STRUCTURAL PROPERTIES OF HYDROGENATED AMORPHOUS SILICON
(a-Si:H) THIN FILM GROWN VIA RADIO FREQUENCY PLASMA
ENHANCED CHEMICAL VAPOR DEPOSITION (RF PECVD)
HASBULLAH BIN ANTHONY HASBI
A thesis submitted in fulfilment of the
requirements for the award of the degree of
Master of Science (Physics)
Faculty of Science
Universiti Teknologi Malaysia
JUNE 2005
PSZ 19:16 (Pind. 1/97)
UNIVERSITI TEKNOLOGI MALAYSIA
BORANG PENGESAHAN STATUS TESIS
JUDUL:
STRUCTURAL PROPERTIES OF HYDROGENATED
AMORPHOUS SILICON (a-Si:H)THIN FILM GROWN VIA
RADIO FREQUENCY PLASMA ENHANCED CHEMICAL
VAPOR DEPOSITION (RF PECVD)
SESI PENGAJIAN: 2004 / 2005
HASBULLAH BIN ANTHONY HASBI _____
Saya
(HURUF BESAR)
mengaku membenarkan tesis ini disimpan di Perpustakaan Universiti Teknologi Malaysia dengan
syarat-syarat kegunaan seperti berikut :
1.
Hakmilik tesis adalah dibawah nama penulis melainkan penulisan sebagai projek bersama dan
dibiayai oleh UTM, hakmiliknya adalah kepunyaan UTM.
Naskah salinan di dalam bentuk kertas atau mikro hanya boleh dibuat dengan kebenaran bertulis
daripada penulis.
Perpustakaan Universiti Teknologi Malaysia dibenarkan membuat salinan untuk tujuan pengajian
sahaja.
Tesis hanya boleh diterbitkan dengan kebenaran penulis. Bayaran royalti adalah mengikut kadar
yang dipersetujui kelak.
*Saya membenarkan/tidak membenarkan Perpustakaan membuat salinan tesis ini sebagai bahan
pertukaran antara institusi pengajian tinggi.
**Sila tandakan (√)
2.
3.
4.
5.
6.
√
SULIT
(Mengandungi maklumat yang berdarjah keselamatan atau
kepentingan Malaysia seperti yang termaktub di dalam AKTA
RAHSIA RASMI 1972)
TERHAD
(Mengandungi maklumat TERHAD yang telah ditentukan oleh
organisasi/badan di mana penyelidikan dijalankan)
TIDAK TERHAD
Disahkan oleh
_____________________________
(TANDATANGAN PENULIS)
Alamat Tetap:
275, KPG. SUNGAI BEDIL BESAR,
PETRA JAYA, 93050, KUCHING,
SARAWAK.
Tarikh: 14 JUN 2005
_____________________________
(TANDATANGAN PENYELIA)
P. M. DR ZULKAFLI OTHAMAN
(NAMA PENYELIA)
Tarikh: 14 JUN 2005
CATATAN: * Potong yang tidak berkenaan.
** Jika Tesis ini SULIT atau TERHAD, sila lampirkan surat daripada pihak berkuasa/
organisasi berkenaan dengan menyatakan sekali tempoh tesis ini perlu dikelaskan
sebagai SULIT atau TERHAD.
“I/We* hereby declare that I/we* have read this thesis and in my/our*
opinion this thesis is sufficient in terms of scope and quality for the
award of the degree of Master of Science (Physics)”
Signature
:
Name of Supervisor I
: P. M. DR. ZULKAFLI OTHAMAN
Date
:
Signature
:
Name of Supervisor II
: PROF. DR. SAMSUDI SAKRANI
Date
:
*Delete as necessary
BAHAGIAN A – Pengesahan Kerjasama*
Adalah disahkan bahawa projek penyelidikan tesis ini telah dilaksanakan melalui
kerjasama antara _______________________ dengan _______________________
Disahkan oleh:
Tandatangan :
Nama
:
Jawatan
(Cop rasmi)
:
Tarikh :
* Jika penyediaan tesis/projek melibatkan kerjasama.
BAHAGIAN B – Untuk Kegunaan Pejabat Sekolah Pengajian Siswazah
Tesis ini telah diperiksa dan diakui oleh:
Nama dan Alamat Pemeriksa Luar
:
Nama dan Alamat Pemeriksa Dalam :
Nama Penyelia Lain (jika ada)
PROF. DR. ABD. HALIM SHAARI
PHYSICS DEPARTMENT,
FACULTY OF SCIENCE,
UNIVERSITI PUTRA MALAYSIA
PROF. MADYA DR. YUSSOF WAHAB
PHYSICS DEPARTMENT,
FACULTY OF SCIENCE,
UNIVERSITI TEKNOLOGI MALAYSIA
:
Disahkan oleh Penolong Pendaftar di SPS:
Tandatangan :
Nama
:
Tarikh :
iv
“I hereby acknowledge this thesis as my own research and endeavor work except for
some quotations and abridgement that have clearly been stated its point of source.
This thesis has not been accepted for any degree and is not concurrently submitted in
candidature of any degree.”
Signature
:
Author’s name
: HASBULLAH BIN ANTHONY HASBI
Date
:
v
To
My parents Anthony Hasbi and Paimah Junaidi
for setting me on the path toward intellectual pursuits
My sisters Hasmah, Hasmiah and Hazalimah
for their continuing support along the way
My friends
for making the journey so enjoyable.
vi
ACKNOWLEDGEMENT
I would like to acknowledge the large number of people, too numerous to
mention individually, who have stimulated my interest in this research over the past
two years. I would particularly like to thank my project supervisor, Associate Prof.
Dr. Zulkafli Othaman and co-supervisor Prof. Dr. Samsudi Sakrani for their
suggestions and indirect encouragement in this venture.
Furthermore, this thesis would not have been possible without the very
pleasant and creative working atmosphere that prevails at the Vacuum and Research
Laboratory, Department of Physics, UTM Skudai. I would like these lines to be an
expression of my gratitude to my colleagues and my pals especially Lau Yee Chen
and Suriani. Not forgetting lab attendance Mr. Nazri, Kak Wani and Mr. Jaafar.
Their help has turn out to be inestimably important.
My thanks to Prof. John Wilson, Academic Head of Physics, Schools of
Engineering and Physical Sciences, Heriot-Watt University, Edinburgh, UK for his
support and guidance. I really appreciate the helpful discussions and assistance along
my attachment program at Heriot-Watt. Not forgetting all the members of Diamond
Group who’ve helped me with the use of some of the equipments. Their constant
help and support made all the months I’ve spent there one of the best experiences in
my life.
The working and writing of this project has been at the expense of many
hours and miles away from my loving parent. It is dedicated to you, Mom and Dad.
vii
ABSTRACT
An investigation of the structural properties of hydrogenated amorphous
silicon (a-Si:H) thin films prepared by plasma enhanced chemical vapour deposition
of silane (SiH4) was done using a combination of atomic force microscopy (AFM),
photoluminescence, infrared and UV spectroscopy. Films were prepared with rf
power ranging from 100-250 W. For every rf power employed, substrate temperature
were varied from room temperature to 300˚C. The deposition rate was found to be
slightly increasing with an increase of rf power while decreasing as the substrate
temperature increases. The AFM images can be classified into three groups: most
smooth (rms: 1.2nm), intermediate rms (2.4-3.6 nm) and highest roughness (rms: 4.9
nm). The transition to rougher films at higher substrate temperature is attributed to a
change in the deposition process. The IR vibrational spectra obtained from FTIR
spectroscopy display modes which can be characterized as predominantly hydrogen
motions. On the basis of these identifications, it is found that samples produced on
high-temperature have SiH, SiH2 and (SiH2)n groups with very little SiH3. In
contrast, the ir spectra of samples produced on room-temperature are dominated by
vibrational modes of SiH3 and (SiH2)n. At low rf power, the spectrum is dominated
by a strong absorption bands at 2000 cm-1 associated with SiH stretching bond and
also 630 cm-1 associated with SiH bending. At high rf power, an additional
absorption band at around 2090 cm-1 which corresponds to (SiH2)n stretching mode
and SiH2 stretching mode becomes more pronounced. The optical energy gap
obtained from UV spectroscopy decreases with increasing of rf power and substrate
temperature. This decrement is due to the drop of hydrogen content. At low substrate
temperature, photoluminescence spectrum of a-Si consists of a relatively broad band
with its main peak around 1.4 eV. The spectrum shifts to lower energies (around 1.37
eV) and its intensity decreases with increasing temperature. It is suggested that this is
due to an activated non-radiative recombination (relaxation) process where exciton
are captured by deep traps and this become more probable as temperature increases.
viii
ABSTRAK
Satu kajian tentang struktur bahan saput tipis amorfus silikon terhidrogenasi
(a-Si:H) yang disediakan melalui kaedah pemendapan wap kimia diperkuat plasma
dengan gas silane (SiH4) telah dijalankan melalui kombinasi kaedah Mikroskopi
Daya Atom (MDA), fotoluminesen serta spektroskopi inframerah (IM) dan
ultralembayung. Saput telah disediakan dengan kuasa frekusi radio (fr) dari 100-250
W. Bagi setiap kuasa fr yang dikenakan, suhu substrate diubah dari suhu bilik ke
300˚C. Kadar pemendapan didapati meningkat dengan setiap penambahan kuasa fr
manakala ianya menurun apabila suhu substrat meningkat. Imej MDA yang dicerap
boleh diklasifikasikan kepada tiga kumpulan: paling halus (rms 1.2 nm), rms
pertengahan (2.4-3.6 nm) dan paling kasar (rms 4.9 nm). Transisi ke saput yang lebih
kasar adalah disebabkan prubahan yang berlaku dalam proses pemendapan.
Spektrum yang diperolehi daripada spektroskopi inframerah memaparkan mod yang
boleh dicirikan sebagai gerakan hidrogen. Melalui identifikasi ini, didapati sampel
yang disediakan dalam suhu tinggi mempunyai kumpulan SiH, SiH2 dan (SiH2)n
dengan sedikit SiH3. Sebaliknya, dalam suhu bilik didapati spektrum didominasi oleh
mod getaran SiH3 dan (SiH2)n. Pada kuasa fr rendah, spektrum didominasi oleh jalur
serapan yang kuat pada 2000 cm-1 dikaitkan dengan ikatan regangan SiH dan 630
cm-1 dikaitkan dengan bengkokan SiH. Manakala pada kuasa fr yang tinggi, satu
jalur serapan sekitar 2090 cm-1 dikaitkan dengan mod regangan (SiH2)n dan SiH2
didapati semakin ketara. Jurang tenaga optik yang diperolehi melalui spektroskopi
ultralembayung menurun dengan peningkatan kuasa fr dan suhu substrat. Ini
disebabkan oleh menurunnya jumlah hidrogen yang terkandung dalam sampel. Pada
suhu substrat yang rendah, spektrum fotoluminesen a-Si memaparkan jalur lebar
dengan puncak utama sekitar 1.4 eV. Puncak spectrum menurun ke tenaga yang lebih
rendah (sekitar 1.37 eV) manakala keamatannya berkurang dengan peningkatan suhu
substrat. Ini adalah disebabkan oleh teraktifnya proses penggabungan semula tanpa
pemancaran (santaian) di mana exciton diperangkap oleh perangkap dalam pada
jurang tenaga dan proses ini menjadi lebih mudah terjadi dengan peningkatan suhu
substrat.
ix
TABLE OF CONTENTS
CHAPTER
1
2
TITLE
PAGE
DECLARATION
ii
ACKNOWLEDGEMENT
iv
ABSTRACT
v
ABSTRAK
vi
TABLE OF CONTENTS
vii
LIST OF TABLES
x
LIST OF FIGURES
xi
LIST OF SYMBOLS
xv
INTRODUCTION
1.1
Amorphous Semiconductor
1
1.2
Potential Applications
2
1.3
Research Objectives
5
1.4
Research Scope
5
1.5
Layout of Thesis
5
LITERATURE REVIEW
2.1
2.2
Hydrogenated Amorphous Silicon (a-Si:H)
7
2.1
8
Hydrogenation of a-Si
Conventional Thin Film Deposition Techniques
13
x
2.2.1 Chemical Vapor Deposition
13
2.2.2 DC Glow Discharge
14
2.2.3 RF Glow Discharge
18
2.3
Plasma Enhanced Chemical Vapor Deposition (PECVD)
19
2.4
Plasma Fundamentals
19
2.4.1
19
2.5
3
Non-equilibrium Glow Discharges
2.4.2 Potentials in rf Glow Discharges
21
Qualitative Model for PECVD Reactions
22
METHODOLOGY
3.1
3.2
Deposition of Amorphous Silicon Films
27
3.1.1 Introduction
27
3.1.2
27
Plasma Enhanced Chemical Vapor Deposition
3.1.3 PECVD Setup
28
3.1.4
PECVD System Deposition Process
33
3.1.5
Preparation of a-Si:H Samples
34
3.1.5.1 Substrate Preparation
34
3.1.5.2
34
Deposition Parameters
Analytical Tools
36
3.2.1
36
Spectroscopy
3.2.2 Absorption
3.2.3
37
3.2.2.1 Optical Absorption
37
3.2.2.2
42
UV-3101-PC Spectrophotometer
Infrared (IR) Absorption Spectroscopy
45
3.2.3.1 Introduction
45
3.2.3.2
45
Mechanism of IR Absorption
3.2.3.3 FTIR – Fourier Transform Infrared
Spectroscopy
3.2.4
3.2.5
46
Photoluminescence
49
3.2.4.1
Basic Theory of Photoluminescence
49
3.2.4.2
Luminescence Spectrometer LS 55
51
Atomic Force Microscopy (AFM)
54
xi
3.2.5.1
Contact Mode
3.2.5.2 Non-contact Mode
57
3.2.5.3
Tapping Mode
57
3.2.5.4
Image Display
57
3.2.5.5 Surface Roughness
3.3
4
5
56
Film Thickness Measurements
58
58
RESULTS AND DISCUSSIONS
4.1
Surface Morphology and Deposition Rate
61
4.2
X-Ray Diffraction (XRD) Analysis
70
4.3
Infrared (IR) Transmission Spectrum
72
4.3.1 Dependence on rf Power
74
4.3.2
79
Dependence on Substrate Temperature
4.4
Optical Energy Gap
84
4.5
Photoluminescence
89
4.5.1
89
Dependence on Substrate Temperature
CONCLUSION
5.1
Summary
96
5.2
Recommendation
98
REFERENCES
99
PRESENTATIONS
105
xii
LIST OF TABLE
TABLE NO.
3.1
TITLE
PAGE
The preparation conditions for a-Si:H thin films under
different RF power and substrate temperature. Note that
4.1
other parameters are kept constant.
35
The observed mode frequency and their assignments.
72
xiii
LIST OF FIGURES
FIGURE NO.
1.1
TITLE
PAGE
Schematic representation of a typical thin film a-Si:H
solar cell on glass. The incoming photons with an energy
larger than the band gap are absorbed in the intrinsic
a-Si:H film creating holes-electron pairs.
3
1.2
Schematic of a general active matrix array.
4
2.1
A model of hydrogenated amorphous silicon. The small
ball represent hydrogen atoms while the large ball
represents silicon atoms.
2.2
Model for the adsorption, diffusion and recombination
of SiH3 on the growing surface.
2.3
9
12
A common implementation of CVD
(AX, X, BX – gases, AB – solid material).
14
2.4
Current-voltage characteristics in a DC glow discharge.
15
2.5
Luminous zones and dark spaces in a DC glow
discharges.
2.6
The potential distribution in a DC glow discharge.
(Vp – plasma potential, Vc – cathode potential).
2.7
17
Schematic of fundamental transport and reaction
processes underlying CVD.
2.8
16
23
Equivalent circuit representation of the sequence of
steps in thermally driven CVD. The Ri (with i = 1-7)
represent the seven steps describe in the text.
2.9
Equivalent circuit representation showing the thermal
CVD path, and the parallel plasma-enhanced path
24
xiv
represented by the Ri*s.
