<%BANNER%>

Fundamental Understanding of Surface Charactersitic and Chromium Contamination on Solid Oxide Fuel Cell Cathodes

Permanent Link: http://ufdc.ufl.edu/UFE0041560/00001

Material Information

Title: Fundamental Understanding of Surface Charactersitic and Chromium Contamination on Solid Oxide Fuel Cell Cathodes
Physical Description: 1 online resource (175 p.)
Language: english
Creator: Oh, Dong
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: cathode, cell, chromium, cobalt, contamination, degradation, elemental, enrichement, fuel, impregnation, mechanism, oxide, segregation, sofc, solid, surface, vaporization
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: A solid oxide fuel cell (SOFC) is a highly efficient and environmental-friendly energy conversion device. But the progress toward commercialization continues to be a slow struggle mostly due to high operation temperatures (800~1000oC). Lowering operation temperatures can bring manufacturing costs down and provide high conversion efficiency, and less difficulty in sealing and heat management. However, substantial increase in internal losses, especially activation overpotential, lowers SOFC performance with reduced operation temperatures. Because the activation overpotential is mostly attributed to the oxygen reduction reaction (ORR) on the cathode, tremendous works have been done in order to improve the cathode performance and understand the ORR mechanisms and degradation mechanisms. A heterogeneous reaction is primarily affected by the interaction between surface and adsorbed species. Therefore, the surface composition and structure are one of the most important factors affecting the cathode performance. Nevertheless, compositional and structural variations of cathode surfaces at high temperatures have not gained much attention. The objective of this study is, firstly, to better understand the surface structure of cathode materials at high temperatures, and, secondly, to improve the SOFC cathode performance based on this knowledge. La0.6Sr0.4Co0.2Fe0.8O3 has been widely used for SOFC cathodes. Chemical and structural variations of La0.6Sr0.4Co0.2Fe0.8O3 (LSCF) surfaces under oxidizing environment were investigated. Cr contamination free LSCF showed the formation of submicron-sized SrOx precipitates on the grain surface. This caused the reduced concentration of transition metals in B sites. The addition of cobalt oxide on the surface of LSCF was found to improve the cathode performance. In contrast, Cr vapor deposition caused the formation of larger SrCrO4 particles on the LSCF surface along grain boundaries, and Sr-deficient matrix. The structural analysis identified the phase transition from rhombohedral to cubic perovskite in due to Sr deficiency. A defect chemistry model was presented based on observed phenomena. Electrical conductivity relaxation, AC impedance spectroscopy and temperature programmed isotopic exchange (TPX) were conducted to study the impacts of Cr contamination on the cathode performance. It was found that Cr contamination was a chemical process for Co-Fe based materials and an electrochemical process for Mn based materials. Unstable Co4+ and Fe4+ lead to the formation of a nucleation agent for the chemical reaction. By contrast, stable Mn4+ does not lead to the formation of a nucleation agent. At last, a new hypothesis for the vaporization of Sr from La0.6Sr0.4Co0.2Fe0.8O3-? (LSCF) was proposed based on previous observations, and the evidence for Sr vaporization was provided.
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 Dong Oh.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Wachsman, Eric D.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041560:00001

Permanent Link: http://ufdc.ufl.edu/UFE0041560/00001

Material Information

Title: Fundamental Understanding of Surface Charactersitic and Chromium Contamination on Solid Oxide Fuel Cell Cathodes
Physical Description: 1 online resource (175 p.)
Language: english
Creator: Oh, Dong
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: cathode, cell, chromium, cobalt, contamination, degradation, elemental, enrichement, fuel, impregnation, mechanism, oxide, segregation, sofc, solid, surface, vaporization
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: A solid oxide fuel cell (SOFC) is a highly efficient and environmental-friendly energy conversion device. But the progress toward commercialization continues to be a slow struggle mostly due to high operation temperatures (800~1000oC). Lowering operation temperatures can bring manufacturing costs down and provide high conversion efficiency, and less difficulty in sealing and heat management. However, substantial increase in internal losses, especially activation overpotential, lowers SOFC performance with reduced operation temperatures. Because the activation overpotential is mostly attributed to the oxygen reduction reaction (ORR) on the cathode, tremendous works have been done in order to improve the cathode performance and understand the ORR mechanisms and degradation mechanisms. A heterogeneous reaction is primarily affected by the interaction between surface and adsorbed species. Therefore, the surface composition and structure are one of the most important factors affecting the cathode performance. Nevertheless, compositional and structural variations of cathode surfaces at high temperatures have not gained much attention. The objective of this study is, firstly, to better understand the surface structure of cathode materials at high temperatures, and, secondly, to improve the SOFC cathode performance based on this knowledge. La0.6Sr0.4Co0.2Fe0.8O3 has been widely used for SOFC cathodes. Chemical and structural variations of La0.6Sr0.4Co0.2Fe0.8O3 (LSCF) surfaces under oxidizing environment were investigated. Cr contamination free LSCF showed the formation of submicron-sized SrOx precipitates on the grain surface. This caused the reduced concentration of transition metals in B sites. The addition of cobalt oxide on the surface of LSCF was found to improve the cathode performance. In contrast, Cr vapor deposition caused the formation of larger SrCrO4 particles on the LSCF surface along grain boundaries, and Sr-deficient matrix. The structural analysis identified the phase transition from rhombohedral to cubic perovskite in due to Sr deficiency. A defect chemistry model was presented based on observed phenomena. Electrical conductivity relaxation, AC impedance spectroscopy and temperature programmed isotopic exchange (TPX) were conducted to study the impacts of Cr contamination on the cathode performance. It was found that Cr contamination was a chemical process for Co-Fe based materials and an electrochemical process for Mn based materials. Unstable Co4+ and Fe4+ lead to the formation of a nucleation agent for the chemical reaction. By contrast, stable Mn4+ does not lead to the formation of a nucleation agent. At last, a new hypothesis for the vaporization of Sr from La0.6Sr0.4Co0.2Fe0.8O3-? (LSCF) was proposed based on previous observations, and the evidence for Sr vaporization was provided.
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 Dong Oh.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Wachsman, Eric D.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041560:00001


This item has the following downloads:


Full Text

PAGE 1

1 FUNDAMENTAL UNDERSTANDING OF SURFACE CHARACTERSITIC AND CHROMIUM CONTAMINATION ON SOLID OXIDE FUEL CELL CATHODES By DONGJO OH A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILL MENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010

PAGE 2

2 2010 D ongjo Oh

PAGE 3

3 To my parent s; wife Chanyoung ; and daughter Narae

PAGE 4

4 ACKNOWLEDGMENTS I would like to thank my advisor Dr. Eric Wachsman for his s upport and guidance throughout this experience. I really appreciate the opportunity that he has provided me I also would like to thank Dr. Juan Nino, Dr. Susan Sinnott Dr. Scott Perry and Dr. Mark Orazem for their advice and participation as my committe e member I acknowledge Dr. Hee Sung Yun and Dr. Jiho Yoo for sharing their invaluable experience and guiding me toward a critical thinking and fundamental understanding of materials Finally, m y wife, Chanyoung, I always appreciate your support and generosity. You and our baby, Narae are the great est gift for me that Jesus has sent to me.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ...................................................................................................... 4 LIST OF TABLES ................................................................................................................ 8 LIST OF FIGURES .............................................................................................................. 9 LIST OF ABBREVIATIONS .............................................................................................. 14 ABSTRACT ........................................................................................................................ 16 CHAPTER 1 INTRODUCTION ........................................................................................................ 18 1.1 Dual C hallenges of E nergy and C limate C hange ............................................ 18 1.2 Solid Oxide Fuel Cell (SOFC) .......................................................................... 18 1.3 Objectives ......................................................................................................... 19 2 BACKGROUND .......................................................................................................... 21 2.1 Basic C omponents and O peration Principle of SOFC .................................... 21 2.2 SOFC Performance .......................................................................................... 22 2.2.1 Equilibrium Voltage ................................................................................. 22 2. 2 2 Activation O verpotential .......................................................................... 23 2. 2 3 Ohmic O verpotential ............................................................................... 25 2. 2 4 Concentration O verpotential ................................................................... 26 2.2.5 Other Losses ........................................................................................... 29 2.3 Efficiency and Degradation .............................................................................. 29 2. 4 Overview of C athode ........................................................................................ 31 2. 4 .1 Requirements .......................................................................................... 31 2. 4 .2 Oxygen Reduction Process .................................................................... 31 2. 4 .3 Mixed I onic Electronic C onductor ........................................................... 32 2. 4 4 La1 xSrxCoyFe1 yO3(LSCF) ................................................................... 34 2.5 Summary ........................................................................................................... 35 3 EXPERIMENTAL ........................................................................................................ 45 3.1 Electrical Cond uctivity Relaxation .................................................................... 45 3.1.1 Theoretical Background .......................................................................... 45 3.1.2 Electrical Conductivity Measurement ..................................................... 48 3.1.3 Effect of Surface Roughness .................................................................. 48 3.2 Electrochemical Impedance ............................................................................. 49 3.3 Temperature Programmed Isotopic Exchange ................................................ 50 3.4 Materials Characterizations .............................................................................. 51

PAGE 6

6 4 MECHANISTIC UNDERSTANDING OF LSCF DEGRADATION ............................. 57 4.1 Introduction ....................................................................................................... 57 4. 2 Experimental ..................................................................................................... 57 4.2.1 Sample preparation ................................................................................. 57 4.2.2 Heat treatment ......................................................................................... 58 4.2.3 Characterization ...................................................................................... 58 4.3 Results and Discussion .................................................................................... 60 4.4 Conclusions ...................................................................................................... 67 5 IMPACT OF CO IMPREGNATION INTO LSCF CATHODE ON SOFC PERFORMANCE ........................................................................................................ 74 5.1 Introduction ....................................................................................................... 74 5.2 Experimental ..................................................................................................... 75 5.2.1 AC I mpedance Analysis .......................................................................... 75 5.2.2 Electrical Conductivity Relaxation .......................................................... 76 5.2.3 Button Cell Fabrication and Tests .......................................................... 77 5.2.4 Catalyst Impregnation ............................................................................. 79 5.2.5 Characterization ...................................................................................... 79 5.3 Results and Discussion .................................................................................... 79 5.3.1 Effect of Co Impregnation on Electrode Impedance .............................. 79 5.3.2 Effect of Co Impregnation on Power Density of a Single Cell ............... 8 1 5.3.3 Characteristic D ependence of Kchem on O xygen P artial Pressure ..... 83 5.4 Conclusions ...................................................................................................... 85 6 MECHANISM OF CR VAP OR DEPOSITION ON SOFC CATHOES ....................... 94 6.1 Introduction ....................................................................................................... 94 6.2 Experimental ..................................................................................................... 95 6.2.1 Sample Preparation ................................................................................ 95 6.2.2 Heat Treatment ....................................................................................... 96 6.2.3 Characterization. ..................................................................................... 96 6.3. Results and Discussion .................................................................................... 97 6.3.1 XRD ......................................................................................................... 97 6.3.2 Cr Vapor Deposition on LSCF ................................................................ 97 6.3.3 Dependence of A -site ions on Cr Vapor Deposition ............................ 101 6. 3.4 Dependence of B -site i ons on Cr V apor D eposition ............................ 103 6.4. Conclusions .................................................................................................... 107 7 DEGRADATION MECHANISM AND THEIR IMPACT ON OXYGEN REDUCTION KINETICS OF LSCF .......................................................................... 117 7.1. Introduction ..................................................................................................... 117 7.2 Experimental. .................................................................................................. 117 7.2 1 AC I mpedance ................................................................ ....................... 117 7.2 2 Electrical C onductivity R elaxation ........................................................ 118

PAGE 7

7 7.2 3 Temperature Programmed I sotopic E xchange (TPX) ......................... 119 7.2 .4 Characteri zation .................................................................................... 120 7.3 Results and Discussion .................................................................................. 120 7.3.1 AC Impedance ....................................................................................... 120 7.3.2 Electrical Conductivity Relaxation ........................................................ 121 7.3.3 Temperature Programmed Isotopic Exchange .................................... 123 7.3.4 Post ECR Characterization ................................................................... 126 7.4 Conclusions .................................................................................................... 128 8 VAPORIZATION OF DOPED ALKALINE EARTH METAL OXIDE ........................ 137 8.1 Introduction ..................................................................................................... 137 8. 2 Experimental ................................................................................................... 138 8.3 Result s and D iscussion .................................................................................. 138 8. 4 Conclusions .................................................................................................... 140 9 CONCLUSIONS ........................................................................................................ 145 APPENDIX A THEORETIC EQUILIBRIUM VOLTAGE .................................................................. 147 B MEASUREMENT EFFECT OF NON -IDEAL STEP CHANGE IN ECR .................. 149 C MEASUREMENT OF THE SURFACE SENSITIVE RATE COEFFICIENT USING ISOTHERMAL ISOTOPIC SWITCHING ..................................................... 152 C.1 Introduction ..................................................................................................... 152 C.2 Background ..................................................................................................... 153 C.3 Experimental ................................................................................................... 154 C.4 Result s and D iscussion .................................................................................. 155 C.5 Conclusions .................................................................................................... 157 D TAPE CAST ING SLURRY COMPOSITION ............................................................ 161 E CR CONTAMINATION ON LSM82 .......................................................................... 162 F Kchem and Dchem at 750C ........................................................................................... 163 G SURFACE VARIATION OF LCCF AND LBCF ........................................................ 165 H COBALT OXIDE DISPERSION ON LSCF SURFACE ............................................ 166 L IST OF REFERENC ES ................................................................................................. 167 BIOGRAPHICAL SKETCH .............................................................................................. 175

PAGE 8

8 LIST OF TABLES Table page 3 -1 List of instruments for m aterial characterizations. ................................................. 56 5 -1 Thermodynamic factors and (3oC.52, 94 .................... 93 5 -2 Characteristic dependencies of the apparent rate coefficients on pO2.95 ............ 93 6 -1 Summary of tested ABO3 perovskite materials ................................................... 115 6 -2 Chemical composition of Crofer22APU ............................................................... 115 6 -3 Elemental a tomic percents (%) of LSCF after heat treatment at 800oC for 50 hrs in the presence of Crofer22APU obtained using EDS .................................. 115 6 -4 AES a tomic percents of LSCF after heat treatment(HT) at 800oC f or 50 hours 115 6 -5 Elemental atomic percents of non-heat treated and heat treated sample s* ...... 116 7 -1 Samples tested in TPX ......................................................................................... 136 7 -2 Elemental atomic percents (%) of LSCF obtained using EDS ........................... 136 7 -3 Elemental atomic percents (%) of LSCF after ECR test at 800oC for 400 hours obtained using AES ................................................................................... 136 A-1 Thermodynamic properties at 298.15K and 1bar ................................................ 148 D -1 Tape casting slurry composition. ......................................................................... 161

PAGE 9

9 LIST OF FIGURES Figure page 2 -1 Schematic diagram of the cross section of a planar type SOFC. ......................... 37 2 -2 Typical current density and voltage characteristic of SOFC. jL is the limiting current density.43 .................................................................................................... 37 2 -3 Contribution of chemical and electrical potential gradients to th e equilibrium state at the electrochemical system. ..................................................................... 38 2 -4 The electrochemical energy change from equilibrium state as a result of applying activation overpotential .......................................................................... 39 2 -5 act) as a function of current density and the effect of exchange current density (jo) on activation overpotential. .................................... 40 2 -6 Schematic diagram of the cross section for (a) electrolyte and (b) anode supported cell. ........................................................................................................ 40 2 -7 (a) Sketch of a diffusion layer between free stream and electrode and (b) c oncentration profile s across a diffusion layer at steady state.10 ......................... 41 2 -8 con) as a function of current density. .......... 41 2 -9 voltage (Ecell) at 800oC 20 ...................................................................................... 42 2 -10 Active reaction sites (painted in red) for oxygen reduction. (a) Pure electronic conductors and (b) mixed ionic and electronic conductors.23 ............................... 42 2 -1 1 Unit cell of pseudo cubic perovskite of ABO3 perovskite ...................................... 43 2 -1 2 Surface exchange coefficients of La0.7Ca0.3CrO3 as a function of mole fraction of oxygen vacancy .30................................................................................. 43 2 -1 3 Effect of B site transition metals on La0.7Sr0.3BO3 cathode over potential .44 ........ 44 2 -1 4 temperature.45 ........................................................................................................ 44 3 -1 Sample dimension and electrical connection between the sample and Lock in Amplifier. ................................................................................................................. 53 3 -2 Schematic of the effect of a rough surface on the boundary conditions. ............. 53 3 -3 The effect of roughness on the measurement of Dchem and Kchem (top), and surface roughness and actual surface area of samples (bottom). ....................... 54

PAGE 10

10 3 -4 Nyquist plot of an RC circuit. .................................................................................. 54 3 -5 Core shell ionization and subsequent deexcitation via (c) Auger process and (d) characteristic X -ray emission.58........................................................................ 55 3 -6 Schematic diagram of escape depth of Auger electr on and characteristic X ray.58 ....................................................................................................................... 55 4 -1 Brief description of sample preparation procedure. .............................................. 68 4 -2 Heat treatment set up. ........................................................................................... 68 4 -3 SEM micrographs of LSCF of (a) before heat treatment and after heat treatment at (b) 600oC, (c) 700oC, (d) 800oC and (e) 900oC for 50 hours in the absence of Fe-Cr alloy. .................................................................................... 69 4 -4 Magnified images of the precipitates on LSCF surfaces after heat treatment at (a) 900oC for 50 hours, (b) 900oC for 50hours,(c) 860oC for 100 hours, and (d) 890oC for 100 hours. ......................................................................................... 70 4 -5 Differentiated Auger electron spectra of the (a) LaMLL, (b) SrLMM, (c) CoLMM, (d) FeLMM, and (e) OKLL before(blue solid line) and after (red dotted line) heat treatment at 800oC for 50 hours. ........................................................................... 71 4 -6 TEM bright field image of the cross section of precipitate (left side) and EDS line scan of selected elements (right side) ............................................................ 71 4 -7 SEM micrographs of LSCF surfaces after heat treatment for 50 hours at (a) 800oC/ N2, (b) 800oC/0.1% O2 (c) 890oC/N2 and (d) 890oC/0.1% O2 ................... 72 4 -8 Schematic diagram of (a) the formation of oxygen vacancies and (b) formation of cation and anion vacancies simultaneously. .................................... 72 5 -1 Sketch of deposition of nano size particles on cathode surface via wet impregnation. .......................................................................................................... 86 5 -2 Schematic of AC impedance measurements. ....................................................... 86 5 -3 Equivalent circuit to fit the measured impedance data. ........................................ 87 5 -4 Current density and voltage measurement setup ................................................. 87 5 -5 AC Impedance spectra of LSCF/GDC/LSCF symmetric cell ............................... 88 5 -6 Impedance plots for symmetric cell of plain LSCF electrode and Co impregnated LSCF electrode at 800oC in air. ....................................................... 88 5 -7 Comparison of total electrode ASRs as a function of temperatures .................... 89

PAGE 11

11 5 -8 Measurement of cell voltage as a function current density at (a) 600oC, (b) 550oC and (c) 500oC. (d) Maximum power densities as a function of temperatures. ......................................................................................................... 90 5 -9 Impedance of a button cell for plain LSCF and Co impregnated cathode at 600oC ...................................................................................................................... 90 5 -10 The area specific resistance of electrode and electrolyte for plain and Co impregnated LSCF cathode cells. ......................................................................... 91 5 -11 (a) SEM micrograph of LSCF/GDC/Ni -GDC button cell and (b) magnified images of Co impregnated LSCF after cell test. ................................................... 91 5 -12 (a) Dchem and Kchem, (b) Dself and Ko, and (c) Dv at 800oC as a function of pO2. ... 92 6 -1 Schematic of experimental design for heat treatment. ....................................... 109 6 -2 XRD patterns of tested perovskite materials ....................................................... 109 6 -3 Surfaces of LSCF (a) before and (b-d) after heat treatment in the presence of Crofe r22APU at 800oC 50 hrs. The leading edge distances between LSCF and Crofer22APU sheets were (b) direct contact, (c) 0.1 cm, and (d) 0.5 cm. All images are the same scale. ............................................................................ 110 6 -4 XRD patterns of fresh and heat treated LSCF on Crofer22APU sheet at 800oC for 50hours. (b ~ d) E nlarged patterns for better comparison. is the XRD pattern measured using a low angle mode. ............................................... 111 6 -5 Surface s of LCCF, LBCF, LSCF and SCF after heat treatment in the presence of Crofer22APU at 800oC for 50 hours. The leading edge distances between samples and Crofer22APU sheet were 0.1 cm and 0.5 cm for the left and right hand sides, respectively ................................................................ 112 6 -6 Surfaces of tested after heat treatment in the presence of Crofer22APU at 600oC and 400oC (Leading edge distance 0.1 cm) ............................................ 113 6 -7 SEM micrographs of LSM surfaces of (a) before and (b) after heat treatment at 800oC for 50 hours in the presence of Crofer22APU (Leading edge distance 0.1 cm) .................................................................................................. 113 6 -8 (a) SEM micrograph of LSM af ter heat treatment in the presence of Crofer22APU at 1050oC for 50 hours, and EDS spectra for (b) the flat region and (c) particle (Leading edge distance 0.1 cm) ................................................ 114 7 -1 Equivalent circuit to fit the measured impedance data ....................................... 130

PAGE 12

12 7 -2 (a) A schematic of the experimental set up used for temperature programmed isotopic exchange and (b) the preheat treatment and measurement conditions ...................................................................................... 130 7 -3 AC Impedance spectra of LSCF/GDC/LDCF in air at 800oC .............................. 131 7 -4 Area specific resistance of LSCF at 800oC in air as a func tion of aging time (a) in the (a) presence and (b) absence of K type thermocouple in the reactor. .................................................................................................................. 131 7 -5 (a) Normalized electrical conductivity relaxation curve after switching oxygen partial pressure from 43 to 67% at t =0 and (b) variation of relaxation curves with aging at 800oC. ............................................................................................. 132 7 -6 Degradation of Kchem and Dchem in Cr contaminated environment at 800oC 132 7 -7 Effect of total surface area on oxygen exchange profiles which are measured by TPX. SA refers to the surface area. ................................................................ 133 7 -8 T PX result of hea t treated LSCF at 800oC for 400 hours with and without exposure to Crofer22APU in comparison to fresh LSCF. ................................... 134 7 -9 SEM micrographs of polished LSCF surfaces (a) before and (b) after ECR at 80 0oC for 400 hours. ............................................................................................ 135 8 -1 Schematic of experimental set up. ...................................................................... 142 8 -2 SEM micrographs of (a) as polished and (b) heat treated LSC F in the presence of Crofer22APU. ................................................................................... 142 8 -3 (a) LSM surface after heat treatment at 1050oC for 50 hours in the presence of Crofer22APU sheet. EDS spectra for (b) the newly formed particle after heat treatment and (c) as polished LSM sample. ............................................... 143 8 -4 (a) Surface of LSM after heat treatment at 800oC for 50 hours in the presence of BaO powder and Crofer22APU. (b) Spot mode EDS spectra fo r the particle. ........................................................................................................... 144 B-1 Reactor flush time measurement set up using mass spectrometer. The flow from ECR to mass spec was limited to 30 sccm by the ball flow meter. ............ 150 B-2 Measured reactor flush time using mass spectrometer (a) and a fit using Eq B-3 (b). .................................................................................................................. 151 C -1 Time dependent relaxation profile of normal oxygen (16O ), oxygen isotope (18O) and scrambled oxygen (16O+18O) at 700oC ................................................ 158

PAGE 13

13 C -2 Normalized concentrations of desorbed oxygen from the sample and incorporated oxygen ............................................................................................. 158 C -3 Kchem measured by ECR in this study and literatures 52, 94, converted using the thermodynamic factor, at 800oC. ................................................................... 159 C -4 N ormalized relaxation curves of inc orporated oxygen isotope at different temperatures. ....................................................................................................... 159 C -5 Arrhenius plot of surface exchange coefficient ( pO2= 0.01) ............................... 160 E-1 ( a) As polished LSM82 surface and (b) heat treated LSM82 at 800oC for 50hrs in the presence of Crofer22APU. .............................................................. 162 F -1 Dchem and Kchem at 750oC as a function of pO2. ................................................... 163 F -2 Degradation of Kchem at 750oC. ............................................................................ 164 G -1 Differentiated A uger electron spectra for pre (blue) and post (red) heat treated samples at 800oC for 50 hrs. ( a) LCCF and (b) LBCF ........................... 165

PAGE 14

14 LIST OF ABBREVIATION S AES Auger electron spectroscopy AFM Atomic force microscopy ASR Area specific resistance BET Brunauer, Emmett, and Teller surface area measurement BOP Balance of Plan t DPB Di -n butyl phthalate ECR Electrical conductivity relaxation EDS Energy dispersive spectroscopy FIB Focused Ion beam GDC Gd2O3 doped CeO2 LBCF La0.6Ba0.4Co0.2Fe0.8O3 LCCF La0.6Ca0.4Co0.2Fe0.8O3 LSCF La0.6Sr0.4Co0.2Fe0.8O3 LSM La0.6Sr0.4MnO3 MIEC Mixed ionic and electronic conductor PEG Polyethylene glycol PVB Polyvinyl butyral SCF SrCo0.2Fe0.8O3 SECA Annual s olid s tate e nergy c onversion a lliance SEM Scanning electron microscopy SOFC Solid oxide fue l cell TEM Transmission electr on microscopy TPB Three phase boundary TPX Temperature programmed isotopic exchange

PAGE 15

15 XRD X -ray diffraction YSZ Y2O3 stabilized ZrO2 pO2 Oxygen partial pressure gradient act Activation overpotential con Concentration overpotential ohm Ohmic overpotential

PAGE 16

16 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfi llment of the Requirements for the Degree of Doctor of Philosophy FUNDAMENTAL UNDERSTANDING OF SURFACE CHARACTERSITIC AND CHROMIUM CONTAMINATION ON SOLID OXIDE FUEL CELL CATHODES By D ONGJO OH May 2010 Chair: Eric. D Wachsman Major: Materials Science an d Engineering A solid oxide fuel cell (SOFC) is a highly efficient and environmental -friendly energy conversion device But the progress toward commercial ization continues to be a slow struggle mostly due to high operation temperatures (800~1000oC) Lowering operation temperatures can bring manufacturing costs down and provide high conversion efficiency, and less difficulty in sealing and heat management. However, substantial increase in internal losses, especially activation overpotential, lowers SOFC pe rformance with reduced operation temperatures. Because t he activation overpotential is mostly attributed to the oxygen reduction reaction (ORR) on the cathode t remendous works have been done in order to improve the cathode performance and understand the O RR mechanism s and degradation mechanism s A heterogeneous reaction is primarily affected by the interaction between surface and adsorbed sp ecies Therefore, the surface composition and structure are one of the most important factors affecting the cathode performance. Nevertheless, compositional and structural variations of cathode surfaces at high temperatures have not gained much attention. The objective of this study is firstly, to better understand the surface

PAGE 17

17 structure of cathode materials at high tem peratures and, secondly, to improve the SOFC cathode performance based on this knowledge. La0.6Sr0.4Co0.2Fe0.8O3 has bee n widely used for SOFC cathodes. Chemical and structural variations of La0.6Sr0.4Co0.2Fe0.8O3 (LSCF) surfaces under oxidizing environm ent were investigated. Cr contamination free LSCF showed the formation of submicron-sized SrOx precipitates on the grain surface. This caused the reduced concentration of transition metals in B sites The addition of cobalt oxide on the surface of LSCF was found to improve the cathode performance. In contrast, Cr vapor deposition caused the formation of larger SrCrO4 particles on the LSCF surface along grain boundaries, and Srdeficient matrix The structural analysis identifie d t he phase transition from r hombohedral to cubic perovskite in due to Sr deficiency A defect chemistry model was presented based on observed phenomena Electrical conductivity relaxation, AC impedance spectroscopy and temperature programmed isotopic exchange (TPX ) were conducted to study the impact s of C r contamination on the cathode performance. It was found that Cr contamination was a chemical process for Co -Fe based materials and an electrochemical process for Mn based materials Unstable Co4+ and Fe4+ lead to the formation of a nucleation agent for the chemical reaction. By contrast, stable Mn4+ does not lead to the formation of a nucleation agent At last, a new hypothesis for the vaporization of Sr from La0.6Sr0.4Co0.2Fe0.8O3 (LSCF) was proposed based on previous observations and the evidence for Sr vaporization was provided.

