©Copyright by Devendra Khatiwada 2019 All Rights Reserved
Fabrication and Optimization of Single-Junction GaAs Thin Film Solar
Cells on Epi-ready Flexible Metal Tapes for Low-cost Photovoltaics
A Dissertation
Presented to
the Faculty of the Department of Mechanical Engineering
University of Houston
In Partial Fulfillment
for the Requirements of the Degree
Doctor of Philosophy
in Materials Science and Engineering
by
Devendra Khatiwada
August 2019 Fabrication and Optimization of Single-Junction GaAs Thin Film Solar
Cells on Epi-ready Flexible Metal Tapes for Low-cost Photovoltaics
Devendra Khatiwada
Approved:
Chair of the Committee Venkat Selvamanickam, Professor, Mechanical Engineering
Committee Members:
Haleh Ardebili, Associate Professor, Mechanical Engineering
James K.Meen, Research Associate Professor, Chemistry
Jae -Hyun Ryou, Associate Professor, Mechanical Engineering M e c Alamgir Karim, Associate Professor, h Material Science and Engineering a n i c
Suresh K.Khator, Associate Dean, Alamgir Karim, Professor and a Cullen College of Engineering Director, Material Science and l Engineering E n g ii i n e e r Acknowledgements
Firstly I would like to acknowledge my thesis advisor professor Venkat
Selvamanickam for his persistent help, guidance and offering me opportunities to involve in various interesting project. The last four years has been an outstanding experience to work with him. His patience in research, scientific knowledge, hard work and technical skills always inspired me to continuously work in different projects. I am very lucky to have him as my advisor.
It was a great pleasure to work with the excellent members in our research group.
Special thanks to Dr.Pavel Dutta, Dr.Monika Rathi, Sicong Son, Carlos Favela, Sahil
Sharma, Sara Pouladi for their great help during my research work. Special thanks to
Sicong Son who was always enthusiastic to help me at any time. I appreciate help and guidance from Dr.Jae-Hyuan Ryou in each step of my research work.
I would like to thank my committee member Dr.Alamgir Karim, Dr. James
K.Meen and Dr.Haleh Ardebili. Thanks to Dr.Alamgir Karim for providing all the support needed from the department.
I would specially thank to my family member for their immense support. My lovely wife, Yojana Ghimire has been extremely supportive in all the circumstances and through this entire process. Without her it would have been impossible to complete my goal of research.
v
Fabrication and Optimization of Single-Junction GaAs Thin Film
Solar Cells on Epi-ready Flexible Metal Tapes for Low-cost Photovoltaics
An Abstract
of a
Dissertation
Presented to
the Faculty of the Department of Mechanical Engineering
University of Houston
In Partial Fulfillment
for the Requirements of the Degree
Doctor of Philosophy
in Materials Science and Engineering
by
Devendra Khatiwada
August 2019
vi
Abstract
Properties like high efficiency, flexibility and light weight, resistance to UV radiation and moisture and a low temperature coefficient make gallium arsenide (GaAs) more favorable than the ubiquitously-used silicon for solar cells. In spite of their high efficiency, GaAs solar cells have found limited use in various application due to high cost of the GaAs or Ge wafer used. In an approach to produce low-cost GaAs solar cells, we have developed a technology to grow epitaxial semiconductor thin films on low-cost flexible epi-ready metal tapes that can replace the expensive wafers. The template layers on the epi-ready metal tapes are grown via a roll-to-roll process using Ion Beam Assisted
Deposition (IBAD).
Metal organic chemical vapor deposition (MOCVD) is used to epitaxially grow
GaAs solar cells structure on ‘single-crystalline-like’ germanium film on epi-ready metal foils. These epitaxial GaAs films exhibit excellent crystalline alignment with high carrier mobility, excellent crystalline alignment and optoelectronic properties. These grown thin films were processed via photo-lithography, etching and contact deposition to fabricate single-junction (1J) GaAs solar cell devices. The fabricated solar cell underwent a process of cap layer removal (passivation) and anti-reflection coating. Efficiency greater than 6% with open circuit voltage (VOC) of 566mV, fill factor (FF) of 68%, short circuit
2 current density (JSC) of 17.4mA/cm was obtained.
Steps were taken to improve the quality of p-n junction by improving the quality of Ge template and incorporating an intrinsic layer with p-i-n solar cell structures. A
2 device efficiency of 11.5 % with VOC of 566mV, FF of 68%, JSC of 17.4mA/cm was
vii obtained at 1 Sun on improved Ge template, using CVD germanium instead of sputtered germanium used before. Solar cells fabricated with the improved p-i-n structure showed a
−2 device efficiency of 13.2% at 1 sun with VOC of 650 mV, JSC of 28 mA cm , and FF of
72 %. These thin film GaAs photovoltaics, with further improvement in quality, can potentially lead to light-weight, inexpensive and scalable solar cell manufacturing.
viii
Table of Contents
Acknowledgement……………………………………...……………………..…...….…..v
Abstract………………………………………………………………….……………….vii
Table of Contents………………………………...………………………..………….…..ix
List of Figures……………………………………..………………...... …...…..xiv
List of Tables……………………………………………...……………………...……..xix
List of Abbreviation……………………………………………..…...……………...... xxi
Chapter 1 Introduction ...... 1
1.1 Introduction to Solar cells ...... 1
1.1.1 Background...... 1
1.2 Photovoltaics (PVs) ...... 1
1.2.1 Semiconductors as photovoltaic materials ...... 3
1.3 The p-n junction in semiconductors ...... 6
1.3.1 Mechanism and characteristic curve for p-n junctions under bias ...... 9
Chapter 2 Solar Cells: Structures and Characterizations ...... 13
2.1 Solar or Photovoltaic Cell ...... 13
2.2 Solar Cell Structure ...... 15
2.3 Solar Cell Characteristics ...... 18
2.3.1 Equivalent circuit for solar cells ...... 19
2.4 Solar cell parameters used for measurements ...... 21
ix
2.4.1 Short circuit current density (Jsc) ...... 22
2.4.2 Open circuit voltage (Voc) ...... 23
2.4.3 Leakage current (Io) ...... 24
2.4.4 Maximum power point (Pmax) ...... 25
2.4.5 Fill Factor (FF) ...... 25
2.4.6 Efficiency ...... 25
2.4.6.1 Air Mass (AM) ...... 26
2.5 Series and shunt resistance in solar cell ...... 27
2.6 Loss mechanism in solar cells ...... 29
2.6.1 Optical Loss ...... 29
2.6.1.1 Absorption loss ...... 29
2.6.1.2 Reflection Loss ...... 31
2.6.2 Recombination loss ...... 33
2.6.2.1 Direct recombination ...... 34
2.6.2.2 Auger recombination ...... 34
2.6.2.3 Trap-assisted recombination ...... 35
2.7 Conclusion ...... 35
Chapter 3 III-V semiconductor material and solar cells fabrication and characterization 36
3.1 Introduction ...... 36
3.2 Growth Methods of III-V Semiconductor ...... 37
x
3.2.1 Liquid-encapsulated Czochralski (LEC) Method ...... 37
3.2.2 Float Zone and Bridgman ...... 38
3.2.3 Liquid Phase Epitaxy (LPE) ...... 38
3.2.4 Molecular Beam Epitaxy ...... 39
3.2.5 Metal Organic Chemical Vapor Deposition (MOCVD) ...... 40
3.3 III-V Single-junction Gallium Arsenide Solar Cells Structure and Design 42
3.4 III-V Multi -junction Solar Cell Structure and Design ...... 44
3.5 Characterization tools for characterizing III-V materials and Solar cells ... 47
3.5.1 X-Ray Diffraction (XRD) ...... 47
3.5.2 Hall Measurement ...... 49
3.5.3 Photoluminescence intensity (PL) ...... 53
3.5.4 Secondary Ion Mass Spectroscopy (SIMS) ...... 55
3.5.5 Solar simulator...... 56
3.5.6 External Quantum Efficiency (EQE) ...... 57
3.5.7 Atomic Force Microscopy (AFM)...... 59
3.5.8 Transmission line method (TLM) ...... 60
3.6 Conclusion ...... 64
Chapter 4 Next-generation solar cells for low cost and high solar cell efficiency ...... 65
4.1 Low-cost substrates for growth of III-V semiconductor materials ...... 65
4.2 Epitaxial lift off process for GaAs substrate reuse ...... 66
xi
4.3 Low-cost and scalable process for III-V solar cells ...... 68
4.4 Growth of epi-ready buffer on metal tape ...... 70
4.4.1 Growth of semiconductor buffer over epi-ready buffer ...... 72
4.4.2 Growth of 1J GaAs solar cell on top of semiconductor buffer...... 75
4.5 Conclusion ...... 76
Chapter 5 Fabrication and characterization of single-junction (1J) GaAs solar cells on
sputtered germanium over epi-ready buffer ...... 77
5.1 Device fabrication process for 1J GaAs solar cells on epi-ready buffer ..... 78
5.2 Measurement of fabricated 1J GaAs solar cells ...... 79
5.3 Impact of citric acid passivation on 1J GaAs solar cells ...... 86
5.4 Impact of hydrogen and hydrogen/phosphine passivation on 1J GaAs films
...... 94
5.5 Conclusion ...... 104
Chapter 6 Improving the quality of a germanium template using R2R PECVD Ge for
high-quality 1J GaAs solar cells ...... 106
6.1 Properties of CVD Ge on flexible substrates ...... 108
6.2 Properties of flexible GaAs on R2R CVD Ge ...... 111
6.3 1J GaAs solar cell results on the sputtered and CVD germanium ...... 113
6.4 Conclusion ...... 119
xii
Chapter 7 Improving the efficiency of single-junction GaAs solar cells by incorporating
an intrinsic layer in the p-i-n solar cell structure...... 121
7.1 Optical characterization of p-n and p-i-n solar cell device film ...... 124
7.2 Fabrication and characterization of single-junction p-i-n solar cells ...... 128
7.3 Transmission electron microscopy analysis for p-i-n and p-n solar cell
stacks ...... 137
7.4 Time-of-Flight SIMS analysis for p-i-n and p-n films ...... 139
7.5 Photoelectric response for p-n and p-i-n solar cells ...... 141
7.6 Conclusion ...... 142
Chapter 8 Summary and Future Work ...... 144
References……………………………………………………………………………....147
xiii
List of Figures
Figure 1. 1 Excitation of an electron in semiconductor material from the valence band to the conduction band with a photon energy greater than the band gap energy...... 4
Figure 1. 2 Fermi energy level in n-type and p-type semiconductors...... 5
Figure 1. 3 Energy diagram of a p-n junction in thermal equilibrium...... 6
Figure 1. 4 Band diagram for a (a) forward and (b) reversed bias p-n junction...... 10
Figure 1. 5 Forward and reverse characteristics of a p-n junction...... 11
Figure 2. 1 Structure of p-n junction (single junction) solar cells...... 16
Figure 2. 2 Band diagram for p-n junction with window layer...... 17
Figure 2. 3 Equivalent circuit of a solar cell...... 19
Figure 2. 4 Current (I) – Voltage (V) curve (red plot) of solar cell with power curve (green plot)...... 22
Figure 2. 5 Impact of series resistance on I–V characteristics of a solar cell...... 28
Figure 2. 6 Impact of shunt resistance on current voltage curve of solar cell...... 28
Figure 2. 7 Spectral Irradiance for solar spectrum (AM1.5, 1000 W/m2)...... 30
Figure 2. 8 Optical losses in solar cell...... 31
Figure 2. 9 Recombination mechanism in solar cell (a) direct (b) trap-assisted or indirect and (c) Auger recombination...... 34
Figure 3. 1 Composition vs lattice constant for III-V materials...... 36
Figure 3. 2 Schematic of liquid-encapsulated Czochralski method...... 37
Figure 3. 3 Schematic of Float Zone and Bridgman process...... 38
Figure 3. 4 Photograph of Liquid Phase Epitaxy system at our facility...... 39
Figure 3. 5 Schematic of MBE process...... 40
Figure 3. 6 Photograph of MOCVD Chamber at University of Houston...... 41
Figure 3. 7 Single-junction GaAs solar cells structure on (a) Conducting buffer (b) Non conducting buffer...... 42
xiv
Figure 3. 8 Ohmic contact formation for n type (a), (b) and p type (c), (d) semiconductor...... 43
Figure 3. 9 Triple-junction solar cell (Monolithic structure)...... 45
Figure 3. 10 NREL Best Research Cell Efficiency...... 46
Figure 3. 11 (a) Schematic of X-Ray Diffraction, (b) Photograph of XRD at UoH...... 48
Figure 3. 12 (a) Schematic of the Hall Measurement method (b) Photograph of Hall measurements at the University of Houston...... 50
Figure 3. 13 (a) Schematic of Photoluminescence, (b) Photograph of PL system...... 54
Figure 3. 14 Schematic of Secondary Ion Mass Spectroscopy ...... 55
Figure 3. 15 (a) Schematic diagram of a solar simulator (b) Set up for solar cell efficiency measurements ...... 56
Figure 3. 16 EQE comparison of ideal solar cell with experimental...... 57
Figure 3. 17 (a) Schematic of EQE system, (b) Photograph of EQE system used in this work...... 59
Figure 3. 18 Schematic of the atomic force microscope...... 60
Figure 3. 19 Illustration of linear contact arrays for top and bottom contact...... 60
Figure 3. 20 Distance vs resistance plot for TLM...... 62
Figure 4. 1 Schematic of 1J GaAs solar cell and buffer layers on a flexible metal tape...... 69
Figure 4. 2 Cross sectional microstructure and selected-are diffraction patterns from an epi-ready buffer architecture...... 70
Figure 4. 3 R2R equipment for (a) Ion Beam Assisted Deposition (IBAD) and (b) Magnetron Sputtering...... 71
Figure 4. 4 Custom-built dual MOCVD tool...... 73
Figure 4. 5 Structure of semiconductor buffer layer...... 74
Figure 4. 6 Structure of (a) p/n and (b) p/i/n solar cells...... 75
Figure 5. 1 Photomask design for solar cell fabrication...... 77
Figure 5. 2 Schematic diagram showing the fabrication steps of a single-junction (1J) GaAs solar cell on flexible epi-ready metal tape...... 78
xv
Figure 5. 3 Transmission line method (TLM) for (a) top and (b) bottom contacts...... 80
Figure 5. 4 (a) Illuminated J-V characteristic and (b) dark current-voltage (I-V) responses of the fabricated GaAs solar cell...... 80
Figure 5. 5 (a) Illuminated J-V characteristic and (b) dark I-V responses of the GaAs solar cell, both as-fabricated and with the cap layer removed...... 81
Figure 5. 6 Reflectance vs. wavelength graph for double-layer ZnS/MgF2 anti- reflective coating (ARC) on the 1J GaAs solar cell...... 83
Figure 5. 7 (a) Illuminated J-V characteristic and (b) dark I-V responses of the as- fabricated, cap-layer-removed, and ARC-coated GaAs solar cells...... 83
Figure 5. 8 Illuminated J-V characteristics for (a) wafer and (b) flexible 1J solar cells...... 85
Figure 5. 9 (a) and (c) AFM images of GaAs before and after passivation; (c) and (d) SEM images of GaAs before and after passivation...... 87
Figure 5. 10 XPS spectra of GaAs film before and after passivation using citric acid. (a) Ga 3d core level with its oxide, (b) As 3d core level with its oxide, and (c) O 1s core level...... 88
Figure 5. 11 Evaluation of solar cells with three different base thicknesses (1140 nm, 840 nm, and 380 nm). (a) and (c) Dark current voltage responses (b) and (d) illuminated J-V characteristics of the devices...... 89
Figure 5. 12 Impact of passivation on solar cells A and D. (a) and (c) Dark I-V responses (b) and (d) illuminated J-V characteristics of solar cells A (500 μm) and D (1250 μm) at the as-fabricated and final stages...... 92
Figure 5. 13 XRD and FWHM for the as-deposited and H2 plasma-treated GaAs films on epi-ready metal substrate...... 96
Figure 5. 14 PL spectra of the as-deposited and H2 plasma-treated GaAs films on epi- ready metal substrate...... 97
Figure 5. 15 XPS analysis of the as-deposited and H2 plasma-treated GaAs films on epi- ready flexible metal substrate...... 97
Figure 5. 16 Kelvin probe force microscopy images and line profiles for the as-deposited and H2 plasma-treated GaAs films on epi-ready metal substrate...... 98
Figure 5. 17 Electrical conductivity and resistivity of GaAs films before and after passivation...... 99
Figure 5. 18 Depth profile for hydrogen in GaAs solar cells...... 101
xvi
Figure 5. 19 Illuminated and dark I-V curves for the as-deposited and plasma-treated GaAs solar cells on epi-ready metal substrate...... 102
Figure 5. 20 Illuminated and dark I-V curves for as-deposited and phosphine- incorporated hydrogen plasma-treated GaAs solar cells on epi-ready metal substrate...... 103
Figure 6. 1 Schematic of PECVD Ge deposition on R2R chamber...... 106
Figure 6. 2 AFM image for CVD and sputtered Germanium...... 108
Figure 6. 3 XRD for sputtered and CVD germanium with a rocking curve...... 109
Figure 6. 4 In-plane texture of sputtered and CVD germanium...... 109
Figure 6. 5 Raman shift for the sputtered and CVD germanium...... 110
Figure 6. 6 (a) XRD and (b) rocking curve for GaAs on the sputtered and CVD germanium (c) (220) pole figure for GaAs on CVD germanium...... 112
Figure 6. 7 Raman spectra for p-GaAs on sputtered and CVD germanium...... 112
Figure 6. 8 STEM Image for GaAs on sputtered and CVD Ge...... 113
Figure 6. 9 J-V characteristic of the solar cell fabricated on CVD germanium...... 114
Figure 6.10 J-V characteristic for the 1J GaAs solar cell on sputtered and CVD germanium ...... 115
Figure 6. 11 Dark I-V for 1J GaAs on sputtered and CVD germanium...... 117
Figure 6. 12 External quantum efficiency for a 1J solar cell...... 118
Figure 7. 1 Band profile for p-n and p-i-n structures...... 122
Figure 7. 2 Schematic of a 1J p-i-n solar cell and buffer layer on flexible metal tape. 123
Figure 7. 3 Optical, SEM and AFM images of p-n films...... 125
Figure 7. 4 HIM images of p-n film surfaces...... 126
Figure 7. 5 Optical, SEM and AFM images of p-i-n film...... 127
Figure 7. 6 HIM images of p-i-n film surfaces...... 128
Figure 7. 7 Transmission line method (TLM) for (a) top and (b) bottom TLM...... 129
Figure 7. 8 Illuminated J-V characteristics of the fabricated p-i-n GaAs solar cells. .. 130
xvii
Figure 7. 9 Variation of JSC, VOC and FF for p-i-n cells with different intrinsic layer thicknesses...... 131
Figure 7. 10 Dark I-V characteristics of the fabricated p-i-n GaAs solar cell...... 131
Figure 7. 11 Forward-bias component of dark curve fit...... 133
Figure 7. 12 (a) Illuminated and (b) dark current-voltage characteristic of a p-i-n solar cell with a 1000-nm intrinsic layer...... 136
Figure 7. 13 TEM analysis for cross-sections of the p-n and p-i-n devices...... 138
Figure 7. 14 Depth profile analysis of elements in p-n (a) and p-i-n (b) films. Comparison of Zn profiles in p-n and p-i-n films are shown in (c) ...... 140
Figure 7. 15 Photoelectric response for p-n and p-i-n solar cells...... 142
xviii
List of Tables
Table 2. 1 Types of different generations of solar cells ...... 13
Table 5. 1 Leakage Current (Io), Open circuit voltage (VOC), Short circuit current density (JSC), Fill Factor (FF), and Efficiency (η) of solar cells at initial stage of device fabrication...... 81
Table 5. 2 Io, VOC, JSC, FF, and η of solar cells at initial stage of fabrication and after cap layer removal...... 82
Table 5. 3 Io, VOC, JSC, FF, and η of solar cells at initial stage of fabrication, after cap layer removal, and after ARC application...... 84
Table 5. 4 VOC, JSC, FF, and η of flexible and wafer 1J solar cells...... 85
Table 5. 5 IO, VOC, JSC, FF, and η at three different base thicknesses, corresponding to before and after passivation...... 90
Table 5. 6 Increase in open circuit voltage (ΔVOC), short circuit current density (ΔJSC), and efficiency (Δη) and decrease in leakage current (ΔIO) in solar cells with three different thicknesses...... 90
Table 5. 7 IO, VOC, JSC, FF, and η of solar cells A (500 μm) and D (1250 μm) at the initial and final stages...... 92
Table 5. 8 ΔVOC, ΔJSC, Δη, and ΔIO in solar cells A (500 μm) and D (1250 μm) at the as-fabricated and final stages...... 93
Table 5. 9 η and IO before and after H2 plasma treatment of GaAs solar cells on epi- ready metal substrate...... 102
Table 5.10 η and IO before and after H2 plasma treatment of GaAs solar cells on epi- ready metal substrate...... 103
Table 6. 1 J-V parameter for the 1J GaAs solar cell fabricated on CVD germanium. 114
Table 6. 2 Device data of GaAs solar cells on CVD and sputtered Ge template showing results of four devices...... 116
Table 6. 3 Representative device data of GaAs solar cells on sputtered Ge template showing results of four different device sizes...... 116
Table 7. 1 VOC, JSC, FF and η of p-i-n solar cells with different intrinsic layer thicknesses ...... 130
Table 7. 2 Leakage current (IO) of p-i-n solar cells with different intrinsic layer thicknesses...... 132 xix
Table 7. 3 Saturation current density from double-diode fitting...... 133
Table 7. 4 Io, VOC, JSC, FF and η of solar cells at the initial stage of fabrication, after cap layer removal and after ARC application...... 137
xx
List of Abbreviations
Vbi Built in potential
Jn,drift Drift current of electrons
Jp,drift Drift current of holes
Jn,drift Diffusion current of electrons
Jp,drift Diffusion current of holes
CdTe Cadmium telluride
CIGS Copper indium gallium arsenide
OPV Organic photovoltaic
DSSC Die synthesized solar cells
GaAs Gallium arsenide
ARC Anti reflection coating
InGaAs Indium gallium arsenide
InAs Indium arsenide
InGaP Indium gallium phosphide
GaInAsP Gallium indium arsenide phosphide
AlAs Aluminum Arsenide
xxi
RHEED Reflection high energy diffraction
MOCVD Metal organic chemical vapor deposition
TMGa Trimethylgallium
TEGa Trimethylgallium
AsH3 Arsine
XRD X-ray diffraction
PL Photoluminescence
SIMS Secondary ion mass spectroscopy
EQE External quantum efficiency
IQE Internal quantum efficiency
AFM Atomic force microscopy
TLM Transmission line method
CVD Chemical Vapor deposition
Ge Germanium
ELO Epitaxial liftoff
LMO Lanthanum manganese oxide
MgO Magnesium oxide
xxii
CeO2 Cerium Oxide
ARC Anti reflection coating
H2SO4 Sulfuric acid
H2O2 Hydrogen peroxide
JSC Short circuit current density
VOC Open circuit volta
FF Fill factor
XPS X-ray photoelectron spectroscopy
(NH4)2S Ammonium sulfide
Ga2O3 Gallium oxide
As2O3 Arsine oxide
KPFM Kelvin probe force microscopy
PECVD Plasma Enhanced chemical vapor deposition
ICP Inductive coupled plasma
GAADS General area detector diffraction
HIM Helium ion microscopy
xxiii
Chapter 1 Introduction
1.1 Introduction to Solar cells
1.1.1 Background
Solar energy is one of the most abundant, renewable sources of energy available on Earth and is comparable to other energy sources, such as wind, geothermal, hydroelectricity, biomass, etc. [1, 2]. Earth receives about 1.8 1017 W of solar radiation each hour, which is 10,000 times greater than the energy consumed by humans in an entire year (2 1013 W). Utilizing only 0.16% of Earth’s land with 10% efficient solar energy conversion system would produce 2 1013 W of solar power [2-5]. Thus, solar energy has the potential to supersede fossil fuels and other renewable energy sources.
However, eighty six percent of Earth’s total energy demand is obtained from fossil fuels
(e.g., oil, coal, natural gas) and some part from nuclear power. Consequently, carbon emissions have become a global problem [6]. China is the largest emitter of fossil fuel bi- products like carbon dioxide (CO2), methane (CH4), nitrous oxide (N2O), and chlorofluorocarbons (CFCs) which can all cause global warming [7, 8]. Solar energy can reduce global warming due to fossil fuels, minimize environment impacts and fulfil the demand of today’s clean and green energy. The conversion of solar photons into electrical energy is accomplished by photovoltaics (PVs) [9].
1.2 Photovoltaics (PVs)
In photovoltaics, sunlight is directly converted into electricity using semiconductor materials. The phenomenon where electromagnetic radiation such as light generates an electromagnetic force when interacting with matter is called the photovoltaic
1 effect [10]. This was first discovered in 1839 by French experimental physicist Edmond
Becquerel when a platinum electrode coated with light-sensitive material [e.g., silver chloride (AgCl) or silver bromide (AgBr)] in electrolytic solution exposed to light.
Becquerel developed a device called ‘actinography’ to record the temperature of a heated body [11]. In 1872, William Grylls Adams and Richard Evans Day discovered that the junction between selenium and platinum showed a photovoltaic effect when exposed to light [12]. The first solid state solar cell was demonstrated by C. Fritts in 1883 with a thin layer of Au on selenium, which formed a Schottky junction that resulted in an internal field separating the charge carriers and produced an efficiency less than 1% [13].
The first modern solar cell was patented by Russell Shoemaker Ohl in 1946, for a photo E.M.F cell formed from a silicon ingot with a rhodium conductive terminal [14].
Later, in 1954, D.M. Chapin, C.S. Fuller and G.L. Pearson at Bell Laboratories demonstrated a new silicon p-n junction photocell that used a thin layer of p-type silicon formed over an n-type base to form p-n junction solar cells [15]. Apart from silicon, different photovoltaic materials were later developed and designed to form p-n solar cells. Materials used for solar cells range from organic, inorganic and hybrid materials.
Organic solar cells use organic compounds like polymer/fullerene [16-18]. Material like cadmium telluride (CdTe), copper indium gallium di-selenide (CIGS), cadmium sulphide
(CdS), or Gallium Arsenide (GaAs) are used as inorganic materials for photovoltaic [19-
21]. Hybrid photovoltaic materials consist of organic and inorganic counter parts [22,
23]. Other types of solar cells include dye-synthesized solar cells (DSSCs), which use liquid electrolytes [24].
2
The phenomenon of photovoltaics can be explained by the quantum theory of light. Light consists of packets of energy called quanta or photons. According to the quantum theory of light, the energy E of a photon corresponding to a wavelength λ is given by
hc E = hν = , (1.1) λ where h = 6.626068 × 10-34 m2kgs-1 is called Planck’s constant. ν, c and λ are the frequency, velocity and wavelength of light, respectively.
When the energy of the incident photon is equal to or larger than the forbidden energy gap of a photovoltaic material, electrons are excited to higher energy states.
Electrical potential is produced due to the separation of charge carriers at two different energy levels.
1.2.1 Semiconductors as photovoltaic materials
Semiconductors are materials whose electrical conductivity can be controlled in various ways. Semiconductor materials are distinguished from insulators and conductors according to their band gap. The conductivity of the semiconductor can be controlled deliberately by controlling the amounts of impurities. Semiconductors are then classified as intrinsic and extrinsic semiconductors [25].
An undoped semiconductor is called intrinsic, whereas a doped semiconductor is called extrinsic. Doping in semiconductors can either be p-type (acceptor like) or n-type
(donor like) [26, 27]. Thus, the properties of semiconductors can be tuned by varying the doping concentration by 13 to 14 orders of magnitude above the intrinsic concentration.
3
For example, the intrinsic doping concentration of a Gallium Arsenide (GaAs) semiconductor at room temperature is 107/cm3, which can be increased by doping to
1020/cm3 [28] .
Photon energy can excite the electron from a lower band energy called the conduction band (EC) to a higher band energy called the valence band (EV), as shown in
Figure 1.1. Electrons are governed by Fermi-Dirac statistics, and we define the probability of occupying an energy level in the valence and conduction band of semiconductor material called the fermi energy (EF) level [29]. For an intrinsic or undoped semiconductor, the number of electrons and holes are equal, and the probabilities of occupying the energy level in the two bands are equal. So, for such a semiconductor, the fermi energy level lies at the center of the energy band, as shown in figure 1.1.
Figure 1. 1 Excitation of an electron in semiconductor material from the valence band to the conduction band with a photon energy greater than the band gap energy.
However, for n-type semiconductors, where the probability of occupancy of the energy level by the electron is greater than that of a hole, the Fermi level lies near the
4 conduction band edge. For p-type semiconductors, where the probability of occupancy of the energy level of the hole is greater than that of an electron, the Fermi level lies near the valance band edge. This is illustrated in Figure 1.2 with Ei as fermi level of undoped or intrinsic semiconductor.
Figure 1. 2 Fermi energy level in n-type and p-type semiconductors.
An n-type semiconductor is obtained by doping the semiconductor with impurities that can provide excess electrons. A p-type semiconductor is formed by doping the semiconductor with impurities that can provide excess holes. For example, n-type silicon is obtained by doping silicon with pentavalent impurities from group V elements of the periodic table. They include phosphorous (P), arsenic (As), etc. Meanwhile, p-type silicon is obtained by doping silicon with trivalent impurities from group III elements of the periodic table like Boron (B), aluminum (Al), gallium (Ga), etc. In n-type semiconductors, electrons are the majority charge carriers, and holes are the minority charge carriers. For p-type semiconductors, holes are majority charge carriers, and electrons are the minority charge carriers.
When two doped semiconductors (n doped and p doped) are bought together, a region in the semiconductor material is formed called a p-n junction. An electric field is
5 formed at the junction, which biases the flow of charge carriers in the opposite direction.
A brief description of the p-n junction is explained below.
1.3 The p-n junction in semiconductors
A p-n junction is formed when one side of semiconductor material is a doped p- type, and the other side is a doped n-type, or when p- and n-doped semiconductors are brought together [30]. The sandwich can be formed by doping, ion implantation, diffusion of dopant with temperature or by epitaxial growth of p-type doped crystal over n-type, or vice versa. The formation of a p-n junction between p-type and n-type semiconductors at equilibrium (no bias) is shown in Figure 1.3 below.
