Appl Phys A (2010) 99: 913919

DOI 10.1007/s00339-010-5680-6
Optical band gap and refractive index dispersion parameters
of As
x
Se
70
Te
30x
(0 x 30 at.%) amorphous lms
Kamal A. Aly
Received: 15 March 2010 / Accepted: 19 March 2010 / Published online: 16 April 2010
Springer-Verlag 2010
Abstract Amorphous As
x
Se
70
Te
30x
thin lms with (0
x 30 at.%) were deposited onto glass substrates by us-
ing thermal evaporation method. The transmission spectra
T () of the lms at normal incidence were measured in the
wavelength range 4002500 nm. A straightforward analysis
proposed by Swanepoel based on the use of the maxima and
minima of the interference fringes has been used to drive the
lm thickness, d, the complex index of refraction, n, and the
extinction coefcient, k. The dispersion of the refractive in-
dex is discussed in terms of the single-oscillator Wemple and
DiDomenico model (WDD). Increasing As content is found
to affect the refractive index and the extinction coefcient of
the As
x
Se
70
Te
30x
lms. With increasing As content the op-
tical band gap increases while the refractive index decreases.
The optical absorption is due to allowed indirect transition.
The chemical bond approach has been applied successfully
to interpret the increase of the optical gap with increasing
As content.
1 Introduction
Chalogenide glasses have been recognized as promising ma-
terials for infrared optical element, infrared optical bers,
and for the transfer of information [13]. They have also
found applications in xerography switching and memory de-
vices, photolithographic process, and in the fabrication of in-
expensive solar cells and more recently as reversible phase
change optical recorders [48]. This has made it important
K.A. Aly ()
Physics Department, Faculty of Science, Al-Azhar University,
Assiut, Egypt
e-mail: kamalaly2001@Gmail.com
to have an insight into their optical and electronic proper-
ties. The addition of an impurity has a pronounced effect
on the conduction mechanism and the structure of the amor-
phous glass and this effect can be widely different for differ-
ent impurities [9]. Therefore, the ternary compounds involv-
ing AsSeTe have interesting properties as well as tech-
nological applications because they form a wide range of
glassy region. Although it is possible to nd in the literature
more papers dealing with AsSeTe thin lms [1016] but,
to the best of our knowledge, the effect of As content on the
optical constants of AsSeTe have not been reported. The
present work deals with investigation of the optical proper-
ties of the As
x
Se
70
Te
30x
(0 x 30 at.%) thin lms. The
well-known Swanepoels method [17, 18] is used to accu-
rately determination of the refractive index and lm thick-
ness in the weakly absorbing and transparent regions of the
spectrum. Also, the absorption coefcient, and therefore the
extinction coefcient, has been determined in the strong ab-
sorption region of the transmission spectra.
2 Experimental details
Different compositions of bulk As
x
Se
70
Te
30x
(0 x
30 at.%) chalcogenide glasses were prepared from their
components of high purity (99.999%) by the usual melt
quenching technique. The elements were heated together in
an evacuated silica ampoule up to 1200 K and then the am-
poule temperature kept constant for about 20 h. During the
course of heating, the ampoule was shaken several times to
maintain the uniformity of the melt. Finally, the ampoule
was quenched into ice-cooled water to avoid the crystalliza-
tion process.
The amorphous thin lms were deposited by evaporat-
ing the alloys from a resistance-heat quartz glass crucible
914 K.A. Aly
Fig. 1 X-ray diffraction patterns of as prepared As
x
Se
70
Te
30x
with
(0 x 30 at.%) thin lms
onto clean glass substrates kept at room temperature and a
vacuum of about 210
6
Torr using a conventional coat-
ing unit (Denton Vacuum DV 502 A). The evaporation rate
as well as the lm thickness was controlled using a quartz
crystal DTM 100 monitor. Mechanical rotation of the sub-
strate holder (30 rpm) during deposition-produced homo-
geneous lm. The temperature rise of the substrate due to
radiant heating from crucible was negligible.
The amorphous nature of the as-deposited lms was
checked using a Philips X-ray diffractometer (1710). The
chemical compositions of the as-deposited lms were mea-
sured using an energy dispersive X-ray spectroscopy (Link
analytical EDS). The compositions so determined agreed
with those of the starting materials to within 0.35 at.%.
The optical transmittance at normal incidence was mea-
sured in the wavelength range 4002500 nm using a double-
beam computer-controlled spectrophotometer (Jasco V-630
combined with PC). The spectrophotometer was set with a
slit width of 1 nm and as this was much smaller than the line
widths it was unnecessary to make slit-width corrections.
The line width is simply taken to be the separation of two
adjacent interference maxima and minima. Without a glass
substrate in the reference beam, the measured transmittance
spectra were used to calculate the optical constants by ap-
plying the envelope method suggested by Swanepoel [17].
3 Results and discussion
Figure 1 represents the XRD patterns for As
x
Se
70
Te
30x
(0 x 30) thin lms; as shown in this gure the lms did
not reveal discrete or any sharp peaks but the characteristic
broad humps of the amorphous materials.
3.1 Calculation of the refractive index and lm thickness
Figure 2a shows the measured transmittance (T ) spectra for
different compositions of As
x
Se
70
Te
30x
thin lms. From
Fig. 2 (a) Transmission spectra for different compositions of
As
x
Se
70
Te
30x
with (0 x 30 at.%) thin lms. (b) Transmission
spectra for Se
70
Te
30
thin lms. The T
M
, T
m
, and T

