1\
INFLUENCE OF SPUTTER POWER AND SPUTTER
PRESSURE ON OPTICAL PROPERTIES OF WOx THIN
FILMS DEPOSITED BY REACTIVE DC
MAGNETRON SPUTTERING ((
By
LWANGATI, CHARLES KIAMA
B. ED. Sc. (Hons.)
A thesis submitted in partial fulfillment of the requirements for the award of the
degree of Masters of Science, in the School of Pure and Applied
Sciences of Kenyatta University.
Department of Physics
APRIL 2008
wangau, Cnanes
Influence of sputter
power and sputter
11
Declaration
This thesis is my original work and has not been presented for award of a degree in
any other University or any other award.
All sources of the information have particularly been acknowledged by means of
references.
-----------------~----------------------------------------------
Wangati, Charles Kiama
We confirm that the candidate, under our supervision, carried out the work
reported in this thesis ..
---------------~-------------------------------------------------
DR. W. K. NJOROGE
DEPARTMENT OF PHYSICS,
KENYA TTA UNIVERSITY,
P. O. BOX 43844,
NAIROBI -00100 GPO
KENYA.
-----------------------~---------------------------------------------------
PROF. 1. OKUMU,
DEPARTMENT OF PHYSICS,
KENY ATTA UNIVERSITY,
P. O. BOX 43844, -
NAIROBI -00100 GPO
KENYA.
111
Dedication
This thesis is dedicated to my family: wife, Waithira and children; Ngima, Wangati and
Wanjiru.
IV
Acknowledgement
I am profoundly grateful to my research supervisors Dr. Njoroge and Prof. Okumu whom
I have constantly consulted throughout this research work. I acknowledge their valuable
scientific support through guidance, corrections and suggestions. Most of this work was
done in the physics laboratory, Department of Physics, Kenyatta University. I would like
to thank all those who work in these laboratories for their assistant and in particular the
chief technician Mr. Simon Njuguna, who introduced me to the working of the Edwards
Auto 306 Sputtering Machine. I also want to acknowledge remarkable scientific
knowledge which I received from my colleagues especially on working with Spectro 320
Optical Analyzer and operation of Scout- 98 application software for spectral simulations.
I appreciate the university assistance in sending my samples to Japan for x-ray diffraction
(XRD) and atomic force microcopy (AFM) measurements.
I am grateful to Dr. S. Venkataraj of Advanced Material Laboratory, National Institute
for Material Science, Ibaraki, Japan for his assistant in structural measurement that is, the
XRD and AFM on the tungsten oxide thin films.
I'
I am highly grateful to my family for their support and encouragement during this period
of study. I am thankful to Almighty God for giving me the energy, opportunity and the
means to pursue this study .To Him I ascribe all the praise and honour
vAbstract
Tungsten oxide thin films playa major role in electrochromism and it's the most
investigated electrochromic material. Studies have shown that structural, electrical and
optical properties of tungsten oxide thin films depend on deposition conditions and
preparation techniques. In this work, the optical properties of tungsten oxide thin films
prepared in reactive de magnetron sputtering of tungsten target with argon in oxygen
atmosphere have been studied. The optical properties of films prepared at different
sputtering power from 300 watts to 400 watts and sputtering pressure in the range (0.65
- 0.90) Pa were investigated through the transmittance spectra recorded by optical
spectroscopy measurements in the wavelength range 300-800 nm. The experimental
curves of transmittance are reproduced by simulations to determine the band gap energy,
refractive index, extinction coefficient, and film thickness of tungsten oxide films. From
these studies the effects of increasing sputtering power from 300 watts to 400 watts was
found to; decrease optical band gap from 3.16 eV to 2.97 eV while refractive index
increased from 2.05 to 2.38 respectively. In addition, the deposition rate increased while
extinction coefficient decreased with sputtering power. Increasing sputtering pressure
from 0.65 Pa to 0.90 Pa resulted to decrease in band gap energy from 2.94 eV to 2.8geV
while refractive index decreased from 2.43 to 2.08 respectively. The surface structure of
the films is smooth as revealed by atomic force microscopy (AFM) measurements.
Further, the x-ray diffraction measurement indicated that tungsten oxide thin films
deposited are amorphous.
VI
Symbols
A Total absorptance
Band gap energy
h Planck constant
N Complex refractive index
nilj Real part of refractive index,(in media i/j)
Intensity reflectance from the front surface (f) or back
surface (b) for s or p polarized light
i jr s,p Fresnel coefficient for reflectance at the boundary between
layer i and j for polarized light.
t:bT s-p Intensity transmittance from the front surface (f) or back
surface (b) for s and p polarized light
ti j s,p Fresnel coefficient for transmittance at the boundary
between layer i and j for polarized light.
Greek
Wavelength phase change
Complex dielectric function
Conductivity
v Frequency
x Susceptibility
a Absorption coefficient
Permeability of free space
Permittivity of free space
AC
AFM
a-Woxide
CVD
D
DC
DOS
E
OD
JDOS
RF
UV
VIS
XRD
Vll
Acronyms and abbreviations
Alternating Current
Atomic Force Microscopy
Amorphous tungsten oxide
Chemical Vapor Deposition
Electric displacement
Direct Current
Electron Density of States
Electric field
Optical Density
Joint Density of states
Radio Frequency
Ultra Violet
Visible
X-Ray Diffraction
Vlll
Table of Contents
Content Page
Title
Declaration
Dedication
Acknowledgement
Abstract
Symbols used in this work
Acronyms and Abbreviations
Table of Content
Lists of figures
List of Table
(i)
(ii)
(iii)
(iv)
(v)
(vi)
(vii)
(viii)
(xi)
(xiii)
CHAPTER ONE
INTRODUCTION
1.1
1.2
1.3
1.3.1
1.3.2
Background to the study
Statement of the problem
Objectives of the research project
Main Objective
Specific Objectives
Rationale of the research project
1
5
6
6
6
61.4
CHAPTER TWO
2.1
2.2
2.3
2.4
LITERATURE REVIEW
Introduction
Density and deposition rate
Sub-stoichiometric phases of tungsten oxide
Optical band gap and Refractive index
7
8
8
9
CHAPTER THREE
THEORY OF OPTICAL PROPERTIES
3.1
3.2
3.3
Introduction
Optical properties of thin films
electromagnetic radiation in Polarization states
12
12
19
IX
3.4 Optical absorption mechanism 25
3.5 Density of states (DOS) 25
3.5.1 Band tails (Urbach Tails) 26
3.6 Microscopic models 26
3.6.1 OJL Interband transition model for amorphous material 27
3.7 Theory of X-ray diffraction measurement 32
3.8 Theory of Atomic Force Microscopy (AFM) 35
3.9 Theory of sputter machine 36
3.9.1 Operation of Edwards Auto 306 sputter machine 36
3.9.2 Sputtering principle 38
3.9.3 The Target 39
CHAPTER FOUR
EXPERIMENTAL PROCEDURES
4.1 Introduction 40
4.2 Deposition of tungsten oxide thin films 40
4.2.1 Substrate cleaning and film deposition arrangement 43
4.2.2 Gas flow set-up 44
4.2.3 Determination of oxygen to argon gas flow levels
during sputtering 45
4.2.4 Target characterization and power deposition 45
4.2.5 Chamber pressure deposition characteristics 46
4.3 Optical measurements 46
4.3.1 Spectrophotometry 46
4.3.2 Spectral Analysis 48
4.3.3 OIL model implementation 48
4.3.4 The Scout-98 parameters for OIL model 49
4.4 XRD measurement 50,-4.5 Atomic Force Microscopy (AFM) 50
CHAPTER FIVE
RESULTS AND DISCUSSION
5.1
5.2
5.3.1
5.3.2
Introduction
Preliminary results
Intensity spectrum for the light source
Thickness distribution along the substrate
Deposition characteristics
Reactive to sputtering gases flow level
Effect of sputtering power on deposition rates
51
51
51
52
55
55
56
5.2.1
5.2.2
5.3
5.4
5.5
5.6
6.1
6.2
7.0
5.3.3
5.4.1
5.4.1.1
5.4.1.2
5.4.1.3
5.4.1.4
5.4.2
5.4.2.1
5.4.2.2
5.4.2.3
5.4.2.4
5.5.1
5.5.2
x
Effect of sputtering pressure on deposition rates
Optical properties of tungsten oxide films
Effect of sputtering power on optical properties
Effect of sputtering power on transmittance
Effect of sputtering power on refractive index
Effect of sputtering power on optical band gap
Effect of sputtering power on extinction coefficient
Effect of sputtering pressure on optical properties
Effect of sputtering pressure transmittance
Effect of sputtering pressure refractive index
Effect of sputtering pressure optical band gap
Effect of sputtering pressure extinction coefficient
X-ray diffraction characteristics
Effect of sputtering power on XRD characteristics
Effect of sputtering pressure on XRD characteristics
Surface microstructure
CHAPTER SIX
CONCLUSIONS AND RECOMMENDATIONS
Conclusions
Recommendations
REFERENCES
58
60
60
60
64
66
67
69
69
72
75
76
78
78
79
80
82
83
85
Figure
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
3.10
4.1
4.2
4.3
4.4
4.Y'
5.1
5.2
5.3
5.4
5.5
5.6(a)
5.6(b)
5.7(a)
Xl
List of figures
Caption Page
An illustration of a wave that is absorbed (damped) in a medium 16
Symbols used in Fresnels equations and some material parameters 20
Illustration of propagation of light intensity in a thin film 22
OJL model, where parabolic bands are assumed with tail states
exponentially decaying into the band gap 27
Illustration of band structure of silicon crystal 29
Illustration of small mass corresponding to large increase
of energy with wavevector 30
Illustration of large mass corresponding to lower increase in
energy with wavevector 31
Illustration of ray's movement in incident and reflected beam 33
Diagram of Edwards auto 306 de and rfsputter coater 36
Schematic set up of sputtering mechanism 38
A picture of Edwards Auto 306 sputter machine and gas
supply set up 42
Substrate - target set up in the sputtering chamber 44
Optical transmittance measurement set up 47
Transmittance measurement configuration 47
A picture ofSpectro 320 used in spectral analysis 47
Intensity spectra for light source used 52
Position on the substrate whose thickness is recorded 53
Simulated spectrum for position 3 on the substrate 54
Thickness distribution a long the substrate 54
Deposition rate as a function of sputtering power 57
Variation of deposition rate for film with increasing and
decreasing deposition pressure 58
Deposition rate as function of sputter pressure 59
Optical 'transmittance of films as a function of wavelength
xu
for varying sputtering power 61
5.7(b) Transmittance as a function of sputtering power at 500nm 62
5.8(a) Simulated and measured spectra for films deposited at 325 watts 62
5.8(b) Simulated and measured spectra for films deposited at 375 watts 63
5.9(a) Refractive index as a function of wavelength for films deposited
at different sputtering power 64
5.9(b) Refractive index as a function of sputtering power at 400nm 65
5.10 Band gap energy as a function of sputtering power 66
5. 11(a) Extinction coefficient as a function of wavelength fro films
deposited at varying sputtering power 67
5.11(b) Extinction coefficient as a function of sputtering power at 700 nm 68
5.12(a) Optical transmittance as a function of wavelength for varying
sputtering pressure 70
5. 12(b) Transmittance as a function of sputter pressure at 500 nm 70
5. 13(a) Simulated and measured spectra for films deposited at 0.65 Pa 71
5. 13(b) Simulated and measured spectra for films deposited at 0.85 Pa 71
5. 14(a) Refractive index as a function of wavelength fro films
deposited at different sputter pressure 73
5.14(b) Refractive index as a function of sputter pressure at 400 nm 73
5.15 Variation of band gap energy with sputtering pressure 75
5. 16(a) Extinction coefficient as a function of wavelength at
I' varying sputter pressure 76
5.16(b) Extinction coefficient as a function of sputter pressure at 600 nm 77
5.17 XRD spectra for films deposited at different sputtering power 78
5.18 XRD spectra for films deposited at 0.75 Pa and 0.85 Pa 79
5.19 Atomic force micrograph for a sample of film surface 81
Tables
3.1
4.1
4.2
Xlll
List of Tables
Page
Properties of tungsten metal (target)
Deposition condition at varying sputter power
Deposition condition at varying sputter pressure
39
41
42
1CHAPTER ONE
INTRODUCTION
1.1 Background to the study
Thin films are applied to the solution of many different problems for a very broad range of
disciplines. These include optical applications (Dobrowolski, 1994; Rempe et al., 1993;
Dobrowolski and Verky, 1993), energy related applications such as in solar cells designs (Wanlass,
1992), architectural and automotive design (Scrosati, 1993), communication technology (Lee,
1985), information display devices (Wei et al., 2001; Lampert, 2003) and computer application
(Dobrowolski, 1995).
Optically active thin films coatings can be useful for different types of devices, which can
reversibly alter their optical properties as a function of changes in external conditions such as
temperature (thermochromic), intensity of irradiation (photo chromic) or applied external potential
(electrochromic) (Bange et al., 1995). Thermochromism has been exhibited in compounds
containing metals such as CU2Hg4 and Ag2Hg4, in addition to the most commonly investigated
oxides of vanadium (Khan and Granqvist, 1989).
Photochromism is found in silver complex oxides thin films such as A~ V04 and A~P04 within
the wavelength range for visible and infrared light (Hirono and Yamada, 1986). It is also observed
in amorphous transition metal oxide films known to be electrochromic. For example in W03 films,
a blue coloration is reported when the films are irradiated with light whose photon energy is greater
than 3.4 eV for several hours (Amoldussen, 1976). From technological point of view, thin
2electrochromic films are the best investigated and they are the most promising candidates for
various applications (Bange et al., 1995).
Tungsten oxide thin films have been investigated widely in companson to other inorganic
electrochromic material such as molybdenum oxide (Mo03), niobium oxide (Nb20S), vanadium
oxide (V205) and iridium hydroxide (Ir (OH)3) (Scrosati, 1993). It is reported to represent the best
inorganic electrochromic material for many technological applications, as its optical properties are
independent of the viewing angles and has an advantage of high optical contrast (Berggren, 2004).