Figure 1. 3 Energy diagram of a p-n junction in thermal equilibrium.
6
Due to a concentration gradient of electron and holes at the junction, electrons from the n-side semiconductor diffuse into the p side and combine with holes. This results in uncharged, immobile, electrically neutral, positive-charged donor ions (ND).
The same type of phenomenon occurs at the p-side junction, which results in uncharged, immobile, electrically neutral negative charged acceptor ions (NA). Thus, an electric field is created at the junction due to these immobile charge carriers, which opposes the flow of charge carriers across the junction. The electric field in turn causes a drift of carriers in the opposite direction. A state of thermal equilibrium is reached when this diffusion current due to the concentration gradient balances the drift current across the junction.
The potential barrier across the junction is called “built-in potential”. The region is defined as the “space charge region” or “depletion region” as all the charge carriers are depleted due to the built-in potential (qVbi). This depletion region creates a barrier to the majority charge carriers, whereas it causes low resistance to minority charge carriers. The region where the electric field is zero on both sides of the depletion region is called the quasi-neutral region. As depicted in the energy vs. distance graph in figure 1.3, at equilibrium (under no external bias), the Fermi energy remains constant throughout the entire p-n region. This is because any change in Fermi energy designates a change in the electric field, which in turn gives rise to a net current, violating the conservation of charge [31].
Due to the above-mentioned phenomenon, after the p-n junction is formed, two types of currents occur: drift and diffusion current [32, 33]. Drift current arises due to the electric field at the junction and affects the majority charge carrier. The drift current of
7 electrons (Jn,drift) in the n region and holes (Jp,dift) in the p region due to the electric field are given by
Jn,drift = qnµn E (1.2)
and
Jp,drift = qpµp E, (1.3) where n and p are the concentrations of the electrons and holes, and µn and µp are the mobilities of the electrons and holes, respectively. The direction of the drift current for the electrons is in the direction opposite to the electric field; for the holes, it is in the direction of the applied electric field.
A diffusion current arises due to the concentration gradient of electrons and holes.
The diffusion current in a p-n semiconductor is governed by minority charge carriers. The diffusion current density of electrons (Jn,diff) in a p-type semiconductor and holes (Jp,diff) in n-type semiconductor is given by
dn J = qD x (1.4) n,diff n dx
and
dp J = −qD x , (1.5) p,diff p dx where q is the electron charge, and n and p are electron concentrations of electrons and holes. Dn and Dp are diffusion coefficients of electrons and holes, respectively. The diffusion of electrons and holes in the p and n sides of semiconductor material are limited by its diffusion length. Diffusion length is defined as the average distance that the
8 minority charge carriers move before being recombined or annihilated. The diffusion of minority charge carriers (electrons) in p-type semiconductors, Ln , is given by
Ln = √Dn τn (1.6) whereas, the diffusion of minority charge carriers (hole) in n-type semiconductors Lp is given by
Lp = √Dp τp , (1.7)
where τn and τp are recombination lifetimes, and Dn and Dp are diffusion coefficients of electrons and holes in the p side and n side, respectively. The properties of the p-n junction can be tuned by applying an external bias.
1.3.1 Mechanism and characteristic curve for p-n junctions under bias
When a p-n junction is forward biased, a negative voltage is applied to the n-type material, a positive voltage is applied to the p-type material and the depletion region width starts to decrease. As a result, the built-in potential (qVbi) across the space charge region decreases (qVbi-VA), as shown in figure 1.4 (a). When the external voltage (VA) becomes higher than the potential barrier at the junction, current starts to flow. This characteristic voltage is called the “knee voltage” or “kick-in voltage”. For example, in silicon p-n junctions, the knee voltage is 0.7 V, for germanium (Ge) it is 0.3 V and it is
1.2 V for gallium arsenide (GaAs). The fermi energy level of p and n type semiconductor at biased condition are denoted by EFp and EFn respectively.
9
Figure 1. 4 Band diagram for a (a) forward and (b) reversed bias p-n junction.
However, when the p-n junction is reversed biased, where a negative voltage is applied to the p-type material and a positive voltage is applied to the n-type material, the depletion width increases due to lack of electrons and holes [25]. Thus, the built-in potential across the junction increases (qVbi+VA), as shown in Figure 1.4 (b). However, a very small current in micro-amperes or nano-amperes flows through the junction, called leakage current (Io), which is caused by minority carriers from the p and n region that are pushed through the depletion zone to the junction. Since the minority charge carriers depend on temperature, the leakage current increases with increasing temperature.
10
Figure 1. 5 Forward and reverse characteristics of a p-n junction.
The p-n diode is a p-n junction that has a variety of applications including photodiodes, solar cells, light emitting diodes, temperature sensors, rectifiers, etc. The characteristics of a p-n diode are shown in Figure 1.5. The flow of current across the p-n junction is unidirectional and is exponential, as shown in Figure 1.5. The current through the p-n diode (ID) is given by
qv I = I exp[( ) − 1] , D o nkT (1.8) where ID is the current through the p-n diode, Io is the leakage current, q is the electronic charge (1.9 10-19 C), k is Boltzmann’s constant (1.38 10-23 J/K), T is the absolute temperature and n is the ideality factor that measures how closely the diode follows the ideal behavior and gives information on its recombination mechanism. A solar cell used
11 for the conversion of light energy into usable electrical energy is actually a p-n junction with a large surface area. A detailed description of a solar cell’s structure and its characterization is provided in chapter 2.
12
Chapter 2 Solar Cells: Structures and Characterizations
2.1 Solar or Photovoltaic Cell
A solar cell is a device consisting of a semiconductor p–n junction, which converts photon energy into usable electrical energy. The working principle of a solar cell is the same as that of a semiconductor diode [34]. Different types of materials with different solar cell structures have been used since the first solar cell was built by Charles
Fritts in 1883 [13]. Today, solar cells are broadly classified into different categories as shown in Table 2.1:
Table 2. 1 Types of different generations of solar cells
Peak conversion Photovoltaic Types Cost efficiency(approx.) 1st Generation Mono C-Si 26.1% Low Poly C-Si 22.3% Low
2nd a-Si 10.2% Low Generation μc or nc-Si 12% Low
CIGS 22.9% Low
CdTe 21.04% Low
Group III-V 28.8% High
DSSC 12% Low
3rd Polymer 12.3% Low Generation Oligomer 10% Low
Perovskites 22.7% Low solar cells
The above table shows different generations of solar cells that have been fabricated with their peak conversion efficiency [22, 35-38] . The first generation of
13 solar cells consist of bulk solar cells known as monocrystalline and polycrystalline cells.
The second and third generations consist of solar cells that use thin film technology.
Monocrystalline solar cells are manufactured via the Czochralski method, which involves chamfering and sawing single crystals [39, 40]. In the case of the monocrystalline silicon, the silicon crystal seed is dipped into a vat of molten silicon, and the seed is slowly pulled from molten silicon to form a large single crystal. This block of silicon is called an ingot. The ingot is then sliced into uniform wafers of different thickness. The efficiency of monocrystalline silicon solar cells is around 25%.
Polycrystalline solar cells consist of crystals with different grain orientations. In the case of polycrystalline silicon solar cells, the silicon crystal seed is placed in a vat of molten silicon and then allowed to cool. Polycrystalline solar cells have lower efficiencies in comparison to monocrystalline silicon ones, but their advantage is a lower price. The efficiency of polycrystalline silicon ranges from 20% to 22% [41].
Thin-film solar cells are fabricated by depositing one or more thin semiconductor layer(s) on a substrate [42] or transferring a completed solar cell to a flexible substrate
[43]. In such a type of solar cell, the thickness of the layer varies from nanometers to micrometers. Cadmium telluride (CdTe) [44], copper indium gallium arsenide (CIGS)
[19], amorphous silicon (a-Si) [45], organic photovoltaic (OPV) [16, 46], dye-synthesized solar cells (DSSC) [47], and gallium arsenide (GaAs) are a few of the thin-film solar cells. These thin film solar cells can be made in bendable forms. The efficiency of single junction a-Si, CdTe, CIGS, OPV, DSSC, and GaAs are 10%, 21%, 22%, 12%, 10%–
14
12%, and 28.8%, respectively [48]. The highest (world record) efficiency was observed for a single junction GaAs due to its advantageous properties [34, 49].
Different structures of solar cells can be designed, modeled, and characterized in order to obtain the best performance. In this dissertation, the focus was on the GaAs thin film solar cell due to its various advantages over other thin films. Despite its high efficiency, its manufacturing costs are very high. Low-cost substrates, high throughput, roll-to-roll manufacturing technique can be applied in order to reduce the costs. The details of these will be discussed in chapter 4.
2.2 Solar Cell Structure
A typical p-n junction (single junction) solar cell structure is shown in Figure 2.1.
It consists of a p-n junction as a building block. The p-n junction can be fabricated using the same types of semiconductor materials (homo-junction) or different types (hetero- junction) [50]. Between the p-n counterparts, one of them, called an emitter, is thinner and doped at a higher level than the other one. The other one, called a base, is thicker and lightly doped and absorbs most of the light. Thus, a solar cell can be constructed either as an n or p base. Figure 2.1 (panels a, b and c, d) shows a solar cell with n- and p-base structures, respectively. The choice depends on the properties of the n- and p-type materials comprising the solar cell. For example, in the case of silicon solar cells, the n- up structure with p-type bases is favored, as the n-type silicon has a higher surface quality than the p-type silicon. However, the n-base (p-up structure) silicon solar cells are also in practice as the n-type base has a higher tolerance to transition metal impurities than p- bases [51-53].
15
For the GaAs solar cell, the mobility of the electrons (minority charge carrier in the p side) is higher than that of the hole; thus, the p-base is preferable [54, 55] as the base has to be thicker to absorb more light. However, for a high quality thin film GaAs solar cell, an n-base has also been used [56]. In order to conduct the extracted charge carrier from base to the contact via the emitter without any resistive loss, the emitter is prepared to be thinner and more highly doped than the base. Very high levels of doping might reduce the quality of the material due to a recombination of charge carriers [57] .
Figure 2. 1 Structure of p-n junction (single junction) solar cells.
In order to mitigate the surface recombination at the front and back end of the p-n junction, a surface field layer is deposited and is commonly called a back-surface field
(BSF) at the rear or back end and a front-surface field (FSF) at the front end. Silicon solar cells with highly-doped n and p layers are grown on top of n- and p-type semiconductors
[58, 59]. Kaminski et al. (2002) used aluminum as a BSF in silicon solar cells [60].
16
For GaAs solar cells, either aluminum or indium gallium arsenide (AlGaAs or
InGaAs) is used for back and front surface fields [61]. The front surface field is also called the window layer. The band gap of this layer is larger than that of the p-n junction, so as to minimize absorption in the top layer. Since the surface field is the interface between high and low-doped semiconductor materials, an electric field formed at the interface introduces barriers for minority charge carriers, thus reducing the leakage current as shown in figure 2.2.
Figure 2. 2 Band diagram for p-n junction with window layer.
A suitable contact, called an ohmic contact, is required to extract current to the external circuit [62]. A good contact is defined as one that interacts with the majority of charge carriers (allows majority charge carriers to flow without any resistance). For an ideal contact to occur, the contact resistance should be negligible. Practically, the resistance between the contact and the semiconductor layer is never zero. Resistance is created between the deposited front and back contacts and the semiconductor in the solar cell. In order to reduce the contact resistance in the solar cell, a highly-doped layer is grown before contact deposition. Attention should be taken not to grow very-highly-
17 doped regions below the contact because this action might led to diffusion of dopant from semiconductor to the contact.
The solar cells can be grown on a conducting or a non-conducting buffer.
Depending on the type of buffer, solar cell contacts can be designed to fit the selected buffer. Figure 2.1 (a, c) shows solar cell structures with n and p bases on a conducting buffer, whereas, Figure (b, d) shows solar cell structures with n and p bases on a non- conducting buffer. Solar cells can also be constructed as double and triple junctions in order to maximize solar cell absorption, which will be discussed in a subsequent chapter.
In research work conducted at our facility, most of the solar cell devices that were grown in this configuration were n-up devices with a p-base. In most of the cases, the experiments were done using non-conducting buffers. However, both conducting and non-conducting buffers were used in order to optimize growth conditions and fabricate single-junction GaAs solar cells.
2.3 Solar Cell Characteristics
The photovoltaic solar cell is based on three important principles: a) excitation of the free mobile charge carrier due to light absorption; b) separation of charge carriers, such as holes and electron; and c) collection of charge carriers. Most of the solar spectrum is absorbed by the absorption layer (base layer) of the solar cell. When light radiation of suitable energy is used, excitation of mobile charge carriers occurs, which creates electrons and holes. These electrons and holes then move to their respective electrodes. The majority charge carriers can easily pass through the external circuit while the minority charge carriers undergo phenomena such as drift and diffusion, before they
18 are collected. As explained in chapter 1, electron drift occurs due to the built-in electrical field at the junction, and diffusion is due to the concentration gradient. Also, these charge carriers have to overcome resistance before collection. Two basic type of resistance are series and shunt (RSH). Series resistance (RS) in solar cells occurs for three reasons: a) resistance during drift and diffusion of charge carriers; b) resistance between semiconductor and metal contacts; and c) resistance of the metal contact itself, whereas, shunt resistance is due to damaged edges or defects in semiconductor material and contact diffusion [63]. Details concerning the equivalent circuit of a solar cell with resistance, solar cell parameters, and resistance effects on solar cell performance are explained below.
2.3.1 Equivalent circuit for solar cells
Based on the above facts, a solar cell can be modeled with a diode in parallel with the light source consisting of two basic types of resistance as shown in figure 2.3
Figure 2. 3 Equivalent circuit of a solar cell. in which Iph is the photon-generated current, ID is current through the diode, RS is the series resistance, Rsh is the shunt resistance, Ish is the current loss due to shunt resistance,
I is the total current through the circuit, and V is the voltage across the device. 19
From Kirchhoff’s current law:
I = Iph − ID − Ish. (2.1)
From the Shockley equation for an ideal diode, the diode current is given by the equation:
V + IRs ID = Io [exp ( ) − 1], (2.2) nVT in which Io is called the saturation or leakage current, n is the ideality factor of the diode, and VT is the thermal voltage given by the equation:
kT V = = 0.026 eV, (2.3) T q in which k is Boltzmann’s constant, q is the electronic charge, and T is the absolute temperature. The shunt current can be obtained from Kirchhoff’s voltage law using the equation:
IRs + V − IshRsh = 0, (2.4) thus, the shunt current is given by the equation:
V + IRs Ish = . (2.5) Rsh
From equations 2.1, 2.2, 2.3, and 2.5:
V+IRs V + IRs I = I − I [exp[q ( ) ] − 1] − . (2.6) ph o nkT Rsh
For ideal solar cells, the series resistance (Rs) is very low, and the shunt resistance is infinitely large. The last term in above equation can be neglected, and the above equation 2.6 can be written:
20
qV I = I − I [exp ( ) − 1 ]. (2.7) ph o nkT
Also, the value of (qV/nkT) is ≥ 1, so the equation can be written:
qV I = I − I exp ( ). (2.8) ph o nkT
For ideal solar cells in the dark, (Iph ~ 0):
qV I = I exp ( ). (2.9) dark o nkT
In order to reduce the effects of area on the current through the solar cells, the current can be expressed in term of current density (J), which is current per unit area of solar cell.
Equation 2.9 can be expressed as
qV J = J exp ( ). (2.10) dark o nkT
Equations 2.8 and 2.9 are fundamental equations for calculating current and voltage in solar cells under illuminated and dark conditions. In all of our calculation used for solar cell current density measurement, the preceding equation is considered, which will be discussed in detail in chapter 4.
2.4 Solar cell parameters used for measurements
As explained in equation 2.8, the solar cell under illumination is given by the equation:
qV I = I − I exp ( ). (2.11) ph o nkT
Equation 2.11 in terms of current density (current per unit area) can be written:
21
qV J = J − J exp ( ). (2.12) ph o nkT
Plotting the above equation gives the current density (J) , voltage (V) characteristic curve as shown in figure 2.4.
Figure 2. 4 Current density (J) – Voltage (V) curve (red plot) of solar cell with power curve (green plot).
The graph in figure 2.4 shows the current–voltage curve under dark (no illumination) condition. The I–V curve under conditions of no illumination and illumination follows the same trend except for the current at zero voltage. The parameter obtained from the I–V curve is described below.
2.4.1 Short circuit current density (Jsc)
The maximum current through the solar cell under short circuit conditions (at zero voltage) is called the short circuit current denoted by Isc or Jsc as indicated in figure 2.4.
Short circuit current density in solar cells depends on the absorption co-efficient (higher
22 is better), minority charge carrier life time (longer is better), spectrum of incident light
(broad spectrum range is better), intensity of light/temperature (increases with light intensity and slight increase with temperature), and defects in semiconductor material
(higher with fewer defects). The short circuit current density under no illumination is given by equation 2.10:
qV J = J exp ( ). (2.13) dark o nkT
Under short circuit conditions, V= 0 for an ideal diode:
JSC = Jph . (2.14)
The equation for short circuit density in solar cells with perfectly passivated and uniform generation is given by the equation:
JSC = qG (Ln + Lp), (2.15) in which G is the carrier generation rate and Ln and Lp are diffusion lengths of electrons and holes, respectively [64].
2.4.2 Open circuit voltage (Voc)
The maximum measured voltage across the solar cell at zero current is called the open circuit voltage and is denoted by Voc. The open circuit in a solar cell decreases with the defect or any other recombination (surface or bulk) in solar cells [65, 66]. It is also affected by temperature; Voc decreases with temperature due to an increase in leakage current. From equation 2.12, the open circuit voltage at J=0, is based on the equations:
23
nkT Jph V = ln [ ( ) + 1] and (2.16) oc q Jo
nkT Jph Voc~ ln ( ). (2.17) q Jo
Thus, from the above equation, it can be seen that the open circuit voltage for a solar cell decreases with the leakage current (Jo).
2.4.3 Leakage current (Io)
Leakage current, also called dark saturation current, is an undesirable current due to recombination that is denoted by Io. The source of leakage currents are un-passivated surfaces, grain boundaries, and/or defects originated in bulk, surface, grain boundaries, or grain itself [67]. The lower the leakage in a solar cell, the better is its performance.
Leakage current for solar cells is never zero and has some absolute value even for the most passivated solar cells.
Due to the leakage current, the operating voltage of a solar cell is always lower than its band gap. As an example, for a silicon solar cell with band gap of 1.1eV, the operating voltage is only 0.7 or 0.8 V considering an ideality factor of n ≤1, leakage
-12 current =10 A, and photo current (JPh)=30mA. The open circuit voltage as given by equation 2.18:
kT Jph 30mA Voc = ln ( ) = 0.026V ∗ ln ( −12 ) = 0.63V. (2.18) q Jo 10 A
Thus, the open circuit voltage in such condition is only 0.63 V, which is much lower than the band gap of the material due to the leakage current from various sources, such as defects, sidewalls, and other factors. In order to obtain the maximum output voltage in solar cells, the leakage current should be kept as small as possible. 24
2.4.4 Maximum power point (Pmax)
The point at which the solar cell should be operated to give maximum output power as shown by the green curve in figure 2.4 is called the maximum power point and is denoted by Pmax. The current and voltage corresponding to maximum power are called
Imax and Vmax, respectively.
2.4.5 Fill Factor (FF)
As shown in the figure 2.4, fill factor is defined as the squareness of the current voltage curve which is the ratio of the area of blue rectangle to the area of golden rectangle in figure 2.4. It is the ratio of maximum power from the solar cell to the product of Voc and Isc. Fill factor of the solar cell is affected by series and shunt resistances, leakage current, and recombination in the space charge region of the solar cell [68]. It is expressed as
Maximum Power point V ∗ J FF = = max max . (2.19) Voc ∗ Jsc Voc ∗ Jsc
2.4.6 Efficiency
The efficiency of solar cells refers to the conversion of solar energy into electrical power. It is used to compare the performances of different solar cells. The efficiency of a solar cell is defined as the ratio of maximum power delivered by the solar cell to the incident or input power entering the solar cell. Thus, the efficiency of solar cell can be expressed as
25
Maximum Output Power FF ∗ V ∗ J Efficiency = = oc sc . (2.20) Input Power Pin
The input power is the power due to solar radiation under standard conditions
(standard intensity and temperature). The standardized conditions for measurement of input power is defined in terms of air mass (AM) at a particular temperature. A brief description of air mass is given below.
2.4.6.1 Air Mass (AM)
Air mass (AM) measures the atmospheric distance that solar radiation has to travel before reaching the surface of the earth [69]. It depends on the angle of inclination of the sun and the position at which the solar cells are present (such as altitude and latitude). For example, at sea level, the AM value varies according to the angle of inclination and is minimum at an inclination of 90o and is called AM1 with the highest intensity. The value of AM at the same sea level increases with the inclination angle, thus causing a decrease in the intensity of light radiation.
The value of AM at high altitudes decreases compared to that at sea level at the same angle of inclination. For example at 90o inclination, the AM at an altitude of 1 mile above sea level is approximately 0.827, compared to 1.0 at sea level [70]. The above two examples are described for a fixed latitude. However, if the latitude of a site varies (with constant altitude and solar inclination), the angle of inclination of solar zenith also varies and hence, so does the AM.
26
Thus, AM is defined as the path length that solar radiation travels through the atmosphere at a given site comparable to the solar radiation when sun is at its zenith. The
AM is expressed according to equation:
1 L AM = = , (2.21) cosθ Lo in which 휃 is the angle from the vertical, L is the path length at zenith, and Lo is the path length through the atmosphere. Terrestrial solar cells are measured under AM1.5 conditions at a temperature of 25oC in order to maintain a single standard at the mid- altitude region, which corresponds to a solar zenith angle 휃=48.2o. The maximum intensity of sunlight under these conditions is 1000 W/m2. The above equation for AM is based on the assumption that the atmosphere is a flat horizontal layer. The above equation should be modified if curvature or spherical models of the earth are taken into consideration.
In the experiment conducted for my research, the standardized AM1.5 solar simulator with an intensity equal to 1000 W/m2 was used. The details of this experiment will be explained in chapter 3.
2.5 Series and shunt resistance in solar cell
No solar cell is perfect and is comprised of series and shunt resistances as previously explained. The effect of series resistance on solar cell characteristics is shown in figure 2.5. As shown in the I–V curve, an increase in series resistance causes a decrease in the fill factor of a solar cell and, at very high values of series resistance, the short circuit current is reduced.
27
Figure 2. 5 Impact of series resistance on I–V characteristics of a solar cell.
The impact of shunt resistance on I–V characteristics of a solar cell is shown in
Figure 2.6. An ideal solar cell is considered to have infinite shunt resistance. As the shunt resistance decreases, the fill factor of a solar cell decreases. At very low values of shunt resistance, the open circuit voltage of a solar cell decreases.
Figure 2. 6 Impact of shunt resistance on current voltage curve of solar cell.
Beyond the above-described parameters affecting solar cell efficiency, solar cells are also impacted by different types of losses. The loss mechanisms in a solar cell are explained below.
28
2.6 Loss mechanism in solar cells
Only a small part of the solar spectrum is converted into electrical power in solar cell devices. The majority of energy loss occurs in different ways [71, 72]. The major types of losses in solar cells are outlined below.
2.6.1 Optical Loss
Optical loss can be classified into absorption and reflection losses in solar cells.
They reduce the short circuit current by reducing the net power obtained from solar cells.
2.6.1.1 Absorption loss
The electromagnetic radiation from the solar spectrum is composed of a range of different wavelengths of light. However, the photons of certain energies have higher intensity than others. At a given wavelength, we can calculate how much power is coming from all of the photons received per unit time. This power is called spectral irradiance:
ϕE Spectral Irradiance = , (2.18) Δλ in which ϕ is the photon flux given by the equation:
Number of photons ϕ = , (2.19) s. m2
E is the energy of photon given by the equation:
29
hc E = , (2.20) λ in which h is the Planck constant, c is velocity of light, and λ is the wavelength of radiation.
The spectral irradiance distribution of solar radiation with wavelengths at sea level and AM1.5 is shown in figure 2.7. The solar spectrum is close to that of a black body at 5778K, which is indicated by the blue curve in the diagram [73].
Figure 2. 7 Spectral Irradiance for solar spectrum (AM1.5, 1000 W/m2).
Figure 2.7 shows that the solar intensity is maximum in the visible region (green- shaded area) and lower along the ultra-violet and near-infrared regions. The fluctuations in the spectrum arise from the absorption of the energy of photons by gases in air, including water vapor (H2O), carbon dioxide (CO2), ozone (O3), and other greenhouse gases. The above graph shows that not all the photons from solar radiation are absorbed by solar cell material. Thus, solar materials should be effectively chosen in order to absorb most of the light spectrum.
30
Figure 2. 8 Optical losses in solar cell.
As shown in Figure 2.8, the energy of the photons that is less than that of the band gap of a solar cell simply pass through the solar cell. Also, the photons with energy greater than the band gap lose their extra energy, a process called thermalization loss.
This prompts us to carefully choose the semiconductor material with an optimal band gap
(not too high or low). Furthermore, we can also design solar cells (multi junction solar cell) in a way so as to absorb all of the wavelengths or energy of a light beam. Therefore, in order to design perfect solar cells, all of the loss factors should be taken into account and resolved.
2.6.1.2 Reflection Loss
The refractive index difference between different layers in a solar cell structure causes reflection of incident light. Reflection from any two surface is given by the equation:
푛1 − n2 Reflectance(R) = [ ]2 , (2.21) 푛1 + n2 in which n1 is the refractive index of air, and n2 is the refractive index of the surface on which light is incident. For example, in the case of a silicon solar cell, n1=1 for air and
31 n2= 3.6 (approximately for green light). Therefore, the reflection from silicon surfaces when light passes from air to silicon is expressed by the equation:
1 − 3.6 2 Reflectance(R) = [ ] = 32%. (2.22) 1 + 3.6
From equation 2.22, it is clear that 32% of the light is reflected from the surface of silicon, which is not desirable. The front surface of a silicon solar cell is coated with antireflection coating (ARC) in order to minimize the reflection. Also, the thickness of the ARC material is important. The thickness of this ARC material can be calculated according to the equation:
λ (2.23) Thickness (T ) = , ARC 4n in which n is the refractive index of intermediate layer used as anti-reflection coating expressed according to the equation:
n = √n1n2 = √1 ∗ 3.6 = 1.89. (2.24)
Thus, for silicon solar cells, the anti-reflection coating material should have a refractive index in the range as shown above. The materials used for anti-reflection coatings include silicon dioxide (SiO2), silicon nitride (Si3N4), titanium dioxide (TiO2), zinc sulfide (ZnS), and magnesium fluoride (MgF2) or combinations of them in order to match the refractive index [74, 75]. The required thicknesses of these antireflection coatings can be calculated using the equation:
32
λ 560 (2.25) Thichness (T ) = = = 74 nm. ARC 4n 7.56
Thus, the approximate thickness for anti-reflection material is around 74 nm. In some cases, two different material layers of desired thickness are used as anti-reflection coatings. However, the thickness of the material and its refractive index should be chosen precisely. For single junction GaAs solar cells fabricated at our facility, double layer antireflection coatings composed of ZnS/48 nm and MgF2 /96 nm are used. Details of the deposition process can be found in chapter 4.
Also, reflection losses can be minimized using a back reflector, such as silver, and texturing the solar cells [76]. Texturing can be a pyramid with bowl-like grooves in a micro-scale regular geometric pattern that can change the direction of light in order to yield total internal reflection [77].
2.6.2 Recombination loss
Recombination of photo-generated charge carriers occurs due to defective semiconductor states, small diffusion carrier lengths, and lower carrier lifetimes [78].
Defects in semiconductors may be surface or bulk defects. Thus, the carrier dies before being collected at its respective electrode due to the defects present in semiconductor material. The mechanism by which recombination in solar cell takes place can be expressed:
a) Direct recombination,
b) Indirect or trap-assisted recombination, and
c) Auger recombination.
33
Figure 2. 9 Recombination mechanism in solar cell (a) direct (b) trap-assisted or indirect and (c) Auger recombination.
2.6.2.1 Direct recombination
In the case of direct recombination, free electrons combine with free holes resulting in photon emission with an energy equal to the potential energy of the free electrons (band gap of semiconductor). Thus, this recombination is also called radiative recombination. Radiative recombination occurs mainly in direct band gap semiconductors, such as gallium arsenide, amorphous-silicon, and others. For indirect band gap semiconductors, it is very unlikely that this process occurs.
2.6.2.2 Auger recombination
Auger recombination occurs when a free electron recombines with a free hole, which results in an intra-band free carrier transition. This excited carrier ultimately thermalizes back to the bottom of the conduction band and top of the valence band.
34
2.6.2.3 Trap-assisted recombination
Trap-assisted recombination occurs when free electrons and holes both are captured in a state deep inside the bandgap. Both electrons and holes are involved in such recombinations, and each transition results in photon emission in order to satisfy energy and momentum conservation. This occurs in both direct and indirect semiconductor materials. The recombination is also called Shockley–Red–Hall (SRH) recombination
[79, 80] in honor of William Shockley, William Thornton Read and Robert N. Hall.
2.7 Conclusion
Different types of solar cells are made with different designs. Solar cells suffer from different losses such as optical, spectral, resistive, and recombination. These losses for single junction solar cells limit the theoretical efficiency to around 30% at 1.1 eV and
33.7% at 1.34 eV (AM1.5 solar spectrum), also called the Shockley–Queisser limit [81,
82]. A multi-junction solar cell can overcome the Shockley–Queisser limit [83]. Single and multi-junction solar cells can be constructed using different type of semiconductor materials, such as silicon (Si), germanium (Ge), perovskite, gallium arsenide (GaAs), and others.
Despite of all the above mentioned losses, the highest efficiency observed for single and multi-junction solar cells has been achieved with the use of III-V semiconductor materials [84, 85]. The details of III-V semiconductor materials and their uses in photovoltaic applications are explained in chapter 3.
35
Chapter 3 III-V semiconductor material and solar cells fabrication and
characterization
3.1 Introduction
III-V semiconductor materials correspond to group III-boron, aluminum, gallium, indium and group V-nitrogen, phosphorous, arsenic and antimony. The combination of these group III and group V elements form a binary semiconductor compound (e.g.,
GaAs InAs, Gap, InP etc.), ternary compound (e.g., InGaAs, InGaP, AlGaAs, etc.) or quaternary compound (e.g., GaInAsP, InGaNAs). The most used combinations with these
III-V materials are AlxGa1-xAs, InxGa1-xAs and AlxIn1-xAs, where x ranges from 0 to 1
[86].
