<%BANNER%>

Conductivity and Stability of Bismuth Oxide-based Electrolytes and Their Applications for IT-SOFCs

Permanent Link: http://ufdc.ufl.edu/UFE0024943/00001

Material Information

Title: Conductivity and Stability of Bismuth Oxide-based Electrolytes and Their Applications for IT-SOFCs
Physical Description: 1 online resource (188 p.)
Language: english
Creator: Jung, Doh
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: bismuth, cell, conductivity, electrolyte, fuel, lattice, oxide, parameter
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: Cubic stabilized ((DyO1.5)x-(WO3)y-(BiO1.5)1-x-y) electrolytes (DWSB) having higher ionic conductivity than (ErO1.5)0.2(BiO1.5)0.8 (20ESB) were developed to achieve higher conductivity. An optimal 2:1 dopant content ratio (Dy:W) was determined based on the solid solubility limit, X-ray diffraction (XRD) pattern and Arrhenius behavior. Various compositions with the same 2:1 dopant content ratio were tested to obtain the conductivity dependence on total dopant concentration. With double doping, stabilization of the cubic phase was achieved with as little as 12 mol% total dopant concentration. Overall, DWSB has a closer inherent structure to pure delta-Bi2O3 than any singly doped compositions. Both lattice parameter and conductivity linearly extrapolate with total dopant concentration to that of pure delta-Bi2O3, resulting in the ability to stabilize delta-phase at lower dopant concentration thus achieving higher conductivity. However, this DWSB composition experienced conductivity degradation like other cubic stabilized bismuth oxides at intermediate temperatures (IT), i.e. 500 ~ 700 oC. Several DWSB compositions with the same 2:1 dopant content ratio (Dy:W) were annealed to observe isothermal conductivity behavior in the IT range with time. All DWSB compositions maintained their initial conductivity at 700 oC, but underwent conductivity degradation at < = 600 oC. The effect of total dopant concentration on conductivity degradation behavior was investigated at 600 oC and 500 oC. Notably, the effect of dopant composition on conductivity behavior with time at 500 oC demonstrates that there is a trade-off between initial conductivity and long term stability at this temperature. Therefore, it is necessary to find an optimal total and relative concentration of dopants to provide the enhanced long term stability needed to make the DWSB electrolyte system feasible for 500 oC operation. To this end, it was found that (DyO1.5)0.25-(WO3)0.05-(BiO1.5)0.70, 25D5WSB, maintained a conductivity of 0.0068 S/cm without appreciable degradation after annealing at 500 oC for 500 hours. Cathode performance can be improved by the introduction of a second ionic conducting phase with high conductivity due to increased triple phase boundary (TPB) lengths. In this study, bismuth oxide-based electrolytes were combined with La1-xSrxMnO3-delta (LSM) to obtain composite cathodes. This composite cathode had better performance on ESB ((Er2O3)0.20(Bi2O3)0.80) than on GDC (Gd0.1Ce0.9O2-delta).
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 Doh Jung.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Wachsman, Eric D.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024943:00001

Permanent Link: http://ufdc.ufl.edu/UFE0024943/00001

Material Information

Title: Conductivity and Stability of Bismuth Oxide-based Electrolytes and Their Applications for IT-SOFCs
Physical Description: 1 online resource (188 p.)
Language: english
Creator: Jung, Doh
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: bismuth, cell, conductivity, electrolyte, fuel, lattice, oxide, parameter
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: Cubic stabilized ((DyO1.5)x-(WO3)y-(BiO1.5)1-x-y) electrolytes (DWSB) having higher ionic conductivity than (ErO1.5)0.2(BiO1.5)0.8 (20ESB) were developed to achieve higher conductivity. An optimal 2:1 dopant content ratio (Dy:W) was determined based on the solid solubility limit, X-ray diffraction (XRD) pattern and Arrhenius behavior. Various compositions with the same 2:1 dopant content ratio were tested to obtain the conductivity dependence on total dopant concentration. With double doping, stabilization of the cubic phase was achieved with as little as 12 mol% total dopant concentration. Overall, DWSB has a closer inherent structure to pure delta-Bi2O3 than any singly doped compositions. Both lattice parameter and conductivity linearly extrapolate with total dopant concentration to that of pure delta-Bi2O3, resulting in the ability to stabilize delta-phase at lower dopant concentration thus achieving higher conductivity. However, this DWSB composition experienced conductivity degradation like other cubic stabilized bismuth oxides at intermediate temperatures (IT), i.e. 500 ~ 700 oC. Several DWSB compositions with the same 2:1 dopant content ratio (Dy:W) were annealed to observe isothermal conductivity behavior in the IT range with time. All DWSB compositions maintained their initial conductivity at 700 oC, but underwent conductivity degradation at < = 600 oC. The effect of total dopant concentration on conductivity degradation behavior was investigated at 600 oC and 500 oC. Notably, the effect of dopant composition on conductivity behavior with time at 500 oC demonstrates that there is a trade-off between initial conductivity and long term stability at this temperature. Therefore, it is necessary to find an optimal total and relative concentration of dopants to provide the enhanced long term stability needed to make the DWSB electrolyte system feasible for 500 oC operation. To this end, it was found that (DyO1.5)0.25-(WO3)0.05-(BiO1.5)0.70, 25D5WSB, maintained a conductivity of 0.0068 S/cm without appreciable degradation after annealing at 500 oC for 500 hours. Cathode performance can be improved by the introduction of a second ionic conducting phase with high conductivity due to increased triple phase boundary (TPB) lengths. In this study, bismuth oxide-based electrolytes were combined with La1-xSrxMnO3-delta (LSM) to obtain composite cathodes. This composite cathode had better performance on ESB ((Er2O3)0.20(Bi2O3)0.80) than on GDC (Gd0.1Ce0.9O2-delta).
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 Doh Jung.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Wachsman, Eric D.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024943:00001


This item has the following downloads:


Full Text

PAGE 1

1 CONDUCTIVITY AND STABILITY OF BISMUTH OXIDE BASED ELECTROLYTES AND THEIR APPLICATIONS FOR IT SOFCS By DOH WON JUNG 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 D oh W on J ung

PAGE 3

3 To my loving family

PAGE 4

4 ACKNOWLEDGMENTS I would like to thank my advisor, Dr. Eric D. Wachsman, for his support and guidance. His encouragement helped me to reach a higher level of success and expand my potential. I would like to thank Dr. Juan C. Nino for invaluable suggestions and discussions. I also would like to thank Dr. Simon Phillpot, Dr. Vale ntin Craciun and Dr. Mark Orazem for their advice, guidance and constructive comments. I would like to thank Dr. Keith Duncan for his encouragement, understanding and friendship. I would like to acknowledge Dr. Heesung Yoon for his guidance and discussion s. I also wish to acknowledge my former and current group member s ; Dr. Matthew Camaratta, Dr. Takkeun Oh, Dr. Jin Soo Ahn, Dr. Shobit Omar, Dr. Sean Bishop, Jianli n Li Byung Wook Lee and other members for providing me excellent research environment and he lpful comments. Special thanks to Kang Taek Lee and Dongjo Oh for giving me opportunity to co-work with them. I would like to thank my parents and my wife s parents for their support and trust. Finally, I thank my beautiful wife, Youjin Han, for her love, encouragement and giving me a lovely son, Jiwoo (Bradley).

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................................... 4 LIST OF TABLES ................................................................................................................................ 8 LIST OF FIGURES .............................................................................................................................. 9 ABSTRACT ........................................................................................................................................ 14 CHAPTER 1 INTRODUCTION ....................................................................................................................... 16 1.1 Motivation .......................................................................................................................... 16 1.2 Objective ............................................................................................................................ 20 1.3 Dissertation Organization ................................................................................................. 20 2 BACKGROUND ......................................................................................................................... 21 2.1 Solid Oxide Fuel Cell ........................................................................................................ 21 2.1.1 Principles of Solid Oxide Fuel Cell ..................................................................... 21 2.1.2 Actual Performance .............................................................................................. 38 2.2 Solid Oxide Fuel Cell Electrolytes for IT -SOFCs .......................................................... 39 2. 2.1 Phases and Conductivity of Pure Bi2O3 .............................................................. 39 2.2.2 Bi2O3 ..................................................................................... 40 2.2. 3 Stabilized Bismuth Oxide ..................................................................................... 42 2.2.3.1 Structure and ionic conductivity of doped bismuth oxides ................. 42 2.2.3.2 Stabilization with double dopants ......................................................... 44 2.2.3.3 The determination of dopant species .................................................... 44 2.2.4 Phase Stability ....................................................................................................... 46 2.2.5 Structura l Stability ................................................................................................ 46 2.3 Impedance Spectroscopy .................................................................................................. 47 2.3. 1 Impedance Spectra for Conventional Ceramics .................................................. 47 2.3.2 The Comparison of Impedance Response ........................................................... 48 2.3.2.1 CeO2-based electrolyte .......................................................................... 48 2. 3.2. 2 Bi2O3-based electrolyte ......................................................................... 49 3 EFFECT OF TOTAL DOPANT CONCENTRATION AND DOPANT RATIO ON CONDUCTIVITY OF DWSB COMPOSITIONS ................................................................... 53 3.1 Introdu ction ....................................................................................................................... 53 3. 2 Experimental Procedure .................................................................................................... 54 3.2.1 Preparation of Electrolyte Samples ..................................................................... 54 3. 2.2 X Ray Diffraction Analysis ................................................................................. 68 3.3 Results and Discussion ..................................................................................................... 68

PAGE 6

6 3. 3.1 Structure ................................................................................................................ 68 3. 3.2 Conductivity .......................................................................................................... 71 4 TEMPERATURE AND TIME DEPENDENT CONDUCTIVITY BEHAVIOR OF DWSB ELECTROLYTE SYSTEMS ........................................................................................ 81 4. 1 Introduction ....................................................................................................................... 81 4.2 Experimental Procedure .................................................................................................... 95 4.2.1 Sample Preparation ............................................................................................... 95 4. 2.2 Temperature Dependent Conductivity Behavior ................................................ 95 4. 2.3 Characterization .................................................................................................... 95 4. 3 Results a nd Discussion ..................................................................................................... 95 4. 3.1 Conductivity .......................................................................................................... 95 4.3.2 Stability ................................................................................................................. 96 4.3.2. 1 Conductivity behavior of DWSB in the IT ranges .............................. 96 4.3.2. 2 Conductivity degradation mechanism ................................................ 100 4.3.2.3 Conductivity behavior of 8D4WSB ................................................... 106 5 ENHANCED LONG TERM STABILITY OF BISMUTH OXIDE BASED ELECTROLYTES FOR IT SOFC OPERATION AT 500 OC .............................................. 109 5. 1 Introduction ..................................................................................................................... 109 5. 2 Experimental Procedure .................................................................................................. 110 5. 2.1 Sample Preparation ............................................................................................. 110 5. 2.2 Time Dependent Conductivity Behavior ........................................................... 110 5. 2.3 Phase Analysis .................................................................................................... 111 5. 3 Results a nd Discussion ................................................................................................... 111 5. 3.1 Conductivity ........................................................................................................ 111 5. 3.2 Stability ............................................................................................................... 121 5. 3.2.1 Long term conductivity behavior ........................................................ 121 5. 3.2. 2 Conductivity behavior of 25DSB ....................................................... 124 6 NEW DOUBLY DOPED BISMUTH OXIDE ELECTROLYTES ....................................... 128 6.1 Introduction ..................................................................................................................... 128 6.2 Experimental Procedure .................................................................................................. 129 6.3 Results and discussion .................................................................................................... 129 6.3.1 (Tb4O7) (Bi2O3) and ( Tb4O7) (WO3) (Bi2O3) System .................................... 129 6.3.2 (Dy2O3) (Gd2O3) (Bi2O3) System ..................................................................... 142 6.3.3 (Dy2O3) (CeO2) (Bi2O3) System ........................................................................ 147 7 HIGH PERFORMANCE LSM ESB AND LSM -DWSB COMPOSITE CATHODES ...... 152 7.1 Introduction ..................................................................................................................... 152 7.2 Experimental Procedure .................................................................................................. 153 7.2.1 Electrode Preparation ......................................................................................... 153 7. 2.2 Electrolyte Preparation ....................................................................................... 153 7.2.3 Characterizations ................................................................................................ 154

PAGE 7

7 7.3 Results and Discussions .................................................................................................. 154 7.3.1 LSM 20ESB C omposite Cathode ...................................................................... 154 7.3.2 LSM 8D4WSB Composite Cathode ................................................................. 162 8 SUMMARY ............................................................................................................................... 166 APPENDIX A CONDUCTIVITY COMPARISON ........................................................................................ 168 B LATTICE PARAMETER CALCULATION .......................................................................... 171 C CHARACTERIZATION OF XRD PATTERNS .................................................................... 176 LIST OF REFERENCES ................................................................................................................. 182 BIOGRAPHICAL SKETCH ........................................................................................................... 188

PAGE 8

8 LIST OF TABLES Table page 2 1 Structur e data of the Bi2O3 phases [18] ................................................................................ 50 2 2 Bi2O3 measured at 778 oC [27] .................................................. 51 2 3 Enthalpies of (Bi2O3)1 x(Er2O3)x [50] .................................................................................. 52 3 1 Conductivity activation energy and lattice parameter of 10D5WSB, 7.5D7.5WSB and 7D3WSB. ......................................................................................................................... 78 3 2 Conductivity activation energies of various DWSB compositions. .................................... 79 3 3 Pre -exponential terms of various DWSB compositions. ..................................................... 80 4 1 Conductivity activation energies of 8D4WSB, 10D5WSB, 12D6WSB, 14D7WSB and 20ESB. ........................................................................................................................... 108 5 1 Conductivity activation energies for 10D5WSB, 15D5WSB, 20D5WSB and 25D5WSB ............................................................................................................................ 126 5 2 Conductivity of various compositions before and after annealing at 500 oC for different hours. ..................................................................................................................... 127 7 1 Activation energy and ASR of three different samples ..................................................... 165 A 1 S ample dimension and the current collector used for 8D4WSB and 12D6WSB to confirm reproducibility. ....................................................................................................... 170 C1 Structure parameters of cubic Bi2O3 [27, 38]. .................................................................... 180 C2 Structure parameters of Bi0.775Dy0.225O1.5 (Rhombohedral) [109]. ................................... 181

PAGE 9

9 LIST OF FIGURES Figure page 1 1 Electrical conductivity of fluorite oxides ............................................................................ 20 2 1 Principle of operation of a solid oxide fuel cell with hydrogen as a fuel and oxygen as an oxidant ........................................................................................................................... 37 2 2 Ideal and actual fuel cell voltage/current characteristic [17] ............................................... 38 2 3 Electrical conductivit 2O3 ................................................ 39 2 4 Structural mode -Bi2O3 ............................................................................................... 40 2 5 Bi2O3 at 7 78 oC ............................. 41 2 6 Xmin, the minimum value of x required to stabilize the fcc phase in (Bi2O3)1x(Ln2O3)x as a function of the ionic radius ( Rion) of Ln3+ [38] ............................................................. 42 2 7 2O3)1x(Ln2O3)x for Xmin as a function of the ionic radius of the substituted Ln3+ [38] ......................................................................................... 43 2 8 Relative conductivity, as a fuction of time, at 500 oC, for different dopants [33] ............. 44 2 9 Relationship between conductivity decay time constant and dopant polarizability [43] .......................................................................................................................................... 45 2 10 Relat ive conductivity as a function of time, at 500 oC, for different Er2O3 concentrations [50] ................................................................................................................. 46 2 11 Complex plane diagram [55] ................................................................................................. 47 2 12 Impedance spectra of CGO20 with 200 ppm and 3000 ppm SiO2, measured at 350 oC in air [60] ................................................................................................................................ 48 2 13 Impedance complex plane plots of 20ESB [61] ................................................................... 49 3 1 Flow chart for the powder synthesis using conventional solid state reaction .................... 67 3 2 X ray diffraction patterns of DWSB ..................................................................................... 68 3 3 X ray diffraction patterns of DWSB with 2:1 mol dopant ratio between Dy and W ........ 69 3 4 Lattice parameters of various DWSB compositions with 2:1 mol dopant ratio measured at room temperature (RT) and 760 oC as a function of total dopant concentration .......................................................................................................................... 70 3 5 Typical impedance spectra of 8D4WSB composition in air .............................................. 71

PAGE 10

10 3 6 Arrhenius plot of conductivities for 10D5WSB, 7.5D7.5WSB and 7D3WSB. ................. 72 3 7 Conductivity versus total dopant concentration of DWSB electrolyte with 2:1 mol dopant ratio ............................................................................................................................ 73 3 8 Conductivity versus lattice parameter of DWSB compositions with 2:1 mol dopant ratio ........................................................................................................................................ 74 3 9 Arrhenius plot of conductivi ties for 8D4WSB, 9D4.5WSB, 10D5WSB, 11D5.5WSB, 12D6WSB, 14D7WSB 20ESB and 10GDC ; the dash ed line represents the conductivity of 10YSZ [75] ............................................................................................. 75 3 10 The conductivity activation energy for low t emperature region ( 600 oC) and for high temperature region ( 600 oC) for DWSB systems as a function of total dopant concentration of DWSB compositions .................................................................................. 76 3 11 The pre -exponential term for low temperature region ( 600 oC) and for high temperature region ( 600 oC) as a function of total dopant concentration of DWSB compositions ........................................................................................................................... 77 4 1 Comparison of conductivity decay of 25YSB at 500 oC and 650 oC [33]. ......................... 94 4 2 Arrhenius plot of conductivities for 8D4WSB, 10D5WSB, 12D6WSB, 14D7WSB 20ESB and 10GDC. ............................................................................................................... 95 4 3 Conductivity behavior o f 8D4WSB, 10D5WSB and 12D6WSB annealed at 700 oC as a function of time ............................................................................................................... 96 4 4 Isothermal comparison of time -dependent c onductivity behavior for 8D4WSB, 10D5WSB 12D6WSB and 14D7WSB annealed at 6 00 oC and 5 00 oC ............................ 97 4 5 Relative time dependent conductivity degradation for 8D4WSB, 10D5WSB 12D6WSB and 14D7WSB annealed at 6 00 oC and 5 00 oC ................................................ 98 4 6 Comparison of relative conductivity change of various DWSB compositions after 100 hour annealing at 600 oC and 500 oC as a function of total dopant concentration ..... 99 4 7 SEM images of 8D4WSB as -sintered annealed at 600 oC for 100 hours and annealed at 500 oC for 100 hours ........................................................................................................ 100 4 8 XRD patterns of 8D4WSB, 10D5WSB, 12D6WSB and 14D7WSB annealed at 600 oC for 10 0 hours ................................................................................................................... 101 4 9 Arrhenius plots of electrical conductivity for (Bi2O3)1x(WO3)x, where x = 0.125, 0.25 and 0.50 [69] ................................................................................................................ 102 4 10 XR D patterns of 8D4WSB, 10D5WSB, 12D6WSB, 14D7WSB and 20ESB annealed at 500 oC for 100 hours ........................................................................................................ 103

PAGE 11

11 4 11 DTA heating curve for 8D4WSB; as -sintered powder, annealed at 600 oC for 100 hours and annea led at 500 oC for 100 hours ....................................................................... 104 4 12 XRD patterns of 8D4WSB annealed at 500 oC for 100 hours before and after DTA measurements ....................................................................................................................... 105 4 13 Time dependent conductivity and relative conductivity for 8D4WSB annealed at various temperatures as a function of time ......................................................................... 106 4 14 XRD patterns of 8D4WSB annealed at various temperatures for 100 hours ................... 107 5 1 X ray diffraction patterns of 10D5WSB, 15D5WSB, 20D5WSB and 25D5WSB; inset represents magnif oC. ..................................................... 118 5 2 Lattice parameter versus (Dy + W) dopant concentration of DWSB compositions measured at room temperature. ........................................................................................... 119 5 3 Arrhenius plot of conductivities for 10D5WSB, 15D5WSB, 20D5WSB and 25D5WSB. ............................................................................................................................ 120 5 4 Conductivity and relative conductivity for 10D5WSB, 15D5WSB, 20D5WSB 25D5WSB and 25DSB annealed at 500 oC as a function of time ..................................... 121 5 5 Typical impedance spectra of 25D5WSB and 10GDC in air at 500 oC ........................... 122 5 6 X RD patterns for 10D5WSB, 15D5WSB, 20D5WSB and 25D5WSB annealed at 500 oC for different time periods; parenthesis represents the total annealing hours for each composition .................................................................................................................. 123 5 7 Relative cond uctivity comparison for 25DSB composition which is taken from between the present work and Jiang et al. [33] at 500 oC as a function of time ............... 124 5 8 XRD patterns of 25DSB composition of as sintered, annealed at 500 oC for 96 hours and 300 hours ....................................................................................................................... 125 6 1 X ray diffraction patterns of (Tb4O7)x(Bi2O3)1x, where x = 0.15, 0.20 and 0.25 ............ 136 6 2 X ray diffraction patterns of 8T4WSB and 10T5WSB ..................................................... 137 6 3 Arrhenius plot of conductivities for 10T5WSB, 25TSB, 8D4WSB and 20ESB ............. 138 6 4 Isothermal comparison of time -dependent conductivity for 10T5WSB and 25TSB annealed at 500 oC; 20ESB data is added for comparison ................................................. 139 6 5 XRD patterns of 25TSB as -s intered and annealed at 500 oC for 120 hours ..................... 140 6 6 XRD patterns of 10T5WSB as sintered and annealed at 500 oC for 100 hours ............... 141

PAGE 12

12 6 7 X RD patterns of 8D4GSB, 6D6GSB and 4D8GSB after first calcined and second calcined at 800 oC for 16 hours ........................................................................................... 142 6 8 X RD patterns of 10D5GSB, 7.5D7.5GSB and 5D10GSB after first calcination and second cal cination at 800 oC for 16 hours .......................................................................... 143 6 9 X RD patterns of 12D6GSB after the first and second calcinations steps at 800 oC for 16 hours ................................................................................................................................. 144 6 10 Arrhenius plot of conductivit y for 8D4GSB, 10D5GSB, 12D6GSB and 20ESB; the conductivities were measured in both direction of temperature measurements, HL (High to Low) and LH (Low to High) for 12D6GSB ........................................................ 145 6 11 Conductivity vs. time for 8D4GSB, 10D5GSB and 12D6GSB annealed at 5 00 oC; 20ESB is shown for comparison ......................................................................................... 146 6 12 XRD patterns of 8D4GSB and 10D5GS B as -sintered and annealed at 500 oC for 100 hours ...................................................................................................................................... 147 6 13 XRD patterns of calcined (CeO2)x(Bi2O3)1x (x=0.15, 0.20, 0.30 and 0.40) ..................... 148 6 14 Conductivity vs.time for 10D2CSB at 500 oC; 20ESB was added for comparison......... 149 6 15 Conductivity vs. time for 10D2CSB, 8D4WSB and 10D5WSB at 600 oC; inset represents the conducti vity behavior of these compositions for 100 hours ...................... 150 6 16 XRD patterns of 10D2CSB as sintered, annealed at 500 oC for 100 hours and annealed at 600 oC for 500 hours ........................................................................................ 151 7 1 XRD patterns of LSM 20ESB mixture s fir ed at 900 oC for 50 hours .............................. 158 7 2 Impedance spectra of Pure LSM on ESB (Cell 1), LSM ESB on GDC (Cell 2) and LSM ESB on ESB (Cell 3) measured at 700 oC in air. ..................................................... 159 7 3 Arrhenius plot of ASR vs. temperature for Pure LSM on ESB (Cell 1), LSM ESB on GDC (Cell 2), LSM ESB on ESB (Cell 3) and Jiang et al. [93]. ...................................... 160 7 4 ASR vs. time for LSM ESB on ESB (Cell 3) at 700 oC. ................................................... 161 7 5 XRD patterns of LSM 8D4WSB mixtures fir ed at 900 oC for 50 hours .......................... 162 7 6 Arrhenius plot of ASR vs. temperature for LSM DWSB on ESB(Cell 4) and LSM DWSB on DWSB (Cell 5) including previous Cell 1 and Cell 3. .................................... 163 7 7 Cross -sectional micrographs of LSM -ESB cathode on ESB (Cell 3) and LSM ESB cathode on GDC (Cell 2) ..................................................................................................... 164 A 1 Reproducibility data for 8 D4WSB (this work and pre vious work [13]) and 20ESB 12D6WSB ............................................................................................................................. 169

PAGE 13

13 B1 Calculated lattice parameter versus Nelson Riley function for 10D5WSB taken at room temperature ................................................................................................................. 173 B2 XRD patterns of 25TSB measured at 25 oC, 300 oC, 500 oC and 600 oC ......................... 174 B3 Calculated lattice parameter versus Nelson Riley function for 25TSB at different temperatures .......................................................................................................................... 175 C1 XRD patterns of observed 25DSB annealed at 500 oC for 300 hours and calculated mixtures of cubic and rhombohedral phases ...................................................................... 178 C2 XRD patterns of observed 10D5WSB annealed at 500 oC for 100 hours and calculated mixtures of cubic, orthorhombic and tetragonal phases .................................. 179

PAGE 14

14 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partia l Fulfillment of the Requirements for the Degree of Doctor of Philosophy CONDUCTIVITY AND STA BILITY OF BISMUTH OXIDE BASED ELECTROLYTES AND THEIR APPLICATIONS F OR IT SOFCS By Doh Won Jung August 2009 Chair: Eric D. Wachsman Major: Materials Science and Engineering Cubic stabilized ((DyO1.5)x(WO3)y(BiO1.5)1xy) electrolytes (DWSB) having higher ionic conductivity than (ErO1.5)0.2(BiO1.5)0.8 (20ESB) were developed to achieve higher conductivity. An optimal 2:1 dopant content ratio (Dy:W) was determin ed based on the solid solubility limit, X ray diffraction (XRD) pattern and Arrhenius behavior. V arious compositions with the same 2:1 dopant content ratio wer e tested to obtain the conductivity dependence on total dopant concentration. With double doping, stabilization of the cubic phase was achieved with as little as 12 mol% total dopant concentration. O verall, DWSB has a closer inherent structure to pure Bi2O3 than any singly doped compositions. Both lattice parameter and conductivity linearly extrapolate with total dopant concentration to that of pure -Bi2O3, resulting in the ability to stabilize -phase at lower dopant concentration thus achieving higher conductivity. However, this DWSB composition experienced conductivity degradation like other cubic stabilized bismuth oxides at intermediate temperature s (IT), i.e. 500 ~ 700 oC. Several DWSB compositions with the same 2:1 dopant content ratio (Dy:W) were annealed to observe isothermal conductivity behavior in the IT range with time. All DWSB compositions maintained their initial conductivity at 700 oC, but underwent conductivity degradation at 600 oC. The

PAGE 15

15 effect of total dopant concentration on conductivity degradation behavior was investigated at 600 oC and 500 oC. Notably, t he effect of dopant composition on conductivity behavior with time at 500 oC demonstrates that there is a trade off bet ween initial conductivity and long term stability at this temperature. Therefore, it is necessary to find an optimal total and relative concentration of dopants to provide the enhanced long term stability needed to make the DWSB electrolyte system feasible for 500 oC operation. To this end, it was found that ( DyO1.5)0.25(WO3)0.05(BiO1.5)0.70, 25D5WSB, maintained a conductivity of 0.0068 S/cm without appreciable degradation after annealing at 500 oC for 500 hours. Cathode performance can be improved by the introduction of a second ionic conducting phase w ith high conductivity due to increased triple phase boundary (TPB) lengths In this study, bismuth oxide -based electrolytes were combined with La1xSrxMnO3 (LSM) to obtain composite cathode s This comp osite cathode had better performance on ESB ((Er2O3)0.20(Bi2O3)0.80) than on GDC ( Gd0.1Ce0.9O2).

