Crystallization kinetics in new Sb14As29Se52Te5 amorphous glass

Article abstract of DOI:10.1016/j.ssc.2006.02.033

Differential scanning calorimetry (DSC) under non-isothermal conditions is used to study the crystallization kinetics of Sb₁₄As₂₉Se₅₂Te₅ chalcogenide glass. In addition, two approaches are used to analyze the dependence of glass transition temperature (T_g) on the heating rate (α). One is empirical linear relationship between (T_g) and ln(α). The second approach is the use of straight line ln ( T_g² / α ) vs. 1/T_g for the evaluation of the activation energy for glass transition. The phases at which the alloy crystallizes after the thermal treatment have been identified by using X-ray diffraction. The diffractogram of the transformed material shows the presence of some crystallites of As, Te, AsSb, As₂Se₃, Sb₂Se₃ and AsSe_.5Te_.5 in the residual amorphous matrix.

Full text of DOI:10.1016/j.ssc.2006.02.033

Solid State Communications 138 (2006) 184–189
www.elsevier.com/locate/ssc
Crystallization kinetics in new Sb As Se Te amorphous glass
29 52
14
5
a
A.A. Othman , K.A. Aly
b,
b
*
, A.M. Abousehly
a
Physics Department, Assiut University, Assiut, Egypt
b
Physics Department, Faculty of Science, Al-Azhar University, Al-Azhar University Street, Assiut 71524, Egypt
Received 16 August 2005; received in revised form 6 February 2006; accepted 24 February 2006 by J.F. Sadoc
Available online 20 March 2006
Abstract
Differential scanning calorimetry (DSC) under non-isothermal conditions is used to study the crystallization kinetics of Sb14As29Se52Te
5
chalcogenide glass. In addition, two approaches are used to analyze the dependence of glass transition temperature (T ) on the heating rate (a).
g
2
One is empirical linear relationship between (T
g
) and ln(a). The second approach is the use of straight line lnðT =aÞ vs. 1/T
g
for the evaluation of
g
the activation energy for glass transition. The phases at which the alloy crystallizes after the thermal treatment have been identified by using X-ray
diffraction. The diffractogram of the transformed material shows the presence of some crystallites of As, Te, AsSb, As
in the residual amorphous matrix.
2 3 2 3
Se , Sb Se and AsSe.5Te.5
q 2006 Elsevier Ltd. All rights reserved.
PACS: 61.10.Nz; 61.43.Fs; 61.43.-J; 65.60.Ca
Keywords: A. Amorphous solid; A. Glasses; C. X-ray diffraction; D. Thermal properties
1
. Introduction
The crystallization kinetics from the exothermic peaks of
the differential scanning calorimetry (DSC) curves under non-
isothermal conditions were studied by various methods [7–10].
In this work, the glass transition and the crystallization
kinetics were studied using DSC with continuous heating of the
sample at various uniform heating rates. The characteristic
Thermal activated transformations in the solid state can be
investigated by isothermal or non-isothermal experiments.
While isothermal experimental analysis techniques are, in
most cases, more definitive but non-isothermal techniques
have several advantages. The non-isothermal techniques have
become particularly prevalent for the determination of the
thermal stability of amorphous alloys in the investigation of
the processes of the nucleation and growth that occur during
the transformation of the metastable phases in the glassy
alloy as it is heated. These techniques provide rapid
information on such parameters as the glass transition
temperature and the activation energy over a wide range of
temperature [1,2].
temperatures T , T , and T as well as the activation energy of
g
c
p
the glass transition and the activation energy for crystallization,
Eg and E , respectively, determined for Sb As Se Te
5
c
14 29 52
glassy alloy. Finally, the crystalline phases are corresponding
to the crystallization process were identified by X-ray
diffraction (XRD) measurements.
2. Experimental
The semi-conducting Sb14.25As28.5Se52.25Te5 glassy was
The activation energy in the glass crystallization phenom-
ena is associated with nucleation and growth processes. The
studies of glass crystallization upon heating can be performed
by several different ways and the crystallization process can be
interpreted in terms of several theoretical models [3–6].
made from their components of 99.999% high purity. The
proper amount for each material was weighed, and then the
weighed materials were introduced into cleaned silica tubes.
To avoid the oxidation of the sample the tube was evacuated to
K5
1
1
0
250 K for 24 h, hand shaking of the constituent materials
Torr. The ampoule was put into a furnace at around
inside the ampoule in furnace was necessary for realizing the
homogeneity of the composition, then the ampoule was
quenched in to ice-cold water to avoid crystallization.
*
Corresponding author. Tel./fax: C20 88 2325647.
E-mail address: kamalaly2001@yahoo.com (K.A. Aly).
The amorphous state of the material was confirmed by a
diffractometric X-ray scan (Philips diffractometer 1710) using
0038-1098/$ - see front matter q 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ssc.2006.02.033