2.10
25
Activation energy diagram for a thermally driven
(solid line) and plasma enhanced (dashed line) reaction.
The activation energy for the plasma enhanced reaction,
∆E*, is typically less than that for the thermally driven
reaction, ∆E.
26
3.1
NPN2 gas.
29
3.2
PN2 gas rack.
29
3.3
Silane (SiH4) gas cabinet.
30
3.4
Scrubber.
30
3.5
Schematic of the rf PECVD reactor system.
31
3.6
The rf PECVD reactor system.
32
3.7
A schematic of the optical absorption curves of
amorphous materials.
3.8
39
Electron transition for semiconductors having a
direct gap and indirect gap
40
3.9
Shimadzu UV-3101-PC Spectrophotometer.
44
3.10
Examples of infrared active and inactive absorption
bands in CO2.
3.11
46
Fourier transform infrared spectrometer (Perkin Elmer
Spectrum GX).
48
3.12
A schematic of the PL process.
50
3.13
Photoluminescence spectrometer LS 55.
52
3.14
Luminescence spectrometer schematic diagram.
54
3.15
Beam deflection system, using a laser and
photodetector to measure the beam position.
55
3.16
Force between tip and sample surface
56
3.17
Colour mapping scheme for height
58
3.18
Dektak3 surface profiler.
59
4.1
Representation of surface morphology of a-Si:H films.
63
4.2
The 10 µm × 10 µm AFM images with surface height
given by the grey scales extending from 0 (black) to
400 nm (white): (a) sample A deposited at 27°C;
(b) sample B deposited at 100°C; (c) sample C
xv
deposited at 200°C; (d) sample D deposited at 300°C.
4.3
Deposition rate of a-Si:H films deposited at different
temperature as a function of rf power.
4.4
69
X-ray diffraction pattern of a-Si:H thin film
deposited at room temperature and rf power of 100 W.
4.7
68
Model of surface reaction where Si-Si is formed
releasing H2.
4.6
67
Deposition rate of a-Si:H films deposited at different
rf power as a function of substrate temperature.
4.5
64
70
X-ray diffraction pattern of a-Si:H thin film deposited
at 200˚C and rf power of 50 W.
71
4.8
Local Si-H vibrations for SiH, SiH2 and SiH3 groups.
72
4.9
IR transmission of a-Si:H thin films deposited at
different rf power. Emphasize on the 2000 cm-1 regime.
4.10
IR transmission of a-Si:H thin films deposited at
different rf power. Emphasize on the 600 cm-1 regime.
4.11
77
IR transmission of a-Si:H thin film prepared under
different rf power by Lucovsky et al. (1979)
4.13
76
IR transmission of a-Si:H thin films deposited at
different rf power. Emphasize on the 800 cm-1 regime.
4.12
75
78
IR transmission of a-Si:H thin films deposited at
different substrate temperatures. Emphasize on the
2000 cm-1 regime. RF power is at 100 W.
4.14
81
IR transmission of a-Si:H thin films deposited at
different substrate temperatures. Emphasize on the
800 cm-1 regime. RF power is at 100 W.
4.15
82
IR transmission of a-Si:H thin films deposited at
different substrate temperatures. Emphasize on the
600 cm-1 regime. RF power is at 100 W.
4.16
The (αhν)1/2 vs hν graphs for a-Si:H deposited at
different rf power.
4.17
84
The (αhν)1/2 vs hν graphs for a-Si:H deposited at
different substrate temperature.
4.18
83
Energy gap of a-Si:H films prepared under different
85
xvi
substrates temperatures as a function of rf power.
4.19
86
Luminescence spectrum of amorphous silicon at
various substrate temperatures with rf power kept
constant at 50W.
4.20
90
Luminescence spectrum of amorphous silicon at
various substrate temperatures with rf power kept
constant at 100W.
4.21
91
Luminescence spectrum of amorphous silicon at
various substrate temperatures with rf power kept
constant at 150W.
4.22
91
Luminescence spectrum of amorphous silicon at
various substrate temperatures with rf power kept
constant at 200W.
4.23
92
Luminescence spectrum of amorphous silicon at
various substrate temperatures with rf power kept
constant at 250W.
4.24
92
Flow diagram representing the interrelation of the
excitation and recombination processes appropriate
to amorphous silicon.
4.25
94
Schematic plot of the relaxation and recombination
probabilities versus energy from midgap.
95
xvii
LIST OF SYMBOL
VB
-
Breakdown voltage
VG
-
Glow discharge voltage
Vp
-
Plasma potential
Vc
-
Cathode potential
Vf
-
Floating potential
A
-
Gas species
A*
-
Reactive species
e-
-
Electrons
e
-
Unit electron charge (1.60 × 10-19 C)
k1
-
Reaction rate coefficient
Te
-
Electron temperature
Ts
-
Substrate temperature
mi
-
Ion mass
me
-
Electron mass
k
-
Boltzmann constant (1.38 × 10-23 J/K)
∆E*
-
Activation energy for plasma enhanced reaction
∆E
-
Thermally driven activation energy
Eg
-
Optical gap
Eo
-
Urbach energy
α
-
Optical absorption
h
-
Planck constant (6.625 × 10-34 J-s)
ν
-
Frequency
B
-
Edge width parameter
A
-
Absorbance
T
-
Transmittance
Io
-
Incident light intensity
xviii
I
-
Intensity of light transmitted through sample
d
-
Film thickness
t
-
Deposition time
c
-
Speed of light (2.998 × 108 m/s)
λ
-
Wavelength of light
Ds
-
Surface diffusion
τs
-
Staying time
dr
-
Deposition rate
CH
-
Bonded H content
XC
-
Volume fraction of microcrystalline
CHAPTER 1
INTRODUCTION
1.1
Amorphous Semiconductor
Amorphous semiconductors are noncrystalline and have significantly different
characteristics than those of crystalline (Street, 1991). They lack long-range periodic
ordering of their constituent atoms. That is not to say that amorphous semiconductors
are completely disordered on the atomic scale.
Local chemistry provides almost rigorous bond-length and a lesser extent,
bond-angle constraint on the nearest-neighbor environment. Unlike amorphous
metals, amorphous semiconductor do not consist of close-packed atoms, but rather
they contain covalently bonded atoms arranged in an open network with correlations
in ordering up to the third or fourth nearest neighbors. The short-range order is
directly responsible for observable semiconductor properties such as optical
absorption edges and activated electrical conductivities.
Amorphous semiconductors are usually fabricated in the form of thin films by
an atomic deposition procedure such as evaporation, sputtering, chemical vapor
deposition, and plasma decomposition on gases or electroplating. Sometimes ion
bombardment of crystals is used to have an amorphous layer in the collision trail of
the ions.
2
1.2
Potential Applications
The commercial potential of amorphous semiconductors has encouraged many
to study their properties and preparation. In particular, hydrogenated amorphous
silicon (a-Si:H) is very versatile, low cost material that has made it desirable for the
use in many device applications. Other special attribute of a-Si:H is the ability to
deposit the material inexpensively over large areas.
In recent years, the development of thin film a-Si:H photovoltaic solar cells has
been extensively pursued because such devices offer the potential of low-cost
electricity, making them attractive as a source of utility and residential electric
power. Single-junction a-Si:H p-i-n solar cells with solar energy conversion
efficiency of 10% have been achieved by several laboratories (Shen et al. 1991). The
basic structure of a single-junction a-Si:H p-i-n cell consists of a very thin (less than
10 nm thick, p-type layer), low-defect, 200 to 600 nm thick intrinsic layer, and a thin
(about 30 nm thick) n-type layer. The construction of a basic single junction thin film
a-Si:H solar cell is illustrated in Figure 1.1. To improve efficiency and stability of aSi:H solar cells, multiple-junction solar cell structures using a-Si:H alloys are being
extensively studied (Yang et al. 1997).
The need for large-area charged particle and X-ray detectors for applications
like medical imaging and calorimetry in high-energy physics experiments have
stimulated significant investigations into using a-Si:H (50 to 70 µm thick) for such
applications (Xi et al. 1991). Other photodiode applications for a-Si:H alloys include
ultraviolet light detectors (Krause et al. 2001), edge detector for application to neural
network image sensors (Sah et al. 1990), and position sensors for telephone terminals
(Brida et al. 2002).
3
Figure 1.1: Schematic representation of a typical thin film a-Si:H solar cell on glass.
The incoming photons with an energy larger than the band gap are
absorbed in the intrinsic a-Si:H film creating holes-electron pairs.
The most attractive applications of a-Si:H technology are active matrix
displays and active matrix flat-panel imagers (AMFPIs) (Zhao et al. 1995) which are
collectively termed as active matrix arrays (Figure 1.2). Active matrix arrays contain
many individual elements commonly known as pixels, which are generally addressed
or read out by a grid structure of interconnecting lines termed gate and data lines. In
these applications, an a-Si:H thin film transistor (TFT) is used as a switching element
or pass transistor. The active matrix arrays require external chips to multiplex and
drive the large number of gate and data lines. Considering the growing applications
of the active matrix displays and imaging arrays, a low cost on-chip solution is
needed for the multiplexer and driver circuitry. Designing of the multiplexer and
driver circuits in a-Si:H technology requires the specific details of the displays and
imaging arrays.
4
Data line
Gate line
Sensor /
Display pixel
TFT
Figure 1.2: Schematic of a general active matrix array
The unique properties of a-Si:H depend primarily on the complex structure
involving different bonding configurations and on the incorporation of hydrogen in
the films. Hydrogen, being a terminator in the carbon network, plays a crucial role in
determining the properties of the films. The structure of, and the incorporation of
hydrogen in a-Si:H films are critically determined by the energy of the ionic species
and the consumption of the gas mixture in the deposition process. The ion energy can
be changed by varying the deposition parameters. Therefore, it is feasible to obtain aSi:H films with a wide range of properties by adjusting the deposition parameters in
the growth process.
5
1.3
Research Objectives
The main purpose of this work is to deposit hydrogenated amorphous silicon
(a-Si:H) thin films using silane (SiH4) gas as film precursor via plasma enhanced
chemical vapor deposition (PECVD) technique. Secondly, to study the structural
characteristics of hydrogenated amorphous silicon (a-Si:H) thin films and to acquire
better understanding of this material by characterisation techniques using surface
profiler, atomic force microscopy (AFM), UV spectrophotometer, Fourier Transform
Infrared (FTIR) spectroscopy, X-ray diffraction (XRD) and photoluminescence
spectrometer. In order to do so, the effect of deposition conditions on the grown film
properties is investigated.
1.4
Research Scope
Hydrogenated amorphous silicon thin films are deposited using rf-PECVD
with silane (SiH4) gas as film precursor under different substrate temperature and rf
power while other parameters kept constant. The films are then structurally
characterized using pre-determined characterization techniques which consist of;
1.5
•
Surface profiler
•
Atomic force microscopy (AFM)
•
UV spectrophotometer
•
Fourier Transform Infrared (FTIR) spectroscopy
•
X-ray diffraction (XRD)
•
Photoluminescence spectrometer
Layout of Thesis
This thesis is organized as follows. In Chapter 1, some of the previous related
works on hydrogenated amorphous silicon (a-Si:H) and its application are reviewed.
This chapter also ruled out the objectives of conducting the research.
6
Following the introduction chapter, the literature survey is presented in
Chapter 2. This will cover the growth of a-Si:H thin film process, the fundamental of
deposition technique which is the plasma enhanced chemical vapor deposition
(PECVD) and the reaction process in this technique.
Details of experimental methods, including fabrication of the films, PECVD
setups are given in the initial part of Chapter 3. This is followed by the
characterization techniques used, namely UV spectroscopy, FTIR spectroscopy and
photoluminescence.
Chapter 4 presents the results obtained in this work. Among the
characterization results that would be presented are the surface morphology and
deposition rate, IR transmission spectrum, optical energy gap and
photoluminescence. Effect of varying deposition parameters on the film
characteristics will also be discussed in this chapter.
Finally, the conclusions of the project are made in Chapter 5. These include the
summarization of the whole project and some recommendations for future work are
also suggested.
CHAPTER 2
LITERATURE REVIEW
2.1
Hydrogenated Amorphous Silicon (a-Si:H)
Since the use of a-Si:H for solar cells began in 1974 (Spear et al. 1975), the
use has extended to other devices such as TFT’s for liquid crystal displays, image
sensors for line scanners and image pickup tube, and photoreceptors for
electrophotography and laser printing. The rapid growth of technology for a-Si:H
based alloys is partly because of its excellent optoelectronic properties which are
made possible of hydrogenation.
Silicon is among today’s technologically most important materials. Its
abundance and unique semiconducting properties make silicon currently the most
attractive material for electronic devices. The most important applications of silicon
are for high-tech devices used in computer technology, flat panel displays and solar
cells. Pure silicon exists in monocrystalline, polycrystalline and amorphous forms. In
the first case, the material exists as one (large) single crystal. In the polycrystalline
form the material consists of many (small) single crystals packed together and
connected by grain boundaries.
On the extreme end of this classification is amorphous silicon (a-Si) that does
not exhibit any crystallinity. Although one of the defining features of the amorphous
state is the absence of long range order, there is generally a local (short range) order,
8
in which the constituting atoms are in environment very similar to the crystalline
counterpart. This means that in amorphous silicon most silicon atoms are fourfold
coordinated in a tetrahedral bonding configuration very similar to crystalline silicon.
Only due to small variations in bond angles and bond lengths some randomness in
the atomic positions is introduced, which leads to the characteristic long range
disorder.
Besides the short range order and long range disorder, defects in the atomic
structure play an important role. In its pure form a-Si contains a very high defect
density of up to 1020 cm-3 (Brodsky et al. 1969), which make the material’s
semiconductor and optical properties inadequate for applications in technology. But
in the late 1960s hydrogenated amorphous silicon (a-Si:H) was deposited from silane
(SiH4) in a plasma (Chittig et al. 1969). The incorporation of a great number of
hydrogen (10 – 20 at. %) reduces the density of defects by several order of
magnitude (Shimizu et al. 1983). High quality a-Si:H contains less than 1015 cm-3
defects. Finally it was discovered that it can be doped (Spear et al. 1974), which
made the development of devices such as solar cells and field effect transistors
possible.
In several aspects the quality of hydrogenated amorphous silicon (a-Si:H) for
use in devices is inferior when compared to the crystalline material (e.g., efficiency
of solar cells or speed of transistors), but on the other hand its production is a lot
cheaper. In addition, a-Si:H contains advantageous properties, such as stronger
absorption in the visible region of the light spectrum (leading to the fabrication of
thinner solar cells).
2.1.1
Hydrogenation of a-Si
In pure form, a-Si (unhydrogenated) is a perfect fourfold-coordinated network,
with the only deviation from the single crystal being the variations in the bonding
angles, no other defects but the tail states would be found in the bandgap. But in
reality, many silicon atoms are threefold-coordinated, leaving one dangling bond (see
9
Figure 2.1). Dangling bonds, considered by many to be the most major defects, from
localized states in the middle of the bandgap. The density of these dangling bonds in
pure a-Si is so high (>1019/cc), that is useless for electronic applications. The
incorporation of hydrogen during growth (thus hydrogenated amorphous silicon),
effectively passivates the dangling bonds and reduces the density of midgap states to
about 5×1015 to 5×1016 cm-3eV-1, thereby making a-Si:H an electronic material.
Figure 2.1: A model of hydrogenated amorphous silicon. The small ball represent
hydrogen atoms while the large ball represents silicon atoms.
a-Si:H can easily be deposited by plasma-enhanced chemical vapor
deposition (PECVD) method. This method, also known as the glow-discharge
method, offers advantages such as low deposition temperature (<300˚C), large area
deposition possibility, a wide choice of substrates, and reproducible high quality aSi:H films. The most common gas used to produce a-Si:H material is silane.
