Size and Composition Effects on Ionic Conductivity

MISSING IMAGE

Material Information

Title:
Size and Composition Effects on Ionic Conductivity Doped Ceria Bulk Ceramics and Thin Film
Physical Description:
1 online resource (113 p.)
Language:
english
Creator:
Kasse, Robert M
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Master's ( M.S.)
Degree Grantor:
University of Florida
Degree Disciplines:
Materials Science and Engineering
Committee Chair:
Nino, Juan C
Committee Members:
Jones, Jacob L
Perry, Scott S

Subjects

Subjects / Keywords:
ceria -- sofc
Materials Science and Engineering -- Dissertations, Academic -- UF
Genre:
Materials Science and Engineering thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
Epitaxial Nd0.1Ce0.9O2-dthin films were deposited on Pt bottom electrodes for the first time and the first across-plane ionic conductivity measurements of these films were performed.  Microstructure was investigated using X-ray diffraction, transmission electron microscopy, and atomic force microscopy.  Conductivity measurements were performed using two-point AC electrochemical impedance spectroscopy (EIS).  Such thin film samples allow for the first time the direct measurement of grain ionic conductivity of doped ceria.  Ionic conductivity data indicates inconsistencies in literature may simply be a result of the experimental conditions used in each study. Samples of SmxNdyCe0.9O2-dwere prepared using solid state processing and conductivity measurements were done using EIS from 250-700°C. Activation energy was calculated from Arrhenius plots and values increased as composition shifted from pure Sm+3 to pure Nd+3in agreement with literature and the effective index concept.  The results of this study conclude that SmxNdyCe0.9O2-dconstitutes a versatile system impervious to potential performance degradation due to preferential dopant segregation and redistribution. NdxCe1-xO2-d(0.05=x= 0.55) was prepared using solid state processing and grain ionic conductivity was measured using EIS.  X-ray and neutron diffraction indicate the highly defective fluorite stabilizes into aC-type rare earth oxide structure at compositions above x=0.50.  The conductivity of the compounds follows an Arrhenius behavior showing a maximum conductivity of 0.054(1) S·cm-1for Nd0.15Ce0.85O2-d at 700°C with an associated activation energy of 0.726(5) eV.  Both activation  energy and pre-exponential factor increase significantly during the phase transition from 35-40% Nd, resulting in an effective gradual decrease of conductivity across the phase transition.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility:
by Robert M Kasse.
Thesis:
Thesis (M.S.)--University of Florida, 2013.
Local:
Adviser: Nino, Juan C.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2013-11-30

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2013
System ID:
UFE0045499:00001


This item is only available as the following downloads:


Full Text

PAGE 1

1 SIZE AND COMPOSITION EFFECTS ON IONIC CONDUCTIVITY : DOPED CERIA BULK CERAMICS AND THIN FILM By ROBERT MARTIN KASSE A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUI REMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2013

PAGE 2

2 2013 Robert Martin Kasse

PAGE 3

3 To my future s elf

PAGE 4

4 ACKNOWLEDGMENTS Though it is written by an individual, it would be nearly impossibl e to complete all of the work that goes into a thesis without significant help from others. Whether helping out with experiments, giving you time and space to struggle and figure th ings out on your own, or helping to keep you sane despite spending far t oo much time in the lab, a thesis is very much a group effort. I could simply list the names of everyone who helped me throughout the last few years but I feel there are a few who deserve more recognition that that. First I would like to thank my advisor D r. Juan C. Nino f or giving me the opportunity to feed the curiosity I had for performing scientific research. Throughout my three years with NRG he has trusted me to conduct my experiments in my own way at my own pace while at the same time providing nece ssary g uidance to prevent me from getting too far off track and wasting my time. Had I not been given this opportunity I may have never discovered how much I enjoy (and at the same time hate) conducting research and would not have realized that this is wh at I would like to someday make a career out of. Of course I also need to thank my parents for helpi ng me get to where I am today. Never once have they questioned the decisions I have mad e regarding my education and their love an d support have made thes e past few years easier than they could have been otherwise. There have also been a number of students and professors who have helped me with my research that deserve special mention, and they are as follows: Dr. Kyeong Won Kim and Dr. David Norton (PLD), Dr. Nicholas Rudawski and Dr. Kevin Jones (TEM), Mina Hanna and Dr. Scott Perry (AFM), and Dr. Jacob Jones and various

PAGE 5

5 students in his research group for the use of their lab and equipment. This work could not have been completed without the use of equipm ent located at MAIC and NRF. I also need to give a special thanks to Microsoft Paint. Along the way there have been many times when I needed to ask my fellow group members for help with something related to my project and they were more than willing (most of the time) to provide that help. This list of people includes, but is not limited to: Chris Turner, Trey Davis, Roberto Esquivel, Don Moore, Paul Joh ns, Hyuksu Han, Brittnee Mound George Baure, Mehrad Mehr, Sasm it Gokhale, Satyajit Phadke, and Wei Qiu However, it is the time spent doing everything except working on science that I will remember most when looking back on my time here. Whether it was Ninolympics, Lab B Chteau or ju st general nonsensical conversations about nothing in particular, it was these moments that kept me from going crazy and getting fed up with research.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF FIGURES ................................ ................................ ................................ ........ 10 LIST OF ABBREVIATIONS ................................ ................................ ........................... 13 LIST OF SYMBOLS ................................ ................................ ................................ ...... 15 ABSTRACT ................................ ................................ ................................ ................... 17 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 19 1.1 Statement of Problem and Motivation ................................ ............................... 19 1.2 Scientific Approach ................................ ................................ ........................... 19 1.3 Organization of Thesis ................................ ................................ ...................... 21 1.4 Contributions to the Field ................................ ................................ .................. 21 2 BACKGROUND ................................ ................................ ................................ ...... 23 2.1 Solid Oxide Fuel Cells ................................ ................................ ....................... 23 2.2 Structure of Relevant Materials ................................ ................................ ......... 30 2.2.1 Ceria ................................ ................................ ................................ ........ 30 2.2.2 Sapphire ................................ ................................ ................................ .. 34 2.2.3 Platinum ................................ ................................ ................................ .. 35 2.3 Electrochemical Impedance Spectroscopy ................................ ....................... 35 2.4 Thin Film Deposition ................................ ................................ ......................... 38 2.4.1 Pulsed Laser Deposition ................................ ................................ .......... 38 2.4.2 DC Sputt ering ................................ ................................ .......................... 44 3 EXPERMENTAL PROCEDURES ................................ ................................ ........... 46 3.1 Preparation of Thin Film Samples ................................ ................................ ..... 46 3.1.1 Substrate Cleaning ................................ ................................ .................. 46 3.1.2 Deposition of Highly Oriented Pt Layer Using DC Sputtering .................. 46 3.1.3 Deposition of Highly Oriente d Nd 0.1 Ce 0.9 O 2 Layer Using Pulsed Laser Deposition ................................ ................................ ................................ ..... 47 3.1.4 Deposition of Pt Top Electrodes Using DC Sputtering ............................. 47 3.2 Preparation o f Bulk Samples ................................ ................................ ............ 47 3.2.1 Powder Synthesis ................................ ................................ .................... 48 3.2.2 Formation of Ceramic Green Body ................................ .......................... 48 3.2.3 Pellet Sintering ................................ ................................ ........................ 48 3.3 Sample Characterization ................................ ................................ ................... 49

PAGE 7

7 3.3.1 Impedance Spectroscopy ................................ ................................ ........ 49 3.3.1.1 Thin Film Samples ................................ ................................ ......... 49 3.3.1.2 Bulk Samples ................................ ................................ ................. 49 3.3.2 X ray Diffraction ................................ ................................ ....................... 50 3.3.3 Profilometry ................................ ................................ ............................. 51 3.3.4 Transmission Electron Microscopy ................................ .......................... 51 3.3.5 Atomic Force Microscopy ................................ ................................ ........ 5 2 3.3.6 Scanning Electron Microscopy ................................ ................................ 52 4 HIGHLY ORIENTED Nd 0.1 Ce 0.9 O 2 THIN FILMS ON Pt BOTTOM ELECTRODES ................................ ................................ ................................ ....... 54 4.1 Introduction ................................ ................................ ................................ ....... 54 4.2 Structural Characterization ................................ ................................ ............... 58 4.3 Ionic Conductivity ................................ ................................ .............................. 63 4.4 Summary and Conclusions ................................ ................................ ............... 65 5 IONIC CONDUCTIVITY OF Sm x Nd y Ce 0.9 O 2 ................................ ......................... 66 5.1 Introduction ................................ ................................ ................................ ....... 66 5.2 Phase Analysis and Microstructural Characterization ................................ ....... 67 5.3 Ionic Conductivity ................................ ................................ .............................. 69 5.4 Conclusions ................................ ................................ ................................ ...... 74 6 IONIC CONDUCTIVITY ACROSS THE DISORDER ORDER PHASE TRANSITION IN THE NdO 1. 5 CeO 2 SYSTEM ................................ ........................ 75 6.1 Introduction ................................ ................................ ................................ ....... 75 6.2 Phase and Structural Analysis ................................ ................................ .......... 78 6.2.1 X ray Diffraction ................................ ................................ ....................... 78 6.2.2 Neutron Diffraction ................................ ................................ ................... 81 6.3 Microstructural Analysis ................................ ................................ .................... 85 6.4 Ionic Conductivity ................................ ................................ .............................. 86 6.5 Conclusions ................................ ................................ ................................ ...... 92 7 SUMMARY AND FUTURE WORK ................................ ................................ ......... 93 7.1 Summary ................................ ................................ ................................ .......... 93 7.2 Future Work ................................ ................................ ................................ ...... 94 7.2.1 Thin Films ................................ ................................ ................................ 94 7. 2.2 Bulk ................................ ................................ ................................ ......... 95 APPENDIX A NDC FILMS ON VARIOUS SUBSTRATES ................................ ............................ 97 B FITTING IMPEDANCE DATA USING ZVIEW ................................ ...................... 101

PAGE 8

8 LIST OF REFERENCES ................................ ................................ ............................. 103 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 113

PAGE 9

9 LIST OF TABLES Table page 5 1 Activation energy, pre exponential coefficient, configurational entropy, and effective index values for all of the compositions studied. ................................ .. 73 6 1 Detailed crystallographic information for Nd 0.55 Ce 0.45 O 1.725 obtained by Rietveld refinement. ................................ ................................ ............................ 82 6 2 Grain sizes of the sintered pellets of different compositions. .............................. 86

PAGE 10

10 LIST OF FIGURES Figure page 2 1 Schematic depiction of the operation of a solid oxide fuel cell. ........................... 23 2 2 Large area electrolyte supported YSZ based Fl exCells by Nextech. .............. 24 2 3 Ragone plot comparing the specific power and specific energy of variou s devices to SOFCs. ................................ ................................ .............................. 26 2 4 The log of ionic conductivity versus temperature for various fluorite structured oxides. ................................ ................................ ................................ ................ 27 2 5 Cross s ectional SEM micrographs of the anode supported single cells with GDC electrolyte films ................................ ................................ .......................... 29 2 6 The s chematic s show the c rystal st r u c ture of doped and undoped ceri a .......... 31 2 7 The C type rare earth structure, space group Ia ................................ ............ 32 2 8 Schematic depiction of the space charge region near grain boundaries in doped ceria. ................................ ................................ ................................ ........ 33 2 9 Concentration profile obtained via atom probe analysis of various speci es near a grain boundary in GDC ................................ ................................ ............ 34 2 10 Various view s of the sapphire crystal structu re ................................ .................. 34 2 11 V arious views of the FC C crystal structure of platinum ................................ ..... 35 2 12 Plot showing the relationship be tween impeda nce and phase angle .................. 36 2 13 Generic Nyquist plot showing the grain core, grain boundary, and electrode impedance arcs for a polycrystalline sample. ................................ ..................... 37 2 14 Equivalent circuit used to fit a Nyqu ist plot of complex impedance ..................... 37 2 15 Schematic view of the PLD process. ................................ ................................ .. 39 2 16 Bright white plume generated by the ablation of an Nd 0.1 Ce 0.9 O 2 target by a 248 nm wavelength KrF laser ................................ ................................ ............. 41 2 17 Schematic view of various growth modes ................................ ........................... 42 2 18 Schematic showing the relationship between substrate temperature, ga s pressure, and film morphology ................................ ................................ ........... 43

PAGE 11

11 4 1 Schematic illustration of films with single crystalline, columnar grain, and polycrystalline microstructur es ................................ ................................ ............ 55 4 2 Schematic illustration of electroding configurations for in plane (left) and across plane (right) electrical measurements of thin film samples. .................... 55 4 3 X ray diffraction pattern of a 43 nm thin film sample as compared to the theoretical patterns for Al 2 O 3 Pt, and 10% Nd doped ceria ............................... 59 4 4 X ray diffraction patterns for thin film samples of different thicknesses ranging from 4 190 nm. ................................ ................................ ................................ ... 60 4 5 Convergent beam electron diffraction (left), selected area electron diffraction ( center inset), and high resolution cross sectional TEM (right) ........................... 61 4 6 Atomic force microscopy topography maps and line scans of a Pt layer (top) and a 43 nm N DC film (bottom) ................................ ................................ .......... 62 4 7 Typical Nyquist plot obtained from impedance measurements of thin film samples ................................ ................................ ................................ .............. 64 4 8 Arrhenius plot of grain ionic conductivity for 10% Nd doped ceria ...................... 64 4 9 Plot of the log of conductivity versus temperature ................................ .............. 65 5 1 X ray diffraction patterns for Sm x Nd y Ce 0.9 O 2 ................................ ................... 68 5 2 SEM image of a fracture surface of Sm 0.025 Nd 0.075 Ce 0.9 O 2 showing the typical microstructure of all samples. ................................ ................................ .. 69 5 3 Arr henius plot of grain ionic conductivity for all nine compositions investigated in the Sm x Nd y Ce 0.9 O 2 system ................................ ...................... 73 5 4 Plot of gra in ionic conductivity versus dopant concentration for Sm x Nd y Ce 0.9 O 2 from 400 700C. ................................ ................................ ...... 74 6 1 The ternary phase diagram of NdO 1.5 SmO 1.5 CeO 2 at room tem perature reported by T. Vandera h ................................ ................................ .................... 76 6 2 XRD pattern s of th e calcined powders of Nd x Ce 1 x O 2 ................................ ....... 79 6 3 Lattice parameters of Nd x Ce 1 x O 2 ... 80 6 4 Observed neutron diffraction ................................ 83 6 5 Observed (cross), calculated (continuous line), and difference neutron diffraction profiles for Nd 0.35 Ce 0.65 O 1.825 are shown. ................................ ............ 83

PAGE 12

12 6 6 Observed (cross), calcul ated (continuous line), and difference neutron diffraction profiles for Nd 0.55 Ce 0.45 O 1.725 are shown. ................................ ............ 84 6 7 SEM micrograph for sintered pellets of Nd x Ce 1 x O 2 ................................ ......... 86 6 8 Nyquist plot of impedance for Nd 0.25 Ce 0.75 O 1.875 measured at 250C in air. ...... 88 6 9 Arrhenius plot of grain conductivity of Nd x Ce 1 x O 2 ..................... 88 6 10 Grain conductivity of Nd x Ce 1 x O 2 concentration (x) at temperatures from 450 700C ................................ ............. 89 6 11 Activation energy ( E A ), pre exponential factor ( A 0 ), and shortest cation anion distance as a function of Nd content. ................................ ................................ .. 90 A 1 XRD p attern of a polycrystalline NDC film deposited on a platini zed silicon substrate using PLD ................................ ................................ .......................... 97 A 2 XRD pattern of a polycrystalline Pt la yer deposited on a c plane (0001) sapphire substrate using DC sputtering. ................................ ............................. 98 A 3 XRD pattern of a highly oriented Pt layer deposited on an a plane sapphire substrate using DC sputtering ................................ ................................ ........... 99 A 4 Omega scan of a highly oriented NDC film with a FHWM = 0.86. ................... 100 A 5 The pole figure of an NDC film ................................ ................................ ........ 100 B 1 Equiva lent circuit fitting window in ZView. ................................ ........................ 101 B 2 Fit resulting from equivalent circuit analysis performed using ZView. ............... 102

PAGE 13

13 LIST OF ABBREVIATIONS AC Alternating cu rrent AFM A tomic force microscopy ALD A tomic layer deposition CBED Convergent beam electron diffraction CPE Constant phase element CTE Coefficient of thermal expansion DC Direct current EIS Electrochemical impedance spectroscopy FCV Fuel cell vehicle FWHM Full width at half max FIB Focused ion beam GDC Gadolinium doped ceria HR XTEM High resolution cross sectional transmission electron microscopy IT Intermediate temperature LOI Loss on ignition MBE Molecular beam epitaxy NDC Neodymium doped ceria PLD Pulsed laser deposition PVA Polyvinyl alcohol PVD Physical vapor deposition R Resistor RMS Root mean square SAED Selected area electron diffraction SCL Space charge layer