A.A. Othman et al. / Solid State Communications 138 (2006) 184–189
185
˚
Cu as target and Ni as filter (lZ1.542 A), showing an absence
of the peaks, which are characteristics of crystalline phases.
Energy dispersive X-ray spectroscopy (link analytical EDS)
was used to measure the elemental composition and indicate
that, the investigating composition is correct up to G0.2 at.%.
The calorimetric measurements were carried out by using
differential scanning calorimeter (Shimadzu 50) with an
accuracy of 0.1 K, keeping a constant flow of nitrogen to
extract the gases generated during the crystallization reactions,
which is the characteristic of chalcogenide materials. The
calorimeter was calibrated, for each heating rate, using the
well-known melting temperatures and melting enthalpies of
zinc and indium supplied with the instrument [11]. Twenty
milligrams powdered samples, introduced into aluminum pans
and scanned at continuous heating rates (aZ2.5, 5, 10, 15, 20,
phenomena are resolved during at the studied temperature
region. The first one (T Z450.53 K) corresponds to the glass
g
transition temperature, the second one (T Z525 K) corre-
c
sponds to the onset temperature of crystallization and the last
one (T Z545.12 K) corresponds to the peak temperature of
p
crystallization of the Sb As Se Te chalcogenide glass. This
1
4
29 52
5
behavior is a typical curve for glass-crystalline transformation.
Fig. 1(b) shows the DSC thermograms of the Sb As Se
29 52-
1
4
Te chalcogenide glass at different heating rates (aZ2.5, 5, 10,
5
15, 20, 25 and 30 K/min). Generally, the three characteristic
temperatures are increases as heating rate increases.
3.1. Glass transition
Two approaches are used to analyze the dependence of glass
transition temperature on the heating rate a. The first is the
empirical relationship
2
crystallization extrapolated onset, T and the crystallization
5 and 30 K/min). The value of the glass transition, T , the
g
c
peak, T , temperature were determined with accuracy G1 K by
p
using the microprocessor of the thermal analyzer.
The fraction crystallized c at any given temperature T is given
Tg Z A CB lnðaÞ (1)
where A and B are constants for any given glass composition
by cZA /A where A is the total area of the exotherm between the
T
[12]. The results shown in Fig. 2 indicate the validity of Eq. (1)
for the Sb As Se Te glassy alloy, the empirical relation-
ship for this glass can be written in the form
temperature, T , where crystallization is just beginning and
i
1
4
29 52
5
the temperature, T , where the crystallization is completed, A is
f
T
the area between T and T, as shown in Fig. 1(a).
i
Tg Z 436:9 C5:42 lnðaÞ (2)
where a least squares line representing T vs. ln(a) has been
3
. Results and discussion
g
fitted to the experimental data.
The other approach concerns to the use of Kissinger’s linear
The DSC obtained under non-isothermal conditions of
Sb As Se Te chalcogenide glass at heating rate of
1
4
29 52
5
dependence in the form [13,14]
!
1
5 K/min, is shown in Fig. 1(a). Three characteristics
a
Eg
ln
ZK
Cconst:
(3)
2
Tg
RTg
2
a straight line between lnða=T Þ and 1/T , whose slope yields a
g
g
value of E /R where the subscript g denotes magnitude values
g
corresponding to the glass transition temperature and R is the
gas constant.
2
g
The variation in lnð1=T Þwith ln(a) is much slower than
ln(1/T ) with ln a [15–18]. Therefore, the Kissinger equation
g
Fig. 1. DSC thermograms for powderd Sb14As29Se52Te
heating rates, (b) DSC traces at aZ15 K/min for the powdered Sb14As29Se52-
Te glass, The hatched area shows A , The area between T and T.
5
glass at different
g 5
Fig. 2. Glass transition temperature T vs. ln a (a in K/S) of Sb14As29Se52Te
glassy alloy.
5
T
i