10
There are three stages in which the amorphous silicon based films grow on
the substrate. The first stage is the electron impact dissociation, ionization of silane
and chemical reactions in the plasma. The second stage consists of chemisorption,
diffusion of neutral radicals and the abstraction of H from the substrate surface. The
final stage is the elimination of H and Si-Si formation in the bulk of the film
(Knights et al. 1985). In the plasma, Silane is excited to higher-energy states through
inelastic collisions with energetic electrons and dissociated into neutral radicals and
atoms. A part of silane excitation involves the formation of ionic and emissive
species.
The radicals and ions generated in the plasma as described above will
undergo a variety of reactions with SiH4 before they reach the substrate. Several of
these reactions are listed below:
Si + SiH4 = Si2H4
(I)
SiH + SiH4 = Si2H5 = Si2H3 + H2
(II)
SiH2 + SiH4 = Si2H6
(III)
SiH3 + SiH4 = SiH4 + SiH3
(IV)
H + SiH4 = H2 + SiH3
(V)
SiH3 + SiH4 = Si2H5 + H2
(VI)
It can be seen from the above reactions especially on reaction (IV), that
radical SiH3 has very long reaction lifetime since it reacts with SiH4 and reforms as
SiH3 again. The H transfer in reaction (IV) is established by using isotopic labeling.
In this reaction, H from SiH4 (left side of the reaction (IV)) is transferred to SiH3
producing SiH4 and leaving another SiH3 (as seen in the right side of reaction (IV)).
This is the reason why that SiH3 has the highest concentration in silane (SiH4) glowdischarge (Drevillon et al. 1980). SiH2 on the other hand, reacts with SiH4 and forms
Si2H6 which has no reactivity on the growing surface (Inoue et al. 1985, Jasinski et
al. 1988). In light of SiH4’s dissociation and radical reactions, SiH3 is the primary
species reaching the a-Si:H growing surface in the SiH4 glow-discharge system, a
fact that has been experimentally confirmed by Longeway and Robertson et al
(Longeway, 1984, Robertson et al. 1986). However, other species, such as Si, SiH,
11
and SiH2, can also reach the growing surface although their concentrations are lower
than that of SiH3 and affect the structural and photoelectric properties of the resultant
film.
Radicals and atoms generated in the plasma travel to the growing surface
through the gas-phase diffusion and collision with SiH4. Some of these species are
adsorbed on the surface (adsorption); some diffuse on the surface (surface diffusion)
and make chemical bonds at their favorite sites (chemisorption); and some are
desorbed depending on the sticking coefficient on the surface. A poor quality
material results when the surface diffusion length (length for the radicals and atoms
to move over the surface before being adsorbed) is dependent on the surface
diffusion coefficient (Ds) and the time for the adsorbed radicals or atoms to stay on
the surface or staying time (τs). H-coverage factor is a critical parameter for
obtaining high Ds.
When the growing surface is totally covered with H, the radicals reaching the
surface have high Ds and find the more stable site to form a denser random network.
At high substrate temperatures, the H-coverage factor decreases due to the increase
of free Si bonds on the growing surface, resulting in the reduction of Ds. Therefore,
the maximum substrate temperature for depositing a large surface diffusion
coefficient is limited by the H-coverage factor.
SiH3 radicals adsorbed on the growing surface are incorporated in the random
amorphous silicon network through a bimolecular-like deposition reaction shown in
Figure 2.2. A hydrogen-abstraction reaction with SiH3 forms SiH4, creating a site for
a second SiH3 to diffuse on. SiHx radicals other than SiH3 react with the H-covered
or uncovered site directly through insertion.
12
SiH4
Figure 2.2: Model for the adsorption, diffusion and recombination of
SiH3 on the growing surface.
The types of species available in the gaseous stage, the transport phenomena,
and the actual growth are all dependent on the deposition parameters such as
deposition temperature, chamber pressure, applied power, gas flow rate, electrode
spacing, and frequency of the applied power.
It has been found that for high quality a-Si:H films, radicals with high surface
mobility are desirable (Luft et al. 1988). The low surface mobility (high sticking
coefficient) radicals cause growth of columnar structures which, in turn, result in
inferior electronic properties. To produce high quality films, SiH3 radicals are
preferred. The SiH3 radical has a low sticking coefficient and high surface mobility.
13
2.2
2.2.1
Conventional Thin Film Deposition Techniques
Chemical Vapor Deposition
In the chemical vapor deposition process (CVD), a solid-phase reactant
product (thin film) is deposited on a heated substrate surface from appropriate
reactant vapor by a chemical reaction involving direct elemental combination. The
substrate material, substrate temperature, composition of the reactant gas mixture,
flow rate, and total pressure of the gas can be varied to deposit material with
different properties. In chemical vapor deposition, gaseous reactants are admitted
into a reactor as shown in Figure 2.3, and the following type of chemical reaction
occurs between gaseous reactants near or at the heated substrate surface (Bunshah,
1993). A simple example of this reaction is when silane gas (SiH4) react with
methane gas (CH4) to form silicon carbide (SiC) solid film and leaving hydrogen gas
(H2) as byproduct.
Gaseous reactants → Solid material + Gaseous products
AX
+
BX
=
AB
+
2X
(VII)
Example of this process is as follows;
SiH4 (g) + CH4 (g) =
SiC (s) + 4H2 (g)
(VIII)
14
AX
2X
BX
AB
substrate
AX (gas)
AX (solid)
BX (gas)
X (gas)
Figure 2.3: A common implementation of CVD
(AX, X, BX – gases, AB – solid material)
There are some drawbacks to CVD, since the process often requires high
temperatures for a chemical reaction to happen. In some cases, the temperature
necessary to achieve acceptable deposition rates may be so high as to lead to
diffusion, alloying, or chemical reactions on the substrate surface. The limitation of
CVD can be overcome if an electric discharge is created in the reactant gases to
produce a significant number of free radicals, which will be much more reactive at
lower temperatures. The latter process is known as plasma enhanced chemical vapor
deposition (PECVD). Depending on the form of excitation used to create the
discharge, there are two distinct types of discharge – DC and RF.
2.2.2
DC Glow Discharge
The term DC glow discharge normally refers to the DC diode glow discharge
configuration. When a sufficiently high DC voltage is applied between the two
electrodes in a gaseous medium at reduced pressure, the atoms and molecules in the
15
medium will break down, creating electron-ion pairs and letting a current flow. A
typical current-voltage characteristics of a DC glow discharge is illustrated in Figure
2.4 (Howatson, 1976).
Figure 2.4: Current-voltage characteristics in a DC glow discharge
At low voltages, the discharge current is due to the primary electrons and ions
produced by external sources such as cosmic radiation. As the voltage is increased,
when all the charges are reaching the electrodes, the current become independent of
voltage. This is referred to as the saturation regime. Beyond the saturation regime,
the current starts to increase with increasing voltage in the Townsend regime. In this
latter regime some primary electrons gain enough energy to ionize the gas species,
producing positives ions and additional electrons.
If the voltage across the tube carrying a Townsend discharge increased, the
current starts to increase sharply, perhaps by several orders of magnitude; this is
referred to as breakdown. At this point, cathode bombardment and secondary
electron emission occur as the positive ions gain sufficient energy to travel to the
cathode and promote secondary emission and a positive feedback occurs due to the
regeneration of electrons and the discharge is self-sustained. As the current is further
16
increased, the voltage across the tube decreases until VG, the voltage required to
sustain the discharge. At the onset of the normal glow regime, a luminous glow starts
to grow at the cathode surface.
No voltage increase is required as the current continues to increase, since the
increase in current does not change the current density but only the cross-sectional
area of the discharge. After the discharge entirely covers the cathode surface, any
further increase in the discharge current results in the increase in the current density,
requiring an increase in the discharge voltage. This last condition is known as
abnormal glow discharge. To obtain uniformity in depositing thin films, the
abnormal glow regime is used for the deposition.
A low pressure glow discharge plasma contains alternate dark and bright
luminous layers, as described in Figure 2.5 (Raizer, 1991).
Cathode dark space
Faraday dark space
Anode dark space
_
+
Cathode glow
Negative glow
Positive column
Anode glow
Figure 2.5: Luminous zones and dark spaces in a DC glow discharges.
All of the luminous zones illustrated in Figure 2.5 are observed only when the
inter-electrode separation is large compared to the size of the electrodes. In practical
glow discharge systems used for thin film deposition, the inter-elecrode separation
needs to be comparable to the size of the electrode. When the electrode separation in
Figure 2.5 is decreased, the positive column shrinks. This process continues until the
17
positive column and also all the other zones disappear except the negative glow,
cathode dark space, and anode dark space.
The potential relative to the cathode increases linearly in the cathode dark
space (also called the cathode sheath), starting from the negative cathode potential
(Vc), and reaches a constant plasma potential (Vp) that is maintained in the negative
glow regime, and then drops to zero in the dark anode dark space (also called the
anode sheath). The potential distribution for such a case is shown schematically in
Figure 2.6 . Most analytical glow discharge sources are designed so that the interelectrode separation is a few times the length of the cathode dark space.
Figure 2.6: The potential distribution in a DC glow discharge. (Vp – plasma potential,
Vc – cathode potential)
The negative glow region in Figure 2.6 contains a partially ionized gas with
equal numbers of positive and negative charges, in addition to the neutral species,
and is characterized as a plasma. In this region, gas molecules undergo inelastic
collisions with energetic electrons, resulting in ionization, dissociation, or excitation
(Smith et al. 1994). The relaxation of excited molecules back towards the ground
18
state is accompanied by emission of ultraviolet or visible photons, which result in the
negative glow region (plasma).
The entire voltage drop in the glow discharge occurs at the boundaries of the
negative glow (in the sheath regions, Figure 2.6). The behavior of these regions is
important to sustain the plasma. For instance, due to the large voltage drop in the
cathode sheath, the ions are accelerated towards the cathode where the bombardment
occurs, resulting in secondary electron emission from the cathode. The same voltage
drop in the cathode sheath accelerates these secondary electrons towards the negative
glow, where they then undergo inelastic collisions to produce ions and radicals. The
sheath region is mostly dark, since the secondary electrons undergo collisions with
the neutral gas molecules only at some distance away from the cathode,
corresponding to their mean free paths. The voltage drop in the anode sheath is
relatively smaller. The opposite electric field repels some of the electrons to control
electron flow to match the electron current required by the external circuit.
2.2.3
RF Glow Discharge
The design of the electrode configuration for RF glow discharges is much the
same as for DC glow discharges, but the electrodes are powered by an AC power
supply. When the polarity of the electrodes is alternated, the mechanism of glow
discharge generation is dependent on the frequency of alternation. At low frequency,
the mechanism is simply the same as that of DC glow discharge of alternate polarity
however, as frequency increases, the motion of ions no longer can follow the changes
in the field polarity. A common operating frequency for RF glow discharge
deposition is 13.56 MHz. At this frequency, due to the large mass difference between
electrons and ions, only electrons can follow the variation in the applied field.
Therefore, the plasma can be described as an electron gas which moves back and
forth in a sea of relatively stationary ions. As the electron cloud approaches one
electrode, the ions are exposed to the other electrode forming a positive sheath where
the most of the voltage drop occurs. In the sheath region, the ions are accelerated and
bombard the electrodes.
19
Since the oscillating electrons in an RF plasma do not reach the electrodes and
no real current flows through the circuit in contrast to the DC case, the RF glow
discharge does not require conducting electrodes in contact with the plasma. The
production of the plasma is more efficient in an RF glow discharge than in a DC
glow discharge, since the ionization is more efficient as the electrons gain higher
energies as they follow oscillatory path between the electrodes.
2.3
Plasma Enhanced Chemical Vapor Deposition (PECVD)
This technique uses a plasma discharge to provide the excitation necessary for
chemical reaction to occur. PECVD is often used for silicon deposition, which will
be discussed in more detail in Chapter 3. Other applications of this technique include
the deposition of diamond, silicides and refractory metals.
2.4
2.4.1
Plasma Fundamentals
Non-equilibrium Glow Discharges
We can define a glow discharge as a partially ionized gas composed of equal
volume concentrations of positively and negatively charged species, and different
concentrations of species in the ground state and excited state (Rand, 1979,
Chapman, 1980). Plasmas can be generated by subjecting gases to very high
temperatures or to strong electric or magnetic fields. In thermal plasmas, electrons,
ions, and neutral species are in local thermodynamic equilibrium. In non-equilibrium
or “cold” plasmas, the electrons and ions are more energetic than the neutral species.
Most of the glow discharges used for thin film plasma deposition are created
by subjecting the gas to radio-frequency (rf) electric field; they are non-equilibrium
glow discharge plasmas. The electric field initially reacts most with the free electrons
20
present in the gas. While the electric field also interacts with ions, these species
initially remain relatively unaffected because of their much heavier mass. The
accelerated electrons do not loss much energy in elastic collisions with gas species
because of their large mass difference. In addition, these electrons do not lose much
energy in elastic collisions, such as excitation and ionization, unless their energies
are higher than the relevant threshold energies.
Inelastic collisions between high-energy electrons and gas species generate
high reactive species, such as excited neutrals and free radicals, as well as ions and
more electrons. By this mechanism, the energy of the electrons creates reactive and
charged species without substantially increasing the gas temperature. The reactive
species that are generated in the plasma have lower energy barriers to physical and
chemical reactions than the parent species, and consequently can react at lower
temperatures. In PECVD, these reactive species are utilized to form thin films at
temperatures lower than those possible with thermally activated CVD. The charged
species in the glow discharge may also affect the properties of the deposited films.
We can estimate the rate at which inelastic collisions generate excited
species, ions, free radicals, etc., by using a rate equation shown in Equation 2.1. For
example, the rate at which A* is created from excitation reaction
A + e- → A* + e-
(IX)
d [ A* ]
= k1[ A][e − ] ,
dt
(2.1)
Is given by
Where d[A*]/dt is the rate of formation of A*, k1 is the reaction rate
coefficient, [A] is the concentration of species A, and [e-] is the electron
concentration
Radio-frequency (RF) glow discharges used for the deposition of thin films
operate at frequencies between 50 kHz and 13.56 MHz, and at pressure of 0.1 to 2.0
torr. The plasma density (i.e., the density of ions and free electrons) is in the range of
21
108 to 1012 cm-3. The degree of ionization is typically ≤ 10-4, i.e., the glow consists
mostly of neutral species. Typical average electron energies are in the range of 1 to 3
eV, but the fastest electrons may reach energies as high as 10 to 30 eV. These highenergy electrons are responsible for generating the reactive species that lead to film
formation at relatively low temperatures. Since in these discharges the average
electron energies are much higher than the ion energies, they are known as nonequilibrium glow discharges, i.e., the PECVD environment is not in thermal
equilibrium. Consequently, thermodynamic calculations cannot reliably predict the
product of a PECVD reaction.
2.4.2
Potentials in rf Glow Discharges
There are several important potentials in glow discharge used in PECVD: (1)
the plasma potential, (2) the floating potential, and (3) the sheath potentials. The
plasma potential (Vp) is the potential of the glow region of the plasma, which is
normally considered nearly equipotential. It is the most positive potential in the
chamber and is the reference potential for the glow discharge. The floating potential
(Vf) is the potential at which equal fluxes of negative and positive charged species
arrive at an electrically floating surface in contact with the plasma. It is represented
approximately by following expression:
Vp − V f =
kTe ⎛ mi ⎞
⎟,
ln⎜
2e ⎜⎝ 2.3me ⎟⎠
(2.2)
Where Te is the electron temperature, e is the unit electron charge, and mi and me are
the ion and electron masses, respectively (Chapman, 1980). Equation 2.2 can be used
to estimate the maximum energy with which positive ion may bombard electrically
insulated chamber walls. Most sputtering threshold energies are in the range of 20 to
40 eV (Sterman, 1984). Therefore, a (Vp ― Vf) that is less than or equal to 20 V is
desirable to avoid sputtering material off the reactor walls, which otherwise may lead
to film contamination.
22
The plasma potential is always positive with respect to any surface in contact
with the plasma, because the mobility of free electrons in the plasma is much greater
than that of ions. The initial electron flux to all surfaces is therefore greater than the
ion flux. The surfaces in contact with the plasma become negatively charged, and a
positive space-charge layer develops in front of these surfaces. Because there are
fewer electrons in the space-charge layer, or sheath, fewer gas species are excited by
electron collisions. Consequently, fewer gas species relax and give off radiation, and
the sheath region is dark relative to the glow discharge.