PAGE 14

14 SEM Scanning electron microscopy SOFC S olid oxide fuel cell TCE Trichloroethylene TEM Trans mission electron microscopy XRD X ray diffraction YSZ Y ttria stabilized zirconia

PAGE 15

15 LIST OF SYMBOLS Ionic conductivity g c Grain ionic conductivity Oxygen vacancy stoichiometry Phase angle a Lattice parameter A 0 Pre exponential coefficient A Sample area C s Spherical aberration coefficient E A Activation energy E m Activation energy of migration E ass. Activation energy of association k Boltzmann constant l Sample thickness Ln Lanthanide N O Number of oxygen sites per unit volume q v Charge of an oxygen vacancy r o Effective oxygen radius r c Average cation radius r d Radius of dopant cation r h Radius of hos t cation R Ideal gas constant S Configurational entropy m Entropy change during oxygen diffusion T Temperature

PAGE 16

16 T m Melting temperature v 0 Jump frequency [V O ] Concentration of o xygen vacanc ies x Dopant fraction Real part of complex impedance Imaginary part of complex impedance

PAGE 17

17 Abstract of Thesis Prese nted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science SIZE AND COMPOSITION EFFECTS ON IONIC CONDUCTIVITY: DOPED CERIA BULK CERAMICS AND THIN FILM By Robert Martin Kasse May 2013 Chair: Juan C. Nino Major: Materials Science and Engineering Highly oriented Nd 0.1 Ce 0.9 O 2 thin films were deposited on Pt bottom electrodes for the first time and the first across plane ionic conductivity measurements of these films were performed. Microstructure was investigated using X ray diffraction transmission electron mic roscopy a nd atomic force microscopy Conductivity measurements were performed using two p oint AC electrochemical impedance spectroscopy (EIS) Such thin film samples allow for the first time the direct measurement of grain ionic conductivity of doped ceria. I oni c conductivity data indicates inconsistencies in literature may simply be a result of the experimental conditions used in each study Samples of Sm x Nd y Ce 0.9 O 2 were prepared using solid state processing and conductivity measurements were done using EIS from 250 700C. Activation energy was calculated from Arrhenius plots and values increase d as composition shift ed from pure Sm +3 to pure Nd +3 in a greement wit h literature and the effective index concept T he results of this study conclude that Sm x Nd y Ce 0.9 O 2 constitutes a versatile system

PAGE 18

18 impervious to potential performance degradation due to preferential dopant segregation and redistribution. Nd x Ce 1 x O 2 (0 x was prepared using solid state processing and grain ionic conductivity wa s measured using EIS X ray and neutron diffraction indicate the highly defective fluorite stabilizes into a C type rare earth oxide structure at compositions above x=0. 50 The conductivity of the compounds follows an Arrhenius behavior showing a maximum conductivity of 0.054(1) Scm 1 for Nd 0.15 Ce 0.85 O 2 at 700 C with an associated activation energy of 0.7 26 (5) eV Both a ctivation energy and pre exponential factor inc rease significantly during the phase transition from 35 40% Nd resulting in a n effective gradual decrease of conductivity across the phase transition.

PAGE 19

19 CHAPTER 1 INTRODUCTION 1.1 Statement of Problem and Motivation Growing worldwide energy demand along w ith a push to reduce greenhouse gas production, which contributes to global warming, has scientist s looking for new sources of clean and renewable energy. Possible solutions include wind, solar, nuclear, and fuel cell technology though a combination of al l of these is likely necessary. Solid oxide fuel cells (SOFCs) have received increased attention in recent years as an environmentally friendly power source, particularly those which operate in the intermediate temperature (IT) range of 400 800C due to their flexibility in design, wider range of useable fuels, and potential uses in distributed and portable power. 1 2 However, current SOFCs do not provide sufficient power at these intermediate temperatures and must operate at or above 1000C which is far too high for use in technologies such as fuel cell vehicles (FCVs) Higher operating temperatures also require higher cost fuel cell components. A key component of an effic ient IT SOFC is a solid oxide electrolyte with a high ionic conductivity at intermediate temperatures. Many investigations have focused on rare earth doped ceria as a potential candidate material for IT applic ations but there is still much work to be done in understanding the effects of composition and microstructure on the ionic conductivity of these materials. 3 1.2 Scientific Approach In recent years the majority of investigations concerning SOFC electrolytes have focused on doped ceria systems. The main objective of this thesis is to determine the effects of size and different dopant strategies on the grain ionic conductivity of rare earth doped ceria systems.

PAGE 20

20 A clear understanding of the effects of grain size on the ionic conductivity of doped ceria is currently unavailable. No studies in the current literature have been able to isolate and direct ly measure either grain core or grain bou ndary conductivity. Thin film samples free of grain boundaries will be synthesized and the ionic conductivity will be measured as a function of film thickness in order to directly determine the effect of size on gr ain ionic conductivity. Uniformly co doped samples in the Sm 2 O 3 Nd 2 O 3 CeO 2 system, with a dopant ratio of 1:1, have been investigated but no studies have reported the effects of changing the dopant ratio at constant total dopant content on grain ionic cond uctivity. A set of samples in the Sm x Nd y Ce 0.9 O 2 system will be synthesized and their electrical properties measured such tha t the effect of co dopant ratio on ionic conductivity can be determined. An anion disorder order phase transition as a function of dopant concentration has been observed in the Nd 2 O 3 CeO 2 system and while a deleterious effect on conductivity is expected, this has not been corroborated experimentally. Samples with Nd dopant concentrations from 0.05 0.55 will be synthesized such that the effect of composition on structure and ion ic conductivity can be determined. Various characterization techniques will be used to complete the stated research goals. Ionic conductivity will be measured using two point AC electrochemical impedance spectroscopy. Microstructure will be investigated using scanning electron microscopy, transmission electron microscopy, profilometry, and atomic force microscopy. Phase purity will be determined using X ray diffraction.

PAGE 21

21 1.3 Organization of Thesis Chapter 2 provides background information necessary for be tter understanding the work presented in subsequent chapters Information regarding solid oxide fuel cells, pulsed laser deposition, defect reactions and conduction mechanisms in doped ceria, and electrochemical impedance spectroscopy is provided. Chapte r 3 discusses the experimental procedures, including sample preparation and materials characterization techniques, used in the execution of the research Chapter 4 discusses work investigating highly oriented thin films of doped ceria deposited on Pt botto m electrodes and presents the structural characterization and across plane ionic conductivity results. Chapter 5 reports the effects of co doping ceria with uneven amounts of Sm and Nd and discusses the compositional flexibility of the Sm x Nd y Ce 0.9 O 2 system Chapter 6 details how activation energy, pre exponential factor, and grain ionic conductivity change across the disorder order phase transition in the NdO 1.5 CeO 2 system. Chapter 7 summarizes the work detailed in the preceding chapters and out lines ideas for future investigations of both bulk and thin film doped ceria electrolyte materials. 1.4 Contributions to the Field This work investigates the size and compositional effects on the ionic conductivity of rare earth doped ceria electrolytes. The main contributions of this work to the field of materials science are as summarized below: 1. Highly oriented doped ceria thin films were successfully deposited on Pt bottom electrodes for the first time.

PAGE 22

22 2. T he first ever across plane ionic conductivity me asurements of highly oriented doped ceria thin films were performed, which directly measures grain ionic conductivity. Measurements were performed as a function of film thickness such that size effects on the grain ionic conductivity could be determined. 3. The ionic conductivity of Sm x Nd y Ce 0.9 O 2 ( x + y = 0.10, with x = 0, 0.0125, to determine the effect of different codoping ratios on ionic conductivity and it was found that the system is very compositionally flexible with regards to sinterability a nd ionic conductivity which follow s the theory of effective index. 4. The effect of the disorder order phase transition in the NdO 1.5 CeO 2 system on grain ionic conductivity was investigated for the first time. At room temperature, XRD and neutron diffractio n determined that this transition occurs at ~50% Nd while conductivity measurements indicate that the transition occurs at ~40% Nd at higher temperatures. Both the activation energy and pre exponential factor increase sharply at this transition causing io nic conductivity to decrease gradually.

PAGE 23

23 CHAPTER 2 BACKGROUND 2.1 Solid Oxide Fuel Cells Fuel cells have received renewed attention in recent decades due to their ability to provide a clean source of energy and operate with a high efficiency. There are man y different types of fuel cells, with different electrolyte materials, different charge carrying mechanisms, and different operating temperatures A schematic of a solid oxide fuel cell (SOFC) can be seen in Figure 2 1 with anode and cathode reactions shown in Equations 2 1 and 2 2. Anode Reaction: (2 1) Cathode Reaction: (2 2) Figure 2 1 Schematic depiction of the opera tion of a solid oxide fuel cell. 4

PAGE 24

24 SOFCs directly convert chemical energy gene rated by the reduction of a gaseous fuel such as H 2 to electricity. Hydrogen gas, or whatever the source of hydrogen in the fuel may be (e.g. CH 4 ) is broken into protons and electrons at the anode. Since the solid oxide electrolyte is an electronic insul ator, the electrons cannot pass through it and are forced through an external circuit, generating a useful current. The electrons react with oxygen upon reaching the cathode to form oxygen ions which then travel through the electrolyte and recombine with protons at the anode to produce water in the form of steam, completing the circuit. Current state of the art SOFCs are generally used at temperatures around 1000C to increase fuel conversion efficiency. A commercial electrolyte supported large area YSZ based FlexCell produced by Nextech is shown in Figure 2 2 below. Figure 2 2 Large area electrolyte supported YSZ based FlexCells by Nextech. FlexCells with electrolytes less than 40 m thick can be manufactured with areas up to 470 cm 2 SOFCs in current applications typically have yttria stabilized zi rconia (YSZ) electrolytes O ther materials currently investigated for use as electrolytes include other fluorite based oxides such as doped Bi 2 O 3 and CeO 2 perovskites such as lanthanum gallate, and La 2 Mo 2 O 9 Based on the conductivity of YSZ and the typi cal thickness (~

PAGE 25

25 ) of these electrolytes when employed in SOFCs, the minimum operating temperature is around 700C. 6 Electrodes, both anodes and cathodes, must be chemically and thermally compatible wi th the electrolyte they are used with and must be stable at the operating temperature. 6 These electrodes are porous composites of a metal or electronically conducting oxide and an electroly te material, with a large proportion of triple phase boundaries between metal, electrolyte, and gas to improve fuel conversion Porous Pt used to be commonly used in electrodes but because of its very high cost, they are now typically made with a Ni/ZrO 2 cermet (anodes) and doped LaMnO 3 (cathodes). 7 Good thermal expansion coef ficient (CTE) matching is also important as cells will be thermally cycled between room and operating temperature, which could build up thermal stresses if the CTE mismatch is too large. Interconnects are metal or ceramic plates which separate individual cells in a fuel cell stack and connect them together in series to add together the power generated by each individual cell. High temperature SOFCs use interconnects made of doped LaCrO 3 or expensive high temperature metal alloys, though vaporization of Cr at high temperatures is an issue that has taken attention away from metal alloy interconnects. Interconnects also need to have a good thermal expansion match with other cell components for the reasons stated above. 8 Unlike heat engines which operate more efficiently as temper ature is increased, the oxidation and reduction reactions in SOFCs actually increase in efficiency when the temperature is lowered. 7 A Ragone plot comparing various energy devices is shown in Figure 2 3. Lowering the operating temperature to the intermediate (IT) temperature range (400 800C) also opens up many more options fo r fuel cell stack materials

PAGE 26

26 selection, which is restricted at higher temperatures. Lowering operating temperatures has the potential to lower production costs in multiple ways, including the fact that at lower temperatures interconnects can be made out of less expensive stainless steel as opposed to the ceramic interconnects used in high temperature SOFCs because stainless steel has a good CTE match with other cell components. A lower power density is generated at lower temperatures, but the lifetime of t he fuel cell is extended and the overall cost of the fuel cell is reduced. One of the major necessary steps towards the expansion of SOFCs into automotive applications is to increase the power density at lower temperatures to the levels achieved with curr ent high temperature SOFCs One of the foc i of research in this area is on developing electrolytes with higher ionic conductivities in the IT range. Figure 2 3 Ragone plot comparing the specific power and specific energy of var ious devices to SOFCs. 2 While a high ionic conductivity is vitally important for SOFC electrolytes high thermal and chemical stability at both room and operating temperatures, the ability to be

PAGE 27

27 thermally cycled without performance degradation, a nd a good thermal expansion match with other cell components are also necessary C ubic fluorite structured oxides are receiving a lot of attention as a pote ntial candidate material for IT SOFC electrolyte application because of their performance in the se areas. Figure 2 4 2 4 shows an Arrhenius plot of the conductivity of various fluorite structured oxides as a function of temperature. The slope of the line on this graph is related to the activation energy for ion ic conduction, and the y intercept is related to the pre exponential coefficient. Figure 2 4 The log of ionic conductivity versus temperature for various fluorite structured oxides. 9 As seen in Figure 2 4 2 4 Bi 2 O 3 has the highest ionic conductivity at intermediate temperatures, however is not st able, as it undergoes a phase transformation Bi 2 O 3 between room and operating temperatures which is not ionically conducting Doped Bi 2 O 3 is stable over a wider temperature range but still poses issues such as poor mechanical strength, volatilizatio n of Bi 2 O 3 and high

PAGE 28

28 corrosive activity. 10 The material with the next highest ionic conductivity, and one that is the focus of many investigations 9 11 25 is rare earth doped CeO 2 Doped ceria is a good IT SOFC electrolyte candidate material because of its high ionic conductivity, low electronic conductivity, good stability, and good thermal expansion coefficient match with stainless steel, which can be used as an in expensive interconnect material as noted above. A large portion of the research being performed on the ionic conductivity of SOFC electrolytes is done by measuring pol ycrystalline bulk ceramic pellets. However, the overall performance of these same materials in thin film form is increased because as electrolyte thickness is decreased, so are the ohmic losses associated with ionic conduction since most ohmic resistance comes from the slow rate of ion migration in the electrolyte relative to electron conduction 26 However, the fi lm still must be thick enough to prevent fuel cross over thus an idea l electrolyte should be around 10 depending on the material, morphology, and how it is deposited. 27 To better understand the effect of electrolyte thickness on SOFC performance, many investigations have focused on thin film SOFCs. 28 From a practical standpoint it m without cracks of pinholes that woul d allow gas crossover. Ding et al. investigated SOFCs with varying electrolyte thicknesses, seen in Figure 2 5 below, and found that for GDC deposited using dry co pressing and spray dry co m produced the greatest e lectrical performance. 28 Thin films can also be used as ultrathin protective layers between electrodes and electrolytes to help prevent the formation of undesired pha ses which

PAGE 29

29 degrade cell performance, such as a layer of GDC between a YSZ electrolyte and a lanthanum strontium magnate cathode. 27 Figure 2 5 Cross sectional SEM micrographs of the an ode supported single cells with GDC electrolyte films of (a) 1 m, (b) 4 m, (c) 16 m and (d) 75 m. 28 Smaller SOFC stacks that could be us ed to power individual buildings in remote locations or in military operations require stacks that operate at intermediate temperatures. IT SOFCs could also possibly be used in portable power generators with similar power generating capabilities as those based on heat engines used today. Though f uel cell vehicles (FCVs) are not thought to be the most promising clean energy technology in the area of personal vehicles and other transportation over the long run, their ability to use any hydrocarbon as fuel which have a much greater energy density than H 2 29 makes them an attractive transition technology 2 30 For this to happen, fuel cells will necessarily need to operate at lower temperatures and be smaller and more

PAGE 30

30 compact. Using pure hydrogen gas as fuel may be more environmentally friendly considering the o nly waste product would be water, but SOFCs using hydrocarbon fuels would allow commercialization to happen much sooner. High temperature SOFCs, when coupled with other fuel cell types or gas turbines can attain even higher efficiencies for large scale ene rgy production. 1 For movement into the area of smaller stacks and portable power, bett er cell materials need to be developed for operation in the intermediate temperature range. Progress has been made in developing better electrolytes, more efficient electrodes, and less expensive interconnects but further improvements are needed before wi despread commercialization is viable. The future of this technology is promising and one day SOFCs may be a part of a larger clean energy infrastructure 2.2 Structure of Relevant Materials As noted in s ection 2.1, a key component of an IT SOFC is an elec trolyte with a high ionic conductivity in the IT range. Being familiar with the structure of ceria is necessary for a complete understanding of the approaches taken to enhance ionic conductivity. This section presents information on the structure of ceri a and other materials relevant to the work in this thesis. 2.2.1 Ceria Undoped ceria has a fluorite structure, space group Fm m and lattice parameter a = 5.4114 31 The lowest energy surface in ceria is the (111). In doped ceria, ionic conductivity is the result of an oxygen vacancy diffusion mechanism. Doping introduces oxygen vacancies into the lattice as shown in Equation 2 3 and Figure 2 6 below ( 2 3 )

PAGE 31

31 Figure 2 6. The s chematic s show the c rystal st r u c ture of doped and undoped ceria. Pure c eria (left) has a fluorite crystal structure with space group Fm m. When doped with rare earth elements the dopant cation Ln 3+ substitutes for Ce 4+ and in order to maintain charge neutrality, oxygen vacancies are formed (right). In the figure, yellow spheres represent Ce 4+ ions, red spheres represent oxygen ions, blue spheres represent Ln 3+ cations, and the white square represents an oxygen vacancy. When heavily doped with Ln 3+ cations a phase transition occ urs as discussed in Chapter 6. The oxygen deficient fluorite structure becomes a C type rare earth structure space group Ia seen in Figure 2 7 below. The general chemical formula for materials with this structure is typically A 2 O 3 It can be imagined as a fluorite cell with two oxygen vacancies along the body diagonal. Eight of these defect fluorite cells compose a C type unit cell, thus a ssociated with the phase transition is a doubling of the lattice parameter For example, t he lattice parameter of Nd 0.55 Ce 0.45 O 2 is a = 10.998 The X ray pattern of the C type structure is only slightly different from that of the fluorite in that it features extra super lattice peaks, which are very weak.

