<%BANNER%>

Chemical Bath Deposited Zinc Cadmium Sulfide and Sputter Deposited Zinc Oxide for Thin Film Solar Cell Device Fabrication

Permanent Link: http://ufdc.ufl.edu/UFE0022644/00001

Material Information

Title: Chemical Bath Deposited Zinc Cadmium Sulfide and Sputter Deposited Zinc Oxide for Thin Film Solar Cell Device Fabrication
Physical Description: 1 online resource (189 p.)
Language: english
Creator: Baran, Andre
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: aluminum, bath, cadmium, chemical, deposition, hydrogen, oxide, sputtering, sulfide, zinc
Chemical Engineering -- Dissertations, Academic -- UF
Genre: Chemical Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Chemical bath deposition, or CBD, is used to successfully deposit ZnCdS buffer layers that have many benefits over the traditional CdS thin films used in photovoltaic devices. The characteristics of films growth such as growth rate, optical transmission, and film composition is analyzed. The addition of Zn to CdS has important benefits that aid in improving device performance and therefore it is important to determine how the incorporation of zinc affects film characteristics. Film thickness is found to increase with deposition time and growth rate is determined to be a function of the prepared bath zinc composition. The zinc composition in the prepared bath is also found to affect the optical transmission of deposited film, most notably in the short wavelength region. The transparent conductive oxide ZnO is deposited via RF magnetron sputtering. The effect of aluminum and hydrogen doping is studied by sputtering an aluminum-doped zinc oxide target with argon (AZO) and an argon mixture with hydrogen (HAZO). The addition of hydrogen to the working gas decreases film resistance in addition to improving other characteristics. The working gas pressure and position under the target are also found to have a significant effect on film properties. A figure of merit calculation allows for a single valued representation of the combined characteristics of the deposited films. The figure of merit calculation takes into consideration film resistivity, optical transmission, and provides a quantitative value of the potential performance in photovoltaic devices. The effect of thermal treatments on sputter deposited AZO and HAZO films are determined by way of rapid thermal annealing. High temperature annealing of ZnO is found to improve film characteristics. The gas ambient used in the thermal treatment process is found to be a critical parameter in post-anneal film quality. Gas ambients of nitrogen, argon, and forming gas, a nitrogen and hydrogen mixture, were employed in the thermal treatment studies. In general, optical transmission and resistivity improve for all films under each ambient, although films annealed in forming gas show the most improvement, with resulting figure of merit values two to three times their as deposited values.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Andre Baran.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Crisalle, Oscar D.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0022644:00001

Permanent Link: http://ufdc.ufl.edu/UFE0022644/00001

Material Information

Title: Chemical Bath Deposited Zinc Cadmium Sulfide and Sputter Deposited Zinc Oxide for Thin Film Solar Cell Device Fabrication
Physical Description: 1 online resource (189 p.)
Language: english
Creator: Baran, Andre
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: aluminum, bath, cadmium, chemical, deposition, hydrogen, oxide, sputtering, sulfide, zinc
Chemical Engineering -- Dissertations, Academic -- UF
Genre: Chemical Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Chemical bath deposition, or CBD, is used to successfully deposit ZnCdS buffer layers that have many benefits over the traditional CdS thin films used in photovoltaic devices. The characteristics of films growth such as growth rate, optical transmission, and film composition is analyzed. The addition of Zn to CdS has important benefits that aid in improving device performance and therefore it is important to determine how the incorporation of zinc affects film characteristics. Film thickness is found to increase with deposition time and growth rate is determined to be a function of the prepared bath zinc composition. The zinc composition in the prepared bath is also found to affect the optical transmission of deposited film, most notably in the short wavelength region. The transparent conductive oxide ZnO is deposited via RF magnetron sputtering. The effect of aluminum and hydrogen doping is studied by sputtering an aluminum-doped zinc oxide target with argon (AZO) and an argon mixture with hydrogen (HAZO). The addition of hydrogen to the working gas decreases film resistance in addition to improving other characteristics. The working gas pressure and position under the target are also found to have a significant effect on film properties. A figure of merit calculation allows for a single valued representation of the combined characteristics of the deposited films. The figure of merit calculation takes into consideration film resistivity, optical transmission, and provides a quantitative value of the potential performance in photovoltaic devices. The effect of thermal treatments on sputter deposited AZO and HAZO films are determined by way of rapid thermal annealing. High temperature annealing of ZnO is found to improve film characteristics. The gas ambient used in the thermal treatment process is found to be a critical parameter in post-anneal film quality. Gas ambients of nitrogen, argon, and forming gas, a nitrogen and hydrogen mixture, were employed in the thermal treatment studies. In general, optical transmission and resistivity improve for all films under each ambient, although films annealed in forming gas show the most improvement, with resulting figure of merit values two to three times their as deposited values.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Andre Baran.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Crisalle, Oscar D.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0022644:00001


This item has the following downloads:


Full Text

PAGE 1

1 CHEMICAL BATH DEPOSITED ZINC CADMIUM SULFIDE AND SPUTTER DEPOSTED ZINC OXIDE FOR THIN FILM SO LAR CELL DEVICE FABRICATION By ANDRE BARAN V A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009

PAGE 2

2 2009 Andre Baran V

PAGE 3

3 To my parents who have supported me in all areas of my life.

PAGE 4

4 ACKNOWLEDGMENTS I wish to thank the chair of m y supervisor y committee, Dr. Crisalle, for his invaluable support and his sincere understandin g throughout the last five years. I would also like to thank Dr. Craciun for all the characteri zation support that he provided and for agreeing to serve on my committee, along with Dr. Hoflund and Dr. Svoronos. This work would have not been possible without the equipment expertise of Dr. Mark Davids on, and I am very grateful for his assistance. Chuck Rowland also helped deal with equipmen t issues and support and I sincerely appreciate his efforts. I would like to thank all the co lleagues that I have had the pr ivilege of working with. I thank Jiyon Song, Ryan Kaczynski, Ryan Acher, Wei Liu, and Woo Kyoung Kim, all of whom participated in our photovoltaics gr oup at the University of Florid a. I would also like to thank Sean Jones and Kim Interliggi-Crim for their inva luable support and discussions. In addition I would like to thank Evan Law for his assistan ce with equipment repair, and Dr. Maggie PugaLambers for the SIMS work. I thank Andrew Gerg er the AFM work. I w ould especially like to thank Jacqueline Hodges for the nume rous nights she spent by my side. Lastly, I would like to thank my parents for al l their hard work and support that has gotten me to this point in my life. They have supporte d me every step of the way, and I doubt Ill ever be able to thank them enough.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ............................................................................................................... 4LIST OF TABLES ...........................................................................................................................9LIST OF FIGURES .......................................................................................................................10LIST OF ABBREVIATIONS ........................................................................................................ 14ABSTRACT ...................................................................................................................... .............16 CHAP TER 1 INTRODUCTION .................................................................................................................. 182 TECHNOLOGY OVE RVIE W ............................................................................................... 28Solar Cell Device Physics .......................................................................................................28Photodevices .................................................................................................................. ..28The Photoelectric Effect ..................................................................................................28Solar Radiation ................................................................................................................29Band Gap Energy ............................................................................................................30Solar Cell Materials .........................................................................................................30Thin Film Solar Cell Materials ............................................................................................... 31Amorphous Silicon Alloys Solar Cells ............................................................................ 32Amorphous Silicon Growth Methods .............................................................................. 33Copper Indium Diselenide ............................................................................................... 34Copper Gallium Diselenide ............................................................................................. 35Copper Indium Gallium Diselenide ................................................................................. 36Cadmium Sulfide ............................................................................................................. 37Zinc Sulfide .....................................................................................................................38Zinc Cadmium Sulfide ....................................................................................................38Zinc Selenide ...................................................................................................................39Indium Tin Oxide ............................................................................................................ 40Zinc Oxide .......................................................................................................................41Chemical Bath Deposition ......................................................................................................42Fundamentals .................................................................................................................. .42Nucleation .................................................................................................................... ....43Film Growth ....................................................................................................................45Ion-by-ion mechanism .............................................................................................. 45Hydroxide cluster mechanism .................................................................................. 46Sputter Deposition ............................................................................................................ ......47Chamber Features ............................................................................................................48Targets ....................................................................................................................... ......48Sputter Sources ................................................................................................................49

PAGE 6

6 Nonmagnetron sputter sources ................................................................................. 49Magnetron sputter sources ....................................................................................... 50Ion beam sputter sources .......................................................................................... 503 METHODS AND TECHNIQUES ......................................................................................... 60Substrate Cleaning Procedure .................................................................................................60Absorber Layer Fabrication ....................................................................................................61Buffer Layer Fabrication ........................................................................................................62Chemical Bath Deposition Apparatus ....................................................................................62Transparent Conductive Oxide Fabrication ............................................................................ 63Sputter Deposition System .....................................................................................................64Photovoltaic Device Fabrication .............................................................................................67Back Contact ...................................................................................................................67Metallization and Anti-Reflective Coating .....................................................................68Characterization Techniques ..................................................................................................68Inductively Coupled Plasma ............................................................................................ 68Scanning Electron Microscopy ........................................................................................70Secondary electron imaging ..................................................................................... 71Backscattered electron imaging ............................................................................... 71Energy dispersive spectroscopy ............................................................................... 72X-Ray Diffraction ............................................................................................................ 72Profilometry .................................................................................................................. ...74Ellipsometry .................................................................................................................. ..75Ultraviolet-Visible Spectroscopy ....................................................................................75X-Ray Photoelectron Spectroscopy .................................................................................77Secondary Ion Mass Spectrometry .................................................................................. 78Auger Electron Spectroscopy ..........................................................................................79Four-Point Probe ............................................................................................................. 81Photovoltaic Device Characterization Techniques ................................................................. 82Current-Voltage ............................................................................................................... 82Current-Voltage Measurement ........................................................................................ 84Quantum Efficiency ......................................................................................................... 854 ZINC CADMIUM SULFIDE BUFFER LAYER GROWTH ................................................ 99Introduction .................................................................................................................. ...........99Chemical Bath Deposition of Zinc Cadmium Sulfide .......................................................... 101Materials ..................................................................................................................... ...102Growth Procedure ..........................................................................................................102Substrate and equipment cleaning ..........................................................................103Preparation of dose solutions ................................................................................. 104Bath temperature conditioning ............................................................................... 104Buffer layer deposition ........................................................................................... 105Termination of growth ........................................................................................... 105Zinc Cadmium Sulfide Buffer Layer Characterization Results ............................................106

PAGE 7

7 Growth Rate ...................................................................................................................106Film Composition .......................................................................................................... 107Optical Characteristics ...................................................................................................108Conclusions ...........................................................................................................................1105 CHARACTERIZATION OF SPUTTER DEPOSITED ZINC OXIDE ...............................124Introduction .................................................................................................................. .........124Experimental .................................................................................................................. .......125Equipment ..................................................................................................................... .125Materials ..................................................................................................................... ...126Zinc Oxide Character ization Results .................................................................................... 127Film Thickness ..............................................................................................................127Aluminum doped .................................................................................................... 128Hydrogen and aluminum doped ............................................................................. 129Comparison of AZO and HAZO ............................................................................ 130Electrical Characteristics ............................................................................................... 131Aluminum doped .................................................................................................... 131Hydrogen and aluminum doped ............................................................................. 133Comparison of AZO and HAZO ............................................................................ 135Optical Characteristics ...................................................................................................136Figure of Merit ..............................................................................................................137Conclusions ...........................................................................................................................1386 THERMAL TREATMENT OF ZINC OXIDE .................................................................... 157Introduction .................................................................................................................. .........157Experimental .................................................................................................................. .......158Equipment ..................................................................................................................... .159Materials ..................................................................................................................... ...160Annealing Results .................................................................................................................161Electrical Characterization ............................................................................................ 161Aluminum doped .................................................................................................... 162Hydrogen and aluminum doped ............................................................................. 163Optical Characterization ................................................................................................163Aluminum doped .................................................................................................... 164Hydrogen and aluminum doped ............................................................................. 164Figure of Merit ..............................................................................................................165Aluminum doped .................................................................................................... 165Hydrogen and aluminum doped ............................................................................. 166Structural Characteristics ............................................................................................... 166Surface Roughness ........................................................................................................167Aluminum doped .................................................................................................... 167Hydrogen and aluminum doped ............................................................................. 167Conclusions ...........................................................................................................................168

PAGE 8

8 LIST OF REFERENCES .............................................................................................................181BIOGRAPHICAL SKETCH .......................................................................................................189

PAGE 9

9 LIST OF TABLES Table page 2-1 Approximate solubility product of comm only chemically bath deposited compounds. ... 523-1 Reference data for NREL device S2117-A2. .....................................................................874-1 Recipe for the preparation of the chemi cal bath solution for five different zinc compositions. ................................................................................................................. ..1114-2 Resulting chemical bath solutions fo r five different zinc compositions. ......................... 1114-3 Zinc cadmium sulfide film thicknesses for five values of prepared fractional composition of zinc at 85C. ............................................................................................ 1114-4 Comparison of measured fractional zi nc composition in films deposited for 45 minutes with five prepared zinc compositions at 85C. .................................................. 1124-5 Average transmission of three wavelengt h ranges for ZnCdS films deposited for 45 minutes with five prepared zinc compositions at 85C. .................................................. 1125-1 Average thickness of AZO films at flowrates of 30.0, 40.0, 50.0, and 60.0 sccm. ......... 1395-2 Average thickness of HAZO films at flowrates of 20.0, 30.0, 40.0, and 50.0 sccm. ...... 1396-1 International Centre for Diffrac tion Data fingerprint for ZnO. ....................................... 1696-2 Thermal treatment re sults for AZO films. ....................................................................... 1696-3 Thermal treatment re sults for HAZO films. .................................................................... 1696-4 Surface roughness of annealed AZO and HAZO measured by AFM. ............................ 170

PAGE 10

10 LIST OF FIGURES Figure page 2-1 A p-n junction photovoltaic device. ...................................................................................532-2 Band diagram of the photoelectric e ffect in a p-n junction photodevice. .......................... 542-3 Solar spectrum at AM0 and AM1.5. .................................................................................. 552-4 Chronological progress of a-Si, CdTe, a nd CIS thin film technologies from 1975 to 2005....................................................................................................................................562-5 Ion-by-ion growth of CdS crystals. ....................................................................................572-6 Growth of CdS crystals via th e hydroxide cluster mechanism. ......................................... 582-7 Typical sputter deposition chamber. .................................................................................. 593-1 Structure of a typical CIGS/Z nCdS device fabricated at UF. ............................................ 883-2 Chemical bath deposition apparatus. .................................................................................893-3 Chemical bath deposition reaction vessel. .........................................................................903-4 Chemical bath deposition appara tus and reaction vessel schematic. .................................913-5 Perkin-Elmer Model 4400 Production Sputtering System................................................. 923-6 Perkin-Elmer Model 4400 Production Sputtering System schematic drawing. ................933-7 Simplified operation of a s canning electron microscope. .................................................. 943-8 Common experimental arrangement for ellipsometry measurements. .............................. 953-9 Simplified UV/vis spectrophotometer schematic. .............................................................963-10 Four-point probe configuration schematic. ........................................................................ 973-11 Example of a typical I-V curve. .........................................................................................984-1 Film thickness as a function of deposition time for ZnCdS chemically bath deposited at five prepared fractional zinc compositions at 85C. .................................................... 1134-2 Measured fractional zinc composition (xf) of five films chemically bath deposited at 85C for 45 minutes at five prepared zinc compositions, xp = 0.1, 0.2, 0.3, 0.4, and 0.5.....................................................................................................................................114

PAGE 11

11 4-3 As deposited fractional zinc composition, xf, as a function of deposition time for five prepared zinc compositions, xp = 0.1, 0.2, 0.3, 0.4, and 0.5, at 85C. .............................1154-4 Comparison of measured zinc compositi on of films deposited at 85C for 45 minutes with prepared fractional zinc compositions of xp = 0.1, 0.2, 0.3, 0.4, and 0.5. ................ 1164-5 Optical transmission as a function of wa velength for films depos ited at 85C with a prepared fractional zinc composition of xp = 0.1, at deposition times of 15, 25, 35, and 45 minutes. ................................................................................................................1174-6 Optical transmission as a function of wa velength for films depos ited at 85C with a prepared fractional zinc composition of xp = 0.2, at deposition times of 15, 25, 35, and 45 minutes. ................................................................................................................1184-7 Optical transmission as a function of wa velength for films depos ited at 85C with a prepared fractional zinc composition of xp = 0.3, at deposition times of 15, 25, 35, and 45 minutes. ................................................................................................................1194-8 Optical transmission as a function of wa velength for films depos ited at 85C with a prepared fractional zinc composition of xp = 0.4, at deposition times of 15, 25, 35, and 45 minutes. ................................................................................................................1204-9 Optical transmission as a function of wa velength for films depos ited at 85C with a prepared fractional zinc composition of xp = 0.5, at deposition times of 15, 25, 35, and 45 minutes. ................................................................................................................1214-10 Optical transmission as a function of wa velength for films deposited at 85C for 45 minutes with five prepared fractional compositions, xp = 0.1, 0.2, 0.3, 0.4, and 0.5. ...... 1224-11 Average transmission of three wavelength ranges for ZnCdS films deposited at 85C for 45 minutes with five prepared zinc compositions, xp = 0.1, 0.2, 0.3, 0.4, and 0.5. ....1235-1 Thickness contour plot of AZO film deposited at 0.4 kW with 40 sccm argon working gas flowrate. ....................................................................................................... 1405-3 Average AZO film thickness deposited at 0.4 kW with working gas flowrates of 30.0, 40.0, 50.0, and 60.0 sccm argon. ............................................................................ 1425-4 Thickness contour plot of HAZO film deposited at 0.4 kW with 40 sccm hydrogendoped argon working gas flowrate. .................................................................................. 1435-5 Average HAZO film thickness deposited at 0.4 kW with working gas flowrates of 20.0, 30.0, 40.0, and 50.0 sccm hydrogen-doped argon. ................................................. 1445-6 Comparison of AZO and HAZO average f ilm thickness as a function of sample position and working gas composition and flowrate. ...................................................... 145

PAGE 12

12 5-7 Sheet resistance contour plot of AZO film deposited at 0.4 kW with 40 sccm argon working gas flowrate. ....................................................................................................... 1465-8 Minimum AZO sheet resistance deposited at 0.4 kW with working gas flowrates of 30.0, 40.0, 50.0, and 60.0 sccm argon. ............................................................................ 1475-9 Maximum AZO sheet resistance deposited at 0.4 kW with working gas flowrates of 30.0, 40.0, 50.0, and 60.0 sccm argon. ............................................................................ 1485-10 Sheet resistance contour plot of HAZO film deposite d at 0.4 kW with 40.0 sccm hydrogen-doped argon working gas flowrate. .................................................................1495-11 Minimum HAZO sheet resistance deposited at 0.4 kW with working gas flowrates of 20.0, 30.0, 40.0, and 50.0 sccm hydrogen-doped argon. ................................................. 1505-12 Maximum HAZO sheet resistance deposited at 0.4 kW with working gas flowrates of 20.0, 30.0, 40.0, and 50.0 sccm hydrogen-doped argon. ................................................. 1515-13 Minimum sheet resistance of AZO a nd HAZO films deposited at 0.4 kW with working gas flowrates of 30.0, 40.0, and 50.0 sccm. .......................................................1525-14 Maximum sheet resistance of AZO a nd HAZO films deposited at 0.4 kW with working gas flowrates of 30.0, 40.0, and 50.0 sccm. .......................................................1535-15 Minimum resistivity of AZO and HAZO films deposit ed at 0.4 kW. ............................. 1545-16 Average transmission of AZO and HAZO films deposit ed at 0.4 kW. ........................... 1555-17 Figure of merit for AZO and HAZO films deposited at 0.4 kW. .................................... 1566-1 Rapid thermal processing schematic. ............................................................................... 1716-2 Minimum and maximum resistivity of AZ O films before and after annealing in forming gas, argon, and nitrogen ambients. ..................................................................... 1726-3 Minimum and maximum resistivity of HA ZO films before and after annealing in forming gas, argon, and nitrogen ambients. ..................................................................... 1736-4 Average transmission of AZO films before and after annealing in forming gas, argon, and nitrogen ambients. ..........................................................................................1746-5 Average transmission of HAZO films befo re and after annealing in forming gas, argon, and nitrogen ambients. ..........................................................................................1756-6 Figure of merit of AZO f ilms before and after annealing in forming gas, argon, and nitrogen ambients. ............................................................................................................1766-7 Figure of merit of HAZO films before a nd after annealing in forming gas, argon, and nitrogen ambients. ............................................................................................................177

PAGE 13

13 6-8 Cross-sectional image of HAZO annealed in argon am bient taken with a fieldemission scanning electron microscope. .......................................................................... 1786-9 Atomic force microscopy of AZO films a) as-deposited, and annealed in b) forming gas, c) argon, and d) nitrogen ambients. ..........................................................................1796-10 Atomic force microscopy of HAZO film s a) as-deposited, and annealed in b) forming gas, c) argon, and d) nitrogen ambients. ............................................................ 180

PAGE 14

14 LIST OF ABBREVIATIONS AES Auger electron spectroscopy AFM Atomic force microscopy AZO Aluminum-doped zinc oxide a-Si Amorphous silicon CBD Chemical bath deposition CIS Copper indium diselenide CIGS Copper indium gallium diselenide CGS Copper gallium diselenide CVD Chemical vapor deposition EDS Energy dispersive (x-ray) spectroscopy EDX Energy dispersive x-ray spectroscopy FWHM Full-width, half maximum GIXRD Grazing incidence x-ray diffraction H-AZO Hydrogen-aluminum-doped zinc oxide IBE Ion beam etching IBS Ion beam sputtering ICP Inductively coupled plasma spectroscopy ITO Indium tin oxide MBE Molecular beam epitaxy MOCVD Metal-organic chemical vapor deposition MSP Magnetron sputtering PACVD Plasma-assisted chemical vapor deposition PCVD Plasma chemical vapor deposition PMEE Plasma migration enhanced epitaxy

PAGE 15

15 PVD Physical vapor deposition RF Radio frequency SEM Scanning electron microscopy SIMS Secondary ion mass spectroscopy TCO Transparent conductive oxide XPS X-ray photoelectron spectroscopy XRD X-ray diffraction

PAGE 16

16 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy CHEMICAL BATH DEPOSITED ZINC CADMIUM SULFIDE AND SPUTTER DEPOSTED ZINC OXIDE FOR THIN FILM SO LAR CELL DEVICE FABRICATION By Andre Baran V August 2009 Chair: Oscar D. Crisalle Major: Chemical Engineering Chemical bath deposition, or CBD, is used to successfully deposit ZnCdS buffer layers that have many benefits over the traditional CdS th in films used in photovoltaic devices. The characteristics of films growth such as growth rate, optical transmission, and film composition is analyzed. The addition of Zn to CdS has im portant benefits that aid in improving device performance and therefore it is im portant to determine how the inco rporation of zinc affects film characteristics. Film thickness is found to incr ease with deposition time and growth rate is determined to be a function of the prepared ba th zinc composition. The zinc composition in the prepared bath is also found to affect the optical transmission of deposited film, most notably in the short wavelength region. The transparent conductive oxide ZnO is de posited via RF magnetron sputtering. The effect of aluminum and hydrogen doping is studied by sputtering an aluminum-doped zinc oxide target with argon (AZO) and an argon mixture with hydrogen (HAZO) The addition of hydrogen to the working gas decreases film resistance in addition to improving other characteristics. The working gas pressure and po sition under the target ar e also found to have a significant effect on film propert ies. A figure of merit calcula tion allows for a single valued representation of the combined characteristics of the deposited films. The figure of merit

PAGE 17

17 calculation takes into consider ation film resistivity, optical transmission, and provides a quantitative value of the potential performance in photovoltaic devices. The effect of thermal treatments on sputter deposited AZO and HAZO films are determined by way of rapid thermal annealing. High temperature annealing of ZnO is found to improve film characteristics. The gas ambient us ed in the thermal treatment process is found to be a critical parameter in post-a nneal film quality. Gas ambients of nitrogen, argon, and forming gas, a nitrogen and hydrogen mixture, were em ployed in the thermal treatment studies. In general, optical transmission and resistivity improve for all films under each ambient, although films annealed in forming gas show the most improvement, with resulting figure of merit values two to three times their as deposited values.

PAGE 18

18 CHAPTER 1 INTRODUCTION Renewables, broadly speaking, are fuels w hose use today do not reduce the supply for tom orrow. The major sources of renewable energy are solar, wind, and hydropower. These energies are by definition nondepletable. They cannot be completely used up as they are continually being replenished. Re newable energies are also sustainable, meaning that they can provide for the energy needs of the present wit hout reducing the availabi lity of energy in the future. Renewable energy is a promising alternative to the fossil fuels that have become vital to our industrialized society supplying the majority of our energy needs. Problems with fossil fuels have led to re-examination of their use and a search for alternatives. Some of the problems that exist with fossil fuels are environmental da mage, unequal distribution of resources, price instability, and supply constraints. Renewable en ergy offers solutions to the problems that have risen with traditional forms of energy by being re latively clean, widely available, and having a virtually unlimited supply. There have been impressive technical a dvances in renewable technology, specifically for electrici ty generation. The costs of renewable energy have dropped considerably and are almost competitive with fossil fuels. Typically, the majority of people seem to lik e the idea of renewables. Polls have shown strong and consistent public support. Trends in the polls show long-standing preference for renewables compared to other ener gy forms. These trends have b een consistent in the poll data for 20 years making it one of the strongest patterns in all the US national poll data on energy and the environment [1]. Renewable energy, however, is not a no problem s solution to replacing fossil fuels. A close look at renewable energy re veals that they too have thei r limitations. Costs have come

PAGE 19

19 down, but they are still more expensive than foss il fuels in several cases In addition, not all renewable resources are widely available or evenly distributed, and each has their own set of environmental impacts. For example, wind can be intermittent and therefore a poor solution for some electricity systems. Given the limitations we need to make the best use of renewables. One approach to achieve this is through a government policy to pr omote greater use of renewable energy. Our energy system is heavily regulated and publicly controlled, and always has been. This is changing, but government still plays a significant role in energy. Changes in government policy are the first step to inducing ch anges in the energy system. The industrialized worlds energy system provides dependable light, heat, and indust rial power and supplies, and is in many ways an impressive success. Though, it does have its share of problems, problems that are becoming more apparent and require attention. These problems again are environmental damage, inequitable distribution of fossil fuel resources, depleting supplie s, and economic damage due to price volatility. Eighty percent of the worlds en ergy is provided by fossil fuels, oil, coal, and natural gas. The burning of these fuels creates several urgent environmental problems. Climate change, the warming of the earth due to human induced increases in certain atmospheric gases, specifically CO2, is due largely to the burning of fossil fuels. Burning also results in emissions of other pollutants such as sulfur oxides, nitrogen oxides, and other particulates. These pollutants raise many public health and environmental concerns Biomass and waste accounts for 11% of the worlds energy consumption, nuclear accounts for 7%, hydropower accounts for 2%, and geothermal, wind and solar accounts for less than 1%.