PAGE 16

16 CHAPTER 1 INTRODUCTION 1.1 Motivation Fuel cells are energy conversion devices th at produce electricity by electrochemical combination of a fuel with an oxidant. Particular ly, solid oxide fuel cells (SOFCs) are potential alternative electric power generation systems becau se of their high electrical efficiency, fuel flexibility and minimal environmental impact [1, 2]. One of the major components of an electrochemical cell is th e electrolyte, which is an ion-c onducting membrane that separates the two electrodes. Solid oxide electrolytes should satisfy numerous requirements, including fast ionic transport, neg ligible electronic conduction and hi gh mechanical and thermodynamic stability over a wide range of temperatur es and oxygen partia l pressures [3]. Yttria stabilized zirconia (YSZ ) is the conventional electrolyte for SOFCs. YSZ exhibits ionic conductivity of about 0.1 S/cm at 1000 oC [4]. Unfortunately, electr olytes based on zirconia have to be operated at > 700 oC to obtain acceptable oxygen ion conductivity [5]. Such high operating temperatures demand the use of ceramics for interconnects and insulation, as well as time and energy to heat up to the operating temper ature. Therefore, if SOFCs could be designed to give a reasonable power out put in the IT range (500-700 oC), it becomes feasible to use lowcost, readily available metallic materials such as stainless steel for the interconnects and other balance-of-plant materials [2]. As shown in Figure 1-1, YSZ electrolyte exhi bits considerably lo w ionic conductivity in the IT range [6]. For lower temperature cell opera tion to be achieved, the electrolyte resistance must be lowered. This lower el ectrolyte resistance can be ach ieved either by selecting new electrolyte materials that po ssess higher ionic conductivity or by decreasing the electrolyte thickness with better cell fabrication technique s [7-9]. In particular the conductivity of -Bi2O3

PAGE 17

17 is one to two orders of magnitude higher than that of stabilized zirconia at corresponding temperatures [10]. Therefore, re placement of YSZ with an inte rmediate temperature oxide ion conductor may provide a significan t reduction in the material a nd fabrication problems [11]. Cubic-stabilized bismuth oxide s are known to exhibit the hi ghest ionic conductivities [1214]. However, the instability of the bismuth oxide-based el ectrolytes under a reducing atmosphere may limit their applications in fuel cell environments. Takahashi et al. showed that the oxygen partial pressure for the decomposition of Bi/Bi2O3 at 600 oC occurs at Po2 = 10-13.1 atm [15]. A bilayered el ectrolyte concept has been devel oped to overcome this thermodynamic instability of cubic-stabilized bi smuth oxides [16]. It has been also reported that cubic-stabilized bismuth oxides undergoes phase and structural ch anges in IT ranges [11]. Therefore, it is required to develop new bismuth oxide-based el ectrolytes which fulfills both criteria of conductivity and stability for SOFC applications. 1.2 Objective Numerous studies have been carried out to develop new bismuth oxide-based electrolytes to achieve higher ionic conductiv ity. Single dopant systems have been mainly employed to stabilize the cubic fluorite structure of -Bi2O3. Recently, in the literature, it was found that a double doping strategy could increase the ionic conduc tivity of bismuth oxide -based electrolytes compared to the single dopant system. However, in previous studies, systematic criteria for selecting dopants were not well established. In this work, we select prospective dopants based on their polarizabil ity and ionic radius. After identifying the dopants, th e effect of total dopant conc entration and dopant ratio on the structure and conductivity of bismuth oxide based el ectrolytes will be investigated to obtain an optimum dopant concentrati on for maximum conductivity.

PAGE 18

18 With respect to stability, it ha s been known that cubic-stabi lized bismuth oxides experience conductivity degradation when annealed in the IT ra nges for long periods of time. Therefore, it is necessary and meaningful to examine conductivity behavior of newly developed electrolyte materials. The conductivity behavior of severa l bismuth oxide-based el ectrolytes will be investigated as a function of temperature, time total dopant concentration and dopant ratio. This study will provide an ideal dopant composition wh ich takes into account stability as well as conductivity at a given operating temperatures Conductivity degradation mechanisms will be examined to identify the reasons for the decrease in conductivity at intermediate temperatures. In addition, a new approach will be explored to enhance long term stability of bismuth oxide-based electrolytes. Finally, a new composite cathode material c ontaining bismuth oxide-b ased electrolytes will be fabricated. To compare cathode perfor mance, Area Specific Resistance (ASR) will be measured using AC impedance spectroscopy. 1.3 Dissertation Organization The contents of this work are categorized into eight chapters. Chapter 1 outlines the motivation and objectives of this work. Chapter 2 covers a brief introduction on principle of solid oxide fuel cell, solid oxide fu el cell electrolytes for IT-SOF Cs and an overview of impedance spectroscopy. Chapter 3 introduces the novel bi smuth oxide-based electrolyte with double dopants of Dy and W (DWSB). The effect of total dopant concentration and dopant ratio on structure and conductivity is inve stigated to obtain maximum conductivity. Chapter 4 discusses temperatureand timedependent conductivity behavior of these DWSB compositions and also covers conductivity degradation mechanism at intermediate temperatures. In Chapter 5, the approach for enhancing the long term stabilit y of DWSB electrolyte compositions at 500 oC are demonstrated. Chapter 6 introduces new doubly doped bismuth oxide-based electrolytes in

PAGE 19

19 addition to DWSB. Chapter 7 presents the applica tion of this bismuth oxide-based electrolyte in composite cathode materials. Finally, chapte r 8 summarizes the obtained results.

PAGE 20

20 Figure 1-1. Electrical conduc tivity of fluorite oxides [Reprinted from H. Inaba and H. Tagawa, Solid State Ionics 83 (1996) 1. w ith permission from Elsevier].

PAGE 21

21 CHAPTER 2 BACKGROUND 2.1 Solid Oxide Fuel Cell 2.1.1 Principles of Solid Oxide Fuel Cell Figure 2-1 is a schematic diagram to show how solid oxide fuel cells (SOFCs) work. SOFCs are solid state energy conversion devices that convert chemical energy into the electrical energy. Cells are composed of an anode, an el ectrolyte and a cathode. At the cathode side, oxygen molecules combine with electrons from an external circuit to form oxygen ions. The oxygen ions move to the anode side through the electrolyte. At the anode side, hydrogen molecules combine with oxygen ions to form water releasing electrons to the external circuit. The overall reaction is to form water as show n below. Therefore, these devices emit minimal pollution and provide efficient electricity. The open circuit voltage (OCV) of the cell can be expressed by the Nernst equation as follows. 2/1)()( )( ln 22 2 2cathode O anodeH anodeOH oP P P F RT EE (2-1) where Eo is standard voltage, F is Faradays constant. In addition, (PH2O)anode, (PH2)anode are the partial pressure of water vapor and hydrogen at the anode, respectively. (PO2)cathode is the oxygen partial pressure at the cathode. Equilibrium constant K for overall reaction can be expressed by the following equation. 2/1)()( )(2 2 2cathode O anodeH anodeOHP P P K (2-2) The standard Gibbs free energy change for this reaction is calculated using K. o onFEKRTG ln (2-3)

PAGE 22

22 where n is the number of electrons particip ating in the reaction. By manipulating above equations, we can obtain the following equation. anodeOH cathode O anodeHP P P K F RT E )( )()( ln 22 2 22/1 (2-4) Under reversible conditions, the dissociation equilibrium equation of water at the anode side is shown below. 2 2 22 1 OHOH (2-5) Therefore, the equilibrium constant fo r water dissociation is as follows. K P P P KanodeOH anodeO anodeH D1 )( )()(2 2 22/1 (2-6) 2/1)( 1 )( )(2 2 2anodeO anodeOH anodeHP P PK (2-7) Inserting (2-7) to (2-4) give s the following equation. 2/1 2/1)( )( ln 22 2anodeO cathode OP P F RT E (2-8) Finally, it is found that the ope n circuit voltage of an SOFC is dependent on the Po2 difference between two electrodes. 2.1.2 Actual Performance The actual cell potential is decreased from its ideal potential due to several types of irreversible losses. This loss is referred to as polarization or overpoten tial. Figure 2-2 shows a schematic diagram which explains multiple phenomena associated with irreversible losses in an actual fuel cell.

PAGE 23

23 Activation Polarization (activation) This stems from the activation energy of the electrochemical reactions at the electrodes. Ohmic Polarization (ohmic) Ohmic losses are caused by ionic resistance in the electrolyte and electrodes, elec tronic resistance in the elect rodes and contact resistance Concentration Polarization (concentration) This type of polariza tion is due to finite mass transport limitation rates of the reactants and strongly depends on the current density, reactant activity and electrode structure. Therefore, actual voltage can be expressed as the following equation. V = V0 activation ohmic concentration (2-9) where V is the actual voltage, Vo is the theoretical voltage, activation is activation polarization, ohmic is ohmic polarization and concentration is concentration polarization. Among three polarization mechanis ms, ohmic polarization occurs because of resistance to the flow of ions in the electro lyte and resistance to flow of electrons through the electrode. Ohmic losses can be expressed as the followi ng equation because both the electrolyte and electrode obey Ohms law. ohmic = IR (2-1 0) where I is the current flowing through the cell, and R is the total cell resistance, which includes electronic, ionic and contact resistance: R = Relectronic + Rionic + Rcontact (2-11) Any of these components can dominate the ohm ic resistance, depending on the cell type. We refer to the ohmic resistance normalized by the active cell area as the Area Specific Resistance (ASR). ASR has the dimension of cm2. The ASR is a function of the cell design,

PAGE 24

24 material choice and fabrication technique. Th e ohmic polarization can be explained by the following equation using ASR. ohmic = IR=i(ASR) = i L= i(1/ )L (2-12) From Eq. 2-12, it is evident that ohmic polariz ation in the electrolyte is determined by its conductivity and thickness. Th erefore, many efforts have been made to reduce ohmic polarization in two ways from a fabrication as pect and from a material selection aspect. As discussed earlier, thin film t echnology using YSZ electrolytes has been rapidly developed and makes it feasible to produce cost-effective thin films for IT-SOFCs [7-9]. However, zirconia based SOFCs still requires high proc essing temperatures (1200 to 1700 oC) for fabrication and this may affect electrode performance, resulting in lower specific surface area and decreased catalytic activity. Therefore, an alternative electr olyte with higher ioni c conductivity for ITSOFCs has many advantages over fabricating thin YSZ films. 2.2 Solid Oxide Fuel Cell Electrolytes for IT-SOFCs Doped ceria and stabilized bism uth oxides with fluorite-type crystal structures exhibit much higher ionic conductivities than YSZ. Ther efore, these materials have obtained much of attentions for IT-SOFC applications. However, these fluorite type solid oxides have limited thermodynamic properties. Even though aliovalent doped ceria is an excellent candidate to replace YSZ, it shows mixed conduction under reducing atmospheres. This will lower open circuit potential and consequently reduce powder density. Cubic-stabilized bismuth oxides are known to have the highest ionic conductivity. However, their phase and structural stability in the IT range still needs to be examined. In this study, bismuth oxide-based electrolytes will be the primary focus.

PAGE 25

252.2.1 Phases and Conductivity of Pure Bi2O3 Bi2O3 exhibits polymorphism and has and phases [18]. The structure of the various Bi2O3 phases was investigated systematically by Harwig [18, 19]. The high temperature fluorite-related -phase of Bi2O3 was found to be an excellent oxide ion conductor [12-14]. This phase exists between 729 oC and the melting point of bismuth oxide at 824 oC. Below 729 oC the -phase transforms to the low temperature monoclinic -phase [18]. When cooling the -phase, one of two intermediate metastable phases may be formed; the tetragonal -phase at 650 oC or the body centered cubic (bcc) -phase at 639 oC [18]. The structure data of all of the Bi2O3 phases are summarized in Table 2-1. The ionic conductivities of and -phases were systematically measured by Harwig et al. [20, 21]. According to their study, -Bi2O3 exhibits high ionic c onductivity as shown in Figure 2-3. Pure -Bi2O3 exhibits high ionic conductivity due to its relatively open structure which can accommodate a high level of atomic disorder [6]. However, -Bi2O3 transforms to monoclinic -phase on cooling below 730 C. Th erefore, the application of -Bi2O3 is limited to the narrow temperature range of 730 oC-824 oC. 2.2.2 Structure of Pure -Bi2O3 -Bi2O3 has a defect fluorite structure where 25% of the sites in the oxygen sub-lattice are vacant. The high intrinsic defect concentration gi ves high ionic conductivity to this material. The unit cell of cubic bismuth oxide consists of four cations occupying the FCC positions and six anions occupying the eight tetrahedral positions. Th e nature of the arrangement of the six anions and the two anion vacancies has been a subject of controversy. Several contrasting models have been proposed to describe the high temperature cubic phase of pure -Bi2O3 [22-24]. However, none of these models by itself is consistent with the experimental neutron diffraction results. Neutron diffraction studies of cubic bismuth oxides by Battle et al. [25] and

PAGE 26

26 Wachsman et al. [26] showed th at a significant fraction of oxygen ions occupy the 32f interstitial positions. Battle et al. [25] showed that -Bi2O3 has a defective fluorite structure in which 43% of the regular tetrahedral site s (8c) are randomly occupied with 3.44 oxygen ions and the remaining 2.56 oxygen ions per unit cell are displ aced from their regular 8c positions along the <111> directions. In addition, Yashima et al. [27, 28] carried out the maximum entropy method (MEM) [29, 30]-based pattern fitting combined with the Ri etveld refinement [31] using neutron-powderdiffraction data to determine the accurate disorder in -Bi2O3. They firstly conducted Rietveld analysis about three structure models including th e Gattow [23], Harwig [18] and Battle models [25] (Figure 2-4). They obtained the best fit with the observ ed neutron diffrac tion data through the Battle model, compared to other models. The calculated profile with th e Battle model was good agreement with the observed one. The refined crystal parameters ba sed on the Battle model are give n in Table 2-2. The structure parameters was fairly consistent with those in the literature [25]. Secondly, Yashima et al. [27] carried out th e MEM analysis with th e structural factors obtained by the Rietveld analysis using the Battle model. The MEM map provided much information on the complicated disorder of oxide ions in -Bi2O3 compared with previous models. This result demonstrates that th e positional disorder of oxide ions in -Bi2O3 is not fully described by previous simple models consis ting atom spheres. Figure 2-5 shows the MEM nuclear density distribution map on (110) plane which visualizes the structure disorder at 778 oC. It shows that the oxide ions are disordered over a wide area compared with Bi ions in -Bi2O3. Recently, Aidhy et al. carried out molecular dyna mics (MD) simulations to determine and characterize the defect structure of pure -Bi2O3 [32]. They identified the effect of ionic

PAGE 27

27 polarizability on the structure and ionic diffusion of pure -Bi2O3 through MD simulations. They found that the polarizability of the ions is the key determinant of the oxygen diffusion behavior. They created polarizable -Bi2O3 and non-polarizable -Bi2O3 with MD simulation and observed the effect of polarizability on the oxygen diffusion of -Bi2O3. It was found that the polarizable -Bi2O3 shows a continuous increase in the mean-squared displacement (MSD), implying the presence of diffusion, while the non-polarizable -Bi2O3 only had a small amount of atomic motion for an initial period of time and ceased. It is believed that this cessation of oxygen diffusion is due to the formation of vacancy-o rdered structure. They concluded that low polarizability results in a combined ordering of the vacancies in <110> and <111> direction, however, high polarizability allows -Bi2O3 to maintain the disordered state of oxygen sublattice with sustained diffusion. This MD result was al so consistent with the experimental results obtained with dopants having differe nt polarizability [33, 34]. 2.2.3 Stabilized Bismuth Oxide 2.2.3.1 Structure and ionic conductivi ty of doped bismuth oxides It is necessary that high temperature cubic phase be stabilized in order to use bismuth oxide-based electrolytes for IT -SOFCs. Numerous studies have shown that the high conductivity -phase in Bi2O3 can be stabilized at lower temperatur es by the addition of several Lanthanide (Ln3+ or Y3+) dopants [35-37]. However, due to the mism atch in ionic radii between the host and dopant cations, structure stabiliz ation has resulted in the redu ction of ionic conductivity. Verkerk et al. assumed that the stabilization of relatively loose high temperature -phase of Bi2O3 occurs by a contraction of this structure due to the dopant [38]. An investigation was conducted regarding the development of a rela tionship between the ionic radius of the Ln3+ and Y3+ substituent ion and the minimum amount of Ln2O3 or Y2O3 (Xmin) required to stabilize the fcc phase of Bi2O3 [38]. Figure 2-6 shows the relationship between the ionic radius of each

PAGE 28

28 dopant and its corresponding Xmin for phase stabilization. Figure 2-7 also shows the effect of these factors on the conductivity. These results demonstrate that minimu m dopant concentration for cubic-phase stabilization is more important than ionic ra dius to achieve maximum ionic conductivity. In this respect, maximum ioni c conductivity is achieved by lowering Xmin. In case of singly doped bismuth oxide system, the highe st ionic conductivity was obtained by the fcc phase of Er2O3 stabilized Bi2O3 as shown in Figure 2-7. 2.2.3.2 Stabilization with double dopants Besides the singly doped Bi2O3 system, several ternary Bi2O3-based oxides with double dopants have been synthesized and characterized [39, 40]. Meng et al. showed that the fcc structure in Bi2O3-based solid solutions could be stabilized down to room temperature using two rare-earth oxide dopants, with much lower total d opant concentration than that of a singly doped oxide [39]. This cooperative effect was attribut ed to the increase in entropy of the resulting ternary system as a consequence of mixing [39] It was observed that the existence of second dopant in smaller concentration stabilized the fcc st ructure and led to an increase in conductivity, especially in the lower temperatur e regions. Hu et al. also report ed the conductivity variation in Bi2O3-based oxides with different tr ivalent co-dopants [41]. 2.2.3.3 The determination of dopant species The effectiveness of prospective dopants de pends on factors such as ionic radius and dopant polarizability. Shirao et al. investigated the corre lation between the electronic polarizability of an ion and the effective ionic ra dius in lanthanide ions. They observed that the polarizabilty was approximately pr oportional to the cube of ionic radius [42]. Furthermore, Jiang et al. showed that dopant ionic radius affected the crystallogr aphic structure and conductivity [33]. They carried out a systematic study of bism uth oxides which were stabilized with different rare earth oxides, RE2O3 (RE = La, Nd, Sm, Gd, Dy, Ho, Er, Tm, or Yb) as well as Y2O3. Most

PAGE 29

29 samples were synthesized with the sa me composition of 25 mol% dopant, e.g. (RE2O3)0.25(Bi2O3)0.75. Among those dopants, the cubic structure was obtained with Dy2O3, Ho2O3, Y2O3, Er2O3, Tm2O3 and Yb2O3. However, larger-radii rare-earth dopants (La3+ through Gd3+) resulted in formation of the rhombohedral st ructure. This result demonstrates that the structure of stabilized bismuth oxides is dependent on dopant ionic radius. Wachsman et al. found that dopants used in stabilizing the cubic structure of Bi2O3 are typically less polarizable than Bi3+ and their degree of polarizabili ty affects the conductivity and stability of the disordered structure of -Bi2O3 [34]. In order to better understand how the polarizability of dopant cation affects long te rm stability of conduc tivity, the respective electrolyte composition was annealed at 500 oC in air. Within the fcc phase-stability window, they found that the stability for the disordered oxygen sublattice was enhanced as dopant polarizability increased as shown in Figure 2-8. Due to its large radii and high polarizability, Dy provided the greatest stability at 500 oC [33, 34]. Further, recent simulation studies found that polarizability is more important than radii to determine oxygen diffusion in Bi2O3 and high polarizability leads to a disordered lattice [32]. Time constant term was introduced to expre ss time-dependent conductivity decay behavior [43]. Conductivity behavior with time can be represente d by an empirical equation, (t) = ( ) + ( (0) ( ))exp[-(t/ )] (2-13) where t is time, (0) is the initial conductivity, ( ) is the conductivity at infinite time, is the pertinent time constant, and is a dimensionless parameter. Th e time constant is an indicator of decay rate and therefore the kinetic stability of the disordered or the ordered structure can be explained by this time constant ( ). A greater indicates a greater degr ee of stability of the disordered structure. The correlation between the dopant polarizability and the time constant for

PAGE 30

30 aging is shown in Figure 2-9. From this figure, it is clear that there is a strong linear dependence between the dopant polarizability and the kinetics of aging. Thus, highly polarizable cations can be incorporated as dopants into bismuth oxi de to enhance conductivity stability. 2.2.4 Phase Stability Datta et al. examined the (Y2O3)x-(Bi2O3)1-x (xYSB) phase diagram [44] and they showed that the -phase in 25YSB is stable down to low temperatures (~ 400 oC). However, recent work shows that the -phase decomposes to a rhombohedral structure when samples of this composition are annealed at temperatures below 700 oC for long periods of time [45]. Phase destabilization was also observed by Takahashi et al. in 20YSB after annealing [36]. Watanabe indicated that all the stabilized phases in the Ln3+ doped bismuth oxide (Ln3+ = rare earth trivalent cation) reported were the quenched high-temperature stable phases [46]. He emphasized that these quenched -Bi2O3 phases are metastable at lower temperatures (~ below 700 oC). Thus, on annealing at such temperatures these phases transform gr adually into the lowtemperature stable phases. Fung et al. also reported that the cubic phase in the Ln2O3-Bi2O3 and Y2O3-Bi2O3 systems has been found to be unstable below 700 oC and undergoes a transformation to a rhombohedral phase [47-49]. Thus, it was co ncluded that the transformation of the cubic phase into a lower-conductivity rhombohedral phase over some composition ranges may be a general feature of rare-earth oxide (or Y2O3)-bismuth oxide systems. In addition, the high ionic conductivity cubic phase in fact may be a metastable phase below ~700 oC. 2.2.5 Structural Stability Jiang et al. demonstrated that in the case of Ln2O3-Bi2O3 (or Y2O3-Bi2O3) systems, the conductivity degradation can be caused by the form ation of ordered structure as well as the cubic-to-rhombohedral transformation when the sa mples are annealed belo w a certain transition temperature [50]. Wachsman et al. extensively inves tigated this orderi ng phenomenon without

PAGE 31

31 involving phase change using differential therma l analysis, XRD, TEM and neutron diffraction [14, 26, 34, 51-53]. Modeling of ordered structur es based on TEM diffraction patterns indicates a <111> vacancy ordering in the an ion sublattice. In addition, neut ron diffraction experiments [26, 53] demonstrate that the vacancy ordering is acco mpanied by the displacement of almost all the oxygen ions from 8c to 32 s ites (positional ordering). Jiang et al. investigated the tr end in conductivity decay with Er2O3 concentration as shown in Figure 2-10 [50]. It is observed th at the aging rate decreases as Er2O3 dopant concentration increases for 15ESB, 20ESB and 25ESB. These compositions still underw ent large conductivity decay. However, the decay rate was consider ably reduced for 30ESB, 35ESB and 40ESB. In addition, there is no significant difference in decay rate with compos ition at the higher doping concentrations. Jiang et al. performed DSC experiments to m easure the enthalpy change for relaxation of the aged sample with different Er2O3 concentrations as shown in Table 2-3 [50]. The largest measured enthalpy change was 7.79 kJ/mol for 15ESB. The enthalpy change for other compositions is even smaller than that of 15E SB. After the sample has been annealed for 100 hours, it was thought that the aging process was nearly complete. Earnest reported the enthalpy change associated with the transition for pure Bi2O3 as 33.2 kJ/mol [54]. Since this value is much larger than the reverse-aging enthalpy ch ange, Jiang et al. rati onalized that the energy change for the aging process is much smaller than that for a phase transformation, implying that the two processes are completely distinct [50]. 2.3 Impedance Spectroscopy In order to measure the electr ical conductivity of so lid oxide electrolyte, A.C. impedance spectroscopy has been widely used. The equivalent circuit which is suitable for these impedance

