USF Libraries
USF Digital Collections

Growth and characterization of ZnO for the front contact of Cu(In,Ga)Se2 solar cells using reactive sputtering techniques

MISSING IMAGE

Material Information

Title:
Growth and characterization of ZnO for the front contact of Cu(In,Ga)Se2 solar cells using reactive sputtering techniques
Physical Description:
Book
Language:
English
Creator:
Bhatt, Rita
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla.
Publication Date:

Subjects

Subjects / Keywords:
ZnO window layer
Semiconducting transparent thin films
Al doped ZnO
Sputtering parameters
Metal target sputtering
Dissertations, Academic -- Electrical Engineering -- Masters -- USF   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Abstract:
ABSTRACT: ZnO window layers for CIGS solar cells are grown with a DC sputtering technique instead of a conventional RF sputtering technique. Transparent window layers and buffer layers are sputtered from the Zn target in the presence of Oxygen. The window layer is doped with Aluminum in order to achieve high electrical conductivity and thermal stability. The effect of different sputtering parameters on the electrical and optical properties of the films is elaborately studied. Sets of annealing experiments are also performed. Combinations of different deposition parameters are examined to design the optimum fabrication conditions. We are able to deposit 85% transparent, Al doped ZnO films having 002-axis orientation and 4e-4 ohm-cm resistivity, which is successfully, used on CIGS solar cells. Resistivity of undoped ZnO buffer layers is varied form 10-2 ohm-cm to unmeasurable by varying the sputtering parameters. The performance of a reactively sputtered window layer and a buffer layer have matched the performance of the RF sputtered ZnO on CIGS solar cells. There has been considerable effort to eliminate Chemical Bath Deposition of the CdS buffer layer from CIS solar cell fabrication. The performance of an undoped DC sputtered ZnO layer is examined on Cd free CIGS solar cells. The ZnO buffer layer is directly sputtered on an underlying CIGS material. The performance of Cd free solar cells is highly susceptible to the presence of Oxygen in the sputtering ambient of the buffer layer deposition 6. As Oxygen is a growth component in reactive sputtering, the growth mechanisms of the DC-sputtered buffer layer are studied to improve the understanding. The performance of all reactively sputtered ZnO devices matched the values reported in the literature and the results for DC sputtered ZnO on Cd-free solar cells were encouraging.
Thesis:
Thesis (M.S.)--University of South Florida, 2007.
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 Rita Bhatt.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 87 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 - 001917605
oclc - 182523257
usfldc doi - E14-SFE0002055
usfldc handle - e14.2055
System ID:
SFS0026373:00001


This item is only available as the following downloads:


Full Text

PAGE 1

Growth and Characterization of ZnO fo r the Front Contact of Cu(In,Ga)Se2 Solar Cells Using Reactive Sputtering Techniques by Rita Bhatt 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: Don L. Morel, Ph.D. Christos Ferekides, Ph.D. Y.L.Chiou, Ph.D. Date of Approval: December 18, 2000 Keywords: ZnO Window Layer, Semico nducting Transparent Thin Films, Al Doped ZnO, Sputtering Parameters, Metal Target Sputtering Copyright 2007, Rita Bhatt

PAGE 2

ACKNOWLEDGMENTS I am deeply indebted to my advisor, Dr Don Morel, for his invaluable guidance and patience throughout this study. I would also like to thank Dr Chris Ferekides for his timely suggestions and Dr He nley for his assistance. I would like to express my speci al thanks to all the fellow researchers for their support and friendliness, which has made working in this lab a memorable experience. I thank Raj, Harish and Zafar for the processing assistance. My special appreciation goes to Sharifa a nd Binita for their support and invaluable friendship. I deeply thank my father for hi s loving guidance and my mother for her support. I am grateful to my brother, Rush i for the constant encouragement. Special thanks go to my loving husband, Hiren, for alwa ys being there for me. And finally, a lots of hugs and kisses go to my sweet daughters, Aashka and Anya, for reviving my spirit.

PAGE 3

i TABLE OF CONTENTS LIST OF TABLES................................................................................................................. ..iii LIST OF FIGURES................................................................................................................ ..v ABSTRACT....................................................................................................................... ......ix CHAPTER 1 INTRODUCTION.............................................................................................1 CHAPTER 2 BA CKGROUND...............................................................................................2 2.1 CuInSe2 Solar Cell........................................................................................................2 2.1.1 Device Structure..................................................................................................2 2.1.2 Operation.............................................................................................................4 2.2 Properties of Semiconducting Transparent Thin Films...............................................7 2.2.1 Semiconductor Properties...................................................................................7 2.2.2 Electrical Properties............................................................................................7 2.2.3 Optical Properties..............................................................................................11 2.2.3.1 Absorption Process in Thin Film Materials............................................12 2.2.3.2 Correlation of Optical and Electrical Properties.....................................15 2.3 Deposition Techniques for Thin Films......................................................................16 2.3.1 Sputtering Deposition.......................................................................................16 2.3.2 Comparison of Gr owth Techniques..................................................................17 CHAPTER 3 MATERIAL PROPERTIES............................................................................21 3.1 Crystal Structure.........................................................................................................2 1 3.2 Optical Pr operties.......................................................................................................22 3.3 Semiconductor Properties..........................................................................................23 3.4 Application of ZnO....................................................................................................24 CHAPTER 4 MATERIAL PROCESSING...........................................................................26 4.1 Device Structure.........................................................................................................26 4.2 Experimental Setup....................................................................................................27 4.3 Experimental Procedure.............................................................................................27 4.4 Processing Parameter for R eactively Sputte red ZnO................................................28 4.4.1 Al Doped ZnO...................................................................................................28 4.4.2 Undoped ZnO....................................................................................................29 4.5 Thin Film Charact erization Methods.........................................................................29

PAGE 4

ii CHAPTER 5 RESULTS AND DISCUSSION.....................................................................32 5.1 Fundamental Mechanism...........................................................................................32 5.1.1 Transport Mechanism.......................................................................................32 5.1.2 Optical Absorp tion Mechanism........................................................................35 5.1.2.1 Effect of Free Carrier Con centration on Optical Absorption.................35 5.1.2.2 Effect of Microstructure on Op tical Absorption of ZnO Thin Film......38 5.2 Effect of Sputtering Parameters on the Growth of a Sputtered ZnO Film...............42 5.2.1 Effect of Spu ttering Voltage.............................................................................44 5.2.2 Effect of Spu ttering Pressure............................................................................44 5.2.3 Effect of O2 Concentration...............................................................................46 5.2.4 Effect of Substr ate Temperature.......................................................................49 5.3 Effect of Sputtering Parameters on the Properties of the Film.................................52 5.3.1 Sputtering Voltage............................................................................................52 5.3.2 Sputtering Pressure...........................................................................................54 5.3.3 Oxygen Concentration......................................................................................55 5.3.4 Effect of Substr ate Temperature.......................................................................61 5.4 Proposed Sputtering Parameters for ZnO..................................................................67 5.5 Comparison between AZO Sputtered from Oxide Target and Metal Target...........70 5.6 Performance of Reactively Sputtered ZnO Window Layers on CIS Solar Cells.....75 5.6.1 CIS/CdS/Zn O Devices......................................................................................75 5.6.2 Direct CIS/ZnO Devices...................................................................................79 CHAPTER 6 C ONCLUSION................................................................................................82 REFERENCES..................................................................................................................... ..84

PAGE 5

iii LIST OF TABLES Table 1 Properties of Zn O films grown using different techniques......................................19 Table 2 Atomic sizes of Zinc a nd Oxygen atoms a nd ions [27]...........................................22 Table 3 Properties of ZnO film [17].......................................................................................24 Table 4 Processing paramete rs for Al doped ZnO.................................................................28 Table 5 Processing parameters for undoped ZnO..................................................................29 Table 6 Electrical properties of undoped ZnO and Al doped ZnO.......................................33 Table 7 Electrical properties of AZO having different microstructures...............................34 Table 8 Electrical properties of undoped ZnO having different microstructures.................35 Table 9 Possible defect leve ls in undoped ZnO films...........................................................41 Table 10 Sputtering yield unde r Argon bombardment [26]..................................................54 Table 11 Comparison of carrier concentr ation and mobility of the undoped ZnO films sputtered at differen t substrate temperatures...............................................64 Table 12 Comparison of carrier concentra tion and mobility of an Al doped ZnO film before and after anne aling in Ar at 200c.....................................................64 Table 13 Electrical properties of an Al doped ZnO film Deposited at 200c.......................64 Table 14 Electrical properties of the AZO film deposited at room temperature and annealed in Oxygen and Ar am bient at 200c for 45 min....................................67 Table 15 Electrical properties of an Al doped ZnO film spu ttered by RF sputtering..........72 Table 16 Electrical properties of Al dope d ZnO film sputtered by DC reactive sputtering...............................................................................................................72 Table 17 (h19-23 H17-08): Diode properties of a CIS solar cell after 200c anneal in Ar and O2 ambient.............................................................................................76

PAGE 6

iv Table 18 (h19-23, h19 r23 ge t h19) Diode properties of th e device fabricated with reactive ZnO and the reference devi ce. Reactive ZnO was sputtered at room temperature and annealed in Ar ambient at 200c for 45 min...................77 Table 19 (7h20-2r,h20-07) Diode properties of a reference device and a device with DC sputtered ZnO. Und oped ZnO was sputtered at room temperature and AZO was sputtered at 100c..........................................................................78 Table 20 (H34 14/H34) Diode properties of CIS/ZnO device s compared with literature results.....................................................................................................80

PAGE 7

v LIST OF FIGURES Figure 1 Structure of CIS solar cell..........................................................................................3 Figure 2 Energy band diagram of an n-on-p heterojuncti on in thermal equilibrium [9]............................................................................................................................ .4 Figure 3 I-V characteris tic for solar cell..................................................................................5 Figure 4 A semiconductor grain boundary in terface in thermodynamic equilibrium [13 fig 3.18]...........................................................................................................9 Figure 5 Variation of h T 1/2 versus 1/T for CVD grown Zn O films [17 fig 3.64].............11 Figure 6 Spectral dependence of a semiconducto r transparent material [19 -fig 4.1]..........12 Figure 7 Absorption processes responsible for absorption of electromagnetic radiation in solids [13 -fonash fig 2.12 ]. a,b-> free carrier absorption. c-> intraband transition. d->band – localize state transition. e->localize statelocalize state transition..........................................................................................12 Figure 8 Plot of band-gap (Eopt) verses N2/3 [24 (fig 4.91)]...................................................15 Figure 9 Optical transmission spectra for undoped, 0.5,1.0 and 2wt% Al doped ZnO films [25 fig 4.96]..................................................................................................15 Figure 10 Schematic drawing showing the gl ow discharge sputtering apparatus of the planar diode ty pe [26 Fi g 19]..........................................................................17 Figure 11 Graphical comparison of electri cal and optical properties of doped and undoped ZnO films as a function of growth process [17 fig 2,89]. Undoped: (1) spray, (2) evaporatio n, (3) sputtering. Doped (4) In-ZnO: spray, (5) In-ZnO: sputtering, (6) Al -ZnO: CVD, (7) Al-ZnO: spray, (8) Al-ZnO: sputtering, (9) Ga-ZnO: sputtering........................................................20 Figure 12 ZnO crystal model [27 fig 1].................................................................................21 Figure 13 Optical properties of Zn O [27 ZnO red pg 24 fig 3]..........................................22 Figure 14 Cross section of CIS solar cell w ith reactively sputtered window layer..............26

PAGE 8

vi Figure 15 Substrate zones w ith respect to target...................................................................30 Figure 16 Variation of free carrier mobility with the carrier concentration.........................32 Figure 17 Absorption coefficient v/ s wavelength for AZO and ZnO...................................36 Figure 18 %Transmission vs. wa velength for AZO and ZnO...............................................37 Figure 19 Calculation of optical band gap for Al doped Zn O (N = 5.4e20 cm-3)................37 Figure 20 Calculation of optical band gap for undoped ZNO (N = 2.2e19 cm-3)................37 Figure 21 Variation of absorption coefficien t as a function of optical wavelength for ZnO films sputtered at differ ent substrate temperatures......................................39 Figure 22 XRD of ZnO deposited at 300c...........................................................................39 Figure 23 XRD of ZnO deposited at 100c...........................................................................39 Figure 24 XRD of ZnO deposited at room temperature........................................................40 Figure 25 Model of Zinc Ox ide [27 pg.58 fig 10].................................................................40 Figure 26 Growth mechanism of reactively sputtered ZnO..................................................42 Figure 27 Variation of growth rate of spu ttered film as a function of sputtering voltage...................................................................................................................44 Figure 28 Variation of growth rate as a function of partial sputtering pressure...................45 Figure 29 Sputtering current as a f unction of sputtering pressure........................................46 Figure 30 Partial pressure vs. run time for 31% and 50% Oxygen concentrated ambient..................................................................................................................47 Figure 31 Sputtering current vs. run time for 31% and 50% Oxygen concentrated ambient..................................................................................................................47 Figure 32 Growth rate as a function of %Oxygen concentration of the films sputtered at 200 c of substrate temperature and at 3 mTorr pressure..................48 .Figure 33 Growth rate as a function of substrate temperature for the films sputtered with sputtering voltage of 400 Vo lt and 480 Volt with 30% Oxygen concentration at 3 mTorr sputtering pressure.......................................................49 Figure 34 Special distributi on of growth rate with different temperatures...........................50

PAGE 9

vii Figure 35 Growth rate as a function of Ox ygen concentration for the films sputtered at 200c and 325c at the sputtering voltage of 480 V and 3 mTorr partial pressure..................................................................................................................51 Figure 36 Resistivity vs. sputtering voltage of an undoped ZnO a nd Al doped ZnO..........52 Figure 37 Carrier concentration and resistiv ity vs. sputtering voltage of the undoped ZnO film................................................................................................................52 Figure 38 Variation of absorption coeffici ent of an undoped Zn O as a function of sputtering voltage at 1350 nm and 550 nm wavelengths.....................................53 Figure 39 Resistivity of AZO as a function of sputte ring pressure.......................................55 Figure 40 Carrier concentration and mobility of undoped ZnO as a function of oxygen content of the sputteri ng ambient at T = 200c.......................................56 Figure 41 Resistivity of th e undoped ZnO as a function of Oxygen content of the ambient..................................................................................................................56 Figure 42 Absorption coefficient of an undoped ZnO films vs wavelength for the films sputtered at 25.93% and 33.33% Oxygen conc entration............................56 Figure 43 Resistivity of undo ped ZnO films vs. Oxygen co ntent of the sputtering ambient, before and af ter the 325c anneal..........................................................57 Figure 44 Carrier concentration of undoped ZnO films vs. Oxygen content of the sputtering ambient, before and after the 325c anneal.........................................58 Figure 45 Mobility of undoped ZnO films vs Oxygen content of the sputtering ambient, before and af ter the 325c anneal..........................................................58 Figure 46 Resistivity of AZO vs. Oxygen content of the sputtering ambient......................59 Figure 47 Carrier concentra tion of AZO vs. Oxygen co ntent of the sputtering ambient..................................................................................................................59 Figure 48 Mobility of AZO vs Oxygen content of the sputtering ambient.........................59 Figure 49 Resistivity vs. substrate te mperature for AZO in zone D.....................................61 Figure 50 Carrier concentration vs. substr ate temperature for AZO in zone D....................61 Figure 51 Mobility vs. su bstrate temperature fo r AZO in zone D........................................62 Figure 52 Absorption coeffici ent of the AZO films sputte red at different substrate temperatures (vis ible range)..................................................................................63

