Study on the stability of GaF3-based glasses by differential scanning calorimetry

Article abstract of DOI:10.1023/A:1013160816648

The glass formation and devitrification of GaF₃-based glasses were studied by differential scanning calorimetry. A comparison of various simple quantitative methods to assess the level of stability of multi-component fluoride glass systems is presented. Most of these methods are based on critical temperatures. In this paper a new parameter k_b(T) is added to the stability criteria. The stability of several GaF₃-based glasses were experimentally evaluated and correlated with the activation energies of crystallization via this new kinetic criterion and compared with those evaluated by other criteria.

Full text of DOI:10.1023/A:1013160816648

Journal of Thermal Analysis and Calorimetry, Vol. 66 (2001) 479–487
STUDY ON THE STABILITY OF GaF3-BASED
GLASSES BY DIFFERENTIAL SCANNING
CALORIMETRY
J. Zhu1, H. Jin1, D. Dong2, D. Qiang1 and F. Ma1
1College of Materials Science and Engineering, Tongji University, Shanghai 200092, P.R. China
2Institute of Technology, Tongji University, Shanghai 200092, P.R. China
(Received July 4, 2000; in revised form June 8, 2001)
Abstract
The glass formation and devitrification of GaF3-based glasses were studied by differential scanning cal-
orimetry. A comparison of various simple quantitative methods to assess the level of stability of multi-
component fluoride glass systems is presented. Most of these methods are based on critical tempera-
tures. In this paper a new parameter kb(T) is added to the stability criteria. The stability of several
GaF3-based glasses were experimentally evaluated and correlated with the activation energies of crys-
tallization via this new kinetic criterion and compared with those evaluated by other criteria.
Keywords: DSC, GaF3-based glasses, glass stability
Introduction
Fluoride-based glasses are being widely investigated because of their potential use as
high-efficiency optical devices, e.g. fiber amplifiers, fast scintillators and up-conver-
sion lasers [1–3]. Recently, it was found that the fluorogallate glasses have a potential
application for infrared transmitting and active fibers because their phonon energy is
lower than that of fluoroindates and their IR transmission range is larger than that of
fluorozirconates [4]. However, improvements in fluorogallate glass stability and pro-
cessing may be expected. For fluorogallate glass, potential applications related to op-
tical fibers require that the devitrification problems during fiber drawing should be
under control.
In this paper we study the glass stability of GaF3-based system containing Al, Ba,
Ca, Gd, Mg, Th and Zr fluoride by different simple quantitative methods [5–13], which
are based on critical temperature [5–9], on the glass crystallization constant k [10, 11],
and on the crystallization activation energy [12~13], respectively. These thermal parame-
ters are easily and precisely obtained by means of DTA or DSC during the heating pro-
cesses of glass samples. Dietzel [5] introduced the first glass criterion, thermal stability
∆T=Tc–Tg. By using the critical temperatures, Hruby [8] developed the Hr criterion,
1418–2874/2001/ $ 5.00
Akadémiai Kiadó, Budapest
© 2001 Akadémiai Kiadó, Budapest
Kluwer Academic Publishers, Dordrecht

