<%BANNER%>

Crytallochemical avenues for enhancement of proton conduction in ceramics

Permanent Link: http://ufdc.ufl.edu/UFE0042525/00001

Material Information

Title: Crytallochemical avenues for enhancement of proton conduction in ceramics phosphate compounds, experiments and simulation
Physical Description: 1 online resource (184 p.)
Language: english
Creator: Phadke, Satyajit
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

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

Notes

Abstract: Current proton exchange membranes (PEMs) are based on polymeric materials such as perfluorosulphonic acid (Nafion ) which have an upper limit to their temperature of operation (approximately 100oC). To overcome this limitation ceramic PEM materials are being investigated for transportation applications, where operating temperatures in the range of 200oC 600oC are desired. In this study, the conductivity behavior and electrical properties of three families of phosphate compounds (namely SnP2O7, LaP5O14, and LaPO4) have been studied in order to better understand crystal structure - proton conduction relationships in this class of ceramic materials. From the phosphate-rich region of the La2O3 - P2O5 binary phase diagram, the orthophosphate (LaPO4) and ultraphosphate (LaP5O14) materials are chosen for a comparative study of the electrical properties. The conductivity of the lanthanum phosphates (doped and undoped) is measured using impedance spectroscopy in the temperature range 300oC 600oC. The conductivity of 5 mol% Sr2+ doped LaP5O14 (1.01 x 10-4 S/cm, 600oC) is found to be an order of magnitude higher than similarly doped LaPO4 (7.00 x 10-6 S/cm, 600oC) which is a well investigated proton conducting material. In addition, it is observed that the activation energy for protonic conduction is much lower for doped LaP5O14 (0.80 plus or minus 0.01 eV) as compared to LaPO4 (1.09 plus or minus 0.01 eV). A hypothesis relating the oxygen to oxygen ion distance in a material to the activation energy for proton conduction is presented and the experimental results obtained are critically examined on the basis of the hypothesis and other relevant literature. From this analysis it is shown that the condensed nature of the phosphate anion in LaP5O14 can provide low energy avenues for proton transport within the material leading to enhanced conductivity. In addition, the electrical properties of undoped and acceptor doped tin pyrophosphate (SnP2O7) are investigated, which is regarded as a potential candidate electrolyte material for PEMFCs. Within the dopants analyzed, Zn0.1Sn0.9P2O7-? shows the highest conductivity of 2.84 x 10-6 S/cm at 600oC. However, this value is several orders of magnitude lower than previously reported for doped tin pyrophosphate compounds but it is consistent with a recent publication. The reason for the discrepancy is investigated and a possible explanation has been proposed based on results obtained from 31P MAS NMR spectroscopy. It is observed that the as-calcined powder of doped SnP2O7 contained a significant fraction of residual phosphoric acid from synthesis procedure which is likely to be responsible for the high conductivity values for acceptor doped SnP2O7 reported previously by other research groups.
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 Satyajit Phadke.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Nino, Juan C.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0042525:00001

Permanent Link: http://ufdc.ufl.edu/UFE0042525/00001

Material Information

Title: Crytallochemical avenues for enhancement of proton conduction in ceramics phosphate compounds, experiments and simulation
Physical Description: 1 online resource (184 p.)
Language: english
Creator: Phadke, Satyajit
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

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

Notes

Abstract: Current proton exchange membranes (PEMs) are based on polymeric materials such as perfluorosulphonic acid (Nafion ) which have an upper limit to their temperature of operation (approximately 100oC). To overcome this limitation ceramic PEM materials are being investigated for transportation applications, where operating temperatures in the range of 200oC 600oC are desired. In this study, the conductivity behavior and electrical properties of three families of phosphate compounds (namely SnP2O7, LaP5O14, and LaPO4) have been studied in order to better understand crystal structure - proton conduction relationships in this class of ceramic materials. From the phosphate-rich region of the La2O3 - P2O5 binary phase diagram, the orthophosphate (LaPO4) and ultraphosphate (LaP5O14) materials are chosen for a comparative study of the electrical properties. The conductivity of the lanthanum phosphates (doped and undoped) is measured using impedance spectroscopy in the temperature range 300oC 600oC. The conductivity of 5 mol% Sr2+ doped LaP5O14 (1.01 x 10-4 S/cm, 600oC) is found to be an order of magnitude higher than similarly doped LaPO4 (7.00 x 10-6 S/cm, 600oC) which is a well investigated proton conducting material. In addition, it is observed that the activation energy for protonic conduction is much lower for doped LaP5O14 (0.80 plus or minus 0.01 eV) as compared to LaPO4 (1.09 plus or minus 0.01 eV). A hypothesis relating the oxygen to oxygen ion distance in a material to the activation energy for proton conduction is presented and the experimental results obtained are critically examined on the basis of the hypothesis and other relevant literature. From this analysis it is shown that the condensed nature of the phosphate anion in LaP5O14 can provide low energy avenues for proton transport within the material leading to enhanced conductivity. In addition, the electrical properties of undoped and acceptor doped tin pyrophosphate (SnP2O7) are investigated, which is regarded as a potential candidate electrolyte material for PEMFCs. Within the dopants analyzed, Zn0.1Sn0.9P2O7-? shows the highest conductivity of 2.84 x 10-6 S/cm at 600oC. However, this value is several orders of magnitude lower than previously reported for doped tin pyrophosphate compounds but it is consistent with a recent publication. The reason for the discrepancy is investigated and a possible explanation has been proposed based on results obtained from 31P MAS NMR spectroscopy. It is observed that the as-calcined powder of doped SnP2O7 contained a significant fraction of residual phosphoric acid from synthesis procedure which is likely to be responsible for the high conductivity values for acceptor doped SnP2O7 reported previously by other research groups.
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 Satyajit Phadke.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Nino, Juan C.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0042525:00001


This item has the following downloads:


Full Text

PAGE 1

1 CRYSTALLOCHEMICAL AVENUES FOR THE ENHANCEMENT OF PROTON CONDUCTION IN CERAMICS: PHOSPHATE COMPOUNDS, EXPERIMENTS AND SIMULATION By SATYAJIT PHADKE A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010

PAGE 2

2 2010 S a tyajit Phadke

PAGE 3

3 To my parents

PAGE 4

4 ACKNOWLEDGMENTS I would like to thank my advisor, Prof. Juan C. Nino for his guidance, co nstant support and encouragement which enabled me to perform to the best of my ability. I would also like to thank him for the fruitful scientific discussions that I had with him during the course of my study. I would like express gratitude towards my com mittee members namely, Prof. Eric D. Wachsman, Prof. Susan Sinnott, Prof. Jacob L. Jones and Prof. Aravind R. Asthagiri for their valuable con tributions at various occasions. I would like to thank Prof. Russ Bowers for his patient guidance and readiness t o help whenever it was required. I would also like to thank Prof. M. Saiful Islam for providing me with the resources and guidance to perform research in the field of simulation. I would like to thank Prof. Scott Perry and Prof. Kevin Jones for agreeing to act as s ubstitute committee members I would like to thank my current and previous group members and colleagues for the innumerable fruitful scientific discussions, for providing a friendly and cordial atmosphere at work as well as for keeping me ente rtained during tough times, namely Wei Qiu Lu Cai, Marta Giachino, Shobit Omar, Peng Xu Tak keun Oh, Andres Molina, Alex Arias, Roberto, Nathan, Robert, Byron, Kevin Tierney, Adam Wilk, Donald Moore and any others that I may have missed. I would like to thank my friends Sonalika Hrishikesh, Shruti, Aniket, Anirban and Ravi for keeping me entertained throughout my stay in Gainesville as well as for providing support and encouragement at times when I needed it. In the end I would like to thank my paren ts without who se guidance nothing would have been possible. They have always instilled me in the virtues of honesty, humility, hard work and patience and have always led by example rather than by instruction,

PAGE 5

5 which has provided me the strength to endure t he toughest times and will continue to do so in the future

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 9 LIST OF FIGURES ................................ ................................ ................................ ........ 11 ABSTRACT ................................ ................................ ................................ ................... 15 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 17 1.1 Statement of Problem and Motivation ................................ ............................... 17 1.2 Scientific Approach ................................ ................................ ........................... 18 1.3 Organization of the Dissertation ................................ ................................ ........ 20 2 BACKGROUND ................................ ................................ ................................ ...... 23 2.1 Solid State Electrochemistry ................................ ................................ ............. 23 2.2 Proto n Exchange Membrane Fuel Cells (PEMFCs) ................................ .......... 25 2.2.1 Introduction ................................ ................................ .............................. 25 2.2.2 Hydrogen Fuel Cell Unit ................................ ................................ .......... 26 2.2.3 Potential Applications ................................ ................................ .............. 27 2.3 Crystalchemistry of Condensed Phosphates ................................ .................... 27 2.3.1 Structur al Comparison with the Silicates ................................ ................. 27 2.3.2 Temperature Dependent Equilibrium in Fused Phosphoric Acid ............. 29 2.3.3 Structural Classif ication of Phosphate Compounds ................................ 30 2.3.4 La 2 O 3 P 2 O 5 Binary Phase Diagram ................................ ...................... 32 2.4 Electrical Characterization ................................ ................................ ................ 34 2.4.1 Impedance Spectroscopy ................................ ................................ ........ 34 2.4.2 Transference Number Measurements ................................ ..................... 39 3 MATE RIALS AND EXPERIMENTAL PROCEDURES ................................ ............ 43 3.1 Synthesis of Materials ................................ ................................ ....................... 43 3.1.1 Lanthanum Ultraphosphate (LaP 5 O 14 ) ................................ ..................... 43 3.1.2 Lanthanum Orthophosphate (LaPO 4 ) ................................ ...................... 45 3.1.3 Tin (IV) Pyrophosphate (SnP 2 O 7 ) ................................ ............................ 47 3.2 Phase Analysis ................................ ................................ ................................ 49 3.2.1 X Ray Diffraction (XRD) ................................ ................................ .......... 49 3.2.2 Energy Dispersive Spectroscopy (EDS) ................................ .................. 50 3.2.3 Scanning Electron Microscopy (SEM) ................................ ..................... 51 3.2.4 Magic Angle Spinning Nuclear Magnetic Resonance (MAS NMR) .... 51 3.3 Electrochemical Impedance Spectroscopy (EIS) and Proton Transference Number Measurement ................................ ................................ ......................... 52

PAGE 7

7 3.3.1 Experimental Setup ................................ ................................ ................. 52 3.3.2 Sample Preparation and EIS ................................ ................................ ... 55 3.3.3 Transference Number Measurement ................................ ....................... 55 3.4 Simulation Methods ................................ ................................ .......................... 57 3.4.1 Potential Model ................................ ................................ ........................ 57 3.4.2 Buckingham Potential ................................ ................................ .............. 57 3.4.3 Harmoni c Angle Bending Potential ................................ .......................... 59 3.4.4 Ionic Polarization (Shell Model) ................................ ............................... 59 3.4.5 Defect Simulation (Mott Littleton Approach) ................................ ......... 61 3.4.6 Proton Oxygen (O H) Interaction ................................ ....................... 62 3.4.7 Energy Minimization ................................ ................................ ................ 63 4 CHARACTERIZATION OF LANTHANUM ULTRAPHOSPHATE (LaP 5 O 14 ) ........... 65 4.1 Ultraphosphates ................................ ................................ ................................ 65 4.1 Crystallochemical Analysis ................................ ................................ ............... 67 4.2 Phase Analysis ................................ ................................ ................................ 73 4.2.1 X Ray Diffraction (XRD) ................................ ................................ .......... 73 4.2.2 Scanni ng Electron Microscopy (SEM) ................................ ..................... 74 4.2.3 Energy Dispersive Spectroscopy (EDS) ................................ .................. 75 4.2.4 Single Crystal X Ray Diffraction ................................ .............................. 75 4.3 Conductivity Measurement ................................ ................................ ............... 77 4.3.1 Effect of Dopant Content ................................ ................................ ......... 77 4.3.2 Effect of Atmospheric Humidity ................................ ............................... 81 4.4 Atomistic Simulation ................................ ................................ .......................... 83 4.4.1 Structural Modeling ................................ ................................ .................. 83 4.4.2 Intrinsic Atomic Defect Formation ................................ ............................ 85 4.4.3 Dopant Incorporation ................................ ................................ ............... 87 4.4.4 Proton Incorporation a nd O H Configuration ................................ ........ 89 4.5 Other Materials ................................ ................................ ................................ 92 4.6 Summary and Conclusion ................................ ................................ ................. 93 5 CHARACTERIZATION OF LANTHANUM ORTHOPHOSPHATE (LaPO 4 ) ............ 95 5.1 Orthophosphates ................................ ................................ .............................. 95 5.2 Crystallochemical Analysis ................................ ................................ ............... 95 5.3 Phase Analysis ................................ ................................ ................................ 99 5.3.1 X Ray Diffraction (XRD) ................................ ................................ .......... 99 5.3.2 S canning Electron Microscopy (SEM) ................................ ................... 100 5.4 Conductivity Measurement ................................ ................................ ............. 101 5.4.1 Effect of Dopant Content ................................ ................................ ....... 104 5.4.2 Humidified Atmosphere Conductivity Measurement .............................. 106 5.5 Atomistic Simulation ................................ ................................ ........................ 109 5.5.1 Structural Modeling ................................ ................................ ................ 109 5.5.2 Intrinsic Atomic Defect Formation ................................ .......................... 111 5.5.3 Dopant Incorporation ................................ ................................ ............. 112 5.5.4 Proton Incorporation and Defect Chemistry ................................ ........... 114

PAGE 8

8 5.6 Summary and Conclusion ................................ ................................ ............... 118 6 CO MPARISON OF LaPO 4 AND LaP 5 O 14 ................................ ............................. 120 6.1 Factors Affecting Protonic Conductivity ................................ .......................... 120 6.2 Effect of Inter Oxygen Distance on Activat ion Energy ................................ ..... 122 6.3 Comparison of Conductivity and Model for Proton Conduction ....................... 125 7 CHARACTERIZATION OF TIN (IV) PYROPHOSPHATE ( SnP 2 O 7 ) ..................... 132 7.1 Pyrophosphates ................................ ................................ .............................. 132 7.2 Crystal Structure ................................ ................................ ............................. 133 7.3 X Ray Diffraction (XRD) ................................ ................................ .................. 136 7.4 Scanning Electron Microscopy (SEM) ................................ ............................. 138 7.5 Conductivity Measurement ................................ ................................ ............. 139 7.5.1 Comparison with Previously Published Data ................................ ......... 139 7.5.2 31 P MAS NMR Spectroscopy ................................ .............................. 142 7.5.3 Acceptor Doped SnP 2 O 7 (D = In, Mg, Sc, Zn, and Ga) .......................... 143 7.6 Proton Transference Number Measurement ................................ ................... 147 7.6 Summary and Co nclusion ................................ ................................ ............... 149 8 SUMMARY AND FUTURE WORK ................................ ................................ ....... 150 8.1 Summary ................................ ................................ ................................ ........ 150 8.2 F uture Work ................................ ................................ ................................ .... 154 8.2.1 Other Ultraphosphates ................................ ................................ .......... 1 54 8.2.2 Other Systems (Germanates, Niobates, Borates and Arsenates) ......... 155 8.2.3 Condensed Hydrogen Phosphates ................................ ........................ 156 8.2.4 Growth of Larger Crystals ................................ ................................ ...... 157 8.2.5 Performance Testing ................................ ................................ ............. 159 APPENDIX A STRUCTURAL CLASSIFICATION OF THE SILICATES ................................ ...... 160 B SYNTHESIS OF LANTHANUM METAPHOSPHATE La(PO 3 ) 3 ......................... 163 C IMPEDANCE CALCULATION FOR A RESISTANCE CAPACITANCE (R C) CIRCUIT ................................ ................................ ................................ ............... 168 D STRUCTURAL PARAMETERS AND ATOMIC POS ITIONS FOR LaPO 4 and LaP 5 O 14 ................................ ................................ ................................ ................. 170 LIST OF REFERENCES ................................ ................................ ............................. 172 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 184

PAGE 9

9 LIST OF TABLES Table page 2 1 Classification of fuel cells. ................................ ................................ .................. 24 2 2 Structural classification of silicates ................................ ................................ ..... 28 3 1 Interatomic potential parameters for LaPO 4 and LaP 5 O 14 ................................ .. 58 3 2 Shell model parameters for LaPO 4 and LaP 5 O 14 ................................ ............... 61 3 3 Parameters for the O H interaction ................................ ................................ ..... 62 4 1 Classification of ultraphosphate compounds with examples. .............................. 66 4 2 Comparison of mean inter and intra tetrahedral oxygen oxygen ion distances in LaPO 4 and LaP 5 O 14 ................................ ................................ ....... 71 4 3 Results from quantitative EDS analysis on the undoped and 1 mol% Sr doped L aP 5 O 14 samples. ................................ ................................ .................... 75 4 4 Peak frequencies and capacitances associated with bulk and electrode impedance for all compositions of LaP 5 O 14 ................................ ........................ 78 4 5 Measured conductivity at 600 o C and calculated activation energy for three different compositions of LaP 5 O 14 ................................ ................................ ...... 79 4 6 Experimental and calculated lattice parameters for LaP 5 O 14 ............................. 83 4 7 Experimental and calculated average bond lengths for LaP 5 O 14 ....................... 84 4 8 Experimental and calculated bond angles for LaP 5 O 14 ................................ ...... 85 4 9 Calculated energies for intrinsic material defects (Frenkel and Schottky) in LaP 5 O 14 ................................ ................................ ................................ ............. 86 4 10 Proton dopant interaction in LaP 5 O 14 ................................ ................................ 91 5 1 Normalized capacitances associated with bulk and grain boundary impedance for all compositions of LaPO 4 ................................ ........................ 103 5 2 Measured conductivity at 600 o C and calculated activation energy for three different compositions of LaPO 4 (unhumidified atmosphere). ........................... 106 5 3 Measured conductivity at 600 o C and calculated activ ation energy for three different compositions of LaPO 4 under humidified atmosphere. ....................... 106 5 4 Comparison of experimental and calculated lattice parameters for LaPO 4 ...... 109

PAGE 10

10 5 5 Comparison of the experimental bond distances and bond angles in LaPO 4 with the calculated values. ................................ ................................ ................ 110 5 6 Calculated energies of Frenkel and Schott ky type intrinsic defects in LaPO 4 .. 111 5 7 Interatomic distances in the pyrophosphate ion, orthophosphate ion and hydrogen phosphate ion in LaPO 4 ................................ ................................ ... 116 5 8 Binding energies and other relevant parameters of proton dopant interaction in LaPO 4 ................................ ................................ ................................ .......... 117 6 1 Obtained activation energies for three different compositions of LaPO 4 and LaP 5 O 14 ................................ ................................ ................................ ........... 127 6 2 Comparison of mean inter and intra tetrahedral oxygen oxygen ion distances in LaPO 4 and LaP 5 O 14 ................................ ................................ ..... 129 7 1 Comparison of lattice parameters for the various doped and undoped materials. ................................ ................................ ................................ .......... 137 7 2 Activation energies obtained for In 3+ doped samples as measured in this study compared with previously pu blished literature. ................................ ....... 141 A 1 Structural classification of the silicates. ................................ ............................ 161 D 1 Atomic positions for lanthanum ultraphosphate (L aP 5 O 14 ). .............................. 170 D 2 Atomic positions of lanthanum orthophosphate (LaPO 4 ). ................................ 171

PAGE 11

11 LIST OF FIGURES Figure page 1 1 Comparison of conductivities of proton conducting materials. 8 ........................... 19 2 1 Individual fuel cell unit of a hydrogen fuel cell ................................ .................... 26 2 2 Thermal dehydration of phosphoric acid leading to formation of poly phosphoric acids of higher complexity. ................................ ............................... 30 2 3 General structural classification of condensed phosphates. ............................... 31 2 4 Binary phase diagram for the system La 2 O 3 P 2 O 5 ................................ .......... 33 2 5 Nyquist plots for (a) R and C, (b), single R C circuit, (c) p artially resolved arcs, and (d) resolved arcs for two R C circuits in series. ................................ .. 37 2 6 Nyquist plots corresponding to single R CPE circuitsFor the CPE, alpha is varied from 0.90 to 1. ................................ ................................ .......................... 38 3 1 (a) Flowchart for synthesis of single crystals of LaP 5 O 14 (b) vitreous carbon crucible, and (c) setup used for dissolution of polyphosphoric acid. ................... 44 3 2 Flowchart for powder synthesis, sample preparation and the various characterization techniques used. ................................ ................................ ...... 47 3 3 (a) Orientation of LaP 5 O 14 crystals during single crystal XRD data collection (b) SEM image of an LaP 5 O 14 crystal showing the (100) plane. ......................... 50 3 4 Setup used for conductivity measurement and proton transference number measurement. ................................ ................................ ................................ ..... 52 3 5 (a) Schematic diagram and (b) picture of quartz reactor used for conductivity measurement and transference number measurement. ................................ ..... 53 3 6 Prepar ed samples of (a) SnP 2 O 7 (b) LaP 5 O 14 and (c) sealed sample of SnP 2 O 7 for proton transference number measurement. ................................ ..... 54 3 7 Diagrammatic representation of shell model parameters. ................................ .. 60 4 1 View of LaP 5 O 14 along the [001] direction. La 3+ ions shown in orange, PO 4 tetrahedra shown in green and oxygen ions are shown in red. .......................... 67 4 2 Eight coordinated La 3+ ion (shown in orange). The eight oxygen ions belong to PO 4 tetrahedra from adjacent phosphate anion ribbons. ................................ 68 4 3 Segment of phosphate anion ribbon composed of P O 4 tetrahedra as viewed along two directions (a) [010] and (b) [001] direction. ................................ ......... 69

PAGE 12

12 4 4 (a) Location of bridging and terminal oxygen in LaP 5 O 14 Pair distribution function for LaP 5 O 4 centered at (b) bridging (c) terminal oxygen ion. ................. 70 4 5 Theoretical powder pattern for LaP 5 O 14 compared with experimental powder patterns for the LaP 5 O 14 samples. ................................ ................................ ...... 72 4 6 Theoretical powder pattern for LaP 5 O 14 compared with experimental powder patterns for the LaP 5 O 14 samples (15 o 35 o ). ................................ .................... 73 4 7 (a) SEM image and (b) EDS spectr um of a typical crystal of undoped LaP 5 O 14 ................................ ................................ ................................ ............. 74 4 8 (a) Single crystal XRD pattern for LaP 5 O 14 samples(b) Relative orientation of the crystal during XRD data collection. ................................ ............................... 76 4 9 Nyquist plots for 5 mol% Sr doped LaP 5 O 14 samples measured at two different temperatures. ................................ ................................ ....................... 77 4 10 Variation of conductivity with dopant concentr ation as measured under unhumidified atmosphere. ................................ ................................ .................. 80 4 11 Conductivity measurement under wet/dry argon atmosphere on 5 mol% Sr doped LaP 5 O 14 ................................ ................................ ................................ .. 82 4 12 Calculated dopant incorporation energy for D 2+ = Mg, Ca, Sr and Ba on the La site. ................................ ................................ ................................ ................ 87 4 13 View of LaP 5 O 14 along (a) c axis and (b) a axis showing the location of hollow c hannels in the structure. (Proton shown in blue). ................................ .. 90 4 14 Conductivity of (a) 5 mol% Ca doped LaP 5 O 14 (b) 5 mol% Ca doped and 5 mol% Sr doped NdP 5 O 14 (unhumidified air). ................................ ....................... 92 5 1 Crystal structure of LaPO 4 as seen along the [001] direction. The PO 4 tetrahedra (green), La 3+ ions (orange) and O 2 (red) are depicted. ..................... 96 5 2 Pair distribution function for LaP 5 O 4 centered at (a) O1, (b) O2, (c) O3, and (d) O4. ................................ ................................ ................................ ................ 97 5 3 Comparison of theoretical and experimental XRD powder pattern of the uncalcined, calcined and si ntered LaPO 4 samples. ................................ ............ 98 5 4 Experimental XRD patterns for the undoped, 1 mol% Sr doped and 5 mol% Sr doped LaPO 4 pellets (W standard). ................................ ............................... 99 5 5 SEM image and EDS spectrum (inset) of the fractured surface of a typical sintered sample of undoped LaPO 4 ................................ ................................ 101

PAGE 13

13 5 6 Nyquist plots obtained under unhumidified air for undoped LaPO 4 samples (a), (b) and for 1 mol% Sr doped LaPO 4 (c), (d). ................................ .............. 102 5 7 Measured conductivity for undoped and doped LaPO 4 (unhumidified air). ....... 105 5 8 Measured conductivity for undoped and doped LaPO 4 as measured in humidified atmosphere (wet air). ................................ ................................ ...... 108 5 9 Calculated dopant incorporation energy for D 2+ = Mg, Ca, Sr and Ba on the La site. ................................ ................................ ................................ .............. 113 5 10 Resulting local relaxed structures of (a) hydrogen phosphate, (b) pyrophosphate, and (c) orthophosphate. ................................ .......................... 115 6 1 Effect of variation of the pre exponential factor ( o ) and activation energy (E a ) on the conductivity. ................................ ................................ .................... 120 6 2 The effect of reducing the oxygen oxygen ion distance on the barrier for proton jump (or activation energy). ................................ ................................ ... 124 6 3 Conductivity of undoped, 1 mol% Sr and 5 mol% Sr doped LaPO 4 ( open symbols ) and LaP 5 O 14 ( closed symbols) ................................ ...................... 125 6 4 Proton jump distances i n (a) LaPO 4 and (b) LaP 5 O 14 ................................ ....... 128 7 1 (a) Ideal crystal structure of SnP 2 O 7 as seen along the [100] direction (b) [P 2 O 7 ] 4 (c) 3 x 3 x 3 superstructure (d) P O P bond angle. ........................ 134 7 2 Powder XRD patterns of crushed as sintered pellets of doped and undoped SnP 2 O 7 ................................ ................................ ................................ ............ 136 7 3 SEM images of fractured surface of SnP 2 O 7 sampl e sintered at 1400 o C for 10 h. (a) Average grain size ~10 m. (b) Porosity at triple grain junctions. .... 138 7 4 Measured conductivity values of 5 mol% Sr:LaPO 4 and 20 mol% Eu:SrCeO 3 as compare d with conductivity values for SnP 2 O 7 samples. ............................. 140 7 5 Comparison of crushed powder and as sintered pellets of doped SnP 2 O 7 by using (a) 31 P MAS NMR spectroscopy and (b) XRD. ................................ ..... 142 7 6 Nyquist plot of 10 mol% In doped SnP 2 O 7 (600 o C). ................................ .......... 144 7 7 Variation of grain boundary, grain ionic and total conductivity in 10 mol% In doped SnP 2 O 7 with inverse temperature. ................................ ......................... 145 7 8 Total conductivity of undoped SnP 2 O 7 and D 0.1 Sn 0.9 P 2 O 7 (D = In, Zn, Mg, Ga, and Sc) plotted as a function of inverse temperature. ................................ 146

PAGE 14

14 7 9 Proton transference number of undoped and 10 mol% In doped SnP 2 O 7 as a function of temperature (300 o C 500 o C). ................................ ......................... 148 8 1 Crystal struc ture of (a) La 2 P 4 O 13 along [010] direction and (b) La(PO 3 ) 3 along [110] direction. ................................ ................................ ................................ .. 155 8 2 Setup for modified hydrothermal synthesis of large ultraphosphate crystals. 151 ................................ ................................ ................................ ........ 157 B 1 Segment of PO 4 chain seen along (a) [010] and (b) [001]. Crystal structure of La(PO 3 ) 3 seen along (c) [110] and (d) [001] direction. ................................ .. 163 B 2 XRD pattern for the as synthesized powder of La(PO 3 ) 3 .3H 2 O using aqueous method. ................................ ................................ ................................ ............ 164 B 3 X ray diffraction pattern for synthesized powder of anhydrous La(PO 3 ) 3 using aqueous method. ................................ ................................ .............................. 165 B 4 XRD pattern of a sintered sample of La(PO 3 ) 3 as compared to the unsintered powder. ................................ ................................ ................................ ............. 166

PAGE 15

15 Abstract of Dissertation Presented to the Graduate School of the Univers ity of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy CRYSTALLOCHEMICAL AVENUES FOR THE ENHANCEMENT OF PROTON CONDUCTION IN CERAMICS: PHOSPHATE COMPOUNDS, EXPERIMENTS AND SIMULATION By Satyajit P hadke December 2010 Chair: Juan C. Nino Major: Materials Science and Engineering Current proton exchange membranes (PEM s ) are based on polymeric materials temperat ure of opera tion (approximately 100 o C). To overcome this limitation ceramic PEM materials are being investigated for transportation applications, where operating temper atures in the range of 200 o C 600 o C are desired. In this study, the conductivity behavior and elect rical properties of three families of phosphate compounds ( namely SnP 2 O 7 LaP 5 O 14 and LaPO 4 ) have been studied in order to better understand crystal structure proton conduction relationships in this class of ceramic materials. From the phosphate rich region of the La 2 O 3 P 2 O 5 binary phase diagram the orthophosphate (LaPO 4 ) and ultraphosphate (LaP 5 O 14 ) materials are chosen for a comparative study of the electrical properties. The conductivity of the lanthanum phosphates (doped and undoped) is measure d using impedance spectroscopy in the temperature range 300 o C 600 o C. The conductivity of 5 mol% Sr 2+ doped LaP 5 O 14 ( 1.01 x 10 4 S/cm, 600 o C) is found to be an order of magnitude higher than similarly doped LaPO 4 ( 7.00 x 10 6 S/cm, 600 o C) which is a well investigated proton conducting

PAGE 16

16 ma terial. In addition, it i s observed that the activation e nergy for protonic conduction i s much lower for doped LaP 5 O 14 ( 0.80 0.01 eV) as compared to LaPO 4 ( 1.09 0.01 eV). A hypothesis relating the oxygen to oxygen io n distance in a material to the activation energy for proton conduction is presented and the experimental results obtained are critically examined on the basis of the hypothesis and other relevant literature From this analysis it is shown that the conden sed nature of the phosphate anion in LaP 5 O 14 can provide low energy avenues for proton transport within the material leading to enhanced conductivity In addition, the electrical properties of u ndoped and acceptor doped tin pyrophosphate (SnP 2 O 7 ) are inv estigated, which is regarded as a potential candidate electrolyte material for PEMFCs. Within the dopants analyzed, Zn 0.1 Sn 0.9 P 2 O 7 shows the highest conductivity of 2.84 x 10 6 S/cm at 600 o C. However, t his value is several orders of magnitude lower tha n previously reported for doped tin pyrophosphate compounds but it is consistent with a recent publication. T he reason for the discrepancy i s investigated and a possible explanation has been proposed based on results obtained from 31 P MAS NMR spectrosco py It i s observed that the as calcined powder of doped SnP 2 O 7 contained a significant fraction of residual phosphoric acid from synthesis p rocedure which i s likely to be responsible for the high conductivity values for acceptor doped SnP 2 O 7 reported prev iously by other research groups.