PAGE 32

32 Figure 2 7. The C type rare e arth structure, space group Ia The image on the right shows a view along the [001] direction. Yellow spheres can be Ce 4+ or Ln 3+ ions and the r ed spheres represent oxygen ion sites Before oxygen vacancy diffusion can occur a mini mum activation energy, E a must be provided which is composed of both association ( E ass. ) and migration ( E m ) enthalpy such that E a = E m + E ass E ass. is the energy need to overcome the association of defect clusters, which form when positively charged oxy gen vacancies and negatively charged dopant cations are attracted to one another. E m is the enthalpy related to the motion of oxygen ions in the material. Equation 2 4 shows the Arrhenius behavior of the ionic conductivity and that the conductivity is no t only dependant on activation energy but also a pre exponential term, A 0 This term is expanded in Equation 2 5 where q V is the charge of an oxygen vacancy, k is the Boltzmann constant, [V O ] is the fraction of free oxygen vacancies, N O is the number of oxygen sites per unit volume, a is the ion jump distance, 0 is the jumping frequency, and m is the entropy change during oxygen diffusion. To a first approximation, all of these terms, except [V O ] are

PAGE 33

33 either constant or independent of total dopant concentration (within the dilute regime) so maximizing the number of mobile oxygen vacancies by using better doping schemes is the ideal way to increase the pre exponential. ( 2 4 ) ( 2 5 ) It is well understood that in microcrystalline doped ceria space charge regions near grain boundaries (depicted schematically in Figure 2 8) block the conduction of oxygen ions and reduce total ionic conductivity. 32 Grain boundaries have positively charged cores due to an excess n umber of oxygen vacancies. To maintain charge neutrality, negatively charged dopant cations segregate to the grain boundaries creating the space charge layer as seen in Figure 2 9 14 Figure 2 8. Schematic depiction of the space charge region near grain boundaries in doped ceria. 32

PAGE 34

34 Figure 2 9. Concentration profile obtained via atom probe analysis of various species near a grain boundary in GDC. 14 2.2.2 Sapphire Sapphire, single crystal Al 2 O 3 has a hexagonal unit cell and is of the space group R c and has lattice parameters a = 4.735 and c = 12.899 33 Various views of the crystal structure can be seen in Figure 2 10 below. Unit cell (0001) (11 0) Figure 2 10 Various view s of the sapphire crystal structure. The cent er image shows the (0001) plane, also referred to as the c plane while the image on the right shows the (11 0) plane, also referred to as the a plane.

PAGE 35

35 2.2.3 Platinum Platinum has an FCC crystal structure, space group Fm m and lattice parameter a = 3.9231 34 Various views of the crystal structure can be seen in Figure 2 11 below including a view of the lowest energy surface in Pt which is the (111) Figure 2 11 Various views of the FCC crystal structure of platinum. The image on the right is a view along the [111] direction. 2.3 Electrochemical Impedance Spectroscopy The most important property of SOFC electrolytes is the ionic conductivity. One method commonly used to measure the ionic conductivity of oxide ceramics is two point alternating current (AC) electrochemical impedance spectroscopy (EIS). EIS is ideal for these types of measurements due to its ability to distinguish between grain, grain boundary, and electrode impedance. In a typical experiment a voltage is applied to an electroded sampled and the corresponding current is measured (or vice versa) as a function of frequency, usually in the range of 10 1 10 7 Hz. 35 The magnitude of the i ), the difference in phase between the current and voltage, which is zero for a pure resistor,

PAGE 36

36 90 for a pure capacitor, and +90 for a pure inductor 36 parts of the complex impedance are related to the phase angle as shown in Figure 2 1 2 below and can be calculated using Equatio ns 2 6 and 2 7 Figure 2 1 2 Plot showing the relationship between impedance and phase angle. 35 (2 6 ) (2 7 ) When the real and the negative of the imaginary part s of the complex imp edance are plotted against one another the result is called a Nyquist plot, seen in Figure 2 13 below. For polycrystalline samples three distinct semicircular arcs are present representing the grain, grain boundary, and electrode impedance, respectively The frequency at the top of each arc corresponds to the characteristic frequency which is related to the time constant for that conduction process.

PAGE 37

37 Figure 2 1 3 Generic Nyquist plot showing the grain core, grain boundary, and electrode impedance arcs for a polycrystalline sample. An equivalent circuit analysis can be used to extract impedance from a Nyquist plot. One such circuit can be seen in Figure 2 1 4 where an inductor is used to model the experimental setup and a parallel resistor constant pha se element (CPE) element is used to model the grain, grain boundary and electrode A CPE is equivalent to a distribution of capacitors in parallel with a phase angle somewhat less than 90 Figure 2 1 4 Equivalent circuit used to fit a Nyquist plot of complex impedance. L1 is an inductor modeling th e experimental setup while the R CPE elements model grain core, grain boundary and electrode impedance. In this thesis the primary focus is on the grain core ionic conductivity which corresponds to the fir st arc of the impedance spectrum. The resistance value obtained from the equivalent circuit fit can simply be plugged in to Equation 2 6 below, along with

PAGE 38

38 the sample dimensions, to calculate the grain ionic conductivity. R GC is the grain core resistivity l is the sample thickness, and A is the sample area. A similar calculation can be done to calculate the grain boundary and total conductivity. (2 6) When grain size decreases into the nano range the grain core and grain boundary arcs can begin to become indistinguishable and very difficult to fit accurately Thus, techniques different from those discussed above must be used to determine the independent contributions. Due to connectivity or shape problems (i.e. grains are not cub es), brick layer models tend to fail when grain size and grain boundary width become comparable. 36 N ewer models such as the nano grain composite model can be used t o extract grain core and grain boundary impedance and dielectric constants Another solution to this problem is to work with thin film samples which allow the measurement of fundamental defect and interfacial properties that are not possible in bulk sampl es. 37 2.4 Thin Film Deposition Two different thin film deposi tion techniques were employed in this work: pulsed laser deposition (PLD) and DC sputtering. This section will describe these processes. 2.4.1 Pulsed Laser Deposition Micro SOFCs employ thin film electrode and electrolyte layers, as discussed in section 2 .1 above Synthesis of materials in the form thin films has been the topic of research for many years. 38 Films on the o rder of only tens or hundreds of nanometers in thickness allow for the measurement of very small scale and fundamental materials properties which cannot be investigated using bulk samples. 39

PAGE 39

39 There are numerous thin film deposition techniques including dip coating, spin coating, DC sputtering, atomic layer deposition (ALD), molecular beam epitaxy (MBE), and pulsed laser deposition (PLD). One of the main disadvantages associated with many of these techniques is that it is difficult to deposit f ilms of materials with complex crystal structures such as superconducting oxides and multilayer materials while maintaining a reasonably high deposition rate Materials with many different cations require many different precursors, precise control of th e relative amounts of precursors, and if the material is an oxide, control of the oxidizing environment for these other processes. PLD is a relatively simple solution to many of these problems along with having other beneficial characteristics. 39 40 Pulsed laser deposition w as first demonstrated to be a useful technique for thin film deposition in 1965. 41 However, this process did not really take off until the 1980s after it had been proven to be a reliable, relatively inexpensive way of producing high quality films with the same stoichiometry as the target material. 42 43 PLD is particularly popular in depositing a wide range of superconducting oxide materials, other electroceramics, and polymers. 39 44 45 Figure 2 1 5 Schematic view of the PLD process.

PAGE 40

40 Conceptually the PLD process is quite easy to understand. 43 A bulk target of the desired film material is ablated with a pu lsed laser in a high vacuum environment creating a plasma plume of the target material The plasma deposit s on a substrate that is typically heated. A schematic of the PLD process is seen in Figure 2 1 5 above. Th e laser is pulsed rather than continuous to increase energy efficiency such that energy is not lost thermally or otherwise. 43 Currently, nanosecond pulses are the most widely used though they are inefficient compared to femtosecond pulse lengths which are garnering a lot of attention. 40 For the deposition of insulating materials, lasers with wavelengths in the 200 400 nm range are most common because photons of this energy are strongly absorbed by insulators. 46 Two of the most common laser types are KrF and XeCl. 47 When the laser hits the target, absorption of photons causes the target to both melt and ablate; surface temperatures can reach as high as 3200 K. 48 Ions, atoms, molecules, and other neutral particles are ejected from the target creating a plume. 48 If the laser passes through the plume it can further ionize particles. The plume then expands adiabatically as thermal energy is converted to kinetic energy. 48 There maybe be a noise associated with the plume expansion as some particles could be travelling faster than the speed of sound. 43 To avoid uneven wear of the target and the formation of grooves or pits, the target is usually rotated and the laser moved back and forth across the target surface. 44 49 This also helps to avoid the formation of droplets of the material being transferred to the substrate. How a target will respond depends on the characteristics of the target as well as the laser. This response is dominated by

PAGE 41

41 elec tronic properties of the target for pulses of less than a picosecond but is dominated by thermal factors when the pulses are on the order of nanoseconds. 42 The plume that can initially be as hot as 10,000 K, may b e a bright white color due to free free transitions within the plasma, as seen in Figure 2 1 6 below 4 2 47 It is interesting to note that t he direction of the plume is always normal to the surface of the target regardless of the incident angle of the laser and is unaffected by the laser if it were to pa ss through the plume. 43 That being said, unless an advanced system is used to raster the plume across the substrate surface the thickness distribution of the deposited films is rather non uniform thus PLD is generally only used to deposit over small areas. 50 The plume is also aff ected by any background gas, such as oxygen, that is introduced into the vacuum chamber. Gas may be used to be incorporated into the film, slow the velocity of the particles in the plume, confine the size of the plume, or decrease the energy of the partic les to limit sputtering away of the growing film. 47 Controlling the average energy of a particle as it strikes a substrate is critical for producing films of certain a stoichiometry and microstructure. Figure 2 1 6 Bright white plume generated by the ablation of an Nd 0.1 Ce 0.9 O 2 target by a 248 nm wavelength KrF laser.

PAGE 42

42 Films deposited using PLD generally have the same structure and composition as the target provided there is no preferential ablation, segregated phases on the target surface, or evaporation of only some of the target components. 51 Other conditions that could cause films to not be stoichiometric include plume species having different sticking coefficients or resputtering of the already deposited film. 52 There are three basic ways in which a film can form, as shown schematically in Figure 2 1 7 : island growth (Volmer Weber ) layer growth (Frank van der Merwe) or a combination of the two known as Stranski Krastanov. 53 Which method of formation occurs is determined by the how particle particle interactions compare to particle substrate interactions. Island growth occurs when particle particle bonds are preferred while layer growth is the result of particle substrate interactions being favorable. Stranski Krastanov is not well understood but occurs when metals are deposited on metal substrates. 53 Figure 2 1 7 Schematic view of various growth mod es : a) island growth (Volmer Weber); b) layer growth (Frank van der Merwe); and c) Stranski Krastanov. 54

PAGE 43

43 Figure 2 18. Schematic showing the relationship between substrate temperature, gas pressure, and film morphology. 55 Films deposited at room temperature are generally amorphous with the crystallinity of the films increasing with increasing substrate temperature. 42 Structure zone models relate film microstr ucture to deposition conditions, such as the one developed my Thornton et al. which relates film morphology to gas pressure and the ratio of substrate temperature to film melting temperature. 55 A schematic of this model

PAGE 44

44 can be seen in Figure 2 1 8 above Morphologies like those shown in Zone 1 result from a lack of surface diffusion ca using atoms to stick where they first hit the substrate yielding columnar structures. Zone T is characterized as having a fibrous structure similar to what is found within an indi vidual column of Zone 1 morphology. Both Zone 1 and Zone T structures are c onsiderably affect by gas pressure. 55 Zone 2 structures occur above about 0.3 T m where T m is the melting point, and can form faceted needles or platelets. These structures are not affected by gas pressure and are the result of adatom diffusion. Zone 3 structures are due to bulk diffusion, such as recrystallization and grain growth, a nd occur at around 0.5 T m Grains may be columnar or equiaxed and this growth is not affected by gas pressure. 55 2.4.2 DC Sputtering Figure 2 19. Schematic representation of the DC sputtering process. 56 During the DC sputtering process, a target (cathode) is bombarded with ions causi ng the formation of a glowing plasma which is deposited on a substrate (anode). Secondary electrons are also produced by the collision and they interact with the ejected particles causing further ionization. A magnetic field is used to trap secondary

PAGE 45

45 ele ctrons near the target surface increasing the probability of particle ionization events. 56 The increased ionization efficiency produces a dense plasma which leads to higher deposition rates than would be possible without the use of a magnetic field and allows for lower operating pressures and voltages. 56 Figure 2 1 9 schematically shows the DC sputtering process.

PAGE 46

46 CHAPTER 3 EXPERMENTAL PROCEDURES This chapter will cover the synthesis of both thin film and bulk ceria samples, discussing starting materials, processing steps, and heat treatments, as well as the characterization techniques used to investigate sample properties. 3.1 Preparation of Thin Film Samples Highly oriented doped ceria thin film samples on Pt bottom electrodes were synthesized using two different deposition techniques: DC sputtering and pulsed laser deposition (PLD). This section will outline the synthesis process in detail. 3.1.1 Sub strate Cleaning Single crystal a plane sapphire ( section 2.2.2) substrates purchased from MTI Corporation were cleaned using ultrasonication in trichloroethylene (TCE) acetone, and methanol for five minutes each and dried using a nitrogen gun. Substrate s measuring 1cm x 1cm or 5 mm x 5 mm were used. 3.1.2 Deposition of Highly Oriented Pt Layer Using DC Sputtering DC sputtering ( section 2.4.2) was used to deposit a 50 nm thick Pt layer on a clean sapphire substrate following Nefedov et al 57 First the substrate was loaded into the deposition chamber (~10 8 Torr) and annealed at 500C for ten minutes T he temperature was then dropped to 300C for the Pt deposition step which was done under an atmosphere of 5 mTorr Ar Pt was deposited for 90 seconds but growth rate depends on the gun and target used and must be calibrated. Heating and cooling rates wer e 10C/min

PAGE 47

47 3.1.3 Deposition of Highly Oriented Nd 0.1 Ce 0.9 O 2 Layer Using Pulsed Laser Deposition The platinized substrates were again clean as described above before being adhered to a substrate holder using Ag paste. PLD ( section 2.4.1) was used to dep osit the Nd 0.1 Ce 0.9 O 2 layer at 720C following Gobel et al. 37 under an atmosphere of 10 mTorr O 2, laser energy of 320 mJ, and a repetition rate of 5 Hz. A KrF laser with a wavelength of 248 nm and a pulse width of 30 ns was used O ne edge of the platinized substrate was masked using a second substrate to ensure a strip of the Pt bottom electrode remained exposed such that it could be used as bottom contact for electrical measurements Oxygen was left in the chamber until the substrate had cooled back down to room temperature. The target used during PLD was syn thesized using solid state processing as detailed in section 3.2 Preparation of Bulk Samples below. 3.1.4 Deposition of Pt Top Electrodes Using DC Sputtering Pt to p electrodes 50 nm thick were deposited using DC sputtering with a shadow mask at room tempe rature. The shadow masks were purchased from Benchmark diameter holes in a hexagonal lattice pattern with the Samples were affixed to the holder using double sided Kapton tape. To attac h the mask, two substrates the same thickness as the sample were affixed to the holder on either side of the sample. The mask was then taped to these substrates to ensure it would lay flat and not bend. 3.2 Preparation of Bulk Samples Bulk samples were sy nthesized using conventional solid state processing. This section will explain the processes used in detail.

