"Applied Science, Faculty of"@en . "Electrical and Computer Engineering, Department of"@en . "DSpace"@en . "UBCV"@en . "Lin, Derek Yun Tsung"@en . "2013-02-19T15:26:36Z"@en . "2013"@en . "Master of Applied Science - MASc"@en . "University of British Columbia"@en . "Since the development of the world\u00E2\u0080\u0099s first practical solar cell in 1954 at Bell Laboratories, researches have been conducted to increase solar cell efficiencies and lower the fabrication cost. Traditional Schottky junction solar cells suffer from the low transparency of metal films and increasing cost of indium tin oxide. In this thesis, p-type and n-type silicon Schottky junction solar cells are fabricated by integrating novel materials with silicon in an attempt to overcome these limitations. \n\tThe p-type solar cells integrate graphene and p-type silicon. Graphene is first synthesized using scotch tape exfoliation method, and then using chemical vapor deposition (CVD) of methane on copper foils to improve its quality. The CVD graphene growth system is custom built in our lab. Graphene films are optically and electrically characterized and solar cells are fabricated. Measured solar cell characteristics results are presented and reasons for the obtained parameters are discussed. Finally, methods for improving the solar cell performance are described.\n\tThe n-type solar cells are fabricated by depositing gold coated Polyacrylonitrile (PAN) nanofiber mesh on top of n-type silicon. Schottky junctions are formed where the nanofibers are in contact with silicon surface, and each junction contributes to the total current. The nanofibers are economically produced by electrospinning and coated with gold by sputtering. The solar cells are characterized and the results suggest this structure can be a promising candidate for photovoltaic application.\n\tIn addition to experimental work, we conduct numerical simulations of graphene based Schottky junction solar cells to identify possible future applications of graphene. Copper indium gallium diselenide, cadmium telluride, and amorphous silicon are chosen as the semiconductor bases because of their high absorption coefficient, high/tunable bandgap, and the possibility for economical fabrication as compared to single crystal silicon technology. The simulation is carried out using MATLAB with material properties obtained from textbooks and published literatures. The simulation results provide an estimate of the relevant photovoltaic parameters. It identifies graphene/p-type cadmium telluride as a potential Schottky junction solar cell that can achieve a conversion efficiency of 11.3%, if the graphene sheet resistance of 30 ohms/square and transmittance of 90% can be attained."@en . "https://circle.library.ubc.ca/rest/handle/2429/43933?expand=metadata"@en . "INTEGRATING GRAPHENE AND NANOFIBERS WITH SILICON TO FORM SCHOTTKY JUNCTION SOLAR CELLS by Derek Yun Tsung Lin B.A.Sc., The University of British Columbia, 2010 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE STUDIES (Electrical and Computer Engineering) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) February 2013 \u00C2\u00A9 Derek Yun Tsung Lin, 2013 ii Abstract Since the development of the world\u00E2\u0080\u0099s first practical solar cell in 1954 at Bell Laboratories, researches have been conducted to increase solar cell efficiencies and lower the fabrication cost. Traditional Schottky junction solar cells suffer from the low transparency of metal films and increasing cost of indium tin oxide. In this thesis, p-type and n-type silicon Schottky junction solar cells are fabricated by integrating novel materials with silicon in an attempt to overcome these limitations. The p-type solar cells integrate graphene and p-type silicon. Graphene is first synthesized using scotch tape exfoliation method, and then using chemical vapor deposition (CVD) of methane on copper foils to improve its quality. The CVD graphene growth system is custom built in our lab. Graphene films are optically and electrically characterized and solar cells are fabricated. Measured solar cell characteristics results are presented and reasons for the obtained parameters are discussed. Finally, methods for improving the solar cell performance are described. The n-type solar cells are fabricated by depositing gold coated Polyacrylonitrile (PAN) nanofiber mesh on top of n-type silicon. Schottky junctions are formed where the nanofibers are in contact with silicon surface, and each junction contributes to the total current. The nanofibers are economically produced by electrospinning and coated with gold by sputtering. The solar cells are characterized and the results suggest this structure can be a promising candidate for photovoltaic application. In addition to experimental work, we conduct numerical simulations of graphene based Schottky junction solar cells to identify possible future applications of graphene. Copper indium iii gallium diselenide, cadmium telluride, and amorphous silicon are chosen as the semiconductor bases because of their high absorption coefficient, high/tunable bandgap, and the possibility for economical fabrication as compared to single crystal silicon technology. The simulation is carried out using MATLAB with material properties obtained from textbooks and published literatures. The simulation results provide an estimate of the relevant photovoltaic parameters. It identifies graphene/p-type cadmium telluride as a potential Schottky junction solar cell that can achieve a conversion efficiency of 11.3%, if the graphene sheet resistance of 30 ohms/square and transmittance of 90% can be attained. iv Preface A version of chapter 2 has been published. D.Y.T. Lin, S. Shambayati, N. Mohseni Kiasari, D.L. Pulfrey and P. Servati, \"Non-idealities in graphene/p-silicon Schottky-barrier solar cells,\" MRS Proceedings 1322 (2011) p. B08-58. I conducted the majority of the device fabrication, all of the device testing, and wrote the majority of the manuscript. Ms. Shambayati has performed photolithography to create the top and bottom contacts for some devices. Mr. Mohseni Kiasari has taken scanning electron microscope images for the devices. Dr. Pulfrey and Dr. Servati have important contribution on the \u00E2\u0080\u009CExperimental Results and Discussion\u00E2\u0080\u009D section, and provide support in revising the manuscript. Part of the works done in Chapter 3 is through collaboration with Dr. Sarah Burke and Mr. Gregory McMurtrie from the Department of Physics and Astronomy, and the Department of Chemistry. Dr. Burke and Mr. McMurtrie have provided great support in setting up and calibrating the CVD graphene growth system. They have also provided the copper foils and quartz boats for the experiments. I was responsible for silicon substrate fabrication, graphene growth and transfer, optical microscope and Raman spectroscopy characterization, transmittance and sheet resistance measurements, solar cell fabrication, data analysis and writing the result into the thesis. The works in chapter 4 is performed by me, Dr. Saeid Soltanian, and Mr. Rowshan Rahmanian in our group. Dr. Saeid Soltanian has prepared the PAN nanofibers for me using electrospinning and sputtering, as well as measuring the transmittance and sheet resistance of those nanofibers. Mr. Rowshan Rahmanian has taken the SEM images. I was responsible for fabricating the silicon substrates in SFU and UBC cleanroom, transferring nanofibers to make v devices, and measuring the device I-V characteristics. I have also analyzed the data and wrote the results in the thesis. vi Table of Contents Abstract .......................................................................................................................................... ii Preface ........................................................................................................................................... iv Table of Contents ......................................................................................................................... vi List of Tables ................................................................................................................................ ix List of Figures .................................................................................................................................x Acknowledgements .................................................................................................................... xiv Dedication ................................................................................................................................... xvi Chapter 1: Introduction ...............................................................................................................1 1.1 Motivation of Schottky Junction Solar Cells .................................................................. 1 1.2 Review of Previously Reported Schottky Junction Solar Cells ...................................... 3 1.2.1 Metal Film/Semiconductor Schottky Junction Solar Cell .......................................... 3 1.2.2 ITO as a Transparent Conducting Electrode in Schottky Junction Solar Cell ............ 4 1.3 Structure of Thesis .......................................................................................................... 6 Chapter 2: Exfoliated Graphene/P-Silicon Schottky-junction Solar Cell ...............................9 2.1 Physics of P-type Schottky Junction ............................................................................. 10 2.2 Experimental Details ..................................................................................................... 12 2.3 Experimental Results and Discussion ........................................................................... 17 2.4 Chapter Summary ......................................................................................................... 20 Chapter 3: CVD Graphene/P-Silicon Schottky-junction Solar Cell ......................................22 3.1 CVD Graphene Growth on Copper ............................................................................... 24 3.1.1 CVD System Setup ................................................................................................... 24 vii 3.1.2 CVD Graphene Growth Procedure ........................................................................... 27 3.1.2.1 Principle of CVD Graphene Growth on Copper Foil ....................................... 27 3.1.2.2 Sample Preparation and Growth ....................................................................... 28 3.1.3 CVD Graphene Transfer Methods ............................................................................ 29 3.1.3.1 Transfer Using PDMS as Physical Support Layer ............................................ 30 3.1.3.2 Transfer Using PMMA as Physical Support Layer .......................................... 33 3.1.4 CVD Graphene Characterization .............................................................................. 35 3.1.4.1 Optical Microscope Characterization ............................................................... 35 3.1.4.2 Raman Spectroscopy Characterization ............................................................. 39 3.1.4.3 Transmittance and Sheet Resistance of CVD Grown Graphene ...................... 41 3.2 Solar Cell Fabrication ................................................................................................... 43 3.3 Experimental Results and Discussion ........................................................................... 46 3.4 Chapter Summary ......................................................................................................... 53 Chapter 4: Gold Coated PAN Nanofibers/N-silicon Schottky-junction Solar Cell ..............55 4.1 PAN Nanofiber ............................................................................................................. 59 4.1.1 PAN Nanofiber Synthesis and Transfer .................................................................... 59 4.1.2 PAN Nanofiber Mesh Optical and Electrical Characterization ................................ 60 4.2 Solar Cell Fabrication ................................................................................................... 63 4.3 Experimental Results and Discussion ........................................................................... 67 4.4 Chapter Summary ......................................................................................................... 75 Chapter 5: Numerical Simulation of Graphene Based Schottky Junction Solar Cells ........77 5.1 MATLAB Modeling Methodology............................................................................... 77 5.2 Material Properties Used in Simulation ........................................................................ 82 viii 5.2.1 Properties of Graphene Film ..................................................................................... 82 5.2.2 Properties of Single Crystal Silicon .......................................................................... 82 5.2.3 Properties of CIGS .................................................................................................... 83 5.2.4 Properties of CdTe .................................................................................................... 84 5.2.5 Properties of Amorphous Silicon .............................................................................. 86 5.3 Modeling Results and Discussion ................................................................................. 87 5.4 Chapter Summary ......................................................................................................... 89 Chapter 6: Conclusion ................................................................................................................90 6.1 Conclusion .................................................................................................................... 90 6.2 Possible Future Research Directions ............................................................................ 91 Bibliography .................................................................................................................................92 Appendices ..................................................................................................................................103 Appendix A Absorption Coefficient of the Semiconductors Used in Chapter 5 .................... 103 Appendix B MATLAB Code for the Simulation Used in Chapter 5 ...................................... 111 ix List of Tables Table 1. Essential PV parameters of the best graphene/semiconductor photovoltaic device ....... 19 Table 2. Essential PV parameters of the best measured solar cell, and the average values from 28 solar cells measured. ..................................................................................................................... 51 Table 3. Comparison of solar cell parameters for the our present works and related works ........ 70 Table 4. Essential PV parameters of the best solar cell and the average parameters.................... 74 Table 5. Relevant simulation parameters of p-type silicon........................................................... 83 Table 6. Relevant simulation parameters of p-type CIGS ............................................................ 84 Table 7. Relevant simulation parameters of p-type CdTe ............................................................ 85 Table 8. Relevant simulation parameters of p-type amorphous silicon ........................................ 86 Table 9. Solar cell parameters obtained from numerical simulation ............................................ 88 x List of Figures Figure 1. Structure of the metal film/semiconductor Schottky junction solar cell ......................... 4 Figure 2. Schematic illustration of the Schottky junction solar cell with ITO as the transparent electrode .......................................................................................................................................... 6 Figure 3. Energy band diagram showing the graphene/p-Si Schottky junction at equilibrium .... 11 Figure 4. Schematic illustration of solar cell fabrication process. a) As received p-type silicon wafer with 300 nm of SiO2. b) Top contacts are created by photolithography. The distance between each contact is 100 \u00C2\u00B5m. The back contact is created by thermal evaporation of aluminum. c) Graphene flakes randomly deposited onto SiO2 by scotch tape transfer method. d) Zoomed in view of one contact and one graphene flake. e) Gold contact created by laser lithography and gold deposition to connect graphene with the top contact. f) Graphene flake making contact with the underlying silicon to form the Schottky junction. ................................. 12 Figure 5. Graphene flakes on SiO2 ............................................................................................... 14 Figure 6. Gold contact connecting graphene to top contact .......................................................... 15 Figure 7. Graphene bending at the edge of a gold contact after etching away the SiO2. It makes contact with the underlying p-type silicon to form the Schottky contact. SEM setting: Voltage: 10kV, Magnification: 25k, 45\u00C2\u00B0 tilted stage. .................................................................................. 16 Figure 8. I-V characteristics, both dark and light, on a linear scale. ............................................ 17 Figure 9. I-V characteristics, both dark and light, on a semi-logarithmic scale. .......................... 18 Figure 10. a) Photo of the CVD system for growing graphene. b) Schematic illustration of CVD System ........................................................................................................................................... 26 xi Figure 11. Schematic illustration of the three stages of CVD graphene growth on copper foil: a) Copper foil with native oxide before annealing. b) Nucleation of graphene islands on the copper surface after exposure to CH4 and H2 gases at 1000 o C. c) Graphene islands grow in size to form domains that connects with neighboring domains with different lattice orientations, to form a continuous monolayer graphene. .................................................................................................. 28 Figure 12. Photographs of copper foil before and after growth of graphene ................................ 29 Figure 13. Schematic illustration of transfer process using PDMS method ................................. 32 Figure 14. Schematic illustration of transfer process using PMMA method ................................ 34 Figure 15. Optical microscope images of graphene transferred onto SiO2/Si wafer using PMMA and PDMS methods, under 5X magnification, 20X magnification, 50X magnification, and 100X magnification (From top to bottom). ............................................................................................ 36 Figure 16. Enlarged optical microscope images of graphene transferred onto SiO2/Si substrate using PMMA method, under 100X magnification. Graphene wrinkles, cracks, single layer, and bi-layer graphene regions are indicated. ....................................................................................... 38 Figure 17. Raman spectra of single layer and bi-layer graphene, corresponding to the area marked by the arrows in figure 16. The position of the Raman G band at 1585cm -1 and 2D band at 2719 cm -1 indicates the presence of graphene. The 2D-to-G band peak Raman intensity ratio of 4.07 indicates single layer graphene, and a ratio of 1.25 indicates bi-layer graphene. The very small D band at 1350cm -1 indicates high quality graphene. ......................................................... 40 Figure 18. Photo of graphene film with area of 1 cm 2 transferred onto glass slide. ..................... 42 Figure 19. The transmittance spectrum of the graphene film ....................................................... 42 xii Figure 20. I-V Measurement of square graphene film with area of 1cm 2 . The slope of the line is given by 1 0.0012 I R V \u00EF\u0080\u00AD\u00EF\u0081\u0084 \u00EF\u0080\u00BD \u00EF\u0080\u00BD \u00EF\u0081\u0084 , and the resistance of the film is 833 ohms. For a square graphene film, length = width and the sheet resistance 833 ohms. .............................................................. 43 Figure 21. a) A schematic illustration of device fabrication: transferring graphene onto substrate. b) Cross section view of solar cell. ............................................................................................... 45 Figure 22. a) SEM image showing graphene film across the Gold/SiO2 step. SEM settings: Voltage: 5kV, Magnification: 19.11kX, 45 o tilted stage. b) The corresponding optical image. .. 46 Figure 23. J-V characteristics, both dark and light, on a linear scale. .......................................... 49 Figure 24. Zoomed in J-V characteristics, both dark and light, on a linear scale. ........................ 49 Figure 25. J-V characteristics, both dark and light, on a semi-logarithmic scale. ........................ 