(CHARACTERIZATION OF CuxOy- ZnO:AI P-N JUNCTION FOR SOLAR
CELL APPLICATIONS
FANUEL MUGWANG'A~ZE (B.Ed. Sc)
I56/CE/15646/2005
A thesis submitted in partial fulfillment of the requirements for the award of the degree of
Master of Science (Electronics and Instrumentation) in the School of Pure and Applied
Sciences of Kenyatta University
Keheze, Fanuel
Characterrization of
CuxOy-ZnO:Al P-N
1~1!~IIIIJ~IIII"IIII"III~III~illllJ2012/383682
OCTOBER 2011
11
DECLARATION
This thesis is my original work and has not been presented for the award of a degree or
any other award in any other University
FANUEL MUGWANG'A KEHEZE
Department of Physics~::~t:~.~ DateO:r.!rO!L!"I .
We confirm that the candidate carried out the work reported in this thesis under our
supervision,
Dr. Patrick M. Karimi
Department of Physics (f\
Kenyatta University ~
Signature . Date ~\~.\.~.U .
Dr. Walter K. Njoroge
Department of Physics
Kenyatta University
signature ~ Date !?:.r.(~lf/. .
Dr. Sebastian M. Waita
Department of physics~i::;::"~~l~O~l fIff Date . .TkO f.~/t. .
111
DEDICATION
This thesis is dedicated to my wife Susan, daughters and my late grandmother Imungu.
IV
ACKNOWLEDGEMENTS
My special thanks go to my supervisors Dr. P. M. Karimi, Dr. W. K. Njoroge and Dr.
S.M. Waita for their tireless scholarly guidance and encouragement during the research of
my work. Their vast knowledge in solar cell materials was a great asset to this work. I am
particularly grateful to Dr. P. M. Karimi who introduced me to the world of
optoelectronics cells and optoelectronics. Thanks to Dr. W. K. Njoroge and Dr. S.M.
Waita for their constant guidance through supervision of this work.
I acknowledged the Vice Chancellor of Kenyatta University, Prof Olive Mugenda for
providing a conducive learning environment beneficial to this work. I am grateful to the
members of staff, Physics Department, led by Dr. C. Migwi for being supportive
throughout the research period. This project involved a lot of laboratory work, and I am
very grateful to all the Physics department technical staff led by Mr. Simon Njuguna who
was always there for me, gave me the required apparatus, and generally did things that
were beyond the call of duty. Not to forget is Mr.Boniface Muthoka, senior technician,
solid state physics laboratory, University of Nairobi, who assisted and trained me on how
to make optical and solar cell measurements and made the equipments available to me.
I will also not forget my staff mates and students in Kisangula secondary school led by
the Principal, Mr Moga for his encouragement to and support during my studies.
vTo my family, thanks for your moral support and love throughout my studies. My due
regards to my colleagues Agumba, Omayio, Muga, Kirwa, Ogaro, Munguti, Kitur,
Masinde, Tuwei and Ketui just to mention a few. Thanks to my friends Dingili and
Igunza who gave me moral support and my late grandmother Selina lmungu, who
remained the source of my inspiration. Above all, thanks to God without whom this
work would not have been successful.
Vi
TABLE OF CONTENTS
TITLE .i
DECLARATION .ii
DEDICATION .iii
ACKNOWLEDGEMENTS .iv
TABLE OF CONTENTS v
LIST OF TABLES xi
LIST OF FIGURES xii
ABBREVIATION AND ACRONyMS xv
SyMBOLS xvii
ABSTRACT xviii
CHAPTER ONE
INTRODUCTION
1.1 Background to the stud 1
1.2 Types of solar cells 2
1.3 Statement of the research problem and justification 6
1.4 Objectives 7
1.4.1 Main objective 7
Vll
1.4.2 Specific objectives 7
1.5 Rationale 8
CHAPTER TWO
LITERATURE REVIEW
2.1 Introduction 9
2.2 Background to solar cell technology 9
2.3 Copper Oxide : 13
2.4 Aluminum doped ZnO .16
CHAPTER THREE
THEORETICAL BACKGROUND
3.1 Introduction 20
3.2 Solar cells 20
3.3 Theory of I-V Characterization 23
3.3.1 Short Circuit Current (lsc) 25
3.3.2 Open Circuit Voltage (Voe) 25
3.3.3 Fill Factor (FF) 26
3.3.4 Efficiency (11) 26
3.3.5 Effect of physical size (area of cell) 27
3.4 Solar radiation 29
3.4.1 Absorption of light by a p-n junction 30
Vlll
3.5 High-efficiency solar cells by material integration 32
3.6 Deposition Techniques 33
3.6.1 Sputtering Technique 34
3.6.2 Vacuum Evaporation Technique 35
3.6.3 Other Deposition Techniques 36
3.6.3.1 Arc Vapour Deposition Technique 36
3.6.3.2 Ion Plating Technique 37
3.6.3.3 Chemical Vapour Deposition Technique 37
3.7 Optical and structural characterization of thin films 38
3.7.1 Optical characterization of thin films 38
3.7.1.1 Optical properties of amorphous and crystalline materials 38
3.7.1.2 Direct and indirect optical transitions .40
3.7.1.3 Absorption edge and Urbach energy .42
3.7.1.4 Optical band gap and physical models .43
3.7.1.5 Structural properties ofthin films .46
CHAPTER FOUR
EXPERIMENTAL PROCEDURE
4.1 Introduction 48
4.2 Thin film preparation .48
4.2.1 Deposition of Copper Oxide (CuxOy) thin films .48
4.2.2 Deposition of Aluminum doped zinc oxide (ZnO:Al) 50
IX
4.3 Copper Oxide and Aluminum doped ZnO thin films characterization 50
4.3.1 Thin film thickness measurements 50
4.3.2 Optical measurements 51
4.4.3 Sheet resistivity measurements 52
4.4 Fabrication ofCuxOyl ZnO:Al heterojunction 52
4.5 Solar cell characterization 53
CHAPTER FIVE
RESULTS AND DISCUSSION
5.1 Introduction 55
5.2 Characterization of ZnO:AI thin film 55
5.2.1 Optical characterization ofZnO:AI thin films 55
5.2.1.1 Simulated and Experimental graphs for AZO transmittance data 57
5.2.1.2 Optical band gap and urbach energy for AZO thin films 60
5.2.2 Electrical characterization ofZnO:AI thin films 66
5.2.3 Composition and structural characterization ofZnO:AI thin films 68
5.2.3.1 X-ray diffraction ofZnO:Al thin film (as-deposited) 68
5.2.3.2 Elemental composition of ZnO:Al thin film (as-deposited) 69
5.3 Characterization of Cu.O, thin films 71
5.3.1 Optical characterization of Cu.O, thiri films 71
5.3.1.1 Simulated and experimental graphs for Copper Oxide transmittance data. 72
x5.3.1.1 Simulated and experimental graphs for Copper Oxide transmittance data.72
5.3.1.2 Optical band gap and urbach energy for Copper Oxide thin films 75
5.3.2 Electrical characterization of Copper Oxide thin films 81
5.3.3 Structural characterization ofCu.O, thin films (as-deposited) 83
5.3.4 Elemental composition of Cu.O, thin films (as-deposited) 84
5.4 Solar cell fabrication 85
5.4.1 Optimized conditions for solar cell fabrication 84
5.4.2 Solar cell characterization 85
CHAPTER SIX
CONCLUSION AND OUTLOOK
6.1 Conclusions 88
6.2 Recommendations 90
REFERENCES 91
Appendix I: Photograph of Edwards Auto 306 vacuum coater. , 97
Appendix II: Photograph of Keithley Source Meter 2400 model. 97
Appendix III: Photograph Sheet Resistivity Measurement system 98
Appendix IV: Photograph of MiniPal2 XRF Spectrometer. 98
Appendix V: Photograph ofPW 3040/60 X'Pert XRD diffractometer. 99
Appendix VI: Solar cell simulator 99
Appendix VII: UV -VIS NIR Spectrophotometer solid state 3700 DUV machine 100
Xl
LIST OF TABLES
Table 1.1: Types of p-n junction solar cells and conveision efficiency. 3
Table 5.1: Values of optical band gap at different doping levels for AZO. 60
Table 5.2: Values of the Urbarch energy diffeent doping levels of ZnO with AI. 65
. .
Table 5.3: A summary of electrical surface sheet resistivity for AZO thin films. 66
Table 5.4: XRF elemental percentage composition of optimized AZO thin films. 70
Table 5.5: Optical bandgap for Copper Oxide at different oxygen flow rate. 77
Table 5.6: A summary of electrical surface sheet resistivity for Copper Oxide. 80
Table 5.7: XRF elemental composition of optimized Copper Oxide thin films. 82
Table 5.8: Deposition conditions for optimized Copper Oxide. 84
Table 5.9: Table showing optimized deposition paraneters for AZO 85
XlI
LISTOFFIGURES
Figure 3.1: Picture showing Nellis power plant in North America 21
Figure 3.2: Schematic diagram of an equipment solar cell circuit 22
Figure 3.3: Schematic diagram ofI-V curves in dark and light modes 24
Figure 3.4: Schematic diagram of an illuminated I-V curve 24
Figure 3.5: Schematic diagram of photo voltaic effect on illumination 30
Figure 3.6: P-n heterojunction solar cell in dark and light mode 31
Figure 3.7: Outline of sputtering phenomenon 34
Figure 3.8: A schematic ofthennal evaporation in vacuum 36
Figure 3.9: Schematic diagram showing direct and indirect band transitions 41
Figure 3.10: Parabolic bands with tail state exponentially decaying into band gap 44
Figure 4.1: Schematic diagram of four point probe 52
Figure 4.2: Schematic diagram of) CuxOy-ZnO:Al solar cell 53
Figure 5.1: Spectral graph of transmittance (%) against wavelength (nm)
various doping levels of AZO. 56
Figure 5.2: Spectral graph of reflectance % in the visible, IR, and UV regions against
wavelength (mn) for different doping levels of AZO. 57
Figure 5.3 (a-e): Transmittance spectra fur ZnO doped with AI at various Al doping
Xlll
Levels
Figure 5.4: Variation of optical band gap for ZnO:AI with various doping levels of
Alwninum (%)
Figure 5.5: Graph of variation of absorption coefficient for various Al doping
levels against photon energy
Figure. 5.6: Graph of Inn against energy for various AI doping levels
Figure 5.7: Graph of Aluminum doping levels with sheet resistivity
Figure 5.8: XRD spectra of optimized AZO thin film
Figure 5.9: XRF spectra of as-deposited optimized AZO thin film
Figure 5.10: Variation of transmittance (%) with Oxygen flow rate
Figure 5.11: Spectral graphs showing variations of reflectance (%) with Oxygen flow
rate in Sccm
Figure 5.12 (a-d): Simulated and experimental graphs of sampled thin films of Copper
Oxide at various Oxygen flow rates (Sccm) 76
Figure 5.13: Variation of a a versus energy for different oxygen gas flow rates 77
Figure 5.14: Graph of optical band gap with various levels of Oxygen flow rate 79
Figure 5.15: Graph for determination of urbach energy for Copper oxide films 80
Figure 5.16: Variation of sheet resistivity with oxygen flow rates 82
Figure 5.17: XRD spectrum for optimized Cu.O, thin film 83
59
62
63
64
67
69
70
71
72
XIV
Figure 5.18: XRF spectrum of as-deposited optimized Cu.O, thin film 84
Figure 5.19: Schematic diagram for CuxOy-ZnO:AI p-n junction solar cell 87
Figure 5.20: Solar cell characteristics for the fabricated Cu.O, - ZnO:AI p-n junction 88
xv
ABBREVIATIONS AND ACRONYMS
AM
AZO
ASTM
CB
CIS
CPV
CVD
DC
E-H
EBSD
FF
HEP
IR
I-V
LabVIEW
LED
LVD
MBE
MOCVD
NASA
PDT
PEC
Air Mass
Aluminum doped Zinc Oxide
The American Society for Testing and Materials
Conduction Band.
Copper Indium Selenide
Cyclic Photovoltametry.
Chemical Vapour Deposition
Direct Current.
Electron-Hole
Electron Backscatter Diffraction
Fill Factor.
Hydro Electric Power
Infra Red
Current versus Voltage
Laboratory Virtual Instruments Engineering Workshop
Light Emitting Diode.
Liquid Vapour Deposition
Molecular Beam Epitaxy
Metal Oxide Chemical Vapour Deposition.
National Aeronautic and Space Administration
Post Deposition Treatment.
Photo Electrochemical
XVI
PVD Physical Vapour Deposition.
PV Photovoltaic
TCO Transparent Conducting Oxide
TL Tauc Lorentz
VB Valence Band
UV Ultra Violet
XRD X-Ray Diffraction
XRF X-Ray Florescence
XVll
SYMBOLS
o.,o,
CU20
CuO
CdTe
Eg
Copper Oxide
Cuprous Oxide
Cupric Oxide
Cadmium Telluride
Band gap energy.
Phonon energy
Fermi level
Germanium
EF
Ge
GaAs Gallium Arsenide
Dark saturated current.
Short circuit current
Open circuit voltage.
Aluminum doped with Zinc Oxide
Conversion efficiency
Sheet resistivity
ZnO: Ai
11
XVlll
ABSTRACT
Semiconductors thin films have found applications in various optoelectronics devices. Of
much interest is the application in solar cells. Consequently, this has led to a continuous
search for new semiconductor materials for solar cell applications. Various elements and
compounds have been used to fabricate thin film semiconductors for solar cell
applications like Si, AI, GaAs, Culnxe», CuO among others. In this work, Cu.Oc and
ZnO:AI thin films on plain glass substrates have been deposited as single thin films for
property characterization by reactive DC Magnetron Sputtering and evaporation
techniques using an Edward Auto 306 Magnetron Sputtering System respectively. A
CuxOy-ZnO:AI p-n junction was fabricated by reactive DC magnetron sputtering and
reactive thermal evaporation tec1mique on a glass slide with silver as contacts.
Transmittance and reflectance data in the range 300 nm-2500 nm were obtained using
UV-VIS NIR Spectrophotometer Solid State 3700 DUV for all the thin films samples that
were prepared. Transmittance values of above 70% for Copper Oxide and above 80 % for
AZO were observed. The optical measurements were simulated using SCOUT 98
software to determine optical constants and optical bad gap of the thin films. Band gap
values of 1.62 eV - 2.54 eV are observed for Copper Oxide and 3.18 eV - 3.42 eV for
AZO. The surface sheet resistivities at room temperature of 298 K were found to vary
with the deposition parameters and film thickness. Urbach energy for AZO was found to
increase with doping levels from 2.08 x 10-4to 2.18x 10-4 and varied between 0.6 x 10-4
to 1.92 x 10-4for Copper Oxide. Current-Voltage (I-V) characteristic of the fabricated p-
n junction was obtained using solar simulator. From the I-V characteristic the following
parameters were determined; open circuit voltage Voc = 0.579 V and the short circuit
current Isc= 1.12 mAlm2 resulting in 0.42 % solar cell efficiency.
1CHAPTER 1
INTRODUCTION
1.1 Background to the study
Solar cells or photovoltaic's (PV) are devices that convert solar energy into electricity.
This source of electrical energy is enviromnent friendly and inexhaustible. However the
energy supply from the mostly exploited sources of energy like hydro electric power
(HEP), geothermal, nuclear and fossil fuels among others are yet to meet the demand.