PAGE 20

20 Renewable forms of energy are considered to be low impact on the environment in comparison to fossil fuels. Wind power, solar power, and hydropower contribute little or no CO2 to the environment. Therefore, they do not di rectly contribute to climate change. However, some will argue that hydropower contributes to climate change due to the methane produced from biological matter in dammed areas. Ther e are no direct air em issions from hydropower, solar power, or wind power fueled f acilities so they are not respons ible for any local or regional air quality problems. Wind and solar energies do not require mining or fuel transport systems. This eliminates them from bei ng associated with environmental damage. Although they are not entirely innocuous, in general, renewables are cleaner and more environmentally friendly than fossil fuels. As mentioned, the distribution of energy, a universal human need, is uneven across the planet. The uneven distribution of what has become a necessary re source results in international tensions, trade deficits, and c onstraints on global development. For example, the US is responsible for 26% of annual petroleum consum ption but only 12% of its production [2]. Approximately two-thirds of the worlds crude oil reserves are in the Middle East, and although Asia has 56 % of the worlds population, it has le ss than 30% of the worlds recoverable coal reserves [2]. Fossil fuel resources are bei ng depleted, and there are disagreements over just how much oil, natural gas, and coal are left and when it will become necessary to find alternatives. What is not in dispute, however, is that fossil fuel res ources are finite, and whether their lifetime is another 30 years, or 300 years, eventually thos e resources will run dry and it will become necessary to find one, or many, replacements. A nother concern of fossil fuels is the economic damage being caused as a result of its price vola tility. The prices of oil and natural gas are

PAGE 21

21 alarmingly volatile. For example, in the United States, the price of elect ric utilities paid for natural gas fluctuated between $2-$3 per 1000 ft3 for most of the 1980s and 1990s. Then, in 2000, gas prices started to climb and reached over $8 per 1,000 cubic feet by December 2000. In January 2001 the price peaked at $9.47 but by December 2001 they had fallen to $3.11 [3]. The behavior of the prices of crude oil behaved simila rly to the prices of na tural gas during that time period as the price of crude oil went from $10.60 in January 1999 to $25.60 in January 2000 [2]. Fluctuations such as these have the ability to cause considerable economic damage and complicate financial planning and forecasting. Technologies go through several stages as they progress from conceptual ideas to widespread use. They can be broken down into three major stages. Stage one is proof-oftechnology. It involves moving fr om a concept to more detailed plans, and eventually to a working model, larger-scale test plant, and then an operating facility. The next stage is designing to market. This means fine-tuning the engineering and design, tailoring the performance to meet users needs, and reducing cost, which is especi ally important for electricity generation. The third and last stage is market penetration. This is the stage in which the technology moves from a market ready idea to widespread use. Some renewable electricity-gen erating technologies are still at the proof-of-technology stage. Howeve r, many are at the third stage of development struggling to make the jump fr om a proven technology to widesp read use. Some of these technologies include wind power, photovoltaics, biomass, landfill methane, and geothermal power. It is misleading to talk about renewable en ergies as a whole wit hout specifying exactly which technology is being discussed. This work focuses on renewable energy converted from incoming radiation from the sun, or solar ener gy. Photovoltaics (PV), convert sunlight into

PAGE 22

22 electricity and have many attractive features. Photovoltaic devices are qu iet, dependable, have no moving parts, no noise or waste products, can be quickly installed, and can be sized to power anything from a single bulb to an entire comm unity. They are, however, quite expensive. Current costs are around $0.20-$0.40 per kWh for grid connected systems compared to $0.03$0.05 per kWh for coal or natural gas. Costs have decreased drastically in recent years and will continue to decrease. Howeve r, at the moment, PV devices are nowhere near being costcompetitive with fossil fuels. Fortunately, these high costs havent kept solar cell installations from booming. Photovoltaic production is growi ng worldwide at about 25% per year. Total electricity production from PV devices is minute in comparison to co al and other fossil fuels, but the use of solar cells in niche markets keeps growing which means cont inued technological and production advances leading to ev entual decreases in price. The solar resource is huge and could solely supply the worlds energy needs many times over. For example, assuming a module conversi on efficiency of 10%, a capacity factor of 22%, and insolation of 1 kW/m2, 14,000 square miles of PV pane ls (about 15% the land area of Nevada) could provide enough electricity for th e entire United States or 800 gigawatts. Obviously, a system like this would be immensely im practical. It wouldnt ge nerate electricity at night and would require massive c onstruction of new transmission li nes. Fortunately, solar cells work anywhere the sun shines, not just Nevada. Clearly they will produce more electricity in sunnier areas with increased levels of solar insolation, but even on a cloudy day, there is enough sunshine to produce electricity. Of all the forms of renewable energy, photovoltaics are the least resource-constrained. There are two basic types of PV modules, crystalline silic on and thin film devices. Crystalline silicon modules are us ed in almost all commercial-s cale PV systems. Commercial

PAGE 23

23 PV systems are those producing elec tricity for resale rather than for direct use. A crystalline silicon system consists of silicon and is mixed, or doped, with a small amount of a substance with a different number of electrons such as boron or phosphorous. As light hits the PV material, electrons dislodge and this movement of electrons cr eates an electric current. Crystalline silicon modules have relatively hi gh conversion efficiencies of approximately 1214%. Efficiency is defined as the percentage of energy in the sunlight st riking the module that is converted into electricity. Crystalline silicon modules are made from readily available materials, often from waste silicon from semiconductor chip manufacturing plants. Unfortunately, crystalline silicon modules are expensive to manuf acture, which hinder the overall cost of solar energy based on those modules. Thin film PV technology works on the same general principle as crystalline silicon modules but has the advantage of generating electric ity from a very thin film. For example, thin film modules can be incorporated into building materials such as roofing til es. Thin film PV devices typically require less mate rial to manufacture than crystalline silicon and are easier to produce on a large scale. Although material and production costs are lower, current thin film photovoltaic devices have lower effi ciencies than crystalline silicon modules. This is expected to change over the next five to ten year s as thin film technologies advance. Photovoltaic modules, also known as panels, can have a peak power output from of 50300 W. Panels can be assembled into arrays, cons isting of as little as two panels (for a small residential system) to thousands of panels for a utility-scale system of 100 kW or more. These panels are the fundamental component of a PV syst em but definitely not the sole component. PV systems consist of mounting brackets, supports, hard ware, and inverters. The costs of these other components are significant often consisting of one-t hird to one-half of th e total system costs.

PAGE 24

24 In general, there is not a definitive answer to the cost of photovoltaic power as costs can vary widely. Some of the reasons for this variation come from th e definition of cost itself, what parts are included, the size and application of th e system, the proposed location of the system, when the system is built, and whether it is an actu al or projected cost. Several studies have been carried out in order to paint a more complete picture of PV power and installation cost. An analysis of 220 grid-connected PV systems in the US found average first costs to be in the range of $7,400 per kW [4]. The majority of system s analyzed are sized for rooftops and were approximately 4.9 kW in size and were instal led around 1998. An analysis of 23 large-scale (>70kW) US PV systems found that costs for larg e PV systems are dropping and that this trend is likely to continue. PV system costs droppe d by 31% from an average of $1,035 per kW in 1996-1997 to an average of $810 per kW in 1999-2000. The PV modules accounted for approximately two-thirds of the cost, while th e remainder was for balance-of-system costs and installation. Operating costs for photovoltaic systems are uncer tain as there are few in operation. Fuel costs are zero and scheduled maintenance consis ts mostly of washing the modules to remove dust and dirt. Technical failures of the panels themselves are rare Inverters have been a source of problems however they are showing improved re liability in recent years. One review found maintenance costs for actual grid-tied systems to vary from 0.4-9.5 US cents per kWh [5]. Making reasonable, but arguable assumptions, these costs add up to about 40 US cents per kWh, which assumes a first cost of $8,000 per kW, $0.01 per kWh in operation and maintenance costs, a 22% capacity factor, and a lifetime of twenty years. This is about ten times more expensive than that of new natural gas turbines. There ar e other published estimates that are lower than 40 cents per kWh, but are still higher than most fossil fuel-based technologies.

PAGE 25

25 Despite the increased cost of solar ener gy, the photovoltaic market is booming. PV production worldwide grew from 70 MW in 1994 to over 200 MW in 1999. Technical advances in thin film production paired with a growing in terest in photovoltaics, thin film devices in particular, will allow for the continued rapid gr owth of the PV market. As mentioned, the current issues hindering industrial-wide thin film deployment are manufacturing costs, outdoor reliability, and most importantly, efficiency. However, much progress and success on these drawbacks is taking place in laboratories around the world. Recently, efficiencies in excess of 17% have been reported for thin film devices based on CuInSe2, CuGaSe2, and CuInGaSe2. The major drawback holding back these devices, from an industrial point of view, lie in the manufacturing costs. Thin film devices are cheap er to manufacture than crystalline silicon cells, however they often require complex deposition pro cedures and strategies. Hence, this research focuses on improvement in the quality of these thin film materials along with growth of thin film solar cell constituents via cost saving methods. Chapter 2 of this work provides a technol ogy overview of photovoltaic devices. The physics of photodevices are discussed, including the photoelectric e ffect and the role of solar radiation and band gap in a solar cell device physics. Typical solar cell materials are discussed beginning with the most common silicon-based photovoltaic devices, followed by the emergence of thin film devices based on CdTe, CuInGaSe2 alloys, CdS buffer layers, and transparent window layers such as ITO and ZnO. An overview of the two main deposition processes discussed in this work, chemical bath depos ition and rf magnetron sputtering, are detailed, including fundamentals of CBD film growth to types of targets and sputter sources, respectively. Chapter 3 details the methods and techniques used in this research. The stringent substrate cleaning procedure used on all samples is expl ained, along with the method for absorber layer

PAGE 26

26 fabrication at UF. The apparatus used to ch emically bath deposit ZnCdS buffer layers is discussed along with the sputtering system res ponsible for ZnO TCO growth. The large number of characterization techniques used in analysis of research in this work are detailed, which include inductively coupled plasma, x-ray di ffraction, and secondary ion mass spectroscopy, for example. Last is a discussion on photovoltaic device performance phys ics and the techniques used to measure device performance, including the I-V measurement and quantum efficiency techniques. Chapter 4 presents work focusing on the growth and characterization of ZnxCd1-xS thin film buffer layers which are deposited via chemi cal bath deposition. The CBD process used to deposit ZnCdS is detailed including substrate and equipment cleaning, preparation of dose solutions, and bath temperature conditioning. Th e difference between the fractional composition of zinc in the prepared solution and the as-deposited films is compared, where the subscript p, b, and f are used to denote fractional zinc concentrat ion in the prepared dose solutions, the actual bath solution, and the as-deposited film, respectively. Prepared fractional concentrations of zinc ranging from xp = 0.1 to xp = 0.5 are deposited on Corning 1737 gl ass substrates and the resulting fractional composition of zinc in the film (xf) is reported. The addition of Zn to buffer layer thin films is important, as it has the ability to ch ange the band gap of the film, thereby altering its optical transmission as ZnS has a larger band gap va lue than CdS. Therefore, it is important to know the correlation between prepared fraction of Zn and the amount of Zn that is actually deposited during the chemical bath process. Chapter 5 details the sputter deposition of ZnO thin film s, one of the most popular transparent conductive oxides used in solar cell applications. A lthough ZnO can be deposited by other methods such as spray pyrolysis or chemic al vapor deposition, sputtering is generally the

PAGE 27

27 preferred method of deposition and has the ability to produce high quality films at low costs. This work intends to improve upon the quality of ZnO, in partic ular H-doped and Al-doped ZnO, films by investigating a number of factors that affect film thic kness, sheet resistance, optical transmission, and structural characteristics. Films deposited at higher base pressures of approximately 2.5-3.0 x 10-6 Torr are studied as higher base pre ssures have the potential to lower the cost of thin film PV devices, as an increa se in industrial throughput can be achieved without the need for lower base pressures. The overall quality of Al:ZnO (AZO ) and H-Al:ZnO (HAZO) sputter deposited films is charac terized by a figure of merit calcu lation which takes into account resistivity and average transmittance from 400 nm to 800 nm. The figure of merit calculation provides an easy way to estimate the proposed eff ectiveness of a sputter deposited ZnO film for use in thin film solar cells or any other photodevice. Chapter 6 takes the sputtering of AZO and HAZ O thin films a step further and aims to investigate the effect of thermal treatment under three different ambient conditions, argon, nitrogen, and forming gas (96% N2, 4% H2). A brief discussion of the thermal annealing equipment is provided. Both AZO and HAZO film s are annealed at 450 C for one hour, with 10 min ramp and 15 min cool down cycles before a nd after, respectively. Structural, electrical, and optical characterization is pe rformed to determine the effect of ambient gas on post-annealed AZO and HAZO films. Atomic force mi croscopy provides a look at surface roughness variations between as-deposited ZnO and those films subjected to thermal annealing treatment. Secondary ion mass spectroscopy is also employed to elucidate any depth profile changes which might result as a factor of the annealing treatments.

PAGE 28

28 CHAPTER 2 TECHNOLOGY OVERVIEW Solar Cell Device Physics The design and m aterials used in fabricati ng a solar cell device is critical to its performance. Understanding the principals of photovoltaic device operation allow for a better understanding of how different materials can pro duce devices with a wide range of performance parameters. This chapter pres ents an overview of photovoltaic devices, the phys ics that drive their operation, and the most common parameters used to describe device performance. Photodevices There are tw o basic types of semiconductor photodevices, which are classified according to whether they convert electri cal energy into photo-energy or vice-versa. Photodevices that convert photo-energy into electr ical energy are termed photodetec tors or solar cells, depending on whether the purpose of the energy conversion is to detect informa tion regarding the photoenergy (photodetectors) or to produce electrical power (solar cells). The purpose of the second type of photodevice is to produce photo-energy from electrical energy. Light emitting diodes (LEDs) and laser diodes are two examples of this type of photodevice [6]. The Photoelectric Effect The first conversion of photo-energy to electr ical power is credited to Edmond Becquerel, who discovered the process in 1839. This convers ion of sunlight to electrical energy at the atom ic level has become known as the photoelectri c or photovoltaic effect The production of electricity by the photoelectric e ffect requires that a photovoltaic de vice, or solar cell, absorb a certain amount of light. The light radiation incident on solar cell s is not always fully absorbed, as a percentage of the radiation can be reflected or pa ss through the device. Electrical energy is only the result of incident light that is absorbed by the solar device.

PAGE 29

29 Once light is absorbed by a sola r cell, its energy is transferre d to electrons that use this energy to escape from their normal positions in the photovoltaic material. These electrons become free to flow as current through an elec trical current. The force to drive the current through an external circuit is provided by a built-in electric field. This el ectric field is created by contacting an n-type semiconductor, which ha s a surplus of electr ons, with a p-type semiconductor, which contains an abundance of hol es. Joining these two materials creates a p-n junction, an example of which is pictured in Figu re 2-1. When the n-type and p-type materials are contacted, a buildup of positive and negative charge occurs on the side of the n-type and ptype layer, respectively. This creates an electr ic field at the p-n junc tion that is capable of moving electrons towards the ne gative surface and holes towards the positive surface, thereby driving current through an external device. A band diagram depicting the creation of electrons and holes in a p-n junction via the photoelectric effect is shown in Figure 2-2 [7]. It can be seen that electrons travel to the n-type layer while ho les travel to the p-type side of the junction. Solar Radiation The Sun is the start at the center of our So lar System and is responsible for providing energy to Earth in the form of sunlight. This sunlight is responsible for all natural processes, such as creation of fossil fuels via photosynthesi s, and can be harnessed for use in synthetic processes, such as heating or electrical convers ion via the photoelectric effect. The Sun has a surface temperature of approximately 5,800 K and ther efore has a stellar clas sification of G2V. The surface of the Sun is believed to consis t primarily of helium and hydrogen, with trace quantities of other elements. The radiation spectrum emitted by the Sun is very closely approximated by blackbody radiation at 5,800 K, as can be s een in Figure 2-3 [6]. The solar radiation spectrum lies in the ultraviolet, visible, and infrared portions of the electromagnetic spectrum spanning 100 nm to

PAGE 30

30 106 nm. The solar constant is defined as the am ount of solar radiation th at hits Earth per unit area. When measured at the outer surface of the Earths atmosphere, the solar constant is approximately 1350 W/m2. This level of radiation is termed Air Mass 0, or AM0, as it has yet to pass through the Earths atmosphere. As the S uns radiation enters the atmosphere it becomes subject to energy loss as a result of absorption and scattering, f actors that are dependent on the incident angle of the incoming solar energy. The level of Air Mass incoming radiation experiences is calculated as Air Mass (AM) L1L2 1 cos (2-1) where is the solar zenith angle, L1 is the path length the sola r energy actually travels, and L2 is the shortest possible path lengt h. An Air Mass level of 1.5 ( = 48.19 ) is most often used for photovoltaic device testing, which correlates to a solar constant of approximately 1000 W/m2. Band Gap Energy Solar radiation incident on photovol taic m aterial has the abilit y to free electrons from the materials crystal lattice. In general, only photons with a certain level of energy can free electrons. This energy level is required to fr ee an electron is known as the band-gap energy, and is unique to each material. In cases where incident photons carry a higher energy than a materials band-gap energy, electrons are releas ed along with the extra energy of the photons, which is given off as heat. This undesired heat is often difficult to avoid as a large majority of incoming solar radiation is either below the band-gap energy of the photovoltaic material, or carries excess energy. Solar Cell Materials Photovoltaic devices can be m anufactured from a variety of different semiconductor materials, including silicon, and pol ycrystalline and single crystallin e thin films. Each of these

PAGE 31

31 materials possesses certain properties that allow their use in solar cell devices. The crystallinity of a photovoltaic material refers to the degree of order found in its crys tal structure. Single crystalline, multicrystalline, polycrystalline, and amorphous se miconductor materials have all found use in photovoltaic devices. Absorption is an important property of solar cell materials and specifies the ability of a material to absorb incident light. The ab sorption coefficient of a material is typically a function no t only of the material itself, but also of the wavelength of the light to be absorbed. If the energy of incoming radiation is below the band gap energy of the material, the light will not be absorbed. The ba nd gap energy of a material refers to the minimum energy needed to free an electron, or more specifically, to move an electron from the low energy valence band to th e high energy conduction band. Thin Film Solar Cell Materials The solar ce ll market is currently dominated by crystalline and polycryst alline silicon. The price of traditional silicon based photovoltaic de vices has been significantly reduced over the past few years as a result of increased producti on volume. Unfortunately, even economies of scale are not expected to aid silicon in mee ting cost per watt effi ciencies demanded by consumers and industry, which is typically in th e range of $3-5/Wp (watt peak power). Solar cell devices based on thin film technologies have made remarkable prog ress in terms of high conversion efficiencies, long-term stability, and industrial scal e manufacturing capabilities, which is why thin film devices are believed to be capable of reaching th e desired cost per watt goals. Thin film technologies are gaining industrial in terested as they have the possibility to drastically cut manufacturing costs as compared with traditional silicon cells. Photovoltaic devices based on thin film technologies use less material, generally require fewer processing steps, allow for simple device processing, and ar e capable of being cheaply manufactured for use

PAGE 32

32 in large-area modules and arrays The manufacturing cost reduc tion stems from the fact that design and manufacture of solar cells based on thin film technologies contain many common elements and steps. In general, large scale manufacture of solar cells aims to choose a given technology based on high efficiency, ease of manu facture, reliability, material av ailability, and environmental sensitivity. Alloys of amorphous silicon (a-Si:H), cadmium telluride (CdTe), and multinary copper indium selenide (CIS) and its alloys with Ga are three thin film technologies that are the current leading contenders for large scale ma nufacture. Figure 2-4 s hows the chronological evolution of conversion efficien cies of a-Si, CdTe, and CuInSe2 thin film technologies from 1975 through 2005 [8]. In the figure it can be se en that CIS based thin film technology shows the promise of exceeding the required cost/performance levels to compete with other sources of energy. All three of these thin film technologies are only begi nning to become commercialized on a large scale. Although leading on paper, CIGS based technology trails in commercialization with a current production capacity of approxima tely 100 kW/y compared to around 1 MW/y and 35 MW/y for CdTe and a-Si technologies, respectively. Amorphous Silicon Alloys Solar Cells Am ong thin film technologies, p-i-n junction-t ype devices based on a-Si:H and its alloys are the most developed and commercially available, a result of specific technical properties of the material. They use the p-i-n type device configuration as a-Si:H and its alloys have intrinsically high defect densitie s and low minority carrier lifetimes, and require field assistance for collection of photoelectric generated carriers. Amorphous sili con and its alloys have high optical absorption coefficients of around 105 cm-1, a band gap that can be adjusted from 1.1 eV to 2.5 eV depending on the level alloying, are fair ly easy to fabricate with many different techniques under low temperatures, and are capable of being used in multijunction devices,

PAGE 33

33 which are extremely efficient at utilizing incomi ng solar radiation. There are also a number of disadvantages of amorphous silicon thin film technology, such as severe cell efficiency degradation by as much as 15-20% Electronic properties of the material also tend to degrade with increased deposition rates larger than a few /s, which limits throughput of device fabrication. Module effici encies are typically lower than other thin film technologies, as seen in Figure 2-4. Significant amounts of research has helped to dimi nish these disadvantages and much progress has been made in achieving incr eased deposition rates a nd higher efficiencies. Amorphous Silicon Growth Methods Many deposition techniques can be used to grow a-Si:H alloy m aterials for use in photovoltaic devices. Some of the most comm on techniques include sputtering, plasmaenhanced chemical vapor deposition (PECVD), eletron-cyclotron resonance glow discharge (ECR), photo CVD, and plasma CVD. Plasma-enha nced chemical vapor deposition is perhaps the most widely used deposition method, however it suffers from very low deposition rates, especially of i-layer amorphous silicon, which is the bottleneck for all deposition techniques. These low deposition rates, around 1-2 /s, are no t desirable for commercial purposes as they negatively affect throughput and do not result in cost reduction. Extensive efforts have been aimed at increasing growth rate of the i-layer, and increased deposition rates of 10 /s [9] and up to 15 /s [10] have been achieved using very high frequency RF plas ma-enhanced chemical vapor deposition (VHF-PECVD). Higher deposition rates of a-S i:H alloys have been achieve d with a hot-wire chemical vapor deposition process (HWCVD), with device quality films grown at rates up to 167 /s. The HWCVD process has been used to create a-Si:H devices with efficiencies of 5.5% at 18 /s, 4.8% at 35 /s, 4.1% at 83 /s, and 3.8% at 127 /s [11]. Although efficiencies decrease, the HWCVD technique is capable of increasing depo sition rates, thereby in creasing throughput and

PAGE 34

34 possibly decreasing cost. Unfort unately, cells fabricated by the HWCVD process have exhibited significant photoinduced degradation. Very high deposition rates of 0.3 m/s to 1.6 m/s have been achieved using an atmospheric pressure plasma CVD technique which employs 150 MHz VHF power to generate a plasma of a He, H2, and SiH4 gas mixture [12]. Despite the high quality films produced at high deposition rates, this atmospheric pressure plasma CVD process has not been used for photovoltaic device fabrication. Copper Indium Diselenide Research of sem iconductor materials for thin f ilm devices has extended to the investigation of ternary and quaternary materials where CuInSe2-based solar cells ar e being considered a promising potential. CuInSe2 is a direct band gap material with excellent photovoltaic properties as a polycrystalline absorber layer for thin films. CIS has a band-gap energy between 1.5 eV and 1.55 eV, smaller than the band-gap of silicon but an excellent match for the solar spectrum [13]. Efficiency ( ) of the conversion process is a basic parameter that is used to evaluate all solar cells. Thin film solar cells based on CIS have achieved 12.2% total area efficiency in 1996 [14] and 11.4% total area efficiency prepared by ra pid thermal processing in 2001 [15]. The maximum theoretical efficiency of CIS is 28.5% [16] The efficiency of CI S devices is assisted by its ability to invert via de position of CdS and become n-type even though the bulk of the sample is p-type. CIS devices normally have a low open circ uit voltage (0.729 V) to band gap ratio compared to silicon devices and a large short ci rcuit current creating a relatively large series resistance loss and compromising cell performa nce [17]. From 1978-1981, efficiencies of CIS cells increased from approximately 4% to 10%. This improvement stemmed from the research

PAGE 35

35 efforts that went into solar cells as a result of the energy crisis of the 1970s. Research has continued to today resulting in the variety of commercial products now available [18]. Absorbance of CIS can be as high as 105 cm-1, making it the most light absorbing semiconductor known. With CIS, it is possible to achieve high efficiencies with considerably less material consumption than traditional PV technology. Only half a micron of CIS is sufficient to absorb 90% of incoming photons [19] PV devices created with CIS exhibit long life and consistent performance as a result of the excellent stabili ty and radiation hardness of CIS [20]. CIS efficiencies have reached 18.8% for lab-sized cells and over 10% for complete modules [21]. Siemens Solar Industries have r ecently entered the market as a commercial product for terrestrial applicati ons with CIS-based devices [22] These applications include generating power for remote locations, genera ting power for developi ng countries without suitable power infrastructures, and the space industry. Copper Gallium Diselenide The ternary com pound, copper gallium diselenide (C GS), has a relatively large band gap of 1.7 eV [13] and a high ab sorption coefficient of = 3 x 104 cm-1 at 1.7 eV [23]. CGS is a member of the I-III-VI2 semiconductors. CGS band gap ener gy will decrease as the Cu/Ga ratio increases [24]. The theoretical limit for CGS device efficiency is 26%, however only recently have CGS-based solar cells overcome efficiencies of 6.2% [25]. Now, efficiencies have reached 9.7% for single crystal devices and 9.3% for thin films [26]. Both of th ese recorded values are still well below the theoretical limit. In comparison to the rapid progress made with CIS and CIGS based solar cells, the efficiency of CGS cells are still low, even t hough CGS cells have been investigated for more than 25 years. The most recognized downfall of CGS based solar cells is the low open-circuit

PAGE 36

36 voltage compared to the band gap. Theoretica lly, CGS-based photovoltaic devices should be capable of producing an open-circuit voltage of approximately 1.2 V. However, in practice, solar cells with CGS absorbers have been lim ited to open-circuit voltages of 0.9 V and 0.75 V, using Ga-rich and Cu-rich absorber layers, respec tively [27, 28]. Decreased defect densities and a decrease in tunneling-enhanced recombination has been determined as the cause for improved device performance of Ga -rich CGS films [29]. CGS wide gap chalcopyrites have material a nd growth related defects that limit device performance. CGS tends to exhibit a tetragonal structure, as a re sult of this arrangement it is susceptible to a large concentration of vacancies and anti-site defects. Slightly Cu-poor CGS absorbers have shown to have better performan ce than stoichiometric ones. For Cu-rich CGS, deviation from stoichiometry is facilitated by th e formation of a second phase, and therefore the material develops more defects. CGS is often used as a top cell in tandem cells. It responds to the short wavelength portion of the solar spectrum. In addition, it allows long wavelength photons to pass through and reach the bottom layer of the tandem cell, often a CIS la yer. It has been estimated that a CGS/CIS tandem structure could reach e fficiencies greater than 30%. Copper Indium Gallium Diselenide Adding gallium to copper indium diselenide forms a CuIn1-xGaxSe2 alloy known as CIGS and can increase the band gap energy to appr oximately 1.7 eV therefore enhancing the opencircuit voltage of the solar cell creating a higher performing devi ce. The Ga content can be adjusted with the goal of optimizing the CIGS absorber band gap profile, improving the opencircuit voltage, and the short-ci rcuit current. High performan ce CIGS-based devices generally require an extremely high temperature to produce the absorber layer, the reasons for which are not entirely clear.

PAGE 37

37 CIGS cells have reached higher efficiencies than CGS-base d thin film devices. The maximum conversion efficiency achieved for a CIGS thin film solar cell is 19.5% (VOC = 0.694 V, JSC = 35.2 mA/cm, FF = 79.7%), which wa s constructed at the National Renewable Energy Laboratory [30]. NREL was also the first to reach efficiency greater than 20% for a CIGS thin film solar cell. The maxi mum theoretical efficien cy for a single junction CIGS solar cell is 27.5% [16]. Cadmium Sulfide In fabrication of thin film photovoltaic devices, deposition of the buffer layer occurs on top of the absorber layer. T he buffer layer of a solar cell is intended to prevent diffusion of impurities into the absorber layer from deposition of the transparent conducting oxide layer. The most common buffer layer for high efficiency devices is a CdS buffer layer, although interface passivation and the establishment of an inverted region in the absorber are other proposed benefits [31]. CdS buffer layers are highly resi stive and are most often deposited via chemical bath deposition, although other techniques are very common such as the successive ionic layer adsorption and reaction (SILAR) technique. The band gap of CdS is very low, 2.4 eV, and is not able to transmit all useful incoming radiation. Therefore, optimization of device stru ctures based on the thickness of the CdS buffer layer is necessary, as unacceptable absorption occurs if the CdS buffer layer is too thick, blocking out a large section of incident photons This absorption results in poor device perforance as a result of decrea sed short-circuit current. On the other hand, the device opencircuit voltage can be hampered if the deposited CdS buffer layer is too thin [32]. Diffusion of cadmium ions into CIGS absorber layers is thou ght to create such high efficiencies by formation of a pn-junction.

