USF Libraries
USF Digital Collections

Characterization of cadmium zinc telluride solar cells

MISSING IMAGE

Material Information

Title:
Characterization of cadmium zinc telluride solar cells
Physical Description:
Book
Language:
English
Creator:
Sivaraman, Gowri
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla.
Publication Date:

Subjects

Subjects / Keywords:
czt
widegap semiconductors
sublimation
tandem solar cells
thin films
Dissertations, Academic -- Electrical Engineering -- Masters -- USF   ( lcsh )
Genre:
government publication (state, provincial, terriorial, dependent)   ( marcgt )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Summary:
ABSTRACT: Currently thin film solar cells have efficiencies in the range of 16-18%. Higher efficiencies of 20% or more can be achieved by two junction solar cells in which two p-n junctions are connected in series one on top of the other in a tandem structure. The ideal bandgaps for optimum efficiency in a tandem structure are about 1eV for the top cell and 1.7 eV for the bottom cell. Copper Indium Gallium di-Selenide (CIGS) with a bandgap of 1 eV is a suitable candidate for the bottom cell and Cadmium Zinc Telluride (CZT) with a tunable bandgap of 1.44-2.26 eV is a suitable candidate for the top cell. This work involves characterization of cadmium zinc telluride films and solar cells prepared by close spaced sublimation. CZT is deposited by co-sublimation of CdTe and ZnTe. The process has been investigated on various wide bandgap semiconductor materials including cadmium sulphide, cadmium oxide and zinc selenide. Different post deposition heat treatments were carried out to determine their effect on film and device properties. Characterization of the CZT devices was done using XRD, EDS, SIMS, J-V and spectral response measurements. CZT (Eg~1.7 eV) /CdS exhibited best performance when compared to the other window layers investigated. The best device exhibited Voc=640mV, FF=40% and Jsc=4.5 mA/cm2. The theoretical performance of CZT based solar cells were investigated using SCAPS. The effect of bulk and interface defects on the device parameters were studied.
Thesis:
Thesis (M.S.E.E.)--University of South Florida, 2003.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Gowri Sivaraman.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 70 pages.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001447479
oclc - 54089530
notis - AJN3924
usfldc doi - E14-SFE0000219
usfldc handle - e14.219
System ID:
SFS0024915:00001


This item is only available as the following downloads:


Full Text
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001447479
003 fts
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 040114s2003 flua sbm s000|0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0000219
035
(OCoLC)54089530
9
AJN3924
b SE
SFE0000219
040
FHM
c FHM
090
TK145
1 100
Sivaraman, Gowri.
0 245
Characterization of cadmium zinc telluride solar cells
h [electronic resource] /
by Gowri Sivaraman.
260
[Tampa, Fla.] :
University of South Florida,
2003.
502
Thesis (M.S.E.E.)--University of South Florida, 2003.
504
Includes bibliographical references.
516
Text (Electronic thesis) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 70 pages.
520
ABSTRACT: Currently thin film solar cells have efficiencies in the range of 16-18%. Higher efficiencies of 20% or more can be achieved by two junction solar cells in which two p-n junctions are connected in series one on top of the other in a tandem structure. The ideal bandgaps for optimum efficiency in a tandem structure are about 1eV for the top cell and 1.7 eV for the bottom cell. Copper Indium Gallium di-Selenide (CIGS) with a bandgap of 1 eV is a suitable candidate for the bottom cell and Cadmium Zinc Telluride (CZT) with a tunable bandgap of 1.44-2.26 eV is a suitable candidate for the top cell. This work involves characterization of cadmium zinc telluride films and solar cells prepared by close spaced sublimation. CZT is deposited by co-sublimation of CdTe and ZnTe. The process has been investigated on various wide bandgap semiconductor materials including cadmium sulphide, cadmium oxide and zinc selenide. Different post deposition heat treatments were carried out to determine their effect on film and device properties. Characterization of the CZT devices was done using XRD, EDS, SIMS, J-V and spectral response measurements. CZT (Eg~1.7 eV) /CdS exhibited best performance when compared to the other window layers investigated. The best device exhibited Voc=640mV, FF=40% and Jsc=4.5 mA/cm2. The theoretical performance of CZT based solar cells were investigated using SCAPS. The effect of bulk and interface defects on the device parameters were studied.
590
Adviser: S.Ferekides, Dr.Christos
653
czt.
widegap semiconductors.
sublimation.
tandem solar cells.
thin films.
690
Dissertations, Academic
z USF
x Electrical Engineering
Masters.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.219



PAGE 1

Characterization Of Cadmium Zinc Telluride Solar Cells by Gowri Sivaraman A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Electrical Engineering Department of Electrical Engineering College of Engineering University of South Florida Major Professor: Christos S. Ferekides, Ph.D. Don L.Morel, Ph.D Yun.L.Chiou, Ph.D Date of Approval: November 12, 2003 Keywords: czt,thin films, tandem solar cells,sublimation,widegap semiconductors Copyright 2003 Gowri Sivaraman

PAGE 2

DEDICATION This thesis is dedicated to my family and friends for their everlasting love and support.

PAGE 3

ACKNOWLEDGMENTS I would like to thank my Major Professor, Dr. Christos S. Ferekides, for his guidance and support during the course of th is work. His enthusiasm and profound knowledge in the subject has been a great sour ce of inspiration. I would like to thank Dr. Don L. Morel and Dr. Yun L. Chiou for their guidance in the fulfillment of this thesis. I would like to express my special thanks to Su Yu whose invaluable guidance has been of great help to me. I would also like to thank all the fellow colleagues of the Thin Film Electronic Materials laboratory for their help and support.

PAGE 4

TABLE OF CONTENTS LIST OF TABLES iv LIST OF FIGURES v ABSTRACT viii CHAPTER 1 INTRODUCTION 1 CHAPTER 2 SEMICONDUCTORS AND SOLAR CELLS 4 2.1 p-n Junction 5 2.1.1 Heterojunction 8 2.2 Solar Cells 9 2.3 PhotoVoltaic Effect 10 2.4 Operation 10 2.5 Equivalent Circuit 12 2.6 Tandem Solar Cells 13 CHAPTER 3 OVERVIEW OF MATERIALS 15 3.1 Cadmium Telluride 16 3.2 Zinc Telluride 16 3.3 Cadmium Zinc Telluride 17 3.4 Cadmium Sulphide 19 3.5 Zinc Selenide 19 3.6 Cadmium Oxide 20 i

PAGE 5

CHAPTER 4 CZT SOLAR CELLS 22 4.1 CdS/CZT Structure 22 4.1.1 Cadmium Sulphide 23 4.1.2 Close Spaced Sublimation 24 4.1.3 Back Contacts 25 4.1.3.1 Zinc Telluride 25 4.1.3.2 Graphite Back Contact 26 4.2 Heterojunction Structures Investigated 26 CHAPTER 5 RESULTS AND DISCUSSION 27 5.1 CZT on ZnSe Films 27 5.1.1 Microstructures of CZT/ZnSe Films 29 5.1.2 SIMS Analysis 29 5.2 CZT/ZnSe Solar Cells 30 5.3. CZT/CdO Cells 33 5.4 CZT/CdS Films 35 5.4.1 Effect of Deposition Conditions 35 5.4.2 SIMS Analysis 38 5.5 Characterization of CZT/CdS Solar Cells 39 5.5.1 Effect of CdS Thickness 39 5.5.2 Effect of Annealing 41 5.5.3 Analysis of the Best Device 43 CHAPTER 6 SCAPS SIMULATION 46 6.1 Effect of Bulk Traps on Device Performance 50 ii

PAGE 6

6.2 Effect of Interface States 51 CHAPTER 7 CONCLUSION 55 REFERENCES 57 iii

PAGE 7

LIST OF TABLES Table 1. Lattice Constants 18 Table 2. Properties of Relevant Materials 21 Table 3. V OC s of CZT/ZnSe Devices 31 Table 4. Effect of Ambient on Composition 35 Table 5.Composition of a CZT Film with E g ~1.6 eV 37 Table 6. CZT Devices in Various Annealing Ambient 42 Table 7. Comparison of As-Deposited & H 2 Annealed Devices 43 Table 8. Tabulation of Results 45 Table 9. Input Values for SCAPS Simulation of CZT 47 iv