PAGE 18

18 CHAPTER 1 INTRODUCTION 1.1 Dual C hallenges of E nergy and C limate C hange The development of clean and sustainable energy sources has recently gained speed due to increased worldwide attention on decreasing oil supplies and climate change .1 3 Due to the potential worldwide impact from the se problem s global efforts are required. For example, there was an histori c Summit on Climate Change in Copenhagen in December 2009. 115 world lead ers participated in the summit and recognize d the catastrophic impacts of climate change. The summit, however, ended in failure without a strong agreement to reduc e green house gas emission because of the conflict of national interests. Decades of activities of wealthy countries are most likely responsible for current climate change while the impacts from the regulations of green house gas emission are more profound for poor countries G reen house gas emissions are closely tied to national economic growth. Poor coun tries do not want to kick away the ladder to reach wealthy countries by reducing emissions Political efforts are the best option to slow down the momentum for global warming in the short term. Long term solutions will require the development of clean and sustainable energy technology 1.2 Solid Oxide Fuel Cell (SOFC) A fuel cell is an energy conversion device that converts chemical fuel s directly into electrical energy t hus providing high efficiency and reduced emission of pollutants, including green house gases .4 8 In addition, the absence of moving parts allows quiet operat ion and long life time s The fuel cell is different from the rechargeable battery since its size can range from small to large and it can be recharged quickly by refueling. There are vario us f uel cell types such as polymers, aqueous alkalis, phosphoric acids,

PAGE 19

19 molten carbonates, and solid oxides T he solid oxide fuel cell ( SOFC ), which uses a solid oxide electrolyte, are marked by fewer problems in electrolyte management extended life time highest efficiencies, low materials costs, fuel flexibility, and a high grade of waste heat .9, 10 The SOFC ha s a high p otential for application in stationary power plant s and auxiliary power unit s of vehicles. There have been several successful demonstrations .6, 11 However, the progress toward a commercial ization of the SOFC is a slow struggle One of the biggest hurdles is high manufacturing costs. The high cost of SOFC stack is in part due to high operating temperatures. SOFCs operate at higher temperatures than the other types due to the solid oxide electrolyte which leads to high material costs. The cost breakdown model shows that the SOFC stack is the one of the major components while the remaining balance of plant (BoP) equipment such as fuel and air supply and power electronic control system holds a minor share of the manuf acturing cost .12 Below 800oC LaCrO3 based ceramic interconnects which contributes over 80% of material cost for the stack, can be replaced with cheaper metallic interconnect s.13 Furthermore, reduced operation temperatures improves thermodynamic conversion efficiency for reformed gas and improves the sealing and heat management situation .9 However, a rapid decrease in SOFC performance must be overcome in order to reduce operation temperatures. 1.3 Objectives A heterogeneous reaction is primarily determined by interaction between surface and gas phases As a result, surface chemical composition is a n important factor affecting the oxygen reduction reaction on cathodes T he chemical composition of metal oxide surfaces may significant ly differ from that of a bulk due to broken bonding at the

PAGE 20

20 surface. The chemical variation of the surface of La0.6Sr0.4Co0.2Fe0.8O3 (LSCF) which ha s been widely used for intermedi ate temperature (600800oC) range SOFC s was investigated. It was found that the catalytically inactive element s tended to segregate to the surface or near -surface region under ox idizing condition. Based on these results, additional cataly st material was dispersed on the surface by impregnation and the effect was examined. In the meanwhile, utilization of metallic interconnects is ad vantageous for cost reduction. However, most interconnects contain high concentration of Cr (10~30%) and out diffused Cr is known to be evaporated into CrO3 or CrO2(OH)2 at high temperatures under oxidizing conditions .14 As these Cr vapor species are transported into the cathode or electrolyte, th ey may affect the catalytic activity of cathodes. This study investigated the effect of Cr contamination on the performance of La0.6Sr0.4Co0.2Fe0.8O3 cathode, and the reaction mechanism between cathodes and Cr vapor species in order to provide fundamental knowledge on the long term stability of SOFC.

PAGE 21

21 CHAPTER 2 BACKGROUND 2.1 Basic C omponents and O peration P rinciple of SOFC The basic components of a single cell are a cathode, anode, and electrolyte as described in Figure 2 1. The solid state elect rolyte is a pure ionic conductor, which spatially separates air from fuel gases and prevents direct combustion between them. The cathode and anode refer to the electrodes where oxygen is reduced into oxygen ion s and hydrogen is oxidized in to water vapor in combination with oxygen ions respectively. While the electrolyte must be dense, the two electrodes are porous in order to increase reaction areas and improve mass transport of reactants. The overall reaction for the SOFC is spontaneous. Once oxygen and fuel gases are delivered to the anode and cathode, respectively, there is a net chemical driving force for oxygen transport due to an pO2) across the electrolyte. Provided that hydrogen is used as a fuel, the two electrochemical half reactions taking place at the cathode and anode are (Cathode) 2 22O 4e (g) O (2 -1) (Anode) 2 24e O(g) 2H (g) H 2O 22 (2 2) Or by using the Krger Vink notation,15 (Cathode) o / O 22O 4e 2V (g) O (2 3) (Anode ) o / 2 2 o2V 4e O(g) 2H (g) H 2O 2 (2 -4) w here Vo represents the oxygen vacanc ies in the lattice and Oo represents the oxygen anions in regular oxygen anion site s The overall reaction is (Overall) O(g) 2H (g) 2H (g) O2 2 2 (2 5)

PAGE 22

22 The internal flow of oxygen ions through the electrolyte must be balanced by the flow of electronic charge through an external circui t providing useful work. 2.2 SOFC P erformance 2.2.1 Equilibrium Voltage Theoretical t hermodynamics limits the maximum voltage that a single cell can overall reaction in Eq. 2 5 is (g) O (g) H O(g) H Reactant Product Rxn2 2 2G 2G 2G G G (2 -6) where G is the chemical e nergy of corresponding species. Under a non-stan dard state, G can be written as p ln T R G Go (2 7) Go is the chemical energy at standard state, R is the gas constant, T is the absolute temperatur e and p is the pressure. The chemical energy change is associated with electrical energy by F n E (2 -8) where E is the voltage, n is the number of electron transferred during the reaction, and F is Faradays constant. Thus the equilibrium voltage (EEq) can be written as16 2 H O 2 O H 0 Eq2 2 2 Tp p p ln nF RT E E (2 -9) where ET 0 is the voltage at standard state and temperature T, and T is the absolute temperature. Eq. 29 is known as the Nernst equation. EE q has been also called Nernstian voltage or open circuit voltage (OCV). At 800oC, the equilibrium voltage of a singe cell is approximately 1.13 V (Appendix A).

PAGE 23

23 At equilibrium the SOFC produces the maximum output voltage but no current. The cell voltage starts to decrease from the equilibrium voltage as the current density increases due to losses from internal cell resistances. Figure 2-2 summarizes the theoretic behavior of the current density and voltage. The key performance of an SOFC is given by its pow er density which is how much electrical energy an SOFC can produce per unit time and unit area. There are four major internal losses: the activation actohmcon), and the loss (EL) due to the partial electronic conductivity of electrolytes or imperfect gas sealing. T he output voltage measured (V) can be written as V= EEq actohmcon-EL (2 -1 0 ) In the following section, the nature of internal loss es is explained. 2. 2 2 Activation O verpotential Activation overpotential is attributed to the limit of electrode reaction kinetic s Even though the reaction spontaneous R xn < 0), the rate is finite because of the activation barrier (EA). O nly species that have gained suff iciently high energies to overcome the activation threshold are able to jump to the next state as Figure 2 -3 a describes.17 According t o statistical thermodynamics, the probability of finding a species (P) in the activated state at temperature T is given by RT EAe P (2 -1 1 ) where EA is the activation energy, R is the gas constant and T is the absolute temperature. T he forward (JF) and backward reaction (JB) rate can be written by RT E F R FAe f C J (2 -1 2 )

PAGE 24

24 RT E B P BRxn Ae f C J (2 -13 ) where CR and CP are the concentration of reactants and products at the surface respectively. fF and fB are the frequency of activated species moving forward and R xn is the chemical free energy ch ange. The net reaction rate (J) is RT E B P RT E F R B FRxn A Ae f C e f C J J J (2 14) Since the reaction at electrode involves charge transfer, the rate is connected to current density (j) by Faradays law (Eq. 215) RT E B P RT E F RRxn A Ae f C F n e f C F n J F n j (2 -15 ) where j is the current density, n is the number of electron transferred during the reaction, and F is Faradays constant. In an electrochemical system, the accumulation or depletion of charged species creates the electrical pot and exerts additional force on charged species as seen in Figure 23b by F n G ~ (2 -16 ) where G ~ is the electrochemical potential. Eq. 215 then becomes RT G nF E B P RT E F RRxn A Ae f C F n e f C F n J F n j (2 17) When R xn+ nF the system reaches the equilibrium as Figure 2 3c describes and n o current flows as the rates of forward and backward reactions are the same. The current density at equilibrium state is called the exchange current density (jo). That is, RT E F R b f oAe f C F n j j j (2 -18 )

PAGE 25

25 The rate of forward and backward reaction can be manipulated by adjusting the electrical potential as described in Figure 2 -4. The applied potential for the reduction of the forward reaction act). T he current density as a result of applying activation overpotential can be written as RT F n (1 RT F n oact acte e j j (2 -19 ) the va lue depends on the symmetry of the activation barrier. Eq. 2-19 is known as the Butler Volmer equation, which defines the relation between the activation overpotential and current density. Figure 2 5 plots the activation overpotential as a function of curr ent density according to Eq. 2-19 In the f igure an i ncrease in exchange current reduce s the activation overpotential. Since t he exchange current density is a function of the reactant and activation barrier s in Eq. 218, improvement of the exchange current density can be achieved by using appropriate catalysts and increasing reactant concentrations. 2. 2 3 Ohmic O verpotential The origin of ohmic overpotential is the ohmic resistances of cell components against the transport of charged species. The overall ohmic overpotential can be expressed by ohmic = ( ASRelectrodes + ASRelectrolyte + ASRcontact ) j (2 2 0 ) where j is the current density and ASR are the area specific resistances, which refers to the resistance normalized by area. Generally the electrode is a good electronic conductor and so ohmic overpotential mostly arises from electrolyte. There is a relationship between ASR, sample dimension and resistivity (Eq. 22 1 ). ASR = RA = L = (1/ ) L (2 2 1 )

PAGE 26

26 the conductivity. T wo approaches have been widely employed to reduce ohmic losses : (1) reducing electrolyte thickness via thin film technique and anode supported cell design (2) reducing resistivity by using faster oxygen ionic conductor. In an anode supported cell, the anode provides the mechanical support so the electrolyte thickness can be much thinner compared to an electrolyte supported cell as described in Figure 2 6 In addition, some materials such as doped CeO2 and doped Bi2O3 have higher ionic conductivity compared to stabilized ZrO2.18 Applicati on of those materials for the electrolyte has been demonstrated for high performance SOFC in the intermediate temperature range. Recently contact resistance has received growing attention. In the level of small size cell test, the magnitude of contact res istance is insignificant. H owever, with larger area stacked cell, the use of metal interconnects can lead to the formation of less conductive secondary phases at the interface between electrode and interconnect .19 Moreover, unbalanced compressive pressure over a large area due to warpage may lead to poor contact and a substantial increase in contact resistance. Proper system design and development of contact materials to provide uniform and reliable electrical contact is important to k eep contact resistance mitigated in the large scale cell. 2. 2 4 Concentration O verpotential The theoretic al calculation of activation overpotential and equilibrium voltage assumes that the reactant concentration remains constant. However, as Figure 2 7 a d ecribes, the concentration in the region in the vicinity of electrode surface varies due to electrode reaction, which consumes reactant s and produce s products. The loss due to reduced concentration of reactants is the origin of concentration overpotential. M ass

PAGE 27

27 transport through the diffusion layer is governed by a concentration gradient and diffusivity. At a steady state condition as described in Figure 2 -7b the diffusion flux (Jdiff) can be described using Ficks 1st law (Eq. 2-22 ). C C D D JR R Diff (2 2 2 ) where D is the apparent diffusivity of reactants in the diffusion layer, CR and CR are the concentrations of reactant s in the free stream and at the electrode surface respectively, the diffusion layer.10 T he diffusion flux reaches maximum value when CR approaches zero. Since the diffusion flux is associated with current density by Faraday s law, the maximum or limiting current density (jL) is C D F n jR L (2 -23 ) jL refers to the theoretical limit of current density that SOFC can produce. R educed concentration of reactants at the electrode surface leads to an increase in activation overpotential and equilibrium voltage. The magnitude of the increase i n activation overpotential is ) (C ) (C (act) R act R act con (2 -24 ) Input t ing the CR into Eq. 2-19 gives RT F n (1 P P o RT F n R R oact acte C C j e C C j j (2 25) At high current density where d epletion of reactants is significant when the forward reaction on the electrod e surface is substantially fast 1 RT F n act

PAGE 28

28 S o Eq. 2-33 can be simplified into R o R actC j C j ln Fn T R (2 -26 ) Input ting Eq. 226 i nto 2 24 yields R R conC C ln F n T R (act) (2 -27 ) This can be expressed in terms of jL based on the relation ship between CR and jL in Eq. 2 -22 and 223. j j j ln F n T R (act) L L con (2 -28 ) Meanwhile, equilibrium voltage also drops due to reduced concentration of reactant. In the same manner above, this provides j j j ln F n T R ) (C E ) (C E ) (E L L R eq R eq eq con (2 -29 ) The summation of Eq. 2-28 and 2-29 provides t he overall concentration overpotential, which can be written as j j j ln 1 1 F n T R L L con (2 -3 0 ) If the concentration overpotential is plotted as a function of current density according to Eq. 2-30, it shows r apid increase as the current density approaches the limiting current density as seen in Figure 2 8 In order to reduce concentration overpotential, the limiting current density needs to be increased. Eq. 2-23 provides that the limiting current can be increased by efficient delivery of fuels and oxygen from appropriate design of the interconnect and porous electrode structure

PAGE 29

29 2.2.5 Other L oss es The equilibrium voltage given in the Nernst equation (Eq. 29) assumes a pure ionic conductor, in which electronic conduction is zero. If there is a partial electronic conduction across the electrolyte, the equilibrium voltage is lower than the ideal case. The ratio of ionic and electronic conduction can be expressed as a transference number (t). The electronic (te) and ionic (ti) transference numbers are defined as i e e e t (2 31) i e i i t (2 32) where e and i are the electronic and ionic conductivity, respectively The summation of te and ti must be the unity. Using the transference number, the loss in equilibrium voltage (EL) can be written as e Eq Lt E E (2 33) Thus, loss in the equilibrium voltage is increased with electronic conduction. For ex ample, doped CeO2 is a fast oxygen ion conductor. However, Ce has a multi valent state so that there is a partial electronic conduction at reduced condition. Therefore, it is difficult for a cell using a thin doped CeO2 electrolyte to obtain an ideal equil ibrium v oltage. In order to overcome this problem there have been attempts to make bi layers, which consist of a fast ionic conductor and a pure ionic conductor (ZrO2 or Bi2O3).20, 21 2.3 Efficiency and Degradat ion T t hermodynamic) of the SOFC can be defined as20, 21 1 energy Total energy Useful mic thermodyna (2 -34 )

PAGE 30

30 Using thermodynamic data in Table A-1, the efficiency of electrochemical conversion is approximately 0.8 at 800oC. However, the efficiency in practical SOFC operation is lower than that because of fuel utilization. When the cell voltage approaches EEq, fuel consumption is reduced and unused fuels in the exhaust reduce the fuel -cell efficiency. The f uel utilization efficiency ( fuel) can be written as fuel fuelJ nF j/ (2 -35) where j is the current density Jfuel is the rate at which fuel is supplied to the cell The overpot ential is another factor affecting the efficiency. In the previous section, t he cell voltage was shown to decrease due to the overpotential. As a result the efficiency of the SOFC is further decreased T voltage) is defined as Eq voltageE V (2 36 ) Generally t he total cell efficiency is represented by the product of the thermodynamic, fuel utilization and voltage efficiency (E q. 237). thermodynamic voltage fuel (2 -37 ) This indicates that the conversion efficiency depends on the operating conditions such as curre nt density or voltage. Figure 9 shows the efficiency of the cell as a function of cell voltage. As the output voltage is close to the equilibrium voltage. T he ideal efficiency ( thermodynamic voltage ) increases with increasing output voltage but the fuel utilization efficiency drops as the voltage approaches to the equilibrium voltage. Therefore, SOFC operation conditions need to be carefully decided to get high efficiency.

PAGE 31

31 The long term stability of the system is of the utmost importance for practical application Solid State Energy Conversion Alliance, which is initiated by Department of Energy has set 2.0 % per 1000 hours as a degradation rate of SOFC system by 2010 and planned to further reduce the degradation rate. T he power density of the cell is c urrent density multiplied by voltage. Therefore, the degradation rate of the cell is proportional to increase in internal losses in constant current density mode. Since t he activation overpotential is the dominant factor affecting the overall internal loss reducing the degradation of cathode s is necessary to minimize the degradation of SOFC. 2. 4 Overview of C athode 2. 4 .1 Requirements The requirements for a SOFC cathode are22 H i gh catalytic activity for oxygen reduction. High electronic and ionic conductivity under oxidizing condition Chemical compatibility with neighboring cell components Matched therm al expansion coefficient Good stability at high temperature Good adhesion to electrolyte surfaces Simple fabrication Adequate porosity Low cost 2. 4 .2 Oxygen R eduction P rocess I n order for oxygen reduction to occur reactants, electron (e-), oxygen vacanc ies (Vo ) and oxygen gas molecule (O2), must be supplied simultaneously according to Eq. 2 -3 Thus, reduction of oxygen takes place at the three phase boundary (TPB) of air, cathode, and electrolyte phase as Figure 2 10 a shows. The overall reaction o f oxygen reduction may involve a number of elementary steps such as

PAGE 32

32 Gas phase diffusion Adsorption of molecular oxygen (O2(g)+Sad2 (ad) ) Dissociation into atomic oxygen (O2(ad) ) Charge transfer (O(ad)+2e/// (ad) ) Incorporation of oxygen ions (O// + VO o x ) Determining the rate limiting step for the oxygen reduction reaction may provide guidelines for improving catalytic activity. A number of papers have been devoted to the mechanistic study of oxygen reduction.2325 However, there still remains uncertainty and disagreement about the sequence of steps as well as the rate determining step. This is in part due to differen ces in cathode compositions and microstructure, complex reaction pro cess, heterogeneity of surface and experimental conditions between different research groups .25, 26 2. 4 .3 Mixed I onic Electronic C onductor For a pure electronic conductor such as Pt or La1 xS rxMnO3 (LSM), the number of active reaction sites is proportional to three phase boundary (TPB) length. One method to red uce cathode overpotential is to extend the TPB length by mixing the electronically conductive cathode and ionically conductive electrolyte, which is called a composite cathode. Another method is to replace the electronic conductor with a mixed ionic and el ectronic conductor (MIEC) material which conducts oxygen ions as well as electrons. Due to the large amount of oxygen vacanc ies of the electrodes, oxygen gas can be reduced into ions over the electrode surface, giving rise to extensive active reaction site s (Figure 210 b) A c omposite cathode between the MIEC and electrolyte was found to further improve oxygen reduction kinetics .27 ABO3 perovskite -structured oxides are typical MIECs, wherein A site and B site cations have 12 and 6 coordination to oxygen ions respectively. The ideal structure of

PAGE 33

33 cubic perovskite is illustrated in Figure 2 1 1 As long as a tolerance fact or ( Eq. 23 4 ) lie s in the proper range ( 0.7 < t <1.0), the perovskite structure can be maintained. ) r (r 2 ) r (r tO B O A (2 3 4 ) where t is the tolerance fa ctor, rA, rB, and rO are the ionic radii of A,B and oxygen respectively. When the tolerance factor is between approximately 0.89 and 1, the perovskite is likely to adopt cubic structure. Out of that range, the structure can be slightly distorted, having lo wer symmetry such as rombohedral and orthorhombic .8 Many studies have attempted to elucidate the relationship between the catalytic property and constituents of materials .30 36 It was shown that oxygen vacanc ies and B site transition metals were the critical factor s affecting catalytic activity of ABO3 perovskite materials .28, 29 Figure 2 1 2 shows that the surface exchange c oeffi cient (k), which refers to apparent rate of oxygen reduction reaction, increases with oxygen vacancy concentration.30 In the figure, open and closed symbols refer to reduction and oxidation runs, respectively T he surface exchange coefficient (k) and exchange current density (jo) are related with each other by31 [O] k F n jo (2 3 5 ) where [O] is the oxygen concentration in the lattice. Thus, increased k indicates reduction in activation overpotential. For SOFC cathodes, a mixture of rare and alkaline earths and transitional metals have been widely tested for A and B sites ions, respectively. Recently, Co based perovskite materials such as La1 xSrxCoO3 and Ba1 xSrxCoO3 have been shown to display better performance than other cathod e materials as shown in Figure 2 1 3 It was found that the electronic state and binding energy of transition metal s were deeply

PAGE 34

34 correlated to catalytic activities .24, 32, 33 There is n o chemical reaction between Co based perovskite cathodes and Gd doped CeO2 (GDC) electrolyte s, b ut a higher thermal expansion coefficient (TEC) than GDC electrolyte is a critical problem .343 6 The TEC mismatch can lead to a substantial increase in interfaci al resistance due to de lamination. In order to match the TEC between the cathode and electrolyte while maintaining the catalytic activity, a mixture of Co and Fe in B site has been tested. 2. 4 4 La1 xSrxCoyFe1 yO3(LSCF) La1 xSrxCoyFe1 yO3(LSCF) is a typical MIEC. It provides high oxygen reduction capability and its TEC match es that of the GDC electrolyte. The structure of LSCF is not cubic but orthorhombic or rhombohedral perovskite depending on composition. L a3+ is partially substituted by Sr2+ in order to create either oxygen vacancies or holes The reaction can be described by (g) O 2 1 V 5O 2B 2Sr O B SrO 22 O o B / La 3 2 3 2O La (2 -36 ) o B / La 2 3 26O 2B 2Sr (g) O 2 1 O B SrO 23 2O La (2 37 ) where B is th e transition metal on the B site. Energetically the relative magnitude of the oxygen vacancy formation energy and valence stability of the transition metals determine s the proportion of ionic and electronic compensation. At room temperatures, electronic compensation through Eq. 2-37 is pronounced. The concentration of oxygen vacancies further increases at higher temperatures This is due to the reduction of transition metals according to O B 2 B oV 2B (g) O 2 1 2B O (2 38)

PAGE 35

35 The stability of the transition metals and bonding strength between metal oxide ions determines the onset temperatures at which oxygen nonstoichiometry rapidly increases .37 The effect from the amount of dopant on the A and B site s on the concentr ation of oxygen vacanc ies is shown in Figure 2 1 4 It shows that oxygen vacancy concentration in LSCF increase s with amount of the Sr dopant An increase with temperatures represents the reduction of transition metals. While LSCF offers improved power densities than pure electronic conductors, it does not exhibit stable performance over long term period of time. Simner el al.38 found the d egradation of cell performance during 500 hours at 750oC as a result of substantial increase in electrode resistance. Similar degradation behavior of LSCF cathode was also reported by Kim et al.39 In a practical application, SOFCs need to be operated in stack in which metal interconnects connects several single cells in series to produce higher voltage. Since such metal interconnects contain high concentration of Cr, the interaction between LSCF cathode and Cr vapor from metal interconnects can lead to more significant degradation in cathode performance.40 2. 5 Summary This chapter covers basic background and operation principle of SOFC s The key performance of SOFC is explained by voltage and cur rent density characteristic, follo wed by various overpotentials. During SOFC operation, the activation overpotential which is primarily determined by the oxygen reduction reaction on cathodes, is the dominant loss Therefore, improvement of catalytic acti vity and stability of cathode materials can contribute to enhancement in cell performance. As s mixed ionic electronic conductor (MIEC) LSCF cathodes ha ve been widely used for a solid oxide fuel cell cathode due to the high ionic and electronic conducing properties22, 41 They

PAGE 36

36 have also exhibited enhanced catalytic activity for oxygen reduction compared to the pure electronic conductor.42 However, an attempt to reduce operating temperatures causes the reduced rate of oxygen reduction reaction (ORR), which results in a substantial increase in activation overpotential. In addit ion LSCF cathodes exhibited substantial degradation during the long term test. Materials performance and properties are deeply related to materials structures. Especially, a heterogeneous reaction is primarily affected by surface structure and composition. Therefore, i nvestigat ion of the chemical and structural variations of LSCF at high temperatures can contribute to better understand degradation mechanism s and improve cathode performance.