PAGE 48

48 3.2.1 Powder Synthesis Staring materials for bulk samples were s toichiometric amounts of oxide powders with 99.99% purity (Alfa Aesar). Stoichiom etric amounts were determined by taking loss on ignition (LOI) data into account. Purchased powders were weighed before and after heating to 800C for 1 h, allowing water and other surface impurities to volatilize to better determine the percentage of ef fective powder. Starting powders were wet ball milled in water for 24 h, using ammonium polyacrylate ( DARVAN 821 A ) as a mesh before calcining at 1450C for 10 h. Calcined powders were again wet ball milled for 24 h, dried at 120C for 16 h, and ground and sieved using a mesh with a aperture opening. 3.2.2 Formation of Ceramic Green Body Up to 2 wt% binder ( polyvinyl alcohol in deionized water) was added to the calcined powder which was then uniaxially pressed at 180 MPa using dies with diameters of 7, 9, and 13 mm. The pellet s were then placed into balloons, which were evacuated using a roughing vacuum pump, and isostatically pressed at 250 MPa for 3 min. 3.2.3 Pellet Sintering Pellets were sin tered at 1600C for 10 h in air. Sacrificial powder of the same composition was pla ced between the pellet and the zirconia setter to prevent any reaction. Heating and cooling rates were 200 C /hour. Pellets were held at 400 C for 1 hour to allow for binder burn out. Densi ty was calculated geometrically and all pellets achieved a relati ve density of at least 95% of theoretical.

PAGE 49

49 3.3 Sample Characterization Microstructural and electrical properties of bulk and thin film samples were characterized using various methods. This section wil l d etail the techniques used 3.3.1 Impedance Spectros copy Due to the differences in sample geometry and microstructure, impedance measurements performed on thin film and bulk samples required the use of different equipment and experimental conditions. This section will describe in detail the impedance spect roscopy techniques used in this work See section 2.3 for background information on EIS. 3.3.1.1 Thin Film Samples Two point AC electrochemical impedance spectroscopy (EIS) was performed on the samples in air from 100 300C. Samples were heated using a d igitally controlled hotplate and surface temperatures were confirmed using an optical pyrometer. Electrical contacts with top and bottom electrodes were made using a Micromanipulator 450 PM Test Station with tungsten probe tips. The test station was conn ected to an Agilent Precision Impedance Analyzer 4924A. Measurements were taken ov er the frequency range 40 Hz 1 MHz using an oscillating voltage of 300 mV. Nyquist plots from the impedance data were fit with ZView software using an equivalent circuit composed of a resistor in paralle l with a constant phase element until the %error of the fit was minimized. 3.3.1.2 Bulk Samples Ionic conductivity measurements were pe rformed using two point AC electrochemical impedance spectroscopy (EIS). A Solartron SI 1260 impedance

PAGE 50

50 analyzer was used to take measurements in air from 250 700C over the frequency range 0.1 Hz 32 MHz. Pellets were polished using SiC paper (up to 600 grit) to achieve smooth parallel surfaces that were then painted with Pt ink (Heraeus CL 11 5349) and finally fired at 900C for 1 h to ensure a good electrode ceramic interface Pt wires with a diameter of 1 paste (SPI Supplies) which was allowed to dry before the samples were installed in a glass reactor for measuring. The thermocouple was kept in close proximity to the samples to ensure an accura te temperature reading and pellets were held at each temperature for at least 20 min to allow them to reach equilibrium. Data was fit with ZView software using equivalent circuit analysis consisting of an inductor, to model the experimental setup, in seri es with two constant phase element resistor parallel components, to model the grain core and grain boundaries. Fits were performed such that the %error of the fit was minimized. Refer to Appendix B for more information. Reported conductivity values are the average of three samples. 3.3.2 X ray Diffraction XRD patterns from the calcined powders were collected using a Philips APD 3720 to ensure the complete dissolution of the dopants into the host ceria lattice. Lattice parameters were determined from XRD data using the Nelson Riley extrapolation technique 58 XRD analysis was performed on thin film samples using a Philips APD 3720 after both the Pt and ceria deposition steps to ensure the layers were highly oriented D spacin gs for both layers were calculated using XRD data to determine if the layers were in tension or compression.

PAGE 51

51 Though phi and omega scans would be helpful in further characterizing the microstructure of thin film samples, the inability to access necessary di ffraction equipment has prevented the inclusion of this data. 3.3. 3 Profilometry Profilometry was performed using a Dektak 150 Surface Profiler to determine the thickness of the ceria films. Scans were moving onto the NDC layer. To get a more reliable measurement, t he probe was set to detect step changes in both the positive and negative directions. Thicknesses were determined based on an average of three separ ate scans. 3.3. 4 Transmission Electron Microscopy While the TEM work was performed by Dr. Nicholas Rudawski under guidance of Prof. Nino and R. M. Kasse, the general process is describe here for completeness. The microstructure of thin film samples were i nvestigated by TEM using a combination of high resolution cross sectional TEM (HR XTEM), selected area electron diffraction (SAED), and convergent beam electron diffraction (CBED) using a JEOL 2010F transmission electr on microscope operating at 200 K eV wit h spherical aberration coefficient C s = 1.0 mm. Samples for TEM analysis were prepared via site specific focused ion beam (FIB) milling using an FEI DB235 dual beam scanning electron microscope/FIB system using an in situ lift out m ethod as described else where 59 ; the samples were prepared such that the beam direction (foil normal) used for TEM analysis was B = [ 1 100]. Prior to FIB processing, the specimens were coated with a ~100 nm thick conductive C layer to protect the surface of the structure during the initial stages of in situ FIB assisted deposit ion of a protective Pt layer 60 Immediately prior to

PAGE 52

52 TEM analysis, the sample was plasma cleaned for 2 min using 75%/25% Ar/O 2 plasma to prevent carbon contamination. T he specimen was etched for 10 s in a 25:1:26 (by volume) HNO 3 :HF:H 2 O solution, which was reported to be highly effective and selective in the removal of ion beam damaged Al 2 O 3 61 This etching strategy completely removed the protective C and Pt layers, but did not affect the thickness of the NDC or Pt layers ; the resulting sample quality was sufficient to allow for high quality HR XTEM imaging. 3.3. 5 Atomic Force Microscopy While the AFM work was performed by Mina Hanna and Alexander Rudy under guidance of Prof. Nino and R. M. Kasse, the general process is describe here for completeness. AFM analysis was performed using an MFP 3D Stand Alone AFM (Asylum Res earch LLC). Images were taken in tapping mode using a model AC240TS (Olympus) silicon probe tip (radius 9 2 nm) with an Al reflex coating Root mean square ( RMS ) roughness data and line scans were analyzed using Igor Pro software (Wavemetrics Inc). 3.3 6 Scanning Electron Microscopy Microstructure analysis of bulk samples was performed using a Jeol JCM 5000 Neoscope sc anning electron microscope (SEM) on both fracture surfaces and polished and thermally etched pellet cross sections. Pellets were prepare d for thermal etching by mechanical polishing using SiC polishing papers first and then finishing with Al 2 O 3 in methanol for 30 min, thermally etched at 1500C for 1 h, and again ultrasonicated for

PAGE 53

53 30 min in methanol Grain size values were determined by examining SEM images o f polished and thermally etched surfaces using ImageJ software 62

PAGE 54

54 CHAPTER 4 HIGHLY ORIENTED Nd 0.1 Ce 0.9 O 2 THIN FILMS ON Pt BOTTOM ELECTRODES 4.1 Introduction Understanding the effects of composition and microstructure on the transport properties of rare earth doped ceria is crucial to enhancing the ionic conductivity of these materials in the intermediate temperature range for use as electrolytes in SOFC applications. The role of grain size in determining the ionic conductivity is of particular interest. Distinguishing between grain and grain boundary ionic conductivity using data from bulk impedance mea surements is difficult especially in samples with nanometer sized grains, making it difficult to precisely determine the effect of grain size on grain ionic conductivity. 63 Nonetheless the development of newer methods for modeling conductivity in these materials has yielded some exciting results. 36 Some studies report that when nanostructured materials are synthesized, grain boundaries and interfaces become highly conductive paths for oxygen ion transport, the reby increasing the total ionic conductivity by orders of magnitude. 64 66 Conversely, other investigations have concluded that since grain boundary concentration i s increased as grain size decreases, and grain boundaries tend to block oxygen ion transport, total ionic conductivity is lower in nanostructured materials. 37 67 69 Tuller suggests that this discrepency may be due to a failure to separate ionic conductivity and total conductivity which includes contributions from electrons and holes, mechanisms that degrade SOFC perfor mance 66 The microstructure of thin film samples varies depending on the deposition m ethod used, substrate material, heat treatment, and other processing variables. Films can be single crystals, have columnar grains, or be polycrystalline, as seen in Figure 4

PAGE 55

55 4 1 below. Since polycrystalline samp les have grain boundaries at many random orientations it is not possible to completely separate grain and grain boundary conductivity. Single crystal films would allow for the direct measurement of grain conductivity due to the lack of grain boundaries. Films exhibiting columnar grain structures also allow for the direct measurement of grain conductivity provided across plane impedance measurements are performed since there would be no grain boundaries perpendicular to the path of conduction. Figure 4 below illustrates potential in plane and across plane electrode geometries for measuring thin film samples. Figure 4 1 Schematic illustration of films with single crystalline, columnar grain, and polycrystalline microstructures, from left to right respectively. Figure 4 2 Schematic illustration of electroding configurations for in plane (left) and across plane (right) electrical mea surements of thin film samples. Gobel et al. investigated the effect of thickn ess on the ionic conductivity of both GDC and undoped ceria thin films with epitaxial and columnar grain microstructures. 37 Grain structure was controlled by depositing on different substrates; films deposited on Al 2 O 3 (0001) were epitaxial while those deposited on SiO 2 (0001) exhibited columnar

PAGE 56

5 6 grains. Since film s were deposited on insulating substrates, in plane electrical measurements were performed using an electrode setup similar to what is depicted in Figure 4 2. The individual contributions from grain and grain boundary could not be separated due to stray c apacitance orders of magnitude larger than that of either the grain or grain boundary capacitance so only total conductivity was reported. 37 Dewetting of the platinum electrodes was noticed after electrical measurements on polycrystalline samples and was attributed to the presence of grain boundaries as this was n ot observed for epitaxial samples. Ionic conductivity was not influenced by the dewetting as electrodes were still electrically interconnected in all cases. 37 Ionic conductivity in epitaxial GDC thin films showed no dependence on thickness. This indicates that the film/substrate interface has no effect on conduc tivity in epit axial samples when measured in plane. Columnar grain GDC samples, however, showed an increase in total ionic conductivity with an increase in film thickness. This is explained by the corresponding decrease in lateral grain size with increas ing film thickness, resulting in a decrease in grain boundary concentration. 37 Important to note is the blocking nature of grain boundaries in these samples even with grain size on the order of tens of nanometers. Huang et al. also measured the ionic conductivity of polycrystalline GDC films as a function of thickn ess and found that total ionic conductivity increases three to four orders of magnitude as thickness decreases. 68 Initially these results may seem to contradict those of Gobel et al. However, once experimental conditions are considered the results actually tel l the same story. Polycrystalline thin film samples, with microstructures similar to that depicted in Figure 4 1, were deposited on platinized

PAGE 57

57 silicon substrates and across plane measurements were performed. Films synthesized in this study had grains on the order of 10 50 nm in diameter. The increase in total ionic conductivity was attributed to the reduction in the number of grain boundaries perpendicular to the path of conduction with decreasing film thickness. 68 Again, grain boundaries in nanocrystalline sa mples are presented as being blocking, limiting the conduction of oxygen ions and reducing total ionic conductivity. Additionally, it is reported that grain conductivity increases with decreasing film thickness, further enhancing conductivity of thinner f ilms. However, since there were grain boundaries present in the samples this could not be directly measured. While the thin film studies discussed above seem to imply that grain boundaries remain blocking even in nanostructured samples, several other stud ies report that grain boundaries become highly conductive paths for oxygen ions when grain size is on the order of nanometers. 64 66 Bellino et al. report an order of magnitude increase in total ionic conductivity in nanocrystalline samples compared to microcrystalline samples due to increased grain boundary ionic diffusivity. 64 A change in transport mechanism with temperature is also reported, with bulk transport being dominant at higher temperatures and grain boundary transpor t dominant at lower temperatures. Similar results are reported by Rupp et al. and Tuller et al. though both mention the possibility of the increased conductivity being attributed to electronic conduction rather than ionic conduction and suggest further st udies to determine which mechanism is responsible. 65 66 Across plane ionic conductivity measurements on highly orient ed films of varying thicknesses would provide critical new information regarding size effects in these materials by allowing grain ionic conductivity to be measured directly, but the

PAGE 58

58 community has thus far been unable to successfully synthesize highly orie nted doped ceria thin films on bottom electrodes. 37 68 70 In this study it has been shown for the first time the preparation of highly oriented thin film samples of doped ceria with a bottom electrode. T he first ever impedance measurements of these samples have been performed as a function of film thickness to directly determ ine size effects on grain ionic conductivity. 4.2 Structural Characterization Highly oriented thin film samples of Nd 0.1 Ce 0.9 O 2 with thicknesses ranging from 4 190 nm were prepared such that the across plane ionic conductivity could be measured as a func tion of film thickness. Figure shows the X ray diffraction (XRD) pattern of a 43 nm sample. It is evident that the film is highly oriented as the only peaks attributed to the ceria layer are the (111) and (222). The other peaks are from the single crystal a plane sapphire substrate and the Pt bottom electrode layer, which is also highly oriented and oriented along the same direction as the ceria layer. That NDC grows so highly oriented on the highly oriented Pt (111) layer is somewhat interesting. Joo et al. deposited GDC on Pt (111) substrates and obtained films with a polycrystalline microstructure 70 while Abid et al. reported oriented growth of Pt (111) on CeO 2 (111). 71 L attice mismatch with the sapphire substrate could cause t he Pt layer to be under strain which may contribute to the NDC layer being so highly oriented Figure shows XRD patterns for films 4 190 nm thick, proving that this process is reproducible and can be used to depo sit films covering a relatively wide range of thicknesses. It is proposed that the NDC layer grow s epitaxially within each grain due to alignment of the Pt (1 0) and NDC (3 ) directions based on coincident site lattice (CSL) calculations.

PAGE 59

59 XRD patterns obtained for films ranging from 4 190 nm thick prove that this process is reproducible and can be used to deposit films covering a relatively wide range of thicknesses. Figure 4 3 X ray diffr action pattern of a 43 nm thin film sample as compared to the theoretical patterns for Al 2 O 3 Pt, and 10% Nd doped ceria.

PAGE 60

60 Figure 4 4 X ray diffraction patterns for thin film samples of different thicknesses ranging from 4 190 nm. Transmission electron m icroscopy (TEM) data confirms that both the Pt and ceria layers are highly oriented along the (111) direction, as seen in Figure 4 5 The high resolution cross sectional TEM (HR XTEM) images show alignment of the different lattice fringes across both interfaces. The selected area electron diffraction (SAED) and convergent beam electron diffraction (CBED) patterns show that certain crystallographic planes in each of the three materials are aligned. D spacings meas ured from the high resolution TEM micrographs are in good agreement with bulk values reported in literature

PAGE 61

61 Figure 4 5 Convergent beam electron diffraction (left), selected area electron diffraction (center inset), and high resolution cross sectional T EM (right) all confirm that both the Pt and NDCC layers are highly oriented in the [111] direction. Results of atomic force microscopy (AFM) analysis are shown in Figure The root mean square (RMS) roughness of t he Pt layer is ~490 pm while the ceria layer has a RMS roughness of ~0.7 1.9 nm. This corresponds to a variation in thickness of less than 2% for even the thinnest ceria films tested.

PAGE 62

62 Figure 4 6 Atomic force microscopy topography maps and line sc ans of a Pt layer ( top ) and a 43 nm NDC film ( bottom ). Note the different scales used on the line scan plots.