curves according
to the text, T
s
is the transmission of the substrate alone
this gure one can note that the addition of As content
at the expense of Te content shifts the optical transmit-
tance to the higher energies (i.e., blueshift of the optical ab-
sorption edge). Figure 2b as a comparative example, shows
the measured transmittance (T ), the created envelopes, T
M
and T
m
, (both the envelopes being computer-generated us-
ing the Origin Lab version 7 program), and the geometric
mean, T

T
M
T
m
, in the spectral region with interference
fringes [18] for Se
70
Te
30
thin lm.
According to Swanepoels method based on the idea of
Manifacier et al. [19], the rst approximate value of the
refractive index of the lm, n
1
, in the spectral region of
medium and weak absorption can be calculated as well as
detailed in Ref. [18]. Using the values of, n
1
, and taking
into account the basic equation for the interference fringes:
2nd =m
o
(1)
where the order number, m
o
, is an integer for maxima and
a half-integer for minima the rst approximate value of the
lm thickness, d
1
, can be expressed as:
d
1
=

1
.
2
2(n
c2

1
n
c1
.
2
)
(2)
where, n
c1
, and, n
c2
, are the refractive indices at two ad-
jacent maxima (or minima) at,
1
, and,
2
. The last value
deviates considerably from the other values and must conse-
quently be rejected. This deviation is an indication that n
c1
is not accurate enough due to the departure of the hypoth-
esis of transparency behind the application of the envelope
method [17]. The average value (d
1
) of d
1
(ignoring the last
value) can now be used along with n
1
to calculate m
o
for
the different maxima and minima using (1). The accuracy
Optical band gap and refractive index dispersion parameters of As
x
Se
70
Te
30x
(0 x 30 at.%) 915
Fig. 3 The plots of l/2 vs. n/,
in order to determine the lm
thickness and the rst-order
number m
1
for As
x
Se
70
Te
30x
with (0 x 30 at.%) thin
lms
of the lm thickness can now be signicantly increased by
taking the corresponding exact integer or half-integer val-
ues of m
o
associated with each extreme point (see Fig. 2b)
and deriving a new thickness, d
2
, using (1), again using the
values n
1
. The values of the thickness in this way have a
smaller dispersion. It should be emphasized that the accu-
racy of the nal thickness, d
2
, is better than 1% as well as
reported elsewhere [18]. With the accurate values of m
o
and
(d = d
2
) expression (1) can then be solved for n at each
and, thus, the nal values of the refractive index, n
2
, are ob-
tained.
Furthermore, a simple complementary graphical method
for deriving the rst-order number m
1
and the lm thick-
ness d, based on (1), was also used. For this purpose (1) is
rewritten as follows for the successive maxima and minima,
starting from the long-wavelength end [18]:
l
2
=2d