PAGE 8

LIST OF FIGURES Figure 1. Electric Field Distribution of a p-n Junction 6 Figure 2. Energy Band Diagram of p-n Junction 7 Figure 3. Energyband of Heterojunction at Equilibrium 8 Figure 4. The Solar Spectrum 9 Figure 5. Dark and Light I-V Curves of an Ideal Solar Cell under AM 1.5 Spectra 11 Figure 6. Equivalent Circuit of Solar Cell 12 Figure 7. Two Terminal Tandem Structure 13 Figure 8. Tandem Structure 14 Figure 9. Zinc Blende Structure 17 Figure 10. Structure of CdS/CZT Solar Cell 22 Figure 11.Experimental Setup for Chemical Bath Deposition 23 Figure 12. Experimental Setup of Close Spaced Sublimation 25 Figure 13. Structure of ZnSe Based Devices 26 Figure 14. Structure of CdO Based Devices 26 Figure 15. Transmission of a CZT Film 28 Figure 16. Energy Band Diagram of CZT Film 28 Figure 17. SEM of CZT Film on ZnSe 29 Figure 18. SIMS of CZT/ZnSe [24] 30 Figure 19. Light and Dark J-V of CZT/ZnSe Devices 32 v

PAGE 9

Figure 20. Spectral Responses of CZT/ZnSe Devices 32 Figure 21. Light and Dark J-Vs of CZT/CdO Devices 33 Figure 22. Spectral Response of CZT/CdO Devices 34 Figure 23. XRD of CZT 36 Figure 24. SEM Image of CZT 37 Figure 25.SIMS on ZnTe/CZT/CdS Structure 38 Figure 26. V OC Vs CdS Thickness 39 Figure 27. FF vs CdS Thickness 40 Figure 28. Spectral Response of CZT Devices 40 Figure 29. Effect of Annealing Ambient on V OC and FF 41 Figure 30. SR of CZT Devices for Various Annealing Ambient 42 Figure 31. SR of H 2 -anneal and As-deposited Devices 43 Figure 32. Spectral Response of the Best Device 44 Figure 33. Dark J-V of the Best Performance Device 44 Figure 34. Light J-V of the Best Device 45 Figure 35. Effect of Work Function on V OC Simulated Results 48 Figure 36. Effect of Work Function on Fill FactorSimulated Results 48 Figure 37. Light I-V Characteristics of Various Work Functions @ Na=10 14 cm -3 49 Figure 38. Energy Band Diagram for Different Work Functions in Dark @ Zero Bias 49 Figure 39. V OC vs Trap Density for Different Capture Cross Section Ratios 50 Figure 40. Effect of N T on V OC Jsc and FFs 51 Figure 41. V OC vs Trap Density for Various Acceptor Interface Defects 52 Figure 42. Current Density vs Trap Density for Various Acceptor Interface Defects 52 vi

PAGE 10

Figure 43. Device Performance for Various Energy Levels 53 Figure 44. Simulated Results of QE for Acceptor Interface Defects 53 vii

PAGE 11

CHARACTERIZATION OF CADMIUM ZINC TELLURIDE SOLAR CELLS Gowri Sivaraman ABSTRACT Currently thin film solar cells have efficiencies in the range of 16-18%. Higher efficiencies of 20% or more can be achieved by two junction solar cells in which two p-n junctions are connected in series one on top of the other in a tandem structure. The ideal bandgaps for optimum efficiency in a tandem structure are about 1eV for the top cell and 1.7 eV for the bottom cell. Copper Indium Gallium di-Selenide (CIGS) with a bandgap of 1 eV is a suitable candidate for the bottom cell and Cadmium Zinc Telluride (CZT) with a tunable bandgap of 1.44-2.26 eV is a suitable candidate for the top cell. This work involves characterization of cadmium zinc telluride films and solar cells prepared by close spaced sublimation. CZT is deposited by co-sublimation of CdTe and ZnTe. The process has been investigated on various wide bandgap semiconductor materials including cadmium sulphide, cadmium oxide and zinc selenide. Different post deposition heat treatments were carried out to determine their effect on film and device properties. Characterization of the CZT devices was done using XRD, EDS, SIMS, J-V and spectral response measurements. CZT (Eg~1.7 eV) /CdS exhibited best performance when compared to the other window layers investigated. The best device exhibited Voc=640mV, FF=40% and Jsc=4.5 mA/cm 2 The theoretical performance of CZT based solar cells were investigated using SCAPS. The effect of bulk and interface defects on the device parameters were studied. viii

PAGE 12

CHAPTER 1 INTRODUCTION Globalization came about as the industrialized countries began to exploit global fossil fuel and mineral reserves in pursuit of greater economic efficiency. There is an increasing demand for a cost effective renewable source of energy. Fossil fuels are assumed to offer low prices and greater potential. Coal, Oil and Gas are called "fossil fuels" because they have been formed from the fossilized remains of prehistoric plants and animals. They provide around 66% of the world's electrical power, and 95% of the world's total energy demands. But the basic disadvantage of fossil fuels is pollution. Burning any fossil fuel produces carbon dioxide, which contributes to the "greenhouse effect" adding to global warming. The fundamental economic reality of fossil fuels is that they are found in relatively small number of locations across the globe, yet consumed everywhere [7]. This has led to an increased interest in renewable sources of energy such as solar, hydroelectric and wind. Solar power is one of the world's fastest growing renewable energy sources, offering a potentially endless supply of power generation capable of meeting the electricity demands of the entire world. Sun is one of the most important non-conventional sources of energy. The radiative energy from the sun derives from a nuclear fusion reaction. The intensity of solar radiation in free space, at the average distance of earth from the sun is defined as the solar constant and has a value of 1353 W/m 2 [1]. The 1

PAGE 13

mechanism of converting sunlight directly into electrical energy is known as the photovoltaic effect. This effect is demonstrated in solid state devices known as solar cells. A solar cell is a p-n junction that absorbs light, releases electrons and holes, thus creating a voltage. For larger applications an array of cells connected in series or in parallel depending on the requirements can be used. To be economically competitive with conventional fossil fuels for terrestrial applications, solar cell arrays with conversion efficiencies in excess of 15% are required [1]. The theoretical band gap for a maximum efficiency of solar cell lies in the range of 1.2-1.8 eV. It has been shown from the solar spectrum and semiconductor characteristics that the conversion efficiencies of homojunction solar cells are maximum at a bandgap of about 1.5 eV. CdTe, a direct band gap semiconductor with a bandgap of 1.45 eV is ideally suited for solar energy conversion. Direct band gap semiconductors have high absorption coefficients and hence can be used in thin films form. Materials like CdTe, CIS, and CdSe are direct band gap semiconductors. CdTe single junction solar cells fabricated from polycrystalline materials have demonstrated efficiencies of more than 15% [19] [26]. Single junctions of CIS exhibit efficiencies of around 18% [18]. But higher efficiencies of 20-30% can be achieved by two junction solar cells in which two p-n junctions are connected in a tandem structure. The ideal band gaps for optimum efficiency in a tandem structure are about 1 eV for the bottom cell and 1.7 eV for the top cell [2]. Since about two thirds of the output comes from the top cell in a tandem structure, this requires a top cell efficiency of 16 18%[2]. Therefore Cadmium Zinc Telluride (CZT) films with a bandgap of 1.7 eV are suited for this application. CdTe and ZnTe, direct gap semiconductors with room temperature 2

PAGE 14

bandgap energies of 1.44 and 2.26 eV respectively, form a continuous series of solid solutions of Cd 1-x Zn x Te with tunable bandgap depending on the composition x. The lattice parameter of the system follows Vegards law [9]. This work involves processing and characterization of CZT solar cells by close spaced sublimation. The close spaced sublimation process offers several advantages for manufacturing purposes such as ease of scale up, high throughput and efficient material utilization. The highest cell efficiencies reported to date for CdS/CdTe solar cells were fabricated by CSS [26]. Various window materials used for this work include wide band gap materials like CdS, CdO and ZnSe. The objective of this thesis is to investigate the performance of CZT on various window layers. Films and devices have been characterized using x-ray diffraction, optical transmission, dark and light J-V and spectral response measurements. Numerical simulation based on SCAPS was carried out to investigate the effect of bulk and interface defects. 3

PAGE 15

CHAPTER 2 SEMICONDUCTORS AND SOLAR CELLS A solar cell is formed by a light sensitive p-n junction semiconductor. Each photon has an energy associated with it. The charge carriers in a semiconductor are holes and electrons. If the energy of the photon incident on the semiconductor is greater than or equal to the energy required to release the electron it will contribute to the output of the solar cell. Semiconductors exist both in elemental and compound form and can be found in single crystalline, polycrystalline or amorphous structures. Atoms in single crystalline materials are arranged in a well defined order. Polycrystalline materials consist of randomly oriented small single crystal grains whereas amorphous materials do not have any defined ordered structure. The specific properties of a semiconductor depend on the impurities, or dopants added to it [4]. A semiconductor doped in such a way that the hole carrier concentration is greater than the electron carrier concentration is said to p-type. An acceptor is an impurity introduced into the semiconductor to generate a free hole by accepting an electron from the semiconductor. An n-type semiconductor has a higher electron carrier concentration than the hole carrier concentration. A donor is an impurity that is added to a semiconductor to generate free electrons by donating an electron. The bandgap (E g ) is the separation between the energy of the lowest conduction band and that of the highest valence band. The Fermi level (E f ) is defined as the energy level at which 4

