<%BANNER%>

Examination of Zno, CuCrO_2, and CuScO_2 Properties for Use in Transparent Electronics and Chemical Sensors

Permanent Link: http://ufdc.ufl.edu/UFE0024867/00001

Material Information

Title: Examination of Zno, CuCrO_2, and CuScO_2 Properties for Use in Transparent Electronics and Chemical Sensors
Physical Description: 1 online resource (166 p.)
Language: english
Creator: Sadik, Patrick
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: adsorption, cucro2, cusco2, delafossite, device, electronic, junction, lumenescense, pl, pn, seebeck, semiconductor, surface, tco, thiol, zno
Materials Science and Engineering -- Dissertations, Academic -- UF
Genre: Materials Science and Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Greater progress in the field of TCO (transparent oxide semiconductor) research has been hampered by a lack of availability of p-type candidates. Though the n-type TCO s are well documented and easily produced only one well documented oxide has spurred continued efforts at producing robust, p-type behavior, namely ZnO. Though p-ZnO has been produced, a coterminous research effort into improving the film qualities of lesser known naturally p-type TCO s could prove fruitful. Thus, this research has examined the thin film properties of two delafossites, CuCr_1-xMg_xO_2 and CuSc_1-xMg_xO_2 as well as the surface properties of ZnO. The delafossites, CuCr_1-xMg_xO_2 and CuSc_1-xMg_xO_2 have been shown to have the superlative properties of high conductivities (220 S cm-1 and 70 S cm-1) and high transparencies in 400 nm thick films of ~40% and ~80% throughout the visible spectrum. We have also been the first to describe a near band edge photo-luminescence in CuSc_1-xMg_xO_2 for 0.00 < x < 0.03, opening the possibility of creating delafossite active layer light emitting diodes. The greatest challenge in creating viable pn-junctions between the delafossites and ZnO is preventing an in situ reaction between CuCr_1-xMg_xO_2 and ZnO that creates a (Mg/Cu)Cr_2O_4 spinel interface. We have found that a strategy of using a CuCr_1-xMg_xO_2/CuSc_1-xMg_xO_2/ZnO on c-Al_2O_3 in two ways may allow the growth of either a buffer/p/n or a p/i/n structure. Specifically, growing CuCrO_2 at 700 masculine ordinalC as a 100 nm nucleation layer followed by a 750 masculine ordinalC growth of CuSc_1-xMg_xO_2 and a 400 masculine ordinalC growth of ZnO secures an epitaxial growth for the entire structure despite large lattice mismatches between c-Al_2O_3/CuCrO_2 (8.7%) and CuCrO_2/CuScO_2 (7.7%). This combined research takes the first steps in making robust pn-heterojunctions possible. Also, in an effort to progress an understanding of the surface chemistry of ZnO we provide results showing marked difference in the chemisorptions of dodecane thiol upon Zn- and O-terminated faces of ZnO.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Patrick Sadik.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Norton, David P.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024867:00001

Permanent Link: http://ufdc.ufl.edu/UFE0024867/00001

Material Information

Title: Examination of Zno, CuCrO_2, and CuScO_2 Properties for Use in Transparent Electronics and Chemical Sensors
Physical Description: 1 online resource (166 p.)
Language: english
Creator: Sadik, Patrick
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: adsorption, cucro2, cusco2, delafossite, device, electronic, junction, lumenescense, pl, pn, seebeck, semiconductor, surface, tco, thiol, zno
Materials Science and Engineering -- Dissertations, Academic -- UF
Genre: Materials Science and Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Greater progress in the field of TCO (transparent oxide semiconductor) research has been hampered by a lack of availability of p-type candidates. Though the n-type TCO s are well documented and easily produced only one well documented oxide has spurred continued efforts at producing robust, p-type behavior, namely ZnO. Though p-ZnO has been produced, a coterminous research effort into improving the film qualities of lesser known naturally p-type TCO s could prove fruitful. Thus, this research has examined the thin film properties of two delafossites, CuCr_1-xMg_xO_2 and CuSc_1-xMg_xO_2 as well as the surface properties of ZnO. The delafossites, CuCr_1-xMg_xO_2 and CuSc_1-xMg_xO_2 have been shown to have the superlative properties of high conductivities (220 S cm-1 and 70 S cm-1) and high transparencies in 400 nm thick films of ~40% and ~80% throughout the visible spectrum. We have also been the first to describe a near band edge photo-luminescence in CuSc_1-xMg_xO_2 for 0.00 < x < 0.03, opening the possibility of creating delafossite active layer light emitting diodes. The greatest challenge in creating viable pn-junctions between the delafossites and ZnO is preventing an in situ reaction between CuCr_1-xMg_xO_2 and ZnO that creates a (Mg/Cu)Cr_2O_4 spinel interface. We have found that a strategy of using a CuCr_1-xMg_xO_2/CuSc_1-xMg_xO_2/ZnO on c-Al_2O_3 in two ways may allow the growth of either a buffer/p/n or a p/i/n structure. Specifically, growing CuCrO_2 at 700 masculine ordinalC as a 100 nm nucleation layer followed by a 750 masculine ordinalC growth of CuSc_1-xMg_xO_2 and a 400 masculine ordinalC growth of ZnO secures an epitaxial growth for the entire structure despite large lattice mismatches between c-Al_2O_3/CuCrO_2 (8.7%) and CuCrO_2/CuScO_2 (7.7%). This combined research takes the first steps in making robust pn-heterojunctions possible. Also, in an effort to progress an understanding of the surface chemistry of ZnO we provide results showing marked difference in the chemisorptions of dodecane thiol upon Zn- and O-terminated faces of ZnO.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Patrick Sadik.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Norton, David P.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024867:00001


This item has the following downloads:


Full Text

PAGE 1

1 EXAMINATION OF Z n O, C u CrO2, AND C u S c O2 PROPERTIES FOR USE IN TRANSPARENT ELECTRONICS AND CHEMICAL SENSORS By PATRICK SADIK A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009

PAGE 2

2 2009 Patrick Sadik

PAGE 3

3 To my wife Carrie and our families for their constant and continued support

PAGE 4

4 ACKNOWLEDGMENTS I thank my parents for being a constant source of encouragement without ever being nagging or overwhelming. I thank my parents in law for their continued patience and belief that I could actually one day contribute to our familys income Most of all I tha nk Carrie. Being married to you has allowed me to take the risks necessary to accomplishing this dissertation. I love you.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................................... 4 LIST OF TABLES ................................................................................................................................ 8 LIST OF FIGURES .............................................................................................................................. 9 ABSTRACT ........................................................................................................................................ 12 CHAPTER 1 LITERATURE REVIEW ........................................................................................................... 14 Interest in Delafossites ................................................................................................................ 14 Traditional TCOs ....................................................................................................................... 14 Binary p Type Oxides ......................................................................................................... 14 TCO Coordination Classes .................................................................................................. 15 Octahedral class ............................................................................................................ 15 Tetrahedral class ........................................................................................................... 15 Linear class ................................................................................................................... 16 Delafossites .................................................................................................................................. 17 Delafossite Background ...................................................................................................... 17 Delafossite Polytypes .......................................................................................................... 19 p TCO Design Requirements .............................................................................................. 19 AI Cations: Effects on the Delafossite Electrical Structure .............................................. 21 Pt and Pd roles in delafossite properties ..................................................................... 21 The use of silver in the delafossite lattice ................................................................... 22 The use of copper in the delafossite lattice ................................................................. 23 Structural Considerations .................................................................................................... 23 c -Axis relationship to MIII cation radius ..................................................................... 24 Negative thermal expansion and links to bonding character ..................................... 25 Nature of A -A interactions .......................................................................................... 26 Minimal appearance of c axis dependence upon MIII radii ....................................... 27 Influence of A upon AMO2 a -axis lattice parameter ................................................. 27 Explanation of deviation of the c axis lattice parameter of AInO2 ........................... 28 The Ro le of Defects in p Type Behavior ........................................................................... 28 Oxygen Intercalation ........................................................................................................... 29 Explanation of Spinel Formation ........................................................................................ 32 Extrinsic Doping of the Delafossite Lattice ....................................................................... 34 Interest in CuCr1xMgxO2 system and CuSc1xMgxO2+ Systems ............................................. 36 2 ELECTRICAL AND STRUCTURAL PROPERTIES OF CuCr1 xMgxO2 THIN FILMS ..... 47 Background .................................................................................................................................. 47 Experimental Background .......................................................................................................... 48

PAGE 6

6 Results and Discussion ............................................................................................................... 49 XRD Analysis ...................................................................................................................... 49 Powder XRD: temperature dependence ..................................................................... 49 Powder XRD: O2 pressure dependence ..................................................................... 50 Thermodynamics of Cu, Cr, and O system ................................................................ 50 Powder XRD Mg dependence ..................................................................................... 50 Epitaxy with 4 -circle XRD .................................................................................................. 52 Surface Roughness: AFM .................................................................................................. 52 4 Point Van der Pauw resistivity Mg dependence ............................................................. 52 Optical Transmission: Mg Dependence ............................................................................ 53 Summary ...................................................................................................................................... 54 3 XPS AND ELECTRICAL CHARACTERIZATION of CuCr1xMgxO2 DELAFOSSITE THIN FILMS GROWN BY PULSED LASER DEPOSITION ............................................... 65 Introduction ................................................................................................................................. 65 Experimental Description ........................................................................................................... 66 Results and Discussion ............................................................................................................... 67 XRD Discussion: Mg dependence ...................................................................................... 67 Small Polaron Discussion .................................................................................................... 68 Summary of activation energies .................................................................................. 68 Temperature dependence of conduction mechanism ................................................. 69 Resistivity: Mg Dependence ............................................................................................... 69 Trans port Analysis: Seebeck Data ...................................................................................... 70 Optical Transmission Analysis ........................................................................................... 71 XPS Data: Mg Dependence ............................................................................................... 72 Explanation of fitting methodology ............................................................................ 72 Analysis: Cu 2p peak .................................................................................................. 72 Analysis: Cr 2p peak .................................................................................................... 73 Analysis: O 1s peak ..................................................................................................... 74 Analysis: Valence states .............................................................................................. 75 Conclusion............................................................................................................................ 75 4 ELECTRICAL TRANSPORT AND STRUCTURAL STUDY OF DELAFOSSITE CuSc1 xMgxO2 THIN FILMS GROWN BY PULSED LASER DEPOSITION ..................... 87 Interest in CuSc1xM gxO2 ............................................................................................................ 87 Experimental Procedure .............................................................................................................. 87 Discussion .................................................................................................................................... 89 Unbuffered CuSc1 xMgxO2 on c -Al2O3: Temperature Dependence ................................ 89 XRD orientation study ................................................................................................. 89 Optical absorption ........................................................................................................ 90 Seebeck study ............................................................................................................... 90 Unbuffered CuSc1 xMgxO2 on c -Al2O3: Mg Dependence ................................................ 91 CuCrO2 Buffered CuSc1xMgxO2 Films: Temperature Dependence ............................... 91 XRD orientation and phase composition study .......................................................... 92 Resistivity study ........................................................................................................... 93

PAGE 7

7 CuCrO2 Buffered CuSc1xMgxO2 Films: Mg Dependence ............................................... 93 XRD orientation and phase composition study .......................................................... 93 4 Circle XRD epitaxy .................................................................................................. 94 Use of Gradient Buffers ...................................................................................................... 95 XRD orientation study ................................................................................................. 95 Optical transmission and phase solubility .................................................................. 96 Buffered Strategy Comparison ........................................................................................... 96 ZnO Device Structures and Buffer Strategies .................................................................... 97 Conclusion ................................................................................................................................... 98 5 LUMINESCENCE AND OPTICAL PROPERTIES of CuSc1 xMgxO2 THIN FILMS GROWN BY PULSED LASER DEPOSI TION ..................................................................... 116 Introduction ............................................................................................................................... 116 Experimental Procedure ............................................................................................................ 117 Discussion .................................................................................................................................. 118 Crystallinity Study: XRD .................................................................................................. 118 Optical Transmission Properties ....................................................................................... 118 Photoluminescence Study ................................................................................................. 119 Comparison of optical transmission and PL data ..................................................... 119 Doping dependence of PL activity at 15 K and 300 K ............................................ 120 Major transitions ........................................................................................................ 120 Minor transitions ........................................................................................................ 124 Summary o f PL Effects ............................................................................................................. 124 6 FUNCTIONALIZING Zn AND O TERMINATED ZnO WITH THIOLS ........................ 136 Background ................................................................................................................................ 136 Possible Biosensor Applications ....................................................................................... 136 Thiol Chemistry ................................................................................................................. 137 Experimental Procedure ............................................................................................................ 138 Results ........................................................................................................................................ 138 Native Surface Chemistry of ZnO .................................................................................... 139 Effects of Heat Treatment of Thiol on ZnO Surface ....................................................... 139 Temperature profiles: O species desorption ............................................................ 139 Temperature profiles: thiol desorption ..................................................................... 140 Nature of Thiol Adsorption ............................................................................................... 140 Crystallinity of Thiol Adsorption ..................................................................................... 141 Summary .................................................................................................................................... 142 APPENDIX: XRD STRUCTURE DATA FOR RELATED CRYSTAL SYSTEMS ............... 151 LIST OF REFERENCES ................................................................................................................. 156 BIOGRAPHICAL SKETCH ........................................................................................................... 166

PAGE 8

8 LIST OF TABLES Table page 1 1 Table of known delafossite lattice parameters. .................................................................... 45 1 2 Compilation of doped delafossites from the literature. ........................................................ 46 3 1 XPS peak fitting values for the Cu 2p peaks. ....................................................................... 84 3 2 XPS values of doublet peak energy differences for Cu2O and CuO ................................... 84 3 3 XPS peak fitting values for the Cr 2p peaks. ........................................................................ 85 3 4 XPS values of doublet peak energy differences for Cr2O3 and CrO3. ................................ 85 3 5 XPS peak fitting values for the O 1s peaks. ......................................................................... 86 4 1 Compilation of various device constructs and their growth conditions. ....................... 115 4 2 Compilation of -scan and rocking curve data for device constructs ...................... 115 5 1 PL fitting data for CuSc1 xMgxO2 taken at 300 K with and features. ......................... 135 5 2 PL fitting data for CuSc1 xMgxO2 taken at 300 K with and features. .......................... 135 5 3 PL fitting data for CuSc1 xMgxO2 taken at 15 K with and features. ........................... 135 5 4 PL fitting data for CuSc1 xMgxO2 taken at 15 K with and features. ............................ 135 A 1. Atomic parameters for CuCr2O4. ........................................................................................ 151 A 2. 2 XRD peak positions for CuCr2O4. ............................................................................. 151 A 3. Atomic parameters CuCrO2. ................................................................................................ 152 A 4. 2 XRD peak positions for CuCrO2. ............................................................................... 152 A 5. Atomic par ameters for 3R CuScO2. .................................................................................... 153 A 6. 2 XRD peak positions for 3R CuScO2. ......................................................................... 153 A 7. Atomic parameters for 2H CuScO2. ................................................................................... 154 A 8. 2 XRD peak positions for 2H -CuScO2. ......................................................................... 154 A 9. Atomic parameters for Cu2Sc2O5. ....................................................................................... 155 A 10. 2 XRD peak positions for Cu2Sc2O5. ............................................................................ 155

PAGE 9

9 LIST OF FIGURES Figure page 1 1 Phase space of ternary oxides in the AMO2 system. ............................................................ 38 1 2 Crystal structure of delafossite 3R and 2H polytypes. ......................................................... 39 1 3 Flat band diagram of delafossite AMO2.. ............................................................................. 39 1 4 Basic construct of octahedral crystal field splitting. ............................................................ 40 1 5 Single [MO6] distorted octahedra. ......................................................................................... 40 1 6 Plots of calculated MIII radii against reported a and c lattice parameters. .......................... 41 1 7 Relationship of a lattice parameter to MIII radius ................................................................ 42 1 8 Triangular lattice of CuI cations with intercalated oxygen anion. ....................................... 42 1 9 Cu O plain for the oxygen intercalated delafossites(a) CuYO2.5 showing sawtooth pa ttern and (b) CuLaO2.66 showing kagome pattern. ........................................................... 43 1 10 Inverse susceptibility versus T curve for CuScO2.5. ............................................................ 44 2 1 XRD 2 scan for CuCr0.95Mg0.05O2 thin films showing the behavior as a function of deposition temperature. ..................................................................................................... 55 2 2 XRD 2 scan for CuCr0.95Mg0.05O2 thin films showing the behavior as a function of deposition pressure. ........................................................................................................... 56 2 3 Thermodynamic plot of the Cu, Cr, and O system. ............................................................. 57 2 4 c -Axis lattice parameter as a function of Mg content (x) in CuCr1xMgxO2. ...................... 57 2 5 Compilation of XRD 2 scans of CuCr1xMgxO2. ............................................................ 58 2 6 Transmittance data for spinel sampl e grown at 450 C and 5 Pa O2. ................................. 59 2 7 XRD -scans of CuCr0.95Mg0.05O2 grown on c -Al2O3 showing. ......................................... 60 2 8 XRD rocking curves for CuCr0.95Mg0.05O2 grown on c -Al2O3 for the (a) CuCrO2 (006) peak and (b) the CuCr2O4 (202) peak. ........................................................................ 61 2 9 AFM image of CuCr0.95Mg0.05O2 film grown at 650 C and 1 Pa O2 pressure. ................. 62 2 10 Resistivity plotted against growth temperature for CuCr0.95Mg0.05O2 films. ..................... 62 2 11 Resistivity plotted against Mg doping composition for CuCr1xMgxO2 films. ................... 63

PAGE 10

10 2 12 Optical transmission data of CuCr1xMgxO2 films. .............................................................. 64 3 1 XRD 2 data from CuCr1xMgxO2 of (006) pe ak for 0.00 x 0.05. ................................ 77 3 2 XRD 2 data from CuCr1xMgxO2 0.00 x 0.05 films. .................................................... 78 3 3 Resistivity versus temperature data for 400 nm CuCr0.99Mg0.01O2 film. ............................ 79 3 4 Van der Pauw 4 -point resistivity of Mg dependence of CuCr1 xMgxO2. ............................ 80 3 5 Seebeck data for thi n films of CuCr1xMgxO2. ..................................................................... 80 3 6 Direct band gaps calculated from Tauc plots of respective films. ...................................... 81 3 7 XPS multiplex for Cu 2p, Cr 2p, and O 1s showing Gaussian peak fits. ........................... 82 3 8 XPS plots showing binding energy changes with an increase in Mg content x for CuCr1xMgxO2 thin films ....................................................................................................... 83 4 1 XRD 2 data for CuSc0.95Mg0.05O2 films grown directly on c -plane Al2O3. ................ 100 4 2 Optical transmission data growth temperature dependence for films of CuSc0.95Mg0.05O2. ................................................................................................................. 101 4 3 Seebeck coefficients growth temperature dependence for CuSc0.95Mg0.05O2 films. ........ 102 4 4 XRD 2 data for grow th temperature dependence of CuSc1xMgxO2 films on c Al2O3. .................................................................................................................................... 103 4 5 XRD 2 scan for growth temperature dependence of CuSc0.95Mg0.05O2 grown on CuCrO2 template. ................................................................................................................. 104 4 6 XRD 2 intensity ratio comparison for growth temperature dependence of CuSc0.95Mg0.05O2 on CuCrO2 buffered substrates. ............................................................. 105 4 7 4 point Van der Pauw resistivities of ~400 nm thick films of CuSc0.95Mg0.05O2 on 50 nm CuCrO2 as a function of growth temperature. .............................................................. 106 4 8 XRD 2 data for CuSc1xMgxO2 films grown on CuCrO2 template. ............................. 107 4 9 Full -width half -max values of rocking curves of the (006) peaks of CuSc1xMgxO2 of buffered and unbuffered films. ....................................................................................... 108 4 10 -scan s of CuScO2 thin film grown with CuCrO2 buffer Al2O3. ...................................... 109 4 11 XRD 2 data of various graded films. ............................................................................. 110

PAGE 11

11 4 12 Optical transmiss ion data for (a) CuCrO2, (b) CuCr1/3Sc2/3O2, (c) CuCr1/3Sc2/3O2, and (d) CuScO2 films. ................................................................................................................. 111 4 13 XRD 2 data of CuSc0.95Mg0.05O2 grown on (a) CuCr2/3Sc1/3O2/CuCr1/3Sc2/3O2 multilayer, (b) CuC rO2 buffer, and (c) directly on sapphire. ............................................ 112 4 14 XRD 2 plot of data from sample A. .............................................................................. 113 4 15 -scan of sample A. .............................................................................................................. 114 4 16 XRD 2 plot of data from sample A. .............................................................................. 114 5 1 Proposed flat band diagram for AMO2. .............................................................................. 126 5 2 XRD 2 data shown for highly textured polycrystalline CuSc1xMgxO2. ..................... 127 5 3 rocking curve FWHM values for CuSc1 xMgxO2 grown directly on c -Al2O3. ............. 128 5 4 Tauc plots with corresponding transmission plots inset for CuSc1 xMgxO2. .................... 129 5 5 Tauc data (right axis) and PL (photo luminescent left ax is) values at 15 K and 300 K for CuSc1xMgxO2. ............................................................................................................ 130 5 6 Compilation of PL data for CuSc1xMgxO2 films at A) 15 K and B) 300 K. .................... 131 5 7 PL scans taken at various temperatures for CuSc0.99Mg0.01O2 400 nm film. .................... 132 5 8 Conduc tivity versus temperature curves for CuSc0.95Mg0.05O2 film. ................................ 133 5 9 Resistivity values for CuSc1xMgxO2 values for textured polycrystalline and buffered, epitaxial 400 nm thin films. ................................................................................................. 134 6 1 XPS plots of the Zn 2p3/2 peak and t he O 1s peak(s) before and after sputtering ........... 143 6 2 XPS plots of the Zn 2p3/2 peak and the O 1s peak(s)at 50 C and after 500 C ............... 144 6 3 XPS plots of O 1s peak(s) for A) O -terminated and B) Zn -terminated substrates. ......... 145 6 4 XPS O 1s peak(s) Gaussian fit of O terminated substrate at 350 C. ............................... 146 6 5 O 1s integrated peak areas for adsorbed and bulk oxygen species for both O terminated and Zn -terminated substrates. ........................................................................... 147 6 6 XPS plots of S 2p3/2 pea k(s) for A) O terminated and B) Zn terminated substrates. ...... 148 6 7 Comparison of S 2p3/2 peak maximum intensities for O terminated and Zn terminated surfaces. .............................................................................................................. 149 6 8 RHEED patterns for Zn -terminated and O terminated substrates. ................................... 150

PAGE 12

12 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfi llment of the Requirements for the Degree of Doctor of Philosophy EXAMINATION OF Z n O, C u CrO2, AND C u S c O2 PROPERTIES FOR USE IN TRANSPARENT ELECTRONICS AND CHEMICAL SENSORS By Patrick Sadik August 2009 Chair: David Norton Major: Materials Science and Eng ineering Greater progress in the field of TCO ( transparent oxide semiconductor ) research has been hampered by a lack of avai lability of p -type candidates. Though t he n -type TCOs are well documented and easil y produced only one well documented oxide has s purred continued efforts at producing robust, p -type behavior namely ZnO. Though p -ZnO has been produced, a coterminous research effort into improving the film qualities of lesser known naturally p -type TCOs could prove fruitful. Thus, this research has examined the thin film properties of two delafossites, CuCr1 xMgxO2 and CuSc1xMgxO2 as well as the surface properties of ZnO. The delafossites CuCr1xMgxO2 and CuSc1xMgxO2 have been shown to have the superlative properties of high conductivities (220 S cm1 and 70 S cm1) and high transparencies in 400 nm thick films of ~40% and ~80% throughout the visible spectrum We have also been the first to describe a near band edge photoluminescence in CuSc1 xMgxO2 for 0.00
PAGE 13

13 CuCrO2 at 700 C as a 100 nm nucleation layer followed by a 750 C growth of CuSc1xMgxO2 and a 400 C growth of ZnO secures an epitaxial growth for the entire structure despite large lattice mismatches between c -Al2O3/CuCrO2 (8.7%) and CuCrO2/CuScO2 (7.7%). This combined research takes the first steps in making robust pn-heterojunct ions possible. Also, in an effort to progress an understanding of the surface chemistry of ZnO we provide results showing marked difference in the chemisorptions of dodecane thiol upon Znand O terminated faces of ZnO.

PAGE 14

14 CHAPTER 1 LITERATURE REVIEW Interest in Delafossite s In recent years, there has emerged significant interest i n functional oxide materials .1 One of the more important functionalities for electronic applications is that of the wide bandgap semicondu ctors .24 Though there has been recent success in growing ZnO pn-homojunctions, the need still exists to examine alternative p type transparent conductive oxides (TCO s) to couple with n -ZnO to create successful top and bottom gated TFTs and transparent pn junctions. Perhaps the most promising group of candidates is that of the copper delafossites, CuMO2, where M is a trivalent metal. Traditional TCOs Binary p -Type Ox ides As a transparent conducting oxide (TCO), the delafossites offer intriguing properties, particularly as it pertains to the formation of a p type material, since the list of possible candidates is quite short. Of the binary conducting oxides only Cu2O a nd NiO exhibit a well established native p type behavior. Cu2O has been shown to have a bandgap of 2.17 eV, a maximum mobility of 100 cm2V1s1 and a carrier concentration of 1x1017 cm3. While its more narrow band gap may make an attractive option to couple with ZnO as a solar cell it excludes its usefulness in a blue/UV LED or even as a true TCO.5 In 1993 Sato et al grew NiO as a p type oxide.6 Though this film showed hole concent rations as high as 1.3x1019 cm3 and a resistivity of 1.4x101 ohm cm, its resistivity scaled with transparency. The determination that the bandgap was due to an electronic transition on individual Ni atoms limited it applicability. A recent article by Miyata et al. has attempted to revive interest in these films by growing Cu2O NiO thin films

PAGE 15

15 on glass plagued with either low carrier concentrations or high resistivities.7 Thus, it seems that a list of native, wide -band gap, p -type binary oxides is effectively non-existent. TCO Coordination Classes Octahedral class Traditional TCOs may typically be grouped according to their cation coordinations with oxygen as octahedral, tetrahedral, or linear. Octahedral coordinated T COs include CdO, In2O3, SnO2, CdIn2O4, and Cd2SnO4. Those which belong to the spinel structure group of formula AIIBIII 2O4 belong to the octahedral class despite having their AII cation in tetrahedral coordination because the A -A cation distance is typica lly too great to affect the electrical transport. This contributes to B B interactions dominating the electrical properties of these materials. Also, spinels typically suffer from poor solubility in acceptor doping and self compensating defects and therefo re have been difficult to dope p-type.8 One candidate that has emerged however is NiCo2O4. It shows p type beha vior, a transparency between 4060%, and a conductivity of 16 Scm1. The typically low acceptor solubilit y was perhaps overcome by t he relatively small ionic diameter difference between Co and Ni allowing Ni3+ to sit on the octahedral sites and act as an acceptor.913 Tetrahedral class ZnO is the sole member of tetrahedrally coordinated TCOs (though two nitrides, AlN and GaN, have an equivalent cr ystal structure). ZnO has sparked a great deal of interest. It has a large exciton binding energy of ~60 meV, a bandgap of 3.35 eV, and a room temperature Hall mobility of more than 200 cm2V1s1.1417 It may also be wet etched. ZnO growth characteristics can produce an abundance of Zn interstitials and O vacancies. These defects give ZnO its native n type semiconductive behavior and have foiled most attempts at making p-type ZnO. Two inherent issue sill plague ZnO. First, despite a more covalent bonding than other TCOs ZnO

PAGE 16

16 still suffers from a localization of acceptors on oxygen atoms known for their extremely high electronegativity. This effectively leads to low hole mobilities. Secondly, it suffers f rom a low valence band maximum (VBM) which contributes to the difficulty creating p type material. The number of acceptor s which can lie close to the valence band edge is limited Thus, much of the early work attempted to dope with the group V element nitr ogen ha s been met with mixed results. The first successful attempt at nitrogen doping of ZnO was achieved by growing ZnO under NH3 and H2 18 The latest attempts at p doping ZnO have not only continued work on incorporating nitrogen into ZnO but also either used group arsenic and phosphorous often alongside cationic codopants which increase anion solubilities in ZnO.19 21 Though in his large 2005 ZnO review Ozgur et al. declared high quality p type ZnO unconfirmed groups since hav e confirmed p ty pe ZnO and ZnO pn -homojunctions.2 22 In many ways these issues still plague ZnO device structures even though reliable pZnO has been created and devices tested. The material still suffers from killer defects and its inherent tendency to form donor states either through Zn interstitials or O vacancies. This is further complicated by the tendency of ZnO to form a tough OH layer on the surface which is conductive and can cause electr ical bypassing. Linear class The group of linearly coordinated TCOs largely consists of delafossites of formula AIMI I IO2 and the interesting species SrCu2O2. T hese take advantage of the Ag and Cu ability to form O -A -O bonds in linear coordinations. These TCOs share their linear cation oxygen coordination with Cu2O but have a more limited A -A coordination. Details of this material will be described below.

