<%BANNER%>

Engineering High Performance Intermediate Temperature Solid Oxide Fuel Cells

Permanent Link: http://ufdc.ufl.edu/UFE0024335/00001

Material Information

Title: Engineering High Performance Intermediate Temperature Solid Oxide Fuel Cells
Physical Description: 1 online resource (126 p.)
Language: english
Creator: Ahn, Jin
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: afl, anode, bilayer, bismuth, ceria, esb, gdc, it, sndc, sofc
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: Solid oxide fuel cells (SOFCs) are an efficient, fuel flexible energy conversion device, capable of operating on fuels ranging from natural gas to gasoline, diesel, and biofuels, as well as hydrogen. However, to this point the marketability of SOFCs has been limited by their high operating temperatures. Achieving high power at intermediate temperatures (IT, 500 700 masculine ordinalC) would be a significant breakthrough, as low temperature operation would result in better stability and allow for a broader range of material options for the SOFC components as well as the balance of plant, such as stainless steel interconnects (which are only viable at < 700 masculine ordinalC). Thus far, power densities on the order of 2 W/cm2 have been limited to temperatures above 800 masculine ordinalC. This dissertation contains a series of works to realize exceptionally high power at IT ranges. First, improved fabrication techniques including anode tapecasting and electrolyte spray coating were developed, and a molecular approach to anode functional layer (AFL) was employed using precursor solutions. This newly developed AFL reduced the ASR of a SOFC sample by 60 % and increased the open circuit potential (OCP) by more than 0.1 V resulting in a 140 % increase in power. Further investigations into this molecular AFL showed that a multilayered AFL can further reduce the ASR and increase the maximum power density. Secondly, the potential use of Sm0.075Nd0.075Ce0.85O2-? as an electrolyte has been investigated. The current-voltage (I-V) performance of the cell exhibits a maximum power density reaching 1.38 W/cm2 with an area specific resistance (ASR) of 0.087 ?cm2 at 650 masculine ordinalC with 90 sccm of air and wet hydrogen. Also, the high OCP achieved at 500 masculine ordinalC (0.96 V) as well as the high performance confirmed the viability of Sm0.075Nd0.075Ce0.85O2-? as an alternative electrolyte material. The cathode used for this study was La0.6Sr0.4Co0.2Fe0.8O3 (LSCF) Gd0.1Ce0.9O2 (GDC) composite. Finally, Er0.8Bi1.2O3 (ESB) / GDC bilayered electrolyte combined with recently developed ESB / Bi2Ru2O7 (BRO) composite cathodes was tested. In this work a maximum power density of 2 W/cm2 was achieved at 650 masculine ordinalC with the help of the novel AFL and tapecast anode supports. This is the highest power yet achieved in the IT range and I believe redefines the expectation level for maximum power under IT-SOFC operating conditions.
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 Jin Ahn.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Wachsman, Eric D.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2009-11-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024335:00001

Permanent Link: http://ufdc.ufl.edu/UFE0024335/00001

Material Information

Title: Engineering High Performance Intermediate Temperature Solid Oxide Fuel Cells
Physical Description: 1 online resource (126 p.)
Language: english
Creator: Ahn, Jin
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: afl, anode, bilayer, bismuth, ceria, esb, gdc, it, sndc, sofc
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: Solid oxide fuel cells (SOFCs) are an efficient, fuel flexible energy conversion device, capable of operating on fuels ranging from natural gas to gasoline, diesel, and biofuels, as well as hydrogen. However, to this point the marketability of SOFCs has been limited by their high operating temperatures. Achieving high power at intermediate temperatures (IT, 500 700 masculine ordinalC) would be a significant breakthrough, as low temperature operation would result in better stability and allow for a broader range of material options for the SOFC components as well as the balance of plant, such as stainless steel interconnects (which are only viable at < 700 masculine ordinalC). Thus far, power densities on the order of 2 W/cm2 have been limited to temperatures above 800 masculine ordinalC. This dissertation contains a series of works to realize exceptionally high power at IT ranges. First, improved fabrication techniques including anode tapecasting and electrolyte spray coating were developed, and a molecular approach to anode functional layer (AFL) was employed using precursor solutions. This newly developed AFL reduced the ASR of a SOFC sample by 60 % and increased the open circuit potential (OCP) by more than 0.1 V resulting in a 140 % increase in power. Further investigations into this molecular AFL showed that a multilayered AFL can further reduce the ASR and increase the maximum power density. Secondly, the potential use of Sm0.075Nd0.075Ce0.85O2-? as an electrolyte has been investigated. The current-voltage (I-V) performance of the cell exhibits a maximum power density reaching 1.38 W/cm2 with an area specific resistance (ASR) of 0.087 ?cm2 at 650 masculine ordinalC with 90 sccm of air and wet hydrogen. Also, the high OCP achieved at 500 masculine ordinalC (0.96 V) as well as the high performance confirmed the viability of Sm0.075Nd0.075Ce0.85O2-? as an alternative electrolyte material. The cathode used for this study was La0.6Sr0.4Co0.2Fe0.8O3 (LSCF) Gd0.1Ce0.9O2 (GDC) composite. Finally, Er0.8Bi1.2O3 (ESB) / GDC bilayered electrolyte combined with recently developed ESB / Bi2Ru2O7 (BRO) composite cathodes was tested. In this work a maximum power density of 2 W/cm2 was achieved at 650 masculine ordinalC with the help of the novel AFL and tapecast anode supports. This is the highest power yet achieved in the IT range and I believe redefines the expectation level for maximum power under IT-SOFC operating conditions.
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 Jin Ahn.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Wachsman, Eric D.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2009-11-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024335:00001


This item has the following downloads:


Full Text

PAGE 1

1 ENGINEERING HIGH PERFORMANCE INTERMEDIATE TEMPERATURE SOLID OXIDE FUEL CELLS By JIN SOO AHN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE D EGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009

PAGE 2

2 2009 Jin Soo Ahn

PAGE 3

3 To my loving wife, In Kyung

PAGE 4

4 ACKNOWLEDGMENTS I would like to thank my advisor, Professor Eric D. Wachsman, for giving me th e opportunity, advice and support for this work. His prudent guidance and deep understanding of electrochemistry and solid state ionics ma de the task of learning a new subject much easier I am confident that I will draw on my experience under his guida nce as I move forward I am also grateful to other members of my supervisory committee Dr. Juan Nino, Dr. Wolfgang M. Sigmund, Dr Susan Sinnott, and Dr. Mark E. Orazem for their time and thoughtful insight. I would like to thank Dr. Nino for inspiring my use of co doped ceria system s which consist s of one chapter of this dissertation I gratefully acknowledge Dr. Heesung Yoon s kind help in setting up many of my experiments on electrochemical performance of solid oxide fuel cells. Without him, all the achievements and most results that I have made would have been impossible. I am grateful to Dr. Matthew Camar a tta for the years of help and co-working alongside me Also, I appreciate all the grammar corrections he has performed on my writing here at UF Id like to thank Dr. Jiho Yoo for kindly providing advice and counsel in research as well as in my personal life. It was also a great pleasure to work with Doh Won Jung, Dong Jo Oh, Kang Taek Lee, Matthew Barnet, Byung Wook Lee, Brian Blackburn, Dan Gos tavic, Sean Bishop, Dr. Takkeun Oh, Eric Macam, Cynthia Kan, Eric Armstrong, Jianlin Lee and other group members. Their suggestions, cooperation, and comments were invaluable, and have helped me to improve my graduate research. I would also like to give a special thank you to Dr. Shobit Omar for his great help in the synthesis and characterization of a new material. Also, he provided many insightful discussions on the ceria systems. I appreciate Dr. Sejin Kim for his help setting up PLD deposition of ESB a nd Dr. Rajiv Singh for use of his PLD system. I thank Dr. Enrico Traversa and Dr. Daniele Pergolesi for their co -work on bilayered electrolytes. Their help enabled me to achieve an

PAGE 5

5 extraordinary result in fuel cell performances. I would also acknowledge the Department of Defense (DO D) and Lynntech, Inc. for their funding (2007NAV4910001) of the project s I have been working on. I thank my wife for years of support love and patience for bearing with me during my unexpectedly long study. Without her, study ing for years would have been impossible. Finally, I would like to dedicate this dissertation to all my family members my sister, father and mother For years, they have supported my study. Without thei r unconditional sacrifice and endless love, I could ha ve not achieved anything.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................................... 4 LIST OF TABLES ................................................................................................................................ 9 LIST OF FIGURES ............................................................................................................................ 10 ABSTRACT ........................................................................................................................................ 14 CHAPTER 1 INTRODUCTION ....................................................................................................................... 16 1.1. Solid Oxi de Fuel Cells ......................................................................................................... 16 1.2. Anode Functional Layers .................................................................................................... 19 1.3. Sm0.075Nd0.075Ce0.85O2as Alternative Electrolyte ............................................................ 22 1.4. ESB/GDC Bilayer Electrolytes for IT -SOFC .................................................................... 28 2 BACKGROUND ......................................................................................................................... 32 2.1. Fundamental Mechanisms of SOFC ................................................................................... 32 2.2. Potential Losses of SOFC .................................................................................................... 34 2.2.1. Activation Pol arization ............................................................................................. 36 2.2.2. Ohmic Polarization .................................................................................................... 37 2.2.3. Concentration Polarization ....................................................................................... 37 2.2.4. Internal Leakage Current .......................................................................................... 38 2.2.5. Combination of Each Polarization ........................................................................... 39 3 EXPERIMENTAL PROCEDURES .......................................................................................... 41 3.1. SOFC Fabrication ................................................................................................................ 41 3.1.1. Anode Tapecasting .................................................................................................... 41 3.1.2. SNDC Elect rolyte Deposition .................................................................................. 41 3.1.3. GDC Electrolyte Deposition ..................................................................................... 43 3.1.4. LSCF Composite Cathode ........................................................................................ 43 3.1.5. AFL Deposition ......................................................................................................... 44 3.2. Fabrication of ESB/GDC Bilayer Electrolytes .................................................................. 44 3.2.1. Co -pressing Proce dure .............................................................................................. 45 3.2.2. Colloidal Route .......................................................................................................... 46 3.2.3. Cold PLD ................................................................................................................... 47 3.2.4. Hot PLD ..................................................................................................................... 49 3.3. Characterization ................................................................................................................... 49

PAGE 7

7 3.3.1. Microstructural Analysis ........................................................................................... 49 3.3.2. XRD ........................................................................................................................... 50 3.3.3. I -V Measurement ....................................................................................................... 50 3.3.4. Impedance Spectroscopy .......................................................................................... 51 4 ANODE FUNCTIONAL LAYER FOR IT -SOFC ................................................................... 54 4.1. Homogeneous Functional Layers ....................................................................................... 54 4.1.1. Microstructural Analysis ........................................................................................... 54 4.1 .2. Electrochemical Performance ................................................................................... 58 4.2. Heterogeneous Functional Layers ....................................................................................... 66 4.2.1. Microstructural Analysis ........................................................................................... 66 4.2.2. Electrochemical Performance ................................................................................... 67 5 CODOPED CERIA AS ALTERNATIVE ELECTROL Y TES ................................................ 73 5.1. XRD ...................................................................................................................................... 73 5.2. Particle Size .......................................................................................................................... 73 5.3. Microstructural Analy sis ..................................................................................................... 75 5.4. Power Density ...................................................................................................................... 76 5.5. Impedance Analysis ............................................................................................................. 78 6 ESB/GD C BILAYER ELECTROLYTE FOR IT -SOFC ......................................................... 81 6.1. Co -pressing Procedure ......................................................................................................... 81 6.2. Colloidal Route .................................................................................................................... 84 6.3. Cold PLD .............................................................................................................................. 88 6.4. Hot PLD ................................................................................................................................ 93 7 CONCLUSIONS ......................................................................................................................... 97 APPENDIX A STACK CELL APPLICATIONS ............................................................................................ 100 B Reproducibility and stability .................................................................................................... 106 1. Slow P rogress toward H igh P erformance ............................................................................ 106 1.1 Electrolyte P article S ize E ffect ................................................................................... 107 1.2. Slurry C omposition .................................................................................................... 109 1.3. Co -pressing ................................................................................................................. 111 1.4. Spray Coating ............................................................................................................. 113 1.5. Reproducibility ........................................................................................................... 114 2. Future W ork .......................................................................................................................... 115 2.1. Sealing and L eak ........................................................................................................ 115 2.2. Long T erm S tability ................................................................................................... 118 2.3. Critical Flaws .............................................................................................................. 119

PAGE 8

8 LIST OF REFERENCES ................................................................................................................. 121 BIOGRAPHICAL SKETCH ........................................................................................................... 126

PAGE 9

9 LIST OF TABLES Table page 4 1 Details of No -AFL, Partial AFL and Full -AFL sam ples at 650 C. The unit for ASR 2. ....................................................................................................................... 59 4 2 Summary of detailed ASR values 650 C including Ni GDC AFL and AFML sample. ................................................................................................................................................. 69 5 1 Comparison between the total ASR obtained from I -V characteristic and impedance measurement. .......................................................................................................................... 80 6 1 Details of co -pressed cell. ...................................................................................................... 83 6 2 Details of tapecast cell without AFL. .................................................................................... 87 6 3 Details of tapecast cell wit AFL. ........................................................................................... 91 6 4 Summary of GDC single l ayer vs. ESB/GDC bilayer results. ............................................ 96

PAGE 10

10 LIST OF FIGURES Figure page 1 1 Schematic diagram of a SOFC with hydrogen fuel and oxygen. ........................................ 16 1 2 SECA phase 1 requirement. ................................................................................................... 18 1 3 Comparison of maximum power densities of thin film GDC of samaria doped ceria (SDC ) electrolyte cells with va rious cathodes. .................................................................... 19 1 4 SOFC depending on operating temperatures. ....................................................................... 23 1 5 Ionic conductivities of various oxides including the conductivity values of GDC, SNDC and dysprosium and tungsten stabilized bismuth oxide (DWSB). .......................... 25 1 6 Conductivity values of the ceria system with gadolinium single dopant and samarium and neodymi um double dopant. ............................................................................................ 27 1 7 Bilayer electrolyte concept with relative thickness ratio to control interfacial oxygen partial pressure. ...................................................................................................................... 29 2 1 A typical current -voltage characteristics of a SOFC .......................................................... 35 2 2 Separation of polarizations using a model developed by Yoon et al. ................................. 39 3 1 PLD system with KrF excimer laser. .................................................................................... 48 3 2 Electrochemical measurement station ................................................................................... 53 4 1 TEM images of raw powders of NiO (top) and GDC (bottom) used for tape casting of anode support. .................................................................................................................... 54 4 2 Backscattered images showing change in anode surface porosity and roughness by GDC functional layer deposition. ....................................................................................... 56 4 3 SEM micrographs showing of the surface view of the GDC electrolyte deposited by spray coating on AFL coated anode. ..................................................................................... 57 4 4 S EM micrograph showing the cross -sectional view of the SOFC with AFL after I -V testing. ..................................................................................................................................... 57 4 5 I-V characteristics of No -AFL, Partial -AFL and Full -AFL samples using 30 sccm of wet hydrogen and air at 650 C. ............................................................................................ 58 4 6 The impedance spectra of No -AFL, Partial -AFL and Full -AFL samples at 650 C under open circuit conditions. ............................................................................................... 60

PAGE 11

11 4 7 I-V characteristics of Full -AFL sample at temperatures from 450 to 650 C using 30 sccm of wet H2 and air. .......................................................................................................... 61 4 8 The impedance spectra of Full -AFL sample at various temperatures under operating conditions. ............................................................................................................................... 62 4 9 The total, electrode and ohmic ASR values calculated from impedance spectra at temperatures from 450 to 650 C. ......................................................................................... 63 4 10 The effect of the gas flow -rate on Partial -AFL sample at 650 C. .................................... 64 4 11 The effect of gas flow rate and gas composition on the performance of Full AFL sample at 650 C. ................................................................................................................... 65 4 1 2 Microstructure of Ni GDC composite AFL on NiO GDC anode ...................................... 6 6 4 13 T he I -V curves for the sample with no AFL, GDC AFL and Ni GDC AFL using 30 sccm wet hydrogen and dry air at 650 C. ............................................................................ 67 4 14 T he impedance spectra under open circuit conditions at 650 C. ....................................... 68 4 15 Schematics for AFL structure for Multilayer approach. ...................................................... 69 4 16 I-V curves of the sample with Ni GDC single layer AFL and multilayered AFL using 30 sccm wet hydrogen and dry air at 650 C ....................................................................... 71 5 1 XRD pattern of Sm0.075Nd0.075Ce0.85O2synthesized using co-precipitation technique. ... 73 5 2 Particle size distribution of Sm0.075Nd0.075Ce0.85O2synthesized using coprecipitation technique and La0.6Sr0.4Co0.2Fe0.8O3. ............................................................ 74 5 3 FE -SEM images of the cross -section view of the electrodes and electrolyte after operation. ................................................................................................................................ 75 5 4 FE -SEM image of the surface of the SNDC electrolyte after operation. ........................... 75 5 5 The I -V characteristics of the prototype SOFC sample with Sm0.075Nd0.075Ce0.85O2electrolyte at various temperatures ranging from 500 to 650 C in 90 sccm of both dry air and wet hydrogen. ...................................................................................................... 77 5 6 Impedance spectrum of the SOFC cell measured at various temperatures ranging from 500 to 650 C. ................................................................................................................ 78 5 7 ASR values at various temperatures ..................................................................................... 79 6 1 Back scattered image showing the cross -section of the GDC ESB composite cathode and Ni -GDC anode after cell testing. .............................................................................................................................. 81

PAGE 12

12 6 2 I-V characteristics of three co -pressed samples at 650 C using 30 sccm of wet hydrogen and dry air. ............................................................................................................. 82 6 3 Backscatter image of the cross -section of the bilayer sample produced by colloidal deposition ............................................................................................................................... 85 6 4 Backscatter images of bilayer samples showing the surface view of the ESB layer produced by colloidal deposition. ......................................................................................... 85 6 5 The I -V characteristics of the samples with GDC single layer and ESB/GDC bilayer electrolyte at 650 C. .............................................................................................................. 86 6 6 As-deposited ESB layer on the GDC electrolyte. ................................................................ 88 6 7 ESB layer after heat tr eatment at 700 C for 4h. After heat treatment, the ESB wasnt fully densified. ........................................................................................................................ 88 6 8 Cross -section image of the cell with the ESB layer by cold PLD after heat treatment at 700 C, 4h. .......................................................................................................................... 89 6 9 XRD patterns of bilayer samples depending on heat treatment conditions. ....................... 90 6 10 I-V characteristics of the samples with GDC sin gle layer and ESB/GDC bilayer electrolyte at 650 C using 90 sccm of air and wet hydrogen. ............................................ 91 6 11 Impedance spectra of the samples with GDC single layer and ESB/GDC bilayer electrolyte at 650 C using 90 sccm of air and wet hydrogen. ............................................ 92 6 12 The bilayer microstructure with the ESB by Hot PLD ....................................................... 93 6 13 XRD pattern of th e bilayer sample. ....................................................................................... 93 6 14 I-V characteristics of the samples with GDC single layer and ESB/GDC bilayer electrolyte at 650 C using 90 sccm of air and wet hydrogen. ............................................ 94 6 15 Impedance spectra of the samples with GDC single layer and ESB/GDC bilayer electrolyte at 650 C using 90 sccm of air and wet hydrogen. ............................................ 95 A 1 SOFC Stack cell development in many countries. ............................................................. 100 A 2 Planar SOFC sample with 6 x 6 cm in size with possible active area of 30 cm2. ............ 101 A 3 Microstructure of planar SOFC sample (a) surface view. (b) cross -section of GDC on NiO -GDC anode. (c) backscattered image of anode cross -section. (d) cross -section of the planar cell with brush painted LSCF GDC composite cathode before test. .......... 102 A 4 I-V characteristics (a) and impedance spectra (b) of a button cell taken from a large stack cell. .............................................................................................................................. 103

PAGE 13

13 A 5 6 cm x 6cm SOFC plana r cell with NiO -GDC anode, GDC electrolyte, GDC AFL, and 30 cm2 LSCF GDC cathode ........................................................................................ 103 B1 I-V characteristics of the sample after 6 months of studying. ........................................... 107 B2 I-V characteristics of GDC powders from three different companies at 650 C. ............ 108 B3 (a) SNDC synthesis by solid state route. (b) SNDC synthesis by co precipita tion route. ..................................................................................................................................... 109 B4 a) Surface of the GDC electrolyte. (b) Back -scattered cross -sectional image of the GDC electrolyte on NiO GDC anode. (c) Surface of the GDC electrolyte with higher density. (d) Cross -sectional image of the GDC electrolyte with higher density. ............. 110 B5 A sample looking acceptable from visually inspection can have poor performance and porous microstructure. .................................................................................................. 111 B6 I-V curves of the copressed sample at various temperatures using 90 sccm dry air and wet hydrogen. ................................................................................................................ 112 B7 (a) Change in the p ower density with increasing H2 flow -rate from 15 to 90 sccm with 30 sccm of air. (b) Change in the power density with increasing air flow rate from 15 to 90 sccm with 30 sccm of H2. ............................................................................ 113 B8 A custom made spray coating setup. ................................................................................... 114 B9 I-V curves from three consecutive three cells showing reproducibility of high performance at 650 C. ........................................................................................................ 115 B10 Effect of ceramabond 1 (a) and ceramabond 2 (b) on the I -V characteristics. ................. 117 B11 Maximum power density of a SOFC without AFL for 700 hours at 550 C. .................. 118 B12 (a) The OCP values and (b) the ASR values at 0.41 V for 700 hours. ............................. 118 B13 Unusually porous region in the GDC electrolyte where othe r area is dense. ................... 120

PAGE 14

14 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy ENGINEERING HIGH PERFORMANCE INTERMEDIATE TEMPERATURE SOLID OXIDE FUEL CELLS By Jin Soo Ahn May 2009 Chair: Eric D. Wachsman Major: Materials Science and Engineering Solid oxide fuel cells ( SOFCs ) are an efficient, fuel flexible energy conversion device, capable of op erating on fuels ranging from natural gas to gasoline, diesel, and biofuels, as well as hydrogen. However, to this point the marketability of SOFCs has been limited by their high operating temperatures. Achieving high power at intermediate temperatures (IT, 500 700 C) would be a significant breakthrough, as low temperature operation would result in better stability and allow for a broader range of material options for the SOFC components as well as the balance of plant, such as stainless steel intercon nects (which are only viable at <700 C) Thus far power densit ies on the order of 2 W/cm2 ha ve been limited to temperatures above 800 C. This dissertation contains a series of works to realize exceptional ly high power at IT ranges. First, improved fabr ication technique s including anode tapecasting and electrolyte spray coating were developed, and a molecular approach to anode functional layer (AFL) was employed using precursor solutions. This newly developed AFL reduced the ASR of a SOFC sample by 60 % and increased the open circuit potential (OCP) by more than 0.1 V resulting in a 140 % increase in power. F urther investigations into this molecular AFL showed that a multilayered AFL can further reduce the ASR and increase the maximum power density. Secon dly, t he potential use of Sm0.075Nd0.075Ce0.85O2 as an electrolyte has been investigated. T he current -voltage (I -V)

PAGE 15

15 performance of the cell exhibits a maximum power density reaching 1.38 W / cm2 with an area specific resistance (ASR) 2 at 650 C with 90 sccm of air and wet h ydrogen. Also, the high OCP achieved at 500 C (0.96 V) as well as the high performance confirmed the via bility of Sm0.075Nd0.075Ce0.85O2a s an alternative electrolyte material The cathode used for this study was La0.6Sr0.4Co0.2Fe0.8O3 (LSCF) Gd0.1Ce0.9O2 (GDC) composite. Finally, Er0.8Bi1.2O3 (ESB) / GDC bilayered electrolyte combined with recently developed ESB / Bi2Ru2O7 (BRO) c omposite c athodes was tested. In this work a maximum power density of 2 W/cm2 was achieved at 650 C with the help of the novel AFL and tapecast anode support s This is the highest power yet achieved in the IT range and I believe redefine s the expectation level for maximum power under IT SOFC operating conditions.

