You are on page 1of 4

REVIEW OF SCIENTIFIC INSTRUMENTS 80, 046107 2009

Determination of energy band gap of nanocrystalline SbSI using diffuse


reflectance spectroscopy
M. Nowak,1,a B. Kauch,2 and P. Szperlich1
1

Institute of Physics, Silesian University of Technology, ul. Krasiskiego 8, 40-019 Katowice, Poland
Department of Biophysics, Agricultural University SGGW, ul. Nowoursynowska 159, 02-776 Warsaw, Poland

Received 11 December 2008; accepted 2 March 2009; published online 21 April 2009
Twelve methods of determining energy band gap Eg of semiconductors using diffuse reflectance
spectroscopy have been applied in investigations of sonochemically produced antimony sulfoiodide
SbSI consisting of nanowires. It has been proved that the best method of determining Eg is based
on simultaneous fitting of many mechanisms of absorption to the spectral dependence of Kubelka
Munk function evaluated from the diffuse reflectance data. It allows determining the values of
indirect forbidden Eg, the Urbach energy, and the constant absorption/scattering of the examined
semiconductor. 2009 American Institute of Physics. DOI: 10.1063/1.3103603
Optical absorption spectra have been extensively used as
one of the most important tools for probing the energy gaps
Eg and band structures of semiconductors.1 There are several methods for measuring them. These include diffuse reflectance spectroscopy DRS. When a material, consisting
of many particles, or nanoparticles, is illuminated, some of
the impinging radiation penetrates the sample and some is
reflected from its surface. The portion that penetrates the
sample is scattered at a large number of points in its path as
well as it is transmitted through the particles a number of
times. Only the part of this radiation that is returned to the
surface of the sample and comes back out is considered to be
diffuse reflection.2
DRS is a suitable, not destructive, and simple method of
investigation, especially important in examinations of porous, nanocrystalline materials and gels.3 It is impossible to
examine such materials applying specular reflection and it is
also extremely difficult to determine the path length in an
optical transmittance of them. The following factors are related to high spectral quality of diffuse reflectance: dilution
of the sample with a nonabsorbing matrix ensures a deeper
penetration of the incident beam into the sample, which increases the contribution of the scattered component in the
spectrum and minimizes the specular reflection component;
and smaller particles improve the quality of DRS spectra.
However, there is no one reliable method for determining Eg while applying the DRS. The already used methods of
DRS investigations have been applied in this paper to determine Eg of ultrasonically prepared nanowires of antimony
sulfoiodide SbSI. The SbSI being a semiconducting ferroelectric has an unusually large number of interesting properties that have been reviewed in a few monographs but they
are still investigated see the literature in, e.g., Ref. 3.
The SbSI xerogel consisting of nanowires was prepared
sonochemically from the elements Sb, S, and I. The used
experimental set up and the applied procedure were the same
as described in Ref. 3. Characterization of the xerogel was
done using powder x-ray diffraction, scanning electron microscopy, energy dispersive x-ray analysis, high-resolution
a

Electronic mail: marian.nowak@polsl.pl. Tel.: 48 32 603-41-67.

0034-6748/2009/804/046107/3/$25.00

transmission electron microscopy and selected area electron


diffraction. Description of the used equipment and the results
of investigations were given in Ref. 3.
The DRS measurements were carried out on a spectrophotometer SP-2000 equipped with an integrating sphere
ISP-REF Ocean Optics Inc.. Spectra were recorded at temperature of 293 K in the mode of exclusion of specular reflectance. The standard WS-1 Ocean Optics Inc. was used
as a reference. The spectrum of diffuse reflectance signal the
ratio Rd = Idiffuse / Istandard of intensities of radiation reflected in
diffuse manner from the sample and from the known standard of SbSI xerogel has been presented in Fig. 1. It corresponds to the ideal diffuse reflectance spectrum consisting of
the following: the nearly flat, low absorbance region at small
photon energies h Rd in this region can be dominated by
reflection and scattering due to the high refractive index of
the investigated material; the abrupt decrease in Rd when
radiation becomes more intensively absorbed with increasing
h it corresponds to an onset of transmission near the optical absorption edge, and; the also nearly flat, strong absorbance region at high h.
There are at least eleven methods in determining Eg using DRS. Method 1 see, e.g., Ref. 4: the onset of the linear
increase in diffuse reflectance with decreasing h Fig. 1a
in the supplementary material5 was taken as a measure of
Eg. Method 2 see, e.g., Refs. 4 and 6: the intersection of
straight-line extrapolations below and above the high photon
energy knee of the spectral characteristics of Rd Fig. 1b in
the supplementary material5 was used to determine Eg.
Method 3 see, e.g., Ref. 7: similarly to the case of
optical transmittance, the diffuse reflectance data were transformed in the so-called absorbance