PAGE 17

17 CHAPTER 1 INTRODUCTION 1.1 Statement of Problem and Motivation There is growing scientific consensus regarding the existence of a direct relation between the increasing levels of greenhouse gas emissions and global warm ing. 1 The natural greenhouse gases include CO 2 N 2 O, CH 4 and ozone which are essential to support life on earth. However since the beginning of the industrial revol ution in 1765, there has been a drastic increase in the production of carbon dioxide. The average temperature of the earth has risen by 1 o C sin ce the mid 19 th century. 2 In addition to the global warming concerns, the limited crude oil reserves and the exponential increase in fuel prices has lead to worldwide interest in cleaner and sustainable sources of energy. 3 Proton exchange membrane f uel cells (PEMFCs) present a viable alternative to obviate the above current problems in the energy sector. 4 Compared to the conventional internal combustion engines which use direct combustion of fuel to produce energy, fuel cells utilize electrochemical combustion to generate elect ricity from fuel which is intrinsically much more efficient and leads to a much lesser wastage of energy. Moreover, the use of h ydrogen as a fuel for fuel cells is an attractive option as it has zero emission of toxic gases and greenhouse gases. 5 The only by products from the electrochemical combustion of hydr ogen in a fuel cell are water and heat. One of the other critical advantages of the fuel cell technology is the ease of scalability so as to suit a variety of applications. Due to this reason hydrogen fuel cells are being looked at as a potential techno logy to provide small scale power to daily electronic devices such as laptops as well as for high power applications such as domestic power and heating and also for transportation applications

PAGE 18

18 H owever, some major obstacles need to be overcome to make th is technology more economical to ensure its widespread use. The currently used materials are polymer based (Nafion TM ) and have an upper limit to their temperature of operation (<100 o C). 6,7 A higher temperature of operation in the range 300 o C 600 o C is highly desirable because of increased efficiency in this temperature r ange as well as increased tolerance of the electrodes to CO (carbon monoxide) poisoning. 7 9 The com mon ly used electrode material contains platinum which is very expensive and hence prevention of electrode poisoning has the potential of lowering the overall cost of the technology significantly 10 According to the targets laid down by DOE (department of energy) for transportation applications, by 2015, a hydrogen fuel cell system needs to be developed with a peak efficiency of 60% with a cost of $30/kW or lower. Thus, there is need for the development of new materials which exhibit optimum properties in this temperature range. The focus of this study is on achieving enhanced conductivity in phosphate based materials by expanding the realm of knowledge t o higher condensed phosphates which have so far received very limited attention as candidate proton conducting electrolyte materials. 1.2 Scientific Approach F or achieving higher temperatures of operation numerous investigations (in the last decade or so ) have focused on ceramic based materials. The conductivities of various proton conducting materials reported in literature are plotted in Figure 1 1. Broadly all proton conducting materials can be classified into three distinct categories on the basis o f their temperature range of operation In the low temperature range there are mainly polymeric materials such as Nafion TM which show very high proton conductivity. However, these materials tend to

PAGE 19

19 dry up as the temperature is increased beyond the boili ng point of water and hence their use is limited to lower temperatures. Solid acidic salts such as CsHSO 4 and KHSO 4 show appreciably high conductivity at slightly higher temperatures than polymer based materials However, their use at higher temperature s is restricted due to a low decomposition temperature (~ 180 o C). Also, their practical applications as electrol yte material are limited due to the high solubility in water. Figure 1 1. Comparison of conductivities of proton conducting materials. 8

PAGE 20

20 In the high temperature range (> 600 o C) the doped perovskite type oxide materials such as BaZrO 3 BaCeO 3 SrCeO 3 exhibit high conductivities This leaves the intermediate temperature range (300 o C 600 o C) where no material with suitably high 1 and was first recognized by Norby et al 7,8 As explained in the previous section this temperature range of operation is de sirable for PEMFCs In this respect, several different orthophosphate materials have been investigated as PEMs, including RPO 4 (R = La, Sm, Nd, Ce), Ba 3 Ce(PO 4 ) 3 and some pyrophosphate materials such as SnP 2 O 7 and TiP 2 O 7 However, the number of studies on higher condensed phosphate materials such as metaphosphates (A(PO 3 ) n ) and ultraphosphates (A 3 (P 5 O 14 ) n ) is quite limited. In the phosphate rich region of the La 2 O 3 P 2 O 5 binary phase diagram four stoichiometrically different lanthanum phosphate materi als exist, namely LaPO 4 (orthophosphate), La 2 P 4 O 13 (tetraphosphate), La(PO 3 ) 3 (metaphosphate/polyphosphate) and LaP 5 O 14 (ultraphosphate). The difference in stoichiometry in the above compounds is acc ommodated crystallochemically by changes in the degree o f interl inking or condensation of the phosphate anion which leads to a different yet closely related structure This project investigates the effect of structural variation across the entire range of condensed phosphate compounds on the proton conductivit y in the material. 1.3 Organization of the Dissertation C hapter 2 covers the background information that is referred to on several occasions throughout the dissertation In the beginning a brief history and introduction to electrochemi stry is presented along with a broad classification of fuel cells. This is followed by a brief review on PEMFCs which include the basic working principle, the

PAGE 21

21 main advantages and potential applications of this technology and also the current materials challenges in this fi eld. A detailed overview of the various approaches towards development of an optimum electrolyte material and a classification of the different types of materials that are being investigated as proton exchange membranes (PEMs) is presented. The remainder of the chapter introduces the rich chemistry of the condensed phosphate materials which is analogous to that observed and well known in the silicates. The phosphate rich region of the La 2 O 3 P 2 O 5 phase diagram is discussed based on which the two materia ls chosen for this comparative study were selected. At the end of the chapter the relevant information about electrical characterization used in this study is presented. In Chapter 3, the materials and experimental procedures and a description of the ex perimental setup used in this study is presented. T he experimental details about all the cha racterization techniques used are also stated. Chapter 4 is dedicated to all the characterization studies performed on lanthanum ultraphosphate (LaP 5 O 14 ). A crys tallochemical analysis of the same is presented which will be crucial to the comparative discussion of electrical behavior presented in Chapter 6. The results from conductivity measurement on the three different compositions of LaP 5 O 14 are presented follo wed by the results from atomistic simulation studies on the material. Chapter 5 contains the results from all the experimental characterization studies performed on lanthanum orthophosphate (LaPO 4 ). The compositions of the prepared samples of LaPO 4 were kept consistent with those of LaP 5 O 14 in terms of dopant cation and concentration to aid the comparative discussion that follows in

PAGE 22

22 Chapter 6. These are followed by the results obtained from the atomistic simulation studies on the same. The electrical p roperties of LaPO 4 and LaP 5 O 14 are compared in Chapter 6. The obtained differences in the conductivity are critically analyzed on the basis of the proposed model for proton conduction in these materials. The enhancement in conductivity achieved in LaP 5 O 1 4 led to the extension of the study to other ultraphosphates which are discussed in Chapter 7. The results from experimental characterization of SnP 2 O 7 are presented in Chapter 8 This is followed by summary of the entire work and description of the poss ibilities for future extension of the work in Chapter 9.

PAGE 23

23 CHAPTER 2 BACKGROUND 2.1 Solid S tate E lectrochemistry Solid state electrochemistry is a broad field that deals with all types of solid state ionically conducting materials. Under the influence o f an electrochemical gradient these materials allow the passage of electric current by the transport of ions such as protons, oxygen ions, silver ions, lithium ions or any other mobile charged species in the material. Ionic conduction in solid materials was reported by Warburg et al. as early as 1884 by carrying out transference number measurements in which the charge transport was equated with the mass transport in the material. 11 The concept of defects (interstitials and vacancies) was introduced by Joffe for the first time to explain to observed high ionic conductivity in some materials. 12 One of the earliest well investigated ionic conductors was AgI (silver iodide) which exhibits very high conductivity (cationic con duction) above 149 o C. 13 The oxygen ion conduction in doped zirconia (ZrO 2 ) was ex tensively investigated by Kiukkola et al. in 1957 and is now commonly used as electrolyte material in SOFCs. 14 Sodium ion conduction was studied by Weber et al. in 1967 in the material NaAl 11 O 17 which is used as the electrolyte material in sodium sulfur batteries. 15 The presence of protons in oxides such as Cu 2 O, CuO and NiO was discussed b y Wagner et al. as early as 1966. However, significantly high proton conduction was reported for the first time by Iwahara et al. in 1981 in doped strontium cerates (SrCeO 3 ). 16 Solid state electrol yte materials have various technological applications which include battery materials (such as lithium batteries and alkaline batteries), gas sensors,

PAGE 24

24 gas separation membranes, efficient gas synthesis methods (such as methane and ammonia), electrolytic com bustion of hydrocarbons in car exhausts, and various types of fuel cells. Table 2 1. Classification of fuel cells. Fuel Cell Type Operating Temperature ( o C) System Output (kW) Efficiency (Electric ,% ) Applications Alkaline (AFC) 90 100 10 100 60 70 Military, space Phosphoric Acid (PAFC) 150 200 50 1000 36 42 Distributed generation Proton Electrolyte Membrane (PEMFC) 50 100 250 50 60 Portable power transportation(cars) small distributed generation Molten Carbonate (MCFC) 600 1000 < 1000 60 Large distributed generation Solid Oxide (SOFC) 650 1000 5 1000 60 Auxillary power, large distributed generation Depending on the type of electrolyte used in the fuel cell they can be classified broadly into five different categories namel y, phosphoric acid fuel cells (PAFCs), alkaline fuel cells (AFCs), proton exchange membrane fuel cells (PEMFCs), solid oxide fuel cells (SOFCs) and molten carbonate fuel cells (MCFCs). The range of temperature of operation and system output of each type of fuel cell are summarized in Table 2 1. In terms of capacity PAFCs, SOFCs and MCFCs can be designed to generate up to 1000 kW and hence are mostly being considered for large scale distributed power generation. On the contrary, PEMFCs which can generate up to 250 kW are being looked at for transportation applications as well as for domestic power generation. Currently employed PEMFCs use polymer based electrolyte materials

PAGE 25

25 which have an upper limit to their temperature of operation (< 100 o C). However, a higher temperature of operation is highly desirable for reasons described in detail in section 1.1. 2.2 Proton Exchange Membrane F uel C ell s (PEMFCs) 2.2.1 Introduction Since the beginning of the industrial revolution in 1765, there has been an exponent ial increase in the production of carbon dioxide. 1 Th e re is worldwide concern regarding the d irect relation between the increasing levels of greenhouse gas emissions and global warming. 2 In addition to the glob al warming conc erns, limited crude oil reserves and the exponential increase in fuel prices has lead to interest in cleaner and sustainable sources of energy. Fuel cells present a viable alternative to obviate the above current problems in th e energy sect or. Hydrogen as a fuel for fuel cells is an attractive option as it has zero emission of toxic gases and greenhouse gases. The only products from the electrochemical combustion of hydrogen in a fuel cell are water and heat. 4 The fuel cell technology also represe nts a paradigm shift from the mechanism by which power is produced from fuel (electrochemical versus conventionally used direct combustion in an internal combustion engine (ICE)). There are significant advantages of a fuel cell over an ICE. The former ha s efficiencies ranging from 40 60% as compared to much lower efficiencies of the latter (~ 20%) This is mainly because : T he electrochemical combustion of fuel inside a fuel cell and subsequent conversion to mechanical energy through an electric motor is inherently much more efficient than the d irect combustion inside an ICE Efficiency of a fuel cell is not limite d by the Carnot cycle as in ICE Redu ced wastage of heat from tail pipe e missions or exhaust gases

PAGE 26

26 Reduced energy wastage from friction because a fuel cell has m uch less number of moving parts No power wastage during idling i f fuel cells are used in combination with an auxiliary power storage unit (such as lithium batteries) Figure 2 1. Individual fuel cell unit of a hydrogen fue l cell 2.2.2 Hydrogen F uel C ell U nit A fuel cell produces electricity directly from the electrochemical reaction between hydrogen (used as fuel) and oxygen from the air as shown in Figure 2 1. An individual fuel cell unit consists of a proton conducting electrolyte sandwiched between two electrodes, an anode and a cathode Hydrogen gas flows through channel s to the anode, where a catalyst causes the hydrogen molecules to separate into protons and electrons The proton conducting electrolyte is insulatin g to electrons and hence allows only the protons to pass through it. The electrons follow an exte rnal circuit to the cathode resulting in an electr onic current which can be utilized to do work. Each such

PAGE 27

27 unit of a fuel cell produces about 0.7 V of potent ial difference By stacking multiple such units in series a fuel cell unit of the desired capacity can be designed 2.2.3 Potential A pplications The power capacity of a PEMFC unit is in the range of ~ 250 kW which makes it ideally suited for use in the transportation industry to provide power to automobiles. However some critical road blocks need to be overcome to make this technology sufficiently economical so as to ensure its widespread use in the automotive industry. Proton exchange membrane fuel ce lls are also being looked at for small scale applications such as to provide portable power for computers, laptops and other daily electronic devices. Their low temperature of operation as compared to other types of fuel cells makes them suitable for this purpose. In addition PEMFCs are also being considered for combined heat and power generation for domestic purposes. 2 .3 Crystalchemistry of C ondensed P hosphates 2. 3 .1 Structural C omparison with the S ilicates The structural complexity of the condensed phosphates is analogous to that observed in the silicates It is important to draw the analogy considering that the structur e of silicate compounds is more well known compared to phosphate compounds The struct ural unit for the anion in all the silicate compounds is an [SiO 4 ] 4 tetrahedron. Structurally the tetrahedron can form diverse anion structures by corner sharing one or more of its oxygen ions with other neighboring tetrahedra. A structural classificati on of silicate compounds is presented in Table 2 2. The orthosilicates are the least complex structurally and consist of isolated [SiO 4 ] 4 tetrahedra which are held together by cations. The most common examples of this type of compounds are the olivines and garnets

PAGE 28

28 According to the arrangement of the tetrahedra, the anion can attain either ring l ike, linear chain like or sheet like morphology leading to the formation of different types of silicates. Table 2 2. St ructural classification of silicates Nomenclature Structural arrangement Chemical formula O/Si ratio Example compounds Mineral names Neosilicates / Orthosilicates Isolated tetrahedra [SiO 4 ] 4 4 (Mg O.9 Fe 0.1 ) 2 Si O 4 Forsterite, Fayalite, Monticellite, Tephroi te Sorosilicates / Pyrosilicates Paired tetrahedra [Si 2 O 7 ] 6 3.5 Na 6 Si 2 O 7 Nd 2 Si 2 O 7 Y 2 Si 2 O 7 Sc 2 Si 2 O 7 Thortveitite, Thalanite Cyclosilicates / Ring silicates Tetrahedra rings [Si n O 3n ] 2n 3 NaAl 3 Al 6 B 3 Si 6 O 30 (OH) Tourmaline, Beryl, Cordierite Inosilica tes / Chain silicates Ribbons [Si 4n O 11n ] 6n 2.75 Ca 2 (Mg,Fe) 5 Si 8 O 22 (OH) 2 Na 2 (Mg,Fe) 3 Al 2 Si 8 O 22 (OH) 2 Actinolite, Glaucophane Sheet silicates Sheets [Si 2 O 5 ] n 2n 2.5 Mg 3 Si 4 O 10 (OH) 2 Muscovite, Biotite, Talc Tectosilicates 3D network [SiO 2 ] 2 CaAl 2 Si 2 O 8 Anorth ite The structure of the silicates has a significant impact on the physical properties of the material. For example, the naturally occurring micas (a class of silicate minerals) have a common property that they can all be easily cleaved along one plane. This is because their silicate anion structure has a sheet like morphology and the intra layer bonds are strong covalent bonds whereas the adjacent layers are held together by only The ring silicates have very low thermal ex pansion coefficient and is thus commonly used to produce refractory material that is highly resistant to thermal shock.

PAGE 29

29 2.3 .2 Temperature D ependent E quilibrium in F used Phosphoric A cid Orthophosphoric which has the chemical formula H 3 PO 4 undergoes a cond ensation reaction when heated above room temperature. Since condensation reaction is thermally aided and occurs with the elimination of a water molecule it is termed as thermal dehydration of phospho ric acid. The reaction proceeds as shown in equation 2 1 starting with two orthophosphoric acid molecules that condense to form py r ophosphoric acid (or diphosphoric acid) with the release of a water molecule. The condensation reaction is shown schematically in Figure 2 2. The reaction can proceed fu r ther by condensation to yield more complex acids such as triphosphoric acid, metaphosphoric acid and so on. (2 1) The extent to which the condensation reaction occurs is temperature dependent and was investigated in detail by Chudino va et al. by studying the effect of temperature on dilute mixtures of bismuth oxide (Bi 2 O 3 ) and orthophosphoric acid (H 3 PO 4 ). 17 The crystallization of four stoichiometric compounds was reported namely BiPO 4 Bi 2 P 4 O 13 BiH(PO 3 ) 4 and BiP 5 O 14 in the temperature ranges <200 o C, 200 o C 280 o C, 280 o C 380 o C, and >400 o C respectively. Within each temperature range the condensation reaction in phosphoric acid continues until equilibrium is reached which is followed by precipitation of the salt c ontaining the cation from the oxide material in the starting mixture. A similar trend has been observed in preparation of several other phosphates and a comprehensive list of the same can found in a compilation by Durif et al 18

PAGE 30

30 A similar trend is observed i n the lanthanum phosphates The synthesis of lanthanum metaphosphate (La(PO 3 ) 3 ) from phosphoric acid was carried out by Znamierowska et al. at 400 o C wh ile crystals of lanthanum ultraphosphate (LaP 5 O 14 ) were prepared by Park et al. at a higher temperature (600 o C). 19 A similar trend is also observed in the preparation of neodymium phosphates by Hong et al 20 Figure 2 2. Thermal dehydration of phosphoric acid leading to formation of poly phosphoric acids of higher complexity. 2.3 .3 Structural C lassification of P hosphate C ompounds The structural units for the condens ed phosphate compounds are PO 4 tetrahedra. Depending on the degree of condensation (corner sharing) among the tetrahedra to form the phosphate anion, the phosphates can be classified as shown in the flowchart in Figure 2 3. 18 The oxyphosphates can be found in the phosphate poor regions of the phase diagram of the metal oxide phosphorus pentoxide phase diagram. The O/P ratio for the oxyphosphates is greate r than 4, which means that there is an excess of oxygen ions in the structure beyond those required to form PO 4 tetrahedra. T hus as the name suggests t he anion in these materials consists of a mixture of isolate d PO 4 tetrahedra and isolated oxygen ions. The orthophosphates (or monophosphates) separate the oxyphosphates from the condensed phosphates in the phase diagram and have and O/P ratio of exactly 4. They

PAGE 31

31 are structurally least complex among the phosphates and their anion consists of isolated PO 4 t etrahedra. There is a large variation in the possible structures and stoichiometries in the category of condensed phosphates depending on the nature of corner sharing of tetrahedra in the anion as seen in Figure 2 3 18 Figure 2 3 General structural classification of condensed phosphates. The formation of a part icular type of phosphate has strong temperature dependence as explained in section 2.3. Broa dly, the condensed phosphates can be classified as polyphosphates, cyclophosphates (or ring phosphates) and ultraphosphates. The anion in polyphosphates with the general formula [P n O 3n+1 ] (n+2) consists of pairs (n = 2) triplets (n = 3) quads (n = 4) an d so on of corner sharing

PAGE 32

32 (linear) tetrahedra The metaphosphates are an extreme case of polyphosphates when theoretically n approaches infinity. In the cyclophosphates the tetrahedra are corner sharing to yield a ring like morphology of the phosphate an ion and can be further classified depending on the ring size. The anion in ultraphosphates in general has a ribbon like morphology but can occur in a variety of structures with minor differences which are describ ed in more detail in section 4 .1 2.3 .4 L a 2 O 3 P 2 O 5 B inary P hase D iagram As shown in Figure 2 4 the phase diagram includes the compounds LaPO 4 La 2 P 4 O 13 La(PO 3 ) 3 and LaP 5 O 14 19 The difference in stoichiometry in the above compounds is accommodated (crystallochemicall y) by changes in the degree of condensation (corner sharing) in the phosphate anion an d thus leads to a different yet closely related crystallochemical arrangement. Th is crystallochemical variation across stoichiometry, and in particular, as a function of O/P ratio, is analogous to that observed in silicates (correspondingly O/Si ratio). 21 Structurally, lanthanum orthosphosphate (LaPO 4 O/P = 4) is least complex and consists of isolated PO 4 tetrahedra which are held together by LaO 9 polyhedra. In lanthanum tetraphosphate (La 2 P 4 O 13 O/P = 3.25) the phosphate anion group consists of a quartet of corner sharing PO 4 tetrahe dra that are held together by LaO 8 polyhedra. As the O/P ratio further decreases to lanthanum metaphosphate (La(PO 3 ) 3 O/P = 3) the phosphate anion group is significantly more complex (with more condensation in PO 4 tetrahedra) and consists of theoretically infinitely long helical chains of corner sharing tetrahedra. In lanthanum ultraphosphate (LaP 5 O 14 O/P = 2.8) the phosphate anion consists of theoretically infinitely long ribbons of corner sharing PO 4 tetrahedra which are parallel to each other and are held together by LaO 8 polyhedra.

PAGE 33

33 Thus, as the O/P ratio decreases from LaPO 4 to LaP 5 O 14 the degree of cor ner sharing of the PO 4 tetrahedra increases leading to a structurally more complex anion structure. As will be evident in the discussion in Chapter 6, the O O distance decreases along the crystallographic direction of condensation in the specific lanthanu m phosphate as the degree of condensation in the phosphate anion increases. Figure 2 4 Binary phase diagram for the system La 2 O 3 P 2 O 5 A comparison of the experimentally measured conductivities and activation energies (across the range of compounds) in such a case can be directly correlated with the oxygen to oxygen distances. In this study, we present a scientific evaluation of this

PAGE 34

34 crystallochemical approach towards the development of enhanced proton conducting materials by contrasting strontium d oped LaPO 4 with strontium doped LaP 5 O 14 2 4 Electrical C haracterization 2. 4 .1 Impedance S pectroscopy Some parts of following text in this section are based on the book by MacDonald et al. which presents a comprehensive discussion on fundamentals and app lications of impedance spectroscopy. 22 At any interface in a material, the local physical properties of the material such as crystal structure, composition, mechanical and electrical properties change quite abruptly. Such discontinuities in the material properties lead to partial obstruction of the flow of charge c arrying species which are being driven by an e lectrochemical gradient. The obstruction of flow leads to preferential charge depletion or acc umulation along the path This causes heterogeneous charge distributions (also termed polarization) at the material interface, which affect s the electrical prop erties of the material. In case of materials that are used for solid state electrochemical devices the three commonly encountered interfaces are grain boundaries, electrolyte electrode and the triple phase electrolyte electrode atmosphere interface. Upo n application of a potential difference to a material charge each interface polarizes in a unique way. However, the rate at which the direction of polarization will reverse when the applied potential difference is reversed is characteristic of the particu lar type of material interface. Generally bulk grain response is the quickest followed by grain boundary polarization and electrode polar ization Essentially this means that if electrical properties are measured as a function of the frequency of the inpu t AC signal over a wide range of frequencies then it can be used to obtain information about the dynamics of the mobile

PAGE 35

35 charge carriers and characteristics of the interfaces in the material. Thus impedance spectroscopy has emerged as a very powerful tech nique for investigating the electrical properties of a material as it has the ability to isolate the contributions from the various microscopic processes occurring within the material which contribute to the overall electrical response. The frequency dep endent impedance data collected can be written as a sum of the real and imaginary parts. (2 2) If one plots the imaginary part of the impedance versus the real part, i.e. ( versus ( ), the resulting locus shows distinc tive features for certain combi nations of circuit elements. Such a plot is commonly referred to as a Nyquist plot. The characteristic Nyquist plots for a purely resistive circuit and a pure capacitance is shown in Figure 2 5 (a). The Nyquist plot for a commonly used R C circuit is shown in Figure 2 5 (b). It can be seen that the shape of the plot is different for different combinations of circuit elements although it is not necessarily unique. 22 Complex impedance plots are useful for determining an appropriate equivalent circuit for a system and for estimating the values of the circuit parameters In gener al polycrystalline ceramic materials exhibit a three arc response which corresponds to bulk, grain boundary and electrode polarization phenomena. Each of the observed arcs is modeled using a separate R C circuit similar to that shown in Figure 2 5 (b). Physically the individual circuit elements can be interpreted as resistance of the material to the flow of the mobile charge carriers and the capacitance due to charge accumulation due to discontinuities at the interfaces within the material as

PAGE 36

36 explained previously. The reason for the polarization in the bulk is explained by saying that the flow of charge carrying species is partially blocked at the grain boundary. In the same way the charge carriers inside the grain boundary are partially blocked at the interface which explains the observed capacitance associated with grain boundary impedance. The interpretation of the origin of each arc observed in the Nyquist plot for a polycrystalline ionically conductive material was provided by Bauerle et al. from conductivity measurements on yttria stabilized zirconia. 23 The area / thickness ratios of the samples were varied by a ratio of about 100 while the thickness was varied independently by a ratio of about 10 b y using wafer specimens and long bars Th e reasoning behind the above experiment was that the electrode impedance arc in the Nyquist plot should be affected only by the change in area while be un affected by the change in length of the sample. Similarly, the arcs corresponding to the electrolyte would be unaffected by the changes in the ratio of area / thickness. Thus it is clear that the impedance spectra (Nyquist plots) are very sensitive to th e microstructure of the material and can be used effectively to study the effect of varying the material processing parameters (which influence the microstructure) on the electrical properties. Examples of two R C circuits in series are shown in Figure 2 5 (c) and (d). These Nyquist plots were generated by inputting the relevant equivalent circuit elements and parameters in ZView TM software. The values of the resistances are kept the same in both the circuits however the capacitance were given differe nt values. It can be seen by comparing Figure 2 5 (c) and (d) that a large difference in the capacitance values leads better resolution of the two arcs. In a more general sense the resolution of the arcs is determined by the differences in the peak frequ p ) of the two

PAGE 37

37 microscopic polarization phenomena responsible for the observance of the arcs in the Nyquist plot. (2 3) If there is a large difference in the capacitance values then good resolution of the arcs is o btained in the impedance spectrum but if there is large difference in the resistance values then good resolution is obtained in the modulus spectrum. (a) (b) (c) (d) Figure 2 5. Nyquist plots for (a) R and C, (b), single R C circuit, (c) partiall y resolved arcs, and (d) resolved arcs for two R C circuits in series. In most of the real material samples neither spectra give very good resolution but the equivalent circuit parameters can be accurately estimated by fitting the data. The above R C circuits use a single value of capacitance for a certain type of polarization

PAGE 38

38 phenomenon over the entire sample which is generally not true F or example the grain boundary in a material cannot be assumed to be completely homogeneous over the entire sample At the atomic level its thickness may be different in some places or the degree of crystallographic mismatch between the two adjacent grains may be higher or lower depending on the relative orientations. Figure 2 6. Nyquist plots corresponding to s ingle R CPE circuits. For the CPE, alpha is varied from 0.90 to 1. Thus the equivalent circuit parameters vary over a certain range about a mean value over the entire material. However, the macroscopic impedance data collected is representative of t he average properties of the interface. As the peak frequency ( p ) is dependent on the equivalent circuit parameters (R and C) it leads to a distribution of peak frequencies in the arc and a corresponding depression in the observed semi circle in the Nyquist plot. Th i s observation is accounted for by considering a constant

PAGE 39

39 phase element (CPE) instead of a capacitance element. The impedance of a CPE is mathematically denoted as follows: (2 4) The effect of varying the parameter from 0.90 to 1 on the shape of the observed Nyquist plo t of and R CPE circuit is shown in Figure 2 6. The numbers along the plots indicate the logarithm of the corresponding frequency for the individual data point. A value of resis tive circuit element. 2. 4 .2 Transference N umber M easurements In most electrochemical materials there is always more than one type of charge carrier species. As such the total conductivity in these materials can be expressed as the sum of the partial con ductivities of the individual species as shown below: (2 4) Usually for application as an electrolyte material the partial conductivity of one of the charge carrying species is much higher than the others and it is the domin ant charge carrier in that material. In such a situation it is important to ascertain the individual contribution of each of the mobile charge carriers and can be accomplished by carrying out transference number measurements. The transference number of a charge carrier d can be represented as: (2 5) Where t d is the transference number, d is the conductivity of the charge carrier d and total is the total measured conductivity. Three different techniques have been

PAGE 40

40 commonly used to perform transference number measurement, namely mass transfer measurements, emf measurements using galvanic cells and polarization measurements. The mass transfer measurements involve the passage of current through the electrolyte material and c omparison of the charge flown with the mass transported between electrodes. This is the oldest method for transference number measurement and its first use dates back to 1888 by Warburg et al 24 The transference number of silver iodide (silver ion conductor) was experimentally calculated by Tubandt et al by calculating the mass of silver transported across the electrolyte by meas uring the weight of the electrodes before and after passing current. 24 The necessary condition for this type of measurement is t hat the electrode must be a good electronic conductor as well as a good conductor of the ion that is being transported through the electrolyte. In the second measurement method the material to be studied is setup as an electrolyte in a galvanic cell. Th is type of measurement w as used in the current work for the measurement of the proton transference number. In a galvanic cell, the electrical work done is equal to the decrease in Gibbs free energy of the system (constant temperature and pressure). The o verall electrical work done is also equal to the product of the pot ential difference and the charge transported. Hence the following equation can be written for a galvanic cell (Nernst equation) 24 : (2 6) Where G is the Gibbs free energy of the cell reaction, E is the emf of the cell, F is z is the number of electrons transferred per cell reaction. For proton transference number measurement the galvanic cell u tilizes a differenti al

PAGE 41

41 hy drogen partial pressure setup Thus the difference in partial pressure (or chemical potential ) of hydrogen across the sample provides the electrochemical gradient required for charge transport. In such a situation the following expression applies to the emf of the cell 24 : (2 7) W here t ion is the transference number of the mobile ionic charge carriers For the case of pure ionic conduction the t ion = 1 and knowing that the dependence of chemical potential on the partial pressure of hydrogen can be represented by the theoretical emf of the cell can be calculated as: (2 8) Where R is the universal gas constant, T is the absolute temperature and P H2 represent the partial pressures of hydrogen in the two chambers of the galvanic cell. However, if the electrolyte material shows some partial electronic conductivity in addition to ionic conductivity then the measured value of emf across the cell is lower than the theoretical emf and the transference number can be calculated as below: (2 8) The third method of calculating the transference number consists of direct measurement of the partial conductivity of each of the mobile charge carrier species by using the appropriate blocking electrodes. The advantage of this method is that the

PAGE 42

42 transference number of the minority cha rge carrier can also be calculated accurately and was used for the first time by Hebb and Wagner. 24

PAGE 43

43 CHAPTER 3 M ATERIALS AND E XPERIMENTAL PROCEDUR ES 3.1 Synthesis of Materials This section describes the synthesis procedure used to prepare the samples of doped and undoped LaP 5 O 14 LaPO 4 and SnP 2 O 7 Due to the comparative nature of the study the dopant content (undoped, 1 mol% an d 5 mol% Sr 2+ ) was kept consistent across the lanthanum phosphate compounds. The dopant cations for SnP 2 O 7 (D = In, Mg, Ga, Sc, and Zn) were chosen on the basis on the minimization of lattice strain upon incorporation of the dopant cation. 3.1.1 Lanthan um U ltraphosphate (LaP 5 O 14 ) Single crystals of LaP 5 O 14 were synthesized by precipitation from concentrated phosphoric acid solutions as previously described by Park and Kreidler. 19 A flowchart of the synthesis process is shown in Figure 3 1 (a). The i nitial mixture was prepared by mixing 1 g of lanthanum oxide ( La 2 O 3 Alfa Aesar, 99.99%) with 30 mL of 85% p hosphoric acid (Fisher Chemical ) which was then allowed to stand at 300 o C for about 4 h until the oxide was completely dissolved and a transparent solution was formed. The oxides are generally insoluble in phosphoric acid at room temperature. However, th e oxides readily dissolve after the initial heat treatment of H 3 PO 4 ( at 300 o C) which is accompanied by vigorous bubbling due to the evaporation of water from the mixture. After the evaporation of water the oxides remain dissolved even if the solution is c ooled down to room temperature. To prepare single crystals of strontium doped LaP 5 O 14 the initial mixture consisted of lanthanum oxide and strontium oxide ( SrO, Alfa Aesar, 99.99%) in the appropriate molar ratio. The above solution was then tr ansferred t o a furnace ( pre heated to 300 o C ) to prevent thermal shock to the crucible After

PAGE 44

44 transferring the contents to the furnace the temperature was allowed to increase to 650 o C at a ramp rate of 100 o C/h and the solution was allowed to stand there for a perio d of 7 to 10 days. All of the above described reactions were carried out in covered vitreous carbon crucibles which are resistant to chemical attack by fused po lyphosphoric acid. (a) (c) (b) Figure 3 1. ( a ) Flowchart for synthesis of single crystals of LaP 5 O 14 (b ) vitreous carbon crucible, and ( c ) setup used for dissolution of polyphosphoric acid. Vitreous carbon crucibles are extremely susceptible to cracking resulting from the existence of a thermal gradient in the material. In additi on, the material undergoes a phase transition at about 100 o C which is also accompanied by a volumetric change.