PAGE 10

viii Figure 53 Absorption coeffici ent of the AZO films sputte red at different substrate temperatures (IR range).........................................................................................63 Figure 54 Effects of 200c anneal on th e absorption coefficient of the films deposited at different substrate te mperatures (visible range)...............................65 Figure 55 Effects of 200c anneal on th e absorption coefficient of the films deposited at different substrat e temperatures (IR range).....................................65 Figure 56 Absorption coefficient of the film s sputtered at 100c and 25c (after the 200c anneal) is compared with that of AZO sputtered at 200c substrate temperature (vis ible range)...................................................................................66 Figure 57 Absorption coefficient of the film s sputtered at 100c and 25c (after the 200c anneal) is compared with that of AZO sputtered at 200c substrate temperature (IR range)..........................................................................................66 Figure 58 Effect of oxygen concentrati on and sputtering voltage on the electrical and optical proper ties of ZnO...............................................................................68 Figure 59 Effect of Oxygen concentrati on and the substrate temperature on the electrical and optical properties of ZnO...............................................................68 Figure 60 Resistivity (x10-3 ohm-cm) of AZO on 10 cm x10 cm substrate relative to the location of s puttering s ource...........................................................................70 Figure 61 Optical properties of AZO in different sputtering zones......................................70 Figure 62 Comparison of the optical proper ties of AZO films spu ttered from Zn and ZnO targets............................................................................................................73 Figure 63 Comparison of the spectral res ponses of a cell annealed in Oxygen and Ar ambient.............................................................................................................76 Figure 64 Spectral response of the reference device and a device fabricated with a reactive ZnO window layer. The re active ZnO layer was deposited at room temperature and ann ealed in Ar ambient....................................................77 Figure 65 Spectral response of the reference device and a device fabricated with a reactive ZnO window layer. The undoped ZnO was sputtered at room temperature. The AZO wa s sputtered at 100c....................................................78 Figure 66 Spectral response of a CIS/Cd S/ZnO device and a CIS/CdS device...................80

PAGE 11

ix GROWTH AND CHARACTERIZATI ON OF ZNO FOR THE FRONT CONTACT OF CU(IN, GA)SE2 SOLAR CELLS USING REACTIVE SPUTTERING TECHNIQUES Rita Bhatt ABSTRACT ZnO window layers for CuInxGax-1Se2 solar cells are grown with a DC sputtering technique instead of a conven tional RF sputtering techniqu e. Transparent window layers and buffer layers are sputtered from the Zn target in the presence of Oxygen. The window layer is doped with Aluminum in order to achieve high electr ical conductivity and thermal stability. The effect of different s puttering parameters on the electrical and optical properties of the films is elaboratel y studied. Sets of ann ealing experiments are also performed. Combinations of different de position parameters are examined to design the optimum fabrication conditions. We are ab le to deposit 85% transparent, Al doped ZnO films having 002-axis orientation and 4e-4 ohm-cm resistivity, which is successfully, used on CIGS solar cells. Re sistivity of undoped ZnO buffer layers is varied form 10-2 ohm-cm to unmeasurable by varying the sput tering parameters. The performance of a reactively sputtered window la yer and a buffer layer have matched the performance of the RF sputtered ZnO on CuInxGax-1Se2 solar cells. There has been considerable effort to eliminate Chemical Bath Deposition of the CdS buffer layer from CIS solar cell fabri cation. The performance of an undoped DC

PAGE 12

x sputtered ZnO layer is examined on Cd free CIGS solar cells. The ZnO buffer layer is directly sputtered on an underlying CIGS materi al. The performance of Cd free solar cells is highly susceptible to the presence of Oxygen in the sput tering ambient of the buffer layer deposition [6]. As Oxygen is a growth component in reactive sputtering, the growth mechanisms of the DC-sputtered buffer layer are studied to improve the understanding. The performance of all reactiv ely sputtered ZnO devices matched the values reported in the literature and the results for DC s puttered ZnO on Cd-free solar cells were encouraging.

PAGE 13

1 CHAPTER 1 INTRODUCTION ZnO is a semi-conducting material, which crysta llizes in the hexagonal wurtzite lattice. A large direct band gap (3.2 eV) makes this ma terial transparent in the visible and IR region. Pure Zinc Oxide films are transparent an d usually highly resistive. Such a coating is widely used as a buffer layer on solar ce lls. Non-stoichiometric ZnO films have high conductivity but they are not very stable at high temperatures. ZnO is highly conductive and stable when doped with fluorine, alumin um, gallium or indium. Low resistivity and high transparency in the visible region has ma de this oxide suitable for the front contact of solar cells. Zinc Oxide based coatings have recently received much attention because they have advantages over the more commonly used indi um (ITO) and tin (TO) based oxide films. ZnO is more transparent in the 400800 nm range compared to ITO and TO films. Moreover, Indium oxide and tin oxide are usua lly more expensive than zinc oxide films. Thin ZnO films can be deposited with MOCVD, spray pyrolysis, evaporation and sputtering. However, sputtering is the most commonly used technique for solar cell applications. The RF magnetron sputtering t echnique has been widely used for ZnO deposition on Cu (In,Ga)Se2 solar cells. Atomic beam sputtering and reactive sputtering are also implemented for front contact preparations.

PAGE 14

2 CHAPTER 2 BACKGROUND 2.1 CuInSe2 Solar Cell The CIS solar cell is a heterojunction PN diode. The bandgap of CIS is 1.0 eV and measured value of the abso rption coefficient is 3-6x105 cm-1. The higher absorption coefficient and the low cost processing makes CIS solar cell a leading candidate for thin film solar cells. The highest efficiency th at is reported in any compound semiconductor thin film is 17.1 %, which was achieved with a CIGS solar cell [1]. Moreover the CIGS solar cell is stable as well as less toxic compared to other thin film solar cells, e.g., CdTe [2]. 2.1.1 Device Structure Figure 1 shows a schematic diagram of a CIS solar cell. Soda lime glass has been successfully used as a substrate material for so lar cells. A back contact of this cell is a thin layer of Molybdenum. Low resistivity in the range of 10-5 ohm-cm and the smooth surface of molybdenum result in low series as we ll as low shunt resistance of the device. A p-type CIS absorber layer is grown on th e Molybdenum surface. CIS has a bandgap of 1.0 eV and an absorption coeffi cient in the range of 3-6 x 105 cm-1. Most of the sunlight is absorbed in less than one-micron thickness of this material.

PAGE 15

3 Figure 1 Structure of CIS solar cell An n-type CdS buffer layer is used between the CIS absorber layer and the ZnO window layer. This layer behaves as an absorber layer for the photons having energy higher than the band gap of CdS i.e. 2.4 eV. On the other hand, for low energy photons, CdS behaves as a window layer material. The CdS laye r protects the underlying CIS from oxygen during the window layer deposition [3]. CBD pr ocess of the CdS deposition also modifies the underlying CIS surface [4]. Many groups have reported improvement of the open circuit voltage (Voc) and the fill factor (FF) when a thin CdS buffer layer was introduced [6, 7]. CBD CdS has been us ed for the best Cu(In,Ga)Se2 devices with efficiency above 16% [5]. A resistive layer of ZnO is deposited betw een a CdS buffer layer and the n-type doped ZnO window layer. Typical resistivity of this layer is 20 ohm-cm. Solar cells without any intrinsic ZnO exhibit the st rong decrease in Voc with h eating duration. Either the diffusion of Al species from the conductive ZnO:Al or an enhanced diffusion of atmospheric species (O2 or H2O) in to the CdS/CIS material could explain the Glass Substrate Molybdenum ( Back Contact ) CdS (Buffer Layer) ZnO (Buffer layer) AZO (Front Contact) Cu(In,Ga )Se2 (Absorber Layer)

PAGE 16

4 degradation behavior [8]. A 100–200 nm thick undoped ZnO layer should be deposited to ensure the stability of the device. Surface resistivity of the front contact of a solar cell must be in the range of 4-10 ohmcm2 in order to control the series resistance of a device. A thin layer of Al doped ZnO has been successfully used as a front contact ma terial of the CIS solar cell. The wide band gap (3.2 eV) of ZnO makes it tr ansparent for the IR and visi ble spectrum of the light. 2.1.2 Operation A heterojunction solar cell is a p-n diode formed with two dissimilar semiconductors. Figure 2 [9] shows the band diagram of a typical heterojunction solar cell. Figure 2 Energy band diagram of an n-on-p he terojunction in ther mal equilibrium [9] The semiconductor material with a higher band gap acts as a wi ndow for light in a heterojunction solar cell. This layer is practica lly transparent to the useful light. Sunlight is absorbed in the second layer of a narrow band gap material. This layer is known as an absorber layer. An electron–hole pair is genera ted in the absorption layer as a result of the absorption of photon energy. Electrons generated in the p-side within the diffusion length of the junction are swept by the electric field to the n-side. Similarly light generated holes

PAGE 17

5 are drifted from the n type material to the p type material. Light generated current is collected at the output contacts. I-V characteristic of a solar cell is give by [9] I = Is( eqv/kt –1 ) IL Where Is is diode sa turation current and IL is a current resulted by the excitation of the excess carriers due to the sola r radiation. This equation is plotted in Figure 3. The curve passing through the fourth quadr ant represents negative power In this mode, power is extracted from the solar cell. Figure 3 I-V characteristic for solar cell The performance of a solar cell is interpre ted from the photovoltaic efficiency. The efficiency of a solar cell is a measure of li ght energy successfully c onverted to electrical energy. Rs=0 Rsh = I Pm Im VmVoc Isc FFDARK LIGHT Io

PAGE 18

6 = Pm/Pin = FF Isc*Voc / Pin Where Voc is open circuit voltage and can be given by Voc = A0 (kT/q) ln[(Isc/I0)+ 1] Isc is short circuit current and Io is reverse sa turation current. Ao is a diode quality factor. The value of Ao can be approximated from th e slope of dark I-V data between 0.2V and 0.6V. Io can be calculated from the y intercept of a dark I-V curve. Short circuit current Isc can be obtained from the y intercept of a light I-V curve. Fill factor FF is given by FF = VmIm/VocIsc Where Vm and Im are the maximum current and the voltage corresponding to the maximum power point. A practical solar cell includes the series resi stance from ohmic loss in the front contact and the shunt resistance from the leakage currents. An ideal solar cell has negligible series resistance and infinite shunt resistance. Decrease in shunt resistance in a solar cell is mainly due to the defects like pinholes. IV characteristics of su ch cells can be given by [9]. The low shunt resistance decreases FF and Vo c of the solar cell whereas the high series resistance decreases FF and Isc of the cell. IRs V kT q 1 IsRsh IRs V Is I + I lnL

PAGE 19

7 2.2 Properties of Semiconducti ng Transparent Thin Films 2.2.1 Semiconductor Properties Semiconductor materials, implied by the name itself, have elec trical conductivity intermediate between metals and insulators. Conductivity of an intrinsic semiconductor material is due to the atomic arrangement of the material. Silicon, Germanium, GaAs are the examples of such materials. Conductiv ity of an extrinsic semiconductor material arises due to the presence of impurities, which may be added in the precisely controlled amount. Al doped ZnO is among the ex trinsic semiconductor materials. 2.2.2 Electrical Properties All known semiconductor oxides have n-type conductivity. The following discussion is mainly for the n type semiconductors where electrons are the majo rity carriers. According to the Ohms law when electric field E is applied to a material, current density is given by J = E Where is known as electrical resistivity. Resist ivity of the material is related to the mobility and the carrier concentration according to following relation. Where N is electron density, is mobility of an electron a nd e is the electron charge. The expression of mobility is 1 eN* / m e

PAGE 20

8 Mobility of free carriers depe nds upon the relaxation time ( ), which further depends upon the drift velocity and the mean free path of the charge carriers. These parameters, in turn, depend upon the mechanism by which the carriers are scattered due to the lattice imperfection. A brief account of the vari ous scattering mechanisms involved in semiconductors is given here. a) Lattice Scattering In addition to the various stat ionary imperfections, lattice vi bration also distorts perfect lattice periodicity. Lattice vibrations are categorized in to the acoustical and optical modes. This scattering is a strong function of temperature, material density, crystal structure and lattice imperfections. Lattice scattering is a dominant scattering mechanism in the single crystal undoped ZnO. b) Neutral impurity Scattering The mobility due to neutral impurity scattering is [10] =2 m*e3/20h (Nn)-1 Where Nn is the concentration of the neutral impurities, m* is the effective mass of an electron and h is Planck’s constant. It is clear from above relation that mobility is inversely proportional to the concen tration of neutral impurities. c) Ionized Impurity Scattering Of all the impurities that may be present in the crystal, the greatest effect on the scattering of the carrier is produced by the ionized impurities. This is because the

PAGE 21

9 electrostatic field due to such impurities rema ins effective even at a great distance. The corresponding relation for mobility is [11] =(4e/h)( /3)1/3N-2/3 Where, N is a concentration of the ionized impurities. Domination of the ionized impurity scatteri ng is reported for the Al doped ZnO films with carrier concentration higher than 4x1020 cm-3 [12]. d) Electron-Electron Scattering Electron-electron scattering has little influen ce on mobility because in this process total momentum of the electr on gas is not changed. e) Grain Boundary Scattering Grain boundary scattering is an important m echanism for the polycrystalline ZnO films. In the polycrystalline thin film, conduction mechanism is dominated by the inherent inter crystalline boundaries (g rain boundaries) rather th an the inter-crystalline characteristics. Figure 4 A semiconductor grain b oundary interface in thermod ynamic equilibrium [13 fig 3.18] Grain boundaries generally contain high densit ies of the interface st ates. Free carriers from the bulk of the grain can be trapped here giving rise to the sp ace charge region in

PAGE 22

10 the grain boundary as shown in Figure 4. Band banding occurs due to this space charge region. This band banding introdu ces potential barriers for the charge transports. In such a situation, mobility of the film ( h) is a sum of contributions from bulk ( bulk) as well as grain boundaries ( g). 1/ h=1/ bulk + 1/ g In a lightly doped material or at high temp erature, thermionic emission over the barrier dominates. Thermionic activated mobility in this case will be given by [14, 17] g= 0T-1/2exp(-e b/KT) 0=M/nckT Where b is a height of the potential barrier, nc is a number of crysta llites per unit length along the film and M is a factor that is ba rrier dependent. Mobility defined in above relation can be approximated as a total mob ility of the film as the mobility in the crystallites is much higher th an that of the grain boundary. In the polycrystalline films, when a semiconductor is heavily doped or at low temperature, the dominating current is due to the thermal field emission (tunneling) of the carriers through the barriers. When E00 >> kT thermal field emission comes in to the picture. Where [15, 16] E00=18.5 X 10-12 (N/m* )1/2 For ZnO films m*=0.38 and =8.5. At the room temperature N=6x1018 satisfies E00=KT. This indicates that mobility of the sample w ith carrier concentration higher than this