PAGE 63

63 4.3 Ionic Conductivity Characteristic ionic conductivity data is shown in the figures below. Figure shows a typical Nyquist plot of the complex impedance obtained from an impedance measurement. Only one semicircle is present indicating that there was no grain boundary contribution to the conductivity. Measured conductivity values for thin film samples of 43 and 190 nm thicknesses are slightly lower but comparable to values obtained from bulk polycrystalline samples, as seen in Figure below The measured conductivity of the films may be lower than that of the b ulk for a few reasons. Liu et al. reported for Sm/Nd co doped ceria films, conduction along the [111] direction may be lower than other directions. 72 Also, Tsuchiya et al. found that the oxygen exchange rate is slower in highly textured films when compared to polycrystalline films. 73 Conductivity appears to be independent of film thickness, similar to that reported by Gobel et al. for in plane measurements of epitaxial GDC films 37 Considering the fact that the space charge layers (SCLs) near interfaces in doped ceria are on the order of 2 5 nm 36 and thus no interaction between SCLs would occur in films of the thicknesses studied, it makes sense that the grain core conductivity remains constant. Interestingly, t he conductivi ty values here obtained for the films fall within broad conductivity values reported in literature for doped ceria thin films with different microstructures, see Figure One possibility is that this is simply a r esult of the temperatures for which the values are reported. Recall that Bellino et al. reported a change in transport mechanism from grain boundary to bulk dominated with increasing temperature. 64 Additional measurements covering a wider range of thicknesses and temperatures may help to better determine the relation ship between film thickness and grain ionic conductivity.

PAGE 64

64 Figure 4 7 Typical Nyquist plot obtained from impedance measurements of thin film samples. Figure 4 8 Arrhenius plot of grain ionic conductivity for 10% Nd doped ceria comparing thin film samp les of two different thicknesses and data obtained from a bulk polycrystalline sample.

PAGE 65

65 Figure 4 9 Plot of the log of conductivity versus temperature. Across plane data from the highly oriented thin film samples measured in this study show good agreeme nt with literature data. The arrows indicate increasing film thickness. 37 68 4.4 Summary and Conclusions Highly orien ted doped cer ia thin films were deposited on Pt bottom electrodes for the first time and the first across plane ionic conductivity measurements of these films were performed. DC sputtering was used to deposit a highly oriented layer of Pt (111) on a plane sapphire substrates. The highly oriented doped ceria layer was then deposited using pulsed laser deposition. Such thin film samples allow for the first time the direct measurement of grain ionic conductivity of doped ceria since there is no grain bounda ry contribution to the across plane conductivity measurements. I onic conductivity data indicates inconsistencies in literature may simply be a result of the experimental conditions used in each study and failure to distingui sh between ionic and electronic conduction.

PAGE 66

66 CHAPTER 5 IONIC CONDUCTIVITY OF Sm x Nd y Ce 0.9 O 2 5.1 Introduction In the dilute region when dopant concentration is small, the migration enthalpy, E m for ionic conduction is independent of dopant concentration so a minimization of defect association enthalpy, E ass is necessary to increase ionic conductivity. 12 While finding the ideal dopant type and concentration is essential to both lowering activation energy and increasing ionic conductivity, the optimum composition may not be the same for both properties. 9 Kilner et al calculated association energies for dopants creating an effective charge of plus one and plus two and concluded that dopants causing an effective charge of plus one resulted in a lower association energy and thus a higher conductivity. 74 The composition with the highest conductivity will be the one with the best balance between a high pre exponential and a low activation energy. Ceria doped with Gd 3 + and Nd 3 + has shown the highest grain ionic condu ctivity of all singly doped materials in the IT range. 16 18 Co doping ceria has been shown to increase ionic conductiv ity to values above those of singly doped materials 19 22 25 75 80 though this is not the case for all systems 81 In co doped materials a reduction in the growth of defect clusters is observed which lowers the association enthalpy. 21 23 The reduction in defect clusters results in less local ordering of the structure which causes an increase in the conductivity. 24 Attempting to address this, d ensity functional theory calculations by Andersson et al. predicted that a dopant with an atomic number near that of Pm 3 + (61), which has an ionic radius of 1.093 would result in the lowest activation energy and thus highest ionic conductivity because oxygen vacancies in this system show almost no preference

PAGE 67

67 for being in nearest neighbor or next nearest neighbor positions with respect to the dopant cations. 11 82 Since Pm 3 + is radioactive and not suitable for use in fuel cell applications a co doping scheme using Sm 3 + (62) and Nd 3 + (60) was suggested such that an average atomic radius near the ideal value could be achieved. Work by Omar et al. showed that 1:1 co doping in Sm x/2 Nd x/2 Ce 1 x O 2 showed conductivity values intermediate between the singly doped materials with Nd doped ceria exhibiting the highest grain ionic conductivity. 16 20 More recent studies have determined that the addition of small amounts of a second dopant enhances conductivity beyond that of singly doped materials and 1:1 co doped. 25 75 For example, Park et al. report that in the Ce 0.8 Gd 0.2 x Dy x O 2 system, while 1:1 co doping actually degrades ionic conductivity, 25 Similarly, Dholabhai et al. found that for ceria doubly doped with Pr 3+ and Gd 3+ Gd rich compositions exhibited the highest conductivity. 75 Therefore a systematic investigation of this effect within the Sm x Nd y Ce 0.9 O 2 system is of significant interest and is the focus of this chapter. 5.2 Phase Analysis and Microstructural Characterization Fig shows the room temperature XRD patterns for all of the measured compositions. All samples show peaks corresponding to the cubic fluorite structure with no additional peaks confirming the complete dissolution of dopants into the host ceria lattice. Lattice parameters calculated from X ray data increased as dopants shifted from pure Sm +3 to pure Nd +3 83 as expected since the ionic radius of Nd +3 is larger than that of Sm +3 82 The secondary SEM image of a fracture surface in Fig shows the typical microstructure of all of the samples tested. All samples showed very low porosity which

PAGE 68

68 confirms the high relative densities calculated from g eometrical measurements. Grai n size analysis using ImageJ software showed that the average grain size was around 11 16 20 24 while others 14 17 25 The larger average grain size reported here is the result of a higher si ntering temperature (1600C) which may also affect dopant segregation to grain boundaries and increase grain ionic conductivity. 84 Fig ure 5 1. X ray diffraction patterns for Sm x Nd y Ce 0.9 O 2

PAGE 69

69 Fig ure 5 2. SEM image of a fracture surface of Sm 0.025 Nd 0.075 Ce 0.9 O 2 showing the typical microstructure of all samples. 5.3 Ionic Conductiv ity Fig shows the plot of versus 1000/T for the nine compositions examined. Note that all of the compositions exhibit Arrhenius behavior and the lines are nearly superimposed upon one another. It is impo rtant to note that all the conductivity values reported are the density corrected averages of three pellets of different geometric aspect ratios. Density corrections involved dividing the measured conductivity by the calculated relative density. 85 The plot of grain ionic conductivity versus dopant concentration at temperatures ranging from 400 700C presented in Fig shows a maximum in conductivity for Sm 0.075 Nd 0.025 Ce 0.9 O 2 with a value of 0.05070.0015 Scm 1 at 700C. However this increase is only ~8% over the 1:1 co doped composition which is statistically insignificant giv en the experimental error. The variation in conductivity also appears to

PAGE 70

70 increase with increasing temperature, but when the error bars, which also get larger, are taken into account, it is clear that this perceived variation is statistically insignificant It is anecdotal to note that Andersson et al. 11 suggested that an ideal dopant should have an effective atomic number between 61 and 62 and this dopant combina tion has an effective atomic number of 61.5. However, it is clear that at the unit cell level a material is composed of individual atoms, not averages of atoms. Also, it is important to note that the critical dopant radius (or atomic number) of the ideal dopant is temperature dependent; thus these values are to be assessed at the operating temperatures of interest as recently shown. 16 Furthermore and interestingly, although it shows the highest conductivity, Sm 0.075 Nd 0.025 Ce 0.9 O 2 is not the composition with the lowest activation energy nor does it have the highest pre exponential coefficient as seen in Table This result s from the fact that there are two competing mechanisms affecting the activation energy. Reduction in the amount of oxygen vacancy ordering due to an increase in configurational entropy will reduce the activation energy. 86 At the same time, the increase in the lattice parameter from that of pure ceria (5.4126 ) tends to increase the activation energy. 77 Configurational entropy values, calculated following Guan et al. 76 using Equation 5 1 below where R is th e ideal gas constant are reported in Table 5 1 As dopants shift from pure samarium to pure neodymium, both the activation energy and pre exponential coefficient increase as seen in 5 1 Sha et al. also reporte d about samples of co doped ceria electrolytes that had higher conductivity values despite also h aving higher activation energy values 77 (5 1 )

PAGE 71

71 As noted above, Park et al. 25 reported that conductivity values increased significantly in Ce 0.8 Gd 0.2 x Dy x O 2 co doped ceria with the addition of small concentrations (2 3%) of the second dopant while larger concentrations a ctually diminished conductivity and Dholabhai et al. 75 report ed a similar increase in conductivity in the Pr Gd co doped ceria system for samples rich in Gd relative to Pr. This enhancement may occur because the addition of small amounts of a second dopant with a similar ionic radius dis rupts local ordering of the cations and oxygen vacancies as suggested by Yamamura et al. 24 without any significant increase in activation energy. However, the compositions studied in this work indicate that this behavior is not exhibited by the Sm x Nd y Ce 0.9 O 2 system as no enhancement in conductivity is observed. Mori et al. 87 proposed th e idea of an effective index to explain ionic conductivity in d oped ceria systems, shown in Equations 5 2 and 5 3 below. Effective index takes into account the ionic radius of the host cation r h the average cation radius r c the average dopant radius r d and the effective oxygen radius r o which is dependent on the fraction of oxygen vacancies Kilner et al. had previously suggested that ionic conductivity in fluorite oxides was related to r d /r h 88 however, this fails to take into account oxygen vacancies. This approach suggests that a material with an effective index approaching 1 has an ideal fluorite crystal structure which leads to a higher ionic conductivity. Effective index values for the compositions studied can be found in Table 5 1 and agree with the experimental data in that both conductivity and effective index incre ase with increasing Nd content. However, it is important to note that the ionic radius values used in these calculations are at room temperature. A more appropriate

PAGE 72

72 meas ure of effective index could be obtained if ionic radius values are operating temperatures were used. Effective Index ( ( 5 2 ) ( (5 3 ) Grain ionic conductivity in rare earth doped ceria has been shown to be a function of both dopant type and concentratio n and increases with temperature following an Arrhenius type relationship. Enhancement of ionic conductivity can be achieved by co doping ceria with different elements. This work has shown that while further increases in conductivity can be realized by u sing a co doing scheme which uses only small additions of a second dopant in some systems, allowing for larger increases in the pre exponential coefficient relative to activation energy, this doping scheme does not enhance conductivity in all systems. The Sm x Nd y Ce 0.9 O 2 system correlates well with the theory of effective index. Finally, it is important to note that in doped ceria systems, segregation of dopants to the grain boundaries and microdomains containing ordered oxygen vacancies tends to lower ion ic conductivity and degrade material performance. In co doped systems, one dopant may preferentially segregate leading to non uniform cation distribution and an increased deleterious effect on grain and grain boundary conductivity. The Sm x Nd y Ce 0.9 O 2 sy stem, exhibiting uniform conductivity across all (x/y) dopant ratios measured, appears to be free of these problems thus presenting a compositionally versatile system for use in IT SOFC applications.

PAGE 73

73 Table 5 1 Activation energy pre exponential coefficien t, configurational entropy, and effective index values for all of the compositions studied. Compound Activation e nergy (eV) Log(p re exponential) Configurational e ntropy, S (JK 1 mol 1 ) Effective i ndex Sm 0.1 Ce 0.9 O 2 0.646 0.002 5.027 0.022 2.703 0.799 Sm 0.0875 Nd 0.0125 Ce 0.9 O 2 0.647 0.007 5.004 0.052 3.016 0.802 Sm 0.075 Nd 0.025 Ce 0.9 O 2 0.654 0.004 5.100 0.029 3.170 0.806 Sm 0.0625 Nd 0.0375 Ce 0.9 O 2 0.657 0.008 5.054 0.050 3.253 0.809 Sm 0.05 Nd 0.05 Ce 0.9 O 2 0.653 0.006 5.067 0.045 3.279 0.812 Sm 0.0375 Nd 0.0625 Ce 0.9 O 2 0.659 0.003 5.077 0.027 3.253 0.815 Sm 0.025 Nd 0.075 Ce 0.9 O 2 0.663 0.003 5.132 0.021 3.170 0.818 Sm 0.0125 Nd 0.0875 Ce 0.9 O 2 0.659 0.007 5.063 0.047 3.016 0.821 Nd 0.1 C e 0.9 O 2 0.660 0.006 5.108 0.047 2.703 0.824 Fig ure 5 3. Arrhenius plot of grain ionic conductivity for all nine compositions investigated in the Sm x Nd y Ce 0.9 O 2 system. There is no apparent variation in conductivity with composition.

PAGE 74

74 Fig ure 5 4. Plot of grain ionic conductivity versus dopant concentration for Sm x Nd y Ce 0.9 O 2 from 400 700C. 5.4 Conclusions The development of materials with higher ionic conductivities in the intermediate temperature range is necessary for the advancement of IT SOF C technology. Co doping using Sm +3 and Nd +3 has demonstrated only intermediate values between those of the respective singly doped materials with Nd doped ceria exhibiting the highest conductivity. This work investigated the Sm x Nd y Ce 0.9 O 2 system and fo und that non 1:1 co doping does not result in higher conductivity values than singly doped or evenly co doped materials but rather follows the theory of effective index. Since co dopant ratio (x/y) variations within the Sm x Nd y Ce 0.9 O 2 system do not lead to significant conductivity changes it constitutes a versatile system impervious to potential performance degradation due to preferential dopant segregation and redistribution.

PAGE 75

75 CHAPTER 6 IONIC CONDUCTIVITY ACROSS THE DISORDER ORDER PHASE TRANSITION IN THE NdO 1.5 CeO 2 SYSTEM 1 6.1 Introduction Rare earth elements are considered the main candidates for doping ceria because of the resulting low activ ation energy for conduction. 11 They are also shown to slightly improve the mechanical properties of ceria 90 which is beneficial during cell fabrication and operation. Through doping, oxygen vacancies are introduced into the ceria lattice to maintain the overal l charge neutrality of the material. At low dopant concentrations, these oxygen vacancies will distribute randomly in the host lattice in a disordered manner. However, in heavily doped ceria the oxygen vacancies tend to occupy preferential lattice sites around the dopant cations 91 leading to the ordering of oxygen vacancies. The short range ordering in ceria based materials has been invest igated by several group s 91 95 For instance, Ou et al 92 observed experimentally the formation of nanosized domains exhibiting a C type rare earth oxide structure (C type structure) in 25 mol% lanthanide (Ln) doped ceria (Ln=Sm, Gd, Dy, and Yb); these ordered domains were believed to be the cause of the degraded ionic conductivity. In heavily doped ceria oxygen vacancies tend to occupy preferential lattice sites around the dopant cations 91 leading to the ordering of oxygen vacancies. The short range ordering in ceria based materials has been invest igated by several groups 91 95 For instance, Ou et al 92 observed experimentally the formation of nanosized domains exhibiting a C type rare earth oxide structure (C type structure) in 25 mol% lanthanide 1 Adapted from reference 89, Li, L. et al. Ionic conductivity across the disorder order phase transition in the NdO1.5 CeO2 system. Solid State Ionics 221 15 21, doi:10.1016/j.ssi.2012.06.007 (2012).