50 Figure 26. Dark I-V characteristic in the semi logarithmic scale. Extrapolation to zero bias gives saturation current (Io) of 8.0x10 -7 A .............................................................................................. 50 Figure 27. Energy band diagram showing the gold/n-Si junction at equilibrium. ........................ 56 Figure 28. Schematic illustration of NF collection and gold coating. a) Apply glue on holder. b) Holder placed on non-woven NF electrospinned on aluminum foil. c) Holder lifted, collecting a square of NF mesh. d) Sputtered gold coated NF mesh. .............................................................. 60 Figure 29. Optical microscope image showing nanofiber meshes with different densities. ......... 62 Figure 30. Transmittance spectrum and sheet resistance of nanofiber meshes with densities corresponding to those shown in Figure 29. ................................................................................. 63 Figure 31. Transfer process of Nanofiber onto Silicon Substrate. a) Electrospinned PAN NF on aluminum foil. b) Copper wire with glue on it, placed onto the non-woven NF. c) Lift up frame to collect a 2x2 cm 2 area of NF mesh. d) Coat NF with gold by sputtering. e) Place frame on xiii patterned silicon substrate, add 1 drop of methanol on NF to adhere NF to the substrate. f) Drag down frame. NF breaks away from the frame and adhere to the silicon and gold electrode. ....... 65 Figure 32. Optical image of an as-fabricated solar cell, showing the nanofiber mesh covering the top contact and silicon. ................................................................................................................. 66 Figure 33. SEM image showing NF in ohmic contact with top Au contact, and in Schottky contact with n-silicon. SEM setting: Voltage: 15.0kV, Magnification: 1.10k. The dash line indicates the boundary at gold and silicon. The inset shows the corresponding optical image. ... 67 Figure 34. a) J-V characteristics, both dark and light, on a linear scale. b) J-V characteristics, both dark and light, on a semi-logarithmic scale. c) Dark I-V characteristic in the semi logarithmic scale. Extrapolation to zero bias gives saturation current (Io) of 2.7x10 -7 A ............. 69 Figure 35. SEM image showing cracks in the gold on nanofiber. SEM settings: Voltage: 15kV, Magnification: 20.67 kX ............................................................................................................... 71 Figure 36. Schematic of the solar cell used in simulation. a) cross section view b) top view. The effective area (overlapping area of graphene and semiconductor) is 1 cm 2 ................................. 81 Figure 37. Current Density-Voltage Characteristics under AM1.5G illumination for graphene/Si (red curve), graphene/CIGS (blue curve), graphene/CdTe (black curve), and graphene/a-Si (green curve) ................................................................................................................................. 88 xiv Acknowledgements I would like to express my gratitude to many people for their support during my master program at UBC. First and most importantly, I would like to thank Dr. Peyman Servati, my thesis supervisor. Dr. Servati has always provided guidance and support on my research in every aspect. He has taught me the methods to conduct research and collaborate with other researchers. He is always encouraging and patient when I encountered obstacles in my research. Also, he has always revised my reports and thesis and shared his thoughts with me. I also want to thank Professor Nicolas Jaeger and Professor Kenichi Takahata for taking time to be my thesis committee members. I would like to thank Dr. David Pulfrey for analyzing my research data with me and helping me to write the results and discussions of my MRS proceeding manuscript. He has also provided me great support for conducting computer simulation for solar cells. I would like to thank all my group mates in the Flexible Electronics and Energy Lab (FEEL). Mr. Nima Mohseni Kiasari has facilitated my research by providing helpful comments and sharing his research experiences with me. He has always helped me to solve problems when working in the lab and office. Also, he has offered his help to take SEM images for my samples. In addition, I would like to thank my group mates Mr. Rowshan Rahmanian and Mr. Yan Wang for SEM imaging. I would like to thank Ms. Shabnam Shambayati for her support in device fabrication and training. She has provided me training on using AutoCAD and CleWin to create patterns used for laser lithography, and fabricated some silicon samples for me as example. I xv would like to thank Ms. Saeedeh Ebrahimi Takalloo for contacting with SFU 4D labs for growing SiO2 on my silicon wafers. Especially, I would like to thank Dr. Bobak Gholamkhass and Dr. Saeid Soltanian for their generous feedbacks on my work, and training on lab equipment. Also, Dr. Saeid Soltanian has prepared the PAN nanofibers for me using electrospinning and sputtering. Without his help, my device fabrication would be much harder to achieve. In addition, I would like to thank my group mate Mrs. Mahshid Karimi, Mrs. Rubaiya Rahman, and Ms. Zenan Jiang for their great comments on my presentations and helpful discussions during the course of my master program. I wish to thank Dr. Guangrui Xia for offering me to use the Raman spectroscope in her lab, and her student Mr. Yiheng Lin for training me to use the Raman spectroscope. I would like to give my great thanks to Dr. Sarah Burke and Mr. Gregory McMurtrie for setting up and calibrating the CVD graphene growth system together with me. Many of my experiments are conducted at the SFU 4D labs and the Advanced Materials and Process Engineering Laboratory (AMPEL). I would like to thank Mr. Nathanael Sieb, Mr. Tom Cherng, Mrs. Grace Li, and Mr. Chris Balicki for their training at 4D labs. I would like to thank Dr. Mario Beaudoin and Ms. Alina Kulpa for training and wafer scribing at UBC AMPEL. Also, I would like to thank Dr. Frank Ko and his student Mr. Ritchie for borrowing their tube furnace for annealing my silicon devices. I would like to thank Natural Sciences and Engineering Research Council (NSERC) of Canada, and Canada Foundation for Innovation (CFI) for funding support. Finally, I would like to thank my parents and my brother for their ultimate support. My father, mother, and brother are always my greatest support during my master program. They have encouraged me to keep working hard and do my best. xvi Dedication To Dad, Mom, and Brother 1 Chapter 1: Introduction 1.1 Motivation of Schottky Junction Solar Cells Conventional electricity generation relies heavily on unsustainable fuels such as electro- mechanical generation that burns fossil fuels, and nuclear power generation that leaves nuclear wastes. One of the most promising sustainable alternatives of meeting the increasing demand of energy is solar energy, due to its abundance, environmental-friendliness and availability. The most popular type of solar cells remains to be crystalline silicon based p-n junction solar cells. However, the cost of today\u00E2\u0080\u0099s silicon solar cells has prohibited wide spread use of this type of solar cell. P-n junction based solar cells require many fabrication steps with the associated fabrication costs. To lower these costs, new materials and fabrication methods are required. One popular alternative to crystalline silicon solar cells is thin film solar cell. Thin film solar cells are less expensive to produce because they require less material for the active layers, has fewer processing steps, and simpler device processing and assembly. The majority of thin film solar cells are amorphous silicon (a-Si) solar cells and cadmium telluride (CdTe) solar cells. A-Si solar cells have a p-i-n junction structure instead of a p-n structure. The p and n layers are used to establish the built-in electric field, while the thicker i-layer is used to absorb light. The advantages of a-Si solar cells include high light absorption coefficient, adjustable bandgap of 1.1 to 1.9 eV, low material cost, and low deposition temperature. However, a-Si solar cells suffer from cell degradation after light illumination, known as the Staebler-Wronski effect (SWE). This effect causes an increase in recombination of electron-hole pairs which reduces the power conversion efficiency of the solar cell [1]. 2 CdTe solar cells are p-n heterojunction thin film solar cells, consisting of a few micron thick p-type CdTe and a less than a half micron thick n-type CdS [2]. In addition to low material cost, CdTe cells has a high absorption coefficient (>10 5 cm -1 ) which allows more than 90% light absorption within a few micron thick CdTe layer. However, CdTe solar cells suffer from two disadvantages. First, the global tellurium reserve is limited therefore it would be difficult to rely solely and largely on CdTe solar cells. Also, the metal cadmium used in the production of CdTe solar cells would be toxic if released, which is an environmental and safety concern of using CdTe solar cells [3]. Another alternative to p-n junction solar cell is Schottky junction solar cell. In comparison to p-n junction, Schottky junction has the advantage of material universality, low- cost, and simple fabrication. Schottky junction solar cells combine metal with n-type or p-type semiconductor as the active layers, to establish the built in electric field. The metal layer serves as both the transparent conducting electrode, and as an active layer that establishes built in field. Since only one type of semiconductor is needed, only one type of dopants is needed in the fabrication, and there is no need for a second doping using ion implantation. In addition, deposition of a metal layer on semiconductor is straightforward and economic, making device fabrication cheaper than conventional solar cells. By choosing metal and semiconductor combination that have a large work-function difference, a built in voltage can be established in the semiconductor. Unlike the p-n junction, there is no emitter layer but a metal layer instead on the top region of the solar cell. The depletion region in metal is almost non-existence, and there are no recombinations of photo-generated carriers in the metal. Recombination of photo- generated carriers in p-n junction emitter has caused p-n junction based PV cell to have a lower collectable current. 3 1.2 Review of Previously Reported Schottky Junction Solar Cells 1.2.1 Metal Film/Semiconductor Schottky Junction Solar Cell Traditionally, Schottky junction solar cell is fabricated by depositing a thin metal film on semiconductor [4, 5, 6]. The metal film has to be thin enough to allow sufficient photon to reach the semiconductor surface, and has to be thick enough to have a low sheet resistance that give high short circuit current. Anderson\u00E2\u0080\u0099s group [4] has fabricated the cell by depositing chromium film on p-type single crystal silicon. Figure 1 shows the schematic of the structure of the Schottky junction solar cell. A p-type single crystal silicon wafer with resistivity of 2 ohm-cm is used as the semiconductor substrate. A 100 nm thick aluminum film is evaporated on the bottom of wafer as ohmic bottom contact. A 4.4 nm Cr film and 5.8 nm Cu film are then deposited to form the combined metal layer. Finally, 100 nm of Al is deposited as top contact and 69 nm of silicon dioxide (SiO2) is deposited as antireflection coating. All depositions are done in vacuum (6-9 \u00C2\u00B5Torr). The chromium film on p-type silicon is used to establish the built in voltage, and the copper film (5.8 nm) on top of the chromium film is used to achieve a combined film sheet resistance that is lower than standalone chromium. The optimized combined film structure gives a sheet resistance to 50 ohms/square, and a transmittance of 55% at 600 nm [4]. Anderson\u00E2\u0080\u0099s report proposed promising Schottky junction solar cells; however, two factors limit the application of this type of solar cell. First, the optimized metal thickness by Anderson\u00E2\u0080\u0099s group is 4.4nm chromium and 5.8nm copper, and the transmittance is at most 55% at 600nm. For the metal film/semiconductor Schottky structure, the metal film needs to be thick enough to form a continuous film, which inevitably absorbs large percentage of the incident light and hence limits the energy conversion efficiency [7]. To further increase the transmittance (to above 90%) 4 and increase the efficiency, a different type of transparent film that has higher transmittance and lower sheet resistance is required. Second, all of the metal films in the metal film/semiconductor Schottky solar cell are deposited using vacuum evaporation at which induces high fabrication cost that prohibits large scale terrestrial application. Figure 1. Structure of the metal film/semiconductor Schottky junction solar cell 1.2.2 ITO as a Transparent Conducting Electrode in Schottky Junction Solar Cell To improve on the transmittance of traditional metal film/semiconductor Schottky junction solar cell, indium tin oxide (ITO) film has been used as a transparent conducting electrode in Schottky junction solar cell [8, 9]. ITO is a solid alloy of 90% indium oxide and 10% tin oxide by weight. It is a heavily doped n-type semiconductor with bandgap around 4 eV. It is commonly deposited by physical vapor deposition, electron-beam evaporation and sputtering onto substrates. In a Schottky solar cell structure, ITO does not act as the active layer. Instead, ITO is used as the window to allow light into the semiconductor base, and provides a 5 low resistive path to transfer charge carriers to external electrodes. The Schottky junction is formed between the semiconductor and the metal on the other side, and the built in electric field is established there. Figure 2 shows the schematic of a typical Schottky junction solar cell with ITO as the transparent electrode. The semiconductor layer is thin to allow sufficient light to reach the Schottky junction at the semiconductor/metal interface. Examples of these semiconductor/metal Schottky structure includes: p-type lead sulphide (PbS) nanocrystal film/aluminum, p-type PbS nanocrystal film/silver, and p-type lead selenide (PbSe) nanocrystal film/calcium [8, 9]. The ITO has two main features that make it desirable as a transparent conducting electrode. First, it has a transmittance of around 90% and sheet resistance of less than 100 ohms/square. Second, it has a large bandgap of around 4eV, making it transparent to light in the visible part of the spectrum [10]. However, ITO has several disadvantages that make it unfavorable to be used in Schottky junction solar cells. The main disadvantage of ITO is the limited supply of indium that leads to high production cost. As of September 2012, the price of 99.99% purity indium ingot per kilogram is USD$535. Second, vacuum deposition of ITO layer also contributes to high production cost. Third, the Schottky junction is formed between semiconductor and the back contact, rather than with the front contact. This forces the semiconductor layer to be made thin, and make sufficient light absorption difficult to achieve [9]. Finally, ITO is fragile and cannot be applied onto flexible substrates in the future. 6 Figure 2. Schematic illustration of the Schottky junction solar cell with ITO as the transparent electrode 1.3 Structure of Thesis To overcome the low transparency in metal films and the increasing cost of ITO, we decide to develop new type of Schottky junction solar cells that use new materials for the active layer/transparent conducting electrode, while maintaining low cost fabrication. The hypotheses of this thesis is to explore and fabricate Schottky junction solar cells by combining single crystal silicon substrate with novel materials, that could have lower fabrication cost than p-n junction cells. This thesis reports two types of Schottky junction solar cells, a p-type Schottky junction solar cell by combining graphene with p-type silicon, and a n-type Schottky junction solar cell by combining gold coated Polyacrylonitrile (PAN) nanofibers with n-type silicon. Both graphene and gold coated PAN nanofibers act as transparent conducting electrode for charge transport, and as active layer for electron-hole pair separation. For our solar cells, silicon is chosen as the 7 semiconductor substrate because of its cheap price, ease of accessibility, and high compatibility with silicon processing technology. It is possible to extend the result of our work onto thin film solar cells and flexible solar cells, to further lower the cost of cell fabrication. This thesis consists of six chapters. Chapter 1 is the introduction including motivation of Schottky junction solar cells and review of previously reported Schottky junction solar cells. Chapter 2 presents exfoliated graphene/p-type silicon Schottky junction solar cells. It includes the physics of p-type Schottky junction, fabrication of graphene by the scotch tape transfer method, characterization of the solar cell, and discussion of results. This chapter confirms the Schottky junction formation between graphene and silicon. Chapter 3 presents chemical vapor deposited graphene/p-type silicon Schottky junction solar cells. It includes graphene synthesis using the chemical vapor deposition method, optical and electrical characterization of graphene, characterization of the solar cell, and discussion of results. Chapter 3 improves on the work of chapter 2 by fabricating solar cells with larger areas and measurable efficiency. Chapter 4 presents gold coated PAN nanofibers/n-type silicon Schottky junction solar cells. It includes physics of n-type Schottky junction, PAN nanofiber synthesis method, optical and electrical characterization of PAN nanofibers, solar cell fabrication and measurement, and discussion of results. Chapter 5 presents numerical simulation of graphene based Schottky junction solar cells. In this chapter, numerical simulations of graphene/silicon, graphene/copper indium gallium diselenide, graphene/cadmium telluride, and graphene/amorphous silicon Schottky junction solar cells are performed. The simulation methodology, semiconductor properties, and the estimated solar cell performance are presented. From the simulation, we identify potential cells that could have better performance and lower fabrication cost than graphene/silicon cell, which can be 8 investigated in future. Chapter 6 is thesis conclusions that identify potential applications of the research works and possible future research directions. 9 Chapter 2: Exfoliated Graphene/P-Silicon Schottky-junction Solar Cell Schottky-junctions are potentially easier and cheaper to fabricate in comparison to p-n junction devices, so they are of interest for cost-sensitive devices such as solar cells. Graphene films are nominally appropriate as a transparent electrode material for Schottky-junction solar cells due to their high optical transparency [11], low sheet resistance [12], and tunable work function [13]. The most reported value of work function of pristine graphene is 4.4 eV [14, 15, 16, 17]. There have been reports of graphene/n-Si Schottky-junction devices [18], but here we use p-silicon as it should yield higher junction heights [19], which, in turn, should lead to lower dark currents and higher open-circuit voltages. In this chapter, Schottky junction solar cells are fabricated by integrating graphene and p-type single crystal silicon. In this chapter, I have conducted the majority of the device fabrication, all of the device testing, and wrote the majority of the experimental results and discussion. Former group member Ms. Shambayati has performed photolithography to create the top and bottom contacts for some devices. Mr. Mohseni Kiasari in our group has taken scanning electron microscope images for the devices. Dr. Pulfrey and Dr. Servati have contributed in the experimental results and discussion section. Graphene is mechanically exfoliated from the bulk of highly oriented pyrolytic graphite (HOPG). The device fabrication and measurement procedure is described, and the results are discussed. The results show s-shaped form of the illuminated I-V characteristic that lead to very low fill-factors. We speculate on the reasons for this and for the occurrence of high ideality factors in our diodes. 10 2.1 Physics of P-type Schottky Junction In this chapter, Schottky junction solar cells are fabricated by depositing graphene on top of p-type silicon. When the graphene makes contact with p-type silicon, electrons transfer from the graphene which has the lower work function to silicon which has the higher work function. The Fermi level in the p-silicon is raised by an amount equal to the difference in work function of graphene and silicon. The two Fermi levels line up to establish thermal equilibrium. The depletion region forms and the built in electric field is established in the p-silicon. Figure 3 shows the energy band diagram of the Schottky junction at equilibrium. \u00CE\u00A6Graphene is the work function of graphene, and is 4.4eV for pristine, non-doped graphene. \u00CE\u00A6p-Si is the work function of silicon, which is determined by the doping density. Work function is the difference in energy between the Fermi level and vacuum level, and is measured in electron volts (eV). \u00CE\u00A6B,p is the p-type Schottky junction barrier height. \u00CF\u0087 is the electron affinity of silicon, which is the energy difference between the conduction band edge and the local vacuum level. E0 is the force free vacuum level. El is the local vacuum level. EC, EV, and EF are the conduction band edge, valence band edge, and Fermi energy, respectively. Eg is the bandgap. Vbi is the built in voltage. WD is the depletion region width. 