Positive ions that enter the sheaths from the glow region by random thermal
motion are accelerated into the electrodes and other surfaces in contact with the
plasma. Similarly, secondary electrons emitted from the surfaces (due to positive ion
bombardment) accelerate through the sheaths into the glow region, is determined by
the difference between the potential of the surface and the plasma potential. This
maximum energy with which positive ions bombard a surface, and the maximum
energy with which secondary electrons enter the glow region, is determined by the
difference between the potential of the surface and the plasma potential. This
potential across the sheath is usually referred to as sheath potential.
2.5
Qualitative Model for PECVD Reactions
In thermal CVD, gaseous precursor reactants that contain the elements of the
film material to be deposited undergo a sequence of essentially seven steps that can
be shown schematically in Figure 2.7 and can be define and listed as follows:
23
(1)
Main Gas Flow Region
(1)
Gas Phase Reactions
(2)
Transport to Surface
(3)
Surface Diffusion
Adsorption of Film Precursor
(3)
Precursor Gas Molecules
(5)
Desorption of
Film
Precursor
(4) Nucleation and
Island Growth
(5)
Desorption of
Volatile Surface
Reaction Products
Step Growth
By-products
Main Radical contributing
to deposition
Figure 2.7:
Schematic of fundamental transport and reaction processes
underlying CVD
(1) Mass transport of the reactants in the gas flow region from the reactor inlet to
the deposition zone. Gas phase reactions leading to the formation of film
precursors and by-products;
(2) Mass transport of film precursors to the growth surface;
(3) Adsorption of the film precursors on the substrate surface. Surface diffusion
of film precursors to growth sites;
(4) Physical-chemical reactions leading to the solid film and reaction byproducts;
(5) Desorption of film precursor and by-products, mass transport of by-products
to the main gas stream and out of the reaction chamber.
An equivalent circuit representation of this sequence of steps is shown in Figure 2.8
(Reif et al. 1979).
24
Gas Phase
R1
Gas Phase
R2
Solid Surface
R3
R4
R7
Figure 2.8:
R6
R5
Equivalent circuit representation of the sequence of steps in thermally
driven CVD. The Ri (with i = 1-7) represent the seven steps describe
in the text.
When plasma is generated for CVD applications, a fraction of the groundstate precursor species in the gas phase undergoes electron impact dissociation and
excitation that generate reactive species. In addition to unchanged ground-state
species, these highly reactive species also diffuse to the surface and undergo a
similar sequence of processes. These highly reactive follow an alternate deposition
pathway that operates in parallel to the thermal pathway. An equivalent circuit
representation of this situation is represented in Figure 2.9 (Reif et al. 1990).
25
Gas Phase
Gas Phase
Solid Surface
R2
R3*
R2
R3
*
R1
R4*
R4
R7
R5
R6
R5*
R6*
Figure 2.9:
Equivalent circuit representation showing the thermal CVD path, and
the parallel plasma-enhanced path represented by the Ri*s.
The plasma kinetic pathway often by passes that of the ground-state species
because the sticking coefficients of the highly reactive species generated by the
plasma are closer to unity (Rand, 1979), and the activation energies for chemical
dissociation are typically lower, as illustrated in Figure 2.10. The diagram compares
the activation energy of a ground-state reaction,
A → B
(X)
which has an activation energy of ∆E, with that of
A + e fast → A* + e slow ,
*
A
*
→ B
(XI)
(XII)
with an activation energy of ∆E* (Reif et al. 1990). Typically, ∆E* < ∆E and,
consequently, the plasma kinetic pathway leads to a higher deposition rate. Some
heat is still needed to drive the reaction over ∆E*, as indicated in Figure 2.10, thereby
providing the energy required to promote surface reactions, desorb byproducts, lower
film contamination, and minimize absorption and inclusion of gasses in the film.
26
A*
∆E*
∆E
A
B*
B
Figure 2.10:
Activation energy diagram for a thermally driven (solid line) and
plasma enhanced (dashed line) reaction. The activation energy for the
plasma enhanced reaction, ∆E*, is typically less than that for the
thermally driven reaction, ∆E.
Ions present in the plasma may bombard the substrate surface, further
modifying the kinetic pathway by breaking down weakly bonded reactive species,
effecting the surface migration of adsorbed atoms (adatoms), and/or removing
undesired contaminants. Ions energies, fluxes, or doses that are too high may also
affect the film quality.
CHAPTER 3
METHODOLOGY
3.1
3.1.1
Deposition of Amorphous Silicon Films
Introduction
Any thin film deposition process consists of three main steps: (1) production of
the appropriate atomic, molecular, or ionic species, (2) their transport to the substrate
through a medium, and (3) deposition on the substrate to form a solid film. When
depositing a thin film, one has to choose appropriate conditions over a broad range of
parameters involved in the above three steps, and a method of deposition to achieve a
film with desirable properties.
3.1.2 Plasma Enhanced Chemical Vapor Deposition
Since the earlier work on plasma enhanced chemical vapor deposition
(PECVD) was reviewed by Hollahan and Rosler et al. (1978), it has become an
established commercial technique for the deposition of a number of important
materials, especially insulating films. The major advantage of PECVD is its lower
temperature capability compared to that of thermally driven CVD. For example,
deposition temperatures of 700˚ to 900˚C are required to deposit silicon nitride films
by thermal CVD, whereas only 250˚ to 350˚C is sufficient to deposit similar films by
28
PECVD (Gorowitz et al. 1985, Adams, 1983). This lower temperature capability is
made possible by the addition of electrical energy to the CVD environment, and the
effective substitution of this electrical energy for thermal energy. Applications of
PECVD thin films and coating range from electronics to optics metallurgy.
3.1.3 PECVD Setup
The whole PECVD setup in this studies consists of different interdependent
part which include non pure nitrogen (NPN2) and pure nitrogen (PN2) gas rack, high
purity argon gas rack, silane (SiH4) gas cabinet, PECVD system and scrubber. The
non pure nitrogen and pure nitrogen are supplied to gas cabinet and PECVD machine
for purging purposes while argon and silane are the reactant gas used for deposition
of thin films.
Figure 3.1 and 3.2 shows the non pure nitrogen and pure nitrogen gas racks.
The silane gas cabinet is shown in Figure 3.3. This gas cabinet is equipped with gas
detector and alarm system to avoid any possibility of gas leakage within the cabinet
or on the exhaust line. All the nasty by product from the PECVD machine is first
channeled through a scrubber shown in Figure 3.4 for treatment before being
released from the exhaust line to the atmosphere.
29
Figure 3.1: NPN2 gas
Figure 3.2: PN2 gas rack
30
Figure 3.3: Silane (SiH4) gas cabinet
Figure 3.4: Scrubber
31
Figure 3.5 shows the schematic of the reactor system used in the experiment.
This reactor is a parallel-plate capacitively-coupled, 13.56 MHz, inward radial flow
type reactor. The plasma is generated between the two parallel, circular electrodes.
The constant power level during deposition is maintained by the equipped
automatic/manual power sensing and tuning circuits. Substrates are loaded on the
bottom, electrically grounded electrode which is made of hard anodized aluminum.
Substrates can be heated by a contact heater and the substrate temperature can
be sensed and controlled by a thermocouple and a temperature controller. The rf
power is applied to the top electrode, which is of bare aluminum, through a matching
network. The precursor gases are introduced into the chamber from the periphery of
the bottom electrode. The gases flow radially inward over the substrates in the
reaction zone and are exhausted through a pumping port at the centre of the bottom
electrode. Gas flow is controlled by electronic mass flow controllers. For the a-Si:H
films being studied, the primary reactant gas was electronic grade silane (SiH4).
Insulated rf input
Samples
Cylinder
wall
Plasma
Aluminum
electrodes
Gas
inlet
Heated
sample
holder
Pump
Gas
inlet
Figure 3.5: Schematic of the rf PECVD reactor system
32
Figure 3.6: The rf PECVD reactor system
33
3.1.4 PECVD System Deposition Process
The a-Si:H thin films in this study were grown in an rf PECVD system
PentaVacuum. The whole deposition process was done following a list of sequence.
Prior starting up the PECVD system, the scrubber should be switched on and NPN2
gas, PN2 gas and SiH4 gas should be made sure opened. The deposition process from
loading of samples to unloading of the samples is as follows;
1. Samples were loaded into the reactor (on a sample stage).
2. Reactor door was closed by pressing the door control button.
3. Choose PECVD OPERATION on the main menu of display screen.
4. The desired reactor pressure (deposition pressure) was set by turning the dial
on throttle valve controller on the PECVD system.
5. Substrate temperature was set on the temperature controller. (Substrate
temperature can only reach a maximum temperature of 600°C)
6. RF power was set by turning a knob on the rf matching panel.
7. Flow rate of precursor gas (silane, SiH4) was set from the flow rate controller.
Prior to that, precursor gas type (silane, SiH4) was set on the PECVD process
parameter screen.
8. Deposition time was set on the display screen.
9. The reactor was pumped down for 15 minutes by touching the ROUGH
CHAMBER icon on the display screen to its base pressure 12-14 mTorr
(roughing process). This is done via mechanical pump in the PECVD system
itself just to ensure no remaining gas or water vapour being trapped in the
chamber and system pipelines.
10. Once this base pressure is reached, the reactor will be further vacuumed down
to 10-6 Torr by selecting SELECT TURBO icon on the display screen. This is
done via turbomolecular pump in PECVD system.
11. Once the ultra-high vacuum pressure is being reached, READY signal will
appear on the display screen.
12. Deposition process was initiated by touching PROCESS START icon on the
display screen. The whole deposition process including admitting the
34
precursor gas into the reactor is done automatically following the
predetermined parameters.
13. When deposition is done, the reactor will be pressurised automatically to
atmospheric pressure. All remaining by-product gas will be treated via
scrubber before being released through an exhaust to atmosphere.
14. The reactor door was open by pressing the door control button and samples
were unloaded from reactor.
3.1.5
Preparation of a-Si:H Samples
3.1.5.1 Substrate Preparation
Corning glass 2947 substrates with the dimension of 76.2 × 25.4 mm were
used as substrates for depositing the films. Since chemical bonding forces extend
only a few tenths of a nanometer, only one monolayer of poorly bonded contaminant
can be sufficient to prevent bonding of the depositing film to the substrate, causing
poor adhesion between the film and the substrate.
Also, contaminants can affect the structure and optoelectronic properties of
the film. Contaminants, largely water and oil, come mainly from the atmosphere and
from machining and handling; they can be removed with degreasing solvents such as
chromic acid and acetone. The substrates for the film depositions were cleaned
ultrasonically in chromic acid and deionized water for 45 minutes in each solvent and
were finally washed with deionized water and blow-dried.
3.1.5.2 Deposition Parameters
The major concern in the deposition of a-Si:H films studied in this thesis was
the effect of substrate temperature and rf power on the structural properties of a-Si:H
thin films
35
The substrate temperature was varied from room temperature which referred
to as 27°C to 300˚C to produce a series of a-Si:H films, as summarized in Table 3.1.
For each substrate temperature, samples were prepared under different rf power
ranging from 50 W to 250 W. Except for the substrate temperature and rf power,
other preparation conditions were the same for all samples; e.g. chamber pressure =
150 mTorr, SiH4 flow rate = 20 standard centimeter cubic per minute (sccm) and
deposition time = 60 minutes.
All deposition parameters was chose based on a typical deposition of a-Si:H
films via PECVD by other researchers like Nishikawa et al. (1985) and Lucovsky et
al. (1979). These parameters were chosen in order to get a suitable and good quality
a-Si:H thin films. According to Lucovsky et al. (1979), smooth and homogeneous
film deposition occurs when gas pressure lies between 100 mTorr to 200 mTorr.
Typical flow rate of SiH4 for a-Si:H deposition are between 10 sccm to 20 sccm. If
the rate is increased further, the film will not evenly deposited onto the substrate and
yellowish powder will be developed around the chamber wall and also on the film.
So chamber pressure of 150 mTorr and silane flow rate of 20 sccm was chosen.
Prior deposition, all samples were pre-cleaned with Argon plasma in the
deposition chamber for about 10 minutes in order to eliminate any contaminants on
the substrate’s surface.
Table 3.1: The preparation conditions for a-Si:H thin films under different RF power
and substrate temperature. Note that other parameters are kept constant.
Saple
RF Power
(Watt)
Substrate
Temperature
(˚C)
Chamber
Pressure
(mTorr)
SiH4 Flow
Rate (sccm)
QR50
50
27
150
20
QR100
100
27
150
20
QR150
150
27
150
20
QR200
200
27
150
20
36
QR250
250
27
150
20
SR50
50
100
150
20
SR100
100
100
150
20
SR150
150
100
150
20
SR200
200
100
150
20
SR250
250
100
150
20
PR50
50
200
150
20
PR100
100
200
150
20
PR150
150
200
150
20
PR200
200
200
150
20
PR250
250
200
150
20
MR50
50
300
150
20
MR100
100
300
150
20
MR150
150
300
150
20
MR200
200
300
150
20
MR250
250
300
150
20
To investigate the repeatability of the samples in Table 3.1, three additional
samples were deposited under the same conditions as for samples QR, SR, PR and
MR. Thickness, Photoluminescence, Infrared and UV measurements yielded results
which were essentially the same.
3.2
Analytical Tools
3.2.1 Spectroscopy
Spectroscopy is the use of the absorption, emission, or scattering of
electromagnetic radiation by atoms or molecules (or atomic or molecular ions) to
qualitatively or quantitatively study the atoms or molecules or to study physical
processes. More recently the definition has been expanded to include the study of the
37
interactions between particles such as electrons, photons, and ions, as well as their
interaction with other particles as a function of their collisions energy. The
interaction of radiation with matter can cause redirection of the radiation and/or
transitions between the energy levels of the atoms or molecules. A transition from a
lower level to a higher level with transfer of energy from incident radiation, to the
atom or molecule is called absorption. A transition from a higher level to a lower
level is called radiative decay if photon energy is transferred or emitted form the
transition. Whereas it is called non-radiative decay if no radiation is emitted.
Redirection of light due to its interaction with matter is called scattering, and may or
may not occur with transfer of energy, i.e. the scattered radiation has a slightly
different or the same wavelength.
3.2.2
Absorption
When atoms or molecules absorb light, the incoming energy excites a
quantized structure to a higher energy level. The type of excitation depends on the
wavelength of the light. Electrons are promoted to higher orbital by ultraviolet or
visible light, vibrations are excited by infrared light, and microwaves excite
rotations.
An absorption spectrum is the absorption of light as a function of wavelength.
The spectrum of an atom or molecule depends on its energy level structure, and
absorption spectra are useful for the identifying of compounds.
3.2.2.1 Optical Absorption
Optical absorption measurements can give information about optical band
gap Eg and Urbach energy Eo which ia related to the width of the band tails or
disorder in the film. The optical absorption curves of amorphous materials exhibit
three distinct regions, as shown in Figure 3.7: a slowly varying region (region A) for
38
α ≤ 1 cm-1, an exponential rise (region B) for 1 cm-1 ≤ α ≤ 103 cm-1, and a shoulder
(region C) for α ≥ 103 cm-1.
Optical transitions between the valence and conduction bands are responsible
for the optical absorption in region A. Therefore, the optical gap can be deduced
from the shape of optical absorption in this region. There is no precise location of the
gap for amorphous materials because the band tail density of states decays
continuously with energy. So the band gap can only be defined in terms of an
extrapolation of the bands. Taking this approach, Tauc et al. (1966) introduced a
simple model for the band gap which is given by
αhν = B (hν − E g )2
(3.1)
This relationship is written as;
(αhν )
1
2
=B
1
2
(hν − E )
g
(3.2)
where α is the absorption coefficient, B is constant proportional to the edge width
parameter, hν is the photon energy, and Eg is the optical gap of the sample. From the
above equation, the optical gap can be obtained by plotting the square root of the
product of the absorption coefficient and the photon energy versus photon energy
((αhν)1/2 vs hν). The intercept on the abscissa is obtained by extrapolating the linear
portion of the plot, yielding the optical band gap.