PAGE 45

45 Thus to prevent cracking during the heat treatment procedure low ramp up and ramp down rates were used and the crucibles were removed from the f urnace only after the samples had cooled down to room temperature. After cooling, the excess phosphoric acid solidifies to form a gel like transparent semi solid High temperature leads to the thermal dehydration of phosphoric acid which consists of the condensation of the isolated PO 4 tetrahedra in acid with the elimination of a water molecule. The condensation gradually leads to the formation of acids of much higher complexity such as di phosphoric acid, tri phosphoric acid, meta phosphoric acid and so on. Pure orthophosphoric acid is a viscous liquid at room temperature but the condensed form of the acid is a soft semi solid. Since the condensed acid is the de hydrated form of orthophosphoric acid it has great affinity for water and can be dissolv e d by placing the crucible in warm water (~70 o C) as shown in Figure 3 1 (c) Upon dissolution of the gel, rhombus shaped crystals in the form of thin platelets were recovered from the bottom of the crucible. The habit planes of the crystal growth are ind icated in Figure 3 3 (b). The obtained crystals were washed several times with deionised water and then dried in air and stored in a desiccator for further experimentation. 3.1.2 Lanthanum O rthophosphate (LaPO 4 ) Lanthanum orthophosphate was prepared by the precipitation from an aqueous solution. Lanthanum nitrate (La(NO 3 ) 3 .6H 2 O, 99.9%, Alfa Aesar) was dissolved in deionised water to produce a 1 M solution In a separate vessel a 1 M solution of di ammonium hydrogen phosphate ((NH 4 ) 2 HPO 4 98%, Alfa Aesa r) was prepared by dissolving the solid in deionised water. Both the above powders are readily soluble in water and the bottom of the vessel was inspected regularly to ensure the completion of

PAGE 46

46 the dissolution process. The di ammonium hydrogen phosphate s olution was added to the lanthanum nitrate solution drop wise which almost instantaneously lead to the formation of a white precipitate. The solution was allowed to stir for 2 h after the mixing of the solution for the reaction to complete. T he solution containing the precipitate was allowed to drip through filter paper to remove the precipitate from the solution by filtration. The obtained precipitate was then dried at 120 o C and ground with mortar and pestle under ethanol. Grinding under ethanol is mor e effective as compared to dry grinding and leads to more efficient reduction of agglomeration in the material. The dried precipitate was then calcined at 1000 o C for 2 h using ramp rate of 200 o C/h. At ea ch step of the synthesis procedure a small part of the powder was kept aside for x ray diffraction analysis to evaluate the evolution of the sample at each step of the synthesis process The calcined powder was then mixed with 2 wt% of polyvinyl alcohol (PVA) using a mortar and pestle. The PVA acts as a binder material and contributes to the strengthening of the green body. The powder was then uniaxially pressed into cylindrical samples using a hydraulic press (~200 MPa) The green pellets were sintered at 1400 o C for 10 h using a ramp up and ramp dow n rate of 200 o C/h. Sacrificial powder of the same composition was placed below the green body to restrict possibility of contamination from the surface of the alumina setter used for sintering. During the ramp up step of the sintering heat treatment proc edure the sample was held at 450 o C for 2 h to allow for the combustion of the organic compound used as binder during green body preparation.

PAGE 47

47 3.1.3 Tin (IV) P yrophosphate (SnP 2 O 7 ) The tin (IV) pyrophosphate powders ( SnP 2 O 7 ) were prepared by heating SnO 2 ( tin (IV) oxide, 99.9%, Alfa Aesar) with some excess of H 3 PO 4 (phosphoric acid, 85%, ACS) as summarized in Figure 3 2 A 1:3 ratio of Sn:P was used instead of 1:2, to compensate for the loss of phosphoric acid during the heating procedure. To prepare the acceptor doped SnP 2 O 7 samples, an appropriate mixture of the oxide of the dopant (D = In, Al, Ga, Zn, and Sc) and SnO 2 was used to react with the phosphoric acid Figure 3 2. Flowchart for powder synthesis, sample preparation and the various character ization techniques used. To the above mixture 100 mL of deionised water was added and the contents were heated and maintained at a temperature of 300 o C The solution was stirred throughout at 300 rpm using a Teflon TM magnetic stirrer to aid the dissolu tion. At temperatures in excess of 200 o C vigorous boiling takes place in the phosphoric acid

PAGE 48

48 due to rapid evaporation of the water in the solution. After evaporation of the water a clear thick viscous solution is formed. The oxides are soluble in pure phosphoric acid at these temperatures. Throughout the process the bottom of the glass beaker was regularly inspected and any oxide powder which was stuck to the bottom of the beaker was scraped off using a spatula In the initially prepared samples SnO 2 impurities could be detected from powder XRD which was most likely due to unreacted SnO 2 The prepared viscous slurry was then transferred to an alumina crucible. The crucible was covered with a ceramic cover and calcined for 2.5 h at 650 o C. Phosphoric acid is volatile above 600 o C and rapidly evaporates above this temperature. The use of a crucible cover and excess phosphoric acid in the starting mixture compensate for the loss of phosphoric acid during calcination. The calcined powder obtained was gr ound using mortar and pestle and then sieved though a 212 m mesh. A small part of the powder was removed for x ray diffraction (XRD) analysis To prepare a green body, the powder sample was mixed with 2 wt. % of PVA which is used as a binder. The synt hesized powder was mixed with the binder using mortar and pestle until the powder was dry. The prepared powder was uniaxially pressed into flat cylindrical samples using a hydraulic press. The pressure during compaction was kept constant at 200 MPa by mo difying the applied force according to the inner diameter of the die used for pressing. F lat cylindrical pellets with diameter of 7 mm and 12 mm with a thickness of about 1 3 mm were synthesized The pellets wi th the larger diameter (12 mm) we re requir ed for transference number measurements. This is because the inner cylindrical ceramic tube inside the quartz reactor needs to be completely covered

PAGE 49

49 and sealed using the sample to cover the tube so that differential partial pressures of hydrogen can be ma intained across the sample. The green pellets were placed in a covered alumina crucible with sacrificial powder of the same material and then the crucible cover was sealed using ceramic cement (EA139 B50, Saint Gobain TM ). The sacrificial powder restrict s potential contamination of the sample from the inner surface of the crucible while sealing of the crucible using ceramic cement is required to prevent decomposition of the sample during sintering. Decomposition of the pyrophosphate to the orthophosphate with the release of volatile phosphorus pentoxide (P 2 O 5 ) during the sintering heat treatment procedure leads to insufficient densification of the sample. The sealed crucibles were then heated to 1400 o C for 10 h using a ramp up and ramp down rate of 200 o C /hr During the ramp up step of the sintering heat treatment procedure, the sample was held at 450 o C for 2 h to ensure complete combustion of the organic binder used for preparing the green body. The relative density of the sintered pellets was above 90% Undoped SnP 2 O 7 pellets had a higher relative density of about 95%. Some pellets were crushed and ground for powder XRD analysis. 3.2 Phase A nalysis 3.2.1 X R ay D iffraction (XRD) A small part of the prepared LaP 5 O 14 crystals were crushed to powder wit h a mortar and pestle under acetone. These powders were then sieved through a 212 m sieve The theoretical powder pattern of LaP 5 O 14 used for comparing with the experimental powder pattern was generated using CrystalDiffract TM from previously reported c rystal structure data. 25 In addition, XRD was also performed on single crystals of the prepared samples. The relative orientation of the crystal during single

PAGE 50

50 crystal XRD data collection is shown in the schematic in Figure 3 3 (a). Th e main face t of the crystal is indicated as the (100) plane on the SEM image shown in Figure 3 3 (b). The synthesized pellets of LaPO 4 and SnP 2 O 7 were crushed and ground under acetone for powder XRD analysis. The powders were sieved through a 212 m sieve The XRD data was collected on the instrument (Phillips APD 3720) using CuK radiation and the operating parameters for the same were set to 40 kV and 20 mA. (a) (b) Figure 3 3 (a) Orientation of LaP 5 O 14 crystals during single crystal X RD data collection (b) SEM image of an LaP 5 O 14 crystal showing the (100) plane. 3.2.2 Energy D ispersive S pectroscopy (EDS) Previous to all further experimentation the crystals of LaP 5 O 14 were dried in air at 600 o C for 3 h. This was done to remove any re maining phosphoric acid that could have been sticking to the surface sample since the crystals were grown from a fused phosphoric acid. Initial samples that were not subjected to this heat treatment showed a very high P:La ratio as compared to the expecte d value in EDS analysis. The pellets of LaPO 4 and SnP 2 O 7 were sintered at temperatures above 1000 o C and hence did not require the drying heat treatment procedure.

PAGE 51

51 Previous to EDS analysis the samples were attached to a cylindrical aluminum holder using conductive double sided carbon tape. The prepared samples were then gold coated using the sputter coating technique for 3 min at 30 mA to deposit a conductive coating on the surface of the sample. 3.2.3 Scanning E lectron M icroscopy (SEM) The ceramic sam ples were ultra sonicated in ethanol for 5 min to remove any surface contamination and dust particles previous to SEM analysis. The samples were then placed on aluminum holders using conductive double sided carbon tape. To deposit a conductive coating on the sample surface was sputter coated with gold for 3 min at 30 mA. SEM analysis was performed on the instrument JEOL 6400. During SEM analysis an accelerating voltage of 15 kV was used and sample working distance was set to 15 mm. 3.2.4 Magic A ngle S p inning N uclear M agnetic R esonance (MAS NMR) P owder samples w ere prepared by crushing sintered samples of SnP 2 O 7 under ethanol. The samples were then packed tightly into 4 mm zirconia rotors for MAS NMR spectroscopy. It was ensured that the powder w as packed firmly and uniformly in the rotor as uneven packing prevents the rotor from spinning at high speeds. Inside the instrument the rotor is levitated using an upward bearing pressure of nitrogen gas while spinning is accomplished by using a transver se drive pressure of flowing nitrogen. Thus a non uniform weight distribution resulting from uneven packing in the rotor leads to improper balancing of the rotor in the chamber which is detrimental to achieving high spinning speeds. The 31 P MAS NMR sp ectra were acquired at a magnetic field of 9.4 T using a Bruker Avance NMR spectrometer operating at 161.974 MHz. Each sample was spun

PAGE 52

52 at approximately 10 kHz. The fully relaxed 31 P spectra were acquired at ambient temperature by signal averaging 128 tra nsients using a recycle delay of 4 s and a /2 pulse width of 5 s. A Gaussian line broadening of 50 Hz was applied prior to Fou rier transformation. Chemical shifts are referenced to 85 % orthophosphoric acid. Figure 3 4. Setup used for conductivity measurement and proton transference number measurement. 3.3 Electrochemical I mpedance S pectroscopy (EIS) and Proton T ransference N umber M easurement 3.3.1 Experimental S etup Ionic conductivity measurements and proton transference measurements were made u sing setup shown in Figure 3 4 This setup includes one furnace ( Lindberg/Blue

PAGE 53

53 M, HTF55322A), three m ass flow controllers (Sierra instruments, C100L ), one custom designed quartz reactor (Analytical Research Systems) and an i mpedance analyzer (Solartron 12 60). A schematic diagram showing the setup is shown in the inset in Figure 3 4 (a) (b) Figure 3 5. (a) Schematic diagram and (b) picture of quartz reactor used for conductivity measurement and transference number measurement. The mass flow contr ollers (located behind the furnace) control the f low and concentration of gases flowing into the quartz reactor which is located inside the furnace. For impedance measurements at different temperatures the sample was placed inside the quartz reactor, its elf placed inside the tube furnace. To minimize the errors in temperature measurement an independent thermocouple was placed right next to the sample.

PAGE 54

54 A schematic of the design of the quartz reactor and the reactor itself are shown in Figure 3 5 The reactor has two separate gas inlets and outlets for two different chambers (flow of gases indicated by arrows) such that the atmosphere in both the chambers can be controlled independently. (a) (b) (c) Figure 3 6. Prepared samples of (a ) SnP 2 O 7 (b ) LaP 5 O 14 and (c ) sealed sample of SnP 2 O 7 for proton transference number measurement. Such a design is needed for proton transference number measure ment since it requires a hydrogen partial pressure gradient to be maintained across the sample. The two chambers are separated by placing the sample to be measured on the alumina tube s o that it produces an airtight junction as described in a lat er section. Within the reactor gold wires are used to make the connection from the sample surface to the BNC connectors attached to the reactor which are then connected to the impedance analyzer. Within the reactor the connecting gold wires are encased in an alumina tube the exterior of which is covered in platinum paint. Temperature regulation of the furnace w as done by connecting the thermocouple within the reactor to the furnace controller.

PAGE 55

55 3.3 .2 Sample P reparation and EIS For conductivity measurement on crystals of LaP 5 O 14 the crystals were gold sputtered on both the main facets (the (100) surfaces) with the sputter coater An SEM image of a coated and wired sample of LaP 5 O 14 is shown in Figure 3 6 (b). To prevent the sputter coater from coating the sides of the sample a masking paper with a hole of appropriate dimensions was used to cover the sample du ring the coating procedure. For conductivity measurement on sintered LaPO 4 and SnP 2 O 7 pellets, the pellets were painted on both sides with platinum paste (CL11 5349, Heraeus) to serve as the e lectrode. The pellets were sintered at 900 o C for 1 h using a r amp rate of 200 o C/h. Platinum wires (99.9%, 0.127 mm dia.) were attached to the electroded surfaces of the coated pellet s/crystals using silver paste (part # 5063, SPI Supplies) as shown in Figure 3 1 Three samples with each composition were used for co nductivity measurement. A typical coated and wired sample of SnP 2 O 7 is shown in Figure 3 6 (a). EIS measurements were done in unhumidified or humidified air (obtained by bubbling nitrogen gas through a saturated solution of MgCl 2 in de ionized water) in the temperature range of 300 o C 600 o C using the two point probe technique (Solartron SI 1260). For humidified argon measurements the gas was flown through a water bubbler before passing through the reactor. The AC amplitude during impedance data collec tion was kept at 1000 mV. Frequency sweeps were taken in the range 32 MHz 0.1 Hz Fitting of the spectra was carried out using ZView TM software. 3.3. 3 Transference N umber M easurement The experimental setup used for transference number measurement is sh own in Figure 3 4 Platinum paste was applied to both sides of the sample and platinum wires

PAGE 56

56 were attached using silver paste as described previously The sample was attached to the tube using ceramic sealing cement (Cermabond 503 Aremco). An optical i mage of the sealed sample is shown in Figure 3 6 (c). The whole setup was heated to 450 o C and held at that temperature overnight for the seal to form. Gas flow into the reactor was controlled by using mass flow controllers Open circuit voltage was meas ured by using a multimeter (model # 34410A, Agilent Technologies). The transference number measurement was carried out by using the following dry hydrogen concentration cell, 80 vol.% H 2 + balance N 2 Pt/D 0.1 Sn 0.9 P 2 O 7 /Pt, 4 vol .% H 2 + balance N 2 The total flow rate was controlled at 100 sccm in both the chambers of the reactor. Under a hydrogen partial pressure gradient t he theoretical electrom otive force can be expressed as follows : (3 1) Where R is the universal gas constant, T is the absolute temperature and F is the If the material is a mixed ionic conductor then the measured value of the electromotive force is less than the theoretical value calc ulated above. The protonic transference number can be expressed as follows. (3 2) Proton transference number measurements were performed on undoped and 10 mol% In 3+ doped SnP 2 O 7 samples in the temperature range 300 o C 50 0 o C.

PAGE 57

57 3.4 Simulation M ethods The starting point for performing lattice simulations on a material is the development of a suitable potential model for the system under consideration. The potential model is a mathematical description of the energy of the s ystem as a function of the particle co ordinates. The reliability of the results predicted from the simulations then depends on how aptly the potential model depicts the material system. 3.4.1 Potential Model In accordance with Born model, the interatomi c forces are separated into the long range Coulombic terms and short range terms that correspond to electron cloud overlap (Pauli repulsion) and attractive Van der Waals interactions. Mathematically the interatomic potential model can be represented as follows 26 (3 3) The first term on the right hand side of the equation is the summation of all the energy terms arising from the electrostatic (or Coulombic) interaction between the component ions of the lattice The second and the third terms represent the summation of the two body and three body interactions which arise due to the short range interac tion between the ions in the system. The importance of the short range interaction terms and their overall contribution to the lattice energy increases with increasing deviation from ionic bonding in the compound under investigation. 3.4.2 Buckingham P otential The two body interactions are generally described by the Buckingha m interatomic potential model which shown below 26 :

PAGE 58

58 (3 4 ) Where A ij ij and C ij are empirically derived parameters for each set of ion ion interactions in the material. Equation 3 4 is the summation of a repulsive and an attractive interaction term. The exponential term always ha s a positive value (since A is positive) and hence represents a repu lsive force in the interaction The second potential energy term always has a negative value (since C is positive) and hence represents an attractive force in the interaction The inter atomic potential parameters used for the lanthanum phosphate compounds are summarized in Table 3 1. The parameters for La O and O O interactions have been transferred from previously published work, but those for P O were empirically derived for this stud y. Ta ble 3 1. Interatomic potential parameters for LaPO 4 and LaP 5 O 14 Interaction A (eV) ( ) C (eV. 6 ) Reference La 3+ 2 4579.23 0.3044 0.00 27 P 5+ 2 1250.00 0.3240 0.00 Empirically derived. O 2 2 22764.30 0.1490 27.88 27 The shape of the Buckingham potential when plotted as a function of the distance r shows a potential well at an equilibrium distance. At distances lower than the equilibrium distance the potential rises rapidly due to a high value of repulsion while at distances greater than the equilibrium distance the potential gradually rises to almost zero at large distances. The two body interaction term represented b y the Buckingham pot ential does not have any angle dependent components unlike the three body interaction term described in the following section.

PAGE 59

59 3.4.3 Harmonic Angle Bending Potential The three body term takes the form of a harmonic angle bending potential about the central ion 26 : (3 5) Where k is the force constant and o is the equilibrium bond angle. The three body interaction terms are generally used for only selective interactions in the material. The y a re commonly used to account for the angle dependent nature and rigidity of the polyhedral structural units that o ccur in some compounds such as silicates (SiO 4 tetrahedra) gallates (GaO 4 tetrahedra), niobates (NbO 4 tetrahedra) and phosphates (PO 4 tetrahedra). 27 For the both lanthanum phosphate compounds (LaPO 4 and LaP 5 O 14 ) the three body interaction term was applied to all the O P O interactions within each PO 4 tetrahedron in the materials. Specifically for LaP 5 O 14 an additional three body term was included in order to account for the P O P bond in the material The overal l contribution from three body interaction terms to the lattice energy is small. However, they are known to have significant impact of the vibrational properties of the material. The P O P bond angle is the angle between two neighboring corner sharing PO 4 tetrahedra which has an average value of ~134 o for all condensed phosphates. Since the structure of LaPO 4 does not have any corner sharing PO 4 tetrahedra the P O P three body interaction term was not included in the input file. 3.4.4 Ionic P olarizat ion (Shell Model) As the presence of charged defects in the material will polarize the ions in the surrounding lattice, the ionic polarizability of the ions must be accounted for in the potential model. The shell model developed by Dick and Overhauser 28 account s for

PAGE 60

60 such effects by treating each ion in terms of a core and a shell which are connec ted by a harmonic spring. The core represents the nucleus and the core electrons. T he mass less shell represents the polarizable valence shell electrons. A schematic representation of the shell model of an ion and the relevant parameters is shown in Fig ure 3 7. The spring constant is denoted by K and the core and shell charges by Y c and Y s (which must add up to the total charge on the ion being modeled). Figure 3 7. Diagrammatic representation of shell model parameters. The shell model parameter to known physical properties of the material such as dielectric constant, elastic modulus or the crystal structure of the material. The polarization of the ion resulting in a dipole moment is cause d due to the displacement of the shell with respect to the core. The shell model parameters used for LaPO 4 and LaP 5 O 14 were obtained from previously published literature and are listed in Table 3 2.

PAGE 61

61 3.4.5 Defect Simulation (Mott Littleton Approach) Th e existence of a charged defect in a lattice causes a significant alteration in the positions of the surrounding lattice ions Thus there is significant rearrangement of the nearby lattice in response to the presence of a defect site which is commonly ref erred to as relaxation. In the case of ionic solids the effect is much more pronounced because the perturbation provided by the defect is mainly Coulombic in origin which is a long range force. Thus to accurately calculate the energies associated with the formation of a defect an intensive treatment of lattice relaxation is necessary while restricting the required computation time to a reasonable value Table 3 2 Shell model parameters for LaPO 4 and LaP 5 O 14 shell model Species Y (e) K (eV. 2 ) Re ference La 3+ 0.25 145.0 27 P 5+ 5.0 0 99999 .0 29 O 2 2.96 65.0 29 For this purpose the region around the defect is divided into two concentric spherical regions. This approach for calculating the defect ene rgy is commonly referred to as the Mott Littleton approach. 30 The inner region which consists of about 250 atoms is atomistically simulated using the specified interatomic potentials Within this region all the coordinates (shell and core coordinates for each ion) are adjusted until the confi guration with the minimum energy is obtained. In this inner region the neighboring lattice experiences strong forces due to the close vicinity of the defect and hence atomistic simulation is necessary to model the relaxation phenomenon accurately By inc reasing the size of the inner region and observing the effect on the calculated energy

PAGE 62

62 of the defect it can be ascertained if the values are sensitive to an expansion of the region. The forces experienced by the ions in the outer region are much weaker a nd hence the ions in the outer region are treated by more approximate quasi continuum methods According to the Mott Littleton approximation the response of the outer region can be regarded as a dielectric response to the charged defect introduced in th e lattice. 30 Table 3 3 Parameters for the O H interaction Morse potential D c (eV) ( 1 ) r o ( ) O ... H 7.05 2.1986 0.9485 Buckingham potential A (eV) ( ) C (eV 6 ) 311.97 0.25 00 0.0 000 3.4.6 Proton Oxygen (O H) Interaction For the protonic defect, the O H interaction was mo deled using an attractive Morse potential The Morse potential can be mathematically expressed according to the following equation: (3 6) Where D is the depth of the potential well r o is the equilibrium distance between the oxygen ion and the associated proton and is an empirically obtained parameter The values used for the above parameters were obtained from previously published literature and are summarized in Table 3 3 27 The values listed in Table 3 3 were derived f rom quantum mechanical cluster calculations by Saul and Catlow on sodium hydroxide (NaOH) 31,32 The surrounding lattice was represented using point charges corresponding to the ions in the lattice. The dipole of the hydroxyl species was distributed across both ions with O assigned 1.4263 and H +0.4263 to give and overall

PAGE 63

63 charge of 1 for the protonic defect. Such a charge distribution across the hydrogen oxygen correctly mimics the dipole moment of the hydroxyl group and is in good agreement with the charge distribution calculated from quantum mechanical studies on perovskite type proton conducting material 33 To describe the interaction between the hydroxyl grou p and the lattice ions an additional Buckingham potential was included which was incorporated from previous studies of water incor poration in silicates. 27 3.4.7 Energy Minimization Energy minimization is used dur ing calculation of defect energies as well as to reproduce the crystal structure of the material using the relevant interatomic potential parameters. The simplest energy minimization procedure consists of scanning the entire para meter space to find the mi nimum energy configuration However, such a process is computationally expensive and almost impossible in most cases due to the large number of parameters involved. Gradient techniques are much more efficient for energy minimization as compared to a sim ple scanning of the parameter space. The Newton Raphson procedure is commonly used to locate the energy minimum in the structure. According to this procedure if the initial value of a parameter is known then the next guess can be made according to the following equation 26 : (3 7) Where x n is the initial guess, x n+1 n ) is the first derivative of The n ) which

PAGE 64

64 corresponds to a minima in the function f(x). The use of this procedure can greatly reduce the number of iterations required for the optimization process.

PAGE 65

65 CHAPTER 4 C HARACTERIZATION OF L ANTHANUM ULTRAPHOSPH ATE (L a P 5 O 14 ) 4.1 Ultraphosphates Ultraphosphate anions have the general formula [P (n+2) O (3n+5) ] n where commonly n = 2, 3, 4, or 6. However, a compound calcium yttrium ultraphosphate (CaYP 7 O 20 ) with n = 5 has also been reported. 34 A list of the several of the synthesized ultraphosphate materials is summarized in T able 4 1 In general, the type of ultraphosphate that is formed depends on valence of the participating cation. For example the rare earths with a valence of +3 generally tend to form ultraphosphate s with the general formula RP 5 O 14 where n = 3. However, exceptions are seen in the case of calcium ultraphosphates which have been synthes ized with at least three different stoichiometries CaP 4 O 11 (n = 2), Ca 3 (P 5 O 14 ) 2 (n = 3) and Ca 2 P 6 O 17 (n = 4) which depend upon the synthesis conditions. 35 37 All the types of ultraphosphates share some similariti es in the general arrangement of PO 4 tetrahedra in the condensed phosphate anion. The arrangement consists of rings of corner sharing PO 4 tetrahedra which yield a ribbon like morphology of the phosphate anion in most cases However the key difference l ies in the number of PO 4 tetrahedra participating in the formation of the condensed rings which is ten ( for n=2) and eight ( for n=3). 25,38 In the rare earth ultraphosphates (RP 5 O 14 where R = rare earth elements ) f our different structural types have been observed. 18 The differences (between the I and II type) mainly lie in the slightly different orientations of some t etrahedra that either lend inversion symmetry or mirror symmetry to the anion. In the materials of type III, the anion structure is considerably different from the first two types.

PAGE 66

66 Some materials such as HoP 5 O 14 exhibit all the three structures depending on the synthesis conditions while only one material belonging to type IV is known (CeP 5 O 14 ). 18 There have been studies on s everal ultraphosphate materials wit h a mixed cation stoichiometry such as Na 3 MP 8 O 23 (M = Cr, V, Mn, Al, Fe or Ga), NdLa(P 5 O 14 ) and YbEr(P 5 O 14 ), CaYP 7 O 20 and NiHP 5 O 14 most of which are only structural investigations 34,39,40 Table 4 1. Classi fication of ultraphosphate compounds with examples. Type Compounds References N = 2 ZnP 4 O 11 CaP 4 O 11 NiP 4 O 11 CuP 4 O 11 Na 2 P 4 O 11 MgP 4 O 11 SrP 4 O 11 CdP 4 O 11 MnP 4 O 11 CoP 4 O 11 SrP 4 O 11 41 47 N = 3 LaP 5 O 14 EuP 5 O 1 4 DyP 5 O 14 MnP 5 O 14 GaP 5 O 14 NdP 5 O 14 GdP 5 O 14 BiP 5 O 14 CeP 5 O 14 HoP 5 O 14 SmP 5 O 14 YP 5 O 14 YbP 5 O 14 LuP 5 O 14 YbEr(P 5 O 14 ) 2 NdLa(P 5 O 14 ) 2 Ca 3 (P 5 O 14 ) 2 NiHP 5 O 14 25,36,48 57 N = 4 Sr 2 P 6 O 17 Cd 2 P 6 O 17 Ca 2 P 6 O 17 37,46 N = 5 CaYP 7 O 20 34 N = 6 Na 3 MP 8 O 23 (M = Cr, V, Mn, Al and Ga), Na 3 FeP 8 O 23 39,40 The synthesis of ultraphosphates has almost exclusively been performed by precipitation from concentrated phosphoric aci d solution s containing the dissolved oxides of the respective cations in the temperature range of 400 o C 800 o C This is because f used orthophosphoric acid (H 3 PO 4 ) tends to form condensed polyphosphoric acids of higher complexity at elevated temperatures by thermal dehydration. Details about the process of thermal dehydration can be found in section 2.4.2. There has been extensive research on single crystals of neodymium ultraphosphates as a potential candidate material for application in lasers. 56,58 Also

PAGE 67

67 alkaline ultraphosphates, mainly those of sodium and calcium have been investigated as food preservative s 44,59 61 From the viewpoint of prospecti ve applications as proton exchange membranes s everal different orthophosphate s have been investigated, including RPO 4 (R = La, Sm, Nd, Ce), Ba 3 Ce(PO 4 ) 3 and some pyrophosphate materials such as SnP 2 O 7 and TiP 2 O 7 62 65 However, the number of studies on higher condensed phosphate materials such as metaphosphates an d ultraphosphates is very limited. 66 Figur e 4 1. View of LaP 5 O 14 along the [001] direction. La 3+ ions shown in orange, PO 4 tetrahedra shown in green and oxygen ions are shown in red. 4.1 Crystallochemical A nalysis The crystal structure of LaP 5 O 14 along the [001] direction is shown in Figure 4 1 (using structural data by Schonherr et al. 67 ). LaP 5 O 14 is isomorphic with other ultraphosphates such as EuP 5 O 14 and GdP 5 O 14 and belongs to a monoclinic crystal system ( s pace group: P2 1 /c (No. 14), a = 8.8206 b = 9.1196 c = 13.1714 and = 90.661 o ) 67 The basic structural units comprising the phosphate anion are PO 4 tetrahedra (shown in green) which are common to all condensed phosphates as shown

PAGE 68

68 in Figure 4 1 The La 3+ ions ( shown in ora nge) are located in an eight coordinated oxygen ion environment. Figure 4 2. Eight coordinated La 3+ ion (shown in orange). The eight oxygen ions belong to PO 4 tetrahedra from adjacent phosphate anion ribbons. The phosphate anion structure can be v isualized as consisting of theoretically infinitely long ribbons of condensed PO 4 tetrahedra (formed by corner sharing an oxygen ion between neighbo ring tetrahedra) which extend along the [100] direction in the cr ystal structure. Such a ribbon like struct ure for the phosphate anion is commonly observed in ultraphosphates. 25 The length of the ribbon is theoretically considered to be infinite, but practic ally its length is obviously limited by material defects and discontinuities in the crystal. The individual anionic ribbons are held together by the La 3+ ions as shown in Figure 4 2 Each La 3+ ion is bound to eight oxygen ions belonging to four pairs of PO 4 tetrahedra from four surrounding ribbons. The structural arrangement of PO 4 tetrahedra in an individual ribbon segment is shown in Figure 4 3 along two orthogonal directions, namely [010] and [001]. Individual ribbons are made up of rings composed of 8 corner sharing PO 4 tetrahedra. It can be

PAGE 69

69 seen by comparing the two figures that the spatial arrangement of the condensed tetrahedra leads to a three dimensionally networked phosphate anion structure. Figure 4 3. Segment of phosphate anion ribbon composed of PO 4 tetrahedra as viewed along two directions (a) [010] and (b) [001] direction. In the following part of this section a crystallochemical analysis of LaP 5 O 14 i s been presented. It is known that proton transport in solid state proton conduc ting materials takes place via the Gr tthuss mechanism, and hence the (O O) distance (or proton jump distance) is a key parameter which governs the magnitude of the activation energy for the proton transfer process. A detailed description of the above can be found in section 6. 2 The oxygen ions in the structure of LaP 5 O 14 can be broadly classified in two categories, namely bridging oxygen and terminal oxygen ions. The former are

PAGE 70

70 oxygen ions that are bound to two P 5+ ions and thus are at the vertex of tw o corner sharing tetrahedra. The terminal oxygen ions, on the contrary are bound to one P 5+ ion and La 3+ ion. The difference is illustrated in Figure 4 4 (a) which shows a small segment of the condensed phosphate anion. (a) (b) (c) Figure 4 4. (a) Location of bridging and terminal oxygen in LaP 5 O 14 Pair distribution function for LaP 5 O 4 centered at (b) bridging (c) terminal oxygen ion.