PAGE 16

16 CHAPTER 1 INTRODUCTION 1.1. Solid Oxide Fuel Cells Due to a variety of factors including heavily fluctuating energy costs and environmental concerns, next generation energy sources require a reduced reliance on simple combustion of fossil fuels. Greenhouse emissions from automobiles and power plants have long been believ ed of contributing to global climate change.1 Particulate matter exhausted from buses and larger vehicles can penetrate into the lungs, c ausing respiratory and cardiac disease.2 T here has been a longstanding need to reduce the polluting byproducts of conventional energy generation There is an expanding list of reasons to explore alternatives to conventional energy generation methods SOFCs are expected to play a significant role in the future of energy t echnology. Development of SOFC technology is a m ajor step towards reaching the goal of sustainable energy production. Due to its potential for high efficiency i t not only conserves valuable natural resources but also assists in reducing pollution and greenhouse gas emissions Figure 1 1 Schematic di agram of a SOFC with hydrogen fuel and oxygen

PAGE 17

17 A SOFC is a device that converts chemical energy to electrical energy at elevated temperature. The need for direct combustion is eliminated, giving fuel cells much higher conversion efficiencies than conventio nal thermo -mechanical methods. The main difference between SOFCs and other fuel cells, such as alkaline fuel cell (AFC), polymer electrolyte membrane fuel cell (PEMFC), phosphoric acid fuel cell (PAFC) and molten carbonate fuel cell (MCFC) is the material used for the electrolyte, and consequently the operating temperature ranges and thus plausible market applications. Due to the use of solid oxide electrolytes, higher operating temperatures are required for SOFCs than any other type of fuel cell. Since the SOFC concept was first introduced, much research has been conducted to realize stable and efficient SOFCs suitable for replacement of current energy generation technologies. Generally SOFCs consists of three main parts: anode, electrolyte and cathode (Fig ure 1 1 ). The anode is where the oxidation of hydrogen (2H2 4H+ + 4e-) (or other fuels) occurs, and the cathode is where the reduction of oxygen (2O2+ 4eO2) takes place. Electrons provided from the anode move to cathode through the outer electrica l circuit. The electrolyte is the component where oxygen ionic conduction takes place, and separates the cathode and the anode from electronic contact. Though most SOFCs currently use hydrogen as fuel ,3 SOFCs can utilize deferent reactions depending on different fuels. Example reactions are the follow ing: 2O2(ion) + H2 (gas) 2H2O+ 4e(electron) CH4 (gas)+ O2(ion) CO(gas) + 2H2(gas)) + 2e(electron) CH4 (gas)+ 4O2 (ion) CO2(gas) + 2H2O(gas)) + 8e(electron) Figure 1 2 shows the Solid State Energy Conversion Alliance (SECA) phase 1 requirement for SOFC operations. To fulfill the SECA requirement of 3 k W power output highly efficient SOFC performance is needed.4 A simple calculation shows that 10000 cm2 of active area with a power density of 0.3 W/cm2 would be required to attain this power, which is considered

PAGE 18

18 relatively high performance these days. This huge active area requires incre ased number of stacked SOFC panels interconnectors and sealant materials, and this will subsequently increase the cost and undermine the viability of SOFC application So, realizing highly efficient SOFC is one of the key technological issues in the SOFC field. Figure 1 2. SECA phase 1 requirement .4 It has been demonstrated that the combination of an y ttria stabilized zirconia (YSZ) electrolyte and a La0.8Sr0.2Mn1O3 (LSM) cathode can generate a power density approaching 2 W/cm2 above 800 C.5 Achieving comparable power densities at the IT range (500 700 C), where lower cost, better stability and broad materials choice, such as stainless steel interconnects are possible, would be a significant breakthrough for SOFC s The benefits of IT operation will be discussed in depth in the following section. Combining a GDC electrolyte and LSCF cathode o ffer s operation temperatures below 700 C (Figure 1 3),6 but does not guarantee a high power output. A recently developed next generation cathode, Ba0.5Sr0.5Co0.8Fe0.2O3 (BSCF), produced a power density of 1 W/cm2 at IT ranges.6,7 However, t hus far a power density of 2 W/cm2 has been limited onl y to temperatures above 800 C.5

PAGE 19

19 Figure 1 3 Comparison of maximum power densities of thin film GDC of samaria doped ceria (SDC ) electrolyte cells with various cathodes.6 1.2. Anode Functional Layers The need for higher efficiency and stability in SOFC performance changes simple cathode, electrolyte and anode layer s into multilayered structure s with functional gradients. A n example of this is the ESB/GDC bilayer electrolyte where the ESB layer helps to increase cell OCP and improve ionic conductivity, while the GDC layer protect s ESB from reducing atmosphere s .8 Also, a GDC/YSZ bilayer electrolyte allows for use of LSCF cathode s as GDC as a protective layer between YSZ and Co from the cathode .9 The multiple functions of each SOFC compartment, a cathode, electrolyte and anode, can be shared by multilayers specializing in a d esignated function. These relatively new approaches allowed for high performance or better stability in SOF Cs .10 It is likely that the future fuel cell will unlikely be a simple stack of three component layers the cathode, electrolyte and anode.

PAGE 20

20 One of the most important application s of these multilayer or interlayer approaches in SOFC field is the anode functional layer (AFL, sometimes called the anode inter layer) .11 13 Overall SOFC performance depends on the particle size of anode cermet because particle size dictates the porosity and the TPB length. Most SOFC electrode reactions take place at triple phase boundaries (TPB) between the ga s phase, an electronic conducting phase, and an ionic conducting phase. Increasing the TPB length increases the number of active reaction sites and therefore enhances the performance of SOFCs. Deposited on an electrolyte of a given surface area, electrod es composed of smaller particles result in larger TPB lengths. Therefore, it is clear that use of smaller particles for the electrode will result in better SOFC performance. However, while smaller particles produce higher TPB lengths, they tend to increa se density. It is difficult for a dense anode to provide fuel rapidly to the reaction sites and to remove water molecules efficiently. Frequently sub micron -sized anode powder is mixed with pore former for the fabrication of anode supported SOFCs. To a chieve a sufficient amount of gas channels, an extensive volume fraction of pore former is required. However, it has been reported that pore former additions result in porous anode surfaces and subsequently induces poor quality of the electrolyte deposite d by colloidal methods .14 Also, large pores on the anode surface result in poor contact between the electrolyte and the anode 14 One a pproach to achieve both high open porosity and large TPB length is to deposit a thin anode interlayer using fine powders on an anode made of coarse particles15 prior to deposition of the electrolyte. In this configuration the anode can successfully satisf y the two distinctive roles of an anode: gas transport and electrochemical oxidation of hydrogen. Becau se an AFL removes abrupt change at the anode/electrolyte interface, it allows for gradual changes in properties such

PAGE 21

21 as thermal expansion coefficient and interfacial resistance. Additionally the AFL reduces surface roughness, thus improving the quality of the electrolyte film deposited by colloidal methods. Recently our study on IT -SOFCs has focused on fabrication of planar cells by colloidal deposition of the GDC electrolyte on tape cast Ni GDC anodes without addition of pore formers. In order to fabri cate stacked, large diameter planar SOFCs with doped ceria electrolytes, it is crucial to prepare a dense and thin ceria electrolyte layer on the anode. Tape casting is widely used for planar and multi -stacked ceramic devices because it is a reproducible and efficient ceramic fabrication process. Planar SOFCs have attractive geometries for simple stacking integration. Tape casting of anode supports combined with colloidal deposition of electrolytes is cost -effective and suitable for mass production. To a chieve sufficient anode porosity, the anodes were tape cast with large micron -sized NiO powder mixed with small sub micron-sized GDC powder. The images of NiO and GDC powder used for tapecasting were taken by TEM and present in Figure 4 1. Though sufficient porosity can be achieved by this method, greater anode polarization loss is expected due to the use of large NiO particles. The anode polarization can be reduced by applying an AFL on the anode surface. In this dissertation we propose a method to dispe rse a very fine anode functional layer at the interface of the GDC electrolyte and Ni GDC anode tape using a GDC precursor solution. To the best of my knowledge all AFL studies done so far have been limited to colloidal deposition of fine powder of the s ame composition as the anode bulk.12,13,16 Compared to the conventional AFLs, which utilize colloidal slurry, the precursor solution is advantageous in many ways. D eposition of an AFL with a precursor solution is ba sed on the following assumptions. First, it is easier to form smaller particles at the interface using precursor solution.

PAGE 22

22 Secondly, penetration depths of precursor solutions are deeper than that of colloidal slurries. The greater penetration depth of A FL expands TPB length and provides a better adhesion between the anode and the electrolyte. Additionally, i t is easier to stabilize the solution as well as spray it onto the anode. The effects of the AFL from GDC precursor on the electrochemical performanc e are presented in this dissertation The effect of gas flow rate on the performance of the fuel cell is also present ed Also, further investigat ion on the composition effect of the molecular AFL by co -spraying Ni and GDC precursors were performed It is well known that the Ni GDC composite can extend the TPB length. Therefore I believe the ideal composition of AFL should include Ni. A comp one nt such as the anode, which has two main functions, is not necessarily limit ed to two layers. Holtappels e t al., 17 demonstrated an anode with four distinct layers with gradually changing porosity. A series of experimental and modeling work on recently-developed functionally gradient electrodes (FGE) ,1822 have demonstrated higher performance or better stability using porosity, particle size, or comp osition gradient s Also, Ni e t al., demonstrated by modeling that composition gradient of the anode can expand TPB length and provide better stability due to removal of abrupt change at the interface.21 T his dissertation demonstrates how to further improve these molecular approach AFLs by layer by layer approach. L SCF GDC composites cathodes were used for this study. 1.3. Sm0.075Nd0.075Ce0.85O2as Alternative Electrolyte T he long -term performance degradation as well as the high cost associated with high operating temperature ( 800 ~ 1000 C) have been substantial obstacle s to the widespread commercialization of SOFC technology .23 Reducing cell operating temperatures to the IT range will extend its application domain to residential power and portable de vices However, with

PAGE 23

23 current state -of the art SOFC materials, it is not possible to obtain sufficient power in the IT range. This is due to high ohmic losses and electrode polarization, which have a detrimental effect on device performance and efficiency Thus, there is a need to develop materials that show improved properties in the IT range. Figure 1 4 SOFC depending on operating temperatures. Automobile APU application (300 500 C), Delphi Residential power (500650 C). Stationary (8001000 C), Siemens Westinghouse Due to the high operating temperature of SOFCs, they have been viewed primarily for stationary applications. However, the ability of SOFCs to run on gaseous and liquid hydrocarbons such as gasoline, diesel, and natural gas suggests great opportunity for use in portable energy source applications particularly for the automotive industry, given that the production, delivery and storage infrastructure of these fuels is already established. Figure 1 4 shows various applications for which a SOFC can be used depending on operating temperatures. For localized power plants, IT SOFCs are more advantageous than high temperature SOFCs.24 The leading candidate for automotive fuel cells is the polymer electrolyte membrane fuel cell ( PEMFC ) due to its low operating temperature (80 ~ 120 ever a PEMFC requires extremely pure hydrogen f uel.25 So, for example, in order to utilize hydrocarbon fuels for a car equipped with a stack of PEMFC s an autothermal/steam reformer, a high temperature shift reactor, a low temperatu re shift reactor, and a preferential oxidation system would be required in order to produce pure hydrogen from these largely carbon-based fuels .26 By contrast SOFCs can

PAGE 24

24 directly utilize hy drocarbon fuels without additional devices to pre -process fuels. Fewer parts mean greater efficiency, lower costs and a smaller overall package .26 There is one major drawback of SOFCs that has hurt their viability in this market their high operating temperatures. This requires extended startup times a period of fuel burning which is needed to reach these operating temperatures. High operating temperatures are also responsible for sealing pro blems and expensive interconnect and balance of plant materials for SOFC stacks. Additionally, high temperature operation can induce thermal stresses at electrolyte -electrode interfaces, as well as cause inter diffusion between cell components. Most of the se problems will be solved if the operating temperature can be lowered to around 500 C. The major benefits of IT SOFCs are listed below: Low heat low cost. Low thermo -mechanical stress in assembly resulting in longer service life. Performance stability due to reduced degradation of electrodes and electrolytes Increased range of materials that can be used including sealants. Fast start up time At these reduced temperatures, conventional SOFCs using YSZ electrolytes have very high ohmic polarization. Usi ng an oxide with higher conductivity can reduce the ohmic polarization. One such electrolyte material is the doped ceria system, which has higher ionic conductivity than has significantly reduced ohmic polarization losses at temperatures between 500 C and 600 C .27 According to Figure 1 5 GDC and ESB are the leading can didates for electrolyte materials at the IT ranges.27 Higher conductivity of the electrolyte produces low ohmic polarization, and it can be understood by a simple equation; Thickness Thickness ohmT i T i ASR i ) / ( *. (1 1 )

PAGE 25

25 w here i, ASR, Tthickness correspond to the ohmic overpotential, current densi ty, area specific resistance, resistivity conductivity, and electrolyte thickness, respectively. In the equation above, the ohmic overpotential is inversely proportional to the conductivity. 1.0 1.1 1.2 1.3 1.4 1.5 700 600 500 400 1.0 1.1 1.2 1.3 1.4 1.5 700 600 500 400 YSZ YSZ GDC SNDC SNDC ESB ESB DWSB DWSBLn T (S -cm1K)0 -1 1 2 7 6 5 4 3 1000/T (K1) Figure 1 5 Ionic conductivities of various oxides including t he conductivity values of YSZ,27 GDC,28 SNDC29 and dysprosium and tungsten stabilized bismuth oxid e (DWSB).28 Th ere is one other important advantage of using faster ionic conductor s specifically when using hydrocarbon fuels. Faster ionic conductor s can provide larger amount s of oxygen from the cathode to the anode reaction sites. This is an import ant aspect for the oxidation of heavy hydrocarbon fuels or liquid alcohol fuels as larger amount of oxygen can reduce coke formation According to Wagner s theory, increase d ionic conductivity increases oxygen flux.30 2 2 2ln 16 ) (ln ln 2 2 O P P ion el ion elP d d F RT O jO O (1 2 )

PAGE 26

26 wh ere R is the gas constant, F is Faraday s constant, d is the electrolyte thickness, el is electronic conductivity and ion is ionic conductivity. The use of faster an ionic conductor such as GDC in the composite anode and cathode is also beneficial. As will be discussed in C hapter 6, a cathode containing a faster ionic conductor exhibits smaller ASR compar ed to a similar cathode containing a slower ionic conductor.31 To date, the simple ox ygen reduction reaction on a pure P t cathode is not fully understood.32 This is true of fuel oxidation on the anode as well One of the main obstacles of anode research is that the mechanism of fuel oxidation at the anode even for simple hydrogen fuel and conventional cermet anode s is not understood yet.3 The anode reaction is even less understood than the cathode reaction because of the higher level complexity (involving removal of the water byproduct and mixed conductivity of certain ionic conductors in reducing atmo sphere ). The mechanistic complexity will dramatically increase for methane and heavi er hydrocarbons. One of the most significant mechanistic studies on the anode was conducted by Mogensen work but the model could n o t incorporate the fact that GDC can have electronic conductivity at reducing conditions and hence, reaction site s can exist at the GDC surface, and failed to explain the reported properties of the anode.33 However, a simpler model exists relating the electrode polarization to the ionic conductivity of the ceramic phase of the electrode Theoretical background and predictions in cathode design can be found in practical models develope d for composite cathodes. Equation 1 3 is a simple model developed for the cathode that has both electronic and ionic conducting phases34 and proved useful for electrode design.35 kr t rk Ri/ 2 coth 2 /2 (1 3 )

PAGE 27

27 where Ri is the cathode resistance, t is the electrode thickness, r is the ionic resistivity of the ionic conducting phase, k is an intrinsic interfacial resistance characteristic of the interfaces between the ionically and electronically conducting phases, and r is the pore radius. According to this model the interfacial resistance of the cathode and the anode with YSZ as an ionic conductor can be suppressed by replacing YSZ with GDC, which is a faster ionic conductor. The reduction in cathode polarization by replacing G DC with ESB will be discussed in Chapter 6. 1.0 1.1 1.2 1.3 1.4 1.5 0.0 0.5 1.0 1.5 700 650 600 550 500 450 400 1.04 1.08 1.12 1.16 1.2 1.4 1.6 1.8 700 675 650 625 600 Log[ grainT(S.cm-1.K)]1000/T(K-1) T(oC)1000/T(K-1) Sm0.075Nd0.075Ce0.85O2Sm0.05Nd0.05Ce0.90O2Gd0.10Ce0.90O2-Log[ grainT(S.cm-1.K)]T(oC) Figure 1 6. Conductivity values of the ceria system with gadolinium single dopant and samarium and neodymium double dopant.29,36

PAGE 28

28 Electrolytes with high ionic conductivit y are critical for the development of SOFC s that can successfully generate reasonable power at IT range .37 In recent years, doped ceria electrolytes have emerged as a potential electroly te candidate material due to its high ionic conductivity .23,29 A mong doped ceria materials, Gd0.1Ce0.9O2was widely accepted to exhibit the highest ionic conductivity .36 However, Omar et al., have shown that co -doping based on Sm3+ and Nd3+ leads to further enhancement in the ionic cond uctivity in ceria systems .29 Optimization of dopant concentration in Smx/2Ndx/2Ce1 xO2system resulted in the development of Sm0.075Nd0.075Ce0.85O2 which exhibits 30 % higher grain ionic conductivity than that of GDC at 550 C in air ( Figure 1 6 ). In the present work, the potential of Sm0.075Nd0.075Ce0.85O2as an alternative electrolyte is further investigated by testing its performance in a SOFC. An anode supported SOFC was fabricated, with Sm0.075Nd0.075Ce0.85O2 as the electrolyte. At thi s point, the effect of incorporating Sm0.075Nd0.075Ce0.85O2 into the anode and cathode will be relegated as future work. The performance of these SOFC s at IT is discussed in C hapter 5 1.4. ESB/GDC Bilayer Electrolyte s for IT-SOFC Operating temperatures can be lowered by reducing the resistance of each cell component : the anode, the cathode, and the electrolyte. Reduced electrolyte resistance has been a topic of enduring study in the SOFC field. One method for reducing electrolyte resistance is to replac e the conventional yttria -stabilized zirconia (YSZ) electrolyte with materials exhibiting higher ionic conductivity. Two of the leading candidates are gadolinia doped ceria (GDC) and erbium stabilized bismuth oxide (ESB). Based on reported conductivit y va lues, 10 m thick electrolyte s of YSZ, GDC and ESB have area specific resistance s (ASR) of 1.259, 0.143 and 0.037 cm2, respectively, at 500 C,28 that is, an order of magnitude drop in electrolyte ASR can be achieved by changing from YSZ to GDC and fr om GDC to ESB

PAGE 29

29 However, these new materials do have drawbacks F or ceria based electrolytes, the reduction of Ce4+ cation s into Ce3+ allows for electronic conduction via a small polaron hopping mechanism yielding a narrow electrolytic domain.38 Thus, under reducing conditions (such as that f ound at the SOFC anode 2OP ~ 1026), GDC will exhibit a degree of electronic conduction which is negligible below 500 C, though .39 This electronic co nduction reduces the cells ionic transference number and hence its Nernstian potential and maximum power output. Second, for ESB t he weak metal -oxygen bonds which make it one of the fastest ionic conductors known, ironically also lead to reduced thermody namic stability at low 2OP .40 Figure 1 7 Bilayer electrolyte concept with relative thickness ratio to control interfacial oxygen partial pressure ( 2OP ).40,41 These issues can be resolved by utiliz ing a bilayer electrolyte with GDC on the fuel side and ESB on the air side. In this configuration, the ESB layer will block electronic conduction from the GDC layer, and a sufficiently thick GDC layer would prevent decomposition of ESB. The interfacial 2OP can then be controlled by varying the thickness ratio of the component layers (Figure 1 7 ).