α (cm-1)
39
Eg
E (eV)
Figure 3.7: A schematic of the optical absorption curves of amorphous materials.
40
The exponential rise in region B reflects the tailing of states into the gap due
to fluctuations of bond lengths and bond angles (Janai et al. 1981). The optical
coefficients in this region is given by (Urbach, 1958)
α (hν ) = α 0 exp[(hν − E1 ) / E0 ]
(3.3)
where E0 is the Urbach energy, and E1 and α0 are experimentally determined factors.
From the above equation, Urbach energy can be obtained from the plot of the
absorption coefficient versus photon energy in the vicinity of the band gap energy.
The origin of absorption in region C is related to midgap defect states. Since
the optical transitions involving defects states are very weak due to low defect
densities as compared to the densities above the bands, optical absorption technique
can not be used in this region due to low sensitivity of this technique.
Direct Gap
Indirect Gap
Conduction
Band
Electron Energy, E
Electron Energy, E
Conduction
Band
Valence
Band
∆k = 0
k
Valence
Band
∆k ≠ 0
k
Figure 3.8: Electron transition for semiconductors having a direct gap and
indirect gap
41
Electron excitation by light radiation (photon energy, hν) from valens band to
conduction band in semiconductor can occur by direct electronic transition or
indirect transition depending on the type of material and band structure. In direct
transition, electron located at the maximum position of valens band can easily
excited to the adjacent minimum conduction band at the same momentum
coordinates while maintaining momentum conservation (transition takes place with
no change in momentum).
In an indirect transition, electron movement from maximum valens band will
deviate to the nearest minimum conduction band. In order for the electron to change
its momentum, another particle which is phonon (lattice vibration) must be involved
for energy and momentum conservation. The difference between the direct and
indirect case is illustrated in Figure 3.8.
In order to determine the band gap energy, these transitions (direct & indirect
transition) can be characterize by Tauc et al.(1966) relation. By using equation (3.1),
the relationship can be rewritten as;
αhν = B(hν − E g )m
(3.4)
Where α is the absorption coefficient, B is the edge width parameter, hν is photon
energy, Eg is photon energy and m is constant. When m in equation (3.4) equal to ½
or 2, the transition would be allowed direct and allowed indirect respectively. While
when m is equal to 3/2 or 3, the transition would be forbidden direct and forbidden
indirect respectively. For a-Si:H thin film, the transition would be allowed indirect
electronic transition (Janai et al. 1981).
Optical absorption measurements were carried out using a conventional two
beam spectrometer (Shimadzu UV-3101-PC). For these measurements, films
deposited on Corning glass substrates were used along with the reference substrate.
The spectrometer actually measures the absorbance, A, or in this experiments the
transmittance, T data was used, which is defined by
42
A = log
T=
I0
I
I
I0
(3.5)
(3.6)
where I0 and I are the incident light intensity and the light transmitted through the
sample, respectively. The light intensity transmitted through the film follows the
Beer-Lambert’s law, which is given by
I = I 0 e − αd
(3.7)
where d is the film thickness. The absorption coefficient, α, of the film can be
obtained from the combination of equations (3.6) and (3.7),
α=
1 ⎛1⎞
ln⎜ ⎟
d ⎝T ⎠
(3.8)
3.2.2.2 UV-3101-PC Spectrophotometer
Shimadzu UV-3101-PC Spectrophotometer (see Figure 3.9) is used to
analyze optical characteristics such as transmittance, absorbance and reflectance.
Ample optical energy is available for various measurements over a wide wavelength
of 190nm (ultraviolet) to 3200nm (near-infrared) with detector of photomultiplier or
PbS cell. A glass substrate is used as reference to compensate the difference of light
path passing through the film.
The spectrophotometer is consists of optical source where the light beam
from the deuterium lamp or halogen lamp is condensed by the light source switching
mirror and reflected to the pre-monochromator. Prior measurement, the light source
should be let for about 15 minutes to obtain thermal equilibrium. Pre-monochromator
is consists of holographic grating which dispersing the light from its source into
different wavelength components.
43
The monochromatic light beam will pass through an intermediate slit and
enters the second monochromator. The light beam dispersed by the second
monochromator si focused on an exit slit as a sharp dispersion image. In this double
monochromator system, the dispersion in the near-infrared region is ¼ times that in
the visible/ultraviolet region. Therefore, the spectral band width in the near-infrared
region is 4 times that in the visible/ultraviolet region if the same slit is used, thereby
providing sufficient optical energy. All gratings are blazed holographic gratings
which provide the monochromator with high optical energy and decrease stray light
to a minimum.
The monochromatic light beam is passed into the double-beam chopper.
Here, the light beam is chopped by the chopper mirror into sample beam and
reference beam (50Hz and 60 Hz), and then passes through the sample compartment
to the detector.
The light beam passing into the detector is detected by the photomultiplier
PM (in the visible/ultraviolet region) or PbS cell (in the near-infrared region) and be
converted into presentable data in the electronic system.
44
Figure 3.9: Shimadzu UV-3101-PC Spectrophotometer
45
3.2.3
Infrared (IR) Absorption Spectroscopy
3.2.3.1 Introduction
IR spectroscopy is the measurement of the wavelength and intensity of the
absorption of the mid-infrared light by example. Mid-infrared light is energetic
enough to excite molecular vibrations to higher energy levels. The wavelength of IR
absorption bands is characteristic of specific types of chemical bonds, and IR
spectroscopy finds its greatest utility for identification of organic and organometallic
molecules.
This technique covers the region of electromagnetic spectrum between the
visible (wavelength of 800 nanometers) and the short-wavelength microwave (0.3
millimeter). The spectra observed in this region are primarily associated with the
internal vibrational motion of molecules, but few light molecules will have rotational
transitions lying in the region. For the infrared region, the wavenumber (the
reciprocal of the wavelength) is commonly used to measure energy. Infrared
spectroscopy historically has been divided into three regions, near infrared (4000 –
12500 cm-1), the mid-infrared (400 – 4000 cm-1) and the far infrared (10 – 400 cm-1).
With the development of Fourier-transform spectrometers, this distinction of areas
has blurred and the more sophisticated instruments can cover from 10 to 25000 cm-1
by an interchange of source, beam splitter, detector, and sample cell.
3.2.3.2 Mechanism of IR Absorption
There must be a change in dipole moment during the vibration for a molecule
to absorb infrared radiation.
46
O=C=O
O = C = O
O = C=O
Equilibrium
position
O = C = O
O = C = O
O=C = O
symmetric stretch
1340 cm-1
asymmetric stretch
2350 cm-1
Figure 3.10: Examples of infrared active and inactive absorption bands in CO2.
As seen in the Figure 3.10 there is no change in dipole moment during the
symmetric stretch vibration and the 1340 cm-1 band is not observed in the absorption
spectrum (the symmetric stretch is called infrared inactive). There is a change in
dipole moment during the asymmetric stretch and the 2350 cm-1 band does absorb
infrared radiation (the asymmetric stretch is infrared active). A related vibrational
spectroscopic method is Raman spectroscopy, which has different mechanism and
therefore provides complementary information to infrared absorption. Modern IR
instruments more commonly use Fourier-transform techniques with a Michelson
interferometer.
3.2.3.3 FTIR – Fourier Transform Infrared Spectroscopy
Fourier Transform Infrared Spectroscopy (FTIR) is a powerful tool for
identifying types of chemical bonds in a molecule by producing an infrared
absorption spectrum that is like a molecular “finger print”.
FTIR is most useful for identifying chemicals that are either organic or
inorganic. It can be utilized to quantitate some components of an unknown mixture.
It can be applied to the analysis of solids, liquids, and gasses. The term Fourier
Transform Infrared Spectroscopy refers to a fairly recent development in the manner
in which the data is collected and converted from an interface pattern to a spectrum.
47
Today’s FTIR instruments are computerized which makes them faster and more
sensitive than the older dispersive instruments.
Qualitative analysis
FTIR can be used to identify chemicals from spills, paints, polymers,
coatings, drugs and contaminants. FTIR is perhaps the most powerful tool for
identifying types of chemical bonds. The wavelength of light absorbed is
characteristic of the chemical bond. By interpreting the infrared absorption spectrum,
the chemical bonds in a molecule can be determined. FTIR spectra of pure
compounds are generally so unique that they are like a molecular “finger print”.
While organic compounds have very rich, detailed spectra, inorganic compounds are
usually much simpler. For most common materials, the spectrum of an unknown can
be identified by comparison to a library of known compounds.
Physical Principal
Molecular bonds vibrate at various frequencies depending on the elements
and the type of bonds. For any given bonds, there are several specific frequencies at
which it can vibrate. According to quantum mechanics, these frequencies correspond
to the ground state (lowest frequency) and several excited states (higher frequencies).
One way to cause the frequency of a molecular vibration to increase is to excite the
bond by having it absorb light energy. For any given transition between two states
the light energy (determined by the wavelength) must exactly equal the difference in
the energy between the two states (usually ground state (Eo) and the first excited state
(E1)).
Difference in Energy States = Energy of Light Absorbed
E1 − E0 =
hc
λ
(3.9)
48
where;
h = Plank constant
c = speed of light
λ = wavelength of light
The energy corresponding to these transitions between molecular vibrational
states is generally 1-10 kilocalories/mole, which corresponds to the infrared portion
of the electromagnetic spectrum.
The infrared spectroscopy characterization of the deposited samples in this
study was done using a Fourier transform infrared spectrometer (Perkin Elmer
Spectrum GX) as shown in Figure 3.11 within the region 400 to 3000 cm-1.
Figure 3.11: Fourier transform infrared spectrometer (Perkin Elmer Spectrum GX)
49
3.2.4
Photoluminescence
Photoluminescence (PL) is a useful technique to study localized states within
the band gap of a semiconductor. Therefore, this technique is particularly applicable
to amorphous semiconductor such as a-Si:H, since most of the optical and electronic
properties of this material are determined by the localized states introduced by the
intrinsic disorder or by specific defects.
For a-Si:H, PL can only be detected at low temperature. Increase in
temperature causes a strong thermal quenching of PL. At room temperature, PL is
totally quenched. There is a universal agreement that in samples with a high quantum
efficiency, there is a single luminescence band. The peak of this band is usually
between 1.25-1.4 eV. This band is attributed to the transition between conduction
and valence bandtail states, because the energy is the correct range of the bandtails.
The second well characterized transition is seen at 0.8-0.9 eV. The PL intensity of
this peak is very low compared to the main peak. This peak is attributed to a
transition between the bandtail states and dangling bond defect states in the midgap
and correlates very well with the defect density.
3.2.4.1 Basic Theory of Photoluminescence
Photoluminescence is the radiation emitted by the recombination process
following by illumination. PL process comprises three distinct events in sequence as
illustrated in Figure 3.12. First, an electron and hole are promoted to the extended
states by the absorption of a photon under illumination. The excess energy of
electron and hole is lost by many transitions within the band. And usually they relax
down in band edge localized states. This process is referred to as thermalization.
Finally, the electron and hole complete recombination either radiatively or nonradiatively.
50
Figure 3.12: A schematic of the PL process.
The thermalization process occurs by the emission of photons whose
frequencies are of the order of ~1012-1013 s-1. Consequently, the thermalization
process in the extended states is extremely fast, and occurs in the time scale of 10-13 s
(Street, 1991). Recombination lifetimes are much longer than the thermalization
times, therefore these two process occur on distinctly different time scales. If the
recombination is radiative, the process is accompanied by the emission of a photon.
Detection of this process is the basis of the PL experiment. The radiative
recombination can be geminate or non-geminate. If the recombination occurs
51
between the electron hole pair created by the same photon, the recombination is
called geminate. This process is caused by the strong Coulomb attraction between the
geminate pairs. If the Coulomb attraction between the electron and hole pair is not
strong enough, the particles diffuse apart. Then the recombination is between distant
pairs. This process is called non-geminate.
There are four types of non-radiative mechanisms that could account for nonradiative recombination. These are thermal quenching, tunneling to defects, Auger
recombination and surface recombination. Thermal quenching accounts for the
decrease in quantum efficiency of PL with increasing temperature. The primary
cause for thermal quenching is the increased mobility of carriers. With an increase in
temperature, the carriers become more mobile and can find non-radiative centers
more easily.
Non-radiative tunneling to defects accounts for the decrease in intensity of PL
in samples with high defect density. With an increase in defect density, the average
separation between the defects decreases. If this separation becomes smaller than the
average separation between the photogenerated carriers, the carriers can easily tunnel
to defects and eventually recombine non-radiatively.
Auger recombination occurs when the recombination of an electron-hole pair
excites a third carrier up into the band, instead of giving up a photon. Auger
recombination is only important at high excitation intensities because it needs a third
carrier to participate. Surface recombination occurs when there is a high density of
recombination centers at the surface to provide a non-radiative recombination path.
3.2.4.2 Luminescence Spectrometer LS 55
Luminescence spectrometer LS 55 (see Figure 3.13) is used to measure
excitation and emission for analyzing samples in the form of solid (approximate size
1.5 cm × 1.5 cm) or liquid and powder (both to be stored in a sample holder). Prior
52
measurement, sample holder should be cleaned with acetone to eliminate any
potential impurities.
Figure 3.13: Photoluminescence spectrometer LS 55
A luminescence spectrometer is consisting of few components which are (see
Figure 3.14);
•
Light source
Light source with high intensity is needed. It can be from the deuterium lamp,
halogen lamp or even laser.
•
Monochromator
The monochromator used is a double monochromator of grating to grating
type which is composed of the first monochromator (pre-monochromator)
having 3 gratings and the second monochromator (main monochromator)
having 3 gratings. It’s able to record excitation spectrum (absorption),
emission (fluorescence), and synchronous (both emission and excitation
spectrum simultaneously)
53
•
Sample cell
Rectangular or cylindrical cell made of glass or silica
•
Detector
A signal multiplier is needed to amplified the small fluorescence signal. For
example photomultiplier and photodiode.
•
Recorder
For each data collection cycle spectral data are obtained from the
photomultiplier. The data signals undergo integration, conversion, averaging,
digital filtering and ratioing before the computer receives the data. The
computer will then process the data using available software to make the data
presentable and to be printed.
Figure 3.14 shows the components diagram of the luminescence
spectrometer. It uses two beams optical system to provide more stability to the light
power source. Light will passes through the first excitation monochromator. The
monochromatic light beam with a certain wavelength will excite the sample.
Fluorescence light generated from the sample will disperse to every direction.
Nevertheless, detection of the light is made 90˚ with the light direction since on the
other direction there will be disturbance from light beam passing through the sample
or light caused by dispersion or reflection.
The fluorescence light will passes through the emission monochromator and
light with a certain wavelength will be transmitted to the detector and be recorded.
Reference light signal will passes through a light switching system to reduce the light
intensity so that it matches with fluorescence light from the sample. The light beam
will then be detected and recorded in the electronic system.
54
Light
source
Excitation
monochromator
Sample
cell
Emission
monochromator
Light
switching
system
Reference
detector
Detector
Recorder
Figure 3.14: Luminescence spectrometer schematic diagram
3.2.5
Atomic Force Microscopy (AFM)
The atomic force microscope (AFM), was invented in 1986 by Binnig, Quate
and Gerber. The atomic force microscope is ideal for quantitatively measuring the
nanometer scale surface roughness and for visualizing the surface nano-texture on
many types of material surfaces. Advantages of the AFM for such applications are
derived from the fact that the AFM is non-destructive and it has a very high three
dimensional spatial resolution.
55
Like all other scanning probe microscopes, the AFM utilizes a sharp probe
moving over the surface of a sample in a raster scan (scanline). In the case of the
AFM, the probe is a tip on the end of a cantilever which bends in response to the
force between the tip and the sample . Today, most AFMs use a laser beam
deflection system where a laser is reflected from the back of the reflective AFM
lever and onto a position-sensitive detector (see figure 3.15). AFM tips and
cantilevers are microfabricated from Si or Si3N4. Typical tip radius is from a few to
tens of nm.