PAGE 23

11 value will be limited by both thermionic and th ermal field emission at the grain boundary [17]. Figure 5 Variation of h T 1/2 versus 1/T for CVD grown ZnO films [17 fig 3.64] From Figure 5, it is clear that mobility is dominated by tunneling below 100K and conduction is due to both tunneling and ther mionic emission above 100K for the ZnO film with N=2.85x1019 (sample 118)[18]. Conduction is due to the thermionic emission beyond 200K. 2.2.3 Optical Properties Absorption and reflection of the incident electromagnetic radiation are the deciding factors of transmittance of a transparent th in film. Transparent semiconducting materials in general, act as a selective transitive layer. They are transparent in the visible and near

PAGE 24

12 infrared range and reflective to the thermal in frared radiation Figure 6 [19]. At very low wavelengths, absorption due to the fundament al band-gap dominates. High reflection due to free electrons (free electron plasma absorpti on) is observed at very high wavelengths. Figure 6 Spectral dependence of a semiconduc tor transparent material [19 -fig 4.1] 2.2.3.1 Absorption Process in Thin Film Materials The following absorption processes are respons ible for the absorption of electromagnetic radiation in solids. (Figure 7) [13] Figure 7 Absorption processes responsible for absorption of electromagnetic radiation in solids [13 -fonash fig 2.12]. a, b-> free carrier absorption. c> intraband transition. d->band – localize state transition. e->loca lize statelocalize state transition

PAGE 25

13 a) Free Carrier Absorption This absorption arises from the transition of an electron from the lower energy band to the higher energy band after ab sorbing a photon within a con duction band. The same kind of transition within the valance band is also possible for a hole. Such transitions are important in the semiconductor whenever th e density of the carrier in the band is significant. Absorption caused by the free carri er is characterized by following relation [20, 21] = c 2N/ Where c is a constant and is the mobility of a charge carrier. This relation predicts that absorption caused by the free carriers beco mes stronger at longer wavelengths. b) Phonon Absorption This absorption normally occurs in the infrar ed region. Absorbed radiation in this case introduces vibration modes. This kind of pro cess is common in the ionic and covalent materials. c) Electron Inter-Band Transition. Electron absorbs photon energy and moves from a single particle stage in the valance band to a single particle stat e in the conduction band (Figure 7 c). In the direct band gap materials like ZnO, no change in a momentum v ector is required for such a transition and phonons are not involved in th is absorption process.

PAGE 26

14 Minimum photon energy for a band-to-band transition depends upon the band gap of a material. A valance band electron absorbs the incident photon that has energy higher than the band gap (Eg) of the material. In other wo rds, the fundamental ab sorption edge occurs for h =Eg, where is a frequency of an incident op tical radiation. Variation of the absorption ( ) with the photon energy is given by (h )= C( h -Eg)1/2 The absorption phenomenon in a semiconductor with high free electron concentration is explained by Burstein-Moss model [22, 23] In heavily doped semiconductors, where states near the bottom of the conduction band are filled, transition of an electron is possible only form the valance band to th e conduction band stat es lying above the degeneration Fermi level. This shifts an op tical band gap from Eg to the higher energy Egd. This degeneration energy gap is given by Egd =Eg + (h2/8 2 m*) (2 2n)2/3 The fundamental absorption edge in this case shifts towa rd the lower value of the wavelength. In a polycrystalline film, localized states ar e present in the forb idden energy band gaps. These localized states arise from broken ch emical bonds, impurities and other structural defects. This could involve the electron transitions from the localized states and a band (Figure 7 d) and the transitions between the localized bandgap (Figure 7 e). Electrons in the valance band can absorb photon energy less then Eg and transit to the localized state. This may also be true for the transition of an electron from one localized state to the other

PAGE 27

15 localized state. Localized states within the forbidden energy band cause the absorption of a photon with energy less than the band gap energy (Eg) of a material. 2.2.3.2 Correlation of Optical and Electrical Properties a) Effect of Carrier Concentration on the Optical Band Gap: According to the Burstein-Moss band-filling mo del, optical band-gap of a film increases as the carrier concentration increases, as show n in Figure 8 [24]. It is clearly seen from Figure 9 [25] that, as the carrier concentra tion increases, band edge shifts towards the lower wavelengths. Figure 8 Plot of band-gap (Eopt) verses N2/3 [24 (fig 4.91)] Figure 9 Optical transmission sp ectra for undoped, 0.5,1.0 an d 2wt% Al doped ZnO films [25 fig 4.96]

PAGE 28

16 b) Effect of Carrier Concentration on IR Absorption Effect of the free carrier absorption is predom inantly seen near the infrared wavelengths in the semiconductors such as ZnO. This phenomenon puts an upper limit on the carrier concentration of the transparent conductive oxides. 2.3 Deposition Techniques for Thin Films Different techniques can be utilized for the deposition of thin films. Deposition processes like spray pyrolysis, sputtering, vacu um evaporation and MOCVD have been successfully used for thin film oxides. ZnO thin films were deposited with reactive sputtering during this study. The following topics briefly describe the sputtering deposition and compare sputtering proce ss with other deposition techniques. 2.3.1 Sputtering Deposition Sputtering is a process where material is di slodged and ejected from the surface of a solid due to the momentum exchange associated with surface bombardment by the energetic particles. Figure 10 [26] shows the sche matic drawing of a sputtering device. The source of the coating material is calle d the target and it is placed in a vacuum chamber along with the substrate. The bombardme nt species are generally ions of a heavy inert gas. Argon is the most commonly used sputtering gas. Potential is applied between anode (substrate) a nd cathode (target) to ignite the glow discharge. DC voltage is ge nerally applied betw een two electrodes when the sputtering target is a good electrical c onductor. Non-conducting targets ar e not sputtered with direct current methods because of the charge accumu lation on the target surface. This difficulty

PAGE 29

17 can be overcome with the use of radio fre quency (13.56 MHz) sput tering. Deposits of poorly conducting metallic compounds can also be formed by DC sputtering the metallic component while injecting other constituents in the gas phase. This is known as reactive sputtering. In magnetrons sputtering, a magnetic field is used with the cathode surface to form electron traps. This increases the deposition ra te at low substrate he ating. Cylindrical-post magnetrons, planner magnetrons and gun type magnetrons are the most commonly used types in the electronics industries. Figure 10 Schematic drawing showing the glow discharge sputtering apparatus of the planar diode t ype [26 Fig 19] 2.3.2 Comparison of Grow th Techniques Properties of the deposited oxides signifi cantly depend on the growth techniques. Comparison of the characteristics of the ZnO films grown by different deposition techniques is shown in Table 1 [17]. This ta ble enables us to draw following observations related to the sput tering technique.

PAGE 30

18 Disadvantages of Sputtering Technique a) Low deposition rates CVD is recommended for the growth of reproduc ible devices where deposition rate is the primary requirement. b) Substantial heating of the substrate due to its bombardment by secondary electrons and high-energy ions Ion assisted growth techniques are prefe rred over the sputteri ng for deposition on the polystyrene like materials where substrate heating is not allowable. c) More complex and expensive process Spry pyrolysis can be employed for the growth of low cost film where uniformity is not required. Advantages of Sputtering Technique a) Higher Purity b) Better controlled composition c) Provides films with be tter adhesive strength d) Permits better control of the film thickness e) Versatile

PAGE 31

19 Table 1 Properties of ZnO films grown using different techniques Process Substrate Temperature (c) Rate (/min) Resistivity (Ohm-cm) Transmission (%) Ref. Remarks ZnO Spray 400 10-2 70 17 Annealing in N2 Reactive Evaporation 150-200 1.5x 10-3 89 17 Sputtering 200-250 10-2 80 17 O2 = 1-2% Sputtering 125 2 x10-3 90 17 Sputtering in H2 Bias Sputtering Room 7 x 10-3 ~80 17 5%H2/Ar ZnO:Al CVD 367-444 3 x10-4 85 17 Spray 500 2 x10-2 80 17 Spray 300-500 10-3 85 17 Annealing in H2 Sputtering Magnetron <100 3-90 5 x 10 -4 85(Visible+ IR) 47 RF sputtering of ZnO, DC sputtering for Al Sputtering Magnetron 350 200 3-6 x 10-4 85(visible) 53 DC Planner – ZnO+Al2O3 Sintered Target Sputtering Magnetron 350 5 x 19 -4 90(visible+I) 12 DC Reactive Sputtering Zn target Sputtering Magnetron Room 150 2 x 10 -4 80 (visible) 46 RF Magnetron ZnO +Al2O3 Target A graphical comparison of the el ectrical and optical properties of ZnO films as a function of the process is given in Figure 11 [17]. It is quite evident that aluminum doped ZnO, prepared by the sputtering tec hnique has the best properties. Moreover, a vast range of optical and electrical properties is achieved when sputtering tec hnique is employed for the growth of the transparent conducting oxides.

PAGE 32

20 Figure 11 Graphical comparison of electrical and optical prope rties of doped and undoped ZnO films as a functi on of growth process [17 fig 2, 89]. Undoped: (1) spray, (2) evaporation, (3) sputtering. Do ped (4) In-ZnO: spra y, (5) In-ZnO: spu ttering, (6) Al-ZnO: CVD, (7) Al-ZnO: spray, (8) Al-ZnO: sputtering (9) Ga-ZnO: sputtering The sputtering technique, although more comp lex and expensive, is preferred over any other technique when reproducibility, unifo rmity and optimum electrical and optical properties of the ZnO film is a prime concern.

PAGE 33

21 CHAPTER 3 MATERIAL PROPERTIES 3.1 Crystal Structure ZnO crystallizes in the hexagonal wurtzite la ttice. The zinc atom s are nearly in the position of hexagonal close packing. Figure 12[ 27], illustrates the manner in which a zinc crystal model may be constructed. In a Zn O Crystal, every oxygen atom lies within a tetrahedral group of four zinc atoms. Figure 12 ZnO crystal model [27 fig 1] The Zn atom gives up two electrons in its outer shell and become s positively charged Zn++. On the other hand oxygen takes two el ectrons and become negatively charged O--. This favors the ionic binding. The evidence of the presence of some covalent forces has also been reported in ZnO bonding [27 pg. 15]. The distance be tween Zn and Oxygen

PAGE 34

22 atoms along ‘c’ axes of the crys tal lattice is found to be s horter than the value along the other two axes. This departure of complete tetrahedral symmetry is considered to be due to a covalent bond involving a sharing of electrons by zinc and oxygen atoms along the ‘c’ axis [27]. Table 2 Atomic sizes of Zinc and Oxygen atoms and ions [27] Zn 1.33 Zn++ 0.74 O 0.64 O-1.40 As shown in the Table 2, O— is a large ion compare to Zn++. As a result, many small atoms can enter a zinc oxide crys tal at elevated temperatures. 3.2 Optical Properties Figure 13 Optical properties of ZnO [27 ZnO red pg 24 fig 3] The fundamental band gap of ZnO is 3.2-3. 3 eV at room temperature [29]. Optical transmission of ZnO is shown in Figure 13 [27] ZnO completely absorbs ultraviolet rays

PAGE 35

23 of the spectrum. In other words, ZnO is black in the ultraviolet light. Refractive index of ZnO is 2.2. 3.3 Semiconductor Properties The semiconductor property of ZnO is main ly due to the imperfection in the ZnO composition. Like all other semiconductor oxid es, imperfection in the crystal structure liberates extra electrons and makes this oxide an n-type semiconductor. Following are the possible types of imperfections s een in ZnO composition [27, 17]: a) Introduction of interstitial atom such as excess Zn At –220 c, a Zn interstitial atom carries no charge but at room temper ature, most of the Zn atoms liberate one electron and be comes positively charged [30, 27]. b) Oxygen vacancies Oxygen vacancy in the film composition results in the extra electrons liberated from the Zn atom. c) Substitution of Zn or O by the impurity atoms d) Impurity atoms at the interstitial sites. Conductivity of undoped ZnO is mainly due to th e imperfections of type a and b [31, 40]. Resistivity of the reactively sput tered undoped film can vary form 102 to 10-3 ohm-cm. All undoped films sputtered during this st udy showed decrement in the carrier concentration when annealed in the inert gas. The decrease in carrier concentration after the heat treatment [chapter 5.3.3] strongly suggested that presen ce of the in terstitial

PAGE 36

24 oxygen, behaving as an acceptor, was negligible in these films. The carrier concentration of the undoped films sputtered during this wo rk was mainly due to the oxygen vacancy or/and interstitial Zn ions. It has also been report ed that the undoped conducting ZnO films have unstable electrical properties in a l ong-term [32]. The imperfections of type c and d can be achieved by doping a film with the trivalent impurities like Al, Ga and In. Resistivity of such films can be as low as 10-4 ohm-cm. Dopent content of 0.2 – 0.4 at% is typical for such films. Doping of ZnO not only improves conductivity of the film but also its electrical stabil ity. Electrical properties of the Al doped films are stable up to 650K in vacuum and 450K in Oxygen ambient [33,34]. Table 3 shows some of the im portant properties of the ZnO film. Table 3 Properties of ZnO film [17] Lattice Parameter (A) a = 3.24, b = 5.20 Dielectric Constant 7.9 Band Gap (eV) 3.23.3 Single Crystal Mobility (cm2/v-s) 180 Refractive Index 2.2 Electron affinity (eV) 4.2 Melting Point (c) 1975 Density (gm/cm3) 5.6 3.4 Application of ZnO High conductivity, stability and transmission in the visible and IR region make ZnO films suitable for the solar cell applications.

PAGE 37

25 ZnO is extensively used in ceramic industries due to its low coefficient of expansion and high melting point. ZnO has a well-established position in cosmetics and pharmaceutical industries. It provides high covering power, screens out ul traviolet rays, and prolongs effectiveness of perfume [27]. Performance and durability of carbon brushes for a sliding elec trical contact are improved by the use of ZnO (~40%). Similarly silver c ontacts for electrical swit ching also use ZnO. Up to 66% ZnO is used in electrical and heat insulation. Rubber, paint, adhesive, lubricants, plastics are few mo re examples of the diversifyi ng use of ZnO in the industry [27].