186
A.A. Othman et al. / Solid State Communications 138 (2006) 184–189
can be simplified to:
Eg
lnðaÞ ZK
Cconst:
(4)
RTg
2
g
Fig. 3 shows plots of lnða=T Þ, curve (a) and ln(a)curve (b)
vs. 1/T for the Sb As Se Te semiconductor glass,
29 52
g
14
5
displaying the linearity of the equations used. The obtained
values of the glass transition activation energy are 291.62 kJ/
mol (a) and 296.8 kJ/mol (b), respectively. From the
experimental values of T one can use both relations (3) and
g
(
4) to evaluate the glass transition activation energy E for
g
Sb As Se Te glass. The average value of hE i is 294.21G
1
4
29 52
5
g
4
kJ/mol.
3
.2. Crystallization
By using Kissinger’s formula [13,14] for the evaluation of
the activation energy of glass transition, E , for the
2
Fig. 4. lnða=T Þ vs. 1000/T
c 5
for Sb14As29Se52Te glassy alloy.
c
vs. 1/T shown in Fig. 6(b) according to the relation [19,22]
p
g
ꢀ
ꢁ
homogeneous crystallization with the spherical nuclei, it has
been suggested that [19,20] the dependence of T on a is given
by
Tp
Ec
ln
ZK
Kln K0
(7)
c
a
RTp
ꢀ
where T is the onset temperature of crystallization. The
ꢁ
from the linear dependence Ec was found to be equal
9
a
Ec
K1
ln
ZK
Cconst
(5)
2
117.93 KJ/mol and the frequency factor K
0
Z4.54!10 s
.
Tc
RTc
The area under the DSC curve is directly proportional to the
total amount of the alloy crystallized. The ratio between the
ordinates and the total area of the peak gives the corresponding
crystallization rates, which make it possible to plot the curves
of the exothermal peaks represented in Fig. 6. It may be
c
2
relationship between ln½a=T ꢀwith 1/T is shown in Fig. 4. The
c
c
calculated value of E was found to be equal to 123.5 kJ/mol.
c
The activation energy of crystallization (E ) can be
c
evaluated by using the variation of T with a Bansal et al.
p
observed that, the (dx/dt) values increases as well as the
p
[
21] developed a method for non-isothermal analysis as follows
heating rate, a property which has been widely discussed in
Ref. [23]. With the aim of correct applying of the preceding
theory, the material was reheated up to 460 K (a temperature
ꢀ
ꢁ
ꢂ
ꢃ
2
Tp
Ec
Ec
R
ln
ZK
Cln
Kln K0
(6)
a
RTp
slightly higher than T ) for 1 h in order to form a large number
g
2
p
of nuclei. It was ascertained by X-ray diffraction that, there is
no crystalline peaks were detected after the nucleation
treatment.
where K is the frequency factor. The plot of ln½T =aꢀ vs. 1/T
0
p
is shown in Fig. 5(a) from the linear dependence E was found
c
to be equal 113.44 KJ/mol. Furthermore, from the intercept of
this linear relation one can deduce the value of the frequency
For non-isothermal crystallization, the volume fraction c of
crystals precipitated in the heated glass at a uniform rate a is
8
K1
factor K (K Z1.76!10 s ). The relation between ln[T /a]
0
0
p
2
g
2
Fig. 3. (a) lnða=T Þ vs. 1000/T
g
, (b) ln(a) vs. 1000/T
g
for Sb14As29Se52Te
5
Fig. 5. (a) lnðTp =aÞ vs. 1000/T
p p p 5
, (b) ln(T /a) vs. 1000/T for Sb14As29Se52Te
alloy.
glass.