PAGE 17

17 Delafossites Delafossite Background An overview of the wider body of both binary and ternary oxides r eveals that most TCOs are de generate n -type semiconductors.23, 2 4 Few TCOs are p type .252 7 While the hist ory of delafossites in the literature goes back more than 130 years ptype conduction in delafossites was first reported in 1997.28 The first mention of delafossites in the literature was from a Siberian mineral s ample in 1873. The sample of CuFeO2 became known as a delafossite after the French crystallographer Gabriel Delafosse.29 30 It was later detected by Rogers in deposits in the United States in 1913 and 1922 but t he crystal structure was not deduced until work by Soller and Thompson in 1935 and work by Pabst in 1946 using a synthetic and mineral sample deduced the unusual structure. .3134 The Kawazoe et al. Nature article, published in 1997, entered delafossites into the semiconducting oxide community by expou nding the ptype properties of CuAlO2.28 Aside from its obvious uses in transparent electronic this wide -bandgap, p -type material has the potential to impact material performance in several areas. Highly conductive p -type TCO s are important for polycrys talline photovoltaics, where junction formation between an ntype absorber and the TCO is desired. Its worth noting that despite delafossites fulfilling a need for p type TCO s b oth n and p type semiconductors have been reported within the delafossite s ystem Also, t he ability to obtain both p and n -type delafossites opens the possibility of fabricating oxide semiconductor pn -heterojunctions. Yanagi et al reported the formation of both nand p -type CuInO2 by doping with Sn or Ca as donors an d acceptor s, respectively a nd even produced a pn junction.35, 3 6 Finally, though other uses undoubtedly exist, there has also been interest in utilizing delafossite semiconductors for hydrogen production via water splitting .37

PAGE 18

18 However, the ceramics field provides the second largest area of delafossite research by focusing on its use as a thermoelectric.3 840 The properties required to create an effici ent thermoelectric are straight -forward. They are a large Seebeck coefficient (to electrically stimulate a large T), a low thermal conductivity (to prevent hot and cold poles of the material from mixing), and a high electrical conductivity ( to minimize Joule heating ). However, m aterials that fit these requirements are difficult to formulate. Early research focus ed on metals but the inherent tendency of metal lic conductivity to increase in -step with thermal conductivity thwarted early efforts. The layered structure of delafo ssites may serve to fulfill these basic design requirement s by making heat transport more dif ficult through the material while it s ability to serve as a near degenerate semiconductor may further promote the high conductivities necessary to decrease Joule heating Specificall y, Seebeck coefficients in excess of 500 V/K (for CuCr0.95Mg0.05O2 and others ) and relatively high electrical conductivities (+100 Scm1) make further research promising. Another unusual candidate is CuRh1 xMgxO2 which actually becomes metallic with an increase in Mg concentration. Mg doping amounts as high as x = 0.20 have been achieved with CuRh0.90Mg0.10O2 showing a ZT as large as 0.15 (dimensionless) near 1000 K.41, 4 2 L imitations in utilizing delafossites as semiconductors inclu de relatively low cond uctivities, low mobilities, and the sometimes observed challenge in achieving phase -pure delafossite thin film .44 While phase -pure delafossite films have been reported in the literature efforts by other groups suggest that the processing con ditions are somewhat narrow .25, 3 7 Delafossites are a subset of the much larger group of AMO2 compounds.23 29, 30, 35, 3 6 4 756 These compounds ( Figure 1 1 ) exist within the four distinct coordinat ion classes : (1) AVIMVIO2 VI (tetragonal -LiFeO2 and rhombohedral NaFeO2), (2) AI VMI VO2 I V (orthorhombic NaFeO2), (3) AV IIIMI VO2 VI (orthorhombic KFeO2), and (4) AI IMVIO2 I V (delafossite). The

PAGE 19

19 coordination classes are determined relatively simply by the diameters of the A1 + and M3+ cations. Sma ll cations simply cannot sustain a large coordination due the oxygen crowding which would result while large cations require large r coordination numbers. Delafossites with the formula AIMIIIO2 and a 2:6:4 coordination scheme have a both a limited (for AI) and wide open (for MIII) constituency AI can be palladium, platinum, silver, or copper in a monovalent state and MIII is a trivalent metal with a diameter equal to or larger than that of boron and equal to or smaller than that of lanthanum essentially an y trivalent metal .57 A list of extant delafossites is displayed in Table 1 1 The AI atom has the very unusual II -coordination which is typically seen only for Cu and Ag c ompounds but in this structure also supports Pd and Pt in the very unusual 1+ oxidation state. This is the only compound which secures Pd or Pt in this 1+ valence and will be addressed more in depth in a later section. Delafossite Polytype s The structure i s one composed of an alternating stack of O -Cu O dumbbells and [MO6] edge sharing octahedral and is divided between two polytypes the 3R rhombohedral crystal structure and the similar hexagonal structure 2 H The point group symmetry for the two polytypes is R3m and P63/mmc respectively. The basic construction of [MO6] polyhedral and O A O dumbbells are the equivalent but the 3R delafossite forms in a ABCABC stacking sequence whereas the 2 H delafossite has its polyhedral layers rotated 180 to each other in a ABABAB sequence. While the 3R is more common, the 2H is seen most often for the IB delafossites, CuScO2, CuYO2, and CuLaO2.45 Both polytypes may be seen in Figure 1 2 p -TCO Design Requirements In his 1997 Nature article Kawazoe et al. claimed to have sought out to design a p -type TCO ba sed on three principles.28, 58 First, the cationic species must have a filled d10s0 character to remain color transparent. An unfilled d10 shell can create optical absorptions states for light

PAGE 20

20 within the visible spectrum. Currently researched n type compounds such as ZnO, TiO2, and In2O3 which contain respectively Zn2 +, Ti4 +, and In3+ fulfill this well understood requirement. Secondly, the cationic d10s0 shell needs to have an energy that overlaps the oxygen 2p shell. This decreases the localization of oxygens 2p electron (often referred to as lone pairs) thereby dispersing the valence band for better hole conduction. Two cations match the above two requirements Cu1+ (3d10) and Ag1+ (4d10). Thirdly, crystal geometry was predicted to play an important role in promoting covalent metal -oxygen bonds by providing a catio n tetrah edron around each oxygen atom. The delafossit e structure has oxygen in quasi tetrahedral arrangement with each surrounded by three MIII cations and one AI cation. This environment has been calculated to provide a lessening of the O 2p3 localization thereby maximizing hole mobility. Seemingly, Cu2O meets all the requirements but fails the optical visibility test with an inadequately small band gap. Therefore, crystal environments were examined that essentially contained Cu2O but that would isolate la yers of Cu2O into 2D quantum wells which increase the band gap. The net result was a thin film delafossite material, CuAlO2, which exhibited native ptype conductivity, an optical band gap of 3.5 eV, a carrier density of 2.3x 1017 cm3, a conductivity of 0.095 S cm1, and a Hall mobility of 10.4 cm2V1s1.5 The crystalline environment of each atom in the delafossite crystal plays a huge part in determining the special properties of the delafossite crystal system. The O -A O layer maintains its two -fold coordination seen in the binary oxide Cu2O. As mentioned, Cu2O h as a bandgap of 2.17 eV due to the three -dimensional proximity of the Cu atoms to one another in the lattice (12 nearest Cu neighbors) .59 This is easily understandable in terms of the low oxidation state of the Cu1+ in Cu2O and the resulting low density of oxygen in the structure. The delafossite maintains this copper oxidation state while placing the Cu in a hexagonal t wo dimensional grid of 6 Cu

PAGE 21

21 atoms .60 This two dimensional Cu lattice promotes less interaction between A atoms thereby maximizing the effective bandgap. It also opens the possibility of tailoring properties throughout the lattice by increasing or decreasing Cu Cu interactions. AI Cations: Effects on the Delafossite Electrical Structure Shannon and Prewitt also determined that those delafossite containing Cu or Ag behave as semiconductors while those containing Pd or Pt behave as metals. The Figure 1 3 band diagram of these materials explains this phenomenon adequately as the dz 2-s valence band is half -filled due to the Pd+1 and Pt+1 oxidation state having a d9 configuration as opposed to the filled as opposed to a filled d10 valency for Cu+1 and Ag+1. Pt and P d roles in delafossite p roperties A great deal of theoretical interest has been generated by the delafossites PdCoO2, PdCrO2, PdRhO2, and PtCoO2 both because of the unique oxidation state for the noble metals and their extremely high conductivities. The unusual oxidation state is accompanied by tendency of these cations to form square planar geometries as in PdO and PtS.61, 6 2 The transport properties are also highly anisotropic. PdCoO2 has a xy resistivity of 6.0 x 106 ohm -cm and a z resistivity of 2.1 x 103 ohm -cm. PtCoO2 has a xy resistivity of 3.0 x 106 ohm -cm and a z resistivity of 1.1 x 103 ohm -cm.63, 64 These xy values surpass even the native metals resistivity of 1.05 x 105 ohm -cm for both palladium and platinum. There is even interest in PdCoO2 as a possible superconductor because of the hexagonal arrangement of its Pd atoms and absence of 3c2e bonding orbitals which would otherwise prevent the metal -metal interactions from dispersing over the entire plane.65 Further work to promote superconductivity may attempt to dope the lattice .3 e per formula unit in order to shift the Fermi energy to a saddle point. It should also be noted that a very limited number of allowed MIII metals which react with Pd or Pt to form the delafossite structure, Rh, Co, and Cr for Pb and only Co for Pt. These MIII

PAGE 22

22 anions have in common low spin states with empty eg orbitals. Further calculations shows that these low spin cations do not contribute appreciably to the conductivi ty and act instead as metal oxide insulating layers.66 The use of silver in the delafoss ite l attice Though they form semiconductors, Cu and Ag delafossites show several nuanced property differences First, the valen ce band intermixing between A d10 and O 2p6 states are different between Ag and Cu lar gely due to the higher energy of the Ag 4d10 states. Kandpal summed up their electronic state calculations claiming the following.67 First, Cu delafossites have the most narrow valence bands while Ag (and their calculated Au ) delafossites have bands that were broader than expected. Secondly, d10-d10 interactions are significant in Ag and Au delafossite at distances found in S c delafossites or smaller. Cu delafossite showed almos t no influence of M diameters on conductivity since its Cu d10-d10 interactions are limited due to the relatively large distances experienced by Cu 3d10 orbitals as compared to their separation s in Cu m etal. Thirdly, the character of the mixed A d10 O 2p6 hybrid valence band is dominated by the A d10 character for the Cu delafossite and the O 2p6 character for the Ag and hypothetical A u delafossites. The implication that hole doping within a band consist ing of greater d10 character will have an enhanced hole mobility favors Cu delafossites. However AgCoO2, which has been shown to have one of the highest Ag -based delafossite conductivities (2.0x101 S cm1), has A -A distances equal to the native metal.68 This along w ith the non-distorted octahedra found in the Co delafossites may play a role in their ubiquitous presence in all four types of delafossite (A = Cu, Ag, Pd, and Pt). It should also be noted that it is the sole 3d transition metal to have its 3+ cation with both empty eg states and completely filled t2g states as demonstrated by the basic field splitting model in Figure 1 4 A greater mixing of the Co 3d10 orbitals with t he O 2p6 orbitals may help substitute for the lack of Ag 4d10 character in

PAGE 23

23 the hybridized valence band. The effect of MIII d10 orbital contributions has been otherwise neglected within this discussion but will be shown to play a larger role in the heightened conductivity of this and other delafossites It will be explained more in depth in a proceeding section. The use of copper in the delafossite l attice T he overall property gaines of Cu delafossites over their Ag analogues is due predominately to aforemen tioned character of the valence band. Both the enhanced hybridization of the Cu 3d10 and O 2p6 states over the Ag 4d10 and O 2p6 states and the greater d -band nature of the former encourage higher hole mobilites. These act to dispel oxygens usual drive toward a localization of charge states and promotes dispersion of the hole states which form and is evidenced by the typically lower conductivities of the Ag base delafossites.69 The resulting band structure pro mote s relatively light hole masse s as compared to conventional TCOs.70 The original valence band diagram Figure 1 3 proposed by Shannon et al. shows that the valence band is dominated by the aforementioned Cu 3d10 and O 2p6 bands as a z oriented antibonding orbital denoted as dz 2-s. This promotes conduction between the bond s within the two dimensional xyplane of the copper sublattice, accounting for the anisotropy of conduct ion of single crystal delafossites .63 64, 71 The conduction band is dominated by a bonding state denoted dz 2+s. Howeve r, mor e recent calculations predict that p type conductivities for at least some delafossites may be greatest along the c axis of the materials, namely for CuScO2, CuYO2, and CuAlO2.70 Structural Considerations It be ars repeating that the delafossite structure is characterized and noted fo r its layered geometry. Both the wide band gap and p type behavior can both be explained by the layered structure of the delafossite. Firstly, the wide band gap is largely determined by the structure

PAGE 24

24 which is comp osed of alternating strata of [ M3+O6] edge sharing polyhedra and Cu+1 ions. This has been modeled electronically as alternating wide band gap trivalent metal oxide and narrow band gap copper (I) oxide layers which effectively increases the band gap due to t he resulting 2D quantum wells. This quantum confinement effect is often noted for the classic CdS and PbS nanoparticle examples.72 Second ly the layered structure for delafossites with a lattice constants larger than or equal to CuScO2 allows extra -stoichiometric oxygen to diffuse into the Cu plain to further oxidize some of the copper from a valence of Cu1+ to Cu2+. c -Axis r elationship to MIII c ation radius It should be noted that the de lafossite c lattice constant is dependent predominately upon the AI diameter and the MI II diameter a ffects the a -lattice constant. The minimal dependence of the c lattice constant on the MIII diameter will be commented upon in a later section. This minimal dependence should not obviously be the case for a perfect octahedra [MO6] as seen in Figure 1 5 and described in the Equation 1 1 h' = 6/6 a (1 1) H ere h is the effective z height of a perfect poly hedra and a is the a lattice constant. Equation 1 1 may be further analyzed to produce the relationship shown in Equation (1 -2 a ) and Equation (1 2b) a = 2 dM O (1 2a) a() = 1.414 rM3+ + 1.92 (1 2 b ) The variable dM O is t he effective MIII to O bond length (or equivalently the addition of the the M3+ and O2 radii). Essentially, if the [MO6] was a perfect octahedral there should be a substantial correlation between the c axis lattice parameter and the MIII radius. However the c lattice constant is almost independent of the MIII cation while the a lattice closely follows the

PAGE 25

25 values predicted by the Equation 1 2b (Figure 1 6 ). Basically, while the xy plane adjust s predictably according to MIII selection, the actual z lattice parameter far short of the predicted geometry with the exception of CuAlO2 and CuCoO2, which aside from the recently created CuBO2, are the delafossites consisting of the smallest MIII anions. This shows that the while horizontal spacing of the oxygen sublattice is dominated by the easily calculable geometry of the [MO6] octahedron the vertical assembly of oxygen atoms is mu ch more compressed than a perfect octahedron [MO6] would suggest. Simply, t he oxygen locations are compressed in the z direction but not in the xy plane. However, a further examination of Figure 1 6 d emonstrates that while a perfect octahedron model fits those cations with radii smaller than Yb3+ for the a lattice parameter, thos e with larger cations exceed their calculated a values by roughly 5 6%. This is true independent of whether Ag or Cu is the AI constituent. These MIII cations share, with the exception of Y, filled inner cores of 4f orbitals which are often noted by scient ists for the lanthanide contraction. This effect is seen as a general contraction of the atomic radius with an increase in atomic number from lanthanum through l utetium. Whether the relatively large size of these cations or the prevalence of 4f interaction s plays a role in this dilation has not been discusse d elsewhere in the literature. Negative thermal expansion and links to bonding c haracter Further analysis of the delafossite structure yields the interesting property of NTE (negative thermal expansion). J. Li et al. examines delafossites and their linear O -AI-O coordinations as materials demonstrating an NTE coefficient.73 The analysis resulting fr om their experiments helps to paint a more accurate picture of the underpi nnings of the unique properties of th e delafossite structure. First, the O -AI bond shortened significantly with an increase in MIII size. The effect was seen for both Cu and Ag delafossites. Since the caxis length is shaped by three quantities the A O di stance, the M -O distance, and the O -M O angle the initial

PAGE 26

26 observation that c axis lattice parameter is invariant with changing MIII diameters l eads to the question whether ei ther the M O distance or O -M O angle (or both) also change to compensate for MIII variations Nature of A A i nteractions Jansen and Li et al. described the above observation as an effective competition of the A O and A -A bond. It should be noted that the A environment is unusual in that while it has a formal coordination of 2 it actu ally has 8 nearest neighbors two O atoms and six Cu atoms, essentially forming a dodecahedron. Since A atoms in the delafossite structure have filled d shells of the quantum configuration nd10(n+1)s0, there is no bond expected between neighboring A atoms.7 3, 7 4 However, as electrical transport measurements have demonstrated for single crystals of delafossites, conduction seems to take place largely in the xy plane with the (MO2)1 layers acting as insulators. Thus Jansen calculated and proposed that the linear O -A O bonds effectively hybridized the s and d orbitals in which charge density could pass from d states to the unoccupied s orbitals which could thereby promote d -d interactions. J. Li et al. found that while possible this weak bond alone could not explain the relatively large decrease in A -O bond length with an increase in MIII diameter. They proposed that the s -dz 2 hybridization polarizes the filled d -shell. As a shrinking MIII diameter for ces As more closely together they in turn are forced to become less polar in turn shifting electron density into the A O bond. This effectively makes the A O bond weaker and longer.73, 7 4 Basical ly, a weaker A O bond encourages stronger A -A interactions. This was initially summarized in Shannon and Prewitts original work in Figure 1 3 flat band diagram and modified slightly since that time M ore explanati on of the relevance of this diagram will be explained in a later section.

PAGE 27

27 Minimal a ppearance of c -axis dependence u pon MIII r adii Secondly, the rate that the c lattice parameter increases slightly with an increasing MIII radius is different for CuMO2 and AgMO2 species. While both seem to have a fairly linearly increasing c lattice parameter with an increasing MIII diameter the AgMO2 compounds show a steeper slope and thus a greater or more rapid effect of the MIII diameter upon the c axis lattice param eter. This relationship is converse for the a lattice parameter where the increasing MIII diameter has a more marked effect upon the lattice dimensions of the xy -plane. The relative strength of the AgO and Cu O bonds also plays a role in the phenomenon be cause of the relationship that a weak A O bond implies a stronger M O bond. This is seen for the marked thickening of the (MO2)1 layers in the CuInO2 species as temperature is raised and the relative stability of the (MO2)1layer in AgInO2 and its shorte r In O bond distance. The weaker bond is far more influenced by temperature than a strong bond. As an example, the In O bond distance is shorter for AgInO2 than CuInO2 for all temperatures. Influence of A upon AMO2 a -axis lattice p arameter Finally, while it might be expected that the a lattice parameter would depend solely upon the MIII cation there exist small disparities between Cu and Ag based delafossites. A closer examination of Figure 1 7 shows that there is crossover for Cu and Ag based delafossites effect upon the xy lattice dimensions. This crossover occurs between Fe3+ and Rh3+ and thus at A -A separations between 3.0351 and 3.074. 67, 7 3 It should be noted that A -A separation in the Ag and Cu native metals is 2.89 and 2.56 respectively. Li et al explained this trend as resulting from enigmatic d10-d10 interaction where the larger Ag atoms result in a greater repulsive force at smaller distances but become slightly attractive at larger distances. 67 7 3 Simply put, Cu Cu interactions are smaller than their Ag -Ag counterparts because of their smaller size.

PAGE 28

28 Explanation of d ev iation of the c axis l attice p arameter of AInO2 Viewing the AMO2 delafossite as a multi layered Cu1+/(MO2)1 stack may help to explain one deviation (that of AInO2) from fairly flat MIII independence of the c axis lattice parameter. Essentially, as the rea l charge on the MIII cation increases the repulsion between edge sharing [MO6] octahedral increases causing a thinning in the z direction of this (MO2)1layer. As electron egativity is tracked it is found to have the highest electronegativity for the avail able MIII delafossite constituents. This may help explain why it has an unexpectedly large c axis lattice parameter as seen in Figure 1 6 Plots of calculated MIII radii against reported a and c lattice parameters .75 The Role of Defects in p -Type Behavior A comprehensive explanation for ptype behavior in what appear to be stoichiometric, undoped delafossites has not been sufficiently explained. The fact that most native delafo ssites exhibit p type conductivity is perhaps best explained by the doping limit rule. The VBM (valence band maximum) is apparently close to the vacuum level for the majority of these materials.76 This was calculated to be true in the cases of CuAlO2, CuScO2, and CuYO2 by Shi et al .70 More extensive study involving experiments and calculations were performed by Ingram et al Essentially they varied the Al/Cu ratio in CuAlO2 powde red samples by synthesizing the delafossite through high and low temperature techniques. This in turn produced Al -poor and Al rich material respectively. Though the ir analysis was far more extensive, the different growth techniques had the most obvious eff ect of providing higher -conductivity material with the low temperature, hydrothermal, Al -rich process. This affirmed their further measurements and calculations that the associated defect (AlCu 2Oi ) was responsible for the heightened p-typ e conductivity of their samples. This is a significant finding since intercalation is largely prevented

PAGE 29

29 in cations with an MIII diameter smaller than Sc3+ and opens up a small window for further improving the p type conductivity in CuAlO2 by tailoring grow th conditions to effect defect concentration8, 50, 77 79 Because of the doping and intercalation limitations imposed by several of the smaller a lattice constant delafossites, best represented in the literature by CuAlO2, interest has increasingly concentrated upon those delafossites which demonstrate high doping solubilities. To prepare delafossites for use in transparent electronics research has focused largely upon either acceptor doping or growing better quality, phase pure material. The first category may be grossly subdivided into t wo major camps: oxygen intercalation and replacing a MIII ion with a divalent cation to create a hole. The first two have attracted much of the attention and have led to large net improvements in carrier concentration and conductivity in delafossite films. The attempt to grow better epitaxial films has also been of keen interest but will be addressed more extensively in Chapter 2 and Chapter 4 Oxygen Intercalation Traditionally it has been noted that oxygen intercalation has played a major role in increasing the p type conductivity of many delafossites such as CuScO2, CuYO2, and CuReO2 (where Re is a rare earth metal).68, 8 0 81 The ability of the delafossite lattice to allow intercalated oxygen is most dependent upon the size of the MIII and is typically only seen for delafossites with MIII larger than or equal to the radius of Sc3+ (0.885 ). The layered nature of delafossites allows fairly s imple diffusion of oxygen through the lattice. The openness of the O -Cu O layer allows this ready diffusion and is well explained by Mugnier et al in Figure 1 8 and Equation 1 3 .82 First, it is noted that the intercalated oxygen lie at the center of corner shared Cu triangles.82 This basic geometric constraint would seem to limit oxygen intercalation to a value

PAGE 30

30 of = 0.5. However, the ver y large lattice constants seen in the delafossite system for MIII = Y and the rare earth metals allows these delafossite to achieve a concentration of =0.67. The basic geometr y seen is shown in Figure 1 8 where a is the a lattice constant, the smaller spheres are Cu1+ cations, and the large circle described by a triangle is the available space for an oxygen anion. This is a very simplistic analysis since the formal valence of the Cu cations will obviously change due the oxidation which will lead to smaller values of the Cu cation radius thereby opening the lattice more while the oxygen radius depends on its coordination. The result actually does well predict observed intercalation trends. Using this simple geomet ry the diameter of the sphere of space in which an oxygen atom may fit can best be described by the simple Equation 1 3. ROx = 3/3 a r(CuII +) = 3/3 a 0.46 (1 3) This results in a no -strain minimum a -lattice parameter of 3.15 for an assumed ROx = 1 .36 Using E quation 1 3 produces an oxygen intercalation boundary of r(MIII) = 0.867 by which MIII radii that are larger are amenable to oxygen intercalation and those which are smaller are much less capab le. Reports of oxygen intercalation are well documented and typically reflect these geometrical constraints. The literature provides several examples of intercalation studies. CuYO2 was an easy candidate for a post growth oxygen treatment because of its very open lattice.83 Notably, it converted from a 2H delafossite structure into an orthorhombic configuration at x = 0.5 for CuYO2+x. For x = 1/3 the oxygen atoms were seen to lie on triangular positions on the copper sublattice. Other studies of intercalation repor t a marked increase in conductivity after annealing for CuYO2, CuScO2, and CuReO2 for Re = Y, La, Pr, Nd, Sm, and Eu.68 8 0, 8 1 Though Ingram et al agreed that the MIII size is th e key factor in determining whether a particular