Some of them have limited availability and pose grave danger to the environment due to
their byproducts. A case in point is the adverse climatic changes attributed to increased
Carbon emission from fossil fuel consumption. Photovoltaic's form one of the most
reliable alternatives to large scale production of energy. This solar energy, using the
cheaply fabricated optoelectronic devices, can be harnessed to produce power both for
home and commercial usage. Compared to other sources of energy, it is relatively easy
and cost effective to install and use solar cells (Markvart, 1998). Photovoltaic power
generation is reliable, and involves no moving parts and the operation and maintenance
costs are very low.
There is a growing need for thin film materials with good optoelectronic properties in the
visible, infra-red and ultra violet spectral regions. These materials have found use in
optoelectronic devices like solar cells. Copper Oxide (CuxOy) and Aluminum doped Zinc
Oxide (ZnO:Al) or AZO offer high potential in solar cell applications. Copper based solar
2cells are very stable and have longer operational lifetimes. These materials are suitable
for solar cell applications due to the following properties.
• Direct energy band gap.
• High absorption coefficient.
• Can be fabricated in thin film format for diversified applications with
minimum strain on material usage and device fabrication costs.
• Low band gap (2.2 eV - 2.9 eV for Cu20 and 1.3 eV - 2.1 eV for CuO),
absorption in the VIS range of solar radiation (Ogwu et al., 2005).
• Transparent polycrystalline semiconductors.
On the other hand, ZnO:AI has been used as a transparent conducting oxides (TCOs)
exhibiting superior properties appropriate for photovoltaic applications such as, good
optical transmission, high conductivity and stability against degradation.
1.2 Types of solar cells
There are two mam types of solar cells: p-n junction (conventional) and photo
electrochemical (Waita, 2008). P-n junction solar cells are further classified into three
main types of solar cells, which are distinguished by the type of crystal used in them.
They are mono crystalline, polycrystalline, and amorphous. The production of single
crystal silicon places high requirements on the seed material. The parameters such as
dislocations densities, point defects and impurity concentrations must be minimal.
3Monocrystalline rods are extracted from molten Silicon through the Czochralski process.
(Stanley and Richard, 1986). This production process guarantees a relatively high level of
efficiency. The production of polycrystalline cells is more cost-efficient. In this process,
liquid silicon is poured into blocks that are subsequently sawed into plates. During
solidification of the material, crystal structures of varying sizes are formed, at whose
borders defects emerge. As a result of this crystal defect, the solar cell is less efficient. If
a silicon film is deposited on gJass or another substrate material, the result is called
amorphous or thin-layer cell. The layer thickness amounts to less than l um (Markavat,
1998; Ali and Rndny, 2007). The production costs of amorphous based Silicon thin film
solar cell are lower due to the lower material volume required and the reduced energy
consumption per unit watt to fabricate these films. However, the efficiency of amorphous
cells is much lower than that of the other two cell types. As a result, they are used mainly
in low power equipment, such as watches and pocket calculators, or as facade elements.
Table 1.1: Types of p-n junction solar cells and conversion efficiency (Gupta et al., 2009)
Material Efficiency in Efficiency of
laboratory (%) industrial cell
(%)
monocrystalline silicon ::::::24 14-17
polycrystalline silicon ::::::18 13-15
Amorphous silicon ::::::13 5-7
4In order to provide suitable power for different applications, solar cells are connected
together to form larger units called modules. Cells connected in series have a higher
voltage, while those connected in parallel produce more current. The interconnected solar
cells are usually embedded in transparent ethylene vinyl acetate, fitted with an aluminum
or stainless steel frame, and covered with transparent glass on the front side to make a
module. Typical peak power ratings of such solar modules range from lOW to 100 W
(Gupta et al., 2009). For international standards, solar cells are tested the standard test
conditions of 1000 W/m2 solar radiation and cell temperatures of 298 K..
Group I-UI-VI2 chalcopyrite-structured compounds and related solid solutions are
currently of interest for use as visible-light absorbing layers in thin-film solar cells.
Depending on the deposition conditions and post-treatment, thin films of CIS (Copper,
Indium Selenide) and related materials might exhibit electronic or mixed ionic-electronic
conductivity (Pethe et al., 2004). They can be p or n-type semiconductors, due to the
presence of point defects. The acceptor defects are Copper- or Indium-ion vacancies
while the donor character is due to the Copper- or Indium - ion interstitials, and/or
Sulphide ion vacancies. In addition to these simple defects, Raman studies by Tang et al.
(2000) showed the presence of anti-site defects. The anti-site defects, related to the
occurrence of Cu-Au ordering, deviate from the chalcopyrite structure and determine the
quality of the CIS layer and the efficiency of the solar cell. The defect chemical
equations, using the Kroger-Vink notation, (Tang et al., 2000) are proposed for the CIS
material as basis for explaining the defect formation and conduction mechanisms. A
5similar approach can be used to describe the defect structures of the related materials, like
CuGaS2, Culn'Se-, Cu (Galn)S2 (Murat 1984)
Copper Oxide is an attractive material for photovoltaic applications. It is naturally a p-
type semiconductor due to negatively charged Copper vacancies with a direct band gap of
approximately 2 eV (Akimoto et al., 2006) which is sufficiently close to the optimal band
gap under AM1.5 radiation spectrum. It is abundant on earth, non toxic, and exhibits
fairly high minority carrier diffusion lengths, high absorption coefficient in the visible
region, and large exciton binding energy (Oba el af., 2005). The theoretical photoelectric
conversion efficiency is 20%. However, there have been relatively few attempts to
fabricate photovoltaic cells with Cuprous Oxide. The optimization of CU20 solar cells is
slowed down by the lack of clear understanding of its electronic and thermodynamic
properties, and by the difficulties in the doping process. Early attempts focused on metal
Cu20 Schottky solar cells because CU20 is natively p-type and is hard to be doped n-type.
However, it is believed that there will not be a significant improvement with this
structure beyond the highest practical photoelectric conversion efficiency currently
reported at 1.76% (Mittiga., et al 2006). When CU20 is placed in contact with a metal to
form a Schottky barrier, most metals reduce CU20 to form a Copper rich region at the
interface, which dictates the barrier -height magnitude (Gupta et al., 1998). This is why
the barrier heights range between 0.7 and 0.9 eV, regardless of the choice of metal. These
oxidation phenomenon's coupled with inter diffusion of Copper results in low
efficiencies in the order of 1%, necessitating design of heterojunction.
6To construct a p-n heterojunction, other n-type materials can be used to mate with the
CU20, such as ZnO, In203, Sn02 and CdO. The highest photoelectric conversion
efficiency reported to date, 2%, was achieved by Mittiga, (2009) using a CU20/ZnO
heterojunction. Due to its low electron affinity, a ZnO layer increases the shunt resistance
and decreases the dark current, resulting in a higher open circuit voltage compared to
indium tin oxide (ITO) (Mitiga et al., 2006). ZnO has a hexagonal (wurtzite) structure
while Cu20 has a cubic (cuprite) structure. The lattice parameter of Cu20 at room
temperature is 4.27 ,which leads to the nearest atomic distance of 3.02A. The ZnO
(001) plane and the CU20 (111) plane have similar atomic arrangements with only 7.1%
lattice mismatch, making these two planes the preferred atomic arrangement at the
interface. For the reasons stated above, in this report, we will focus on CuxOylZnO
heterojunction solar cells and explore methods to improve upon the current structure.
Thin film p-Cu.Oc/n-Znt) solar cells may provide a solution for a potentially cheap, large
area, stable photovoltaic structures. However, their conversion efficiency is still low. The
low efficiency results from the presence of defects at copper oxide layer (Mittiga et al.,
2009).
1.3 Statement of the research problem and justification
Various elements and compounds have been studied to fabricate thin film semiconductors
for solar cell applications. Some of the materials include; silicon (Si), gallium arsenide
(GaAs) and Copper (II) Oxide (CuO) among others. Silicon based solar cells are bulky
and require a lot energy and high purity to fabricate. This increases the cost of Silicon
based solar cells hence reducing their affordability. Thin film p-Cu.Oc/n-Zni) solar cells
provide a solution for a potentially cheap, large area, stable photovoltaic structures. They
are direct band gap materials whose constituent materials are easy to fabricate. In this
research study, both Cu.O, and ZnO:Al are deposited and studied for solar cell
applications. Zinc Oxide is doped with Aluminum to improve its optical and electrical
properties for solar cell application.
1.4 Objectives
1.4.1 Main objective.
To characterize thin films of Copper Oxide (CuxOy) and Aluminum doped Zinc Oxide
(ZnO:Al) deposited by reactive de magnetron sputtering and reactive thermal evaporation
respectively for fabrication of a p-n junction for solar cell applications.
1.4.2 Specific objectives
(i) To deposit Cu.O, thin films on glass substrates by means of reactive de
magnetron sputtering.
(ii) To deposit ZnO:Al thin film semiconductor by reactive thermal evaporation.
(iii) To study the optical and electrical properties and of Cu.O, and ZnO:AI thin
films.
(iv) To study the structural properties of Cu.O, and ZnO:Al thin films.
8(v) To fabricate CuxOy-ZnO:Al p-n junction.
(vi) To study the I-V characteristics of the Cu.Oc-Znfr.Al p-njunction
1.5 Rationale
The world primarily depends on hydroelectric power for its energy needs. However, as it
has become evident lately, this energy source is no longer able to meet the world's energy
needs due to rising population and unreliable rainfall patterns. The world therefore needs
other alternative sources of energy. We have a lot of solar energy hence the need to
search for new materials for solar cell applications. Copper Oxide (Cu.Oc) and
Aluminwn doped Zinc Oxide (ZnO:Al) semiconductor materials have found considerable
interest in solar cell applications due to their interesting properties such as direct band
gap and high absorption coefficient.
9CHAPTER 2
LITERATURE REVIEW
2.1 Introduction
A solar ceU is a device that is gaining widespread application particularly in the realms
of renewable energy due to the conversion of light via photovoltaic effect to electricity.
This chapter gives a background on the recent developments and highlights the ongoing
research areas in solar cell technology. Special attention is focused on characteristics of
various generations of solar cells devices including the more recent chalcopyrite based
solar cells which have been characterized with high efficiencies. Towards the end of this
chapter, the research on unconventional materials that are applicable in solar cell research
particularly those based on Copper Oxide and AZO will be presented and discussed to
show the open questions in research on these materials.
2.2 Background to solar cell technology
In 1839, a French physicist A. E. Becquerel recognized the first photovoltaic effect
(Becquerel 1839, Williams 1960). However in 1883, Charles Fritts built the first solar
cell by coating selenium with an extremely thin layer of gold to form the junction. The
device was about 1% efficient. The modem junction semiconductor solar cell was
patented by Russell Oh1in 1946 (Chapin et al., 1954).
10
Development of solar cells can be classified into three generations. The classification is
based on the order of which each became important. The first generation solar cells
consist of large-area, high quality and single junction devices. They emphasize on high
energy and labour inputs which hinder any progress in reducing production costs. Second
generation materials are developed to address energy requirements and production costs
of solar cells. They are based on thin film technology. These involve using other
fabrication techniques such as solution deposition, vapour deposition, electroplating, and
use of Ultrasonic Nozzles. These techniques are advantageous since they reduce
temperature effects during fabrication. The most successful second generation materials
have been cadmium telluride (CdTe), copper indium Gallium Selenide, amorphous
silicon and micromorphous silicon (Hirshman et al., 2008). These materials are applied in
a thin film to a supporting substrate such as glass or ceramics, hence reducing material
mass and subsequently reducing production costs.
Third generation technologies aim to enhance poor electrical performance of thin-film
technologies while maintaining very low production costs. Current research is targeting
increasing conversion efficiencies while retaining low cost materials and manufacturing
techniques (Green, 2002). To achieve this, various fabrication techniques are employed.
This includes;
(i) The use of Multijunction photovoltaic cells.
(ii) Concentration of the incident spectrum.
(iii) The use of a wider electromagnetic spectrum such as thermal generation by
11
ultra violet (UV) light to enhance voltage or carrier collection, or the use of
the infrared spectrum for night-time operation.
(iv) Use of thin film materials.
Current research aims at bringing down the cost of producing solar cells at the same
time increasing the conversion efficiency. This involves increasing the efficiency, and
decreasing the cost of the solar cells per generated unit of power.
Various studies have been done on both Copper Oxide and AZO as possible materials for
solar cell applications. ZnO:Al was deposited on-surfaces of Sn02:F substrate by Zhao
and others. The extent of degradation was monitored, its role as a back reflector were also
studied. Samples which had been coated with ZnO:AI showed little degradation and very
good solar cell properties (Zhao et al., 2002). Doped films of ZnO were studied and
found to exhibit high transmittance in visible spectra region, high reflectance in infra red
(LR) region and relatively good conductivity. Optical, electrical and structural properties
heavily depended on deposition parameters (Tang et al., 2000). In both the above cases
the photoelectric conversion efficiency was very low in range 11 :s 0.2 (Yo
Copper based solar cells were studied and found to exhibit good solar cell properties.
CulnSe, has a direct band gap of 1 eV and a high absorption coefficient compared to Si
and GaAs (Pethe et al., 2004) thus making it the lead candidate for solar cell device
fabrication due to high solar conversion efficiency and low processing costs. Other
12
alternative class of materials consists of the ZnOICd/CulnSe thin films which have been
found to yield a record efficiency of 14.1% (Mitchell et al., 1991).
Jaeyoung and yongsug, (2000) studied selective electrodepositing of ZoO onto CU.20.
The co-deposition of Cuprous Oxide and Zinc Oxide on indium tin oxide (ITO) substrate
was executed by two different electrochemical methods and formation mechanism of
ZnO onto Cu20 were investigated. The optical properties were found to depend on
deposition parameters. The electrical properties and the interfaces of CU20/ZnO/ITO p-i-
n were fabricated by electro-chemical deposition method. Its current-voltage (I-V)
measurements and characteristics were studied and found to have smaller turn on voltage
than the barrier potential. This was due to interface defects (Zhang et al., 2004). Sheng et
al., (2006) prepared a polycrystalline p-Cu20/n-ZnO junction for use in solar cells. Two
deposition sequences were studied, ZnO deposited on Cu20 and Cu20 deposited on ZnO.
It was found that crystallographic orientation and I-V characteristics of the interface were
highly influenced by deposition sequences. Both being far superior for junctions with
CU20 on ZnO than for inverse structure. They successfully obtained photo-response for
the first time in deposited thin film of CU20/ZnO however the efficiency recorded was
less than 1 %. Su Sheng et at. (2006) prepared transparent p-n junction diode using p-
type SrCu202 and n-type ZnO:AI on glass substrate. The junction exhibited non linear
and rectifying 1-V characteristics. Electrodeposited thin films of ZnOlCu20
heterojunction solar cells have rectifying interface and suitable for oxide based solar cells
or diodes (Ozawa, et al., 2009).
13
The challenge of increasing the photovoltaic efficiency is thus of great interest both
from the academic and economic points of view. This study aims at using thin film
technology to fabricate solar cells from cheap and available materials.