PAGE 38

26 CHAPTER 4 MATERIAL PROCESSING 4.1 Device Structure Figure 14 Cross section of CIS solar cell with reactively sputtered window layer Figure 14 shows the structure of CIS solar cell fabricated under the University Of South Florida solar cell project. The Molybdenum back contact is sp uttered on a clean 7059 soda lime glass substrate. The CIGS absorber layer is formed by a two-stage process. In the first stage, the precursor is formed by the sequential deposition of Cu and Ga followed by the co-evaporation of In and Se from the elemental sources. In the second stage, selenization of the precursor is achieved at high temperatures in the presence of Se flux. A thin buffer layer of CdS is deposited by chemical bath deposition process. A thin layer of undoped ZnO is sputtered before the front contact deposition. Al doped ZnO is sputtered to form the wi ndow layer of the cell. Glass ( 2mm) Mo ( 1.0 ) CIS ( 2.0 ) CdS ( 0.02 ) ZnO ( 0.09) AZO ( 0. 3 )

PAGE 39

27 4.2 Experimental Setup ZnO was reactively sputtered using the Veeco 3000 planer magnetron sputtering system. ZnO film was grown on a 4” x 4” soda lime glass substrate. The substrate was mounted on a graphite plate holder and placed 10-15 cm above the target. The substrate was kept slightly off centered in order to avoid the racetrack region of a target. Surface temperature of the substrate was controll ed with a heat lamp. A thermocouple was inserted inside the graphite holder to monitor the substrate temperature. The target material was 3” in diameter and 99.999% pure Zn. Small Aluminum pieces were placed on the target su rface to dope a film. A water-cooling system was used to prevent overheating of the metal target. 4.3 Experimental Procedure Chamber is pumped down to 100 mTorr with a mechanical pump after loading a substrate. The chamber is further evacuated with a diffusion pump. Substrate heating is provided when the pressure drops below 10–6 Torr. When the preferred substrate temperature is reached, ultra high purity Argon is flown in to the chamber. DC power is then turned on and the voltage is adjusted. As soon as the glow discharge initiates, the mass flow controller is set fo r a desired flow of Oxygen a nd Argon. The flow of Oxygen and Argon is adjusted such th at the sputtering pressure is maintained around 3mTorr and the value of sputtering current is maintained within the specified limit. Ratio of the masses of both of the gasses is also kept in the proximity of the selected value.

PAGE 40

28 The shutter is opened when the glow discharge stabilizes. Significant fluctuations of the DC voltage are often noticed during the depos ition process. DC voltage is constantly monitored to ensure the stability of the gl ow discharge. Sputtering voltage is held constant by frequent adjustments of the power source during the film deposition. Occasionally, the flow of the gases has been slightly changed to stabilize the process. Switch off the DC power and heater when desi red thickness of the film is grown. Cut off the gas flow and let the substrate cool down in vacuum. Vent the system when substrate temperature falls below 50c. 4.4 Processing Parameter fo r Reactively Sputtered ZnO 4.4.1 Al Doped ZnO Doping is achieved by placing Al pieces on th e target. Table 4 gives the guidelines for the processing parameters suitable for the deposition of AZO. Table 4 Processing parameters for Al doped ZnO Base Pressure 1-6 x 10-6 Torr Deposition pressure 2.5 – 4.4 mTorr Substrate Temperature 100 – 300 c DC Voltage 480 V DC Current 140 170 mA % Oxygen 33% % Argon 67% Thickness on CIS Solar cell 2500 3000

PAGE 41

29 4.4.2 Undoped ZnO Undoped ZnO is sputtered from the Zn target without the Al pieces on it. When Al doped ZnO is sputtered with Al pieces on the target, some portion of the Al liquefies and penetrates into the Zn metal target. The traces of Al are detected in the sputtered film when the same target is used without Al piec es on it. Separate Zn targets are used to sputter doped and undoped layers in order to avoid such cross contamination. Table 5 is a list of recommended processing parameter for ZnO deposition. Table 5 Processing parameters for undoped ZnO Base Pressure 1 x 10-6 Torr Deposition pressure 2.5 – 4.0 mTorr Substrate Temperature 100 – 300 c DC Voltage 400 440 V DC Current 130 155 mA % Oxygen 38% % Argon 62% Thickness on CIS solar cell 900 4.5 Thin Film Characterization Methods Properties of the film vary significantly according to its position with respect to the sputtered target. As shown in the Figure 15 substrate was divided into five zones according to its position. The region right a bove the target was labeled ‘Zone A’. The zone away from the target was labeled ‘Zone E’. Damages caused by the high-energy particle bombardment are the least in a region away from the target. Due to this reason,

PAGE 42

30 Zone D and E are used to deposit the window layer on the CIS solar cell. Most of the films discussed in following chapters were deposited in Zone D and E. Figure 15 Substrate zones with respect to target Electrical resistivity of th e film was checked by four-point probe method. Following formula was used to calculate sheet re sistance and resistivity of the film. Sheet resistance Rsh = 4.53 (V/I ) ohm Resistivity = Rsh Thickness of the film in cm Varian Cary 17 D Photospectrometer was used to measure optical transmission of the film. Absorption coefficient was calculated fr om the measured value of the transmission with following relation. Absorption coeff. (cm-1) = ln(measured transmission/10,000) / thickness of the film(cm) Hall Effect measurement was used to determine carrier concentration and mobility of the majority carriers. Zn Target A B CD E

PAGE 43

31 Film composition was determined by standa rd energy dispersive x-ray spectroscopy (EDS). X-ray diffraction (XRD) was obtai ned on a Nicolet P3/R3 single crystal diffractometer using Mo radi ation with the wavelength of 0.71. Film thickness was measured on Alpha-step 200 profilometer.

PAGE 44

32 CHAPTER 5 RESULTS AND DISCUSSION 5.1 Fundamental Mechanism 5.1.1 Transport Mechanism The transport mechanism in solids gives rise to the electrical curr ent. The electrical properties of a conductive film are determined by the mechanism by which free electrons transport within the material. Ionized im purity scattering and grain boundary scattering are the most significant scattering mechanisms affecting the mobili ty of free carriers within the sputtered ZnO film. Figure 16 Variation of free carrier mob ility with the carrier concentration 0 20 40 60 80 100 120 140 0.00E+001.00E+202.00E+203.00E+204.00E+205.00E+206.00E+207.00E+208.00E+209.00E+20 Carrier concentrationMobility Measured Calculated Power (Calculated)

PAGE 45

33 Mobility of free electrons is studied to understand the scattering mechanism of the majority carrier within the ZnO film. The m obility affected by ionized impurity scattering is given by the following expression [11]. Free carrier mobility = (4e/h) ( /3)1/3 N -2/3 = 9.80 x 10 14 N –2/3 In Figure 16, the mobility of free carriers is calculated by using the above relation and compared with the measured data. It is cl ear from the plot that, at higher carrier concentrations, measured data follows the simulated relation for ionized impurity scattering. As depicted in Figure 16, the number of free electrons affected the mobility of the carriers when carrier concentration exceeded 5x1020 cm-3. These observations support the reported presumption that the transport mechanism enters into the ionized impurity scattering regime when free carrier concentration exceeds 4 x 1020cm-3 [12]. The transport mechanism is gove rned by grain boundary scattering as well as defect induced scattering at the lower carrier concentrations. Table 6 Electrical pr operties of undoped ZnO and Al doped ZnO ZnO AZO Carrier Concentration (n\ cm2) 8.0 x 10 18 5.10 x 10 20 Mobility 36 18 As shown in Table 6 Al doped ZnO has highe r carrier concentration and lower mobility compared to the values for the undoped Zn O films deposited with similar sputtering

PAGE 46

34 settings. Higher concentration in AZO films is attributed to the contribution from Al+3 ions on substitutional sites of Zn+2 ions and from the interstitial aluminum in the ZnO lattice. Aluminum atoms that produce ioni zed impurity scattering centers may occupy interstitial positions and deform the structur e of the film. Scattering by ionized impurities and defects in the crystal result in the lower mobility seen in doped ZnO films [35, 36, 46]. Table 7 Electrical prop erties of AZO having di fferent microstructures Substrate Temperature Carrier Concentration (cm –1) Mobility (cm2/V-s) 25c 7 x 10 19 3.2 100c 5 x 10 20 11 200c 6 x 10 20 17.8 325c 5 x 10 20 18 Typical carrier concentration of the Al Doped ZnO was 5x 10 20 cm-3. The mobility of these films showed dependence on carrier concen tration as well as microstructure of the film. Table 7 shows the electr ical properties of films spu ttered at different substrate temperatures. The films sputtered at lower s ubstrate temperatures (2 5c – 100c) had poor orientation and smaller grain size. The mobility of these films showed more dependence upon the microstructure of the film. This indicated the prevalence of grain boundary scattering and defect-induced scattering. Films that s puttered at higher substrate temperatures had better crystalline stru cture and showed predominance of ionized impurity scattering.

PAGE 47

35 Typical free carrier concentra tion of an undoped ZnO film was observed in the range of 5 x 1019 cm-3. Full-width at half-maxima (FWHM) of the 002 peaks for the films that sputtered at different temper atures was obtained from XRD measurements and are listed in Table 8. FWHM is a good indication of st ructural integrity and the grain size of the film. It is clear from this data that mobility improved as crystallinity of the film was improved. From this observation, it can be concluded that the mobility of carriers in undoped ZnO films shows sole dependence on grain boundary scattering and defectinduced scattering. As expected, these films did not show a consistent relation between carrier concentration and mobility. Table 8 Electrical proper ties of undoped ZnO having different microstructures Substrate Temperature ( c) Carrier Concentration (cm-3) Mobility (cm2/V-s) FWHM () 25 9.2 x 10 19 6.0 100 5.7 x 1019 30 0.64 200 5.10 x1019 35 0.54 325 8.0 x1018 36 0.40 5.1.2 Optical Absorption Mechanism The optical characteristics of polycrystal line ZnO films depend predominantly upon the free carrier concentration, struct ure, and composition of the film. 5.1.2.1 Effect of Free Carrier Con centration on Optical Absorption The following are the two main effects of hi gher carrier concentration on the optical properties of a film:

PAGE 48

36 a) Increase in optical absorption at longer wave lengths Free carrier absorption is responsible for optic al absorption at the longer wavelengths of the spectrum [37]. Films with higher carrier concentrations s howed significant increase in absorption at wavelengths longer than 1000 nm (Figure 17). Such reduction in transmission puts an upper limit on the carrier concentration of a conductive transparent film. Figure 17 Absorption coefficient v/s wavelength fo r AZO and ZnO b) Decrease in optical absorption at shorter wave lengths Undoped ZnO showed a sharp absorption profile near 350nm, whereas doped ZnO remained transitive at 350nm and became opaque around 325nm (Figure 18). The optical band gaps for both of these films were determ ined by plotting the square of absorption coefficient’s value vs. energy of the incident ra diation (Figure 18, ). These plots illustrate that the optical band gap of the undoped ZnO f ilm shifted from 3.3 eV to 3.52 eV when doped with Al (Figure 19, ). The wi der bandgap seen in the doped film 1 10 100 1000 10000 100000 1000000 02004006008001000120014001600 Wavelength (nm)Abs.Coeff.(cm-1) Undoped ZnO(N=2.5e19) Al Doped ZnO(N=5.2e20)

PAGE 49

37 suggests a Burstein-Moss shift [37]. The opening of the band gap explains the decrease in optical absorption at shorter wavelengths. Figure 18 %Transmission vs. wavelength for AZO and ZnO Figure 19 Calculation of optical band gap for Al doped Zn O (N = 5.4e20 cm-3) Figure 20 Calculation of optical band gap for undoped ZNO (N = 2.2e19 cm-3) 0 20 40 60 80 100 120 02004006008001000120014001600 Wabelength (nm) %Transmission ZnO(N=2.0e19 cm-3) AZO(N=5.2e20 cm-3) 0 1 2 3 4 5 6 7 3.253.303.353.403.453.503.55 evabsorption coeff 2 X 1e10 + reflection losses ZNO : N=2.2N19 cm-3 Linear (ZNO : N=2.2N19 cm-3)

PAGE 50

38 5.1.2.2 Effect of Microstructure on Opti cal Absorption of ZnO Thin Film Stoichiometry of the film had considerable effects on the optical properties of the ZnO thin film. ZnO films with different inte grity were grown by va rying the substrate temperature. Figure 21 compares the optical behavior of undoped ZnO films with various microstructures. XRD analysis of these film s is shown in Figure 22,Figure 23 and Figure 24. We analyzed free carrier concentration and mobility along with the XRD data to see the overall picture of the microstructure of thes e films. Oxygen vacancies and interstitial Zn generate free carriers in undoped ZnO films [31] which, when deposited at lower substrate temperatures, showed a high higher carrier concentration, which suggested the presence of excess Zn in the composition of the film (Table 8). Thus, lower free carrier concentration in undoped ZnO films indicates bette r Stoichiometry of the thin layer. The improvement of half width full maxima of the 002 peak seen in less conductive undoped films affirms the enhanced integrity of the material (Figure 22,Figure 23 and Figure 24). Figure 21 distinctly shows that films with better Stoichiometry were more transparent than the inferior films sputtered at room temperature. Less transmittance of the inferior films can be attributed to the localized states arising from broken chemical bonds, defects, and impur ities [38]. Electrons re siding in localized states absorb the incident photon energy and transit it to the conduction band. Electron transitions between two localized states a nd a localized state a nd the valance band are also feasible [13]. Such transi tions were responsible for more absorption of the incident

PAGE 51

39 light. Furthermore, poor microstructure increas ed scattering of the incident light. Such scattering also contributed to the reduction of transmittance of the film. Figure 21 Variation of absorptio n coefficient as a function of optical wavelength for ZnO films sputtered at different substrate temperatures Figure 22 XRD of ZnO deposited at 300c Figure 23 XRD of ZnO deposited at 100c 0 20000 40000 60000 80000 100000 4006008001000120014001600 25 C 100 C 300 C 0 200 400 600 800 1000 051015202530 2Intensity 0 100 200 300 400 500 051015202530 2Intensity

PAGE 52

40 Figure 24 XRD of ZnO deposite d at room temperature It is interesting to note that the effect of stoichiometry on the optical transmission is prominent in the visible range of the spectru m. Shifting of the abso rption band into the visible region has been reported with the introduction of excess Zi nc into ZnO. This effect has been attributed to the strain caused by excess Zn atoms in the interstitial positions [38]. In order to understand this eff ect, we studied possible donor levels of the Zn ion in ZnO films. A model of non-stoich iometric ZnO, shown in Figure 25 [27], was used to explain the above observations. Figure 25 Model of Zinc Oxide [27 pg.58 fig 10] 0 200 400 600 800 051015202530 2intensity

PAGE 53

41 Table 9 Possible defect le vels in undoped ZnO films Donor/Acceptor Levels Min Energy Required for Electron Transition from VB to Defect Levels Min Energy Required for Electron Transition from Defect Levels to CB eV nm eV nm Zn + 3.26 380 .04 3100 Zn++ 1.1 1127 2.2 563 Table 9 summarizes the possible defect leve ls in ZnO and the fundamental absorption edge associated with these levels. Donor levels produced either because of oxygen vacancies or incorporation of hydrogen, indium lithium, or Zn are reported in the range of 0.02-0.05 eV below the conduction band [17]. In undoped ZnO film, an acceptor level has been reported at 0.80 eV below the conduction band [27]. Wavelengths associated with these energy levels fall in the visibl e region (380nm – 563 nm) and higher IR region (1550-3100 nm) of the spectrum (Table 9). As sh own in Figure 21,the optical behavior of ZnO at frequencies higher than 1500nm was not observed during this study. Influence of microstructure on the transmission of visible li ght can be partly attributed to the presence of defect levels. This hypothesis may prove useful to explain the predominant effect of stoichiometry on optical absorption of the visible radiation. B ecause of the limited nature of the collected data, this presumption requires further research for the justification.