Figure 3. 1 Composition vs lattice constant for III-V materials.
Illustrated in figure 3.1, GaAs has a perfect lattice match with AlAs at any graded composition [87]. Small amounts of In added to GaAs can allow lattice matching to
GaAs and such a material can grow without a strain relaxation or an associated need for defect reduction procedures. Growth of III-V compound semiconductor materials has
36 progressed from wafer to thin film growth. This growth can be either homogeneous
(same material) or heterogeneous (different material) [88]. Some growth methods for III-
V semiconductors are described below.
3.2 Growth Methods of III-V Semiconductor
3.2.1 Liquid-encapsulated Czochralski (LEC) Method
In this method, a single crystal seed of known orientation is placed in a melt of a material of the same composition in a quartz crucible that is heated via a graphite susceptor and is slowly pulled from the melt, which produces a single crystal ingot or boule [89, 90]. The shape of the crystal is controlled by adjusting the heating power
(using RF coil) and the pulling and rotation rates of the seed crystal. The vapor pressure of arsinic (As) is higher than gallium (Ga) at high temperatures. Thus, during growth of
GaAs, a boron trioxide (B2O3) capped layer is used to prevent evaporation of As at high temperatures and above atmospheric pressure. A schematic of the process is shown in figure 3.2.
Figure 3. 2 Schematic of liquid-encapsulated Czochralski method. 37
3.2.2 Float Zone and Bridgman
In this process, the melt is moved away from the seed such that recrystallization occurs behind the moving heater and melt zone. This basically creates a two-zone furnace, one zone with a higher temperature (hot zone) than the other (cold) zone; the zones are separated by a muffle as shown in figure 3.3, to control the temperature gradient. A benefit of this method is that impurities expelled from the solidified material are segregated further from the seed to obtain a high-purity crystal. For GaAs growth, precursor GaAs material (GaAs melt) is kept in the higher temperature zone at its melting point under arsenic overpressure. The seed crystal is kept at a point such that one of its ends melts and is slowly moved toward the low temperature zone, which produces single crystal GaAs.
Figure 3. 3 Schematic of Float Zone and Bridgman process.
3.2.3 Liquid Phase Epitaxy (LPE)
This is the solution growth technique for the deposition of a crystallographically- oriented thin film on a seeding or template substrate from a molten solution.. The molten solution is contained in a graphite or quartz boat that is placed in a quartz reactor tube at a 38 high temperature and hydrogen flow levels. The deposited film follows the same a crystallographic orientation as the substrate by epitaxial growth.
Figure 3. 4 Photograph of Liquid Phase Epitaxy system at our facility.
3.2.4 Molecular Beam Epitaxy
Molecular beam epitaxy is an evaporation technique where atoms or molecules are deposited on a substrate at a desired temperature under ultra-high vacuum conditions
[91-93]. The precursor solids are placed in graphite Knudsen cells or effusion cells, which are radiatively heated until they reach their sublimation point to emit molecular beams that are aligned toward the substrate heated by RF power. Each Knudsen cell uses a shutter to close and open the molecular beam. The composition of the film depends on the temperature of the film, angle at which the beam impinges, flux ratio of individual components that reach the substrate etc. Thin films of different elements can be grown layer by layer and characterized in situ. To characterize the sample during growth, reflection high energy electron diffraction (RHEED) pattern is used. This pattern helps to reveal the quality of the deposited film and its nature (amorphous or crystalline). A mass
39 spectrometer is used to monitor the rate of multiple molecular beams. A schematic of the
MBE process is shown in figure 3.5.
Figure 3. 5 Schematic of MBE process.
3.2.5 Metal Organic Chemical Vapor Deposition (MOCVD)
MOCVD is a chemical vapor deposition technique in which metal organic precursor vapors with dopant species and reactant gases are used to grow epitaxial films
[94-98]. The desired substrate or wafer is kept at a high temperature inside a reactor and the precursor vapors and reactant gases are impinged on the wafer through a showerhead.
High purity gases (e.g., high purity hydrogen passed through palladium purifier) are used as carrier gases. The chemical reaction takes place on the surface of wafer, thus forming a monolayer of the desired compound. For example, trimethlygallium (TMGa) or triethylgallium (TEGa), arsine (AsH3) with hydrogen as the carrier gas undergo the following chemical reaction at substrate temperature around 700oC to form epitaxial
GaAs,
Ga (CH3)3 + AsH3 → GaAs + 3 CH4 . (3.1)
40
Figure 3. 6 Photograph of MOCVD Chamber at University of Houston.
The III-V material grown by MOCVD can be used to fabricate different devices like light emitting diodes (LED), thin film transistors, lasers etc. MOCVD is mostly used to grow III-V materials due to its relatively low cost compared to other techniques; also the growth by MOCVD can be controlled well. A photograph of the MOCVD system in our facility for roll to roll and stationary processes is shown in figure 3.6. This technique is used to grow epitaxial stacks of different III-V materials with the desired level of doping to create a single- or multi-junction solar cell structure.
The most commonly used single-junction III-V semiconductor material for solar cell applications is GaAs. Some of the qualities of GaAs that are advantageous over silicon are:
a) The saturation electron velocity (Vd) is higher than that of silicon, which results in
higher mobility and is given as
Vd = µE , (3.2) where Vd is the drift velocity of electron, µ is the mobility of electron and E is the applied electric field.
41 b) It is a higher band gap material, which makes it less sensitive to high temperature.
Hence, it can be used in outer space applications and as an optical window in high-
power applications. c) The temperature coefficient of GaAs is very low i.e. degradation of solar cell
efficiency at higher temperatures is much less than that of silicon. d) The direct band gap helps to absorb and emit light efficiently making it suitable for
opto-electronic properties. e) The absorption co-efficient of GaAs is 10 times higher than silicon. So, it can be
grown very thin (in the nanometer-scale range) adding the advantage of flexibility
and light weight.
Due to the favorable features GaAs, it has been used to fabricate high efficiency solar cells.
3.3 III-V Single-junction Gallium Arsenide Solar Cells Structure and Design
The structure of single-junction (1J) GaAs solar cell fabricated in this work is shown in figure 3.7.
(
Figure 3. 7 Single-junction GaAs solar cells structure on (a) Conducting buffer (b) Non conducting buffer.
42
Single-junction GaAs solar cells can be of either p-n or n-p geometry. Highly- doped n++ or p++ type GaAs is grown for ohmic contact in the solar cell. For a conducting buffer, the bottom contact can be made at the back of the solar cell, whereas for a non- conducting buffer, a front contact geometry is preferable for both top and bottom contact.
Aluminum gallium arsenide (AlGaAs) or indium gallium arsenide (InGaAs) is used as a window (front surface field or FSF) and back surface field (BSF). AlGaAs or InGaAs has a higher bandgap than GaAs, so it can be used as a window layer which allows the photon to pass through it into the base for absorption. They also block the back flow of electrons and hole as explained in chapter 2.2.
Figure 3. 8 Ohmic contact formation for n type (a), (b) and p type (c), (d) semiconductor.
Ohmic metal contact is required for solar cells, which allows free movement of charge carriers, which is illustrated in figure 3.8. Electron states at metal-semiconductor interfaces cause band bending in the conduction and valence bands near the interface. For n type contact, the work function of metal (Φm) should be less than that of the
43 semiconductor and for p-type contact, the opposite is true. Nickel (Ni)/Germanium (Ge)/
Gold(Au) is the dominant contact metallurgy for n-type ohmic contact in single-junction
GaAs solar cells [99-102]. Other types of contacts are Pd/Ge, Ni/Ge, Ta/Ge and Mo/Ge that [103].
Nickel is used as a wetting layer and has various advantages: it avoids forming a surface oxide because it reliably reacts with different types of surfaces; avoids possible formation of the gold-germanium complex that might arise if only Au/Ge is present
(spiking effect), it forms of Ni-GaAs complexes that drive the diffusion process of Ge-Au to form better contact and improve the uniformity of the contacts.
For good ohmic contacts, it is necessary for Ge to diffuse along the interstitial sites until Ge finds a vacancy in the GaAs lattice, which requires very high temperatures
(600oC). This is facilitated at low temperatures (100oC) by Ni-GaAs complexes [101,
104]. Au on the other hand, forms a stable intermetallic reaction product, Au-Ga, which facilitates the diffusion by disturbing the GaAs crystal lattice. Ni on GaAs also facilitates formation of Au-Ga, which plays an important role in the formation of the ohmic contact.
Sometimes, thermal annealing might be useful to get a better ohmic contact after the deposition of Ni/Au/Ge. In this work on fabricating a single-junction GaAs solar cell, Ni
(5nm)/Ge (10nm) and Au (250nm) were used. Details of the device fabrication are provided in chapter 4.
3.4 III-V Multi -junction Solar Cell Structure and Design
Multi-junction or Tandem solar cell architecture comprises two or more solar cells, combined to harvest more sunlight and increase solar cell efficiency [105, 106].
44
The top absorber has the highest band gap, the bottom with the lowest band gap and the middle one with an intermediate band gap. While designing the tandem solar cell, there are important considerations. The top absorber layer must be transparent for the low energy photon to pass to the layers beneath. Secondly, a tunnel junction should be applied to eliminate mismatch between different absorber layers.
Figure 3. 9 Triple-junction solar cell (Monolithic structure).
Two basic designs of tandem solar cell are monolithic and the stacked tandem
[107-109]. In a monolithic structure, one absorber layer is grown over the other in a continuous fashion (series connection) as shown in figure 3.9. A spectral mismatch between the two absorber layers is likely to occur and the domination of the lower one reduces the overall efficiency of the solar cell. A tunnel junction is used to assemble the monolithic stack. A very narrow space charge region is created due to the interactions between these layers through which tunneling of electrons and holes takes place. It is constructed by the interface of highly doped p++ and n++ layers which is illustrated in
45
Figure 3.9. Whereas, on the other hand in the stacked tandem, one absorber layer is glued to the other. Higher efficiency tandem solar cells are made using crystalline III-V materials where the flexibility in band gap is obtained by tuning the elemental composition.
The highest efficiency has been obtained in III-V single-junction and multi- junction solar cells. For a single junction III-V GaAs solar cell, the world record efficiency is 28.9% at AM1.5G, fabricated by Alta Devices [43, 110]. Whereas, for multi junction solar cells (3 junction), an efficiency of 37.5% has been obtained at AM1.5G, and 44% at 947suns (concentrated light) has been achieved. An efficiency of 44.7% at
297suns has been demonstrated in a 4-junction III-V solar cell fabricated by Fraunhofer
Institute [111]. The chart of the highest efficiencies of research solar cell, maintained by
NREL, is shown below.
Figure 3. 10 NREL Best Research Cell Efficiency.
46
Optimization of defect, doping, design etc. is required to achieve high efficiency in III-V solar cells. For this reason, different characterization tools are required to analyze
III-V semiconductors and single/multi-junction solar cells.
3.5 Characterization tools for characterizing III-V materials and Solar cells
After III-V materials films are grown for solar cells, they are characterized to investigate the quality of the material grown, defects and other necessary properties.
Some instruments used for characterization are listed below,
3.5.1 X-Ray Diffraction (XRD)
X-Ray diffraction is a non-destructive technique to study the elastic scattering of
X-Ray photon atoms in a periodic lattice [112-114]. This is used to identify the crystal structure, phases of various materials, residual stress, material composition, unit cell lattice parameters, orientation, crystal size etc. Diffraction occurs when the wavelength of radiation, which is comparable to atomic spacing, is scattered by the atoms in the crystal lattice with constructive interference. The phenomenon is explained by Braggs model of diffraction and is shown in figure 3.10 (a).
47
(a) (
(b) (
Figure 3. 11 (a) Schematic of X-Ray Diffraction, (b) Photograph of XRD at UoH.
When a monochromatic X-ray of wavelength ‘λ’ is incident on a material with interatomic distance ‘d’ at an angle of ‘휃’ and reflects with the same angle of scattering, then according to Braggs law of diffraction, for constructive interference, the path difference between the X-ray is equal to an integral multiple of wavelength i.e.,
2dsinθ = nλ , (3.3) where n is the integer, d is the lattice spacing and λ is the wavelength of light.
The instrument consists of some important components like the X-ray source, sample holder, goniometer, filter and the detector. The detector counts the number of x- rays diffracted by the sample. During the scan, the detector or the sample is rotated over a range of angles to detect the band of diffracted x-rays produced by the crystal in the sample. Both in-plane and out-of-plane crystal measurements can be performed using
48
XRD. A photograph of the Bruker XRD system used in this work is shown in figure 3.10
(b).
3.5.2 Hall Measurement
Hall measurement uses the principle of Hall effect observed by E.H. Hall in 1879
[115-117]. According to the Hall Effect, “an electrical potential perpendicular to the direction of current/voltage and applied magnetic field is developed in current-carrying sample, placed in such magnetic field’’. Hall measurement is used to measure the resistivity (ρ), conductivity (σ), mobility (μ) and the Hall coefficient (RH) of the sample.
It basically uses a long, narrow Hall bar geometry or nearly square/circular Van der Pauw geometry [118]. For the traditional bar-shaped sample, six contacts are needed. Four- probe Hall measurements using the van-der-pauw geometry is one of the common methods where the current is applied at two ends with voltage measurements taken at next two ends. Then, the current and voltage polarity is changed (shown in Figure 3.12
(a)) and all four resistances are measured and averaged. Additionally, temperature- dependent Hall measurements (using the same method) can further help provide information on activation energies. The HMS-5000 Hall effect measurement system
(shown in Figure 3.12 (b)) was used to measure GaAs samples fabricated using MOCVD.
49
( (a)
(b)
Figure 3. 12 (a) Schematic of the Hall Measurement method (b) Photograph of Hall measurements at the University of Houston.
The current (Ix) through the conducting sample (GaAs) is given by
Ix = −qvxAn , (3.4) where vx is called the drift velocity of charge through the conductor of area A and n is the charge carrier density.
So, current density (Jx) is given by,
Jx = −qvxn (3.5) and, the drift (vx) is given by,
50
J (3.6) v = − x . x nq
The Lorentz force is the mechanism that drives hall measurement. When a sample carrying charge ‘q’ is placed perpendicular in a magnetic field ‘B’, the force experienced is called the Lorentz force given by,
(3.7) F⃗ = −q(v⃗⃗⃗⃗x⃗ × B⃗⃗⃗⃗z⃗ ) = −qvxBz,
Where the magnetic field is applied in the Z-direction and the current is flowing in x- direction with a drift velocity, vx.
As charge is built up in the sample, the electric field or Hall voltage (EH) is developed perpendicular to both the magnetic field and direction of current. The force (FH), due to
Hall voltage (EH), is given by the equation:
FH = qEH, (3.8)
Under steady conditions, the Lorentz force balances the force due to hall voltage (force due to electric field),
|F| = |FH|. (3.9)
From equation 3.7 and 3.8,
EH (3.10) vx = − , BZ therefore,
EH = −vx Bz (3.11) and, from equation 3.10 and 3.11,
51
1 (3.12) E = − J B = R J B , H nq x z H x z where RH is called the Hall co-efficient given by
1 (3.13) R = − . H nq
The mobility of a conducting sample is given by
v µ = x . Ex (3.14)
From equation 3.10, 3.13 and 3.14,
vx jx jxRH µ = = − = . (3.15) Ex nqEx Ex
If σ is the conductivity of the sample then,
Jx = σEx , (3.16)
From equation 3.15 and 3.16,
µ = σRH . (3.17)
From the above equation 3.17, the mobility of the sample can be determined.
Since,
1 (3.18) E = − J B = R J B H nq x z H x z and
V E = H , H w (3.19) from equation 3.18 and 3.19,
52
Jx = Ix tw, (3.20) where, t is the thickness of the sample, w is the width of the sample and Ix is the current flowing through the sample. From equation 3.19 and 3.20
EH VHt (3.21) RH = = . JxBz IxBz
The above equation shows that the Hall coefficient depends on the thickness of the sample. After growth of the III-V semiconductor layer of the solar cell, each solar cell carrier concentration, conductivity and mobility is measured using the Hall measurement.
3.5.3 Photoluminescence intensity (PL)
Photoluminescence involves two processes. The first is excitation of atoms or molecules to a higher energy state by absorbing light or photons, and the second is radiation of photons while returning to a lower energy state [119, 120].
53
(a)
(b)
Figure 3. 13 (a) Schematic of Photoluminescence, (b) Photograph of PL system.
The PL spectrometer shown in the schematic in figure 3.13 (a) consists of a light source that can excite the specimen (Xenon arc lamp, laser source), beam splitter (set of mirrors) to direct the light into the desired path, an excitation monochromator, if required, in order to get specific monochromatic light, neutral density filter if the intensity of light is high, sample holder for the specimen, emission monochromator for the selection of a specific monochromatic light, a detector (photomultiplier tube), and a recorder (computer system with PL software),. A photograph of the PL system used for GaAs measurement is shown in Figure 3.13 (b). It uses a Mai Tai 2.5W laser connected to a power supply and uses a mirror to focus light on the sample via a neutral density filter, CS-260 monochromator and detector (Hamamatsu).
54
Time resolved photo-luminescence can also be performed with the same instrument with a short pulse laser and fast detector that can help to measure the lifetime and relaxation processes.
3.5.4 Secondary Ion Mass Spectroscopy (SIMS)
Secondary ion mass spectroscopy is a surface analytical technique that is capable of determining the elemental concentration with a sensitivity less than 1 part per million
(ppm) at the surface layer and around 1 part per billion (ppb) in the bulk [121]. Beams of positive (e.g., CS) ions or negative (e.g., O) ions are focused on a sample surface to produce secondary ions that are transferred by a high electrostatic potential in a vacuum environment to a mass spectrometer. A variety of mass spectroscopy techniques has been developed that can be classified as static SIMS used for sub-micrometer elemental analysis and dynamic SIMS used for analysis of the bulk to acquire compositional information. A high yield of the secondary ions is produced in dynamic SIMS mode where the primary ions beam exceed the static limit (1012 ions/cm2). Thus, SIMS is applied for elemental analysis, depth profiling, interface analysis and image analysis.
Figure 3. 14 Schematic of Secondary Ion Mass Spectroscopy
55
Different types of secondary ions mass spectroscopy includes imaging SIMS used for spatially-resolved elemental analysis, time of flight (TOF) SIMS [122] that provides elemental and molecular information about surface thin layers and the interface by measuring the time of flight from the sample surface to the detector and Helium ion microscopy (HIM) SIMS [123, 124] with high lateral resolution of around 0.5 nm.
3.5.5 Solar simulator
A solar simulator is a light source that provides an illumination spectrum that matches the sun’s spectrum [125]. It is used to test solar cells under controlled and repeatable conditions. According to its required application, its detection range (ultra violet, visible, infrared .) and intensity can be varied.
(a)
(b)
Figure 3. 15 (a) Schematic diagram of a solar simulator (b) Set up for solar cell efficiency measurements
56
A schematic diagram of a solar simulator is shown in Figure 3.15 (a). A photograph of the solar simulator and the schematic of the set up used for conversion efficiency measurements on GaAs solar cells made in this work is shown in Figure 3.15
(b). The light from the Xenon arc lamp (Oriel Sol3A class AAA solar simulator) is reflected by an ellipsoidal reflector which further reflects through a mirror to the spectral correction filter (AM filter). The transmitted light (through the filter) then passes to the integrated lens setup, a mirror, collimating lens and to the sample. The solar cell to be tested is connected to a source meter (Keithley 2400) via a probe to measure current and voltage.
3.5.6 External Quantum Efficiency (EQE)
External Quantum efficiency is the ratio of extracted charge carriers collected to the number of incident photons of a given energy on the solar cell. Ideally, EQE curve is square in shape, as shown in figure 3.16, however some wavelength are reduced in intensity due to different types of recombinations in the solar cell. EQE gives an idea of where the losses in the solar cell come from as described in Figure 3.16.
Figure 3. 16 EQE comparison of ideal solar cell with experimental.
57
EQE reveals the effect of optical losses. Sometimes QE is measured after the reflected and transmitted light have been accounted, which is called internal quantum efficiency (IQE). So, IQE tells us what the fraction of the absorbed photons are converted to electrons in the device. Focused light from the lens, via a monochromator, is incident on the sample which is connected to a trans-impedance amplifier (used to amplify the signal), semiconductor parameter analyzer and to the computer hardware for data output. A schematic of EQE is shown in figure 3.17 (a) and a photograph of the EQE system used in this work is shown in figure 3.17 (b). EQE and IQE is expressed as
electrons/sec EQE = photons/sec (3.22) and
electrons/sec EQE IQE = = . absorbed photons/sec 1 − Reflection (3.23)
58
(a)
(b)
(b)
Figure 3. 17 (a) Schematic of EQE system, (b) Photograph of EQE system used in this work.
3.5.7 Atomic Force Microscopy (AFM)
Atomic force microscopy is a type of scanning force microscopy with a resolution in the nanoscale range. This is used to measure local properties of the material, such as its roughness (topography image), height distribution (height distribution image), composition (phase image), electrical and potential distribution (Kelvin probe force microscopy design) etc. It operates via the measure of the force between the probe
(cantilever and tip) and the sample. The deflection (vertical and lateral) of the cantilever is detected by a four-segment position-sensitive photo detector (PSPD) using a laser beam. The feedback signal is provided between the PSPD and the piezo electrical scanner, which reveals three dimensional positioning of the sample. The imaging mode,
59 called the tapping mode, using an Agilent 5100 AFM was used in this work to get the
AFM image for the GaAs samples. A schematic of an AFM is shown in Figure 3.18.
Figure 3. 18 Schematic of the atomic force microscope.
3.5.8 Transmission line method (TLM)
Transmission line method is used to determine the contact resistance (Rc), sheet resistance (Rsh) and the specific contact resistance (ρc) of solar cells [126, 127]. In this method, ohmic contact pads at variable distances (d) on each contact layer (n and p) are fabricated, as shown in figure 3.19.
(
(
Figure 3. 19 Illustration of linear contact arrays for top and bottom contact.
For a semiconductor solar cell with contacts on two side, the total resistance is given by,
60
R = 2Rc + Rsemi + Rm , (3.22) where Rsemi is the resistance of the semiconductor between the two contacts, Rm is the resistance of the metal contact on two sides and Rc is the contact resistance (resistance between metal semiconductor interface).
The resistance of the metal contact is very low compared to the semiconductor and the contact resistance. So, Rm can be neglected. Equation 3.22 can be written as
R = 2Rc + Rsemi . (3.23)
The resistance of semiconductor of length ‘d’ and area ‘A’ is given by
ρd ρd ρ d d (3.24) R = = = ohms = R . semi A tz t z sh z
Where ‘z’ is the width of the contact, ‘t’ is the thickness of the semiconductor, d is the length of semiconductor between the contacts and ρ is the resistivity.
The unit of resistance is ohms (Ω), d/z has no unit, so the unit of ρ/t should be ohms. But ρ/t is not the resistance of the sample. So, to distinguish between R and ρ/t, its unit is expressed as ohms/square and is called sheet resistance (Rsh) between the contacts.
From equation 3.24.
d R (3.25) R = R + 2R = R + 2R = sh d + 2R , semi c sh z c z c plotting resistance for different values of d (d1,d2……..dn), we obtained the graph shown in figure 3.20.
61
Figure 3. 20 Distance vs resistance plot for TLM.
Compared with equation 3.25, the slope and intercept are given by
ΔR R (3.26) Slope = = sh , Δd z
so, the sheet resistance is given by
Rsh = Slope x z . (3.27)
From equation 3.25,
Intercept = 2Rc . (3.28)
From the above equation, the contact resistance (RC) can be calculated. However, the contact resistance depends on the size of the contact. It is not a good point of comparison, so a term called contact resistivity (ρc) or specific contact resistance is defined. Considered to be a small region at the contact, where resistance (Rc) is given by:
62
′ Δx (3.29) Rc = ρ , Ac where Δx, is the small region in the vicinity of the contact with a resistivity ρ and Ac is the area of the contact. At the point of contact, the contact resistivity (ρc) or specific contact resistance is described as
lim ′ ρc = Δx→0 (ρ Δx) = Rc Ac . (3.30)
The unit of contact resistivity is Ωm2 or Ωcm2. For a contact of infinite length (L), current in semiconductor flows at a finite length near the contact, called the transfer length (LT), is shown in figure 3.19, which is the average distance that an electron or hole travels in the semiconductor beneath the contact before it flows up into the contact and is given by
(3.31) ρc LT = √ , Rsh therefore, from equation 3.24,
ρc Rsh LT (3.32) Rc = = = Slope LT , LTz z
R (3.33) L = c , T Slope and from equation 3.27,
63
ρc = RcAc = RczLT . (3.34)
Also, from equation 3.25 and 3.29,
d R R L (3.35) R = R + 2R = R + 2R = sh d + 2 sh T semi c sh z c z z R = s (d + 2L ). z T
Extrapolating back to the x-axis in the graph between L and R, at R=0 we can calculate the transfer length (LT) from the intercept of graph 3.13.
Intercept = −2LT . (3.36)
Thus, we can determine the contact resistance (RC), sheet resistance (RSH) and specific contact resistance (ρc) of the solar cell using TLM.
3.6 Conclusion
The above characterization tools were used for testing the III-V semiconductor solar cells to understand the material quality, defects in material, its epitaxial nature etc.
Solar cells were fabricated after optimization of the quality of the material. The fabricated solar cells are then tested to determine the device’s characteristics as explained in chapter
4.
64
Chapter 4 Next-generation solar cells for low cost and high solar cell
efficiency
As mentioned in the previous chapter, the highest efficiency among all solar cells has been observed in III-V semiconductor materials. However, the cost of III-V materials is still high compared to other types of cells. Two basic approaches are being pursued for low cost III-V solar cells:
a) Use of low-cost substrates (other than GaAs wafers) for growth of III-V
semiconductor materials and
b) Epitaxial lift-off processes for reuse of GaAs wafer substrate.
4.1 Low-cost substrates for growth of III-V semiconductor materials
Non-GaAs substrates have been used for growth of III-V semiconductors, such as germanium (Ge), silicon (Si), or compositional graded SiGe. Single-junction GaAs solar cell efficiency >17% at AM1.5 has been reported on Ge wafers using the chemical vapor deposition (CVD) [128], molecular beam epitaxial (MBE) [129], and metal organic chemical vapor deposition (MOCVD) techniques [130]. The lattice mismatch between
GaAs/Ge is just 0.07% at room temperature and results in compressive bi-axial stress in the growth plane which is compensated with an enlargement of the Ge lattice in the direction of the growth [131]. However, the GaAs (polar)/Ge (non-polar) interface results in an antiphase boundary or disorder due to two possible phases of GaAs grown on Ge that could result from the switch of cation and anion sub-lattices [132].
65
For Si as substrate, the lattice mismatch between GaAs/Si of approximately 4% as well as differences in thermal expansion coefficients result in nucleation of high density of threading dislocation greater than 1018/cm2 [133]. Techniques such as graded Ge-Si buffers have been suggested to reduce lattice mismatch strain between GaAs/Si lattices that could result in reduction of the threading dislocation densities [134, 135]. However, the GaAs solar cells fabricated on such substrate have not exhibited very high efficiency.
4.2 Epitaxial lift off process for GaAs substrate reuse
An efficiency >15% was achieved in large area III-V solar cells during the late
1970’s using techniques such as liquid phase epitaxy (LPE) [136]. However, this process used GaAs wafers that could not be reused. In order to reuse these wafers, an epitaxial lift-off (ELO) technique was suggested for single-junction (1J) solar cells grown using
LPE on GaAs wafer using Ga0.3Al0.7As (5 μm) as a release layer over GaAs (30 μm).
Hydrofluoric (HF) acid was used as a solvent in order to etch the release layer. The solar cell was then mounted on the aluminum (Al) foil after ELO, and the measured efficiency obtained at maximum concentration (109Suns) was 9.4% [137]. In the late 1980’s,
Yablonovitch et al. introduced 100 nm aluminum arsenide (AlAs) as a release or sacrificial layer with double heterostructure structure (GaAlAs/GaAs/GaAlAs/AlAs) using HF as the sacrificial layer etchant and compressive strain induced via Apiezon W wax (25 g of wax dissolved in 100ml of trichloroethylene sprayed to GaAs and cured).
Carrier lifetime studies using double hetero-structures before and after ELO were found to be exactly the same; thus, the quality of GaAs was unaltered by ELO [138]. In order to improve the epitaxial lift-off process, Geelen et al. developed a modified ELO method with AlGaAs as release layer. He used a 50 μm Apiezon W wax on 1J solar cell with an
66 additional 100 μm plastic foil attached to it. This was then hung on one side with weights for processing by ELO in HF. A solar cell efficiency of 9% at AM1.5 was obtained via this process [139]. The substrate was unaffected and was reused several times, but film cracking was induced during repeated lift-off process. In order to further improve the process, Chermer et al. used AlAs as a release layer in a hot HF solution (80ºC) to yield a high etching rate with cm-size, crack-free, single crystal indium gallium phosphide
(InGaP) films [140]. In the next set of experiments conducted by Chermer, the author was able to lift off the film from the entire wafer (50 mm) using the same ELO technique
[141]. Large-area ELO on a 100 mm wafer with a solar cell area of 1 cm2 was demonstrated by Tatavarti et al. with an efficiency >21% at AM1.5 [142].
In 2009, Bauhuis et al. fabricated a GaAs 1J solar cell with record efficiency of
26.1% under 1 sun using the ELO process. The authors used an improved ELO process with a high-quality active layer material after optimization, a low-temperature annealed front contact (lead/germanium/gold [Pd/Ge/Au])/back contact (Au), and anti-reflective coating (zinc sulfide/magnesium fluoride [ZnS/MgF2]) [143]. Today’s world record efficiency of 29.3% for 1J GaAs via ELO was demonstrated by Alta Devices [43]. A repeatable ELO process with multi-stack architecture (a multi-layer 1J solar cell separated by a sacrificial layer) was designed by Choi et al. for transfer printing of 1J
GaAs microcells from GaAs to foreign substrates. First, microcells were fabricated on the first stack and transferred to a foreign substrate using polydimethylsiloxane (PDMS) by repeating the same process for the entire stack. Efficiency >10% has been reported in such solar cells [144]. The rate at which the sacrificial layer undergoes lift-off plays an
67 important role in the ELO process. Approximately, a five-time increase in the ELO rate has been obtained by mixing HF with hydrophilic solvents, such as acetone [145].