2.3 Copper Oxide
Copper Oxide is a semiconductor which shows varying optical behaviour because of
stoichiometric deviations arising from its methods of preparation and parameters (Ogwu
et al., 2005). It has been reported that many of the methods of growing Copper Oxide
result in a combined growth of Copper (1) Oxide (CU20) and Copper (II) Oxide (CuO)
(Ogwu et al., 2005). A range of direct optical band gap energies has also been reported
for Cu20 and CuO (Balamuruga and Mehta, 2001) semiconductor films in the literature,
depending on the method of fabrication and stoichiometry. Sputter deposition of Copper
Oxide films on glass are reported to have high transparency, with a slightly yellowish
appearance, and absorbs usually at wavelengths below 600 nm, whilst CuO absorbs
strongly throughout the visible spectrum and is black in appearance. The current
applications areas of Copper Oxide thin films include solar cells and electro-chromic
devices (Richardson et al., 2000). Copper Oxide films have been reported to have band
gap energy values which make them suitable for application as absorbers for solar energy
conversion (Ogwu et al., 2005)
14
Copper Oxide semiconductors have been studied for several reasons such as: the natural
abundance of starting material Copper (Cu); the ease of production by Copper oxidation;
their non-toxic nature and the reasonably good electrical and optical properties exhibited
by Cu20. Cupric Oxide (CuO) is a p-type semiconductor having a band gap of 1.21-2.1
eV and monoclinic crystal structure (Balamurunga and Mehta, 2001). Cuprous oxide
(Cu20) is also a p-type semiconductor having a band gap of approximately 2.2 - 2.9 eV
and a cubic crystal structure (Balamurunga and Mehta, 2001). 'Its high optical absorption
coefficient in the visible range and reasonably good electrical properties constitute
important advantages and render CU20 as the most interesting phase of copper oxides
(Ogwu et al., 2007). Several methods such as: thermal oxidation (Ozawa, 2009);
electrodeposition (Ogwu et al., 2005); chemical conversion (Laurie, 1989); chemical
brightening (Ogwu et al., 2005); spraying (Balamurunga and Mehta, 2001); chemical
vapor deposition (Maruyama, 1998); plasma evaporation (Balamurunga and Mehta,
2001); reactive sputtering (Drobny and Pulfrey , 1979 ); and molecular beam epitaxy
(Kita et al., 1994) have been used to prepare Copper Oxide thin films. In most of these
studies, a mixture of phases of Cu, CuO and CU20 is generally obtained and this is one of
the nagging problems for non-utilizing CU20 as a semiconductor. Pure CU20 films can be
obtained by oxidation of Copper layers within a range of temperatures followed by
annealing for a small period of time. CU20 is one of the oldest semiconducting materials
which are useful in solar cell applications.
15
Balamuruga and Mehta, (2001) studied nanocrystalline Cu20 thin films. The films were
synthesized using an activated reactive evaporation technique. Structural and optical
characterizations of these films were carried out using, glancing angle X-ray
diffractometer; Fourier transform infrared spectrometer; transmission electron
microscope; and UV-VIS-NIR spectrophotometer. The nanocrystalline size in these films
was varied by changing the deposition parameters. Furthermore, optical spectroscopy
studies on nanocrystalline CU20 showed a direct allowed transition and a shift in the
optical absorption edge from the bulk value with nanocrystalline size and stoichiometry
of these films. These results show that a single phase nanocrystalline CU20 thin films can
be synthesized at a relatively low substrate temperature using the activated reactive
evaporation technique. These studies indicate that nanocrystallinity results in the stability
of cubic Cu20 phase in these films.
Papadimitropoulos et al., (2006) studied Copper Oxide films which were grown by
oxidation of vacuum evaporated Copper layers on silicon substrates. Oxidations were
performed at atmospheric pressure, in a nitrogen-oxygen mixture 10% in oxygen and at
temperatures varying between 185°C and 450 DC, the optical properties of films were
studied with spectroscopic ellipsometry measurements within the energy range 1 to 3.5
eV. These measurements were analysed using the Tauc-Lorentz (TL) model
(Papadimitropoulos et al., 2006) to simulate the dispersion of the complex refractive
index of disordered films. It was shown that the TL model describes satisfactorily the
refractive index dispersion of these Copper Oxide films. The band gap, as defined by the
16
TL model, was found equal to 2.3 eV for CU20 and between 1.05 eV and 1.2 eV for CuO.
It was shown that the gap of the CU20 films was free of localized states, which was not
the case for the gap of CuO.
Ogwu et al., (2005) sputtered Copper Oxide films on glass substrates by reactive radio
frequency (rf) magnetron sputtering, using a solid Copper target and an Argon-Oxygen
gas atmosphere. Optical transmission in the prepared films was measured by
spectrophotometer in the 400-850 run wavelength regions. A maximum transmission of
between 40% and 80% for Copper Oxide films prepared at a low rf power of 200 W was
observed, for the Oxygen flow rates investigated. The optical band gap values of the
films ranged between 2.05 and 2.4 eV. Hailing Zhu et al., (2009), prepared Cu20 thin
films on quartz substrate by reactive direct current magnetron sputtering. The influences
of Oxygen partial pressure and gas flow rate on the structures and properties of deposited
J
films were investigated. The as-deposited Cu20 films were found to have a very high
optical absorption in the visible spectra under 600 run and were endowed with
photocatalytic reactivity under the visible light
2.4 Aluminum doped ZnO
Zinc Oxide (ZnO) films have also become technologically important due to their range of
electrical and optical properties, together with their high chemical and mechanical
stabilities, which make them suitable for a variety of applications such as flat panel
17
display electrodes and gas sensors. Moreover, these films can be used as surface acoustic
wave devices, because of their large piezoelectric constant, and also as solar cells, since
their optical band gap (3.3 eV) is wide enough to transmit most of the useful solar
radiation (Selmi et al., 2008). ZnO is an n-type semiconductor and its conductivity can be
controlled by thermal treatment or by adequate doping (Kyungsoo Jang et al., 2009). The
doping of ZnO films with the group III elements can increase the conductivity of the
films. In comparison with other elements, Aluminum and Gallium are the best dopants
because their ionic radii are similar to that of Zn2+ (Rodrigo et al., 2003 Atanas et al.,
2010).
Many techniques have been used for fabricating ZnO films, such as chemical vapour
deposition, pulsed laser deposition, de reactive sputtering, spray pyrolysis and the sol-
gel process (Rodrigo et al.. 2003). The structural, optical, and electrical properties of
ZnO and ZnO :Al films prepared by thermally evaporating zinc acetate and AICh in
vacuum have been investigated in detail, together with the effects of heat treatment in air
and vacuum. The properties of the deposited ZnO and ZnO:Al films depend on the
deposition parameters such as substrate temperature, evaporation rate of zinc acetate and
Aluminum concentration (Jin Ma et al., 2000).
The Al-doped ZnO film exhibits remarkable electrical conductivity, together with high
charge carrier density and mobility (Su-Shia Lin et af., 2005). The ZnO doped with Ae+
is used extensively for photo-electronic devices (Su-Shia Lin et al., 2005), Theoretically,
18
spatially organized ZnO doped with A13+ could result in improving electrical properties.
For this reason, only few studies of the conduction mechanism in heavily Al-doped ZnO
films have been reported. Slightly doping was explained by a limited incorporation of Al
into the ZnO lattice, and Aluminum acts as a donor (Su-Shia Lin et al., 2005).
Selmi et al., (2008) studied deposition time and its effects on the properties of ZnO:AI
films. It is shown that films grow with the hexagonal c-axis perpendicular to the substrate
surface. The morphological characteristics show a granular and homogenous surface and
the cristallinity of the films are enhanced with increased deposition time. The deposited
films show good optical transmittance (80%-90%) in the visible and near infrared
spectrum.
Transparent conductive oxides (TCO) films are degenerate wide band gap
semiconductors with low resistance and high transparency in the visible range. For these
reasons these materials, are widely used in optoelectronic applications such as flat panel
displays, solar cells and electro chromatic devices. Usually, the TCO films are n-type
semiconductors such as Indium Tin Oxide (ITO), Tin Oxide (Sn02) and Zinc Oxide
(ZnO), whereas ITO film is the one most used in these devices up to now. Recently, Al
doped ZnO (ZnO:Al) film is one of the materials which could replace the ITO films. The
direct optical band gap of ITO films is generally greater than 3)5 eV although a range of
values from 3.5 to 4.06 eV have also been reported in the literature (Gupta et al., 1989)
19
This research uses two materials, namely ZnO:Al and Copper Oxide. ZnO:Al is widely
used because the films have electrical and optical properties similar to those of ITO, and
because it is stable in a hydrogen atmosphere. Thin ZnO films should be doped by
aluminum, since it has been remarked that extrinsic donors due to the dopant atom are
more stable than intrinsic donors due to the native defects. The electrical conductivity in
ZnO:Al film is higher due to the Al3+ ions in substitutional sites of the Zn2+ions and the
Aluminum interstitial atoms, in addition to Oxygen vacancies and Zinc interstitials
(Atanas et al.. 2010). The Bond enthalpy of Zn-O is 159±4 kJ/moI while that of Al-O is
511±3kJ/mol (Atanas et al.. 2010).
Copper Oxide is an attractive material for photovoltaic applications. It is a naturally a p-
type semiconductor due to negatively charged Copper vacancies with a direct band gap of
approximately 2 eV (Mittiga et al., 2009) which is sufficiently close to the optimal band
gap under AM 1.5 radiation spectrum. It is abundant on earth, non toxic, and exhibits
fairly high minority carrier diffusion lengths, high absorption coefficient in the visible
region, and large exciton binding energy (Mittiga et al., 2009). The theoretical conversion
efficiency is 20%.
20
CHAPTER 3
THEORETICAL BACKGROUND
3.1lntroduction
This chapter outlines the theory on deposition techniques and more specifically on
sputtering and thermal evaporation as the techniques adopted in this study. Also
presented are theoretical backgrounds on solar cells, optical characterization, thin film
resistivity measurements, and structural properties of th:in films.
3.2 Solar cells
Solar cells represent the fundamental power conversion unit of a photovoltaic system.
Solar energy is non-exhaustible (sun's life span is estimated at 10 billion years), non-
polluting and free (Deshmukh and Deshmuck, 2008). Solar cells are basically
multijunction semiconductor devices made from thin films which are usually p-n or p-i-n
diode. They convert sunlight directly into electricity by exploiting photovoitaic effect
(Markvat, 1998). In 1967 Horace de Sasura built the first world solar collector. In 1807,
C Gunter invented the solar boiler, sparking the current research into utilization of solar
energy (Markavat and Castener, 2005). P-njunction solar cells are common for domestic
and industrial use. For practical operation solar cells are usually assembled into modules.
Solar ceUs must overcome the following challenges; performance, reliability, durability,
efficiency, availability and afford ability. Photovoltaic energy conversion in solar cells
consists of three essential steps namely
21
(i) Charge generation: This entails photon absorption to create electron hole
pairs.
(ii) Charge separation and
(iii) Charge transport.
Figure 3.1: Picture showing Nellis Solar Power Plant inNorth America.
(httpllsolarcooking.orglsaussaure.h~ 22nd Jan 2011)
An ideal solar cell can be represented by a current source in parallel with a rectifying
diode as shown in the equivalent diode circuit shown in .figure3.2.
22
ID ISH +
v
Figure 3.2: Schematic diagram of an equivalent solar cell circuit (Markvart, 1998).
From the equivalent circuit it is evident that the current produced by the solar cell (1) is
equal to the current produced by the current source (IJ, minus the current flowing
through the diode (ID) and the shunt resistor (ISH) (Markvart, 1998).
1=h -lD - ISH (3.1)
The current through these elements is governed by the voltage across them:
Vj=V+IRs (3.2)
Where V;= voltage across both diode and resistor RS!1, V = voltage across the output
terminals I = output current (Amperes) and Rs = series resistance. Using the Shockley
diode equation the current diverted through the diode is given by (Markvart, 1998).
1=10 {exp [!!!L] -I}nKT (3.3)
Where 10= reverse saturation current, n = diode ideality factor q = elementary charge
23
k = Boltzmann's constant and T= absolute temperature. By Ohm's law, the current
diverted through the shunt resistor is (Markvart, 1998).
v.
I.e;}'f = _._.J_
RSlf (3.4)
Where RSH = shunt resistance. Substituting these into the first equation produces the
characteristic equation of a solar cell, which relates solar cell parameters to the output
current and voltage. (Markvart, 1998).
I - I I- {, [q(F +- I Irs)] 1} l.f+- I Irs
- L. - U . cxp nk7~ - - [lS'1I . (3.5)
Since the parameters fa, n, Rs, and RSH cannot be measured directly, the most common
application of the characteristic equation is nonlinear regression to extract the values of
these parameters on the basis of their combined effect on solar cell behaviour.
3.3 Theory of I-V Characterization
Photovoltaic (PV) cells are modeled to look like a current source in parallel with a diode.
When there is no light present (dark mode), the PV cell behaves like a diode. In the
presence of light (light mode) current is generated by the PV cell. This is illustrated i11
Figure 3.3
24
Voltage (V)
Diode (In dark)
Solar cell (On U1omination)
Figure 3.3: Schematic diagram of l-V curves in showing both light and dark mode ofa
Diode (Markvart, 1998)
The current-voltage (I-V) curve of PV cell in light mode has the shape shown in Figure
3.4. Voltage across the measuring load increases from zero to open circuit voltage Voc.
Performance parameters for the PVcells can be determined from this data
I~~--------------------------------I.-c:
aximum pow r
re I
Voltage Ym Yoc
Figure 3.4: Schematic diagram of an illuminated I-V Curve in light mode for ideal
Solar cell (Markvart and Castener, 2005)
25
3.3.1 Short Circuit Current (Ise)
This is the largest current which may be drawn from a solar cell. The short circuit current
(Ise) shown in figure 3.4 corresponds to condition when the impedance is low and is
calculated when the voltage equals zero. For an ideal solar cell, this maximum current
value is the total current produced in the solar cell by photon excitation. Ise = IMi\x = Ie
for forward-bias power quadrant (Markvart and Casterner, 2005). The short circuit
current (Ise) depends on;
(i) Area of the solar cell.
(ii) Number of photons of incident light.
(iii) Spectrum of incident light.
(iv) Optical properties of material and,
(v) Collection probabilities of the solar cell.
3.3.2 Open Circuit Voltage (Voc)
This is maximum voltage available from a solar cell. The open circuit voltage (Voc)
occurs at point on the curve when there is no current passing through the cell (figure 3.4).
When the cell is operated at open circuit, I = 0 and the voltage across the output terminals
is defined as the open-circuit voltage. Assuming the shunt resistance is high enough to
neglect the final term of the characteristic equation, the open-circuit voltage Vac is given
as,
voc
(3.6)
26
Voc depends on saturation current of a solar cell, recombination of charge carriers and
doping levels within a material.
3.3.3 Fill Factor (FF)
The Fill Factor (FF) is a measure of squareness of solar cell and also area of the largest
rectangle that will fit the I-V curve. FF is the measure of quality of the solar cell. It is
calculated by comparing the maximum power (PMAX) to the theoretical power (PI') that is
product ofVoc and Isc. A larger fill factor is good, and corresponds to an I-V curve that is
more square-like. Typical values of fill factor range from 0.5 to 0.82. Fill factor is also
represented as a percentage (Markvart and Castaner, 2005);
FF=
I Vm m
I Vsc oc
(3.7)
3.3.4 Efficiency (11)
Efficiency is measure of the ratio of the energy output (Pout) from the PV cell to input
from the sun (Pin) into the PV cell. Pout can be taken to be Pmox since the solar cell can be
operated up to its maximum power output to get the maximum efficiency (Markvart arid
Castaner, 2005).
