MECHANICAL, THERMAL, AND ELECTROCHEMICAL PROPERTIES OF MIXED IONIC ELECTRONIC CONDUCTOR S FROM WAFER CURVATUR E MEASUREMENTS By Yuxi Ma A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Materials Science and Engineering ŠDoctor of Philosophy 2019 ABSTRACT MECHANICAL, THERMAL, AND ELECTROCHEMICAL PROPERTIES OF MIXED IONIC ELECTRONIC CONDUCTORS FROM WAFER CURVATURE MEASUREMENTS By Yuxi Ma Mixed ionic electronic conductors (MIECs) are a group of materials that have been widely used in various applications including Solid Oxide Fuel Cells, gas separation membranes, memristors, electrostrictive actuators, chemical s ensors and catalytic converters. The functionality of these materials are based on their large, quickly -changeable point defect concentrations, which also produces a mechanical response in the material. Unfortunately, considerable disagreement over the ion ic point defect concentrations, surface exchange coefficients, and mechanical properties of even the most widely -used MIECs exists in the literature. This dissertation demonstrates that the Young™s modulus, thermo -chemical expansion coefficient, oxygen no nstoichiometry, oxygen surface exchange coefficient, oxygen surface exchange resistance and stress state of MIEC thin films can all be obtained as a function of temperature and/or oxygen partial pressure using an in-situ , non-contact, current -collector -fre e wafer curvature measurement platform. The validity of this wafer curvature technique was evaluated by experiments on identically -prepared SrTi 0.65 Fe0.35 O3-x films which showed that nearly identical oxygen surface exchange coefficients could be obtained from optical relaxation (another current -collector -free technique), and experiments on a single Pr 0.1 Ce0.9 O1.95 -x film which showed that nearly identical Young ™s Moduli could be obtained from more traditional X -ray diffraction based techniques. Wafer curvature experiments performed on Pr 0.1 Ce0.9 O1.95 -x films with Si surface impurities showed that Si can reduce the oxygen surface exchange coefficient of Pr 0.1 Ce0. 9O1.95 -x by several orders of magnitude, suggesting that Pr 0.1 Ce0.9 O1.95 -x may not be suitable for real -word Solid Oxide Fuel Cell operation in dusty conditions. Additionally, experiments performed on Pr0.1 Ce0.9 O1.95 -x films with intentionally -added Pt sur face impurities showed that the precious metal current collectors used to measure oxygen surface exchange coefficient via traditional techniques (such as electrical conductivity relaxation, electrical impedance spectroscopy, etc.) artificially enhance the Pr0.1 Ce0.9 O1.95 -x oxygen surface exchange coefficient, and hence likely contribute to the large oxygen surface exchange coefficient discrepancies observed in the literature. iv ACKNOWLEDGEMENTS This work was supported by Department of Ene rgy Award Number DE -FE0023315. The microscopy work was conducted at the Michigan State Composites Center, which is supported by the NSF Major Instrumentation Program and Michigan State University. This work made use of the Pulsed Laser Deposition Shared Fa cility at the Materials Research Center at Northwestern University supported by the National Science Foundation MRSEC program (DMR -1720139) and the Soft and Hybrid Nanotechnology Experimental (SHyNE) Resource (NSF ECCS -1542205). I would like to thank my a dvisor Dr. Jason Nicholas for his patience and guidance throughout the period of my doctoral study . The critical thinking and perseverance I learnt from him helped me a lot when research doesn™t go as planned. I would also like to thank my dissertation com mittee members Dr. Scott Barton, Dr. Thomas Bieler, and Dr. Thomas Schuelke for their guidance and support. I would like to acknowledge the tremendous help from Dr. Timothy Hogan and Mr. Karl Dersch of Michigan State University for assistance with the thin film deposition and photolithography. I would like to acknowledge the help from Mr. Per Askeland for assistance with the XPS measurements. I also appreciate the assistance of Mr. D Bruce Buchholz of Northwestern for the pulsed laser deposition. I would like to acknowledge the help from Dr. Ting Chen and Dr. Nicola Perry with the help of optical relaxation measurements. I would also like to acknowledge the help from Demetrios Tzelepis for the high temperature X -Ray diffraction measurements. I wou ld also like to thank my fellow graduate students for their support. The discussion with you really inspired me and give me a new perspective of looking at the problems. Finally, I would like to express my sincere gratitude to my parents for their selfles s support and understanding, this dissertation is dedicated to you. v TABLE OF CONTENTS LIST OF TABLES ......................................................................................................................... ix LIST OF FIGURES ........................................................................................................................ x KEY TO SYMBOLS AND ABBREVIATIONS ........................................................................ xiv 1. Introduction ................................................................................................................................. 1 1.1 Energy and Environment Challenges .................................................................................... 1 1.2 Fuel Cells and Applications .................................................................................................. 1 1.3 Advantages and Progres s of Solid Oxide Fuel Cells (SOFCs) ............................................. 2 1.4 Challenges and Objectives .................................................................................................... 4 2. Literature Review ........................................................................................................................ 8 2.1 Conventional Young™s Modulus Measurement Techniques ................................................. 8 2.1.1 Existing Young™s Modulus Measurement Techniques .................................................. 8 2.1.2 Limitation of Existing Techniques .............................................................................. 10 2.2 Mechano -Chemical Coupling and Oxygen Surface Exchange of Mixed Ionic Electronic Conductors (MIECs) ................................................................................................................. 11 2.2.1 Chemical Expansion in MIECs .................................................................................... 11 2.2.2 Oxygen Surface Exchange ........................................................................................... 12 2.3 Oxygen Surface Exchange Coefficient Measurements from Literature Studies ................ 14 2.3.1 Oxygen Permeation ...................................................................................................... 14 2.3.2 Electrical Impedance Spectroscopy (EIS) ................................................................... 14 2.3.3 Isotope Depth Profiling ................................................................................................ 14 2.3.4 Electrical Conductivity Relaxation (ECR) .................................................................. 15 2.3.5 Isot hermal Isotope Exchange ....................................................................................... 15 2.3.6 Electrical Titration ....................................................................................................... 16 2.3.7 Thermogravimetric Relaxation .................................................................................... 17 2.3.8 Optical Relaxation ....................................................................................................... 17 2.3.9 Strain Relaxation .......................................................................................................... 18 2.3.10 Limitations of k chem Measurement Techniques Reported in the Literatures .............. 18 3. Dual Substrate Measurements ................................................................................................... 20 3.1 Introduction ......................................................................................................................... 20 3.2 Theory ................................................................................................................................. 21 3.2.1 Wafer Curvature to Measure In -Situ Film Stress ........................................................ 21 3.2.2 Dual Substrate Stress -Temperatu re Measurements to Determine Film Elastic and Expansion Coefficients ......................................................................................................... 22 3.2.3 Extraction of the Film Strains, Oxygen Nonstoichiometry, and Film Stresses from Dual Substrate Stress -Temperature Measurements .............................................................. 23 3.3 Experimental Methods ........................................................................................................ 24 3.3.1 Sample Preparation ...................................................................................................... 24 3.3.2 Film Microstructure and Crystallographic Orientation Characterization .................... 27 vi 3.3.3 Du al Substrate Measurements ..................................................................................... 27 3.3.4 Chemical Strain Determination ................................................................................... 28 3.4 Results and Discussion ....................................................................................................... 29 3.4.1 Crystallography and Morphology of the Film ............................................................. 29 3.4.2 Stress vs. Temperature Data ........................................................................................ 31 3.4.3 Young™s Modulus Measurements from Dual Substrate Measurements ...................... 35 3.4.4 Thermo -Chemical Expansion Coefficient from Dual Substrate Measurements ......... 36 3.4.5 Chemical Strain & Oxygen Nonstoichiometry ............................................................ 37 3.5 Summary ............................................................................................................................. 39 4. Curvature Relaxation Measurements ........................................................................................ 40 4.1 Introduction ......................................................................................................................... 40 4.2 Theory ................................................................................................................................. 40 4.2.1 Curvature Relaxation Measurements ........................................................................... 40 4.2.2 Surface Polarization Resistance ................................................................................... 42 4.3 Experimental Methods ........................................................................................................ 43 4.3.1 Sample Preparation ...................................................................................................... 43 4.3.2 Crystallography and Morpho logy Characterization ..................................................... 43 4.3.3 Curvature Relaxation Measurements ........................................................................... 43 4.4 Results and Discussion ....................................................................................................... 44 4.4.1 Crystallography and Morphology of the Film ............................................................. 44 4.4.2 Relaxation Data and Curve Fitting .............................................................................. 44 4.4.3 Oxygen Surface Exchange Coefficient of 10PCO Thin Film ...................................... 47 4.4.4 Oxygen Surface Exchange Resistance ......................................................................... 50 4.5 Summary ............................................................................................................................. 51 5. HTXRD (High Temperature X -Ray Diffraction) - MOSS Combined Test ............................. 53 5.1 Introduction ......................................................................................................................... 53 5.2 Theory ................................................................................................................................. 54 5.3 Experimental Details ........................................................................................................... 55 5.3.1 Sample Preparation ...................................................................................................... 55 5.3.2 Microstructural and Crystallographic Characterization ............................................... 55 5.3.3 High Temperature X -Ray Diffraction Measurements ................................................. 55 5.3.4 Stress Measurements .................................................................................................... 56 5.4 Results and Discussion ....................................................................................................... 56 5.4.1 Microstructural and Crystallographic Characterzation ................................................ 56 5.4.2 HTXRD Characterization ............................................................................................ 59 5.4.3 Out -of-Plane Strain of the 10PCO Thin Film .............................................................. 63 5.4.4 Thermal Expansion Coefficient of YSZ ...................................................................... 64 5.4.5 Thermo -Chemical Expansion Coefficient via HTXRD ............................................... 65 5.4.6 Young™s Modulus via HTXRD -MOSS ........................................................................ 66 5.5 Summary ............................................................................................................................. 69 6. kchem Measurements Compared with Other Electrode -Free Techniques .................................. 70 6.1 Introduction ......................................................................................................................... 70 6.2 Theory ................................................................................................................................. 70 6.3 Experimental Details ........................................................................................................... 71 vii 6.3.1 Sample Preparation ...................................................................................................... 71 6.3.2 Crystallographic Characterization ............................................................................... 72 6.3.3 Curvature Relaxation Measurements ........................................................................... 72 6.3.4 Optical Relaxation Measurements ............................................................................... 72 6.4 Results and Discussion ....................................................................................................... 72 6.4.1 XRD Characterization .................................................................................................. 72 6.4.2 Curvature and Optical Relaxation of STF35 Thin Films ............................................. 73 6.4.3 Oxygen Surface Exchange Coefficient Comparison ................................................... 76 6.5 Summary ............................................................................................................................. 78 7. 7. Effect of Silicon Contaminants ............................................................................................. 79 7.1 Introduction ......................................................................................................................... 79 7.2 Experimental Method .......................................................................................................... 80 7.2.1 Sample Fabrication ...................................................................................................... 80 7.2.2 Microstructural and Crystallographic Characterization ............................................... 81 7.2.3 XPS Measurements ...................................................................................................... 81 7.2.4 ToF -SIMS Measurements ............................................................................................ 81 7.2.5 Curvature Relaxation Measurements ........................................................................... 81 7.3 Results and Di scussion ....................................................................................................... 82 7.3.1 Crystallography and Morphology of the Film ............................................................. 82 7.3.2 Near -Surface Si Content Characterization ................................................................... 86 7.3.3 Curvature Relaxation Measurements ........................................................................... 89 7.3.4 Oxygen Surface Exchange Kinetics ............................................................................. 93 7.4 Summary ............................................................................................................................. 96 8. Effect of Surface Platinum Coverage ....................................................................................... 97 8.1 Introduction ......................................................................................................................... 97 8.2 Experimental Details ........................................................................................................... 98 8.2.1 Pulsed Laser Deposition .............................................................................................. 98 8.2.2 Crystallographic Characterization ............................................................................... 98 8.2.3 Photolithography .......................................................................................................... 98 8.2.4 Pt Deposition and Photoresist Removal ....................................................................... 99 8.2.5 Curvature relaxation measurements ............................................................................. 99 8.3 Results and discussion ...................................................................................................... 100 8.3.1 X-Ray Diffraction Analysis ....................................................................................... 100 8.3.2 Curve Fitting of Pt|PCO|YSZ .................................................................................... 101 8.3.3 Effect of Pt surface Coverage .................................................................................... 102 8.3.4 XPS Analysis ............................................................................................................. 104 8.4 Conclusions ....................................................................................................................... 104 9. Dissertation Summary ............................................................................................................. 106 10. Future Work .......................................................................................................................... 109 APPENDICES ............................................................................................................................ 111 APPENDIX A: Deriva tion of Fitting Equation Used in Curvature Relaxation Data Processing ................................................................................................................................................. 112 viii APPENDIX B: Error Analysis for Dual Substrates Techniques ............................................ 114 APPENDIX C: Error Analysis for HTXRD -XRD Measurements ......................................... 118 BIBLIOGRAPHY ....................................................................................................................... 119 ix LIST OF TABLES Table 1.1 Fundamentals of Different Types of Fuel Cells 5-8 .......................................................... 7 Table 2.1 Characteristic thickness of common MIEC materials .................................................. 12 Table 2.2 Pros and Cons for Oxygen Surface Exchange Coef ficient Measurement Techniques 37, 44, 47 -61 ................................................................................................................................................ 19 Table 4.1 Comparison of PCO|YSZ samples between this work and Zhao et al. ........................ 48 x LIST OF FIGURES Figure 1.1 Schematics of an operating solid oxide fuel cell ........................................................... 2 Figure 1.2 Gravimetric and volumetric power density comparison of different energy conve rsion techniques 9,10 ................................................................................................................................... 3 Figure 1.3 Reported oxygen surface exchange coefficient of lanthanum strontium ferrite ............ 4 Figure 1.4 Graphic summary of objectives of this work ................................................................ 5 Figure 3.1 AFM images (a) before and (b) after the etching proces ............................................. 26 Figure 3.2 Reproducibility of the 0.2 oC/min stress -temperature data on cooling for two representative a) PCO|MgO and b) PCO|YSZ samples. ............................................................... 27 Figure 3.3 Representative X -Ray Diffraction (XRD) results for a) 10PCO|MgO and b) 10PCO|YSZ indexed using CeO 2, MgO, and YSZ JCPDS card numbers 81 -0792, 87-0653, and 70-4436, respectively. The asterisks denote impurity peaks n ot caused by the 10PCO, as proven in Figure 3.4b .................................................................................................................................... 29 Figure 3.4 Representative a) survey and b) detailed X -Ray Diffraction (XRD) scans of a bare (100) oriented MgO wafer showing that the 34.2 o and 35.2 o peaks shown in Figure 3.3 (denoted by asterisks) were not from the 10PCO film. .................................................................................... 