PAGE 32

32 spectra will be di scussed. Impedance spectra of CeO2based electrolytes and Bi2O3-based electrolyte will be compared. 2.3.1 Impedance Spectra for Conventional Ceramics It is generally reported that a capacitive doubl e layer is formed at the interface between electrolyte and electrode when a potential gradient is applied to the both electrodes of the cell. The reason for the double layer is attributed to the fact that the electrochemical reaction of interfaces is dependent on time. The equivale nt circuit and impedance spectra for interface reaction are given in Figure 2-11(a). The impedance can be expressed as follows. "'1jYYZY (2-14) Y is the admittance and is the reciprocal of impedance Z Y and Y are the real part and imaginary part of admittance, respectively. Therefore, Y represents resistive component and Y represents capacitive compone nt of the cell response. 1' R Y (2-1 5) dl dlCXY 1" (2-16) R, Cdl and Xdl represent ohmic resistance, capacitan ce of the electrical double layer and reactance, respectively. In addition, represents angular frequency. R RCj XR jRX jXRjYYZYdl dl dl dl 1 "'1 1 1 (2-17) dljwRC R Z 1 (2-18) When electrolyte resistance, Ru is included, the equivalent circuit and impedance spectra are shown in Figure 2-11(b ) and the impedance can be expressed as follows. 222 2 2221 1 1dl dl dl u dl uCR RC j CR R R RCj R RZ (2-19)

PAGE 33

33 From this equation, impedance Z can be separated into real (Z ) and imaginary (Z ) components. 2221 'dl uCR R RZ (2-20) 222 21 "dl dlCR RC Z (2-21) From this equation, the following relation between Z and Z is obtained. )'( 1 )"()'(222 2 2 2 u dl uRZR CR R ZRZ (2-22) 0)"()'()'(2 2 ZRZRRZu u (2-23) 2 2 2) 2 ()"() 2 '( R Z R RZu (2-24) From this equation, when we plot Z on the X axis and Z on the Y axis, (Z Z ) is located on a circle of radius R/2 centered at (Ru+R/2, 0) Conductivity measurement was performed by electrochemical impedance spectroscopy (EIS). The impedance of a system can be calculated by the following equation, )( )( )( tI tV (2-25) where I is current and V is a.c. voltage. The equation can be rewritten in terms of real (Z ) and imaginary (Z ) components. "' iZZZ (2-26) By plotting Z vs. Z one acquires a semicircle where the magnitude of Z along the x-axis indicates the resistance R of th e electrolyte sample. That is, R Z (2 -27)

PAGE 34

34 Electrical conductivity of this electrolyte can be calc ulated from sample geometry and resistance. RA d (2-28) where d is the specimen thickness and A is th e cross-sectional area perpendicular to the current direction in the sample. Information about oxygen ion conduction can be obtained by plotting the conductivity as a function of temperature (Arr henius equation below). )exp( kT E ATa (2-29) The pre-exponential factor A is a function of geometry, crystal lattice spacing, attempt frequency of ion hopping and en tropy. The activation energy (Ea) is a function of vacancy mobility and vacancy distribution. The slope of this Arrhenius plot represents the activation energy of ionic conductivity for the system. This slope should be linear for single conduction processes. 2.3.2 The Comparison of Impedance Response Impedance spectroscopy is often used to sepa rate the bulk, grain boundary and electrode processes of polycrystalline ceramic materials with ionic or mixed c onduction [56]. The high frequency component usually corresponds to the tr ue bulk properties, as measured for single crystals [57], the intermediate frequency range is ascribed to resistive grain boundaries and the low frequency range corresponds to electrode processes or processes occurring at the material/electrode interface. The bulk relaxation frequency fB is often at least two orders of magnitude higher than the grain boundary relaxation frequency fgb. The electrode relaxation frequency fel is also much smaller than fgb. In these conditions, the Nyquist plot should show nearly separate semicircles for the bulk grain boundary and electrode terms.

PAGE 35

35 For example, in case of CeO2-based electrolytes, we usually see three-arc responses which are related to the bulk, grain boundary and electrode s. If the frequency range is not limited, we may see every process which is related to this electrochemical reaction by impedance spectroscopy. However, due to experimental lim itation, we only see some parts of the whole response depending on the temperature. For modeling of this impedance spectrum, an equivalent circuit was made which is suitable for this system. When we have a perfect semicircle, a capacitor can be used in the equiva lent circuit model. However, in real circumstances, the arc is often depressed and this leads to the necessity of a constant phase element (CPE). In this case, a constant phase element (CPE) was used instead of a pure capacitor for modeling in order to apply this equivalent circuit to real impedance data. A constant phase element is equivalent to a distribution of capacitors in parallel and it arises due to th e microstructural inhomogeneities within the sample [58, 59]. 2.3.2.1 CeO2-based electrolyte Zhang et al. examined the ionic conductivity of ceria-based solid solutions as a function of silica contents [60]. To a larg e extent, the blocking behavior of grain boundaries is mainly attributed to the presence of th in siliceous films. The deleterious grain boundary behavior arising from SiO2 impurities has been recognized in zirconia-based electrolytes for several decades and in ceria-based ceramics for over ten years. The impedance characteristics of the Ce1-xGdxO2ceramics, with different levels of SiO2, were investigated. Figure 2-12 shows the impedance spectra of the Ce0.8Gd0.2O2(CGO20) ceramics with di fferent levels of SiO2, measured at 350 oC in air. It is found that the grain boundary impedance increases with SiO2 content. Specifically, the sample CGO20 with ~ 3000 ppm SiO2 has a very large grain bound ary arc compared to the sample with relatively low SiO2 content, indicating that SiO2 impurity is extremely detrimental to the grain boundary conduction of ceria-based electrolytes.

PAGE 36

362.3.2.2 Bi2O3-based electrolyte Duran et al. developed Bi2O3-Y2O3 (or Er2O3) systems and examined the electrical properties of these oxygen ion conductors [61]. Concerning the AC impe dance plane behavior they found for this material the absence of a grain boundary response. Two arc responses have been reported for bismuth oxides doped with Er2O3, Y2O3 and MoO3 [61-63]. Figure 2-13 shows impedance plots of 20 mol% Er2O3 stabilized Bi2O3 (20ESB) at 350 oC and 500 oC obtained from Duran et al [61]. Based on th is figure, in contrast to CeO2-based electrolyte, bismuth oxidebased electrolytes show no eviden ce of a grain boundary contributi on to their conductivity. This could be due to very small grain boundary impe dance or to the similar time constants for both grain interior and grain boundary impedances. Unlike zirconia a nd ceria, bismuth oxide is very reactive and has considerable solid solubility with silica and alumina, two common constituents of grain boundary phases in ceramics. Thus it is possible that bism uth oxide segregated at grain boundaries has a self-cleaning effect [59]. Due to this property of bismuth oxide, the impedance response of bismuth oxide based electroly te is clearly different from that of ceria based electrolyte. This can be considered a striking feature of bismuth oxi de-based electrolyte.

PAGE 37

37 2 22 2 1 OeO eOHOH 22 2 2 OHOH2 2 22 1 Cathode Anode Cell 2 22 2 1 OeO eOHOH 22 2 2 OHOH2 2 22 1 Cathode Anode Cell Anode Cathode Electrolyteair fuel O2-VO 2e-2eA Anode Cathode Electrolyteair fuel O2-VO VO 2e-2eA Figure 2-1. Principle of operation of a solid oxide fuel cell with hydrogen as a fuel and oxygen as an oxidant.

PAGE 38

38 Figure 2-2. Ideal and actual fuel ce ll voltage/current characteristic [17]

PAGE 39

39 Figure 2-3. Electri cal conductivity, log versus temperature of Bi2O3 [Reprinted from N.M. Sammes, G.A. Tompsett, H. Nafe and F. Aldinger, J Eur Ceram Soc. 19 (1999) 1801. with permission from Elsevier].

PAGE 40

40 Figure 2-4. Structural models for -Bi2O3. (a) Gattow model (b) Harwig model (c) Battle model and (d) Rietveld refinement and the ME M-based pattern fitting. Red and yellow spheres denote the bismuth and oxygen atom sites, respectively [Reprinted from M. Yashima and D. Ishimura, Chem Phys Lett 378 (2003) 395. with permission from Elsevier].

PAGE 41

41 Figure 2-5. Scattering amplitude distribution on (110) plane of -Bi2O3 at 778 oC with white contours in the range from at 0.3 to 4.0 fm/3 (0.2 fm/3 step). Oxide ions have a disorder along [111] directions. [Reprinted from M. Yashima and D. Ishimura, Chem Phys Lett 378 (2003) 395. with permission from Elsevier].

PAGE 42

42 0.098 0.1000.1020.104 0.106 0.15 0.20 0.25 0.30 0.35Rion(nm)XminYb3+Er3+Y3+Dy3+Gd3+ 0.098 0.1000.1020.104 0.106 0.15 0.20 0.25 0.30 0.35Rion(nm)XminYb3+Er3+Y3+Dy3+Gd3+ Figure 2-6. Xmin, the minimum value of x required to stabilize the f cc phase in (Bi2O3)1-x(Ln2O3)x as a function of the ionic radius (Rion) of Ln3+ [38].

PAGE 43

43 0.098 0.1000.1020.104 0.106Rion(nm) -3 -2 -1 Log (S/cm)700oC 500oCYb3+Er3+Y3+Dy3+Gd3+ 0.098 0.1000.1020.104 0.106Rion(nm) -3 -2 -1 Log (S/cm)700oC 500oCYb3+Er3+Y3+Dy3+Gd3+ Figure 2-7. Electri cal conductivity, log of (Bi2O3)1-x(Ln2O3)x for Xmin as a function of the ionic radius of the substituted Ln3+ [38].

PAGE 44

44020406080100120 0.0 0.2 0.4 0.6 0.8 1.0 Dy Yb Tm Er Ho Relative conductivity (t/o)Time (hour)Y Figure 2-8. Relative co nductivity, as a fuction of time, at 500 oC, for different dopants [33].

PAGE 45

45 Figure 2-9. Relationship between conductivity decay time constant and dopant polarizability [43]. 1 10 100 1000 1.801.902.002.102.202.302.40TIME CONSTANT (h)POLARIZABILITY AYb Er Ho Dy

PAGE 46

46 020406080100120 0.0 0.2 0.4 0.6 0.8 1.0 15ESB 20ESB 25ESB 30ESB 35ESB 40ESB Relative conductivity (t/o)Time (hour) Figure 2-10. Relative conductivity as a func tion of time, at 500 oC, for different Er2O3 concentrations [50].

PAGE 47

47 Figure 2-11. Complex plane diagram (a) for a paralle l RC circuit, and (b) in (a) with the addition of Ru, an uncompensated electr olyte resistance [55]. )(" ohmsZ )(' ohmsZ(b) Ru R+Ru )(" ohmsZ )(' ohmsZR (a) =1/RCdl

PAGE 48

48 051015202530 0 5 10 15 20 25 Grain Boundary 200 ppm SiO2 3000 ppm SiO2Z''103 (cm)Z'103 (cm) 350oCGrain Figure 2-12. Impedance spectra of CGO20 with ( ) 200 ppm, and ( ) 3000 ppm SiO2, measured at 350 oC in air [60].

PAGE 49

49 012345 0 1 2 3 4 ''103 (cm)''103 (cm)'103 (cm) (a) 350oC 0246810 0 2 4 6 8 ''10 (cm)'10 (cm) (b) 500oC Figure 2-13. Impedance complex plane plots of 20ESB [61].

PAGE 50

50 Table 2-1. Structure data of the Bi2O3 phases [18] Phase Phase stability temperature range (oC) < 729 729-824 330-650 500-639 Temperature 298 1047 916 298 Structure monoclinic fcc tetragonal bcc Space Group c/P21 Fm 3m c24P1 I23 Space Group Number 14 225 114 197 a () 5.8496 5.6595 7.738 10.268 b () 8.1648 c () 7.5101 5.731 () 112.977 May persist to room temperature.

PAGE 51

51 Table 2-2. Structure parameters of -Bi2O3 measured at 778 oC [27] Atom Site occupancy Occupation number per unit cell x y z U (2) Bi 4 a 1 4 0 0 0 0.078(2) O1 8 c 0.23(22)a 1.84 0.25 0.25 0.25 0.08(5) O2 32f 0.13(5)a 4.16 0.335(21) = x (O2) = x (O2) 0.121b

PAGE 52

52 Table 2-3. Enthalpies of (Bi2O3)1-x(Er2O3)x [50] X (%) HR (kJ/mol) H (T<600 o C) (kJ/mol) H (T>600 o C) (kJ/mol) H (T<600 o C)H (T>600 o C) (kJ/mol) 15 7.79(47) 121.(0) 21.(4) 99.(6) 20 5.80(46) 119.(6) 62.(8) 56.(8) 25 3.54(34) 117.(7) 75.(8) 41.(9) 30 0.700(2) 117.(5) 87.(7) 29.(8) 35 0.346(2) 115.(3) 100.(8) 14.(5) 40 0.214(1) 115.(5) 111.(1) 4.(4) For relaxation of the order state ( HR), following a 500 oC anneal for 100 h; conductivity below H (T<600 o C) and below H (T>600 o C), the order-disorder transition; and the difference between the two, H (T<600 o C)H (T>600 o C)

PAGE 53

53 CHAPTER 3 EFFECT OF TOTAL DOPANT CONC ENTRATION AND DOPANT RATIO ON CONDUCTIVITY OF DWSB COMPOSITIONS 3.1 Introduction Cubic -Bi2O3 exhibits the highest known oxygen-ion conductivity [12-14, 34]. The conductivity of -Bi2O3 is two orders of magnitude higher th an that of stabilized zirconia at corresponding temperatures [10]. -Bi2O3 has high ionic conductivity du e to its large number of highly mobile oxygen vacancies. The high mobility is attributed to the weak Bi-O bond and the high polarizability of Bi3+ with its lone-pair 6s2 electrons [34]. However, pure -Bi2O3 transforms to monoclinic -phase on cooling below 730 C, resulting in a greatly reduced conductivity and melts above 824 oC [64]. Therefore, the application of -Bi2O3 is limited to the narrow temperature range of 730 oC-824 oC. Nevertheless, the hi gh temperature phase of -Bi2O3 can be stabilized down to room temperature by d oping with other oxides [65]. However, due to the mismatch in ionic radii be tween the host and dopant cations, phase stabilization lowers the ionic conductivity. Verkerk et al. found that, to st abilize the cubi c phase of Bi2O3 to room temperature, it was necessary to dope Bi2O3 with cations with smaller ionic radii than Bi3+. The stabilization occurs because such dopant cations induce a c ontraction of the open structure of -Bi2O3 [38]. They obtained the highest ioni c conductivity with (ErO1.5)0.2(BiO1.5)0.8 (20ESB) because doping with Er3+ allowed for the lowest dopant concentration needed to stabil ize the cubic structure in a single dopant system [38, 66]. In addition, several doubly doped Bi2O3 electrolyte systems have been synthesized and characterized to determine the optimal dopant concentration and the solid solution range [39, 40]. In one such study, Meng et al. showed that with two dopants, the cubic phase of bismuth oxide could be stabilized with a lower total dopant concentrat ion compared to the single dopant

PAGE 54

54 system [39]. It is thermodynamically favorable to maintain its high-temper ature fcc phase as long as the dopant concentration is sufficient because a large negative entropy change is accompanied when highly symmetric structure with random distribution of atoms at high temperatures transitions to a lower symmetric structure at lower temperatures. The use of two stabilizing dopants increases the entropies significantly. Therefore, double dopant strategy allows much lower dopant concentration to stabilize the fcc phase down to room temperature [39]. Among prospective dopants, Dy and W were select ed for dopants to st abilize the structure and improve the conductivity. Based on the results obtained from Wachsman et al. [33, 34], Dy was chosen as a dopant. Other stud ies have shown that tungsten also stabilizes the fcc structure of Bi2O3 and has high polarizability [67-69]. Watanabe et al. also showed that Bi7WO13.5 (=7Bi2O32WO3, 22.22 mol% WO3) has high oxygen ion conductivity with a lower activation energy compared with other cubic stabilized bismuth oxides [69]. Accordingly, in our previous work [13, 70] we developed bismut h oxide electrolytes doubly doped with Dy2O3 and WO3 (DWSB) that were found to have higher c onductivity than 20ESB. However, the effect of total dopant c oncentration and dopant ratio on structure and conductivity of bismuth oxide based electrolytes was still not fully investigated. In this study, dopant ratio and total dopant c oncentration were systematical ly manipulated to obtain an optimum dopant concentrati on for maximum conductivity. 3.2 Experimental Procedure 3.2.1 Preparation of Electrolyte Samples Figure 3-1 shows the flow ch art of the typical powder synthesis. Disks of (DyO1.5)x(WO3)y-(BiO1.5)1-x-y (xDyWSB) composition were fabricated by solid state synthesis. Disks of (ErO1.5)0.2(BiO1.5)0.8 (20ESB) and (GdO1.5)0.1(CeO2)0.9 (10GDC) were also prepared for comparison of conductivity results. First, six different compositions were investigated;

PAGE 55

55 (DyO1.5)0.15-x(WO3)x(BiO1.5)0.85 (where x=0.05, 0.075, 0.10) and (DyO1.5)0.10-x(WO3)x(BiO1.5)0.90, (where x=0.03, 0.05, 0.07). The three former compositions, with a fixed 15 mol% total dopant concentrations, are 10D5WSB, 7.5D7.5WSB and 5D 10WSB. The latter th ree compositions, with a fixed 10 mol% total dopant concentrat ions, are 7D3WSB, 5D5WSB and 3D7WSB. Subsequently, various compositions with 2:1 mo l dopant ratio between Dy and W were also prepared to determine the effect of total dopant concentration on conductivity: (DyO1.5)2x(WO3)x(BiO1.5)1-3x (where x = 0.040, 0.045, 0.050, 0.055, 0.060 and 0.070). These compositions are referred as 8D4WSB, 9D 4.5WSB, 10D5WSB, 11D5.5WSB, 12D6WSB and 14D7WSB. A stoichiometric mixture of Bi2O3 (99.9995% pure), Er2O3 (99.99% pure), Dy2O3 (99.99% pure) and WO3 (99.8% pure), from Alfa Aesar, were mixed and ball-milled with zirconia ball media in a high-density polyethyl ene bottle for 24 hours. Similarly a stoichiometric mixture of CeO2 (99.9% pure) and Gd2O3 (99.99% pure) from Alfa Ae sar were used for 10GDC composition. After drying, the mixed powders of DWSB or 20ESB were calcined at 800 C for 16 hours. For 10GDC, the mixed powders were calcined at 1450 oC for 10 hours. Agglomerated powders were ground using mortar and pestle and sieved using a 325 m mesh to get uniform particle sizes. The powders were pressed uniaxially into a disk -shaped 8 mm diameter die using 40 MPa to get disk-shaped pellets and pressed again to increase th e density of the pellets through cold isostatic pressing at 200 MPa. The pellets of DWSB or 20ESB were subsequently sintered in air at 890 C for 16 hours. For 10GDC the pellets were sintered at 1550 oC for 10 hours. The sintered pellets of bismuth oxide base d electrolytes had densities 94% 2% of theoretical densities and were polished to acqui re an even surface and the desired thickness (3

PAGE 56

56 mm). Engelhard gold paste mixed with isopropanol was brushed onto both sides of the DWSB or 20ESB electrolytes and the organic additives were evaporated at 120 oC for 1 hour. Subsequently, the electrode s were sintered at 800 oC for 1 hour. For 10GDC, Pt paste (CL115349, Heraeus) was used as the electrode and sintered at 900 oC for 1 hour. Pt wires with circular Ag mesh were attached to the cells by mechanical contact. 3.2.2 X-Ray Diffraction Analysis The lattice structure of the calcined powders was identified by means of X-ray diffraction analysis (XRD, Philips APD 3720). XRD pattern was obtained using CuK radiation at room temperature between 20o and 70o (2 ). Extrapolation method using Nelson-Riley function was performed to calculate the lattice parameter of cubic-stabilized DWSB compositions. Details on this method are given in Appendix B. CrystalMak er & CrystalDiffract Programs were used for better phase identification. Details on the use of this program are given in Appendix C. In addition, High Temperature XRD (HT-XR D, Philips APD 3720 HT with heat supply under He gas condition) was used to obt ain precise lattice parameter at 500 oC, 700 oC and 760 oC. The continuous scan was performed to obtain (111) peak of each composition between 26o and 30o (2 ) with step size of 0.01o. The ProFit For Windows program (Philips Profile Fit) was used to remove the contribution from K 2 radiation and lattice parameter was calculated with the wavelength of CuK 1. 3.2.3 Conductivity Measurements Conductivity measurements were obtained from a two-point probe electrochemical impedance spectroscopy (EIS) using a Solartron 1260 over the frequency range of 0.1 Hz to 10 MHz. The frequency response anal yzer was used in standalone mode for unbiased testing and interfaced to a computer using Zplot software. Due to the small sample impedances at high

PAGE 57

57 temperature, a nulling technique was necessary to remove artifacts caused by inductive responses of the test leads and the equipment. This was achieved by measuring the impedance of the leads without a sample and then subtracting it from th e sample measurements. The measurements were performed between 200 oC and 700 C in air. For each composition, three different samples were used to determine the error for measured conductivity. 3.3 Results and Discussion 3.3.1 Structure To find the ideal DWSB composition, total dopa nt concentration was varied according to three preset dopant content ratios of 2:1, 1:1 and 1:2. Figure 3-2 shows XRD patterns of various calcined DWSB compositions. Of these six compositions, single phase was obtained for 10D5WSB, 7.5D7.5WSB and 7D 3WSB. Mixed phases were f ound for 5D10WSB, 5D5WSB and 3D7WSB. Figure 3-2 shows th at total dopant concentration and dopant ratio affect phase purity of calcined DWSB compositions. It was analyzed that the secondary phases from 5D5WSB, 3D7WSB an d 5D10WSB are the mixtures of a tetragonal phase (7Bi2O3WO3 Bi14WO24, S.G. 88) and an orthorhombic (Bi2O3WO3 Bi2WO6, S.G. 61) phase. Depending on the composition, relative amount of each secondary phase was different. There is a tendency to have single fcc phase as total dopant concentrat ion and dopant ratio (Dy/W) increases as shown in Figure 3-2. This indicates that there exis ts a minimum dopant concentration necessary to stabilize the cubic structure of Bi2O3. It also shows that WO3 has more limited solubility than Dy2O3 into Bi2O3. Indeed, Verkerk et al. examined Dy2O3 doped Bi2O3 system (DSB) and obtained fcc structure with 28.5-50 mol% Dy2O3 [38], while Takahashi et al. investigated WO3 doped Bi2O3 system (WSB) and attained fcc structure with 22-28 mol% WO3 [67]. From a viewpoint of solid solution range, the WO3 doped Bi2O3 system had a narrower range for fcc phase stabilization

PAGE 58

58 than the Dy2O3 doped Bi2O3 system [38, 67]. The XRD results confirm the literature finding. Therefore, it is expected that Dy content should be larger than th e W content needed to attain a pure fcc phase in doubly-doped bismuth oxide. Based on the XRD pattern of Figure 3-2, DWSB electrolytes with the same 2:1 mol dopant ratio between Dy and W were then prepar ed (6D3WSB, 7D3.5WSB, 8D4WSB, 9D4.5WSB, 10D5WSB, 11D5.5WSB, 12D 6WSB and 14D7WSB). The total dopant concentration ranges from 9 mol% up to 21 mol%. Figure 3-3 shows XR D patterns of these DWSB electrolytes. All compositions were found to be pure fcc pha se except for 6D3WSB and 7D3.5WSB. Therefore, 8D4WSB is the composition whic h has the minimum dopant concentration for stabilizing fcc structure, with a 2:1 dopant content ratio. This re sult shows that the reduction of total dopant concentration needed to stabilize the cubic struct ure was achieved down to 12 mol% for DWSB compared with DSB (28.5 mol%), WSB (22 mol%) and ESB (20 mol%). The lattice parameters of various DWSB co mpositions with 2:1 mo l dopant ratio were determined from the (111) peak position of XRD patterns at room temperature and high temperatures. Figure 3-4 shows the lattice parame ters of various DWSB compositions measured at room temperature (RT) and 760 oC. At both temperatures, the latt ice parameters decrease with increasing total dopant concentration becau se both dopants are smaller than the Bi3+ radii [71, 72]. We also found that the decrease of lattice pa rameter was linear with dopant concentration, obeying Vegards law, within the fcc stability window. The lattice parameters of DWSB compositions were also compared with those of other singly doped Bi2O3 compositions at room temperature in Figure 3-4. For single dopant compositions, (DyO1.5)0.285(BiO1.5)0.715 (28.5DSB) and (WO3)0.22(BiO1.5)0.78 (22WSB) have the minimum amount of dopant to stabilize the fluorite structure, respectively. From the literature,