PAGE 16

the probabability of occupation by an electron is half [4]. For a p-type semiconductor, the fermi level is close to the valence band whereas for an n-type semiconductor the Fermi level is close to the conduction band. 2.1 p-n Junction p-n junctions are the basic building blocks of electronic devices A p-n junction is formed when a p-type material and an n-type material are placed in contact. Free electrons from the n-type semiconductor diffuse into the p-type semiconductor and holes diffuse from the p-type semiconductor into the n-type semiconductor. This movement of charge carriers constitutes the diffusion current. As the electrons diffuse into the p-type semiconductor, they leave behind positively charged atoms causing part of the n-type semiconductor as a whole to become positively charged. Similarly, as holes diffuse out of the p-type semiconductor, they leave the p-type material negatively charged. Thus, as diffusion of holes and electrons proceeds, static charges build up both in the p-type and n-type materials close to the junction. This charged area is called the space charge region. This charge also sets up an electric field that leads to a second current component called the drift current. The drift current flows in the opposite direction of the diffusion current. Initially the diffusion current dominates the drift current, but with further diffusion, the potential causing the drift current builds up, causing larger drift currents until eventually the drift current matches the diffusion current, and they balance each other out. The electric field developed at the junction by the ionized donors and acceptors is given by E V (1) For one dimension, the electric field is given by dxdVE (2) 5

PAGE 17

The figure below shows the electric field and the space charge distribution in a p-n Junction. P N (a) Ionized Donors Q + Q x (b) Ionized Acceptors x (c) E(x) Figure 1. Electric Field Distribution of a p-n Junction This electric field limits the diffusion process bringing the junction to equilibrium. Under equilibrium conditions J diff + J drift =0. The total current in the p-n junction [4] accounting for both the electrons and holes is given by 1exp0nAkTqVII (3) 6

PAGE 18

where I 0 is the reverse saturation current and is given by npnpnpLnDLpDqAI000 (4) Where D is the diffusion constant and L is the diffusion length of each carrier type, P no and n p0 are the majority carrier concentrations on the p and n side; A is the diode quality factor. The value of A depends on the current transport. For injection/diffusion mechanism, n=1 and for recombination mechanism n=2.The energy band diagram of a p-n homojunction is shown in the figure below. Figure 2. Energy Band Diagram of p-n Junction At equilibrium the fermi level lines up by band bending with a built in electric field created between the junctions. V bi is the built-in potential and E i is the intrinsic fermi energy. The built in potential V bi with carrier concentration is given by the relationship n i is given by 7

PAGE 19

V bi = qKT ln 2indaNN (5) 2.1.1 Heterojunction When a narrow bandgap p-type semiconductor and a larger bandgap n-type semiconductor are brought in contact a heterojunction is formed. Figure 3. Energyband of Heterojunction at Equilibrium The energy band diagram of the heterojunction is shown in figure 3. Two materials with different bandgaps E g1 & E g2 electron affinities x1& x2 and work functions m1 and m2 are brought together forming a heterojunction. In the above figure E c =(x1-x2) and E v =E g1 -E g2 E c represent the discontinuities in the conduction and the valence band 8

PAGE 20

induced due to the different electron affinities, work functions and the bandgaps of the materials. In solar cells the wider bandgap materials form the window layers and the narrow bandgap materials form the absorber. 2.2 Solar Cells The Sun radiates its energy from its surface to the outer space. The solar cell is based on the photovoltaic (PV) effect in which light falling on a two layer device produces a photo voltage or potential difference between the layers. This voltage is capable of driving a current through an external circuit and thereby producing useful energy. Figure 4. The Solar Spectrum The suns radiation can be described by a black body radiator at about 6000 K. The spectrum as seen from outside the earths atmosphere is referred to as the Air Mass Spectrum (AM0) and this closely fits the black body spectrum at 5800 K. The total power 9

PAGE 21

density is 1385 W/m 2 The solar spectrum as observed on the earth is modified due to absorption in the atmosphere and is known as AM 1.0 as shown in the figure 4. For AM1 the power density is reduced to 925 W/m 2 Another standard AM 1.5 yields a power density of 844 W/m 2 The irregularities in the AM spectra are due to the absorption of gasses including O 2 H 2 CO 2 which occurs at specific photon energies. Solar cells are tested depending on type of application. For example, cells developed for space applications are tested in AM 0 spectra and terrestrial applications in AM 1.5 spectra. 2.3 PhotoVoltaic Effect If a photon is incident upon a semiconductor and if the photon energy is greater than the bandgap energy, the photon is absorbed and an electron is elevated from the valence band to conduction band. This creates an electron-hole pair. Mobile carriers produced in this manner are called photo-carriers and are in excess of equilibrium concentration. Excess carriers in a p-n junction can recombine or get collected. Recombination is the annihilation of an electron hole pair. The term collection refers to electrons and holes separated by an electric field leading to photocurrent (I L ). 2.4 Operation In dark, a solar cell is just a diode. The current-voltage characteristics of a diode has been given earlier. Figure 5 shows the I-V characteristics of an ideal solar cell. Under illumination the J-V characteristics are shifted due to the light generated current I L Therefore under illumination the current-voltage relationship is given by the following equation. 10

PAGE 22

LIAkTqVII]1)[exp(0 (6) The open-circuit voltage (V OC ) is the voltage output at the device terminals with infinite load attached to it and the short circuit current is the current at zero voltage. The figure below shows the V OC I sc and peak power P max at V p and I p. Voc Dark I (A) Vp Ip Light Isc V (V) Figure 5. Dark and Light I-V Curves of an Ideal Solar Cell under AM 1.5 Spectra The outer rectangle in the fourth quadrant has an area of V OC Isc and inner rectangle has the area of V p I p The Fill Factor is defined as the ratio of areas of these two rectangles and is given by the following formula scIocVPFFmax (7) scOCpPIVIV (8) 11

PAGE 23

The efficiency of the solar cell is a measure of conversion of light energy into electricity and is given by = inppPIV (9) Where Pin is the incident power. 2.5 Equivalent Circuit The figure below shows the equivalent circuit of a practical solar cell. An ideal solar cell should have an infinite shunt resistance (R sh ) and zero series resistance (R s ). Unlike an ideal solar cell, R s and R sh are present in a practical solar cell. The leakage currents across the perimeter or in the bulk can lead to a low shunt resistance. A low shunt resistance leads to poor fill factors and lower V OC s. In single crystals the leakage current is low and insignificant. But in thin films, grain boundaries act as sources of shunting paths leading to high leakage currents. Series resistance (R s ) is mainly affected by the bulk resistivity and the contact resistance. Rs V IL Rsh Figure 6. Equivalent Circuit of Solar Cell 12

PAGE 24

2.6 Tandem Solar Cells The ideal efficiency of single junction solar cells is limited. Higher efficiencies for thin films in range of 20-30% can be obtained by the fabrication of tandem structures. Figure 7 shows a two terminal tandem structure of InP/InGaAsP [17]. The disadvantage of a two terminal device is that the substrate and the superstrate configuration are always connected in series. This leads to the need of transparent tunneling contacts between the two cells and the need to match the currents. If one of the cells gives a lower current the overall current from the tandem structure is low. Figure 7. Two Terminal Tandem Structure Tandem junction solar cells can consist of a superstrate structure (top cell) and a substrate structure (bottom cell) glued together. Figure 8 shows a four terminal tandem structure with CIGS as the bottom cell and CZT as the top cell [2]. CZT with a tunable 13

PAGE 25

band gap of 1.45-2.2 eV can be the material of choice for the top cell. When light is incident on the top cell photons with energies greater than 1.7 eV are absorbed. The photons with less energy are transmitted and absorbed by the bottom cell which has an absorber with a band gap of 1 eV. The projected efficiencies are 30-40% for these tandem structures [2]. Figure 8. Tandem Structure 14