A cleaning process is required for wafer reuse after ELO. Traditional chemical- mechanical polishing could be used for cleaning the wafer, but the cost would be greater than 20% of a regular substrate cost. So, an inexpensive wafer cleaning is needed. A detailed analysis of wafer cleaning steps after ELO and solar cell parameters of 1J GaAs after fabrication on reused wafers (50 mm) is given by Bauhuis et al. The authors used
AlAs (10 nm) as a sacrificial layer that was later etched using HF for ELO. The best cleaning process was obtained with a solution containing NH4OH/ H2O2/H2O (1:1:50) resulting in a root mean square (RMS) roughness ≤0.4 nm after cleaning. The overall efficiencies before and after the wafer reuse were comparable [146].
Despite all of these efforts, wafer reuse for ELO processing cannot be done repeatedly as several defective areas develop on each wafer and lead to a gradual decrease in efficiency. In addition, all of the processes for ELO are conducted in a prototype format and may not be suitable for large scale production. Thus, there is a requirement for an alternate method for low-cost substrates for III-V solar cells. We have taken an approach for large scale, low cost, thin-film GaAs solar cells based on roll-to- roll-(R2R) processed flexible epi-ready metal tape. This process is explained below.
4.3 Low-cost and scalable process for III-V solar cells
We have developed a novel epitaxial approach for growing low-cost and scalable
GaAs solar cells on light-weight and robust substrates via a roll-to-roll process. The process starts with fabrication of a Ge template on a flexible metal substrate using R2R
68 process [147]. The metal substrate is first coated with a highly biaxially-oriented film by ion beam assisted deposition (IBAD), which then enables epitaxial growth of a Ge film.
The Ge template on the epi-ready metal substrate can then be used to grow single junction GaAs thin film solar cell architectures [148, 149]. These GaAs films are single crystalline-like with structural and optoelectronic properties comparable to single crystal
GaAs. Single-junction GaAs solar cells have been fabricated with these single- crystalline-like GaAs on metal substrates. The details of the buffer layers and an architecture of a single junction GaAs solar cell on metal substrate are shown in figure
4.1.
Figure 4. 1 Schematic of 1J GaAs solar cell and buffer layers on a flexible metal tape.
The growth of 1J thin film GaAs on roll-to-roll-processed buffer consists of the following steps:
a) Growth of epi-ready buffer layer on metal tape
69
b) Growth of semiconductor buffer over epi-ready buffer
c) Growth of 1J GaAs solar cell on top of semiconductor buffer
The grown GaAs thin film structure then undergoes the processes of fabrication, metallization, passivation, and anti-reflective coating. The details of processing are explained below.
4.4 Growth of epi-ready buffer on metal tape
The IBAD technique, which uses a seed and epitaxy approach, was used to develop the epi-ready buffer layer. The challenge was to grow those buffer layers, which are lattice-matched, in a thermally and structurally suitable way for epitaxial growth of
GaAs. The detailed structure of an epi-ready buffer layer is shown in figure 4.2.
Figure 4. 2 Cross sectional microstructure and selected-area diffraction patterns from an epi-ready buffer architecture.
Briefly, the process begins with an electropolished Hastelloy C-276 tape for the roll-to-roll process as shown in Figure 4.2. C-276 Hastelloy consists of nickel, molybdenum, and chromium with small amounts of tungsten. The material has excellent resistance (due to a small amount of tungsten) toward sulfuric, hydrochloric, acetic, and
70 formic acids, wet chloride gas, hypochlorite and chorine solutions. Additionally, the material has high temperature tolerance, is mechanically robust, and amenable to growth at elevated temperatures. The Hastelloy used in our process was 50 μm thick and 1.2 cm wide.
The choice of material for growth of III-V GaAs was Ge (as discussed earlier) due to its lattice-matched properties with GaAs. However, Ge could not be grown directly on the polished Hastelloy.
(a) (b)
Figure 4. 3 R2R equipment for (a) Ion Beam Assisted Deposition (IBAD) and (b) Magnetron Sputtering.
We started with the IBAD technique (figure 4.3 a) in order to produce a biaxially- textured magnesium oxide (MgO) film. For this process, an 80 nm amorphous aluminum oxide (Al2O3) layer was deposited by R2R reactive radio frequency (RF) sputtering, which served as the diffusion barrier preventing metallic element diffusion from the substrate to the upper layer. Next, a nucleation layer consisting of amorphous yttrium oxide (Y2O3), 7 nm thick, was deposited using the R2R IBAD method. Eventually, a
MgO film, 10 nm thick, was deposited at room temperature, which resulted in single crystal-like, highly-(00l)-oriented MgO films that provided the platform for heteroepitaxy. The alignment occurred within 10 nm of the MgO thickness during the
71 rapid roll-to-roll process. Next, a 60 nm thick epi-MgO layer was grown via reactive sputtering (RS) at higher temperatures (~500 ºC) on the IBAD MgO layer in order to improve the crystalline quality.
In order to enable Ge epitaxy, an intermediate lanthanum manganese oxide
(LaMnO3/perovskite structure) with basal plane lattice constant, a=5.53 Aº, b=5.71 Aº, of 50 nm thickness was grown using R2R RF sputtering. LaMnO3 was epitaxially grown with a 45º rotation such that the pseudo plane (lattice constant, a=0.41) matched with
MgO with lattice mismatchs of 2.1% and 1.1% along the a and b directions, respectively.
Cerium oxide (CeO2) with a lattice constant of 5.4 Aº and a fluorite structure
(cubic) was grown on LaMnO3 via R2R magnetron sputtering in spite of the –4.5% lattice mismatch between Ge (5.65 Aº) and CeO2 (5.41 Aº) because of the good structural match between these materials.
Ge was then grown on top of the CeO2 tape using a R2R RF magnetron sputtering system. The optimized Ge tape with good in- and out-of-plane texture was then used as an epi-ready buffer layer. A semiconductor buffer layer was then grown over the epi- ready buffer layer using MOCVD growth techniques.
4.4.1 Growth of semiconductor buffer over epi-ready buffer
As explained previously, Ge and CeO2 have a lattice mismatch of >4%, which produces threading dislocations. Etch pit experiments conducted on the Ge film over the epi-ready layer showed substantial defects. Increasing the thickness of Ge layer from 1 to
6 μm, reduced the defect density significantly on the surface, but a large number of threading dislocations still persisted at the interface [150]. In another approach, a
72 semiconductor buffer grown before the growth of the 1J solar cell reduced threading dislocations that would have otherwise propagated to the p/n junction. The semiconductor buffer was grown via a custom built MOCVD tool with a dual chamber as shown in figure 4.4 and explained in chapter 3. Segments of these epi-ready Ge templates were used as substrates for growth of the semiconductor buffer layer.
Figure 4. 4 Custom-built dual MOCVD tool.
The MOCVD tool used to grow III-V materials is shown in figure 4.4. The precursors consisted of organometallic compounds: trimethyl gallium (TMGa), trimethyl aluminum (TMAl), and trimethly indium (TMIn) as Ga, Al, and In sources and an inorganic arsine (AsH3) as an As source. Because of their low memory effects, the dopants used for the process were silane (SiH4) and diethyl zinc (DEZn) for n and p types, respectively. Ultra-high purity hydrogen (H2) purified additionally with palladium to ppb levels was used as the carrier gas. The sample was mounted on a custom-designed graphite susceptor employed to hold the flexible substrate to ensure uniform heating of samples in the MOCVD reactor. The structure selected for the semiconductor buffer layer in order to reduce defects from the flexible epi-ready Ge tape is shown in figure 4.5.
73
Figure 4. 5 Structure of semiconductor buffer layer.
The flexible Ge template was loaded in chamber on a graphite susceptor. All of the conditions, such as the temperature, pressure and growth rate, were optimized for each layer before growing the actual device. First, the chamber pressure was fixed at 1
Torr, and 200 nm plasma-assisted GaAs was grown at a temperature of 550 oC and plasma power of 100 W. After plasma-assisted growth of GaAs, the pressure of the chamber was increased to 70 Torr and 1.8 μm undoped GaAs was grown at a temperature of 650 oC. The pressure of the chamber was kept constant for the growth all of the other remaining layers. On top of undoped GaAs, consecutive layers of GaAs/InGaAs (called superlattice) were grown with each layer thickness increment of 5 nm. A total of 30 periods of superlattice with total thickness of 150 nm were grown. Furthermore, the process was repeated with a growth of undoped GaAs (1.2 μm), superlattice (150 nm), and an undoped GaAs of 1.2 μm. The role of the semiconductor buffer layer was to act as a Ge diffusion barrier and reduce the defect density of the GaAs layer prior to deposition of the p-n junction. The buffer was ready for growth of p/i/n solar cells starting with the p++ layer. The details of the p/n layer grown using MOCVD is explained below.
74
4.4.2 Growth of 1J GaAs solar cell on top of semiconductor buffer
After growth of the semiconductor buffer layer, either p/n or a p/i/n solar cell structure was grown over the semiconductor buffer layer without taking samples out of the chamber. Details of the solar cell structures are shown in Figure 4.6.
(a) (b)
Figure 4. 6 Structure of (a) p/n and (b) p/i/n solar cells.
The p/n or p/i/n junction solar cell structures started with a highly doped p++
GaAs layer of bulk carrier concentration (BC) of 1 1019 cm-3 and 1.4 µm thickness as a lateral current conduction layer for p contacts (which is the back contact in the above structure). Subsequently, a 33 nm p++ AlGaAs layer as a back-surface field (B.C = 1
1019 cm-3) was grown at 750 oC under same pressure. A higher temperature for AlGaAs was used in order to avoid any oxide formation due to Al metal organics. The temperature was again decreased to 650 oC and 810 nm p-GaAs base (BC = 1 1017 cm-
3) was grown as the base layer, which was the absorber layer in the case of the p/n solar cell. Whereas, for the p/i/n solar cell, the p-GaAs layer is thin (67 nm), as the layer above it (intrinsic layer), can act as an absorber layer due to its greater thickness (~1μm). So, for p/i/n solar cells, an intrinsic layer of ~1 μm thick is grown above the p-GaAs (67nm).
75
Then, n+ GaAs (50 nm) as an emitter (BC = 5 1018 cm-3) was grown at the same temperature and pressure. Again, the temperature of the susceptor was increased to grow
33 nm of an n++ AlGaAs window layer (BC = 1 1019 cm-3). Finally, a highly doped 30 nm, n ++ GaAs cap layer (BC = 1 1019 cm-3) was deposited. The resulting thin-films were taken out of the chamber and further fabricated via lithography.
4.5 Conclusion
A single-junction (1J) GaAs solar cell structure was successfully grown by
MOCVD on a low-cost, flexible, light weight, R2R-processed epi-ready buffer on a metal substrate. The process started with the growth of a epi-ready buffer consisting of amorphous aluminum oxide (Al2O3) as the diffusion barrier / amorphous yttrium oxide as a nucleation layer/ IBAD MgO for hetero epitaxial growth/ reactive sputtering MgO to improve the crystalline quality/ Buffer suitable to grow germanium (RF sputtering
LaMnO3 and CeO2) and/ the germanium on 50 μm thick and 1.2 cm wide Hastelloy.
Next, segments of this epi-ready germanium buffer were used to grow semiconductor buffer with superlattice structure, to prevent diffusion of germanium. The layer consisted of plasma assisted undoped GaAs/ 30 period of superlattices (GaAs/InGaAs)/ undoped
GaAs/30 period of superlattices/ undoped GaAs. 1J GaAs solar cell structures with p/n or p/i/n structure were then grown on top of these semiconductor buffer without taking sample out of the chamber. The grown films were then characterized and fabricated. The details of the characterization and fabrication process with device results are shown in chapter 5.
76
Chapter 5 Fabrication and characterization of single-junction (1J)
GaAs solar cells on sputtered germanium over epi-ready
buffer
The single-junction (1J) GaAs solar cell on sputtered Ge over epi-ready metal tape was further processed for fabrication via photolithography to form solar cells of various sizes. The photomask pattern was designed using software (AutoCAD and L-edit) and made on a quartz mask using chrome (called the chrome mask). The structure for the
AutoCAD-designed photomask is shown in Figure 5.1.
Figure 5. 1 Photomask design for solar cell fabrication.
The above schematic shows the different sizes and shapes of the 1J solar cells that were designed. The designs and sizes were developed to help cope with many challenges that are common when fabricating solar cells, like the low carrier lifetime and shadowing by the contact, as well as allowing ease of measurement of the device.
Photolithography was used to fabricate 1J solar cells with the designed mask, as shown in the schematic in Figure 5.2.
77
5.1 Device fabrication process for 1J GaAs solar cells on epi-ready buffer
Figure 5. 2 Schematic diagram showing the fabrication steps of a single-junction (1J) GaAs solar cell on flexible epi-ready metal tape.
Achieving the appropriate sample surface flatness during photolithography is essential to obtain the desired pattern, resolution, and device yield, the flexible solar cell film was adhered with a conductive silver paste onto a silicon wafer to make it easy to fabricate. The adhered sample then underwent the following device fabrication steps:
1) mesa patterning using photolithography and etching using the etchant solution; 2) n- type and p-type contact deposition; 3) cap layer removal (parasitic removal); and 4) application of the anti-reflective coating (ARC).
First, the attached sample is patterned with the mesa pattern from the mask using photolithography. The patterning steps are optimized to get a sharp, photoresist pattern and the mesa pattern is formed on the flexible sample. The mesa-patterned sample then underwent etching with an etchant that is suitable for GaAs. In this study, a chemical
78 etchant consisting of sulfuric acid (H2SO4), hydrogen peroxide (H2O2), and water (H2O)
(1:8:280 by volume) was used for mesa etching at an etch rate of approximately 130 nm/min. The etch thickness was measured using a profilometer until the desired thickness for etching was achieved. The sample is then cleaned and again patterned for top contact.
E-beam evaporation of nickel (Ni) (5 nm), germanium (Ge) (6 nm), and gold (Au) (250 nm) is used for the top contact. The contact-deposited sample then undergoes a lift-off process using acetone followed by IPA cleaning and drying. The cleaned sample is again patterned with the bottom contact, followed by contact deposition with chrome (Cr) and
Au using e-beam evaporation. The contact-deposited sample is then lifted up using acetone followed by IPA cleaning and drying. Finally, a solar simulator is used to evaluate the fabricated 1J solar cells.
5.2 Measurement of fabricated 1J GaAs solar cells
First, the ohmic nature of the top and bottom contacts was confirmed by determining the sheet resistance and specific contact resistivity using the transmission line method (TLM) described in chapter 3. The electrode configurations of the top and bottom TLM are the same as those of the solar cell contact. The specific resistivities of the n- and p-type contacts were 8 10-4 Ωcm2 and 7 10-5 Ωcm2, respectively. The detailed calculation is as shown in the inset of figure 5.3.
79
(b) (a)
Figure 5. 3 Transmission line method (TLM) for (a) top and (b) bottom contacts.
After measuring the contact resistance, the solar cell efficiency was measured from the current-voltage (I-V) curve using an NREL-calibrated Oriel 200 series solar simulator at 1 sun. The measured illuminated and dark curves for the solar cell are shown in figure 5.4.
(b) (a)
Figure 5. 4 (a) Illuminated J-V characteristic and (b) dark current-voltage (I-V) responses of the fabricated GaAs solar cell.
80
Table 5. 1 Leakage Current (Io), Open circuit voltage (VOC), Short circuit current density (JSC), Fill Factor (FF), and Efficiency (η) of solar cells at initial stage of device fabrication.
Stage of Device VOC JSC Fill Efficiency Leakage 2 Fabrication (mV) (mA/cm ) Factor (%) current (Io) (%) (mA) As Fabricated 521 10 68 3.5 3 10-2
The measured solar cells showed an open circuit voltage (VOC) of 521 mV, short
2 circuit current density (JSC) of 10 mA/cm , Fill Factor (FF) of 68%, leakage current (IO) of 3×10-2, and efficiency of 3.5%. Further, the highly doped (cap) layer was removed to reduce parasitic absorption and enhance absorption in the base layer. Citric acid mixed with hydrogen peroxide (at a 1:1 ratio at room temperature) was used as the wet etchant at an etch rate of 5 nm/min for all the solar cell samples. The samples were re-measured after the cap layer removal as shown in figure 5.5.
14
0 As Fabricated ) As Fabricated 2 12 (b) Cap Layer Removal Cap Layer Removal (Passivation)
-1 (Passivation) 10 mA/cm ( 8 V = 521/560mV -2 OC 6 2 JSC = 10/12.6mA/cm log(Current) 4 -3 (a) FF = 68/68% 2 Eff. = 3.5/4.8%
-4 Density Current 0 -1000 -500 0 500 0 100 200 300 400 500 600 Voltage (mV) Voltage (mV) Figure 5. 5 (a) Illuminated J-V characteristic and (b) dark I-V responses of the GaAs solar cell, both as-fabricated and with the cap layer removed.
81
Table 5. 2 Io, VOC, JSC, FF, and η of solar cells at initial stage of fabrication and after cap layer removal.
Stage of Device VOC JSC Fill Efficiency Leakage 2 Fabrication (mV) (mA/cm ) Factor (%) current (Io) (%) (mA) As Fabricated 521 10 68 3.5 3 10-2 Cap Layer Removed 560 12.6 68 4.8 7 10-3
The measured solar cell’s efficiency was enhanced from 3.5% to 4.8% due to the
2 increases in VOC, from 521mV to 560mV, and current density, from 10 mA/cm to 12.6
2 mA/cm . Thus, after citric acid treatment, the VOC increased by ~30 mV and the JSC increased by 2.3 mA/cm2, which are attributed to a decrease in the reverse saturation
-2 -3 current (Io) of the device from 3 10 mA to 7 10 mA and GaAs cap-layer removal, respectively; this result was observed in all the fabricated devices. This result implies that the citric acid helped to passivate the side walls and the surface dangling bonds while etching the cap layer.
As discussed in chapter 2, solar cell devices suffer from reflection losses. To minimize reflection losses, solar cells are coated with ARCs. Zinc sulfide (ZnS) and magnesium fluoride (MgF2) were the ARCs chosen for this study. The thicknesses of the
ARC materials were optimized on a GaAs film while considering a minimum reflectance at a 1-sun reference spectrum. Figure 5.6 shows the wavelength vs reflectance graph obtained from the double-layer ARC on the GaAs film.
82
25 MgF2 (48nm)/ZnS (102nm) 20
15
10
5 Reflectance(%)
0 400 500 600 700 800 900 Wavelength (nm) Figure 5. 6 Reflectance vs. wavelength graph for double-layer ZnS/MgF2 anti- reflective coating (ARC) on the 1J GaAs solar cell.
As shown in figure 5.6, a reflectance of less than 5% was obtained in the wavelength range from 450 nm to 850 nm with 48 nm of ZnS and 102 nm of MgF2. This double-layer ARC was then applied to the fabricated solar cell devices and re-measured.
The illuminated J-V and dark I-V graphs of the solar cell after ARC application are shown in figure 5.7.
18
) 16 As Fabricated (a) As Fabricated 2 0 Cap Layer Removal Cap Layer Removal 14 ARC ARC
12 -1 mA/cm ( 10 (b) 8 -2 VOC= 521/560/566mV 6 2 J = 10/12.6/17.4mA/cm log(Current) 4 SC -3 FF = 68/68/68% 2 Eff. = 3.5/4.8/6.8% Current Density Current 0 -4 0 100 200 300 400 500 600 -1000 -500 0 500 Voltage (mV) Voltage (mV) Figure 5. 7 (a) Illuminated J-V characteristic and (b) dark I-V responses of the as- fabricated, cap-layer-removed, and ARC-coated GaAs solar cells.
83
Table 5. 3 Io, VOC, JSC, FF, and η of solar cells at initial stage of fabrication, after cap layer removal, and after ARC application.
Stage of Device VOC JSC Fill Efficiency Leakage current 2 Fabrication (mV) (mA/cm ) Factor (%) (Io) (mA) (%) As Fabricated 521 10 68 3.5 3 10-2 Cap Layer Removed 560 12.6 68 4.8 7 10-3 ARC Applied 566 17.4 68 6.8 7 10-3
Figure 5.7 shows the dark response and illuminated J-V curves of a GaAs solar cell as fabricated, after cap layer removal, and after ARC application, as measured at 1
2 sun. After ARC application, the solar cell’s JSC increased to 17.4 mA/cm due to the decrease in reflectance to 5%. The solar cell’s efficiency increased to 6.8% at 1 sun. No difference in IO was observed after ARC application.
Hence, we successfully fabricated a 1J GaAs solar cell with three stages of device fabrication. The final stage of device fabrication showed a promising VOC of 566 mV and an efficiency of 6.8% at 1 sun; those values are significantly lower than those of the single-crystal wafer-based GaAs solar cells fabricated as reference samples and grown under the same conditions. The measured efficiency of the 1J GaAs solar cell on the wafer was > 20% with a VOC of ~ 900 mV, as shown in figure 5.8.
84
19 )
(a) 2 17 (b) 15 V = 566mV
13 OC
mA/cm ( 11 J = 17.4mA/cm2 9 SC 6 FF = 68% 4 Eff. = 6.8% 2 Flexible
Current Density Current 0 0 100 200 300 400 500 600 Voltage (mV)
Figure 5. 8 Illuminated J-V characteristics for (a) wafer and (b) flexible 1J solar cells.
Table 5. 4 VOC, JSC, FF, and η of flexible and wafer 1J solar cells.
Stage of Device VOC JSC Fill Efficiency Fabrication (mV) (mA/cm2) Factor (%) (%) Wafer 900 26.2 85 20
Flexible 566 17.4 68 6.8
Therefore, after comparing the performance of the wafer and flexible solar cells, it is clear that the flexible GaAs cells’ performance must be enhanced before they are viable for practical applications. To further improve the device’s efficiency, it was necessary to understand the impact of passivation on solar cells and to study different means of passivating the GaAs film and device. The impact of citric acid passivation on the 1J
GaAs solar cell and other techniques, like hydrogen and hydrogen/phosphine passivation were studied. First, to understand the effects of citric acid passivation, 1J GaAs solar cells with different base thicknesses were fabricated and passivated. These solar cells were
85 analyzed via atomic force microscopy (AFM), scanning electron microscopy (SEM), X- ray photoelectron spectroscopy (XPS), and current-voltage (I-V) characterization. For the
AFM, SEM, and XPS analyses, passivation was performed on 1J GaAs films. The impact of citric acid passivation is further described below.
5.3 Impact of citric acid passivation on 1J GaAs solar cells
Different studies have described the effects of several passivation methods (using chalcogens like sulfur, selenium, etc. in the gas or solution phase) on the GaAs surface
[151]. For effective device passivation, the use of ammonium sulfide (NH4)2S [152, 153], sodium sulfide, and trioctylphosphine sulfide [154-156] have been reported; however, these passivating solutions contain sulfur, which is toxic. In addition, the passivation of
GaAs films with sulfur [151], selenium [157, 158], and H2S [159] was only moderately successful.
In this research, non-toxic citric acid was used for the passivation of GaAs solar cells. Citric acid, mixed with hydrogen peroxide at a 1:1 volume ratio, has been proven effective for the selective wet etching of the GaAs surface [160, 161]; however, no one has reported or studied whether this citric acid:H2O2 mixture can effectively passivate
(sidewall and surface) GaAs solar cells and thus improve their overall efficiency.
To understand the impact of citric acid passivation, 1J GaAs passivated films were studied using AFM, SEM, and XPS. The AFM and SEM images before and after passivation are shown in figure 5.9.
86
Figure 5. 9 (a) and (c) AFM images of GaAs before and after passivation; (c) and (d) SEM images of GaAs before and after passivation.
AFM topography scanning was conducted over a scale of 40 μm 40 μm. The average root mean square (RMS) roughnesses of the GaAs films before and after passivation are 107 nm and 75 nm respectively, taken at different location. This decrease in RMS roughness was attributed to the etching of the GaAs surface by the passivating solution. The corresponding SEM images show that the GaAs film was composed of grain sizes that ranged from 2 μm to 20 μm. Further analysis was performed using XPS to investigate the effects of oxidation before and after passivation. S
Figures 5.10 (a) and 5.10 (b) show the XPS spectra of the Ga 3d and As 3d core -2 levels of the GaAs films, detected at a takeoff angle of 45o from normal. Figure 5.10 (a) shows two XPS peaks (at binding energies of 19 eV and about 21 eV) for the film before passivation. The former peak can be assigned to the photoemission from Ga bonding in bulk GaAs and the latter to a photoelectron signal from oxidized gallium (Ga2O3). After passivation of the GaAs surface with citric acid, the intensity of the Ga 3d peak increased
87 and the oxide peak disappeared, indicating that the passivation step effectively removed the native oxide of Ga.
Figure 5. 10 XPS spectra of GaAs film before and after passivation using citric acid. (a) Ga 3d core level with its oxide, (b) As 3d core level with its oxide, and (c) O 1s core level.
Figure 5.10 (b) shows three XPS peaks that correspond to binding energies of 41 eV, 44.4 eV, and 44.9 eV for the sample before passivation. The peak at 41 eV can be assigned to the photoemission from As bonding in bulk GaAs. The peaks at 44.4 eV and
44.9 eV can be assigned to the photoelectron signals from As bonding with oxygen as
As2O3 and As2O5, suggesting the presence of As sub-oxides. After passivation, the As2O5 peak disappeared but the As2O3 peak remained, indicating that the passivation step is only partially effective in removing the As oxide. Figure 5.10 (c) shows the reduction in intensity of the O 1S peak from the GaAs surface after passivation. The overall reduction in oxides on the device is clear evidence that citric acid was an active passivant on the
GaAs surface and prevented further oxidation. This removal of surface oxides results in an unpinning of the Fermi level, thus suppressing surface recombination [162] and enhancing the open-circuit voltage.
Our results confirm that citric acid treatment is a critical step in the fabrication of
GaAs solar cells on flexible single-crystalline-like templates. To further probe this result,
88 the effect of citric acid passivation on the 1J GaAs solar cell’s structure (with three different base thicknesses: 1140 nm, 840 nm, and 380 nm) was investigated.
Figure 5. 11 Evaluation of solar cells with three different base thicknesses (1140 nm, 840 nm, and 380 nm). (a) and (c) Dark current voltage responses (b) and (d) illuminated J-V characteristics of the devices.
Figures 5.11 (a) and 5.11 (b) show the dark I-V responses and illuminated J-V characteristics, respectively, of as-fabricated solar cells with three different base thicknesses. Figures 5.11 (c) and 5.11 (d) display the data for the corresponding cells at the final stage of fabrication, i.e., after parasitic cap layer removal and ARC application.
The VOC, JSC, and Io values are similar between all three as-fabricated solar cells with different base thicknesses; however, the cells at the final fabrication stage exhibited increases in VOC, JSC, and efficiency as well as a decrease in Io. The values of each parameter at each different thickness are shown in table 5.5.
89
Table 5. 5 IO, VOC, JSC, FF, and η at three different base thicknesses, corresponding to before and after passivation.
Device Parameter 1140 nm 840 nm 380 nm
-2 -3 -2 -3 -2 -4 Leakage Current (IO) 2x10 / 6x10 2x10 /4x10 1x10 /4x10
VOC (mV) 454/500 450/518 440/560
2 JSC (mA/cm ) 9.9/15 9.9/15 9.9/16.7
FF (%) 60/60 61/65 64/69
η (%) 2.7/4.5 2.7/5 2.8/6.5
All cells, regardless of the base thickness, followed a similar trend after final fabrication, i.e., they exhibited increases in VOC, JSC, and η and a decrease in IO after final processing; however, the cell with a 380-nm base thickness showed significantly greater changes between the as-fabricated stage and the final stage, compared to the thicker cells.
The absolute and percentage changes in the VOC, JSC, η, and IO of solar cells with three base layer thicknesses, from the as-fabricated stage to the final stage after parasitic cap layer removal and ARC application, are tabulated in table 5.6 below.
Table 5. 6 Increase in open circuit voltage (ΔVOC), short circuit current density (ΔJSC), and efficiency (Δη) and decrease in leakage current (ΔIO) in solar cells with three different thicknesses.
Device Parameter 1140 nm 840 nm 380 nm
ΔIo -0.014 (-70%) -0.016 (-80%) -0.0096(-96%)
ΔVOC (mV) 46 (10%) 68 (15%) 120 (27%)
2 ΔJSC (mA/cm ) 5.1(51%) 5.1 (51%) 6.8(69%)
Δη (%) 1.8 (66%) 2.3 (85%) 3.7(132%)
90
The changes in the VOC of the cells with base thicknesses of 1140 nm and 840 nm were about 10% and 15%, respectively, whereas the change in VOC in the cell with the thinner base of 380 nm was about 27%. This substantially higher improvement was attributed to a higher reduction in the IO, which was 96% reduced in the thinner base cell.
2 The increase in JSC was similar and within the error margin (±1mA/cm ) of the measurement in all three types of solar cells (1140 nm, 840 nm, and 380 nm). The slightly higher JSC in the thinner cell at the final fabrication stage could be attributed to decreased recombination and increased carrier collection since the carriers had to travel smaller distances, compared to those of the thicker base cells. Overall, compared to cells with base thicknesses of 1140 nm and 840 nm, the cell with the thinner base thickness of
380 nm showed better passivation after citric acid treatment, which yielded better- performing devices. We attribute the passivation of the thinner base to the small surface area (584 μm2) exposed to the passivating solution, compared to the larger exposed surface areas of the cells with thicker bases (1750 μm2 area).
Next, the impact of passivation on devices of different sizes with the same base thickness (380 nm) was explored. For this purpose, solar cells A and D (with diameters of
500 μm and 1250 μm, respectively) were chosen. Figures 5.12 (a) and (c), (b) and (d) show the dark I-V responses and illuminated J-V curves, respectively, for solar cells A and D in the as-fabricated and final stages. The parameters for solar cells A and D are shown in table 5.7.
91
Figure 5. 12 Impact of passivation on solar cells A and D. (a) and (c) Dark I-V responses (b) and (d) illuminated J-V characteristics of solar cells A (500 μm) and D (1250 μm) at the as-fabricated and final stages.
Table 5. 7 IO, VOC, JSC, FF, and η of solar cells A (500 μm) and D (1250 μm) at the initial and final stages.