m
1
, l =0, 1, 2, 3, . . . (3)
where, m
1
is the order number of the rst (l = 0) extreme
considered, an integer for a maximum and a half integer
for a minimum. Therefore, by plotting (l/2) versus (n/)
a straight line with slope 2d and cut-off on the Y-axis at
m
1
. Figure 3 shows this plot, in which the values obtained
for (d =d
2
=0.5 slope value) and m
1
for each sample of
the As
x
Se
70
Te
30x
thin lms as well as denoted on the same
graph.
Now the values of n
2
can be tted to a reasonable func-
tion such as the two-term Cauchy dispersion relationship
[18]:
n() =a +b/
2
(4)
where a and b are constants, then (4) can be used to ex-
trapolate the wavelength dependence beyond the range of
measurement [17, 18]. Figure 4 illustrates the dependence
of the refractive index, n, on wavelength for different com-
positions of the amorphous As
x
Se
70
Te
30x
thin lms. The
relative error in n, n/n, does not exceed the precision of
916 K.A. Aly
Fig. 4 Refractive index dispersion spectra for As
x
Se
70
Te
30x
with
(0 x 30 at.%) thin lms. The solid curves were determined ac-
cording to Cauchy dispersion relationship [18]
the measurements T/T (1%). The least-squares t of n
2
values (solid lines of Fig. 3) for the different samples, yields
n = 3.21 + (3.31 10
5
/
2
), n = 3.02 + (3.07 10
5
/
2
),
n = 2.90 + (2.53 10
5
/
2
), n = 2.74 + (2.12 10
5
/
2
),
n = 2.62 + (1.98 10
5
/
2
), n = 2.49 + (1.87 10
5
/
2
),
and n =2.41 +(1.55 10
5
/
2
) for x =0, 5, 10, 15, 20, 25,
and 30 at.%, respectively. As shown in Fig. 4 the refractive
index, n, decreases with increasing wavelength of the inci-
dent photon, while at higher wavelengths the refractive in-
dex, n, tends to be constant for all compositions under study.
Here the values of refractive index for all compositions can
be tted according to the WempleDiDomenico (WDD) dis-
persion relationship [20];
n
2
(h) =1 +
E
0
E
d
E
2
0
(h)
2
(5)
where E
0
is the single-oscillator energy and E
d
is the disper-
sion energy or single-oscillator strength where the refractive
index factor (n
2
1)
1
can be plotted as a function of (h)
2
and tting straight lines as shown in Fig. 5, the values of the
E
0
and E
d
can be determined from the intercept E
0
/E
d
and
the slope (E
0
E
d
)
1
. As mentioned before by Tanaka [21]
that the oscillator energy (E
0
) is an average energy gap and
to a good approximation, scales with the optical band gap
(E
g
), E
0
2E
g
as shown in Table 1.
Figure 5 also shows the values of the refractive index
n(0) at h = 0 of the As
x
Se
70
Te
30x
thin lms. The ob-
tained values of E
0
, E
d
, and n(0) are listed in Table 1. It
was observed that the single-oscillator energy E
0
increases
while both the dispersion energy E
d
and n(0) decrease with
the increase of As content. An important achievement of the
WDD model is that it relates the dispersion energy, E
d
, to
other physical parameters of the material through the fol-
lowing empirical relationship [20]:
E
d
=N
c
Z
a
N
e
(eV) (6)
Fig. 5 Plots of refractive index factor (n
2
1)
1
vs. (h)
2
for
As
x
Se
70
Te
30x
with (0 x 30 at.%) thin lms
where N
c
is the effective coordination number of the cation
nearest neighbor to the anion, Z
a
is the formal chemical va-
lency of the anion, N
e
is the effective number of valence
electrons per anion, and is a two-valued constant with ei-
ther an ionic or a covalent value (
i
= 0.26 0.03 eV and

c
= 0.37 0.04 eV, respectively). Therefore, in order to
account for the compositional trended of E
d
it is suggested
that the observed decrease in E
d
with increasing As content
is primarily due to the change in the ionicities (homopolar
SeSe bonds are introduced together with extra Se atoms),
which decreases with increasing As content (see Table 1).
The values of the single-oscillator energy, the dispersion en-
ergy, the static refractive index, and the excess of SeSe ho-
mopolar bonds for the As
x
Se
70
Te
30x
thin lms are listed
in Table 1. In addition, the fundamental electron excitation
spectrum of a substance is generally described in terms of a
frequency-dependent complex electronic dielectric constant
(() =
1
() + i
2
()) either the real part
1
() or the
imaginary part
2
() contains all desired response informa-
tion since causality arguments relate the real and imaginary
parts. Therefore, the single-oscillator and dispersion energy
parameterization given by (5) are dened by Ref. [22],
E
2
0
=
M
1
M
3
and E
2
d
=
M
3
1
M
d
(7)
The oscillator energy E
0
is independent of the scale of