PAGE 59

59 the lattice parameter of 28.5DSB is about 5.49 [38], while the lattice parameter of 22WSB is about 5.50 at room temperature [73]. Extrapolating the lattice pa rameter of DWSB to the same total dopant concentrations, it is clear that the double dopant re sults in a significantly larger lattice parameter than is obtained from either of these dopants when they are singularly doped. The lattice parameters of various ESB co mpositions (15ESB, 20ESB, 25ESB and 30ESB) at room temperature were also added for comparison [33]. It is interesti ng to note that the slope of the linear fit is the same for the DWSB compositions and the ESB compositions at room temperature. However, again we can see that the double doping results in a structure with a larger lattice parameter than singly doped Bi2O3, resulting in a larger intercept when the dopant concentration is extrapolated to zero ( -Bi2O3). In addition, the lattice parameters were then determined at a temperature where -Bi2O3 is thermodynamically stable (760 oC) through High Temperature XRD (HT-XRD) for 8D4WSB, 10D5WSB, 12D6WSB and 14D7WSB compositions. The slop e of the linear fit at 760 oC is comparable to, but slightly higher than that of the linear fit at room temperature for DWSB compositions. This may result from a larger ther mal expansion of the structure with decreasing total dopant concentration at 760 oC. When this linear fit was extrapolated to pure -Bi2O3 (dopant concentration = 0) at 760 oC, the calculated la ttice parameter was 5.660 This value is comparable to the lattice parameter of pure -Bi2O3 ( = 5.654 ) obtained from Yashima et al. at 760 oC [74]. Therefore, this indi cates that the doubly doped DWSB more closely matches the inherent structure of -Bi2O3 than any of the singly doped compositions. 3.3.2 Conductivity Figure 3-5 shows the typical impeda nce spectra for 8D4WSB at (a) 200 oC and (b) 500 oC. The conductivities were calcula ted from bulk resistance values measured at the frequency corresponding to the appropriate minima in the Nyquist plots. Figure 3-5(a) shows the typical

PAGE 60

60 bulk impedance consisting of a single semicirc le and the electrode impedance at lower frequency. As the temperature increased, the high frequency semicircle attributable to the bulk resistance was no longer resolvable as shown in Figure 3-5(b). The bulk resistance values were then taken from the high frequency intercept. Electrolyte conductivity of each composition was converted from the bulk resistance measured and sample dimensi ons. It is worthwhile to note that two different current range s are used when conductivity was measured using a Solatron over the entire frequency range. Above the transition frequency, the high frequency current range is used, while below the transition frequency, the lo w frequency range is used. In general, if the current range is set to Auto, the equipment will determine the correct current range based on the actual current present in the cell. The scattering of data can sometimes be seen at a certain frequency because of integration failure with Au to default setup. By selecting the adequate current range manually in both high frequency re gion and low frequency re gion, we could obtain better impedance spectra w ithout scatting of data. The bulk conductivities of three DWSB electrolytes ( 10D5WSB, 7.5D7.5WSB and 7D3WSB) are plotted in Figure 3-6. Of these compositions, 10D5WSB showed the highest conductivity. The conductivity of 10D5WSB was 0.487, 0.066 and 2.23-4 S/cm at 700, 500 and 300 oC, respectively. Table 3-1 shows the conductiv ity activation energy and lat tice parameter of 10D5WSB, 7.5D7.5WSB and 7D3WSB. It is reported that cubic stabili zed bismuth oxides go through an order-disorder transition, which is reflected by a change in conductivity activation energy, at a temperature in the neighborhood of 600 oC [12-14, 34, 66]. That is, above the order-disorder transition temperature, the disord ered state of oxygen sublattice is maintained. On the other hand, below this transition temperature, the oxygen sublattice becomes ordered. Even though

PAGE 61

61 10D5WSB and 7.5D7.5WSB have the same tota l dopant concentrations the 10D5WSB has a higher conductivity than 7.5D7.5WSB throughout the entire temperature range. On the other hand, 10D5WSB had a slightly larger disparity between the two activation energies than 7.5D7.5WSB. This may be attributed to the rela tively smaller lattice parameter of 7.5D7.5WSB composition and difference in bond strength be tween Dy-O and W-O. This observation demonstrates that the dopant ratio plays an im portant role in determin ing ionic conductivity. Figure 3-6 shows that 7D3WSB experienced an abrupt change in conductivity and activation energy. Above 600 oC, the conductivity is closer to 10D5WSB and below 600 oC, the conductivity is closer to 7.5D7.5WSB. This implies that 7D 3WSB is not stable in the lower temperature range due to the small total dopant c oncentration. Therefore, from the XRD pattern and conductivity behavior, we concluded that it was necessary to have >10 mol% total dopant concentration in order to obtain DWSB stabilized in the fcc structure. Figure 3-7 shows the conductivity values of various DWSB compositions with 2:1 mol dopant ratio as a function of tota l dopant concentra tion at (a) 700 oC and (b) 500 oC. Both figures show a trend wherein th e highest conductivity is achieve d with the lowest total dopant concentration. In addition, it is apparent that conductivity increases linearly as total dopant concentration decreases with fixe d dopant ratio for both temperatures. When this linear fit of the data in Figure 3-7(a) was extrapolated to pure -Bi2O3 at 700 oC, the calculated conductivity value was 1.1 S/cm. Harwig et al measured the conductivities of and phases of Bi2O3 [21]. From his plot, the conductivity of pure -Bi2O3 at 700 oC is about 1 S/cm which is very comparable to our calculated value from extrapolation. However, extrapolation at 500 oC is not possible because the conductivity of pure -Bi2O3 is not available at that temperature due to the cubic to monoclinic phase transition.

PAGE 62

62 Among DWSB formulations tested in this study, 8D4WSB is the composition with minimum total dopant concentration for stabilizing the fcc Bi2O3 phase down to room temperature. Single dopant systems of (Dy2O3)-(Bi2O3) and (WO3)-(Bi2O3) have also been studied and the highest conductivity was achieved with the lowest dopant concentration for cubic structure stabilization in either cas e [38, 67]. The conductivity of 28.5DSB is 0.144 S/cm and 0.0071 S/cm at 700 and 500 oC, respectively [38]. Likewise, the conductivity of 22WSB is 0.062 S/cm and 0.01 S/cm at 700 and 500 oC, respectively [67]. Figure 3-8 shows the conductivity depende nce of DWSB compositions on lattice parameter at 700 oC and 500 oC for both room temperature and HT XRD. The lattice parameters of 8D4WSB, 10D5WSB, 12D6WSB and 14D7WSB compositi ons were obtained at 700 oC and 500 oC through HT-XRD. The conductivity at 700 oC increases linearly as the lattice parameter at 700 oC and room temperature increases. Similarly the conductivity at 500 oC increases linearly as the lattice parameter increases at 500 oC. In order to compare DWSB with -Bi2O3 we compared these results with that of Yashima et al. [74] and Harwig et al. [21] at 700 oC. Harwig measured the conductivity of pure Bi2O3 in the temperature range 300 800 oC in the heating and cooling directions. In the heating direc tion, due to the monoclinic to cubic phase transition at 729 oC, the conductivity of pure -Bi2O3 was not available at 700 oC. However, in the cooling direction, due to th e temperature delay of the cubic to monoclinic phase transition, the conductivity of pure -Bi2O3 was obtained at 700 oC. Yashima determined the lattice paramete r as a function of temperature for pure -Bi2O3 from 778 oC to 740 oC. Extrapolating Yashimas data to 700 oC results in a lattice parameter of 5.648 Extrapolating our DWSB conductivity re lationship to a lattice parameter of 5.648 results in a calculated co nductivity of 1.02 S/cm at 700 oC. This compares very favorably with

PAGE 63

63 Harwigs conductivity of 0.99 S/cm at 700 oC and demonstrates that the conductivity of DWSB linearly approaches that of -Bi2O3 as the lattice parameter incr eases, without a shift in the intercept. The bulk conductivities of various DWSB electrolytes are plotted as log vs. 1000/T in Figure 3-9. Note the conductivity of 8D4WSB is two times highe r than we previously reported (See Appendix A for details). For comparison we prepared 20ESB and 10GDC by a similar solid state synthesis route and measured their conductivities as descri bed earlier. The conductivity of (YO1.5)0.1(ZrO2)0.9 (10YSZ) [75] was also added to Figure 3-9 as a contextual reference. Figure 3-9 shows that above 400 oC DWSB compositions with the sa me 2:1 mol dopant ratio have higher conductivities than 20ESB, 10GDC and 10YSZ. 20ESB in this work had comparable conductivi ties as those reported by Verkerk et al. [66] and, 10GDC in this work also had similar grain ionic conductivity as those reported by Omar et al. [58]. This not only indicates that the conduc tivity values of 20ESB and 10GDC prepared in this work are comparable to literature value but also validates our e xperimental results for DWSB. The conductivities of 8D4WSB, the highest conductivity composition, are 0.569, 0.098 and 2.50-4 S/cm at 700, 500 and 300 oC, respectively. At 500 oC, 8D4WSB is 4 times more conductive than 20ESB, 10 times more conductive than 10GDC and 100 times more conductive than 10YSZ. These results coul d allow significant re duction in SOFC operation temperature thereby making this DWSB composition a very promising electrolyte for low temperature applications. The temperature dependence of ionic conduc tivity can be expre ssed by the following empirical equation.

PAGE 64

64 )exp(kT E ATa (3-1) where is the oxygen ion conductivity, A is a pre-exponential constant, T is the absolute temperature, k is the Boltzmann constant and Ea is the activation energy for oxygen migration. Therefore, activation energy and pre-exponential terms are crucial determinants of the ionic conductivity of a solid. As mentioned earlier, cubi c stabilized bismuth oxides undergo a change in conductivity activ ation energy at ~ 600 oC [34, 66]. DWSB compositions also show a change in conductivity activatio n energy. In this study, activati on energy was calculated in two temperature regions at 600 oC and at 600 oC. Some studies for si ngly doped bismuth oxides have calculated two different activation energies at 600 oC and 550 oC [38, 66]. Orderdisorder transition temperatur e in Arrhenius plot could be slightly different around 600 oC depending on dopant composition. It could be m eaningful to obtain the activation energy at 600 oC and 550 oC and to identify the effect of dopant concentration on th e activation energy for these DWSB compositions. Table 3-2 shows the values of activat ion energy measured below and above 600 oC. The conductivity activation ener gy was plotted as a function of to tal dopant concentration in Figure 3-10. The conductivity activation en ergy in the high temperature re gion exhibits both an opposite trend and an order of magnitude greater effect as a function of dopant concentration compared with that in the low temperature region. The disparity between these two activation energies below and above 600 oC decreases when the total dopant conc entration increases. When the data in the high and low temperature regions are extrap olated to higher dopant concentration, we find that DWSB should have a single activation energy of 1.035 eV at 30.8 mol%. Extrapolation of the high temperature region data (12 mol% ~ 18 mol%) to 0 mol% dopant concentration (pure Bi2O3) results in a calculated conduc tivity activation energy value of 0.29 eV. This value is

PAGE 65

65 comparable to the activation energy of pure -Bi2O3 ( 0.3 eV) measured by Harwig et al. [21]. Thus, at high temperature when the anion sublattice is disordered the conductivity activation is due primarily to migration and as the dopant co ncentration decreases to 0 mol% the conductivity activation energy approaches that of pure -Bi2O3. Verkerk et al. showed that the activation ener gy of stabilized bismut h oxides is influenced by the dopant type and dopant concentration [76] and believed this was because the Ln-O (Ln = rare earth or Y) bond is much st ronger than the Bi-O bond. Howeve r, Figure 3-8 also shows that the conductivity increa ses with lattice parameter and that the conductivity in the high temperature region extrapolates to that of -Bi2O3. Therefore, in this temperature region it may be that lattice strain plays a larger role than bond strength. In contrast the lower temperature region exhibits different trends and is also the region where def ect ordering occurs [26, 52]. Therefore, the bond strength issues pointed out by Verkerk et al. [ 76] may be more prevalent in this region. The relation between pre-exponential terms and total dopant concentration was also investigated. Table 3-3 shows pre-exponen tial terms measured for th e low and high temperature regions. The pre-exponential term is also plotted as a function of total dopant c oncentration in Figure 3-11. What is surprising is that even though the cond uctivity is higher in the high temperature region, the pre-exponen tial term is in fact lower in this temperature region for all compositions investigated. As with the conductivity activation energy, the trend with dopant concentration is opposite for the low and high temperature regions. The pr e-exponential term decreases with increasing total dopant concentration in the low temperature region, wh ile it increases in the high temperature region. For both regions the c onductivity decreases with increasing dopant

PAGE 66

66 concentration. Therefore, in the high temperat ure region the increase in pre-exponential term with dopant concentration is overshadowed by the increase in conductivity activation energy with dopant concentration. In contrast, in th e low temperature region the small decrease in conductivity activation ener gy with dopant concentration is overshadowed by the much larger decrease in pre-exponential term with dopant concentration.

PAGE 67

67 Figure 3-1. Flow chart for the powder synthesi s using conventional so lid state reaction.

PAGE 68

68203040506070 (a)3D7WSB 5D5WSB 7D3WSB(400) (222) (311) (220) (200) (111) Intensity (arb. unit) 2(degree) 203040506070 (b)5D10WSB 7.5D7.5WSB 10D5WSB(400) (222) (311) (220) (200) (111) Intensity (arb. unit) 2(degree) Figure 3-2. X-ray diffraction pattern s of DWSB (a) with a fixed 10 mol% and (b) with a fixed 15 mol% total dopant concentration.

PAGE 69

69203040506070 7D3.5WSB 6D3WSB(400) (222) (311) (220) (200) (111) 14D7WSB 12D6WSB 11D5.5WSB 10D5WSB 9D4.5WSB 8D4WSBIntensity (arb. unit)2 (degree) Figure 3-3. X-ray diffraction patte rns of DWSB with 2:1 mol dopant ratio between Dy and W.

PAGE 70

70 036912151821242730 5.46 5.48 5.50 5.52 5.54 5.56 5.58 5.60 5.62 5.64 5.66 Y = 5.581 0.0038X28.5DSBY = 5.612 0.0038X Y = 5.660 0.0043XTotal dopant concentration (mol %)Measured lattice parameter () DWSB (760 oC) DWSB (RT) ESB (RT) 22WSB -Bi2O3 Figure 3-4. Lattice parameters of various DW SB compositions with 2:1 mol dopant ratio measured at room temperature (RT) and 760 oC as a function of total dopant concentration; the lattice parameters of 22WSB, 28.5DSB and various ESB compositions at room temperature (RT) are added for comparison [33, 38, 73].

PAGE 71

71 0.0 5.0x1041.0x1051.5x1052.0x1052.5x1050.0 -5.0x104-1.0x105-1.5x105-2.0x105 (a) 0.1Hz 1KHz 1MHz200oC Z'' (ohm)Z' (ohm) 1MHz0102030405060 0 -10 -20 -30 -40 500oC(b) Z'' (ohm)Z' (ohm) 1KHz 1Hz 0.1Hz 1MHz0102030405060 0 -10 -20 -30 -40 500oC(b) Z'' (ohm)Z' (ohm) 1KHz 1Hz 0.1Hz Figure 3-5. Typical impedance spectra of 8D4WSB composition in air (a) 200 oC and (b) 500 oC.

PAGE 72

72 1.01.11.21.31.41.51.61.71.8 -5.0 -4.5 -4.0 -3.5 -3.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 10D5WSB 7.5D7.5WSB 7D3WSBT (oC) Log (S/cm) 1000/T (K-1)700600500400 300 Figure 3-6. Arrhenius plot of conductivitie s for 10D5WSB, 7.5D7.5W SB and 7D3WSB.

PAGE 73

73 12151821 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 036912151821 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.1 (a)14D7WSB 12D6WSB 11D5.5WSB 10D5WSB 9D4.5WSB 8D4WSB700oC Conductivity (S/cm)Total dopant concentration (mol %) 12151821 0.00 0.02 0.04 0.06 0.08 0.10 0.12 (b)500oC 14D7WSB 12D6WSB 11D5.5WSB 10D5WSB 9D4.5WSB 8D4WSB Conductivity (S/cm)Total dopant concentration (mol %) Figure 3-7. Conductivity versus total dopant concentration of DWSB elect rolyte with 2:1 mol dopant ratio at (a) 700 C ; inset represents the extrapolation to pure -Bi2O3 and (b) 500 C If not shown, error bars are smaller than the data points.

PAGE 74

74 5.525.545.565.585.605.625.645.66 0.2 0.4 0.6 0.8 1.0 700oC Lattice parameter at 700 oC Lattice parameter at RT Y = 62.2 + 11.3XConductivity (S/cm)Y = 46.2 + 8.36X(a) 1.02 Lattice paramater () 5.555.565.575.585.595.60 0.00 0.02 0.04 0.06 0.08 0.10 0.12 Conductivity (S/cm)Lattice paramater ()Y = 10.4 + 1.87X(b)500oC Figure 3-8. Conductivity versus lattice paramete r of DWSB compositions with 2:1 mol dopant ratio at (a) 700 oC for both room temperature (RT) and 700 oC XRD and (b) 500 oC.

PAGE 75

75 1.01.11.21.31.41.51.61.7 -4.0 -3.5 -3.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 8D4WSB 9D4.5WSB 10D5WSB 11D5.5WSB 12D6WSB 14D7WSB 20ESB 10GDC10YSZ T (oC) Log (S/cm) 1000/T (K-1)700600500400 300 Figure 3-9. Arrhenius plot of conductivities for 8D4WSB, 9D4. 5WSB, 10D5WSB, 11D5.5WSB, 12D6WSB, 14D7WSB, 20ESB and 10GDC; the dashed line represents the conductivity of 10YSZ [75].

PAGE 76

7612151821 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 Y = 0.219 + 0.0265XHigh temp region Low temp region Activation energy (eV)Total dopant concentration (mol %)Y = 1.137 0.00329X Figure 3-10. The conductivity activation energy for low temperature region ( 600 oC) and for high temperature region ( 600 oC) for DWSB systems as a function of total dopant concentration of DW SB compositions.

PAGE 77

77 12151821 10 12 14 16 18 20 22 Y = 23.463 0.227X Y = 10.488 + 0.201XHigh temp region Low temp region Ln A (S cm-1 K) Total dopant concentration (mol %) Figure 3-11. The pre-exponential te rm for low temperature region (600 oC) and for high temperature region ( 600 oC) as a function of total dopa nt concentration of DWSB compositions.

PAGE 78

78 Table 3-1. Conductivity activation energy and lattice parameter of 10D5W SB, 7.5D7.5WSB and 7D3WSB. Composition Ea (eV) ( 600 oC) Ea (eV) ( 600 oC) Ea (eV) Lattice parameter () 10D5WSB 1.0770.007 0.6260.018 0.4510.016 5.5540.001 7.5D7.5WSB 1.0400.007 0.7640.042 0.2760.046 5.5490.001 7D3WSB 1.4270.003 0.5710.019 0.8560.018 5.5340.002

PAGE 79

79 Table 3-2. Conductivity activation ener gies of various DWSB compositions. Composition Ea (eV) ( 600 oC) Ea (eV) ( 600 oC) Ea (eV) 8D4WSB 1.108 0.544 0.564 9D4.5WSB 1.086 0.583 0.503 10D5WSB 1.077 0.626 0.451 11D5.5WSB 1.084 0.637 0.447 12D6WSB 1.082 0.677 0.405 14D7WSB 1.069 0.796 0.273

PAGE 80

80 Table 3-3. Pre-exponential terms of various DWSB compositions. Composition Ln A (S cm-1 K) ( 600 oC) Ln A (S cm-1 K) ( 600 oC) 8D4WSB 20.70 12.81 9D4.5WSB 20.33 13.28 10D5WSB 19.90 13.64 11D5.5WSB 19.88 13.67 12D6WSB 19.74 14.11 14D7WSB 18.43 14.72

PAGE 81

81 CHAPTER 4 TEMPERATUREAND TIMEDEPENDENT CONDUCTIVITY BEHAVIOR OF DWSB ELECTROLYTE SYSTEMS 4.1 Introduction Cubic stabilized bismuth oxides exhibit high conductivity th rough structure stabilization over a wide temperature range [37, 66, 67]. However, due to an orderdisorder transition, a change in conductivity activati on energy is observed at ~ 600 oC [12, 14, 34]. Previous reports have shown that the polarizability of the dopant cations, which are typically less polarizable than Bi3+, affects the conductivity and th e stability of the disordered fluorite structure [34]. Cubic stabilized bismuth oxides experience c onductivity degradation when annealed in intermediate temperature (IT) ranges, i.e. 500 ~ 700 oC for long periods of time [26, 38, 47-49, 52, 77]. Many studies have been conducted to investigate the factor s influencing this conductivity degradation of cubicstabilized bismuth oxides with time in IT ranges, and it has been reported that both phase transformation [47-49] and aging effects [26, 33, 52] might contribute to conductivity degradation of cubic stab ilized bismuth oxides. However, there is still uncertainty about the temperature regime and the kinetics of each process. Fung et al. and Watanabe et al. repo rted that the cubic phase in the Ln2O3-Bi2O3 (Ln3+ = rare earth trivalent cation) and Y2O3-Bi2O3 systems are unstable 700 oC and undergoes a phase transformation from cubic to rhombohedral phase [47-49, 77]. The phase change and the concomitant lattice restructuring resulted in conductivity degradat ion. Likewise, Verkerk et al. observed that 25 mol% Dy2O3 doped Bi2O3 transformed almost completely from the cubic structure to a rhombohedral st ructure after annealing at 607 oC for 2000 hours [38], again accompanied by conductivity degradation. Indeed, Watanabe reports that all the stabilized phases in the Ln3+-doped bismuth oxide series (Ln3+ = rare earth trivalent cation) reporte d are quenched high temperature phases and

PAGE 82

82 thereby metastable at temperatures, below ~700 oC [77]. Moreover, Watanabe indicates that there exists a critical transformation temperatur e which varies with comp osition and dopant type below which the cubic-phase transforms gra dually into a stable non-cubic phase [77]. In addition to phase transformation, ordered st ructure of oxygen subla ttice has also been shown to play a crucial role in conductivity degradation. Wachsman and co-workers have shown that cubic-stabilized bismuth oxides undergo an order-disorder transition of the oxygen sublattice without phase transformation at about 600 C [1 2, 14, 26, 33, 34, 51, 52, 78]. This phenomenon leads to a non-linear decay in conductivity (aging effect) below this temperature with time. Jiang et al. indicated that although phase tr ansformation and aging lead to conductivity degradation, the kinetic s of these two processes are differe nt [33]. They showed that each process dominates in one of two different temp erature regimes. To find the time-dependence of the conductivity degradation, they comp ared the relative conductivity of (YO1.5)0.25(BiO1.5)0.75 (25YSB) annealed at 500 oC and 650 oC. The conductivity of 25YSB at 500 oC decreased monotonically with time, without apparent phase change. In contrast, the conductivity of this composition at 650 oC was maintained for about 50 hours and then decreased abruptly afterwards due to a cubic-to-rhombohedral phase transf ormation which they observed using X-ray diffraction. Based on these results, they dem onstrated that the fcc-rhombohedral phase transformation was most rapid at ~ 650 oC, while the aging phenomenon toward an ordered structure was most rapid at ~ 500 oC. In Chapter 3, we developed cubic-stabilized DWSB electro lytes having higher conductivity than 20ESB [79]. Through double doping with Dy and W, a reduction of total dopant concentration needed to stabilize the cubic structure wa s achieved at 12 mol% and it resulted in higher conductivity than 20ESB, the most conductive of the cubic-stabilized bismuth

PAGE 83

83 oxides reported [79]. However, due to the long te rm stability issues outlined above, it is crucial to assess to the performance of this DWSB co mposition with respect to temperature and time. Hence, the objectives of this chapter are to examine conductivity behavior of the DWSB system as a function of temperature, time, dopant ratio and total dopant co ncentration. This study provides insight into the steps needed to obtai n optimal dopant composition at a given operating temperature by considering both initial conduc tivity and long term st ability at various temperatures. 4.2 Experimental Procedure 4.2.1 Sample Preparation In Chapter 3, flow chart of the powder synthe sis was demonstrated. In the same way, the electrolyte pellets were prepared for conductiv ity measurements. In order to investigate temperatureand timedependent conductivity behavior, four DWSB compositions of the preferred (highest conduc tivity) 2:1 mol dopant ratio were select ed for long-term stability tests: (DyO1.5)2x(WO3)x(BiO1.5)1-3x (where x = 0.04, 0.05, 0.06 and 0.07); these compositions are referred as 8D4WSB, 10D5 WSB, 12D6WSB and 14D7WSB, respectively. 4.2.2 Temperature Dependent Conductivity Behavior Conductivity measurements were performed through two-point probe electrochemical impedance spectroscopy (EIS) using a Solartron 1260 over the frequency range of 0.1 Hz to 10 MHz. Details on conductivity measurements were explained in Chapter 3. For the long term stability test, each composition was annealed at a given temperature to observe conductivity behavior with time. The measurements were performed between 300 oC and 700 C in air for duration of 100 hours. Each electr olyte pellet was only used for one temperature measurements to identify conductivity mechanism through post-analysis.