Device Parameter Device A Device D
-2 -4 -1 -3 Leakage Current (IO mA ) 1x10 /4x10 1x10 /4x10
VOC (mV) 465/560 386/550
2 JSC (mA/cm ) 9.8/16.7 7.9/15
FF 64/69 56/58
η (%) 2.9/6.4 1.7/4.8
The parameters observed for solar cells A (500 μm) and D (1520 μm) at the initial and final stages of device fabrication are tabulated in table 5.7. The smaller device shows better performance than the larger device; however, the net increments from the initial to
92 final stages of device fabrication were the similar (see Table 5.8). In the as-fabricated stage, cell D (1250 μm in diameter) showed higher leakage values, which resulted in a lower VOC; however, at the final stage, substantial increases in VOC, JSC, and efficiency and a decrease in Io were observed. The reverse saturation current at the final stage was two orders of magnitude lower for both cells, which resulted in higher final VOC values after citric acid treatment. In conclusion, there was no size dependence of the VOC, JSC, η, and Io values after device passivation.
Table 5. 8 ΔVOC, ΔJSC, Δη, and ΔIO in solar cells A (500 μm) and D (1250 μm) at the as-fabricated and final stages.
Device Parameter Device A Device D
ΔIo -0.0096 (96%) -0.096(96%)
ΔVOC (mV) 95 (21%) 164 (42%)
2 ΔJSC (mA/cm ) 6.9 (70%) 7.1 (89%)
Δη (%) 3.5 (120%) 3.1 (182%)
Thus, 1J GaAs solar cells with different base thicknesses and device sizes were fabricated and the impact of passivation was studied. Citric acid, which is commonly used for etching GaAs, acted as an effective passivant during device fabrication. The devices’ Io values decreased after passivation, resulting in increases in VOC for all the solar cells, which ranged in size from 500 µm to 1250 µm. In addition, the use of citric acid successfully prevented the oxidation of the GaAs solar cells by removing the Ga and
As oxides from the film’s surface. A comparative passivation study of solar cells with different base thicknesses (1140 nm, 840 nm, and 380 nm) revealed that the passivation
93 more effectively reduced the Io (with a corresponding increase in VOC) in cells with a base thickness of 380 nm.
After clearly delineating the impact of citric acid passivation, which passivated the surface and side wall, the next step was to passivate the grain boundaries (GBs) and bulk of the 1J GaAs solar cell using hydrogen and phosphine-incorporated hydrogen passivation, as described below.
5.4 Impact of hydrogen and hydrogen/phosphine passivation on 1J GaAs films
As explained in chapter 4, the GaAs films grown on the epi-ready template are
“single-crystalline-like” and have structural and optoelectronic properties comparable to those of single-crystal GaAs. One of the main differences between the “single-crystalline- like” GaAs films on metal substrates and single-crystal GaAs films on bulk GaAs wafers is the presence of 2- to 4-micrometer grains with a high density of low-angle GBs. Even though the grains are highly oriented in the (00l) direction with ultra-low grain misorientation (< 1 degree), the GBs are still expected to be highly electrically-active defect centers. GBs play a critical role in reducing the VOC and limiting the performance of the devices, as observed in other polycrystalline semiconductor devices [163-165].
Post-passivation treatment on 1J GaAs films before fabrication can make GBs benign and electrically neutral, which may be an effective way to improve the device’s performance
[166].
One particularly effective post-passivation treatment for GBs is hydrogen plasma treatment. Hydrogen plasma passivation treatment passivates the shallow acceptors
(electrically active dislocations), removes trap states, reduces surface recombination
94 velocities, and effectively improves the optoelectronic properties of GaAs and other semiconductors, such as Si and Ge [164, 165]. However, all the previously-reported hydrogen plasma treatments were carried out on wafer-based GaAs devices and therefore mainly passivated the surface defect states and reduced surface recombination velocities.
Unlike polycrystalline Si, Ge, and CdTe, where GB passivation has been widely reported,
GB passivation of poly-GaAs has not been studied extensively and few reports are available on the passivation of GaAs GBs.
First, hydrogen plasma passivation was performed on GaAs films in a PECVD chamber. For this process, hydrogen (at 150 standard cubic centimeters per minute
(SCCM)) was used at a substrate temperature of 150 oC and ICP power of 250 W. The passivation was done for an hour under a chamber pressure of 150 mTorr. XRD was then conducted to check the film’s crystalline quality. The strong (00l) peaks of GaAs confirmed no degradation of epitaxy of the GaAs films. The peak width from the rocking curve of the (00l) peak of GaAs was 1.04º before and after treatment. No significant changes were observed in the out-of-plane peak widths and the intensity of the peaks which indicates that hydrogenation had little to no tangible effect on the film’s structural properties.
Figure 5.13 shows the XRD 2-theta plots of GaAs films before and after plasma treatment. The XRD plot shows the crystalline orientation of GaAs and the underlying template. The strong (00l) peaks of GaAs confirm epitaxial growth of the GaAs films and no disruption of epitaxy after plasma treatment. The peak widths from the rocking curves
(shown in the inset) before and after plasma treatment were 1.27 and 1.24 degrees,
95 respectively. Again, no significant change was observed in either the out-of-plane peak widths or peak intensities.
Figure 5. 13 XRD and FWHM for the as-deposited and H2 plasma-treated GaAs films on epi-ready metal substrate.
Further analysis was done using photoluminescence (PL) on the as-deposited and passivated films at room temperature, as shown in figure 5.14. The results showed a significant enhancement in the PL peak intensity after hydrogenation. In addition, the peak width also marginally decreased, from 63 to 62 eV, and the band-to-band peak position shifted from 1.40 to the typical 1.41 eV for n-type GaAs films, indicating that the film’s optical quality was improved and shallow defect states were removed from the band gap due to the hydrogenation. For the p-type contact, the peak width decreased from
52 to 50 eV and the band-to-band peak position shifted from 1.41 to the typical 1.42 eV.
It has been reported that the Ga-H and As-H bonds created during hydrogen plasma treatment play a critical role in removing mid-band gap states in GaAs and other semiconductors [167, 168].
96
Figure 5. 14 PL spectra of the as-deposited and H2 plasma-treated GaAs films on epi- ready metal substrate.
XPS was performed to determine the effect of the hydrogen plasma treatment on the surface states of GaAs and the results are shown in figure 5.15. A substantial decrease
in oxygen content and reduction in gallium oxide (Ga2O3) and arsine trioxide (As2O3) indicate the removal of native oxides from the GaAs surface, which suggests that hydrogen plasma was responsible for removing these native oxides from the surface.
Overall, no significant changes in the structural properties during hydrogenation were observed.
Figure 5. 15 XPS analysis of the as-deposited and H2 plasma-treated GaAs films on epi- ready flexible metal substrate.
97
Next, Kelvin probe force microscopy (KPFM) was used to understand the effect of passivation on the GBs of the epitaxial GaAs on epi-ready metal substrate. Figures 5.16
(a) and (c) show the surface morphologies of the as-deposited and hydrogen plasma- exposed n-GaAs films. The surface morphology consisted of layered grains with distinct
GBs. The average RMS surface roughness, measured from multiple images at several locations on the surface, were 21 and 22 nm for the as-deposited and plasma-treated samples, respectively. The morphology appeared to be largely unchanged and the roughness variation was within one standard deviation. It can be safely suggested that the plasma treatment did not noticeably modify the surface.
Figure 5. 16 Kelvin probe force microscopy images and line profiles for the as- deposited and H2 plasma-treated GaAs films on epi-ready metal substrate.
Figure 5.16 (b) and (d) show the corresponding surface potential images of the as- deposited and untreated samples. Using line profile analysis, modulation along the GBs was observed in the as-deposited sample. A significant decrease in surface potential, from
98
620 mV to 522 mV, was observed across the entire 6 µm 6 µm area after plasma treatment. Moreover, the narrowing of the peak suggested that the surface potential was more tightly distributed after plasma treatment, i.e., with smaller variations from point to point. Figure 5.16 (f) shows the modulation of the surface potential across the line profiles shown in figure (b) and (d) by a discontinuous line. The modulation across the
GBs is much more pronounced for the as-deposited sample, whereas the modulation is significantly reduced after plasma treatment. A potential drop (about 400 mV) occurred along the GBs, as shown by an arrowhead in figure 5.17. This is attributed to the passivation of the energetic defect states at the GBs by reducing the potential barrier height at the GBs. The decrease in GB potential due to hydrogen plasma passivation may potentially result in more efficient charge transport and far fewer recombination events, thereby improving the devices’ current carrying and collecting capabilities.
Figure 5. 17 Electrical conductivity and resistivity of GaAs films before and after passivation.
99
Next, to confirm the hypothesis that hydrogen plasma passivation enhanced electrical conduction, temperature-dependent I-V measurements (77–350 K) were conducted on the n-GaAs film before and after plasma treatment to identify any changes in electrical conductivity. Figure 5.17 (a) shows the temperature-induced variations in electrical conductivity of the as-deposited and plasma-treated n-GaAs film on metal foil.
A significant increase in electrical conductivity was observed after plasma treatment at all temperatures, which can be correlated to the decrease in overall surface potential and GB barrier height, as observed in the local surface potential results. Figure 5.17 (b) shows the
Arrhenius plot (lnσ vs 1/kT) of the as-deposited and plasma-treated samples. The electrical activation (Eb) energy was determined from the exponential dependence of conductivity with temperature α exp (-Eb/kT). Activation energies of 41 and 30 meV were estimated for the as-deposited and hydrogenated samples, respectively. An 11-meV drop in activation energy from the bulk I-V measurement provided additional support to findings from the local surface potential measurements. The decrease in activation energy
(Δϕ) indicated decreased surface potential, decreased GB barrier height (as explained in the GBT model [169]), and increased carrier conduction.
To understand the diffusion of hydrogen through the GaAs solar cell structure, secondary ion mass spectroscopy (SIMS) analysis was conducted. A quantitative analysis of the concentration of hydrogen atoms diffused into the GaAs structure is shown in figure 5.18. A cesium (Cs+) primary ion source was used for ion bombardment. The hydrogenated sample was compared with a non-hydrogenated sample to better understand the hydrogen diffusion. Compared to the as-deposited sample, hydrogen was
100 concentrated more in the hydrogenated sample and diffused to a depth of 600 nm. In the region beyond 600 nm, the hydrogen concentration appeared to be almost flat.
Figure 5. 18 Depth profile for hydrogen in GaAs solar cells.
After detailed study of hydrogen passivation on GaAs films, 1J GaAs solar cells were fabricated and measured at 1 sun. Figure 5.19 shows the illuminated and dark I-V curves for the as-fabricated and hydrogen-treated solar cells. The VOC increased from 380 mV to 806 mV and the Io decreased by two orders of magnitude for the hydrogen-treated cell, as shown in table 5.9. The solar cell’s FF increased from 52% to 68%, indicating the
2 2 impact of passivation; however, the JSC decreased from 14.2 mA/cm to 8.5 mA/cm .
This decrease in current is attributed to some arsenic vacancies that can be caused during the hydrogenation process.
101
16
) As Fabricated (No Plasma) 0 As Fabricated (No Plasma) 2 14 Hydrogen Plasma -1 Hydrogen Plasma 12
-2 mA/cm
( 10 8 -3 6 -4 4 log(current(mA)) -5 2 -6
Current Density Density Current 0 0 150 300 450 600 750 900 -1000 -500 0 500 Voltage (mV) Voltage (mV) Figure 5. 19 Illuminated and dark I-V curves for the as-deposited and plasma-treated GaAs solar cells on epi-ready metal substrate.
Table 5. 9 η and IO before and after H2 plasma treatment of GaAs solar cells on epi- ready metal substrate.
Device Before Treatment After Treatment Parameter
Leakage 1x10-2 1x10-4 Current (Io)
VOC (mV) 380 860
2 JSC (mA/cm ) 14.2 8.5
FF 52 68
η (%) 2.8 4.7
It has been suggested that the effect of the lower JSC in the hydrogenated sample can be mitigated by phosphine (an n-type dopant)-incorporated hydrogen passivation.
Specifically, the incorporation of phosphine has led to the passivation of surface defects, such as arsenic vacancies, thereby reducing the formation of As2O3 and free As from the
GaAs surface [170, 171]. Phosphine-incorporated hydrogen passivation has been explained in the literature for wafer devices; for the first time, the process for a “single- crystalline-like” 1J GaAs film has been implemented.
102
16 0
) As Fabricated (No Plasma) 2 14 -1 Phosphine/Hydrogen Plasma 12
-2 mA/cm
( 10 8 As Fabricated No Plasma -3 Phosphine/Hydrogen Plasma 6 -4
4 log(current(mA)) -5 2 -6
Current Density Density Current 0 0 150 300 450 600 -1000 -500 0 500 Voltage (mV) Voltage (mV)
Figure 5. 20 Illuminated and dark I-V curves for as-deposited and phosphine- incorporated hydrogen plasma-treated GaAs solar cells on epi-ready metal substrate.
Table 5.10 η and IO before and after H2 plasma treatment of GaAs solar cells on epi- ready metal substrate.
Device Parameter Before After Treatment Treatment
-2 -3 Leakage Current (Io) 1x10 1x10
VOC (mV) 380 560
2 JSC (mA/cm ) 14.8 14.2
FF 52 58
η (%) 2.8 4.8
Figure 5.19 shows the illuminated and dark I-V curves for the as-fabricated and phosphine-incorporated hydrogen-treated cells. The VOC increased from 380 mV to 560 mV and the IO decreased by one order of magnitude for the hydrogen-treated cell, as shown in table 5.10. The solar cell’s FF increased from 52% to 58%, indicating the impact of passivation; however, the JSC remained almost constant without any significant decrease, as occurred with hydrogen-only passivation. This shows that phosphine- incorporated hydrogen passivation can lead to a significant enhancement in device 103 efficiency without a reduction in current density. To further enhance the efficiency of the
1J GaAs solar cell, in-depth hydrogen and phosphine-incorporated hydrogen diffusion analysis is required.
Overall, hydrogen and phosphine-incorporated plasma treatment proved to be an effective method to improve the optoelectronic properties and device performance of granular biaxially-textured GaAs films. It is believed that the passivation of GBs and reduction of GB barrier heights, as evidenced by the Kelvin probe force microscopy studies, played an important role in decreasing interface and sidewall recombination, reducing the IO and increasing the VOC of the solar cells. In addition, the removal of oxide states may also be an important factor when improving metal-semiconductor interface properties.
5.5 Conclusion
First, a 1J GaAs solar cell was fabricated with three different stages of device fabrication: initial, after passivation with citric acid, and after ARC application. To understand the impact of citric acid passivation, 1J GaAs solar cells with different base thicknesses (380 nm, 840 nm and 1140 nm) and device sizes (500 µm and 1250 µm) were fabricated and the impact of passivation was studied. Citric acid, which is commonly used for etching GaAs, acted as an effective passivant during device fabrication. The IO of the devices decreased after passivation, resulting in increased VOC
(380 mV to 860 mV) values for both the 500 µm and 1250 µm solar cells. In addition, the use of citric acid successfully prevented the oxidation of the GaAs solar cells by removing the Ga and As oxides on the surface and sidewall. A comparative passivation
104 study of solar cells with different base thicknesses (1140 nm, 840 nm, and 380 nm) revealed that the passivation was more effective in reducing the IO (with a corresponding increase in VOC) in cells with a thinner base thickness.
In addition to citric acid passivation of the surface and side wall on fabricated solar cells, bulk passivation techniques, such as hydrogen and hydrogen-incorporated phosphine passivation, were studied on the 1J GaAs film to further enhance the device’s efficiency and determine the impact of passivation. Hydrogen passivation led to an increase in the open-circuit voltage by 480 mV. Meanwhile, the JSC decreased by 5.7 mA/cm2, which was attributed to the reduction in the level of surface arsenic (n-type dopant). Therefore, to overcome the adverse effects of hydrogen passivation, phosphine was incorporated. The results showed no reduction in current, as observed in hydrogen passivation, and a 180-mV increase in VOC. Further optimization of hydrogen/phosphine diffusion on 1J GaAs films and understanding their distribution will help to improve the solar cell’s overall efficiency.
In addition, to enhance the efficiency of the solar cell further, other routes should be investigated. One such route is to increase the efficiency of the solar cells by improving the quality of the GaAs. The GaAs quality was improved via two different processes:
1) Improving the quality of the Ge template; and
2) Incorporating an intrinsic layer with a PIN solar cell structure.
The details regarding these two methods are explained in chapters 6 and 7, respectively.
105
Chapter 6 Improving the quality of a germanium template using R2R
PECVD Ge for high-quality 1J GaAs solar cells
To further improve the quality of the sputtered Ge layer, a PECVD Ge layer was coated over the sputtered Ge layer. First, sputtered Ge was deposited, as explained in chapter 4, via an R2R RF magnetron sputtering system on top of an epi-ready CeO2 buffer. A small section of the sputtered Ge was then spot welded to a Hastelloy leader tape and spooled in an R2R PECVD Ge deposition chamber. A schematic of the roll-to- roll process chamber with a photograph of CVD Ge over the epi-ready Ge is shown in figure 6.1.
Figure 6. 1 Schematic of PECVD Ge deposition on R2R chamber.
A custom-designed R2R PECVD chamber was developed to enable continuous growth of Ge. It comprises a continuous deposition chamber with dispensing and collecting spools. The continuous spool can hold over a kilometer-long substrates for multi-layer deposition by rolling the tape back and forth in the reactor between the two spools. A polymeric interleaf film is used to protect the deposited films from contamination.
106
The deposition chamber is fitted with IR lamps with a heating capability up to
1300 °C in the deposition zone. High-purity Germane (GeH4), with a purity of 99.999% and hydrogen purified by a palladium cell to a ppb level, were used as process gases. An
RF inductive-coupled hydrogen plasma (ICP) at a power of 250 W and pressure of 2 Torr was used. The substrate temperature was 650 °C. For R2R deposition, a tape speed of 6 cm per minute was used to achieve a Ge thickness of ~1 µm. The deposited PECVD Ge film was then used for deposition of the GaAs thin film via MOCVD. The MOCVD growth process for the single junction GaAs solar cell structure was the same for both substrates (sputtered and CVD germanium), as explained in chapter 4.
The PECVD Ge and MOCVD GaAs films were analyzed by several characterization tools, as explained in chapter 3. First, the crystallography orientation and microstructure of the deposited film were analyzed. Theta-2theta and rocking curve measurements were performed using HRXRD (a triple axis Bruker D8 diffractometer) for out-of-plane texture. X-ray pole figures were obtained using GAADS (Bruker 2D
General Area Detector Diffraction) XRD. In-plane texture was determined from the phi scans extracted from the pole figure. Raman spectra (T64000 spectrometer from Horiba) were obtained using a blue (488nm) excitation laser (<3mW) wavelength in the backscattering geometry.
The morphology of the film was studied using AFM, SEM and TEM. SEM imaging was carried out at different sample locations using 55VP Zeiss SEM at 20kV and
10 - 15nA current in a variable pressure mode. AFM imaging (Agilent 5100 AFM) in tapping mode was performed for measuring the RMS roughness. Transmission electron microscopy (JEOL JEM-2100 LaB6) was used for plane-view imaging at 200 keV at
107 several locations. The carrier concentration and Hall measurement were conducted in
Van der Pauw geometry (using HMS-5000 Ecopia), as explained in chapter 3. All the above mentioned characterization analyses were used to study the properties of CVD Ge and GaAs film on CVD germanium in comparison to the GaAs films on sputtered germanium.
6.1 Properties of CVD Ge on flexible substrates
Figures 6.2 (a) and (b) show the AFM surface topography image, and (c) and (d) show the SEM image of the sputtered and CVD germanium, respectively. As observed from the AFM and SEM images, the CVD Ge comprises of interconnected grains with elongated grain morphology, whereas sputtered Ge has a granular grain morphology. The root mean square (RMS) roughness of sputtered and CVD germanium is 10±0.5 nm and
9±0.5 nm, respectively, showing smoother morphology for CVD germanium.
Figure 6. 2 AFM and SEM image for CVD and sputtered Germanium.
108
Figure 6. 3 XRD for sputtered and CVD germanium with a rocking curve.
Figure 6.3 shows a theta-2theta scan with a rocking curve (inset of figure) of the
R2R sputtered and CVD Ge, respectively. For both types of Ge, strong (004) Ge peaks and absence of any other peak indicates a high degree of alignment towards the (00l) direction. The out-of-plane texture measured via the rocking curve of the sputtered and
CVD Ge for the (004) germanium peak are 1.70 ° and 1.50 °, respectively, indicating improved crystalline quality for CVD germanium.
Figure 6. 4 In-plane texture of sputtered and CVD germanium.
109
Further, the in-plane texture was determined using GAADS via a pole figure.
Figure 6.4 (b) shows the XRD (220) pole figure of the sputtered and CVD germanium, indicating strong four-fold symmetry and confirming biaxial texture. The phi scan (Δϕ) obtained from the pole figure for the sputtered and CVD germanium is shown in figure
6.4 (a). The phi scan measurement shows an in-plane texture spread of 2.41 ° and 3.04° for CVD and sputtered germanium, respectively, suggesting higher in-plane alignment of grains in CVD germanium compared to the sputtered germanium.
Figure 6. 5 Raman shift for the sputtered and CVD germanium.
The crystal quality of the film was further examined using Raman spectroscopy.
Figure 6.5 compares Raman spectra for the sputtered and CVD germanium. They show intense and narrow Raman peaks (transverse optical mode) at 302 cm-1, corresponding to crystalline germanium. The Raman peak-width determined using a Lorentzian fit of the peaks shows peak widths of 5.8 and 5.5 cm-1, corresponding to the sputtered and CVD germanium and indicating improved crystallinity for CVD germanium.
Finally, Hall mobility at room temperature using Van der Pauw geometry was used to measure carrier concentration and mobility of the sputtered and CVD germanium.
110
Both the germanium types were n-type in character, with average carrier concentration of
(2±0.5) × 1018 cm-3. The measured mobility of the sputtered and CVD germanium were
80±20 cm2/Vs and 230±20 cm2/Vs, respectively. The mobility of the sputtered germanium was substantially lower than the CVD germanium. Thus, depositing CVD germanium over the sputtered germanium improves the quality of the CVD germanium, providing a platform for an improved GaAs film. Next, a GaAs film was grown by
MOCVD on top of the sputtered Ge, and the CVD germanium and GaAs properties were characterized.
6.2 Properties of flexible GaAs on R2R CVD Ge
X-Ray diffraction theta-2theta with a rocking curve analysis was performed to understand the crystalline orientation of GaAs grown on the sputtered and CVD germanium. Figure 6.6 (a) demonstrates the diffraction pattern with (004) and (002) peaks of zinc blende GaAs. This confirms the epitaxial growth with a strong (00l) orientation for the single crystalline nature of GaAs. The rocking curve analysis, as shown in figure 6.6 (b), reveals out-of-plane widths (Δω) of 1.20 and 1.10 for GaAs on the sputtered and CVD germanium, respectively. The narrow out-of-plane width for
GaAs on CVD Ge is attributed to improved texture and crystalline quality of CVD germanium. The in-plane texture (ΔΦ) obtained from the (220) pole figure for GaAs on
CVD germanium is 1.78 ° (figure 6.6) compared to the sputtered germanium 2.2°.
111
Figure 6. 6 (a) XRD and (b) rocking curve for GaAs on the sputtered and CVD germanium (c) (220) pole figure for GaAs on CVD germanium.
Next, room temperature photoluminescence (PL) as shown in figure 6.7 data was acquired on p-GaAs to investigate the optical properties. Measurement was done using a fixed excitation wavelength of 690 nm and an incident power of 10 mW. The peak locations for GaAs on the sputtered and CVD germanium were obtained at 1.41 and
1.42eV with peak widths (fitted using Lorentz function) of 55 and 49 meV, respectively.
Shifting of the peak position for GaAs on the sputtered germanium (1.41 eV) from the ideal band gap transition energy 1.42 eV suggested lowering film quality.
Figure 6. 7 Raman spectra for p-GaAs on sputtered and CVD germanium.
Next, the dislocation density on GaAs was calculated using plan-view STEM.
Plan-view STEM images of an undoped GaAs (2 μm) film on the sputtered and CVD germanium are shown in Figure 6.8. The grain size of GaAs ranged over from 2-4
112 microns. The figure inset shows a selected area electron diffraction pattern, exhibiting the single crystalline nature of the film corresponding to the zinc blende structure. The threading dislocation density for GaAs on CVD germanium is less than that for GaAs on sputtered germanium. The majority of defects in the sputtered germanium were confined at the grain boundaries regions. Several plan-view images were used to calculate the
8 threading dislocation density (TDD). The average TDD was counted manually as 8×10
-2 8 -2 cm and 4×10 cm for GaAs on the sputtered and CVD germanium, respectively. This clearly shows improved quality of GaAs on CVD germanium.
SAED
Figure 6. 8 STEM Image for GaAs on sputtered and CVD Ge.
6.3 1J GaAs solar cell results on the sputtered and CVD germanium
A single junction GaAs solar cell device structure was grown via MOCVD on the sputtered and CVD germanium. A schematic of the solar cell structure grown is shown in
Figure 6.9. Eachindividual layers were optimized before growing the actual device. The thin film device structure on flexible tape was then processed for fabrication of the solar cell with different device sizes from 500 to 1250 μm. The detailed parameters obtained from different-sized devices are tabulated in table 6.3.
Figure 6.9 shows the J-V characteristic for the fabricated solar cell device on
CVD germanium at three different stages of device fabrication, as explained in chapter 5
113
(initial, after cap layer removal/parasitic and ARC). All of the process conditions were kept consistent for the solar cell as that fabricated for the sputtered germanium. The open circuit voltage (VOC) at the initial stage of device fabrication was 592 mV, which is higher than the VOC (561mV) obtained for GaAs on sputtered germanium (chapter 5 section 5.2). This is attributed to the one order of magnitude lower leakage current (1×10-
3mA), as shown in Figure 6.10 (b), for the solar cell on CVD germanium.
(
Figure 6. 9 J-V characteristic of the solar cell fabricated on CVD germanium.
Table 6. 1 J-V parameter for the 1J GaAs solar cell fabricated on CVD germanium.
Stage of Device VOC JSC Fill Leakage Stage of Device Fabrication (mV) (mA/cm2) Factor current Fabrication (%) (Io) (mA) As Fabricated 592 11.4 65 4.5 110-3 Cap Layer Removal 634 22 67 9.2 210-4 ARC 642 25 70 11.02 810-5
Table 6.1 shows the device parameter extracted from the J-V plot for the solar cell fabricated on CVD germanium at three different stages of device fabrication. The VOC of
114
2 the solar cell increased from 592 mV to 642 mV, and JSC increased from 11.4 mA/cm to
25 mA/cm2, with overall increase in efficiency from 4.5% to 11.02%.
24)
2 GaAs on Sputtered Ge 20 GaAs on CVD Ge
16mA/cm ( VOC= 566/642mV 12 J = 17.4/25mA/cm2 8 SC FF = 68/70% 4 Eff. = 6.8/11.02%
Current Density Current 0 0 100 200 300 400 500 600 Voltage (mV)
Figure 6.10 J-V characteristic for the 1J GaAs solar cell on sputtered and CVD germanium
Figure (6.10) shows a comparison of J-V plots for a single junction (1J) GaAs solar cell fabricated on the sputtered germanium (black curve) and CVD germanium (red curve). All of the parameters of the solar cell have higher values for the 1J solar cell on
CVD germanium compared to the sputtered germanium. This improved performance of the solar cells on CVD germanium templates may be attributed to higher material quality- that is, low defect density, higher crystalline quality and better grain alignment for CVD germanium. The above result was demonstrated for the champion device. Measurements were also performed on a few other devices of same size whose comparisons are shown in Table 6.2.
115
Table 6. 2 Device data of GaAs solar cells on CVD and sputtered Ge template showing results of four devices.
GaAs device on CVD Ge template GaAs device on sputtered Ge template
Device Eff Voc Jsc FF Device Eff Voc Jsc FF # (%) (mV) (mA/cm2) (%) # (%) (mV) (mA/cm2) (%) CVD1 11.02 642 25 70 Sput1 6.8 566 17.4 68 CVD2 8.8 606 19 70 Sput2 5.2 531 15 65 CVD3 6.8 622 18.2 64 Sput3 6.4 571 16.5 66 CVD4 8.1 612 18.5 72 Sput4 6.5 550 17.1 69 The first row shows the parameter obtained for the champion device, whereas other rows are for other devices made during the same period of time. To ensure a consistent comparison, devices of the same size (i.e., 500 μm) were compared in above
Table 6.2.
Table 6. 3 Representative device data of GaAs solar cells on sputtered Ge template showing results of four different device sizes.
Jsc Eff Device ID Dia (µm) Voc (mV) 2 FF (%) (mA/cm ) (%) A 500 549 16.8 72 6.7 B 750 543 15.8 64 5.5 C 1000 542 15 65 5 D 1250 534 15.4 62 5.1
The fabricated solar cell comprised devices with four different sizes: 500 μm, 750
μm, 1000 μm and 1250 μm. The result for the 1J GaAs solar cell (champion device) on sputtered germanium comprising four different sizes is shown in Table 6.3.
116
The decrease in VOC, JSC, and overall efficiency in the case of bigger devices is due to the increase in leakage current. The path for leakage current is the increasing grain boundaries along with an increase in device size. This trend was observed for all the devices fabricated independently of the type of substrate used. Techniques such as passivation of grain boundaries (Chapter 5), use of other buffers with a larger grain size
[e.g., Titanium nitride (TiN), Molybdenum (Mo) etc.], selective area epitaxy (SAE) and other methods would provide better pathways for increasing efficiency in larger devices.
Figure 6. 11 Dark I-V for 1J GaAs on sputtered and CVD germanium.