PAGE 26

CHAPTER 3 OVERVIEW OF MATERIALS Cadmium telluride and its ternary alloy Cd 1-x Zn x Te are important semiconductor materials used in solar cells, gamma ray detectors, electro-optical modulators and other optoelectronic devices [10]. Cadmium Zinc Telluride layers have promising applications as optoelectronic devices since the bandgap is tunable from the visible to the infrared portion of electromagnetic spectrum. High efficiency thin film solar cells with efficiencies greater than 20% can be achieved with tandem structures. A two terminal tandem solar cell is composed of two single junction solar cells with different bandgaps. Considerable progress has been made in the area of the bottom cell with reported efficiencies of 18% [18]. The material for the top cell is the area of interest. CZT and CdSe are two wide bandgap materials of choice for the top cell. Efficiencies as high as 16% have been reported for CdTe solar cells. CZT, an extension of CdTe is the typical material of choice for the top cell. The top cell by itself is a single junction solar cell. To make a good heterojunction partner it is necessary to understand the properties of various window layers. A window layer is a transparent conducting semiconductor used as a heterojunction partner in solar cells. Different wide bandgap materials like CdS, ZnSe, and CdO are n-type and can be considered as candidates for the window layer. The p-type CZT is the absorber. 15

PAGE 27

3.1 Cadmium Telluride Cadmium Telluride (CdTe) forms crystals in zincblende structure with a lattice parameter of 6.48. Each Cd atom is tetrahedrally surrounded by four Te atoms and vice versa. Theoretically semiconductors with a bandgap of 1.5 eV are ideally suited for solar cell applications [4]. CdTe, a direct bandgap semiconductor is a promising material for high efficiency thin film solar cells because of its near optimum 1.45 eV bandgap, ease of deposition and strong optical absorption. A 1-2 m thick film of CdTe is sufficient for conversion of sunlight into electricity. CdTe is a defect semiconductor with a dielectric constant of 10.2. It can be doped both p and n-type. CdTe can be deposited by different techniques like physical vapor deposition, electro deposition, sputtering, close spaced sublimation, screen printing and metal organic chemical vapor deposition. CdTe films grown by closed space sublimation have reported the efficiency of higher than 15% [19]. Single Crystalline CdTe finds applications in nuclear detectors, solar cells and electro-optical modulators. 3.2 Zinc Telluride Zinc Telluride (ZnTe) is a reddish crystalline solid with two different structures namely hexagonal and zinc blende structure. It has a wide bandgap of 2.4 eV with a dielectric constant of 8.7. ZnTe is widely used in IR optics, solar cells, gamma ray detectors and waveguide modulators. Nitrogen doped ZnTe produced by reactive sputtering can be used as a back contact for CdTe solar cells. Since ZnTe is transparent to photons below 2.2 eV it is a candidate for the back contact/tunnel junction in tandem solar cells. Efficiencies of 10-11% have been achieved with ZnTe as a back contact on 16

PAGE 28

CdTe solar cells [23]. ZnTe can be deposited by sputtering, closed space sublimation, MOCVD etc. 3.3 Cadmium Zinc Telluride Cadmium Zinc Telluride crystallizes in zinc blende structure showing crystallographic polarity along the <111> direction as shown in figure 9. Te Figure 9. Zinc Blende Structure There are two types of {111} alternate planes, one formed by randomly distributed Cd and Zn atoms and the other formed by Te atoms[10]. The control and monitoring of Zn concentration, x is crucial for obtaining a CZT alloy with the required 17

PAGE 29

lattice parameter. The Zn concentration x is governed by Vegards law which follows the lattice parameter. The concentration of zinc x can be calculated from the lattice constant relation given by a=6.481-0.3837(x) (10) Where a is the lattice constant. The lattice parameter for CdTe and ZnTe are 6.481 and 6.104 respectively. Some lattice constants of CZT with different zinc concentrations are as follows. Table 1. Lattice Constants Compound Lattice Parameter() Cd 0.95 Zn .05 Te 6.495 Cd .7 Zn .3 Te 6.364 Cd .5 Zn .5 Te 6.287 By alloying CdTe with ZnTe, the bandgap and the lattice parameter can be tuned over a wide range. The CdTe bond is strengthened by the addition of few atomic percent of ZnTe. CZT can be deposited by different techniques. The most commonly used bulk methods for the growth of CZT are Bridgman method and the Travelling Heater method. Single Crystalline CZT grown by High pressure Bridgman method can be used to fabricate high performance nuclear radiation detectors. For solar cell applications CZT has been deposited by sputtering, molecular beam epitaxy [11]. To make a good heterojunction partner with this absorber it is necessary to understand the properties of various window materials like CdS, ZnSe and CdO. These are discussed in the sections below. 18

PAGE 30

3.4 Cadmium Sulphide Cadmium Sulphide (CdS) exists in two crystalline forms namely zinc blende and wurtzite structure with alternating Cd and S atoms. Each Cd atom is surrounded by four S atoms in a hexagonal close packed structure. The S vacancies act as donors and make the CdS n-type. CdS is a wide bandgap material with a bandgap of 2.42 eV and a lattice constant of 5.8 The material chosen for the window layer must be transparent to most of the solar spectrum. The n-type conductivity and transparency makes the CdS a well suited window material for thin film CdTe solar cells. To minimize the absorption of light in CdS, the layer is made as thin as possible. CdS films can be prepared by a variety of techniques, such as chemical bath deposition, vacuum evaporation, electro deposition, laser evaporation, molecular beam epitaxy, RF sputtering and spray pyrolysis [14]. Of these, Chemical Bath Deposition (CBD) has proved to be simple, inexpensive and safe, and has been used to produce large area homogeneous films. The deposition parameters such as temperature, pH and composition, control the stoichiometry and structure of the films. CdS is grown in a bath containing cadmium acetate, ammonium acetate, ammonium hydroxide and thiourea [21]. 3.5 Zinc Selenide Zinc Selenide (ZnSe) is an opto-electronic semiconducting material with a wide bandgap of 2.7 eV. The properties such as large bandgap, low resistivity and high photosensitivity make this material highly attractive. It can substitute CdS in photovoltaic solar cell devices, aiming at higher cell efficiency by means of admission of more photons to the absorber layer, because of its larger band gap. ZnSe exists in two 19

PAGE 31

crystalline forms: wurtzite (hexagonal) and zinc blende (cubic), of which the cubic phase is believed to be stable. ZnSe is widely used in thermal imaging systems, ultrasonic transducers, photo detectors, solar cells, semiconductor diode lasers and blue light emitting devices. A variety of methods can been employed for the growth of high quality ZnSe thin films such as molecular beam epitaxy, solution growth, organometallic chemical vapor deposition, spray pyrolysis, sputtering and close spaced sublimation. ZnSe is used as a buffer layer in CdTe solar cells. The ZnSe/CdTe solar cells grown by close spaced sublimation have yielded fairly good results with an efficiency of 11% [5]. The grain size of the CSS ZnSe depends on the surface properties of the substrate, the substrate temperature, the total reactor pressure and the initial growth conditions [5]. 3.6 Cadmium Oxide Cadmium Oxide (CdO) is an n-type semiconductor with NaCl structure. It has a direct optical bandgap of 2.3 eV. It has a high electrical conductivity that is attributed to moderate electron mobility and higher carrier concentration due to the contribution from shallow donors resulting from inherent non-stoichiometry. Nonstoichiometric undoped CdO thin films usually exhibit low resistivity due to native defects of oxygen vacancies and cadmium interstitials [13]. Optically, it is a transparent material over energies of 540 nm in the solar spectrum. CdO films find wide applications in the field of optoelectronic devices such as solar cells, photo-transistors, photo-diodes and transparent electrodes. Thin films of CdO can be prepared by employing various physical and chemical thin film deposition techniques such as reactive evaporation, spray pyrolysis, solution growth and sputtering [22]. CdO/CdTe heterostructure solar cells with efficiencies of 9.1% have been 20

PAGE 32

fabricated by DC magnetron sputtering [22]. The optical properties of various window materials used in this work are tabulated in the table below. Table 2. Properties of Relevant Materials Material CdS ZnSe CdO Structure Zincblende Zincblende NaCl Lattice Constant () 5.8 5.6 4.7 Bandgap (eV) 2.42 2.7 2.3 Electron Affinity (eV) 4.5 4.1 4.5 21

PAGE 33

CHAPTER 4 CZT SOLAR CELLS Tandem solar cells are the path to the highest efficiencies for thin film photovoltaics. By alloying CdTe with ZnTe the bandgap of CdZnTe and the lattice parameter can be tuned over a large range. The window layers used for this work include Cadmium Sulphide, Zinc Selenide and Cadmium Oxide. 4.1 CdS/CZT Structure The structure of the CZT solar cell with CdS as the window layer is given below. Graphite Back Contact ZnTe CdZnTe CdS Sno 2 7059 Glass Light Figure 10. Structure of CdS/CZT Solar Cell Light enters the cell through the transparent glass substrate (Corning 7059).This type of configuration is called superstrate configuration. The first step in processing is the deposition of tin oxide on glass. This is done by metal organic chemical vapor deposition 22