30 Figure 3.5 SEM Images of (a) 10PCO|MgO and (b) 10PCO|YSZ ............................................... 31 Figure 3.6 Representative averaged stress vs. temperature plots for a) 10PCO|MgO and b) 10PCO|YSZ. The red lines are fits to the measured data. ............................................................. 33 Figure 3.7 The effect of cooling rate on the stress vs. temperature measurements of representative 10PCO|YSZ samples. Note, samples tested with either 0.1 or 0.1 oC/min cooling rates were considered to be in thermal equilibrium f rom 280 -700oC since the additional tim time at each temperature with a slower 0.1 oC/min cooling rate had no effect on the stress -temperature trajectory. ...................................................................................................................................... 34 Figure 3.8 PCO Young™s moduli values measured by the Dual Substrate method compared to the literature measurements. (DS stands f or Dual Substrate, NI stands for Nano -Indentation, MOSS stands for Multi -beam Optical Stress Sensor, and XRD stands for X -Ray Diffraction) .............. 35 Figure 3.9 10PCO thermo -chemical expansion coefficients from the Dual Substrate method compared to the literature measurem ents of Bishop et al. (DS stands for Dual Substrate, XRD stands for X -Ray Diffraction, and Dil stands for Dilatometry). ................................................... 37 Figure 3.10 10PCO (a) chemical strain and (b) oxygen nonstoichiometry values from the 10PCO|YSZ sample compared to the literature measurements .................................................... 38 xi Figure 4.1 Representative raw curvature rela xation data for a 10PCO|YSZ sample at 600 oC ..... 44 Figure 4.2 a) Normalized curvature fits to the 10PCO|YSZ data, and b) ln(1 -Normalized Curvature) plots for the 10PCO|YSZ samples. Note that the generally good single time constant (red line) fits to the data of part a) and a single slope in the ln(1 -Normalized Curvature) plots in part b) at times before the equilibrium state is reached suggest only one mechano -chemically active process is active with redox cycling between synthetic air and 10% diluted synthetic air. .......................... 45 Figure 4.3 Raw wafer curvature relaxation data for the 10PCO|YSZ samples from all tested temperatures (500, 525, 550, 575, and 600 oC) ............................................................................. 46 Figure 4.4 10PCO chemical oxygen surface exchange coefficients from the curvature relaxation B stands for Micro -Balance, EIS stands for Electric Impedance Spectroscopy and OR stands for Optical Relaxation). The curvature -determined k error is less than the size of the symbol ...................... 47 Figure 4.5 10PCO oxygen surface polarization resistance (R S) values obtained here compared to the literature values for 10PCO, LSCF, LSF, LSC, and SSC ....................................................... 50 Figure 5.1 X -Ray diffraction survey scan of 10PCO|YSZ at 25 oC, the asterisk denotes the (400) YSZ diffraction peak from tungsten radiation .............................................................................. 57 Figure 5.2 Pole figure and zoomed -in image of (111) peak of preferred -oriented 10PCO thin film at room temperature ...................................................................................................................... 58 Figure 5.3 Raw data summary of HTXRD measurements from 2 5 to 700 oC .............................. 59 Figure 5.4 Ful l width half max of (200) 10PCO peak at different temperatures during HTXRD measurements ................................................................................................................................ 60 Figure 5.5 XRD scans of (400) YSZ peak at different temperatures during HTXRD measurements ....................................................................................................................................................... 61 Figure 5.6 Peak fitting of HTXRD data using Pearson VII function ............................................ 62 Figure 5.7 Out -of-plane thermo -chemical strain and its fitting of 10PCO|YSZ sample .............. 63 Figure 5.8 Thermal expansion coefficient of YSZ substrate ........................................................ 64 Figure 5.9 Thermo -chemical expansion coefficient of 10PCO measured from this work, compare d with other literature studies ........................................................................................................... 65 Figure 5.10 Stress change vs. strain change plot of 10PCO thin film .......................................... 67 Figure 5.11 Young™s modulus measured in this work, compared with other literature studies. .. 68 Figure 6.1 Representative XRD scan for STF35|YSZ sample ...................................................... 73 Figure 6.2 Raw data of (a) curvature relaxation and (b) optical relaxation at 700 oC ................... 74 xii Figure 6.3 Representative curve fitting of (a) curvature relaxation and (b) optical relaxation of reduction relaxation at 700 oC........................................................................................................ 75 Figure 6.4 Comparison of kchem values measured from curvature relaxation (kR) and optical relaxation (OTR) ........................................................................................................................... 77 Figure 7.1 Representative XRD scans of (a) etched and (b) non -etched 1 0PCO|YSZ samples using using CeO 2, and YSZ JCPDS card numbers 81 -0792 and 70 -4436, respectively. ....................... 83 Figure 7.2 Cross -sectional back -scattered electron image of 10PCO|YSZ .................................. 84 Figure 7.3 A FM scans of 10PCO|YSZ sample (a) before and (b) after etching ........................... 85 Figure 7.4 XPS spectra of (a) non -etched and (b) etched sample aged at different temperatures, noting that below 600 oC, there is no Si contamination detected for etched samples .................. 86 Figure 7.5 Near -surface Si concentration of etched and 600 oC-aged 10 PCO|YSZ measured by ToF -SIMS ..................................................................................................................................... 88 Figure 7.6 Representative curvature relaxation data of Si -free and Si -contaminated samples. The relaxation time for the Si -contaminated sample is significantly longer than Si -free sample (note difference in horizontal time scale) ............................................................................................... 89 Figure 7.7 Curvature relaxation raw data at different temperatures for (a) Si -free sample and (b) Si-contaminated sample, note that the relaxation time of the Si -contaminated sample is significantly longer than etched sample ........................................................................................ 90 Figure 7.8 Multiple 750oC Curvature relaxations performed on the same 10PCO thin film by switching between (a) synthetic air (pO 2=0.21) and a 10 times diluted air (10% synthetic air - 90% Ar, pO 2=0.021) mixture, and (b) synthetic air (pO 2=0.21) and a 5 times diluted air (20% synthetic air Œ 80% Ar, pO 2=0.042) mixture ................................................................................................ 91 Figure 7.9 Fittings of curvature relaxation measurements at different temperatures for non -etched samples .......................................................................................................................................... 92 Figure 7.10 ln(1 -Normalized Curvature) plo ts for the non -etched sample. Note that the generally good single time constant (red line) fits to the data of Figure 7.9 and a single slope in the ln(1 - Normalized Curvature) plots of Figure 7.10 at times before the equilibrium state is reached suggest only one mechano -chemically active process is active with redox cycling between synthetic air and 10% diluted synthetic air. ....................................................................................................... 93 Figure 7.11 (a)10PCO oxygen surface exchange coefficient from curvature relaxation comparing with other literature studies (b) 10PCO surface polarization resist ance (R S) values compared to literature reporting 10PCO R S values and other R S values of conventional solid oxide fuel cell cathode materials .......................................................................................................................... 94 Figure 8.1 Image of bare YSZ wafer, PCO|YSZ and Pt|PCO|YSZ .............................................. 99 xiii Figure 8.2 XRD scan of PCO|YSZ sample ................................................................................. 100 Figure 8.3 Representative curve fitting of Pt|PCO|YSZ at 725oC ............................................. 101 Figure 8.4 k chem comparison between PCO|YSZ and Pt|PCO|YSZ ............................................ 103 Figure 8.5 XPS scan for as -deposited PCO|YSZ ........................................................................ 104 Figure 9.1 Graphic summary of objectives of this work ............................................................ 106 xiv KEY TO SYMBOLS AND ABBREVIATIONS A Area a Lattice parameter AFC Alkaline fuel cell AFM Atomic force microscopy thermochemical Thermo -chemical expansion coefficient Chemical expansion coefficient Thermo -chemical expansion coefficient and Thermal expansion coefficients of substrate 1 and 2 Change in optical absorption coefficient b Width Dimensionless constant regarding the geometry of the indenter CHP Combined heat and power Concentration of oxygen vacancy Oxygen concentration D Diffusion coefficient D* Tracer diffusion coefficient Dchem Chemical diffusion coefficient Oxygen nonstoichiometry E Young™s modulus ECR Electrical conductivity relaxation EIS Electrical impedance spectroscopy Young™s modulus of the film xv Young™s modulus of the substrate Effective Young™s modulus Strain Chemical strain F Faraday™s constant Flexural frequency Thermodynamic factor Film thickness Substrate thickness HTXRD High temperature X -Ray diffraction I Transmitted light intensity IET Impulse excitation technique Exchange current density kchem Chemical oxygen surface exchange coefficient k Oxygen surface exchange coefficient Tracer oxygen surface exchange coefficient Electrically measured oxygen surface exchange coefficient Curvature of the sample R Curvature relaxation l Film thickness Characteristic thickness L Sample length Initial gauge length xvi Change in gauge length LSM Lanthanum strontium manganese Surface permeability Biaxial modulus of the thin film Biaxial modulus of the substrate MCFC Molten carbonate fuel cell MgO Magnesium oxide MIEC Mixed ionic electronic conductor MOSS Multi-beam optical stress sensor Poisson™s ratio Surface oxygen chemical potential Avogadro ™s constant OTR Optical relaxation Ox Oxidation PAFC Phosphoric acid fuel cell PEMFC Proton exchange membrane fuel cell pO2 Oxygen partial pressure PCO Praseodymium doped ceria PLD Pulsed laser deposition Rs Surface polarization resistance R Ideal gas constant Red Reduction Density xvii S Unloading stiffness SAFC Solid acid fuel cell SEM Scanning electron microscopy SOFC Solid oxide fuel cell Stress Chemical stress Thermal stress Total stress ToF -SIMS Time -of-flight secondary ion mass spectroscopy v Poisson™s ratio Poisson™s ratio of the substrate Poisson™s ratio of the fil m Wave velocity at longitude direction Molar volume Thermodynamic factor XRD X-Ray Diffraction XPS X-Ray photoelectron spectroscopy YSZ Yttria -doped zirconia 1 1. Introduction 1.1 Energy and Environment Challenges With the growth of population and the development of industry , the energy demand is steadily increasing over the past decade and will continue to grow in the future 1. Also , the emission of carbon dioxide increased accordingly 1. Utilizing the existing non -renewable fossil fuel efficiently ha s become a must to meet the energy demand as well as reducing the emission of carbon dioxide. Conventional energy conversion technique s like internal combust ion not only have low efficiencies , but combustion exhaust also contains pollutant s such as carbon monoxide, nitric oxide , and sulfur dioxide. 2 Therefore, new energy conversion technique s that are both efficient and environmentally friendly is ne eded to tackle the se energy and environment challenges. 1.2 Fuel Cells and Applications A fuel cell is an energy conversion techn ology that converts chemical energy into electrical energy. There are different types of fuel cells , including the proton exchange membrane fuel cell (PEMFC), phosphoric acid fuel cell (PAFC), solid acid fuel cell (SAFC), alkaline fuel cell (AFC), solid oxide fuel cell (SOFC), and molten carbonate fuel cell (MCFC). Based on the operating temperature and the power output , fuel cells can be utilized for a variety of applications , ranging from portable devices (100 W) to a centralized power generator (100 MW) .3 Because of its versatility, 62, 000 fuel cell systems ha ve been shipped worldwide in 2016, and the total megawatt s of fuel cells shipped worldwide increased from 300 MW at 2015 to over 500 MW at 2016. 4 2 1.3 Advantages and Progress of Solid Oxide Fuel Cells (SOFCs) Figure 1.1 shows the schematic of a SOFC. At high temperature (600~1000°C), one oxygen ion gets incorporated into the cathode materials, transported through electrolyte, and then react s with the fuel (H 2, hydrocarbons , CO) at the anode side. Meanwhile, to maintain the electroneutrality, two electrons are transported through the interconnect between anode and cathode , which generates electricity. Compar ed to other types of fuel cell s, as shown in Table 1.1 SOFC has the highest combined head and power (CHP) e fficiency and the most flexible fuel choices. 5-8 When compar ed to other ene rgy conversion techniques, as shown in Figure 1.2 9, 10 , SOFC has the highest gravimetric and volumetric power density. However, SOFC requires high operating temperature for the transport of oxygen ion s. Degradation s processes such as particle coarsening, Figure 1.1 Schematics of an operating solid oxide fuel cell Electrolyte Cathode: ()++2Anode: ()++2+()23 interdiffusion between different components, and dopant segregation can shorten the lifetime of SOFC and therefore hinder ing the further development of SOFC . To mitigate the degradation issue, one solution is to decrease the operating temperature of SOFC. Over the past decade, a variety of studies have been conducted to develop new cathode materials or different cathode microstructure s to reduce the operating temperature .11 -18 As a result, with the help of the high -performance cathodes that have the fast oxygen surface exchange kinetics, the operating temperature of SOFC s have been brought down to ~500 oC,17 a significant advance from the previous prototype using a lanthanum strontium manganese (LSM) cathode, which operates at 1000oC. Figure 1.2 Gravimetric and volumetric power density comparison of different energy conversion techniques 9,10 4 1.4 Challenges and Objectives In order to incorporate a cathode material into a SOFC, multiple physical/chemical properties must be known in order to meet the performance and stability requirements. For example, Young™s modulus and the thermo -chemical expansion coefficient are needed to evaluate the mechanical stability at the interfa ce between cathode and electrolyte. The o xygen surface exchange coefficient is needed to quantify the performance of cathode materials. Th ese properties have to be measured in-opera ndo to be meaningful for SOFC applications. However, due to the difficulty of in-situ measurements, there is a lack of studies about Young™s modulus at high temperature. Additionally, as shown in Figure 1.3, 19 -22 for a conventional SOFC cathode material, there is ~5 orders of magnitudes discrepancy in the oxygen surface exchange coefficient. This discrepancy makes it even harder to compare the performance across different cathode materials. Figure 1.3 Reported oxygen surface exchange coefficient of lanthanum strontium ferrite k1 k2 k1 There is a 5 order of magnitude discrepancy in k at SOFC operating temperatures !5 Therefore, the objective of t his work , as shown in Figure 1.4, 23 is to develop an in-situ technique that can measure stress ( ), strain ( ), Young™s modulus (E), biaxial modulus ( M), thermo -chemical e xpansion coefficient (thermochemical ), thermal expansion coefficient ( thermal ) oxygen surface exchange coefficient (kchem ), oxygen nonstoichiometry (), and surface polarization resistance (Rs); Then us e this technique to detect the effect of surface contaminants on the oxygen surface exchange process. Of the following chapters, Chapter 2 will review basic knowledge of mix ed ionic electronic conductors (MIECs), chemical expansion, oxygen surface exchange process, and the conventional Figure 1.4 Graphic summary of objectives of this work 6 characterization techniques to measure Young™s modulus . Chapter 3 will introduce the curvature measurement platform, multibeam optical stress sensor (MOSS), and its application using the dual substrate method to obtain Young™s modulus, the thermo -expansion coefficient and oxygen nonstoichiometry from thin film materials. Chapter 4 will examine the curvature relaxation technique and its applications in measuring the oxygen surface exchange coefficient and surface polarization resistance. Chapter s 5 and 6 will illustrate the results of Young™s modulus, thermo -chem ical expansion coefficient , and oxygen surface exchange coefficient from other measurement techniques for cross -checking the reliability of the MOSS platform. Chapter s 7 and 8 will report the effect of surface contaminants on the oxygen surface exchange pr ocess. Chapter 9 will sum up the conclusions from this dissertation and present recommendations for future work . 7 Table 1.1 Fundamentals of Different Types of Fu el Cells 5-8 Fuel Cell Electrolyte Operating Temperature Fuel Charge Carriers Electrode Reactions Electrical Efficiency CHP Efficiency Problems PMEFC Nafion 70-110°C H2H+, e-40-50% NAH2O Management, CO intolerant AFC Aqueous KOH100-250°C H2OH-, e-50%NACO2 intolerant PAFC H3PO4150-250°C H2H+, e-40%NAElectrolyte Leaks MCFC (Na,K) 2CO3500-700°C HCs, COCO32-, e-50%80%Long Start-up, corrosive electrolyte SOFC Y2O3-ZrO 2, Gd2O3-CeO 2600-1000°C H2, CO, HCs O2-, e-45-65% 90%Long start-up, high temperature degradation 12()++2()2+212()++2()2+212()+()+22 ()+2 2()+212()+()+2()+()+()+212()+2()+()+28 2. Literature Review 2.1 Conventional Young™s Modulus Measurement Technique s 2.1.1 Existing Young™s Modulus Measurement Techniques Tensile test and nano -indentation are two of the most commonly used measurement techniques to measure Young™s modulus. For tensile test, the sample is place d in the testing instrument and then extended . The elongation of the sample and the force are record ed during the test. To calculate the Young™s modulus, one needs to know the stress and s train of the tested sample, which are defined as: =(2.1) and =(2.2) where is the stress, is the tensile force, A is t he cross -section area, is the strain, is the change in gauge length, and is the initial gauge length. Based on the equation: E=(2.3) Young™s modulus can be calculated. Nano -indentation, on the other hand, uses the slope of the load -displacement curve during unloading to calculate the Young™s modulus of the sample. Specifically, b ased on the slope of loading -displacement curve, namely the unloading stif fness, the effective Young™s modulus can be calculated: 24 S=2(2.4) 9 where S is the unloading stiffness, is a dimensionless constant regarding the geometry of the indenter, is the contact area, and is the effective Young™s modulus, which is defined as: 24 1=1+1(2.5) where and are the Young™s modulus and Poisson™s ratio of the sample, and are the Young™s modulus and Poisson™s ratio of the indenter. The i mpulse excitation technique (IET) and ultrasonic wave technique are two of the methods to measure Young™s modulus based on the propagation of the wave . The IET me asures the resonant frequency in order to extract Young™s modulus of the measured sample. The excitation is induced by using a small projectile to tap the sample. Then the result ing vibration signal is recorded with a microphone /piezoelectric sensor/laser vibrometer/accelerometer. The acquired data in the time domain is processed with fast Fourier transformation . Then the resonant frequency is determined to calculate Young™s modulus. The equations for calculating the Young™s modulus depends on the geometry of the sample. For example, for a rectangular bar, the Young™s modulus can be calculated by: 25 E=0.9465(2.6) where E is the Young™s modulus, m is the mass, is the flexural frequency, b is the width, L is the length, t is the t hickness, and T is the correction factor, which is defined as: T=1+6.585(2.7) Note that the correction factor will be necessary only if L/t 20. 10 The u ltrasonic wave technique determines the elastic properties of the material by detecting the propagation speed of ultrasonic wave s. Specifically, the elastic modulus and density determines the velocity of the ultrasonic wave :26 =(2.8)where V is the velocity of the wave propagation, M is the elastic moduli, and is the density of the sample. For isotropic materials, the Young™s modulus can be determined by :26 =(1+)(12)/(1)(2.9) where is the wave velocity at longitude direction, is the Young™s modulus, and is the Poisso n™s ratio. 2.1.2 Limitation of Existing Techniques For tensile test s and nano -indentation, the first limitation is the destructive testing procedure. The second limitation is that it is hard to incorporate the testing instrument into a high temperature environm ent. For nano -indentation, specifically, as described in Eqn 2.5, the change of Young™s modulus at high temperature could also happens to the indenter, which may lead to the inaccura te results. For IET , although it is non -destructive, the constant used in determining Young™s modulus, as shown in Eqn 2.6 varies among different materials and sample geometries .25 Additionally, it takes special design of the test rig for the IET to measure Young™s modulus a t high temperature. 25 For the ultrasonic wave technique, Young™s modulus is calculated under the assumption of stress -free sample , which doesn™t hold for the case of a thin film on a rigid substrate . Again, it is hard to incorporate the whole system into a high temperature environment. 