11 Figure 3. Energy band diagram showing the graphene/p-Si Schottky junction at equilibrium When the device is illuminated, light passes through the graphene film to reach the silicon surface and into the silicon bulk. Then, electron-hole pairs are generated in the silicon bulk, and separated and collected by the graphene/p-Si junction built in field. The graphene film acts as an active layer and as a transparent electrode, by establishing built-in field in silicon for electron/hole separation and allow electron transport. 12 2.2 Experimental Details Schottky junction solar cells are fabricated on a 3\u00E2\u0080\u009D p-type silicon wafer substrate with 300 nm of silicon dioxide (SiO2). The 300 nm thick SiO2 layer is required to see the graphene under optical microscope through light interference [20, 21]. The wafer has a resistivity range between 0.5 and 0.75 ohm*cm. Figure 4 shows a schematic illustration of the solar cell fabrication process. Figure 4. Schematic illustration of solar cell fabrication process. a) As received p-type silicon wafer with 300 nm of SiO2. b) Top contacts are created by photolithography. The distance between each contact is 100 \u00C2\u00B5m. The back contact is created by thermal 13 evaporation of aluminum. c) Graphene flakes randomly deposited onto SiO2 by scotch tape transfer method. d) Zoomed in view of one contact and one graphene flake. e) Gold contact created by laser lithography and gold deposition to connect graphene with the top contact. f) Graphene flake making contact with the underlying silicon to form the Schottky junction. After standard RCA cleaning, electrode patterns are created using photolithography on the top side of the wafer. Then, 10nm of chromium and 100nm of gold are evaporated, followed by liftoff to create the front contacts for the solar cell. By coating the top surface with a photoresist, a 10:1 buffered oxide etch is used to etch away the SiO2 on the bottom, followed by evaporation of 100nm of aluminum on the back of wafer for creation of the back contact. Graphene flakes are peeled from the bulk of HOPG and transferred to the substrate by the scotch tape transfer method [22]. This process is repeated until a few graphene flakes are identified under optical microscopy by a slight shift in color, as shown in Figure 5. P-type silicon is used as the substrate as it offers the possibility of a higher Schottky junction barrier height and, therefore, a lower dark current and a higher open-circuit voltage than n-type silicon [19]. HOPG is used because its contact causes minimal disturbance at the semiconductor surface. This is because the van der Waals force of attraction is relatively weak and the graphene sheets of the graphite are robustly impervious to diffusion of impurity atoms [23]. HOPG is also very stable in thin film configuration because the carrier mobility in HOPG can remain high even when it\u00E2\u0080\u0099s only one-atom thick (thin film) [22]. 14 Figure 5. Graphene flakes on SiO2 After identifying graphene by optical microscopy, gold contacts that connect graphene flakes to front contact are then created on the sides of the graphene by laser lithography and gold deposition. AutoCAD and CleWin software are used to design the pattern for these contacts, and a Heidelberg \u00C2\u00B5PG 101 Laser Writer is used to write the pattern onto the resist coated wafer. After laser lithography, 5nm of chromium and 100nm gold are deposited followed by liftoff to create the contacts. The continuity of the deposited contacts is verified under optical microscope, as shown in Figure 6. 15 Figure 6. Gold contact connecting graphene to top contact The SiO2 under the graphene are then etched away by immersing the wafer in a 10:1 buffered oxide etch for approximately six minutes. This allows selectively etching the oxide under graphene to be performed while oxides under gold contact and top contact remain, and as a result the graphene flake collapses onto the underlying p-type silicon and forms a Schottky junction. To verify that graphene flake has survived after etching and also to check the graphene bending at the edge of gold contact, scanning electron microscope (SEM) photomicrographs are taken using a Hitachi S4700 Field Emission, as depicted in Figure 7. 16 Figure 7. Graphene bending at the edge of a gold contact after etching away the SiO2. It makes contact with the underlying p-type silicon to form the Schottky contact. SEM setting: Voltage: 10kV, Magnification: 25k, 45\u00C2\u00B0 tilted stage. The solar cell is then characterized within a probe station by attaching a probe on top gold contact and probe to the back contact. The dark characteristics are obtained by isolating the solar cell in the dark probe station chamber at room temperature (300K). To obtain the light characteristics, the device is illuminated by a Newport solar simulator with an AM1.5 spectrum at a power density of 200mW/cm 2 as measured at the devices\u00E2\u0080\u0099 surface by a Newport optical power meter. The data from these measurements are recorded with a Keithley Model 4200-SCS Semiconductor Characterization System. 17 2.3 Experimental Results and Discussion Results for our best solar cell are presented in Figure 8 and Figure 9. Figure 8 illustrates the linear current-voltage characteristics, both in the dark and under AM1.5 illumination, signifying an s-shape device behavior under illumination. The solar cell exhibits a short-circuit current of 142.66 nA, an open-circuit voltage of 0.45 V, and a fill factor of 8.72%. The dark and light current-voltage characteristics are displayed in a logarithmic scale in Figures 9, demonstrating ON and OFF characteristics for the Schottky diode and the shift in current under illumination. The calculated generated power (product of current and voltage) as a function of voltage at the maximum power point is 5.598. Figure 8. I-V characteristics, both dark and light, on a linear scale. 18 Figure 9. I-V characteristics, both dark and light, on a semi-logarithmic scale. The photovoltaic device parameters, are extracted from light and dark characteristics using conventional diode equations [24, 25], as summarized in Table 1. The estimated area of the device is based on the size of graphene flake, however, due to the uncertainty in estimating the effective light-absorbing area of the device, the conversion efficiency cannot be quoted, but the low fill factor would indicate a low efficiency <1%. The effective light-absorbing area is larger than the size of graphene flake because the entire surface of the silicon is illuminated during measurement, rather than only illuminating the graphene. Therefore, we cannot use the graphene area to calculate current density and efficiency. The current and power obtained from this measurement are overestimates. However, a significant photoresponse is observed for this Schottky solar cell, which suggests that the graphene/p-silicon structure could be used to fabricate Schottky junction solar cell. 19 Parameter Value Input optical power density 200 mW/cm 2 Area of graphene 2.837E-11 m 2 Ideality factor (n) 3.8 Short circuit current (Isc) -142.66 nA Open circuit voltage (Voc) 0.45 V Maximum power (Pmp) 5.5984E-9 W Fill factor (FF) 8.72% Table 1. Essential PV parameters of the best graphene/semiconductor photovoltaic device The low fill-factor is a consequence of the concave-downwards illuminated I-V characteristic shown in the 4 th quadrant of Figure 5. The full I-V curve, extending into the 3 rd and 1 st quadrants, reveals an s-shaped characteristic, similar to what has been reported before for some organic [26, 27] and amorphous/crystalline silicon diodes [16]. The consensus of opinion is that such characteristics can be due to unintentional barriers that result in the blocking of one or both types of photo-carriers, thereby preventing their extraction. In our case, the reason may be a poor-quality graphene/silicon interface, which could arise from the non-optimized wet-etching step that has been used to remove the silicon dioxide and to enable the collapse of the graphene film onto the semiconductor. Interface effects may also be responsible for the high diode ideality factor reported here. It is quite likely that we have, in reality, a Metal-Insulator-Semiconductor (MIS) junction [29] because some silicon dioxide remains under the graphene after etching. Tunneling via interface states has been claimed to be responsible for diode ideality factors up to 3.8 in MIS Schottky 20 diodes. Chen\u00E2\u0080\u0099s group [30] has also reported large values of ideality factor for various layer graphene-silicon Schottky diodes ranging from 4.89 to 33.5. However, there is no obvious relationship between ideality factor and graphene layers. It is likely that a thin interfacial layer is present in our diodes, since the etchant must reach silicon dioxide under the graphene sheet that is gradually collapsing on silicon. As a result, the etchant may not be able to effectively reach sections of the oxide blocked by the collapsed graphene resulting in a non-uniform and incomplete etching. To improve on this solar cell, we identify several approaches. The size of graphene should be reproducible and scalable to larger scale, and the location of graphene deposition should be controlled. This would allow more precise definition of the effective light absorbing area, which allows short circuit current density, Schottky junction barrier height, and efficiency to be calculated. To increase the fill factor and reduce the ideality factor, we would need to improve the interface between the graphene and p-silicon. This could be done by direct placement of graphene on cleanly etched silicon surface. To achieve these, graphene should be synthesized by chemical vapor deposition on metal substrates, and a new transfer method would need to be developed and implemented. 2.4 Chapter Summary In this chapter, we present a detailed fabrication process to prepare a graphene/silicon Schottky junction solar cell. The best-fabricated solar cell is characterized and its relevant PV parameters are reported. A significant photoresponse is observed for this Schottky device, but the fill factor is low, which limits the conversion efficiency. The low fill-factor, and the high diode ideality factor, are both believed to be due to an imperfect interface between the graphene and 21 silicon wafer. Furthermore, methods for improving the interface between silicon and graphene and PV device parameters are identified for fabrication of graphene/silicon Schottky junction solar cells. 22 Chapter 3: CVD Graphene/P-Silicon Schottky-junction Solar Cell Although the scotch-tape exfoliation method provides a method to obtain graphene and allow Schottky-junction solar cells to be fabricated, there are many draw backs associated with this method. The graphene flake thickness (number of layers), size, and location are random and uncontrollable by using the scotch tape method. More importantly, because one has to first find the desired graphene flake for fabrication of solar devices, the scotch tape exfoliated highly orientated pyrolytic graphene can only be deposited on limited types of substrates, particularly silicon wafer with 300 nm thick of SiO2. In order to use graphene as a transparent conductor in our Schottky solar cell, we would need to have control over the graphene thickness (by controlling the number of layers), to control its transparency and conductance. Furthermore, it is important to obtain graphene with large and controllable area to efficiently fabricate the Schottky solar cells. The exfoliate graphene are deposited randomly on the oxidized silicon wafer, and one has to look for its location utilizing alignment marks and optical interference. This is a very inefficient way to find graphene, and can only produce small area (micron size) with non- repeatable size and location. To increase the size of graphene and therefore make Schottky-junction solar cells with larger effective area, graphene is synthesized using the chemical vapor deposition (CVD) method. Several methods have been used to attain scalable and reproducible graphene on substrates. Covalent or non-covalent exfoliations of graphite in liquids, and the conversion of SiC (0001) to graphene by means of sublimation of silicon atoms at high temperatures, are two examples [31, 32, 33, 34]. However, covalent and non-covalent exfoliation of graphite in liquids introduce electronic disorder and structural disorder in the graphene produced, [35, 36, 37] and 23 the high cost of SiC wafer prohibits large volume graphene production.[38] Among all of the methods of graphene synthesis, chemical vapor deposition of graphene onto transition metal substrates is the most inexpensive and readily accessible method. Several research groups have developed CVD graphene on Copper, Nickel, Iridium, Palladium, and Ruthenium substrates [39, 40, 41, 42]. Graphene grown on these metals need to be transferred onto other substrates for characterization and application. Within these metal substrates, etching of copper, nickel, and ruthenium can be easily achieved, whereas wet etching of Iridium is difficult because it is one of its fundamental properties. [43] Moreover, copper and nickel are comparatively cheaper than ruthenium and iridium, which favors larger volume production. Both copper and nickel can be used as catalyst to grow large area graphene films. However, graphene films grown on nickel consist of single and multilayer regions that are few tens of microns large, and are non-uniform over the entire substrate. This is because that graphene is grown on nickel by precipitation process, and not surface-catalyzed process. Nickel has a high solubility of carbon, and therefore when it is cooled down to room temperature, the diffusion of excess carbon atoms from the bulk to the surface, followed by surface segregation of these atoms, cause the formation of thick graphite films consisting of multiple graphene layers. On the contrary, copper has a low solubility of carbon and that graphene grown on polycrystalline copper is by a surface catalyzed process. Graphene film grown on copper consists of mostly single layer graphene, with less than 5% multi-layers. [44] The growth process is self- limiting because graphene stops to grow when the copper surface is fully covered with graphene and that no more copper catalyst is available. This surface catalyzed process guarantees the production of mostly single layer graphene, regardless of growth time, heating, and cooling rates. 24 Overall, copper foils are inexpensive and can be etched by common laboratory solvents, making graphene growth and transfer process attractive. In this chapter, Schottky junction solar cells are fabricated by integrating graphene and p- type single crystal silicon. CVD graphene growth and graphene transfer methods are presented. Graphene films are optically and electrically characterized and solar cells are and measured. Measurement results are presented and reasons for the obtained parameters are discussed. Finally, methods for improving the solar cell performance are given. In this chapter, the CVD graphene growth system was setup by me, Dr. Sarah Burke and Mr. Gregory McMurtrie. Dr. Burke has provided the copper foils and quartz boats for the experiments. I was responsible for silicon substrate fabrication, graphene growth and transfer, optical microscope and Raman spectroscopy characterization, transmittance and sheet resistance measurements, solar cell fabrication, data analysis and writing the result into the thesis. 3.1 CVD Graphene Growth on Copper 3.1.1 CVD System Setup The CVD system is made up of a furnace, one rotary pump, one quartz tube, two mass flow controllers (MFC), one MFC readout/controller, one pressure transducer, one pressure controller, two gas bottles, and necessary valves and gas tubes. Figure 10 shows the photo and schematic illustration of the CVD system. The two gases are hydrogen (Praxair 5.0 ultrahigh purity 100.00%) and methane (Praxair 3.7 ultrahigh purity 99.97%). Each gas is connected to a MFC, which is used to control the flow rate of the gases. The flow rate is monitored and controlled by using a CCR-400 MFC readout/controller. The MFC readout/controller can be programmed to set and maintain the flow rate of the two gases. It works by sending a voltage 25 value to the MFCs, which is then converted to the corresponding flow rate in Standard Cubic Centimetres per Minute (s.c.c.m) by the MFCs. To monitor the flow rates, the MFCs convert the detected flow rate (in s.c.c.m) into a voltage value, and send this value to the MFC readout/controller for display. The rotary pump is a Varian DS302 1HP rotary vane pump, used to pump down the CVD system pressure to a base pressure of 6mTorr. The system pressure is controlled by a MKS 253B exhaust throttle valve positioned between the quartz tube outlet and the pump. The pressure of the system is detected using a MKS 626B Barotron capacitance manometer pressure transducer. The exhaust throttle valve and the pressure transducer are both connected to a MKS Type 651C pressure controller. The pressure controller can be programmed to set the exhaust throttle valve at specified positions to achieve the desired system pressure, and monitors pressure with pressure transducer. A 3-Zone Carbolite TZF furnace is used to achieve the growth temperature of 1000 o C. The quartz tube acts as the chamber that holds the copper foil during growth. 26 Figure 10. a) Photo of the CVD system for growing graphene. b) Schematic illustration of CVD System 27 3.1.2 CVD Graphene Growth Procedure 3.1.2.1 Principle of CVD Graphene Growth on Copper Foil The graphene synthesis by chemical vapor deposition of methane on copper foils can be divided into three stages. Figure 11 illustrates the three stages of growth process. Prior to growth, the copper foil is annealed in a hydrogen reducing atmosphere at 1000 o C to facilitate growth of graphene flakes in later stages. This annealing stage removes the native oxide on copper (CuO, Cu2O) surface, increases the copper grain size, and eliminates most of the copper surface defects that might be present [45, 46]. After annealing, the methane gas is introduced and catalytically decomposed on copper to form hydrocarbon species (CxHy). The temperature, methane partial pressure, and hydrogen partial pressure are then set to achieve supersaturation of CxHy species on the copper surface, which leads to nucleation of graphene islands on the copper surface [47]. As the growth time is increased, more carbon atoms attach to the island surface and edges and the islands grow in size to form domains that connects with neighboring domains, to form a continuous monolayer graphene. After the copper surface is fully covered with graphene, further exposure to methane for longer time does not lead to formation of multilayered graphene, because the carbon species are supplied by the available copper-catalyzed decomposition of methane [47]. Therefore, CVD graphene growth on copper foil is a surface-catalyzed and self- limiting process. 28 Figure 11. Schematic illustration of the three stages of CVD graphene growth on copper foil: a) Copper foil with native oxide before annealing. b) Nucleation of graphene islands on the copper surface after exposure to CH4 and H2 gases at 1000 o C. c) Graphene islands grow in size to form domains that connects with neighboring domains with different lattice orientations, to form a continuous monolayer graphene. 