480
ZHU et al.: STABILITY OF GaF3-BASED GLASSES
Hr=(Tc–Tg)/(Tm–Tc). On the basis of Hr criterion, Saad and Poulain [9] obtained other
′
′
two criteria, the weighed thermal stability H and S criteria, H =(Tc–Tg)/Tg and
S=(Tp–Tc)(Tc–Tg)/Tg. Having compared the existing criteria based on critical tempera-
tures, k(T) criteria, and Branda’s E/RT criterion, we consider synthetically both the ki-
netic parameters and characteristic temperatures and define a kb(T) stability criteria.
Experimental
Glass preparation
The starting materials were AlF3, BaF2, CaF2, CdF2, GdF3, MgF2, SrF2, ThF4, ZnF2,
ZrF4 and Ga2O3. Gallium oxide was first fluorinated at 400°C with NH4FHF in a plati-
num crucible. A batch of 10 g was mixed with excess ammonium bifluoride
(NH4FHF) as a precaution to convert trace oxide impurities to fluoride. Glasses were
melted for 4–5 h in a platinum crucible with a resistance furnace, after complete melt-
ing, the glass melts were cast by pouring into open brass moulds. Glass samples were
annealed at 300°C for 15 min. GaF3-based glasses had the following mole composi-
tion formulae: 30GaF3–25SrF2–25CdF2–15ZnF2–MFn (GSCZM), where MFn is AlF3,
BaF2, CaF2, GdF3, MgF2, ThF4 and ZrF4.
Differential scanning calorimetry (DSC)
DSC scans were carried out by a CDR-1A Thermal Analyzer (SBIF, Shanghai) with an
accuracy of ±0.5°C. For non-isothermal experiments, glass samples were heated from
room temperature to 700°C at heating rates varying from 5 to 20°C min–1. Bulk glass
pieces (5–15 mg), which were water clear, instead of powder were used to avoid prob-
lems with particle size distribution, and Al2O3 powder was used as reference material.
Theoretical analysis
The relationship between crystallization fraction and the nucleation and growth rate
is given by the Johnson–Mehl–Avrami equation:
x =1–exp[–(kt)n]
(1)
where x is the volume fraction crystallized after time, t, n is the Avrami exponent, and k is
the reaction-rate constant, encompassing nucleation and crystal growth factors, whose
temperature dependence is generally expressed by the Arrhenian-type equation:
k =νexp(–E/RT)
(2)
where ν is the frequency factor, E is apparent activation energy, R is the ideal gas
constant and T is the isothermal temperature in Kelvin.
In a non-isothermal DSC experiment, the temperature is changed linearly with
time at a known scan rate β (=dT/dt):
T=T0+βt
(3)
J. Therm. Anal. Cal., 66, 2001

ZHU et al.: STABILITY OF GaF3-BASED GLASSES
481
where T0 is the starting temperature, and T is the temperature after time t, as the tem-
perature constantly changes with time, k is no longer a constant but varies with time
in a more complicated form, and Eq. (1) becomes:
x =1–exp[–(k(T–T0)/β)n]
(4)
n
t
or
x =1–exp − k(t)dt
(5)
∫
0
If the rate of transformation is maximum at the peak of DSC curve, then (T=Tp):
d2x/dt2= 0
(6)
After rearranging Eq. (5), Bansal et al. [14] developed a method for a non-iso-
thermal analysis of devitrification. The final relation is as follows:
ln(T 2/β) = ln(E/R) – lnν+E/RTp
(7)
p
Therefore, a plot of the logarithm of the peak temperature squared divided by the
heating rate, β, should be proportional to the reciprocal of Tp, with slope, E/R, and inter-
cept, ln(E/R)–lnν. Then one can obtain E, ν and k(T). Surinach et al. [10] introduced a
k(Tg) criterion, and Hu and Jiang [11] developed the k(Tp) criterion respectively:
k(Tg) = νexp(–E/RTg)
k(Tp) = νexp(–E/RTp)
(8)
(9)
Thus the values of these two parameters indicate the tendency of glass to devitrify
on heating. The larger their values are, the greater the tendency is to devitrify. However,
the formation of glass is a thermodynamic process and a kinetic process as well. A new
criterion including both the kinetic parameters and characteristic temperature is being
searched, and hoped to evaluate the glass stability over a broad temperature range. So,
here, the Hr stable factor [8] is introduced to the k(T), and define kb(T) criterion:
kb(T) = νexp(–HrE/RT)
(10)
where T is the temperature Tg, Tc or Tp. Just like the k(T) criteria, the smaller the val-
ues of kb(T), are, the greater the thermal stability of the glass is. The obvious advan-
tage of this method is that it can evaluate the glass stability over a broad temperature
range, because it takes into account the inflexion point on the kink of the basic line,
the first symptoms of crystallization occur and the appearance of the liquid phase [8].
Results and discussion
After preparing GaF3-based glass samples, at first, the glass-forming ability, or the
glass stability was estimated by a metallographic microscope. The observed results
are listed in Table 1.
J. Therm. Anal. Cal., 66, 2001