11 2.2 Mechano -Chemical Coupling and Oxygen Surface Exchange of Mixed Ionic Electronic Conductors (MIECs) 2.2.1 Chemical Expansion in MIECs Mixed ionic electronic conductor s are a group of materials that can conduct both ion and elect rons. For the MIEC used as cathode material under SOFC operation conditions, it goes through the following defect reaction, expressed in Kröger -Vink notation: 27 12++2(2.10) This reaction can be triggered by increasing the sample temperature and/or decreasing the oxygen partial pressure. After the formation of an oxygen vacancy, the lattice tends to expand due to the expulsion resulting in a positively charged cation and the oxygen vacancy. This behavior is described as mechano -chemical coupling. The resulting strain from a changing chemical environment can be expressed by the following equation 28 : =(2.11) where is the chemical strain, is the chemical expansion coefficient, is the oxygen nonstoichiometry of the material, and is the change in the oxygen nonstoichiometry of the material. is related to the concentration of oxygen vacanc ies : =(2.12) where is the concentration of oxygen vacancy, is the molar volume of the material , and is the Avogadro™s number. For conventional MIEC used in SOFC applications, the chemical expansion coefficient ranges from 0.01 to 0.1, 29 even with a typical change of 0.02 in oxygen nonstoichiometry, the result ing chemical expansion could be quite substantial compar ed to the thermal expansion , leading to internal strain energy that can drive microstructural changes and 12 damage . Therefore, being able to measure chemical expansion of a MIEC material can be of great importance for the mechanical stability of SOFC applications. 2.2.2 Oxygen Surface Exchange For oxygen transport in MIECs, it is typically a mixed combination of oxygen diffusion and oxygen surface exchange processes. The oxygen exchange process, specificall y, consists of oxygen adsorption and dissociation. The competition between those two processes can be described with the characteristic thickness (detailed derivation can be found in Appendix A): 30 =(2.13) where is the characteristic thickness, D is the diffusion coefficient, an d k is the oxygen surface exchange coefficient. If the sample thickness l is less than 1/100 th of the characteristic thickness, the oxygen transport is predominately controlled by oxygen surface exchange kinetics. 19 Table 2. 1 shows the characteristic thickness of commonly used MIEC materials for SOFC applications .18, 31 -Table 2.1 Characteristic thickness of common MIEC materials Temperature ( oC)pO2 (kPa) Characteristic Thickness (cm) La0.5 Sr0.5 MnO 70070.02*10-7La0.8 Sr0.2 CoO 70070.03*10-3La0.9 Sr0.1 CoO 9004.52*10-3La0.9 Sr0.1 FeO 9006.56*10-3La0.6 Sr0.4 FeO 10006.55*10-2800101.32*10-380020.76*10-38005.12*10-28001.35*10-2Pr 0.1 Ce0.9 O 67020.70.4La0.6 Sr0.4 Co0.2 Fe 0.8 O 13 33 With the current geometries of SOFC cathodes, the electrode size is much smaller than the characteri stic thickness. Therefore, oxygen surface exchange process is the rate limiting step of oxygen transport in the MIECs used for SOFC applications . Depending on the measurement techniques, there are three types of oxygen exchange coefficients in the literat ure, which include chemical (ambipolar) exchange coefficient (k chem ), tracer exchange coefficient (k*), and electrically determined exchange coefficient (k q). The chemical exchange coefficient is defined as: 34 =4==(2.14) where is the exchange current density, is the oxygen concentration, is the surface oxygen chemical potential, R is the ideal gas constant, T is the temperature, is the surface permeability, and is the thermodynamic factor, which is defined as (detailed derivation of thermodynamic factor can be found in Ref 34) :35 =1 (2.15) The chemical exchange coefficient repres ents the exchange process involving oxygen vacancies and electrons without the applied current or voltage . The tracer exchange coefficient is defined as: 34 =(2.16) which only represents the exchange process of oxygen vacancies without the involvement of electrons . The electrically determined exchange coefficient is defined as: 34 =(2.17) 14 which also represents the exchange process containing oxygen vacancies and electrons. However, in this cas e, there is an external circuit, therefore the oxygen flux is not related with the electron flux. 2.3 Oxygen Surface Excha nge Coefficient Measurement s from Literature Studies 2.3.1 Oxygen Permeation The o xygen permeation technique requires a membrane material to be sealed inside a chamber with feed gas on one side and sweep gas on the other side, and the flux of oxygen is calculated via the composition results analyzed by gas chromatography. The measured c hemical oxygen exchange coefficient (k chem ) is related to the permeability as described in Eqn 2.14. 2.3.2 Electrical Impedance Spectroscopy (EIS) The EIS requires a symmetrical cell in order to measure the oxygen surface exchange coefficient. The symmetrical c ell is composed of one electrolyte and two identical electrodes at each side of the electrolyte. After applying an AC voltage across the cell and measuring the impedance spectroscopy of the sample , the resistance of the electrode is equal to the diameter o f the semi -circle impedance signal divided by two (considering two identical electrodes in the measurement). Since the oxygen surface exchange process is the rate limiting step of the oxygen transport process in the electrode, therefore, the area l specific resistance of the electrode is related to the electrically determined oxygen exchange coefficient (k q):36 =4(2.18) where F is the Faraday™s constant, is the area specific resistance. 2.3.3 Isotope Depth Profiling Isotope depth profiling u ses the 18 O isotope to determine both isotope oxygen surface exchange (k*) and diffusion (D*) coefficient of a material. Specifically, the sample pre -annealed 15 in normal gas atmosphere (normal 18 O/16 O abundance) is exposed to 18 O2 at designated temperature. Then afte r quenching the sample to room temperature, the composition profile of 18 O is determined by secondary ion mass spectroscopy (SIMS). According to Fick™s second law, the 18 O profile is fitted with the following equation: 37 (,)=(,)= 2exp (+) 2+(2.19) where g(x, t) is the normalized concentration of oxygen isotope, c(x, t) is the instantaneous concentration of oxygen isotope, is the oxygen isotope concentration at the background, is the oxygen isotope concentration in the gas, t is the expos ure time, and =. 2.3.4 Electrical Conductivity Relaxation (ECR) The ECR technique relies on the relaxation of the conductivity after an abrupt change in oxygen partial pressure in the surrounding gas atmosphere of the sample. Under the assumption of a li near relationship between oxygen concentration and conductivity (assuming constant mobility of oxygen ion), with the sample thickness significantly smaller than the characterization thickness (), the k chem value can be determined by fitting the cond uctivity relaxation curve to a simplified version of Fick™s second law: 38 -41 ()=()=()=1exp (2.20) where () is the instantaneous conductivity, is the initial conductivity and is the final conductivity. (), , represents the instantaneous, initial and final oxygen concentration, respectively. The detailed derivation of Eqn 2.20 can be found in Appendix A. 2.3.5 Isothermal Isotope Exchange Isothermal isotope exchange is an in-situ technique to measure the isotope oxy gen surface exchange coefficient of powder materials . It monitors the change in 18 O concentration at the gas 16 outlet after an abrupt change in oxygen partial pressure. The concentration relaxation of 18 O can be fitted with one of the solution s to Fick™s sec ond law: 42 ()=()=16exp [+(1)](2.21) where +1=0(2.22)and d ==(2.23) is the radius of particles which are assumed to be spherical , and is the characteristic thickness. 2.3.6 Electrical Titration Electrical titration uses the current relaxation of a n electrochemical cell after a small step change of potential to determines the chemical oxygen surface exchange coefficient . Specifically, the electrochemical cell is surrounded b y the gas atmosphere of interest, and sealed in a testing chamber. After a step change in potential, the oxygen vacanc ies were generated, therefore creating the driving force for oxygen exchange. The exchange process is accompanied with current relaxation, by fitting the current relaxation to a solution of Fick™s second law in order to obtain the chemical oxygen surface exchange coefficient can be acquired :43 ()=2exp (++)(2.24) where Q is the amount of passing charge within the time of t, is defined as: =(2.25)and 17 =(2.26) 2.3.7 Thermogravimetric Relaxation Thermogravimetric relaxation measures the weight change of the sample after an abrupt change in oxygen partial pressure in order to obtain chemical oxygen surface exchange coefficient. The relaxation of weight can be fitted with :44 ()=()=()=12exp [++](2.27) where m is the mass of the sample , and is defined as: ===(2.28) 2.3.8 Optical Relaxation Optical relaxation measures the relaxation of transmitt ed light intensity to determine the chemical oxygen surface exchange coefficient. Specifically, some MIEC materials (e.g. pra seodymium doped ceria) will a bsorb light with certain wavelengths due to the effect of the dopant. The oxidation state of the dopant is related to the level of a bsorption and is also related to the concentration of oxygen vacanc ies . Therefore, after an abrupt change of oxygen partial pressure, by monitoring the relaxation of transmitted light intensity , the chemical oxygen surface exchange coefficient can be determined by fitting the relaxation curve with the following equation: 45 ()= () =()=1exp (2.29) where is the transmitted light intensity. 18 2.3.9 Strain Relaxation Strain Relaxation measures the lattice parameter relaxation to obtain chemical oxygen sur face exchange coefficient. Specifically, a time -resolved X -Ray diffractometer is needed to measure the simultaneous lattice parameter change (chemical strain change) after an abrupt change in oxygen partial pressure. 46 Assuming a linear relationship between chemical strain and oxygen nonstoichiometry (const ant chemical expansion coefficient), the kchem value can be acquired by fitting the relaxation curve with Eqn 2.27 and 2.28. 2.3.10 Limitations of k chem Measurement Techniques Reported in the Literatures As shown in Table 2.1, these techniques have their own limitations .37, 44, 47 -61 An affordable, current collector -free, contact -free, versatile, in-situ technique that can measure oxygen surface exchange coefficient at different temperature and gas atmosphere is n eeded to be able to evaluate the performance of various MIEC materials 19 Table 2.2 Pros and Cons for Oxygen Surface Exchange Coefficient Measurement Techniques 37, 44, 47 -61 Conductivity Relaxation Permeation Electrical Titration Impedance Spectroscopy Gravimetric Optical Relaxation Strain Relaxation Isotope Depth Profiling Isothermal Isotope Exchange kchem /kq/k* kchem kchem kchem kqkchem kchem kchem k*k*Current Collector Needed? Yes NoYes Yes NoNoNoNoNoin-situ Yes Yes Yes Yes Yes Yes Yes NoNoContact- Free? NoNoNoNoYes Yes Yes Yes Yes Thin Film Capable? Yes Yes Yes Yes NoYes Yes Yes Yes Versitality Good Good Good Good Good Bad Good Good Good Cost $$$$$$$$$$$$$$$$$$$$20 3. Dual Substrate Measurements (This chapter has been adapted from a published paper) 3.1 Introduction A variety of advanced functional materials, including those used in batteries, 62, 63 fuel cells, 64 -68 electrolysis cells, 69 -72 memristors, 73, 74 electrostrictive actuators, 75, 76 gas separation units, 77 chemical sensors, 78 electrochromic windows, 79 catalytic converters, 80 etc., obtain their functionality from a high concentration of ionic point defe cts. Since point defect concentration changes typically result in lattice parameter changes, 67, 81, 82 a coupling exists between the mechanical and electrochemical states of most high performance mixed ionic electronic conducting (MIEC) materials. 83 In traditional batteries, solid oxide fuel cells (SOFCs) and other electrochemical devices made from bulk particles (i.e. grain sizes >> ~100 nm), this mechano -chemical coupling i s problematic because it produces stress gradients that crack and mechanically pulverize the material if, and when, a material experiences compositional changes during device operation. 84 -86 However, the higher fracture toughnesses and higher Griffiths critical crackin g stresses exhibited by thin film materials 87 have spurred intere st in using externally applied stress to intentionally increase the point defect concentrations and electrochemical performance of thin film MIEC devices. 88 -94 For either situation, knowledge of a material™s in situ/ in operando mechanical, thermal and electrochemical performance is critical for engineering the stress profiles that help determine device performance and/or durability. Unfortuna tely, such data is scarce in the existing literature, especially at the elevated temperatures often encountered during device operation. This Chapter demonstrates that the biaxial modulus ( ), Young™s Modulus ( ), thermal expansion coefficient ( ), thermo -chemical expansion coefficient ( ), oxygen 21 nonstoichiometry ( ), of Pr 0.1 Ce0.9 O1.95 (10PCO) can all be obtained as a function of simultaneously measured 10PCO chemical stress ( ), chemical strain ( ), thermal stress ( ), thermal str ain ( ), total stress ( ), total strain ( ), temperature ( ), and oxygen partial pressure ( ) conditions using in situ , non-contact, current -collector -free wafer curvature measurements. Doped ceria was chosen for study based on its importance as a catalytic converter oxidation catalyst, 80, 95 oxygen sensor material, 96 water -splitting/alternative fuel production catalyst, 72 ,97 and SOFC material. 18, 98, 99 10PCO, in particular, was chosen because of its conveniently large chemical expansion coefficient, 100 easily access ible mixed ionic electronic conducting state (i.e. mechano -chemically active regime), 101 well -established point defect model 102 and status as a model material. 102 3.2 Theory 3.2.1 Wafer Curvature to Measure In -Situ Film Stress Mechanics theory indicates that the average biaxial stress ( ) within a dense thin film on top of a dense thick substrate (such that the film thickness ( ) to sub strate thickness ( ) ratio is less than 0.001) can be extracted from the wafer curvature ( ) (without knowledge of the film elastic properties) using Stoney™s Equation: =6(3.1) where is the substrate biaxial modulus defined as /(1), is the substrate Young™s modulus, and is the substrate Poisson™s Ratio. 103 -105 Hence Eqn. [1] was used to extract the in situ film stress from the wafer curvature using the procedures described in the Experimental Methods of Section 3.3 . Detailed equipment set -up and curvature measurement guide can be found in our previous work .106 22 3.2.2 Dual Substrate Stress -Temperatu re Measurements to Determine Film Elastic and Expansion Coefficients Previous studies have shown that the stress -temperature behavior of electrochemically inactive thin films on multiple substrates can be used to determine thin film elastic constants and thermal expansion coefficients. 107 -109 Here, this approach was extended to also measure the thin film thermo -chemical expansion coefficients of mechano -chemically active materials. Specifically, the stress -temperature derivatives of 10PCO thin films atop two mechano -chemically inactive substrates with different thermal expansion coefficients (i.e. (Y 2O3)0.095 (ZrO 2)0.905 (YSZ) and MgO with 280 -700oC averaged values of 9.5 and 14.3 ppm/ oC, respectively) were measured and related to the film biaxial modulus ( ), the substrate thermal expansion coefficients ( and ), and the film thermo -chemical expansion coefficient () using the relationships: =()(3.2) and =()(3.3) Application of temper ature -dependent substrate thermal expansion data calculated from the literature 110, 111 to Equations 3.2 and 3 .3, allowed and to be determined by solving these two equations (bo th with two unknowns) simultaneously. For those temperatures where mechano -chemical coupling was inactive as indicated by previous 10PCO oxygen nonstoichiometry measurements (i.e. below ~400 oC), 101, 102, 112 -115 the thermochemic al expansion coefficient was treated as simply representing the thermal (not thermo -chemical) expansion coefficient. The film Young™s modulus ( ) was then determined from the measured data using the definition of the biaxial modulus: 23 =1(3.4) by assuming a 10PCO film Poisson™s ratio ( ) of 0.33, as has been done previously in the literature. 116, 117 (Note, could also have been measured directly by performing experiments on anisotropic substrates, as has been done in the literature, 118 but this was not attempted here due to the minor temperature variation in Poisson™s Ratio observed for most materials, even as they encounter oxygen nonstoichiometries simila r to those encountered here 119, 120 ). 3.2.3 Extraction of the Film Strains, Oxygen Nonstoichiometry, and Film Stresses from Dual Substrate Stress -Temperature Measurements The total film strain ( ) was extracted from the measured film stress ( ) by assuming the film and substrate behaved as elastic solids and applying the thin film version of Hooke™s Law: 104 = (3.5) The ceria, YSZ, and MgO in this stu dy were elastic over the entire 280 -700 oC temperature range evaluated here as demonstrated by the reproducible stress -curvature trajectories in in Figure S2 of the Supplemental Materials. This is consistent with the disappearance of ceria™s oxygen -vacancy -induced elastic dipole anelasticity 76 above 250 oC.121 Since literature studies have shown that 10PCO exhibits insignificant oxygen nonstoichiometry below 400 oC, 101, 102, 112 -115 and 114 =+(3.6) the 280 -400oC chemical strain ( ) was assumed to be zero such that the 280 -400oC repr esented only the thermal strain ( ). A similar argument was made for the 280 -400oC chemical, thermal, and total stress. As a rough approximation, the 280 -400oC was assumed to vary linearly 24 with temperature, in keeping with previous reports of the n ear -linear thermal expansion of 10PCO, 112, 122 doped ceria, 123, 124 8YSZ 123 and MgO 125 over the 280 -700oC temperature range. was then extrapolated to temperatures > 400 oC using this fit so that could be extracted from via Eqn. 6. A similar treatment was given to the 400 -700oC thermal and chemical stress data. With knowledge of , the thin film 10PCO oxygen nonstoichiometry ( ) was determined using the relationship: 81, 114, 126 ==(3.7) using a (100) oriented 10PCO chemical expansion coef ficient ( ) of 0.067 76 and a =0 below 400oC (as has been assumed in other studies 114, 115 ). Use of a constant was warranted over the 280-700oC range because of the dilute nature of the oxygen vacancies encountered here ( <0.015) and the fact that past ceria experimental 114, 127 -131 and modeling 76, 81, 132 studies have shown that the lattice strain per oxygen vacancy (i.e. ) is constant for <0.03 and temperatures up to 1000oC (Note, studies on 10 PCO have shown that remains constant to at least =0.055).112, 114 A value of 0.067 was chosen because that is the DFT -predicted value for (100) oriented ceria 76 and is consistent with the 650 -800oC thin film 10PCO =0.064±0.005 measured previously 131 (Note, although convenient, a constant is not required to reliably extract materials properties from wafer curvature measurements). 3.3 Experimental Methods 3.3.1 Sample Pr eparation One -side polished, (100) oriented, circular, 200 thick, 25 mm diameter (Y2O3)0.095 (ZrO 2)0.905 (YSZ) and magnesium oxide (MgO) single crystal (Crystec GmbH, Berlin, Germany) were used as PLD substrates. Prior to deposition, all the substrates were annealed at 1450oC for 20 hours with a 5oC/min nominal heating and cooling rate to relieve any residual 25 internal stress . Afterwards, t he 25-700oC curvature change s of only the substrates were measured to ensure that any substrate residual stresses capable of producing unwanted curvature changes during later film stress -temperature measurements were adequately removed. Only substrates exhibiting 25 -700oC curvature changes less than 0.005 m -1 were used for subsequent Pulsed Laser PLD depositi on. Targets for PLD deposition were produced by pellet pressing and sintering Pr0.1 Ce0.9 O2- powders. These powders were produced using the g lycine nitrate combustion method 133 using 18.2 M water (Millipore, Burlington, MA) , Pyrex glassware (Sigma Aldrich, St. Louis, MO), Teflon coated stir bars (Fischer Scientific, Pittsburgh, PA), a stainless steel reaction vessel (Polar Ware, Kiel, WI), 99.9% pure praseodymium nitrate (Strem Chemicals, Newburyport, MA), 99.9% pure cerium nitrate (Strem Chemicals, Newburyport, MA), and 99% pure glycine (Sigma Aldrich, St. Louis, M A) with a 1:1 glycine to nitrate ratio . After synthesis, t he powder was calcined in an 99.8% pure alumina crucible (CoorsTek, Golden, CO) at 1000 oC in air using a 5oC/min nominal heating and cooling rate. Then , the powder was transferred to a 38 mm diamete r stainless steel die (MTI, Richmond, CA) and uniaxially compacted to ~63 MPa of pressure. The pressed target pellet was then sintered at 1450 oC for 20 hrs with 3 oC/min nominal heating and a 10oC/min nominal cooling rate to produce a 25 mm diameter PLD target. PLD was conducted with a XeF laser emitting at 353 nm . The chamber was first pumped down to 10 -6 torr and then heated to a substrate temperature of ~5 80oC. After the substrate temperature was stable, the chamber was backfilled with oxygen until the pressure reache d 9*10-3 torr. 10PCO was then simultaneously deposited onto the previously described YSZ and MgO substrates for 20 minutes using a 350 mJ laser power, a 10 Hz pulse frequency , a ~50 rpm sample rotation, and a substrate to target distance of ~6 cm . 26 After removal from the PLD , the samples were re -equilibrated in air under protective 99.9% alumina crucibles at 1000 oC for 1 hour with 3 oC/min nominal heating and cooling rate s. Given the known effect of surface impurities on the oxygen exchange properties of 10PCO, 134 the 10PCO|YSZ samples used for later oxygen surface exchange measurements were surface etched using the procedures described in Zhao et al. 134 Specifically, the samples were placed in 65oC 50% NaOH -50% H 2O solution for 24 hours with 100 rpm stirring speed . As shown Figure 3.1 AFM images (a) before and (b) after the etching proces 27 in Figure 3.1, this procedure was capabl e of removing Si surface impurities without significantly altering the surface roughness. 