3.1.2.2 Sample Preparation and Growth Using the cleaned scissors and tweezers, copper foils are cut into 1cm by 1cm squares, and placed onto the cleaned quartz boat. The quartz boat is then loaded into the quartz tube and placed in the center of the furnace. Then, the CVD system is pumped down to low pressure using the rotary pump, while monitoring the base pressure using the pressure controller. After base pressure (6 mTorr) is reached, the furnace is heated to 1000 o C with flowing 8 s.c.c.m H2 at 90 mTorr. After the furnace temperature reaches 1000 o C, the copper foils are annealed for 30 minutes under the same hydrogen flow and pressure, to increase the copper grain size and remove the copper oxide on the surface. As reported by other research groups, large grain size yields higher-quality graphene films. After annealing, the growth process is carried out by flowing 8 s.c.c.m of H2 and 24 s.c.c.m of CH4 at a system pressure of 460 mTorr for 30 minutes. When the growth is done, the furnace is turned off and allowed to cool down to room temperature, with 8 s.c.c.m. of H2 under a pressure of 90 mTorr. At room temperature (23 o C), the 29 graphene/copper/graphene stack samples are taken out and stored in plastic Petri dishes. Figure 12 shows the copper foil before and after growth of graphene. Figure 12. Photographs of copper foil before and after growth of graphene 3.1.3 CVD Graphene Transfer Methods In order to characterize the CVD grown graphene and to fabricate the Schottky solar cells, graphene must be transferred from the copper foil to other substrates. This involves using a physical support layer to hold and cover the graphene, and then etching away the underneath copper foils with copper etchant. Two methods are used to transfer the graphene from the copper foils, by transfer using spin coated and cured poly-methyl methacrylate (PMMA) solution and transfer using polydimethylsiloxane (PDMS) polymer stamp as the physical support layer. The quality of graphene transferred using these two methods are then compared, and one method is chosen to be used for transferring graphene in fabricating the Schottky solar cell. 30 3.1.3.1 Transfer Using PDMS as Physical Support Layer Since graphene is only a thin fragile layer, direct etching of the copper foil to free the graphene would break the graphene film. Therefore, we first use PDMS polymer stamp as the physical support to transfer graphene. The PDMS ingredients are the Sylgard 184 kit manufactured by Dow Corning Corporation. In a glass bottle, the elastomer base and the curing agent is combined in a 10:1 ratio, and mixed thoroughly using a pipette, to make the PDMS mixture. This step generates bubbles in the PDMS mixture. The PDMS mixture is then degassed in a desiccator under vacuum until all of the air bubbles are removed. The degassed PDMS mixture is then poured into a plastic petri dish, with a RCA cleaned 4 inch silicon wafer inside as a clean flat bottom surface. The use of a silicon wafer ensures a flat PDMS stamp surface after curing. The thickness of the PDMS stamp is made to be 3mm to ensure that it can be easily handled by tweezers in later steps. After the degassed PDMS is casted into the petri dish, it is cured for 3 hours in a 60 o C oven. After curing, the PDMS layer is carefully peeled off from the silicon wafer, and cut into 1 cm 2 square stamps by using a scalpel and tweezers. The copper etching solution used to etch away the copper foil is Marble\u00E2\u0080\u0099s reagent solution. The solution is prepared by mixing together 10g of Cupric Sulphate (CuSO4), 50mL of Hydrochloric acid, and 50mL of DI water in a glass petri dish. This solution has been used by other research groups to etch copper foils under CVD grown graphene and is safe to graphene [48, 49]. Figure 13 shows a schematic illustration of the transfer process using PDMS as physical support layer. For each 1 cm 2 graphene/copper foil stack, a 1 cm 2 PDMS stamp is used as the physical support layer. The graphene/copper foil stack is placed on a clean glass slide, the PDMS 31 stamp is placed on it, and then another glass slide is pressed on top of the PDMS to bond the stamp to the graphene. Then, the glass slides are removed using tweezers, and the PDMS/graphene/copper/graphene stack is floated PDMS side up on the etchant. Since graphene grows on both sides of the copper foil, two films are exfoliated when the copper is etched. The graphene on the bottom of copper foil that is not covered by the PDMS stamp breaks into small pieces during the etching process and deposits at the bottom of the etching solution. When all copper is etched (about 35 minutes for a 1 cm 2 , 25 micron thick copper foil), the PDMS/graphene stack is picked up by an acid resistant tweezer and floated onto a DI water bath for rinsing away the etchant residue. The stack is rinsed for 10 minutes, and then transferred to another fresh DI bath for a second rinse of 10 minutes. After rinsing, the PDMS/graphene stack is taken out and dried inside a vacuum oven overnight to remove any liquid on the graphene surface. After the graphene is dried, the PDMS/graphene stack is stamped graphene side down onto a SiO2 (300nm thick)/Si substrate and baked on a hotplate at 80 o C for 30 minutes to increase adhesion of graphene to the silicon dioxide surface [45]. Finally, the PDMS stamp is gently removed with tweezers and the graphene on silicon substrate can be characterized. 32 Figure 13. Schematic illustration of transfer process using PDMS method 33 3.1.3.2 Transfer Using PMMA as Physical Support Layer Figure 14 shows a schematic illustration of transfer process using PMMA as the physical support layer. The top surface of the graphene/copper/graphene stack is spin coated with 500nm thick of PMMA (Microchem 950 PMMA C4), under a spin speed of 3500rpm for 45 seconds, and then cured at 180 o C for 1 minute. The stack is then floated on the etchant solution. Similar to the transfer process using PDMS, two films are exfoliated during the etching process, as the graphene on the bottom of copper foil that is not covered by the PMMA layer breaks into small pieces and deposits at the bottom of the etching solution. After all copper is etched, the PMMA coated graphene is scooped up using a plastic spoon, while keeping some etchant under graphene, and transferred to a DI water bath for wash. After washing in two DI water baths, the graphene is scooped up with the desired substrate and dried in vacuum oven overnight to remove any liquid trapped in between the graphene and substrate. After the graphene on substrate is dried, the substrate is heated at 180 o C to for 30 minutes on a hot plate to soften the PMMA and allow the graphene to make better contact with the substrate. Finally, the PMMA is dissolved in a 50 o C acetone bath for 1 hour followed by IPA rinse and blown dry with Nitrogen. The heat treatment step before removing the PMMA helps to reduce the density of cracks in the transferred graphene. The copper foils used in CVD have rough surface that is further roughened due to surface reconstruction during annealing and growth [50]. When graphene grows, it follows the surface morphology of the copper foils, and when it is transferred onto other substrates it does not lie flat on the surface. After transferred onto a substrate, small gaps between graphene and the substrate forms, and cracks are introduced when the PMMA is removed. By heating the PMMA and making it more flexible, the gap between the graphene and the substrate are reduced, and the density of cracks is also reduced [50]. 34 Figure 14. Schematic illustration of transfer process using PMMA method 35 3.1.4 CVD Graphene Characterization Graphene is verified after growth and transferred onto SiO2/Si substrate, silicon substrate, and glass slides. The techniques used to characterize graphene include optical microscopy, scanning electron microscopy (SEM), Raman spectroscopy, and transmittance and sheet resistance measurements. 3.1.4.1 Optical Microscope Characterization Optical microscopy is used to macroscopically verify the presence of graphene film, the film uniformity, and to estimate the number of layers of graphene. Graphene is transferred using both PDMS and PMMA as physical support layers onto 300nm thick SiO2/Si substrates. The 300 nm thick SiO2 layer is required to see the graphene under optical microscope through light interference [51, 52]. Figure 15 shows optical microscope images, comparing graphene transferred using the two methods. For each microscope image, filters are used and contrasts are adjusted to best show the graphene and distinguish it from the background. 36 Figure 15. Optical microscope images of graphene transferred onto SiO2/Si wafer using PMMA and PDMS methods, under 5X magnification, 20X magnification, 50X magnification, and 100X magnification (From top to bottom). 37 Figure 15 clearly shows that graphene transferred by the PDMS method is much more cracked compared to graphene transferred by the PMMA method. The copper etchant used, etching time, DI water rinsing time, and drying time are kept the same for both transfer methods. Therefore, the cracks in PDMS transferred graphene are created when bonding PDMS stamps to graphene on copper, and bonding to and peeling off the PDMS/graphene stack from silicon. Graphene transferred by the PMMA method has better continuity across the entire film, as evident from the optical color uniformity. Under 5X magnification, the edge of the transferred graphene can be easily defined through the color contrast to SiO2 surface. Under magnifications of 20X, 50X, and 100X, cracks in the graphene film can be identified using the color contrast of graphene and the exposed SiO2 surface. For graphene transferred by the PMMA method, some PMMA residues are also present on the transferred graphene. We have tried removing the PMMA layer using acetone baths for 1 hour, 12 hours, and 24 hours, but have obtained similar PMMA residues. However, these residues do not affect the quality of the graphene as will be shown by Raman spectra later. Under high magnification (100X), graphene wrinkles and cracks, as well as single and bi-layer graphene films can also be seen through color contrast. Single layer graphene has the lightest color, whereas bi-layer graphene has darker colors. Figure 16 shows the enlarged, 100X magnification, optical microscope image of graphene film transferred by PMMA method. The majority of the film consists of single layer graphene, with few bi-layer graphene. Graphene wrinkles and cracks can also be seen in this figure. The wrinkles are due to the difference in thermal expansion coefficients of graphene and copper foil. [43] The thermal expansion coefficient of graphene is -7 \u00C2\u00B5m/mK and is 17 \u00C2\u00B5m/mK for copper. As a result of the difference in thermal expansion coefficients, graphene expands and copper foil contracts during the cooling down process after growth, thus creating the wrinkles. 38 The presence of wrinkles serves as a proof for continuous graphene film. The cracks in the graphene film is introduced while handling the graphene during the transfer process, but the number of cracks and the size of cracks are much smaller compared to graphene films transferred using PDMS method. Figure 16. Enlarged optical microscope images of graphene transferred onto SiO2/Si substrate using PMMA method, under 100X magnification. Graphene wrinkles, cracks, single layer, and bi-layer graphene regions are indicated. 39 3.1.4.2 Raman Spectroscopy Characterization To characterize the structural properties transferred graphene layers, Raman spectroscope is used. Raman spectroscopy has the advantage of being a non-destructive, high resolution, and fast measurement to give structural information of graphene. The Raman spectra are obtained by a high resolution confocal micro-Raman system: LabRam HR by Horiba Scientific. The Raman spectra of graphene transferred to SiO2/Si substrate is obtained using a 442 nm excitation laser with a 100X objective lens and a 2400 line/mm grating. The laser spot is a circle with diameter of 1 \u00C2\u00B5m, and the laser power at the surface of graphene is <1 mW to avoid graphene damage or laser induced heating. Raman spectroscopy is used to accurately determine the presence of graphene and multilayer graphene films, the number of graphene layers, and the defects in graphene. The position of the Raman G band at ~1585 and 2D band at ~2720 indicates the presence of graphene. The 2D-to-G band peak Raman intensity ratio ~4 indicates single layer graphene, and a ratio of ~1.2 indicates bi-layer graphene [43, 44, 53, 54, 55, 56, 57]. In addition, single layer graphene has 2D band with a Full Width Half Maximum of ~30 cm -1 [54]. The intensity of the D band located at ~1350 cm -1 is a measure of defects in the graphene [43, 55, 56]. Figure 17 shows the Raman spectra of single and bi-layer graphene regions corresponding to the area marked by the arrows in figure 16. These spectra show the D, G, and 2D bands that are representative for single layer graphene (blue plot) and bi-layer graphene (red plot). For the blue plot, the 2D band is symmetric and sharp with a Full Width Half Maximum (FWHM) of 30.8 cm -1 , indicating that single layer graphene. The peak of the 2D band is located at 2719 cm -1 , and the 2D-to-G band intensity ratio is 4.07, both also indicate that the graphene is single layer. For the blue plot, the D band is very small and under the Raman detection limit of the system, indicating high quality graphene. For the red plot, the 2D-to-G band intensity ratio is 1.25, and 40 its D band intensity is also negligible, indicating that the area is covered by high quality bi-layer graphene [40, 44, 58, 59, 60]. Each G and 2D band is fitted with a single Lorentzian to obtain the peak position, peak intensity, and FWHM [57, 61]. We analyzed the optical image over the whole graphene film (1cm 2 ) for all transferred graphene films, and the area with the lightest color constitute for more than 90% and all Raman spectra taken at these areas show monolayer graphene. Figure 17. Raman spectra of single layer and bi-layer graphene, corresponding to the area marked by the arrows in figure 16. The position of the Raman G band at 1585cm -1 and 2D band at 2719 cm -1 indicates the presence of graphene. The 2D-to-G band peak Raman intensity ratio of 4.07 indicates single layer graphene, and a ratio of 1.25 indicates bi-layer graphene. The very small D band at 1350cm -1 indicates high quality graphene. 41 3.1.4.3 Transmittance and Sheet Resistance of CVD Grown Graphene Graphene film grown by CVD with 1 cm 2 area is transferred onto a glass slide using the PMMA method. Figure 18 shows graphene film with area of 1 cm 2 transferred onto glass slide. To obtain the transmittance spectrum of the graphene film, a Cornerstone\u00E2\u0084\u00A2 130 1/8 m Monochromator is used to generate the monochromatic light with wavelengths from 400 nm to 800nm. The light is shine through the graphene film/glass substrate, and the optical power of the transmitted light at each wavelength is measured by a Newport Model 842 optical power meter. Figure 19 shows the transmittance of the graphene film as a function of wavelength of the incident light with a blank glass slide used for background subtraction. The graphene film has a transmittance of 97% at wavelength of 550nm. This value agrees with the transmittance of graphene reported by others [40, 62]. To measure the sheet resistance of graphene film transferred onto glass slides, silver paint is applied on two parallel sides of the graphene film, and wires are connected from the two sides for measurement. The measurement is done by using a Keithley Model 2400 Source Meter, by sweeping the voltage from -1V to 1V and measuring current. The inverse of the slope of the obtained current vs. voltage plot gives the resistance of the graphene film. Then, the sheet resistance is obtained by multiplying resistance by the ratio of width to length of the graphene film. Figure 20 shows the I-V characteristic for measuring the resistance of a square graphene film with area of 1cm 2 . Several graphene films are measured and the sheet resistance is between 833 ohms/square and 2166 ohms/square. The lowest sheet resistance obtained is 833 ohms/square. These values for sheet resistance and transmittance are in agreement with other reported values [43, 50]. The variation in the sheet resistance could be due to cracks in graphene introduced during the transfer process, which gives higher sheet resistance. 42 Figure 18. Photo of graphene film with area of 1 cm 2 transferred onto glass slide. Figure 19. The transmittance spectrum of the graphene film 43 Figure 20. I-V Measurement of square graphene film with area of 1cm 2 . The slope of the line is given by 1 0.0012 I R V \u00EF\u0080\u00AD\u00EF\u0081\u0084 \u00EF\u0080\u00BD \u00EF\u0080\u00BD \u00EF\u0081\u0084 , and the resistance of the film is 833 ohms. For a square graphene film, length = width and the sheet resistance 833 ohms. 3.2 Solar Cell Fabrication Schottky solar cells are fabricated on a 4\u00E2\u0080\u009D p-type silicon wafer substrate with 300nm of silicon dioxide (SiO2). The wafer has a resistivity of 3.37 ohm*cm, measured by a 4 point probe, which corresponds to an P-type doping concentration of 4.03x10 15 cm -3 . After standard RCA cleaning, electrode patterns are created using photolithography on the front side of the wafer. Then, 10nm of chromium and 100nm of gold are evaporated, followed by liftoff to create the front contacts for the solar cell. The wafer is then immersed in a 10:1 buffered oxide etch to etch away the SiO2 on the front area not covered with gold/SiO2 pattern, and all oxides on the back. To verify the complete etch of SiO2, the pattern thickness of the front contact is measured by a 44 profilometer. After etching the oxide, 300nm of aluminum are evaporated on the back of wafer to create the back contact. The wafer is then annealed at 400 o C for 15 minutes in a tube furnace under Nitrogen flow to improve the contact resistance between aluminum and silicon. Post deposition annealing of aluminum creates p-p + back surface field at the silicon-aluminum junction that reduce the carrier recombination at the back surface. After annealing, the wafer is cut into cells with dimensions shown in Figure 21a. To make the Schottky junction, a 1 cm 2 graphene film is transferred onto the cell to cover the window defined by the front contact, using the PMMA method. After drying in vacuum, graphene bends and makes contact with the exposed P-silicon to form a Schottky junction, and makes contact with the gold electrode to make an ohmic junction. Figure 21b shows the schematic illustration of transferring graphene onto the patterned substrate, to make the Schottky junction solar cell. To verify the graphene bending at the edge of the gold contact after transfer, SEM images is taken, as depicted in Figure 22a. 45 Figure 21. a) A schematic illustration of device fabrication: transferring graphene onto substrate. b) Cross section view of solar cell. 46 Figure 22. a) SEM image showing graphene film across the Gold/SiO2 step. SEM settings: Voltage: 5kV, Magnification: 19.11kX, 45 o tilted stage. b) The corresponding optical image. The solar cells are characterized with a probe station by attaching a probe on front gold electrode and a probe on back aluminum electrode. The dark characteristics are obtained by isolating the solar cells in the dark probe station chamber at room temperature (300K). To obtain the light characteristics, the solar cells are illuminated by a Newport solar simulator with an AM1.5 spectrum at a power density of 100mW/cm 2 as measured at the devices\u00E2\u0080\u0099 surface by a Newport Model 842 optical power meter. The data from these measurements are recorded with a Keithley Model 2400 Source Meter. 3.3 Experimental Results and Discussion The results for the best solar cell are presented in Figures 23 to 26. Figure 23 and Figure 24 illustrate the linear current density-voltage characteristics, both in the dark and under AM1.5 illumination. The solar cell exhibits a short-circuit current density of 1.26 mA/cm 2 , an open- 47 circuit voltage of 0.22 V, a fill factor of 36.4%, and a power conversion efficiency of 0.1%. The dark and light current density-voltage characteristics are displayed in a logarithmic scale in Figures 25, demonstrating ON and OFF characteristics for the Schottky diode and the shift in current under illumination. Figure 26 shows the semi-log scale dark I-V curve with a line fit by the equation to the linear regime of the forward characteristic. According to [24], the forward I-V characteristic (for V> 3kT/q) is represented by , and Io is the saturation current obtained by extrapolating the current from the semi-log scale dark I-V curve to V=0. The extracted saturation current (Io) and the ideality factor (n) are 8.0x10 - 7 A and 1.8, respectively. The calculated generated power at the maximum power point is 9.04 \u00C2\u00B5W. The photovoltaic device parameters are extracted from light and dark characteristics using conventional diode equations [24, 25], as summarized in Table 2. The total current of the Schottky junction, consisting of both thermionic emission and tunnelling, can be expressed as [24]: ( ( ) ) ( ( ) )( ( ) ) where A is the device area, A* is the effective Richardson constant [ 79.2 A cm-2 K-2 for p-type silicon], T is the absolute temperature in Kelvin, q is the electron charge, \u00CE\u00A6B,p is the Schottky barrier height, k is the Boltzmann constant, V is the applied voltage bias, and n is the diode ideality factor. We assume that there is little or no tunnelling current due to lightly doped silicon 48 (NA\u00E2\u0089\u00A610 17 cm -3 ) and room temperature (T=300K) measurement environment, and that the dominant current is thermionic current. The Schottky barrier height is calculated using the saturation current equation [24]: ( ) which can be rearranged into ( ) The extrapolated Io is 8.0x10 -7 A, and the area is 0.09 cm 2. From this equation, \u00CE\u00A6B,p is calculated to be 0.71eV. This value is close to the barrier height estimated by \u00CE\u00A6B,p=Eg+ \u00CF\u0087\u00E2\u0080\u0093\u00CE\u00A6Graphene = 1.12+4.05-4.4 = 0.77eV. For Schottky junction, the depletion width is [24]: \u00E2\u0088\u009A and the built in voltage is [24]: where \u00CE\u00B5s=11.9\u00CE\u00B5o is the permittivity of silicon, NV is the effective density of states in the valence band, NA is the silicon doping concentration, and \u00CE\u00B5o is the permittivity of free space. Using the calculated barrier height and the measured doping concentration is calculated to be 0.55 V, and to be 0.43\u00C2\u00B5m established in the p-silicon. 