PAGE 38

38 Although CdS is perhaps the most widely used buffer layer, one of the main problems of CdS technology is its relatively low band gap value of 2.4 eV. This reduces the current that can be collected as there is a cons traint on the short wavelength por tion of the solar spectrum that reaches the absorber. For higher band gap cells, th is limit of the short wavelength part of the solar spectrum can become a severe hindrance for device performance [33]. Zinc Sulfide Zinc sulf ide is an attractive alternative buffer layer on CGS-based thin film solar cells as it exhibits a wide optical band-gap energy (3.8 eV) [34] coupled with the fact that ZnS raises fewer ecological concerns than CdS. Photovoltaic devices fabricated with ZnS buffer layers have reached efficiencies of 18.6% [35] Unfortunately, the ZnS comes with some downfalls, such as the difficulty of ZnS diffusion into CIGS films co mpared to CdS. An external force, such as heat, is required to enhance the diffusion in to CIGS to create the buried p-n homojunction. Compared to the best performing CdS/CIGS sola r cells with a total ar ea efficiency of 19.5% AM1.5G, ZnS-based buffer layers are a very pr omising replacement for CdS buffer layers on CIGS cells [36]. Zinc Cadmium Sulfide Zinc cadm ium sulfide is used as an alternativ e to CdS buffer layers in an effort to produce a more environmentally friendly CIGS solar cell by reducing the amount of the toxic cadmium component used. These ZnCdS buffer layers also have the best lattice matching to CGS absorber layers. Addition of Zn to CdS results in a band-gap energy that is between CdS (2.4 eV) and ZnS (3.8 eV). The result of this increased bandgap energy is increased sh ort-circuit current in the short wavelength region. In an effort to maximize photovoltaic device po tential, absorber and buffer layers electron affinities should match, yielding an increase in the open-circuit voltage. A loss of current occurs

PAGE 39

39 in cells fabricated with CdS buffer layers and CG S absorber layers as CdS has a larger electron affinity, 4.5 eV for CdS and 3.9 eV for CGS. Addi ng Zn to CdS buffer layers results in electron affinity ranges from 3.7 eV to 4.5 eV. Substituting ZnCdS for CdS buffer layers minimizes current loss in the cell and maximizes the open-circuit voltage. Another benefit of the addition of Zn to CdS is the improved lattice matching to CIS and CGS based absorbers. Addition of zinc decreases the lattice parameter of ZnCdS from 5.82 (CdS) to 5.42 (ZnS), which more closely matches that of CGS (5.61 ) and CIS (5.78 ). Although increased lattice matching is beneficial, the increase in Zn content also results in films with increased resistivity [37]. It has been proposed that a ZnCdS buffer layer film w ith 30% zinc content should be a successful buffer layer for CGS absorbers [38]. Zinc Selenide Zinc selenide is another bu ffer layer that is good for use in a wide band-gap CGS solar cell. It p rovides another alternative buffer la yer to replace the use of toxic CdS in CIGS photovoltaic devices. Zinc selenide is a wide band-gap materials with a band-gap energy of 2.7 eV [34]. The deposition of ZnSe (5.668 ) buffer layers on CGS (5.614 ) absorber material creates a practically strain -free interface with a lattice mismatch of less than 1%. This is a large improvement in comparison to the latti ce mismatch of approximately 4.2% between CdS and CGS. Unfortunately, conversions efficiencies using ZnSe are much lower than those using CdS, as CGS solar cells using ZnSe buffer laye rs deposited by MOVPE have only reached 3.3% conversion efficiency [39]. Using CIGS absorber layers, conversion efficiencies of 13.7% have been obtained with ZnO/ZnSe/Cu(In,Ga)(S,Se)2 device structures [40]. Further experimentation of ZnSe buffer layers is needed in order to increase solar cell conversion efficiencies.

PAGE 40

40 Indium Tin Oxide Indium oxide, doped with tin (ITO), was deve loped in the early 1950s and has the lowest resistivity, on the order of ~1 x 10-4 cm. Indium tin oxide is the most widely used transparent conductive oxide in optoelectronic devices [41], as they have been studied extensively because of their combined unique transp arent and conductive properties. Indium tin oxide thin films were the first TCOs that were actually useful in any pract ical application. A stable supply of ITO may be difficu lt to achieve for the growing market of optoelectronic devices due to the expense and availa bility of indium [42]. Half of the indium worldwide is used in flat panel display applications [41]. This problem is compounded by the fact that the amount of indium available in nature is very small cau sing it to be quite and expensive material. There are no mines built specif ically to remove indium from nature due to the trivial amount available, it is only obtained as a byproduct from zinc mines and other metals. In 1995, the price of indium spiked at $550 pe r kilogram, creating economical as well as technically motivation for the creation of alternative tr ansparent conductive oxides. Indium tin oxide thin films are highly de generate n-type semiconductors with low electrical resistivity. The low resistivity of ITO is due to a high carrier co ncentration as a result of the Fermi level being located above the conduction level [41]. Several deposition techniques can be used to deposit ITO thin films, such as chemical vapor deposition, magnetron sputtering, evaporation, and spray pyrolysis. These techniqu es require either a high substrate temperatures during deposition or high temperature thermal tr eatment, annealing, after deposition. These high temperatures often cause damage to the surface of both the substrate and the film. Deposition of ITO in the manufacturing environment is usua lly performed by dc-magnetron sputtering. The choice of target materials, ceramic or metal, is dependant of the film quality sought and the amount of process control available. Variables that are adjusted during process optimization

PAGE 41

41 include oxygen partial pressure, total gas working pressure, residual water vapor partial pressure, substrate and target temperature, sputter pow er, and target compositi on and configuration. Zinc Oxide Zinc oxide thin film s were developed in the 1980s as an alternative to ITO with similar resistivity. Zinc oxide is one of the few metal oxides that can be used as a transparent conductive oxide. It has advantages over other materials such as In2O3 and SnO2 due to its unique combination of interesting properties. Zinc is abundant, inexpensive, non-toxic, has good electrical, optical, and piezoelectri c behavior, and is stable in a hydrogen plasma atmosphere. It is a wide band-gap semiconductor with a band -gap energy of 3.43 eV. Techniques such as sputtering, spray pyrolysis, and chemical vapor de position have been used to deposit intrinsic zinc oxide thin films. Aluminum and galliu m doped zinc oxide films have recently received attention for use as transparent electrodes in thin film solar cells with resistivity on the order of 1 x 10-4 cm have been achieved. It has been found th at at temperatures greater than 150C the electrical properties of zinc oxide become unsta ble. This problem can be resolved by using extrinsically doped films. Thin film ZnO can be deposited with extrem ely high transparency while still maintaining high conductivity, especially when doped with aluminum or galliu m. The high transparency of zinc oxide films allows for increased transmission in solar cells allowing more light to reach the absorber layer, resulting in th e creation of more el ectron-hole pairs. Conductive ZnO films easily form ohmic contacts with aluminum. This ohmic contact is essential for collecting the current generated from photovoltaic devices. The use of ZnO has been researched extensively for use in other applications such as organic light-emitting diodes and transistors. Thin films of zinc oxide have most often been prepared using various techniques su ch as spray pyrolysis, evaporation, chemical vapor deposition, and sputtering.

PAGE 42

42 Chemical Bath Deposition The deposition of m aterials via ch emical bath deposition is a t echnique that has been in use as far back as 1835, when CBD was used to deposit silver in the fabricatio n of mirrors. In CBD, chemical deposition occurs on a solid substrate from a reaction that occurs in a chemical bath solution. The deposition can proceed along different pathways depending on whether the reaction in solution occurs homoge nously, heterogeneously, or vi a decomposition of a complex. The type of reaction pathway must be known in or der to understand how to control the rate of growth, especially when stoichiometry of the deposited material is important. Fundamentals The solubility product of a salt is a fundam ent al concept behind the mechanics of chemical bath deposition. A sparingly soluble salt, such as CdS, when dissolved in water will result in the following equilibrium CdS( s ) Cd2 S2 (2-2) where CdS is in the solid phase The small concentration of Cd2+ and S2ions that are present in the solution is given by the solubility product Ksp Cd2 S2 (2-3) which is defined as the product of the ion concentrations. As a salt become more soluble, the concentra tion of dissolved ions increases which results in a larger solubility product Ksp. Since the number of ions disso lved affects the size of the solubility product, it can be more accurately written as Ksp Mn aXm b (2-4) where the dissolution of salt MX results in the following equilibrium MaXb aMn bXm (2-5)

PAGE 43

43 The solubility products of many popular chemically bath depos ited compounds are listed in Table 2-1. Cadmium sulfide and zinc sulfide, two compounds studied in this research, have solubility products of Ksp = 1 10-28 and Ksp = 3 10-25, respectively. The majority of chemical bath deposition reac tions are conducted in alkaline solutions. Under these conditions the concentration of OHcan impede deposition of the desired compound. For example, water at pH = 10 has an OHconcentration of approximately 9 10-4 at 60C. With such a high concentration of ions the precipitation of metal hydroxides removing metal ions from solution becomes a problem. Ther efore, a complexing agen t, or ligand, is often added to the solution in order to prevent prec ipitation of the metal hydroxide. Addition of the complexing agent also serves the secondary purpose of impeding bulk precipitation of the desired compound by reducing the c oncentration of the free metal ions. Ammonium hydroxide is often used as a complexing agent for the chem ical bath deposition of cadmium, and forms a cadmium tetraamine complex given as Cd2 4 NH3 CdNH34 2 (2-6) with a solubility product of Ksp CdNH34 2 Cd2 NH34 1.3 107. (2-7) Nucleation Nucleation is a physical proce ss that describes how precipita te particles of a com pound will form once the solubility product is surpasse d by the concentration of ions in solution. Thermodynamics and kinetics play a large role in determining the precipitation of dissolved ions. This precipitation can occur via homogenous or he terogeneous nucleation, which refer to nuclei formation in the solution or on a substrate, respectively.

PAGE 44

44 Homogeneous nucleation generally refers to precipitate in the bulk solution. First, embryos are formed from the collision of mo lecules or individual i ons. Embryos are not considered nuclei at this stage as they are typi cally unstable and have the ability to redissolve. The embryos can grow in size as a result of additional collisions with ions or molecules, or with other embryos, which necessary in order to form stab le nuclei. If an embryo is too small, it will be thermodynamically unstable as a result of high surface energy and have an increased likelihood of redissolution. Low temperatures aid in the formation of thermodynamically stable nuclei by kinetically stabilizing smaller embryos long enough to grow in size A critical radius exists where particles have an equal chance of fo rming stable nuclei or re dissolving into solution, and is a function of volume and surface energies of the embryo. The surface energy required for embryo formation is E s 4 R2 (2-8) where is surface energy per area and R is embryo radius. When an embryo is formed the amount of energy released is Ev 4 3 HR3 (2-9) where H is the heat of solution and is solid density. In general, the average critical radius is around 100 molecules, which is approximately 0.5 nm to 1 nm. Heterogeneous precipitation occurs when ions or unstable embryos pr ecipitate, or absorb, onto a surface or substrate. In this case, nucleation occurs on an interf ace and there requires less energy than homogeneous nucleation, which tends to require a high degree of supersaturation in order to form stable nuclei. As a result, heterogeneous nucleation is energetically favored and generally dominates homogenous nucleation. Nucl ei absorbed onto a surface can increase in size

PAGE 45

45 by addition of particles from the solution or su rface diffusion. In many cases it is possible for individual ions to chemisorb ont o substrates to create nuclei that can further react and grow. Solid precipitate in the so lution is not always eviden ce of homogenous nucleation and growth. Inherently, heterogeneous nucleation will result in film growth on all surfaces in a chemical bath deposition reaction vessel, which includes any contaminatio n that may be present in the system. Any foreign material present in the solution will act as heterogeneous nucleation sites. These particles can then collide with other embryos or surf aces, which leads to the possibility of sample contamination. Therefore, it is extremely important that chemical bath solutions be well filtered to prevent pollution of the sample with foreign particles. Film Growth The film growth mechanism of chemically bath deposited films is generally divided into two main categories. The first involves reaction of free ions in solution, known as the ion-by-ion mechanism. The second involves the clustering of molecules that even tually result in the formation of film particles or crystals known as the hydroxide cluster mechanism. Ion-by-ion mechanism The sequential reaction of free i ons in solution is the basis of the ion-by-ion mechanism, and is generally considered the mechanism res ponsible for the majority of film growth. Heterogeneous nucleation occurs on all surfaces pr esent in the chemical bath deposition reaction vessel. Van der Waals and electr ostatic interactio n are primarily responsible for adhesion of one particle to surfaces or one anothe r. These interactions result in film adhesion to a variety of surfaces, including hydrophobic materials such as Te flon. Film growth is slowest during the initial nucleation stage, as deposition occurs fast er after nuclei have formed on a given surface. Once nuclei are formed, crystals grow until terminated by absorption of foreign particles from solution or steric hindra nce by nearby crystals.

PAGE 46

46 A schematic showing the deposition and growth of cadmium sulfide crystals via the ionby-ion mechanism is shown in Figure 2-5 [43]. First, cadmium and sulfide ions diffuse through the solution towards the surface or substrate. On ce reached, the surface aids in nucleation of the ions and CdS nuclei form. The flux of cadmium and sulfide ions to the surface continues which results in formation of new nuclei or further growth of CdS crystals. Finally, crystals continue to grow in size until growth is terminated by steric hindrance or absorption of foreign particles. Hydroxide cluster mechanism Nucleation via the hydroxide cluster mechanis m is considered simpler than ion-by-ion growth, as the solid phase is already present in the form of a metal hydroxide. Deposition begins with the adhesion of this metal hydroxide to the surface or substr ate. These clusters are ten converted in deposits of the desired compound. Film growth proceeds by the adhesion and conversion of additional metal hydroxide clusters to bare substrate or deposits that have already been converted. Also, metal hydroxide clusters exist in the bulk solution where they can be converted without adhering to any surface or subs trate. In general, the hydroxide cluster mechanism results in a small crystal size distri bution as compared to ion-by-ion growth where crystal growth can proceed via nuclei already present. A schematic depicting the hydroxide cluster mechanism applied to the chemical bath deposition of cadmium sulfide is shown in Figur e 2-6 [43]. First, cadmium hydroxide clusters present in the solution slowly diffuse towards and a dhere to the surface or substrate. Sulfide ions react with cadmium hydroxide clus ters, both in solution and adhere d to the surface. The reaction results in the inward exchange of hydroxide ions by sulfide ions, and continues un til essentially all of the hydroxide ions are replaced. The resulting clusters aggregate on the surface or substrate to form the cadmium sulfide film.

PAGE 47

47 Sputter Deposition Sputtering is a complex subject that involve s many physical and chemical processes that are all interrelated. Although many complex details of the process remain unknown, sputter deposition is a widely used tec hnology that finds many applications in research and industry, including optical coatings for mirrors and filters solar cells, magneto-optical storage media, transparent conducting electrodes, and wear-resistant co atings. There are many advantages of sputter deposition that are the cause for its popularity including the ease w ith which researchlevel results can be scaled to industrial processes. Sputter de posited films generally exhibit excellent film uniformity, even over larger substrates, and film thickness and roughness can generally be well controlled. Film uniformity is dependant on source geometry and size, operating pressure, and target-to-substrate dist ance. Substrate motion and aperture masking (uniformity shields) are both used to improve film uniformity. Films can be deposited with nearly bulk-like properties at high growth rate s and show excellent adhesion to a variety of substrates. Sputtering is a kinetic process involving momentum exchange, not a chemical process, and is therefore extremely versatile in its implementation. There are two main sputter sources used in sputtering, glow discharge sources and ion beam sources. Glow discharge sputter sources in clude diode, triode, and magnetron sputtering. Although each source has its own co nfiguration and set of advant ages, they all function on the same basic mechanism, exchange of momentum between energetic particles and surface atoms, in order to eject particles. Planar magnetron so urces are perhaps the most widely used sputter source due to their high efficien cy and scale-up ability, however nonmagnetron sources, such as ion beam and rf planar diode sources, are ex tremely common. Although ion beam sources are typically used for sputter deposition, they can also be used for etching applications.

PAGE 48

48 Chamber Features Sputtering is a fairly mature technology that is used in a number of industrial applications. Its industrial popularity stems from the straight forward setup of these sputtering system, which all feature very similar components, regardless of application or materi al being deposited. A schematic depicting the most comm on features of sputtering system s is shown in Figure 2-7 [44]. Most sputtering vacuum chambers are composed of stainless steel or mild steel. A high vacuum pump is attached to the system to achieve a base system pressure on the order of 1 10-6 Torr or lower, depending whether ultrahigh vacuum conditions are necessary. Multiple pressure gauges, including thermocouple, convectr on, capacitance monomers, and i on gauges, are typically used to provide accurate measurement in both high and low vacuum conditions. Mass flow controllers and throttle valves are two methods employed to maintain the working pressure necessary for sputter deposi tion. Generally, downstream pressure control (throttle valves) is considered s uperior to upstream control of gas flow (mass flow controllers), although in practice throttle valves are set at a fixed value and the working pressure is controlled upstream. Most systems can generally accommodate a large number of samples and are capable of substrate heating during deposition. Sputtering of numerous samples is accomplished by placing samples on a circular platen that allows multiple samples to be deposited by rotating the platen under the target. Bombardment, or etch ing, of samples can be accomplished by applying a bias to the substrate holder. Targets The sputter target is an integral part of th e system that dictates what materials can be deposited and under what conditions. Planar targets are composted of a wide range of materials, and are fabricated by mechanical, sintering, and metallurgical techniques. Two-source sputtering is possible, where two targets are used to sput ter deposit a film, which requires substrate rotation

PAGE 49

49 through both target deposition regions. However, the best results are obtained when using targets with alloy compositions identical to those of the desired film. Perhaps the largest concern related to targets is incomplete utilization of target material, especially in magnetron setups. Water-cooling is required for most magnetron s ources, where oversize targ ets are required in order to achieve sufficient cooling. As a result, erosion tracks are generally significantly smaller than target diameters, and therefore target ut ilization of 25% or le ss is extremely common. Sputter Sources There are two main categories of sputter sour ces, glow discharge and ion beam sources. Among them, they can be further separated into nonmagnetron and magnetron sputter sources, which use a magnetic field in order to alter the s putter source, target, and substrate interactions. Nonmagnetron sputter sources The simplest sputter source is the planar diode a disc-shaped cathode ta rget consisting of the material of interest, approximately 5 to 10 cm in diameter. The target is bonded, either by solder or conducting epoxy, and generally water-c ooled. The target is protected by a ground shield, which acts to protect the target from sputtering of the s upport structure and side of the target. The efficient use of target material in diode s puttering is one of its la rgest advantages. The diode electrodes are fairly larg e and the resulting electric fi eld between them is extremely uniform, which results in an ion flux that is almost constant over the entire target surface. On the other hand, the inefficient use of secondary el ectrons is a large disadvantage of the diode technique. With a diode sputter source, substrates will be bombarded with energetic electrons, which can result in increased substrate temperatures. As a result, diode sputtering is limited to decreased deposition rates as compared to ot her sputter sources capab le of low-pressure operation.

PAGE 50

50 A triode source improves on the diode source by adding a heated filament. This filament provides electrons that are independently able to sustain the glow discharge, which greatly increases ionization efficiency. The advantage of th e triode source is that the discharge is able to operate at lower pressures and vol tages, and therefore is capable of much higher deposition rates as compared to planar diodes. In some cases, systems that operate with reactive gases can shorten the lifetime of the filament. Magnetron sputter sources Magnetron sputtering is capable of coating he at-sensitive materials as its generally considered a cold deposition process. Magnetron sputter sources operate by combining the electric field with a magnetic field of approximately 50 to 500 gauss. This combination is able to control secondary electrons, and confines them to drift in front of th e target surface, which results in greatly increased efficiency. As a re sult, magnetron sputter sources can operate at low pressures and voltages. However, a large disa dvantage of the magnetr on configuration is the poor utilization of target materi al, as the erosion of the targ et is non-uniform. The typical magnetron setup utilizes at most 30% of the original target material. By rotating the target with respect to the magnets, its possible to increase target utilization to va lues of 90% or greater. Ion beam sputter sources Ion beam sputter sources are unique for a numbe r of reasons. Most se tups offer complete isolation of the substrates from the ion gene ration process and therefore minimal interaction between processes occurring at the target and su bstrate, respectively. As a result, substrate heating is kept to a minimum. Ion beam sources also offer control of the ion impact angle and spot size along with independent control of i on energy and current density. The sources require low background pressure, which results in le ss scattering of sputte red particles and gas incorporation. The most common ion beam sour ce is the Kaufman source, which is capable of

PAGE 51

51 providing an Ar ion beam up to 10 cm in diam eter, with an energy range between 500 and 2,000 eV.

PAGE 52

52 Table 2-1. Approximate solubi lity product of commonly chemi cally bath deposited compounds. Compound Solubility Product ( K sp ) Compound Solubility Product ( K sp ) AgCl 2 10-10 CuSe 2 10-40 AgBr 8 10-13 In(OH)3 6 10-34 AgI 1 10-16 PbS 1 10-28 Cd(OH)2 2 10-14 SnS 1 10-26 CdS 1 10-28 SnS2 6 10-57 CdSe 4 10-35 SnSe 5 10-34 CdTe 1 10-42 Zn(OH)2 1 10-16 Cu2S 1 10-48 ZnS 3 10-25 CuS 5 10-36 ZnSe 1 10-27

PAGE 53

53 + Electrical Energy Photovoltaic Device n-type p-type Light Energy Figure 2-1. A p-n junc tion photovoltaic device.

PAGE 54

54 Figure 2-2. Band diagram of the photoelect ric effect in a p-n junction photodevice.

PAGE 55

55 Figure 2-3. Solar spectrum at AM0 and AM1.5.

PAGE 56

56 Figure 2-4. Chronological progre ss of a-Si, CdTe, and CIS thin film technologies from 1975 to 2005.

PAGE 57

57 Figure 2-5. Ion-by-ion gr owth of CdS crystals.

PAGE 58

58 Figure 2-6. Growth of CdS crystals via the hydroxide cluster mechanism.

PAGE 59

59 Figure 2-7. Typical sput ter deposition chamber.

PAGE 60

60 CHAPTER 3 METHODS AND TECHNIQUES This chapter serves to introduce and describe the m ethods and techniques used in this research. The process of device fabrication begins with proper substrate cleaning along with the deposition of the absorber layer. The procedur e for deposition of the buf fer layer via chemical bath deposition is presented which is follo wed by sputter depositi on of the transparent conductive oxide layer. Completion of a photovoltaic device is achieved w ith the addition of top and bottom contacts. A number of thin film characterization techniques are also presented followed by the characterization of solar cell devices. Substrate Cleaning Procedure A strict cleaning procedure is adopted to ensure that all s ubstrates are thoroughly cleaned before their use in deposition or characterization. The first pr eparation step involves thoroughly washing each substrate with deionized water to remove any visible contaminants that may be present on the surface. Once each substrate has been individually rinsed they are placed in a Teflon substrate holder and transp orted to a hot deionized water ba th held at a temperature of 80 C. The substrates remain in this first hot de ionized water bath for at least 20 minutes before being removed. Each substrate is individua lly removed from the holder and mechanically cleaned with an Alconox detergent solution. The substrates are t horoughly rinsed with deionized water to ensure complete removal of the deterg ent solution. Once all substrates have been mechanically scrubbed and rinsed they are placed back in the substrate holder and delivered to an ultrasonic cleaner for a period of at least 20 minutes. Upon completion of the ultrasonic cleaning, the samples are again placed in an 80 C deionized water bath for at least 20 minutes. The substrates are then removed and placed in a three-step cascade of deionized water and

PAGE 61

61 nitrogen for 30 minutes. Lastly the substrates are removed from the cascade wash and dried thoroughly with nitrogen gas, and are stored in a clean and dry environment until use. Once a substrate has been properly cleaned it is ready for use in film deposition, characterization, or any other processing step. All cleaned substrates are placed in vacuumsealed bags in order to prevent contamination. Samples are removed from the sealed bags as needed for each processing step, thoroughly clean ed with nitrogen gas, and vacuum sealed immediately afterwards. Sample s are sealed in 3.5 mm thick 8" by 10" mylar bags with a Mighty Mutt MPV-18 industrial vacuum sealer. Absorber Layer Fabrication Absorber layers are often deposited by both vacuum and non-vacuum processes. Nonvacuum processes generally pr oduce photovoltaic devices with lo wer efficiencies, however the cost of production is quite typi cally far less than the vacuum process alternative. Typically, vacuum processes, such as co -evaporation or sputtering, are employed in order to achieve optimal device performance. Currently, cham pion CIGS devices are fabricated using an absorber layer deposited through a co-evaporation process [45]. A high quality and well-controlled absorber laye r is essential to the fabrication of high efficiency solar cells. At the University of Florida, CIGS layers are deposited using a PMEE, or plasma migration enhanced epitaxy, system [46] The PMEE system is a variant of molecular beam epitaxy (MBE), which is a more traditional approach. With PMEE, the substrate is placed on a rotating platen and is sequentially passed under each source. Metal components such as copper, indium, and gallium are sequentially depos ited during the exposure to each source. After the metal is deposited, the substrate then rotate d into the selenium zone where it encounters a surplus of selenium vapor. Temperatures can re ach up to 490C and at these high temperatures the metals components and selenium vapor react and polycrystallin e CIGS is formed. A typical

PAGE 62

62 CIGS photovoltaic device fabricated at UF is pictur ed in Figure 3-1. During each rotation cycle the substrate is only exposed to each source once. Plasma migration enhanced epitaxy has a low throughput in compared to other PVD processes such as co-evaporation. It has been fairly well established that this process is not economically viable for la rge-scale photovoltaic production, however it is a useful techniques for exploring th e single crystal regime of CIGS-based devices. Buffer Layer Fabrication Deposition of a buffer layer is the next step in photovoltaic device fabrication. Deposition of this thin film is most typically achieved vi a a wet chemical process known as chemical bath deposition, or CBD. Chemical ba th deposition is a technique of controlled precipitation of a compound from an aqueous solution and is extrem ely popular as it does not require a vacuum nor high temperatures, thus making it an extrem ely economical option for deposition of thin films. Zinc cadmium sulfide, ZnCdS, is a typical buffer layer deposited at the University of Florida in the fabrication of photovoltaic device s. The incorporation of zinc allows for a reduction in the amount of the toxic cadmium component found in the commonly deposited CdS buffer layer. The recipe used for deposition of ZnCdS at UF is based off a patent by the Boeing company [47]. Chemical Bath Deposition Apparatus Buffer layers, namely zinc cadmium sulfide, are deposited via the chemical bath deposition apparatus shown in Figure 3-2. This photo shows both the hot water circulator and the chemical bath reaction vessel, which is pictured close-up in Figure 3-3. The CBD apparatus is capable of depositing up to four substrates simultaneously (each 2" x 2") at temperatures up to 90C using deionized water as the hot bath liquid.