2
and is consequently an average energy gap, whereas
E
d
depends on the scale of
2
and thus serves an inter-
band strength parameter. Since the M
1
and M
3
moments
are involved in computation of E
0
and E
d
,
2
spectrum is
weighted most heavily near the interband absorption thresh-
old. As a result, the dispersion energy may depend upon
the detailed charge distribution within each unit cell, conse-
quently, would then be closely related to chemical bonding
that may lie within a nearly localized orbital theory.
Optical band gap and refractive index dispersion parameters of As
x
Se
70
Te
30x
(0 x 30 at.%) 917
Table 1 The WempleeDiDomenico dispersion parameters, E
0
, E
d
, M
1
, M
3
, the values of the refractive index, n(0), extrapolated at h =0,
the average coordination number, N
co
, the excess of SeSe homopolar bonds, the optical band gap, E
g
, E
0
/E
g
ratio, and the cohesive energy, CE,
for As
x
Se
70
Te
30x
with (0 x 30 at.%) thin lms
Composition eV n(0) N
co
Excess of E
g
E
g
/E
0
CE
E
0
E
d
M
1
M
3
SeSe (eV) (eV/atom)
Se
70
Te
30
2.924 27.58 9.188 1.075 3.23 2.00 80 1.40 2.094 1.911
As
5
Se
70
Te
25
2.939 24.39 8.410 0.974 3.05 2.05 75 1.44 2.041 1.942
As
10
Se
70
Te
20
3.094 22.92 8.069 0.843 2.90 2.10 70 1.49 2.073 1.974
As
15
Se
70
Te
15
3.233 21.33 7.691 0.736 2.76 2.15 65 1.56 2.075 2.006
As
20
Se
70
Te
10
3.382 20.25 7.429 0.650 2.64 2.20 60 1.64 2.062 2.038
As
25
Se
70
Te
5
3.450 18.24 6.929 0.582 2.51 2.25 55 1.72 2.007 2.070
As
30
Se
70
3.622 17.75 6.805 0.519 2.43 2.3 50 1.81 1.997 2.102
3.2 Determination of the extinction coefcient and optical
band gap
Since the values of the refractive index, n, are already known
over the whole spectral range 4002500 nm, the absorbance
x
a
() can be calculated using the interference-free transmis-
sion spectrumT

(see Fig. 2) using the well-known equation

suggested by Connell and Lewis [23]:
X
a
=
P +[P
2
+2QT

(1 R
2
R
3
)]
Q
(8)
where P =(R
1
1)(R
2
1)(R
3
1) and Q=2T

(R
1
R
2

R
1
R
3
2R
1
R
2
R
3
), R
1
is the reectance of the airlm
interface (R
1
= [(1 n)/(1 + n)]
2
), R
2
is the reectance
of lmsubstrate interface (R
2
= [(n s)/(n + s)]
2
), and
R
3
is the reectance of the substrateair interface (R
3
=
[(s l)/(s + 1)]
2
). Moreover, since d is known, the rela-
tion x
a
=exp(d) can then be solved for the values of the
absorption coefcient, . In order to complete the calcula-
tion of the optical constants, the extinction coefcient, k, is
calculated using the values of and through the already-
mentioned formula, k =/4.
Figure 6 illustrates the dependence of the absorption co-
efcient, , on the wavelength for As
x
Se
70
Te
30x
(0 x
30 at.%) thin lms. For 10
5
cm
1
, the imaginary part of
the complex index of refraction is much less than n, so that
the previous expressions used to calculate the reectance is
valid. In the region of strong absorption, the interference
fringes disappear; in other words, for a very large , the
three curves T
M
, T