PAGE 37

37 Figure 2 1. Schematic diagram of the cross section of a planar type SOFC. Figure 22 Typical current density and voltage characteristic of SOFC. jL is the limiting current density.43

PAGE 38

38 Figure 2 3. Contribution of chem ical and electrical potential gradient s to the equilibrium state at the electrochemical system.

PAGE 39

39 Figure 2 4. The electrochemical energy change from equilibrium state as a result of applying activation overpotential

PAGE 40

40 Figure 2 5. Activation oact) as a function of current density and the effect of exchange current density (jo) on activation overpotential. Figure 2 6. Schematic diagram of the cross section for (a) electrolyte and (b) anode supported cell.

PAGE 41

41 Figure 2 7. (a) Sketch of a diffusion layer between free stream and electrode and (b) c oncentration profiles across a diffusion layer at steady state.10 Figure 2 8 con) as a function of current density.

PAGE 42

42 Figure 2 9 voltage (Ecell) at 800oC 20 Figure 2 10. Active reaction sites (painted in red) for oxygen reduction. (a) P ure electronic conductors and ( b ) mixed ionic and electronic conductors.23

PAGE 43

43 Figure 2 1 1 Unit cell of pseudo cubic perovskite of ABO3 perovskite Figure 2 1 2 Surface exchange coefficients of La0.7Ca0.3CrO3 as a function of mole fraction of oxygen vacancy .30

PAGE 44

44 Figure 2 1 3 Effect of B site transition metals on La0.7Sr0.3BO3 cathode over potential .44 Figure 2 1 4 temperature.45

PAGE 45

45 CHAPTER 3 EXPERIMENTAL 3 1 Electrical C onductivity R elaxation 3.1. 1 Theoretical Background Electrical conductivity relaxation measures the transient electrical conductivity change of the sample in response to a step change of oxygen partial pressur e The measured data is fit to obtain the chemical diffusion coefficient of oxygen in the lattice and the chemical exchange coefficient on the surface.46 49 The kinetic model applied for ECR is the solution to non-s teady state diffusion, which is known as Ficks 2nd law. When the thickness (2 x ) of the sample is much smaller than both the width (2y) and the length (2z) the spatial and time dependence of the oxygen concentration can be described by 2 2 chemx C D t C l x l (3 1) where C is the oxygen concentration, t is the time, and Dchem is the chemical diffusion coefficient. The initial and boundary condition s are given in Eq s 3 2 and 3 -3 which mean that the sample is initially at an equilibrium and the derivative of oxygen concentration with respect to position at the center of the sample (x = 0) is zero, respectively. oC C(x) when t = 0 (3 -2) Co is the initial equilibrium concentration. 0 x C 0 x (3 -3) The oxygen exchange process at the surface is assumed to proceed at a rate proportional to the difference between the concentration at the new equilibrium and the

PAGE 46

46 transient surface concentration. This provides boundary condition at the surface (Eq. 34). ) C (C K x C Dt chem l x chem (3 4) Kchem is the chemical exchange coefficient that refers to the apparent surface reaction coefficient. On the basis of initial and boundary con ditions, the mathematical solution to Eq. 3 -1 is given by50 m 2 2 m 2 chem 2 m m 1 m 0 0 t cos ) L L ( ) l t D ( exp ) l x ( cos L 2 1 C C C C (3 5) l is half the thickness of the sample. Subscripts 0, t and and final equilibrium time s respectively. m is the positive root of the transcendental equation L D K l tan chem chem m m The integration of Eq. 35 in the range of l x l provides the total amount of diffusing oxygen species over the entire sample. L) L ( ) l t D ( exp L 2 1 M M g(t)2 m 2 2 m 2 chem 2 m 2 1 m t (3 -6) tM and M are the total amount of diffused component s up to time t and at the new equilibrium, respectively. For (A3+ 1 xA2+ x)BO3 (B=Co, Fe, Ni), the relative mass change can be correlated to the electrical conductivity change via defect chemistry. The overall electroneutrality condition is ] [V p ] [V ] [V ] [D nO /// B /// A /3 (3 -7)

PAGE 47

47 where n and p are the electron and hole concentrations respectively. [D/], [VA3+ ///], [VB ///], and [Vo ] are the concentrations of A site dopant, A site vacancy, B site vacancy and oxyg en vacancy, respectively. In the high pO2 egion, where [D/] >> n+[VA3+ ///]+ [VB ///], Eq. 3 -7 can be written in a simple r form .51 ] [V p ] [DO / (3 -8) Since [D/] is constant, the following equation is satisfied. p p p 0 0 tp o ] o [V ] o [V o ] o [V t ] o [V (3 9) The electrical conductivity of a p type conductor is defined as F p (3 10) w her s Assuming that the mobility is constant in small pO2 range s Eq. 3 -9 can be written as p p p 0 0 t 0 0 tp o ] o [V ] o [V o ] o [V t ] o [V (3 -1 1 ) From the site relationship of oxygen vacanc ies and oxygen in the lattice (Eq. 31 2 ), [Vo ]+ [Oo x] = 3 (3 1 2 ) Mt / M can be re written in terms of an ele ctrical conductivity change as given by p p p M M0 0 t 0 0 t t p o ] o [V ] o [V o ] o [V t ] o [V o ] o [O ] o [O o ] o [O t ] o [O (3 -1 3 ) Technically m easurement of electrical conductivity changes is easier and more accurate compared to measurement of small mass changes at high temperatures. Theref ore, the

PAGE 48

48 ele ctrical conductivity changes are measured and fit by Eq. 3 -6 in order to obtain Dchem and Kchem. 3.1. 2 Electrical Conductivity Measurement The Lock in amplifier serves as a voltage source. By connecting a large external resistor in series as shown in Figure 3 -1 a constant current is applied to the sample according to external amplifer in Lock external sample amplifer in LockR V R R V i when Rexternal >> Rsam ple (3 1 4 ) where V is the output voltage from the Lock in amplifier, and Rsample and Rexternal are the resistances of the sample and external resist or, respectively. From Ohm s law the electrical conductivity of the sample was obtained according to b a c i V measured sample (3 1 5 ) w here Vmeasured is the measured voltage at a given current (i). From Figure 3-1 a, b, and c define sample geometries of thickness, width and measurement distance respectively. 3.1. 3 Effect of Surface R oughness Apparent rate coe fficients for oxygen diffusion in the lattice and oxygen exchange reaction at the surface have been reported.4649, 52 Generally there are agreements in reported values of the oxygen diffusion coefficient s but the oxygen surface exchange coefficient s have presented a rather wide range of difference s This is in part due to different sample preparation. Figur e 3-2 schematically describes the effect of surface roughness on the boundary conditions at the surface In the figure, J1 and J2 are the total amount of diffused and exchanged oxygen and at the surface, respectively. A1 and A2 are the areas of the sub -su rface and the top surface J1 and J2 should be same if the

PAGE 49

49 desorption is neglect For a flat surface, t he area of A1 and A2 are equivalent and therefore this requirement is same as the boundary condition used in the derivation of kinetic model for elec tri cal conductivity relaxation (Eq. 3-4 ). By contrast A2 is much larger than A1 when the surface is rough. In this case, Eq. 3 -4 is no longer adequate but the boundary condition needs to include the contribution of the higher surface area. As a result, when the sample having a rough surface is tested, it is likely to be an overestimated value in the form of KA2/A1. Figure 33 compares Dchem and Kchem of polished and unpolished LSCF samples at 800oC. The sample dimension was approximately 2x2x20 mm3, therefo re the two dimensional diffusion model was applied to fit the data. The pO2 step s were 10 21 44% with a total flow rate of 500 ccm. W hile Dchem is roughly constant between the two samples, Kchem from the unpolished sample is approximately five times higher than that of the polished sample. T he surface area of the unpolished sample, which was measured using a tomic force microscope, was five times larger than that of the polished sample It shows that an overestimated value due to surface roughness is close t o a factor of A2/A1. Therefore, for accurate measurement of the apparent oxygen exchange coefficient, the sample must have a flat surface or the surface roughness and area need to be measured. 3.2 Electrochemical I mpedance Electrochemical impedance spec troscopy is a widely used technique for de convoluting different reaction processes .53 It measures a time dependent current in response to a small sinusoidal voltage perturbation. In a complex plane, i mpedance (Z) is given by

PAGE 50

50 img real o i oZ Z ) isin (cos Z e Z Z (3 -16 ) where Zo is the ratio of Vo and io the a mplitude of voltage and current, respectively. is the phase shift Zo cos is a real component and i Zosin is an imaginary component A plot of Zreal on the x axis and Zimg on the y axis is called a Nyquist plot. Figure 3 4 shows an example of a Nyquist plot for an electric circuit, in which a resi stor and a capacitor are connected in parallel The capacitor and resistor are pure imaginary and real components, respectively T he capacitor works as a short circuit at high frequencies and current flow s via the resistor at low frequencies. In the interm ediate frequency range, impedance is the sum of imaginary and real component s It is important to understand the behavior of an RC circuit because many electrochemical reactions exhibit similar behavior 3.3 Temperature P rogrammed I sotopic E xchange In a dynamic equilibrium, the net flux of oxygen exchange between metal oxide and surrounding gas is zero, but there is continuous exchange. Temperature programmed isotopic exchange technique (TPX) is designed to distinguish lattice oxygen, which comes out of a sample, from gaseous oxygen, which incorporates into a sample, in a dynamic equilibrium condition. This is done by using oxygen isotope (18O2) in conjunction with a mass spectrometry (MS). One can assume that a preheat treated sample in normal oxygen (16O2) possesses only 16O in the lattice and surface. If the feed stream is switched from 16O2 to 18O2 at room temperature, 18O2 incorporates into the sample while 16O2 comes out of the lattice. However, due to slow exchange kinetics, essentially no change in MS signals is observed at room temperatures, and thereby the input concentration of 18O2 is the same as the output concentration of 18O2. Since the

PAGE 51

51 oxygen exchange kinetic s are a function of temperature, oxygen exchange reaction increases with increasing temperatures. The onset temperature is called the temperature where the concentration of 18O2 substantially drops. At this temperature, the MS signal of 16O2 and 16O18O (scrambled species) starts to increase. In TPX analysis, comparison of the catalytic ac tivity for oxygen reduction is made by comparing the onset temperature. A lower onset temperature implies a higher catalytic activity. Details of theoretical background can be found elsewhere.54 Note that TPX tests a powder sample Therefore, it can avoid the impact of electrical connections and sintering, which may affect surface catalytic activity.55 In addition, TPX is a surface sensitive technique due to high surface area of powder sample. 3. 4 Materials Characterizations The instruments which were used in this study for material characterizations and their purpose of usage are listed in Table 3-1. Two chemical analyses were applied in this study : Energy dispersive spectroscopy a nd Auger electron spectroscopy. While the former provides the chemical composition of a bulk, the latter gives a surface composition. As th e Core level electron is removed by X -ray photo ionization or electro n -impact ionization (Figure 3-5 a), the atom is left in an excited state, and subsequently deexcitation occurs (Figure 3-5 b). The energies of electrons in the shells are quantized and therefore, the energy difference between electron shells is a specific value for each element. In the Auger process, an outer level electron fills the core level and the excess energy ejects another electron in an outer level with a spec ific kinetic energy (F igure 3-5 c). This electron is called Auger electron.565 8 In the characteristic Xray process, the difference in the energy level is emitted as a photon of ele ctromagnetic radiation (Figure 3 -5 d). The Auger electron and characteristic X -ray present

PAGE 52

52 substantially different escape depth as illustrated in Figure 3 -6 As an Auger electron has low kinetic energy, an escape depth of Auger electron is less than a few nano meters. Therefore, Auger electron spectroscopy which is based on Auger process, is a surface sensitive technique. And Energy dispersive spectroscop y which is based on characteristic X -ray process provides bulk composition.

PAGE 53

53 Fig ure 3 1 Sample dimension and electrical connection between the sample and Lock in Amplifier. Figure 32 Schematic of the effect of a rough surface on the boundary conditions.

PAGE 54

54 Figure 33 The effect of roughness on the measurement of Dchem and Kchem (top), and surface roughness and actual surface area of samples (bottom). Figure 34. Nyquist plot of an RC circuit.

PAGE 55

55 Figure 35 C ore shell ionization and subsequent de excitation via (c) Auger process and (d) characteristic X -ray emission.58 Figure 36 Schematic diagram of escape depth of Auger electron and characteristic X ray.58

PAGE 56

56 Table 3 1. List of instruments for material characterizations. Instrument Purpose Model and company AES Chemical analysis of top surface Perkin -Elmer PHI 660 Scanning Auger Multiprobe AFM Surface roughness Digital Instruments Dimension 3100 BET Surface area of powder Quantachrome NOVA 12 00 FIB Preparation of TEM specimen Dual Beam Focused Ion Beam (FIB) Strata DB 235 SEM Surface morphology JEOL JSM 6400 / 6335F TEM Analysis of the cross section for nano precipitate JEOL TEM 2010F EDS Chemical analysis XRD Crystal structure of secondary phase s and synthesized powder s XRD Philips APD 3720 Philips MRD X'Pert System

PAGE 57

57 CHAPTER 4 MECHANISTIC UNDERSTA NDING OF LSCF DEGRAD ATION 4.1 Introduction La0.6Sr0.4Co0.2Fe0.8O3 (LSCF) has been widely used as a solid oxide fuel cell ca thode. Due to partial substitution of Sr for La and reduction of transition metals, it has a substantial oxygen vacancy concentration at high temperatures and therefore exhibits high ionic and electronic conductivity. However, several studies report that Sr in LSCF tend s to segregate to electrode surface s or interface s 38, 59 61 As a result the transition metal concentration was found to substantially decrease in the outermost surface layers. It is generally believed that B site transition metals play a critical role for catalytic activity of ABO3 perovskite materials,26, 66 68 and heterogeneous interaction is primarily determined by surface composition and structure.62 Consequently, the formation of Sr enriched layer on LSCF surfaces can be a potential degradation mechanism of LSCF cathode -based cell s L ongterm stability is an important requirement for the commercialization of SOFCs. Thus, the development of LSCF for a practic al SOFC cathode requires an understanding of Sr enrichment behavior on the surface. In the work presented herein, surface chemical compositional change s of LSCF at high temperatures w ere investigated by Auger electron spectroscopy (AES) and transmission el ectron microscopy (TEM). Dense and polished LSCF samples were examined to clearly identify surface morphological change s 4. 2 Experiment al 4.2.1 Sample preparation Figure 4 1 describes the sample preparation procedure. A dense LSCF sample

PAGE 58

58 was prepared for analysis It was prepared by pressing La0.6Sr0.4Co0.2Fe0.8O3 powder (Praxair Specialty Ceramics, USA) uni axially and sintering a pellet at 1400oC for 4 hours in air. The density of the sintered body measured by the Archimedes method was ~ 97 % of the theoretical value (6.36 g/cm3). The sintered body was cut and then Several samples were mounted on a holder, which has a flat surface, using crystal bond. C rystal bond is an adhesive that melt s at around 100oC so it need ed to be heated on a hot plate prior to application. After polishing, samples were detached either by reheating above 100oC or submerging in acetone. By doing so, sample handling was much easier and multiple samples were polished simultaneously. All samples were cleaned with DI water and acetone before heat treatments. Final sample thickness es were around 1 mm. 4.2.2 Heat treatment Samples were located on a quartz plate in the middle of a quart reactor as s een in Figure 42 They were exposed to stagnant air between 6009 00oC in the absence of Fe -Cr alloy. As K -type thermocouples (TC) could be a source of Cr contamination, the thermocouple was placed outside of the quartz reactor. Heat treatment was conducted for 50 hours and 100 hours in air and heating and cooling rates were established at 5oC/min. LSCF was also heat treated in N2 and 0.1% O2 for 50 hrs at 800 and 890oC. Gas flow rate was controlled by mass flow controller (MKS) and the total flow rate was fixed at 10 SCCM. 4.2.3 Characterization X -ray diffraction patterns of the samples before and after heat treatment were measured (Philips APD 3720) with Cu ). Low angle mode XRD

PAGE 59

59 (Philips MRD X'Pert System) was applied in order to analyze the structural variation of the surface region. All data were compared to the literature and the Joint Committee on Powder Diffraction Standards database (JCPDS). The morphological and compositional change was characterized using scanning electron mic roscopy (SEM, JSM 6335F) combined with energy dispersive spectroscopy (EDS, Oxford, UK). ZAF correction was used for semi quantitative analysis, where Z is the atomic number, A is the absorption correction factor, and F is the fluorescence correction facto rs. The standard materials used in the analysis are quartz(SiO2), Wollastonite(CaSiO3), Cr, Fe, Co, SrF2, BaF2, and LaB6 were used for O K, Ca K, Cr K, Fe K, Co K, Sr L, Ba L, and La L, respectively. The cross sectional specimen of the precipitate, which w as formed on the surface region after heat treatment, was prepared by focused ion beam (FIB, FEI Strata DB235) and analyzed using transmission electron microscopy (TEM, JEOL 2010F) and EDS. Surface chemical composition of LSCF before and after heat treatm ent was measured by Auger electron spectroscopy (AES, AES Perkin Elmer PHI 660 ). The accelerating voltage was 10 kV and the scanned energy ranges were between 50 to 2050 eV, which covered the characteristic peaks of LaMNN (625 eV), SrLMM (1649 eV), Co (775 eV), and FeLMM (703 eV). These characteristic peaks were selected in order to minimize interference with each other. Obtained AES spectrum was differentiated, and elemental atomic percents were calculated based on the differentiated data using AugerScan 3.2.0 software (RBD Instrument). The atomic percents are calculated according to m m a a/S I /S I % Atomic ( 4 1)

PAGE 60

60 where I is the peak intensity and S is the relative elem ental sensitivity factor for the Auger line used. Relative sensitivity factors, which are provided by RBD instruments, are 0.06, 0.045, 0.19, 0.13, and 0.35 for LaMNN, SrLMM, FeLMM, CoLMM and OKLL, respectively. This study calculated the relative change ba sed on the ratio of atomic percents before and after heat treatment which were normalized by oxygen concentration. The enrichment factor was calculated according to HT pre oxygen cation HT post oxygen cation] X / [X ] X / [X (E.F.) Factor Enrichment (4 2) 4.3 Results and D iscussion In Figure 4 3 are r epresent ative SEM micrographs of the LSCF surfaces before and after heat treatment in air. Before heat treatment, LSCF exhibited dense and flat surfaces as shown in Figure 4 -3 a. However, after heat treatment, submicron pre cipitates were formed on the surfaces. These precipitates were more pronounced and exhibited higher areal density with increasing heat treatment temperatures from 600-900oC. Interestingly, i n the figure, the precipitates were not formed along the grain boundaries but on grain surfaces at 800 and 900oC. The magnified images in Figure 4 -4 show more clearly the shapes of the precipitates, which are different from grain to grain. The typical shapes of precipitates are rods, triangles, or irregular polygons. For the rod type, the direction of the precipitate was either well aligned in one direction or vertically crossed. The precipitates were found not only on polished surfaces but also on unpolished surface s as shown in Figure 4 4 d. It shows that the direction o f precipitates was well aligned along the facets of the grains. This observation indicates that the formation of precipitates corresponds to the surface orientation of the underlying grains. The characteristic of rod type precipitates is close to the Widmanstatten morphology, of

PAGE 61

61 which an incoherent curved interface exhibits a fast growth rate while a semi -coherent facet has slower growth.63 However, the formation of precipitates observed in this study is different from the Widmanstatten morphology in that the precipitates grow out of the surface. Chemical analysis was conducted using EDS after heat treatment and the results were compared to the as -fresh sample. As Table 4 -1 shows, EDS data indic ates roughly constant atomic percents of all elements before and after heat treatment. This indicates that the bulk composition remains unchanged and the formation of precipitates is limited to the top surface or near the surface region. AES analysis was p erformed to investigate the compositional change of the surface region. Figure 4 5 shows differentiated AES spectra of LaMNN, SrLMM, CoLMM, FeLMM and OKLL for as -fresh and heat treated LSCF. After heat treatment, a substantial increase in peak intensity was observed in Sr while that of La, Co and Fe was decreased. The elemental atomic percents obtained from AES are given in Table 4 -1 While the concentration of cations varied, the oxygen concentration remained essentially constant after heat treatment. The enrichment factor, which represents the relative change of concentrations in the surface region, is calculated based on Eq. 4 -2. Sr concentration doubled while Fe concentration halved. It should be stressed that the concentration of Co, which has been thou ght to be catalytically active, was not detected on the surface. Chemical analysis of the precipitate was conducted using transmission electron microscopy (TEM). Figure 4 6 a is a bright field image of the cross section of the precipitate on the LSCF surfa ce. The TEM micrograph shows the presence of a transition region below the precipitate on the top surface, which is marked by the

PAGE 62

62 brightest region in the figure. The EDS line scan was performed from the inner bulk region to the precipitate on the surface a s shown in Figure 4 -6 a. It shows that the intensity of Sr peak increases in the precipitate region compared to that in bulk and the oxygen peak decreases slightly. By contrast, the peak intensities of other cations decrease substantially in the precipitate. This means that the precipitate is a Sr -O. The decrease of Sr intensity in sub -surface regions indicates that Sr diffuses from the subsurface region to the surface region. F igure 4 7 shows the pO2 dependence of the formation of precipitates. Samples we re heat treated at 800oC for 50 hrs in pure N2 and 0.1 % O2, which was balanced by Ar. Compared to heat treated sample s in air at the same temperature (Figure 4 -3 d) the formation of precipitates was reduced considerably in 0.1 % O2 (Figure 4 7 a). And the fo rmation of precipitates was not found in pure N2 (Figure 4 -7 b). Such trends were consistently observed when samples were heat treated at higher temperatures in same environment s Figure s 4 -7 c and 4 -7 d show the heat treated LSCF surface at 890oC in N2 and 0.1% O2, respectively. It shows also the reduced formation of Sr -O precipitates. This indicates that in addition to temperatures, the oxygen partial pressure is a critical factor affect ing the formation of Sr -O precipitate s on LSCF surface s Sr enrichment o n the LSCF surface after heat treatment at 800oC is evident. As a result the transition metal concentration decreases. This is in good agreement with previous works Bucher et al.59 found Sr enrichment on the surfaces of LSCF after measuring surface exchange coefficients at 600-700oC. Similarly, Heide60 observed Sr enrichment on La1 xSrxCoyFe1 yO3 surfaces after heat treatment at 1000oC in 1% O2. While these previous studies on Sr enrichment mostly postulated the formation of SrO