PAGE 30

30 Previously it has been demonstrated that the OCP can be increased with a bilayer electrolyte over a wide range of temperatures, and that it h as good long term stability.8,41,42 It should be noted that the increase in OCP is a function of the ESB layer densit y and thickness .41,42 Fabrication of a thin and dense ESB and GDC layer s is key for increasing OCP In addit ion to increasing cell OCP, the bilayer electrolyte can reduce the effective total electrolyte ASR. By modeling the bilayer electrolyte as two resistors in series, electrolyte ASR can be described as equation (1 4 ), is the thickness ratio of ESB to GDC and i is the ionic conductivity of each material Park et al. demonstrated that the ASR can be decreased by utilizing a bilayer electrolyte .43 ) 1 ( *GDC ESB GDC e ElectrolytL ASR (1 4 ) An additional advantage of having an ESB layer on the air side of the cell is that as frequently reported the surface of bismuth oxide -based electrolytes is active for the adsorption of oxygen species. It is believed that bismuth strongly enhances the surface oxygen exchange rate .44 46 This is evidenced by the factor of 103 larger sur face oxygen exchange rate for bismuth oxide based solids compared to YSZ solids. In addition, bismuth-based electrolytes have been reported to be catalytically active towards oxygen dissociation and charge transfer .47,48 It is thus possible that the use of ESB at the cathode/electrolyte interface may result in reduced cathode activation overpotentials. This phenomenon is discussed further in chapter 6. However, because of their weak metal -oxide bonds, bismuth oxide based electrolytes are highly reactive towards many conventional cathode systems. Recently, some composite systems which are chemically compatible with bismuth oxide based electrolytes have been developed. Ag bismuth oxide show significantly lower values of ASR compared with other frequently studied cathode systems such as LSCF GDC, and have been modified to show relatively good

PAGE 31

31 long -term performance.49,50 Even lower ASR values were achieved from microstructurally optimized composites of Bi2Ru2O7 (B RO7) and ESB .51,52 While the b ilayer electrolyte concept has proven sound, bilayer cells exhibiting high performance have not been reported to this point. The previous results wi th ~ 800 m thick ESB/samaria doped ceria (SDC) electrolyte, Pt anode and Au cathode achieved a higher OCP and lower ASR resulting in a 33% increase in power density over a single layer SDC electrolyte under the same conditions .41 However, the power density achieved was only 50 mW/cm2 at 800 C. Similarly Leng et al fabricated SOFCs utilizing yttria -stabilized bismuth oxide ( YSB ) / GDC bilayer electrolyte s with total thicknesses above ~ 800 m and Pt electrodes .53 Their bilayer study also demonstrated higher OCP and power density than with a single layer GDC electrolyte but the power density did not exceed 20 mW/cm2 at 600 C. The poor performance is due in part to the difficulty in deposit ing thin and dense ESB layers. In addition, low resistance ESB -compatible cathodes have only recently been developed The SOFCs util izing single layer GDC electrolytes have made much progress toward achieving high power densities .54,55 T his dissertation demonstrate s that a bilayer ESB/GDC electrolyte can achieve exceptionally high power densities when combined with high performance GDC electrolytes as well as recently developed BRO7 -ESB composite cathodes. Various fabricat ions were employed for deposition of dense ESB including pulsed laser deposition (PLD)

PAGE 32

32 CHAPTER 2 BACKGROUND 2.1. Fundamental Mechanisms of SOFC The simplest way of conceptualizing an SOFC operating with hydrogen and oxygen is that the cell burns hydroge n and converts the energy generated from this process into useful work. O H O H2 2 22 2 (2 1) The main difference between a SOFC and a conventional combustion engines is that in SOFCs, hydrogen m olecules are burn ed by electrochemical oxidation reactions. So, instead of generating heat, a SOFC generates electrical energy. The benefit of electrochemical oxidation of fuel is that the efficiency in energy conversion is not limited by the Carnot effici ency limit where there is a heat loss.56 When an SOFC is at equilibrium, in other words at zero current, its potential is called open circuit potential (OCP). The OCP can be determin ed by Nernst potential (E). E t OCPi (2 2) where, ti is the transference number for ionic conduction. The following is a derivation of Nernstian potential relevant for SOFC systems. eK RT G G ln (2 3 ) where G is G ibbs free e nergy, G is Gibbs free energy at standard state, R is the gas constant, T is the absolute temperature and Ke is equilibrium constant. Since, nFE G (2 4 ) where n is number of electr ons per mole of oxygen molecules, F is Faraday s constant and E is potential, the potential of the cell can be described by anode O cathode OP P F RT E E2 2ln 4 (2 5 )

PAGE 33

33 Here, G or E should be regarded as zero for SOFC rea ctions since the driving force for a SOFC is the chemical potential difference of oxygen molecules across the electrolyte. At standard state, chemical potential of oxygen molecules at the cathode and anode can be treated as unity. So, the potential of the SOFC is now, anode O cathode OP P F RT E2 2ln 4 (2 6 ) Equation (2 6 ) shows that the potential is a function of temperature and chemical potential (partial pressure difference) of oxygen molecules Using this equation one can easily calculate the theoretical potential. The following is a calculation of the theoretical potential of a SOFC operating at 650 C with the anode partial pressure of oxygen set by a bubbler at room temperature. When H2Oliq and H2Ovap is at equi librium ,57 732 19 log 65 4 2900 log2 T T PO H (2 7 ) According to this equation, at room temperature, a bubbler is sending humidified hydrogen gas to the anode with O HP2 = 0.03 and we can assume that 2HPis close to 1 for simplicity of the calculation since 2OP is negligible The 2OP can then be calculated using the equilibrium constant and the Gibbs free en ergy of a water forming reaction. The water forming reaction used for this calculation, the equilibrium constant for the reaction, and the Gibbs free energy in terms of temperature are as follows ,57 O H O H2 2 22 1 (2 8 ) 2 / 12 2 2O H O H eP P P K (2 9 ) Since G = 247500 + 55.85T (J ) for E quation (2 8 ), and G = RT ln Ke,57

PAGE 34

34 3144 8 85 55 3144 8 247500 ln T Ke (2 10) Since 2HP is close to 1 2 / 12 2O O H eP P K (2 1 1 ) 22 2 e O H OK P P (2 1 2 ) If the operating temperature of a SOFC is 650 C, 1110 2278 1 3144 8 85 55 923 3144 8 247500 ln eK (2 1 3 ) Since Ke is calculated, we can calculate the anode 2OP 2610 9697 52 OP (2 1 4 ) The Nernst potential at 650 C is then V P P F RT Eanode O cathode O1239 1 10 9697 5 21 0 ln 3399 96485 4 3144 8 ln 4262 2 (2 14) Th is equation shows how increases in operating temperature can decrease the OCP. As T increases, the RT term increases, but the an o de 2OP rapidly increases and the overall OCP decreases. 2.2. Potential Losses of SOFC The Nernstian potential above assumes no change in potential as a function of current drawn. However, t he performance of a SOFC can be measured by the voltage output as a function of current density drawn and t hermodynamics can no longer predict the potential in this case If current is drawn from a SOFC, the voltage output will deviate from the equilibrium potential predicted by thermodynamics. The voltage will be lowered by various mechanisms

PAGE 35

35 depending on the amount of current draw n. Figure 2 1 shows a typical I -V and a power density curve ( which is a voltage multiplied by current densit y) The measured vo ltage, V, can be written as, . conc ohm act o conc ohm act oE IR IR IR E V (2 15) w here Ract ., Rohm and Rcon c. stand for activation, ohmic and concentration polarizati on resistances, respectively and which can be described as a potential loss due to each polarization mechanism Figure 2 1. A typical current -voltage characteristics of a SOFC .25 Activation polarization ensues whenever reacting chemical (inc luding electrochemical) species are involved. An activation energy barrier must be overcome in order for the reaction to proceed. This results in a potential drop which may be regarded as the extra potential necessary to overcome the energy barrier of the rate -determining step of the reaction, and is related to the current passing through the cell by the phenomenological Butler -Volmer relation.58 Ohmic

PAGE 36

36 polarization is caused by resistan ce to the flow of ions in ionic conductors and electrons in electronic conductors, and by contact resistances between cell components. Concentration polarization occurs when reacting species are supplied to reaction sites slower than they are consumed, or when reaction products are not removed fast enough so that they block the reaction sites. Note that the slope of current density vs. voltage curve is the ASR. 2.2.1. Activation Polarization The energy necessary to overcome a reaction at an electrode increases exponentially with curren t This was first observed by Tafel in 1905.59 The relationship b etween activation polarization and activation energy is not yet understood, but an empirical equation was developed and used in the form of the Butler -Volmer equation RT F n iRT F n i iact c o act a o exp exp (2 16) w here io is the exchange cu rrent density, the apparent transfer coefficient, n the number of electrons transferred and F is Faraday s constant. For hydrogen electrodes the apparent transfer coefficient, 0.5. For oxygen electrodes, varies from 0.1 to 0.5. However, for SOFC reaction s ca n be treated as 0.5 assuming change in makes little impact.60 Interesting ly, for SOFCs, when the number of electrons, n is 1 good fits to the Butler -Volmer equation are obtained ,61 while most SOFC reactions obviously show n values of either 2 or 4 depending on the convention used for writing the chemical equations involved. The Tafel equation can be used t o look at the electrochemical reaction at an electrode with over potential, in terms of current .56 Th e Tafel equation can be viewed as another expression of the Butler -Volmer equation using the constant A. o oi i F n RT i i A log log (2 17)

PAGE 37

37 The constant A is higher for an electrochemical reaction which is slower. The exchange current density, io, is higher if the reaction is faster. Now with the Tafel equation, one can derive the actual cell voltage, with respect to E and oi i A E V log (2 18) Thus, reducti ons in activation polarization (that is, reductions in V in the equation) can only be achieved by increasing the exchange current density or by increasing T. Therefore increasing the actual number of reactions sites (since io is normalized with area) and using catalytically more active materials can reduce the activation polarization. 2.2.2. Ohmic Polarization The ohmic polarization was already described in C hapter 1. Here, note that the equation only deals with the resistance caused by ionic migration th rough electrolytes, but the anode and cathode can also exhibit substantial ohmic polarization.62 Thickness Thickness ohmT i T i ASR i ) (. (1 1 ) O hmic polarization is simplest type of loss to understand and it is clear that the deposition of a thin electrolyte or the use of materials exhibiting higher ionic conductivity can reduce the ohmic polarization. Additionally, to reduce the ohmic re sistance from electrodes, deposition of thick electrodes to remove sheet resistance, or use of current collect ors is necessary .52 2.2.3. Concentration Polarization When current is drawn from a SOFC, consumption of gas at electrodes cause s a decrease in partial pressure of that gas This phenomenon changes the potential across the electrolyte. By m od ify ing the equation for Nernst potential, one can express the change in potential as a function of the change in partial pressure,

PAGE 38

38 1 2ln 4 P P F RT V (2 19) The limiting current, il, is the current at which fuel is being consumed as quickly as it is being supplied At this point, the cell voltage will be zero. Rewriting the equation (2 19) gives, li i P P 11 2 (2 2 0 ) Substituting the equation (220) into equation (2 19), li i F RT V ln 4 (2 21) E quation (2 21) represents the potential loss es due to mass transpor t phenomena A m ore rigorous concentration polarization model for the anode and cathode were developed by Kim et al.,63 and Zhu et al.61 2.2.4. Internal Leakage Current The use of a mixed ionic electronic conducting (MIEC) electrolyte such as GDC will additionally lower the potential by internal leakage currents through the electrolyte. T he pot ential across a MIEC electrolyte is given by the following equation, anode O cathode O iP P F RT t OCP2 2ln 4 (2 22) where ti is the ionic transference number. It is often mentioned that the potential drop caused by interna l leakage currents leads to total efficiency drops This point of view arises from calculating the efficiency of a SOFC using the following equation: 100 H G Efficiency (2 23)

PAGE 39

39 where, H is enthalpy of formation of the fuel and G is Gibbs free energy of the fuel oxidation reaction. G is a function of the potential of the cell ( G = -nEF ). However, using this equation for measuring the overall SOFC efficiency can be misleading since a ccording to the equation above, the maximum efficiency of a SOFC using hydrogen fuel and oxygen can go above 80%.56 In other words, the efficiency loss predicted by t hermodynamics is only 20%. However, overall performance of SOFC is strongly affected by irreversibility such as polar ization mechanisms. An electrolyte with larger amount of leak age current can still be considered for use as an electrolyte if it has a very high ionic conductivity which can compensate voltage loss es caused by internal current s 2.2.5. Combination of Each Polarization Figure 2 2. Separation of polarizations using a model developed by Yoon et al .64 (a) Sc stabilized zirconia (ScSZ) electrolyte based SOFC. (b) YSZ electrolyte based SOFC.

PAGE 40

40 In this section, a combination of each polarization mechanism is present ed to better understand the I -V behavior of a SOFC, which is a main characterization tool of this dissertation. As seen in Figure 2 2, the I -V curve represents a potential drop from the theoretical OCP by various mechanisms. For an electrolyte with ti approaching unity, such as YSZ, combining equations developed for each polarization mechanism is relatively easy compared to MIEC electrolytes. A simple way of expressing pote ntial drop is to combine simple equations above in this dissertation. l n o n ni i i B i i i A ASR i i E V 1 ln ln ) ( (2 23) where, in is the internal leakage current density, A is the constant in E quation (2 17), and B is the constant in E quation (2 2 1). A better expression for potential drop in YSZ based SOFCs, which can be used for fitting to an actual I -V curve, were demonstrated by Yoon et al.64 They demonstrated a s imple mathematic conversions of each equation lead to a equation where cell voltage is a function of applied current.64 It should be noted here that their work is only valid for pure ionic conductors. A more rigorous performance modeling study has been discussed by Amphlett .65,66 The most advanced and sophisticated equation, which can predict the I -V behavior of MIEC electrolyte based SOFCs, w as developed by Duncan et al., where the cathode and anode activation polarizations and the cathode and anode ohmic polarizations were separated.62 The I -V characteristic of the Full GDC AFL sample at 650 C in Figure 4 5 was analyzed using his model and it can be found in his publication.62

PAGE 41

41 CHAPTER 3 EXPERIMENTAL PROCEDURES 3.1. SOFC Fabrication 3.1.1. Anode Tapecasting The Ni GDC anode supports were synthesized by tape casting a mixture of NiO (Alfa Aesar, 99% purity, CAS 1313) and GDC (Rhodia, LOT H 050708) powders. The powder mixt ure contains 65 weight % of NiO. Both raw oxide powders were weighed (to obtain 150 g batch of powder mixture), and were mixed in toluene (Fisher Scientific) (13.6 wt %) and ethanol (Fisher S cientific) (11.3 wt %). A 1 wt % of S olsperse (Air Products and Chemicals) was added to the slurry as dispersant. Further, a mixture of di n butyl p hthalate ( DBP, Alfa Aesar) (4.3 wt %), and polyethylene glycol ( PEG, Fisher Scientific) (0.8 wt %) was added as plasticizer, while polyvinyl butyral ( PVB, Acros Organics) (8.4 wt %) was added to the slurry as binder. Mixing was performed using ball milling using zirconia ball media for 24 h The resulting slurry was then transferred to a vacuum chamber, for de airing. During this process, the slurry was constantly magne tically stirred to avoid any solidification. The slurry was then tape -cast using a tape cast er (Procast from DHI, Inc.) on G10JRM Silicone coated Mylar (R. E. Mistler, Inc.), with a caster speed of 10 cm /min The substrate temperature was maintained at 8 0 C during this process with a doctor blade gap was 800 m The NiO GDC tape was subsequently dried at 100 C for 2 h in an oven (~ 5 00 m after drying drying) Circular green tapes with 32 mm diameter were then punched out from the tape, and partially sintered at 900 C for 2 h 3.1.2. SNDC Electrolyte Dep osition The co precipitation technique was used to synthesize phase pure powder of Sm0.075Nd0.075Ce0.85O2 One of the main objectives of using wet chemical routes is to obtain fine particle size powder to enhance the sintering kinetics so that highly de nse electrolyte s can be

PAGE 42

42 synthesized at lower sintering temperature and time. Highly pure cerium nitrate (Ce(NO3)3.6H2O, Alfa Aesar, 99.99%), samarium nitrate (Sm(NO3)3.6H2O, Aldrich, 99.999%), and neodymium nitrate (Nd(NO3)3.6H2O Alfa Aesar, 99.9%) were u sed as starting raw materials. They were weighed in the stoichiometric proportions and dissolved in de ionized water to produce an aqueous solution. Excess ammonia solution (Acros Organics, 2830% of NH3 solution in water) was added to the stirred soluti on by 5 ml with 30 minutes time interval. T he pH was stabilized at 12 to make sure all the cations are precipitated. The addition of ammonia solution resulted in the formation of yellowish brown color precipitate. The precipitate was filtered, and then s ubsequently dried at 80 C for 12 h The agglomerated powder was then ground to fine particles using mortar and pestle. The powder was then calcined at 900 C for 10 h in air. After calcination the agglomerated powder was ground using mortar and pestle to fine A n Intel CPS 120 diffractometer was used to obtain the X -ray diffraction (XRD) pattern of the Sm0.075Nd0.075Ce0.85O2 using Cu K radiation. A monochromator crystal was used to sep arate out Cu K from the incident X ray beam. For the deposition of the SNDC electrolyte on the anode support, the Sm0.075Nd0.075Ce0.85O2 powder was ball milled for 24 h using zirconia ball media in ethanol medium with Solsperse as a dispersing agent For a typical 10 g batch of Sm0.075Nd0.075Ce0.85O2powder, 26.3 wt % of ethanol and 1 wt % of Solsperse were used. After 24 h of ball milling, 10 wt % of PVB and 3 wt % of DBP were added into the slurry. The ceramic slurry was again ball milled for another 24 h using the same ball media Before the deposition, ceramic slurry was sonicated for 10 minutes. The ceramic slurry of Sm0.075Nd0.075Ce0.85O2was then deposited twice onto the anode (NiO GDC) surface using a pipette I n between coats 30 minu tes of drying

PAGE 43

43 step was applied The electrolyte deposited anode samples were subsequently, heat treated at 120 C for 5 h in an oven. The bi -layered structure of the electrolyte and anode was then cosintered at 1550 C for 4 h using a 3 C / min ramp rate in air. 3.1.3. GDC Electrolyte Deposition For deposition of the GDC electrolyte, the GDC powder from Anan was ball milled for 24 h with Solsperse in ethanol. PVB and DBP were added after the first ball -milling step and ball milled using zirconia ball med ia for another 24 h For deposition of the GDC electrolyte, the GDC slurry was sprayed onto the anode surface using a paint spray gun (Excell) The setup for spray coating can be found in APPENDIX B. The samples were moved to a vacuum oven and heat treated at 120 C for 5 h The bilayered structure of the electrolyte and anode was cosintered at 1450 C for 4 h using a 3 C / min ramp rate in air. 3.1.4. LSCF Composite Cathode The cathode ink was prepared by mixing La0.6Sr0.4Co0.2Fe0.8O3(Praxair Specialty Ceramics, 99.9% purity) and the GDC powder in a 1:1 weight ratio, using mortar and pestle. A lpha terpiniol (40 wt % of the powder) was added as a solvent, and DBP (6.7 wt % of the powder) was used as a plasticizer. Further, PVB (2.3 wt % of the powder) dissolved in ethanol (21 wt % of the powder) was added as a binder. After mixing and grinding the cathode ink in mortar and pestle for 1 h, the ink was brush-painted evenly onto the electrolyte (deposited on NiO/GDC tape). The first layer of cathode ink was dried in an oven at 120 C for 1 h. The second layer of the same cathode ink was then evenly brush painted on top of the first layer. After applying the cathode on top of the electrolyte, samples were then fired at 1100 C for 1 h As a current collector, Pt paste (CL11 5349, Heraeus) was brush-painted onto both electrodes

PAGE 44

44 along with a Pt mesh and Au connecting wires. Samples with current collectors and connecting wires were then finally heat -treated at 900 C for 1 h. 3.1.5. AFL Deposition For the GDC functional layer, 1 M of GDC precursor solution w as prepared in an ethanol solvent. Upon complete dissolution of the precursors, the solution was transferred to a paint spray gun (Excell), sprayed onto the anode substrate, and h eat -treated at 900 C for 1 h For the Ni GDC functional layer, Ni precursor solution was prepared with same concentration as GDC solution. Ni precursor solution and GDC precursor solution was sprayed in turns to deposit composite Ni GDC composite AFL. Deposition rate was controlled by spray time which was maintained to be same to deposit same amount of Ni and GDC. The composite AFL was heat treated at 900 C for 1 h For the multilayered AFL, firstly, the same procedures for Ni GDC composite AFL was don e, and the sample was placed on a hot plate (Corning). Upon desired amount of GDC precursor was deposited, the sample was heat treated at 900 C for 1 h To quantify deposition rate, mass change after spraying and heat treatment was measured for all thre e AFL samples. 3.2. Fabrication of ESB/GDC Bilayer Electrolytes The fabrication of SOFCs with single GDC electrolyte is improved sequentially starting from a co pressing procedure to tapecasting anode support, to anode functional layer (AFL) on tapecast a node. Various ESB depositions on sintered GDC are evaluated to form a dense ESB layer from screen printing to colloidal drop coating, to cold PLD to hot PLD. Note that there is no terminology as cold PLD and hot PLD. In this study these terms were used to distinguish PLD without substrate heating (cold PLD) and with substrate heating (hot PLD).