49 Figure 23. J-V characteristics, both dark and light, on a linear scale. Figure 24. Zoomed in J-V characteristics, both dark and light, on a linear scale. 50 Figure 25. J-V characteristics, both dark and light, on a semi-logarithmic scale. Figure 26. Dark I-V characteristic in the semi logarithmic scale. Extrapolation to zero bias gives saturation current (Io) of 8.0x10 -7 A 51 Parameter Best Measured Value Average Measured Value of 28 Cells Input optical power density 100 mW/cm 2 100 mW/cm 2 Estimated effective area 0.09 cm 2 0.09 cm 2 Short circuit current density (Jsc) 1.26 mA/cm 2 0.991 mA/cm 2 Open circuit voltage (Voc) 0.22 V 0.22 V Maximum power (Pmp) 9.04 \u00C2\u00B5W 8.2 \u00C2\u00B5W Fill factor (FF) 36.4% 41% Power conversion efficiency (\u00CE\u00B7) 0.1% 0.091% Table 2. Essential PV parameters of the best measured solar cell, and the average values from 28 solar cells measured. The area used in calculating the current density is estimated to be the entire silicon window surface area. This area includes overlapping areas of continuous graphene and silicon where the Schottky junctions are formed, as well as the areas of broken graphene films that are isolated from the top contact. Broken graphene are seen in optical images of transferred graphene on SiO2/Si substrates and SEM images of transferred graphene on Si substrates. These broken graphene that are isolated from the top contact form Schottky junctions and contribute to electron-hole pair generation, but do not contribute to the collected current. These broken graphene further increase the sheet resistance of the graphene film, and reduce short circuit current, because less conducting pathways are available for electrons to reach the top electrode. The high sheet resistance (between 833 ohms/sq and 2166 ohms/sq) limits the electron collection, and results in low fill factor for the fabricated cells. The size and density of cracks varies by each transferred graphene, and therefore the efficiency of each fabricated solar cells also varies. The most probable source of graphene cracks is during transfer onto substrates. When graphene is transferred onto the patterned silicon substrate, it does not immediately make contact with the silicon surface. DI water from rinsing the graphene film is trapped in between 52 the graphene and silicon interface, within the patterned window. As the device is dried in vacuum, water is gradually removed, and the graphene film gradually makes contact with the silicon surface. During this process, it is likely that some parts of the graphene cracks as the PMMA/graphene film bends and makes contact to the patterned gold/silicon step. The reason for low open circuit voltage can be due to presence of interface states. Schottky junction solar cell generally has a high surface recombination probability, which is caused by the large quantity of interface states that are formed at the metal/semiconductor interface. These interface states reduce open circuit voltage and the conversion efficiency of the solar cell [63]. The quantity of interface states can be significantly reduced by intentional deposition of an appropriate passivating layer on the semiconductor surface. [63] One example to passivate the semiconductor surface is demonstrated in the fabrication of metal-insulator- semiconductor (MIS) solar cells. In a MIS solar cell, an insulating layer is intentionally grown on the semiconductor surface before depositing the metal layer, to improve the Schottky junction height, open circuit voltage, and efficiency [64]. We believe similar ideas could be applied on graphene/silicon Schottky junction solar cells to improve Voc and efficiency. The reason for the low fill factor is due to the high sheet resistance of graphene that limits electron collection, as seen in the space charge limiting current in the semilog J-V characteristic. This space charge limiting current can be due to mobility-reducing defects in graphene such as wrinkles, cracks, and edges that act as scattering-centers, which are introduce during transfer [65]. Two approaches can be used to lower the sheet resistance of graphene and increase the fill factor. First, there is a trade-off between graphene transparency and sheet resistance that need to be optimized. Sheet resistance of graphene decreases as the number of graphene layers increase [66]. On the other hand, transparency of graphene decreases with increasing number of graphene 53 layers [67]. Therefore, to optimize the fill factor and the efficiency, the number of graphene layers used in solar cell needs to be optimized. Another approach to reduce the sheet resistance of graphene is by N-doping graphene films. Pristine graphene and unintentionally doped graphene have sheet resistance of several hundred to several thousand ohms per square. Consequently, the solar cell has high series resistance and low fill factor. A controllable, reproducible, and facile method of doping graphene is required to lower its sheet resistance and tune its work function. Some methods to N-dope graphene film have been reported [68, 69, 70, 71, 72, 73]. So far, the most effective N-doping of graphene film has been achieved by N + -ion irradiation of graphene to introduce vacancy defects in the honeycomb atomic configuration, followed by annealing in ammonia (NH3) gas at 800 o C to 1100 o C to introduce atomic N in the graphene and restore its structure [73]. However, this process is complex because it needs precise choice of the N + -ion irradiation fluence, as well as optimization of dopant atoms, dopant energy, and dopant doses in order to restore the graphene structure after annealing. The ideality factor greater than one can be due Schottky junction inhomogeneity. Schottky junction inhomogeneity may be present at the interface due to charge puddles in graphene. Charge puddles are unintentionally formed when graphene is exposed in ambient conditions. They are electron density inhomogeneity and causes electron scattering [74]. 3.4 Chapter Summary In this chapter, we have fabricated Schottky solar cells by integrating graphene and p- type single crystal silicon. The graphene is synthesized by CVD of methane on copper foils, which provides larger graphene samples that allow more control in device fabrication. PDMS and PMMA are used to transfer the grown graphene, and the quality of the transferred graphene 54 is compared. Graphene films are characterized using optical microscopy, SEM, Raman spectroscopy, transmittance measurement, and sheet resistance measurement. Solar cells are fabricated by transferring graphene films onto silicon, and are characterized. The best fabricated solar cell exhibits a short-circuit current density of 1.26 mA/cm2, an open circuit voltage of 0.22 V, a fill factor 36.4 %, and a power conversion efficiency of 0.1%. The reasons for low Voc, fill factor, and high ideality factor are discussed, and methods for improving these parameters are presented. 55 Chapter 4: Gold Coated PAN Nanofibers/N-silicon Schottky-junction Solar Cell In addition to p-type silicon based Schottky junction solar cells, we also investigate in n- type silicon based Schottky junction solar cells that could have lower fabricating cost and higher power conversion efficiency. In this project, we fabricate Schottky junction solar cells by depositing gold coated PAN nanofiber mesh on top of n-type silicon. Schottky junctions are formed where the gold coated PAN nanofibers are in contact with silicon surface, and each junction contributes to the total current. When the gold makes contact with n-type silicon, electrons transfer from silicon to the gold. The Fermi level in the n-silicon is lowered by an amount equal to the difference in work function of gold and silicon. The two Fermi levels line up to establish thermal equilibrium. The depletion region forms and the built in electric field is established in the n-silicon. Figure 27 shows the energy band diagram of the Schottky junction at equilibrium [25]. \u00CE\u00A6Gold is the work function of gold. \u00CE\u00A6n-Si is the work function of silicon, which is determined by the doping density. Work function is the difference in energy between the Fermi level and vacuum level, and is measured in electron volts (eV). \u00CE\u00A6B,n is the n-type Schottky barrier height. \u00CF\u0087 is the electron affinity of silicon, which is the energy difference between the conduction band edge and the local vacuum level. E0 is the force free vacuum level. El is the local vacuum level. EC, EV, and EF are the conduction band edge, valence band edge, and Fermi energy, respectively. Eg is the bandgap. Vbi is the built in voltage. WD is the depletion region width. 56 Figure 27. Energy band diagram showing the gold/n-Si junction at equilibrium. When the device is illuminated, light passes through the gaps in the nanofiber mesh to reach the silicon surface. Electron-hole pairs are then generated in the silicon bulk, and separated and collected by the nanofiber/n-Si junction built in field. Here, the nanofiber mesh serves as a transparent window for light transmission, and also as an active layer for electron/hole separation and hole transport. The nanofiber mesh has several advantages compared to the graphene film we fabricated. First, a continuous single-layer graphene film with uniform thickness, crack-less surface is difficult to synthesize, especially in large area as we found out in our CVD grown graphene. Other groups have also synthesized larger area graphene films by chemical vapor deposition [18, 44, 47, 49]. However, it is often challenging to transfer CVD graphene film to other substrates 57 without suffering from cracks in the film, as single layer graphene is very fragile to handle during the transfer process. The cracks significantly increase the sheet resistance of graphene, and results in solar cells with low short circuit current density, low fill factor and, therefore low efficiency. In addition, graphene synthesis by CVD requires high temperature and vacuum environment, which can be expensive and not economically accessible to all research groups. Although graphene can be economically synthesized by mechanical cleavage, the device fabricated has limited size which could be challenging to implement in large area. Chen\u00E2\u0080\u0099s group has fabricated Schottky diodes by depositing mechanically exfoliated graphene on top of silicon substrates, and has yielded a device area of only 92 um 2 [30]. In our first project, we have also fabricated Schottky solar cells with small area. In comparison, our nanofiber mesh can be synthesized in arbitrary sizes and the cost of synthesis is much lower. Our nanofibers are produced by conventional electro spinning, which doesn\u00E2\u0080\u0099t require high temperature, vacuum environment, and expensive equipment compared to processes that requires a CVD furnace. The nanofibers are coated with gold by sputtering which contributes to a lower cost fabrication compared to gold evaporation. The size of our nanofiber mesh is limited only by the size of the square holder used in transfer, as will be discussed in the experimental details section. The nanofiber mesh is also easier to transfer onto substrates due to its thicker structure (200-400nm in diameter, compared with 1\u00C3\u0085 thick single layer graphene). Furthermore, there is no need to cover the nanofiber mesh with a physical support layer and having to remove it afterwards, which makes nanofiber characterization and device fabrication simpler. These advantages allow solar cells with large area to be more easily and more economically fabricated as compared to graphene based cells and other cells requiring CVD. 58 There have been reports on Schottky junction solar cells using gold (Au) as the metal layer. Yu\u00E2\u0080\u0099s group used Au/graphene stack as the metal layer with single n-CdS nanowire as the substrate, producing an energy conversion efficiency of up to 1.65% [75]. However, the fabrication of both graphene and n-CdS nanowire requires the use of CVD furnace, and the solar cell size is relatively small (10\u00C2\u00B5m*100nm) compared to our size which may limit its application in larger area. Li\u00E2\u0080\u0099s group used Au as the metal layer with ZrS2 nanobelt networks as the substrate, and achieved a conversion efficiency as high as 1.2% [76]. However, their device suffers from a low open circuit voltage (0.19V) and small area (0.01 cm 2 ). For our solar cells, silicon is chosen as the substrate because of its cheap price, high abundance, and high compatibility with silicon processing technology. We have noticed that some researchers have used silicon-nanowire-array and patterned silicon-pillar-array as the substrate to enhance light harvesting by suppressing light reflection [77, 78, 79, 80]. In this work, we choose to use planar silicon in an attempt to maximize the number of Schottky junctions formed between nanofiber and silicon, to generate more current, and to achieve higher conversion efficiency. In this chapter, Schottky junction solar cells are fabricated by integrating gold coated PAN nanofiber mesh and n-type single crystal silicon. PAN nanofiber mesh synthesis procedure, transfer procedure, and solar cell assembly procedure are reported. The nanofibers are optically and electrically characterized and the solar cells are characterized in dark and under standard AM1.5 illumination to evaluate their performance. Results are then discussed and compared with our CVD graphene/p-silicon Schottky junction solar cells and reported results for other Schottky solar cells. Finally, methods for improving the solar cell performance are given. The works in this chapter is performed by me, Dr. Saeid Soltanian, and Mr. Rowshan Rahmanian in our group. Dr. Saeid Soltanian has prepared the PAN nanofibers for me using electrospinning and 59 sputtering, as well as measuring the transmittance and sheet resistance of those nanofibers. Mr. Rowshan Rahmanian has taken the SEM images. I was responsible for fabricating the silicon substrates in SFU and UBC cleanroom, transferring nanofibers to make devices, and measuring the device I-V characteristics. I have also analyzed the data and wrote the results in the thesis. 4.1 PAN Nanofiber 4.1.1 PAN Nanofiber Synthesis and Transfer Polyacrylonitrile (PAN) nanofibers were synthesized using the conventional electrospinning method. Polyacrylonitrile (average molecular weight of 100000 g/mol) was dissolved in Dimethylformamide with a concentration of 10 wt% and stirred at 60 oC for 24 hours to form a homogenous solution. It is then loaded into a plastic syringe with a blunted G18 needle. A nanofiber electrospinning unit by KATO TECH CO., LTD. was used to prepare the nanofibrous samples from the solution. The solution was delivered to the needle at a constant flow rate of 0.5 ml/h while applying a potential of 17 kV to the needle. The PAN nanofibers are then electrospinned onto aluminum foil collector. The non-woven fibers are collected on the surface of a 2x2 cm2 square holder to form a square area of nanofiber (NF) mesh as schematically shown in Figure 28a to 28c. The holder is made by a copper wire bend into a square shape. Glue is then applied on the holder, and the holder is placed onto the non-woven NF and left on for 15 minutes. The holder is subsequently lifted along with a square area of NF mesh, and the NF mesh can be thinned by using a metal tweezer to desired density. The NF meshes are then coated with 40nm of gold by sputtering with an Edward S150A sputter coater, which provides a uniform conformal coating on the surface of nanofibers as schematically shown in Figure 28d. We chose sputter coating instead of 60 evaporating gold to have uniform coating on all PAN NFs, as evaporation coating only partially coats the NFs and the conductivity of the coated NFs is low. After sputter coating, the NF meshes are optically and electrically characterized. Figure 28. Schematic illustration of NF collection and gold coating. a) Apply glue on holder. b) Holder placed on non-woven NF electrospinned on aluminum foil. c) Holder lifted, collecting a square of NF mesh. d) Sputtered gold coated NF mesh. 4.1.2 PAN Nanofiber Mesh Optical and Electrical Characterization The morphologies of NF meshes with different fiber densities are studied using confocal 61 microscope, as shown in Figure 29. Figure 30 shows the transmittance spectra and sheet resistance of the NFs corresponding to those shown in Figure 29a to 29e. The transmittance spectra are measured by using a Cornerstone\u00E2\u0084\u00A2 130 1/8 m Monochromator and a Newport Model 842 optical power meter. To measure the transmittance spectra, the NF meshes collected on the square holders are placed between the output of the Monochromator and the optical power meter. The transmittance of air is used as the background. The sheet resistance of the NF meshes with densities corresponding to those shown in Figure 29, are measured by the four point probe Van Der Pauw method to avoid contact resistance. The NF meshes are transferred onto clean glass slides for sheet resistance measurements. The transfer process is identical to the process used to transfer NF meshes onto silicon surface, as shown in Figure 31a to Figure 31e, with glass slides being the substrates. The measurements were carried out using a Keithley 2400A source meter and a Tektronix DMM4040 Digital Precision Multimeter. 62 Figure 29. Optical microscope image showing nanofiber meshes with different densities. 63 Figure 30. Transmittance spectrum and sheet resistance of nanofiber meshes with densities corresponding to those shown in Figure 29. 4.2 Solar Cell Fabrication Schottky solar cells are fabricated on a 4\u00E2\u0080\u009D n-type silicon wafer substrate with 300nm of silicon dioxide (SiO2) on both sides. The wafer has a resistivity of 1.54 ohm*cm, measured by a MICROWORLD Model 280 4-point probe meter, which corresponds to an N-type doping concentration of 3.04x10 15 cm -3 . After standard RCA cleaning, electrode patterns are created using photolithography on the front side of the wafer. Then, 10nm of chromium and 100nm of gold are evaporated, followed by liftoff to create the front contacts for the solar cell. The wafer is then immersed in a 10:1 buffered oxide etch to etch away the SiO2 on the front area not covered with gold/SiO2 pattern, and all oxides on the back. To verify the complete etch of SiO2, the pattern thickness of the front contact is measured by a profilometer. After etching the oxide, 64 10nm of chromium and 300nm of aluminum are evaporated on the back of wafer to create the back contact. Finally, the wafer is cut into 1.5 cm x 1.5 cm square cells. To make the Schottky junction, a 2x2 cm 2 square copper wire holder is used to transfer NF mesh onto the top of the cell. To deposit the NF mesh onto the substrate, one drop of methanol is dropped onto the center of the NF mesh, while holding it close to the cell\u00E2\u0080\u0099s surface (Figure 31e). The wet nanofibers break away from the holder\u00E2\u0080\u0099s edge, and makes contact with the gold front contact and silicon surface (Figure 31f). After the methanol vaporizes, the nanofiber mesh makes ohmic contact with the gold front contact and Schottky contact with the n-silicon surface. Figure 32 shows optical images of an as fabricated solar cell. Figure 33 shows an SEM image of the as-transferred NF mesh onto the patterned silicon substrate. It can be seen that the NF mesh makes contact with gold and n-silicon, and is continuous across the patterned step. It should be noted that the size of the NF mesh can be conveniently scaled up or down by changing the size of the square holder, making them applicable to different sizes of solar cell. 65 Figure 31. Transfer process of Nanofiber onto Silicon Substrate. a) Electrospinned PAN NF on aluminum foil. b) Copper wire with glue on it, placed onto the non-woven NF. c) Lift up frame to collect a 2x2 cm 2 area of NF mesh. d) Coat NF with gold by sputtering. e) Place frame on patterned silicon substrate, add 1 drop of methanol on NF to adhere NF to the substrate. f) Drag down frame. NF breaks away from the frame and adhere to the silicon and gold electrode. 66 Figure 32. Optical image of an as-fabricated solar cell, showing the nanofiber mesh covering the top contact and silicon. 67 Figure 33. SEM image showing NF in ohmic contact with top Au contact, and in Schottky contact with n-silicon. SEM setting: Voltage: 15.0kV, Magnification: 1.10k. The dash line indicates the boundary at gold and silicon. The inset shows the corresponding optical image. 4.3 Experimental Results and Discussion The solar cells are characterized with a probe station by attaching a probe on front gold electrode and a probe on back aluminum electrode. The dark characteristics are obtained by isolating the solar cells in the dark probe station chamber at room temperature (300K). To obtain the light characteristics, the solar cells are illuminated by a Newport solar simulator with an AM1.5 spectrum at a power density of 100mW/cm2 as measured at the devices\u00E2\u0080\u0099 surface by a 68 Newport Model 842 optical power meter. The data from these measurements are recorded with a Keithley Model 2400 Source Meter. The results for the best solar cell are presented in Figures 34a to 34c. Figure 34a illustrates the linear current density-voltage (J-V) characteristics, both in the dark and light (under AM1.5 illumination). The solar cell exhibits a short-circuit current density of 11.72 mA/cm2, an open-circuit voltage of 0.54 V, a fill factor of 28%, and a power conversion efficiency of 1.78%. These values are higher than our graphene/silicon based Schottky solar cells, and comparable to the values measured from gold layer based Schottky solar cells reported by other groups [75, 76]. Table 3 summarizes the relevant solar cell values reported by other groups and by our group. 