PAGE 84

844.2.3 Characterization The crystal phase of as-sintered and anneal ed samples was identified by means of X-ray diffraction analysis (XRD, Philips APD 3720). Afte r we observe conductivity behavior of each composition at a given temperature for 100 hour annealing, X-ray diffraction (XRD) patterns were obtained at room temperature to check phase stability. XRD pattern was obtained using CuK radiation at room temperature between 20o and 80o (2 ). In order to obtain phase identification and the relative amount of each phas e in mixtures, CrystalMaker & CrystalDiffract Programs were used. CrystalDiffract program can generate the calculated diffraction patterns using the structure parameters which include the types and positions of atoms in a unit cell of a crystal. This program also provides the way to work with observed and calculated data. By comparing an observed diffraction pattern with a nu mber of calculated patterns, we can identify an unknown substance and further know the relative amount of each phase in case of mixtures. Details on the use of this progr am are given in Appendix C. The microstructures of as-sintered and a nnealed samples were observed by scanning electron microscopy (SEM, JEOL JSM 6400). Ther mal etching was not conducted for sample preparation in order to mainta in internal structure obtained from long term annealing. To confirm the conductivity mechanism, differential thermal analysis (DTA) was performed with TGA/SDTA851e fr om Mettler Toledo using alumina crucible as sample holders from room temperature to 820 oC. As-sintered and annealed pe llets were ground using mortar and pestle to get fine powders for DT A measurements. The heating rate was 10 oC/min. Nitrogen gas was used as the purging gas, with a flow rate of 45 cm3/min.

PAGE 85

854.3 Results and Discussion 4.3.1 Conductivity Figure 4-2 shows the bulk conductivities of se lected DWSB compositions with 2:1 mol dopant ratio, 20ESB and 10GDC as log vs. 1000/T. As this figure shows, 8D4WSB exhibits the highest conductivity of all compositions over the temperature range shown. As described in Chapter 3, cubic-stabilized bismuth oxides go through an order-disorder transition, which is reflected by a change in Ea at ~ 600 oC [12, 14, 34, 66]. It has been reported that the size of the change in the activation energy is related to the aging rate of cubic-stabilized bismuth oxides [33]. Table 4-1 shows conductivity activation en ergy of selected DWSB compositions and 20ESB in both low and high temperature regions. Each of the DWSB compositions shows a change in conductivity activation energy, and we will explore the relationship between this change and stability in the following sections. 4.3.2 Stability 4.3.2.1 Conductivity behavior of DWSB in the IT ranges In order to observe the effect of annealing temperature on conductivity behavior with time, various DWSB compositions were annealed between 500 and 700 oC for 100 hours. The effect of total dopant concentration on conductivity at each annealing temperature was examined. Figure 4-3 shows the conductivity behavior of three different DWSB compositions annealed at 700 oC with time. When the samples were annealed at 700 oC, which is significantly higher than the expected order-disor der transition temperature of ~ 600 oC, each DWSB composition maintained its init ial conductivity. Also, 8D4WSB with a lo wer total dopant concentration maintained higher conductivity than 10D5WSB and 12D6W SB. This observation demonstrates that all DWSB compositions under consideration are stable at 700 oC.

PAGE 86

86 However, these DWSB compositions experi enced a decrease in conductivity when annealed 600 oC. Figure 4-4 shows time-dependent conduc tivity behavior for various DWSB compositions annealed at (a) 600 oC and (b) 500 oC. 20ESB was also annealed at 500 oC for comparison with the DWSB compositions. Although conductivity degradation was observed in the DWSB compositions at 600 oC, the conductivity degradation trend with respect to total dopant c oncentration was different at 600 oC versus 500 oC. Of these DWSB compositions, 8D4WSB maintained the highest conductivity of 0.10 S/cm after 100 hour annealing at 600 oC. On the other hand, of all DWSB compositions as well as 20ESB, 10D5WSB had the highe st conductivity of 0.0051 S/cm after 100 hour annealing at 500 oC due to its relatively high initial c onductivity and low degradation rate. The relative conductivity ( t/ o) was plotted on a semi-log scale in order to clearly observe the time-dependent conductivity degradation trend at both temper atures. Figure 4-5(a) shows the relative conductivity versus time for various DWSB compositions annealed at 600 oC. This figure shows that conductivity degradation is observed in all compositions and increases as total dopant concentration increases (with a fixed dopant ratio). Conversely, when these DWSB compositions are annealed at 500 oC, the trend reverses as shown in Figure 4-5(b) and the conductivity degradation lessens as tota l dopant concentration increases. To compare the relative amount of c onductivity decrease after 100 hours, log ( 100/ 0) was plotted at 600 oC and 500 oC as a function of total dopant conc entration as shown in Figure 4-6. Log ( 0/ 0) was chosen as a baseline for the initial conductivity at both temp eratures. From these plots, it is obvious that conductiv ity degradation exhibits an oppos ite trend as a function of total dopant concentration at 600 oC versus 500 oC. In addition, DWSB compositions experienced a

PAGE 87

87 relatively larger conductivity decrease at 500 oC compared to 600 oC except for 14D7WSB (total dopant concentration = 21 mol%), wh ich has the highest total dopant concentration in this study. Clearly this shows 2 diff erent phenomena. At 500 oC, it is ordering and higher total dopant concentration decreases activ ation energy difference and rate of ordering. At 600 oC, it is phase stability which is observed for so me (not all dopants) singly doped bismuth oxides. It is believed that increasing W concentration makes it less phase stable at 600 oC. Details on the phase stability of these DWSB compositions will be discussed in the following section. 4.3.2.2 Conductivity degradation mechanism SEM was carried out to compare their microstructures between as-sintered sample and annealed sample. Figure 4-7 shows SEM images of fractured 8D4WSB pellets. As shown in Figure 4-7(a), assintered 8D4W SB has well-sintered grains a nd many pores are trapped within the grains. However, Figure 4-7(b) and Figure 4-7(c) show that the 8D4WSB annealed at 600 oC (or 500 oC) for 100 hours has the distortion of the latti ce grain. This could be due to the phase decomposition while annealing at 600 oC and 500 oC. In order to investigate phase stability during annealing, the structure of each composition was characterized using X-ray di ffraction analysis after each stability test. Figure 4-8 shows XRD patterns of various DWSB compositions annealed at 600 oC for 100 hours. This figure shows that all DWSB compositions tested at this temperature experienced varying amounts of phase change with the conductivity degradation. Figure 4-8(b) s hows the magnified version of XRD patterns which are composed of the main peak of each phase between 26 ~ 30o (2 ). It appears that the amount of secondary phase te nds to increase as tota l dopant concentration increases. Many studies have shown that some singly doped bismuth oxides undergo a phase transformation from cubic to rhombohedral phase at 600 oC [47-49]. However, the secondary

PAGE 88

88 phase which is observed from DWSB compositi ons in this study was not identified as a rhombohedral phase. Instead, the secondary phases of Figure 4-8 were well matched with a (Bi rich) tetragonal phase (7Bi2O3WO3 Bi14WO24, S.G. 88) and an (W rich) orthorhombic (Bi2O3WO3 Bi2WO6, S.G. 61) phase. CrystalDiffract program was used to obtain phase identification and the relative am ount each phase in mixtures. Fi rst, the calculated mixture pattern of cubic [27], tetragonal (Bi14WO24) [80] and orthorhombic (Bi2WO6) [81] were generated using respective structure parameter. By manipulating the relative amount of each phase in calculated mixtures, it was analyzed that the observed pattern of 14D7WSB annealed at 600 oC for 100 hours is approximately composed of 45% cubic phase, 30% orthorhombic phase and 25% tetragonal phase. Indeed, Watanabe et al. found that Bi7WO13.5 (22.22 mol% WO3) decomposed gradually into the tetragonal and or thorhombic phases below 670 oC [69]. They measured the electrical conductivity for (Bi2O3)1-x(WO3)x electrolyte compositions. Figure 4-9 shows the electrical conductivities of for (Bi2O3)1-x(WO3)x, where x = 0.125 (tetragonal) 0.25 (cubic) and 0.50 (orthorhombic) obtained by Watana be et al. [69]. This figure de monstrates that the neighboring phases, Bi14WO24 (12.5 mol% WO3) and Bi2WO6 (50 mol% WO3) had far lower conductivity than Bi7WO13.5 (22.22 mol% WO3), which has a cubic structure [ 69]. Therefore, we deduce that secondary phase formation is likely a strong co mponent of isothermal conductivity degradation of DWSB at 600 oC. It is also believed that W concen tration as a second dopant should be minimized since the secondary phase formation is due to a decomposition of (WO3)-(Bi2O3) cubic phase. On the other hand, when these DWSB compositions were annealed at 500 oC for 100 hours, the trend of conductivity degradation and phase change according to total dopant

PAGE 89

89 concentration was different Unlike annealing at 600 oC, at 500 oC the conductivity degradation rate of the DWSB compositions lessens as tota l dopant concentration increases as shown in Figure 4-5(b) and Figure 4-6. Figure 4-10 shows XRD patterns of various DWSB compositions including 20ESB annealed at 500 oC for 100 hours. Figure 4-10(b) clearly shows that the amount of secondary phase tends to decrease as the total dopant concentration increases. Although secondary phases still exist at the lowest dopa nt concentration, the amount of secondary phase diminished considerably as total dopant concentration incr eases. Especially, 14D7WSB maintained a cubic structure after the 100 hour annealing at 500 oC. It is interesting to note that 20ESB did not undergo a phase tr ansformation even with large conductivity decay as shown in Figure 4-10. The st ability observed for 20ESB is consistent with literature findings [33]. Specifical ly, Wachsman et al. found that this conductivity decay without a phase change was due to the ordering of the oxygen sublattice [26, 33, 51, 52, 78]. For DWSB compositions, conductivity degradation was s till observed for 12D6WSB and 14D7WSB at 500 oC although the phase change was negligible, which indicates that this conductivity degradation at 500 oC is due to anion ordering. This assertio n is further supported by the fact that each DWSB composition undergoes a larg er conductivity degradation at 500 oC than at 600 oC in spite of the reduced secondary phase formation at 500 oC. To confirm the difference in conductivity degradation mechanism at 600 oC vs. at 500 oC, differential thermal analysis (DTA ) was performed to investigate th e effect of thermal history on the samples from room temperature to 820 oC. Figure 4-11 shows the DTA results of 8D4WSB which was (a) as-sintered, (b) annealed at 600 oC for 100 hours and (c) annealed at 500 oC for

PAGE 90

90 100 hours. The as-sintered 8D4WSB sample did not exhibit a p eak, while the 8D4WSB samples annealed at 600 oC and 500 oC exhibited endothermic peaks. Watanabe et al. also used DTA to characterize (Y2O3)0.225(Bi2O3)0.775 (22.5YSB) with rhombohedral phase and observed a drastic endothermic peak at 720 oC [82]. They demonstrated that the endothermic peak which corresponds fro m rhombohedral to cubi c phase transformation occurs near 720 oC in the (Y2O3)x(Bi2O3)1-x series [82]. In our study, we observed an endothermic peak at ~ 767 oC for 8D4WSB compositi on annealed at 600 oC. Therefore, it is believed that the endothermic peak at ~ 767 oC corresponds to the transformation from secondary phase formed during annealing at 600 oC to high-temperature cubic phase. The temperature difference for the endothermic peak position betw een 22.5YSB and 8D4WSB could be attributed to the nature of the different structure of secondary phases. For 8D4WSB composition annealed at 500 oC, two endothermic peaks at ~ 630 oC and at ~ 750 oC were observed. The first endothermic peak at ~ 630 oC is consistent with previous DTA results of 20ESB reported by Wachsman et al [51] and previous differential scanning calorimetry (DSC) results of 20ESB reported by N. Jiang et al. [33] They report that the ordered structure of oxygen sublattice is rela xed to a disordered state at ~ 630 oC with absorption of energy upon heating. Therefore, this result clearly demonstrates 8D4WSB undergoes the ordering of the oxygen sublattice as well as se condary phase formation when annealed at 500 oC for 100 hours and that this anion ordering contributed to a larger decrease in conductivity at 500 oC. The second endothermic peak temperature (~ 750 oC), associated with phase transformation, for 8D4WSB annealed at 500 oC, was relatively lower than the e ndothermic peak temperature (~ 767 oC) of 8D4WSB annealed at 600 oC. This fact provides further ev idence that 8D4WSB experience less

PAGE 91

91 secondary phase formation at 500 oC than at 600 oC. Therefore, the DTA results confirm that anion ordering played a dominant role in conductivity degradation at 500 oC for the 8D4WSB composition. For 8D4WSB annealed at 500 oC for 100 hours, XRD patterns were obtained before and after DTA measurement as shown in Figure 4-12. Th is result also identifies that cubic fluorite structure with disordered oxygen sublatti ce can be achieved upon heating to ~ 800 oC, supporting the DTA results. Therefore, it can be concluded that the diso rdered state of the oxyge n sublattice for DWSB compositions is maintained at 700 oC, which is significantly higher than order-disorder transition temperature. Hence, conductivity degradati on was not observed while annealing at this temperature. Conversely, phase destabilization and/or ordering of oxygen sublattice occur at temperatures 600 oC, resulting in conductivity degradation. Especia lly, anion ordering is a dominant source for conduc tivity degradation at 500 oC, while phase transformation is dominant at 600 oC. The effects of the difference in bond character between Bi-O a nd dopant (Dy or W)-O may be obscured at 700 oC due to the higher thermal energies However, the differences in bond character which are caused by the dopant type and dopant concentr ation starts to influence the structure and therefore the conductiv ity as temperature decreases. 4.3.2.3 Conductivity behavior of 8D4WSB Figure 4-2 shows that 8D4WSB is the highest conductivity composition, but experienced the most rapid conductivity degradation at 500 oC as shown in Figure 4-4. Indeed, the conductivity of 8D4WSB after 100 hour annealing at 500 oC was less than that of other DWSB compositions. It is believed that 8D4WSB has th e highest initial conductivity since it has more open structure with larger lattice parameter [79] but its stability decreases when there is insufficient thermal energy to maintain the hi gh oxygen vacancy concentration. Therefore,

PAGE 92

92 though the double dopant strategy using Dy and W was successful in achieving high initial conductivity, it struggled to deliver phase a nd structural stability between 500 and 600 oC. To obtain a clearer picture of the long term stability of 8D4WSB, additional anneals were conducted at 400 oC, 300 oC and 650 oC to observe the conductivity de gradation. Figure 4-13(a) shows the conductivity behavior of 8D4WSB annealed at each di fferent temperature including previous 500 ~ 700 oC temperature ranges. The relative conductivity ( t/ o) of 8D4WSB was plotted as a function of time at variou s temperatures in Figure 4-13(b). In contrast to 500 oC operation, conductivity degradat ion was greatly reduced at 400 oC and 300 oC. Figure 4-13(b) clearly shows th at the degradation rate at 300 oC was much smaller than that at 400 oC. This suggests that exponentially slow er kinetics at decreased temperatures influence the conductivity degradation rate. That is, slow kinetics, for phase transformation or anion ordering, retard the conduc tivity degradation rate as te mperatures decreases below 500 oC. X-ray diffraction analysis was used to identify the phase stability in the samples after a 100 hour anneal at various temperatures. Fi gure 4-14 shows the XRD patterns of 8D4WSB annealed at 300 ~ 700 oC for 100 hours as well as a referenc e pattern for an as -sintered pellet. This figure also includes th e XRD pattern of cubic Bi2O3 which has Fm 3m space group (225). The cubic fluorite structure was mainta ined when 8D4WSB was annealed at 650 oC and 400 oC. On the other hand, secondary pha ses were formed between 500 and 600 oC. That is, 8D4WSB is phase stable at temperatures 650 oC and 400 oC. Yaremchenko et al. have also reported that Bi2O3-ZrO2-Y2O3 and Bi2O3-Nb2O5-Ho2O3 solid solutions experienced phase instability at 600 oC and 500 oC [83]. Moreover, they reported that the new secondary phase could not be assigned as rhombohedr al phase. This result is consistent with our findings for the Bi2O3-Dy2O3-WO3 system (DWSB).

PAGE 93

93 Figure 4-13 and Figure 4-14 show that the small degradation of conductivity is still observed below 500 oC without phase change. Since there was no observed phase change below 500 oC, we believe that the slight conductivity de gradation below this temperature is due to ordering of oxygen sublattice [26, 33, 52]. Practically, 8D4WSB, which is the highest initial conductivity composition, can be a promising electrolyte compos ition when operated above 650 oC. There is a possibility that this composition could be also used for low temperature ( 400 oC) applications if electrode kinetics is fast enough, since it undergoes only slight conductivity degradation without phase change. Lowering the total dopant concentration to obtain higher co nductivity is reas onable above the order-disorder transition temperature (~ 600 oC) or below 500 oC for DWSB compositions in this study, but further study will be necessary to enha nce the long term stab ility of this DWSB electrolyte compositi on between 500 and 600 oC.

PAGE 94

94050100150200250 0.0 0.2 0.4 0.6 0.8 1.0 500oC Relative conductivity (t/o)Time (hour)650oC Figure 4-1. Comparison of conduc tivity decay of 25YSB at 500 oC and 650 oC [33].

PAGE 95

95 1.01.11.21.31.41.51.61.7 -4 -2 0 2 4 6 8 8D4WSB 10D5WSB 12D6WSB 14D7WSB 20ESB 10GDCT (oC) Ln T (Scm-1K) 1000/T (K-1)700600500400 300 Figure 4-2. Arrhenius plot of conductivities for 8D4WSB, 10 D5WSB, 12D6WS B, 14D7WSB, 20ESB and 10GDC.

PAGE 96

96 020406080100 -0.75 -0.50 -0.25 0.00 8D4WSB 10D5WSB 12D6WSB Log (S/cm)Time (hour) Figure 4-3. Conductivity behavior of 8D4W SB, 10D5WSB and 12D6W SB annealed at 700 oC as a function of time.

PAGE 97

97 020406080100 -2.0 -1.5 -1.0 -0.5 8D4WSB 10D5WSB 12D6WSB 14D7WSB(a) 600oC Log (S/cm)Time (hour)020406080100 -3.0 -2.5 -2.0 -1.5 -1.0 8D4WSB 10D5WSB 12D6WSB 14D7WSB 20ESB Log (S/cm)Time (hour) (b) 500oC Figure 4-4. Isothermal comp arison of time-dependent conduc tivity behavior for 8D4WSB, 10D5WSB, 12D6WSB and 14D7WSB annealed at (a) 600 oC and (b) 500 oC; 20ESB is added in (b)

PAGE 98

98020406080100 -1.5 -1.0 -0.5 0.0 Log (t/o)(a) 600oC 8D4WSB 10D5WSB 12D6WSB 14D7WSB Time (hour)020406080100 -2.0 -1.5 -1.0 -0.5 0.0 Log (t/o) 8D4WSB 10D5WSB 12D6WSB 14D7WSB 20ESB Time (hour) (b) 500oC Figure 4-5. Relative time-dependent conductivity degradation fo r 8D4WSB, 10D5WSB, 12D6WSB and 14D7WSB annealed at (a) 600 oC and (b) 500 oC; 20ESB is added in (b)

PAGE 99

9912 15 18 21 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 600oC 500oC Log (100/o)Total dopant concentration (mol %) Figure 4-6. Comparison of rela tive conductivity change of va rious DWSB compositions after 100 hour annealing at 600 oC and 500 oC as a function of total dopant concentration.

PAGE 100

100 Figure 4-7. SEM images of 8D4WSB (a) as-sintered (b) annealed at 600 oC for 100 hours and (c) annealed at 500 oC for 100 hours.

PAGE 101

10120304050607080 (a)14D7WSB 12D6WSB 10D5WSB 8D4WSB As sintered Intensity (arb. unit)2 (degree) 600oC(420) (331) (400) (222) (311) (220) (200) (111)2627282930 orthorhombic tetragonal14D7WSB 12D6WSB 10D5WSB 8D4WSB Intensity (arb. unit)2 (degree)(b)cubic Figure 4-8. (a) XRD patterns of 8D4WSB, 10 D5WSB, 12D6WSB and 14 D7WSB annealed at 600 oC for 100 hours; the pattern of as-sintere d 8D4WSB was added as the reference and (b) magnified version between 26 ~ 30o (2 ).

PAGE 102

1021.01.21.41.61.82.02.2 -5 -4 -3 -2 -1 x = 0.25 x = 0.125 x = 0.50 orthorhombic tetragonalT (oC) Log (S/cm) 1000/T (K-1)cubic(Bi2O3)1-x(WO3)x800700600500400300 200 Figure 4-9. Arrhenius plots of electrical conductivity for (Bi2O3)1-x(WO3)x, where x = 0.125, 0.25 and 0.50 [69].

PAGE 103

10320304050607080 14D7WSB 8D4WSB(420) (331) (400) (222) (311) (220) (200) (111)20ESB 12D6WSB 10D5WSB Intensity (arb. unit)2 (degree) 500oC(a) 2627282930 cubic14D7WSB 8D4WSB 20ESB 12D6WSB 10D5WSB Intensity (arb. unit)2 (degree)(b) orthorhombic Figure 4-10. (a) XRD pattern s of 8D4WSB, 10D5WSB, 12D6WSB, 14D7WSB and 20ESB annealed at 500 oC for 100 hours and (b) magnified version between 26 ~ 30o (2 ).

PAGE 104

104 500550600650700750800 -0.5 -0.4 -0.3 -0.2 -0.1 0.0 (a) As sintered Temperature Differential (oC)Temperature (oC)500550600650700750800 -0.5 -0.4 -0.3 -0.2 -0.1 0.0 0.1 767(b) 600oC Temperature Differential (oC)Temperature (oC)500550600650700750800 -0.5 -0.4 -0.3 -0.2 -0.1 0.0 (c) 500oC Temperature Differential (oC)Temperature (oC)634.5 749.3 Figure 4-11. DTA heating curve for 8D4WSB; (a) as-sintered powder, (b) annealed at 600 oC for 100 hours and (c) annealed at 500 oC for 100 hours.

PAGE 105

10520304050607080 After Before Intensity (arb. unit)2 (degree) Figure 4-12. XRD patterns of 8D4WSB annealed at 500 oC for 100 hours before and after DTA measurements.

PAGE 106

106 0102030405060708090100 -4.0 -3.5 -3.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 (a) 700oC 650oC 600oC 500oC 400oC 300oC Log (S/cm)Time (hour)0102030405060708090100 0.0 0.2 0.4 0.6 0.8 1.0 700oC 650oC 600oC 500oC 400oC 300oC(b) Relative conductivity (t/o)Time (hour) Figure 4-13. (a) Time dependent conductivity and (b) relative conduc tivity for 8D4WSB annealed at various temperatures as a function of time.

PAGE 107

10720304050607080 JCPDS (Cubic, S.G. 225) 650oC 700oC 600oC 500oC 400oC 300oC As sintered Intensity (arb. unit)2 (degree)(111) (200) (220) (311) (222) (400) (331) (420) Figure 4-14. XRD patterns of 8D4WSB annealed at various temperatures for 100 hours.

PAGE 108

108 Table 4-1. Conductivity activation energies of 8D4WSB, 10D5WSB, 12 D6WSB, 14D7WSB and 20ESB. Composition Ea (eV) ( 600 oC) Ea (eV) ( 600 oC) Ea (eV) 8D4WSB 1.108 0.544 0.564 10D5WSB 1.077 0.626 0.451 12D6WSB 1.082 0.677 0.405 14D7WSB 1.069 0.796 0.273 20ESB 1.235 0.733 0.502

PAGE 109

109 CHAPTER 5 ENHANCED LONG TERM STABILITY OF BISMUTH OXIDE BASED ELECTROLYTES FOR IT-SOFC OPERATION AT 500 OC 5.1 Introduction The popular yttria stabilized zirc onia (YSZ) electrolyte is known to be quite stable in Solid Oxide Fuel Cells (SOFCs) due to its therm odynamic stability in both oxidizing and reducing atmospheres [11]. However, YSZ require s an operation temperature of > 700 oC to obtain acceptable ohmic resistance [1, 2]. Lower temper ature operation allows a wider choice of interconnect materials and improvement in cell reliability during prol onged operation [10, 11]. Therefore, it would be of great benefit if SOFC s could be designed to give a reasonable power output in IT ranges (500 ~ 700 oC) [84]. In the last few decades, cubi c-stabilized bismuth oxide has received attention as an alternative electrolyte material to conventional YSZ. Cubic-stabilized bismuth oxides exhibit the highest known oxide ion conductivity over a wide temperature range [66, 79] In the IT range, its ionic conductivity is two orders of magnitude higher than that of YSZ at corresponding temperatures [66, 79]. Therefore, it is expected that a replacement of YSZ by cubic-stabilized bismuth oxide would allow signifi cant reduction in SOFC operation temperature. On the other hand, when annealed in the IT range for long pe riods of time, cubic-stab ilized bismuth oxides are known to experience conduc tivity degradation due to phase and structural instability [33, 48]. In Chapter 3 and Chapter 4, it was found that (DyO1.5)0.08-(WO3)0.04-(BiO1.5)0.88, 8D4WSB, had the highest conductivity due to the reduction of total dopant concentration and concomitant larger lattice parameter [79] and showed good l ong term stability when it was annealed above ~ 650 oC [85]. However, 8D4WSB had the grea test decrease in conductivity at 500 oC, resulting from the phase change and ordering of the oxygen sublattice [85]. It was also found that conductivity degradation decrea sed as total dopant concentration increased for DWSB

PAGE 110

110 compositions [85]. This result demonstrates that more stable DWSB compositions can be achieved at 500 oC by increasing total dopant co ncentration. Enhancing th e long term stability is crucial for making stabilized bismuth oxides amen able to SOFC operation over time periods in excess of several hundred hour s, in the IT range. The objectives of this Chapte r 5 are to examine the effect of dopant concentration on conductivity behavior of various DW SB compositions with time at 500 oC, and to obtain an optimal dopant composition which has enhanced long term stability at this temperature. Phase stability was investigated for each DWSB co mposition before and after annealing at 500 oC, to determine the factors contributing to conductivity degradation. 5.2 Experimental Procedure 5.2.1 Sample Preparation All samples were synthesized by the solid stat e reaction of a stoichiometric mixture of oxides. In this study, four (DyO1.5)x(WO3)0.05(BiO1.5)1-x-0.05 (where x = 0.10, 0.15, 0.20 and 0.25) (DWSB) compositions with the same 5 mol% W concentration were prepared. These compositions are referred as 10D5W SB, 15D5WSB, 20D5WSB and 25D5WSB. (DyO1.5)0.25(BiO1.5)0.75 (25DSB) and (GdO1.5)0.1(CeO2)0.9 (10GDC) were also prepared for conductivity comparison. 5.2.2 Time Dependent Co nductivity Behavior Conductivity measurements were performed through two-point probe electrochemical impedance spectroscopy (EIS) using a Solartron 1260 over the frequency range of 0.1 Hz to 10 MHz. Conductivity measurements were preformed for each composition at 500 oC for 300 ~ 500 hours to observe conductivity behavior with time The conductivity measurements were carried out on an hourly basis at this temperature in air.