Isc Voc FF
Pi
(3.8)
Pin is taken as the product of the irradiance of the incident light, measured in W/m2 (1000
W/m2) or in suns (1 sun = 1000 W/m2, with the surface area of the solar cell (m"). The
maximum efficiency (11max) found from a light test is not only an indication of the
27
performance of the solar cell under test, but, like all of the I-V parameters, can also be
affected by ;
(i) Temperature on the solar cell.
(ii) The intensity of incident light and,
(iii) Spectrum of the incident light.
For this reasons, it is recommended to test and compare PV cells using same temperature
and lighting conditions.
3.3.5 Effect of physical size (area of cell)
The values of fv, Rs, and RSH are dependent upon the physical size of the solar cell. In
comparing otherwise identical cells, a cell with twice the surface area of another will, in
principle, have double the fa because it has twice the junction area across which current
can leak. It will also have half the Rs and RSH because it has twice the cross-sectional area
through which current can flow. For this reason, the characteristic equation is frequently
written in terms of current density, or current produced per unit cell area (Markvart and
Castaner, 2005):
{ ..'..·[q(\f ..+t: J.rs:)] 1} _ 'V +'" Jr5'J = ,1L --- ,10 .. ex.p .... -. n.k'T J SlI (3.9)
Where 1 = current density, JL = photo generated current density, 10= reverse saturation
current density, 1'5 = specific series resistance, rSH = specific shunt resistance. This
formulation has several advantages. One is that since cell characteristics are referenced to
a common cross-sectional area they may be compared for cells of different physical
28
dimensions. While this is of limited benefit in a manufacturing setting, where all cells
tend to be the same size, it is useful in research and in comparing cells between
manufacturers. Another advantage is that the current density scales the parameter values
to similar orders of magnitude to make their numerical extraction much simpler and more
accurate even with naive solution methods (Markvart and Castaner, 2005).
There are however practical limitations of this formulation namely,
• Parasitic effects.
• Recombination and
• Contamination of the junction.
Very small cells may exhibit higher values of Jo or lower values of RS/I than larger cells
that are otherwise identical. In such cases, comparisons between cells must be made
cautiously and with these effects in mind. This approach should only be used for
comparing solar cells with comparable layout. For instance, a comparison between
primarily quadratic solar cells like typical crystalline silicon solar cells and narrow but
long solar cells like typical thin film solar cells can lead to wrong assumptions caused by
the different kinds of current paths and therefore the influence of distributed series
resistance Rs.
29
3.4 Solar radiation
The solar energy received varies with geographical locations. A higher solar energy to the
tune of 583.3W/nl per day is received in the Arabian Peninsula and Sahara Desert as
compared to northern countries and lower America (291.65W/m2 per day) (Waita, 2008).
The amount received in Kenya on average falls in between the two zones above. If the
total desert land of area 50 x 107 Km2 receives solar radiation for 8 hours a day, then the
solar energy received is 163.2 x 1012 Kwb/day or 60 x 1012 Kwh/year (Waita, 2008). If
each solar panel is 1% efficient, this would translate to 60 x 1010 Kwh/year (Waita,
2008,). The above figures are too big for the world's energy requirements. If only ten
minutes is utilized per day then it wouid be enough to provide us with the worlds annual
needs (Markvart and Castaner, 2005).
Solar radiation is produced by fusion reaction like combination of hydrogen with helium
which takes place in sun's interior (Markvart and Castaner, 2005). It is then propagated to
the earth by radiation. The energy from the sun per unit time received on a unit area
perpendicular to the direction of propagation of radiation to the earth mean distance from
the sun outside the atmosphere is called solar constant. The National Aeronautics and
Space Administration (NASA) and The American society for testing and materials
(ASTM) standard value for solar cell constant is 1353.W/m2 (Waita, 2008).
30
3.4.1 Absorption of light by a p-n junction
Photovoltaic energy conversion relies on quantum nature of light which carry the energy.
E (A)=hcPH A (3.10)
Where h is the plank's constant, c is the speed of light and A is the wavelength. When
solar energy falls on a p-n junction (figure 3.5), photons with energy (hn) greater than
band gap of the semiconductor are absorbed. The absorbed photons create electron-hole
(E-H) pairs to generate a photocurrent upon charge separation that flows from the n-type
to the p-type material (Wieder, 1982). The E-H pairs are separated by an internal junction
electric field. Holes drift to one electrode and electrons to the other (Shar et al., 1999).
Direct current is produced by solar cells as shown in figure 3.5.
e
load
,..-- antireflection coating
front contact
emitter
base
eIectron-hoIe
pair
+
Figure 3.5: Schematic diagram of photo voltaic effect when light is illuminated on a solar
( http:www.pveducation.orglpvcdrom, 12thFeb,20 11)
31
VB2
CB2~---CB2
hv
VBl-----
(a) (b)
Figure 3.6: (a) p-n heterojunction solar cell at thermal equilibrium in dark. (b) P-n
junction solar cell under illumination, open circuit conditions (Kemell,2003)
Figure 3.6, where: 1 and 2 refers to n-type and p-type semiconductor in the p-n junction
respectively, Eg is band gap, EF is fermi level, CB is conduction band, VB is valence
band and Voc is open circuit voltage. When the solar cell is illuminated, E-I-I pairs
increase minority carrier concentration. The potential energy barrier decreases, allowing
the current to flow and a voltage Voc (photo voltage under open circuit voltage) is
generated across the junction (Stone 1993, Tsubomura et al., 1993). Voc is limited by the
band gap energy (Eg) of the absorber material. The maximum value of Voc is calculated
using the following relation (Kemell, 2003);
Emax .Voc =--.L
e (3.11 )
where e is electron charge.
32
3.5 High-efficiency solar cells by material integration
The efficiency of solar cell is low because the energy densities of solar light are low both
per wavelength and per area. A properly designed semiconductor p-n junction can
convert the incident photon energy to electric power with almost 100% efficiency, if the
photon energy is same as the band gap of the semiconductor. However, even if the
incident photon energy increased, it generates the same electric power and the conversion
efficiency decreases. If the incident photon energy is lower than the band gap of the
semiconductor, it cannot generate any electric power and the efficiency becomes zero. As
the spectrum of the solar radiation is fairly broad from UV to lR, only 20% and a little
more of solar energy is converted to electric power by the semiconductor p-n j unction. A
tandem solar cell which has two semiconductor p-n junctions with different band gap can
generate more electric power than a simple solar cell, because it generates electric power
using both shorter and longer wavelength of light, at the top wide gap and at the bottom
narrow gap semiconductor, respectively (Tsubomura et al., 1993). This type of solar cell
is expected to achieve the conversion efficiency higher than 30%. In general, both the
operating voltage (V) and the operating current (I) are lower than the open circuit voltage
Voc and the short circuit current Isc, respectively. The ratio ofVl/Vj; Isc is called fill factor
(FF) and, in general, it increases with the input light intensity. Therefore, if the solar
radiation is concentrated by the use of an optical lens, the efficiency (11) of the solar cell
increases. It also implies that if we use a multi-junction tandem cell without a light
concentrator, the fill factor (FF) will decrease and the total efficiency will not increase.
When a triple-junction cell is fabricated properly, the conversion efficiency will be
33
expected to exceed 40% (Markvart and Cas tener, 2005). However, further increase in the
conversion efficiency is fairly difficult even if a solar cell with more junctions is
prepared. The key technology for multi-junction solar cell is to prepare a composite
material with wide gap and narrow gap semiconductors.
3.6 Deposition Techniques
Deposition techniques for thin films broadly fall in five categories: Physical vapour
deposition (PVD), Chemical vapour deposition (CVD), Oxidation, Spin coating and
Plating. In PVD technique, films are formed by atoms that are directly transported from
source to the substrate through gas phase and they include sputtering, evaporation
(Thermal evaporation and E-Beam evaporation) and Reactive PVD. On the other hand, in
CVD technique, films are formed by chemical reaction on the surface of the substrate and
they include Low-Pressure CVD (LPCVD), Plasma-Enhanced CVD (PECVP),
Atmosphere-Pressure CVD (APCVD) and Metal-Organic CVD (MOCVD). Evaporation
and sputtering are the two main techniques employed in physical vapour deposition. In
this study, due to its advantages such as proper control on the film chemical composition,
high deposition rate and low substrate heating during film deposition, DC reactive
magnetron sputtering technique has been adopted and is discussed in the next section.
34
3.6.1 Sputtering Technique
Sputtering is a physical vapor deposition (PVD) technique whereby bombarding particles
incident on a target collide with surface atoms thus dislodging them from the lattice
through a transfer of energy (Ohring, 1992). The displaced lattice atoms as well as the
bombarding particles (projectile) then undergo collisions with other lattice atoms,
dislodging them and a chain reaction of collision ensues. Atoms with sufficient energy
required to overcome the surface potential called the surface binding energy (Uo) will
escape (Matsunan mi et al., 1980). The schematic diagram below shows sputtering
phenomenon.
1 S··· .. · A
•..... puu ..ero.o,. Atom
Gas
•
Trapped Neutral
Figure 3.7: Outline of sputtering phenomenon (Matsunanmi et al., 1980).
35
There is a minimum projectile kinetic energy needed to induce sputtering called threshold
energy (Et/,) and is given by Bohdansky (1984) as:
(3.12)
where M, is projectile mass, M2 is mean molecular mass per atom of a target and fJ is
maximum fractional energy transfer possible in a head-on collision given by:
/3= 4MJM2
(MJ +M2Y (3.13)
(3.14)
Sputtering processes are wide and varied. They can be divided up into four categories;
DC, RF, magnetron and reactive. There are also important variants within each category
for example DC bias and even hybrids between categories (Almen and Bruce, 1961).
3.6.2 Vacuum Evaporation Technique
Vacuum evaporation is also a widely used PVD technique. It is where materials from a
thermal vaporization source reach the substrate with little or no collision with gas
molecules in the space between the source and substrate. The vacuum environment
during evaporation provides the ability to reduce gaseous contamination in the deposition
system to a low level. This technique is generally done using thermally heated sources
36
such as tungsten WIre coils or by high energy electron beam heating of the source
material itself. Generally, the substrates are mounted at an appreciable distance away
from the evaporation source to reduce radiant heating of the substrate by the vaporization
source. Figure 3.8 depicts a schematic of evaporation phenomenon.
Gas inlet Rotating substrate
..• ..• Vapour~\~4----J-r-
Boat (beam evaporator)
Glass substrate
Movable shutter
Vacuum
Figure 3.8: A schematic ofthennal evaporation in vacuum
3.6.3 Other Deposition Techniques
3.6.3.1 Arc Vapour Deposition Technique
Arc vapour deposition uses a high current, low-voltage arc to vaporize a cathodic
electrode (cathodic arc) or anodic electrode (anodic arc) and deposit the vaporized
material on a substrate. The vapourized material is highly ionized and usually the
substrate is biased so as to accelerate the ions to the substrate surface.
37
3.6.3.2 Ion Plating Technique
Ion plating utilizes concurrent or periodic bombardment of the depositing film by atomic-
sized energetic particles, to modify and control the properties of the depositing film. In
ion plating the energy, flux and mass of the bombarding species along with the ratio of
bombarding particles to depositing particles are important processing variables.
3.6.3.3 Chemical Vapour Deposition Technique
Thermal chemical vapour deposition is the deposition of atoms or molecules by the high
temperature reduction or decomposition of a chemical vapour precursor species which
contains the material to be deposited. Reduction is normally accomplished by hydrogen
at an elevated temperature. Decomposition is accomplished by thermal activation. The
deposited material may react with other gaseous species in the system to give compounds
(e.g. oxides, nitrides). CVD technique has numerous other names and adjectives
associated with it such as Vapour Phase Epitaxy (VPE) when CVD is used to deposit
single crystal films, Metalorganic CVD (MOCVD) when the precursor gas is a metal-
organic species, Plasma Enhanced CVD (PECVD) when a plasma is used to induce or
enhance decomposition and reaction, and Low Pressure CVD (LPCVD) when the
pressure is less than ambient.
38
3.7 Optical and structural characterization of thin films
3.7.1 Optical characterization of thin films
Once photons are incident on a film, they interact with electrons of the films. A photon
can be reflected, transmitted, scattered, undergo luminescence or absorption. From these
optical effects, we obtain optical constants. A graph of absorption spectrum as a function
of photon energy, a number of optical properties can be studied. At high energies photons
are absorbed by the transitions of electrons from filled valence band states to empty
conduction band states. Radiations are absorbed due to the formation of excitons, and
electron transitions between band and impurity states for energies just below the lowest
forbidden energy gap, (Rndar and Ali, 2007). The transitions of free carriers within
energy bands produce an absorption continuum which increases with decreasing photon
energy. The crystalline lattices absorb radiation, with the energy being given off as
optical ~nd acoustic phonons. Finally, at low energies, electronic transitions are observed
between impurities and bands associated to them. These processes have important
optoelectronics applications (Davis and Mott, 1970).
3.7.1.1 Optical properties of amorphous and crystalline materials
Mott and Davis (1970) have summarized the general theory of the experimental results
on amorphous semiconductors. These results have shown two important features namely:
(i) A band structure related to these materials.
(ii) The replacement of the sharp cutoff most attributed to crystalline structures
39
with tailor defect states in the forbidden energy gap.
They show energy band structure, and also that the normally sharp cut off in the density
of states curves at the band edges is replaced by a tailing into the normally forbidden
energy gap. Hence, a difference in the absorption spectra is expected, particularly at the
fundamental absorption edge, between samples of crystalline and amorphous though
from same basic material. From this features, mobility or pseudo gap is defined. It is
larger for amorphous materials than for crystalline materials having the same chemical
compositions. The equivalent band gaps from the optical properties will depend on the
form of excitation process taking place in the material when photons are absorbed. This
means various possibilities will arise, depending on whether the transitions involved are
direct or indirect. The theory for such transitions has been discussed by Davis and Mott
(Davis and Mott, 1970). They consider the localized electronic states in the mobility or
pseudo-gap. In amorphous materials the conservation of K-space rule breaks down and
hence K is not a good quantum number. If we take the matrix element for optical
transitions to have the same value regardless of whether or not the initial and final states
are localized, and also the densities of states at the band edges are linear functions of the
energy, then we can derive a ( optical absorption coefficient ) at a given angular
frequency w (Davis and Mott, 1970):
(3.15)
40
Where A =(4n/c)CJ)n/'J.E, 0"0 is the electrical conductivity at absolute zero, !::.E the
width of the tail of localized states in the normally forbidden band gap, no the refractive
index and E opt the optical energy gap.
The optical energy band gap can be determined from the extrapolation to (ancv)~ = 0 of
a spectral graph of (ancv)~ vs. (ncv). This describes optical absorption associated with
forbidden indirect transitions.
3.7.1.2 Direct and indirect optical transitions
The energy gap in a semiconductor is responsible for the fundamental optical absorption
edge. The fundamental absorption process is one in which a photon is absorbed and an
electron is excited from an occupied valence band state to an unoccupied conduction
band state. If the photon energy (hro) is less than the band gap energy, then absorption is
impeded and in such cases the material is said to be transparent to incoming
electromagnetic radiation for which (hro < Egap) (Rndar and Ali, 2007).
For (luo > E gap) on the other hand, such inter band absorption processes are possible. In
high quality semiconductor crystals at low temperatures, the density of states, geE) rises
sharply at the band edge and consequently the absorption. rises very rapidly when the
photon energy reaches the gap energy. Observation of the optical absorption edge is the
most common means of measuring the energy gap in semiconductors.