PAGE 63

63 layers on the surface based on the results from chemical analysis, this study clearly show s the formation of submicron-sized Sr -O precipitates on the surface using SEM and TEM. The forma tion of AO precipitates on the surface was previously shown in other ABO3 perovskite materials,64 66 but the formation of the Sr -O precipitate on LSCF surfaces was visualized for the first time in the present study. Generally, elemental enrichment and precipitation is related to defect structures, especially the presence of point defects.74 77 As La1 xSrxCoyFe1 yO3 exhibits a substantial oxygen deficiency above 600oC,34, 45 oxygen vacancies may contribute to Sr O precipitation. Even though oxygen vacancies improve the oxygen reduction reaction on the surface and oxygen diffusion in the lattice, their presence reduce s the stability of th e structure by increasing repulsive forces between like atoms. The effect of oxygen vacancies on bonding strength was experimentally demonstrated by Wang et al.67 They found that the elastic modulus of ceria decreases drastically as the oxygen vacancy concentration increases. This is because oxygen vacancies weaken the attractive forces. The formation of oxygen vacancy in the lattice can be written as Eq. 4 3 using Kroger -Vink notation. o 2 oV (g) O 2 1 2h O ( 4 3) Figure 4 8 a shows the schematic of the (100) plane of a cubic perovski te structure. Due to originally broken bonding, atoms on the surface are unstable compared to those in bulk. If oxygen ions are removed from the lattice according to Eq. 4 -3, the shielding effect between cations will be removed and atoms in the surface region will become more unstable. Thus, i n order to reduce the excess surface energy corresponding to unstable surfaces it is probable that some cations near the surface region precipitates

PAGE 64

64 on the surface, yielding cation vacancies in the lattice as demonstr ated in Figure 4 -8 b. As the binding of A -O is generally weaker than the binding energy of B -O in the ABO3 perovskite structure, the precipitation of A site cation s will be more energetically favorable than the precipitation of B site cation s.68, 69 Therefore, enriched Sr in LSCF is likely to precipitate into Sr -O to reduce the excess surface energy caused by oxygen vacancy formation. LSCF was heat treated at the same temperatures over various pO2 range s from pure N2 to air. At lower pO2, LSCF exhibits a higher oxygen vacancy concentration, resulting in increased probability for precipitate formation However, LSCF shows reduced formation of Sr -O precipitates with decreasing pO2. A similar pO2 dependence of elemental enrichment was reported by Dufour et al.70 T hey found that the ratio of Ba to Ti on the surface s of BaTiO3 bec a me doubled after heat treatment at 1300oC in air but the ratio was decreased to 1.4 when surfaces were heat treated in N2. Th is indicates reduced Ba enrichment at lower pO2. Szot et al.68 studied the nature of the surface layer in ABO3 perovskite structure materials such as SrTiO3 and observed the formation of A O layers at elevated temperatures, which is known as the Ruddlesden-Popper (RP) phase.71 They propose that the RP phase is likely to be more stable at high pO2 and therefore the formation of A -O layers can be inhibited at reduced pO2. The formation of Sr-O precipitate s on LSCF surface s can be understood by a two -step mechanism:(i) the formation of oxygen vacancies, and (ii) subsequent precipitation of enriched Sr. Step (i) increases with decreasi ng pO2 but step (ii) occurs at oxidizing conditions i.e. at high pO2. As LSCF exhibits a high oxygen vacancy concentration in oxidizing condition s the overall reaction is likely to be controlled by step (ii). Konigstein et al.72 reported that

PAGE 65

65 thermodynamically stable compounds of Sr -O are SrO and SrO2. T he step (ii) for Sr-O precipitation, then, can be written as 2h V SrO(ppt) (g) O 2 1 Sr/// La 2 / La ( 4 -4 ) o /// La 2 x O 2 / LaV V (ppt)SrO O (g) O 2 1 Sr ( 4 5 ) The crystal structure of precipitates was analyzed using low angle XRD but the presence of either SrO or SrO2 could not be identified due to the limited amount of the secondary phase. So, it can not be concluded decisively which reaction is more dominant. However, the reaction increasing hole concentration is likely to be thermodynamically unfavorable because the formation of holes increases the concentrations of Co4+ and Fe4+, which are unstable species at high temperatures.8 A further systematic quantum mechanical study is required to conclude the dominant reaction mechanism. Recently, Simner et al.38 measured power density of anode -supported cells utilizing LSCF cathodes over long term periods at 750oC and observed decrease in the cell perf ormance. Similarly, Tietz et al.73 also reported substantial degradation of LSCF cathode based cells during long term testing. Obvi ously the formation of Sr -O precipitates on the surfaces of LSCF cathodes can contribute to such degradation of cell performance during the test. The oxygen reduction reaction at the cathode consists of a series of elementary steps such as i Gas phase diffusion ii. Adsorption of molecular oxygen (O2(g) + sad2 (ad) ) iii. Dissociation into atomic oxygen (O2(ad)

PAGE 66

66 iv Charge transfer (O(ad)+2e/// (ad) ) v Incorporation of oxygen ions (O// + VO o x) where sad is the oxygen adsorption sites on the surface. Sr-O is an electronically insulator.74 The collected data of the surface exchange coefficients and the oxygen tracer diffusion coefficients o f various metal oxides evidence that the electronic species on the surface are essential for oxygen exchange reaction.42 I t is expected that the formation of Sr -O precipitate on the surface of LSCF form s a passive layer reducing the concentration of electronic species and inhibiting oxygen reduction reaction, especially charge transfer. In addition, f or ABO3 perov ksite materials, it is generally believed that B site transition metals play a critical role for catalytic activity.26, 66 68 Takeda et al.44 measured the cathodic polarization of La0.7Sr0.3MO3 (M=Cr, Mn, Fe, Co) and found that La0.7Sr0.3CoO3 exhibited the lowest polarization at 800 oC in air. Kilner et al.31 investigated the dependence of surface exchange coefficients (Ko) on the composition of La0.8Sr0.2Mn1 yCoyO3 and found that Ko increased with increasing Co concentration. Kan et al.54 studied the effect of catalyst impregnation on the catalytic activity for oxyg en exchange for La0.8Sr0.2MnO3(LSM). They found oxygen exchange reaction was improved with the addition of Co on LSM surfaces via wet impregnation. The results in literatures indicate that Co is a catalytically active element for oxygen reduction reaction. Consequently, d eficiency of catalytically active transition metals due to Sr enrichment is likely to further reduce oxygen reduction reaction. Thus, Sr -O precipitation on LSCF surfaces can cause substantial increases in nonohmic resistance. It can also l ead to increases in ohmic resistance if Sr is enriched at the interface of LSCF and other cell components, forming secondary phases such as SrZrO3 or SrCrO4. 46, 85

PAGE 67

67 4 4 Conclusions LSCF surface s exhibit substantial compositional variations in oxidizing condition at elevated temperatures. This heterogeneous surface structure involves Sr enrichment and subsequent precipitat ion into submicron sized Sr-O on the s urface. A t wo -step mechanism was proposed for Sr -O precipitation: (i) the formation of oxygen vacancies, and (ii) subsequent precipitation of enriched Sr. As LSCF exhibits high oxygen nonstoichiometry at elevated temperatures, it is likely that step (ii) c ontrol s the overall reaction. Sr -O is electrically insulat ing and Sr enrichment results in the depletion of transition metals Consequently, the heterogeneous surface structure can lead to substantial degradation of the surface catalytic activity for LSCF.

PAGE 68

68 Figure 41. Brief description of sample preparation procedure. Figure 42. Heat treatment set -up.

PAGE 69

69 Figure 4 3 SEM micrographs of LSCF of (a) before heat treatment and after heat treatment at (b) 600oC, (c) 700oC, (d) 800oC a nd (e) 900oC for 50 hours in the absence of Fe -Cr alloy.

PAGE 70

70 Figure 4 4 Magnified images of the precipitates on LSCF surfaces after heat treatment at (a) 900oC for 50 hours (b) 900oC for 50 hours (c) 860oC for 100 hours and (d) 890oC for 100 hours

PAGE 71

71 Fi gure 45 Differentiated Auger electron spectra of the (a) LaMLL, (b) SrLMM, (c) CoLMM, (d) FeLMM, and (e) OKLL before(blue solid line) and after (red dotted line) heat treatment at 800oC for 50 hours Figure 4 6 TEM bright field image of the c ross se ction of precipitate ( left side) and EDS lin e scan of selected elements ( right side)

PAGE 72

72 Figure 4 7 SEM micrographs of LSCF surfaces after heat treatment for 50 hours at (a) 800oC/ N2, (b) 800oC/0.1 % O2 (c) 890oC/N2 and (d) 890oC/0.1 % O2 Figure 4 8 Sch ematic diagram of (a) the formation of oxygen vacancies and (b) formation of cation and anion vacancies simultaneously.

PAGE 73

73 Table 4 1. Elemental atomic percents (%) of preand post -heat treated LSCF at 800oC obtained from EDS and AES EDS La Sr Co Fe O As -fresh 12 9 4 14 61 After heat treatment 13 9 4 15 59 AES La Sr Co Fe O As fresh 30 12 5 11 42 After heat treatment 28 22 0 7 43 Enrichment factor 0.9 1.8 0 0.6

PAGE 74

74 CHAPTER 5 IMPACT OF CO IMPREGN ATION INTO LSCF CATHODE ON SOFC PERFORMANCE 5.1 Introduction Recently the mixed ionic electronic conductor (MIEC) has been widely used for a solid oxide fuel cell cathode due to the high ionic and electronic conducing properties22, 41 They have also exhibited enhanced catalytic activity for oxygen reduction compared to the pure electronic conductor.42 Ho wever, an attempt to reduce operating temperatures causes the reduced rate of oxygen reduction reaction (ORR), which results in a substantial increase in activation overpotential. Therefore, tremendous efforts have been done in order to improve the cathode performance.42, 7579 Since the ORR is a heterogeneous reaction, the interaction between the cathode and oxygen is primarily affected by surface characteristics. Thus, enhancing surface catalytic activity by disper sing nano -sized particles on the surfaces of the porous electrodes can lead to an improvement in cathode performance.91, 92 Previously, precious metals were impregnated into the LSCF cathode in order to enhance the LSCF performance. This method is used to deposit catalytically active fine particles on the surface of a porous electrode backbone as described in Figure 5-1 Sahibzada et al .80 investigated the effect of the addition of Pd into the porous La0.6Sr0.4Co0.2Fe0.8O3 (LSCF) cathode and observed 3 4 times lower cathodic impedance in the temperature range of 400 750oC after Pd impregnation. Similar enhanced cat hode performance by addition of Pd or Pt was also reported by Simner et al.81 and Uchida et al.82 However, from the economi c point of view, it is advantageous to use abundant elements for the reduction of manufacturing costs. Yamahara et al.83 tested impact of Co i mpregnation into the La0.85Sr0.15MnO3 cathode and found

PAGE 75

75 reasonable increase in maximum power density and decrease in electrode resistance after impregnation. However, the impact of Co impregnation into the LSCF cathode has not been tested yet. It is repo rted that the chemical composition of the LSCF surface is substantially different from the stoichiometric values. Viitanen et al.61 studied the chemical compositional change of LSCF after an oxygen permeation experiment at 900oC using X -ray Photoelectron Spectroscopy (XPS). They observed the abs ence of Co and Fe in the outermost atomic layer from contamination-free LSCF. Bucher et al.59 found Sr enrichment on the surface using XPS after measuring surface exchange coefficient of LSCF at 600700oC. Oh et al.84 showed that Sr in LSCF not onl y enriched on the surface but also precipitated into SrO under an oxidizing condition. They observed reduced concentration of transition metals on the surface in accordance with SrO precipitation. Vance et al.85 and Lee et al.86 reported that excess Sr increases the electrode resistance. It is generally believed that B site transition metals play a cri tical role for catalytic activity of ABO3 perovskite materials.24, 69, 87, 88 Therefore, Sr enrichment, can degrade the surface catalytic activity. In this study, cobalt oxide was additionally dispersed into the porous LSCF cathode instead of precious metals in order to improve LSCF performance. The impact of Co i mpregnation was tested by measuring the electrode resistance and cell performance. 5.2 Experimental 5.2.1 AC I mpedance A nalysis A symmetric cell, which consists of LSCF / GDC/LSCF was prepared for the impedance measurement. A dense Gd0.1Ce0.9O2 (GDC, A nan Kasei, Japan) electrolyte was made by sintering a pressed pellet at 1450oC for 4 hours A LSCF ink was

PAGE 76

76 prepared by mixing LSCF powder with texanol based polymer vehicle (ESL electro science, USA) using Thinky mixer (Thinky, Japan). The ink was symmetri cally painted on both sides of the electrolyte, each having a same electrode area of 0. 28 cm2. Pt wires were connected as a current collector and LSCF ink was applied for better contact. After that, the cell was sintered at 1100oC for 1 hr with heating and cooling rate of 5oC/ mins. Figure 52 shows the LSCF/GDC/LSCF symmetric cell and the schematic of AC impedance measurement set -up. AC impedance of the symmetric cells was measured in air in the temperature range between 600 and 800oC with 50oC intervals using a Solartron SI 1260 frequency response analyzer. The amplitude of the excitation voltage was 50 mV and the AC frequency range was from 0.1 to 105 Hz. The experimental data was fit using the equivalent circuit shown in Figure 5 -3 in order to de convolute the impedance response and extract the electrode resistance. It consists of a series of resistors (R) and constant phase elements (CPE) CPEs were used instead of capacitors to fit the depressed semi -circle behavior.89 In the figure, RE is the electrolyte o hmic resistance and RH F and RLF refer to high and low frequency components respectively. The overall electrode resistance is the sum of RH F and RLF The area specific resistance of the electrode is obtained by 2 Area ) R (R ASRElectrode LF HF Electrode (5 1) 5.2.2 Electrical Conductivity Relaxation A dense LSCF sample was prepared for analysis It was prepared by pressing La0.6Sr0.4Co0.2Fe0.8O3 powder (Praxair Specialty Ceramics, USA) uni axially and sintering a pellet at 1400oC for 4 hours in air. The density of the sintered body

PAGE 77

77 measured by the Archimedes method was ~ 97 % of the theoretical value (6.36 g/cm3). The sintered body was cut and then polished with a series of SiC papers and diamond All samples were cleaned with DI water and acetone before ECR test The sample dimension was approximately 0.1x2x20 mm3. A bar of LSCF was positioned on the alumina tube i n the middle of the reactor and a K -type thermocouple was placed below it. A N2 and O2 gas mixture was fed into the reactor through a mass flow controller (MKS & Edwards). The total flow rate was fixed at 300 ccm. The electrical conductivity was measured v ia the four point probe method. The conductivity change was monitored by a Lock in amplifier (Stanford, SR830) automated with Labview software. The sample was heated to 800oC at a rate of 5oC/min in air. Before the ECR measurement, the feed stream was swit ched to 3% O2. Once it reached steady state the feed stream was switched via four way switching valve. The oxygen partial pressure ( pO2) step was programmed to increase by the following steps: 3 % 5.2. 3 Button C ell F abrication and T ests NiO -GDC anode supports were fabricated by the tape-casting method. A tap e c as t ing slurry was prepared in two steps First, a mixture of NiO (Alfa Aesar) Ce0.9Gd0.1O1.95 (Anan) and a dispersant Solsperse was ball milled in a toluene and ethanol solvent system for 24 hours. After that, a plasticizer di -n butyl phthalate (DBP) and polyethylene glycol 5000 (PEG5000) and a binder polyvinyl butyral (PVB) were added to the slurry After 24 hours t he slurry was de gased to remove trapped air bubbles, whic h cause cracks or defects during casting. The NiO -GDC anode tape was tape cas t ed using a Procast tape casting system (DHI) and then the tape was dried for 2 hours at 100oC. C ircular shape s of green tape (d ia. 32 mm) were cut out and partially

PAGE 78

78 sintered at 900oC for 2 hours to remove polymer vehicles as well as provide an appropriate mechanical strength for the electrolyte deposition. An anode functional layer was added to increase SOFC performance An e thanol based GDC precursor solution was spray coated on the surface of a pre -sintered anode and heat treated at 900oC for 1 hour. An electrolyte was then deposited on the top functional layer. A GDC electrolyte slurry which was prepared by ball -milling GDC powder Solsperse PVB and DBP in ethanol, was spin -coated on the pre-sintered body and dried at room temperature for 10 hours. Finally it was sintered at 1450oC for 4 hours using a heating rate of 3oC/ min. A LSCF cathode ink was brush-painted on the electrolyte. The cathode was fired at 1100 C for 1 ho ur. Pt paste, as a current collector, was brushpainted onto both electrodes along with Pt mesh and gold leads Current collectors were heat treated at 900 C for 1 hour. Figure 5 4 shows a diagram of the current density and voltage measurement set up. Th e button cell was loaded in a ho use built fuel cell testing set up and gas sealed using a two -part ceramabond sealant (a mixture of 517powder and 517liquid from Aremco). 30 ccm of dry air and 30 ccm of wet hydrogen were supplied to the cathode and anode s ide s respectively. The cell voltage was measured as a function of current density using a Solartron 1287 potentiostat. I mpedance was measured at open circuit conditions by two -point probe methods using a PARSTAT 2273 frequency response analyzer ( Princeto n Applied Research) with a frequency range of 0.1 to 105 Hz The e lectrode area specific resistance ( ASR) was calculated from the area below the semi circle in the Nyquist impedance plot.

PAGE 79

79 5.2. 4 Catalyst I mpregnation A Co precursor solution was prepared by dissolving Co(NO3)26H2O (Alfa Aesar 99.999% ) in ethanol and its pH was measured using a pH meter (Seven multi, Mettler Toledo). The Co precursor was impregnated into the porous body of the cathode by dipping the cell into the Co nitrate solution while a vacuum was pulled NiO -GDC Anode was dense before reduction in H2 condition, and so Co solution was barely impregnated into anode side. Co impregnation was continued until no more air bubbles came out from the surface of the cathode. It was taken from t he solution and dried at 70oC. The electrolyte area around the electrode was cleaned in ethanol and then heat -treated at 850oC for 30 mins 5.2. 5 Characterization The morphological change was characterized using scanning electron microscopy (SEM, JSM 6335F ) combined with energy dispersive spectroscopy (EDS, Oxford, UK). ZAF correction is used for semi quantitative analysis, where Z is the atomic number, A is the absorption correction factor, and F is the fluorescence correction factors. 5. 3 Results and Dis cussion 5. 3 .1 Effect of Co Impregnation on Electrode Impedance ASR -AC impedance spectroscopy measurements were carried out in the temperature range of 600 800oC. Figure 5 -5 shows the AC impedance spectra for LSCF/GDC/LSCF symmetric cell at three different temperatures. The area below the semi -circles corresponds to the electrode resistance, and the high frequency intercepts of the semi -circles were normalized for easier comparison. There are approximately two semi -circles, and each semi -circle corresponds to a different reaction step.90 The greatest contribution to the electrode resistance is the high frequency component. As

PAGE 80

80 temperature decreases, the size of the high frequency semi -circle increases, while that of the low frequency semi -circle changes slightly. However, two semi circles were apparently combined into one below 700oC and so they could not be de -convoluted each other. Esquirol et al.91 and Sahibzada et al.80 measured the impedance of LSCF/GDC/LSCF symmetric cell and observed similar behavior. They attributed the high and low frequency components to the surface reaction and gas phase diffusion, respectively. It is expected that the resis tance for the surface reaction increases substantially as the temperatures decrease while the mass transport of oxygen gas depends little on temperature. And so their explanation justifies the observed dependence of AC impedance spectra on temperature. Fi gure 5 6 shows the impedance spectra for the symmetric cells of plain LSCF and Co impregnated LSCF. It shows that Co impregnation into LSCF reduces RHF while RLF is unaffected, indicating that it does not retard oxygen gas diffusion but improves the surfac e reaction rate. At temperatures below 700oC it is difficult to distinguish the contribution of the high and low frequency component and so the total electrode ASRs are compared each other. Figure 5 7 shows the ASRElectrode of plain and Co impregnated LSCF plotted as a function of temperature in log scale. For the plain LSCF 2 at 800oC, which is consistent with literature values.91 2 at the same temperature after Co impregnation, providing a 32 % reduced value. The activation energies were obtained from Arrhenius plots. They were 1.44 eV and 1.55 eV for plain and Co impregnated LSCF electrodes, respecti vely. The activation energy is only slightly changed after Cr impregnation. It seems that the rate limiting step does not change with Co impregnation.

PAGE 81

81 ASRElectrode is related to the surface exchange coefficient (ko) according to100, 102 o Electrode 2 2 oC ASR F z T R K (5 2 ) where R is the gas constant, T is the absolute temperature, F is faraday constant, n is the number of electron transferred during the reaction, and Co is oxygen concentration. Ko is the apparent rate coefficient of the ORR, in which a series of elementary steps occurs such as92 1 Gas phase diffusion 2 Adsorption of molecular oxygen (O2(g) + sad2 (ad) ) 3 Dissociation into atomic oxygen (O2(ad) 4 Charge transfer (O(ad)+2e/// (ad) ) 5 Incorporation of oxygen ions (O// + VO o x) where sad is the oxygen adsorption sites on the surface. And so the reduction in ASRElectrode means that one of the elementary reactions becomes faster due to Co impregnation. It is expected that incorporation of oxygen ions remains unaffected. And the resistance for gas phase diffusion was found to remain roughly constant after impr egnation. Therefore, one of adsorption, dissociation, or charge transfer process is likely to be activated by Co impregnation. 5.3.2 Effect of Co Impregnation on Power Density of a S ingle Cell Figure 58 shows the output voltage and power density of single cells as a function of current density. One has a plain LSCF cathode and the other has a Co impregnated LSCF cathode. They showed similar open circuit voltage at 600oC. However the maximum power density is significantly different. The maximum power density of the cell with plain LSCF cathode was around 0.5 W/cm2. After Co impregnation, it increased

PAGE 82

82 up to 1.2 W/cm2. The gas flow rates were 90 SCCM of wet H2 (3 % H2O) on anode side and 30 SCCM of air on cathode side. Co impregnation also increased cell performance at 550 and 500oC as seen in F igure 5-8 The maximum power density of button cells was plotted as a function of temperatures between 500-600oC in Figure 5-8 d. Note that these are not composite, but single component cathodes. The maximum power density is high compared to the composite c athode in literatures.93, 104 Factors affecting SOFC performance are complex. In this study, the contribution of electrolyte and electrode was distinguished by measuring the impedance in an open circuit condition. Figure 5 -9 compares the impedance spectra of the two cells at 600oC. The reduction in the size of semi -circles is pronounced after Co impregnation, which is in agreement with symmetric cell test. The X axis intercept is slightly reduced in Co impregnated cell. It seems that the two cells have slightly different mic rostructure even though they were fabricated at the same time. Post test analysis shows that they have different thickness of the electrolyte, which was around 15 and 10 and Co impregnated cells, respectively. Figure 5 10 is the ASR of electrode and electrolyte, plotted as a function of temperatures. Co impregnation brings a more significant reduction in electrode resistance while electrolyte ASR was decr eased slightly. The magni tude of reduction in electrode ASR was higher at lower temperatures. This result indicates that the improvement in cell performance is mostly brought by reduction in electrode resistance. Since Co impregnation does not affect the bulk properties of LSCF cathode such as electronic conduction, the observed reduction in electrode resistance can be attributed to enhanced ORR at the surface and not the reduced in ohmic resistance.