3.3.2 Film Microstructure and Crystallographic Orientation Characterization X-Ray diffraction (XRD) was conducted using a Rigaku SmartLab diffractometer with a 44 kV voltage and a 40 mA current. Scans were carried out between 20 and 80 o with a 0.01 o/min scan rate and a 1 second dwell time. Scanning electron microscopy (SEM) imag ing was conducted on fractured sample cross -sections coated with ~5 nm of Pt using a TESCAN MIRA3 Field Emission SEM (TESCAN Inc.) using a 20 kV beam voltage. 3.3.3 Dual Substrate Measurements For Dual Substrate measurements, the curvature of both 10PCO|YSZ and 10PCO|MgO sample s were measure d from 280 to 700 oC with 5oC/min heating and 0.2 oC/min cooling rate s in 25 sccm synthetic air . Analysis temperatures 280oC and synthetic air were chosen to avoid stress changes caused by water adsorption 135 and to avoid complications introduced by potentially orientable elastic dipoles present in ceria present below 250 oC.121 As shown in Figure 3.2, two Figure 3.2 Reproducibility of the 0.2 oC/min stress -temperature data on cooling for two representative a) PCO|MgO and b) PCO|YSZ samples. 28 ther mal cycles were conducted in synthetic air to ensure reproducibility , and the stress -temperature results were averaged together to produce the values in Figure 3. 6. The 10PCO|MgO samples went directly from the PLD chamber to the 1000 oC re -oxidation furnace to the XRD and then directly to the synthetic -air -flushed Multibeam Optical Stress Sensor (MOSS) test rig in an attempt to minimize hydration of the MgO substrate. The simultaneously produced 10PCO|YSZ samples were stored in a CaCl 2 containing desiccator for ~2 weeks while the 10PCO|MgO samples were being MOSS tested, before being analyzed. 3.3.4 Chemical Strain Determination The =0.21 chemical strain data was determined by first fitting the 10PCO|YSZ stress -temperature curve with a 3 rd order polynomial e quation over its entire 280 -700oC range. The 3rd order polynomial fits used here don not have any physical meaning, it was chosen because it can represent the stress vs. temperature relationship on 10PCO|YSZ without picking up the little disturbances in th e signal. The low temperature (280 -400oC) 10PCO|YSZ stress vs. temperature curve was then fitted with a linear equation , assuming a near constant thermal expansion coefficient for 10PCO and the YSZ . The difference between the linear extrapolation and 3 rd order polynomial at high temperature was used to measure the amount of temperature -induced chemical stress. The chemical strain was then calculated using the fitted chemical stress and the temperature dependent biaxial moduli obtained from the Dual Substrate method of Section 3.2.2. 10PCO|YSZ was chosen instead of 10PCO|MgO for this purpose because of the lower sample noise shown in Figure 3.6. Since Dual Substrate measurements were not taken in =0.021, the =0.021 chemical s train values of Figure 5a were determined by combining the measured chemical stresses generated by switching from a of 0.21 to 0.021 with the temperature dependent biaxial moduli 29 obtained from the Dual Substrate method of Section 3.2.2, and adding th e result to the =0.21 chemical strain values shown in Figure 3.6. 3.4 Results and Discussion 3.4.1 Crystallography and Morphology of the Film Figure 3.3 Representative X -Ray Diffraction (XRD) results for a) 10PCO|MgO and b) 10PCO|YSZ indexed using CeO 2, MgO, and YSZ JCPDS card numbers 81 -0792, 87-0653, and 70-4436, respectively. The asterisks denote impurity peaks not caused by the 10PCO, as proven in Figure 3.4b 30 Figure 3.3 shows representative X -ray Diffraction (XRD) scans of the oxygen re -equilibrated 10PCO films . These results indicate that the 10PCO films were phase pure and highly crystall ized on both (100) oriented MgO and (100) oriented YSZ substrates. Further, the 10 PCO films on both substrates had a similar, predominantly (100) preferred orientation. Specifically, the 10PCO on YSZ films displayed only the (100) orientation, while ~97% of the 10PCO on MgO grains were (100) oriented and ~3% of the 10PCO on MgO grains w ere (111) oriented (based on the ~100:9 intensity ratio of the 10PCO (200):(111) in Figure 3.3a and the 28.5:100 CeO 2 JCPDS PDF #34 -394 (200):(111) intensity ratio 136 for a randomly oriented polycrystal). This grain orientation behavior was identical to that reported in the literature for CeO 2-x on (100) MgO, 137 CeO 2-x on (100) YSZ, 137 -140 and 10PCO on (100) YSZ. 18 , 141 A Scherrer Equation 142 analysis indicated that the average 10PCO grain size on the MgO and YSZ substrates was ~28 nm and ~21 nm, respectively. (Note, the limited number of XRD peaks resulting from the 10PCO preferred orientation prevented a more accurate Williamson -Hall 143 grain size determination). Figure 3.4 Representative a) survey and b) detailed X -Ray Diffraction (XRD) scans of a bare (100) oriented MgO wafer showing that the 34.2 o and 35.2o peaks shown in Figure 3.3 (denoted by asterisks) were not from the 10PCO film. 31 As indicated by the cross -section scanning electron microscopy (SEM) images of Figure 3.5, the 10PCO films on MgO and YSZ were dense and 235 ± 2 nm and 230 ± 5 nm in thickness, respectively. Post analysis SEM and XRD scans (not shown) did not reveal any changes in the crystallographic or microstructural character of the 10PCO films caused by the 25 -700oC thermal cycling and elevated temperature holds encountered during wafer curvature testing. 3.4.2 Stress vs. Temperatu re Data Figure 3.6 shows representative stress -temperature curves for the 10PCO|MgO and 10PCO|YSZ samples taken with a 0.2 oC/min cooling rate. The red lines are 3 rd order polynomial fits to the measured data. For 10PCO|MgO, =(486.0±1.638)+(2.389±0.001059 )×+0.007670±2.201×10)×+(7.020×10±1.479×10)× and for 10PCO|YSZ, =(1241.7±4.065)+(3.788±0.02622)×+(0.008090 ±5.446×10)×+(7.562×10±3.654×10)×. The 3rd order polynomial fits used here do n not have any physical mean ing , they were chosen because they can represent the stress vs. temperature relationship on both samples without picking up the little disturbances in the Figure 3.5 SEM Images of (a) 10PCO|MgO and (b) 10PCO|YSZ 32 signal. As demonstrated in Figure 3.7, heating rates faster than 0.2 oC/min (0.5 oC/min, 1 oC/min) gave different stress vs. temperature response s due to the thermal equilibrium issue. However, the stress vs. temperature curve collected with 0.1 oC/min is near identical as the data collected with 0.2oC/min. This indicates that 0.2oC/min was slow e nough to ensure that the samples remained in thermal equilibrium as the stress -temperature data was collected . The initial increase in film stress with increasing temperature displayed by the Figure 3.6a 10PCO|MgO sample is consistent with the fact that fr om 280 -500oC the 10PCO value (which ranges from 8 to 14 ppm/K) 114 is less than the 280 -500oC MgO (which ranges from 13 t o 14 ppm/K). 110 Similarly, the subsequent decrease in film stress with increasing temperature above ~500 oC is consist ent with the fact that the 500 -700oC 10PCO (which ranges from ~14 to 24 ppm/K) 114 is greater than the 500 -700oC MgO (whi ch ranges from 14 to 15), 110 due to the onset of chemical expansion in 10PCO. The constant decrease in film stress wi th temperature for the 10PCO|YSZ sample of Figure 3.6b is consistent with the fact that the 280 -700oC 10PCO (which ranges from 8 to 24 ppm/K) 114 is always larger than the 280 -700oC YSZ (which ranges from 9 to 10 ppm/K). 111 33 Figure 3.6 Representative averaged stress vs. temperature plots for a) 10PCO|MgO and b) 10PCO|YSZ. The red lines are fits to the measured data. 34 It is interesting to note that these thermal -expansion -mismatch induced stresses were in addition to tensile ~580 oC 10PCO growth stresses of ~300 and ~250 MPa on MgO and YSZ, respectively (even larger 10PCO growth stresses have been observed in the literature 131 ). Figure 3.7 The effect of cooling rate on the stress vs. temperature measurements of representative 10PCO|YSZ samples. Note, samples tested with either 0.1 or 0.1 oC/min cooling rates were considered to be in thermal equilibrium from 280 -700oC since the additional tim time at each temperature with a slower 0.1 oC/min cooling rate had no effect on the stress -temperature trajectory. 35 3.4.3 Young™s Modulus Measurements from Dual Substrate Measurements Figure 3 .8 shows the temperature -dependent Young™s Modulus valu es measured here in comparison to all the 10PCO E measurements presently available in the literature. The essentially 170-180 GPa constancy of 280 -700oC GPa Ce 0.9 Gd0.1 O1.95 - (CGO) E values is likely related to the small magnitude of the 10PCO encounter ed here (and discussed later). The slight dip in the Figure 3 .8 E values is likely an artifact of the 10PCO|MgO ~400 -500oC fitting error shown in Figure 3.8 PCO Young™s moduli values measured by the Dual Substrate method compared to the literature measurements. (DS stands for Dual Substrate, NI stands for Nano -Indentation, MOSS stands for Multi -beam Optical Stress Sensor, and XRD stands for X -Ray Diffraction) 36 Figure 3.6. The ± ~10% error bars shown in Figure 3 .8, calculated using the procedures described in Appendix B, are similar to those reported in other Dual Substrate literature studies. 109 The E values obtained here agree well with the 750 oC MOSS stress/XRD stain determined (100) 10PCO E value from Sheth et al .117 However, they do not agree as well with the 600 oC nano -indentation determined E value from Swallow et al .116 This may result from the inherent difficulty in performing reliable high temperature nano -indentation ex periments or the fact that unlike all the other studies in Figure 3 which were performed on 10PCO, Swallow et al .116 examined 20PCO. Room temperature extrapolations of the E values obtained here agree with the fast (but not the slow) 25 oC 10PCO nano -indentation measurements in the literature. 116, 144 This is consistent with the idea that nano -indentation -determined E values taken too quickly to be impacted by reorientation of the anelasti city -inducing oxygen -vacancy -generated elastic dipoles 76 present in ceria below 250 oC121 should be simi lar to the E values extrapolated from high temperature ceria samples not containing orientable, oxygen -vacancy -generated elastic dipoles. 3.4.4 Thermo -Chemical Expansion Coefficient from Dual Substrate Measurements Figure 3.9 shows the temperature -dependent values measured here in comparison to all the 10PCO measurements presently available in the literature. Similar to the Figure 3.8 results which do not display a systematic difference between the E values obtai ned from bulk/micro -grained samples compared to those from thin film and/or nano -grained samples, the thin film values obtained here agreed very well with previous in situ XRD literature measurements on bulk, micro -sized grain samples. 114 The ± ~4% error bars shown in Figure 3.9, calculated using the procedures described in Appendix B, are similar to those reported for in other Dual Substrate literature studies. 109 37 3.4.5 Chemical Strain & Oxygen Nonstoichiometry Figure 3.10a and 3.10b show the measured 10PCO chemical strain, and the oxygen nonstoichiometry extracted from it, respectively, compared to literature values . Interestingly, the measured thin films of Figure 3.10a experience less in -plane chemical strain than bulk 10PCO, but experience a similar to bulk PCO. This is caused by the lower values of ~0.07 for (100) oriented, thin film 10PCO 122, 131 compared to ~0.09 for bulk 10PCO. 114 The good agreement Figure 3.9 10PCO thermo -chemical expansion coefficients from the Dual Substrate method compared to the literature measurements of Bishop et al. (DS stands for Dual Substrate, XRD stands for X -Ray Diffraction, and Dil stands for Dilatometry). 38 Figure 3.10 10PCO (a) chemical strain and (b) oxygen nonstoichiometry values from the 10PCO|YSZ sample compared to the literature measurements 39 between the Figure 3.10b =0.21 thin film values obtained here and the bulk 10PCO data of Bishop et al. 114 may be caused in part, by the relatively low (i.e 300 to -50 MPa) 600 -700oC 10PCO film stress stresses encountered here. The values obtained under a of 0.21 and 0.021 are also both in good agreement with other thin film 10PCO studies that did, 115 and did not, 131 utilize precious metal current collectors to determine . The unphysical (i.e. slightly negative) ~375 -500oC chemical strain and oxygen nonstoichiometry values in Figure 3.10 result from spurious differences between the linear and third -order polynomial fits to the data caused by signa l drift (i.e. waviness) in the Figure 3.6b data below ~500 oC. 3.5 Summary This chapter demonstrates for the first time that wafer curvature measurements can be used to directly measure a variety of disparate and technologically -relevant thin film physicochemi cal properties (i.e. the oxygen nonstoichiometry, biaxial modulus, thermo -chemical expansion coefficient, and thermal expansion coefficient) under well -characterized film stress states , strain states, temperatures and atmospheric conditions. With application of an externally -derived Poisson™s ratio, the Young™s modulus was also determined (Note, studies have shown that wafer curvature experiments can also be used to directly measure the Poisson™s ratio 118 ). 40 4. Curvature Relaxation Measurements 4.1 Introduction Mixed ionic electronic conductors (MIECs) have been used in various applications including sol id oxide fuel cell s,3, 145 -148 gas separation membrane s,149 oxide memristors, 150, 151 electrostrictive actuators, 152, 153 chemical sensors, 154 catalytic converters, 155 and electrochromic windows. 156 The functionality and performance of MIECs are based on the oxygen transport through/within the material. The o xygen surface exchange coefficient (k chem ) is one of the kinetic constants that determines the speed of oxygen transport process. Being able to accurately measure the k chem enables performance evaluation and material selection when designing a device involving a MIEC. H owever, even for a conventional MIEC like lanthanum strontium ferrite, there is still an ~5 orders of magnitude difference in k chem .19 The u sage of current collector s during in -situ k chem measurements may be one of the reasons for this variation. 23, 157 Therefore, an in-situ , contact -free and current collector -free technique is needed to measure k chem without the influence of a sur face -altering noble metal. This chapter demonstrates that with known oxygen nonstoichiometry ( ), the chemical oxygen surface exchange coefficient ( chem ) and oxygen surface exchange resistance ( ) of Pr0.1 Ce0.9 O1.95 (10PCO) can all be obtained as a function temperature ( ), and oxygen partial pressure ( ) using in situ , non-contact, current -collector -free wafer curvature measurements. 4.2 Theory 4.2.1 Curvature Relaxation Measurements According to Eqn 3.5 (Stoney™s equat ion), the biaxial stress of a dense thin film on a dense thick substrate/wafer can be calculated from the curvature of the sample without knowing the biaxial modulus of the film. For a mechano -chemical active material, after an abrupt change in 41 oxygen part ial pressure, the oxygen nonstoichiometry in the thin film changes accordingly, which can be described by the solution of Fick™s Second Law 158 : =1exp (4.1) where is the instantaneous oxygen nonstoichiometry, is the initial oxygen nonstoichiometry, is the new -pO2-equilibrated oxygen nonstoichiometry , k is the chemical oxygen surface exchange coefficient, and t is time. The change in oxygen nonstoichiometry induces chemical strain, which can be described by equation 159 : ==(4.2) where is the chemical strain, l is the sample length, is the chemical expansion coefficient. The change in chemical strain induces a change in chemical stress according to Hooke™s Law: =(4.3) Combining Eq n 4.2 and 4.3, the film stress changes can be expressed as: =(4.4) Then combining Equation 3.5 and 4.4, it yields the following r elationship: 6=(4.5) where is the curvature of the sample. Since , , and are all constant, the change in curvature is proportional to the change in oxygen nonstoichiometry. Therefore, Equation 4.2 can also be expressed as 19, 106, 160 : ==1exp (4.6) 42 The oxygen surface exchange coefficient of thin film with thickness 1000 times smaller than the characteristic thickness (=/) can then be obtained by monitoring the curvature change of a bilayer sample. 4.2.2 Surface Polarization Resistance The surface polarization resistance ( ) of 10PCO can be expressed as: =4(4.7) where R is the ideal gas constant, T is the absolute temperature, F is Faraday™s constant, is the lattice oxygen concentration, and is the electrical oxygen surface exchange coefficient. was calculated from reported in the literature 161 with the following relationship: =4(2)(4.8) where is Avogadro™s Number a nd is the 10PCO lattice constant reported in the literature at different temperatures. 112 is calculated from the chemical oxygen surface exchange coefficient (kchem ) using a simplified version of Equation 6.102 from ref 162 : =(4.9) where is the transference number of the electrons, is the thermodynamic factor with the definition (detailed derivation of thermodynamic factor can be found in Ref 163): 163 =12ln()ln(2)(4.10) The oxygen nonstoichiometry data used for the R s calculation are from chapter 3 .161 43 4.3 Experimental Methods 4.3.1 Sample Preparation Details about sample preparation can be found in Section 3.3.1. In this work, only 10PCO|YSZ w as measured by the curvature relaxation technique. 4.3.2 Crystallography and Morphology Characterization The crystallography and morphology characterization of 10PCO thin film was conducted with XRD and SEM. Details about the paramete rs of those measurements can be found in Section 3.3.2. 4.3.3 Curvature Relaxation Measurements Prior to high temperature curvature relaxation measurements, the 10PCO|YSZ samples were heated up to 500 oC in synthetic air with a 5 oC/min heating rate. Curvature re laxation measurements were then conducted on 10PCO|YSZ samples from 500 to 600 oC with 25 oC increments, following the procedures described in detailed previously. 19, 160, 164, 165 Wafer curvature relaxations were triggered by switching between 100 sccm of sy nthetic air (20% O 2-80% Ar) and 100 sccm of 10% synthetic air -90% Ar (i.e. 10 times diluted air). To minimize possible Si contamination from the fused quartz curvature relaxation test rig (which is a time -dependent process), only 1 reduction and 1 oxidatio n cycle were measured at each temperature between 500 -575oC. Multiple oxidation/reduction cycles were then tested at 600 oC, as shown in Figure 4. 1, to determine if the curvature measurements were reproducible with redox cycling. 44 4.4 Results and Discussion 4.4.1 Crystallography and Morphology of the Film Details about the film orientation and film thickness can be found in Section 3.4.1. The 10PCO t hin film has a (100) preferred orientation on YSZ substrate and has a uniformed thickness of ~230 nm. 4.4.2 Relaxation Data and Curve Fitting Figure 4.1 shows representative stress redox cycle data for a 10PCO film kchem tested at 600oC. While a steady -state equilibrium was obtained after each cycle (allowing a reliable k Figure 4.1 Representative raw curvature relaxation data for a 10PCO|YSZ sample at 600 oC 45 determination), the equilibrium film stress after each oxidation and reduction cycle was slightly altered (shown in the red dotted guideline) . This behavior is similar to an alteration in the Figure 4.2 a) Normalized curvature fits to the 10PCO|YSZ data, and b) ln(1 -Normalized Curvature) plots for the 10PCO|YSZ samples. Note that the generally good single time constant (red line) fits to the data of part a) and a single slope in the ln(1 -Normalized Curvature) plots in part b) at times before the equilibrium state is reached suggest only one mechano -chemically active process is active with redox cycling between synthetic air and 10% diluted synthetic air. 46 10PCO|YSZ equilibrium film stress with redox cycling reported in previous 10PCO measurements 131, 134 and attributed to stress -relaxing alterat ions of the grain boundary structure. 131 Figure 4.2 shows representative curve fitting and the alternatively plotted relaxation data . The fact that only one physical process is observed in Figure 4.2b suggests that whatever the mechanism, it is purely the result of, and occurs on the same timescale, as oxygen exchange into/out of the film . Additionally, Figure 4.3 shows the relaxation data from all tested temperature , which shows that the 10PCO thin film is mechano -chemical active and the curvature relaxations have good signal to noise ratio from 500 t o 600oC. Figure 4.3 Raw wafer curvature relaxation data for the 10PCO|YSZ samples from all tested temperatures (500, 525, 550, 575, and 600 oC) 47 4.4.3 Oxygen Surface Exchange Coefficient of 10PCO Thin Film Figure 4.4 shows the oxygen surface exchange coefficient values, in comparison with other literature measurements. As seen in many other studies on various oxyg en exchange materials, 160, 166, 167 the oxidation kinetics were faster than the reduction kinetics. As postulated in other studies, 166, 167 this is likely due to the larger at the beginning of the oxidation process than at the Figure 4.4 10PCO chemical oxygen surface exchange coefficients from the curvature Relaxation. MB stands for Micro -Balance, EIS stands for Electric Impedance Spectroscopy and OR stands for Optical Relaxation). The curvature -determined k error is less than the size of the symbol 48 beginning of the reduction process , which speed s up the initial oxygen exchange. The measured k values displayed Arrhenius behavior over the entire 500 -600oC range, which is consistent with only one oxygen exchange process being active. Tests on select samples performed with smaller step sizes u sing air and 5 times (instead of 10 times) dilut ed air gave k values with the same 0.9 ± 0.1 eV activation energy as those in Figure 4.4, but with absolute values between the reduction and oxidation data of Figure 4.4, suggesting that the oxygen surface exchange kinetics remained linear at a 10 times di lution, as observed previously in the literature for other materials. 