PAGE 63

63 A schematic showing the details of the chemical bath deposition apparatus is portrayed in Figure 3-4. A Julabo F32-HC hot wa ter circulator filled with water serves as the heating agent. The hot water bath is connected to a 1 L Chemgl ass jacketed beaker that serves as the CBD reaction vessel. The jacketed beaker is placed on a Fisher Scientific hotplate with magnetic stirring capabilities in conjunction with a magnetic mixer placed in the reaction vessel. The heating agent, in this case water, is circulated into the bottom inlet of the jacketed beaker and out of the top. A mercury thermometer is used to measure the temperature of the chemical bath inside the reaction vessel. Manual adjustments are then made to the Julabo hot water circulator control unit set point until the thermometer rev eals that the bath has attained its target temperature. A Julabo Pt100 thermocouple sensor is capable of providing communication to the Julabo F32-HC control unit and can be used in place of the mercury thermometer. In this closedloop setup, the control unit is able to automatically adjust its se t point to provide and maintain the desired bath temperature. Entegris D110215 Teflon single-substrate dippers are secured to the reaction vessel cover to maintain a vertical position inside the react ion vessel during film growth. Transparent Conductive Oxide Fabrication Following deposition of the buffer layer, a tran sparent conductive oxide layer is deposited, most often via sputtering. An intrinsic zinc oxide layer is first deposited followed by an aluminum-doped zinc oxide (AZO) layer. There are two characteristics of TCOs that have a significant impact on the performa nce of CIGS-based photovoltaic devices, optical transmission and electrical resistance. Both of these characteristics are impacted by TCO thickness in conflicting ways. Transmission determines the quant ity of photons that are able to enter into the device and resistance determines the ability of generated current to flow through the device. Therefore, the thickness of the TCO layer must be adjusted in order to maintain good optical

PAGE 64

64 transmission while achieving low sheet resistance. Decreasing the thickness of the TCO layer would help to improve transmission but would also increase resistance and result in substandard device performance. Sputter Deposition System Transparent conductive oxides are deposited vi a sputter deposition, namely the PerkinElmer Model 4400 Production Sputtering System photographed in Figure 3-5. The Model 4400 System operates at room temperature and is cap able of depositing a wide variety of materials onto substrates such as ceramics, glass, plas tics, metals, and semiconductors, in thickness ranging from a few angstroms to hundreds of micr ons. Sputter etching can also be performed thereby reversing the process in order to remove material from the surface of the substrate. The Perkin-Elmer Model 4400 Production Sputtering System consists of a RF impedance matching network, system control unit, target and substrate support asse mblies inside a large vacuum chamber, and a load lock with a pall et transfer mechanism. A schematic of the sputtering system can be seen in Figure 3-6. The top of the system houses the impedance matching network. This network serves to ma tch the impedance of th e cathode/discharge/anode system to the 50 impedance of the incoming RF power line. In this system a Randex 2 kW RF generator serves as the RF power source, at a frequency of 13.56 MHz. Without this impedance matching network the impedance mismatch can cause standing waves and heating of cables and certain power generator component s as a result of reflected wave s traveling back into coaxial cable connected to the power generator. Load and tune controls se rve to match real and imaginary parts of the impedance which allows proper matching even under wide variations in deposition parameters. Three sputter modes are available depending on where the RF power is applied. Sputter deposition can be achieved by grounding the substr ate pallet and applying all the RF power to

PAGE 65

65 the target, meaning only the target is bombarded by ions and subseque ntly sputtered. In the bias sputter mode RF power is again applied to the target, but a small amount of RF power is also applied to the substrate pallet. This results in substrates also being bombarded by ions, but to a far lesser degree than th e target. Compared to RF sputte ring with grounded substrates, bias sputtering typically yields higher quality deposit ions. Lastly, applying all RF power to the substrate pallet results in spu tter etching, where substrate material is etched away by ion bombardment. A target selector is used to connect the output of the matching circuitry to any of the three targets installed in the system. Brazing, soldering, or epoxy bonding can be used to bond target materials onto the cathode assembly. Eight-inch molybdenum, zinc, and aluminum oxide doped zinc oxide (2wt%) targets are currently installe d in the system. Dark space shields are required for targets less than 8 inches in diameter in orde r to limit discharge to the target material only, protecting the backing plate from ion bombardment. Target to substrate distance has a large influence on the uniformity of deposited film th ickness, and can be adjusted between 5 cm and 9 cm. Water is used for cooling of the targets and substrate pallet. The substrate pallet has a diameter of 23. 1", and can accommodate up to 30 three inch diameter substrates. In fixed m ode the substrate pallet is held st ationary and sputter deposition is limited to the exposed area of the substrate pallet. The rotary mode allows for the substrate pallet to be rotated according to user specified dir ection, angular velocity, and oscillatory motion. In this case a uniformity shaper can produce film uniformity within 5% resulting in highly identical results from deposition to deposition, as suming all other parameters are held constant. Transfer of substrates into the main chambe r is accomplished via a load lock that allows the sputtering chamber to remain isolated from the atmosphere during loading and unloading of

PAGE 66

66 samples. Use of a load lock results in shorter pump-down times and eliminates the need for repeated cleaning of the targets after introductio n of new samples, both of which lead to higher production rates. A transport mechanism carries th e substrate pallet from the load lock to the anode, or substrate table, inside the sputtering chamber. An elevator mechanism is then responsible for raising the substrat e table to its correct target to substance distance. The system features automatic or manual control of all load lock functions. The Perkin-Elmer Model 4400 Sputtering Syst em includes two pumps. A Leybold Trivac rotary vane pump (60 ft3/min) is used for chamber roughing, and the high vacuum pump is a CTI CryoTorr 8 cryopump, which is attached to the system via a throttle valve. The venetian blind style throttle is open during regul ar pumping and has a conductance greater than 1100 L/s. The valve is closed during sputtering reducing pumping speed thereby allowing the sputtering gas to fill the chamber. The throttle valve can be adjust ed to provide the desired level of pumping. Mass flow controllers (0-100 sccm) are used to regulate the flow of sputtering gas into the system, namely Ar, O2, and an Ar /H2 mix. Normal working pre ssures are in the range of 10 mTorr to 150 mTorr, which are easily altered by adjustment of the throttle valve and gas flow into the chamber. Accurate system pressure is measured by co mbining several different types of pressure gauges. Three thermocouple gauges are em ployed, effective between 10 Torr to 10-3 Torr, which operate via a thermocouple that measures the temp erature of a heated wire that is cooled by surrounding gas. Two convectron gauges, a type of thermocouple gauge, are attached to the system, which are effective in measuring pressures between 10-3 Torr and 1000 Torr. A conversion chart is required when measuring a mixed species gas environment, as convectron gauges are sensitive to the type of gas being m easured. To combat this problem a MKS Baratron

PAGE 67

67 capacitance manometer is used which measures the change in capacitance of a diaphragm as a result of flexure of the diaphragm caused by a ch ange in pressure. Capacitance manometers are direct measurement gauges and therefore can accurately measure mixed species gas environments in the range of 10-5 Torr to 1 Torr, depending on the thickness of the diaphragm. In order to maintain low substrate temperatures an RF magnetron is used to take advantage of magnetically confined plasma sputtering. The magnetic field induced by the RF magnetron deflects secondary electrons produced in the sputtering pr ocess away from the substrates. This magnetic field is superimposed on the plasma, how ever is sufficiently low enough to only deflect secondary electrons but not strong enough to substantially effect th e plasma, thereby resulting in low substrate temperature operation. Water is used to cool the RF magnetron assembly. Photovoltaic Device Fabrication Photovoltaic device fabrication requires deposition of additional components such as the back contact, top contacts, and anti-reflective co atings. The back contac t usually consists of molybdenum and is deposited on the substrate before the absorber layer. The anti-reflective coating applied is usually MgF2 and is deposited on top of the TCO, which is then followed by the evaporation of the top contact, us ually a mix of nickel and aluminum. Back Contact A thin layer of molybdenum is deposited, most often by sputtering, on a bare substrate to serve as the back contact for a photovoltaic de vice. Molybdenum is used as it forms a good ohmic contact with CIGS and at the same time re sists reaction with the CIGS film. The Mo back contact also remains stable in the highly corrosive selenium va por to which it is exposed during absorber growth. For devices fabricated at UF molybdenum-coated soda lime glass (SLG) is provided by Shell Solar Industries (SSI).

PAGE 68

68 Metallization and Anti-Reflective Coating A dual layer contact grid of nickel and aluminum is deposited onto the transparent conductive oxide, or AZO in the ca se of devices fabricated at UF, layer to complete device fabrication. Nickel serves as a diffusion barrier preventing aluminum oxide from entering the zinc oxide layer. The grid covers only a very small portion of the device, about 5%, to allow a sufficient amount of light to pa ss through the contact la yer and into the device. While no antireflective coating is used in the work at the Univ ersity of Florida, in an ideal situation, a MgF2 anti-reflective coating could be evaporated onto the device to reduce li ght reflection and permit maximum light entrance into the device. Characterization Techniques To fully characterize the thin films under anal ysis, it becomes necessary to employ a range of modern measurement techniques to probe topics such as surface structure, surface morphology, surface region composition, crystal stru cture, and film thickness, among others. The following subsections present brief descriptions of techniques that are of particular relevance to this research. Inductively Coupled Plasma Inductively Coupled Plasma (ICP) is an analytic al technique used for the detection of trace metals in environmental samples. The primary goal of ICP is to cause elements to emit light of a characteristic wavelength which can then be measured. The technology for the ICP method was first employed in the early 1960's with the intention of improving crystal growing techniques. ICP hardware is designed to generate a plas ma, a gas in which atoms are present in an ionized state. The basic set up of an ICP consists of three concentric tubes, most often made of silica. These tubes, denoted outer loop, interm ediate loop, and inner l oop, collectively make up the torch of the ICP. The torch is situated with in a water-cooled coil of a radio frequency (RF)

PAGE 69

69 generator. As flowing gases are introduced into the torch, the RF field is activated and the gas in the coil region is made electrically conductive. This sequence of events forms the plasma. In order to prevent possible short-circuiting as well as meltdown, the plasma must be insulated from the rest of the instrument. Insu lation is achieved by the concurrent flow of gases through the system. Three gases flow through the system--the outer gas, intermediate gas, and inner or carrier gas. The outer gas is typically Argon or Nitrogen. The outer gas serves several purposes, including maintaining the plasma, stabili zing the position of the plasma, and thermally isolating the plasma from the out er tube. Argon is commonly used for both the intermediate gas and inner or carrier gas. The purpose of the carrier gas is to convey the sample to the plasma. The ICP analysis technique requires that the elements that are to be analyzed be in solution. An aqueous solution is preferred ove r an organic solution, as the former requires special manipulation prior to injection into the ICP. Solid samples are also discouraged, as clogging of the instrumentation can occur. A nebu lizer transforms the aqueous solution into an aerosol. The light emitted by the atoms of an element in the ICP must be converted to an electrical signal that can be measured. This is accomplished by resolving the light into its component radiation, by means of a diffraction gr ating, and then measuring the light intensity with a photomultiplier tube at the specific wa velength for each element line. The light emitted by the atoms or ions in the ICP is converted to electrical signals by the photomultiplier in the spectrometer. The intensity of the electron signal is compared to previous measured intensities of known concentration of the element, and a con centration is computed. Each element typically has several specific wavelengths in the spectrum that could be used for analysis. Thus, the selection of the best line for the analytical a pplication in hand requires considerable experience with ICP wavelengths.

PAGE 70

70 The advantages of using the ICP analysis incl ude its ability to identify and quantify all elements, with the exception of Argon. Since many wavelengths of varied sensitivity are available for any one element, ICP is suitable fo r all concentrations from ultra-trace levels to major components. The detection limits are generally low for most elements with a typical range of 1 100 g/L. Probably the largest advantage of employing ICP for quantitative analysis is that multi-elemental analysis can be successfully accomplished, and the process is particularly rapid. A complete multi-element analysis can be comp leted in a period as short as 30 seconds, consuming only 0.5 mL of sample solution. A lthough in theory all elements except Argon can be determined via ICP, certain unstable elements require spec ial facilities for handling the radioactive fume of the result ing plasma. Also, ICP has diffi culty handling halogens where special optics for the transmissi on of the very short wavelengths of these elements become necessary. Scanning Electron Microscopy Scanning electron microscopy (SEM) is possi bly the most widely used thin film characterization instrument as a result of its reas onable cost and the wide range of information it is capable of providing. A sca nning electron microscope allows one to see the surface of a sample visually by providing a highly magnified im age. The method used to create these images can greatly simplify the image, and care should be taken to consider th e interpretation of SEM images. Magnification of sophisticated e quipment can range from 10x to 300,000x, with resolution on the order of a few nanometers. Most SEMs are capable of accelerating voltages ranging from 1 keV to 50 keV, howev er most are operated around 25 keV. Figure 3-7 portrays a schematic describi ng the operation of a scanning electron microscope [48]. A load-lock system is most often used to introduce the sample into the SEM, which is operated under high vacuum conditions. Electrons are emitted from a source, typically

PAGE 71

71 a tungsten or LaB6 filament, and focused by deflection coils in to a fine probe that is rastered over the surface of the sample. A number of interactio ns take places as these electrons are scanned over and penetrate the surface of the sample, many of which result in the emission of photons or electrons from, or through, the surface. Detectors are used to collect the emissions, and the output is fed to a display device. The x and y inputs of the display device are rastered in sync with the rastering of th e electron beam on the sample. The em ission of each point on the sample is directly mapped to the corresponding point on the display device. Repeating this for all measured points on the sample results in an image being produced on the display device. Magnification is achieved by reduci ng the area of the rast er pattern on the sample while keeping the display mapped to a fixed level. Secondary electron images, backs cattered electron images, and elemental x-ray maps are the three main imag es types that a SEM is capable of producing. Secondary electron imaging Secondary electron imaging is the most comm on imaging mode, and relies on detection of secondary electrons emitted from the sample. These are very low energy electrons, which originate from a subsurface depth no larger than seve ral angstroms. A detector is used to capture these electrons and produce an output signal. This output signal serves to modulate the intensity of an external display device, which is rastered in sync with the raster ing of the primary beam impinging upon the specimen. Images can be magni fied by increasing the ratio of scan lengths on the display device to that on the specimen. Sophisticated SEM systems are capable of imaging samples with a resolution on the order of a few nanometers. Backscattered electron imaging Backscattered electron imaging relies on detection of backscattered electrons emitted from the sample. These are high energy electrons, which possess nearly the same energy as impinging electrons, that are elastically scat tered by the sample. The atomic number of the sample material

PAGE 72

72 determines the level of backscattering, as back scattering of electrons in creases with increasing atomic number. Although specific elemental identif ication is not possible, regions of the sample that exhibit wide differences in atomic number can be easily contrasted. Energy dispersive spectroscopy X-rays can often be emitted as a result of inelastic collisions of primary beam electrons with the specimen. Continuum and characteristic x-rays are the two types of x-rays that can be emitted during such a collision. Characteristic x-rays are of greatest importance, however continuum x-rays, analogous to noise, can play a large role in rende ring energy dispersive spectroscopy (EDS) spectrum analysis difficult. Ch aracteristic x-rays have unique energies that are related to electronic transitions in every atom Analysis of these energies is achieved with the use of an energy dispersive spectrometer, which is a detector that is capable of discriminating between x-ray energies. This discrimination allows atoms in the specimen to be identified and, via ZAF (atomic number, absorption, and fluorescence) correction factors, qualified quantitatively. X-Ray Diffraction X-ray Diffraction (XRD) is a popular method that is capable of identifying and measuring the structural properties of crys talline phases that are present in a given sample. Grain size, phase composition, epitaxy, strain preferred orientation, and defect structure of crystalline phases can all be accurately identified. Thickness determination of thin films and multi-layers can also be achieved with XRD. Thin film analys is requires very little sample preparation, does not contact the sample, and is non-destructive, wh ich makes it an ideal in situ characterization technique. Most all materials can be successful ly analyzed, however the measurement sensitivity depends on the material in question. Elemen ts with a high atomic number provide strong diffraction intensities and therefor e provide increased XRD sensitiv ity. Sophisticated equipment

PAGE 73

73 can easily measure thin films less than 50 and can measure film structure as a function of depth. X-Ray Diffraction (XRD) is becoming an extremely popular method for determining crystalline phases present in solid samples as a result of the techni ques versatile and nondestructive nature, as well as the extremely simple sample pr eparation requirements. Unknown sample identification is achieved by comparing xray diffraction patterns against an international database of known phases. In XRD, a collimated beam of x-rays, with wavelength = 0.5-2 is incident on a specimen and is diffracted by the crystalline phases in the specimen according to Braggs law 2 d sin (3-1) where d is the spacing between atomic planes in the crystalline phase. The intensity of the diffracted x-rays is measured as a function of the diffraction angle 2 and the specimens orientation. This diffr action pattern is used to identify the specimens crystalline phase and to measure its structural properties, including stra in (which is measured with great accuracy), epitaxy, and the size and orientation of crystallit es (small crystalline regions). XRD can also determine concentration profiles, film thickne sses, defects, and atomic arrangements in amorphous materials and multilayers. To collect th is structural and physi cal information from thin films, XRD instruments and techniques are designed to maximize the diffracted x-ray intensities, since the diffracting power of thin films is small. The full width at half maximum values, easil y obtained from high resolution XRD scans, can be used to calculate grain size information. The average grain size L can be estimated using the Scherrer formula [49], L K cos (3-2)

PAGE 74

74 where is the full width at half maximum, is the x-ray wavelength, and is the diffraction angle. The shape factor K is a constant and its value is often taken as 1.0 [50]. Profilometry Profilers are widely used instruments and routin ely used as an altern ative to in-situ film thickness measurements. Mechanic al profilers operate by recordi ng height displacements as a stylus runs over a surface. By recording th e vertical displacement as a function of sample positioning, a profile of the surface can be cr eated. The biggest drawback of surface profilometry is the need for physical interaction with the sample surface, since the stylus must come in contact with the surface in order to perf orm the measurement. Non-contact alternatives focus around reflectometric and time-of-flight meas urements of light passing through the film. In order to perform thickness measurements on thin films with a surf ace profiler, a portion of bare substrate is required. Two methods ex ist to remove a portion of the film. The first method is a type of lift-off etching where a portion of the substrate is masked before film deposition. After deposition this mask can be removed either chemically or mechanically, resulting in lift-off of the film and exposure of the substrat e. The second method involves using a mask to protect a portion of the deposited film rather than remove it. Once the mask is applied to the film, the substrate is placed in an et ching solution, resulting in any part of the film not covered by the mask being etched away. A different chemical solution can then be used to remove the mask material from the remaining film. Care must be taken with this method as excessive etching times can resu lt in horizontal etching of the protected film under the mask, which can alter surface profiler measurements. For this work, a Veeco Dektak 150 Surface Profiler is used for measuring thin film thickness in this work. The Dektak 150 is cap able of accommodating samples up to four inches

PAGE 75

75 thick, and can perform long scan of up to 55 mm. The profiler has a vertical measurement range of 512 m and can perform precise step height measurements of thin films down to 100 with a repeatability of 6 The system also features an X-Y auto stage that can be programmed to provide automated sample mapping. A low for ce stylus (sensitivity of 0.03 mg) enables nondestructive characteriza tion of delicate surfaces. Ellipsometry The technique of ellipsometry, also known as polarimetry and polarization, is over a century old and is used to obtain the thickness and optical constants of films. The method consists of measuring and interpreting the chan ge of polarization state that occurs when a polarized light beam is reflected at non-normal incidence from a film surface. Figure 3-8 shows a schematic depicting a common experimental arra ngement for ellipsometry measurements [51]. The light source is first made monochromatic, co llimated, and then linearly polarized. Upon passing through the compensator, the light is ci rcularly polarized and then impinges on the specimen surface. After reflection, the light is transmitted through a second polarizer that serves as the analyzer. Finally, the light intensity is measured quantitatively by a photomultiplier detector. The polarizer and analyzer are rota ted until light extinction occurs. The extinction readings enable the phase difference e and amplitude ratio tan of the two components of reflected light to be determined. From these, either the film thickness or the index of refraction can be obtained [52]. Ultraviolet-Visible Spectroscopy Ultraviolet-visible spectroscopy is most of ten used for measuring the transmittance and absorbance of liquid and solid samples. A si mple schematic of a UV/vis spectrophotometer can be seen in Figure 3-9. A spectrophotometer m easures transmission with a UV/vis source, most

PAGE 76

76 often tungsten and deuterium, which produces ligh t that is then selectively separated by a monochromator. A detector colle cts the diffracted light that passes through the sample and the signal is passed on to software which then co mpares the intensity of the transmitted light, I to the intensity of the incident light, Io. The transmittance of the sample is given by T I Io (3-3) where the transmittance is expressed as a percen tage. Once the transmittance of the sample is known, the absorbance, A can be calculated as A log10IoI (3-4) The relationship of the absorbance coefficient, to the intensity of the transmitted light I can be derived according to the Beer-Lambert equation, and is given by I Ioe t (3-5) where t is the film thickness of th e sample. When the thickne ss of the film is known, the transmittance can be used to calculate the absorption coefficient as ln 1 T t (3-6) Once the absorption coefficient is known, the opti cal band gap of the film can be calculated by the Tauc equation, hn Bh Eg (3-7) where B is a constant, and n = 0.5 for an indirect transition or n = 2 for a direct transition semiconductor [53, 54]. The optical band gap en ergy can then be obtained from a plot of ( h )n

PAGE 77

77 vs. ( h ), where a straight line is observed over the absorption onset region. This fitted line can be extrapolated to ( h )n = 0 to obtain the optical band gap energy, Eg. A Perkin-Elmer Lambda 800 UV/Vis spectropho tometer is used to collect optical transmission, absorbance, and band gap data. The system uses twin beams, tungsten and deuterium, to collect transmittance spectra fr om 200 nm to 900 nm. The preparation of a reference sample is suggested to provide ade quate background correcti on of measured spectra. X-Ray Photoelectron Spectroscopy X-ray photoelectron spectroscopy (XPS) is a surf ace sensitive technique that is capable of detecting binding energies of atoms in different ch emical states. XPS is one of the most broadly used surface analysis techniques as a result of its surface sensitivity and its quantitative and chemical state analysis capabili ties. Detection of all elements is possible with a sensitivity variation across the periodic table of approxi mately 30, except for hydrogen and helium. Gaseous, liquid, and solid samples can be meas ured, although most systems are specifically tailored for solid samples. XPS is most capab le of measuring the surface of samples (top two atomic layers) but can also measure from 15 to 20 layers deep, with a resolution of up to 70 m. It is generally considered a non-destructive technique and a variety of samples can be accommodated, including biological, or ganic, and polymeric materials. XPS is based on the Einstein photoelectric law. High energy photons have the ability to ionize atoms and eject free electrons. The kine tic energy of the electron depends on the energy of the photon h which can be expressed by th e Einstein photoelectric law KE hv B E (3-8) where KE is the kinetic energy of the photoelectron and BE is the binding energy of the electron to the atom. The kinetic ener gy can therefore be measured to determine the binding energy,

PAGE 78

78 since the energy of the photon ( h ) is known. In a neutral atom, the number of electrons equals the number of protons. These electrons are arra nged in pairs of two in orbitals around the nucleus, bound by electrostatic attrac tion. Each orbital has a discre te energy level, which varies for all elements since the electrostatic attraction will vary with differing nuclei. The binding energy of an electron can therefore be determined by the amount of energy required to remove it from the atom. Experimental determination of the binding energy approximately relates to a discrete energy value, which is specific to a part icular atom, thereby identifying that atom [48]. X-ray photoelectron spectroscopy, Auger elect ron spectroscopy, and secondary ion mass spectrometry are all very widely used surface anal ysis techniques. XPS is similar to AES in terms of elemental analysis capability and absolute sensitivity. The main advantages of XPS are its more developed chemical state analysis capability, somewhat more accurate elemental analysis, and far fewer problems with induced sample damage and charging effects for insulators [48]. Secondary Ion Mass Spectrometry Secondary ion mass spectrometry (SIMS) is an analytical technique that is capable of trace-level contaminant detecti on and can provide quantitative measurements of major and minor components. The technique is destructive, as mate rial must be removed from the sample in order to perform the measurement. Material is sputtered from the sample in the form of neutral and ionized atoms and molecules by a focused ion beam. This sputtered material is accelerated into a mass spectrometer, and is then separated accordi ng to mass-to-charge rati os. Depth profiling is the most common application of SIMS, where elem ental impurity is measured as a function of depth. Quantitative measurements are pos sible but require the use of standards.