PAGE 54

42 5.2 Effect of Sputtering Parameters on th e Growth of a Sputtered ZnO Film Figure 26 Growth mechanism of reactively sputtered ZnO The growth mechanism of reac tively sputtered ZnO is depict ed in Figure 26. Compounds are synthesized at the target or at the substrate in reactive sputtering. Reaction in the gas phase for the most part can be ruled out, as ions cannot be neutrali zed in the gas phase. The heat liberated in a chemical reaction ca nnot be dissipated in a two-body collision. Simultaneous conversion of energy and moment um requires the reaction to occur at a surfaceeither at the target or the substrate [52,26] At low O2 injection rate, virtually all of the O2 is getter pumped by the condensing Zn coating. Consequently the O2 partial pressure remains relatively low, and the cathode process remains primarily one of the simple Ar sputtering of Zn. Most of the ZnO is synthesized on the substrate. The coatings deposited under these conditions are metallic in nature. In this mode, deposition rate linea rly increases w ith the Oxygen injection rate. As the O2 injection rate approaches that requir ed to produce a stochiometric ZnO, the O2 Ar +x Zn O-x ZnO eZn Target Substrate -Ve ( Sputtering Voltage) A Sputtering Current O ZnO

PAGE 55

43 partial pressure rises because of the reduced getter pumping rate [26]. In this mode, formation of ZnO on the target accelerates. As a consequence, the target cathode develops a surface oxide and deposition rate re duces. It is possible that ZnO decomposes when sputtered. Such poisoned targets can produce Zn atoms, Oxygen atoms, ZnO compounds as well as secondary electrons [26] The same effect can be seen when the sputtering rate of Zn atoms is decreased while the oxygen injec tion rate is kept constant. As shown in Figure 27, the deposition rate of sputtered films is governed by the following factors: a) Sputtering rate of metallic target Affecting Parameters: Sputtering voltage, gas pressure and O2 concentration b) Fraction of the sputtered atom (Z n) that can reach the substrate Affecting Parameters: Gas pressure, O2 concentration c) Fraction of the sputtered at oms that are oxidized (Zn1+xO) before they reach the substrate. Affecting Parameters: Gas pressure, O2 concentration, sputtering voltage d) Oxidation rate of sputtere d atom on the substrate Affecting Parameters: O2 concentration, substrate temperature

PAGE 56

44 5.2.1 Effect of Sputtering Voltage The following results were taken at differe nt sputtering voltages while keeping other parameters constant. Sputteri ng current was restrained in the range of 150mA to 200mA. In this case, sputtering power can be consider ed directly proportional to the sputtering voltage. Figure 27 Variation of growth rate of sputte red film as a function of sputtering voltage As shown in Figure 27, sputtering yield of Zn was achieved between 360 Volts and 380 Volts at 3 mTorr of ambient pressure. The sputtering rate of Zn increased with the sputtering voltage. This resulted in increased deposition rate. Change in the slope of the linear plot at 400 Volts indicates slower growth at voltages higher than 400 Volts. Limited supply of O2 was responsible for the slower growth at higher voltages. 5.2.2 Effect of Sputtering Pressure The effect of sputtering pressure was studi ed by increasing the flow rates of both, O2 and Ar, without altering their relati ve concentrations. Data show n in Figure 28 has been taken at constant O2 to Ar ratio (0.85). 0 0.2 0.4 0.6 0.8 1 300350400450500550 Sputtering Voltage (V)Grwoth Rate A/s

PAGE 57

45 Figure 28 Variation of growth rate as a function of partial sputtering pressure As shown in Figure 28 a monotonic increase in deposition rate was observed when pressure was raised from 1 mTorr to 4 mTo rr. The faster growth of a film can be attributed to the enhancement of O2 and Ar fluxes. The higher concentration of Ar flux increased the sputtering rate of Zn atoms from the target. The rise in sputtering current confirmed more sputtering activities at higher ambient pressures (Figure 29). The oxidation ra te of sputtered Zn atoms/Ions was also increased at higher sputtering pressu res due to the enhancement of O2 flux. As a result, the deposition rate increased by a factor of two when pressure was increased by 3 mTorr (Figure 28). The mean free path of a sputtered atom (Zn) de creases at higher pressures. This results in back scattering of the sputtered atom [ 40]. Enhanced sputtering current could not contribute to the faster growth of the film due to the back scattering. As a result, deposition slowed down above 4.0 mTorr (Fig ure 29). At higher pressures, unoxidized Zn atoms were scattered all over the sputte ring chamber in the form of fine powder. O2/Ar Ratio=0.85 temp=200C volt = 480 v 0.00 0.20 0.40 0.60 0.80 1.00 0.001.002.003.004.005.006.007.00 Partial Sputtering pressure mTorrGrowth Rate A/s

PAGE 58

46 Figure 29 Sputtering current as a function of sputtering pressure The minimum sputtering pressure, which coul d sustain the glow di scharge was found in the range of 1 mTorr. Belo w this pressure, secondary electrons could not undergo a sufficient number of ionizi ng collisions before they struck the anode [26]. The fluctuation of sputtering pressure increas ed at higher pressures due to the enhanced O2 flux. The glow discharge became unstable fo r partial pressure more than 10mTorr. Instability of the discharge in the oxygen rich ambient was also observed when oxygen concentration was increased beyond a certain limit (50%), even at lower sputtering pressures (4 mTorr). Other groups have also reported an influence of oxygen concentration on the stability of the glow discharge. [41,40]. 5.2.3 Effect of O2 Concentration Oxygen flow rate had been the most impor tant sputtering parame ter due to reactive nature of the sputtering. Experiments were carried out with constant sputtering voltage and pressure to study the effect of O2 concentration on the growth of ZnO Films. Various concentrations of O2 were achieved by altering the O2 to Ar ratio at constant pressure. In order to maintain the pressure, the flow of Ar was decreased when the flow of oxygen was increased. 0 100 200 300 0.001.002.003.004.005.006.007.00 Partial pressure (mTorr)Sputtering Current (mAmp)

PAGE 59

47 Figure 30 Partial pressure vs. run time fo r 31% and 50% Oxygen concentrated ambient Figure 31 Sputtering current vs. run time fo r 31% and 50% Oxygen concentrated ambient When ZnO is reactively sputtered from a Zn target in the presence of Oxygen, there are two competing oxidation proce sses occurring during the depo sition process [41]. One is the oxidation of a deposited film on the substrat e, while the other is the oxidation of the Zn metal cathode. Higher Oxygen flow causes ra pid oxidation of the Zn target, forming a thin oxide layer on the target. A fraction of the DC applied vo ltage drops across this layer resulting in lower cathode voltage available to sustain the discharg e [26]. Moreover, the sputtering yield of ZnO is lower than that of Zn because the oxide has a higher binding energy and mass than the metal atom [26]. As a result, the increasing depth of the oxidized layer slows down sputtering activit ies. The deposition rate decreases. This effect is clearly seen in Figure 30 and Fi gure 31. Sputtering curre nt gradually dropped 2 2.5 3 3.5 4 01020304050 Run Time (min)Partial Pressure (mTorr) 31 % O2 50% O2 135 155 175 01020304050 Run Time (min)Sputtering Current (m Amps) 31 % O2 50% O2

PAGE 60

48 from 180mA to 140mA in 45 minutes when the process was stared with 50% Oxygen concentration. Partial pressure increased by 0.69 mTorr during this period. On the other hand, the process carried out with 30% O xygen concentration main tained practically constant current with 0.47 mTorr increment in the partial pressure. In order to avoid excessive oxidation of the Zn target, an uppe r limit of the Oxygen concentration was set to 40% during this work. Figure 32 Growth rate as a f unction of %Oxygen concentrati on of the films sputtered at 200 c of substrate temperature and at 3 mTorr pressure Figure 32 shows the dependence of growth ra te on oxygen concentration for the films sputtered at 200 c. Acceleration in th e deposition rate was observed when oxygen concentration was increased from 22% to 29% The sputtering rate of Zn could be considered constant for this range of oxygen concentrations, as th e sputtering power and pressure were unaltered during this set of experiments. Thus, the increase in deposition rate was the consequences of faster oxida tion of the sputtered Zn atoms [42,43,44]. O2 concentration higher than 29% slowed downs the deposition rate. This is attributed to oxidation of the target itself. Since the binding energy of Zn O is greater than Zn, the reduction in sputtering rate was observed at higher gas phase concentrations. This is 0.65 0.70 0.75 0.80 0.85 0.90 23.0825.6528.5729.8231.0333.3337.50 % oxygenGrowth Rate ( A/sec)

PAGE 61

49 further confirmed by the reported fact that the gr owth rates obtained using a Zinc target at higher O2 concentrations were more comparable to the values for a ZnO target [45]. Initial acceleration in growth rate with increasing oxygen concentration and drastic decrement of deposition rate in highly oxygen c oncentrated ambient is the characteristic of reactive sputtering. Several groups have also repor ted such behavior. [42,43,44]. Films become oxygen deficient when Oxygen flow is limited. This in troduces the lower limit of O2 concentrations, which was 28.5% (at 3m Torr pressure) with our experimental set up. 5.2.4 Effect of Substr ate Temperature Deposition governed by reaction on the substrat e surface depends upon the arrival rate of the sputtered Zn and oxidation rate of this atom ON the substrate surface. .Figure 33 Growth rate as a function of substrate temperatur e for the films sputtered with sputtering voltage of 400 Volt and 480 Volt with 30% Oxygen concen tration at 3 mTorr sputtering pressure The relation between substrate te mperature activated growth a nd arrival rate of sputtered Zn was studied with the different combinati ons of substrate temperature and sputtering voltage. As shown in Figure 33, film grow th was more temperature dependent when 0 0.2 0.4 0.6 0.8 1 050100150200250300350 Substrate Temperature cGrowth Rate A/sec 400 v 480 v

PAGE 62

50 films were sputtered at higher voltages. Whereas, when the arrival rate of Zn to the substrate was slower at lower voltages, de position was less temperature dependent. This suggested that oxidation rate of arriving Zn atoms was suffi cient enough to oxidize all the approaching Zn atoms on the substrate surface when the sputtering rate of Zn was slow. Moreover, the rate of ZnO formation ON the target increased when the sputtering rate was decreased [52, pg.42]. Grow th of the film was more te mperature dependent when the flux of sputtered Zn approached the substrat e surface rapidly. At high target sputtering rates, it is well established that virtually all of the compound synt hesis occurs at the substrate [52]. At low substr ate temperatures and high spu ttering voltages, the reaction rate was not fast enough to oxidize all the ar riving metal atoms. Accelerated oxidation rate of metallic Zn at higher substrate temp eratures resulted in a faster film growth. Figure 34 Special distributi on of growth rate with different temperatures This kind of effect was also observed for different substrate positions (Figure 34). The arrival rate of Zn was slow at the region away from the target (region E) whereas Zn atoms rapidly approached the re gion right above the racetrack (region A). It is clear from Figure 34 that growth is more temperature dependent in region A compared to that of 0 0.5 1 1.5 2 2.5 0100200325 subsgtrate Temperature cgrowth rate A/s A B C D E

PAGE 63

51 region E. The effect of high-energy particle bombardment might also have contributed to this result. Figure 35 Growth rate as a f unction of Oxygen concentration for the films sputtered at 200c and 325c at the sputtering voltage of 480 V and 3 mTorr partial pressure Oxygen flux approaching the grow ing film is another significant factor, which plays an important role in determining the temperat ure activated reaction regime. As shown in Figure 35, the substrate temperature did not affect deposition rate when the film was sputtered in oxygen concentra tions lower than 30%. This was the ‘Oxygen starved’ regime and oxygen flux was not sufficient e nough to oxidize all the arriving metal atoms [28]. In the Oxygen starved regime, a higher rate of reaction did not re sult in faster film growth. The drop in the gr owth rate at higher O2 concentrations can be attributed to the oxidation of the target. From the above observations, it is safe to c onclude that growth was more temperature dependent when sufficient flux of Zn and Oxygen was available near the substrate. 0.6 0.7 0.8 0.9 1 2025303540 % O2 concentrationGrowth rate A/s 200 c 325 c

PAGE 64

52 5.3 Effect of Sputtering Parameters on the Properties of the Film The primary purpose of this work was to de velop an optimum spu ttering technique for the window layer ZnO. In order to get the be st combinations of electrical and optical properties of the film, the effect of differ ent sputtering parameters on deposited films was elaborately studied. 5.3.1 Sputtering Voltage The DC voltage applied between a Zn targ et and a substrate is known as sputtering voltage. Figure 36 Resistivity vs. sputtering volta ge of an undoped Zn O and Al doped ZnO Figure 37 Carrier concentration and resistivity vs. sputtering voltage of the undoped ZnO film 0.0001 0.001 0.01 0.1 1 10 100 380390400410420430440450460470480490500 Sputtering Voltage (volt)Resistivity (ohm-cm) Zno AZO 0 2 4 6 8 10 460480500Sputtering Voltage (Volt)Carrier concentrationxE19 Resistivity x E-3 carrier concentration xE 19 Resistivit X E -3

PAGE 65

53 Figure 38 Variation of absorptio n coefficient of an undoped Zn O as a function of sputtering voltage at 1350 nm and 550 nm wavelengths As shown in Figure 37, carrier concentr ation of the undoped film increased with sputtering voltage, which resulted in less resistive films. Transmission of the ZnO also decreased when sputtered with higher voltage s (Figure 38). This suggested that film sputtered at higher sputtering voltages were Zn rich, i.e. oxygen deficient. This observation confirmed that conductivity of the undoped films is due to the nonstochiometry in the form of interstitial Zn atom and/or deficient oxygen sites. It is interesting to note th at Al doped ZnO showed thresh old effects similar to the undoped ZnO films in above plot. The tran sition from high resistivity to high conductivity was achieved around 410 Volts for AZO (Figure 36). Undoped films showed such transitions at 425 Volt. This sugge sted the possibility of oxidation of Al at lower voltages. At lower voltages, the spu ttering rate of Al was slow. The relatively higher concentration of Oxygen could be suffi cient to partially oxidize Al atoms/ ions. Oxides of Aluminum do not behave as a dopent and do not contribute much to the conductivity of the film (Figur e 36). Similar effect was s een when AZO was sputtered with higher oxygen concentrations with 480 sputtering voltage. 0 2000 4000 6000 8000 440460480500 Sputtering Voltage (Volt)Absorption coefficient (cm-1) + reflection losses 1350nm 550nm

PAGE 66

54 As shown in Table 10, Al is a low yield mate rial compared to Zn. The presence of Al in the sputtered film was noticed around 60W with our experi mental set up. Deposition of ZnO was initiated approximately at 54W sputtering power. Due to the significant difference between sputtering yi elds of these two materials, films sputtered at the low voltages had negligible Al incorporation. ED S analysis had indicated a small amount of Al (0 % 0.3% wt) in films sputtered at lowe r voltages. At higher voltages, the sputtering rate of Al increased along w ith Zinc. This resulted in more dopant incorporation inside the growing film. Higher conductivit y of the film, seen at higher sputtering voltages, can be attributed to the increased dopant concentration [47 fig 4]. Table 10 Sputtering yield under Argon bombardment [26] Target Ion Energy (eV) Yield (Atom/Ion) Al 600 1.5 Zn 600 5.07 5.3.2 Sputtering Pressure The effect of sputtering pressure on the resist ivity of the film was not significant (Figure 39) .The increase in resistivity of Al doped Zn O at high pressures may be attributed to the increased oxidization rate of sputtered Al atoms. Moreover, at high pressures the mobility of the free carriers might have degraded due to the higher Al concentration within the film.