PAGE 111

1115.2.3 Phase Analysis The phase analysis of as-sintered and annealed samples was identified by means of X-ray diffraction analysis (XRD, Philips APD 3720) XRD pattern was obtained using CuK radiation at room temperature between 20o to 80o (2 ). Extrapolation method us ing Nelson-Riley function was carried out to estimate the lattice paramete r of cubic-stabilized bismuth oxides at room temperature. In addition, in order to obtain phase identification and th e relative amount of each phase in mixtures, CrystalMaker & CrystalDiffract Programs were used. 5.3 Results and Discussion 5.3.1 Conductivity In our previous work [85], although a ll DWSB compositions experienced large conductivity degradation at 500 oC, (DyO1.5)0.10-(WO3)0.05-(BiO1.5)0.85, 10D5WSB, maintained the highest conductivity af ter 100 hour anneal at this temper ature including 20E SB. It has been shown that ordering decrease with increasing dopan t concentration at this temperature [33, 85]. Hence higher dopant concentrations are required to achieve a stable DWSB composition, even though this results in a lo wer initial conductivity. It was also determined that W concentrati on should be minimized since the secondary phase formation is due to a decomposition of (WO3)-(Bi2O3) cubic phase. Theref ore, considering both phase stability and electrical properties, we increased the Dy dopant concentration with a fixed 5 mol% W dopant concentration in order to attain improved conductivity stability. Figure 5-1 shows XRD patterns of various DWSB compositions, each with the same 5 mol% W content. It was found that initially al l DWSB compositions had single cubic structure. The lattice parameters of these compositions were determined from XRD patterns using NelsonRiley function at room temperat ure. Figure 5-2 shows that the lattice parameter measured at room temperature linearly decreas es with increasing Dy concentr ation, obeying Vegards law.

PAGE 112

112 When this linear fit is extrapolated to pure -Bi2O3 (dopant concentrat ion = 0), the calculated lattice parameter is 5.609 and is very compar able to that of our previous work [79]. The bulk conductivities of various DWSB comp ositions with different Dy concentration (10D5WSB, 15D5WSB, 20D5W SB and 25D5WSB) are plotted in Figure 5-3. We observed that the conductivity decreased as Dy dopant concentration increased. Table 5-1 shows the activation energy values for these DWSB compositions in the low and high temperature regions. The change in activ ation energy slope of plot below and above 600 oC decreases with increasing Dy dopant concentration. This behavior is consis tent with our previous findings [79]. It has been known that the obser ved change in conductivity activation energy is generally consistent with the trend in aging ra te of cubic stabilized bismuth oxides at 500 oC [33, 79]. Of those DWSB compositi ons, 25D5WSB has considerab ly lower activation energy difference below and above 600 oC. 5.3.2 Stability 5.3.2.1 Long term conductivity behavior Long term stability tests were perfor med for 15D5WSB, 20D5WSB, 25D5WSB and 25DSB at 500 oC for about 300 hours. Particularly 25D5WSB was annealed at 500 oC for 500 hours. Figure 5-4(a) shows time-dependent conductivity behavior for these compositions as well as, for comparison, 10D5WSB at 500 oC. This figure shows that as the total dopant concentration (Dy concentration in this case) increased, the long term stability was greatly enhanced although the initial conduc tivity decreased as th e total dopant concentr ation increased. In order to examine the time-dependent conductiv ity behavior in more detail, the relative conductivity ( (t)/ (t=0)), was plotted. Figure 5-4(b) show s relative conductivity behavior for various DWSB compositions and 25DS B as a function of time at 500 oC.

PAGE 113

113 DWSB compositions usually showed slight con ductivity degradation in the initial period of time and underwent large conductivity degradati on afterwards except for 25D5WSB. Figure 5-4 shows that the onset time of large degradation in conductivity is delayed and the overall degradation rate decreases as dopant concentra tion increases. Consequently, we observed that conductivity degradation at 500 oC was significantly minimized by increasing the total dopant concentration while retaining a constant 5 mol% W concentration for DWSB compositions. Therefore, this result demonstrates that higher d opant concentrations provide better stability at this temperature. Table 5-2 shows the conductiv ity of various DWSB compositions and 25DSB before and after annealing at 500 oC for various time periods. In part icular, for 25D5WSB, no appreciable degradation was observed for 500 hours other than a small depression of conductivity in the initial period of time from 0.0081 S/cm to 0.0070 S/cm. The conductivity (0.0068 S/cm at t = 500 hours) of 25D5WSB after 500 hour annealing was even higher than that (0.0051 S/cm at t = 100 hours) of 10D5WSB after 100 hour annealing and comparable to the initial grain conductiv ity of 10GDC (0.0068 S/cm). Figure 5-5 shows impedance spectra of 25D5WSB vs. 10GDC measured at 500 oC in air. This figure clearly shows that grain boundary resistance of 25D5WSB is negligible, but 10GDC has a large grain boundary resistance. Duran et al. developed Bi2O3-Y2O3 (or Er2O3) systems and examined the electrical properties of these oxygen ion conductors [61]. Concerning the AC impe dance plane behavior, they observed the absence of a grain boundary re sponse for this material. Two arc responses consisted of bulk and electrode s also have been reported fo r bismuth oxides doped with Er2O3, Y2O3 and MoO3 [61-63]. Based on these re sults, bismuth oxide-based electrolytes show no

PAGE 114

114 evidence of a grain boundary contribution to their conductivity. This could be due to very small grain boundary impedance or to the similar time constants for both grai n interior and grain boundary impedances. Unlike zirconiaand ceri a-based electrolytes, it is well known that bismuth oxide-based electrolytes have negligib le grain boundary resistances because bismuth oxide has considerable solid solubility w ith silica and alumina, two common impurity constituents of grain boundary phases in ceramics [ 59]. This can be consid ered a striking feature of bismuth oxide-based electrolyte. On the other hand, Zhang et al. examined the ionic conductivity of cer ia-based electrolytes as a function of silica contents [60]. They reported that SiO2 impurity is extremely detrimental to the grain boundary conduction in ceria-based electrolytes [60]. This impedance spectrum of 10GDC is comparable to that of Omar et al. [58, 86]. It is be lieved that large grain boundary resistance of 10GDC is due to the SiO2 impurity content in this study. Therefore, w ith respect to total conductivity, 25D5WSB has 84 times greatly higher conduc tivity than 10GDC even after 500 hour annealing at 500 oC. To verify the phase stability of these DWSB compositions after annealing at 500 oC, X-ray diffraction analysis was perf ormed. Figure 5-6 shows XRD patterns of various DWSB compositions annealed at 500 oC. This figure also includes the JCPDS pattern of -Bi2O3 which has a space group of Fm 3m (225). In order to compare the inte nsity of the secondary phase, all XRD patterns were normalized based on the intensity of (111) main peak of fcc structure. 10D5WSB and 15D5WSB had mixed phases of tetragonal and orthorhombic from the decomposition of (WO3)-(Bi2O3) cubic phase, similar to that found in 8D4WSB. Using CrystalDiffract program, it was analyzed that annealed 10D5WSB compos ition is approximately composed of 61% cubic phase, 31% orthohombic pha se and 8% tetragonal phase. Details on this

PAGE 115

115 analysis are given in Appendix C. We also obs erved that the secondary phases in 15D5WSB were reduced, even after a 300 hour anneali ng. In particular, 20D5WSB and 25D5WSB maintained cubic fluorite struct ure throughout the anneal time. Of the two compositions, 20D5WSB had a sma ll depression in conduc tivity initially and then maintained its conductivity until a gradua l decrease in conductivity began after ~ 100 hour annealing. When we compare the XRD pattern s between 20D5WSB and 25D5WSB in Figure 56, it was observed that the crystallinity of 20D5WSB was reduced as revealed by peak broadening, although the fcc stru cture was maintained. This obs ervation provides clues that phase destabilization begins occurring graduall y after ~ 100 hour and consequently results in conductivity degradation for 20D5W SB. However, 25D5WSB mainta ined its conductivity even after 500 hour annealing and we did not observe any indication of phase instability for this composition. Therefore, this resu lt demonstrates that 25D5WSB is the stable composition with respect to phase and structural stability. In this study, other DWSB compositions which have total dopant concentration above 30 mol% were not prepared for l ong term stability tests because of their lower initial conductivities. 5.3.2.2 Conductivity behavior of 25DSB Jiang et al. have shown that the aging phenom enon (without phase tran sformation) is most dominant at ~ 500 oC for cubic stabilized bismuth oxides with RE2O3 (RE = Dy, Ho, Er, Tm and Yb) as well as Y2O3 [33]. These compositions had an exponential decrease in conductivity for an initial period of time and reached a plateau. Also, the amount of conductivity degradation was different depending on the polarizability of dopant cation [33, 34]. Among the dopants, Dy provided the greatest conductivity stability with time at 500 oC due to its higher polarizability than other dopants [33, 34].

PAGE 116

116 To confirm previous result and identify c onductivity degradation mechanism, another 25DSB composition wa s annealed at 500 oC for 96 hours similar to Jia ng et al. [33]. Figure 5-7 shows the comparison of relative conductivity between this work and Jiang et al. [33] for 25DSB with time. This figure shows that 25DSB in this work has the same conductivity trend as that of Jiang et al. That is, 25DSB has initial conductivity degradation for a short period of time before reaching a plateau. This conductivity degradation was analyzed and explained by Jiang et al. as due to the ordering of the oxygen sublattice [12, 26, 33, 52]. However, Figure 5-4 shows that the conductivity of 25DSB gradually decreases after ~ 96 hours with higher degradation rate than before when 25DSB was annealed at this temperature for 300 hours. Figure 5-8 shows the XRD patte rns of two 25DSB compositions which were annealed at 500 oC for different time periods. This figure also includes the JCPDS pattern of cubic phase (Fm 3m) and rhombohedral phase (R 3m) of Bi2O3. The 25DSB which was annealed for ~ 96 hours maintained cubic fluorite structure, while 25DSB had mixed phases of cubic and rhombohedral after 300 hour annealing at 500 oC. It was analyzed that the mixtures were consisted of cubic phase (Fm 3m, Space Group 225) and rhombohedral phase (Bi0.775Dy0.225O1.5, R 3m, Space Group 166). Similarly CrystalDiffract Programs were used for phase analysis. We generated calculated XRD pattern s of mixtures of two phases a nd compared that with observed XRD pattern. It was found that the observ ed pattern of 25DSB annealed at 500 oC for 300 hours is approximately composed of 60% cubic phase and 40% rhombohedral phase. It is believed that ordering phenomenon results in c onductivity degradation initially and the phase change from cubic to rhombohedral is taking e ffect resulting in a gradual d ecrease in conduc tivity for 25DSB composition after ~ 96 hours.

PAGE 117

117 It is interesting to note that the nature of secondary phase is different between doubly doped DWSB and singly doped 25DSB. 25DSB unde rgoes phase change from cubic to rhombohedral, while DWSB experiences phase in stability resulting from the decomposition of (WO3)-(Bi2O3) cubic phase when the total dopant concentration is not sufficient. However, it is shown that W as a second dopant c ontributes to the improvement of phase stability for long term annealing at 500 oC compared with singly doped 25D SB. Consequently, 25D5WSB or 20D5WSB showed enhanced conduc tivity stability than 25DSB. However, further study will be necessary to find another doubly doped bismuth oxide electrolyte composition. Though high dopant concentration provided better conductivity stability for DWSB, it resulted in a lowe r initial conductivity. In Chapter 6, in addition to DWSB new doubly doped bismuth oxide based elect rolytes will be discussed.

PAGE 118

118 27.027.528.028.529.0 20304050607080 (420) (331) (400) (222) (311) (220) (200) (111)25D5WSB 20D5WSB 15D5WSB 10D5WSB Intensity (arb. unit)2 (degree)27.027.528.028.529.0 20304050607080 (420) (331) (400) (222) (311) (220) (200) (111)25D5WSB 20D5WSB 15D5WSB 10D5WSB Intensity (arb. unit)2 (degree) Figure 5-1. X-ray diffracti on patterns of 10D5W SB, 15D5WSB, 20D5WSB and 25D5WSB; inset represents magnified version of 2 = 27 ~ 29 oC.

PAGE 119

119 1015202530 5.48 5.50 5.52 5.54 5.56 5.58 Y = 5.609 0.0035X(Dy + W) dopant concentration (mol %)Measured lattice parameter () Figure 5-2. Lattice parameter versus (Dy + W) dopant concentration of DWSB compositions measured at room temperature.

PAGE 120

120 1.01.11.21.31.41.51.61.7 -4 -2 0 2 4 6 10D5WSB 15D5WSB 20D5WSB 25D5WSBT (oC) Ln T (Scm-1K) 1000/T (K-1)700600500400 300 Figure 5-3. Arrhenius plot of conductivi ties for 10D5WSB, 15D5WSB, 20D5WSB and 25D5WSB.

PAGE 121

1210100200300400500 -3.0 -2.5 -2.0 -1.5 -1.0 10D5WSB 15D5WSB 20D5WSB 25D5WSB 25DSB Log (S/cm)Time (hour)(a)10GDC 0100200300400500 0.0 0.2 0.4 0.6 0.8 1.0 10D5WSB 15D5WSB 20D5WSB 25D5WSB 25DSB Relative conductivity (t/o)Time (hour)(b) Figure 5-4. (a) Conductivity and (b) relative conductivity for 10D5WSB, 15D 5WSB, 20D5WSB, 25D5WSB and 25DSB annealed at 500 oC as a function of time; the initial grain conductivity of 10GDC was added in (a).

PAGE 122

12250 100150200 50 0 -50 -100 (a)0.1HzElectrode Grain50KHz Z'' (ohm)Z' (ohm) 0500010000150002000025000 5000 0 -5000 -10000 -15000 -20000 0100200300400 100 0 -100 -200 -300 Grain Z'' (ohm)Z' (ohm) Grain0.1Hz 30Hz 3KHz Z'' (ohm)Z' (ohm)Grain Boundary Electrode(b) Figure 5-5. Typical impedance spectra of (a) 25D5WSB and (b) 10GDC in air at 500 oC.

PAGE 123

12320304050607080 (420)JCPDS (Cubic, S.G. 225) 25D5WSB(500h) 20D5WSB(250h) 15D5WSB(300h) 10D5WSB(100h) Intensity (arb. unit)2 (degree)(111) (200) (220) (311) (222) (400) (331) Figure 5-6. XRD patterns for 10D5WSB, 15D5WSB, 20D5WSB and 25D5WS B annealed at 500 oC for different time periods; parenthesis represents the total a nnealing hours for each composition.

PAGE 124

124 050100150200250300 0.0 0.2 0.4 0.6 0.8 1.0 25DSB This work (96 hours) 25DSB This work (300 hours) 25DSB Jiang et al. Relative conductivity (t/o)Time (hour) Figure 5-7. Relative conductivity comparison for 25DSB composition which is taken from between the present work a nd Jiang et al. [33] at 500 oC as a function of time.

PAGE 125

125 20304050607080 (420)As sintered 96hJCPDS (Rhombohedral, S.G. 166) JCPDS (Cubic, S.G. 225) 300h Intensity (arb. unit)2 (degree)(111) (200) (220) (311) (222) (400) (331) Figure 5-8. XRD patterns of 25DSB compos ition of as sintered, annealed at 500 oC for 96 hours and 300 hours.

PAGE 126

126 Table 5-1. Conductivity activation energies for 10D5WSB, 15D5WSB, 20D5WSB and 25D5WSB. Composition Ea (eV) ( 600 oC) Ea (eV) ( 600 oC) Ea (eV) 10D5WSB 1.0770.007 0.6260.018 0.4510.016 15D5WSB 1.1280.017 0.7130.047 0.4150.052 20D5WSB 1.1270.008 0.7820.053 0.3450.045 25D5WSB 1.1220.007 0.8950.036 0.2270.030

PAGE 127

127 Table 5-2. Conductivity of various compos itions before and after annealing at 500 oC for different hours. Composition Conductivity (S/cm) (t=0h), (log ) (t=100h) (t=250h) (t=300h) (t=500h) 10D5WSB 0.065 (-1.18) 0.0051 (-2.29) 15D5WSB 0.029 (-1.53) 0.017 (-1.77) 0.0025 (-2.60) 0.0024 (-2.62) -------20D5WSB 0.016 (-1.79) 0.015 (-1.82) 0.0092 (-2.04) --------------25D5WSB 0.0081 (-2.09) 0.0070 (-2.15) 0.0068 (-2.16) 0.0068 (-2.16) 0.0068 (-2.16) 25DSB 0.018 (-1.75) 0.014 (-1.84) 0.0075 (-2.12) 0.0048 (-2.32) -------

PAGE 128

128 CHAPTER 6 NEW DOUBLY DOPED BISMUTH OXIDE ELECTROLYTES 6.1 Introduction In our previous study of DWSB compositions [79], we found that double doping made it possible to reduce total dopant c oncentration for the stabilization of the cubic phase resulting in the highest ionic conductivity reported to date. However, DWSB compositions with relatively low total dopant concentration experien ced large conductivity degradation at 500 oC [85]. Therefore, it is meaningful to develop new el ectrolyte systems and examine their conductivity stability. Iwahara et al. demonstrated that depending on the dopant ionic radius of added oxides, Bi2O3-Ln2O3 (Ln = La Yb) exhibits the rhombohedral phase at relatively large Ln3+ radius and exhibits the fcc phase for comp aratively small radius of Ln3+ [87]. They also reported that Bi2O3Ln2O3 could be composed of mixed fcc and rhom bohedral phases with a cationic radius of intermediate size, depend ing on composition [87]. As discussed earlier, oxygen vacancies tend to order in cubic-stabilized bismuth oxide below 600 oC. Our studies of cubic-stabilized bismut h oxide showed that the polarizability and concentration of the dopant cati on played an important role in the ordering rate [33]. For lanthanide dopant, it was found th at the ordering rate was slowest for Dy and fastest for Yb since Dy3+ was the most polarizable and Yb3+ the least polarizable of the dopant cations within the fccstability window [33]. In our previous work, othe r dopants with larger ioni c radii than Dy were not considered because a pure cubic phase could not be obtained using these dopants. However, the success of the DWSB electrolyte indicates that cubic-phas e stabilization is achievable using the double dopant strategy. Therefore, we should consider lanthanides dop ants such as Ce, Gd and Tb since they have even gr eater polarizability than Dy.

PAGE 129

129 The objective of this Chapter 6 is to de velop a new optimal double dopant electrolyte system which satisfies both goals of high initia l conductivity and long-term stability in the IT range. For this purpose, the total dopant concen tration used to stabi lize bismuth oxides was limited to below 20 mol%. Depending on the dopant ratio and the relati ve amount of dopants, phase purity was investigated. Subsequently, the conductivity behavior and stability in IT ranges were examined. 6.2 Experimental Procedure In this study, three new electrolyte systems were developed. These are (Tb4O7)x(WO3)y(Bi2O3)1-x-y (xTyWSB), (Dy2O3)x(Gd2O3)y(Bi2O3)1-x-y (xDyGSB) and (Dy2O3)x(CeO2)y(Bi2O3)1-x-y (xDyCSB). Details on electrolyt e pellet fabricati on and conductivity measurements are shown in Chapter 3 and Chapter 4. 6.3 Results and discussion 6.3.1 (Tb4O7) (Bi2O3) and (Tb4O7)-(WO3)-(Bi2O3) System The terbium oxide-bismuth oxide system has been investigated for the phase composition and conductivity [88-90]. However, it has also been noted in the literature that Tb-stabilized bismuth oxide shows mixed conductivity due to the multivalent state of the Tb cation [89]. In this study, we paid more attention to the phase stability and overall time -dependent conductivity trend of this material system. Fi rst, terbium doped bismuth oxides ((TbO1.75)x(BiO1.5)1-x) containing 15-25 mol% Tb4O7 were synthesized. Figure 6-1 s hows the XRD patterns of 15TSB, 20TSB and 25TSB. Among the three compositions, a single cubic phase was obtained for 25TSB. Mixed phases of cubic ( -Bi2O3) and rhombohedral (Bi0.775Tb0.225O1.5) were observed for 15TSB and 20TSB. Therefore, we found that 25 mol% Tb was necessary to stabilize cubic structure of TSB system. On the basis of the success of double doping in DWSB system, Bi2O3based electrolytes using Tb and W as dopants ((TbO1.75)x (WO3)y(BiO1.5)1-x-y) were prepared in

PAGE 130

130 order to enhance long term stability as well as reduce total dopant concentrations. Figure 6-2 shows the XRD patterns of 8T4W SB and 10T5WSB. In the case of 8T4WSB, a small amount of secondary phase was found, while 10T5WSB had a pure cubic structure. This result demonstrates that using the double doping st rategy, we could reduce the total dopant concentration for cubic-phase stabilization as was the case for the DWSB system. Conductivity testing was carried out for 25T SB and 10T5WSB. The bulk conductivities of 25TSB and 10T5WSB ar e plotted as log vs. 1000/T in Figure 6-3. Conductivity values of these compositions are compared with 8D4WSB and 20ES B in this Arrhenius plot. For the case of 25TSB, its initial conductivity wa s lower than that of other DWSB and ESB systems. However, the disparity between activation energies was much smaller than that of the other system. For 10T5WSB, even though the conductiv ity was lower than that of 8D 4WSB at high temperature, it had comparable conductivity at low temperature. Long term stability testing was perf ormed on 25TSB and 10T5WSB at 500 oC for 100 hours in order to determine the effect of Tb as a dopant cation on conductivity and stability. Figure 6-4 compares the time-dependent conducti vity behavior for 25TSB and 10T5WSB with 20ESB under isothermal operation of 500 oC. It is observed that 25TSB was fairly stable for about 80 hours, but experienced a gradual decrease in conductivity. On th e other hand, 10T5WSB has hi gher initial c onductivity than 25TSB with less tota l dopant concentration a nd the initial con ductivity was maintained for about 35 hours before degrading. Phase stability was examined afte r long term stabili ty tests at 500 oC. Figure 6-5 and Figure 6-6 show the XRD patterns of 25TSB and 10T5WSB annealed at 500 oC, respectively. 25TSB experienced cubic to rhombohedral phase transformation. This phase change may be the

PAGE 131

131 cause of the observed gradual decrease in conductivity after 80 hours of annealing. For 10T5WSB, the cubic phase was fairly well mainta ined, but the crystallinity was considerably reduced. It is complicated to under stand the precise conductivity degradation mechanism at 500 oC. Previous results on a few DWSB compositions demonstrate that phase change and ordering can contribute to conductivity degrad ation together. Therefore, th e ordering phenomenon may also affect the conductivity degrada tion of 25TSB and 10T5WSB. 6.3.2 (Dy2O3)(Gd2O3)-(Bi2O3) System Doping with larger-radii la nthanides such as Gd result in the formation of the rhombohedral phase [33]. Nevertheless, since double doping reduces the dopant concentration needed to stabilize cubic Bi2O3 (as demonstrated in the DWSB electrolyte system) [79], we anticipate that Bi2O3 doped with the correct ratio of Dy3+ and Gd3+ should attain the desired fcc structure. Moreover, it is e xpected that Dyand Gd-doped Bi2O3 electrolytes (DGSB) will have improved long term stability because Gd ha s even greater polariz ability than Dy. Verkerk et al. examined Dy2O3 doped Bi2O3 system (DSB) and achieved the fcc structure with 28.5-50 mol% Dy2O3 [38], while Takahashi et al. investigated Gd2O3 doped Bi2O3 system (GSB) and attained the fcc structure with 35-50 mol% Gd2O3 [36]. In our previous study of the DWSB electrolyte system, we achieved cubic ph ase-stabilization with a lower total dopant concentration compared to the single dopant system, resulting in higher conductivity [79]. Therefore, we expect to achieve th e cubic phase thr ough double doping of Dy3+ and Gd3+ at lower total dopant concentrations. Various (DyO1.5)x(GdO1.5)y(BiO1.5)1-x-y (xDyGSB) compositions were prepared to find optimal dopant ratio and dopant concentration. Figure 6-7(a) shows the XRD patterns of 8D4GSB, 6D6GSB and 4D 8GSB which were first calcined at 800 oC for 16 hours. As Gd

PAGE 132

132 concentration increases, the amount of rhombohed ral phases increases. In particular, 8D4GSB has almost single cubic phase, but the rhombohedral phase is dominant for 4D8GSB. An additional calcination was carried out under the same condition to ensure the phase of each composition. We found that after the second calcination, 8D4GSB has a pure cubic phase and the other compositions tend to contain more cubic phase as compared with the single calcination powders. This may be due to the slightly larger ionic radius of Gd3+ than Dy3+. As was the case for DWSB, it appears that the content of Dy2O3 should be larger than that of Gd2O3 in order to achieve cubic-phase stabilization. Three compositions with 15 mol% total dopant concentrations were also prepared. Figure 6-8 shows the XRD patterns of 10D5GSB, 7. 5D7.5GSB and 5D10GSB The rhombohedral phase is more dominant than the cubic phase for all three compositions when a single calcinations step is carried out After the second calcination, 10D5GSB is nearly cubic phase pure, but the rhombohedral pha se was still dominant for 7.5D7.5GSB and 5D10GSB. Based on phase observations from other DGSB compositions, 12D6GSB was lastly prepared. As shown in Figure 6-9, mixed cubi c and rhombohedral phases were observed even after a second calcination. For the Dyand Gd-doped Bi2O3 electrolyte system (DGSB), there is a tendency to form the rhombohedral phase as Gd concentration or tota l dopant concentration increases. In this work, various DGSB compositions were prepared to find the correct ratio and net concentration of Dy3+ and Gd3+ needed to obtain the fcc phase. However, it seems that DGSB system has narrow fcc-stability window based on the observed XRD results. Here, 8D4GSB, 10D5GSB and 12D6GSB were selected to examine conductivity and long term stability.