PAGE 31

31 delafossite is either resistant or amenable to oxygen intercalation, some studies suggest that delafossites CuAlO2 and CuFeO2 may be intercalated as well despite having the 2nd and 7th smallest delafossites MIII ionic diameters respectively. 75 84, 85 These are both much smaller than the 0.867 boundary previously mentioned. The most obvious expression t o explain the defect chemistry on the conductivity is shown in Equation 1 4 .86, 87 O2(g) = 2OO x + VCu + VAl + 4H (1 4) The acceptor doping found with oxygen i ntercalation may be expre ssed in Equation 1 5. O2(g) + 2 CuCu x = Oi + 2 CuCu (1 5) It has been found that for 0.5 in AMO2+ Cu triangles with a n O center are typically found which order into corner sharing tiling.88 Similar systems have been described as a sawtooth pattern At = 0.5 these chains saturate to form a regular periodic triangular tiling in the xy plain as seen in Figure 1 9 (a). The three principal effects of intercalation are pha se change, a rise of conductivity through acceptor doping, and an enhancement of magnetic properties. The instance of phase change in the yttrium system has been discussed. The lanthanide series (with lanthanum the most studied species) provides a similar ability as yttrium to be intercalated to extremely high levels likely because of its huge lanthanum a lattice parameter of 3.83 This large lattice parameter allows the crystal to intercalate as much as 33% more oxygen than its basic formula to form a st able LaCuO2.66 delafossite phase.88, 89 This is seen to promote in CuLaO2.66 a Cu Cu superexchange mediated by the O atoms (to) yield and effec tive Cu network kogom lattice, with the extra oxyge n content of = 0.66 the fo rmal Cu valence becomes 2.33 composed of 1/3 nonmagnetic Cu 3+ and 2/3 Cu 2+ with s = 1/2.90 These first Cu 3+ cations act as hole dopants and may further

PAGE 32

32 hinder the magnetic properties of the delafoss ite system. These Cu 2+ states are theorized to form 1 D chains to which is displayed as a kogam patter n seen in Figure 1 9 (b) At saturations of = 0.5 CuYO2+ and CuLaO2+ show a formal copper valency of Cu2+ which all show a s = 1/2 spin state. The t hree -fold symmetry of the lattice naturally frustrates alignments of spin. These effects have largely been studied in La and Y delafossites as natural carry -overs from research in layered BayLa2yCuO4 materials wh ich were the first high Tc materials reported.9194 However, more recent studies have shown intercalation in other delafossites. In 2004 Garlea e t al showed interesting and analogous magnetic behavior in CuInO2.5 and CuScO2.5 bulk samples. In both cases structural refinement based on X ray and neutron diffraction data showed the presence of corner -sharing, oxygen containing triangles of Cu.95 This resulted in the magnetic behavior described in Figure 1 10. The behavior is characterized by an initial quick increase to ~20 K, followed by a linear increase to ~120 K, furth er followed by a non-linear increase resulting in a maximum of 6000 mol uem1 at 225 K, and a local minimum at 275 K. E vident ly a strange phenomenon is seen in these materials. Explanation of Spinel Formation Unexpectedly, the aforementioned species CuAlO2 and CuFeO2 have been reportedly intercalated as well despite Al3+ and Fe3+ radii (0.675 and 0.785 respectively) falling far below Sc3+ (0.885 ). In the case of CuAlO2+ the research lacked a comparison of before and after XRD data and seemed to be plotted against a linear intensity scale. It is therefore difficult to determine whether a non -percolated CuAl2O4 phase could be playing a role in the physical properties of t he system. Previous calculations had placed the maximum amount of intercalated

PAGE 33

33 oxygen at = 1/50,000 due to size and thermodynamic constraints.28 The other source cited in the Banerjee et al text is a report by Wang et al on CVD grown polycrystalline/XRD amorphous thin films which showed neither XRD data n or other data to show that the spinel reaction had not taken place. 2 CuAlO2+ + (1 2 )/2 O2(g) CuAl2O4 + CuO (1 6) Instead, it was assumed that the increase in conductivity with longer annealing durations was due solely to intercalation of oxygen into the delafossite lattice.96 The simple observation that delafossites with MIII smaller than Sc 3+ do not intercalate appreci able oxygen is not due merely to the size available but also to the thermodynamic driving force which produces spinels at low temperat ures and high oxygen pressures. There are few discussions in the literature but the spinel phenomenon has been addressed b y Mugnier et al .75 They noted that for CuFeO2+ powder samples an annealing temperature above 480 C in air produced a mixture of spinel and delafossite phases while lower temperature were able to achieve truly intercalated delafossite. They also performed a literature review and found that spinel phases are only stable for MIII radii are smaller than 0.665 In this case the rule may be expanded as the following schematic (Equation 1 7). r(MIII) (A) 0.665 r(MIII) (B) 0.885 r(MIII) (C) (1 7) The three possible cases are case (A) which is a spinel phase, case (B) which is a delafossite phase with limited intercalation and strain and case (C) which is either an intercalated delafossite AMO2+ or different phase such as those seen in Figure 1 1 The spinel species may present huge problems in determining the properties of delafossite materials. Our own work has shown that (Cu,Mg)Cr2O4 likely acts as a p type metal thereby lowered optical transmittance values and frustrating device constructs The lack of notation in the

PAGE 34

34 literature to this problem is rather surprising Its an effect that is exacerbated by doping attempts with M2+ such as magnesium which essentially have only the divalent oxidation state thereby providing a larger driving force toward enhanced oxidation into spinel phases Extrinsic Doping of the Delafossite Lattice The M cation site can also be doped to influence magnetic properties or acceptor concentrations. CuAl1 xMnxO2 has been shown to have bulk paramagnetic behavior.97 CuCr1 xCaxO2 and CuCr1xMgxO2 have been shown to have ptype behavior with the latter having the highest reported conductivity f or any delafossite = 220 S/cm for a doping (alloying) level of 5%.98, 8 0 Of note are the small to immeasurable Hall mobilities for the entire set of delafossites. Also, Marquardt et al notes that the low conduct ivities, with the obvious exceptions of CuCrO2, CuScO2, and AgInO2, are due to the inability of many of the delafossite systems to accept large doping concentrations.29 A few systems are exceptional in their n t ype character: CuFe1x SnxO2, AgIn1x SnxO2, and Cu In1x SnxO2. The CuInO2 compound is even more unique since it has been shown to be able to be doped both n-type and p-type. Ohta et al. reported that a pn -homojunction was created with CuInO2 by replacing In3+ with Sn4 + or Ca2 + as dopants.99 Notably, AgIn1x SnxO2 has not been doped p-type despite being a promising n-type delafossite. Table 1 2 provides a table of acceptor d oped delafossites and their respective resistivities. The three highest conductivities are for CuCr0.95Mg0.05O2, CuSc0.95Mg0.05O2, and AgIn0.95Sn0.05O2 where AgIn0.95Sn0.05O2 is an n type delafossite. This later finding of high conductivity is surprising c onsidering the more typical ptype nature of the vast majority of delafossites. Also, of note is the equally surprising CuRh0.9Mg0.1O2 delafossite. As a 4d analogue to the 3d candidate CuCoO2 one might expect this material to have similar properties. However, Kuriyama et al notes the CuRh1 xMgxO2 system undergoes a semiconductor to metallic

PAGE 35

35 tra nsition somewhere near x = 0.10, a dopant saturation and transition not reported for any other delafossite. It could be a more ubiquitous phenomenon but most delafos site studies have far less than an x = 0.10 doping cap for the AM1xDxO2. This property could also be due to some electronic stabilization due to the larger 4d6 orbitals configuration.100 The Krger -Vink expression f or this substitutional dopant is DM where D is the dopant and M the MIII lattice position for AMO2. This represents a geometrical ly interesting doping possibility. The idea of introducing DM d opants into the [MO6] layer while current is largely conducte d in the Cu Cu layer should allow dopants to improve the carrier concentration while enabling the mobility to remain almost unchanged. This is true in the larger class of ABCh2 new material known as an oxychalcogenides with formula LnCuChO (Ln2Ch2 2 + and Cu2O2 2-) where Ln is a 3+ cation and Ch is S, Se, or Te. This material was first described by Marcel Palazzi in 1981 and was latter found to have the properties induced by its superlattice These properties include a wide, tunable bandgap, optical transpar ency, native p type behavior, a large carrier concentration, and a low resistivity.97, 1 01104 Conduction in these materials takes place in the Cu Ch sheets while dopants sit on the Ln site. They have even spun-off research in a new area of superconducting materials known as oxypnictides with a similar layered structure and the analogous formula LnFePnO This compound description is more aptly described as a layered stack of La2O2 2and Fe2Pn2 2+ where Pn = P, As, Sb, or Bi.105 107 However, as Table 1 2 demonstrates the mobilities of most delafossites are extremely low which may have more to do with the nature of the conductivity in the material than lattice scattering. Ingram et al observes small polaron conduction in CuAlO2 thereby limiting theoretical mobility to <1 cm2/Vs.47, 7 7 This is essentially a thermally activated site to site

PAGE 36

36 jumping of charge i n the lattice and may prove to b e an insurmountable hurdle towards creating robust device structures. The literature skirts this deficiency by seldom providing Hall plots To date, despite several claims of successful Hall measurements and studies there has been no reliable Hall plot showing p type conductivity a side from a study of the metallic CuRh1xMgxO2.108 This r eport was interesting in that it proposed an xy -conduction anisotropy for the system but instead of the usual A -A conduction path the RhO2 layer was cited as being the primary conduction path. Also, difficulties in doping are limited by the relatively small subset of choices for cationic doping and the correspondingly small size of many MIII cations. Such acceptor dopant choices include Mg2+ (0.86 ), Ca2+ (1. 14 ), Ni2+ (0.83 ), Zn2+ (0.88 ), and Cd2+ (1.09 ). All of these are effectively too large for all delafossite with the diameter of Rh2+ (0.805 ) or smaller and Ca and Cd would likely cause a great deal of strain on any lattice smaller than La (1.17 ). Notably, CuAlO2 and CuCoO2 have never been able to be extrinsically doped, due likely to their small MIII sizes, though Ca has been found to be soluble in the other group V compounds (M = Ga, In), CuYO2, CuLaO2, and most surprisingly CuCrO2.10911 1 No reports have been found which attempt to dope the A lattice position with a n atom other than Cu or Ag and it remains unclear whether this lattice position is being doped th ereby further hurting p-conductivity within the lattice by either negating acceptor doping with the introduction of donors into the A -A layer or impurities scattering due to deep charge traps in the A -A layer. Shibasaki et al did show that for the Cu1yAgyRh1 xMgxO2 system mobility was minimally affected by increases in Ag content.108 Interest in CuCr1xMgxO2 system and CuSc1xMgxO2+ Systems Though these materials will be the focus of this research and discussed far more extensively in further chapters, a brief synopsis of their important properties is described here.

PAGE 37

37 For their potentials uses for optoelectronic and TCO applications, the delafossite CuCr1xMgxO2 and CuSc1xMgxO2+ stand out for different reasons The CuCrO2 system currently holds the highest delafossite conductivity of 220 Scm1 likely due to the added influence of Cr 3d orbital hybridization in its electrical transp ort and has visible light transparencies of around 40% for 400 nm thin films It can also be readily doped with Mg and has been found to grow epitaxially upon c -Al2O3 due to an 8.7% lattice mismatch where a(CuCrO2) = 2.99 and a (Al2O3) = 2.75 This a( Al2O3) is affected by a 30 crystal rotation upon the sapphire substrate. It also holds an economic advantage in utilizing relatively cheap starting materials CuScO2 though composed the less economical Sc metal has a comparable conductivity of 70 Scm1 an d has the added flexibility of being both doped with Mg and intercalated with O2 though effective doping has proved more difficult compared with the CuCrO2 system. Sc is also in the middle of the continuum, with its a lattice parameter large enough to int ercalate oxygen and small enough to allow some Cu Cu interaction, promoting electrical transport. Also, It has the enviable position of having an extremely small lattice mismatch with ZnO (~0.3% for a(CuScO2) = 3.24 and a(ZnO) = 3.25 ) and will not form a problematic spinel phase (with proposed formula (Cu/Mg ) Sc2O4) that is seen for CuCrO2.

PAGE 38

38 Figure 1 1 Phase space of ternary oxides in the AMO2 system R eprinted with permission from Figure 2 of R. D. Shannon, C. T. Prewitt, and D. B. Rogers, Inorg. Chem. 10, 722 (1971).

PAGE 39

39 F i gure 1 -2 Crystal structure of delafossite 3R and 2H polytypes Figure 1 -3 Flat band diagram of delafossite AMO2. M odified and reprinted with permission from Figure 4 of R. D. Shannon, D. B. Rogers, C. T. Prewitt, and J. L. Gi llson, Inorg. Chem. 10, 725 (1971).

PAGE 40

40 Figure 1 4 Basic construct of octahedral crystal field splitting Figure 1 5 Single [MO6] distorted octahedra with the corner sphere representing oxygen and the M at om is internal to the octahedra.

PAGE 41

41 Figure 1 6. Plots of calculated MIII radii against reported a and c lattice parameters Reprinted with permission of Elsevier and D. P. Cann from Figure 4 of M. A. Marquardt, N. A. Ashmore, D. P. Cann, Thin Solid Films 496, 151 (2006).

PAGE 42

42 Figure 1 7 Relationship of a lattice parameter to MIII radius with inset of aAg(MIII)-aCu(MIII). Figure 1 8 Triangular lattice of CuI cations with intercalated oxygen anion a R Cu +R Ox

PAGE 43

43 Figure 1-9. Cu-O plain for the oxygen intercalated delafossites A) CuYO2.5 showing sawtooth pattern and B) CuLaO2.66 showing kagome pattern with 1/3 holes showing 1-D magnetic lines drawn as dark triangles throughout the st ructure. Reprinted with permission of Elsevier from Figure 1 and 2 of A. Olariu, D. Bono, F. Bert, P. Mendels, C. Darie, P. Bordet, A. D. Hillier, C. Baines, and A. Amato Solid State Sci. 5, 1101 (2003).

PAGE 44

44 Figure 1 10 Inverse susceptibility versus T curve for CuScO2.5. R eprinted with permission from Figure 3 of O. Garlea, P. Bordet, C. Darie, O. Isnard, and R. Ballou, J of Phys Cond Matter 16, S816 (2004)

PAGE 45

45 Table 1 1 Table of known delafossite lattice parameters Compound A I CR () M III CR () Type a () c () AgAlO 2 Ag 0.81 Al 0.675 3R 2.8729 18.336 AgCoO 2 Ag 0.81 Co 0.685 3R 2.89 18.27 AgCrO 2 Ag 0.81 Cr 0.755 3R 2.9843 18 .511 AgFeO 2 Ag 0.81 Fe 0.785 3R 3.0391 18.59 AgGaO 2 Ag 0.81 Ga 0.760 3R 2.9889 18.534 AgInO 2 Ag 0.81 In 0.940 3R 3.2772 18.881 AgNiO 2 Ag 0.81 Ni 0.700 3R 2.936 18.35 AgRhO 2 Ag 0.81 Rh 0.805 3R 3.0684 18.579 AgScO 2 Ag 0.81 Sc 0.885 3R 3.2112 18.538 A gTlO 2 Ag 0.81 Tl 1.025 3R 3.568 18.818 AgYbO 2 Ag 0.81 Yb 1.008 3R 3.3717 18.214 CuAlO 2 Cu 0.60 Al 0.675 3R 2.8571 16.94 CuAlO 2 Cu 0.60 Al 0.675 2H 2.863 11.314 CuBO 2 Cu 0.60 B 0.410 3R 2.84 16.52 CuCoO 2 Cu 0.60 Co 0.685 3R 2.8488 16.92 CuCrO 2 Cu 0.60 Cr 0.755 3R 2.975 17.096 CuEuO 2 Cu 0.60 Eu 1.087 3R 3.63 17.08 CuFeO 2 Cu 0.60 Fe 0.785 3R 3.0351 17.166 CuGaO 2 Cu 0.60 Ga 0.760 3R 2.975 17.096 CuInO 2 Cu 0.60 In 0.940 3R 3.2922 17.338 CuLaO 2 Cu 0.60 La 1.172 3R 3.83 17.1 CuNdO 2 Cu 0.60 Nd 1.123 3R 3.71 17.09 CuPrO 2 Cu 0.60 Pr 1.130 3R 3.75 17.05 CuRhO 2 Cu 0.60 Rh 0.805 3R 3.074 17.094 CuScO 2 Cu 0.60 Sc 0.885 3R 3.2204 17.0999 CuScO 2 Cu 0.60 Sc 0.885 2H 3.223 11.413 CuSmO 2 Cu 0.60 Sm 1.098 3R 3.65 17.03 CuYO 2 Cu 0.60 Y 1.040 3R 3.533 17.136 C uYO 2 Cu 0.60 Y 1.040 2H 3.531 11.418 PdCoO 2 Pd 0.73 Co 0.685 3R 2.83 17.43 PdCrO 2 Pd 0.73 Cr 0.755 3R 2.9239 18.087 PdRhO 2 Pd 0.73 Rh 0.805 3R 3.0209 18.083 PtCoO 2 Pt 0.74 Co 0.685 3R 2.83 17.84 Reprinted with permission of Elsevier and P. Cann from Table 2 of M. A. Marquardt, N. A. Ashmore, D. P. Cann, Thin Solid Films 496, 150 (2006).

PAGE 46

46 Table 1 2 Compilation of doped delafossites compiled from the literature Species x Type (Scm 1 ) Type (cm 2 V 1 s 1 ) Ref erences CuCr 1 x Mg x O 2 0.05 Film 220 p N/A 1 0 9, 1 1 2 1 1 5 CuCr 1x Mg x O 2 0.05 Bulk 0.01 0.077 p <0.1 1 0 9, 1 1 3, 1 1 4 CuCr 1x Ca x O 2 0.04 Poly N/A N/A N/A 1 09 CuAl 1 x Mn x O 2 0.03 Bulk <10 6 p N/A 1 16 1 17 CuAl 1 x Mg x O 2 0.02 Film 8.3x10 2 p 0.1 118 1 19 CuY 1x Ca x O 2 0.02 Film 1 p N/A 1 1 0 CuIn 1x Ca x O 2 0.07 Film 2.8x10 3 p N/A 36 1 10 CuIn 1x Sn x O 2 0.05 Film 3.8x10 3 n N/A 36 1 10 CuGa 1x Ca x O 2 0.05 Bulk 5.6x10 3 p 0.1 50 CuFe 1 x Mg x O 2 0.02 Bulk 8.9 p 0.1 1 21 CuFe 1 x Sn x O 2 0.05 Bulk 2.4x10 4 n 1x10 6 1 21 CuRh 1x Mg x O 2 0.10 Bulk 167 m etal N/A 10 0 CuSc 1 x Mg x O 2 0.05 Film 30 p N/A 1 1 2, 1 22 1 23 AgIn 1x Sn x O 2 0.05 Film 20 70 n 0.5 1.4 1 24 1 27 Reprinted with permission of Elsevier and P. Cann from Table 6 of M. A. Marquardt, N. A. Ashmore, D. P. Cann, Thin Solid Films 496, 154 (2006).

PAGE 47

47 CHAPTER 2 ELECTRICAL AND STRUCTURAL PROPERTIES OF Cu Cr1 XM gXO2 THIN FILMS Background In recent years, there has emerged significant interest i n fu nctional oxide materials .1 One of the more important functionalities for electronic applications is that of the w ide bandgap semiconductors .24 Included among the wide bandgap semiconducting oxides are materials known as several delafossites.23, 3 5 3 6, 29, 30, 4456 92 Delafossites are ternary oxides with the basic formula AIMIIIO2, where AI can be palladium, platinum, silver, or copper in a monovalent state and MIII is a trivalent metal with a diamete r equal to or larger than that of aluminum and equal to or smal ler than that of lanthanum .4, 3 5 Those composed of palladium and platinum are metals, while those composed of silver or copper are s emiconductors. Some delafossite -type oxides show n -type conduction; others have narrow band gap. However, i n the case of several cuprate delafossites, the layered crystal structure behaves much like a multi quantum well structure, with Cu2O layers (Eg = 2. 1 eV) sandwiched between insulating layers of MIII 2O3. This dilates the energy bandgap of the Cu2O layers, yielding large r optical band gap s The evolution of these properties has been discussed in detail elsewhere (Chapter 1 ).30 Both nand p type semiconductors have been reported within the delafossite oxide s. The ability to obtain both pand n type delafossites opens the possibility of fabricating oxide semiconduct or pn heterojunctions. In addition, Yanagi et al. reported the formation of both nand p type CuInO2 by doping with Sn or Ca as donors a nd acceptors, respectively .35, 3 6 As a transparent conducting oxide (TCO), the delafossites offer intriguing properties, particularly as it pertains to the formation of a p type material. Most TCOs are degenera te n type semiconductors .24, 2 5 There ar e few TCOs that are p type .26, 2 7 1 28 P -type conduction in delafossites was first reported for CuAlO2.46 Several other delafossites have since also been

PAGE 48

48 shown to be ptype. Highly conductive p -type TCOs are important for polycrystalline photovoltaics, where junction formation between an n type absorber and the TCO is desired. For transparent thin -film electronics, there is a generic need t o develop p-type transparent semiconductor s for field -effect transistors and pn -junctions .37, 129132 There is also interest in utilizing dela fossite semiconductors for hydrogen production via water splitting .131 However, limitations in utilizing delafossites as semiconductors include relatively low conductivities, low mobilities, and, the sometimes observed challenge in achieving phase -pure delafossite thin films .43 There are a lso a lack of reports of improving crystal growth to achieve epitaxy for device structures. While phase -pure delafossite films have been rep orted in the literature efforts by other groups suggest that the processing con ditions are somewhat narrow .25 4 3, 1 15 Recently, CuCr1xMgxO2 has been shown to exh ibit relatively high conductivity (220 Scm1) p type conduction and optical transparency in the visible spectrum (though reports of indirect band gaps of 1.32 eV and 2.6 eV exists) .92, 115, 133137 Here, the synthesis and properties of epitaxial CuCr1xMgxO2 thin films are examined. In particular, we examined the phase stability and resultant pro perties as a function of deposition parameters and Mg content for films grown by pulsed laser deposition. Experimental Background Our films of CuCr1 xMgxO2 were gro wn by pulsed laser deposition. The substrates were single crystal c -plane sapphire substrate s The ablation targets were made by mixing and grinding stoichiometric amounts of 99.999% Cu2O, 99.998% MgO, and 99.97% Cr2O3. The powder mixture was then uniaxially pressed and sintered at 1300 C for 12 hours. The sapphire substrates were cleaned in son icated baths of trichloroe thylene, acetone, and methanol successively for 5 minutes and subsequently blown dry with nitrogen. The substrates were secured to the chambers heater surface using silver paint. The base pressure of the deposition

PAGE 49

49 chamber was le ss than 103 Pa and the high purity oxygen pressure used during the growth w as varied from 0.1 Pa to 5 Pa. The range of deposition temperatures varied from 400 C to 800 C. A KrF excimer laser was used as the ablation source with a laser repetition rate o f 5 Hz. After growth, the substrate temperature was lowered at a rate of 10 C /min in the ambient pressure used during deposition. Film thickness ranged from 250 nm to 400 nm as determined by surface profilometer scans. The crystallinity of the films was c haracterized using 2 -circle and 4 -circle X ray diffraction (XRD). The 2 -circle XRD measurements were performed with a Philips APD 3720. The 4 -circle XRD characterization was done with a Philips X'Pert Mate rials Research Diffractometer. Both employ a Cu K x ray source. Surface morphology was examined using atomic force microscopy (AFM) with a Digital Instruments Dimension 3100. The transport properties were determined by four -point resistivity measurements. Majority carrier type was also examined using Seeb eck measurements as Hall measurements proved to be challenging due to the low carrier mobility. Mg concentrations as reported are taken from the nominal atomic concentration of the targets. The Seebeck measurements were performed on a Quantum Design Physic al Property Measurement System. Results and Discussion XRD Analysis Powder XRD: temperature dependence The orientation and phase purity of the films was examined as a function of deposition parameters and Mg content. Figure 2 1 shows the XRD results for CuCr0.95Mg0.05O2 films grown at 1 Pa of O2 pressure and at temperatures ranging from 600 C to 750 C. The majority phase is the delafossite structure, although minority phases, including a CuCr2O4 phase possessing the spinel crystal structure, are o ften observed. The relatively narrow temperature window of 650 C to 700 C was found to produce phase pure delafossite material though this was often found to

PAGE 50

50 be difficult to reproduce The outlying conditions both promote spinel and the related Cu2O formation as readily evidenced by their (202) and (111) peaks respectively. Powder XRD: O2 pressure dependence The subsequent XRD 2 scans seen in Figure 2 2 are for films grown with a constant temperature of 700 C throughout the growth runs with an O2 pressure varied between 0.1 Pa and 5 Pa. While the films consist predominately of the de lafossite phase, there exists the secondary spinel phase, (Cu,Mg)Cr2O4, as seen by the (Cu,Mg)Cr2O4 (101), (202), and (303) diffraction peaks at both the 0.1 Pa and 5 Pa growth pressures. In coa rse terms, The Cu2O formation seems to be a function solely of the growth temperature and the spinel formation is a function of both temperature and oxygen pressure, as discussion in the next section. T he phase space for phase pure delafossite films seems to exist primarily between 650 C and 700 C around 1 Pa of O2 pressure. Thermodynamics of Cu, Cr, and O system The phase diagram derived from the Barin et al thermodynamic tables shown in Figure 2 3 explains the thermodynamic driving force toward producing the spinel phase during film growth .138 Unless the oxygen pressure is reduced during cool down, any thin -film growth technique will involve a cool down step that includes a path through the spinel stable region to reach standard temper ature and pressure. The effect of adding Mg to the growth recipe exacerbates this issue and is likely due to the nature of its 2+ valence state Powder XRD Mg dependence Increasing the alloy percentage of Mg decreases the c axis lattice parameter of the Cu Cr1 xMgxO2 phase in a basically linear fashion, except for x = 0.005 which acts as a significantly contracted state Figure 2 4 shows this trend, despite the fact Mg forms a larger Mg(II) cation in 6 -fold coordinat ion (with a 0.86 radius) than the Cr(III) cation in 6 -fold coordination (with a

PAGE 51

51 0.75 radius) otherwise sitting on its site. In addition, increasing the alloy percentage of Mg generally increased the amount of the spinel impurity phase present within the films. Figure 2 5 shows the XRD scans for CuCr1 xMgxO2 films grown at 700 C and 1 Pa O2 pressure where the Mg content is varied from 0 to 5 at.%. All XRD 2 scans are calibrated to the substrate Al2O3 (006) peak at 41.71 and a value of 100,000 counts. T he relative intensities of the spinel (101) peaks increase with an increase in Mg content. Both the copper and magnesium components can act as divalent A -site ions in the A M2O4 spinel crystal lattice. One can also observe Cr2O3, Cu2O, and CuO as impurity phases within some Mg doped CuCrO2 films. Figure 2 5 ( b ) shows the XRD data over the region of 2 where peaks for these minority phases are best seen. In the absence of any Mg doping, Cr2O3 becomes the abundant minority phase. A further increase in the Mg percentage increases the dominance of the Cu2O and Cu/MgCr2O4 phases and largely eliminates the Cr2O3 phase. The n ature of the spinel is evident in phase pure films grown at 450 C and 5 Pa of O2 pressure from an x = 0.05 CuCr1 xMgxO2 target. It was found to have a Seebeck coefficient of 128 V/K and a resistivity of 0.49 Ohm -cm. An Arrhenius plot (not provided) showed the film to have an activation energy of 147 eV and demonstrated its semiconducting behavior. The optical transmission ( Figure 2 6 ) dat a shows a much less transparent film than its delafossite cousin ( the optical transmittance < 2% throughout the visible spectrum for 400 nm th ick films ) which likely points to a metallic nature or strongly adsorbing Cu d9 and Cr d3 states Also, the inset Tauc plots of Figure 2 6 show that the indirect fit (E1/2) likely provides a better picture of the optical transition than the direct fit (E2).