PAGE 45

45 3.2.1. Co -pressing Procedure The first method used to fabricate bilayer cells involved co pressing fine GDC powders onto composite NiO GDC anode supports. The anode supports were prepared by uniaxial pressing a well -mixed powder of NiO (Alpha Aesar), a very fine GDC (Rhodia), and a PVB (Alfa Aesar) binder (3 wt%) in a 1 1/8 cylindrical die at ~14 MPa. Next, ~0.35 g de agglomerated GDC powder was added to the die, carefully and uniformly spread across the anode substrate surface, and pressed at ~42 MPa. The pellets were then pressed isostatically at 250 MPa, and fired at 1450 C for 4 h using a 3 C/min heating rate and a 400 C, 1 h binder burnout step. The resulti m ). An ESB layer was then screen printed on top of the sintered GDC surface of one cell. The ESB powder used in the screen printing ink was prepared by the solid state technique. Erbium oxide (Alfa Aesar) and bi smuth oxide (Alfa Aesar) powders were weighed in stoichiometric amounts, ball milled for 24 h and calcined at 800 C for 10 h to yield Er0.4Bi1.6O3. ESB inks were then prepared by mixing ~1 g of the prepared ESB powder (slightly wetted with an ethanol solvent) with alpha terpiniol (Alfa Aesar), DBP ( Alfa Aesar), and a solution of 10 wt% PVB in ethanol using a 3:1:2 volume ratio in a mortar and pestle until the ink reached a honeylike consistency. The ink was then screen printed onto the GDC side of the cel l and fired at 890 C for 4 h. For cathoding, two different composite materials were used : La0.6Sr0.4Co0.2Fe0.8O3(Praxair) -GDC and a low ASR BRO7 ESB. Synthesis of BRO7 powders and optimization of BRO7 -ESB composite cathodes have been described previo usly ,52 and involve sedimentation to narrow the size and distribution of pa rticles which were prepared by solid state synthesis. Cathode inks for both materials were prepared in a similar manner to that described above for the ESB bilayer electrolyte screen printing ink. For LSCF -GDC composites, a 1:1 weight ratio was

PAGE 46

46 used, and for BRO7 ESB, the optimized c omposition52 was used. The previous section de scribes application of LSCF GDC composite cathode in details. BRO7 -ESB cathodes were brush painted onto the electrolyte surface, dried at 120 C, and a second layer was applied. BRO7 ESB cathodes were fired at 800 C for 2 h. A Pt current collector (Heraeu s) was paint brushed onto both electrodes of cells utilizing LSCF GDC composite cathodes, while a pure BRO7 current collector (also prepared from the ink synthesis method described above) was applied to both electrodes of cells utilizing BRO7 ESB cathodes. Finally, lead wires and meshes were attached to the electrodes (using Pt paste on the anode side and the same current collector ink used on the cathode side) and fired in-situ with the testing apparatus. Three types of samples were studied using the co-pr essing technique : two utilizing a single -layer GDC electrolyte with either an LSCF GDC cathode or a BRO7 ESB cathode, and one utilizing a bilayer ESB/GDC electrolyte with a BRO7 ESB cathode. 3.2.2. Colloidal R oute The second stage of SOFC fabrication invol ved GDC spray coating on tape -cast anodes followed by ESB colloidal drop coating The a node support was prepared by tapecasting as previously described. Anode tapes were presintered at 900 C for 2 h and GDC electrolytes were deposited by spray coating. Details on anode tapecasting and GDC spray coating were described by Ahn et al ,54,55,67 and previous section of this dissertation. For colloidal ESB deposition, in order to increase the yield of fine ESB powders, a co precipitation route was employed. Pure Ce Nitrate and Er Nitrate were used as starting raw materials. They were weighed in the stoichiometric proportions a nd dissolved in 70 % nitric acid to produce a solution. Excess ammonia solution (Acros Organics 28 30 % of NH3 solution in water) was added to the stirred solution to increase the pH value to 12. The addition of ammonia

PAGE 47

47 solution resulted in the formation of yellowish brown color precipitate. The precipitate was filtered, and then subsequently dried at 80 C for 12 h. The agglomerated powder was then ground to fine particles using mortar and pestle. The powder was then calcined at 900 C for 10 h in air. For the colloidal slurry, ESB powder was ball milled with Solsperse (Air Products and Chemicals), PVB, and DBP in ethanol for 24 h, and drop -coated onto the sintered GDC electrolyte surface. The drop coating was repeated until a desired thickness was achieved. The ESB layer was sintered at 890 C for 4 h using a 400 C 1h binder burnout step, and a 3 C/min ramp -rate. The same BRO7 -ESB composite cathode as described above was applied onto the ESB surface A LSCF -GDC composite cathode was applied to a tape-cast cell utilizing a single GDC layer as a baseline. I -V measurement s were carried out on the two different samples in the same way with a Solartron 1287. 30 sccm of dry air and 30 sccm of wet hydrogen were supplied to the cathode and anode side, respectively. For the bilayer sample 90 sccm of wet hydrogen was applied to examine the effect of gas f low -rate on I -V maintaining 30 sccm of air. ASR values were calculated from the initial slopes of the I -V curves due to their nonlinear nature at most current regions. 3.2.3. Cold PLD PLD is a thin film deposition technique using evaporation of plasma va por from a solid target by impinging the target with high power pulsed laser. The evaporated plasma, which is referred as plume, migrates to the substrate and deposit an ultra thin layer. Use of plasma plume in PLD requires ultra high vacuum usually but oxygen filling is frequently used for deposition of fully oxidized oxide layers. PLD of ESB on spray coated GDC was used in order to enhance ESB layer density. Figure 3 1 shows the schematics of the PLD system used. The a node support was prepared by tape

PAGE 48

48 ca sting as described above. The important addition in this stage of fabrication is an AFL deposited between the GDC electrolyte and the Ni GDC anode. The AFL was prepared by spraying GDC precursor onto a presintered anode and heat treated at 900 C, 1 h. Substrate GDC/NiO ESB target KrF Excimer Laser Vacuum chamber Plasma plume UV transport window Substrate GDC/NiO ESB target KrF Excimer Laser Vacuum chamber Plasma plume Substrate GDC/NiO ESB target KrF Excimer Laser Vacuum chamber Plasma plume UV transport window Figure 3 1 PLD system with KrF excimer laser. For PLD deposition the target was made by uniaxial pressing of ESB powder and sintering at 890 C for 4 h. ESB powder was prepared by solid state route as previously described in this manuscript. A KrF eximer 2, and a frequency was 5 Hz. The distance between target and substrate was 5 cm. O2 filling was 0.3 Torr vacuum level. The substrate (GDC surface on a NiO GDC anode) was maintained at room temperatur e. The deposition was made for 45 min. To achieve a pure ESB phase, an annealing step of 700 and 890 C for 4h was used and the film was examined by XRD. The Crystallinity of heat treated ESB layers w as compared with the ESB layer without heat treatment. The thickness of the ESB layer by cold PLD was hard to measure because of severe roughness of the film (Figure 6 8).

PAGE 49

49 3.2.4. Hot PLD To further improve ESB layer quality PLD deposition was performed with substrate (GDC electrolyte and AFL on the Ni GDC ano de) heating to 630 C. The target (ESB pellet) and substrate were the same as those used for Cold PLD. The parameters used for PLD were the same as with Cold PLD, but there was no heat treatment step after the deposition. XRD was employed to check crystal linity of the as -deposited ESB layer. I -V characteristics were carried out by a Solartron 1287 and the impedance analysis was performed by a Par -stat using 90 sccm of air and wet hydrogen. Again, ASRIV and ASRImpedance were used to distinguish the total ce ll ASR obtained by the two different methods. 3.3. Characterization 3.3.1. Microstructural Analysis Microstructures of each comp onent of the SOFC before and after test ing can be investigated by SEM. Porosity of the bulk of the anode and cathode can be determined. Also, SEM can easily measure other parameters that may be of interest such as anode thickness and pore sizes Another important reason for SEM being one of the major characterization tool s is that it can measure the penetration of unwanted impurity into the matrix by back -scattered imag ing When images in high resolution (up to 100,000x) are required, f ield emission scanning electron microscope (FE SEM) can be used to characterize the detailed microstructure of the fuel cell sample. In order to vi ew the cross -section of all three layers (anode, electrolyte, and cathode) of the cell, the test sample was fractured and some times, embedded in epoxyresin, and mechanically polished.

PAGE 50

50 3.3.2. XRD XRD is a valuable tool for confirming acquisition of a des ignated phase. In addition, phase purity (both before and after electrochemical testing) can be judged based on the peak spectrum it produces. For use with a SOFC study, XRD can be used as a post analysis tool. The mean sizes of the crystallites are deter mined from the x ray diffractograms, using the Scherer equation (3 1) and assuming that the particles are spherical ; max 2cos / 9 0 B L (3 1) ray wavelength ( 0. 154056 nm radiation), B is the width of the diffraction peak at half -max is the angle at the peak maximum position. 3.3.3. I -V Measurement Key characteristics of SOFC can be measured using the plot of v oltage output as a function of electrical current density recorded by a potentiostat. The theoretical background of the I -V characteristics are provided in chapter 2 of this dissertation. Measurable properties using pote ntiostat include I -V characteristics, power density as a function of current density, open circuit potential and degradation of voltage output depending on given current. A sample with silver lead wire s and pla t inum current collector s was placed on a zirco nia tube in a custom -made testing apparatus The sample leads were connected to gold leads from the air -side and fuel -side in the reactor. The fuel cell sample was sealed (anode side) using a two part ceramabond sealant (a mixture of 517powder and 517-liq uid from Aremco). Ceramabond was applied around the edge of the sample as a sealant, and w as cured inside the testing furnace. The s eal ing was completed while the furnace was ramped to the operation temperature Once operating temperature was reached hydr ogen was flown through a bubbler as the fuel to the

PAGE 51

51 anode and dry air gas was fed to the cathode. Flow rates were maintained using mass flow controllers. Depending on the chosen power generation level, 30 sccm or 90 sccm of dry air and wet hydrogen were s upplied to the cathode and anode, respectively. Note that the flow rates should be maintained symmetric during all measurements, because the OCP can be lowered by asymmetric flow rate s due to external gas leakage. The cell OCP was monitored using a Solartr on 1287 potentiostat until a stable value was reached, and the I -V measurements were taken with the same instrument. As it was mentioned in chapter 2, the slope of the I -V curves give s the ASR. T he ASR near the open circuit region can be compared with the ASR measured by the impedance spectroscope under no bias. Taking the initial slope of the I -V curves, total cell ASR was estimated Throughout this dissertation, the regions from 0 to 0.02 A/cm2 in most I -V curves were used to measure the ASRIV and the va lues were compared with ASRImpe. Less discrepancy between ASRIV and ASRImpe means higher credibility of the data, since they were measured by different devices. 3.3.4. Impedance Spectroscopy Impedance spectroscopy is a valuable tool for characterization of electrochemical processes. A small AC potential (across a range of frequencies) is applied to the sample, and the current response (impedance using Ohms Law) through the sample is measured over a range of frequencies. This response is usually represent ed as a Nyquist or Cole Cole plot where the real part of the impedance is the abscissa and the imaginary part of the impedance the ordinate The response of the cell is usually modeled in terms of equivalent circuits, i.e., a group of electrical circui t elements (resistors, capacitors, inductors) that are connected in a way that would give the same response as the cell.

PAGE 52

52 After the I -V measurement, impedance analysis was carried out at open circuit conditions using twopoint probe (cathode and anode) meas urements with a Par -stat 2273 (Princeton Applied Research) with a frequency range of 10 KHz to 0.01 Hz and an ac signal strength of 10 mV at various temperatures from 500 to 650 C. Impedance spectra were used to calculate the total ASR. From the high fr equency complex -plane intercepts of the impedance spectra with the real axis, the ohmic ASR values were calculate d normalizing the resistance according to cathode area. Electrode ASR values were calculated from the difference between low and high frequenc y intercepts, also normalizing the difference according to cathode area. From the IV characteristic measurement s electrolyte potential loss es can be easily separated because electrolyte loss es are purely ohmic. However, the cathodic polarization and anodi c polarization are not easy to separate. So, three point measurement s are necessary However, even in an impedance study, separating cathode and anode contributions is very difficult because of the reference electrode location on the thin electrolyte.68,69 The best way to measure the anode polarization is to use a thick electrolyte with an imbedded reference electrode. However, if we fabricate such a cell, it is different from the actual cell we are targeting wh ich has a thin electrolyte It is safe to assume that reported anodic polarizations using three -point measurements with a reference electrode placed on the thin electrolyte only measures the polarization for that specific cell geometry. This configuration should be used to measure the difference between polarization measurements depending on the research parameter s not to measure the actual anodic polarizations. Throughout this dissertation, all impedance measurements wer e done using the two -point pro b e method under open circuit condition s which only separates ohmic and electrode polarization s Knowing values of cathodic polarization losse s of LSCF GDC cathodes

PAGE 53

53 on GDC electrolyte s from previous measure ments using a symmetry cell setup, the anodic polarizat ion can be acquired by subtracting out the cathodic polarization from the overall electrode polarization. The calculated anodic polarizations may not be actual values, but will provide a standard to compare each sample to, assuming that the same cathode on both t hin film and thick pellet GDC electrolytes will have similar cathodic polarizations. Figure 3 2 shows the measurement setup where most samples discussed in this dissertation were tested. Figure 3 2 Electrochemical measurement station

PAGE 54

54 CHAPTER 4 ANODE FUNCTIONAL LAYER FOR IT -SOFC 4.1. Homogeneous Functional Layers 4.1.1. Microstructural Analysis 0.2 m 0.2 m 0.2 m 0.2 m 0.5 m 0.5 m 0.5 m 0.5 m 0.5 m Figure 4 1. TEM images of raw powders of NiO (top) and GDC (bottom) used for tape casting of anode support. Figure 4 1 shows TEM images to compare the relative particle sizes of NiO and GDC powders used for tape casting of the anode. NiO particles are mostly micron -sized, while GDC particles are less than 100 nm forming aggregates due to their high surface area. L arge micron -

PAGE 55

55 sized NiO powder was used for tape casting t o achieve sufficient anode porosity without addition of pore formers. The use of large NiO powder which will form nickel catalysts for hydrogen oxidation, can cause a large anodic polarization. Reducing the additional anode polarization by applying an AFL was the focus of this study Aggregates of GDC particles must be well dispersed especially for GDC electrolyte deposition as slurries with aggregated particle clusters form a porous film after sintering. As described in the experime ntal section, a polymeric dispersant, Solsperse, was used for this purpose. The microstructure of the GDC electrolyte prepared with Solsperse is shown in Figure 4 4 Since the GDC functional layer is the same material as the electrolyte, after sintering, it is indistinguishable from the electrolyte in cross -sectional SEM images. Conclusive evidence of the AFL and its effect on the performance can be only acquired by electrochemical measurements, which will be discussed in this manuscript. However, to ver ify formation of the AFL visually, anode surface views with and without the AFL were taken before GDC electrolyte deposition. Figure 4 2 shows the change in anode surface porosity and roughness by introduction of a GDC functional layer. Images of the ano de surface with AFL were taken after heat treatment at 900 C for 1 h As seen in Figure 4 2, the anode surface without a GDC functional layer is very porous and rough. On the anode surface it is easy to find large voids formed by use of large NiO partic les. Formation of a dense membrane by colloidal deposition is difficult to achieve on this type of porous surface especially with materials that are difficult-to sinter, such as ceria In Figure 4 2 a partially -sprayed GDC functional layer on the anode su ccessfully blocks most large pores on the surface. It should be noted that a thick layer of GDC covering the entire

PAGE 56

56 anode surface does not easily form unless a significant amount of GDC precursor is sprayed. Up to certain level of precursor spraying, the precursor solution is sucked into the pores and coagulates inside the pores filling the voids of anode surface. This result s in bare nickel particles on the anode surface. Bare anode surface Partial spraying of AFL Full spraying of AFL Bare anode surface Partial spraying of AFL Full spraying of AFL Figure 4 2. Backscattered images showing change in anode surface porosity and roughness by GDC functional layer deposition. Images were taken after pre -sintering the AFL at 900 C for 1 h The top left figure shows bare anode surface, the top right figure shows partial spray coating of the GDC functional layer, and bottom figure shows the fully sprayed GDC functional layer. Figure 4 2 shows that the fully sprayed GDC functional layer covers the entire anode surface with GDC, creating a relatively smooth and uniform surface for high quality GDC electrolyte deposition. NiO particle s on the anode surface are fully covered with GDC particles

PAGE 57

57 when a functional layer is used. This GDC functional layer is expected to provide better adhesion between the anode and the electrolyte. Figure 4 3 SEM micrographs showing of the s urface vie w of the GDC electrolyte deposited by spray coating on AFL coated anode. Figure 4 4 SEM micrograph showing the cross -sectional view of the SOFC with AFL after I -V test ing Cathode Electrolyte Anode

PAGE 58

58 Figure 4 3 shows the surface of the GDC electrolyte deposit ed by spray coating onto anode surface (fully covering GDC functional layer). The electrolyte deposition was followed by sintering at 1450 C for 4 h As seen in Figures 4 3 and 4 4, small pores are observed on the surface and in the bulk of the electrol yte. However, the cross -sectional view of GDC electrolyte (Figure 4 4 ) shows no open porosity. The thickness of the GDC electrolyte is estimated to be 10 m from the SEM image. As seen in Figure 4 4 after reduction of NiO to Ni during operation, the an ode became very porous, sufficient for gas transport. 4.1.2. Electrochemical Performance 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0 0.2 0.4 0.6 0.8 1 0 0.5 1 1.5 2Potential / VPower density / W cm-2Current density / A cm-2No-AFL Partial-AFL Full-AFL 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0 0.2 0.4 0.6 0.8 1 0 0.5 1 1.5 2Potential / VPower density / W cm-2Current density / A cm-2No-AFL Partial-AFL Full-AFL Figure 4 5 I -V characteristics of No -AFL Partial -AFL and Full -AFL samples using 30 sccm of wet hydrogen and air a t 650 C.

PAGE 59

59 Figure 4 5 shows the I -V characteristi cs of the No -AFL Partial -AFL and Full-AFL samples All I -V measurements utilized 30 sccm of wet hydrogen on the anode side and 30 sccm of dry air on the cathode side. At 650 C, the OCP and the maximum power density of the No AFL sample were 0.677 V and 407 mW / cm2, respectively. Despite the very low total ASR calculated from linear fit of this I -V curve (0.23 7 cm2), the maximum power density was not high due to the low OCP value. For comparison, the total ASR from the impedance spectrum of the No -AFL sample in Figure 4 6 was 0.218 cm2. ASR values from both I -V and impedance measurements were comparable within 10 % for all samples The low OCP value of the sample without an AFL suggests that the 10 m thick GDC fabricated by spray coating is not de nse enough to block H2 gas or electronic leakage currents. It is clear that the spray coating of the GDC electrolyte without an AFL needs improvement. However, without any changes to the spray coated GDC, the OCP value was increased from 0.677 to 0.719 V by a partial AFL coverage at the interface and dramatically increased from 0.677 to 0.796 V by a full AFL coverage. This gives strong evidence that the quality of the GDC electrolyte can be enhanced by using an AFL at the interface of the electrolyte an d the anode. The Partial -AFL sample resulted in a maximum power density of 738 mW / cm2 and the Full -AFL produced a maximum power density of 994 mW / cm2 at the same temperature. This is an 81 % increase in the power density for Partial -AFL sample and a 144 % increase for Full -AFL sample Th e dramatic increase in the maximum power density by application of the AFL is not only due to the increased OCP but also due to reduced total ASR. Table 4 1. Details of No -AFL, Partial AFL and Full -AFL samples at 650 C The u nit for ASR values cm2. Total ASR (IV) Total ASR (Impe.) Ohmic ASR Electrode ASR No AFL 0.237 0.218 0.104 0.114 Partial AFL 0.130 0.128 0.062 0.066 Full AFL 0.089 0.087 0.051 0.036

PAGE 60

60 0 0.05 0.1 0 0.05 0.1 0.15 0.2 0.25ASRIm / cm2ASRre / cm2 Full AFL Partial AFL No -AFL Figure 4 6 The impedance spectra of No -AFL Parti al -AFL and Full -AFL samples at 650 C under open circuit conditions The i mpedance spectra were obtained using t wo point probe measurements. The change in the ASR for the No -AFL, Partial -AFL and Full -AFL samples was further analyzed by impedance measuremen ts (Figure 4 6 ). Table 4 1 shows the total, ohmic and electrode ASR values from the three samples. The No -AFL sample has a total ASR value of 0.218 cm2 at 650 C. 48 % of the total ASR value is from the Ohmic ASR and 52 % is from the electrode ASR. The Partial -AFL sample had only 59% of the total ASR of No-AFL sample This implies that a partially sprayed AFL successfully reduced the total cell AS R by 41 %. The reduction in the total ASR is due to a 41 % reduction in the Ohmic ASR and a 42 % reduction in the electrode ASR. At this point, the electrode ASR was not further separated into the cathodic and the anodic resistances due to an issue with reference electrode configurations for thin electrolytes .68,69 However, the difference in the electrode ASR by use of this AFL can be attributed to the difference in the anodic polarization, since an AFL will n ot affect the cathodic polarization.