167 The 600 oC k values measured here agreed exactly with those obtained from the optical relaxation studies of Zhao et al. 134 even though the samples were subjected to different thermal histories before testing (Zhao et al. 134 do not report re -equilibrating their samples in 1000 oC air before testing, as was done here) and the specimens were grown in different Pulsed Laser Deposition (PLD) chambers that likely contain different impurities. The likely source of this agreement is that, as shown in Table 4.1, the samples used for both studies were produced in the same manner, had similar final microstructures, had similar final stress states, did not u se precious metal current collectors, and were both surface etched before testing. In contrast, the measured k values were much lower t han those determined from microbalance and EIS studies. Although Simons et al .166 attributed their observed k enhancement to grain boundary effects in their randomly -oriented polycrystalline films, 166 this seems unlikely Table 4.1 Comparison of PCO|YSZ samples between this work and Zhao et al. Samples Deposition Technique Orientation Current Collectors Deposition Temperature ( oC) PCO|YSZ from This Work Pulsed Laser Deposition (200) No 680 PCO|YSZ from Zhao et al. Pulsed Laser Deposition (200) No 725 49 to be abl e to explain the ~1000 times difference with the present results because other literature studies have only observed k differences within an order of magnitude for ceria 168 or lanthanum strontium cobalt iron oxide 169 when examining thin films with intentio nally -varied crystallographic orientations. Instead, based on the documented ability of precious metals such as Pt, 170 -172 Ag, 170, 173 Au, 171 to catalyze the o xygen exchange reaction on a variety of MIEC materials (including PCO), it is more likely that catalytically active Pt migrated onto the surface of Simons et al. ™s 10PCO films from the underlying Pt current collector during subsequent 10PCO deposition and/ or testing. Catalytic enhancement provided by the porous Au current collectors covering Chen et al .™s 10PCO films may also explain why their EIS -measured k™s were higher than those measured here. Interestingly, the similar activation energies obtained for Chen et al. ™s study and the present work suggests that surface decorating Au particles do not change the 10PCO oxygen incorporation pathway, but instead simply increased the surface area available for incorporation via a spill -over mechanism. Although pr ecious metal electrode k-enhancement may be an exciting pathway to realize improved device performance, the 10PCO k results in Figure 7 contribute to the growing body of evidence 170, 171, 173 indicating that k measurements obtained from past studies perform ed with significant amounts of intentional or inadvertent precious metal surface coverage may be artificially enhanced. These include 1) electrochemical impedance spectroscopy, 18 mechanical impedance 174 or other experiments performed with porous, surface -coating precious metal current collectors, 2) electrical conductivity relaxation experiments performed with interdigitated precious metal electrodes, 175 -178 3) resonant microbalance experiments u sing supporting Pt layers, 166 etc. As such, the k values obtained from such studies would not be expected to accurately describe the performance of conventional MIEC micro -porous, MIEC micro -comp osite, or MIEC infiltrated 50 electrodes (since the MIEC surfaces in those electrodes are unlikely to be decorated with significant amounts of precious -metal catalyst). 4.4.4 Oxygen Surface Exchange Resistance Figure 4.5 shows the oxygen surface exchange resistance values measured here, compar ed with the literature. Consistent with the behavior observed in Figure 4.4, the wafer curvature measured 10PCO RS values were significantly higher than those obtained from the precious -metal -contaminated EIS experiments of Chen et al .18 The slightly higher wafer -curvature -determined RS Figure 4.5 10PCO oxygen surface polarization resistance (R S) values obtained here compared to the literature values for 10PCO, LSCF, LSF, LSC, and SSC 51 activation energy, compared to that from Chen et al., 18 comes from t he slightly different oxygen nonstoichiometries shown in Figure 3.10. That being said, the 1.5 ±0.1 eV 10PCO wafer -curvature -determined RS activation energy measured here is similar to the activation energy of other many MIEC oxygen exchange materials. 179, 180 Figure 4.5 also shows that the wafer -curvature measured 10PCO RS values were two orders of magnitude higher than the RS values of the SOFC material La 0.6 Sr0.4 Co0.8 Fe0.2 O3- (LSCF). Unlike EIS measurements performed on large surface area samples requiring current collectors to distribute the charge, the small lateral dimensions of the microelectrode samples used to measure the La 0.6 Sr0.4 FeO 3- (LSF), 179 La0.6 Sr0.4 CoO 3- (LSC), 179 LSCF, 179, 180 Sr0.5 Sm0.5 CoO 3- (SSC), 179 and Ba 0.5 Sr0.5 Co0.8 Fe0.2 O3- (BSCF) 179 RS values reported in Figure 8 ensured that these EIS microelectrode RS values could be measured without the presence of precious metal current collectors. The trustworthiness of these EI S microelectrode RS values is also indicated by the fact that the open -circuit performance of LSF, LSFC, LSCF, LSC, and SSC infiltrated SOFC cathodes can be successfully modeled using them. 181 -184 Taken together, the data in Figure 8 suggests that low -stress, precious -metal -free, (100) -oriented 10PCO is a poor choice as a SOFC cathode material. However, with future strain engineering, grain boundary engineering, precious metal surface decoration, etc. PCO may become SOFC cathode material. 4.5 Summary This chapter demonstrates that an in-situ , contact -free, current -collector -free technique can be used to evaluate the oxygen surface exchange kinetics of 10PCO thin film. With the knowledge of oxygen nonstoichiometry from chapter 3, the oxygen surface exchange resistance can also be calculated. The difference between EIS and curvature relaxation measurements suggests that previous measurements u sing precious -metal current collectors, support layers, etc. may report 52 spuriously large oxygen surface exchange coefficient values and spuriously low oxyg en surface exchange resistances. The comparison of oxygen surface exchange resistance between 10PCO and other commonly used MIECs suggests that 10PCO may not be an effective material for solid oxide fuel cell applications. 53 5. HTXRD (High Temperature X -Ray Diffraction) - MOSS Combined Test 5.1 Introduction Ceria and doped ceria have been widely used in various applications including solid oxide fuel cells (SOFCs), 146, 185, 186 solid oxide electrolysis cells (SOEC s),69 catalytic converters, 155 chemical sensors, 187, 188 and electrochromic devices. 189 There are extensive studies exploring the thermal, 114, 190 -194 mechanical, 144, 193 -196 electrical 197 -200 and chemical properties 18, 114, 201 -204 of these materials in order to characterize, and improve the performance and mechanical stability of su ch devices. However, at high temperature, due to the limitation of measurement platform s and the geometry of thin film samples, there is a lack of data for Young™s modulus of thin film ceria and doped ceria at high temperature. 205, 206 Chapter 3 showed that the Young™s modulus and the thermo -che mical expansion coefficient can be obtained by combining two sets of stress vs. temperature data of two identical thin films on top of two different substrates. 194 The results showed that the Young™s modulus stays constants at high temperature. However, this method requires two near -identical film to have a similar orientation on different substrates, which is not always applicable for every material. Additionally, the error ge nerated from both sets of data significantly boost the error bar of the Young™s modulus, which is ~15% of the actual value. Therefore, using another approach to measure high -temperature Young™s modulus and the thermo -chemical expansion coefficient with one thin film sample can be beneficial for general measurements on other thin film materials in the future . The Young™s modulus and the thermo -chemical expansion coefficient from an alternative measurement method can also provide a validation to the dual subs trate work. In this chapter , a Pr0.1 Ce0.9 O2- (10PCO) | 9.5% yttria stabilized zirconia (YSZ) bilayer sample was measured using high temperature X -Ray diffraction (HTXRD) and multi -beam optical 54 stress sensor (MOSS). It shows that by combing the strain measurement from HTXRD and the stress measurement from MOSS, the thermo -chemical expansion coefficient and Young™s modulus of the ( 100) -oriented 10PCO thin film can be extracted at different temperatures between 500~700 oC. 5.2 Theory By monitoring the position change of the 10PCO (200) peak, according to Bragg™s law, the lattice constant of 10PCO thin film at different temperature s can b e extracted. Then, the out -of-plane strain can then be obtained through the following equation: =(5.1) where is the instantaneous lattice constant of 10PCO, and is the lattice constant at room temperature. The out -of-plane strain can also be expressed as the sum of thermal strain and the strain from Poisson™s expansion/contraction (i.e. the contribution of strain along the in -plane directions), which can be expressed as: =21()(5.2) where is the thermal ex pansion coefficient of the substrate and is the Poisson™s ratio of 10PCO. Differentiating Eqn 6 with respect to temperature leads to: =21()(5.3) The vs. temperature data was fitted with the following equation: =(5.4) 55 where A and B are all fitting constants. The exponential part ( ) represents the chemical expansion of the 10PCO thin film at high temperature, which corresponds to the following relationship: exp (5.5) where is the change in chemical strain, is the change in oxygen nonstoichiometry, and is the formation energy of oxygen vacancy in 10PCO. By differentiating the fitting equation with respect to temperature, can be obtained. The thermal expansion coefficient of YSZ substrate ( ) was also measured via HTXRD. With known v (v=0.33 for 10PCO), 205 the thermo -chemical expansion coefficient of 10PCO can be extracted. Detailed error analysis for this methodology is described in Appendix C of the supplement ary material. 5.3 Experimental Details 5.3.1 Sample Preparation Detailed sample preparation for 10PCO|YSZ can be found in Section 3.3.1. 5.3.2 Microstructural and Crystallographic Characterization The cross -sectional images were taken via a TESCAN MIRA3 Field Emission Scanning Elect ron Microscope (SEM) (TESCAN Inc.) with a 20 kV beam voltage on the sample with ~5 nm Pt coating. XRD measurements were conducted on a Rigaku SmartLab Diffractometer with a 44 kV voltage and a 40 mA current. A survey scan was conducted between 20 and 80 o with a 0.01° step size and a 1 second dwell time. 5.3.3 High Temperature X -Ray Diffraction Measurements 10PCO|YSZ was first heated up to 700 °C, then cooled down with a 1 °C/min cooling rate, during which the positions of the (200) 10PCO peak were measured with 25°C increments in air. 56 The range of the scans was 31~34 o, the increment was 0.01 o and the dwelling time was 1 second. The sample dwell time was set for 10, 20, 30, 40, 40, 50, 60, 60 min before each measurement at 675, 650, 625, 600, 575, 550, 525, 500 oC, respectively. For each temperature, 3 measurements were taken to check for thermal/chemical equilibrium (25 and 700 oC excluded). 5.3.4 Stress Measurements Detailed information of stress measurements via multi -beam optical stress sensor (MOSS) can be found in Se ction 3.3.3 for 10PCO|YSZ sample. 5.4 Results and Discussion 5.4.1 Microstructural and Crystallographic Characterzation The XRD survey scan of the 10PCO|YSZ sample in Figure 5.1 shows that the 10PCO film has a preferred (100) orientation on the (100) -oriented YSZ s ubstrate. The orientation of the 10PCO thin film is consistent with other studies of 10PCO thin films on (100) -oriented YSZ substrates. 18, 23, 57, 204 -207 The full width half max (FWHM) of the (200) 10PCO peak is 0.682° and the FWHM of the (200) YSZ is 0.079°. Since the single crystal YSZ substrate is strain free and has a grain size of 2.54 cm, the FWHM of the (200) YSZ is considered as instrumental broad ening . According to the equation: =+(5.6) where is the measured peak broadening, is the broadening from the instrument, and is the peak broadening from the sample. 208 Applying the and values mentioned above to Eqn 5.6, a value of 0.677° was obtained. Based on Sherrer™s equation 209 :an estimated crystallite size of ~21 nm was obtained on the 10PCO|YSZ sample. Although a more accurate Williamson -Hall method would be more suitable for the grain size estimation, 210 the limited number of peak s resulted from the preferred orientation , which hinders the implementation of this approach. 57 Therefore, the grain size reported here is the lower limit without considering the strain broadening. Additionally, the film thickness of the 10PCO|YSZ sample is ~230 nm according to the SEM image shown in Figure 3.5. Comparable grain size was obtained for 10PCO (400) peak with a ~39 nm grain size. Figure 5.1 X-Ray diffraction survey scan of 10PCO|YSZ at 25 oC, the asterisk denotes the (400) YSZ diffraction peak from tungsten radiation 58 As shown in Figure 5.2, the pole figure of 10PCO|YSZ sample indicates the presence of low -angle grain boundaries within th e 10PCO thin film. Therefore, the brick -layer -model, as reported by Sheth et al ,117 was not used here for the analysis of thermo -chemical expansion coefficient and Young™s modulus because the grain interior and the low -angle grain boundary of doped ceria have similar structure and near -identical oxygen vacancy formation energy. 211 Therefore, the brick -layer -model, which typically requires significantly different grain boundary and grain inter ior properties (2.5~5 times difference in expansion coefficient) ,212 are not used in this situation. Figure 5.2 Pole figure and zoomed -in image of (111) peak of preferred -oriented 10PCO thin film at room temperature 59 5.4.2 HTXRD Characterization Figure 5.3 shows the representative XRD scans of the ( 200) 10PCO peak at different temperatures during cooling of the sample. As the temperature increas es, the (200) peak shifts to a lower 2 -theta value, which indicates an expansion of the lattice in PCO thin film . Figure 5.3 Raw data summary of HTXRD measurements from 25 to 700 oC 60 Furthermore, as shown in Figure 5.4, the full width half max of the peak at each temperature didn™t change, suggesting no grain growth or grain growth related strain during the measurements. Therefore, the peak shift in Figure 5.3 is mainly caused by the thermo -chemical expansion of the lat tice. Figure 5.4 Full width half max of (200) 10PCO peak at different temperatures during HTXRD measurements 61 Figure 5.5 shows the XRD scans of (400) YSZ peak at different temperatures. As the temperature increas es, the ( 400) peak shifts to a lower 2 -theta value, which indicates an expansion of the YSZ lattice . Figure 5.5 XRD scans of (400) YSZ peak at different temperatures during HTXRD measurements 62 Figure 5.6 Peak fitting of HTXRD data using Pearson VII function 63 Figure 5.6 shows the peak fitting for 10PCO thin films. All the HTXRD peaks were fitted with the Pearson VII function to obtain the peak position for strain calculation. As shown in the plot, the Pearson VII function yields a good fit of HTXRD peaks at each temperature. 5.4.3 Out -of-Plane Strain of the 10PCO Thin Film Figure 5.7 shows the out -of-plane strain values calculated from HTXRD data. The strain values increase as the temperature rises from 500 to 700 oC, which agrees with the trend observed in other studies, 114, 213 indicating thermal and/or chemical expansion of the 10PCO film. The dashed line is the fitting line using Eqn 5.4. The data are well -represented with a 0.999 R 2 value. Figure 5.7 Out -of-plane thermo -chemical strain and its fitting of 10PCO|YSZ sample 64 Note that the out -of-plane strain value reported in Figure 5.7 is a combination of the grain interior expansion and Poisson™s expansion/contraction (in -plane strain contribution). Therefore, a direct comparison with strain values from dilatometry is not ap propriate in this situation. 5.4.4 Thermal Expansion Coefficient of YSZ Figure 5.8 shows the thermal expansion coefficient of YSZ based on the (400) peak position change of the YSZ substrate. The thermal expansion coefficient stays almost constant at ~ 10.3 ppm/ oC from 25 to 700 oC. Figure 5.8 Thermal expansion coefficient of YSZ substrate 65 5.4.5 Thermo -Chemical Expansion Coefficient via HTXRD Figure 6.9 shows the thermo -chemical expansion coefficient of 10PCO measured in this work, compared with other literature studies. 23, 114 The thin film value measured in this work is comparable with both microcrystalline and thin film data in the literature studies. 23, 114 The similarity between bulk and thin film sample is consistent with what Sheth et al. reported at pO2=0.21. 206 Additionally, for film with preferred -orientation , it is reported that there is no difference for oxygen vacancy formation energy for between bulk and thin film sample s due to the structural similarity between grain interior and low angle grain boundary. 211 Therefore, be cause of the high oxygen partial pressure and the low angle grain boundaries, the difference between the Figure 5.9 Thermo -chemical expansion coefficient of 10PCO measured from this work, compared with other literature studies 66 thermo -chemical expansion coefficient of bulk and thin film samples is either negligible or non -existing. The thermo -chemical values from the HTXRD mea surements is slightly lower than the values from dual substrate measurements, th is variation is expected since: (i) Unlike the dual substrate method, which requires two samples (namely 10PCO|YSZ and 10PCO|MgO) , the HTXRD measurements were conducted on one sample (10PCO|YSZ). Besides , the calculations to obtain thermo -chemical expansion coefficient are different for th ese two method s. Therefore, a small variation is expected between th ese two data sets. (ii) The increment of the HTXRD measurements is 25 oC while the increment of dual substrate method is 0.1 oC. Therefore, the fitting for the stress vs. temperature data used in the dual substrate method is more accurate than the out -of-plane strain fitting used in HTXRD measurements. As a result, the differenc e in curve fitting propagates and then generates the difference in thermo -chemical expansion coefficient. 5.4.6 Young™s Modulus via HTXRD -MOSS Figure 5.10 shows that by combining the stress data from MOSS with the strain data from HTXRD, the Young™s modulus of 10PCO thin film can be determined. The slope of the linear fit between stress change and strain change is the Biaxial modulus of the 10PCO thin film. Assuming a Poisson™s ratio of 0.33, the Young™s modulus can be calculated. 67 Figure 5.10 Stress change vs. strain change plot of 10PCO thin film 68 Figure 5.11 shows Young™s modulus of 10PCO measured in this work in comparison with other literature data .23, 144, 205 Young™s modulus is nearly constant with temperature at a value of ~165 GPa and it agrees well with the data reported by Sheth et al. at 750 oC.206 However, it does not agree with the nano -indentation data at either 600 oC or 25 oC. The difference at 600 oC might be caused by the difficulty in performing the nano -indentation at high temper ature or a different dopant level (20PCO was used by Swallow et al.). 205 The difference between slow -loading room temperature nano -indentation and this work might result from the reorientation of the elastic dipole caused by oxygen vacancy formation , where the dipoles resulted from the formation of oxygen Figure 5.11 Young™s modulus measured in this work, compared with other litera ture studies. 69 vacancies were re-oriented towards a preferred direction due to the effect of an external electric field or force .144 The difference between fast -loading room -temperature nano -indentation and this work might be that the loading speed of a fast -loading nanoindentation is still not able to get rid of the dipole reorientat ion effect. 5.5 Summary This Chapter demonstrate s that the thermo -chemical expansion coefficient and Young™s modulus of a thin film can be extracted by combining the HTXRD strain measurements and MOSS stress measurements. It also shows that with the HTXRD -MOSS method, the magnitude of the error bar for thermo -chemical expansion coefficient and Young™s modulus were significantly reduced. The data from HTXRD -MOSS is comparable with other literature studies, especially with the work in Chapter 3 , which validate s the reliability of dual substrate method. 70 6. kchem Measurements Compar ed with Other Electrode -Free Techniques 6.1 Introduction As discussed in previous sections, MIECs have been widely used in various applications. The fun ctionality and performance of MIECs are based on the oxygen transport through/within the material. The o xygen surface exchange coefficient (k chem ) is one of the kinetic constants that determines the speed of the oxygen transport process. Being able to accu rately measure the k chem enables performance evaluation and material selection when designing a device involving MIEC. However, even for a conventional MIEC like lanthanum strontium ferrite, there is still an ~5 orders of magnitudes difference in k chem . Th e us e of a current collector during in -situ k chem measurements may be one of the reasons for this variation. 23, 157 Therefore, an in-situ , contact -free and current -collector -free technique is needed to measure k chem without the influence of surface -altering noble metal. In this work, the k chem of two near identical SrTi 0.65 Fe0.35 O3- (STF35) thin film s were deposited on yttria doped zirconia (YSZ) single crystal s and measured in-situ by curvature optical relaxation (OTR) to validate the consistency of various contact -free, current collector -free, in-situ kchem characterization techniques. 6.2 Theory For STF35, when there is no interaction between the oxygen vacancies (the dilute case) , there is a li near relationship between the change of Fe 4+ and the change of the optical absorption coefficient. 214, 215 The transmitted light intensity is related with the change of the optical absorption coefficient: =exp ()(7.1) 71 where is the change in transmitted light intensit y, is the incident light intensity, is the change in optical absorption coefficient, and L is the thickness of the sample. Similar to curvature relaxation, the relaxation curve of transmitted light intensity can be fitted by a solution of Fick™s s econd law: =lnlnlnln=1exp (7.2) where the subscripts have similar meanings as mentioned above. 