PAGE 52

52 Epitaxy with 4 -c ircle XRD The epitaxial orientation and crystalline q uality of the CuCr1xMgxO2 films was examined with -scans as seen in Figure 2 7 The figure shows the -scan through the ( 018) CuCrO2 reflections. A six -fold symmetry in the CuCr0.95Mg0.05O2 diffraction peaks is e vident. The in plane axes of the delafossite are rotated 30 relative to t hat for the sapphire substrate. The expected 3 -fold geometry of the delafossite (108) plane is seen twinned into its 6 -fold twinned analogue. This is the same orientation as seen in epitaxial relationship for CuGaO2 and CuAlO2 on sapphire, with out -of -plane (001) CuCrO2 // (0001) -Al2O3, in -plane (110) CuCrO2 -Al2O3.139 1 40 Figure 2 8 (a ) shows the rocking curve FWHM (full width half maximum) for the CuCrO2 (006) pea k to be 1.13 The rocking curve of the c -oriented CuCr2O4 (202) plane is also shown in Figure 2 8 ( b ). The results show that the minority spinel phase is also highly textured /epitaxial with a FWHM of 2.306 in i ts rocking curve. Surface Roughness: AFM The surface morphology of the films was exa mined using tapping mode AFM. Figure 2 9 shows an image for a CuCr0.95Mg0.05O2 film grown at 650 C and 1 Pa O2. This film has a root mean squared roughness of 10 nm. One interesting aspect of the AFM image is the 3-fold symmetry demonstrated in the morphology. This is consistent with the multiple in -plane variants in orientation seen in the X ray diffraction data. The crystallite sizes are typically 400 nm in length and 100 nm in width. 4 -Point Van der Pauw resistivity Mg dependence Four -point resistivity measurements were performed for the various CuCr1xMgxO2 thin films at 25 C. Contacts on the film surface were made with indiu m metal which was found to be Ohmic. Resistivity varied from ~102 to 104 ohm -cm, dependent on growth conditions and Mg

PAGE 53

53 content. Figure 2 10 shows film resistivity plotted as a function of growth temperature for va rious O2 pressures. For these measurements, 300 nm thick films of CuCr0.95Mg0.05O2 were employed. The resistivity generally increased as the dep osition temperature increased. The generally lessened presence of the spinel phase ( Figure 2 5 ) may contribute to the trend. Figure 2 11 provides the resistivity for CuCr1xMgxO2 thin films grown at 700 C, 1 Pa O2 as a function of Mg concentration. The results show a decrease in re sistivity as the Mg concentration is increased. This is consistent with the Mg serving as a p type dopant in the CuCrO2 matrix. Notably the resistivity appears to saturate for 3 at.% Mg concentration, suggesting that the solid solubility limit may have be en reached, although this conclusion is inconsistent with the continued de crease in c lattice constant with dopant concentration. The resistivity saturation may simply indicate the onset of compensating donor defects. Attempts to extract a Hall voltage fro m van der Pauw Hall measurements proved to be unsuccessful due to a low carrier mobility. An upper limit on the H all mobility was on the order of 0.1 cm2/V -s. Carrier type was examined using Seebeck measurements. The sign of the Seebeck coefficient was pos itive (except for the x = 0.005 film) consis tent with p type conductivity. More detailed analysis of the carrier transport is provided in Chapter 3 Optical Transmission: Mg D ependence Optical properties of th e films were examined by optical transmission over a wavelength of 200 900 nm. Figure 2 12 shows the results for CuCr1 xMgxO2 films grown at 700 C and 1 Pa O2. Figure 2 1 2 (A ) provides a Tauc plot for four samples with varying Mg concentrations. The Tauc plots show direct bandgap behavior with the band gap ranging from 3.19 eV to 3.02 eV for 0.5% Mg and 1.0% samples. Figure 2 12(B) shows that the optical transmission is greater than 40% for the range of visible light.

PAGE 54

54 Summary The growth and properties of highly conductive epitaxial optically transparent films of CuCr1xMgxO2 is reported. Con ductivities on the order of 70 S -cm1 w ere realized with bandgaps ranging from 3.02 eV to 3.19 eV. Epitaxy on c -sapphire was confirmed with -scans sh owing twinned 3-fold symmetry. The issue of spinel formation is addressed and more work is necessary to implement different strategies to suppres s this problematic phase as it may play a role in quashing clean CuCrO2/ZnO interfaces. For implementation in thin -film electronics, future efforts should address the issues related to junction form ation and device process .141, 1 42

PAGE 55

55 Figure 2-1. XRD -2 scan for CuCr0.95Mg0.05O2 thin films showing the behavior as a function of deposition temperature with a fixed pressure of 1 Pa and temperatures of (A) 750 C, (B) 700 C, (C) 650 C, and (D) 600 C at a range between A) 2 = 15 to 70 and in more detail at a range of B) 2 = 32 to 42 to more obviously show the minority phase peaks.

PAGE 56

56 Figure 2-2. XRD -2 scan for CuCr0.95Mg0.05O2 thin films showing the behavior as a function of deposition pressure with a fixed temperat ure of 700 C at a range between A) 2 = 15 to 70 and in more detail at a range of B) 2 = 32 to 42 to more obviously show the minority phase peaks.

PAGE 57

57 Figure 2-3. Thermodynamic plot of the Cu, Cr, and O system. Figure 2-4. c-Axis lattice parameter as a function of Mg content in CuCr1-xMgxO2 (x = 0.0, 0.005, 0.01, 0.03, 0.05) thin films grown at 700 C and 1 Pa O2 pressure.

PAGE 58

58 Figure 2-5. Compilation of XRD -2 scans of CuCr1-xMgxO2 (x = 0.0, 0.005, 0.01, 0.03, 0.05) grown at 700 C and 1 Pa O2 pressure shown at A) larg e scale and B) in detail.

PAGE 59

59 Figure 2-6. Transmittance data for spinel sample grown at 450 C and 5 Pa O2 with inset E1/2 (indirect) and E2 (direct) Tauc plots.

PAGE 60

60 Figure 2-7. XRD -scans of CuCr0.95Mg0.05O2 grown on c-Al2O3 showing twinned epitaxy of both CuCrO2 and CuCr2O4 phases.

PAGE 61

61 Figure 2-8. XRD -rocking curves for CuCr0.95Mg0.05O2 grown on c-Al2O3 for the A) CuCrO2 (006) peak and B) the CuCr2O4 (202) peak.

PAGE 62

62 Figure 2-9. AFM image of CuCr0.95Mg0.05O2 film grown at 650 C and 1 Pa O2 pressure showing a 3-fold crystallite symmetry. Figure 2-10. Resistivity plotted against growth temperature for CuCr0.95Mg0.05O2 films grown under 0.1 Pa, 1 Pa, and 5 Pa of O2 pressure.

PAGE 63

63 Figure 2 11. Resistivity plotted against Mg doping composition for CuCr1xMgxO2 films grown at 700 C under 1 Pa O2.

PAGE 64

64 Figure 2-12. Optical tran smission data of CuCr1-xMgxO2 films grown at 700 C and 1 Pa O2 pressure graphed in A) a Tauc plot form at showing direct ba ndgap behaviors and in B) an optical transmittance plot.

PAGE 65

65 CHAPTER 3 XPS AND ELECTRI CAL CHARACTERIZATION OF C u Cr1XM gXO2 DELAFOSSITE THIN FILMS GROWN BY PULSED LASER DEPOSITION Introduction As a transparent conducting oxide (TCO), the delafossites offer intriguing properties, particularly as it pertains to the f ormation of a p type materi al. Most TCOs are degenerate n type sem iconductors .24, 2 5 There are few TCOs that are robust p -type candidates due often to the localization of charge upon oxygen atoms in the lattice .26, 2 7 1 28 Thus, when ptype conduction in delafossites was first reported for CuAlO2 it created a great deal of interests in other delafossites where properties could be further ta ilored according choices of the MIII metal and dopants .44 Several other delafossites have since also been shown to be p -type (Table 1 -2 ). A desire exists for creating polycrys talline TCO photovoltaics, where junction formation between an n typ e absorber and the TCO is advantageous for transparent thin -film electronics, for field -effect transistors for pn -junctions and even for semiconductors for hydrogen production via water splitting .37 1 29132 However, limitations in utilizing delafossites for such applications include relatively low conductivities, low mobiliti es, and, the sometimes observed challenge in achieving phase -pure delafossite thin films.4 3 While phase -pure delafossite films have been reported in the literature efforts by other groups suggest that the proces sing conditions are restrictive .24; 4 3 Recently, CuCr1 xMgxO2 has been shown to exhibit relatively high conductivity, high carrier concentration (as evidenced by low mob ilities), p type conduction and optical transparency i n the visible spectrum 92 1 15137 In this chapter the synthesis and properties of epitaxial CuCr1 xMgxO2 thin films is examined in depth with extra attention given to an analysis of dopant characteristics with XPS optical transmission and electrical data.

PAGE 66

66 Experimental Description Epitax ial films of CuCr1xMgxO2 were gr own by pulsed laser deposition. The substrates were single crystal c -plane sapphire. The ablation targets were made by mixing and grinding stoichiometric amounts of 99.999% Cu2O, 99.998% MgO, and 99.97% Cr2O3. The powder mi xture was then uniaxially pressed and sintered at 1300 C for 12 hours. The sapphire substrates were cleaned in sonicated baths of trichloroethylene, acetone, and methanol each for 5 minutes and subsequently blown dry with nitrogen. The substrates were sec ured to the chambers heater surface using silver paint. The base pressure of the deposition chamber was less than 103 Pa and the high purity oxygen pressure used during the growth was kept at 1.0 Pa. The films were grown at 700 C. A KrF excimer laser wa s used as the ablation source with a laser repetition rate of 5 Hz. After growth, the substrate temperature was lowered at a rate of 10 C/min in the ambient pressure used during deposition. Film thickness ranged from 250 nm to 400 nm as determined by surf ace profilometer scans. The lattice parameters of the films were characterized using 2-circle XRD. The 2 circle XRD was performed with a Philips APD 3720 em ploying a Cu K x ray source. The resistivity was determined by four -point Van der Pauw resistivity measurements in a Quantum Design PPMS (Physical Property Measurement System). Majority carrier type was also examined using Seebeck measurements in the PPMS as Hall measurements proved to be challenging due to the low carrier mobility. Mg concentrations as reported are taken from the nominal atomic concentration of the targets. Optical transmission measurements were taken on a UV/Vis P erkin Elmer Lambda 800 system. XPS measurements were taken from a Perkin Elmer PHI 5100 system using a Mg anode.

PAGE 67

67 Results a nd Discussion XRD Discussion: Mg dependence The c axis lattice constant was examined as a function of Mg doping percentage. Figure 3 1 shows the XRD results for CuCr0.95Mg0.05O2 films for an scan with particul ar focus on the delafossite (006) peak. The double peaks seen for the x = 0.00, 0.03, and 0.05 cases likely correspond to a structural relaxation of delafossite lattice due to the 8% lattice mismatch between the substrate and the films. The majority phase is the delafossite structure, although minority phases, including small amounts of a (Mg, Cu )Cr2O4 phase possessing the spinel crystal structure, were observed. Chapter 2 deals with this spinel phenomenon in far greater detail and is most evident by examining the Cu/MgCr2O4 (101) peak ( Figure 2 5 ).114 The relative amount of the spinel phase as represented by the (101) peak in the samp les was seen to increase with an increase in Mg concentration ( Figure 2 5 ). Other authors have explained that the small size of MIII cation delineates between those delafossites which may be intercalated with oxygen and those which become spinels.7 5, 1 45 Equation 1 7 provides a n approximate upper limit of 0.665 for the MIII diameter for creating CuM2O4 spinels despite literature references and our own e xperience at growing phase -pure spinel CuCr2O4 which corresponds to a n octahedral cation radius of 0.755 for Cr3+. This structure also has a n approximate Cu2+ tetrahedral cation radius of 0.71 as compared to a tetrahedrally coordinated radius of 0. 71 for Mg2+ allowing it to sit well on the tetrahedral site.189 As for an examination of Figure 3 1 two features seem most striking. First, the two -peaked nature of the (006) is consistent for each films except the x = 0.005 and x = 0.01 cases These seem to correspond to either a mixing of 3R and 2H polytypes or a relaxation of CuCrO2 due to its lattice mismatch with c -sapphire. Secondly, t he more dominant peak for each film shows a fairly linear progression toward s maller c lattice values with an increase in Mg concentration.

PAGE 68

68 This seems to verify magnesiums solubility in the CuCrO2 lattice. It also yields c lattice parameters of 17.30 17.17 17.25 17.23 and 17.21 for x = 0.00, 0.005, 0.01, 0.03, and 0.0 5 all of which are well above the literature reported value of 17.1 for CuCrO2. Since the c lattice parameter is almost entirely unaffected by changes of MIII, it is likely that a substantial proportion of the magnesium is substituting on the Cu site. This will be analyzed more fully in the coming discussion on Seebeck coefficients as well as the severe contraction of the x = 0.005 sample. However, Figure 3 2 will be useful in explaining the deviation of propertie s for CuCr0.97Mg0.03O2 in the discussion which follows. Small Polaron Discussion The nature of small polaron materials has been explained by theoreticians elsewhere.47, 1 46 A polaron is an elect ron which carries its lattice distortion, due to its polarizing effect, with it as i t moves through the lattice. This particular category of polarons is small in the sense that the size of their interaction with the lattice is limited to a few nearest neig hbors. They are also described as hoping from lattice site to lattice site. This inherent transport inefficiency limits their th eoretical hole mobility to a maximum of ~1 cm2V1s1 due to this inefficient conduction mechanism This is obviously a bane for achieving high mobility p -type materials. Several delafossites have been noted to have their conductivity dominated by this transport mechanism, most notably below 220 K.28, 68 92, 147 Summary of activation energies Activation energies calculated for Arrhenius plots of the conductivity versus temperature imply the more typical thermally activated carrier conduction for delafossites at room temperature. P erhaps the best summary of the literature is provided by Ashmore et al showing activation energies for various delafossites ranging from 0.02 eV for a film of CuCrO2:Mg to

PAGE 69

69 0.36 eV for bulk CuCrO2.68 1 47 1 48 Most values have been found to be within t he range of 0.10 eV to 0.25 eV. Temperature dependence of conduction mechanism Figure 3 3 A shows our own conductivity versus temperature plot of a CuCr0.99Mg0.01O2 thin film. Though an Arrhenius plot may be appropriate for the temperature range of ~220 K to 300 K, Figure 3 3 B shows that the classic small polaron plot ( log( ) (1/T)1/4) fits the da ta linearly for all temperatures extremely well with three distinct regions: I, II, and III. The temperature boundaries are marked by double dashed lines around ~ 220 K and ~90 K between regions I and II and II and III respectively. Regions I, II, and III are fit to have linear fits with slopes of BI = 64.3, BI I = 51.0, and BI II = 59.9 respectively. The 220 K boundary may be well explained by the transition from small polaron to thermally activated carrier transport. The 90 K boundary has not, to our knowle dge, been examined in the literature in application to delafossites but may be due to a polaron band that may exists at low temperatures. Schmid explained the phenomenon that the transport of polarons is facilitated by polaron sites of equivalent energy, basically forming a band, which would naturally be the case at temperatures approaching absolute zero.149 Thus, as temperatures are raised above ~90 K ( kT = 7.8 meV) the spread of site energies increases and the opp ortunity of polarons to jump between sites of equivalent energy drops precipitously disrupting transport within this band However, on par with the previous discussion of the high likelihood of some Mg doping on Cu lattice sites, this transition could be due to the activation of carriers from these Mg donor sites. Resistivity: Mg Dependence Four -point resistivity measurements were performed for the various CuCr1xMgxO2 thin films at 300 K Contacts on the film surface were made with indium metal solder wh ich was found by linear I -V curves to be Ohmic. Resistivity varied from ~102 to 104 ohm cm, dependent

PAGE 70

70 on growth conditions and Mg content. Figure 3 4 shows the resistivity for CuCr1xMgxO2 thin films grown at 700 C, 1 Pa O2 as a function of Mg concentration. The results show a decrease in resistivity as the Mg concentration is increased which is consistent with the Mg serving as a ptype dopant in the CuCrO2 matrix. Not ably, the resistivity appears to saturate fo r 3 at.% Mg concentration, suggesting that the solid solubility limit may have been reached, although this conclusion is inconsistent with the continued de crease in lattice constant with dopant concentration. The presence of both CuxO (1 x 2) and Mg/CuCr2O4 spinel phases at the x = 0.03 growth condition may explain this resistivity minimum in the Mg doping sequence. However, Ono et al showed a minimum of resistivity at x = 0.03 for powder samples .113 The presence of Mg in the lattice, while creating holes in the lattice, de creases the overlap of Cu 3d orbitals by shortening Cu O bonds since t he strength of xy interactions (dxy and dx 2-y 2) are inversely proportional to the strength of the z bond (dz 2-s) The compet ition seems to reach a maximum at x = 0.03 though the lattice seems able to accept higher concentrations of Mg Transport Analysis: Seebeck Data Seebeck data is presented in Figure 3 5 Look et al explained the im portance of Seebeck measurement s to elucidate the dominant carrier type in low mobility films since Hall measurements depend on a 2p 2n relationship while Seebeck has a p dependency.1 51 This makes achieving accur ate Hall measurements extremely difficult for low mobility or mixed carrier samples. Figure 3 5 shows a general increase of Seebeck coefficient with an increase of Mg percentage with two notable issues at the x = 0 .005 and x = 0.03 which have coefficients of 32 V/K and 16 V/K respectively. This general trend is in marked difference to results reported by Ono where 350 K Seebeck values for bulk polycrystalline samples exceeded 1000 V/K for the x = 0.00 case and dropped precipitously for a minimum of 200 V/K fo r the three cases x =

PAGE 71

71 0.03, 0.04, and 0.05.113 The negative value at x = 0.005 and small value at x = 0.03 deviate from the large values of 192 V/K and 513 V/K for x = 0.01 and x = 0.05 respectively. The latter two cases seem to correspond to the high hole concentrations expected for the CuCrO2:Mg system. The presence of the aforementioned CuxO (1 x 2) and spinel phases for the x = 0.03 case may act in myriad ways to compensate the hole contributions from the del afossite system with corresponding donor states. The negative Seebeck coefficient for the x = 0.005 growth condition doesnt show an exceptional increase in the presence of a spinel or CuxO (1 x 2) phase and thus an alternate phase presence cannot be u sed to explain the negative Seebeck coefficient. A report on polycrystalline samples from Okuda et al confirms that the solubility of Mg extends from x = 0.00 to x = 0.05 and specifically included x = 0.005 with no marked effects for this doping level.152 It may indicate the presence of Mg sitting on the Cu lattice as the species MgCu which effectively places extra electrons into the lattice thereby acting as a donor. This seems to be partnered with a sharp constrictio n of the c lattice parameter which should accompany a weakening of Cu Cu interactions. Optical Transmission Analysis Optical transmission data of the films presented in a previous publication provides strong evidence for a direct bandgap of in CuCr1 xMgxO2 for 0.00 x 0.05 and though lacking a strong trend displays a hiatus between those films of x 0.005 and those of x 0.01.114 Figure 3 6 shows a drop in band gap between a value of 3.19 eV for x = 0.005 and 3.02 eV for x = 0.01. The figure also shows an increase of 0.02 eV between the undoped and x = 0.005 states which may be attributed to the existence of donor states ( from MgCu ) low levels of Mg substitution which are masked with higher Mg concentrations. It may be worthwhile to examine much lower doping levels of x < 0.005 to attempt to establish a trend. Also, t he typical value of 3.08 eV

PAGE 72

72 given f or the bandgap of CuCrO2 may be dilated here (x 0.005) by an increase o f the lattice in the z direction over the standard value of 17.1 and may have much to do with the significant presence of Mg donor states XPS Data: Mg Dependence Explanation of fitting methodology An examination of XPS data for varying Mg content shows a consistent shift in Fermi energy with an increase in Mg concentration. All data was taken from samples sputtered for five minutes with argon in sit u and calibrated against a C 1s value of 284.9 eV for the small amounts of adventitious carbon. The presen ce of the additional peak at 297 eV represents carbonyl species not seen in the samples prior to the argon sputter likely due to high energy reactions between released oxygen and surface carbon. A multiplex analysis was performed on three energy windows to capture detailed binding en ergy information for Cu 2p (960 925 eV), Cr 2p (605570 eV), and O 1s (535 528 eV). All three conditions involved 20 sweeps, a 0.1 eV/s sweep speed (0.2 eV/s for O 1s), and 22.36 eV pass energies. The Mg 2p peak was difficult to detect due to much more represented Fe 3d peak, from the sample puck, and is here neglected in our analysis. Figure 3 7 shows a multiple Gaussian peak fit for CuCr0.99Mg0.01O2. Each CuCr1 xMgxO2 data set was ana lyzed using the same methodology The fits were collectively subtracted from the raw data and the resulting delta values are plotted at the base of the plots. Analysis: Cu 2p peak Table 3 1 tabulates the peak fit values for Cu 2p. The Cu states are divided into AI and AII which correspond to equivalent Cu2O and CuO oxidation states found in the literature. Notably, CuO Cu 2p states (AII) are typically marked by broad peaks and extremely broad shake up satellites ( seen in Figure 3 7 around 944 eV as the extremely broad peak in the middle of the plot) close to 9 eV higher in binding energy than their respective 2p3/2 and 2p1/2 peaks. The Cu

PAGE 73

73 AIIa 2p3/2 values have a high of 9 34.4 eV with x = 0.00 to variable values for the other doping levels in the range of 933.4933.8 eV.153 This Cu oxidation state demonstrated a 2p3/2 to 2p1/2 peak splitting of greater than 20.0 eV for all doping lev els ( Figure 3 8 Table 3 1 and Table 3 2 ). Cu2O Cu 2p states are much more narrow with the most prominent peak seen being AIA. T his peak exhibited a shift of ~0.5 eV between x = 0.00 and the films containing Mg and narrow from 1.7 eV to 1.31.4 eV for the same films. The AIB peaks seem to display a trend (2.0 eV to 1.6 eV for x = 0.00 to x = 0.05) as they narrow with an increase in Mg percentage. The peak splitting between AIA and AIB (2p3/2 and 2p1/2) exhibited a consistency of 19.8 eV for all doping levels. This peak is the largest contributor to the added trend line exhibited in Figure 3 8 which runs roughly through the general 2p3/2 maximums. This evaluation provides two insights. First, the consistency of the Cu2O binding energies, peak splitting, and peak FWHM values demonstrates the default presence of the Cu1+ value within the delafos site lattice as expected by stoichiometry. The presence of the CuO oxidation states for all samples including the undoped film may indicate the default presence of additional copper oxidation within the lattice. Also, the relative presence of the Cu2+ oxidation states (seen as the Area percentage) shows no dependence upon Mg level showing the possible presence of a native oxygen intercalation. This is also augmented by the presence of two oxygen 1s states shown in a following section. Analysis: Cr 2p peak Cr 2p values also reveal two basic oxidation states denoted as BI and BII in Figure 3 7 and Table 3 3 The literature provides a common estimate of 580.0 eV for the 2p3/2 bi nding energy for Cr6+ and 576.8 eV for Cr3+, a good fit with the BIA. BIIA can therefore represent either a Cr4+ or a Cr(3+ )+ oxidation state.154 The peak splitting seen of both states stands between 9.59.7 eV

PAGE 74

74 for every film and the full width half max (FWHM) values average 2.5 eV for BIA, 2.7 eV for BIB, 3.65 eV for BIIA, and 3.6 eV for BIIB. The Cr 3d states have been calculated to play a substantial role in effecting the extremely high conductivities found for the CuCrO2:Mg system and warrant an examination of Cr 2p binding energies and their implied oxidation states. The feature labels BI corresponds to the Cr2O3 (Cr3+) oxidation state which is the dominant state achieved by the stoichiometry of the delafossit e system and BII represents CrO1.5+ oxidation state (Cr(3+ )+). The Cr states are represented in higher proportion by the more expected Cr 3+ oxidation state a total of 54 65% of the oxidation states, with 65% corresponding to the undoped film. This seems to correspond to the high acceptor levels achieved on the doped films contributing possibly to a Cr4+ oxidation state but also seems to show this heightened oxidation state exists in the native material. The most noticeable trend is perhaps the BI oxidati on state. It shifts from a high of 576.8 eV at x = 0.00 to a low of 576.0 eV at x = 0.05. Analysis: O 1s peak Oxygen 1s and valence states were also examined for all films and showed similar binding energy shifts as found for the previous elements. Speci fically, O 1s demonstrated two bonding states denoted CI (~530 eV) and CII (~531.4 eV) where CI represents the bulk O atoms and CII is traditionally assigned to the OH species (Table 3 5 and Figure 3 8 ). Considering that the films were sputtered i n situ there should be a substantial reduction in adsorbed oxygen species such as this possible hydroxide layer The CII does shift consistently with increases in doping levels fro m 531.8 eV at x = 0.00 to 531.3 eV at x = 0.05 and peaks in its predominance (24%) at x = 0.03 and thus seems to coincide with a native state which tracks the resistivity trend.

PAGE 75

75 Analysis: Valence states The valence states were measured coarsely at either 1 .0 or 0.5 eV/s over 10 sweeps with a pass energy of 89.45 eV. Though the coarse measuring prevents a precise picture of the valence states of these films there still seems to be a general trend mirrored in the other described bonding states. Approximate pe ak center values for x = 0.00, 0.005, 0.01, 0.03, and 0.05 are 3.2 eV, 2.75 eV, 2.5 eV, 2.5 eV, and 2.5 eV respectively. These peak values place the right edge (the valence band edge) of the curves within 0.5 eV of the Fermi energy which may then be direct ly compared to the 0.090 eV activation energy previously described. Conclusion The research has sought to describe anomalies in the resistivity of the doping sequence of CuCr1xMgxO2 0.00 x 0.05, namely the negatively Seebeck coefficient for x = 0.005 and the increase of Seebeck coefficient with the doping level. The presence of significant oxygen intercalation in the undoped films is indicated by the two states seen of both O 1s and Cu 2p. XPS measurements confirm that doping effectively lowers the Fer mi energy at all doping levels with decreases of 0.6 eV, 0.8 eV and 0.5 eV for O 1s, Cr 2p3/2, and Cu 2p3/2 states respectively regardless of the magnitude of the Seebeck coefficients. This seems to correspond with a dramatic effect of doping on the Cr 2p state which corroborates reports that CuCrO2 resistivity benefits from a substantial hybridization of the Cr 3d states with the expected Cu 3d and O 2p states. The XPS measurements also confirm the predominant shift of binding energies takes place between the undoped and doped films with a small shift at higher doping levels. The CuCr0.97Mg0.03O2 film show s both a surprisingly low Seebeck coefficient and the lowest resistivity among these films likely due to both a high incidence of CuxO and spinel phase im purities and an effective Mg saturation /c lattice constriction limit while the CuCr0.995Mg0.005O2

PAGE 76

76 film likely has a negative Seebeck coefficient likely do to either its c -lattice contraction (possibly from less than stoichiometric levels of Cu) and/or the presence of Mg as a donor

PAGE 77

77 Figure 3-1. XRD -2 data from CuCr1-xMgxO2 of (006) peak for 0.00 x 0.05.

PAGE 78

78 Figure 3-2. XRD -2 data from CuCr1-xMgxO2 0.00 x 0.05 films with an emphasis placed upon spinel and CuxO phases. W K peak is seen at 37.5.