, and T
m
converge to a single curve.
According to Taucs relation [24, 25] for allowed indirect
transitions, the photon energy dependence of the absorption
coefcient can be described by
(h)
1/2
=B
1/2
(h E
g
) (9)
where B is a parameter that depends on the transition prob-
ability and E
g
is the optical energy gap. Figure 7 shows the
Fig. 6 The absorption coefcient, , as a function of the wavelength,
, for As
x
Se
70
Te
30x
with (0 x 30 at.%) thin lms
Fig. 7 The absorption coefcient in the form of (h)
1/2
versus pho-
ton energy (h) for As
x
Se
70
Te
30x
with (0 x 30 at.%) thin lms.
from which the optical band gap (E
g
) is estimated (Taucs extrapola-
tion)
absorption coefcient in the form of (h)
1/2
versus h for
the As
x
Se
70
Te
30x
thin lms. The intercepts of the straight
lines with the photon energy axis yield values of the optical
band gap, E
g
.
918 K.A. Aly
Table 2 Bond energies and the relative probabilities of formation of
various bonds in AsSeTe glasses, taking the probability of SeTe
bond as unity
Bond Bond energy Relative probability
(kcal mol
1
) (at T =298.15 K)
SeTe 44.197 1
SeSe 44.04 0.763
AsSe 41.71 0.015
TeTe 33.00 5.872 10
9
AsTe 32.74 3.782 10
9
AsAs 32.10 1.28 10
9
According to the chemical-bond approach [26, 27],
bonds are formed in the sequence of decreasing bond en-
ergy until the available valence of atoms is satised. The
bond energies D(AF) for heteronuclear bonds have been
calculated by using the empirical relation;
D(AF) =[D(AA) D(F F)]
1/2
+30(
A

F
)
2
(10)
proposed by Pauling [28], where D(AA) and D(F F)
are the energies of the homonuclear bonds (44.04, 30.22,
and 33 kcal/mol for As, Se, and Te, respectively) [18, 29],

A
and
F
are the electronegativity values for the involved
atoms [28].
The energies of various possible bonds in the AsSeTe
system are given in Table 2. Depending on the bond en-
ergy (D), the relative probability of its formation was cal-
culated [30] using the probability function exp(D/kT )and
listed in Table 2. Bonds such as TeTe, AsTe and AsAs
have insignicant probability of formation because of their
low bond energies. Therefore, only SeTe, AsSe, and Se
Se bonds exist with high priority in the AsSeTe system.
The observed increase of the E
g
with increasing the As
content can be attributed to the formation of AsSe bonds
(E
g
= 1.55 eV) increases at the expense of SeTe bonds
(E
g
= 1.3 eV) and also a shortage of homopolar SeSe
bonds.
Knowing the bond energies we can estimate the cohe-
sive energy (CE), i.e., the stabilization energy of an innitely
large cluster of the material per atom, by summing the bond
energies over all the bonds expected in the systemunder test.
The CE of the prepared samples is evaluated from the fol-
lowing equation [31];
CE =