A.A. Othman et al. / Solid State Communications 138 (2006) 184–189
187
Fig. 6. Crystallization rate (dc/dt) vs. temperature of the exothermal peaks at
Fig. 8. ln[Kln(1Kc)] vs. 1000/T plots for Sb14As29Se52Te
5
glass at different
different heating rates.
heating rates.
related to E by the expression [24,25]
c
between ln[Kln(1Kx)] and ln a for as-quenched glass at fixed
temperature 540 K and reheated glass at 529 K. According to
Eq. (8), the slopes of these lines give the n-values, and it was
found that, nZ2.885 for the as-quenched glass and nZ2.888
for the reheated glass. Allowing to experimental error, both
values are close to three. This indicates that a large number of
nuclei exist already in the material, and, therefore, mZnZ3,
that is consistent with the mechanism of volume nucleation
with two dimensional growth.
ꢂ
ꢃ
mEc
ln½Klnð1KcÞꢀ ZKn ln aK1:052
RT
Cconst:
(8)
where rn and n are the numerical factors already mentioned.
When nuclei formed during the heating at constant rate (a) are
dominant, n is equal to mC1 and when nuclei formed during
any previous heat treatment prior to thermal analysis are
dominant, n is equal to m [24,25].
On the other hand, Fig. 8 shows the plots of ln[Kln(1Kx)]
vs. 1/T for all heating rates (2.5 up to 30 K/min). From the
slopes of these linear relations, the average value of the
crystallization activation energy is 120.5 KJ/mol.
From the experimental data, the plots of ln[Kln(1Kx)] vs.
ln a at different specific temperatures have been drawn, both
for the as-quenched glass and for the reheated glass. It has been
observed that, the correlation coefficients of the corresponding
straight regression lines show a maximum value for a given
temperature, which was considered as the most adequate one
for the calculation of parameter n. Fig. 7 shows the relation
The activation energy for crystallization (E ) was deter-
c
mined by using Coats–Redfern–Sestak [26] equation in the
following form:
ꢀ
ꢁ
ꢂ
ꢃ
lnð1KcÞ
nE =R
c
ln K
ZK
Cconst:
(9)
2
n
T
T
2n
Fig. 9 shows the linear relation between ln½Klnð1KcÞ=T ꢀ wit
/T at different heating rates. The slope of the linear portions of
1
at different heating rates (with nZ3), Fig. 9, equals nE /R; thus
c
the average activation energy of crystallization E Z
c
126.158 kJ/mol.
Finally, Augis and Bennett [27] suggested a method to
deduce The crystallization activation energy, E , in the form
c
ꢀ
ꢁ
a
ln
Z ln K KE =RT
p
(10)
0
c
Tp KT0
where T is the initial temperature, when ln[a/(T KT )] is
0
p
0
plotted vs. 1/T a straight line is obtained whose slope is E /R as
c
p
shown in Fig. 10(a),The value of E for Sb As Se Te glass
c
14 29 52
5
calculated by using this method is 117.67 kJ/mol. The value of
the crystallization activation energy, E , could be calculated
using Takhor method [28] via the following equation with the
Fig. 7. Variation of ln[Kln(1Kc)] with logarithm of heating rate (a in K/s), (a)
as quenched glass 560 K, and (b) reheating glass 548 K.
c

188
A.A. Othman et al. / Solid State Communications 138 (2006) 184–189
5
Fig. 11. (a) Diffractogram of glassy alloy Sb14As29Se52Te (b) diffraction beaks
of the alloy after thermal treatment (annealing at T for 2 h).
p
with d and I, respectively, while there remain also a residual
amorphous phase.
6
Fig. 9. ln[Kln(1Kc)/T ] vs. 1000/T for Sb14As29Se52Te
5
glass at different
heating rates.
assumption ln(T KT ) is constant [4,29]
p
5. Conclusions
0
Ec
lnðaÞ ZK
Cconstant
(11)
(
1) Two approaches were used to analyzing the glass
transition. The first one is the linear dependence of the
glass transition temperature on the logarithm of the heating
RTp
On this basis, the slope of ln(a) vs. 1/Tp relation,
Fig. 10(b)), yields the activation energy, E for Sb As Se
Te5 glass. The value of E deduced from this method is
(
rate. The second is the linear relationship between the
2
c
14 29 52-
c
logarithm of the quotient ða=T Þ and the reciprocal of the
g
1
22.43 KJ/mol.
glass transition temperature.
(
2) The numerical factors, n and m depend on the mechanism
of nucleation, growth, and the dimensionality of the
crystal. In addition, nZmC1 for as quenched glass
containing no nuclei while nZm for a glass containing a
sufficiently large number of nuclei. The kinetic parameters
were deduced based on the mechanism of crystallization.
The n value of the kinetic exponent (nZ3) of the as
quenched glass is consistent with the mechanism of
volume nucleation with two dimensional growth.
4
. Identification of the crystalline phases
The crystalline phases during the thermal treatment applied
to the sample can be identified by means of XRD
measurements. For this purpose, Fig. 11(a) shows the absence
of any crystalline peaks, which are characteristic as quenched
glass. The diffractogram of the transformed material after heat
treatment (Fig. 11(b)) suggests the presence of some crystal-
lites of As, Te, AsSb, As Se , Sb Se and AsSe Te indicated
(
3) The crystallization kinetics of the glassy alloy Sb As
29-
1
4
2
3
2
3
.5 .5
Se Te with and without previous reheating. According to
2
5
5
the study carried out, it is possible to establish that, the
reheating did not cause the appearance of nuclei, while the
as-quenched material already contains a sufficient number
of them.
(
4) Finally, the identification of the crystalline phases was
made by recording the X-ray diffraction pattern of the
transformed material. This pattern shows the existence of
some crystallites of As, Te, AsSb, As Se , Sb Se and
2
3
2
3
AsSe Te in an amorphous matrix.
.5
.
5
References
[
1] Z. Altounian, J.O. Strom-Olsen, in: R.D. Shull, A. Joshi (Eds.), Thermal
Analysis in Metallurgy, The Minerals, Metals and Materials Society,
Warrendale, PA, 1992, p. 155.
Fig. 10. (a) ln[a/(T
Se52Te glass.
p 0 p p
KT )] vs. 1000/T , (b) ln(a) vs. 1000/T for Sb14As29-
5
[2] D.W. Henderson, J. Non-Cryst. Solids 30 (1979) 301.