PAGE 67

55 Figure 39 Resistivity of AZO as a function of sputtering pressure High electrical resistance of th e film, sputtered below 2 mTorr, could be attributed to the enhancement of the bombardment by high-energy oxygen, resulting from the increase in the mean free path of these species [48]. 5.3.3 Oxygen Concentration Figure 40 and Figure 41 shows th e electrical properties of the undoped ZnO films as a function of oxygen content of the sputtering ambient. When Oxygen concentration was increased from 22% to 33%, free carrier concentration declined from 6 x 10 19 cm-3 to 2.5 x 10 19 cm-3. Mobility of the free carri ers appeared to be unaffect ed in this range of the concentration. The reduction in free carrier concentration s uggested fewer Oxygen deficient sites within the undoped film. Improved optical transmissi on, both in IR and visible region of the spectrum, also implied less metal rich stochi ometry of the film (Figure 42). In other words, films sputtered within Oxygen rich ambient had fewer neutral and ionized Zn atoms. The lower carrier concentration of these films could not produce drastic improvement in free carrier mobility becau se ionized impurity scattering is not a dominant transport mechanism in the lightly doped polycrystalline films [chapter 5.1.1]. The damaging effects of Oxygen bombardment may also be a factor preventing the 0.00E+00 2.00E-04 4.00E-04 6.00E-04 8.00E-04 1.00E-03 0.001.002.003.004.005.006.007.00 Sputtering Pressure (mTorr)Resistivity ( ohm-cm)

PAGE 68

56 enhancement in mobility, when sputtered with in the higher Oxygen concentration [48]. In order to improve the electrical properties of these films, we annealed them for 45 minutes in Ar at 325c. Figure 40 Carrier concentration and mobility of undoped ZnO as a function of oxygen content of the sputtering ambient at T = 200c Figure 41 Resistivity of the un doped ZnO as a function of Oxyg en content of the ambient Figure 42 Absorption coefficien t of an undoped ZnO films vs wavelength for the films sputtered at 25.93% and 33.33% Oxygen concentration 0.00E+00 2.00E-03 4.00E-03 6.00E-03 8.00E-03 1.00E-02 20.0022.0024.0026.0028.0030.0032.0034.00 % OxygenResistivity (Ohm cm) % Oxygen-4,000 1,000 6,000 11,000 16,000 3505507509501150135015501750 Wavelength (nm)Absorption Coeff. (cm-) + reflection losses 25.93 33.33 0 5 10 15 20 25 30 20.0022.0024.0026.0028.0030.0032.0034.00 % OxygenCarrier Concentration(cm-3 x 10 19) Mobility (cm^2/vs) Nxe19 mobility Linear (Nxe19) Linear (mobility)

PAGE 69

57 As shown in Figure 44 and Figure 45, all an nealed films showed decrease in carrier concentration and increase in the mobility. The resistance of all ZnO films was increased after the heat treatment (Figure 43). The improvement of electrical mobility was prominent for films sputtered in an Oxygen rich ambient. The mobility increased from 28 cm2 v-s to 30 cm2 v-s and carrier concentration decrea sed by a factor of two for the films sputtered with lower oxygen concentration (b elow 26%). Whereas, films sputtered with higher Oxygen concentration (33%) showed an increment of 10 cm2 v-s in the mobility. Carrier concentration decreased by an order of a magnitude for th ese films. In other words, films sputtered with higher Oxygen concentration showed a significant decrement in carrier concentration and im provement in the electrical mobility when annealed. This can be attributed to the tr ansfusion of Oxygen into the Oxygen deficient sites of the microstructure. As these films were a nnealed in Oxygen free ambient, such chemisorption of oxygen was negligible from the surface of the film. The decrease in carrier concentration can be due to the chem isorption of Oxygen, which was trapped at defects such as grain boundaries [49, 50,46]. Films sputtered at higher Oxygen concentration had more Oxyge n trapped at defect sites, which resulted in better stoichiometric films when annealed [28]. Figure 43 Resistivity of undoped ZnO films vs. Oxygen content of the sputtering ambient, before and after the 325c anneal 0.00E+00 2.00E-02 4.00E-02 6.00E-02 8.00E-02 2022242628303234 % OxygenResistivity (Ohm cm) Before Annealing After Annealing

PAGE 70

58 Figure 44 Carrier concentration of undoped ZnO films vs. Oxyge n content of the sputtering ambient, before and after the 325c anneal Figure 45 Mobility of undoped ZnO films vs. O xygen content of the sputtering ambient, before and after the 325c anneal The decrease in carrier concen tration after the heat treatment strongly suggested that the presence of intersti tial/ trapped oxygen behaving as an acceptor was negligible in all films and that the carrier concentration of the undoped films sputte red during this work was mainly due to the oxygen vacan cies or/and inters titial Zn ions. The above observations imply that the hi gher Oxygen concentrati on in the sputtering ambient minimizes the Oxygen deficient sites in the deposited films. Th is results in lower free carrier concentrations. The improvement in mobility, after the heat treatment, can be attributed to less grain boundary scattering of the free carriers. Most likely, less scattering due to defects associated with trapped oxygen complexes. 0 10 20 30 40 2022242628303234 % OxygenMobility ( cm^2/vs ) Before Annealing After Annealing Poly. (After Annealing ) Poly. (Before Annealing) 0.00E+00 1.00E+19 2.00E+19 3.00E+19 4.00E+19 5.00E+19 2022242628303234 % OxygenCarrier Concentration (cm-3) Before Annealing After Annealing

PAGE 71

59 Figure 46 Resistivity of AZO vs. Oxyge n content of the sputtering ambient Figure 47 Carrier concentration of AZO vs. O xygen content of the sputtering ambient Figure 48 Mobility of AZO vs. Oxygen content of the sputtering ambient Figure 46, Figure 47 and Figure 48 shows the el ectrical properties of the Al doped ZnO films. Typical carrier con centration of the doped film s was in the range of 10 20 cm-3, which was 10 times higher than the value for undoped films. In such a range, it is 0.00E+00 2.00E-04 4.00E-04 6.00E-04 8.00E-04 1.00E-03 20.0022.0024.0026.0028 .0030.0032.0034.00 %OxygenResistivity (ohm-cm) 0.00E+00 2.00E+20 4.00E+20 6.00E+20 8.00E+20 1.00E+21 15.0020.0025.0030.0035.00 % OxygenCarrier Concentration (/ cm^3) 0 5 10 15 20 25 30 15.0020.0025.0030.0035.00 % OxygenMobility (cm^2/vs)

PAGE 72

60 unlikely that the significant influence of intr insic carrier concentration on the total free carrier concentration can be obs erved. Thus the decrease in carrier concentration from 8.5 X 10 20 to 5.10 X 10 20, when sputtered within the Oxygen rich ambien t, can be attributed to the reduced dopant efficien cy. The resistively of Al in creases significantly due to oxidation even when it is evaporated in oxyge n [43 14-347SP]. It is reasonable to accept that sputtered Al goes into a film as a part ially or fully passivated oxide when deposited in an Oxygen rich ambient. Doped films spu ttered in 42% Oxygen s howed resistivity in the range of 10 2 ohm-cm, which was comparable with the resistivity of undoped ZnO films. This observation indicates the presence of Al as an oxide within the film when sputtered in an oxygen rich ambient. The mobility of the free carriers increased by the factor of two when sputtered in the Oxygen rich ambient. Less scattering of fr ee electrons by ionized impurities can be a major factor responsible for such improvement. The peak for Al and its oxides had not b een detected by XRD m easurement due to the very small amount of the dopant concentration (~ 0.60%). The degradation of dopant efficiency preven ted us from using Oxygen concentrations more than 35% within our experimental setup.

PAGE 73

61 5.3.4 Effect of Substr ate Temperature Microstructure of reactively sputtered film is highly su sceptible to the substrate temperature [12]. Consequently, Optical and electrical properties of the sputtered films showed strong dependence on the substrate temp erature. ZnO deposited during this work was intended to be used on the substrate-confi gured solar cell. It was necessary to deposit the window layer at the substrat e temperature that was safe enough to retain properties of the underlying material. The effect of substr ate temperature was st udied in order to establish the lowest possible temperature at which, reasonable optical and electrical properties can be achieved. The range of te mperature was confined to be from room temperature to 350c. Figure 49 Resistivity vs. substrate temperature for AZO in zone D Figure 50 Carrier concentration vs. subs trate temperature fo r AZO in zone D 0.00E+00 1.00E-02 2.00E-02 3.00E-02 4.00E-02 5.00E-02 050100150200250300350 Substrate Temperature cResistivity (ohm-cm) 0.00E+00 2.00E+20 4.00E+20 6.00E+20 8.00E+20 050100150200250300350 Substrate Temperature cCarrier Concentration ( /cm^3)

PAGE 74

62 Figure 51 Mobility vs. su bstrate temperature for AZO in zone D As shown in Figure 49, resistivity of the Al doped film in region D decreased when deposition temperature was increased from room temperature to 100c. Low resistivity in the range of 5x10-4 was independent of substrate te mperature between 200c and 325c. Such behavior has also been noted by ot her groups [51]. Figure 50 and Figure 51 shows the electrical properties of the film deposited in zone D. Th is zone was practically free from the bombardment damages at selected sputtering parameters. Carrier concentration and mobility significantly increased when subs trate heating was provided (Figure 50 and Figure 51). Consequently, the resistivity of the film dropped from 4.3 x 10-2 ohm-cm to 1.09 x 10-3 ohm-cm. The increase in mobility with the increase in temperature (up to 200c) was due to improved microstructure and crystallinity of the film [51]. 0 5 10 15 20 050100150200250300350 Substrate Temperature cmobility (cm^2/v-s

PAGE 75

63 Figure 52 Absorption coefficien t of the AZO films sputtere d at different substrate temperatures (visible range) Figure 53 Absorption coefficien t of the AZO films sputtere d at different substrate temperatures (IR range) Table 11 shows the electrical properties of undo ped ZnO films. Carrier concentration of these films decreased when the substrate temperature was elevated. The decrease in carrier concentration indicated better stochiometry of the film. Al doped as well as undoped films became more transparent, especia lly in the visible sp ectrum of light, when substrate temperature was elevat ed ( Figure 52 and Figure 53). 0 20000 40000 60000 80000 350450550650 Wavelength(nm)Absorption Coeff. ( /cm) + Reflection losses 25 c 100 c 200 c 325 c 0 5000 10000 15000 700800900100011001200130014001500 Wavelength(nm)Absorption Coeff.( /cm) + Reflection losses 25 c 100 c 200 c 325 c

PAGE 76

64 Analysis of the electrical a nd optical properties of these f ilms suggested that substrate temperature improved stochiometry and crysta llinity of the film [53]. The observation that films prepared at higher substrate temp erature were less Oxyge n deficient confirmed that substrate temperature stimulates oxidati on of the film on the substrate [52 pg.59]. Table 11 Comparison of carrier c oncentration and mobility of the undoped ZnO films sputtered at different substrate temperatures Substrate Temperature (c) Carrier Concentration (cm-3) Mobility (cm2/vs) 25 9.2 x 1019 6 100 5.7 x 1019 30 200 5.10 x 1019 35 325 8 x 1018 36 Table 12 Comparison of carrier c oncentration and mobility of an Al doped ZnO film before and after annealing in Ar at 200c Deposition Temperature Carrier Concentration (cm–3) Mobility (cm2/Vsec) Before Annealing After Annealing Before Annealing After Annealing 25c 7.0 E 19 4.40 E 20 1.8 6.20 100c 5.0 E 20 5.5 E 20 11.2 12.6 Table 13 Electrical properties of an Al doped ZnO film Deposited at 200c Deposition Temperature Carrier Concentration (cm–3) Mobility (cm2/Vsec) 200c 6.20 e20 17.8

PAGE 77

65 Figure 54 Effects of 200c ann eal on the absorption coeffici ent of the films deposited at different substrate temperatures (visible range) Figure 55 Effects of 200c ann eal on the absorption coeffici ent of the films deposited at different substrate temperatures (IR range) Films sputtered at 25c and 100c were annealed at 200c in Ar ambient for 45 minutes to study the effect of heat treatment on these films. As shown in Figure 54 and Figure 55, transmission of the film was slightly improve d after the heat treatment. From Table 12, it is clear that mobility and dopant efficiency of the film were enhanced due to the heat treatment. 0 20000 40000 60000 80000 350450550650 Wave length (nm )absorption coeff.( cm1) + Reflection losses As made (sub.temp = 100c) Annealed (sub.temp = 100c) As made (sub.temp = 25c) Annealed (sub.temp = 25c) 0 2000 4000 6000 8000 10000 70090011001300 Wave length (nm )absorption coeff.( cm1) + Reflection losses As Made ( temp = 100c) Annealed (sub.temp = 100c) As made (sub.temp = 25c) Annealed (sub.temp = 25c)

PAGE 78

66 Figure 56 Absorption coefficient of the films sputtered at 1 00c and 25c (after the 200c anneal) is compared with that of AZO sputtered at 200c su bstrate temperature (visible range) Figure 57 Absorption coefficient of the films sputtered at 1 00c and 25c (after the 200c anneal) is compared with that of AZO sputtered at 200c subs trate temperature (IR range) Optical properties of the annealed films a nd a film deposited at 200c are compared in Figure 56 and Figure 57. The electr ical properties of a film prep ared at 200c are listed in Table 13. It is clear that th e electrical and optical propertie s of the annealed films could not reach the values for the film sputtere d at 200c. Annealin g temperature was not sufficient enough to improve the crystallinity of the film and XRD analysis could not detect any significant change after the heat treatment. The enhancement of optical and electrical properties can solely attributed to the improveme nt of the stochiometry and defect structure due to the heat treatment. 0 20000 40000 60000 350450550650 Wavelength (nm)Absorption Coeff. + Reflection losses (/cm) Annealed (sub.temp = 100c) Annealed (sub.temp = 25c) As made (sub.temp = 200c ) 0 5000 10000 15000 700900110013001500 Wavelength (nm)Absorption Coeff. + reflection losses(/cm) Annealed (sub.temp = 100c) Annealed (sub.temp = 25c) As made (sub.temp = 200c )

PAGE 79

67 Table 14 Electrical properties of the AZO film deposited at room temperature and annealed in Oxygen and Ar ambient at 200c for 45 min Deposited at 25c Annealed at 200c in 100%Ar Annealed at 200c in 50% O2 +50%Ar Step 0 Step 1 Step 2 Carrier Concentration (cm-3) 6.0E 19 3.1 E20 3.0 E 20 Mobility (cm2/V sec) 2.2 6.8 8.2 Absorption coeff (/cm) at 500nm 10050 11830 12460 AZO sputtered at room temperature was anneal ed at 200c in the presence of Oxygen to promote chamiabsorption of Oxygen from the ambient. Table 14 show s the electrical and optical properties of these films. It is cl ear from the observations that an annealing temperature of 200c was not sufficient enough to initiate significant chamiabsorption of the Oxygen from the surface of a film. The improvement of electrical properties was entirely temperature enhanced. 5.4 Proposed Sputtering Pa rameters for ZnO Deposited ZnO films are categorized into three groups according to their optical and electrical properties (Figure 58 and Figure 59). Films belonging to first group are opaque and conductive. Such films are characterized as a mixture of Zinc and Zinc oxide. These films are prepared at lower substrate temp erature or/and higher s puttering voltages with less oxygen flow rates. Films of the third group have a composition close to stoichiometry of ZnO and are transparent and non-conductive. Films of this group are most suitable for the resistiv e layer of zinc oxide on sola r cells. Group two films have intermediate properties. Tran sparent as well as conductive f ilms fall in this category.