PAGE 133

133 Figure 6-10 shows the Arrhenius behavior of the three DGSB compositions. 20ESB was added for comparison. Particularly, 12D6GSB showed a discontinuity at around 550 oC when conductivity was measured from hi gh temperature to low temperatur e. It is thought that this discontinuity corresponds to th e phase transition from the fcc phase to the rhombohedral phase. For this composition, it was not possible to ach ieve a pure fcc phase even after the second calcination. This indicates that the fcc phase is not stable and easily transforms to the rhombohedral phase on cooling. We again measur ed conductivity from low to high temperature and obtained an almost straight line without recovering to the original conductivity at high temperature. The long term tests were performed fo r these 8D4GSB, 10D5GSB and 12D6GSB compositions at 500 oC for about 100 hours. Figure 6-11 sh ows the conductivity behavior of these compositions as well as 20ESB at 500 oC as a function of time. 8D4GSB and 10D5GSB underwent large conductivity degrad ation initially without delay and reached a plateau. This conductivity trend is very simila r to that of 20ESB. On the other hand, 12D6GSB showed low initial conductivity, but almost ma intained this initial conductivity for 100 hours. It is believed that the phase change from cubic to rhombohe dral is almost complete during conductivity measurements. Afterwards, no conductivity degrada tion was observed during long term stability tests at 500 oC. Phase stability was identified by means of X -ray diffraction after long term annealing at 500 oC. Figure 6-12 shows XRD patterns of 8D4GSB and 10D5GSB before and after annealing. Even though both compositions experienced similar conductivity degradation trends with time, XRD results were totally different. 8D4GSB maintained the cubic phase, while 10D5GSB underwent a complete transformation from cubic to rhombohedral after 100 hour annealing at

PAGE 134

134 500 oC. This result indicates that ordering of oxyge n sublattice is involved in the conductivity degradation of 8D4GSB [14]. However, 10D5GSB experienced an almost complete phase change from cubic to rhombohedral during only 24 hour annealing at 500 oC as shown in Figure 6-12(b). This XRD result demonstrates rhombohed ral phase is more stable than cubic for 10D5GSB even though a pure cubic phase was achievab le with two calcinations steps. It is also noteworthy that after the rhombohedral phase is formed, no conductivity degradation was observed for 10D5GSB. Cubic-stabilized bi smuth oxides are known to undergo occupancy ordering and positional ordering which resu lt in conductivity degradation at 500 oC [26]. Therefore, this observation s uggests that rhombohedral phas e may not experience ordering phenomenon. However, it is still unclear how dopa nt ionic radius and dopant concentration affect ordering of the oxygen sublattice or phase change. 6.3.3 (Dy2O3)-(CeO2)-(Bi2O3) System Fung et al. have reported that phase change from cubic to rhombohedr al can be suppressed by the addition of a small amount of ZrO2 or ThO2 to stabilize Y2O3-Bi2O3 (YSB) electrolyte system [47, 49]. With additions of less than 5 mol% of these additives, the cubic phase was retained at 650 oC for over 1000 hours. In addition, Huang et al. reported that small additions of CeO2 suppress aging of stabilized Y2O3-Bi2O3 oxides [91]. In this study, 2 mol% of CeO2 was added to Dy2O3-Bi2O3 oxides to enhance phase stability. (Dy2O3)0.10(CeO2)0.02(Bi2O3)0.88 (10D2CSB) was prepared through solid state synthesis. Long te rm conductivity results were examined at 500 oC and 600 oC for this composition. First, (CeO2)x-(Bi2O3)1-x (xCSB) electrolyte system we re prepared. Four compositions having 15 mol%, 20mol%, 30 mol% and 40 mol% CeO2 as a dopant were fabricated and are referred to as 15CSB, 20CSB, 30CSB and 40C SB. Figure 6-13 shows the XRD patterns of calcined 15CSB, 20CSB, 30CSB and 40CSB. All compositions had mixed phases of -Bi2O3

PAGE 135

135 and CeO2. No change was observed even with two calci nations steps. Therefore, it is believed that the mixed phases are very stable and the sole use of a Ce dopa nt is unable to stabilize the cubic structure. Based on the success with DWSB com positions, 10D2CSB with 2 mol% CeO2 was prepared and examined. It was found that 10D2C SB had a pure cubic phase. To examine the time-dependent conductivity behavior at 500 oC and 600 oC, samples with this composition were annealed at these temperatures for various peri ods of time, respectively. Figure 6-14 shows that 10D2CSB experienced an initial co nductivity drop and reaches to a plateau. This trend is similar to that of 20ESB. While the conductivity of 10D2CSB was fa irly stable at 600 oC for about 200 hours with the highest initial conductivity, but it did subsequently experience a gradual decay as shown in Figure 6-15. After long term stability tests, phases of 10D2CSB samples annealed at 500 oC and 600 oC were examined by XRD as shown in Figure 6-16 Surprisingly the cubic phase was maintained for 10D2CSB which had been annealed at 500 oC. Therefore, again it can be deduced that the ordered structure of the oxygen subl attice resulted in an initial conductivity drop and the addition of CeO2 was not effective at mitigating the ordering phenomenon at 500 oC. In addition, XRD result of 10D2CSB annealed at 600 oC demonstrates that this composition underwent a phase change during an nealing for 500 hours. From Figure 6-15 and Figure 6-16, it was believe d that the use of CeO2 as a dopant was an effective way to suppress phase transformation at 600 oC, but the conductivity degradation was accompanied by a phase change from cubic to rhombohedral (Bi0.775Dy0.225O1.5) at the end. A slightly higher concentration of CeO2 (> 2 mol%) is expected to s uppress the phase transformation at 600 oC, ultimately.

PAGE 136

13620304050607080 20TSB 25TSB 15TSB Bi0.775Tb0.225O1.5 (Rhombohedral) Intensity (arb. unit) 2 (degree) (111) (200) (220) (311) (222) (400) (331) (420) Figure 6-1. X-ray diffr action patterns of (Tb4O7)x(Bi2O3)1-x, where x = 0.15, 0.20 and 0.25.

PAGE 137

137 20304050607080 (420) (331) (400) (222) (311) (220) (200) (111) 10T5WSB 8T4WSBIntensity (arb. unit) 2 (degree) Figure 6-2. X-ray diffraction pattern s of 8T4WSB a nd 10T5WSB.

PAGE 138

138 1.01.11.21.31.41.51.61.71.8 -4.5 -4.0 -3.5 -3.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 10T5WSB 25TSB 8D4WSB 20ESBT (oC) Log (S/cm) 1000/T (K-1)700600500400 300 Figure 6-3. Arrhenius plot of conductivities for 10T5WSB, 25 TSB, 8D4WSB and 20ESB.

PAGE 139

139 020406080100120 -3.0 -2.5 -2.0 -1.5 -1.0 -0.5 10T5WSB 25TSB 20ESB Log (S/cm)Time (hour) Figure 6-4. Isothermal comp arison of time-dependent conductivity for 10T5WSB and 25TSB annealed at 500 oC; 20ESB data is added for comparison.

PAGE 140

140 20304050607080 Bi0.775Tb0.225O1.5 (Rhombohedral)25TSB_annealed 25TSB_as sintered Intensity (arb. unit) 2 (degree) (111) (200) (220) (311) (222) (400) (331) (420) Figure 6-5. XRD patterns of 25TSB as-sintered and annealed at 500 oC for 120 hours.

PAGE 141

141 20304050607080 10T5WSB_annealed 10T5WSB_as sinteredIntensity (arb. unit) 2 (degree)(111) (200) (220) (311) (222) (400) (331) (420) Figure 6-6. XRD patterns of 10T5WSB as sintered and annealed at 500 oC for 100 hours.

PAGE 142

142 20304050607080 (a)4D8GSB 6D6GSB 8D4GSBJCPDS (Cubic, S.G. 225) JCPDS (Rhombohedral, S.G. 166) Intensity (arb. unit)2 (degree)(111) (200) (220) (311) (222) (400) (331) (420)20304050607080 4D8GSB 6D6GSB 8D4GSB(b) Intensity (arb. unit)2 (degree)(111) (200) (220) (311) (222) (400) (331) (420) Figure 6-7. XRD patterns of 8D4GSB, 6D6GSB and 4D8GSB (a) first calcined (b) second calcined at 800 oC for 16 hours.

PAGE 143

14320304050607080 (a)JCPDS (Rhombohedral, S.G. 166) JCPDS (Cubic, S.G. 225) 5D10GSB 7.5D7.5GSB 10D5GSBIntensity (arb. unit)2 (degree)(111) (200) (220) (311) (222) (400) (331) (420)20304050607080 (b)5D10GSB 7.5D7.5GSB 10D5GSBIntensity (arb. unit)2 (degree)(111) (200) (220) (311) (222) (400) (331) (420) Figure 6-8. XRD patterns of 10D 5GSB, 7.5D7.5GSB and 5D10GSB after (a) first calcination and (b) second calcination at 800 oC for 16 hours.

PAGE 144

14420304050607080 JCPDS (Cubic, S.G. 225) JCPDS (Rhombohedral, S.G. 166) 12D6GSB-2nd 12D6GSB-1st Intensity (arb. unit)2 (degree)(111) (200) (220) (311) (222) (400) (331) (420) Figure 6-9. XRD patterns of 12D 6GSB after the first and second calcinations steps at 800 oC for 16 hours.

PAGE 145

145 1.01.11.21.31.41.51.61.71.8 -4.5 -4.0 -3.5 -3.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 8D4GSB 10D5GSB 12D6GSB_HL 12D6GSB_LH 20ESBT (oC) Log (S/cm) 1000/T (K-1)700600500400 300 Figure 6-10. Arrhenius plot of conductivity for 8D4GSB, 10D5 GSB, 12D6GSB and 20ESB; the conductivities were measured in both dire ction of temperature measurements, HL (High to Low) and LH (Low to High) for 12D6GSB.

PAGE 146

146 020406080100120 -3.0 -2.5 -2.0 -1.5 -1.0 8D4GSB 10D5GSB 12D6GSB 20ESB Log (S/cm)Time (hour) Figure 6-11. Conductivity vs. time for 8D4GSB 10D5GSB and 12D6GSB annealed at 500 oC; 20ESB is shown for comparison.

PAGE 147

147 20304050607080 Annealed for 100h As sintered(a) Intensity (arb. unit) 2(degree) 8D4GSB(111) (200) (220) (311) (222) (400) (331) (420)20304050607080 Annealed for 24h Annealed for 100h Intensity (arb. unit)2 (degree)As sintered(b)10D5GSBJCPDS (Cubic, S.G. 225) JCPDS (Rhombohedral, S.G. 166) (111) (200) (220) (311) (222) (400) (331) (420) Figure 6-12. XRD patterns of (a) 8D4GSB and (b) 10D5GSB as-sintered and annealed at 500 oC for 100 hours.

PAGE 148

148 20304050607080 -Bi2O3+CeO2 15CSB 20CSB 30CSB 40CSB CeO2 Intensity (arb. unit)2 (degree) Figure 6-13. XRD patterns of calcined (CeO2)x(Bi2O3)1-x (x=0.15, 0.20, 0.30 and 0.40).

PAGE 149

149 020406080100 -3.0 -2.5 -2.0 -1.5 -1.0 10D2CSB 20ESB Log (S/cm)Time (hour) 500oC Figure 6-14. Conductivity vs.time for 10D2CSB at 500 oC; 20ESB was added for comparison.

PAGE 150

150 0100200300400500600 -1.25 -1.00 -0.75 -0.50 -0.25 020406080100 -1.25 -1.00 -0.75 -0.50 -0.25 600oC 10D2CSB 8D4WSB 10D5WSB Log (S/cm)Time (hour) Figure 6-15. Conductivity vs. time for 10D2CSB, 8D4WSB a nd 10D5WSB at 600 oC; inset represents the conductivity behavior of these compositions for 100 hours.

PAGE 151

151 20304050607080 Bi0.775Dy0.225O1.5 (Rhombohedral)600oC (500h) 500oC (100h) Intensity (arb. unit)2 (degree)As sintered(111) (200) (220) (311) (222) (400) (331) (420) Figure 6-16. XRD patterns of 10D2CSB as sintered, annealed at 500 oC for 100 hours and annealed at 600 oC for 500 hours.

PAGE 152

152 CHAPTER 7 HIGH PERFORMANCE LSM-ESB AND LSM-DWSB COMPOSITE CATHODES 7.1 Introduction It has been recognized that La1-xSrxMnO3(LSM) is the most promising cathode material for high-temperature solid oxide fuel cells (SOFCs) due to its good thermal and chemical stability [92, 93]. Currently, considerable effo rts have been devoted to reducing cell operating temperature because lower temperature operation a llows for a wider choice of interconnect and other balance-of-plant materials and also results in reduced pr oblems with sealing and thermal degradation [10, 94]. However, LSM is not suitable for intermediate temperature (IT) SOFCs because of its negligible ionic conductivity as well as its high activation energy for oxygen dissociation. To improve cathode performance, an ionic conducting phase such as gadoliniadoped ceria (GDC) and yttria-stab ilized zirconia (YSZ) was added in to LSM to form a composite cathode, and resulted in a significa nt decrease in the interfacial resistance compared with pure LSM [95-98]. It is well unde rstood that cathode properties can be enhanced using ionic conductors with high conductivity by increasi ng triple phase boundary (TPB) lengths [98]. Murray et al. showed that the area specific resistance (ASR) of LSM-GDC composite cathodes was two to three times lower than that of LSM-YSZ composite cat hodes on YSZ electrolytes [96]. Recently, bismuth based oxides have been examin ed as a promising material for electrolyte and cathode in IT-SOFCs because of their high oxygen ionic conductivity [20, 99]. It has also been shown that bismuth oxides improve catal ytic effects on the oxygen dissociation reaction which is often considered to be the rate-limiti ng step in the oxygen re duction reaction of SOFC cathodes [100-103]. Therefore, many studies were car ried out using stabili zed bismuth oxides as

PAGE 153

153 the ionic conducting phase to fu rther reduce interfacial polarization resistance of composite cathodes [92, 93, 99, 104-106]. In this study, two LSM-cont aining composite cathodes: LS M-ESB and LSM-DWSB were designed to investigate cathode performance and chemical compatibility. The composition details of each composite cathode are (La0.80Sr0.20)MnO3-(Er2O3)0.20(Bi2O3)0.80 (LSM-ESB) and (La0.80Sr0.20)MnO3(Dy2O3)0.08(WO3)0.04(Bi2O3)0.88 (LSM-DWSB). LSMESB cathodes were used on ESB and Gd0.1Ce0.9O2(GDC) electrolytes, respective ly, to identify the effect of electrolytes on cathode polarization. 7.2 Experimental Procedure 7.2.1 Electrode Preparation Conventional LSM powder with a surface area of 5.6 m2/g was purchased from Fuel Cell Materials. To obtain LSM-ESB (or DWSB) el ectrode slurries, LSM and ESB (or DWSB) powders of the same weight ratio were mixed with alpha terpineol (Alfa Aesar), Di-n-butyl phthalate (DBP) and ethanol. Once an appropriate viscosity was reached, the electrode slurry was applied to both sides of the electrolyte substrates by brush pain ting. After drying the symmetric cells at 120 oC for 1 hour, another coating of elect rode slurry was applied to the electrolyte substrates. The doubly-coated cells were then sintered at 800 oC for 2 hours in air. After sintering the electrode, silv er mesh current collectors and pl atinum lead wires were pressed against the samples in a quartz reactor using a ceramic screw-and-bolt assembly. 7.2.2 Electrolyte Preparation In this work, three different electrolyte substrates 20ESB, 8D4WSB and 10GDC were fabricated in order to compare the effect of the electrolyte on cathode performance. To get dense Gd0.1Ce0.9O2(10GDC) pellets, 10GDC (Anan Kasei Co. Ltd., Japan) was used. Details on the fabrication of bismuth oxide-based and 10GDC elect rolyte pellets were described in Chapter 3.

PAGE 154

1547.2.3 Characterizations Electrochemical performance of the electr odes was carried out through electrochemical impedance spectroscopy (EIS) using a Solartron 1260. The impedance of a symmetric cell was measured with an ac voltage amplitude of 50 mV over the frequency range of 0.1 MHz to 0.1 Hz in air. The frequency response analyzer was us ed in standalone mode and interfaced to a computer using Zplot software. Measurements were made from 500 oC to 750 oC in 50 oC increments. XRD patterns (XRD, Philips APD 3720) were obtained using CuK radiation at room temperature between 20o and 80o (2 ). Electrode microstructure s were analyzed by scanning electron microscopy (SEM). Secondary and backs cattered microstructural images were observed through a JEOL JSM 6400 SEM. 7.3 Results and Discussions 7.3.1 LSM-20ESB Composite Cathode Figure 7-1 shows XRD patterns of LSM and ESB powder mixtur es before and after heattreatment at 900 oC for 50 hours. LSM and ESB solid soluti ons are consisted of perovskite and cubic structures. No additional phases were observed after heat treatment, suggesting these materials are chemically compatible. It is assumed that better connectivity of the ionic c onduction phase and electronic conduction phase effectively extends the triple phase boundary (TPB), resulting in enhanced electrochemical activity of the composite cathode [92]. In fact, studies have shown that the optimal composition of composite cathodes with ESB as an ionic conduction phase was achieved with about 50wt.% ESB content [92, 99]. Theref ore, in this study, LSM-ESB mixtures were prepared by mixing LSM and ESB powde rs in equal amounts (50wt.%).

PAGE 155

155 Figure 7-2 shows impedance spec tra measured in air at 700 oC for the three different cells. Electrolyte ASR has been subtracted from th e real component of each data point for easy comparison of cathode ASR. In this way, the lo w-frequency intercept in real-axis corresponds to the cathode polari zation resistance. Table 7-1 shows the activation energy and area specific resistance (ASR) of these cells. At 700 oC, the ASR of Cell 1 was 0.26 cm2, while that of Cell 3 was 0.08 cm2 So, the polarization of LSM-ESB is about four times lower than that of pure LSM with the same ESB electrolyte. This result demonstrates that th e addition of ESB phase increases the TPB and consequently leads to improved cathode performance. It is also important to understand the effect of the electrolyte substrate on the interfacial resistance. Figure 7-3 shows the ASR for three different cells as a function of temperature. Murray et al. reported that ASR of pure LSM is 7.82 cm2 and 2.67 cm2 on YSZ and GDC electrolyte at 700 oC, respectively [96]. In this study, the ASR of pure LSM on ESB electrolyte was about 30 times lower than that of pure LS M on YSZ electrolyte. This result clearly shows that pure LSM has better cathode performance on ESB electrolyte than on either YSZ or GDC electrolytes. This may be attribut ed to the fact that bismuth oxide -based electrolytes enhance the oxygen dissociation and surface oxygen exchange rate [99, 102, 103, 107]. In this study, the ASR of LSM-ESB com posite cathode was obtai ned on two different electrolytes (ESB and GDC). The same cathode exhibited about 70% smaller polarization resistance on ESB than on GDC. The stability of cell 3 (LSM-ESB cathode) was also examined. Figure 7-4 shows that LSM-ESB composite cathode maintained the sa me performance for 100 hours in the range of

PAGE 156

156 0.08 0.001 cm2 while that of Ag-ESB increased by more than 70% after 100 hours at 650 oC [106]. 7.3.2 LSM-8D4WSB Composite Cathode In previous Chapters, we found that 8D4WSB has higher conductivity than single doped 20ESB. Therefore, it is expected that even better cathode performance can be achieved by replacing 20ESB with 8D4WSB in composite cathode. To check compatibility between LSM and 8D4WSB, heat treatment was carried out. Figure 7-5 shows XRD patterns of LSM and 8D4WSB powder mixtures before and after heat-treatment at 900 oC for 50 hours. As was the case for LSM-ESB composite cathodes, no additi onal peaks were observed except for a small peak at ~ 28.3o (2 ). In the same way, LSM-8D4WSB (LSM-DWSB) mixtures were prepared by mixing equal amounts of LSM and 8D4WSB powders (50wt.%) Figure 7-6 shows th e ASR of LSM-DWSB on ESB (Cell 4) and LSM-DWSB on DWSB (Cell 5) along with previous results for Cell 1 and Cell 3 for comparison. Cell 5 had almost same cathode ASR as Cell 3 even though electrolyte ASR of Cell 5 was lower than that of Cell 3 due to the higher c onductivity of 8D4WSB compared with 20ESB. In addition, the LSM-DWSB composite cathode had higher ASR than LSM-ESB composite cathode on the same ESB elect rolyte substrate. De tails on ASR of these cells at each temperature are shown in Table 7-1. This may be attributed to a lattice mismatch between 8D4WSB and 20ESB in cathode-electrolyte interface. 8D4WSB has higher ionic conductivity than 20ESB, but further study will be necessary to use this composition for the composite cathode material to achieve better performance than LSM-ESB. Figure 7-7 shows cross-sectional backscattered images of two different cells. Both samples show good adhesions between composite cathode and electrolyte. LSM and ESB also appear to be uniformly dispersed. Microstructural optim ization based on partic le size and relative

PAGE 157

157 composition of two constituents in composite ca thodes is expected to further enhance the cathode performance.

PAGE 158

158 20304050607080 (128) (134) (208) (220) (018) (214) (024) (006) (202) (104) (110) (012) (c) LSM+ESB before annealing (d) LSM+ESB after annealing (a) Pure LSM Intensity (arb. unit)2 (degree)(b) 20ESB(111) (200) (220) (311) (222) (400) (331) (420) Figure 7-1. XRD patterns of (a) Pure LSM (b ) 20ESB (c) a mixture of LSM and ESB (d) a mixture of LSM and ESB, after firing at 900 oC for 50 hours

PAGE 159

159 0.000.050.100.150.200.25 0.00 -0.05 -0.10 -0.15 -0.20 Pure LSM on ESB (Cell 1) LSM-ESB on GDC (Cell 2) LSM-ESB on ESB (Cell 3) Z'' ( cm2)Z' ( cm2) Figure 7-2. Impedance spectra of Pure LSM on ESB (Cell 1), LSM-ESB on GDC (Cell 2) and LSM-ESB on ESB (Cell 3) measured at 700 oC in air.

PAGE 160

160 1.0 1.1 1.2 1.3 0.1 1 10 Pure LSM on ESB (Cell 1) LSM-ESB on GDC (Cell 2) LSM-ESB on ESB (Cell 3) LSM-YSB on SDC, Jiang et al. ASR ( cm2)1000/T (K-1) T (oC) 700650600550500 Figure 7-3. Arrhenius plot of ASR vs. temper ature for Pure LSM on ESB (Cell 1), LSM-ESB on GDC (Cell 2), LSM-ESB on ESB (Cell 3) and Jiang et al. [93].

PAGE 161

161 020406080100 0.00 0.02 0.04 0.06 0.08 0.10 ASR (cm2)Time (h) Figure 7-4. ASR vs. time for LS M-ESB on ESB (Cell 3) at 700 oC.

PAGE 162

162 20304050607080 (c) LSM+8D4WSB before annealing (d) LSM+8D4WSB after annealing (a) Pure LSM Intensity (arb. unit)2 (degree)(b) 8D4WSB (128) (134) (208) (220) (018) (214) (024) (006) (202) (104) (110) (012)(111) (200) (220) (311) (222) (400) (331) (420) Figure 7-5. XRD patterns of (a) Pure LSM (b) 8D4WSB (c) a mixture of LSM and 8D4WSB (d) a mixture of LSM and 8D4W SB, after firing at 900 oC for 50 hours

PAGE 163

163 1.0 1.1 1.2 1.3 0.1 1 10 Pure LSM on ESB (Cell 1) LSM-DWSB on ESB (Cell 4) LSM-ESB on ESB (Cell 3) LSM-DWSB on DWSB (Cell 5)T (oC) ASR ( cm2)1000/T (K-1)700650600550500 Figure 7-6. Arrhenius plot of ASR vs. temp erature for LSM-DWSB on ESB(Cell 4) and LSMDWSB on DWSB (Cell 5) including previous Cell 1 and Cell 3.

PAGE 164

164 ESB electrolyte LSM ESB (a) ESB electrolyte LSM ESB ESB electrolyte LSM ESB (a) LSM ESB GDC electrolyte (b) LSM ESB GDC electrolyte LSM ESB GDC electrolyte (b) Figure 7-7. Cross-sectional mi crographs of (a) LSM-ESB cathode on ESB (Cell 3) (b) LSM-ESB cathode on GDC (Cell 2).