In architectural engineering, tungsten oxide thin film represent the best working electrode in
"smart" (electrochromic) window systems (Rowley and Mortimer, 2002). Electrically controlled
electrochromic windows can moderate the room temperature through variable reflectance and
transmittance of solar and thermal infrared radiation (Raul, 1999). This results to higher energy
efficiency due to reduction in energy used for air conditioning in order to cool down buildings or
vehicles during warm seasons (Hummel and Guenther, 1995; Lee and DiBartolomeo, 2002).
In motot industry, it is possible to transform the variable transmittance window described above to
a variable reflectance mirror if a mirror is placed behind the device hence the reflected intensity can
be varied through the electrochromic device. This application can be used in the making of side
mirrors that can control the glare intensity in vehicles, therefore adding to safety and comfort of
driving in the dark (Lampert, 1998; Bange, 1999). If a mirror behind the electrochromic device for
the case of anti glare mirrors above is replaced by a flexible patterned and diffusely scattering
pigmented surface, optical information display can be achieved (Wei et al., 2001; Lampert, 2003).
3In addition, there is strong interest in using electrochromic devices for the thermal control of
satellites in orbit. Remittance modulation of surfaces and components is normally achieved by use
of bulky type radiators, which comprise of series of highly reflective vanes. The miniaturization of
satellites has led to reduced mass and volume available for large systems to handle temperature
variations onboard. Towards this, a flexible electro chromic reflectance device based on tungsten
oxide for infrared emissive control, which is necessary for thermal control of satellite in orbit, has
been reported (Devries et al., 1999; Larsson, 2004; Rowley and Mortimer, 2002).
Furthermore, there is a recent report concerning the use of electro chromic materials incorporated
into paper. In an attempt to produce "electrochromic writing papers" where an electrochromic
image is generated on paper when a stylus electrode touches it, a series of organic electro chromic
materials were dispersed in paper. Out of the materials studied, tungsten oxide yielded the best
permanent electrochromic writing medium (Monk et al., 2001). These and many other applications
make electrochromic materials highly sought after for commercial use. These varied
electrochromic properties of tungsten oxide film are in agreement with research findings that, the
optical, electrical, structural and electrochromic properties of the deposited films are sensitive to
the prevailing deposition conditions (Ohring, 1992; Granqvist, 1993).
There are many physical deposition methods available to produce thin films. These include vacuum
processes such as reactive and ion assisted evaporation (Swann, 1988) in addition to reactive and
non-reactive sputtering (Amoldussen, 1976; Kaneko et al., 1982; Deneuville and Gerard, 1978).
Others are sol-gel (Mucke et al., 1990) and spray pyrolysis (Hurditch, 1975).
4Various techniques have been used to prepare tungsten oxide thin film. Evaporation and sputtering
are the main physical vapor deposition techniques for the deposition of tungsten oxide thin films. In
the evaporation methods, the film is produced by target vaporization under heating in vacuum. The
weakness of thermal evaporation method is that the layer properties of the films formed suffer from
imperfections like low packing density, low hardness, and low resistance against environmental
attacks, which are mainly due to low thermal energy of the evaporated material (Swann, 1988).
Sputtering, which involves use of high-energy ions such as argon ions (Ar") to strike a solid and
knock off atoms from the surface overcomes most of these problems in the deposition of thin films.
The magnetic field created in a de magnetron sputtering system retains the electrons around the
target therefore increasing the rate of ionization of argon atoms and hence the sputtering yields.
Magnetron dc sputtering is a better method for quality and economical production of tungsten
oxide film (Hummel and Guenther, 1995). Although it is well known that the structure and
composition of the films are responsible for the optical characteristics, the composition of the film
is however difficult to control by evaporation and sol-gel processes. The sputtering deposition
technique is of special interest because the constituents of the reactive gas can control the
composition of the films (Yeon-Gon et al., 1999). For these reasons, reactive de magnetron
sputtering is often the technique of choice in depositing oxide films on large scale.
As stated earlier, in a given deposition technique many parameters can be set to produce film of
different characteristics. Among the deposition parameters that affect the thin films prepared by
sputtering technique include; the sputtering power, target-substrate distance, substrate temperature,
chamber pressure, reactive gas flow rate and deposition angle (angle between normal to the
5substrate surface and deposition flux) as well as annealing temperature among others (Salinga,
2002; Wendt et al., 1997).
Although tungsten oxide thin film is one of the most investigated electro chromic material, among
the challenges is the need to determine a precise understanding of the intimate relationship between
deposition parameters and film properties such that researchers can make an intelligent prediction
of the film properties from the deposition conditions. Further, research reports indicate that present
electrochromic devices takes a long time to switch between transparent and colored states (optical
response), thereby hindering the realization of their full potential, (Cummins at al., 2000; Gratzel,
2001).
1.2 Statement of the research problem
Optical constants reveals the interaction of electromagnetic waves with the material, hence accurate
knowledge of the constants are a pre-cursor to successful determination of specific technological
solutions for optically active devices. Though various studies have been done on tungsten oxide,
thin films there is need to determine other constants aimed at improving the efficiency of their
products. Hence, this study is set to investigate the optical constants for WOx films prepared by
reactive de magnetron sputtering using Edwards Auto 306 sputtering system. The specific
deposition parameters applied in this study have not been reported. Hence, from these study
specific optical constants for tungsten oxide, thin films will be introduced by varying sputtering
power above 300 watts at a constant sputtering pressure of 0.8 Pa and by varying sputtering
pressure below 0.90 Pa at a constant sputtering power of 350 watts. In both conditions, the
substrate target distance being fixed at 0.17 metres.
61.3 Objectives of the research project
1.3.1 Main Objective
The general aim is to prepare tungsten oxide thin films by reactive de magnetron sputtering and
study the influence of sputtering power and sputtering pressure on their optical properties.
1.3.2 Specific Objectives
The specific objectives of this research work are to determine the influence of:
(i) Sputtering pressure below 1.0 Pa,
(ii) Sputtering power above 300 watts on the optical properties of tungsten oxide thin film.
1.4 Rationale of the research project
The properties of thin films prepared by a given condition (set parameters) can differ greatly from
those produced by another condition. Due to the numerous applications of the thin films,
understanding their optical properties resulting from different deposition parameters is of
importance in the industries. The focus of this study is the correlation of optical properties of
tungsten oxide thin films with the deposition parameters as specified in section 1.2. This study will
I'
contribute towards greater understanding of the optical properties of tungsten oxide thin films
produced by varying the parameters of sputtering power and pressure within the set range using a
reactive de magnetron sputtering.
7CHAPTER TWO
LITERATURE REVIEW
2.1 Introduction
Tungsten oxide (W03) is an indirect band gap semiconductor and exhibits excellent electrochromic
properties (Wooten, 1972). Studies conducted during the last several decades, report that tungsten
oxide thin film seems to be the best inorganic electrochromic compound with high colorations
efficiency and hence accepted to represent the best option for the working electrode within smart
window system (Lee and DiBartolomeo, 2002). Certain applications of thin films critically depend
on their physical structure. It has been reported (Salinga, et al., 2002) that tungsten oxide films
would have to be nonporous and amorphous for effective use as gasochromic coatings in solar
control glazing.
Various authors (Kaneko et al., 1982; Berggren, 2004; Deneuville and Gerald, 1978) report the
effect of chamber pressure in de magnetron sputtering in an atmosphere of argon-oxygen mixture.
Report on this study showed that increased chamber pressure reduces the rate of deposition. In
addition films deposited under a constant operating pressure higher than 2x 10-2 torr are found to be
I"transparent and have a spectral transmittance of about 85 percent in the visible region. It was also
noted that the total pressure and relative oxygen gas flow influence the film transmittance, and
much less the reflectance in the visible wavelength region. The absorption coefficient increases for
independently decreasing pressure and oxygen to argon gas flow ratios. These results suggest that
good electrochromic properties, such as coloration efficiency, are achieved at low pressure and
intermediate or low oxygen to argon gas ratio that gives a high oxygen deficiency in the WOx
composition.
8The fact that the rate of coloring and bleaching are limited by the rate of ion transport into and out
of the metal oxide film, which can be influenced by grain morphology, and distribution of defects
(Risser and Ferris, 1990) is supported in a report by Bellac et al. (1995) and Larsson (2004) that
show infra-red reflectance increases with higher intercalation levels, and also, ion- intercalation
must extend into the bulk of the metal oxide film, in order to obtain sufficient optical absorption.
2.2 Density and deposition rate
The density of amorphous tungsten oxide is lower than for crystalline W03 in bulk form, which is
around 7.16 g/crrr'. This implies that substantial porosity and considerable defect structure can be
found in amorphous samples (Berggren, 2004). It has been reported (Kaneko et al., 1986) that the
density for sputtered amorphous tungsten oxide from a compressed W03 powder is about 6.1g/cm3
and it is found to increase to 6.7g/cm3 with decreasing sputter pressure (Miyake et al., 1983).
Sputtering rate for deposition of W03-z in argon and oxygen plasma shows that the rate goes down
and approaches a constant value of about 0.1 nmIsec with increasing pressure. The decrease in the
sputter rate with increasing sputter pressure is associated to the oxidation of the target (Kitao et al.,
1992). Higher sputtering rates can be accomplished by increasing the power density on the target.
The highest sputter rate in the literature is 3.4 nmIs, which was reported for reactive magnetron
sputtering at different oxygen contents for power density of 34 W/cm2 (Hichwa et al., 1987).
2.3 Sub stoichiometric phases of tungsten oxide
The sub stoichiometric phases in crystalline tungsten oxides have a tendency to form organized
phases in the series Wm03m-1 and W03m-2 (m=l, 2, 3) but there are reports of slightly different
9phases like W 18049 (Berak and Sienka, 1970). It is evident that sub-stoichiometry is connected
with the occurrence of tungsten atom in the oxidation states below 6+ as indicated by stoichiometry
deviation probe using x- ray photoelectron spectroscopy (XPS) (Deneuville and Gerard, 1978).
Films that are definitely crystalline in as- deposited state can be obtained by sputtering on a
substrate at high enough temperature (Goldner et al., 1983). At different compositions in thin films
the color and color changes are much weaker than in the bulk form. This is attributed to much
smaller thickness and hence smaller absorption. Sputtered amorphous WOx films have been
found to be transparent when x > 2.6, blue when 2.5 < x > 2.6 and metallic like when x < 2.5. As-
sputtered films can show different optical properties depending on the deposition conditions
(Deneuville and Gerard, 1978). Their data is from films by rf sputtering from tungsten as well as
W03 target in argon and oxygen mixtures while the oxygen non-stoichiometry is through
Rutherford Back Scattering spectroscopy (RBS). The electrochromism is found to deviate from
ideal tungsten oxide across the full 2.0 >hro > 0.5 eV range when x = 0.40 in W03-x. For large non-
stoichiometry (x > 0.4), the absorption goes up. Tungsten oxide (WOx) films fabricated by reactive
magnetron sputtering at room temperature were found to be amorphous and denser than those
resulting from reactive thermal evaporation. This is revealed by x- ray photoelectron spectroscopy,
x-ray diffraction and density measurements (Ong et al., 2001).
2.4 Optical band gap and refractive index
The optical band gap in reactively sputtered amorphous tungsten oxide films from a tungsten
target, have been reported to be around 3.1 eV (Daniel et al., 1988) and the refractive index around
2.3 - 2.4 for good electrochromic films while poor electrochromic films (with higher oxygen gas
flow) shows a slightly smaller band gap (Kaneko et al., 1982). The band gap has therefore been
10
found to decrease when oxygen gas flow increases but the refractive index is found to be about
the same. In addition another study of the refractive index in sputtered tungsten oxide films showed
an index that was close to 1.9 (Huang et a/., 2004). The optical band gap has also been determined
to be 3.25 eV for amorphous tungsten oxide sputtered films (Green, 1990).
Quantitative values ofEg can be derived by applying a relation of the type ah v = (h v - E g J in
the spectral range where the absorption is strong and the exponent n depends on the kind of optical
transitions that prevail while the thermal broadening is neglected. For crystalline semiconductors, n
is Yz, 3/2, 2, and 3 when the transitions are direct allowed, direct forbidden, indirect allowed, and
indirect forbidden respectively (Wooten, 1972). In the case of an amorphous semiconductor n
should be 2 as a result of the non-conservation of wave vectors (Tauc, 1974). For sputter
deposition, most studies have reported 3.0 < Eg< 3.4 eV with a tendency that high pressure of
sputter gas and a high oxygen admixture in reactive sputtering give a low Eg (Kaneko et al., 1982;
Kaneko et a/.,1988; Kitao et a/.,1992; Miyake et al., 1983). Green (1990) reported refractive index
for films made by reactive de magnetron sputtering where his data indicate a refractive index of
about 2 with a tendency to decrease with wavelength.
From the fact that the optical response of dielectric films can be influenced by grain morphology
and presence and distribution of defects (Risser and Ferris, 1990) which depends on depositing
techniques and prevailing parameters there is room for further investigation. Tungsten oxide having
been accepted as a leading electrochromic material becomes the best candidate for this
investigation. This study is therefore set to investigate the influence of sputter pressure less than 1.0
Pa and sputtering power in the above 300 watts on the optical properties of tungsten oxide thin
films aimed at developing a precise understanding of the intimate relationship between deposition
\
11
parameters and film properties. This information will enhance the technologists' ability to predict
the film or product performance from knowledge of deposition conditions.
12
CHAPTER THREE
THEORETICAL BACKGROUND
3.1 Introduction
The study of optical constants of materials is important, as it gives information related to material
atomic structure and physical properties. In addition, accurate knowledge of optical constants over
wide ranges of wavelength is necessary for materials used in optical fibers and reflecting coatings.
This chapter highlights the theory of main optical properties of thin films, theory of OJL model
used in spectra simulation and that of structural analytical methods; x-ray diffraction (XRD) and
surface topology through atomic force microscopy (AFM) involved in this study.
3.2 Optical properties of thin films
The unifying concept that entrances all optical properties is the interaction of electromagnetic
radiation with the electrons of the material. On this basis, optical properties are interpretable from
what is known of the electronic structure and how it is affected by atomic structure, bonding,
impurities and defects (Ohring, 1992; Theiss, 2004). Equations 3.1 to 3.22 highlight the
interrelationship between optical constants and the other optical properties. The velocity of
electromagnetic wave through a solid is given by the frequency dependent complex refractive index
N= n- ik (3.1)
where the real part n, is related to the velocity, and k, the extinction coefficient is related to the
decay, or the damping of the oscillation amplitude of the incident electric field.