PAGE 34

(MOCVD) technique at 450C.Tetramethyl Tin serves as the source for tin and halocarbon 13B1 serves as the source for fluorine. The transparent conducting oxide, tin oxide with a bandgap of 3.5 eV serves as the front contact for the solar cell. 4.1.1 Cadmium Sulphide The polycrystalline cadmium sulphide is n-type doped and serves as the n-type heterojunction partner. Being a wide bandgap material with an E g of 2.4 eV, it is transparent down to wavelengths of around 515 nm and is referred to as the window layer. Figure 11. Experimental Setup for Chemical Bath Deposition Chemical bath deposition has proven to be the most efficient method to produce homogeneous CdS thin films since it is an efficient, cost effective and large scale method. The solution growth process is based on the reaction of cadmium salt, a complexing agent and a sulphur compound in a heated aqueous solution. Figure 11 shows the Chemical Bath Deposition (CBD) setup .A solution containing measured amounts of 23

PAGE 35

Cadmium acetate (CdAc), ammonium acetate (NH 4 Ac) and ammonium hydroxide (NH 4 OH) is prepared along with thiourea. The SnO 2 coated substrates are immersed in the bath containing water and the solution is heated from 80 to 90C by a coolant. The deposition parameters such as temperature, time and composition control the stoichiometry and structure of the film [21]. The amount of particulates on the surface of the solution is minimized by continuous stirring. The films deposited on Sno 2 : F are more adherent than on glass. The thickness of CdS increases with increase in time of deposition. This work involves the use of various thicknesses of CdS. 4.1.2 Close Spaced Sublimation Close Spaced sublimation is a process by which source and the substrate are separated by a small distance, in which the source is sublimed at higher temperatures. For this work CZT has been deposited by the co-sublimation (CO-CSS) of CdTe and ZnTe. Figure12 shows the CSS chamber used for CdZnTe deposition. The graphite plate with holes is used as the source. The CdTe and ZnTe are filled in the holes in such a way that uniform deposition in good stochiometric form could be obtained. The substrate is supported by a graphite plate. The thicknesses of the spacers used were typically in the range of 7-8 mm. The thermocouple inserted in the graphite plates read the temperature of the source and the substrate. The source is heated to higher temperatures of 600-700C and the substrate is heated to temperatures in the range of 400-600C by a quartz lamp. The deposition takes place in Helium (He) ambient. Prior to the deposition process, the chamber is evacuated using a mechanical pump and purged with He several times. Subsequently the necessary amount of He is filled in the chamber, 24

PAGE 36

which is made up of quartz tubing. The rate of deposition is controlled by temperature of the source and the substrate, pressure of the gas and the distance between the source and the substrate. Figure 12. Experimental Setup of Close Spaced Sublimation 4.1.3 Back Contacts 4.1.3.1 Zinc Telluride Zinc Telluride is deposited by closed space sublimation in an inert ambient of He over CZT. The ZnTe source is prepared from 99.99% pure Zinc Telluride chunks. For the deposition of ZnTe by CSS, with T src =600-700C; T sub =500-600C. 25

PAGE 37

26 4.1.3.2 Graphite Back Contact Following the ZnTe/CZT deposition they are contacted with doped graphite paste. This provides the back contact for the cell. The Sno 2 is exposed by scribing the area around the graphite that forms the front contact of the cell. 4.2 Heterojunction Structures Investigated Graphite Back Contact CZT ZnSe Sno2 7059 Glass Light Figure 13. Structure of ZnSe based Devices Graphite Back Contact CZT CdO CdS Sno2 705 9 G lass Light Figure 14. Structure of CdO based Devices

PAGE 38

CHAPTER 5 RESULTS AND DISCUSSION This Chapter discusses the properties of CZT thin films and solar cells prepared by Co-CSS of CdTe and ZnTe. Characterizations of CZT films were done using Scanning Electron Microscopy (SEM), Optical Transmission, X-ray Diffraction Spectroscopy (XRD), Energy Diffraction Spectroscopy (EDS) and Secondary Ion Mass Spectroscopy (SIMS). The thickness of the CZT film was obtained using a Dektak profilometer. Transmission measurements were performed using a spectrophotometer. Transmission of the films was referenced with zinc selenide/cadmium sulphide/tin oxide coated substrates. Characterization of devices was done using J-V and spectral response (SR) measurements. J-V measurements were done using an AM 1.5 simulator. 5.1 CZT on ZnSe Films All CZT/window structures were characterized using optical transmission as a quality control to ensure that the CZT absorber was in the 1.6-1.7 eV range. This was done due to the fact that aging of the source material after several depositions could result in loss of composition control. The transmission of a typical CZT film is shown in figure 15. This particular CZT film was deposited on zinc selenide. The transmission curve shows a sharp cutoff at 710 nm. The CZT/ZnSe film had an overall transmission response of 60%. Figure 16 shows a plot of 2 vs Energy used to determine the optical 27

PAGE 39

bandgap. The Bandgap (E g ) is determined by the extrapolation of the tangent of the curve to the x axis and in this case it was found to be 1.67 eV. 0510152025303540455055606570500550600650700750800850900950100010501100Wavelength(nm)T(%) Figure 15. Transmission of a CZT Film 0.E+001.E+082.E+083.E+084.E+085.E+086.E+087.E+088.E+089.E+081.601.621.641.661.681.701.721.741.761.781.801.821.841.861.881.90Energy(eV)2( cm-2) Figure 16. Energy Band Diagram of CZT Film 28

PAGE 40

5.1.1 Microstructures of CZT/ZnSe Films Cadmium Zinc Telluride films on ZnSe were deposited at a substrate temperature of 550C. Figure 17 shows the SEM image of a CZT film on ZnSe. From the SEM image it is clear that the CZT film is dense and pinhole free. The grain size was approximately 1.5-2 m. Only films with E g in 1.6-1.75 range were fully processed into cells. Figure 17. SEM of CZT Film on ZnSe 5.1.2 SIMS Analysis The SIMS depth profile of CZT on ZnSe film was done at the National Renewable Energy Laboratory (NREL) and the results are shown in figure 18[24]. The profile shows elemental composition of the CZT/ZnSe films. Elements such as S, Fe, Na etc. were below the detection limits of the measurement and were removed for clarity. The interface between CZT and ZnSe can be identified by a sharp increase in the Se and Zn signal at about 4 m which was the targeted thicknesses of these films. The Te profile 29

PAGE 41

remains constant but the cadmium and zinc show an increase from the top surface into the film. It is not clear whether this grading is an artificat of the measurement. With the tellurium counts remaining constant, an increase in the zinc concentration should be accompanied by a decrease in cadmium counts for stochiometric films. The high counts of selenium near the surface suggest the formation of Cd x Zn 1-x TeSe. 1E+001E+011E+021E+031E+041E+051E+061E+071E+0802468Depth (microns)Count 10 s 64Zn80Se111Cd119Sn125TeCZT-ZnSe.SWFCZT/ZnSe NREL Figure 18. SIMS of CZT/ZnSe [24] 5.2 CZT/ZnSe Solar Cells The processing of CZT/ZnSe devices was categorized into four groups. Type I as-deposited, Type IIHe annealed, Type III He+O 2 annealed and Type IV-H 2 annealed. 30

PAGE 42

All annealing experiments were performed at 400C. Table 3 shows the V OC range of all four groups. The as-deposited devices had the lowest V OC range when compared to the other types. This suggests that the interface/junction properties that to a great extent determine V OC may have improved after annealing. Table 3. V OC s of CZT/ZnSe Devices Types V OC (mv) Type I(as-deposited) 400-450 Type II(He anneal) 570-600 Type III(He+O 2 anneal) 570-600 Type IV(H 2 anneal) 400-500 The Light and Dark J-V curves are shown in figure 19. The fill factors of the devices for all types were in range of 24-25%. The J-V curves show excessive shunting and high series resistances leading to poor fill factors. The high series resistance could be due to highly compensated CZT resulting in high resistivity and/or contact barrier leading to some of the J-V curves in figure 19. It is not clear what causes the shunt since the films were pinhole free in CZT. Pinholes in the ZnSe surface could act as a source of shunting paths leading to poor junction properties. The spectral response of all the types of devices processed is shown in figure 20. The J sc of all the devices calculated from the SR was less than 2 mA/cm 2 which is only 10% of the ideal value based on E g As-deposited devices had the lowest value when compared with the other types indicating that annealing slightly improves the junction properties. He-annealed devices had the overall better 31