A.A. Othman et al. / Solid State Communications 138 (2006) 184–189
189
[
3] S. Mahadevan, A. Giridhar, A.K. Singh, J. Non-Cryst. Solids 88
[16] M.M. Hafiz, N. Afify, M. Osman, M. Dongol, J. Therm. Anal. 36 (1990)
2417.
(1986) 11.
[
[
[
[
[
4] N. Afify, J. Non-Cryst. Solids 142 (1992) 247.
[17] N. Afify, A.H. Moharram, M. Dongol, M.M. Hafiz, Phys. B 176 (1992) 301.
[18] A. El-Korashy, N. Afify, J. Therm. Anal. 37 (1991) 507.
[19] H.S. Chen, J. Non-Cryst. Solids 27 (1978) 257.
[20] J.E. Shelby, J. Non-Cryst. Solids 34 (1979) 111.
[21] N.P. Bansal, R.H. Doremus, A.J. Bruce, C.T. Moynihan, J. Am. Ceram.
Soc. 66 (1983) 233.
5] N. Rysava, T. Spasov, L. Tichy, J. Therm. Anal. 32 (1987) 1015.
6] H.E. Kissinger, Anal. Chem. 29 (1957) 1702.
7] Z.G. Ivanova, E. Cernoskova, Thermochim. Acta 411 (2004) 177–180.
8] K.F. Kelton, Crystal Nucleation in Liquids and Glasses, Solid State
Physics, vol. 45, Academic Press, New York, 1991.
[
9] N. Clavaguera, J. Non-Cryst. Solids 162 (1993) 40.
[22] T. Ozawa, Bull. Chem. Soc. Jpn. 38 (1965) 1881.
[23] Y.Q. Gao, W. Wang, F.Q. Zheng, X. Liu, J. Non-Cryst. Solids 81 (1986)
135.
[
[
10] S.A. Khan, M. Zulfequar, M. Husain, J. Phys. Chem. Solids 123 (2002)
63–468.
4
11] Perkin Elmer, PC Series, Thermal Analysis System, DSC 7 Differential
Scanning Calorimeter, Operators Manual, Norwalk, Connecticut, 1989.
12] A.B.A. El-Moiz, N. Afify, M.M. Hafiz, Phys. B 182 (1992) 33.
13] H.S. Chen, J. Non-Cryst. Solids 27 (1978) 257.
[24] N. Afify, Phys. B 179 (1992) 48.
[25] K. Matusita, T. Komatsu, R. Yokota, J. Mater. Sci. 19 (1984) 291.
[26] A.W. Coats, J.P. Redfern, Nature 201 (1964) 68.
[27] J.A. Augis, J.E. Bennett, J. Therm. Anal. 13 (1978) 283.
[28] R.L. Takhor, Advances in Nucleation and Crystallization of Glasses,
American Ceramic Society, Columbus, 1972. p. 166.
[29] H. Yinnon, D.R. Uhlmann, J. Non-Cryst. Solids 54 (1983) 253.
[
[
[
[
14] J.A. Macmillan, J. Phys. Chem. 42 (1965) 3497.
15] S. Mahadevan, A. Giridhar, A.K. Singh, J. Non-Cryst. Solids 88
(1986) 11.