The electric field E of a plane wave of frequency f propagating through a solid with v.elocity v in a
direction defined by x is described as
\
E = Eo exp{ 2mf(t - ~)}v (3.2)
13
where Eo is the incident electric field vector and
27rif(t-!...-)
v
is the displacement at time t after a disturbance, created by the electric field at a point situated at
'x' along the line of propagation. From Maxwell's equations on electromagnetic theory, the speed
of light in a vacuum 'c' is related to the permittivity of free space (£0) and the permeability of free
space( J.1o) by the equation.
1 (3.3)C=---l
The velocity of propagation through the solid of complex refractive index N is related to the speed
of light in a vacuum 'c', by
Cv =-N
(3.4)
Then by substituting N from equation (3.1) equation (3.5) is obtained
1 n ik:-=--- (3.5)
v C C
Therefore substituting lIv into equation (3.2) produces
E = Eo exp(2mft) exp( - 2mjx n) exp( -27ifk x)
C C
(3.6)
where the last term is a measure of the damping factor incorporating the extinction coefficient (Ie).
Since the power (P) or the intensity of an incident wave through a solid is the conductivity (0-) of
14
the solid multiplied by the square of the electric field vector, power (P) can be expressed as in
equation (3.7)
P = CTE2 (3.7)
then using the damping factor term the fraction of the incident power that has been propagated
from position '0' to a distance x through the material of conductivity (J" is given by
_p_(x_) = _ CTE2(X) = exp{_-_47ifkx~}
p(o) CTE2(O) c (3.8)
from which the absorption coefficient (a) can be expressed in terms of the extinction coefficient (k)
as
tvifka=-c (39)
Since the velocity oflight in a vacuum (c) is given as,
c=fo. (3.10)
Then by substituting c in equation (3.9) a can be expressed as
41rka=-
A
(3.11)
Optical measurements on matter using electromagnetic radiation are easiest performed by detecting
the outgoing radiation intensities from the sample. These intensities are the reflectance and the
transmittance. They can be diffuse where imperfections on the surface are greater than the
15
wavelength of the incident radiation or specular if the imperfection on the surface is less than the
wavelength of the incident wave.
All these properties are often measured in fractions of the incoming intensity and depend on the
wavelength (A), or alternatively the frequency (m), then intensity ratio contribution of the
reflectance (R), transmittance (1), scattered (S) and absorptance (A) at each wavelength are,
according to energy conservation, equal to unity:
R(A) + A(A) + T(A) +S(A) = 1 (3.12)
where
R(l)=IIIIo, ra )=/yIIo, A(l )=IAlIo, S(JJ =Is 110 (3.13)
10 [W/m2] is here the total incident intensity while IR, Ir, Is and IA are the total intensities of the
reflected, transmitted, scattered and absorbed radiation respectively. The absorptance in relation
3.13 also gives the spectral emittance. When the medium is in thermal equilibrium then the fraction
of absorptance and spectral emittance at a given wavelength are equal.
Figure 3.1 illustrate the damping of the wave amplitude when the electromagnetic wave radiation
propagates through an absorbing material. Equation (3.14) shows the relationship between the
incident intensity, the fraction of intensity that is reflected on the surface and the intensity that
enters the medium.
16
Medium
Incident wave
Transmitted wave
Air Air
x.__ .~
Figure 3.1: An illustration of a wave that is absorbed (damped) in a medium. The dashed line is a
guide to show the exponential behavior (Born and Wolf, 1983)
I 0 = RIo + I T ,0 (3.14)
In this the equation R is the surface reflectance, and Ir. 0 is the incident intensity corrected for the
surface reflectance (at x=O). In the diagram x (m) is the distance along the path of the wave in the
medium (Born and Wolf, 1983). The absorption in a layer of the material is also described by the
dimension unit ad (where x=d, i.e. the film thickness), that is called the optical density (OD).
Equation Ll S shows the expression of the surface reflectance for light passing through a medium
of refractive index no impinging normally on a transparent film of index n,
(3.15)
If the film is absorbing with index of absorption k) then surface reflectance is expressed as in
equation (3.16).
R = (n) -no)2 +k.2
(n. + no)2 + k.2 (3.16)
17
The excellent transmission of dielectric materials in the visible region of the spectrum is terminated
at short wavelengths with the onset of the ultraviolet absorption edge. The critical radiation
wavelength ').at which this occurs is given by the equation (3.17).
~ he ~ 1.24 V
/L = -or/L = --e
Eg Eg
(3.17)
where h is the Planck constant and Eg is the band gap energy . The critical wavelengths
physically correspond to electronic transitions from the filled valence band level across the energy
gap (Eg) to the unfilled conduction -band state.
The complex dielectric constant e can be introduced as
(3.1Sa)
Where E\ is the real part and &2 is the imaginary part.
since (3.1Sb)
,.
then substituting N in equation (3 .ISb) results to an expression of E as given in equation (3.ISc)
& = n2-2nik-e (3.lSc)
such that the real part (&1) can be expressed as in equation (3.19a) while the imaginary parte &2) is
expressed in equation (3.19b)
(3. 19a)
18
(3.19b)
From equation (3. 19a), for weakly absorbing films, (k is very small) then n can be expressed as
n=F: (3.20a)
and from equation (3 .19b)
k=5..
2n (3.20b)
The microscopic models usually enable us to calculate 62 rather than N and the determined value
G2 is used to generate a eI values using the Kramers- Kronig relations. Measurable optical
properties can then be obtained by converting 6) and 62 to n and k.
Substituting k = ~ in equation (3.19a), then n is expressed in terms of real and imaginary2n
dielectric function as shown in equation (3.21).
(3.21)
19
From the expression n = E 2 ,w hen n is substituted in the equation (3 .19a) then k can be2k
expressed as shown in equation (3.22).
(3.22)
The equations (3.21) and (3.22) above show that, from complex dielectric function (E), the values
of n and k for the material can be obtained. Hence, these equations form the basis of theoretical
determination of optical constants.
3.3 Electromagnetic radiation in polarization states
When the film thickness is comparable or less than the wavelength, interference effects may occur,
the electromagnetic radiation needs to be separated in s- and p- polarized light.
If an electromagnetic radiation propagates in medium i with refractive index n, at an angle 8j with
respect to a line perpendicular to an interface (figure 3.2) then radiation will change its propagation
direction, after the interface, to an angle, 8j according to Snell's law:
n i sm () i = n j sm () j (3.23)
20
Definition of symbols in the Frenel's equations and some material parameters are shown in figure
3.2.
J
n, ri,js,p
1
Figure 3.2: Symbol used in Fresnels equations and some material parameters.
The refractive indices ni and nj of the media can be characterized by the speed of light cij in the
media or the complex dielectric function with reference to media (i and j) as Gij, and magnetic
I'permeability, f.1ij , by relation (in the case of medium i):
n cc = (e i f.i i )+- (3.24)
Further, electromagnetic radiation can be polarized in s- and p- polarization states. The electric
field vectors that are normal to the plane of incidence (the plane spanned by the incident, reflected
and transmitted beams) denote the s polarization state while the electric field vectors that are in the
plane of incidence denote the p polarization state. The amplitude ratios for the reflected (r) and the
21
transmitted (t) radiation including their polarization states (s and p) have been derived in a slightly
less general form, than given below, by Fresnel in 1823 (Born and Wolf, 1983).
ij nj cos OJ-njcos OJrs nj cos () j + n j cos {)j
ij n j cos {)j-njcos {)jrp n j cos () j + n j cos {)j
tij 2 n i cos ()i
s n i cos ()i + n j cos ()j
t ij 2 n j cos ();
p n j cos {)j+nj cos {)j
(3.25)
(3.26)
(3.27)
(3.28)
An expansion of this discussion can be made to include a thin homogenous film on a substrate as
shown in figure 3.3. The numbers in the figure denote the environment 1, the film 2, with thickness
d and the substrate 3 while f denotes light incident from the front side and b denotes light incident
from the substrate. In this description all layers are considered as non-magnetic and the speed of
light in the environment is approximately the same as in vacuum i.e. the refractive index is equal to
1.
22
Figure 3.3 :Illustration of propagation of light intensity in a thin film 2 of thickness d on a substrate
3·OBerggren,2004)
R2 denotes reflectance from the interfaces 1-2 and 2-3 of the film where f is reflectance from the
front side and b from the backside. T2f denotes transmittance from 1 to 3 while T2 b denotes the
transmittance in the opposite direction. R3 and T3 is the reflectance and transmittance from and
through-the backside of the substrate respectively. Multiple reflections in layer 2 are not illustrated
in the figure. The amplitudes can, in this case be considered as an addition of reflected amplitudes
since the light can be reflected several times between the two interfaces of the film 2. The resulting
amplitudes are now also functions of the film thickness and the wavelength since the phase change
of the traversing light beam in the film has to be considered.
The phase change J of a beam with wavelength A traveling through the film can be expressed as
23
(3.29)
The amplitude reflectance (r2) and transmittance (t2) for the film on a semi-infinite substrate are
expressed as follows
fr 2 s
r 12 + r 23 e 2 i iss s (3.30)1 + r 12 r 23 e 2 i iss s
ft 2 s t
12 t 23 i 8sse
(3.31)
In these equations, the same relations also hold for p polarization (when s is replaced by p) and
have the same form as for light incident from the backside/substrate when indices 12 and 23 are
replaced by 32 and 21 respectively.
The reflected and transmitted intensities written here by capital letters are proportional to the
squared amplitudes. The light intensities, in the substrate, when the influence of the back surface of
the substrate is not included, are given by the following equations.
(3.32a)
7'1 _I f 12 n3COSB3J.2S - t2S ~ cosB) (3.32b)
(3.32c)
24
For a transparent substrate, multiple reflections in the substrate have also to be taken into account.
The intensity of the reflectance and the transmittance at the back surface of the substrate are
R 3 S , P I 31 12rS,p (3.33a)
(3.33b)
T 3 p (3.33c)
When the effect of a non-absorbing substrate IS incorporated the final expressions for the
reflectance and the transmittance are;
(3.34)
(3.35)
Analogous equations can be obtained for p polarized light.
25
3.4 Optical absorption mechanism
Optical absorption is the excitation of electrons from one state to another. The extended states
describes the collective electron band in the infinite undistorted solid and localized states describe
the electron bands caused by localized point defects, impurity lattice atoms or other imperfections.
The structural bonding between the neighbors determines the optical properties such as absorption
and transmission of the amorphous material. An arrangement is considered disordered when it is
not possible to explain its characteristics with an infinite lattice with ideal long-range order and
with addition of perturbation theory including the dynamic and static perturbations. There are
different types of disorder in solids where each type gives rise to different properties.
3.5 Density of States (DOS)
In amorphous semiconductors, local defects such as dangling bonds can cause impurity bands or
levels to appear in the gap. Impurities, defects and non-stoichiometric asymmetry of the lattice
cause an extension of the band edge of the conduction and valence band. These levels give the
DOS a characteristic structure. The properties of amorphous semiconductors are largely determined
by such states. High acceptor concentrations (negative ion doping) will give energy levels close to
the top of the valence band while high donors concentrations (positive ion concentrations) will give
energy levels close to the bottom of the conduction band. The Fermi level, Ef, where the
occupation probability of an electron is one half, will adjust itself, towards the conduction band for
increased donor doping and towards the valence band for increased acceptor doping, as the
concentrations are changed.
26
3.5.1 Band tails (Urbach Tails)
Localized states deep in the band tail are also small-polaric. That is, when occupied, these states,
near the band edge, are both severely localized (small) and accompanied by significant
displacements of surrounding atoms (polaronic). Since these small-polaronic states have their
origin in the band tail their density is quite low. The separation between the band tail and small
polaron band states are much larger than the usual interatomic separation of the small polaron. As a
result, the overlap between the small-polaric band tail states may not be large enough to produce
adiabatic hopping motion. That is, the small- polaron jump rate and the mobility within the band
tail are smaller than for the usual small- polarons that are localized in a band. (Berggren, 2004)
3.6 Microscopic models
Optical spectroscopy is a non-destructive and standard method to characterize a layer or layers
system. It involves the determination of microscopic quantities such as resonance frequencies of
oscillating atoms, impurity concentrations; or thin film thicknesses from macroscopic experiments.
The complex dielectric function e or its square root, the complex refractive index is the key
property of the material representing the connection between the dielectric displacement D and the
electric field vector E. In addition, materials dielectric function is closely related to the
susceptibility Xas shown in the equations (3.36) and (3.37) below.
D = 80E8 (3.36)
e = 1 + X (3.37)
Therefore, the frequency dependence of the susceptibility is very characteristic for a material since
it incorporates vibrations df the electronic systems and the atomic cores as well as contribution
27
from free charge carriers. Optical spectra can be easily interpreted through simulation of the
experimental transmission or reflectance data based on a model where the model parameters are
adjusted to fit the measured data. Various models have been developed that fit to simulation of
certain material spectra. The so-called OJL model due to O'Leary et al. (1997) where expressions
for the density of states (DOS) are given for the optical transition from the valence band to the
conduction band is applied in this analysis as the ideal dielectric function model for amorphous
material.
3.6.1 OJL interband transition model for amorphous materials
In this model, parabolic bands are assumed with tail states exponentially decaying into the band
gap as sketched in figure 3.4.
Density of states N(E)
Valence band
EnergyE
~exp«E-Fc)l'fc)r ~exp( -(E-Ev)l'fv)
Figure 3.4: OJL models, where parabolic bands are assumed with tail states exponentially decaying
into the band gap
The original parameters of the OJL density of states model are the valence band energy Ev and
conduction band energy Ee, the 'damping constants' of the valence and conduction band Yv and vc,
respectively, as well as the masses of the valence and conduction band mv and me.
28
The mobility edges of valence band energy (EMV) and conduction band energy (EMc) are given in
the equations 3.38 and 339 respectively.
1
EMV =Ei, =z r»2 (3.38)
(339a)
The mobility gap in the OJL model is therefore the difference between EMC and EMV is given by
(339b)
The meaning of the term 'mass' in this context can be considered from figure 3.5, which shows the
band structure of a silicon crystal.