Next, series and shunt resistance calculations were performed from dark I-V characteristics (at final stage) near the short circuit current and open circuit voltage. The
I-V curves for the 1J GaAs solar cell on the sputtered and CVD germanium are shown in figure 6.11. Figure 6.11 shows the dark current-voltage (I-V) plots of the GaAs solar cells fabricated on CVD and sputtered Ge templates. The I-V curve was fitted with the diode equation I = Io[exp(qv/nkT)-1] as described in chapter 2, to determine the saturation current (Io) and ideality factors (n), where I is the current across the diode, Io is the reverse saturation current, V is the applied voltage, e is the charge of electron, k is the
117
Boltzmann constant and T is the absolute temperature. The diode ideality factor (n) and reverse saturation current (Io), determined by fitting the dark I-V curve using standard diode equation, were 2.6 (for voltages between 0.5 to 0.6 V) and 2.2 10-8 A,
-9 respectively, which are higher than the previously reported n of ~ 2 and Io of ~ 10 A for champion GaAs devices on wafer substrates. The high values of n and Io of the flexible
GaAs cell suggest an increased non-radiative recombination that may be attributed to the high-density of grain boundaries, intra-grain defects and un-passivated sidewalls and perimeters of the flex-GaAs cells.
The average series (Rs) and shunt resistances (Rsh) of both devices were 260±50 and 1150±50 Ohms, respectively. The devices were thus limited by the low shunt resistance, which may be due to the shunt path created via the grain boundaries where side wall and bulk leakage occur. However, the series resistance was small which is likely due to the low contact resistance of the top and bottom contacts.
Figure 6. 12 External quantum efficiency for a 1J solar cell.
An Oriel IQE-200 setup with a tungsten filament light source and Newport monochromator calibrated using silicon and germanium reference detectors was used to determine the EQE spectra response for 1J GaAs solar cells. The quantum efficiency
118 yield was moderate with a peak of ~ 58 % at the lower wavelength region. The EQE curve between 500-800 nm shows recombination in the absorber or base region of the solar cell. The downward slope in this region indicates insufficient absorption in the base, as explained from ideal curve deviation in Chapter 3 (section 3.5.7). This is due to the short diffusion length of the minority charge carriers in the solar cells. For GaAs on flexible metal foil, the previous measured carrier life time was 2 ns (using double hetero structure), and the estimated diffusion length is around 1.2 μm . The base thicknesses of the fabricated solar cells were thus designed to be 0.8-1.2 μm to effectively collect the carriers. However, the low EQE might be due to recombination at the grain boundaries, sidewalls, interfaces and low absorption coefficient of GaAs films on metal tape, which are not completely understood.
Approaches such as improvement in the defect density, passivation technique, incorporation of light management processes and multi-junction solar cells will further improve the photovoltaic conversion efficiency of the flexible GaAs solar cells on epi- ready metal tape.
6.4 Conclusion
Single-junction GaAs solar cells were successfully grown and fabricated on the sputtered and CVD germanium templates. A detailed analysis was performed on both types of template and GaAs grown on them. The CVD germanium and GaAs grown on them had higher quality and better performance compared to the sputtered germanium and the GaAs grown on it. A conversion efficiency of 11.02% for champion devices at 1 sun was obtained for 1J GaAs solar cells on CVD germanium compared to 6.8% on the
119 sputtered germanium. The higher efficiency on CVD germanium is attributed to improved quality of GaAs, which resulted in lower leakage current and increased VOC.
However, further improvement in film quality is required with higher absorption to enhance efficiency above 11%. For this approach, we incorporated a thick intrinsic layer
(i-layer) GaAs between the p-n junction, resulting in the p-i-n junction solar cell. GaAs film quality and device quality differences between a p-n and p-i-n solar cells will be discussed in Chapter 7.
120
Chapter 7 Improving the efficiency of single-junction GaAs solar cells
by incorporating an intrinsic layer in the p-i-n solar cell
structure
The solar cell structures grown and fabricated on flexible epi-ready metal tape in
Chapter 5 and Chapter 6 were single-junction p-n solar cells. The solar cell efficiency was mainly limited by absorption and leakage/recombination (e.g., from the side wall, grain boundary, defects). In addition to grain boundaries acting as recombination centers in p-n solar cells, threading dislocations, the density of which was on the order of 108 cm-
2 due to a 4% lattice mismatch between Ge and CeO2, and the quality of germanium itself are other sources of defects [149]. These threading dislocations act as diffusion tracks for the dopant. Also, in the case of the p-n solar cell structure on epi-ready metal tape, we observed surface defects in the form of pits and particles. In addition, the minority charge-carrier lifetime of GaAs fabricated at our facility on flexible metal tape is on the order of nanoseconds (~2 ns) with an estimated diffusion length of 1.2 μm as determined from its double hetero-structure [172]. The EQE result obtained for p-n solar cells, as explained in chapter 6 (figure 6.12), with the sharp fall of the EQE curve above the wavelength of 500 nm, also supported the fact of lower absorption in the base due to the short minority carrier diffusion length.
In order to address the above-mentioned issues (e.g., diffusion, surface defects, and absorption), p-n solar cell with an intrinsic layer, that is, p-i-n solar cell structure was developed. The electric field in a p-i-n type structure extends over a wider region compared to the p-n type structure and the increase in potential is the same with the same dop-i-ng level as shown in figure 7.1. This electric field drifts the charge carriers that are
121 photo-generated in the intrinsic layer (towards the contact) such that they can survive for a longer distance. The intrinsic layer thickness also increases the absorption. This type of structure is common for an amorphous silicon solar cell structure where the diffusion length is short [173].
Figure 7. 1 Band profile for p-n and p-i-n structures.
In case of p-i-n solar cells, the thickness of the intrinsic layer (electric field associated due to thickness at a particular bias) plays a critical role to deplete all the photo-generated charge carriers. A thicker intrinsic layer absorbs more light, but the defects in the region reduce the electric field across it. For a thinner intrinsic layer, the absorption might not be sufficient. In addition, the photo-generated charge carriers will not be completely depleted by the electric field at an applied bias. Attention should also be given to the charged background doping of the intrinsic layer, which should be very low. High background doping of the intrinsic layer will split the electric field in the region to the p-i and i-n interfaces due to the formation of a junction region on either side.
For this, the thickness of the intrinsic layer plays an important role and must be optimized.
122
The above-mentioned facts were taken into account in the fabrication of single- junction p-i-n GaAs solar cells. The dop-i-ng concentration in the intrinsic layer was optimized for device growth. Details of the single-junction p-i-n GaAs solar cell structure is shown in figure 7.2.
Figure 7. 2 Schematic of a 1J p-i-n solar cell and buffer layer on flexible metal tape.
The details of the buffer layers used for fabrication of the solar cell have already been explained in chapter 4, section 4.3. The doping concentration and fabrication condition for all of the layers in the p-i-n structure were kept the same as that of p-n solar cells explained in Chapter 4 and 5. The base thickness of the p-i-n structure was varied to obtain maximum current generation. Single-junction p-i-n solar cell devices with base thicknesses of 250 nm, 500 nm, 1000 nm, 1500 nm and 2000 nm were grown by
MOCVD and fabricated using a photolithography process. The fabricated solar cells were
123 measured, and the trend in variation of the device parameter was studied. After optimizing the base thickness, the optimized solar cell was used for other stages (i.e., parasitic layer removal/passivation and ARC) of device fabrication. Different characterization tools were used to test and compare p-n and p-i-n solar cells. A comparative study was conducted between p-n and p-i-n solar cells.
7.1 Optical characterization of p-n and p-i-n solar cell device film
Optical, SEM and AFM images were obtained to determine the nature of the particles on the surface of p-n GaAs film, and were compared with the p-i-n versions as shown in figure 7.3 and 7.4. The optical image showed a higher concentration of random- sized black-spotted feature (defects) resembling pits and particles. To better understand whether the feature that appeared on the surface were pits or particles, a high-resolution imaging technique was used. The average root-mean-square roughness obtained from the
AFM surface topography scan of an area of 40 μm × 40 μm is 180 nm taken over different areas of the sample. The 3D image from a scan over a 10 μm × 10 μm area shows a pit with a depth of approximately ~750 nm. This suggested that many pits were formed on the p-n solar cell surfaces. Each pits is the center for diffusion of particles and a reason for high leakage in p-n solar cells. However, the pit widths and the features inside the pits are still unknown.
124
Figure 7. 3 Optical, SEM and AFM images of p-n films.
Helium ion microscopy was used to image the surface of p-n solar cell films to better understand the widths of the pits as well as their features. Helium ion microscopy is a type of Field Ion Microscopy (FIM) that uses a Gaseous Field Ionization Source
(GFIS) gun with helium gas at the vicinity [174]. It typically operates between 20 keV to
45 keV and can produce a helium beam current in the range of 10-14 to 10-11 A. Changing the background pressure of the imaging gas, the beam current can be modulated [175].
For beam energy in the given range (20-45 keV), the wavelength of the helium ion is several hundred times shorter than that of electron beams of the same energy (used in
SEM), generating high momentum to reduce diffraction. The depth-of-field for HIM is thus 10x to 20x larger than that of SEM [176]. ORION NF equipment (at Oak Ridge
National Laboratory) that produces a beam (convergence angle of 0.5mrad) with a probe size of 0.4 nm and a beam current in picoamperes was used for our experiment. For imaging the p-n solar cell surfaces at a working distance of 17.12 mm, a beam current of
0.589 pA was used for all sizes of images.
125
Figure 7. 4 HIM images of p-n film surfaces. Figure 7. 4 shows the HIM images of p-n solar cell surfaces at different locations.
In the first row of images in Figure 7.4, the first image of 50 µm shows a concentration of spots on the surface. The high magnification images clearly show that such surface spots are pits. The 1- µm image observed from HIM in the second row of Figure 7.4 is clear evidence for the formation of such pits. The widths of those pits varied from approximately 400 nm to 650 nm. This high concentration of pits in the p-n solar cell led to high leakage, low current and lower open circuit voltage. When metal contacts are deposited during the fabrication process, the pit’s area will short the device. The third row of images in Figure 7.4 supports the roughness observed on the surface by AFM of the p-n GaAs film along with the pit’s feature. The growth mode appears to be layer-by- layer with some localized step features leading to a rougher film.
The surface morphology of the grown p-i-n solar cells with different intrinsic thicknesses were of the same nature as shown in optical, SEM and AFM images in figure
126
7.5. The optical and SEM images show a very clean surface with a very low concentration of black spots. The average root-mean-square roughness obtained from a scan area of 40 μm × 40 μm was 50 nm. The 3D image scan of the 10 μm × 10 μm area clearly shows that the surface of the film is clean with no such pits observed within that scan range. This is clear evidence of a cleaner surface in the p-i-n compared to the p-n film. Still, few pits and some particles persisted in the larger area scan.
Figure 7. 5 Optical, SEM and AFM images of p-i-n film.
The HIM analysis of the p-i-n film is shown in figure 7.6. The scan sizes between
250 µm and 1.3 µm at different locations supports the observation from SEM and AFM that the concentration of pits is greatly reduced. HIM surface imaging of p-i-n surfaces was performed at the same working distance and beam current as applied for the p-n sample. The growth mode appears to be layer-by-layer.
127
Figure 7. 6 HIM images of p-i-n film surfaces.
Thus, the p-i-n films showed a smoother morphology than p-n films. As mentioned above, surface analysis of the p-i-n film done for different intrinsic thicknesses showed nearly identical morphologies. These p-i-n device films with different base thicknesses were then fabricated to check the device results. Details on the device result and analysis are shown below.
7.2 Fabrication and characterization of single-junction p-i-n solar cells
Solar cells with films that have p-i-n structures comprising of different intrinsic thicknesses were fabricated. The fabrication process for the p-i-n solar cells were exactly the same as for the p-n solar cells described in chapter 5 (section 5.1), except for the etching process for the different thickness of the intrinsic layers. The flexible samples were pasted on a wafer and processed via lithography and contact deposition. The ohmic contact nature of the fabricated samples was tested using TLM, as shown in Figure 7.7.
The specific contact resistivities of n- and p-type contacts were 1.5 10-4 Ωcm2 and 9
10-5 Ωcm2, respectively. The detailed calculation is shown in the inset of Figure 7.7.
128
Comparing the p-n device film from chapter 5, a lower contact resistance (12 Ω in p-n and 1.35 Ω in p-i-n) and sheet resistance (360 Ω/sq in p-n and 31.3 Ω/sq in p-i-n) were observed for the p-i-n device film. This is attributed to a smooth p-i-n surface with negligible surface pits and particles compared to the p-n surface.
(a) (b)
Figure 7. 7 Transmission line method (TLM) for (a) top and (b) bottom TLM.
After measuring the TLM, the solar cells’ efficiencies were measured from J-V curves. An NREL-calibrated Oriel 200 series solar simulator at 1 sun was used to obtain illuminated J-V behavior. Figure 7.8 shows the illuminated curve for p-i-n solar cells with intrinsic layer thicknesses of 250 nm, 500 nm, 1000 nm, 1500 nm and 2000 nm. The parameter obtained from the J-V curve is shown in Table 7.1. All devices were measured after the initial stage of device fabrication.
129
18
) 16 2 14
12
mA/cm ( 10 250nm 8 500nm 6 1000nm 4 1500nm Current Density Current 2 2000nm 0 0 100 200 300 400 500 600 Voltage(mV) Figure 7. 8 Illuminated J-V characteristics of the fabricated p-i-n GaAs solar cells.
Table 7. 1 VOC, JSC, FF and η of p-i-n solar cells with different intrinsic layer thicknesses
Intrinsic layer VOC JSC Fill Efficiency thickness (nm) (mV) (mA/cm2) Factor (%) (%) 250 334 13.8 0.51 2.8 500 568 14.7 0.67 5.7 1000 615 17.4 0.71 7.6 1500 564 15 0.63 5.6 2000 567 13.2 0.66 5.0
As observed from the above J-V curves in Figure 7.8 and Table 7.1, the VOC, JSC,
FF and η values all followed certain patterns as the thickness of intrinsic layer was varied. The best result is observed for the intrinsic layer of 1000 nm. It clearly signifies the importance of appropriate intrinsic layer thickness. Figure 7.9 shows the variation of
VOC, JSC and FF values for different intrinsic layer thicknesses.
130
Figure 7. 9 Variation of JSC, VOC and FF for p-i-n cells with different intrinsic layer thicknesses.
As observed from Figure 7.9, the JSC increased as the intrinsic layer thickness increased from 250 nm; it reached a maximum value at 1000 nm and then decreased for thicker intrinsic layers (1500 and 2000 nm, respectively). As explained previously, the lower JSC for the thinner base is due to insufficient absorption of light. As the thickness increases, the absorption increases initially but then decreases for thicker bases due to increased defects (low carrier lifetimes in thicker base limit the diffusion length). The open circuit voltage and fill factor is maximum for 1000 nm and minimum for 250 nm intrinsic thicknesses. This effect can be explained from dark I-V characteristics for different intrinsic layer thicknesses.
0 2 4 6 8 10 -1 250nm -1 250nm 500nm 500nm 1000nm -2 (a) 1000nm -2 1500nm 1500nm 2000nm -3 2000nm -3
-4
log(Current) -4 log(Current) -5 -5 (b) -6 -6 -1000 -750 -500 -250 0 250 500 750 100 200 300 400 500 600 700 Voltage (mV) Voltage(mV) Figure 7. 10 Dark I-V characteristics of the fabricated p-i-n GaAs solar cell.
131
Figure 7.10 (a) shows the log-scale plot of the dark I-V curves of p-i-n cells with different intrinsic layer thicknesses, and Figure 7.10 (b) shows the same plot with the component of forward-biased dark current. The dark curve in the forward-biased case comprises two regions (i.e., the lower voltage and higher voltage regions) distinguished by the indicated vertical line. The performance can be approximated by two diodes in parallel with parasitic resistance Rs and can be expressed as
V−퐽 R V−퐽 R 푑푎푟푘 s 푑푎푟푘 s Jdark(V) = J01 [exp{q ( kT ) } − 1] − J02 [exp{q ( 2kT )} − 1], (7.1) where J01 and J02 are the saturation current densities (recombination component) at higher and lower voltage regions, respectively. T is the operating temperature (25 °C for standard solar cell test condition), q is electronic charge and k is Boltzmann’s constant. In the low-voltage region, the second diode with the 2kT component dominates, and in the high-voltage region, the first diode with the 1kT component dominates. The low-voltage region in the forward-bias case gives information on the recombination at the space charge region (SCR), and the high-voltage region gives the information on recombination at the quasi-neutral region (QNR). The leakage current calculated from the dark I-V plot for different intrinsic layer thicknesses is tabulated in Table 7.2 below.
Table 7. 2 Leakage current (IO) of p-i-n solar cells with different intrinsic layer thicknesses.
Intrinsic layer 250 500 1000 1500 2000 thickness (nm) -3 -3 -5 -4 -5 Leakage 4 10 2 10 8 10 1 10 6 10 current (Io) (mA)
132
From the above Table 7.2, the leakage current in the range of 10-5 mA was obtained for a cell with an intrinsic layer thickness of 1000 nm, whereas, for a cell with an intrinsic layer thickness less than 1000 nm, the leakage current was two orders of magnitude higher. For a cell with an intrinsic layer thickness of 1500 nm, the leakage current increases by an order of magnitude from that for the 1000 nm cell. The leakage current decreases in the cell with an intrinsic layer thickness of 2000 nm. To understand the trend in leakage current due to the intrinsic layer thickness, the dark current-voltage curve was fitted with a double-diode model, as explained in equation 7.1. The double- diode fit at the higher-voltage region for an intermediate base thickness of 1000 nm and thickness of 250 nm is shown in Figures 7.11 (a) and (b), respectively.
(b) (a)
Figure 7. 11 Forward-bias component of dark curve fit.
Table 7. 3 Saturation current density from double-diode fitting.
Intrinsic Layer 250 500 1000 1500 2000 thickness (nm) 2 -9 -9 -10 -9 -9 J01 (A/cm ) 2.3×10 2.6×10 3.9×10 5.8×10 2.5×10 2 -7 -8 -9 -8 -8 J02 (A/cm ) 6.3×10 4.4×10 3.1×10 1.42×10 1.1×10
133
Table 7.3 shows the extracted saturation current density obtained from the fit for different intrinsic layer thicknesses. From the double diode fit in Figure 7.11(a), the extracted values of J01 and J02 for an intermediate intrinsic layer thickness of 1000 nm are
3.9×10-10A/cm2 and 3.1×10-9A/cm2, respectively. As observed from the leakage current, for an intrinsic layer of thickness 250 nm, the leakage current is two orders of magnitude higher than that of the cell with 1000 nm intrinsic layer thickness. This can be explained from the above-mentioned fact (illuminated J-V showing lower current) that the field generated within the intrinsic layer of thickness 250 nm is not sufficient to deplete the photo-generated charge carriers generated. So, these charge carriers recombine before collection, causing higher leakage. The forward-biased region of the dark curve, as shown in Figure 7.10 (b), also supports the higher recombination mechanism with the intrinsic layer thickness of 250 nm. From the double-diode fit in Figure 7.11(b), the
- extracted value of J01 and J02 values for a thinner intrinsic layer of 250 nm are 2.3×10
9 2 -7 2 A/cm and 6.3×10 A/cm , respectively. The value of J01 is one order of magnitude higher, and the value of J02 is two orders of magnitude higher than that for intrinsic layer thickness of 1000 nm. This clearly shows that the recombination in the space-charge region and the quasi-neutral region is higher for an intrinsic layer thickness of 250 nm.
We observed a similar trend for an intrinsic layer with a thickness of 500 nm. The
-9 2 -8 2 J01 and J02 values were 2.6×10 A/cm and 4.4×10 A/cm , respectively. The value of J01 is thus comparable to the cell of 250-nm thick intrinsic layer, whereas the value of J02 is one order of magnitude higher than the 1000-nm thick layer cell and one order of magnitude lower than the 250-nm thick layer cell. Thus, the recombination in SCR for
500-nm thick layer cell is higher than that of the 1000-nm thick layer cell but lower than
134 that with the 250-nm thick layer cell. This is supported by the fact that the leakage current follows the same trend.
As the thickness of the intrinsic layer increases from 1000 nm to 1500 nm, the charged defects increase, and this reduces the net electric field in the region and increases
-9 2 the recombination. This fact is also supported by higher values of J01 (5.8×10 A/cm )
-8 2 and J02 (1.42×10 A/cm ) for the 1500 nm thick layer cell compared to the 1000-nm thick layer cell, resulting in higher leakage.
As the intrinsic layer thickness increases to 2000 nm, the net electric field increases, which is higher than in the cell with intermediate layer thickness. However, the recombination due to the charged defects increases. Due to this trade off nature, a lower leakage was observed but with reduced current in the cell with an intrinsic layer thickness of 2000 nm. Also, careful comparison of the low-voltage region on the forward-biased
-8 2 side shows that the value of J02 is higher (1.1×10 mA/cm ) for the 2000-nm-thick intrinsic layer compared to the 1000 nm thick layer (3.1×10-9 mA/cm2), indicating higher recombination occurring at the space-charge region.
Therefore, the optimal thickness for the intrinsic layer is 1000 nm, since this thickness exhibited the lowest leakage current and highest values of VOC, JSC, FF and η.
To further enhance the device efficiency, p-i-n solar cells with this intermediate thickness
(1000nm) of the intrinsic layer underwent further steps of device fabrication (cap layer removal/passivation and ARC).
135
) 2 28 (a) -1 Initial As Fabricated 24 Cap layer removal (pasivation) -2 Cap layer Removal (Passivation) ARC mA/cm 20 ARC ( -3 16 (b) VOC= 634/634/654mV 12 -4 J = 19/21.2/27.6mA/cm2 SC log(Current) -5 8 FF = 71/71/73%
4 Eff. = 8.5/9.6/13.2% -6 Current Density Current 0 0 100 200 300 400 500 600 700 -1000 -500 0 500 Voltage(mV) Voltage(mV)
Figure 7. 12 (a) Illuminated and (b) dark current-voltage characteristic of a p-i-n solar cell with a 1000-nm intrinsic layer.
Figure 7.12 shows the illuminated J-V characteristic for a fabricated p-i-n solar cell with an intrinsic layer thickness of 1000 nm at three different stages of device fabrication. For cap layer removal, a 1:1 mixture of citric acid (C6H8O7) and hydrogen peroxide (H2O2) at room temperature was used. After cap layer removal, the JSC of solar cell increased from 19 mA/cm2 to 21.2 mA/cm2, whereas the leakage current remained the same. To further enhance the current density by minimizing the surface reflection,
ARC was deposited. Zinc sulfide (48 nm) and magnesium fluoride (96 nm) were used as
ARC for consistency with p-n solar cells. The short circuit current density further increased after ARC coating from 21.2 to 27.6 mA/cm2, keeping the leakage current constant. Table 7.4 provides details regarding the extracted parameters for solar cells at three different stages of device fabrication. The fill factor of the solar cell at different stages of device fabrication remained almost constant. The leakage current remained almost constant with a value of 5× 10-5mA.
136
Table 7. 4 Io, VOC, JSC, FF and η of solar cells at the initial stage of fabrication, after cap layer removal and after ARC application.
Stage of Device VOC JSC Fill Efficiency Leakage current 2 Fabrication (mV) (mA/cm ) Factor (%) (Io) (mA) (%) As Fabricated 633 19.2 71 8.5 5 10-5 Cap Layer Removed 633 21.2 71 9.6 5 10-5 ARC Applied 653 27.6 73 13.2 5 10-5
An efficiency of 13.2% was obtained for p-i-n solar cells with an intrinsic thickness of 1000 nm. To better understand the p-i-n stack, a cross-sectional transmission electron microscopy (TEM) analysis was performed. We used a p-i-n solar cell device with an optimized intrinsic layer thickness of 1000 nm. Analysis was also done for a p-n sample, and the two samples were compared.
7.3 Transmission electron microscopy analysis for p-i-n and p-n solar cell stacks
First, the sample for the cross-sectional TEM was prepared using a FEI 235 dual beam focused ion beam system. The prepared sample was transferred to a TEM chamber.
Transmission electron cross-sectional imaging was carried out using a JEOL JEM
2000FX microscopy at 200 keV. Figure 7.13 shows the complete architecture of p-n and p-i-n starting from the germanium buffer. The multilayered architecture consists of germanium, semiconductor buffer comprising of 2-μm undoped GaAs/ 30period of superlattice (GaAs, 5 nm/ InGaAs, 5 nm)/ 1.2μm of undoped GaAs/ 30period of superlattice/ 1.2 μm of undoped GaAs, and a single-junction solar cell stack with p-n and p-i-n solar cell structures (active layer). A high density of defects was observed in the germanium film, starting from the interface of cerium oxide/ germanium, which is due to
137 a 4.2% lattice mismatch between the two layers as observed in previous observation
[148, 150, 177]. The cross section analyzed by TEM shows that the incorporation of the strained layer superlattice reduces the density of threading dislocation on both the p-n and p-i-n structures. This commonly used technique including a strained layer superlattice is capable of bending over the dislocation lines, creating barriers against the dislocation
[178, 179]. As observed from the image, many dislocations are terminated or reflected at the superlattice, which result in a low dislocation density in the upper part of film above the superlattice. The active region comparison between the p-n and p-i-n shows that the active region of the p-i-n GaAs is relatively clean with fewer defects.
Figure 7. 13 TEM analysis for cross-sections of the p-n and p-i-n devices.
The above TEM analysis showed that cleaner surfaces with less defects are obtained via introduction of an intrinsic layer for p-i-n type solar cell structures in comparison to p-n. The dislocations observed in the p-n and p-i-n structures acts as a potential diffusion path for the dopant. To analyze the elemental species and diffusion of
138 dopant, if any, time-of-flight secondary ion mass spectroscopy (TOF SIMS) was conducted for both p-n and p-i-n samples.
7.4 Time-of-Flight SIMS analysis for p-i-n and p-n films
Time-of-flight secondary ion mass spectrometry (TOF-SIMS) was used to determine elemental species present in p-n and p-i-n device films. TOF-SIMS is an analytical technique that focuses a pulsed beam of primary ions to produce secondary ions from the surface in a sputtering process. TOF-SIMS measurements in this study
+ were performed using TOF.SIMS.5 NSC instrument (ION.TOF Gmb). Bi3 liquid metal ion gun (energy 30 keV, current 30 nA and spot size 5 m) was used as a primary source rastered over an area of 100 100 m for extraction of the secondary analyte ions. An additional Cs+ ion gun (energy 1 keV, current 75 nA and spot size ~25 m) rastered over
300 300 m was used as a sputter source for depth profiling. A time-of-flight mass analyzer was operated in positive ion detection mode and enabled mass resolution m/Δm=2,000 - 10,000. Measurements were performed in a non-interlaced mode, where
+ + each analysis scan with Bi3 were followed by 5 s of the sputtering with Cs . Chemical data were further averaged in the x-y direction and represented as 1D depth profiles for different chemical elements (e.g., Ga+, Al+, As+).
139
( (a)
(b)
( (c)
Figure 7. 14 Depth profile analysis of elements in p-n (a) and p-i-n (b) films. Comparison of Zn profiles in p-n and p-i-n films are shown in (c)
Figure 7.14 (a) shows the depth profile of gallium, arsine and zinc up to 2500 nm in p-n films. Gallium and arsine appear throughout the etch depth, whereas the zinc distribution starts at 80 nm (at n+ GaAs) and stays constant to 800 nm (at p GaAs/base), after which its intensity increases to a depth of 850 nm (at p GaAs/base). From depths of
850 nm to 1970 nm (at p++ GaAs), the Zn level remains constant and then the intensity decreases. This shows that a small amount of zinc diffused in to the emitter region which starts from depths of 60 nm up to 133 nm (approx.). Thus, this disturbs the p-n junction region (i.e., the depletion width decreases with decreases in the effective electric field) leading to lower VOC. This depth of each layer is approximately determined with an error margin of ±5 nm.
140
Figure 7.14 (b) shows the depth profile of a p-i-n device film. The distribution of gallium and arsinic in the p-i-n cell is the same as that in the p-n cell. The zinc distribution starts from 1080 nm, which is almost the depth after the intrinsic layer. This clarifies that there is no zinc diffusion above or to the intrinsic layer, and this layer helps to stop the diffusion. The intensity of the zinc profile increases from 1080 nm
(approximately after intrinsic layer) and becomes constant after 1130 nm (after p-GaAs) up to 2400 nm. Figure 7.13 (c) shows the comparison between the zinc distribution depth profiles in p-n and p-i-n devices. The results explain why leakage current in the case of a p-n junction (~10-3mA) is higher than that of a p-i-n junction (~10-5mA). In the case of a p-n junction, due to the diffusion of the p-dopant (zinc), the effective depletion width
(electric field) decreases, which increases the tunneling current near the depletion layer and causes more leakage. With an optimized thickness of the intrinsic layer, the diffusion of species completely stops with overall improvement in device efficiency. Further, to understand the absorption difference in the p-n and p-i-n devices, the photoelectric response was analyzed.
7.5 Photoelectric response for p-n and p-i-n solar cells
The photoelectric response for p-n and p-i-n solar cells under zero bias and up to the wavelength of 700 nm is shown in figure 7.15. For p-n solar cells, the photocurrent increases with wavelength and reaches its maximum at a wavelength of 550 nm. Then, it remains constant until 600 nm and rapidly decreases. For p-i-n solar cells, the photocurrent increases with wavelength and reaches its maximum at 660 nm. Then, it remains constant until 700 nm. The maximum photo-response is found to increase to longer wavelengths in the case of the p-i-n solar cell structure. The current response for
141 both p-n and p-i-n structures at the lower-wavelength region is due to insufficient extraction of carriers in window and emitter layers.
Figure 7. 15 Photoelectric response for p-n and p-i-n solar cells.
The constant photo response at longer wavelengths is attributed to the longer lifetimes of carriers in p-i-n solar cells compared to p-n solar cells . This point is also supported by the fact that the current density for p-i-n solar cells is higher than that of p-n solar cells. Thus, the overall efficiency of p-i-n solar cells exceeds the efficiency of p-n solar cells.
7.6 Conclusion
In order to improve the device efficiency of p-n solar cells, single-junction GaAs solar cells with p-i-n structures comprising different intrinsic layer thicknesses (250 nm,
500 nm, 1000 mm, 1500 nm and 2000 nm) were successfully grown on flexible metal buffers. The p-i-n films were fabricated to obtain an optimized thickness of the intrinsic layer. Data obtained from the I-V characteristic showed that the cell with an intrinsic layer of thickness 1000 nm had the least leakage current with the highest values of VOC and JSC compared to cells with other intrinsic thicknesses. The solar cell device with an
142 intrinsic layer of thickness of 1000 nm was further processed with another step of device fabrication (i.e., cap layer removal and ARC). An efficiency of 13.2% was obtained for the optimized intrinsic thickness. A comparison between the surface morphology of p-n and p-i-n solar cell devices using AFM, SEM and HIM revealed that p-i-n solar cells were free of pits and particles. An analysis of cross-sectional TEM images showed the diffusion of zinc in the p-n film, while it was not observed for the p-i-n film due to the intrinsic layer, which stopped diffusion. This improved the leakage current in the p-i-n film. The optical response of the p-i-n solar cell device with an intrinsic layer of thickness
1000 nm showed less recombination and absorption to longer wavelengths compared to the response of its p-n counterpart. In conclusion, p-i-n solar cells with optimized intrinsic thickness resulted in better morphology, device results and showed a better pathway to improve solar cell efficiency than p-n solar cells.