PAGE 71

71 T he differences in the oxygen co ordination environment of the two types of oxygen ions are shown using pa ir distribution functions in Figure 4 4 (b) and (c). Around the bridging oxygen ion there are six neighboring oxygen ions available for the proton jump with distances in the range 2.39 2.58 This can be correlated easily with the illustration in Figu re 4 4 (a) where it can be seen that the six corresponding nearest oxygen ions would be O5, O7, O25 and O4, O14, and O34 belonging to the two corner sharing tetrahedra at whose vertex O45 is located. At larger distances in the range 3.0 3.2 three othe r oxygen ions can be seen which belong to either second nearest neighbor tetrahedra or to tetrahedra from the adjacent phosphate ribbon As the activation energy for the proton transfer process is directly related to the (O O) distance, the jump coordinat ion for the bridging oxygen ion is denoted as six in Figure 4 4 (b). Table 4 2 Comparison of mean inter and intra tetrahedral oxygen oxygen ion distances in LaPO 4 and LaP 5 O 14 LaP 5 O 14 Mean oxygen oxygen ion distance ( ) Intra tetrahedra Inter tet rahedra P1O 4 2.5150 0.0 5 3.0706 0.14 P2O 4 2.5055 0.07 3.1 382 0.1 2 P3O 4 2.5217 0.0 5 3.0683 0.12 P4O 4 2.5032 0.07 3.1176 0. 10 P5O 4 2.5155 0.04 3.0807 0. 11 Similarly the pair distribution function for the terminal oxygen ion shown in Figure 4 4 (c) can be correlated to the illustration in Figure 4 4 (a). Around the terminal oxygen ion O4 there are three nearest oxygen ions (O4, O14 and O45) with distances in the range 2.57 2.60 At larger distances in the range 2.85 3.0 th ere are two additional oxygen ions which either belong to the second nearest neighbor tetrahedra or to tetrahedra from the adjacent phosphate ribbon. Thus the jump coordination number is denoted as 3 for the proton transfer process.

PAGE 72

72 Figure 4 5. Theor etical powder pattern for LaP 5 O 14 compared with experimental powder patterns for the LaP 5 O 14 samples. Although the above analysis is only presented for the P4O 4 tetrahedron it applies to all the five crystallographically different tetrahedra in the stru cture. The resulting mean (O O) distances for all the tetrahedra are summarized in Table 4 2 Thus, i t can be seen that within the same tetrahedron the O O distances are always < 2.6 Since, the tetrahedra are linked together (by corner sharing an oxy gen ion) the jump distances encountered by a diffusing proton will always be less than this value, provided the proton diffuses along the three dimensional network of condensed PO 4 tetrahedra In light of the above analysis it can be inferred that proton conduction along the networked

PAGE 73

73 anion structure is more energetically favorable as compared to other possibilities available for conduction (detailed discussion in section 6.2) Figure 4 6. Theoretical powder pattern for LaP 5 O 14 compared with experimen tal powder patterns for the LaP 5 O 14 samples (15 o 35 o ). 4.2 Phase A nalysis 4.2.1 X R ay D iffraction (XRD) The experimental powder diffraction pattern (powder prepared from crushed crystals) for all the prepared compositions of LaP 5 O 14 is compared with t he theoretical pattern for t he same in Figure 4 5 The same plot can be seen in greater detail in Figure 4 6. The theoretical pattern was generated (using Crystal Diffract software) from the structural data reported previously by Schonherr et al 25 for LaP 5 O 14 All the peaks

PAGE 74

74 observed correspond to the lanthanum ultraphosphate phase. The obtained XRD patterns of both the doped materials were compared with other compounds containing Sr, La, P and O namely, SrP 4 O 11 Sr 2 P 6 O 17 and Sr 3 La (PO 4 ) 3 to see if the addition of the dopant had lead to the formation of any additional phases. However, no secondary phase formation could be detected from the obtained XRD pattern. The geometrical density of the undoped crystals measured was 3.23 g.cm 3 (99.4% of the relative density) while the Archimedes density was 3.20 g.cm 3 (98.5% of the relative density). These values are close to the theoretical density which is 3.25 g.cm 3 The density measured by the Archimedes method was slightly lower than the geometrically measured density most likely due to the presence of minute air bubbles on the crystal surface. (a) (b) Figure 4 7. (a) SEM image and (b) EDS spectrum of a typical crystal of undoped LaP 5 O 14 4.2.2 Scanning Electron Microscopy (SEM) Scanning electron microscopy (SEM) image of a typical sample of undoped LaP 5 O 14 crystal is shown in Figure 4 6 (a) Undoped LaP 5 O 14 crystals were in the form

PAGE 75

75 of rhombus shaped flat sheets. The main facet of the crystals (which can be seen in the tw o SEM images) was placed parallel to the surface of the sample holder for single crystal XRD data collection. Table 4 3. Results from quantitative EDS analysis on the undoped and 1 mol% Sr doped LaP 5 O 14 samples. Undoped LaP 5 O 14 1 mol% Sr doped LaP 5 O 14 La 13.15 0.29 La 13.27 0.45 P 66.65 0.69 P 67.01 0.70 Au 20.21 0. 40 Au 19.72 1.15 P:La 5.07 0. 12 P:La 5.05 0. 12 4.2.3 Energy Dispersive Spectroscopy (EDS) In the electron dispersive spectra (EDS) obtained from the undoped a nd doped LaP 5 O 14 samples peaks corresponding to La, P, O, and Au (since samples were sputter coated with Au previous to EDS) were detected as shown in Figure 4 6 (b) The P/La ratio for both the doped and the undoped samples was ~ 5 as expected for LaP 5 O 1 4 as shown in Table 4 3 The signal from strontium atoms could not be distinctly observed since their expected concentration is 0.05 atomic % which is below the detection limit of the instrument (~1 atomic %). Previous to sputter coating the crystal sam ples for EDS, the crystals were dried in air at 600 o C for 2 hours. This was done to ensure that the crystals were free of any surface phosphoric acid which may interfere with the quantitative EDS analysis. The initial samples that were not subjected to t his heat treatment showed very high phosphorus content. 4 .2.4 Single C rystal X R ay D iffraction The XRD pattern obtained for LaP 5 O 14 single crystals is shown in Figure 4 8 (a) For XRD data collection the crystal was placed on the holder with its surface with

PAGE 76

76 maximum area facing the surface of the holder. The relative orientation of the crystal with respect to source and detector is shown in Figure 4 8 (b) In this orientation peaks corresponding to the (100) plane ( from several orders of reflection ) co uld be seen in the XRD pattern obtained from crystals of undoped LaP 5 O 14 (a) (b) Figure 4 8. (a) Single crystal XRD pattern for LaP 5 O 14 samples (b) Relative orientation of the crystal during XRD data collection. Similarly peaks from the (1 00) plane were also observed for 1 mol% and 5 mol% Sr doped LaP 5 O 14 (Figure 4 8 (a )). It is clear from the above information that the main facet of the crystal is the (100) plane for the doped and the undoped crystals. The relative orientation of the pho sphate anion ribbons with respect to the (100) plane in the crystal can be seen in the sche matic representation in Figure 4 8 (b ). T o investigate the effect of structural anisotropy in the material on the electrical properties conductivity measurements we re made along the [100] direction as well as along a transverse direction. All the peaks seen in the single crystal XRD pattern of the doped and the

PAGE 77

77 undoped samples were accompanied by an adjacent satellite peak with about half the relative intensity. Th ese peaks were due to the presence of K ( = 1.544390 ) in the incident x ray beam. Figure 4 9. Nyquist plots for 5 mol% Sr doped LaP 5 O 14 samples measured at two different temperatures. 4.3 Conductivity M easurement 4.3.1 Effect of D opant C ontent Single crystals with three different compositions, undoped LaP 5 O 14 1 and 5 mol% Sr doped LaP 5 O 14 were grown for conductivity measurements. Three crystals of each composition were measured and the average conductivity values at each temperature were calculated. Typical Nyquist plots for 5 mol% Sr doped LaP 5 O 14 samples at two different temperatures are shown in Figure 4 9 All the plots obtained for all samples of LaP 5 O 14 exhibited two impedance arcs corresponding to the bulk impedance (at higher

PAGE 78

78 frequency) and the electrode impedance (low er frequency). In general ceramic materials exhibit a three arc response, where each arc corresponds to bulk, grain boundary and electrode impedance in decreasing order of frequency. 23 However, due to absence of grain boundaries in the single crystals of LaP 5 O 14 we observe a two arc response in the Nyquist plot. An equivalent circuit consisting of a two R (resistance) CPE (constant phase element) circuit elements in series was used to fit the obtained impedance spectra. The conductivity of the sample was calculated by normalizing the obtained resistance with the known dimensions of the crystal sample. Table 4 4. Peak frequencies and capacitances associated with bulk and electrode impedance for all compositions of LaP 5 O 1 4 Dopant content C bulk (pF) Undoped 13.31 0.43 1 mol% 2.79 0.23 5 mol% 8.70 0.29 The capacitance associated with any impedance arc can be calculated by knowing the peak frequency for the arc which is shown in Figure 4 9 The capacitance associated with the bulk impedance for 5 mol% Sr doped LaP 5 O 14 was found to be independent of the temperature and the calculated value was 8.70 0. 29 pF (see Table 4 4) The value is similar to the capacitance for bulk impedance for other ionic conductors thus confirming that the arc observed in the Nyquist plot was correctly ascribed to the bulk impedance 23 On the contrary, the capacitance c alculated for the ele ctrode impedance was much lower as expected The measured conductivity of the LaP 5 O 14 samples increased as the dopant content was increased upto 5 mol% as seen in Figure 4 10 For the undope d and 1 mol% Sr doped LaP 5 O 14 samples the highest value of conductivity which was obtained

PAGE 79

79 at 600 o C, was 3.04 x 10 6 S/cm and 1.86 x 10 5 S/cm respectively. For the 5 mol% Sr doped LaP 5 O 14 sample s t he highest conductivity measured was 1.01 x 10 4 S/cm at 600 o C. The activation energies obtained for the undoped, 1 mol% Sr and 5 mol% Sr doped samples were 0.85 0.01 eV, 0.76 0.02 eV and 0.80 0.01 eV respectively which are summarized in Table 4 5 Within 1 mol% doped LaP 5 O 14 samples the conductivity mea surements were made in two different crystal orientations, that is along the ribbon direction and along the transverse direction (along the z axis in Figure 4 3 (a)). A long the transverse direction the obtained conductivity was 6.64 x 10 6 S/cm ( at 600 o C) This value is slightly lower than measured conductivity along the phosphate anion ribbon direction in 1 mol% Sr doped crystals The activation energy calculated for protonic conduction along the transverse direction was only slightly higher than along the phosphate anion ribbon direction (0.76 eV and 0.89 eV) Table 4 5 Measured conductivity at 600 o C and calculated activation energ y for three different compositions of LaP 5 O 14 Material Activation energy (eV) Conductivity (S/cm) Undoped 0.85 0.0 1 3.04 x 10 6 1 mol% Sr 0.76 0.02 / 0.89 0.02 (t) 1.86 x 10 5 / 6.64 x 10 6 (t) 5 mol% Sr 0.80 0.01 1.01 x 10 4 With regards to the mechanism of proton conduction in undoped ultraphosphates there is an alternate model that has been proposed by Hammas et al. According to the model protonic conduction shown by condensed phosphate materials occurs along chemically adsorbed water molecules which lie all along the condensed phosph ate anion in the material structure rather than along the phosphate anion itself. 68 The proton transfer process is described as occurring by cleavage and formation of hydrogen bonds between a hydronium ion and an adjacent water molecule. This could

PAGE 80

80 probably explain the measured conductivity in the undoped material reported at low temperatures, but it does not explain how protonic conduction can be observed in phosp hates at much higher temperatures (~700 o C) where chemically adsorbed water does not exist in the material. Also, since the material measured by Hammas et al. is undoped (and it does not contain any structural protons) it is unclear how a proton would be i ncorporated in the material leading to the formation of a hydronium ion from a water molecule. Figure 4 10. Variation of conductivity with dopant concentration as measured under unhumidified atmosphere It can be envisioned that the defect chemistr y in LaP 5 O 14 is likely to be similar to LaPO 4 69 Analogous to LaPO 4 oxygen deficiencies created from acceptor doping can be expected to lead to addi tional local condensation in the ultraphosphate structure.

PAGE 81

81 The locally condensed phosphate group can then by hydrolyzed in the presence of ambient water vapor to form protonic defects in the material. 4 .3 .2 Effect of A tmospheric H umidity It is known tha t p roton incorporation in those ceramic materials (which are intrinsically devoid of protons) can be accomplished by acceptor doping as represented by the following equation 62 : (4 1) Acceptor doping leads to the formation of charge compensating oxygen ion vacancies which may be fil led up by protonated oxygen ion species ( ) formed by the dissociative absorption of ambient water vapor. These protons (which are associated to an oxygen ion) are free to hop around (Gr tthuss mechanism) between oxygen ions maintaini ng a dynamic equilibrium within the material. Under suitable electro chemical gradient these protons can diffuse along the gradient leading to charge transport within the material. 8 Alternatively, holes can also be generated at the expe nse of oxygen ion vacancies in an atmosphere containing oxygen according to the following equation 70 : ( 4 2 ) In general, holes a re more dominant at higher tem peratures (generally above 800 o C) while protonic defects are the majority carriers at lower temperatures. T he overall charge neutrality equation can be written as follows: ( 4 3 )

PAGE 82

82 Where p represents hole concentration and square brackets indicate the co ncentration of the respective charged species. Figure 4 11. Conductivity measurement under wet/dry argon atmosphere on 5 mol% Sr doped LaP 5 O 14 Samples of 5 mol % Sr doped LaP 5 O 14 showed a higher conductivity when the measurements were carried out i n a humidified argon atmosp here as can be seen in Figure 4 11 An increase in conductivity of a material with increase in humidity is an indication that the charge transport is significantly protonic. An increase in concentration of protonic defects is e xpected with an increase in humidity in accordance

PAGE 83

83 with equation ( 4 1), leading to a higher conductivity and has been observed previously in other phosphates. 62,69,71 4. 4 Atomistic S imulation 4.4 .1 Structural M od eling The experiment ally observed crystal structure of LaP 5 O 14 was reproduced prior to carrying out the defect calculations. The experimental and calculated lattice parameters are compared in Table 4 6 The differences in the observed and calculated latt ice parameters are small indicating the potentials used to reproduce these structures modeled the structures accurately Table 4 6. Experimental and calculated lattice parameters for LaP 5 O 14 Parameter Exp ( ) Calc ( ) a 8.8206 8.8809 0.06 b 9.1 196 9.0823 0.03 c 13.1714 12.9774 0.19 90.66 o 90.20 o 0.46 o As mentioned previously the structural units comprising the phosphate anion in LaP 5 O 14 are PO 4 tetrahedra. Within the se tetrahedra the P O (phosphorus oxygen) bonds can be classified into either O bonds. If the oxygen ion is located at the vertex of two corner sharing PO 4 tetrahedra ( so that it is bonded to two P 5+ ions) then the P P O bond If the oxygen ion is bonded t o one La 3+ and one P 5+ In all the five different tetrahedra in LaP 5 O 14 the terminal P O bonds are considerably shorter than the bridging P O bonds as can be seen in Table 4 7. Such dif f erence in bond lengt hs is commonly o bserved in condensed phosphates. For example in P 2 O 5 each P atom is bonded to three bridging oxygen ( P O P bonds) while the fourth which is a terminal bond has considerable double bond character (P=O) and

PAGE 84

84 hence is much shorter ( ~1.60 and ~1.45 respectively). This difference in bond lengths is correctly reproduced in the calculated structure for LaP 5 O 14 and the results are summarized in Table 4 7 The mean bond length of terminal P O bonds is 1.519 0.01 while that of the brid ging P O bonds is 1.615 0.04 Table 4 7. Experimental and calculated average bond lengths for LaP 5 O 14 Parameter Average bond length ( ) Diff (abs) EXP SIM P O t 1.474 0.01 1.519 0.01 0.02 P O b 1.593 0.03 1.615 0.04 0.02 P1O 4 1.54 7 0.08 1.582 0.07 0.04 P2O 4 1.540 0.05 1.564 0.03 0.02 P3O 4 1.553 0.08 1.591 0.08 0.04 P4O 4 1.540 0.05 1.564 0.04 0.02 P5O 4 1.548 0.09 1.584 0.08 0.04 La O 2.423 0.05 2.503 0.04 0.08 In addition to comparing the lattice par ameters of the experimental and calculated structure, it is also important to compare the bond lengths and bond angles in the materials to gauge the accuracy of the potential parameters used in describing the structure. The average calculated bond length for P O bonds (in the 5 crystallographically distinct PO 4 tetrahedra) and La O bonds in the LaO 8 polyhedron is compared with the average experimental value in Table 4 7. The difference between the values is very small for all the bond le ngths compared. It is interesting to note that t he length of the P O bond is very similar across a variety of different ultraphosphates like EuP 5 O 14 (1.5427 0.06 ), GdP 5 O 14 (1.5450 0.06 ) and also metaphosphates such as Pr(PO 3 ) 3 (1.5418 0.06 ) and C sMn(PO 3 ) 4 (1.5417 0.06 ). 25,72,73 The average calculated intra tetrahedron O P O bond angles are compared with the experimental values in Table 4 8 The difference between the values are small

PAGE 85

85 (< 0.4 o ) indicating t hat the bond angles were modeled accurately. The average values of the bond angles in each PO 4 tetrahedron are close to the expected value for an ideal tetrahedron 109 o and simulated values. This is because the PO 4 tetrahedra in the ultraphosphate are quite distorted. Table 4 8. Experimental and calculated bond angles for LaP 5 O 14 Parameter Average bond angle ( o ) Diff (Abs) EXP SIM O P O P1O 4 109.12 7.39 108.94 10 .14 0.17 P2O 4 109.20 7.01 108.91 09.36 0.29 P3O 4 109.05 7.59 108.73 10.41 0.32 P4O 4 109.12 7.56 108.84 10.15 0.27 P5O 4 109.13 7.22 108.95 09.70 0.17 P O P 133.63 1.79 134.62 03.64 0.98 The average calculated values of the P O P bonds are also compared with the experimental values in as shown in Table 4 8. The average value calculated was 134.62 3.64 o which is close to the experimental value. 4.4 .2 Intrinsic A tomic D efect F ormation The energies of isolated point defects such as vacancies and interstitials were calculated for all the anions and cations in the structure of LaP 5 O 14 To evaluate the energy of a vacancy, an ion was removed from its original lattice position and then placed at an infinite distance. Th e structure around the vacancy was then allowed to relax and the energy of the relaxed structure was then compared with the original structure to calculate the vacancy energy. Similarly the energy of the interstitials was also calculated. The energies of the isolated defects were then combined to evaluate the energies of formation of Frenkel and Schottky defects that are summarized in Table

PAGE 86

86 4 9 To evaluate the partial Schottky defect energies (La Schottky type and P Schottky type) the lattice energies f or La 2 O 3 and P 2 O 5 were obtained by modeling their respective crystal structure using the same interatomic potential that was employed to model the lanthanum phosphates. Table 4 9. Calculated energies for intrinsic material defects (Frenkel and Schottky) in LaP 5 O 14 Type Defect equilibrium E (eV) Full Schottky 114.46 P Frenkel 19.76 La Frenkel 14.69 O Frenkel 7.31 La Schottky type 40.95 P Schottky type 34.23 As seen from the Table 4 9 the energies of formation of Frenkel defects of P, La and O are 19.76 eV, 14.69 eV and 7.31 eV respectively. All the above values are very high indicating that formation of such defects is highly unlikely in the temperature range of operation of the material. In some materials such as Ba 2 In 2 O 5 a low oxygen Frenkel formation energy (0.91 eV) is regarded as the explanation for the high oxygen ion conductivity observed at low t emperatures. 74 76 Due to the ease of formation of oxygen vacancies a significant concentration is expected to be present even at low temperatures which can contribute to oxygen ion transport. However, in the case of LaP 5 O 14 the O Frenkel formation energy is much higher and thus the concentration of intrinsic oxygen ion vacancies can be expected to be almost negligible. The high defect energies for the formation of the P Frenkel and O Frenkel can also be taken as an indication of the structural integrity of the PO 4 tetrahedra in the material.

PAGE 87

87 The formation energy for partial Shottky defects of two types namely, La Schottky type and P Schottky type were calculated to be 40.95 eV and 34.23 eV respectively. The se values are also very large indicating that such defects are very unlikely to be formed in significant concentration in the temperature range of stability of the material. Overall high formation energies were obtained for all the types of intrinsic defe cts considered in these calculations which suggests that the degree of intrinsic disorder in LaP 5 O 14 is almost negligible in the temperature range of thermal stability of the material. Thus the extrinsic defects formed upon acceptor doping are likely to m ainly control the defect chemistry in doped LaP 5 O 14 samples Figure 4 12. Calculated dopant incorporation energy for D 2+ = Mg, Ca, Sr and Ba on the La site. 4.4 .3 Dopant I ncorporation In solid state proton conducting materials it is known that oxygen ion vacancies can lead to the production of protonic defects by the dissociative absorption of ambient

PAGE 88

88 water vapor (Equation 4 1) The protons thus incorporated are mobile and can lead to charge transport via the Gr tthuss mechanism. However, to achieve a significant concentration of oxygen ion vacancies, extrinsic acceptor dopant substitution at the cation site is required which leads to the formation of oxygen ion vacancies as represented by the following equation in Kr ger Vink notation. ( 4 4 ) In the above equation is the negatively charged defect formed due to dopant incorporation at the La site, which leads to the formation of oxygen ion vacancies ( ). By calculating the energie s of these isolated defects and using the equation above t he energetics of divalent dopant (D =Mg 2+ Ca 2+ Sr 2+ Ba 2+ ) substitution on the La 3+ site were calculated. The interatomic potentials for the dopant cations are obtained from the res pective metal oxides. The obtained solution energies for LaP 5 O 14 are shown in Figure 4 1 2 plotted as a function of the dopant ionic radius. For comparison the ionic radius of La 3+ is also shown. Out of the four dopant ions selected Ca 2+ and Sr 2+ seem to be most favo r able while Ba 2+ and Mg 2+ seem to be most unfavo rable The trend obtained is in good agreement with that observed in acceptor doped LaGaO 3 which is also an ionic conductor. 77 In addition, calculations were perf ormed to evaluate the possibility of dopant ion incorporation on the P site instead of the La site. However, the obtained energies for such an arrangement were too high indicating that dopants are likely to be incorporated almost exclusively on the La sit e. Thus according to the trend shown below Ca 2+ can be expected to be a suitable ion for extrinsic dopant substitution in LaP 5 O 14 However, to the best of our knowledge no experimental study has been

PAGE 89

89 reported on the investigation of electrical properties in calcium doped LaP 5 O 14 to compare the above calculated results 4.4 .4 Proton I ncorporation and O H C onfiguration It is known that protons incorporated in ceramic materials by acceptor doping are closely associated to an oxygen ion. Infrared spectrosc opy and NMR spectroscopy have been used previously to probe the oxygen proton interaction in proton conducting materials 69,78 89 However, these techniques cannot provide detailed information on the location of su ch protons in the structure. Neutron diffraction can be used to obtain information regarding location of the proton in the material and has been used to study some materials which intrinsically contain protons such as CsH 2 (P O 4 ) 86,90 However the concentration of protons in acceptor doped ceramic materials is very low making such studies difficult. Atomistic simulations can be used to probe the local environment around a proton to determine the most energeticall y favorable configuration for the proton within the structure. In this study a proton was placed at a distance of ~1 from an oxygen ion in the structure and the entire structure was allowed to relax so that it could attain the minimum energy configurati on. Such calculations were performed for all the 14 crystallographically different oxygen ions in the structure. In the relaxed structure an equilibrium O H distance of ~ 0.98 was obtained for all the calculations. However, common to all the calcul ations, it was noticed that in the relaxed structure the O H bond was directed in such a way that the equilibrium position of the proton was located in either of the two hollow channels located along the a axis or the c axis. The equilibrium positio ns for the protons bonded to oxygen ions O1 and O14 are shown in Figure 4 13 (a) and (b) respectively Along both crystallographic directions

PAGE 90

90 it can be seen that the proton is located within the hollow channels that exist in the crystal structure. This i s most likely due to the electrostatic repulsion exerted on the proton by the La 3+ and P 5+ ions in the structure. (a) (b) Figure 4 13. View of LaP 5 O 14 along (a) c axis and (b) a axis showing t he location of hollow channels in the structure. (Proton shown in blue). A protonic defect ( ) is essentially a positively charged defect which is generated to maintain charge neutrality in the material upon acceptor doping on the ca tion site which leads to the formation of negatively charged defects ( ). If there is association between the two oppositely charged defects it can adversely affect the proton conductivity in the material and hence it is of interest t o calculate the energetics of such an interaction. The cluster binding energy for such an interaction can be calculated by considering the energy of the cluster (E cluster ) and comparing it with the sum of the energies of isolated individual defects as sho wn in the general equation below: ( 4 5 )

PAGE 91

91 A negative E bind in such a case would indicate that such an association is energetically favorable. In the case of the lanthanum ultraphosphate compound the relevant equation would be a s follows: ( 4 6 ) The calculated binding energies obtaine d by varying the dopant cations are listed in Table 4 10 It can be seen that for all the dopants the values are negative which implies that such an association is favorab le. The lowest energy calculated was for the strontium dopant ( 0.39 eV) while calcium dopant exhibited the highest binding energy ( 0.50 eV). Table 4 10. Proton dopant interaction in LaP 5 O 14 D 3+ E binding (eV) D H ( ) D H ( ) ( o ) Mg 0.4 4 2.50 0.387 15.7 00 Ca 0.50 2.69 0.199 15.099 Sr 0.39 2.82 0.068 14.283 Ba 0.4 5 3.30 0.411 14.014 If the charge carrying species ( protonic defect in this case ) dopant site, th en it would have to ov ercome a higher e nergy barrier which would be due to a contribution from the binding energy term to the activation energy for conduction. For these calculations the dopant ion and protonic defect were kept at nearest neighboring positions for which the de gree of interaction would be the highest. However, it can be envisioned that as the distance between the two are increased the interaction energy would gradually decrease. Indeed it has been found that the activation energy for protonic conduction increa ses slightly with the increase of dopant concentration in the material. The high value of binding energy for the calcium do pant

PAGE 92

92 cation suggests that a greater change in the activation energy of the dopant with increase of dopant concentration in the mater ial can be expected 4.5 Other Materials Based on the energetics of dopant incorporation in LaP 5 O 14 which were calculated from atomistic simulations the conductivity of Ca doped LaP 5 O 14 was measured. In addition the conductivity of NdP 5 O 14 samples usi ng Ca and Sr as dopants (independently) was also measured to investigate the effect of change of the host cation on the electrical properties of the ultraphosphate materials. (a) (b) Figure 4 14. Conductivity of (a) 5 mol% Ca doped LaP 5 O 14 (b) 5 m ol% Ca doped and 5 mol% Sr doped NdP 5 O 14 (unhumidified air). The measured conductivity of 5 mol% Ca doped LaP 5 O 14 samples is shown in Figure 4 14 (a) where conductivities of other compositions of LaP 5 O 14 are also shown for reference. The conductivity o f 5 mol% Ca doped (4.50 x 10 5 S/cm 600 o C ) samples was slightly lower than the conductivity of 5 mol% Sr doped LaP 5 O 14 This is in good agreement with the dopant incorporation energies calculated from atomistic simulations according to which the solution energy for Ca 2+ was found to be slightly higher than

PAGE 93

93 Sr 2+ A higher dopant incorporation energy implies that a low er concentration of substitutional defects ( ) may have been formed in the bulk, leading to a lower concentration of pr otonic defects and hence the lower conductivity in the material. The measured conductivity of 5 mol% Ca doped NdP 5 O 14 and 5 mol% Sr doped NdP 5 O 14 is shown in Figure 4 14 (b). The Sr doped NdP 5 O 14 samples showed higher conductivity as compared to the Ca doped samples This trend is similar to that observed in doped LaP 5 O 14 samples However, t he conductivity of 5 mol% Sr doped NdP 5 O 14 (5.55 x 10 5 S/cm, 600 o C) was slightly lower than similarly doped LaP 5 O 14 samples. A probable explanation for this could be smaller ionic radi us of the Nd 3+ cation (1.11 ) in comparison to the La 3+ cation (1.16 ) which causes a larger lattice strain in the material upon incorporation of the dopant cation Sr 2+ which has an ionic radius of 1.26 4.6 Summary and Conclusi on S ynthesis of lanthanum ultraphosphate (LaP 5 O 14 ) single crystals of three different compositions (undoped, 1 mol% and 5 mol% Sr 2+ doped) was carried out by precipitation from concentrated phosphoric acid solutions. Phase purity of the prepared crystals was established by comparing the experimental powder diffraction pattern with the theoretical pattern and also from quantitative EDS analysis. Conductivity measurements were performed on crystal samples of all the above prepared compositions along the [10 0] direction in the temperature range 300 o C 600 o C using electrochemical impedance spectroscopy. Within the 1 mol% Sr doped LaP 5 O 14 samples conductivity was also measured along the transverse direction.