E 6
[eV] 5
4
3
2
1
0
-1
-2
-3
-4
-5
L (111)
29
r (100) X K
wavevector k
(110) r
Figure 3.5: Illustration of band structure of a silicon crystal
The black rectangle shows a region of one of the conduction bands, which may be described by a
I'parabola approximately. That means that the energy is proportional to the square of the wave vector
and can be expressed as in equation 3040 where a is constant that determines the curvature of the
conduction band and b is the lowest energy of the band.
(3.40)
This relation between wavevector k and energy E is similar to the one between momentum p and
energy of a classical particle of mass m moving with velocity v. Since momentum p is given as
\
30
p=mv (3.41a)
then (3.41b)
therefore E can be expressed as in equation (3.41 (c))
(3.41c)
The coefficient a that determines the curvature of the conduction band can be associated with the
inverse mass of a classical particle, which is said to represent an electron in that conduction band:
Ii Iim=-=>a=-2a 2m
(3.42)
A small mass corresponds to a large increase of energy with k, i.e. large slopes in the band
structure:
10
~ 8'--'
~
Iij
6~
4
2
0
-5 -4 -3 -2 -1 0 1 2 3 4 5
k [11m]
Figure 3.6: Illustration of small mass corresponding to large increase of energy with k
31
On the other hand, a large mass stands for a flat band as shown in figure 3.7.
10 ~------------------------------~
4
" '" "
,-»:
2
'" -~,-~-.--------~-
o +-~--~--~~--~~~~--~~--~
-5 -4 -3 -2 -1 0 234 5
k [11m]
Figure 3.7: Illustration of large mass lower increase in energy with k
For the optical properties, only the joint density of states (mOS) denoted here as J is important
which is a certain combination of the density of states of the valence and conduction band:
J( n L~~ = IN CtJiIill&ti(!/fbflJlll ( E)' N VoJeIf(;dflJlll (E - n UJ) dE
-.0 (3.43)
The integration collects all possible transitions from the valence band to the conduction band with
energy separation equal to the energy of the absorbed photon. It is shown by O'Leary et al.(1997)
that J can be written as a product of a function, which is independent of the masses and a pre-factor
(called M here) which contains the masses of the conduction and valence bands besides some
constant factors:
32
M 3 3V m c )+- (3.44)
The model is well suited to describe amorphous materials or crystalline phases with a high defect
density. However, the complex dielectric function of all materials in the system must be known for
reflectance or transmittance spectra to be computed. If the imaginary part of the dielectric function
is known for all fre9uencies, which is the case in a model, the real part can be constructed uniquely
employing the Kramer-Kronig relation (KKR) between a real and imaginary part of a causality
quantity. From the dielectric function obtained through simulated spectra, optical band gap,
refractive index and extinction coefficient for the deposited tungsten oxide thin films are
determined.
3.7 Theory of X-Ray Diffraction measurement
x - Ray diffraction measurement is based on Bragg's law that refers to the simple equation
nA.=2dsin9 (3.45a)
This equation explains why the cleavage faces of crystals appear to reflect X-ray beams at certain
angles of incidence (theta,8). The variable d is the distance between atomic layers in a crystal, and
the variable lambda A is the wavelength of the incident X-ray beam while n is an integer.
Figure 3.8 is an example of X-ray wave interference, commonly known as X-ray diffraction
(XRD), and it gives direct evidence for the periodic atomic structure of crystals. It is based on
Bragg's law, which explains the interference pattern of X-rays scattered by crystals and has been
developed to study the structure of all states of matter with any beam, e.g., ions, electrons,
33
neutrons, and protons, with a wavelength similar to the distance between the atomic or molecular
structures of interest
d
--~---
B
z
A B
Figure 3.8: Illustration of rays' movement in incident and reflected beam (Bragg's law)
Bragg's law can easily be derived by considering the conditions necessary to make the phases of the
beams coincide when the incident angle equals reflecting angle. The rays of the incident beam are
always in phase and parallel up to the point at which the top beam strikes the top layer at atom z.
\
34
The second beam continues to the next layer where it is scattered by atom B. The second beam
must travel the extra distance AB + BC if the two beams are to continue traveling adjacent and
parallel. This extra distance must be an integral (n) multiple of the wavelength (A.) for the phases of
the two beams to be the same:
nA..=AB+BC (3.45b)
The lower beam must travel the extra distance (AB + BC) to continue traveling parallel and
adjacent to the top beam. Recognizing d as the hypotenuse of the right triangle ABz, we can use
trigonometry to relate d and e to the distance (AB + BC). The distance AB is opposite e so,
AB = d sin(:) (3.45c)
Because AB = BC equation (3.45b) becomes
nA..=2AB (3.45d)
Substituting equation (3.45c) then equation (3.45d), becomes nA..= 2 d sinfl.
The location of the surface does not change the derivation of Bragg's law (Schields, 2004).
I'
The principle of x-ray diffraction therefore reveals a samples crystalline or amorphous nature
through characteristic interferences pattern obtained due to the phase difference between x-ray
photons scattered at each individual lattice plane (Ngaruiya, 2004). The shape and intensity of the
reflection peaks reveals necessary information for determination of the structural properties of the
sample.
35
3.8 Theory of Atomic Force Microscopy (AFM)
Atomic force microscopy uses two modes of operation that is contact and tapping modes. In
contact mode, a sharp tip of radius less than 10 nm at the end of the cantilever scans the surface of
a sample. The sharp tip is scanned over a surface with feed back mechanisms that enable the piezo-
electric scanners to maintain the tip at constant force (to obtain height information) or constant
height (to obtain force information). The force between the tip and the sample is less than 10-9 N.
An optical detection system is employed to map out the surface morphology. A quadrant photo
detector detects the reflection of the laser beam at the backside of the cantilever. The signal is used
as an input for a feedback loop, which keeps the deflection of the laser beam from the center of the
detector constant by moving the cantilever up and down with the help of a piezo transducer. The
voltage applied to the piezo is proportional to the height.
In tapping mode, the cantilever oscillates as it is driven by a piezo crystal. The drive frequency is
close to the resonance frequency of the cantilever leading to amplification of small amplitudes.
When the tip approaches the surface, damping is increased and the amplitude of the oscillation
decreases. This reduction in oscillation amplitude is used to identify and measure surface
topologies. A feedback loop adjusts the tip-sample separation to maintain constant amplitude and
force on the sample (Ngaruiya, 2004). The topological image of the surface obtained together with
the x-ray diffraction pattern would enable identification of the sample as either crystalline or
amorphous.
36
3.9Theory of sputter machine
Figure 3.9 shows the major parts of Edwards Auto 306 Sputtering machine that is used in this
investigation.
Rotary workholder drive
RF 1 W kh ldpower supp or 0 ••I I Deposition IX '\~II chamber/\ '\ VDC power SUPP1~ ~
) Maonerron IJI I 00 @1\\: /Emergency / Control
off button ~ r-. ...•....... ) ...- console......•.@
~
F'II Distribution panel r 0 . . ..
I ControlMFC Control I •••- cabinet
I Equipment Rack
~
I'
Figure 3.9: Edwards Auto 306 DC and RF Sputter Coater
3.9.1 Operation of Edwards Auto 306 sputtering machine.
The machine has two 0.075 m diameter EPM magnetron-sputtering sources that can hold the target
material, which is clamped with the clamping ring. D.c. power source is manually adjusted and
37
supplies power to the magnetron. Source shutter, which is manually controlled through an external
lever, is set above the target holder. It enables the operator to control the timing of sputter-
deposition on the substrate and shields the substrate from contamination during the initial cleaning
(pre-sputtering using argon gas).
Three tripod legs fitted to the base plate supports baffle plate, which shield the base plate from
deposition and prevent contamination and debris from entering the high pumping aperture. Base
plate provides the ports for the connection of gauge heads, gas control valves, source shutters and
other accessories. The two pressure gauges that is the penning gauge and pirani gauge measures the
chamber pressure during fine pumping and rough pumping respectively. Rotary workholder that
consists of a 0.260 m diameter of aluminum plate attached to the chamber top plate by a rotary
headthrough supports the substrate on which the film is deposited. De motor and gear assembly
continuously rotate the workholder through 3600 at a speed of 40 revolutions per minute to ensure
uniform distribution of the film on the substrate. A window on the side of the sputtering chamber
enables a physical check on the rotation of the workholder, closing and opening of the shutter as
well as the plasma colour changes in the chamber.
38
3.9.2 Sputtering principle
Figure 3.10 Illustrate a simple schematic set-up showing the sputtering mechanism.
Vacuum pump Substrate Gas inlet
.•.•.+ ••"'."., ... '"'" T •• • •+. . . " •+. . . ..•..•.++. Plasma • • +,. .. .. ,.: .~~;~:\~f~f :;'.
HV Power supply
Figure 3.10: Schematic set up of sputtering mechanism.
The processes include mounting the target firmly on the sputtering source using a clamping ring
while the substrate is attached to workholder above the target. The chamber is evacuated to prevent
I'
pollution of the film by unwanted atoms. A chemically non-reactive working gas (argon) with a
constant flow is then introduced in the chamber. When a high negative potential (cathode) is
applied to the target material above the breakdown voltage of the working gas while zero or
positive (anode) is applied to the rest of the chamber and substrate, electrons will be emitted and
accelerated from the target therefore ionizing the working gas atoms. This further amplifies the
plasma as new electrons are created under ionization. These ions created in the plasma accelerate
39
towards the target and knock off target atoms and other secondary electrons. These electrons
further ionize the working gas resulting to sustained glow discharge. The knocked out atoms are
then deposited onto the substrate as a thin film. By allowing oxygen, gas to flow constantly into the
chamber with argon a compound of the tungsten and oxygen can be formed. In this case, WOx
where x varies with deposition conditions (Berggren, 2004).
3.9.3 The target
The tables below show the specifications of tungsten metal used as target in this work. The table
highlights the main characteristics of the metal. In addition, the different possible oxidation states
of the metal are indicated.
Table 3.1: Properties of tungsten metal
Properties of tungsten Description
Physical appearance Hard, grey metallic element
Chemical symbol W
Atomic weight 183.85
Melting point 34100c
Boiling point 56600c
Specific gravity 19.3 at 20°c
Oxidation states 6,5,4,3,2
40
CHAPTER FOUR
EXPE~NTALPROCEDURES
4.1 Introduction
This chapter describes the main experimental techniques applied in this study. It gives the basics
of the sputtering chamber preparation, sputtering process, the gas flow process, and substrate
cleaning and mounting. In addition, the basics of the transmittance measurement and simulation
processes are explained.
4.2 Deposition of tungsten oxide thin film
Tungsten oxide films were prepared using Edwards Auto 306 Magnetron sputtering system
connected to external gas supply as shown in figure 4.1. The films were prepared by reactive de
method from a metallic tungsten target. The sputtering vacuum chamber was initially vented and
the cylindrical chamber walls cleaned using a cleaning detergent. All other parts housed by the
chamber were cleaned using cotton wool and laboratory alcohol. In order to ensure airtight
condition on the top and bottom boundary of the chamber, silicon grease was applied on the rubber
enclosing the chamber. The target was fixed at the base of vacuum chamber and substrate
(microscope glass slide) on the rotary workholder, which consist of a 260 mm diameter aluminum
plate attached to the chamber top plate by a rotary headthrough. The chamber was then pumped
down to base pressure of 5.0 x 10 -4 Pa. Argon gas was allowed into the chamber until the chamber
pressures reached 0.55 Pa. Sputtering power was then adjusted and fixed at the required value. The
tungsten target was then pre-sputtered by argon ions for about 10 minutes before introducing
oxygen in order to clean the surface of tungsten. A common knob controls the entry of argon and
oxygen gases' into the sputtering chamber while flow rate of each gas was set through individual
\
41
gauges. To ensure uniform deposition of the film on the substrate the d.c. motor rotates the
workholder continuously through 360°. When the system was in steady state the shutter was
opened, and deposition parameters recorded and maintained constant for the required duration of
deposition. These parameters are sputtering pressure, oxygen to argon gas flow levels, target-
substrate distance and sputtering power. All other deposition parameters are maintained constant
during the preparation of the films while varying power and pressure are given in the table 4.1 and
4.2 respectively. In the table the base pressure is the lowest pressure attained on evacuation, while
sputtering pressure refer to sum of the base pressure, argon and oxygen reactive gas pressures. The
shutter, which is manually controlled by a lever, enables the operator to control the timing of
sputter-deposition on substrate and shield the substrate from contamination during initial cleaning.
The film thickness during deposition was monitored using the quartz crystal oscillator fixed inside
the sputtering chamber. All films were allowed to reach the same thickness as displayed on the
control console of the Edwards auto 306 sputtering chamber and the time taken using a stopwatch
in order to estimate the deposition rate. In order to investigate the effect of sputtering power all
other parameters were maintained constant while sputtering power was adjusted from 300 watts to
400 watts at intervals of 25 watts for different film deposition. Similarly, the sputtering pressure
was varied from 0.65 Pa to 0.90 Pa while all other parameters were kept constant in order to
investigate the effect of sputtering pressure on films optical properties.
Table 4.1: Deposition conditions at varying sputter power
Sputtering power
Base chamber pressure
Argon gas + base chamber gas pressure
Sputtering pressure
Substrate -target distance
300W-400W
5.0 x 10-4Pa
0.55 Pa
0.80 Pa
0.17m
42
Table 4.2: Deposition conditions at varying sputter pressure
Sputtering chamber pressure
Base gas pressure
Sputtering power
Substrate -target distance
Argon gas + base gas pressure
(0.65- 0.90) Pa
5.0 x 10-4Pa
350W
0.17m
0.55 Pa
The figure below illustrates the set up of Edwards Auto 306 Sputtering machine. From the left side
of the picture; the power control section, the sputtering chamber, the workholder speed control and
quartz crystal display section, the cooling control component and external gas sources are clearly
displayed.
Figure 4.1: Picture of Edwards Auto 306 sputtering machine and gas supply set up.
43
4.2.1 Substrate cleaning and film deposition arrangement
Microscope glass slides of the dimensions 76.2 x 25.4 x 1 mm were used as substrates. The
substrate was soaked in a cleaning detergent for about 30 minutes then cleaned using cotton wool
followed by dipping in distilled water before rinsing with alcohol and allowing it to dry. The
substrate was fixed on the rotating workholder as shown (figure 4.2). The central rotating substrate
and an eccentrically positioned source (target) off the axis of rotation are as shown in figure 4.2.