PAGE 79

79 SIMS is an extremely sensitive analytical te chnique, capable of detection limits in the parts-per-million to parts-per-billion range. De pth resolution is generally around 2 nm with a lateral resolution be tween 50 nm and 2 m. SIMS can accurately measure any elemental impurity, even elemental isotopes, with a detection limit between 1012 and 1016 atom/cm3 for most impurities. Secondary ion mass spectrometry is one of th e dominant surface analysis techniques along with x-ray photoelectron spect roscopy and Auger electron spec troscopy. Although SIMS has a lower spatial resolution and speed compared with the other two techniques, neither can match the superior trace analysis capabiliti es of SIMS. SIMS also has th e ability to detect hydrogen and helium, which neither AES nor XPS can detect. Auger Electron Spectroscopy Auger electron spectrosc opy (AES) is a versatile techniqu e that is used to identify elemental composition and chemical bonding of atoms in the surface region of thin films. It is often combined with ion beam sputtering to provide composition and chemistry analysis as a function of sample depth. AES measures the en ergy distribution of secondary electrons, notably the Auger electron component, that are released as a result of an electron beam probing the sample surface. J.J. Lander was the first to report the detection of Auger electrons in the secondary electron energy spectra [55]. The technique is surface sensitive and exhi bits a sampling depth between 10 and 100 depending on the energy of the A uger electrons measured. Generall y, lateral spatial resolution is as low as 300 and depth resolution as low as 20 depending on the electron and ion guns used, respectively. Detectability under good conditions is approximately 100 ppm for most

PAGE 80

80 elements. Unfortunately, the electron beam probe can be destructive to some samples in addition to the ion beam damage that takes pl ace during depth profiling measurements. The basic Auger process involve s the production of an atomic inner shell vacancy, most often by electron bombardment, and the decay of the atom from this excited state by an electronic rearrangement and emission of an en ergetic electron rather than by emission of electromagnetic radiation. Th ere are four basic contributions responsible for producing Auger electrons found in the energy distribution resulting from the bombardment of a surface with an electron beam. These contributi ons are the creation of inner sh ell vacancies in atoms of the sample, the emission of electrons as a result of Auger processes resulting from these inner shell vacancies, the transport of those electrons out of the sample, and the detection and measurement of the energy distribution of the electrons coming from the sample. Auger electrons are generated in transitions back to the ground stat e of atoms with inner shell vacancies, no matter what process produces the inner shell vacancy. As a result, Auger peaks are observed in energy spectra generated by electron excitation, x-ray excita tion, ion exchange, and certain nuclear reactions [48]. There are three main modes of AES analysis. Po int analysis mode is the simplest and most often used mode, in which the primary electron b eam is positioned in a particular location and an Auger survey spectrum is collected. Another common mode is the dept h profiling mode, which is identical to point analysis mode with the addi tion of an ion beam directed at the sample. The ion beam is used to sputter material off the surface of the sample so that the analysis can measure composition variation as a function of depth. In depth profiling mode, Auger data can be acquired either as a survey spectrum or can be gathered in a narrow scan window in order for detection of a specific element. The final m ode of AES analysis, mapping mode, operates with

PAGE 81

81 the goal of specific element detection. In mappi ng mode, the Auger peak of a specific element is monitored while the electron beam is scanned ove r the sample area. This results in a twodimensional horizontal spatial distribution of a specific element, compared to the vertical distribution produced by the dept h profiling mode. A three-dimens ional spatial distribution of any element can be created by combining the mapping and depth profiling modes. Auger electron spectroscopy, x-ray photoelectron spectroscop y, and secondary ion mass spectrometry are a very comm on set of techniques used in thin film analysis and characterization. They are capable of produci ng similar results, however each has its own strengths and weaknesses. The main advantage of AES its superior spatia l and depth resolution, along with absolute sensitivity to many important elements. Four-Point Probe The four-point probe technique was originally developed to measure the earths resistivity [56], and is currently one of the most commonly used methods for measuing the resistivity of semiconductor materials [57]. A schematic of th e four-point probe configuration is seen in Figure 3-10 and shows four finite metal tips with uniform spacing s. The outer two metal tips serve to introduce current I into the sample with the resulting electric potential V being measured by the inner two tips, or electrodes. The use of a four-probe set up eliminates the appearance of contact resistance in the resistivity measurement since separate electrodes are used for current supply and voltage detection compared to a two-pr obe system [58, 59]. As a result, the fourpoint probe technique is much preferred for its ease of data interpretation [60]. The resistivity of bulk and thin film sample s can be measured via the four-point probe technique, however a different expression is used in each case. In the case of a bulk sample, where the sample thickness t is generally much larger than the probe spacing s, the expression for bulk resistivity is given as

PAGE 82

82 2s V I (3-9) assuming equal probe spacing. In the cas e of a thin film, the film thickness t is usually much smaller than the probe spacing s, in which case the sheet resistivity is given as Rs k V I 4.53 V I (3-10) where k = 4.53 is the geometric factor in the case of a semi-infinite thin film [61]. Given the film thickness t the resistivity can be calculated by Rst 4.53 V I t. (3-11) The Alessi four-point probe system features 0.05 spaced tungsten-carbide probe tips, which are 0.002 in diameter. Currents of 1 A to 100 mA are supplied by a Crytronics Model 120 current source. The voltage resulting across the inner probes is measured by a Keithley Model 181 Nanovoltmeter which is cap able of 10 nV sensitivity. Photovoltaic Device Characterization Techniques Current-voltage (I-V) characte rization is an important part of analyzing photovoltaic device performance. A solar cells I-V characteristics are directly related to a devices maximum power conversion efficiency. The standard Air Mass 1.5 Global Spectrum (AM1.5G), which is normalized at 1000 W/m2, is used for I-V measurement of devices. In addition, quantum efficiency (QE) measurements show a devices ab ility to adsorb radiation at specific incoming wavelengths. Current-Voltage Current-voltage measurements are used to predict the performance of a photovoltaic device. An I-V curve can be generated by measuring the current produced by varying load

PAGE 83

83 resistance upon illumination of a solar cell device, most often under AM1.5G conditions. A typical I-V curve is shown in Figure 3-11, where volta ge is plotted on the horizontal axis and current on the vertical axis. The ma ximum voltage, or open-circuit voltage Voc, exists when the current across the cell is zero, under infinite load resistan ce. The maximum current, or shortcircuit current Isc, exists when the voltage across the cell is zero, under zero load resistance. Photovoltaic devices cannot be ope rated at either the open-circu it voltage nor the short-circuit current, as no power would be produced since no current or voltage exist at the two points, respectively. The maximum power Pm the cell is capable of generating is given by P m I mVm (3-12) where Im and Vm are the values that maximize the maxi mum power as portrayed in Figure 3-x. The degree to which the maximum power fills the I-V curve is given by the fill factor parameter which describes the degree to whic h the current and voltage at maximum power ( Im and Vm) match Isc and Voc, respectively. The fill factor is given as FF P mIscVoc I mVmIscVoc (3-13) and is generally given as a percentage. The conversion efficiency of a photovoltaic device describes its ability to co nvert incident photon energy, Pin, into electrical energy, and is calculated as P mPin I mVmPin I scVocFF Pin (3-14) where is the percentage of incident light converted to electricity. The conversion efficiency of a solar cell is a ffected by a number of factors that hamper the conversion of incident photon ener gy to electricity. The band gap of a photovoltaic device refers to the minimum energy required to free electrons, and differs depe nding on the absorber material

PAGE 84

84 used. The reason that photovoltaic devices are not very efficien t is that they are unable to convert all wavelengths of incomi ng radiation. Photons with en ergies below the band gap do not possess the required energy and th erefore only generate heat in the device. Likewise, photon energy above the band gap is also wasted a nd ends up being emitted as heat or light. Another factor affecting solar cell conversion efficiency is recombination, where charge carriers recombine before contributing to the ce lls current. Combina tion of photon-generated electrons and holes is termed direct recombination whereas indirect recombination occurs as a result of impurities or defects in the crystal structure. Temperature can also have a negative affect on devices as conversion efficiency tends to decrease with increasing temperatures. Reflection of light away from a photovoltaic devices surface is another factor that negatively affected conversion efficiency. The two most common me thods for reducing the reflection of incident photons are the application of antireflective coatings and surface texturing. Single or multiple antireflective layers are deposit ed in order to minimize reflection at different wavelengths. Perhaps the most common antireflec tive coating applied to CIGS-based solar cells is MgF2 [62, 36, 63-66]. Adding texture to the surf ace of a device increases the probability that incident light will be reflected and strike the su rface again, thereby increasing the possibility of absorption. Current-Voltage Measurement The I-V measurement system employs a tungstenhalogen lamp to produce incident photon energy, which compares well with the energy spectr um of AM1.5G radiation. A reference cell is used in order to calibrate the intensity of the tungsten-halogen lamp, which is adjusted by altering the distance between the lamp and the device being characterized. This distance is modified so that the measured short-circuit current matches the known value of the reference cell used in the calibration. The I-V measurement system is calibrated using reference data from a

PAGE 85

85 CIGS solar cell obtained from the National Re newable Energy Laboratory (NREL). This CIGS device, S2117-A2, was fabricated on August 22, 2003, and contains seven cells, each with an area of 0.430 cm2. The reference data obtained from NREL for each cell of the S2117-A2 device is listed in Table 3-1. For example, under th e AM1.5G radiation spectrum used by NREL, cell 4 of the device exhibits a short-circuit current Jsc = -33.43 mA/cm2, an open-circuit voltage Voc = 0.653 V, a fill factor FF = 74.67 %, and a conversion efficiency = 16.291 %. The I-V measurement system is controlled via a personal computer and data is collected and analyzed using a custom LabVIEW software package [67]. As mentioned, increased devi ce temperature tends to have a negative effective on overall device performance and conversion efficiency. An increase in operating temperature typically results in a decrease in the open-circuit vol tage of CIGS devices. To ensure optimal performance, a cooling system is employe d to control device temperatures during I-V measurement. Cooling water ci rculates through the measurement assembly during measurement and a controller adjusts the flow rate to maintain an operating temperature of 25 C 1 C. Quantum Efficiency The quantum efficiency (QE) of a solar cell re lates the number of char ge carriers collected by the device to the number of photons incident u pon the device. The QE of a solar cell is given as a function of radiation wavelength, and theref ore relates the response of a device to various wavelengths of the incident spectrum. Quantum efficiency exhibits an upper limit of one for the case where all photons of a dist inct wavelength are absorbed and the corresponding minority carriers are collected. An incoming photon with energy less than the band gap will not be absorbed, resulting in a QE of zero. In the ideal case, quantum efficiency would be constant for all wavelengths of incoming radiation, resulting in a square shape. Recombination is considered

PAGE 86

86 the main cause of reduced QE. High energy photons are generally absorbed closer to the front surface of devices, where recombination negativ ely affects the blue portion of the QE spectrum. On the other hand, the green portio n of the QE spectrum is negatively affected by the recombination of carriers produced by the ab sorption of low energy photons in the bulk of devices. The quantum efficiency of a photovoltaic device can be calculated under two conditions. The first condition includes losses in incident ra diation that occur as a result of reflection and transmission through the device. This is generally referred to as the external QE, which takes into account all external losses of radiation. The second condition excludes these losses and focuses solely on light that has not been reflected or transmitted through the device, and is termed internal QE. External QE curves can ge nerally be converted to internal QE curves by correcting for the reflection a nd transmission of the device.

PAGE 87

87 Table 3-1. Reference da ta for NREL device S2117-A2. Cell Area (cm2) Voc (V) Jsc (mA/cm2) Isc (mA) FF (%) Eff (%) Vmp (V) Imp (mA) Pmax (mW) 1 0.430 0.645 -31.83 -13.687 76.66 15.737 0.538 -29.26 15.737 2 0.430 0.649 -32.43 -13.945 75.30 15.858 0.532 -29.79 15.858 3 0.430 0.653 -30.11 -12.947 74.92 14.719 0.534 -27.54 14.719 4 0.430 0.653 -33.43 -14.374 74.67 16.291 0.532 -30.61 16.291 5 0.430 0.655 -33.59 -14.443 74.85 16.458 0.534 -30.80 16.458 6 0.430 0.654 -31.45 -13.524 74.88 15.400 0.532 -28.93 15.400 7 0.430 0.644 -32.97 -14.177 74.27 15.775 0.524 -30.08 15.775

PAGE 88

88 ZnO:Al Ni/Al Ni/Al 0.05/3 m 0.5-1.5 m SLG Mo CuIn1-yGaySe2 ZnxCd1-xS 0.05 m 1.0-2.0 m 0.5-1.0 m i-ZnO0.05 m Figure 3-1. Structure of a typical CIGS/ZnCdS device fabricated at UF.

PAGE 89

89 Figure 3-2. Chemical ba th deposition apparatus.

PAGE 90

90 Figure 3-3. Chemical bath deposition reaction vessel.

PAGE 91

91 Figure 3-4. Chemical bath deposition apparatus and reaction vessel schematic. Hot Water Bath Controller Thermocouple Feedback Outlet Inlet Stirrer Sample Sample Holder Thermocouple Jacketed Beaker Chemical Bath Beaker Outlet Beaker Inlet

PAGE 92

92 Figure 3-5. Perkin-Elmer Model 4400 Production Sputtering System.

PAGE 93

93 Figure 3-6. Perkin-Elmer Model 4400 Productio n Sputtering System schematic drawing.

PAGE 94

94 Figure 3-7. Simplified operation of a scanning electron microscope.

PAGE 95

95 Figure 3-8. Common experimental arra ngement for ellipsometry measurements.

PAGE 96

96 UV/Vis Source Monochromator Sample Detector Software/ Readout Figure 3-9. Simplified UV/vi s spectrophotometer schematic.

PAGE 97

97 + V I s s s t sample Figure 3-10. Four-point pr obe configuration schematic.

PAGE 98

98 V I Isc Voc Vm Im Pm = ImVm Figure 3-11. Example of a typical I-V curve.

PAGE 99

99 CHAPTER 4 ZINC CADMIUM SULFIDE BUF FER LAYER GROWTH Introduction The chemical bath deposition technique for thin film growth is gaini ng attention as it does not require expensive nor sophis ticated equipment such as the vacuum systems required by other deposition processes. In fact, chemical bath deposition can be performed with a few beakers, a hot plate, and a magnetic stirrer, along with the proper starting chemicals, which are usually quite common and therefore genera lly inexpensive. These are important factors when it comes to cost reduction of photovoltaic devices; however there are ma ny other advantages propelling the renewed interest in chemical bath deposition of thin films. For exam ple, a large number of substrates can be processed simultaneously with chemical bath deposition. Also, there are few constraints regarding substrate requirements, a nd therefore any insoluble surface is most likely suitable for deposition. Low co st, substrate flexibility, and large throughput are some of the factors that make the chemical bath deposit ion technique such an attractive choice in photovoltaic device fabrication. Zinc cadmium sulfide (ZnxCd1-xS) thin films are good candidates for development as a wide-bandgap window layer in photovoltaic devices. Dependi ng on the amount of zinc, the bandgap of ZnCdS can be tailored from 2.4 eV (CdS) to 3.8 eV (ZnS). Many laboratory devices have reached high efficiencies of 16.5% [68], 19.5% [36], 15.0% [36], and 18.6% [62] using designs with CdS/CdTe, CdS/CIGS, CdS/CIS, and ZnS/CI GS structures, respectively, where the second material in each pair de notes the photovoltaic absorber la yer. Using CdS as the buffer layer in these devices limits the attainable devi ce efficiencies, as CdS buffer layers have a low bandgap energy, which causes considerable abso rption in the short-wa velength region. One method to improving the blue response is to reduce the CdS buffer layer thickness, thereby

PAGE 100

100 achieving increased short-circuit current ( JSC). However, reducing the CdS thickness can adversely impact device open-circuit voltage ( VOC) and fill factor ( FF) performance parameters as a result of the increased potential for pinhole formation in thinner films. Depositing a buffer layer material with a higher bandgap would help alleviate the problems encountered with CdS thin films. The addition of Zn to CdS is an attractive alternative as it yields a more favorable conduction band alignment and provid es a lattice parameter that more closely matches that of CIGS absorber material. Therefore, devel oping an alternative window layer with a higher bandgap based upon ZnCdS is a promising approach. A number of deposition techniques have been used in creating cadmium sulfide and zinc sulfide thin films, including vacuum chemical vapor deposition [69-71], spray pyrolysis [72], chemical bath deposition [64, 73-75, 38, 76-81] evaporation [82, 83], successive ionic layer adsorption and reaction (SILAR) [84-86], and the di p technique [87, 88]. In addition to its low cost, chemical bath deposition (CBD) has a dist inct advantage as it forms homogeneous films adequate for use in solar cell devices. Zinc cadmium sulfide film growth in the CBD technique is based upon the slow release of S2and the controlled release of Zn2+ and Cd2+ ions in the solution. Two common sulfiding agents norma lly used in this process are thiourea [(NH2)2CS] and thioacetamide [CH3CSNH2]. To date, chemical bath deposition has proven more successful for CdS film deposition than for ZnCdS films. The addition of Zn in the chemical bath introduces problems, including the common problem of oxide formation in the ch emical deposition bath The process is additionally complicated as the solubility products of CdS ( Ksp = 1 10-28) and ZnS ( Ksp = 3 10-25) differ by several orders of magnitude. Several complexing agents have been utilized to study the conditions fo r co-deposition of Cd and Zn su lfide, such as triethanolamine

PAGE 101

101 (TEA), NH3, and hydrazine (N2H4). The different stability constants for the dominant cadmium and zinc precursors further narrow down the proper chalcogenide ope rating window and make control of the deposition process particularly difficult [89]. In some cases a second complementary comple xing agent is preferred for enhanced Zn incorporation in the film [90]. Another approach is the depos ition of CdS/ZnS multilayers for subsequent processing [91]. There is much disagreement concerning the best approach to achieve an optimal deposition of ZnCdS thin film s, justifying the need for additional studies concerning ZnCdS deposition. In this study we hope to prov ide additional insight to the incorporation of zinc in ZnCdS thin f ilms deposited by chemical bath deposition. Chemical Bath Deposition of Zinc Cadmium Sulfide In this work, chemical bath deposition of zi nc cadmium sulfide thin films is carried out with the use of four source solutions, namely a cadmium s ource solution, a sulfur source solution, a zinc source solution, and an ammonia source solution. The fractional composition of zinc in the chemical bath is controlled by mixi ng dose solutions of all four source solutions in specific proportions. The amount of source solution required for buffer layer deposition depends on the desired fractional composition of zinc re lative to cadmium, defined by the fractional composition formula xp b f ZnZn Cd (4-1) where [ Zn ] and [Cd ] are the respective molar concentrations of zinc and cadmium. The subindex p b or f on the fractional composition, x, is used to specify the zinc composition in the prepared bath solution, the actual bath so lution, and in the deposited film. The relaxed restrictions on subs trates means that chemical bath deposition can be used to deposit thin films on a variety of substrates. Whereas device fabricat ion requires buffer layer

PAGE 102

102 deposition on Mo-coated soda-lime glass (SLG) substrates covered with some type of CIGS absorber, for development and characterizati on purposes zinc cadmium sulfide films are deposited on bare SLG substrates. The chemical bath deposition procedure invo lves a protocol for cleaning substrates, preparation of cadmium, sulfur, zinc, and ammonia dose solutions, mixing of dose solutions to carry out the layer growth, and termination of growth. Materials Four source solutions serve as the basis of th is chemical bath deposition technique, which finds its roots in a patent from Boeing. Fi rst, a cadmium source solution is prepared by combining 3.84 g cadmium chloride hydrate (A lfa Aesar Item # 10661) and 2.78 g ammonium chloride (Alfra Aesar Item # 10632) with 2000 mL deionized water. Second, a sulfur source solution is prepared by mixing 12.69 g thi ourea (Alfa Aesar Item # 36609) with 2000 mL deionized water. Third, a zinc source solution is created by th e addition of 10 g zinc chloride hydrate (Alfa Aesar Item # 35782) to 400 mL of deionized water. Ammonium chloride (Alfa Aesar Item # 10632) is then added until the white di ssipation dissolves, leaving a clear solution. Fourth, an ammonia source so lution is prepared by mixing 50 mL of 28%-38% ammonium hydroxide (Fisher Scientific Item # AC20584-0010) with 350 mL de ionized water. Each source solutions is thoroughly agitated for at least 60 minutes after prepara tion, and is also agitated for a minimum of 15 minutes prior to its use. Growth Procedure The growth of zinc cadmium sulfide thin films via chemical bath deposition follows a strict procedural sequence. A ll substrates and equipment must be thoroughly cleaned before deposition. The chemical bath solution is form ulated through the extrac tion of a dose volume of each source solution selected so th at the mixing of all four dose solutions results in a bath that

PAGE 103

103 has a target fractiona l composition of zinc, xp. The dose solutions are added to the reaction vessel in a step-wise fashion under stringent control of the chemi cal bath temperature. Once a desired growth duration is achieved the depos ition is terminated and the substrates are thoroughly cleaned of any visi bly loose powder by rinsi ng with deionized water. For deposition on glass substrat es, Corning 1737 is chosen for its superior characteristics over typical soda-lime glass, which boasts anneal and strain points almost 200 C higher. The lower thermal expansion coefficient of the Corn ing 1737 glass results in less warp and distortion in substrates exposed to thermal gr adients during substrate processing. Substrate and equipment cleaning Glass substrates must be thoroughly cleaned before their use in deposition. The first preparation step involves thoroughl y washing each substrate with deionized water to remove any visible contaminants that may be present on the surface. Once each substrate has been individually rinsed they are pla ced in a Teflon substrate holder a nd transported to a hot deionized water bath held at a temperature of 80 C. The substrates remain in this first hot deionized water bath for at least 20 minutes before being removed. Each substrate is individually removed from the holder and mechanically clean ed with an Alconox detergent solution. The substrates are thoroughly rinsed with dei onized water to ensure complete re moval of the dete rgent solution. Once all substrates have been mechanically sc rubbed and rinsed they are placed back in the substrate holder and delivered to an ultrasonic cleaner for a peri od of at least 20 minutes. Upon completion of the ultrasonic cleaning, the samples are again placed in an 80 C deionized water bath for at least 20 minutes. The substrates ar e then removed and placed in a three-step cascade of deionized water and nitrogen for 30 minutes. Lastly, the s ubstrates are removed from the

PAGE 104

104 cascade wash and dried thoroughly with nitrogen gas, and are stored in a clean and dry environment until use. Device fabrication requires less preparation be fore each deposition step, as the substrates are originally cleaned before CGS or CIGS absorb er deposition. To keep these samples clean, all device substrates are vacuum sealed after each deposition step. Samples are removed as needed for each processing step, and sealed again immediately af ter processing. Device samples are thoroughly cleaned with nitrogen ga s before CBD buffer layer deposition. Before buffer layer deposition, it is necessary to ensure that all equipment, including the jacketed reaction vessel, graduated cylinders, pipettes, beakers, and stirrers, are thoroughly and repeatedly washed in deionized water. Preparation of dose solutions The desired fractional composition of zinc in the CBD bath is obtained by mixing specific amounts of dose solutions. Table 4-1 lists the do se solution volumes that are required to prepare chemical baths with various fractional compositions of zinc, xp. For example, a chemical bath with fractional composition of zinc xp = 0.3 can be prepared us ing 106.7 ml of cadmium dose solution, 2.09 ml of zinc dose solution, 100 ml of sulfur dose solution, and 0.25 ml of ammonia dose solution. Table 4-2 takes in to account 490 mL of deionized water that is initially added to the reaction vessel, and lists fina l bath volumes and concentrations. Again, for the example of xp = 0.3, the final chemical bath solution ha s a total volume of 709 mL, and has cadmium and zinc concentrations of 1.264 mM and 0.542 mM, respectively. Bath temperature conditioning The chemical bath should be kept at the desired target temper ature at all times during film growth. This can be achieved either through ma nual adjustment or auto matic feedback from a closed-loop thermocouple system. An external temperature sensor is used to report the chemical

PAGE 105

105 bath temperature to the control unit. The closed-loop system is able to automatically make all adjustments to the hot water circ ulation in order to maintain th e given set point temperature. Buffer layer deposition The procedure for chemical bath depositi on involves the step-w ise addition of dose solutions, and the timely immersion and removal of substrates. First, 490 mL of deionized water is placed in the reaction vessel and the hot water circulator control unit is set to the desired growth temperature. All openings of the reaction vessel are covere d with a plastic film that is only temporarily and briefly removed when adding e ither a substrate or dose solution. Mixing of the chemical bath solution is achieved by the use of a magnetic stirrer. Once the target growth temperature has been reached, the cadmium dose solution is slowly added to the reaction vessel. Substrates are firs t cleaned with nitrogen gas, placed in Teflon holders, and the holders are imme rsed in the reaction vessel. Next, after the bath temperature returns to within two degrees of the target value, the sulfur dose solution is slowly added to the reaction vessel, followed by the zinc dose solution. In addition, the ammonia dose solution is added to 10 mL deionized water in a separate beaker. Then, when the target growth temperature recove rs to within two degrees, the diluted ammonia dose solution is quickly poured in to the reaction vessel, starting th e reaction. The chemical bath eventually becomes cloudy, with a milky-white co lor. Depending on the growth temperature and the bath composition, this cloud iness can persist for as long as 40 minutes, yet depending on the bath composition generally transitions to a tr ansparent yellow color after 20 or 25 minutes. Termination of growth After the desired deposition time has elapsed, the Teflon holders are removed from the chemical bath solution and immediately rinsed in deionized water. The substrates must be thoroughly rinsed until no loosely-bound powder rema ins on the surface. If a visual inspection

PAGE 106

106 reveals that powder remains, the substrates are ultrasonically cleaned until powder is no longer evident. The substrate is then thoroughly dried with nitrogen gas a nd vacuum sealed or stored in a clean, dry environment. Zinc Cadmium Sulfide Buffer L ayer Characterization Results The effect of zinc concentration on chemical bath deposition of zinc cadmium oxide is presented. The thickness of ZnCdS films is determ ined via profilometry. The zinc incorporation of deposited film is determined by inductiv ely coupled plasma spectroscopy and x-ray photoelectron spectroscopy. UV/vis spectroscopy is employed to determine the optical characteristics of CBD grown ZnCdS in the range of 300 nm to 800 nm. Growth Rate The growth rate of ZnCdS thin films deposit ed via CBD is determined by measuring film thickness at different deposition times. Afte r deposition, AZ Electronic Materials 1529 Photoresist is applied to the por tions to be measured. The covered films are soft-baked for approximately 10 minutes at 120C. The exposed Zn CdS film is then etched with a 10% nitric acid solution, until only the film covered by the photor esist is left on the substrate. The substrate is then rinsed with deionized water, the re maining photoresist removed with acetone, and the sample is again thoroughly rinsed in deionized water. Film thickness is then measured using a Veeco Dektak 150 profilometer. Zinc cadmium sulfide average film thickness is reported in Table 4-3 and shown in Figure 4-1 as a function of deposition time at 85C fo r films grown with prep ared fractional zinc compositions of xp = 0.1, 0.2, 0.3, 0.4, and 0.5. For each prepared zinc composition, four samples are deposited simultaneously. Deposition time is controlled by removing samples at 15, 25, 35, and 45 minutes. Measured ZnCdS average film thickness is reported on the vertical axis, with increasing deposition time on the horizontal axis. Each mark er shape represents a specific

PAGE 107

107 prepared fractional zinc compositi on as denoted in the legend. For example, average thicknesses of 26.7 nm and 55.0 nm were measured for a prepared zinc composition of xp = 0.2 at deposition times of 25 and 45 minutes, respectively. Fo r all values of prepared zinc composition, xp, ZnCdS film thickness is seen to increase with increasing deposition time. Film growth rate appears inversely related to the prepared zinc composition, as film thickness decreases with increasing content of zinc in the prepared bath solution. Film Composition The film composition of ZnCdS thin films ( xf) deposited via CBD is determined by ICP and XPS. A Perkin-Elmer Plasma 3200 is used for ICP measurements and a Perkin-Elmert PHI 5100 ESCA System is used for XPS measurements Films are deposited for 45 minutes at 85C with prepared fractional zinc compositions of xp = 0.1, 0.2, 0.3, 0.4, and 0.5. For XPS analysis, samples are cut and measured as received and afte r seven minutes of sputtering. For ICP, a 10% nitric acid solution is used to dissolve the deposited ZnCdS films. Table 4-4 reports the comparison between the two measurement tec hniques for films deposited for 45 minutes at 85C. The film fractional zinc compositions measur ed by ICP for films deposited for 45 minutes is shown in Figure 4-2. The fr actional zinc composition measured in the film is reported on the vertical axis while the prepared fractional zinc composition is reported on the horizontal axis. The red dashed line corresponds to a 1:1 ratio of zinc composition in the prepared solution and in the as-deposited film. Prep ared zinc compositions of xp = 0.1 and xp = 0.2 result in less incorporation of zinc into the film, where film zinc compositions of xf = 0.05 and xf = 0.02 were measured by ICP. The opposite of this trend is seen with prepared zinc compositions of xp = 0.4 and xp = 0.5, where film zinc compositions ( xf) were measured to be xf = 0.65 and xf = 0.81. However, a prepared zinc composition of xp = 0.3 results in a film fr actional zinc composition of

PAGE 108

108 xf = 0.32 which means that Zn and Cd are depositi ng in the same ratios as found in the chemical bath solution. To determine the effect of deposition time on film composition, ICP was used to measure films deposited at 85C for 15, 25, 35, and 45 minu tes. Figure 4-3 presents the results of the measured fractional zinc composition (xf) as a function of depositi on time for prepared zinc compositions of xp = 0.1, 0.2, 0.3, 0.4, and 0.5. The measured fractional zinc composition is graphed on the vertical axis while increasing deposition time is gra phed on the horizontal axis. It appears that for values of xp 0.2 zinc is incorporated into th e film faster than cadmium. For example, for xp = 0.3, the film zinc composition is xf = 0.79 after 15 minutes, xf = 0.65 after 25 minutes, xf = 0.49 after 35 minutes, and finally xf = 0.32 after 45 minutes of deposition. A visual comparison of the measured film zinc fraction ICP and XPS data in Table 4-4 is seen in Figure 4-4. The measured fractional zinc composition if plotted on the y-axis and the five values of prepared fractional zinc composition are plotted on the x-axis. Each measurement technique is assigned a specific shape as de noted in the legend. The two measurement techniques produce similar results at the high and low end of the prepared zinc composition spectrum. However, there is disagreement at a prepared zinc composition of xp = 0.3 where XPS reports values of xf = 0.62 as received and xf = 0.46 after seven mi nutes of sputtering. Optical Characteristics The optical transmission of ZnCdS films is measured using a Perkin-Elmer Lambda 800 UV/Vis spectrophotometer. Films are deposited at 85C with various deposition times of 15, 25, 35, and 45 minutes with prepared fractional zinc compositions of xp = 0.1, 0.2, 0.3, 0.4, and 0.5. In order to accurately measure the transmission of a single ZnCdS film, a 10% nitric acid solution is used to remove the film deposited on the back of all substrates. Figures 4-5 through 4-9 show the optical transmission as a function of wavelength for deposition times of 15, 25, 35,

PAGE 109

109 and 45 minutes. For each figure, transmission is reported on the vertical axis and wavelength is reported on the horizontal axis. Ea ch deposition time is assigned a line type as denoted in the legend, where SLG denotes bare soda lime glass with no ZnCdS deposition. For each prepared fractional zinc composition, the transmission be low 500 nm decreases with increasing deposition time. This is expected as film thickness increases with increasing deposition time thereby decreasing optical transmission. The measured transmission for films deposited for 45 minutes at 85C as a function of prepared fractional zinc composition ( xp) is shown in Figure 4-10. Transmission is reported on the vertical axis and wavelength if reported on the horizontal axis. Each prepared fractional zinc composition is assigned a line type as denoted in the legend. A distin ct variation between prepared values of zinc can be seen at wavelengths less than 500 nm as xp = 0.4 produces a film with increased transmission compared to th e film grown with a prepared bath of xp = 0.2. To further analyze the difference in transmission be tween these films, the average transmission for each prepared zinc value is calculated over three wavelength ranges, 300 nm to 800 nm, 300 nm to 550 nm, and 500 nm to 800 nm. Figure 4-11 plots the average transmission in each wavelength range for the ZnCdS films shown in Fi gure 4-10. Average tran smission is plotted on the vertical axis and the prepar ed fractional zinc composition is plotted on the horizontal axis. The three wavelength range are assigned markers as denoted in the legend. The same data is listed in Table 4-5. Over the entire 300 nm to 800 nm range, the sample deposited with prepared zinc composition xp = 0.4 reports the highest average tr ansmission value of 82.32%. Comparing the remaining two wavelength ranges, it is eviden t that the high average transmission value for xp = 0.4 is a result of its increased transmissi on at wavelengths less than 550 nm, with a 73.52%

PAGE 110

110 average transmission compared to 64.29% aver age transmission for the film deposited with a prepared zinc composition of xp = 0.2. Conclusions Chemical bath deposition is an extremely a ttractive process as it provides a low cost method to producing high quality films with high throughput. In particular, chemical bath deposition is an excellent candidate for deposition of ZnCdS thin films for use as buffer layers in photovoltaic devices. The addition of Zn to the tr aditional CdS buffer laye r results in a larger band-gap energy that aids to increase transmi ssion in the short-wavelength region. Zinc cadmium sulfide also yields a more favorable conduction band alignment a nd provides a lattice parameter that more closely matches CIGS absorber material. Experimental studies show that thickness of ZnCdS film increases with increasing deposition time up to 45 minutes. Thickness is seen to be a function of the prepared fractional zinc composition, where film thickness decreases with increasing zinc content in the prepared bath. Deposited film zinc content is also seen to be dependant on the pr epared zinc composition, where values of xp = 0.1 and 0.2 show less zinc incorpor ation into the film and values of xp = 0.4 and 0.5 result in fractional zinc composition of xf = 0.65 and 0.81 as reported by ICP, respectively. ICP results also show that a prepared zinc composition of xp = 0.3 results in a film with a zinc composition of xf = 0.32, although slightly contra dicted by XPS results. Zinc incorporation into the film is seen to be a function of deposition time, as the zinc-to-cadmium incorporation on the substrate is increased at reduced deposition times. Films deposited with prepared zinc compositions of xp = 0.3 and xp = 0.4 show average optical transmissions greater than 80% for wavelengths from 300 nm to 800 nm. In particular, it is critical to maximize transmission at wavelengths less than 550 nm to increase device short-circ uit current and both of these films show average transmission of at least 70% in this region.