PAGE 43

performance of 1.1 mA/cm 2 The wavelength dependant SR suggests poor collection especially for deeply absorbed light. -0.0003-0.0002-0.00010.00000.00010.00020.00030.0004-1.00-0.500.000.501.00Voltage [Volts]Current Density [A/cm2] 7-8-2-He 7-8-2-He+O2 7-8-2-H2 asdep -0.003-0.003-0.002-0.002-0.001-0.0010.0000.0010.0010.002-0.50-0.250.000.250.500.751.00Voltage [Volts]Current Density [A/cm2] 7-8-2-He 7-8-2-He+O2 7-8-2-H2 asdep Figure 19. Light and Dark J-V of CZT/ZnSe Devices 0%5%10%15%400450500550600650Wavelength, (nm)Q.E. Type II Type I Type III Type IV Figure 20. Spectral Responses of CZT/ZnSe Devices 32

PAGE 44

5.3 CZT/CdO Cells CZT/CdO cells were categorized in two groups namely type A and B. A: SnO 2 /CdO/CZT structure, B: SnO 2 /CdS/CdO/CZT structure. CZT was deposited at a substrate temperature of 500C for both types. However the majority of type A films did not adhere well to the substrate. This is believed to be due to CdO partially evaporating during this CSS of CZT. Few devices were made from films that adhered to the substrate. The V OC s for Type A were in the range of 100-190 mv. The reason for lower V OC is what is stated above (i.e. Poor adhesion). Evaporation of CdO led to the formation of CdO/SnO 2 junctions which have been observed to have V OC s in the range of 100-250 mv [25]. The light and Dark J-Vs of CZT/CdO devices are shown in figure 21. The J-V curve shows shunting and series resistance effects leading to poor fill factors in the range of 24-26%. The current densities of devices made on CZT/CdO were in the range of 4-6 ma/cm 2 -0.010.000.010.010.020.020.030.030.040.040.05-0.500.000.501.001.502.002.503.00Voltage [Volts]Current Density [A/cm2] -0.020-0.0100.0000.0100.0200.0300.040-2.00-1.000.001.00Voltage [Volts]Current Density [A/cm2] Figure 21. Light and Dark J-Vs of CZT/CdO Devices 33

PAGE 45

Type B devices had good adhesion to the CdO/CdS surface and were found to generate higher current densities (~8 mA/cm 2 ) The V OC s were in the range of 200-280 mv. The FFs were in the range of 24-27%. The limiting performance factors were excessive shunt resistance and high series resistance. The figure below shows the comparison of spectral response for Type A and B devices. 0%10%20%30%40%400500600700800900Wavelength, (nm)Q.E. Type B Type A Figure 22. Spectral Response of CZT/CdO Devices The spectral response behavior shows that type B devices have better collection. The CdS absorption edge is visible for Type B devices while the CdO absorption edge is not clearly present in SR reinforcing the vaporization of CdO during deposition of CZT. The wavelength dependant response suggests that weakly absorbed carriers are not effectively collected. 34

PAGE 46

5.4 CZT/CdS Films 5.4.1 Effect of Deposition Conditions To investigate the effect of ambient composition (targeted CZT composition 0.3
PAGE 47

in the range of 0.15-0.45. The XRD of a CZT film is shown in figure 23. It is apparent from the figure that CZT peaks lay between CdTe and ZnTe peaks. The concentration of zinc was calculated using Vegards law given by the equation 10 given in chapter 3. The lattice constant of CZT was calculated from the XRD measurements using the following equation. d spacing = 222lkha (11) Where d spacing is the interplanar distance and (hkl) are the miller indices. Table 5 displays the composition of one of the CZT film as determined by three methods; XRD, EDS, Transmission. They are found to be in good agreement. Figure 23. XRD of CZT 36

PAGE 48

Table 5. Composition of a CZT Film with E g ~1.6 eV XRD EDS X estimated from Optical Transmission .14 .13 .14 Despite being able to achieve the desired composition x for CZT films on CdS, the films could not be processed into devices as they revealed a significant concentration of pinholes. The SEM image shown in the figure 24(a) shows the as-deposited CZT film on CdS. The grain size was approximately 2 m. The SEM image demonstrates loosely packed CZT grains. The reason for this can be attributed to the fact that the nucleation process is affected by the substrate. In this case it results in low density films. CZT/CdS devices in most cases shorted during contact application. 2 m (a) CZT Film (b) ZnTe/CZT Figure 24. SEM Images of CZT 37

PAGE 49

In order to overcome this difficulty ZnTe was deposited on the CZT film to plug in the pin holes. Zinc Telluride was deposited by CSS in inert ambient (He). Zinc Telluride surface is shown in figure 24 b. It is clear that ZnTe film is denser than CZT and appears to plug in the pinholes in the underlying CZT layer. The ZnTe can also be used as a transparent back contact for the top cell. 5.4.2 SIMS Analysis Figure 25 shows a SIMS depth profile for ZnTe/CZT/CdS structure. The CZT/CdS interface is identified by a sharp increase in the sulphur signal at about 5 microns. At the surface Cd decreases and Zn increases marking the ZnTe film. The cadmium signal shows a gradual decrease near the surface. It is not clear whether this grading is caused by an artifact of the measurement due to the surface roughness. 1E+001E+011E+021E+031E+041E+051E+061E+071E+0802468Depth (microns)Count s 34S64Zn111Cd119Sn125TeCZT-CdS.SWFCZT/CdSNREL 10 Figure 25.SIMS on ZnTe/CZT/CdS Structure 38

PAGE 50

5.5 Characterization of CZT/CdS Solar Cells 5.5.1 Effect of CdS Thickness To find out suitable thickness of the n-type heterojunction partner, CZT/CdS devices were fabricated on approximately 900 800 700 and 600 CdS substrates. It was observed that thicker CdS resulted in a decrease in performance. Figure 26 shows the plot of average V OC vs. CdS thickness for several CZT/CdS/Graphite devices with E g in the range of 1.6-1.75 eV. The fill factors did not exhibit a consistent trend as figure 27 indicates. CZT devices on 60 minute cadmium sulphide on the average exhibited the highest FF. It is difficult to identify the mechanism responsible for the observed device performance. It is possible that CZT behaves in a similar manner as CdTe forming an interdiffusion compound at the junction influencing both junction properties and the nucleation process. 0100200300400500600500550600650700750800850900950CdS thickness()Voc(mv) Figure 26. V OC Vs CdS Thickness 39

PAGE 51

051015202530354045500550600650700750800850900950CdS thickness ()FF (%) Figure 27. FF vs CdS Thickness Spectral responses of the devices are shown in figure 28. From the response it is clear that the 600 CdS devices exhibit higher collection. The bandgap shift in SR is apparent. This could be due to run to run compositional variation or variation in collection efficiency. Absence of CdS absorption edge in the short wavelength region makes it difficult to use this measure as a qualitative indicator of any CdS interdiffusion. 0%5%10%15%20%25%400450500550600650700750800850900Wavelength, (nm)Q.E. 80 70 60 90 Figure 28. Spectral Response of CZT Devices 40

PAGE 52

5.5.2 Effect of Annealing Several heat treatments were tried since the as-deposited devices yielded poor performance initially. Heat treatments similar to CdTe processing (CdCl 2 -based) yielded very poor performance. Following chloride based treatments a series of annealing experiments in He, He + O 2 and H 2 were carried out at 400C. Figure 29 shows the effect of annealing ambient on V OC and FF. Although variations are small, H 2 -annealed devices always resulted in better performance. Nevertheless all performance parameters are significantly lower than targeted values. The spectral response of the three types discussed above is shown in figure 30. It is evident from the figure that H 2 annealed devices exhibit higher collection. The H 2 annealed device shown in figure 31 exhibits a J sc of 6 mA/cm 2 Table 6 lists the values of Voc, Jsc and FF for devices fabricated in various ambients. The critical issue during this work was the difficulty to reproduce device performance. H 2 -annealed devices almost always outperformed as-deposited or devices annealed in He/O 2 ambient. Nevertheless the results exhibited a significant degree of scattering. 67069071033.335.437.4650660670680690700710720AmbienceVoc(mv)3132333435363738FF( % ) FF VocHeHe+O2H2 Figure 29. Effect of Annealing Ambient on V OC and FF 41