PAGE 61

61 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.5 1 1.5 2Potential (V)Power density (W/cm2)Current density (A/cm2)650 oC 600 oC 550 oC 500 oC 450 oC Figure 4 7 I -V characteristic s of Full -AFL sample at temperatures from 450 to 650 C us ing 30 sccm of wet H2 and air The total ASR of No -AFL sample was reduced by 60.1 % for the sample with a fully sprayed AFL. Th e total ASR of Full-AFL sample was only 0.089 cm2 at 650 C resulting in 994mW / cm2 at the same temperature. The ohmic ASR values decreased by 51.0 % with this AFL, and the electrode ASR was reduced by 68.4 %. As expected, a large anodic polarization is occurring due to the use of large NiO particle s for anode fabrication. In general the cathode polarization is considered to be the dominant contributor to the overall polarization in SOFCs. However, the fact that more than 60 % of the electrode resistance can be removed by the use of an AFL shows th at depending on anode design the dominant electrode resistance can be the anode polarization for anode -supported cells.

PAGE 62

62 Figure 4 7 shows the I -V behavior of Full -AFL sample at various temperatures from 450 C to 650 C. The OCP values are 0.796 V 0.830 V 0.874 V 0.913 V and 0.950 V at 650 C, 600 C, 550 C, 500 C, and 450 C, respectively. The maximum power densities are 994, 913, 627, 440, and 241 mW / cm2 at 650 C, 600 C, 550 C 500 C and 450 C, respectively. The sample has a 0.49 cm2 active ar ea In this measurement we limited gas flow rates to 30 sccm and n ote that the I -V curve at 650 C shows an increase in ASR with increasing current density T he anode shows concentration polarization at high currents even though large NiO particles were used to enhance anode porosity. The effect of gas flow rate on the anode concentration polarization will be further discussed in this manuscript 0 0.5 1 1.5 2 2.5 3 3.5 0 0.5 1 1.5 2 2.5 3 3.5 450 C 500 C 550 C 600 C 650 CZim / Zre / 0 0.1 0.2 0.3 0.4 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 550 C 600 C 650 CZim / Zre / Figure 4 8. The impedance spectra of Full -AFL sample at various temperatures under operating conditions.

PAGE 63

63 0 0.5 1 1.5 2 400 450 500 550 600 650 700 Total Electrode OhmicASR / cm2Temperature / oC Figure 4 9. The total, electrode and ohmic ASR values calculated from impedance spectra at temperatures from 450 to 650 C. Figure 4 8 shows the impedance spectra at each temperature for which I -V was measured. Figure 4 9 shows the total ohmic and elec trode ASR values at various temperatures calculated from the impedance data The ohmic and electrode ASR values were obtained from the low and high frequency intercepts of the spectra with the real axis, respectively. At 650 C the major contributions to the total cell resistance are ohmic and electrode polarization losses. Electrode resistance from the anode and cathode constitutes 41.38 % of the total resistance. However, as temperature decreases, electrode resistance increases, and at 550 C the elec trode ASR is greater than the ohmic ASR. At 500 C the electrode resistance is 68.79 % of the total ASR. However, the total ASR at 500 C is still lower than 1 cm2, resulting in a maximum power density of 440 mW / cm2. At 450 C electrode resistance dram atically increases, constituting 72.64 % of the

PAGE 64

64 total resistance of the cell. We believe this is due to the high activation energy of the oxygen reduction reaction causing higher resistance from the cathode at low temperatures. 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0 0.2 0.4 0.6 0.8 1 1.2 0 1 2 3 4 5Potential / VPower density / W cm2Current density / A cm2H2 100 / ccm H2 30 / ccm 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0 0.2 0.4 0.6 0.8 1 1.2 0 1 2 3 4 5Potential / VPower density / W cm2Current density / A cm2H2 100 / ccm H2 30 / ccm Figure 4 10. The e ffect of the gas flow -rate on Partial -AFL sample a t 650 C. The P artial -AFL produced 1.03 W/cm2 when it is not limited to 30 s ccm of gas flow rate. To investigate the effect of gas flow rate on the cell performance, hydrogen and air gas flow rates were varie d from 30 to 100 sccm for Partial -AFL sample and 30 to 90 sccm for Full AFL sample at 650 C as shown in Figure 4 10. The Partial -AFL s ample produced 738 mW / cm2 at 650 C using 30 sccm of wet hydrogen and dry air. The reason we limited gas flow rate to 3 0 sccm is to compare the data with that of No -AFL sample, which was acquired using 30 sccm Increasing wet hydrogen and dry air flow rate to 100 sccm resulted in 1.01 W / cm2 a t the same temperature (Figure 4 10). Currently, many I -V results and impedance measurements from the literature are not limited to 30 sccm gas flow rate s The fact the power density and impedance

PAGE 65

65 (ASR) can change by a large degree requires extra caution when comparing electrochemical properties between various cells. 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 0 0.5 1 1.5 2 2.5Potential / VPower density / W cm2Current density / A cm2 Figure 4 11. The e ffect of gas flow rate and gas composition on the performance of Full -AFL sample at 650 C. For Full -AFL sample, i ncreasing the H2 flow rate to 90 sccm maintaining 30 sccm air flow led to a reduction in ASR at high currents but a decrease in open cir cuit potential from 0.796 to 0.779 V (Figure 4 11). The fact that asymmetric flow rates cause a decrease in the OCP suggests that H2 gas either leaks through the electrolyte or through the ceramabond sealant. An increase in the air flow -rate from 30 to 90 sccm increased the OCP value back to 0.796 V but did not change the slope significantly. Changing air to pure O2 as the oxidant gas increased OCP to 0.836 V while ASR was not affected significantly. The maximum power density was increased up to 1.57 W / cm2 using 90 sccm of H2 and O2.

PAGE 66

66 4.2. Heterogeneous Functional L ayer s 4.2.1. Microstructural Analysis Ni GDC co spraying Figure 4 1 2 Microstructure of Ni GDC composite AFL on NiO -GDC anode T hree different AFLs were used in this study: a GDC functional layer (GDC AFL), a N i GDC functional layer (Ni GDC AFL) and an anode functional multilayer (AFML ), which is a bilayer of the GDC and Ni -GDC AFL s Figure 4 1 2 shows the surface co-sprayed with Ni and GDC precursor solutions T he surface of the bare anode after presintering ste p and the anode surface fully covered with GDC AFL can be found in Figure 4 2 T he cross section of a sample that has GDC AFL can be found in Figure 4 4 Compared to the bare anode surface, large voids on the GDC AFL sample surface are completely filled wi th GDC particles, providing smooth and homogeneous surface. Figure 4 1 2 shows that Ni GDC AFL can provide smooth surface, but the contrast in color indicates the composition is not homogeneous. The mass change after applying the GDC AFL was 3.4 mg/cm2 and 2.1 mg/cm2 for Ni GDC AFL. A simple calculation predicts that 3.4 mg/cm2 of GDC yields a film with ~ 4.7 m thick ness if 100 % dense. Though the entire surface is covered with GDC (Figure 4 2 ), the layer

PAGE 67

67 is substantially thinner In fact it is so thin that it is hard to estimate the thickness with cross sectional SEM. This is because the GDC precursors penetrate the presintered anode through capillary action Note that further spraying of the GDC precursor does not form a dense electrolyte quality GDC layer The data is not provided here but heavy spraying of GDC precursor s only produce s a porous poor quality film. Forming a dense layer using precursor solution may provide a continuous process from an AFL to electrolyte deposition, but it is beyond the scop e of this manuscript. 4.2.2. Electrochemical Performance 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0 0.2 0.4 0.6 0.8 1 0 0.5 1 1.5 2 No AFL Ni-GDC functional layer GDC functional layer Potential (V)Power density (W/cm2)Current density (A/cm2) 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0 0.2 0.4 0.6 0.8 1 0 0.5 1 1.5 2 No AFL Ni-GDC functional layer GDC functional layer Potential (V)Power density (W/cm2)Current density (A/cm2) Figure 4 13. T he I -V curves for the sample with no AFL, GDC AFL and Ni -GDC AFL using 30 sccm wet hydrogen and dry air at 650 C. Figure 4 13 shows I -V curves for sample s with no AFL, GDC AFL and Ni -GDC AFL using 30 sccm wet hydrogen and dry air at 650 C. T he OCP and the maximum power density of the N o -AFL sample were 0.677 V and 407 mW/cm2, respectively. The OCP and the maximum

PAGE 68

68 power density of the GDC AFL sample were 0.796 V and 994 mW/cm2, respe ctively The change in the ASR for the N i -AFL and GDC AFL samples was further analyzed by impedance measurements. 0 0.05 0.1 0.15 0.2 0.25 0 0.05 0.1 0.15 0.2 0.25ASRIm ( cm2)ASRre ( cm2) Ni-GDC functional layer GDC functional layer No AFL 0 0.01 0.02 0.05 0.06 0.07 0.08 0.09ASRIm ( cm2)ASRre ( cm2) Figure 4 14. T he impedance spectra under open circuit conditions at 650 C. Figure 4 14 shows the impedance spectra of the sample with no AFL, GDC AFL and Ni GDC AFL under open circuit conditions at 650 C, and Table 4 2 shows the total, ohmic and electrode ASR values calculated from the impedance data The No AFL sample has a total ASR value of 0.218 cm2. 48 % of the total ASR value of the No -AFL sample is from o hmic contributions and 52 % is from non -ohmic contributions (primarily electrode processes) The GDC AFL sample exhibited a total ASR of 0.087 cm2, which is 40 % of that of the No -AFL sampl e This implies that the GDC AFL successfully reduced the total cell ASR by 60 % The reduction in total ASR is due to a 51 % reduction in the o hmic ASR and a 68 % reduction in the electrode ASR. The details on GDC AFL performance can was discussed in th e previous section of this dissertation but it is notable that the GDC AFL reduces not only the electrode ASR but

PAGE 69

69 also the ohmic ASR. We believe lower ohmic ASR due to an AFL support the idea of its ability to increase the OCP by improving the electrolyte quality. Table 4 2. Summary of detailed ASR values 650 C including Ni GDC AFL and AFML sample. Total ASR IV Total ASR Impedance Ohmic ASR Electrode ASR No AFL 0.237 0.218 0.104 0.114 GDC AFL 0.089 0.087 0.051 0.036 Ni GDC AFL 0.077 0.078 0.057 0.021 A FML 0.077 0.075 0.055 0.020 In theory, performance should be improved if the AFL were a composite It is well know that the anode made with a mixture of Ni and GDC can provide larger TPB extending deeper into the bulk. However, F igure 4 1 4 shows the perf ormance of the Ni GDC composite AFL is lower than that of GDC AFL. This is not contradictory to conventional belief that the AFL should be a composite. As evidenced by the I -V curve, the lower performance was due to lower OCP value but note that in fact, t he slope (the ASR) is lower for Ni GDC AFLs. This is clear evidence that the mixture of Ni and GDC at the interface can extend TPB lengths and reduces ASR. At first this lower OCP was regarded as experimental error but repeated tests yielded reproducible r esults. Figure 4 15. Schematics for AFL structure for Multilayer approach. Porous and rough anode surface was first covered with Ni -GDC composite AFL, followed by GDC AFL.

PAGE 70

70 Impedance test ing was also conducted to further investigate the ASR performance. The figure shows that the sample using a Ni GDC AFL has a smaller total ASR mainly due to the small electrode ASR (high to low frequency intercepts). On the other hand, the difference in ohmic ASR values (high frequency intercept) for the two AFLs is rel atively small Detailed values are tabulated in Table 4 2 The c omposite AFL has a 42 % lower electrode ASR, while the o hmic difference is less than ~ 11 %. This means Ni GDC AFL clearly provides a greater TPB length at the interface and therefore a lowe r ASR. We believe that the GDC AFL provides a smooth and homogeneous surface for better quality GDC colloidal deposition. However, in terms of providing longer TPB s at the interface and reducing the electrode ASR, Ni GDC composite AFL is better. One can easily infer that a combination of higher OCP of the GDC AFL and lower ASR of Ni GDC AFL can generate even higher performance. In an effort to realize this assumption, a multilayered AFL was proposed with a Ni GDC AFL on the anode side and a GDC AFL on the electrolyte side In this configuration, GDC AFL provides a better surface for colloidal deposition of the electrolyte and the Ni -GDC AFL expands TPB length more at the anode surface. Figure 4 -1 5 illustrates the idealized microstructure for this AFL nea r the anode/electrolyte interface. Since these precursor solutions yield nano -sized particles and very thin films cross sectional SEM does not reveal any valuable information on the actual microstructure. However, the evidence of the bilayer existence is clear from the electrochemical performance. Figure 4 1 6 shows the performance of the AFML (anode functional multilayer) in comparison with a Ni GDC AFL sample. The I -V curves of AFML and Ni -GDC AFL samples are almost identical except for the OCP values. T his indicates that the AFML sample produced a similar ASR value as the Ni GDC AFL sample due to the TPB effect Impedance was also

PAGE 71

71 measured for the AFML sample but the semicircle is not present in this manuscript simply because it is superimposed on the se micircle of the Ni GDC AFL sample. However, detailed ASR values calculated from the impedance data is shown in Table 4 2 The OCP values for the GDC AFL, Ni GDC AFL and AFML sample s are 0.796 V, 0.731 V and 0.795 V respectively. The fact that the difference in OCP for the AFML sample and GDC AFL sample is n egligible indicates that the ability of AFML to increase the OCP (provide better surface for colloidal coating) is comparable with that of the GDC AF L Combining the high OCP and low ASR values, the AFML sample produced a maximum power density of 1.10 W/cm2 at 650 C. This is 10 % increase over the maximum power density of the GDC AFL sample 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0 0.2 0.4 0.6 0.8 1 1.2 0 0.5 1 1.5 2 2.5Potential (V)Power density (W/cm2)Current density (A/cm2) 650 oC Anode functional multilayer Ni GDC Single layer AFL Figure 4 16. I -V curves of the sample with Ni GDC single layer AFL and multilayered AFL using 30 sccm wet hydrogen and dry air at 650 C.

PAGE 72

72 Regarding with the fabrication of these precursor AFL, one important aspect should be noted. The pre -sintering of the AFL is crucial at this point, although it is an additional heat treatment step. We were able to observe, durin g deposition of the electrolyte on the AFL without pre -sintering, the GDC slurry doesnt spread well It seems that t he GDC precursor layer without heat treatment makes the anode surface hydroph obic. The resulting GDC green body film contained large GDC ag gregations scattered on surface, and remained porous after sintering. In order to eliminate the AFL pre -sintering step, further study on altering surface wettability is required. The AFML deposition requires sequential deposition of Ni GDC composite AFL and GDC AFL. This means 3 presintering steps are needed including anode presintering. In fact GDC AFL spray on Ni GDC AFL without heat -treatment does not easily form a pure GDC layer covering the entire Ni GDC surface. It seems that the hydrophobic surface caused by precursor spray ing prohibits uniform spreading of another precursor solution. Fortunately, as described in the experimental section, for the AFML deposition, the Ni -GDC composite AFL was heated on a hot plate while the GDC AFL was sprayed, and t hat seemed to provide better GDC precursor spreading. For practical applications and reduced fabrication costs, a dditional heat treatment steps should be avoided, and it seems likely that precursor deposition on a precursor layer does not require high temperature heat treatment.

PAGE 73

73 CHAPTER 5 CODOPED CERIA AS ALT ERNATIVE ELECTROL Y TES 5.1. XRD Figure 5 1 shows the XRD profile of the calcined powder of Sm0.075Nd0.075Ce0.85O2taken at room temperature. The p owder i s phase pure with cubic fluorite structure. The best estimate of the lattice constant (ao) was calculated using the least -squares extrapolation method .70 The lattice parameter of Sm0.075Nd0.075Ce0.85O2synthesized using coprecipitation technique was 5.4340 0.0017 The esti mated lattice parameter value is close to the value (5.4351 ) obtained from the lattice parameter (a) and dopant concentration (x) relationship for SmxNdxCe1xO2 presented elsewhere by Omar et al .36 30 40 50 60 70 80 331 420 400 222 311 220 200Intensity (Arbitary Units)2 (Degrees)111XRD pattern of Sm0.075Nd0.075Ce0.85O2synthesized using Co-precipitation method Figure 5 1 XRD pattern of Sm0.075Nd0.075Ce0.85O2 synthesized using co -precipitation technique. 5.2. Particle Size Figure 5 2 shows the particle size distribution of the Sm0.075Nd0.075Ce0.85O2and the commer cially obtained LSCF powders. The particle size distribution of the phase pure Sm0.075Nd0.075Ce0.85O2 powder synthesized using co-precipitation technique was measured

PAGE 74

74 using a Beckman Coulter LS13320 particle size analyzer The particle size distributio n of LSCF powder is provided by the company. In order to achieve high density ceramic electrolyte, it is and monodisperse. It can be seen that the Sm0.075Nd0.075Ce0.85O2particles exhibit size less than (d50) monodisperse Differential Number (%) Figure 5 2 Particle size distribution of Sm0.075Nd0.075Ce0.85O2synthesized using co precipi tation technique and La0.6Sr0.4Co0.2Fe0.8O3(obtained from Praxair Specialty Ceramics). The GDC powder obtained from Rhodia comprised of very fine particles of size less than 100 nm .54 On the other hand, the NiO particles are mostly micron -sized .54 Large particles of NiO, aft er reduction in H2, results in the porous anode. This is essential for the rapid fuel delivery to the reaction sites (i.e., triple phase boundaries of electrolyte, electrode, and air), and efficient water molecule removal. The LSCF powder used for the ca thode exhibit the number

PAGE 75

75 particle size distribution which is bi-modal, with most of the particles a (shown in Figure 5 2 ). 5.3. Microstructural Analysis Anode SNDC Cathode Figure 5 3 FE -SEM images of the cross -section view of the electrodes and electrolyte after operation Figure 5 4 FE -SEM image of the surface of the SNDC electrolyte after operati on

PAGE 76

76 F igure 5 3 shows the microstructure of the cross -section of the fuel cell sample, and Figure 5 4 shows the surface of Sm0.075Nd0.075Ce0.85O2 electrolyte. It can be seen both in the cross sectional and surface micrographs that the electrolyte is den sely sintered except for a few isolated residual pores. From the SEM image, while irregular, the thickness of the SNDC deposited with ceramic processes such as colloidal deposition. Further, the nickel (reduced form of NiO) particles in the anode are large compared to LSCF particles in the cathode. This is mainly due to the large particle size of the NiO starting powder. Although, the hydrogen oxidation in the anode side is kinetically more favorable than oxygen reduction in the cathode side, the significantly large particle size of the Ni near the electrolyte and anode interface can cause high anodic activation polarization at low temperatures. We have previ ously demonstrated that the anode polarization can be significantly lowered using an AFL between anode and electrolyte layers .13,15 In this work, the AFL which can reduce the additional anode polarization caused by the use of large NiO powder was employed. Details on the change in the microstructure for the same anode support using AFL can be found in the previous chapter and elsewhere .54,55 5.4. Power Density Figure 5 5 shows the I -V characteristics of the SOFC with Sm0.075Nd0.075Ce0.85O2electrolyte at temperatures ranging from 500 C to 650 C. The OCP values obtained w ere 0.86, 0.89, 0.93 and 0.96 V at 650 C, 600 C, 550 C and 500 C respect ively. The obtained OCP values were higher than previous typical OCP values achieved in the cell with GDC as an electrolyte, using similar fabrication route and the same experimental setup (0. 8 V at 650 C).54,55

PAGE 77

77 0 0.2 0.4 0.6 0.8 1 0 0.25 0.5 0.75 1 1.25 1.5 0 1 2 3 4 5Potential (V)Power density (W cm-2)Current density (A cm-2)650 oC 550 oC 600 oC 500 oC 0 0.2 0.4 0.6 0.8 1 0 0.25 0.5 0.75 1 1.25 1.5 0 1 2 3 4 5Potential (V)Power density (W cm-2)Current density (A cm-2)650 oC 550 oC 600 oC 500 oC Figure 5 5 The I -V characteristics of the prototype SOFC sample with Sm0.075Nd0.075Ce0.85O2 electrolyte at various temperatures ranging from 500 to 650 C in 90 sccm of both dry air and wet hydrogen. Figure 5 5 also sho ws the power density as a function of current density The maximum power densities achieved in the test cell were 1.43, 1.10, 0.73 and 0.32 W / cm2 at 650 C, 600 C, 550 C, and 500 C, respectively. For the intermediate temperatures, the obtained power d ensities were exceptionally high and higher than those obtained from the SOFC samples with 10 m thick GDC electrolyte (1 W/cm2 at 650 C) This c ould be due in part to the high ionic conductivity of Sm0.075Nd0.075Ce0.85O2 electrolyte. However it is important to note that the I -V characteristics and the maximum power density of the SOFC is a f unction of numerous processing and material variables. For this reason, the potential of Sm0.075Nd0.075Ce0.85O2

PAGE 78

78 electrolyte cannot be directly compared with that of GDC using I -V characteristics. Comparison between the ionic conductivity of Sm0.075Nd0.075Ce0.85O2and GDC electrolyte is reported elsewhere .29,36 However, the performance testing results clearly suggest that the Sm0.075Nd0.075Ce0.85O2 electrolyte material can successfully generate high power dens ity in SOFCs operating in the i ntermediate temperature range. 5.5. Impedance Analysis Figure 5 6. Impedance spectrum of the SOFC cell measured at various temperatures ranging from 500 to 650 C. Figure 5 6 shows the impedance spectrum of the SOFC at 650 C, 600 C, 550 C and 500 C. Using impedance spectroscopy, it is possible to separate out the electrode and ohmic contributions to the total ASR value at each temperature. From the high frequency intercept of the impedance spectrum with the real axis, the ohmic ASR value was calculated (after normalizing the resistance to the cathode area of 0. 48 cm2). Electrode ASR was determined from the difference between the low and high frequency intercepts ( also after normalizing the resistance to the cathode ar ea). Figure 5 7 shows the electrode and the ohmic ASR values separated from the total ASR at different temperature s. The rapid increase in the total ASR with decreas in g temperature is mainly due to the significant increase in the electrode polarization.