41
Now considering an inter band electronic transition, we see that such transitions must be
essentially vertical on the band diagram. This is required if the process is to conserve
momentum.
hk photon = hl'1k electro II . (3.16)
If the maximum of the valence band (VB) and the minimum of the conduction band (CB)
occur at the same k-value (often k = 0) then this condition is satisfied. The gap is said to
be direct as shown in figure 3.9 (a). Such semiconductors are suitable for optoelectronics
applications. The gap is said to be indirect band gap as in figure 3.9 (b) when the VB
maximum and CB minimum do not occur at same k-value, in indirect transition energy
and momentum cannot be conserved without the participation of another particle, usually
a phonon. Because of the participation of a third particle, their transition probabilities are
much lower than those of direct processes.
{a) Du-ect baud gal':
C .B. minimum at k = 0
(b) In.dh-cct buud gllp
C.R minimum at k ;= ()
E conduction baud E
valence band
~~,.
1II:"I ~~
___ 1.
n o
Figure 3.9: Schematic diagram showing direct and indirect band gap transitions (Rndar
and Ali, 2007)
The most important approach to determining the band structures of semiconductors is by
optical spectroscopy. This is based on the fact that Photo induced transitions can occur
42
between different bands, this leads to the determination of the energy band gap. Optical
measurements can also be used to study lattice vibrations. The reflection coefficient R
and the transmission coefficient T are the two important quantities generally measured.
Given normal incidence transmittance is given as (Ohring, 1992);
T = (1- R2 )exp(-4Jlx/ A)
1 - R 2 exp(-8Jlx/ A) (3.17)
T = (1- n)2 +k2
(1 +11)2 +k2 (3.18)
Where, A is the wave length, 11 the refractive index, k the absorption constant, and x the
thickness of the sample. The equation of absorption coefficient per unit length a is given
by (Ohring 1992);
a 4JrkA (3.19)
Transmittance and reflectance data at normal incidence can be simulated using various
models to derive both 11 and k. Absorption coefficient near the absorption edge the can be
expressed as:
(3.20)
Where (11m) is the photon energy, (Eg) is the band gap, and 'Yis a constant. 'Yequals 1/2
and 3/z for allowed direct transitions and forbidden direct transitions, respectively.
3.7.1.3 Absorption edge and Urbach energy
The study of the absorption edge is a useful method for investigating optical! y induced
transitions and the band structure of crystalline and amorphous semiconductors materials.
43
This is based on the absorption of photons with energies greater than the band gap.
Urbach rule is fundamental in study of absorption edge in many disordered materials
given by;
(3.21)
Where a (co) is the absorption coefficient at an angular frequency of to = 2nv, and Eu is
Urbach energy. In logarithms form then,
(
Ii OJ )Ina(OJ)=C+ Eu (3.22)
Eu = [~(~)]-I
d In a (3.23)
By plotting a graph of the natural logarithm of the absorption coefficient as a function of
energy, the Urbach energy can be determined indirectly. This is done by evaluating the
reciprocal of the slope from this graph.
3.7.1.4 Optical band gap and physical models
OJL (O'Leary et al., 1997) model has been employedto describe inter band transition in
amorphous materials; while Drude model has been used to account for the movement of
free electrons. OJL inter band model give expressions for the density of states for the
44
optical transition from the valence band to conduction band. Parabolic bands are assumed
with tail states exponentially decaying into the band gap.
Density of states NCE)
Valence band
+exp «E-Fc) lYe)
-exp( -(E-Ev)lYv)
\ EnergyE
Figure 3.10: Parabolic bands with tail states exponentially decaying into the band gap
(Theiss, 2000).
The original parameters of OJL density of states model are energy, E, and Ee, and the
'damping constants' of the valence and conduction bands, "[v and 1e respectively. Also
included are the effective masses of the electrons and holes bands, m, and m;
The expressions
1
EM,Y =Ey--Yv2 (3.24)
and
1EMC =Ec +-yc, 2 (3.25)
45
denote the mobility edges of the valence and conduction bands respectively. The mobility
gap, Eo, in the OJL is therefore given by:
E +.!.r - [E -.!. r ] = Ec 2c v 2v 0 (3.26)
The band gap energy, Eg, is the difference between Ec and Ev values, i.e. the band gap in
the case of no disorder, which is realized when both E, and Ec, are zero. This gap energy
is different from the mobility band gap which depends on the values of the disorder
parameters E, and Ec. The optical band gap energy is simulated by SCOUT 98 software
(Theiss, 2000).
The energy gap (or band gap) Eg and its structure as a function of the wave vector are key
characteristics of the semiconductor material and of fundamental importance to the
operation of the solar cell (see figure 3.5). The principal features of interest are the
temperature variation of the band gap energy Eg and the magnitude of wave vector
associated with low-energy transitions (Markvart and Castener, 2005).
To free an electron, the energy of a photon must be greater than the band gap energy.
However, photons with more energy than the band gap energy will expend that extra
amount as heat when freeing electrons. So, it's important for a solar cell to be "tuned" -
through slight modifications to the silicon's crystal structure - to optimize the photon
energy. A key to obtaining an efficient solar cell is to convert as much sunlight as
possible into electricity. Crystalline silicon has indirect band gap energy of 1.1 electron-
volts (eV) (Markvart, 1998). The band gap energies of other effective photo voltaic
46
semiconductors range from 1.0 to 1.6 eV. In this range, electrons can be freed without
creating extra heat. The photon energy of light varies according to the different
wavelengths of the light. The entire spectrum of sunlight, from infrared to ultraviolet,
covers a range of about 0.5 eV to about 2.9 eY. For example, red light has energy of
about 1.7 eV (Markvart, 1998, Wooten, 1972), and blue light has energy of about 2.7 eV.
Most solar cells CaJ1l10tuse about 55% of the energy of sunlight, because this energy is
either below the bandgap of the material or carries excess energy (Wooten, 1972).
3.7.1.5 Structural properties of thin films
Texture can be determined by various methods. Some of them allow a qualitative analysis
of the texture; others are only quantitative. Among the quantitative techniques, the most
widely used is X-ray diffraction using triple axis goniometers, followed by EBSD-
method (electron backscatter diffraction) in Scanning Electron Microscopes (SEM).
Qualitative analysis can be done by Laue photography, simple X-ray diffraction or with
the polarized microscope. Neutron and synchrotron high-energy X-ray diffraction allow
access to textures of bulk material and in-situ analysis, whereas laboratory x-ray
diffraction instruments are more appropriate for thin film textures. X-ray scattering
techniques are a family of non-destructive analytical techniques which reveal information
about the crystallographic structure, chemical composition, and physical properties of
materials and thin films. These techniques are based on observing the scattered intensity
of an X-ray beam hitting a sample as a function of incident and scattered angle,
polarization, and wavelength or energy. X-ray diffraction yields the atomic structure of
47
materials and is based on the elastic scattering of X-rays from the electron clouds of the
individual atoms in the system.
X-Ray fluorescence (XRF) IS a multi-elemental analysis technique that IS non-
destructive, multi-elemental, fast and cost effective. High energy photons (X-rays)
displace inner shell electrons. Outer shell electrons then fall into the vacancy left by the
displaced electron: In doing so, they emit light (f1uoresce) equivalent to the energy
difference between the two states. Since each element has electrons with more or less
unique energy levels, the wavelength of light emitted is characteristic of the element and
the intensity of light emitted is proportional to the elements concentration. There are two
types of XRF spectrometers: wavelength dispersive and energy dispersive. Wavelength
dispersive system uses a diffraction crystal to focus specific wavelengths onto a detector.
Energy dispersive spectrometer focuses all the emitted x-rays onto an energy analyzing
detector. While this is faster and less expensive, wavelength dispersive spectrometers are
more sensitive and have higher resolution. For this reason, a wavelength dispersive
system was used.
48
CHAPTER 4
EXPE~ENTALPROCEDURE
4.1 Introduction
This chapter presents a detailed description of the materials used in the research study,
their purity and how they are processed to fabricate a p-n junction solar cell. The
parameters for deposition and experimental procedures for fabrication of the solar cell are
discussed.
4.2 Thin film preparation
In this research study, p-type Cu.O, and n-type ZnO:AI films were separately deposited
using reactive de magnetron sputtering and reactive thermal evaporation techniques,
respectively. Multiple deposition was employed in fabricating Cu.O, - ZnO:Al p-n
junction. Silver was used as a rear contact electrode. An Edwards auto 306 coater was
used for the thin film deposition.
4.2.1 Deposition of Copper Oxide (CuxOy) thin films
The glass substrates were cleaned in dilute chromic acid to remove metallic particles and
organic stains or surface contaminants and thereafter rinsed thoroughly with dislilled
water and ethanol to remove any acidic residues and allowed to dry completely. The
substrates were then introduced into the vacuum chamber. Thin films of Copper Oxide
49
were deposited on microscope glass substrates (25 mm x 25 mm x 1 mm). The distance
between the source and substrate was 15 ern. The starling materials were a solid copper
target (99.99 % purity) and two gases namely, Oxygen (reactive gas) and Argon
(sputtering gas). The deposition chamber was evacuated to a base pressure of 5 x 10-6
mbar. Argon gas was then introduced into the sputter chamber at fixed flow rate of 20
Seem and the target was pre-sputtered in pure argon atmosphere for 15 min to remove
oxide layers and other contaminants on the surface of the target. After pre-sputtering and
turning-on' of dc power, Oxygen gas was admitted to reach the required sputtering
pressure and the de power supply was turned on again to start the thin-film deposition.
The flow of Oxygen into the deposition chamber was monitored using mass now
controllers and gas regulators interfaced to a computer. The pressure during deposition
was maintained at 5.5 x 10-3 mbar. Sputtering was done [or 15 minutes. Pure Argon
(99.99% purity) was used as the sputter gas and pure Oxygen (99.999% purity) as the
reactive gas. Copper Oxide thin films were prepared under different Oxygen now rates
(2.5 ... 25 Seem) and the other deposition parameters such as power (200W), substrate
temperature (325 K) and sputtering pressure were held constant. The sputtering
conditions maintained during the growth of Copper Oxide thin films were set after some
preliminary experimental studies. Quartz crystal monitor mounted in the sputtering unit
was used to measure the film thickness. This process was repeated for all samples in this
research study
50
4.2.2 Deposition of Aluminum doped zinc oxide (ZnO:Al)
Zinc (99.9 % purity) and Aluminum (99.99 % purity) were mixed at varying doping
percentages of Aluminum (0 - 6 %) and then heated in closed gJass tube until they melted
to form a compound. Glass substrates were cleaned to remove stains on them by boiling
in dilute chromic acid to remove surface contaminants and rinsing thoroughly with
distilled water and ethanol and allowed to dry completely. The substrate was then
mounted on a rotating substrate holder and the compound was then placed in a
Molybdenum boat. The chamber was covered tightly and pumped down to 5.0 x 10-6
mbars. A current of 4.0 A was supplied to the heater to evaporate the materials at a
temperature of about 800 K. The shutter was removed to permit deposition on glass
substrate in the presence of oxygen which was let into the chamber. Since Zinc is more
reactive than Aluminum, ZnO:Al thin films were formed. The Bond enthalpy of Zn-O is
159±4 kJ/mol while that of AI-O is 5l1±3kJ/moi (Atanas et al. 2010).
4.3 Copper Oxide and Aluminum doped ZnO thin films characterization
The details of the techniques used for characterization of Copper Oxide and ZnO:Al thin
films are described from section 4.3.1 to 4.3.5.
4.3.1 Thin film thickness measurements
Thin film thickness was estimated using Tencor alpha step surface profilometry
(resolution of 5 ) equipment with a diamond stylus of radius 12.5 )..UTI. During
51
measurement, the stylus was moved across the film surface while keeping the sample and
the sample stage stationary. The step created during the deposition process enabled the
film's thickness to be read directly as the step height. SCOUT 98 software was also used
to simulate the film thickness. This was used to validate the measurements obtained by
Tencor alpha step surface profilometry equipment with comparisons to thickness with
Quartz crystal monitor.
4.3.2 Optical measurements
Optical measurements (reflectance and transmittance) in the spectral range from 300 nm -
2500 nm were carried out using UV/VIS/NIR 3700 double beam Shimanzu
spectrophotometer. Photons of selected wavelengths and beam intensity 10 (photons/emf-
s) were directed at the film of thickness (t) and their relative transmissions observed.
Wavelengths of photon are selected by the spectrophotometer. Photons with energies
greater than band gap (Eg) are absorbed while those with energies less than Eg are
transmitted. The spectrophotometer had two radiation sources; a deuterium lamp for UV
range and a halogen lamp for visible (VIS) and near infrared (NIR) range. The radiation
source changed automatically to access the wavelength range during measurements.
During transmission measurements, samples were placed in front of the integration
sphere and behind it during reflection measurements SCOUT 98 software was used to
simulate transmittance data to get the optical constants like absorption coefficient among
others. Drude, OlL, Tauch Lourntz, Extended Drude and Harmonic Oscillator models
were used to simulate the data. These models are inbuilt in the SCOUT 98 software
52
(Theiss, 2000). The models simulate refractive index, dielectric function, absorption
coefficient real and imaginary parts and energy loss parameters.
4.4.3 Sheet resistivity measurements
•
•....
Figure 4.1: Schematic diagram of a four point probe used to measure surface sheet
Resistivity (Brown and Jakeman, 1996)
The four point probe technique (figure 4.2) was used to measure the sheet resistivity of
the Cuprous Oxide and Aluminum doped ZnO semiconductor thin film samples. With a
symmmetrical square geometry adopted, the four leads from the probe head were
connected to Keithley SourceMeter via relay switching circuit as per the Van del' Pauw
set-up for Voltage and Current measurements (Agumba, 2010).
4.4 Fabrication of CuxOyl ZnO:AI heterojunction
The Cu.O, solar cell with glass/ Cu.O, /ZnO:AlI Ag was fabricated as in figure 4.3 in
stages. First, a p-type layer of transparent conducting Cu.O, film with thickness 350nm
53
was deposited onto a glass substrate at 328 K by sputtering method at pressure of 5.5 x l O'
3 mbar. Using aluminum foil to mask section of the p-type film, ZnO:Al film of thickness
200 run was deposited onto Cu.O, on the same substrate at a temperature of 780 K. An n-
type layer of AZO was formed. The temperature was measured using thermocouple
attached to the backside of molybdenum boat. Fabrication was completed by pasting 0.3
em' silver contact on AZO film. The film obtained was CuxOylZnO:AI which forms a
heterojunction. The junction was fabricated using optimized deposition parameters based
on optical and electrical characterization of various samples of AZO and CuxOy.
Glass Substrate
ce,o,
ZnO:Al
Silver contact -e +
Figure 4.2: Schematic diagram for Cu.O, - ZnO:AI p-n junction solar cell.
4.5 Solar cell characterization
A solar cell simulator was used to characterize the solar cell. In the I-V measurements, a
constant beam of monochromatic light was incident on the cell. The active area of the cell
was 3 ern". A Keithly Meter recorded the current as the voltage across the cell was varied.
54
I-V measurements were done using a microwave powered light drive 1000 W lamp solar
simulator while data acquisition was computerized. The simulator was calibrated using a
standard solar cell before any measurements were done. The lamp was standardized to
one sun. The results of the experiments carried out to optimize the AZO and Copper
Oxide thin films samples and the fabricated p-n junction solar cell are presented in the
next chapter.