PAGE 80

68 Undoped films sputtered at lower voltages (440 V), moderate temperature (200c) and higher oxygen concentration (~50%) can be plac ed in category three. These films showed resistivity more than 10 ohmcm and average absorption coe fficient in the range of 1600 cm-1. Change in the electrical resistivity was not detected after the 325c anneal. These stoichiomentric films showed good thermal stability. Figure 58 Effect of oxygen concentration and sputtering voltage on the electrical and optical properties of ZnO Figure 59 Effect of Oxygen con centration and the substrate te mperature on the electrical and optical properties of ZnO %O2 360380400 420 440460480500 10% 20% 30% 40% 50% Sputtering Voltage in Volt Opaque & Conductive Transparent & Conductive Transparent & NonConductive 25100200325 10% 20% 30% 40% %O2Substrate Temperature oC Opaque & Conductive Transparent & Conductive Transparent and non conductive

PAGE 81

69 Transparent and non-conductive ZnO (category 3) should result in an ideal conductive transparent film when doped with Aluminum As shown in Figure 58 and Figure 59, such film can be deposited with higher su bstrate temperature or within the O2 rich ambient. An upper limit of the substrate temperature wa s set to 350c as AZ O films studied during this work were intended to be used on CIS solar cells. In this range of temperatures, stochiometric film can be deposited either w ith lower sputtering volta ges or with higher oxygen concentrations. As dopant effect of Al is achieved around 440 V, the only way to sputter doped films in category three is to use at least 40% oxygen concentration in the sputtering ambient. On the other hand, spu ttered aluminum goes into a film as a passivated oxide if sputtered with oxygen c oncentration higher than 29%. Due to these practical limitations, all Al doped films sputtered during this work fall in the group two categories. The optimum combination of optical and elec trical properties was repetitively shown by the films sputtered with 480 Volt, 3 mTorr pressure, 30% O2 concentration and 325c substrate temperature. Spatial distribution of resistivity and deposition rate with respect to the target position are shown in Figure 60. Figure 61 shows op tical properties of this film. The best-achieved properties of reactivel y sputtered AZO reported by other groups are comparable with these results [47,12].

PAGE 82

70 Zone A B C D E Deposition Rate 2.56A/s 1.71A/s 1.23A/s 0.85A/s .53A/s Resistivity: A B C D E Figure 60 Resistivity (x10-3 ohm-cm) of AZO on 10 cm x 10cm substrate relative to the location of sputtering source Figure 61 Optical properties of AZ O in different sputtering zones 5.5 Comparison between AZO Sputtered from Oxide Target and Metal Target ZnO window layers have been successfully sputtered form Oxide targets with the RF sputtering technique in the USF semiconducto r lab. Properties of reactively sputtered ZnO were compared with such films in orde r to examine its prospective for CIS solar cell. 0.63 0.45 0.50 0.55 0.71 1.60 0.44 0.42 0.44 0.60 0.90 0.47 0.48 0.54 0.66 0.59 0.48 0.56 0.67 0.68 0.48 0.52 0.64 0.75 0.83 Target 0 2,000 4,000 6,000 8,000 10,000 12,00040 0 5 00 6 00 70 0 8 00 900 1 000 1100 1200 1300 1400 150 0Wavelength (nm)Absorption Coeff. /(cm) + reflection losses A C E

PAGE 83

71 The window layer of CIS solar cells has been spu ttered from a Zinc Oxide target at 100c. Electrical properties of a typical film are shown in Table 15. Properties of a film sputtered from a zinc target using reactive dc sputtering are listed in Table 16. In order to make the fair comparison, both films were deposited at 100c using similar experimental setups. Faster deposition rate is the main advantag e of DC sputtering. Lower heat conductivity and heat dissipation in Oxide targets puts an upper limit to RF sputtering power. On the other hand, metal targets withstand highe r power densities w ithout heat-induced damages. Sputtering power as high as 200 W was used without damaging the Zn target during this study. We can probably go to k ilowatts in power without heating the Zn target. Whereas, the safe power limit for a ZnO target with the same experimental set up was under 100W. As deposition rate is dir ectly proportional to the sputtering power, higher growth rate was achievable wi th the DC sputtering technique. The Films compared in Tables 15 and 16 we re deposited with sputtering power under 100W. When sputtered with the same electrical power, growth of the film was faster with the Zn metal target. This can be attributed to the difference in the sputtering yields of both the materials. Since the bi nding energy of ZnO is greater than that of the Zn, the sputtering yield of Zn is highe r than its oxide [26]. In othe r words, the higher deposition rate of a film sputtered from a Zn target can be attributed to the f act that oxides sputter much more slowly than a pure metal.

PAGE 84

72 Table 15 Electrical properties of an Al do ped ZnO film sputtere d by RF sputtering Deposition Rate (/s) A B C D E 1.22 1.01 0.76 0.58 0.41 Resistivity (x 1e-3 Ohm-cm) 8.3 2.4 1.4 1.9 4.3 24 2.8 1.2 1.7 4.7 89 4.2 1.2 1.5 4.5 90 5.3 1.2 1.6 3.7 93 4.5 1.2 1.6 3.9 34 2.9 1.2 1.8 4.1 15 2.2 1.2 1.9 4.7 8.2 1.8 1.2 2.1 5.7 Table 16 Electrical properties of Al doped Zn O film sputtered by DC reactive sputtering Deposition Rate ( /s) A B C D E 1.65 1.52 1.07 0.69 0.55 Resistivity (x 1e-3 Ohm-cm) 1.7 1.2 1.5 1.6 1.3 2.7 1.3 0.97 0.92 1.1 4.0 1.6 0.94 0.84 1.3 6.2 1.7 0.94 0.80 1.0 7.7 1.7 0.97 0.81 1.0 7.6 2.7 0.45 0.83 .98 4.6 2.0 0.94 0.81 1.0 3.3 1.8 0.96 0.94 1.0

PAGE 85

73 Spatial distribution of electrical resistivity was uniform when sputtered with reactive sputtering (Table 15 and 16). Positional depend ence of resistivity of AZO sputtered from the oxide target has been attributed to the bombardment of highly energetic oxygen species [53,54]. The origin of these high-energy Oxygen species can be the oxygen atoms in the oxide target [48]. The flux of high-energy neutrals is smaller for Zn metal targets than for Oxides target [40]. Moreover, th e ratio of atomic to molecular oxygen was reported to be higher in the case of th e RF sputtering mode [55]. Less position dependence of resistivity, seen in DC reac tively sputtered films suggests less high-energy oxygen bombardment on the grounded substrate. As mentioned previously (chapter 5.5), most of the Al doped films were sputtered in the Oxygen deprived sputtering ambient. Less Oxygen concentration in the sputtering ambient might also have prevented the damage caused by high-energy Oxygen species, wh ich resulted in less sp atial variation in the resistivity of the reactively sputtered films. Figure 62 Comparison of the op tical properties of AZO films sputtered from Zn and ZnO targets AZO Films sputtered with reactive sputtering were mo re conductive than the RF sputtered films. Optical parameters of both films are also comparable (Figure 62). More flexibility of sputtering parameters offe red by reactive sputtering provided a large 0 5000 10000 15000 20000400 5 00 600 700 800 900 1000 1100 1200 1300 1400 1500Wavelength (nm )Absorption Coeff. + reflection losses (/cm) ZnO target Zn target

PAGE 86

74 spectrum of deposition combinations. The best combinations of el ectrical and optical properties were achieved with DC reactive sputtering. Compared to the metal target, control of film stoichiometry was easier when the Oxide target was sputtered. Undoped films sputtered from the oxide target often showed high resistivity and were less metal rich. Films must be sputtered in high Oxygen concentration and with slower deposition rate from the Zn target to achieve similar properties. Use of an Oxide target is r ecommended over a metal target for undoped ZnO deposition. Metal targets are cheaper than Oxide target s. Besides, a variety of compound metal targets can be easily fabricated. Faster deposition rate and less positional dependence of the properties result in high throughput when a Zn target is used. Moreover, the high cost of RF generators, matchboxes and waveleng th interference make s the RF sputtering technique more costly [8]. Sputtering from a Zn target is less simplifie d but more economic. DC reactive sputtering offers better flexibility to establish optimum combinations of process parameters. Such flexibility increases the opportunity of achieving the best combinations of desirable properties for window layer ZnO.

PAGE 87

75 5.6 Performance of Reactively Sputtered ZnO Window Layers on CIS Solar Cells 5.6.1 CIS/CdS/ZnO Devices Deposition of ZnO window layers is the final step of CIS solar cell fabrication. The process of window layer deposition should not only grow a tran sparent and conductive oxide, but should also cons erve the properties of th e underlying material. The critical parameter for preparation of the front contact is the substrate temperature. The interdiffusion of Cd into Cu ternaries has been reported at temperature greater than 300c. A substrate temperature above 200c cause s Interdiffusion of Se into S [8]. The ZnO window layer must be prepared under 200 c to prevent the adverse effects on the underlying CIS layer. To start with both, undoped and doped, layers of ZnO were sputtered at 200c. All of the devices completed with this window layer di d not show diode characteristics during IV measurement. In other words, the devices were shorted. These shorted devices suggested that other than inter diffu sion of Sulfur and Selenium, The reactive nature of the ZnO deposition coul d also be a decisive parameter for the substrate temperature. To st udy the effect of sputtering am bient on the performance of solar cells, CIS solar cells with conventional RF sputtered ZnO were annealed at 200c in Ar and O2 for 45 min.

PAGE 88

76 Table 17 (h19-23 H17-08): Diode properties of a CIS solar cell after 200c anneal in Ar and O2 ambient Annealed in Ar at 200c Annealed in 75%Ar 25% O2 at 200c Voc Before Annealing 433mV 419mV Voc After Annealing 420mV 383mV Jsc before annealing 39.2 mA/cm2 36.1 mA/cm2 Jsc after annealing 40.5 mA/cm2 42.9 mA/cm2 Figure 63 Comparison of the spectral responses of a cell annealed in Oxygen and Ar ambient As shown in Table 17, a 9 % decrease in Vo c was observed when the film was annealed in 75% Ar and 25% oxygen, whereas the decreas e in Voc was 3% when annealed in pure Ar atmosphere. This suggested that the CdS layers in these devices were not thick enough to protect diffusion of O2 into the CIS absorber layer at 200c [j.Kessler 1996]. Spectral response for these devices is shown in Figure 63. In order to minimize the adverse effect of O xygen, we tried to deposit window layers at lower substrate temperatures. Doped and undoped layers of ZnO were sputtered at room 0 0.5 1 1.5400 480 560 640 720 800 880 960 1040 1120 1200 1280 1360Wavelength (nm)Q.E After Annealed in O2 After Annealed in Ar

PAGE 89

77 temperature on CIS absorption layer. The comp leted device was then annealed in Ar at 200c for 45 minutes in order to improve the el ectrical and optical properties of the ZnO layer. Open circuit voltages of this device a nd reference device are compared in Table 18. Table 18 (h19-23, h19 r23 get h19) Diode properties of the device fabricated with reactive ZnO and the reference device. Reactive ZnO was sputtered at room temperature and annealed in Ar ambient at 200c for 45 min Reference cell with RF sputtered ZnO Cell with reactively sputtered ZnO Voc 433mV 400 mV Jsc 39.2 mA/cm2 28.2 mA/cm2 FF 60% 56% % EFF 10% 6.2% Figure 64 Spectral response of the reference device a nd a device fabricated with a reactive ZnO window layer. The reac tive ZnO layer was deposited at room temperature and annealed in Ar ambient As shown in Figure 64, the current density of the device completed with reactively sputtered ZnO was 28.2 mA/cm2 which was 10 mA/cm2 less than the value for the reference cell. Lower transmission (70 %) and higher sheet re sistance (36 ohm-cm2) of DC sputtered ZnO were responsible for poor performance of the device. This device 0 0.2 0.4 0.6 0.8 1 4006008001000120014001600 Wavwlength (nm) Q. E Referance Deposited

PAGE 90

78 suggested that ZnO should be sputtered at least at 100c wh en used as a transparent conductive oxide. Figure 65 Spectral response of the reference device a nd a device fabricated with a reactive ZnO window layer. The undope d ZnO was sputtered at room temperature. The AZO was sputtered at 100c Table 19 (7h20-2r,h20-07) Diode properties of a re ference device and a device with DC sputtered ZnO. Undoped ZnO was sputtered at room temperature and AZO was sputtered at 100c Reference Cell with RF Sputtered ZnO Cell with Reactively Sputtered ZnO Voc 367mV 376mV Jse 35.9 mA/cm2 33.6 mA/cm2 FF 56% 57% % Eff 7.3% 7.2% Duration of Window layer deposition ~ 2hrs 30 min ~ 1 hrs 30 min Thickness of I-ZnO layer 500 900 Thickness of AZO 5000 3000 In another set of experiments, undoped ZnO wa s sputtered at room temperature to avoid the exposure of under lying CdS to Oxygen at high temperature. A 2000A thick Al doped ZnO layer was reactively sputtered at 100c on the undoped ZnO layer. 0 0.2 0.4 0.6 0.8 1 350550750950115013501550 Wavelength (nm)Q.E. Reference Deposited

PAGE 91

79 Performance of this device and the reference de vice were practically id entical (Table 19). Spectral response of a cell wi th a reactively sputtered Zn O window layer is compared with that of a RF sputtered ZnO window layer in Figure 65. The deposition of the reactively sputtered ZnO was faster than RF sputtered ZnO. Moreover, a thinner layer of doped ZnO of fered acceptable surface conductivity. Use of RF sputtered ZnO was not only costly but also time consuming compared to the use of reactively sputtered ZnO. There is enough room for other process co mbinations, which can further improve the performance of CIS solar cells with the r eactively sputtered ZnO window layer. These results encourage us to believe that reac tively sputtered ZnO can not only replace RF sputtered ZnO window layer on CIS solar cell but also be proven superior to the conventional window layer. 5.6.2 Direct CIS/ZnO Devices Before the ZnO deposition, a CdS buffer layer is deposited using a chemical bath deposition (CBD) technique. Device fabricati on, without a wet chemical step of the highly toxic CdS deposition, is an on going i nvestigation. Many groups have reported the significance of the CdS layer in preventing the CIS layer from the exposure of oxygen during the front contact formation [3,6,8]. Some of the groups have successfully fabricated Cd free CIS solar cells with 10.5% efficiency [6]. These reports suggested the use of a non-O2 containing RF plasma process. Addition of Oxygen in the sputtering ambient severely degraded the performance of the Cd free cell [6,3]. We tried to explore the effect of the react ive nature of DC sputtering on the Cd free solar cell. Reactively

PAGE 92

80 sputtered ZnO was deposited on the CIS layer, without CdS deposition to examine the future of this deposition tec hnique for Cadmium free devices. Deposition of intrinsic ZnO is the most critic al process in Cd free devices, as the undoped ZnO layer is directly deposited on the absorp tion layer. This layer was deposited with reactive sputtering, whereas the doped ZnO layer was deposited with standard RF sputtering. The undoped ZnO layer was sputte red at room temperature with 33% oxygen concentration. Doped ZnO was sputtered us ing conventional RF s puttering at 100c. Table 20 compares the device properties of al l reactively sputtered ZnO junction devices with the literature results. Table 20 (H34 14/H34) Diode prop erties of CIS/ZnO devices compared with literature results CIS/ZnO (Literature)[28] CIS/ZnO (USF) Voc (mV) 390 350 Jsc (mA/cm2) 34.6 32.1 FF 39% 35% Figure 66 Spectral response of a CIS/ CdS/ZnO device and a CIS/CdS device 0 0.2 0.4 0.6 0.8 1 4006008001000120014001600 Wave Length (nm)Q.E. Cis/Cds/ZnO CIS/ZnO

PAGE 93

81 Spectral responses of these devices are compared in (Figure 66). The CdS free cell showed 350 mV Voc and 32.1 mA/ cm2 of Jsc. During the study, the Voc’s of all Cd free devices were within 10-20% of the values for the conventional ce lls. However, Jsc was affected most when the Cds layer was eliminat ed. Almost 10% loss in Jsc was seen in the spectral response (Figure 66). Ma ny devices showed shunting in the dark IV curve. Fill factor was also low compared to CdS contro l devices. Low open circuit voltage and fill `factor degraded the efficiency of this de vice. Other groups have also noticed such degradation in CdS free cells with RF sputtered ZnO window layers [6]. Results of this experiment encourage us to conclude that despite the use of oxygen in the sputtering ambient, reactive sputtering can still be a candidate for direct CIS/ZnO devices. Less damage caused by energetic oxygen spices and O2 deficient plasma processes of DC sputtering may favor th e endeavor of Cd free fabrications.