PAGE 83

83 Post -test analysis was conducted using SEM. Figure 5 11 a is the SEM micrograph of the cross section of post -tested button cell. And Figure 5 -11b is the magnified images for a cathode side. After Co impregnation, deposition of nano sized particles was found on the surface of LSCF cathodes. However, atomic percents obtained from EDS do not provide any noticeable difference between the plain and Co impregnated LSCF cathode. This is because the penetration depth of electron beam is much deeper than the size of the nano sized particle. Previously Yamahara et al.83 impregnated Co nitrate solution into an LSM cathode in similar way to this study, and they observed the formation of Co3O4 nano particles using Transmission electron microscopy. And sot i t is likely that the newly formed nano particle on LSCF is also Co3O4. The reduced electrode resistance is mostly attributed to the addition of cobalt oxide on LSCF surface. However, the removal of Sr enriched sur face layers can also contribute to the improvement in electrode performance. Miura et al.93 tested the oxygen permeability of La0.6Sr0.4Co0.8Fe0.2O3 membrane and obtained improved permeability with acid treatment. They performed XPS analysis to identify the reason for activation, and found a reduced concentration of excess SrO on the surface. In this study, a cell was dipped into a nitrate solution during impregnation process. The pH of Co nitrate solution measured was 2.15. This is not a strong acid, but the possibility that a few surface layers were etched during the Co im pregnation process cannot be ruled out and may contribute to enhance d ORR. 5.3.3 Characteristic D ependence of Kchem on O xygen P artial P ressure The chemical exchange coefficient (Kchem) and chemical diffusion coefficient (Dchem) of LSCF were measured at 80 0oC using electrical conductivity relaxation. And they are plotted as a function of pO2 on a log scale in Figure 5 -12a Obtained Kchem and

PAGE 84

84 Dchem values can be converted into surface exchange coefficients (Ko) and the self diffusion coefficients (Dself) usi ng Eqs. 5 3 and 54 .94 K Kchem o ( 5 3 ) D Dchem self (5 4 ) o 2C ln pO ln 2 1 ). In addition, the oxygen vacancy diffusion coefficient (Dv) can be calculated according to53 (3 D C C D Dself V o self V ( 5 -5 ) where Co and Cv are the concentration of oxygen and oxygen vacancies, La0.6Sr0.4Co0.2Fe0.8O3 The thermodynamic factors and (3800oC are given in Table 5 1.47, 94 Ko, Dself, and Dv were calculated using the values in Table 5 1 and plotted as a function of pO2 seen in Figures 5 12b and 5 -12 c, respectively. The slope in the figure indicates the dependence of K and D values on pO2. Such dependence must reflect the behavior of oxygen transport mechanisms. For example, DV is proportional to pO2 0.073, which is roughly independent of oxygen partial pressure. This indicates that the mobility of oxygen vacancies is constant in this pO2 range, which is a reasonable result. Maier assumed a thin layer, charge-transfer, or adsorption process were ratelimiting steps for the ORR, and set the kinetic models for each case.95 The kinetic model provided the characteristic dependencies of apparent rate coefficients on oxygen partial

PAGE 85

85 pressure as Table 5 -2 shows. The characteristic dependencies of Ko and Kchem on pO2, which were obtained in this study, were 0.1 and 0.56 respectively. A comparison of them sug gests that the rate -limiting step of the ORR is likely to be a charge transfer process. Therefore, it is likely that the addition of cobalt oxide on the surface of LSCF reduces the electrode resistance by expediting charge transfer from the cathode to the oxygen gas. 5. 4 Conclusion s Based on understanding surface characteristic of LSCF, Co was impregnated into a porous LSCF cathode. After Co impregnation, the electrode ASR was reduced and the maximum power density of a single component LSCF cathode cell w as significantly improved, giving 1.2 W/cm2 at 600oC. This result suggests that Co impregnation into LSCF surface effectively improves deteriorated surface catalytic activity caused by Sr enrichment and transition metal depletion. It seems that the oxygen reduction reaction is limited primarily by a charge transfer process and addition of cobalt oxide via wet impregnation improves this process Such activation is mainly attributed the addition of cobalt oxide on the surface but t he impact of the removal of Sr enriched surface layers during impregnation can not be ruled out. The long-term stability of the impregnated electrode structure needs to be tested.

PAGE 86

86 Figure 51. Sketch of deposition of nano size particles on cathode surface via wet impregnation. Figure 5 2 Schematic of AC impedance measurement s

PAGE 87

87 Figure 5 3 Equivalent circuit to fit the measured impedance data. Figure 5 4 Current density and voltage measurement set up

PAGE 88

88 Figure 5 5 AC Impedance spectra of LSCF/GDC/LSCF symmetric ce l l Figure 5 6 Impedance plots for symmetric cell of plain LSCF electrode and Co impregnated LSCF electrode at 800oC in air.

PAGE 89

89 Figure 5 7 Comparison of total electrode ASRs as a function of temperatures

PAGE 90

90 Figure 58 Measurement of cell voltage as a function current density at (a) 600oC, (b) 550oC and (c) 500oC. (d) Maximum power densities as a function of temperatures. Figure 5 9 Impedance of a button cell for plain LSCF and Co impregnated cathode at 600oC

PAGE 91

91 Figure 5 10 The area specifi c resistance of electrode and electrolyte for plain and Co impregnated LSCF cathode cells. Figure 5 11 (a) SEM micrograph of LSCF/GDC/Ni -GDC button cell and (b) magnified images of Co impregnated LSCF after cell test.

PAGE 92

92 Figure 512. (a) Dchem and Kchem, (b) Dself and Ko, and (c) Dv at 800oC as a function of pO2

PAGE 93

93 Table 5 1. Thermodynamic factors and ( 3 oC. 52, 94 pO 2 0.05 0.08 0.12 0.18 0.28 0.43 125.9 135.8 158.5 184.9 232.8 293.1 (3 0.0128 0.0101 0.0094 0.0074 0.0067 0.0060 Table 5 2. C haracteristic dependencies of the apparent rate coefficient s on pO2.95 Controlling parameters lnP lnko lnP lnkchem Thin layer -1/2 0 Charge transfer 0 1/2 Adsorption 1/2 1

PAGE 94

94 CHAPTER 6 MECHANISM OF CR VAPO R DEPOSITION ON SOFC CATHOES 6.1 Introduction Solid oxide fuel cell s (SOFC) need to be operated as a stack in order to generate practical (high) voltages. The stack consists of multiple single cells which are electronically connected in series via interconnects. The interconnect also separates oxygen at the cathode side of one cell from the fuel at the anode side of the adjacent cell. Thus, high electronic conductivity, gas tight ness good stability both in ai r and fuel environment, and low reactivity with other cell components are necessary properties for interconnects .96 While ceramic interconnects are generally used for high temperature SOFCs, reduction in the operating temperatures allows utilization of metal alloys. M etal interconnects are advantageous over ceramic interconnects because of reduced manufacturi ng costs, enhanced machinability and thermomechanical stability .97, 98 D espite such benefits, metal alloy interconnect s contain high concentrations of Cr14, which leads to out -diffusion of Cr to surface region, formation of Cr scale and Cr vaporization from the Cr scale at high temp eratures in oxidizing conditions (Eq. 6 -1 and 6 -2 ). (g) 2CrO (g) O 2 3 (s) O Cr3 2 3 2 (6 1 ) (g) (OH) 2CrO O(g) 2H (g) O 2 3 (s) O Cr2 2 2 2 3 2 ( 6 -2 ) M ass transport of such Cr vapor species to the cathode and the formation of a solid chromium oxide can degrade the cathode performance either by blocking electrochemically active sites or by decomposing cathodes.40, 99 SOFCs utilizing metallic interconnects undergo significant degradation during long term operation.39, 100

PAGE 95

95 The long-term stability is an important requirement for the commercialization of SOFC. Thus, the development of a practical SOFC cathode requires an understanding of the degradation mechanism caused by Cr vapor deposition. One group has hypothe sized that the degradation mechanism is based on the electrochemical reduction of high valent vapor species of chromium oxide (CrO3) and oxyhydroxide (CrO2(OH2)) to solid phase Cr2O3.40, 101 In contrast, others sugg est that the nature of Cr poisoning is a chemical process in which a nucleation agent on the cathode interacts with Cr vapor species.114116 In this study, Cr vapor deposition behaviors on various ABO3 perovskite materials were systematically investigated in order to better understand the reaction mechanism between Cr vapor species and solid oxide fuel cell cathodes. Dense and polis hed samples were examined to clearly identify surface morphological changes. A special emphasis was placed on La0.6Sr0.4Co0.2Fe0.8O3 (LSCF). 6.2 Experimental 6.2.1 Sample Preparation La0.6Sr0.4Co0.2Fe0.8O3 powder was obtained from Praxair Specialty C eramics, USA and the rest of the compounds, La0.6A0.4Co0.2Fe0.8O3 (A=Ca, Ba), SrCo0.2Fe0.8O3, and La0.6Sr0.4MnO3, were synthesized by a solid state reaction process. La2O3 (Cerac, 99.999%), SrCO3 (Johnson Matthey, 99%), BaO (Acros, 90%), Co3O4 (Alfa Aesar, 99.7%) and Fe2O3 (Alfa Aesar, 99.945%) powders, were ball milled for 24 hrs in ethanol, and dried at 70oC. They were calcined at 900oC for 6 hrs. LSCF and LSM were sintered at 1400oC for 4 hours and 1450oC for 6 hours, respectively. And the others were si ntered at 1200oC for 6 hours. The sintered body was cut and then polished with a

PAGE 96

96 cleaned with DI water and acetone before heat treatments. Final sample thicknesses were around 0.1 cm. The compositions of metal oxides and preparation method are listed in Table. 6 -1. 6.2.2 Heat T reatment Samples were heat treated under ambient conditions in the presence of Crofer22APU as described in Figure 6 1. Crofer22APU (ThyssenKrupp VDM, W erdohl, Germany) i s a ferritic stainless steel for solid oxide fuel cell interconnects. The composition of the steel, given by the supplier, is listed in Table 6 -2. Samples were located on a Crofer22APU sheet, thus the distance between the sheet and sample surfaces is 0.1 cm, which was equivalent to sample thickness. They were also located 0.5 cm away from the sheet. Heat treatments were carried out for 50 hours at 400, 600 and 800oC with heating and cooling rates of 5oC/min. Because no Cr vapor deposition was found at 800oC on LSM, it was heat treated additionally at 1050oC for 50 hours. 6.2.3 Characterization. X -ray diffraction patterns (Philips APD 3720) of the samples were measured with Cu -K ). Low angle mode XRD (Philips MRD X'Pert System ) was applied in order to analyze the structural variation of LSCF in surface region after heat treatment. All data were compared to the literatures and the Joint Committee on Powder Dif fraction Standards Data (JCPDS). Morphological and compositional changes were characterized using scanning electron microscopy (SEM, JSM 6335F) combined with energy dispersive spectroscopy (EDS, Oxford, UK). ZAF correction was used for semi quantitative analysis. In addition, surface chemical composition of LSCF before and after heat treatment was measured using Auger electron spectroscopy (AES, AES Perkin Elmer PHI 660 ). The accelerating

PAGE 97

97 voltage was 10 kV and i t scanned the energy ranges between 50 to 205 0 eV, which covered the characteristic peaks of LaMNN (625 eV), SrLMM (1649 eV), Co (775 eV), FeLMM (703 eV), and OKLL (510 eV) These characteristic peaks were selected in order to minimize interference with each other Obtained AES spectrum was different iated, and elemental atomic percents were calculated using AugerScan 3.2.0 software (RBD Instrument). It calculates the atomic percents (X) according to m m a a/S I /S I X (6 1) where I is the peak intensity and S is the relative elemental sensitivity factor for the Auger line used. The relative sensitivity factors, which are provided by RBD instruments, are 0.06, 0.045, 0.19, 0.13, and 0.35 for LaMNN, SrLMM, FeLMM, CoLMM and OKLL, respectively. This study calculated the relative change based on the ratio of atomic percents before and after heat treatment which were normalized by oxygen concentration. The enrichment factor was calculated according to HT pre oxygen cation HT post oxygen cation] X / [X ] X / [X (E.F.) Factor Enrichment (6 2) 6. 3. Results and Discussion 6. 3.1 XRD Figure 6 2 shows the XRD patterns of the sintered pellets LSCF SCF LBCF, LSM are identified in the pure perovskite phase102 while small quantities of secondary phases were detected in solid state synthesized LCCF 6. 3.2 Cr Vapor D eposition on LSCF In Figure 6 3 are representative SEM micrographs of LSCF surfaces before and after heat treatment at 800oC for 50 hours in the presence of Crofer22APU Compared

PAGE 98

9 8 to the flat surface of non-heat treated samples in Figure 6 -3 a all the surfaces showed the formation of a few micron sizes of particles after heat treatment. Heat treatment with direct contact to Crofer22APU sheet (Figure 6 3b) resulted in uniform formation of secondary phases on grain boundary and grain surfaces while the particles were formed along the grain boundary when LSCF surface was not in contact with the sheet (Figure 6 3c and 6 3d) The distances between Crofer22APU and leading edge of LSCF were 0.1 cm and 0.5 cm in Figure 6 3c and 6 3d, respectively. As the distance increased, apparently the formation of secondary phases decreased due to reduced mass transport of Cr vapor species to LSCF Elemental atomic percents for a newly formed particle and a flat region between the particles after heat treatment were obtained using energy dispersive spectroscopy (EDS). As T able 6 -3 shows, the flat region did not present noticeable difference in atom ic percents c ompared to the nonheat treated sample. By contrast, the particle exhibited substantially increased concentration of Sr and Cr indicating that the particles are Sr -Cr -Ox. Compositional variation of surface layers as a result of the formation of Sr Cr -Ox was investigated using Auger electron spectroscopy (AES) The flat region between SrCr -Ox particles in Figure 6 3d was selected for the analysis, and the values were compared to nonheat treated sample. The elemental atomic percent obtained b y AES is given in Table 6 -4. It shows that atomic percent of cations varied after heat treatment while that of oxygen remained essentially constant. The enrichment factors were calculated using Eq. 6 -2. While La concentration was slightly increased, the c oncentration of the rest of the cations became halved. In our previous study, Sr enrichment after heat treatment was evident on the surface of Cr

PAGE 99

99 contamination free LSCF.103 Notably the formation of submicron sized Sr -O precipitate on the grain surface was clearly visualized. For comparison, the elemental atomic percents and enrichment factor of the LSCF surface, which was heat treated at 800oC in the absence of C rofer22APU, were added in Table 6 -4. Sr concentration became doubled while transition metal concentration decreased after heat treatment. Such Sr enrichment phenomena on LSCF were also reported by Bucher et al.59 and Heide et al.60 In Table 6 4, the major difference between the two samples, which were heat treated in the presence and absence of Crofer22APU, arises from the A site cation while the B site transition metals show similar trends. This implies that Sr -O precipitate s on LSCF surfaces act as nucleation agent s for the formation of Sr -Cr -Ox particles. 104 As a Sr -O precipitate is formed on grain surfaces while Sr -Cr -Ox particle is preferentially formed along the grain boundaries, it is likely that Sr -O on grain surface s diffuse to grain boundary region s As a consequence, the grain surface of LSCF exhibited a Sr deficiency. Figure 6 4 shows the XRD patterns of pre and post heat treated LSCF in the presence of Crof er22APU. A single phase rhombohedral perovskite structure was confirmed for the pre-heat treated sample.105 After heat treatment, no obvious change was found in the standard normal XRD pattern. But low angle XRD, which provides intensified peaks from the surface region, identified the formation of SrCrO4. The change is pronounced only in low angle XRD because the structural variation occurs in the near -surface region. In addition to the formation of SrCrO4, low angle XRD revealed the formation of Fe2O3 and the combination of split characteristic peaks of rombohedral into a single peak (Figure 6 4b~d). Similar structural changes of LSCF was observed by

PAGE 100

100 Waller et al.106 They tested the effect of Sr deficiency on the struct ural change of LSCF and found the combination of split characteristic peaks of the rhombohedral phase. They suggested that this is evidence of a phase transition from rhombohedral to cubic structure. In their study, Fe2O3 appeared as Sr deficiency increased, which was also observed in this study. Therefore, the structural analysis in this study indicates that the LSCF surface undergoes a phase transition under Cr contamination conditions as a result of a Sr deficiency. This is in good agreement with the chemical analysis performed in this study. For ABO3 perovksite materials, it is generally believed that B site transition metals play a critical role for catalytic activity.24, 69, 87, 88 Kilner et al.31 investigated the dependence of surface exchange coefficients (Ko) on the composition of La0.8Sr0.2Mn1 yCoyO3 and found that Ko increased with increasing Co concentration. Takeda et al.44 measured the cathodic polarization of La0.7Sr0.3MO3 (M=Cr, Mn, Fe, Co) and found that La0.7Sr0.3CoO3 exhibited the lowest polarization at 800 oC in air. In addition, Kan et al.54 studied the effect of catalyst impregnation on the catalytic activity for oxygen exchange. They found oxygen exchange reaction of La0.8Sr0.2MnO3(LSM) was improved with the addition of Co on the surfaces via wet impregnation. These results in literatures indicate that Co is a catalytically active element for oxygen reduction reaction. Therefore, reduction in Co c oncentration on the surface of LSCF is likely to degrade oxygen exchange reaction. Furthermore, the collected data of the surface exchange coefficients and the oxygen tracer diffusion coefficients of various metal oxides evidences that the electronic speci es on the surface are essential for oxygen exchange reaction. The formation of near insulator Fe2O3,107 and less conductive Sr deficient LSCF,108 reduces

PAGE 101

101 the concentration of electronic species an d so contributes to degradation of oxygen exchange reaction at the surface. Consequently, the formation of Sr and transition metal deficient layers on LSCF surfaces as a result of the formation of SrCrO4 particle is likely to cause a significant degradation in oxygen reduction reaction. In our group, the impact of Cr contamination on the oxygen reduction reaction of LSCF was examined using electrical conductivity relaxation (ECR), AC impedance, and temperature programmed isotopic exchange.109 They consistently showed degradation of oxygen exchange reaction at the surface as a result of Cr vapor depo sition. The same chemical variation of LSCF surface as observed in this study was found after ECR test at 800 oC. The result confirms that the formation of Sr and transition metal deficient phases reduces the catalytic activity for oxygen reduction. 6. 3. 3 Dependence of A site ions on Cr V apor D eposition LCCF, LSCF, LBCF, and SCF, which have different A site but identical B site ions, were heat treated at 800oC for 50 hours in the presence of Crofer22APU. Figure 6 5 shows representative surfaces of these s amples after heat treatment. All samples present significant morphological changes after heat treatment. The formation of secondary phase decreased as the separation distance increased due to reduced mass transport of Cr vapor species. Table 6 -5 shows the elemental atomic percents of newly formed particles obtained using EDS. The concentration of alkaline earth metal ions was substantially higher than the other cations compared to nonheat treated samples, and considerable amount s of Cr were found. This res ult provides that the secondary phases are Ca -Cr -Ox, Ba -Cr -Ox, and Sr Cr -Ox for LCCF, LBCF, and SCF respectively, in which alkaline earth metal oxides worked as a nucleation agent. It is likely that Cr vapor

PAGE 102

102 deposition on these materials is governed by the same Cr vapor deposition mechanism as for LSCF. Figure 6 6 shows the influence of temperature. Compared to heat treated surfaces at 800oC, these surfaces exhibit reduced formation of secondary phases when heat treated at 600oC. When heat treatment was co nducted at 400oC, a dramatic drop in Cr vapor deposition on metal oxides was observed. Since Cr vapor pressures increase with increasing temperatures, this is primarily due to the reduced Cr vapor pressure at lower temperatures. The effect of A site cations on the reactivity to Cr vapor species was examined by comparing the amount of secondary phases. Previously, Yokokawa et al.110 provided the formation enthalpy difference between metal chromates and metal ox ides. They showed a strong tendency of basic oxides to form stable metal chromates, in which BaO has a stronger tendency than SrO, which has a stronger tendency than CaO. This indicates that the tendency to form a metal chromate increases with increasing b asicity of metal oxides. LBCF shows higher areal density of secondary phase than LSCF at 800oC (Figure 6 -5) and 600oC (Figure 6 6). However, it is difficult to tell whether LSCF is more reactive than LCCF based on the formation of secondary phase. At 800oC LCCF exhibited fewer particles with higher growth rate while at 600oC showed similar areal density of particles after heat treatment. In addition, it appears that the reactivity increases with increasing molar concentration of alkaline earth metal ions In Figure 6 -5 and Figure 6 6, SCF presents a substantially higher amount of secondary phase compared to LCCF, LSCF and LBCF. In part, this can be explained by increased amount of Sr, which has a higher basicity

PAGE 103

103 than La. However higher reactivity of SCF i s also attributed to higher concentration of unstable transition metal species. This will be explained in following section. 6. 3.4 Dependence of B site i ons on Cr V apor D eposition The effect of B site transition metals on the reactivity to Cr vapor was i nvestigated by comparing Cr vapor deposition behavior on LSCF and LSM. They have the same A site composition but different B site transition metals. LSM was placed on the Crofer22APU sheet and heat treated 800oC for 50 hours. Figure 6 7a and 6 -7b show the surfaces of non -heat treated and post -heat treated LSM, respectively. After heat treatment, any noticeable change, e.g. the formation of particles, was not found on the LSM surface. This indicates that LSM does not react with Cr vapor species at 800oC. Th e formation of secondary phase was observed when LSM was heat treated at higher temperatures. Figure 6 8a shows the SEM micrograph of LSM after heat treatment at 1050oC for 50 hours. It shows the formation of facetted particles on the surface. EDS spectra for nonheat treated LSM and facetted particles are given in Figure 6 8b and 6 8c, respectively. It reveals that the secondary phase is a Cr Mn -Ox. Multiple particles were analyzed but the formation of Sr Cr -Ox was not found on LSM. The temperature dependence of Cr vapor deposition on LSM is in good agreement with the literature. Tucker at al. 111 tested Cr vapor deposition on La0.65Sr0.3MnO3 and found that Cr vapor deposition did not occur until 1000oC. The observation of Cr vapor deposition on LSM without application of cathodic or anodic polarization supports t h e theory that the nature of Cr vapor deposition on LSM at 1050oC is a chemical process, in which Mn acts as a nucleation agent. It was previously reported that the crystal structure of Cr Mn -Ox, which was formed as a result of Cr vapor deposition on LSM, was (CrMn)3O4.14, 126 The most prominent

PAGE 104

104 form of (CrMn)3O4 is cubic MnCr2O4 spinel in which Mn2+ and Cr3+ occupy the tetrahedral and octahedral sites, respectively.112, 113 Consequently, it is likely that Mn2+ acts as a nucleation agent for the formation of (CrMn)3O4 as proposed by Jiang et al.114 In order to form Mn2+ species, LSM needs to undergo charge disproportion of Mn3+ according to115 2Mn3+ 2+ +Mn4+ (6 -3) Since the formation of (CrMn)3O4 does not occur until 1050oC, charge disproportion of Mn3+ is likely to be inhibited below this temperature. Note that the formation of SrCrO4 on LSCF was found at 600 and 800oC, i n which Sr acted as a nucleation agent. Also, other perovskite materials containing Co and Fe in B site exhibited similar Cr vapor deposition behavior. By contrast, no formation of SrCrO4 was detected on LSM at 800oC, and Cr vapor deposition occurred at much higher temperatures. Since the chemical composition of A site concentration is the same between LSCF and LSM, the significant difference in Cr vapor deposition behavior is attributed to B site transition metals. While LSCF exhibits a substantial oxygen deficiency in air ,40, 49 LSM shows oxygen excess. Such high oxygen vacancy concentration in LSCF is due to the instability of Co4+ and Fe4+, which are generated by a partial substitution of La for Sr. They are unstable species at high temperatures8 and thus, LSCF achieves a more stable state by reducing them through oxygen vacancy formation. In contrast, the valence stability of Mn4+ is relatively higher than that of Mn3+,116, 117 and therefore, LSM does not exhibit oxygen deficiency. Even though oxygen vacancies improve the oxygen reduction reaction on the surface and oxygen diffusion in the lattice,23 their presence weakens the stability of the structure by

PAGE 105

105 removing the shielding effect between like atoms. Previously Wang et al.67 demonstrated t he effect of oxygen vacancies on bonding strength experimentally They found that the elastic modulus of ceria decreases drastically as the oxygen vacancy concentration increases. This proves that the bonding strength becomes weakened i n accordance with the formation of oxygen vacancies. At elevated temperatures, LSCF has substantial oxygen vacancies in the lattice45 and thus Sr is likely to be loosely bound in the lattice. Then enriched Sr in surface region can be precipitated into Sr -O and react to Cr vapor deposition. In contrast, Sr in LSM, which is not oxygen deficient in air, is tightly bound in the lattice and so inert in Cr contaminated environment s In the previous section, SCF exhibited substantially higher Cr vapor deposition than LCCF, LSCF, and LBCF. The difference between SCF and other materials is the concen tration of alkaline earth metal ions in A site. As explained, increase in alkaline earth metal ions increases the basicity of A site. But it also raises the concentration of unstable Co4+ and Fe4+ species. As a result, SCF has more oxygen vacancies in the lattice and becomes more reactive to Cr vapor species. Matsuzaki et al.118 compared Cr vapor deposition behaviors of LSCF and LSM cathode based cells. They found that Cr vapor deposition occurred on the electrode surface of LSCF during the cell test at 800 oC. Cr contamination was intense on elect rode surface near the interconnect while it decreased at the three phase boundary (TPB) of air, electrode and electrolyte. In contrast, in the case of LSM cathode based cell, Cr vapor deposition was found in TPB region while it was significantly reduced on the electrode surface. Such different Cr contamination behaviors indicate that the nature of Cr vapor deposition process on LSCF and LSM during cell operation is not identical.