PAGE 79

79 T his increase in electrode polarization is most likely due to the cathode reaction which is a thermally activated process.71 In contrast ohmic ASR value s remain relatively small even at lower temperatures. 0 0.2 0.4 0.6 0.8 450 500 550 600 650 700ASR ( cm2)Temperature (oC) Total ASR Electrode ASR Ohmic ASR 0 0.2 0.4 0.6 0.8 450 500 550 600 650 700ASR ( cm2)Temperature (oC) Total ASR Electrode ASR Ohmic ASR Figure 5 7. ASR values at various temperatures T he total ASR value s were also calculated from the gradient of the linearly fitte d I -V curves. Table 5 1 shows the comparison between the total ASR value s obtained from the impedance measurement s and I -V characteristics. I mpedance measurement s w ere done under open circuit condition therefore while fitting t he I -V curve the region near zero curre nt (0 0.2 A/cm2) was taken into account. Both ASR values from the I -V characteristics and the impedance measurement are comparable at all temperatures The ohmic and electrode ASR values separated from the total ASR at each temperature a re also presented in Table 5 1 At 650 C the ohmic contribution to the total ASR is 3 1 %.

PAGE 80

80 With decreasing temperature the ohmic ASR percentage contribution remains relatively small. At 500C, the ohmic contribution towards the total ASR value is reduce d to 29 %. It is obvious that further improvement in the performance of this fuel cell can be done by reducing the electrode ASR especially for lower operating temperatures. Table 5 1. Comparison between the total ASR obtained from I -V characteristic and impedance measurement. The ohmic contribution towards the total ASR is also shown. Temperature ( C ) ASR I V ( cm 2 ) ASR Impedance ( cm 2 ) ASR Electrode ( cm 2 ) ASR Ohmic ( cm 2 ) 650 0. 106 0. 105 0.072 0.033 600 0. 166 0. 167 0.103 0.064 550 0. 323 0. 318 0.201 0.117 500 0 783 0 776 0.552 0.224 We believe the electrode ASR is mostly cathode ASR at this point by following reasons: First, the high activation energy of the oxygen reduction reaction results in higher resistance from the cathode at low temperatures. Secondly, with the recently -developed novel AFL, the anode polarization from hydrogen oxidation reaction caused by use of relatively large particles of Ni at the anode -electrolyte interface is effectively removed by 60 %. However, due to the limitation of the two -point impedance measurement, it is difficult to deconvolute the contribution from each of the electrodes It is also well known that introducing a highly ionic conducting phase in an electrode composite also enhances the performance of SOFC s .35 Since Sm0.075Nd0.075Ce0.85O2exhibits higher ionic conductivity than GDC, it is expected that replacing GDC with Sm0.075Nd0.075Ce0.85O2 will further improve SOFC performance and is the subject of fut ure work.

PAGE 81

81 CHAPTER 6 ESB/GDC BILAYER ELECTROLYTE FOR IT -SOFC 6.1. Co -pressing Pro cedure It has been demonstrated that a YSZ electrolyte and LaSrMgO3 cathode can generate the maximum power density of 2 W/cm2 at 800 C.5 This chapter is dedicated to reproduction of the same level power densities at IT ranges. Figure 6 1 Bac k scattered image showing the cross -section of the GDC ( ) and ESB (20 ) bilayer electrolyte with BRO7 ESB composite cathode and Ni -GDC anode after cell testing. BROcc refers to a pure Bi2Ru2O7 current collector. Figure 6 1 shows the m icrostructure of the co -pressed sample. Great effort was made to reduc e the thickness of the electrolyte while spreading GDC powder on NiO GDC anode support for higher performance. However, the minimum thickness obtained by co-pressing route was ~50 Also, as seen in the Figure 6 1 the thickness of GDC throughout the cel l is not uniform. This is one limitation of the co -pressing method used for this study. Electrolyte uniformity can be improved by spray coating GDC onto the anode support prior to pressing distributing a thin

PAGE 82

82 layer of GDC particles across the surface .6 Alternatively, foamy GDC powder can be used to better control GDC thickness while spreading across the surface .72 Howeve r, in this work the co -press ing approach was not further studied as it is not suitable for mass production. Rather, it was used for proof -of -concept purposes. More practical SOFC fabrication routes involving tape cast anodes and thin electrolyte depositio n is discussed later in this article. 0 0.2 0.4 0.6 0.8 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0 0.5 1 1.5Potential (V)Power density (W/cm2)Current density (A/cm2)LSCF-GDC on GDC BRO7-ESB on GDC BRO7-ESB on ESB/GDC 650 C Figure 6 2 I -V characteristics of three co -pressed samples at 650 C using 30 sccm of wet hydrogen and dry air. The result shows that BRO7 ESB cathode has lower ASR than LSCF GDC cathode on a single layer GDC electr olyte. The performance of the BRO7 -ESB cathode was further improved by using a bilayer electrolyte resulting in 588 mW/cm2. Figure 6 1 reveals that ESB layer is porous while co -pressed GDC is very dense. The porous nature of the ESB layer is a result of th e screen printing of coarse powder synthesized by the solid state process. The ESB layer thickness was ~20 resulting in a total bilayered

PAGE 83

83 electrolyte thickness greater than 70 Hence a substantial portion of the total cell resistance is expected to be attributable to the electrolyte ohmic resistance The results of current -voltage testing at 650 C are shown in Figure 6 2 as well as in T able 6 1 Figure 6 2 shows that BRO7 -ESB performs better than LSCF GDC composite cathode on the same GDC electrolyt e. The total cell ASR and maximum power density is 0.43 cm2 and 460 mW /cm2 for BRO7 cm2 and 338 mW /cm2 for LSCF GDC, respectively. Since the anode supports and the electrolytes were all prepared at the same time, resistances due to the anode and electrolyte are expected to be similar for both cel ls, so the improvement can be attributed mainly to the cathode. Table 6 1. Details of co pressed cell. Cell Type OCP (V) ASR ( cm 2 ) Maximum Power Density (W/cm 2 ) LSCF GDC on GDC 0.91 0.62 338 BRO 7 ESB on GDC 0.90 0.43 460 BRO 7 ESB on ESB/GDC 0.87 0.32 588 The cell utilizing a bilayer electrolyte had significantly better performance compared with the other cells 588 mW/cm2 2 ASR despite the added resistance caused by inserting an ESB layer between the GDC and the c athode. This is partly due to a degree of catalytic activity present at the surface of the ESB electrolyte, and the role the cathode/electrolyte interface plays in the cathode resistance, as discussed previously .4448 This co -pressing study demonstrated that lower ASRs can be achieved using BRO7 ESB cathodes and using bilayer electrolytes. However, an increase in OCP for the bilayer samp l e was not observed. This should not be an undermining result to the bilayer con cept. A series of studies have already shown that an ESB layer can increase the low GDC OCP .8,4042 Note that these studies were all conducted using thick GDC pellets and thick layers of ESB, which can block leakage current relatively easily. The main reason for the low bilayer OCP can be seen in the

PAGE 84

84 microstructure. As seen in Figure 6 1 the screen printed ESB layer didnt produce a dense membrane. Porous ESB requires higher thickness to block electronic leakage cur rent from GDC and even with a thickness The next section discusses ESB synthesis by co precipitation in order to reduce particle size and thus reduce porosity. Further, there may be some stability issues at this temperature as post -mortem SEM analysis reveals formation of an interlayer between the cathode and the ESB layer (Figure 6 1 ). The nature and formation mechanism of this interlayer is not know at present, and may likely also be due to the porous na ture of the ESB layer as an interlayer was not observed for samples with more dense ESB layers, discussed below. It is clear that these cathodes have potential for use in lower temperature SOFCs. We emphasize that the bilayer electrolyte increased the max imum power density by 74% and lowered the ASR by 49%. 6.2. Colloidal R oute Co -pressing produced high OCP (~ 0.9 V) due mostly to high GDC density and thickness. However, co -pressing is not a practical fabrication route To overcome the limitations of copr essing, spray coating of GDC onto tape cast anode substrates was performed. This process reduced the entire sample thickne According to Kim et al. ,63 a thick anode support frequently exhibit s high anode concentration polarization Reduction in thickness significantly reduces ASR resulting in higher power density. Details on tapecasting anode and GDC electrolyte depos ition h ave already been described.67 Figur e 6 3 shows the GDC thickness is ~ 20 m Due to the mixed ionic -electronic conducting behavior of ceria electrolytes the OCP decreases as thickness decreases .62 Therefore,

PAGE 85

85 the OCP of this cell is only 0.68 V (Figure 6 5 ) compared to 0.91 V with the co-pressed sample under same conditions.55 GDC ESB BRO7 -ESB Ni -GDC Figure 6 3 Backscatter image of the cross -section of the bilayer sample produced by colloidal deposition Note that there is the unclear interface between the BRO7 -ESB cathode and the ESB layer. Figure 6 4 Backscatter images of bilayer sample s showing the surface view of the ESB layer produced by colloidal deposition. Note the high degree of roughness and porosity of the ESB layer

PAGE 86

86 Figure 6 4 shows the surface microstructure of ESB deposited on the GDC. Note that the magnifications are different. The ESB surface is rough and porous. At this point it is clear that further improvement in ESB deposition is required. However, we present this preliminary co precipitation ESB study because the result shows the nature of the bilayer electrolyte 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0 0.2 0.4 0.6 0.8 0 0.5 1 1.5 2 2.5 3 3.5Potential (V)Power density (W/cm2)Current density (A/cm2)GDC single electrolyte ESB/GDC Bilayer electolyte Bilayer with 90 sccm H2 Figure 6 5 The I -V characteristics of the samples with GDC single layer and ESB/GDC bilayer electrolyte at 650 C The bilayer electrolyte improved the maximum power density from 407 to 614 mW/cm2 using 30 sccm of air and wet hydrogen. The maximum power density of the bilayer sample was further improved to 808 mW/cm2 using 90 sccm of wet hydrogen. The Lower OCP of 90 sccm sample is due to asymmetric flow rate; the air flow -rate was maintained at 30 sccm. At 650 C the single layer SOFC resulted in 407 mW/cm2 even though spray -coated GDC produced low OCP (Figure 6 5 ). High flow rates can substantially improve cell performance (~200 sccm are frequently used.) so care must be used with comparing cell ASR and maximum

PAGE 87

87 power density values Considering the low gas flow -rate of 30 sccm this is result is quite good, and the low ASR (0.237 cm2) can be attributed to reduced electrolyte and anode thickness es. The ESB/GDC bilayer was prepared and tested at this stage, and was shown to successfully blocked some portion of electronic conduction resulting in a higher OCP of 0.75 V (Figure 6 5) Considering the fact that ESB is almost a purely ionic conductor73 the bilayer OCP is still low. However, the 0.07 V increase in OCP is significant and supports the bilayer concept With increased OCP the bilayer electrolyte sample resulted in a maximum power density of 614 mW/cm2. This 51 % increase in the maximum power density in not only due to higher OCP but also due to the lower ASR. Though the GDC electrolyte used for bilayer sample (~20 m) was thicker than that for the single layer GDC sample by 10 m, the I -V curve shows that the bilayer cell exhibits a lower ASR resulting in higher performance using a BRO7 -ESB cathode As seen in Figure 6 5 when H2 flow rate was increased to 90 sccm ( air was maintained at 30 sccm), the maximum power density and the ASR was 808 mW/cm2 and 0.133 cm2, respectively. Note that the asymmetric flow rate to the cathode and the anode resulted in lower a lower cell OCP, indicating an external gas leak .55 Table 6 2. Details of tapecast cell without AFL Cell Type OCP (V) ASR ( cm 2 ) Maximum Power Density (W/cm 2 ) GDC single layer 0.68 0.237 407 ESB (colloidal) /GDC 0.75 0.158 614 T able 6 2 summarizes the preliminary results obtained by colloidal deposition of ESB prepared by co -precipitation. The bilaye r showed a 0.07 V increase in the OCP and a 33 % decrease in the ASR resulting in a 51 % improvement in maximum power density. Though these results are promising, the microstructure of ESB shows that further improvement s in film

PAGE 88

88 deposition are required Th erefore an alternate deposition technique PLD, was employed to confirm that bilayer electrolyte s can produce high performance 6.3. Cold PLD Figure 6 6 As-deposited ESB layer on the GDC electrolyte. Figure 6 7 ESB layer after heat treatment at 700 C for 4h. Figure 6 6 shows the asdeposited ESB layer on sintered GDC layer. As seen in this figure ESB precursors are not densely nor uniformly deposited. Typically, PLD is used to deposit

PAGE 89

89 ultrathin, high quality oriented films on a semiconductor substr ate In this study PLD parameters are set for an unusually fast deposition rate so that the layer can grow micro -level thickness. This fast deposition rate resulted in a porous and lumpy layer. The poor quality of ESB was not improved after heat treatment resulting in porous and rough ESB (Figure 6 -7 ). The SEM image s shows that the electrolyte surface is fully covered by ESB and i n fact no pin holes showing the underlying GDC electrolyte were observed However, the layer is not sufficiently dense enough to block electronic current from GDC from cross -sectional view ( Figure 6 8 ). It is easier to grow a thick and dense ceramic membrane with unusually fast deposition rate by heating the substrate during PLD, and this will be addressed in the following section. However, the poor quality of ESB was still tested because while the ESB layer may not increase the OCP it may to decrease the ASR as discussed above GDC ESB Ni -GDC Anode Figure 6 8 Cross -section image of the cell with the ESB layer by cold PLD after heat treatment at 7 00 C, 4h. Figure 6 9 shows XRD data of ESB/GDC bilayer samples prepared by cold PLD Asdeposited ESB on GDC produced GDC peaks and no additional peaks. Instead, it shows a broad

PAGE 90

90 amorphous background probably from the ESB precursor Thus a calcinations st ep was performed on these samples in order to yield the fluorite phase for the ESB layer. ESB fluorite peaks appear after heat treatment at 890 C for 4h which is typical sintering temperature and time for solid state ESB fabrication. The cubic fluorite ESB formation was confirmed even at lower temperature, 700 C for 4h, by PLD deposition. 20 30 40 50 60 70 80 Intensity (arb. unit)2 (degree) 700oC, 4h 890oC, 4h As deposited GDC ESB Figure 6 9 XRD patterns of bilayer samples depending on heat treatment conditions. The bilayer sample with as -deposited ESB shows only GDC peaks. As mentioned in the previous section, spray coating of GDC onto tapecast anode produced low OCP (Figure 6 5 ). Improvement in GDC single layer fabrication was made b y a simple step involving the application of an AFL to the pre -sintered anode substrate surface Details on the a pplication and mechanism of improvement in performance by the addition of this AFL ha ve been discussed previously in the chapter 4. Figure 6 10 show s that the OCP has been increased slightly to 0.72 V and that the maximum power density has reached 1.03 W/ cm2 at 650 C using the AFL

PAGE 91

91 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0 0.25 0.5 0.75 1 1.25 1.5 0 1 2 3 4 5 6Potential (V)Power density (W/cm2)Current density (A/cm2)650 oC GDC single electrolyte ESB/GDC Bilayer electolyte Figure 6 10. I -V characteristics of the samples with GDC single layer and ESB/GDC bilayer electrolyte at 650 C using 90 sccm of air and wet hydrogen. Figure 6 10 shows the improvement in performance achieved for the bilayer cell over the single layer GDC cell As exp ected from the microstructure and minimal thickness of the ESB layer, the OCP was not improved by the bilayer. However, the maximum power density was improved to 1.45 W/cm2. W e can derive the total cell ASR, ASRI V, by taking a slope of the I -V curves near the open circuit region These values are tabulated in T able 6 3 The table shows that ASRIV was reduced from 0.125 to 0.084 cm2 when using a bilayer electrolyte. The bilayer cell exhibited a 44 % increase in t he maximum power density and a 33 % reduction in the ASR Table 6 3. Details of tapecast cell wit AFL. ASR unit is cm2. Cell Type Total ASR IV Total ASR Impedance Electrode ASR Ohmic ASR GDC Single layer 0.125 0.126 0.064 0.062 ESB(Cold PLD)/GDC 0.084 0.0 84 0.044 0.040 ESB(Hot PLD)/GDC 0.075 0.079 0.033 0.046

PAGE 92

92 The effect of the bilayer electrolyte on ASR was further investigated with impedance spectroscopy (Figure 6 1 1 ). The total ASR obtained from impedance spectroscopy was labeled as ASRI mpedance and de tailed values of various ASR are also listed in Table 6 3 First, we emphasize that the ASRIV produced by a Solartron 1287 and the ASRI mpedance by Par -stat are in good agreement within 1% showing high credibility of the data. The result shows that the bila yer reduced the total ASR of the cell from 0.126 to 0.084 cm2. The bilayer reduced both ohmic and electrode ASR 35 % and 31 %, respectively 0 0.05 0 0.05 0.1 0.15ASRIm ( cm2)ASRre ( cm2)650 oC ESB/GDC Bilayer electrolyte GDC single electrolyte Figure 6 11. Impedance spectra of the samples with GDC single layer and ESB/GDC bilayer electrolyte at 650 C u sing 90 sccm of air and wet hydrogen. It should be noted here that the ohmic ASR is lower in a bilayer sample than in the single layer sample. Equation (6 1 ) predicts that the lower ohmic ASR using higher ESB thickness for a given total bilayer thickness However, in this study, the total thickness is not fixed Rather, t he ESB layer is added to the 10 m GDC layer and should be regarded as an added ohmic resistance This was the case for every bilayer sample measured One possibility is that ESB penetrates into the GDC grain boundaries resulting in lower grain boundary resistances .43 However, further study is needed to resolve this matter

PAGE 93

93 6.4. Hot PLD Ni -GDC Anode GDC ESB Ni -GDC Anode GDC ESB Figure 6 12. The bilayer microstru cture with the ESB by Hot PLD Figure 6 13. XRD pattern of the bilayer sample. A p oor quality ESB film was formed from cold PLD as a result of the very fast PLD deposition rate, which is crucial for attaining micron thick films Figure 6 1 2 shows that t he quality of the ESB film was substantially improved by heating the substrate (GDC electrolyte on

PAGE 94

94 the anode) to 630 C during PLD It is due to an increase in the sticking coefficient of ablated molecules during PLD The ESB membrane thickness was ~ 4 m achieving a 0.4 thickness ratio of ESB to GDC (GDC electrolyte is ~ 10 m in Figure 6 1 2 ). An additional advantage of hot PLD is that an ESB calcination step is not necessary. Figure 6 1 3 shows that cubic fluorite ESB has formed on the GDC layer without additional heat treatment step. 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0 0.25 0.5 0.75 1 1.25 1.5 1.75 2 0 1 2 3 4 5 6 7 8Potential (V)Power density (W/cm2)Current density (A/cm2)GDC single electrolyte ESB/GDC Bilayer electolyte Figure 6 14. I -V characteristics of the samples with GDC single layer and ESB/GDC bilayer electrolyte at 650 C using 90 sccm of air and wet hydrogen. Figure 6 1 4 shows the I -V characteristics of both GDC single layer and ESB/GDC bilayer samples at 650 C. The bilayer electrolyte and BRO7 -ESB cathode increased the maximum power density from 1.03 to 1.95 W/cm2 (93 % increase). The slope of I -V curves shows that the bilayer electrolyte achieved an ASR of 0.075 cm2; a 40 % reduction compared with the single layer sample. The increase in OCP from 0.72 to 0.77 V also contribute d to the dramatic

PAGE 95

95 improvement in power density. The OCP of the hot PLD sample, 0.77 V, is higher than that of cold PLD sample, 0.71 V. This indicates that the increase in OCP is a function of layer densities and thickness and that hot PLD resulted in denser ESB layer. 0 0.05 0 0.05 0.1 0.15ASRIm ( cm2)ASRre ( cm2)650 oC GDC single electrolyte ESB/GDC Bilayer electolyte Figure 6 15. Impedance spectra of the samples with GDC single layer and ESB/GDC bilayer electrolyte at 650 C using 90 sccm of air and wet hydrogen. Figure 6 1 5 shows the effect of a bilayer electrolyte on ASR measured by impedance spectroscopy, and the values are given in T able 6 3 Again, the total ASRImpedance matches well with the ASRIV within 5%. Table 6 3 shows the reduc tion in total ASRImpedance is due to a 48% reduction in the electrode ASR and a 26 % reduction in the ohmic ASR. As was the case for the cold PLD sample, the hot PLD sample also produced lower ohmic ASR than the single layer. The thickness of the GDC layer was ~ 10 m for both cold and hot PLD sample. However, t he reduction in ohmic ASR is more significant in the sample prepared from cold PLD compared with hot PLD. This is most likely due to the fact that the ESB layer was much thicker in the hot PLD (Figure 6) resul ting in a proportionally larger electrolyte resistance. Due to the rough nature of the ESB layer produced by cold PLD, the thickness was difficult to characterize It is safe to assume that cold PLD only provide very thin and porous ESB interlayer. However thicker ESB