PAGE 94

94 The conductivity was found to increase with incre ase of dopant concentration upto 5 mol% Sr doping. The highest measured conductivity was f or the 5 mol% Sr doped LaP 5 O 14 sample s which was 1.01 x 10 4 S/cm at 600 o C. The activation energies obtained for the undoped, 1 mol% Sr and 5 mol% Sr doped samples were 0.85 0.01 eV, 0.76 0.02 eV and 0.80 0.01 eV respectively. The activation energy calculated for protonic conduction along the transverse direction was only slightly higher than along the phosphate anion ribbon direction (0.76 eV and 0.89 eV) ind icating that the structural anisotropy did not have any significant impact on the electrical properties of the material. Humid atmosphere measurements on the doped samples indicate that the charge transport is proton dominated in these materials. The ex perimental ly obtained crystal structure of LaP 5 O 14 was accurately reproduced using atomistic simulations and previously reported interatomic potential parameters. Defect calculations on LaP 5 O 14 indicate that intrinsic defects of all types are energeticall y unfavorable and hence the defect chemistry in the material is likely to be entirely controlled by extrinsic dopant substitution. Dopant dissolution studies indicated that strontium and calcium were the most suitable dopants for conductivity enhancement in LaP 5 O 14 while barium and magnesium were unfavorable. Proton dopant interaction studies indicated that strontium doping in LaP 5 O 14 was most suitable for conductivity enhancement because of lowest interaction energy

PAGE 95

95 CHAPTER 5 C HARACTERIZATION OF L ANTHANUM ORTHOPHOSPH ATE (L a PO 4 ) 5.1 Orthophosphates Lanthanum orthophosphate (LaPO 4 ) occurs at 50 mol% P 2 O 5 in the La 2 O 3 P 2 O 5 binary phase diagram The structure of anion in LaPO 4 is least complex among all the phosphate compounds that occur in the phos phate rich region of the phase diagram, the most structurally intricate being LaP 5 O 14 I n the previous chapter the conductivity of three different compositions of LaP 5 O 14 samples (undoped, 1 mol% and 5 mol% Sr doped) was measured. Since the main objectiv e of this study was to evaluate the effect of structural variation across the range of lanthanum phosphate compounds on the electrical properties of the same, the conductivity of LaPO 4 samples with exactly the same dopant composition was also measured. T his was done to eliminate the possibility of experimental variation among different research groups so that the observed differences in properties could be attributed to differences in material conduction behavior. The current chapter discusses the result s from characterization of lanthanum orthophosphate (LaPO 4 ). The measured properties of the two lanthanum phosphate compounds are contrasted and critically analyzed on the basis of crystallochemical concepts in the following chapter. 5.2 Crystallochemic al A nalysis The crystal structure of LaPO 4 as seen along the z axis is shown in Figure 5 1. The structure is isomorphic with other rare earth orthophosphate s (RPO 4 where R = Pr, Nd, Eu, Gd and Sm) and belongs to a monoclinic crystal system with space gro up: P2 1 /n (No. 14), with lattice parameters a = 103.27 o ). 91,92 The structure of LaPO 4 consists of isolated PO 4 tetrahedra (shown in

PAGE 96

96 green) which are common to all cond ensed phosphates. However, unlike LaP 5 O 14 the tetrahedra are not corner shar ed and thus do not form a phosphate anion network The La 3+ ions (shown in orange) are located in a nine coordinated oxygen ion environment. Figure 5 1. Crystal structure of LaPO 4 as seen along the [001] direction. The PO 4 tetrahedra (green), La 3+ ions (orange) and O 2 (red) are depicted Here, a crystallochemical analysis of LaPO 4 is presented similar to the analysis for LaP 5 O 14 described in the previous chapter It is kn own that proton transport in solid state proton conducting materials takes place via the Gr tthuss mechanism, and hence the (O O) distance (or proton jump distance) is a key parameter which governs the magnitude of the activation energy for the proton tran sfer process. A detailed description of the above can be found in section 6 1. The pair distribution functions for LaPO 4 centered on each of the four crystallographically different oxygen ions are shown in Figure 5 2. It can be seen that around each of the four oxygen ions there are three other oxygen ions in the range 2.44 2.58 These are the oxygen ions that belong to the s ame PO 4 tetrahedron as the centered oxygen ion. The O O distances within the same tetrahedron are almost the

PAGE 97

97 same as the tet rahedra in LaP 5 O 14 (~ 2.38 2.58 ). At larger distances ( > 2.85 ) there are other ions which belong to the neighboring tetrahedra. (a) (b) (c) (d) Figure 5 2. Pair distribution function for LaP 5 O 4 centered at (a) O1, (b) O2, (c) O3, and ( d) O4. As the activation energy for the proton transfer process is directly related to the (O O) distance, the jump coordination for all the oxygen ions is denoted as three in Figure 5 2 Details regarding the relation between the activation energy and (O O) distance are presented in section 6.2. In LaP 5 O 14 the oxygen ions were classified as either bridging or terminal oxygen ions. The former had a jump coordination of six

PAGE 98

98 whereas the latter had a jump coordination of three. It can be seen that all the oxygen ions in LaPO 4 are terminal oxygen ions since the tetrahedra are all isolated from each other and hence the lower jump coordination number Figure 5 3. Comparison of theoretical and experimental XRD powder pattern of the uncalcined, calcin ed and sintered LaPO 4 samples. I t can be seen that within the same tetrahedron the O O distances are always less than 2.6 However, for long range proton transport to occur in LaPO 4 a jump must be accomplished by the proton between oxygen ions belong ing to two adjacent tetrahedra. The jump distance for such an event is much higher (> 2.8 ). On the basis of the above, LaPO 4 may be expected to show an overall higher activation energy (and correspondingly lower conductivity) in comparison to LaP 5 O 14

PAGE 99

99 5. 3 Phase A nalysis 5.3 .1 X R ay D iffraction (XRD) T he powder XRD pattern obtained for as precipitated powder, calcined powder and sintered pellet of undoped LaPO 4 is compared with the theoretical powder pattern in Figure 5 3 As can be seen from the figur e the as precipitated powder shows broad peaks which indicates that the powder is considerably amorphous and not homogeneous By comparison the XRD profile of the calcined powder shows much narrower peaks indicating that crystalline powder was obtained a fter the calcination process. The experimental powder XRD pattern of the sintered pellet of undoped LaPO 4 was found to be in good agreement with the theoretical powder pattern indicating that the synthesized material was phase pure within the instrument r esolution (a) (b) Figure 5 4. (a) Experimental XRD patterns and (b) calculated lattice parameters for pellets of the three different LaPO 4 (W standard) compositions The experimental powder XRD pattern obtained for sintered samples of undoped, 1 mol% Sr doped and 5 mol% Sr doped LaPO 4 is compared with the theoretical pattern

PAGE 100

100 in Figure 5 4. Tungsten was used as an internal standard for the above measurement (labeled in the graph). All the obtained peaks in the XRD patterns match well with the t heoretical pattern indicating that the synthesized materials were phase pure within the instrument resolution Further, the lattice parameters were calculated by fitting the experimental data using the MAUD software using previously know lattice parameter s and atomic positions for LaPO 4 as the starting point The variation of unit cell volume and the lattice parameters (a, b and c) is plotted as a function of dopant concentration in Figure 5 4 (b). It can be seen that lattice expansion occurs on dopant i ncorporation as expected since the ionic radius of strontium is larger than the lanthanum cation. Moreover the lattice expansion is uniform along all the three axes. The trend shown by the lattice parameters indi cate that lattice saturation limit of dopa nt incorporation is achieved around slightly above 1 mol% dopant concentration. The obtained XRD patterns of both the doped materials were compared with other compounds containing Sr, La, P and O namely, Sr 2 P 2 O 7 and Sr 3 La(PO 4 ) 3 to see if the addition of t he dopant had lead to the formation of any additional phases. However, no secondary phase formation could be detected from the obtained XRD pattern. 5.3 .2 Scanning Electron Microscopy ( SEM) A typical microstructure of the fractured surface of the sinter ed pellet is shown in Figure 5 5 Some closed porosity can be observed from the SEM image. The average grain size was estimated by observing the microstructure inside one of the closed pores similar to the one indicated by the arrow. The average grain s ize was about 1.5 m. The EDS (energy dispersive spectroscopy) spectra for undoped LaPO 4 samples is shown in an in set in Figure 5 5. In the spectrum peaks belonging to La, P, O and C (since the samples were carbon coated prior to EDS analysis ) were seen as expected.

PAGE 101

101 Figure 5 5. SEM image and EDS spectrum (inset) of the fractured surface of a typical sintered sample of undoped LaPO 4 5. 4 Conductivity M easurement The N yquist plots obtained for undoped and 1 mol% Sr doped LaPO 4 obtained under unhumi dified atmosphere are shown in Figure 5 6 In general ceramic materials exhibit a three arc response, where each arc corresponds to bulk, grain boundary and electrode impedance in decreasing order of frequency. 23 Further background details regarding the Nyquist plots of ceramic materials can be found in se ction 2.5.1. In t he plots for undoped LaPO 4 obtained at 600 o C, 520 o C, 440 o C, and 360 o C ( shown in Figure 5 6 (a) and (b) ) only one distinct arc can be seen with a slight bulge in the low frequency region (right side of the arc). The distinct arc is due to the bulk impedance while the bulge is most likely to be due to the grain boundary impedance which is not well

PAGE 102

102 res olved from the bulk impedance. The relative magnitude of the grain bounda r y resistance is much smaller compared to the bulk resistance due wh ich it is mostly obscured by the bulk impedance arc. The grain boundary arc is more prominently observed at lower temperatures. (a) (b) (c) (d) Figure 5 6. Nyquist plots obtained under unhumidified air for undoped LaPO 4 samples (a), (b) and for 1 mol% Sr doped LaPO 4 (c), (d). On the contrary two distinct and reasonably well resolved arcs can be seen in the Nyquist plots for 1 mol% doped LaPO 4 at all temperatures as shown in Figure 5 6 (c)

PAGE 103

103 and (d) Similarly the arcs were also well resolv ed for the 5 mol% Sr doped samples. The arc in the high frequency region is due to bulk impedance while the arc seen at lower frequencies is due to the grain boundary impedance in the sample. By comparing the relative sizes of the arcs in Figure 5 6 (c) and (d) it can be observed that the bulk impedance is more prominent at higher temperatures. However, at lower temperatures (360 o C) the grain boundary impedance is higher than the bulk impedance. A partial electrode impedance arc can be seen in Figure 5 6 (c) at low frequency. A comparison of the normalized Nyquist plots of the 1 mol% and 5 mol% doped samples shows that the magnitude of the grain boundary impedance of the 5 mol% doped sample is much higher as compared to the 1 mol% doped sample. This m ay be due to the grain boundary segregation of dopant ions since the lattice parameter calculations show that saturation of dopant ion incorporation is achieved at a dopant concentration slightly above 1 mol%. Table 5 1. Normalized capacitances associat ed with bulk and grain boundary impedance for all compositions of LaPO 4 Dopant C bulk (pF) C grain boundary (pF) d / D ratio G.B. thickness Undoped 28 1.98 N.A. N.A. 1 mol% 7.76 0.69 977.88 237.35 0.01 15 nm 5 mol% 25.89 1.33 2400.79 283.61 0.01 15 nm The capacitance associated with any impedance arc can be calculated by knowing the peak frequency for the arc which is shown in Figure 5 6 All the calculated values o f the capacitances are summarized in Table 5 1. The capacitance associated with the bulk impedance arc for both the 1 mol% and 5 mol% Sr doped LaPO 4 samples (7.76 0.69 and 25.89 1.33 pF) is much smaller as compared to the capacitance of the grain boun dary impedance arc (977.88 237.35 and 2400.79 283.61 pF) The values

PAGE 104

104 obtained are similar to the capacitance for bulk and grain boundary impedance for other ionic conductors thus confirming that the arcs observed in the Nyquist plots were correctly as cribed to the respective polarization phenomena. 23 D ue to the poor resolution of the arcs corresponding to the grain boundary and bulk impedance the capacitance associated with grain boundary impedance could not be calculated for the undoped samples. Using the brick layer model to describe the electrical properties of a polycrystalline material, the relative thickness of the grain boundary can be estimated accord ing to the equation shown below 22 : ( 5 1) Where d and D are the thickness and diameter of the grain boundary and grain respectively, C bulk and C gb are the capacit ances associated with the grain and grain boundary, and bulk and gb are the dielectric constants of the bulk and grain boundary phases. The average value of the grain diameter has been estimated from the SEM images of the fractured cross section of the samples. The calculated values of the d / D ratio and the est imated grain boundary thickness are shown in the Table 5 1 5.4 .1 Effect of D opant C ontent The temperature variation of conductivity for all the samples of doped and undoped LaPO 4 is shown in Figure 5 7 All the measurements were made in unhumidified at mosphere. Undoped LaPO 4 exhibited the lowest conductivity in the measured temperature range (5.15 x 10 7 S/cm 600 o C). Its conductivity was about an order of magnitude lower than 1 mol% Sr doped LaPO 4 over the entire temperature range Overall 5 mol%

PAGE 105

105 S r doped LaPO 4 showed the highest conductivity (7.00 x 10 6 S/cm, 600 o C ) The conductivity enhancement from 1 mol% to 5 mol% dopant concentration was much lower than that from undoped to 1 mol% dopant concentration. This is in good agreement with the latt ice parameter calculations which indicate that the dopant saturation is achieved in the lattice slightly above 1 mol%. Figure 5 7. Measured conductivity for undoped and doped LaPO 4 (unhumidified air). The activation energy for conduction was calcul ated from the plot by fitting the data using the Arrhenius equation as shown below: (5 2)

PAGE 106

106 Where, is the conductivity, T is the absolute temperature, A is the pre exponential, k and E a is the act ivation energy for conduction. The activation energy for conduction for undoped LaPO 4 (1.27 0.03 eV) was much higher than that for 1 mol% Sr doped LaPO 4 (1.04 0.01 eV) and 5 mol% Sr doped LaPO 4 (1.09 0.01 eV) as summarized in Table 5 2 Table 5 2 Measured conductivity at 600 o C and calculated activation energ y for three different compositions of LaPO 4 (unhumidified atmosphere) Material Activation energy (eV) Conductivity (S/cm) Undoped 1.27 0.03 5.15 x 10 7 1 mol% Sr 1.04 0.01 1.77 x 10 6 (560 o C) 5 mol% Sr 1.09 0.01 7.00 x 10 6 The overall conductivity for all the synthesized compositions of LaPO 4 was lower than the corresponding LaP 5 O 14 samples at all temperatures. In addition, the activation energy for conduction obtained for all th e LaP 5 O 14 samples were lower than those obtained for the LaPO 4 samples. These results will be analyzed and discussed in detail in the following chapter. Table 5 3. Measured conductivity at 600 o C and calculated activation energ y for three different comp ositions of LaPO 4 under humidified atmosphere Material Activation energy (eV) Conductivity (S/cm) Undoped 1.28 0.01 6.33 x 10 7 1 mol% Sr 1.07 0.02 9.98 x 10 6 5 mol% Sr 1.05 0.02 3.24 x 10 5 5.4 .2 Humidified A tmosphere C onductivity M easurement Con ductivity measurements were made on the three prepared compositions of LaPO 4 samples under humidified atmosphere. The obtained results are shown in Figure 5 8 along with the measured conductivity under unhumidified atmosphere for comparison. The mea sured conductivities and calculated activation energies have

PAGE 107

107 been summarized in Table 5 3. Both doped samples showed an increase in conductivity under h umid ified atmosphere conditions. Such a trend has been observed previously in several other proton con ducting materials and is explained as follows. 62,70,91 It is known that proton incorporation in those ceramic materials (which are intrinsically devoid of protons) can be accomplished by acceptor doping as repres ented by the following equation 62 : (5 3) Acceptor doping leads to the formation of charge compensating oxygen ion vacancies which may be filled up by protonated oxygen ion species ( ) formed by the dissociative absorption of ambient water vapor. The above equa tion is analogous to the equation shown below which is more applicable to the phosphate based proton conducting materials 69 : (5 4) In orthophosphates acceptor doping leads to the formation of charge compensating oxygen ion vacancies, which is equivalent to formation of a pyrophosphate group. The pyrophosphate is formed by the local conden sation of two adjacent PO 4 tetrahedra, one of which is missing an oxygen ion. In the presence of am bient water vapor the hydrolysis of the pyrophosphate leads to the formation of protonic defects as re presented by the equation 5 4. It can be seen from bo th the equations that an increase in humidity would lead to an increase in concentration of protonic defects which contribute to the charge transport leading to an increase in conductivity under humidified conditions.

PAGE 108

108 Further it was observed that the enh ancement in conductivity in humidified atmosphere was higher for LaPO 4 samples with higher dopant content. It can be seen from Figure 5 8 that the conductivity of undoped LaPO 4 is almost independent of the humidity while in 5 mol% Sr doped LaPO 4 almost an order of magnitude increase in conductivity is observed under humidified atmosphere. Figure 5 8. Measured conductivity for undoped and doped LaPO 4 as measured in humidified atmosphere (wet air). This is in good agreement with the expected trend and is explained as follows. In the doped material with a higher dopant content, a higher concentration of oxygen vacancies are formed in accordance with the charge neutrality equation shown below: (5 5)

PAGE 109

109 Where p represents hole c oncentration and square brackets indicate the concentration of the respective charged species. Thus under humid atmosphere, the concentration of charge carrying protonic defects formed in the sample with higher dopant content is higher leading to a higher enhancement in conductivity. For the same reason almost no difference in conductivity is observed in the undoped sample under the two different atmospheres. Table 5 4. Comparison of experimental and calculated lattice parameters for LaPO 4 Parameter Exp ( / o ) Calc ( / o ) a 6.8313 6.8302 0.02 b 7.0705 6.9922 0.08 c 6.5034 6.6258 0.12 103.27 104.42 1.15 o 5.5 Atomistic S imulation 5.5.1 Structural M odel ing Atomistic simulations were performed using the GULP (Generalized Utility Lattice Program) technique to s tudy the dopant incorporation, intrinsic defects, defect chemistry, and proton dopant interactions in LaPO 4 Simulation studies are a useful tool for probing the local atomic environment within the material and also defect structures which are crucial for understanding the electrical properties of a material. Thus they can prove useful in explaining some of the experimentally obtained results as well as give directions for further experimentation in certain cases. The experimentally observed crystal stru cture of LaPO 4 was reproduced prior to carrying out the defect calculations. The experimental and calculated lattice para meters are compared in Table 5 4 The differences in the observed and calculated lattice parameters are small indicating the potenti als used to reproduce these structures modeled the structures

PAGE 110

110 accurately. In addition to comparing the lattice parameters of the experimental and calculated structure, it is also important to compare the bond lengths and bond angles in the materials to ga uge the accuracy of the potential parameters used in describing the structure. The average calculated bond length for P O bonds (in the 5 crystallographically distinct PO 4 tetrahedra) and La O bonds in the LaO 9 polyhedron is compared with the average experimental value in Table 5 5 The difference between the values is very small for all the bond lengths compared. It is interesting to note the structural similarity of PO 4 tetrahedra across the range of condensed phosphate compounds in the La 2 O 3 P 2 O 5 binary phase diagram For example the average P O bond length in LaPO 4 ( 1.564 0.03 ) is quite similar to the same in LaP 5 O 14 ( 1.577 0.01 ) Table 5 5. Comparison of the experimental bond distances and bond angles in LaPO 4 with the calculated values. P O EXP ( ) SIM ( ) Diff (Abs) O P O ( o ) ( o ) Diff (Abs) P O1 1.523 1.545 0.02 O1 P O2 104.59 103.22 1.37 P O2 1.537 1.547 0.01 O1 P O3 105.53 105.45 0.08 P O3 1.541 1.549 0.01 O1 P O4 108.31 107.49 0.82 P O4 1.553 1.615 0.06 O2 P O3 112.28 110.99 1.29 P O (mean) 1.539 0.01 1.564 0.03 0.03 O2 P O4 112.99 114.25 1.26 O3 P O4 113.23 115.32 2.09 La O (mean) 2.579 0.10 2.621 0.11 0.03 109.49 3.9 109.45 4.9 0.04 As mentioned previously that within the PO 4 tetrahedra in La P 5 O 14 the P O bonds O bonds depending upon the nature of bonding of the oxygen ion. However, in LaPO 4 there is no corner sharing of oxygen ions between two neighboring tetrahedra and hence all t he P O bonds are can be considered terminal bonds.

PAGE 111

111 5.5 .2 Intrinsic A tomic D efect F ormation The energ y of isolated point defects such as vacancies and interstitials was calculated for all the anions and cations in the structure of LaPO 4 To evaluate th e energy of a vacancy, an ion was removed from its original lattice position and then placed at an infinite distance. The structure around the vacancy was then allowed to relax and the energy of the relaxed structure was then compared with the original st ructure to calculate the vacancy energy. Similarly the energy of the interstitials wa s also calculated. The energy of the isolated defects was then c ombined to evaluate the energy of formation of Frenkel and Schottky defects that are summarized in Table 5 6 Table 5 6. Calculated energies of Frenkel and Schottky type intrinsic defects in LaPO 4 Type Defect equilibrium E (eV) Full Schottky 46.09 P Frenkel 41.33 La Frenkel 17.0 6 O Frenkel 12.74 La Schottky type 29.94 P Schottky type 74.51 To evaluate the partial Schottky defect energies (La Schottky type and P Sch ottky type) the lattice energy for La 2 O 3 and P 2 O 5 was obtained by modeling their respective crystal structure using the same interatomic potential that was employed to model the lanthanum phosphates. As seen from the Table 5 6 the energy of formation of Fr enkel defects of P, La and O is 41.33 eV, 1 7.06 eV and 12.74 eV respectively. All the above values are very high indicating that formation of such defects is highly unlikely in the temperature range of operation of the material. In some materials such as Ba 2 In 2 O 5 low oxygen Frenkel formatio n energy (0.91 eV) is regarded as the

PAGE 112

112 explanation for the high oxygen ion conductivity observed at low temperatures. Due to the ease of formation of oxygen vacancies a significant concentration is expected to be present even at low temperatures which can contribute to oxygen ion transport. 74 However, in the case of LaPO 4 the O Frenkel formation energy is much higher and thus the concentration of intrinsic oxygen ion vacancies can be expected to be almost negli gible. The high defect energy for the formation of the P Frenkel and O Frenkel can also be taken as an i ndication of the structural integrity of the PO 4 tetrahedra in the material The formation energy for partial S c hottky defects of two types namely, La Schottky type and P Schottky type was calculated to be 29.94 eV and 74.51 eV respectively. These v alues are also very large indicating that such defects are very unlikely to be formed in significant concentration in the temperature range of stability of the material. Overall high formation energies were obtained for all the types of intrinsic defects considered in these calculations which suggests that the degree of intrinsic disorder in LaPO 4 is almost negligible in the temperature range of thermal stability of the material. Thus the defect chemistry in LaPO 4 is likely to be entirely controlled by ex trinsic dopant substitution in the material The results from atomistic simulation studies on LaP 5 O 14 are presented and discussed in section 4.4. 5.5 .3 Dopant I ncorporation In solid state proton conducting materials it is known that oxygen ion vacancies c an lead to the production of protonic defects by the dissociative absorption of ambient water vapor (Equation 5 3 and 5 4 ). The protons thus incorporated are mobile and can lead to charge transport via the Gr tthuss mechanism. However, to achieve a signi ficant concentration of oxygen ion vacancies, extrinsic acceptor dopant substitution

PAGE 113

113 at the cation site is required which leads to the formation of oxygen ion vacancies as represented by the following equation in Kr ger Vink notation. (5 6 ) In the above equation is the negatively charged defect formed due to dopant incorporation at the La site, which leads to the formation of oxygen ion vacancies ( ). By calculating the energies of th ese isolated defects and using the equation above the energetics of divalent dopant (D =Mg 2+ Ca 2+ Sr 2+ and Ba 2+ ) substitution on the La 3+ site were calculated. Figure 5 9. Calculated dopant incorporation energy for D 2+ = Mg, Ca, Sr and Ba on the La site. The interatomic potentials for the dopant cations are obtained from the respective metal oxides. The obtained solution energies for LaPO 4 are shown in Figure 5 9 plotted as a function of the dopant ionic radius. For comparison the ionic radius of La 3+ is also

PAGE 114

114 shown. Out of the four dopant ions selected Ca 2+ and Sr 2+ seem to be most favorable while Ba 2+ and Mg 2+ seem to be unfavorable. The trend obtained is in good agreement with that observed in acceptor doped LaGaO 3 which is also an ionic condu ctor. 77 In addition, calculations were performed to evaluate the possibility of dopant ion incorporation on the P site instead of the La site. However, the obtained energies for such an arrangement were too hi gh indicating that dopants are likely to be incorporated almost exclusively on the La site. Thus according to the trend shown below Ca 2+ can be expected to be a suitable ion for extrinsic dopant substitution in LaPO 4 5.5 .4 Proton I ncorporation and D efe ct C hemistry The equilibrium between water, pyrophosphate ion (which is formed in response to acceptor doping) and the hydrogen phosphate ion (formed by dissociative adsorption of ambient water vapor) is represented by equation 5 4 which was proposed by Am ezawa et al. on the basis of NMR investigations 69 The formation of the pyrophosphate ion in phosphates is analogo us to an oxygen vacancy in proton conducting perovskite type oxides. It is formed by corner sharing an oxygen ion between two adjacent PO 4 tetrahedra one of which is missing an oxygen ion. Using atomistic simulations the local structure around the releva nt defect structures was investigated and the results are depicted in Figure 5 10 The relevant defect (such as an oxygen vacancy or a protonic defect) was introduced in the structure and the ions around the defect were allowed to relax to a minimum energ y configuration The normal distance between phosphorus ions from two neighboring [PO 4 ] 3 groups is 4.187 as shown in Figure 5 10 (c) and the average P O distance within each tetrahedron is 1.564 0.03 However, on the formation of an oxygen vacanc y

PAGE 115

115 due to acceptor doping the two adjacent tetrahedra are pulled closer together and share an oxygen ion to form the pyrophosphate anion as shown in Figure 5 10 (b). This result is in good agreement with experimental observation of the pyrophosphate group in doped LaPO 4 69 The inter phosphorus ion distance in this arrangement is reduced to 3.175 while the P O b (bri dging oxygen ion) bond distance increases to 1.719 (a) (b) (c) Figure 5 10. Resulting local relaxed structures of (a) hydrogen phosphate, (b) pyrophosphate, and (c) orthophosphate.

PAGE 116

116 This is in accordance with the observation for higher condensed phosphates such as LaP 5 O 14 that the bridging P O bonds are always longer than the terminal P O bonds. The relevant interatomic distances for the defect structures shown in Figure 5 10 are summariz ed in Table 5 7. Table 5 7. Interatomic distances in the pyrophosphate ion, orthophosphate ion and hydrogen phosphate ion in LaPO 4 Ion Ion d () Ion Ion d () Ion Ion d () P 1 P2 3.175 P P adj 4.050 P P adj 4.038 P1 Ob 1.719 P O1 1.574 P O proton 1.882 P2 Ob 1.719 P O2 1.576 P O2 1.558 P1 O2 1.550 P O3 1.583 P O3 1.542 P1 O3 1.564 P O4 1.604 P O4 1.541 P1 O4 1.552 P2 O2 1.547 P2 O3 1.553 P2 O4 1.572 To evaluate the local structure around the hydrogen phosphate [HPO 4 ] 2 ion which is formed by hydrolysis of a pyrophosphate ion in the presence of ambient water vapor a proton was placed close to an oxygen ion and the resulting struct ure was allowed to relax to attain the minimum energy configuration. It can be seen in Figure 5 10 (a) that the proton prefers to lie outside the PO 4 tetrahedron. In addition, its inclusion also leads to a higher P O distance (1.882 ) for the oxygen i on to which the proton is associated, most likely due to electrostatic repulsion from the P 5+ ion. A protonic defect ( ) is essentially a positively charged defect which is generated to maintain charge neutrality in the material upon acceptor doping on the cation site which leads to the formation of negatively charged defects ( ). If there is association between the two oppositely charged defects it can adversely affect the

PAGE 117

117 proton conductivity in the material and hence it is of interest to calculate the energetics of such an interaction. The cluster binding energy for such an interaction can be calculated by considering the energy of the cluster (E cluster ) and comparing it with the sum of the energies of isolated individual defects as shown in the general equation below: (5 7 ) A negative E bind in such a case would indicate that such an association is energetically favorable. In the case of the lanthanum ortho phosphate compound the rel evant equation would be as follows: (5 8 ) The calculated binding energies obtained by varying the dopant cations are listed in Table 5 8 It can be seen that for all the dopants the values are negative which implies that such a n association is favorable. The highest binding energy calculated was for the magnesium dopant ( 0. 75 eV) while for all the other dopants the values were quite similar (~ 0.55 eV) Table 5 8. Binding energies and other relevant parameters of proton dop ant interaction in LaPO 4 D 3+ E binding (eV) D H ( ) D H ( ) ( o ) Mg 0.75 2.94 0.054 19.051 Ca 0.59 3.04 0.152 0.98 4 Sr 0.57 3.1 9 0.296 3.69 5 Ba 0.56 3.3 7 0.478 8.72 3 The dopant ion proton distances (D H) listed in Table 5 8 show that the same increases with increase of the ionic radius of the dopant cation (R Ba > R Sr > R Ca > R Mg ).

PAGE 118

118 The angular deviation ( ) in the location of a proton upon substitution of a neighboring La site with a dopant cation also increases with the increase of the dopant ionic radius. 5.6 Summary and Conclusion Three different composition s of LaPO 4 were synthesized namely undoped, 1 mol% and 5 mol% Sr doped LaPO 4 Phase purity of the prepared compositions was confirmed by x ray diffraction (XRD) and energy disper sive spectroscopy (EDS). Scanning electron microscopy (SEM) was used to observe the microstructure of the sintered pellets and to confirm the densification of the samples. Conductivity measurements were performed on all the prepared samples in the temper ature range 300 o C 600 o C under humidified and unhumidified atmosphere separately. The conductivity was found to increase with increasing dopant concentratio n. The highest conductivity obtained for 5 mol% Sr doped LaPO 4 samples was 3.24 x 10 5 S/cm (humi dified atmosphere, 600 o C) and 7.00 x 10 6 S/cm (unhumidified atmosphere, 600 o C). For both the doped compositions the conductivity was found to increase under humidified atmosphere which is in good agreement with the known model for proton conduction in ac ceptor doped ceramic proton conducting materials. The activation energy for proton conduction in the undoped, 1 mol% and 5 mol% Sr doped samples was 1.27 eV, 1.04 eV and 1.09 eV. The experimentally obtained crystal structure of LaPO 4 was accurately repr oduced using atomistic simulations and previously reported interatomic potential parameters. Defect calculations on LaPO 4 indicate that intrinsic defects of all types are energetically unfavorable and hence the defect chemistry in the material is likely t o be entirely controlled by extrinsic dopant substitution. Dopant dissolution studies indicated that strontium and calcium were the most suitable dopants for conductivity enhancement in

PAGE 119

119 LaP 5 O 14 while barium and magnesium were unfavorable. The local struc tures around oxygen vacancies and protonic defects were investigated using atomistic simulations and the results obtained support the formation of pyrophosphate ion and hydrogen phosphate ion in acceptor doped LaPO 4 The experimental results obtained for LaPO 4 are discussed in comparison with LaP 5 O 14 and critically analyzed on the basis of crystallochemical concepts in the following chapter.