55
CHAPTERS
RESULTS AND DISCUSSION
5.1 Introduction
This chapter presents the results and a detailed discussion of the results of the study.
Firstly, optical, electrical and structural properties of AZO are discussed Then followed
by optical, electrical and structural properties of Copper Oxide. Optical and electrical
properties are used as the basis for optimizing AZO and Cu.O, thin films for fabrication
of the Cu.Oy-Znfr.Al p-n junction solar cell. Then the structural properties of the
optimized AZO and CUxOy are discussed. Lastly the I-V characteristic of the CUxOy -
ZnO:Al p-n junction solar cell formed is discussed.
5.2 Characterization of ZnO:Al
In this res~arch study, optical constants from near normal reflectance and transmittance
data for ZnO:Al are studied. SCOUT 98 software (Theiss, 2000) was used to simulate the
transmittance data to generate the corresponding optical constants. Drude and OJL
models are essential in simulations of transmittance data.
5.2.1 Optical characterization of ZnO:Al thin films
The optical transmittance spectra of ZnO: Al films as a function of wavelength in the
range (300 - 2500 om) were plotted in figure 5.1. The high transmission (2: 70 %) is
56
understood because ZnO is a semiconductor with wide direct band gap of 3.3 eV
(Rodrigo et al., 2003, Selmi et al., 2008). Due to the high transmission, these films have
good optical properties for solar cells window applications. It was observed that the
transmission over the visible range decreases as the concentration of Aluminum
increases. This is due to free carriers coupling to the electric field hence increasing the
reflection.This agrees very well with Elmin and others (Elmin et al., 2009) who reported
significantly reduced transmission when ZnO was doped with higher percentages of
Aluminum.
100~----'---~----r---~---.----~---.----~---.---.
80
,-,
~ 60 .•'-'
C1J
U§+-':§ 40 ~'" •cro .•>-<t--<
20
•0
500
o
2
3
4
• 5.• 6
1000 2000 25001500
Wavelength (nm)
Figure 5.21: Spectral graph of transmittance (%) against wavelength (nm)
various doping levels of AZO.
57
From the transmittance graph on figure 5.1 the absorption edge shifted toward higher
wavelength as the doping levels increased. All films exhibited sharp fundamental
absorption edge. Ref1ectance data (figure 5.2) show that the average reflectance is below
45% within the visible range. All films exhibit low absorption in the visible spectrum
range. The film had thickness range between 95 nm to 130 nm.
50
45
40
~ 35Q)
0I=i 30cO+-'0
Q)4=1 25
Q)
~ 20
15
10
5
Al doping levels (%)
3
• 2
.••.4
.••• 5•
•••••••• •• ••• •• •••• ••• •• •• •••
. ~.- -... .. ..•.. "". ~•• - ". ••••A ..•..••• - ..•...•. "lto,._y..- ..•.. .... ~ .-.r:••- ..•..•. .•. .;....... ••••••••••• .•..•.•.. ..•...•.. ......G.....•. .•. ..•. .•. ..................... r..•....•.. ..•.. ..•....•....•....•.. ..•.. ..•....•....•.. ..•.. '"..•.. ..•....•....•.. "' ..•....•..~ ..•....•....•..
400 1200600 800
Wavelength 11m
1000
Figure 5.22: Spectral graph of reflectance % in the visible, IR, and UV regions against
wavelength (nm) for different doping levels of AZO.
5.2.1.1 Simulated and Experimental graphs for AZO transmittance data
The experimental and simulated spectral for transmittance data were plotted against
wavelength for the different samples. Using the SCOUT 98 software, the simulated
curves fitted perfectly in to the experimental curves as shown in figure 5.3 (a) - (e) below.
From the graphs, the optical band gap energies and thickness of the films were simulated.
58
2% Doped
100
80 r~~ • SIMULATED~ 60 0
<1) 0o 0 0 EXPERIMENTAL~ 0ro 0<-'<-' 0·8 40
Ul 0~ 0ro•... 0f- 0
20 0
0
0
0f}a
200 400 600 800 1000 1200
Wavelength (nm)
Figure 5.3(a): ZnO:Al (2 % Al doping concentration)
100,-----~---.----~---.----~--_,r_---.----._--~--__,
3 % Doped
60
0
40 0
0
20
0
Q
0
0
200 400
80
r~mrwJXIJJ!Jlf!l!l!!JJJiJJrrriJJlJ.lJiJJmiJ!JJJrililiJ!JlIJlIJ!lJJJ.rI!J!JNill((((W((((rrrramrrarra(rr=
o
o
o
• Simulated
o Experimental
600 800 12001000
Wavelength (nm)
Figure 5.3(b): ZnO:Al (3 % Al doping concentration)
90
80
70
-'"""' 60
~
OJ 50u
hcr:::: 40·s
V)cro 30•...f-<
20
10
0
200
59
4 % Doped
0
0
0
0
0
0
0
400
• Simulated
o Experimental
1200
5 % Doped
100
~1iNl!m:m.~1KJ
80 r
0
0 Simulatedr>: •~ 60 0 0 Experimental
OJu§ 0::::·s 40 0'"1ii
~ 0
20
0
0
0
0
200 400 600 800 1000 1200
600 800 1000
Wavelength (nm)
Figure 5.3(c): ZnO:AI (4 % Al doping concentration)
Wavelength (nrn )
Figure 5.3(d): ZnO:AI (5% Al doping concentration)
60
G % Doped
100
80
~
~ 60
OJo0:fl§ 40Vl0:'"r::
20
0
200
o Experimental
o Simulated
~rmrmmmmi'!'(""['f'f'Ciiililiilll:mr~~
o
a
o
o
o
a
o
o
o
400 600 800
W avelcngth (n m )
1000 1200
Figure 5.3(e): ZnO:AI (6% AI doping concentration)
Figure 5.23 (a-e): Transmittance spectra for ZnO doped with Al at various Al doping
levels
5.2.1.2 Optical band gap and urbach energy for AZO thin films
Optical band gap for various samples of Aluminum doped thin films as simulated from
SCOUT 98 software are shown in table 5.1. The reduction in transmittance also led to
variation in band gap. This may be due to the Oxygen vacancies and the behaviour of free
carrier's concentration with increasing the ZnO doping. This compares very well with
studies conducted by Wang (Wang, 2004) and Shadia (Shadia. et al 2009) who both got
values of band gap ranging between 3.2 eV and 3.5 eV ..Between 0 % to 3 % the optical
band gap reduces. This is followed with widening of the band gap for doping between 4
% to 6 %.
61
Table 5.1: Values ofthe optical band gap at different doping levels for AZO
Doping % of Optical Thickness Urbach energy ± 0.01
ZnO With Al bandgap ±0.2 (nm) ± 5nm (xl0-4) eV
eV
0 3.34 113 2.02
2 3.28 115 2.04
3 3.18 112 2.07
4 3.21 108 2.09 -r-
5 3.32 101 2.12
6 3.42 98 2.18
From table 5.1, Urbach energy gradually increased with increasing band gap. This is
consistent with the variation of the conductivity of the film with doping concentration.
When band gap was reducing due to formation of localized states near the conduction
band, it corresponded to increase in Urbach energy. Increased impurity formed more
localized states within the band gap despite that beyond the 3 % doping there was
increased scattering. Increased doping at 4 % to 6 % did not enhance conductivity
despite Urbach energy increasing.
62
3.45
3.40
,-.,
3.35> .~OJ'-'P..,robfJ
"0 3.30cro..0
~
0 3.25'';::::;P..,
0
3.20
3.15
0 2 3 4 5 6
Aluminium doping levels %
Figure 5.24: Variation of optical band gap for ZnO:AI with various doping levels of
Aluminum (%).
From figure 5.4, there was a general drop in the band gap up to a doping level of 3 % .
Decrease in optical band gap energy can be attributed to creation of new donor levels in
the forbidden zone; and a shift in the fermi level causing a change in the band structure of
the films. At room temperature aluminum atoms occupy the zinc sites in the ZnO lattice.
They are singly ionized donors giving one extra electron. There after the band gap starts
to widen, this can be explained by Burstein Moss effect (Yaodong Liu et al., 2009).
63
Increased concentration of donor atoms causes more electrons to occupy states at the
bottom of the conduction band causing it to be filled up with donor electrons which
results in widening of band gap. It is clear that the variation of a (cm') versus photon
energy (eV), figure 5.5 which is near straight line, indicate the presence of direct optical
transitions. Excess doping also destroys the structure stoichometry hence reducing the
conductivity of the ZnO films (Ogwu et al., 2007).
4.00E+012
Al doping levels (%)
3.50E+0 12 - 2
3.00E+012 • 3
rr-; ..•. 4
S 2.50E1012 .•. 5~ • 6+-'c:: 2.00E+O 12<U ..•.u::tH
<U •0 1.50E+0 12oc ..•. ••.2+-' 1.00E+012e- ..•. -..0
t/J •••• • •.D 5.00E+0 II
•••• • •<C ..•.••*....,t
O.OOE+OOO ...••., ..................
-5.00E+0 I I
2.5 3.0 3.5 4.0
Energy (eV)
Figure 5.25: Graph of variation of absorption coefficient for various Al doping
levels against photon energy.
The Urbach energy which is interpreted as the width of tails of localized states in the gap
region were calculated from the following relationship, Eu= (dlna/dfzvyi.
It is obvious when (lno.) is plotted against (nw), the inverse of slope will be the value of
64
tail localized state (Rnjdar and Ali, 2007). The values of Urbach energy Eu for all
composition are tabulated in table 5.1. This is extracted from spectral graph in figure 5.6.
12 A ~..•. I· •A ~. · •. ..: •..
-
10 · · I...~ ... .. •8 ... •.......... . •(j
E 6 f-
a •..... aa •AA4 f- .A"" •.•..•..•..•..•..•. .It.: • • •... •
2 l- ..... •~ ... .... •
-
•
•AI doping levels
2
3
4
5
6
o ~~~~--~--~--~--~--~--~--~--~--~--~
1.5 2.0 2.5 3.0
Energy (eV)
3.5 4.0 4.5
Figure 5.26: Graph ofIna against energy for various Al doping levels.
It is observed that, the Urbach energy increased gradually with increasing the Aluminum
doping levels (table 5.2). This may be due to the fact that, ZnO doping causes a shift in
the optical absorption edge therefore change in the band structure of the films. Variation
in the optical absorption edge of the films with increasing the ZnO ratio indicate that, the
dopant ratio is responsible for the width of localized states in the optical band of the films
and causes an increase in the energy width of localized states thereby affecting the optical
energy gap.
65
Table 5.2: Values of Urbarch energy at different doping levels ofZnO with N.
Doping % of Urbach energy ± 0.01
ZnO With Al (xlO-4) eV
0 2.02
2 2.04
3 2.07
4 2.09
5 2.12
6 2.18
The band tailing (Urbach energy) is the effect of impurity or disorder and any other
defects. In the exponential-edge region (figure 5.5), the absorption coefficient is
expressed by the Urbach relationship. The Urbach's absorption edge is formed in the
region of photon energies below the forbidden band gap. The interaction between lattice
vibrations and localized states in the tail of the band gap of the compound has a
significant effect on the optical properties of the thin film. The increase in aluminum
doping creates crystal strain in ZnO crystal since AI-O has a higher bond enthalpy of
511Kj/mol compared to 159kJ/moi of Zn-O. This increases structural defects and thereby
increasing Urbach energy
66
5.2.2 Electrical characterization of ZnO:AI thin films
Using a four point probe (Agumba , 2010) the following results for sheet resistivity of
AZO thin films were realized as in table 5.3.
Table 5.3: A summary of electrical surface sheet resistivity for AZO thin films.
Aluminium doped Zinc Oxide
Doping % Sheet resistivity
±2 Q-cm
0 48.23
2 44.26
3 36.24
4 44.71
5 51.72
6 55.21
Surface sheet resistivity at room temperature was as in table 5.3. The electrical properties
greatly depend on deposition parameters and film thickness. During the deposition, the
substrate temperature was kept at 780 K. Figure 5.7 shows the variation of doping
concentration with sheet resistivity. The mobilities of the ZnO:Al reduce with the doping
concentration, Such behaviour was expected as a result of subs'titutional doping of Al3+ at
the Zn2+ site creating one extra free carrier in the process. As the doping level is
increased, more dopant atoms occupy lattice sites of Zinc atoms resulting in more charge
67
carriers. This Jeads to a higher polarization of the electron by the addition of Al and thus
increasing the electron phonon coupling. Strong scattering implies short carrier lifetimes
and thus lower mobility.
55
50
v
~ 40
if)
35
o 2 3 4 5
Al doping concentration (%)
Figure 5.27: Graph of Aluminum doping levels with sheet resistivity.
6
Lower electron mobility cause reduction in conductivity hence high sheet resistivity.
Thus, the resistivity increases with the doping concentration. However, after a certain
level of doping, the dopant atoms in the crystal grain and grain boundaries tend to
saturation. In this case, high doping concentration will lead to a large quantity of ionized
impurity. This ionized impurity provides strong scattering centre's for charge carriers.
According to the Conwell- Weisskoft theory (Ma Jin et al., 1999), when degenerate
charge carriers are scattered by impurity ions, the energy dependence of ionized impurity
68
scattering mobility due to a high doping concentration will lead to a larger quantity of
ionized impurity, resulting in a decrease in the mobilities of the ZnO:Al films (Jin Ma et
al., 2000). This result compares well with other studies (Zhao et al., 2002; Jinsu Yoo et al
2004) who reported sheet resistivity of 40 - 60 n- ern.
5.2.3 Composition and structural characterization of ZnO:AI thin films
5.2.3.1 X-ray diffraction of ZnO:AI thin film (as-deposited)
XRD measurements show structural make-up and size of crystalline structures for AZO.
The X-ray diffraction (XRD) spectra of all the AZO films show the presence of a sharp
peak indicating that the films are highly oriented as shown in figure 5.8. The XRD
spectra of the films reveal that they have crystalline structure with the main diffraction
peak located at angle 28 = 36°. Neither metallic Zinc or Aluminum peaks nor Zinc Oxide
peaks were observed from the XRD patterns. This implies that Aluminum atoms replace
zinc in the hexagonal crystal lattice. Peak intensity of the XRD diffraction reflections is
determined by the crystalline grain, size and structure and axis orientation. These results
are in agreement with those of Mujdat, (Mujdat et al., 2008). XRD analysis reveals that
the film exhibit only the (002) peak, indicating that they have preferred orientation,
implying a c-axis growth perpendicular to the substrate surface. The dominant (002) peak
becomes sharper, indicating the well-established c-axis orientation of ZnO:Al thin films.
This shows the crystallinity of the AZO thin film (Ogwu et al., 2005).
69
zinc sample_1
C 002
0 Zinc20000
U
N
10000 -T
S
0 I I I I I I
10 20 30 40 50 60
Position [°28] (Zn)
Figure 5.28: XRD spectra of optimized AZO thin film
5.2.3.2 Elemental composition of ZnO:AI tbin film (as-deposited)
Table 5.4 shows the elemental percentage concentration of the thin films as obtained by
the X-ray florescence (XRF) MiniPal 2 machine. The analysis was carried out to
ascertain the elemental composition of the thin films. The peak-based analysis technique
is used where elemental intensities of thin films are calculated and respective spectral
background obtained. Figure 5.9 shows the peak based analysis of the elemental
composition of thin film samples.
70
Table 5.4: XRF elemental percentage composition of optimized AI doped ZnO thin
films.