PAGE 79

79 Figure 3-3. Resistivity data for 400 nm CuCr0.99Mg0.01O2 film plotted as a A) log( ) v [1/T] graph showing Arrhenius dependence and as B) log ( ) v [1/T]1/4 showing small polaron dependence.

PAGE 80

80 Figure 3 4 Van der Pauw 4-point resistivity of Mg dependence of CuCr1xMgxO2 at room temperature. Fi gure 3 5 Seebeck data for thin films of CuCr1 xMgxO2.

PAGE 81

81 Figure 3-6. Direct band gaps calculated from Tauc plots of respective films.

PAGE 82

82 Figure 3-7. XPS multiplex for Cu 2p, Cr 2p, and O 1s showing Gaussian peak fits and resulting difference (right axis) between actual and calculated values.

PAGE 83

83 Figure 3-8. XPS plots showing bi nding energy changes with an in crease in Mg content x for CuCr1-xMgxO2 thin films

PAGE 84

84 Table 3 1 XPS peak fitting values for the Cu 2p pea ks representing four key features, AIa, AIb, AIIa, and AIIb which are classically assigned as Cu2O 2p3/2, Cu2O 2p1/2, CuO 2p3/2, and CuO 2p1/2. BE is the binding energy of said peaks, Area is the percentage of the area under the calculated Gaussian fit com pared with the total fitted area, and FWHM is the full width half max value of the peak A Ia A Ib A IIa A IIb Cu 2 O 2p 3/2 Cu 2 O 2p 1/2 CuO 2p 3/2 CuO 2p 1/2 x for CuCr 1 x Mg x O 2 BE (eV) Area FWHM (eV) BE (eV) Area FWHM (eV) BE (eV) Area FWHM (eV) BE (eV) Area FWHM (eV) 0 .000 933.0 44% 1.7 952.8 21% 2.0 934.4 17% 4.2 954.4 18% 5.4 0.005 932.5 44% 1.4 952.3 20% 1.7 933.8 21% 4.7 954.6 15% 4.7 0.01 0 932.6 58% 1.3 952.4 16% 1.9 933.4 8% 3.9 954.7 18% 4.8 0.03 0 932.3 45% 1. 4 952.2 22% 1.8 933.6 18% 4.2 954.0 16% 6.2 0.05 0 932.5 43% 1.4 952.3 17% 1.6 933.8 21% 4.7 953.6 20% 4.7 Table 3 2 XPS v alues of doublet peak energy differences ( BE (eV)) between the Cu2O 2p1/2 and 2p3/2 and CuO 2p1/2 and 2p3/2 oxidation states. A I A I I x for CuCr 1 x Mg x O 2 Cu 2 O CuO BE (eV) BE (eV) 0 .000 19.8 20.0 0.005 19.8 20.8 0.010 19.8 21.3 0.030 19.8 20.5 0.050 19.8 19.9

PAGE 85

85 Table 3 3 XPS peak fitting values for the Cr 2p peaks representing four key features, BIa, BIb, BIIa, and BIIb which are classically assigned as Cr2O3 2p3/2, Cr2O3 2p1/2, CrO3 2p3/2, and CrO3 2p1/2. BE is the binding energy of said peaks, Area is the percentage of the area under the calculated Gaussian fit compared with the total fitted area, and FWHM is the full width half max value of the peak. B Ia B Ib B IIa B IIb Cr 2 O 3 2p 3/2 Cr 2 O 3 2p 1/2 CrO1.5+ 2p3/2 CrO1.5+ 2p1/2 x for CuCr 1 x Mg x O 2 BE (eV) Area FWHM (eV) BE (eV) Area FWHM (eV) BE (eV) Area FWHM (eV) BE (eV) Area FWHM (eV) 0 .000 576.8 41% 2.6 586.4 24% 2.8 579.0 21% 3.4 588.6 14% 3.4 0.005 576.2 36% 2.5 585.7 21% 2.5 578.0 26% 3.9 587.6 17% 3.5 0.01 0 576.2 35% 2.5 585.7 19% 2.6 578.0 27% 4.0 587.7 19% 3.8 0.030 576.1 40% 2.6 585.5 23% 2.7 578.2 22% 3.4 587.7 15% 3.4 0.050 576.0 36% 2.4 585.6 23% 2.9 578.0 24% 3.6 587.7 17% 3.8 Table 3 4 XPS v alues of doublet peak energy differences ( BE (eV)) between the Cr2O3 2p1/2 and 2p3/2 and CrO1.5+ 2p1/2 and 2p3/2 oxidation states. B I B I I x for CuCr 1 x Mg x O 2 C r 2 O 3 C r O 1.5+ BE (eV) BE (eV) 0 .000 9.6 9.6 0.005 9.5 9.7 0.010 9.5 9.7 0.03 0 9.5 9.5 0.050 9.6 9. 7

PAGE 86

86 Table 3 5 XPS peak fitting values for the O 1s peaks representing two key features, CI and CI I, which are classically assigned as bulk oxygen and other oxygen species respectively BE is the binding e nergy of said peaks, Area is the percentage of the area under the calculated Gaussian fit compared with the total fitted area, and FWHM is the full width half max value of the peak. C I C II O (bulk) 1s O (other) 1s x for CuCr 1 x Mg x O 2 BE (eV) Area FWHM (eV) BE (eV) Area FWHM (eV) 0.000 530.4 83% 1.8 531.8 17% 2.1 0.005 529.9 81% 1.7 531.5 19% 2.0 0.010 530.0 78% 1.6 531.4 22% 1.9 0.030 529.8 76% 1.7 531.4 24% 2.1 0.050 529.8 81% 1.7 531.3 19% 1.9

PAGE 87

87 CHAPTER 4 ELECTRICAL TRA NSPORT AND STRUCTURAL STUDY OF DELAFOSSITE Cu S c1XM gXO2 THIN FILMS GROWN BY PULSED LASER DEPOSITION Interest in CuSc1xMgxO2 Recent efforts in p type oxide materials have identified the class of ternary metal oxides known as delafossites as promising semi conductor candidates.28, 29, 4453, 92 This class of compounds is represente d by the formula A1+M3+O2 where M3+ has a radii between Al3+ (0.535 ) and La3+ (1.032 ) and A1+ must be Cu1+, Ag1+, Pt1+, or Pd1+.29, 35 In general, the delafossites composed of silver or copper act as semiconductors and f or the chromate and scandate co pper delafossites, the optical bandgap is dilated from 2.1 eV for Cu2O to greater than 3.1 eV for CuCrO2 and 3.6 eV for CuScO2. Given the p-type properties of these materials, there is significant interest in the thin film growth of delafossite semiconductors. Here, the formation and prope rties of CuSc1 xMgxO2 thin films grown by pulsed laser deposition (PLD) are addressed. CuScO2 has been found to be the smallest delafossite which can be inter calated with oxygen and readily doped. Doping a divalent metal such as Mg, on the trivalent site can introduce effective acceptor states and mobile holes. Little work has been reported for CuScO2 alloys and heterojunctions despite the fact that the a axis lattice p a rameter of CuScO2 (a = 3.22 ) matches well with that for ZnO (a = 3.25 ). There has also been little discussion on using buffer techniques which may allow the growth of epitaxial films at lower temperatures than 9 2 5 C.139 Specific issues that are addressed in this study include achieving highly oriented epitaxial films through buffer techniques r e alizing high p type carrier concentrations, and avoiding the formation of competing secondary phases. Experimental Procedure Thin films of CuSc1xMgxO2 were grown by pulsed l a ser deposition. The substrates were single crystal c -plane sapphire. The ablation targets were made by mixing and grinding

PAGE 88

88 stoichiometric amounts of the various 99.999% Cu2O, 99.998% MgO, 99.97% Cr2O3, and 99.99% Sc2O3 powders ap propriate for each target. The powder mixtures for the delafossites were then uniaxially pressed and sintered at 1300 C for 12 hours which resulted in black ta r gets. All films were grown using a KrF excimer laser (wavelength = 2 48 nm). Prior to film depos ition, the c -plane Al2O3 substrates were cleaned by sonicating in successive baths of electronic purity grade trichloroethylene, acetone, and methanol, followed by blow -drying with compressed n i trogen. The substrates were adhered to the heater surface with silver paint that was allowed to dry under ambient conditions. The background pressure of the growth system was below 1.0 x 103 Pa The substrates were heated to the film growth temperature at a rate of ~10 C /min under 1 Pa of O2. In addition to the di rect growth of CuSc1 xMgxO2 on sapphire, the use of intermediate buffer te m plate layers to assist in nucleation was also examined. The template layers included CuCrO2 or graded CuCr2/3Sc1/3O2/CuCr1/3Sc2/3O2 multilayer s Previous work has shown that the gro wth of c axis oriented epitaxial growth of CuCrO2 on sapphire could be a c complished over a reasonably wide processing window as compared to other delafossite semiconductors.1 14 For CuCrO2 template growth, the temper ature was held at 700 C to perform a 30 minute deposition of CuCrO2 at a laser energy of 185 mJ (fluence of 0.185 J/mm2) a laser repetition rate of 5 Hz, and a growth pre s sure of 1 Pa The growth of the graded CuCr2/3Sc1/3O2/CuCr1/3Sc2/3O2 template multilayers was per formed under similar conditions After growth of the template, the O2 pressure was decreased to 0. 1 Pa the temperature was changed to that desired for CuSc1xMgxO2 growth, and a film was grown for 90 minutes at a laser repetition rate of 2 H z and laser energy density of 185 mJ /mm2. The difference in the laser repetition rate was used to keep the observed growth rate at approximately 0.5 /s. After film growth, the samples were cooled at ~20 C /min at the system base pressure of 103 Pa, large ly

PAGE 89

89 comprised of water vapor The properties of the films were examined by powder X ray diffraction (XRD), 4 -circle XRD, four -point Van der Pauw resistivity, and optical adsorp tion. Discussion Unbuffered CuSc1 xMgxO2 on c Al2O3: Temperature Dependence XRD orientation study Initial studies focused on the growth behavior of CuScO2 and CuSc0.95Mg0.05O2 directly on c -plane Al2O3. Figure 4 1 (a ) shows X -ray diffraction scans for CuSc0.95Mg0.05O2 grown at various growth te mperatures with an oxygen pressure of 0.1 Pa directly on c -plane Al2O3. For all XRD 2 plots the regions from 30 38 and 63 73 were plotted in detail since they are of the greatest utility for examining minority phases and off axis peaks. From the (00l ) peaks, the c axis lattice parameter for the films was determined to be constant at 16.96 The preferred orientation of the CuSc0.95Mg0.05O2 was the rhombohedral 3R perpendicular c -plane as demonstrated by the (006), (009) ( not shown), and (00 12 ) peaks. Unlike CuCrO2 crystal structure the CuScO2 lattice shows little to no (003) peak as its ideal peak intensity is far less than 5% of the intensity of the (006) peak as compared with 7% for the CuCrO2 system. However, for all deposition temperatures considered, additional orientations of the CuScO2 phase were evident (marked most strongly by 3R (101), 2H (101), and 3R (012) peaks ). This was expected due to the large a -lattice mismatch of 14.6% between CuScO2 (a = 3.2 2 ) and Al2O3 (a = 2.75 with the 30 rotation ). Films grown at 600 C and 700 C showed a complete absence of the Cu2O (1 11) peak and thus the Cu2O phase. Figure 4 1 (B) more clearly shows that, by comparing the ratio of respected peak intensities to the CuScO2 (006) peak, the 2H (101) and 3R (012) peaks were also of prominence in the scans with a gradual increase of the prevalence of the (012) over the 3R (101) peak intensities. The intensity ratios of the CuSc0.95Mg0.05O2 3R (101) /(006) diffraction peaks decrease linearly with an increase in deposition temperature while

PAGE 90

90 the ratio of the CuSc0.95Mg0.05O2 (012)/(006) remains relatively constant with the lowest value seen at 600C. Optical absorption The optical absorption properties of the CuSc0.95Mg0.05O2 films were measured over the wavelength range 200 900 nm. Figure 4 2 ( A ) shows the resulting Tauc plots for the various films indicating a bandgap value of 3.6 eV for the films grown at 400 C, 500 C, and 600 C. A plot of the transmittance for these same films is shown in Figure 4 2 (B) The film grown at 700 C showed a slightly lower bandgap of 3.45 eV and, open inspection, a visible cloudiness of the samples from an unknown origin. The resulting Tauc plots suggest that all growth temperatures produc ed direct band gaps of the above values due to the linear relationship between wavelength energy and E2 at the band transition. The nature of the delafossite bandgap is a topic of theoretical interest. X. Nie et al calculated that there is a much smaller non -optically active bandgap for the group III delafossites (CuMO2 where M = Al, Ga, and In) .52 A deeper look into this phenomenon was prompted by the observation that these delafossites had bandgaps that increased with atomic numbe r of the trivalent cation, with Eg = 3.5, 3.6, 3.9 eV for M = Al, Ga, and In .36, 158, 159 Interestingly, the CuScO2 system may be unique in maintaining a direct bandgap as suggested by reports from Gilliland et al with a excitonic binding energy of 100 meV .160, 1 61 162 An analysis of the plotoluminescence of CuSc1xMgxO2 is presented in Chapter 5 Seebeck study Carrier type for the deposited films was determined by measuring the Seebeck coefficient Figure 4 3 shows the room temperature Seebeck coefficient of CuSc0.95Mg0.05O2 films as a function of deposition temperature. The positive sign of the Seebeck coefficients shows that all films are p type. Due to t he low mobility nature of these films, Hall measurements could not be

PAGE 91

91 used to extract carrier density or mobility. It is notable that there is a nearly two order of magnitude difference in the Seebeck coefficient for films grown at 500 C as compare d to fi lms grown at 600 C. This may reflect a greater incorporation of Mg into the delafossite lattice (as inferred from Figure 4 7 ) or the presence of more Cu2O phase dispersed in the films but may also indicate a kinet ic hindrance of creating self compensating defects. Unbuffered CuSc1 xMgxO2 on c Al2O3: Mg Dependence T his study also examined the effects of Mg dopin g on film orientation. There is a slight shift in the c axis lattice parameter with Mg doping, with c = 1 6.99 for CuScO2 and c = 16.96 for CuSc0.95Mg0.05O2. This was expected due to the difference of diameters between Mg2+ (0.86 ) and Sc3+ (0.885 ) Films with x = 0 and 0.01 showed Cu2O (111) peak while those with a higher Mg concentration did not. Both Figure 4 1 ( A ) and Figure 4 4 ( A ) lead us to the conclusion that low temperatures and Mg concentrations lead to Cu2O formation. Figu re 4 4 (b ) shows that the I(012)/I3R(006) diffraction peak ratio remained relatively consistent across dopant levels while the I(101)/I3R(006) diffraction peak ratio is notably high for the undoped film and low for the 1% Mg film. The resulting films cope with the large lattice mismatch with c -Al2O3 with very little dilation or contraction of the (00 l ) pla nes dimensions seen most readily in the 3R (006) and (0012 ) peak positions. CuCrO2 Buffered CuSc1xMgxO2 Films : Temperature Dependence In an effort to enhance the degree of (00 l ) orientation in the deposited films, the effects of an intermediate CuCrO2 template layer on the epitaxial growth and properties of CuSc0.95Mg0.05O2 were examined. Previous work showed that highly c axis oriented CuCrO2 can be g rown on c -Al2O3 over a relatively large processing window.114 For all films, the CuCrO2 buffer layer was grown at 700 C and 1 Pa O2 for 30 minutes at 5 Hz, yielding a buffer layer of an estimated 50100 nm thick

PAGE 92

92 XRD orientation and phase composition stud y Figure 4 5 shows the XRD 2 scans for CuSc0.95Mg0.05O2 grown at 0.1 Pa O2 on this CuCrO2 template. An increase in growth temperature increased the prevalence of the CuSc0.95Mg0.05O2 (012) film orientation. As with the unbuffered films, the optimal growth region lies around 600 C and also around 750 C. The Cu2O peaks are likely due to the partial decomposition of the CuCrO2 buffer layer. Figure 4 6 shows the XRD peak intensity ratio comparison of the buffered and unbuffered films. Figure 4 6 ( A ) shows the intensity ratio of the CuScO2 3R (101), 2H (101), and (012) peaks to the peak delafossite (006) maximum values. While the (101) 3R ratio increases linearly with an increase in growth temperature t he 2H polytype (represented in the 2H (101) peak) disappears entirely at 700 C and the prevalence of the (012) orientation increases dramatically to a value of 3.76. The 750 C film shows a dramatic improvement in all respects and may be addressed by an in terfusion of Cr2+ into Sc lattice sites. This will be discussed in a proceeding section. The buffered films showed a significant decrease in the relative intensity of off -axis (012) and (101) CuScO2 peaks compared with unbuffered films. Also, buffered fil ms showed a marked decrease in the prevalence of non(00 l ) CuScO2 peaks over films grown directly on c -Al2O3 except for those films grown at 700 C which matched quite well. The similarities of phase behavior of buffered and unbuffered films at 700 C may be the result of the higher growth kinetics promoting interdiffusion of the two layers. Figure 4 6 ( b ) shows the relatively minimal occurence of the Cu2O phase for 400600 C region. It increases dramatically for th e 700 C growth condition. It has not been determined whether this is due to the decomposition of the CuCrO2 buffer layer or the CuScO2 film.

PAGE 93

93 Resistivity s tudy The resistivity of CuSc0.95Mg0.05O2 films grown on CuCrO2 buffered Al2O3 was measured. The meas urements were made using a 4 -point Van der Pauw configuration. Buffered samples grown at 400 C, 500 C, 600 C, and 700 C are shown to have resistivities displayed in Figure 4 7 The resistivity varied from appr oximately 102 to 104 ohm cm. While the resistivity reached a low value at 500 C of 73 ohm cm, this was accompanied by the strong presence of a Cu2O phase. One can not discount the possible effect of the CuxO secondary phase dispersed throughout the films on the transport measurement. CuCrO2 Buffered CuSc1xMgxO2 Films : Mg Dependence XRD orientation and phase composition study The crystallinity and orientation of CuSc1 xMgxO2 films with varying Mg content grown on the CuCrO2 buffer was examined using 2 scans. The CuSc1 xMgxO2 films were grown according to the optimal 600 C and 0.1 Pa growth conditions to yield 400 nm thick films. The issues of the Tauc plot for the 700 C growth condition have not been resolved so the 700 C and 750 C were avoided Figure 4 8 (A ) shows the X ray diffraction results for films grown with varying Mg content. The splitting of the (00 l ) peaks corresponds to the d -spacing of the CuCrO2 and CuScO2 phases. An increase in Mg alloying percentage increases the prevalence of the off axis (012) peak. Of note is a complete absence of the Cu2O (111) peak for all films. Figure 4 8 (B) shows the intensity of CuScO2 (101) and (012) peaks relative to the CuScO2 3R (006) peak. The figure shows an increase in relative (012) peak intensity with an increase in Mg percentage and a general increase in relative 3R (101) peak in tensity with the Mg percentage. The crystallinity of the CuSc1 xMgxO2 films as a funct ion of Mg content was further considered by examining rocking curves. Figure 4 9 shows a comparison of rocking curves

PAGE 94

94 for CuSc1xMgxO2 films grown on sapphire (no buffer) and on the CuCrO2 buffer. Films were grown at 600 C and 0.1 Pa O2 with different Mg concentr ations. As expected, the CuSc1xMgxO2 phase was seen to grow with better crystallinity on the CuCrO2 buffer as compared to films directly on sapphire. CuSc1xMgxO2 films grown on the CuCrO2 buffer showed rocking curve full -width half -maximum (FWHM) less th an 1 for 0 x 0.05. For films directly on sapphire, the FWHM was between 1.5 and 3.0 In order to further examine film crystallinity, the in -plane alignment of the CuScO2 films on buffered sapphire was examined using four circle X ray diffraction. 4 -Circle XRD epitaxy Figure 4 10 shows the phi scans confirming this epitaxial relationship as represented by the structure labeled sample F in Table 4 1 and Table 4 2 The growth conditions are described in these tables ( Table 4 1 and Table 4 2 ) and it should be noted that structure includes a ZnO film g rown as the structures top layer which is not included in this discussion. The Cu Sc O2 (018) phi peak at = 150 ha s a FWHM value of 3.2 Likely, for a combination of atomic and structure factor reasons, the (018) peak was more easily discerned for the sc andium delafossite than the (104) peak used for the chromium delafossite and was thus used Six peaks are easily seen and distinguishable, which confirms a twinning epitaxy for the film and a well defined alignment with the buffer layer (CuCr0.99Mg0.01O2) and a 30 rotation from the underlying c -Al2O3 substrate as predicted by the better match to the this sublattice (c = 2.75 ) No peaks were seen for a phi scan of the unbuffered films consistent with the wide rocking curves (FWHM> 2 ) and their designation as textured polycrystalline

PAGE 95

95 Use of Gradient Buffers XRD orientation study In addition to CuCrO2 templates, we also examined the use of a graded CuCr2/3Sc1/3O2/CuCr1/3Sc2/3O2 multilayer template for growin g CuSc0.95Mg0.05O2 on sapphire. By alloying CuCrO2 and CuScO2 in various proportions, the a axis lattice constant of CuCrO2 (2.99 ), which can be grown epitaxially on c -plane Al2O3 (30 rotation yields an in-plane relevant spacing of 2.75), can be graded up to that for CuScO2 (a = 3.22 ). Assuming that Vegards rule holds for the Cu(Cr,Sc)O2 solid solution, the lattice parameters for CuCr2/3Sc1/3O2 and CuCr1/3Sc2/3O2 should be 3. 07 and 3.14 respectively. The CuCr2/3Sc1/3O2/CuCr1/3Sc2/3O2 multilayer provides a graded transition from the c plane Al2O3 (a = 2.75 ) to CuScO2 (a = 3.22 ). For the CuCr2/3Sc1/3O2/CuCr1/3Sc2/3O2/CuSc0.95Mg0.05O2 multilayer, the thickness of the CuCr2/3Sc1/3O2 and CuCr1/3Sc2/3O2 layers was approximately 20 nm and the outer was 400 nm. The films were grown at 700 C and 1 Pa O2 pressure. Figure 4 11(A ) shows XRD plots of the sequence of a CuSc0.95Mg0.05O2 fil m grown on the graded buffer (a ), a CuCr1/3Sc2/3O2 film (b ), a CuCr2/3Sc1/3O2 film (c ), and a CuCrO2 film ( d ). Implementing the graded multilayer resulted in better delafossite (00 l ) texture as measured by the peak intensity of the (00l ) peaks relative to that for other orientations as seen in the inset graph within Figure 4 11(B). The nature of the phases was further examined by optical transmission measurements. The CuCr1yScyO2 films demonstrate that the y = 2/3 film (labeled film (b)) shows a low incidence of off axis peaks while film ( c) demonstrates a high incidence of the (012) delafossite peak and a broad double peak at (006). This indicates that the Cr cation is more soluble in the CuScO2 lattice than Sc is in the CuCrO2 lattice which is corroborated by the following data.