The height h (distance between the substrate plane and the target level) was set at 0.17 m while R
(approximately 9cm) is dictated by the geometry of the sputtering machine. The transmittance
spectra for different positions, along the substrate, are recorded and then fitted with the simulation
curves using scout software, to determine their thickness distribution along the substrate. The
positions are marked from the end of the substrate near the axis of rotation. A clip supports the
substrate on the rotating workholder at the end A.
44
,--- ---,,
I
I
,
\
I
I
,/+-- Axis of rotation
Rotating workholder---+
Substrate arrangement
h
Target lever···································· · · ···· · · ······ · · ···· · ········r=t ==-t~Static target
Figure 4.2: Substrate -target set up in the sputtering chamber.
4.2.2 Gas flow set-up
A gas control valve regulates both oxygen and argon into the sputtering chamber since both gases
,-
enters the chamber through a common tube. However, oxygen and argon gas flow are separately
externally controlled through control gauges before joining the common tube. The gas flow rate
could not be determined accurately due to lack of flow meters. The control gauges along the
individual gas flow tubes were only capable of maintaining the gas flow constant by fixing the flow
level. Due to this limitation, the partial pressures of the gases could not be used reliably as a
variable parameter in this work. Hence, the total chamber pressure (Ar + 02) mixture was used as
the variable parameter.
45
4.2.3 Determination of oxygen to argon gas flow levels during sputtering:
Reactive sputtering processes are highly unstable. A magnetron target can very quickly switch from
metallic state to fully poisoned condition (Swann, 1988). This change unless it is prevented will
make the process unworkable and results in large variations in the sputtering rate. Possible solution
includes use of the target in fully poisoned mode. However, the compounds on the target sputter at
a much lower rate than a pure metal hence it is not desirable. A different solution is to use a
feedback control system that can very quickly adjust the reactive gas flow in response to the plasma
conditions, in order to hold the process in high rate metallic or transition mode. In order to establish
the appropriate flow levels for oxygen and argon gas, necessary to produce an oxide different flow
trial rates were done. For each flow rate the characteristics of the film formed was determined
through observation of plasma colour changes during sputtering, the film deposition rate, physical
appearance and the transmittance level of the film deposited in light.
4.2.4 Target characterization and power deposition characteristics
In order to remove oxides and contamination on the target surface, pre-sputtering was done using
argon gas alone for about 10 minutes at a chamber pressure of 0.55 Pa. The oxygen gas was
introduced and the flow of oxygen to argon gases maintained at 70:30 as determined in 4.2.3, while
the total sputtering chamber pressure was adjusted to 0.80 Pa. The sputtering power was set at 300
watts and The deposition rate was determined by dividing thickness of the film formed as recorded
by quartz crystal oscillator by time taken. Sputtering power was set at 300 watts initially then
adjusted upwards at intervals 25 watts up to 400 watt. The process was repeated as power
decreased from 400 watts to 300 watts.
46
4.2.5 Chamber pressure deposition characteristics
Having identified the gases flow levels through the control gauges that is adequate for deposition of
tungsten oxide, the characteristic behavior for different chamber pressure was monitored. The flow
level for oxygen to argon was maintained at 70:30 and sputtering power fixed at 350 watts. The
deposition rate was investigated as chamber pressure increased from 0.6 Pa to 0.90 Pa and the
process repeated as it decreased from 0.90 Pa to 0.60 Pa.
4.3 Optical measurement
4.3.1 Spectrophotometry
Figure 4.3 shows the arrangement of the varIOUS components used in optical transmittance
measurement. In this study, a Spectro 320 Optical Analyzer was used to detect and record the
transmittance spectra for various tungsten oxide thin film samples. It has a charge coupled device
detector (photomultiplier) for the range 190-880 nm. This adequately covers the ultra violet and the
visible regions of the spectrum while the second detector; an InGaAs semiconductor diode array
covers the near-infra red region 800-1700 nm. A 12V, 0.25A De power lamp (Nippon Keiki Works
(Japan) Ltd.) powered at 6.75 watts was used as source of radiations for all measurements. On
optical transmission measurement, the light radiations from the lamp were recorded at normal
incidence to the tungsten oxide thin films surface. The separation between the lamp and detector
was maintained at 20 tnm. The spectral intensity of the lamp detected without a sample between the
lamp and the detector was recorded as a reference measurement after which measurements for
other films were made. Figure 4.4 shows the configuration setup while figure 4.5 show a picture of
the Spectro 320 used in detection of transmittance spectra in this work.
47
~ Enclosed wooden box painted black inside
Transmission measurement configuration
Spectro 320 Analyzer C mputer
Power Source
Figure 4.3: Optical transmittance measurement set up.
Thin film on glass
From pow
Source
~~ -.----er
Ct. o T~
1\
~ - I'--
Light source- bulb Light detector
o Spectro 320
Figure 4.4: Transmittance measurement configuration.
L-.~L, k<~'
~ rf .
"<,., •..• ,;1
• • J"c.,,,"~~ ~
"'" .,~t
)' IIIISTRUMENTSYSTEMS
Figure 4.5: A picture ofSpectro 320 used in spectral analysis
48
4.3.2 Spectral Analysis
Response of material on applied electric fields is determined by the frequency dependent dielectric
function. Using the transmittance spectra measurement made for different films as input, a
simulation approach using Scout-98 software produces an optical spectrum (transmittance) that can
be compared directly to the measured data. From the simulated spectra, the thickness, band gap
energy and dielectric function are determined. The simulation process was also conducted for
transmittance spectra on different positions a long a substrate in order to determine the film
thickness distribution a long a substrate. OJL model by O'Leary et al. (1997) is adapted for
computer simulation as it adequately covers electronic transition between valence and conduction
bands. In addition, it describes both amorphous material and crystalline phases with high defect
density quite well. Further, OJL model gives expressions for joint density of states for optical
transitions from the valence band to the conduction band. From the expression for the density of
states, the imaginary part of the dielectric function is modeled and then the real part is calculated by
Fast Fourier Transformation to satisfy causality (Kramer- Kronig relations). In this model,
parabolic states that are characteristics of the defects in the film are assumed with tail states
decaying exponentially into the band gap. These model parameters fit, are adjusted in order to
determination the dielectric function of the sputtered tungsten oxide thin films from which the
refractive index and extinction coefficient are obtained. The relationship between dielectric
function and extinction coefficient are given in equations (3.21) and (3.22).
4.3.3 OJL model implementation
Since there is no easy way to determine the absorption coefficient of a thin film system the
absorption coefficient is not a good 'interface quantity' between theory and experiment. Instead, a
49
simulation approach is recommended which ends up in a prediction of optical spectra (reflectance
and transmittance, usually) that can be compared directly to the measured data. A model parameter
fit then adjusts the parameters of interest to the recorded spectra.
To compute reflectance or transmittance spectra the complex dielectric function of all materials in
the system must be known. Unfortunately, in case of a known joint density of states only the
imaginary part of the dielectric function can be computed, but not the real part. Fortunately, if the
imaginary part is known for all frequencies, which is the case in a model, the real part can be
constructed uniquely employing the Kramer-Kronig relation (KKR) between a real and imaginary
part of a causal quantity. To do this, however, the imaginary part must approach zero for high
frequencies, which is not the case in the O'Leary expressions. Hence, a decay of the imaginary part
is enforced by a term, which is introduced in addition to the OJL paper.
4.3.4 The SCOUT _98 parameters for the OJL model
The model applies the energy gap Eg, between valency and conduction band, gamma factor YII and
Yc for the exponential decay from the valence and conduction band respectively into the band gap;
in the case of no disorder, both 'Yv and 'Yc, are zero. However, gap energy is different from the
mobility gap, which depends on the values of the disorder parameters 'Yv and "fc. For traditional
reasons the gap energy must be specified in wavenumbers. To convert wavenumbers given in
reciprocal centimeters to eV, the wavenumbers is divided by 8065. The mass refer to collection of
all constants and the square root of the third power of the valence and conduction band masses; the
mass factor m represents the strength of the optical transition. The ratio "f v / "f c is used here
because it simplifies working with a constant ratio of valence and conduction band exponential
50
decays. This number determines the way in which the imaginary part decays to zero for high
frequencies.
4.4 X- Ray Diffraction (XRD) measurement
X-ray diffraction measures crystal structure, grain size, texture and/or residue stress of materials
and compounds through the interaction of x-ray beams and the sample. In this study, the XRD
measurement performed helped to measure the intensity of the diffracted beam. The sample was
fixed at a grazing incidence angle ro of 0.75 0 and the detector scan performed using Goniometer,
RINT2000 Vertical goniometer. The advantage of this grazing angle for incident beam is that the
effective penetration of the x-rays is reduced and the diffraction pattern is from the near surface
resulting to less attenuation of the signal to noise ratio.
4.5 Atomic Force Microscopy ( AFM)
Surface microstructure of thin film coating significantly affects important properties of the optical
scattering. Since AFM can advantageously measure these microstructure, it is used in this study to
examine the topology of the film surface and a topological image of the surface is obtained from
which the nature of the material (amorphous or crystalline) can be detected. The information
I'
obtained can substantially contribute to successful development of low-scatter optical coatings.
From scattering theories, the amount of optical scatter depends on the roughness height of a
structure and on its lateral distribution. At a randomly rough surface, many different spatial
frequencies are present and they are quantitatively expressed by the Power Spectral Density (PSD),
giving the relative strength of each roughness component of a surface microstructure as a function
of spatial frequency. AFM was performed on the surface of these films to reveal the degree of
roughness and hence determine whether the film is amorphous or crystalline.
51
CHAPTER FIVE
RESULTS AND DISCUSSION
5.1 Introduction
This chapter reports on the results of the practical work done in this research. The results of
deposition rate, transmittance, band gap energy, refractive index, and XRD as well as AFM
measurements are analyzed and discussed in line with the existing theory. In particular, this
research focuses on the influence of two parameters:
(i) Sputtering power in the range of 300 watts to 400 watts,
(ii) Sputtering pressure in the range 0.65 - 0.90 Pa,
on the optical characteristics of tungsten oxide thin films. The analysis was done in the spectral
range of 300 run to 800 nm. Report on characterization of the deposition rate in the sputtering
chamber for both varying sputtering power and sputtering pressure, flow ratio of gases (oxygen and
argon), intensity spectrum of the light used as well as distribution of the film thickness along the
substrate are also discussed.
5.2 Preliminary results
5.2.1 Infensity spectrum for the light source
From figure 5.1, the intensity spectrum of the lamp used in this study show that the detected
radiations becomes negligible for wavelength less than 300 nm. Hence, the lamp used as the source
of radiations cannot effectively be used to investigate the optical properties of the film within the
range of ultra violet spectrum despite the fact that the spectrum analyzer, Spectro 320 used in this
measurement, can detect radiation of as low as 190 nm. Although this photo spectrometer can
detect transmittance spectra, from 190 nm to 1700 nm by combining the range of the first detector
52
(190- 880) run with the second (800-1700) run, the spectrum from the different detectors do not
match. Fortunately, this can be done by selecting a position on one spectrum within the overlap
region (800-880) run at which the computer can match up the two spectra automatically by
multiplying data from one spectrum by an appropriate factor so that they can match up and overlap
at this point. However, this matching was found to give inconsistent results (through simulation)
for films deposited at a sputtering power of 400 watts and those deposited at sputtering pressure of
0.65 Pa. Hence, transmittance spectra for this study were limited to the range of 300 run to 800 run.
1.0
0.8
0.6 Lam p spectrum I
~·inc:
~ 0.4c:
0.2
o 200 400 600 800 1000 1200 1400 1600 1800
Wavelength (nm)
Figure 5.1: Intensity spectra for the light source used. The lamp spectrum range from 300 run to
1700 run.
5.2.2 Thickness distribution along the substrate
Figure 5.2 show positions on the film whose thickness was determined by simulation of
transmittance data. Position 1 indicates the region nearest the axis of rotation of the workholder on
53
which the substrate was attached. Although the length of the substrate was about 7.6 em, only
about 5.5 em can effectively be investigated as the end marked A near the edge of workholder was
covered with the clip attaching substrate to the workholder. Figure 4.2 illustrates this set up.
III 7.6cm •
(0 CJ 8 881 I I I I•EndO lcm 2cm 3.5cm 4.5cm 5.2cm
III 5.5cm •
Figure 5.2: Positions on the substrate whose thickness is plotted in figure 5.4.
Figure 5.3 show the measured and simulated transmittance spectra at position 3 as representative
spectra used to determine the thickness of the film for the other positions. The agreement between
the simulated and experimental spectra is reasonably good. This suggests that the OJL model
applied in this simulation adequately covers the transition involved in this film. The presence of
oscillations on the transmittance spectra are as a result of interference fringes that occur when the
I'
film thickness is uniform otherwise the spectra would appear as a line. The presence of these
oscillations contributes greatly to proper simulation of the film.
54
100
90
80- 70:::R0- 60Q)
0c: 50('0;:;
E 40
U)c:
('0 30~I-
20
10
0
Thickness 297 nm
--+ --Simulated spectrum
-- .•..Experimental spectrum
400 500 600 700
Wavelength (nm)
800
Figure 5.3: Simulated and measured spectra for position 3 on the substrate
Figure 5.4 shows the thickness distribution of the film a long the substrate as determined through
simulations.
310
300
290
280
,-... 270
~~ 260
'"'"Q) 250..e
u 240
~ 230
220
210
200
2
Length from
3 4 5
end of substrate (em)
Figure 5.4: Thickness distribution a long the substrate
55
From the graph (figure 5.4), the thickness along the length of the substrate varies slightly, with a
maximum at the centre of the substrate. However, from the error bars shown based on the standard
deviation there is no significant difference between the thicknesses recorded along the film ( Error
bars are small and they all overlap). This is in agreement with Bach and Krause (1997) report that,
for a central rotating substrate, the thickness variation a long the substrate should not be more than
14 nm as is the case in this work. This can be attributed to the fact that the source -to- substrate
geometry of the sputtering chamber consists of rotating workholder and an eccentrically positioned
source (target) off the axis of rotation as explained in section 4.2.1. The use of microscope slide as
a substrate ensures high transparency in the visible wavelength range while film coating add
interference effects to the surface, which changes the transmittance, reflectance and absorptance of
the glass medium therefore providing a means of relating thickness from the simulation of the
transmittance spectra as in this study.