CONWOUNDffiLEMrnNTAL PERCENTAGE LEVELS
±0.05%
Al 1.23
Si 59.0
Zn 39.77
Si02 53
ZnO 46.8
Ah02 0.2
I
IZn KA,
A "'-
ZnlKB
I I I I I I I I6 B 18 12 14 16 1B 28
keU
I
I Si KA
I Al,~Aj .~
J2nm'\.
I2 I4
Figure 5.29: XRF spectra of as-deposited optimized AZO thin film
71
5.3 Characterization of Cu.O, thin films
5.3.1 Optical characterization of Cu.O, thin films
The optical transmittance of Copper Oxide films prepared at 200 W, de power and 5,
7.5 ... 22.5 Seem Oxygen now rate is shown in figure 5.10. The optical transmittances of
the films show a strong dependence on the flow rate during deposition; the optical
transmittance decreased with an increase in Oxygen flow rate during deposition. This
behaviour was observed for all the films deposited. Figure 5.10 summarizes the inter-
relationship between the Oxygen flow rate and the optical transmission. There is
observable shift in absorption edge (figure 5.10).
90
80
70
60
,-.s. 50
'"o<:: 40§
E 30'"<ll:!f- 20
10
0
-10
200 400 600 800 1000 1200 1400 1600 1800 2000 2200
W a v e l e n g tl: (n m )
Figure 5.30: Variation of transmittance (%) with Oxygen now rate
72
The reflectance spectrum for Copper Oxide films figure 5.11 shows low reflectance
values in the visible spectrum. This when compared to the above transmittance indicates
that it has a high absorption coefficient making it a good absorber material for solar cell
applications. This agrees very well with Ogwu (Ogwu et al., 2005) who observed low
values of reflectance (%) bellow 45 % for Copper Oxide thin films.
70
65
60
55
50
45;i~ 40
Q)o 35coj
+-' 30o
Q)
!:H 25Q)~
20
15
10
5
0
0
-0-5.
-0-10
-"-15
-"'-20
-"'-22.5
Oxygen flowrate in Seem
500 1000 1500
WaveJength (nm)
Figure 5.31: Spectral graphs showing variations of reflectance (%) with Oxygen now
rate in Sccm.
5.3.1.1 Simulated and experimental graphs for Copper Oxide transmittance data
The experimental and simulated data for transmittance data were plotted against
wavelength for the different gas flow rates. Figure 5.12 (a., d) are samples of the fitted
data using SCOUT 98 software.
80
70
60~2?-
'"--'
(!) 50
(,)s::
(\jt:: 40Os
VJs::
(\j
30H[-<
20
10
200
73
• SIMULATED
EXPERIMENT AL
60
50 -
40 -
,-,~0~
<l) 30 -u
~§
'" 20 -§
~
10 -
o -
o
•r-.
400 600 800 .1000 1200
Wavelength (nm)
Figure 5012(a): Oxygen flow rate of 5 Seem
o
o
o
o
o • Experimental
o Sim ul a te d
-
-
-
o
o
o
o
o
o
00c
300 400 500 600 700 800
Wavelength (11m)
Figure 5.12(b): Oxygen flow rate 01'7.5 Seem
900
60
50 -
40 -~
~<:»
Q)o 30 -~ro-+-'-+-'.,...,
S 20 -r.f)~roHt--<
10-
0
74
o
o
o
o
o
o
o -.- EXPERIMENTAL
-0-- SIMULATED
o
o
o
oo
00
300 600 800 900700400 500
80 - o~o
• 0Q LlLl
0
~ 00
60 -
0000~ ClLlClOooooQa
~ ~"-"
Q)
0
C 40 - 0ro • SIMULATED.~
S \;/ a EXPERIMENT ALtr:
~ 0H 20 ~E--< ~a°iii0°·0 00(0);/ ••
200 400 600 800 1000 1200
Wavelength (nm)
Figure 5.l2(e): Oxygen flow rate of 12.5 Seem
Wavelength (nm)
Figure 5.12(b): Oxygen flow rate of7.5 Seem
75
80
60 -
o
• Simulated
o ExperimentalE</)t::ro•...
f--< 20-
o
200 400 600 800 1000
Wavelength (nm)
Figure 5.12(d): Oxygen flow rate of 15 Seem
1200
Figure 5.32 (a-d): Simulated and experimental graphs of sampled thin films of Copper
Oxide at various Oxygen flow rates (Seem).
5.3.1.2 Optical band gap and Urbach energy for Copper Oxide thin films
From figure 5.13 it was observed that the variation ofa (cm') versus energy (eV) which
is near straight line indicates the presence of direct optical transition. The absorption edge
tail width is long showing levels of defects called Urbach energy. The Copper Oxide
formed had relatively low transmittance with low values of optical band gap.
1.2
1.0
---
E~ 0.8
"'0
xc: 0.6OJ·0
!B
OJuc 0.40'R
0
Vl.D
oj 0.2
0.0
76
Oxygen flow rate levels (Seem)
7.50
10.0
12.5
15.0
• 17.5
20.0 ..· . .· .·
•
1.0 1.5 4.52.0 2.5 3.0 3.5 4.0
Energy (eV)
Figure 5.33: Variation of a (cm') versus energy (eV) for different oxygen gas flow rates
The optical band gap of the thin film as simulated by SCOUT 98 software shows a
decrease in band gap as the Oxygen flow rate increased up to at 10 Sccm (table 5.5 and
figure 5.14) then increases again. This can be attributed to recombination and variation
in Oxygen doping levels. This also corresponds to the variation of Urbach energy with
flow rate as indicated in figure 5.15 and table 5.5
77
Table 5.5: Optical band gap for Copper Oxide at different oxygen flow rate.
Gas flow rate Optical band Urbach
(Sccm) gap eV ±0.02 energy eV
±0.02 x 10-4
5 2.09 1.82 7
7.5 2.02 0.63
10 1.62 0.61
12.5 1.82 1.07
15 2.11 1.13
17.5 2.32 1.23
20. 2.43 1.43
22.5 2.54 1.92
78
2.6
T/!
2.4 /1> /1Q)0.. 2.2ro
b/)
-0 I_~_! /1§..D 2.0
(;J
.~ \/!p, 1,80
1,6 !
4 6 8 10 12 14 16 18 20 22 24
Oxygen flowrate (Seem)
Figure 5.34: Variation of optical band gap with various levels of Oxygen now rate
(Seem)
The values of the band gaps were 1.62 eV and 2.54 eV. This band gap depends on the
kinetics of formation of the oxides. The kinetics of the formation of oxides of Copper
during thin film deposition is dependent on following factors.
(i) The nucleation rates of Cu, Cu20 and CuO during the
(ii) The sticking coefficient / sticking probability of the particles
(iii) Re-evaporation and migration by the impinging copper and
(iv) The different growth rates of the nucleated species
All of the above factors depend on the power and gas pressure in the deposition chamber
during the sputtering process. At 10 Sccm the band gap is 1.62 eV. At this band gap the
Copper Oxide thin film makes a good absorber material for solar cell application.
13.5
13.0
12.5
12.0
~ 11.5.s
11.0
10.5
10.0
9.5
9.0
79
...• .
•• 'I .• : •... ....;.' . ..- ,.. ......:.......... ........., .....,.,........... .---~- ..+. ..
.•.• .." ·O;ygen flow rate (Seem)
5
7.5
15
20
22.5
.•
•
1.0 1.5 2.0 2.5 3.0
EnergyeV
3.5 4.0 4.5
Figure 5.35: Spectral graph for determination of urbach energy for Copper oxide films
According to Kawaguchi and Muthe (Kawaguchi et al., 1994 and Muthe et al., 1998), the
effective sticking probability is important in deducing the type of oxide formed i.e. Cu-O
or CuO. This depends on the Oxygen to Copper nux ratio, but not on the individual nux.
values during deposition. It is also observed that the kinetics of CuO formation is much
slower than that of Cu20. They further proposed that CuO formation is based on the
reaction
CUzO + 0 -- .••2CuO
Since the extent of dissociation of the starting oxygen molecules during deposition into
atomic Oxygen and Oxygen ions depends on the applied power during deposition, the
ratios of Oxygen ions to Copper atoms and atomic Oxygen to Copper atoms, which
80
determine the phase stability regime for CU20 and CuO formation (Muthe et al., 1998)
depended on the power applied during deposition.
5.3.2 Electrical characterization of Copper Oxide thin films
Using a four point probe (Agumba, 2010) the following measurements for sheet
resistivity of Copper Oxide thin films were as indicated in table 5.6.
Table 5.6: A summary of electrical surface sheet resistivity for Copper Oxide
Copper Oxide (Oxygen now rate)
Oxygen now rate ( Seem) Sheet resistivity
±2.a-em
5 33.62
7.5 36.25
10 40.45
12.5 53.93
15 55.64
17.5 56.34
20 58.22
22.5 61.21
81
High film sheet resistivity at room temperature can be attributed to crystal defects during
deposition. It should be noted that sheet resistivity depends on deposition parameters like
pressure (Agumba, 2010); Bias voltage (Reddy et al., 2001), temperature (Karanja, 2009;
Agumba, 2010) among others. Figure 5.16 shows variation of sheet resistivity with
Oxygen flow rates. The sheet resistivity increased with increasing Oxygen flow rates. We
observed metallic like conduction on films produced at low Oxygen flow rate, which is
attributed to a high copper content of the films produced under this condition. This may
lead to increased Copper vacancies (Ogwu et al., 2007)
65
60
55
SoS 50
C.:>:~45
VJ
OJ•...
0)
OJ 40.ctr:
35
30
4 6 8 10 12 14 16 18 20
Oxygen flow rate (Seem)
Figure 5.36: Variation of sheet resistivity with oxygen flow rates.
5.3.3 Structural characterization of Cu.O, thin films (as-deposited)
22 24
CuO is found to be predominantly amorphous (figure 5.17). The background of the
diffraction pattern indicates the presence of a significant content of amorphous phase of
the Copper Oxide thin films. The structure of these amorphous films, as determined by
82
X-ray diffraction, consists of a random packing of quasi-layers with fourfold coordinated
Copper atoms mixed with Oxygen atoms.
10 30 40 50
Position r,lTheta] (Coppertcu)
60
Figure 5.37: XRD spectrum for optimized Cu.O, thin film
5.3.4 Elemental composition ofCuxOy thin rdms (as-deposited)
Table 5.7: XRF elemental percentage composition of optimized Copper Oxide thin.
CONWOUNDffiLEMrnNTAL PERCENTAGELEVELS±l
Si 88
Cu 12
Si02 95
CuO 4.9
'iT
....;N
QJ
C
C
IUI.£)..c •....•
U<,toQ.
U
CD
83
2 4 6 8 18 12 14 16 18 28
keU
Figure 5.38: XRF spectrum of as-deposited optimized Cu.O, thin film
Tables 5.7 shows the elemental percentage concentration of the thin films as obtained by
the X-Ray Florescence (XRF) MiniPal 2 machine. Figure 5.18 shows the peak. based
analysis of the elemental composition of thin film samples. The yellow coloured
spectrum shows the composition of the elements present in the thin films at ~, Kp, La
and Lp lines. The bottom line represents the background radiation which shows the
radiations detected by the XRF machine but are not present in the thin film samples.
These radiations may have originated from the machine itself, or/and the room in which
the measurements were taken. The highest colourless peak.represents the detector escape
which is a spectrum for detector peak. (Rhodium) element that the machine is made of
(the machine detects itself). From the XRF results, the Copper Oxide phase of CuO was
predominant.
84
5.4 Solar cell fabrication
5.4.1 Optimized conditions for solar cell fabrication
At an Oxygen flow rate of 10 Seem, Cu.O, thin films samples exhibited lowest optical
bandgap of 1.62 eV. This band gap makes it a good absorber material for solar cell
application. The optical band gap of ZnO:AI at 3 % Al doping was found to be 3.18 eY.
The electrical results showed high conductivity at 3 % AI doping. Based on these results,
Cu.O, - ZnO:Al p-n junction solar cell was fabricated using conditions or deposition of
Cu.O, and ZnO:AI at Oxygen flow rate of 10Sccm and Al doping of ZnO at 3 %
respectively. Tables 5.8 and 5.9 show the optimized conditions used to fabricate a CuxOy-
ZnO:AI p-n j unction solar cell (figure 5.] 9).
Table 5.8: Deposition conditions for optimized Copper Oxide thin film.
Argon flow rate 20 Sccm
Oxygen flow rates 10 Seem
Power 200W
Deposition time 15 min
Chamber temperature 328 K
Base pressure 5 x 10-6 mbar
Sputtering pressure 5.5 x lO-J mbar
85
Table 5.9: Table showing optimized deposition parameters for AZO thin film.
Doping (AI) % 3
Gas flow rate (02) 20Sccm
Deposition pressure 3.0 x 10-' mbar
Base pressure 5 x 10-0mbars
Chamber temperature 780K
Deposition current 4A
Agcontact
Agcontact
+
Figure 5.39: Schematic diagram for CuxOy-ZnO:A1p-njunction
solar cell
5.4.2 Solar cell characterization
Using a solar cell simulator, data was extracted which was used to plot the I-V
characteristic in figure 5.20. From the I-V characteristic the open circuit voltage was
computed as Voc = 0.57 V; the short-circuit current Isc = 1.12 mA/cm2. The short circuit
current was slightly below some other reported values. This was attributed to high
86
resistance within the cell as a result of impurities during the deposition process. From
maximum power point, the fill factor (FF) can be calculated. Fill factor measures quality
of device through squareness of current -voltage curve and was given as 0.6423. With
total radiation (Pin=1000 W/m2) incident on the thin film solar cell, conversion efficiency
(I]) was 0.42 %.
This study observes a solar cell efficiency of 0.42%. Previous studies reported p-Cu-Ozn-
ZnO prepared by Sheng et al (2006) for use in solar cells. Ish =4 mAlcm2 ,Voc= 0.35 V.
ZnOlCu20 heterojunction solar cells have rectifying interface and suitable for oxide
based solar cells (Balamurunga and Mehta,200 1; Ozawa, et al., 2009). Ish =6 mAlcm2
,Voc= 0.4 V. Mitiga et al, (2009) used same arrangement and reported Voc=0.595,
FF=0.5 Isc=678mAlcm2 and 11=2 %.
2.0
1.8
1.6
rr-; 1A'" Eo--- 1.2<E'--'
+-> 1.0c
Q)l: 0.8:::lU
0.6
OA
0.2
87
I
(){){){){){){){){){){){)
() 4)
()
e
4)
()
-
()
O.O~--~----,---~----,,----,----,,----,----,,--~
0.0 0.2 OA 0.6 0.8
Voltage (V)
Figure 5.40: Solar cell characteristics for the fabricated Cu.O, - ZnO:Al p-njunction
88
CHAPTER SIX
CONCLUSIONS AND OUTLOOK
6.1 Conclusion
Deposition of thin films of Cu.O, and ZnO:AI was done by de magnetron reactive
sputtering and reactive thermal evaporation techniques respectively. The re£1ectance and
transmittance data of the films were measured. Aluminum-doped ZnO films have high
transmittance above 70 % and the corresponding reflectance of below 45% within the
visible range. For Cu.O, films, transmittance decreased with increase in gas flow rates
between 30 % to 90 %. Reflectance was generally low within the visible range. Optical
band gap depended on Oxygen flow rate for Copper Oxide recording values of 1.82 eV to
2.54 eV. AZO observed optical band gap ranging between 3.18 eV and 3.42 eV. In both
cases conductivity increased with Oxygen flow rate from 2.09 eV at Oxygen flow rate of
5 Seem to minimum of 1.82 eV at flow rate of 10 Seem and doping levels for AZO from
3.34 eV for undoped ZnO to minimum of 3.18 eV at 4 % Aluminum doping. The XRD
spectra indicate that the films were amorphous for CuO and crystalline in nature for
AZO.