PAGE 120

120 CHAPTER 6 C OMPARISON OF L a PO 4 AND L a P 5 O 14 6.1 Factors Affecting Protonic Conductivity Since the diffusion of ch arged species (proton jump which leads to electrical conduction in this case) is a thermally activated process, it generally follows the well known Arrhenius relationship which is shown below: (6 2) Where is the conductivity o is the pre exponential factor, E a is the activation energy, R is the universal gas constant and T is the absolute temperature. Figure 6 1. Effect of variation of the pre exponential factor ( o ) and activation energy (E a ) o n the conductivity. T as shown in Figure 6 1. In such a plot a linear dependence is obtained if the pre

PAGE 121

121 exponential factor and the activation energy are constant over the entire temperature range of measurement (which is generally true) Using such a plot the activation energy can be directly interpreted from the slope of the graph For achieving an enhancement in conductivity at a particular temperature a higher valu e of the pre exponential factor as well as lower activation energy is desirable. This is shown schematically in Figure 6 1. The activation energies and pre exponentials for the individual lines are indicated alongside on the graph. It can be seen th at as the pre exponential is m ultiplied by a factor of 10 while the activation energy is kept constant the line shifts towards the higher conductivity region parallel to itself However, if the activation energy is varied keeping the pre exponential constant then the slope of the line changes leading to a higher conductivity corresponding to a lower activation energy and vice versa. Taking into consideration the atomistic view of diffusion, the pre exponential term can be expressed by the following equation 93 : (6 3) W here C i is the concentration of charge carriers (protonic defects) k is the B q v is the charge on the conducting species, v o is the pre exponential of the attempt jump frequency, is the jump distance, is a geometrical constant dependant on the crystal structure of the material, is the defect coordination number, and m is the entropy difference between ground states and activated states of the proton It is clear from the above coordination number would contribute towards higher proton conductivity. In addition, a higher concentration of mobile charge carriers C i (which can be controlled by varying

PAGE 122

122 the dopant concentration) also contributes to a higher conductivity. In this case, t he defect coordination number is the number of nearest neighbor oxygen ions available for the proton jump. 6. 2 Ef fect of Inter Oxygen Distance on Activation Energy Ba sed on computational investigations it is known that the proton spends most of its time performing rotational diffusion around an oxygen ion without much restriction si nce this type of diffusion has low activation energy ~ 0.1 eV. 94 96 From quantum molecular dynamics (QMD) simulations and from quasielastic neutron scattering experiments on perovskite type oxides it is understood that the proton resides in one of several potential wells which are located around an oxygen ion, preferentially along the direction of the neighboring oxygen ion. 90,95 102 Rotational diffusion refers to the dynamic transfer process of the proton around a single oxygen ion across these potential wells. The proton jump (between two neighboring oxygen ions), which actually contributes to electrical conduction (charge transport) is a much rarer event (since it requires a much higher activation energy ). Typical activation energies for proton conduction in ceramics range from 0.4 eV to 1.3 eV. 62,63,69,70,95,96,103 106 Table 6 1. Oxygen to oxygen ion separation and corresponding activation energy calculated b y Cherry et al. for perovskite oxides. 33 O O separation () Activation energy (eV) 2.67 0.29 2.76 0.40 2.90 0.66 The relation between the activation energy fo r the proton jump and the inter oxygen distance in perovskite oxides has been previously investigated by Cherry et al using quantum mechanical techniques. 33 They calculated the activation energy for the proton jump by calculating the energy dif ference between the ground state and the

PAGE 123

123 barrier or the transition state. The oxygen to oxygen separation and corresponding activation energies obtained are shown in Table 6 1 It can be seen that the activation energy for the proton jump decreases as th e separation between the two oxygen ions decreases. The relationship between the activation energy for the proton jump and the oxygen to oxygen distance can also be qualitatively understood from the plot shown in Figure 6 2. The authors here wish to cla rify that the Coulombic interaction between the ions were not considered in the preparation of this graph and this graph is shown solely for the purpose of a qualitative understanding of the relationship between the activation energy and the oxygen to oxyg en ion distance. A more detailed analysis involving the Coulombic interactions and quantum mechanical considerations is more appropriate but is outside the scope of this study. The parameters used for plotting the graph are listed in section 3.4 and a de tailed description of the same can be found in Saul et al. 32 According to the interaction function there is a potential well close to an oxygen ion while the value of the potential increases to a constant value at larger distances from the oxygen ion. 33,107 The proton is generally located along the line joining two neighboring oxygen ions (although it may be slightly displaced due to repulsion from neighboring cat ions 95 ) in a potential well close to one of the oxygen ions. To jump to the adjacent potential well (close to the neighboring oxygen ion) the proton has to cross the region between the two oxygen atoms where the potential is high because the proton is far away f rom either of the oxygen ions. 33 The above behavior also implies that if the oxygen to oxygen distance is low ered (assuming the other factors affecting the shape of the well remains unchanged) then it will lead to lower activation energy.

PAGE 124

124 Such a decrease in the activation energy for the proton jump on lowering the oxygen to oxygen distance is reported for a pero vskite type oxide from computational investigations which we hereby extend to the phosphate materials 33,108 Figure 6 2. The effect of reducing the oxygen oxygen ion distance on the barrier for proton jump ( or activation energy) It is thereby here proposed that tailoring a material to lower the jump distance for the proton or a lower O O distance, sh ould lead to a smaller activation energy and higher proton conductivity, provided all other relevant factors in the material are minimally disturbed. In addition, it is also expected that other factors such as a higher oxygen to oxygen coordination concentration of protonic defects and proton jump distance would favor a larger pre exponential factor also enhan cing the proton conducti vity in a material. This work investigates the combined effect that these two enhancement avenues have on the proton conductivity of ceramic compounds.

PAGE 125

125 6. 3 Comparison of C onductivit y and Model for Proton Conduction In LaPO 4 and in LaP 5 O 14 three different compositions were synthesized for conductivity measurement. The measured conductivity of the LaP 5 O 14 samples was found to increase as the dopant content was increased up to 5 mol% as seen in Figure 6 3 The highest conductivity measured for the 5 mol% Sr doped LaP 5 O 14 sample s was 1.01 x 10 4 S/cm at 600 o C. Figure 6 3. Conductivity of undoped, 1 mol% Sr and 5 mol% Sr doped LaPO 4 ( open symbols ) and LaP 5 O 14 ( closed symbols) Conductivity measurements on LaPO 4 exhibited a similar trend of increasing conductivity with in creasing dopant concentration. The maximum conductivity measured for the 5 mol% Sr doped LaPO 4 sample was 7.00 x 10 6 S/cm at 600 o C.

PAGE 126

126 However, for all compositions synthesized, the conductivity of LaP 5 O 14 samples was higher than the respective LaPO 4 samples in the entire temperature range It can also be observed from Figure 6 3 that a greater enhancement in conductivity is observed in LaP 5 O 14 as compared to LaPO 4 as we move from 1 mol% to 5 mol% dopant concentration. This could be due to differences in the homogeneity of dopant distribution within the single crystals of La P 5 O 14 as compared to the sintered samples of LaPO 4 Considering the synthesis procedure for LaPO 4 and given the polycrystalline nature of the sample, dopant segregation at the grain boundary is quite probable. This would result in a lower dopant concent ration in the bulk grain leading to a poorer enhancement in bulk conductivity. Previously dopant segregation and secondary phase formation has been observed by Norby et al. in acceptor doped samples of LaPO 4 109 On the contrary, the crystals of LaP 5 O 14 were grown slowly from a phosphoric acid melt containing dissolved oxides of the host and the dopant cation which may have led to a more homogeneous di stribution of the dopant and correspondingly a higher enhancement in conductivity upon increasing dopant concentration from 1 mol% to 5 mol% Another possibility to consider could be the relative solubility limit for dopant incorporation which may be high er for LaP 5 O 14 as compared to LaPO 4 The activation energies for the undoped, 1 mol% and 5 mol% Sr 2+ doped LaPO 4 samples were 1.29 eV, 1.04 eV and 1.09 eV respectively. The activation energies for the similarly doped LaP 5 O 14 samples were much lower and had values of 0.85 eV, 0.76 eV and 0. 80 eV as summarized in Table 6 1 The difference in the activation energies

PAGE 127

127 of the two materials (~ 0. 3 eV) can be correlated with the respective crystal structures and is explained as follows Table 6 1 Obtained a ctivation energies for three different compositions of LaPO 4 and LaP 5 O 14 LaPO 4 LaP 5 O 14 Material Activation energy (eV) Material Activation energy (eV) Undoped 1.27 Undoped 0.85 1 mol% Sr 1.04 1 mol% Sr 0.76 / 0.89 5 mol% Sr 1.09 5 mol% Sr 0.80 ndicates conductivity measurement along the transverse direction as indicated The crystal structures of the two materials have been discussed previously in sections 4.2 and 5.2. The basic difference between the two structures lies in the isolated versu s condensed structure of the phosphate anion which presents vastly different avenues for proton transport within the material. The various possibilities for proton migration in LaPO 4 and LaP 5 O 14 have been graphically depicted in Figure 6 4 In both the s tructures La 3+ cations are not shown for purposes of clarity. The jump distances for the migrating proton are also depicted in Figure 6 4 (a) and (b). Due to the close dimensional resemblance of the PO 4 tetrahedra in both the phosphates, the energetics o f proton migration within a tetrahedron which includes rotational motion around a single oxygen ion or a jump betwe en adjacent oxygen ions can be expected to be similar. This is obvious when comparing the mean intra tetrahedron oxygen to oxygen ion dist ances for LaPO 4 and LaP 5 O 14 which are very similar and have the values 2.5113 and 2.5150 respectively as summarized in Table 6 2 However, in LaPO 4 the jump distance is much higher when the proton has to hop between oxygen ions belonging to two neigh boring tetrahedra (> 2.8 ). On the contrary in LaP 5 O 14 the condensed nature of the tetrahedra in the phosphate anion

PAGE 128

128 allow for an easier pathway for conduction. Thus a migrating proton will encounter a smaller jump distance and hence a smaller energy ba rrier (activation energy) for proton conduction which is in agreement with the experimental results Figure 6 4 Proton jump distances in (a) LaPO 4 and (b) LaP 5 O 14 Since the corner sharing PO 4 tetrahedra in LaP 5 O 14 form a three dimensional network ed structure multiple paths are available for the protons to travel along different directions in the material. This can probably explain the similar activation energy calculated for protonic conduction along two different directions in 1 mol% Sr doped La P 5 O 14 (0.76 0.02 eV and 0.89 0.02 eV). However, one must treat the above analysis with some caution as there could be some obvious problems with treating the structure as a static rigid lattice. For example it has been observed previously in the cas e of perovskite oxide materials that the picture at the atomic level is significantly distorted by the local alterations caused in the material due to atomic vibrations. 33 Compared to the perovskite oxides the situation in these complex phosphate materials could be much more intricate which would entail a much deeper analysis. Nonetheless, our current analysis which relies on the average

PAGE 129

129 spatial arrangeme nt of ions in a lattice (as can be obtained from XRD measurements) can be used as the fundamental basis for detailed investigations in the future. Table 6 2. Comparison of mean inter and intra tetrahedral oxygen oxygen ion distances in LaPO 4 and LaP 5 O 1 4 System Mean oxygen oxygen ion distance ( ) Intra tetrahedra Inter tetrahedron LaPO 4 2.5113 0.06 2.8410 0.04 LaP 5 O 14 2.5150 0.05 In addition to the justification presented above, there could be other factors that also affect the mobility of protonic defects (and hence the activation energy) such as protonic defect dopant site association ( ). There have been computational investigations on such defect interactions for perovskite type oxide conductors where it has been mentioned that higher interaction energy could lead to a lower conductivity by defect trapping. 110,111 In case of lanthanum phosphates, higher interaction energy for the proton dopant association in LaPO 4 could b e a potential explanation for the lower conductivity and higher activation energy observed as compared to LaP 5 O 14 The calculated value of association energy for Sr doped LaPO 4 ( 0.57 eV) is indeed lower than that for similarly doped LaP 5 O 14 ( 0.39 eV) an d is i n support of the above potential explanation. According to DFT studies by Jonghe et al. on energetics of proton migration in LaPO 4 both inter tetrahedral and intra tetrahedral proton transfer processes have almost similar but relatively high activ ation energies (~ 0.8 eV). 112 However, the activation energy for rotational motion around an oxygen ion is comparatively much lower (0.15 eV) Such low energy values for rotational motion are consistent with other computational investigations on proton conducting perovskite oxides. 94 96 The rotat ional

PAGE 130

130 motion of a proton does not contribute to proton transport directly but can substantially assist in the process in LaP 5 O 14 Essentially this means that an inter tetrahedral proton jump which has a high energy barrier in LaPO 4 can be substituted for by an energetically inexpensive low barrier rotation around an oxygen ion in LaP 5 O 14 With regards to the mechanism of proton conduction in undoped ultraphosphates there is an alternate model that has been proposed by Hammas et al. According to the mode l protonic conduction shown by condensed phosphate materials occurs along chemically adsorbed water molecules which lie all along the condensed phosphate anion in the material structure rather than along the phosphate anion itself. 66 The proton transfer process is described as occurring by cleavage and formation of hydrogen bonds between a hydronium ion and an adjacent water molecule. This could proba bly explain the measured conductivity in the undoped material reported at low temperatures, but it does not explain how protonic conduction can be observed in phosphates at much higher temperatures (~700 o C) where chemically adsorbed water does not exist in the material. Also, since the material measured by Hammas et al. is undoped (and it does not contain any structural protons) it is unclear how a proton would be incorporated in the material leading to the formation of a hydronium ion from a water molecul e. Although not focus of the current study, it can be envisioned that the defect chemistry in LaP 5 O 14 is quite similar to LaPO 4 69 Analogous to LaPO 4 oxygen deficiencies created from acceptor doping can be expected to lead to additional local condensation in the ultraphosphate structure. The locally condensed phosph ate group

PAGE 131

131 can then by hydrolyzed in the presence of ambient water vapor to form protonic defects in the material. In this study it has been observed that the condensation of the phosphate anion facilitates the proton transfer process in the material lead ing to enhanced proton conductivity. Although the obtained conductivity is otherwise modest it shows a proof of concept which can be applied to other phosphate systems (Ln 2 O 3 P 2 O 5 Ln = lanthanides) and other material systems such as germanates which al so exhibit a rich and diverse anion chemistry similar to the phosphates. 17,20,113,114

PAGE 132

132 CHAPTER 7 C HARACTERIZATION OF T IN (IV) PYROPHOSPHATE (S n P 2 O 7 ) 7 .1 Pyrophosp hates According to the structural classification of condensed phosphates presented in section 2.4.3 the pyrophosphates belong to the category of oligophosphates with n = 2. Thus they are also sometimes referred to as diphosphates. Several pyrophosphate materials have been investigated for a variety of applications. Mixed cation pyrophosphate materials of the type ABP 2 O 7 such as LiM 1.5 P 2 O 7 (M = Fe, Ni, Co and Cu) have been investigated as electrode material for lithium ion batteries. 115 Some binary phosphates such as CuFeP 2 O 7 have been investigated for ferromagnetic properties while others such as K 2 Cu 3 (P 2 O 7 ) 2 .3H 2 O have be en studied as mic ronutrient material to boost fertilizers. 116,117 However, the number of investigations on pyrophosphate materials focusing on the conductivity of the same are somewhat limited and can be summarized as follows. T he conductivity of p yrophosphate s o f the type (AP 2 O 7 where A = V, Ti, Zr) has been studied for potential application of the material as a catalyst for oxidative dehydrogenation of orga nic compounds. 65,118 The sodi um ion conducti on in undoped sodium pyrophosphate (Na 4 P 2 O 7 ) is well investigated For application as proton conducting electrolyte materials the following pyrophosphate materials have been investigated. In the low temperature range (0 100 o C) and under humid conditions z irconium pyrophosphate based materials show moderately high conductivity. 119 121 The proton conduction in these materials is believed to be mainly occurring along the surface via adsorbed water molecules or surface

PAGE 133

133 hydroxyl groups and as such many of the investigations have focused on conductivity enhancement by increasing the surface area of the synthesized material. Solid acid orthophosphate materials exhibit high proton conductivity but low thermal stability at high temperatures. To enhance their thermal and chemical stability at high temperatures s ome pyrophosphate solid acid orthophosphate composite materials of the type CsH 2 (PO 4 ) SiP 2 O 7 were studied by Eguchi et al. 122 124 In general the pyrophosphate materials that have been studied so far for proton conduction do not contain any stoichiometric protons in the chemical formula The protons are incorporated as charge compensating defects by a cceptor doping on the cation site in such materials However, it could be interesting to study some solid acid pyrophosphate materials (which contain intrinsic protons) such as GdHP 2 O 7 LaHP 2 O 7 SmHP 2 O 7 and LuHP 2 O 7 which have all been synthesized previou sly. 125 Such pyrophosphates conta ining an acidic proton are analogous to the well investigated solid acid sulfates such as CsHSO 4 It is known that the solid acid sulf ate and selenate salts show very high protonic conductivity but their practical applications are very limited due to high solubility in water and low decomposition temperature (~150 o C). The hydrogen phosphates on the contrary are thermally stable to much higher temperatures (~500 o C) and are insoluble in water. 125 In this chapter, the results from the synthesis and characterization of acceptor doped tin py rophosphate (SnP 2 O 7 ) are discussed. 7. 2 Crystal S tructure In the last few years there have been several structural studies on compounds belonging to the family of pyrophosphates with the general formula AP 2 O 7 and

PAGE 134

134 ABP 2 O 7 126 131 This family of materials was initially thought of having a simple cubic structure as shown in Figure 7 1 (a) consisting of AO 6 octahedra and PO 4 tetrahedra. (a) (b) (c) (d) Figure 7 1. (a) Ideal crystal structure of SnP 2 O 7 as seen along the [100] direction (b ) [P 2 O 7 ] 4 (c) 3 x 3 x 3 superstructure (d) P O P bond angle

PAGE 135

135 The space group of such a structure is Pa (No. 205). The Wyckoff position of Sn is 4a and as shown in Figure 7 1 (a) and it occupie s the corners and face centers of the unit cell. The P and the O atoms take up the positions 8c and 4b, 24d respectively. The octahedra of SnO 6 and the tetrahedra of PO 4 are arranged in a corner sharing pattern to produce a three dimensional network. Th e two PO 4 tetrahedra are paired together by corner sharing one oxygen ion to form a P 2 O 7 group. This is an ideal structure for the general family of compounds of the type AM 2 O 7 and restricts the P O P bond in the P 2 O 7 group to 180 o However, this is an e nergetically unfavorable orientation as it lead s to very short P O P bond lengths. The required bond lengths are accommodated in the structure by bending at the bridging oxygen in the P 2 O 7 group and this is shown clearly in Figure 7 1 (c) and (d). The va lue of the average P O P bond angle for SnP 2 O 7 is 134 o which is quite similar across the entire range of condensed phosphate materials including metaphosphates and ultraphosphates. 25,72,73,132 134 The energetical ly favorable structure for SnP 2 O 7 is slightly different from the ideal crystal structure described earlier and can be more accurately represented by considering a 3 x 3 x 3 cubic superstructure. A comparison of the two structures is shown in Figure 7 1 T he ideal cubic unit cell has the P O P bond angles restricted to 180 o (shown in Figure 7 1 ( a)) while in the superstructure the same is no longer 180 o A blown up image of the same can be seen in Figure 7 1 (d) where it can be seen clearly that the two a djacent PO 4 tetrahedra are bent with re spect to each other. The relative size of the ideal unit cell with respect to the superstructure is shown in Figure 7 1 (c)

PAGE 136

136 7. 3 X R ay D iffraction (XRD) The powder XRD pattern obtained for all the doped and undoped samples of SnP 2 O 7 is shown in Figure 7 2. All the pe aks observed for all of the undoped and doped samples (except Ga doped) were indexed according to the compound SnP 2 O 7 (JCPDS 29 1352). The powder pattern from the Ga doped sample showed one very low in tensity impurity peak which matched with the tallest peak from the theoretical XRD pattern of gallium metaphosphate (Ga(PO 3 ) 3 ). Figure 7 2. Powder XRD p atterns of crushed as sintered pellets of doped and undoped SnP 2 O 7 The crystal structure was Sn P 2 O 7 originally proposed to be cubic (Pa ), however recent studies by Gover et al. 130 suggest that the room temperature symmetry may be lower than cubic due to reasons described in the previous section The peaks in the

PAGE 137

137 XRD were labeled a ccording to JCPDS 29 1352 and the unit cell parameter for SnP 2 O 7 was determined as 7.9331 0. 00 1 ( the peak positions were obtained by fitting the data with the Pseudo Voigt function using Xfit TM software) The effect of sample displacement during XRD data collection was accounted for as described elsewhere. 135 The calculated lattice parameters are summarized in Table 7 1 and are in accordance with the expected trend since the Ga 3+ ion has a smaller ionic radius as compared to Sn 4+ while all the other dopa nt cations have higher ionic radii. The ionic radii are also listed in Table 7 1 for comparison. Table 7 1. Comparison of lattice parameters for the various doped and undoped materials. Dopant (10 mol %) Lattice parameter () Ionic radii () Ga 7.9327 0.003 0.620 Undoped 7.9331 0.001 0.690 Mg 7.9350 0.001 0.720 Zn 7.9407 0.003 0.740 Sc 7.9412 0.001 0.745 In 7.9429 0.002 0.800 In the samples that were initially prepared i mpurity peaks corresponding to unreacted SnO 2 were observed which w ere due to the loss of phosphoric acid during calcination. This was overcome and phase pure samples were obtained by increasing the (Sn+D):P ratio from 1:2 to 1:3 in the starting mixture. Sintering in an open crucible resulted in insuffi cient densification of the pellet which was also accompanied by weight loss. Such weight loss has been previously reported during sintering of other phosphate materials such as GeP 2 O 7 and CeP 2 O 7 In a study on the sintering of cerium pyrophosphate (CeP 2 O 7 ) conventional sintering, spark plasma sintering and hot pressing techniques were used by Tanaka et al. 136 It was found that the former two methods led to insufficient densification in the material

PAGE 138

138 However according to the study the weight loss issues were largely avoided by using hot pressing Decomposition associated weight loss has also been ob ser ved during the high temperature heat treatment o f higher condensed phosphates of lanthanum such as LaP 5 O 14 and La(PO 3 ) 3 by Park et al 19,126 In these cases, the weight loss was attributed to the loss of P 2 O 5 (due to decomposition) and the issue was overcome by using a sealed vessel. The problems of weight loss and insufficient densification in the current work were avoided when sealed alumina crucibles were used for sintering heat treatments (as described in section 3.1.3 ). (a) (b) Figure 7 3. SEM imag es of fractured surface of SnP 2 O 7 sample sintered at 1400 o C for 10 h. (a) Average grain size ~10 m. (b) Porosity at triple grain junctions. 7. 4 Scanning E lectron M icroscopy (SEM) A typical microstructure of the fractured surface of the sintered pelle t is shown in Figure 7 3 (a) The average grain size wa s about 10 m (as seen in Figure 7 3) which is much larger than the initial average particle size (before sintering) which is less than 1 m. This indicates that there is significant grain growth occ urring during the sintering process. Some porosity can be seen mainly concentrated at the triple grain junctions (as seen in Figure 7 3 (b)).

PAGE 139

139 7. 5 Conductivity M easurement 7. 5 .1 Comparison with P reviously P ublished D ata Given the interest of the present work in elucidating the nature of the contrasting conductivity data for tin pyrophosphate reported in literature, and as a foundation for data reliability and calibration of the measurement setup, the conductivity of 20 mol% Eu doped SrCeO 3 and 5 mol% Sr d oped LaPO 4 was measured. The conductivity was measured under dry atmospheric conditions (under flowing dry N 2 ). The measured conductivity for these two materials is shown in Figure 7 4 and compared with previously published results in literature measured under similar atmospheric conditions 62,69,137 The latter was synthesized by using the experimental procedure described by Norby et al. 62 The measured values were found to match well (within 5%) with published values. Since the purpose of this work is to contrast t he obtained conductivity data with previously reported data, the atmosphere conditions (unhumidified air) were kept co nsistent throughout this study with those used in previously reported studies. 64,138 T he conduc tivity values for undoped SnP 2 O 7 and 10 mol% In 3+ doped SnP 2 O 7 as measured here are also shown in Figure 7 4 These values compare well with the conductivity values as reported by Tao for undoped SnP 2 O 7 and 8 mol% In doped SnP 2 O 7 in the temperature range 300 o C 900 o C. 138 Dopant addition increases the conductivity of SnP 2 O 7 by about an order of magnitude as observed from reported values of Tao and also res ults from the current study. By contrast Hibino et al report a much higher conductivity for 10 mol% In 3+ doped SnP 2 O 7 (1.95 x 10 1 S/cm)

PAGE 140

140 Moreover, there is also a significant difference in the activation energies for conduction. The activation energy for conduction was calculated by fittin g the temperature variation of conductivity with the well known Arrhenius relationship: ( 7 1) Where is the conductivity, A is the pre exponential, E a is the activation energy, R is the universal gas constant, and T is the absolute temperature. Figure 7 4. Measured conduct ivity values of 5 mol% Sr: LaPO 4 and 20 mol% Eu : SrCeO 3 as compared with conduc tivity values for SnP 2 O 7 samples. The values of activation energy obtained in this study are compared with previously published values in Table 7 2 The activation energ y values obtained are as

PAGE 141

141 follows: 1.36 eV for 10 mol% In doped SnP 2 O 7 (current work ), 1.02 eV for 8 mol% In doped SnP 2 O 7 (Tao) and 0.21 eV for 10 mol% In doped SnP 2 O 7 (Hibino et al. ). The drastically different activation energy value obtained by Hibino et al. is a clear indication that the mechanism of conduction in the latter (Hibino e t al. ) is significantly different from the previous two (Tao and current work). Table 7 2. Activation energies obtained for In 3+ doped samples as measured in this study compared with previously published literature. Material Activation energy (eV) Sinte ring temperature ( o C) Reference In 0.08 Sn 0.92 P 2 O 7 1.02 1000 Tao (2009) 138 In 0.1 Sn 0.9 P 2 O 7 1.36 1400 Current work In 0.1 Sn 0.9 P 2 O 7 0.21 *650 Hibino et al (2006) 139 H 3 PO 4 (fused) 0.29 N.A. Greenwoo d et al. ( 1959) 140 T he temperatures used fo r sintering heat treatment on the pellets previous to conductivity measurement are also shown *Sintering was not performed on the pressed pellets. Temperature shown represents the calcination temperature. High temperature ceramic protonic conductors generally have activation energ y values in the range 0.4 eV 1.31 eV. 62,63,69,95 One of the possible explanations for the above difference in activation energy values could be that leftover phosphoric acid from synthesis (due to insufficient annealing time) played a major role in the charge transport in samples prepared by Hibino et al It appears to be so since the conductivity reported by Hibino et al is close to the conductivity of fused phosphoric acid in a similar temperature range (also shown in Table 7 2 ). 140 The activation energy for H 3 PO 4 was calculated to be 0.29 eV which is also close to the value (0.21 eV) obtained for samples prepared by Hibino et al It has to be remembered that Hibino et al did not use any sintering procedure but rather measured conductivity on pressed powder (pellets). By contrast, high sintering temperatures used by Tao (1000 o C) and in the current work

PAGE 142

142 (1400 o C) eliminate the possibility of any leftover phosphoric acid from synthesis in these samples. 7. 5 .2 31 P MAS NMR S pectroscopy To investigate the reason for the differences in obtained conductivity data by different research groups 31 P MAS NMR spectroscopy was used to compare the as calcined and sintered samples. T he NMR spectra obtained for crushed powder made by grinding a sintered sample are compared with that obtained for the as calcined powder sample in Figure 7 5 (a) Spectrum of sintered sample shows a large number of closely spaced peaks with chemical shifts between 25 ppm and 50 ppm. Figure 7 5. Comparison of c rushed powder and as sintered pellet s of doped SnP 2 O 7 by using (a) 31 P MAS NMR spectroscopy and (b) X RD These peaks correspond to the phosphorus atoms belonging to the pyrophosphate group (P 2 O 7 4 ) in SnP 2 O 7 as previously reported by Gover et al. 130 The spectrum for the as calcined powder shows only a single broad peak centered at the same chemical shift and with the same overall shape as the pyrophosphate group in the sintered sample. It can be assumed to be due to the pyrophosphate group in the calcined powder. It could be that the annealing effect during sintering has reduced the disorder

PAGE 143

143 in the crystal lattice, permitting most of the crystallographically distinct phosphorous sites to be resolved. This interpretation is consistent with the reduced line broadening observed in the XRD of the sintered sample c ompared to the as ca lcined sample. Furthermore, the calcined powder shows an additional prominent peak with a chemical shift of almost 0 ppm. Since the reference standard used was orthophosphoric acid this peak can be safely ascribed to phosphorus atoms belonging to the ort hophosphate group (PO 4 3 ). T he XRD p atterns for the sintered and the as calcined samples are compared in Figure 7 5 (b) The two XRDs show peaks only corresponding to SnP 2 O 7 hence it can be safely assumed that the orthophosphate group (seen in NMR spect rum) belongs to some amorphous phase in the as calcined powder. This amorphous phase is most likely leftover phosphoric acid (surface adsorbed) from the synthesis procedure due to insufficient calcination time. Thus it was confirmed that an additional ph ase was present in the as calcined samples which was absent in the sintered samples. 7. 5 .3 Acceptor D oped SnP 2 O 7 (D = I n M g S c Z n and Ga) The Nyquist plot obtained for 10 mol% In 3+ doped SnP 2 O 7 at 600 o C is shown in Figure 7 6 Three distinct but not completely resolved arcs can be seen representing grain, grain boundary and electrode polarization phenomena. Since, the arc correspondi ng to the electrode impedance was unstable at lower temperatures, for the sake of consistency only the part of the dat a representing the grain and grain boundary impedance was considered during the fitting process at all temperatures. The low frequency data representing the electrode impedance was not used in the fitting process. Therefore, the equivalent circuit used t o fit the data consisted of 2 R CPE (constant phase element) circuits in series with an inductance (shown in Figure 7 6). Inset shows Bode plot where Im(Z) is plotted as a function of the frequency which

PAGE 144

144 three different polarization phenomena. As mentioned previously, only the part of the data representing the grain and grain boundary impedance was used in the data fitting process. Since the grain and grain boundary arcs (in the Nyquist plots) are not w ell resolved the Bode plots were used to ensure that the fitting process had accounted correctly for the grain and grain boundary impedance. Figure 7 6. Nyquist plot of 10 mol% In doped SnP 2 O 7 (600 o C). T he temperature variation of grain boundary, g rain and total conductivity of 10 mol% In doped SnP 2 O 7 is shown in Figure 7 7 At temperatures below 325 o C the impedance of the sample was too large and fell outside the measurement limit of the instrument. The grain ionic conductivity was higher than th e grain boundary conductivity in the entire temperature range of measurement.

PAGE 145

145 It is well known that introduction of dopant cation in a host lattice produces strain in the host lattice due to ionic radii mismatch of the dopant cation and the host cation. For this study dopant cations were shortlisted on the basis of their ionic radii to minimize lattice strain produced upon dopant incorporation in the host lattice. Since the crystal structure of SnP 2 O 7 consists of SnO 6 octahedra, the ionic radii of dopan t cations in six fold coordination were used. 129,141 Based primarily on these two criteria, In, Ga, Zn, Mg, and Sc, were selected as dopant for this study. Multivalent cations were excluded from the list of dopan ts. Figure 7 7. Variation of grain boundary, grain ionic and total conductivity in 10 mol% In doped SnP 2 O 7 with inverse temperature. Figure 7 8 shows the temperature dependence of total conductivity for undoped SnP 2 O 7 and D 0.1 Sn 0.9 P 2 O 7 (D = In, Zn, Mg, Ga, and Sc). Among the dopants analyzed

PAGE 146

146 Zn 0.1 Sn 0.9 P 2 O 7 shows the highest conductivity of 2.84 x 10 6 S/cm at 600 o C. Below 400 o C, Mg 0.1 Sn 0.9 P 2 O 7 shows the highest conductivity. Sc 0.1 Sn 0.9 P 2 O 7 and In 0.1 Sn 0.9 P 2 O 7 exhibit the low est conductivity among the dopants analyzed. Figure 7 8. Total conductivity of undoped SnP 2 O 7 and D 0.1 Sn 0.9 P 2 O 7 (D = In, Zn, Mg, Ga, and Sc) plotted as a function of inverse temperature. This result is expected since partial substitution of the t in cation by a cation of lower valence produces negatively charged defect sites which can be compensated by the formation of oxygen ion vacancies in the structure. Acceptor doped SnP 2 O 7 shows a much higher conductivity than undoped SnP 2 O 7 In the presenc e of ambient moisture these vacancies may be filled up to form protonic defects leading to a higher conductivity.