482
ZHU et al.: STABILITY OF GaF3-BASED GLASSES
Table 1 Glass-forming ability of the GaF3-based systems
Sample
Glass-forming ability
GSCZAl
GSCZBa
GSCZCa
GSCZGd
GSCZMg
GSCZTh
GSCZZr
the devitrification occurred in glasses
one or two little isolated crystals in glasses
several little isolated crystals in glasses
the devitrification occurred in glasses
few devitrification for glasses
water clear
water clear
From the DSC scans of GaF3-based glasses, the characteristic temperatures such
as the glass transition temperature, Tg, the onset crystallization temperature, Tc, the
exotherm maximum temperature, Tp, and the melting temperature, Tm are given in Ta-
ble 2. The stability of these seven GaF3-based glass samples can be estimated by us-
ing these characteristic temperatures, Tg, Tc, Tp, Tm. The existing stability criterion pa-
′
rameters (∆T, Hr, H and S) based on these characteristic temperatures are shown in
Figs 1a and 1b.
Fig. 1 The stable orders of the GaF -based glasses by (a) ∆T and Hr, (b) H and S crite-
′
3
ria ( – 5°C min , – 10°C min , ∆ – 15°C min , – 20°C min–1)
–1
–1
–1
n
¡
J. Therm. Anal. Cal., 66, 2001

ZHU et al.: STABILITY OF GaF3-BASED GLASSES
483
Table 2 Characteristic temperatures of the GaF3-based glasses
Tg/°C
Tc/°C
Tp/°C
Tm/°C
Sample
β=5°C min–1
β=10°C min–1
β=15°C min–1
β=20°C min–1
GSCZAl
GSCZBa
GSCZCa
GSCZGd
GSCZMg
GSCZTh
GSCZZr
318.5
318.0
311.0
329.5
315.0
321.5
333.5
382.5
379.0
369.0
393.5
383.5
404.5
407.5
388.0
381.5
374.0
397.5
387.0
410.0
412.0
612.5
613.0
614.5
632.5
645.0
594.0
599.0
GSCZAl
GSCZBa
GSCZCa
GSCZGd
GSCZMg
GSCZTh
GSCZZr
322.0
322.5
314.5
332.0
323.5
328.0
339.5
389.0
384.0
371.5
401.5
389.0
413.0
414.0
396.0
388.0
376.5
405.5
394.5
419.5
420.0
613.5
614.0
615.0
630.0
644.0
594.5
602.0
GSCZAl
GSCZBa
GSCZCa
GSCZGd
GSCZMg
GSCZTh
GSCZZr
325.5
325.5
320.5
335.5
324.0
329.0
337.0
391.0
388.0
374.0
402.0
392.0
418.0
418.0
399.0
392.5
379.0
409.0
399.0
427.5
425.0
614.5
615.5
616.0
627.5
645.5
593.5
602.0
GSCZAl
GSCZBa
GSCZCa
GSCZGd
GSCZMg
GSCZTh
GSCZZr
326.0
332.0
322.0
336.0
326.5
330.5
342.5
393.5
388.0
375.5
404.0
397.0
420.5
420.5
403.0
394.0
382.5
412.5
404.0
431.0
430.0
615.0
616.0
618.0
627.0
647.0
590.0
602.0
It is known that these criteria of glass stability allow a prediction of the tendency
to devitrification. It is suggested that the higher values are, the higher the glass ther-
′
mal stability should be. According to these suggestions, parameters, ∆T, Hr, H and S
J. Therm. Anal. Cal., 66, 2001