6.3 Experimental Details 6.3.1 Sample Preparation The STF35 target was manufactured by a solid state diffusion process. Fe 2O3, TiO 2, and SrCO 3 were grounded together, pressed uniaxially and then isostatically at 300 MPa at room temperature . The pressed pellet was sintered at 1425 oC for 6 hours with 5 oC/min heating and cooling rates. The substrates for cu rvature relaxation measurements were annealed at 1450 oC for 20 hours to release the residual stress in the substrate. A ~200 nm thick STF35 thin film was deposited on a 200-micron thick , one side polished, (100) Œoriented, 1 inch diameter , 13% YSZ substrate (MTI Corporation, CA) via PLD for curvature relaxation measurements . A ~70 nm thick STF35 thin film was deposited on a two -sides polished , (100) -oriented, 10*10*0.5 mm, 13% YSZ substrate (Dalian Keri Optoelectronic Technology Co. Ltd., Dalia n, China) via PLD for optical relaxation measurements. The films were deposited at 700 oC with 5 Pa O 2 wit h 75~80 mJ laser power and 5 Hz pulse frequency. The target rotation speed was 25o/s and the substrate rotation speed was 10o/s. The target to substra te distance was 55 mm. After the deposition, the samples were cooled in 20 Pa O 2 72 with 10 oC/min cooling rate. The samples for both curvature relaxation and optical relaxation were aged at 850 oC at pO 2=0.21 to re -equilibrate their oxygen content. 6.3.2 Crystallographic Characterization The XRD measurements were made with a Rigaku SmartLab 9kW AMK instrument . The XRD signal was measured from 10 to 90 o with a scan rate of 1 o/min with 45 kV voltage and 40 mA current. 6.3.3 Curvature Relaxation Measurements The relaxation of curvature was triggered by an abrupt change in oxygen partial pressure, between compressed air (pO 2=0.21) and 10% air diluted in N 2 (pO 2=0.021). The k chem values were measured from 850 oC to 700 oC with 25 oC increments. 6.3.4 Optical Relaxation Measu rements The optical relaxation was performed by switching between synthetic air (20% O 2 - 80% N2) and 5 times diluted synthetic air (4% O 2 Œ 96% N 2). The k chem values were measured from 750 to 580 oC. 6.4 Results and Discussion 6.4.1 XRD Characterization Figure 6.1 shows the representative X-ray diffraction scan of one of two STF35|YSZ sample s, indicating a well -crystallized STF35 thin film. The peaks at ~32 and 68 o show that the film is (110) preferred -oriented. 73 6.4.2 Curvature and Optical Relaxation of STF35 Thin Films Figure 6.2 shows the raw data for the curvature relaxation (200 nm STF35 thin film) and optical relaxation measurements (70 nm STF35 thi n film) at 700 oC. After an abrupt oxygen partial pressure change, the curvature and the transmitted light intensity changed accordingly, indicating an oxygen transport process in the STF35 thin film. The difference of the equilibration times between the optical and curvature relaxation s is the result of different STF35 film thickness es. The Figure 6.1 Representative XRD scan for STF35|YSZ sample 74 oscillation in the curvature relaxation was caused by the vibration of the testing environment, and the oscillation in the optical relaxation was caused b y the thermal vibration at high temperature . Figure 6.2 Raw data of (a) curvature relaxation and (b) optical relaxation at 700 oC 75 Figure 6.3 Representative curve fitting of (a) curvature relaxation and (b) optical relaxation of reduction relaxation at 700 oC 76 Figure 6.3 shows the representative curve fit of the and the OTR using Eqn 4.6 and Eqn 7.2, respectively. Additionally, the two relaxation curves can both be fitted with one relaxation time constant, indicating there is only one relaxation process happening during the relaxation period. 6.4.3 Oxygen Surface Exchange Coefficient Comparison Figure 6.4 shows kchem measure d chem limitation (when the rate limiting step is the process of gas refilling the test ing chamber) appeared above 725 oC for OTR measurements and due to the slow curvature relaxation response at temperature s below 700 o OTR. Although the film thicknesses were different for (70 nm), the oxygen transport process was dominated by the oxygen surface exchange because the film thicknesses for both samples are well below the characteristic thickness of the STF35 (no actual characteristic thic kness data for STF35, 10 -1~10 -3 cm for conventional MIEC materials) .18, 31 Therefore, similar kchem value s should be expected for the the values at 70 0oC are comparable for both techniques, but it is still not solid proof to validate the current -collector -free techniques. d air may introduce carbon dioxide, which could potentially form a Sr(CO 3)2 layer at the surface, thus deteriorating the oxygen surface exchange kinetics. In conc lusion , and OTR showed a promising start in val idating different current collector -free techniques. 77 Figure 6.4 Comparison of kchem values measured from curvature relaxation (kR) and optical relaxation (OTR) 78 6.5 Summary In this work, kchem measurements were conducted on two near identical STF35 thin film with two current collector -free techniques, namely curvature relaxation and optical relaxation. The kchem values from the two data sets are comparable , which indicates a promising start for validating the kchem measurements of different current -collector -free techniques. However, the reasons for the difference between two data sets are still unclear. Further investigation is needed to identify the cause of this variation. 79 7. 7. Effect of Silicon Contaminants 7.1 Introduction The oxygen surface exchange coefficient ( kchem ) is crucial to a variety of applications. It determines the current density, efficiency, and response time for applications like solid oxide fuel cells , 64 -66, 68 electrolysis cells , 69 -72 gas separation devices ,77 catalytic converters ,80 gas sensors ,78 and memristors .216 Typically, a group of materials called mixed electronic conductors (MIECs) are u sed for th ese applications for their fast oxygen surface exchange pr ocess. Praseodymium doped ceria (PCO), for example, is one of the most studied of MIEC materials because (a) it can exchange oxygen with the surrounding gas atmosphere in air at elevated temperatures 18 (b) it has the well -studied defect chemistries ,57, 204, 207 mechanical properties , 144, 161, 205 and electrochemical properties .18 Those advantages ma ke it an ideal model material for studying oxygen surface exchange behavior. During the oxygen exchange process, PCO goes through the following defect reaction :29, 217 2+..+122+(5.1) However, surface exchange reactions of MIEC materials can be affected by impurities ,218 -232 Siliceous contaminants, for example, are common impurit ies that can be introduced during fabrication and operation . 226, 227, 233 -235 It can hinder the surface exchange reactions of commonly used MIEC materials like lanthanum strontium cobalt iro n oxide 228 -230 and lanthan um strontium cobalt oxide 231 . It was also reported that the bulk conductivity of ceria -based materials can be easily affected by Si .236 Furthermore , there is no systematic, quantitative study about the effect of Si contaminants in the literature. 232 Especially for the surface exchange process, knowledge of the chemical environment at the surface is crucial for kchem measurements. 80 Therefore, the objective s of this work w ere to determine the surface content of the PCO thin film s after the aging process in the Si-rich environment, and to measure the kchem values for the aged PCO thin film s. These objectives were achieved by doing the in-situ , contact -free, current -collector -free a nd non-destructive curvature measurements , accompanied by the surface characterization methods like X -ray photoelectron spectroscopy (XPS) and time -of-flight secondary ion mass spectroscopy (ToF -SIMS). 7.2 Experimental Method 7.2.1 Sample Fabrication Detailed thin film deposition information can be found in Section 3.3.1. After deposition, all three samples were re -oxidized at 1000 oC for 1 hour with a 3 oC/min heating and cooling rate to alleviate the extremely high oxygen nonstoichiometry in 10PCO thin film result ing from the deposition process. Then two 10PCO|YSZ samples were etched with 50% NaOH solution at 65 oC for 24 hours with 100 rpm stirring speed in order to remove any possible Si contaminants that is on the surface of the sample before the aging process .161 To ensure there was no Si contamination on the etc hed samples during curvature relaxation measurements, one of the etched 10PCO|YSZ samples was sectioned into 4 pieces. Later, they were aged at 525 oC, 525 / 550 oC, 525/550/575 oC, and 525/550/575/600 oC with 5 oC/min heating rate. The dwell time was 1.5 hour at 525 oC, 1 hour at 550 oC, 50 min at 575 oC, and 30 min at 600 oC. Then th ese aged samples were characterized ex-situ with XPS to determine if Si contaminants were on the surface. To further quantify the amount of Si contamination, the near -surface Si concentration of Si was measured by the ToF -SIMS. 81 7.2.2 Microstructural and Crystallographic Characterization Scanning electron microscopy (SEM) images were taken via a TESCAN MIRA3 Field Emission SEM (Tescan Inc) o f the cross section of a fractured samp le. Prior to the imaging process, the fractured samples were coated with ~5 nm of Ti. X-Ray Diffraction (XRD) was conducted using a Rigaku SmartLab diffractometer with 40 o 80o with 0.01o step size and 1s dwell time. 7.2.3 XPS Measurements XPS measurements w ere conducted using a Physical Electronics (PHI) 5400 X -Ray photoelectron spectroscopy system. An aluminum X -Ray was operated at 300 W and 23.50 eV pass energy. The peak positions were calibrated with the C 1s peak at 284.8 eV. Survey scans of binding energy were conducted from 70 eV to 180 eV. 7.2.4 ToF -SIMS Measurements The ToF -SIMS depth profiling was performed by EAG Labs (East Windsor, NJ, USA). Only the sample aged at 600 oC was measured using ToF -SIMS due to the cost of performing depth profiling measurements . 7.2.5 Curvature Relaxation Measurements For the Si-free sample, the curvature relaxation signals were measured from 500 oC to 600oC in synthetic air in 25 oC increments. For Si-contaminated sample, the curvature relaxation signals were measured from 675 oC to 725 oC in synthetic air in 25 oC increments. The relaxations were triggered by switching between synthetic air (20% O 2-80% Ar) and diluted synthetic air (10% synthetic air - 90% Ar). Noted that the temperature ranges of the measurements are different. It is 82 because for the Si-contaminated sample, the relaxation process was too slow to measure between 500~600 oC. To prove that an oxygen partial pressure from 0.21 to 0.021 did not change the oxygen surface exchange coefficient, the relaxations between synthetic air and 5 times diluted synthetic air (20% synthetic air Œ 80% Ar) were also measured . 7.3 Results and Discussion 7.3.1 Crystallography and Morphology of the Film Figure 7.1 shows representati ve XRD scans of the Si -free and Si -contaminated samples . It indicates that both 10PCO films are phase pure and well -crystallized. Both 10PCO thin films have the (100) preferred -orientation on (100) oriented YSZ substrates. This preferred orientation is ide ntical to what was reported in the literature studies using (100) YSZ substrate .18, 141, 161, 237 -240 The averaged grain size is 24 nm for the Si -free sample and 21 nm for the Si -contaminated sample, which were calculated using Scherrer™s Equation .209 Due to the limited peaks for a preferred oriented 10PCO thin film, the more accurate Williamson -Hall method 210 was not used for grain size determination. Therefore, the peak broadening generated becaus e of the strain effect was not accounted for. The reported grain sizes here are only the lower limits. Note that the peak intensit ies of the Si-free and Si-contaminated sample are similar, the peak height difference is caused by different intensity ratios between 10PCO (200) and YSZ (200) peaks (no amorphous peak). This ratio variation can be caused by different batches with different sample height calibrations. Therefore, the crystallinit ies of the Si-free and the Si -contaminated sample s are similar, which is not a variable in the discussion below. 83 Figure 7.1 Representative XRD scans of (a) etched and (b) non -etched 10PCO|YSZ samples using using CeO 2, and YSZ JCPDS card numbers 81 -0792 and 70-4436, respec tively. 84 Figure 7.2 shows the representative SEM image of 10PCO thin film after the deposition , which indicates that the 10PCO thin films were dense and uniform with a thickness of 230 ± 5 nm. The film thickness is significantly smaller than the characteristic thickness of 10PCO 18 , ensuring that oxygen transport was dominated by the surface exchange process. After te sting, the SEM scan did not show any microstructural change in the 10PCO thin films. Figure 7.3 shows the representative AFM images of 10PCO|YSZ samples before and after the etching process, showing that the surface roughness of the film did not have a si gnificant change after the etching process. This was expected because ceria is not reactive to a KOH solution 241 . Therefore, the difference in oxygen surface exchange coefficients between the etched and non -etched samples would not be a result of surface roughness . Figure 7.2 Cross -sectional back -scatte red electron image of 10PCO|YSZ 85 Figure 7.3 AFM scans of 10PCO|YSZ sample (a) before and (b) after etching 86 7.3.2 Near -Surface Si Content Characterization Figure 7.4. shows the XPS spectrum of the etched, aged, and non -etched samples. The XPS spectrum of the non-etched sample at 25 oC shows that the as -deposited sample have surface Si contaminants , which could be introduced during the deposition process since the deposition chamber was used to deposit silicate material s prior to the 10PCO deposition . Therefore, the etching process is needed to guarantee a 10PCO thin film with Si -free surface. Compari ng the XPS Spectra of etched and non -etched sample at 25 oC, the etching process completely removed surface Figure 7.4 XPS spectra of (a) non -etched and (b) etched sample aged at different temperatures, not ing that below 600 oC, the re is no Si contamination detected for etched samples 87 Si-contaminants. The XPS results are in good agreement with the etching rate of Si and SiO2 reported by Seidel et al. 242 The XPS peak patterns of etched and non -etched 10PCO also match well with the XPS data report ed previously for Si -free and Si -contaminated 10PCO thin films .232 Comparing the XPS spectrum of aged and etched samples, for curv ature relaxation measurements, there are no Si contaminants detected at 525, 550 and 575 oC. The Si contaminants were detected after the curvature relaxation measurements at 600 oC. Those results were expected since the optical relaxation work of Zhao et al. reported noticeable Si -contamination after aging the sample in a Si -based test -rig for 60 hrs at 600 oC 232 . Since the deposition speed of Si became lower with lower temperatures, and with the exposure of a short amount of time, the surface of the sample is therefore free from detectable Si -contamination below 600 oC. Figure 7.5 shows the ToF -SIMS depth profil e near the etched and 600oC-aged 10PCO|YSZ sample surface. The concentration of Pr, Ce, and O rises significantly as the depth increases while the concentration of Si drops from 3*10 21 to 4*10 20 atoms/cm 3 at a depth of ~10 nm, indicating a surface covered with Si contaminants. The slight increase in Si concentration between 10~20 nm could be that there is a Si -rich region in the bulk of the film, which could form during the deposition in a Si -contaminated d eposition chamber. However, the Si in the bulk of the film do not dominate the oxygen transport in the 10PCO thin films since the oxygen surface exchange process is the rate limiting step. Therefore, the effect of Si in the bulk is not considered in later discussion. On the other hand, Zr and Y are also detected at the surface and stays relatively constant as the depth increases . However, compar ed to the concentration of Si contaminants, the contribution of Zr and Y are not significant. For the layer contai ning silicate contaminants, the oxygen to cerium ratio is significantly larger than the stoichiometry of ceria , indicating a lack of oxygen vacanc ies , which may hinder the oxygen surface exchange process of the 10PCO thin film. Additionally, the 88 thickness of the contaminants is ~10 nm. Therefore, an overnight etching at 65 oC in 50% NaOH is sufficient to remove the surface silicate contaminants, according to previous Si and SiO 2 etching studies 242 . Figure 7.5 Near -sur face Si concentration of etched and 600 oC-aged 10PCO|YSZ measured by ToF -SIMS 89 7.3.3 Curvature Relaxation Measurements Figure 7.6 shows the representative curvature relaxation data for Si-free and Si-contaminated samples. The data from the Si-free sample was measured at 600 oC while the data for the Si-contaminated sample was measured at 7 50oC. Measurements on both samples have Figure 7.6 Representative curvature relaxation data of Si-free and Si-contaminated samples. The relaxation time for the Si -contaminated sample is significantly longer than Si-free sampl e (note difference in horizontal time scale) 90 reproducible time constants after reduction and re -oxidation. Both samples reached a steady -state equilibrium after being exposed to a different oxygen partial pressure. Curvature relaxation data measured at other temperatures is shown in Figure 5.7. Figure 7.7 Curvature relaxation raw data at different temperatures for (a) Si-free sample and (b) Si-contaminated sample, note that the relaxation time of the Si-contaminated sampl e is significantly l onger than etched sample 91 Comparing Figure 7.6(a) and Figure 7.6(b), it is obvious that even if the Si-contaminated sample was measured at a higher temperature, its relaxation process is significantly slower than the Si-free sample. The l arger relaxation time constant of the non -etched sample indicates that oxygen surface e xchange process of 10PCO was blocked by surface Si contaminants. The increase of relaxation time was also observed by Zhao et al .232 Figure 7.8 Multiple 750oC Curvature relaxations performed on the same 10PCO thin film by switching between (a) syntheti c air (pO 2=0.21) and a 10 times diluted air (10% synthetic air - 90% Ar, pO 2=0.021) mixture, and (b) synthetic air (pO 2=0.21) and a 5 times diluted air (20% synthetic air Œ 80% Ar, pO 2=0.042) mixture 92 As shown in Figure 7.8, curvature relaxation measurements were also conducted between 0.21 and 0.042. They were conducted to prove that the step size of oxygen partial pressure (0.21 to 0.021) is small enough to ensure a linear relationship between cur vature and oxygen nonstoichiometry. Detailed fittings for all the relaxation curves can be found in Figure 7.9, 7.10 and 4.2, indicating that there is only one mechano -chemically active process during the redox cycling. Figure 7.9 Fittings of curvature relaxation measurements at different temperature s for non -etched samples 93 7.3.4 Oxygen Surface Exchange Kinetics Figure 7.11 (a) shows the oxygen surface exchange coefficient measured in this work compared to those measured from other literature studies. As reported in other studies, the oxidation process measured here is kinetically faster than the reduction process. A larger oxygen vacancy concentration at the beginning of the oxidati on process might be responsible for a faster oxygen transport process .160, 243, 244 The k chem values for both Si -free and Si -contaminated samples showed Arrhenius behavior over the temperature range of their Figure 7.10 ln( 1-Normalized Curvature) plots for the non -etched sample. Note that the generally good single time constant (red line) fits to the data of Figure 7.9 and a single slope in the ln(1 -Normalized Curvature) plots of Figure 7.10 at times before the equilibrium s tate is reached suggest only one mechano -chemically active process is active with redox cycling between synthetic air and 10% diluted synthetic air. 94 Figure 7.11 (a)10PCO oxygen surface exchange coefficient from curvature relaxation comp aring with other literature studies (b) 10PCO surface polarization resistance (R S) values compared to literature report ing 10PCO R S values and other R S values of conventional solid oxide fuel cell cathode materials 95 measurements, indicating that there w as only one mechano -chemically active process on both samples. Most importantly, the k chem value of Si -free sample s is ~3 orders of magnitudes higher than Si -contaminated samples . The discrepancy between those two data sets agrees with the variation report ed by Zhao et al . between Si -free and Si -contaminated samples ,232 which is likely to be a result of Si contaminants covering the surface reaction sites , which comes from the fused silica test -rig during high temperature measurements . Additionally, the activation energy of the oxygen surface exchange process was almost d oubled for the Si-contaminated sample, indicat ing that not only do the Si contaminants cover the surface reaction sites, it hinders the migration of oxygen ions. When comparing the kchem from this work with the values measured by electrical impedance spect roscopy (EIS) ,18 the values measured by R) are ~5 times smaller. As indicated by other literature studies ,157, 161, 245 this discrepancy can be a result of applying surface -altering Au/Pt current collector. It is wort h noting that the 600 oC kchem data measured in this study is very likely to be affected by surface Si contaminants according to XPS data. However, the value of kchem agrees well with what Zhao et al. 232 reported in the Si -free film. This result is expected because Zhao et al. aged their samples for 60 hrs at 600 oC to get a sign ificant change (~3 orders of magnitudes) in the oxygen surface exchange coefficient. With only 2 hours of aging at 600 oC. It is possible that the film is not fully covered by Si contaminants judging by the Arrhenius behavior between 500~600 oC, resulting in a similar kchem value as Zhao et al. On the other hand, even for the Si -free sample from Zhao et al , it is still possible that there were already Si -contaminants on the surface during their optical relaxation measurements (No XPS results were provided aft er the measurements for the Si -free samples) .232 Therefore, with the reasons mentioned above, the k chem value at 600 oC still repre sents the response of the PCO thin film. 96 Figure 7.11 (b). shows the surface polarization resistanc e, compared to those measured in other literature studies. It shows that the Si -contaminated sample has a higher resistance than the Si-free sample. There ar e ~3 orders of magnitudes of difference between th ese two data sets. This variation is mainly caused by a difference in kchem as it was shown in Figure 5.11 (a) ., which also have ~3 orders of magnitudes difference between Si -free and Si -contaminated samples. Consistent with kchem , there are also differences between EIS and curvature relaxation measured data for the Si-free samples . As mentioned above, this discre pancy can be explained by the catalytic effect of the Au current collector used in EIS measurements. 