143
Chapter 8 Summary and Future Work
Despite the high efficiency of III-V solar cells based on GaAs, their usage in large-scale terrestrial application is very limited due to the excessive cost of GaAs and
Germanium (Ge) wafers. This work was aimed to develop GaAs solar cells on high- quality epitaxial semiconductor thin films on inexpensive flexible metal tapes to overcome the wafer cost as well as benefit from the lower manufacturing cost by roll-to- roll processing.
Metal organic chemical vapor deposition (MOCVD) is used to grow high-quality epitaxial GaAs solar cell structure on ‘single-crystalline-like’ Ge film on epi-ready metal tape. The epi-ready buffer comprised of Magnetron sputtered Ge/CeO2/LaMnO3/ homo epi-MgO, Ion-Beam Assisted Deposition (IBAD) MgO/Y2O3 as seed layer and Al2O3 as diffusion barrier on Hastelloy (C-276). A thick epitaxial undoped semiconductor GaAs buffer and various defect reduction techniques such as InGaAs/GaAs superlattice (SL), two step GaAs growth and thermal annealing were incorporated using MOCVD to significantly reduce defects prior to the growth of the active layers. Single junction (1J) p-n solar cell structures comprising of aluminum gallium arsenide (AlxGa1-xAs) front/back surface field (BSF), window and highly-doped GaAs cap layer were grown.
These grown thin films were further processed via photo-lithography, etching and contact deposition to fabricate 1J GaAs solar cell devices. Top contact shadowing was minimized to increase current density by special designed extended contact. To boost light absorption, cap layer etching using citric acid was performed which resulted in passivation of the GaAs surface. To reduce the reflection from the GaAs solar cell surfaces, anti-reflection coating (ZnS/ MgF2) was deposited. Single-junction GaAs p-n 144 devices on flexible metal tapes with 6.8% efficiency was obtained at 1 sun with open
-2 circuit voltage (VOC) of 566 mV, current density (JSC ) of 17.4 mA cm , and fill factor
(FF) of 68%. Citric acid, which is commonly used for etching GaAs, acted as an effective passivant during device fabrication. To understand the impact of citric acid passivation,
GaAs solar cell with different base thickness (1140nm, 840nm and 380nm) were grown and fabricated. Comparative passivation study on solar cells with different base thickness
(1140nm, 840nm and 380nm) revealed that the passivation was more effective in reducing the leakage current (and corresponding increase in Voc) in cells with thinner base thickness. However, this resulted only in surface and side wall passivation. For bulk passivation hydrogen and phosphine-incorporated-hydrogen plasma treatment was applied on GaAs film. In-depth analysis were conducted to understand the influence of passivation on the structural and opto-electronic properties.
To further enhance efficiency of 1J GaAs solar cells the quality of the sputtered
Ge template was improved via roll-to roll-processed Ge using chemical vapor deposition
(CVD). Power conversion efficiency of 11.5% was achieved in these devices at 1 sun
−2 with VOC of 642 mV, JSC of 25 mA cm , and FF of 72%. The improved efficiency is attributed to high quality GaAs film on roll-to-roll Ge.
Further, the quality of p-n junction GaAs films was improved by incorporating an intrinsic layer with p-i-n solar cell structure. Gallium arsenide solar cells with intrinsic layer thickness 250nm, 500nm, 1000nm, 1500nm and 2000nm were grown using
MOCVD. The grown solar cells were then fabricated and tested. The intermediate thickness of 1000nm resulted in best device with higher VOC, JSC, FF and efficiency compared to other intrinsic layer thickness. Fabricated p-i-n solar cells (intrinsic layer
145 thickness of 1000nm) after cap layer removal and ARC coating at 1 sun after ARC
−2 yielded a device efficiency of 13.2% with VOC of 654 mV, JSC of 27.6 mA cm , and FF of 73 %. The efficiency improvement is attributed to improved quality of GaAs film and increased absorption due to optimized intrinsic layer thickness.
To further improve the efficiency of single junction GaAs solar cell on the flexible template, the future work could be focused on:-
1) Optimization of roll-to-roll CVD germanium substrate and growth of p-i-n
solar cell on CVD germanium as this study focused only on the p-i-n solar cell
on sputtered germanium template.
2) Increasing the grain size of GaAs as the smaller grain size consisting of
numerous grain boundaries act as sites for recombination resulting in higher
leakage. For this purpose, an alternative template such as nickel tungsten
(NiW) with larger grain size can be used.
3) Implementing GaAs on flexible substrate for other applications such as high
frequency integrated circuits, monolithic microwaves integrated circuits etc.
4) As the flexible substrate consists of oxide buffer it can be used to fabricate
microcells consisting of many small solar cells which can be joined in series
to get high voltage output.
5) Using light concentration with solar cell on metal templates. Recently, PDMS
microlens has been used as a microconcentrator which needs further
optimization. This can also be implied on the microcells connected in series to
get better performance.
146
References:
[1] M. A. Green, K. Emery, Y. Hishikawa, W. Warta, and E. D. Dunlop, "Solar Cell
Efficiency Tables (Version 45)," Progress in photovoltaics: research and
applications, vol. 23, pp. 1-9, 2015.
[2] C. E. Fritts, "On a New Form of Selenium Cell, and Some Electrical Discoveries
Made by Its Use," American Journal of Science, pp. 465-472, 1883.
[3] S. Sharma, K. K. Jain, and A. Sharma, "Solar Cells: In Research and
Applications—a Review," Materials Sciences and Applications, vol. 6, p. 1145,
2015.
[4] Q. An, F. Zhang, J. Zhang, W. Tang, Z. Deng, and B. Hu, "Versatile Ternary
Organic Solar Cells: A Critical Review," Energy & Environmental Science, vol.
9, pp. 281-322, 2016.
[5] R. Bergmann, "Crystalline Si Thin-Film Solar Cells: A Review," Applied physics
A, vol. 69, pp. 187-194, 1999.
[6] J. Gong, J. Liang, and K. Sumathy, "Review on Dye-Sensitized Solar Cells
(Dsscs): Fundamental Concepts and Novel Materials," Renewable and
Sustainable Energy Reviews, vol. 16, pp. 5848-5860, 2012.
[7] M. Wright and A. Uddin, "Organic—Inorganic Hybrid Solar Cells: A
Comparative Review," Solar energy materials and solar cells, vol. 107, pp. 87-
111, 2012.
[8] W. Zulehner, "Czochralski Growth of Silicon," Journal of Crystal Growth, vol.
65, pp. 189-213, 1983.
147
[9] M. Tilli, M. Paulasto-Krockel, T. Motooka, and V. Lindroos, Handbook of
Silicon Based Mems Materials and Technologies: William Andrew, 2015.
[10] P. Rai‐Choudhury and P. Hower, "Growth and Characterization of Polycrystalline
Silicon," Journal of The Electrochemical Society, vol. 120, pp. 1761-1766, 1973.
[11] A. Shah, P. Torres, R. Tscharner, N. Wyrsch, and H. Keppner, "Photovoltaic
Technology: The Case for Thin-Film Solar Cells," science, vol. 285, pp. 692-698,
1999.
[12] B. M. Kayes, H. Nie, R. Twist, S. G. Spruytte, F. Reinhardt, I. C. Kizilyalli, and
G. S. Higashi, "27.6% Conversion Efficiency, a New Record for Single-Junction
Solar Cells under 1 Sun Illumination," in 2011 37th IEEE Photovoltaic Specialists
Conference, 2011, pp. 000004-000008.
[13] B. E. McCandless and J. R. Sites, "Cadmium Telluride Solar Cells," Handbook of
Photovoltaic Science and Engineering, pp. 617-662, 2003.
[14] R. W. Miles, G. Zoppi, and I. Forbes, "Inorganic Photovoltaic Cells," Materials
today, vol. 10, pp. 20-27, 2007.
[15] D. E. Carlson and C. R. Wronski, "Amorphous Silicon Solar Cell," Applied
Physics Letters, vol. 28, pp. 671-673, 1976.
[16] H. Hoppe and N. S. Sariciftci, "Organic Solar Cells: An Overview," Journal of
materials research, vol. 19, pp. 1924-1945, 2004.
[17] A. Pivrikas, N. S. Sariciftci, G. Juška, and R. Österbacka, "A Review of Charge
Transport and Recombination in Polymer/Fullerene Organic Solar Cells,"
Progress in Photovoltaics: Research and Applications, vol. 15, pp. 677-696, 2007.
148
[18] V. Sugathan, E. John, and K. Sudhakar, "Recent Improvements in Dye Sensitized
Solar Cells: A Review," Renewable and Sustainable Energy Reviews, vol. 52, pp.
54-64, 2015.
[19] K. Chopra, P. Paulson, and V. Dutta, "Thin‐Film Solar Cells: An Overview,"
Progress in photovoltaics: research and applications, vol. 12, pp. 69-92, 2004.
[20] W. Guter, J. Schöne, S. P. Philipps, M. Steiner, G. Siefer, A. Wekkeli, E. Welser,
E. Oliva, A. W. Bett, and F. Dimroth, "Current-Matched Triple-Junction Solar
Cell Reaching 41.1% Conversion Efficiency under Concentrated Sunlight,"
Applied Physics Letters, vol. 94, p. 223504, 2009.
[21] D. L. Meier and A. Rahatgi, "Solar Cell Having Crystalline Silicon Pn
Homojunction and Amorphous Silicon Heterojunctions for Surface Passivation,"
ed: Google Patents, 2009.
[22] V. D. Mihailetchi, Y. Komatsu, and L. Geerligs, "Nitric Acid Pretreatment for the
Passivation of Boron Emitters for N-Type Base Silicon Solar Cells," Applied
Physics Letters, vol. 92, p. 063510, 2008.
[23] D. Macdonald and L. Geerligs, "Recombination Activity of Interstitial Iron and
Other Transition Metal Point Defects in P-and N-Type Crystalline Silicon,"
Applied Physics Letters, vol. 85, pp. 4061-4063, 2004.
[24] A. Cuevas, M. J. Kerr, C. Samundsett, F. Ferrazza, and G. Coletti, "Millisecond
Minority Carrier Lifetimes in N-Type Multicrystalline Silicon," Applied Physics
Letters, vol. 81, pp. 4952-4954, 2002.
[25] R. Sahai and A. Milnes, "Heterojunction Solar Cell Calculations," Solid-State
Electronics, vol. 13, pp. 1289-1299, 1970.
149
[26] R. H. Parekh and A. M. Barnett, "Improved Performance Design of Gallium
Arsenide Solar Cells for Space," IEEE Transactions on electron devices, vol. 31,
pp. 689-695, 1984.
[27] K. Lee, J. D. Zimmerman, T. W. Hughes, and S. R. Forrest, "Non‐Destructive
Wafer Recycling for Low‐Cost Thin‐Film Flexible Optoelectronics," Advanced
Functional Materials, vol. 24, pp. 4284-4291, 2014.
[28] E. F. Schubert, Doping in Iii-V Semiconductors: E. Fred Schubert, 2015.
[29] O. Von Roos, "A Simple Theory of Back Surface Field (Bsf) Solar Cells," Journal
of Applied Physics, vol. 49, pp. 3503-3511, 1978.
[30] V. A. Dao, J. Heo, H. Choi, Y. Kim, S. Park, S. Jung, N. Lakshminarayan, and J.
Yi, "Simulation and Study of the Influence of the Buffer Intrinsic Layer, Back-
Surface Field, Densities of Interface Defects, Resistivity of P-Type Silicon
Substrate and Transparent Conductive Oxide on Heterojunction with Intrinsic
Thin-Layer (Hit) Solar Cell," Solar Energy, vol. 84, pp. 777-783, 2010.
[31] A. Kaminski, B. Vandelle, A. Fave, J. Boyeaux, R. Monna, D. Sarti, and A.
Laugier, "Aluminium Bsf in Silicon Solar Cells," Solar Energy Materials and
Solar Cells, vol. 72, pp. 373-379, 2002.
[32] B. Galiana, I. Rey-Stolle, M. Baudrit, I. García, and C. Algora, "A Comparative
Study of Bsf Layers for Gaas-Based Single-Junction or Multijunction
Concentrator Solar Cells," Semiconductor science and technology, vol. 21, p.
1387, 2006.
[33] H. Berger, "Contact Resistance and Contact Resistivity," Journal of the
Electrochemical Society, vol. 119, pp. 507-514, 1972.
150
[34] K. Bouzidi, M. Chegaar, and A. Bouhemadou, "Solar Cells Parameters Evaluation
Considering the Series and Shunt Resistance," Solar Energy Materials and Solar
Cells, vol. 91, pp. 1647-1651, 2007.
[35] J. Nelson, The Physics of Solar Cells: World Scientific Publishing Company,
2003.
[36] H. Sakai, T. Yoshida, T. Hama, and Y. Ichikawa, "Effects of Surface Morphology
of Transparent Electrode on the Open-Circuit Voltage in a-Si: H Solar Cells,"
Japanese Journal of Applied Physics, vol. 29, p. 630, 1990.
[37] D. Arumí, R. Rodríguez-Montañés, J. Figueras, S. Eichenberger, C. Hora, and B.
Kruseman, "Gate Leakage Impact on Full Open Defects in Interconnect Lines,"
IEEE Transactions on Very Large Scale Integration (VLSI) Systems, vol. 19, pp.
2209-2220, 2011.
[38] T. R. Johnson, G. Ganguly, G. S. Wood, and D. E. Carlson, "Investigation of the
Causes and Variation of Leakage Currents in Amorphous Silicon Pin Diodes,"
MRS Online Proceedings Library Archive, vol. 762, 2003.
[39] M. A. Green, "Accuracy of Analytical Expressions for Solar Cell Fill Factors,"
Solar Cells, vol. 7, pp. 337-340, 1982.
[40] C. Riordan and R. Hulstron, "What Is an Air Mass 1.5 Spectrum?(Solar Cell
Performance Calculations)," in IEEE Conference on Photovoltaic Specialists,
1990, pp. 1085-1088.
[41] D. L. King, J. A. Kratochvil, and W. E. Boyson, "Measuring Solar Spectral and
Angle-of-Incidence Effects on Photovoltaic Modules and Solar Irradiance
151
Sensors," in Conference Record of the Twenty Sixth IEEE Photovoltaic
Specialists Conference-1997, 1997, pp. 1113-1116.
[42] T. Tiedje, E. Yablonovitch, G. D. Cody, and B. G. Brooks, "Limiting Efficiency
of Silicon Solar Cells," IEEE Transactions on electron devices, vol. 31, pp. 711-
716, 1984.
[43] L. C. Hirst and N. J. Ekins‐Daukes, "Fundamental Losses in Solar Cells,"
Progress in Photovoltaics: Research and Applications, vol. 19, pp. 286-293, 2011.
[44] C. Strümpel, M. McCann, G. Beaucarne, V. Arkhipov, A. Slaoui, V. Švrček, C.
Del Cañizo, and I. Tobias, "Modifying the Solar Spectrum to Enhance Silicon
Solar Cell Efficiency—an Overview of Available Materials," Solar energy
materials and solar cells, vol. 91, pp. 238-249, 2007.
[45] J. Zhao and M. A. Green, "Optimized Antireflection Coatings for High-Efficiency
Silicon Solar Cells," IEEE transactions on electron devices, vol. 38, pp. 1925-
1934, 1991.
[46] D. Bouhafs, A. Moussi, A. Chikouche, and J. Ruiz, "Design and Simulation of
Antireflection Coating Systems for Optoelectronic Devices: Application to
Silicon Solar Cells," Solar Energy Materials and Solar Cells, vol. 52, pp. 79-93,
1998.
[47] H. Sai, Y. Kanamori, K. Arafune, Y. Ohshita, and M. Yamaguchi, "Light
Trapping Effect of Submicron Surface Textures in Crystalline Si Solar Cells,"
Progress in Photovoltaics: Research and Applications, vol. 15, pp. 415-423, 2007.
152
[48] Y.-J. Lee, D. S. Ruby, D. W. Peters, B. B. McKenzie, and J. W. Hsu, "Zno
Nanostructures as Efficient Antireflection Layers in Solar Cells," Nano letters,
vol. 8, pp. 1501-1505, 2008.
[49] J. Oh, H.-C. Yuan, and H. M. Branz, "An 18.2%-Efficient Black-Silicon Solar
Cell Achieved through Control of Carrier Recombination in Nanostructures,"
Nature nanotechnology, vol. 7, p. 743, 2012.
[50] A. Cuevas, "The Recombination Parameter J0," Energy Procedia, vol. 55, pp. 53-
62, 2014.
[51] G. Hurkx, D. Klaassen, and M. Knuvers, "A New Recombination Model for
Device Simulation Including Tunneling," IEEE Transactions on electron devices,
vol. 39, pp. 331-338, 1992.
[52] S. Rühle, "Tabulated Values of the Shockley–Queisser Limit for Single Junction
Solar Cells," Solar Energy, vol. 130, pp. 139-147, 2016.
[53] T. Trupke and P. Würfel, "Fundamental Limits of Solar Energy Conversion,"
Photovoltaic Modeling Handbook, pp. 7-25, 2018.
[54] S. Kurtz, D. Myers, W. McMahon, J. Geisz, and M. Steiner, "A Comparison of
Theoretical Efficiencies of Multi‐Junction Concentrator Solar Cells," Progress in
Photovoltaics: research and applications, vol. 16, pp. 537-546, 2008.
[55] K. Tanabe, "A Review of Ultrahigh Efficiency Iii-V Semiconductor Compound
Solar Cells: Multijunction Tandem, Lower Dimensional, Photonic up/Down
Conversion and Plasmonic Nanometallic Structures," Energies, vol. 2, pp. 504-
530, 2009.
153
[56] N. Asim, K. Sopian, S. Ahmadi, K. Saeedfar, M. Alghoul, O. Saadatian, and S. H.
Zaidi, "A Review on the Role of Materials Science in Solar Cells," Renewable
and Sustainable Energy Reviews, vol. 16, pp. 5834-5847, 2012.
[1] S. R. Bull, "Renewable Energy Today and Tomorrow," Proceedings of the IEEE,
vol. 89, pp. 1216-1226, 2001.
[2] N. Panwar, S. Kaushik, and S. Kothari, "Role of Renewable Energy Sources in
Environmental Protection: A Review," Renewable and Sustainable Energy
Reviews, vol. 15, pp. 1513-1524, 2011.
[3] N. S. Lewis, "Toward Cost-Effective Solar Energy Use," science, vol. 315, pp.
798-801, 2007.
[4] M. Thirugnanasambandam, S. Iniyan, and R. Goic, "A Review of Solar Thermal
Technologies," Renewable and sustainable energy reviews, vol. 14, pp. 312-322,
2010.
[5] N. S. Lewis and D. G. Nocera, "Powering the Planet: Chemical Challenges in
Solar Energy Utilization," Proceedings of the National Academy of Sciences, vol.
103, pp. 15729-15735, 2006.
[6] J. Antonanzas, N. Osorio, R. Escobar, R. Urraca, F. Martinez-de-Pison, and F.
Antonanzas-Torres, "Review of Photovoltaic Power Forecasting," Solar Energy,
vol. 136, pp. 78-111, 2016.
[7] P. Ciais, C. Sabine, G. Bala, L. Bopp, V. Brovkin, J. Canadell, A. Chhabra, R.
DeFries, J. Galloway, and M. Heimann, "Carbon and Other Biogeochemical
Cycles," in Climate Change 2013: The Physical Science Basis. Contribution of
154
Working Group I to the Fifth Assessment Report of the Intergovernmental Panel
on Climate Change, ed: Cambridge University Press, 2014, pp. 465-570.
[8] Y. Pan, R. A. Birdsey, J. Fang, R. Houghton, P. E. Kauppi, W. A. Kurz, O. L.
Phillips, A. Shvidenko, S. L. Lewis, and J. G. Canadell, "A Large and Persistent
Carbon Sink in the World’s Forests," Science, vol. 333, pp. 988-993, 2011.
[9] N. S. Lewis and G. Crabtree, "Basic Research Needs for Solar Energy Utilization:
Report of the Basic Energy Sciences Workshop on Solar Energy Utilization, April
18-21, 2005," ed: US Department of Energy, Office of Basic Energy Science,
2005.
[10] R. Williams, "Becquerel Photovoltaic Effect in Binary Compounds," The journal
of Chemical physics, vol. 32, pp. 1505-1514, 1960.
[11] A.-E. Becquerel, "Recherches Sur Les Effets De La Radiation Chimique De La
Lumiere Solaire Au Moyen Des Courants Electriques," CR Acad. Sci, vol. 9, pp.
145-149, 1839.
[12] W. G. Adams and R. E. Day, "V. The Action of Light on Selenium," Proceedings
of the Royal Society of London, vol. 25, pp. 113-117, 1877.
[13] C. E. Fritts, "On a New Form of Selenium Cell, and Some Electrical Discoveries
Made by Its Use," American Journal of Science, pp. 465-472, 1883.
[14] R. S. Ohl, "Electrical Translating Device Utilizing Silicon," ed: Google Patents,
1946.
[15] D. M. Chapin, C. Fuller, and G. Pearson, "A New Silicon P‐N Junction Photocell
for Converting Solar Radiation into Electrical Power," Journal of Applied
Physics, vol. 25, pp. 676-677, 1954.
155
[16] H. Hoppe and N. S. Sariciftci, "Organic Solar Cells: An Overview," Journal of
materials research, vol. 19, pp. 1924-1945, 2004.
[17] C. Brabec, U. Scherf, and V. Dyakonov, Organic Photovoltaics: Materials, Device
Physics, and Manufacturing Technologies: John Wiley & Sons, 2011.
[18] J. Nelson, "Organic Photovoltaic Films," Current Opinion in Solid State and
Materials Science, vol. 6, pp. 87-95, 2002.
[19] R. W. Miles, G. Zoppi, and I. Forbes, "Inorganic Photovoltaic Cells," Materials
today, vol. 10, pp. 20-27, 2007.
[20] A. Slaoui and R. T. Collins, "Advanced Inorganic Materials for Photovoltaics,"
MRS bulletin, vol. 32, pp. 211-218, 2007.
[21] L. El Chaar and N. El Zein, "Review of Photovoltaic Technologies," Renewable
and sustainable energy reviews, vol. 15, pp. 2165-2175, 2011.
[22] M. Wright and A. Uddin, "Organic—Inorganic Hybrid Solar Cells: A
Comparative Review," Solar energy materials and solar cells, vol. 107, pp. 87-
111, 2012.
[23] D. Khatiwada, S. Venkatesan, N. Adhikari, A. Dubey, A. F. Mitul, L.
Mohammad, A. Iefanova, S. B. Darling, and Q. Qiao, "Efficient Perovskite Solar
Cells by Temperature Control in Single and Mixed Halide Precursor Solutions
and Films," The Journal of Physical Chemistry C, vol. 119, pp. 25747-25753,
2015.
[24] M. Grätzel, "Dye-Sensitized Solar Cells," Journal of photochemistry and
photobiology C: Photochemistry Reviews, vol. 4, pp. 145-153, 2003.
[25] D. A. Neamen, "Semiconductor Physics and Devices," 1992.
156
[26] D. K. Schroder, Semiconductor Material and Device Characterization: John Wiley
& Sons, 2006.
[27] K. Seeger, Semiconductor Physics: Springer Science & Business Media, 2013.
[28] A. G. Baca and C. I. Ashby, Fabrication of Gaas Devices: IET, 2005.
[29] C. Mead and W. Spitzer, "Fermi Level Position at Semiconductor Surfaces,"
Physical Review Letters, vol. 10, p. 471, 1963.
[30] S. Fonash, Solar Cell Device Physics: Elsevier, 2012.
[31] A. Fahrenbruch and R. Bube, Fundamentals of Solar Cells: Photovoltaic Solar
Energy Conversion: Elsevier, 2012.
[32] S. M. Sze and K. K. Ng, Physics of Semiconductor Devices: John wiley & sons,
2006.
[33] A. S. Lin and J. D. Phillips, "Drift-Diffusion Modeling for Impurity Photovoltaic
Devices," IEEE Transactions on Electron Devices, vol. 56, pp. 3168-3174, 2009.
[34] M. A. Green, K. Emery, Y. Hishikawa, W. Warta, and E. D. Dunlop, "Solar Cell
Efficiency Tables (Version 45)," Progress in photovoltaics: research and
applications, vol. 23, pp. 1-9, 2015.
[35] S. Sharma, K. K. Jain, and A. Sharma, "Solar Cells: In Research and
Applications—a Review," Materials Sciences and Applications, vol. 6, p. 1145,
2015.
[36] Q. An, F. Zhang, J. Zhang, W. Tang, Z. Deng, and B. Hu, "Versatile Ternary
Organic Solar Cells: A Critical Review," Energy & Environmental Science, vol.
9, pp. 281-322, 2016.
157
[37] R. Bergmann, "Crystalline Si Thin-Film Solar Cells: A Review," Applied physics
A, vol. 69, pp. 187-194, 1999.
[38] J. Gong, J. Liang, and K. Sumathy, "Review on Dye-Sensitized Solar Cells
(Dsscs): Fundamental Concepts and Novel Materials," Renewable and
Sustainable Energy Reviews, vol. 16, pp. 5848-5860, 2012.
[39] W. Zulehner, "Czochralski Growth of Silicon," Journal of Crystal Growth, vol.
65, pp. 189-213, 1983.
[40] M. Tilli, M. Paulasto-Krockel, T. Motooka, and V. Lindroos, Handbook of
Silicon Based Mems Materials and Technologies: William Andrew, 2015.
[41] P. Rai‐Choudhury and P. Hower, "Growth and Characterization of Polycrystalline
Silicon," Journal of The Electrochemical Society, vol. 120, pp. 1761-1766, 1973.
[42] A. Shah, P. Torres, R. Tscharner, N. Wyrsch, and H. Keppner, "Photovoltaic
Technology: The Case for Thin-Film Solar Cells," science, vol. 285, pp. 692-698,
1999.
[43] B. M. Kayes, H. Nie, R. Twist, S. G. Spruytte, F. Reinhardt, I. C. Kizilyalli, and
G. S. Higashi, "27.6% Conversion Efficiency, a New Record for Single-Junction
Solar Cells under 1 Sun Illumination," in 2011 37th IEEE Photovoltaic Specialists
Conference, 2011, pp. 000004-000008.
[44] B. E. McCandless and J. R. Sites, "Cadmium Telluride Solar Cells," Handbook of
Photovoltaic Science and Engineering, pp. 617-662, 2003.
[45] D. E. Carlson and C. R. Wronski, "Amorphous Silicon Solar Cell," Applied
Physics Letters, vol. 28, pp. 671-673, 1976.
158
[46] A. Pivrikas, N. S. Sariciftci, G. Juška, and R. Österbacka, "A Review of Charge
Transport and Recombination in Polymer/Fullerene Organic Solar Cells,"
Progress in Photovoltaics: Research and Applications, vol. 15, pp. 677-696, 2007.
[47] V. Sugathan, E. John, and K. Sudhakar, "Recent Improvements in Dye Sensitized
Solar Cells: A Review," Renewable and Sustainable Energy Reviews, vol. 52, pp.
54-64, 2015.
[48] K. Chopra, P. Paulson, and V. Dutta, "Thin‐Film Solar Cells: An Overview,"
Progress in photovoltaics: research and applications, vol. 12, pp. 69-92, 2004.
[49] W. Guter, J. Schöne, S. P. Philipps, M. Steiner, G. Siefer, A. Wekkeli, E. Welser,
E. Oliva, A. W. Bett, and F. Dimroth, "Current-Matched Triple-Junction Solar
Cell Reaching 41.1% Conversion Efficiency under Concentrated Sunlight,"
Applied Physics Letters, vol. 94, p. 223504, 2009.
[50] D. L. Meier and A. Rahatgi, "Solar Cell Having Crystalline Silicon Pn
Homojunction and Amorphous Silicon Heterojunctions for Surface Passivation,"
ed: Google Patents, 2009.
[51] V. D. Mihailetchi, Y. Komatsu, and L. Geerligs, "Nitric Acid Pretreatment for the
Passivation of Boron Emitters for N-Type Base Silicon Solar Cells," Applied
Physics Letters, vol. 92, p. 063510, 2008.
[52] D. Macdonald and L. Geerligs, "Recombination Activity of Interstitial Iron and
Other Transition Metal Point Defects in P-and N-Type Crystalline Silicon,"
Applied Physics Letters, vol. 85, pp. 4061-4063, 2004.
159
[53] A. Cuevas, M. J. Kerr, C. Samundsett, F. Ferrazza, and G. Coletti, "Millisecond
Minority Carrier Lifetimes in N-Type Multicrystalline Silicon," Applied Physics
Letters, vol. 81, pp. 4952-4954, 2002.
[54] R. Sahai and A. Milnes, "Heterojunction Solar Cell Calculations," Solid-State
Electronics, vol. 13, pp. 1289-1299, 1970.
[55] R. H. Parekh and A. M. Barnett, "Improved Performance Design of Gallium
Arsenide Solar Cells for Space," IEEE Transactions on electron devices, vol. 31,
pp. 689-695, 1984.
[56] K. Lee, J. D. Zimmerman, T. W. Hughes, and S. R. Forrest, "Non‐Destructive
Wafer Recycling for Low‐Cost Thin‐Film Flexible Optoelectronics," Advanced
Functional Materials, vol. 24, pp. 4284-4291, 2014.
[57] E. F. Schubert, Doping in Iii-V Semiconductors: E. Fred Schubert, 2015.
[58] O. Von Roos, "A Simple Theory of Back Surface Field (Bsf) Solar Cells," Journal
of Applied Physics, vol. 49, pp. 3503-3511, 1978.
[59] V. A. Dao, J. Heo, H. Choi, Y. Kim, S. Park, S. Jung, N. Lakshminarayan, and J.