(C
i
D
i
/100) (11)
where C
i
and D
i
are the numbers of the expected chemical
bonds and the energy of each corresponding bond, respec-
tively. The calculated values of the cohesive energies for all
compositions are presented in Table 1. It is observed that the
values of CE increases with the increase of As content. The
increase in the CE with increasing As content is due to the
decrease in the excess of SeSe homopolar bonds (see Ta-
ble 1). This result is in a good agreement with many authors
[32, 33].
4 Conclusions
Optical characterization of As
x
Se
70
Te
30x
thin lms with
(0 x 30 at.%) have been analyzed using the Swanepoels
method, which is based on the generation of the envelopes
of the interference maxima and minima of the transmis-
sion spectrum. Allowed indirect electronic transitions are
mainly responsible for the photon absorption in the investi-
gated lms. Fitting of the refractive indices according to the
single-oscillator model WempleDiDomenico (WDD) rela-
tionship results in the dispersion of parameters that are di-
rectly related to the structure of these lms. It was found
that the optical band gap (E
g
) and the single oscillator
energy (E
0
) increase, while the refractive index (n) and
the dispersion energy (E
d
) decrease on increasing the As
content. A chemical-bond approach has been applied suc-
cessfully to interpret the increase of the optical gap of the
As
x
Se
70
Te
30x
lms with increasing As content.
Acknowledgements The author wish to thank the Optics Lap. at the
Physics Department of the Faculty of Science, Al-Azhar University,
Assuit, Branch, for achieving the optical measurements and the nan-
cial support for the (XRD) measurements, also the Author would like to
acknowledge Dr. A. Dahshan Dep. of Phys., Faculty of Science, Suez
Canal University, Port Said, Egypt, for his help and advice throughout
this work.
References
1. E. Marquez, J.M. Gonzalez-Leal, R. Jimenez-Garay, M. Vlcek,
Thin Solid Films 396, 183 (2001)
2. T. Ohta, J. Opto-electron. Adv. Mater. 3, 609 (2001)
3. E. Marquez, P. Villars, R. Jimenez-Garay, J. Mater. Res. 3, 314
(1988)
4. A. Zakery, S.R. Elliott, J. Non-Cryst. Solids 330, 1 (2003)
5. S.R. Ovshinsky, Phys. Rev. Lett. 21, 1450 (1986)
6. N.F. Mott, Philos. Mag. 24, 911 (1971)
7. D.E. Carlson, C.R. Wronski, Appl. Phys. Lett. 28, 671 (1976)
8. J. Fusong, M. Okuda, Jpn. J. Appl. Phys. 30, 97 (1991)
9. N.E. Mott, Philos. Mag. 19, 835 (1969)
10. M.M. El-Nahass, M.B. El Den, Opt. Laser Technol. 33, 31 (2001)
11. M.B. El-Den, M.M. El-Nahass, Opt. Laser Technol. 35, 335
(2003)
12. V. Lyubin, T. Tada, M. Klebanov, N.N. Smirnov, A.V. Kolobov,
K. Tanaka, Mater. Lett. 30, 79 (1997)
13. L.A. Wahab, S.A. Fayek, Solid State Commun. 100, 345 (1996)
14. R.A. Ligero, M. Casas-Ruiz, A. Orozco, M.P. Trujillo, R. Jimnez,
Thermochim. Acta 249, 221 (1995)
15. M. Roilos, J. Non-Cryst. Solids 6, 5 (1971)
16. T. Takahashi, J. Non-Cryst. Solids 34, 307 (1979)
17. R. Swanepoel, J. Phys. E 16, 1214 (1983)
18. A. Dahshan, H.H. Amer, K.A. Aly, J. Phys., D. Appl. Phys. 41,
215401 (2008) (7pp)
Optical band gap and refractive index dispersion parameters of As
x
Se
70
Te
30x
(0 x 30 at.%) 919
19. J.C. Manifacier, J. Gasiot, J.P. Fillard, J. Phys. E 9, 1002 (1976)
20. S.H. Wemple, M. DiDomenico, Phys. Rev. B 3, 1338 (1971)
21. K. Tanaka, Thin Solid Films 66, 271 (1980)
22. S.H. Wemple, M. DiDomenico Jr., Phys. Rev. Lett. 23, 1156
(1969)
23. G.A.N. Connell, A.J. Lewis, Phys. Status Solidi b 60, 291 (1973)
24. E.A. Davis, N.F. Mott, Philos. Mag. 22, 903 (1970)
25. H. Fritzsche, Philos. Mag. B 68, 561 (1993)
26. J. Bicermo, S.R. Ovshinsky, J. Non-Cryst. Solids 74, 75 (1985)
27. B. Jozef, O. Stanford, S. Mahadevan, A. Gridhar, A.K. Singh,
J. Non-Cryst. Solids 74, 75 (1985)
28. J. Pauling, Nature of the Chemical Bond Ithaca (Cornell Univer-
sity Press, Ithaca, 1960)
29. A. Dahshan, K.A. Aly, Acta Mater. 56, 4869 (2008)
30. D.R. Goyal, A.S. Maan, J. Non-Cryst. Solids 183, 182 (1995)
31. S.A. Fayek, J. Phys. Chem. Solids 62, 653 (2001)
32. S.A. Fayek, M. El-Ocker, A.S. Hassanien, Mater. Chem. Phys. 70,
231 (2001)
33. S.S. Fouad, A.E. Bekheet, A.M. Farid, Physica B 322, 163 (2002)