484
ZHU et al.: STABILITY OF GaF3-BASED GLASSES
in Figs 1a and 1b show that the glass sample GSCZTh is the most stable in the seven
samples. It is also found that GSCZZr is more stable than the other five glass samples
′
and GSCZCa is the most unstable in the seven samples based on ∆T, Hr and H at dif-
ferent heating rates. The stable order of these samples (except GSCZTh) changed
with heating rates. According to Branda’s criterion, E/RTp, higher values of E/RTp
should be indicative of greater tendency to devitrify. GSCZTh is the most stable, and
the stability orders at different heating rates are GSCZTh>GSCZZr>GSCZMg>
GSCZAl>GSCZGd>GSCZBa>GSCZCa (Fig. 3b).
In order to compare the tendencies to devitrification in this fluoride glass sys-
tem, it is useful to point out that the devitrification is a kinetic process, thus it is rea-
sonable to assess the glass stability via the kinetic parameters. To obtain the kinetic
parameters of crystallization, Eq. (7) is applied. Figure 2 presents the evolution of
ln(Tp2/β) vs. 1/Tp. The plots were found to be straight lines in accordance with Eq. (7).
The slope E/R, and frequency factor ν are then evaluated by a least-squares fitting
method. When the values of E/R and ν are known, the kinetic parameters can be cal-
culated. The kinetic parameters k(T) and kb(T) of GaF3-based glasses are shown in
Figs 3a–c.
Surinach et al. [10] and Hu and Jiang [11] suggested that the lower the value of
k(Tg) or k(Tp) is, the higher the thermal stability of the glass should be. Accordingly,
data on k(Tp) in Fig. 3a indicate that the glass sample GSCZTh is the most stable and
Fig. 2 Determination of the activation energy of crystallization by plotting ln(Tp2/β) vs.
1/Tp for the GaF3-based glasses: a – for GSCZCa, GSCZGd, GSCZMg, GSCZZr
and b – for GSCZAl, GSCZBa, GSCZTh
J. Therm. Anal. Cal., 66, 2001

ZHU et al.: STABILITY OF GaF3-BASED GLASSES
485
Fig. 3 The stable orders of the GaF3-based glasses by a – k(Tg) and k(Tp), b – E/RTp and
¡
–1
–1
n
kb(Tg), and c – kb(Tc) and kb(Tp) criteria ( – 5°C min , – 10°C min ,
∆ – 15°C min–1, – 20°C min–1)
GSCZCa is the least stable of these glass samples at different heating rates. However,
it is difficult to obtain the same stability sequence of GSCZAl and GSCZGd for GaF3-
based glasses by this criterion.
Surinach’s criterion k(Tg) is contrary to Hu’s criterion k(Tp) for this fluoride
glass system shown in Fig. 3a and also does not accord with the actual melting pro-
cess of the glass system. No consistent order of stability at different heating rates can
be obtained (Fig. 3a). Therefore, this criterion does not fit for this glass system.
Data on kb(Tg), kb(Tc) and kb(Tp) in Figs 3b and 3c also show that the glass sample
GSCZTh is the most stable, and the stability sequence is GSCZTh>GSCZZr>
GSCZMg>GSCZAl>GSCZGd>GSCZBa>GSCZCa at various heating rates. This
J. Therm. Anal. Cal., 66, 2001