PAGE 96

96 Optical t ransmission and phase solubility Figure 4 12(A ) inset shows that the bandgap of the films increase with the proportion of Sc alloyed in the films from a low of 3.20 eV for the CuCrO2 film (a ) to the 3. 64 eV for the CuScO2 film ( d ). The intermediat e films show a bandgap of ~3.5 eV and thus seem to exhibit some saturation limit of Cr into the CuScO2 lattice for the condition of 1/3 y 2/3 It is notable that this bandgap is expressed well by equation 4 1 which pred icts a bandgap of 3.49 eV for y = 1/3 E g(Cu Sc1-yCryO2) = y Eg+Eg(CuCrO2) (4 1) The variable y represents the proportion of Cr in the lattice for the film CuSc1yCryO2 while the band gap difference( Eg) is the band gap difference between the y = 0 (CuScO2) and y = 1 (CuCrO2) films. There is an additional feature at 3.07 3.08 eV which corresponds to CuCr1 xMgxO2 (x = 0.03 and 0.05) fil ms we have report elsewhere but may here represent similar chromium rich delafossite phase .114 This feature likely represents a phase where s c andium acts as an impurity within the CuCrO2 crystal structure The other dominant phase seems to be a truly alloyed CuSc1 yCryO2 phase This accords with the idea that Cr2+ with a covalent radius of 0.87 (low spin) cation may have a good lattice match w ith the Sc3+ cation (covalent radius = 0.885 ) and may act as an effective p -type dopan t in the CuScO2 lattice. It may be conjectured that the presence of 1/3 high spin chromium atoms would be especially stable be producing a coordinated spin system, analo gous to the described intercalated delafoss ites and the stability of CuYO2.33.90 Further Seebeck and magnetic test will hopefully provide more insight into the nature of this interaction Buffered Strategy Comparison Figure 4 13( A ) shows 2 XRD scan comparisons of CuSc0.95Mg0.05O2 films grown at 600 C on the graded buffer ( a ), CuCrO2 ( b ), and directly on c plane sapphire ( c ). All films are

PAGE 97

97 400 nm in thickness. Figure 4 13( A ) shows clearly that the intensity of the CuScO2 (006) peak has a maximum for the film with the gradient buffer. As may be expected, Figure 4 13(B) shows the CuSc0.95Mg0.05O2 film grown upon successive layers of CuCr2/3Sc1/3O2 and CuCr1/3Sc2/3O2 respectively demonstrate better quality CuSc0.95Mg0.05O2 films with the most minimal out of plane CuScO2 peaks (namely (101), (012), (202), and (024)). The 700 C/750 C CuCrO2/CuScO2 strategy was not used in this comparis on. ZnO Devic e Structures and Buffer Strategies The ultimate goal of this study is to provide a template for the growth of epitaxial ZnO on CuSc1 xMgxO2. This takes advantage of the close lattice match between ZnO (a = 3.24 ) and CuScO2 (a = 3.22 ) and i n combination with a CuCr1xMgxO2 base layer opens the possibility of either creating a buffered pn junction or a pin -structure. The t able of growth techniques (Table 4 1 ) shows that while gradient buffers are effe ctive in decreasing the incidence of off axis peaks the aforementioned gradient recipes were unable to eliminate them entirely. However the growth condition of using a simple 50 100 nm CuCrO2 buffer grown at 700 C followed by the growth of CuScO2 grown a t 750 C virtually eliminates off axis peaks. Some spinel and Cu2O was detected for the sample ( Figure 4 14), the later likely being attributed to the CuCrO2 buffer layer since it is seen at 700 C growths of buffe red films ( Figure 4 5 ) but not unbuffered films ( Figure 4 1 ). The (006) peaks for both CuCrO2 and CuScO2 were examined by rocking curves to show FWHM values of 0.80 and 1.09 considerably low values for these films This was likely due to the effective anneal of the buffer film caused by the 45 minute 750 C growth The odd looking -scan ( several of the peaks were substantially reduced) seen in Figure 4 15 does confirm a twinned epitaxy of both films on c -Al2O3. The poor intensity of several of the peaks is likely due to a slightly tilted attachment of the samples to the substrate holder making the film

PAGE 98

98 surface out of parallel wi th the substrate holder surface. A slight tilting of the films may allow between lattice mismatched layers. Similar films labeled sample E and sample F in Table 4 1 and Tabl e 4 2 substantiated these results showing epitaxy with a subsequent layer of ZnO. The middle layer of CuScO2 allows seems to benefit from the longer grow th times in improving its crysta l linity The rocking curve FWHM values of CuScO2 (006) peaks for the se device structures correspond to growth times of 45 minutes, 30 minutes, 15 minutes as 1.09 1.92 and 2.12 (samples A, F, and E in Table 4 2 ). According to previous calibrations of growth rates of these mat erials these layers are approximately 200 nm, 130 nm, and 60 nm thick. Though thicker layers co uld improve the crystallinity further this work demonstrates that it may be possible to achieve reasonable quality CuScO2/ZnO junctions with further tweaking of this recipe. Figure 4 16 shows the c axis orientation of ZnO as only (002) and (004) peaks may be seen. Values (FWHM) of rocking curves for ZnO (002) peaks of samples E (grown at 400 C) and F (grown at 650 C) a re 1.16 and 1.03 respectively. These represent a substantial improvement over the graded buffer films which have ZnO rocking curve FWHM values of 2 and 7 for 600 C and 700 C growth temperatures. This presents the issue that the graded buffer films se em to provide higher quality CuScO2 while the CuScO2 grown at 750 C on a CuCrO2 buffer allows for the growth of higher quality ZnO. Conclusion The epitaxial growth of CuSc0.95Mg0.05O2 on c -sapphire is reported. We have successfully used a heteroepitaxial buffer layer to grow epitaxial thin films of CuSc0.95Mg0.05O2 on c plane sapphire. X ray diffraction results determined that the growth conditions of 600 C and 0.1 Pa O2 produce the best phase pure material while higher temperatures produce higher resist ivity and mu ch lower Seebeck coefficients. The nature of CuSc2/3Cr1/3O2 may represent a solubility limit

PAGE 99

99 and a magnetically active material. The CuSc1 xCrxO2 phase seems to show that Sc and Cr delafossites are at least partly miscible and may set the stage for acceptor doping in CuScO2 with divalent chromium at high doping levels In -plane x -scans show epitaxy for films grown on various buffer layers with a strategy of using 700 C CuCrO2 buffers followed by 750 C CuSc1xMgxO2 layer provi ding the best epitaxy of ZnO and eliminates off axis delafossite peaks but the nature of doping in this case is unestablished. With further improvement ptype CuSc0.95Mg0.05O2 should prove useful in the forma tion of oxide pn junctions and Schottky barriers for interf aces with n -type oxide films (especially ZnO) or nanowires .24, 37, 163 166

PAGE 100

100 Figure 4-1. XRD -2 data for CuSc0.95Mg0.05O2 films grown directly on c-plane Al2O3 graphed as A) a function of temperature and as B) ratios of maximum peak intensities of offaxis (012) and (101)3R peaks compared to the corresponding (006) peak for each film.

PAGE 101

101 Figure 4-2. Optical transmission data for 400 nm films of CuSc0.95Mg0.05O2 grown at 400 C, 500 C, 600 C, and 700 C directly upon c-Al2O3 plotted as A) Tauc plots showing direct bandgap behavior and as B) optical transmission plots.

PAGE 102

102 Figure 4-3. Seebeck coefficients for CuSc0.95Mg0.05O2 films grown at 400 C, 500 C, 600 C, and 700 C directly upon Al2O3.

PAGE 103

103 Figure 4-4. XRD -2 data for CuSc1-xMgxO2 films grown on c-Al2O3 for (x = 0, 0.01, 0.03, and 0.05) plotted as A) typical -2 plot and as B) ratios of maximum peak intensities of off-axis 3R(012), 2H(101), and 3R(101) peaks compared to the corresponding on-axis (006) peak intensity.

PAGE 104

104 Figure 4-5. XRD -2 scan for CuSc0.95Mg0.05O2 grown on CuCrO2 template at 400 C, 500 C, 600 C, and 700 C. The feature at 3 7.5 is an artifact from W K, not from the sample.

PAGE 105

105 Figure 4-6. Comparison for CuSc0.95Mg0.05O2 on CuCrO2 buffered substrates grown at 400 C, 500 C, 600 C, and 700 C. of ratios of maxi mum peak intensities of A) off-axis (101) 2H, (101) 3R, and (012) peaks compared to the on-axis (006) peak intensity and of B) the Cu2O(111) to CuScO2 (006) intensity.

PAGE 106

106 Figure 4-7. 4-point Van der Pauw resistiv ities of ~400 nm th ick films of CuSc0.95Mg0.05O2 on 50 nm CuCrO2 as a function of growth temperature.

PAGE 107

107 Figure 4-8. XRD -2 data for CuSc1-xMgxO2 films grown on CuCrO2 template for (x = 0, 0.01, 0.03, and 0.05) plotted as A) typical -2 plot and as B) ratios of maximum peak intensities of off-axis 3R(012) and 3R( 101) peaks compared to the corresponding onaxis (006) peak intensity.

PAGE 108

108 Figure 4-9. Full-width half-max values of -rocking curves of the (006) peaks of CuSc1-xMgxO2 thin films grown with CuCrO2 buffers and without buffer layers at 600 C and 0.1 Pa O2.

PAGE 109

109 Figure 4 10. -scans of CuScO2 thin film (Sample F from Table 4 1 ) grown with CuCr1xMgxO2 buffer Al2O3 showing CuScO2 (018), CuCrO2 (104), and Al2O3 (101) peaks

PAGE 110

110 Figure 4-11.XRD -2 data of (a) 200 nm CuSc0.95Mg0.05O2 grown on a graded 20 nm CuCr2/3Sc1/3O2/ 20 nm CuCr1/3Sc2/3O2 multilayer, (b) 200 nm CuCr1/3Sc2/3O2, (c) 200 nm CuCr2/3Sc1/3O2, and (d) a 200 nm CuCrO2 films plotted as A) a typical -2 plot and as B) ratios of maximum peak intens ities of off-axis (012) and 3R(101) peaks compared to the corresponding on-axis (006) peaks.

PAGE 111

111 Figure 4-12. Optical transm ission data for (a) CuCrO2, (b) CuCr1/3Sc2/3O2, (c) CuCr1/3Sc2/3O2, and (d) CuScO2 films plotted as A) Tauc plots showing direct bandgap character and as B) transmittance plots.

PAGE 112

112 Figure 4-13. XRD -2 data of CuSc0.95Mg0.05O2 grown on (A) CuCr2/3Sc1/3O2/ CuCr1/3Sc2/3O2 multilayer, (B) CuCrO2 buffer, and (C) directly on sapphire plotted as A) typical -2 plots and as B) ratios of maximum peak intensities of off-axis (012) and 3R(101) peaks compared to the on-axis (006) peak for CuSc0.95Mg0.05O2.

PAGE 113

113 F i gure 4 14. XRD -2 plot of data from sample A ( Table 4-1 and Table 4-2 ) showing a detailed plot of (006) peak (top) and broad data of range = 10 80 (bottom)

PAGE 114

114 Fi g ure 4 15. -scan of Sample A ( Table 4-1 and Table 4-2 ) showing CuScO2 (018) and CuCrO2 (104) peaks. Fi g ure 4 16. XRD -2 plot of d ata from S ample E (Table 4-1 and Table 4-2).

PAGE 115

115 Table 4 6 Compilation of various device constructs and their growth conditi ons Buffer CuSc 1 x Mg x O 2 ZnO:Al 0.01 Sample Buffer Type Temp ( C) P(O 2 ) (Pa ) Time (min) Type Temp ( C) P(O 2 ) ( Pa ) Time (min) Film Temp ( C) P(O 2 ) ( Pa ) Time (min) A CuCr 0.99 Mg 0.01 O 2 700 1 60 (5 Hz) x=0.00 750 4x10 2 45 (2 Hz) NA B Gradient 7 00 4x10 2 10 (2 Hz) x=0.05 600 4x10 2 90 (2 Hz) yes 400 5x10 3 60 (3 Hz) C Gradient 700 4x10 2 10 (2 Hz) x=0.05 7 00 4x10 2 95 (2 Hz) yes 400 5x10 3 60 (3 Hz) D CuCr 0.95 Mg 0.05 O 2 700 1 60 (5 Hz) x=0.05 700 4x10 2 10 (2 Hz) yes 650 1x10 3 60 (1 Hz) E CuC rO 2 700 1 60 (5 Hz) x=0.0 5 750 4x10 2 15 (2 Hz) yes 400 5x10 3 60 (3 Hz) F CuCr 0.99 Mg 0.01 O 2 700 1 60 (5 Hz) x=0.00 750 3x10 2 30 (2 Hz) yes 650 1x10 3 30 (2 Hz) Gradient refers to a graded growth of CuCr2/3Sc1/3O2/CuCr1/3Sc2/3O2 buffer grown each in s uccession for 10 minutes and 2 Hz each under the assigned growth temperatures and O 2 pressures. Table 4 7 Compilation of -scan and rocking curve data for device constructs Scan Proper t i es Rocking Curve Cr Sc ZnO Cr Sc ZnO Sample FWHM (deg) FWHM (deg) FWHM (deg) FWHM (deg) FWHM (deg) FWHM (deg) A 0.8 1.2 NA 0.80 1.09 NA B NA 2.5 10 NA 1.72 7 C NA 1.7 1.2 NA 0.94 2 D 1 NA 1.8 1.05 NA 1.35 E 1.9 1.9 1.93 2.12 1.16 F 1 1.3 1 1.76 1.92 1.03

PAGE 116

116 CHAPTER 5 LUMINESCENCE AND OPTICAL PROPERTIES OF C u S c1 XM gXO2 THIN FILMS GROWN BY PULSED LASER DEPOSITION Introduction Transparent conducting oxides have garnered attention due their unusual combination of optical transparency and high conductivity among the functional oxide materials .1 Though the literature describes a number of wide bandgap semiconducting oxides that can be doped ntype few have been doped p-type. For thin -film elect ronics, there is a large need to develop p-type transparent semiconductors fo r field -effect transistors, sensors, and especially blue/UV pn junction LEDs .37 129, 132, 163 The few metal oxides that demonstrate room temperature p type conduction (e.g. NiO and Cu2O) have disappointingly small bandgaps and /or low conductivities.155, 156 The delafossites composed of palladium and platinum act as metals while those composed of silver or copper typically act as semiconductors. In the Cu -based delafossites, the Cu atoms 3dz 2 orbital is hybridized by its nearest neighbor anisotropy (2 oxygen atoms and 6 copper atoms) between 3dz 2+4s and 3dz 24s bands (see Figure 5 1 ). Both Pt and Pd based delafossites act as metals precis ely because their Fermi energies lie within the 4dz 25s (5dz 26s) band. A direct transition can take place between the hybridized oxygen 2sp3 band and the 3dz 2+4s band with filled 3dxy3dy 2 z 2, 3dxz3dyz and 3dz 24s bands that would otherwise cause optical a bsorption. Cu and Ag systems both have Fermi energies lying within the band gap between 3dz 2+4s and 3dz 24s and thus b ehave as semiconductors with an indirect transition between these bands. Given the p-type properties of these materials, there is significant interest in the thin film growth of delafossite semiconductors. There is a recent report of a LED construction made with CuAlO2/ZnO/ZnO nanowires with the ZnO thin film acting as t he active layer with both a stro ng

PAGE 117

117 band edge and green EL electrolumin esence .190 In this paper, optical and structural properties of CuSc1 xMgxO2 (x = 0, 0.01, 0.03, 0.05) thin films are investigated to show that they seem to demonstrate direct bandgap behavior. This finding is especially surprising considering that delafoss ites have been known more recently for their non -optically active indirect bandgaps. Though they are commonly known as optically transparent materials X. Nie explored two major issues with this designation thereby revealing a hidden indirect bandgap .52 Firstly, it is experimentally observed that the bandgaps of CuMIIIO2 increase with atomic number of the trivalent cation, with Eg = 3.5, 3.6, 3.9 eV for M = Al, Ga, and In, respectively .108, 158, 159 This finding was unexpected considering the analogous III -V semiconductor series shows the opposite trend due largely to the greater ionic character of the compounds Secondly, the species with t he largest bandgap, namely CuInO2, is the only member shown to be amenable to both p and n type doping. First principle calculations predict that there should be a decreasing indirect, non -optically active bandgap with an increase in MIII site atomic numb er which contributes to both effects .52 This new set of bandgaps corresponds to popular reasoning as it scales against increasing atomic number down the group III elements, Al, Ga, and In where Eg = 1.97, 0.95, and 0.41 eV. Interestingly, the CuMIIIO2 system (M = Sc, Y, and La) may be unique in maintaining a true direct bandgap as suggested by reports from Gilliland et al .160, 161 This could open doors for a first generation of delafossite LEDs. Experimental Procedure Films of CuSc1xMgxO2 films were grown by pulsed laser deposition (PLD). The targets were made by massing, grinding, and pressing the appropriate mixtures of 99.99% Sc2O3, 99.999%Cu2O, and 99.998% MgO on a metals basis. The pressed substrates were then sintered at 1300 C for 12 hours. The substrates were cleaned through a series of sonicated baths of tri -

PAGE 118

118 chloroethylene, acetone, and methanol respectively for 5 minut es each. The substrates were then blown dry. Next, the substrates were adhered to the heater surface using silver paint and allowed to dry under ambient conditions. Upon loading the sample into the growth chamber and allowing the base pressure to fall belo w 103 Pa the substrate was heated at 10 C/minute to 600 C under a 0. 1 Pa O2 pressure. The dominant background gas was found to be water vapor. Growth was carried out by ablating the target at 2 Hz 180 mJ /mm2 for 90 minutes yielding 400 nm thick films. The samples were quenched under vacuum. Discussion Crystallinity Study: XRD The samples were initially examined using X -ray diffraction to determine phase purity and alignment Figure 5 2 (A ) demonstrates that for a ll Mg percentages highly textured c -oriented films resulted. Though there was a predominance of c -plane orientation Figure 5 2 (B) shows that ratio of offaxis 3R (101), 2H (101), and 3R (012) peaks also were presen t and varied greatly in their intensity between growth runs. A small Cu2O (111) peak is seen in both 0% Mg and 1% Mg runs. Notably, t he 1% Mg film showed the lowest relative intensities of all off axis peaks. Figure 5 3 gives the relative FWHM (full width half max) of rocking curves for these films which range in their FWHM intensities from 1.646 to 2.710 for the CuSc0.99Mg0.01O2 and CuSc0.97Mg0.03O2 films. Subsequent -scans revealed no rotation symmetry leading to the determination that our films were highly textu red polycrystalline. Optical Transmission Properties Next, the optical absorption properties of the CuSc1xMgxO2 films were measured over the wavelength range 200900 nm. Figure 5 4 shows the resulting Tauc plots w ith inset transmission plots for the sequence of films. All films had a bandgap of between 3.563.60 eV 2 plots show the bandgap

PAGE 119

119 types to be direct. The relative consistency of the bandgaps between films shows the aforemention ed oxygen 2sp3 3dz 2+4s tra nsition to be largely responsible for the optical bandgap of this material regardless of the bonding state of the Mg within the material. Notably, the CuSc0.99Mg0.01O2 film ( Figure 5 4 (B)) revealed two sub -Eg adsor ptions at 3.21 eV and 3.53 eV while the CuSc0.97Mg0.03O2 film ( Figure 5 4 (C )) hold one adsorption at 3.46 eV. Neither the undoped n or 5% Mg film showed any unique absorptions within their bandgaps. Through the use 2 which is evidence for a direct bandgap. This, paired with the high optical transparencies in the visible spectrum, makes a strong case for these delafossites being direct bandgap m aterials. Photoluminescence Study Comparison of optical transmission and PL data Further an alysis using He = 325 nm for photoluminescence (PL) revealed the fairly surprising occasion of luminescence ( Figure 5 5 ) from the material CuSc1 xMgxO2 for x = 0.00, 0.01, and 0.03 but not x = 0.05. Though no papers discuss luminescence in the CuScO2 system a relatively small number of papers have described the PL activity of delafossite materials (CuYO2 and CuL aO2).167 172 Interestingly, these papers detail an active -center photoluminescence from CuYO2 and CuL aO2, both of which are composed of MIII column Ib dblock transition metals as is CuScO2. However, CuScO2 is known to favor the 3R as opposed to the 2H favored by both CuYO2 and CuLaO2. The papers show luminescence from doping with group II cations (as an acceptor dopant) and/or rare earth cations such as Eu, Tb, or Tm. The latter take advantage of the wide band gaps of the parent material to create a phosphor. Other delafossites have been explored and have failed to produce photoluminescence presumably due to the previously mentioned smaller non -optically active indirect band gaps. The selection rules

PAGE 120

120 governing this behavior deny same parity states from the adsorption of light thus leaving the films optically transparent. Doping dependence of PL activity a t 15 K and 300 K Figure 5 6 provides plots of photoluminescent intensity for each film at both 15 K and 300K. These plots show luminescent features for each film except the x = 0.05 film for which no PL was detecte d Using a log plot for Figure 5 6 (B) shows the overlap in the luminescent intensity at 300 K between x = 0.00 and x = 0.01 CuSc1xMgxO2. There is sharp fall off in luminescent intensity at x > 0.01 with a complete absence of luminescence at x = 0.05. This is likely due to self quenching effects which cause self -excitation and the internal reabsorbance of luminescent light. An overview of the temperature dependence of the PL profiles is provided in Figure 5 7 Major transitions Also, Figure 5 6 (Table s 5 1 4) shows essentially two easily discernible luminescent feature labeled as and The luminescent transition is centered around 3.35 eV at 300 K and 3.2 eV around 15K. The luminescent transiti on is centered near 2.6 eV at 300 K and 15 K. Two other peaks were found by performing a multi peak Gaussian fit to the sets of raw PL data. Thes e peaks were iterated until they yielded R2 > 0.99. The b asic fitting method used the standard Gaussian equation (Equation 5 1 ). y = y0 + A/(w ( /2)) e2(x x 0 ) 2 /w 2 (5 1) The variable A is the area under the curve, w is the 2 value of the luminescent wavel ength in nm which represents approximately 0.849 of the FWHM (full width half max) value, x is the luminescent wavelength in nm, and y0 is the baseline intensity offset in arbitrary units.

PAGE 121

121 In broad terms, the peaks shifted to higher energy values with a rise in temperature while their peak intensities decreased. While the peak locations at 300 K for x = 0.00, 0.01, and 0.03 were 3.33 eV, 3.37 eV, and 3.35 eV at 300 K which represented energy increases from the 15 K measurements of 0.13 eV, 0.18 eV, and 0. 06 eV. Opposing trends for the intensities of the peaks were found as well. At 300 K the intensity of the transmission increased with a corresponding increasing in Mg concentration. The opposite trend is seen for the 15 K measurements. Basically, lattice vibrations that increase with an increase in te mperature seem to enhance the x = 0.03 case while decreasing the luminescent intensity by 93.6% and 77.7% for the x = 0.00 and x = 0.01 cases. This may be merely due to the greater abundance of thermally activated acceptors in the x = 0.03 sample or perhaps either a site dependent saturation which results around x = 0.02. It should be noted that our in house attempts at Seebeck measurements yielded extremely small (S < 1.0 V/K) values regardless of Mg concentration leaving the location of the Mg in the lattice difficult to discern As a substitutional in the Sc lattice site it forms an acceptor expressed b y MgSc In the Cu site it can form as a donor of the form MgCu Though it is widely seen that the Cu position poorly accepts deviation from stoichiometry, it is precisely this phenomenon which controls the conductivity in CuAlO2, through the defect complex [AlCu 2Oi ].47, 78 Figure 5 5 seems to show that the luminescence is related to a band edge transition. At 300 K the luminescent peak center for all doping levels is within 0.25 eV of the direct bandgap measured in these films. This corresponds approximately with the literature value of 95 m eV and 160 meV for the activation energy of holes in CuSc0.95Mg0.05O2.68, 1 47 Our own plot may be seen in Figure 5 8 (A) which shows an Arrhenius dependence fits well at temperatures approaching room temperature (EA = 160 meV) and Figure 5 8 (B) shows a transition at ~ 220 K of the small polaron behavior of the film, similar to the previously described

PAGE 122

122 CuCr0.99Mg0.01O2 film in Chapter 4 A quick exami nation of the plots and the FWHM values of the peak width, labeled E in Table 5 1 Table 5 2 Table 5 3 and Table 5 4 shows that the onset of luminescence begins very close to the band edge. The position is, like the transmission, ubiquitous for all doping levels both at 15 K and 300 K. It is also the strongest source of luminescent li ght, known as sky blue in color, emis sion At 15 K a trend in the Y and A values may be observed. The intensity of both values decreases as the doping level increases with a precipitous decline seen between x = 0.01 and x = 0.03 samples. At 300 K the x = 0.01 sample literally shines by having a luminescent area and intensity an order of magnitude above the other two samples. Its peak position of 2.73 eV at 300 K exceeds the other positions by at least 0.12 eV. The x = 0.00, 0.01, and 0.03 samples each show a decrease of intensity between 15 K and 300 K of 98.7%, 88.3%, and 39.2% respectively. The positions of the peaks between 15 K and 300 K states has a positive shift of 0.03 eV, 0.15 eV, and 0.06 eV for x = 0.00, x = 0.01, and x = 0.03 respectively. The nature of th e transmission is more speculatively deduced than the transmission. Because it is seen for the x = 0.00 sample it must come from some feature in the native lattice and not directly dependent on the presence of the Mg. It should be noted that these are polycrystalline samples which are c axis textured implying a high density of grain boundaries. Qualitatively it has been noticed that these samples show a decrease in resistivity with time under desiccant storage without any detectable change in phase or l attice dimensions. This has become so pronounced that polycrystalline samples measured after a number of months are less resistive than their corresponding counterparts in buffered epitaxial CuSc1 xMgxO2 films. Figu re 5 9 shows two effects of doping. First, the doping effectively lowers the resistivity of doped samples. Secondly, the single crystal samples labeled as buffered show a higher resistivity for

PAGE 123

123 all doping concentrations. Freshly grown unbuffered, polycr ystalline samples of CuSc0.95Mg0.05O2 are too resistive to measure. For these reasons the presence of the luminescence may be ascribed to a gradual room temperature oxygen intercalation which obeys the chemical equation 5 2 2Cu1+ +1/2 O2 2Cu2+ + Oi (5 2) This has been observed by our measurements most demonstrably by resistivity values of polycrystalline thin films exceeding those of a higher crystalline quality. Figure 5 9 shows both this phenomenon for samples measured after three months of storage of films in a desiccant environment and demonstrates that Mg doping decreases resistivity for all doping levels. Other authors have described the effects of Ca and Sr doping for both CuLaO2 and CuYO2.111, 173 CuLa1xSrxO2 showed a yellow -green lumine scent peak centered at 575 nm (at 300 K) with a slight negative shift of 10 nm between x = 0.01 and x = 0.00 states, and 625 nm (at 77 K) with a larger negative shift of 50 nm for x = 0.00 and x = 0.01 states. The delafossite CuLa1 xCaxO2 showed a similar luminescent peak centered around 580 nm (at 300 K) with very little location dependence on x and 620 nm (at 77 K) with 25 nm shift between x = 0.00 and x = 0.01 doping levels.111 It seems it can be accurately sta ted th e location of this luminescent feature depends far more on the parent lattice composition rather than the choice of dopant. Another report shows CuY1xCaxO2 produced a photo-luminescent peak centered around 510 nm, 520 nm, and 550 nm for x = 0.05, 0. 03, and 0.00. This study used a larger range of doping concentrations than the equivalent CuLa1xCaxO2 samples but seemed to show a large shift in peak location of 40 nm at 300 K. This native Y -delafossite peak position is a 30 nm negatively shifted in com parison to the native La -delafossite peak position. This would imply a further negative shift

PAGE 124

124 for CuScO2 over the peak position for CuYO2 as seen for 300 K (455 nm to 476 nm) and 15 K (480 nm to 489 nm). Minor transitions The and emission are even m ore difficultly assigned. Firstly, the peak is seen for only for doped samples, at 300 K for x = 0.01 and x = 0.03, and at 15 K for x = 0.03. It thus appears that this peak is induced by the presence of Mg within the CuScO2 lattice and is increased in in tensity with thermal activation. Both the area, A, and the intensity, Y, for CuSc0.97Mg0.03O2 increase with a rise in temperature while the position is lowered in energy by 0.21 eV (3.14 eV at 15 K 2.93 eV at 300 K). This seems to correspond well to a therma lly activated activation of Mg a s a dopant. The emission seen for all three samples at various temperatures is most prominent at 15 K and is centered around 2.47 eV for x = 0.00 and x = 0.01 samples while the 300 K condition reliable only shows a emission for the x = 0.03 sample. The emission for the 300 K condition for x = 0.00 has an E value of 2.16 eV and may represent a completely different luminescent process. Regardless of the doping concentration the data for both I and A show an order of magnitude lower value at 300 K than for the 15 K values. Also there appears to be no shift in the peak E between 15 K and 300 K conditions. The scope of the possible causes includes lattice defects, surface states, and grain bou ndary species. More tes ting of epitaxial film s should refine the cause but is not discernable at this time Summary of PL Effects We have provided substantial evidence for band edge PL emission in CuSc1 xMgxO2 films with 0.00 x 0.03 while x = 0.05 films showed a complete la ck of PL emission. This seems to confirm previous reports that CuScO2:Mg has a direct bandgap which could be utilized

PAGE 125

125 in LED devices as an active emission layer. Also present were substantial contributions of a ~477 nm, sky blue, emission from a Cu1+ Cu2+ transition likely due to oxygen intercalation in the lattice. This effect may offer further opportunities as CuScO2:Mg use as a phosphor or EL (electroluminescent) active layer for further lighting applications. Further work should focus on determining whe ther EL can be achieved with an appropriate device structure with either nZnO or an n delafossite such as AgInO2:Sn.

PAGE 126

126 Figure 5 1 Proposed f lat band diagram for AMO2. Shown are the proposed indirect (EI) and direct (ED) energy transitions Reprinted an d modified with permission from Figure 4 of R. D. Shannon, D. B. Rogers, C. T. Prewitt, and J. L. Gillson, Inorg. Chem. 10, 725 (1971).