A = d = log

Istandard
= logRd,
Idiffuse

where d is absorption path length and represents absorption coefficient of radiation. The Eg was derived as the intersection point between lines that extrapolate the absorbances
in the small h range and at the linear absorption edge Fig.
2 in the supplementary material5.
Method 4 see, e.g., Ref. 8: the Eg was determined by
the intersection point Fig. 3a in the supplementary

80, 046107-1

2009 American Institute of Physics

046107-2

Rev. Sci. Instrum. 80, 046107 2009

Nowak, Kauch, and Szperlich

TABLE I. Comparison of the values of energy gap of SbSI nanowires determined using different techniques of diffusive reflection spectroscopy.
Eg
eV

FIG. 1. Spectra of the diffuse reflectance coefficient and Kubelka


Munk function calculated for the DRS data of SbSI nanowires. Solid
curve represents the least square fitted theoretical dependence for the sum of
indirect forbidden absorption without excitons and phonon statistics, Urbach
ruled absorption, and constant absorption term description in the text, values of the fitted parameters are given in Table II.

material5 between h axis and the line extrapolated from the


linear portion of the absorption edge of so-called Kubelka
Munk function known as re-emission function, too2
FKMRdh =

1 Rdh2 h
=
,
2Rdh
S

where S is the scattering factor. In some cases, S is approximately equal to unity for very small participles or it is
constant e.g., when thickness of the sample is much bigger
than the individual participle see, e.g., Refs. 4 and 812.
Method 5 see, e.g., Ref. 9: sometimes, also in the small h
range, FKM has values much larger than zero.9 Therefore, Eg
was determined as the intersection point between lines that
extrapolate FKM values in the small h range and at the linear
absorption edge Fig. 3b in the supplementary material5.
Method 6 see, e.g., Refs. 4 and 8: the Eg was determined as
the energy coordinate of the point on the low energy side of
FKM curve at which the linear increase starts Fig. 3c in the
supplementary material5. However, this method is very
ambiguousit is difficult to determine uncertainty of Eg
value. Method 7 see, e.g., Ref. 10: The Eg was evaluated as
the intersection point between lines that extrapolate FKM values in the high h range and at the linear absorption edge
Fig. 3d in the supplementary material5.
Method 8 see, e.g., Ref. 10: some more information
about Eg can be obtained by McLean analysis of the absorption edge. The threshold of continuous optical absorption can
be explained by the quantum mechanical perturbation theory
of electron-photon interaction. For different transition
mechanisms, it has been proved that along the sharp absorption edge, the energy of incident photons, and Eg of a semiconductor can have, i.e., such simple relations as

n = A1h Eg

3a

hn = A2h Eg,

3b

or
where index n depends on the interband transition mechanism. The values of n equal 2 or 2/3 imply a direct allowed
or direct forbidden transition, respectively. If n = 1 / 2 or
n = 1 / 3, the interband transition is indirect allowed or indirect
forbidden, respectively. Generally the best fitting for differ-