Cantilever
Tip
Figure 3.15: Beam deflection system, using a laser and photodetector
to measure the beam position.
Since the cantilever obeys Hooke's Law for small displacements, the
interaction force between the tip and the sample can be found (see Figure 3.16). The
movement of the tip or sample is performed by an extremely precise positioning
device made from piezo-electric ceramics, most often in the form of a tube scanner.
The scanner is capable of sub-angstrom resolution in x-, y- and z-directions. The zaxis is conventionally perpendicular to the sample.
56
Figure 3.16: Force between tip and sample surface
The way in which image contrast is obtained can be achieved in many ways.
The three main classes of interaction are contact mode, tapping mode and non-
contact mode.
3.2.5.1 Contact Mode
In contact mode the AFM tip makes soft physical contact with the sample. As
the tip approaches the sample surface the interatomic forces become very strongly
repulsive and, since the cantilever has a low spring constant (lower that the effective
spring constant holding the atoms of the sample together), the forces will cause the
cantilever to bend following the topography of the sample. Therefore, the detection
of the position of the cantilever leads to a topographic map of the sample surface
57
3.2.5.2 Non-contact Mode
In non-Contact AFM the cantilever is vibrated near the surface of the sample.
The spacing between the tip and the sample is of the order of tens of nanometres.
Non-contact AFM provides a means for measuring sample topography with little or
no contact between the tip and the sample: this is advantageous for studying soft or
elastic samples like bio-molecules.
The system vibrates a stiff cantilever near its resonant frequency (typically in
the range from 50 to 200 kHz) with amplitude of a few tens of nanometres. Then it
detects changes in the resonant frequency or vibration amplitude as the tip comes
near to the sample surface and this changes in resonant frequency can be used to
measure changes in the sample topography.
3.2.5.3 Tapping Mode
When operated in air or other gases, the cantilever is oscillated at its resonant
frequency (often hundreds of kilohertz) and positioned above the surface so that it
only taps the surface for a very small fraction of its oscillation period. This is still
contact with the sample in the sense defined earlier, but the very short time over
which this contact occurs means that lateral forces are dramatically reduced as the tip
scans over the surface. When imaging poorly immobilized or soft samples, tapping
mode may be a far better choice than contact mode for imaging.
3.2.5.4 Image Display
Height image data obtained by the AFM is three-dimensional. The usual
method for displaying the data is to use a colour mapping for height, for example
58
black for low features and white for high features. A popular choice of colour
scheme is shown in Figure 3.17.
Figure 3.17: Colour mapping scheme for height
3.2.5.5 Surface Roughness
Surface and area roughness parameters are meant to help quantify the surface
texture of a material. Initially such parameters were used for characterizing machined
surfaces, however; now they are used for characterizing all types of high technology
and nano-materials. The values of roughness of film surface are often termed as rootmean-square surface roughness or rms. Small value of rms (example rms: 0.5 or 1.2)
indicates smooth film. As the value of rms increases, it indicated that the surface
roughness of film also increases.
In this study, Atomic Force Microscopy (AFM) (Digital Instruments
Nanoscope II System) was used to characterize the morphology of film surface. It
digitally displays images of film surface and calculated the surface roughness values
through software provided.
3.3
Film Thickness Measurements
Thickness of the films were measured using Dektak3 surface profiler as shown
in Figure 3.18. It consists of stylus (with a 25 µm tip diameter) which actually probes
59
the surface of the samples and accurately measures step heights from below 100 Å to
over 50 microns. Measurements are made electromechanically by moving the sample
beneath a diamond-tipped stylus. The high precision stage moves a sample beneath
the stylus according to a user-programmed scan length and speed. The stylus is
mechanically coupled to the core of an LVDT (Linear Variable Differential
Transformer).
Figure 3.18: Dektak3 surface profiler
As the stage moves the sample, stylus rides over the sample surface. Surface
variations cause the stylus to be translated vertically. Electrical signals corresponding
to the stylus movement are produced as the core position of the LVDT changes
respectively. An analog signal proportional to the position change is produced by the
LVDT, which in turn conditioned and converted to a digital format through a high
precision, integrating analog to digital converter. The digitized signals from a single
scan which gives the thickness measurements are stored in computer memory for
display, manipulation, measurement and print.
60
Prior measurement, a step must first be created in order to provide surface
variations on the sample surface. Two methods can be used, I) adhesive tape II)
buffered oxide etch (BOE). If the film is thick enough, a step can be created using
adhesive tape (such as Scotch tape) by peeling off some of the film from the sample
deposited on the glass substrate.
However, if the above method fails, BOE can be used to create a step. Masking
the Si wafer was done using nail polish. Raising the temperature of the BOE
increases the etch rate. The nail polish was removed using acetone. The samples
were blown dry with N2.
For this work the first method which uses adhesive tape was used to create a
step on each samples deposited.
CHAPTER 4
RESULTS AND DISCUSSIONS
4.1
Surface Morphology and Deposition Rate
The film thickness was measured using a Dektak 3 surface profiler machine
as described in Chapter 3. The film thickness measured in this work was between
521 Å to 3961 Å. The color of the film was transparent yellowish and the film’s
surface macroscopically appeared smooth. This is in agreement with the result
reported by Dutta et al. (1982).
From the results obtained, it was also observed that the surface morphology
of the deposited films as functions of the rf power and substrate temperature and can
be classified into three categories;
a) Mirror-like
b) Partially cloudy
c) Totally cloudy and soft powder-like
Here, the substrate dimensions are 76.2 × 25.4 mm and the surface morphology is
classified visually (as shown in Figure 4.1).
Mirror-like films evolved when the grain size of the film are considerably
small and the film formation on the substrate is homogeneous. This in turn would
62
make the surface become smooth and mirror-like. Partially cloudy on the other hand,
are formed when powder formation starts to develop on the gas phase of the plasma
but this formation is less than that of films with totally cloudy and soft powder-like
morphology. The films however, are still consist of relatively small-size grain or
clusters other than powder incorporation on the grown film (Nishikawa et al. 1985).
The cloudy morphology is probably produced by the powder formation in
the gas phase. When the powder is attached to the substrate, the growth of the films
becomes inhomogeneous and the surface becomes cloudy. When the powder
produced is strongly attached, the film becomes soft powder-like. Therefore, the
surface morphology reflects the amount of powder formation in the gas phase
(Nishikawa et al. 1985).
As the rf power increases, the gas-phase polymerization commences, leading
to the cloudy surface morphology. The reduction of temperature, correspond with an
enhancement of secondary reactions in the silane plasma which encourage the
formation of powders (Hadjadj et al.2001). This is the reason why the films are
observed to be cloudy and soft powder-like at low substrate temperature. As the
substrate temperature increases, the region where the cloudy morphology appears
shifts to the higher rf power region.
Substrate Temperature (˚C)
63
300
200
100
27
50
Mirror like
100
150
200
RF power (W)
250
Totally cloudy with soft powder like
Partially cloudy
Figure 4.1: Representation of surface morphology of a-Si:H films.
64
(a)
(b)
(c)
(d)
Figure 4.2: The 10 µm × 10 µm AFM images with surface height given by
the grey scales extending from 0 (black) to 400 nm (white): (a) sample A
deposited at 27°C; (b) sample B deposited at 100°C; (c) sample C deposited
at 200°C; (d) sample D deposited at 300°C.
65
Figure 4.2 presents the AFM images of the a-Si:H thin films, which shows a
root-mean-square (rms) surface roughness in the range of 1.2-4.9 nm. The surface
roughness was measured by atomic force microscopy (AFM) (Digital Instruments
Nanoscope II system) as mentioned in Chapter 3 in terms of the standard deviation of
the measured heights within a surface area of 10 µm × 10 µm. Tapping mode was
used in this AFM analysis and it was only made on samples prepared at different
substrate temperatures.
From the AFM images obtained, it is obviously seen that the shape of the
grains on the surfaces is spherical. These spherical grains are attributed to an
amorphous cluster or weakly crystalline phase, since XRD measurements did not
reveal any crystalline phase. Based on the surface morphology all samples can be
classified into three groups : (a) most smooth, sample A deposited at 27 °C (rms: 1.2
nm); (b) samples B and C deposited at 100 and 200 °C respectively are of
intermediate rms (in the range of 2.4-3.6 nm); (c) sample D has the highest
roughness (rms: 4.9nm) and on it exist some grains similar to those on sample A and
some similar to samples B and C (see figure 4,2(d)).
The transition to rougher films at higher substrate temperature has been
reported in the deposition of diamond-like carbon films (Lifshitz et al. 1994), and
this transition was attributed to a change in the deposition process, in which the
subplanted ions were able to migrate back to the surface. As the deposition
temperature (Ts) is increased, the mobility of some of the film-forming ions trapped
in subsurface positions of the evolving layer increases until the mobility is high
enough for the species to migrate back to the surface where they follow surface
growth processes.
Successive incorporation of film-forming atoms in subsurface positions of the
evolving film leads to high internal stress and to the formation of a dense phase in
films (Lifshitz et al. 1994). Since stresses associated with the incorporation of these
species in subsurface position no longer exist, rough films are formed. Again the
mobility of the ions increases with temperature leading to higher surface roughness.
Therefore, the much higher rms values for the samples deposited at 200°C and 300°C
than that of the sample deposited at 27°C becomes understandable.
66
At low substrate temperature, surface diffusion coefficient of the precursor
radical would increase. This would in turn reduce the mobility of growing radicals on
the surface. Thus making the film-forming radicals more probable to be adsorbed and
fill up the microvoids and columnar morphology available on the film surface. Films
would become more homogeneous and smoother films are formed (Knights et
al.1979).
The average deposition rate (dr) was obtains by dividing the thickness over
deposition time. It can simply be defined as;
dr =
Where;
d
t
(4.1)
d = The film thickness
t = Deposition time
The deposition rates for films prepared under different rf power are shown in
Figure 4.3. From the result shown, there is an increase of deposition rate for all
samples as rf power increases from 50 to 250 W. Goh (1992) also reported an
increase in the deposition rate with an increase of rf power.
Patel et al. (1986) observed a linear increase in the deposition rate from 0.18
nm/s to 0.5 nm/s as rf power increases from 5 to 50 W. Matsuda et al. (1982) stated
that in his report, as a primary process, the dissociation of SiH4 is enhanced by an
increase in rf power. At low power, less SiH4 molecules were dissociated, resulting
in lower deposition rate as seen in Figure 4.3.
67
Deposition Rate vs rf Power
1.2
Deposition Rate (Å/s)
Deposition Rate(A/s)
1
0.8
RT
300°C
100
200°C
˚C
100°C
200
˚C
RT˚C
300
0.6
0.4
0.2
0
50
100
150
200
250
rf Power (W)
Figure 4.3: Deposition rate of a-Si:H films deposited at different temperature as a
function of rf power.
68
The effect of substrate temperature was also investigated. Films were
deposited with the parameters as previously listed in Table 3.1. The substrate
temperature was varied from 27˚C to 300˚C. Figure 4.4 shows a general deposition
rate from 27˚C to 300˚C for sets of films prepared at different rf power.
Deposition Rate vs Substrate Temperature
1.2
Deposition Rate (A/s)
Deposition Rate (Å/s)
1
0.8
50 W
100 W
150 W
200 W
250 W
0.6
0.4
0.2
0
27
100
200
300
Substrate Temperature (C)
Figure 4.4: Deposition rate of a-Si:H films deposited at different rf power as a
function of substrate temperature.
69
H2
H H
H
H
Si
H
Si
Si
Si
Si
Si
H
H
H
Si
H
Si
Si
Si
Si
Si
Si
H
Si
Figure 4.5: Model of surface reaction where Si-Si is formed releasing H2.
It is obviously seen that the deposition rate decreases as substrate
temperatures increases for every rf power employed. Due to the increase in energy
(as substrate temperature increases), it is speculated that the growing radicals and
atoms generated in the plasma at the growing surface have high mobilities resulting
in a decrease in the deposition rate.
At high temperatures, Si-rich films may be deposited. This higher energy may
break the Si-H bonds, liberating the hydrogen and leaving Si-Si bonds behind (Goh,
1992). This can be illustrated in Figure 4.5. Matsuda et al. (1982) have reported that
binding energy (dissociation energy of a diatomic molecule) of Si-H bonds is lower
than Si-Si bonds where binding energy for Si-Si bond is 3.10 eV and Si-H bond is
3.06 eV.
70
4.2
X-Ray Diffraction (XRD) Analysis
The X-ray diffraction (XRD) analysis carried out on the a-Si:H thin films
confirmed that all samples are amorphous and no crystalline phase is observed due to
the presence of low-angle broad peak in the X-ray diffraction patterns. Figure 4.6 and
Intensity (a.u.)
4.7 depict typical XRD patterns for some representatives of a-Si:H thin film samples.
10
20
30
40
50
60
70
80
90
100
2θ (°)
Figure 4.6: X-ray diffraction pattern of a-Si:H thin film deposited at room
temperature and rf power of 100 W.
Intensity (a.u.)
71
10
20
30
40
50
60
70
80
90
100
2θ (°)
Figure 4.7: X-ray diffraction pattern of a-Si:H thin film deposited at 200˚C and
rf power of 50 W.
72
4.3
Infrared (IR) Transmission Spectrum
Before discussing the vibrational spectra, it is useful to establish the nature of
the vibrational modes expected at sites containing one, two or three H atoms: SiH,
SiH2, and SiH3 respectively. Figure 4.8 illustrates the atomic motions of these modes
(Brodsky et al. 1977, Lucovsky et al. 1979). There are basically two types: those
involving changes in the Si-H bond length (bond stretching) or H-Si-H bond angle
(bond bending), and those involving the rotation of these groups as a rigid unit
(rocking, wagging or twisting modes). H can be incorporated into a-Si in a number of
different ways as in this work, by the glow discharge decomposition of silane (SiH4).
Figure 4.8: Local Si-H vibrations for SiH, SiH2 and SiH3 groups.
73
The IR measurements in this work were made on a Perkin Elmer Spectrum
GX Spectrophotometer as mentioned in Section 3.2.3.3. The samples used in these
transmission measurements were deposited onto Corning glass substrates. The
measurements were made using a double-beam mode of operation in which a plain
Corning glass was used in the reference channel.
Summary of the IR peak assignment for the result obtained in this work and
its comparison with the results acquired by other researchers can be listed in Table
4.1.
Table 4.1: The observed mode frequency and their assignments.
Absorption Peak (cm-1)
This Work
2000
2090
630
636
890
845
850
880
592
Reference
Lucovsky et al. (1979)
2000
Assignment
SiH stretching
Shanks et al. (1980)
(SiH2)n and SiH2
2100
stretching
Lucovsky et al. (1979)
630
Brodsky et al. (1977)
635
Lucovsky et al. (1979)
890
Lucovsky et al. (1979)
845
Brodsky et al. (1977)
850
Lucovsky et al. (1979)
900
Brodsky et al. (1977)
590
SiH bending
SiH wagging
(SiH2)n bending
(SiH2)n wagging
SiH2 bending
SiH3 bending
SiH rocking
74
4.3.1
Dependence On RF Power
Figure 4.9, 4.10 and 4.11 illustrate the IR absorption spectra for films of a-Si
produced from the glow discharge decomposition of SiH4 and deposited onto
Corning glass substrates held at 200˚C with different rf power concentrating in the
2000 cm-1, 600 cm-1 and 800 cm-1 regime respectively.
75
2000
2090
50 W
100 W
T (arb. units)
150 W
200 W
250 W
2200
2000
cm-1
Figure 4.9: IR transmission of a-Si:H thin films deposited at different rf
power. Emphasize on the 2000 cm-1 regime.
76
630
50 W
100 W
T (arb. units)
150 W
200 W
250 W
800
800
600
-1
cm
Figure 4.10: IR transmission of a-Si:H thin films deposited at different rf
power. Emphasize on the 600 cm-1 regime.