PAGE 147

147 This equilibrium ca n be represented as shown below, ( 7 2 ) Among the other pyrophosphate materials investigate d f or conductivity are TiP 2 O 7 ZrP 2 O 7 and CeP 2 O 7 65,136 The conductivity of TiP 2 O 7 was 2.51 x 10 8 S/cm while that of ZrP 2 O 7 was 7.94 x 10 9 S/cm at 500 o C These values are slightly lower than those obtained for undoped SnP 2 O 7 at the same temperature. It has been shown for acceptor doped LaPO 4 that protonic defects are the dominant defects at low temperatures. 62 At temperatures in excess of about 900 o C the formation of oxygen ion vacancies is favored since reaction shown in equation 7 2 is generally exothermic. How ever, in this study conductivity measurements were carried out in the intermediate temperature range, hence protonic defects can be expected to be the majority charge carriers in the material as compared to oxide ion vacancies. Alternatively, the oxide io n vacancies may also adsorb oxygen from ambient atmosphere leading to the formation of holes (p type electronic conduction) as expr essed in the following equation. ( 7 3 ) High te mperatures and parti al pressures of oxygen promo te p type electronic conduction as obvious from equation 7 3 Such behavior has been observed in doped LaPO 4 and also in doped SrCeO 3 and BaCeO 3 69,70 7. 6 Proton T ransference N umber M easurement Proton transferen ce number measurements were carried out on the undoped and 10 mol% In doped SnP 2 O 7 samples as described previously in the temperature range from 300 o C 500 o C as shown in Figure 7 9 Proton transference number

PAGE 148

148 measurements are required to ascertain the ex act contribution of protonic conduction to the overall charge transport in the material. Figure 7 9. Proton transference number of undoped and 10 mol% In doped SnP 2 O 7 as a function of temperature (300 o C 500 o C). The undoped sample shows a low tran sference number of (~0.3) in the measured temperature range. A low transference number is expected since the undoped sample itself does not contain any structural protons or protonic defects. However, the doped sample shows a much higher transference num ber of about (~0.7) which is due to the contribution to charge transport from protonic defects incorporated in the material according to equation 7 2. It is likely that p type electronic conduction (according to equation 7 3) is responsible for the remain ing charge transport (~30%) in the material in the measured temperature range.

PAGE 149

149 7. 6 Summary and Conclusion Materials with the composition, D 0.1 Sn 0.9 P 2 O 7 (D = In, Ga, Mg, Zn, and Sc) in addition to undoped SnP 2 O 7 were prepared. Sufficient densification (>90%) was achieved and weight loss issues were minimized by optimizing the sintering heat treatment procedure. The conductivity of the system D 0.1 Sn 0.9 P 2 O 7 was measured in the intermediate temperature range from 300 o C 600 o C. Acceptor doping led to an increase in conductivity of about an order of magnitude for all the dopants in comparison to undoped SnP 2 O 7 Proton transference number measurements i ndicated that the charge transport in acceptor doped SnP 2 O 7 was mostly protonic (~70%) as compared to undoped SnP 2 O 7 in which electronic conduction was the dominant charge transport mechanism. Within the dopants analyzed Zn 0.1 Sn 0.9 P 2 O 7 showed the highes t conductivity of 2.84 x 10 6 S/cm at 600 o C. This conductivity value was several orders of magnitude lower than the previously reported conductivity for acceptor doped SnP 2 O 7 but consistent with values recently reported in literature. The results of 31 P MAS NMR spectroscopy revealed that the as calcined powder of doped SnP 2 O 7 contained a significant fraction of leftover phosphoric acid from synthesis procedure. The difference in conductivity values in the previously reported studies was ascribed to th e presence of leftover phosphoric acid in the as calcined powder which results in a higher conductivity. Thus adequate measures need to be taken while dealing with samples for which the synthesis method involves phosphoric acid, so as to ensure that surfa ce contamination does not adversely affect the characterization of the material.

PAGE 150

150 CHAPTER 8 S UMMARY AND FUTURE WO RK 8 .1 Summary Proton exchange membrane fuel cells (PEMFCs) present a viable alternative to obviate the problems in the energy sector posed by increasing levels of greenhouse gas emissions. However, some major obstacles need to be overcome to make this technology more economical so as to ensure its widespread use. T o bring down the overall cost of the technology electrolyte materials need to be developed which can function in a higher temper ature range ( 300 o C 600 o C ) than currently possible. At these temperatures i ncreased efficiency and increased tolerance of the platinum electrodes to CO (carbon monoxide) poisoning can be achieved in fuel cells The focus of this study wa s on exploring the class of phosphate ceramic materials for achieving enhanced conductivity in the desired temperature range I n addition, the aim of the study was also to expand the realm of knowledge to higher condense d phosphates which have so far received very limited attention as candidate proton conducting electrolyte materials. In this respect three different phosphate compounds namely, LaP 5 O 14 LaPO 4 and SnP 2 O 7 were investigated (about 14 different compositions) From the phosphate rich region of the La 2 O 3 P 2 O 5 binary phase diagram the orthophosphate (LaPO 4 ) and ultraphosphate (LaP 5 O 14 ) materials we re chosen for a comparative study of the electrical properties to investigate the structure property relationsh ips in phosphates The s ynthesis of lanthanum ultraphosphate (LaP 5 O 14 ) single crystals of three different compositions (undoped, 1 mol% and 5 mol% Sr 2+ doped) was carried out by precipitation from concentrated phosphoric acid solutions. Phase purity of t he prepared crystals was established by comparing the experimental

PAGE 151

151 powder diffraction pattern with the theoretical pattern and also from quantitative EDS analysis. Conductivity measurements were performed on crystal samples of all the above prepared compo sitions along the [100] direction in the temperature range 300 o C 600 o C using electrochemical impedance spectroscopy. Within the 1 mol% Sr doped LaP 5 O 14 samples conductivity was also measured along the transverse direction. The conductivity was found t o increase with increase of dopant concentration upto 5 mol% Sr doping. The highest measured conductivity was for the 5 mol% Sr doped LaP 5 O 14 samples which was 1.01 x 10 4 S/cm at 600 o C. The activation energies obtained for the undoped, 1 mol% Sr and 5 m ol% Sr doped samples were 0.85 0.01 eV, 0.76 0.02 eV and 0.80 0.01 eV respectively. The activation energy calculated for protonic conduction along the transverse direction was only slightly higher than along the phosphate anion ribbon direction (0.7 6 eV and 0.89 eV) indicating that the structural anisotropy did not have a significant impact on the activation energy for protonic conduction in the material. The conductivity of the samples was higher when the measurements were performed under h umid atm osphere This implies that the materials are pre dominantly proton conducting because in the presence of moisture additional protonic defects are formed by the dissociative absorption of water vapor which leads to an enhancement in conductivity. To aid the comparison between the two lanthanum phosphates the three different compositions of LaPO 4 synthesized were kept consistent with those of LaP 5 O 14 in terms of the type of dopant cation and concentration. Phase purity of the prepared compositions was con firmed by x ray diffraction (XRD) and energy dispersive spectroscopy (EDS). Scanning electron microscopy (SEM) was used to observe the

PAGE 152

152 microstructure of the sintered pellets and to confirm the densification of the samples. Conductivity measurements were performed on all the prepared samples in the temperature range 300 o C 600 o C under humidified and unhumidified atmosphere separately. The conductivity was found to increase with increasing dopant concentration. The highest conductivity obtained for 5 mol % Sr doped LaPO 4 samples was 3.24 x 10 5 S/cm (humidified atmosphere, 600 o C) and 7.00 x 10 6 S/cm (unhumidified atmosphere, 600 o C). For both the doped compositions the conductivity was found to increase under humidified atmosphere which is in good agreeme nt with the known model for proton conduction in acceptor doped ceramic proton conducting materials. The activation energy for proton conduction in the undoped, 1 mol% and 5 mol% Sr doped samples was 1.27 eV, 1.04 eV and 1.09 eV respectively Overall th e conductivity of all the LaPO 4 samples was lower than the LaP 5 O 14 samples in the entire temperature range while the estimated activation energy for conduction was significantly higher in the former To explain the above observation a crystallochemical pr inciple relating the oxygen to oxygen ion distance in a material to the activation energy for proton conduction was pr oposed and the experimental results obtained this work and other relevant literature were critica lly examined and further validated this p rinciple From this analysis it is shown that the condensed nature of the phosphate anion in LaP 5 O 14 can provide low energy avenues for proton transport within the material leading to enhanced conductivity in the material. The relevant properties of the lanthanum phosphates were further investigated by using atomistic simulations. Using a combination of empirically derived and previously reported interatomic potential parameters the crystal structures of LaPO 4 and LaP 5 O 14

PAGE 153

153 were accurately reproduced. De fect calculations indicated that intrinsic defects of all types are energetically unfavorable in the lanthanum phosphates and hence the defect chemistry in the se materials is likely to be entirely controlled by extrinsic dopant substitution. Dopant dissol ution studies indicated that out of the four dopants considered (D = Sr, Mg, Ca, and Ba) strontium and calcium were the most suitable dopants for conductivity enhancement while barium and magnesium were unfavorable. Proton dopant cation interaction stud ies indicated that strontium doping in LaP 5 O 14 was most suitable for conductivity enhancement because of lowest interaction energy. The local structures around oxygen vacancies and protonic defects were also investigated using atomistic simulations. In b oth materials it was found that the proton tends to lie outside the PO 4 tetrahedron most likely due to repulsion from the close ly vicinity of the P 5+ ion. In LaP 5 O 14 it was found that for all the fourteen crystallographically different oxygen ions in the structure the associated protons tend to lie in the hollow channels that exist along the a axis and c axis due to the open framework structure of the material. In LaPO 4 the results from atomistic simulations on protonic defects support the formation of p yrophosphate ion and hydrogen phosphate ion upon acceptor doping. In addition to the lanthanum phosphates the electrical properties of undoped and acceptor doped tin pyrophosphate (SnP 2 O 7 ) we re investigated, which is regarded as a potential candidate el ectrolyte material for PEMFCs. Materials with the composition, D 0.1 Sn 0.9 P 2 O 7 (D = In, Ga, Mg, Zn, and Sc) in addition to undoped SnP 2 O 7 were prepared by the conventional solid state route Sufficient densification (>90%) was achieved and weight loss issues were minimized by optimizing the sintering heat

PAGE 154

154 treatment procedure. T he conductivity of the system D 0.1 Sn 0.9 P 2 O 7 was measured in the intermediate temperature range from 300 o C 600 o C. Acceptor doping led to an increase in conductivity of about an order of magnitude for all the dopants in comparison to undoped SnP 2 O 7 Pr oton transference number measurements indicated that the charge transport in acceptor doped SnP 2 O 7 was mostly protonic (~70%) as compared to undoped SnP 2 O 7 in which electronic conduction was the dominant charge transport mechanism. Within the dopants anal yzed Zn 0.1 Sn 0.9 P 2 O 7 showed the highest conductivity of 2.84 x 10 6 S/cm at 600 o C. This conductivity value was several orders of magnitude lower than the previously reported conductivity for acceptor doped SnP 2 O 7 but consistent with values recently reported in literature The results of 31 P MAS NMR spectroscopy revealed that the as calcined powder of doped SnP 2 O 7 contained a significant fraction of leftover phosphoric acid from synthesis procedure. On the basis of results from NMR spectroscopy and EIS measurements t he difference in conductivity values in the previously reported studies was ascribed to the presence of leftover phosphoric acid in the as calcined powder which results in a higher conductivity. 8. 2 Future W ork 8. 2.1 Other U ltraphosphates The current study focuses on the comparison of electrical properties of the orthophosphate and the ultraphosphate material. According to the La 2 O 3 P 2 O 5 binary phase diagram two other phosphate materials exist at intermediate phosphate concentrations namely, lanthanum te traphosphate (La 2 P 4 O 13 ) and lanthanum metaphosphate (La(PO 3 ) 3 ) in addition to lanthanum orthophosphate (LaPO 4 ) and lanthanum ultraphosphate (LaP 5 O 14 ) The crystal structures of the same are shown in Figure 8 1.

PAGE 155

155 To complete the trend of proton conductivi ty shown in this study across the phosphate compounds strontium doped materials of similar composition need to be synthesized and characterized. In addition the trend could also be extended to the phosphate poor regions of the phase diagram which includes oxyphosphate compounds such as La 3 PO 7 and La 7 P 3 O 18 (a) (b) Figure 8 1. Crystal structure of (a) La 2 P 4 O 13 along [010] direction and (b) La(PO 3 ) 3 along [110] direction. 8. 2.2 Other S ystems ( Germanates N iobates, Borates and Arsenates ) The rich and diverse crystal chemistry exhibited by the phosphate s (which is analogous to the well known silicates) is also found in other systems such as germanates, niobates borates and arsenates 142 144 In germanates the condensation of the GeO 4 tetrahedra can lead to a variety of anion structures, such as Li 2 Ge 2 O 5 (sheet structure), Li 2 GeO 3 (chain structure), Li 6 Ge 2 O 7 (paired tetrahedra), Li 4 GeO 4 (isolated tetrahedra) and Li 8 O 2 GeO 4 (isolated tetrahedra). 145 Some ortho germanates, ortho niobates (such as acceptor doped LaNbO 4 and GdNbO 4 ) and ortho tantalates (ac ceptor doped RETaO 4 where RE = La, Nd Gd, Tb

PAGE 156

156 and Er) have been studied previously for pro ton conduction 27,146 150 An approach similar to the one used in this study for ach ieving enhancement in conductivity in the phosphates can thus be extended to the niobates, tantalates and germanates for the optimization of conductivity in these m aterials. 8. 2.3 Condensed Hydrogen P hosphates In general the pyrophosphate materials that have been studied so far for proton conduction do not contain any stoichiometric protons in the chemical formula. The protons are incorporated as charge compensating defects by acceptor doping on the cation site in such materials. However, it could be in teresting to study some solid acid pyrophosphate materials (which contain intrinsic protons) such as GdHP 2 O 7 LaHP 2 O 7 SmHP 2 O 7 and LuHP 2 O 7 which have all been synthesized previously. 125 Such pyrophosphates containing an acidic proton are analogous to the well investigated solid acid sulf ates such as CsHSO 4 It is known that the solid acid sulfate and selenate salts show very high protonic conductivity but their practical applications are very limited due to high solubility in water and low decomposition temperature (~150 o C). The hydroge n phosphates on the contrary are thermally stable to much higher temperatures (~500 o C) and are insoluble in water and thus are more suitable for potential applications as proton conducting electrolyte material in the intermediate temperature range (300 o C 600 o C) 125 I n addition higher condensed hydrogen phosphates have also been synthesized previously such as LnH(PO 3 ) 4 (methaphosphates where Ln = Bi, Sm, Eu, Gd, Tb, Dy, Ho, and Er ) and NiHP 5 O 14 (ultraphosphate ). 18,54 By contrasting the electrical prope rties across the range of condensed hydrogen phosphates of varying structural complexity, it can be ascertained if the condensation of the phosphate anion leads to enhancement in

PAGE 157

157 conductivity similar to the condensed phosphates which were the focus of the current study (not containing a structural proton). Figure 8 2. Setup for modified hydrothermal synthesis of large ultraphosphate crystals. 151 8. 2.4 Growth of L arger C rystals The size of the largest defect free crystal that was grown in this study was about 6 x 6 x 0.5 mm. However, to be used as a material for practical applications crystals of much larg er size will need to be grown. T he ultraphosphate materials are susceptible to decomposition at high temperatures (decompose to orthophosphate) prior to melting. Hence the commonly used techniques such as Czochralski method cannot be used to grow single crystals of these materials.

PAGE 158

158 However, hydrothermal single crystal growth technique which has been previously used for synthesizing single crystals of several different inorganic compounds is suited for the ultraphosphates. The slight modification to the technique is that instead of using water as the liquid medium for dissolution and growth of single crystals, phosphoric acid is used. This modified crystal growth technique has been used previously for the growth of crystals of phosphates such as AlPO 4 a nd Ga PO 4 152 In the current study the relevant oxide powder was dissolved in phosphoric acid and the ultraphosphate compound was allowed to crystallize at the crystallization temperature. However since the crystal is being grown from a saturated solution this leads to t he formation of several nucleation sites within the crucible. Thus several small crystals are formed instead of a few large ones. To obviate this problem for the growth of single crystals a setup simila r to the one shown in the Figure 8 2 can be used T he main benefit of using such a setup is that the chamber is partitioned into a distinct dissolution zone and a crystal growth zone. The dissolution of the oxide powder in phosphoric acid can occur at temperatures above 200 o C (after the evaporation of wat er) while the condensation of the phosphoric acid and crystal growth occurs at higher temperature of about 600 o C. Due to the physical separation between these two zones the solute has to diffuse through the length of the tube to reach the other zone unde r a concentration gradient. Thus the concentration of the solute in the crystal growth zone can be restricted to a suitable value such that the growth occurs exclusively on the seed crystal. The main parameters to control in such a growth technique are e xternal pressure, temperature of the two zones, length of the chamber and temperature gradient.

PAGE 159

159 8. 2.5 Performance T esting The growth of large single crystals of the ultraphosphate will enable the performance testing of the material in a fuel cell. In th e experimental setup used in the current work the sample is sealed to an alumina tube so that it can be subjected to different a tmospheres on opposite faces. The diameter of the cylindrical ceramic tube is ~2.5 cm and hence the crystal sample must suffici ently large so as to cover the tube. The I V characteristics of the sample are measured and the open circuit potential (V) is plotted as a function of the current density (A / cm 2 ). From the data obtained the power density (W/cm 2 ) can be calculated and plotted as a function of the current density which is indicative of the fuel cell performance.

PAGE 160

160 APPENDIX A STRUCTURAL CLASSIFIC ATION OF THE SILICAT ES The structural classification of the silicate compounds is presented in Table A 1. The silicates are n have been studied extensively previously. The classification of the silicates is based on the O/Si ratio in the materials. The structural unit in the anion is the [SiO 4 ] tetrahedron. The extent to which the tetrahedra are corner sharing in the anion structure leads to different structures and different O/Si ratios in these materials. Structurally the simples compounds are the neosilicates in which the O/Si ratio is 4 and their struct ure consists of isolated tetrahedra of SiO 4 which are held together by cations. As the structural complexity in the anion structure increases due to a higher degree of corner sharing in the anion the O/Si ratio decreases as can be seen in Table A 1. The corner sharing of the tetrahedra can lead to either a linear morphology of the anion (sorosilicates) or a ring morphology (cyclosilicates) or a planar morphology (inosilicates). Within these broad categories the structure can be further classified on the basis of the extent of linear corner sharing or size of the ring or the type of planar structure in the material. Structurally the most complex are the tectosilicates in which the corner sharing leads to a three dimensional networked structure of the anio n. The structure of the silicates is analogous to the phosphates in which the structural unit comprising the structure is the PO 4 tetrahedron. Corner sharing in the phosphates anion structure between adjacent tetrahedra by sharing a corner oxygen ion lea ds to complex anion structures in these materials.

PAGE 161

161 Table A 1. Structural classification of the silicates. Type O/Si ratio Archetype Neosilicates 4 Isolated tetrahedra [SiO 4 ] 4 Sorosilicates 3.5 Paired tetrahedra [Si 2 O 7 ] 6 Cyclosilicates 3 Tetrahedra rings [Si n O 3n ] 2n

PAGE 162

162 Inosilicates 2.75 Ribbon [Si 4n O 11n ] 6n Tectosilicates 2 3D framework [SiO 2 ]

PAGE 163

163 APPENDIX B SYNTHESIS OF LANTHANUM METAPH OSPHATE L a ( PO 3 ) 3 (a) (b) (c) (d) Figure B 1. Segment of PO 4 chain seen along (a) [010] and (b) [001]. Crystal structure of La(PO 3 ) 3 seen along (c) [110] and (d) [001] direction. Lanthanum metaphosphate occurs in the phosphate rich region of the La 2 O 3 P 2 O 5 phase diagram at 75 mol% P 2 O 5 The crystal structure of La(PO 3 ) 3 is shown along the [110] direction and [001] direction in Figure B 1 (c) and (d) where LaO 8 polyhedra

PAGE 164

164 are shown in orange whereas PO 4 tetrahedra are depicted in green. The PO 4 tetrahedra are corner sharing in the anion so as to attain a helical chain like morphology that extends along the [001] direction. The top view and side view of a segment of an individual chain of tetrahedra in helical arrangement are shown in Figure B 1 (a) and (b). The degree of condensation of the phosphate anion in the material can be said to be intermediate between the orthophosphate and the ultraphosphate. Thus it is interesting to investigate the electrical properties of this material from the view point of comparison with the other lanthanum phosphate materials. Figure B 2. XRD pattern for the as synthesized powder of La(PO 3 ) 3 .3H 2 O using aqueou s method. The synthesis of La(PO 3 ) 3 was carried out by two different methods. In the aqueous m ethod the powder was prepared by mixing two aqueous solutions (1 M) of lanthanum nitrate (La(NO 3 ) 3 ) and NaP 3 O 9 The solution was allowed to stir for 2 hrs for

PAGE 165

165 the precipitation reaction to complete. The obtained precipitate was then filtered out of the s olution and washed with deionised water and allowed to dry at 120 o C. To confirm the phase of the samples the theoretical XRD pattern was compared with the experimental powder pattern of the samples. The patterns were found to match well indicating the ob tained powder was phase pure as shown in Figure B 2. Due to precipitation from aqueous solution the formed phase contained water of crystallization and had the chemical formula La(PO 3 ) 3 .3H 2 O. Figure B 3. X ray diffraction pattern for synthesized powd er of anhydrous La(PO 3 ) 3 using a queous method. The second method used for the preparation involved mixing powders of lanthanum oxide (La 2 O 3 ) and di ammonium hydrogen phosphate ((NH 4 ) 2 HPO 4 ) in the appropriate molar ratio. The mixture was then heated to a temperature of 900 o C in a sealed alumina vessel for 3 h. After the completion of the heat treatment the powders

PAGE 166

166 were reground under acetone and the heat treatment was repeated two times to ensure completion of the solid state reaction. To ensure the phas e purity of the samples the experimental powder XRD pattern was compared with the theoretical powder pattern for the same and the patterns were found to match well as shown in Figure B 3. However, since this method involved a high temperature synthesis pr ocedure the powder obtained was anhydrous. Figure B 4 XRD pattern of a sintered sample of La(PO 3 ) 3 as compared to the unsintered powder. There are two main problems with synthesis of consolidated La(PO 3 ) 3 samples by conventional sintering. The f i rst is that when heated to temperatures above 1235 o C the material undergoes a phase transformation to LaPO 4 and a phosphorus rich liquid (refer to Figure 2 4) Moreover when heated to temperatures above 1000 o C the kinetics of the following decomposition reaction become appreciable.

PAGE 167

167 Thus to prepare consolidated samples of La(PO 3 ) 3 spark plasma sintering (SPS) technique was employed which can provide extremely fast heating rates which would minimize the time available for the decomp osition at temperatures above 1000 o C. The sintering procedure was carried out at 1200 o C for 5 min. The powder XRD patterns of the initial powder and sintered sample are compared in Figure B 4. It can be seen that the powder XRD pattern of the sintered s ample is considerably different from that of the initial powder implying that decomposition had taken place during the synthesis process. There is a possibility that the kinetics of the decomposition could be slowed down by using a process such as hot pr essing. This is because the decomposition reaction leads to the formation of a volatile product (P 2 O 5 ). The use of a high pressure and a sealed vessel for the process could in theory be restrict the volatilization and hence decomposition of the material.

PAGE 168

168 APPENDIX C IMPEDANCE CALCULATIO N F OR A RESISTANCE CAPACITANCE (R C ) CIRCUIT An R C circuit is commonly used as the equivalent circuit for interpreting Nyquist plots obtained for ionically conductive materials. The circuit consists of a resista nce and a capacitance in parallel. The values obtained by fitting the data can then be used to calculate the relevant properties of the material by normalizing the material with the dimensions of the samples. The total impedance of an R C circuit is derived as follows: In the Nyquist plot the imaginary part of impedance is plotted as a function of the real part. Hence to obtain the relation between the real and imaginary parts, the input frequency must be eliminated from equations (1) and (2) which is worked out as follows. Using equation (1),

PAGE 169

169 The total impedance can be expressed as a sum of the imag inary and real parts as follows. ----------(1) --------(2) The equations (1) and (2) represent the real an d imaginary parts of impedance as a function of the applied frequency. The above equation represents the equation of a circle whose center is displaced along the x axis by a distance a as shown below. However, since cannot have negative values the imaginary part of impedance can only have negative values (refer to equa tion (2)). Thus the Nyquist plot only shows a semicircle as theoretically varies from 0 to

PAGE 170

170 APPENDIX D STRUCTURAL PARAM ETERS AND ATOMIC POS ITIONS FOR L a PO 4 AND L a P 5 O 14 The atomic positions for a unit cell of lanthanum ultra phos phate are summarized in Table D 1 below. The crystal structure belongs to the P 2 1 /c space group and the lattice parameters are, a = 8.8206 b = 9.1196 c = 13.1714 and = 90.661003 o The structure contains 14 crystallographically distinct oxygen ion sites, 5 phosphorus sites and 1 lanthanum site. The labeling scheme for the oxygen sites is such that the terminal oxygen ions have a single digit in their label while the corner shared oxygen ions have two digits in their label. Thus the oxygen ion that is shared between P1O4 and P2O4 tetrahedra is labeled as O12 and so on. The only exception is O11 which is a terminal oxygen ion. Table D 1. Atomic positions for lanthanum ultraphosphate (LaP 5 O 14 ). Atom Label Fractional coordinates x y z La L a 0.77697 0.6903 0.99852 O O1 0.9168 0 0.619 00 0.1536 0 O O2 0.7172 0 0.9114 0 0.8911 0 O O3 0.0357 0 0.8156 0 0.9929 0 O O4 0.7287 0 0.9071 0 0.1159 0 O O5 0.5831 0 0.6144 0 0.8713 0 O O6 0.7891 0 0.4127 0 0.0016 0 O O7 0.5764 0 0.6208 0 0.1175 0 O O11 0.9138 0 0.6132 0 0.8454 0 O O12 0.9108 0 0.0678 0 0.7956 0 O O14 0.132 0 0.9013 0 0.7755 0 O O23 0.7229 0 0.192 00 0.8997 0 O O25 0.6378 0 0.0801 0 0.7441 0 O O34 0.7025 0 0.1838 0 0.0865 0 O O45 0.5863 0 0.0827 0 0.2336 0 P P1 0.00965 0.9929 0 0.7077 0 P P2 0.74624 0.04856 0.83606 P P3 0.81164 0.25206 0.00034 P P4 0.72357 0.05249 0.16222 P P5 0.49426 0.00245 0.32262

PAGE 171

171 The atomic positions for a unit cell of LaPO 4 are summarized in Table D 2. The crystal structure belongs to the space group P 2 1 /n with lattice parameters, a = 6.831 3 b = 7.0705 c = 6.5034 and = 103.2 7 o The structure has 4 crystallographically different oxygen ions, 1 phosphorus ion and 1 lanthanum ion. Thus each of the four oxygen ions constituting the PO 4 tetrahedra are crystallographically unique. Table D 2. Atomic positions of lantha num orthophosphate (LaPO 4 ). Atom Label Fractional coordinates x y z La La 0.2815 0.1603 0.1007 O O1 0.2503 0.0077 0.4477 O O2 0.3799 0.3315 0.4964 O O3 0.4748 0.1071 0.8018 O O4 0.1277 0.2168 0.7101

PAGE 172

172 LIST OF REFERENCES 1 N. Oreskes, "Beyond the ivory tower The scientific consensus on climate change," Science, 306 [5702] 1686 86 (2004). 2 M. I. Hoffert, "Farewell to Fossil Fuels?," Science, 329 [5997] 1292 94 (2010). 3 M. I. Hoffert, K. Caldeira, G. Benford, D. R. Criswell, C. Green, H. Herzog, A. K. Jain, H. S. Kheshgi, K. S. Lackner, J. S. Lewis, H. D. Lightfoot, W. Manheimer, J. C. Mankins, M. E. Mauel, L. J. Perkins, M. E. Schlesinger, T. Volk, andT. M. L. Wigley, "Advanced technology paths to global climate stability: Ener gy for a greenhouse planet," Science, 298 [5595] 981 87 (2002). 4 M. Z. Jacobson, W. G. Colella, andD. M. Golden, "Cleaning the air and improving health with hydrogen fuel cell vehicles," Science, 308 [5730] 1901 05 (2005). 5 M. G. Schultz, T. Diehl, G. P. B rasseur, andW. Zittel, "Air pollution and climate forcing impacts of a global hydrogen economy," Science, 302 [5645] 624 27 (2003). 6 S. M. Haile, D. A. Boysen, C. R. I. Chisholm, andR. B. Merle, "Solid acids as fuel cell electrolytes," Nature, 410 [6831] 91 0 13 (2001). 7 T. Norby, "The promise of protonics," Nature, 410 [6831] 877 78 (2001). 8 T. Norby, "Solid state protonic conductors: principles, properties, progress and prospects," Solid State Ionics, 125 [1 4] 1 11 (1999). 9 T. E. Springer, T. Rockward, T. A. Zawodzinski, andS. Gottesfeld, "Model for polymer electrolyte fuel cell operation on reformate feed Effects of CO, H 2 dilution, and high fuel utilization," J. Electrochem. Soc., 148 [1] A11 A23 (2001). 10 J. W. Tye, M. B. Hall, andM. Y. Darensbourg, "Better than platinum? Fuel cells energized by enzymes," Proc. Natl. Acad. Sci. U. S. A., 102 [47] 16911 12 (2005). 11 E. Warburg and F. Tegetmeier, Wiedemann Ann. Phys., 32 455 (1888). 12 A. Joffe, "Electricity passage through crystals," Annalen Der Physik 72 [22] 461 500 (1923). 13 C. Tubandt, "Electricity conduction in solid crystalised compounds Second Announcement Conveyance and ambulation of ions in consistent solid electrolytes," Z. Anorg. Allg. Chem., 115 [1/2] 105 26 (1921). 14 K. Kiukkola and C. Wagner, "GALVANIC CELLS FOR THE DETERMINATION OF THE STANDARD MOLAR FREE ENERGY OF FORMATION OF METAL HALIDES, OXIDES, AND SULFIDES AT ELEVATED TEMPERATURES," J. Electrochem. Soc., 104 [5] 308 16 (1957).

PAGE 173

173 15 J. T. Kummer and N. Weber, "SODIUM SULFUR SECONDA RY BATTERY," Sae Transactions, 76 88 & (1968). 16 H. Iwahara, T. Esaka, H. Uchida, andN. Maeda, "Proton Conduction in Sintered Oxides and Its Application to Steam Electrolysis for Hydrogen Production," Solid State Ionics, 3 4 [Aug] 359 63 (1981). 17 N. N. C hudinova, A. V. Lavrov, andI. V. Tananaev, "Reaction of bismuth oxide with phosphoric acid during heating," Izvestiya Akademii Nauk SSSR, Neorganicheskie Materialy, 8 [11] 1971 76 (1972). 18 A. Durif, "Crystal Chemistry of Condensed Phosphates. Plenum Pres s: New York, (1995). 19 H. D. Park and E. R. Kreidler, "Phase equilibria in the system La 2 O 3 P 2 O 5 ," Journal of the American Ceramic Soceity, 67 [1] 23 26 (1984). 20 H. Y. P. Hong, "Crystal Structures of Neodymium Metaphosphate (NdP 3 O 9 ) and Ultraphosphate ( NdP 5 O 14 )," Acta Crystallographica B, B30 468 (1974). 21 D. W. Kingery, H. K. Bowen, andD. R. Uhlmann, "Introduction to ceramics. Wiley Interscience, (1976). 22 E. Barsoukov and J. R. Macdonald, "Impedance Spectroscopy: Theory, Experiment and Applications 2nd ed. John Wiley and Sons, (2005). 23 J. E. Bauerle, "Study of solid electrolyte polarization by a complex admittance method," Journal of Physics and Chemistry of Solids, 30 [12] 2657 & (1969). 24 H. Rickert, "Electrochemistry of Solids An Introduction Springer Verlag Berlin Heidelberg: New York, (1982). 25 J. M. Cole, M. R. Lees, J. A. K. Howard, R. J. Newport, G. A. Saunders, andE. Schonherr, "Crystal structures and magnetic properties of rare earth ultraphosphates, RP5O14 (R = La, Nd, Sm, Eu, Gd)," Journal of Solid State Chemistry, 150 [2] 377 82 (2000). 26 A. K. Cheetham and P. Day, "Solid State Chemistry: Techniques. Clarendon Press: Oxford, (1987). 27 P. M. Panchmatia, A. Orera, E. Kendrick, J. V. Hanna, M. E. Smith, P. R. Slater, andM. S. Islam, Protonic defects and water incorporation in Si and Ge based apatite ionic conductors," J. Mater. Chem., 20 [14] 2766 72 (2010). 28 B. G. Dick and A. W. Overhauser, "THEORY OF THE DIELECTRIC CONSTANTS OF ALKALI HALIDE CRYSTALS," Physical Review, 112 [1] 90 10 3 (1958).