PAGE 96

96 should still be ideal for high power density since it significantly reduced the electrode ASR and increased the cell OCP The OCP and the ohmic ASR depend strongly on the ESB quality. Table 6 4 Summary of GDC single layer vs. ESB/GDC bilayer results. Cell Description Anode Support Fabrication GDC Fabrication ESB Fabrication GDC Single Layer ESB/GDC bilayer Co press Co press ed Co press ed Screen print 0.338 0.588 Colloidal Tape cast Spray coat Drop coat 0.407 0.614 Cold PLD Tape cast wit h AFL Spray coat Cold PLD 1.03 1.45 Hot PLD Tape cast with AFL Spray coat Hot PLD 1.03 1.95 *Maximum power density at 650 C ( W/cm2). At this point, even with the high performance result, it is hard to demonstrate the full potential of the electrochemica l performance of the bilayer electrolyte, mainly due to low OCP of the thin ESB layers. It is obvious that a thin and dense electrolyte is the key to higher OCP and lower ASR, and with it the bilayer electrolyte is expected to present even higher performan ce. T able 6 4 summarizes the fabrication technique used for the GDC and ESB layers and the maximum power densit y obtained for each sample

PAGE 97

97 CHAPTER 7 CO NC LUSIONS First a simple and effective fabrication method to produce an AFL on an anode surface prepared by tape casting was proposed and tested Without addition of pore formers to the anode we were able to produce reasonably high power density. The high anodic polarization generated by using large NiO as anode catalyst was successfully reduced using an A FL fabricated by the proposed method. The AFL can be deposited by various colloidal techniques. In this study spray -coating was employed to deposit the AFL since spray-coating was also used to deposit the electrolyte which means continuous procedure fro m the electrolyte coating to the AFL coating Two -point impedance measurements revealed that the AFL affects the electrode ASR and the ohmic ASR. The fully sprayed AFL reduced the total ASR from 0.218 to 0.087 cm2 and the electrode ASR from 0.114 to 0. 036 cm2. The maximum power density is increased from 407 to 994 mW / cm2 at 650 C by use of the Full -AFL. The dramatic increase in the maximum power density is not only due to the lower ASR but due to the higher OCP by the AFL. In this study, the low O CP across the GDC electrolyte was improved by 0.12 V. T he fact that the o hmic resistance was reduced by 51 % also supports this conclusion It should be noted here that increase s in cell OCP is not the typical purpose of the AFL. Rather, t his can occur when the GDC electrolyte coating is not optimized. In this case, a simple AFL step can help increase the low OCP by helping improve the quality of the GDC coating T wo different single layer AFLs were studied : a GDC AFL and a Ni GDC composite AFL. The res ult shows that GDC AFL s have larger electrode ASR value s but have higher OCP s W e assume that the smooth and homogeneous surface that the GDC AFL provides is ideal for colloidal deposition of GDC electrolyte. A Ni GDC AFL is ideal in terms of decreasing

PAGE 98

98 el ectrode ASR but not for increasing OCP. A m ultilayer approach was proposed to combine the advantages of these two single layer AFLs. The so -called AFML successfully achieved the relatively high OCP of the GDC AFL and the low electrode ASR of the Ni -GDC AFL resulting in the highest observed maximum power density of 1.10 W/cm2 at 650 C. The next work was to test the performance of SOFC s based on a novel co-doped ceria electrolyte material under real operating conditions The anode supported SOFC was fabri cated, with Sm0.075Nd0.075Ce0.85O2electrolyte deposited on NiO GDC composite anode using colloidal process ing Previously developed AFL was applied in fabrication of SNDC based SOFCs at the interface of SNDC electrolyte and NiO -GDC anode. The microstructure results show a dense ceramic electrolyte of thickness ~ 10 The anode microstructure consists of fine GDC grains with coarse Ni particles The current voltage characteristics were measured at various temperatures, using 90 sccm of dry air and wet h ydrogen in cathode and anode sides, respectively. At 650 C, a maximum power density of 1. 43 W / cm2 was achieved The total ASR of the cell was 0. 105 2 at 650 C resulting in the exceptionally high power density. The impedance analysis revealed that t he decrease in power density at lower temperatures is mainly due to the electrode overpotential T he ohmic ASR remain s relatively low at most temperatures contributing only ~ 30 % to the total ASR Th e last study examine d the incremental improvement in SOFC performance comprised of components prepared from a wide range of techniques from pressed anodes to tape -cast anodes, from GDC single layer electrolytes to ESB/GDC bilayer and from LSCF GDC composite cathodes to optimized BRO7-ESB composites. GDC sin gle layer electrolyte based SOFC s were prepared from f our different fabrications and exhibit maximum power densit ies

PAGE 99

99 ranging from 0.338 to 1.03 W/cm2, at 650 C. At each fabrication stage, an ESB layer was applied to form a bilayer electrolyte. ESB was de posited by a range of techniques including colloidal deposition and PLD. The result confirms that depending on a fabrication route the bilayer electrolyte can reduce the total ASR 33~49 % and increase the maximum power density 44~93 %. A bilayer ESB (hot P LD) /GDC electrolyte produced an exceptionally high power density of 1.95 W/cm2 at 650 C. T he results herein show that utilization of a bilayer ESB/GDC electrolyte can reduce total cell ASR by 49 % and increase the maximum power density by 93% giving cred ence to its use in practical SOFC applications, particularly at reduced operating temperatures Further improvement in colloidal deposition of thin, dense ESB layers is expected to increase OCP and yield even higher power densit ies

PAGE 100

100 APPENDIX A STACK CELL APPLICATIONS SOFC s have gained attention for many years as clean energy conversion device s Due to high operating temperature and high efficiency, the most suitable SOFC application is believed to be stationary applications. To be able to generate reasona bly high voltage and power for practical applications, stack s of a large number of highly efficient individual cells are required. Figure A 1 shows research activities toward fabrication of SOFC stacks applicable to the industry. Though theses are the mos t advanced SOFCs there are many obstacles still preventing successful commercialization of SOFCs such as long term stability, low fabrication cost, and high efficiency, etc. (a) (b) (c) (d) Figure A 1. SOFC Stack cell development in many countries. (a) Fuel Cell Energ y, Inc.74 in the U.S. (b) NTT Energy75 in Japan. (c) German Aerospace Center, DLR76 in Germany. (d) KIST77 in South Korea. Recently, the Florida Institute of Sustainable Energy (FISE) was built at the University of Florida to research SOFCs that can operate at lower t emperatures with various fuels, such as

PAGE 101

101 hydrogen, biofuels, ethanol, diesel and so on. Recently FISE was successful in development of a novel AFL application of a new co -doped ceria, high performance using ESB/GDC bilayer electrolyte, ESB compatible low A SR cathodes, use of JP8 and other hydrocarbon fuels and so on. After a progressive series of successes, a maximum power density of 2W/cm2 at the IT range was achieve d. Since these results were based on tapecast anode support s scaleup to larger size shoul d be relatively easy Sintered with electrolyte Presintered anode Figure A 2. Planar SOFC sample with 6 cm x 6 cm in size with possible active area of 30 cm2. The sample consists o f NiO GDC anode, GDC electrolyte by spin coating, and GDC AFL. Figure A 2 shows 6 x 6cm planar cell with possible acti ve area of 30 cm2 (5 .5 cm x 5 .5 cm) The se cells used the same NiO -GDC anode tapes as the previously tested button cells. Instead of punching out circular support s, a square 7.5 cm x 7.5 cm anode tape was cut out assuming 15% lateral shrinkage. After presi ntering, previously developed GDC AFL was applied using precursor solution and GDC electrolyte was applied by spin coating method (Chemat technology Inc.). Detailed process on tapecasting and AFL deposition can be found in the chapter 3 of this dissertatio n.

PAGE 102

102 Figure A 3 shows the microstructure of the planar SOFC sample. From the surface and cross -sectional view a few pores were observed, but they were mostly isolated. Figure A 3 (c) shows the back -scatter image of an anode cross -section by tapecasting. As e xpected large NiO particles were observed, and they were surrounded with we ll dispersed micro -sized GDC particles. Judging from the microstructural images, we can conclude that scale up to 30 cm2 was successfully achieved. However, the most convincing test for a SOFC is I -V measurement GDC NiO -GDC LSCF -GDC AFL (a) (b) (c) (d) Figure A 3 Microstructure of planar SOFC sample (a) surface view. (b) cross -section of GDC on NiO GDC anode. (c) backscattered image of anode cross -section. (d) cross -section of the planar cell with brush painted LSCF GDC composite cathode before test. Even though FISE has the capability of measuring large sized SOFC samples, the setup is not done yet. In the mean time, a circular button was taken out from the planar sample. The

PAGE 103

103 button cell was mechanically polished to 1 i nch in diameter, and LSCF GDC composite cathode was applied as described in chapter 3. Figure A 3 shows the cross -sectional view of the sample. Note that brush painting is not a reproducible technique and is therefore not applicable for larger cells. It wa s only applied to the button cell for this experiment. A better deposition technique using screen printing is currently being developed for future cells 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0 0.2 0.4 0.6 0.8 1 1.2 1.4 0 1 2 3 4 5 6Potential (V)Power density (W/cm2)Current density (A/cm2)650 OC 30 ccm 90 ccm 0 0.04 0 0.02 0.04 0.06 0.08 0.1 0.12ASRIm ( cm2)ASRre ( cm2)650 oC90 ccm 30 ccm (a) (b) Figure A 4. I -V characteristics (a) and impedance spectra (b) of a button cell taken from a large st ack cell. 30 and 90 sccm symmetric flow rates were used for hydrogen and air.

PAGE 104

104 Figure A 4 shows the I -V characteristics and impedance data of the SOFC sample at 650 C. Since the planar cells will eventually be used in a stack which has different testing e nvironment in terms of fuel and oxidant flow, both 30 and 90 sccm flow -rates were tested. The OCP showed a typical partial GDC AFL value (chapter 4). The maximum power densities were 1.07 and 1.33 W/cm2 for 30 and 90 sccm, respectively. Simple comparison w ith the result for the partial AFL reported in chapter 4, shows that the performance of the SOFC was enhanced. Figure A 4 also shows impedance spectra for 30 and 90 sccm. It indicated that the reduction in the ASR due to higher gas flow rate is mainly due to reduction in gas conversion limitations. However, further study is required to analyze impedance data to determine the processes ascribed to each semicircle. The button -cell test confirmed that the fabrication of the planar cell for a stack application resulted in a high quality SOFC samples, satisfying high efficiency requirements for actual applications. In order t o test the planar cell s under real stack operating conditions a better technique for uniform cathode deposition over a large area should be developed The initial choice was the spray coating, which was well studied for the electrolyte deposition previously. Figure A 5 shows a stack cell that has spray -coat LSCF -GDC cathode. Recently we found that the spray-coating of LSCF GDC cathode is a good cathode deposition method without any delamination or crack ing found. However, to prevent accumulation of slurry at the center of the cell, a large spray diameter is required. To deposit a thick and laterally continuous cathode with that large target area, too much slurry was wasted. More recent results on screen print ed cathodes demonstrated a thick cathode can be achieved with good uniformity. Currently the application of screen print ed cathode s using an automated screen printer at FISE is being inv estigated.

PAGE 105

105 Figure A 5. 6 cm x 6 cm SOFC planar cell with NiO GDC anode, GDC electrolyte, GDC AFL, and 30 cm2 LSCF GDC cathode. At the University of Florida alternative electrolytes have been developed with higher ionic conductivities such as Dy and W doped bismuth oxide28 and Sm and Nd doped ceria The a pplication of these new materials into an actual fuel cells and stack s with bilayer electrolyte configuration shown in chapter 6 with novel molecular AFLs and optimized Bi2Ru2O7 composite cathode, wi ll like ly demonstrate an exceptionally high performance. I believe that this dissertation made a significant contribution to realization of highly efficient SOFC fabrications.

PAGE 106

106 APPENDIX B REPRODUCIBILITY AND STABILITY The SOFC performance results in this dissertation were exceptional However, the length of time in processing and testing trials required to achieve these good results was not discussed. Also, it would be difficult for an inexperienced SOFC researcher to reproduce the same level of power dens it y Similarly this dissertation did not cover stability issues, which is a crucial factor for real -world applications. This chapter is designed to describe the multitued of failures that have occurred over the past 2 years, and how there are still proble ms that need to be solved 1. Slow P rogress toward H igh P erformance It should be noted here that the previous chapters of this dissertation deal with results that were relatively reproducible and reliable. This section deals with the results that were gat hered when reproducible and stable operation was not achieved. So, most results present ed in this section should not be referenced for any scientific matters The anode of the first cell was prepared by uniaxial pressing of a mixture of NiO (J.T.Baker) and GDC (Rhodia) powder. GDC electrolyte slurry was prepared without plastisizer, and the dispersant and binder quantity was not optimized. The composition is given in the following subsection. The GDC electrolyte was dip coated onto the anode and LSCF GDC ca thode was brush painted on the electrolyte. The first test resulted in 7 m W/cm2 at 650 C with an OCP of 0.55 V The main problem of having extremely low power density is that it is difficult to determine the main polarization mechanisms. Also, unstable OC P allows only a limited number of tests at a limited temperature range. At that time, impedance measurements were not employed for the performance testing, which would separate ohmic and non-ohmic resistances giving better information regarding the dominan t polarizaiton mechanism

PAGE 107

107 0 0.1 0.2 0.3 0.4 0.5 0 0.005 0.01 0.015 0.02 0.025 0.03 0 0.05 0.1 0.15 0.2 0.25Potential (V)Power density (W/cm2)Current density (A/cm2) 30 sccm 60 sccm 90 sccm 0 0.1 0.2 0.3 0.4 0.5 0 0.005 0.01 0.015 0.02 0.025 0.03 0 0.05 0.1 0.15 0.2 0.25Potential (V)Power density (W/cm2)Current density (A/cm2) 30 sccm 60 sccm 90 sccm Figure B 1. I -V characteristics of the sample after 6 months of studying. Increasing H2 from 3 0 to 9 0 sccm dramatically decreased the OCP and the power density indicating external gas leak Figure B 1 shows that even after stu dying similar cells more than 6 months the results were very poor at 650 C In this figure three main problems were observed : low OCP, high resistance and external gas leak. The OCP of 0.5 V was a normal value at that time mainly due to the porous nature of the electrolytes developed The porosity of the electrolyte was so great that large NiO grains were commonly observed in the electrolyte, which will reduce to Ni metal during operation conducting electrons through the electrolyte ( Figure B 4 (b) ). Also while maintaining the air flow rate a t 30 sccm, and increasing H2 flow rate from 30 to 90 sccm yielded a significant drop in OCP indicating a large external gas leak. Finally the ASR estimated from the I -V curve is ~4 cm2, which is extremely high value at 650 C. 1.1 Electrolyte P article S ize Effect The first step toward the pursuit of higher performance involved selecting the right powder from various particle sizes. Three different GDC powders were obtained from diff erent

PAGE 108

108 companies (Nextech, Rhodia and Anan) and the mean particle size was measured by a Beckman Coulter LS13320 particle size analyzer : 304.7 12.3 nm for GDC (Nextech) 278.2 29.1 nm for GDC (Rhodia) and 236.0 8.4 nm for GDC(Anan) Figure B 2 shows the I -V curves of the three samples with three different GDC powders. It is hard to conclude that simply changing particle size from 304.7 to 236.0 nm increases the power density 165 % especially when the reproducibility of the performance was questionable. Howev er, it is obvious that decreasing particle size increased the OCP and increased the maximum power density 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0 0.05 0.1 0.15 0.2 0.25 0.3Potential (V)Power density (W/cm2)Current density (A/cm2)Anan GDC Nextech GDC Rhodia GDC Figure B 2. I -V characteristics of GDC powders from three different companies at 650 C. The same trend for particle size effect was observed in t he SNDC electrolyte tests. Chapter 5 only presented SNDC powder made from co-precipitation C hanging SNDC synthesis route from solid state to co -precipitation yielded smaller particle size and therefore a major change in the microstructure was achieved a s seen in Figure B 3 (a) and (b). Note that Figure B 3 (b) shows the SNDC deposition on the GDC AFL, which as described previously, provides a better

PAGE 109

109 surface for colloidal deposition. However, the significant change in microstructure of the electrolyte is evident Anode SNDC Cathode Anode SNDC(a) (b) Figure B 3. (a) SNDC synthesis by solid state route. (b) SNDC synthesis by co -precipitation route. 1.2. Slurry C omposition Figure B 4 (a) and (b) shows surface and cross sectional views of the GDC electrolyte on the NiO -GDC anode. Note that th e magnifications are different in the figure. As seen in F igure B4 (b ), large NiO grains are found in the GDC electrolyte. This NiO penetration was a frequent problem before better electrolyte quality was achieved. The initial GDC electrolyte slurry composition was as following: GDC (10 g), Corn oil (0.3g) and PVB (1g) and ethanol 50cc. The initial composition was not taken from any literature and judged purely by eye To achieve the change in microstructure from (a ) and (b) to ( c ) and ( d ), it took more th an 2 years of trial anderror experiment ation Among many trials, no systematic research was taken to optimize the slurry composition mainly due to instability of the performance and therefore any reasonable conclusion on each component s concentration eff ect wa s hard to establish However, the final composition is given in Chapter 3 of this dissertation with addition of DBP as plasticizer A t horough study is not present ed here, but in general PVB prevent s delamination of the film,

PAGE 110

110 dispersant reduces agglo merates in the slurry and DBP prevents crack formation during the drying process of the green body (a) (b) (c) (d) GDC NiO -GDC GDC GDC NiO -GDC GDC Figure B 4. (a) Surface of the GDC electrolyte (b) Back -scattered cross -sectional image of the GDC electrolyte on NiO GDC anode (c) Surface of the GDC electrolyte with higher density. (d) Cross -sectional image of the GDC electrolyte with higher density It should be noted that the use of extensive amount s of these three organic components yields an excellent green body and sometimes even after sintering the film looks dense However, I -V test ing still shows either low OCP or high ohmic resistance in this case The typical microstructure of this type of sample is given in Figure B 5 The sample in the Figure B 5 looked shiny and transparent In general s hininess means fewer defects in the film light scattering can be used as a qualitative estimation method of electrolyte density. Samples with

PAGE 111

111 e xtensive amount s of organic compounds emit a greater volume of fumes around 400 C during sintering, and it is likely that full densification is difficult to achieve since the green body loses such a large fraction of its volume The total amount of organic compound has been limited less than ~10 wt% through out the studies in this dissertation A better electrolyte was achieved by particle size optimization and the slurry composition study, but the poor performance with very low OCP was not solved to this point. Figure B 5. A sample looking acceptable from visually inspection can have poor performance and porous mi crostructure. 1.3. Co -press ing The first significant improvement in terms of performance was made by changing the fabrication route to co -pressing. This involves pressing the anode with a uniaxial press followed by spreading the electrolyte powder on top b y opening the cap of the die, and co-pressing the anode electrolyte bilayer again. Details on co -pressing method are present in Chapter 3. The major disadvantage of this method is that the anode and the electrolyte thicknesses can not be reduced, which is not suitable for achieving high performance. However, Figure B 6 shows the I V curves of the co -pressed sample. The maximum power density values achieved was much

PAGE 112

112 lower than what would be considered acceptable for publication but the major contribution tha t the co pressing route provided was a high and stable OCP. The co -pressed cells still exhibited external gas leak ranging from 2.5 to 5 sccm using 30 sccm of inlet gas. However, high and stable OCP enabled study of the effect of flow rate on performance. 0 0.2 0.4 0.6 0.8 1 1.2 0 0.025 0.05 0.075 0.1 0.125 0.15 0.175 0.2 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7Potential (V)Power density (W/cm2)Current density (A/cm2)500 C 550 C 600 C 650 C Figure B 6, I -V curves of the copressed sample at various temperatures using 90 sccm dry air and wet hydrogen. Figure B 7 shows the H2 and air flow rate on the I -V curves at 650 C. The result shows that changing in the gas flow rate on the cathode s ide does not give a significant change in performance, while changing the flow rate on the anode side yielded a substantial change in performance From this result it can be concluded that the gas flow through a thick and dense anode pellet is a significan t problem. While co -pressing resulted in an in -house record high performance of 187 mW/cm2 at 650 C and further improvement on performance is possible, further study was not performed mainly because of the inherent limitation of the pressing method. At that time we decided to move on to tapecasting the anode support, where thin anode s

PAGE 113

113 and electrolyte s are more feasible. The details on the anode tapecasting were covered in Chapter 3 of this dissertation. 0.25 0.30 0.35 0.40 0.45 0.50 0.160 0.165 0.170 0.175 0.180 0.185I (Amps/cm2)Power (Watts/cm2)iv 650C 90ccmh2.cor iv 650C 60ccmh2.cor iv 650C 50ccmh2.cor iv 650C 45ccmh2.cor iv 650C 40ccmh2.cor iv1 650C 30ccmh2.cor iv650C 20ccmh2.cor iv650C 15ccmh2.cor H2 Flow rate increasing From 15 to 90 ccm 0.25 0.30 0.35 0.40 0.45 0.50 0.160 0.165 0.170 0.175 0.180 0.185I (Amps/cm2)Power (Watts/cm2)iv 650C 90ccmh2.cor iv 650C 60ccmh2.cor iv 650C 50ccmh2.cor iv 650C 45ccmh2.cor iv 650C 40ccmh2.cor iv1 650C 30ccmh2.cor iv650C 20ccmh2.cor iv650C 15ccmh2.cor H2 Flow rate increasing From 15 to 90 ccm 0 0.25 0.50 0.75 0 0.25 0.50 0.75 1.00I (Amps/cm2)E (Volts) 0 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20Power (Watts/cm2) 650 C O2 Flow rate increasing from 15 to 90 ccm 0 0.25 0.50 0.75 0 0.25 0.50 0.75 1.00I (Amps/cm2)E (Volts) 0 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20Power (Watts/cm2) 650 C O2 Flow rate increasing from 15 to 90 ccm (a) (b) Figure B 7 (a) Change in the power density with in creasing H2 flow -rate from 15 to 90 sccm with 30 sccm of air. (b) Change in the power density with increasing air flow rate from 15 to 90 sccm with 30 sccm of H2. 1.4. Spray Coating Though a high performance SOFC using dip-coating method is possible, it t akes for a researcher long time to make a good coating by dip -coating based process Many times when equipment for fabrication is limited, experiments at a university lab scale rely heavily on the experience of the researcher At this point, a rt is a mor e dominant factor in the result than s cience. While experiments relying on devoted human skill can be reproducible, still difference between batches are not avoidable in general. Also, the transfer of knowledge of the same technique is not easy. Frequently, detailed descriptions on an e xperiment dont promise good reproducibility by other researchers Therefore, advanced process ing requires minimal change in performance between batches and should be independent of the researcher conducting the study In an effort to have a better and reproducible method, spray coating was introduced While dip coating and drop coating requires a human hand to deposit the slurry and to remove the excess