Fabrication of the solar cell was done successfully and from the I-V characteristics the
open circuit voltage was observed to be Voc = 0.579 V, the short-circuit current Isc =
1.1193 mAlcm2 with an efficiency of 0.42 %.
89
6.2 Recommendations
This research mainly investigated the structural and compositional, optical and electrical,
properties of thin films using spectrophotometer 3700 machine, SCOUT 98 program,
XRD and XRF techniques. However, further research is recommended by use of methods
such as Scanning Electron Microscopy (SEM) and Transmission Electron Microscopy
(TEM) in order to obtain detailed results on the thin film properties. There is also need to
study the effect of etching, post-deposition annealing and dye sensitizing on optical and
electrical properties of the thin films and solar cell. This will enable production of thin
films with improved properties for solar cell applications. Also recommended is the use
of appropriate anti-reflective coating (ARC) so as to comp.ensate for high reflection
losses of thin film of CuxOyl ZnO:AI solar cell. We recommend modification of device
geometry to establish whether optical properties and interface defects would be
minimized by sputtering copper oxide on AZO.
90
REFERENCES
Agumba, J. O. (2010). Design and fabrication of a simple four point probe system for
electrical characterization of thin films. Thesis, Department of Physics. Kenyatta
University.
Akimoto, K., Ishizuka, S., Yanagita, M., Nawa, Y., Paul, G. K. and Sakurai, T,(2006)
Solar EnergyJournal of applied physics 80:715
Almen, O. and Bruce, G. (1961). Collection and Sputtering experiments with nobble gas
ions. Nuclear Instruments and Methods. 11(1): 257-278.
Atanas Tanusevskia., Verka Georgieva (2010). Optical and electrical properties of
nanocrystal zinc oxide films prepared by de magnetron sputtering at different sputtering
pressures. Applied Surface Science .19964-19965
Balamurugan ,B. and Mehta, B.R. (2001) .Optical and structural properties of
nanocrystalline Copper Oxide thin films prepared by activated-reactive evaporation, Thin
Solid Films 396: 90-96.
Bohdansky, 1. (1984). A Universal relation for the sputtering yield of a monoatomic
solids at normal incidence. Japan Nuclear Inst. Met. Phys. Res. B. 2: 578-591.
I
Brown, M. and Jakeman, F. (1996). Theory of four point probe technique as applied to
film layers on conducting substrates. British Journal ofapplied Physics. 17: 1146-1149.
Chapin,D.M., Fuller.C.S., and Pearson.G.L.(1954). Jounal of applied physics. 25: 676
Davis, E.A and Mott, N.F(1970).Conduction in non-crystalline systems conductivity,
optical absorption and photoconductivity in amorphous semiconductors. Philosophical
Magazine 22: 903-922.
Deshmukh, M.K. and Deshmukh, C.S., (2008).Modeling of hybrid renewable energy
systems, renewable and sustainable energy reviews. Thin Solid Films 12 235-249
Drobny Y.F.,. Pulfrey, D.L.(1979) . Properties of reactively-sputtered Copper Oxide thin
films. Thin Solid Films 61: 89-98.
Elmin Bacaksiz, Serdar Aksu, Salih Yylmaz , Mehmet Parlak , Mustafa Altunba. (2009).
Structural, optical and electrical properties of Al-doped ZnO micro~'o'dsprepared by
spray pyrolysis.Thin Solid Films .133 :245-253.
Green, M. A. (April 2002). "Third generation photovoltaic solar cells for 2020 and
beyond", Low -diniensional Systems and Nanostructures, 14: 65-67.
91
Gupta, N., Alapatt, G. F., Podila, R., Singh, R. and Poole, K.F. (2009). Prospects of
Nanostructure-Based Solar Cells for Manufacturing Future Generations of Photovoltaic
Modules. International Journal of Photoenergy: 10: 1155.
Hailing Zhu., Junying Zhang., Chunzhi Li ., Feng Pan .,Tianmin Wang and Baibiao
Huang. (2009) .Cu20 thin Flms deposited by reactive direct current magnetron sputtering.
Thin Solid Films 517: 5700--5704.
Hirshman, J., William, P., Hering, W., Schmela, T. and Michael, Y. (2008). Market
Survey: Cell & Module Production http://www.photon-magazine.com/Photon
International], pp. 152
Jaeyoung, L. and Yongsug, T. (2000). Electrodeposition of thin films for solar energy.
London. Academic Press Limited.
Jinsu Yoo, Jeonghul Lee, Seokki Kim, Kyunghoon Yoon, .TunPark, Dhungela S.K.,
Karunagarana B., Mangalaraja D. and .TunsinYia. (2005). High transmittance and low
resistive ZnO:AI films for thin film solar cells. Thin Solid Films 480:213- 217
Jin Ma, Feng Ji, Hong-lei Ma and Shu-ying Li. (2000). Preparation and properties of
transparent conducting zinc oxide and Aluminum-doped zinc oxide films prepared by
evaporating method. Solar energy materials & Solar Cells 60: 341-348.
Karanja, .T.O. (2009). Study of (Sn.Se.) thin films for phase change memories (PCM)
applications. Thesis, Department of Physics. Kenyatta University.
Kawaguchi ,K,. Kita, R,. Nishiyama ,M. and Morishita ,T.(1994). Molecular beam
epitaxy growth of CuO and CuO films with controlling the oxygen content by the flux
ratio ofCu/O+.Journal of Crystal Growth 143 :221-226.
Kemell, M. (2003). Electrodeposition of CuInSe2 and doped ZnO thin films for solar
cells. University of Helnsiki, Faculty of Science, Finland. 15: 59-61.
Kita, R., Hase, T,.1tti ,R,.Sasaki, M,. Morishita, T. and Tanaka .S. (1992). Synthesis of
cupric oxide films using mass-separated low-energy beam. Applied Physica Letters 60
:2624-2630.
Kyungsoo .Tang, Hyeongsik Park, Sungwook .Tung,Nguyen Van Duy , Youngkuk Kim,
Jaehyun Cho Hyungwook Choi , Taeyoung Kwon , Wonbaek Lee, Daeyeong Gong,
Seungman Park, .Tunsin Yi ,Doyoung Kim , Hyungjun Kim. (2010). Optical and
electrical properties of 2 wt.% Ah03-doped ZnO films and characteristicsof Al-doped
ZnO thin-film transistors with ultra-thin gate insulators. Thin Solid Films 518: 2808-
2811.
92
Matsunanmi, A., Yamamura, Y., Itikawa, Y, Itoh, N., Kazumata, Y., Miyagawa, S.,
Morita, K., Shimizu, R. (1980). A serniempirical formula for the energy dependenc of the
sputtering yield. Radiation Effects Letters. 57: 15-21.
Markvat, T. and Castaner,L. (2005). Solar Cells: Materials,Manufacture and Operation.
New York. John Wiley & Sons Ltd.
Markvart, T. (1998). Solar Electricity, John Wiley & Sons. New York. 34-35.
Mittiga, A, Biccari, F.,and Malerba, e. (2009). Thin Solid Films. Journal of applied
physics 517: 2469 -2472
Mitchell, M., Canet, P. and Xiong, K. (1991). Solar cells of high efficiency, Journal of
non-crystalline solids; 280: 116-121.
Mittiga, A, Salza, E., Sarto, E, Tucci, M.and Vasanthi, R. (2006). Applied .Physics.
Letters 88, 163502,
Mudjat, C, Saliha, 1., Yasemin, e. and Fahrettin, Y (2007). The effects of Aluminium
doping on the optical constants of ZnO thin films. Journal of Material Science. 19: 704-
708.
Murat Nezir Eron. (1984).Thin film CulnSe2/Cd(Zn)S Heterojunction Solar Cell:
Characterization and Modeling. PhD. Theseis, Drexel University, 1984, Philadelphia
Markvart, T. and Castener (2005). Solar cell I1W teria Is, New York, Canada: John Wiley
& Sons Ltd. 23-24.
Muthe, K.P., Vyas, J.e., Narang, S.N,. Aswal, D.K. Gupta, S.K., Bhattacharya ,D., Pinto
,R., Kothiyal, G.P, and Sabharwal, S.C. ( 1998) . A study of the CuO phase formation
during thin film deposi tion by molecular beam epitaxy, Thin solid tilms324 :37-43.
Oba, F., Ernst, F., Yu, Y, Liu, R., Kothari, H. M and. Switzer, J. A (2005). Thin Solid
Films. Joumal of applied physics 88:2
O'Leary, J., Johnsons, R. and Lim, P. (1997). Thin film Physical models. Journal of
Applied Physics., 8 (7).
Ohring, M. (1992). TIle Materials Science of thin films. London. Academic Press Ltd, 79-
109.
Ogwu, A A, Danna, T. T. and Bourquerel, E. (2007). Electrical resistivity of copper
oxide thin films prepared by reactive magnetron sputtering. Journal of achievements in
materials and manufacturing engineering 24(1): 172
93
Ozawa, K.,Oba Y.,Edamota, K. (2009).Fonnation and characterization of the CU20
overlayer on Zn(OOO1) Journal of science 603(13): 2163-2170.
Ogwu, AA Bouquerel, E., Ademosu, 0.,Moh, S., Crossan ,E, and Placido, F.(2005).
An investigation of the surface energy and optical transmittance of copper oxide thin .lrns
prepared by reactive magnetron sputtering Thin Film Centre, Electronic Engineering and
Physics Division,.Acta Materialia 53 :5151-5159
Ogwu, A A , Bouquerel, E., Ademosu, 0., Moh, S., Crossan ,E andPlacido, F (2005).
The influence of rf power and oxygen flow rate during deposition on the optical
transmittance of Copper Oxide thin films prepared by reactive magnetron sputtering
Journal of Physics. 38: 266.
Papadimitropoulos ,G., N. Vourdas, N., Vamvakas, E., Davazoglou ,D. (2006). Optical
and structural properties of copper oxide thin films grown byoxidation of metal layers
Institute of Microelectronics. Thin Solid Films 515 :2428-2432.
Pethe, S. (2004). Optimization of the Two Stage Process for Cuiln, Ga)Se2 Solar Cells,
M.Sc (Physics) Thesis. University of South Florida. 1-16.
Reddy, A Sivasankar, A, Uthanna, S. and Sreedhara Reddy, P. (2001). Structural and
electrical characterization of de reactive magnetron sputtered Cu20 films. Department of
Physics, Sri Venkateswara University, Tirupati 396: 90-96.
Richardson, T.J.,. Slack, J.L. Rubin, M.D. (2000). Electrochromism of copper oxide thin
films, Applied Physica Letters 98: 262-263.
Rodrigo, F. Silva, Maria, E.D. and Zaniquelli. (2004). Aluminum-doped zinc oxide films
prepared by an inorganic sol.gel routeThin Solid Films 449: 86-93.
Rnjdar Rauff, M. and Ali Bakr Ali (2007). Optical Properties of Thin Film Sulaimani
University College of Science Physics Department volume 24 issue I
Selmi, M., Chaabouni, F., Abaab, M. and Rezig, B. (2008). Superlattices and
Microstructures Studies on the properties of sputter-deposited Al-doped ZnO films
Superlattices and Microstructures 44 :268-275.
Shadia, J., Naseem, M. and Riyad, N. (2009). Electrical and Optical properties of ZnO:AI
thin films prepared by pyrolysis technique. Faculty of Science; Physics department,
University of Jordan; Ammam, 11942, Jordan.
Sheng S., Guoji, F., Zhigiang, C. and Linggang, F. (2006). Semiconductor science and
technology ;145: 86-589.
94
Shar, A., Torres, P., Tscharner, R., Wyrsch, N., Keppner, H. (1999). Science, 285: 692-
698. Stone, J. (1993). Physics Today, 22-29.
Stanley Wolf and Richard N. Tauber, (2000). Silicon Processing for the VLSI Era:
Process Technology 1:635
Su Sheng, Guoji, F.,Chunli, Zhigiang, C.,S. ma, Linggang, F and Zhao,X.(2006).
Semiconductor science and technology. 586-589.
Su-Shia Lina, Jow-Lay Huanga and Sajgalik .P. (2005). Effects of substrate temperature
on the properties of heavily Al-doped ZnO films by simultaneous r.f. and d.c. magnetron
sputtering Surface & Coatings Technology 190: 39- 47
Tang, N., Tomoyuki, K. and Takahiro, M. (2000). Thin Film Solar Cells with ZnO Buffer
Layers, Journal of Applied Physics, 37: 499-501.
Theiss, W. (2000). Scout thin films analysis software handbook, edited by Theiss M
(Hand and Software Aachen German) www.mtheiss.com.
Tsubomura, H., Kobayashi, H., (1993). Critical Revolution of Solid State Matter. Science
18: 261-326.
Waita,S.M., (2008). Dye sensitized Ti02 solar cells. PhD. thesis, university of Nairobi.
Wang, Zhong, L. (2004). Zinc Oxide Nanostructures: Growth, Properties and
Applications. Journal of Physics_lG: 829-858.
Williams, R. (1960). "Becquerel Photovoltaic Effect 111 Binary Compounds". 171e
Journal of Chemical Physics 32 (5): 1505-1514.
Wooten, F. (1972). Optical properties of solids, Academic Press, New York. Pp 41-80.
Yaodong Liu,Qiang Li,and Huiliang Shao, (2009). Optical and photoluminescent
properties of AI-doped zinc oxide thin films by pulsed laser deposition .journal of alloys
and compounds. Thin solid physics 485: 529-531
Zhang, D., Liu, L. and Hyang, G. (2004). The Electrical properties and interface of
CU20/ZnO/ITO P-i-N heterojunction, Journal of Applied Physics, 5: 178-183.
Zhao, Y., Xinhua,. G. and Wang, W. (2002). R&D Activities of Silicon-based thin film
solar cells in China. Thin Film Device and Technology, 203: 714-'720.
Sze, S. and Ng, K. K. (2006) Physics and Properties of Semiconductors-A Review, in
Physics of Semiconductor Devices, John Wiley & Sons, Inc., Hoboken, N.J, USA.
doi: 10.1002/9780470068328.chl
9S
(http//solarcooking.org/saussaure.htm, (22nd Jan 2011)
(http:www.pveducation.org/pvcdrom, (12tJ1 Feb 2011)
96
APPENDICES
Appendix I: Photograph of Edwards Auto 306 vacuum coater
Appendix II: Photograph of Keithley Source Meter 2400 model
97
Appendix Ill: Photograph of the Designed and Fabricated Sheet Resistivity
Measurement system (Agumba,2010)
Appendix IV: Photograph of MiniPal2 XRF Spectrometer
98
Appendix V: Photograph ofPW 3040/60 X'Pert XRD diffractometer
Appendix VI: Solar cell simulator
99
Appendix VII: UV-VIS NIR Spectrophotometer solid state 3700 DUV.
_:,'" ._....~.·~ __ .!if:!lf. -:1'.1:' ~1ii;;:.... :..1- .:-~.. , P ''OClr •• to".
ENY ITA UNIVERSITY LIBRARY