Yi, "Simulation and Study of the Influence of the Buffer Intrinsic Layer, Back-
Surface Field, Densities of Interface Defects, Resistivity of P-Type Silicon
Substrate and Transparent Conductive Oxide on Heterojunction with Intrinsic
Thin-Layer (Hit) Solar Cell," Solar Energy, vol. 84, pp. 777-783, 2010.
[60] A. Kaminski, B. Vandelle, A. Fave, J. Boyeaux, R. Monna, D. Sarti, and A.
Laugier, "Aluminium Bsf in Silicon Solar Cells," Solar Energy Materials and
Solar Cells, vol. 72, pp. 373-379, 2002.
160
[61] B. Galiana, I. Rey-Stolle, M. Baudrit, I. García, and C. Algora, "A Comparative
Study of Bsf Layers for Gaas-Based Single-Junction or Multijunction
Concentrator Solar Cells," Semiconductor science and technology, vol. 21, p.
1387, 2006.
[62] H. Berger, "Contact Resistance and Contact Resistivity," Journal of the
Electrochemical Society, vol. 119, pp. 507-514, 1972.
[63] K. Bouzidi, M. Chegaar, and A. Bouhemadou, "Solar Cells Parameters Evaluation
Considering the Series and Shunt Resistance," Solar Energy Materials and Solar
Cells, vol. 91, pp. 1647-1651, 2007.
[64] J. Nelson, The Physics of Solar Cells: World Scientific Publishing Company,
2003.
[65] H. Sakai, T. Yoshida, T. Hama, and Y. Ichikawa, "Effects of Surface Morphology
of Transparent Electrode on the Open-Circuit Voltage in a-Si: H Solar Cells,"
Japanese Journal of Applied Physics, vol. 29, p. 630, 1990.
[66] D. Arumí, R. Rodríguez-Montañés, J. Figueras, S. Eichenberger, C. Hora, and B.
Kruseman, "Gate Leakage Impact on Full Open Defects in Interconnect Lines,"
IEEE Transactions on Very Large Scale Integration (VLSI) Systems, vol. 19, pp.
2209-2220, 2011.
[67] T. R. Johnson, G. Ganguly, G. S. Wood, and D. E. Carlson, "Investigation of the
Causes and Variation of Leakage Currents in Amorphous Silicon Pin Diodes,"
MRS Online Proceedings Library Archive, vol. 762, 2003.
[68] M. A. Green, "Accuracy of Analytical Expressions for Solar Cell Fill Factors,"
Solar Cells, vol. 7, pp. 337-340, 1982.
161
[69] C. Riordan and R. Hulstron, "What Is an Air Mass 1.5 Spectrum?(Solar Cell
Performance Calculations)," in IEEE Conference on Photovoltaic Specialists,
1990, pp. 1085-1088.
[70] D. L. King, J. A. Kratochvil, and W. E. Boyson, "Measuring Solar Spectral and
Angle-of-Incidence Effects on Photovoltaic Modules and Solar Irradiance
Sensors," in Conference Record of the Twenty Sixth IEEE Photovoltaic
Specialists Conference-1997, 1997, pp. 1113-1116.
[71] T. Tiedje, E. Yablonovitch, G. D. Cody, and B. G. Brooks, "Limiting Efficiency
of Silicon Solar Cells," IEEE Transactions on electron devices, vol. 31, pp. 711-
716, 1984.
[72] L. C. Hirst and N. J. Ekins‐Daukes, "Fundamental Losses in Solar Cells,"
Progress in Photovoltaics: Research and Applications, vol. 19, pp. 286-293, 2011.
[73] C. Strümpel, M. McCann, G. Beaucarne, V. Arkhipov, A. Slaoui, V. Švrček, C.
Del Cañizo, and I. Tobias, "Modifying the Solar Spectrum to Enhance Silicon
Solar Cell Efficiency—an Overview of Available Materials," Solar energy
materials and solar cells, vol. 91, pp. 238-249, 2007.
[74] J. Zhao and M. A. Green, "Optimized Antireflection Coatings for High-Efficiency
Silicon Solar Cells," IEEE transactions on electron devices, vol. 38, pp. 1925-
1934, 1991.
[75] D. Bouhafs, A. Moussi, A. Chikouche, and J. Ruiz, "Design and Simulation of
Antireflection Coating Systems for Optoelectronic Devices: Application to
Silicon Solar Cells," Solar Energy Materials and Solar Cells, vol. 52, pp. 79-93,
1998.
162
[76] H. Sai, Y. Kanamori, K. Arafune, Y. Ohshita, and M. Yamaguchi, "Light
Trapping Effect of Submicron Surface Textures in Crystalline Si Solar Cells,"
Progress in Photovoltaics: Research and Applications, vol. 15, pp. 415-423, 2007.
[77] Y.-J. Lee, D. S. Ruby, D. W. Peters, B. B. McKenzie, and J. W. Hsu, "Zno
Nanostructures as Efficient Antireflection Layers in Solar Cells," Nano letters,
vol. 8, pp. 1501-1505, 2008.
[78] J. Oh, H.-C. Yuan, and H. M. Branz, "An 18.2%-Efficient Black-Silicon Solar
Cell Achieved through Control of Carrier Recombination in Nanostructures,"
Nature nanotechnology, vol. 7, p. 743, 2012.
[79] A. Cuevas, "The Recombination Parameter J0," Energy Procedia, vol. 55, pp. 53-
62, 2014.
[80] G. Hurkx, D. Klaassen, and M. Knuvers, "A New Recombination Model for
Device Simulation Including Tunneling," IEEE Transactions on electron devices,
vol. 39, pp. 331-338, 1992.
[81] S. Rühle, "Tabulated Values of the Shockley–Queisser Limit for Single Junction
Solar Cells," Solar Energy, vol. 130, pp. 139-147, 2016.
[82] T. Trupke and P. Würfel, "Fundamental Limits of Solar Energy Conversion,"
Photovoltaic Modeling Handbook, pp. 7-25, 2018.
[83] S. Kurtz, D. Myers, W. McMahon, J. Geisz, and M. Steiner, "A Comparison of
Theoretical Efficiencies of Multi‐Junction Concentrator Solar Cells," Progress in
Photovoltaics: research and applications, vol. 16, pp. 537-546, 2008.
[84] K. Tanabe, "A Review of Ultrahigh Efficiency Iii-V Semiconductor Compound
Solar Cells: Multijunction Tandem, Lower Dimensional, Photonic up/Down
163
Conversion and Plasmonic Nanometallic Structures," Energies, vol. 2, pp. 504-
530, 2009.
[85] N. Asim, K. Sopian, S. Ahmadi, K. Saeedfar, M. Alghoul, O. Saadatian, and S. H.
Zaidi, "A Review on the Role of Materials Science in Solar Cells," Renewable
and Sustainable Energy Reviews, vol. 16, pp. 5834-5847, 2012.
[86] S. Adachi, Physical Properties of Iii-V Semiconductor Compounds: John Wiley &
Sons, 1992.
[87] S. Adachi, Properties of Semiconductor Alloys: Group-Iv, Iii-V and Ii-Vi
Semiconductors vol. 28: John Wiley & Sons, 2009.
[88] P. Zanella, G. Rossetto, N. Brianese, F. Ossola, M. Porchia, and J. Williams,
"Organometallic Precursors in the Growth of Epitaxial Thin Films of Iii-V
Semiconductors by Metal-Organic Chemical Vapor Deposition (Mocvd),"
Chemistry of Materials, vol. 3, pp. 225-242, 1991.
[89] D. Holmes, R. Chen, K. Elliott, and C. Kirkpatrick, "Stoichiometry‐Controlled
Compensation in Liquid Encapsulated Czochralski Gaas," Applied Physics
Letters, vol. 40, pp. 46-48, 1982.
[90] D. Holmes, R. Chen, K. R. Elliott, C. Kirkpatrick, and P. W. Yu, "Compensation
Mechanism in Liquid Encapsulated Czochralski Gaas: Importance of Melt
Stoichiometry," IEEE Transactions on Microwave Theory and Techniques, vol.
30, pp. 949-955, 1982.
[91] J. Neave, B. Joyce, P. Dobson, and N. Norton, "Dynamics of Film Growth of
Gaas by Mbe from Rheed Observations," Applied Physics A, vol. 31, pp. 1-8,
1983.
164
[92] J. Neave, B. Joyce, and P. Dobson, "Dynamic Rheed Observations of the Mbe
Growth of Gaas," Applied Physics A, vol. 34, pp. 179-184, 1984.
[93] T. Nishinaga and K.-I. Cho, "Theoretical Study of Mode Transition between 2d-
Nucleation and Step Flow in Mbe Growth of Gaas," Japanese journal of applied
physics, vol. 27, p. L12, 1988.
[94] V. K. Jain, "Synthesis and Purification of Some Main Group Organometallic
Precursors for Compound Semiconductors," Bulletin of Materials Science, vol.
28, pp. 313-316, 2005.
[95] F. Maury, "Organometallic Molecular Precursors for Low‐Temperature Mocvd of
Iii–V Semiconductors," Advanced Materials, vol. 3, pp. 542-548, 1991.
[96] C. Larsen, N. Buchan, S. Li, and G. Stringfellow, "Gaas Growth Using
Tertiarybutylarsine and Trimethylgallium," Journal of Crystal Growth, vol. 93,
pp. 15-19, 1988.
[97] G. Haacke, S. Watkins, and H. Burkhard, "Metalorganic Chemical Vapor
Deposition of High‐Purity Gaas Using Tertiarybutylarsine," Applied Physics
Letters, vol. 54, pp. 2029-2031, 1989.
[98] R. Bhat, M. Koza, and B. Skromme, "Growth of High‐Quality Gaas Using
Trimethylgallium and Diethylarsine," Applied physics letters, vol. 50, pp. 1194-
1196, 1987.
[99] V. Rideout, "A Review of the Theory and Technology for Ohmic Contacts to
Group Iii–V Compound Semiconductors," Solid-State Electronics, vol. 18, pp.
541-550, 1975.
165
[100]
Films on N-Type Gaas," Solid-State Electronics, vol. 18, pp. 331-338, 1975.
[101] J. Gyulai, J. Mayer, V. Rodriguez, A. Yu, and H. Gopen, "Alloying Behavior of
Au and Au–Ge on Gaas," Journal of Applied Physics, vol. 42, pp. 3578-3585,
1971.
[102]
-Gaas," Solid-State Electronics, vol. 17, pp. 835-837,
1974.
[103] A. Piotrowska, A. Guivarc'h, and G. Pelous, "Ohmic Contacts to Iii–V Compound
Semiconductors: A Review of Fabrication Techniques," Solid-State Electronics,
vol. 26, pp. 179-197, 1983.
[104] M. Yoder, "Ohmic Contacts in Gaas," Solid-State Electronics, vol. 23, pp. 117-
119, 1980.
[105] F. Dimroth and S. Kurtz, "High-Efficiency Multijunction Solar Cells," MRS
bulletin, vol. 32, pp. 230-235, 2007.
[106] M. Yamaguchi, T. Takamoto, K. Araki, and N. Ekins-Daukes, "Multi-Junction
Iii–V Solar Cells: Current Status and Future Potential," Solar Energy, vol. 79, pp.
78-85, 2005.
[107] M. W. Wanlass, "Monolithic Tandem Solar Cell," ed: Google Patents, 1991.
[108] M. W. Wanlass, "Monolithic Tandem Solar Cell," ed: Google Patents, 1994.
[109] A. Bett, C. Baur, R. Beckert, F. Diimroth, G. Letay, M. Hein, M. Muesel, S. van
Riesen, U. Schubert, and G. Siefer, "Development of High-Efficiency
Mechanically Stacked Gainp/Gainas-Gasb Triple-Junction Concentrator Solar
166
Cells," in Conference Record, 17th European Solar Energy Conference, Munich,
2001, pp. 84-87.
[110] B. M. Kayes, L. Zhang, R. Twist, I.-K. Ding, and G. S. Higashi, "Flexible Thin-
Film Tandem Solar Cells with> 30% Efficiency," IEEE Journal of Photovoltaics,
vol. 4, pp. 729-733, 2014.
[111] F. Dimroth, M. Grave, P. Beutel, U. Fiedeler, C. Karcher, T. N. Tibbits, E. Oliva,
G. Siefer, M. Schachtner, and A. Wekkeli, "Wafer Bonded Four‐Junction
Gainp/Gaas//Gainasp/Gainas Concentrator Solar Cells with 44.7% Efficiency,"
Progress in Photovoltaics: Research and Applications, vol. 22, pp. 277-282, 2014.
[112] W. Marra, P. Eisenberger, and A. Cho, "X‐Ray Total‐External‐Reflection–Bragg
Diffraction: A Structural Study of the Gaas‐Al Interface," Journal of Applied
Physics, vol. 50, pp. 6927-6933, 1979.
[113] B. E. Warren, X-Ray Diffraction: Courier Corporation, 1990.
[114] A. Guinier, X-Ray Diffraction in Crystals, Imperfect Crystals, and Amorphous
Bodies: Courier Corporation, 1994.
[115] H. Ogawa, M. Nishikawa, and A. Abe, "Hall Measurement Studies and an
Electrical Conduction Model of Tin Oxide Ultrafine Particle Films," Journal of
Applied Physics, vol. 53, pp. 4448-4455, 1982.
[116] R. Chwang, B. Smith, and C. Crowell, "Contact Size Effects on the Van Der
Pauw Method for Resistivity and Hall Coefficient Measurement," Solid-State
Electronics, vol. 17, pp. 1217-1227, 1974.
[117] W. Versnel, "The Geometrical Correction Factor for a Rectangular Hall Plate,"
Journal of Applied Physics, vol. 53, pp. 4980-4986, 1982.
167
[118] D. W. Koon, "Effect of Contact Size and Placement, and of Resistive
Inhomogeneities on Van Der Pauw Measurements," Review of Scientific
Instruments, vol. 60, pp. 271-274, 1989.
[119] S. A. Empedocles, R. Neuhauser, K. Shimizu, and M. G. Bawendi,
"Photoluminescence from Single Semiconductor Nanostructures," Advanced
Materials, vol. 11, pp. 1243-1256, 1999.
[120] G. S. Pomrenke, H. Ennen, and W. Haydl, "Photoluminescence Optimization and
Characteristics of the Rare‐Earth Element Erbium Implanted in Gaas, Inp, and
Gap," Journal of applied physics, vol. 59, pp. 601-610, 1986.
[121] A. Benninghoven, J. Okano, R. Shimizu, and H. Werner, "Secondary Ion Mass
Spectrometry Sims Iv," in Proc. 3rd Int. Conf., Budapest, 1981, pp. 438-42.
[122] A. Benninghoven, "Chemical Analysis of Inorganic and Organic Surfaces and
Thin Films by Static Time‐of‐Flight Secondary Ion Mass Spectrometry (Tof‐
Sims)," Angewandte Chemie International Edition in English, vol. 33, pp. 1023-
1043, 1994.
[123] G. Hlawacek and A. Gölzhäuser, Helium Ion Microscopy: Springer, 2016.
[124] H. Gnaser and H. Oechsner, "Noval Detection Scheme for the Analysis of
Hydrogen and Helium by Secondary Ion Mass Spectrometry," Surface and
interface analysis, vol. 17, pp. 646-649, 1991.
[125] C. Domínguez, I. Antón, and G. Sala, "Solar Simulator for Concentrator
Photovoltaic Systems," Optics express, vol. 16, pp. 14894-14901, 2008.
168
[126] G. Reeves and H. Harrison, "Obtaining the Specific Contact Resistance from
Transmission Line Model Measurements," IEEE Electron device letters, vol. 3,
pp. 111-113, 1982.
[127] U. Hoefer, K. Steiner, and E. Wagner, "Contact and Sheet Resistance of Sno2
Thin Films from Transmission-Line Model Measurements," Sensors and
Actuators B: Chemical, vol. 26, pp. 59-63, 1995.
[128] J. C. Fan, C. O. Bozler, and B. J. Palm, "Calculated and Measured Efficiencies of
Thin‐Film Shallow‐Homojunction Gaas Solar Cells on Ge Substrates," Applied
Physics Letters, vol. 35, pp. 875-878, 1979.
[129] D. Miller and J. Harris Jr, "Molecular Beam Epitaxial Gaas Heteroface Solar Cell
Grown on Ge," Applied Physics Letters, vol. 37, pp. 1104-1106, 1980.
[130] M. Hudait, P. Modak, S. Hardikar, K. Rao, and S. Krupanidhi, "Comparative
Studies of Si-Doped N-Type Movpe Gaas on Ge and Gaas Substrates," Materials
Science and Engineering: B, vol. 55, pp. 53-67, 1998.
[131] A. N. Quezada, A. Rodríguez, M. Vidal, G. H. Sosa, and H. N. Contreras,
"Critical Thickness of Ge/Gaas (001) Epitaxial Films," Superficies y vacío, vol.
16, pp. 42-44, 2003.
[132] S. Ting and E. Fitzgerald, "Metal-Organic Chemical Vapor Deposition of Single
Domain Gaas on Ge/Ge X Si 1− X/Si and Ge Substrates," Journal of Applied
Physics, vol. 87, pp. 2618-2628, 2000.
[133] M. Yamaguchi and C. Amano, "Efficiency Calculations of Thin‐Film Gaas Solar
Cells on Si Substrates," Journal of applied physics, vol. 58, pp. 3601-3606, 1985.
169
[134] M. E. Groenert, C. W. Leitz, A. J. Pitera, V. Yang, H. Lee, R. J. Ram, and E. A.
Fitzgerald, "Monolithic Integration of Room-Temperature Cw Gaas/Algaas
Lasers on Si Substrates Via Relaxed Graded Gesi Buffer Layers," Journal of
Applied Physics, vol. 93, pp. 362-367, 2003.
[135] W.-C. Kuo, H.-C. Hsieh, W. Chih-Hung, H. Wen-Hsiang, C.-C. Lee, and J.-Y.
Chang, "High Quality Gaas Epilayers Grown on Si Substrate Using 100 Nm Ge
Buffer Layer," International Journal of Photoenergy, vol. 2016, 2016.
[136] G. Kamath, J. Ewan, and R. Knechtli, "Large-Area High-Efficiency (Alga) as—
Gaas Solar Cells," IEEE Transactions on Electron Devices, vol. 24, pp. 473-475,
1977.
[137] M. Konagai, M. Sugimoto, and K. Takahashi, "High Efficiency Gaas Thin Film
Solar Cells by Peeled Film Technology," Journal of crystal growth, vol. 45, pp.
277-280, 1978.
[138] E. Yablonovitch, T. Gmitter, J. Harbison, and R. Bhat, "Extreme Selectivity in the
Lift‐Off of Epitaxial Gaas Films," Applied Physics Letters, vol. 51, pp. 2222-
2224, 1987.
[139] A. Van Geelen, P. Hageman, G. Bauhuis, P. Van Rijsingen, P. Schmidt, and L.
Giling, "Epitaxial Lift-Off Gaas Solar Cell from a Reusable Gaas Substrate,"
Materials Science and Engineering: B, vol. 45, pp. 162-171, 1997.
[140] J. Schermer, G. Bauhuis, P. Mulder, W. Meulemeesters, E. Haverkamp, M.
Voncken, and P. Larsen, "High Rate Epitaxial Lift-Off of Ingap Films from Gaas
Substrates," Applied Physics Letters, vol. 76, pp. 2131-2133, 2000.
170
[141] J. Schermer, P. Mulder, G. Bauhuis, M. Voncken, J. Van Deelen, E. Haverkamp,
and P. Larsen, "Epitaxial Lift‐Off for Large Area Thin Film Iii/V Devices,"
physica status solidi (a), vol. 202, pp. 501-508, 2005.
[142] R. Tatavarti, G. Hillier, A. Dzankovic, G. Martin, F. Tuminello, R.
Navaratnarajah, G. Du, D. Vu, and N. Pan, "Lightweight, Low Cost Gaas Solar
Cells on 4 ″Epitaxial Liftoff (Elo) Wafers," in 2008 33rd IEEE Photovoltaic
Specialists Conference, 2008, pp. 1-4.
[143] G. Bauhuis, P. Mulder, E. Haverkamp, J. Huijben, and J. Schermer, "26.1% Thin-
Film Gaas Solar Cell Using Epitaxial Lift-Off," Solar Energy Materials and Solar
Cells, vol. 93, pp. 1488-1491, 2009.
[144] W. Choi, C. Z. Kim, C. S. Kim, W. Heo, T. Joo, S. Y. Ryu, H. Kim, H. Kim, H.
K. Kang, and S. Jo, "A Repeatable Epitaxial Lift‐Off Process from a Single Gaas
Substrate for Low‐Cost and High‐Efficiency Iii‐V Solar Cells," Advanced Energy
Materials, vol. 4, p. 1400589, 2014.
[145] R.-H. Horng, F.-L. Wu, S.-L. Ou, and Y.-C. Kao, "High Separation Rate of
Epitaxial Lift-Off Using Hydrophilic Solvent for Iii–V Solar Cell and Reusable
Applications," in 2015 IEEE 42nd Photovoltaic Specialist Conference (PVSC),
2015, pp. 1-3.
[146] G. Bauhuis, P. Mulder, E. Haverkamp, and J. Schermer, "Substrate Reuse for
Epitaxial Lift-Off of Iii-V Solar Cells," in Proceedings of the 22nd European
Photovoltaic Solar Energy Conference: Milan, 2007, pp. 106-110.
[147] V. Selvamanickam, S. Sambandam, A. Sundaram, S. Lee, A. Rar, X. Xiong, A.
Alemu, C. Boney, and A. Freundlich, "Germanium Films with Strong in-Plane
171
and out-of-Plane Texture on Flexible, Randomly Textured Metal Substrates,"
Journal of Crystal Growth, vol. 311, pp. 4553-4557, 2009.
[148] P. Dutta, M. Rathi, N. Zheng, Y. Gao, Y. Yao, J. Martinez, P. Ahrenkiel, and V.
Selvamanickam, "High Mobility Single-Crystalline-Like Gaas Thin Films on
Inexpensive Flexible Metal Substrates by Metal-Organic Chemical Vapor
Deposition," Applied Physics Letters, vol. 105, p. 092104, 2014.
[149] M. Rathi, P. Dutta, N. Zheng, Y. Yao, D. Khatiwada, A. Khadimallah, Y. Gao, S.
Sun, Y. Li, and S. Pouladi, "High Opto-Electronic Quality N-Type Single-
Crystalline-Like Gaas Thin Films on Flexible Metal Substrates," Journal of
Materials Chemistry C, vol. 5, pp. 7919-7926, 2017.
[150] R. Wang, S. Sambandam, G. Majkic, E. Galstyan, and V. Selvamanickam, "High
Mobility Single-Crystalline-Like Germanium Thin Films on Flexible, Inexpensive
Substrates," Thin Solid Films, vol. 527, pp. 9-15, 2013.
[151] V. N. Bessolov, M. V. Lebedev, N. M. Binh, M. Friedrich, and D. R. Zahn,
"Sulphide Passivation of Gaas: The Role of the Sulphur Chemical Activity,"
Semiconductor science and technology, vol. 13, p. 611, 1998.
[152] C. Sandroff, R. Nottenburg, J. C. Bischoff, and R. Bhat, "Dramatic Enhancement
in the Gain of a Gaas/Algaas Heterostructure Bipolar Transistor by Surface
Chemical Passivation," Applied Physics Letters, vol. 51, pp. 33-35, 1987.
[153] B. Brennan, M. Milojevic, C. Hinkle, F. Aguirre-Tostado, G. Hughes, and R.
Wallace, "Optimisation of the Ammonium Sulphide (Nh 4) 2 S Passivation
Process on in 0.53 Ga 0.47 As," Applied surface science, vol. 257, pp. 4082-4090,
2011.
172
[154] M. T. Sheldon, C. N. Eisler, and H. A. Atwater, "Gaas Passivation with
Trioctylphosphine Sulfide for Enhanced Solar Cell Efficiency and Durability,"
Advanced Energy Materials, vol. 2, pp. 339-344, 2012.
[155] R. Ghita, C. Negrila, C. Cotirlan, and C. Logofatu, "On the Passivation of Gaas
Surface by Sulfide Compounds," Digest Journal of Nanomaterials &
Biostructures (DJNB), vol. 8, 2013.
[156] G. Wang, T. Ogawa, M. Umeno, T. Soga, and T. Jimbo, "Surface and Bulk
Passivation of Gaas Solar Cell on Si Substrate by H 2+ Ph 3 Plasma," Applied
Physics Letters, vol. 76, pp. 730-732, 2000.
[157] K. Tsuchiya, M. Sakata, A. Funyu, and H. Ikoma, "Selenium Passivation of Gaas
with Se/Nh4oh Solution," Japanese journal of applied physics, vol. 34, p. 5926,
1995.
[158] C. Sandroff, M. Hegde, L. Farrow, R. Bhat, J. Harbison, and C. Chang,
"Enhanced Electronic Properties of Gaas Surfaces Chemically Passivated by
Selenium Reactions," Journal of applied physics, vol. 67, pp. 586-588, 1990.
[159] J. Herman and F. Terry Jr, "Hydrogen Sulfide Plasma Passivation of Gallium
Arsenide," Applied physics letters, vol. 60, pp. 716-717, 1992.
[160] J.-H. Kim, D. H. Lim, and G. M. Yang, "Selective Etching of Algaas/Gaas
Structures Using the Solutions of Citric Acid/H 2 O 2 and De-Ionized H 2
O/Buffered Oxide Etch," Journal of Vacuum Science & Technology B:
Microelectronics and Nanometer Structures Processing, Measurement, and
Phenomena, vol. 16, pp. 558-560, 1998.
173
[161] G. C. DeSalvo, W. F. Tseng, and J. Comas, "Etch Rates and Selectivities of Citric
Acid/Hydrogen Peroxide on Gaas, Al0. 3ga0. 7as, In0. 2ga0. 8as, In0. 53ga0.
47as, In0. 52al0. 48as, and Inp," Journal of The Electrochemical Society, vol.
139, pp. 831-835, 1992.
[162] A. Callegari, P. Hoh, D. Buchanan, and D. Lacey, "Unpinned Gallium
Oxide/Gaas Interface by Hydrogen and Nitrogen Surface Plasma Treatment,"
Applied Physics Letters, vol. 54, pp. 332-334, 1989.
[163] C. Seager and D. Ginley, "Passivation of Grain Boundaries in Polycrystalline
Silicon," Applied Physics Letters, vol. 34, pp. 337-340, 1979.
[164] N. Johnson, D. Biegelsen, and M. Moyer, "Deuterium Passivation of Grain‐
Boundary Dangling Bonds in Silicon Thin Films," Applied Physics Letters, vol.
40, pp. 882-884, 1982.
[165] N. Nickel, N. Johnson, and W. Jackson, "Hydrogen Passivation of Grain
Boundary Defects in Polycrystalline Silicon Thin Films," Applied Physics Letters,
vol. 62, pp. 3285-3287, 1993.
[166] C. Dube, J. Hanoka, and D. Sandstrom, "Hydrogen Diffusion Along Passivated
Grain Boundaries in Silicon Ribbon," Applied Physics Letters, vol. 44, pp. 425-
427, 1984.
[167] N. Johnson, R. Burnham, R. Street, and R. Thornton, "Hydrogen Passivation of
Shallow-Acceptor Impurities in P-Type Gaas," Physical Review B, vol. 33, p.
1102, 1986.
174
[168] N. Pan, S. Bose, M. Kim, G. Stillman, F. Chambers, G. Devane, C. Ito, and M.
Feng, "Hydrogen Passivation of C Acceptors in High‐Purity Gaas," Applied
physics letters, vol. 51, pp. 596-598, 1987.
[169] W. T. Read and W. Shockley, "Dislocation Models of Crystal Grain Boundaries,"
Physical review, vol. 78, p. 275, 1950.
[170] T. Sugino, T. Yamada, and J. Shirafuji, "Effect of Surface Phosphidization on
Gaas Schottky Barrier Junctions," Japanese Journal of Applied Physics, vol. 29, p.
L864, 1990.
[171] T. Sugino, T. Yamada, K. Matsuda, and J. Shirafuji, "Schottky Junctions on
Phosphidized Gaas Surfaces," Applied surface science, vol. 56, pp. 311-316,
1992.
[172] P. Dutta, M. Rathi, D. Khatiwada, S. Sun, Y. Yao, B. Yu, S. Reed, M. Kacharia,
J. Martinez, and A. Litvinchuk, "Flexible Gaas Solar Cells on Roll-to-Roll
Processed Epitaxial Ge Films on Metal Foils: A Route Towards Low-Cost and
High-Performance Iii–V Photovoltaics," Energy & Environmental Science, vol.
12, pp. 756-766, 2019.
[173] J. Meier, R. Flückiger, H. Keppner, and A. Shah, "Complete Microcrystalline P‐I‐
N Solar Cell—Crystalline or Amorphous Cell Behavior?," Applied Physics
Letters, vol. 65, pp. 860-862, 1994.
[174] B. Ward, J. A. Notte, and N. Economou, "Helium Ion Microscope: A New Tool
for Nanoscale Microscopy and Metrology," Journal of Vacuum Science &
Technology B: Microelectronics and Nanometer Structures Processing,
Measurement, and Phenomena, vol. 24, pp. 2871-2874, 2006.
175
[175] J. Notte, R. Hill, S. McVey, L. Farkas, R. Percival, and B. Ward, "An
Introduction to Helium Ion Microscopy," Microscopy and Microanalysis, vol. 12,
pp. 126-127, 2006.
[176] D. C. Joy and B. J. Griffin, "Orion Nanofab-2nd Generation Helium Ion
Microscope," Microscopy and Microanalysis, vol. 19, pp. 854-855, 2013.
[177] V. Selvamanickam, S. Sambandam, A. Sundaram, S. Lee, G. Majkic, A. Rar, and
A. Freundlich, "Novel, Single-Crystalline-Like Templates on Low-Cost, Flexible
Substrates for High Efficiency Photovoltaics," in 2010 35th IEEE Photovoltaic
Specialists Conference, 2010, pp. 001048-001051.
[178] R. Fischer, D. Neuman, H. Zabel, H. Morkoc, C. Choi, and N. Otsuka,
"Dislocation Reduction in Epitaxial Gaas on Si (100)," Applied physics letters,
vol. 48, pp. 1223-1225, 1986.
[179] G. Osbourn, "Strained-Layer Superlattices-a Brief Review," IEEE journal of
quantum electronics, vol. 22, pp. 1677-1681, 1986.
176