PAGE 174

174 29 C. A. J. Fisher, V. M. H. Prieto, andM. S. Islam, "Lithium battery materials LiMPO4 (M = Mn, Fe, Co, and Ni): Insights into defect association, transport mechanisms, and doping behavior," Chem. Mater., 20 [18] 5907 15 (2008). 30 M. F. Mott and M. J. Littleton, "Conduction in polar crystals. I. Electrolytic conduction in solid salts," Transactions of the Faraday Society, 34 [1] 0485 99 (1938). 31 A. N. Cormack, P. Saul, andC. R. A. Catlow, "MODELING HYDROGEN IN OXIDES," Physica B & C, 131 [1 3] 139 44 (1985). 32 P. Saul, C. R. A. Catlow, andJ. Kendrick, "Theoretical Studies of Protons in Sodium Hydroxide," Philosophical Magazine B Physics of Condensed Matter Statistical Mechanics Electronic Optical and Magnetic Properties, 51 [2] 107 17 (1985). 33 M. Cherry, M. S. Islam, J. D. Gale, andC. R. A. Catlow, "Computational Studies of Protons in Perovskite Structured Oxides," Journal of Physical Chemistry, 99 [40] 14614 18 (1995). 34 A. Hamady and T. Jouini, "A new ultraphosphate, CaYP 7 O 20 preparation and c rystal structure," Journal of Solid State Chemistry, 111 [2] 443 46 (1994). 35 I. Tordjman, M. Bagieubeucher, andR. Zilber, "Crystal structure of calcium ultraphosphate CaP 4 O 11 ," Zeitschrift Fur Kristallographie, 140 [3 4] 145 53 (1974). 36 G. Gozel, M. Kizilyalli, andR. Kniep, "Characterization of a new calcium ultraphosphate, Ca 3 (P 5 O 14 ) 2 ," Journal of Solid State Chemistry, 129 [2] 196 99 (1997). 37 D. Stachel, H. Paulus, I. Svoboda, andH. Fuess, "Crystal structure of calcium ultraphosphate, Ca 2 P 6 O 17 ," Z eitschrift Fur Kristallographie, 202 [1 2] 117 18 (1992). 38 M. Funke, M. Blum, R. Glaum, andB. ElBali, "Synthesis, crystal structures, and spectroscopic characterization of NiP 4 O 11 and CaNiP 2 O 7 ," Zeitschrift Fur Anorganische Und Allgemeine Chemie, 630 [7] 1 040 47 (2004). 39 N. N. Chudinova, K. K. Palkina, N. B. Karmanovskaia, S. I. Maximova, andN. T. Chibiskova, "Double ultraphosphate of iron and sodium Na 3 FeP 8 O 23 having a new anion shape," Doklady Akademii Nauk Sssr, 306 [3] 635 38 (1989). 40 N. N. Chudinova E. V. Murashova, andN. B. Karmanovskaya, "NA 3 MP 8 O 23 binary ultraphosphate where (M = Cr, V, Mn, Al, Ga) synthesis and crystalline structure," Zhurnal Neorganicheskoi Khimii, 39 [11] 1769 72 (1994). 41 C. Baezdoelle, D. Stachel, I. Svoboda, andH. Fuess, "Crystal structure of zinc ultraphosphate, ZnP 4 O 11 ," Zeitschrift Fur Kristallographie, 203 282 83 (1993).

PAGE 175

175 42 A. Olbertz and D. Stachel, "Nickel ultraphosphate, NiP 4 O 11 ," Acta Crystallographica Section C Crystal Structure Communications, 51 1047 49 (1995) 43 R. Glaum, M. Weil, andD. Ozalp, "Studies on the thermal behavior and crystal chemistry of anhydrous phosphates .15. Synthesis, crystal structure, and properties of copper(II) ultraphosphate CuP 4 O 11 ," Zeitschrift Fur Anorganische Und Allgemeine Chemie, 622 [11] 1839 46 (1996). 44 Y. B. Kim, S. K. Lee, K. H. Kim, andI. J. Yoo, "Effect of the addition of lysosome and sodium ultraphosphate on the shelf life of press ham," Korean Journal of Animal Science, 33 [2] 176 84 (1991). 45 D. Stachel, H. Paulus, C. Gu enter, andH. Fuess, "Crystal structure of magnesium ultraphosphate, MgP 4 O 11 ," Zeitschrift Fur Kristallographie, 199 [3 4] 275 76 (1992). 46 A. S. Antsyshkina, M. A. Poraikoshits, L. K. Minacheva, A. V. Lavrov, andV. G. Ivanova, "Structure of stronitum and cadmium ultraphosphate crystals of composition M 2 IIP 6 O 17 ," Doklady Akademii Nauk Sssr, 229 [4] 896 97 (1976). 47 A. Olbertz, D. Stachel, I. Svoboda, andH. Fuess, "Crystal structure of manganese ultraphosphate, MnP 4 O 11 ," Zeitschrift Fur Kristallographie, 20 9 [11] 899 99 (1994). 48 F. Chehimi Moumen and M. Ferid, "DyP 5 O 14 ," Acta Crystallographica Section E Structure Reports Online, 63 i89 i91 (2007). 49 D. Tranqui, M. Bagieu, andA. Durif, "Crystalline structure of ultraphosphate of samarium, SmP 5 O 14 ," Acta Cry stallographica Section B Structural Science, B 30 [JUL15] 1751 55 (1974). 50 A. R. Grimmer and N. N. Cudinova, "Nuclear magnetic resonance studies on bismuth phosphates detection of crystal ultraphosphate," Zeitschrift Fur Chemie, 12 [4] 149 & (1972). 51 A Mbarek, A. Oudahmane, M. El Ghozzi, andD. Avignant, "An orthorhombic polymorph of the ultraphosphate YP 5 O 14 ," Acta Crystallographica Section E Structure Reports Online, 65 I22 U104 (2009). 52 I. Szczygiel and T. Znamierowska, "The synthesis and therman s tability of cerium ultraphosphate CeP 5 O 14 ," Journal of Thermal Analysis, 36 [6] 2195 98 (1990). 53 K. Horchani Naifer and M. Ferid, "Lutetium ultraphosphate," Acta Crystallographica Section E Structure Reports Online, 64 I33 U64 (2008). 54 A. Olbertz, D. St achel, I. Svoboda, andH. Fuess, "NiHP 5 O 14 ultraphosphate," Acta Crystallographica Section C Crystal Structure Communications, 52 1603 05 (1996).

PAGE 176

176 55 H. Y. P. Hong and J. W. Pierce, "Crystal structure of ytterbium ultraphosphate, YbP 5 O 14 ," Materials Researc h Bulletin, 9 [2] 179 89 (1974). 56 H. Y. P. Hong, "CRYSTAL STRUCTURES OF NEODYMIUM METAPHOSPHATE (NDP3O9) AND ULTRAPHOSPHATE (NDP5O14)," Acta Crystallographica Section B Structural Science, B 30 [FEB15] 468 74 (1974). 57 I. F. Balashov, V. A. Berenberg, A. V. Lavrov, V. A. Pismennyi, V. S. Terpugov, andN. N. Chudinova, "Synthesis and generation characteristics of crystals of neodymium lanthanum ultraphosphate," Inorganic Materials, 15 [1] 132 33 (1979). 58 Danielme.Hg and H. P. Weber, "Fluorescence in neodymi um ultraphosphate," Ieee Journal of Quantum Electronics, QE 8 [10] 805 & (1972). 59 R. M. Lee, P. A. Hartman, D. G. Olson, andF. D. Williams, "Bactericidal and bacteriolytic effects of selected food grade phosphates, using staphylococcus aureus as a model system," Journal of Food Protection, 57 [4] 276 83 (1994). 60 S. K. Lee, S. P. Hong, Y. B. Kim, S. Y. Choi, andI. J. Yoo, "Effect of lysozyme and sodium ultraphosphate on the lysis of lactobacillus plantarum," Korean Journal of Animal Science, 33 [1] 73 77 ( 1991). 61 S. K. Lee, K. H. Kim, K. S. Kim, Y. B. Kim, andI. J. Yoo, "Effect of lysozyme and sodium ultraphosphate on the shelf life of vienna sausage," Korean Journal of Animal Science, 33 [1] 78 84 (1991). 62 T. Norby and N. Christiansen, "Proton Conductio n in Ca Substituted and Sr Substituted LaPO 4 ," Solid State Ionics, 77 240 43 (1995). 63 G. J. Zhang, E. D. Wachsman, J. C. Nino, S. J. Song, andJ. M. Rhodes, "Proton conducting material Ba 3 Ce(PO 4 ) 3 synthesized by coprecipitation," J. Electrochem. Soc., 154 [7] H566 H71 (2007). 64 M. Nagao, A. Takeuchi, P. Heo, T. Hibino, M. Sano, andA. Tomitab, "A proton conducting In 3+ doped SnP 2 O 7 electrolyte for intermediate temperature fuel cells," Electrochemical and Solid State Letters, 9 [3] A105 A09 (2006). 65 I. C. M arcu, J. M. M. Millet, andJ. M. Herrmann, "Semiconductive and redox properties of Ti and Zr pyrophosphate catalysts (TiP 2 O 7 and ZrP 2 O 7 ). Consequences for the oxidative dehydrogenation of n butane," Catalysis Letters, 78 [1 4] 273 79 (2002). 66 I. Hammas, K. Horchani Naifer, andM. Ferid, "Conduction properties of condensed lanthanum phosphates: La(PO3)(3) and LaP5O14," Journal of Rare Earths, 28 [3] 321 28 (2010).

PAGE 177

177 67 J. M. Cole, M. R. Lees, J. A. K. Howared, R. J. Newport, G. A. Saunders, andE. Schonherr, "Crystal Structures and Magnetic Properties of Rare Earth Ultraphosphates, RP 5 O 14 (R = La, Nd, Sm, Eu, Gd)," J. Solid State Chem., 150 377 82 (2000). 68 I. Hammas, K. Horchani Naifer, andM. Ferid, "Conduction properties of condensed lanthanum phosphates: La(PO 3 ) 3 and LaP 5 O 14 ," J Rare Earth, 28 [3] 321 28 (2010). 69 K. Amezawa, H. Maekawa, Y. Tomii, andN. Yamamoto, "Protonic conduction and defect structures in Sr doped L aPO4," Solid State Ionics, 145 [1 4] 233 40 (2001). 70 E. Fabbri, T. K. Oh, S. Licoccia, E. Traversa, andE. D. Wachsman, "Mixed Protonic/Electronic Conductor Cathodes for Intermediate Temperature SOFCs Based on Proton Conducting Electrolytes," J. Electroche m. Soc., 156 [1] B38 B45 (2009). 71 S. Gallini, M. Hansel, T. Norby, M. T. Colomer, andJ. R. Jurado, "Impedance spectroscopy and proton transport number measurements on Sr substituted LaPO 4 prepared by combustion synthesis," Solid State Ionics, 162 167 73 ( 2003). 72 A. Jouini, M. Ferid, J. C. Gacon, L. Grosvalet, A. Thozet, andM. Trabelsi Ayadi, "Crystal structure and optical study of praseodymium polyphosphate Pr(PO 3 ) 3 ," Materials Research Bulletin, 38 [11 12] 1613 22 (2003). 73 E. V. Murashova and N. N. Chu dinova, "Synthesis and crystal structure of the double polyphosphate CsMn(PO 3 ) 4 ," Inorganic Materials, 36 [12] 1277 81 (2000). 74 J. F. Shin, L. Hussey, A. Orera, andP. R. Slater, "Enhancement of the conductivity of Ba 2 In 2 O 5 through phosphate doping," Chemi cal Communications, 46 [25] 4613 15 (2010). 75 R. Hui, R. Maric, C. Deces Petit, E. Styles, W. Qu, X. G. Zhang, J. Roller, S. Yick, D. Ghosh, K. Sakata, andM. Kenji, "Proton conduction in ceria doped Ba 2 In 2 O 5 nanocrystalline ceramic at low temperature," Jou rnal of Power Sources, 161 [1] 40 46 (2006). 76 R. Maric, C. Deces Petit, R. Hui, X. Zhang, D. Ghosh, K. Sakata, andM. Kenji, "Preparation and characterization of nanocrystalline Ba 2 In 2 x M x O 5 d (M = Ce, Zr)," Journal of the Electrochemical Society, 153 [8] A 1505 A10 (2006). 77 M. S. Khan, M. S. Islam, andD. R. Bates, "Dopant Substitution and Ion Migration in the LaGaO 3 Based Oxygen Ion Conductor," J. Phys. Chem. B, 102 3099 14 (1998). 78 L. Q. Wang, X. D. Zhou, G. J. Exarhos, L. R. Pederson, C. Wang, andC. F. Windisch, "Proton dynamics in ZnO nanorods quantified by in situ solid state H 1 nuclear magnetic resonance spectroscopy," Appl. Phys. Lett., 91 [17] (2007).

PAGE 178

178 79 R. Q. Fu, Z. R. Ma, andJ. P. Zheng, "Proton NMR and dynamic studies of hydrous ruthenium ox ide," J. Phys. Chem. B, 106 [14] 3592 96 (2002). 80 Y. Aihara, A. Sonai, M. Hattori, andK. Hayamizu, "Ion conduction mechanisms and thermal properties of hydrated and anhydrous phosphoric acids studied with H 1, H 2, and P 31 NMR," Journal of Physical Chemi stry B, 110 [49] 24999 5006 (2006). 81 T. Dippel, K. D. Kreuer, J. C. Lassegues, andD. Rodriguez, "Proton Conductivity in Fused Phosphoric Acid a H 1 P 31 PFG NMR and QNS Study," Solid State Ionics, 61 [1 3] 41 46 (1993). 82 M. Fetchtelkord, "Proton Dynami cs in Letovicite, (NH 4 ) 3 H(SO 4 ) 2 : A 1 H and 14 N NMR Spectroscopic Study," Solid State Nucl. Magn. Reson., 17 76 88 (2000). 83 Y. Yoshida, K. Hisada, Y. Matsuo, andS. Ikehata, "NMR measurements on proton conductor CsHSeO 4 below T(c)e=401 K," Solid State Ionic s, 178 [37 38] 1869 71 (2008). 84 A. R. Lim, "A study of the phase transitions and proton dynamics of the superprotonic conductor Cs 5 H 3 (SO 4 ) 4 0.5H 2 O single crystal with H 1 and Cs 133 nuclear magnetic resonance," J. Solid State Chem., 181 [4] 796 801 (2008 ). 85 J. Kanellopoulos, C. Gottert, D. Schneider, B. Knorr, D. Prager, H. Ernst, andD. Freude, "NMR investigation of proton mobility in zeolites," J. Catal., 255 [1] 68 78 (2008). 86 A. Ishikawa, H. Maekawa, T. Yamamura, Y. Kawakita, K. Shibata, andM. Kawai "Proton dynamics of CsH 2 PO 4 studied by quasi elastic neutron scattering and PFG NMR," Solid State Ionics, 179 [40] 2345 49 (2008). 87 L. S. Cahill, U. A. Rana, M. Forsyth, andM. E. Smith, "Investigation of proton dynamics and the proton transport pathway in choline dihydrogen phosphate using solid state NMR," Physical Chemistry Chemical Physics, 12 [20] 5431 38 (2010). 88 G. J. Zhang, R. Yu, S. Vyas, J. Stettler, J. A. Reimer, G. Harley, andL. C. De Jonghe, "Proton conduction and characterization of an La(P O 3 ) 3 Ca(PO 3 ) 2 glass cerarnic," Solid State Ionics, 178 [35 36] 1811 16 (2008). 89 T. Scherban and A. S. Nowick, "Protonic Conduction in Fe Doped KTaO 3 Crystals," Solid State Ionics, 53 6 1004 08 (1992). 90 M. Pionke, T. Mono, W. Schweika, H. Schober, andT. Springer, "Proton diffusion in Ba[Ca (1+x/3) Nb (2 x/3) ]O 3 x/2 studied by quasielastic neutron scattering," Physica B, 234 95 96 (1997).

PAGE 179

179 91 N. Kitamura, K. Amezawa, Y. Tomii, andN. Yamamoto, "Protonic conduction in rare earth orthophosphates with the monazite structure," Solid State Ionics, 162 161 65 (2003). 92 Y. X. Ni, J. M. Hughes, andA. N. Mariano, "CRYSTAL CHEMISTRY OF THE MONAZITE AND XENOTIME STRUCTURES," Am. Mineral., 80 [1 2] 21 26 (1995). 93 M. W. Barsoum, "Fundamentals of ceramics. Taylor and Franc is, Inc, (2002). 94 E. Matsushita and T. Sasaki, "Theoretical approach for protonic conduction in perovskite type oxides," Solid State Ionics, 125 [1 4] 31 37 (1999). 95 K. D. Kreuer, "Proton conducting oxides," Annu. Rev. Mater. Res., 33 333 59 (2003). 96 T Norby and Y. Larring, "Concentration and transport of protons in oxides," Current Opinion in Solid State & Materials Science, 2 [5] 593 99 (1997). 97 M. Pionke, T. Mono, W. Schweika, T. Springer, andH. Schober, "Investigation of the hydrogen mobility in a mixed perovskite: Ba[Ca (1+x/3) Nb (2 x/3) ]O 3 x/2 by quasielastic neutron scattering," Solid State Ionics, 97 [1 4] 497 504 (1997). 98 E. Matsushita, "Tunneling mechanism on proton conduction in perovskite oxides," Solid State Ionics, 145 [1 4] 445 50 (2001). 99 Q. F. Zhang, G. Wahnstrom, M. E. Bjorketun, S. W. Gao, andE. G. Wang, "Path Integral Treatment of Proton Transport Processes in BaZrO3," Physical Review Letters, 101 [21] (2008). 100 F. Shimojo, K. Hoshino, andH. Okazaki, "Stable positions and probable migration paths of a proton in Y doped SrCeO 3 : an ab initio molecular dynamics simulation," J. Phys. Soc. Jpn., 67 [6] 2008 11 (1998). 101 R. Hempelmann and C. Karmonik, "Proton diffusion in proton conducting oxides," Phase Transitions, 58 [1 3] 175 84 (199 6). 102 T. Matzke, U. Stimming, C. Karmonik, M. Soetratmo, R. Hempelmann, andF. Guthoff, "Quasielastic thermal neutron scattering experiment on the proton conductor SrCe 0.95 Yb 0.05 H 0.02 O 2.985 ," Solid State Ionics, 86 8 621 28 (1996). 103 K. Amezawa, Y. Nish ikawa, Y. Tomii, andN. Yamamoto, "Electrical and mechanical properties of Sr doped LaPO 4 prepared by spark plasma sintering," J. Electrochem. Soc., 152 [6] A1060 A67 (2005). 104 H. H. Huang, M. Ishigame, andS. Shin, "Protonic Conduction in the Single Crysta ls of Y Doped SrZrO 3 ," Solid State Ionics, 47 [3 4] 251 55 (1991).

PAGE 180

180 105 S. Shin, H. H. Huang, M. Ishigame, andH. Iwahara, "Protonic Conduction in the Single Crystals of SrZrO 3 and SrCeO 3 Doped with Y 2 O 3 ," Solid State Ionics, 40 1 910 13 (1990). 106 Y. Larrin g and T. Norby, "Protons in LaErO 3 ," Solid State Ionics, 70 305 10 (1994). 107 M. S. Islam and M. Cherry, "Protons in LaMO3: Atomistic modelling and ab initio studies," Solid State Ionics, 97 [1 4] 33 37 (1997). 108 E. Ruiz Trejo and R. A. De Souza, "Dopant substitution and oxygen migration in the complex perovskite oxide Ba 3 CaNb 2 O 9 : A computational study," J. Solid State Chem., 178 [6] 1959 67 (2005). 109 F. Tyholdt, J. A. Horst, S. Jorgensen, T. Ostvold, andT. Norby, "Segregation of Sr in Sr doped LaPO 4 cer amics," Surf Interface Anal, 30 [1] 95 97 (2000). 110 R. Glockner, M. S. Islam, andT. Norby, "Protons and other defects in BaCeO 3 : a computational study," Solid State Ionics, 122 [1 4] 145 56 (1999). 111 S. J. Stokes and M. S. Islam, "Defect chemistry and pr oton dopant association in BaZrO3 and BaPrO3," Journal of Materials Chemistry, 20 [30] 6258 64 (2010). 112 R. Yu and L. C. De Jonghe, "Proton transfer mechanism in LaPO4," J Phys Chem C, 111 [29] 11003 07 (2007). 113 G. Z. Liu, H. X. Zhang, Z. E. Lin, S. T. Zheng, J. Zhang, J. T. Zhao, G. M. Wang, andG. Y. Yang, "Germanates of 1D chains, 2D layers, and 3D frameworks built from Ge O clusters by using metal complex templates: Host guest symmetry and chirality transfer," Chem. Asian J., 2 [10] 1230 39 (2007). 11 4 A. Jouini, M. Ferid, J. C. Gacon, L. Grosvalet, A. Thozet, andM. Trabelsi Ayadi, "Crystal structure and optical study of praseodymium polyphosphate Pr(PO3)(3)," Materials Research Bulletin, 38 [11 12] 1613 22 (2003). 115 C. V. Ramana, A. Ait Salah, S. Utsu nomiya, A. Mauger, F. Gendron, andC. A. Julien, "Novel lithium iron pyrophosphate (LiFe1.5P2O7) as a positive electrode for li ion batteries," Chemistry of Materials, 19 5319 24 (2007). 116 B. Boonchom and N. Vittayakorn, "Synthesis and ferromagnetic prope rty of new binary copper iron pyrophosphate CuFeP2O7," Mater. Lett., 64 [3] 275 77 (2010). 117 M. Ciopec, C. Muntean, A. Negrea, L. Lupa, P. Negrea, andP. Barvinschi, "Synthesis and thermal behavior of double copper and potassium pyrophosphate," Thermochim. Acta, 488 [1 2] 10 16 (2009).

PAGE 181

181 118 M. Havecker, R. W. Mayer, A. Knop Gericke, H. Bluhm, E. Kleimenov, A. Liskowski, D. Su, R. Follath, F. G. Requejo, D. F. Ogletree, M. Salmeron, J. A. Lopez Sanchez, J. K. Bartley, G. J. Hutchings, andR. Schlogl, "In situ investigation of the nature of the active surface of a vanadyl pyrophosphate catalyst during n butane oxidation to maleic anhydride," Journal of Physical Chemistry B, 107 [19] 4587 96 (2003). 119 G. Alberti, M. Casciola, S. Cavalaglio, andR. Vivani, "Proton conductivity of mesoporous zirconium phosphate pyrophosphate," Solid State Ionics, 125 [1 4] 91 97 (1999). 120 G. Alberti, M. Casciola, U. Costantino, A. Peraio, andT. Rega, "PROTON CONDUCTING SOLID DISPERSIONS OF SILICA AND ZIRCONIUM PHOSPHATE PYROPHOSPHA TE," Journal of Materials Chemistry, 5 [11] 1809 12 (1995). 121 G. Alberti, M. Casciola, E. D'Alessandro, andM. Pica, "Preparation and proton conductivity of composite ionomeric membranes obtained from gels of amorphous zirconium phosphate sulfophenylenphos phonates in organic solvents," Journal of Materials Chemistry, 14 [12] 1910 14 (2004). 122 T. Matsui, T. Kukino, R. Kikuchi, andK. Eguchi, "An intermediate temperature proton conducting electrolyte based on a CsH2PO4/SiP2O7 composite," Electrochemical and S olid State Letters, 8 [5] A256 A58 (2005). 123 T. Matsui, T. Kukino, R. Kikuchi, andK. Eguchi, "Composite effects of silicon pyrophosphate as a supporting matrix for CsH5(PO4)(2) electrolytes at intermediate temperatures," Electrochimica Acta, 51 [18] 3719 2 3 (2006). 124 H. Muroyama, K. Kudo, T. Matsui, R. Kikuchi, andK. Eguchi, "Electrochemical properties of MH2PO4/SiP2O7 based electrolytes (M=alkaline metal) for use in intermediate temperature fuel cells," Solid State Ionics, 178 [27 28] 1512 16 (2007). 125 F. Chehimi Moumen, P. Llewellyn, F. Rouquerol, G. Vacquier, D. B. Hassen Chehimi, M. Ferid, andM. Trabelsi Ayadi, "Constant transformation rate thermal analysis of HGdP2O7 center dot 3H(2)O," Journal of Thermal Analysis and Calorimetry, 82 [3] 783 89 (2005) 126 E. R. Losilla, A. Cabeza, S. Bruque, M. A. G. Aranda, J. Sanz, J. E. Iglesias, andJ. A. Alonso, "Syntheses, structures, and thermal expansion of germanium pyrophosphates," J. Solid State Chem., 156 [1] 213 19 (2001). 127 V. V. Lisnyak, N. V. Stus, N. S. Slobodyanik, N. M. Belyavina, andV. Y. Markiv, "Crystal structure of a novel cubic pyrophosphate WP2O7," J. Alloys Compd., 309 [1 2] 83 87 (2000).

PAGE 182

182 128 G. R. Levi and G. Peyronel, "Crystallographic structure of the isomorphic group (Si4+, Ti4+, Zr4+, Sn4+ Hf4+)P2O7," Z. Kristallogr., 92 [3/4] 190 209 (1935). 129 R. K. B. Gover, N. D. Withers, S. Allen, R. L. Withers, andJ. S. O. Evans, "Structure and phase transitions of SnP2O7," Journal of Solid State Chemistry, 166 [1] 42 48 (2002). 130 F. Fayon, I. J. Ki ng, R. K. Harris, R. K. B. Gover, J. S. O. Evans, andD. Massiot, "Characterization of th Room Temperature Structure of SnP 2 O 7 by 31 P Through Space and Through Bond NMR correlation Spectroscopy," Chem. Mater., 15 2234 39 (2003). 131 D. M. Poojary, R. B. Bor ade, F. L. Campbell, andA. Clearfield, "Crystal Structure of Silicon Pyrophosphate (Form I) from Powder Diffraction Data," J. Solid State Chem., 112 [1] 106 12 (1994). 132 K. Palkina and K. H. Jost, "Crystal Structure of the Polyphosphate [BiH(PO 3 ) 4 ] x ," Act a Crystallographica, B31 2285 (1975). 133 J. Amami, M. Ferid, andM. Trabelsi Ayedi, "Crystal structure and spectroscopic studies of NaGd(PO3)(4)," Materials Research Bulletin, 40 [12] 2144 52 (2005). 134 M. Bagieu Beucher and M. T. Averbuch Pouchot, "Crysta l data and crustal structure of bismuth tetraphosphate: Bi 2 P 4 O 13 ," Z. Kristallogr., 180 165 70 (1987). 135 A. Pramanick, S. Omar, J. C. Nino, andJ. L. Jones, "Lattice parameter determination using a curved position sensitive detector in reflection geometry and application to Sm Nd x/2 Ce x/2(1 x) O2 delta ceramics," J. Appl. Crystallogr., 42 490 95 (2009). 136 H. Onoda, Y. Inagaki, A. Kuwabara, N. Kitamura, K. Amezawa, A. Nakahira, andI. Tanaka, "Synthesis and electrical conductivity of bulk tetra valent ce rium pyrophosphate," Journal of Ceramic Processing Research, 11 [3] 344 47 (2010). 137 T. Oh, "Determination of Optimal Eu Concentration in SrCeO 3 Mixed Conducting Ceramic Membrane for Maximum Hydrogen Production," Ph.D. Dissertation, University of Florida 47 (2008). 138 S. Tao, "Conductivity of SnP 2 O 7 and In doped SnP 2 O 7 prepared by and aqueous solution method," Solid State Ionics, 180 148 53 (2009). 139 M. Nagao, T. Kamiya, P. Heo, A. Tomita, T. Hibino, andM. Sano, "Proton conduction in In 3+ doped SnP 2 O 7 a t intermediate temperatures," J. Electrochem. Soc., 153 [8] A1604 A09 (2006). 140 N. N. Greenwood and A. Thompson, "The mechanism of electrical conduction in fused phosphoric acid and trideuterophosphoric acids," Journal of Chemical Society 3485 (1959).

PAGE 183

183 14 1 R. D. Shannon, "Revised Effective Ionic Radii and Systematic Studies of Interatomic Distances in Halides and Chalcogenides," Acta Crystallographica, A32 751 (1976). 142 S. Natarajan and S. Mandal, "Open framework structures of transition metal compounds," Angew. Chem. Int. Edit., 47 [26] 4798 828 (2008). 143 O. A. Gurbanova and E. L. Belokoneva, "Topology symmetry analysis of the mixed frameworks in the structures of phosphates, germanates, gallates, and silicates based on fundamental building blocks," Crys tallogr. Rep., 51 [4] 577 83 (2006). 144 K. H. Breuer, W. Eysel, andM. Behruzi, "COPPER(II) SILICATES AND GERMANATES WITH CHAIN STRUCTURES .2. CRYSTAL CHEMISTRY," Z. Kristall., 176 [3 4] 219 32 (1986). 145 G. D. Ilyushin and L. N. Dem'yanets, "Crystal chemis try of germanates: Characteristic structural features of Li,Ge germanates," Crystallogr. Rep., 45 [4] 626 32 (2000). 146 H. Fjeld, K. Toyoura, R. Haugsrud, andT. Norby, "Proton mobility through a second order phase transition: theoretical and experimental s tudy of LaNbO4," Physical Chemistry Chemical Physics, 12 [35] 10313 19 (2010). 147 R. Haugsrud, B. Ballesteros, M. Lira Cantu, andT. Norby, "Ionic and electronic conductivity of 5% Ca doped GdNbO4," Journal of the Electrochemical Society, 153 [8] J87 J90 (20 06). 148 R. Haugsrud and T. Norby, "Proton conduction in rare earth ortho niobates and ortho tantalates," Nature Materials, 5 [3] 193 96 (2006). 149 R. Haugsrud and T. Norby, "High temperature proton conductivity in acceptor doped LaNbO4," Solid State Ionic s, 177 [13 14] 1129 35 (2006). 150 A. Orera and P. R. Slater, "Water incorporation studies in apatite type rare earth silicates/germanates," Solid State Ionics, 181 [3 4] 110 14 (2010). 151 C. B. Carter and M. G. Norton, "Ceramic Materials Science and Engine ering. Springer Science Business Media: New York, (2007). 152 E. D. Kolb and R. A. Laudise, "PRESSURE VOLUME TEMPERATURE BEHAVIOR IN THE SYSTEM H2O H3PO4 ALPO4 AND ITS RELATIONSHIP TO THE HYDROTHERMAL GROWTH OF ALPO4," Journal of Crystal Growth, 56 [1] 83 9 2 (1982).

PAGE 184

184 BIOGRAPHICAL SKETCH Satyajit R. Phadke was born in 1984 in Nagpur which is located in central India. He grew up mostly in the historic town of Lucknow which is located in northern India in the foothills of great Himalayas and is famous for the great cultura l heritage as well as K He earned his B.Tech in m etallurgical a nd m aterials e ngineering from the Indian Institute of Technology (IIT) Roorkee which was the first engineering college of India established in the year 1847. During his stay at IIT Roorkee i n addition to academics he was also able to pursue several of his other interests like playing cricket and tennis as well as hone his artistic skills by working as the designe magazine. He then joined the highly regarded Materials Science and Engineering Department at the University of Florida to pursu e further studies. Here, he was introduced to the rapidly emerging and exciting field of fuel c ells and went on to pursue a Ph.D. in this field.