PAGE 53

0%5%10%15%20%25%30%35%400425450475500525550575600625650675700725750775800825850Wavelength, (nm)Q.E. H2 He He+O2 Figure. 30 SR of CZT Devices for Various Annealing Ambient Table 6. CZT Devices in Various Annealing Ambient Sample # Ambience Voc (mv) FF (%) Jsc(mA/cm 2 ) He 150 24.6 .87 6-7-1B H 2 250 27.7 6.34 He 690 30.7 .73 He+O 2 550 37.2 .45 6-7-1C H 2 710 37.4 1.2 At this time the most significant limitation appear to be low J sc s as indicated throughout this chapter. An instance of it is shown in Figure 31. It can be suggested that the change in the electronic properties of CZT could have resulted in lower J sc s. As indicated in table 7 the H 2 -annealed devices always had a better performance than as-deposited. 42

PAGE 54

0%5%10%15%20%25%400500600700800900Wavelength, (nm)Q.E. H2 anneal As Dep Figure 31. SR of H 2 -Anneal and As-deposited Devices Table 7. Comparison of As-Deposited & H 2 Annealed Devices Sample # V OC (mv) FF (%) Jsc (mA/cm 2 ) 2-25-1-H 2 anneal 640 40 4.5 2-25-1-as-dep 630 38 1.8 2-25-2-H 2 anneal 590 36.4 3.5 2-25-2as-dep 510 41.7 2.17 2-25-3-H 2 anneal 580 33 1.65 2-25-3-as-dep 590 33.6 1.26 2-25-4-H 2 anneal 579 42.1 2.48 2-25-4as-dep 570 46.1 1.57 5.5.3 Analysis of the Best Device Figure 32 shows the spectral response of the H 2 -annealed quarter that had the best overall performance. It is clear from SR that bandgaps of the devices are low. The 43

PAGE 55

transmission response of the film shows a bandgap approximately 1.6 eV. The loss of zinc could be attributed to the nucleation process. 0%10%20%30%400500600700800900Wavelength, (nm)Q.E. 2-25-1-Q1_1 2-25-1-Q1_2 2-25-1-Q1_3 2-25-1-Q1_4 Figure 32. Spectral Response of the Best Device -0.00004-0.000020.000000.000020.000040.000060.000080.000100.000120.000140.00016-2.00-1.50-1.00-0.500.000.501.00Voltage [Volts]Current Density [A/cm2] 2-25-1_1 2-25-1_2 2-25-1_3 2-25-1_4 Figure 33. Dark J-V of the Best Performance Device 44

PAGE 56

45 The Dark J-V of these devices as shown figure 33 shows shunting. Pinholes on the CZT surface could have acted as source of shunting paths. The fourth quadrant light J-V is shown in figure 34. The devices show both shunting and series resistance effects. The difficulty of establishing a suitable contact with the CZT results in high series resistance resulting in lower fill factors. Table 8 lists the values of all the devices made in this quarter. -0.0075-0.0050-0.00250.00000.00250.0050-2.00-1.50-1.00-0.500.000.501.001.50Voltage (v)Current Density [A/cm2] 2-25-1_1 2-25-1_2 2-25-1_3 2-25-1_4 Figure 34. Light J-V of the Best Device Table 8. Tabulation of Results Serial # Voc (mv) FF (%) Jsc(mA/cm 2 ) 2-25-1-Q1_1 670 32.3 4.3 25-1-Q1_2 650 35.9 3.65 2-25-1-Q1_3 640 39.5 4.5 2-25-1-Q1_4 650 35.8 4.4

PAGE 57

CHAPTER 6 SCAPS SIMULATION The internal structure of polycrystalline heterojunction thin film solar cells is more complex than that of homojunction crystalline cells. The band alignment can have discontinuities and interface states may be present between the layers leading to poor device performance. A numerical simulation program called SCAPS (a solar cell capacitance simulator in one dimension) was developed at the University of Gent, Belgium to realistically simulate various characteristics of a thin film solar cell. SCAPS is a windows application software with codes written in C language. SCAPS can be used to model up to six semiconductor layers. The program calculates the current-voltage (JV), energy band diagram, capacitance voltage (CV), capacitance frequency (CF) and spectral response (SR). The program consists of several panels. The main panel is the action panel where the user sets the problem. In this panel the user can set an operating point namely temperature, voltage, frequency, illumination and an action list to be carried out. Input parameters can be loaded from a file or created before running the simulation. The second panel is the solar definition panel in which various layers of the solar cell are defined. The modelling of polycrystalline semiconductor solar cells is discussed more in detail in the literature [15]. The basic device physics and the effect of material properties involved in CZT solar cells can be better understood using device simulations. The 46

PAGE 58

simulations discussed were done at room temperature under AM1.5 illumination. The values of the input parameters are shown in table 9. Table 9. Input Values for SCAPS Simulation of CZT Parameters Cadmium Sulphide Cadmium Zinc Telluride Thickness 0.1 micron 5 micron Band gap(eV) 2.42 1.7 Electron Affinity(eV) 4.5 4 Electron Mobility(cm 2 /V-S) 50 100 Hole Mobility(cm 2 /V-S) 20 50 Shallow Donor Density(N d ) 1e+16 0 Shallow Acceptor Density(N a ) 0 1e+14 Absorption Constant 1e+05 1e+05 The work function (WF) of the metal is the minimum energy necessary for an electron to escape from an initial energy at the Fermi level into vacuum. For metals, the work function is of the order of (3-5) eV. The work function of gold, graphite and molybdenum were chosen for these simulations. The net acceptor concentration of the absorber (CZT) was varied. Figures 35 & 36 show the plot of V OC and FF for different acceptor concentrations (N a ) & work function. Increasing the acceptor concentration results in lower V OC The higher doping concentration shifts the fermi level closer to valence band, leading to a higher built-in potential increasing the open circuit voltage 47

PAGE 59

(V OC ). Increasing the WF causes the light J-V curve to shift in the third quadrant leading to higher FF as is shown in figure 37. 00.20.40.60.811.21.44.64.855.25.45.6Work Function (eV)Voc (mv) Na=1e+14 Na=1e+15 Na=1e+13 Na=1e+12 Figure 35. Effect of Work Function on V OC Simulated Results 010203040506070804.74.84.955.15.25.35.45.55.6Work Function (eV)FF (%) Na=1e+14 Na=1e+15 Na=1e+13 Na=1e+12 Figure 36. Effect of Work Function on Fill FactorSimulated Results 48

PAGE 60

-30-20-100102030-2-1.5-1-0.500.511.52Voltage(volts)Current density(mA/cm2) WF=5 WF=5.5 WF=4.8 Figure 37. Light I-V Characteristics of Various Work Functions @ Na=10 14 Cm -3 -3-2.5-2-1.5-1-0.500.511.52-10123456Distance(micrometer)Energy Level (eV) Ec5 Ef5 Ev5 Ec5.5 Ef5.5 Ev5.5 Ec4.8 Ef4.8 Ev4.8 Figure 38. Energy Band Diagram for Different Work Functions in Dark @ Zero Bias From figure 38 it is clear that with as WF increases, the conduction band shifts leading to an increase in V OC Even though there was a considerable change in V OC and FF with increase in N a the current density was essentially constant as expected. From 49

PAGE 61

these simulations, the work function of back contact was chosen to be 5 for further studies of device performance since the work function of graphite is 5 eV. 6.1 Effect of Bulk Traps on the Device Performance The performance of the device was studied by introducing bulk defects in the CZT layer. Donor defects were introduced with the acceptor concentration kept constant at 10 14 The capture cross section () used for this simulation were n =10 -15 cm 2 and p =10 -15 cm 2 for n / p =1. The traps were positioned at mid gap (i.e.0.85 eV for CZT). The trap density and the capture cross section ratio were varied to study the device performance of CZT solar cells. It was found that increasing the density of trap (N T ) resulted in poor performance as expected. Figure 39 shows the plot of trap density versus V OC for various n / p The V OC drops with increase in N T and n / p ratio. 00.10.20.30.40.50.60.70.80.91.00E+131.00E+141.00E+15NT(cm-3)Voc(mv) n/p=1 n/p=10 n/p=100 Figure 39. V OC vs Trap Density for Different Capture Cross Section Ratios 50

PAGE 62

-30.00-20.00-10.000.0010.00-2-1.5-1-0.500.511.522.5Voltage (V)Current Density (mA/cm2) NT 13 NT 14 Nt 5e14 acc def Nt1e13 nodef WF=5, Na 1014 Figure 40. Effect of N T on V OC Jsc and FFs Figure 40 shows the J-V characteristics of devices under the influence of defects, no defects. The shape of the curve remains the same for the acceptor defect of N t =10 13 and no defect indicating no change. The kink in the curve for increasing values of trap density suggests the effect of back contact on the device performance reducing the fill factor and current density. 6.2 Effect of Interface States To find out the effect of interface states on the device performance, interface states were introduced. Acceptor defects were introduced at the interface with n =10 -15 cm 2 and p =10 -11 cm 2 The variation of V OC and J are shown in the figures 41 & 42. The position of the trap was varied at the interface to understand the influence of the position of the energy level. The trap density was varied for different concentration to analyze its effect. It can be concluded that with lower energy levels, the V OC s and FFs increase 51