PAGE 127

12 Figure 5-2. XRD -2 data shown for highly textured polycrystalline CuSc1-xMgxO2 grown on cAl2O3 as A) -2 plots and as B) the resulting peak ratios to quantify relative presence of off-axis peaks( ) 3R(101)/(006), ( ) 2H(101)/(006), and ( ) 3R(012)/(006).

PAGE 128

12 Figure 5-3. -Rocking curve FWHM (full-width at half max) values for CuSc1-xMgxO2 grown directly on c-Al2O3.

PAGE 129

1 Figure 5-4. Tauc plots with corresp onding transmission plots inset for CuSc1-xMgxO2 A) x = 0.00, B) x = 0.01, C) x = 0.03, and D) x = 0.05.

PAGE 130

13 Figure 5-5. Tauc data (right axis) and PL (photo-luminescentleft axis) values at 15 K and 300 K for CuSc1-xMgxO2 for A) x = 0.00, B) x = 0.01, C) x = 0.03, and D) x = 0.05. The values labeled ( ) are UV/vis-absorbance values while ( ) are PL values measured at 15 K and ( ) are PL values measured at 300 K.

PAGE 131

13 Figure 5-6. Compilation of PL data for CuSc1-xMgxO2 films for A) 15 K and B) 300 K measurements for x = 0.00, x = 0.01, x = 0.03, and x = 0.05

PAGE 132

13 Figure 5-7. PL scans taken at various temperatures for CuSc0.99Mg0.01O2 400 nm film grown on c-Al2O3.

PAGE 133

13 Figure 5-8. Conductivity versus temperature curves for CuSc0.95Mg0.05O2 film showing A) Arrhenius dependence (detail inse t) and B) small polaron plot.

PAGE 134

13 Figure 5-9. Resistivity values for CuSc1-xMgxO2 values for ( ) textured polycrystalline and ( ) buffered, epitaxial 400 nm thin films grown on c-Al2O3 determined by 4-point Van der Pauw measurements

PAGE 135

135 Table 5 1 PL fitting data for CuSc1xMgxO2 taken at 300 K with and features. x for CuSc 1 x Mg x O 2 E (eV) E (eV) A (eV) Y (arb.) E (eV) E (eV) A (eV) Y (arb.) 0.00 3.33 0.56 0.0073 0.0141 0 .01 3.37 0.26 0.0102 0.0411 2.93 0.32 0.1283 0.3708 0.03 3.35 0.32 0.0163 0.0510 2.93 0.39 0.0510 0.1220 Table 5 2 PL fitting data for CuSc1xMgxO2 taken at 300 K with and features. x for CuSc 1 x Mg x O 2 E (eV) E (eV) A (eV) Y (arb.) E ( eV) E (eV) A (eV) Y (arb.) 0.00 2.61 0.44 0.0152 0.0330 2.16 0.59 0.0101 0.0166 0.01 2.73 0.47 0.1297 0.2560 0.03 2.60 0.31 0.0090 0.0292 2.47 0.61 0.0226 0.0350 Table 5 3 PL fitting data for CuSc1xMgxO2 taken at 15 K with and features. x for CuSc 1 x Mg x O 2 E (eV) E (eV) A (eV) Y (arb.) E (eV) E (eV) A (eV) Y (arb.) 0.00 3.20 0.35 0.0838 0.2200 0.01 3.19 0.37 0.0735 0.1840 0.03 3.29 0.23 0.0069 0.0330 3.14 0.39 0.0221 0.0550 Table 5 4 PL fitting data for CuS c1xMgxO2 taken at 15 K with and features. x for CuSc 1 x Mg x O 2 E (eV) E (eV) A (eV) Y (arb.) E (eV) E (eV) A (eV) Y (arb.) 0.00 2.58 0.24 0.6749 2.6170 2.44 0.27 0.3218 1.1100 0.01 2.58 0.22 0.5360 2.1800 2.46 0.29 0.3532 1.1300 0.03 2.54 0.55 0.0283 0.0480

PAGE 136

136 CHAPTER 6 FUNCTIONALIZING Z n AND O T ERMINATED Z n O WITH THIOLS Background ZnO is a II -VI compound semiconductor that has been extensively investigated for chemical sensor application.174176 In recent years, research activities in chemical sensors have increasingly considered the detection of specific bio-molecules for biological and/or medical applications. ZnO and Zn1xMgxO alloys have properties that appear attractive for biologically related sensors. Zn and Mg elements are essential minerals for neurotransmitter production and enzyme functioning. Toxicology studies indicate that ZnO presents minimal health hazard to h umans, thus opening the possibility of in vivo biomedical device applications.166 In aqueous environments, ZnO based devices have already been shown to be effective as pH sensors with high sensitivity and resolu tion.177 Similar opportunities might be anticipated for using ZnO and related alloys as a sensor for biomolecules. Possible Biosensor Applications Research on the use of ZnO as a biosensor has been limited to only a few repor ts involving device constructs. ZnO -based biosensors employing enzyme -modified field effect transistors have been investigated. Through chemical bonding on the ZnO surfaces, in particular the adsorption of organic molecules for the immobilization of enzymes, the device characteristics of a field effect transistor can be modified via the adsorption of selected biomolecules .178, 179 In addition, the electrochemical reaction of hemoglobin with flavin adenine dinucleotide -modified zinc oxide self assembled films has been explored for hemoglobin sensing applications.180 Biomolecule detection may also be realized by combining ZnO -based sensors with surface a coustic wave (SAW) devices.181 E nhancement of DNA immobilization has been reported using ZnO nanostructures on a lithium niobate SAW device.182 In this case, the ZnO nanotip array

PAGE 137

137 served to immobilize the bioreceptor, and the SAW device was used as the transducer to detect the mass loading effect resulting from the analyte receptor reaction. For ZnO based biosensor applications, a key aspect is the functionalization of the ZnO surf ace with selected molecular species. Biomolecular immobilization on the ZnO film surface has been reported for a few systems.183 In one example, a two -step functionalization was used that included binding of menthyl hexyl phosphonate through an amide bond formation, between an amine group and the carboxylic group in menthyl hexyl phosphonate on ZnO epilayers. Surface functionalization was monitored by XPS. In some cases, the ZnO -molecule interactions appear dependent on the specific crystalline surfaces. Thiol Chemistry In this chapter, we report on the interaction of thiols with Zn a nd O terminated ZnO surfaces. The use of normal thiols to functionalize surfaces has been employed for noble metals, most notably gold.184 185 Because of golds otherwise chemical inertness, the atomic gold -sulfur attraction has been exploited as a common template for studying surface chemistry. By using long, straight chain ( C>11) thiols, an ordered SAM ( S elf -A ssembled M onolayer ) can form. Further changing the terminal methyl groups on the nalkanethiols to another -functional group, the surface properties of the SAM may be altered. Commonly, amine groups, a carboxylic acid g roups, and a hydroxyl groups have all been used to replace the methyl groups from the n alkanethiol. Other functionalities could be applied. Thus, these -functionalized thiols can provide a worthwhile pathway for the development of low -temperature surface coatings, biomolecule manipulators, and selective chemical sensors. Multiple strategies have been previously explored to functionalize the ZnO surface. However, two problems commonly arise: the short chains have little ability to form any alkyl chain su rface crystallinity (limiting their enthalpic

PAGE 138

138 drive toward monolayer formation) and the functional head groups may self -polymerize through an SN 2 substitution. The first problem leads to low surface density and thus a lower sensitivity. The second problem creates multiple chemical layers, thereby decreasing the possibility of a charge transfer event being realized directly into ZnO where it may be a recorded event useful in surface sensing. Our efforts have therefore focused upon using long chain thiol c hemistry to functionalize ZnO, aimed at forming structures analogous to those seen with thiol addition to gold. We have examined the absorption of dodecanethiol on both oxygenand zinc terminated ZnO surfaces. Experimental Procedure Both zinc and oxygenterminated ZnO substrates were purchased from Cermet, Inc. Dodecanethiol (98%) and isopropanol (electronic grade) were purchased from Sigma Aldrich and used without further processing. One pair of substrates was examined with X ray Photoelectron Spectrosco py (XPS) both as -received, after argon sputtering, and then after exposure to a 10 mM dodecanethiol/isoproponal solution. Both the zinc and oxygen-terminated ZnO substrates were suspended in the solution for one week. Afterward, they were rinsed with isop ropanol and blown dry with high -purity nitrogen. The treated substrates were examined with XPS, then subjected to various thermal cycles to determine the stability of the functionalized surface. For each thermal cycle, the samples were heated to the desire d temperature in a tube furnace under a high -purity argon ambient for one hour, then slowly cooled. They were then examined using XPS, removed from the XPS system, and heated to the next temperature. Temperatures used progressed in 50 C increments from 50 C to 500 C. Results Our analysis focused on the intensity and position of the XPS binding energy peaks for Zn 2p3/2, O 1s, C 1s, and S 2p3/2 for both zinc and oxygenterminated substrates at each

PAGE 139

139 temperature. The C 1s peaks were used to calibrate the other binding energy positions. Figure 6 1 show s peak intensities for a pair of as -received substrates and substrates after Ar sputtering; Figure 6 2 show treated substrate s after a 50 C heat treatment and those same substrates after their final 500 C heat treatment. Native Surface Chemistry of ZnO Both zinc and oxygen terminated substrates showed order of magnitude Zn 2p3/2 and large O 1s ( b) ) peak intensity differences b etween their virgin and sputtered states (Figure 6 1 ). By sputtering the virgin surface, the 531 eV binding energy peak (O 1s peak ( )) is revealed to represent the bulk substrate oxygen while the higher energy O 1s ( ), binding energy ~533 eV, represents adsorbed species, namely O H. This binding energy difference is due to the higher binding energy arising from the higher ionizatio n of the adsorbed oxygen species. Effects of Heat Treatment of Thiol on ZnO Surface For both the Zn -terminated and O terminated surfaces, the O 1s peak shifted to higher energy when the thiol treated samples were heated to 500 C as seen Figure 6 2 This is constituted with desorption of a surface oxygen species. It is notably that the S 2p3/2 peak is reduced in intensity but does not completely disappear upon annealing, indicating a high degree of thermal stability for sulfur -substrate bonding. Temperature profiles: O species desorption Figure 6 3 shows the plots of the adsorbed and bulk oxide peaks over the temperature range 50 C to 500 C. The gradual decrease of the ads orbed oxygen species and the growth of the bulk oxygen peak clearly show the degradation of the surface species with increasing temperature. Figure 6 4 shows how Gaussian fits were assembled to match the raw data. The integrated area was used to assemble Figure 6 5 ( A ) to compare relative amounts of adsorbed O species against the bulk intensities. More specifically, both substrates show a strong increase in

PAGE 140

140 bulk oxygen peak i ntensity from 350 C to 400 C. This is also mirrored in the plot of O 1s peak ratios seen in the bottom of Figure 6 5 (B) Both the oxygen and zinc surfaces show a marked increase in the ratio of bulk to adsorbed O 1s peaks bet ween 350 C and 400 C. Temperature profiles: t hiol desorption The XPS spectra of the S 2p3/2 peak taken after sequential anneals are shown in Figure 6 6 Sulfur adsorption is remarkably stable with thermal treatment. The Zn terminated substrate shows a significant decrease in S 2p3/2 peak intensity between 400 C and 450 C while the O terminated substrate show a strong downturn in S 2p3/2 peak intensity between 350 C a nd 400 C as seen in Figure 6 7 Also, Figure 6 6 indicate s that both substrates show some sulfur remaining on the surface at 500 C. It should also be noted that the S 2p3/2 peak position at ~164163 e V represents the more reduced form of sulfur as the thiol moiety while the higher energy 170 eV peak represents an oxidized sulfur moiety. In our experiments we find a minimal amount of sulfur oxidation with an extremely small peak, seen in Figure 6 6 (B) for the oxygen surface, which is almost completely undetectable upon heating to 400 C. This implies a good resistance to oxygen degradation contrary to similar thiol -gold systems.186 It also implies that the adsorb ed oxygen and thiol species are not reactive. Nature of Thiol Adsorption Two conclusions can also be drawn from an analysis of Figure 6 7 First, the thiol surface coverage is higher on the Zn-terminated surface. T ypical intensities of the S 2p3/2 peaks on the Zn terminated surface before their decline after 400 C are higher than peak intensities on the O terminated surface before its more gradual decline after 350 C. This difference corresponds to a higher surfac e coverage for the thiol on the zinc surface. Secondly, the thiol binds more strongly to the zinc surface. The higher temperature of thiol degradation from the zinc surface shows a stronger enthalpic bond with the zinc terminated surface. The relative abse nce of the S 2p3/2 peak

PAGE 141

141 at 170 eV shows that the Zn-S is the dominant bond in this absorption event. It is unclear how the thiol absorbs onto the oxygen terminated surface but it also appears to be dominated by the lower binding energy sulfur moiety at 164 eV. A recent fluorescence study indicated adsorption of an amine on the Zn -terminated ZnO surface with a subsequent reaction of PDA (polydiacetylene) on the amine.175 Previous X ray photoelectron spectroscopy studies for octanethiol absorption on ZnO show weak absorption of sulfonate and thiolate groups that is dependent on crystal orientation.187 The interaction of pyridine (C5H5N) with O and Zn terminated ZnO surfaces was been investigated using XPS, thermal desorption spectroscopy, and x ray absorption spectroscopy. The binding energy of pyridine was twice as high on the Znterminated ZnO (0001) surface as compared to the O -terminated ZnO (000-1).188 The attachment of 1 -propanethiol (CH3CH2CH2SH) was examined for the O -terminated ZnO (0001) surface. F ormation of S O bonds between the 1-propanethiol and ZnO was seen in XPS measurements. The propanethiol/ZnO structures were stable at temperature as high as 400 C, although surface coverage was only a few percent.178 Crystallinity of Thiol Adsorption To f urther characterize the thiol absorption we performed Reflection H igh E nergy E lectron D iffraction (RHEED) measurements on both treated substrates. The Zn -terminated and O terminated substrates show equivalent surface crystallization in their untreated stat e as seen in Figure 6 8 After thiol exposure and thermal treatment of the substrates to 250 C, the zinc terminated substrate revealed an amorphous pattern. However, the thiol functionalized oxygen terminated subs trate showed some degree of surface ord er even after thiol treatment. This suggests that, though the surface coverage is lower, the oxygen surface absorbs the thiol in a more crystalline fashion with the zinc surface being more amorphous.

PAGE 142

142 Summary We invest igated the adsorption of dodecanethiol on the surface of both zinc and oxygen terminated ZnO. A strong enthalpic adsorption is demonstrated by the stability of sulfur on both ZnO surfaces up to 350 C and 400 C for oxygen terminated and zinc terminated su bstrates, respectively. The minimal presence of the S 2p3/2 170 eV peak indicates absorption of the sulfur as an unoxidized thiol. The results also show the higher surface coverage of the thiol on the zinc terminated surface. Evidence for surface crystalli nity after thiol treatment of the oxygen terminated substrate suggests that dodecanethiol can adsorb in a highly crystalline manner.

PAGE 143

14 Figure 6-1. XPS plots of the Zn 2p3/2 peak at 1023 eV and the O 1s peak(s) at 534 eV and 531.5 eV for both A) O-terminated and B) Zn-ter minated substrates as received and after argon sputtering. (A) (B)

PAGE 144

14 Figure 6-2. XPS plots of the Zn 2p3/2 peak at 1023 eV, the O 1s peak(s) at 534 eV and 531.5 eV, and the S 2p3/2 peak at for both A) O-terminated and B) Zn-terminated substrates after 50 C and 500 C heat treatments. (A) (B)

PAGE 145

14 Figure 6-3. XPS plots of O 1s peak(s) for A) Oterminated and B) Zn-terminated substrates. (A) (B)

PAGE 146

14 Figure 6-4. XPS O 1s peak(s) Gaussi an fit for peak separation of O-terminated substrate at 350 C. The Absorbed peak represents adso rbed O species (integrated area 5356) and the Bulk peak represents the bulk O (integrated area is 10766).

PAGE 147

14 Figure 6-5. O 1s integrated peak areas for adsorbed and bulk oxygen species for both Oterminated and Zn-terminated substrates.

PAGE 148

14 Figure 6-6. XPS plots of S 2p3/2 peak(s) for A) O-terminated and B) Zn-terminated substrates. (A) (B)

PAGE 149

1 Figure 6-7. Comparison of S 2p3/2 peak maximum intensities for O-terminated and Zn-terminated surfaces.

PAGE 150

150 Figure 6 8 RHEED patterns for Zn termin ated and O -terminated substrates after 250 C temperature treatment.

PAGE 151

151 APPENDIX XRD STRUCTURE DATA FOR RELATED CRYSTAL SYSTEMS Table A 1 A tomic parameters for CuCr2O4. System Space Group Lattice () Atomic Coordinates Tetragonal I4 1 /amd a b c x y z (141) 6.0316 6.0316 7.7837 Cu 0 0.25 0.125 Cr 0 0 0 90 90 90 O 0 0.5344 0.2555 Table A 2. XRD peak positions for CuCr2O4. Intensity h k l d () 18.606 11 ( 1 0 1 ) 4.765 29.572 23 ( 2 0 0 ) 3.018 31.064 34 ( 1 1 2 ) 2.876 35.159 100 ( 2 1 1 ) 2.550 37.692 67 ( 2 0 2 ) 2.384 37.692 67 ( 1 0 3 ) 2.384 42.317 20 ( 2 2 0 ) 2.134 46.628 8 ( 3 0 1 ) 1.946 46.628 8 ( 0 0 4 ) 1.946 48.649 1 ( 2 1 3 ) 1.870 53.427 13 ( 3 1 2 ) 1.713 56.172 39 ( 3 2 1 ) 1.636 57.965 9 ( 3 0 3 ) 1.590 61.419 28 ( 4 0 0 ) 1.508 64.795 35 ( 4 1 1 ) 1.438 64.795 35 ( 2 2 4 ) 1.438 66.408 2 ( 4 0 2 ) 1.407 66.408 2 ( 3 2 3 ) 1.407 69.635 3 ( 4 2 0 ) 1.349 70.42 3 ( 3 3 2 ) 1.336 74.366 14 ( 4 2 2 ) 1.274 74.366 14 ( 4 1 3 ) 1.274 76.755 3 ( 1 1 6 ) 1.241 77.473 4 ( 3 0 5 ) 1.231 80.513 5 ( 4 0 4 ) 1.192 85.76 2 ( 5 1 2 ) 1.132 88.019 6 ( 4 2 4 ) 1.109 89.505 7 ( 4 3 3 ) 1.094 91.801 1 ( 3 1 6 ) 1.073 92.512 7 ( 4 1 5 ) 1.066 97.058 3 ( 2 1 7 ) 1.028

PAGE 152

152 Table A 3 Atomic parameters CuCrO2. System Space Group Lattice () Atomic Coordinates Rhombohedral R3m a b c x y z (166) 2.9761 2.9761 17.102 Cu 0 0 0 Cr 0 0 0.5 90 90 120 O 0 0 0.1079 Table A 4. XRD peak positions for CuCrO2. Intensity h k l d () 15.525 4 ( 0 0 3 ) 5.703 31.35 60 ( 0 0 6 ) 2.851 35.178 14 ( 1 0 1 ) 2.549 36.369 100 ( 0 1 2 ) 2.468 40.852 30 ( 1 0 4 ) 2.2 07 47.842 4 ( 0 0 9 ) 1.900 55.824 20 ( 0 1 8 ) 1.645 62.346 35 ( 1 1 0 ) 1.488 65.411 16 ( 0 0 12 ) 1.426 65.411 16 ( 1 0 10 ) 1.426 71.451 16 ( 1 1 6 ) 1.319 74.382 8 ( 2 0 2 ) 1.274 77.271 2 ( 0 2 4 ) 1.234 88.519 2 ( 2 0 8 ) 1.104 88.519 2 ( 0 1 14 ) 1.104

PAGE 153

153 Table A 5 Atomic parameters for 3R CuSc O2. System Space Group Lattice () Atomic Coordinates Rhombohedral R3m a b c x y z (166) 3.223 3.223 16.976 Cu 0 0 0 Cr 0 0 0.5 90 90 120 O 0 0 0.1079 Approximate value taken from CuCrO 2 Table A 6. XRD peak positions for 3R CuScO2. Intensity h k l d () 15.648 <1 ( 0 0 3 ) 5.658 31.597 21 ( 0 0 6 ) 2.829 32.472 12 ( 1 0 1 ) 2.755 33.768 100 ( 0 1 2 ) 2.652 38.567 36 ( 1 0 4 ) 2.332 41.856 1 ( 0 1 5 ) 2.156 48.206 1 ( 0 0 9 ) 1.886 49.76 <1 ( 1 0 7 ) 1.831 54.252 24 ( 0 1 8 ) 1.689 57.091 26 ( 1 1 0 ) 1.612 59.586 0 ( 1 1 3 ) 1.550 64.152 17 ( 1 0 10 ) 1.450 65.982 4 ( 0 0 12 ) 1. 415 66.73 14 ( 1 1 6 ) 1.401 67.234 1 ( 0 2 1 ) 1.391 68 11 ( 2 0 2 ) 1.377 69.543 1 ( 0 1 11 ) 1.351 71.022 6 ( 0 2 4 ) 1.326 73.254 <1 ( 2 0 5 ) 1.291 77.892 13 ( 1 1 9 ) 1.225 79.088 <1 ( 0 2 7 ) 1.210 81.266 <1 ( 1 0 13 ) 1.183 82.673 6 ( 2 0 8 ) 1.166 85.787 <1 ( 0 0 15 ) 1.132 87.67 3 ( 0 1 14 ) 1.112

PAGE 154

154 Table A 7 Atomic parameters for 2H CuSc O2. System Space Group Lattice () Atomic Coordinates Hexagonal P6 3 /mmc a b c x y z (194) 3.223 11.413 16.976 Cu 1/3 2/3 1/4 Cr 0 0 0 90 90 120 O 1/2 2/3 0.088* Approximate value taken from similar 2H CuCrO 2 system Table A 8. XRD peak positions for 2H -CuScO2. Intensity h k l d () 31.339 30 ( 0 0 4 ) 2.852 33.014 100 ( 1 0 1 ) 2.711 35.79 70 ( 1 0 2 ) 2.507 40.034 10 ( 1 0 3 ) 2.250 51.71 20 ( 1 0 5 ) 1.766 57.101 30 ( 1 1 0 ) 1.612 58.71 30 ( 1 0 6 ) 1.571 65.38 10 ( 0 0 8 ) 1.426 66.349 10 ( 1 0 7 ) 1.408 66.611 20 ( 1 1 4 ) 1.403 67.57 10 ( 2 0 1 ) 1.385 69.264 10 ( 2 0 2 ) 1.355

PAGE 155

155 Table A 9 Ato mic parameters for Cu2Sc2O5. System Space Group Lattice () Atomic Coordinates Orthorhombic Pna2 1 a b c x y z (33) 10.439 3.222 16.976 Cu NA NA NA Sc NA NA NA 90 90 90 O NA NA NA Table A 10. XRD peak positions for Cu2Sc2O5. Intensity h k l d () 14.703 25 ( 0 0 2 ) 6.020 18.469 5 ( 2 0 1 ) 4.800 28.036 12 ( 2 0 3 ) 3.180 28.68 30 ( 0 1 1 ) 3.110 29.919 8 ( 1 1 1 ) 2.984 32.618 25 ( 2 1 0 ) 2.743 32.618 25 ( 1 1 2 ) 2.743 33.497 70 ( 2 1 1 ) 2.673 34.371 100 ( 4 0 0 ) 2.607 34.371 100 ( 2 0 4 ) 2.607 35.728 16 ( 0 1 3 ) 2.511 37.538 8 ( 4 0 2 ) 2.394 37.999 5 ( 3 1 0 ) 2.366 39.8 12 ( 2 1 3 ) 2.263 41.265 16 ( 4 0 3 ) 2.186 41.265 16 ( 2 0 5 ) 2.186 45.312 16 ( 4 1 1 ) 2.000 47.271 20 ( 4 1 2 ) 1.921 48.592 20 ( 2 0 6 ) 1.872 50.391 25 ( 4 1 3 ) 1.809 50.391 25 ( 2 1 5 ) 1.809 54.514 14 ( 3 1 5 ) 1.682 54.514 14 ( 5 1 2 ) 1.682 57.151 12 ( 0 2 0 ) 1.610 57.723 6 ( 6 0 3 ) 1.596

PAGE 156

156 LIST OF REFERENCES 1 D. P. Norton, Mater. Sci. Eng. R 43, 39 (2004). 2 U. Ozgur, Y. I. Alivov, C. Liu, A. Teke, M. A. Reshchikov, S. Dogan, V. Avrutin, S. J. Cho, and H. Morkoc, J. Appl. Phys. 98, 041301 (2005). 3 C. G. Granqvist and A. Hultaker, Thin Solid Films 411, 1 (2002). 4 D. S. Ginley and C. Bright, Mater. Res. Bull. 25, 8 (2000). 5 M. Izaki, K. Mizuno, T. Shinagawa, M. Inabi, and A. Tasaka, J. Electrochem. Soc. 153, C668 (2006) 6 H. Sato, T. Minami, S. Takata, and T. Yamada, Thin Solid Films 236, 27 (1993). 7 T. Miyata, H. Tanaka, H. Sato, and T. Minami, J. Mater. Sci. 41, 5531 (2006). 8 B. J. Ingram, G. B. Gonzalez, D. R. Kammler, M. I. Bertoni, and T. O. Mason, J. Electroceram. 13, 167 (2004). 9 G. J. Exarhos, C. F. Windisch Jr., K. F. Ferris, and R. R. Owings, Appl. Phys. A 89, 9 (2007). 10. R. R. Owings, P. H. Holloway, G. J. Exarhos, and C. F. Windisch Jr., Surf. Interface Anal. 37, 424 (2005). 11. J. A. K. Tareen, A. Malecki, J. P. Doumerc, J. C. Launay, P. Dordor, M. Pouchard, and P. Hagenmuller, Mater. Res. Bull. 19, 989 (1984). 12. J. Windisch C.F., G. J. Exarhos, K. F. Ferris, M. H. Engelhard, and D. C. Stewart, Infrared Transparent Spinel Films with p-Type Conductivity, (Elsevier, San Diego, CA, United states, 2001), p. 45. 13. C. F. Windisch Jr., K. F. Ferris, G. J. Exarhos, and S. K. Sharma, Thin Solid Films 420421, 89 (2002). 14. E. Senadim, H. Kavak, and R. Esen, J. Phys.: Condens. Matter 18, 6391 (2006). 15. S. H. Jeong, I. S. Kim, S. Kim, J. K. Kim, and J. T. Lee, J. Cryst. Growth 264, 110 (2004). 16. S. J. Pearton, C. R. Abernathy, M. E. Overberg, G. T. Thaler, D. P. Norton, N. Theodoropoulou, A. F. Hebard, Y. D. Park, F. Ren, J. Kim, an d L. A. Boatner, J. Appl. Phys. 93, 1 (2003). 17. A. R. Hutson, Phys. Rev. 108, 222 (1957).