PAGE 111

111 Table 4-1. Recipe for the preparation of the chemical bath solution for five different zinc compositions. Prepared zinc composition ( x p ) Cd Source (mL) Zn Source (mL) S Source (mL) NH4OH Source (mL) 0.1 137.10 0.70 100 0.25 0.2 121.90 1.40 100 0.25 0.3 106.70 2.09 100 0.25 0.4 91.40 2.79 100 0.25 0.5 76.20 3.49 100 0.25 Table 4-2. Resulting chemical bath solutio ns for five different zinc compositions. Prepared zinc composition ( x p ) Overall bath volume (mL) Cadmium concentration (mM) Zinc concentration (mM) 0.1 738 1.560 0.173 0.2 724 1.415 0.354 0.3 709 1.264 0.542 0.4 694 1.106 0.737 0.5 680 0.941 0.941 Table 4-3. Zinc cadmium sulf ide film thicknesses for five values of prepared fractional composition of zinc at 85C. Thickness (nm) Deposition Time (min) xp = 0.1 xp = 0.2 xp = 0.3 xp = 0.4 xp = 0.5 15 22.5 17.5 13.3 12.5 8.3 25 37.5 26.7 20.0 17.5 13.3 35 60.0 37.5 31.7 27.5 16.7 45 80.0 55.0 43.3 37.5 22.5

PAGE 112

112 Table 4-4. Comparison of m easured fractional zinc compos ition in films deposited for 45 minutes with five prepared zinc compositions at 85C. Fractional Composition (%) Prepared Fractional Zinc Composition (x p ) ICP XPS (as received) XPS (7 min sputter) 0.1 0.05 0.00 0.00 0.2 0.02 0.00 0.00 0.3 0.32 0.62 0.46 0.4 0.65 0.69 0.65 0.5 0.81 0.76 0.77 Table 4-5. Average transmissi on of three wavelength ranges fo r ZnCdS films deposited for 45 minutes with five prepared zinc compositions at 85C. Wavelength Range Prepared Fractional Zinc Composition (x p ) 300nm 550nm 550nm 800nm 300nm 800nm 0.5 70.03 90.66 80.31 0.4 73.62 91.07 82.32 0.3 70.01 90.76 80.36 0.2 64.29 91.25 77.73 0.1 67.29 92.69 79.95

PAGE 113

113 Figure 4-1. Film thickness as a function of de position time for ZnCdS chemically bath deposited at five prepared fractional zinc compositions at 85C.

PAGE 114

114 Figure 4-2. Measured fractional zinc composition (xf) of five films chemically bath deposited at 85C for 45 minutes at five prepared zinc compositions, xp = 0.1, 0.2, 0.3, 0.4, and 0.5.

PAGE 115

115 Figure 4-3. As deposited fr actional zinc composition, xf, as a function of deposition time for five prepared zinc compositions, xp = 0.1, 0.2, 0.3, 0.4, and 0.5, at 85C.

PAGE 116

116 Figure 4-4. Comparison of meas ured zinc composition of films deposited at 85C for 45 minutes with prepared fractional zinc compositions of xp = 0.1, 0.2, 0.3, 0.4, and 0.5.

PAGE 117

117 0 10 20 30 40 50 60 70 80 90 100 300350400450500550600650700750800Transmission (%)Wavelength (nm) SLG 15 min 25 min 35 min 45 minFigure 4-5. Optical tran smission as a function of wavelength for films deposited at 85C with a prepared fractional zinc composition of xp = 0.1, at deposition times of 15, 25, 35, and 45 minutes.

PAGE 118

118 0 10 20 30 40 50 60 70 80 90 100 300350400450500550600650700750800Transmission (%)Wavelength (nm) SLG 15 min 25 min 35 min 45 minFigure 4-6. Optical tran smission as a function of wavelength for films deposited at 85C with a prepared fractional zinc composition of xp = 0.2, at deposition times of 15, 25, 35, and 45 minutes.

PAGE 119

119 0 10 20 30 40 50 60 70 80 90 100 300350400450500550600650700750800Transmission (%)Wavelength (nm) SLG 15 min 25 min 35 min 45 minFigure 4-7. Optical tran smission as a function of wavelength for films deposited at 85C with a prepared fractional zinc composition of xp = 0.3, at deposition times of 15, 25, 35, and 45 minutes.

PAGE 120

120 0 10 20 30 40 50 60 70 80 90 100 300350400450500550600650700750800Transmission (%)Wavelength (nm) SLG 15 min 25 min 35 min 45 minFigure 4-8. Optical tran smission as a function of wavelength for films deposited at 85C with a prepared fractional zinc composition of xp = 0.4, at deposition times of 15, 25, 35, and 45 minutes.

PAGE 121

121 0 10 20 30 40 50 60 70 80 90 100 300350400450500550600650700750800Transmission (%)Wavelength (nm) SLG 15 min 25 min 35 min 45 minFigure 4-9. Optical tran smission as a function of wavelength for films deposited at 85C with a prepared fractional zinc composition of xp = 0.5, at deposition times of 15, 25, 35, and 45 minutes.

PAGE 122

122 0 10 20 30 40 50 60 70 80 90 100 300350400450500550600650700750800Transmission (%)Wavelength (nm) SLG xp = 0.5 xp = 0.4 xp = 0.3 xp = 0.2 xp = 0.1Figure 4-10. Optical transmission as a function of wavelength for films deposited at 85C for 45 minutes with five prepared fractional compositions, xp = 0.1, 0.2, 0.3, 0.4, and 0.5.

PAGE 123

123 Figure 4-11. Average transmission of three wa velength ranges for ZnCdS films deposited at 85C for 45 minutes with five prepared zinc compositions, xp = 0.1, 0.2, 0.3, 0.4, and 0.5.

PAGE 124

124 CHAPTER 5 CHARACTERIZATION OF SPUTTER DE POSITED ZINC OXIDE Introduction Zinc oxide is a wide band-gap (3.3 eV) semiconductor with many technological applications such as gas sensors, phosphors, piezoelectric transducers, and transparent conducting electrodes in optoelectronic devices [92-94]. Zinc is inexpensive, non-toxic, and zinc oxide has several advantages over other TCO mate rials including high transmittance in the shortwavelength region and is abundant in natural resource. Several deposition techni ques such as pulsed laser deposition [95], spu ttering [96-100], spray pyrolysis [101, 102], and chemical vapor deposition [103-105] have been used to deposit in trinsic zinc oxide thin films. Among these, sputtering is the most widely us ed as it has outstanding advantages such as simple apparatus, high deposition rates, low deposition temperature, large area deposition, good adhesion, and easy control of doping concentration. Unfortunately, it is very difficult to deposit intrinsic ZnO films that exhibit sufficiently high conductivit y for use in optoelectronic devices. Aluminum-doped zinc oxide (AZO) is a popular a lternative [99, 106, 107] to intrinsic zinc oxide as it exhibits high transmittance, low resi stivity, can be easily deposited, and its band-gap energy can be altered by changing the amount of aluminum dopant. The addition of hydrogen during sputtering is seen to signi ficant reduce the resist ivity of AZO films [108]. It is believed that hydrogen acts as a shallow donor in ZnO thin films to achieve high performance, greatly improving the Al doping efficien cy thereby enhancing the conduc tivity of the AZO thin films [109, 110]. For sputter deposited AZO, aluminum and hydrogen cont ent, substrate temperature, working gas pressure, RF power, and the working di stance are all parameters that can affect the structural, optical, and electrical properties of doped zinc oxide.

PAGE 125

125 There is no single guaranteed r ecipe or deposition technique th at yields Al:ZnO films with superior optoelectronic properties. Therefore, it is necessary to further investigate which knobs can be turned to produce the most optimal deposition conditions for the sputtering system at the University of Florida. The objective of this work is to determine the film characteristics of Al and H doping of ZnO sputter deposited films and intends to improve upon the quality of AZO and HAZO by investigating a number of factors that influence film thickness, sheet resistance, optical transmission, and structural characteristics. In addition, it is desired to deposit zinc oxide at higher base pressures in hopes of achieving the potential for increased industrial throughput while maintaining or improving film properties. Experimental This work consists of sputter deposition of Al:ZnO (AZO) and H-Al:ZnO (HAZO) thin films on glass substrates. Samples are deposit ed via rf magnetron sputtering using argon and hydrogen-doped argon working gas. Four-point probe and uv/vi s spectroscopy are used to measure film resistivity and optical transmission. Film thickness is measured using profilometry and powder x-ray diffraction is used to analyze the structural characte ristics of AZO and HAZO films. A figure of merit calculation is calcu lated for each film to determine which process conditions achieve optimal film characteristics. Equipment A Perkin-Elmer Model 4400 Production Sputtering system is used to sputter deposit zinc oxide thin films on Corning 1737 glass substr ates with a base pressure of 2.6 x 10-6 Torr. Aluminum-doped zinc oxide is deposited using a zinc oxide target that is doped with Al2O3 in an ultra-high purity Ar working gas environment. The content of Al2O3 in the target is 2% in weight and measures eight inches in diamet er. Hydrogen-aluminum-doped zinc oxide is

PAGE 126

126 deposited using the Al2O3 doped target along with an ultra-high purity Ar and H2 (0.1%) sputtering gas mixture. An Alessi four-point probe system is used to measure the resistivity of the sputter deposited films. The system features 0.05 spaced tungsten-carbide probe tips, which are 0.002 in diameter. Currents of 1 A to 100 mA are supplied by a Crytronics Model 120 current source. The voltage resulting across the inner pr obes is measured by a Keithley Model 181 Nanovoltmeter which is capable of 10 nV sensiti vity. Optical transmission of the films is measured from 400 nm to 800 nm using a Perk in-Elmer Lambda 800 spectrophotometer. The system uses twin beams, tungsten and deuterium, to collect transmittance spectra and a reference sample to provide background corr ection of measured spectra. The structural characteristics of the film s are analyzed with a Philips APD 3720 powder diffractometer operating at 40KV and 20mA, which uses Cu radiation ( = 0.154056 nm). Survey spectra scans are run with a 0.03 step size and a step time of 0.5 second per step. High resolution spectra scans are produce d by reducing the step size to 0.01 and increasing the scan rate to 1.0 second per step. The thickness of the films is obtained using a Veeco Dektak 150 Surface Profiler, which has a ver tical measurement range of 512 m and can perform precise step measurements of thin films down to 100 Materials Zinc oxide thin films are sputter deposite d on Corning 1737 glass substrates, which must be thoroughly cleaned before use. Glass substrat es are washed with deio nized water to remove visible contaminants and subjected to a series of hot deionized water ba ths. Substrates are mechanically scrubbed with an Alconox detergent solution, placed in a hot deionized water bath,

PAGE 127

127 then ultrasonic cleaned, placed in a three-step cas cade of deionized water, and finally dried with nitrogen gas. The glass subs trate cleaning procedure is furt her detailed in Chapter 3. Aluminum-doped and hydrogen-aluminum-doped film s are sputtered from an 8-inch zinc oxide target doped with Al2O3 (2wt%). Al-doped ZnO is deposited in an UHP Ar working gas environment while a mixture of UHP Ar and H2 (0.1%) serves as the working gas for H-Aldoped ZnO. Acetone is used to remove Sharpi e ink when creating tren ches for subsequent thickness measurements. Zinc Oxide Characterization Results The effect of sputter deposition parameters such as sample positioning under the target and working gas flowrate is pres ented. The thickness of AZO and HAZO is determined via profilometry. Four-point probe measurements ar e used for electrical char acterization of sputter deposited films in the form of sheet resistance and used in combination with film thickness to calculate resistivity. UV/Vis sp ectroscopy is used to determine the optical characteristics of annealed ZnO in the range of 400 nm to 800 nm. Structural characteri zation of AZO and HAZO thin films is performed via x-ray diffraction (XRD ). A figure of merit ca lculation is used to select the growth parameters that produce f ilms with superior optoe lectronic properties. Film Thickness The thickness of Al:ZnO and HAl:ZnO sputter deposited fi lms is determined using a Veeco Dektak 150 profilometer. Average thickne ss is calculated by taking the mean of three measurements along the etched tren ch. Unless otherwise stated, the target is conditioned with a pre-sputter before opening the shutter to start deposition on the SLG substrates. The effect of flowrate and positioning under the target is analyzed for each working gas. The base pressure of the sputter deposition chamber is kept constant for all thickness experiments, on the order of

PAGE 128

128 2.6 x 10-6 Torr. Films are sputtered from the Al2O3-doped zinc oxide target both for AZO and HAZO. Aluminum doped A contour plot of aluminum-doped zinc oxide (AZO) film thickness is shown in Figure 51. The film is deposited for 10 minutes with 10 minutes pre-sputter of th e target before opening the shutter to begin the depositi on. The vertical axis represen ts the y-coordinate location under the sputter target and the horizontal axis repr esents the x-coordinate location. The color assignments denoted in the legend represent th ickness equal to or less than their assigned thickness value. The RF power is kept at 0. 4 kW during deposition with the argon working gas flowrate maintained at 40 sccm. The measured fi lm thickness on the 6" x 6" sample is seen to reach a minimum of around 160 nm (2,3) and a maximum close to 300 nm (-2,-3). Film thickness is also seen to decrease toward the lo wer right hand area of the sample (2,-3). These lower areas of film thickness correspond to areas that are positioned closer to the edge of the target. In order to consistently compare film s between experiments, four sample locations measuring 1" x 2" are defined as seen in Figure 5-2. These four sample positions are kept constant for all experiments by placing a grid on the substrate platen to ensure samples are accurately positioned for each deposition. The average thickness of AZO deposited at four working gas flowrates is shown in Figure 5-3, and is calculated by averag ing three measurements along the tr ench. The vertical axis plots the average thickness while the horizontal axis categorizes the deposited AZO by sample location under the target. Each wo rking gas flowrate is assigned a marker style as denoted in the legend. For example, with an argon working ga s flowrate of 40.0 sccm the average measured thickness at sample position 3 is 236.7 nm. The RF power is kept constant at 0.4 kW for all conditions, with a base pressu re of approximately 2.6 x 10-6 Torr. The flowrate of argon

PAGE 129

129 working gas is adjusted from 30.0 sccm to 60. 0 sccm in 10.0 sccm increments. The average AZO measured thickness at each flowrate is rou ghly a function of working gas flowrate, which is directly related to working gas pressure. At position 4, 60.0 sccm Ar gives an average thickness of approximately 247 nm compared to 238 nm deposit ed at 30.0 sccm. This represents a 3.8% variation that is fairly consis tent at all sample positions. Al though the variation percentage is similar, the amount of material deposited at sample position 1 is increased compared to all other sample positions. An argon flowrate of 60. 0 sccm resulted in a measured thickness of approximately 273 nm while only 259 nm were deposited at 30.0 sccm. Hydrogen and aluminum doped A contour plot of hydrogen and aluminum-dope d zinc oxide (HAZO) film thickness is shown in Figure 5-4. The vertical axis repres ents the y-coordinate location under the sputter target and the horizontal axis represents the x-coordinate location. The color assignments denoted in the legend represent th ickness equal to or less than their assigned thickness value. The films are deposited for 10 minutes with 10 mi nutes pre-sputter of the target before opening the shutter to begin the deposition. The RF power is kept at 0.4 kW during deposition with the hydrogen-doped argon working gas flowrate main tained at 40 sccm. The measured film thickness on the 6" x 6" sample is seen to reach a minimum of approximately 160 nm (2,3) and a maximum close to 298 nm (-2,-3). Film thickness is also seen to decrease toward the lower right hand area of the sample (2,-3), where thickness decreases towards 175 nm. These lower areas of film thickness correspond to areas that are posit ioned closer to the edge of the target and therefore see less sputtered material. The average thickness of HAZO deposited at f our working gas flowrates is pictured in Figure 5-5. The average thickness is plotted on th e vertical axis while the sample location under the target is categorized on the hor izontal axis. Each working gas flowrate is assigned a marker

PAGE 130

130 style as denoted in the legend. For example, with a hydrogen-doped argon working gas flowrate of 30.0 sccm the average measured thickness at sample position 2 is 228.3 nm. The RF power is kept constant at 0.4 kW for all conditions, w ith a base pressure of approximately 2.6 x 10-6 Torr. For sputtered HAZO films, a hydrogen-doped argon flowrate of 20.0 sccm seems to provide alternating results between sample positions, wi th thickness of 225 nm, 261.7 nm, 205 nm, and 245 nm reported for sample positions 1 through 4, respectively. The thickness at all other flowrates is much more consistent, with 40.0 sccm hydrogen-doped argon c onsistently providing the most deposition and 50.0 sccm the least at all four sample positions. Comparison of AZO and HAZO A comparison of AZO and HAZO average film thickness at each of the four sample positions under the target is shown in Figure 5-6. The average thickness is graphed on the vertical axis and the horizontal axis categorizes the deposited films by sample location under the target. Each working gas is assigned a line t ype and each working gas flowrate is assigned a marker style as denoted in the legend. Fo r example, hydrogen-doped argon working gas is assigned a dashed line, and 30.0 sccm flowrate of working gas is assigned a square marker, with a 228.3 nm average HAZO thickness at sample position 2. The RF power is kept constant at 0.4 kW for all conditions, with a base pressure of approximately 2.6 x 10-6 Torr. In general, the average film thickness is highest at sample position 1 for both types of working gas. For hydrogen-doped argon working gas the maximum deposition occurs at 40.0 sccm, with thickness at 50.0 sccm decreasing below that of 30.0 sccm This is not seen for the standard argon working gas used to deposit AZO films. The addition of the hydrogen changes the plasma characteristics and drops the sputter depositi on rate at some point between 40.0 sccm and 50.0 sccm.

PAGE 131

131 Electrical Characteristics The electrical characteristics of Al:ZnO and H-Al:ZnO sputter deposited films is determined using an Alessi four-point probe coupled with a Keithley Model 181 Nanovoltmeter. Measurements are taken at eight equally spaced locations across the sample surface. The minimum and maximum sheet resist ance values are reported rather than the average, which can be heavily influenced by outliers. Unless otherw ise stated, the target is conditioned with a presputter before opening the shutter to start depos ition on the SLG substrates. The effect of flowrate and positioning under the target is analyzed for AZO and HAZO sputter deposited films. The base pressure of th e sputter deposition chamber is kept constant for all experiments, on the order of 2.6 x 10-6 Torr. Films are sputtered from the Al2O3-doped zinc oxide target both for AZO and HAZO. Aluminum doped A contour plot of aluminum-doped zinc oxide (AZO) sheet resistan ce is shown in Figure 5-7. The film is deposited for 10 minutes with 10 minutes pre-sputter of the target before opening the shutter to begin the deposition. The vertical axis represents the y-coordinate location under the sputter target a nd the horizontal axis represents the x-coordi nate location. The color assignments denoted in th e legend represent sheet resistan ce equal to or less than their assigned sheet resistance value. The RF power is kept at 0.4 kW during deposition with the argon working gas fowrate maintained at 40.0 sccm The measured film sheet resistance on the 6" x 6" sample is seen to main tain a very low value less than 1000 /sq in the central location of the sample, with a large area, be tween (-2,-1.5) and (0, -1.5), re porting sheet resistance values less than 300 /sq. A x-coordinate values of 2 to 2.5 a large dramatic increase in sheet resistance is seen, with values reaching 13,000 /sq at (2.5, 2). As seen with film thickness,

PAGE 132

132 these areas are positioned closer to the edge of the sputter target a nd therefore receive less deposition. The minimum sheet resistance of AZO deposited at four working gas flowrates is shown in Figure 5-8. The vertical axis plots the minimum sheet resistance while the horizontal axis categorizes the deposited AZO by sample location unde r the target, as specified in Figure 5-2. Each working gas flowrate is a ssigned a marker style as denoted in the legend. For example, with an argon working gas flowrate of 30.0 sccm the minimum sheet resistance at sample position 2 is 140 /sq. The RF power is kept constant at 0.4 kW for all conditions, with a base pressure of approximately 2.6 x 10-6 Torr. The flowrate of argon working gas is adjusted from 30.0 sccm to 60.0 sccm in 10.0 sccm increments It can be seen that the minimum sheet resistance is a function of working gas flowra te. The minimum argon flowrate results in the highest sheet resistance values at all sample positions, reaching 341 /sq at sample position 4. As the argon flowrate increases, the minimum sheet resistance decr eases, with an argon flowrate of 60.0 sccm achieving the lowest sheet resist ance values at all posit ions, with position 2 providing the lowest of all sheet resistance, 98.2 /sq. Figure 5-9 shows the maximum sheet resistan ce of AZO deposited at four working gas flowrates. The maximum sheet resistance is plo tted on the vertical axis and the sample location under the target is categorized on the horizontal axis. Each work ing gas flowrate is assigned a marker style as denoted in the legend. Fo r example, a maximum sheet resistance of 2240 /sq is measured for 30.0 sccm argon flowrate at samp le position 1. The maximum sheet resistance follows the same trend as the minimum sheet re sistance, where increasing working gas flowrate results in decreased sheet resistance. Sample positions 1 and 4 exhibit maximum sheet resistance values that are 500% and 800%, respectively, of those measured at sample positions 2 and 3.