PAGE 106

106 Cr vapor deposition on electrodes can occur either by chemical or electrochemi cal reaction. This study shows that Cr vapor species deposited on LSCF at 800 oC in the absence of polarization currents, which is an evidence of chemical reaction. It was found that Cr vapor deposition via chemical reaction decreased with increasing separ ation distance between metallic interconnect and electrode surface due to the reduced mass transport of Cr vapor species. This well justifies the observed Cr vapor deposition on LSCF in literature. Thus, it is likely that Cr vapor deposition on LSCF electr ode during the cell test is chemical reaction. In contrast, this study shows no Cr vapor deposition on LSM at 800 oC in the absence of polarization currents. This indicated that Cr vapor deposition on TPB of LSM during the cell test at 800 oC is not chemic al reaction. Previously, Ricoult119 found that cathodic polarization during the single cell oper ation caused strong enrichment of Mn2+ at the TPB of LSM and YSZ at 700 oC. As mentioned, Mn2+ is likely to act as a nucleation agent for the formation of (CrMn)3O4 particles. Consequently, the formation of Mn2+ due to cathodic polarization leads to prefer ential deposition of Cr vapor species along the TPB region during the cell operation. This is in agreement with the observation in literature. Therefore, the dominant reaction process of Cr vapor deposition on LSM at 800 oC is likely to be the electrochemi cal reaction. At last, it was observed that the grain surface of LSCF exhibited a Sr deficiency in Cr contaminated environment because enriched Sr on LSCF diffused from grain to grain boundary regions, forming SrCrO4. Two possible diffusion mechanisms are surface diffusion and evaporation -condensation. Recently Bishop et al.120 reported mass loss of La0.6Sr0.4Co0.2Fe0.8O3 at 800 oC in air using thermogravimetry. Mass reduction

PAGE 107

107 continued over 200 hrs, which is a substantially longer time scale compared with oxygen diffusion and surface reaction. Such a mass loss phenomenon was difficult to be justified by Sr enrichment to surface regions or the formation of Ruddlesden-Popper (RP) phases wh ich were frequently found from other perovskite materials .68, 71 One can deduce that Sr -O on LSCF is vaporized at oxidizing conditions at 800 oC This result puts more weight on evaporation -condensation for the diff usion mechanism of Sr -O on LSCF surface. 6. 4. Conclusions This study improves the fundamental understanding of the reaction between SOFC cathode materials and Cr vapor species. It turns out Cr vapor deposition causes substantial chemical and structural ch anges of LSCF surface layers above 600oC, which involves the formation of SrCrO4 particles, and Sr and transition metal deficient matri ces This may lead to a significant degradation in surface catalytic activity of LSCF. LCCF, LBCF, and SCF exhibited simi lar Cr vapor deposition behavior, in which alkaline earth metal ions in the A site acted as nucleation agent s In contrast, LSM presented considerably different behavior in Cr vapor deposition. Mn2+, which is formed as a result of charge disproportion of M n3+, seems to act as a nucleation agent on LSM. Consequently, chemical reaction between LSM and Cr vapor species occurred at higher temperatures (1050oC). The results imply that transition metals in the B site, rather than the basicity of A site ions, prim arily determines the reactivity to Cr vapor specie s It is proposed oxygen vacancy formation due to instability of Co4+ and Fe4+ increases the repulsive forces between cations, and allows the formation of nucleation agent at lower temperatures. Therefore b elow 1000oC the dominant Cr vapor deposition mechanism

PAGE 108

108 on LSCF and LSM during the cell operation is likely to be chemical and electrochemical reactions respectively.

PAGE 109

109 Figure 6 1. Schematic of experimental design for heat treatment. Figure 6 2 XRD patterns of tested perovskite materials

PAGE 110

110 Figure 6 3 Surfaces of LSCF (a) before and (b d) after heat treatment in the presence of Crofer22APU at 800oC 50 hrs. The leading edge distance s between LSCF and Crofer22APU sheets were (b) di rect contact, (c) 0.1 cm, and (d) 0.5 cm. All images are the same scale.

PAGE 111

111 Figure 6 4 (a) XRD patterns of fresh and heat treated LSCF on Crofer22APU sheet at 800oC for 50hours. (b ~ d) E nlarged patterns for better comparison. is the XRD pattern measur ed using a low angle mode.

PAGE 112

112 Figure 6 5. Surfaces of LCCF, LBCF, LSCF and SCF after heat treatment in the presence of Crofer22APU at 800oC for 50 hours. The leading edge distances between samples and Crofer22APU sheet were 0.1 cm and 0.5 cm for the left and right hand sides, respectively

PAGE 113

113 Figure 6 6. Surfaces of tested after heat treatment in the presence of Crofer22APU at 600oC and 400oC (Leading edge distance 0.1 cm) Figure 6 7. SEM micrographs of LSM surfaces of (a) before and (b) after heat tr eatment at 800oC for 50 hours in the presence of Crofer22APU ( Leading edge distance 0.1 cm)

PAGE 114

114 Figure 6 8. (a) SEM micrograph of LSM after heat treatment in the presence of Crofer22APU at 1050oC for 50 hours, and EDS spectra for (b) the flat region and (c) particle (Leading edge distance 0.1 cm)

PAGE 115

115 Table 6-1. Summary of tested ABO3 perovskite materials Composition Abbreviati on Preparation method La0.6Ca0.4Co0.2Fe0.8O3 La0.6Sr0.4Co0.2Fe0.8O3 La0.6Ba0.4Co0.2Fe0.8O3 SrCo0.2Fe0.8O3 La0.6Sr0.4MnO3 LCCF LSCF LBCF SCF LSM Solid State synthesis Praxair specialty Solid State synthesis Solid State synthesis Solid State synthesis Table 6-2. Chemical com position of Crofer22APU Cr Fe Mn Si Cu La 20~24 Bal. 0.3~0. 8 0.5 0.5 0.04~0.2 Table 6-3. Elemental atomic percents (%) of LSCF after heat treatment at 800oC for 50 hrs in the presence of Crof er22APU obtained using EDS Table 6-4. AES atomic per cents of LSCF after heat treatment(HT) at 800oC for 50 hours The enrichment factors are given in parenthesis. A flat surface region was selected for the analysis. La Sr Co Fe O Cr Non-heat treated 12 9 4 14 61 Post-heat treatment Flat surface17 8 4 20 51 Particles 5 16 1 5 58 15 La Sr Co Fe O Non-heat treated 30 12 5 11 42 HT in the presence of Crofer22APU* 38 (1.2)8 (0.6) 2 (0.4) 8 (0.7) 44 HT in the absence of Crofer22APU 28 (0.9)22 (1.8)0 (0) 7 (0.6) 43

PAGE 116

116 Table 6-5. Elemental atom ic percents of non-heat tr eated and heat treated samples* LCCF La Ca Co Fe Cr Non-heat treated 12 8 4 16 Post heat treatment 1 18 1 2 19 LBCF La Ba Co Fe Cr Non-heat treated 17 11 4 21 Post heat treatment 1 16 1 4 14 SCF Sr Co Fe Cr Non-heat treated 27 6 25 Post heat treatment 21 3 12 12 *The newly formed particles after heat treat ment were selected for the analysis.

PAGE 117

117 CHAPTER 7 DEGRADATION MECHANIS M AND THEIR IMPACT O N OXYGEN REDUCTION KINETICS OF LSCF 7.1 Introduction The advantages of using Fe -Cr metal alloys as interconnects for solid oxide fuel cell (SOFC) stacks i nclude reduction in manufacturing costs, good machinability and enhanced thermo mechanical stability .109, 110 However, Cr vapor species generated from the Cr scales of these alloys are known to react with the cathod e, forming a less catalytically active phase .40, 104, 121 The long -term stability is an important requirement for the commercialization of SOFC s Thus, there is a need to understand cathode performance in Cr contami nated environments over long time periods. In this study, the influence of Cr contamination on the oxygen reduction reaction of La0.6Sr0.4Co0.2Fe0.8O3 (LSCF) was examined using electrical conductivity relaxation (ECR),48, 49, 52 AC impedance spectroscopy23, 80, 91 and temperature programmed isotopic exchange (TPX).54, 122 The impact of Cr vapor deposition on surface morpholog y and compo sition of LSCF was clearly identified from post ECR characterization 7.2 Experimental. 7.2 1 AC I mpedance A symmetric cell, which consists of LSCF/GDC/LSCF, was prepared for the impedance measurement. A dense Gd0.1Ce0.9O2 (GDC, Anan Kasei, Japan) electrolyte was made by sintering a pressed pellet at 1450 oC for 4 hours. A LSCF ink was prepared by mixing LSCF powder with texanol based polymer vehicle (ESL electro science, USA) using Thinky mixer (Thinky, Japan). The ink was symmetrically painted on both sides of the electrolyte, each having a same electrode area of 0.28 cm2. Pt wires were connected as a current collector and LSCF ink was applied for better

PAGE 118

118 contact. After that, the cell was sintered at 1100 oC for 1 h our w ith heating and cooling rate of 5 oC/ mins. AC impedance of the symmetric cells was measured at 800 oC in the presence and absence of K -type thermocouple using a Solartron SI 1260 frequency response analyzer. The amplitude of the excitation voltage was 50 mV and the AC frequency range was from 0.1 to 105 Hz. Measurement was repeated every 1 hour using ZPlot software (Scribner). The experimental data was fit using the equivalent circuit shown in Figure 1 using ZView software (Scribner) in order to de-convol ute the impedance response and extract the electrode resistance. It consists of a series of resistors (R) and constant phase elements (CPE). In the figure, RE is the electrolyte ohmic resistance, and RHF and RLF refer to high and low frequency resistance r espectively. The electrode area specific resis tance (ASRp) is obtained by 2 Area ) R (R ASRElectrode LF HF p (7 1 ) 7.2 2 Electrical C onductivity R elaxation A dense LSCF sample was prepared for analysis It was prepared by pressing La0.6Sr0.4Co0.2Fe0.8O3 powder (Praxair Specialty Ceramics, USA) uni axially and sintering a pellet at 1400oC for 4 hours in air. The density of the sintered body measured by the Archimedes method was ~ 97% of the theoretical value (6.36g/cm3). The sintered body was cut and t hen polished with a series of SiC papers and diamond After polishing, samples were cleaned with DI water and acetone before ECR measurement The sample dimension was approximately 0.1x2x20 mm3.

PAGE 119

119 A bar of LSCF was positioned on the alumina tube in the middle of the reactor and a K -type thermocouple (OMEGA, Part# KMQSS -040E -12) was placed below the sample for accurate measurement of temperature. A N2 and O2 gas mixture was fed into the reactor through a mass flow controller (MKS & Edwards) and the total flow rate was fixed at 300sccm. The electrical conductivity was measured via the four point probe method using a Lock -in amplifier (Stanford, SR830) automated with Labview software. The sample was heated to 800oC at a rate of 5oC/ min in air. Before the ECR measurement, the feed stream was switched to 3% O2. Once it reached steady state, the feed stream was switched via four way switching valve. The oxygen partial pressure (pO2) step was programmed to increase by the following steps : 3 7.2 3 Temperature P rogrammed I sotopic E xchange (TPX) A schematic of the TPX test system is seen in Figure 7 -2 a. The custom made quartz reactor (i.d. = 4 mm) was used and a powder sample was placed atop the quartz frit in the middl e of the reactor. The flow rate of the feed gas was controlled by mass flow controllers (MKS & Alicat) with a total flow rate of 25sccm. The reactor effluent was sampled by a quadrupole mass spectrometer (MS, Extrel QMS) in which Masses 16 (16O), 18 (18O) 28 (N2), 32 (16O2), 34 (16O18O), 36 (18O2) and 40 (Ar, inert tracer) were continuously monitored. The K type thermocouple, which was shielded in a quartz capillary tube, was placed above the sample to monitor temperatures. The heating schedule and condi tion are briefly summarized in Figure 7 -2 b. Before a TPX run began, the sample was pre-heat treated at 650oC in 1% 16O2 for 10 mins and cooled down to room temperature. Once the MS signal reached a steady state, the feed

PAGE 120

120 gas was switched to 0.25% 18O2 (Cam bridge Isotope, 95% pure), followed by sample heating at a linear rate of 30oC/mins. All gas mixtures were balanced by helium. 7.2 .4 Characterization The microstructure and element distribution before and after ECR measurement was characterized by scanni ng electron microscopy (SEM, JEOL JSM6400) combined with energy dispersive spectroscopy (EDS, Oxford, UK). ZAF correction was used for semi quantitative analysis for EDS. In addition, surface chemical composition of LSCF before and after ECR test was meas ured using Auger electron spectroscopy (AES, AES Perkin -Elmer PHI 660 ) The accelerating voltage was 10 kV and i t scanned the energy ranges between 50 to 2050 eV, which covered the characteristic peaks of LaMNN (625 eV), SrLMM (1649 eV), Co (775 eV), FeLMM (703 eV), and OKLL (510 eV) These characteristic peaks were selected in order to minimize interference with ea ch other. Obtained AES spectr a were differentiated, and elemental atomic percents were calculated using AugerScan 3.2.0 software (RBD Instrument ). T he specific surface area of powders was measured by BET an alysis (Quantachrome NOVA 1200) 7.3 Results and Discussion 7.3 1 AC I mpedance Figure 7 3 shows the impedance spectra of a LSCF /GDC/LSCF symmetric cell at 800oC. There are approximately two semi -circles, which consists of high frequency (HF) and low frequency (LF) components as seen in the figure. As the measurement was conducted at 800oC, the size of semi -circles changed with aging For easier comparison the HF intercepts of the semi -circles were subtracted in Figure 7 -3 After 190 hours, t he size of the HF semi -circle was found to increase, while that of the LF semi -circle remained roughly constant. HF and LF components were de -convoluted by fitting the

PAGE 121

121 equivalent circuit (Figure 7 -1 ) to experimental data, and the related a rea specific resistance (ASR) was plotted as a function of aging time in Figure 7 4 a. This shows that the electrode ASR, which is the sum of ASRHF and ASRLF, increased relatively faster for the first 50 hours due to the incr ease in ASRHF. T hen the degradation rate became slightly reduced, but the electrode ASR increased continuously. The overall electrode ASR increased approximately 57 % after 190 hours. In contrast, if the K -type thermocouple inside the reactor was removed du ring impedance measurement the symmetric cell exhibited roughly constant ASRHF and ASRLF for 190 hours at 800oC as shown in Figure 4 b. Since the temperatures were maintained by the same furnace temperature controller, which was located outside of the quar tz react or the temperatures were essentially the same in both cases. The temperature differences between inside and outside of the reactor were less than 5 oC. Each semi -circle corresponds to different reaction mechanism s Previously Esquirol et al.91 and Sahibzada et al.80 measured the impedance of LSCF/GDC/LSCF symmetric cell and observed the formation of two semi -circles at 800 oC. They attributed the HF and LF components to the surface reaction and gas phase diffusion, respectively. In this study, the ASRHF was found to increas e while ASRLF remained roughly constant at 800 oC when the symmetric cell was tested in the presence of Ktype TC i n the reactor. This indicates that resistance for surface reaction increased due to K type TC. 7.3 2 Electrical C onductivity R elaxation Fig ure 7 5 a shows an electrical conductivity relaxation curve after switching the pO2 from 43 to 67% at 800oC which is normalized by initial and final equilibrium

PAGE 122

122 electrical conductivities It shows that the kinetic model fits well to experimental data. Figu re 7 5 b shows the variation of relaxation curves with aging when the pO2 was switched from 43 to 67 %. As the ECR run was repeated over time at isothermal condition the relaxation time, which was required to reach a new equilibrium, gradually increased. Th is indicates degradation of oxygen transport on the surface or in the bulk. Chemical diffusion (Dchem) and chemical exchange (Kchem) coefficients in the pO2 range from 5 to 67% at 800oC were obtained by fitting kinetic model to relaxation curves. They were plotted in log scale as a function of aging time in Figure 7 6 It shows that the degradation rate of Dchem was not significant over the entire pO2 range, while Kchem exhibited relatively fast reduction up to 400 hours. Note that Kchem in log scale linearly dropped. This means that the degradation rate of Kchem was relatively faster at the beginning and decreased at longer time. Since Kchem represents the apparent rate for oxygen exchange reaction at the surface, the results indicate degradation of surfa ce layers during the measurement Since K -type thermocouple (TC) was located under the LSCF bar, it is likely the source of the Cr contamination on LSCF surface. Post ECR characterization clearly showed how it degraded the surface morphology and composition. The relationship between electrode ASR (ASRp) and surface exchange coefficient (Ko) is 100, 102 o P 2 2 oC ASR F z T R K ( 7 -2 ) where R is the gas constant, T is the absolute temperature, F is faraday constant, n is the number of electron transferred during the reaction, and Co is oxygen concentration. Furthermore, Ko can be converted into KChem by94

PAGE 123

123 o chemK K (7 3 ) o 2lnC lnpO 2 1 ). Thus, the decrease in KChem from ECR is consistent with the increase in ASRp from AC impedance under essentially identical conditions. 7.3 3 Temperature P rogrammed I sotopic E xchange Figure 7 7 sh ows the TPX results of fresh LSCF samples. Each oxygen concentration, C(16O2), C(18O2) and C(16O18O), was normalized by input oxygen concentration, C(18O2). Bulk oxygen evolution was observed above 550oC. Previously Stevenson et al.45 and Tai et al.34 measured the oxygen nonstoichiometry of La1 xSrxCoyFe1 yO3 and showed that oxygen nonstoichiometry increased above 600oC for LSCF having same chemical composition to this study. The difference in temperature s between this study and previous works is attributed to sample condition. While TPX tested a powder sample, a dense sample was used in literature. A powder sample possesses substantially higher surface area than dense sample, and thus it can provide more surface sensitive behavior. Since the bonding strength o f oxygen in near surface region is weaker than oxygen in the bulk due to broken bonding and more point defects,120 it is likely that oxygen in near surface region evolves more rapidly at lo wer temperatures than oxygen in bulk. Even though overall oxygen concentration does not change until 550oC, TPX reveals that the oxygen exchange reaction is activated at substantially lower temperatures, thereby concentration of 18O2 decreases and that of 16O2 increases. Since oxygen diffusivity in the lattice is insignificant around 300oC, the oxygen

PAGE 124

124 exchange reaction mostly occurs on the surface or near the surface region. In Figure 7 7, TPX results of two samples are compared. They are exactly the same except for the sample weight In the figure, the LSCF sample with a higher total surface area exhibits lower onset temperature s for oxygen exchange. T his is because the total amount of oxygen exchanges per unit time is proportional to total surface area and oxygen exchange kinetic s Therefore, in order for direct comparison of oxygen exchange kinetic s between different samples, the total surface area must be the same. For the test of Cr contaminat ed sample, s amples were prepared by annealing LSCF powder at 800oC for 400 hours with and without Crofer22APU present. The specific surface area of the powders is given in Table 7 -1 Heat treated LSCF powders exhibited reduced specific surface area due to coarsening than fresh LSCF and so s ample sizes were selected to provide the same total surface area. TPX result s of these samples are given in Figure 7 8. While they showed identical temperature dependency for overall oxygen concentration changes, the oxygen exchange behavior was considerably changed. The onset tem perature for the heat treated LSCF sample is higher than the fresh LSCF sample, and it further increases when heat treated in the presence of Crofer22APU. A higher onset temperature indicates decreased oxygen exchange reaction at the surface Therefore th e TPX result implies that heat treating LSCF at 800oC degrade d the surface reactivity and the extent of degradation is increased by Cr contamination. This result is in agreement with the r esults of ECR and AC impedance and further confirm s that LSCF cathod e degradation is due to decreased surface oxygen exchange kinetics.

PAGE 125

125 It is reported that the chemical composition of the LSCF surface is different from that of bulk at high temperatures Viitanen et al.61 observed the absence of Co and Fe in the outermost atomic layer from contamination-free LSCF after an oxygen permeation experiment at 900oC Bucher et al.59 found Sr enrichment on the surface after meas uring chemical exchange coefficient of LSCF at 600700oC. Furthermore, Furthermore, Oh et al.84 showed that enriched Sr in LSCF precipitated into submicron sized Sr-O particles under an oxidizing condition in temperature range of 600 -900oC They also observed reduced concentration of transition metals on the surface in accordance with Sr -O precipitation at 800oC Similarly, enrichment of A site cation is also found in other perovskite materials such as La1 xSrxMnO3,123 BaTiO3,70 and SrTiO3 6466 at elevated temperatures. As a result concentration of B site transition metals became reduced in the surface region. Rahmati et al.64 observed t he formation of Sr-O island s on the surface of SrTiO3. They found that the areal density of surface islands under re oxidized conditions (1200oC for 30 hours) increased with increasing surface energies of underlying grains This indicates that formation of Sr -O islands is due to the stabilization of surface. Szot. et al68 investigat ed the chemical variations of several ABO3 perovskite materials at elevated temperatures (5001000oC) and consistently found formation of AO layers on the surface which is known as Ruddlesden-Popper Phase (RP).71 They propose that the formation of RP phase is stable in oxidizing condition. When LSCF powder was heat treated at 800oC for 400 hours in the absence of Crofer22APU sheets, the surface is likely to have Sr enriched and transition metal deficient phase. S ince B site transition metals play a critical role for catalytic activity of ABO3 perovskite materials26, 6668 and Sr -O is a electrically insulator74,

PAGE 126

126 such Sr enrichment and the formation of Sr -O precipitates on the surface degrade s the surface catalytic activity. Thus, it is likely the cause for the increase in the onset temperature when LSCF is heat treated without Crofer22APU present The reason for higher onset temperature in Cr co ntaminated LSCF will be explained in following section. 7.3 .4 Post ECR C haracterization Figure 7 9 a and 7 9 b are SEM micrographs of LSCF before and after ECR test respectively. While the surface before ECR test ing is flat and unblemished, it becomes rug ged and the formation of facetted particles is found along grain boundary surfaces after ECR testing Elemental atomic percents for a newly formed particle on grain boundaries as well as grain surfaces between the particles were obtained using energy dispersive spectroscopy (EDS). As table 7 2 shows, the grain surface did not present noticeable difference in atomic percents compared to the fresh sample. By contrast, the particle exhibited substantially increased concentration of Sr and Cr, indicating that t he particles are Sr -Cr -O. This is in good agreement with literatures.40, 104 Since K type thermocouple (OMEGA) uses 304 stainless steel, which contains around 8 wt% Cr, as a sheath,124 it is likely that Cr vapor species is formed and deposits on the LSCF electrode during ECR measurement. The kinetic measurements in this study consistently show the degradation of surfa ce catalytic activities Therefore, the c ompositional variation of surface layers was investigated using Auger electron spectroscopy (AES). The grain surface between Sr Cr -O particles was selected for the analysis The elemental atomic percent which was o btained using Eq. 11, is given in Table 7 -3 m m a a/S I /S I X (7 4 )

PAGE 127

127 where I is the peak intensity and S is the relative elemental sensitivity factor for the Auger line used. C ompared to the fresh sample, atomic percent of cations varied substantially after heat treatment. It shows Sr deficiency on LSCF surface rather than Sr enrichment. In addition, the transition metals concentrations became halved. In literatures, Sr enrichment at elevated temperatures was evident on the surface of LSCF.59, 60 For the comparison, the elemental atomic percents of the LSCF surface, which was heat treated at 80 0oC in the absence of Cr contamination were added in Table 7 3 .103 Sr concentration became doubled and transition metal concentration decreased after heat treatm ent in the absence of Cr contamination. In Table 7 3 the major difference between the two samples arises from A site cation while B site transition metals show same trends. This implies that Sr -O precipitate on LSCF surfaces acts as a nucleation agent for the formation of Sr -Cr -O particles As Sr-Cr -O particle is preferentially formed along the grain boundaries, it is likely that Sr -O on grain surface diffuses to grain boundary region either by surfa ce diffusion or by evaporationcondensation As a consequ ence, grain surface of LSCF exhibit s Sr deficiency as observed in this study. The heterogeneous catalytic activity is primarily determined by interaction between surface and surrounding gases, and thus surface composition and structure is an important fac tor affecting the activity. Even though the rate determining step of oxygen reduction reaction is not well understood, it is no doubt that B site transition metals play a critical role for oxygen reduction reaction.26, 66 68 Previously De Souza et al.42 measured surface exchange coefficient of La1 xSrxMn1 yCoyO3 and foun d that the activation energy of Ko decreased with increasing Co concentration. And Teraoka et

PAGE 128

128 al.37 found that the rate of the oxygen permeation of La1 xSrxCoyFe1 yO3 increased with increasing Co. Kremenic et al. reported the oxygen adsorption properties of LaMO3 (M=Cr, Mn, Co, Fe, Ni) depended on the transition metals, in which Co and Ma exhibited the highest oxygen adsorption. In addition, Kan et al.54 studied the effect of catalyst impregnation on the catalytic activity for oxygen exchange. They found oxygen exchange reaction of La0.8Sr0.2MnO3(LSM) was improved with the addition of Co on the surfaces via wet impregnation. These results in literatures indicate that Co is a catalytically active element for oxygen reduction reaction. Therefore, the formation of Sr and transition metal deficient layers on the grain surface of LSCF can bring a significant degradation in oxygen reduction reaction and eventually cell performance. In the TPX result, a heat treated LSCF sample in the presence of Crofer22APU exhibited higher onset temperatures than heat treated LSCF in the absence of Crofer22APU. This indicates the formation of Sr and transition metal deficient surface more dramatically reduces the surface catalytic activity than Sr enriched surface. It should be stressed that the Cr contamination source for ECR and AC impedance test was a K type TC, which contains Cr alloys as a sheath. In many studies the TC is located near the sample in order for accurate measurement of temperatures. However, this may lead to unwanted Cr contamination on tested sam ples during the measurement. In order to prevent it, the TC must be shielded by other materials such as Zirconia or quartz. 7.4 Conclusions This study improved degradation mechanism and their impact on the oxygen reduction reaction of LSCF. T he temperatur e programmed isotopic exchange (TPX) result showed that heat treating LSCF at 800oC degraded the surface reactivity due to

PAGE 129

129 Sr enrichment and the extent of degradation was further increased by Cr contamination. Electrical conductivity relaxation and AC impe dance study consistently showed that oxygen reduction reaction at the surface was degraded as a result of Cr vapor deposition Post ECR characterization revealed that Cr vapor deposition deteriorated surface properties by forming a catalytically inactive s urface layer, which consists of Sr Cr -O particles and Sr deficient surface layers. The surface of LSCF also exhibited reduced concentration of transition metals. It seems that enriched Sr on LSCF surface acts as a nucleation agent for the formation of Sr -C r-O particles, and as a result LSCF surface exhibit s Sr deficiency.