7.4 Summary In this work, a systematic aging and characterization processes was conducted to investigate the influence of Si contaminants on the oxygen surface exchange process of the 10PCO thin films. The XPS results show that Si contaminants were introduced after the aging process at 600oC due to the usage of fused silica for the c urrent curvature relaxation test -rig. The oxygen surface exchange coefficient, surface polarization resistance, and activation energy of 10PCO thin films are all affected by Si contamination. With the measurement via curvature relaxation technique, ~3 orde rs of magnitudes degradation of th ese properties have been observed. Also, ~3 orders of magnitudes higher Rs and ~2 times larger activation energy has been measured from Si -contaminated samples, in comparison with Si -free samples. Th ese phenomena indicate that Si contaminants block the surface reaction sites for the oxygen surface exchange process, which hinders the oxygen migration process in 10PCO thin film s. 97 8. Effect of Surface Platinum Coverage 8.1 Introduction The oxygen surface exchange coefficient ( kche m) is a material property determining the speed of oxygen exchange in the mixed ionic electronic conductors (MIEC) used in oxygen sensors, Solid Oxide Fuel Cells, Solid Oxide Electrolysis Cells, mechano -chemical actuators, and other electrochemical devices . Historically, in-situ techniques like electrical impedance spectroscopy (EIS) and electrical conductivity relaxation (ECR) have been used to measure kchem .18, 21 Unfortunately, these techniques require noble -metal electronic current collectors th at may alter the underlying MIEC stress state (potentially altering MIEC kchem performance through mechano -chemically -induced point defect concentration changes) or interfering with an accurate MIEC kchem measurement through catalysis of the oxygen exchang e reaction .23, 157 Recently there are studies which implementing current -collector -free, contact -free, in -situ techniques to study the oxygen surface exchange coefficient with the purpose of excluding the effect of Pt enhancement .23, 157, 246 However, the effect of platinum enhancement was only shown with the comparison between different studies, there is a lack of systematic studies for the effect of platinum surface coverag e on the oxygen surface exchange process of a MIEC. In this work, two identical PCO|YSZ samples were manufactured using PLD. One of the samples went through a photolithography process and has a Pt patterns deposited on the surface of the sample (which is shown in Figure 8.1) . The kchem of two samples were then measured via curvature relaxation technique and then compared with each other in order to show the influence of Pt surface coverage. 98 8.2 Experimental Details 8.2.1 Pulsed Laser Deposition ~200 nm thick P r0.1 Ce0.9 O2- (PCO) thin films were deposited onto (100) oriented, 200 micron thick, 1 -inch diameter, one -side -polished, 9.5% YSZ substrate (Crystec GmbH, Berlin, Germany) . Thin films were deposited at 700 oC (setpoint of the substrate thermostat) w ith 30 mTorr oxygen for 15000 pulses. The target to substrate distance was 100 mm. The pulse frequency was 10 Hz. The sample s were cooled with a 10oC/min cooling rate in 30 mTorr oxygen . After the deposition, the samples were annealed in air at 1000 oC for 1 hour in order to re -equilibrate its oxygen content. 8.2.2 Crystallographic Characterization The detailed XRD parameter s can be found in Chapter 3.3.2. 8.2.3 Photolithography The surface of PCO|YSZ were cleaned with acetone wash followed by a methanol wash, dipped in water, and N2 gas drying. The sample w as pre -heated to 115 oC for 5~10 min to get rid of the residual water on the surface. Then the sample was placed in a spin coa ter. The spinning program was set to rotate at 700 rpm for 10 seconds and them 3000 rpm for 30 seconds. The Photoresists S1813 ( Microchem Corp, Westborough, MA) was dripped onto the sample with a pipette when the sample was rotating at 700 rpm. After the s pin coating process, the coated sample was then baked at 116 oC for 1 min. Then the sample was transferred into a Karl Suss (MJB3) m ask aligner (SUSS MicroTec SE, Garching, Germany) with a mask from Photoscience (Photoscience Inc, Lexington, KY). The coated photoresist was exposed to 274 watts of ultra -violet light for 90 seconds and then developed with photoresist remover MF 319 (Microchem Corp, Westborough, 99 MA) for 45 seconds. The developed sample was then rinsed with water and then baked at 116 oC again fo r 5 min. 8.2.4 Pt Deposition and Photoresist Removal The Pt was deposited by a sputtering process. The PCO|YSZ sample masked with photoresist w as placed into the deposition chamber, pump ed down to 10 -7 torr before deposition. Then the chamber was filled with A r to 10 mTorr. The Pt was deposited with 200W power for 2 min with a sample rotation speed of 30 rpm. The remaining photoresist w as cleaned after the Pt deposition. The sample w as dipped into acetone to remove the photoresist and then cleaned with isopropanol to remove the remaining acetone on the surface of the sample. Figure 8.1 shows the image of bare wafer, PCO|YSZ and Pt|PCO|YSZ. 8.2.5 Curvature relaxation measurements The curvature relaxations were triggered by switching oxygen partial pressure from 0.21 (21% O 2-79% Ar) to 0.021 (2.1% O 2-97.9% Ar) . The platinum coated PCO|YSZ (Pt|PCO|YSZ) Figure 8.1 Image of bare YSZ wafer, PCO|YSZ and Pt|PCO|YSZ Bare Wafer PCO|YSZ Pt|PCO|YSZ 100 was measured from 500 to 725 oC with 25 oC incr ements. The PCO|YSZ sample w as measured from 675 to 725 oC with 25 oC increments. To ensure a linear response between oxygen nonstoichiometry change and curvature change, curvature relaxation s between 0.21 and 0.042 oxygen partial pressure were also measured for Pt|PCO|YSZ at 725 oC 8.3 Results and discussion 8.3.1 X-Ray Diffraction Analysi s Figure 8.1 shows the XRD results from the PCO|YSZ sa mple. The PCO thin film has a (100) preferred orientation on YSZ substrate, which agrees with the result s in Chapter 3, 4, 5, and 6 despite a difference in deposition conditions. Figure 8.2 XRD scan of PCO|YSZ sample 101 8.3.2 Curve Fitting of Pt|PCO|YSZ Figure 8.2 shows the representative fitting of the relaxation curve of Pt|PCO|YSZ at 725 oC. The fitting showed that although there is Pt surface coverage, Eqn 4.6 is still sufficient to fit the relaxation signal , meaning there is only one oxygen surface exchange process during the relaxation period. Figure 8.3 Representative curve fitting of Pt|PCO|YSZ at 725oC 102 8.3.3 Effect of Pt surface Coverage Figure 8.4 shows the comparison of kchem values between PCO|YSZ and Pt|PCO|YSZ . Both data sets show Arrhenius behavior, indicating only one dominant oxygen transport process dominating the relaxation. When comparing the kchem between PCO|YSZ and Pt|PCO|YSZ, the kchem value of P t|PCO|YSZ is 2~3 times higher than the kchem value for PCO|YSZ . Additionally, the activation energy reduced from 0.8 eV to 0.4 eV. Th is behavior indicates that the Pt improves the oxygen migration process at the surface of the PCO thin film. The boost of oxygen sur face exchange is expected due to the similar mechanism as the spillover effect, where the gas molecule can be adsorbed and dissociate d on a transition metal surface, and then diffused into a non -metal substrate. 247 The current oxygen surface exchange coefficient measurement t echnique can only determines the overall t ime constant for oxygen surface exchange process at the gas/solid interface, therefore, a reduction of energy barrier for oxygen adsorption and dissociation does not get detected in the relaxation curve fit but it still giv es an overall boost of oxygen sur face exchange coefficient. Additionally, the kchem values measured between 0.21~0.021 pO 2 is similar to the value measured between 0.21~0.042 pO 2, indicating that there is still a linear response between the curvature and the oxygen nonstoichiometry changes of 10PCO with the pO 2 step size from 0.21 to 0.021, making sure that Eqn 4.6 is applicable to fit the relaxation curves. 103 Figure 8.4 kchem comparison between PCO|YSZ and Pt|P CO|YSZ 104 8.3.4 XPS Analysis Figure 8.5 shows the XPS scan of the as -deposited PCO|YSZ sample. The XPS signal shows that there are Mo, and Cs on the surface of the PCO thin film. This may be the reason for the ~1 order of magnitude lower kchem value measured in this chapter in contras t to the data from Chapter 4. The Mo may come from the heating element of the PLD system. The Cs may come from the previous depositions that deposited Cs -containing materials. 8.4 Conclusions This chapter investigated the effect of Pt surface coverage on the oxygen surface exchange process of PCO thin film. Two identical PCO|YSZ samples were used with one of them deposited with Pt pattern on the surface. The kchem results from curvature relaxation shows that the Pt Surface coverage not only boost the k chem val ues 2~3 times, it also decrease s the activation energy required for oxygen migration. Therefore, techniques that require a current collector (ECR, EIS, etc.) at the surface can overestimate the oxygen surface exchange coefficient of MIEC, which could be on e Figure 8.5 XPS scan for as -deposited PCO|YSZ 105 of the sources of kchem variation in the literature studies among different research groups. However, the relationship between kchem enhancement and area coverage of Pt is still unclear, which will be investigated in the future work. 106 9. Dissertation Summa ry The objective of this work, as shown in Figure 9.1,23 is to develop an in-situ technique that can measure Young™s modulus (E), thermo -chemical expansion coefficient thermochemical ), oxyge n surface exchange coefficient (k chem ), oxygen nonstoichiometry (ö), and surface polarization resistance (R s). Then using this technique to detect t he effect of surface contaminants on the oxygen surface exchange process . The conclusions from this dissertation are shown below: Figure 9.1 Graphic summary of objectives of this work 107 (1) wafer curvature measurements can be used to directly measure a variety of disparate and technologically relevant thin film physicochemical properties (i.e. the oxygen nonstoichiometry, biaxial modulus, thermo -chemical expansion coefficient, and thermal expansion coefficient) under well -characterized film st ress states , strain states, temperatures and atmospheric conditions. Using an externally -derived Poisson™s ratio, Young™s modulus was also determined (Note, studies have shown that wafer curvature experiments can also be used to directly measure the Poisso n™s ratio 118 ). (2) A n in-situ , contact -free, current -collector -free technique can be used to evaluate the oxygen surface exchange kinetics of 10PCO thin film. With the knowledge of oxygen nonstoichiometry from dual substrate measurements , the oxygen surface exchange resistance can also be calculated. (3) The comparison of the oxygen surface exchange resistance between 10PCO and other commonly used MIECs suggests that 10PCO may not be an ideal material for solid oxide fuel cell applications. (4) The t hermo -chemical expansion coefficient and Young™s modulus of a thin film can be extracted by c ombining the HTXRD strain measurements and MOSS stress measurements. Similar results from HTXRD -MOSS and the dual substrate method validate the reliability of dual substrate method. (5) With HTXRD -MOSS method, the magnitude of the error for thermo -chemica l expansion coefficient and Young™s modulus were significantly reduced compar ed to the data from dual substrate measurements. 108 (6) A cross -check with other electrode -free kchem measurement techniques shows that the data from curvature relaxation technique is comparable with the data from optical relaxation technique . (7) Si contaminants on the surface not only degrade the value of kchem , but also increase the migration energy o f surface oxygen transport . (8) Pt surface coverage boost s the value of kchem 2~3 times while decreasing the migration energy for surface oxygen transport. 109 10. Future Work High temperature X -Ray diffraction was use d to verify the reliability of Young™s mod ulus and the thermo -chemical expansion coefficient measured by the dual -substrate method. It also provides an alternative option for measuring Young™s and the thermo -chemical expansion coefficient of the thin film samples. However, in the case of PCO , the films have a (100) preferred orientation. Due to the structural similarity between low -angle grain boundary and grain interior and the near -atmospheric oxygen partial pressure (pO 2=0.21), there is no significant difference between the expansion coefficient of the grain interior and grain boundary. But in real SOFC applications, a polycrystalline microstructure is often seen in a MIEC electrode . As reported by Sheth et al, 206 an ~200 MPa compositional stress difference was observed between two PCO|YSZ samples with different grain size s (27 nm vs. 72 nm) in PCO thin film s. The change in compositional stress could be crucial for the mechanical stability of the devices. Therefore, measuring the grain boundary, grain interior, and overall thermo -chemical expansion coeff icient of the PCO thin films with different grain sizes can be a useful guide to improve the manufacturing process and the design of SOFC related applications. The results from curvature relaxation measurements showed that the Pt surface coverage provides 2~3 times enhancement for the oxygen surface exchange coefficient of the PCO thin films. However, the relationship between the area of surface coverage and the magnitude of enhancement is still unclear. Additional curvature relaxation experiments with dif ferent surface coverage areas are therefore needed to systematically quantify the influence of Pt surface coverage. Given that the oxygen surface exchange coefficient of PCO can be boosted by Pt, the effect of other noble metal (Au, Ag, etc.) surface cov erages can also be important since the silver and gold paste are commonly used in the electrochemical measurements. On the other hand, since Pt, 110 Au, and Ag are all transition metal s, the effect of non -precious transition metal s (Fe, Ni, Co, Cu, etc.) could also be explored and potentially provide an alternative way to boost the performance of the MIECs. 111 APPENDICES 112 APPEND IX A : Derivation of Fitting Equation Used in Curvature Relaxation Data Processing The oxygen surface exchange coefficient measurement of a thin film MIEC can be modeled as a case of surface evaporatio n. The rate of losing evaporating substance on a planar sheet can be described as: DCx=()(A.1) where is the concentration of oxygen in the gas atmosphere, is the oxygen concentration in the bulk of the thin film. The concentration of oxygen at the beginning of the relaxation was defined as: =C(x,t)(A.2) Since the oxygen concentration of surrounding gas atmosphere didn™t change: C(0,t)t=0(A.3) Additionally, the oxygen transport through the thin film can be describe via Fick™s Second Law: C(x,t)t=(,)(A.4) with A.1, A.2, A.3, and A.4, the solution to Fick™s Second Law can be expressed as: =12exp (++)(A.5) where is the amount of oxygen transported into the film at a certain time, is the amount of oxygen transported into the film after infinite amount of time, and: tan=(A.6) and 113 L==(A.7) where l is the film thickness. In the case of k chem measurements, for most of the MIEC thin films, . Therefore, L is a very small number, in which case: tan==L(A.8) Applying A.7 and A.8 to A.5, A.5 can then b e simplified into: =()==1exp (A.9) 114 APPENDIX B : Error Analysis for Dual Substrates Techniques The following equations were used for fitting the stress (y) Œ temperature (x) data: =+++(B.1) =3+2+(B.2) Separating out the temperature from the fitting parameters a, b, and c, the slope of the stress -temperature curve (denoted S) was expressed as: =(,,)=3+2+(B.3) where the error bars of a, b and c is automatically calculated when fitting the raw data in the origin computer program. Mathematical error analy sis 248 indicates that the error in (i.e. S) can then be expressed as: =++=(3)+(2)+(B.4) Based on Equations 2 and 3 in the literature, the equation for the calculation of thermo -chemical expansion coefficient is : = (B.5) where is the slope of the stress vs. temperature curve for 10PCO|YSZ, is the slope of stress vs. temperature curve for 10PCO|MgO, is the thermal expansion coeffi cient of MgO substrate from the literature, 125 is the thermal expansion coefficient of YSZ substrate from the literature, 111 and is the thermo -chemical expansion coefficient of PCO. For an equation like S5 of the form of (in Equation S5 = and = ) mathematical error analysis 248 indicates that the error for A can be calculated by: 115 = + + (B.6) and the error for B can be calculated by: =+ (B.7) assuming that there is no error in and (an assumption based on the fact that no errors were reported for these literature studies 111, 125 ). Plugging in Equation S4 to S6 -7, the error for the error equals: = +(B.8) The equation for calculation of Young™s Modulus is: = 10000.67(B.9) Using the same approach as that described previously for the thermo -chemical expansion coefficients, the error for Young™s modulus is: = 10000.67+ (B.10) Mathematical error analysis 248 indicates that the error caused by fitting the total stress with an equation for the form: =+++(B.11) is: =+++=()+()+()+(.12) 116 Mathematical error analysis 248 indicates that the error caused by fitting the thermal stress with an equation for the form: =+(B.13) is: = + =()+(B.14) The chemical film strain ( ) is calculated by the equation: = (B.15) where is the total stress on the film, is the thermal stress on the film, and is the biaxial modulus of the film. Considering that Equation S18 has the form of (where = and =), the error of can be calculated as: =+ (B.16) and =(B.17) Hence, then the error on the chemical strain can be calculated as: =+(B.18) The Oxygen nonstoichiometry is determined from the chemical strain via the equation: == [S22] 117 Ass uming chemical expansion coefficient is 0.067, remains constant, 114 and has no error associated with it (based on the fact that none was provided in its measurement 114 ) the error on the oxygen nonstoichiometry can be calculated with the equation: =. [S23] 118 APPENDIX C : Error Analysis for HTXRD -XRD Measurements The strain vs. temperature data was fitted by: =(C.1) The first -order derivatives of the strain values with respect to temperature can be expressed as: ==(C.2) The error of the fitting can be expressed as: =()+(+)(C.3) According to Eqn 3, the error of thermo -expansion coefficient is: =11+()+21(C.4) 119 BIBLIOGRAPHY 120 BIBLIO GRAPHY 1. E. I. Administration and E. Department, Annual Energy Outlook 2015: With Projections to 2040 , Government Printing Office, 2015. 2. S. Hirota and T. Tanaka, Journal , 1999. 3. E. D. Wachsman, C. A. Marlowe and K. T. Lee, Energy & Environmental Science , 2012, 5, 5498-5509. 4. S. Curtin and J. Gangi, Fuel Cell and Hydrogen Energy Association, Washington , 2016. 5. S. M. Haile, Acta Materialia , 2003, 51, 5981-6000. 6. R. O'hayre, S. -W. Cha, W. Colella and F. B. Prinz, Fuel cell funda mentals , John Wiley & Sons, 2016. 7. L. Carrette, K. A. Friedrich and U. Stimming, ChemPhysChem , 2000, 1, 162 -193. 8. M. Bischoff, Journal of Power Sources , 2006, 160 , 842-845. 9. J. D. Nicholas, The Electrochemical Society Interface , 2013, 22, 49-54. 10. E. D. Wachsman and K. T. Lee, Science , 2011, 334 , 935-939. 11. F. Baumann, J. Fleig, H. Habermeier and J. Maier, Solid State Ionics , 2006, 177, 1071-1081. 12. F. Baumann, J. Fleig, G. Cristiani, B. Stuhlhofer, H. -U. Habermeier and J. Maier, Journal of The Electrochemical Society , 2007, 154, B931 -B941. 13. F. S. Baumann, J. Fleig, M. Konuma, U. Starke, H. U. Habermeier and J. Maier, Journal of the Electrochemical Society , 2005, 152, A2074 -A2079. 14. Q. Liu, K. A. Khor and S. Chan, Journal of Power Sources , 2006, 161, 123-128. 15. T. E. Burye and J. D. Nicholas, Journal of Power Sources , 2016, 301, 287-298. 16. M. J. Jørgensen and M. Mogensen, Journal of The Electr ochemical Society , 2001, 148, A433 -A442. 17. Z. Shao and S. M. Haile, in Materials for Sustainable Energy: A Collection of Peer -Reviewed Research and Review Articles from Nature Publishing Group , World Scientific, 2011, pp. 255-258. 18. D. Chen, S. R. Bish op and H. L. Tuller, Journal of Electroceramics , 2012, 28, 62-69. 121 19. Q. Yang and D. J. D. Nicholas, Journal of the Electrochemical Society , 2014, 161, F3025 -F3031. 20. M. Mosleh, M. Søgaard and P. V. Hendriksen, Journal of The Electrochemical Society , 2009, 156, B441 -B457. 21. J. E. ten Elshof, M. Lankhorst and H. J. Bouwmeester, Journal of the Electrochemical Society , 1997, 144, 1060-1067. 22. M. Søgaard, P. V. Hendriksen and M. Mogensen, Journal of Solid State Chemistry , 2007, 180, 1489-1503. 23. Y. Ma a nd J. D. Nicholas, Physical Chemistry Chemical Physics , 2018, 20, 27350-27360. 24. W. C. Oliver and G. M. Pharr, Journal of materials research , 2004, 19, 3-20. 25. G. Roebben, B. Bollen, A. Brebels, J. Van Humbeeck and O. Van der Biest, Review of scientifi c instruments , 1997, 68, 4511-4515. 26. E. P. Papadakis, Journal of testing and evaluation , 1998, 26, 240-246. 27. F. Kröger and H. Vink, in Solid state physics , Elsevier, 1956, vol. 3, pp. 307 -435. 28. S. Bishop, D. Marrocchelli, C. Chatzichristodoulou, N . Perry, M. B. Mogensen, H. Tuller and E. Wachsman, Annual Review of Materials Research , 2014, 44, 205-239. 29. H. L. Tuller and S. R. Bishop, Annual Review of Materials Research , 2011, 41, 369-398. 30. H. J. Bouwmeester, H. Kruidhof and A. Burggraaf, Soli d State Ionics , 1994, 72, 185-194. 31. S. Carter, A. Selcuk, R. Chater, J. Kajda, J. Kilner and B. Steele, Solid State Ionics , 1992, 53, 597-605. 32. T. Ishigaki, S. Yamauchi, K. Kishio, J. Mizusaki and K. Fueki, Journal of Solid State Chemistry , 1988, 73, 179-187. 33. J. Lane, S. Benson, D. Waller and J. Kilner, Solid State Ionics , 1999, 121, 201-208. 34. M. Leonhardt, R. De Souza, J. Claus and J. Maier, Journal of the Electrochemical Society , 2002, 149, J19 -J26. 35. W. Weppner and R. A. Huggins, Journal of The Electrochemical Society , 1977, 124, 1569-1578. 