PAGE 133

133 Overall, a working gas flowrate of 60.0 sccm seems to provide the optimal sheet resistance conditions with an average maximum value of 269.3 /sq and a minimum value of 106.1 /sq over sample positions 1 through 3. Hydrogen and aluminum doped A contour plot of hydrogen and aluminum-doped zinc oxide (HAZO) sheet resistance is shown in Figure 5-10. The film is deposited fo r 10 minutes with 10 minutes pre-sputter of the target before opening the shutter to begin film deposition. The vertical axis represents the ycoordinate location under the sput ter target and the horizontal axis represents th e x-coordinate location. The color assignments denoted in the legend represent sh eet resistance equal to or less than their assigned sheet resistance value. Th e RF power is kept at 0.4 kW during deposition with the hydrogen-doped argon working gas flowrate maintained at 40.0 sccm. The measured sheet resistance on the 6" x 6" substrate is very similar to AZO, as a large majority of the sample exhibits relatively uniform and low sheet resist ance. The sheet resistance achieves a minimum value at the (-2, -2) location, and maximums at (2, -2) and (2, 2), which is similar to AZO. However, HAZO sheet resistance stays low at the (2, 0) location, which is not seen in the case of AZO. The minimum sheet resistance of HAZO deposited at four working gas flowrates is shown in Figure 5-11. The vertical axis plots the mi nimum sheet resistance while the horizontal axis categorizes the deposited AZO by sample location unde r the target, as specified in Figure 5-2. Each working gas flowrate is a ssigned a marker style as denoted in the legend. For example, with a hydrogen-doped argon working gas flowrate of 30.0 sccm the minimum sheet resistance at sample position 1 is 110 /sq. The RF power is kept cons tant at 0.4 kW for all conditions, with a base pressure of approximately 2.6 x 10-6 Torr. The flowrate of hydrogen-doped argon working gas is adjusted from 20.0 sccm to 50.0 sccm in 10.0 sccm increments. As is the case

PAGE 134

134 with AZO, it appears that the minimum sheet resistance of HAZO films is a function of the working gas flowrate. The sheet resistance decrea ses with increasing working gas flowrate from 20.0 sccm and reaches a minimum at 40.0 sccm, with max and min values of 136 /sq and 95.1 /sq at sample position 2, respectively. A slight increase in sheet resistance is then seen upon increasing the hydrogen-doped argon working gas flowrate from 40.0 sccm to 50.0 sccm. For sample position 2, the minimum sheet resistance increase from 95.1 /sq to 101 /sq when increasing from 40.0 sccm to 50.0 sccm. Figure 5-12 portrays the maximum sheet resistance of HAZO deposited at four working gas flowrates. The maximum sh eet resistance is plotted on the vertical axis and the sample lo cation under the target is categori zed on the horizontal axis. Each working gas flowrate is assigne d a marker style as denoted in the legend. For example, a maximum sheet resistance of 1440 /sq is measured for 30.0 sccm hydrogen-doped argon flowrate at sample position 4. The maximum sheet resistance of HAZO films follows a similar trend where increasing working gas flowrate resu lts in diminished sheet resistance. The only exception to this is 20.0 sccm which reports valu es almost idental to 30.0 sccm at all sample position except position 4, where a mu ch reduced sheet resistance of 793 /sq is measured compared to 1440 /sq at 20.0 sccm hydrogen-doped argon flowrate. The maximum sheet resistance at sample positions 2 and 3 is on aver age three to four times smaller than the sheet resistance at positions 1 and 4. These positions e xhibit a very tight spread for the entire range of working gas flowrate, varying from 137 /sq to 227 /sq and 217 /sq to 367 /sq at sample positions 2 and 3, respectively. Sample positions 1 and 4 report ranges of 691 /sq and 647 /sq, respectively. Overall, working gas fl owrates of 40.0 sccm and 50.0 sccm hydrogendoped argon appears to provide to provide optimal sheet resi stance with an minimum sheet

PAGE 135

135 resistance in the range of 95.1 /sq to 101 /sq and maximum values of 183 /sq and 136 /sq at sample position 2. Comparison of AZO and HAZO A comparison of minimum sheet resistance for AZO and HAZO sputter deposited films is shown in Figure 5-13. The minimum sheet resist ance is graphed on the ve rtical axis and the horizontal axis categorizes th e deposited films by sample loca tion under the target. Each working gas is assigned a line type and each worki ng gas flowrate is assigned a marker style as denoted in the legend. For example, a hydroge n-doped argon working gas is assigned a dashed line, and 30.0 sccm flowrate of working ga s is assigned a square marker, with a 103 /sq minimum sheet resistance measured at sample pos ition 2. The RF power is kept constant at 0.4 kW for all conditions, with a base pressure of approximately 2.6 x 10-6 Torr. The overall distribution of minimum sheet re sistance is much tighter across sample positions for HAZO films than those films deposited with argon work ing gas. In particular, the minimum sheet resistance at 30.0 sccm Ar is markedly higher th an other argon working gas flowrates. A similar comparison of maximum sheet resistance for AZO a nd HAZO films can be seen in Figure 5-14. The maximum sheet resistance is plotted on the vert ical axis and the horiz ontal axis categorizes the deposited films by sample location under the targ et. Each working gas is assigned a line type and each working gas flowrate is assigned a marker style as denoted in the legend. For example, an argon working gas is assigned a solid line, and 40.0 sccm of woring gas is assigned a triangle marker, with a 1600 /sq maximum sheet resistance measured at sample position 1. A decrease in the maximum sheet resistance is seen with an increase in the working gas flowrate. Also, the addition of hydrogen to the working gas is shown to decrease the sheet resist ance at each flowrate and sample position. As seen before, sample positions 2 and 3 appear to produce the

PAGE 136

136 highest quality films as the sheet resistance variat ion between working gas flowrates is kept to a minimum. Film resistivity is calculated with the sheet resistance data using the average measured thickness reported earlier this chapter. Figure 5-15 shows the minimum re sistivity of sputter deposited AZO and HAZO films. The minimum re sistivity is graphed on the vertical axis and the horizontal axis categorizes the deposited films by sample locat ion. Each working gas is assigned a line type and each working gas flowrate is assigned a market style as denoted in the legend. For example, a hydrogen-doped argon wo rking gas is assigned a dashed line, and 50.0 sccm flowrate of working gas is a ssigned a circle marker, with a 2.47 x 10-3 cm minimum resistivity measured at sample position 1. In general, sample position 2 achieves good resistivity a number of flowate and working gas conditi ons. A minimum resistance of 2.28 x 10-3 cm is achieved at sample position 2 with a 50.0 sccm flowrate of hydrogendoped argon working gas. Optical Characteristics The optical transmission of AZO and HAZO thin films is studied using UV/Vis spectroscopy in the 300 nm to 900 nm range. Th e transmittance spectra are measured without the use of an integrating sphere. A bare glass substrate with no deposited ZnO is measured in order to factor out transmission loss from the s ubstrate itself. The average transmission between 400 nm and 800 nm of AZO and HAZO films before and after thermal trea tment is presented. The effects of working gas flowrate and sa mple position under the target on the average transmission of AZO and HAZO fi lms are shown in Figure 5-16. The average transmission is graphed on the vertical axis, and the horizonta l axis categorizes the films by sample position under the target. Each working gas flowrate is as signed a marker color and shape, as denoted in the legend. Filled-in markers correspond to argon working gas while empty markers correspond to hydrogen-doped argon working gas. For example, the average transmittance of HAZO

PAGE 137

137 deposited at 40.0 sccm in sample position 1 is 82.3%. For HAZO, the highest transmission is observed at 50.0 sccm hydrogendoped argon except for sample position 2 where 20.0 sccm results in 86.4% transmission. At all other sa mple positions 20.0 sccm only produces the second highest transmission. The vari ation in transmission is much smaller for AZO films., where transmission is observed to increase with working gas flowerate. The highest transmittance is measured at sample position 2 for both AZO and HAZO is 86.4% and 83.7%, respectively. Figure of Merit High transmission and low electrical resistivity are two parameters that are extremely important to transparent conductors. In order to evaluate the quali ty of an optoelectronic film, a figure of merit performance parameter is calculated which is based off the ratio of film resistivity to the visible transmission. Th e figure of merit is defined as FOM 1ln T (5-1) where is the film resistivity and T is the average transmission in the 400 nm to 800 nm range [111-113]. A larger figure of merit (FOM) va lue indicates superior optoelectronic film properties. The effect of working gas flowrate and samp le position under the targ et on the figure of merit of AZO and HAZO is shown in Figure 5-17. The vertical axis plots the figure of merit and the horizontal axis categorizes the films by sample position. Each working gas flowrate is assigned a marker color and shape, as denoted in the legend. Filled-in markers correspond to argon working gas while empty markers correspo nd to hydrogen-doped argon working gas. For example, the FOM of AZO deposited at 30.0 sccm in sample position 3 is 1234 -1cm-1. The figure of merit is seen to be the highest for th e films deposited at sample position 2. This is mostly a result of the superior resistivity exhibited by the films deposited in that sample location.

PAGE 138

138 Sample positions 1 and 3 produce similar result s with sample position 4 exhibiting the worst FOM values. The figure of merit increases with increasing hydrogen-doped argon working gas flowrate. This same trend is seen for AZO fi lms, with 60.0 sccm of argon working gas produced the highest figure of merit at each sample posit ion, and 30.0 sccm the lowest. Deposition at 50.0 sccm and 60.0 sccm at sample position 2 produces HAZO and AZO, respectively, with superior optoelectronic propertie s compared to lower flowrates or sample locations under the target. Conclusions The sputter deposition of aluminum-doped and hydrogen-aluminum-doped zinc oxide is carried out via rf magnetron sputtering. The possi bility of deposition at a higher than usual base pressure is studied in hopes of achieving the po tential for increased i ndustrial throughput while maintaining or improving on film properties. Both AZO and HAZO films are sputtered at 0.4 kW and a base pressure of 2.6 x 10-6 Torr. The effect of gas working, which effects the working pressure, is also studied. The averag e AZO film thickness is seen to range from 235.0 nm to 273.3 nm and 225.0 nm to 265.0 nm for HAZO, depending on the positioning under the target. Sheet resistance is also affected by sample positioning and working gas flowrate. The minimum AZO resistivity of 2.37 x 10-3 cm is achieved at 60.0 sccm argon, while the minimum HAZO resistivity of 2.28 x 10-3 cm is achieved at 50.0 s ccm hydrogen-doped argon. Transmission of AZO and HAZO films vary be tween 81.0% and 86.4% depending on working gas flowerate and sample positioning. The figure of merit calculation is used for determining the optoelectronic superiority of the deposited film s. The best AZO and HAZO films deposited exhibit FOM values of 2376 -1cm-1 and 2672 -1cm-1, respectively.

PAGE 139

139 Table 5-1. Average thickness of AZO films at flowrates of 30.0, 40.0, 50.0, and 60.0 sccm. Position 1 Thickness (nm) Position 2 Thickness (nm) Position 3 Thickness (nm) Position 4 Thickness (nm) 30.0 sccm Ar 258.33 235.00 240.00 238.33 40.0 sccm Ar 258.33 240.00 236.67 241.67 50.0 sccm Ar 261.67 236.67 236.67 241.67 60.0 sccm Ar 273.33 241.67 236.67 246.67 Table 5-2. Average thickness of HAZO films at flowrates of 20.0, 30.0, 40.0, and 50.0 sccm. Position 1 Thickness (nm) Position 2 Thickness (nm) Position 3 Thickness (nm) Position 4 Thickness (nm) 20.0 sccm Ar+H2 225.00 261.67 205.00 245.00 30.0 sccm Ar+H2 258.33 228.33 230.00 235.00 40.0 sccm Ar+H2 265.00 246.67 238.33 248.33 50.0 sccm Ar+H2 248.33 225.00 220.00 230.00

PAGE 140

140 Figure 5-1. Thickness contour plot of AZO film deposited at 0.4 kW with 40 sccm argon working gas flowrate.

PAGE 141

141 Figure 5-2. Four defined sample positions under the Al2O3-doped zinc oxide target.

PAGE 142

142 Figure 5-3. Average AZO film thickness deposite d at 0.4 kW with working gas flowrates of 30.0, 40.0, 50.0, and 60.0 sccm argon.

PAGE 143

143 Figure 5-4. Thickness contour pl ot of HAZO film deposited at 0.4 kW with 40 sccm hydrogendoped argon working gas flowrate.

PAGE 144

144 Figure 5-5. Average HAZO film thickness deposited at 0.4 kW with working gas flowrates of 20.0, 30.0, 40.0, and 50.0 sccm hydrogen-doped argon.

PAGE 145

145 Figure 5-6. Comparison of AZ O and HAZO average film thickness as a function of sample position and working gas composition and flowrate.

PAGE 146

146 Figure 5-7. Sheet resistance c ontour plot of AZO film deposite d at 0.4 kW with 40 sccm argon working gas flowrate.

PAGE 147

147 Figure 5-8. Minimum AZO sheet resistance deposited at 0.4 kW with working gas flowrates of 30.0, 40.0, 50.0, and 60.0 sccm argon.

PAGE 148

148 Figure 5-9. Maximum AZO sheet resistance deposite d at 0.4 kW with working gas flowrates of 30.0, 40.0, 50.0, and 60.0 sccm argon.

PAGE 149

149 Figure 5-10. Sheet resistance contour plot of HAZO film deposited at 0.4 kW with 40.0 sccm hydrogen-doped argon working gas flowrate.

PAGE 150

150 Figure 5-11. Minimum HAZO sheet resistance deposited at 0.4 kW with working gas flowrates of 20.0, 30.0, 40.0, and 50.0 sccm hydrogen-doped argon.

PAGE 151

151 Figure 5-12. Maximum HAZO sheet resistance deposited at 0.4 kW with working gas flowrates of 20.0, 30.0, 40.0, and 50.0 sccm hydrogen-doped argon.

PAGE 152

152 Figure 5-13. Minimum sheet re sistance of AZO and HAZO films deposited at 0.4 kW with working gas flowrates of 30.0, 40.0, and 50.0 sccm.

PAGE 153

153 Figure 5-14. Maximum sheet re sistance of AZO and HAZO films deposited at 0.4 kW with working gas flowrates of 30.0, 40.0, and 50.0 sccm.

PAGE 154

154 Figure 5-15. Minimum resistivity of AZ O and HAZO films deposited at 0.4 kW.

PAGE 155

155 Figure 5-16. Average transmission of AZ O and HAZO films deposited at 0.4 kW.

PAGE 156

156 Figure 5-17. Figure of merit for AZ O and HAZO films deposited at 0.4 kW.

PAGE 157

157 CHAPTER 6 THERMAL TREATMENT OF ZINC OXIDE Introduction The application of thermal treatments has l ong been a method to increase the quality of thin films. Thermal treatments in reducing ambien ts have been shown to lower the resistivity of Al-doped ZnO thin films by an order of magnitude to the 10-3 to 10-4 cm range [114-117]. Hydrogen treatments tend to see dramatic improvement in film resitivity and carrier concentration due to desorption of negatively charged oxygen species from grain boundaries, which form potential barriers by acting as trapping sites [103, 118]. Al-doped zinc oxide (AZO) has gained significa nt attention as an abundant, low cost, and non-toxic alternative to indium tin oxide (ITO) AZO thin films, with band gap energy of 3.3 eV, exhibit low resistivities (10-4 cm) and high optical transmittance (>90%) in the visible light spectrum [119, 120]. As a result, AZO finds use in many advanced applications such as flat panel display electrode s [121], solar cells [66, 122, 123] optical waveguides [124], surface acoustic devices [125], and gas sensors [126]. Transparent conductive oxides ar e generally desired for their high transparency and excellent electrical conductivity, and are most often deposited via chemical vapor deposition [127, 128], spray pyrolysis [129], reactive sputte ring [130, 131], ion beam sputtering [132], or the sol-gel method [114, 133]. Sn-doped In2O3 (ITO), SnO2, and ZnO are examples of transparent conductive oxide s that have been widely studied [134]. Zinc oxide TCOs are most often doped, with dopants such as Ga [135], Si [136], and Al [137] The properties of these Aldoped ZnO thin films can be altered by post-deposition annealing at elevated temperatures under various ambient atmospheric compositions [111, 114-117, 138-144].

PAGE 158

158 The objective of this work is to determine th e effect of thermal treatment on the properties on Al:ZnO (AZO) and H-Al:ZnO (HAZO) thin f ilms. Sputter deposited AZO and HAZO films are annealed at 450 C for 60 minutes. The effect of thermal treatment is examined as a function of three ambient gases, ultra-hi gh purity argon, ultra-high purity nitrogen, and forming gas (96% nitrogen, 4% hydrogen). The change in structural electrical, and optical characteristics of the Al:ZnO and H-Al:ZnO thin films as a function of the three ambient gases is investigated. Anneal temperatures of 300 C to 600 C under varying ambient gases have been found to significantly improve electrical prope rties of AZO films. These el evated temperatures are often incompatible with the thermal budgets of most applications su ch as photovoltaics and display devices. Regardless, the application of these th ermal treatments provides insight into increasing the quality of as-deposited ZnO. Experimental This work consists of rapid thermal anneal ing of Al:ZnO (AZO) and H-Al:ZnO (HAZO) thin films that have been sputter deposited on glass substrates. Samples are subjected to thermal treatment at 450 C for 60 minutes under three ambient ga ses, ultra-high purity argon, nitrogen, and forming gas (96% N2, 4% H2). The effect of thermal treatm ent on film characteristics before and after thermal treatment is investigated. F our-point probe and uv/vis spectroscopy are used to measure film resistivity and optical transmission. Film thickness is measured using profilometry. Powder x-ray diffraction and atomic force micr oscopy are used to analyze the structural characteristics and surface roughness of the th ermally treated AZO and HAZO films. Atomic force microscopy is used to probe surface roughne ss of as-deposited and annealed films. Depth profiling of the thermally annealed films is provided by secondary ion mass spectrometry.

PAGE 159

159 Equipment A Perkin-Elmer Model 4400 Production Sputtering system is used to sputter deposit zinc oxide thin films on Corning 1737 glass substr ates with a base pressure of 2.6 x 10-6 Torr. Aluminum-doped zinc oxide is deposited using a zinc oxide target that is doped with Al2O3 in an ultra-high purity Ar working gas environment. The content of Al2O3 in the target is 2% in weight and measures eight inches in diamet er. Hydrogen-aluminum-doped zinc oxide is deposited using the Al2O3 doped target along with an ultra-high purity Ar and H2 (0.1%) sputtering gas mixture. A schematic of the rapid thermal processing appara tus is pictured in Figur e 6-1. It features a quartz tube chamber and sample holder that is able to accommodate samples up to 1" x 3". Three gas cylinders can be connect ed to the gas inlet and the flow of gases are controlled by the regulator on each cylinder. A halogen lamp array is located above and below the quartz chamber. The power to these lamp arrays is controlled by the temperature controller, which measures the sample temperature via a thermoc ouple and adjusts the lamp power to achieve the desired set point. The controller settings can be adjusted to provide accurate control of the sample temperature during ramp hold, and cool down phases. An Alessi four-point probe system is used to measure the resistivity of the sputter deposited films before and after therma l treatment. The system features 0.05 spaced tungstencarbide probe tips, which are 0.002 in diameter. Currents of 1 A to 100 mA are supplied by a Crytronics Model 120 current sour ce. The voltage resulting acro ss the inner probe s is measured by a Keithley Model 181 Nanovoltmeter which is capable of 10 nV sensitivity. Optical transmission of the films is measured from 300 nm to 900 nm using a Perkin-Elmer Lambda 800 spectrophotometer. The system uses twin beams, tungsten and deuterium, to collect

PAGE 160

160 transmittance spectra and a reference sample to provide background correction of measured spectra. The structural characteristics of the film s are analyzed with a Philips APD 3720 powder diffractometer operating at 40KV and 20mA, which uses Cu radiation ( = 0.154056 nm). Survey spectra scans are run with a 0.03 step size and a step time of 0.5 second per step. High resolution spectra scans are produce d by reducing the step size to 0.01 and increasing the scan rate to 1.0 second per step. The thickness of the films is obtained using a Veeco Dektak 150 Surface Profiler, which has a ver tical measurement range of 512 m and can perform precise step measurements of thin films down to 100 The surface morphology of the films are analyzed with a Digital Instruments Dimension 3100 atomic force microscope. The AFM allows for large samples up to six inches, a large scan size of up to 100 m, with a sub-angstrom vertical resolution. A JEOL JSM-6335F field emission scanning electron microscope is used to produce visual imag es of the deposited films. The FE-SEM is capable of accelerating voltage s of 0.5 to 30 KV. Both secondary and backscatter electron imaging m odes are available and can produce digital images up to 2048 x 2048 pixels. Materials Zinc oxide thin films are sputter deposite d on Corning 1737 glass substrates, which must be thoroughly cleaned before use. Glass substrat es are washed with deio nized water to remove visible contaminants and subjected to a series of hot deionized water ba ths. Substrates are mechanically scrubbed with an Alconox detergent solution before being placed in hot deionized water bath, then ultrasonic cleaned, placed in a th ree-step cascade of deionized water, and finally dried with nitrogen gas. The gla ss substrate cleaning procedure is further detailed in Chapter 3.

PAGE 161

161 Aluminum-doped and hydrogen-aluminum-doped film s are sputtered from an 8-inch zinc oxide target doped with Al2O3 (2wt%). Al-doped ZnO is deposited in an UHP Ar working gas environment while a mixture of UHP Ar and H2 (0.1%) serves as the working gas for H-Aldoped ZnO. Acetone is used to remove Sharpi e ink when creating tren ches for subsequent thickness measurements. Annealing Results The effect of ambient gas on annealing of tran sparent conductive zinc oxide is presented. Structural characterization of the annealed zi nc oxide thin films is performed via x-ray diffraction (XRD). Four-point probe measurement is used for electrical characterization of annealed ZnO thin films. Sample resistivity is determined using film thicknesses measured via profilometry. UV/Vis spectroscopy is used to determine the optical characteristics of annealed ZnO in the range of 400 nm to 800 nm. A figure of merit calculation is used to select the growth parameters that produce films with superior optoelectronic properties. Electrical Characterization Four-point probe is used to investigate the effect of thermal treatment parameters on AZO and HAZO electrical characterist ics. The sheet resistance of thin films can be obtained as Rs4.53V I (6-1) where I and V are the current and voltage, respectively, and 4.53 is the geometric factor as discussed in Chapter 3. The thickness of the fi lm can be used to calculate film resistivity Rst 4.53V I t (6-2) where resistivity is given as the product of sheet resistance and film thickness. Films are annealed for 60 minutes at 450C in forming gas, argon, and nitr ogen ambients, with ramp and

PAGE 162

162 cool down times of 10 minutes a nd 15 minutes, respectively. Th e thickness and sheet resistance are measured before and after annealing to determin e the change in film resistance as a result of the ambient gas used during the thermal treatme nt process. Both AZO and HAZO films are deposited at 0.4 kW with a base pressure of 2.6 x 10-6 Torr. A working gas flowrate of 40.0 sccm is used to deposit both AZO and HAZ O films used in this study. Samples are deposited for 10 minutes with 10 minutes pre-sputte r of the target before opening the shutter to expose the substrate to the sputtered material. A complete list of electrical characterization results for AZO and HAZO is found in Table 6-2 and 6-3, respectively. Aluminum doped The effect of thermal treatment with thr ee ambient gases on AZO minimum and maximum resistivity is shown in Figure 62. The minimum film resistivity is plotted using the left-hand vertical axis and the maximum resistivity using th e right-hand vertical axis. The horizontal axis categorizes the ambient gas used during annealing. Pre-anneal measurements are plotted with a dashed line, while post-anneal measurements ar e plotted with a solid line. The minimum and maximum values are assigned a marker style as de noted in the legend. For example, the film annealed in argon has a maximum film resistivity of 4.03 x 10-3 cm after thermal treatment. In terms of maximum film resi stivity, both forming gas and n itrogen ambients show a large decreases in resistance of 82.5% and 94.4%, resp ectively. Forming gas and nitrogen ambients also had a large impact on mi nimum resistivity, which decreases by 63.1% and 57.0% for the two gases. The use of argon ambient also resu lts in a decrease in resistance for both minimum and maximum values, however the change was not as significant compared to forming gas and nitrogen. The forming gas ambient yielded th e best results from a minimum and maximum standpoint, 9.4 x 10-4 cm and 1.04 x 10-3 cm, respectively. It is important to note that Figure 6-2 shows that an as-d eposited film with poor electr ical characteristics can be

PAGE 163

163 significantly improved by thermal treatments, as can be seen in the case of the film annealed in nitrogen. Hydrogen and aluminum doped Figure 6-3 shows the effect of forming gas, argon, and nitrogen ambients on the minimum and maximum resistivity of HAZO sputter deposited f ilms. The left-hand vertical axis is used to measure minimum resistivity and the right-hand vertical axis is used to measure maximum resistivity. A dashed line is used for pre-an neal measurements, while a solid line reflects measurements that are acquire d after annealing. The mini mum and maximum values are assigned a marker style as denot ed in the legend. For example, the HAZO film annealed in nitrogen has a maximum film resistivity of 4.81 x 10-3 cm after thermal treatment. Table 6-3 gives the film resistance data for all ambient gases, including the thickness and sheet resistance measurements used in th e resistivity calculation. It is evident that thermal tr eatment of HAZO in argon ambient is extremely detrimental to the films electrical character istics. Annealing in the ar gon ambient increased the HAZO minimum film resistivity 175.4% from 2.66 x 10-3 cm to 7.33 x 10-3 cm and the maximum resitivity increased 469.3%, from 5.974 x 10-3 cm up to 3.40 x 10-2 cm. Similar to AZO, the forming gas and nitrogen ambients decreased both minimum and maximum film resistivity. Annealing in forming gas produced a high quality HAZO film with a minimum resistivity of 1.15 x 10-3 cm, and a maximum of 1.33 x 10-3 cm. Optical Characterization The optical transmission of annealed ZnO th in films is studied using UV/Vis spectroscopy in the 300 nm to 900 nm range. The transmittance spectra are measured without the use of an integrating sphere. A bare glass substrate with no deposited ZnO is measured in order to factor out transmission loss from the substrate itself. The average transmission between 400 nm and

PAGE 164

164 800 nm of AZO and HAZO films before and after thermal treatment is presented. A complete list of optical characterizati on results for AZO and HAZO is found in Table 6-2 and 6-3, respectively. Aluminum doped The effect of thermal treatment on the average transmission of AZO films is shown in Figure 6-4. The average transmission is graphe d on the vertical axis, and the horizontal axis categorizes the ambient gas used during thermal tr eatment. A circle marker is assigned to preanneal values and a square marker to post-anneal values, as denoted in the legend. The average transmission is seen to increase 6.0%, 5.0%, and 5.6% for the AZO film annealed in forming gas, argon, and nitrogen, respectively. Most likely the transmission increase is most heavily influenced by the thermal budget of the anneal ing process rather than the gas ambient. Hydrogen and aluminum doped Figure 6-5 presents the pre-anneal and postanneal average transmission for HAZO films treated in forming gas, argon, and nitrogen ambi ents. The vertical axis plots the average transmission, and the horizontal axis categorizes the ambient gas used during thermal treatment. A circle marker is assigned to pre-anneal values and a square marker to post-anneal values, as denoted in the legend. As with AZO, the av erage transmission of HAZO films is seen to increase in all three ambient gases. However, the argon ambient only increases the HAZO transmission by 1.9%, whereas the other ambien ts each increase the transmission by 5.0%. Given the post-anneal resistivit y results of HAZO films, its po ssible that argon provides some counter-productive mechanism that is outweighe d by the increase in transmission provided by the thermal budget.

PAGE 165

165 Figure of Merit High transmission and low electrical resistivity are two parameters that are extremely important to transparent conductors. In order to evaluate the quali ty of an optoelectronic film, a figure of merit performance parameter is calculated which is based off the ratio of film resistivity to the visible transmission. Th e figure of merit is defined as FOM 1ln T (6-3) where is the film resistivity and T is the average transmission in the 400 nm to 800 nm range [111-113]. A larger figure of merit (FOM) va lue indicates superior optoelectronic film properties. The effect of annealing ambient on AZO and HAZO is determined by use of the figure of merit calculatio n. A complete list of the FOM results for AZO and HAZO is found in Table 6-2 and 6-3, respectively. Aluminum doped The effect of ambient on the figure of merit of thermally treated AZO is shown in Figure 6-6. The vertical axis plots th e figure of merit and the horizontal axis categorizes the ambient gas used during thermal treatment. A square marker is assigned to pre-anneal values and a circle marker to post-anneal values, as denoted in the legend. The FOM is seen to increase as a result of thermal treatment for all ambient conditions Forming gas and nitrogen ambients see the largets improvements with 286% and 220% increases, respectively. From these results it is determined that annealing AZO in forming gas provides the best optoelectronic film properties. The FOM of AZO annealed in forming gas increased from 2012 -1cm-1 to 7763 -1cm-1. This large increase in figure of merit is mostly due to its low resistivity as all three ambients produce fairly equivalent average transmission post-anneal.

PAGE 166

166 Hydrogen and aluminum doped Figure 6-7 presents the change in figure of merit for HAZO films annealed in forming gas, argon, and nitrogen ambients. The figure of merit is plotted on the vertical axis and the horizontal axis categorizes the ambient gas used during thermal treatment. A square marker is assigned to pre-anneal values and a circle marker to post-anneal values, as denoted in the legend. The film annealed in argon ambient shows a 60% d ecrease in FOM, which is a result of the large increase in film resistivity as a result of the thermal treatment. The nitrogen ambient results in a 61% increase in FOM, a result of both improved tr ansmission and decreased resistivity. Again, as with AZO, the film annealed in forming gas shows the largest increase in FOM. The figure of merit increases176% from 2303 -2cm-1 to 6366 -1cm-1 for the HAZO film annealed in forming gas. It is observed that th e benefits of hydrogen incorporatio n must hit some limit, as the forming gas annealed AZO produces a larger FOM than the HAZO annealed under identical conditions, especially given that the HAZO film is considered the superior film according to its pre-anneal FOM value. Structural Characteristics An SEM micrograph depicting th e structure of HAZO annealed in argon ambient is shown in Figure 6-8. The image shows a vertical view of the substrate. In areas where the thick gold coating has broken free it can be observed that the HAZO film develops in a columnar grain structure. Often it can be seen that the surface morphology appears textured or cratered. This is a consequence of the nucleation of oriented c-ax is grains that grow vertically and impinge laterally. This is the result of competition betw een the arrival rate of new Al:ZnO species on the surface and redistribution over the surface by diffusion [143]. This non-equilibrium growth results in the surface texturing of the film, which has the potentia l for enhanced light trapping in photovoltaic devices [145].