PAGE 76

76 (Ln) doped ceria (Ln= Sm, Gd, Dy, and Yb); these ordered domains were believed to be the cause of the degraded ionic conductivity. Figure 6 1. The ternary phase diagram of Nd O 1.5 Sm O 1.5 Ce O 2 at room temperature reported by T. Vanderah 96 The dots on the graph represent experime ntal compositions examined by XRD; F indicates the region crystallizing with the fluorite structure; A, B, and C indicate the regions with A B and C type rare earth oxide structure, respectively. While short range ordering causes defect clustering and nanosized domains in the lattice, long range ordering result s in phase change of the bulk material. Pure ceria has a cubic fluorite structure and at certain Nd concentration the NdO 1.5 CeO 2 system

PAGE 77

77 will stab ilize into the C type structure as mentioned in section 2.2.1 Samples with different Nd content have been reported for the onset of C type phase. For instance, Chavan et al 97 found C type phase started to show up at 52.5% Nd. Parks et al 98 reported the first composition exhibiting weak C type reflections to be 50% Nd and 40% Nd was type peaks. Quite recently, the phase equilibrium relations of the Nd O 1.5 Sm O 1.5 Ce O 2 system were investigated by T. Vanderah 96 and t he reported ternary phase diagram (redrawn for clarity ) is presented in Figure 6 1 It is seen th at the disorder order transition happen s at 40.5 mol% Nd and 35.5 mol% Sm concentrations along the NdO 1.5 CeO 2 and SmO 1.5 CeO 2 tie lines, respectively. The conductivity of neodymi um doped ceria (NDC) has been studied by different groups 99 104 However the conductivity of NdO 1.5 CeO 2 system and in particular the conductivity across the disorder order phase transition has not been systematically examined. Furthermore, although heavy doping of ceria was believed to affect the conductivity adversely, in the work by Omar et al 20 the grain conductivity of Sm 3+ and Nd 3+ co doped ceria continued to increase up to 18% total dopant concentration. More recent work by Fu et al 104 also showed that the total conductivity was still increasing at 25% NDC, posing the possibility of furthe r conductivity enhancement upon increased Nd content. In this work, the disorder order phase transition of the NdO 1.5 CeO 2 system is examined by XRD and neutron diffraction at room temperature in order to elucidate the phase transition within the NdO 1.5 Ce O 2 system. Th e ionic conductivity and the changes of pre exponential factor and activation energy across the disorder order

PAGE 78

78 transition i s evaluated in order to better understand the impact of oxygen vacancy ordering on the ionic conductivit y of ceria bas ed electrolytes. 6.2 Phase and Structural Analysis 6.2.1 X ray Diffraction The as calcined powders were characterized by XRD a nd the results are shown in Figure 6 2 (a and b). In Figure 6 2 (a) i t is shown that all XRD patterns of 0.05 x 0.4 5 corres pond to cubic fluorite structure, which is the same structure as pure ceria. It is thus confirmed that Nd was fully dissolved into ceria and solid solutions were formed. the increased expansion of the host lattice, which resulted from the fact that the ionic radius of Nd is larger than that of Ce (r Ce 4+ = 0.97 and r Nd 3+ = 1.109 82 ). Figure 6 2 (b) shows the XRD profiles for x = 0.50 and 0.55 in addition to the theoretical pattern of the C type structure for Nd 0.50 Ce 0.50 O 1.75 It is clear th at the XRD profiles for x = 0.50 and 0.55 match the theoretical pattern. In order to identify the minor peaks of the C type structure, the inset in Fig ure 6 2 (b) shows a close up of the theoretical positions, the experimental XRD profiles for x = 0.50 an d 0.55, and the profiles for x = 0.40 and 0.45 for comparison. While for x = 0.40 and 0.45 no peaks can be identified, the reflections for x = 0.50 and 0.55 correspond well with the theoretical positions. According to Grover et al 105 and as discussed in section 2 .2.1 t he XRD profile of C type structure only differs from the fluorite in that the former has extra super lattice peaks in addition to the reflections by the cubic structure and t hese extra super lattice peaks are especia lly weak. Thus it is inferred that at about 50% Nd the compound crystallizes with the C type structure.

PAGE 79

79 Figure 6 2. XRD pattern s of the calcined powders of Nd x Ce 1 x O 2 (a) (0 x show ing fluorite structure. (b ) for x = 0.50 and 0.55 exhibiting C type structure. The inset in (b) shows a close up of the theoretical positions of the C type structure, and the experimental patterns for x = 0.40 0.55.

PAGE 80

80 Figure 6 3. Lattice parameters of Nd x Ce 1 x O 2 (0 x half lattice parameters are plotted for x = 0.50 and 0.55, for which C type structure is observed. The results reported by Omar et al 16 Aneflous et al 103 Stephens et al 102 Chakraborty et al 106 and Parks et al 98 are added for comparison. Note the lattice parameters for 0 x overlapped with those by Aneflous. In order to quantify the lattice expansion, the cell parameters were calculated using the Ne lson Riley extrapolation method 58 The results are presented as a function o f Nd concentration in Fig ure 6 3, with lattice parameters reported in literature are also included for comparison. It should be noted that in Fig ure 6 3 half lattice parameters are plotted for x = 0.50 and 0.55, after transition to C type structure and do ubling of the unit cell as discussed in section 2.2.1 For 0.05 x 5 t he lattice parameters

PAGE 81

81 increase linearly with increasing Nd content 83 The fitted linear functio n can be expressed as (6 1) where x = Nd concentration, and a = lattice parameter ( ) The lattice constant determined in this work had excellent agreement with that of Aneflous et al 103 On the other hand, the results by Stephens et al could be fitted by a 2 nd order polynomial equation and was explained as having attractive interactions among def ects 102 The differences may be due to varying purity of starting materials : C hemicals used in this work and th at of Aneflous et al. are 99.9% pure or above, while the purity for ceria nitrate in the work by Stephens et al. is only above 99%. Another factor that may have caused the disparity in lattice parameters is sample preparation. In this work solid state re actions are used, while Aneflous et al. and Stephens et al. used sol gel and nitrate decomposition techniques, respectively. A substantial deviation from the linear increase of lattice parameters is observed above 0.50. The contraction of the lattice may be attributed to the possibly enhanced defective interactions of C type structure, which is already forming by x = 0.50 as evidenced by XRD. 6.2.2 Neutron Diffraction Neutron diffraction was carried out on powders with compositions where 0 .35 x 0. 5 5 The observed patterns for all sa mples tested are plotted in Figure 6 4. The calculated peak positions for fluorite and C type structures are also shown a t the bottom and the top of Figure 6 4. For samples with 35 45% Nd, the neutron diffraction profil es can be indexed with fluorite structure and no reflections of the C type structure were

PAGE 82

82 observed. C type reflections began to appear in the sample containing 50% Nd, and these reflections became more prominent with 55% Nd. As an example for the fluorit e structure, the observed, calculated, and difference neutron diffraction patterns of Nd 0.35 Ce 0.65 O 1.825 a re presented in Figure 6 5 The calculated peak positions are also added as ticks in Fig ure 6 5 Figure 6 6 shows the obtained, calculated, and diff erence neutron diffraction profiles of Nd 0.55 Ce 0.45 O 1.725 having the C type structure. The crystallographic structure reported by Chakraborty et al. 106 was used as the initial parameters for the Rietveld refinement of C type structure. The background wa s modeled by f unction 1 (Shifted Chebyschev) with a sixth order polynomial. Profile function 4 with Stephens asymmetry incorporated 107 was used to fit the diffraction peak profile. Subsequently, unit cell parameter, atomic positions, isotropic displacement parameters, and site occupancies were refined The refinement re sults for Nd 0.55 Ce 0.45 O 1.725 are listed in Table 6 1. The fitted lattice parameter (10.99 80 ( 3 ) ) is quite close to what was calcul ated in Sec. 3.1 (10.998(3) ). Table 6 1. Detailed crystallographic information for Nd 0.55 Ce 0.45 O 1.725 obtained by Rietvel d refinement. Space Group: Ia 3 Lattice parameter: a = 10.99 80 ( 3 ) Name X Y Z Ui/Ue*100 Multiplicity Occupancy Nd1 0.250000 0.250000 0.250000 1. 50 ( 8 ) 8 0.56 6 (3) Ce1 0.250000 0.250000 0.250000 1. 50 ( 8 ) 8 0.4 27 (3) Nd2 0.016 6 ( 9 ) 0.000000 0.250000 1. 30 ( 5 ) 24 0.5 48 ( 7 ) Ce2 0.016 6 ( 9 ) 0.000000 0.250000 1.30 ( 5 ) 24 0.4 52 ( 7 ) O1 0.382 8 ( 7 ) 0.137 4 ( 5 ) 0.378 6 ( 4 ) 2.0 8 ( 7 ) 48 0. 89 (1) O2 0.38 68 ( 8 ) 0.38 68 ( 8 ) 0.38 68 ( 8 ) 4.6 ( 1 ) 16 0. 85 ( 2 )

PAGE 83

83 Fig ure 6 4. Observed neutron diffraction profiles for 0. 3 5 Nd 5 5 Calculated peak positions for fluorite and C type structure are added as ticks at the bottom and t op of the graph, respectively. Fig ure 6 5. Observed (cross), calculated (continuous line), and difference neutron diffraction profiles for Nd 0. 35 Ce 0.65 O 1.825 are shown.

PAGE 84

84 Fig ure 6 6 Observed (cross), calculated (continuous line), and difference neutron diffraction profiles for Nd 0.55 Ce 0.45 O 1.725 are shown. While T. Vanderah 96 and Chavan et al 97 found the composition that started to exhibit C type structure had 4 0.5% and 52.5% Nd respectively, both neutron diffraction and XRD in this work indicates the formation of C type structure happens at 50% Nd, which is also the value reported by Parks et al. 98 as noted earlier. Since the formation of C type phase is kinetically controlled 97 the disparity may be due to different sample preparation, including synthesis method, heat treatment, etc., considering that T. Vanderah 96 used solid state reaction s and the samples were calcined at 1400C for 40 h twice, followed by a third calcination at 1400C for 64 h, with an intermediate grinding of samples between each heat treatment. Chavan et al 97 also used solid state reaction method with slow cooling (2C/min) after heat treatment. In this work all samples were ball milled for 24 h, calcined at 1450C for 10 h and cooled dow n at ~3C/min. Parks et

PAGE 85

85 al 98 prepared samples by co precipitation technique and various heat treatments were used, followed by air quenching. 6.3 Microstructural Analysis The microstructure of the sintered pellets was characterized by SEM and micrographs for x = 0.05, 0.1 5 a nd 0.25 are presented in Figure 6 7 where (a c) are thermally etched surfaces and (d f) represent cross sections. It can be seen that the pellets are well densified, which is in accordance with the measured high relative densities. The individual grains on the cross section SEM micrographs are barely 108 109 us ing thermally etched large SEM images and the results are listed in Table 6 2. In the current study the Nd marked impact on grain size, which is in agreement with observations reported by Fu and Chen 104 who sintered samples of Nd x Ce 1 x O 2 (0.05 x For other lanthanide doped ceria systems, both increases 13 and decreases 110 in grain size with an increase in trivalent ion s have been reported. However, the grain ionic conductivity is expected to be independent of grain size in bulk samples since it is an intrinsic property of the material.

PAGE 86

86 Figure 6 7 SEM micrograph for sintered pellets of Nd x Ce 1 x O 2 : a d) x = 0.05 ; b e ) x = 0.15 ; and c f ) x = 0.25. The images on the left hand side are for thermally etched surfaces, while the images on the right hand side represent fractured cross sections. Table 6 2. Grain sizes of the sintered pellets of dif ferent compositions. Composition Grain size (m) Nd 0.05 Ce 0.95 O 1.975 12.3(0.4) Nd 0.15 Ce 0.85 O 1.925 14.0(0.7) Nd 0.25 Ce 0.75 O 1.875 5.7(0.1) 6.4 Ionic Conductivity The grain ionic conductivity was measured from 250 700C in air and the Nyquist plot of impe dance for Nd 0.25 Ce 0.75 O 1.875 (at 250C) is shown in Figure 6 8 Two well defined semi circles, representing the grain and grain boundary resistances respectively, can be easily identified. Electronic conduction in the materials was assumed to be negligib le since it has been shown the reduction of Ce 4+ to Ce 3+

PAGE 87

87 happens only at reducing atmospheres and high temperatures (~ 1000C) 111 112 Fig ure 6 9 shows the Arrhenius plot and it is seen that the conductivity follows Arrhenius type behav ior. The straight lines In Figure 6 9 are the least square linear fitting results and the adjusted coefficients of determination ( R 2 adj. ) are above 0.999 for all linear fittings. Although Arrhenius plots are sometimes fit using two straight lines (i.e., one line for high temperatur e region and the other for low temperature), minimum deviation from linearity in the whole temperature range was observed in this work. Therefore fits using only one straight line per composition were performed. Similar practices ha ve recently been repor ted for the NdO 1.5 CeO 2 system 10 4 as well as for the SmO 1.5 CeO 2 113 and GdO 1.5 CeO 2 114 systems. The isothermal plot of ionic conductivity as a function of Nd content at temperatures between 450 700 C is shown in Figure 6 8. The maximum conductivity is found to be 0.054 (1) Scm 1 exhibited by Nd 0.15 Ce 0.85 O 1.925 at 700 C with an associated activation energy of 0.7 26 (5) eV In Figure 6 10 the grain conductivity reported by Stephens et al 102 and Omar et al 20 is also added for comparison. It is observed that at a fixed temperature, as the Nd content increases, the conductivity first increased and reached its maximum values at 10% or 15% Nd content (depending on the temperature), then starts to decrease steadily The initial increase in conductivity is attributed to the increase of charge carriers (oxygen vacancies) into the host lattice However, as more Nd is added the fraction of trivalent ion oxygen vacancy associates increases le ading to the immobilization of these oxygen vacancies Therefore the decrease in conductivity is seen at higher Nd regions in Fig ure 6 9 In contrast to this work, Fu et al 104 found that Nd 0. 25 Ce 0. 7 5 O 1.875 delivered the highest total conductivity among all compositions tested. It s hould be stressed that grain conductivity is

PAGE 88

88 examined in this work while Fu et al evaluated total conductivity (sum of grain and grain boundary conductivity). Consequently the different trend observed is explained by the fact that grain conductivity is a fundamental material property while microstructure and impurity, in addition to chemical formation, affect grain boundary conductivity as well. 20 Figure 6 8 Nyquist plot of impedance for Nd 0.25 Ce 0.75 O 1.875 measured at 250C in air. Figure 6 9 Arrhenius plot of grain conductivity of Nd x Ce 1 x O 2 ( 0 x 0. 55).

PAGE 89

89 Figure 6 10 Grain conductivity of Nd x Ce 1 x O 2 ( 0 x 0. 55) as a function of Nd concentration (x) at temperatures from 450 70 0C. The results reported by Stephens et al. 102 and Omar et al. 20 are added for comparison. It is worthwhile to mention that across the expected disorder order transition region (either 40.5% or 50%), no drastic change of conductivity is observed which is in direct contrast to phase transition of Bi 2 O 3 Harwig et al. 115 found the conductivity Bi 2 O 3 Bi 2 O 3 at 729C. Bi 2 O 3 which has fluorite type structure with a quarter of the oxygen anion sublattice unoccupied, has th e highest oxide ion conductivity among all known fluorite based oxides 116 (~ 1 Scm 1 above 730C 115 ), which is ascribed to highly disordered state and high mobility of anion vacancies. 115 T o further understand the observed change in grain conductivity across the phase transition, A 0 and E A for different Nd concentrations were calculated from the ordinate intercept and the s lope of the fitted lines in Figure 6 9 respectively. The results are shown in Fig ure 6 9, where the shortest cation anion distance is also presented. It should be noted that in Figure 6 9 l og ( A 0 ) is used as one of the ver tical

PAGE 90

90 axes instead of A 0 It is seen both l og ( A 0 ) and E A exhibit similar trend: they fi rst increased with increasing Nd concentration and then start ed to drop a bove 40% Nd Most notably, when Nd content went from 3 5 % to 40 %, both l og ( A 0 ) and E A experien ced drastic increases, which can be attributed to the disorder order transition that happened at around 40% Nd, which is the composition reported by Bevan 117 that started to exhibit C type peaks. Although both XRD an d neutron diffraction established earlier that Nd 0. 50 Ce 0. 50 O 1.75 is the first composition to have C type reflection at room temperature, all impedance measurements were carried out at 250C or above; therefore it is likely that at higher temperatures the d isorder order transition in the NdO 1.5 CeO 2 system may be shifted to lower Nd content. Such shift in the NdO 1.5 CeO 2 system was also reported by others 97 Figure 6 11 Activation energy ( E A ), pre exponential factor ( A 0 ), and shortest cation anion distance as a functi on of Nd content.

PAGE 91

91 In Fig ure 6 11 l og(A 0 ) increase s from 6.11(4) to 7.12(4) between 35% and 40% Nd, indicating an order of magnitude change for A 0 This large increase i s consistent with the change in the local crystallographic environment (such as coordi nation number bond distance, and change in concentration of vacancies due to emergence of vacant crystallographic sites [in the ordered structure]) for the anion when transitioning from fluorite to C type structure P ossible sources for the increase of A 0 include the change of mobile oxygen vacancy fraction [V O ] and configurational entropy (part of ) due to the ordering of oxygen vacancies in C type structure. During the phase transition E A shows a marked increase as seen in Fig ure 6 11 T h e observed increase is due to the stabiliz ation of oxygen vacancies in C type structure. These stabilized o xygen vacancies are likely to have higher binding energies 118 and form deep traps, which prevent oxygen migration. 93 It is also shown in Fig ure 6 11 the increase of E A is accompanied by the decrease of the shortest cation anion distance, which may contribute to higher E A since stronger columbic attraction can be expected when the cation and anion come closer. It is not surprising to observe the marked increase of E A during the phase transition of the bulk material, given that E A is already seen to increase greatly when the size of the micro C type domain became larger in YO 1.5 CeO 2 system 119 Although both A 0 and E A vary noticeably across the disord er order transition, as seen in Fig ure 6 10 they seem to counteract each other resu lting in an overall gradual de crease in conductivity. In Figure 6 9 at low temperatures the gap between x = 0.35 and 0.40 is clearly larger than those between the other concentrations, indicating greater changes in conductivity during the phase transitio n. However, given the small

PAGE 92

92 conductivity values at 450C these changes are not readily noticeable in Fig ure 6 10 For NdO 1.5 CeO 2 system the optimal conductivities in the intermediate temperature range are achieved with materials containing 10 to 15% of Nd 6.5 Conclusions Samples of Nd x Ce 1 x O 2 ( 0 x 0. 55) were prepared by conventional solid state reaction s and the grain conductivity for all samples was evaluated. The lattice parameters for 0 x 0.4 5 increased linearly with increasing Nd content, complying and neutron diffraction suggested that the first composition that exhibited C type reflections at room temperature was Nd 0.50 Ce 0.50 O 1.75 Although both the pre exponential factor ( A 0 ) and activation energy ( E A ) increased noticeably when Nd content went fr om 35% to 40%, they counteracted each other and the overall effect is that the conductivity declined gradually. The disorder order transition during the impedance measurements was shown to happen at around 40% Nd concentration, deviating from the XRD and neutron diffraction results due to different testing conditions. No unusual change of conductivity was observed either around 40.5% or 50% Nd, which are the possible disorder order transition compositions reported in literature. The maximum conductivity of 0.054(1) Scm 1 was exhibited by Nd 0.15 Ce 0.85 O 1.925 at 700 C and the associated activation energy was 0.7 26 (5) eV.