5.3 Deposition characteristics
This section report on, the reactive to sputtering gas flow levels and deposition rate as influenced
by sputtering power and sputtering pressure.
"
5.3.1 Reactive (oxygen) to sputtering (argon) gases flow levels
The rate of sputtering in fully poisoned mode was found to be a bout 0.3 nmlmin in the sputtering
range 300 -400 watts. At this rate, more than three hours would be required to deposit a film of
reasonable thickness necessary for fitting of the transmittance spectra during simulations. In
addition, the chamber temperature was found to rise due to the prolonged sputtering time resulting
to erratic fluctuation of chamber pressure. The low sputtering rate can be associated with high
56
binding energy of the compound due to the oxide formed on the surface resulting to low sputter
yield. When the supply of reactive gas (oxygen) is very small it was found that, the deposited film
was almost metallic. After various trials with different gas flow levels, it was found that at a ratio
of 70:30 for oxygen to argon gases respectively, the plasma colour changed from blue to yellow
suggesting presence of enough oxygen mixture; the film deposition rate increased, and a
transparent tungsten oxide film was deposited. Hence, oxygen to argon gases flow levels were
fixed at 70:30 for this study.
5.3.2 Effect of sputtering power on deposition rate
From figure 5.5 shown below, the variation of deposition rate as a function of increasing (square
symbols) and decreasing (circular symbols) sputtering power show good agreement. That is, at a
given sputtering power the deposition rate was maintained almost constant with time. The
deposition rate was also found to increase with increase in sputtering power. This implies that
deposition of these films was done in a relatively stable state and the results could therefore be
reproduced (no hysteresis). The process involved monitoring the film deposition rate using quartz
crystal oscillator fixed in the Edwards Auto 306 Sputtering system in order to determine the target
I'behavior upon increasing and decreasing the sputtering power. This deposition rate is only an
estimate as crystal oscillator gives a relative thickness and not the actual. It was observed that for
sputtering power less than 300 watts the rate of deposition was very low while above 400 watts the
films formed were metallic. Hence, the sputtering power investigated in this study was set between
300 watts and 400 watts. The triangular symbols show the deposition rates as determined through
simulations from measured transmittance spectra of the tungsten oxide thin film samples deposited
at different sputtering power.
15
14,..-.. 13~.-] 1211'-'
Q) 10~~ 9~0.- 8.•....-000 7fr0 6
5
57
•
-.- Increasing Power data
-e- Decreasing Power data
-A- Simulation data
300 340 380 400320 360
Sputtering Power (Watts)
Fig.5.5: Deposition rate of tungsten oxide thin film as function sputtering power
The deposition rates as determined through simulation differ from those calculated from quartz
crystal data but both results indicate the same trend where deposition rate was found to increase
with increase in sputtering power. This can be attributed to the fact that, increase in sputtering
power results to an increase in momentum of the argon ions, which are responsible for knocking
out the tungsten atoms from the target. Hence, the sputtering yield increases (Swann, 1988)
resulting to higher rate of deposition as compared to lower sputtering power. In addition, since the
chamber pressure is maintained constant, the mean free path for the tungsten atoms at higher
sputtering power is not affected. Therefore, more of the sputtered tungsten atoms sputtered at
higher power reaches the substrate resulting to increase in the deposition rate.
58
5.3.3 Effect of sputtering pressure on deposition rate.
Figure 5.6(a) show the effect of increasing and decreasing sputtering pressure on deposition rate of
tungsten oxide thin film as monitored using Quartz crystal oscillator, fixed inside the chamber. The
values were obtained when the flow ratio of oxygen to argon gas was maintained at 70:30
respectively. It was observed that films deposited at sputtering pressures below 0.65 Pa were
metallic while above 0.90 Pa, the deposition rate was very low and the plasma could not be
sustained for long period necessary to deposit film of reasonable thickness. For this reason, the
sputtering pressure that could be investigated in this study was (0.65 - 0.90) Pa. The sputtering
power for this range of pressure was fixed at 350 watts. From figure 5.6(a) very slight variation of
the deposition rate was found when the deposition pressure was increasing as compared to the case
of decreasing pressure within the chamber. This indicates that the results were produced at a
relatively stable state and can easily be reproduced. Hence, no reasonable hysteresis was reported .
8.0 •.•.~-c: 7.8'E--E.S-
a> 7.6tV0::: I'c:
0 7.4 .•. Pressure Increasing~
II) ~ Pressure Decreasing0a.
a> 7.20
7.0
0.65 0.70 0.75 0.80 0.85 0.90
Sputtering Pressure (Pa)
Figure 5.6 (a) Variation of deposition rate for tungsten oxide thin film as a function of
increasing and decreasing sputtering pressure.
59
The deposition rate as determined using quartz crystal oscillator figure 5.6(a) show the same trend
but differ from those determined through simulations (figure 5.6 (b). An increase in deposition rate
was observed from 0.65 Pa to 0.75 Pa followed by a pronounced decrease from 0.80 Pa to 0.90 Pa.
11.5
11.4
;-0.,.S /i 11.3 ~~/11.2"-" •! 11.1
I::
0.- 11.0.•....-en
0fr '10.9 -,0
10.8
0.65 0.70 0.75 0.80 0.85 0.90
Sputtering Pressure (pa)
Figure 5.6(b): Deposition rate as a function sputtering pressure.
The decrease in sputtering rate at 0.65 Pa can be associated with resputtering at low pressures
(Ohring, 1992). This is due to the fact that at low sputtering pressures, the mean free path is high;
enabling the sputtered tungsten atoms avoid multiple collisions with gas molecules and hence
knock the substrate with high energy resulting to resputtering. The decreased deposition rate at
sputtering pressures above 0.80 Pa can be attributed to a reduction in the mean free path due to
increased pressure resulting to an increased particle scattering. Therefore, few sputtered atoms
reach the substrate resulting to low deposition rate. Decrease in deposition rate as sputtering
60
pressure increases during sputtering is also reported by Kaneko et at. (1982), Berggren (2004) and
Deneuville and Gerald (1978).
5.4 Optical properties of tungsten oxide thin films
Sputtering power and sputtering pressure were found to affect the deposition rate (section 5.3) and
therefore influence the optical properties of the films. Their effect on transmittance, refractive
index, and band gap energy and extinction coefficient of the deposited tungsten oxide thin films are
reported and discussed in this section.
5.4.1 Effect of sputtering power on optical properties
The effect of sputtering power on transmittance, refractive index, and band gap energy and
extinction coefficient are reported and discussed in this section.
5.4.1.1 Effect of sputtering power on transmittance
From figure 5.7(a), the optical transmittance spectra of tungsten oxide thin films deposited at
sputtering power of 300 watts to 400 watts shows high transmittance within the wavelength range
I'
300-800 nm for different sputtering power. All spectra reveal very pronounced fringe interference
effects as demonstrated by oscillations on transmittance spectra for photon energies below the
absorption edge (wavelength greater than 350 nm). The periods of oscillation for films deposited at
different sputtering power are approximately equal. The dependence of transmittance on
wavelength is shown where transmittance peaks for films sputtered at different power increases
with wavelength. Figure 5.7(b) is an extract from figure 5.7(a) which clearly shows the variation of
transmittance on different films at a wavelength of 500 nm. From this figure, transmittance is found
61
to decrease with increase in sputtering power. The optical transmittance of the films at 500 nm
decreased from approximately 87 % to 70 % with increase in sputtering power from 300 watts to
400 watts.
Figure 5.8 (a) and (b) shows typical transmittance simulation spectra for tungsten oxide thin films
deposited at 325 watts and 375 watts respectively. These sample spectra represents the simulation
results for films deposited at varying sputtering power. The fit between the simulation spectrum
and the experimental transmittance spectrum is reasonably good. The deviation of experimental
spectrum from the simulation spectrum was 4.66 x 10-5 nm at 325 watts and 3.68 x 10-5nm at 375
watts as indicated during simulation.
100
80
-:::R 600••.....
Q)occo~ 40·e
I/)cco•... '20.-
0
200
'*' 300 watts
-<>- 325 watts
--- 350 watts
.•. 375 watts
----l~ 400 watts
300 400 500 600 700 800
Wavelength (nm)
Figure 5.7(a) optical transmittance as a function of wavelength for varying sputtering power
62
95
90
•
85--..'$.
'-'d) 800a •t: •.- 75eIZla •
~ 70 •
65
60
300 320 340 360 380 400
Sputtering Power (Watts)
Figure 5.7(b): Transmittance as a function of sputtering power at 500 nm. This data was obtained
from the experimental spectra shown in figure 5.7(a)
100
90
80
70-~ 60Q)
0c: 50co~
E 40fI)c:
~ 30I-
20
10
0
300 400
-------..-Simulation
-- Measured at 325watts
500 600 700 800
Wavelength nm
Figure 5.8 (a) Simulated and measured spectra for film deposited at 325 watts
\
100
90
80
70;?e......
60CI>0c 50m~'e 40(/Jcm 30'-~
20
10
0
300
63
-+- Simulation
-- Measured at 375watts
400 500 600 700 800
Wavelength (nm)
Figure 5.8(b): Simulated and measured spectra for film deposited at 375 watts
From the results shown in figure 5.7(a) the deposited tungsten oxide thin films are all transparent
with their transmittance in the range of (60-87) percentage in the visible regions of the
electromagnetic spectrum. Thisobservation is supported by Gordon et al. (2001) who reported that
in its' fully oxidized state W03 is transparent throughout the visible region of the spectrum. The
presence of oscillations in the transmittance spectrum suggests that the film thickness is fairly
uniform. The decrease in transmittance as shown in figure 5.7(b) can be attributed to the fact that at
higher sputtering power, the sputter yield is higher. As a result, more tungsten atoms reach the
substrate and hence the ratio of tungsten to oxygen atoms in the resulting film is greater. The
resulting sub-stochiometric WOx ,where x decreases with sputtering power show less transmittance.
The good fit between simulation spectrum and experimental transmittance spectrum (figure 5.8(a)
and (b)) implies that the optical constants obtained are reliable.
64
5.4.1.2 Effect of sputtering power on refractive index
Figure 5.9(a) shows the variation of refractive index with the wavelength for films deposited at
different sputtering power. The dependence of refractive index with wavelength is clearly shown
where refractive index decrease with increase in wavelength. Figure 5.9(b) show the variation of
refractive index with sputtering power at 400 nm where refractive index was found to increase with
increase in sputtering power. However, the change in refractive index at 325 watts, 350 watts and
375 watts was found to be quit small but the change was notably high for films deposited at 300
watts and 400 watts.
2.60
2.55
2.50
2.45
2.40
2.35
>< 2.30<U 2.25'"0 2.20oS
<U 2.15.:: 2.10- 2.05o'"<l:l 2.00
<U 1.95~ 1.90
1.85
1.80
1.75
1.70
300
I'
.- 300Watts
-+- 325Watts
• 350Watts
-s-375Watts
~OOWatts
~- ,.---.
'Y-- ~ .•.... *...•. ~ ~~ •...•. .•...
..•.
400 500 600 700 800
W a ve Ie n 9 th (n m )
Figure 5.9 (a): Refractive index as a function of wavelength for films deposited at different
sputtering power.
65
2.40
2.35
2.30
x 2.25Q)"0
C
Q) 2.20>:a 2.15
IV•...- 2.10Q)0:::
2.05 •2.00
300
•
Refractive Index at 400 nm
•• •
320 340 360 380 400
S puttering Power (Watts)
Figure 5.9(b): Refractive index as a function of sputtering power at 400 nm. This data is
obtained from figure 5.9(a)
The increase of refractive index with increase in sputtering power can be associated with higher
energy of the sputtered atoms at higher sputtering power resulting to a higher density of the
deposited film. The fact that refractive index for films deposited at 325 watts, 350 watts and
375 watts was approximately the same suggests that that the densities of these films do not vary
reasonably. At sputtering power of 400 watts, the sputter yield is high as is the energy of the
sputtered tungsten atoms. This results to greater packing density of the film and hence high
refractive index. The refractive index for film deposited at 400 watts was found to be more than
2.3 for wavelengths less than 500 nm. Electrochromic films have their refractive index around
(2.3 - 2.4) Kaneko et al. (1982). Therefore, from this study films formed at an increased
sputtering power (400 watts) with refractive index more than 2.3 are likely to be
electrochromic.
66
5.4.1.3 Effect of sputtering power on optical band gap
The band gap energy for the films deposited decreases with increasing sputtering power as shown
in figure 5.10. At a sputtering power of 300 watts the band gap energy was a bout 3.16 eV and
decreases to 2.99 eV at 400 watts. However, at 375 watts the band gap was 2.97 eV which is lower
than that at 400 watts .
3.15 •
>'Q)"-"
~ 3.10
J-.
Q)
~
~ 3.05c
]
t:O 3.00
•
•
•
•
300 320 340 360 380 400
Sputtering Power (watts)
Figure 5.10: Band gap energy as a function of sputtering power
At high sputtering power, the ratio of tungsten to oxygen increases due to increased sputter yield
resulting to increased deficiency of oxygen in W03-x, where (x>O) hence forming substochiometric
oxide with pseudo bands due to incomplete bonds. Such defects produced localized states in the
67
amorphous solids. The presence of these bands is associated with lower band gap energy (Theye,
1974). The general decrease in the band gap energy with sputtering power can be attributed to
presence of these localized states. The band gap values in this study (2.97 -3.16) eV compares
resonably with report by Daniel et al. (1988) that the band gap energy for reactively sputtered
amorphous tungsten oxide from tungsten target is around 3.10 eV. The variation of band gap
energy at 375 watts may be due to changes in composition, defects and impurities on the deposited
film that affects the band gap.
5.4.1.4 Effect of sputtering power on extinction coefficient
Figure 5.11(a), show the variation of extinction coefficient with wavelength where the extinction
coefficient was found to increase exponentially with wavelength within the visible region of the
spectrum. Figure 5.11(b) shows the variation of extinction coefficient with sputtering power. This
data was extracted from figure 5.11(a) at 700 nm. The extinction coefficient decreased with
increases in sputtering power; however, the value at 400 watts was almost the same as that at 325
watts.
