeO2 gives a rapid decrease in Tg , but the substitution of WO3 for TeO2
gives an increase in Tg . Further, the formation of
TeOW bonds might retard the rearrangements
of Te4 or W6 structures necessary for crystallization.
Uchino and Yoko [24] reported that the TeO
(axial) bonds in TeO4 are much weaker than the
TeO (equatorial) bonds. Himei et al. [25] reported
that only a single component was observed in O1s
photoelectron spectra in R2 OTeO2 glasses (R: Li,
Na, Rb) and non-bridging oxygen (NBO) atoms
could not be resolved from bridging oxygen (BO)
atom. They suggest that the electronic density of
the valence shell on an NBO is almost equal to that
of a BO atom. Further, Tatsumisago et al. [26]
reported from the high-temperature Raman spectra for TeO2 -based glasses that TeO4 units change
to TeO3 as the temperature is increased above Tg .
These studies strongly suggested that the bond
strength in the network consisting of TeO4 in
TeO2 -based glasses is substantially weak, almost
irrespective of the TeO4 /TeO3 ratio. But, it is obvious from the present study that the characteristic
of weak network bonds does not mean low thermal stability against crystallization in TeO2 -based
glasses.
4.2. Heat capacity changes in the glass transition
region
As proposed by Angell [18], glass forming liquids having small DCp or small Cpl =Cpg are called
strong liquids, while those showing large DCp or
large Cpl =Cpg are called fragile liquids. It is well
recognized that SiO2 and GeO2 having small DCp
and Cpl =Cpg ; 1:1 are strong glass-forming liquids.

162

T. Kosuge et al. / Journal of Non-Crystalline Solids 242 (1998) 154164

SiO2 and GeO2 have generally tetrahedrally coordinated network structures with strong covalent
bonds that are expected to experience relatively
little disruption during heating. It is obvious that
the xK2 O  xWO3  100 2xTeO2 samples having
the values of DCp 4858 J mol1 K1 and
Cpl =Cpg 1.61.8 are included in the category of
fragile liquids, implying that an inferred rapid
breakdown of their congurational structure occurs with increasing temperature.
Lee et al. [6] reported that two measures of
fragility, DCp and Eg=Tg , were increased with increasing Li2 O content in the Li2 OTeO2 system.
They observed a good correlation between DCp
and Eg=Tg , indicating that the glasses became more
fragile as the Li2 O content increased. On the other
hand, the present authors [8,10,11] found that
DCp at Tg in mixed-alkali tellurite glasses of
20 xLi2 O  xNa2 O  80TeO2 is almost the same
irrespective of the Na2 O=Li2 O ratio, but the degree
of fragility, m, estimated from Eq. (1), depends on
the Na2 O=Li2 O ratio, i.e. the mixed-alkali tellurite
glasses are more `strong' in comparison with single
alkali tellurite glasses:

dlogg
Eg

:
1
m

dT =T
2:303 RT
g

T Tg

Angell [18] has classied alcohols as thermodynamically fragile but kinetically strong liquids,
because they show low fragility (temperature dependence of viscosity) despite large DCp . Recently,
Senapati and Varshneya [27] have reported that
there is a lack of correlation between viscositybased and heat capacity-based classication of
strong and fragile liquids in the overconstrained
region of GeSe and GeSbSe glass-forming
systems. As proposed by Angell [18,28], therefore,
one should consider the strong/fragile concept in
supercooled liquids from kinetic and thermodynamic behaviors. In other words, for example, one
should use the terms `kinetic fragility', estimated
from the temperature dependence of viscosity at
Tg , and `thermodynamic fragility', estimated from
DCp , to avoid any confusion with the use of the
term `fragile'.
We estimated the degree of fragility of
xK2 O  xWO3  (100 ) 2x)TeO2 samples using
Eq. (1), and the obtained values are given in Ta-

ble 2. The value of m are 4750, being almost the


same irrespective of TeO2 content. That is, the
kinetic fragility of, for example, 10K2 O  10WO3 
80TeO2 is almost the same as that in 25K2 O 
25WO3  50TeO2 . We can say, therefore, that the
thermodynamic fragility (i.e DCp ; Cpl =Cpg in Fig. 9)
in these samples changes with TeO2 content, but
the kinetic fragility (i.e. m in Table 2) does not
appear to change. We consider these phenomena
using the structural relaxation model (the congurational entropy model) proposed by Adam and
Gibbs [29], which has been successfully applied to
the structural relaxation behavior of various
glasses [18,27,2933].
According to Adam and Gibbs [29], the structural relaxation time, s, involving the congurational entropy Sc is expressed by Eq. (2),