No. of the
method

2.234

2.111

2.08

11

2.05

2.04

1.981

1.971

1.921

1.901

1.821

1.811

12

1.761

10

Base of the method of determining


Eg using the DRS
The intersection of the straight-line extrapolation
below and above the high photon energy knee
of the FKM curve
The intersection of the straight-line extrapolation
below and above the high photon energy knee
of the curve of diffuse reflectance
Maximum of numerical derivative of
KubelkaMunk reemission function
The onset of linear increase in reflectance
with decreasing photon energy h
The onset of linear increase in FKM function
with increasing photon energy
The intersection of the straight-line extrapolations
below and above the small photon energy knee
of the FKM curve;
The intersection point between the photon energy
axis and the line extrapolated from the linear
portion of the absorption edge in a plot
of FKM function
The intersection point between the line that
extrapolates A = logRd values in the small
h range and the line extrapolated from the
linear portion of the absorption edge in the
A vs h eV plot
The intersection of the straight-line extrapolations
below and above the small photon energy knee
of the FKM1/3 curve
The intersection point between the h axis and
the line extrapolated from the linear portion
of the absorption edge in a plot of
FKM1/3 function
Multi absorption fitting of the KubelkaMunk
reemission function
The intersection point between the h axis
and the line extrapolated from the linear portion
of the absorption edge in a plot of
lnRmax Rmin / Rd Rminn function n = 1 / 3

ent n was looked for to determine the absorption mechanism


in the investigated material or the absorption mechanism
cited in literature was taken into account. Since indirect forbidden band gap exists in SbSI single crystals,13 in this paper
FKM1/3 values were plotted versus h Fig. 4a in the supplementary material5. The intersection of the fitting straight
line with h axis gave the value of Eg. Method 9 see, e.g.,
Ref. 11: the Eg was determined by the intersection of the
straight-line extrapolations below and above the small h
knee of FKM1/3 curve Fig. 4b in the supplementary
material5.
Method 10 see, e.g., Ref. 10: the Eg was obtained
from the intersection point between h axis and the line extrapolated from the linear portion of the absorption edge
Fig. 5 in the supplementary material5 in a plot of
lnRmax Rmin / Rd Rminn function where Rmax and Rmin
are maximum and minimum values of diffusive reflectance,
Rd is the reflectance at a given h, and n = 1 / 3 for indirect
forbidden transition.
Method 11 see, e.g., Ref. 12: the value of Eg was obtained from the maximum of a numerical derivative of FKM
with respect to h Fig. 6 in the supplementary material5.
The values of Eg of SbSI nanowires derived using the
presented above DRS methods are in a very large range
from 2.23 to 1.76 eV Table I. It must be underlined that
only methods 8, 9, and 10 are based on theoretical dependences that include explicitly Eg. Methods 1, 2, and 7 are

046107-3

Rev. Sci. Instrum. 80, 046107 2009

Nowak, Kauch, and Szperlich

decidedly incorrect because Eg must be near the small energy


knee of the reflectance spectra as it is in the case of the
transmittance one.
It must also be noted that the values of Eg determined
from the intersections of h axis and the extrapolations of
some functions i.e., in methods 8 and 10, can be influenced
by the absolute value of the appropriate function in the small
h range, e.g., due to strong free carriers absorption or high
value of refractive index of the investigated material. As a
result, too small a value of the determined Eg can be obtained. In order to avoid it, Eg13 = 1.801 eV was calculated
from the intersection of the straight-line extrapolations below
and above the small h knee of the lnRmax Rmin / Rd
Rminn curve Fig. 5 in the supplementary material5.
However, one can see that a more reliable method of investigations is needed.
Since not only various mechanisms of absorption of
electromagnetic radiation can be observed in semiconductors, nevertheless some of them can coexist in the same
spectral range. Therefore another way for fitting the spectral
dependence of the KubelkaMunk function has been suggested. In this method, the following least square function
has been minimized as follows:
n

2 = FKMhi B jhi
i=1

where i represents photons of different energy, j describes


various mechanisms of light absorption, and B is the proportionality factor. Some 32 of the known mechanisms of absorption in semiconductors have been considered.
Figure 1 presents the spectrum of FKM of the investigated SbSI nanowires and the least square fitted theoretical
dependence appropriate for the sum of indirect forbidden absorption without excitons and phonon statistics 1, Urbach
ruled absorption 2, and constant absorption term
3 = A0 see reference cited in Ref. 3