486
ZHU et al.: STABILITY OF GaF3-BASED GLASSES
stability result agrees with that of the E/RTp criterion (Fig. 3b) and the actual melting
processes of GaF3-based glasses.
According to Portier [15], ThF4, ZrF4 and AlF3 as good glass formers are charac-
terized by a high enthalpy of formation and a small ionic radius. These two features
correspond to a strong metal-fluorine bond with certain covalence favorable to a
polymerization in the liquid state. MgF2 is a good structural intermediate [16]. It is
able to participate in the continuous glass network and can increase the glass stability
and chemical durability. BaF2, CaF2 and GdF3 are glass modifiers in terms of the ratio
of the standard heat of formation over the cationic radius [15]. According to Cottrant
et al. [17], glass modifiers have two contradictory effects on glass formation in
multicomponent system: 1. some of them tend to induce crystallization, while 2.
some of them enhance the covalence of the glass former, increasing the polymeriza-
tion of the species in the liquid phase and favouring glass formation. As shown above,
CaF2, BaF2 and GdF3 have a higher tendency to crystallization. However, in the
InF3–ZnF2–BaF2–SrF2 basic system, the addition of small quantities of GdF3 (less
than 4%) improves the stability [18].
Conclusions
A new criterion kb(T) has been defined for the evaluation of GaF3-based glass stability
from DSC data. It includes both the kinetic parameters and characteristic tempera-
tures. The criterion is more accurate and reasonable than the existing criteria based on
′
critical temperatures, ∆T, Hr, H , S, and k(T) criteria. A high value of kb(T) means
poor stability of the glass. Study of the non-isothermal devitrication of seven
GaF3-based glasses at different heating rates and various temperatures has verified
that the criterion kb(T) is neither affected by the heating rate, nor by the temperature,
′
while the results of ∆T, Hr, H , S, or k(T) criterion vary with the heating rates and
temperatures. Among seven glass samples, kb(T) of GSCZTh glass sample is the
smallest, so this glass composition is the most stable. The stability order of these
seven fluoride glasses is GSCZTh>GSCZZr>GSCZMg>GSCZAl>GSCZGd>
GSCZBa>GSCZCa. This result fits in well with the actual melting processes of the
glasses, and can also be explained by the correlation, which is the heat of formation of
the compounds forming the glass and the ionic strength of the cations in the glass.
* * *
The authors wish to thank Associate Prof. Zhonglin Bo for helpful discussions.
References
1 M. Poulain, J. Non-Cryst. Solids, 184 (1995) 103.
2 W. Seeber, E. A. Downing, L. Hesselink, M. M. Fejer and D. Ehrt, J. Non-Cryst. Solids, 189
(1995) 218.
3 Zhang Guoyin and M. Poulain, J. Chinese Ceram. Soc., 26 (1998) 103.
4 A. Soufiane and M. Poulain, J. Non-Cryst. Solids, 161 (1993) 206.
J. Therm. Anal. Cal., 66, 2001

ZHU et al.: STABILITY OF GaF3-BASED GLASSES
487
5 A. Dietzel, Glasstech. Ber., 22 (1968) 41.
6 D. R. Uhlmann, J. Non-Cryst. Solids, 7 (1972) 337.
7 D. R. Uhlmann, J. Non-Cryst. Solids, 25 (1977) 43.
8 A. Hruby, Czech. J. Phys., B22 (1972) 1187.
9 M. Saad and M. Poulain, Mater. Sci. Forum., 19-20 (1987) 11.
10 S. Surinach, M. D. Barof, M. T. Clavaguera-Mora and N. Clavaguera, J. Mater. Sci.,
19 (1984) 3005.
11 L. Hu and Z. Jiang, J. Chinese Ceram. Sci., 18 (1990) 315.
12 X. Zhao and S. Sakka, J. Non-Cryst. Solids, 95-96 (1987) 487.
13 F. Branda, A. Marotta and A. Buri, J. Non-Cryst. Solids, 134 (1987) 123.
14 N. P. Bansal, R. H. Doremus, A. J. Bruce and C. T. Moynihan, J. Am. Ceram. Soc.,
66 (1983) 233.
15 J. Portier, J. Non-Cryst. Solids, 112 (1989) 15.
16 C. M. Baldwin and J. D. Mackenzie, J. Am. Ceram. Soc., 62 (1979) 537.
17 J. F. Cottrant, B. Dubois and J. Portier, Mat. Res. Bull., 20 (1985) 203.
18 Y. Messaddeq, A. Delben, M. Boscolo, M. A. Aegerter, A. Soufiane and M. Poulain,
J. Non-Cryst. Solids, 161 (1993) 210.
J. Therm. Anal. Cal., 66, 2001