77
845
50 W
890
T (arb. units)
100 W
150 W
200 W
250 W
1000
800
-1
cm
Figure 4.11: IR transmission of a-Si:H thin films deposited at different rf
power. Emphasize on the 800 cm-1 regime.
78
Note that at low rf power, the spectrum for all samples contain a strong
absorption bands at 2000 cm-1 which is associated with SiH stretching bond and also
at 630 cm-1 associated with SiH bending.
At high rf power, the spectrum is dominated by an additional absorption band
at around 2090 cm-1 which is corresponds to (SiH2)n stretching mode and SiH2
stretching mode. Scott et al. (1983) have reported that films prepared from SiH4 at
high rf glow discharge have higher hydrogen concentration and enhanced (SiH2)n
mode absorption compared with those prepared from SiH4 at low rf power.
Consequently, the concentration of SiH2 and (SiH2)n increases. The absorption band
at around 630 cm-1 becomes more pronounced as rf power increases. The same
results have been reported by Lucovsky et al. (1979) as rf power increases to 300 W
(see Figure 4.12).
630
10 Watt
100 Watt
200 Watt
T%
300 Watt
800
cm-1
600
Figure 4.12: IR transmission of a-Si:H thin film prepared under different rf
power by Lucovsky et al. (1979).
79
It is speculated by Tong Li et al. (2000) that the hydrogen content in the film
is directly related to the rf power. With the increase of the rf power, the densities of
the absorption peak located at 630 cm-1 which corresponds to hydride modes (SiH
bending (Lucovsky et al. 1979)) have increased.
Figure 4.11 exhibit the IR absorption of the samples prepared at different rf
power in the 800 cm-1 regime. It can be seen that the spectrum consist of doublet 890
cm-1 and 845 cm-1 which are assigned to (SiH2)n bending mode and wagging mode
respectively. As the rf power is increased, the absorption at 845 cm-1 increases
dramatically but there is a weak absorption evident at approximately 890 cm-1. This
is in accordance with the result obtained by Lucovsky et al. (1979). Since the (SiH2)n
groups (absorption at 845 cm-1) is enhanced as rf power increases, the films contain
an excess of (SiH2)n chains. These (SiH2)n will produce a polymer chain in the
structure of the film. These results show that the structure of the films becomes
polymeric-chain-like (with (SiH2)n chains) with increasing of rf power.
4.3.2
Dependence On Substrate Temperature
It is interesting to note that samples produced by plasma decomposition of
SiH4 onto high temperature substrates (Ts > 200˚C) have local environments
containing SiH groups, SiH2 groups and (SiH2)n groups. Samples produced at room
temperature display features that can be attributed to SiH3 groups as well as (SiH2)n
(Lucovsky et al. 1979).
This can be seen from Figure 4.13; from samples prepared at room
temperature, spectrum shows peak at 2090 cm-1 which is associated with (SiH2)n
stretching and SiH3 stretching mode (Brodsky et al. 1977) while samples prepared at
higher temperature shows a strong absorption bands peaked near 2000 cm-1 attribute
to SiH stretching mode.
There are also strong absorption bands peaked around 850 cm-1 and 880 cm-1
for samples prepared at room temperature, 27°C (see Figure 4.14) which is
80
associated with SiH2 and SiH3 bending modes respectively (Brodsky et al. 1977). At
higher substrates temperature the spectrum shows weak absorption peaks.
Thompson et al. (1972) stated in his report that films produced at substrate
temperatures exceed 200˚C do not exhibit SiH3 vibrations; however SiH3 modes are
clearly evident at samples produced on room-temperature substrates as what is
observed in this work.
For spectrum near 600 cm-1 regime as what is depicted in Figure 4.15, sample
prepared at room temperature shows peak around 636 cm-1 and 592 cm-1 which is
associated with Si-H wagging mode and rocking mode absorption band respectively.
As temperature increases, it shows slight shifts in the frequencies to higher values
peaked at around 650 cm-1 and 600 cm-1 for samples at 300˚C. The same increment
has also been observed by Brodsky et al. (1977) where in all their samples, wagging
mode are seen with slight shifts in the frequencies to higher values as the substrate
temperature is increased from room temperature to 250˚C.
Further note that the Si-H bond-stretching frequencies in a-Si:H alloys are
smaller than those found in silane molecules which is around 2100 – 2360 cm-1
(Bellamy, 1975) but fall within the range of frequencies observed for H vibrations in
crystalline silicon which is in the range of 2000 – 2175 cm-1 (Stein, 1975).
81
2000
2090
27˚C
T (arb. units)
100˚C
200˚C
300˚C
2200
2000
cm-1
Figure 4.13: IR transmission of a-Si:H thin films deposited at different
substrate temperatures. Emphasize on the 2000 cm-1 regime. RF
power is at 100 W.
82
850
27˚C
880
T (arb. units)
100˚C
200˚C
300˚C
1200
1000
800
cm-1
Figure 4.14: IR transmission of a-Si:H thin films deposited at different
substrate temperatures. Emphasize on the 800 cm-1 regime. RF
power is at 100 W.
83
590
635
27˚C
T (arb. units)
100˚C
200˚C
300˚C
800
800
600
-1
cm
Figure 4.15: IR transmission of a-Si:H thin films deposited at different
substrate temperatures. Emphasize on the 600 cm-1 regime. RF
power is at 100 W.
84
4.4
Optical Energy Gap
The optical energy gap obtained from the intercept in the (αhν)1/2 vs hν plot
ruled out from the Tauc’s expression discussed in the Section 3.2.2.1. The graph of
(αhν)1/2 vs hν are plotted for both samples prepared under different rf powers and
substrate temperatures as shown in Figure 4.16 and 4.17 respectively. The optical
energy gap measured from UV-3101-PC Spectrometer is plotted as a function of the
rf power in Figure 4.18. From the result obtained the optical energy gaps were found
to vary between 1.59-1.97 eV.
(αhν)1/2 vs hν (RF Power Dependency)
14000
50 W
(αhν)1/2 (x 109 cm -1/2 eV 1/2)
12000
100 W
10000
150 W
200 W
8000
250 W
6000
4000
2000
0
1
1.5
2
2.5
3
3.5
hν (eV)
Figure 4.16: The (αhν)1/2 vs hν graphs for a-Si:H deposited at different rf power.
50 W
100 W
150 W
200 W
250 W
85
(αhν)1/2 vs hν (Substrate Temperature Dependency)
(αhν)1/2 (x 109 cm -1/2 eV 1/2)
12000
RT
10000
100°C
8000
200°C
RT
100˚C
200˚C
300˚C
6000
300°C
4000
2000
0
1
1.5
2
2.5
3
3.5
4
4.5
hν (eV)
Figure 4.17: The (αhν)1/2 vs hν graphs for a-Si:H deposited at different substrate
temperature.
86
Energy Gap (eV) vs RF Power (W)
2.2
Energy Gap (eV)
2
1.8
RT
100˚C
200˚C
300˚C
1.6
1.4
1.2
1
50
100
150
200
250
RF Power (W)
Figure 4.18: Energy gap of a-Si:H films prepared under different substrates
temperatures as a function of rf power.
87
The effect of rf power on the optical band gap of the films deposited from
silane glow discharge is a decrease in the band gap as the power level increases. This
is well in agreement with the result obtained by Sichanugrist et al. (1986). According
to Tsuo et al. (1987), the decreases in the optical band gap with rf power is due to the
increase in the hydrogen content, since the band gap increases linearly with
increasing hydrogen content.
Values of the optical band gap of a-Si:H depend on the hydrogen content,
following the empirical relation Eg = 1.56 + 1.27 CH, where CH is the atomic fraction
of the bonded hydrogen (Ross and Jaklik et al. 1984, Tsuo et al. 1987). Tsuo et
al.(1987) reported that the optical band gap of a-Si:H thin film deposited decreases
from 1.73 eV to 1.59 eV as the total H content decreases from 13 at. % to 2 at. %.
This equation is for glow discharge-deposited a-Si:H deposited at high temperature.
At a very high power levels, the formation of microcrystalline silicon causes
a sharp decreases in both the hydrogen content and the optical band gap (Hata et al.
1981). Hata et. al. (1981) suggest that it is reasonable to consider that bonded
hydrogens are distributed mainly in the amorphous phase at a constant concentration,
and are present in much lower concentration in the microcrystalline phase. This can
be express in an empirical relation;
CH
= const
(1 − X C )
where;
(4.2)
CH = bonded H content
XC = volume fraction of microcrystalline
Therefore as the film appears to be having some microcrystalline phase, there
will be a decrease to hydrogen content. This would thus leads to a decrease in optical
band gap.
88
From Figure 4.17 it shows that the optical band gap of the samples deposited
decreases as the substrate temperature increases. This is primarily due to the
decreased in hydrogen content. Hata et al. (1981) in his work reported that the
optical gap for a-Si:H thin films deposited via rf glow discharge decreases from 1.72
eV to 1.65 eV as hydrogen content decreases from 12 at. % to 7 at. %. The total
hydrogen content decreases with increasing substrate temperature in films both from
monosilane (Knights et al. 1979) or disilane (Ross and Jaklik et al. 1984). Ross and
Jaklik et al. (1984) in their study reported that as substrate temperature increases
from 100°C to 300°C, the total H content decreases from 23 at. % to 8 at. % for film
deposited in SiH4 plasma. Since optical gap is predominantly related to H content,
consequently the optical gap would also decreases from 1.85 eV to 1.66 eV.
89
4.5
Photoluminescence
Photoluminescence measurement gives a contactless, nondestructive method of
probing the electronic structure of the amorphous silicon thin films deposited.
Besides UV spectrometer, band gap determination could also be done using
Photoluminescence results. The most common radiative transition in semiconductor
is between states in the conduction and valence bands, with the energy difference
being known as band gap.
The return of photo-excited electrons within a material to equilibrium state,
also known as “recombination” can involve both radiative and nonradiative
processes. The amount of photoluminescence and its dependence on the level of
photo-excitation and temperature are directly related to the dominant recombination
process. Analysis of photoluminescence helps to understand the underlying physics
of the recombination mechanism.
In this study, all photoluminescence characterization was done using
photoluminescence spectrometer LS 55 as described previously in detail in Chapter
3.
4.5.1
Dependence On Substrate Temperature
Figure 4.19 to 4.23 show the luminescence spectrum of amorphous silicon
prepared at different deposition temperatures varied from 27˚C to 300˚C with rf
power fixed at 50, 100, 150, 200 and 250W. The excitation energy for all the samples
in this measurement was around 1.95-2.0 eV which is much higher than the bandgap
value obtained from Section 4.4.
90
0.9
0.8
0.7
Intensity (arb. units)
RT
0.6
100˚C
RT
200˚C
300˚C
0.5
0.4
100°C
0.3
0.2
200°C
0.1
300°C
0
1.25
1.3
1.35
1.4
1.45
1.5
1.55
1.6
Energy (eV)
Figure 4.19: Luminescence spectrum of amorphous silicon at various substrate
temperatures with rf power kept constant at 50W
91
0.8
0.7
RT
Intensity (arb. units)
0.6
0.5
RT
100˚C
200˚C
300˚C
0.4
100°C
0.3
0.2
200°C
0.1
300°C
0
1.25
1.3
1.35
1.4
1.45
1.5
1.55
1.6
Energy (eV)
Figure 4.20: Luminescence spectrum of amorphous silicon at various substrate
temperatures with rf power kept constant at 100W
0.8
0.7
RT
Intensity (arb. units)
0.6
0.5
RT
100˚C
200˚C
300˚C
0.4
100°C
0.3
0.2
200°C
0.1
300°C
0
1.25
1.3
1.35
1.4
1.45
1.5
1.55
1.6
Energy (eV)
Figure 4.21: Luminescence spectrum of amorphous silicon at various substrate
temperatures with rf power kept constant at 150W
92
0.8
0.7
RT
Intensity (arb. units)
0.6
0.5
RT
100˚C
200˚C
300˚C
0.4
100°C
0.3
0.2
200°C
0.1
0
1.25
300°C
1.3
1.35
1.4
1.45
1.5
1.55
1.6
Energy (eV)
Figure 4.22: Luminescence spectrum of amorphous silicon at various substrate
temperatures with rf power kept constant at 200W
0.9
0.8
RT
Intensity (arb. units)
0.7
0.6
100°C
0.5
RT
100˚C
200˚C
300˚C
0.4
200°C
0.3
0.2
300°C
0.1
0
1.25
1.3
1.35
1.4
1.45
1.5
1.55
1.6
Energy (eV)
Figure 4.23: Luminescence spectrum of amorphous silicon at various substrate
temperatures with rf power kept constant at 250W
93
From all figures (Figure 4.19 – 4.23) it can be seen that at lower substrate
temperature (Ts = 27°C and 100°C), luminescence spectrum of amorphous silicon
consists of a relatively broad band with its main peak around 1.4 eV. This result is in
well accordance with Pankove and Carlson et al. (1976) and Pincik and Kobayashi et
al. (2003) which Pankove and Carlson describe their spectrum as peaking at 1.44 eV.
Note that the energy where the luminescence spectrum peak is slightly lower than the
energy band gap obtained in Section 4.4. Nevertheless it is still within the energy
range. It is believed that most of the radiative transitions are not entirely between
states in the conduction and valence bands.
As temperature increases (Ts = 200°C and 300°C), the centre of mass of the
luminescence band moves to lower energy around 1.37 eV. While the intensity of the
luminescence band decreases as substrate temperature (Ts) increases. The same result
has been reported by Engemann and Fischer et al. (1977) where they found the
photoluminescence intensity decreases as temperatures increases above 80 K. It is
suggested that this is due to an activated non-radiative recombination process.
When the temperature increases, the probability for a bound electron-hole
pair to be separated become larger. These carriers can recombine in either of two
ways. They can re-form bound pairs and recombine radiatively, and they can
recombine non-radiatively at a centre. It is assumed that recombination at a centre is
more likely than re-forming of pairs. This is the case if, for example, after
dissociation the hole is quickly captured by a deep trap and thereby generates an
effective recombination centre. These excitation and recombination mechanism can
be illustrated in a schematic diagram as shown in Figure 4.24.
At high enough temperature, the higher mobility of the exciton makes its
capture by deep traps more probable. This mechanism is clearly responsible for the
temperature quenching of luminescence which indicates that above 200˚C the
radiative channel rapidly becomes insignificant. In other word, this would explain
why the luminescence intensity of the samples observed decreases as substrate
temperature increases from 100 to 300˚C.
94
PHOTOEXCITATION
EXCITON
FORMATION
CAPTURE BY SHALLOW
TRAPS OR BAND TAILS
ELECTRONS AND
HOLES CAPTURED
BY DEEP TRAPS
RADIATIVE
RECOMBINATION
NONRADIATIVE
RECOMBINATION
GROUND STATE
Figure 4.24: Flow diagram representing the interrelation of the excitation and
recombination processes appropriate to amorphous silicon (Street, 1978).
95
The shift of the centre mass of the luminescence band can also be explained
by the recombination edge model as described by Fischer et al. (1971). In this model,
the electrons excited into states well above the mobility edge (conduction band),
relax to lower states by emission of phonons until they reach an energy where
recombination is more probable than further relaxation.
Probability
Relaxation
Recombination
T1, T2
T2
T1
T2 > T1
Energy (eV)
Figure 4.25: Schematic plot of the relaxation and recombination
probabilities versus energy from midgap.
In Figure 4.25 the probabilities for thermal relaxation and radiative
recombination are plotted schematically versus energy from midgap. It is assumed
that relaxation at any particular energy becomes more probable when the temperature
increases. This assumption is reasonable since relaxation needs the contribution of
phonons. So as temperature increases, relaxation (non-radiative recombination) is
more probable than radiative recombination, giving the centre mass of luminescence
spectrum move to lower energy.