PAGE 114

114 slurry on the surface, spray coating utilize spray time and air pressure to deposit and substrate temperature to remove the excess slurry Figure B 8 shows a custom -made spray coating setup in a hood, which was used for deposition of GDC electrolytes used in Chapter s 4 and 6 of this dissertation. Figure B 8. A custom made sp ray coating setup. 1.5 Reproducibility It took more than 2 years and 126 pressed and tested samples before achiev ing 0.3 W/cm2 at 650 C. Including non tested samples, roughly more than 200 pressed samples were fabricated. The 127th tested sample is the f irst tapecast SOFC with dip -coated GDC electrolyte and it resulted in the maximum power density over 0.3 W/cm2. However, i t took less time to achieve 1 W/cm2 at the same temperature. So far, 55 tapecast samples were tested and the 26th tapecast sample achi eved 1 W/cm2. Figure B 9 shows th at high performance was reproduced for 3 consecutive samples though they were not fabricated for reproducibility test which requires all experimental paramters kept the same. The repeatable high performance means most exper imental parameters were in good control. A lso, the fact that less time was required to

PAGE 115

115 achieve the next break through supports this idea. For example, the maximum power density of 2 W/cm2 present ed in Chapter 6 was achieved by making 1 sample of ESB/GDC bi layer electrolyte. However, the stability of this cell was poor and did not last even 24 hours. Reproducibility comes from good control in experimental parameters and repeatable fabrication methods, but the stability of the SOFC is another subject that req uires extensive study. Figure B 9 I -V curves from three consecutive three cells showing reproducibility of high performance at 650 C. In each experimental parameters are different but high performance was reproduced. 2. Future W ork This subsection dea ls with the important issues that havent been solved. In order to solve the problems listed here, th o rough research over a long time span is required 2.1. Sealing and Leak By m easuring the gas flow rate of the inlet and the outlet of the SOFC testing app aratus, one can quantify the gas leak. GDC based cells (GDC electrolyte on NiO GDC anode) with P571

PAGE 116

116 (Aremco) ceramabond usually exhibit leakage ranging from 2.5 to 5 sccm using 30 sccm of inlet gas. Seal testing with various glass o rings was previously st udied by James Rhodes .78 However, no gas tight seal was achieved for GDC based cells from any glass o rings used in the above study, and the gas leak age was comparable to that observed for ceramabond. In this dissertation ceramabond was used because the maximum curing temperature is 372 C, where as glass o rings require furnace temperature s usually above 800~900 C for 1~3 hours. 5 sccm of gas leak corresponds to 17% hydrogen gas loss to the cathode side, which is significant and may in fact lower the 2OP of the cathode side. The CTE for ceramabond P571 is 12.6*106/ C and the CTE for GDC is 12.7*106/ C. However, since GDC is a thin membrane and only a small volume portion of the sample, the cell CTE will be dominated by the anode, which is cermet of NiO and GDC. M ismatch e s in CTE between the sealant and the sample can be a problem. However, i t should be noted here that the same ceramabond (and even various glass o rings) was used f or YSZ based cells (YSZ electrolyte on NiO/YSZ anode) and in the case of YSZ based cells, no external gas leakage was detected So, it is obvious that sealing GDC based samples is much more difficult Also, typically no gas leak is observed at the start of H2 flow to the anode, but the gas leak gradually increas es with time. This indicates that changes in the sample when exposed to reducing conditions may in fact cause the leak. According to Bishop et al., the chemical expansion of GDC is al most the same order of thermal expansion.79 I believe the chemical expansion from the bottom of the anode progressing toward the electrolyte is the main problem. This was confirmed by the fact that GDC based samples do not exhibit leakage when Ar ga s which has 2OP of 106 ( Airgas ) is used on the fuel side To address this problem Ag -paste (DAD87, Sh anghi RISR) used commonly for I/C circuit s

PAGE 117

117 and Ag conductor paste (9907, ESL) were applied and compressed by the same ceramic tube from the cathode side, but the preliminary test confirmed the same level of leak age At this point no solution has been sugge sted. To solve this problem a compliant sealant is required. Compressive sealing using metallic o rings can be a solution, but much study is required since each company has large number of pastes depending on purpose of use While gas -tight seal hasn t bee n available the application of the ceramabond for reduced leakage was studied. Ceramabond 1 powder P571 and a solvent, L571 were mixed in a 1.5 to 1 wt% ratio and appl ied to the sample. Then one curing step at 94 C for 1 hour was applied as per the compa ny instruction s Ceramabond 2 consists of the same P571 and L571 mixture applied to the sample. However, after standing at room temperature for 2 hours the liquid component was applied onto the ceramabond surface. Then a 3 -step cure was conducted at 94, 2 60 and 374 C for 2 hour eac h The external gas leak was reduced from 5 to 2.5 sccm with Ceramabond 2. Figur e B 10 shows Ceramabond 1 and Ceramabond 2 produced different results in I -V characteristics at all temperatures. However, it should be noted that c eramabond 2 steal has 8.3 % gas leak, and this is a problem that needs to be resolved. 0 0.1 0.2 0.3 0.4 0.5 0 0.2 0.4 0.6 0.8 1.0I (Amps/cm2)E (Volts)iv1 650C.cor iv1 600C.cor iv1 550C.cor iv1 500C.cor 0 0.025 0.050 0.075 0.100 0.125 0.150Power (Watts/cm2) 134 mW @ 650C 101 mW @ 600C 68 mW @ 550C 36 mW @ 500C 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1I (Amps/cm2)E (Volts) 0 0.05 0.10 0.15 0.20Power (Watts/cm2) 187 mW @ 650C 152 mW @ 600C 102 mW @ 550C 58 mW @ 500C(a) (b) Figure B 10. Effect of ceramabond 1 (a) and ceramabond 2 (b) on the I -V characteristics.

PAGE 118

118 2.2. Long Term S tability 0 50 100 150 200 250 300 350 0 100 200 300 400 500 600 700Max Power Density (mW/cm2)HoursDegradation rate = 0.248 mW / hour Y = 127.66 + 167.3exp(-0.01346 X) R = 0.9508 Figure B 11. Maximum power density of a SOFC witho ut AFL for 700 hours at 550 C. An earlier tape cast anode supported SOFC without AFL was tested at 550 C and the initial maximum power density was 316 mW/cm2 at 0.41 V. The cell voltage of 0.41 V was maintained potentio -statically and the maximum power d ensities were monitored for 700 hours at the same temperature. Figure B 11 shows a rapid degradation in cell performance for the first 200 hours of operation followed by stabilization at 130 mW/cm2. The total degradation rate of the maximum power density over 700 hours was 0.248 mW/hour. 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0 100 200 300 400 500 600 700OCP (V)HoursDegradation rate = 0.207 mV / hour Y = 0.57747 + 0.14436exp(-0.0050523 X) R = 0.93597 0 0.2 0.4 0.6 0.8 1 0 100 200 300 400 500 600 700ASR ( cm2)HoursDegradation rate = 0.335 m cm2/ hour Y = 0.67647 0.22685exp(-0.011567 X) R = 0.72261 (a) (b) Figure B 12. (a) The OCP values and (b) the ASR values at 0.41 V for 700 hours.

PAGE 119

119 Figu re B 12 shows that the main cause of the power degradation was a decrease in OCP. The degradation rates of the OCP and the ASR were 0.207 mV/hour and 0.335 m cm2/hour, respectively. However, the OCP values decrease continuously with time while the ASR tends to level off and did not contribute that much to the performance degradation. The decrease in OCP is most likely due to gas leakage. S ince this was an early cell before development of the AFL, electrolyte density was poor. So, internal leakage is quite likely resulting in a decrease in OCP with time. In addition, the observed variability in OCP with flow rate confirmed we had external se aling issues. In fact to apply SOFCs developed in this dissertation to practical fields, not only long -term stability but also thermal cycling and red -ox cycling test are necessary, which has neither been discussed n or present ed in this dissertation. It is certain that application of AFL and multilayered AFL would improve the stability of the sample, but extensive long term studies are required to attack all obstacles. 2.3. Critical Flaws Critical flaws in this section mean unusual spots or region of poor e lectrolyte quality. Sometimes, this kind of defect can be observed by the naked eye if it is large enough. This kind of defect can be formed when large agglomerates in the colloidal slurry are not broken. Since agglomerates densify more readily than the r est of the film, cracks form around agglomerates .80 Also, dust in the electrolyte green body can cause these defects to form Critical flaws can be reduced but hardly removed completely because of the limitation of simple colloidal process ing The number of critical flaws increases with the active area of the cell. For example, if we assume the number of critical flaws of current technique is 1/cm2, a sample with a 0.5 cm2 active area in a button cell has a 50 50 chance of existing in a region with a critical flaw However, a cell w ith an active area of 30 cm2, for example, should statistically contain 30 critical flaws and hence a very low probability of containing zero critical flaws.

PAGE 120

120 Figure B 13. Unusually porous region in the GDC electrolyte where other area is dense. A sudde n drop in OCP soon after H2 flow to the anode has been frequently observed during many I -V measurements over the past 4 years, and we believe this is due to critical flaws in the electrolyte. Also, chronic decrease s in OCP are in part due to this type of d efect. It is believed the critical flaws open up due to chemical expansion from the anode and lose gas tightness. When this happens during the I -V measurement, the sample fails before a day and sometimes even less than 1 hour. A b etter technique in stabili zing the GDC colloidal slurry is required. No measurements of the isoelectric point of the GDC particles used in this study have been measured and therefore pH of the solution was not optimized accordin gly. Currently the addition of 1 wt% of inorganic disp ersant to the slurry has been the only way to stabilize the slurry. Additionally maintaining a clean lab environment is very important during deposition in order to minimize dust exposure to the green body.

PAGE 121

121 LIST OF REFERENCES 1 M. Inma n, Nature Reports Climate Change 18 (2009). 2 K. Donaldson, P. W. Hadoke, N. A. Boon, W. MacNee, F. R. Cassee, T. Sandstrm, A. Blomberg, D. E. Newby, and N. L. Mills, Nature Clinical Practice Cardiovascular Medicine 6, 36 (2008). 3 A. Atkinson, S. A. Bar nett, R. J. Gorte, J. T. S. Irvine, A. J. Mcevoy, M. Mogensen, S. C. Singhal, and J. Vohs, Nature Materials 3, 17 (2004). 4 N. Minh General Electric Energy, Hybrid Power Generation Systems in http://www.netl.doe.gov/publications/proceedings/04/seca -wrkshp/seca wrkshp04.html#Industrial_Teams (2004). 5 S. d. Souza, S. J. Visco, and L. C. D. Jonghe, J. electrochem. Soc. 144, L36 (1997). 6 Q. L. Liu, K. A. Khor, and S. H. Chan, J. Power Sources 161, 123 (2006). 7 Z. Shao and S. M. Haile, Nature 431, 170 (2004). 8 E. D. Wachsman, Solid State Ionics 152-253, 657 (2002). 9 Q. L. Liu, K. A. Khor, S. H. Chan, and X. J. Chen, J. Power Sources 162, 1036 (2006). 10 C. Torres Garibay, D. Kovar, and A. Manthiram, J. Power Sources 187, 480 (2009). 11 M. D. Gross, J. M. Vohs, and R. J. Gorte, Journal of The Electrochemical Society 154, B694 (2007). 12 N. Ai, Z. Lu, J. Tang, K. Chen, X. Huang, and W. Su, J. Po wer Sources 185, 153 (2008). 13 K. F. Chen, X. J. Chen, Z. Lu, N. Ai, X. Q. Huang, and W. H. Su, Electrochimica Acta 53, 7825 (2008). 14 N. Ai, Z. Lu, K. F. Chen, X. Q. Huang, X. B. Du, and W. H. Su, J. Power Sources 171, 489 (2007). 15 A. A. E. Hassan, N. H. Menzler, G. Blass, and M. E. Ali, Adv. Eng. Mater. 4, 125 (2002).

PAGE 122

122 16 Z. R. Wang, J. Q. Qian, S. R. Wang, J. D. Cao, and T. L. Wen, Solid State Ionics 179, 1593 (2008). 17 P. Holtappels, C. Sorof, M. C. Verbraeken, S. Rambert, and U. Vogt, Fuel Cells 06, 113 (2006). 18 N. T. Hart, N. P. Brandon, M. J. Day, and N. Lapena -Rey, J. Power Sources 106, 42 (2002). 19 Y. Liu, C. Compson, and M. L. Liu, J. Power Sources 138, 194 (2004). 20 S. W. Zha, Y. L. Zhang, and M. L. Liu, Solid State Ionics 176, 25 (2005). 21 M. Ni, M. K. H. Leung, and D. Y. C. Leung, Chem. Eng. Tchnol. 30, 587 (2007). 22 X. Xu, C. Xia, G. Xiao, and D. Peng, Solid State Ionics 176, 1513 (2005). 23 B. C. H. Steele, Solid State Ionics 129, 95 (2000). 24 B. C. H. Steele and A. Heinzel, Nature 4 14, 345 (1990). 25 S. M. Haile, Acta Materialia 51, 5981 (2003). 26 B. C. H. Steele, Nature 400, 12 (1999). 27 B. C. H. Steele, Solid State Ionics 75, 157 (1995). 28 D. Jung, K. Duncan, and E. D. Wachsman, submitted to Solid State Ionics. 29 S. Omar, E. D. Wachsman, and J. C. Nino, Applied Physics Letters 91, 144106 (2007). 30 P. J. Gellings and H. J. M. Bouwmeester, The CRC handbook of Solid state electrochemistry (CRC press, Inc., 1997). 31 J. S. Ahn, M. A. Camaratta, D. Pergolesi, K. T. Lee, H. Yoon, B. W. Lee, D. W. Jung, E. Traversa, and E. D. Wachsman, Submitted to J. Elctrochem. Soc. (2009). 32 J. Nowotny, T. Bak, M. K. Nowotny, and C. C. Sorrell, Advances in Applied Ceramics 104, 154 (2005).

PAGE 123

123 33 M. Mogensen and T. Lindegaard, 1993 (The electrochemical society), p. 484. 34 T. Kenjo, S. Osawa, and K. Fujikawa, J. Electrochem. Soc. 138, 349 (1991). 35 E. P. Murray and S. A. Barnett, Solid State Ionics 143, 265 (2001). 36 S. Omar, E. D. Wachsman, and J. C. Nino, Solid State Ionics 178, 1890 (2008). 37 S. O mar, E. D. Wachsman, and J. C. Nino, Solid State Ionics 177, 3199 (2006). 38 K. Eguchi, T. Setoguchi, T. Inoue, and H. Arai, Solid State Ionics 52, 165 (1992). 39 B. C. H. Steele, Solid State Ionics 86-88, 1223 (1996). 40 E. D. Wachsman, G. R. Ball, N. Jia ng, and D. A. Stevenson, Solid State Ionics 52, 213 (1992). 41 E. D. Wachsman, P. Jayaweera, N. Jiang, D. M. Lowe, and B. G. Pound, J. Electrochem. Soc. 144, 233 (1997). 42 J. Y. Park and E. D. Wachsman, Ionics 11-2, 15 (2006). 43 J. Y. Park and E. D. Wach sman, J. Am. Ceram. Soc. 88, 2402 (2005). 44 B. A. Boukamp, K. J. d. Vries, and A. J. Burggraaf, Non -Stoichiometric Compounds: Surfaces, Grain Boundaries and Structural Defects (1989). 45 J. C. Boivin, C. Pirovano, G. Nowogrocki, G. Mairesse, P. Labrune, a nd G. Lagrange, Solid State Ionics 113-115, 639 (1998). 46 B. A. Boukamp, Solid State Ionics 136-137, 75 (2000). 47 M. Dumelie, G. Nowogrocki, and J. C. Boivin, Solid State Ionics 28-30, 524 (1988). 48 I. C. Vinke, K. Seshan, B. A. Boukamp, K. J. d. Vries, and A. J. Burggraaf, Solid State Ionics 34, 235 (1989). 49 M. A. Camaratta and E. D. Wachsman, Solid State Ionics 178, 1411 (2007). 50 M. A. Camaratta and E. D. Wachsman, Solid State Ionics 178, 1242 (2007).

PAGE 124

124 51 A. Jaiswal and E. D. Wachsman, J. Electrochem. Soc. 152, A787 (2005). 52 M. A. Camaratta and E. D. Wachsman, J. Electrochem. Soc. 155, B135 (2008). 53 Y. J. Leng, S. H. Chan, and K. A. Khor, p. 1110. 54 J. S. Ahn, H. Yoon, and E. D. Wachsman, ECS Trans. 11, 99 (2008). 55 J. S. Ahn, H. Yoon, K. T. Le e, M. A. Camaratta, and E. D. Wachsman, Submitted to Fuel Cells (2009). 56 J. Larminie and A. Dicks, Fuel Cell Systems Explained (Wiley). 57 D. R. Gaskell, Introduction to the Thermodynamics of Materials (Taylor & Francis Books, Inc., 2003). 58 J. Newman a nd K. E. Thomas -Alyea, Electrochemical systems 3rd Ed. ed. (Wiley, NewJersey, 2004). 59 A. J. Bard and L. R. Faulkner, Electrochemical Methods. Fundamentals and Applications 2nd Ed. ed. (Wiley, New York, 2001). 60 S. H. Chan, K. A. Khor, and Z. T. Xia, J Power Sources 93, 130 (2001). 61 H. Zhu and R. J. Kee, J. Power Sources 117, 61 (2003). 62 K. L. Duncan and s. E. D. Wachsman Submitted to J. Electrochem. Soc. 63 J. -W. Kim, A. V. Virkar, Z. F. K, K. Mehta, and S. C. Singhal, J. Electrochem. Soc. 146, 69 (1999). 64 K. J. Yoon, P. Zink, S. Gopalan, and U. B. Pal, J. Power Sources 172, 39 (2007). 65 J. C. Amphlett, R. M. Baumert, R. F. Mann, B. A. Peppley, P. R. Roberge, and T. J. Harris, J. Elctrochem. Soc. 142, 1 (1995). 66 J. C. Amphlett, R. M. Baumer t, R. F. Mann, B. A. Peppley, P. R. Roberge, and T. J. Harris, J. Elctrochem. Soc. 142, 9 (1995).

PAGE 125

125 67 J. S. Ahn, S. Omar, H. Yoon, J. C. Nino, and E. D. Wachsman, submitted to J. Power Sources (2008). 68 S. B. Adler, J. Elctrochem. Soc. 149, E166 (2002). 69 J. Winkler, P. V. Hendriksen, N. Bonanos, and M. Mogensen, J. Elctrochem. Soc. 145, 1184. 70 A. Pramanick, S. Omar, J. C. Nino, and J. L. Jones, J. App. Crystall. in Press (2009). 71 M. Godickemeier, K. Sasaki, L. J. Gauckler, and I. Riess, Solid State I onics 86-8, 691 (1996). 72 S. Zha, A. Moore, H. Abernathy, and M. Liu, J. Elctrochem. Soc. 151, A1128 (2004). 73 P. Shuk, H. D. Wiemhofer, U. Guth, W. Gopel, and M. Greenblatt, Solid State Ionics 89, 179 (1996). 74 H. Ghezel -Ayagh, in 9th Annual SECA Works hop (Pittsburgh, PA, 2008). 75 M. Yokoo, Y. Tabata, Y. Yoshida, H. Orui, K. Hayashi, Y. Nozaki, K. Nozawa, and H. Arai, J. Power Sources 184, 84 (2008). 76 M. Lang, C. Auer, A. Eismann, P. Szabo, and NorbertWagner, Electrochimica Acta 53, 7509 (2008). 77 H Y. Jung, S. H. Choi, H. Kim, J. -W. Son, J. Kim, H. W. Lee, and J. H. Lee, J. Power Sources 159, 478 (2006). 78 J. M. Rhodes, Thesis, University of Florida, 2005. 79 S. R. Bishop, K. L. Duncan, and E. D. Wachsman, Electrochim. Acta 54, 1436 (2009). 80 C. Lu, W. L. Worrell, C. Wang, S. Park, H. Kim, J. M. Vohs, and R. J. Gorte, Solid State Ionics 152-153, 393 (2002).

PAGE 126

126 BIOGRAPHICAL SKETCH Jin Soo Ahn was born in Seoul, Korea in 1976. After graduating f rom Un -nam High School in Seoul Korea in 1995 He stu died in the d epartment of c eramic e ngineering at Yonsei University in Korea During his stud ies he was enlisted in the Republic of Korea Army (ROKA) as a computer in a mortar team in 1996, and finished military service in 1998 Then, he returned to the Yonsei University and received his Bachelor of Science degree in c eramic e ngineering in February 2002. He entered the Michigan State University as a graduate student i n August 2002, and received the Master of Science in m aterials s cience and e ngineering in summer 2004 Finally he entered the University of Florida in 2004 and joined in Dr. Wachsman s group and received his Ph.D. from the University of Florida in the spring of 2009. S ince then, he ha s been working on my graduate research in the field of solid -state ionics with a specialty in fabrication and characterization of SOFC s