PAGE 93

93 CHAPTER 7 SUMMARY AND FUTURE WORK 7.1 Summary The development of materials with higher ionic conductivity in the intermediate temperatur e range is necessary for the advancement of IT SOFC technology. The work reported in this thesis investigates size and compositional effects on ionic conductivity in doped ceria. Highly oriented doped ceria thin films were deposited on Pt bottom electrod es for the first time and the first across plane ionic conductivity measurements of these films were performed. DC sputtering was used to deposit a highly oriented layer of Pt (111) on a plane sapphire substrates. The highly oriented doped ceria layer wa s then deposited using pulsed laser deposition. Such thin film samples allow for the first time the direct measurement of grain ionic conductivity of doped ceria since there is no grain boundary contribution to the across plane conductivity measurements. I onic conductivity data indicates inconsistencies in literature may simply be a result of the different experimental conditions used in each study Prior to this work, c o doping using Sm +3 and Nd +3 has demonstrated only intermediate values between those o f the respective singly doped materials with Nd doped ceria exhibiting the highest conductivity. This work investigated the Sm x Nd y Ce 0.9 O 2 system and found that non 1:1 co doping does not result in higher conductivity values than singly doped or evenly c o doped materials but rather follows the theory of effective index. Since co dopant ratio (x/y) variations within the Sm x Nd y Ce 0.9 O 2 system do not lead to significant co nductivity changes it constitutes a versatile system impervious to potential performa nce degradation due to preferential dopant segregation and redistribution.

PAGE 94

94 Samples of Nd x Ce 1 x O 2 ( 0 x 0. 55) were prepared by conventional solid state reaction s and the grain conductivity for all samples was evaluated. The lattice parameters for 0 x 0.4 5 increased linearly with increasing Nd content, complying composition that exhibited C type reflections at room temperature was Nd 0.50 Ce 0.50 O 1.75 Although both the pre exponential factor ( A 0 ) and activation energy ( E A ) increased noticeably when Nd content went from 35% to 40%, they counteracted each other and the overall effect is that the conductivity declined gradually. The disorder order transition during the impedan ce measurements was shown to happen at around 40% Nd concentration, deviating from the XRD and neutron diffraction results due to different testing conditions. No unusual change of conductivity was observed aroun d either possible disorder order transition compositions reported in literature 40.5% or 50% Nd The maximum conductivity of 0.054(1) Scm 1 was exhibited by Nd 0.15 Ce 0.85 O 1.925 at 700 C and the associated activation energy was 0.7 26 (5) eV. 7.2 Future Work 7.2.1 Thin Films Investigating the across plane ionic conductivity of highly oriented ceria thin films over a wider range of dopants, thicknesses, and temperatures is necessary to better understand the behavior in these promising electrolyte materials. Finding ways to deposit higher quality film s could provide the abi lity to measure thinner films, perhaps accomplished using atomic layer deposition. Measuring films as thin as 4 10 nm, the typical size of the space charge layer (SCL) in doped ceria materials being 2 5 nm 36

PAGE 95

95 would allow for the direct observation of the effect of SCL interactions on grain ionic conductivity. D evelop ing a better setup for the measurement of ionic conductivity of thin films, p articularly focusing on sample heating and atmospheric control would also be helpful The ability to heat the sample up to 800C under various oxygen partial pressures while ensuring the film does not move (potentially using vacuum to secure the film) wou ld provide data that could more easily be compared to typical bulk data while increasing reproducibility. Measuring conductivity of the films as a function of oxygen partial pressure could confirm that the conductivity is ionic. It would also be interesti ng to use atom probe analysis to investigate the segregation of various species to surfaces and interfaces in th in film samples to determine whether or not this segregation is similar to what is observed in bulk samples. This would also be use ful in deter min ing if segregatio n is affected by film thickness, and thus grain size. 7.2.2 Bulk The primary focus of this thesis was on enhancing the grain ionic conductivity of doped ceria. However, the presence of grain boundaries in bulk ceria limits the total io nic conductivity and is thus also very important. Additional work relating various doping strategies to the grain boundary and total ionic conductivity is need for a complete understanding. It has been demonstrated that when bulk samples are synthesized u sing coprecipitated nanopowder and microwave sintering, grain boundary impedance is reduced. 120 However, due to the smaller grain size, the grain boundary density is

PAGE 96

96 significantly higher thus the total ionic conductivity is lower since the grain boundaries are still blocking. Synthesizing samples using micron sized coprecipitated powder and microwave sintering could produce bulk material with both low grain boundary impedance and grain boundary density resulting it a higher total ionic conductivity.

PAGE 97

97 APPENDIX A NDC FILMS ON VARIOUS SUBSTRATES Highly oriented thin films of NDC were successfully deposited on platinized substrates, as reporte d above, but there were many prior unsuccessful attempts. The parameters which were varied include substrate material and substrate orientation The simplest approach involved using platinized silicon substrates since they could be used as purchased witho ut the need to deposit a Pt layer. However, the resulting films were polycrystalline as evidenced by the appearance of additional peaks in the XRD pattern, as seen in Figure A 1 below. This is attributed to the fact that the Pt layer on the platinized si licon substrates was also not preferentially oriented Figure A 1. XRD pattern of a polycrystalline NDC film deposited on a platinized silicon substrate using PLD The unlabeled peaks are attributed to Si and Pt layers.

PAGE 98

98 In order to deposit an highly ori ented film of NDC, it was determined that it is necessary to start with a highly oriented layer of Pt. The first attempt used RF sputtering to deposit Pt on c plane (0001) sapphire substrates. However, the equipment used was not capable of matching the t emperature and pressure conditions stated by Bachelet et al. 121 and the Pt layer was polycrystalline, as seen in Figure A 2 below. Figure A 2. XRD pattern of a polycrystalline Pt layer deposited on a c plane (0001) sapphire substrate using DC sputtering The next attempt used DC sputtering to deposit Pt on a plane (11 20) sapphire subs trates following Nefedov et al. 57 The absence of the (200) and (220) peaks in Figure A 3 below indicate th at the Pt layer i s highly oriented F ilms deposited on these

PAGE 99

99 platinized substrates were highly oriented confirming that a highly oriented Pt layer is a necessary condition for the deposition of highly oriented NDC films. Figure A 3. XRD pattern of a highly oriented Pt layer deposited on an a plane sapphire substrate using DC sputtering. The peak aro und 81 is the (22 0) peak attributed to the substrate. These films are highly oriented but not epitaxial. The high degree of orientation is confirmed by the omega scan seen in Figure A 4. The full width at ha lf max (FWHM) is below one degree, indicating a high level of orientation. However, the pole figure seen in Figure A 5 exhibits a ring of intensity and not discrete spots, indicating the film is not epitaxial.

PAGE 100

100 Figure A 4. Omega scan of a highly oriented NDC film with a FHWM = 0.86. Figure A 5. The pole figure of an NDC film exhibits a ring of intensity indicating the film is not epitaxial. The pole was the [022] direction and the scan covered chi angles 29 41

PAGE 101

101 APPENDIX B FITTING IMPEDANCE DATA USI NG ZVIEW Fitting impedance data using equivalent circuit analysis is made quite simple by the use of ZView software. Once the data file is opened in ZView, the range of data points which is to be fit is selected and initial guesses are made for circuit co mponents; only data points for the first two arcs (grain and grain boundary) are fit. To minimize the error of the fit, it may be necessary to fix certain parameters for initial fitting runs, which can be done by clicking the button until it changes to an X. For example, fixing the values of L1, R1, and R2 can help better fit the values of CPE1 T and CPE2 T. Once this is complete, changing all values back to and running another fit usually produces lower error. An example for an Nd 0.1 Ce 0.9 O 2 sampl e at 250C is seen in Figure B 1. Figure B 1. Equivalent circuit fitting window in ZView. The values of R1 and R2 correspond to the diameters of the semicircular arcs associated with grain and grain boundary impedance. Values for CPE1 T are typically

PAGE 102

102 on the order of 10 11 10 13 while values for CPE2 T are typically on the order of 10 6 10 8 depending on composition and temperature. CPE1 P and CPE2 P are typically between 0.8 1.1. L1, which is a measure of the inductance of the experimental setup, is us ually around 10 6 10 8 Figure B 2 shows the result of the equivalent circuit fit whose component values are presented in Figure B 1. Figure B 2. Fit resulting from equivalent circuit analysis performed using ZView.

PAGE 103

103 LIST OF REFERENCES 1 Steele, B. C. H. Material science and engineering: The enabling technology for the commercialisation of fuel cell systems. Journal of Materials Science 36 1053 1068 (2001). 2 Wachsman, E. D. & Lee, K. T. Lowering the Temperature of Solid Oxide Fuel Cells. Science 334 9 35 939, doi:10.1126/science.1204090 (2011). 3 Boivin, J. C. & Mairesse, G. Recent material developments in fast oxide ion conductors. Chemistry of Materials 10 2870 2888 (1998). 4 Labo ratory, N. E. T. < http://www.netl.doe.gov/technologies/coalpower/ fuel cells/ seca/ primer/cell.html > 5 Phadke, S. R. & Nino, J. C. Conductivity Enhancement in Lanthanum Phosphates. Journal of the American Ceramic Society 94 1817 1823 (2011). 6 Chemistry of Materials 22 660 674, doi:10.1021/cm902640j (2009). 7 Tarancn, A. Strategies for Lowering Solid Oxide Fuel Cells Operating Temperature. Energies 2 1130 1150 (2009). 8 Steele, B. C. H. Materials for IT SOFC stacks: 35 years R&D: the inevitability of gradualness ? Solid State Ionics 134 3 20, doi:10.1016/s0167 2738(00)00709 8 (2000). 9 Inaba, H. & Tagawa, H. Ceria based solid electrolytes Review. Solid State Ionics 83 1 16 (1996). 10 Kharton, V. V., Marques, F. M. B. & Atkinson, A. Transport properties of so lid oxide electrolyte ceramics: a brief review. Solid State Ionics 174 135 149, doi: http://dx.doi.org/10.1016/j.ssi.2004.06.015 (2004). 11 Andersson, D. A., Simak, S. I., Skorodumova, N. V., Abrikosov, I. A. & Johansson, B. Optimization of ionic conducti vity in doped ceria. Proceedings of the National Academy of Sciences of the United States of America 103 3518 3521 (2006). 12 Fuda, K., Kishio, K., Yamauchi, S., Fueki, K. & Onoda, Y. O 17 Nmr Study of Y2o3 Doped Ceo2. J. Phys. Chem. Solids 45 1253 1257 (1984). 13 Kuharuangrong, S. Ionic conductivity of Sm, Gd, Dy and Er doped ceria. Journal of Power Sources 171 506 510 (2007).

PAGE 104

104 14 Li, F. et al. Laser assisted three dimensional atom probe analysis of dopant distribution in Gd doped CeO(2). Scr. Mater. 63 332 335 (2010). 15 Li, F., Ohkubo, T., Chen, Y. M., Kodzuka, M. & Hono, K. Quantitative atom probe analyses of rare earth doped ceria by femtosecond pulsed laser. Ultramicroscopy 111 589 594, doi:10.1016/j.ultramic.2010.12.011 (2011). 16 Omar, S., W achsman, E. D., Jones, J. L. & Nino, J. C. Crystal Structure Ionic Conductivity Relationships in Doped Ceria Systems. Journal of the American Ceramic Society 92 2674 2681 (2009). 17 Ou, D. R. et al. Evidence of intragranular segregation of dopant cations in heavily yttrium doped ceria. Electrochem. Solid State Lett. 10 P1 P3 (2007). 18 Steele, B. C. H. Appraisal of Ce1 yGdyO2 y/2 electrolytes for IT SOFC operation at 500 degrees C. Solid State Ionics 129 95 110 (2000). 19 Omar, S., Wachsman, E. D. & N ino, J. C. A co doping approach towards enhanced ionic conductivity in fluorite based electrolytes. Solid State Ionics 177 3199 3203 (2006). 20 Omar, S., Wachsman, E. D. & Nino, J. C. Higher conductivity Sm3+ and Nd3+ co doped ceria based electrolyte mat erials. Solid State Ionics 178 1890 1897 (2008). 21 Ou, D. R., Ye, F. & Mori, T. Defect clustering and local ordering in rare earth co doped ceria. Physical Chemistry Chemical Physics 13 9554 9560 (2011). 22 van Herle, J., Seneviratne, D. & McEvoy, A. J. Lanthanide co doping of solid electrolytes: AC conductivity behaviour. Journal of the European Ceramic Society 19 837 841 (1999). 23 Wang, F. Y., Wan, B. Z. & Cheng, S. Study on Gd<sup>3+</sup> and Sm<sup>3+ </sup>co doped ceri a based electrolytes. Journal of Solid State Electrochemistry 9 168 173, doi:10.1007/s10008 004 0575 0 (2005). 24 Yamamura, H., Katoh, E. & Kakinuma, K. Electrical conductivity in rare earth cation co doped ceria solid solution systems. Journal of the Ce ramic Society of Japan 110 1021 1024 (2002). 25 Park, K. & Hwang, H. K. Electrical conductivity of Ce0.8Gd0.2 xDyxO2 delta (0 <= x <= 0.2) co doped with Gd3+ and Dy3+ for intermediate temperature solid oxide fuel cells. Journal of Power Sources 196 4996 4999 (2011).

PAGE 105

105 26 Lee, K. T. C o mprehensive D e velopment of H igh P e rformance S o lid O x ide F u el C e lls f or I n termediate a nd L o w T e mperature A p plications PhD thesis, University of Florida, (2010). 27 Chourashiya, M. G., Bharadwaj, S. R. & Jadhav, L. D. Synthesi s and characterization of electrolyte grade 10%Gd doped ceria thin film/ceramic substrate structures for solid oxide fuel cells. Thin Solid Films 519 650 657 (2010). 28 Ding, C. S. et al. Effect of thickness of Gd0.1Ce0.9O1.95 electrolyte films on electr ical performance of anode supported solid oxide fuel cells. Journal of Power Sources 195 5487 5492 (2010). 29 McWhorter, S., Read, C., Ordaz, G. & Stetson, N. Materials based hydrogen storage: Attributes for near term, early market PEM fuel cells. Curren t Opinion in Solid State and Materials Science 15 29 38, doi:10.1016/j.cossms.2011.02.001 (2011). 30 MacLean, H. L. & Lave, L. B. Evaluating automobile fuel/propulsion system technologies. Progress in Energy and Combustion Science 29 1 69, doi:10.1016/s 0360 1285(02)00032 1 (2003). 31 Data, J. I. C. f. D. Vol. 34 0394. 32 Guo, X. & Waser, R. Electrical properties of the grain boundaries of oxygen ion conductors: Acceptor doped zirconia and ceria. Prog. Mater. Sci. 51 151 210 (2006). 33 Data, J. I. C. f. D. Vol. 10 0173. 34 Data, J. I. C. f. D. Vol. 4 802. 35 Barsoukov, E. & Macdonald, J. R. Impedance Spectroscopy: Theory, Experiment, and Applications (Wiley, 2005). 36 Kidner, N. J., Perry, N. H., Mason, T. O. & Garboczi, E. J. The brick layer model revisited: introducing the nano grain composite model. Journal of the American Ceramic Society 91 1733 1746 (2008). 37 Gobel, M. C., Gregori, G., Guo, X. X. & Maier, J. Boundary effects on the electrical conductivity of pure and doped cerium oxide thin films. Physical Chemistry Chemical Physics 12 14351 14361 (2010). 38 Beckel, D. et al. Thin films for micro solid oxide fuel cells. Journal of Power Sources 173 325 345, doi: http://dx.doi.org/10.1016/j.jpowsour.2007.04.070 (2007).