69 Figure 34. a) J-V characteristics, both dark and light, on a linear scale. b) J-V characteristics, both dark and light, on a semi-logarithmic scale. c) Dark I-V characteristic 70 in the semi logarithmic scale. Extrapolation to zero bias gives saturation current (Io) of 2.7x10 -7 A Table 3. Comparison of solar cell parameters for the our present works and related works The Voc is higher than our graphene/p-silicon solar cell due to a higher Schottky junction height. The Voc is higher than those reported for Au/CdS nanobelt solar cell and Au/ZrS2 nanobelt solar cells due to a larger difference between the work function of metal and semiconductor. The Jsc is higher than our graphene/p-silicon and other group\u00E2\u0080\u0099s graphene/n-CdSe solar cells, due to the lower sheet resistance of the nanofiber mesh compared to the graphene films. The low fill factor of our solar devices is due to the degrading contacts between gold and silicon surface. During the cell fabrication, one drop of methanol is added on the nanofiber mesh Schottky Junction Materials Jsc Voc FF \u00CE\u00B7 Device Area Reference [mA/cm2] [V] [%] [%] [cm2] Mechanically exfoliated graphene/n-silicon 0.0015 mA 0.25 Not reported Not reported Not reported 30 Schottky junction solar cell CVD Graphene/n-Silicon 0.003 0.25 19 0.04 Not reported 80 Schottky junction solar cell CVD Graphene/n-CdSe single Nanobelt 5.75 0.51 42.7 1.25 1.99E-05 81 Schottky junction solar cell CVD Graphene/Au/n-CdSe single Nanowire 27.5 0.15 40 1.65 1.00E-08 75 Schottky junction solar cell Au/ZrS2 Nanobelt networks 21.8 0.19 30 1.2 0.01 76 Schottky junction solar cell CVD Graphene/p-Silicon 1.26 0.22 36.4 0.1 0.09 Present Work Schottky junction solar cell Au coated PAN Nanofibers/n-Silicon 11.72 0.543 28 1.78 0.64 Present Work Schottky junction solar cell 71 to adhere it to the silicon surface. The solar cell measurements are performed right after the methanol vaporizes. However, we noticed that there are cracks in the gold after methanol vaporizes, as shown in SEM image in Figure 35. These cracks reduce the current conductivity of the NFs as there are fewer pathways for collecting charge carriers. Therefore the short circuit current and the fill factor reduce. Figure 35. SEM image showing cracks in the gold on nanofiber. SEM settings: Voltage: 15kV, Magnification: 20.67 kX The short circuit current densities of the fabricated solar cells vary because of the variation in the number of Schottky junctions formed on each cell, which is due to the variation 72 in the density of NFs. Table 4 summarizes the values for the solar cell parameters of our best solar cell and the average values from all fabricated cells. It should be noted that the nanofiber meshes themselves are not transparent, and light passes through the gaps between nanofibers to reach and be absorbed by silicon. Electron-hole pairs are then generated in the silicon bulk, and separated and collected by the gold/n-Si junction built in field. There is a trade-off between the transparency and sheet resistance of nanofiber meshes, and a trade-off between the transparency and number of Schottky contacts formed. More transparent nanofiber meshes allow more photons to be absorbed by the silicon, but form fewer number of Schottky contacts with silicon. Conversely, less transparent nanofibers allow less photons to be absorbed by the silicon, but form greater number of Schottky contacts with silicon. For the best solar cell, a nanofiber mesh with transparency of 65% and sheet resistance of 18 ohm/sq is used. We are currently working on optimizing the efficiency of the solar cells through choosing the nanofibers with the best combination of transparency, sheet resistance, and number of Schottky contacts formed. We chose to use PAN NFs because of its insolubility in water and high mechanical strength. The low solubility of PAN NFs in water makes them stable in air atmosphere and allows device fabrication without vacuum. The high mechanical strength of PAN NFs is important when handling the NFs. Collecting non-woven NFs, transporting NFs in and out of sputter chamber, and placing them onto substrates with methanol all induce high stress on the NFs. The high mechanical strength makes it less likely to break and make device fabrication easier. The dark and light current density-voltage characteristics are displayed in a logarithmic scale in Figures 8b), demonstrating ON and OFF characteristics for the Schottky diode and the shift in current under illumination. Figure 8c) shows the semi-log scale dark I-V curve with a line 73 fit by the equation to the linear regime of the forward characteristic. The extrapolated saturation current and the ideality factor are 2.7x10-7A and 2.5 respectively. The calculated generated power gives a maximum power of 1.10 mW. The photovoltaic device parameters are extracted from light and dark characteristics using conventional diode equations [24, 25], as summarized in Table II. The area used in calculating the current density is estimated to be the entire silicon window surface area. This area includes overlapping areas of the nanofibers and silicon where the Schottky junctions are formed, and the exposed areas where light absorbing occurs. The total current of the Schottky junction, consisting of both thermionic emission and tunneling, can be expressed as: ( ( ) ) ( ( n ) )( ( ) ) where A is the device area, A* is the effective Richardson constant [ 252 A cm-2 K-2 for n-type silicon], \u00CE\u00A6B,n is the n-type Schottky barrier height, k is the Boltzmann constant. We assume that there is little or no tunneling current due to lightly doped silicon (ND\u00E2\u0089\u00A610 17cm-3) and room temperature (T=300K) measurement environment, and that the dominant current is thermionic current. The Schottky barrier height is calculated using the equation: ( ) The extrapolated Io is 2.7x10 -7 A, and the area is 0.64 cm2. From this equation, \u00CE\u00A6B is calculated to be 0.82eV. This value is similar to the highest Au/Si Schottky barrier height 74 reported and tabulated in [18]. For Schottky junction, the depletion width is: \u00E2\u0088\u009A D and the built in voltage is: \u00F0\u009D\u0090\u00B6 where \u00CE\u00B5s=11.9\u00CE\u00B5o is the permittivity of silicon, NC is the effective density of states in the conduction band, ND is the silicon doping concentration, and \u00CE\u00B5o is the permittivity of free space. Using the calculated barrier height and the measured doping concentration, is calculated to be 0.58 V, and to be 0.50 \u00C2\u00B5m established in the n-silicon. Parameter Value Average value from 18 solar cells Input optical power density 100 mW/cm2 100 mW/cm2 Estimated effective area 0.64 cm2 0.64 cm2 Short circuit current density (Jsc) 11.72 mA/cm 2 6.98 mA/cm2 Open circuit voltage (Voc) 0.54 V 0.52 Maximum power density (Pmp) 1.78 mW/cm 2 1.23 mW/cm2 Fill factor (FF) 28% 34% Power conversion efficiency (\u00CE\u00B7) 1.78% 1.23% Table 4. Essential PV parameters of the best solar cell and the average parameters. Comparing sheet resistance, the resistance of our gold coated NF mesh is better than the sheet resistances of graphene films grown by us and those reported by others as transparent electrode/active layer of solar cells. The lowest sheet resistance achieved by our nanofiber mesh that has a 90% transmittance at 550nm, is only 90 ohm/sq. This value is lower compared to the reported value of pristine CVD graphene film between 350 ohm/sq for four layer graphene to 2100 ohm/sq for single layer graphene [43]. This resistance is also superior than the lowest sheet 75 resistance of our graphene film (833 ohm/sq), and can be a promising transparent conducting electrode for other applications. 4.4 Chapter Summary In this chapter, we have fabricated Schottky junction solar cell by integrating gold coated PAN nanofiber mesh and single crystal silicon. The nanofiber mesh is examined by optical microscope, SEM, and confocal microscope. Nanofiber meshes with different transparencies and sheet resistances are also synthesized and compared. The best fabricated solar cell exhibits a high short-circuit current density of 11.72 mA/cm2, an open circuit voltage of 0.543V, a fill factor 28%, and a power conversion efficiency of 1.78%. The solar cell efficiency could be improved by using nanofiber meshes with optimized combination of sheet resistance and transparency, as well as patterning the silicon surface to reduce light reflection. The fill factor could be improved by improving the contact between nanofiber and silicon. Our experimental results suggest that the gold/Silicon Schottky junction structure can be a promising candidate for photovoltaic application. 76 77 Chapter 5: Numerical Simulation of Graphene Based Schottky Junction Solar Cells To explore the potential application of graphene in Schottky junction solar cells, we conduct numerical simulations of several graphene based Schottky junction solar cells. In particular, we have chosen single crystal silicon, copper indium gallium diselenide (CIGS), cadmium telluride (CdTe), and amorphous silicon (a-Si) as the semiconductor bases. The simulation is carried out using MATLAB with semiconductor material properties obtained from textbooks and published literatures as modeling parameters. The simulation gives an estimate of the performance of solar cells, with relevant PV parameters including short circuit current density, open circuit voltage, fill factor, and conversion efficiency. J-V characteristics of the four solar cells are produced and discussed. From these simulations, we identify possible combinations of graphene and semiconductor to form economical Schottky junction solar cells, providing graphene application in future. I was responsible for completing this chapter, including MATLAB simulation, data analysis, and writing of the results into the thesis. 5.1 MATLAB Modeling Methodology The global, AM1.5G solar spectral irradiance data is obtained from American Society for Testing and Material [82]. The spectral irradiance (S) at each wavelength is converted into the photon flux by dividing by the photon energy at that wavelength. The photon energy at each wave length is given by: photon hc E \u00EF\u0081\u00AC \u00EF\u0080\u00BD 78 where h is the Plank's constant, c is the speed of light in vacuum. The photon flux (\u00CE\u00A60) is then given by: \u00EF\u0080\u00A8 \u00EF\u0080\u00A90 photon S E \u00EF\u0081\u00AC\u00EF\u0081\u0086 \u00EF\u0080\u00BD with the units of photons/m 2 /second. The range of wavelength used in this simulation is from 280nm to 1200nm, with 10nm increments. To calculate the total photocurrent, we first calculate the photocurrent generated in the depletion region and the base. The photocurrent is the current resulting from the absorption of photons. Photons absorbed create electron-hole pairs, and are separated by the built in field and made to flow in opposite directions. From [25], the base photocurrent density is given by: - - ' - ' 0 2 2 (cosh - ) sinh ( ) - -1 sinh cosh W B B g eB e e e e e e e e e e qT L e H Q e Q L e J L L H Q Q \u00EF\u0081\u00A1 \u00EF\u0081\u00A1 \u00EF\u0081\u00A1\u00EF\u0081\u00A1 \u00EF\u0081\u00A1 \u00EF\u0081\u00AC \u00EF\u0081\u00A1 \u00EF\u0081\u00A1 \u00EF\u0081\u0086 \u00EF\u0083\u00A9 \u00EF\u0083\u00B9\u00EF\u0080\u00AB \u00EF\u0080\u00AB \u00EF\u0080\u00BD \u00EF\u0083\u00AA \u00EF\u0083\u00BA \u00EF\u0080\u00AB\u00EF\u0083\u00AB \u00EF\u0083\u00BB where Tg is the transmittance of graphene, \u00CE\u00A60 is the photon flux at AM1.5G, \u00CE\u00B1(\u00CE\u00BB) is the absorption coefficient of the semiconductor, as function of wavelength. Appendix A tabulates the absorption coefficient of all four semiconductors, at corresponding wavelengths. \u00CE\u00BB is the wavelength. The range of wavelength used in the simulation is from 280nm to 1200nm Le is the electron diffusion length, W is the width of the depletion region, B is the thickness of the semiconductor, 'B B W\u00EF\u0080\u00BD \u00EF\u0080\u00AD , 79 B e e e S L H D \u00EF\u0080\u00BD , and ' e e B Q L \u00EF\u0080\u00BD . The depletion region photocurrent density is given by [25]: - 0( ) [1- ] D W e gJ qT e \u00EF\u0081\u00A1\u00EF\u0081\u00AC \u00EF\u0080\u00BD \u00EF\u0081\u0086 The total photocurrent density integrated over the wavelengths is given by [25]: ( ) ( )B Dph e eJ J J d\u00EF\u0081\u00AC \u00EF\u0081\u00AC \u00EF\u0081\u00AC\u00EF\u0080\u00BD \u00EF\u0080\u00AB\u00EF\u0083\u00B2 The Schottky junction dark current is given by [25]: / 0[ 1] L thV V DarkJ J e\u00EF\u0080\u00BD \u00EF\u0080\u00AD where VL is the load voltage. J0 is the dark saturation current density and is given by [25]: , / 0, , B p thV Dark V R holeJ qN e v \u00EF\u0080\u00AD\u00EF\u0081\u0086 \u00EF\u0080\u00BD where NV is the effective density of states in the valence band, vR,hole is the mean unidirectional velocity of holes, and \u00CE\u00A6B,p is the Schottky barrier height. The Schottky barrier height is obtained by \u00CE\u00A6B,p=Eg+ \u00CF\u0087\u00E2\u0080\u0093\u00CE\u00A6Graphene where Eg and \u00CF\u0087 are the semiconductor bandgap and electron affinity, and \u00CE\u00A6Graphene is the work function of graphene. The load current density at the output of the solar cell is the superposition of the photocurrent density and dark current density, Load ph DarkJ J J\u00EF\u0080\u00BD \u00EF\u0080\u00AD The open circuit voltage of the solar cell is given by [25]: 0 0 ln ph oc th J J V V J \u00EF\u0080\u00AB\u00EF\u0083\u00A6 \u00EF\u0083\u00B6 \u00EF\u0080\u00BD \u00EF\u0083\u00A7 \u00EF\u0083\u00B7 \u00EF\u0083\u00A8 \u00EF\u0083\u00B8 80 The fill factor is given by [25]: mp mp mp sc oc sc oc J V P FF J V J V \u00EF\u0080\u00BD \u00EF\u0080\u00BD Where Jmp, Vmp, Pmp, are the current density, voltage, and power at the maximum power point. Jsc is the short circuit current density. The photovoltaic conversion efficiency is given by: mp in P P \u00EF\u0081\u00A8 \u00EF\u0080\u00BD where Pin is the input optical power density. The effective area of the solar cell is set to be 1 cm by 1cm. Figure 36 shows the schematic of the solar cell used in this simulation. Resistance-less front contacts along two opposite sides of graphene, and resistance-less back contact on the semiconductor are used. 81 Figure 36. Schematic of the solar cell used in simulation. a) cross section view b) top view. The effective area (overlapping area of graphene and semiconductor) is 1 cm 2 82 5.2 Material Properties Used in Simulation 5.2.1 Properties of Graphene Film In this simulation, we choose the value for graphene sheet resistance and transmittance to estimate the highest achievable efficiency. HNO3 doped graphene film with 90% transmittance at 550nm and 30 ohms/square is used. This corresponds to the lowest reported sheet resistance to- date for four-layer doped graphene film [62]. The work function is 4.4 eV. In the simulation the optical transmittance of graphene is taken to be 90% across the entire AM1.5G spectrum, to show the estimated performance of the cells. 5.2.2 Properties of Single Crystal Silicon A 4-inch single crystal p-type silicon is used. Since the electrons (minority carriers) generated in the semiconductor bulk (base) have to travel to the depletion region to be collected, we choose the semiconductor to be p-type doped. The electrons (minority carrier) have higher mobility and diffusivity compared to holes for common semiconductors. The doping density is chosen to be 1x10 16 cm -3 to avoid tunneling current. The properties of p-type single crystal silicon used in the simulation are taken or calculated using equations from [25] and are listed in Table 5. The absorption coefficient data are taken from those listed in the work of [83] and tabulated in Appendix A. Symbol Description Value Units Reference Eg,Si Bandgap of Si 1.12 eV [25] \u00CF\u0087,Si Electron affinity of Si 4.05 eV [25] NV,Si Effective density of states in valence band of Si 1.8 x 10 9 cm -3 [25] \u00CE\u00B5r,Si Relative permittivity 11.9 [25] 83 Symbol Description Value Units Reference of Si NA,Si Doping concentration of silicon of Si 1 x 10 16 cm -3 De,Si Minority electron diffusivity of silicon 0.0033 m 2 /s [25] \u00CF\u0084e,Si Electron minority carrier lifetime of silicon 4.17 x 10 -7 s [25] Le,Si Electron diffusion length of silicon 3.685 x 10 -5 m [25] B,Si Thickness of 4\u00E2\u0080\u009D silicon wafer 525 \u00C2\u00B5m Table 5. Relevant simulation parameters of p-type silicon 5.2.3 Properties of CIGS CIGS is being used in low-cost, thin film solar cells. It has a higher absorption coefficient and electron affinity than single crystal silicon, and a bandgap that is between 1.04eV and 1.7eV depending on the mole fraction of gallium and indium. The higher electron affinity gives higher Schottky barrier height, lower dark current density, and higher built in voltage, and therefore possible higher conversion efficiency. The absorption coefficient data is calculated from the imaginary component of the refractive index (kr) of CuIn1-xGaxSe2, listed in the work of [84]. The absorption coefficient can be calculated by: 4 rk\u00EF\u0081\u00B0\u00EF\u0081\u00A1 \u00EF\u0081\u00AC \u00EF\u0080\u00BD and the result is tabulated in appendix A.In this simulation, the absorption coefficient with 0.2 mole fraction of gallium is calculated, in order to cope with other parameters obtained from different references. The thickness of the CIGS absorber is chosen to be 3\u00C2\u00B5m, which is the 84 typical thickness for CIGS thin film solar cells. CIGS can be p-doped by incorporating vacancies in the deposited film and used as the solar cell base. It is conventionally combined with n-type cadmium sulfide (CdS) window layer to form the heterojunction solar cell. Here we replace the CdS layer with graphene, deposited on top of CIGS to form the Schottky junction. Table 6 lists relevant properties of p-type CIGS that are used as simulation parameters. Symbol Description Value Units Reference Eg,CIGS Bandgap of CIGS 1.1085 eV [85] \u00CF\u0087, CIGS Electron affinity of CIGS 4.1 eV [86] NV,CIGS Effective density of states in valence band of CIGS 1.78 x 10 19 cm -3 [87] \u00CE\u00B5r,CIGS Relative permittivity of CIGS 13.6 [87] NA,CIGS Doping concentration of silicon of CIGS 1 x 10 16 cm -3 [87] De,CIGS Minority electron diffusivity of CIGS 2.59 x 10 -4 m 2 /s [87] \u00CF\u0084e,CIGS Electron minority carrier lifetime of CIGS 1.3 x 10 -6 s [88] Le,CIGS Electron diffusion length of CIGS 1.83 x 10 -5 m [87, 88] B, CIGS Thickness of CIGS 3 \u00C2\u00B5m [86] Table 6. Relevant simulation parameters of p-type CIGS 5.2.4 Properties of CdTe CdTe is another semiconductor that is being used as a thin film absorber in p-n junction solar cells. CdTe has high absorption coefficient and only a few micron thick configuration is needed to absorb most of the photons. The bandgap is 1.5 eV, which is close to the optimum 85 bandgap of 1.4 eV for single-junction solar cell. CdTe is usually combined with a n-type CdS window layer to create the p-n junction solar cell. Here, we replace the n-type CdS layer with graphene, and replaces p-n junction with Schottky junction. Table 7 lists relevant properties of p- type CdTe that are used as simulation parameters. CdTe thin film can be economically deposited by various methods including: closed-space sublimation, spray deposition, screen printing, electro-deposition, CVD, and galvanic deposition [89]. The low deposition cost combined with low material used (few microns) indicate that the graphene/CdTe Schottky junction solar cell could have lower fabrication cost compared to single crystal silicon based p-n junction cells. The absorption coefficient data are taken from those listed in the work of [87], and is tabulated in appendix A, and the thickness of the CdTe layer is chosen to be 3 \u00C2\u00B5m. Symbol Description Value Units Reference Eg,CdTe Bandgap of CdTe 1.5 eV [19] \u00CF\u0087,CdTe Electron affinity of CdTe 4.05 eV [19] NV,CdTe Effective density of states in valence band of CdTe 1.8 x 10 19 cm -3 [87] \u00CE\u00B5r,CdTe Relative permittivity of CdTe 9.4 [87] NA,CdTe Doping concentration of silicon of CdTe 1 x 10 16 cm -3 [19] De,CdTe Minority electron diffusivity of CdTe 8.29 x 10 -4 m 2 /s [90] \u00CF\u0084e,CdTe Electron minority carrier lifetime of CdTe 0.5 x 10 -9 s [90] Le,CdTe Electron diffusion length of CdTe 6.44 x 10 -7 m [90] B,CIGS Thickness of CdTe 3 \u00C2\u00B5m [19] Table 7. Relevant simulation parameters of p-type CdTe 86 5.2.5 Properties of Amorphous Silicon Amorphous silicon (a-Si) has several advantages compared to single crystal silicon, that makes it a favorable material for solar cell. It has higher absorption coefficient, higher bandgap (1.7 eV for regular hydrogen passivated a-Si), lower fabrication temperature (<300 o C) and lower material