CHAPTER 5
CONCLUSION
5.1
Summary
In this work, samples of hydrogenated amorphous silicon (a-Si:H) thin films
have been deposited using silane (SiH4) as film precursor via plasma enhanced
chemical vapor deposition (PECVD) technique. Deposition was done under different
substrate temperatures and rf powers while other parameters kept constant.
Characterization results obtained reveal that the surface morphology of the
deposited films as a function of the rf power and substrate temperature can be
visually classified into three categories which are; mirror like, partially cloudy and
totally cloudy/soft powder-like. These morphology transitions is subjected to gasphase polymerization as deposition parameters change which lead to powder
formation that would alter film’s surface structure.
From the AFM images obtained, the surface morphology of all samples can be
classified into three groups; most smooth (rms 1.2 nm), intermediate (rms 2.4-3.6
nm), highest roughness (rms 4.9 nm). The transition to rougher films at higher
substrate temperature is attributed to the mobility of the subplanted ions in growing
film surface.
97
The deposition rate of the films increases with increasing rf power, which is
associated with SiH4 dissociation process. It is also speculated that the growing
radicals and atoms generated in the plasma at the growing surface as substrate
temperature arises, have high mobilities resulting in a decrease in the deposition rate.
From the IR measurement in this work, it is shown that at low rf power, the IR
spectrum for all samples contain a strong absorption bands at 2000 cm-1 which is
associated with SiH stretching mode and also 630 cm-1 associated with SiH bending.
At high rf power, the spectrum is dominated by an additional absorption band at
around 2090 cm-1 which is corresponds to (SiH2)n and SiH2 stretching mode. It is
speculated that the hydrogen content in the film is directly related to the rf power
which explains the increased of absorption peak located at 630 cm-1 which
corresponds to hydride modes as rf power increased.
It is interesting to note that samples produced by plasma decomposition of
SiH4 onto high temperature substrates (Ts > 200˚C) have local environments
containing SiH groups, SiH2 groups and (SiH2)n groups. Samples produced at room
temperature on the other hand displayed features that can be attributed to SiH3
groups as well as (SiH2)n.
The effect of rf power on the optical band gap of films deposited from silane
glow discharge is a decrease in the band gap as the rf power level increases. At a
very high power levels, the formation of microcrystalline phase causes a sharp
decrease in hydrogen content, thus leading to a decrease in optical band gap. As
substrate temperature increases, a decrease of the band gap was being observed. This
is also due to drop of hydrogen content.
At lower substrate temperature, luminescence spectrum of amorphous silicon
consists of a relatively broad band with its main peak around 1.40 eV. As
temperature increases, the centre of mass of the luminescence band moves to lower
energy around 1.37 eV. While the intensity of the luminescence band decreases as
substrate temperature (Ts) increases. It is suggested that this is due to an activated
non-radiative recombination process. At high enough temperature, the mobility of
98
the exciton makes its capture by deep traps more probable; hence, radiative channel
becomes insignificant. In addition, as temperature increases, relaxation (nonradiative recombination) is more probable than radiative recombination, giving the
centre mass of luminescence spectrum moved to lower energy.
5.2
Recommendation
In this study, the hydrogenated amorphous silicon thin films were prepared
under different rf powers and substrate temperatures. For future works, it is
suggested that imposing other parameters such as chamber pressure and precursor
gas flow rate would also be considered as these parameters are also believed to give a
significant change in the structure of the amorphous film grown.
In addition, instead of only focusing on the intrinsic amorphous silicon thin
film, study can also be done on doped amorphous silicon thin film since doped
materials are pretty much important mainly in developing thin film devices such as
solar cells. This can be made possible by mixing doping gas precursor namely
diborane (B2H6) and phosphane (P2H4) with silane (SiH4) during deposition process.
There are many ways to study the structural characteristic of amorphous thin
film. Other than the techniques mentioned in this thesis, the amorphous silicon thin
films can also be characterised by using Scanning Electron Microscopy (SEM) or
Tunneling Electron Microscopy (TEM) since these characterisation equipments are
able to give detail on the surface morphology of the films. This would in turn
compliment the AFM results that have been obtained.
99
REFERENCES
Adams, A. C. (1983). Plasma Deposited Hydrogenated Amorphous Silicon Nitride
Thin Films. Solid State Technology. 26(4): 135-140.
Bellamy, L. J. (1975). The Infra-Red Spectra of Complex Molecules. London:
Chapman and Hall.
Brida, D., Fortunato, E., Aguas, H., Silva, V., Marques, A., Pereira, L., Ferreira, I.
and Martins, R. (2002). New Insights on Large Area Flexible Position Sensitive
Detectors. J. Non-Cryst. Solids, 299: 1272-1276.
Brodsky, M. H., Cordona, M., Cuomo, J. J. (1977). Infrared and Raman Spectra of
The Silicon-hydrogen bonds in a-Si Prepared by Glow Discharge and Sputtering.
Phys. Rev. B. 16(8): 3556-3571.
Brodsky, M. H. and Title, R. S. (1969). On The Deposition of Amorphous Silicon
Films from Glow Discharge Plasmas of Silane. Phys. Rev. Lett. 23: 581-585.
Bunshah, R. F. (1993). Handbook of Deposition Technologies for Films and
Coatings. Park Bridge, N. J.: Noyes Publications.
Chapman, B. (1980). Glow Discharge Processes. New York: John Wiley.
Chittig, R. C., Alexander, J. H., Sterling, H. F. (1969). The Preparation and
Properties of Amorphous Silicon. J. of Electrochem. Soc. 116: 77-81.
100
Drevillon, G., Huc, J., Lort, A., Perrin, J., de Rosny, G., and Schimtt, J. P. M. (1980).
Silane Dissociation Mechanisms and Thin Film Formation in a Low Pressure
Multipole DC Discharge. Appl. Phys. Lett. 37:646-648.
Dutta, R., Banarjee, P. K. and Mitra, S. S. (1982). Optical and Electrical Properties
of Hydrogenated Amorphous Silicon Carbide. Phys. Stat. Sol. B. 113(1): 277-284.
Engemann, D. and Fischer, R. (1977). Photoluminescence in Amorphous Silicon.
Phy. Stat. Sol. (B), 79: 195-202.
Fischer, R., Heim, U., Stern, F., and Weiser, K. (1971). Photoluminescence of
Amorphous 2As2Te3 · As2Se3 Films. Phys. Rev. Lett., 26(19): 1182-1185.
Goh, F. H. C. (1992). Wide Bandgap Amorphous Silicon-based Alloys Using
PECVD. University of Arkansas: Ph.D. Thesis.
Gorowitz, B., Gorczyca, T. B. and Saia, R. J. (1985). Amorphous Silicon Nitride
Deposition Rates in RF Glow Discharges. Solid State Technology. 28(6): 197-200.
Hadjadj, A., Beorchia, A.,Rocca I Cabarrocas, P., Boufendi, L., Heut, S. and
Budendorff, J. L. (2001). Effects of the Substrate Temperature on the Growth and
Properties of hydrogenated nanostructured silicon thin films. J. Phys. D: Appl. Phys.,
34: 690-699.
Hata, N., Yamasaka, S., Oheda, H., Matsuda, A., Okushi, H. and Tanaka, K. (1981).
A Photoluminescence Study of Amorphous-Microcrystalline Mixed-Phase Si:H
Films. Jpn. J. Appl. Phys., 20(11): L793-796.
Hollahan, J. R. and Rosler, R. S. (1978). In: Vossen, J. L. and Kern, W. ed. Thin
Film Processes. New York: Academic Press. Ch. IV-1.
Howatson, A. M. (1976). An Introduction to Gas Discharges. Oxford: Pergamon
Press.
101
Inoue, G. and Suzuki, M. (1985). Reactions of SiH2 (X’A1) with H2, CH4, C2H4, SiH4
and Si2H6 at 298 K. Chemical Phys. Lett. 122: 361-364.
Janai, M., Weil, R. and Levin, H. K. (1981). Optical Properties of Fluorinated GlowDischarge Amorphous Silicon. Appl. Phys. 52: 3622-3625.
Jasinski, J. M. and Chu, J. O. (1988). Absolute Rate Constants for the Reaction of
Silylene with Hydrogen, Silane and Disilane. J. Chem. Phys. 88: 1678-1687.
Knights, J. C., Lucovsky, G. and Nemanich, R. J. (1979). Defects in PlasmaDeposited a-Si:H. J. Non-Crystal. Solids, 32: 393-403.
Knights, J. C. (1988). Plasma Deposition of a-Si:H. Mat. Res. Soc. Symp. Proc. 38:
371-381.
Krause, M., Topic, M., Stiebig, H. and Wagner, H. (2001). Thin Film UV Detectors
Based on Hydrogenated Amorphous Silicon and Its Alloys. Phys. Stat. Sol. A, 185:
121-127.
Lifshitz, Y., Lempart, G. D., Grossman, E. (1994). Substation of Subplantation
Model for Diamond-like Film Growth by Atomic Force Microscopy. Phys. Rev. Lett.
72: 2753-2756.
Longeway, P. A. (1984). Plasma Kinetics. Semiconductors and Semimetals. 21: 179193.
Lucovsky, G., Nemarich, R. J. and Knights, J. C. (1979). Structural Interpretation of
The Vibrational Spectra of a-Si:H Alloys. Phys. Rev. B. 19(4): 2064-2073.
Luft, W. and Tsuo, S. (1988). Plasma Deposition of Hydrogenated Amorphous
Silicon Films. Appl. Phys. Comm. 8(1): 1-74.
Matsuda, A. and Tanaka, K. (1982). Plasma Spectroscopy-Glow Discharge
Deposition of Hydrogenated Amorphous Silicon. Thin Solid Films. 92(1): 171-187.
102
Nishikawa, S., Kakinuma, H., Watanabe, T. and Nihei, K. (1985). Influence of
Deposition Conditions on Properties of Hydrogenated Amorphous Silicon Prepared
by RF Glow Discharge. Jpn. J. Appl. Phys. 24(6): 639-645.
Pankove, J. I. and Carlson, D. E. (1976). Electroluminescence in Amorphous Silicon.
Appl. Phys. Lett., 29(9): 620-628.
Patel, R. I., Olsen, D. J., Shrick, J. R. and Tran, N. T. (1986). Comparison Studies of
Hydrogenated Amorphous Silicon Films Prepared from Silane-Hydrogen and SilaneHelium Mixtures. Mat. Res. Soc. Symp. Proc., 70: 55-58.
Pincik, E., Kobayashi, H., Gleskova, H., Kucera, M., Ortega, L., Jergel, M., Falcony,
C., Brunner, R., Shimizu, T., Nadazdy, V., Zeman, M., Mikula, M., Kumeda, M. and
R.A.C.M.M. van Swaaij (2003). Photoluminescence Properties of a-Si:H Based Thin
Films and Corresponding Solar Cells. Thin Solid Films, 433: 344-351.
Raizer, Y. R. (1991). Gas Discharge Physics. Berlin: Springer-Verlag.
Rand, M. (1979). Plasma-promoted Deposition of Thin Organic Films. J. Vac. Sci.
Tech.. 16(2): 420-427.
Reif, R., Kamins, T. I. and Saraswat, K. C. (1979). Surface Reactions in Discharge
and CVD Deposition of Silane. J. Electrochem. Soc. 126: 644-652.
Reif, R. (1990). In: Rossnagel, S., Cuomo, J. and Westwood, W. ed. Handbook of
Plasma Processing Technology. Park Ridge, N. J.: Noyes Publications. Ch. 10.
Robertson, R. and Gallagher, A. (1986). Reaction Mechanism and Kinetics of Silane
Pyrolysis on a Hydrogenated Amorphous Silicon Surface. J. Chem. Phys. 85: 3623.
Ross, R. C., Jaklik, J. (1984). Plasma Polymerization and Deposition of Amorphous
Hydrogenated Silicon from rf and dc Silane Plasmas. J. Appl. Phys., 55(10): 37853794.
103
Sah, W. J., Lee, S. C., Tsai, H. K. and Chen J. H. (1990). Amorphous Silicon Edge
Detectors for Application to Neural Network Image Sensors. Appl. Phys. Lett., 56:
2539-2541.
Scott, B. A., Reimer, J. A. and Longeway, P. A. (1983). Growth and Defect
Chemistry of Amorphous Hydrogenated Silicon. J. Appl. Phys. 54: 6853-6863.
Shen, D. S., Chatham, H. and Bhat, P. K. (1991). High-deposition Rate Amorphous
Silicon Solar Cells: Silane or Disilane?. Solar Cells, 30: 271-275.
Sherman, A. (1984). Plasma-assisted Chemical Vapor Deposition Processes and
Their Semiconductor Applications. Thin Solid Films. 113(2): 135-149.
Shimizu, T., Nakazawa, K., Kumeda, M., Ueda, S. (1983). Physica (Utrecht) 117B118B: 926.
Sichanugrist, P., Suzuki, H., Konagai, M. and Takahashi, K. (1986). High-Rate
Preparation of Amorphous-Silicon Solar Cells with Monosilane. Jap. J. Appl. Phys.,
25(3): 440-443.
Smith, D. L. (1994). Thin Film Deposition. New York: McGraw-Hill Inc..
Spear, W. E. and LeComber, P. G. (1975). Substitutional Doping of Amorphous
Silicon. Solid State Comm. 17(9): 1193-1196.
Stein, H. J. (1975). Bonding and Thermal Stability of Implanted Hydrogen in Silicon.
J. Elec. Mater. 4(1): 159-174.
Street, R. A. (1978). Recombination in Amorphous Semiconductors. Phys. Rev. B,
17: 3984-3995.
Street, R. A. (1991). Hydrogenated Amorphous Silicon. Cambridge: Cambridge
University Press.
104
Tauc, J., Grigorovici, R. and Vancu, A. (1966). Optical Properties of Solids. Phys.
Stat. Sol. 15: 627-638.
Thompson, L. F., Mayhan, K. G. (1972). The Plasma Polymerization of Vinyl
Monomers. I. The Design, Construction and Operation of an Inductively Coupled
Plasma Generator and Preliminary Studies With Nine Monomers. J. Appl. Polym.
Sci. 16(9): 2291-2315.
Tong Li (2000). Advanced Processes and Characterization of Amorphous Silicon
Thin Film Transistors. University of Michigan: Ph.D. Thesis.
Tsuo, T. S., Smith, E. B., Deb, S. K. (1987). Ion Beam Hydrogenation of Amorphous
Silicon. Appl. Phys. Lett., 51(18): 1436-1438.
Urbach, F. (1958). The Long-wavelength Edge of Photographic Sensitivity and the
Electronic Absorption of Solids. Phys. Rev. 92(5): 1324.
Xi, J., Hollingsworth, R. E., Buitrago, R., Oakley, D., Cumalot, J. P., Naunberg, U.,
McNeil, J., Anderson, D. F. and Perez-Mendez, V. (1991). Minimum Ionizing
Particle Detection Using Amorphous Silicon Diodes. Nuclear Instruments &
Methods in Physics Research, A301: 219-222.
Yang, J., Banerjee, A. and Guha, S. (1997). Triple-junction Amorphous Silicon
Alloy Solar Cell With 14.6% Initial and 13.0% Stable Conversion Efficiencies. Appl.
Phys. Lett., 70: 2975-2977.
Zhao, W. and Rowlands, J. A. (1995). X-ray Imaging Using Amorphous Selenium:
Feasibility of a Flat Panel Self-scanned Detector for Digital Radiology. Med. Phys.,
22(10): 1595.
105
PRESENTATIONS
1. Conference On Advanced Material, Advanced Technology Congress
2003 (ATC2003), 20-21 May 2003, Putrajaya Marriot Hotel, IOI
Resort, Putrajaya.
2. Conference on Public Institutions Of Higher Learning (IPTA), UPM, 910 October 2003, PWTC, Kuala Lumpur.
3. XX Regional Conference On Solid State & Technology (MASS), 12-14
December 2003, Lumut, Perak.
4. Annual Fundamental Science Seminar 2004 (AFSS 2004), 14-15 June
2004, Ibnu Sina Institute For Fundamental Science Studies, UTM
Skudai, Johore.