40 /
,- 35 • 300Watts »
~ • 325Watts0 30 ~ 350WattsT"""x .- ~C 25 ... 375Watts
Q) --+--400 Watts
~
..,.
·0
IE 20
~ ..,..
U 15ca13 10c
~W 5
0
500 600 700 800
Wavelength (nm)
68
Figure 5.11 (a) Extinction coefficient as a function of wavelength for films deposited at different
sputtering power
•
,., 2.0'0•.....•
x • •.•...s:: •Q).-oS
Q)
0U 1.5 •s::0.-.•...os::.-t<~
1.0
300 320 340 360 380 400
Sputtering Power (Watts)
Figure 5.11(b): Extinction coefficient as a function of sputtering power for films deposited at 700
nm. This data is obtained from figure 5.11(a)
The trend for extinction coefficient as function of sputtering power is similar to that of band gap
energy. At 400 watts, the coefficient remains the same as that of325 watts (figure 5.11(b). Increase
I'in extinction coefficient for films deposited at 400 watts can be attributed to the changes in
structure and composition of tungsten oxide formed. This is supported by the x-ray diffraction
pattern (figure 5.17) that show film deposited at 400 watts having more defined peak at 25° as
compared to others. Ong et al. (2001) reported that for large non-stoichiometry W03-x where (x>
0.4), the absorption goes up. From the equation a =41fk1A the extinction coefficient is proportional
to absorption coefficient and hence the extinction coefficient is found to increase at 400 watts.
69
5.4.2 Effects of sputtering pressure on optical properties
The effect of sputtering pressure on transmittance, refractive index, optical band gap, and
extinction coefficient of tungsten oxide thin film are reported and discussed in this section.
5.4.2.1 Effects of sputtering pressure on transmittance
Figure 5.12(a) show the optical transmittance for deposited tungsten oxide films with wavelength.
The transmittance peaks for films deposited at different sputtering pressure were found to increase
with increase in wavelength. The transmittance spectrum also indicates reasonable oscillation
where the period of the oscillation for different films was approximately equal. Figure 5.12(b) has
it's data extracted from figure 5. 12(a).It shows clearly the variation of transmittance with sputtering
pressure at 500 nm. From the graph, the transmittance was found to increase with increase in
sputtering pressure. Figure 5.13 (a) and (b) show the spectral fit between the experimental
transmittance spectrum and the simulated spectrum for films samples deposited at a sputtering
pressure of 0.65 Pa and 0.85 Pa. The spectral fit as indicated, reveal a good agreement between the
experimental and simulated spectra. These samples are typical examples of the other films
deposited under varying sputtering pressure.
70
100
00
00
70-~0'-' 60Q)
0c 50cu.•..•.•..•·E 40CJ)ccu 30•...I-
20
10
0
300 400
-T-O.90Pa
-III- 0.85Pa
~0.75Pa
~O.65Pa
500 600 700 800
Wavelength (nm)
Figure 5.12(a) Optical transmittance as a function of wavelength for films deposited at different
sputter pressure
95
•
_90
~ •~ I'
Q)
~ •.-
~85
~E-<
•
80
0.65 0.70 0.75 0.80 0.85 0.90
Sputtering Pressure (Pa)
Figure 5.12 (b): Transmittance as function of sputter pressure at 500 nm. This data is obtained from
experimental spectra shown-in figure 5.12 (a).
100
90
80
70...-
~ 60Q)
0c: 50CO~'E 40fJ)c:
CO 30L.~
20
10
0
300 400
71
••••••
+ Simulated
-a- Measured at 0.65 Pa
500 600 800700
Wavelength (nm)
Figure 5.13(a) Simulated and measured spectra for film deposited at 0.65 Pa
100
90
80
70.•..•...~0'-' 60Q)
0c ~oro~'E 40
(/)
Cro 30L.t-
20
10
0
300 400
+ Simulated
-a- Measured at 0.85 Pa
500 600 700 800
Wavelength (nm)
Figure 5.13(b): Simulated and measured spectra for film deposited at 0.85 Pa
72
The recorded spectra show that tungsten oxide thin films deposited at sputtering pressures ranging
from 0.65 Pa to 0.90 Pa are generally transparent with their maximum transmittance ranging
between (81-90) percent within the visible spectrum. The increase in transmittance at higher
pressures (figure 5.12(b)) can be attributed to the fact that at high pressures reduced mean free path;
ensures high rate of collision between tungsten atoms and gas molecules, resulting to reduced
momentum of sputtered atoms and therefore low packing density of the deposited film. At low
chamber pressures, the sputtered tungsten atoms strike the substrate with high energy due to
reduced collision as a result of high mean free path (Swann, 1988).The porosity of the films
deposited is reduced and the refractive index increases resulting to decreased transmittance.
Therefore, transmittance is found to increase with increase in sputtering pressure. The good fit
(figure 5.13 (a) and (b)) between the experimental spectrum and simulated spectrum suggest that
the model used in this simulation was adequate and hence the results are reliable.
5.4.2.2 Effects of sputtering pressure on refractive index
Figure 5.14(a) shows spectral curves for refractive index as a function of wavelength for films
deposited at different sputtering pressure. From the figure, refractive index was found to decrease
with increase in wavelength. However, for films deposited at 0.75 Pa and 0.65 Pa their refractive
index differ slightly as compared to those deposited at sputtering pressure of 0.90 Pa and 0.85 Pa,
Figure 5.14(b) show the variation of refractive index at 400 nm for films deposited at different
sputtering pressure. This data was obtained from figure 5.14(a) and clearly indicates that refractive
index decreased with sputtering pressure. The refractive index decreased from a bout 2.43 to 2.08
73
as sputtering pressure increased from 0.65 Pa to 0.90 Pa respectively.
2.4 -F====~====A:==~ __
~ 2.2Q)~~-Q)> 2.0.-.•...
0dtl::
Q)
~ 1.8
0.65 Pa
••• 0.75 Pa
• 0.85 Pa• 0.90 Pa
400 500 600
1.6~~==~==~===-~--~--~--~----~--~---r---'
700 800
Wavelength (nm)
Figure 5.14 (a): Refractive index as a function of wavelength for films deposited at different sputter
pressure
2.45
•2.40 •
2.15
x 2.30Q) •"0c:- 2.25
Q).2:
~ 2.20'-'+-Q)a:: 2.15
2.10
•2.05
0.65 0.70 0.75 0.80 0.85 0.90
Sputtering Pressure (Pa)
Figure 5.14(b): Refractive index at 400 nm as function of chamber pressure
74
Decrease in refractive index with increase in wavelength (figure 5.14(a)) agree with research
finding reported by Green (1990), which indicate that refractive index for tungsten oxide films
made by reactive de magnetron sputtering is about 2 with a tendency to decrease with wavelength.
Figure 5.14(b) shows that the refractive index of tungsten oxide thin film was found to decrease
with increase in sputtering pressure. This observation can be attributed to decrease in density at
high sputtering pressures, since refractive index increases with density (Miyake et al. 1983). At
high sputtering pressure, the mean free path is small and there are many collisions, which reduces
the number and the energy of the sputtered atoms that settle on the substrate. Hence, packing
density is low. This may lead to formation of porosities in the film at high pressures resulting in
films of low refractive index. The observation made in this work is supported by report on density
variation of tungsten oxide with decreasing pressure by Miyake et al. (1983). The report indicates
that the density for sputtered amorphous tungsten oxide is about 6.1 g/cnr' and is found to increase
to 6.7 g/cnr' with decreasing sputter pressure. The results obtained suggests that low sputtering
pressure are likely to produce better electrochromic films as their refractive index values range
from 2.3 - 2.4. This is in reference to Kaneko et al. (1982) finding that good electrochromic
I'materials have their refractive index in the range 2.3-2.4.
75
5.4.2.3 Effects of sputtering pressure on optical band gap
Figure 5.15 shows a decrease of the optical band gap with increase in sputtering pressure. The band
gap energy was found to decrease from approximately 2.95 eV to 2.89 eV as sputtering pressure
increased from 0.65 Pa to 0.90 Pa respectively. At high sputtering pressure the mean free path
decreases resulting to multiple collision of the sputtered atoms. Therefore, fewer tungsten atoms
with less energy reach the substrate. The resulting film has a stoichiometry with less oxygen
deficiency in comparison with those deposited at lower sputtering pressure. Although this film is
amorphous and disorderly, the grain sizes are considered wider than for those at lower pressures
resulting to reduce band gap. This was pointed out by Green and Hussain (1991) in their
explanation for band gap widening in disordered film.
2.95 • •
2.94 •->Q)........-
>- 2.93
0>~
Q)c
W 2.92a.roo
"U 21}1croCD
2.90 •
2.89
0.65 0.70 0.75 0.80 0.85 0.90
Sputtering Pressure (Pa)
Figure 5.15: Variation of band gap energy with sputtering pressure
76
The variation of band gap energy with sputtering pressure as shown in this study is supported by
research finding reported by Kaneko et al. (1988), Kitao et al. (1992), and Miyake et al. (1983).
They reported 3.0 < Eg< 3.4 eV for tungsten oxide thin film with a tendency that at high pressure
of the sputter gas and a high 02 admixture (total sputtering pressure) in reactive sputtering giving a
low Eg. The lower band gap reported in this study may be due to possible contamination in
sputtering chamber, and incomplete bonds resulting to formation of localized bonds.
5.4.2.4 Effects of sputtering pressure on extinction coefficients
From figure 5.16(a), the extinction coefficient was found to increase exponentially with
wavelength within the visible region of the spectrum. The coefficient is also found to increase with
sputtering pressure from (0.75 - 0.90) Pa (figure 5. 16(b)). From figure 5.15 the band gap energy in
this study is found to be lower for films deposited at higher pressures.
30
25
~ 20
C
Q)·uIE 15
Q)o
0"
C 10oUc:;:J 5
X
W
"P- 0.90 Pa
---0.85 Pa
• 0.75 Pa
0.65 Pa
•
••
O~~----.---~----.---~-----.----~---.--~
500 600 700 800
Wavelength (nm)
Figure 5. 16(a): Extinction coefficient as a function of wavelength for tungsten oxide thin films
deposited at different sputtering pressure
77
1.5 •
... 1.4'0 •~
X.•..c:Q) 1.3'0:;::
Q)
0 •o 1.2c:
015 •c:
~ 1.1W
1.0
0.65 0.70 0.75 0.80 0.85 0.90
Sputtering Pressure (Pa)
Figure 5,16(b): Extinction coefficient as a function of sputtering pressure at 600 nm.
The absorption is therefore expected to be higher at high- sputtering pressure deposited films.
Since extinction coefficient (k) relates to absorption coefficient (u) through the equation a =47dd1
then the observed trend agrees with the theoretical predictions. However, film deposited at 0.65 Pa
is found to have greater extinction coefficient contrary the trend exhibited by the others. This can
be associated with various processes that take place during this film deposition, such as
resputtering, which affects the tungsten to oxygen atoms ratio, in addition to presence of impurities
and defects in the resulting film.
78
5.5:X-ray diffraction characteristics
This section reports on x- ray diffraction characteristics of the deposited films as influenced by
sputtering power and sputtering pressure.
5.5.1 Effect of sputtering power on x - ray diffraction characteristics
Figure 5.17 shows the intensity spectra for films deposited at 300 watts, 350 watts and 400 watts.
The intensity spectra for all the samples show two broad peaks at approximately 25° and 55°.
However, at 25° the peak intensity for film deposited at 400 watts is relatively pronounced in
comparison to the other. In addition, the intensity of the peak at 300 watts was relatively low in
comparison to the others.
1200
1000
800
enQ.
U 600
>-:!:enc:Q) 400-c:
I"
200
0
--Intensity spectra at 350 watts
--Intensity spectra at 400 watts
--Intensity spectra at 300 watts
20 30 40 50 60 70 80
Figure 5.17: XRD spectra for tungsten oxide thin film deposited at different sputtering power.
79
The absence of any intensive peaks indicate that the x-ray scattered from different crystal plane in
the film are not in phase and therefore do not interfere constructively. Hence, the films formed are
amorphous in structure. The peak at 25° appear to become more pronounced as sputtering power
increases from 300 watts to 400 watts. This agrees with previous suggestion based on variation of
transmittance and refractive index that the film stoichiometry changes with increase in sputtering
power although within the range of this study the deposited films were amorphous.
5.5.2 Effect of sputtering pressure on x-ray diffraction characteristics
Figure 5.18 show the x -ray diffraction pattern for films deposited at sputtering pressures of 0.75
Pa and 0.85 Pa .These two spectra are typical examples of the other films deposited under varying
sputtering pressure ..
900
800
700
600
II)a.
U 500
~·in 400c2E 300
200,-
100
o
--Intensity spectra at O.85Pa
Intensity spectra at O.75Pa
20 30 40 50 60 70 80
Figure 5.18: XRD spectra for tungsten oxide thin films deposited at 0.85 Pa and 0.75 Pa
All the patterns reveal two broad peaks. The main peaks are at approximately at 25° and the minor
peaks at 55°. The peak intensities for the two spectra are comparable at approximately 800 cps and
80
650 cps for films deposited at 0.75 Pa and 0.85 Pa respectively. The presence of the two broad
peaks and the absence of any intensive peaks indicate, that the x-ray scattered from different crystal
plane in the film are not in phase and therefore, do not interfere constructively hence the films are
amorphous in structure. There is no major difference in the XRD pattern, incase of varying
sputtering pressure within the range investigated in this work. This compares well with report of
studies by (Ong et al., 2001) that tungsten oxide thin films are amorphous.
5.6 Surface microstructure
Figure 5.19 shows an AFM image (1000 x 1000 nm scan range) of a sample of the tungsten oxide
films deposited at 325 watts in this study. The image reveals very few white protrusions that are
sparsely populated. Hence, the film is generally smooth. This implies that the deposited tungsten
oxide thin films are amorphous in agreement with the results of XRD measurements reported in
section.5.5. The surface features results show the columnar structure that is typical for many
dielectric thin films. Surface microstructure of thin film coating significantly affects important
properties of the optical scattering. Knowledge of optical scattering is important in determining the
quality of optical components used in complex applications such as lasers, microscopes, and
lithography systems. From scattering theories, the amount of optical scatter depends on the
roughness height of a structure and on its' lateral distribution. Atomic force microscope (AFM) can
measure the surface microstructure and the information obtained can substantially contribute to
successful development of low-scatter optical coatings.