usage. Table 8 lists relevant properties of p-type a-Si that are used as simulation parameters. From [91], a 1 \u00C2\u00B5m a-Si layer can absorb about 90% of usable solar energy. The a-Si thin film can be deposited by plasma-enhanced chemical vapor deposition over large area. Here we combine graphene with p-type a-Si to form the Schottky junction. The absorption coefficient data are taken from those listed in the work of [92], and is tabulated in appendix A, and the thickness of the a-Si layer is chosen to be 1 \u00C2\u00B5m. Symbol Description Value Units Reference Eg,a-Si Bandgap of a-Si 1.7 eV [93] \u00CF\u0087, a-Si Electron affinity of a-Si 3.93 eV [94] NV, a-Si Effective density of states in valence band of a-Si 2 x 10 22 cm -3 [93] \u00CE\u00B5r, a-Si Relative permittivity of a-Si 11 [95] NA, a-Si Doping concentration of silicon of a-Si 1 x 10 16 cm -3 [96] De, a-Si Minority electron diffusivity of a-Si 1.295 x 10 -7 m 2 /s 5.18 \u00CF\u0084e, a-Si Electron minority carrier lifetime of a-Si 1 x 10 -6 s 5.18 Le, a-Si Electron diffusion length of a-Si 3.60 x 10 -7 m 5.18 B, a-Si Thickness of a-Si 1 \u00C2\u00B5m 5.18 Table 8. Relevant simulation parameters of p-type amorphous silicon 87 5.3 Modeling Results and Discussion Matlab code for the simulation is included in the Appendix B. Figure 37 shows the current density-voltage characteristic under AM1.5G illumination for the four solar cells. Table 9 tabulates the short circuit current density, open circuit voltage, fill factor, conversion efficiency, and base bulk resistance for the four solar cells. Graphene/CIGS solar cell has higher short circuit current density than graphene/silicon because of higher absorption coefficient. The performance of graphene/CIGS solar cell can be further optimized by tuning the mole fraction of gallium to tune the bandgap of CIGS to close to the optimum bandgap (1.4 eV). In this simulation the bandgap is not optimized due to the unavailability of some of the modeling parameters that have a dependency on the mole fraction of gallium. Graphene/CdTe solar cell also has higher open circuit voltage than graphene/silicon cell because of the higher Schottky barrier height, which is due to the higher bandgap. It also has higher short circuit current density because of the higher absorption coefficient, and the lower base bulk resistance due to the thinner base (3 \u00C2\u00B5m compared to 525 \u00C2\u00B5m). Graphene/a-Si solar cell also has higher open circuit voltage than graphene/silicon cell because of the higher bandgap. However, the low electron mobility of a-Si results in low short circuit current density and very high semiconductor (base) bulk resistance, leading to a low fill factor and low efficiency. From this simulation, we identify a potential Schottky junction solar cell that can achieve a conversion efficiency of 11.3%, by combining graphene with p-type CdTe. A method to improve the open circuit voltage and short circuit current density of this cell is to lower the work function of graphene. To-date, the most effective N-doping of graphene film has been achieved by N + -ion irradiation of graphene to introduce vacancy defects in the honeycomb atomic 88 configuration, followed by annealing in ammonia (NH3) gas to introduce atomic N in the graphene and restore its structure. Figure 37. Current Density-Voltage Characteristics under AM1.5G illumination for graphene/Si (red curve), graphene/CIGS (blue curve), graphene/CdTe (black curve), and graphene/a-Si (green curve) Semiconductor Short circuit current density [mA/cm 2 ] Open circuit voltage [V] Fill factor [%] Conversion efficiency [%] Base bulk resistance [\u00E2\u0084\u00A6] Silicon 28 0.25 48.8 3.48 0.074 CIGS 38.1 0.31 49.5 5.83 0.0075 CdTe 23.6 0.64 75.2 11.3 0.0047 a-Si 4.8 0.52 7.15 0.63 104 Table 9. Solar cell parameters obtained from numerical simulation 89 5.4 Chapter Summary In this chapter, we perform numerical simulations of several graphene based Schottky junction solar cells in an attempt to identify potential cells that can outperform graphene/silicon cell. We have chosen CIGS, CdTe, and a-Si because of their high absorption coefficient and high bandgap, which can give higher short circuit current density and open circuit voltage. Also, these materials can be more economically fabricated into solar cells compared to single crystal silicon. From the simulation results, we identify graphene/CdTe as a potential Schottky junction solar cell that can achieve a conversion efficiency of 11.3%, if the graphene sheet resistance of 30 ohms/square and transmittance of 90% can be attained. 90 Chapter 6: Conclusion 6.1 Conclusion In this thesis work, we have fabricated p-type and n-type silicon Schottky junction solar cells by integrating graphene and gold coated PAN nanofibers with silicon, in an attempt to develop new cells that can overcome the low transparency of traditional metal films and increasing cost of ITO. Graphene can be synthesized by chemical vapour deposition of methane on copper foils, and the scale is principally unconstrained, limited only by the size of growth chamber. The transparency of graphene is above that of ITO. However, transfer of grown graphene from copper to other substrates inevitably induces cracks that results in large sheet resistance and low fill factor. The fabricated graphene/p-type silicon Schottky junction solar cell has conversion efficiency of 0.1%. This can be improved by optimizing the trade-off between transparency and sheet resistance, improving the graphene transfer process, and applying controllable and repeatable methods of graphene doping. We have also fabricated gold coated PAN/nanofiber Schottky junction solar cells that show a conversion efficiency of 1.78%. Compared to ITO, the PAN nanofiber can be more economically and more easily produced by conventional electrospinning and coated by sputtering. The nanofibers has sheet resistance and transparency comparable to ITO, while having a lower fabrication cost. The solar cell efficiency could be improved by optimizing the sheet resistance and transparency of the nanofiber mesh, by optimizing the density of the nanofibers. Patterning silicon surface to reduce light reflection could also improve the cell performance. In addition, the gold coated PAN nanofibers are flexible and can be applied onto flexible solar cells unlike the brittle ITO. 91 In addition to experimental work, we have also performed numerical simulations of various graphene based Schottky junction solar cells to identify possible future research directions. We have chosen CIGS, CdTe, and a-Si as the possible candidates for the semiconductor because of their high absorption coefficient, high/tunable bandgap, and the ability to be economically fabricated compared to single crystal silicon. The simulation results identify graphene/p-type CdTe as a potential Schottky junction solar cell that can achieve a conversion efficiency of 11.3%, if the graphene sheet resistance of 30 ohms/square and transmittance of 90% can be attained. 6.2 Possible Future Research Directions From this thesis work, we identify several future research directions. To use graphene based Schottky junction solar cell as nano-scale power sources in integrated nano-systems, scalable graphene patterning processes should be developed. Graphene on CdTe Schottky junction solar cells should be investigated as it can achieve higher efficiency than graphene/silicon, as estimated in our numerical simulation. For gold coated PAN nanofiber based solar cell, better contact method to improve the interface between nanofibers and semiconductor should be developed. Also, a top antireflection layer that encapsulates the solar cell should be developed to make modules. 92 Bibliography [1] J. Nelson, The Physics of Solar Cells. World Scientific, 2003. [2] T. L. Chu and S. S. Chu, \"Thin film II\u00E2\u0080\u0093VI photovoltaics,\" Solid-State Electronics, vol. 38, pp. 533-549, 3, 1995. [3] J. Zayed and S. Philippe, \"Acute oral and inhalation toxicities in rats with cadmium telluride,\" Int. J. Toxicol., vol. 28, pp. 259-265, 2009. [4] W. Anderson, A. Delahoy and R. Milano, \"An 8% efficient layered Schottky\u00E2\u0080\u0090barrier solar cell,\" J. Appl. Phys., vol. 45, pp. 3913-3915, 1974. [5] W. A. Anderson and R. Milano, \"IV characteristics for Silicon Schottky solar cells,\" Proc IEEE, vol. 63, pp. 206-208, 1975. [6] W. Anderson, S. Vernon, A. Delahoy, K. Ng, P. Mathe and T. Poon, \"Variables which influence silicon schottky solar cell performance,\" in Electron Devices Meeting, 1975 International, 1975, pp. 217-220. [7] B. Sharma, \"Metal--Semiconductor Schottky Barrier Junctions and Their Applications,\" Plenum Press, pp. 370, 1984. [8] J. Tang, X. Wang, L. Brzozowski, D. A. R. Barkhouse, R. Debnath, L. Levina and E. H. Sargent, \"Schottky quantum dot solar cells stable in air under solar illumination,\" Adv Mater, vol. 22, pp. 1398-1402, 2010. [9] J. M. Luther, M. Law, M. C. Beard, Q. Song, M. O. Reese, R. J. Ellingson and A. J. Nozik, \"Schottky solar cells based on colloidal nanocrystal films,\" Nano Letters, vol. 8, pp. 3488-3492, 2008. 93 [10] H. Kim, C. Gilmore, A. Pique, J. Horwitz, H. Mattoussi, H. Murata, Z. Kafafi and D. Chrisey, \"Electrical, optical, and structural properties of indium\u00E2\u0080\u0093tin\u00E2\u0080\u0093oxide thin films for organic light-emitting devices,\" J. Appl. Phys., vol. 86, pp. 6451-6461, 1999. [11] F. Nelson, V. Kamineni, T. Zhang, E. Comfort, J. Lee and A. Diebold, \"Optical properties of large-area polycrystalline chemical vapor deposited graphene by spectroscopic ellipsometry,\" Appl. Phys. Lett., vol. 97, pp. 253110, 2010. [12] P. Blake, P. D. Brimicombe, R. R. Nair, T. J. Booth, D. Jiang, F. Schedin, L. A. Ponomarenko, S. V. Morozov, H. F. Gleeson and E. W. Hill, \"Graphene-based liquid crystal device,\" Nano Letters, vol. 8, pp. 1704-1708, 2008. [13] H. Liu, Y. Liu and D. Zhu, \"Chemical doping of graphene,\" J.Mater.Chem., vol. 21, pp. 3335-3345, 2010. [14] L. Xie, X. Wang, H. Mao, R. Wang, M. Ding, Y. Wang, B. Ozyilmaz, K. Ping Loh, A. T. S. Wee and W. Chen, \"Electrical measurement of non-destructively p-type doped graphene using molybdenum trioxide,\" Appl. Phys. Lett., vol. 99, pp. 012112-012112-3, 2011. [15] Y. Zhu and J. M. Tour, \"Organic light-emitting diodes: Non-oxide boost,\" Nature Photonics, vol. 6, pp. 72-73, 2012. [16] J. Wu, H. A. Becerril, Z. Bao, Z. Liu, Y. Chen and P. Peumans, \"Organic solar cells with solution-processed graphene transparent electrodes,\" Appl. Phys. Lett., vol. 92, pp. 263302- 263302-3, 2008. [17] H. Ago, T. Kugler, F. Cacialli, W. R. Salaneck, M. S. P. Shaffer, A. H. Windle and R. H. Friend, \"Work functions and surface functional groups of multiwall carbon nanotubes,\" The Journal of Physical Chemistry B, vol. 103, pp. 8116-8121, 1999. 94 [18] X. Li, H. Zhu, K. Wang, A. Cao, J. Wei, C. Li, Y. Jia, Z. Li, X. Li and D. Wu, \"Graphene\u00E2\u0080\u0090 On\u00E2\u0080\u0090Silicon Schottky Junction Solar Cells,\" Adv Mater, vol. 22, pp. 2743-2748, 2010. [19] S. Shambayati, D. Lin, P. Servati and D. Pulfrey, \"performance predictions for Graphene- Window schottky-barrier solar cells,\" 2011. [20] I. Jung, J. S. Rhyee, J. Y. Son, R. S. Ruoff and K. Y. Rhee, \"Colors of graphene and graphene-oxide multilayers on various substrates,\" Nanotechnology, vol. 23, pp. 025708, 2011. [21] S. Roddaro, P. Pingue, V. Piazza, V. Pellegrini and F. Beltram, \"The optical visibility of graphene: Interference colors of ultrathin graphite on SiO2,\" Nano Letters, vol. 7, pp. 2707-2710, 2007. [22] X. Du, I. Skachko, A. Barker and E. Y. Andrei, \"Approaching ballistic transport in suspended graphene,\" Nature Nanotechnology, vol. 3, pp. 491-495, 2008. [23] S. Tongay, T. Schumann and A. Hebard, \"Graphite based Schottky diodes formed on Si, GaAs, and 4H-SiC substrates,\" Appl. Phys. Lett., vol. 95, pp. 222103-222103-3, 2009. [24] S. M. Sze and K. K. Ng, Physics of Semiconductor Devices. Wiley-interscience, 2006. [25] D. L. Pulfrey, Understanding Modern Transistors and Diodes. Cambridge University Press, 2010. [26] A. Kumar, S. Sista and Y. Yang, \"Dipole induced anomalous S-shape IV curves in polymer solar cells,\" J. Appl. Phys., vol. 105, pp. 094512-094512-6, 2009. [27] W. Tress, A. Petrich, M. Hummert, M. Hein, K. Leo and M. Riede, \"Imbalanced mobilities causing S-shaped IV curves in planar heterojunction organic solar cells,\" Appl. Phys. Lett., vol. 98, pp. 063301-063301-3, 2011. [28] T. Wang, E. Iwaniczko, M. Page, Q. Wang, Y. Xu, Y. Yan, D. Levi, L. Roybal, R. Bauer and H. Branz, \"High-efficiency silicon heterojunction solar cells by HWCVD,\" in Photovoltaic 95 Energy Conversion, Conference Record of the 2006 IEEE 4th World Conference on, 2006, pp. 1439-1442. [29] H. Card and E. Rhoderick, \"Studies of tunnel MOS diodes I. Interface effects in silicon Schottky diodes,\" J. Phys. D, vol. 4, pp. 1589, 2002. [30] C. C. Chen, M. Aykol, C. C. Chang, A. Levi and S. B. Cronin, \"Graphene-silicon Schottky diodes,\" Nano Letters, vol. 11, pp. 1863-1867, 2011. [31] S. Stankovich, D. A. Dikin, G. H. B. Dommett, K. M. Kohlhaas, E. J. Zimney, E. A. Stach, R. D. Piner, S. B. T. Nguyen and R. S. Ruoff, \"Graphene-based composite materials,\" Nature, vol. 442, pp. 282-286, 2006. [32] G. Eda, G. Fanchini and M. Chhowalla, \"Large-area ultrathin films of reduced graphene oxide as a transparent and flexible electronic material,\" Nature Nanotechnology, vol. 3, pp. 270- 274, 2008. [33] Y. Hernandez, M. Lotya, D. Rickard, S. D. Bergin and J. N. Coleman, \"Measurement of multicomponent solubility parameters for graphene facilitates solvent discovery,\" Langmuir, vol. 26, pp. 3208-3213, 2009. [34] W. A. De Heer, C. Berger, X. Wu, P. N. First, E. H. Conrad, X. Li, T. Li, M. Sprinkle, J. Hass and M. L. Sadowski, \"Epitaxial graphene,\" Solid State Commun, vol. 143, pp. 92-100, 2007. [35] C. Mattevi, G. Eda, S. Agnoli, S. Miller, K. A. Mkhoyan, O. Celik, D. Mastrogiovanni, G. Granozzi, E. Garfunkel and M. Chhowalla, \"Evolution of electrical, chemical, and structural properties of transparent and conducting chemically derived graphene thin films,\" Advanced Functional Materials, vol. 19, pp. 2577-2583, 2009. 96 [36] S. De, P. J. King, M. Lotya, A. O'Neill, E. M. Doherty, Y. Hernandez, G. S. Duesberg and J. N. Coleman, \"Flexible, Transparent, Conducting Films of Randomly Stacked Graphene from Surfactant\u00E2\u0080\u0090Stabilized, Oxide\u00E2\u0080\u0090Free Graphene Dispersions,\" Small, vol. 6, pp. 458-464, 2009. [37] G. Eda and M. Chhowalla, \"Chemically Derived Graphene Oxide: Towards Large\u00E2\u0080\u0090Area Thin\u00E2\u0080\u0090Film Electronics and Optoelectronics,\" Adv Mater, vol. 22, pp. 2392-2415, 2010. [38] C. Mattevi, H. Kim and M. Chhowalla, \"A review of chemical vapour deposition of graphene on copper,\" J.Mater.Chem., vol. 21, pp. 3324-3334, 2010. [39] K. S. Kim, Y. Zhao, H. Jang, S. Y. Lee, J. M. Kim, K. S. Kim, J. H. Ahn, P. Kim, J. Y. Choi and B. H. Hong, \"Large-scale pattern growth of graphene films for stretchable transparent electrodes,\" Nature, vol. 457, pp. 706-710, 2009. [40] J. Coraux, Alpha T. N'Diaye, C. Busse and T. Michely, \"Structural coherency of graphene on Ir (111),\" Nano Letters, vol. 8, pp. 565-570, 2008. [41] S. Y. Kwon, C. V. Ciobanu, V. Petrova, V. B. Shenoy, J. Bare o, V. Gambin, I. Petrov and S. Kodambaka, \"Growth of semiconducting graphene on palladium,\" Nano Letters, vol. 9, pp. 3985, 2009. [42] P. W. Sutter, J. I. Flege and E. A. Sutter, \"Epitaxial graphene on ruthenium,\" Nature Materials, vol. 7, pp. 406-411, 2008. [43] X. Li, Y. Zhu, W. Cai, M. Borysiak, B. Han, D. Chen, R. D. Piner, L. Colombo and R. S. Ruoff, \"Transfer of large-area graphene films for high-performance transparent conductive electrodes,\" Nano Letters, vol. 9, pp. 4359-4363, 2009. [44] X. Li, W. Cai, J. An, S. Kim, J. Nah, D. Yang, R. Piner, A. Velamakanni, I. Jung and E. Tutuc, \"Large-area synthesis of high-quality and uniform graphene films on copper foils,\" Science, vol. 324, pp. 1312-1314, 2009. 97 [45] M. Butt, \"Effect of hydrogen attack on the strength of high purity copper,\" J. Mater. Sci. Lett., vol. 2, pp. 1-2, 1983. [46] Y. Trehan, \"The Reduction of Copper Oxides by Molecular Hydrogen,\" Zeitschrift F\u00C3\u00BCr Anorganische Und Allgemeine Chemie, vol. 318, pp. 107-112, 1962. [47] X. Li, C. W. Magnuson, A. Venugopal, J. An, J. W. Suk, B. Han, M. Borysiak, W. Cai, A. Velamakanni and Y. Zhu, \"Graphene films with large domain size by a two-step chemical vapor deposition process,\" ArXiv Preprint arXiv:1010.4731, 2010. [48] M. Peng, J. Zhi, G. Jian-Nan, P. Hong-Liang, L. Xin-Yu, Y. Tian-Chun, W. Hong and W. Guan-Zhong, \"Chemical Vapour Deposition Graphene Radio-Frequency Field-Effect Transistors,\" Chinese Physics Letters, vol. 29, pp. 057302, 2012. [49] L. Ren, Q. Zhang, J. Yao, Z. Sun, R. Kaneko, Z. Yan, S. Nanot, Z. Jin, I. Kawayama and M. Tonouchi, \"Terahertz and Infrared Spectroscopy of Gated Large-Area Graphene,\" Nano Letters, vol. 12, pp. 3711-3715, 2012. [50] J. W. Suk, A. Kitt, C. W. Magnuson, Y. Hao, S. Ahmed, J. An, A. K. Swan, B. B. Goldberg and R. S. Ruoff, \"Transfer of CVD-grown monolayer graphene onto arbitrary substrates,\" ACS Nano, vol. 5, pp. 6916-6924, 2011. [51] I. Jung, J. S. Rhyee, J. Y. Son, R. S. Ruoff and K. Y. Rhee, \"Colors of graphene and graphene-oxide multilayers on various substrates,\" Nanotechnology, vol. 23, pp. 025708, 2011. [52] S. Roddaro, P. Pingue, V. Piazza, V. Pellegrini and F. Beltram, \"The optical visibility of graphene: Interference colors of ultrathin graphite on SiO2,\" Nano Letters, vol. 7, pp. 2707-2710, 2007. 98 [53] A. Ferrari, J. Meyer, V. Scardaci, C. Casiraghi, M. Lazzeri, F. Mauri, S. Piscanec, D. Jiang, K. Novoselov and S. Roth, \"Raman spectrum of graphene and graphene layers,\" Phys. Rev. Lett., vol. 97, pp. 187401, 2006. [54] Y. Wang, Z. Ni, T. Yu, Z. X. Shen, H. Wang, Y. Wu, W. Chen and A. T. Shen Wee, \"Raman studies of monolayer graphene: the substrate effect,\" The Journal of Physical Chemistry C, vol. 112, pp. 10637-10640, 2008. [55] J. Lee, L. Tao, Y. Hao, R. S. Ruoff and D. Akinwande, \"Embedded-gate graphene transistors for high-mobility detachable flexible nanoelectronics,\" Appl. Phys. Lett., vol. 100, pp. 152104-152104-4, 2012. [56] H. Ji, Y. Hao, Y. Ren, M. Charlton, W. H. Lee, Q. Wu, H. Li, Y. Zhu, Y. Wu and R. Piner, \"Graphene Growth Using a Solid Carbon Feedstock and Hydrogen,\" ACS Nano, vol. 5, pp. 7656- 7661, 2011. [57] A. Soudi, G. Aivazian, S. F. Shi, X. Xu and Y. Gu, \"Probing transconductance spatial variations in graphene nanoribbon field-effect transistors using scanning gate microscopy,\" Appl. Phys. Lett., vol. 100, pp. 033115-033115-4, 2012. [58] C. Y. Su, D. Fu, A. Y. Lu, K. K. Liu, Y. Xu, Z. Y. Juang and L. J. Li, \"Transfer printing of graphene strip from the graphene grown on copper wires,\" Nanotechnology, vol. 22, pp. 185309, 2011. [59] Y. Lee, S. Bae, H. Jang, S. Jang, S. E. Zhu, S. H. Sim, Y. I. Song, B. H. Hong and J. H. Ahn, \"Wafer-scale synthesis and transfer of graphene films,\" Nano Letters, vol. 10, pp. 490-493, 2010. 99 [60] S. J. Kang, B. Kim, K. S. Kim, Y. Zhao, Z. Chen, G. H. Lee, J. Hone, P. Kim and C. Nuckolls, \"Inking Elastomeric Stamps with Micro\u00E2\u0080\u0090Patterned, Single Layer Graphene to Create High\u00E2\u0080\u0090Performance OFETs,\" Adv Mater, vol. 23, pp. 3531-3535, 2011. [61] T. J. Ha, D. Akinwande and A. Dodabalapur, \"Hybrid graphene/organic semiconductor field-effect transistors,\" Appl. Phys. Lett., vol. 101, pp. 033309-033309-3, 2012. [62] S. Bae, H. Kim, Y. Lee, X. Xu, J. S. Park, Y. Zheng, J. Balakrishnan, T. Lei, H. R. Kim and Y. I. Song, \"Roll-to-roll production of 30-inch graphene films for transparent electrodes,\" Nature Nanotechnology, vol. 5, pp. 574-578, 2010. [63] A. G. Aberle, \"Surface passivation of crystalline silicon solar cells: a review,\" Prog Photovoltaics Res Appl, vol. 8, pp. 473-487, 2000. [64] D. L. Pulfrey, \"MIS solar cells: A review,\" Electron Devices, IEEE Transactions on, vol. 25, pp. 1308-1317, 1978. [65] K. Ellmer, \"Past achievements and future challenges in the development of optically transparent electrodes,\" Nature Photonics, vol. 6, pp. 809-817, 2012. [66] S. Pang, Y. Hernandez, X. Feng and K. M\u00C3\u00BCllen, \"Graphene as transparent electrode material for organic electronics,\" Adv Mater, vol. 23, pp. 2779-2795, 2011. [67] F. Bonaccorso, Z. Sun, T. Hasan and A. Ferrari, \"Graphene photonics and optoelectronics,\" Nature Photonics, vol. 4, pp. 611-622, 2010. [68] Y. C. Lin, C. Y. Lin and P. W. Chiu, \"Controllable graphene N-doping with ammonia plasma,\" Appl. Phys. Lett., vol. 96, pp. 133110-133110-3, 2010. [69] A. K. Singh, M. W. Iqbal, V. K. Singh, M. Z. Iqbal, J. H. Lee, S. H. Chun, K. Shin and J. Eom, \"Molecular n-doping of chemical vapor deposition grown graphene,\" Journal of Materials Chemistry, 2012. 