36. F. S. Baumann, J. Fleig, H. -U. Habermeier and J. Maier, Solid State Ionics , 2006, 177, 1071-1081. 37. R. De Souza and J. Kilner, Solid State Ionics , 1999, 126, 153-161. 122 38. I. Yasuda and T. Hikita, Journal of the Electrochemical Society , 1994, 141, 1268 -1273. 39. G. Kim, S. Wang, A. Jacobson and C. Chen, Solid State Ionics , 2006, 177, 1461-1467. 40. K. Kerman, C. Ko and S. Ramanathan, Physical Chemistry Chemical Physics , 2012, 14, 11953-11960. 41. C. Ko, A. Karthikeyan and S. Ramanathan, The Journal of chemical physics , 2011, 134, 014704. 42. E. Armstrong, K. Duncan, D. Oh, J. Weaver and E. Wachsman, Journal of The Electrochemical Society , 2011, 158, B492 -B499. 43. S. Diethelm, A. Closset, K. Nisançioglu, A. J. McEvoy and T. M. Gün, Journal of The Electrochemical Society , 1999, 146, 2606-2612. 44. M. Katsuki, S. Wang, K. Yasumoto and M. Dokiya, Solid S tate Ionics , 2002, 154, 589-595. 45. J. J. Kim, S. R. Bishop, N. J. Thompson and H. L. Tuller, Ecs Transactions , 2013, 57, 1979-1984. 46. S. R. Bishop, N. H. Perry, D. Marrocchelli and B. W. Sheldon, Electro -Chemo -Mechanics of Solids , Springer, 2017. 47. J. E. ten Elshof, M. Lankhorst and H. J. Bouwmeester, Solid State Ionics , 1997, 99, 15-22. 48. M. Tsuchiya, N. A. Bojarczuk, S. Guha and S. Ramanathan, The Journal of chemical physics , 2009, 130, 174711. 49. J. B. Smith and T. Norby, Journal of the Electroc hemical Society , 2006, 153 , A233 -A238. 50. S. Kim, Y. Yang, A. Jacobson and B. Abeles, Solid State Ionics , 1998, 106, 189-195. 51. M. H. Lankhorst and H. J. Bouwmeester, Journal of the Electrochemical Society , 1997, 144, 1268-1273. 52. S. Diethelm, Solid S tate Ionics , 2004, 174, 127-134. 53. M. Sahibzada, W. Morton, A. Hartley, D. Mantzavinos and I. S. Metcalfe, Solid State Ionics , 2000, 136, 991-996. 54. J. Jamnik and J. Maier, Journal of the Electrochemical society , 1999, 146, 4183 -4188. 55. L. Wang, R. M erkle and J. Maier, Journal of The Electrochemical Society , 2010, 157, B1802 -B1808. 56. M.-B. Choi, S. -Y. Jeon, H. -S. Yang, J. -Y. Park and S. -J. Song, Journal of the Electrochemical Society , 2011, 158, B189 -B193. 123 57. J. J. Kim, S. R. Bishop, N. J. Thompson, D. Chen and H. L. Tuller, Chem Mater , 2014, 26, 1374-1379. 58. F. Noll, W. Münch, I. Denk and J. Maier, Solid State Ionics , 1996, 86, 711-717. 59. R. Moreno, P. García, J. Zapata, J. Roqueta, J. Chaigneau and J. San tiso, Chem Mater , 2013, 25, 3640-3647. 60. E. M. Hopper, E. Perret, B. J. Ingram, H. You, K. -C. Chang, P. M. Baldo, P. H. Fuoss and J. A. Eastman, The Journal of Physical Chemistry C , 2015, 119, 19915-19921. 61. R. Chater, S. Carter, J. Kilner and B. Steel e, Solid State Ionics , 1992, 53, 859-867. 62. J. Maier, Angewandte Chemie -International Edition , 2013, 52, 4998-5026. 63. S. Levasseur, M. Menetrier, Y. Shao -Horn, L. Gautier, A. Audemer, G. Demazeau, A. Largeteau and C. Delmas, Chemistry of Materials , 2003, 15, 348-354. 64. S. B. Adler, Chemical Reviews , 2004, 104, 4791-4843. 65. A. Atkinson, S. A. Barnett, R. J. Gorte, J. T. S. Irvine, A. J. McEvoy, M. B. Mogensen, S. Singhal and J. Vohs, Nature Materials , 2004, 3, 17-27. 66. Y. L. Lee, J. Kleis, J. Rossm eisl, Y. Shao -Horn and D. Morgan, Energy Environ. Sci. , 2011, 4, 3966-3970. 67. T. Das, J. D. Nicholas and Y. Qi, Journal of Materials Chemistry A , 2017, 5, 4493-4506. 68. K. D. Kreuer, Chemistry of Materials , 1996, 8, 610-641. 69. S. D. Ebbesen, S. H. Jensen, A. Hauch and M. B. Mogensen, Chemical Reviews , 2014, 114, 10697-10734. 70. G. M. Wang, H. Y. Wang, Y. C. Ling, Y. C. Tang, X. Y. Yang, R. C. Fitzmorris, C. C. Wang, J. Z. Zhang and Y. Li, Nano Letters , 2011, 11, 3026-3033. 71. M. G. Walter, E. L. Warren, J. R. McKone, S. W. Boettcher, Q. X. Mi, E. A. Santori and N. S. Lewis, Chemical Reviews , 2010, 110, 6446-6473. 72. W. C. Chueh, C. Falter, M. Abbott, D. Scipio, P. Furler, S. M. Haile and A. Steinfeld, Science , 2010, 330, 1797-1801. 73. J. J. S. Yang, D. B. Strukov and D. R. Stewart, Nat. Nanotechnol. , 2013, 8, 13-24. 74. F. Messerschmitt, M. Kubicek, S. Schweiger and J. L. M. Rupp, Adv. Funct. Mater. , 2014, 24, 7448-7460. 75. R. Korobko, A. Patlolla, A. Kossoy, E. Wachtel, H. L . Tuller, A. I. Frenkel and I. Lubomirsky, Advanced Materials , 2012, 24, 5857-5861. 124 76. T. Das, J. D. Nicholas, B. W. Sheldon and Y. Qi, Phys. Chem. Chem. Phys. , 2018, Accepted Manuscript: DOI: 10.1039/C8CP01219A . 77. H. J. M. Bouwmeester, Catalysis Today , 2003, 82, 141-150. 78. M. W. Ahn, K. S. Park, J. H. Heo, J. G. Park, D. W. Kim, K. J. Choi, J. H. Lee and S. H. Hong, Applied Physics Letters , 2008, 93. 79. C. G. Granqvist, Solar Energy Materials and Solar Cells , 2000, 60, 201-262. 80. A. Trovarelli, Catalysis Reviews: Science and Engineering , 1996, 38, 439-509. 81. D. Marrocchelli, S. R. Bishop, H. L. Tuller and B. Yildiz, Adv. Funct. Mater. , 2012, 22, 1958-1965. 82. T. Das, J. D. Nicholas and Y. Qi, Journal of Materials Chemistry A , 2017, 5, 25031-25043. 83. J. D. Nicholas, Y. Qi, S. R. Bishop and P. P. Mukherjee, Journal of the Electrochemical Society , 2014, 161, Y10 -Y11. 84. W. H. Woodford, Y. M. Chiang and W. C. Carter, Journal of the Electrochemical Society , 2010, 157, A1052 -A1059. 85. A. Atkinson, Solid State Ionics , 1997, 95, 249-258. 86. X. H. Liu, L. Zhong, S. Huang, S. X. Mao, T. Zhu and J. Y. Huang, ACS Nano , 2012, 6, 1522-1531. 87. R. W. Hertzberg, Deformation and Fracture Mechanics of Engineering Materials , John Wiley & Sons, New York, 4th edn., 1996. 88. A. Kushima and B. Yildiz, Journal of Materials Chemistry , 2010, 20, 4809-4819. 89. H. L. Tuller and S. R. Bishop, Annual Review of Materials Research , 2011, 41, 369-398. 90. M. Kubicek, Z. H. Cai, W. Ma, B. Yildiz, H. Hutter and J. Fleig, ACS Nano , 2013, 7, 3276-3286. 91. A. Fluri, D. Pergolesi, V. Roddatis, A. Wokaun and T. Lippert, Nature Communications , 2016, 7. 92. R. A. De Souza, A. Ramadan and S. Horner, Energy Environ. Sci. , 2012, 5, 5445 -5453. 93. A. K. Sagotra and C. Cazorla, ACS Appl. Mater. Interfaces , 2017, 9, 38773-38783. 94. N. Muralidharan, C. N. Brock, A. P. Cohn, D. Schauben, R. E. Carter, L. Oakes, D. G. Walker and C. L. Pint, ACS Nano , 2017, 11, 6243-6251. 95. S. D. Park, J. M. Vohs and R. J. Gorte, Nature , 2000, 404, 265-267. 125 96. T. S. Stefanik and H. L. Tuller, Journal of the European Ceramic Society , 2001, 21, 1967-1970. 97. A. Primo, T. Marino, A. Corma, R. Molinari and H. Garcia, Journal of the American Che mical Society , 2011, 133, 6930-6933. 98. M. Mogensen, N. M. Sammes and G. A. Thompsett, Solid State Ionics , 2000, 129, 63-94. 99. H. Uchida, N. Osada and M. Watanabe, Electrochemical and Solid State Letters , 2004, 7, A500 -A502. 100. S. R. Bishop, D. Marroc chelli, C. Chatzichristodoulou, N. Perry, M. B. Mogensen, H. Tuller and E. Wachsman, Annual Review of Materials Research , 2014, 44, In Press. 101. H. L. Tuller, S. R. Bishop, D. Chen, Y. Kuru, J. J. Kim and T. S. Stefanik, Solid State Ionics , 2012, 225, 194-197. 102. S. R. Bishop, T. S. Stefanik and H. L. Tuller, Phys. Chem. Chem. Phys. , 2011, 13, 10165- 10173. 103. G. G. Stoney, Proceedings of the Royal Society of London Series A -Containing Papers of a Mathematical and Physical Character , 1909, 82, 172-175. 104. L. B. Freund and S. Suresh, Thin Film Materials: Stress, Defect Formation and Surface Evolution , Cambridge University Press, New York, NY, 2003. 105. C. A. Klein, Journal of Applied Physics , 2000, 88, 5487-5489. 106. J. D. Nicholas, Extreme M ech Lett , 2016, 9, 405-421. 107. T. F. Retajczyk and A. K. Sinha, Applied Physics Letters , 1980, 36, 161-163. 108. M. M. de Lima, R. G. Lacerda, J. Vilcarromero and F. C. Marques, Journal of Applied Physics , 1999, 86, 4936. 109. J. Thurn and M. P. Hughey, Journal of Applied Physics , 2004, 95, 7892-7897. 110. I. Suzuki, S. -i. Okajima and K. Seya, Journal of Physics of the Earth , 1979, 27, 63-69. 111. H. Hayashi, T. Saitou, N. Maruyama, H. Inaba, K. Kawamura and M. Mori, Solid State Ionics , 2005, 176, 613-619 . 112. C. Lenser, F. Gunkel, Y. J. Sohn and N. H. Menzler, Solid State Ionics , 2018, 314, 204-211. 113. R. Chiba, H. Taguchi, T. Komatsu, H. Orui, K. Nozawa and H. Arai, Solid State Ionics , 2011, 197, 42-48. 126 114. S. R. Bishop, H. L. Tuller, Y. Kuru and B. Yildiz, Journal of the European Ceramic Society , 2011, 31, 2351-2356. 115. D. Chen, S. R. Bishop and H. L. Tuller, Chemistry of Materials , 2014, 26, 6622-6627. 116. J. G. Swallow, J. J. Kim, M. Kabir, J. F. Smith, H. L. Tuller, S. R. Bishop and K. J. Van Vliet, Acta Materialia , 2016, 105, 16-24. 117. J. Sheth, D. Chen, H. L. Tuller, S. T. Misture, S. R. Bishop and B. W. Sheldon, Phys. Chem. Chem. Phys. , 2017, 19, 12206 -12220. 118. J. H. Zhao, Y. Du, M. Morgen and P. S. Ho, Journal of Applied Physics , 2000, 87, 1575-1577. 119. T. Kushi, K. Sato, A. Unemoto, S. Hashimoto, K. Amezawa and T. Kawada, Journal of Power Sources , 2011, 196 , 7989-7993. 120. Y. Kimura, T. Kushi, S. -i. Hashimoto, K. Amezawa and T. Kawada, Journal of the American Ceramic Society , 2012, 95, 2608-2613. 121. A. Kossoy, Y. Feldman, R. Korobko, E. Wachtel, I. Lubomirsky and J. Maier, Adv. Funct. Mater. , 2009, 19, 634-641. 122. Y. Kuru, S. R. Bishop, J. J. Kim, B. Yildiz and H. L. Tuller, Solid State Ionics , 2011, 193, 1-4. 123. P. Gao, A. Bolon, M. Taneja, Z. Xie, N. Orlovskaya and M. Radovic, Solid State Ionics , 2017, 300, 1-9. 124. J. Zhang, C. Ke, H. Wu, J. Yu, J. Wang and Y. Wang, Journal of Alloys and Compounds , 2017, 718, 85-91. 125. I. Suzuki, Journal of Physics of the Ear th, 1975, 23, 145-159. 126. X. Y. Chen, J. S. Yu and S. B. Adler, Chemistry of Materials , 2005, 17, 4537-4546. 127. S. Hull, S. T. Norberg, I. Ahmed, S. G. Eriksson, D. Marrocchelli and P. A. Madden, Journal of Solid State Chemistry , 2009, 182, 2815-2821. 128. S. R. Bishop, K. L. Duncan and E. D. Wachsman, Acta Materialia , 2009, 57, 3596-3605. 129. S. R. Bishop, K. L. D uncan and E. D. Wachsman, Electrochemica Acta , 2009, 54, 1436-1443. 130. H. W. Chiang, R. N. Blumenthal and R. A. Fournelle, Solid State Ionics , 1993, 66, 85-95. 131. J. Sheth, D. Chen, J. J. Kim, W. J. Bowman, P. A. Crozier, H. L. Tuller, S. T. Misture, S . Zdzieszynski, B. W. Sheldon and S. R. Bishop, Nanoscale , 2016, 8, 16499-16510. 127 132. D. Er, J. Li, M. Cargnello, P. Fornasiero, R. J. Gorte and V. B. Shenoy, Journal of the Electrochemical Society , 2014, 161, F3060 -F3064. 133. L. A. Chick, L. R. Pederson, G. D. Maupin, J. L. Bates, L. E. Thomas and G. J. Exarhos, Materials Letters , 1990, 10, 6-12. 134. L. Zhao, N. H. Perry, T. Daio, K. Sasaki and S. R. Bishop, Chemistry of Materials , 2015, 27, 3065-3070. 135. M. Shirpour, G. Gregori, R. Merkle and J. Maier , Phys. Chem. Chem. Phys. , 2011, 13, 937-940. 136. International Center for Diffraction Data Joint Committee on Powder Diffraction Standards, Journal , 2012. 137. N. Savvides, A. Thorley, S. Gnanarajan and A. Katsaros, Thin Solid Films , 2001, 388, 177-182. 138. C.-M. Wang, S. Thevuthasan and C. H. F. Peden, Journal of the American Ceramic Society , 2003, 86, 363-365. 139. D. Q. Shi, M. Ionescu, J. McKinnon and S. X. Dou, Physica C , 2001, 356, 304-310. 140. N. L. Edleman, A. Wang, J. A. Belot, A. W. Metz, J. R . Babcock, A. M. Kawaoka, J. Ni, M. V. Metz, C. J. Flaschenriem, C. L. Stern, L. M. Liable -Sands, A. L. Rheingold, P. R. Markworth, R. P. H. Chang, M. P. Chudzik, C. R. Kannewurf and T. J. Marks, Inorganic Chemistry , 2002, 41, 5005-5023. 141. A. Cavallaro, F. Sandiumenge, J. Gazquez, T. Puig, X. Obradors, J. Arbiol and H. C. Freyhardt, Adv. Funct. Mater. , 2006, 16, 1363-1372. 142. A. L. Patterson, Physical Review , 1939, 56, 978-982. 143. G. K. Williamson and W. H. Hall, Acta Metallurgica , 1953, 1, 22-31. 144. R. Korobko, S. K. Kim, S. Kim, S. R. Cohen, E. Wachtel and I. Lubomirsky, Advanced Functional Materials , 2013, 23, 6076-6081. 145. S. B. Adler, Journal of the American Ceramic Society , 2001, 84, 2117-2119. 146. V. Kharton, F. Marques and A. Atkinson, Solid State Ionics , 2004, 174, 135 -149. 147. T. Das, J. D. Nicholas and Y. Qi, Journal of Materials Chemistry A , 2017, 5, 4493-4506. 148. K.-D. Kreuer, Chem Mater , 1996, 8, 610-641. 149. H. J. Bouwmeester, Catalysis Today , 2003, 82, 141-150. 150. J. J. Yang, D. B. Strukov and D. R. Stewart, Nature Nanotechnology , 2013, 8, 13-24. 128 151. F. Messerschmitt, M. Kubicek, S. Schweiger and J. L. Rupp, Advanced Functional Materials , 2014, 24, 7448-7460. 152. R. Korobko, A. Patlolla, A. Kossoy, E. Wachtel, H. L. Tuller, A. I. Frenkel and I. Lubomirsky, Advanced Materials , 2012, 24, 5857-5861. 153. T. Das, J. D. Nicholas, B. Sheldon and Y. Qi, Physical Chemistry Chemical Physics , 2018, DOI: 10.1039/C8CP01219A. 154. M.-W. Ahn, K. -S. Park, J. -H. Heo, J. -G. Park, D. -W. Kim, K. J. Choi, J. -H. Lee and S. -H. Hong, Applied physics letters , 2008, 93, 263103. 155. A. Trovarelli, Catalysis Reviews , 1996, 38, 439-520. 156. C. G. Granqvist, Solar Energy Materials and Solar C ells , 2000, 60, 201-262. 157. N. H. Perry, J. J. Kim and H. L. Tuller, Science and Technology of Advanced Materials , 2018, 19, 130-141. 158. J. Crank, The Mathematics of Diffusion , Oxford University Press, London, 1975. 159. D. Marrocchelli, S. R. Bishop, H. L. Tuller and B. Yildiz, Advanced Functional Materials , 2012, 22, 1958-1965. 160. Q. Yang, T. E. Burye, R. R. Lunt and J. D. Nicholas, Solid State Ionics , 2013, 249Œ250, 123-128. 161. Y. Ma and J. D. Nicholas, Phys. Chem. Chem. Phys. , 2018, 20, 27350-27360. 162. J. Maier, Physical chemistry of ionic materials: ions and electrons in solids , John Wiley & Sons, 2004. 163. M. Liu, Journal of the Electrochemical Society , 1997, 144, 1813-1834. 164. J. D. Nicholas, in Electro -Chemo -Mechanics of Solids , eds. S. Bishop, D. Marrocchelli, N. Perry and B. Sheldon, Springer, New York, 2017, DOI: http://link.springer.com/chapter/10.1007/978 -3-319 -51407-9_5, ch. 5, pp. 103 -136. 165. J. D. Nicho las, Extreme Mechanics Letters , 2016, 9, 405-421. 166. P. Simons, H. I. Ji, T. C. Davenport and S. M. Haile, Journal of the American Ceramic Society , 2017, 100, 1161-1171. 167. S. Wang, A. Verma, Y. L. Yang, A. J. Jacobson and B. Abeles, Solid State Ionics , 2001, 140, 125-133. 168. M. Tsuchiya, N. A. Bojarczuk, S. Guha and S. Ramanathan, Journal of Chemical Physics , 2009, 130, 174711-174711 to 174711 -174719. 129 169. K. Develos -Bagarinao, H. Kishimoto, J. De Vero, K. Yamaji and T. Horita, Solid State Ionics , 2016, 288, 6-9. 170. M. Leonhardt, R. A. De Souza, J. Claus and J. Maier, Journal of the Electrochemical Society , 2002, 149, J19 -J26. 171. N. H. Perry, J. J. Kim and H. L. Tuller, Science and Technology of Advanced Materials , 2018, 19, 130-141. 172. H. Borch ert, Y. V. Frolova, V. V. Kaichev, I. P. Prosvirin, G. M. Alikina, A. I. Lukashevich, V. I. Zaikovskii, E. M. Moroz, S. N. Trukhan, V. P. Ivanov, E. A. Paukshtis, V. I. Bukhtiyarov and V. A. Sadykov, The Journal of Physical Chemistry B , 2005, 109, 5728-5738. 173. Y. H. Gong, C. Y. Qin and K. Huang, Ecs Electrochemistry Letters , 2013, 2, F4 -F7. 174. J. G. Swallow, J. J. Kim, J. M. Maloney, D. Chen, J. F. Smith, S. R. Bishop, H. L. Tuller and K. J. Van Vliet, Nature Materials , 2017, 16, 749. 175. A. B ieberle -Hutter, J. L. Hertz and H. L. Tuller, Acta Materialia , 2008, 56, 177-187. 176. A. Bieberle -Hutter, M. Sogaard and H. L. Tuller, Solid State Ionics , 2006, 177, 1969-1975. 177. A. B. Bieberle -Hutter and H. L. Tuller, Journal of Electroceramics , 2006, 16, 151-157. 178. M. Sogaard, A. Bieberle -Hutter, P. V. Hendriksen, M. Mogensen and H. L. Tuller, Journal of Electroceramics , 2011, 27, 134-142. 179. F. S. Baumann, J. Fleig, G. Cristiani, B. Stuhlhofer, H. U. Habermeier and J. Maier, Journal of the Electrochemical Society , 2007, 154, B931 -B941. 180. H. Xiong, B. K. Lai, A. C. Johnson and S. Ramanathan, Journal of Power Sources , 2009, 193, 589-592. 181. J. D. Nicholas and S. A. Barnett, Journal of the Electrochemical Society , 2010, 157, B536 -B541. 182. M. Shah, J. D. Nicholas and S. A. Barnett, Electrochemistry Communications , 2009, 11, 2-5. 183. J. D. Nicholas, L. Wang, A. V. Call and S. A. Barnett, Phy s. Chem. Chem. Phys. , 2012, 14, 15379-15392. 184. T. E. Burye, H. Tang and J. D. Nicholas, Journal of the Electrochemical Society , 2016, 163, F1017 -F1022. 185. Z. Gao, L. V. Mogni, E. C. Miller, J. G. Railsback and S. A. Barnett, Energy Environ. Sci. , 2016, 9, 1602-1644. 130 186. J. W. Fergus, Journal of Power Sources , 2006, 162 , 30-40. 187. P. Jasinski, T. Suzuki and H. U. Anderson, Sensors and Actuators B: Chemical , 2003, 95, 73-77. 188. J. G. Swallow, J. J. Kim, M. Kabir, J. F. Smith, H. L. Tuller, S. R. Bis hop and K. J. Van Vliet, Acta Materialia , 2016, 105, 16-24. 189. N. Özer, Solar Energy Materials and Solar Cells , 2001, 68, 391-400. 190. H. Hayashi, M. Kanoh, C. J. Quan, H. Inaba, S. Wang, M. Dokiya and H. Tagawa, Solid State Ionics , 2000, 132, 227-233. 191. R. Ko, M. Ricken, J. No and I. Riess, Journal of Solid State Chemistry , 1989, 78, 136-147. 192. T. Hisashige, Y. Yamamura and T. Tsuji, Journal of alloys and compounds , 2006, 408, 1153-1156. 193. S. Sameshima, T. Ichikawa, M. Kawaminami and Y. Hirata, Materials Chemistry and Physics , 1999, 61, 31-35. 194. Y. Ma and J. D. Nicholas, Physical Chemistry Chemical Physics , 2018, Submitted . 195. M. Morales, J. J. Roa, X. Capdevila, M. Segarra and S. Piñol, Acta Materialia , 2010, 58, 2504-2509. 196. A. Atkinson and A. Selcuk, Solid State Ionics , 2000, 134, 59-66. 197. I. Kosacki, T. Suzuki, V. Petrovsky and H. U. Anderson, Solid State Ionics , 2000, 136, 1225-1233. 198. H. Inaba and H. Tagawa, Solid State Ionics , 1996, 83, 1-16. 199. H. Yahiro, K. Eguc hi and H. Arai, Solid State Ionics , 1989, 36, 71-75. 200. Y. Takasu, T. Sugino and Y. Matsuda, Journal of applied electrochemistry , 1984, 14, 79-81. 201. S. R. Bishop, T. S. Stefanik and H. L. Tuller, Journal of Materials Research , 2012, 27, 2009-2016. 202. S. Bishop, K. Duncan and E. Wachsman, Electrochimica Acta , 2009, 54, 1436-1443. 203. S. R. Bishop, K. Duncan and E. Wachsman, Ecs Transactions , 2006, 1, 13-21. 204. D. Chen, S. R. Bishop and H. L. Tuller, Advanced Functional Materials , 2013, 23, 2168-2174. 131 205. J. G. Swallow, J. J. Kim, M. Kabir, J. F. Smith, H. L. Tuller, S. R. Bishop and K. J. Van Vliet, Acta Materialia , 2016, 105, 16-24. 206. J. Sheth, D. Chen, H. L. Tuller, S. T. Misture, S. R. Bishop and B. W. Sheldon, Phys Chem Chem Phys , 2017, 19, 12206 -12220. 207. D. Chen, S. R. Bishop and H. L. Tuller, Chem Mater , 2014, 26, 6622-6627. 208. B. D. Cullity, 2001. 209. A. Patterson, Physical review , 1939, 56, 978. 210. G. Williamson and W. Hall, Acta metallurgica , 1953, 1, 22-31. 211. W. C. Chueh and S. M. Haile, Physical Chemistry Chemical Physics , 2009, 11, 8144-8148. 212. H. Klam, H. Hahn and H. Gleiter, Acta Metallurgica , 1987, 35, 2101-2104. 213. C. Lenser, F. Gunkel, Y. J. Sohn and N. H. Menzler, Solid State Ionics , 2017. 214. R. Waser, T. Bieger and J. Maier, Solid state communications , 1990, 76, 1077-1081. 215. J. H. Yu, J. S. Lee and J. Maier, Angewandte Chemie International Edition , 2007, 46, 8992-8994. 216. J. J. Yang, D. B. Strukov and D. R. Stewart, Nat Nanotechnol , 2013, 8, 13-24. 217. S. R. Bishop, T. S. Stefanik and H. L. Tuller, Physical Chemistry Chemical Physics , 2011, 13, 10165-10173. 218. S. Taniguchi, M. Kadowaki, H. Kawamura, T. Yasuo, Y. Akiyama, Y. Miyake and T. Saitoh, Journal of Power Sources , 1995, 55, 73-79. 219. S. Badwal, R. Deller, K. Foger, Y. Ramprakash and J. Zhang, Solid State Ionics , 1997, 99, 297-310. 220. S. Jiang, J. Zhang, L. Apateanu and K. Foger, Journal of the Electrochemical Society , 2000, 147, 4013-4022. 221. S. Jiang, J. Zhang and K. Foger, Journal of The Electrochemical Society , 2000, 147, 3195-3205. 222. S. P. Jiang and Y. Zhen, Solid State Ionics , 2008, 179, 1459-1464. 223. S. P. Jiang, S. Zhang and Y. Zhen, Journal of The Electrochemical Society , 2006, 153, A127 -A134. 224. S. P. Simner, M. D. Anderson, G. -G. Xia, Z. Yang, L. R. Pederson and J. W. Stevenson, Journal of The Electrochemical Society , 2005, 152 , A740 -A745. 132 225. T. Komatsu, H. Arai, R. Chiba, K. Nozawa, M. Arakawa and K. Sato, Electrochemical and solid -state letters , 2006, 9, A9 -A12. 226. Y. Liu and C. Jiao, Solid State Ionics , 2005, 176, 435-442. 227. A. Hagen, Y. -L. Liu, R. Barfod and P. V. Hendriksen, Journal of the Electrochemical Society , 2008, 155, B1047 -B1052. 228. E. Bucher, C. Gspan, F. H ofer and W. Sitte, Solid State Ionics , 2013, 230, 7-11. 229. E. Bucher, W. Sitte, F. Klauser and E. Bertel, Solid State Ionics , 2011, 191, 61-67. 230. M. Viitanen, R. v. Welzenis, H. Brongersma and F. Van Berkel, Solid State Ionics , 2002, 150, 223-228. 231. E. Bucher, W. Sitte, F. Klauser and E. Bertel, Solid State Ionics , 2012, 208, 43-51. 232. L. Zhao, N. H. Perry, T. Daio, K. Sasaki and S. R. Bishop, Chem Mater , 2015, 27, 3065-3070. 233. Y.-L. Liu, S. Primdahl and M. Mogensen, Solid State Ionics , 2003, 161, 1-10. 234. M. S. Schmidt, K. V. Hansen, K. Norrman and M. Mogensen, Solid State Ionics , 2008, 179, 1436-1441. 235. K. Haga, S. Adachi, Y. Shiratori, K. Itoh and K. Sasaki, Solid State Ionics , 2008, 179, 1427-1431. 236. T. Zhang, J. Ma, S. Chan, P. Hing and J. Kilner, Solid state sciences , 2004, 6, 565-572. 237. N. Savvides, A. Thorley, S. Gnanarajan and A. Katsaros, Thin Solid Films , 2001, 388, 177-182. 238. C. M. Wang, S. Thevuthasan and C. H. Peden, Journal of the Amer ican Ceramic Society , 2003, 86, 363-365. 239. D. Shi, M. Ionescu, J. McKinnon and S. Dou, Physica C: Superconductivity , 2001, 356, 304-310. 240. N. L. Edleman, A. Wang, J. A. Belot, A. W. Metz, J. R. Babcock, A. M. Kawaoka, J. Ni, M. V. Metz, C. J. Flasche nriem and C. L. Stern, Inorganic chemistry , 2002, 41, 5005-5023. 241. G. Smith, Corrosion of materials in fused hydroxides , Oak Ridge National Lab., Tenn., 1956. 242. H. Seidel, L. Csepregi, A. Heuberger and H. Baumgärtel, Journal of the Electrochemical Society , 1990, 137, 3626-3632. 133 243. P. Simons, H. I. Ji, T. C. Davenport and S. M. Haile, Journal of the American Ceramic Society , 2017, 100, 1161-1171. 244. S. Wang, A. Verma, Y. Yang, A. Jacobson and B. Abeles, Solid State Ionics , 2001, 140, 125-133. 245. J. J. Kim, S. R. Bishop, N. J. Thompson and H. L. Tuller, Solid Oxide Fuel Cells 13 (Sofc -Xiii) , 2013, 57, 1979-1984. 246. Q. Yang, T. E. Burye, R. R. Lunt and J. D. Nicholas, Solid State Ionics , 2013, 249-250, 123-128. 247. W. Karim, C. Spreafico, A. Kleibert, J. Gobrecht, J. VandeVondele, Y. Ekinci and J. A. van Bokhoven, Nature , 2017, 541, 68. 248. I. Hughes and T. Hase, Measurements and their Uncertaintie s: A Practical Guide to Modern Error Analysis , Oxford University Press, Oxford, United Kingdom, 2010.