DlSc
;
2
s s0 exp
kTSc
where s0 is constant, Dl is the potential barrier
against rearrangement and k is Boltzmann's constant Sc is the congurational entropy of the
smallest cooperatively group of molecules (cluster)
that can undergo a rearrangement; generally
Sc ; k ln2 is assumed [30]. In their model, the heat
capacity dierence between supercooled liquids
and glasses, DCp , is considered to be the congurational heat capacity, Cpconf , i.e. DCp Cpconf
Cpl Cpg . That is, DCp is the thermodynamic
measure of temperature-induced structural changes such as the change in distribution of diering
chemical elements onto analogous sites of the
structure and the change in the topology of the
structure. Richet et al. [3133] have analyzed the
heat capacity changes in various silicate glasses by
assuming that DCp is considered to be the congurational heat capacity, because the congurational entropy in silicate liquids is usually twice or
three times as great as the entropy frozen-in at the
glass transition.
The congurational entropy is given by Eq. (3):
ZT
Sc T
T2

DCp
dT ;
T

where DCp is equal to Cpl Cpg , and T2 is the


temperature at which Sc 0. The gradual increase

T. Kosuge et al. / Journal of Non-Crystalline Solids 242 (1998) 154164

in DCp (Fig. 9) in the xK2 O  xWO3  (100 )


2x)TeO2 samples suggests that the congurational
entropy increases gradually with decreasing TeO2
content, although the value of Sc has not been
determined in this study. This would be supported
by the structural information obtained from Raman spectra, i.e. the breaking of TeOTe network
bonds occurs due to the addition of K2 O, because
the increase in the congurational entropy means
that cooperative rearrangements of the structure
can take place independently in smaller and
smaller regions of the liquid [29,33].
Since the viscosity is proportional to the structural relaxation time, i.e. g sG1 , where G1 is the
high frequency shear modulus of the liquid, one
can obtain the viscosity equation from Eq. (2)
[30,32],
log g A

DlSc
;
kTSc

where A is a pre-exponential term. Applying the


Adam and Gibbs model to the concept of fragility,
the degree of fragility, m, is expressed by Eq. (5),
because the activation energy for viscous ow
corresponds to DlSc =Sc ;


1
DlSc
:
5
m
Sc
2:303 RTg
In the structural relaxation model proposed by
Adam and Gibbs, therefore, the kinetic fragility,
m, is determined from the value of DlSc =Sc =Tg ,
meaning that the degree of fragility is a function of
Dl; Sc and Tg (here we assume Sc ; K ln2). As can
be seen in Table 2, the value of Eg in the
xK2 O  xWO3  (100 ) 2x)TeO2 samples decreases with decreasing TeO2 content. This means that
the value of (DlSc =Sc ) decreases with decreasing
TeO2 content. But, since the glass transition temperature decreases steeply with decreasing TeO2
content, the value of Eg=Tg is almost constant irrespective of TeO2 content as given in Table 2,
meaning that the value of DlSc =Sc Tg is also almost constant. The product of DCp and Tg increases slightly with decreasing TeO2 content (see
in Table 2). It is, therefore, considered that the
value of Dl would increase slightly, but would not
change much with decreasing TeO2 content in the
xK2 O  xWO3  (100 ) 2x)TeO2 samples.

163

The present results indicate that the kinetic


fragility estimated from the activation energy for
viscous ow at and near the glass transition temperature must be carefully analyzed and the
appearance of the dierence between the thermodynamic fragility estimated from DCp and the kinetic fragility estimated from Eg=Tg is realistic
even for oxide glass forming liquids. It is known
that there are some liquids (such as glycerol) with
large DCp (i.e. thermodynamically fragile), but
relative strong viscosity behavior (i.e. kinetically
strong) [18,27,29]. In the previous papers [8,10,11],
it was found that the mixed-alkali tellurite glasses
of (20 ) x)Li2 O  xNa2 O  80TeO2 are more
stronger kinetically in comparison with single alkali telurite glasses, although the values of DCp are
almost constant irrespective of the Na2 O=Li2 O
ratio. It is desired to examine the relationship between DCp and Eg=Tg in other various TeO2 -based
glasses, giving a clearer understanding of the unique features of structure, bonding, thermal stability against crystallization and structural
relaxation behaviors in TeO2 -based glasses.
5. Conclusion
As described in the introduction, information
on thermal stability against crystallization and
structural relaxation behavior in TeO2 -based
glasses is lacking compared to the study of structure and optical properties. The present study
clearly demonstrated that TeO2 -based glasses show
unique behaviors even in thermal stability and
structural relaxation. For example, some glasses
such as 15K2 O  15WO3  70TeO2 show large heat
capacity changes, DCp , in the glass transition region and at the same time show very high thermal
stability against crystallization. The compositional
dependence of the thermodynamic fragility estimated from DCp and the kinetic fragility estimated
from the activation energy for viscous ow at Tg in
xK2 O  xWO3  (100 ) 2x)TeO2 samples is largely
dierent. These behaviors have been analyzed
using the structural relaxation model proposed by
Adam and Gibbs. In order to get more detailed
information on the structural relaxation in TeO2 based glasses, the evaluation of congurational