PAGE 130

130 Figure 7 1 Equivalent circuit to fit the measured impedance data Figure 7 2 (a) A schematic of the experimental set up used for temperature programmed isotopic ex change and (b) the preheat treatment and measurement conditions

PAGE 131

131 Figure 7 3 AC Impedance spectra of LSCF/GDC/LDCF in air at 80 0oC Figure 7 4 Area specific resistance of LSCF at 800oC in air as a function of aging time (a) in the (a) presence and (b) absence of K type thermocouple in the reactor

PAGE 132

132 Figure 7 5 (a) Normalized electrical conductivity relaxation curve after switching oxygen partial pressure from 43 to 67% at t =0 and (b) variation of relaxation curves with aging at 800oC. Figure 7 6 Degradation of Kchem and Dchem in Cr contaminated environment at 800oC

PAGE 133

133 Figure 7 7 Effect of total surface area on oxygen exchange profile s which are measured by TPX. SA refers to the surface area.

PAGE 134

134 Figure 7 8 TPX result of h eat treat ed LSCF at 800oC for 400 hours with and without exposure to Crofer22APU in comparison to fresh LSCF.

PAGE 135

135 Figure 7 9 SEM micrographs of polished LSCF surfaces (a) before and (b) after ECR at 800oC for 400 h ours.

PAGE 136

136 T able 7 1 Samples tested in TPX Heat treatment conditions Specific SA (m 2 /g) Sample size(mg) Total SA (m 2 ) Fresh LSCF LSCF LSCF w/ Crofer22APU 800oC / 400 hours 800 o C / 400 hours 6.53 4.38 3.97 15.3 22.8 25.2 0.1 0.1 0.1 SA is the surface area Table 7 2 Elemental atomic percents (%) of LSCF obtained using EDS Table 7 3 Elemental atomic percents (%) of LSCF after ECR test at 800oC for 400 hours obtained using AES The enrichment factors are given in parenthesis. A grain surface was selected for the analysis. La Sr Co Fe O Cr Fresh 12 9 4 14 61 Post ECR test at 800oC for 400 hours Grain surface 17 9 4 19 51 Particles 2 13 1 2 72 10 La Sr Co Fe O Fresh 30 12 5 11 42 Post ECR test at 800 o C for 400 hours* 38 (1) 0 (0) 2 (0.3) 8(0.6) 52 Heat treatment at 800 o C for 50 hours in the absence of Cr contamination. 28 (0.9) 22 ( 1.8) 0 (0) 7 (0.6) 43

PAGE 137

137 CHAPTER 8 VAPORIZATION OF DOPED ALKALINE EARTH MET AL OXIDE 8.1 Introduction D oped Sr has been found to segregate in the surface region at high temperatures.59, 60 Bucher el al.59 found Sr enrichment on the surfac e of LSCF using X ray photoelectron spectroscopy ( XPS ) after electrical conductivity relaxation measurement at 600-700oC. Similarly Simner et al.38 found Sr enrichment at the cathode electrolyte and cathode -current collector interfaces after testing a button cell at 750oC for 500 hours. This study observed that Sr in LSCF not only enriched to the surface but also precipitated into SrO under oxid izing condition between 600900oC. As a heterogeneous catalyst, the surface reaction is primarily affected by surface composition, and therefore such surface segregation of catalytically inactive elem ent can lead to degradation of cathode performance. Recently, Bishop et al.120 reported a slow mass reduction of a dense La0.6Sr0.4Co0.2Fe0.8O3 bar under isothermal condition s in air duri ng thermogravimetric measurement s The mass reduction was small, but it continued over 200 h ou rs, of which this time scale is substantially longer than that of oxygen diffusion or oxygen exchange reaction at the surface. Elemental segregation phenomena in the surface region may occur during long term heat treatments, but such a mechanism is not suitable for justifying mass reduction.71, 80 Previously, Mai et al.125 tested the effect of a Ce0.8Gd0.2O2 (GDC) interlayer between LSCF cathode and a Zr0.92Y0.08O2 (YSZ) electrolyte in order to prevent chemical reactions between the LSCF and YSZ d uring sintering. They found that SrZrO3 was formed at the YSZ interface when the GDC interlayer was porous while no Sr diffusion occurred when the interlayer was dense.

PAGE 138

138 This indicates that Sr was likely to transport from the LSCF cathode to the YSZ electro lyte via vapor phase diffusion. The objective of this study was to improve the understanding of the effect of doped alkaline earth metal ions on surface structure at high temperature. B ased on observations in literatures it was hypothesized that doped Sr was vaporized in the vicinity of the surface region and a simple heat treatment experiment was conducted to prove the hypothesis. 8. 2 Experimental LSM was exposed to a stagnant dry atmosphere at 800oC for 50 hours Samples were heat treated in the presence of Crofer22APU sheets and BaO (or LSCF powder) as described in Figure 8 -1 a Also, LSM was heat treated with exposure to LSCF powder and Crofer22APU as described in Figure 8 1b in order to reduce the separation distance between the LSM surface and the L SCF powder 8. 3 Result s and D iscussion Figure 8 2 shows the SEM micrographs of as -polished and heat treated LSM. Heat treatment was conducted in the same conditions as for the LSCF. In contrast to the result s for LSCF, LSM shows no formation of secondary phase after heat treatment. This means that LSM must not react with Cr vapor species at 800oC. However, when LSM was heat treated at higher temperatures, the formation of secondary phases was found. Figure 8 3 a shows the SEM micrographs of L SM after heat treatment at 1050oC in the presence of Crofer22APU. T he formation of facetted particles is clearly observed. The EDS spectra for the LSM matrix and secondary phase are provided in Figure 8 -3 b and 5c, respectively. The characteristic peaks of Mn (K 1 5.899 keV) and Cr are pronounced while no Sr and La is found. This means that the secondary phase is a Mn -

PAGE 139

139 Cr -Ox. It seems that Sr2+ in LSM is in a stable state, and a s a result LSM reacts with Cr vapor differently than LSCF Since A -site cations are the same for LSCF and LSM, the results indicate that the transition metals in B -site play a critical role for the reactivity to Cr vapor deposition. Figure 8 4 a represents the SEM micrograph of LSM surf ace after heat treatment at 800oC for 50 hours in the pr esence of Crofer22APU and BaO together They are located as illustrated in Figure 8-1a. Compared to heat treated LSM in the absence of BaO, the formation of secondary phases was observed on the LSM surface. Figure 8 -4 b is the s pot mode EDS spectrum for the particle Compared to the EDS spectrum of as polished LSM (Figure 8 -3c ), it shows the characteristic peaks of Ba (L 1 4.466 keV) and Cr proving the presence of Ba and Cr for the particle. It seems that the penetration depth of accelerated electron beam o f SEM is deeper than the size of the particles and so EDS spectrum for the particle contains the characteristic peaks of LSM. During heat treatment, LSM was located above BaO powder without any direct contact Therefore the only way for BaO to deposit on LSM surface is through vapor phase diffusion, in which a vapor phase is formed from the BaO powder and mass transported to the LSM surface. As Ba vapor species are basic,126 they interact with the acidic Cr vapor species, forming a salt on LSM surface.110 T his study therefore, shows that the vaporization of BaO cannot be neglect ed at s ubstantially lower temperatures than boiling temperatures. LSM was heat treated with exposure to LSCF powder and Crofer22APU together as described in Figure 8 -1 b. Figure 8-5 a is the SEM micrographs of LSM after heat treatment at 800oC for 50 hours. It shows the fo rmation of particles on grain boundary

PAGE 140

140 after heat treatment. Figure 8 -5 b is the EDS spectrum for the particle. The intensity of the characteristic peak of Sr substantially increased compared to those of La and Mn. In addition, Cr was also found. It indicat es that the particle is a Sr-Cr -Ox and t he formation of Sr Cr -Ox means the mass transport of Sr vapor from LSCF powder to LSM as same as BaO. Sr vapor species are also basic, and thus react with Cr vapor species, forming Sr-Cr -Ox on the surface of the LSM. This result supports that Sr in LSCF is vaporized in the vicinity of surface region at 800oC. Furthermore this helps explain the slow mass reduction, which was observed by Bishop et al.120 The formation of Ba -Cr -Ox was found over the entire surface of LSM when it was heat treated with exposure to BaO and Crofer22APU. By contrast, the formation of Sr-Cr -Ox was found on the LSM surface only i n the region near the LSCF powder, which is marked by the dotted box in Figure 1b, and the most area remained unchanged. This means vaporization of Sr from LSCF occurs in the vicinity of surface region. 8. 4 Conclusion s Sr vaporization from LSCF was tested by comparing Cr contamination behavior on LSCF an d LSM at high temperature. Heat treatment of LSCF with exposure to Crofer22APU at 800oC for 50 hours led to the formation of Sr -Cr -Ox particle on LSCF surface. By contrast, no Cr vapor deposition occurred on LSM at 800oC, and the formation of Cr -Mn -Ox part icle was found at 1050oC. It seems that Sr in LSM is in stable state and so it does not react with Cr vapor species When LSM was exposed to BaO and Crofer22APU at 800oC, the formation of Ba-Cr -Ox particles was found. Since there was no direct contact betw een BaO powder and LSM, it indicated that BaO was vaporized and transported to LSM surface. In a same way, the formation of Sr -Cr -Ox was found when LSM was heat treated in the presence of LSCF powder and

PAGE 141

141 Crofer22APU at 800oC. The result well supports that Sr in LSCF is vaporized in the vicinity of surface region. As a heterogeneous catalyst, such vaporization of Sr from LSCF may degrade the catalytic activity and long term performance.

PAGE 142

142 Figure 8 1. Schematic of experimental set up. Figure 8 2 SEM micrographs of (a) as -polished and (b) heat treated LSCF in the presence of Crofer22APU.

PAGE 143

143 Figure 8 3 (a) LSM surface after heat treatment at 1050oC for 50 hours in the presence of Crofer22APU sheet. EDS spectra for (b) the newly for med particle after heat treatment and (c) as polished LSM sample.

PAGE 144

144 Figure 8 4 (a) Surface of LSM after heat treatment at 800oC for 50 hours in the presence of BaO powder and Crofer22APU. (b) Spot mode EDS spectra for the particle Figure 85 (a) SE M micrograph of heat treated LSM surface with exposure to

PAGE 145

145 CHAPTER 9 CONCLUSIONS Surface characteristic of La0.6Sr0.4Co0.2Fe0.8O3 in the absence of and presence of Cr contamination ha ve been investigated in order to better understand degradation m echanisms and improve the SOFC cathode performance. In many studies t he oxygen non-stoichiometry at high temperatures is a major concern for SOFC cathode materials and thus A site ions on ABO3 perovskite materials are partially substituted with the ions having a lower oxidation state to create more oxygen vacancies in the lattice. But t his study showed that some doped Sr tended to segregate in the surface region at high temperatures and formed catalytically inactive layers. AES and TEM provided that segregated Sr in La0.6Sr0.4Co0.2Fe0.8O3 (LSCF) precipitated into SrOx and grew on top of the surface. It seems that excess surface energy due to increase in oxygen vacancies lead to the precipitation of cations. Sr segregation led to reduction in the concentration of transit ion metal s on the surface region, which might bring significant reduction in catalytic activity for oxygen reduction. In order to improve its performance, cobalt oxide was impregnated into a porous LSCF cathode via wet impregnation and its e ffects were tested. The ASR of the electrode was reduced and the maximum power density of a single LSCF cathode cell was significantly improved. It seems that increased concentration of cobalt oxide on the surface enhances surface reaction, especially char ge transfer process. But the possibility of removal of Sr enriched surface layer cannot be ruled out. While a catalyst impregnation has been widely used previously mostly expensive materials such as Pt and Pd were applied. In this study, based on the result of surface characterization,

PAGE 146

146 SOFC performance was improved through Co impregnation without incurring significant increase in material costs. In the meanwhile, SOFCs need to be operated in a stack, in which Fe-Cr alloys are used for the interconnect. At high temperatures, Cr was found to vaporize from these metal alloys and deposit on the cathodes. Electrical conductivity relaxation, AC impedance spectroscopy and temperature programmed isotopic exchange (TPX) consistently showed that surface reaction kin etics of LSCF was substantially decreased in Cr contamination environment. This degradation is attributed to the formation of SrCrO4 particles and Sr deficient matrix on LSCF surface in Cr contamination condition. A phase transition from rhombohedral to cu bic structure occurred and Fe2O3 was formed in the near surface region due to Sr deficiency. It was found that t he reactivity between ABO3 perovskite materials and Cr vapor species was not primarily determined by basicity of A site ions but by B site trans ition metals. Unstable Co4+ and Fe4+ lead to the formation of a nucleation agent for the chemical reaction while stable Mn4+ does not form a nucleation agent. Therefore, Cr contamination is a chemical process for CoFe based materials and an electrochemica l process for Mn based materials. Finally this study identifies that doped Sr in LSCF n ot only precipitates on the surface but also vaporiz es in the vicinity of the surface region. As LSCF is a heterogeneous catalyst, this effect can play crucial role in its electrochemical performance.

PAGE 147

147 APPENDIX A THEORETIC EQUILIBRIU M VOLTAGE At 800oC the Nernst equation can be written as 2 H 2 O H 0 923K 2 H O 2 O H 0 Eq2 2 2 2 2 TP 0.21 P ln 96485C/mol 4 1073K K l 8.3144J/mo E p p p ln nF RT E E (A1 ) If it is a ssumed that gaseous pressures at anode and cathode sides are the same as the at mos pheric pressure, the pressure (p) is replaced by unit -less partial pressure (P). In order to s olv e Eq. A-1 the standard voltage variation with temperature ( 0 TE ), and partial pressure of H2O (g) and H2 (g) are necessary. First, 0 TE may be derived from the variation of standard free energy with temperatures. At constant pressures, the standard o) is written as o p o dT G d (A -2 ) where So is the standard entropy change. Generally, the temperature dependence of So is insignificant and thus, i f So is assumed to be independent of temperature, integral of Eq. 2 -11 from 298.15 K to an arbitrary temp erature (T) gives ) T (T 298.15K o 298.15K o 298.15K o T T ) T G (o 298.15K 298.15K o 298.15K o 298.15K (A3 ) This is in the form of o T =A+B T (A4 )

PAGE 148

148 Table A-1 provides the thermodynamic properties of reactants and produc t s o 298.15K and o 298.15K are 228.57 kJ/mol and 88.86 J/mol K, respectively. Thus, A is 484,000 J/mol and B is 88.86 J/mol K. Therefore, oTE at 800oC is 1.01V 96485C/mol 4 K 1 K 88.86J/mol l 484000J/mo nF Eo T oT 073 (A5 ) Approximate values of O H2P and 2HP can be obtained from the equilibrium vapor pressure of H2O (g) using Clausius -Clape yron equation. 19.732 T log 4.65 T 2900 (atm) p logO H2 (A-6 ) Since hydrogen gas flows thr ough the anode via a water bath at room temperature (298.15K), O H2p = 0.032 atm and subsequently 2HP = 1 0.032=0.968. From Eq. A -5 and A 6 the equilibrium voltage for an ideal case at 650oC is 1.13V 0.968 0.21 0.03 ln 96485C/mol 4 K 1 K l 8.3144J/mo 1.01V E2 2 Eq 073 (A7 ) Table A-1. Thermodynamic properties at 298.15K and 1bar Go (kJ/mol) So (J/mol K) O2 (g) 0 205.14 H2 (g) 0 130.68 H2O (g) -228.57 188.82

PAGE 149

149 APPENDIX B MEASUREMENT EFFECT OF NON -IDEAL STEP CHANGE IN ECR A kinetic model of electrical conductivity relaxation ( ECR ) assumes the ideal step change of pO2 inside of the reactor In reality it takes certain amou nt of time to switch the gas concentration. Den otter et al. calculated the reactor flush time based on the continuous stirred tank reactor model (B 1)127 r STP tot v, fT T Vr (B1) where f is the characteristic flush time, Vr it the dead volume, v,tot= total flow rate TSTP=room temp erature, and Tr= reactor tem p erature. They explained that 99% of gas inside the reactor would be replaced after the time of 4 x f has passed. Based on Eq. B-1, the characteristic flush time of our reactor at 800oC with total flow rate of 500ccm was calculated according to 0.2sec 1073K 273K cc/sec 60 500 4.4cc T T Vr r STP v,tot f (B2) It takes about 0.8 sec to switch 99% gas inside of the reactor. The actual flush time was measured by connecting the output of ECR to Mass spectrometer (MS, Extrel QMS ). I n order to avoid damage on MS, the gas flow to MS was controlled to 30 sccm using the ball flow meter between two set ups as seen in Figure B1. Figure B 2 shows the oxygen concentration profile after switching a gas. Eq. B 3 was used to obtain characteristic flush time128 ) t )exp( p (p p (t) pf 0 (B -3)

PAGE 150

150 where t is the time, f is the characteristic flush time, Po, Pt and P are the initial, transient, and final pressures, respectively. The measured flus h time was about 2.5 sec at 800oC. Based on this value and a kinetic model which account s for the non ideal step change127, expe rimental data was fit and compared with ideal step change model in Figure B 3. The non ideal step change kinetic model provided a more accurate fit at the early stage of exchanging reaction, however it did not affect much on the final values of Dchem and Kchem. Figure B1 Reactor flush time measurement set up using mass spectrometer. The flow from ECR to mass spec was limited to 30 sccm by the ball flow meter.

PAGE 151

151 Figure B2 Measured reactor flush time using mass spectrometer (a) and a fit using Eq B-3 (b). Figure B 3. A fit into experimental data using ideal step change model (a) and nonideal step change (b) model.

PAGE 152

152 CHAPTER C MEASUREMENT OF THE SURFACE SENSITIVE RAT E COEFFICIENT USING ISOTHERMAL ISOTOPIC SWITCHING C.1 Introduction T here were several attempts to get surface sensitive values by reducing the sample thickness. Chen et al .143, 144 made an epitaxial thin film of metal oxides with pulsed laser deposition (PLD) and measured the chemical exchange coefficient ( Kchem). They found that Kchem was much smaller than that of the bulk polycrystalline coefficient As Kim et al129 observed, this is because a preferentially oriented thin film has dissimilar surface properties than a polycrystalline sample. In addition, when a thin film grown on the single crystal substrate is tested at high temperatures, inter diffusion across the interface of film and substrate can occur, changing the oxygen transport property. To overcome this limitation, Ganeshananthan et al .130 prepared porous bulk polycrystalline samples with an average grain size o f a few microns. They found that the magnitude of Kchem of La0.6Sr0.4CoO3 was close to other data in the literature, but it increased as temperature decreased, suggesting an ex othermic surface reaction process. The objective of this study is to introduce t he application of a new experimental method to measure the surface sensitive reaction coefficient: isothermal isotopic switching. One of the benefits of this technique is easy sample preparation. Either fresh or heat treated powder can be tested without a dditional processing (such as pressing and polishing). Since mass spectrometry monitors the gas composition of the reactor in situ and in real time, electrode contacts to the samples arent necessary. This avoids the potential contribution to the catalytic activity from triple phase boundaries along the contact131. Small particle size and high surface area powder samples can provide highly surface sensitive kinetic values. As an initial test, the surface exchange coefficient of

PAGE 153

153 La0.6Sr0.4Co0.2Fe0.8O3 (LSCF), an extensively studied material for intermediate temperature SOFC, was measured and compared to the literature. C .2 Background Cathode materials maintain a dynamic equilibrium with the external atmosphere at sufficiently high temperatures When an oxygen isotope (18O2) is introduced into the sample environment, which has been equilibrated with normal oxygen (16O2), under constant overall oxygen partial pressure, the dynamic equilibrium will not be disturbed, but normal oxygen in the lattice will be displaced by isotopic oxygen. If isotopic effects are ignored, this can be expressed as o 18 k 2 18O 2 2V e 4 (g) Oin O (C -1 ) O out2V e 4 (g) O O 22 16 k o 16 (C 2 ) Direct measurement of the rate constant (kin and kout) is not possible due to lack of understanding of the mechanism. Instead, apparent rate coefficients can be measured assuming a first order reaction such that the flux i n and out can be expressed by130 Jin= V S C(t)] C [ k dt dC(t)18 o 18 in (C 3 ) Jout= V S C(t)] C [ k dt dC(t)16 o 16 out (C 4 ) where J is the flux, Co is the new equilibrium concentration, C(t) is the concentration at time t, V is the volume and S is the surface area. kin* and kout* are the surface exchange coefficients for incorporation and desorption, respectively. Since the sample is in quasi equilibrium, it should satisfy Joverall = Jin+Jout = 0. Accordingly, kin* is equivalent to kout* when C (18O2 ) = C (16O2 ). If a spher ical particle shape is assumed, integration of Eq. C 3 gives

PAGE 154

154 in in 18 o 18 18 o 18k 3 r t exp k S V t exp C(0) C C(t) C g(t) (C 5) where, C(0) is the initial equilibrium concentration and r is the average particle radius. T hus, fitting Eq. C 5 to the con centration profile of oxygen isotope obtained by mass spectrometry will provide surface exchange coefficients (kin*) values. C.3 Experimental Commercially available LSCF powder (Praxair) was used without additional modification. The specific surface area was 6.52 m2/g and the volume averaged particle radius was 1.4 m based on manufacturer provided particle size distribution data. The reactor schematic is the same at temperature programmed isotopic exchange set up. Due to small sample volume and no electrode connection, reactor size (i.d. = 4.12 mm, length = 10 cm and volume ~ 1.33 cm3) can be reduced and this ensures fast gas changing out time in the reactor with a relatively slow gas flow rate. If it follows a continuously stirred tank reactor model as su ggested by den Otter et al127, it takes less than 1 sec at 800oC with 20CCM. It would take even less time using the plug flow model. LSCF powder ( 0.015 g ) was placed atop the glass frit in the middle of the reactor. The flow rate of the feed gas was controlled by mass flow controllers (MKS & Alicat) adjusted to yield a total flow rate of 20 CCM. Gas composition in feed stream was switched using an automatic switching valve (Veeco) and the reactor effluent was constantly sampled by a downstream mass spectrometer (Extrel). First the sample was pretreated under 16O2 diluted with Helium for 30 mins at 800oC. Once the mass spectrometry signal reached steady state, the feed gas was switched to 18O2. The total oxygen concentration was fixed to 10000 PPM. Time dependent relaxation curves were

PAGE 155

155 monitored until a new steady state was reached. The measurements wer e performed at 700, 750 and 800oC. C .4 Result s and D iscussion Fig ure C -1 shows time dependent relaxation curve of normal and isotope oxygen after switching the gas composition from 16O2 to 18O2 at 700oC. 16O+18O represents a scrambled species formed on the surface by combining dissociatively adsorbed species (18Oads) and residual lattice oxygen (16O). Its concentration increased rapidly but decreased as adsorbed 16O diminishes. In Fig ure C 2, the amount of oxygen isotope incorporating into the lattice will be (isotope input 18O +1/2 (16O+18O)) ppm and that of desorbed normal oxygen is equivalent to (16O +1/2 (16O+18O)). Relaxation curves reach a plateau in appr oximately 3000 seconds, indicating that the surface exchange reaction reached isotopic steady state. Right after gas switching, the concentration of 16O2 appeared to overshoot, but this is probably due to pressure fluctuations during gas switching, which w as also observed by Lane et al .48 Despite the initial pressure fluctuation, the normalized relaxation curves of desorbed oxygen and incorporated oxygen isotopes were superimposed and were fit to Eq. C 5 (solid line) as shown in Fig ure C -2 giving kin* = kout*. The measured surface exchange coefficient is converted to the chemical exchange coefficient for the comparison using the thermodynamic factor .95 As shown in Fig ure C -3 the magnitude is lower than values reported in the literature and by our own measurements with the electrical conductivity method. It might be due to particle size distribution. Another possible explanation is different sample geometries and preparation procedures. Samples for both conductivity relaxation and IEDP were solid bars which were sintered at high temperatures to obtain suffic ient density and then

PAGE 156

156 polished with fine abrasives to produce a flat, well defined surface for more accurate estimation of surface kinetics .49 This is in c ontrast this with our sample, which is a fine powder with a high surface area to volume ratio. As a result, the surface composition or structure of powder sample can be different from the dense polycrystalline bar, causing the difference in catalytic activities. In order to calculate the activation energy of the surface reaction, experiments were conducted at three different temperatures. Figure C 4 shows the normalized relaxation curves of 18O2 which shows that the relaxation time decreases as temperature decreases. Fig.6 shows the Arrhenius plot of kin which y ields an activation energy of 23 kJ/mol for the surface reaction. This means the surface reaction is exothermic while in most literatur e an endothermic surface reaction for perovskite materials with similar chemical compositions has been observed .31, 132 Bouwmeester et al .52 reported an activation energy of 214 kJ/mol for the surface reaction of bulk polycrystalline LSCF and Chen et al .133 observed the activation energy of approximately 160kJ/mol for La0.5Sr0.5CoO3 epitaxial film. However, Ganeshananthan et al .130 found that the activation energy of porous La0.6Sr0.4CoO3 was around -23 kJ/mol, which is very close to that found from this study. There is one similar aspect between the powder and porous bulk sample. Both have substantially higher surface area than dense bulk polycrystalline. As shown in Figure C -4, the surface reaction coefficient increa ses with pO2 this means that the rate expre ssion for oxygen reduction should include a surface coverage term. As surface area becomes larger, surface coverage of adsorbed oxygen species may play a more important role in determining the apparent activatio n energy for surface reaction. Thus, the increase in surface coverage at reduced temperatures

PAGE 157

157 might enhance the apparent reaction coefficient, giving negative activation energy. Whether this is due to high surface area or another factor s affect on measurement needs further analysis. C.5 Conclusions A new application for the isothermal isotope switching method was applied to measure surface sensitive reaction coefficients of LSCF. An in situ powder sample was tested without sintering and electrode contacts The small sample volume allowed for a small reactor design, which reduced the gas switch time with relatively lower total gas flow than other experimental techniques. It shows that time dependent relaxation curves of normal and isotope oxygen can be fitt ed to a first order kinetic model. The magnitude of the surface exchange coefficient of the powder type sample used in this study was smaller than values reported in literature. That might be due to the different surface reactivity of fresh powder from den se bulk polycrystalline sample. Relaxation time for 18O2 increased as temperature decreased, giving the activation energy of 23 kJ /mol for the oxygen reduction process. This is substantially different from previous results for dense perovskite materials but it is close to that of porous bulk La0.6Sr0.4CoO3 Surface coverage of oxygen appears to play a larger role in high surface area samples.