PAGE 94

82 CHAPTER 6 CONCLUSION This work demonstrated that the reactive s puttering technique could be successfully used for window layer deposition on Cu(In,Ga)Se2 solar cells. Films sputtered with reactive sputtering showed better electrical and optical properties compared to the films sputtered with conventional RF sputteri ng. Reactive sputtering techni que was less time consuming and cheaper compared to the RF sputtering technique. Conductivity of the undoped ZnO was a result of oxygen vacancies and excess Zn. Typical carrier concentrati on of such films was 5 x 1019 cm-3. Mobility of carriers in undoped ZnO showed dependence solely on grain boundary scattering and defectinduced scattering. Unlike Al doped ZnO; these films did not show a consistent relationship between carrier c oncentration and mobility. Typi cal carrier concentration of Al Doped ZnO was 5x 10 20 cm-3. Mobility of these films showed dependence on carrier concentration as well as on mi crostructure of the film. Undoped films sputtered at lower voltages (440 v), moderate temperature (200c) and higher oxygen concentration (~50% ) were stoichiometric and suitable for the buffer layer application. These films showed resistivit y more than 10 ohm-cm and average absorption coefficient in the range of 1600 cm-1. These ZnO films were stable at higher temperatures. The optimum combination of optical and electrical properties were

PAGE 95

83 repetitively shown by the AZO films sputte red with 480 Volt, 3 mTorr pressure, 30% oxygen concentration and 325c substrate temperature. Growth of the film was more temperature dependent when sufficient flux of Zn and Oxygen was available near the substrate surfac e. The film sputtered at lower substrate temperature was inferior. The heat trea tment improved the optical and electrical properties of such films. Elect rical and optical properties of the film sputtered at higher temperature could not be achieved by anneali ng the film sputtered at lower temperature. The minimum temperature, which could produce reasonable electrical and optical properties, was 100c. Oxygen in the sputtering ambient degraded the absorbent layer during front contact preparation. The adverse e ffect of oxygen was minimized when the undoped layer was deposited at room temperature. Performance of devices prepared with such procedures was comparable with devices fabricat ed with a conventional ZnO layer. The elimination of a CdS buffer layer degraded the Fill Factor and Jsc of the device. However, we were successful in improving Vo c. Results of this study encouraged us to conclude that despite the use of oxygen in th e sputtering ambient, reactive sputtering can still be a candidate for direct CIS/ZnO devices.

PAGE 96

84 REFERENCES 1. J.R. Tuttle, J.S. Ward, A Duda T.A.Bern es, M.A.Contretas, K.R,Ramanathan,A.L. Tennant,J.Keane, E.Dcole, K.Emery and R.Noufi “ The Performance of Cu(InGa)Se2 Based Solar Cells in Conventional and Concentrator applications ”, Proceeding of th e 1996 spring MRS meeting San Fran, Ca, 1996. 2. K.Zweible, “ Progress in Photovoltaics ”, Special issue on thin films (1995). 3. J.Kessler, S.Wiedeman, L.Russel, J. Fogleb och, S.Skibo, R.Arya, and D.Carlson, “ Front Contact Optimization for Cu(InGa)Se2 (SUB) Modules ”, 1996 IEEE 25th PVSC May 1996 Washington D.C. 4. J.Kessler, K.O.Velthaus, M.Rouck, R.Laichinger, H.W.Schock, D.Lincot, R.Ortega Borges and J.Vedel. Proc 6th International PVSEC New Delhi, Oxfprd and IBH Publ., New Delhi 1992 p.1005. 5. J. Hedstrom, M.Bodegard, A.Kylner, L.Stolt, D.Heriskos, M. Ruch and H.W.Shock IEEE 1993. 6. J.Kessler, M.Ruckh, D.Hariskos, U.Ruhle, R.Menner and H.W.Schock, “Interface Engineering Between Cu(InGa)Se2 and ZnO” IEEE 1993. 7. R.W.Birkmire, B.E.McCandles, W.N.Sh afarman and R.D. Varrin, Jr.Proc 9th EC PVSEC, Freiburg Kluwer, Dord rect (1989) p.134. 8. M.Ruckh, D.Hariskot, U.Ruhle and H.W.Schock. “ Application of ZnO in Cu(InGa)Se2 Solar Cells” 1996 IEEE 25th PVSC May, 1996, Washington D.C. 9. Physics of Semiconductor Devices by S.M.Sze. 10. Erginsoy C 1950 Phys.Re 79 1013 (Semiconductor Thin Films). 11. Johnson VA and Lark, Horovitz K 1947 Phys. Rev 71 – 374 (Semiconductor Thin Films ). 12. S.Zafar, C.S.Ferekides and D.Morel. “Characterization and Analysis of ZnO Al Deposited by Reactive Magnetron Sputtering”, J.Vac.Sci.Technol A13 (4) 1995. 13. Solar Cell Device Physics by Stephen J. Fonash. 14. Seto JYW 1975 J.Appl.Phy 46 – pg 5247 (Semiconductor Thin Films).

PAGE 97

85 15. Padovani FA and Stratton R, 1966 Solid State Electron 9 pg 695 (Semiconductor Thin Films). 16. Crovell CR and Rideout VL 1969 Solid State Electron 12 pg 89 (Semiconductor This Files). 17. Semiconducting Transparent Thin Films by H.L.Hartangel, A.L.Dawar, A.K.Jain and C Jagadish. 18. Roth AP and Williams DF 1981 J. Appl Phys. 52 pg86685.[17]. 19. Sinois F, Leiji MV and Hoogendoorn CJ 1979 Solar Energy Mater 1 pg 221. 20. Smith RA 1978 Semiconductors (Cambridge, Cambridge University Press) pg 294.[17]. 21. Krokoszinski HJ and Oesterlein R1990 Thin Solid Films 187 pg179 [17]. 22. Burstein E 1954 Phys. Rev 93 Pg632, [17]. 23. Moss TS 1964 Proc. Phys.Soc B67 Pg 755, [17]. 24. Sarkar A, Ghosh S, Chaudhri S and Pal AK 1991, Thin Solid Films 204 pg 255, [17]. 25. Minami T, Oshashi K, Takata S, Mouri T and Ogawa N 1990 Thin Solid Films 193/194 Pg721. 26. Semiconductor Materials and Process Technology Handbook by G.E.McGuire Noyes Publications 1988. 27. Zinc Oxide Rediscovered by Harvey Brown, The New Jersey Zinc Company 1957. 28. R.Bhatt, H.Shankernarayan, C. S.Ferkides and D.L.Morel 26th IEEE Photovoltic Specialists Conference 383 ,IEEE1997. 29. Chelikowasky JR 1977 Solid State Commun. 22 pg 361. 30. Hustson Bul A.M. 1956 Phys.Soc .Series 2. Vol 1 381 [27]. 31. Hagemark KL and Chaccka LC 1975, J.Solid State Chem. 15 pg 261. 32. Gopel W and Lanoe U 1980 Phys.Rev B 22 pg 6447. 33. Akarussaman AF, Sharma GL and Malhotra LK 1985, Thin Solid Films 193/194. 34. Minami T, Sato H Sonada T, Nato H and Takata S 1989 Thin Solid Fils 171 pg.307. 35. Hu J and Gordon R 1992 J.Appl.Physics 71 pg.880, [17]. 36. Chi BH, Song JS, Yoon Kh 1990 Thin Solid Films 193/194 pg 712,[17].

PAGE 98

86 37. Sarkar A, Ghosh S, Chudhari S and Pal AK 1991, Thin Solid Films 204/255. 38. Gary, J Amer, Ceramic Soc 1954, 37:534,[27]. 39. Barns Jo, Leary Dj and Jordan AG 1980 J. Electrochem. Soc. 127 pg 1636 [17] 40. S,B Krupanidhi and M.Sayer “Position and Pressu re Echen ffects in RF Magnetron Reactibe Sputter Deposition of Piezo electric Zinc Oxide”, J. Appl.Phys, 56 pg 3308. 41. Wen S Chen J.M. Stewart, R.A. Mieckelsen, W. E. Devaney and B.J. Stanbery “ Research of Phlycrystalline This n Film CuGaInSe2 Solar Cells “, Boeing defense and space group final technical report for the period May 3rd 1991 through May21 1993. 42. Leja E, Budzynska K, Pisarkiewicz T and Stapinski T 1983 This Solid Films 100 pg.203. 43. Donaghy L.F. and Gerghty K.G. 1976 Thin Solid Films 38 pg.835. 44. Shinoki F and Itoh A 1975 Appli. Phys. 46 pg.3381. 45. Tsuji N, Komiyama H and Tanaka L 1990 Japan J. Appl. Phys 29 pg.835. 46. T.Minami, H.Nata and S.Takata “Highly Conductive and Transparent Aluminum Doped Zinc Oxide Thin Films Prepared by RF Magnetrone Sputtering, Japanese J. Apply.Phys 1984 23, pg L280. 47. Z.C. Jin, I.Hamberg and C.G.Granqivst, “Optical Properties of Sputter Deposited ZnO:Al Thisn Films”, J.Appl.Phys. 1988 pf5117. 48. Kikuo Tominaga, Satoshi Iwamura, Yoshihiro Shinitani and Osamu Tada, “Energy Analysis of High Energy Neutral Atoms in The Sputtering of ZnO and BatiO3 ” Japanese J. Appl. Phys. 1982, 21 pg 688. 49. T.Minami H.Nato,S.Shooji and S.Takata 1984 ,Thin Solid Films 111, pg.167. 50. Schoenes J, Lanazawa K and Kay E 1977 J.Appl.Phys 48 pg.2537. 51. Minami T, Nato H and Takata S, 1985, Thin Solid Films 193/194. 52. Thin Film Process by Jhon Vossen and Werner Kern. Academic Press 1978. 53. T.Minami, H.Sato, H.Imamoto and S.Takata, “Substrate Temperature Dependence of Transparent Conducting Al Doped ZnO Thin Films Prepared by Magnetron Sputtering” J.Apl. Phys. 1992, 31, pg L257. 54. Yong Eui Lee and Jae Bin Lee “ Microstructure Evolution and Preffered Orientation Changes of Radio Frequency Magnetron Sputtered ZnO Thin Films”, J.Vac.Sci Technol.1996 A 14 pg 1943.

PAGE 99

87 55. K.Elimer, R.Wendt, R.Cebullas “ZnO/ZnO:Al Win dow and Contact Layer for Thin Film Solar Cells: High Rate Deposition By Simultaneous RF and DC Magnetron Sp uttering” 1996 IEEE 25th PVSC Washington D.C.


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 001917605
003 fts
005 20071126163724.0
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 071126s2007 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002055
040
FHM
c FHM
035
(OCoLC)182523257
049
FHMM
090
TK145 (ONLINE)
1 100
Bhatt, Rita.
0 245
Growth and characterization of ZnO for the front contact of Cu(In,Ga)Se2 solar cells using reactive sputtering techniques
h [electronic resource] /
by Rita Bhatt.
260
[Tampa, Fla.] :
b University of South Florida,
2007.
3 520
ABSTRACT: ZnO window layers for CIGS solar cells are grown with a DC sputtering technique instead of a conventional RF sputtering technique. Transparent window layers and buffer layers are sputtered from the Zn target in the presence of Oxygen. The window layer is doped with Aluminum in order to achieve high electrical conductivity and thermal stability. The effect of different sputtering parameters on the electrical and optical properties of the films is elaborately studied. Sets of annealing experiments are also performed. Combinations of different deposition parameters are examined to design the optimum fabrication conditions. We are able to deposit 85% transparent, Al doped ZnO films having 002-axis orientation and 4e-4 ohm-cm resistivity, which is successfully, used on CIGS solar cells. Resistivity of undoped ZnO buffer layers is varied form 10-2 ohm-cm to unmeasurable by varying the sputtering parameters. The performance of a reactively sputtered window layer and a buffer layer have matched the performance of the RF sputtered ZnO on CIGS solar cells. There has been considerable effort to eliminate Chemical Bath Deposition of the CdS buffer layer from CIS solar cell fabrication. The performance of an undoped DC sputtered ZnO layer is examined on Cd free CIGS solar cells. The ZnO buffer layer is directly sputtered on an underlying CIGS material. The performance of Cd free solar cells is highly susceptible to the presence of Oxygen in the sputtering ambient of the buffer layer deposition [6]. As Oxygen is a growth component in reactive sputtering, the growth mechanisms of the DC-sputtered buffer layer are studied to improve the understanding. The performance of all reactively sputtered ZnO devices matched the values reported in the literature and the results for DC sputtered ZnO on Cd-free solar cells were encouraging.
502
Thesis (M.S.)--University of South Florida, 2007.
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 87 pages.
590
Advisor: Don L. Morel, Ph.D.
653
ZnO window layer.
Semiconducting transparent thin films.
Al doped ZnO.
Sputtering parameters.
Metal target sputtering.
690
Dissertations, Academic
z USF
x Electrical Engineering
Masters.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.2055