164

T. Kosuge et al. / Journal of Non-Crystalline Solids 242 (1998) 154164

entropy and the kinetics of heat capacity changes


are strongly desired.
Acknowledgements
This work was supported from a Grant of the
Asahi Glass Foundation and a Grant-in-Aid for
Scientic Research of the Ministry of Education,
Sports and Culture in Japan.
References
[1] J.S. Wang, E.M. Vogel, E. Snitzer, Opt. Mater. 3 (1994)
187.
[2] S.H. Kim, T. Yoko, J. Am. Ceram. Soc. 78 (1995) 1061.
[3] K. Shioya, T. Komatsu, H.G. Kim, R. Sato, K. Matusita,
J. Non-Cryst. Solids 189 (1995) 16.
[4] H.G. Kim, T. Komatsu, K. Shioya, K. Matusita, K.
Tanaka, K. Hirao, J. Non-Cryst. Solids 208 (1996) 303.
[5] M. Imaoka, T. Yamazaki, J. Ceram. Soc. Jpn. 76 (1968)
160.
[6] S.K. Lee, M. Tatsumisago, T. Minami, Phys. Chem.
Glasses 35 (1994) 226.
[7] C.Y. Zahra, A.M. Zahra, J. Non-Cryst. Solids 190 (1995)
251.
[8] T. Komatsu, T. Noguchi, R. Sato, J. Am. Ceram. Soc. 80
(1997) 1327.
[9] K. Tanaka, K. Hirao, K. Kashima, Y. Benino, N. Soga,
Phys. Chem. Glasses 38 (1997) 87.
[10] T. Komatsu, T. Noguchi, Y. Benino, J. Non-Cryst. Solids
222 (1997) 206.
[11] T. Komatsu, R. Ike, R. Sato, K. Matusita, Phys. Chem.
Glasses 36 (1995) 216.
[12] T. Sekiya, N. Mochida, A. Ohtsuka, M. Tonokawa, J.
Non-Cryst. Solids 144 (1992) 128.

[13] T. Sekiya, N. Mochida, S. Ogawa, J. Non-Cryst. Solids 176


(1994) 105.
[14] I. Shaltout, Y. Tang, R. Braunstein, A.M. Abu-Elazm, J.
Phys. Chem. Solids 56 (1995) 141.
[15] V. Dimitrov, M. Arnaudov, Y. Dimitriev, Monatsh. Chem.
115 (1984) 987.
[16] F. Knee, R.A. Condrate, J. Phys. Chem. Solids 40 (1979)
1145.
[17] V. Kozhukharov, S. Neov, I. Gerasimova, P. Mikula, J.
Mater. Sci. 21 (1986) 1707.
[18] C.A. Angell, J. Non-Cryst. Solids 131133 (1991) 13.
[19] T. Komatsu, H. Tawarayama, H. Mohri, K. Matusita, J.
Non-Cryst. Solids 135 (1991) 105.
[20] M. Zhang, S. Mancini, W. Bresser, P. Boolchand, J. NonCryst. Solids 151 (1992) 149.
[21] J. Heo, D. Lam, G.H. Sigel, E.A. Endoza, D.A. Hensley, J.
Am. Ceram. Soc. 75 (1992) 277.
[22] T. Komatsu, K. Shioya, H.G. Kim, Phys. Chem. Glasses
38 (1997) 188.
[23] L.G. Van Uitert, H.M. O'Bryan, M.E. Lines, H.J. Guggenheim, G. Zydzig, Mater. Res. Bull. 12 (1977) 261.
[24] T. Uchino, T. Yoko, J. Non-Cryst. Solids 204 (1996) 243.
[25] Y. Himei, Y. Miura, T. Nanba, A. Osaka, J. Non-Cryst.
Solids 211 (1997) 64.
[26] M. Tatsumisago, T. Minami, Y. Kowada, H. Adachi,
Phys. Chem. Glasses 35 (1994) 89.
[27] U. Senapati, A.K. Varshneya, J. Non-Cryst. Solids 197
(1996) 210.
[28] R. Richet, A. Blumen, Disorder Eects on Relaxational
Processes, Springer, Berlin, 1994, p. 22.
[29] G. Adam, J.H. Gibbs, J. Chem. Phys. 43 (1965) 139.
[30] G.W. Sherer, J. Non-Cryst. Solids 123 (1990) 75.
[31] P. Richet, D.R. Neuville, Adv. Phys. Geochem. 10 (1992)
132.
[32] P. Richet, M.A. Bouhifd, P. Courtial, C. Tequi, J. NonCryst. Solids 211 (1997) 271.
[33] A. Sipp, D.R. Neuville, P. Richet, J. Non-Cryst. Solids 211
(1997) 281.

Вам также может понравиться