100 [70] D. Wei, Y. Liu, Y. Wang, H. Zhang, L. Huang and G. Yu, \"Synthesis of n-doped graphene by chemical vapor deposition and its electrical properties,\" Nano Letters, vol. 9, pp. 1752-1758, 2009. [71] X. Wang, X. Li, L. Zhang, Y. Yoon, P. K. Weber, H. Wang, J. Guo and H. Dai, \"N-doping of graphene through electrothermal reactions with ammonia,\" Science, vol. 324, pp. 768-771, 2009. [72] T. Cui, R. Lv, Z. H. Huang, H. Zhu, J. Zhang, Z. Li, Y. Jia, F. Kang, K. Wang and D. Wu, \"Synthesis of nitrogen-doped carbon thin films and their applications in solar cells,\" Carbon, vol. 49, pp. 5022-5028, 2011. [73] B. Guo, Q. Liu, E. Chen, H. Zhu, L. Fang and J. R. Gong, \"Controllable N-doping of graphene,\" Nano Letters, vol. 10, pp. 4975, 2010. [74] Y. Zhang, V. W. Brar, C. Girit, A. Zettl and M. F. Crommie, \"Origin of spatial charge inhomogeneity in graphene,\" Nature Physics, vol. 5, pp. 722-726, 2009. [75] Y. Ye, Y. Dai, L. Dai, Z. Shi, N. Liu, F. Wang, L. Fu, R. Peng, X. Wen and Z. Chen, \"High- performance single CdS nanowire (nanobelt) Schottky junction solar cells with Au/graphene Schottky electrodes,\" ACS Applied Materials & Interfaces, vol. 2, pp. 3406-3410, 2010. [76] L. Li, H. Wang, X. Fang, T. Zhai, Y. Bando and D. Golberg, \"High-performance Schottky solar cells using ZrS2 nanobelt networks,\" Energy & Environmental Science, vol. 4, pp. 2586- 2590, 2011. [77] G. Fan, H. Zhu, K. Wang, J. Wei, X. Li, Q. Shu, N. Guo and D. Wu, \"Graphene/silicon nanowire Schottky junction for enhanced light harvesting,\" ACS Applied Materials & Interfaces, vol. 3, pp. 721-725, 2011. 101 [78] T. Feng, D. Xie, Y. Lin, Y. Zang, T. Ren, R. Song, H. Zhao, H. Tian, X. Li and H. Zhu, \"Graphene based Schottky junction solar cells on patterned silicon-pillar-array substrate,\" Appl. Phys. Lett., vol. 99, pp. 233505-233505-3, 2011. [79] T. Feng, D. Xie, Y. Lin, H. Zhao, Y. Chen, H. Tian, T. Ren, X. Li, Z. Li and K. Wang, \"Efficiency enhancement of graphene/silicon-pillar-array solar cells by HNO3 and PEDOT- PSS,\" Nanoscale, vol. 4, pp. 2130-2133, 2012. [80] C. Xie, P. Lv, B. Nie, J. Jie, X. Zhang, Z. Wang, P. Jiang, Z. Hu, L. Luo and Z. Zhu, \"Monolayer graphene film/silicon nanowire array Schottky junction solar cells,\" Appl. Phys. Lett., vol. 99, pp. 133113-133113-3, 2011. [81] Y. Ye, L. Gan, L. Dai, Y. Dai, X. Guo, H. Meng, B. Yu, Z. Shi, K. Shang and G. Qin, \"A simple and scalable graphene patterning method and its application in CdSe nanobelt/graphene Schottky junction solar cells,\" Nanoscale, vol. 3, pp. 1477-1481, 2011. [82] Anonymous \"ASTM G173-03 Reference Spectra,\" vol. 2012, 8 Nov 2012, 2012. [83] M. A. Green and M. J. Keevers, \"Optical properties of intrinsic silicon at 300 K,\" Prog Photovoltaics Res Appl, vol. 3, pp. 189-192, 1995. [84] M. I. Alonso, M. Garriga, C. Durante Rinc\u00C3\u00B3n, E. Hern\u00C3\u00A1ndez and M. Le\u00C3\u00B3n, \"Optical functions of chalcopyrite CuGa x In 1-x Se 2 alloys,\" Appl. Phys. A, vol. 74, pp. 659-664, 2002. [85] H. K. Singh, \"CIGS Thin Film Photovoltaic,\" 2011. [86] M. Gloeckler, Numerical Modeling of CIGS Solar Cells: Definition of the Baseline and Explanation of Superposition Failure, 2003. [87] M. Gloeckler, Device Physics of Cu (in, Ga) Se2 Thin-Film Solar Cells, 2005. 102 [88] J. H. Werner, J. Mattheis and U. Rau, \"Efficiency limitations of polycrystalline thin film solar cells: case of Cu (In, Ga) Se< sub> 2,\" Thin Solid Films, vol. 480, pp. 399-409, 2005. [89] D. Bonnet, \"The CdTe thin film solar cell-an overview,\" Int J Sol Energy, vol. 12, pp. 1-14, 1992. [90] D. Pulfrey, J. Dell and L. Faraone, \"Performance predictions for monolithic thin-film CdTe/Ge tandem solar cells,\" in Proceedings of SPIE- the International Society for Optical Engineering, 2010, . [91] M. Zeman, \"Thin-Film Silicon PV Technology,\" 2010. [92] Anonymous \"Optical constants of a-Si (Amorphous silicon),\" 2012. [93] S. K. OLeary and S. M. Malik, \"A simplified joint density of states analysis of hydrogenated amorphous silicon,\" J. Appl. Phys., vol. 92, pp. 4276-4282, 2002. [94] H. Matsuura, T. Okuno, H. Okushi and K. Tanaka, \"Electrical properties of n\u00E2\u0080\u0090amorphous/p\u00E2\u0080\u0090 crystalline silicon heterojunctions,\" J. Appl. Phys., vol. 55, pp. 1012-1019, 1984. [95] N. A. Bakr, A. Funde, V. Waman, M. Kamble, R. Hawaldar, D. Amalnerkar, S. Gosavi and S. Jadkar, \"Determination of the optical parameters of a-Si: H thin films deposited by hot wire\u00E2\u0080\u0093 chemical vapour deposition technique using transmission spectrum only,\" Pramana, vol. 76, pp. 519-531, 2011. [96] A. Moore, \"Electron and hole drift mobility in amorphous silicon,\" Appl. Phys. Lett., vol. 31, pp. 762-764, 1977. 103 Appendices Appendix A Absorption Coefficient of the Semiconductors Used in Chapter 5 Absorption Coefficient [m -1 ] wavelength [nm] Si CIGS CdTe a-Si 280 236000000 56505420.74 0 162672195 290 224000000 48481259.9 0 158490614.5 300 173000000 43686345.95 35000000 153163713 310 144000000 41360750.9 33428571.43 146617115.7 320 128000000 40099358.37 31000000 140193158.7 330 117000000 39482360.11 29446428.57 130509734.5 340 109000000 38891204 28000000 121468016.9 350 104000000 37699111.44 26553571.43 112058726.7 360 102000000 36541018.11 25000000 102667942.3 104 Absorption Coefficient [m -1 ] wavelength [nm] Si CIGS CdTe a-Si 370 69700000 36080884.09 23000000 93498933.63 380 29300000 35964958.6 21000000 84747790.06 390 15000000 35538220.91 19371428.57 76613337.62 400 9520000 35154095.16 18000000 69022660.44 410 6740000 34617255.05 16950000 62139688.61 420 5000000 33443395.17 16000000 55907626.95 430 3920000 31324468.89 15000000 50286567.7 440 3110000 28702393.65 14000000 45230550.96 450 2550000 26261448.54 13000000 40678284.47 460 2100000 24061750.68 12000000 36589290.64 470 1720000 22132757.91 10916666.67 32955333.73 480 1480000 20426789.7 10000000 29687948.48 490 1270000 18925509.65 9458333.333 26718720.75 105 Absorption Coefficient [m -1 ] wavelength [nm] Si CIGS CdTe a-Si 500 1110000 17531069.79 9000000 24061749.73 510 970000 16171667.53 8500000 21664540.85 520 880000 15069584.08 8000000 19519339.82 530 785000 14241333.62 7468750 17585423.7 540 705000 13505515.13 7000000 15832774.24 550 639000 12771023.75 6681250 14236943.46 560 578000 12085975.93 6400000 12769529.31 570 532000 11488772.37 6093181.818 11405043.27 580 488000 10953970.51 5800000 10130745.52 590 449000 10478812.33 5537373.737 8942703.025 600 414000 10058022.72 5300000 7828359.295 610 381000 9678317.682 5094444.444 6780623.883 620 352000 9342523.487 4900000 5806569.982 106 Absorption Coefficient [m -1 ] wavelength [nm] Si CIGS CdTe a-Si 630 327000 9045492.734 4700000 4892357.939 640 304000 8777010.083 4500000 4032529.677 650 281000 8539402.594 4292857.143 3245417.776 660 258000 8334299.447 4100000 2503726.796 670 238000 8155157.501 3937142.857 1765658.11 680 221000 7991340.798 3800000 1144135.229 690 205000 7832226.933 3700000 755967.693 700 190000 7677544.126 3600000 509820.5527 710 177000 7530667.84 3450000 310624.9733 720 166000 7385166.116 3300000 169452.2238 730 154000 7234602.301 3195000 97192.58583 740 142000 7076052.934 3100000 61302.12143 750 130000 6912982.109 3011627.907 32518.96809 107 Absorption Coefficient [m -1 ] wavelength [nm] Si CIGS CdTe a-Si 760 119000 6747303.986 2900000 12114.74589 770 110000 6580914.221 2673705.167 1361.074948 780 101000 6415122.566 2326392.073 0 790 92800 6245536.277 1915882.943 0 800 85000 6074008.168 1500000 0 810 77500 5904546.693 1032570.423 0 820 70700 5741160.301 600000 0 830 64700 5587897.134 274920.1862 0 840 59100 5448158.135 80000 0 850 53500 5318000.639 22780.35764 0 860 48000 5191455.225 100 0 870 43200 5062552.47 0 0 880 38300 4925322.952 0 0 108 Absorption Coefficient [m -1 ] wavelength [nm] Si CIGS CdTe a-Si 890 34300 4773323.864 0 0 900 30600 4595313.885 0 0 910 27200 4397115.567 0 0 920 24000 4190052.976 0 0 930 21000 3985450.178 0 0 940 18300 3794631.239 0 0 950 15700 3628920.225 0 0 960 13400 3497971.817 0 0 970 11400 3393351.87 0 0 980 9590 3305749.805 0 0 990 7920 3227738.481 0 0 1000 6400 3151890.756 0 0 1010 5110 3070779.487 0 0 109 Absorption Coefficient [m -1 ] wavelength [nm] Si CIGS CdTe a-Si 1020 3990 2976977.533 0 0 1030 3020 2863057.752 0 0 1040 2260 2711614.322 0 0 1050 1630 2460518.584 0 0 1060 1110 2125599.407 0 0 1070 800 1737406.054 0 0 1080 620 1326487.789 0 0 1090 470 923393.8728 0 0 1100 350 558673.5697 0 0 1110 270 262876.1419 0 0 1120 200 66550.85206 0 0 1130 150 0 0 0 1140 100 0 0 0 110 Absorption Coefficient [m -1 ] wavelength [nm] Si CIGS CdTe a-Si 1150 68 0 0 0 1160 42 0 0 0 1170 22 0 0 0 1180 6.5 0 0 0 1190 3.6 0 0 0 1200 2.2 0 0 0 111 Appendix B MATLAB Code for the Simulation Used in Chapter 5 %Graphene/Semiconductor Schottky barrier solar cell modelling% %Derek Lin clear all; close all; format long %All units in Meters/Kilogram/Seconds unless otherwise specified %************************************************************************** %Physical Constants Temp=300; %Temperature = 300 Kelvin kB = 1.38e-23; %Boltzmann's constant [J/K] h = 6.63e-34; %Plank's constant [J*s] h_bar = h /(2*pi); %Dirac's constant unit: [J*s] q=1.6e-19; %Elementary charge [C] m0=9.1e-31; %Electron rest mass [kg] eps0 = 8.854e-12; %Permittivity of free space [F/m] c=3e8; %Speed of light in vacuum [m/s] kT=kB*Temp; %Thermal energy at 300K [J] Vth= 0.0259; %Thermal voltage at 300K [V] %************************************************************************** % Read in Wavelength and AM1.5G data Si_data=dlmread('Si.txt',''); wavelength=Si_data(:,1); %Wavelength from 280nm to 1200nm with 10nm increment [nm] S=Si_data(:,3); %Solar Spectral irradiance, global, AM1.5G on 37 degree tilted surface. [W/m2/nm] E_photon=h*c./(wavelength*1e-9); %Photon energy at particular wavelength [J] phi0=S./(E_photon); %Photon flux [photons/m^2/s] %************************************************************************** %Semiconductor Data %Silicon Eg_Si=1.12; %Bandgap of Silicon [eV] chi_Si = 4.05; %Electron Affinity of Silicon [eV] ni_Si=9.5e9*1e6; %Intrinsic carrier concentrations for Silicon, converted to [m^-3] Nv_Si=1.8e19*1e6; %Effective density of states in the valence band, converted to [m^-3] eps_Si=11.9*eps0; %Permittivity of Silicon Na_cm_Si=1e16; %Boron doping concentration [cm-3] mobE_Si = 88 + 1252/(1+6.984e-18*Na_cm_Si); %Mobility of electrons, depending on Boron doping concentration. [cm^2/V/s] mobE_Si = mobE_Si*1e-4; %Mobility [m^2/V/s] mobH_Si = 54.3 + 407/(1+3.745e-18*Na_cm_Si);%Mobility of holes [cm^2/V/s] mobH_Si = mobH_Si*1e-4; %Mobility of holes [m^2/V/s] De_Si = Vth*mobE_Si; %Electron diffusivity[m^2/s] 112 Te_Si = 5e-7/(1+2*Na_cm_Si*1e-17); %Electron recombination lifetime [s] Le_Si = sqrt(De_Si*Te_Si); %Electron diffusion length [m] mhcon_Si=0.4; %Conductivity effective-mass for holes, for Silicon vRh_Si=sqrt(kT/(2*pi*mhcon_Si*m0)); %Mean, unidirectional thermal velocities for holes at 300K [m/s]. For low doping density, fermi dirac integral terms = 1 alpha_Si=Si_data(:,2)*1e2; %Absorption Coefficient of Silicon, converted to [m^-1] %************************************************************************** %CIGS x_Ga = 0.2; %Mole fraction of Gallium chosen to be 0.2 to fit in with other parameters from different sources. Data from Alonso CIGS_data=dlmread('CIGS.txt',''); %Data from Alonso wavelength_CIGS=CIGS_data(:,1); %Wavelength [nm] CIGS_alpha = CIGS_data(:,2); % [m^-1] alpha_CIGS = interp1(wavelength_CIGS,CIGS_alpha,wavelength,'cubic'); %Absorption coefficient after interpolation [m^-1] Eg_CIGS = 1.010 + 0.626*x_Ga - 0.167*x_Ga*(1-x_Ga); %Bandgap of CIGS [eV] chi_CIGS = 4.1; %electron affinity of CIGS [eV] ni_cm_CIGS = 4.7e8; %Intrinsic carrier concentrations for CIGS [cm^-3] ni_CIGS = 4.7e8*1e6; %Intrinsic carrier concentrations for CIGS, converted to [m^-3] Nv_CIGS = 1.78e19*1e6; %Effective density of states in the valence band, converted to [m^-3] eps_CIGS = 13.6*eps0; %Permittivity of CIGS Na_cm_CIGS = 1e16; %Boron doping concentration for CIGS [cm-3] mobE_CIGS = 100*1e-4; %%Mobility of electrons [m2/V/s] mobH_CIGS = 25*1e-4; %%Mobility of holes [m2/V/s] De_CIGS = Vth*mobE_CIGS; %Electron diffusivity [m^2/s] Te_rad_CIGS = 1.3e-6; %radiative carrier lifetime [s] Te_CIGS = Te_rad_CIGS; Le_CIGS = sqrt(De_CIGS*Te_CIGS); %Electron diffusion length[m] mhcon_CIGS = 0.73; %effective mass for holes in CIGS vRh_CIGS = sqrt(kT/(2*pi*mhcon_CIGS*m0)); %Mean, unidirectional thermal velocities for holes in CIGS. For low doping density, fermi dirac integral terms = 1 %************************************************************************** %CdTe CdTe_data=dlmread('CdTe.txt',''); %Data from Gloecker wavelength_CdTe=CdTe_data(:,1); %Wavelength [nm] CdTe_alpha = CdTe_data(:,2); alpha_CdTe = interp1(wavelength_CdTe,CdTe_alpha,wavelength,'cubic'); %[m-1] Eg_CdTe = 1.5; %[eV] chi_CdTe = 4.05; %[eV] ni_CdTe=1.126e6*1e6; %[m^-3] Nv_CdTe = 1.8e19*1e6; %[m^-3] eps_CdTe = 9.4*eps0; Na_cm_CdTe = 1e16; %[cm^-3] mobE_CdTe = 320*1e-4; %[m2/V/s] mobH_CdTe = 40*1e-4; %[m2/V/s] De_CdTe = Vth*mobE_CdTe; %[m^2/s] Te_CdTe = 0.5e-9; %[s] Le_CdTe = sqrt(De_CdTe*Te_CdTe); mhcon_CdTe = (Nv_CdTe/2)^(2/3)*h^2/(2*pi*kT)/m0; 113 vRh_CdTe=sqrt(kT/(2*pi*mhcon_CdTe*m0)); %************************************************************************** %Amorhphous Silicon Eg_aSi=1.7; %eV chi_aSi = 3.93; %eV Nv_aSi=2e22*1e6; %[m^-3] eps_aSi=11*eps0; Na_cm_aSi=1e16; mobE_aSi = 5e-2; %[cm^2/V/s] mobE_aSi = mobE_aSi*1e-4; %[m^2/V/s] mobH_aSi = 6e-4; %[cm^2/V/s] mobH_aSi = mobH_aSi*1e-4; %[m^2/V/s] De_aSi = Vth*mobE_aSi; %[m^2/s] Te_aSi = 1e-6; Le_aSi = sqrt(De_aSi*Te_aSi); mhcon_aSi=0.5; vRh_aSi=sqrt(kT/(2*pi*mhcon_aSi*m0)); aSi_data=dlmread('aSi.txt',''); %Data from SOPRA 2 wavelength_aSi=aSi_data(:,1)*1e3; %Wavelength [nm] k_aSi=aSi_data(:,3); %Extinction coefficient aSi_alpha = (4*pi.*k_aSi)./(wavelength_aSi*1e-9); % [m-1] alpha_aSi = interp1(wavelength_aSi,aSi_alpha,wavelength,'cubic'); %unit: m-1 %************************************************************************** %************************************************************************** %Graphene data TR=0.9; %Transmittance of 4 layer doped graphene from Bae. et al phi0_semi=phi0.*TR; %Photon flux at semiconductor surface is attenuated by graphene layer phim=4.4; %Work function of single layer graphene [eV] %************************************************************************** %************************************************************************** %Select Cell Data %************************************************************************** %Cell data for p-Si phiB_Si=Eg_Si+chi_Si-phim; %Schottky Barrier Height for p-type Schottky Barrier [eV] xj_Si=0; B_Si=525e-6; %No emitter current, depletion thickness on the metal side is 0. xj=0, B=525um for 4 inch Si wafers SF_Si=1; SB_Si=1e12; %SF= 1, blocking contact for holes. SB=1e12, ohmic contact for holes Na_Si=Na_cm_Si*1e6; %Doping Concentration of Silicon wafer [m^-3] Vbi_Si=phiB_Si -(Vth*log(Nv_Si/Na_Si)); %Built in voltage [V] W_Si=sqrt(2*eps_Si/q/Na_Si*Vbi_Si); %Depletion Width [m] %************************************************************************** %Cell data for p-CIGS phiB_CIGS=Eg_CIGS+chi_CIGS-phim; xj_CIGS=0; B_CIGS=3e-6; %Base thickness = 3um for thin film PV SF_CIGS=1; SB_CIGS= 1e7*1e-2; Na_CIGS=Na_cm_CIGS*1e6; 114 Vbi_CIGS = phiB_CIGS - (Vth*log(Nv_CIGS/Na_CIGS)); W_CIGS=sqrt(2*eps_CIGS/q/Na_CIGS*Vbi_CIGS); %************************************************************************** %Cell data for p-CdTe phiB_CdTe=Eg_CdTe+chi_CdTe-phim; xj_CdTe=0; B_CdTe=3e-6; SF_CdTe=1; SB_CdTe=1e12; Na_CdTe=Na_cm_CdTe*1e6; Vbi_CdTe = phiB_CdTe - (Vth*log(Nv_CdTe/Na_CdTe)); W_CdTe=sqrt(2*eps_CdTe/q/Na_CdTe*Vbi_CdTe); %************************************************************************** % Cell data for p-a-Si:H phiB_aSi=Eg_aSi+chi_aSi-phim; xj_aSi=0; B_aSi=1e-6; SF_aSi=1; SB_aSi=1e12; Na_aSi=Na_cm_aSi*1e6; Vbi_aSi=phiB_aSi - (Vth*log(Nv_aSi/Na_aSi)); W_aSi=sqrt(2*eps_aSi/q/Na_aSi*Vbi_aSi); %************************************************************************** %Select Silicon for the base alpha=alpha_Si; vRh=vRh_Si; Nv=Nv_Si; eps=eps_Si; Le=Le_Si; De=De_Si; phiB = phiB_Si; xj=xj_Si; B=B_Si; SF=SF_Si; SB=SB_Si; Vbi=Vbi_Si; W=W_Si; % %Select CIGS for the base % alpha=alpha_CIGS; % vRh=vRh_CIGS; % Nv=Nv_CIGS; % eps=eps_CIGS; % Le=Le_CIGS; % De=De_CIGS; % phiB = phiB_CIGS; % xj=xj_CIGS; % B=B_CIGS; % SF=SF_CIGS; % SB=SB_CIGS; % Vbi=Vbi_CIGS; % W=W_CIGS; % %Select CdTe for the base 115 % alpha=alpha_CdTe; % vRh=vRh_CdTe; % Nv=Nv_CdTe; % eps=eps_CdTe; % Le=Le_CdTe; % De=De_CdTe; % phiB = phiB_CdTe; % xj=xj_CdTe; % B=B_CdTe; % SF=SF_CdTe; % SB=SB_CdTe; % Vbi=Vbi_CdTe; % W=W_CdTe; % %Select Amorphous Silicon for the base % alpha=alpha_aSi; % vRh=vRh_aSi; % Nv=Nv_aSi; % eps=eps_aSi; % Le=Le_aSi; % De=De_aSi; % phiB = phiB_aSi; % xj=xj_aSi; % B=B_aSi; % SF=SF_aSi; % SB=SB_aSi; % Vbi=Vbi_aSi; % W=W_aSi; %************************************************************************** %************************************************************************** %Depletion region photocurrent% JeD=q*phi0_semi.*exp(-alpha.*xj).*(1-exp(-alpha.*W)); %[A/m^2/nm] %Base photocurrent% B_prime=B-(xj+W); Qe=B_prime/Le; He=SB*Le/De; JeB= ((q*phi0_semi.*alpha*Le.*exp(-alpha.*(xj+W)))./(alpha.^2*Le^2-1)).*(alpha.*Le-((He.*(cosh(Qe)-exp(- alpha.*B_prime)))+sinh(Qe)+(alpha.*Le.*exp(-alpha.*B_prime)))./(He.*sinh(Qe)+cosh(Qe))); %[A/m^2/nm] plot(wavelength,JeD*1e3,'k'); %Black line is JeD. Wavelength in nm, JeD multiplied by 1000 to convert to [A/m^2/um] hold on plot(wavelength,JeB*1e3,'b'); %Blue line is JeB. JeB multiplied by 1000 to convert to [A/m^2/um] xlabel('Wavelength [nm]') ylabel('Spectral Photocurrent Density [A*m^-2*um^-1]') hold off % Total photocurrent J_Total=JeD+JeB; % spectral current density [A/m2/nm] 116 J_ph= trapz(wavelength,J_Total) % Total photocurrent density [A/m2] %Dark current and Open Circuit Voltage J0=q*Nv*vRh*exp(-phiB/Vth); Voc= Vth*log((J_ph+J0)/J0) %Open circuit voltage [V] V_load=0:0.0001:Voc; % Sweep load voltage from 0V to Voc Jdark=J0.*(exp(V_load/Vth)-1); % Dark Current density [A/m^2] J_load=J_ph-Jdark; % Load Current density [A/m^2] P_load=-J_load.*V_load; % Load Power Density [W/m^2] figure plot(V_load,-J_load.*0.1) %Plot Load Current density vs Voltage, change unit to mA/cm^2 xlabel('Load Voltage [V]') ylabel('Load Current Density [mA/cm^2]') A_base = 0.01 * 0.01; %Effective area = 1cm * 1cm figure plot(V_load,P_load.*A_base) %Plot Generated Power vs Voltage xlabel('Load Voltage [V]') ylabel('Load Power [W]') %Code to extract Jmp and Vmp [Y I]= max(abs(P_load)); Pmp=Y; %Power density at maximum power point Vmp=V_load(I); %Load Voltage at maximum power point Jmp=J_load(I); %Load Current density at maximum power point %Fillfactor FF=Pmp/(J_load(1)*Voc) %Conversion efficiency eta=Pmp/(1000) % AM1.5G Pin = 1000W/m2 %************************************************************************** %Including Parasitic Resistance Rshunt=1e12; %Assuming no short circuit at vertical edge of device %Series Resistance = base resistance plus graphene spreading resistance %Selecting Base resistance for semiconductor %p-Silicon base resistance rho_base_Si = 1/(q*Na_Si*mobH_Si); %[ohm*m] R_base_Si = (rho_base_Si*B)/A_base; %[ohm] R_base = R_base_Si; % %p-CIGS base resistance % rho_base_CIGS = 1/(q*Na_CIGS*mobH_CIGS); %[ohm*m] % R_base_CIGS = (rho_base_CIGS*B)/A_base; %[ohm] 117 % R_base = R_base_CIGS; % %p-CdTe base resistance % rho_base_CdTe = 1/(q*Na_CdTe*mobH_CdTe); %[ohm*m] % R_base_CdTe = (rho_base_CdTe*B)/A_base; %[ohm] % R_base = R_base_CdTe; % %p-Amorhpous Silicon base resistance % rho_base_aSi = 1/(q*Na_aSi*mobH_aSi); %[ohm*m] % R_base_aSi = (rho_base_aSi*B)/A_base; %[ohm] % R_base = R_base_aSi; %Graphene spreading resistance h=1e-2; %length between contacts [1 cm] b=1e-2; %length of contacts [1 cm] Rsheet=30; %lowest doped ohms/sq R_spread=Rsheet*h/(12*b); %equiv to base-spreading resistance 2 PARALLEL FINGERS R_series = R_base + R_spread; A=h*b; %cell area between fingers (Same as effective area, not shadowed by contacts) V_diode=0:0.0001:Voc; %sweep diode voltage from 0 to Voc Iph=J_ph.*A; I0=J0.*A; Idark_resistance=I0.*(exp(V_diode./Vth)-1); I_load_resistance=Iph-Idark_resistance-V_diode./Rshunt; % [A] V_load_resistance=V_diode-I_load_resistance.*R_series; % [V] P_load_resistance=-I_load_resistance.*V_load_resistance; % [W] figure plot(V_load_resistance,-I_load_resistance./A.*0.1,'r') %Plot JV graph with resistance hold on plot(V_load,-J_load.*0.1) %Plot JV graph without resistance, change unit to mA/cm^2 hold off axis([0 Voc -J_load(1).*0.1 0]) %Show only relevant part of graph xlabel('Load Voltage [V]') ylabel('Load Current Density [mA/cm^2]') figure plot(V_load_resistance,P_load_resistance,'r') %Plot Power vs Voltage with resistance hold on plot(V_load,P_load.*A_base) %Plot Power vs Voltage without resistance hold off axis([0 Voc -Pmp*A_base 0]) %Show only relevant part of graph xlabel('Load Voltage [V]') ylabel('Load Power [W]') [Yl Xl]= min(P_load_resistance); Pmp_load_resistance=Yl; Vmp_load_resistance=V_load_resistance(Xl); %Fill factor with parasitic resistance 118 FF_resistance = -Pmp_load_resistance/(I_load_resistance(1)*Voc) %conversion efficiency with parasitic resistance eta_resistance = -Pmp_load_resistance/(1000*A) % AM1.5G Pin = 1000W/m2 "@en . "Thesis/Dissertation"@en . "2013-05"@en . "10.14288/1.0073579"@en . "eng"@en . "Electrical and Computer Engineering"@en . "Vancouver : University of British Columbia Library"@en . "University of British Columbia"@en . "Attribution-NonCommercial-NoDerivatives 4.0 International"@en . "http://creativecommons.org/licenses/by-nc-nd/4.0/"@en . "Graduate"@en . "Integrating graphene and nanofibers with silicon to form Schottky junction solar cells"@en . "Text"@en . "http://hdl.handle.net/2429/43933"@en .