PAGE 63

considerably. This is shown in the figure 43. Even though the changes in Jscs are not visible in the figure, for higher energy levels the performance of the Jscs decreased significantly. The trap densities used for these simulations are unrealistically high and the main reason is the fact that these simulations were carried out to force the J sc s to match the experimental data. 3153203253303353401.00E+141.00E+15Nt(1/cm3)Voc(mv) 200 100 50 Figure 41. V OC Vs Trap Density for Various Acceptor Interface Defects 234567891.00E+141.00E+151.00E+16Nt(1/cm3)J(mA/cm2) 200 100 50 Figure 42. Current Density Vs Trap Density for Various Acceptor Interface Defects 52

PAGE 64

The spectral response simulation results are shown in figure 44 for four different trap densities. The QE is decreased in long wavelength region. The energy band diagrams shows spikes indicating the presence of interface traps suggesting pinning in the fermi level. With the donor defects the performance of the device is not affected significantly [15]. 00.20.40.60.811.21.40.250.30.350.40.450.50.550.6Energy Level(eV)Normalized Valu e Voc FF Jsc Figure 43. Device Performance for Various Energy Levels 01020304050607080300350400450500550600650700750800850900Wavelength(nm)QE(%) 1.00E+15 2.50E+15 5.00E+15 1.00E+16 Figure 44. Simulated Results of QE for Acceptor Interface Defects 53

PAGE 65

Comparing the SR of the N T and N I with N T ranging from 10 13 to 9*10 14 cm -3 and N I ranging from 10 15 to 10 16 cm 2 the spectral response decreases with interface defects. The bulk defects could not be further explored due to convergence problems. Even though lower current densities were achieved by the interface defects, the interface trap densities used were extremely high. This suggests that the defect in CZT could be a combination of bulk and interface defect. 54

PAGE 66

CHAPTER 7 CONCLUSION Cadmium Zinc Telluride was successfully deposited on different wide band gap semiconductors such as CdO, ZnSe and CdS with tin oxide as the transparent conducting oxide. CZT films with E g in the range of (1.6-1.75) eV were characterized using optical transmission, XRD, EDS, and SIMS. The XRD revealed single phase films and spatial compositional uniformity was confirmed with EDS. The CZT/ZnSe solar cells were processed by annealing CZT in various ambients. The annealing in He had better performance with Vocs in the range of 570-600; however FFs in the range of 24-25% with very low current densities. The maximum current density obtained was 1.1 mA/cm 2 CZT/CdO/CdS structure yielded reasonable current densities (7.9 mA/cm 2 ). However these structures only produced V OC s in the range of 200-250 mv. The best device for this structure had; Voc=210 mV, FF=25% and Jsc=7.85 mA/cm 2 Cadmium Sulphide window layer was found to demonstrate a better performance than the other window layers. CZT on CdS resulted in different nucleation leading to pinholes. ZnTe was used as a top contact to plug in the pinholes. The ZnTe/CZT/CdS devices yielded higher performance for CdS thickness of 60 minutes. Devices annealed in H 2 -ambient had; V OC =640 mv, J sc =4.5 mA/cm 2 FF=40%. Modelling of CZT heterojunction devices was initiated in SCAPS to understand the limitations of devices. It 55

PAGE 67

is believed that CZT based solar cells exhibit poor characteristic due to high defect concentrations at the bulk/interface levels. 56

PAGE 68

REFERENCES [1] H.Scheer, A.Ketley, The Solar Economy, Earth Scan Publications, 2002. [2] C.S.Ferekides & D.L Morel, Development Of II-VI Based High Performance High Band Gap Device For Thin Film Tandem Solar Cells, 29 th IEEE Photovoltaic Specialists Conference, pp.724-727, 2002. [3] R.Hill & J.D.Meakin, Current Topics In Photovoltaics, Academic Press, 1985. [4] S.M.Sze, Physics Of Semiconductor Devices, Wiley Eastern Limited, 1991. [5] S.G.Marinskaya, CdTe/ZnSe Junctions, Thesis, University of South Florida, 1997. [6]C.S. Ferekides, Thin Films And Solar Cells Of Cadmium Telluride And Cadmium Zinc Telluride, Dissertation, University of South Florida, 1991. [7] J.K.R.Williams, Renewable Energy Sources For Fuels And Electricity, Island press, 1993. [8] A.R.Aicha, E.Riad, F.D.Barlow, Thin Film Technology Handbook, McGraw Hill, 1998. [9] P.Capper, Properties Of Narrow Gap Cadmium Based Compounds, Inspec publication, 1994. [10] A.Iller, G. Karczewski, G. Kolmhofer, E. Eusakowska, H. Sitter, AES Investigation Of Chemical Treatment Effect On CdTe And CdZnTe Surfaces, Crystal Research Technology, 33, pp.401-409,1998. 58

PAGE 69

[11] A.Rohatgi, R.Sudharsanan, S.A.Ringel, Growth And Process Optimization Of CdTe And CdZnTe Polycrystalline Films Of High Efficiency Solar Cells, Solar Cells, Volume 30, pp.109-122, 1991. [12] C.Amano, H.Sugiura, M.Yamachi, Fabrication And Numerical Analysis Of AlGaAs/GaAs Tandem Solar Cells With Tunnel Intersections, IEEE Transactions on Electron Devices, 36, pp.1026-1035, 1989. [13] Z.Zhao, D.L. Morel, C.S. Ferekides, Electrical And Optical Properties Of Tin-Doped CdO Films Deposited By Atmospheric Metal Organic Chemical Vapor Deposition, Thin Solid Films, 413, pp.203-211, 2002. [14] N.G.Dheere, Melendez, Solution Grown Cds Layers For Polycrystalline Thin Film Solar Cells, 23 rd IEEE Photovoltaic Specialists Conference, pp. 566-571, 1993. [15] M.Burgelman, P.Nollet, S.Degrave, Modelling Polycrystalline Semiconductor Solar Cells, Thin Solid Films, 361, pp. 527-532, 2000. [16]S.Yu, Characterization Of CdTe/Cds Solar Cells, Thesis, University of South Florida, 2002. [17] C.C Shen, P.T.Chang & K.A. Emery, Two Terminal Monolithic InP InGaAsP Tandem Solar Cells With Tunneling Intercell Ohmic Connections, Photovoltaic Specialists Conference, pp. 381-387, 1991. [18] T. Wada, Y. Hashimoto, High Efficiency CIGS Solar Cells With Modified CIGS Surface, Solar Energy Materials and Solar Cells, 67, pp.305-310, 2001. [19] C.S.Ferekides, J.Britt, Y, Ma & L.Kellen, High Efficiency CdTe Solar Cells By Close Spaced Sublimation, 23 rd Photovoltaic Specialists Conference, pp.389-393, 1993. [20] S.Tobenas, E.M.Larramedi, Growth Of Cd (1-X) Zn x Te Epitaxial Layers By Isothermal Closed Space Sublimation, Journal of Crystal Growth, 2002. 58

PAGE 70

[21] K.S.Ramiah, P.D.Pilkington, A.E.Hill, R.D.Tomlinson and A.KBhatnagar, Structural And Optical Properties Of Cds Thin Films Grown By Chemical Bath Technique, Materials Chemistry and Physics, 68, Issue 1-3, pp. (22-30), 2001. [22] T.K. Subramanyama, G. Mohan Rao, S. Uthanna,Process Parameter Dependent Property Studies On CdO Films Prepared By DC Reactive Magnetron Sputtering, Materials Chemistry and Physics, 69, pp.(133-142),2001. [23] T.A.Gessert, P.Sheldon, X.Li. D.Dunlavy, Studies Of Znte Back Contacts To Cds/CdTe Solar Cells, 26 th IEEE Photovoltaic specialists conference, pp. 419-422. 1997 [24] SIMS analysis was done by Dr.Sally Ashers group @ NREL. [25] CZT/SnO 2 junction evaluations will be the subject of a thesis by M.R.Ramalingam which at the time of this defense was in writing. [26] X. Wu, R.G. Dhere, High-Efficiency CTO/ZTO/Cds/CdTe Polycrystalline Thin-Film Solar Cells, NCPV Program Review Meeting, Colorado, NREL Report. no cp-52034063, 2001. 59