PAGE 157

157 18. K. Minegishi, Y. Koiwai, Y. Kikuchi, K. Yano, M. Kasuga, and A. Shimizu, Jpn. J. Appl. Phys., 2 36, 1453 (1997). 19. P. Wand, N. Chen, Z. Yin, F. Yang, C. Peng, R. Dai, and Y. Bai, J. Appl. Phys. 100, 043704 (2006). 20. G. Hu, H. Gong, E. F. Chor, and P. Wu, Appl. Phys. Lett. 89, 021112 (2006). 21. Y. Nakano, T. Morikawa, T. Ohwaki, and Y. Taga, Appl. Phys. Lett. 88, 172103 (2006). 22. J. H. Lim, C. K Kang, K. K. Kim, I. K. Park, D. K. Hwang, and S. J. Park, Adv. Mater. 18, 2720 (2006). 23. R. E. Stauber, J. D. Perkins, P. A. Parilla, and D. S. Ginley, Electrochem. Solid -State Lett. 2, 654 (1999). 24. S. J. Pearton, D. P. Norton, K. Ip, Y. W. Heo, and T. Steiner, Superlattices Microstruct. 34, 3 (2003). 25. E. Fortunato, D. Ginley, H. Hosono, and D. C. Paine, Mater. Res. Bull. 32, 242 (2007). 26. A. Kudo, H. Yanagi, H. Hosono, and H. Kawazoe, Appl. Phys. Lett. 73, 220 (1998). 27. Y. W. Heo, Y W. Kwon, Y. Li, S. J. Pearton, and D. P. Norton, Appl. Phys. Lett. 73, 220 (1998). 28. H. Kawazoe, M. Yasukawa, H. Kyodo, M. Kurita, H. Yanagi, and H. Hosono, Nature 389, 939 (1997). 29. M. A. Marquardt, N. A. Ashmore, and D. P. Cann, Thin Solid Film s 496, 146 (2005). 30. C. R. Friedel and C. R. Hebard, Acad. Sci. 77, 211 (1873). 31. A. F. Rogers, Am. Mineral. 7, 102 (1922). 32. A. F. Rogers, Am. J. Sci. 35, 290 (1913). 33. A. Pabst, Am. Mineral. 31, 539 (1946). 34. W. Soller and A. J. Thomp son, Phys. Rev. 47, 644 (1935). 35. H. Ohta, K. Nomura, H. Hiramatsu, K. Ueda, T. Kamiya, M. Hirano, and H. Hosono, Solid -State Electron. 47, 12 (2003). 36. H. Yanagi, T. Hase, S. Ibuki, K. Ueda, and H. Hosono, Appl. Phys. Lett. 78, 1583 (2001). 3 7. K Ip, Y. W. Heo, D. P. Norton, S. J. Pearton, J. R. LaRoche, and F. Ren, Appl. Phys. Lett. 85, 1169 (2004).

PAGE 158

158 38. G. J. Snyder and E. S. Toberer, Nature Materials 7, 105 (2008). 39. C. B. Vining, Nature Materials 8, 83 (2009). 40. C. B. Vining, Nature Materials 7, 765 (2008). 41. H. Kuriyama, M. Nohara, T. Sasagawa, K. Takubo, T. Mizokawa, K. Kimura, and H. Takagi, High -Temperature Thermoelectric Properties of Delafossite Oxide CuRh1xMgxO2, (Institute of Electrical and Electronics Engineers Inc, Vien na, Austria, 2006), p. 97. 42. S. Shibasaki, W. Kobayashi, and I. Terasaki, Mg Substitution Effects of New Thermoelectric Rh Oxides, (Institute of Electrical and Electronics Engineers Inc, Jeju, Korea, Republic of, 2007), p. 133. 43. V. Varadarajan and D. P. Norton, Appl. Phys. A 85, 117 (2006). 44. H. Yanagi, H. Kawazoe, A. Kudo, M. Yasukawa, and H. Hosono, J. Electroceram. 4, 407 (2000). 45. T. Koyanagi, H. Harima, A. Yanase, and H. Katayama -Yoshida, J. Phys. Chem. Solids 64, 1443 (2003). 46. Y. Wang, H. Gong, and L. Liu, Int. J. Mod. Phys. B 16, 308 (2002). 47. B. J. Ingram, T. O. Mason, R. Asahi, K. T. Park, and A. J. Freeman, Phys. Rev. B 64, 155114 (2001). 48. J. Robertson, P. W. Peacock, M. D. Towler, and R. Needs, Thin Solid Films 411, 9 6 (2002). 49. A. Buljan, P. Alemany, and E. Ruiz, J. Phys. Chem. B 103, 8060 (1999). 50. F. A. Benko and F. P. Koffyberg, Phys. Status. Solidi A 94, 231 (1986). 51. S. Boudin, C. Felser, and F. Studer, Sol. State Sci. 5, 741 (2003). 52. X. Nie, S. H. Wei, and S. B. Zhang, Phys. Rev. Lett. 88, 066405 (2002). 53. L. Dloczik, Y. Tomm, R. Konenkamp, M. C. Lux Steiner, and T. Dittrich, Thin Solid Films 451/452, 116 (2004). 54. A. N. Banerjee, R. Maity, and K. K. Chattopadhyay, Mater. Lett. 58, 10 (2004). 55. N. Tsuboi, Y. Takahashi, S. Kobayashi, H. Shimizu, K. Kato, and F. Kaneko, J. Phys. Chem. Solids 64, 1671 (2003). 56. M. Ohashi, Y. Lida, and H. Morikawa, J. Am. Ceram. Soc. 85, 270 (2002).

PAGE 159

159 57. M. Snure and A. Tiwari, Appl. Phys. Lett. 91, (2007). 58. H. Kawazoe, H. Yanagi, K. Ueda, and H. Hosono, Mater. Res. Bull 25, 28 (2000). 59. H. Yanagi, S. Inoue, K. Ueda, H. Kawazoe, H. Hosono, and N. Hamada, J. Appl. Phys. 88, 4159 (2000). 60. T. Arnold, D. J. Payne, A. Bourlange, J. P. Hu, R. G Egdell, L. F. J. Piper, L. Colakerol, A. De Masi, P. -. Glans, T. Learmonth, K. E. Smith, J. Guo, D. O. Scanlon, A. Walsh, B. J. Morgan, and G. W. Watson, Phys. Rev. B 79, (2009). 61. H. H. Kan and J. F. Weaver, Surf. Sci. 602, L53 (2008). 62. M. And ersson, J. Grins, and M. Nygren, J. Solid -State Chem. 146, 428 (1999). 63. R. D. Shannon, D. B. Rogers, and C. T. Prewitt, Inorg. Chem. 10, 713 (1971). 64. R. D. Shannon, D. B. Rogers, C. T. Prewitt, and J. L. Gillson, Inorg. Chem. 10, 723 (1971). 65. C. Felser, K. Thieme, and R. Seshadri, J. Mater. Chem 9, 451 (1999). 66. V. Eyert, R. Frsard, and A. Maignan, Chem Mater. 20, 2370 (2008). 67. H. C. Kandpal and R. Seshadri, Solid -State Sci 4, 1045 (2002). 68. J. Tate, M. K. Jayaraj, A. D. Draes eke, T. Ulbrich, A. W. Sleight, K. A. Vanaja, R. Nagarajan, J. F. Wager, and R. L. Hoffman, Thin Solid Films 411, 119 (2002). 69. W. C. Sheets, E. S. Stampler, M. I. Bertoni, M. Sasaki, T. J. Marks, T. O. Mason, and K. R. Poeppelmeier, Inorg. Chem. 47, 2 696 (2008). 70. L. Shi, Z. Fang, and J. Li, J. Appl. Phys. 104, 073527 (2008). 71. R. D. Shannon, C. T. Prewitt, and D. B. Rogers, Inorg. Chem. 10, 719 (1971). 72. Y. Wang and N. Herron, J. Phys. Chem. 95, 525 (1991). 73. J. Li, A. W. Sleight, C. Y Jones, and B. H. Toby, J. Solid -State Chem. 178, 285 (2005). 74. M. Jansen, Angew. Chem. Int. Ed. 26, 1098 (1987). 75. E. Mugnier, A. B arnab, and P. Tailhades, Solid -State Ionics 177, 607 (2006). 76. A. Zunger, Appl. Phys. Lett. 83, 57 (2003). 77. B. J. Ingram, B. J. Harder, N. W. Hrabe, T. O. Mason, and K. R. Poeppelmeier, Chem. Mater. 16, 5623 (2004).

PAGE 160

160 78. B. J. Ingram, G. B. Gonzalez, T. O. Mason, D. Y. Shahriari, A. Barnabe, D. Ko, and K. R. Poeppelmeier, Chem. Mater. 16, 5616 (2004). 79. B. J. Ingram, M. I. Bertoni, K. R. Poeppelmeier, and T. O. Mason, Thin Solid Films 486, 86 (2005). 80. R. Nagarajan, N. Duan, M. K. Jayaraj, J. Li, K. A. Vanaja, A. Yokochi, A. Draeseke, J. Tate, and A. W. Sleight, Int. J. Inorg. Mater. 3, 265 (2001). 8 1 K. Isawa, Y. Yaegashi, M. Komatsu, and M. Nagano, Phys. Rev. B 56, 3457 (1997). 82. V. Simonet, R. Ballou, A. P. Murani, O. Garlea, C. Darie, and P. Bordet, J. Phys. : Condens. Matter 16, S805 (2004). 83. K. Isawa, M. Nagano, and K. Yamada, J. Crys t. Growth 237-239, 783 (2002). 84. A. N. Banerjee, C. K. Ghosh, S. Das, and K. K. Chattopadhyay, Phys. B 370, 264 (2005). 85. A. N. Banerjee, C. K. Ghosh, and K. K. Chattopadhyay, Sol. Energy Mater. Sol. Cells 89, 75 (2005). 86. P. Kofstad, Nonstoichiometry, Diffusion and Electrical Conductivity in Binary Metal Oxides, (Wiley Interscience, Canada, 1972) 87. K. Koumoto, H. Koduka, and W. Seo, J. of Mater. Chemistry 11, 251 (2001). 88. R. J. Cava, H. W. Zandbergen, A. P. Ramirez, H. Takagi, C. T. C hen, J. J. Krajewski, W. F. Peck, J. V. Waszczak, G. Meigs, R. S. Roth, and L. F. Schneemeyer, J. Solid State Chem. 104, 437 (1993). 89. R. J. Cava, W. F. Peck Jr., J. J. Krajewski, S. Cheong, and H. Y. Hwang, J. Mater. Res. 9, 314 (1994). 90. C. Lacro ix, B. Canals, and M. Nunez -Regueiro, Phys. B 230-232, 500 (1997). 91. J. G. Bednorz and K. A. Mueller, Phys. B 148, 166 (1987). 92. J. G. Bednorz, M. Takashige, and K. A. Mueller, Mater. Res. Bull. 22, 819 (1987). 93. J. G. Bednorz, M. Takashige, an d K. A. Mueller, Europhys. Lett. 3, 379 (1987). 94. T. Riesterer, J. G. Bednorz, K. A. Mueller, and B. Reihl, Appl. Phys. A A44, 81 (1987). 95. O. Garlea, P. Bordet, C. Darie, O. Isnard, and R. Ballou, J. Phys.: Cond. Matter 16, S811 (2004). 96. Y. W ang and H. Gong, CVD 6, 285 (2000). 97. K. Ueda, H. Hosono, and N. Hamada, J. Appl. Phys. 98, 043506 (2005).

PAGE 161

161 98. N. Duan, A. W. Sleight, M. K. Jayaraj, and J. Tate, Appl. Phys. Lett. 77, 1325 (2000). 99. H. Ohta, M. Orita, M. Hirano, and H. Hosono, J. Appl. Phys. 89, 5720 (2001). 100. H. Kuriyama, M. Nohara, T. Sasagawa, K. Takubo, T. Mizokawa, K. Kimura, and H. Takagi, High -Temperature Thermoelectric Properties of Delafossite Oxide CuRh1xMgxO2, (Institute of Electrical and Electronics Engineers Inc Vienna, Austria, 2006), p. 97. 101. M. Palazzi, Acad. Sci. Paris. C. R. 292, 789 (1981). 102. H. Hiramatsu, K. Ueda, K. Takafuji, H. Ohta, M. Hirano, T. Kamlya, and H. Hosono, J. Appl. Phys. 94, 5805 (2003). 103. H. Hiramatsu, H. Yanagi, T. Kamiya, K. Ueda, M. Hirano, and H. Hosono, Chem. Mater. 20, 326 (2008). 104. K. Horigane and J. Akimitsu, Science and Technology of Advanced Materials 7, 6 (2006). 105. G. Giovannetti, S. Kumar, and D. B. Van, Phys. B 403, 3653 (2008). 106. I. R. Shein and A L. Ivanovskii, Scr. Mater. 59, 1099 (2008). 107. C. Shekhar, S. Singh, P. K. Siwach, H. K. Singh, and O. N. Srivastava, Supercond. Sci. Technol. 22, (2009). 108. S. Shibasaki, W. Kobayashi, and I. Terasaki, Phys. Rev. B 74, 235110 (2006). 109. T. O kuda, T. Onoe, Y. Beppu, N. Terada, T. Doi, S. Miyasaka, and Y. Tokura, J. Magn. Magn. Mater. 310, 890 (2007). 110. M. K. Jayaraj, A. D. Draeseke, J. Tate, and A. W. Sleight, Thin Solid Films 397, 244 (2001). 111. N. Tsuboi, T. Hoshino, H. Ohara, T. Su zuki, S. Kobayashi, K. Kato, and F. Kaneko, J. Phys. Chem. Solids 66, 2134 (2005). 112. R. Kykyneshi, B. C. Nielsen, J. Tate, J. Li, and A. W. Sleight, J. Appl. Phys. 96, 6188 (2004). 113. Y. Ono, K. Satoh, T. Nozaki, and T. Kajitani, Jpn. J. Appl. Phy s., Part 1 46, 1071 (2007). 114. P. W. Sadik, M. Ivill, V. Craciun, and D. P. Norton, Thin Solid Films 517, 3211 (2009). 115. R. Nagarajan, A. D. Draeseke, A. W. Sleight, and J. Tate, J. Appl. Phys. 89, 8022 (2001). 116. J. Topfer, M. Trari, and J. D oumerc, Solid State Sciences 9, 236 (2007).

PAGE 162

162 117. S. K. Kim, B. S. Lee, S. H. Ji, H. Kim, D. Kim, Y. E. Ihm, and W. K. Choo, Phys. Status Solidi B 241, 1545 (2004). 118. G. Dong, M. Zhang, W. Lan, P. Dong, and H. Yan, Vacuum 82, 1321 (2008). 119. R. E Stauber, P. A. Parilla, J. D. Perkins, and D. S. Ginley, Mater. Res. Soc. Symp.265 (2000). 120. H. Hosono, H. Ohta, M. Orita, K. Ueda, and M. Hirano, Vacuum 66, 419 (2002). 121. F. A. Benko and F. P. Koffyberg, J. Phys. Chem. Solids 48, 431 (1987). 122. Z. Qingnan, Z. Xiujian, A. W. Sleight, and J. Li, Preparation and Electrical, Optical Properties of CuScO2:Mg (4.5%) Films on Quartz Glass by Rf Sputtering, (SPIE, Shanghai, China, 2002), p. 248. 123. Y. Kakehi, K. Satoh, T. Yotsuya, K. Masuko, and A. Ashida, Jpn. J. Appl. Phys., Part 1 46, 4228 (2007). 124. M. Yasukawa, K. Ikeuchi, T. Kono, K. Ueda, and H. Hosono, J. Appl. Phys. 98, 1 (2005). 125. T. Otabe, K. Ueda, A. Kudoh, H. Hosono, and H. Kawazoe, Appl. Phys. Lett. 72, 1036 (1998). 126. H. Yanagi, K. Ueda, and H. Hoson o, Trans. Mater. Res. Soc. Jpn. 26, 997 (2001). 127. S. Ibuki, H. Yanagi, K. Ueda, H. Kawazoe, and H. Hosono, J. Appl. Phys. 88, 3067 (2000). 128. H. S. Kim, S. J. Pearton, D. P. Norton, and F. Ren, J. Appl. Phys. 102, 1 04904 (2007). 129. Y. Kwon, Y. Li, Y. W. Heo, M. Jones, P. H. Holloway, D. P. Norton, Z. V. Park, and S. Li, Appl. Phys. Lett. 84, 2685 (2004). 130. H. Ohta, K. Kawamura, M. Orita, M. Hirano, N. Sarukura, and H. Hosono, Appl. Phys. Lett. 77, 475 (2000) 131. S. Saadi, A. Bouguelia, and M. Trari, Sol. Energy 80, 272 (2006). 132. H. Yabuta, M. Sano, K. Abe, T. Aiba, T. Den, H. Kumomi, K. Nomura, T. Kamiya, and H. Hosono, Appl. Phys. Lett. 89, 112123 (2006). 133. J. Tate, M. K. Jayaraj, A. D. Draesek e, T. Ulbrich, A. W. Sleight, K. A. Vanaja, R. Nagarajan, J. F. Wager, and R. L. Hoffman, J. Appl Phys. 89, 8022 (2001). 134. D. Li, D. Fang, Z. H. Deng, S. Zhou, R. H. Tao, W. W. Dong, T. Wang, Y. P. Zhao, G. Meng, and X. B. Zhu, J. Phys. D: Appl. Phys. 40, 4910 (2007).

PAGE 163

163 135. T. Okuda, T. Onoe, Y. Beppu, N. Terada, T. Doi, S. Miyasaka, and Y. Tokura, J. Magn. Magn. Mater. 40, 4910 (2007). 136. K. Tonooka and N. Kikuchi, Thin Solid Films 310, 890 (2006). 137. S. Mahapatra and A. Shivashankar, Chem. V ap. Deposition 9, 238 (2003). 138. I. Barin, and F. Sauert, Thermochemical Data of Pure Substances, (Wiley -VHC, New York, New York, 1989) 139. Y. Kakehi, K. Satoh, T. Yotsuya, S. Nakao, T. Yoshimura, A. Ashida, and N. Fugimura, J. Appl. Phys. 97, 8 3535 (2005). 140. K. Ueda, T. Hase, H. Yanagi, H. Kawazoe, H. Hosono, H. Ohta, M. Orita, and M. Hirano, J. Appl Phys. 89, 1790 (2001). 141. W. T. Lim, L. Stafford, P. W. Sadik, D. P. Norton, S. J. Pearton, Y. L. Wang, and F. Ren, Appl. Phys. Lett. 90, 142101 (2007). 142. W. T. Lim, L. Stafford, P. W. Sadik, D. P. Norton, S. J. Pearton, Y. L. Wang, and F. Ren, Electrochem. Solid -State Lett. 10, H178 (2007). 143. K. T. Jacob, T. H. Okabe, T. Uda, and Y. Waseda, Bull. Mater. Sci. 22, 741 (1999). 144. K. Isawa, and M. Nagano, Synthesis of ACuO2 (A=Al, Ga, Fe and Rh) Delafossites Containing Two-Dimensional Triangular Lattices. 2000), p. 128. 145. R. Yu, C. Lu, D. Tasi, S. Liang, and F. Shieu, J. Electrochem. Soc. 154, H838 (2007). 146. A. J. Bosman and H. J. Van Daal, 19, 1 (1970). 147. N. A. Ashmore and D. P. Cann, J. Mater. Sci. 40, 3891 (2005). 148. F. A. Benko and F. P. Koffyberg, Mater. Res. Bull. 21, 753 (1986). 149. A. P. Schmid, J. Appl. Phys. 39, 3140 (1968). 150. K. Hayashi, K. Sato, T Nozaki, and T. Kajitani, Jpn. J. Appl. Phys. : Part 1 47, 59 (2008). 151. D. C. Look, Semicond. Sci Tech 20, 55 (2005). 152. T. Okuda, Y. Beppu, Y. Fujii, T. Onoe, N. Terada, and S. Miyasaka, Phys. Rev. B 77, (2008). 153. S. Poulston, P. M. Parle tt, P. Stone, and M. Bowker, Surf. Interface Anal. 24, 811 (1996). 154. C. Battistoni, J. L. Dormann, D. Fiorani, E. Pa parazzo, and S. Viticoli, Solid -State Commun. 39, 581 (1981).

PAGE 164

164 155. M. Nolan and S. D. Elliott, Phys. Chem. Chem. Phys. 8, 5350 (2006) 156. H. J. Van Daal and A. J. Bosman, Phys. Rev. 158, 736 (1967). 157. A. N. Banejee, R. Maity, and K. K. Chattopadhyay, Mater. Lett. 58, 10 (2004). 158. H. Yanagi, T. Hase, S. Ibuki, K. Ueda, and H. Hosono, J. Appl. Phys. 88, 4159 (2000). 159. K Ueda, T. Hase, H. Yanagi, H. Kawazoe, H. Hosono, H. Ohta, M. Orita, and M. Hirano, J. Appl. Phys. 89, 1790 (2001). 160. S. Gilliland, J. Pellicer Porres, A. Segura, A. Munoz, P. Rodriguez -Hernandez, D. Kim, M. S. Lee, and T. Y. Kim, Phys. Status Solidi B 244, 309 (2007). 161. S. Gilliland, J. F. Sanchez -Royo, J. Pellicer Porres, A. Segura, A. Munoz, P. Rodriguez Hernandez, and J. Lopez -Solano, Thin Solid Films 516, 1431 (2008). 162. A. Y. Polyakov, N. B. Smirnov, E. A. Kozhukhova, V. I. Vdovin, K. I p, Y. W. Heo, D. P. Norton, and S. J. Pearton, Appl. Phys. Lett. 83, 1575 (2003). 163. H. Yang, Y. Li, D. P. Norton, S. J. Pearton, S. Jung, F. Ren, and L. A. Boatner, Appl. Phys. Lett. 86, 172103 (2005). 164. Y. W. Heo, L. C. Tien, D. P. Norton, B. S. Kang, F. Ren, B. P. Gila, and S. J. Pearton, Appl. Phys. Lett. 85, 2002 (2004). 165. H. T. Wang, B. S. Kang, F. Ren, L. C. Tien, P. W. Sadik, D. P. Norton, S. J. Pearton, and J. Lin, Appl. Phys. Lett. 86, 243505 (2005). 166. L. C. Tien, P. W. Sadik, D P. Norton, L. F. Voss, S. J. Pearton, H. T. Wang, B. S. Kang, F. Ren, J. Jun, and J. Lin, Appl. Phys. Lett. 87, 222106 (2005). 167. H. Chermette and F. Rogemond, NATO ASI Ser., Ser. B: Phys. 283, 587 (1992). 168. J. P. Doumerc, C. Parent, Z. J. Chao, G. Le Flem, and A. Ammar, J. LessCommon Met. 148, 333 (1989). 169. A. Jacob, C. Parent, P. Boutinaud, G. Le Flem, J. P. Doumerc, A. Ammar, M. Elazhari, and M. Elaatmani, Solid State Commun 103, 529 (1997). 170. A. Katsui, Y. Takahashi, and H. Matsush ita, Funtai Oyobi Funmatsu Yakin 55, 101 (2008). 171. N. Tsuboi, T. Hoshino, S. Kobayashi, K. Kato, and F. Kaneko, Phys. Status Solidi A 203, 2723 (2006). 172. N. Tsuboi, H. Ohara, T. Hoshino, S. Kobayashi, K. Kato, and F. Kaneko, Jpn. J. Appl. Phys., Part 1 44, 765 (2005).

PAGE 165

165 173. A. Katsui, Y. Takahashi, and H. Matsushita, Jpn. J. Appl. Phys., Part 2 46, L546 (2007). 174. G. C. Yi, C. Wang, and W. Park, Semicond. Sci. Technol. 20, S22 (2005). 175. C. W. Lee, H. Choi, M. K. Oh, D. J. Ahn, J. Kim, J. M. Kim, F. Ren, and S. J. Pearton, Electrochem. Solid -State Lett. 10, J1 J3 (2007). 176. J. M. Beeckmans and J. R. Brown, Arch. Environ. Health 7, 346 (1963). 177. B. S. Kang, F. Ren, Y. W. Heo, L. C. Tien, D. P. Norton, and S. J. Pearton, Appl. Phys. Lett. 86, 112105 (2005). 178. K. Ogata, T. Hama, K. Hama, K. Koike, S. Sasa, M. Inoue, and M. Yano, Appl. Surf. Sci. 241, 126 (2005). 179. K. Ogata, K. K. Koike, T. Tanite, T. Komuro, S. Sasa, M. Inoue, and M. Yano, Sens. Actuators, B B100, 209 (2004) 180. S. M. Chen and K. C. Lin, Biosens. Bioelectron. 21, 1737 (2006). 181. G. B. Kim, W. S. Chong, T. K. Kwon, K. Hohkawa, C. U. Hong, and N. G. Kim, Jpn. J. Appl. Phys. 44, 2868 (2005). 182. Z. Zhang, N. W. Emanetoglu, G. Saraf, Y. Chen, P. Wu, J. Zhong, Y. Lu, J. Chen, O. Mirochnitchenko, and M. Inouye, IEEE Transactions Ultrason. Ferroelectrics, and Frequency Control 53, 786 (2006). 183. C. Vahlberg, G. R. Yazdi, R. M. Petoral, M. Syvajarvi, K. Uvdal, A. L. Spetz, R. Yakimova, and V. Khranovs ky, IEEE Sensors 4 (2005). 184. A. Ulman, SelfAssembled Monolayers of Thiols, Thin Films, (Academic Press, San Diego, CA, 1998) 185. R. Colorado, and T. R. Lee, Thiol -Based Self -Assembled Monolayers: Formation and Organization, (Science and Technolog y Elsevier Science Ltd., 2001) 186. M. Nakano, T. Ishida, T. Numata, Y. Ando, and S. Sasaki, Jpn. J. Appl. Phys. 42, 4734 (2003). 187. N. S. Peshika, Z. Hu, K. J. Stebe, and P. C. Searson, J. Phys. Chem. B 106, 6985 (2002). 188. S. Hovel, C. Kolczeqski, M. Wuhn, J. Albers, K. Weiss, V. Staemmler, and C. Woll, J. Phys. Chem. 112, 3909 (2000). 189. G. M. Kale, J. Mater. Sci. 30, 1420 (1995). 190. B. Ling, X. W. Sun, J. L. Zhao, S. T. Tan, Z. L. Dong, Y. Yang, H. Y. Yu, and K. C. Qi, Phys. E 41, 635 (2009).

PAGE 166

166 BIOGRAPHICAL SKETCH Patrick essentially stayed put in the subu rb of Humble, Texas from the time he was two until he was twenty five years old. While there, he attended the Oaks Elementary School and Atascoscita Middle School. In 1996 he graduated 5th in his class from Humble High Schoo l out of roughly 500 students. He next pursued a chemical engineering degree from t he University of Houston where he graduated Sum m a cum Laude with a 3.6 GPA and University Honors and Honors in Ma jor in the spring of 2000. His senior thesis was entitled, A Study of Pluronic Gels for Use in Drug Delivery. Between 2000 and Fall of 2004 He attended law school, worked as a petroleum research engineer for Halliburton, and as a network on-site technici an for the National Multiple Scle rosis Society. In fall of 2004 he enrolled in the Materials Science and Engineering as a Ph.D. student and has continued in the program since that time