PAGE 165

165 Table 7-1. Activation energy and AS R of three different samples Cell Type Composition Ea (KJ/mol) ASR ( cm2) at 700 oC ASR ( cm2) at 600 oC ASR ( cm2) at 500 oC Cell 1 Pure LSM on ESB 130 0.26 1.63 16.4 Cell 2 LSM-ESB on GDC 120 0.20 1.13 9.1 Cell 3 LSM-ESB on ESB 120 0.08 0.46 4.0 Cell 4 LSM-DWSB on ESB 115 0.10 0.51 4.2 Cell 5 LSM-DWSB on DWSB 120 0.08 0.45 3.9

PAGE 166

166 CHAPTER 8 SUMMARY Dysprosiumand tungsten-stab ilized bismuth oxide (DWS B) was developed and its structure and conductivity were examined. Through double doping, the total dopant concentration required to stabilize the cubic structure was reduced down to 12 mol% for this DWSB electrolyte system. We found that the 8D4WSB composition had the highest conductivity, achieving = 0.57 and 0.098 S/cm at 700 and 500 oC, respectively. The conductivity and lattice parameter of DWSB com positions increased linearly as the total dopant concentration decreased with fi xed 2:1 dopant ratio within the fcc phase-stability window. In addition, the disparity in conduc tivity activation energies abov e and below the order-disorder transition decreased as the total dopant concentrati on increased. Time-dependent conductivity behavior of th ese DWSB compositions was investigated between 300 oC and 700 oC as a function of anneal temp erature. The effect of dopant concentration on long term conductivity was al so examined. All DWSB compositions were stable and maintained thei r initial conduc tivity at 700 oC, but displayed conductivity degradation between 400 oC and 650 oC. Phase transformation and ordering of the oxygen sublattice were identified as the source of conductivity de gradation. 8D4WSB, th e highest conductivity composition, displayed phase instability between 500 and 600 oC and experienced the fastest conductivity degradation rate at 500 oC. The degradation rate for 8D4WSB was markedly lower below 500 oC because the kinetics of ordering and/ or phase degradation decreased with temperature. In order to enhance long term stability, new DWSB compositions were fabricated. Of these DWSB compositions, 25D5WSB maintained the highest conductivity of 0.0068 S/cm with phase and structural stability while annealing at 500 oC for 500 hours. With respect to total

PAGE 167

167 conductivity, 25D5WSB had much higher total conductivity than 10GDC even after long term annealing due to its negligib le grain boundary resistance. On the other hand, singly doped 25DSB experienced phase change from cubic to rhombohedral during long periods of annealing at 500 oC, however, the present cubic phase of 25D5WSB had only negligible conductivity degradation with the true cubic phase stabilization at this temperature. Further, new doubly doped bismuth oxides-based electrolytes were developed to enhance long term stability while maintaining the high initial conductivity at 500 ~ 600 oC. 25D5WSB had excellent long term stability for 500 hours, but had relatively low conductivity due to its higher total dopant c oncentrations. (Tb4O7)-(WO3)-(Bi2O3), (Dy2O3)-(Gd2O3)-(Bi2O3) and (Dy2O3)-(CeO2)-(Bi2O3) electrolyte systems were prep ared. Among those systems, (Dy2O3)0.10(CeO2)0.02-(Bi2O3)0.88 was a promising composition which sa tisfies both goals of conductivity and stability at 600 oC. This composition was very stable for first ~ 200 hours, but experienced a gradual decrease in conduc tivity due to a phase transformation. It is expected that long-term stability can be enhanced with slig htly higher concentrations of CeO2 without compromising its initial conductivity. Lastly, LSM-ESB and LSM-DWSB composite cathodes were developed. These composite cathodes had significantly better performance than pure LSM on ESB electrolytes. It was also found that LSM-ESB had better performance on ESB than on GDC due to the high catalytic effect of bismuth oxide-based el ectrolytes on oxygen dissociation.

PAGE 168

168 APPENDIX A CONDUCTIVITY COMPARISON The conductivity of 8D4WSB was observed to be higher in this study than in our previous work [13]. To find out why, the conductivities of several 8D4WSB compositions of different geometries were measured with different cu rrent collectors. Each nulling file was made depending on the current collector used. Table A-1 shows sample dimension and the current collector used for 8D4WSB a nd 12D6WSB. Figure A-1(a) shows that all 8D4WSB compositions had similar conductivity w ithin experimental error over the entire temperature range, but had a slightly higher conductivity than our previous work for 8D4W SB [13]. We susp ect that artifacts caused by inductive response of the test leads an d the equipment was not fully compensated in our previous results [13]. However, the fact th at our currently measured ESB conductivity is the same as the literature gives us confidence in our current results. In additi on, the fact that the linear fit of Figure 3-7(a) corresponds to pure -Bi2O3 at 700 oC also validates our current results. As shown in Figure A-1(b), the conductivity va lue is also almost the same for 12D6WSB composition regardless of sample dimension ev en though there is little deviation at high temperature due to increased fractional error in the measurement for small resistance values.

PAGE 169

1691.01.11.21.31.41.51.61.71.8 -4.5 -4.0 -3.5 -3.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 (a)T (oC) Log (S/cm) 1000/T (K-1) 8D4WSB-1 8D4WSB-2 8D4WSB-3 8D4WSB-4 8D4WSB-Previous work 20ESB700600500400 300 1.01.11.21.31.41.51.61.71.8 -4.5 -4.0 -3.5 -3.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 (b) 12D6WSB-1 12D6WSB-2T (oC) Log (S/cm) 1000/T (K-1)700600500400 300 Figure A-1. Reproducibility data for (a) 8D4WSB (this work and previous work [13]) and 20ESB (b) 12D6WSB.

PAGE 170

170 Table A-1. Sample dimension and the current collector used for 8D4WSB and 12D6WSB to confirm reproducibility. Composition Sample dimension (mm) (Thickness Diameter) Current collector 8D4WSB-1 2.87 7.34 Ag wire + Ag mesh 8D4WSB-2 2.98 7.19 Au wire 8D4WSB-3 2.96 7.19 Pt wire + Ag mesh 8D4WSB-4 7.19 7.25 Pt wire + Ag mesh 12D6WSB-1 3.04 7.12 Ag wire + Ag mesh 12D6WSB-2 6.61 7.14 Pt wire + Ag mesh

PAGE 171

171 APPENDIX B LATTICE PARAMETER CALCULATION X-ray diffraction patterns were obtained by XRD Philips APD 3720 in Major Analytical Instrumentation Center. Extrapolation method using Nelson-Rile y function was carried out to estimate the lattice parameter of cubic-stabilized bismuth oxide s [108]. For a given XRD pattern, lattice parameter (a) was determined for each individual pe ak using the following relationship for cubic structure. 222sin2 lkh a (B-1) The measured lattice parameters were then pl otted as a function of Nelson-Riley function. 2 2cos sin cosK a a d d (B-2) This equation holds quite accurately down to very low values of and not just at high angles. The bracketed terms are called the Nelson-R iley function. The fitted line is in the form of the following equation, KNaaa00 ( 2 2cos sin cos N ) (B-3) where 0a is the true estimation of the lattice parameter and a is the apparent lattice parameter calculated from each peak position. Figure C-1 shows the apparent measured lattice parameters as a function of Nelson-Riley function for (Dy2O3)0.10(WO3)0.05(Bi2O3)0.85 (10D5WSB) taken at room temperature. From the ex trapolation, the true lattice parameter of this composition is 5.553 This extrapolation method was also applied to high temperature XRD patterns. Figure B-2 shows the XRD patterns of (Tb4O7)0.25(Bi2O3)0.75 (25TSB) which was taken at 25 oC, 300 oC, 500 oC and 600 oC. Using Nelson-Riley function, Figure B-3 shows the apparent lattice parameter

PAGE 172

172 versus Nelson-Riley function plots for 25TSB at di fferent temperatures. The increase in intercept is observed with the increase in temperature. This indicates the thermal lattice expansion of cubic-stabilized bismuth oxides.

PAGE 173

173 02468 5.542 5.544 5.546 5.548 5.550 5.552 Y = 5.553 0.0015XNelson-Riley function (N) Measured lattice parameter () Figure B-1. Calculated lattice parameter versus Nelson-Riley function for 10D5WSB taken at room temperature.

PAGE 174

174 30405060 600oC 500oC 300oC 25oC Intensity (arb. unit)2 (degree) Figure B-2. XRD patterns of 25TSB measured at 25 oC, 300 oC, 500 oC and 600 oC.

PAGE 175

175 345678 5.48 5.49 5.50 5.51 5.52 5.53 5.54 5.55 5.56 600oC 500oC 300oCNelson-Riley function (N) Measured lattice parameter ()25oC Figure B-3. Calculated lattice parameter versus Nelson-Riley function for 25TSB at different temperatures.

PAGE 176

176 APPENDIX C CHARACTERIZATION OF XRD PATTERNS In order to obtain phase iden tification and the relative amount of each phase in mixtures, CrystalMaker & CrystalDiffract Programs were us ed. CrystalMaker enables us to build real crystal structure with structure parameters. Crys talDiffract program can generate the diffraction patterns using the structure parameters which incl ude the types and positio ns of atoms in a unit cell of a crystal. CrystalDiffract also allows us to work with observed and calculated data. By comparing an observed diffraction pattern with a nu mber of calculated patterns, we can identify an unknown substance and further k now the relative amount of each phase in case of mixtures. For example, it was found that 25DSB had the mi xtures of two phases after it was annealed at 500 oC for 300 hours. It was analyzed that the mi xtures were consisted of cubic phase (Fm 3m, Space Group 225) and rhombohedral phase (Bi0.775Dy0.225O1.5, R 3m, Space Group 166). To obtain the calculated diffraction pa ttern of two phases, structural parameters of cubic phase and rhombohedral phase were obtained from Yashima et al. [27] and Dr ache et al. [109], respectively. In case of cubic phase, lattice parameter at room temperature was chosen from Verkerk et al. [38]. The details of these structural paramete rs are given at Table C-1 and Table C-2. Using CrystalDiffract program, we can genera te calculated XRD patterns of mixtures of two phases and compare this pattern with observed XRD pattern. Figure C-1 shows the calculated XRD pattern of mixtures (Blue) and the observed XRD pattern (Red). By manipulating the relative amount of each phase of calculated mixtures, it was found that the observed pattern of 25DSB is approximately composed of 60% cubic phase and 40% rhombohedral phase. As was the case for 25DSB, this analysis was performed for aged DWSB compositions. Figure C-2 shows the observed XRD pattern of 10D5WSB which was annealed at 500 oC for 100

PAGE 177

177 hours (Red). This pattern is compared with th e calculated mixtures of cubic, tetragonal (Bi14WO24) [80] and orthorhombic (Bi2WO6) [81]. By manipulating the re lative amount of each phase of calculated mixtures, it was analyzed that the observed pa ttern of 10D5WSB is approximately composed of 61% cubic phase, 31% or thorhombic phase and 8% tetragonal phase.

PAGE 178

178 Figure C-1. XRD patterns of obs erved 25DSB annealed at 500 oC for 300 hours and calculated mixtures of cubic and rhombohedral phases.

PAGE 179

179 Figure C-2. XRD patterns of obs erved 10D5WSB annealed at 500 oC for 100 hours and calculated mixtures of cubic, ort horhombic and tetragonal phases.

PAGE 180

180 Table C-1. Structure parameters of cubic Bi2O3 [27, 38]. a b c Lattice Parameter 5.49 5.49 5.49 90 90 90 Atoms in the Asymmetric Unit # Label Site Occupancy x y z 1 Bi Bi 1.00 0 0 0 2 O1 O 0.23 0.25 0.25 0.25 3 O2 O 0.13 0.335 0.335 0.335

PAGE 181

181 Table C-2. Structure parameters of Bi0.775Dy0.225O1.5 (Rhombohedral) [109]. a b c Lattice Parameter 3.9649 3.9649 3.9649 90 90 120 Atoms in the Asymmetric Unit # Label Site Occupancy x y z 1 Bi 2 Bi 1.00 0 0 0.2252 2 Bi 1 Dy 0.68 Bi 0.32 0 0 0 3 O 1 O 1.00 0 0 0.3 4 O 2 O 0.72 0 0 0.0910 5 O 3 O 0.53 0 0 0.4520

PAGE 182

182 LIST OF REFERENCES [1] O. Yamamoto, Electrochim Acta 45 (2000) 2423. [2] B. C. H. Steele, J Mater Sci 36 (2001) 1053. [3] V. V. Kharton, F. M. B. Marques and A. Atkinson, Solid State Ionics 174 (2004) 135. [4] Y. Arachi, H. Sakai, O. Yamamoto, Y. Takeda and N. Imanishai, Solid State Ionics 121 (1999) 133. [5] S. J. Skinner and J. A. Kilner, Materials Today 6 (2003) 30. [6] H. Inaba and H. Tagawa, Solid State Ionics 83 (1996) 1. [7] Z. H. Wang, K. N. Sun, S. Y. Shen N. Q. Zhang, J. S. Qiao and P. Xu, J Membrane Sci 320 (2008) 500. [8] S. D. Souza, S. J. Visco and L. C. D. Jonghe, Solid State Ionics 98 (1997) 57. [9] J. Schoonman, J. P. Dekker, J. W. Broers and N. J. Kiwiet, Solid State Ionics 46 (1991) 299. [10] N. Q. Minh, J Am Ceram Soc 76 (1993) 563. [11] A. M. Azad, S. Larose and S. A. Akbar, J Mater Sci 29 (1994) 4135. [12] E. D. Wachsman, J Eur Ceram Soc 24 (2004) 1281. [13] N. Jiang, E. D. Wachsman and S. Jung, Solid State Ionics 150 (2002) 347. [14] S. Boyapati, E. D. Wachsman and N. Jiang, Solid State Ionics 140 (2001) 149. [15] T. Takahashi, T. Esaka and H. Iwahara, Journal of Applied Electrochemistry 7 (1977) 299. [16] J. Y. Park, H. Yoon and E. D. Wachsman, J Am Ceram Soc 88 (2005) 2402. [17] DOE, "Fuel Cell Handbook" (EG& G Technical Services, Inc., 2004). [18] H. A. Harwig, Z Anorg Allg Chem 444 (1978) 151. [19] H. A. Harwig and J. W. Weenk, Z Anorg Allg Chem 444 (1978) 167. [20] N. M. Sammes, G. A. Tompsett, H. Nafe and F. Aldinger, J Eur Ceram Soc 19 (1999) 1801. [21] H. A. Harwig and A. G. Gerards, Journal of Solid State Chemistry 26 (1978) 265.

PAGE 183

183 [22] L. G. Sillen, Ark. Kemi. Mineral. Geol. 12A (1937) 1. [23] G. Gattow and H. Schroder, Z Anorg Allg Chem 318 (1962) 176. [24] B. T. M. Willis, Acta Crystallogr 18 (1965) 75. [25] P. D. Battle, C. R. A. Catlow, J. Drennan and A. D. Murray, J Phys C Solid State 16 (1983) L561. [26] S. Boyapati, E. D. Wachsman and B. C. Chakoumakos, Solid State Ionics 138 (2001) 293. [27] M. Yashima and D. Ishimura, Chem Phys Lett 378 (2003) 395. [28] M. Yashima and D. Ishimura, Appl Phys Lett 87 (2005). [29] D. M. Collins, Nature 298 (1982) 49. [30] M. Takata, B. Umeda, E. Nishibori, M. Sakata, Y. Saito, M. Ohno and H. Shinohara, Nature 377 (1995) 46. [31] H. M. Rietveld, J Appl Crystallogr 2 (1969) 65. [32] D. S. Aidhy, J. C. Nino, S. B. Sinnott, E. D. Wachsman and S. R. Phillpot, J Am Ceram Soc 91 (2008) 2349. [33] N. Jiang and E. D. Wachsman, J Am Ceram Soc 82 (1999) 3057. [34] E. D. Wachsman, S. Boyapati and N. Jiang, Ionics 7 (2001) 1. [35] T. Takahashi and H. Iwahara, Journal of Applie d Electrochemistry 2 (1972) 97. [36] T. Takahashi, T. Esaka and H. Iwahara, Journal of Applied Electrochemistry 5 (1975) 197. [37] T. Takahashi, H. Iwahara and T. Arao, Journal of Applied Electrochemistry 5 (1975) 187. [38] M. J. Verkerk and A. J. Burggraaf, Journal of the Electrochemical Society 128 (1981) 75. [39] G. Y. Meng, C. S. Chen, X. Han, P. H. Yang and D. K. Peng, Solid State Ionics 28 (1988) 533. [40] D. Mercurio, M. Elfarissi, B. Frit, J. M. Reau and J. Senegas, Solid State Ionics 39 (1990) 297. [41] K. Hu, C. H. Chen, D. K. Peng and G. Y. Meng, Solid State Ionics 28 (1988) 566. [42] K. Shirao, T. Iida, K. Fukushima and Y. Iwadate, Journal of Alloys and Compounds 281 (1998) 163.

PAGE 184

184 [43] E. D. Wachsman, S. Boyapati and N. X. Jiang, Ionics 7 (2001) 1. [44] R. K. Datta and J. P. Meehan, Zeitschrift Fur Anorganische Und Allgemeine Chemie 383 (1971) 328. [45] A. V. Joshi, S. Kulkarni, J. Nachla s, J. Diamond, N. Weber and A. V. Virkar, Journal of Materials Science 25 (1990) 1237. [46] A. Watanabe, Solid State Ionics 40-1 (1990) 889. [47] K. Z. Fung and A. V. Virkar, J Am Ceram Soc 74 (1991) 1970. [48] K. Z. Fung, J. Chen and A. V. Virkar, J Am Ceram Soc 76 (1993) 2403. [49] K. Z. Fung, H. D. Baek and A. V. Virkar, Solid State Ionics 52 (1992) 199. [50] N. X. Jiang and E. D. Wachsman, Journal of the American Ceramic Society 82 (1999) 3057. [51] E. D. Wachsman, N. Jiang, D. M. Mason and D. A. Stevenson, Proceedings Electrochemical Society 89 (1989) 15. [52] E. D. Wachsman, S. Boyapati, M. J. Kaufman and N. Jiang, J Am Ceram Soc 83 (2000) 1964. [53] N. Jiang, R. M. Buchanan, D. A. Steven son, W. D. Nix, J. Z. Li and J. L. Yang, Mater Lett 22 (1995) 215. [54] C. M. Earnest, Thermochimica Acta 137 (1989) 365. [55] D. Pletcher, R. Greef, R. Peat, L. M. Peter and J. Robi nson, "Instrumental Methods in Electrochemistry" (Ellis horwood, 1985). [56] J. E. Bauerle, Journal of Physics and Chemistry of Solids 30 (1969) 2657. [57] I. Denk, J. Claus and J. Maier, J Electrochem Soc 144 (1997) 3526. [58] S. Omar, E. D. Wachsman and J. C. Nino, Solid State Ionics 177 (2006) 3199. [59] J. R. Macdonald, "Impedance Spectroscopy" (John Wiley & Sons, 1987). [60] T. S. Zhang, J. Ma, S. H. Chan, P. Hing and J. A. Kilner, Solid State Sci 6 (2004) 565. [61] P. Duran, J. R. Jurado, C. Moure, N. Valverde and B. C. H. Steele, Mater Chem Phys 18 (1987) 287. [62] M. J. Verkerk and A. J. Burggraaf, J Electrochem Soc 130 (1983) 78. [63] L. S. Wang and S. A. Barnett, J Electrochem Soc 139 (1992) 2567.

PAGE 185

185 [64] T. Takahashi, T. Esaka and H. Iwahara, J Solid State Chem 16 (1976) 317. [65] T. Takahashi, H. Iwahara and Y. Nagai, Journal of Applie d Electrochemistry 2 (1972) 97. [66] M. J. Verkerk, K. Keizer and A. J. Burggraaf, Journal of Applied Electrochemistry 10 (1980) 81. [67] T. Takahashi and H. Iwahara, Journal of Applie d Electrochemistry 3 (1973) 65. [68] S. N. Hoda and L. L. Y. Chang, J Am Ceram Soc 57 (1974) 323. [69] A. Watanabe and A. Ono, Solid State Ionics 174 (2004) 15. [70] S. Jung, E. D. Wachsman and N. Jiang, Ionics 8 (2002) 210. [71] R. D. Shannon, Acta Crystallographica, Section A A32 (1976) 751. [72] B. D. Cullity, "Elements of X -ray diffraction" (Addison-Wesley, 1977). [73] V. V. Kharton, E. N. Naumovich, A. A. Yaremchenko and F. M. B. Marques, J Solid State Electr 5 (2001) 160. [74] M. Yashima, D. Ishimura and K. Ohoyama, J Am Ceram Soc 88 (2005) 2332. [75] B. C. H. Steele, "High Conductivity So lid Ionic Conductors" (World Scientific, 1989). [76] M. J. Verkerk, G. M. H. Van de Ve lde, A. J. Burggraaf and R. B. Helmholdt, Journal of Physics and Chemistry of Solids 43 (1982) 1129. [77] A. Watanabe, Solid State Ionics 40-41 (1990) 889. [78] E. D. Wachsman, G. R. Ball, N. Jiang and D. A. Stevenson, Solid State Ionics 52 (1992) 213. [79] D. Jung, K. L. Duncan and E. D. Wachsman, Acta Mater (2009) submitted. [80] C. D. Ling, R. L. Withers, J. G. Thompson and S. Schmid, Acta Crystallogr B 55 (1999) 306. [81] K. S. Knight, Mineral Mag 56 (1992) 399. [82] A. Watanabe and T. Kikuchi, Solid State Ionics 21 (1986) 287. [83] A. A. Yaremchenko, V. V. Kharton, E. N. Naumovich, A. A. Tonoyan and V. V. Samokhval, J Solid State Electr 2 (1998) 308. [84] K. Zheng, B. C. H. Steele, M. Sahibzada and I. S. Metcalfe, Solid State Ionics 86-8 (1996) 1241.

PAGE 186

186 [85] D. Jung, K. L. Duncan, M. A. Camaratta, K. Lee, J. C. Nino and E. D. Wachsman, (2009) to be submitted. [86] S. Omar, E. D. Wachsman and J. C. Nino, Solid State Ionics 178 (2008) 1890. [87] H. Iwahara, T. Esaka, T. Sato and T. Takahashi, J Solid State Chem 39 (1981) 173. [88] H. T. Cahen, T. G. M. V. d. belt, J. H. W. D. wit and G. H. J. Broers, Solid State Ionics 1 (1980) 411. [89] T. Esaka and H. Iwahara, Journal of Applied Electrochemistry 15 (1985) 447. [90] P. P. Shuk, A. A. Vecher and V. V. Samochval, Thermochim Acta 93 (1985) 461. [91] K. Q. Huang, M. Feng and J. B. Goodenough, Solid State Ionics 89 (1996) 17. [92] J. L. Li, S. R. Wang, Z. R. Wang, R. Z. Liu, T. L. Wen and Z. Y. Wen, J Power Sources 179 (2008) 474. [93] Z. Y. Jiang, L. Zhang, K. Feng and C. R. Xia, J Power Sources 185 (2008) 40. [94] E. P. Murray, M. J. Sever and S. A. Barnett, Solid State Ionics 148 (2002) 27. [95] T. Kenjo, S. Osawa and K. Fujikawa, J Electrochem Soc 138 (1991) 349. [96] E. P. Murray and S. A. Barnett, Solid State Ionics 143 (2001) 265. [97] E. P. Murray, T. Tsai and S. A. Barnett, Solid State Ionics 110 (1998) 235. [98] H. Zhao, L. H. Huo and S. Gao, J Power Sources 125 (2004) 149. [99] M. Camaratta and E. Wachsman, J Electrochem Soc 155 (2008) B135. [100] J. C. Boivin and G. Mairesse, Chem Mater 10 (1998) 2870. [101] P. Shuk, H. D. Wiemhofer, U. Guth, W. Gopel and M. Greenblatt, Solid State Ionics 89 (1996) 179. [102] I. C. Vinke, K. Seshan, B. A. Bouka mp, K. J. Devries and A. J. Burggraaf, Solid State Ionics 34 (1989) 235. [103] B. A. Boukamp, I. C. Vinke, K. J. Devries and A. J. Burggraaf, Solid State Ionics 32-3 (1989) 918. [104] C. R. Xia, Y. Zhang and M. L. Liu, Appl Phys Lett 82 (2003) 901. [105] M. Camaratta and E. Wachsman, Solid State Ionics 178 (2007) 1411. [106] M. Camaratta and E. Wachsman, Solid State Ionics 178 (2007) 1242.

PAGE 187

187 [107] J. C. Boivin, C. Pirovano, G. Nowogroc ki, G. Mairesse, P. Labrune and G. Lagrange, Solid State Ionics 115 (1998) 639. [108] J. B. Nelson and D. P. Riley, P Phys Soc Lond 57 (1945) 160. [109] M. Drache, S. Obbade, J. P. Wignacourt and P. Conflant, J Solid State Chem 142 (1999) 349.

PAGE 188

188 BIOGRAPHICAL SKETCH Doh Won Jung was born in Seoul, South Korea. He earned a bachelors degree from the Department of Materials Science and Engineerin g, Korea University, Seoul, South Korea, in Feburary 2001. He entered the graduate program in the Department of Materials Science and Engineering at Korea University and experi enced the molten carbonate fuel cell under the guidance of Professor Dokyol Lee. He finished his masters de gree in Feburary 2003. In August 2004, he enrolled in the Department of Materials Sc ience and Engineering, University of Florida, to pursue Ph.D under the guidance of Professor Eric D. Wachsman.