PAGE 167

167 Surface Roughness The change in the surface morphology of the AZO and HAZO sputter deposited films after thermal treatment in forming gas, argon, and ni trogen ambients is inve stigated using atomic force microscopy (AFM). A Digital Instrument s Dimension 3100 atomic force microscope is used to provide digital images and calculate arithmetic roughness (Ra) and the root mean square (rms) roughness (Rq) of the measured films. Aluminum doped The atomic force micrographs of as-deposit ed and annealed AZO films are shown in Figure 6-9. Each AFM micrograph corresponds to a specific anneal condi tion as denoted in the figure caption, while the arithmetic and rms r oughness values for each condition are listed in Table 6-4. The as-deposited AZO film exhibits and rms roughness of 2.94 nm. Annealing in argon and nitrogen ambients results in increased rms roughness values of 4.53 nm and 4.21 nm, respectively. Generally, although increased su rface roughness can lead eventually lead to reduced light trapping, the transmission of AZ O annealed in argon and nitrogen increases. However, the rms roughness of AZO annealed in forming gas decreases to 2.80 nm. Although the grain size has decreased, this film shows a large increase in conductivity and the highest FOM of any film as seen in Figure 6-6. Hydrogen and aluminum doped Figure 6-10 shows the AFM micrographs of as -deposited and annealed HAZO. Each AFM micrograph corresponds to a specific anneal condition as denoted in the figure caption, while the arithmetic and rms roughness values for each condition are listed in Table 6-4. An rms roughness (Rq) of 2.19 nm is reported for as-deposited HAZO. The results for HAZO contrast with the AZO results. Surface roughness decreases when annealed in argon nitrogen ambients,

PAGE 168

168 with measured rms roughness of 1.79 nm and 1.91 nm, respectively. Annealing in forming gas increases the rms roughness to 2.62 nm. Conclusions Transparent conductive oxides such as aluminum -doped zinc oxide are desired for use in photovoltaic devices as a result of their high transparency and excel lent electrical conductivity. These AZO films have gained significant inte rest as a low cost, abundant, and non-toxic alternative to the benchmark indium tin oxide f ilms used in many optoelectronic applications. Thermal treatments such as rapid thermal anneali ng have been identified as possible methods to increase the quality of AZO and HAZO thin films. Experimental studies show that the resist ivity of sputter deposited AZO and HAZO is significantly reduced upon rapid thermal annealing in forming gas, argon, and nitrogen ambients. Annealing HAZO in a forming gas atmosphere results in a resistivity of 9.4 x 10-4 cm. Thermal treatments also result in optical transm ission increases up to 6.0% for AZO annealed in forming gas. A figure of merit calculation correlates the resistivity and transmission of these films and provides a quantitative value to each fi lms ability to provide superior optoelectronic characteristics. In all but one case the anneal ing of AZO and HAZO results in an increase of 61% to 286% in the figure of me rit depending on the ambient used. A cross-sectional scanning electron micrograph shows that annealed HAZO e xhibits a very columnar grain structure as expected. A look at surface roughness with AFM shows that argon and nitrogen ambients tend to increase the roughness of AZO films while decreasing the roughness of HAZO films.

PAGE 169

169 Table 6-1. International Centre for Diffraction Data fingerprint for ZnO. Plane (hkl) d Spacing () 2 () Normalized Intensity (101) 2.47592 36.253 100 (100) 2.81430 31.770 57 (002) 2.60332 34.422 44 (110) 1.62472 56.603 32 (103) 1.47712 62.864 29 (004) 1.30174 72.562 2 Table 6-2. Thermal treatment results for AZO films. Forming Gas Argon Nitrogen Before After Before After Before After Average Thickness (nm) 240.0 238.3 236.7 238.3 241.7 248.3 Average Thickness -0.7% 0.7% 2.8% Low Resistivity ( cm) 2.54 0.94 3.00 1.91 5.15 2.21 Low Resistivity -63.1% -36.5% -57.0% High Resistivity ( cm) 5.98 1.04 10.76 4.03 62.11 3.49 High Resistivity -82.5% -62.5% -94.4% Average Transmission (%) 82.21 87.14 82.02 86.15 81.85 86.44 Average Transmission 6.0% 5.0% 5.6% Figure of Merit ( -1cm-1) 2012 7763 1681 3516 970 3101 Figure of Merit 286% 109% 220% Table 6-3. Thermal treatment results for HAZO films. Forming Gas Argon Nitrogen Before After Before After Before After Average Thickness (nm) 246.7 238.3 238.3 241.7 248.3 245.0 Average Thickness -3.4% 1.4% -1.3% Low Resistivity ( cm) 2.35 1.15 2.66 7.33 3.82 3.22 Low Resistivity -50.9% 175.4% -15.9% High Resistivity ( cm) 4.51 1.33 5.974 34.01 20.51 4.81 High Resistivity -70.5% 469.3% -76.6% Average Transmission (%) 83.11 87.25 83.02 84.59 83.00 87.15 Average Transmission 5.0% 1.9% 5.0% Figure of Merit ( -1cm-1) 2303 6366 2018 815 1403 2258 Figure of Merit 176% -60% 61%

PAGE 170

170 Table 6-4. Surface roughness of annealed AZO and HAZO measured by AFM. As-deposited Annealed Forming Gas Argon Nitrogen AZO R q (nm) 2.94 2.80 4.53 4.21 Ra (nm) 2.31 2.24 3.64 3.44 HAZO R q (nm) 2.19 2.62 1.79 1.91 Ra (nm) 1.75 2.00 1.46 1.48

PAGE 171

171 Ar N2 N2:H2 Power Generator Gas Outlet Gas Inlet Lamp Array Thermocouple Sample Feedback to Lamp Power Lamp Circuitry Temperature Controller Figure 6-1. Rapid thermal processing schematic.

PAGE 172

172 Figure 6-2. Minimum and maximu m resistivity of AZO films before and after annealing in forming gas, argon, and nitrogen ambients.

PAGE 173

173 Figure 6-3. Minimum and maximu m resistivity of HAZO films before and after annealing in forming gas, argon, and nitrogen ambients.

PAGE 174

174 Figure 6-4. Average transmission of AZO films before and afte r annealing in forming gas, argon, and nitrogen ambients.

PAGE 175

175 Figure 6-5. Average transmission of HAZO films before and afte r annealing in forming gas, argon, and nitrogen ambients.

PAGE 176

176 Figure 6-6. Figure of merit of AZO films before and after annealing in forming gas, argon, and nitrogen ambients.

PAGE 177

177 Figure 6-7. Figure of merit of HAZO films before and after anneal ing in forming gas, argon, and nitrogen ambients.

PAGE 178

178 Figure 6-8. Cross-sectional image of HAZO an nealed in argon ambient taken with a fieldemission scanning electron microscope.

PAGE 179

179 a) b) c) d) Figure 6-9. Atomic force micros copy of AZO films a) as-deposited, and annealed in b) forming gas, c) argon, and d) nitrogen ambients.

PAGE 180

180 a) b) c) d) Figure 6-10. Atomic force microscopy of HAZO films a) as-deposited, and annealed in b) forming gas, c) argon, an d d) nitrogen ambients.

PAGE 181

181 LIST OF REFERENCES [1] B.C. Farhar, T.C. Coburn, Colora do Homeowner Preferences on Energy and Environmental Policy (National Re newable Energy Laboratory, 1999). [2] U.S. DOE, Energy Information Agency International Energy Annual 2000 (US. Department of Energy, 2000). [3] U.S. DOE, Energy Information Agency, M onthly Energy Review (US. Department of Energy, 2002). [4] J. Mortensen, Factors Associated with P hotovoltaic System Costs (Topical Issues Brief) (National Renewable Energy Laboratory, 2001). [5] M. Thomas, H. Post, R. DeBlasio, Progress in Photovoltaics: Resear ch and Applications 7 (1999) 1-19. [6] R.F. Pierret, Semiconductor Device Fundame ntals (Addison-Wesley Publishing Company, New York, 1996). [7] W.K. Kim, Study of Reaction Pathways and Kinetics in Cu(InxGa1-x)Se2 Thin Film Growth, Ph.D. Thesis, University of Florida, 2006. [8] B. von Roedern, K. Zweibel, H.S. Ullal, Conference Record of the Thirty-first IEEE Photovoltaic Specialists C onference, 2005, pp. 183-188. [9] S.J. Jones, T. Liu, X. Deng, M. Izu, E.C.D. Inc, M.I. Troy, Conference Record of the Twenty-Eighth IEEE Photovoltaic Specialists Conference, Anchorage, AK, 2000, pp. 845848. [10] A. Shah, J. Dutta, N. Wyrsch, K. Prasad, H. Curtins, F. Finger, A. Howling, C. Hollenstein, Materials Research Societ y Symposium Proceedings, 1992, pp. 15-26. [11] B.P. Nelson, E. Iwaniczko, A.H. Mahan, Q. Wang, Y. Xu, R.S. Crandall, H.M. Branz, Thin Solid Films 395 (2001) 292-297. [12] Y. Mori, K. Yoshii, K. Yasutake, H. Kakiuc hi, S. Kiyama, H. Tarui, Y. Domoto, Journal of the Japan Society for Precisi on Engineering 65 (1999) 1600-1604. [13] B. Tell, J.L. Shay, H.M. Kasper, Physical Review B 4 (1971) 2463-2471. [14] D. Braunger, T. Durr, D. Hariskos, C. Koble, T. Walter, N. Wieser, H.W. Schock, Conference Record of the Twenty Fifth IEEE Photovoltaic Specialists Conference, Washington, DC, 1996, pp. 1001-1004. [15] K. Siemer, J. Klaer, I. Luck, J. Bruns, R. Klenk, Solar Energy Materials & Solar Cells 67 (2001) 159-166.

PAGE 182

182 [16] S. Siebentritt, Thin Solid Films 403-404 (2002) 1-8. [17] B.J. Stanbery, Conference Record of th e 31st IEEE Photovoltaic Sp ecialists Conference, Lake Buena Vista, FL, 2005, pp. 355-358. [18] H.J. Moller, Semiconductors for Solar Cells (Artech House, Inc., Norwood, MA, 1993). [19] H.W. Schock, R. Noufi, Progress in Photovoltaics Resear ch and Applications 8 (2000) 151-160. [20] M. Yamaguchi, Journal of Applied Physics 78 (1995) 1476. [21] M.A. Contreras, E. Egaas, K. Ramanathan J. Hiltner, A. Swartzlander, J. Hasoon, R. Noufi, Progress in Photovo ltaics: Research and App lications 7 (1999) 311-316. [22] H.S. Ullal, K. Zweibel, B.G. von Roedern, Conference Record of the Twenty-Eighth IEEE Photovoltaic Specialists Conferen ce, Anchorage, AK, 2000, pp. 100. [23] W. Horig, H. Neumann, B. Schumann, G. Kuhn, Physica Status Solidi 85 (1978) K57. [24] K. Yoshino, N. Mitani, T. Ikari, P. Fons S. Niki, A. Yamada, Solar Energy Materials and Solar Cells 67 (2001) 173-178. [25] C.H. Henry, Journal of A pplied Physics 51 (1980) 4494-4500. [26] M. Saad, A. Kassis, Solar Energy Ma terials and Solar Cells 77 (2003) 415-422. [27] D.L. Young, J. Keane, A. Duda, J.A. Abushama, C.L. Perkins, M. Romero, R. Noufi, Progress in Photovoltaics: Research and Applications 11 (2003) 535-541. [28] V. Nadenau, U. Rau, A. Jasenek, H.W. Schock, Journa l of Applied Physics 87 (2000) 584593. [29] A. Jasenek, U. Rau, V. Nadenau, H.W. Schock, Journa l of Applied Physics 87 (2000) 594602. [30] M.A. Contreras, K. Ramanathan, J. A bushama, F. Hasoon, D.L. Young, B. Egaas, R. Noufi, Progress in Photovo ltaics: Research and App lications 13 (2005) 209-216. [31] S. Kundu, L. Olsen, Conference Reco rd of the Twenty-Nin th IEEE Photovoltaic Specialists Conference, New Orleans, LA, 2002, pp. 648-651. [32] J.S. Ward, K. Ramanathan, F.S. Hasoon, T.J. Coutts, J. Keane, M.A. Contreras, T. Moriarty, R. Noufi, Progress in Photovoltaics: Research and Applications 10 (2002) 41-46. [33] K. Ramanathan, F.S. Hasoon, S. Smith, D.L. Young, M.A. Contreras, P.K. Johnson, A.O. Pudov, J.R. Sites, Journal of Physics and Chemistry of Solids 64 (2003) 1495-1498.

PAGE 183

183 [34] A. Bauknecht, U. Blieske, T. Kampschulte, J. Albert, H. Sehnert, M.C. Lux-Steiner, A. Klein, W. Jaegemann, Applied P hysics Letters 74 (1999) 1099-1101. [35] D. Hariskos, S. Spiering, M. Powa lla, Thin Solid Films 480-481 (2005) 99-104. [36] J. AbuShama, R. Noufi, S. Johnston, S. Ward, X. Wu, Phot ovoltaic Specialis ts Conference (2005) 299-302. [37] K.T. Ramakrishna, P. Jayarama, Journal of Physics D: Applied Physics 25 (1992) 13451348. [38] J. Song, S. Li, S. Yoon, W. Kim, J. Ki m, J. Chen, Photovoltaic Specialists Conference, 2005. Conference Record on the Thir ty-first IEEE (2005) 449-452. [39] S. Siebentritt, A. Bauknecht, A. Gehaed U. Fiedeler, T. Kampschulte, S. Schuler, W. Harneit, S. Brehme, J. Albert, M.C. Lux-St einer, Solar Energy Materials and Solar Cells 67 (2002) 344-348. [40] A. Bauknecht, U. Blieske, T. Kampschulte, A. Enaoui, V. Nadenau, H.W. Schock, A.N. Tiwari, M. Krejci, S. Duchemin, M.C. Artaud, L.M. Smith, S. Rushworth, J. Sollner, M.C. Lux-Steiner, Proceedings of 2nd World Conf erence on Photovoltaic Energy Conversion, 1998, pp. 2436-2439. [41] B.G. Lewis, D.C. Pain e, MRS Bulletin 25 (2002) 22-27. [42] T. Minami, Semiconductor Sc ience and Technology 20 (2005) 34-35. [43] G. Hodes, Chemical Solution Deposition of Semiconductor Films (Marcel Dekker, Inc., New York, 2003). [44] J.L. Vossen, W. Kern, Thin film processes II (Academic Press, Inc., Boston, 1991). [45] M.A. Contreras, M.J. Romero, R. Noufi, Thin Solid Films 511 (2006) 51. [46] R. Kaczynski, Copper Gallium Diselenide Thin Film Absorber Growth for Solar Cell Device Fabrication, Ph.D. Thesis, University of Florida, 2007. [47] W. Chen, US Patent 5,112,410 (1992). [48] R. Brundle, S. Wilson, Encyclopedia of Ma terials Characterization: Surfaces, Interfaces, Thin Films, 1992). [49] R. Jenkins, R.L. Snyder, Introduction to X-Ray Powder Diffractometry (WileyInterscience, New York, 1996). [50] H.P. Klug, L.E. Alexa nder, X-Ray Diffraction Procedures (Wiley, New York, 1974). [51] W.A. Pliskin, S.J. Zanin, Handbook of Thin Film Technology (McGraw-Hill, New York, 1970).

PAGE 184

184 [52] M. Ohring, The Materials Science of Thin Films (Academic Press, Inc., San Diego, CA, 1992). [53] J. Robertson, Philosophical Magazine Part B 63 (1991) 47-77. [54] T.V. Herak, Journal of Non Crystalline Solids 69 (1984) 39-48. [55] J.J. Lander, Physical Review 91 (1953) 1382-1387. [56] F. Wenner, A Method of Measuring Ea rth Resistivity (Govt. Print. Off., 1916). [57] D.K. Schroeder, Semiconductor material and device characteri zation (Wiley, 1990). [58] B. Jones, The Physic s Teacher 31 (1993) 48-49. [59] L.B. Valdes, Proceedings of the IRE 42 (1954) 420-427. [60] R.E. Hummel, Electron ic Properties of Materials (Springer, New York, 2001). [61] F.M. Smits, BSTJ 37 (1958) 711-718. [62] M.A. Contreras, T. Nakada, M. Hongo, A. Pudov, J. Sites, Proceedings of 3rd World Conference on Photovoltaic Energy Convers ion, Washington, DC, 2003, pp. 570-573. [63] D. Biro, R. Preu, O. Schultz, S. Peters, D. Huljic, Solar Energy Materials and Solar Cells 74 (2002) 35-41. [64] T. Yamaguchi, Y. Yamamoto, T. Tanaka, A. Yoshida, Thin Solid Films 343-344 (1999) 516-519. [65] T. Wada, Y. Hashimoto, S. Nishiwaki, T. Satoh, S.H. ldots, Solar Energy Mater. Solar Cells (2001). [66] K. Ramanathan, M. Contreras, C. Perkins, S. Asher, Progress in Photovoltaics: Research and Applications 11 (2003) 225-230. [67] C.H. Huang, Alternative Buffer Layers Device Modeling and Char acterization of CopperIndium-Diselenide-Based Thin-Film Solar Cells, Ph.D. Thesis, University of Florida, 2002. [68] X. Wu, R. Dhere, D. Albin, T. Gessert, C. Dehart, J.C. Keane, A. Duda, T.J. Coutts, A. Asher, D.H. Levi, H.R. Moutinho, Y. Yan, T. Moriarty, S. Johnston, K. Emery, P. Sheldon, Proceedings of NCPV Progr am Review Meeting (2001) 995. [69] P. Smith, Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films 10 (1992) 897-902. [70] J. Ge, Y. Li, Advanced F unctional Materials 14 (2004) 157-162.

PAGE 185

185 [71] T. Chu, S. Chu, J. Brit t, C. Ferekides, C. Wu, Journa l of Applied Physics 70 (1991) 26882693. [72] S.Y. Yin, A.L. Fahrenbruch, R.H. Bube, Journal of Applied Physics 49 (1978) 1294. [73] X. Li, W. Li, X. Dong, Japanese Jo urnal of Applied Physics 45 (2006) 9108-9110. [74] D. Okoli, A. Ekpunobi, C. Okeke, Acad emic Open Internet Journal 18 (2006) 1-12. [75] P. Roy, J. Ota, S. Srivastava, Thin Solid Films 515 (2006) 1912-1917. [76] Y. Chang, C. Munsee, G. Herman, J. Wager, P. Mugdur, D.H. Lee, C.H. Chang, Surface and Interface Analysis 37 (2005) 398-405. [77] J. Cheng, D. Fan, H. Wang, B. Liu, Y. Zhang, H. Yan, Semiconductor Science and Technology 18 (2003) 676-679. [78] D. Hariskos, M. Powalla, N. Chevaldonnet, D. Lincot, A. Schindl er, B. Dimmler, Thin Solid Films 387 (2001) 179-181. [79] A. Nayeem, G. Vajralingam, K. Yagaiah, M. Nagabhooshanam, Modern Physics Letters B 14 (2000) 733-742. [80] J. Doa, J. Herrero, Journal of Th e Electrochemical So ciety 144 (1997) 4081-4091. [81] D. Lincot, R. Borges, Journal of The Electrochemical Society 139 (1992) 1880-1889. [82] S. Mahmoud, A. Ibrahim, A. Riad, Thin Solid Films 372 (2000) 144-148. [83] L.C. Burton, T.L. Hench, A pplied Physics Letters 29 (1976) 612. [84] M. Valkonen, S. Lindroos, T. Kanniaine n, M. Leskel, Applied Surface Science 120 (1997) 58-64. [85] M. Valkonen, S. Lindroos, M. Lesk el, Applied Surface Science (1998). [86] M. Valkonen, T. Kanniainen, S. Lind roos, M. Leskel, Applied Surface Science 115 (1997) 386-392. [87] S. Ray, M. Karanjai, D. DasGupta, Thin Solid Films 322 (1998) 117-122. [88] I. Oladeji, L. Chow, J. Liu, W. C hu, A.B. ldots, Thin Solid Films (2000). [89] J. Zhou, X. Wu, G. Teeter, B. To, Y. Yan, R. Dhere, Physica Stat us Solidi (b) Applied Research 241 (2004) 775-778. [90] P. O'Brien, J. McAleese, Journa l of Materials Chemistry 8 (1998) 2309-2314.

PAGE 186

186 [91] I.O. Oladeji, L. Chow, C.S. Ferekides, V. Viswanathan, Z. Zhao, Solar Energy Materials and Solar Cells 61 (2000) 203-211. [92] Y.P. Wang, W.I. Lee, T.Y. Tseng, Applied Physics Letters 69 (1996) 1807. [93] D.C. Reynolds, D.C. Look, B. Jogai, Solid State Communica tions 99 (1996) 873. [94] D.M. Bagnall, Y.F. Chen, Z. Zhu, T. Ya o, S. Koyama, M.Y. Shen, T. Goto, Applied Physics Letters 70 (1997) 2230. [95] B. Dong, H. Hu, G. Fang, X. Zhao, D. Zheng, Y. Sun, Journal of Applied Physics 103 (2008) 073711-1-8. [96] W. Lin, R. Ma, W. Shao, B. Ka ng, Z. Wu, Rare Metals 27 (2008) 32-35. [97] S. Lin, J. Huang, P. Sajgalik, Surface & Coatings Technology 190 (2005) 39-47. [98] L.Y. Chen, W.H. Chen, J.J. Wang, F.C.N. Hong, Y.K. Su, Applied Physics Letters 85 (2004) 5628. [99] W. Gao, Z. Li, Ceramics International 30 (2004) 1155-1159. [100] E. Fu, D. Zhuang, G. Zhang, Z. ming, W. Yang, J. Liu, Microelect ronics Journal 35 (2004) 383-387. [101] C.H. Lee, K.S. Lim, J. Song, Solar Energy Materials & Solar Cells 43 (1996) 37. [102] C.H. Lee, K.S. Lim, J. Song, Japanese Journal of Applied Physics 36 (1997) 4418. [103] S.J. Baik, J.H. Jang, C.H. Lee, W.Y. C ho, K.S. Lim, Applied Physics Letters 70 (1997) 3516. [104] R.G. Jianhua, Solar Cells 30 (1991) 437. [105] M. Yoshino, W.W. Wenas, A. Yamada, M. Konagai, K. Takahashi, Japanese Journal of Applied Physics 32 (1993) 726. [106] M. Tzolov, N. Tzenov, D. Dimova-M alinovska, Thin Solid Films 379 (2000) 28-36. [107] Q. Li, D. Zhu, W. Liu, Y. Liu, X. Ma, Applied Surface Science 254 (2008) 2922-2926. [108] S. Lee, T. Lee, K. Lee, B. Cheong, Y. Kim, W. Kim, Journal of Physics D: Applied Physics 41 (2008) 1-7. [109] C.G. Van De Walle, Ph ys Rev Lett 85 (2000) 1012-1015. [110] C.G. Van de Walle, J. Neugebauer, Nature 423 (2003) 626-628. [111] P. Nunes, D. Costa, E. Fortunato R. Martins, Vacuum 64 (2002) 293-297.

PAGE 187

187 [112] M. Martinez, J. Herrero, M. Gutirrez, So lar Energy Materials and Solar Cells 45 (1997) 75-86. [113] J. Lee, K. Kang, S. Kim, K. Yoon, I. Pa rk, J. Song, Solar Energy Materials and Solar Cells 64 (2000) 185-195. [114] V. Musat, B. Teixeira, E. Fortunato, R. Monteiro, Thin Solid Films 502 (2006) 219-222. [115] J. Chang, W. Lin, M. Hon, Applied Surface Science 183 (2001) 18-25. [116] B. Oh, M. Jeong, D. Kim, W. Lee, J. Myoung, Journal of Crys tal Growth 281 (2005) 475480. [117] G. Fang, D. Li, B. Yao, Vacuum 68 (2003) 363-372. [118] V. Srikant, D.R. Clarke, Jour nal of Applied Physics 81 (1997) 6357. [119] K. Kim, K. Park, D. Ma, Journa l of Applied Physic s 81 (1997) 7764-7772. [120] J. Chang, M. Hon, Thin Solid Films 386 (2001) 79-86. [121] M. Ritala, T. Asikainen, M. Leskela, J. Skarp, Thin Films for Photovoltaic and Related Device Applications: Symposium Held April 8-11 426 (1996) 513. [122] J.A.A. Selvan, H. Keppner, A. Shah, Th in Films for Photovoltaic and Related Device Applications: Symposium Held April 8-11, 1996, San Francisco, California, USA (1996). [123] A. Gupta, A. Compaan, Applie d Physics Letters 85 (2004) 684-686. [124] M. Bertolotti, M.V. Laschena, M. Rossi, A. Ferrari, L.S. Qian, F. Quaranta, A. Valentini, Journal of Materials Re search 5 (1990) 1929-1932. [125] V. Gupta, A. Mansingh, Journa l of Applied Physics 80 (1996) 1063-1073. [126] K.S. Weissenrieder, J. Mller, Thin Solid Films 300 (1997) 30-41. [127] A.P. Roth, D.F. Williams, Journal of The Electrochemical Society 128 (1981) 2684. [128] J.L. Deschanvres, B. Bochu, J.C. J oubert, Journal de Physique 4 3 (1993) 3-485. [129] J. Lee, B. Yeo, B. Park, Thin Solid Films 457 (2004) 333-337. [130] J.H. Webb, D.F. Williams, M. Buchan an, Applied Physics Letters 39 (1981) 640. [131] L. Vasanelli, A. Valentini, A. Losacco, Solar Energy Materials 16 (1987) 91. [132] Y. Suzuki, T. Yotsuya, K. Takiguchi, M. Yoshitakey, S. Ogawa, Appl. Surf. Sci 33 (1988) 1114.

PAGE 188

188 [133] S. Fujihara, C. Sasaki, T. Kimura, Applied Surface Scienc e 180 (2001) 341-350. [134] K. Park, D. Ma, K. Kim, Thin Solid Films 305 (1997) 201-209. [135] B.H. Choi, H.B. Im, others, Thin Solid Films 193 (1990) 712-720. [136] T. Minami, H. Sato, H. Nanto, S. Takata, Japanese journal of applie d physics. Pt. 2, Letters 25 (1986) L776-L779. [137] G.A. Hirata, J. McKittrick, T. Cheeks, J.M. Siqueiros, J.A. Diaz, O. Contreras, O.A. Lopez, Thin Solid Films 288 (1996) 29-31. [138] S.A. Studenikin, N. Golego, M. Cocive ra, Journal of Applied Physics 87 (2000) 2413. [139] Y. Igasaki, H. Saito, others, Thin Solid Films 199 (1991) 223-230. [140] D. Zhang, D. Brodie, Thin Solid Films 238 (1994) 95-100. [141] F.H. Lu, H.Y. Chen, Thin Solid Films 355 (1999) 374-379. [142] S. Chu, W. Water, J. Liaw, Journal of the European Ceramic Society 23 (2003) 1593-1598. [143] S. Lin, J. Huang, P. Sajgalik, Surface & Coatings Technology 185 (2004) 254-263. [144] K. Kim, H. Tampo, J. S ong, T. Seong, S. Park, Japanese Journal of Applied Physics 44 (2005) 4776-4779. [145] P.M. Verghese, D.R. Clarke, Journal of Materials Research 14 (1999) 1039-1045.

PAGE 189

189 BIOGRAPHICAL SKETCH Andre Baran V, the son of Andre Baran IV and Candace Elizabeth Harte Baran, was born in 1980 in Deland, Florida. As an only child, he grew up primarily in Port Orange, Florida, graduating from the International Baccalaureate program at Spruce Creek High School in 1999. He chose to continue his educat ion in Atlanta, Georgia. He was honored with the prestigious Presidents Scholarship for outstanding leadersh ip and academic achievements from the Georgia Institute of Technology where he earned a Bachelor of Science in chemical engineering in 2002. In 2003 he returned to Florida and entered the Ph.D. program in the Department of Chemical Engineering at the University of Florid a in Gainesville. He pursued his research under Dr. Oscar D. Crisalle in the area of thin f ilm photovoltaics. Upon completion of his Ph.D. studies Andre will be relocating to Oregon where he will work for the Intel Corporation as a part of their Portland Technology Development center.