1 = A60h EgIf3

2 = AU exp

h
,
EU

for h EgIf ,

5a
5b

where EgIf represents the indirect forbidden energy gap, EU is


the Urbach energy, and A60 and AU are constant parameters.
The constant absorption term A0 is an attenuation coefficient
that is considered as the sum of the scattering and absorption
independent of h near the absorption edge.
The fitting presented in Fig. 1 is rather good. Values of
the fitted parameters are given in Table II.
It is clear to see that the best approach to determine Eg
using DRS is to perform the multiabsorption fitting of the
KubelkaMunk function method 12 because it allows to
determine not only Eg and the main mechanism of absorption, but also gives information on the coexisting phenomena
Table II. The value Eg12 = EgIf = 1.811 eV of the indirect
forbidden energy band gap of SbSI nanowires, which was
determined using this method, is close to the value evaluated
for powdered SbSI single crystals EgIf = 1.7931 eV.3 It is

TABLE II. Comparison of the values of indirect forbidden energy gap


Eg12 = EgIf, the Urbach energy EU and the other parameters of SbSI nanowires with the data of powdered SbSI single crystals Ref. 3 that were
determined from the fitting of the spectrum of KubelkaMunk function.
Values obtained for
Parameter

SbSI nanowires

Powdered bulk
single crystals of SbSIa

B A60, 1 / eV3 m
EgIf, eV
B AU, 1/m
EU, eV
B A0, 1/m

45.2824
1.811
2.796 104
0.5717
0.04881

35.2027
1.7931
2.527 109
0.1035
0.014117

Reference 3.

also comparable to the bulk values of the band gap of SbSI


reported in the literature: 1.82 eV at 301 K,14 for the light
polarized in direction of the c-axis of the SbSI, i.e., in the
direction of growth of the SbSI nanowires.3 The optical energy gap for light polarized in perpendicular direction is
slightly bigger.14
The penetration depth of the incident radiation is determined by properties of the sample and its surface as well as
by the angle of incidence of the incoming photons. Therefore, it is extremely difficult to determine the path length and
absorption coefficient of light in the diffuse reflectance experiment. However, it seems to be less important for determining of Eg using the spectral dependence of the Kubelka
Munk function than in the case of optical transmittance
measurements.
This paper was partially supported by the MNiSzW
Poland under Contract No. N N507 1577 33.
J. I. Pankove, Optical Processes in Semiconductors Prentice-Hall, New
Jersey, 1971.
2
B. Philips-Invernizzi, D. Dupont, and C. Caze, Opt. Eng. 40, 1082 2001.
3
M. Nowak, P. Szperlich, . Bober, J. Szala, G. Moskal, and D. Str,
Ultrason. Sonochem. 15, 709 2008.
4
F. Jahan, M. H. Islam, and B. E. Smith, Sol. Energy Mater. Sol. Cells 37,
283 1995.
5
See EPAPS Document No. E-RSINAK-80-007904 for the complete set of
figures that present elaboration of the experimental data using different
methods. For more information on EPAPS, see http://www.aip.org/
pubservs/epaps.html.
6
H. Hanmei, M. Maosong, Y. Baojun, Z. Xuanjun, L. Qiaowei, Y. Weichao,
and Q. Yitai, J. Cryst. Growth 258, 106 2003.
7
J. Llanos, C. Mujica, and A. Buljan, J. Alloys Compd. 316, 146 2001.
8
F. Zhou, K. Kang, T. Maxisch, G. Ceder, and D. Morgan, Solid State
Commun. 132, 181 2004.
9
R. Patschke, X. Zhang, D. Singh, J. Schindler, C. R. Kannewurf, N.
Lowhorn, T. Tritt, G. S. Nolas, and M. G. Kanatzidis, Chem. Mater. 13,
613 2001.
10
G. A. Martinez-Castanon, M. G. Sanchez-Loredo, H. J. Dorantes, J. R.
Martinez-Mendoza, G. Ortega-Zarzosa, and F. Ruiz, Mater. Lett. 59, 529
2005.
11
A. Escobedo Morales, E. Sanches Mora, and U. Pal, Rev. Mex. Fis. S53,
18 2007.
12
O. Raymond, H. Villavicencio, V. Petranovskii, and J. M. Siqueiros,
Mater. Sci. Eng., A 360, 202 2003.
13
M. Nowak, P. Szperlich, A. Kidawa, M. KJpiska, P. Gorczycki, and B.
Kauch, Proc. SPIE 5136, 172 2003.
14
V. M. Fridkin, Ferroelectric Semiconductors Consultants Bureau, New
York, 1980.
1