PAGE 158

158 Fig ure C -1 Time dependent relaxation profile of normal oxygen (16O), oxygen isotope (18O) and scrambled oxygen (16O+18O) at 700oC 0 0.2 0.4 0.6 0.8 1 1.2 0 1000 2000 3000 4000 5000 6000 Desorbed oxygen Incorporated isotope Fitg(t)Time(sec) Figure C -2. Normali zed c oncentrations of desorbed oxygen from the sample and incorporated oxygen

PAGE 159

159 -6.5 -6 -5.5 -5 -4.5 -4 -3.5 -3 -2.5 -3 -2.5 -2 -1.5 -1 -0.5 0 Our group Lane et al Bouwmeester et al This studylog kchem (cm/sec)log pO2 Figure C -3. Kchem measured by ECR in this study and literatures 52, 94, converted using the thermodynamic factor at 800oC. 0 0.2 0.4 0.6 0.8 1 0 1500 3000 4500 6000 800C 750C 700CNormalized concentration Time(sec) Fig ure C -4 Normalized relaxation curves of incorporated oxygen isotope at different temperatures.

PAGE 160

160 Fig ure C -5 Arrhenius plot of surface exchange coefficient (pO2= 0.01 )

PAGE 161

161 APPENDIX D TAPE CASTING SLURRY COMPOSITION The tape-casting procedure was introduced in Chapter 5 The slurry compositions for 8YSZ, 10GDC, NiO and glass powder are given in Table D 1. Table D -1 Tape casting slurry composition Weight (g) wt% Weight (g) wt% 8YSZ (Tosho) 100 52.1 10GDC (Anan) 50 40.4 Fish oil* 1 0.5 Fish oil 1.5 1.2 DBP 14 7.3 DBP 7 5.7 PVB 10 5.2 PVB 5 4.0 Toluene 38. 2 19.9 Toluene 36. 7 29.7 ETOH 25.7 13.4 ETOH 22 17.8 PEG8000 3 1.6 PEG8000 1.5 1.2 Total 192 100 Total 123. 7 100 Weight (g) wt% Weight (g) wt% NiO 140 55.0 Glass 110 89.4 Fish oil 1.5 0.6 Fish oil 2.1 1.7 DBP 9.8 3.9 DBP 9.8 8.0 PVB 17 6.7 PVB 12 9.8 Toluene 47 18.5 Toluene 41 33.3 ETOH 37 14.5 ETOH 27 22.0 PEG8000 2.2 0.9 PEG8000 5.1 4.1 Total 254.5 100.0 Total 207 168 Menhaden fish oil

PAGE 162

162 APPENDIX E CR CONTAMINATION ON LSM82 La0.8Sr0.2MnO3 (LSM82) is the composition, which is more widely used for SOFC cathodes than La0.6Sr0.4MnO3 (LSM64) which was tested in this study. L SM 82 was tested in a same way that LSM 64 was tested. The powder (Fuel C ell Materials) was uniaxially pressed and then sintered at 1450oC. Figure E 1a shows the as polished surface of LSM82. EDS analysis provided that the darker regions were Mn rich oxide. Figure E 1b represents the LSM after heat treatment with exposure to Cr ofer22APU at 800oC for 50 hours. It shows no Cr vapor deposition on the surface This is in agreement with LSM64, which indicates that (La1 xSrx)MnO3 does not react with Cr vapor species at 800oC. Figure E 1 (a) As -polished LSM82 surface and (b) heat treated LSM82 at 800oC for 50hrs in the presence of Crofer22APU.

PAGE 163

163 APPENDIX F KCHEM AND DCHEM AT 750C The Dchem and Kchem at 750oC were obtained using electrical conductivity relaxation and plotted as a function of oxygen partial pressure in Figure F -1 The characteristic dependencies of Kchem and Dchem are close to that at 800oC meaning that the rate limiting step of LSCF is a charge transfer between LSCF and oxygen gas. Figure F 2 shows the degradation rate of Kchem at 750oC either. Surface reaction rate decreased at a relative fast rate during the 70 hours and at a slower rate after 70 hours. Interestingly, the degradation rate was increased with increasing pO2. Figure F 1. Dchem and Kchem at 750oC as a function of pO2

PAGE 164

164 Figure F 2. Degradation of Kchem at 750oC.

PAGE 165

165 APPENDIX G SURFACE VARIATION OF LCCF AND LBCF F igure G -1 shows the differentiated AES spectra of La0.6Ca0.4Co0.2Fe0.8O3 (LCCF) and La0.6Ba0.4Co0.2Fe0.8O3LBCF. In comparison to the AES spectrum of the pre-heat treated sample, LBCF showed an increase in the characteristic peak of Ba after heat treatment at 800oC for 50 hrs. By contrast, the AES spectra of LCCF present ed weak change after heat treatment. There is no doubt that the concentration of Ba increased while that of r est elements decreased but the elemental atomic percents of LBCF may not be accurate due to peak overlapping. LSCF and LBCF showed similar behavior, but LCCF showed weak segregation of elements. The tendency of segregation in doped alkaline earth metal ox ides is Ca
PAGE 166

166 APPENDI X H COBALT OXIDE DISPERSION ON LSCF SURFACE Co nitrate solution was prepared by dissolving Co(NO3)2 6H2O and PEG 14g in mixture of ethanol and water. It was dip-coated and heat treated at 700 and 900oC, respectively. Figure G -1 shows the formation of nano cobalt oxide on the surface of LSCF. Figure H -1 LSCF sur face after heat treating Co dip-coated LSCF

PAGE 167

167 L IST OF REFERENCES 1. R. E. Smalley, Mrs Bulletin 2005, 30 412-417. 2. C. J. Campbell and J. H. Laherrere, Scientific American 1998, 278, 77 -83. 3. N. Stern, The economics of climate change : the Stern review Cambridge University Press, Cambridge, UK, 2007. 4. B. C. H. Steele and A. Heinzel, Nature, 2001, 414, 345352. 5. S. M. Haile, Acta Materialia, 2003, 51, 5981 -600 0. 6. O. Yamamoto, Electrochimica Acta, 2000, 45 2423 -2435. 7. R. M. Ormerod, Chem. Soc. Rev. 2003, 32 17 -28. 8. T. Ishihara, Perovskite Oxide for Solid Oxide Fuel Cells Springer, New york, 2009. 9. P. Singh and N. Q. Minh, International Journal of Appl ied Ceramic Technology 2004, 1 5 15. 10. R. O'hayre, S.-W. Cha, W. Colella and F. B. Prinz, Fuel Cell Fundamentals John Wiley & Sons Newyork, 2006. 11. M. C. Williams, J. P. Strakey, W. A. Surdoval and L. C. Wilson, Solid State Ionics 2006, 177 2039 20 44. 12. E. Fontell, T. Kivisaari, N. Christiansen, J. B. Hansen and J. Palsson, Journal of Power Sources 2004, 131, 4956. 13. E. Riensche, U. Stimming and G. Unverzagt, Journal of Power Sources 1998, 73 251-256. 14. J. W. Fergus, Materials Science and Engineering A 2005, 397, 271283. 15. M. W. Barsoum, Fundamentals of ceramics McGraw -Hill, London, 2003. 16. H. Ricket, Electrochemistry of Solids: An Introduction Spring -Verlag, New York 1982. 17. S. B. Adler, X. Y. Chen and J. R. Wilson, Journal of Cat alysis 2007, 245, 91109. 18. S. P. S. Badwal and K. Foger, Ceramics International 1996, 22 257265. 19. W. Z. Zhu and S. C. Deevi, Materials research bulletin 2003, 38 957972. 20. H. Y. Zhu and R. J. Kee, Journal of Power Sources 2006, 161, 957964.

PAGE 168

168 21. R. W. Sidwell and W. G. Coors, Journal of Power Sources 2005, 143, 166172. 22. K. C. Wincewicz and J. S. Cooper, Journal of Power Sources 2005, 140, 280-296. 23. S. B. Adler, Chemical Reviews 2004, 104, 47914843. 24. T. Shimizu, Chemistry Letters 1980, 1-4. 25. J. Nowotny, T. Bak, M. K. Nowotny and C. C. Sorrell, Advances in Applied Ceramics 2005, 104 154-164. 26. R. Hammami, H. Batis and C. Minot, Surface Science, 2009, 603, 30573067. 27. Y. J. Leng, S. H. Chan and Q. L. Liu, International Jo urnal of Hydrogen Energy 2008, 33 38083817. 28. T. Nitadori, T. Ichiki and M. Misono, Bulletin of the Chemical Society of Japan 1988, 61 621-626. 29. T. Nakamura, M. Misono and Y. Yoneda, Bulletin of the Chemical Society of Japan, 1982, 55, 394399. 3 0. I. Yasuda and T. Hikita, Journal of the Electrochemical Society 1994, 141, 1268 1273. 31. J. A. Kilner, R. A. DeSouza and I. C. Fullarton, Solid State Ionics 1996, 86 -8 703 709. 32. J. M. D. Tascon and L. G. Tejuca, Reaction Kinetics and Catalysis Le tters 1980, 15, 185191. 33. H. Takahashi, F. Munakata and M. Yamanaka, Physical Review B 1998, 57, 15211-15218. 34. L. W. Tai, M. M. Nasrallah, H. U. Anderson, D. M. Sparlin and S. R. Sehlin, Solid State Ionics 1995, 76 273 -283. 35. L. W. Tai, M. M. N asrallah, H. U. Anderson, D. M. Sparlin and S. R. Sehlin, Solid State Ionics 1995, 76 259 -271. 36. E. Girdauskaite, H. Ullmann, M. Al Daroukh, V. Vashook, M. Bulow and U. Guth, Journal of Solid State Electrochemistry 2007, 11 469-477. 37. Y. Teraoka, H M. Zhang, S. Furukawa and N. Yamazoe, Chemistry Letters 1985, 17431746. 38. S. P. Simner, M. D. Anderson, M. H. Engelhard and J. W. Stevenson, Electrochemical and Solid State Letters 2006, 9 A478-A481.

PAGE 169

169 39. J. Y. Kim, V. L. Sprenkle, N. L. Canfield, K D. Meinhardt and L. A. Chick, Journal of the Electrochemical Society 2006, 153, A880-A886. 40. J. W. Fergus, International Journal of Hydrogen Energy 2007, 32 3664 3671. 41. C. Sun, R. Hui and J. Roller, Journal of Solid State Electrochemistry 2009. 42. R. A. De Souza and J. A. Kilner, Solid State Ionics 1999, 126, 153161. 43. J. Larminie and A. Dicks, Fuel Cell Systems Explained John Wiley and Sons, 2003. 44. Y. Takeda, R. Kanno, M. Noda, Y. Tomida and O. Yamamoto, Journal of the Electrochemical So ciety 1987, 134, 26562661. 45. J. W. Stevenson, T. R. Armstrong, R. D. Carneim, L. R. Pederson and W. J. Weber, Journal of the Electrochemical Society 1996, 143 27222729. 46. M. W. den Otter, L. M. van der Haar and H. J. M. Bouwmeester, Solid State Ionics 2000, 134 259264. 47. S. Wang, P. A. W. van der Heide, C. Chavez, A. J. Jacobson and S. B. Adler, Solid State Ionics 2003, 156 201 -208. 48. J. A. Lane and J. A. Kilner, Solid State Ionics 2000, 136, 9971001. 49. R. Ganeshananthan and A. V. Virk ar, Journal of the Electrochemical Society 2006, 153 A2181-A2187. 50. J. Crank, Mathematics of diffusion, Oxford University Press, London, 1956. 51. R. A. De Souza and J. A. Kilner, Solid State Ionics 1998, 106, 175187. 52. H. J. M. Bouwmeester, M. W. Den Otter and B. A. Boukamp, Journal of Solid State Electrochemistry 2004, 8 599605. 53. J. Maier, Pysical Chemistry of Ionic Materials Wiley, West Sussex, 2004. 54. C. C. Kan and E. D. Wachsman, Journal of the Electrochemical Society 2009, 156, B695-B 702. 55. J. Nowotny, T. Bak, M. K. Nowotny and C. C. Sorrell, Advances in Applied Ceramics 2005, 104 147-153. 56. M. Thompson, M. D. Baker, A. Christie and J. F. Tyson, Auger Electron Spectroscopy John Wiley and Sons, Newyork, 1985. 57. G. E. Mcguire, A uger Electron Spectroscopy Reference Manual Plenum, New York, 1979.

PAGE 170

170 58. J. Goldstein, D. Newbury, D. Joy, C. Lyman, P. Echlin, E. Lifshin, L. Sawyer and J. Michael, Scanning Electron Microscopy and X -ray Microanlysis Springer, New York, 2003. 59. E. Buche r and W. Sitte, Solid State Ionics 2010. 60. P. A. W. v. d. Heide, SURFACE AND INTERFACE ANALYSIS, 2002, 33 414 425. 61. M. M. Viitanen, R. G. von Welzenis, H. H. Brongersma and F. P. F. van Berkel, Solid State Ionics 2002, 150 223 -228. 62. V. E. Henri ch, Reports on Progress in Physics 1985, 48, 1481 -1541. 63. D. A. Porter and K. E. Eastering, Phase Transformations in Metals and Alloys CRC press, FL, 1992. 64. B. Rahmati, J. Fleig, W. Sigle, E. Bischoff, J. Maier and M. Ruhle, Surface Science, 2005, 5 95 115 -126. 65. R. Meyer, R. Waser, J. Helmbold and G. Borchardt, Journal of Electroceramics 2002, 9 103112. 66. K. Gomann, G. Borchardt, A. Gunhold, W. Maus -Friedrichs and H. Baumann, Physical Chemistry Chemical Physics, 2004, 6 3639 3644. 67. Y. L. Wang, K. Duncan, E. D. Wachsman and F. Ebrahimi, Solid State Ionics 2007, 178 53 -58. 68. K. Szot, M. Pawelczyk, J. Herion, C. Freiburg, J. Albers, R. Waser, J. Hulliger, J. Kwapulinski and J. Dec, Applied Physics a -Materials Science & Processing 1996, 6 2 335343. 69. H. Tanaka and M. Misono, Current Opinion in Solid State & Materials Science, 2001, 5 381387. 70. L. C. Dufour, G. L. Bertrand, G. Caboche, P. Decorse, A. El Anssari, A. Poirson and M. Vareille, Solid State Ionics 1997, 101, 661666. 71. S. N. Ruddlesden and P. Popper, Acta Crystallographica 1957, 10 538540. 72. M. Konigstein and C. R. A. Catlow, Journal of Solid State Chemistry 1998, 140 103-115. 73. F. Tietz, A. Mai and D. Stover, Solid State Ionics 2008, 179, 1509 -1515. 74. W. D. Copeland and R. A. Swalin, Journal of Physics and Chemistry of Solids 1968, 29 313-&.

PAGE 171

171 75. E. P. Murray, M. J. Sever and S. A. Barnett, Solid State Ionics 2002, 148, 2734. 76. H. J. Hwang, M. B. Ji -Woong, L. A. Seunghun and E. A. Lee, Journal of Power Sources 2005, 145, 243248. 77. F. Qiang, K. N. Sun, N. Q. Zhang, X. D. Zhu, S. R. Le and D. R. Zhou, Journal of Power Sources 2007, 168, 338345. 78. H. Y. Jung, K. S. Hong, H. G. Jung, H. Kim, H. R. Kim, J. W. Son, J. Kim, H. W. Lee and J. H. Lee, Journ al of the Electrochemical Society 2007, 154 B480-B485. 79. V. A. C. Haanappel, A. Mai and J. Mertens, Solid State Ionics 2006, 177, 20332037. 80. M. Sahibzada, S. J. Benson, R. A. Rudkin and J. A. Kilner, Solid State Ionics 1998, 115 285290. 81. S. P. Simner, J. R. Bonnett, N. L. Canfield, K. D. Meinhardt, J. P. Shelton, V. L. Sprenkle and J. W. Stevenson, Journal of Power Sources 2003, 113 1 10. 82. H. Uchida, S. Arisaka and M. Watanabe, Solid State Ionics 2000, 135, 347-351. 83. K. Yamahara, C. P. Jacobson, S. J. Visco, X. F. Zhang and L. C. de Jonghe, Solid State Ionics 2005, 176, 275279. 84. D. Oh, D. Gostovic and E. D. Wachsman, In preparation, 2010. 85. A. A. Vance and S. McIntosh, Journal of the Electrochemical Society 2008, 155 B1 -B7. 8 6. Y. K. Lee, J. Y. Kim, Y. K. Lee, I. Kim, H. S. Moon, J. W. Park, C. P. Jacobson and S. J. Visco, Journal of Power Sources 2003, 115 219 -228. 87. J. M. Serra, V. B. Vert, M. Betz, V. A. C. Haanappel, W. A. Meulenberg and F. Tietz, Journal of the Electr ochemical Society 2008, 155 B207-B214. 88. M. A. Pena and J. L. G. Fierro, Chemical Reviews 2001, 101, 1981 2017. 89. F. S. Baumann, J. Fleig, H. U. Habermeier and J. Maier, Solid State Ionics 2006, 177, 10711081. 90. M. J. L. Ostergard and M. Mogensen, Electrochimica Acta, 1993, 38, 20152020. 91. A. Esquirol, N. P. Brandon, J. A. Kilner and M. Mogensen, Journal of the Electrochemical Society 2004, 151, A1847 -A1855. 92. J. W. N. I. Chorkendorff, Concepts of Modern Catalysis and Kinetics WILEY -VCH, 2 003.

PAGE 172

172 93. N. Miura, Y. Okamoto, J. Tamaki, K. Morinaga and N. Yamazoe, Solid State Ionics 1995, 79 195 -200. 94. J. A. Lane, S. J. Benson, D. Waller and J. A. Kilner, Solid State Ionics 1999, 121 201-208. 95. J. Maier, Solid State Ionics 1998, 112 197 -228. 96. X. Montero, F. Tietz, D. Stover, M. Cassir and I. Villarreal, Corrosion Science, 2009, 51 110-118. 97. X. Montero, F. Tietz, D. Stover, M. Cassir and I. Villarreal, Journal of Power Sources 2009, 188, 148155. 98. W. Z. Zhu and S. C. Deevi, Materials Research Bulletin, 2003, 38 957 -972. 99. E. Konysheva, H. Penkalla, E. Wessel, J. Mertens, U. Seeling, L. Singheiser and K. Hilpert, Journal of the Electrochemical Society 2006, 153 A765-A773. 100. S. Taniguchi, M. Kadowaki, H. Kawamura, T. Yasuo, Y. Akiyama, Y. Miyake and T. Saitoh, Journal of Power Sources 1995, 55 7379. 101. S. P. Simner, M. D. Anderson, G. G. Xia, Z. Yang, L. R. Pederson and J. W. Stevenson, Journal of the Electrochemical Society 2005, 152, A740A745. 102. S. G. Li, N. P. X u, J. Shi, M. Z. C. Hu and E. A. Payzant, Journal of Materials Science Letters 2001, 20 1631 -1633. 103. D. Oh, D. Gostovic and E. Wachsman, to be submitted for publication. 104. S. P. Jiang, S. Zhang and Y. D. Zhen, Journal of the Electrochemical Society 2006, 153 A127-A134. 105. Ohbayash.H, T. Kudo and T. Gejo, Japanese Journal of Applied Physics 1974, 13, 1 7. 106. D. Waller, J. A. Lane, J. A. Kilner and B. C. H. Steele, Materials Letters 1996, 27 225-228. 107. K. H. Kim, S. H. Lee and J. S. Choi, Journal of Physics and Chemistry of Solids 1985, 46 331-338. 108. G. C. Kostogloudis and C. Ftikos, Solid State Ionics 1999, 126 143 151. 109. D. Oh, E. Armstrong, D. Jung, J. Yoo and E. Wachsman, to be submitted for publication 110. H. Yokokawa, T. H orita, N. Sakai, K. Yamaji, M. E. Brito, Y. P. Xiong and H. Kishimoto, Solid State Ionics 2006, 177, 3193 -3198.

PAGE 173

173 111. M. C. Tucker, H. Kurokawa, C. P. Jacobson, L. C. De Jonghe and S. J. Visco, Journal of Power Sources 2006, 160 130 -138. 112. I. H. Jung, Solid State Ionics 2006, 177 765-777. 113. E. Povoden, A. N. Grundy and L. J. Gauckler, International Journal of Materials Research, 2006, 97 569578. 114. S. P. Jiang, J. P. Zhang, L. Apateanu and K. Foger, Journal of the Electrochemical Society 2000 147, 4013 -4022. 115. S. P. Jiang, Journal of Materials Science 2008, 43 67996833. 116. J. A. M. Vanroosmalen and E. H. P. Cordfunke, Journal of Solid State Chemistry 1994, 110 109112. 117. S. Piskunov, E. Spohr, T. Jacob, E. A. Kotomin and D. E. El lis, Physical Review B 2007, 76 118. Y. Matsuzaki and I. Yasuda, Journal of the Electrochemical Society 2001, 148 A126-A131. 119. M. Backhaus -Ricoult, Solid State Ionics 2006, 177 2195 2200. 120. S. R. Bishop, K. L. Duncan and E. D. Wachsman, Jour nal of the Electrochemical Society 2009, 156, B1242 -B1248. 121. K. Hilpert, D. Das, M. Miller, D. H. Peck and R. Weiss, Journal of the Electrochemical Society 1996, 143, 36423647. 122. Y. F. Chang, G. A. Somorjai and H. Heinemann, Journal of Catalysis 1995, 154 2432. 123. P. Decorse, G. Caboche and L. C. Dufour, Solid State Ionics 1999, 117, 161169. 124. M. Shimada, H. Kokawa, Z. J. Wang, Y. S. Sato and I. Karibe, Acta Materialia 2002, 50 23312341. 125. A. Mai, V. A. C. Haanappel, S. Uhlenbruck, F. Tietz and D. Stover, Solid State Ionics 2005, 176, 1341 -1350. 126. H. Flood and T. Forland, Acta Chemica Scandinavica 1947, 1 592604. 127. M. W. den Otter, H. J. M. Bouwmeester, B. A. Boukamp and H. Verweij, Journal of the Electrochemical Society 2 001, 148, J1 -J6. 128. P. Trambouze, H. V. Landeghem and J. P. Wauquier, Editions Technip, Paris, 1988.

PAGE 174

174 129. G. Kim, S. Wang, A. J. Jacobson and C. L. Chen, Solid State Ionics 2006, 177 14611467. 130. R. Ganeshananthan and A. V. Virkar, Journal of the El ectrochemical Society 2005, 152 A1620-A1628. 131. J. Nowotny, T. Bak and C. C. Sorrell, Advances in Applied Ceramics 2005, 104 174-180. 132. T. Ramos and A. Atkinson, Solid State Ionics 2004, 170, 275286. 133. X. Chen, S. Wang, Y. L. Yang, L. Smith, N. J. Wu, B. I. Kim, S. S. Perry, A. J. Jacobson and A. Ignatiev, Solid State Ionics 2002, 146, 405413.

PAGE 175

175 BIOGRAPHICAL SKETCH D ongjo Oh as also known as DJ, was born in Haenam, South Korea and mov ed to Incheon at 5 years old. From very young ages, he liked a robot animation and so wanted to be a scientist. He graduated Bu-pyung elementary school, Bu -pyung seo middle school, and Bu kwang high sch ool. After he entered Inha U niversity in Incheon and majored materials science and engineering. He voluntarily worked as undergraduate research assistance from sophomore and participated in lots of activity in the university. And he started military service from Oct. 2000 to Dec. 2002 as a KATUSA (Korean Augmentati on to U.S. Army) In the army he learned leadership and professionalism. During that period, he wished to study abroad. After finishing the army service he prepared TOEFL and GRE during the senior and applied materials science and engineering program in th e University of Florida W hile at UF his knowledge on basic material science was improved significantly. In Dec 2008, Chanyoung Yun and DJ became a family and Michelle was born in Sep. 2009. Life in U.S. brought him not only scientific knowledge but also c onfidence He is not fully fledged but ready to come true his dream of young ages. He always appreciates Jesus for giving this chance.