PAGE 106

106 39 Eason, R. Pul sed Laser Deposition of Thin Films: Applications Led Growth of Functional Materials (Wiley Interscience, 2006). 40 Willmott, P. R., Spillmann, H. & Huber, J. R. Design and synthesis of novel thin films and structures by reactive crossed beam laser ablat ion. Journal of Materials Chemistry 12 397 402 (2002). 41 Smith, H. M. & Turner, A. F. Vacuum Deposited Thin Films Using a Ruby Laser. Appl. Optics 4 147 & (1965). 42 Ashfold, M. N. R., Claeyssens, F., Fuge, G. M. & Henley, S. J. Pulsed laser ablation and deposition of thin films. Chemical Society Reviews 33 23 31 (2004). 43 Boyd, I. W. Thin film growth by pulsed laser deposition. Ceramics International 22 429 434, doi:10.1016/0272 8842(95)00086 0 (1996). 44 Buerle, D. et al. Pulsed laser depositio n. Applied Physics A: Materials Science & Processing 69 S48 (1999). 45 Christen, H. M. & Eres, G. Recent advances in pulsed laser deposition of complex oxides. Journal of Physics: Condensed Matter 20 264005 (2008). 46 Ohring, M. Materials Science of Th in Films (Elsevier Science, 2001). 47 Chrisey, D. B. & Hubler, G. K. Pulsed Laser Deposition of Thin Films (John Wiley & Sons, 1994). 48 Singh, R. K., Holland, O. W. & Narayan, J. Theoretical Model for Deposition of Superconducting Thin Films Using P ulsed Laser Evaporation Technique. J. Appl. Phys. 68 233 247 (1990). 49 Panzner, M., Dietsch, R., Holz, T., Mai, H. & Vllmar, S. Scale up of pulsed laser wafer coating. Applied Surface Science 96 98 643 648, doi:10.1016/0169 433 2(95)00536 6 (1996). 50 Greer, J. A. & Tabat, M. D. Large Area Pulsed Laser Deposition Techniques and Applications. J. Vac. Sci. Technol. A Vac. Surf. Films 13 1175 1181 (1995). 51 Dam, B., Rector, J. H., Johansson, J., Kars, S. & Griessen, R. Stoichi ometric transfer of complex oxides by pulsed laser deposition. Applied Surface Science 96 8 679 684 (1996). 52 Arnold, C. B. & Aziz, M. J. Stoichiometry issues in pulsed laser deposition of alloys grown from multicomponent targets. Applied Physics a Mate rials Science & Processing 69 S23 S27 (1999).

PAGE 107

107 53 Ohring, M. The Materials Science of Thin Films (Acad. Press, 1992). 54 Stokes, R. J. & Evans, D. F. Fundamentals of Interfacial Engineering (Wiley, 1997). 55 Thornton, J. A. High Rate Thick Film Grow th. Annu. Rev. Mater. Sci. 7 239 260 (1977). 56 Kelly, P. J. & Arnell, R. D. Magnetron sputtering: a review of recent developments and applications. Vacuum 56 159 172 (2000). 57 Nefedov, A., Abromeit, A., Morawe, C. & Stierle, A. High resolution x ray scattering study of platinum thin films on sapphire. J. Phys. Condes. Matter 10 717 730 (1998). 58 Nelson, J. B. & Riley, D. P. An Experimental Investigation of Extrapolation Methods in the Derivation of Accurate Unit Cell Dimensions of Crystals. Proceed ings of the Physical Society of London 57 160 177 (1945). 59 Giannuzzi, L. A. & Stevie, F. A. A review of focused ion beam milling techniques for TEM specimen preparation. Micron 30 197 204 (1999). 60 Kempshall, B. W., Giannuzzi, L. A., Prenitzer, B. I ., Stevie, F. A. & Da, S. X. Comparative evaluation of protective coatings and focused ion beam chemical vapor deposition processes. J. Vac. Sci. Technol. B 20 286 290 (2002). 61 McHargue, C. J. et al. Etching of amorphous Al2O3 produced by ion implantat ion. Nucl. Instrum. Methods Phys. Res. Sect. B Beam Interact. Mater. Atoms 127 596 598 (1997). 62 Wucherer, L. Synthesis and characterization of electroceramic foams for 3 3 piezoelectric composites [electronic resource] / by Laurel Wucherer (Universit y of Florida, 2008). 63 Kidner, N. J., Homrighaus, Z. J., Ingram, B. J., Mason, T. O. & Garboczi, E. J. Impedance/dielectric spectroscopy of electroceramics Part 1: Evaluation of composite models for polycrystalline ceramics. J. Electroceram. 14 283 29 1 (2005). 64 Bellino, M. G., Lamas, D. G. & de Reca, N. E. W. Enhanced ionic conductivity in nanostructured, heavily doped ceria ceramics. Adv. Funct. Mater. 16 107 113 (2006). 65 Rupp, J. L. M. & Gauckler, L. J. Microstructures and electrical conductiv ity of nanocrystalline ceria based thin films. Solid State Ionics 177 2513 2518 (2006).

PAGE 108

108 66 Tuller, H. L. Ionic conduction in nanocrystalline materials. Solid State Ionics 131 143 157, doi:10.1016/s0167 2738(00)00629 9 (2000). 67 Swanson, M., Sunder, M., Tangtrakarn, N., Krishna, L. & Moran, P. D. Impact of the surface/interface on the ionic conductivity of thin film Gd doped CeO(2). Solid State Ionics 189 45 49 (2011). 68 Huang, H., Gur, T. M., Saito, Y. & Prinz, F. High ionic conductivity in ultrathin nanocrystalline gadolinia doped ceria films. Applied Physics Letters 89 (2006). 69 Rodrigo, K. et al. Electrical characterization of gadolinia doped ceria films grown by pulsed laser deposition. Applied Physics a Materials Science & Processing 101 601 6 07 (2010). 70 Joo, J. H. & Choi, G. M. Electrical conductivity of thin film ceria grown by pulsed laser deposition. Journal of the European Ceramic Society 27 4273 4277 (2007). 71 Abid, M., Paul Boncour, V. & Touroude, R. Pt/CeO2 catalysts in crotonalde hyde hydrogenation: Selectivity, metal particle size and SMSI states. Appl. Catal. A Gen. 297 48 59 (2006). 72 Liu, W., Li, B., Liu, H. Q. & Pan, W. Electrical conductivity of textured Sm(3+) and Nd(3+) Co doped CeO(2) thin film electrolyte. Electrochim. Acta 56 3334 3337 (2011). 73 Tsuchiya, M., Bojarczuk, N. A., Guha, S. & Ramanathan, S. Microstructural effects on electrical conductivity relaxation in nanoscale ceria thin films. J. Chem. Phys. 130 (2009). 74 Kilner, J. A. & Waters, C. D. The Effects of Dopant Cation Oxygen Vacancy Complexes on the Anion Transport Properties of Nonstoichiometric Fluorite Oxides. Solid State Ionics 6 253 259 (1982). 75 Dholabhai, P. P., Adams, J. B., Crozier, P. A. & Sharma, R. In search of enhanced electrolyte materi als: a case study of doubly doped ceria. Journal of Materials Chemistry 21 18991 18997 (2011). 76 Guan, X. F., Zhou, H. P., Wang, Y. & Zhang, J. Preparation and properties of Gd3+ and Y3+ co doped ceria based electrolytes for intermediate temperature sol id oxide fuel cells. Journal of Alloys and Compounds 464 310 316 (2008). 77 Sha, X. Q. et al. Study on La and Y co doped ceria based electrolyte materials. Journal of Alloys and Compounds 428 59 64 (2007). 78 Zheng, Y. F. et al. Effect of Dy on the pro perties of Sm doped ceria electrolyte for IT SOFCs. Journal of Alloys and Compounds 509 1244 1248 (2011).

PAGE 109

109 79 Dikmen, S. Effect of co doping with Sm(3+), Bi(3+), La(3+), and Nd(3+) on the electrochemical properties of hydrothermally prepared gadolinium dop ed ceria ceramics. Journal of Alloys and Compounds 491 106 112 (2010). 80 Zheng, Y. F. et al. The effect of Sr on the properties of Y doped ceria electrolyte for IT SOFCs. Journal of Alloys and Compounds 486 586 589 (2009). 81 Singh, N., Singh, N. K., Kumar, D. & Parkash, O. Effect of co doping of Mg and La on conductivity of ceria. Journal of Alloys and Compounds 519 129 135 (2012). 82 Shannon, R. D. Revised Effective Ionic Radii and Systematic Studies of Interatomic Distances in Halides and Chalcoge nides. Acta Crystallographica Section A 32 751 767 (1976). 83 Vegard, L. & Dale, H. Tests on mixed crystals and alloys. Zeitschrift Fur Kristallographie 67 148 162 (1928). 84 Tian, C. in MRS Fall Meeting. 85 Brugner, F. S. & Blumenth.Rn. Electrical Co nductivity of Single Crystal Ceo2. Journal of the American Ceramic Society 54 57 & (1971). 86 Yamamura, H. et al. Multiple doping effect on the electrical conductivity in the (Ce(1 x y)La(x)My)O2 delta (M = Ca, Sr) system. Electrochemistry 68 455 459 (2 000). 87 Mori, T., Ikegami, T. & Yamamura, H. Application of a crystallographic index for improvement of the electrolytic properties of the CeO2 Sm2O3 system. J. Electrochem. Soc. 146 4380 4385 (1999). 88 Kilner, J. A. & Brook, R. J. A Study of Oxygen I on Conductivity in Doped Nonstoichiometric Oxides. Solid State Ionics 6 237 252 (1982). 89 Li, L. et al. Ionic conductivity across the disorder order phase transition in the NdO1.5 CeO2 system. Solid State Ionics 221 15 21, doi:10.1016/j.ssi.2012.06.007 (2012). 90 Zhang, T. S., Ma, J., Kong, L. B., Hing, P. & Kilner, J. A. Preparation and mechanical properties of dense Ce0.8W0.2O2 delta ceramics. Solid State Ionics 167 191 196 (2004). 91 Ou, D. R. et al. Microstructures and electrolytic properties of yttrium doped ceria electrolytes: Dopant concentration and grain size dependences. Acta Materialia 54 3737 3746 (2006).

PAGE 110

110 92 Ou, D. R. et al. Oxygen vacancy ordering in heavily rare earth doped ceria. Applied Physics Letters 89 171911 171913 (2006). 93 Zh ang, T. S., Ma, J., Kong, L. B., Chan, S. H. & Kilner, J. A. Aging behavior and ionic conductivity of ceria based ceramics: a comparative study. Solid State Ionics 170 209 217 (2004). 94 Mori, T. et al. Influence of particle morphology on nanostructural feature and conducting property in Sm doped CeO2 sintered body. Solid State Ionics 175 641 649 (2004). 95 Mori, T., Drennan, J., Lee, J. H., Li, J. G. & Ikegami, T. Oxide ionic conductivity and microstructures of Sm or La doped CeO2 based systems. Solid State Ionics 154 461 466 (2002). 96 Vanderah, T. in Electronic Materials and Applications (Orlando, FL, 2011). 97 Chavan, S. V., Mathews, M. D. & Tyagi, A. K. Phase relations and thermal expansion studies in the CeO2 NdO1.5 system. Materials Researc h Bulletin 40 1558 1568 (2005). 98 Parks, T. C. & Bevan, D. J. M. Mixed Oxides of Type Mo2 (Fluorite) Mo1.5 .6. Phase Relationships in Systems Ceo2 Smo1.5 and Ceo2 Ndo1.5. Revue De Chimie Minerale 10 15 28 (1973). 99 Li, B., Liu, Y., Wei, X. & Pan, W. Electrical properties of ceria Co doped with Sm3+ and Nd3+. Journal of Power Sources 195 969 976 (2010). 100 Liu, Y., Li, B., Wei, X. & Pan, W. Citric Nitrate Combustion Synthesis and Electrical Conductivity of the Sm3+ and Nd3+ Co Doped Ceria Electrolyt e. Journal of the American Ceramic Society 91 3926 3930 (2008). 101 Dikmen, S., Aslanbay, H., Dikmen, E. & Sahin, O. Hydrothermal preparation and electrochemical properties of Gd3+ and Bi3+, Sm3+, La3+, and Nd3+ codoped ceria based electrolytes for inter mediate temperature solid oxide fuel cell. Journal of Power Sources 195 2488 2495 (2010). 102 Stephens, I. E. L. & Kilner, J. A. Ionic conductivity of Ce1 xNdxO2 x/2. Solid State Ionics 177 669 676 (2006). 103 Aneflous, L., Musso, J. A., Villain, S., G avarri, J. R. & Benyaich, H. Effects of temperature and Nd composition on non linear transport properties in substituted Ce1 xNdxO2 delta cerium dioxides. Journal of Solid State Chemistry 177 856 865 (2004).

PAGE 111

111 104 Fu, Y. P. & Chen, S. H. Preparation and cha racterization of neodymium doped ceria electrolyte materials for solid oxide fuel cells. Ceramics International 36 483 490 (2010). 105 Grover, V., Achary, S. N. & Tyagi, A. K. Structural analysis of excess anion C type rare earth oxide: a case study with Gd1 xCexO1.5+x/2 (X = 0.20 and 0.40). Journal of Applied Crystallography 36 1082 1084 (2003). 106 Chakraborty, K. R., Krishna, P. S. R., Chavan, S. V. & Tyagi, A. K. A neutron diffraction study on ceria neodia solid solutions. Powder Diffraction 21 36 39 (2006). 107 Stephens, P. W. Phenomenological model of anisotropic peak broadening in powder diffraction. Journal of Applied Crystallography 32 281 289 (1999). 108 ASTM. (1996). 109 Heyn, E. The Metallographist 5 39 64 (1903). 110 Chourashiya, M. G., Patil, J. Y., Pawar, S. H. & Jadhav, L. D. Studies on structural, morphological and electrical properties of Ce1 xGdxO2 (x/2). Materials Chemistry and Physics 109 39 44 (2008). 111 Etsell, T. H. & Flengas, S. N. Electrical Properties of Solid Oxid e Electrolytes. Chemical Reviews 70 339 & (1970). 112 Balazs, G. B. & Glass, R. S. Ac Impedance Studies of Rare Earth Oxide Doped Ceria. Solid State Ionics 76 155 162 (1995). 113 Reis, S. L., Souza, E. C. C. & Muccillo, E. N. S. Solid solution formatio n, densification and ionic conductivity of Gd and Sm doped ceria. Solid State Ionics 192 172 175 (2011). 114 Perez Coll, D., Sanchez Lopez, E. & Mather, G. C. Influence of porosity on the bulk and grain boundary electrical properties of Gd doped ceria. Solid State Ionics 181 1033 1042 (2010). 115 Harwig, H. A. & Gerards, A. G. Electrical Properties of Alpha, Beta, Gamma and Delta Phases of Bismuth Sesquioxide. Journal of Solid State Chemistry 26 265 274 (1978). 116 Yashima, M. & Ishimura, D. Crystal structure and disorder of the fast oxide ion conductor cubic Bi2O3. Chemical Physics Letters 378 395 399 (2003). 117 Bevan. in 4th Conference on Rare Earths Research. P441.

PAGE 112

112 118 Ye, F. et al. Simulation of ordering in large defect clusters in gadolinium doped ceria. Solid State Ionics 179 1962 1967 (2008). 119 Toshiyuki Mori, J. D., Yarong Wang, Graeme Auchterlonie, Ji Guang Li, Anya Yago. Influence of nano structural feature on electrolytic properties in Y2O3 doped CeO2 system. Science and Technology o f Advanced Materials 4 213 (2003). 120 Omar, S. Enhanced Ionic Conductivity of Ceria based Compounds for the Electrolyte Application in SOFCS Ph.D. thesis, University of Florida, (2008). 121 Bachelet, R. et al. Enhanced thermal stability of Pt electrode s for flat epitaxial biferroic YMnO3/Pt heterostructures. Applied Physics Letters 95 (2009).

PAGE 113

113 BIOGRAPHICAL SKETCH Robert Kasse was born at Fort Wainwright Alaska in 1989. He has since lived in North Carolina, Indiana, Georgia, Pennsylvania, and Germany before ending up in Florida where he graduated from Niceville Senior High School in 2008. Robert began attending the University of Florida the following August and graduated with a B.S. in Materials Science and Engineering specializing in ceramics, and a minor in Business Administration in May of 2012. During the summer of 2011 Rob ert was awarded the SMART S cholarship for Service by the Department of Defense, with the 96 th Test Wing at Eglin Air Force Base as his sponsoring facility. He will begin work ing at Eglin in fulfillment of his two year obligation shortly after graduating with his m