81
Figure 5.19: Atomic force micrograph of the surface of tungsten oxide thin film.
82
CHAPTER SIX
CONCLUSIONS AND RECOMMENDATIONS
6.1 Conclusion
Tungsten oxide thin films were successfully prepared by reactive de magnetron sputtering of
tungsten target on glass substrate at room temperature. A correlation between the deposition
parameters and optical properties of the films has been established. In particular, the influence of
sputtering power between 300 watts and 400 watts and sputtering pressure between 0.65 Pa and
0.90 Pa were investigated. Changes in sputtering power were found to affect the deposition rate,
refractive index, optical band gap and extinction coefficient such that, deposition rate and refractive
index increases with sputtering power while band gap energy and extinction coefficient decreases
with sputtering power. We can conclude that deposition rate highly influences the optical properties
of tungsten oxide thin films deposited by reactive de magnetron sputtering.
For the case of changing sputtering pressure, the refractive index and optical band gap decreases as
the sputter pressure increase while the extinction coefficient increases with increase in chamber
pressure. Therefore, the sputtering pressure influences the optical properties of the resulting
"
deposited films. Variation of refractive index among films deposited at different sputtering power
and sputtering pressure suggest structural or compositional differences (Smith D. L., 1995).
Therefore, deposition parameters used in this study influenced the optical as well as the structural
properties of the deposited films. The amorphous nature of films deposited at these conditions of
sputtering power and pressure are shown by their x-ray diffraction spectra. In addition, AFM
images reveal a relatively smooth surface that is a characteristic of amorphous structure. Further,
the transmittance spectra reveal highly transparent films within the visible range. From the results
83
obtained, we can conclude that the optical properties of tungsten oxide thin films are highly
influenced by the sputtering power and the prevailing sputtering pressure.
6.2 Recommendations
From the results of this study, it is clear that deposition conditions influence the optical properties
of deposited films; however, the challenge remains the ability to maintain reproducibility and allow
the transfer of process from one system to the other. To overcome this challenge it is important for
more research in order to understand the influence of deposition parameters on type and energy of
deposited atoms, film bombardment and substrate temperature. This can be done by applying
different characterization methods such as Elastic Recoil Detection Analysis (ERDA) to gather
reliable information.
Among the main applications of tungsten oxide thin films is its use as an electrochromic material.
Although e1ectrochromism has not been investigated in this study, report by Kaneko et al. (1982)
indicates that good electrochromic materials have their refractive index between 2.3-2.4. From this
study films deposited at high sputtering power of 400 watts at a constant sputter pressure of 0.80
Pa, and those deposited at lower sputter pressures of 0.65 Pa and 0.75 Pa while sputtering power
was maintained at 350 watts were found to have their refractive index within the range of 2.3-2.4.
We would therefore recommend that further studies be done on films deposited at similar
conditions to determine their e1ectrochromic efficiency.
In addition films deposited at low sputtering pressure and high sputtering power are found to have
lower transmittance suggesting that they are W03-x where x>O implying a high oxygen deficiency.
It is therefore necessary for further analytical test such as Rutherford back scattering to be done in
order to establish the stoichiometry of these films. This would therefore give a clear relationship
84
between the deposition parameters, the optical properties and possible electrochromism of the film
deposited.
85
REFERENCES
Arnoldussen, T.C. (1976). Electrochromism and Photochromism In MoO) Fnm~,JaWnal
Electrochem. Soc., 123, 527-528.
Bach, H. and Krause, D. (1997). Thin Film on Glass. Springer, 404 pp.
Bange, K. (1999). Colouration of tungsten oxide films; a model for optically active coatings. Solar
Energy material & Solar cells, 58, 1-131.
Bange, K., Gambke, T., and Sparschuh, G. (1995). Optically active Thin Film Coatings in
Handbook of optical properties volume 1, Reviewed by Bach, II, CRC Press Inc., pp 105-131.
Bellac, D. L., Azens, A., and Granqvist, C.G. (1995). Angular selective transmittance through
electrochromic tungsten oxide film made by oblique angle sputtering. Applied physics letters,
66(14),1715-1716.
Berak J.M. and Sienko M. J. (1970). Effect of oxygen deficiency on electrical transport of tungsten
trioxide crystals. Journal Solid State Chemistry, 2, 109-133.
Berggren, L. (2004). Optical Absorption and Electrical Conductivity in Lithium Intercalated
Amorphous Tungsten Oxide Films. Summaries of Uppsala University Dissertations from the faculty
of Science and Technology, 1045, 73pp.
Born, M. and Wolf, E. (1983). Principles of Optics (6th Edition). Pergamon: Oxford, pp44.
Cummins, D., Boschloo, G., Ryan, M., Corr, D., Rao, S.N. and Fitzmaurice, D. (2000). Ultrafast
electrochromic windows based on redox - chromophore modified nanostructured semi conducting
and conducting films. Journal of Physical Chemistry- B ,104(48), 11449-11459.
Daniel, M. F., Desbat, B., Lassegues, J.C., and Garie, J. (1988). Infrared and Raman spectroscopies
ofrf sputtered tungsten oxide films. Journal of Solid State Chemistry, 73, 127-139.
Deneuville, A., and Gerard, P. (1978). Influence of substoichiometry, hydrogen content and
crystallinity on the optical and electrical properties. Journal Electronic Materials, 7,559-560.
DeVries, M. J., Trimble, C., Thomas, E. T., Thompson, D. W., Woollam, J.A and Hale J.S. (1999).
Optical constants of crystalline WO) deposited by magnetron sputtering. Journal of Vacuum
Science Technology- A, 17(5),2906-2910.
Dobrowolski, J.A. (1994). Optical Properties of Films and Coatings, in Handbook of Optics (2nd
Ed.). New York: MCGraw-Hill,pp130.
Dobrowolski, J.A and Verky P.G. (1993). Inhomogenous and Quasi-Inhomogeneous Optical
Coatings. Proceedings of a conference (SPIE) held in Quebec 19-20 August, 2046, 2-8.
86
Dobrowolski J.A. (1995). Usual and unusual application of optical Thin Films in Handbook of
Optical Properties Volume II, reviewed by Carniglia C. K, CRC Press Inc., pp 6-32.
Goldner, R B., Mendelsohn, D. H., Alexander J., Henderson W. R, Fitzpatrick, D., Hass, T. E;
Sample H .M., Rauh, R D., Parker and Rose T. L. (1983). High near -Infrared reflectivity
modulation with polycrystalline electrochromic W03 films. Applied Physics Letter, 43, 1093-1095.
Gordon, RG., Barry, S., Barton, J. T., Randy N. R, and Broomhall,D. (2001). Atmospheric
pressure chemical vapor deposition of electrochromic tungsten oxide films, Thin Solid Films 392,
231-235.
Granqvist, C.G. (1993). Physics of thin films edited by Francobe M.H. and Vossen J. L. Academic,
San Diego, 17,301-370.
Gratzel, M. (2001). Ultrafast Colour displays. Nature 409, 575-576.
Green, D. (1990). Optical constants of sputtered WO[3].Applied Optics, 29, 4547-4549.
Green, M., and Hussain, Z. (1991). Optical properties of dilute hydrogen tungsten bronze thin
films. Journal of Applied Physics, 69, 7788.
Hichwa B. P., Caskey, G., Betz, D. F. and Harlow J. D. (1987). Electrical and optical properties of
thin Ta and W oxide films produced by reactive direct current magnetron sputtering. Journal of
vacuum Science Technology, 5, 1775-1777.
Hirono, T., and Yamada, T. (1986). Photo chromic Silver complex oxide thin film synthesized by
anodic oxidation. Journal Applied Physics, 59,948-949.
Huang, C. C., Tung, J.C., and Tao, W.H. (2004). Optical properties of tungsten and titanium oxide
thin films prepared by plasma sputter deposition. Solar Energy Material. & Solar Cells, 83, 15-28.
Hurditch, R (1975). Eectrochromism in hydrated tungsten-oxide films. Electronic Letter, 11, 142-
144.
Hummel, RE. and Guenther, K.H. (1995). Handbook of Optical Properties, Volume 1, ICRC Press,
pp 106-109.
Kaneko, H., Miyake, K. and Teromoto, T.J. (1982). Preparation and Properties of reactively
sputtered tungsten oxide film. Applied physics, 53(4),3070-3075.
Kaneko, H., Nishimoto, S., Miyake, K. and Suedomi N. (1986). Physical and electrochromic
properties of rf sputtered tungsten oxide films. Journal of Applied Physics, 59(7), 2526-2534.
Kaneko, H., Nagao, F. and Miyake, K. (1988). Preparation and properties of de reactively
sputtered tungsten oxide films. Journal of Applied Physics, 63(2), 510-517.
\
87
Khan, K.A., and Granqvist, C.G. (1989). Thermochromic sputter-deposited Vanadium oxyfloride
coatings with low luminous absorptance. Applied Physics letter, 55,4- 5.
Kitao, M. S., Yamada, S., Yoshida, S., Akram, H., Urabe, K. (1992) Preparation conditions of
sputtered electrochromic W03 films and their Infrared absorption spectra. Solar Energy Material &
Solar cells, 25, 241 -255.
Lampert, C. M. (1998). Smart switchable glazing for solar energy and daylight control. Solar
Energy Material & Solar Cells, 52, 207-221.
Lampert, C. M. (2003). Large area smart glass and integrated photovoltaic. Solar Energy Material
& Solar Cells, 76, 489 --499.
Larsson, A. L (2004). All Thin Film electrochromic Device for optical and Thermal Modulation.
Comprehensive Summaries of Uppsala Dissertation from Faculty of Science and Technology,
Uppsala University, 54 pp.
Lee, E.S., and DiBartolomeo, D.L. (2002). Application issues for large-Area electro chromic
windows in commercial buildings. Solar Energy material and Solar cells, 71,465-495.
Lee, W. (1985). Thin films for optical data storage. Journal Vacuum Science Technology, A3, 640-
646.
Miyake,S., Kaneko, H., Suedomi, N. and Nishimoto, S.(1983). Physical and electrochromic
properties ofrf sputtered tungsten oxide films. Journal of Applied Physics, 54(9),5256-5261.
Monk, P.M.S., Delege, F. and Vierra, S.M.C. (2001). Electrochromic paper: Utility of
electrochromes incorporated in paper. Electrochim. Acta., 46 (13-14), 2195-2202.
Mucke, K., Bohm, F., Gambke, T., Ottermann, C., Bange, K. (1990). Deposition methods and
process techniques for the fabrication of electrochromic all solid-state devices (ASSDs), SPIE ,
1323, 188.
Ngaruiya, J.M. (2004). Fundamental Processes in Growth of Reactive DC Magnetron Sputtered
Thin Films. Tag der mundlichen Prufung, pp 50-51.
Ohring, M. (1992). The Material Science of Thin Films. Academic Press Inc., pp128-581.
O'Leary, S K., Johnson, S. R., Lim, P.K. (1997). The relationship between the distribution of
electronics states and the optical absorption spectrum of an amorphous semi-conductors -an
empirical analysis. Journal of Applied Physics, 82, 3334-3340.
Ong, C. W., Wong H. Y., Pang, G. K. H., Baba-Kishi K. Z. and Choy C. L. (2001) Relationship
between the macrostructure and nanoindentation hardness of thermally evaporated and magnetron
sputtered electrochromic tungsten oxide films. Journal of Material Research, 16, 1541-1548.
\
88
Raul, RD. (1999). Electrochromic Windows: an overview. Electrochim. Acta, 44(18), 3165-3176.
Rempe, G., Thompson, RJ., Kimble, H. J., and Lal ezari , R (1993). Measurement of Ultralow
losses in an optical interferometer. Optical letter, 17,363-365.
Risser, S. M. and Ferris, K. F. (1990) Theoretical calculations of the non-linear function of
inhomogeneous thin films. Conference paper Pacific North West Lab., Richardland, WA (USA),
8pp.
Rowley, M.N. and Mortimer, J.R (2002). New electrochromic Material. www.scilet.com, pp 244-
261.
Salinga, C., Weis, H., Wuttig, M. (2002). Gasochromic switching of tungsten oxide films: A
correlation between film properties and coloration kinetics. Thin Solid Films, 414 (2),288-295.
Schield, J. (2004) Bragg's law and Diffraction: How waves reveal the atomic structures of crystals.
State University of New Yolk, Stoney Brook, NY 117994 -2100.
Scrosati, B. (1993). Application of Electro active Polymers. Chapman & Hall, UK, pp58-83.
Smith, D. L. (1995). Thin Film Deposition: Principles and Practice. Mcgraw-Hill Professional,
616pp
Swann, S. (1988). Magnetron Sputtering. Journal Physics Technology ,19, 67-75.
Tauc, J. (1974). Amorphous and Liquid semiconductors, Edited by Tauc, J., New York, Plenum
Press. 18Opp.
Thesis, W. (2004). Spectrum Interpretation by Simulation, edited by M. Theiss (Hard and software,
Aachan, Germany) www.mtheiss.com.
Theye, M. L. (1974). Proceedings of Fifth International Conference on Amorphous and Liquid
Semiconductors, Volume 1, Taylor and Francis, London, pp 479.
Wanlass, M. (1992). A new approach to high-efficiency solar cells. Photonic spectra,
November Issue, 159-166.
Wei, C., Baudrin, E., Dunn, B., and Zink, I.J. (2001). Synthesis and electrochromic properties of
mesoporous tungsten oxide. J. Material Chern., 11,92-97.
Wendt, K., Ellmer, K. and Wiesemann, K. (1997). Thermal power at a substrate during ZnO:Al
thin film deposition in a planar magnetron sputtering system. J. Applied Physics 82(5), 2115-
2122.
Wooten, F. (1972). Optical properties of Solids. Academic Press, New York, pp 49.
\
89
Yeon-Gon, M., Dillon R. O. and Snyder P. G. (1999). Visible and infrared photo chromic properties
of amorphous W03-x films. Journal of Vacuum Science Technology A ,17(5), 2933-2938